Biocatalytic transamination with near-stoichiometric inexpensive amine donors mediated by bifunctional mono- and di-amine transaminases

James L. Galman a, Iustina Slabu a, Nicholas J. Weise a, Cesar Iglesias b, Fabio Parmeggiani a, Richard C. Lloyd c and Nicholas J. Turner *a
aManchester Institute of Biotechnology (MIB), School of Chemistry, The University of Manchester, Manchester, UK. E-mail: nicholas.turner@manchester.ac.uk
bFacultad de Quimica, Universidad de la República, Montevideo, Uruguay
cDr. Reddy's Laboratories, Chirotech Technology Centre, Cambridge, UK

Received 29th July 2016 , Accepted 12th August 2016

First published on 12th August 2016


The discovery and characterisation of enzymes with both monoamine and diamine transaminase activity is reported, allowing conversion of a wide range of target ketone substrates with just a small excess of amine donor. The diamine co-substrates (putrescine, cadaverine or spermidine) are bio-derived and the enzyme system results in very little waste, making it a greener strategy for the production of valuable amine fine chemicals and pharmaceuticals.


Biocatalytic strategies for the synthesis of high added-value optically pure chiral intermediates provide extraordinary benefits to the pharmaceutical and agrochemical industries, due to the excellent activity, stereo-/regioselectivity and stability of many biocatalysts, often unmatched by their corresponding chemical equivalents.1 There are obvious sustainability benefits to such approaches, including renewable resourcing of recombinant biocatalysts, solvent- and metal-free chemical transformations and ambient temperature/pressure of operation. In spite of this highly cited “greenness”, the concept of any given biocatalytic transformation being effectively environmentally benign has been questioned in several instances.2 Many widely used biotransformation protocols require, for example, expensive cofactor supplementation and/or large excesses of co-substrate to combat poor atom efficiency and unfavourable equilibria.

The asymmetric amination of ketones to enantiomerically enriched amines mediated by ω-amine transaminases (ω-ATAs, E.C. 2.6.1.18) is undoubtedly one of the most studied and exploited biotransformations which has become integrated into modern industrial practice. The broad substrate tolerance of this class of enzymes is one of their most appealing features in synthetic applications. A prominent example of this is the commercial manufacture of sitagliptin – probably one of the most cited success stories of industrial biocatalysis.3

For these reactions, the thermodynamic equilibria are often unfavourable and hence different methods for shifting the equilibrium have been sought to address this long-standing challenge (Scheme 1).4 One of the earliest approaches, still heavily relied on, was the coupling of the enzymatic reaction with a secondary concomitant irreversible reaction (e.g., conversion of pyruvate by-product to L-lactate with lactate dehydrogenase, paired with an oxidoreductase to recycle NAD(P)H, Scheme 1a).5,6 Whilst producing naturally-degradable biochemicals as by-products, this method has the obvious disadvantage of requiring two additional enzymes and an expensive cofactor to achieve high conversion of the starting ketone.


image file: c6gc02102f-s1.tif
Scheme 1 Different approaches to shifting the thermodynamic equilibrium of ω-ATA transaminations.

On the other hand, for industrial scale application, the approach of choice is the use of a sacrificial amine donor in large excess to drive the equilibrium towards the desired product (Scheme 1b). Isopropylamine is widely used for this purpose, because of its effectiveness and availability.7 This approach has proven very cost-efficient in many cases, however, a 100-fold excess (or higher) is required to drive equilibrium. Also, in order to improve the yields, the removal of the volatile acetone waste product by evaporation or stripping needs to be implemented using specialised temperature platforms and heat-stable enzymes.8

More recently, the use of synthetic diamines such as o-xylylenediamine9 and but-2-ene-1,4-diamine10 has been the focal point to displace the equilibria of ω-ATA reactions by spontaneous cyclisation and subsequent ring aromatisation of the aminocarbonyl by-product (Scheme 1c). However, these donors are usually expensive, highly toxic and often form difficult to remove polymers following aromatisation, adding to downstream processing costs.

In contrast, biogenic terminal diamines present themselves as renewably-sourced alternative amino donors. In addition to their simple bio-based production and relative low cost, upon transamination they are converted into reactive amino aldehydes which spontaneously convert to cyclic imines, thus driving the equilibrium towards amination of ketones. Enzymes active on such compounds, known as α,ω-diamine transaminases (α,ω-DTAs), have exploited these advantages for the synthesis of N-heterocycles precursors.11 However, these enzymes were found to have strict preference for pyruvate over other ketones, making them unsuitable for the synthesis of a broad range of amines. Nevertheless, if the substrate specificity could be addressed, biogenic diamines would offer a distinct approach to equilibrium issues (Scheme 1d), intrinsically “greener” than previous ones, since near-stoichiometric loadings of sacrificial diamines could be applied, in principle.

Herein, we report the discovery and characterisation of a panel of bifunctional α,ω-DTA/ω-ATA enzymes that readily accept cheap and easily-accessible mono-/diamine donors as well as possessing broad ketone acceptor scope, allowing an equilibrium shift to reach theoretical yield of 100% aminated product with almost stoichiometric donor loadings.

Results and discussion

The lack of activity of commonly used ω-ATAs towards simple aliphatic diamines such as putrescine (1,4-diaminobutane, 1a), cadaverine (1,5-diaminopentane, 1b) and spermidine (1,8-diamino-4-azaoctane, 1c) has been previously reported.10,11 Indeed, the only relevant example is a single commercial transaminase (ATA256) recently found through extensive screening of a large biocatalyst panel from Codexis, highlighting the rarity of such activity.12 This transaminase was shown to have activity with a narrow range of ketone acceptors, and its commercial nature precludes evolutionary, sequence and structural insight into its substrate scope or optimisation through engineering. To address this gap in the transaminase toolbox, we undertook a search for new candidate enzymes able to accept diamine donors.

As a promising starting point for the development of an enzyme with broader substrate range we identified the putrescine transaminase gene spuC, part of the polyamine uptake and utilization pathway in Pseudomonas aeruginosa PAO1.13 The spu (spermidine utilisation) operon, consisting of 9 genes, was previously characterised as responsible for spermidine transport and the catabolic route of putrescine in the arginine decarboxylase pathway. Genes with analogous functions are found in E. coli such as the complementary ygjG gene encoding for putrescine transaminase that suggests a common pathway between the two microorganisms.14

To probe the evolutionary relationships between known putrescine transaminases and well-characterised ω-ATAs, a multiple sequence alignment was performed, from which a cladogram could be constructed to infer the order of divergence events (Fig. 1). Interestingly, it was found that sequences with reported diamine transaminase activity clustered within the ω-ATA sequences as polyphyletic groups. SpuC from P. aeruginosa (Pa-SpuC) was revealed to have high sequence identity with the well-studied Cv-ATA from Chromobacterium violaceum (60%)15,16 and only 28% identity with the more functionally similar E. coli YgjG protein.14 This indicated two separate acquisitions of such activity in this family, with one convergence (Pa-SpuC) occurring relatively recently in evolutionary history compared to the other. As previous reports with the more ancient putrescine transaminases (Ec-, Bme- and Bmy-YgjG) revealed minimal acceptance of monoamine substrates,11 it is probable that any ancestral ω-ATA function of these enzyme has been lost. Due to the higher sequence similarity between the Pa-SpuC and characterised ω-ATA biocatalysts, this enzyme was chosen for investigation of potential latent ω-ATA side-activity. Previous experiments carried out with Pa-SpuC revealed that, although it was likely to be involved in the catabolism of diamines,13 minimal activity was observed with putrescine and derived amides.17 In light of this, a selection of predicted orthologues were mined from publically-available sequence data, cloned and expressed to access the natural diversity of SpuC enzymes with respect to activity and potential monoamine/diamine acceptance.


image file: c6gc02102f-f1.tif
Fig. 1 Cladogram showing evolutionary relationships between various wild-type transaminase enzymes as inferred from amino acid sequence identity.

A Basic Local Alignment Search Tool (BLAST) analysis of the characterised spuC gene from the pathogenic P. aeruginosa PAO1 strain revealed other homologous genes from Pseudomonas species, in particular P. putida (Pp-spuC), P. chlororaphis subsp. aureofaciens (Pc-spuC) and P. fluorescens (Pf-spuC) all with an aligned protein sequence identity of 69% (UNIPROT). The three putative spuC genes were cloned from our in-house NCIMB culture collection and subcloned into a pET-28b vector followed by overexpression in E. coli BL21(DE3) cells. All three enzymes were successfully overproduced and purified as recombinant N-terminal His6-tagged protein for screening against a representative panel of diamine and monoamine donors.

In order to determine the activity of the SpuC enzymes, an adapted version of the previously described L-amino acid dehydrogenase colorimetric assay18 was employed, using pyruvate as the amino acceptor, L-alanine dehydrogenase from Bacillus megaterium to regenerate pyruvate with concomitant reduction of NAD+, and the phenazine–tetrazolium system (MPMS + INT) for colour development (Scheme 2). The results for amines 1a–f are summarised in Table 1 (additional mono- and diamines were tested, and kinetic constants were determined, see ESI).


image file: c6gc02102f-s2.tif
Scheme 2 Activity assay used for the characterisation of ω-ATAs and α,ω-DTAs.
Table 1 Specific activities of the novel SpuC enzymes against amine donors 1a–f, compared with Ec-YgjG and Cv-ATA
  Specific activitya (U mg−1)
Subs. Pp-SpuC Pf-SpuC Pc-SpuC Ec-YgjG Cv-ATA
Expt. Cond.: 5 mM 1a–1f, 1 mM 2, 1 mM PLP, 0.2 mg mL−1 purified ATA, 100 mM sodium phosphate buffer, pH 8.0, 30 °C.a One unit of enzyme activity was defined as the amount of enzyme that produced 1 μmol of alanine per min at 30 °C. n.d.: not determined (too low activity).
1a 1.3 0.4 1.4 2.7 n.d.
1b 27.3 16.5 15.1 2.7 7.7
1c 15.7 16.5 21.5 2.1 n.d.
1d 23.2 25.7 30.4 n.d. 28.9
1e 20.4 18.5 18.7 0.05 5.6
1f 16.5 19.3 13.5 n.d. 14.1


Our initial findings showed that all the recombinant SpuC proteins displayed high activity with the biogenic diamine donors 1a–1c of interest with comparable specific activities within the different Pseudomonas species. Despite a 2-fold increase in activity of Ec-YgjG with putrescine 1a, Pp-SpuC exhibited a 10-fold greater activity with cadaverine 1b, Pc-SpuC with spermidine 1c, with little or no activity displayed with Cv-ATA. Surprisingly, the SpuC enzymes also gave promiscuous activity with industrially relevant monoamines (S)-methylbenzylamine 1d, isopropylamine 1e and butylamine 1f, which have been extensively studied with ω-ATAs, but none has been reported to have high activity with biogenic amines. Further studies on the diaminoalkane series were in close agreement with Cv-ATA as the carbon backbone increased in size (>C6), in contrast to the diminished activity presented in Ec-YgjG (see ESI). This pattern is in part due to the narrow hydrophobic channel approaching the enzyme active site elucidated from recent crystal structures.19

Encouraged by these preliminary studies, we turned to the investigation of non-keto acid prochiral acceptors for preparative applications. As a model keto acceptor for reaction optimisation we chose o-fluoroacetophenone 4a, since the chiral amine (S)-6a produced after transamination can be coupled to rhodanine scaffolds resulting in heterocycle analogues (e.g., 7) with a range of pharmacological activities such as the clinical treatment of type II diabetes mellitus.20,21 Biotransformations were conducted using our SpuC enzymes using different amounts of the diamine donors 1a–c. The best conversions (Table 2) were obtained with Pp-SpuC (that also showed the highest specific activity for 1a–b before), although the values for Pf-SpuC and Pc-SpuC were very similar.

Table 2 Testing of different diamine loadings in the transamination of 4a mediated by Pp-SpuC

image file: c6gc02102f-u1.tif

Diamine, equiv. 1a 1b 1c
C (%) eea (%) C (%) eea (%) C (%) eea (%)
Expt. Cond.: 5 mM 4a, 1 mM PLP, 1% v/v DMSO, 2 mg mL−1 purified SpuC, 50 mM HEPES buffer, pH 9.0, 40 °C, 250 rpm, 24 h.a Measured by HPLC on a chiral stationary phase. C = conversion. ee = enantiomeric excess.
1.5 84 >99 95 >99 83 >99
5 91 >99 96 >99 78 >99
10 87 >99 95 >99 73 >99
25 77 >99 83 >99 63 86


Remarkably, high conversions were observed at near-stoichiometric amounts of diamine donors 1a–1c, indicating an efficient displacement of the equilibrium via in situ cyclisation, providing a greener substitute than 50 equivalents of volatile isopropylamine. The transamination of 1a–b gave ω-aminoaldehydes which underwent spontaneous cyclisation to 1-pyrroline 5a and 1-piperideine 5b, respectively (as confirmed by GC-MS). Biotransformations were conducted at pH 9.0 and it was previously shown via1H NMR that the imine monomers are stable at conditions between pH 7–13.22 Amino donor 1c, instead, afforded the corresponding amino-imine that spontaneously cyclises a second time to yield the fused bicyclic structure 5c under biotransformation conditions. Interestingly, increasing the concentration of 1c not only did not improve the conversion to 6a, it was even detrimental, possibly due to inhibition by the N-heterocycle side-product. The optical purity of the product was almost invariably >99% with all three enzymes tested.

In order to test the feasibility of larger scale processes we also performed a preparative biotransformation with 4a and 2 equivalents of 1b, purifying the product by column chromatography after extraction. The procedure afforded (S)-6a with an isolated yield of 79% and >99% ee.

Finally, to prove the wide applicability of this procedure, we also tested a panel of prochiral ketones 4a–n that afford synthetically useful chiral amines of interest for the production of fine chemicals and pharmaceuticals (Scheme 3). Even though the monomeric imines 5a and 5b are known to oligomerise in aqueous environment,12 the corresponding products (as well as the bicyclic product 5c) are water soluble and do not interfere with the standard work-up. Therefore, the chiral amine products of the transamination 6a–n can be easily recovered by extraction (together with residual ketone) for HPLC analysis. The highest conversions obtained for each substrate among the three SpuC enzymes are listed in Table 3. The (S)-enantiomer of the product was formed with all substrate/enzyme pairs, most often with perfect enantioselectivity. Only substrates 4d, i, k and n afforded in some cases slightly lower ee values (56–96%), however, at least one of the three SpuC enzymes tested provided >99% ee. This emphasises the importance of screening several different members of the family for each application.


image file: c6gc02102f-s3.tif
Scheme 3 Panel of aromatic and aliphatic ketones tested with SpuC enzymes.
Table 3 Amination of ketones 4a–n with diamine donors 1a–c (only the best conversion obtained among the three SpuC enzymes is shown)
  Diamine 1a Diamine 1b Diamine 1c
Subs. SpuC C (%) eea (%) SpuC C (%) eea (%) SpuC C (%) eea (%)
Expt. Cond.: 5 mM 4a–n, 1.5 equiv. 1a–c, 1 mM PLP, 1% v/v DMSO, 2 mg mL−1 purified SpuC, 50 mM HEPES buffer, pH 9.0, 40 °C, 250 rpm, 24 h.a Determined by HPLC on a chiral stationary phase.b In these cases 5 equiv. of amine donor were used to increase the conversion values. C = conversion. ee = enantiomeric excess.
4a Pf 92 >99 Pc 97 >99 Pc 97 >99
4b Pc 89b >99 Pc 93 >99 Pc 93 >99
4c Pf 75 >99 Pc 91 >99 Pc 96 >99
4d Pf 34b 75 Pc 72b 74 Pp 47 >99
4e Pf 12b >99 Pp 42b >99 Pc 30 >99
4f Pc 35b >99 Pc 62b >99 Pc 74 >99
4g Pp 63b >99 Pp 80b >99 Pc 86 >99
4h Pc 99b >99 Pc 93 >99 Pc 99 >99
4i Pp 12 >99 Pc 35 >99 Pf 14 >99
4j Pc 37b >99 Pc 80b >99 Pc 38 >99
4k Pp 92 >99 Pc 87 >99 Pf 81 >99
4l Pp 79 >99 Pp 96 >99 Pf 91 >99
4m Pp 62b >99 Pp 95 >99 Pp 55 >99
4n Pp 99b 92 Pf 99b >99 Pp 91 95


The sterically hindered ortho-substituted ketone substrates 4a–c afforded high conversions with high enantioselectivities (>99%). This high conversion is presumably due to a stabilising interaction between the amine formed and the halogen atom at the ortho-position, that drives the transamination reaction more readily,23 than solely based on the electrostatic activation of the halogen substituent (as evident from the contrasting low conversion attained for the para-fluoro derivative 4i and the high conversion for the para-nitro derivative 4h). Likewise, a similar stabilising effect of the intramolecular H-bond with the newly formed amine and vicinal oxygen atom of 4k has been suggested,24 that would account for the good conversion (81–92%) of this substrate.

For non-activated ketone substrates (4d–g, j and n) the conversions appeared considerably lower, therefore we tested also a slightly higher loading of amine donors 1a–b (5 equiv.), leading to modest to excellent conversions. Interestingly, however, with spermidine 1c comparable conversions were reliably obtained at near-stoichiometric amounts (1.5 equiv.) compared to the better amino donor 1b (5 equiv.). This effect can be rationalised by the higher stability of the bicyclic amine product 5c.

A few examples of the pharmaceutical relevance of the chiral amine products thus obtained are several N-methyl-D-aspartate (NMDA) glycine-site antagonists used in the treatment neuropathic pain (i.e., Parkinson's disease), synthesised from (S)-6l–m,25 or the analogues of the antiarrythmic mexiletine obtained from (S)-6k.26 All these compounds were obtained in almost quantitative conversion and excellent enantioselectivity (Table 3).

Conclusions

In summary, we report the identification, heterologous production and characterisation of SpuC orthologues, a class of transaminases that uniquely allow the conversion of numerous ketone substrates at the expense of natural diamines. The enzymes were found to afford high conversion of a broad range of substrates with a modest excess of amine donor, giving good to excellent enantiomeric excess values for pharmaceutically relevant chiral amines.

The application of these bifunctional SpuC enzymes effectively combines the appealing substrate breadth of traditional ω-ATAs with the equilibrium shifting potential of diamine donor acceptance by α,ω-DTAs. This shift, requiring just 1.5–5.0 equivalents of bio-derived putrescine or cadaverine, represents a clean, ‘green’ and effective synthetic strategy, as opposed to the widespread use of a 50-fold excess of conventional, often poorly-accepted monoamines. The method also mitigates the need for expensive cofactor supplementation and complex regeneration systems often employed in conjunction.

The approaches developed in this work are particularly relevant to recent advances in metabolic engineering of host strains for enhanced fermentation of putrescine,27 potentially enabling the creation of an integrated microbial cell catalyst for chiral amine synthesis.

Acknowledgements

J. L. G. acknowledges the support of the BIOINTENSE project, financed through the European Union 7th Framework Programme (Grant agreement no. 312148). I. S. acknowledges a CASE award from BBSRC and Dr Reddy's (Grant code BB/K013076/1). N. J. W. was funded by the European Union's 7th Framework program FP7/2007–2013 under grant agreement no. 289646 (KYROBIO). N. J. T thanks the Royal Society for a Wolfson Research Merit Award.

Notes and references

  1. N. J. Turner and E. O'Reilly, Nat. Chem. Biol., 2013, 9, 285–288 CrossRef CAS PubMed.
  2. Y. Ni, D. Holtmann and F. Hollmann, ChemCatChem, 2014, 6, 930–943 CrossRef CAS.
  3. C. K. Savile, J. M. Janey, E. C. Mundorff, J. C. Moore, S. Tam, W. R. Jarvis, J. C. Colbeck, A. Krebber, F. J. Fleitz, J. Brands, P. N. Devine, G. W. Huisman and G. J. Hughes, Science, 2010, 329, 305–309 CrossRef CAS PubMed.
  4. P. Tufvesson, J. Lima-Ramos, J. S. Jensen, N. Al-Haque, W. Neto and J. M. Woodley, Biotechnol. Bioeng., 2011, 108, 1479–1493 CrossRef CAS PubMed.
  5. M. D. Truppo, J. D. Rozzell, J. C. Moore and N. J. Turner, Org. Biomol. Chem., 2009, 7, 395–398 CAS.
  6. V. Prachayasittikul, S. Ljung, C. Isarankura-Na-Ayudhya and L. Bülow, Int. J. Biol. Sci., 2006, 2, 10–16 CrossRef CAS PubMed.
  7. K. E. Cassimjee, C. Branneby, V. Abedi, A. Wells and P. Berglund, Chem. Commun., 2010, 46, 5569–5571 RSC.
  8. G. Matcham, M. Bhatia, W. Lang, C. Lewis, R. Nelson, A. Wang and W. Wu, Chimia, 1999, 53, 584 CAS.
  9. A. P. Green, N. J. Turner and E. O'Reilly, Angew. Chem., Int. Ed., 2014, 53, 10714–10717 CrossRef CAS PubMed.
  10. L. Martínez-Montero, V. Gotor, V. Gotor-Fernández and I. Lavandera, Adv. Synth. Catal., 2016, 358, 1618–1624 CrossRef.
  11. I. Slabu, J. L. Galman, N. J. Weise, R. C. Lloyd and N. J. Turner, ChemCatChem, 2016, 8, 1038–1042 CrossRef CAS.
  12. A. Gomm, W. Lewis, A. P. Green and E. O'Reilly, Chem. – Eur. J., 2016 DOI:10.1002/chem.201603188.
  13. C.-D. Lu, Y. Itoh, Y. Nakada and Y. Jiang, J. Bacteriol., 2002, 184, 3765–3773 CrossRef CAS PubMed.
  14. N. N. Samsonova, S. V. Smirnov, A. E. Novikova and L. R. Ptitsyn, FEBS Lett., 2005, 579, 4107–4112 CrossRef CAS PubMed.
  15. U. Kaulmann, K. Smithies, M. E. B. Smith, H. C. Hailes and J. M. Ward, Enzyme Microb. Technol., 2007, 41, 628–637 CrossRef CAS.
  16. M. E. B. Smith, B. H. Chen, E. G. Hibbert, U. Kaulmann, K. Smithies, J. L. Galman, F. Baganz, P. A. Dalby, H. C. Hailes, G. J. Lye, J. M. Ward, J. M. Woodley and M. Micheletti, Org. Process Res. Dev., 2010, 14, 99–107 CrossRef CAS.
  17. F. Steffen-Munsberg, C. Vickers, H. Kohls, H. Land, H. Mallin, A. Nobili, L. Skalden, T. van den Bergh, H. J. Joosten, P. Berglund, M. Höhne and U. T. Bornscheuer, Biotechnol. Adv., 2015, 33, 566–604 CrossRef CAS PubMed.
  18. M. Bommer and J. M. Ward, Enzyme Microb. Technol., 2013, 52, 218–225 CrossRef CAS PubMed.
  19. H. J. Cha, J. Jeong, C. Rojviriya and Y. Kim, PLoS One, 2014, 9, e113212 Search PubMed.
  20. D. J. S. Jean, C. Yuan, E. a. Bercot, R. Cupples, M. Chen, J. Fretland, C. Hale, R. W. Hungate, R. Komorowski, M. Veniant, M. Wang, X. Zhang and C. Fotsch, J. Med. Chem., 2007, 50, 429–432 CrossRef PubMed.
  21. V. R. Solomon and H. Lee, Curr. Med. Chem., 2011, 18, 1488–1508 CrossRef CAS PubMed.
  22. C. Christophersen and C. Struve, Heterocycles, 2003, 60, 1907 CrossRef.
  23. C. E. Paul, M. Rodríguez-Mata, E. Busto, I. Lavandera, V. Gotor-Fernández, V. Gotor, S. García-Cerrada, J. Mendiola, Ó. de Frutos and I. Collado, Org. Process Res. Dev., 2014, 18, 788–792 CrossRef CAS.
  24. F. G. Mutti, C. S. Fuchs, D. Pressnitz, J. H. Sattler and W. Kroutil, Adv. Synth. Catal., 2011, 353, 3227–3233 CrossRef CAS.
  25. D. G. Brown, R. A. Urbanek, T. M. Bare, F. M. McLaren, C. L. Horchler, M. Murphy, G. B. Steelman, J. R. Empfield, J. M. Forst, K. J. Herzog, W. Xiao, M. C. Dyroff, C. M. C. Lee, S. Trivedi, K. L. Neilson and R. A. Keith, Bioorg. Med. Chem. Lett., 2003, 13, 3553–3556 CrossRef CAS PubMed.
  26. A. Carrieri, M. Muraglia, F. Corbo and C. Pacifico, Eur. J. Med. Chem., 2009, 44, 1477–1485 CrossRef CAS PubMed.
  27. Z. G. Qian, X. X. Xia and S. Y. Lee, Biotechnol. Bioeng., 2009, 104, 651–662 CAS.

Footnotes

This manuscript is dedicated to Professor Romas Kazlauskas on the occasion of his 60th birthday.
Electronic supplementary information (ESI) available: Complete Experimental section, molecular biology protocols, analytical methods, product characterisation. See DOI: 10.1039/c6gc02102f

This journal is © The Royal Society of Chemistry 2017