The enzymes of β-lactam biosynthesis

Contents

• 1 Introduction

• 2 Naturally-occurring β-lactams

• 3 Overview of β-lactam biosynthesis

• 4 Penicillin and cephalosporin biosynthesis pathways
  • 4.1 Introduction – overview of the pathways
  • 4.2 Biosynthesis of l-α-aminoadipic acid
  • 4.3 δ-(l-α-Aminoadipoyl)-l-cysteinyl-d-valine synthetase (ACVS)
  • 4.4 Isopenicillin N synthase
  • 4.5 Biosynthesis of the hydrophobic penicillins – isopenicillin amidohydrolase/acyltransferase
  • 4.6 Isopenicillin N/penicillin N epimerase
  • 4.7 Deacetoxycephalosporin C synthase (DAOCS) and deacetylcephalosporin C synthase (DACS)
  • 4.8 Introduction of the 7α-methoxy and formamido-groups
  • 4.9 Modifications at the cephem C-3′ position
  • 4.10 Production of semisynthetic penicillins and cephalosporins
  • 4.11 Protein and metabolic engineering studies on the cephalosporin/cephamycin biosynthesis pathway

• 5 Clavulanic acid biosynthesis
  • 5.1 Carboxyethylarginine synthase
  • 5.2 β-Lactam synthetases in clavam and carbapenem biosynthesis
  • 5.3 The trifunctional role of CAS in clavaminic acid biosynthesis
  • 5.4 Proclavaminic acid amidino hydrolase (PAH)
  • 5.5 Clavaminic acid, a branch point in (2R,5R) clavulanic acid and (3S,5S) clavam biosynthesis
  • 5.6 From clavaminic acid to clavaldehyde: Stereoisomerisation and oxidative deamination of clavaminic acid
  • 5.7 N-Glycyl-clavaminic acid synthetase, GCAS (Orf17)
  • 5.8 From clavaldehyde to clavulanic acid: clavulanic acid dehydrogenase
  • 5.9 The oligopeptide binding proteins (Orf7 and Orf15) of clavulanic acid biosynthesis
  • 5.10 Orf13 and Orf14 of clavulanic acid biosynthesis
  • 5.11 Role of ornithine acetyl transferase in clavulanic acid biosynthesis
  • 5.12 Summary of genetic studies on clavulanic acid biosynthesis

• 6 Carbapenem biosynthesis
  • 6.1 Subfamilies of naturally-occurring carbapenems
  • 6.2 Overview of carbapen-2-em-3-carboxylate biosynthesis
  • 6.3 The role of individual enzymes in C3C biosynthesis
  • 6.4 Thienamycin biosynthesis

• 7 Monocyclic β-lactam biosynthesis
  • 7.1 Monobactam biosynthesis
  • 7.2 Nocardicin biosynthesis
  • 7.3 Tabtoxin biosynthesis

• 8 Summary, conclusions and future prospects
  • 8.1 Mechanisms of structural diversity introduction into β-lactam biosynthesis pathways
  • 8.2 The future of β-lactam biosynthesis studies

• Acknowledgements

• References

---

Abstract

Covering: up to May 2012

The β-lactam antibiotics and related β-lactamase inhibitors are amongst the most important small molecules in clinical use. Most, but not all, β-lactams including penicillins, cephalosporins, and clavulanic acid are produced via fermentation or via modification of fermented intermediates, with important exceptions being the carbapenems and aztreonam. The desire for more efficient routes to existing antibiotics and for access to new and synthetically challenging ones stimulates continued interest in β-lactam biosynthesis. We review knowledge of the pathways leading to β-lactam antibiotics focusing on the mechanisms, structures and biocatalytic applications of the enzymes involved.

---

Refaat B. Hamed

Refaat Hamed got his BSc (in Pharmaceutical Sciences) and MSc (in Natural Products Chemistry) degrees in 1999 and 2004, respectively, from Assiut University, Egypt. In 2010, he finished his D.Phil. in the Chemistry Department, University of Oxford, under Prof. Chris Schofield's supervision. He then continued there as a postdoctoral research fellow. In 2011, he was appointed as a lecturer in Assiut University. His D.Phil. and postdoctoral studies targeted the mechanisms, structures and biocatalytic applications of β-lactam biosynthesis enzymes. Engineering enzymes and biosynthetic pathways redesign for the production of unnatural products of pharmaceutical interest is a main objective for Refaat's research.

J. Ruben Gomez-Castellanos

J. Rubén Gómez Castellanos is a doctoral candidate in the Chemical Biology programme of the Department of Chemistry, University of Oxford, under the supervision of Prof. Chris Schofield. He holds a BSc in Pharmaceutical and Biological Chemistry from La Salle University in Mexico City (2005) and an MSc in Drug Discovery from The School of Pharmacy – University of London (with distinction, 2006). He has previous experience in clinical research working for Eli Lilly and Co. in Mexico City, and is currently working as a Clinical Research Associate II at Pharm-Olam, Intl., an international contract research organisation with offices in Mexico.

Luc Henry

Luc Henry was born in Switzerland and studied biological chemistry in Lausanne (Switzerland), Uppsala (Sweden) and Oxford (UK), earning a M. Sc. in biological chemistry from the Ecole Polytechnique Fédérale de Lausanne (EPFL) in 2007. He then spent four years as a DPhil student in the laboratory of Prof. Christopher J. Schofield at the University of Oxford, working on the enzymology of thienamycin biosynthesis. He recently completed his doctoral studies and has returned to Lausanne to carry out postdoctoral work with Prof Kai Johnsson. His interests lie in the development of chemical tools to study biological systems.

Christian Ducho

Christian Ducho was born in 1976 and studied Chemistry at the University of Hamburg, Germany, from 1996 to 2001. He completed his Ph.D. in Organic Chemistry in the group of Chris Meier at the University of Hamburg in 2005. He then moved to the University of Oxford, UK, for postdoctoral research (biosynthesis of carbapenem antibiotics) in the group of Christopher Schofield from 2005 to 2007. In 2007, he commenced his independent academic career in Germany as an Assistant Professor (‘Juniorprofessor’) at the Georg-August-University Göttingen. In 2011, he moved to the University of Paderborn, Germany, as Associate Professor of Organic Chemistry.

Michael A. McDonough

Michael McDonough received his PhD in Biophysics, with a focus in X-ray crystallography, from the University of Connecticut in 2000 where he made significant contributions to the fundamental understanding of mechanisms by which beta-lactam antibiotic targets bind to their substrates. He then spent two years at the University of Copenhagen honing his structural skills at the Centre for Crystallographic Studies before settling in Oxford in 2003 where he has since been instrumental in the efforts to structurally characterise enzymes involved in clavulanic acid and carbapenem biosynthesis.

Christopher J. Schofield

Chris Schofield is currently Head of the Organic Chemistry, and a Fellow of Hertford College, University of Oxford. His current research interests include antibiotic biosynthesis and resistance, regulation of gene expression by oxygen, and epigenetics with a common theme being enzymes that catalyse chemically interesting reactions of biomedicinal importance.

---

1 Introduction

The family of β-lactam antibiotics (BLAs, Fig. 1) is the most important class of clinically used antibiotics with more than half of the global antibiotic market share.^1,2 However, over the past two decades, research on BLAs by most major pharmaceutical companies has decreased substantially compared to the half-century following their clinical introduction in the 1940s. The same trend is true for other classes of established antibiotics. The reasons for this decrease in activity are probably primarily commercial with the size of potential revenue being insufficient to justify the high cost of antibiotic development.³ However, the increasing problems of antibiotic resistance⁴ and Gram-negative infections⁵ coupled to a critical shortage of new antibiotics⁶ are now stimulating renewed interest in antibiotic development, both in the public and private sectors.^4,7 Given their track record of efficacy and safety, it may be that the timing is appropriate for a resurgence of interest in BLAs.

Fig. 1 Major subfamilies of β-lactams. The names of synthetic β-lactam antibiotic (BLA) subfamilies are in blue. Lactivicin (in the dashed box) is a naturally-occurring γ-lactam antibiotic, which, like BLAs, inhibits penicillin binding proteins (PBPs). Clavulanic acid (a clavam with the (5R)-stereochemistry) is a serine β-lactamase inhibitor; clavams (with the (5S)-stereochemistry) have antibacterial and antifungal activities (the latter not via PBP inhibition).

The targets of BLAs are transpeptidase enzymes which are involved in bacterial cell wall biosynthesis.⁸ The transpeptidases, or penicillin binding proteins (PBPs),⁹ catalyse peptidoglycan cross-linking (Fig. 2) and are likely amenable to modern medicinal approaches that would complement the vast structure–activity relationship data sets acquired in the period leading up to the 1990s. Such studies will benefit from advances in molecular and structural understandings of cell-wall biosynthesis and the enzymes involved (for reviews, see ref. 10, 11). However, a problem with some BLAs subfamilies is that their commercially viable synthesis remains challenging, because of the high degree of functionalisation and the enantioenriched and reactive nature of the core bicyclic ring-structures. Indeed, the issue of production costs is more of an issue with antibiotics than many other classes of pharmaceuticals. Another issue with BLAs is the need for specialised production facilities due to allergy concerns.

Fig. 2 Transpeptidase-catalysed reactions during bacterial cell wall biosynthesis are inhibited by β-lactam antibiotics. Formation of the mature peptidoglycan matrix involves transpeptidase-catalysed cross-linking of linear polysaccharide chains by short peptides. The reaction is exemplified in outline for Staphylococcus aureus.

With the exception of aztreonam and the carbapenems, clinically used β-lactams are produced either by isolation of fermented compounds (e.g. clavulanic acid) or by synthetic or enzymatic modification of fermented compounds (e.g. many penicillins and all such cephalosporins). For this reason, β-lactam biosynthesis has been a focal point for the application of new methodologies in molecular biology with a view to optimising production procedures. Presently, there is interest in optimising sustainable routes to pharmaceuticals. It would seem likely that if there is resurgence in interest in BLAs research, manipulation of their biosynthesis pathways might be an important tool for the efficient and sustainable production of new antibiotics.

Following the pioneering demonstration that β-lactams are elastase inhibitors,¹² β-lactams are also being explored in other non-antibiotic therapeutic applications.¹³ β-Lactams have been used to inhibit various proteases (including HIV,¹⁴ cytomegalovirus¹⁵ and rhomboid intramembrane¹⁶ proteases), to act as cholesterol absorption inhibitors¹⁷ (ezetimibe is used clinically^18,19), as vehicles for anticancer drug delivery,¹³ as anticancer drugs,^20–22 and in the regulation of the excitatory neurotransmitter glutamate.²³ The clinical success and potential of some of these compounds is stimulating continued interest in developing efficient routes to β-lactams. Aside from improving practical routes to existing antibiotics and providing access to new β-lactam structures, a further justification for work on BLA biosynthesis is that many of the reactions (and enzymes that catalyse them) are of basic (bio)chemical interest. Such enzymes include the synthetases involved in peptide and β-lactam formation and, of particular interest to our group, the oxygen-utilizing enzymes that catalyse the formation of the bicyclic ring structure (e.g. in penicillins and clavams) and other modifications. Studies on these enzymes have not only increased our understanding of how nature synthesises molecules that are synthetically inaccessible but has also impacted on fields as far varying as plant hormone signalling, collagen biosynthesis, DNA repair, epigenetics and human oxygen-sensing (for reviews, see ref. 24–28).

Here we review mechanistic and structural studies on β-lactam biosynthesis focusing on the enzymes involved and in particular on recent work (for previous reviews, see ref. 29–34). In several cases we also highlight the potential for protein engineering of β-lactam biosynthesis enzymes, both for the purpose of pathway engineering and more generally for the biocatalytic production of useful compounds.

2 Naturally-occurring β-lactams

BLAs isolated from natural origin may be classified into the penicillin, cephalosporin, carbapenem/carbapenam, clavam, and monocyclic β-lactam subfamilies (Fig. 1). All clinically used BLAs likely work by inhibition of PBPs via formation of relatively stable acyl-enzyme complexes (Fig. 2–4). PBP-inhibition results in impaired peptidoglycan biosynthesis and consequent cytolysis. It is notable that whilst many naturally-occurring BLAs which inhibit PBPs have been identified, only one naturally-occurring non β-lactam compound, lactivicin (Fig. 1), that targets PBPs, has been isolated.^35–37 It is proposed that the apparently special ability of BLAs to inhibit PBPs (and possibly other enzymes which employ nucleophilic serine/threonine/cysteine residues), results from a combination of (i) efficient binding as mediated by appropriate functionalisation and stereochemistry, (ii) possible mimicry of the “tetrahedral” transition state occurring during amide hydrolysis (the Tipper–Strominger hypothesis³⁸), (iii) irreversible reaction with the nucleophilic serine at the active site of PBPs, in part because of the highly strained nature of the β-lactam ring, and (iv) formation of an acyl-enzyme complex that is stable with respect to hydrolysis/nucleophilic attack.^8,38–40 In the case of lactivicin, a γ-lactam, cycloserine derivative, the unusual fused spirocyclic ring system is proposed to enable irreversible reaction with PBPs.^37,41,42

Fig. 3 Views from crystal structures showing examples of relatively stable acyl-enzyme complexes formed by reaction of β-lactams with PBPs and β-lactamases. A: View from a structure obtained by reaction of a serine β-lactamase SHV-1 E166A with sulbactam (PDB 2A3U).^679–681 Glu166, which is involved in hydrolysis of the acyl-enzyme intermediate, was substituted for alanine to enable trapping of the complex; B: View from a complex resulting from reaction of the Mycobacterium tuberculosis β-lactamase BlaC with clavulanic acid (PDB 3CG5). Following nucleophilic attack by Ser70 and β-lactam ring opening, the complex decarboxylates to generate the shown adduct with the clavulanate-derived fragment (CDF,⁶⁸² which is also shown in blue in Fig. 4iii); C: View from a structure of a serine β-lactamase (SHV-1)/meropenem acyl-enzyme complex (PDB 2ZD8). Two conformations were observed, one with the carbonyl of the acyl-enzyme complex in the oxyanion hole (formed by NH amides of Ser70 and Ala237), and another with the carbonyl directed away from the oxyanion hole. Only part of the C-2 substituent of meropenem is shown; D: View from a PBP3 (Pseudomonas aeruginosa)/aztreonam acyl-enzyme complex (PDB 3PBS).⁶⁸³ The white arrows point to the carbonyl carbon of the (hydrolysed) β-lactam. Structures of sulbactam and aztreonam are given in Fig. 5, while those of clavulanic acid and meropenem are in Fig. 6 and Fig. 46, respectively.

Fig. 4 Comparison of outline reaction mechanisms of different β-lactams with PBPs and β-lactamases. (i) β-Lactams bind to the transpeptidase domain of PBPs to form relatively stable acyl-enzyme complexes; (ii) β-Lactams react with serine β-lactamases to form acyl-enzyme complexes labile to hydrolysis; (iii) Clavulanic acid, and other β-lactam-based serine β-lactamase inhibitors, react with serine β-lactamases to form stable acyl-enzyme complexes;⁶⁸⁴ (iv) Thienamycin derivatives (e.g. imipenem) react with some serine β-lactamases to form a stable acyl-enzyme complex;^56,685 (v) β-Lactams react with zinc-dependent metallo-β-lactamases forming a complex that is labile to hydrolysis (aztreonam is an exception⁴⁸). Note that the reaction mechanisms are more complex than shown here and variations on the shown outline mechanisms occur. The water molecule shared between Zn^II ions likely exists as a hydroxide ion, which can act as the nucleophile during metallo-β-lactamase catalysis.^686,687

Resistance to BLAs was identified very soon after their first clinical use in the 1940s,⁴³ and is now known to be mediated via processes including⁴⁴ (i) mutation of PBPs to block inhibitor binding,⁴⁵ (ii) ejection of β-lactams by efflux pumps,⁴⁶ (iii) reduced permeability to β-lactams,⁴⁷ and (iv) the production of β-lactamases that catalyse β-lactam ring hydrolysis.^48–50 The largest family of identified β-lactamases likely evolved from PBPs and similarly employ a serine-residue as a nucleophilic catalyst.^51–53 Metallo-β-lactamases,⁵⁴ although not yet as clinically relevant as their serine-relative, employ one or two zinc ions to activate a “hydrolytic” water molecule and are an increasing threat because of their ability to hydrolyse almost all types of BLAs (a notable exception being aztreonam^54,55). The BLAs show different degrees of stability with respect to β-lactamases, with the cephalosporins and, particularly, the carbapenems showing, in general, increased stability compared to the penicillins (like β-lactamase inhibitors, carbapenems can undergo fragmentation on reaction with serine β-lactamases, Fig. 3 and 4).⁵⁶

Notably, clinically used serine β-lactamase inhibitors are themselves β-lactams, e.g. clavulanic acid (naturally-occurring), sulbactam and tazobactam (Fig. 5). These inhibitors react with the β-lactamase active site serine to form an acyl-enzyme complex, but the initially formed complex fragments to form an ester that is unusually stable with respect to hydrolysis. Because these compounds (Fig. 5) are insufficiently potent as antibiotics, they are used in combination with a BLA (e.g. amoxicillin and clavulanic acid are combined as Augmentin^®). As with the antibacterial activity of BLAs, one current objective for β-lactamase inhibitor development is to increase the spectrum of activity. The currently used β-lactamase inhibitors^57,58 only efficiently inhibit Class A of the three (A, C and D) classes of serine β-lactamases and do not inhibit the metallo-β-lactamases (Class B). AM-112,⁵⁹ an oxapenem, and NXL-104 (Avibactam),^60–62 a synthetic bridged γ-lactam (Fig. 5), have both been developed as broad spectrum β-lactamase inhibitors; Avibactam is presently in clinical trials.

Fig. 5 Structures of aztreonam and some β-lactamase inhibitors. Note that some of the inhibitors (e.g. sulbactam, tazobactam and AM-112) also show antibacterial activity. Clavulanic acid (Fig. 6), sulbactam⁶⁸⁸ and tazobactam⁶⁸⁹ are used clinically in combination with β-lactam antibiotics to protect the latter against the effects of β-lactamases. AM-112⁵⁹ and NXL104 (Avibactam)^60–62 are examples of broad spectrum serine β-lactamase inhibitors.

3 Overview of β-lactam biosynthesis

An important step in the development of clinically useful penicillins was the discovery that different penicillins could be produced by modification of growth media and strains. The combination of corn-steep liquor as a growth medium, a high yielding strain of Penicillium chrysogenum and use of aerated deep fermentation had a historically important role in the development of penicillins. It was found that supplementing the fermentation medium with aromatic acids could result in the production of specific penicillins, e.g. addition of phenoxyacetic acid results in formation of penicillin V, which has a 6-phenoxyacetamido-side chain (Fig. 6), as the major product.⁶³ Penicillin V was developed for oral clinical use because of its increased resistance to acid-catalysed hydrolysis in the stomach compared to penicillin G (the presence of the electron-withdrawing phenoxy group hinders attack of the side chain carbonyl-oxygen on the β-lactam carbonyl).⁶⁴ The development of procedures to produce penicillins with neutral hydrophobic side chains (via supplementing the fermentation medium with appropriate precursors)^63,65 was important because it enables the purification of penicillins from the fermentation medium by extraction. This work was a pioneering example of metabolic engineering. However, the type of penicillins that could be produced by this technique was limited. A crucial further advancement was the development of procedures for the production of 6-aminopenicillanic acid (6-APA)⁶⁶ from extracted hydrophobic penicillins (e.g. penicillin G). This procedure is catalysed by penicillin acylases⁶⁷ (Section 4.10). The 6-APA can then be acylated to give “semi-synthetic” penicillins with diverse side chains – many have found clinical applications, e.g. piperacillin, amoxicillin, and ampicillin (Fig. 7A).

Fig. 6 Prototypical examples of different subfamilies of naturally-occurring β-lactam antibiotics.

Fig. 7 Examples of semisynthetic penicillins (A) and cephalosporins (B) derived from 6-aminopenicillanic acid (6-APA) and 7-aminocephalosporanic acid (7-ACA)/7-ACA derivative, respectively.

The need to combat penicillin resistance provoked a search for new BLAs, which resulted in the isolation of cephalosporin C⁶⁸ from Acremonium chrysogenum (previously known as Cephalosporium acremonium). Cephalosporin C (Fig. 6), like other cephalosporins, shows resistance to some serine β-lactamases (Class A). Procedures were developed for the production of 7-aminocephalosporanic acid (7-ACA, Fig. 7), though these were hindered by the fact that, in contrast to penicillins, it is not yet possible to (at least efficiently) produce cephalosporins with hydrophobic side chains by the simple addition of hydrophobic acetic acid derivatives to the wildtype cephalosporin-producing fermentations (see Section 4.11). One successful approach to address this problem has been the development of chemical methods for the ring-expansion of penicillins with hydrophobic side chains to give cephalosporins.⁶⁹ Enzymic or chemical deacylation is then used to hydrolyse the side chain at the 7-position of the resultant cephalosporins (Section 4.10, where other approaches to producing 7-ACA are also described). Compared to the penicillins, the cephalosporins have an additional site for modification, i.e. at the C-3′ position. Esterases have been identified that efficiently catalyse the hydrolysis of the C-3′ ester of cephalosporin C to give an allylic alcohol which can be subsequently modified, so enabling the production of many cephalosporin derivatives. Thousands of cephalosporin derivatives, modified at C-7 and/or C-3′, have been produced by semi-synthesis with many achieving clinical utility (Fig. 7B).^70,71

The enzymes used in the production of 6-APA, 7-ACA (Section 4.10) and C-3′ hydroxymethyl-cephalosporins are highly efficient on industrial scale. Because these enzymes are not part of the BLA biosynthesis pathways, they are not the subject of this review (see ref. 1, 2, 72 for reviews), but it should be noted that some are structurally and mechanistically related to enzymes involved in BLA biosynthesis (Section 4.5). For example, the penicillin acylases, used in the commercial production of 6-APA from hydrophobic penicillins, and isopenicillin N acyl-transferase, which is involved in the biosynthesis of hydrophobic penicillins, are both members of the N-terminal nucleophilic (Ntn) hydrolase family.^73,74

In the late 1970s, new strategies in which supersensitive strains of bacteria, some with β-lactamase inhibitory activities, were used to screen massive number of samples containing soil micro-organisms for BLAs.^75,76 The new screens coupled to advances in chromatography led to the discovery of three new subfamilies of BLAs: the clavams, monobactams and carbapenems (Fig. 6).

Clavulanic acid (Fig. 6) was the first member of the clavam subfamily of BLAs to be isolated (from Streptomyces clavuligerus), and it was found to be a potent inhibitor of Class A β-lactamases (Fig. 3 and 4) but a relatively weak antibiotic.⁷⁷ The clavams are distinguished from the penicillins by replacement of the sulfur atom of the thiazolidine ring with an oxygen atom to give an oxazolidine ring.

Nocardicin A (Fig. 6) was the first naturally-occurring monocyclic β-lactam to be isolated (from Nocardia uniformis subsp. Tsuyamanensis).⁷⁸ Nocardicin A has weak to moderate activity against Gram-negative and Gram-positive bacteria, and some resistance to β-lactamases; to date the nocardicins have not found clinical application because of their relatively poor antibacterial properties.⁷⁹ The development of monobactams, with the aim of increasing their binding affinity to PBPs, resulted in the discovery of aztreonam (Fig. 5) which is effective against Gram-negative bacteria and unusually resistant to hydrolysis by β-lactamases.⁷⁹

Thienamycin (Fig. 6), named for its novel β-thioenamine group,⁸⁰ was the first of the clinically important carbapenem antibiotics to be isolated (from Streptomyces cattleya).^81,82 The carbapenems are distinguished from the penicillins by replacement of the sulfur atom in the five-membered ring with a methylene group, and by desaturation between C-2 and C-3. Thienamycin has been described as one of the most potent broad-spectrum antibacterial agents isolated from natural sources;⁸³ it has good activity against many Gram-positive and Gram-negative bacteria, and inhibits some serine β-lactamases. Despite its activity, the chemical and metabolic instability of thienamycin have prevented its clinical use. Instead, stable derivatives of thienamycin, e.g. ertapenem, meropenem and doripenem, have been developed and are now widely used as antibiotics (Section 6.1).⁸⁴

An important difference between the penicillin and cephalosporin development stories and that of the clavams, monobactams, and carbapenems is that in the latter three cases the semisynthetic approach has been much less productive. Whilst some clavulanic acid derivatives have been prepared by modifications of the C-2 allylic alcohol,^85–88 modifications at the C-6 position have been limited by the lack of a functional group at this position and inefficiency of synthetic methodologies for C-6 modification. This is a pity because it is known that clavulanic acid derivatives functionalised at C-6 (e.g. with the hydroxyethyl group of carbapenems) have a useful spectrum of β-lactamase inhibitory activity,⁸⁹ as do the related synthetic oxapenems (e.g. AM-112, Fig. 5).⁵⁹ The situation with fermented carbapenems has been similarly frustrating. Despite extensive efforts, it was not possible to increase fermentation titres of carbapenems to useful large-scale levels. As a consequence, total synthesis methods were developed to produce clinically used carbapenems. The development of these methods enabling the production of densely functionalised and relatively labile carbapenems is one of the triumphs of the 20th century synthetic chemistry. Whilst the routes are remarkably efficient (for reviews, see ref. 80, 90), the available synthetic methodology limits the range of accessible compounds, with one particular limitation being the reliance on 4-acetoxy-3-[1-(tert-butyldimethylsilyloxy)ethyl]azetidin-2-one (AOSA, Section 6.1) as a synthetic intermediate.^90–92 The problem of efficient and inexpensive biotechnological production of modified clavams and carbapenems has stimulated efforts to understand and manipulate their biosynthesis pathways through exploiting the techniques of molecular and structural biology. As with the related pioneering studies on penicillin and cephalosporin biosynthesis, this work may have come too late to have a major impact in the first golden age of BLAs, but it is possible that it may help enable production of new and existing antibiotics by efficient and sustainable methods.

Here we review knowledge of the pathways leading to the BLAs with an emphasis on recent work on the biosynthesis enzymes and reactions of chemical interest.

4 Penicillin and cephalosporin biosynthesis pathways

4.1 Introduction – overview of the pathways

To date, the formation of hydrophobic penicillins (Fig. 8) has only been reported in fungi, e.g. P. chrysogenum. In contrast, the hydrophilic cephalosporins are produced by both fungi (e.g. A. chrysogenum), and bacteria. The cephamycins have been isolated from Gram-positive bacteria, e.g. Streptomyces sp., while the cephabacins and related structures (Fig. 8) have been isolated from Gram-negative bacteria, e.g. Flavobacterium sp.,^93–95Xanthomonas lactamgena^96–99 and Lysobacter lactamgenus.^98,99 The Gram-positive actinomycetes are probably able to produce the most diverse array of β-lactam containing structures, including not only the penicillins, cephalosporins and cephamycins but also carbapenems, clavams, and nocardicins.

Fig. 8 Penicillins and cephalosporins isolated from natural sources. Compounds in red are obtained as the result of specific precursor addition to the fermentation medium: phenoxyacetic acid for penicillin V,⁶³L-S-carboxymethylcysteine for RIT 2214⁶⁹⁰ and aromatic thiols for the oganomycins (F-I).¹¹³ Most of the shown penicillins (except penicillin N and KPN) are produced by Penicillium chrysogenum and/or P. notatum in varying ratios depending on the strain and culture conditions. Penicillin F, a major product of P. notatum, is designated so to identify it as the penicillin discovered by Fleming.⁶⁹¹ Penicillin G is a major product of P. chrysogenum grown on corn-steep liquor. Penicillin KPN is produced by strains of the genus Paecillomyces.⁶⁹² Cephalosporins with α-ketoglutaryl side chain (e.g. compounds A, B and C)⁶⁹³ were detected in culture broth of some strains belonging to the genus Acremonium, and are predicted to be produced by an oxidase-mediated deamination of the C-7 D-amino-adipate side chain.³⁶⁷ The 3′-thiomethylcephems F-1¹¹¹ and C-43219¹¹² were isolated from culture broth of mutants of Acremonium chrysogenum, while SF-1623¹¹⁰ was isolated from culture broth of Streptomyces chartreusis. Cephalosporin PA-41937 was isolated from the PA-41937 Streptomyces strain.⁶⁹⁴ Cephamycin PA-32413-I was isolated from the PA-32413 S. clavuligerus strain.⁶⁹⁵ 7α-Methoxycephems with cinnamoyl-side chains at C-3′ were isolated from Streptomyces spp.^{320,345,696,697} The cephabacins are produced by some Gram-negative bacteria e.g. Flavobacterium sp. (SQ 28516/7, chitinovorin C D⁻¹),^93–95Xanthomonas lactamgena (M_1–6, F_4–9, H_4–6)^96–99 and Lysobacter lactamgenus (F_1–3 and H_1–3).^98,99

The biosynthetic routes leading to the core nuclei of penicillins and cephalosporins (Fig. 9) have been extensively investigated (for prior reviews, see ref. 1, 30, 31, 100–103). The committed step in these routes is the assembly of the tripeptide LLD-ACV from L-α-aminoadipic acid, L-cysteine and L-valine (Fig. 9).¹⁰⁴LLD-ACV biosynthesis is catalysed by the multidomain, multimodular, non-ribosomal peptide synthetase ACVS – which in addition to catalysing formation of two peptide bonds, also catalyses inversion of the C-α stereochemistry of L-valine to the D-valine residue in LLD-ACV¹⁰⁵ (Fig. 9). The first β-lactam to be formed, isopenicillin N, is then produced from LLD-ACV in a remarkable oxidative cyclisation reaction catalysed by the Fe^II-dependent oxidase isopenicillin N synthase^31,106 (Fig. 9). In some fungi, e.g. P. chrysogenum, acyl-transferases act upon isopenicillin N to give various acyl-penicillin derivatives, e.g. penicillin G (Fig. 8). In some fungi and bacteria, e.g. Streptomyces spp, the side chain epimerisation of isopenicillin N to penicillin N, catalysed by isopenicillin N epimerase, enables branching of the penicillins/cephalosporins pathway (Fig. 9). Penicillin N transformation into cephalosporin C proceeds in 3 steps (i) expansion of the 5-membered thiazolidine ring of penicillin N into the 6-membered dihydrothiazine ring of deacetoxycephalosporin C (DAOC), as catalysed by the 2-oxoglutarate (2OG) dependent oxygenase deacetoxycephalosporin C synthase (DAOCS); (ii) hydroxylation of DAOC to give deacetylcephalosporin C (DAC), as catalysed by DAOC hydroxylase; and (iii) acetylation of DAC to give cephalosporin C, as catalysed by DAC acetyltransferase (Fig. 9).

Fig. 9 Biosynthetic pathways leading to penicillins and cephalosporins in fungi and bacteria. Whilst bacterial cephamycin-producers have a specific pathway to form L-α-aminoadipic acid (L-AAA) from lysine (shown in blue), L-AAA is an intermediate of lysine biosynthesis in penicillin-producing fungi. The isopenicillin N epimerase in cephalosporin-producing bacteria (i.e. CefD) is pyridoxal phosphate (PLP)-dependent; an alternative isopenicillin N epimerisation system (comprising CefD1 and CefD2) operates in fungi.^248,253 In prokaryotes, the DAOC synthase and DAOC hydroxylase reactions are carried out by separate enzymes (encoded by cefE and cefF, respectively), whereas in fungi, a bifunctional enzyme (encoded by cefEF) catalyses both reactions. CmcJ was previously not considered as belonging to the stereotypical 2OG oxygenase superfamily;¹⁰⁹ however, bioinformatic analyses suggest that it is likely a 2OG-dependent hydroxylase.³³⁰ Acronyms for the genes encoding for enzymes are in parentheses (see Fig. 10, Table 1). Selected cofactors and co-substrates are shown. A non-ribosomal peptide synthetase (NRPS)/polyketide synthase (PKS) hybrid system (CpbI/CpbK) is proposed to be involved in the assembly of the C-3′ side chain of the cephabacins.^114,115

The cephamycins are 7α-methoxy-cephem derivatives with different side chains at C-3′ (for reviews, see ref. 30, 107). Cephamycin C biosynthesis gene cluster (Fig. 10, Table 1) contains specific genes related to L-α-aminoadipic acid formation, and genes encoding for enzymes involved in DAC modification.^30,107 Late steps in cephamycin C biosynthesis (Fig. 9) involve carbamoylation of the C-3′-hydroxyl group, as catalysed by the O-carbamoyltransferase CmcH,¹⁰⁸ and 7α-methoxylation, as catalysed by CmcJ and CmcI (Fig. 9).¹⁰⁹ The side chain at the 3′-position of cephamycins and oganomycins (Fig. 8) can be acetyl (e.g. A-16884A), carbamoyl (e.g. cephamycin C), as well as esters of p-coumaric acid derivatives (e.g. cephamycin A and B, oganomycins). Derivatives of 3-thiomethylcephems (Fig. 8) have also been reported in the culture broth of S. chartreusis (e.g. SF-1623¹¹⁰), mutants of A. chrysogenum (e.g. F-1¹¹¹ and C-43219¹¹²), and after specific precursor (e.g. heterocyclic thiols) addition to the culture broth of S. oganonensis (e.g. oganomycins F-I¹¹³).

Fig. 10 Examples of penicillin and cephalosporin biosynthesis gene clusters from fungi (Penicillium chrysogenum and Acremonium chrysogenum) and bacteria (Streptomyces clavuligerus and Lysobacter lactamgenus).^266,698,699 For the role of the biosynthetic proteins encoded by the genes shown, see Table 1 and Fig. 9. The protein encoded by orf12 (L. lactamgenus) shows similarity to methyltransferases;¹¹⁵ however, the reported cephabacins from L. lactamgenus are not methoxylated at C-7, hence the role of orf12 is unclear. The orf2 gene (S. clavuligerus) is part of the clavulanic acid gene cluster (Fig. 28). The L. lactamgenus cluster, the only cluster reported so far for cephabacin biosynthesis, appears to be incomplete (e.g. the genes encoding for the proteins responsible for the introduction of the C-7 formylamino moiety and for L-aminoadipate biosynthesis are not yet identified^115,700).

Table 1 Genes constituting the reported penicillin/cephalosporin gene clusters (Fig. 10) in Penicillium chrysogenum (^a), Acremonium chrysogenum (^b), Streptomyces clavuligerus (^c) and Lysobacter lactamgenus (^d) and the (predicted) roles of the (putative) proteins that they (may) encode for. The predicted number of amino acid residues (AA) for each of the (putative) proteins is shown. Proteins with an experimentally assigned function are in bold. The order of genes in Table 1 is according to their order in Fig. 10

Gene	AA	(Proposed) function of encoded protein
pcbAB^a–d	3746^a	ACV Synthetase (ACVS); formation of LLD-ACV tripeptide.^134–137
pcbC^a–d	331^a	Isopenicillin N synthase (IPNS); LLD-ACV oxidative cyclisation.^106,172
penDE^a	357	Isopenicillin N acyltransferase (AT).^225–228
cefT^b	561	Efflux pump protein for the export of cephalosporin C.^701
orf3^b		Unknown.
cefD2^b	383	Isopenicillin N-CoA epimerase (CefD2).^253
cefD1^b	609	Isopenicillin N-CoA synthetase (CefD1).^253
cefEF^b	332	Deacetylcephalosporin C synthase/hydroxylase (DAOCS/DACS).^255–258
cefG^b	385	Acetyl-coenzyme A:DAC O-acetyltransferase (CefG).^337
pcbR^c	551	Protein involved in BLA resistance.^702
lat^c	457	L-Lysine-ε-aminotransferase (LAT).^124,703
blp^c	182	Extracellular β-lactamase inhibitory protein.^704
orf11^c		Unknown.
ccaR^c	256	Transcriptional activator (CcaR) essential for the biosynthesis of cephamycin, clavulanic acid and clavams.^704
cmcH^c	521	O-Carbamoyltransferase (CmcH).^108,703
cefF^c,d	318	Deacetylcephalosporin C hydroxylase (DACS).^259–262
cmcJ^c	310	7α-Hydroxylase (CmcJ, 2OG oxygenase).^109,330
cmcI^c	236	Methyl transferase (CmcI).^109
cefD^c,d	398^c	Isopenicillin N epimerase (CefD); epimerisation of L- to D-α-aminoadipoyl side chain.^246–248
cefE^c,d	311	Deacetylcephalosporin C synthase (DAOCS).^259–262
pcd^c	496	Piperidine-6-carboxylate dehydrogenase (Pcd).^125,704
cmcT^c	489	Efflux pump protein for the export of cephamycin C.^125,704,705
pbpA^c	409	Protein involved in resistance against β-lactam antibiotics.^704
orf12^d	302	Methyltransferase.^115
cpbK^d/cpbI^d	1056/5049	Non-ribosomal peptide synthetase/polyketide synthase hybrid involved in biosynthesis of the C-3′-tetrapeptide of cephabacins.^114,115
t1^d		ABC-transporter protein.^115,700
t2^d		ABC-transporter protein.^115,700
t3^d		ABC-transporter protein involved in secretion of cephabacins.^700,706
bla^d	385	Protein involved in BLA resistance.^115

---

The Gram-negative bacteria Flavobacterium sp.,^93–95X. lactamgena^96–99 and L. lactamgenus^98,99 are able to further modify DAC to produce cephabacins which have a substituent at C-3′ comprising acetate and oligopeptide moieties, and may (e.g. cephabacin F group) or may not (e.g. cephabacin H group) contain 7α-formylamino⁹⁸ or 7α-methoxy groups (e.g. cephabacin M group^96,97) (Fig.8). The substituent at C-3′ of the cephem nucleus of cephabacins is proposed to be assembled by a non-ribosomal peptide synthetase/polyketide synthase hybrid encoded for by cpbI/cpbK.^114,115

4.2 Biosynthesis of L-α-aminoadipic acid

The L-α-aminoadipoyl side chain of the tripeptide ACV is crucial for the biosynthesis of penicillins as substantiated by the observation that a mutant of P. chrysogenum blocked in the biosynthesis of L-α-aminoadipic acid only produced penicillin G when L-α-aminoadipic acid was added to the growth medium.¹¹⁶L-α-Aminoadipic acid is an intermediate in lysine biosynthesis in (penicillin-producing) fungi;^117–119 penicillin production in P. chrysogenum is inhibited by lysine due to end-product control exerted by lysine decreasing the availability of α-aminoadipate for penicillin biosynthesis.¹²⁰ Cephamycin-producing bacteria (e.g. Amycolatopsis lactamdurans¹²¹ and S. clavuligerus¹²²) form lysine via the diaminopimelic acid pathway¹²³ which does not involve L-α-aminoadipic acid as an intermediate. Thus, these organisms have a specific system to form L-α-aminoadipic acid from lysine via two enzyme-catalysed steps involving L-lysine-ε-aminotransferase (LAT, encoded by the lat gene)¹²⁴ and piperideine-6-carboxylate dehydrogenase (P6CD, encoded by the pcd gene)¹²⁵ (Fig. 9). Lysine-supplemented cultures of A. lactamdurans showed higher titres of cephamycin C,¹²⁶ as did the overexpression of the lat gene from strong heterologous promoters.¹²⁷ LAT employs 2OG, as an amino-group acceptor, and pyridoxal phosphate, as a cofactor, to enable deamination of lysine to give P6C.^121,122,128 LAT also accepts L-ornithine and, to a lesser extent, N-acetyl-L-lysine as amino-group donors.¹¹⁹ P6CD converts P6C/NAD⁺ into L-α-aminoadipic acid/NADH.^125,129 The specific activity of LAT peaks before cephamycins production and decreases as cephamycins are formed.¹²⁴ P6CD does not accept pyrroline-5-carboxylate, an analogous intermediate in proline/carbapenem biosynthesis, as a substrate analogue.¹²⁹ In S. clavuligerus, whilst the lat gene mutants lose the ability to produce cephamycins, the pcd disrupted mutants are still able to form 30–70% of wildtype levels, implying the existence of an alternative mechanism to convert P6C into L-α-aminoadipic acid.^130–132 Cephamycin production by pcd mutants can be restored to wildtype levels either by addition of α-aminoadipic acid to the fermentation medium, or by complementation of the mutated strain with a copy of the pcd gene.¹³¹

4.3 δ-(L-α-Aminoadipoyl)-L-cysteinyl-D-valine synthetase (ACVS)

The ATP-dependent ACVS reaction is common to the biosynthesis of all penicillins and hence all cephems (Fig. 9). ACVS is a monomeric¹³³ multidomain nonribosomal peptide synthetase (NRPS) with a molecular mass of 404–426 kDa (depending on the source organism);^134–137 for general reviews on NRPSs, see ref. 138–142. ACVS was the first fungal NRPS to be identified. At least under some conditions, LLD-ACV formation is proposed to be the rate-limiting step^143,144 in the penicillin/cephalosporin biosynthesis.

The ACVS encoding gene was designated as pcbAB (Fig. 10) because 2 genetic loci were initially thought to be responsible for ACV synthesis;^145,146 however, genetic evidence indicates that a single gene encodes all the activities needed to form ACV.^147–149 Comparison of the predicted amino acid sequence of ACVSs to those of NRPSs with biochemically-characterised domains¹⁵⁰ reveals that ACVS possesses three modules (and probably at least ten catalytic domains). These modules contain different types of domains: (A) adenylation domain, (T) thiolation/pantothenylation/peptidyl carrier domain (which has a covalently attached 4′-phosphopantetheine prosthetic group), (C) condensation/peptide-bond formation domain, (E) epimerisation domain, and (TE) thioesterase domain. The domains within the two N-terminal modules of ACVS are arranged as (C/A/T) units. The C-terminal module of ACVS contains the (E) and (TE) domains (Fig. 11B).^151,152

Fig. 11 Outline potential mechanism for δ-(L-α-aminoadipoyl)-L-cysteinyl-D-valine synthetase (ACVS). (A) Amino acid substrates are activated by adenylation followed by transfer to the thiol of the 4′-phosphopantetheinyl cofactor(s); (B) Proposed domain organisation of ACVS outlining the multiple carrier thiotemplate mechanism. A1–A3: adenylation domains, T1–T3: thiolation domains, C1 and C2: condensation domains, E: epimerisation domain, and TE: thioesterase domain. Each amino acid is recognized and activated by its cognate adenylation domain (A), and attached as a thioester to the 4′-phosphopantetheinyl cofactor at the corresponding thiolation domain (T). Peptide bond formation is catalysed by the condensation domain (C). The L-valine residue of the tripeptide LLL-ACV is isomerised, at the α-carbon, by the epimerisation domain (E) followed by release of the final LLD-ACV by the thioesterase domain (TE); (C) Proposed role of the 4′-phosphopantetheinyl moiety as a “swinging arm” in peptide bond formation.

Unlike (normal) ribosomal peptide synthesis, NRPS can accommodate modified residues, as exemplified in the case of ACV/ACVS by the presence of an L-α-aminoadipate and a D-residue. The mechanism of ACVS catalysis (Fig. 11), like those of other NRPSs, is proposed to operate via a thiotemplate mechanism involving the recognition and activation of the substrate (L-amino acid) carboxylate group via adenylation (in (A) domain), followed by thiolation (in (T) domain) where transfer of the activated amino acid is coordinated by the 4′-phosphopantetheine cofactor (which acts as a “swinging arm”) to the condensation domain (C) for peptide bond formation. There is evidence that epimerisation of L-valine to the D-configuration, catalysed by the (E) domain, takes place at the tripeptide LLL-ACV level.¹⁵³ Finally, release of the enzyme-bound LLD-ACV occurs through the action of the (TE) domain.^105,151

Thiol-blocking agents (e.g. N-ethylmaleimide and iodoacetamide) inhibit ACVS activity revealing the importance of thiol groups in ACVS catalysis.^136,154 The conversion of the apo-form of ACVS into its active form, via transfer of a phosphopantetheinyl moiety from coenzyme A to a conserved serine residue in the three (T) domains of ACVS, is catalysed by a 4′-phosphopantetheinyl transferase (PPTase). A gene encoding for a specific PPTase in Aspergillus nidulans has been reported to be essential for penicillin biosynthesis.^155,156

Studies on the mechanism of the ACVS-catalysed L-valine epimerisation reveal that D-valine is not a substrate for ACVS and does not significantly induce ATP/PPi exchange.^154,157,158 To date, enzyme-bound thioester intermediates have not been detected in ACVS-catalysis. However, substrate analogue studies in which cysteine was replaced by L-O-methylserine led to isolation of both L-O-methylserinyl-L-valine and L-O-methylserinyl-D-valine dipeptides, implying that epimerisation occurs after formation of (at least one) peptide bond.^159,160 Modification of the (TE) domain of A. nidulans ACVS (by substituting Ser3599 in the conserved GXSXG motif for alanine, aiming to block the final hydrolytic step and trap an enzyme-bound intermediate) led to a ∼95% decrease in penicillin production.¹⁵³ The purified ACVS S3599A variant showed a ∼50% reduction in peptide formation rate, with LLL-ACV being the major product implying that (i) the (TE) domain activity is linked to the (E) domain, (ii) Ser3599 is not essential for peptide release but likely involved in selecting the isomer to be released,¹⁵³ and (iii) the intermediacy of LLL-ACV in penicillin biosynthesis. The epimerisation step possibly takes place at the enzyme-bound thioester level of LLL-ACV^105,137,161 (similarly to that in actinomycin biosynthesis¹⁶²). As the epimerisation step is likely to be reversible,^137,141,161 a selection mechanism for LLD-ACV hydrolysis over LLL-ACV is proposed (note that LLL-ACV is not a substrate for the next enzyme in the pathway, i.e. IPNS). Mutagenesis studies on P. chrysogenum ACVS¹⁶³ revealed that residues E3371, H3373, R3375 and E3376 are essential for the epimerase activity. Overexpression of the sequence encoding the (TE) domain of ACVS complemented mutants lacking the (TE) domain suggested that a “stand-alone” (TE) domain can recognize and interact with the other ACVS domains.¹⁶³

Wildtype ACVS has been found to accept a variety of substrate analogues,¹⁰⁵e.g. ACVS has been reported to substitute L-S-carboxyethylcysteine for L-α-aminoadipic acid, DL-homocysteine, L-S-methylcysteine, L-cystathionine, allylglycine and vinylglycine for L-cysteine (demonstrating that the thiol group of cysteine is not essential for peptide formation), and L-allo-isoleucine, L-α-aminobutyrate, L-norvaline, L-allylglycine or L-leucine for L-valine.^{154,158,164–166} These findings suggest that the biosynthesis of unnatural peptides for use as IPNS-substrates might be possible via the use of ACVS engineered variants.

Structural information on ACVS is of interest not only to reveal how an unusually small NRPS operates, but also because it may guide the rational redesign of ACVS to produce unnatural peptides. Progress has been made with structural studies on other NRPSs (e.g. the X-ray structure of a termination module (four domains) in Bacillus subtilis,^167–169 (T/C) didomain structure in tyrocidine synthetase, and (T/TE) in enterobactin synthetase^170,171), suggesting that ACVS may be amenable to crystallographic analyses.

4.4 Isopenicillin N synthase

The IPNS-catalysed reaction is exceptional in that a densely functionalised bicyclic ring is formed in a single four-electron oxidation step from a simple tripeptide (Fig. 9).^106,172 Only the reduced form of LLD-ACV is a substrate for IPNS, and there is evidence that, in P. chrysogenum, LLD-ACV is maintained in the reduced form by a thioredoxin reductase system.¹⁷³ The pcbAB and pcbC^174–178 genes, which encode for ACVS and IPNS, respectively, are always adjacent in biosynthesis gene clusters (Fig. 10), suggesting potential co-evolution. IPNS is one of few identified non-heme iron enzymes that do not require a co-factor for oxygen activation (unlike other studied 2OG oxygenases involved in BLA biosynthesis; Section 4.7).²⁴ Reducing agents (e.g. DTT and ascorbate) stimulate the activity of IPNS which is sensitive to thiol-specific inhibitors (e.g. N-ethylmaleimide).^179,180 Metal ions (e.g. Co^II, Zn^II and Mn^II) inhibit IPNS activity probably by competition with Fe^II.^180,181

4.4.1 Structure of IPNS. A crystal structure of IPNS in complex with Mn^II was the first to be solved, by Baldwin and coworkers, from the largest family of non-heme iron oxygenases/oxidases.¹⁰⁶ Anaerobic crystallisation methods were developed to obtain crystals of the IPNS:Fe^II:ACV complex and subsequently applied to study complexes after reaction with O₂.¹⁸² The various IPNS structures (Fig. 12) reveal a “core” double-stranded β-helix (DSBH) fold which is also present in the related 2OG oxygenases¹⁸³ including those involved in BLA biosynthesis. The DSBH fold supports the highly conserved 2-His-1-Asp binding motif that coordinates a single Fe^II at the active site (Fig. 12).^184–186 The DSBH topology normally comprises 8 β-strands that form a distorted β-sandwich structure arranged as two four-stranded antiparallel β-sheets (Fig. 12).²⁴ The remainder of IPNS structure (8 β-strands and 10 α-helices) lies outside the jelly-roll core. A C-terminal α-helix enters between the two layers of the jelly-roll enabling the side chain of Gln330 to ligate Fe^II in the resting state^187,188 (Fig. 12A and D). In the resting enzyme, the metal is octahedrally-coordinated by His214, Asp216, His270 and Gln330 and 2 water molecules¹⁸⁹ (Fig. 12A and D). Like other oxygenase enzymes, the IPNS active site sits in a largely hydrophobic cavity proposed to protect reactive intermediates from the external environment²⁴ and hinder self-oxidation by reactive intermediates.¹⁹⁰

Fig. 12 Structural views of isopenicillin N synthase (IPNS). A and D: IPNS-Mn^II complex (PDB 1IPS). Mn^II, which substitutes for Fe^II, is in violet; B and E: IPNS-Fe^II-ACV-nitric oxide complex (PDB 1BLZ).¹⁸⁹ Fe^II is in orange; C and F: IPNS-Fe^II-IPN complex (with ACV superimposed in case of F, PDB 1QJE). Note that the double stranded β-helix core fold (in yellow) – which supports the 2-histidine-1-aspartate iron binding motif – is a conserved structural feature for the non-heme iron(II) oxygenase/oxidase superfamily. The coordination site occupied by NO is highly likely to be the same as that for O₂. It is proposed that the presence of O₂trans to Asp216 prompts the valinyl-isopropyl group rotation about its Cα-Cβ bond (before/coupled to β-lactam formation) to direct the valine β-proton towards the Fe^IV=O for reaction.^{182,189,196,214} The contracted bicyclic structure of IPN relative to that of ACV is proposed to reduce the efficiency of interactions including at the carboxylate termini, and so promote product release.⁷⁰⁷

4.4.2 Mechanism of IPNS. A detailed mechanism for IPNS catalysis has been proposed¹⁸⁹ on the basis of kinetic,¹⁹¹ spectroscopic,^192–194 modelling,¹⁹⁵ crystallographic, and substrate analogues studies^{106,189,196–200} (Fig. 13). Upon binding to the active site Fe^IIvia its cysteinyl thiolate, ACV is proposed to displace Gln330 and (one of) the two water molecules ligating Fe^II (Fig. 12F). This results in a five-coordinate iron site¹⁸⁹ and creates a vacant Fe^II coordination site into which O₂ can bind (the valine isopropyl side chain is proposed to inhibit the ligation of water molecules to this site). Like other related non-heme iron(II) enzymes, in the resting state, iron is in the high-spin Fe^II state¹⁹² probably reflecting the weak ligand field of the coordinating residues.²⁰¹ ACV binding may increase the affinity of IPNS for dioxygen by increasing the hydrophobicity of the active site¹⁹⁹ and/or by reducing the Fe^II/Fe^III redox potential (as formation of the S–Fe^II bond likely increases the electron density on Fe^II).¹⁹⁵ Although an O₂-bound Fe^II-complex has not been experimentally observed, a nitric oxide-bound Fe^II-ACV complex is proposed to serve as a model for O₂ binding.^106,194,202 The structure reveals NO bound trans to Asp216, and it is assumed that O₂ binds in the same position (Fig. 12E). Kinetic isotope effect studies²⁰³ and the crystallographic observation of a monocyclic β-lactam in case of the substrate analogue δ-(L-α-aminoadipoyl)-L-cysteinyl-S-methyl-D-cysteine (ACMC)¹⁹⁶ imply that formation of the β-lactam occurs prior to formation of the C–S bond of the 5-membered thiazolidine ring. An iron-bound dioxygen species (superoxide) is proposed to first abstract the pro-3S proton of the cysteinyl β-methylene group, leading to a ferrous-hydroperoxide intermediate [Fe^II–OOH]. The peroxide can then abstract the valine N–H proton to generate a ferryl-oxo [Fe^IV=O] species; concomitantly, the valine nitrogen can nucleophilically attack the cysteinyl β-carbon, resulting in formation of a monocyclic β-lactam (Fig. 13A).²⁰⁴ Generation of a high-valent Fe^IV=O species is a common theme among non-heme iron oxidases/oxygenases and is normally generated by oxidative decarboxylation of the stereotypical co-substrate 2OG.²⁰⁵ Uniquely in IPNS catalysis, the Fe^IV=O species is generated by the 2-electron oxidative cyclisation of a thiol-containing substrate, i.e. ACV. After β-lactam ring formation, the Fe^IV=O species is proposed to react with the C–H bond of the valine β-carbon to generate a carbon-radical²⁰⁶ and a ferric-hydroxy [Fe^III–OH] species. Notably, structural studies imply that the valine β-C–H bond is directed away from the iron centre (Fig. 12E, 13A) and rotation of the valine isopropyl group must occur (likely coupled to β-lactam ring formation) to direct the valine β-proton towards the Fe^IV=O for productive reaction.^207,208 The valine-β-carbon radical can then react with the cysteine-thiolate to form the thiazolidine ring of isopenicillin N with retention of configuration at the valinyl β-carbon (Fig. 13A).^209,210 Independent deuterium kinetic isotope effects on both the cysteinyl β-carbon and the valinyl β-carbon suggest that both C–H bond cleavage steps are (at least partially) rate-limiting.²¹¹ The ring formation sequence is proposed to be determined in part by the relative strengths of the two C–H bonds with the Fe^IV=O species being preferred for cleavage of the stronger valine β-C–H bond.¹⁹⁵ The question of why IPNS catalyses ring formation rather than hydroxylation of C–H bonds has been addressed by density functional modelling studies, which suggest that the barrier for substrate hydroxylation is ∼4 kcal mol⁻¹ higher than that for ring closure.¹⁹⁵

Fig. 13 Proposed mechanism for isopenicillin N synthase catalysed conversion of the natural substrate LLD-ACV into isopenicillin N (A) and that of the substrate analogue δ-(L-α-aminoadipoyl)-L-cysteine D-α-hydroxyisovaleryl ester (ACOV) into a thiocarboxylate product (B).^196,224,708 R = δ-(L-α-aminoadipoyl). ACOV is near isosteric to LLD-ACV. Note that in the absence of its natural reaction partner (the N–H proton of the L-cysteinyl-D-valine amide bond), the proposed hydroperoxide intermediate can react with the putative thioaldehyde intermediate.

4.4.3 IPNS substrate analogues studies. Substrate analogue studies on IPNS have been highly productive, both from the mechanistic and in terms of realising the biocatalytic potential of non-heme oxygenases (for a review, see ref. 191). Here we highlight some interesting examples. Each of the three residues in LLD-ACV can be replaced by analogues, though replacement of the D-valine has been the most productive in terms of generating new bicyclic β-lactam nuclei (Fig. 14F to 14M).^210,212,213 Notably, however, the δ-(L-α-aminoadipoyl) residue of LLD-ACV can be replaced with hydrophobic side chains to give clinically useful penicillins (e.g. penicillin G and V) albeit in reduced yields (Fig. 14A).^214–217DLD-ACV is also a substrate for IPNS (Fig. 14B). Substitution of the L-cysteine residue has also been productive: e.g. (i) substitution of L-cysteine for (2R)- methylcysteine and (3R)-methylcysteine enabled production of 6α-methyl-isopenicillin and 5α-methyl-isopenicillin, respectively (Fig. 14C);²¹⁸ (ii) for production of γ-lactams (Fig. 14D);²¹⁹ and (iii) in case of 3,3-difluoro-L-homocysteine, where IPNS-catalysis resulted in the production of a thiocarboxylic acid (Fig. 14E) revealing that the full oxidising power of oxygen can be utilised for a four-electron oxidation of the carbon α- to the sulfur atom of the substrate analogue.²²⁰ Particularly, interesting results were obtained when the D-valine was replaced with unsaturated or cyclopropane containing residues, e.g.D-allylglycine, D-propargylglycine, and D-cyanoalanine (Fig. 14G, 14H and 14M, respectively).^210,212 In some cases, multiple bicyclic products are produced from a single tripeptide, e.g. in case of D-allylglycine- and cyclopropyl-containing tripeptides (Fig. 14G and 14 K, respectively).^213,221 Further examples are described in ref. 191, 196, 215–218, 222–224.

Fig. 14 The biocatalytic versatility of isopenicillin N synthase.^{191,196,210,213,215–218,222–224} These reactions exemplify the oxidase and oxygenase activities of IPNS. The approximate ratios of observed products are given below/beside the structures; in some cases, these ratios were perturbed by use of deuterated substrate analogues. Note in the cases of (N) and (O), the shown products were observed in crystal as part of IPNS-product complexes.^708,709 The reaction of IPNS with its natural substrate is boxed. R = δ-(L-α-aminoadipoyl); DAA = δ-(D-α-aminoadipoyl).

Although studies on oxygenases related to IPNS reveal that they too can accept a variety of substrate analogues (e.g. DAOCS and CAS, Sections 4.7 and 5.3, respectively), the available evidence is that IPNS is unusually promiscuous; this may be because, unlike the 2OG oxygenases, it only uses O₂ as a co-substrate, and that its prime substrate (i.e.LLD-ACV) is actually complexed to the active site iron. It should also be noted that almost all the substrate analogue studies on IPNS have been carried out on wildtype proteins, thus there is considerable chance for mutational studies to further expand the scope of IPNS reactivity, as well to optimise the formation of specific products (as has been demonstrated for DAOCS, Section 4.7).

4.5 Biosynthesis of the hydrophobic penicillins – isopenicillin amidohydrolase/acyltransferase

The final stages in the biosynthesis of penicillin G and other hydrophobic penicillins in P. chrysogenum are catalysed by an acyl-coenzyme A: isopenicillin N acyltransferase (AT, encoded by penDE) (Fig. 15A). AT catalyses the exchange of the hydrophilic α-aminoadipoyl side chain of isopenicillin N for an aromatic acyl side chain via 6-APA (Fig. 9 and 15A).^225–228 In order for the acyl-transfer to occur, prior activation of the (aromatic) acid to a CoA thioester by a specific ligase is required^229–231 (AT also accepts glutathione-activated aromatic acids²³²). Note that AT is distinct from a penicillin amidohydrolase (from P. chrysogenum) which catalyses the hydrolysis of penicillin V, and to a lesser extent penicillin G, to give 6-APA.²³³ AT thus has (IPN) amidohydrolase, (6-APA) acyl-transferase, as well as penicillin transacylase activities^228,234 (Fig. 15). The intermediacy of 6-APA in AT catalysis is supported by the finding that 6-APA is a better substrate than isopenicillin N to produce penicillin G, in the presence of phenylacetyl-CoA.²²⁵ 7-Aminocephalosporanic acid (7-ACA) is not a substrate for AT.²²⁵ AT accepts a range of acyl-CoA derivatives of both aromatic and aliphatic nature.^235–237 Only aliphatic carboxylic acids substrates with 6–8 carbon chain length were accepted in vitro by AT, consistent with the observation that penicillin K, F and dihydropenicillin F (Fig. 8) occur in the fermentation broths of P. chrysogenum together with penicillin G, when no phenylacetic acid was added as a precursor.²³⁶

Fig. 15 Reactions catalysed by acyl-coenzyme A: isopenicillin N acyltransferase (AT) and outline AT mechanisms. A: AT catalyses amidohydrolase and 6-APA acyl-transferase reactions;^228,234B: View from mature AT:6-APA complex (PDB 2X1E);²⁴¹C: Outline mechanism of AT autoproteolysis; D: The amidohydrolase activity of AT; E: The acyltransferase activity of AT. The oxyanion hole forming residues are in green and the Ntn residue is in blue. R² can be a variety of aliphatic or aromatic side chains. See Fig. 43 and 62 for related Ntn enzyme mechanisms.

Like the N-acetylornithine:glutamic acid acetyltransferase OAT2²³⁸ and the pantetheine hydrolase ThnT,²³⁹ from the clavulanic acid and thienamycin biosynthesis pathways (Sections 5.11 and 6.4.3, respectively), AT is a member of the N-terminal nucleophile (Ntn)-hydrolase superfamily.^73,74 Despite the often low sequence homology among Ntn-hydrolases, the enzymes share a conserved αββα-core structure (Ntn fold, Fig. 16) and conserved catalytic residues.⁷⁴ AT, like other Ntn-hydrolases, is produced as a single polypeptide chain (40 kDa inactive precursor);²⁴⁰ autocatalytic cleavage of the Gly102–Cys103 peptide bond produces the active α,β-heterodimeric protein (11 kDa α- and 29 kDa β-subunits) in which the N-terminal cysteine of the β-subunit acts as a nucleophile during AT-catalysis (Fig. 15B).^240,241 It is notable that the penicillin acylases (which are used industrially to produce 6-APA from penicillin G) and cephalosporin acylases (which produce 7-ACA)^242–245 (Section 4.10) are also members of the Ntn-hydrolase superfamily.⁷⁴

Fig. 16 Members of the N-terminal nucleophile (Ntn) family of hydrolases involved in β-lactam antibiotic biosynthesis. A: α,β-Monomer of the mature acyl coenzyme A:isopenicillin N acyltransferase (AT, PDB 2X1E) from Penicillium chrysogenum;²⁴¹B: 1 subunit/2 chains of ornithine acetyltransferase (OAT-2, PDB 2YEP, Section 5.11) from Streptomyces clavuligerus;⁴⁹¹C: Monomer of the uncleaved ThnT T282C (PDB 3S3U, Section 6.4.3) from S. cattleya.⁵⁸⁷ Note the conserved αββα-core structure.⁷⁴ The N-terminal nucleophilic residue (or T282C variant for C) is shown in space-filling mode.

A recently reported crystal structure of P. chrysogenum AT²⁴¹ (Fig. 16A) is of interest because it suggests how AT can accommodate a variety of substrate side chains, and, importantly, why it does not accept cephalosporins.²⁴¹ Modelling studies based on the AT structure suggest that there is insufficient room to accommodate the cephalosporin carboxylate which is located at the sp²-hybridised C-2 of the cephem nucleus, but there is room for the penicillin carboxylate which is located at the sp³-hybridised C-2 of the penam nucleus (Fig. 5 in ref. 241). It is proposed that the structural work may guide protein engineering studies aiming at producing AT variants that accept cephalosporin substrates.²⁴¹

4.6 Isopenicillin N/penicillin N epimerase

Prokaryotic isopenicillin N epimerase (CefD) catalyses the reversible epimerisation of the L-α-aminoadipoyl side chain in isopenicillin N to the D-α-aminoadipoyl side chain in penicillin N, in a pyridoxal phosphate dependent manner, likely via a mechanism involving imine/enamine type intermediates, as proposed for related epimerases (Fig. 17).^246–248 CefD catalyses an essential step in cephalosporin biosynthesis because isopenicillin N is not a substrate for DAOCS/DACS (CefEF) or DAOCS (CefE), and thus may have acted to enable a “branch point” between the biosyntheses of cephalosporins and the hydrophobic penicillins (Fig. 9). The CefD encoding gene (cefD) is located adjacent to that encoding for DAOCS (cefE) in both S. clavuligerus (Fig. 10) and Amycolatopsis lactamdurans.^249–251 The two genes are co-transcribed so as to, it is proposed, ensure that penicillin N availability is not limiting for DAOCS-catalysed ring expansion.^249,251 The net result is proposed to shift the isopenicillin N/penicillin N equilibrium towards the latter. The conversion of ACV analogues to cephalosporins using partially purified S. clavuligerus extracts provides evidence that the prokaryotic epimerase may tolerate modifications in the penam nucleus.²⁵²

Fig. 17 Comparison of the proposed mechanisms for conversion of the L-aminoadipoyl side chain of isopenicillin N into the D-aminoadipoyl side chain of penicillin N in bacterial and fungal cephalosporin-producers. The isopenicillin N epimerase in cephalosporin-producing bacteria (i.e. CefD) is pyridoxal phosphate (PLP)-dependent; an alternative isopenicillin N epimerisation system (comprising both CefD1 and CefD2) operates in fungi.^248,253 CefD1 is homologous to long chain acyl-CoA synthetases⁷¹⁰ and CefD2 is homologous to acyl-CoA racemases.^711,712 Penicillin N release is proposed to be catalysed by an unidentified thioesterase.⁷¹³ The fungal isopenicillin N epimerisation system is likely similar to those involved in phytanic acid and ibuprofen racemisation.^253,714

The fungal epimerisation system (e.g. from A. chrysogenum) is proposed to be different from that involved in prokaryotic BLA biosynthesis and to involve three enzyme-catalysed steps (Fig. 17): activation of isopenicillin N to isopenicillinyl-CoA by a CoA-ligase (encoded by the cefD1gene), epimerisation to penicillinyl-CoA by an acyl CoA-epimerase (encoded by the cefD2 gene), and finally hydrolysis to penicillin N by, as yet, an unidentified thioesterase.^253,254

4.7 Deacetoxycephalosporin C synthase (DAOCS) and deacetylcephalosporin C synthase (DACS)

In the eukaryote A. chrysogenum, penicillin N ring expansion (to give deacetoxycephalosporin C, DAOC) and the subsequent C-3′-hydroxylation (to give deacetylcephalosporin C, DAC) are catalysed by a single bifunctional enzyme DAOCS/DACS (encoded by cefEF) (Fig. 9).^255–258 In contrast, in the prokaryote S. clavuligerus, the ring expansion and hydroxylation reactions are catalysed by two distinct but closely related enzymes (DAOCS and DACS, respectively),^259–262 suggesting that a gene duplication event may have occurred during evolution which resulted in two homologous enzymes catalysing different reactions.^262,263 In fact DAOCS displays a low level of DACS activity and vice versa.²⁶² Further, the genes encoding for DAOCS (cefE) and DACS (cefF) from S. clavuligerus show extensive homology to cefEF from A. chrysogenum¹⁸⁷ consistent with a proposed horizontal transfer of the entire pathway from bacteria to fungi.^{101,148,264–266}

4.7.1 Structure of DAOCS. In 1998, a crystal structure for recombinant S. clavuligerus DAOCS was reported²⁶⁷ – the first for a 2OG-dependent oxygenase.²⁶⁸ Subsequently reported DAOCS structures include complexes with penicillin G (substrate analogue),²⁶⁹ 2OG and succinate.^270–272 Recombinant DAOCS occurs in solution predominately as a monomer, which is in equilibrium with a trimeric form, and crystallises as a trimer.^267,268 The trimer is formed in part by the interpenetration of the flexible C-terminal arm (residues 308–311) of one monomer into the active site of the neighbouring monomer in a cyclic fashion. The trimeric oligomerisation in this crystal form fixes the C-terminal arm and partially blocks access to the active site hindering formation of complexes of DAOCS bound to penicillins. Addition of Fe^II or 2OG²⁶⁷ (or attaching an N-terminus His-tag,²⁶⁹ or truncation of the C-terminus^270,273) shifts the equilibrium toward the monomeric form that appears to be the active form in solution. The sequence and length of the C-terminus is important in determining substrate selectivity, e.g. truncation of the C-terminus alters its selectivity from penicillin N, which has a polar side-chain, towards penicillin G, which has a hydrophobic side chain.^268,270 The overall fold of DAOCS is similar to that of IPNS and shows a double-stranded β-helix core with flanking helices (Fig. 18A). The DAOCS structures reveal the presence of a number of arginine residues within the active site, with Arg258 (part of a conserved RXS motif) being involved in 2OG binding (Fig. 18E).^267,268 The equivalent RXS motif in IPNS (Arg281, and Ser283, in case of A. nidulans IPNS) binds the valinyl-carboxylate of ACV.^189,274 2OG ligates to Fe^II in the active site of DAOCS in a bidentate manner via its 2-oxo group and one of its 1-carboxylate oxygens.^267,268 His183, Asp185 and His243 provide the 2-His-1-Asp iron coordinating motif; the sixth ligand in the DAOCS:2OG complex is water (Fig. 18C). For other 2OG oxygenases (Fig. 19), this water is proposed to be displaced upon binding of the “prime” substrate, allowing oxygen binding and initiation of the “typical” catalytic cycle (involving oxidative decarboxylation of 2OG by reaction with oxygen to generate carbon dioxide, succinate and a reactive oxidising species that mediates substrate oxidation) (Fig. 20).^24,268,275 However, analysis of certain DAOCS crystal structures has led to the proposal that the binding sites of the prime substrate (penicillin N) and co-substrate (2OG) overlap (Fig. 18E).²⁶⁹ Therefore, a modified mechanism for DAOCS-catalysed ring expansion has been proposed in which an Fe^IV=O intermediate is generated by the oxidative decarboxylation of 2OG to succinate which leaves the active site before penicillin N binding and reaction (see below).^269,273 For more detailed descriptions of structural and mechanistic studies on 2OG oxygenases, see ref. 24, 25.

Fig. 18 Views from crystal structures of DAOCS. A: view from the overall fold of a monomer of DAOCS:Fe^II:2OG:penicillin G (a substrate analogue) complex showing the conserved double stranded β-helix (in yellow) (PDB 1UOB); B to G: views from a DAOCS:Fe^II complex (B, PDB 1RXF), a DAOCS:Fe^II:2OG complex (C, PDB 1RXG), a DAOCS:Fe^II:penicillin G complex (D, PDB 1UOF), a DAOCS:Fe^II:2OG:penicillin G complex (E, PDB 1UOB), a DAOCS:Fe^II:DAOC complex (F, PDB 1UOG) and a DAOCS:Fe^II:succinate complex (G, PDB 1UO9). Fe^II (in orange), conserved active site residues, substrates/co-factors and products are shown. Note that in case of (E) the reported analysis of the electron density reveals that the penicillin G and 2OG binding sites overlap, leading to the proposal of an atypical 2OG oxygenase mechanism (Fig. 19).²⁷³

Fig. 19 Views from crystal structures of Fe^II-dependent oxidases/oxygenases involved in β-lactam biosynthesis. In all cases, note the highly conserved 2-histidine-1-carboxylate motif. A: IPNS active site with ACV and NO (an O₂ analogue) complexed (PDB 1BLZ);¹⁸⁹B: DAOCS active site with penicillin G and 2OG showing the reported substrate/cosubstrate overlapping binding sites (PDB 1UNB);²⁷³C: Clavaminic acid synthase (CAS, Section 5.3) active site with NO and its substrate for hydroxylation (PDB 1GVG);⁴²⁰D: The carbapenem synthase (CarC, Section 6.3.3) active site with the substrate analogue, N-acetyl-proline (NAP), and 2OG bound (PDB 1NX8).⁵⁴² Overall, the figure illustrates how a conserved binding chemistry can be used to mediate different types of oxidation reactions. *The carbon(s) to be oxidised.

Fig. 20 Proposed outline mechanisms of DAOCS/DACS catalysis. In the ring-expansion of penicillin N to DAOC, an Fe^IV=O intermediate is proposed to be generated in a “typical” 2-electron oxidation process characteristic for 2OG oxygenases. In the modified proposed mechanism for ring expansion as catalysed by DAOCS, succinate is proposed to leave the active site before penicillin N binding.^269,273 Further validation is required for this mechanism. The 3β-hydroxy-3α-methylcepham (in the dashed box) is formed by DAOCS/DACS catalysis as a minor “shunt” product in case of the ring expansion of penicillin N. Its yield is increased by the introduction of a deuterium atom at the C-2 position of penicillin N (due to a kinetic isotope effect) indicating a common intermediate prior to the branch point.^299,300 In the DAOC hydroxylation by DACS, another 2-electron process results in the C-3′ hydroxylation of DAOC to produce DAC.

4.7.2 Mechanism of DAOCS catalysis. Whilst the DACS hydroxylation reaction is typical of 2OG-dependent oxygenases²⁴ (though unusually occurs at an allylic position), the DAOCS catalysed oxidative ring expansion (Fig. 20) is unprecedented in enzymology. It does however have precedent in an analogous chemical process for ring expansion, in which a penicillin β-sulfoxide is converted into a cephalosporin where the (pro-S) β-methyl group of a penicillin is incorporated into the dihydrothiazine/cephem ring.²⁷⁶ Both in vitro²⁷⁷ and in vivo^278,279 studies have shown that the β-methyl group of penicillin N forms the endocyclic C-3 of DAOC during cephalosporin biosynthesis (Fig. 20). Early mechanistic studies were carried out on DAOCS/DACS and used labeled valines/penicillin N,¹⁴³ but recent structural and engineering work has mainly targeted DAOCS.^{268–271,280–293} Here we summarise important results – see Section 4.11.1 and ref. 143 for a more detailed review of this work.

Cell-based feeding studies employing (3-pro-R)-valine, with a stereogenic methyl group at the position that forms the endocyclic C-3 of cephalosporin C, reveal complete loss of stereochemical integrity during ring expansion (Fig. 21IIA).^294–296 In contrast, experiments employing (3-pro-S)-valine, with a stereogenic methyl group at the position that forms the C-3′ exocyclic methylene of cephalosporin C, reveal (partial) retention of stereochemistry (Fig. 21IIB).^294–296 The observed loss of stereochemistry during the ring expansion process suggested the existence of a penicillin N-derived β-methylene radical intermediate, which can undergo rotation before ring expansion. This suggestion is supported by biomimetic studies.^297,298 Although it may be related to the allylic nature of the intermediate, the retention of stereochemistry in the hydroxylation reaction is of interest with regard to 2OG oxygenases catalysis because it implies that hydroxylation can occur via a process not involving a “long-lived” methylene radical (e.g. direct insertion or a very fast recombination process). Incubation of [3-²H]-penicillin N and DAOCS/DACS has also shown that, in the ring expansion process, one hydrogen from the β-methyl group is lost prior to the hydrogen at C-3 position.^299,300 Notably, the incubation resulted in an increased level of 3-β-hydroxy-3-α-methylcepham,^299,300 a normally minor by-product, which has been previously isolated from A. chrysogenum fermentation broths.³⁰¹ A summary of mechanistic proposals for DAOCS/DACS is given in Fig. 20.

Fig. 21 I: The biocatalytic versatility of DAOCS/DACS catalysis. For further examples see ref. 304 and ref. 280 for wildtype and engineered DAOCS-catalysed reactions, respectively. The reaction of DAOCS/DACS with its natural substrate is boxed; II: Feeding studies employing valine labeled with a chiral methyl group at the (3-pro-R) position reveal loss of stereochemical integrity during ring expansion;^294–296 experiment using valine with a chiral methyl group at the (3-pro-S) position imply retention of stereochemistry.^294–296 Where reported, the approximate ratios of observed products are given below the structures; in the case of the [4-²H]-exomethylene analogue reaction (H), the ratio of products varies over time.^313,317

4.7.3 DAOCS substrate analogues studies. Substrate analogue studies on DAOCS/DACS have indicated that, like IPNS, it is rather promiscuous (Fig. 21I). The promiscuity of IPNS (Section 4.4) may be related to the fact that its prime substrate is complexed to the active site Fe^II. In contrast, DAOCS/DACS, like other 2OG oxygenases,³⁰² has the potential to uncouple “prime” substrate oxidation from that of 2OG.²⁶⁸ Whilst wildtype DAOCS is only able to efficiently utilise 2OG and 2-oxoadipate, as 2-oxoacid co-substrates, the R258Q variant can accept shorter 2-oxoacids (e.g. 2-oxo-3-methylbutanoate) as co-substrates.²⁷¹ Examples of prime substrate analogues accepted by DAOCS/DACS include both modifications to the penam/cepham/cephem nucleus and to the C-6 side chain of penicillin N analogues (Fig. 21I). These early studies were important because they encouraged subsequent protein and pathway engineering studies on DAOCS with a view to optimising its acceptance of substrates with hydrophobic side chains (Section 4.11.1).

Here we highlight some of the interesting substrate analogue studies. Mutant strains of A. chrysogenum are reported to produce a cephem-3-aldehyde, probably originating from DAC oxidation.^303,304 DAOCS/DACS has been shown to catalyse the oxidation of DAC to the corresponding aldehyde (Fig. 21IA).³⁰⁴ Incubation of 6α-methylpenicillin N with DAOCS/DACS resulted in the formation of the corresponding cephem aldehyde (Fig. 21IA); however, incubation with DACS resulted only in uncoupled turnover of 2OG.³⁰⁵ While IPNS accepts peptide substrates with both L- and D- N-terminal δ-(α-aminoadipoyl) side chains, DAOCS/DACS does not accept isopenicillin N (with δ-(L-α-aminoadipoyl) side chain), but does accept the adipoyl-penicillin analogues (Fig. 21IB).^{252,306–308} A 2β-methyl-penam derivative (equivalent to penicillin N without the 2α-methyl group) was converted by DAOCS/DACS into a C-3-demethylated DAOC (Fig. 21IC).¹⁴³ As for IPNS, chemically interesting substrate analogues studies have comprised alkene (Fig. 21ID and E),³⁰⁹ C-3 exo-alkene (Fig. 21IF, H and I),^310–314 and cyclopropane (Fig. 21IG)³¹⁵ containing substrate analogues. In case of the exomethylene substrate analogue reaction, DAC was the main observed product (Fig. 21IF);³¹² however, in case of the [4-²H]-exomethylene analogue reaction, DAC, a 3β-spiro-epoxide cepham, and an aldehyde were also detected (Fig. 21IH), suggesting that the first irreversible event in this conversion involves Fe^IV=O addition to the double bond (and not C-4 hydrogen abstraction).^{310,313,316,317} In case of the Z-3-ethylidene substrate analogue reaction, two diastereomeric alcohols were detected (Fig. 21II).³¹⁸ In the case of a cyclopropane-containing analogue, a rearranged product was obtained (Fig. 21IG).³¹⁵

4.8 Introduction of the 7α-methoxy and formamido-groups

Two types of 7α-functionalised cephalosporins have been identified: the cephamycins,^319–321 which contain a 7α-methoxy group, and (some) cephabacins, which contain a 7α-formylamino group (Fig. 8).³²² Both sub-classes show potent antibacterial activity and enhanced resistance to serine β-lactamases.^323,324 Whilst the origin of the 7α-formylamino group is unknown, pioneering labeling studies by Abraham et al. demonstrated that the methyl group and the oxygen of the 7α-methoxy group are derived from methionine and O₂, respectively.³²⁵ The methoxylation reaction, in cell-free extracts of S. clavuligerus, was shown to be dependent on Fe^II, 2OG and S-adenosylmethionine (SAM).³²⁶O-Carbamoyl DAC was found to be a better substrate for 7α-methoxylation than DAOC; DAC was apparently not methoxylated.³²⁶ These results suggest that the sequence of events in cephamycin C biosynthesis proceeds via conversion of DAOC to DAC to O-carbomoyl DAC^327,328 to cephamycin C³²⁶ (Fig. 9). Hood et al. provided evidence that the methoxylation reaction is a two-step process, by isolation and characterisation of the 7α-hydroxy-O-carbamoyl DAC using S. clavuligerus extracts but excluding SAM.³²⁹ Bioinformatic analyses of the cephamycin biosynthesis gene cluster have revealed that cmcI and cmcJ likely encode for a methyl-transferase and a 7α-hydroxylase, respectively.³³⁰ However, as yet, no activity has been reported for the recombinant enzymes, and it may be that they work in concert.^331,332 A crystal structure has been reported for CmcI in complex with SAM (Fig. 22).¹⁰⁹ CmcI was initially annotated as a cephalosporin hydroxylase,³³³ but no evidence for Fe^II/2OG binding could be found.¹⁰⁹ Despite its low sequence identity (8–23%) to SAM-dependent methyltransferases, the CmcI structure is related to that of catechol O-methyltransferase with overlapping nucleotide-binding residues.¹⁰⁹ Attempts to obtain a CmcI structure in complex with cephalosporin C or O-carbamoyl DAC have, to date, been unsuccessful and potential substrates were therefore docked (Fig. 7 in¹⁰⁹); the docking results show a binding mode for 7α-hydroxy-O-carbamoyl DAC which would allow methyl transfer to the C-7 hydroxyl group.¹⁰⁹

Fig. 22 Enzymes responsible for the C-7 methoxylation of cephalosporins. A: Enzymes proposed to catalyse C-7 methoxylation of O-carbamoyl-DAC; B and C: Views from a crystal structure of CmcI¹⁰⁹ from S. clavuligerus (PDB 2BR4). B: The hexameric quaternary structure of CmcI; C: A CmcI monomer. The S-adenosylmethionine (SAM) cofactor, shown in space-filling mode, is employed by CmcI to methylate the 7α-hydroxyl group of the O-carbamoyl-DAC to give cephamycin C and S-adenosylhomocysteine (SAHC). Magnesium(II) is in yellow. The 7α-hydroxyl group is likely introduced by a 2OG oxygenase, CmcJ.³³⁰ The geometry of the SAM/Mg^II binding site is similar to that found in cathechol O-methyltransferases.¹⁰⁹ R = δ-(D-α-aminoadipoyl).

4.9 Modifications at the cephem C-3′ position

Studies on the biosynthesis of cephalosporins and cephamycins substituted at the 3′-position have revealed DAC as a common precursor.³³⁴ However, in case of the cephamycins, the sequence of events leading to 3′-hydroxyl and C-7 modifications is as yet unclear. Docking studies employing a CmcI crystal structure (Section 4.8) suggest that, in the biosynthesis of cephamycin C, the 3′-hydroxyl carbamoylation may occur prior to the C-7 methoxylation.¹⁰⁹

Acetylation of the C-3-hydroxymethyl function of DAC, the final step in cephalosporin C biosynthesis, is catalysed by acetyl-CoA:DAC O-acetyltransferase (DAC-AT/CefG, encoded by cefG, Fig. 9). The presence of an acetyltransferase in the cephalosporin C biosynthesis pathway was initially predicted on the basis of accumulation of DAC in the culture broth of A. chrysogenum mutants.^335,336 The CefG activity was first demonstrated in cell-free extracts of A. chrysogenum where DAOC was converted into cephalosporin C with incorporation of [¹⁴C]-label when acetyl-1-¹⁴C-CoA was added.³³⁷ Subsequent studies enabled purification of CefG^338–341 and the cloning of cefG.^342,343 Initially CefG was reported as a heterodimer of two subunits with molecular weights of ∼27 and 14 kDa;³⁴⁰ however, the enzyme has been shown later to be active only as a single polypeptide chain (i.e. is not processed into subunits).³⁴¹ Crystal structures of CefG (Fig. 23) reveal that the enzyme belongs to the α/β-hydrolase class of acetyltransferases.³⁴⁴ Sequence³³⁴ and structural comparisons indicate that CefG is similar to homoserine-O-acetyltransferases. The CefG active site (Fig. 23B) has a conserved His362-Asp333-Ser149 catalytic triad involved in the acetyl group transfer; Ser149 is proposed to be acetylated by acetyl coenzyme A followed by transfer of the acetyl group to DAC through a tetrahedral intermediate stabilised in an oxyanion hole formed by the backbone –NHs of Leu59 and Met150 (Fig. 23C).³⁴⁴

Fig. 23 Deacetylcephalosporin C acetyltransferase (CefG). A: View of the overall structure of the CefG monomer; B: View from the active site of CefG showing the catalytic triad (Asp333-His362-Ser149) with acetylated Ser149, the oxyanion hole forming residues (Leu59 and Met150) and part of deacetylcephalosporin C (DAC);³⁴⁴C: Proposed mechanism for CefG catalysis.³⁴⁴D: View of the monomer of the GCN5-related acetyltransferase Orf14 involved in clavulanic acid biosynthesis.⁴⁸⁰

Carbamoylation of the C-3-hydroxymethyl function of DAC gives the intermediate O-carbamoyl-DAC en route to cephamycin C (Fig. 9), likely in an ATP-dependent manner, as catalysed by CmcH (encoded by cmcH).^327,328 CmcH sequences analyses imply the presence of a conserved carbamoyl-phosphate-binding amino acid motif.¹⁰⁸

Cephamycins A and B, and oganomycins (A, B, GA and GB, Fig. 8) contain a 3′-p-coumaric acid ester group. The biosynthesis of p-coumaric acids can occur either from L-phenylalanine or L-tyrosine (Fig. 24). The production of oganomycins can be stimulated by the addition of p-coumaric acid to S. oganonensis cultures.³⁴⁵ To our knowledge, there have been no studies on the enzymology of this modification nor on the enzymology of the formylamino moiety at C-7 of some cephabacins¹¹⁵ (Fig. 8). Identification of the genes encoding for the proteins responsible for both modifications should open up the way for the biosynthetic introduction of new functionalities to the cephem nucleus.

Fig. 24 Putative biosynthetic routes to the p-coumaric- and caffeic acid moieties at the 3′-position of cephamycins (Fig. 8). In Streptomyces maritimus, L-phenylalanine is deaminated to trans-cinnamic acid in a reaction catalysed by phenylalanine ammonia lyase (PAL) EncP.^715,716 In plants, trans-cinnamic acid can then be hydroxylated at the C-4 position, as catalysed by cinnamate 4-hydroxylase (C4H), but no homologous enzyme has been described in prokaryotes. Alternatively, p-coumaric acid can be derived from L-tyrosine through the activity of the tyrosine ammonia lyase Sam8.⁷¹⁷ The presence of a caffeic acid ester in C-2081 X⁶⁹⁷ (Fig. 8) suggests that a p-coumaric acid hydroxylase, similar to Sam5,⁷¹⁷ may be present in the producer strains S. heteromorphus and S. panayensis.

4.10 Production of semisynthetic penicillins and cephalosporins

Strains used industrially for the production of penicillins and cephalosporins (i.e. P. chrysogenum and A. chrysogenum, respectively) secrete their products into the fermentation medium. The recovery process starts by filtration of the fermentation broth followed by solvent extraction steps. Whilst penicillins are mainly purified via multiple re-extractions in buffer and solvent at varying pH (employing counter-current chromatography), cephalosporins are mainly purified via solid-phase extraction due to their hydrophilic nature.²

6-APA, 7-ACA and, to a lesser extent, other C-7 deacylated cephalosporin intermediates are key intermediates for the production of semi-synthetic penicillins and cephalosporins, respectively, and are obtained by removal of the acyl side chain of the purified fermented products. The originally used chemical deacylation method³⁴⁶ has now been superseded by enzymatic processes for economic and ecological reasons.² The enzymatic deacylation of penicillins to 6-APA is an established process, employing acylases, including from genetically manipulated E. coli or Bacillus megaterium.⁶⁷ Like the biosynthetic enzyme AT, penicillin acylases are Ntn-hydrolases (Section 4.5). The D-α-aminoadipoyl side chain of the cephalosporins is less susceptible to direct in vitro enzyme-catalysed deacylation than that of the hydrophobic side chain of fermented penicillins,⁷⁰ and efforts have been made to develop efficient enzyme-catalysed routes in vitro or in cells. A two-step enzymatic method has been developed wherein cephalosporins are first subjected to oxidative deamination by a D-amino acid oxidase (DAO, e.g. from Trigonopsis variabilis), then the resulting α-keto-adipoyl-7-ACA undergoes decarboxylation (induced by the peroxide released during the oxidase reaction) into glutaroyl-7-ACA. Addition of small quantities of peroxide following DAO treatment ensures full decarboxylation of α-keto-adipoyl-7-ACA into 7-ACA⁷² (DAO is a flavin oxido-reductase which catalyses the oxidation of D-amino acids to their corresponding keto-acids and ammonia). In the second step, the side chain of glutaroyl-7-ACA is deacylated by a glutarylamidase (e.g. from Pseudomonas diminuta,^347,348 for review on industrial aspects of the enzymatic cleavage of cephalosporins, see ref. 70, 72). It is noteworthy that the glutarylamidases from P. diminuta also exhibit a low cephalosporin C acylase activity (5% of their glutarylamidase activity).³⁴⁹ Enhancing the activity of these amidases for “direct” (one step) splitting of the C-7 side chain of cephalosporins, employing various protein engineering strategies, may help optimise the enzymatic deacylation processes; however, a more efficient approach may be to directly ferment cephalosporins with hydrophobic side chains (Section 4.11).

4.11 Protein and metabolic engineering studies on the cephalosporin/cephamycin biosynthesis pathway

Classical strain improvements technologies (via iterative random strain mutagenesis and screening for improved productivity) have resulted in significant improvements in industrial fermentation titres of BLAs;^102,350,351 however, these technologies do not (at least normally) result in the directed production of new products. One goal of “post-genomic” metabolic and protein engineering studies on BLA biosynthesis is to achieve rational optimisation of the production of BLAs of choice.^352–355

4.11.1 Deacetoxycephalosporin C synthase engineering studies. The employment of various protein engineering strategies (e.g. directed evolutionary approaches, rational redesign, and computational approaches^356–360) to BLA biosynthesis is a potential strategy for enhancing BLA production and the generation of new BLAs. S. clavuligerus DAOCS has been subjected to protein engineering studies aimed at improving its capacity to accept penicillins with hydrophobic side chains in order to generate hydrophobic cephalosporins (for a review, see ref. 280). The availability of crystal structures for DAOCS (Section 4.7) was important as they enabled engineering strategies aimed at developing DAOCS variants with expanded substrate selectivities. Here we summarise selected results. The C-terminus of DAOCS is involved in ordering substrate binding and maintaining the coupling of penicillin and 2OG oxidation.^268,269 On the basis of computational, biochemical, and structural analyses, C-terminal modifications have been proposed to be useful for improving DAOCS activity for ring expansion of unnatural substrates.^268,288 The DAOCS/DACS R308 residue (A. chrysogenum) is located in close proximity to the C-terminus and was proposed to play a role in controlling substrate entry/product release.³⁶¹ Comprehensive substitution at the R308 position has resulted in variants displaying significant improvement in ring-expanding penicillins with hydrophobic side chains.³⁶¹ The R308-based variants with aliphatic side chains are proposed to improve hydrophobic interactions with substrate analogues;³⁶¹ for example, with penicillin G, the R308I variant has a 7-fold increase in relative activity.³⁶¹ The DAOCS (S. clavuligerus) N304-based variants (i.e. N304A, N304L, N304K, and N304R) also display improved activity for conversion of hydrophobic penicillins compared to wildtype DAOCS.^284,288 In the case of penicillin V, the N304L variant has shown an ∼4-fold increase in relative activity.²⁸⁸ In the cases of carbenicillin and phenethicillin, the C281Y I305M double variant has ∼13- and 11-fold increases in relative activity, respectively, likely due to enhanced substrate binding.²⁸² Overall, the engineering studies on DAOCS are pioneering because, although not yet used in production strains, they demonstrate how the selectivity of a 2OG oxygenase can be modified by “semi-rational” structure-based protein engineering.

4.11.2 Metabolic engineering studies. Directed genetic modifications of BLA producing strains have resulted in considerable enhancements in production levels (for reviews, see ref. 102, 362–364). Pathway extension, by heterologous expression of new genes in host strains, can result in the formation of new analogues that can help the purification and/or enzymatic deacylation processes. Despite the intensive strain improvement programs in A. chrysogenum for increasing cephalosporin C production, the relative productivity of cephalosporin C producing strains is still significantly lower than that for penicillin production in P. chrysogenum.³⁶⁵ The reason for the high BLA production levels in industrial strains of P. chrysogenum is proposed to be (at least in part) due to an elevated copy number and high transcription levels of the penicillin biosynthesis gene cluster.^366–369 Therefore, pathway engineering/extension efforts have been directed towards genetically engineering P. chrysogenum for the direct fermentation of cephalosporins.
4.11.2.1 Cephalosporins production in engineered A. or P. chrysogenum. Overexpression of another copy of cefG in A. chrysogenum resulted in a 2–3 fold increase in cephalosporin C production, consistent with the effect of “gene dosage” on cephalosporin production.^370,371 Overexpression of cefE (S. clavuligerus) simultaneously with disruption of cefEF gene in A. chrysogenum results in the production of DAOC at levels comparable to the total BLA biosynthesised by A. chrysogenum; the resulting DAOC can then be isolated and deacylated (as discussed in Section 4.10) to give 7-ADAOCA (Fig. 25A).^372,373

Fig. 25 Metabolic and protein engineering studies aimed at modifying the penicillin-cephalosporin pathways to directly ferment cephalosporins with side chains useful for efficient purification or direct deacylation (e.g. by Pseudomonas diminuta glutarylamidase). A: production of 7-adipoyl-cephems (boxed) via adipoyl-6-aminopenicillanic acid (adipoyl-6-APA) in engineered P. chrysogenum strains fed with adipic acid; B: Proposed pathway for the production of cephems with hydrophobic or clinically useful side chains employing engineered DAOCS variants which have altered substrate specificities^280,283 (see ref. 280 for review). Adipoyl-7-ACCCA: adipoyl-7-amino-3-carbamoyloxymethyl-3-cephem-4-carboxylic acid.

Overexpression of cefD1, cefD2, cefEF and cefG (all from A. chrysogenum) in P. chrysogenum lacking AT activity results in the production and release of significant amounts of DAC into the culture broth as well as intracellular accumulation of cephalosporin C (for the acronyms used in this section, see Table 1).³⁷⁴

4.11.2.2 Adipoyl-cephalosporin production in engineered P. chrysogenum. Together with feeding adipic acid into the culture medium, overexpression of cefE (from S. clavuligerus)^375,376 or cefEF with or without cefG (from A. chrysogenum)^376,377 in P. chrysogenum results initially in the production of adipoyl-6-APA and subsequently, depending on the cloned activities, adipoyl-7-aminodeacetoxycephalosporanic acid (adipoyl-7-ADAOCA, in case of cefE), adipoyl-7-aminodeacetylcephalosporanic acid (adipoyl-7-ADACA in the case of cefEF), or adipoyl-7-ACA (in the case of cefEF and cefG, Fig. 25A). The removal of the adipoyl side chain, which is more susceptible to glutarylamidase cleavage than the α-aminoadipoyl side chain of natural cephalosporins,³⁷⁷ yields the corresponding 7-ADAOCA, 7-ADACA, and 7-ACA, respectively (Fig. 25A). Overexpression of the cefEF gene (A. chrysogenum) and the cmcH gene (S. clavuligerus) in P. chrysogenum, together with feeding of adipic acid, results in the production of adipoyl-7-amino-3-carbamoyloxymethyl-3-cephem-4-carboxylic acid (adipoyl-7-ACCCA, Fig. 25A).³⁷⁸ The precursor adipic acid fed to the culture medium is proposed to be introduced into the modified biosynthesis pathway via the activity of a promiscuous acyl-CoA ligase (encoded by aclA in P. chrysogenum).³⁷⁹ Overexpression of cefT (A. chrysogenum, which encodes for a cephalosporin C transporter belonging to the major facilitator superfamily) in the adipoyl-7-ACCCA-producing P. chrysogenum variant strain results in an ∼2-fold increase in cephalosporin production with a concomitant decrease in penicillin formation.³⁷⁹

Taken together, these results suggest that cephalosporin production by engineered P. chrysogenum strains is possible but may be limited by factors including the ability of the fungus to secrete “unnatural” cephalosporins.

5 Clavulanic acid biosynthesis

Clavulanic acid (CA) and its β-hydroxypropionyl derivative (isolated as its benzyl ester)³⁸⁰ are unusual amongst the family of naturally-occurring clavams because they have the (5R)-ring junction stereochemistry, rather than the (5S)-stereocentre observed in other naturally-occurring clavams (Fig. 26). Previous reviews have summarised pioneering labeling and enzymology studies on CA biosynthesis.^{32,33,143,381} Here we summarise the current state of knowledge on the pathways to CA and the (5S)-clavams, with an emphasis on the role of the enzymes involved. From the perspective of biosynthesis of the β-lactam ring itself, the work on CA biosynthesis was pioneering because it established that nature can synthesise β-lactams by routes other than the oxidative cyclisation chemistry employed by IPNS (Section 4.4).

Fig. 26 Clavams isolated from natural sources.³² In addition to clavulanic acid (CA, with (5R)-stereochemistry) and its β-hydroxypropionyl derivative 1, Streptomyces clavuligerus produces the clavams 6–9 (with the (5S)-stereochemistry).⁷¹⁸ CA has also been isolated from S. jumonjinensis⁴⁴⁴ and S. katsurahamanus.⁴⁴⁵ The N-acyl derivatives 3–5 of clavaminic acid 2 accumulate in a S. clavuligerus mutant blocked in CA production. The hydroxyethyl clavam 10, valclavam 12, and Tü 1718B were isolated from S. antibioticus ssp. antibioticus Tü 1718 which does not produce CA. Valclavam was initially assigned the structure 11^437,438 but this was reassigned to the regio-isomer 12.⁴³⁹ Tü 1718B (boxed) is a degradation product of 12.^439,440 The structure of valclavam is related to that of clavamycins A–F which are produced by S. hygroscopicus.³²

The biosynthesis route to CA can be divided into two stages (Fig. 27A) – those leading to (3S,5S)-clavaminic acid, which are reasonably well established, and those involved in the conversion of clavaminic acid to CA, which are much less well understood. A combination of labeling,^382–388in vitro enzyme assays and genetic studies^388,389 established the six steps leading to clavaminic acid. Fig. 28 displays the genes in the three clusters involved in the biosynthesis of CA and clavams in S. clavuligerus; Table 2 shows the proposed roles of the proteins encoded by these genes.

Fig. 27 Proposed biosynthetic pathways leading to (5R)-clavulanic acid and (5S)-clavams in Streptomyces clavuligerus (A), and those leading to (5S)-clavams in S. antibioticus (B).^32,381,516 Note that clavaminic acid (boxed) acts as a branch point between clavulanic acid and clavam biosynthesis. Selected cofactors and co-substrates are shown. The hydrogens to be removed during the CAS-catalysed oxidations are in red. Putative (i.e. not isolated) intermediates are in grey.

Fig. 28 The three gene clusters encoding for proteins involved in the biosynthesis of clavulanic acid (CA) and clavams in Streptomyces clavuligerus. A: “CA” gene cluster; B: “Paralog” gene cluster; C: “Clavam” gene cluster.^353,381,503 For the (predicted) role of the (putative) proteins encoded by the genes shown, see Table 2 and Fig. 27A. The potential involvement of orf24 to orf28 in CA/clavam biosynthesis is to be assessed.⁵⁰⁶ The pcbR gene (CA gene cluster) is part of the penicillin/cephalosporin gene cluster (Fig. 10).

Table 2 Genes constituting the “Clavulanic acid” gene cluster (^a), the “Paralog” gene cluster (^b) and the “Clavam” gene cluster (^c) in Streptomyces clavuligerus and the (predicted) roles of the (putative) proteins that they (may) encode for. The predicted number of amino acid residues (AA) for each of the (putative) proteins is shown. Proteins with an experimentally assigned biochemical function are in bold

Gene	AA	(Proposed) function of encoded protein
ceas1^b/2^a	547/586	Carboxyethylarginine synthase (CEAS).^390–392
bls1^b/2^a	527/513	β-Lactam synthetase (βLS).^381,382
cas1^c/2^a	324/352	Clavaminic acid synthase (CAS).^420
pah1/2^b,a	457/313	Proclavaminic acid amidinohydrolase (PAH).^412
oat1/2^b,a	452/399	Ornithine acetyltransferase (OAT2).^238
cvm6^c/6P^b	452/507	Pyridoxal phosphate-dependant aminotransferase.^503
cvm7^c/7P^b	1114/915	Transcriptional regulator.^503,515
oppA1^a	564	ABC-type dipeptide transport system.^475
claR^a	432	Transcriptional activator.
cad^a	247	Clavaldehyde dehydrogenase (CAD).^451
cyp^a	407	Cytochrome P-450.^454
fd^a	71	Ferredoxin.^454
orf12^a	458	β-Lactamase-like protein.^454
orf13^a	340	Amino acid (derivative) export pump.^438,442
orf14^a	339	Acetyltransferase.^438,442,480
oppA2^a	562	Peptide transport protein.^475,476
orf16^a	401	N-Acetyltranferase.^453
gcas^a	529	ATP-dependent ligase (GCAS).^459
pbpA^a	484	Penicillin-binding protein.^719
pbp2^a	696	Penicillin-binding protein.^719
orf20^a	389	Cytochrome P-450.^506
orf21^a	201	RNA polymerase σ factor.^508
orf22^a	571	Two-component system histidine kinase.^506
orf23^a	270	Two-component system response regulator.^506,507
orfL^b	229	Transcriptional regulator.^489,504
orfK^b	277	Hydrophobic ligand-binding SRPBCC domain of Micromonospora echinospora.^477,494
orfJ^b	200	Dihydrofolate reductase.^477,494
orfI^b	356	LysR-type transcriptional regulator.^477,494
orfH^b	416	Arabinose efflux permease.^477,494
orfG^b	364	Kinase.^477,494
orfF^b	253	Amino acid permease.^477,494
orfE^b	102	Amino acid permease.^477,494
orfD^b	351	Dehydratase.^477,494
orfC^b	393	Aminotransferase.^477,494
orfB^b	126	Amino acid biosynthesis regulator.^477,494
orfA^b	390	Hydroxymethyltransferase.^477,494
cvmG^c	240	Unknown.^503
cvmP^c	687	Arginine deiminase.^503
cvmH^c	318	Hydrolase.^503
cvm13^c	390	β-Aspartyl peptidase.^503
cvm12^c	445	Transcriptional regulator.^503
cvm11^c	216	Efflux protein.^503
cvm3^c	170	Flavin reductase.^503
cvm2^c	151	Isomerase.^514,515
cvm1^c	344	Aldo/keto reductase.^503
cvm4^c	328	Acetyltransferase.^503
cvm5^c	394	Oxido-reductase.^503
cvm9^c	182	Transcriptional regulator.^503
cvm10^c	308	Kinase.^514,515

---

5.1 Carboxyethylarginine synthase

The first step in clavulanic acid (CA) biosynthesis is catalysed by carboxyethylarginine synthase (CEAS), a member of the thiamine diphosphate (ThDP)-dependent superfamily.^390–392 CEAS catalyses the condensation of L-arginine with glyceraldehyde-3-phosphate to give N²-(2-carboxyethyl)arginine (CEA), the β-amino acid precursor for β-lactam formation (Fig. 27A, 29). Crystal structures of CEAS reveal its structural homology to pyruvate decarboxylase and related ThDP-dependent enzymes (Fig. 30).^391,392 The CEAS active site is located at a dimer interface across which ThDP is bound in a V-shaped conformation, as is characteristic for many ThDP-dependent enzymes. CEAS is unusual amongst the superfamily of ThDP-utilising enzymes because it catalyses a C–N bond formation in a mechanism proposed to occur via Michael-type reaction (Fig. 29). Analogous to other ThDP-dependent enzymes (e.g. pyruvate decarboxylase), the initial step in CEAS-catalysis is proposed to involve reaction of a ThDP ylide with the aldehyde group of glyceraldehyde-3-phosphate. Following proton transfer (from the aldehyde-derived carbon at the position α- to the thiamine ring), it is proposed that elimination of the β-hydroxyl group occurs to give an enol, which undergoes further elimination of the γ-phosphate group to give a ThDP-bound acryloyl-group.³⁹⁰ The latter then undergoes Michael reaction with the α-amino group of L-arginine. Finally, hydrolysis of the ThDP-bound Michael addition product occurs to give CEA. This mechanism is consistent with the available labeling and crystallographic information,^390–392 though kinetic analyses are needed to verify the proposed intermediates in catalysis. It should also be noted that, at least in our hands, isolated CEAS is not highly active; it cannot be ruled out that there is an alternative to glyceraldehyde-3-phosphate as the 3C-component of the reaction. One interesting proposal arising from the structural work on CEAS is that the acid–base catalysis required for the reaction is derived either from ThDP itself or from a species (e.g. hydroxide) generated from substrates – this is because of an apparent lack of suitably-positioned general acid/base side chains in the active site of CEAS.^391,392 To date, the CEAS-catalysed reaction represents a novel and unique route to β-amino acids. Engineering studies on CEAS may be productive with respect to producing modified β-amino acids for conversion into β-lactams.

Fig. 29 Proposed outline mechanism for carboxyethylarginine synthase (CEAS). The enol(ate)-ThDP intermediate has been provisionally assigned in crystalline CEAS following soaking with DL-glyceraldehyde-3-phosphate (G3P) (Fig. 30E).³⁹¹ Details of proton transfers and eliminations are uncertain.^390,391

Fig. 30 Carboxyethylarginine synthase (CEAS), a thiamine diphosphate (ThDP)-dependent enzyme. A: The tetrameric oligomerisation of CEAS.³⁹² The CEAS active site is located at a dimer interface across which ThDP (shown in space-filling mode) is bound; B and C: Views of CEAS (B) and yeast pyruvate decarboxylase (C) monomers reveal their structural homology. Note the similar three-domain structure of each monomer;³⁹²D: View from subunit A of a crystal structure of CEAS, with ThDP bound in V-shaped conformation, following soaking with 5-guanidinovaleric acid (GVA) (PDB 2IHV).³⁹¹ GVA is a nonreactive L-arginine analogue that lacks an α-amino group;³⁹¹E: A view from a crystal structure of CEAS (PDB 2IHU) showing the putative trapped enol(ate)-ThDP intermediate (TDP-G3P, Fig. 29) that has been provisionally assigned in subunit C following soaking with DL-glyceraldehyde-3-phosphate (G3P).³⁹¹ The structures (D and E) suggest (partially) overlapping binding sites for the substrates D-G3P and L-arginine. This is consistent with the proposed CEAS mechanism (Fig. 29). Mg^II is shown as a yellow sphere. *Residues from a neighbouring monomer.

5.2 β-Lactam synthetases in clavam and carbapenem biosynthesis

The second step in the clavam biosynthesis pathway is catalysed by the first identified β-lactam synthetase (β-LS) which catalyses the cyclisation of N²-(2-carboxyethyl)arginine (CEA) to give the monocyclic β-lactam deoxyguanidino-proclavaminic acid (DGPC, Fig. 27A).^393,394 β-LS is closely related to the carbapenam synthetases (CarA^395,396 and ThnM,³⁹⁷Fig. 31) which are involved in carbapenem biosynthesis via catalysing the cyclisation of (2S,5S)-carboxymethylproline (t-CMP) to yield the (2S,5S)-carbapenam nucleus (Section 6.3.2). In this section, we consider structural, mechanistic, and substrate analogue studies for both β-LS and CarA/ThnM.

Fig. 31 Partial sequence alignment for known and putative β-lactam synthetases involved in the biosynthesis of clavams (β-LS), carbapenems (CarA and ThnM), and monobactams (TbIS). Catalytically important residues are highlighted. The highlighted lysine residue is proposed⁷²⁰ to assist in ring cyclisation via stabilisation of the proposed “tetrahedral” intermediate (in case of clavams and carbapenams, Fig. 33). The highlighted Tyr-Glu dyad (in case of clavams and carbapenems) is proposed to deprotonate the amine involved in intramolecular β-lactam formation.⁷²¹ α-Helices (cyan cylinders) and β-strands (red arrows) represent the assigned secondary structure of β-LS (PDB 1MB9).³⁹⁹

5.2.1 Structure of β-LS and CarA, and their evolutionary relation to Asn-B. Both β-LS and CarA/ThnM are related to asparagine synthetase type A (Asn-B) which catalyses the ATP-dependent transfer of ammonia from glutamine to aspartate to give asparagine and glutamate (Fig. 33). Crystal structures have been reported for Asn-B,³⁹⁸ β-LS^399,400 and CarA³⁹⁵ (Fig. 32). Together with mechanistic studies (Fig. 33), these structures provide insights into how a primary metabolic enzyme (Asn-B) that catalyses primary amide bond formation has (likely) evolved into β-lactam synthetases with different substrate selectivities. A β-lactam synthetase may also be involved in the biosynthesis of other β-lactams (e.g. tabtoxin, Section 7.3).

Fig. 32 Views from crystal structures of β-lactam synthetases. A: Views from the structures of the β-lactam synthetases CarA and β-LS (PDB 1Q19 and 1MB9, respectively) in comparison to that of asparagine synthetase (PDB 1CT9)³⁹⁸ showing the N- and C-terminal domains; B: View from a crystal structure of β-LS (PDB 1MBZ) showing the acyl-adenylate N²-(2-carboxymethyl)arginine-AMP (CMA-AMP) trapped species generated by reaction of ATP with the substrate analogue N²-(2-carboxymethyl)arginine.³⁹⁹ The latter is one carbon shorter than the natural substrate; thus, CMA-AMP does not undergo cyclisation to give the highly strained 3-membered ring; C: View from a crystal structure of CarA (PDB 1Q19)³⁹⁵ complexed with the substrate (2S,5S)-5-carboxymethylproline (t-CMP) and an ATP analogue α,β-methyleneadenosine-5′-triphosphate (AMP-CPP) with t-CMP positioned in apparently “productive” conformation for adenylation and subsequent β-lactam formation.

Fig. 33 How nature converts modified amino acids/peptides into β-lactams employing oxidative and non-oxidative mechanisms. A: Outline proposed mechanisms for CarA and β-LS (involving intramolecular nucleophilic attack) in comparison to that of Asn-B (which involves intermolecular nucleophilic attack),³⁹⁸ showing the common mechanism of carboxylate activation, via adenylation, followed by formation of a tetrahedral intermediate; B: Outline mechanism for the oxidase isopenicillin N synthase showing the enzyme-bound monocyclic ferryl intermediate; R = δ-(L-α-aminoadipoyl).

Crystallographic analyses of Asn-B³⁹⁸ (Fig. 32) reveal that it comprises two domains: an N-terminal nucleophile (Ntn) glutaminase domain responsible for production of ammonia from glutamine, and a C-terminal synthetase domain which catalyses formation of β-aspartyl-AMP. It is proposed that ammonia travels from the N-terminal hydrolase domain through a tunnel that links the N- and C-terminal domains where it reacts with the β-aspartyl-AMP to form asparagine (Fig. 33A).^398,401 Structural comparisons of the β-lactam synthetases and Asn-B (Fig. 32) reveal that the formers have maintained the characteristic two domain fold;⁴⁰⁰ however, as β-lactam ring formation is an “intramolecular” process, it does not require the release of ammonia from glutamine, rendering the glutaminase reaction of the N-terminal domain redundant. This analysis is consistent with the replacement of the, catalytically important, nucleophilic Cys1 of Asn-B with Phe1 in β-LS, together with the presence of nine additional N-terminal residues in β-LS which occupy the corresponding glutamine binding pocket in Asn-B.⁴⁰⁰ In CarA, the Cys1 of Asn-B is replaced by a serine which occupies a similar position; however, other important residues for glutamine binding are missing, reflecting the lack of a glutamine binding pocket.³⁹⁵ In the C-terminal domain of all three enzymes, the substrate-carboxyl group is activated, via ATP-mediated adenylation, at a largely conserved active site. Subsequent nucleophilic attack by NH₃/secondary amine yields the amide/β-lactam product (respectively, Fig. 33A).^{395,398–400}

Whilst both β-LS and Asn-B crystallise as dimers, CarA crystallises as a tetramer.³⁹⁵ For each of the three proteins: (i) the N-terminal domain consists of two antiparallel β-sheets that form a sandwich, flanked on each side by (relatively) short α-helices; (ii) the C-terminal domain comprises (11–14) α-helices surrounding a 5-stranded parallel β-sheet; and (iii) the active site is located in a cleft in the C-terminal domain formed by 4 β-strands and 5 α-helices; however, the β-LS/CarA substrate binding cleft is relatively extended to accommodate CEA (β-LS)/t-CMP (CarA) which are larger than aspartic acid (Asn-B).

5.2.2 Mechanisms of β-LS and CarA. The mechanisms of β-LS and CarA involve the ordered binding of ATP/Mg^II and CEA_β-LS/t-CMP_CarA to the apoenzyme (Fig. 32B and C, respectively); the reaction proceeds via the formation of an acyl-adenylate intermediate followed by intramolecular 4-exo-trig-cyclisation to give a β-lactam, likely via a tetrahedral intermediate (Fig. 33A).^395,399,400 Pyrophosphate is the last product to be released.^395,396,399 In the case of β-LS, cyclisation of CEA occurs to give the monocyclic β-lactam deoxyguanidino-proclavaminic acid, which undergoes further oxidative reactions (catalysed by CAS) to give the bicyclic clavam structure. In carbapenem biosynthesis, t-CMP, the substrate for CarA (and likely ThnM), undergoes CarA/ThnM-catalysed cyclisation to yield the (2S,5S)-carbapenam nucleus.^29,395,397

The close structural and mechanistic relation between Asn-B, β-LS and CarA/ThnM coupled to their different substrate specificities suggests that it may have been relatively easy to evolve different types of β-lactam (or indeed potentially other lactams/lactones) synthetases based on the Asn-B structural platform. Whether β-LS and CarA/ThnM evolved separately from Asn-B or from one another is unknown.

5.2.3 Substrate analogues studies. Considering the studies on engineering of DAOCS²⁸⁰ and CarB^402–404 to alter their substrate specificities (Section 4.7, 6.3.1, respectively), it would seem likely that β-LS/CarA/ThnM would be amenable to similar approaches. β-LS has been shown to accept an epimeric mixture of 2-methyl-CEA to give a mixture of epimers of 2-methyl-DGPC (Fig. 34A).^405,406 Notably, CarA has been shown to catalyse the ring closure of at least three of the four possible stereoisomers of t-CMP⁴⁰⁷ (to give the corresponding carbapenams), t-carboxyethylproline (to the corresponding γ-lactam), and some α-amino-diacids (e.g. glutaric acid, α-aminoadipic acid, and α-aminopimelic acid to the corresponding lactams) (Fig. 34B and C).³⁹⁶ Recently, the ability of CarA to form β-lactam bicyclic systems other than carbapenams from 6- and 7-membered ring analogues of t-CMP (i.e. (2S,6S)-6-carboxymethyl-pipecolic acid (t-CMPi) and (2S,7S)-7-(carboxymethyl)-azepane-2-carboxylic acid, respectively) has been demonstrated (Fig. 34B and C).⁴⁰⁴ Furthermore, CarA has also been reported to accept t-CMP derivatives methylated at C-2, C-3, C-4, C-5 or C-6 (to give the corresponding methylated carbapenams) (Fig. 34B and C).⁴⁰² Interestingly, the C-1 and C-6 dimethyl-substituted-(3S,5S)-carbapenams, carbacephams, and their 7-membered analogues generated by CarA catalysis, are reported to be of higher stability (with respect to hydrolysis) than the unsubstituted (3S,5S)-carbapenam.⁴⁰² CarA also has been demonstrated to exercise stereoselective bias. One example is the case of the CarA-catalysed reaction of a 2 : 1 mixture of (7R)-7-methyl-t-CMPi and (7S)-7-methyl-t-CMPi, where selective conversion of the former to give (4S,6S,7R)-7-methyl-carbacepham was observed (Fig. 34C).⁴⁰⁴ Further, in the CarA-catalysed reaction of a 1 : 1 mixture of (4R)-4-methyl-t-CMP and (4S)-4-methyl-t-CMP, preferential (1 : 2) conversion of the latter to give (1S,3S,5S)-1-methyl-carbapenam was observed (Fig. 34B and C).⁴⁰² Combined with studies on CarB/ThnE (Section 6.3.1), these investigations suggest that the carbapenem biosynthesis pathways are amenable to engineering.

Fig. 34 The biocatalytic versatility of the β-lactam synthetases β-LS (A)^405,406 and CarA (B, C).^396,402,404 In the case of substituted t-CMP substrate analogues, CarA can accept derivatives methylated at C-2, C-3, C-4, C-5 or C-6 to produce the corresponding β-lactams.⁴⁰² CarA can also accept at least three of the four possible isomers of CMP.³⁹⁶ Compounds which are reported to be of higher stability compared to the unsubstituted carbapenams are boxed.^402,404 The compounds with asterisked stereocentre represent the preferred stereoisomer at that position for CarA catalysis.^402,404 The incubation of t-carboxyethylproline, a substrate analogue of t-CMP, with ATP/Mg^II/CarA results in the formation of a γ-lactam (shown in a dashed box).³⁹⁶

5.3 The trifunctional role of CAS in clavaminic acid biosynthesis

Four steps are involved in the conversion of deoxyguanidino-proclavaminic acid (DGPC) to clavaminic acid; three of these steps are catalysed by a single 2OG oxygenase, clavaminic acid synthase (CAS, Fig. 27A). It is not unprecedented that 2OG oxygenases catalyse sequential two electron oxidations, as exemplified by the bifunctional role of DAOC/DAC synthase in cephalosporin biosynthesis (Section 4.7), the TET enzymes in the hydroxylation and formylation of human DNA,⁴⁰⁸ and in the 3-stage oxidation of a methyl group catalysed by the plant gibberellin C-20 oxidase.^409–411 However, in case of CAS, two of the three oxidative steps are separated from the first by the action of proclavaminic acid amidino-hydrolase (PAH, Fig. 27A), which is structurally and mechanistically unrelated to CAS.⁴¹² Another feature of CAS catalysis is that the three reactions that it catalyses involve three different types of oxidative reactions (i.e. hydroxylation, oxidative cyclisation, and desaturation) (Fig. 27A).

5.3.1 CAS-catalysed reactions and their mechanisms. The initial CAS-catalysed step involves the stereospecific hydroxylation of DGPC at C-3 to give (2S,3R)-guanidino-proclavaminic acid (GPC) in a hydroxylation reaction that is typical of the 2OG oxygenases (Fig. 35A).⁴¹³ GPC is neither a substrate nor an (efficient) inhibitor of CAS. However, rather than mutate CAS to evolve a new function, it appears that the Mn^II-dependent PAH⁴¹² may have been “hijacked” to enable forward evolution of the pathway via catalysing the hydrolysis of the guanidino group of GPC to give proclavaminic acid (PCA), which is the substrate for the next CAS-catalysed reaction (Fig. 27A).^414–416 Thus, the proposed role of PAH is to “mutate” the side chain of GPC such that further CAS-catalysed oxidations can continue the pathway. To our knowledge, this is highly unusual in studied biosynthetic pathways – such swapping “in and out” of enzymes is likely common in catabolism (e.g. proteolysis) and surely must occur in other biosynthetic pathways.

Fig. 35 Proposed outline mechanisms for the three clavaminic acid synthase (CAS)-catalysed reactions (i.e. hydroxylation (A), oxidative ring cyclisation (B1/2), and desaturation (C), respectively). The hydroxylation reaction is separated from the other two reactions by the activity of proclavaminic acid amidino-hydrolase which catalyses the conversion of guanidino-proclavaminic acid to proclavaminic acid (Fig. 38). For the cyclisation reaction, the alternative B2 mechanism has been proposed in part on the basis of computational studies;⁷²² it involves (1) oxidation of the hydroxyl group of proclavaminic acid to give an O-radical, (2) retro-aldol-like decomposition of the O-radical to an aldehyde and a C-centered radical, which is stabilised by the captodative effect, (3) abstraction of a hydrogen atom from the C-4 (pro-S) position of the C-centered radical by the Fe^III–OH species to yield an azomethine ylide, and (4) 1,3-dipolar cycloaddition to the ylide with the aldehyde acting as a dipolarophile. There is synthetic precedent for the B2 mechanism as in the synthesis of oxapenams via 1,3-dipolar cycloaddition reactions of aldehydes and ketones.⁷²³

The second CAS-catalysed step comprises dihydroclavaminic acid formation by a desaturative ring closure reaction, which involves removal of the β-lactam 4′-pro-S hydrogen and the hydrogen of the hydroxyl group at C-3 of PCA (Fig. 35B). Evidence that the dihydroclavaminic acid oxazolidine ring is formed prior to formation of the exocyclic alkene of clavaminic acid came initially from NMR analyses of the mixture of products resulting upon incubation of PCA and partially purified CAS, which revealed the presence of a minor resonance at δ ∼ 5.4 corresponding to the C-5 proton of 2,8-dihydroclavaminic acid.^415,417 Repeating the experiment with C-3-deuterated PCA enabled the isolation of the saturated bicyclic [²H₂]-dihydroclavaminic acid, because of the operation of a primary isotope effect which slows down the exocyclic double bond formation. Upon incubation with CAS, the purified dihydroclavaminic acid was converted to clavaminic acid demonstrating the intermediacy of dihydroclavaminic acid in the cyclisation of PCA to clavaminic acid. Non-deuterated dihydroclavaminic acid was subsequently shown to be converted into clavaminic acid using recombinant CAS.^416,418 The CAS-catalysed oxidative cyclisation reaction that forms the bicyclic ring structure of dihydroclavaminic acid is reminiscent of the IPNS-catalysed cyclisation of ACV (Section 4.4). However, the CAS-catalysed cyclisation is a two-electron oxidation in which only one ring is formed, whereas the IPNS reaction is a four-electron bicyclisation reaction.

The third CAS-catalysed step (i.e. dihydroclavaminic acid desaturation to give clavaminic acid) occurs (at least predominantly) with loss of hydrogens via a syn-elimination (Fig. 35C).⁴¹⁹

5.3.2 CAS structure. The overall fold of CAS is similar to that of other 2OG oxygenases with a distorted double stranded β-helix core fold flanked by α-helices (Fig. 36A).^24,420,421 Crystal structures of CAS have been reported in complex with Fe^II, 2OG and substrates/substrate analogues (DGPC, PCA and N-α-acetylarginine (NAA), Fig. 36B, C and D, respectively).^420,421 Comparison of CAS structures to those of other 2OG oxygenases (Fig. 19) has revealed that (i) the 2-His-1-carboxylate Fe^II binding motif characteristic of 2OG oxygenases can have a glutamyl rather than an aspartyl Fe^II-ligand⁴²⁰ (aspartate may be generally preferred because its side chain is more rigid than that of glutamate, which may reflect the need of CAS for flexibility to perform its trifunctional catalytic role);⁴²⁰ and (ii) the RXS motif which has been observed in IPNS/DAOCS (Fig. 18E) is absent in CAS; instead the 2OG 5-carboxylate is bound to the hydroxyl of Thr172 and the guanidino side chain of Arg293 (Fig. 36B). These observations were important because they revealed the presence of more than one structural family of 2OG oxygenases.²⁴

Fig. 36 Views from crystal structures of clavaminic acid synthase (CAS).^420,421A: The CAS monomer (PDB 1DRT); B: The active site of CAS (PDB 1GVG) showing the binding sites for Fe^II, the CAS-substrate for hydroxylation (deoxyguanidino-proclavaminic acid, DGPC), nitric oxide (NO, an O₂ analogue), and 2OG; C: The active site of CAS (PDB 1DRT) with proclavaminic acid bound; D: Superimposition of the active sites of CAS:Fe^II:2OG:DGPC:NO complex (PDB 1GVG) and that with the substrate analogue N-α-acetylarginine (NAA, without NO but with a water molecule ligating Fe^II). Note: the superimposed structures (D) reveal that the C1-carboxylate oxygen of 2OG (green) undergoes a “flip” to the other possible coordination site (cyan) in the presence of NO (B).^420,421

The reported structures also provide detailed insights into how CAS catalyses its three oxidative reactions.^420,421 Interestingly, studies using nitric oxide (NO), as an O₂ analogue, reveal the potential for the 2OG 1-carboxylate to rearrange its iron binding position (Fig. 36D). Thus, in the absence of NO as in case of the NAA complex (Fig. 36C), a water molecule occupies the coordination position trans to His279 and adjacent to the substrate C–H bond to be oxidised. Coordination of O₂ to this site should generate the Fe^IV=O in an appropriate position to effect substrate oxidation.^24,420,421 However, in the presence of NO, the 2OG 1-carboxylate moves to this position enabling NO to bind trans to His144 (Fig. 36B). If O₂ binds similarly to NO, the subsequently generated Fe^IV=O would be in an inappropriate position for substrate oxidation, and it must rearrange.^410,411 Alternatively, if a 5-coordinate persuccinate intermediate is formed (Fig. 20),⁴²² simultaneously with CO₂ loss, it may collapse to produce the Fe^IV=O in the (apparently) correct position for substrate oxidation. The incorporation of oxygen from water as well as from O₂ into the alcohol products in some 2OG oxygenase-catalysed hydroxylations may reflect accessibility of a five coordinate Fe^IV=O to react with water.⁴²³ Therefore, the position of O₂ binding in CAS, and in other 2OG oxygenases, is unclear.

5.3.3 Substrate analogues studies. From a biocatalytic perspective, relatively little has been reported on CAS. However, CAS trifunctional role coupled to some substrate analogue studies suggests that it will be amenable to protein engineering studies to elicit new reactions. Incubation of the γ-lactam analogue of PCA with CAS yields a mixture of the saturated and unsaturated γ-lactam products – compounds that are not trivial to access synthetically (Fig. 37A).⁴²⁴ Replacement of the β-lactam rings of DGPC and PCA by an acetamido-group resulted in the following observations:^413,425,426 (i) incubation of N-α-acetyl-L-ornithine with CAS gave a mixture of the C-3 alcohol and the C-2/C-3-E-alkene products (alcohol : alkene, 3 : 1) (Fig. 37C); (ii) in contrast, the β-lactam deoxy-PCA gave predominantly the alkene product (alkene : alcohol, >10 : 1) (Fig. 37B); and (iii) incubation of NAA, like DGPC, gave only the alcohol product (Fig. 37D).

Fig. 37 The biocatalytic versatility of clavaminic acid synthase (CAS).^{413,424–426} The reactions of CAS with its natural substrates are boxed.

Arginine hydroxylation is a common feature in the biosynthesis of some secondary metabolites. Subsequent to the work on CAS, VioC, a 2OG oxygenase from S. vinaceus, has been found to catalyse the hydroxylation of L-arginine to give (2S,3S)-3-hydroxyarginine during viomycin biosynthesis.⁴²⁷ Arginine hydroxylation, likely catalysed by 2OG oxygenases, plays a role in the biosynthesis of streptothricin F^428,429 and the heptapeptide antibiotic K-582 which consists of two components with threo-γ-hydroxy-L-arginine as a constituent of both.⁴³⁰

5.4 Proclavaminic acid amidino hydrolase (PAH)

Proclavaminic acid amidino hydrolase (PAH, Orf4)⁴¹² catalyses the hydrolysis of the guanidino side chain of guanidino-proclavaminic acid (GPC) to give proclavaminic acid and urea (Fig. 38A). Disruption of pah leads to loss of CA production, reflecting its obligatory requirement in CA biosynthesis.⁴³¹ PAH is a member of the arginase family of enzymes which hydrolyse arginine to ornithine and urea; however, PAH does not accept arginine as a substrate.⁴³²

Fig. 38 Mechanism and crystal structures of amidino-hydrolases.^432,433A: Proposed outline mechanism for guanidino-proclavaminic acid (GPC) hydrolysis by proclavaminic acid amidino-hydrolase (PAH).⁴³²B: Monomer of PAH from S. cattleya⁴³² (PDB 1GQ7); C: Monomer of the arginase from Bacillus caldovelox⁴³³ (PDB 3CEV); D: Superimposed active site views of the two amidino-hydrolases showing the residues involved in catalysis (blue for arginase), the conserved Mn^II-binding site (violet for PAH, and yellow for arginase), and the guanidino moiety of the substrate (L-arginine, the substrate for arginase). Note that neither Ser44 nor Arg11 of PAH, which are proposed to be responsible for binding the carboxylate and hydroxyl groups of GPC, respectively, are conserved in the arginase active site.^422,423

5.4.1 PAH structure and mechanism. PAH exists in solution and crystallises as a hexamer.⁴³² Like the closely related arginases (which catalyse arginine hydrolysis to ornithine and urea), the overall fold of PAH comprises a central core of β-strands surrounded by α-helices (Fig. 38B and C). The PAH active site contains two asymmetrically arranged Mn^II ions at ∼3.3 Å apart and bridged by a water molecule (or hydroxide ion) which is proposed to be responsible for hydrolytic attack at the carbon atom of the guanidino group of the substrate (Fig. 38A and D).⁴³² The conserved active-site arrangement of PAH is shown (Fig. 38D) in comparison to that of an arginase from Bacillus caldovelox.⁴³³

PAH catalyses the hydrolysis of the substrate analogues deoxyguanidino-proclavaminic acid, N-acetyl-L-arginine and (3R)-hydroxy-N-acetyl-L-arginine, albeit at a lower rate than the natural substrate GPC.⁴³² These results, together with comparison of the structures of PAH and the B. caldovelox arginase, reveal that the lack of arginine binding to PAH may be rationalised by (i) the presence of the more polar amino group rather than an amide and/or the absence of the C-3 hydroxyl group;⁴³² (ii) Arg11 (from an adjacent PAH monomer) and Ser44 are proposed to be suitably positioned to bind the PAH-substrate carboxylate and hydroxyl groups, respectively. Neither Ser44 nor Arg11 of PAH is conserved in the arginase from B. caldovelox.⁴³³

Comparing the crystal structures of PAH and CAS also reveals an apparent convergence of the arginine and serine residues involved in binding the C-3 alcohol of their substrates. In a crystal structure of CAS complexed with proclavaminic acid (Fig. 36C), the product of PAH, the hydroxyl of proclavaminic acid is positioned to hydrogen-bond the side chains of Arg297 and Ser134 (CAS numbering). In PAH, the side chains of Arg11 and Ser44 (Fig. 38D) are similarly positioned and Ser44 is a likely candidate for substrate hydroxyl binding.⁴³²

5.5 Clavaminic acid, a branch point in (2R,5R) clavulanic acid and (3S,5S) clavam biosynthesis

A series of labeling studies^434–436 has provided evidence that clavaminic acid is a branch point between the (5R)-CA and (5S)-clavams biosynthesis pathways in S. clavuligerus (Fig. 27A). For valclavam^437,438 and Tü 1718B (Fig. 26)^439,440 biosynthesis, labeling studies and the determination of the presence of PAH and CAS in S. antibioticus ssp antibioticus Tü 1718B^{436,441–443} also provided evidence for the common steps with CA biosynthesis up to the point of clavaminic acid. A biosynthetic path to the clavams has been proposed⁴⁴¹– a modified version is shown in Fig. 27A. As genomic sequences become available for more β-lactam producers, it is likely that differences between (3R,5R)-CA and (3S,5S)-clavam producers will become clearer.

5.6 From clavaminic acid to clavaldehyde: Stereoisomerisation and oxidative deamination of clavaminic acid

It seems that CA producers (S. clavuligerus, S. jumonjinesis,⁴⁴⁴ and S. katsurahamanus⁴⁴⁵), but not (5S)-clavam only producers, possess the ability to catalyse the “double-epimerisation” of (3S,5S)-clavaminic acid to (3R,5R)-clavaldehyde. The mechanism of this unprecedented process is unknown; one possibility is via C-9 oxidation, followed by oxazolidine ring opening to form an iminium ion and ring closure to give the (3R,5R)-clavulanate system (Fig. 39).^446,447 Evidence has been reported for the exchange of the C-8-hydrogen during CA biosynthesis from clavaminic acid.⁴⁴⁸S. clavuligerus growth in an ¹⁸O₂-rich atmosphere results in the incorporation of ¹⁸O into both the oxazolidine ring and the allylic hydroxyl group of CA.⁴⁴⁹ The incorporation of ¹⁸O into the oxazolidine ring likely results from CAS-activity (CAS-catalysed hydroxylation of deoxoguanidinoproclavaminic acid results in ∼70% incorporation of ¹⁸O from ¹⁸O₂ into proclavaminic acid).⁴²⁵ The incorporation of ¹⁸O into the allylic group of CA rules out the possibility of a water-mediated deamination and suggests the involvement of an oxygenase (possibly a P450 enzyme, see below) in the conversion of clavaminic acid to CA. However, this is unlikely to involve an oxygenase-catalysed conversion of clavaminic acid to (3S,5S)-clavaldehyde because there is only evidence for the production of (3R,5R)-clavaldehyde.⁴⁵⁰ The (3R,5R)-stereochemistry of clavaldehyde was assigned by synthesis.^450,451 Clavaldehyde, the aldehyde of which may act as an electron sink for decarboxylation, was suggested as an intermediate in (5S)-non-carboxylated clavam biosynthesis on the basis of the observation that proclavaminic acid is a common precursor of both CA and (5S)-non-carboxylated clavams.^434,435,452 It is less certain that (3R,5R)-clavaldehyde is involved in the biosynthesis of the (5S)-clavams.⁴⁵⁰

Fig. 39 Possible outline mechanism for the double epimerisation process involved in clavulanic acid (CA) biosynthesis.^446,447 Note that the oxygen at C-9 of CA originates from O₂.⁴⁴⁹ R₁ could be H, glycyl or N-acetyl-glycyl.

Presently, it is unclear as to how many genes are essential for the conversion of clavaminic acid to CA. The products of the (orf10–orf17) genes of the CA biosynthesis gene cluster are proposed to be involved in the conversion of (3S,5S)-clavaminic acid to (3R,5R)-clavaldehyde.^29,32,453 The genes orf10 (cyp) and orf11 (fd) apparently encode for a cytochrome P-450 and a ferredoxin, respectively. Inactivation of cyp resulted in complete loss of CA production, while production was only reduced in the case of an fd-disrupted mutant, suggesting that cyp is essential and fd is beneficial, but not absolutely required, for CA biosynthesis.^{381,453–455} It is possible that Orf10 acts on a derivative of clavaminic acid (e.g. N-glycyl-clavaminic acid or N-acetyl-glycyl-clavaminic acid) to introduce the O₂-derived oxygen at C-9 (Fig. 39). Disruption of orf15 or orf16 blocks CA production and leads to accumulation of N-acetyl-glycyl-clavaminic acid and N-glycyl-clavaminic acid.⁴⁵³ Accumulation of clavaminic acid was observed in case of a claR disrupted mutant; claR encodes for a transcriptional regulatory protein that is proposed to regulate genes involved in the late steps of CA biosynthesis (e.g. oppA1, cyp and cad).^456–458

5.7 N-Glycyl-clavaminic acid synthetase, GCAS (Orf17)

S. clavuligerus mutants with defects in orf17 are unable to produce detectable levels of CA.⁴⁵³ The gene product of orf17 (N-glycyl-clavaminic acid synthetase, GCAS/Orf17) catalyses the ATP-dependent ligation of the amino group at C-9 of clavaminic acid to the carboxylate moiety of glycine to give N-glycyl-clavaminic acid (Fig. 40). Under the same conditions, N-acetyl-glycine did not react suggesting that CA biosynthesis pathway may proceed via clavaminic acid to N-glycyl-clavaminic acid and then to N-acetyl-glycyl-clavaminic acid.⁴⁵⁹ GCAS shares homology with proteins from the ATP grasp fold superfamily (which includes, e.g., E. coli biotin carboxylase).^460,461 The mechanism of GCAS, similarly to other members of the ATP grasp fold superfamily,^460,461 likely proceeds via an enzyme-bound O-glycyl-phosphate intermediate (Fig. 40). GCAS seems to be selective for the combination of glycine and clavaminic acid;⁴⁵⁹ this is unlike some members of the ATP grasp fold superfamily (e.g. the D-alanine-D-alanine ligases, which, despite being specific for D-amino acids, do display some flexibility in their substrate specificity⁴⁶²). It is proposed that N-glycyl-clavaminic acid and N-acetyl-glycyl-clavaminic acid are intermediates en route to CA, but there is no firm evidence yet to support this.⁴⁵⁹

Fig. 40 Proposed outline mechanism for the ATP-grasp family enzyme Orf17 that catalyses the condensation of the amino group of clavaminic acid to the (activated) carboxylate of glycine.⁴⁵⁹

5.8 From clavaldehyde to clavulanic acid: clavulanic acid dehydrogenase

The last step in CA biosynthesis in S. clavuligerus is catalysed by CA dehydrogenase (CAD/Orf9), an NADP(H)-dependent oxido-reductase that catalyses the (reversible) conversion of (3R,5R)-clavaldehyde into (3R,5R)-CA (Fig. 41B).⁴⁵¹ Sequence analyses of CAD reveal homology to proteins belonging to the short chain dehydrogenases/reductases (SDR) family.^463–466

Fig. 41 Mechanism and structural views of the NADPH-dependent reductase clavulanic acid dehydrogenase (CAD). A: Overview of the labeling studies⁴⁷⁴ aiming to reveal the origin of the C-9 hydrogens of clavulanic acid with respect to its precursor ornithine/arginine; B: Proposed reaction mechanism for CAD showing a possible proton relay involving water molecules. R = adenosine 2′-phosphate-5′-diphosphate;⁴⁶⁹C: A CAD monomer showing the Rossmann fold, which is characteristic of dinucleotide-binding enzymes (two repeats of βαβαβ motifs, one shown in yellow and the other in blue, each binds one NADPH nucleotide);⁴⁶⁹D: View from the CAD active site showing the possible proton relay involving water molecules (B).

5.8.1 Structure of CAD. CAD, similarly to other SDR enzymes, exists as a dimer and tetramer in solution. A CAD monomer comprises a single domain possessing the Rossmann fold (two repeats of βαβαβ motif) characteristic of dinucleotide-binding enzymes,⁴⁶⁷ and overall consists of seven β-strands, creating a parallel β-sheet, surrounded by eight α-helices (Fig. 41C). The conserved catalytic triad/tetrad, which defines the reductase activity of SDR enzymes,⁴⁶⁵ is formed by residues Ser142, Tyr155, Lys159 and Asn115;^466,468 a conserved water molecule associated with Asn115 is also present (Fig. 41C) and may have functional importance (see below). The absolute dependence on NADP(H) in preference to NAD(H)⁴⁶⁹ is due to the lack of CAD for a conserved acidic residue, which forms hydrogen bonds to the 2′- and 3′-hydroxyl groups of the adenine ribose, in case of NAD(H)-preferring enzymes.⁴⁷⁰ NADP(H)-preferring enzymes have two basic residues (Arg/Lys) which bind the 3′-phosphate;^463,464 in case of CAD, only one residue, Arg39, is present. Thus, CAD is classified as a member of the cP2 subfamily.⁴⁶³ A conserved glycine-rich motif (Gly-X₃-Gly-X-Gly) forms hydrogen bonds to the pyrophosphate of NADP(H). The nicotinamide ring binding pocket has a hydrophobic floor and polar residues on the opposite side to enable substrate binding and aldehyde reduction. Lys159 forms bifurcated hydrogen bonds to 2′- and 3′-hydroxyl groups of the nicotinamide ribose, coordinating the cofactor as well as helping to lower the pK_a of the catalytic base Tyr155.^469,471 Unlike some SDR enzymes,⁴⁶⁴ crystalline CAD does not undergo a significant conformational change upon substrate binding; the only noticeable change is that of Arg208 which moves to a position so that it closes the entrance of external solvent to the substrate binding pocket. Several hydrogen bonds are apparently responsible for binding of CA within the active site (e.g. Ser142 and Tyr155 form hydrogen bonds to the hydroxyl group at C-9 of CA, Fig. 41C). This positions C-9 of CA ∼2.5 Å above the pro-S hydrogen of nicotinamide, with a dihedral angle of 123°. The β-lactam C7 carbonyl, interestingly, points toward a hydrophobic pocket and does not participate in hydrogen bonding which may justify for the ability of CAD to process such a labile intermediate without initiating β-lactam hydrolysis.

5.8.2 Mechanism of CAD. The position of CA/NADPH within the active site is consistent with the consensus mechanism for SDR enzymes in which the 4-pro-S hydrogen of NADPH is transferred to the re-face of the aldehydic carbon of clavaldehyde, resulting in the transferred hydrogen occupying the pro-R position at C-9 of CA (Fig. 41B).^464,465 Hydrogen-bonding of clavaldehyde aldehydic oxygen to Ser142 and Tyr155 probably promotes nucleophilic attack by the hydride; the resulting positive charge on the oxidised nicotinamide ring may help reduce the pK_a of Tyr155, allowing proton transfer from the Tyr155-OH group to the alkoxide formed during the reduction step. Subsequently, Tyr155 could be re-protonated via a proton relay system, where protons are transferred from surrounding water molecules located in a small hydrophilic pocket (Fig. 41B).^466,468 The observed conformation of CA in CAD active site (Fig. 41C) reveals the C9–OH bond adopting an eclipsed conformation (i.e. projecting towards the oxazolidine oxygen, and almost coplanar with the O₁–C₂ and exocyclic alkene bonds); this conformation contrasts with that observed in small molecule crystal structures of clavulanate in which the C9–OH bond points away from the clavam nucleus.^472,473 The observed eclipsed conformation may enable the reversible nature of the CAD-catalysed reaction. The proposed stereochemical course of the CAD-catalysed clavaldehyde reduction rationalises earlier labeling studies on the origin of the hydrogens at the C9 of CA. Work using ornithine (a precursor of arginine, Section 5.11) labeled at its C-5 pro-R and pro-S positions demonstrated that only the pro-R hydrogen is incorporated into C9 of CA in a process occurring with overall inversion of configuration.⁴⁷⁴ This implies that the C5 pro-S hydrogen of ornithine/arginine must be lost from glycyl-clavaminic/clavaminic acid en route to clavaldehyde (Fig. 41A).

5.9 The oligopeptide binding proteins (Orf7 and Orf15) of clavulanic acid biosynthesis

Two of the CA biosynthesis cluster genes, orf7 and orf15, are closely related in sequence (∼48% identity) but apparently have little functional cross complementation.⁴⁷⁵ Orf7 and Orf15 (referred to as OppA1 and OppA2, respectively) belong to the oligopeptide-binding protein (OPP)/substrate-binding protein (SBP) family which are involved in the import of peptides in S. clavuligerus.⁴⁷⁵ Mutants of S. clavuligerus defective in orf7 or orf15 show loss of CA production, which can be restored by complementation.⁴⁵³ Genetic studies on CA biosynthesis⁴⁵³ led to the proposal that they might be involved in sensing signalling peptides that regulate CA biosynthesis.⁴⁷⁵ On the other hand, peptide binding assays suggest that OppA1/2 could be involved in binding and/or transport of small-molecules/peptides across the cell membrane of S. clavuligerus, as they have been found to bind di-/tri-peptides containing arginine and certain nona-peptides including bradykinin.⁴⁷⁶

Crystal structures of apo-OppA2 and in complex with arginine or bradykinin have been reported.⁴⁷⁶ OppA2 crystallises as a pear-shaped protein (Fig. 3 in⁴⁷⁶) comprising two loops and three domains, with a large cleft (between the two loops) running the entire length of the molecule, as observed for other oligopeptide-binding proteins.^476,477 The cleft is proposed to capture the substrate with subsequent domain movement, where lobes 1 and 2 close around the ligand (peptide/arginine) in a mechanism that has been likened to a “Venus flytrap”.⁴⁷⁸ Crystal structures of OppA2 in complex with arginine or bradykinin reveal that the C-terminal arginine of bradykinin binds similarly to arginine. At present, the exact role of OppA1 and OppA2 in CA biosynthesis is unclear, but they may play a role in transport of CA/intermediates or in the conversion of clavaminic acid to CA.

5.10 Orf13 and Orf14 of clavulanic acid biosynthesis

The putative proteins encoded for by orf13 and orf14^453,455 show similarity to (amino acid) export pump proteins and acetyltransferases, respectively. Mutation (or deletion) resulted in substantial reduction (orf13) or ablation (orf14) of CA and (5S)-clavam production. It is proposed that Orf14 catalyses the formation of acylated derivatives (e.g. N-acetyl-glycyl clavaminic acid) and that Orf13 may be involved in their transport/excretion.^455,479

5.10.1 Orf14 structure. Orf14 exists, in solution, predominantly as a monomer with some homodimer being observed.⁴⁸⁰ Structural studies reveal Orf14 as a member of the tandem GCN5-related acetyltransferase (GNAT) family proteins which often co-purify with a single acetyl-CoA molecule bound in their N-terminal GNAT domains;⁴⁸⁰ GNAT enzymes catalyse the transfer of an acetyl/acyl group from acetyl/acyl-CoA to an acceptor amine (for review, see ref. 481, 482). GNAT enzymes employ a range of “acceptor” substrates, including proteins (e.g., histones) and small molecules (e.g., tabtoxin, Section 7.3).⁴⁸³ The consensus GNAT mechanism proceeds via a “tetrahedral” intermediate which is stabilised by binding in an oxyanion hole. Both N- and C-terminal domains of Orf14 possess the characteristic GNAT superfamily mixed α,β-fold which comprises a central β-sheet flanked by helices (Fig. 23D).^481,482 The reported Orf14 crystal forms⁴⁷⁶ have one molecule of acetyl-CoA bound to the N-terminal GNAT domain, with the C-terminal domain being ligand-free; however, MS analyses have revealed that a second acetyl/acyl-CoA molecule can bind to Orf14, likely in the C-terminal domain.⁴⁸⁰ The N-terminal domain of Orf14 has no clear hydrogen-bonding interactions with the thioester-oxygen of acetyl-CoA (i.e. there is no clear oxyanion hole); the acetyl group is bound in a hydrophobic pocket with no residues nearby that could be obviously involved in general acid/base catalysis.⁴⁸⁰ Although the C-terminal domain of Orf14 has no CoA-derivative bound in the reported crystal structures, there is sufficient space for one to bind; a potential oxyanion hole is formed by the backbone amides of Met268 and possibly Thr269.⁴⁸⁰ Non-denaturing MS analyses have revealed that CoA derivatives can bind to Orf14 monomer without displacing the already-bound acetyl-CoA, suggesting that the C-terminal domain of Orf14 is the “catalytic” domain.⁴⁸⁰ Thus, the acetyl-CoA bound in the N-terminal domain may be rather a “structural” element than a co-substrate. The finding that Orf14 is a GNAT protein may assist in functional assignments of its role in CA biosynthesis and, in particular, whether or not it is involved in production of N-acetyl-clavaminic acid or N-acetyl-glycyl-clavaminic acid.

5.11 Role of ornithine acetyl transferase in clavulanic acid biosynthesis

The oat2 (orf6) gene of the clavulanic acid (CA) gene cluster (Fig. 28) encodes for an ornithine acetyl transferase (OAT2). Ornithine acetyltransferases (OATs) catalyse the reversible transfer of an acetyl group from N-α-acetyl-L-ornithine to L-glutamate with the corresponding production of L-ornithine and N-α-acetyl-L-glutamate.²³⁸ Although a direct role in CA biosynthesis is possible, especially given that N-α-acetyl-L-ornithine is a substrate for CAS (Section 5.3), it seems likely that OATs are involved in optimising arginine production⁴⁸⁴ as a “feedstock” for CA biosynthesis likely via increased channeling of glutamic acid towards the arginine/CA pathway (Fig. 42). The occurrence of a functional arg box (arginine-controlled regulatory sequence), which is characteristic of genes regulated by the arginine repressor ArgR, upstream of oat2 is supportive of such a role for oat2.^485–488 The disruption of argJ, which encodes for another ornithine acetyl transferase (ArgJ, the fifth enzyme of the arginine biosynthesis pathway in S. clavuligerus),⁴⁸⁷ reduced OAT activity in S. clavuligerus by 24%; however, disruption of oat2 had almost no effect on OAT activity (but slightly reduced CA production), suggesting that the OAT2 activity may be of little relevance to the total OAT cellular activity. Disruption of both argJ and oat2 resulted in residual 69% of the total OAT activity in the cellular extracts, suggesting that uncharacterised enzyme(s), e.g. OAT1 of the clavaminic acid gene cluster (Fig. 28),⁴⁸⁹ might additionally contribute to the acetyltransferase reaction in S. clavuligerus.

Fig. 42 The proposed role of OAT2 in clavulanic acid (CA) biosynthesis. OAT2 is an Ntn-hydrolase proposed to be involved in optimising arginine production as a “feedstock” for CA biosynthesis. Note that N-α-acetyl-L-glutamate can be converted to N-α-acetyl-L-ornithine, via N-α-acetyl-γ-L-glutamyl phosphate and N-α-acetyl-L-glutamate semialdehyde, during L-arginine biosynthesis.⁷²⁴ The ornithine acetyl-transfer reaction (boxed) as catalysed by OAT2 is proposed to proceed via an acyl-enzyme intermediate involving a ping-pong bi–bi mechanism.^492,724

5.11.1 Structure and mechanism of OAT2. Crystallographic studies have revealed that OAT2 has the conserved αββα-fold of the N-terminal nucleophile (Ntn) family of enzymes (Section 4.5, Fig. 16). Like other Ntn enzymes, nascent OAT2 undergoes autoproteolysis, at a canonical motif, to yield two chains (α- and β-subunits) which fold to give the active protein.²³⁸ OAT2 crystallises and exists in solution as α₄β₄-heterotetramer, with four αβ monomers per asymmetric unit (Fig. 44I).^{238,490–492} The α- and β- subunits are associated by non-covalent bonds, with the whole of the α-subunit and the N-terminal part of the β-subunit forming a large sub-domain with a smaller sub-domain being made up of the remainder of the β-subunit. In the apo-OAT2 structure, the OAT2 molecules are packed such that the active site of one monomer is close to an intermolecular interface; in contrast, in the acyl-OAT2 structure, the four OAT2 monomers are arranged such that their active sites, including the C-terminal regions, are directed towards solvent (Fig. 44I).^491,492 The binding of a substrate has been suggested to stabilise the conformation of the mobile C-terminus toward the active site,⁴⁹² possibly leading to the observed reduction in the size of the active site pocket as observed in the case of a Mycobacterium tuberculosis OAT:L-Orn complex structure.⁴⁹³

¹³C NMR and IR analyses imply the presence of an acetyl-enzyme intermediate in OAT catalysis; MS and crystallographic studies have identified Thr181 as the residue being acetylated. The carbonyl oxygen of the acyl-enzyme complex is located in an oxyanion hole and positioned to hydrogen bond with the backbone amide NH of Gly112 and the alcohol of Thr111 (Fig. 43, 44II).^491,492 There is evidence that the ester link of the acetyl-enzyme intermediate, similarly to those for other reported acyl-enzyme complexes,^494–496 adopts the thermodynamically preferred Z-geometry.⁴⁹⁷ While the crystallographic analyses reveal only one structure, IR studies demonstrate the presence of two distinct acyl-enzyme complexes; modeling studies suggest that one structure correlates with that observed crystallographically and the other with only a single oxyanion hole hydrogen bond to the backbone amide of Gly112.⁴⁹² The two structures can interconvert by movement of the Thr111 side-chain away from the oxyanion hole to hydrogen bond with the backbone carbonyl of the acetylated residue Thr181.^491,492 The combined crystallographic and modeling studies reveal that whilst the side chains of L-Glu and L-Orn likely bind in different orientations, the L-Orn and L-Glu α-amino groups occupy similar positions relative to the acetylated Thr181 (Fig. 44III), consistent with a common mechanism for N-acetyl transfer.^491–493 The L-Glu α-amino group is positioned close to the N-terminal α-amino group of Thr181, consistent with the proposed role of the Thr181 α-amino group in general acid–base catalysis during acetylation/deacetylation (Fig. 43).^491,492 From a molecular enzymology perspective, OAT2 is proposed to serve as a model for studying the stereoelectronics of catalysis via acyl-enzyme complexes, which occur as intermediates for many enzymes including proteases, transpeptidases and esterases.

Fig. 43 Mechanistic proposal for the Ntn-hydrolase ornithine acetyltransferase (OAT2). The oxyanion hole forming residues are in green. There is evidence for conformational changes during catalysis. NAG: N-α-acetyl-L-glutamate; NAO: N-α-acetyl-L-ornithine.⁴⁹¹

Fig. 44 Structural views of ornithine acetyltransferase (OAT2). I: The tetrameric oligomerisation of an acetyl-OAT2-glutamate crystal structure (PDB 2YEP) showing the four subunits/eight chains of OAT2 acyl-enzyme complex.⁴⁹² The AB, CD, EF and GH molecules are shown in different colours corresponding to the eight different chains; II: View from the active site of an acyl-OAT2-glutamate complex (AB molecule). The acetylated Ntn-residue Thr181 and the oxyanion hole forming residues Thr111 and Gly112 are shown; III: Superimposition of the acetyl-OAT2-glutamate complex (AB molecule) and non-acetylated OAT from Mycobacterium tuberculosis in complex with L-Orn (PDB 3IT4)⁴⁹³ comparing the binding modes for L-Orn and L-Glu (OAT2 numbering is employed). Note that the positions of the α-amino groups of L-Glu and L-Orn are similar, suggesting a closely related mechanism for N-acetylation/deacetylation of the two substrates, but that the binding of their side chains is different. Backbone (main chain) of some residues are omitted for clarity.

Employing N-acetyl-ornithine as an acetyl donor, D-glutamate, L-aspartate, L-leucine, L-alanine, and L-α-aminoadipic acid were not substrates for OAT2. A low acetyltransferase activity was observed with L-arginine, L-glutamine and L-lysine as acceptors.²³⁸ The OAT2-catalysed acetylation of L-arginine is of interest because N-acetyl-L-arginine is a good substrate for CAS⁴²⁵ (Fig. 37D). These results suggest that OAT2 might be amenable to protein engineering approaches to modify its substrate specificity/acceptance range, e.g. to resolve racemic mixtures of N-acetylated α-amino acids.

5.12 Summary of genetic studies on clavulanic acid biosynthesis

Genetic studies have been instrumental in deciphering known steps in CA biosynthesis – for detailed descriptions of these, the reader is referred to other reports.^{381,498–500} In summary, in S. clavuligerus, three gene clusters are apparently involved in CA/clavam biosynthesis; these have been labeled the “clavulanic acid (CA)”, “clavam” and alanyl-clavam/“Paralog” gene clusters, which are located in 3 unlinked loci in the genome (Fig. 28).⁵⁰¹ Interestingly, the CA gene cluster is located adjacent to that encoding for cephamycin C biosynthesis forming a “super-cluster”.⁵⁰² Like the clavam gene cluster, the super-cluster is located on the chromosome, but the paralog gene cluster is located on a megaplasmid (pSCL4),^381,503,504 which is densely packed with a large number of gene clusters for the potential production of secondary metabolites, including putative antibiotics, such as moenomycin, β-lactams, and enediynes. Interestingly, cross-regulation occurs between chromosomal and plasmid-encoded genes.⁵⁰³

5.12.1 Clavulanic acid gene cluster. Early studies suggested that only orf2–9are required for CA production;^{454,455,486,505} later studies suggested that orf10, orf11, and orf12 (which encode for a putative cytochrome P450, a ferredoxin-like protein, and a putative esterase, respectively) are also required.⁴⁵⁴ However, recent analyses imply that the CA gene cluster consists of up to 23 orfs involved in biosynthesis, transport and regulation (Fig. 27A and Table 2).^{453,455,506,507}

The CA gene cluster likely contains all the non-primary metabolic genes essential for CA production. Homologues of genes encoding for CEAS, β-LS, PAH, and OAT2 are found on the paralog gene cluster, and a cas homologue (CAS1) is present on the clavam gene cluster (Fig. 28). Mutants defective in only one copy of the paired genes (i.e. cas1/cas2, ceas1/ceas2, bls1/bls2, oat1/oat2, and pah1/pah2) are able to produce CA and (5S)-clavams to variable degrees. Double mutants, defective in both copies of the paired genes, are blocked in CA and (5S)-clavam biosynthesis, except for the oat1/oat2 double mutant,⁴⁵³ which produces reduced levels of CA and some (5S)-clavams, suggesting that OAT1/OAT2 serve only to increase arginine levels available for the biosynthesis of these metabolites (Section 5.11). Mutants defective in oppA1 (orf7, Section 5.9), claR (Section 5.6), cad (Section 5.8), or cyp (orf 10, Section 5.6) produced no CA, while production of (5S)-clavam metabolites was relatively unaffected, suggesting the absence of paralogous genes encoding functionally equivalent proteins.⁵⁰⁵ Recently, orf21, orf22, and orf23, were identified to encode a putative sigma factor, a sensor kinase, and a response regulator, respectively.^506–508 Disrupted mutants of these genes displayed modest reductions of CA production, while their overexpression resulted in increased CA production.^506–508 Overexpression of ceas2, bls2, pah2, and cas2 resulted in >8 fold increase in CA production;⁵⁰⁹ for other attempts aiming to improve CA production, see ref. 381, 506, 508, 510, 511.

5.12.2 Paralog gene cluster. Some of the genes of the paralog gene cluster (i.e. ceas1, bls1, oat1, and pah1, Fig. 28B) encode for isozymes of proteins involved in the biosynthesis of clavaminic acid which is common to both CA and (5S)-clavam biosynthesis pathways.^{489,505,512,513} The paralog gene cluster also contains genes involved in the latter stages of (5S)-clavam biosynthesis. Genes similar to cvm6 and cvm7 from the clavam cluster (see below) also appear in the paralog gene cluster (i.e. cvm6P and cvm7P); these genes encode for a putative aminotransferase and a transcriptional regulator, respectively. While mutants defective in cvm6 and cvm7 have no effect on clavam production, those defective in cvm6P and cvm7P completely lost the ability to produce (5S)-clavams,⁵¹⁴ suggesting that the paralog gene cluster might include genes specific for the late stages of (5S)-clavam biosynthesis.⁵⁰⁴ The orfA, orfB, orfC, and orfD genes of the paralog gene cluster encode for a putative hydroxymethyl-transferase, a regulatory protein, an aminotransferase, and a dehydratase, respectively; mutants defective in these genes were unable to produce alanylclavam but had no effect on the production of other (5S)-clavams and CA. Disruption of orfC resulted in accumulation of 8-hydroxy-alanyl-clavam (a proposed intermediate in alanyl-clavam biosynthesis). These results suggest that the four genes (orfA to orfD) are likely involved in the conversion of 2-hydroxymethylclavam to alanylclavam via 8-hydroxyalanylclavam (Fig. 27A).⁵⁰⁴

5.12.3 Clavam gene cluster. The number of species reported to produce non-CA clavams exceeds those producing CA, implying that the ability to biosynthesise CA is more restricted.⁵⁰⁰ In S. clavuligerus, disrupted mutants of cvm1, cvm2, and cvm5 (Fig. 28C) did not affect CA production but displayed altered abilities to produce (5S)-clavams.^514,515 Disruption of cvm1, which encodes for a putative aldo-keto reductase, resulted in loss of (5S)-clavams production.⁵¹⁵ Disruption of cvm2, a putative isomerase, resulted in substantial reduction in 2-hydroxymethylclavam and alanylclavam production and abolished clavam-2-carboxylate production.⁵¹⁴ These results suggest the possible involvement of Cvm1 and Cvm2 in the late uncharacterised steps of (5S)-clavams biosynthesis. Cvm5 and Cvm3 are a putative flavin-dependent monooxygenase and a putative flavin-reductase enzymes, respectively, which are proposed to form a coupled redox system.^381,514 While disruption of cvm3 had no effect on (5S)-clavam production, disruption of cvm5 resulted in abolishment of all known (5S)-clavams and accumulation of the not previously observed metabolite 2-carboxymethylideneclavam (Fig. 27A).⁵¹⁴

The clavam gene cluster of S. antibioticus, which is reported to produce only two clavam metabolites, 2-hydroxyethylclavam and valclavam (Fig. 27B, Table 3), but not CA, has recently been sequenced, analysed and shown to resemble the clavam gene cluster of S. clavuligerus.⁵¹⁶S. antibioticus cluster also contains genes encoding for apparently new/different proteins, e.g. a putative oxido-reductase and a putative ligase (Table 3). Deletion of the putative oxido-reductase encoding gene abolished clavam production.⁵¹⁶

Table 3 Genes constituting the reported clavam biosynthesis gene cluster in Streptomyces antibioticus and the (predicted) roles of the (putative) proteins that they (may) encode for.⁵¹⁶ The predicted number of amino acid residues (AA) for each of the (putative) proteins is shown. The homologous proteins in S. clavuligerus (Table 2, Fig. 27A) are in brackets

Gene	AA	(Proposed) function of encoded protein
psr	843	Pathway specific transcriptional regulator (Cvm7p).
atr	467	Aminotransferase (Cvm6p).
akr	340	Aldo-keto reductase (Cvm1).
hmt	416	Hydroxymethyl transferase (OrfA).
ceaS3	571	Carboxyethylarginine synthase (CEAS1/CEAS2).
bls3	507	β-Lactam synthetase (βls1/βls2).
pah3	320	Proclavaminic acid amidinohydrolase (PAH1/PAH2).
cas3	324	Clavaminic acid synthase (CAS1/CAS2).
oat3	394	Ornithine acetyltransferase (OAT1OAT2).
oxr	214	Enoyl-reductase.
trn	405	Clavam transporter.
lig	478	Ligase.
pah4	309	Proclavaminic acid amidinohydrolase (PAH1/PAH2).
ctr	456	Transcriptional regulator.

---

6 Carbapenem biosynthesis

6.1 Subfamilies of naturally-occurring carbapenems

Since the discovery of thienamycin, the number of naturally-occurring carbapenems has risen to over 45. Naturally-occurring carbapenems can be sub-divided into the thienamycins, olivanic acids, epi-thienamycins, carpetimycins, asparenomycins, pluracidomycins, carbapenems of the PS group, and those of the OA-6129 group (Table 4, Fig. 45).⁵¹⁷ The majority of identified carbapenems were isolated from the fermentation broths of Streptomyces spp. with two notable exceptions: the simplest carbapen(am/em)s, which were isolated from the fermentation broths of some Enterobacteriaceae spp. e.g. the plant pathogen Pectobacterium carotovorum,⁵¹⁸ and AB-110-D, the only known carbapenem with (Z)-geometry at its exo-alkenyl C-2 side chain (Fig. 45), which was isolated from the actinomycete Kitasatosporia papulosa.⁵¹⁹

Table 4 Major sub-groups of carbapenems isolated from natural sources. The C-2 substituent (R³) is generally a derivative of cysteamine or dehydrocysteamine, in which the sulfur atom can be oxidised to a sulfoxide oxidation state. Note that two numbering systems are in use for naming carbapenems.⁷²⁵ In one system, the bicyclic nucleus is considered as a derivative of penicillin (i.e. desthia-l-carbapen-2-em). The other system is based on IUPAC nomenclature, 7-oxo-l-azabicyclo[3.2.0]hept-2-ene-2-carboxylate. The former numbering scheme has been adopted in this review. The figure has been adapted from ref. 517

Group	R^1	R^2	H5/H6	C-6	C-8
Thienamycins	OH	H	trans	S	R
Olivanic acids	OH	H	cis	R	S
Pluracidomycins	OSO3^−	H	cis	R	S
Carpetimycins	OH	CH3	cis	R	—
PS	H	H/CH3	trans	R	—

---

Fig. 45a

Fig. 45b Carbapen(em/am)s isolated from natural sources.⁵²⁵ Some stereochemical assignments are provisional and some of the sulfoxide stereochemistries are unknown. Thienamycin, N-acetyl-thienamycin, N-acetyl-dehydro-thienamycin, 8-epi-thienamycin, northienamycin, and NS-5 were all isolated from wildtype or a mutant of S. cattleya.^{81,82,726–728} Thienamycin was also isolated from S. penemijaciens.⁷²⁹epi-Thienamycins A–F were isolated from S. flavogriseus^571,730 and from S. olivaceus.^{572,731–735} The olivanic acids MM 4550^572,733,734 and MM 27696⁷³⁶ were isolated from S. olivaceus. The PS subfamily of carbapenems (PS-5 to PS-8) were isolated from S. cremeus subsp. auratilis.^737–739 Carpetimycins A–D and KA-6643-G were isolated from Streptomyces sp. KC-6643.^740–744 Asparenomycins A–C were isolated from S. tokumonensis, S. argentealus and another Streptomyces sp.^745,746 Mutation of the carpetimycin producer Streptomyces sp. KC-6643 resulted in a strain that no longer produces carpetimycins but instead produces 6643-X.⁷⁴⁷ The OA-6129 subfamily of carbapenems (A, B₁, B₂, C, D, E), characterised with a pantetheinyl moiety at C-2, were isolated from Streptomyces sp. OA-6129 and of S. fulvoviridis mutant.^{582,748–750} The seven pluracidomycins (A₁, A₂, B, C₁, C₂, C₃ and D), characterised by containing 2- or 3-acidic moieties in their structures, were isolated from S. pluracidomyceticus.^751–753 In addition, the culture co-produced epithienamycin A, epi-thienamycin B sulfoxide, epi-thienamycin D sulfoxide, epi-thienamycin F, MM 4550.⁷⁵² The carbapenam 17927 D (or its enantiomer) was isolated from S. fulvoviridis, and exhibited no antimicrobial activity.⁷⁵⁴

Medicinal chemistry efforts have resulted in the development of the current commercially available carbapenems (see Fig. 46 for examples) which differ from thienamycin in, at least, two important aspects: (i) capping of the nucleophilicity of the C-2 amino side chain, e.g. as in imipenem,^520,521 which reduces the intermolecular hydrolysis of the β-lactam; (ii) the introduction of a 1β-methyl substituent, e.g. as in ertapenem, meropenem, and doripenem (Fig. 46), which results in derivatives with superior antibacterial activity and increased resistance to inactivation by renal dehydropeptidases.^91,522 The α-methyl isomers are also resistant to dehydropeptidase hydrolysis but their antibacterial activities are reduced.⁵²²

Fig. 46 Examples of clinically used carbapenems and the common intermediate used in their preparation, (3R,4R)-4-acetoxy-3-[(R)-1-(tert-butyldimethylsilyloxy)ethyl]azetidin-2-one (AOSA). The part of skeleton corresponding to that of thienamycin is in blue.

6.2 Overview of carbapen-2-em-3-carboxylate biosynthesis

The isolation and characterisation of carbapen-2-em-3-carboxylate (C3C, the simplest carbapenem) was an important advance in understanding carbapenem biosynthesis due to the complexity of the pathways leading to the C-2 and C-6 functionalised carbapenems (e.g. thienamycin, Fig. 45). Pioneering labeling studies in the 1980s established that labeled acetate and glutamate are incorporated into the β-lactam carbons and the fused pyrroline ring, respectively, of both C3C and thienamycin.^523,524 Labeled cysteine and methionine have been reported to be incorporated into the C-2 and C-6 side chains of thienamycin, respectively.⁵²³ For previous reviews on carbapenem biosynthesis, see ref. 525, 526. We begin by summarising recent studies on the biosynthesis of C3C, then return to studies on thienamycin.

C3C was first isolated, in 1982, from P. carotovorum and from a Serratia sp.⁵¹⁸ and was subsequently found to be a product of the insect pathogen Photorhabdus luminescens.⁵²⁷ C3C co-occurs with two saturated carbapenams: (3S,5S)- and (3S,5R)-carbapenam-3-carboxylate.⁵²⁸ The isolation, sequencing, and initial assignment of the P. carotovorum C3C biosynthetic gene cluster have revealed 9 genes (carA-H and carR) that are potentially involved in C3C biosynthesis and regulation (Fig. 47B, Table 5).^529,530 CarA and CarC show homology to the clavam biosynthesis enzymes β-LS and CAS (Sections 5.2 and 5.3, respectively).²⁹ After some uncertainty with respect to the stereochemistry of the carbapenam intermediates,^{528,531–533} it was established that only three enzymes are absolutely required for C3C biosynthesis (i.e. CarA, CarB and CarC).^529,534

Fig. 47 The biosynthetic pathway leading to the simplest carbapenem, (5R)-carbapen-2-em-3-carboxylate (C3C), in Pectobacterium carotovorum (A) and the gene clusters (potentially) involved in C3C biosynthesis (B). The production of C3C by D. zeae and Pantoea sp. is yet to be assessed. Note that in case of Pantoea sp., carD and carE exist as a single gene, unlike other producers of C3C where separate genes exist for carD and carE. See Table 5 for the proposed role of the encoded proteins. Selected co-substrates/co-products are shown.

Table 5 Genes constituting the reported (5R)-carbapen-2-em-3-carboxylate (C3C) gene cluster⁵¹⁸ in Pectobacterium carotovorum and the (predicted) roles of the (putative) proteins that they (may) encode for. The predicted number of amino acid residues (AA) for each of the (putative) proteins is shown. Proteins with an experimentally assigned biochemical function are in bold

Gene	AA	(Proposed) function of encoded protein
carR	244	Transcriptional regulator (CarR).^535
carA	503	CarA; β-Lactam synthetase.^395
carB	250	CarB (crotonase); t-CMP formation.^537,538
carC	273	Carbapenem synthase (CarC).^542
carD	376	Proline dehydrogenase (CarD).^530
carE	92	2Fe-2S ferredoxin (CarE).
carF	288	Protein involved in resistance to C3C (CarF).^529
carG	177	Protein involved in resistance to C3C (CarG).^529
carH	184	Unknown.

---

CarR has been identified as a DNA-binding LuxR homologue transcriptional regulator that functions in the presence of N-(3-oxohexanoyl)-L-homoserine lactone (OHHL).⁵³⁵ The CarR:OHHL complex binds to the carA promoter and activates the expression of the carA–H cluster. CarF and CarG are proposed to be involved in intrinsic resistance to C3C via an uncharacterised mechanism.⁵²⁹ CarI is involved in a quorum sensing system involving the synthesis of OHHL. The reader is referred to reviews^526,536 for the details of pioneering studies on the regulation of carbapenem biosynthesis, which were of importance far beyond the BLA field. CarD (a putative proline dehydrogenase) and CarE (a putative 2Fe-2S ferredoxin) may act in concert to catalyse proline oxidation⁵³⁰ to give an equilibrating mixture of isomers (L-glutamate semi-aldehyde, L-GSA/L-5-hydroxyproline, L-5HP/L-pyrroline-5-carboxylate, L-P5C), collectively abbreviated as L-GHP, which is a co-substrate for the carboxymethylproline synthase CarB. The crotonase CarB catalyses the formation of (2S,5S)-carboxymethylproline (t-CMP) from malonyl-CoA and L-GHP (L-P5C).^537,538 CarB (and likely ThnE in thienamycin biosynthesis) are proposed to act as a “stereochemical gateway” for carbapenem biosynthesis; only in the presence of L-GHP is t-CMP formed with concomitant CoASH production; D-GHP stimulates malonyl-CoA uncoupled turnover and CoASH production but no C–C bond formation product can be detected.^539–541 The ATP-dependent carbapenam synthetase CarA (see Section 5.2) then catalyses the β-lactam closure.³⁹⁵ Finally, the carbapenem synthase CarC, a 2OG oxygenase, catalyses the highly unusual C-5 epimerisation of (2S,5S)-carbapenam to a (2S,5R)-derivative which then undergoes a CarC-catalysed desaturation reaction to afford the biologically active antibiotic C3C (Fig. 47A).^29,542

6.3 The role of individual enzymes in C3C biosynthesis

6.3.1 The carboxymethylproline synthase CarB (and its homologue in thienamycin biosynthesis, ThnE). The carB gene, in P. carotovorum (Fig. 47B), and its homologue in thienamycin biosynthesis thnE, in S. cattleya (Fig. 48B and 49), encode for 250- and 294-residue proteins, respectively, that have been recombinantly expressed in E. coli, purified, and characterised as carboxymethylproline synthases (CMPSs).^{397,537–539,541} CarB and ThnE are unique members of the crotonase superfamily (CS) of enzymes^543,544 because they catalyse three different functional group interconversions (i.e. malonyl-CoA decarboxylation, C–C bond formation, and thioester hydrolysis) within the same active site. The CS enzymes catalyse a diversity of reactions, many of which have precedent in the “enolate” chemistry of organic synthesis.⁵⁴³

Fig. 48 Proposed biosynthetic pathway leading to thienamycin (A) and the reported thienamycin gene cluster in Streptomyces cattleya⁵⁶⁵ (B). The thienamycin-like gene cluster identified in S. flavogriseus (B) has been reported to be unable to produce thienamycin under laboratory conditions.⁷⁵⁵ SAHC: S-adenosylhomocysteine. The enzymes catalysing the experimentally-reported steps are in red in (A). The precursors for the skeleton of thienamycin are in blue. PS-7 sulfoxide was detected only in in vitro assays with ThnG (i.e. it was not isolated from S. cattleya). See Table 6 for (proposed) roles of the proteins. Selected co-substrates/co-products are shown.

Fig. 49 Partial sequence alignment for known and putative CMPSs catalysing the formation of t-CMP in carbapenem biosynthesis. The oxyanion hole forming residues are in pink and the catalytically important (at least in CarB)⁵⁴⁶ glutamate residue in orange (see Fig. 50). The figure was generated using Clustal W⁷⁵⁶ and Genedoc.⁷⁵⁷ α-Helices (cyan cylinders) and β-strands (red arrows) represent the assigned secondary structure of CarB.⁵⁴¹

The CarB-catalysed formation of t-CMP is considered as the committed step in C3C biosynthesis in P. carotovorum.²⁹ In the early stages of thienamycin biosynthesis in S. cattleya, the role of ThnE is likely to provide t-CMP, or a derivative thereof.^397,539,545In vitro, in addition to its ability to catalyse t-CMP formation (from malonyl-CoA and L-P5C), ThnE catalyses the conversion of epimeric methylmalonyl-CoA to give (6R)-methyl-t-CMP as the major product, with 60% diastereomeric excess, compared to 10% in case of CarB, suggesting that the C-6 alkyl group could be introduced at an early stage in thienamycin biosynthesis (see below). However, the lack of in vitro conversion of ethylmalonyl-CoA to 6-ethyl-t-CMP by ThnE suggests subsequent further methylation and hydroxylation steps are required if ThnE were to be responsible for initial methylation at C-6; though, it seems probable that the C-6 ethyl group is introduced at a later stage (see below).⁵³⁹

6.3.1.1 CarB Structure. CarB crystallises as a homotrimer, a dimer of trimers and a trimer of trimers. However, in solution, both CarB and ThnE exist predominantly as trimers.^539,541 While sequence similarities between individual CS proteins are often low, all members share a similar overall fold formed by repeated ββα-motifs that assemble into two approximately perpendicular β-sheets surrounded by α-helices (Fig. 50).⁵⁴³ The catalytic machinery of most CS enzymes operates to stabilise enolate/oxyanion intermediates via an “oxyanion hole” (OAH) comprising two backbone NH groups.⁵⁴³ In the case of CarB and ThnE, the active site residues forming the canonical OAH are proposed to be Met108_CarB/Val153_ThnE and Gly62_CarB/Gly107_ThnE (Fig. 49, 50B).^539,541 A structure of the CarB:acetyl CoA complex⁵⁴¹ reveals the CoA thioester derivative bound in a characteristic U-shaped conformation (Fig. 50B) and a conserved tunnel that binds the pantetheine part of coenzyme A leading to the binding pocket of L-GHP. The L-GHP (L-P5C) co-substrate is predicted to be anchored – via its carboxylate group – into the bottom of the active site pocket.⁵⁴¹ Trp79 forms part of the hydrophobic face of the CarB active site⁵⁴¹ (Fig. 50C) and was important in CarB engineering studies^402,403 (see below). The conserved side chain of Glu131 is suitably positioned to catalyse the hydrolysis of the t-CMP-CoA thioester intermediate^541,546 (Fig. 50B).

Fig. 50 Structural views of carboxymethylproline synthase CarB.⁵⁴¹A: A CarB monomer (PDB 2A81) displaying the repeated ββα-motif which is characteristic of the crotonase superfamily;⁵⁴³B: The substrate binding pocket with dimethylmalonyl-CoA derived enolate and L-pyrroline-5-carboxylate (L-P5C) modelled in. Note the characteristic U-shaped conformation of CoA, the location of the thioester carbonyl in the oxyanion hole (OAH), and that the side chain of Glu131 is suitably positioned (∼4 Å from the thioester carbonyl) to activate a water molecule for hydrolysis of a t-CMP-CoA intermediate (Path 3, Fig. 51); C and D: Models of the C-2 epimers of methylmalonyl-CoA (2S epimer: C, and 2R epimer: D) in the substrate binding pocket. Note that, in both cases, an orthogonal relationship between the methylmalonyl-CoA carboxylate and the OAH-stabilised carbonyl is maintained, as required for the stereo-electronically favoured decarboxylation. The enolate (E/Z) that would be generated following decarboxylation is shown in brackets, in each case.⁴⁰³

6.3.1.2 CarB and ThnE mechanism. CarB/ThnE catalysis is proposed to proceed via decarboxylation of malonyl-CoA to give an OAH-stabilised enolate (Fig. 50B). C–C bond formation can then proceed via reaction of the enolate with potentially any of the forms of L-GHP (Fig. 51): (i) with L-P5C (Path E); (ii) with L-5HP in an S_N2 reaction (Path D); or (iii) with L-GSA in an aldol reaction followed by a substitution reaction (Path B), or via elimination of water across C₄–C₅ (Path A), or via elimination of water across C₅–C₆ followed by Michael addition (Path C). The ability of CarB to catalyse the formation of 6,6-dimethyl-[²H]₆-t-CMP from [²H]₆-L-GHP and dimethylmalonyl-CoA argues against Path A and Path C.^546,547 Path B cannot be entirely eliminated but seems unlikely (at least in case of the CarB-catalysed formation of 6,6-dimethyl-[²H]₆-t-CMP) because it involves a substitution reaction at the sterically hindered “neopentylic” position (C-5). Reaction of the enolate directly with L-5HP (Path D) seems unlikely because: (i) an S_N2 type mechanism would have to occur on only one of the two epimeric hemiaminals of L-5HP to produce t-CMP; (ii) both of the C-5 hemiaminal epimers of L-5HP can readily eliminate water to form protonated L-P5C likely giving a lower energy reaction path (Path E).⁵⁴⁶ The observation that the hindered dimethylmalonyl-CoA derived enolate reacts (at all) is also an indirect evidence for this mechanism because the S_N1 iminium ion route (Path E) is sterically less demanding than an S_N2 route (Path B/D). Another supportive observation for Path E is that at pH 7.7 (the pH of the final CarB assay mixture), the imine form of L-GHP (L-P5C) is the predominant one (>95% by ¹H-NMR).^404,548 Analysis of the CarB structure (Fig. 50B) coupled to the observation that t-CMP is the only detectable diastereisomer imply that nucleophilic attack of the enolate anion occurs stereospecifically on the re face of L-P5C to give the intermediate t-CMP-CoA (Fig. 50 and 51).^538,546

Fig. 51 Possible C–C bond forming reactions leading to the formation of the t-CMP-CoA intermediate as catalysed by carboxymethylproline synthases. The green box highlights the most likely path.⁵⁴⁶ For [²H]₆-L-GHP, R′ = ²H.

For hydrolysis of the t-CMP-CoA intermediate, three mechanisms were considered (Fig. 52): a mechanism involving a ketene intermediate (Path 1), an anhydride intermediate (Path 2), or through direct attack of water (activated by Glu131) onto the thioester carbonyl of t-CMP-CoA (Path 3). The CarB-catalysed conversion of dimethylmalonyl-CoA to 6,6-dimethyl-t-CMP rules out a ketene intermediate (Path 1) at least for this substrate.^546,549 Incubations in buffered H₂¹⁸O led to the incorporation of a single ¹⁸O into the 6,6-dimethyl-t-CMP product; trypsin digest analyses on CarB from the same incubations revealed no ¹⁸O incorporation from solvent into the side chains of Glu131. The lack of ¹⁸O incorporation from solvent into the side chains of Glu-131 eliminates Path 2B, and reveals that CarB thioester hydrolysis occurs by a different mechanism than that proposed for the crotonase 3-hydroxyisobutyryl-CoA hydrolase.⁵⁵⁰ Path 2A cannot be entirely ruled out, but the available evidence argues against it because, at least with 6,6-dimethyl-t-CMP-CoA, this path requires hydrolysis adjacent to a sterically hindered quaternary carbon atom.⁵⁴⁶ Therefore, the CarB-catalysed thioester hydrolysis is proposed to occur via direct attack of a water molecule on the carbonyl carbon of the t-CMP-CoA intermediate (Path 3).⁵⁴⁶

Fig. 52 Possible thioester hydrolysis mechanisms for CarB catalysis to produce 6-alkyl-t-CMP derivatives. The green box highlights the most likely mechanism.⁵⁴⁶ R = H or CH₃.

6.3.1.3 CarB/ThnE engineering studies. Structure- and homology-guided CMPS engineering studies were carried out with the aim of deepening understanding of the CMPS mechanism of catalysis as well as to produce new CMPS variants that can catalyse reactions inaccessible for wildtype enzymes, e.g. to catalyse stereoselective quaternary centre formation.⁴⁰²

The CarB E131 A/Q/C/S variants are unable to catalyse the production of t-CMP or t-CMP-CoA from L-GHP and malonyl-CoA, but can catalyse decarboxylation of malonyl-CoA to acetyl-CoA.⁵⁴⁶ In contrast, the more conservative CarB E131D variant catalyses the formation of t-CMP, albeit in a low yield. These results reveal Glu131 as likely important for both the C–C bond formation and thioester hydrolysis steps but not important for decarboxylation.⁵⁴⁶

Analogues of L-GHP (L-P5C), substituted at C-2, C-3, C-4, and C-5 of the L-P5C skeleton, are accepted by wildtype and engineered CMPSs to produce the corresponding functionalised N-heterocycles in a stereoselective fashion (Fig. 53B). Notable results include the formation of products with quaternary centres at C-2 or C-5 of t-CMP and the stereoselective formation of (4S)-4-methyl-t-CMP which shares the same pattern of substitution and C-4 stereochemistry as many clinically used carbapenems (e.g. meropenem, Fig. 46).⁴⁰²

Fig. 53 The biocatalytic versatility of wildtype and engineered carboxymethylproline synthases (CMPSs). A: The reaction of wildtype CarB and ThnE; B: A variety of functionalised N-heterocycles is produced from appropriate analogues of L-GHP (L-P5C) and analogues of malonyl-CoA as catalysed by wildtype and engineered CMPSs;^{402–404,538,539,546}C: Coupling CarB catalysis to that of crotonyl-CoA carboxylase reductase (Ccr, which produces (2S)-ethylmalonyl-CoA^551,552) results in the stereoselective formation of (6R)-6-ethyl-t-CMP.⁴⁰³ d.e. refers to the diastereomeric excess observed with the shown CMPSs at the asterisked positions.

Analogues of L-GHP (L-P5C) with different chain lengths (i.e.L-aminoadipate semialdehyde and L-aminopimelate semialdehyde) are accepted by wildtype and/or variant CMPSs to stereoselectively produce the corresponding 6- and 7-membered N-heterocycles (Fig. 53B).⁴⁰⁴

The proposed enolate intermediacy in CMPS catalysis led to the suggestion that the catalytic machinery of these enzymes can be employed to generate “tri-substituted enolates” (from alkyl-malonyl-CoA derivatives), which can react with L-GHP/L-GHP analogues in a stereoselective manner (Fig. 54).⁴⁰³ While wildtype CarB and ThnE display weak to mild stereoselectivity, structure-guided engineering has resulted in CMPS variants that can exercise near complete stereocontrol at C-6/C-7 of the produced functionalised N-heterocycles (Fig. 53B).⁴⁰³ In this regard, two active site residues have played a crucial role: (i) Met108_CarB (one of the OAH forming residues) where its substitution with β-branched residues (i.e. Val or Ile) yields variants that favour formation of products with the (6R)-stereochemistry through stereoselective formation of the (Z)-enolate (Fig. 50C, 54); and (ii) Trp79_CarB (one of the residues in the hydrophobic part of the CarB active site) where its substitution with a less bulky residue (i.e. Phe or Ala) yields variants that favour formation of products with the (6S)-stereochemistry through stereoselective formation of the (E)-enolate (Fig. 50D, 54). It has been demonstrated that coupling CMPS catalysis to that of crotonyl-CoA carboxylase reductase^551–553 (Ccr, from Rhodobacter sphaeroides, which produces the single epimer (2S)-ethylmalonyl-CoA) leads to stereoselective formation of (6R)-6-ethyl-t-CMP (Fig. 53C).⁴⁰³ Although the available evidence is that (6R)-6-ethyl-t-CMP is not an intermediate in thienamycin biosynthesis, these results suggest ways in which the pathway may be re-engineered. It is notable that Ccrs are present in some Streptomyces spp.^554–557

Fig. 54 Stereoselective enolate formation and reaction by wildtype and engineered CMPSs.^403,404,546 In the case of alkylmalonyl-CoA substrate analogues, the reaction is proposed to proceed via the decarboxylation of a specific epimer to give either the (E)- or (Z)-enolate, which reacts with the imine form of L-amino acid aldehyde to give a specific diastereomer of the CoA-thioester intermediate. Hydrolysis of the CoA-thioester intermediate via an “activated” water molecule (by Glu131 in case of CarB)⁵⁴⁶ results in the formation of a specific diastereomer of carboxymethyl-substituted N-heterocycles.⁴⁰³ For a model of the possible modes of binding of alkylmalonyl-CoA to CMPS, see Fig. 50.

Overall, the results of CMPS engineering studies have pioneered the applicability of crotonases in biocatalysis and suggest that crotonases may provide a useful platform technology for catalysing synthetically challenging reactions, e.g. the stereoselective generation of enolates and their reaction with electrophiles of choice.

6.3.2 The carbapenam synthetase CarA (and its homologue in thienamycin biosynthesis, ThnM). The P. carotovorum carbapenam synthetase CarA,³⁹⁵ and its homologue in thienamycin biosynthesis in S. cattleya, ThnM,³⁹⁷ catalyse β-lactam formation in (2S,5S)-carbapenam from t-CMP (Fig. 47, 48, respectively). Structural, mechanistic and substrate analogue studies on CarA are described in Section 5.2, together with the related β-lactam synthetase of CA biosynthesis.

6.3.3 Carbapenem synthase (CarC). The 2OG oxygenase CarC catalyses the C-5 epimerisation of (3S,5S)- to (3S,5R)-carbapenam followed by desaturation of the latter to give C3C (Fig. 47A).^531,542 CarC is apparently well conserved in C3C biosynthesis pathways (Fig. 55). Whilst the CarC-catalysed desaturation is precedented in reactions catalysed by other 2OG oxygenases, including CAS and enzymes of flavonoid biosynthesis,⁵⁵⁸ its epimerisation reaction is unique.²⁹ CarC accepts all four possible isomers of its carbapenam substrate.⁴⁰⁷ The relaxed substrate specificities of CarA and CarC coupled to the fact that CarB accepts only L-GHP, but not D-GHP, imply that CarB controls the stereochemical course of C3C biosynthesis. Given that, at least in our current assays, CarC-catalysed epimerisation is less efficient than desaturation, it would be interesting to engineer a carbapenem biosynthesis pathway that does not involve C-5 epimerisation via modification of CarB to produce cis-CMP (corresponding to 5R at C-5 in the carbapenam). If the relative in vitro efficiencies of the CarC-catalysed epimerisation/desaturation are relevant in vivo, the nature of the pathway would seem to be an argument against “backwards evolution” as proposed by Horowitz.⁵⁵⁹

Fig. 55 Sequence alignments for known and putative 2OG and iron-dependent carbapenem synthases catalysing the formation of (5R)-carbapen-2-em-3-carboxylate. The residues involved in the conserved motif of a 2-histidine-1-carboxylate triad⁷⁵⁸ are in orange. Other residues proposed to be important in carbapenem synthase catalysis are in pink.⁵⁴² The figure was generated using Clustal W⁷⁵⁶ and Genedoc.⁷⁵⁷ Note the high degree of homology between the proteins (≥78% identity).

6.3.3.1 CarC structure and mechanism. The overall CarC structure is similar to those of other 2OG oxygenases and contains a double stranded β-helix core fold (Fig. 19). A crystal structure of CarC in complex with N-acetyl-L-proline (as an analogue of (3S,5S)-carbapenam) and modelling studies have provided insights into CarC substrate binding (Fig. 56). Two orientations of (3S,5S)-carbapenam in the CarC active site are proposed: orientation I positions C-3 and C-2 close to the iron centre with C-5 more exposed to the solvent (analogous to Fig. 56B); and orientation II which positions C-5 close to Fe^II.⁵⁴²

Fig. 56 Structural views of the carbapenem synthase CarC.⁵⁴²A: A CarC monomer (PDB 1NX8) displaying the conserved double stranded β-helix core (yellow strands) that supports ligands binding a single Fe^II to which 2OG complexes in a bidentate manner; B: The active site of CarC (monomer B) showing the binding sites for iron, the prime substrate analogue (N-acetyl-L-proline) and 2OG. Note that another orientation of N-acetyl-L-proline, related by a ∼180° rotation to the one shown, is possible.

Various mechanistic possibilities have been considered for the CarC-catalysed reactions; all of the proposed mechanisms involve a radical intermediate (Fig. 57). Labeling studies have shown loss of C-5 hydrogen during CarC catalysis.⁵⁶⁰ It should be noted that the epimerised (3S,5R)-carbapenam can be released from the CarC active site and that the source of the C-5 hydrogen of (3S,5R)-carbapenam is unknown. Ab initio calculations on the radicals derived by H abstraction from C-1, C-2, C-3, C-5, and C-6 of (3S,5S)-carbapenam suggest that the C-3 radical is the lowest in energy, because of the captodative effect.⁵⁶¹ Loss of C-3-H is likely the favoured scenario in orientation I; C-5 epimerisation could then occur through β-scission of C-3˙ and ring opening to eventually give (5R)-C3C (Fig. 57i). However, in case of C-5 epimerisation mechanisms involving initial C-3, C-1, C-2 and C-6 radical formation, the large barrier associated with ring opening and the unfavourable entropy for ring closure imply that significant “enzyme catalysis” would be required. In case of C-5 hydrogen abstraction mechanisms, C-5 epimerisation is predicted to be exothermic. If a ferryl intermediate was to both abstract the C-5 hydrogen and re-hydrogenate at the same position, re-orientation of the substrate would be required (along with a mechanism for hydrogen exchange). Conformational changes upon substrate binding have been observed for other 2OG oxygenases,^268,562 and modelling studies also suggest that (3S,5R)-carbapenam could be accommodated in the active site,^561,563 but there is no clear driving force for such a conformational change in the case of CarC catalysis. Therefore, further studies on the CarC mechanism are required.

Fig. 57 Proposed possibilities for the CarC-catalysed epimerisation/desaturation process.⁵⁴² The 5-endo-trig (or 4-exo-trig) radical cyclisations in (i) and (ii) have synthetic precedent,⁷⁵⁹ including with an imine substrate.⁷⁶⁰ Note the possible intermediacy of datively stabilised radicals. Epimerisation via hydrogen abstraction at C-1 (as is initial electron loss from nitrogen) in a process analogous to (i) is also a possibility, but modelling studies suggest this is less likely. The mechanism proceeding via initial C-5 abstraction (iii) is considered most likely.

6.4 Thienamycin biosynthesis

Labeling studies,^523,525,564 and the structures of isolated C-2- and C-6-functionalised carbapenems (Fig. 45) led to early proposals for the pathway of thienamycin biosynthesis.⁵²³ Sequencing of the thienamycin biosynthesis gene cluster (from S. cattleya)^354,565 has opened up the pathway for delineated genetic and biochemical investigations. In this section we describe the thienamycin gene cluster (Fig. 48B, Table 6) and then describe the reported labeling studies and recent biochemical investigations^{239,397,539,566} in the light of the genetic information^545,565,567 (Fig. 48A).

6.4.1 Overall description of the S. cattleya thienamycin gene cluster. The thienamycin gene cluster comprises more than 20 genes (Fig. 48B, Table 6).⁵⁶⁵ Analysis of the cluster reveals genes homologous to C3C biosynthesis genes in P. carotovorum: (i) thnE and thnM are homologues of carB and carA, respectively.⁵⁶⁵ Indeed, ThnE catalyses the same reaction as CarB (i.e. t-CMP formation from L-GHP and malonyl-CoA)^397,539 and, like CarA, ThnM catalyses the ATP-dependent β-lactam formation³⁹⁷ (Fig. 48A); (ii) thnG and thnQ encode for the 2OG oxygenases ThnG and ThnQ. Recent reports reveal ThnG and/or ThnQ as neither able to catalyse the epimerisation nor the desaturation reaction catalysed by CarC in C3C biosynthesis; instead they have been reported to be involved in oxidising the C-2 and C-6 side chains, respectively (Fig. 48A).⁵⁶⁶

Table 6 The genes constituting the reported⁵⁶⁵ thienamycin gene cluster and the (predicted) roles of the (putative) proteins that they (may) encode for. The predicted number of amino acid residues (AA) for each of the (putative) proteins is shown. Provisional assignment of function to the different proteins is adapted from^523,565 and is in part based on recent reports.^{239,539,545,566} Proteins with an experimentally assigned biochemical function are in bold

Gene	AA	(Proposed) function of encoded protein
thnA	259	Oxidoreductase, similar to 3-oxoacyl-[acyl-carrier-protein] reductase.
thnB	259	Lactone-dependent transcriptional regulator.^565
thnC	259	Lactone-efflux transmembrane protein.^565
thnD	259	Alcohol dehydrogenase.^565
thnE	294	ThnE (crotonase); t-CMP formation.^397,539
thnF	327	ThnF (N-Acetyltransferase); N-acetylthienamycin formation.^239
thnG	263	ThnG (2OG oxygenase); oxidation of C-2 side chain of thienamycins.^566
thnH	224	ThnH cleaves 4-phosphopantetheine to produce pantetheine.^239
thnI	476	Transcriptional activator essential for thienamycin biosynthesis.^567
thnJ	483	Transport protein involved in thienamycin secretion.^565
thnK	681	Methyltransferase.^565
thnL	474	Methyltransferase.^565
thnM	458	ThnM; β-Lactam synthetase.^397
thnN	367	Carboxylate reductase component (similar to griC).^568
thnO	472	Carboxylate reductase component (similar to griD).^568
thnP	484	Methyltransferase.^565
thnQ	259	ThnQ (2OG oxygenase); hydroxylation of C-6 side chain of thienamycins.^566
thnR	240	ThnR cleaves CoA to produce 4-phosphopantetheine.^239
thnS	329	β-Lactamase; probably involved in a resistance mechanism.^565
thnT	399	ThnT; pantetheine hydrolase.^239
thnU	268	Transcriptional activator for cephamycin C biosynthesis genes.^567
thnV	137	Cysteine transferase.^565

---

The transcriptional activator ThnI, encoded for by the thnI gene, is essential for thienamycin biosynthesis and regulates the expression of nine genes involved in thienamycin assembly and export (thnH, thnJ, thnK, thnL, thnM, thnN, thnO, thnP and thnQ); deletion of thnI ablates thienamycin biosynthesis but increases cephamycin C biosynthesis (through an up-regulation of pcbAB and cmcI transcription).⁵⁶⁷ ThnU activates cephamycin C biosynthesis genes which are not part of the thienamycin gene cluster.⁵⁶⁷ These results not only reveal some levels of cross talk between thienamycin and cephamycin C biosynthesis pathways, but also raise a question regarding the precise boundaries of the two gene clusters.⁵⁶⁷

6.4.2 Assembly of the bicyclic core of thienamycin (ThnN, ThnO, ThnE, ThnM). ThnN and ThnO show sequence similarity to GriC and GriD from the grixazone biosynthesis pathway in Streptomyces griseus;⁵⁶⁸ ThnN shows 33% identity and 49% similarity to GriC and ThnO shows 29% identity and 39% similarity to GriD. GriC and GriD have been proposed to reduce 3-amino-4-hydroxybenzoic acid into 3-amino-4-hydroxybenzaldehyde (Fig. 58).⁵⁶⁸ GriC shows homology to AMP-binding proteins and GriD shows homology to NAD(P)-dependent aldehyde dehydrogenases; together, GriC and GriD constitute an ATP- and NAD(P)-dependent carboxylic acid reductase. Thus, both ThnN and ThnO may be involved in catalysing the reduction of glutamic acid (derivative) to L-GHP as a thienamycin precursor. ThnE can catalyse the conversion of L-GHP and malonyl-CoA into t-CMP,^397,539 which is a substrate for ThnM to produce the (3S,5S)-carbapenem (Fig. 48A).³⁹⁷ The timing and nature of enzymes involved in C-5 epimerisation and C₂–C₃ desaturation are unclear.

Fig. 58 The possible role of the reductase GriC/GriD in grixazone biosynthesis in Streptomyces griseus.⁵⁶⁸

6.4.3 The C-6 hydroxy-(m)ethyl side chain (ThnL?, ThnK?, ThnP?, ThnQ). It is proposed (i) that the two carbon atoms of the C-6 hydroxyethyl side chain of thienamycin are introduced through separate methyl transfers from methionine,^523,525,564 and (ii) that the C-6 side chain hydroxylation occurs after introduction of at least one of the C-6 carbons because C-6 hydroxymethyl, ethyl and isopropyl carbapenems (i.e. northienamycin, PS-5, and PS-6, respectively) have been isolated from Streptomyces spp. (Fig. 45).
6.4.3.1 C-6 side chain labeling studies. The methyl group of [methyl-^l4C]methionine is incorporated into northienamycin and thienamycin.⁵²³ In a double labeling experiment, it has been demonstrated that [methyl-^l4C,³H]methionine is incorporated into thienamycin with 58% tritium retention, relative to ¹⁴C, corresponding to 87% of the maximum value for retention of 4 of the 6 hydrogens of the two methyl groups.⁵²³ Feeding [methyl-²H₃¹³C]-methionine to S. cattleya resulted in ¹³C-incorporation into thienamycin C-9 with retention of all three deuterium atoms and into C-8 with retention of one deuterium atom.⁵⁶⁴ Feeding (methyl-R)- and (methyl-S)-[methyl-²H_l,³H₁]-L-methionine to S. cattleya resulted in the production of thienamycin in which the methyl group has the same configuration as that in the starting methionine (Fig. 59),⁵⁶⁴ implying that the transfer of the C-9 methyl group (from methionine/S-adenosylmethionine) occurs with net overall retention of configuration. This result is unusual because most of the studied S-adenosylmethionine transferases catalyse methyl transfer with inversion of configuration, apparently, via an S_N2-type process.⁵⁶⁹ One possibility is the introduction of C-9 of thienamycin via a process involving two sequential methyl transfers, each proceeding with inversion.⁵⁶⁴ Since thienamycin fermentation has been reported to require cobalt,⁵²³ it was proposed⁵⁶⁴ that the two sequential methyl-transfers probably take place from 5-methyltetrahydrofolate, by a B₁₂-dependent synthase, analogous to the biosynthesis of methionine from 5-methyltetrahydrofolate, which also proceeds with net retention of configuration.^570–573 Vitamin B₁₂ has been reported to be excreted from a blocked mutant that does not produce thienamycin, and also to restore the thienamycin producing capacity to other blocked mutants.⁵⁷⁴ It is also noteworthy that methylmalonyl-CoA mutase, which interconverts (2R)-methylmalonyl-CoA and succinyl-CoA, is a vitamin B₁₂ dependent enzyme;^575–577 the dependence of thienamycin biosynthesis on cobalt/vitamin B₁₂ may imply that methylmalonyl-CoA is a potential in vivo substrate. However, there is no obvious vitamin B₁₂-dependent protein in the thienamycin gene cluster. Thus it seems probable that the methyl transfer occurs by a different mechanism (see below).

Fig. 59 Summary of the labeling studies on the side chains at C-2 (A) and C-6 (B) of thienamycin in Streptomyces cattleya. *Also, feeding L-[methyl-¹³C,²H₃]methionine to S. cattleya resulted in retention of all 3 deuterium atoms at ¹³C-9 and retention of one deuterium atom at ¹³C-8.^523,525,564

6.4.3.2 Putative roles of ThnK, ThnL and ThnP. In thienamycin biosynthesis, at least two of the three related putative radical SAM superfamily enzymes (i.e. ThnK, ThnL, and ThnP) are proposed to be involved in the methyl-transfer catalysis to afford the C-6 ethyl side chain of thienamycin.^545,565 SAM-dependent enzymes catalyse diverse reactions (including e.g. functionalisation of relatively unactivated C–H bonds, isomerisation, sulfur insertion, and elimination; for review on radical SAM enzymes, see ref. 578–580) employing the oxidising power of the generated 5′-deoxyadenosyl radical to abstract a hydrogen atom. Thus, ThnK, ThnL or ThnP may catalyse reactions beyond the predicted methyl-transfer in thienamycin biosynthesis (e.g. C-5 epimerisation, C₂-C₃ desaturation, ligation of the C-2 cysteaminyl/pantetheinyl side chain). Insertional inactivation of thnL results in a thienamycin non-producing strain.⁵⁶⁵ Insertional inactivation of thnP resulted in a thienamycin non-producing strain that accumulates a compound with a mass corresponding to that of carbapenam-3-carboxylate.⁵⁴⁵ This result, if confirmed by definitive characterisation of the accumulated metabolite, suggests the intermediacy of carbapenam-3-carboxylate in thienamycin biosynthesis, and also suggests that ThnP catalysis precedes that of ThnL.
6.4.3.3 ThnQ: hydroxylation of the C-6 (m)ethyl side chain. Recombinant ThnQ, a 2OG oxygenase, catalyses the stereoselective hydroxylation of the C-6 ethyl side chain of (i) PS-7 to produce N-acetyl-dehydro-thienamycin; and (ii) PS-5 to produce N-acetyl-thienamycin (Fig. 48A).⁵⁶⁶

6.4.4 The C-2 cysteaminyl side chain (ThnR, ThnH, ThnT, ThnF, ThnG). The C-2 side chain of many carbapenems is generally a derivative of cysteamine or dehydrocysteamine, in which the sulfur atom is occasionally at the sulfoxide oxidation state (Fig. 45). Several of the antibiotics in the pluracidomycin group have more oxidised/truncated C-2 substituents. The members of the OA-6129 subfamily are characterised by a pantetheinyl moiety at C-2 (Fig. 45).
6.4.4.1 C-2 side chain labeling studies. Feeding radiolabeled cystine to resting cells of S. cattleya resulted in a high level of incorporation of radioactivity into the cysteaminyl side chain of thienamycin.⁵²³ On the other hand, [³⁵S]cystamine and [³⁵S]pantethine are poorly incorporated compared to [³⁵S]cystine,⁵²³ implying that they are not (direct) precursors of the C-2 cysteaminyl side chain. Feeding a mixture of [3,3′-³H₂]-cystine and [³⁵S]cystine to S. cattleya resulted in the production of thienamycin with the same ³H/³⁵S ratio as the starting mixture (Fig. 59),⁵²³ consistent with the proposal that a dehydro-thienamycin compound is not an intermediate in thienamycin biosynthesis. Radiolabeled β-alanine is incorporated into the C-2 pantetheinyl side chain of the OA-6129 carbapenems, but labeled pantothenate was not incorporated.⁵⁸¹ On the basis of these findings and biochemical studies on ThnR, ThnH and ThnT (see below) it would seem likely that β-alanine is taken up by the mycelia and converted sequentially into (phospho)pantothenate, then (phospho)pantetheine and finally to coenzyme A (see below).
6.4.4.2 The acylase from Streptomyces fulvoviridis. Several naturally-occurring carbapenems contain a C-2 pantetheinyl side chain (Fig. 45). It is proposed that these are likely precursors of the analogous 2-cysteaminyl-carbapenems.⁵⁸² Mutation of S. fulvoviridis, which mainly produces the N-acetyl-2-cysteaminyl-carbapenems PS-5, and epi-thienamycins A, C, and F, has resulted in a strain that instead produces the 2-pantetheinyl-carbapenems OA-6129A, B₁, B₂ and C (Fig. 60A).⁵⁸² The mutant strain also lacks the acylase activity which removes the N-pantothenyl side chain of OA-6129A to give NS-5 (Fig. 60B).⁵⁸² The acylase in S. fulvoviridis (A933 acylase) which is involved in exchange of the pantothenyl moiety of OA-6129 carbapenems with acetyl-CoA (Fig. 60A) has been partially purified and characterised.^583,584 The A933 acylase also catalyses (i) the depantothenylation of OA-6129A to NS5 (Fig. 60B);⁵⁸² (ii) the acylation of 2-cysteaminyl-carbapenems with acyl-CoA (Fig. 60B); (iii) the acylation of 6-aminopenicillinic acid with acyl-CoAs;⁵⁸³ and (iv) the deacylation of N-acetyl-L-amino acids, but not that of N-acetyl-D-amino acids.⁵⁸⁴ The A933 acylase does not catalyse the deacetylation of PS-5 to NS-5.⁵⁸³ The acylase is inhibited by cobalt ions and p-chloromercunbenzoate, implying that it has an active thiol.⁵⁸⁴ A similar acylase activity has also been detected in the following carbapenem producers: S. cattleya (see below), S. cremeus and S. argentcolis.⁵⁸³ The acylase from S. cattleya has been reported, by Kubo et al., as labile and not amenable to purification, but with a similar pattern of specificity to that of S. fulvoviridis acylase.⁵⁸³

Fig. 60 Transacylase-catalysed reactions with substituted carbapenems. A and B: A933 (Streptomyces fulvoviridis) catalysed reactions;^582–584 and C: ThnT (S. cattleya) catalysed reaction.²³⁹

6.4.4.3 (Putative) roles of ThnR, ThnH, ThnT, ThnF and ThnG. Efforts aiming at elucidating the enzymology of the C-2 side chain assembly have shown that recombinant ThnR, ThnH and ThnT incrementally cleave CoA to 4-phosphopantetheine, pantetheine, and cysteamine, respectively (Fig. 48A).²³⁹ ThnR, a CoA pyrophosphatase, shows similarity to the Nudix hydrolase family.⁵⁸⁵ ThnH, which hydrolyses the phosphate of 4-phosphopantetheine, shows weak similarity to the haloacid dehalogenase superfamily of hydrolases.⁵⁸⁶ ThnT, an Ntn hydrolase⁵⁸⁷ (Section 4.5), catalyses the depantothenylation of both cis- and trans- 2-pantetheinyl-carbapenams into the 2-cysteaminyl analogues²³⁹ (Fig. 60C), suggesting that ThnT might be a homologue of the A933 acylase (see above). However, Blanco et al. reported that insertional inactivation of thnR and thnT does not affect thienamycin production in the mutant strains, raising the question whether they are involved in thienamycin biosynthesis;⁵⁴⁵ one possibility is that the mutant strains have another set of CoA-processing enzymes. The mechanism and timing of ligation of CoA/CoA-derivative into the bicyclic carbapen(am/em) remain to be determined.

ThnF, a putative acetyltransferase, which is weakly related to the GNAT superfamily (Section 5.10), has been predicted to catalyse the conversion of thienamycin to N-acetyl-thienamycin. ThnF has been shown to catalyse N-acetylation of a model compound containing cysteamine in the presence of acetyl-CoA as a co-substrate (Fig. 61).²³⁹

Fig. 61 ThnF-catalysed reaction with a substrate analogue.²³⁹ Note that the putative substrate for ThnF is thienamycin (Fig. 44).²³⁹

ThnG, a 2OG oxygenase, catalyses the oxidation of the N-acetyl-2-cysteaminyl side chain of (i) N-acetyl-thienamycin to produce N-acetyl-dehydro-thienamycin (desaturation); and (ii) PS-5 to produce PS-7 and PS-7 sulfoxide (desaturation followed by sulfoxidation, Fig. 48A).⁵⁶⁶ Insertional inactivation of thnG resulted in a mutant strain that showed (i) 2.5-fold increase in thienamycin production; and (ii) accumulation of a metabolite with a mass corresponding to that of 2,3-dihydro-thienamycin in the mutant strain.⁵⁴⁵ Overall, these results imply that ThnG is not essential for thienamycin biosynthesis, but is required for C-2 oxidation of thienamycin derivatives.

6.4.4.4 ThnT mechanism of autoproteolysis and structure. The autoproteolysis of the Ntn hydrolase ThnT has been studied in detail, and is proposed to be initiated by attack of the activated alcohol of Thr282 onto the carbonyl carbon of the scissile bond, which is located in an oxyanion hole (OAH, Fig. 62C).⁵⁸⁷ Two water molecules in ThnT active site are proposed to facilitate autoproteolysis via deprotonation of Thr282-OH, protonation of the amine leaving group in the oxazolidine intermediate, and hydrolysis of the ester intermediate to release the Ntn nucleophile Thr282 (Fig. 62A).⁵⁸⁷ A crystal structure of ThnT Thr282Cys variant led to the observation of two conformations of the backbone sequence in the region of the scissile bond (Fig. 62B); one of the two conformations has the carbonyl oxygen of the scissile bond in the OAH (cleavage-competent form, Fig. 62C).⁵⁸⁸

Fig. 62 Proposed mechanism and structural views of the pantetheine hydrolase ThnT.⁵⁸⁷A: Proposed autoproteolysis mechanism of ThnT; B: Dual occupancy of the ThnT Thr282Cys variant active site by shown residues (residues in cyan represent the proposed cleavage-competent conformer, while those in white represent the inactive conformer); C: Model of the cleavage-competent form of ThnT. Note the position of the carbonyl-oxygen of the scissile amide bond in the oxyanion hole (OAH).

7 Monocyclic β-lactam biosynthesis

The two most populated subfamilies of naturally-occurring monocyclic β-lactams are the monobactams (3-aminomonobactamic acid derivatives, Fig. 63A) and the nocardicins (3-aminonocardicinic acid derivatives, Fig. 63B). In addition, there are the tabtoxins (with unsubstituted nitrogen of the β-lactam core, Fig. 63C), and the “conjugate” β-lactams (where the β-lactam nucleus is N-linked to a terpenoid, Fig. 63D). The discovery of conjugate β-lactams is notable as it marks the apparent ability of higher plants to produce β-lactams.

Fig. 63 Monocyclic β-lactams isolated from natural sources: The monobactams (A), the nocardicins (B), the tabtoxins (C), and the conjugate β-lactams (D). Some stereochemical assignments are provisional. The nuclei for monobactam and nocardicin subfamilies are boxed. Sulfazecin and its epimer isosulfazecin are produced by Pseudomonas acidophila and P. mesoacidophila, respectively.^761,762 Sulfazecin has also been isolated from Gluconobacter and Acetobacter.^763,764 SQ 26,180 has been isolated from Chromobacterium violaceum ATCC 31532.^763–765 SQ 26,700, SQ 26,812, SQ 26,823, SQ 26,875 and SQ 26,970 (all characterised by a C-3-substituent with an aromatic side chain) were isolated from Agrobacterium radiobacter ATCC 31700.^763,766,767 The monobactams SQ 28,332,⁷⁶⁸ SQ 28,502 and SQ 28,503, with the latter two having longer C-3 oligo-peptide side chains, were isolated from Flexibacter sp.⁷⁶⁹ PB-5266 A, B and C were isolated from Cytophaga johnsonae.^770,771 MM 42842, the only known naturally-occurring monobactam with a 4β-methyl-substituent, was isolated from P. cocovenenans.^772,773 Nocardicins A–G were isolated from Nocardia uniformis.^{78,609,610,774,775} Nocardicins are also produced by Actinosynnema mirum,⁷⁷⁶Nocardiopsis atra⁷⁷⁷ and Microtetraspora caesia.⁷⁷⁸ Chlorocardicin was isolated from a Streptomyces sp.^605,606 Formadicins A–D were isolated from Flexibacter alginoliquefaciens.^607,608 Nocardicin A and B are stereoisomers differing only in the oxime stereochemistry.^611,775 Isotabtoxin (red dashed box) is a stable product of tabtoxin rearrangement; tabtoxinine (blue dashed box) is the hydrolysis product of tabtoxinine-β-lactam.

The monobactams and nocardicins are inhibitors of bacterial cell wall biosynthesis.⁵⁸⁹ Only three members of the tabtoxin subfamily have been isolated: tabtoxinine β-lactam, its precursor tabtoxin, and N-((S)-alanyl)-(5S)-5-chloro-tabtoxinine-β-lactam (Fig. 63C). Tabtoxin and its hydrolysis product, tabtoxinine β-lactam, are involved in the induction of the wildfire disease of tobacco plant likely through inhibition of glutamine synthetase (see below). N-((S)-Alanyl)-(5S)-5-chloro-tabtoxinine-β-lactam has been isolated from a previously unidentified Streptomyces species 372A⁵⁹⁰ and found to inhibit the growth of strains of gram-positive and gram-negative bacteria. This inhibition can be relieved by addition of L-glutamine,⁵⁹⁰ suggesting that, like tabtoxin, it is likely a glutamine synthetase inhibitor. There are few examples of the conjugate β-lactams (Fig. 63D); however, nothing has been reported yet about the biosynthesis of their β-lactam part.⁵⁹¹ The antiulcer^592,593 and anticancer⁵⁹⁴ steroidal-β-lactams pachystermine A and pachystermine B have been isolated from Pachysandra terminalis.^595,596 16α-Hydroxy-pachystermine A has been isolated from Pachysandra procumbens.⁵⁹⁷ The antiestrogen-binding site inhibitor dehydro-pachystermine A and its 16α-hydroxy-derivative have also been isolated from Pachysandra procumbens.⁵⁹⁸ The antimalarial diterpene-β-lactam monamphilectine A has been isolated from the marine sponge Hymeniacidon sp.⁵⁹⁹

In the following sections, we outline biosynthetic studies on monobactams, nocardicins, and tabtoxins.

7.1 Monobactam biosynthesis

All of the identified monobactams possess an N-sulfonated-β-lactam, but differ in their C-3 acylamido side chain and in the presence/absence of a C-3-α-methoxy group (Fig. 63A). The N-sulfonic acid group is proposed to activate the β-lactam carbonyl for reaction with nucleophiles and enable binding to the PBP active site in an analogous way to the penicillin C-2 carboxylate.^600,601 The 4-methyl group in MM 42842 (Fig. 63A) is proposed to enhance the stability of the β-lactam ring towards β-lactamase attack and to improve antibacterial activity (as occurs in the case of aztreonam).⁶⁰⁰

Pioneering studies on the biosynthesis of the β-lactam of sulfazecin, SQ 26,180 and SQ 26,812 using double-labeled amino acids imply that the carbon atoms of the β-lactam of these compounds are derived from serine.⁶⁰² Feeding experiments using 3-[³H]₂-serine indicate that ring closure occurs without change of oxidation state of the serine β-carbon as little [³H]-loss was observed.⁶⁰² These observations led to the proposed mechanism for β-lactam formation (Fig. 64), analogous to that of nocardicin biosynthesis (see below), including an S_N2-type displacement of an activated serine hydroxyl group.

Fig. 64 Proposed biosynthetic pathway leading to naturally-occurring monobactams (on the basis of labeling studies as reviewed in ref. 589). The order of steps is uncertain. APS: adenosine-5′-phosphosulfate. SAM: S-adenosylmethionine; SAHC: S-adenosylhomocysteine.

Experiments to investigate the origin of the sulfamate sulfur of some naturally-occurring monobactams have revealed that inorganic sulfur is the only source utilised in the studied organisms.⁶⁰³ It is proposed that the sulfamate group of monobactams is produced via an activated sulfate ion; sulfate activation might be achieved via the formation of adenosine-5′-phosphosulfate (APS, which is produced from ATP and sulfate as catalysed by sulfate adenylyl-transferase) or 3′-phosphoadenosine-5′-phosphosulfate (PAPS, as catalysed by adenylyl-sulfate kinase). Cell free extracts of Agrobacterium radiobacter produce APS upon incubation of sulfate and ATP with the appropriate cofactors, in support of sulfate activation preceding sulfamate formation.⁶⁰³ The question of whether sulfur introduction precedes β-lactam formation is unanswered. The observation of a sulfated peptide in a sulfazecin producer suggests sulfation might occur prior to ring closure; the peptide contains N-terminal glutamate, alanine, serine (incorporated with retention of its C-3 hydrogens), and sulfur.⁵⁸⁹ The MS and electrophoretic data of the peptide suggest a molecule with properties that fit a substrate for a β-lactam synthetase (Fig. 64). However, incubation with cell free extracts of Acetobacter did not lead to β-lactam formation.⁵⁸⁹

The absence of the 3α-methoxy group in some naturally-occurring monobactams suggests that the methoxy group is introduced after β-lactam formation, as occurs in cephamycin biosynthesis (Section 4.8). The methyl of the 3α-methoxy group has been shown to originate from methionine,⁶⁰² likely via methylation of the 3α-hydroxyl group with S-adenosylmethionine (SAM). Again, by analogy to cephamycin biosynthesis, it is reasonable to propose that hydroxylation of the β-lactam is catalysed by a 2OG oxygenase.

The isolation of isosulfazecin, which differs from sulfazecin in the presence of an L- rather than D-alanyl residue in its C-3 side chain, may suggest that isosulfazecin is an intermediate in the biosynthesis of sulfazecin, and that epimerisation occurs, at a late stage, following both β-lactam formation and C-3α-methoxylation. If so, this represents a further example of production of structural diversity by epimerisation in β-lactam biosynthesis, as demonstrated in the penicillin, cephalosporin, clavam, carbapenem, and nocardicin (see below) biosynthesis pathways (Fig. 9, 27, 47 and 66, respectively).

Fig. 65 A penicillin G acylase from E. coli catalyses the acylation of 3-amino-monobactamic acid and its 4-methyl derivative.⁶⁰⁴

Fig. 66 Nocardicin A biosynthesis. A: Proposed biosynthetic pathways leading to naturally-occurring nocardicins; B: The cluster of genes⁶²⁴ encoding for the proteins proposed to be involved in nocardicin A biosynthesis in Nocardia uniformis. See Table 7 for the proposed role of the encoded proteins. pHPP: p-hydroxyphenylpyruvic acid; pHPG: p-hydroxyphenylglycine.

The variety of C-3 acyl-amido side chains observed in naturally-occurring monobactams has led to the hypothesis that they are derived from a common biosynthetic precursor. As yet, however, a monobactam acyl-transferase, similar to the one in penicillin biosynthesis, has not been identified.⁵⁸⁹ O'Sullivan and Aklonis⁶⁰⁴ have reported that an immobilised penicillin acylase from E. coli can catalyse the acylation of both 3-aminomonobactamic acid and 4α-methyl-3-aminomonobactamic acid (SQ 26 771), when incubated with a variety of acylating agents (e.g. phenylacetic acid and 2-formamidothiazol-5-yl acetic acid, Fig. 65) at pH 4.5. However, in a deacylase reaction, none of the naturally-occurring monobactams has been found to be a substrate for the acylase,⁶⁰⁴ consistent with biosynthetic reports that addition of possible side chain precursors failed to enhance the levels of monobactam produced.⁶⁰²

7.2 Nocardicin biosynthesis

The nocardicin subfamily of monocyclic β-lactams bear two characteristic aromatic side chains, one attached through the β-lactam nitrogen, and the other is part of a p-hydroxy-phenylacetamido system attached to the β-lactam at C-3. The p-hydroxy group can be functionalised by an ether linked L- or D-homoserinyl moiety (Fig. 63B). The structural diversity of nocardicins originates partly from the nature of the substituents at C-2′ which can be an amine, a ketone, or syn/anti-oxime function (Fig. 63B). The p-hydroxy-phenylacetic acid moiety, attached at its α-carbon to the β-lactam nitrogen, is another source of introducing diversity into the nocardicins (Fig. 63B): (i) Chlorocardicin^605,606 is identical to nocardicin A except for the presence of a m-chlorine; (ii) formadicins A and B are p-O-glycosylated nocardicins with formadicin A bearing a 3α-formamido group; formadicins C and D are non-glycosylated analogues with formadicin C bearing a 3α-formamido group.^607,608

The antibacterial activities of nocardicins B, C, D and E, and formadicins B and D are weaker than those of nocardicin A, chlorocardicin, and formadicins A and C, while nocardicins F and G are devoid of antibiotic activity.^609–611 Therefore, the presence of an oxime, and its syn-relationship to the acylamino group, as well as the presence of a D-homoserinyl ether side chain are important for antimicrobial activity.⁶⁰⁹ Note that the formadicins and some cephabacins (Fig. 8) have an α-formamido substituent at analogous positions on the β-lactam ring; the 6α-/7α-formamido substituent has been shown to improve the β-lactamase stability and activity of some penicillins and cephalosporins.^612,613

7.2.1 Pioneering labeling studies. Initial feeding experiments using radiolabeled amino acids supplemented to Nocardia uniformis growing cultures established that the nocardicin core is assembled from tyrosine (viaL-(p-hydroxyphenyl)-glycine (L-pHPG) into the two aromatic moieties), serine (into the β-lactam ring) and L-methionine (into the homoserinyl moiety).^614,615 As in β-lactam formation in monobactam biosynthesis (Section 5.2), experiments with 3-[³H]₂-serine revealed that β-lactam formation occurs without loss of [³H], hence does not involve an oxidative ring closure mechanism⁶¹⁵ (Fig. 66A) as occurs in penicillin biosynthesis (Fig. 13). Investigations employing L- and D-serine stereoselectively labeled with tritium at the β-carbon revealed that L-serine is much better incorporated and that β-lactam ring closure proceeds with inversion of the stereochemistry at this carbon,⁶¹⁵ suggesting that ring closure occurs via an S_N2-type mechanism with nucleophilic displacement of an activated serine hydroxyl group. Such a proposed mechanism is supported by biomimetic transformations of serine-containing peptides under Mitsunobu-type conditions leading to nocardicins and related structures.^616–618

Experiments employing (2S,4R)- and (2S,4S)-[4-²H]methionine revealed that the ether-linked D-homoserinyl moiety of nocardicin A is efficiently precursored by L-methionine via direct nucleophilic displacement of S-adenosylmethionine (SAM, Fig. 66A).^615,619,620

Feeding of racemic [2-¹³C,¹⁵N]-pHPG demonstrated that pHPG provides the nitrogen atoms of both the β-lactam and of the oxime in nocardicin A.⁶²¹ Employing L-[2-³H, 1-¹⁴C]-pHPG, it was found that tritium is not incorporated into the 5-position of nocardicin A.⁶¹⁹ Feeding a labeled sample of (DLD)-nocardicin G618 and (LLD)-epinocardicin G revealed intact incorporation of nocardicin G into nocardicin A,⁶²² suggesting the intermediacy of the tripeptide D-pHPG-L-Ser-D-pHPG and consistent with epimerisation of the pHPG residues during/after peptide formation (Fig. 66A), possibly in an analogous fashion to that of the valinyl residue by ACV synthetase during the tripeptide LLD-ACV formation in penicillin biosynthesis (Fig. 11). Furthermore, it was reported that nocardicin E is converted into isonocardicin A (with L-configuration in its homoserinyl side chain) and, consequently, via epimerisation, into nocardicin A in cell-free extracts supplemented with SAM.⁶²³ Taken together, these observations imply that the simplest nocardicin, nocardicin G, is the first formed nocardicin β-lactam, and that other nocardicins are derived from nocardicin G following the introduction of the homoserinyl moiety (Fig. 66A).

7.2.2 Nocardicin A biosynthesis gene cluster. The nocardicin A biosynthesis gene cluster in Nocardia uniformis has been sequenced (Fig. 66B, Table 7).⁶²⁴ Recently, the genome of Actinosynnema mirum, another nocardicins producer, has been sequenced³⁵⁵ and found to contain a near identical biosynthetic cluster (Identities = 33363/33461 (99%), Gaps = 43/33461 (0%)). Both clusters comprise 14 open reading frames involved in biosynthesis, resistance and transport. Three genes (nocF, nocG and nocN) encode for proteins likely involved in pHPG biosynthesis (Fig. 66A).⁶²⁴ Two genes (nocA and nocB) encode for non-ribosomal peptide synthetases (NPRSs) and likely catalyse formation of an oligopeptide precursor (Fig. 66A). Three genes (nocC, nocJ and nocL) encode for enzymes involved in the late steps of nocardicin A biosynthesis (see below).^625–627 The predicted product of nocH is related to membrane transport proteins and is proposed to be involved in exporting nocardicin A from the cell.⁶²⁸ The product of nocD is a putative acetyltransferase, possibly involved in the resistance mechanism.⁶²⁴ The product of nocR is a positive transcriptional regulator of the nocardicin A biosynthesis pathway coordinating the initial steps of nocardicin A biosynthesis for the production of the pHPG precursor.⁶²⁸ Inactivation of nocR resulted in a variant unable to produce detectable levels of nocardicin A or the early precursor p-hydroxybenzoyl formate.⁶²⁸ The function of the three remaining genes (nocE, nocI and nocK) is unclear.^627,628

Table 7 The genes constituting the reported nocardicin A gene cluster⁶²⁴ in Nocardia uniformis and the (predicted) roles of the (putative) proteins that they (may) encode for. The predicted number of amino acid residues (AA) for each of the (putative) proteins is shown. Proteins with an experimentally assigned biochemical function are in bold

Gene	AA	(Proposed) function of encoded protein
nocN	376	p-Hydroxymandelate oxidase.
nocR	582	Transcriptional regulator.^628
nocL	398	NocL (Cytochrome P-450 oxidase).^623,636
nocK	344	Hydrolase/esterase.
nocJ	327	NocJ (Epimerase).^626
nocI	73	Unknown.
nocH	408	Membrane transport protein.^628
nocG	431	p-Hydroxyphenylglycine transaminase.
nocF	345	p-Hydroxymandelate synthase.
nocA	3692	Non-ribosomal peptide synthetase.^624,779
nocB	1925	Non-ribosomal peptide synthetase.^624,779
nat	301	3-Amino-3-carboxypropyl transferase.^625
nocD	185	Acetyltransferase.
nocE	1414	Unknown.

---

7.2.2.1 NocF, NocN and NocG (pHPG biosynthesis genes). NocF, NocG, and NocN are p-hydroxymandelate synthase, putative p-hydroxyphenylglycine transaminase, and p-hydroxymandelate oxidase, respectively.⁶²⁴ By analogy to other clusters which encode for proteins involved in pHPG biosynthesis,^629–631 the reported nocardicin biosynthesis cluster is lacking a gene encoding for a prephenate dehydrogenase.⁶³² NocF stereospecifically catalyses conversion of p-hydroxyphenylpyruvate to (S)-p-hydroxymandelate; NocN oxidises the latter into p-hydroxybenzoylformate (Fig. 66A).⁶²⁴ The predicted transaminase activity of NocG⁶³² awaits experimental testing.
7.2.2.2 NocA and NocB (NRPSs enzymes). On the basis of initial labeling studies (see above) and by analogy to ACVS catalysis (Fig. 11), NocA and B were predicted to catalyse the formation of the tripeptide L-pHPG-L-Ser-L-pHPG and epimerisation into its DLD-isomer, which is a proposed precursor for an S_N2 ring closure reaction to give nocardicin G (Fig. 66A).

Analyses of the predicted NocA and NocB sequences (3692 and 1925 amino acid residues, respectively) for NRPS motifs¹⁵⁰ imply that together they possess a 5-module structure (3 in NocA and 2 in NocB) including 5 adenylation domains (A1–A5), 4 condensation domains, and a single epimerisation domain, with the 4th module being split between NocA and NocB: the condensation domain being located in NocA while the adenylation and thiolation domains are located in NocB.⁶²⁴ This organisation is unusual as it implies that the NRPS machinery of NocA and NocB does not follow the “colinearity paradigm”.⁶³³ On the basis of sequence analyses and homologies, the substrate specificities of the NocA and NocB adenylation domains have been predicted as follows: A1 L-pHPG, A2 L-ornithine (L-Orn) or L-N^ε-hydroxyornithine (L-N^εhOrn), A3 L-pHPG, A4 L-serine and A5 L-pHPG.^634,635 Consequently, a pentapeptide product with the sequence pHPG-X-pHPG-Ser-pHPG has been predicted (X refers to Orn or N^εhOrn). However, the proposed precursor for the predicted S_N2-type ring closure is the tripeptide D-pHPG-L-Ser-D-pHPG. Further analysis of the NocA sequence has revealed two short unusual amino acid repeats near A1 and T2, suggesting that the first two modules of NocA might be inactive due to a change in their tertiary structure, and that the remaining three modules are involved in the formation of the proposed precursor.⁶²⁴ Another possibility is that the pentapeptide is processed to yield a tripeptide precursor, either through self-editing via intramolecular attack of the (hydroxy)ornithine side chain onto the X-pHPG peptide bond, via a 6-exo trig cyclisation, (Fig. 66A) or as a result of peptidase activity. A similar cyclisation reaction had originally been proposed for coelichelin biosynthesis⁶³⁶ prior to elucidation of the correct structure of coelichelin.⁶³⁷

Only one epimerisation domain has been identified in module 3, and this may catalyse epimerisation of the L-pHPG substrate of A3 upon peptide formation. The necessary epimerisation of the L-pHPG substrate of A5 has been proposed to occur⁶²⁴via acid–base catalysis, exploiting the base-labile nature of the C-5 benzylic proton, or following completion of the peptide product, by a separate peptide epimerase, possibly as in case of serine epimerisation in a 48-amino acid peptide in funnel web spider venom, Agelenopsis aperta.⁶³⁸

To date, attempts to produce active NocA and NocB by recombinant expression have been unsuccessful.⁶²⁴ Two endogenous proteins of 200 kDa (identified as NocB, calculated mass of NocB is 206 kDa) and 150 kDa (identified as a fragment of NocB) have been purified from N. uniformis.⁶²⁴ NocB catalyses ATP/PP_i exchange substantially faster in the presence of L-pHPG than in the case of D-pHPG, supporting the proposal that L-pHPG is the substrate for the A5 adenylation domain and suggesting that epimerisation takes place during or after, but not prior to, peptide formation.⁶²⁴ Exchange in the presence of L-Ser was less efficient. NocB has been shown to contain a reactive sulfhydryl group and 4′-phosphopantetheine prosthetic group; however the nocardicin gene cluster is missing a gene encoding for a phosphopantetheinyl transferase required in the conversion of apo-NocA/B to holo-NocA/B.⁶²⁴

Thus, important aspects of NocA/B catalysis remain unclear including the stereochemical course of tripeptide formation and, of particular interest, whether NocA/B are directly involved in β-lactam formation.

7.2.2.3 Nat (NocC). Nat, S-adenosylmethionine (SAM):nocardicin 3-amino-3-carboxypropyltransferase, catalyses the transfer of the 3-amino-3-carboxypropyl moiety of SAM to the phenolic –OH of nocardicin G to form isonocardicin C (Fig. 66A).⁶²⁵ This S_N2-type of transfer of the sterically hindered methionine-derived moiety of SAM, rather than the methyl group, is unusual; the structural basis for this preference is proposed to reside on the orientation of the substrate binding to fulfil the geometric requirements for an S_N2 reaction, possibly via shielding the methyl group while exposing the 3-amino-3-carboxypropyl moiety for nucleophilic attack.⁶²⁵ Nat displays weak sequence similarity to bacterial SAM-utilising enzymes.⁶²⁵ The C-9′ epimerisation of isonocardicin C to nocardicin C is unrelated to Nat activity (see below). Kinetic analyses revealed a sequential mechanism for Nat and that nocardicin G is a better substrate than nocardicins E/F, supporting the proposed sequence of events in nocardicin A biosynthesis (Fig. 66A).
7.2.2.4 NocJ. The pyridoxal 5′-phosphate (PLP)-dependent enzyme NocJ catalyses the reversible C-9′ epimerisation at the homoserinyl side-chain of nocardicins (Fig. 66A).⁶²⁶ In order to avoid the limitation that isonocardicin A and nocardicin A are indistinguishable by some techniques, e.g. reverse phase HPLC, and to confirm the epimerisation at C-9′ of nocardicin A, the NocJ-reaction has been carried out in deuterated buffer and resulted in a loss of the H-9′ signal in the ¹H NMR spectrum.⁶²⁶ Insertional inactivation of nocJ abolishes nocardicin A production but does not affect that of isonocardicin A, consistent with the proposed function of NocJ.⁶²⁶ Despite their functional and co-factor dependence similarity, NocJ shows no significant sequence similarity to prokaryotic isopenicillin N epimerases (Section 4.6),⁶²⁶ but shows similarity to PLP-dependent deaminases.⁶³⁹
7.2.2.5 NocL. NocL is a cytochrome P450 oxidase which catalyses oxidation of the nocardicin C C-2′ amine to produce the syn-oxime of nocardicin A (Fig. 66A) in the presence of spinach ferredoxin, spinach ferredoxin-NADP⁺ reductase and NADPH.^627,640 The reported nocardicin A cluster does not possess genes encoding for NAD(P)H-dependent reductase or a ferredoxin likely required for NocL catalysis. The NocL sequence possesses both a putative heme binding motif (FGHGxHxCLG),⁶⁴⁰ and an oxygen binding motif (LLxAGHET).^641,642 Nocardicin G is not a substrate for NocL in vitro, suggesting the importance of the homoserinyl side chain for NocL activity, and raising the question of how nocardicins E and F are biosynthesised;⁶⁴⁰ one possibility is through the removal of the homoserinyl side chains of nocardicins A and B, respectively. The order of the C-9′ epimerisation/oxime formation is unclear. Disruption of nocL results in a strain unable to produce nocardicin A but still able to produce nocardicin C.⁶²⁷ The mechanism of NocL catalysis is proposed to proceed via two successive 2′-N-hydroxylations, followed by elimination of water, facilitated by the delocalisation of the neighbouring amide bond, and tautomerisation of the resulting nitroso species to yield an oxime (Fig. 67).⁶⁴⁰ The greater abundance of the syn-oxime (nocardicin A) relative to the anti-oxime (nocardicin B) isolated from N. uniformis fermentation broth^78,610 is proposed to result from intramolecular hydrogen bonding, during the course of the NocL-catalysed oxidation, which can facilitate the dehydration reaction and stabilise the conformation of the nitroso group leading to preferential formation of the syn oxime (Fig. 67).⁶²⁷

Fig. 67 Proposed outline mechanism for the cytochrome P450 NocL-catalysed oxime formation.^627,640

7.3 Tabtoxin biosynthesis

Tabtoxin is a structurally unique monocyclic β-lactam dipeptide because it is not functionalised at its β-lactam nitrogen (Fig. 63C). Elucidation of the tabtoxin structure⁶⁴³ and stereochemistry^{590,644–646} took a relatively long time, largely due to its instability; tabtoxin undergoes a facile intramolecular transacylation to give the stable, inactive δ-lactam, isotabtoxin (Fig. 63C). Tabtoxin is an exotoxin precursor produced by the phytopathogenic bacterium Pseudomonas syringae pv. tabaci and is involved in the induction of the wildfire (infectious leafspot) disease of tobacco (Nicotiana tabacum, which occasionally obliterated whole fields of tobacco plants).^643,646,647 In the infected host, tabtoxin is proposed to be hydrolysed by a peptidase⁶⁴⁸ to release the tabtoxinine-β-lactam (TBL, the actual wildfire toxin,⁶⁴⁹Fig. 63C) which inhibits the host glutamine synthetase, consequently resulting in the accumulation of toxic concentrations of ammonia and the characteristic chlorosis.^649–653 It should be noted that TBL production is linked to the presence of zinc in the growth media⁶⁵⁴ and zinc is required for the aminopeptidase activity leading to the cleavage of tabtoxin to yield TBL.⁶⁴⁸ Self-protection of the producing bacterium has been associated with the adenylation of its glutamine synthetase, which renders the modified enzyme less susceptible to inactivation by TBL.⁶⁵⁵ A second potential self-resistance mechanism involves the production of β-lactamases which hydrolyse the β-lactam of TBL to liberate the nontoxic metabolite, tabtoxinine (Fig. 63C).^656,657 Recently, a tabtoxin-resistance gene (ttr) has been characterised⁶⁵⁸ and used to enable the development of transgenic tobacco cultivars resistant to P. syringae.⁶⁵⁹ TTR is an acetyltransferase and contains the four motifs conserved in the GNAT superfamily (Section 5.10).⁴⁸³

The mechanism of β-lactam ring closure in tabtoxin biosynthesis is of interest; analyses of the biosynthetic gene cluster of tabtoxin⁶⁶⁰ (see below) suggests that β-lactam formation is likely mediated by a β-lactam synthetase, TblS, which is less closely related to the identified β-lactam synthetases than they are to each other (Fig. 31). The β-lactam C-3 hydroxylation is likely catalysed by a 2OG oxygenase. It has been proposed that an N-formylated α-amino carbonyl compound might be an intermediate during the formation of the 3-hydroxy-β-lactam nucleus of tabtoxin;⁶⁶¹ such compounds have been shown to undergo ring closure to yield 3-hydroxy-β-lactams in a potential biomimetic fashion.⁶⁶²

7.3.1 Tabtoxin labeling studies. Independent feeding experiments (using isotopically-labeled amino acids,⁶⁶¹ glucose⁶⁶³ and pyruvate⁶⁶⁴) have revealed potential precursors for tabtoxin biosynthesis in P. syringae: (i) the C-3 side chain is derived from L-threonine and L-aspartate; (ii) the C-2 carbonyl of the β-lactam is derived, uniquely to date, from the methyl group of L-methionine; and (iii) C-3 and C-4 of the β-lactam ring originate from C-3 and C-2 of pyruvate, as an intact unit (Fig. 68A). The incorporation of both aspartate and pyruvate has led to the hypothesis that tabtoxin biosynthesis is related to that of lysine (Fig. 68A);⁶⁶⁴ the observation that lysine itself is not efficiently incorporated suggests that tabtoxin biosynthesis deviates from that of lysine at an unknown common intermediate stage (see below).⁶⁶⁴

Fig. 68 Tabtoxin biosynthesis. A: Proposed biosynthetic pathway leading to tabtoxin (on the basis of reported labeling studies^661,663,664); B: The gene cluster encoding for the proteins proposed⁶⁶⁰ to be involved in tabtoxin biosynthesis in Pseudomonas syringae. The enzymes involved in lysine biosynthesis are in red. See Table 8 and text for full description of the cluster. Steps in lysine biosynthesis are shown to illustrate that 2-amino-6-oxopimelate (AOP) is proposed as a “branch point” between lysine and tabtoxin biosynthesis.

7.3.2 Tabtoxin biosynthesis gene cluster. The tabtoxin biosynthesis gene cluster (Fig. 68B) in P. syringae has been sequenced and 20 open reading frames have been predicted.⁶⁶⁰ It has been demonstrated that a fragment containing orf3–15 leads to tabtoxin production as well as TBL resistance. A fragment comprising orf12–15 is associated with TBL resistance alone.⁶⁶⁰ For description of the (putative) roles of the products of genes of the tabtoxin gene cluster, see Table 8. Notably, the tabP (orf4) gene product is a putative zinc metallopeptidase possibly involved in tabtoxin hydrolysis to give TBL. The gene product of tblS (orf10) is a putative β-lactam synthetase (see Fig. 29). The product of tblC (orf11) is a putative 2OG oxygenase, further revealing the diverse roles of these enzymes in BLA biosynthesis.⁶⁶⁰

Table 8 The genes constituting the reported tabtoxin biosynthesis gene cluster⁶⁶⁰ in Pseudomonas syringae and the (predicted) roles of the (putative) proteins that they may encode for. The predicted number of amino acid residues (AA) for each of the putative proteins is shown

Gene	AA	Proposed function of encoded protein
tabP	388	Zinc metallo-peptidase.
tabD	397	PLP-dependent aminotransferase.
tabB	276	Acetyltransferase.
tabA	420	PLP-dependent decarboxylase.
tblA	252	SAM-dependent methyltransferase.
tabC	412	Zinc-binding protein.
tablS	621	β-Lactam synthetase.
tblC	287	2OG oxygenase.
tblD	694	Fused oxido-reductase and GNAT-acetyltransferase.
tblE	129	Membrane protein.
tblF	213	D-Ala-D-Ala ligase.
tblR	873	Major facilitator superfamily transporter.

---

The orf6–8 genes are important for tabtoxin biosynthesis. The tabA (orf7) gene⁶⁶⁵ is not required for lysine formation despite the significant sequence similarity of its product to prokaryotic diaminopimelate decarboxylase, suggesting that the TabA substrate may be a DAP analogue (Fig. 68A).⁶⁶⁵ It has also been demonstrated that the dapB gene product is required for both DAP and tabtoxin formation, providing evidence for an overlap between the tabtoxin and lysine biosynthesis pathways, and suggesting that the branch point of the pathways is likely after THDPA, but prior to DAP, biosynthesis (Fig. 68A).⁶⁶⁶ The tabB (orf6) gene product⁶⁶⁷ shows sequence similarity to N-succinyl-transferase which catalyses N-succinyl-2-amino-6-oxopimelate formation in lysine biosynthesis (Fig. 68A), suggesting that TabB might be an acetyltransferase;⁶⁶⁷ acetylated intermediates as part of tabtoxin biosynthesis have already been proposed.⁶⁶⁸ The tblA (orf8) gene is required for tabtoxin formation, though its function is unclear.⁶⁶⁹ Regulation of tblA has been shown to be controlled by the lemA gene in P. syringae, probably at the transcriptional level.^670–672

A preliminary proposal for tabtoxin biosynthesis, consistent with the reported labeling studies, is shown in Fig. 68A. Key steps are β-lactam synthetase (TblS)-mediated β-lactam formation and 2OG oxygenase (TblC)-catalysed hydroxylation at C-3 of the β-lactam.

8 Summary, conclusions and future prospects

In this article, we have reviewed studies on the enzymology of BLA biosynthesis which have been conducted over the past half century or so. Given the compact nature of the β-lactam nucleus and the plethora of synthetic methods for β-lactam preparation, it is perhaps surprising that only two β-lactam biosynthesis strategies have been identified to date – the oxidative peptide cyclisation reaction of IPNS (in penicillin/cephalosporin biosynthesis) and the ATP-dependent β-amino acid cyclisation reaction of the β-lactam synthetases (in clavam/carbapenem biosynthesis). It seems most probable that nocardicin biosynthesis presents a third strategy involving a synthetase catalysed N-1/C-4 bond formation. Thus, although it seems that (some of) the pathways may have arisen from (a) common evolutionary origin(s), the available evidence suggests that certain types of enzymes may be particularly suited to β-lactam formation. In contrast to the limited ways for β-lactam formation, there are multiple ways by which structural diversity of BLAs is achieved.

8.1 Mechanisms of structural diversity introduction into β-lactam biosynthesis pathways

As highlighted previously,²⁹ once the β-lactam has been biosynthesised, diversity is achieved both by branching from common intermediates and by multiple sequential reactions (Table 9). 6-APA acts as a branch point for the biosynthesis of various penicillins with different C-6 acetamido side chains, as does deacetylcephalosporin C (DAC, or an analogue thereof) in the formation of C-3′ cephalosporins. A multiplicity of carbapenems is produced by modification of the C-2 and C-6 substituents, though the details are only emerging. Common themes in the modification of the β-lactam include methylation, acetylation, hydrolysis, addition of peptides by NRPS catalysis, and redox reactions (Table 9). The latter are often catalysed by 2OG oxygenases, though dehydrogenases and P450 enzymes also play roles. The use of epimerisation (isomerisation is the most efficient way of creating diversity from an atom economy perspective) is also notable as it occurs at least in the penicillin/cephalosporin, carbapenem, clavulanic acid and nocardicin pathways, and possibly others. For example, the inversion of the α-stereochemistry of L-valine during LLD-ACV formation and isopenicillin N/penicillin N side chain epimerisation are central to the formation of the penicillins and cephalosporins.

Table 9 Some of the enzyme families (non-systematic names) present in one or more of the biosynthetic pathways of β-lactams for which genome sequences have been reported. Enzyme families in blue are involved in providing the precursor for β-lactam formation (i.e. a β-amino acid in the case of clavams and carbapenems, and oligopeptides in the case of penicillins and, probably, nocardicins). Enzyme families in red are involved in β-lactam formation. Enzyme families in green are involved in diversity introduction. NRPSs: non-ribosomal peptide synthetases; ThDP: Thiamine diphosphate dependent; IPNS: isopenicillin N synthase; Ntn: N-terminal nucleophile; 2OG: 2-oxoglutarate; SAM: S-adenosylmethionine; GNAT: GCN5-related acetyltransferase

---

8.2 The future of β-lactam biosynthesis studies

Pioneering studies on BLA fermentation were vital by enabling commercially available production procedures. Subsequent studies on the molecular basis of BLA biosynthesis, which came after the “golden age” of BLAs, have been of (bio)chemical interest and have had impact far beyond the BLA biosynthesis, in fields including quorum sensing, oxygen sensing, DNA repair, and epigenetics. Studies on β-lactam biosynthesis have also been a testing ground for new techniques, particularly in molecular biology. However, the studies on BLA biosynthesis enzymes have not yet had a major impact on BLA production or on the development of new BLAs. This is probably a question of timing.

There remain chemically interesting questions in β-lactam biosynthesis; these include the mechanisms of the epimerisation in CA biosynthesis, the detailed mechanisms of CarC and DAOCS catalysis, and as to how the reactions of the NRPS machinery of ACVS are coordinated. The mechanisms by which the β-lactam of the monocyclic BLAs is formed are still unclear. No doubt genome sequencing of (new) β-lactam producers will also reveal other interesting enzymes/reactions yet to be considered. It would be rather interesting if new ways for Nature to make β-lactams were to be discovered.

Although there are many secondary metabolites of chemical interest, we (and a few others) like to think of BLAs as special and believe that there is still considerable scope for the development of the BLAs. Most, if not all, clinically used BLAs were developed prior to structural and functional insights into their PBP targets. Thus, there is interest in applying “modern” medicinal chemistry approaches to PBP inhibition. Whether this should be centred around β-lactams, or other chemical templates, including alternative acylating agents, is unclear. However, the track record of BLAs is exceptional, despite resistance issues, as for other antibiotics. If there were renewed interest in a new generation of BLAs, biosynthetic studies could be useful. The scarce discovery of new BLAs over the past 30 years coupled to the need to secure green and sustainable routes for the production of clinically useful ones implies that new strategies need to be pursued.⁶⁷³ One possibility is to make the fermentation of carbapenems, or carbapenem intermediates, viable with a view to avoiding reliance on the AOSA intermediate (Fig. 46). Another would be the development of methods for biocatalytic production of “hybrid” antibiotics, e.g. penams with carbapenem side chains and carbapenems with acetamido side chains, possibly employing engineered enzymes and “mixing and matching” of components from different pathways. The ACVS/IPNS coupled system is one unexplored opportunity, as is building on recent studies on the engineering of carbapenem biosynthesis. The ability of β-lactam biosynthesis enzymes to accept alternative substrates, to date best exemplified by the range of heterocycles produced by IPNS and CMPSs (Fig. 14 and 53, respectively), also presents them as candidates for combinatorial biosynthesis approaches. Other opportunities include, for example, exploiting metagenomic cloning,^674,675 biasing pathways via epigenetic control^676,677 and the activation of apparently silent gene clusters.⁶⁷⁸

9 Acknowledgements

We gratefully acknowledge the efforts of all the scientists and their supporters who have worked on the β-lactam biosynthesis story. We apologise for lack of comprehensive citations and any errors or omissions. We also thank the agencies who have funded our work, especially the Biotechnology and Biological Sciences Research Council and the Wellcome Trust. CD thanks the Deutsche Akademie für Naturforscher Leopoldina, Germany, for a postdoctoral fellowship.

10 References

1. R. P. Elander, Appl. Microbiol. Biotechnol., 2003, 61, 385–392.
2. F.-R. Schmidt, in Industrial Applications, ed. M. Hofrichter, SpringerBerlin Heidelberg, 2010, vol. 10, pp. 101–121.
3. D. M. Livermore, J. Antimicrob. Chemother., 2011, 66, 1941–1944.
4. C. T. Walsh and G. Wright, Chem. Rev., 2005, 105, 391–394.
5. I. Chopra, C. Schofield, M. Everett, A. O'Neill, K. Miller, M. Wilcox, J.-M. Frère, M. Dawson, L. Czaplewski, U. Urleb and P. Courvalin, Lancet Infect. Dis., 2008, 8, 133–139.
6. L. Freire-Moran, B. Aronsson, C. Manz, I. C. Gyssens, A. D. So, D. L. Monnet and O. Cars, Drug Resist. Updates, 2011, 14, 118–124.
7. A. R. M. Coates and Y. Hu, Br. J. Pharmacol., 2007, 152, 1147–1154.
8. D. J. Waxman and J. L. Strominger, Annu. Rev. Biochem., 1983, 52, 825–869.
9. E. Sauvage, F. Kerff, M. Terrak, J. A. Ayala and P. Charlier, FEMS Microbiol. Rev., 2008, 32, 234–258.
10. W. Vollmer, D. Blanot and M. de Pedro, FEMS Microbiol. Rev., 2008, 32, 149–167.
11. W. Vollmer and J.-V. Holtje, J. Bacteriol., 2004, 186, 5978–5987.
12. J. B. Doherty, B. M. Ashe, L. W. Argenbright, P. L. Barker, R. J. Bonney, G. O. Chandler, M. E. Dahlgren, C. P. Dorn, P. E. Finke, R. A. Firestone, D. Fletcher, W. K. Hagmann, R. Mumford, L. O'Grady, A. L. Maycock, J. M. Pisano, S. K. Shah, K. R. Thompson and M. Zimmerman, Nature, 1986, 322, 192–194.
13. J. Hamilton-Miller, J. Antimicrob. Chemother., 1999, 44, 729–734.
14. T. Sperka, J. Pitlik, P. Bagossi and J. Tözsér, Bioorg. Med. Chem. Lett., 2005, 15, 3086–3090.
15. C. Yoakim, W. W. Ogilvie, D. R. Cameron, C. Chabot, C. Grand-Maitre, I. Guse, B. Hache, S. Kawai, J. Naud, J. A. O'Meara, R. Plante and R. Deziel, Antiviral Chem. Chemother., 1998, 9, 379–387.
16. O. A. Pierrat, K. Strisovsky, Y. Christova, J. Large, K. Ansell, N. Bouloc, E. Smiljanic and M. Freeman, ACS Chem. Biol., 2010, 6, 325–335.
17. W. D. Vaccaro, R. Sher and H. R. Davis, Bioorg. Med. Chem. Lett., 1998, 8, 319–322.
18. J. Earl and P. Kirkpatrick, Nat. Rev. Drug Discovery, 2003, 2, 97–98.
19. M. Garcia-Calvo, J. M. Lisnock, H. G. Bull, B. E. Hawes, D. A. Burnett, M. P. Braun, J. H. Crona, H. R. Davis, D. C. Dean, P. A. Detmers, M. P. Graziano, M. Hughes, D. E. MacIntyre, A. Ogawa, K. A. O'Neill, S. P. N. Iyer, D. E. Shevell, M. M. Smith, Y. S. Tang, A. M. Makarewicz, F. Ujjainwalla, S. W. Altmann, K. T. Chapman and N. A. Thornberry, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 8132–8137.
20. B. K. Banik, F. F. Becker and I. Banik, Bioorg. Med. Chem., 2004, 12, 2523–2528.
21. B. K. Banik, I. Banik and F. F. Becker, Eur. J. Med. Chem., 2010, 45, 846–848.
22. P. Perez-Faginas, M. T. Aranda, M. T. Garcia-Lopez, A. Francesch, C. Cuevas and R. Gonzalez-Muniz, Eur. J. Med. Chem., 2011, 46, 5108–5119.
23. J. Rothstein, M. Dykes-Hoberg, C. Pardo, L. Bristol, L. Jin, R. Kuncl, Y. Kanai, M. Hediger, Y. Wang and J. Schielke, Neuron, 1996, 16, 675–686.
24. I. J. Clifton, M. A. McDonough, D. Ehrismann, N. J. Kershaw, N. Granatino and C. J. Schofield, J. Inorg. Biochem., 2006, 100, 644–669.
25. M. A. McDonough, C. Loenarz, R. Chowdhury, I. J. Clifton and C. J. Schofield, Curr. Opin. Struct. Biol., 2010, 20, 659–672.
26. C. Loenarz and C. J. Schofield, Trends Biochem. Sci., 2011, 36, 7–18.
27. R. Chowdhury, A. Hardy and C. J. Schofield, Chem. Soc. Rev., 2008, 37, 1308–1319.
28. K. S. Hewitson, N. Granatino, R. W. D. Welford, M. A. McDonough and C. J. Schofield, Philos. Trans. R. Soc. London, Ser. A, 2005, 363, 807–828.
29. N. J. Kershaw, M. E. C. Caines, M. C. Sleeman and C. J. Schofield, Chem. Commun., 2005, 4251–4263.
30. P. Liras and A. L. Demain, in Methods Enzymol., ed. A. H. David, Academic Press, 2009, vol. 458, pp. 401–429.
31. C. J. Schofield, J. E. Baldwin, M. F. Byford, I. Clifton, J. Hajdu, C. Hensgens and P. Roach, Curr. Opin. Struct. Biol., 1997, 7, 857–864.
32. K. H. Baggaley, A. G. Brown and C. J. Schofield, Nat. Prod. Rep., 1997, 14, 309–333.
33. C. A. Townsend, Curr. Opin. Chem. Biol., 2002, 6, 583–589.
34. C. A. Townsend, Chem. Biol., 1997, 4, 721–730.
35. S. Harada, S. Tsubotani, T. Hida, K. Koyana, M. Kondo and H. Ono, Tetrahedron, 1988, 44, 6589–6606.
36. S. Harada, S. Tsubotani, T. Hida, H. Ono and H. Okazaki, Tetrahedron Lett., 1986, 27, 6229–6232.
37. Y. Nozaki, N. Katayama, H. Ono, S. Tsubotani, S. Harada, H. Okazaki and Y. Nakao, Nature, 1987, 325, 179–180.
38. D. J. Tipper and J. L. Strominger, Proc. Natl. Acad. Sci. U. S. A., 1965, 54, 1133–1141.
39. N. J. Westwood, T. D. W. Claridge, P. N. Edwards and C. J. Schofield, Bioorg. Med. Chem. Lett., 1997, 7, 2973–2978.
40. R. C. Wilmouth, N. J. Westwood, K. Anderson, W. Brownlee, T. D. W. Claridge, I. J. Clifton, G. J. Pritchard, R. T. Aplin and C. J. Schofield, Biochemistry, 1998, 37, 17506–17513.
41. P. Macheboeuf, D. S. Fischer, T. Brown, A. Zervosen, A. Luxen, B. Joris, A. Dessen and C. J. Schofield, Nat. Chem. Biol., 2007, 3, 565–569.
42. T. Brown, P. Charlier, R. l. Herman, C. J. Schofield and E. Sauvage, J. Med. Chem., 2010, 53, 5890–5894.
43. W. M. M. Kirby, Science, 1944, 99, 452–453.
44. K. Poole, Cell. Mol. Life Sci., 2004, 61, 2200–2223.
45. B. G. Spratt, Science, 1994, 264, 388–393.
46. K. Poole, Ann. Med., 2007, 39, 162–176.
47. A. H. Delcour, Biochim. Biophys. Acta, Proteins Proteomics, 2009, 1794, 808–816.
48. A. Matagne, A. Dubus, M. Galleni and J. M. Frere, Nat. Prod. Rep., 1999, 16, 1–19.
49. S. Bratu, S. Brooks, S. Burney, S. Kochar, J. Gupta, D. Landman and J. Quale, Clin. Infect. Dis., 2007, 44, 972–975.
50. J.-M. Frere, Beta-Lactamases, Nova Science Pub Inc, New York, US, 2012.
51. I. Massova and S. Mobashery, Curr. Pharm. Des., 1999, 5, 929–937.
52. Antimicrobial Drug Resistance: Mechanisms of Drug Resistance, ed. D. L. Mayers, Human Press, 2009.
53. S. O. Meroueh, G. Minasov, W. Lee, B. K. Shoichet and S. Mobashery, J. Am. Chem. Soc., 2003, 125, 9612–9618.
54. T. R. Walsh, M. A. Toleman, L. Poirel and P. Nordmann, Clin. Microbiol. Rev., 2005, 18, 306–325.
55. D. L. Paterson and R. A. Bonomo, Clin. Microbiol. Rev., 2005, 18, 657–686.
56. S. M. Drawz, M. Babic, C. R. Bethel, M. Taracila, A. M. Distler, C. Ori, E. Caselli, F. Prati and R. A. Bonomo, Biochemistry, 2010, 49, 329–340.
57. M. G. P. Page, Drug Resist. Updates, 2000, 3, 109–125.
58. S. M. Drawz and R. A. Bonomo, Clin. Microbiol. Rev., 2010, 23, 160–201.
59. C. E. Jamieson, P. A. Lambert and I. N. Simpson, Antimicrob. Agents Chemother., 2003, 47, 1652–1657.
60. A. Endimiani, K. M. Hujer, A. M. Hujer, M. E. Pulse, W. J. Weiss and R. A. Bonomo, Antimicrob. Agents Chemother., 2011, 55, 82–85.
61. T. Stachyra, M.-C. Pechereau, J.-M. Bruneau, M. Claudon, J.-M. Frere, C. Miossec, K. Coleman and M. T. Black, Antimicrob. Agents Chemother., 2010, 54, 5132–5138.
62. T. Stachyra, P. Levasseur, M.-C. Péchereau, A.-M. Girard, M. Claudon, C. Miossec and M. T. Black, J. Antimicrob. Chemother., 2009, 64, 326–329.
63. O. K. Behrens, J. Corse, J. P. Edwards, L. Garrison, R. G. Jones, Q. F. Soper, F. R. Van Abeele and C. W. Whitehead, J. Biol. Chem., 1948, 175, 793–809.
64. J. Nayler, Proc. R. Soc. London, Ser. B, 1971, 179, 357–367.
65. O. K. Behrens, J. Corse, J. P. Edwards, L. Garrison, R. G. Jones, Q. F. Soper, F. R. Van Abeele and C. W. Whitehead, J. Biol. Chem., 1948, 175, 771–792.
66. F. R. Batchelor, F. P. Doyle, J. H. C. Nayler and G. N. Rolinson, Nature, 1959, 183, 257–258.
67. A. K. Chandel, L. V. Rao, M. L. Narasu and O. V. Singh, Enzyme Microb. Technol., 2008, 42, 199–207.
68. G. G. F. Newton and E. P. Abraham, Nature, 1955, 175, 548–548.
69. R. B. Morin, B. G. Jackson, R. A. Mueller, E. R. Lavagnino, W. B. Scanlon and S. L. Andrews, J. Am. Chem. Soc., 1969, 91, 1401–1407.
70. V. Sonawane, Crit. Rev. Biotechnol., 2006, 26, 95–120.
71. J. Berdy, in Bioactive Metabolites from Microorganisms, ed. M. E. Bushell and U. Grafe, Elsevier, Amsterdam, 1989, vol. 27, pp. 3–25.
72. M. S. Barber, U. Giesecke, A. Reichert and W. Minas, Adv. Biochem. Eng. Biotechnol., 2004, 88, 179–215.
73. Ö. D. Ekici, M. Paetzel and R. E. Dalbey, Protein Sci., 2008, 17, 2023–2037.
74. C. Oinonen and J. Rouvinen, Protein Sci., 2000, 9, 2329–2337.
75. R. B. Sykes and J. S. Wells, J. Antibiot., 1985, 38, 119–121.
76. K. Kitano, K. Kintaka, S. Suzuki, K. Katamoto, K. Nara and Y. Nakao, J. Ferment. Technol., 1975, 29, 668–669.
77. A. Brown, D. Butterworth, M. Cole, G. Hanscomb, J. Hood, C. Reading and G. Rolinson, J. Antibiot., 1976, 29, 668–669.
78. H. Aoki, H. Sakai, M. Kohsaka, T. Konomi and J. Hosoda, J. Antibiot., 1976, 29, 492–500.
79. R. B. Sykes and D. P. Bonner, Am. J. Med., 1985, 78, 2–10.
80. S. Coulton and E. Hunt, in Prog. Med. Chem., ed. G. P. Ellis and D. K. Luscombe, Elsevier, 1996, vol. 33, pp. 99–145.
81. J. S. Kahan, F. M. Kahan, R. Goegelman, S. A. Currie, M. Jackson, E. O. Stapley, T. W. Miller, A. K. Miller, D. Hendlin, S. Mochales, S. Hernandez, H. B. Woodruff and J. Birnbaum, J. Antibiot., 1979, 32, 1–12.
82. G. Albers-Schönberg, B. H. Arison, O. D. Hensens, J. Hirshfield, K. Hoogsteen, E. A. Kaczka, R. E. Rhodes, J. S. Kahan, F. M. Kahan, R. W. Ratcliffe, E. Walton, L. J. Ruswinkle, R. B. Morin and B. G. Christensen, J. Am. Chem. Soc., 1978, 100, 6491–6499.
83. A. L. Demain and R. P. Elander, Antonie van Leeuwenhoek, 1999, 75, 5–19.
84. M. Sunagawa and A. Sasaki, Heterocycles, 2001, 54, 497–528.
85. G. Brooks, G. Bruton, M. F. Finn, J. B. Harbridge, M. A. Harris, T. T. Howarth, E. Hunt, I. Stirling and I. I. Zomaya, in Recent Advances in the Chemistry of β-Lactam Antibiotics, ed. A. G. Brown and S. M. Roberts, Royal Society of Chemistry, London, 1985, p. 221.
86. J. S. Davies, Tetrahedron Lett., 1982, 23, 5089–5092.
87. C. E. Newall, in Recent Advances in the Chemistry of β-Lactam Antibiotics, ed. G. I. Gregory, Royal Society of Chemistry, London, 1980, p. 151.
88. P. C. Cherry and E. Newall, in Chemistry and Biology of β-Lactam Antibiotics, ed. R. B. Morin and M. Gorman, Academic Press, New York, 1982, p. 361.
89. C. Reading and M. Cole, J. Enzyme Inhib. Med. Chem., 1986, 1, 83–104.
90. A. Sasaki and M. Sunagawa, Chem. Heterocycl. Compd., 1998, 34, 1249–1265.
91. D. H. Shih, F. Baker, L. Cama and B. G. Christensen, Heterocycles, 1984, 21, 29–40.
92. A. H. Berks, Tetrahedron, 1996, 52, 331–375.
93. J. Shoji, T. Kato, R. Sakazaki, W. Nagata, Y. Terui, Y. Nakagawa, M. Shiro, K. Matsumoto and T. Hattori, J. Antibiot., 1984, 37, 1486–1490.
94. J. Shoji, R. Sakazaki, T. Kato, Y. Terui, K. Matsumoto, T. Tanimoto, T. Hattori, K. Hirooka and E. Kondo, J. Antibiot., 1985, 38, 538–540.
95. P. D. Singh, M. G. Young, J. H. Johnson, C. M. Cimarusti and R. B. Sykes, J. Antibiot., 1984, 37, 773–780.
96. S. Tsubotani, T. Hida, H. Ono and S. Harada, J. Antibiot., 1985, 38, 1152–1165.
97. Y. Nozaki, N. Katayama, S. Tsubotani, H. Ono and H. Okazaki, J. Antibiot., 1985, 38, 1141–1151.
98. S. Tsubotani, T. Hida, F. Kasahara, Y. Wada and S. Harada, J. Antibiot., 1984, 37, 1546–1554.
99. S. Harada, S. Tsubotani, H. Ono and H. Okazaki, J. Antibiot., 1984, 37, 1536–1545.
100. T. R. M. Barends, H. Yoshida and B. W. Dijkstra, Curr. Opin. Biotechnol., 2004, 15, 356–363.
101. A. A. Brakhage, Q. Al-Abdallah, A. Tüncher and P. Spröte, Phytochemistry, 2005, 66, 1200–1210.
102. S. S. Weber, R. A. L. Bovenberg and A. J. M. Driessen, Biotechnol. J., 2012, 7, 225–236.
103. R. D. G. Cooper, Bioorg. Med. Chem., 1993, 1, 1–17.
104. T. R. van der Lende, M. van de Kamp, M. v. d. Berg, K. Sjollema, R. A. L. Bovenberg, M. Veenhuis, W. N. Konings and A. J. M. Driessen, Fungal Genet. Biol., 2002, 37, 49–55.
105. M. F. Byford, J. E. Baldwin, C. Y. Shiau and C. J. Schofield, Chem. Rev., 1997, 97, 2631–2650.
106. P. L. Roach, I. J. Clifton, V. Fulop, K. Harlos, G. J. Barton, J. Hajdu, I. Andersson, C. J. Schofield and J. E. Baldwin, Nature, 1995, 375, 700–704.
107. P. Liras, Antonie van Leeuwenhoek, 1999, 75, 109–124.
108. J. J. R. Coque, F. J. Perez-Llarena, F. J. Enguita, J. L. Fuente, J. F. Martin and P. Liras, Gene, 1995, 162, 21–27.
109. L. M. Öster, D. R. Lester, A. Terwisscha van Scheltinga, M. Svenda, M. van Lun, C. Généreux and I. Andersson, J. Mol. Biol., 2006, 358, 546–558.
110. S. Inouye, M. Kojima, T. Shomura, K. Iwamatsu, T. Niwa, Y. Kondo, T. Niida, Y. Ogawa and K. Kusama, J. Antibiot., 1983, 36, 115–124.
111. T. Kanzaki, T. Fukita, H. Shirafuji, Y. Fujisawa and K. Kitano, J. Antibiot., 1974, 27, 361–362.
112. T. Kanzaki, T. Fukita, K. Kitano, K. Katamoto, K. Nara and Y. Nakao, J. Ferment. Technol., 1976, 54, 720–725.
113. T. Osono, S. Watanabe, T. Saito, H. Gushima, K. Murakami, I. Takahashi, H. Yamaguchi, T. Sasaki, K. Susaki, S. Takamura, T. Miyoshi and Y. Oka, J. Antibiot., 1980, 33, 1074–1078.
114. A. V. Demirev, C. H. Lee, B. P. Jaishy, D. H. Nam and D. D. Y. Ryu, FEMS Microbiol. Lett., 2006, 255, 121–128.
115. Y. S. Sohn, D. H. Nam and D. D. Y. Ryu, Metab. Eng., 2001, 3, 380–392.
116. S. Goulden and F. Chattaway, Biochem. J., 1968, 110, 55P–56P.
117. T. M. Zabriskie and M. D. Jackson, Nat. Prod. Rep., 2000, 17, 85–97.
118. H. Xu, B. Andi, J. Qian, A. West and P. Cook, Cell Biochem. Biophys., 2006, 46, 43–64.
119. E. Valmaseda, S. Campoy, L. Naranjo, J. Casqueiro and J. Martín, Mol. Genet. Genomics, 2005, 274, 272–282.
120. A. L. Demain, Arch. Biochem. Biophys., 1957, 67, 244–246.
121. J. J. R. Coque, P. Liras, L. Laiz and J. F. Martin, J. Bacteriol., 1991, 173, 6258–6264.
122. K. Madduri, C. Stuttard and L. C. Vining, J. Bacteriol., 1991, 173, 985–988.
123. G. Scapin and J. S. Blanchard, Adv. Enzymol. Relat. Areas Mol. Biol., 1998, 72, 279–324.
124. B. A. Kern, D. Hendlin and E. Inamine, Antimicrob. Agents Chemother., 1980, 17, 679–685.
125. F. J. Perez-Llarena, A. Rodriguez-Garcia, F. J. Enguita, J. F. Martin and P. Liras, J. Bacteriol., 1998, 180, 4753–4756.
126. A. L. Leitão, F. J. Enguita, J. F. Martín and J. F. Santos Oliveira, Appl. Microbiol. Biotechnol., 2001, 56, 670–675.
127. V. K. Chary, J. L. De La Fuente, A. L. Leitão, P. Liras and J. F. Martín, Appl. Microbiol. Biotechnol., 2000, 53, 282–288.
128. M.-L. Wu, J.-H. Chen, C.-T. Ho and T.-C. Huang, J. Agric. Food Chem., 2007, 55, 1767–1772.
129. J. L. De la Fuente, A. Rumbero, J. F. Martin and P. Liras, Biochem. J., 1997, 327, 59–64.
130. H. Yu, E. Serpe, J. Romero, J.-J. Coque, K. Maeda, M. Oelgeschläger, G. Hintermann, P. Liras, J.-F. Martín, A. L. Demain and J. Piret, Microbiology, 1994, 140, 3367–3377.
131. D. C. Alexander, C. L. Anders, L. Lee and S. E. Jensen, J. Bacteriol., 2007, 189, 5867–5874.
132. D. C. Alexander, M. J. Brumlik, L. Lee and S. E. Jensen, J. Bacteriol., 2000, 182, 348–356.
133. T. Schwecke, Y. Aharonowitz, H. Palissa, H. von Dohren, H. Kleinkauf and H. van Liempt, Eur. J. Biochem., 1992, 205, 687–694.
134. H. B. A. Theilgaard, K. N. Kristiansen, C. M. Henriksen and J. Nielsen, Biochem. J., 1997, 327, 185–191.
135. Y. Aharonowitz, J. Bergmeyer, J. M. Cantoral, G. Cohen, A. L. Demain, U. Fink, J. Kinghorn, H. Kleinkauf, A. MacCabe, H. Palissa, E. Pfeifer, T. Schwecke, H. Van Liempt, H. von Dohren, S. Wolfe and J. Zhang, Bio/Technology, 1993, 11, 807–810.
136. J. Zhang and A. L. Demain, Crit. Rev. Biotechnol., 1992, 12, 245–260.
137. H. von Döhren, in Molecular Biotechnolgy of Fungal beta-Lactam Antibiotics and Related Peptide Synthetases, ed. A. A. Brakhage, Springer Berlin/Heidelberg, 2004, vol. 88, pp. 217–264.
138. U. Linne and M. A. Marahiel, Methods Enzymol., 2004, 388, 293–315.
139. J. Grünewald and M. A. Marahiel, Microbiol. Mol. Biol. Rev., 2006, 70, 121–146.
140. M. A. Marahiel and L. O. Essen, Methods Enzymol., 2009, 458, 337–351.
141. T. J. Montavon and S. D. Bruner, in Comprehensive Natural Products II, ed. M. Lew and L. Hung-Wen, Elsevier, Oxford, 2010, pp. 619–655.
142. S. A. Sieber and M. A. Marahiel, Chem. Rev., 2005, 105, 715–738.
143. J. E. Baldwin and C. J. Schofield, in The Chemistry of β-lactams, ed. M. I. Page, Blackie Academic & Professional, Glasgow, 1992, pp. 1–78.
144. J. Kennedy and G. Turner, Mol. Gen. Genet., 1996, 253, 189–197.
145. G. Banko, S. Wolfe and A. L. Demain, Biochem. Biophys. Res. Commun., 1986, 137, 528–535.
146. J. F. Martín, T. D. Ingolia and S. W. Queener, Molecular Industrial Mycology, 1990, pp. 149–195.
147. J. Martin, T. Ingolia and S. Queener, in Molecular Industrial Mycology: Systems and Applications for Filamentous Fungi, ed. S. Leong and R. Berka, Marcel Dekker, New York, 1991, pp. 149–196.
148. Y. Aharonowitz, G. Cohen and J. F. Martin, Annu. Rev. Microbiol., 1992, 46, 461–495.
149. B. Díez, J. Barredo, E. Alvarez, J. Cantoral, P. Solingen, M. Groenen, A. Veenstra and J. Martin, MGG, Mol. Gen. Genet., 1989, 218, 572–576.
150. M. A. Marahiel, T. Stachelhaus and H. D. Mootz, Chem. Rev., 1997, 97, 2651–2674.
151. J. F. Martin, J. Antibiot., 2000, 53, 1008–1021.
152. H. Kleinkauf and H. von Döhren, Eur. J. Biochem., 1996, 236, 335–351.
153. W. Kallow, J. Kennedy, B. Arezi, G. Turner and H. von Döhren, J. Mol. Biol., 2000, 297, 395–408.
154. J. Zhang, S. Wolfe and A. L. Demain, Biochem. J., 1992, 283, 691–698.
155. D. Keszenman-Pereyra, S. Lawrence, M.-E. Twfieg, J. Price and G. Turner, Curr. Genet., 2003, 43, 186–190.
156. J. E. Baldwin, J. W. Bird, R. A. Field, N. M. O'Callaghan, C. J. Schofield and A. C. Willis, J. Antibiot., 1991, 44, 241–248.
157. J. E. Baldwin, J. W. Bird, R. A. Field, N. M. O'Callaghan and C. J. Schofield, J. Antibiot., 1990, 43, 1055–1057.
158. J. E. Baldwin, C. Y. Shiau, M. F. Byford and C. J. Schofield, Biochem. J., 1994, 301, 367–372.
159. C.-Y. Shiau, J. E. Baldwin, M. F. Byford, W. J. Sobey and C. J. Schofield, FEBS Lett., 1995, 358, 97–100.
160. C. Y. Shiau, M. F. Byford, R. T. Aplin, J. E. Baldwin and C. J. Schofield, Biochemistry, 1997, 36, 8798–8806.
161. L. Luo and C. T. Walsh, Biochemistry, 2001, 40, 5329–5337.
162. A. Stindl and U. Keller, Biochemistry, 1994, 33, 9358–9364.
163. X. Wu, C. García-Estrada, I. Vaca and J.-F. Martín, Biochimie, 2012, 94, 354–418.
164. G. Banko, A. L. Demain and S. Wolfe, J. Am. Chem. Soc., 1987, 109, 2858–2860.
165. S. E. Jensen, D. W. S. Westlake and S. Wolfe, FEMS Microbiol. Lett., 1988, 49, 213–218.
166. J. J. R. Coque, J. L. De La Fuente, P. Liras and J. F. Martín, Eur. J. Biochem., 1996, 242, 264–270.
167. A. Tanovic, S. A. Samel, L. O. Essen and M. A. Marahiel, Science, 2008, 321, 659–663.
168. M. Strieker, A. Tanovic and M. A. Marahiel, Curr. Opin. Struct. Biol., 2010, 20, 234–240.
169. A. Koglin and C. T. Walsh, Nat. Prod. Rep., 2009, 26, 987–1000.
170. S. A. Samel, G. Schoenafinger, T. A. Knappe, M. A. Marahiel and L. O. Essen, Structure, 2007, 15, 781–792.
171. D. P. Frueh, H. Arthanari, A. Koglin, D. A. Vosburg, A. E. Bennett, C. T. Walsh and G. Wagner, Nature, 2008, 454, 903–906.
172. C. Pang, B. Chakravarti, R. Adlington, H. Ting, R. White, G. Jayatilake, J. Baldwin and E. Abraham, Biochem. J., 1984, 222, 789–795.
173. G. Cohen, A. Argaman, R. Schreiber, M. Mislovati and Y. Aharonowitz, J. Bacteriol., 1994, 176, 973–984.
174. B. K. Leskiw, Y. Aharonowitz, M. Mevarech, S. Wolfe, L. C. Vining, D. W. S. Westlake and S. E. Jensen, Gene, 1988, 62, 187–196.
175. O. Landman, D. Shiffman, Y. Av-Gay, Y. Aharoniwitz and G. Cohen, FEMS Microbiol. Lett., 1991, 84, 239–244.
176. J. J. R. Coque, J. F. Martin, J. G. Calzada and P. Liras, Mol. Microbiol., 1991, 5, 1125–1133.
177. M. Garcia-Dominguez, P. Liras and J. F. Martin, Antimicrob. Agents Chemother., 1991, 35, 44–52.
178. P. Loke, C. P. Ng and T. S. Sim, Can. J. Microbiol., 2000, 46, 166–170.
179. S. E. Jensen, B. K. Leskiw, L. C. Vining, Y. Aharonowitz, D. W. Westlake and S. Wolfe, Can. J. Microbiol., 1986, 32, 953–958.
180. D. Perry, E. P. Abraham and J. E. Baldwin, Biochem. J., 1988, 255, 345–351.
181. J. M. Castro, P. Liras, L. Laiz, J. Cortes and J. F. Martin, J. Gen. Microbiol., 1988, 134, 133–141.
182. P. L. Roach, I. J. Clifton, C. M. Hensgens, N. Shibata, A. J. Long, R. W. Strange, S. S. Hasnain, C. J. Schofield, J. E. Baldwin and J. Hajdu, Eur. J. Biochem., 1996, 242, 736–740.
183. C. J. Schofield and Z. H. Zhang, Curr. Opin. Struct. Biol., 1999, 9, 722–731.
184. I. Borovok, O. Landman, R. Kreisberg-Zakarin, Y. Aharonowitz and G. Cohen, Biochemistry, 1996, 35, 1981–1987.
185. R. Kreisberg-Zakarin, I. Borovok, M. Yanko, F. Frolow, Y. Aharonowitz and G. Cohen, Biophys. Chem., 2000, 86, 109–118.
186. T. S. Sim and P. Loke, Appl. Microbiol. Biotechnol., 2000, 54, 1–8.
187. A. G. Prescott, J. Exp. Bot., 1993, 44, 849–861.
188. M. Sami, T. J. N. Brown, P. L. Roach, C. J. Schofield and J. E. Baldwin, FEBS Lett., 1997, 405, 191–194.
189. P. L. Roach, I. J. Clifton, C. M. Hensgens, N. Shibata, C. J. Schofield, J. Hajdu and J. E. Baldwin, Nature, 1997, 387, 827–830.
190. M. Mantri, Z. Zhang, M. A. McDonough and C. J. Schofield, FEBS J., 2012, 279, 1563–1575.
191. J. E. Baldwin and M. Bradley, Chem. Rev., 1990, 90, 1079–1088.
192. V. J. Chen, A. M. Orville, M. R. Harpel, C. A. Frolik, K. K. Surerus, E. Munck and J. D. Lipscomb, J. Biol. Chem., 1989, 264, 21677–21681.
193. R. A. Scott, S. Wang, M. K. Eidsness, A. Kriauciunas, C. A. Frolik and V. J. Chen, Biochemistry, 1992, 31, 4596–4601.
194. C. R. Randall, Y. Zang, A. E. True, L. Que, Jr., J. M. Charnock, C. D. Garner, Y. Fujishima, C. J. Schofield and J. E. Baldwin, Biochemistry, 1993, 32, 6664–6673.
195. M. Lundberg, P. E. M. Siegbahn and K. Morokuma, Biochemistry, 2008, 47, 1031–1042.
196. N. I. Burzlaff, P. J. Rutledge, I. J. Clifton, C. M. Hensgens, M. Pickford, R. M. Adlington, P. L. Roach and J. E. Baldwin, Nature, 1999, 401, 721–724.
197. J. M. Elkins, P. J. Rutledge, N. I. Burzlaff, I. J. Clifton, R. M. Adlington, P. L. Roach and J. E. Baldwin, Org. Biomol. Chem., 2003, 1, 1455–1460.
198. A. J. Long, I. J. Clifton, P. L. Roach, J. E. Baldwin, C. J. Schofield and P. J. Rutledge, Biochem. J., 2003, 372, 687–693.
199. A. J. Long, I. J. Clifton, P. L. Roach, J. E. Baldwin, P. J. Rutledge and C. J. Schofield, Biochemistry, 2005, 44, 6619–6628.
200. A. Daruzzaman, I. J. Clifton, R. M. Adlington, J. E. Baldwin and P. J. Rutledge, ChemBioChem, 2006, 7, 351–358.
201. E. I. Solomon, T. C. Brunold, M. I. Davis, J. N. Kemsley, S. K. Lee, N. Lehnert, F. Neese, A. J. Skulan, Y. S. Yang and J. Zhou, Chem. Rev., 2000, 100, 235–349.
202. C. D. Brown, M. L. Neidig, M. B. Neibergall, J. D. Lipscomb and E. I. Solomon, J. Am. Chem. Soc., 2007, 129, 7427–7438.
203. J. E. Baldwin, R. M. Adlington and S. E. Moroney, J. Chem. Soc., Chem. Commun., 1984,(15), 984–986.
204. J. E. Baldwin, M. Bradley, R. M. Adlington, W. J. Norris and N. J. Turner, Tetrahedron, 1991, 47, 457–480.
205. M. Costas, M. P. Mehn, M. P. Jensen and L. Que Jr, Chem. Rev., 2004, 104, 939–986.
206. J. E. Baldwin, E. P. Abraham, R. M. Adlington, J. A. Murphy, N. B. Green, H. H. Ting and J. J. Usher, J. Chem. Soc., Chem. Commun., 1983, 1319–1320.
207. L. Que Jr and R. Y. N. Ho, Chem. Rev., 1996, 96, 2607–2624.
208. J. E. Baldwin, R. M. Adlington, D. G. Marquess, A. R. Pitt, M. J. Porter and A. T. Russell, Tetrahedron, 1996, 52, 2515–2536.
209. J. E. Baldwin, R. M. Adlington, B. P. Domaynehayman, H. H. Ting and N. J. Turner, J. Chem. Soc., Chem. Commun., 1986, 110–113.
210. J. M. Blackburn, J. D. Sutherland and J. E. Baldwin, Biochemistry, 1995, 34, 7548–7562.
211. J. E. Baldwin and E. Abraham, Nat. Prod. Rep., 1988, 5, 129–145.
212. J. E. Baldwin, M. Bradley, S. D. Abbott and R. M. Adlington, Tetrahedron, 1991, 47, 5309–5328.
213. A. R. Howard-Jones, J. M. Elkins, I. J. Clifton, P. L. Roach, R. M. Adlington, J. E. Baldwin and P. J. Rutledge, Biochemistry, 2007, 46, 4755–4762.
214. P. J. Rutledge, N. I. Burzlaff, J. M. Elkins, M. Pickford, J. E. Baldwin and P. L. Roach, Anal. Biochem., 2002, 308, 265–268.
215. A. L. Demain, Y. Q. Shen and S. E. Jensen, J. Antibiot., 1986, 39, 1007–1010.
216. J. M. Castro, P. Liras, J. Cortes and J. F. Martin, FEMS Microbiol. Lett., 1986, 34, 349–353.
217. J. M. Luengo, M. T. Alemany and F. Salto, Bio/Technology, 1986, 4, 44–47.
218. J. E. Baldwin, R. M. Adlington, N. Moss and N. G. Robinson, J. Chem. Soc., Chem. Commun., 1987, 1664–1667.
219. J. E. Baldwin, J. M. Blackburn, M. Sako and C. J. Schofield, J. Chem. Soc., Chem. Commun., 1989, 970–972.
220. J. E. Baldwin, G. P. Lynch and C. J. Schofield, Tetrahedron, 1992, 48, 9085–9100.
221. J. E. Baldwin, R. M. Adlington, B. P. Domayne-Hayman, G. Knight and H.-H. Ting, J. Chem. Soc., Chem. Commun., 1987, 1661–1663.
222. J. E. Baldwin, W. J. Norris, R. T. Freeman, M. Bradley, R. M. Adlington, S. Long-Fox and C. J. Schofield, J. Chem. Soc., Chem. Commun., 1988, 1128–1130.
223. J. E. Baldwin, R. M. Adlington, A. E. Derome, H. H. Ting and N. J. Turner, J. Chem. Soc., Chem. Commun., 1984, 1211–1214.
224. J. M. Ogle, I. J. Clifton, P. J. Rutledge, J. M. Elkins, N. I. Burzlaff, R. M. Adlington, P. L. Roach and J. E. Baldwin, Chem. Biol., 2001, 8, 1231–1237.
225. E. Alvarez, J. M. Cantoral, J. L. Barredo, B. Diez and J. F. Martin, Antimicrob. Agents Chemother., 1987, 31, 1675–1682.
226. J. L. Barredo, P. van Solingen, B. Diez, E. Alvarez, J. M. Cantoral, A. Kattevilder, E. B. Smaal, M. A. Groenen, A. E. Veenstra and J. F. Martin, Gene, 1989, 83, 291–300.
227. M. B. Tobin, M. D. Fleming, P. L. Skatrud and J. R. Miller, J. Bacteriol., 1990, 172, 5908–5914.
228. P. A. Whiteman, E. P. Abraham, J. E. Baldwin, M. D. Fleming, C. J. Schofield, J. D. Sutherland and A. C. Willis, FEBS Lett., 1990, 262, 342–344.
229. F.-Q. Wang, J. Liu, M. Dai, Z.-H. Ren, C.-Y. Su and J.-G. He, Biochem. Biophys. Res. Commun., 2007, 360, 453–458.
230. Z.-L. Yu, J. Liu, F.-Q. Wang, M. Dai, B.-H. Zhao, J.-G. He and H. Zhang, Folia Microbiol., 2011, 56, 246–252.
231. M. Lamas-Maceiras, I. Vaca, E. Rodriguez, J. Casqueiro and J. F. Martin, Biochem. J., 2006, 395, 147–155.
232. M. A. Ferrero, A. Reglero, J. Martin-Villacorta, J. M. Fernandez-Canon and J. M. Luengo, J. Antibiot., 1990, 43, 684–691.
233. P. A. Whiteman and E. P. Abraham, FEBS Lett., 1996, 394, 31–33.
234. E. Alvarez, B. Meesschaert, E. Montenegro, S. Gutierez, B. Diez, J. L. Barredo and J. F. Martin, Eur. J. Biochem., 1993, 215, 323–332.
235. J. M. Luengo, J. Antibiot., 1995, 48, 1195–1212.
236. J. M. Luengo, J. L. Iriso and M. J. Lopez-Nieto, J. Antibiot., 1986, 39, 1754–1759.
237. H. Martinez-Blanco, A. Reglero and J. M. Luengo, J. Antibiot., 1991, 44, 1252–1258.
238. N. J. Kershaw, H. J. McNaughton, K. S. Hewitson, H. Hernández, J. Griffin, C. Hughes, P. Greaves, B. Barton, C. V. Robinson and C. J. Schofield, Eur. J. Biochem., 2002, 269, 2052–2059.
239. M. F. Freeman, K. A. Moshos, M. J. Bodner, R. Li and C. A. Townsend, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 11128–11133.
240. R. T. Aplin, J. E. Baldwin, S. C. J. Cole, J. D. Sutherland and M. B. Tobin, FEBS Lett., 1993, 319, 166–170.
241. M. Bokhove, H. Yoshida, C. M. H. Hensgens, J. Metske van der Laan, J. D. Sutherland and B. W. Dijkstra, Structure, 2010, 18, 301–308.
242. Y. Li, J. Chen, W. Jiang, X. Mao, G. Zhao and E. Wang, Eur. J. Biochem., 1999, 262, 713–719.
243. M. A. Arroyo, I. d. l. M. de la Mata, C. A. Acebal and M. P. C. Pilar Castillón, Appl. Microbiol. Biotechnol., 2003, 60, 507–514.
244. Y. Kim, K.-H. Yoon, Y. Khang, S. Turley and W. G. J. Hol, Structure, 2000, 8, 1059–1068.
245. V. K. Sudhakaran, B. S. Deshpande, S. S. Ambedkar and J. G. Shewale, Process Biochem., 1992, 27, 131–143.
246. N. Esaki and C. T. Walsh, Biochemistry, 1986, 25, 3261–3267.
247. A. Watababe, Y. Kurokawa, T. Yoshimura, T. Kurihara, K. Soda and N. Esaki, J. Biol. Chem., 1999, 274, 4189–4194.
248. S. E. Jensen, D. W. S. Westlake and S. Wolfe, Can. J. Microbiol., 1983, 29, 1526–1531.
249. F. J. Enguita, J. J. R. Coque, P. Liras and J. F. Martin, J. Bacteriol., 1998, 180, 5489–5494.
250. J. J. Coque, J. F. Martin and P. Liras, MGG, Mol. Gen. Genet., 1993, 236, 453–458.
251. S. Kovacevic, M. B. Tobin and J. R. Miller, J. Bacteriol., 1990, 172, 3952–3958.
252. R. J. Bowers, S. E. Jensen, L. Lyubechansky, D. W. S. Westlake and S. Wolfe, Biochem. Biophys. Res. Commun., 1984, 120, 607–613.
253. R. Ullan, J. Casqueiro, O. Banuelos, F. Fernandez, S. Gutierrez and J. Martin, J. Biol. Chem., 2002, 277, 46216–46225.
254. J. F. Martín, R. V. Ullán and J. Casqueiro, in Molecular Biotechnolgy of Fungal beta-Lactam Antibiotics and Related Peptide Synthetases, ed. A. A. Brakhage, Springer, Berlin/Heidelberg, 2004, vol. 88, pp. 91–109.
255. A. Scheidegger, M. T. Kuenzi and J. Nuesch, J. Antibiot., 1984, 37, 522–531.
256. J. E. Baldwin, R. M. Adlington, J. B. Coates, M. J. Crabbe, N. P. Crouch, J. W. Keeping, G. C. Knight, C. J. Schofield, H. H. Ting, C. A. Vallejo, M. Thorniley and E. P. Abraham, Biochem. J., 1987, 245, 831–841.
257. J. E. Dotzlaf and W. K. Yeh, J. Bacteriol., 1987, 169, 1611–1618.
258. S. M. Samson, J. E. Dotzlaf, M. L. Slisz, G. W. Becker, R. M. Van Frank, L. E. Veal, W. K. Yeh, J. R. Miller, S. W. Queener and T. D. Ingolia, Bio/Technology, 1987, 5, 1207–1208.
259. S. E. Jensen, D. W. S. Westlake and S. Wolfe, Antimicrob. Agents Chemother., 1983, 24, 307–312.
260. S. E. Jensen, D. W. S. Westlake and S. Wolfe, J. Antibiot., 1985, 38, 262–265.
261. S. Kovacevic, B. J. Weigel, M. B. Tobin, T. D. Ingolia and J. R. Miller, J. Bacteriol., 1989, 171, 754–760.
262. B. J. Baker, J. E. Dotzlaf and W. K. Yeh, J. Biol. Chem., 1991, 266, 5087–5093.
263. S. Kovacevic and J. R. Miller, J. Bacteriol., 1991, 173, 398–400.
264. P. Liras, A. Rodriguez-Garcïa and J. F. Martin, Int. Microbiol., 1998, 1, 271–278.
265. C. García-Estrada, F. Fierro and J. F. Martín, in Current Research, Technology and Education Topics in Applied Microbiology and Microbial Biotechnology, ed. A. Mendez-Vilas, Formatex Research Center, Badajoz, Spain, 2010, vol. 1, pp. 577–588.
266. A. A. Brakhage, M. Thön, P. Spröte, D. H. Scharf, Q. Al-Abdallah, S. M. Wolke and P. Hortschansky, Phytochemistry, 70, pp. 1801–1811.
267. K. Valegård, A. C. Van Terwisscha Scheltinga, M. D. Lloyd, T. Hara, S. Ramaswamy, A. Perrakis, A. Thompson, H. J. Lee, J. E. Baldwin, C. J. Schofield, J. Hajdu and I. Andersson, Nature, 1998, 394, 805–809.
268. M. D. Lloyd, H. J. Lee, K. Harlos, Z. H. Zhang, J. E. Baldwin, C. J. Schofield, J. M. Charnock, C. D. Garner, T. Hara, A. C. T. van Scheltinga, K. Valegård, J. A. C. Viklund, J. Hajdu, I. Andersson, A. Danielsson and R. Bhikhabhai, J. Mol. Biol., 1999, 287, 943–960.
269. L. M. Öster, A. C. T. van Scheltinga, K. Valegård, A. M. Hose, A. Dubus, J. Hajdu and I. Andersson, J. Mol. Biol., 2004, 343, 157–171.
270. H. J. Lee, M. D. Lloyd, K. Harlos, I. J. Clifton, J. E. Baldwin and C. J. Schofield, J. Mol. Biol., 2001, 308, 937–948.
271. H. J. Lee, M. D. Lloyd, I. J. Clifton, K. Harlos, A. Dubus, J. E. Baldwin, J. M. Frere and C. J. Schofield, J. Biol. Chem., 2001, 276, 18290–18295.
272. G. Cicero, C. Carbonera, K. Valegård, J. Hajdu, I. Andersson and G. Ranghino, Int. J. Quantum Chem., 2007, 107, 1514–1522.
273. K. Valegård, A. C. Terwisscha Van Scheltinga, A. Dubus, G. Ranghino, L. M. Öster, J. Hajdu and I. Andersson, Nat. Struct. Mol. Biol., 2004, 11, 95–101.
274. P. Loke and T. S. Sim, FEMS Microbiol. Lett., 1999, 179, 423–429.
275. J. Zhou, M. Gunsior, B. O. Bachmann, C. A. Townsend and E. I. Solomon, J. Am. Chem. Soc., 1998, 120, 13539–13540.
276. R. B. Morin, B. G. Jackson, R. A. Mueller, E. R. Lavagnino, W. B. Scanlon and S. L. Andrews, J. Am. Chem. Soc., 1963, 85, 1896–1897.
277. J. E. Baldwin, R. M. Adlington, N. P. Crouch, N. J. Turner and C. J. Schofield, J. Chem. Soc., Chem. Commun., 1989, 1141–1143.
278. H. Kluender, C. H. Bradley, C. J. Sih, P. Fawcett and E. P. Abraham, J. Am. Chem. Soc., 1973, 95, 6149–6150.
279. N. Neuss, C. H. Nash, J. E. Baldwin, P. A. Lemke and J. B. Grutzner, J. Am. Chem. Soc., 1973, 95, 3797–3798.
280. K.-S. Goo, C.-S. Chua and T.-S. Sim, J. Ind. Microbiol. Biotechnol., 2009, 36, 619–633.
281. S. G. Kian, S. C. Chun and T. S. Sim, Proteins: Struct., Funct., Genet., 2008, 70, 739–747.
282. K. S. Goo, C. S. Chua and T.-S. Sim, Appl. Environ. Microbiol., 2008, 74, 1167–1175.
283. M. D. Lloyd, S. J. Lipscomb, K. S. Hewitson, C. M. H. Hensgens, J. E. Baldwin and C. J. Schofield, J. Biol. Chem., 2004, 279, 15420–15426.
284. H. S. Chin, K. S. Goo and T. Sim, Appl. Environ. Microbiol., 2004, 70, 607–609.
285. H. J. Lee, Y. F. Dai, C. Y. Shiau, C. J. Schofield and M. D. Lloyd, Eur. J. Biochem., 2003, 270, 1301–1307.
286. S. J. Lipscomb, H. J. Lee, M. Mukherji, J. E. Baldwin, C. J. Schofield and M. D. Lloyd, Eur. J. Biochem., 2002, 269, 2735–2739.
287. H. J. Lee, C. J. Schofield and M. D. Lloyd, Biochem. Biophys. Res. Commun., 2002, 292, 66–70.
288. H. S. Chin and T. S. Sim, Biochem. Biophys. Res. Commun., 2002, 295, 55–61.
289. J. Sim and T. S. Sim, Enzyme Microb. Technol., 2001, 29, 240–245.
290. H. S. Chin, J. Sim and T. S. Sim, Biochem. Biophys. Res. Commun., 2001, 287, 507–513.
291. J. Sim and T. S. Sim, Biosci., Biotechnol., Biochem., 2000, 64, 828–832.
292. H. J. Lee, M. D. Lloyd, K. Harlos and C. J. Schofield, Biochem. Biophys. Res. Commun., 2000, 267, 445–448.
293. A. Dubus, M. D. Lloyd, H. J. Lee, C. J. Schofield, J. E. Baldwin and J. M. Frère, Cell. Mol. Life Sci., 2001, 58, 835–843.
294. C. A. Townsend, A. B. Theis, A. S. Neese, E. B. Barrabee and D. Poland, J. Am. Chem. Soc., 1985, 107, 4760–4767.
295. C. A. Townsend, J. Nat. Prod., 1985, 48, 708–724.
296. C. P. Pang, R. L. White, E. P. Abraham, D. H. Crout, M. Lutstorf, P. J. Morgan and A. E. Derome, Biochem. J., 1984, 222, 777–788.
297. J. E. Baldwin, R. M. Adlington, T. W. Kang, E. Lee and C. J. Schofield, J. Chem. Soc., Chem. Commun., 1987, 104–106.
298. J. E. Baldwin, R. M. Adlington, T. W. Kang, E. Lee and C. J. Schofield, Tetrahedron, 1988, 44, 5953–5957.
299. J. E. Baldwin, R. M. Adlington, R. T. Aplin, N. P. Crouch, G. Knight and C. J. Schofield, J. Chem. Soc., Chem. Commun., 1987, 1651–1654.
300. J. E. Baldwin, R. M. Adlington, N. P. Crouch, C. J. Schofield, N. J. Turner and R. T. Aplin, Tetrahedron, 1991, 47, 9881–9900.
301. R. D. Miller, L. L. Huckstep, J. P. McDermott, S. W. Queener, S. Kukolja, D. O. Spry, T. K. Elzey, S. M. Lawrence and N. Neuss, J. Antibiot., 1981, 34, 984–993.
302. J. Myllyharju and K. I. Kivirikko, EMBO J., 1997, 16, 1173–1180.
303. Y. Fujisawa and T. Kanzaki, J. Antibiot., 1975, 28, 372–378.
304. J. E. Baldwin, K. C. Goh and C. J. Schofield, J. Antibiot., 1992, 45, 1378–1381.
305. C. S. Hamilton, A. Yasuhara, J. E. Baldwin, M. D. Lloyd and P. J. Rutledge, Bioorg. Med. Chem. Lett., 2001, 11, 2511–2514.
306. J. E. Baldwin, R. M. Adlington, J. B. Coates, M. J. C. Crabbe, J. W. Keeping, G. C. Knight, T. Nomoto, C. J. Schofield and H. H. Ting, J. Chem. Soc., Chem. Commun., 1987, 374–375.
307. J. E. Baldwin, R. M. Adlington, M. J. Crabbe, G. Knight, T. Nomoto, C. J. Schofield and H. H. Ting, Tetrahedron, 1987, 43, 3009–3014.
308. N. Shibata, M. D. Lloyd, J. E. Baldwin and C. J. Schofield, Bioorg. Med. Chem. Lett., 1996, 6, 1579–1584.
309. J. E. Baldwin, R. M. Adlington, N. P. Crouch, J. B. Coates, J. W. Keeping, C. J. Schofield, W. A. Shuttleworth and J. D. Sutherland, J. Antibiot., 1988, 41, 1694–1695.
310. N. Morgan, I. A. C. Pereira, I. A. Andersson, R. M. Adlington, J. E. Baldwin, S. C. J. Cole, N. P. Crouch and J. D. Sutherland, Bioorg. Med. Chem. Lett., 1994, 4, 1595–1600.
311. J. E. Baldwin, R. M. Adlington, N. P. Crouch and C. J. Schofield, Tetrahedron, 1988, 44, 643–650.
312. J. E. Baldwin, R. M. Adlington, C. J. Schofield, N. P. Crouch and H. H. Ting, J. Chem. Soc., Chem. Commun., 1987, 1556–1558.
313. J. E. Baldwin, R. M. Adlington, N. P. Crouch and I. A. C. Pereira, Tetrahedron, 1993, 49, 4907–4922.
314. J. E. Baldwin, R. M. Adlington, N. P. Crouch and I. A. C. Pereira, Tetrahedron, 1993, 49, 7499–7518.
315. J. E. Baldwin, R. M. Adlington, N. P. Crouch, J. W. Keeping, S. W. Leppard, J. Pitlik, C. J. Schofield, W. J. Sobey and M. E. Wood, J. Chem. Soc., Chem. Commun., 1991, 768–770.
316. J. E. Baldwin, R. M. Adlington, N. P. Crouch, R. J. Heath, I. A. C. Pereira and J. D. Sutherland, Bioorg. Med. Chem. Lett., 1992, 2, 669–672.
317. J. E. Baldwin, R. M. Adlington, N. P. Crouch and I. A. C. Pereira, J. Chem. Soc., Chem. Commun., 1992, 1448–1451.
318. J. E. Baldwin, R. M. Adlington, N. P. Crouch, R. L. Hill, M. E. Wood and M. D. Lloyd, Bioorg. Med. Chem. Lett., 1997, 7, 593–596.
319. R. Nagarajan, L. D. Boeck, M. Gorman, R. L. Hamill, C. E. Higgens, M. M. Hoehn, W. M. Stark and J. G. Whitney, J. Am. Chem. Soc., 1971, 93, 2308–2310.
320. E. O. Stapley, M. Jackson, S. Hernandez, S. B. Zimmerman, S. A. Currie, S. Mochales, J. M. Mata, H. B. Woodruff and D. Hendlin, Antimicrob. Agents Chemother., 1972, 2, 122–131.
321. G. Gauze, V. Chugasova, L. Terekhova, N. Lomakina and G. Fedorova, Antibiotiki, 1976, 21, 1059–1062.
322. Y. Nozaki, K. Okonogi, N. Katayama, H. Ono, S. Harada, M. Kondo and H. Okazaki, J. Antibiot., 1984, 37, 1555–1565.
323. E. O. Stapley, J. Birnbaum, A. K. Miller, H. Wallick, D. Hendlin and H. B. Woodruff, Rev. Infect. Dis., 1979, 1, 73–89.
324. K. Okonogi, A. Sugiura, M. Kuno, H. Ono, S. Harada and E. Higashide, J. Antibiot., 1985, 38, 1555–1563.
325. J. O'Sullivan, R. Aplin, C. Stevens and E. Abraham, Biochem. J., 1979, 179, 47–52.
326. J. O'Sullivan and E. P. Abraham, Biochem. J., 1980, 186, 613–616.
327. S. J. Brewer, P. M. Taylor and M. K. Turner, Biochem. J., 1980, 185, 555–564.
328. S. J. Brewer, T. T. Boyle and M. K. Turner, Biochem. Soc. Trans., 1977, 5, 1026–1029.
329. J. D. Hood, A. Elson, M. L. Gilpin and A. G. Brown, J. Chem. Soc., Chem. Commun., 1983, 1187–1188.
330. L. M. Iyer, S. Abhiman, R. F. de Souza and L. Aravind, Nucleic Acids Res., 2010, 38, 5261–5279.
331. J. J. R. Coque, F. J. Enguita, J. F. Martin and P. Liras, J. Bacteriol., 1995, 177, 2230–2235.
332. F. J. Enguita, P. Liras, A. L. Leitão and J. F. Martín, J. Biol. Chem., 1996, 271, 33225–33230.
333. X. Xiao, G. Hintermann, A. Häusler, P. J. Barker, F. Foor, A. L. Demain and J. Piret, Antimicrob. Agents Chemother., 1993, 37, 84–88.
334. S. Gutierrez, J. Velasco, F. J. Fernandez and J. F. Martin, J. Bacteriol., 1992, 174, 3056–3064.
335. S. W. Queener, J. J. Capone, A. B. Radue and R. Nagarajan, Antimicrob. Agents Chemother., 1974, 6, 334–337.
336. Y. Fujisawa, K. Kitano and T. Kanzaki, Agric. Biol. Chem., 1975, 39, 2049–2055.
337. Y. Fujisawa and T. Kanzaki, Agric. Biol. Chem., 1975, 39, 2043–2048.
338. H. R. Felix, J. Nüesch and W. Wehrli, FEMS Microbiol. Lett., 1980, 8, 55–58.
339. A. Scheidegger, A. Gutzwiller, M. T. Küenzi, A. Fiechter and J. Niiesch, J. Biotechnol., 1985, 3, 109–117.
340. K. Matsuyama, H. Matsumoto, A. Matsuda, H. Sugiura, K. Komatsu and S. Ichikawa, Biosci., Biotechnol., Biochem., 1992, 56, 1410–1412.
341. J. Velasco, S. Gutierrez, S. Campoy and J. F. Martin, Biochem. J., 1999, 337, 379–385.
342. A. Matsuda, H. Sugiura, K. Matsuyama, H. Matsumoto, S. Ichikawa and K. I. Komatsu, Biochem. Biophys. Res. Commun., 1992, 186, 40–46.
343. A. Matsuda, H. Sugiura, K. Matsuyama, H. Matsumoto, S. Ichikawa and K.-i. Komatsu, Biochem. Biophys. Res. Commun., 1992, 182, 995–1001.
344. S. Lejon, J. Ellis and K. Valegård, J. Mol. Biol., 2008, 377, 935–944.
345. H. Gushima, S. Watanabe, T. Saito, T. Sasaki, H. Eiki, Y. Oka and T. Osono, J. Antibiot., 1981, 34, 1507–1512.
346. R. B. Morin, B. G. Jackson, E. H. Flynn, R. W. Roeske and S. L. Andrews, J. Am. Chem. Soc., 1969, 91, 1396–1400.
347. R. G. Binder, K. Numata, D. A. Lowe, T. Murakami and J. L. Brown, Appl. Environ. Microbiol., 1993, 59, 3321–3326.
348. C. F. Sio and W. J. Quax, Curr. Opin. Biotechnol., 2004, 15, 349–355.
349. T. Isogai, M. Fukagawa, I. Aramori, M. Iwami, H. Kojo, T. Ono, Y. Ueda, M. Kohsaka and H. Imanaka, Bio/Technology, 1991, 9, 188–191.
350. J. Lein, in Overproduction of Microbial Metabolites: Strain Improvement and Process Control Strategies, ed. Z. Vanek and Z. Hostalek, Butterworth Publishers, Boston, 1986, pp. 105–140.
351. G. J. M. Hersbach, C. Van der Beek and P. VanDijck, in Biotechnology of Industrial Antibiotics. Drugs and the Pharmaceutical Sciences, ed. E. J. Vandamme, Marcel Dekker, New York, 1984, vol. 22, pp. 45–140.
352. M. A. van den Berg, R. Albang, K. Albermann, J. H. Badger, J. M. Daran, A. J. M. Driessen, C. Garcia-Estrada, N. D. Fedorova, D. M. Harris, W. H. M. Heijne, V. Joardar, J. A. K. W. Kiel, A. Kovalchuk, J. F. Martin, W. C. Nierman, J. G. Nijland, J. T. Pronk, J. A. Roubos, I. J. van der Klei, N. N. M. E. van Peij, M. Veenhuis, H. von Dohren, C. Wagner, J. Wortman and R. A. L. Bovenberg, Nat. Biotechnol., 2008, 26, 1161–1168.
353. J. Y. Song, H. Jeong, D. S. Yu, M. A. Fischbach, H. S. Park, J. J. Kim, J. S. Seo, S. E. Jensen, T. K. Oh, K. J. Lee and J. F. Kim, J. Bacteriol., 2010, 192, 6317–6318.
354. V. Barbe, M. Bouzon, S. Mangenot, B. Badet, J. Poulain, B. Segurens, D. Vallenet, P. Marlière and J. Weissenbach, J. Bacteriol., 2011, 193, 5055–5061.
355. M. Land, A. Lapidus, S. Mayilraj, F. Chen, A. Copeland, T. G. Del Rio, M. Nolan, S. Lucas, H. Tice, J. F. Cheng, O. Chertkov, D. Bruce, L. Goodwin, S. Pitluck, M. Rohde, M. Goker, A. Pati, N. Ivanova, K. Mavromatis, A. Chen, K. Palaniappan, L. Hauser, Y. J. Chang, C. C. Jeffries, T. Brettin, J. C. Detter, C. Han, P. Chain, B. J. Tindall, J. Bristow, J. A. Eisen, V. Markowitz, P. Hugenholtz, N. C. Kyrpides and H. P. Klenk, Stand. Genomic Sci., 2009, 1, 46–53.
356. T. W. Johannes and H. Zhao, Curr. Opin. Microbiol., 2006, 9, 261–267.
357. R. A. Chica, N. Doucet and J. N. Pelletier, Curr. Opin. Biotechnol., 2005, 16, 378–384.
358. G. J. Williams, A. S. Nelson and A. Berry, Cell. Mol. Life Sci., 2004, 61, 3034–3046.
359. R. Chen, Trends Biotechnol., 2001, 19, 13–14.
360. U. T. Bornscheuer and M. Pohl, Curr. Opin. Chem. Biol., 2001, 5, 137–143.
361. X. B. Wu, X. Y. Tian, J. J. Ji, W. B. Wu, K. Q. Fan and K. Q. Yang, Biotechnol. Lett., 2011, 33, 805–812.
362. J. Thykaer and J. Nielsen, Metab. Eng., 2003, 5, 56–69.
363. M. A. Van den Berg, R. A. L. Bovenberg, W. T. A. M. De Laat and A. G. Van Velzen, Antonie van Leeuwenhoek, 1999, 75, 155–161.
364. M. Rodríguez-Sáiz, J. L. D. L. Fuente, J. L. Barredo and M. C. Flickinger, in Encyclopedia of Industrial Biotechnology, John Wiley & Sons, Inc., 2009.
365. A. A. Brakhage, Microbiol. Mol. Biol. Rev., 1998, 62, 547–585.
366. J. L. Barredo, B. Diez, E. Alvarez and J. F. Martin, Curr. Genet., 1989, 16, 453–459.
367. F. Fierro, C. García-Estrada, N. I. Castillo, R. Rodríguez, T. Velasco-Conde and J.-F. Martín, Fungal Genet. Biol., 2006, 43, 618–629.
368. F. Fierro, J. L. Barredo, B. Díez, S. Gutierrez, F. J. Fernández and J. F. Martín, Proc. Natl. Acad. Sci. U. S. A., 1995, 92, 6200–6204.
369. R. W. Newbert, B. Barton, P. Greaves, J. Harper and G. Turner, J. Ind. Microbiol. Biotechnol., 1997, 19, 18–27.
370. L. Mathison, C. Soliday, T. Stepan, T. Aldrich and J. Rambosek, Curr. Genet., 1993, 23, 33–41.
371. S. Gutiérrez, J. Velasco, A. T. Marcos, F. J. Fernández, F. Fierro, J. L. Barredo, B. Díez and J. F. Martín, Appl. Microbiol. Biotechnol., 1997, 48, 606–614.
372. J. Velasco, J. L. Adrio, M. A. Moreno, B. Diez, G. Soler and J. L. Barredo, Nat. Biotechnol., 2000, 18, 857–861.
373. M. Rodriguez-Sáiz, J.-L. d. l. Fuente and J.-L. Barredo, Methods Biotechnol., 2005, 18, 41–64.
374. R. V. Ullán, S. Campoy, J. Casqueiro, F. J. Fernández and J. F. Martín, Chem. Biol., 2007, 14, 329–339.
375. C. A. Cantwell, R. J. Beckmann, J. E. Dotzlaf, D. L. Fisher, P. L. Skatrud, W. K. Yeh and S. W. Queener, Curr. Genet., 1990, 17, 213–221.
376. L. Crawford, A. M. Stepan, P. C. McAda, J. A. Rambosek, M. J. Conder, V. A. Vinci and C. D. Reeves, Bio/Technology, 1995, 13, 58–62.
377. M. Conder, N. Almeida, S. Behrens, L. Crawford, S. Delawder, T. Hoerner, P. McAda, J. Rambosek, C. Reeves, T. Schimmell, T. Stepan, R. Stieber, I. T. Tong and V. Vinci, presented in part at the Proceedings of the Conference on Applied Biocatalysis, Brighton, UK, 1994.
378. D. M. Harris, I. Westerlaken, D. Schipper, Z. A. van der Krogt, A. K. Gombert, J. Sutherland, L. M. Raamsdonk, M. A. van den Berg, R. A. L. Bovenberg, J. T. Pronk and J.-M. Daran, Metab. Eng., 2009, 11, 125–137.
379. M. J. Koetsier, A. K. Gombert, S. Fekken, R. A. L. Bovenberg, M. A. van den Berg, J. A. K. W. Kiel, P. A. Jekel, D. B. Janssen, J. T. Pronk, I. J. van der Klei and J.-M. Daran, Fungal Genet. Biol., 2010, 47, 33–42.
380. C. Reading, GB Pat., 1547222, 1979.
381. J. Y. Song, S. E. Jensen and K. J. Lee, Appl. Microbiol. Biotechnol., 2010, 88, 659–669.
382. J. Romero, P. Liras and J. F. Martin, Appl. Environ. Microbiol., 1986, 52, 892–897.
383. C. A. Townsend and M. F. Ho, J. Am. Chem. Soc., 1985, 107, 1065–1066.
384. C. A. Townsend and M. F. Ho, J. Am. Chem. Soc., 1985, 107, 1066–1068.
385. J. Pitlik and C. A. Townsend, Chem. Commun., 1997, 225–226.
386. J. E. Thirkettle, J. E. Baldwin, J. Edwards, J. P. Griffin and C. J. Schofield, Chem. Commun., 1997, 1025–1026.
387. B. P. Valentine, C. R. Bailey, A. Doherty, J. Morris, S. W. Elson, K. H. Baggaley and N. H. Nicholson, J. Chem. Soc., Chem. Commun., 1993, 1210–1211.
388. N. Khaleeli, R. Li and C. A. Townsend, J. Am. Chem. Soc., 1999, 121, 9223–9224.
389. R. Perez-Redondo, A. Rodriguez-Garcia, J. F. Martin and P. Liras, J. Bacteriol., 1999, 181, 6922–6928.
390. M. Merski and C. A. Townsend, J. Am. Chem. Soc., 2007, 129, 15750–15751.
391. M. E. C. Caines, J. L. Sorensen and C. J. Schofield, Biochem. Biophys. Res. Commun., 2009, 385, 512–517.
392. M. E. C. Caines, J. M. Elkins, K. S. Hewitson and C. J. Schofield, J. Biol. Chem., 2004, 279, 5685–5692.
393. H. J. McNaughton, J. E. Thirkettle, Z. Zhang, C. J. Schofield, S. E. Jensen, B. Barton and P. Greaves, Chem. Commun., 1998, 2325–2326.
394. B. O. Bachmann, R. Li and C. A. Townsend, Proc. Natl. Acad. Sci. U. S. A., 1998, 95, 9082–9086.
395. M. T. Miller, B. Gerratana, A. Stapon, C. A. Townsend and A. C. Rosenzweig, J. Biol. Chem., 2003, 278, 40996–41002.
396. B. Gerratana, A. Stapon and C. A. Townsend, Biochemistry, 2003, 42, 7836–7847.
397. M. J. Bodner, R. Li, R. M. Phelan, M. F. Freeman, K. A. Moshos, E. P. Lloyd and C. A. Townsend, ChemBioChem, 2011, 12, 2159–2165.
398. T. M. Larsen, S. K. Boehlein, S. M. Schuster, N. G. J. Richards, J. B. Thoden, H. M. Holden and I. Rayment, Biochemistry, 1999, 38, 16146–16157.
399. M. T. Miller, B. O. Bachmann, C. A. Townsend and A. C. Rosenzweig, Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 14752–14757.
400. M. T. Miller, B. O. Bachmann, C. A. Townsend and A. C. Rosenzweig, Nat. Struct. Biol., 2001, 8, 684–689.
401. S. K. Boehlein, J. D. Stewart, E. S. Walworth, R. Thirumoorthy, N. G. J. Richards and S. M. Schuster, Biochemistry, 1998, 37, 13230–13238.
402. R. B. Hamed, L. Henry, J. R. Gomez-Castellanos, J. Mecinović, C. Ducho, J. L. Sorensen, T. D. W. Claridge and C. J. Schofield, J. Am. Chem. Soc., 2012, 134, 471–479.
403. R. B. Hamed, J. R. Gomez-Castellanos, A. Thalhammer, D. Harding, C. Ducho, T. D. W. Claridge and C. J. Schofield, Nat. Chem., 2011, 3, 365–371.
404. R. B. Hamed, J. Mecinovic, C. Ducho, T. D. W. Claridge and C. J. Schofield, Chem. Commun., 2010, 46, 1413–1415.
405. M. C. Sleeman, C. H. MacKinnon, K. S. Hewitson and C. J. Schofield, Bioorg. Med. Chem. Lett., 2002, 12, 597–599.
406. J. W. Labonte, F. Kudo, M. F. Freeman, M. L. Raber and C. A. Townsend, Med. Chem. Commun., 2012, 3, 960.
407. M. C. Sleeman, P. Smith, B. Kellam, S. R. Chhabra, B. W. Bycroft and C. J. Schofield, ChemBioChem, 2004, 5, 879–882.
408. Y.-F. He, B.-Z. Li, Z. Li, P. Liu, Y. Wang, Q. Tang, J. Ding, Y. Jia, Z. Chen, L. Li, Y. Sun, X. Li, Q. Dai, C.-X. Song, K. Zhang, C. He and G.-L. Xu, Science, 2011, 333, 1303–1307.
409. T. Lange, Planta, 1994, 195, 108–115.
410. P. Hedden, Physiol. Plant., 1997, 101, 709–719.
411. P. Hedden, J. Exp. Bot., 1999, 50, 553–563.
412. S. W. Elson, K. H. Baggaley, M. Fulston, N. H. Nicholson, J. W. Tyler, J. Edwards, H. Holms, I. Hamilton and D. M. Mousdale, J. Chem. Soc., Chem. Commun., 1993, 1211–1212.
413. J. E. Baldwin, M. D. Lloyd, B. Wha-Son, C. J. Schofield, S. W. Elson, K. H. Baggaley and N. H. Nicholson, J. Chem. Soc., Chem. Commun., 1993, 500–502.
414. S. W. Elson, K. H. Baggaley, J. Gillett, S. Holland, N. H. Nicholson, J. T. Sime and S. R. Woroniecki, J. Chem. Soc., Chem. Commun., 1987, 1736–1738.
415. J. E. Baldwin, R. M. Adlington, J. S. Bryans, A. O. Bringhen, J. B. Coates, N. P. Crouch, M. D. Lloyd, C. J. Schofield, S. W. Elson, K. H. Baggaley, R. Cassels and N. Nicholson, Tetrahedron, 1991, 47, 4089–4100.
416. S. P. Salowe, E. N. Marsh and C. A. Townsend, Biochemistry, 1990, 29, 6499–6508.
417. J. E. Baldwin, R. M. Adlington, J. S. Bryans, A. O. Bringhen, J. B. Coates, N. P. Crouch, M. D. Lloyd, C. J. Schofield, S. W. Elson and K. H. Baggaley, J. Chem. Soc., Chem. Commun., 1990, 617–1236.
418. S. Salowe, W. Krol, D. Iwata-Reuyl and C. Townsend, Biochemistry, 1991, 30, 2281–2373.
419. J. E. Baldwin, R. M. Adlington, N. P. Crouch, D. J. Drake, Y. Fujishima, S. W. Elson and K. H. Baggaley, J. Chem. Soc., Chem. Commun., 1994, 1133–1134.
420. Z. Zhang, J. Ren, D. K. Stammers, J. E. Baldwin, K. Marios and C. J. Schofield, Nat. Struct. Biol., 2000, 7, 127–133.
421. Z. Zhang, J.-S. Ren, K. Harlos, C. H. McKinnon, I. J. Clifton and C. J. Schofield, FEBS Lett., 2002, 517, 7–12.
422. T. Borowski, A. Bassan and P. E. M. Siegbahn, Chem.–Eur. J., 2004, 10, 1031–1041.
423. R. W. D. Welford, J. M. Kirkpatrick, L. A. McNeill, M. Puri, N. J. Oldham and C. J. Schofield, FEBS Lett., 2005, 579, 5170–5174.
424. J. E. Baldwin, R. M. Adlington, J. S. Bryans, M. D. Lloyd, T. J. Sewell, C. J. Schofield, K. H. Baggaley and R. Cassels, Tetrahedron, 1997, 53, 7011–14031.
425. M. D. Lloyd, K. D. Merritt, V. Lee, T. J. Sewell, B. Wha-Son, J. E. Baldwin, C. J. Schofield, S. W. Elson, K. H. Baggaley and N. H. Nicholson, Tetrahedron, 1999, 55, 10201–10220.
426. J. E. Baldwin, V. Lee, M. D. Lloyd, C. J. Schofield, S. W. Elson and K. H. Baggaley, J. Chem. Soc., Chem. Commun., 1993, 1694–1696.
427. V. Helmetag, S. A. Samel, M. G. Thomas, M. A. Marahiel and L.-O. Essen, FEBS J., 2009, 276, 3669–3682.
428. K. J. Martinkus, C.-H. Tann and S. J. Gould, Tetrahedron, 1983, 39, 3493–3505.
429. M. Jackson, S. Gould and T. Zabriskie, J. Org. Chem., 2002, 67, 2934–2975.
430. H. Kawauchi, M. Tohno, Y. Tsuchiya, M. Hayashida, Y. Adachi, T. Mukai, I. Hayashi, S. Kimura and S. Kondo, Int. J. Pept. Protein Res., 1983, 21, 546–554.
431. K. Aidoo, A. Wong, D. Alexander, R. Rittammer and S. Jensen, Gene, 1994, 147, 41–47.
432. J. M. Elkins, I. J. Clifton, H. Hernández, L. X. Doan, C. V. Robinson, C. J. Schofield and K. S. Hewitson, Biochem. J., 2002, 366, 423–434.
433. M. C. Bewley, P. D. Jeffrey, M. L. Patchett, Z. F. Kanyo and E. N. Baker, Structure, 1999, 7, 435–448.
434. D. Iwata-Reuyl and C. A. Townsend, J. Am. Chem. Soc., 1992, 114, 2762–2763.
435. J. W. Janc, L. A. Egan and C. A. Townsend, Bioorg. Med. Chem. Lett., 1993, 3, 2313–2316.
436. L. A. Egan, R. W. Busby, D. Iwata-Reuyl and C. A. Townsend, J. Am. Chem. Soc., 1997, 119, 2348–2355.
437. H. Peter, J. Rabenhorst, F. Röhl and H. Zähner, in Recent advances in chemotherapy, ed. J. Ishigami, University of Tokyo Press, Tokyo, 1985, pp. 237–238.
438. F. Röhl, J. Rabenhorst and H. Zähner, Arch. Microbiol., 1987, 147, 315–320.
439. J. E. Baldwin, T. D. W. Claridge, K.-C. Goh, J. W. Keeping and C. J. Schofield, Tetrahedron Lett., 1993, 34, 5645–5648.
440. H. T. Postels and W. A. König, Tetrahedron Lett., 1994, 35, 4535–4538.
441. J. E. Baldwin, Y. Fujishima, K.-C. Goh and C. J. Schofield, Tetrahedron Lett., 1994, 35, 2783–2786.
442. J. E. Baldwin, K. C. Goh and C. J. Schofield, Tetrahedron Lett., 1994, 35, 2779–2782.
443. J. W. Janc, L. A. Egan and C. A. Townsend, J. Biol. Chem., 1995, 270, 5399–5404.
444. J. S. Box, DE 2646001, 1977.
445. K. Kitano, K. Kazuaki, K. Kintaka, K. Katamoto, JP 53104796, 1978.
446. S. W. Elson, in Recent Advances in the Chemistry of β-Lactam Antibiotics, ed. P. H. Bentley and R. Southgate, RSC, Cambridge, UK, 1988, p. 303.
447. C. E. Newal, in Recent Advances in the Chemistry of the β-Lactam Antibiotics, ed. G. I. Gregory, RSC, London, UK, 1980, p. 142.
448. B. W. Bycroft, A. Penrose, J. Gillett and S. W. Elson, J. Chem. Soc., Chem. Commun., 1988, 980–981.
449. C. A. Townsend and W. J. Krol, J. Chem. Soc., Chem. Commun., 1988, 1234–1236.
450. N. H. Nicholson, K. H. Baggaley, R. Cassels, M. Davison, S. W. Elson, M. Fulston, J. W. Tyler and S. R. Woroniecki, J. Chem. Soc., Chem. Commun., 1994, 1281–1282.
451. M. Fulston, M. Davison, S. W. Elson, N. H. Nicholson, J. W. Tyler and S. R. Woroniecki, J. Chem. Soc., Perkin Trans. 1, 2001, 1122–1130.
452. D. Iwata-Reuyl, A. Basak, L. S. Silverman, C. A. Engle and C. A. Townsend, J. Nat. Prod., 1993, 56, 1373–1396.
453. S. Jensen, A. Paradkar, R. Mosher, C. Anders, P. Beatty, M. Brumlik, A. Griffin and B. Barton, Antimicrob. Agents Chemother., 2004, 48, 192–202.
454. R. F. Li, N. Khaleeli and C. A. Townsend, J. Bacteriol., 2000, 182, 4087–4095.
455. E. Mellado, L. M. Lorenzana, M. Rodriguez-Saiz, B. Diez, P. Liras and J. L. Barredo, Microbiology, 2002, 148, 1427–1438.
456. A. S. Paradkar, K. A. Aidoo and S. E. Jensen, Mol. Microbiol., 1998, 27, 831–843.
457. R. Perez-Redondo, A. Rodriguez-Garcia, J. F. Martin and P. Liras, Gene, 1998, 211, 311–321.
458. P. Liras, J. P. Gomez-Escribano and I. Santamarta, J. Ind. Microbiol. Biotechnol., 2008, 35, 667–676.
459. H. Arulanantham, N. J. Kershaw, K. S. Hewitson, C. E. Hughes, J. E. Thirkettle and C. J. Schofield, J. Biol. Chem., 2006, 281, 279–287.
460. M. Fawaz, M. Topper and S. Firestine, Bioorg. Chem., 2011, 39, 185–276.
461. M. Galperin and E. Koonin, Protein Sci., 1997, 6, 2639–2682.
462. F. C. Neuhaus, C. V. Carpenter, J. L. Miller, N. M. Lee, M. Gragg and R. A. Stickgold, Biochemistry, 1969, 8, 5119–5124.
463. Y. Kallberg, U. Oppermann, H. Jörnvall and B. Persson, Eur. J. Biochem., 2002, 269, 4409–4417.
464. N. Tanaka, T. Nonaka, K. T. Nakamura and A. Hara, Curr. Org. Chem., 2001, 5, 89–111.
465. H. Joernvall, B. Persson, M. Krook, S. Atrian, R. Gonzalez-Duarte, J. Jeffery and D. Ghosh, Biochemistry, 1995, 34, 6003–6013.
466. U. Oppermann, C. Filling, M. Hult, N. Shafqat, X. Wu, M. Lindh, J. Shafqat, E. Nordling, Y. Kallberg, B. Persson and H. Jornvall, Chem.-Biol. Interact., 2003, 143–144, 247–253.
467. M. G. Rossmann, D. Moras and K. W. Olsen, Nature, 1974, 250, 194–199.
468. C. Filling, K. D. Berndt, J. Benach, S. Knapp, T. Prozorovski, E. Nordling, R. Ladenstein, H. Jornvall and U. Oppermann, J. Biol. Chem., 2002, 277, 25677–25684.
469. A. K. MacKenzie, N. J. Kershaw, H. Hernandez, C. V. Robinson, C. J. Schofield and I. Andersson, Biochemistry, 2007, 46, 1523–1533.
470. R. K. Wierenga, P. Terpstra and W. G. J. Hol, J. Mol. Biol., 1986, 187, 101–107.
471. Z. Chen, J. C. Jiang, Z. G. Lin, W. R. Lee, M. E. Baker and S. H. Chang, Biochemistry, 1993, 32, 3342–3346.
472. A. G. Brown, D. F. Corbett, J. Goodacre, J. B. Harbridge, T. T. Howarth, R. J. Ponsford, I. Stirling and T. J. King, J. Chem. Soc., Perkin Trans. 1, 1984, 635–650.
473. T. T. Howarth, A. G. Brown and T. J. King, J. Chem. Soc., Chem. Commun., 1976, 266b–267.
474. C. A. Townsend, M.-F. Ho and S.-S. Mao, J. Chem. Soc., Chem. Commun., 1986, 638–639.
475. L. M. Lorenzana, R. Pérez-Redondo, I. Santamarta, J. F. Martín and P. Liras, J. Bacteriol., 2004, 186, 3431–3438.
476. A. K. Mackenzie, K. Valegård, A. Iqbal, M. E. C. Caines, N. J. Kershaw, S. E. Jensen, C. J. Schofield and I. Andersson, J. Mol. Biol., 2010, 396, 332–344.
477. J. R. H. Tame, G. N. Murshudov, E. J. Dodson, T. K. Neil, G. G. Dodson, C. F. Higgins and A. J. Wilkinson, Science, 1994, 264, 1578–1581.
478. F. A. Quiocho and P. S. Ledvina, Mol. Microbiol., 1996, 20, 17–25.
479. S. W. Elson, J. Gillett, N. H. Nicholson and J. W. Tyler, J. Chem. Soc., Chem. Commun., 1988, 979–980.
480. A. Iqbal, H. Arunlanantham, T. Brown, R. Chowdhury, I. J. Clifton, N. J. Kershaw, K. S. Hewitson, M. A. McDonough and C. J. Schofield, Proteins: Struct. Funct. Bioinform., 2010, 78, 1398–1407.
481. M. Vetting, L. S. de Carvalho, M. Yu, S. Hegde, S. Magnet, S. Roderick and J. Blanchard, Arch. Biochem. Biophys., 2005, 433, 212–238.
482. F. Dyda, D. Klein and A. Hickman, Annu. Rev. Biophys. Biomol. Struct., 2000, 29, 81–184.
483. H. He, Y. Ding, M. Bartlam, F. Sun, Y. Le, X. Qin, H. Tang, R. Zhang, A. Joachimiak, J. Liu, N. Zhao and Z. Rao, J. Mol. Biol., 2003, 325, 1019–1049.
484. W. Dou, M. Xu, D. Cai, X. Zhang, Z. Rao and Z. Xu, Appl. Biochem. Biotechnol., 2011, 165, 845–900.
485. A. Rodríguez-García, M. Ludovice, J. F. Martín and P. Liras, Mol. Microbiol., 1997, 25, 219–228.
486. J. E. Hodgson, A. P. Fosberry, N. S. Rawlinson, H. N. M. Ross, R. J. Neal, J. C. Arnell, A. J. Earl and E. J. Lawlor, Gene, 1995, 166, 49–55.
487. A. Rodríguez-García, Á. de la Fuente, R. Pérez-Redondo, J. F. Martín and P. Liras, J. Mol. Microbiol. Biotechnol., 2000, 2, 543–550.
488. A. de la Fuente, J. F. Martin, A. Rodriguez-Garcia and P. Liras, J. Bacteriol., 2004, 186, 6501–6507.
489. K. Tahlan, H. U. Park, A. Wong, P. H. Beatty and S. E. Jensen, Antimicrob. Agents Chemother., 2004, 48, 930–939.
490. J. M. Elkins, N. J. Kershaw and C. J. Schofield, Biochem. J., 2005, 385, 565–573.
491. A. Iqbal, I. J. Clifton, R. Chowdhury, D. Ivison, C. Domene and C. J. Schofield, Org. Biomol. Chem., 2011, 9, 6219–6225.
492. A. Iqbal, I. J. Clifton, M. Bagonis, N. J. Kershaw, C. Domene, T. D. W. Claridge, C. W. Wharton and C. J. Schofield, J. Am. Chem. Soc., 2009, 131, 749–757.
493. R. Sankaranarayanan, M. M. Cherney, C. Garen, G. Garen, C. Niu, M. Yuan and M. N. James, J. Mol. Biol., 2010, 397, 979–990.
494. R. C. Wilmouth, K. Edman, R. Neutze, P. A. Wright, I. J. Clifton, T. R. Schneider, C. J. Schofield and J. Hajdu, Nat. Struct. Biol., 2001, 8, 689–694.
495. G. Katona, R. C. Wilmouth, P. A. Wright, G. I. Berglund, J. Hajdu, R. Neutze and C. J. Schofield, J. Biol. Chem., 2002, 277, 21962–21970.
496. E. S. Radisky, J. M. Lee, C. J. Lu and D. E. Koshland, Jr, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 6835–6840.
497. P. Deslongchamps, Stereoelectronic effects in organic chemistry, Pergamon Press, New York, 1983.
498. P. Liras, I. Santamarta and R. Perez-Redondo, in Streptomyces: Molecular Biology and Biotechnology, ed. P. Dypson, Caister Academic Press, Wymondham, UK, 2011, pp. 167–178.
499. P. Liras and A. Rodriguez-Garcia, Appl. Microbiol. Biotechnol., 2000, 54, 467–475.
500. S. E. Jensen and A. S. Paradkar, Antonie van Leeuwenhoek, 1999, 75, 125–133.
501. K. Tahlan, H. U. Park and S. E. Jensen, Can. J. Microbiol., 2004, 50, 803–810.
502. J. M. Ward and J. E. Hodgson, FEMS Microbiol. Lett., 1993, 110, 239–242.
503. M. H. Medema, A. Trefzer, A. Kovalchuk, M. van den Berg, U. Müller, W. Heijne, L. Wu, M. T. Alam, C. M. Ronning, W. C. Nierman, R. A. L. Bovenberg, R. Breitling and E. Takano, Genome Biol. Evol., 2010, 2, 212–224.
504. N. J. Zelyas, H. Cai, T. Kwong and S. E. Jensen, J. Bacteriol., 2008, 190, 7957–7965.
505. S. E. Jensen, K. J. Elder, K. A. Aidoo and A. S. Paradkar, Antimicrob. Agents Chemother., 2000, 44, 720–726.
506. H. Jnawali, T.-J. Oh, K. Liou, B. Park and J. Sohng, J. Antibiot., 2008, 61, 651–660.
507. J. Song, E. Kim, D. Kim, S. Jensen and K. Lee, J. Ind. Microbiol. Biotechnol., 2009, 36, 301–312.
508. H. Jnawali, K. Liou and J. Sohng, Microbiol. Res., 2011, 166, 369–448.
509. H. Jnawali, J. Yoo and J. Sohng, Biotechnol. Lett., 2011, 33, 1221–1227.
510. M. Medema, M. Alam, W. Heijne, M. van den Berg, U. Müller, A. Trefzer, R. Bovenberg, R. Breitling and E. Takano, Microb. Biotechnol., 2011, 4, 300–305.
511. R. F. Li and C. A. Townsend, Metab. Eng., 2006, 8, 240–252.
512. K. Tahlan, C. Anders and S. E. Jensen, J. Bacteriol., 2004, 186, 6286–6297.
513. S. E. Jensen, A. Wong, A. Griffin and B. Barton, Antimicrob. Agents Chemother., 2004, 48, 514–520.
514. K. Tahlan, C. Anders, A. Wong, R. H. Mosher, P. H. Beatty, M. J. Brumlik, A. Griffin, C. Hughes, J. Griffin, B. Barton and S. E. Jensen, Chem. Biol., 2007, 14, 131–142.
515. R. H. Mosher, A. S. Paradkar, C. Anders, B. Barton and S. E. Jensen, Antimicrob. Agents Chemother., 1999, 43, 1215–1224.
516. S. Goomeshi Nobary and S. E. Jensen, Can. J. Microbiol., 2012, 58, 413–425.
517. W. Dürckheimer, J. Blumbach, R. Lattrell and K. H. Scheunemann, Angew. Chem., Int. Ed. Engl., 1985, 24, 180–202.
518. W. L. Parker, M. L. Rathnum, J. S. Wells, Jr., W. H. Trejo, P. A. Principe and R. B. Sykes, J. Antibiot., 1982, 35, 653–660.
519. Y. Nakamura, K. Ishii, E. Ono, M. Ishikara, T. Kohda, Y. Yokogawa and H. Shibai, J. Antibiot., 1988, 41, 707–711.
520. F. M. Kahan, H. Kropp, J. G. Sundelof and J. Birnbaum, J. Antimicrob. Chemother., 1983, 12, 1–35.
521. W. J. Leanza, K. J. Windonger, T. W. Miller and B. G. Christensen, J. Med. Chem., 1979, 22, 1435–1436.
522. D. H. Shih, L. Cama and B. G. Christensen, Tetrahedron Lett., 1985, 26, 587–590.
523. J. M. Williamson, E. Inamine, K. E. Wilson, A. W. Douglas, J. M. Liesch and G. Albers-Schönberg, J. Biol. Chem., 1985, 260, 4637–4647.
524. B. W. Bycroft, C. Maslen, S. J. Box, A. Brown and J. W. Tyler, J. Antibiot., 1988, 41, 1231–1242.
525. J. M. Williamson, Crit. Rev. Biotechnol., 1986, 4, 111–173.
526. S. J. Coulthurst, A. M. L. Barnard and G. P. C. Salmond, Nat. Rev. Microbiol., 2005, 3, 295–306.
527. S. Derzelle, E. Duchaud, F. Kunst, A. Danchin and P. Bertin, Appl. Environ. Microbiol., 2002, 68, 3780–3789.
528. B. W. Bycroft, C. Maslen, S. J. Box, A. G. Brown and J. W. Tyler, J. Chem. Soc., Chem. Commun., 1987, 1623–1625.
529. S. J. McGowan, M. Sebaihia, S. O'Leary, K. R. Hardie, P. Williams, G. S. A. B. Stewart, B. W. Bycroft and G. P. C. Salmond, Mol. Microbiol., 1997, 26, 545–556.
530. S. J. McGowan, M. Sebaihia, L. E. Porter, G. S. A. B. Stewart, P. Williams, B. W. Bycroft and G. P. C. Salmond, Mol. Microbiol., 1996, 22, 415–426.
531. A. Stapon, R. Li and C. A. Townsend, J. Am. Chem. Soc., 2003, 125, 8486–8493.
532. B. W. Bycroft and S. R. Chhabra, J. Chem. Soc., Chem. Commun., 1989, 423–425.
533. H. Tanaka, H. Sakagami and K. Ogasawara, Tetrahedron Lett., 2002, 43, 93–96.
534. R. F. Li, A. Stapon, J. T. Blanchfield and C. A. Townsend, J. Am. Chem. Soc., 2000, 122, 9296–9297.
535. S. J. McGowan, M. Sebaihia, S. Jones, B. Yu, N. Bainton, P. F. Chan, B. W. Bycroft, G. S. A. B. Stewart, P. Williams and G. P. C. Salmond, Microbiology, 1995, 141, 541–550.
536. S. J. McGowan, M. T. G. Holden, B. W. Bycroft and G. P. C. Salmond, Antonie van Leeuwenhoek, 1999, 75, 135–141.
537. M. C. Sleeman and C. J. Schofield, J. Biol. Chem., 2004, 279, 6730–6736.
538. B. Gerratana, S. O. Arnett, A. Stapon and C. A. Townsend, Biochemistry, 2004, 43, 15936–15945.
539. R. B. Hamed, E. T. Batchelar, J. Mecinović, T. D. W. Claridge and C. J. Schofield, ChemBioChem, 2009, 10, 246–250.
540. J. L. Sorensen, M. C. Sleeman and C. J. Schofield, Chem. Commun., 2005, 1155–1157.
541. M. C. Sleeman, J. L. Sorensen, E. T. Batchelar, M. A. McDonough and C. J. Schofield, J. Biol. Chem., 2005, 280, 34956–34965.
542. I. J. Clifton, L. X. Doan, M. C. Sleeman, M. Topf, H. Suzuki, R. C. Wilmouth and C. J. Schofield, J. Biol. Chem., 2003, 278, 20843–20850.
543. R. B. Hamed, E. T. Batchelar, I. J. Clifton and C. J. Schofield, Cell. Mol. Life Sci., 2008, 65, 2507–2527.
544. H. M. Holden, M. M. Benning, T. Haller and J. A. Gerlt, Acc. Chem. Res., 2001, 34, 145–157.
545. M. Rodriguez, L. E. Nunez, A. F. Brana, C. Mendez, J. A. Salas and G. Blanco, Antimicrob. Agents Chemother., 2011, 55, 1638–1649.
546. E. T. Batchelar, R. B. Hamed, C. Ducho, T. D. W. Claridge, M. J. Edelmann, B. Kessler and C. J. Schofield, Angew. Chem., Int. Ed., 2008, 47, 9322–9325.
547. C. Ducho, R. B. Hamed, E. T. Batchelar, J. L. Sorensen, B. Odell and C. J. Schofield, Org. Biomol. Chem., 2009, 7, 2770–2779.
548. S. L. Bearne and R. Wolfendon, Biochemistry, 1995, 34, 11515–11520.
549. K. T. Douglas, Acc. Chem. Res., 1986, 19, 186–192.
550. B. J. Wong and J. A. Gerlt, J. Am. Chem. Soc., 2003, 125, 12076–12077.
551. T. J. Erb, V. Brecht, G. Fuchs, M. Muller and B. E. Alber, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 8871–8876.
552. T. J. Erb, I. A. Berg, V. Brecht, M. Muller, G. Fuchs and B. E. Alber, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 10631–10636.
553. A. S. Eustáquio, R. P. McGlinchey, Y. Liu, C. Hazzard, L. L. Beer, G. Florova, M. M. Alhamadsheh, A. Lechner, A. J. Kale, Y. Kobayashi, K. A. Reynolds and B. S. Moore, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 12295–12300.
554. M. Wilson and B. Moore, Nat. Prod. Rep., 2012, 29, 72–158.
555. B. Alber, Appl. Microbiol. Biotechnol., 2011, 89, 17–42.
556. H.-G. Yoo, S.-Y. Kwon, S. Kim, S. Karki, Z.-Y. Park and H.-J. Kwon, Biosci., Biotechnol., Biochem., 2011, 75, 1191–1194.
557. N. Quade, L. Huo, S. Rachid, D. Heinz and R. Müller, Nat. Chem. Biol., 2011, 8, 117–141.
558. J. J. Turnbull, J. I. Nakajima, R. W. D. Welford, M. Yamazaki, K. Saito and C. J. Schofield, J. Biol. Chem., 2004, 279, 1206–1216.
559. N. H. Horowitz, Proc. Natl. Acad. Sci. U. S. A., 1945, 31, 153–157.
560. A. Stapon, R. Li and C. A. Townsend, J. Am. Chem. Soc., 2003, 125, 15746–15747.
561. M. Topf, G. M. Sandala, D. M. Smith, C. J. Schofield, C. J. Easton and L. Radom, J. Am. Chem. Soc., 2004, 126, 9932–9933.
562. I. K. Leung, T. J. Krojer, G. T. Kochan, L. Henry, F. von Delft, T. D. Claridge, U. Oppermann, M. A. McDonough and C. J. Schofield, Chem. Biol., 2010, 17, 1316–1324.
563. T. Borowski, E. Broclawik, C. J. Schofield and P. E. M. Siegbahn, J. Comput. Chem., 2006, 27, 740–748.
564. D. R. Houck, K. Kobayashi, J. M. Williamson and H. G. Floss, J. Am. Chem. Soc., 1986, 108, 5365–5366.
565. L. E. Nunez, C. Mendez, A. F. Brana, G. Blanco and J. A. Salas, Chem. Biol., 2003, 10, 301–311.
566. M. J. Bodner, R. M. Phelan, M. F. Freeman, R. Li and C. A. Townsend, J. Am. Chem. Soc., 2010, 132, 12–13.
567. M. Rodriguez, L. Nunez, A. Brana, C. Mendez, J. Salas and G. Blanco, Mol. Microbiol., 2008, 69, 633–645.
568. H. Suzuki, Y. Ohnishi and S. Horinouchi, J. Antibiot., 2007, 60, 380–387.
569. D. O'Hagan and J. W. Schmidberger, Nat. Prod. Rep., 2010, 27, 900–918.
570. T. M. Zydowsky, L. F. Courtney, V. Frasca, K. Kobayashi, H. Shimizu, L. D. Yuen, R. G. Matthews, S. J. Benkovic and H. G. Floss, J. Am. Chem. Soc., 1986, 108, 3152–3153.
571. E. O. Stapley, P. J. Cassidy, J. Tunac, R. L. Monaghan, M. Jackson, S. Hernandez, S. B. Zimmerman, J. M. Mata, S. A. Currie, D. Daoust and D. Hendlin, J. Antibiot., 1981, 34, 628–636.
572. D. Butterworth, M. Cole, G. Hanscomb and G. N. Rolinson, J. Antibiot., 1979, 32, 287–294.
573. A. Imada, Y. Nozaki, K. Kintaka, K. Okonogi, K. Kitano and S. Harada, J. Antibiot., 1980, 33, 1417–1424.
574. T. S. Chen, B. H. Arison, C. L. Ruby, A. W. Dombrowski and E. S. Inamine, J. Ind. Microbiol. Biotechnol., 1993, 12, 66–67.
575. A. J. Brooks, M. Vlasie, R. Banerjee and T. C. Brunold, J. Am. Chem. Soc., 2005, 127, 16522–16528.
576. A. J. Brooks, M. Vlasie, R. Banerjee and T. C. Brunold, J. Am. Chem. Soc., 2004, 126, 8167–8180.
577. B. Zagalak and J. Retey, Eur. J. Biochem., 1974, 44, 529–535.
578. E. Marsh, D. Patterson and L. Li, ChemBioChem, 2010, 11, 604–625.
579. P. Frey, A. Hegeman and F. Ruzicka, Crit. Rev. Biochem. Mol. Biol., 2008, 43, 63–88.
580. M. Challand, R. Driesener and P. Roach, Nat. Prod. Rep., 2011, 28, 1696–2417.
581. K. Kubo, T. Ishikura and Y. Fukagawa, J. Antibiot., 1985, 38, 904–911.
582. Y. Fukagawa, M. Okabe, S. Azuma, I. Kojima, T. Ishikura and K. Kubo, J. Antibiot., 1984, 37, 1388–1393.
583. K. Kubo, T. Ishikura and Y. Fukagawa, J. Antibiot., 1984, 37, 1394–1402.
584. K. Kubo, T. Ishikura and Y. Fukagawa, J. Antibiot., 1985, 38, 622–630.
585. L. W. Kang, S. B. Gabelli, M. A. Bianchet, W. L. Xu, M. J. Bessman and L. M. Amzel, J. Bacteriol., 2003, 185, 4110–4118.
586. E. V. Koonin and R. L. Tatusov, J. Mol. Biol., 1994, 244, 125–132.
587. A. R. Buller, M. F. Freeman, N. T. Wright, J. F. Schildbach and C. A. Townsend, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 2308–2313.
588. H. B. Burgi, J. D. Dunitz and E. Shefter, J. Am. Chem. Soc., 1973, 95, 5065–5067.
589. W. L. Parker, J. O'Sullivan and R. B. Sykes, in Adv. Appl. Microbiol., ed. I. L. Allen, Academic Press, 1986, vol. 31, pp. 181–205.
590. J. P. Scannell, D. Pruess, J. F. Blount, H. A. Ax, M. Kellett, F. Weiss, T. C. Demny, T. H. Williams and A. Stempel, J. Antibiot., 1975, 28, 1–6.
591. K. P. Devkota, B. N. Lenta, P. A. Fokou and N. Sewald, Nat. Prod. Rep., 2008, 25, 612–630.
592. H. Watanabe, K. Watanabe, M. Shimadzu, T. Kikuchi and Z. Liu, Planta Med., 1986, 56–58.
593. H. de Sousa Falcao, J. A. Leite, J. M. Barbosa-Filho, P. F. de Athayde-Filho, M. C. de Oliveira Chaves, M. D. Moura, A. L. Ferreira, A. B. de Almeida, A. R. Souza-Brito, M. de Fatima Formiga Melo Diniz and L. M. Batista, Molecules, 2008, 13, 3198–3223.
594. S. Funayama, T. Noshita, K. Shinoda, N. Haga, S. Nozoe, M. Hayashi and K. Komiyama, Biol. Pharm. Bull., 2000, 23, 262–264.
595. T. Kikuchi and S. Uyeo, Chem. Pharm. Bull., 1967, 15, 549–570.
596. T. Kikuchi and S. Uyeo, Tetrahedron Lett., 1965, 39, 3473–3485.
597. L. C. Chang, K. P. L. Bhat, H. H. S. Fong, J. M. Pezzuto and A. D. Kinghorn, Tetrahedron, 2000, 56, 3133–3138.
598. L. C. Chang, K. P. L. Bhat, E. Pisha, E. J. Kennelly, H. H. S. Fong, J. M. Pezzuto and A. D. Kinghorn, J. Nat. Prod., 1998, 61, 1257–1262.
599. E. Avilés and A. D. Rodríguez, Org. Lett., 2010, 12, 5290–5293.
600. R. B. Sykes and D. P. Bonner, Rev. Infect. Dis., 1985, 7, S579–593.
601. S. Han, N. Caspers, R. P. Zaniewski, B. M. Lacey, A. P. Tomaras, X. Feng, K. F. Geoghegan and V. Shanmugasundaram, J. Am. Chem. Soc., 2011, 133, 20536–20545.
602. J. O'Sullivan, A. M. Gillum, C. A. Aklonis, M. L. Souser and R. B. Sykes, Antimicrob. Agents Chemother., 1982, 21, 558–564.
603. J. O'Sullivan, M. L. Souser, C. C. Kao and C. A. Aklonis, Antimicrob. Agents Chemother., 1983, 23, 598–602.
604. J. O'Sullivan and C. A. Aklonis, J. Antibiot., 1984, 37, 804–806.
605. L. J. Nisbet, R. J. Mehta, Y. Oh, C. H. Pan, C. G. Phelen, M. J. Polansky, M. C. Shearer, A. J. Giovenella and S. F. Grappel, J. Antibiot., 1985, 38, 133–138.
606. J. A. Chan, E. A. Shultis, J. J. Dingerdissen, C. W. DeBrosse, G. D. Roberts and K. M. Snader, J. Antibiot., 1985, 38, 139–144.
607. N. Katayama, Y. Nozaki, K. Okonogi, H. Ono, S. Harada and H. Okazaki, J. Antibiot., 1985, 38, 1117–1127.
608. T. Hida, S. Tsubotani, N. Katayama, H. Okazaki and S. Harada, J. Antibiot., 1985, 38, 1128–1140.
609. J. Hosoda, T. Konomi and N. Tani, Agric. Biol. Chem., 1977, 41, 2013–2020.
610. M. Kurita, K. Jomon, T. Komori, N. Miyairi and H. Aoki, J. Antibiot., 1976, 29, 1243–1245.
611. W. L. Parker, in Journal of Chromatography Library, ed. H. W. Gerald and C. Raymond, Elsevier, 1989, vol. 43, ch. 7, pp. 259–292.
612. P. H. Milner, A. W. Guest, F. P. Harrington, R. J. Ponsford, T. C. Smale and A. V. Stachulski, J. Chem. Soc., Chem. Commun., 1984, 1335–1336.
613. M. J. Basker, R. A. Edmondson, S. J. Knott, R. J. Ponsford, B. Slocombe and S. J. White, Antimicrob. Agents Chemother., 1984, 26, 734–740.
614. J. Hosoda, N. Tani and T. Konomi, Agric. Biol. Chem., 1977, 41, 2007–2012.
615. C. A. Townsend and A. M. Brown, J. Am. Chem. Soc., 1981, 103, 2873–2874.
616. C. A. Townsend and L. T. Nguyen, J. Am. Chem. Soc., 1981, 103, 4582–4583.
617. C. A. Townsend, A. M. Brown and L. T. Nguyen, J. Am. Chem. Soc., 1983, 105, 919–927.
618. C. A. Townsend, G. M. Salituro, L. T. Nguyen and M. J. DiNovi, Tetrahedron Lett., 1986, 27, 3819–3822.
619. C. A. Townsend and A. M. Brown, J. Am. Chem. Soc., 1983, 105, 913–918.
620. C. A. Townsend, A. M. Reeve and G. M. Salituro, J. Chem. Soc., Chem. Commun., 1988, 1579–1581.
621. C. A. Townsend and G. M. Salituro, J. Chem. Soc., Chem. Commun., 1984, 1631–1632.
622. C. A. Townsend and B. A. Wilson, J. Am. Chem. Soc., 1988, 110, 3320–3321.
623. B. A. Wilson, S. Bantia, G. M. Salituro, A. M. Reeve and C. A. Townsend, J. Am. Chem. Soc., 1988, 110, 8238–8239.
624. M. Gunsior, S. D. Breazeale, A. J. Lind, J. Ravel, J. W. Janc and C. A. Townsend, Chem. Biol., 2004, 11, 927–938.
625. A. M. Reeve, S. D. Breazeale and C. A. Townsend, J. Biol. Chem., 1998, 273, 30695–30703.
626. W. L. Kelly and C. A. Townsend, J. Biol. Chem., 2004, 279, 38220–38227.
627. W. L. Kelly and C. A. Townsend, J. Bacteriol., 2005, 187, 739–746.
628. J. M. Davidsen and C. A. Townsend, J. Bacteriol., 2009, 191, 1066–1077.
629. A. M. van Wageningen, P. N. Kirkpatrick, D. H. Williams, B. R. Harris, J. K. Kershaw, N. J. Lennard, M. Jones, S. J. Jones and P. J. Solenberg, Chem. Biol., 1998, 5, 155–162.
630. P. P. Chong, S. M. Podmore, H. M. Kieser, M. Redenbach, K. Turgay, M. Marahiel, D. A. Hopwood and C. P. Smith, Microbiology, 1998, 144, 193–199.
631. H. T. Chiu, B. K. Hubbard, A. N. Shah, J. Eide, R. A. Fredenburg, C. T. Walsh and C. Khosla, Proc. Natl. Acad. Sci. U. S. A., 2001, 98, 8548–8553.
632. B. K. Hubbard, M. G. Thomas and C. T. Walsh, Chem. Biol., 2000, 7, 931–942.
633. S. C. Wenzel and R. Müller, Curr. Opin. Chem. Biol., 2005, 9, 447–458.
634. G. L. Challis, J. Ravel and C. A. Townsend, Chem. Biol., 2000, 7, 211–224.
635. T. Stachelhaus, H. D. Mootz and M. A. Marahiel, Chem. Biol., 1999, 6, 493–505.
636. G. L. Challis and J. Ravel, FEMS Microbiol. Lett., 2000, 187, 111–114.
637. S. Lautru, R. J. Deeth, L. M. Bailey and G. L. Challis, Nat. Chem. Biol., 2005, 1, 265–269.
638. Y. Shikata, T. Watanabe, T. Teramoto, A. Inoue, Y. Kawakami, Y. Nishizawa, K. Katayama and M. Kuwada, J. Biol. Chem., 1995, 270, 16719–16723.
639. P. K. Mehta and P. Christen, Adv. Enzymol. Relat. Areas Mol. Biol., 2000, 74, 129–184.
640. W. L. Kelly and C. A. Townsend, J. Am. Chem. Soc., 2002, 124, 8186–8187.
641. I. Molnar, J. F. Aparicio, S. F. Haydock, L. E. Khaw, T. Schwecke, A. Konig, J. Staunton and P. F. Leadlay, Gene, 1996, 169, 1–7.
642. T. L. Poulos, B. C. Finzel and A. J. Howard, J. Mol. Biol., 1987, 195, 687–700.
643. W. W. Stewart, Nature, 1971, 229, 174–178.
644. D. L. Lee and H. Rapoport, J. Org. Chem., 1975, 40, 3491–3495.
645. P. A. Taylor, H. K. Schnoes and R. D. Durbin, Biochim. Biophys. Acta, Gen. Subj., 1972, 286, 107–117.
646. D. Woolley, R. Pringle and A. Braun, J. Biol. Chem., 1952, 197, 409–426.
647. F. Wolf and A. Foster, Science, 1917, 46, 361–363.
648. C. Levi and R. D. Durbin, Physiol. Mol. Plant Pathol., 1986, 28, 345–352.
649. M. D. Thomas, P. J. Langston-Unkefer, T. F. Uchytil and R. D. Durbin, Plant Physiol., 1983, 71, 912–915.
650. J. G. Turner, Physiol. Plant Pathol., 1981, 19, 57–67.
651. J. Turner and J. Debbage, Physiol. Plant Pathol., 1982, 20, 223–456.
652. P. L. Langston-Unkefer, P. A. Macy and R. D. Durbin, Plant Physiol., 1984, 76, 71–74.
653. P. J. Langston-Unkefer, A. C. Robinson, T. J. Knight and R. D. Durbin, J. Biol. Chem., 1987, 262, 1608–1613.
654. R. D. Durbin and T. F. Uchytil, Experientia, 1985, 41, 136–137.
655. T. J. Knight, R. D. Durbin and P. J. Langston-Unkefer, J. Bacteriol., 1986, 166, 224–229.
656. T. J. Knight, R. D. Durbin and P. J. Langston-Unkefer, J. Bacteriol., 1987, 169, 1954–1959.
657. R. H. Coleman, J. Shaffer and H. True, Curr. Microbiol., 1996, 32, 147–150.
658. J. Liu, Y. Le, B. Ye, Y. Zhen, C. Zhu, J. Shen and R. Zhang, Protein Expression Purif., 2002, 24, 439–444.
659. R. Batchvarova, V. Nikolaeva, S. Slavov, S. Bossolova, V. Valkov, S. Atanassova, S. Guelemerov, A. Atanassov and H. Anzai, Theor. Appl. Genet., 1998, 97, 986–989.
660. T. G. Kinscherf and D. K. Willis, J. Antibiot., 2005, 58, 817–821.
661. B. Müller, A. Hädener and C. Tamm, Helv. Chim. Acta, 1987, 70, 412–422.
662. H. Wehrli, Helv. Chim. Acta, 1980, 63, 1915–1919.
663. C. J. Unkefer, R. E. London, R. D. Durbin, T. F. Uchytil and P. J. Langston-Unkefer, J. Biol. Chem., 1987, 262, 4994–4999.
664. P. Roth, A. Hädener and C. Tamm, Helv. Chim. Acta, 1990, 73, 476–482.
665. K. Engst and P. D. Shaw, Mol. Plant-Microbe Interact., 1992, 5, 322–329.
666. L. Liu and P. D. Shaw, J. Bacteriol., 1997, 179, 507–513.
667. L. Liu and P. D. Shaw, J. Bacteriol., 1997, 179, 5922–5927.
668. G. J. Feistner, T. F. Uchytil, K. K. Knoche and R. D. Durbin, J. Org. Chem., 1991, 56, 2922–2925.
669. T. M. Barta, T. G. Kinscherf, T. F. Uchytil and D. K. Willis, Appl. Environ. Microbiol., 1993, 59, 458–466.
670. E. M. Hrabak and D. K. Willis, J. Bacteriol., 1992, 174, 3011–3020.
671. T. M. Barta, T. G. Kinscherf and D. K. Willis, J. Bacteriol., 1992, 174, 3021–3029.
672. J. J. Rich, S. S. Hirano and D. K. Willis, Appl. Environ. Microbiol., 1992, 58, 1440–1446.
673. J. Spizek, J. Novotna, T. Rezanka and A. L. Demain, J. Ind. Microbiol. Biotechnol., 2010, 37, 1241–1248.
674. V. Miao and J. Davies, in Uncultivated Microorganisms, ed. S. S. Epstein, SpringerBerlin/Heidelberg, 2009, vol. 10, pp. 161–180.
675. C. Schmeisser, H. Steele and W. R. Streit, Appl. Microbiol. Biotechnol., 2007, 75, 955–962.
676. R. H. Cichewicz, Nat. Prod. Rep., 2010, 27, 11–22.
677. R. B. Williams, J. C. Henrikson, A. R. Hoover, A. E. Lee and R. H. Cichewicz, Org. Biomol. Chem., 2008, 6, 1895–1897.
678. A. A. Brakhage and V. Schroeckh, Fungal Genet. Biol., 2011, 48, 15–22.
679. P. S. Padayatti, M. S. Helfand, M. A. Totir, M. P. Carey, P. R. Carey, R. A. Bonomo and F. van den Akker, J. Biol. Chem., 2005, 280, 34900–34907.
680. M. A. Totir, M. S. Helfand, M. P. Carey, A. Sheri, J. D. Buynak, R. A. Bonomo and P. R. Carey, Biochemistry, 2007, 46, 8980–8987.
681. R. Li, J.-M. Liao, C.-R. Gu, Y.-T. Wang and C.-L. Chen, J. Phys. Chem. B, 2011, 115, 10298–10310.
682. L. W. Tremblay, J.-E. Hugonnet and J. S. Blanchard, Biochemistry, 2008, 47, 5312–5316.
683. S. Han, R. P. Zaniewski, E. S. Marr, B. M. Lacey, A. P. Tomaras, A. Evdokimov, J. R. Miller and V. Shanmugasundaram, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 22002–22007.
684. R. P. A. Brown, R. T. Aplin and C. J. Schofield, Biochemistry, 1996, 35, 12421–12432.
685. J.-E. Hugonnet, L. W. Tremblay, H. I. Boshoff, C. E. Barry, 3rd and J. S. Blanchard, Science, 2009, 323, 1215–1218.
686. N. O. Concha, B. A. Rasmussen, K. Bush and O. Herzberg, Structure, 1996, 4, 823–836.
687. M. I. Page and A. Badarau, Bioinorg. Chem. Appl., 2008, 576297.
688. J. D. Williams, Clin. Infect. Dis., 1997, 24, 494–497.
689. C. M. Perry and A. Markham, Drugs, 1999, 57, 805–843.
690. H. Troonen, P. Roelants and B. Boon, J. Antibiot., 1976, 29, 1258–1267.
691. F. Robinson, Analyst, 1947, 72, 274–276.
692. K. Kitano, K. Kintaka, S. Suzuki, K. Katamoto, K. Nara and Y. Nakao, J. Ferment. Technol., 1976, 54, 705–711.
693. K. Kitano, Y. Fujisawa, K. Katamoto, K. Nara and Y. Nakao, J. Ferment. Technol., 1976, 54, 712–719.
694. J. Shoji, R. Sakazaki, K. Matsumoto, T. Tanimoto, Y. Terui, S. Kozuki and E. Kondo, J. Antibiot., 1983, 36, 167–169.
695. K. Tanaka, Y. Kawamura, M. Hattori, E. Kondo, K. Matsumoto, A. Terui, J. Shoji, Japan Pat., JP55143996, 1980.
696. H. Aoki and M. Okuhara, Annu. Rev. Microbiol., 1980, 34, 159–181.
697. H. Fukase, T. Hasegawa, K. Hatano, H. Iwasaki and M. Yoneda, J. Antibiot., 1976, 29, 113–120.
698. A. A. Brakhage, P. Spröte, Q. Al-Abdallah, A. Gehrke, H. Plattner and A. Tüncher, in Molecular Biotechnolgy of Fungal beta-Lactam Antibiotics and Related Peptide Synthetases, ed. A. A. Brakhage, SpringerBerlin / Heidelberg, 2004, vol. 88, pp. 45–90.
699. P. Liras and J. F. Martin, Int. Microbiol., 2006, 9, 9–19.
700. D. H. Nam, S. K. Lim, M. H. Chung, E. S. Lee, Y. S. Sohn and D. D. Y. Ryu, J. Microbiol. Biotechnol., 2001, 11, 153–159.
701. R. Ullán, G. Liu, J. Casqueiro, S. Gutiérrez, O. Bañuelos and J. Martín, Mol. Genet. Genomics, 2002, 267, 673–683.
702. A. Paradkar, K. Aidoo, A. Wong and S. Jensen, J. Bacteriol., 1996, 178, 6266–6274.
703. M. B. Tobin, S. Kovacevic, K. Madduri, J. A. Hoskins, P. L. Skatrud, L. C. Vining, C. Stuttard and J. R. Miller, J. Bacteriol., 1991, 173, 6223–6229.
704. D. C. Alexander and S. E. Jensen, J. Bacteriol., 1998, 180, 4068–4079.
705. J. F. Martin, J. Casqueiro and P. Liras, Curr. Opin. Microbiol., 2005, 8, 282–293.
706. M. J. Park, J. O. Yon, S. K. Lim, D. D. Y. Ryu and D. H. Nam, J. Microbiol. Biotechnol., 2004, 14, 635–638.
707. Y. Zhang and E. Oldfield, J. Am. Chem. Soc., 2004, 126, 9494–9495.
708. W. Ge, I. J. Clifton, A. R. Howard-Jones, J. E. Stok, R. M. Adlington, J. E. Baldwin and P. J. Rutledge, ChemBioChem, 2009, 10, 2025–2031.
709. W. Ge, I. J. Clifton, J. E. Stok, R. M. Adlington, J. E. Baldwin and P. J. Rutledge, J. Am. Chem. Soc., 2008, 130, 10096–10102.
710. K. Turgay, M. Krause and M. A. Marahiel, Mol. Microbiol., 1992, 6, 529–546.
711. W. Schmitz, R. Fingerhut and E. Conzelmann, Eur. J. Biochem., 1994, 222, 313–323.
712. W. Schmitz, C. Albers, R. Fingerhut and E. Conzelmann, Eur. J. Biochem., 1995, 231, 815–822.
713. R. D. Knihinicki, R. O. Day and K. M. Williams, Biochem. Pharmacol., 1991, 42, 1905–1911.
714. M. D. Lloyd, D. J. Darley, A. S. Wierzbicki and M. D. Threadgill, FEBS J., 2008, 275, 1089–1102.
715. J. Piel, C. Hertweck, P. R. Shipley, D. M. Hunt, M. S. Newman and B. S. Moore, Chem. Biol., 2000, 7, 943–955.
716. L. Xiang and B. S. Moore, J. Bacteriol., 2005, 187, 4286–4289.
717. M. Berner, D. Krug, C. Bihlmaier, A. Vente, R. Muller and A. Bechthold, J. Bacteriol., 2006, 188, 2666–2673.
718. D. Brown, J. R. Evans and R. A. Fletton, J. Chem. Soc., Chem. Commun., 1979, 282–283.
719. K. Ishida, T. V. Hung, K. Liou, H. C. Lee, C. H. Shin and J. K. Sohng, Biotechnol. Lett., 2006, 28, 409–417.
720. M. L. Raber, A. Castillo, A. Greer and C. A. Townsend, ChemBioChem, 2009, 10, 2904–2912.
721. M. L. Raber, S. O. Arnett and C. A. Townsend, Biochemistry, 2009, 48, 4959–4971.
722. T. Borowski, S. de Marothy, E. Broclawik, C. J. Schofield and P. E. M. Siegbahn, Biochemistry, 2007, 46, 3682–3691.
723. M. D. Andrews, G. A. Brown, J. P. H. Charmant, T. M. Peakman, A. Rebello, K. E. Walsh, T. Gallagher and N. J. Hales, Chem. Commun., 1999, 249–250.
724. F. Marc, P. Weigel, C. Legrain, Y. Almeras, M. Santrot, N. Glansdorff and V. Sakanyan, Eur. J. Biochem., 2000, 267, 5217–5226.
725. K. E. Wilson, in Journal of Chromatography Library, ed. H. W. Gerald and C. Raymond, Elsevier, 1989, vol. 43, pp. 293–346.
726. D. Rosi, M. L. Drozd, M. F. Kuhrt, L. Terminiello, P. E. Came and S. Daum, J. Antibiot., 1981, 34, 341–343.
727. K. E. Wilson, A. J. Kempf, J. M. Liesch and B. H. Arison, J. Antibiot., 1983, 36, 1109–1117.
728. J. S. Kahan, U. S. Pat., 4,162,323, 1979.
729. K. Tanaka, N. Tsuiji, E. Kondo and Y. Kawamura, Eur. Pat. Appl., 1981, 38, 534.
730. P. J. Cassidy, G. Albers-Schonberg, R. T. Goegelman, T. Miller, B. Arison, E. O. Stapley and J. Birnbaum, J. Antibiot., 1981, 34, 637–648.
731. S. J. Box, J. D. Hood and S. R. Spear, J. Antibiot., 1979, 32, 1239–1247.
732. A. G. Brown, D. F. Corbett, A. J. Eglington and T. T. Howarth, J. Antibiot., 1979, 32, 961–963.
733. J. D. Hood, S. J. Box and M. S. Verrall, J. Antibiot., 1979, 32, 295–304.
734. A. G. Brown, D. F. Corbett, A. J. Eglington and T. T. Howarth, J. Chem. Soc., Chem. Commun., 1977, 523–525.
735. D. F. Corbett, A. J. Eglington and T. T. Howarth, J. Chem. Soc., Chem. Commun., 1977, 953–954.
736. S. J. Box, D. F. Corbett, K. G. Robins, S. R. Spear and M. S. Verrall, J. Antibiot., 1982, 35, 1394–1396.
737. K. Okamura, S. Hirata and A. Koki, J. Antibiot., 1979, 32, 262–271.
738. N. Shibamoto, A. Koki, M. Nishino, K. Nakamura, K. Kiyoshima, K. Okamura, M. Okabe, R. Okamoto, Y. Fukagawa and Y. Shimauchi, J. Antibiot., 1980, 33, 1128–1137.
739. N. Shibamoto, M. Nishino, K. Okamura, Y. Fukagawa and T. Ishikura, J. Antibiot., 1982, 35, 763–765.
740. M. Nakayama, A. Iwasaki, S. Kimura, T. Mizoguchi, S. Tanabe, A. Murakami, I. Watanabe, M. Okuchi, H. Itoh, Y. Saino, F. Kobayashi and T. Mori, J. Antibiot., 1980, 33, 1388–1390.
741. M. Nakayama, S. Kimura, S. Tanabe, T. Mizoguchi, I. Watanabe, T. Mori, K. Miyahara and T. Kawasaki, J. Antibiot., 1981, 34, 818–823.
742. M. Nakayama, S. Kimura, T. Mizoguchi, S. Tanabe, A. Iwasaki, A. Murakami, M. Okuchi, H. Itoh and T. Mori, J. Antibiot., 1983, 36, 943–949.
743. M. Okuchi, M. Nakayama, A. Iwasaki, S. Kimura, T. Mizoguchi, S. Tanabe, A. Murakami, H. Ito, T. Mori, U. S. Pat., 4468350, 1984.
744. S. Harada, S. Shinagawa, Y. Nozaki, M. Asai and T. Kishi, J. Antibiot., 1980, 33, 1425–1430.
745. Y. Kawamura, Y. Yasuda, M. Mayama and K. Tanaka, J. Antibiot., 1982, 35, 10–14.
746. J. Shoji, H. Hinoo, R. Sakazaki, N. Tsuji, K. Nagashima, K. Matsumoto, Y. Takahashi, S. Kozuki, T. Hattori, E. Kondo and K. Tanaka, J. Antibiot., 1982, 35, 15–23.
747. S. Tanabe, M. Okuchi, M. Nakayama, S. Kimura, A. Iwasaki, T. Mizoguchi, A. Murakami, H. Itoh and T. Mori, J. Antibiot., 1982, 35, 1237–1239.
748. M. Okabe, S. Azuma, I. Kojima, K. Kouno, R. Okamoto, Y. Fukagawa and T. Ishikura, J. Antibiot., 1982, 35, 1255–1263.
749. T. Yoshioka, I. Kojima, K. Isshiki, A. Watanabe, Y. Shimauchi, M. Okabe, Y. Fukagawa and T. Ishikura, J. Antibiot., 1983, 36, 1473–1482.
750. T. Yoshioka, A. Watanabe, I. Kojima, Y. Shimauchi, M. Okabe, Y. Fukagawa and T. Ishikura, J. Antibiot., 1984, 37, 211–217.
751. N. Tsuji, K. Nagashima, M. Kobayashi, Y. Terui, K. Matsumoto and E. Kondo, J. Antibiot., 1982, 35, 536–540.
752. N. Tsuji, M. Kobayashi, Y. Terui, K. Matsumoto, Y. Takahashi and E. Kondo, J. Antibiot., 1985, 38, 270–274.
753. T. Ito, N. Ezaki, K. Ohba, S. Amano, Y. Kondo, H. S. Miyado, T. Shomura, M. Sezaki, T. Niwa, M. Kojima, S. Inouye, Y. Yamada and T. Nida, J. Antibiot., 1982, 35, 533–535.
754. T. Haneishi, M. Nakajima, N. Serizawa, M. Inukai, Y. Takiguchi, M. Arai, S. Satoh, H. Kuwano and C. Tamura, J. Antibiot., 1983, 36, 1581–1584.
755. G. Blanco, Arch. Microbiol., 2012, 194, 549–555.
756. J. D. Thompson, D. G. Higgins and T. J. Gibson, Nucleic Acids Res., 1994, 22, 4673–4680.
757. K. B. Nicholas, H. B. Nicholas and D. W. Deerfield, EMBNEW. NEWS, 1997, 4, 1–4.
758. L. Que, Nat. Struct. Biol., 2000, 7, 182–184.
759. H. Ishibashi, T. Sato and M. Ikeda, Synthesis, 2002, 695–713.
760. D. D. Tanner and P. M. Rahimi, J. Org. Chem., 1979, 44, 1674–1677.
761. M. Asai, K. Haibara, M. Muroi, K. Kintaka and T. Kishi, J. Antibiot., 1981, 34, 621–627.
762. A. Imada, K. Kitano, K. Kintaka, M. Muroi and M. Asai, Nature, 1981, 289, 590–591.
763. R. B. Sykes, C. M. Cimarusti, D. P. Bonner, K. Bush, D. M. Floyd, N. H. Georgopapadakou, W. M. Koster, W. C. Liu, W. L. Parker, P. A. Principe, M. L. Rathnum, W. A. Slusarchyk, W. H. Trejo and J. S. Wells, Nature, 1981, 291, 489–491.
764. W. L. Parker, W. H. Koster, C. M. Cimarusti, D. M. Floyd, W. C. Liu and M. L. Rathnum, J. Antibiot., 1982, 35, 189–195.
765. J. S. Wells, W. H. Trejo, P. A. Principe, K. Bush, N. Georgopapadakou, D. P. Bonner and R. B. Sykes, J. Antibiot., 1982, 35, 184–188.
766. W. L. Parker and M. L. Rathnum, J. Antibiot., 1982, 35, 300–305.
767. J. S. Wells, W. H. Trejo, P. A. Principe, K. Bush, N. Georgopapadakou, D. P. Bonner and R. B. Sykes, J. Antibiot., 1982, 35, 295–299.
768. P. D. Singh, J. H. Johnson, P. C. Ward, J. S. Wells, W. H. Trejo and R. B. Sykes, J. Antibiot., 1983, 36, 1245–1251.
769. R. Cooper, K. Bush, P. A. Principe, W. H. Trejo, J. S. Wells and R. B. Sykes, J. Antibiot., 1983, 36, 1252–1257.
770. T. Kato, H. Hinoo, Y. Terui, J. Nishikawa, Y. Nakagawa, Y. Ikenishi and J. Shoji, J. Antibiot., 1987, 40, 139–144.
771. T. Kato, H. Hinoo, J. Shoji, K. Matsumoto, T. Tanimoto, T. Hattori, K. Hirooka and E. Kondo, J. Antibiot., 1987, 40, 135–138.
772. S. J. Box, A. G. Brown, M. L. Gilpin, M. N. Gwynn and S. R. Spear, J. Antibiot., 1988, 41, 7–12.
773. M. N. Gwynn, S. J. Box, A. G. Brown and M. L. Gilpin, J. Antibiot., 1988, 41, 1–6.
774. M. Hashimoto, T. Komori and T. Kamiya, J. Antibiot., 1976, 29, 890–901.
775. M. Hashimoto, T. A. Komori and T. Kamiya, J. Am. Chem. Soc., 1976, 98, 3023–3025.
776. K. Watanabe, T. Okuda, K. Yokose, T. Furumai and H. B. Maruyama, J. Antibiot., 1983, 36, 321–324.
777. W. D. Celmer, L. H. Huang, M. T. Jefferson, H. Maeda, K. Inoue and R. Shibakawa, US Pat., 4212944, 1980.
778. K. Tomita, H. Tsukiura and H. Kawaguchi, US Pat., 4320199, 1982.
779. J. M. Davidsen and C. A. Townsend, Chem. Biol., 2012, 19, 297–306.

---

Footnote

† Present address: Department of Chemistry, University of Paderborn, Warburger Str. 100, 33 098 Paderborn, Germany.

---