Total Chemical Synthesis of Lassomycin and Lassomycin-amide

Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains. Accepted Manuscript Organic & Biomolecular Chemistry


Introduction
The pressing issue of antimicrobial resistance continues to drive the search for novel antibiotic scaffolds. 1 Emerging threats such as extensively drug-resistant tuberculosis require the use of innovative approaches in the search for potential leads. 2,3 Lassomycin (1), a natural product belonging to the 'lasso' peptide class, was recently isolated from an Actinomycete (Lentzea kentuckyensis sp.) in a screen of a library of previously uncultured soil bacteria. 4 The lasso peptide was found to possess potent and specific bactericidal activity against mycobacteria, including drug-resistant forms of Mycobacterium tuberculosis. The specificity of activity arises due to the fact that lassomycin targets the ClpC1 ATPase, an essential enzyme in mycobacteria. ClpC1 ATPase is a key component of protein degradation together with the ClpP1P2 proteolytic complex. 5 The previously reported solution-state NMR structure of lassomycin is shown in Fig. 1. The peptide does not adopt the characteristic knot conformation reported for other homologous lasso peptides: here the C-terminus packs tightly against the N-terminal ring instead of passing through the macrolactam. 4 In light of this discovery, and the potential of lasso peptides as novel drug scaffolds in general, [6][7][8] we set out to chemically synthesize the natural product.
Herein, we report the first chemical synthesis of the recently identified lasso-peptide lassomycin (1) and a Cterminal amide derivative, lassomycin-amide (2). Synthetic peptides 1 and 2 were also evaluated for activity against M. tuberculosis. Fig. 1 Amino acid sequence and proposed solution-state NMR structure of lassomycin (1). 4,9 The N-and C-terminus are labelled.

Results and Discussion
Two key structural features that need to be incorporated during a chemical synthesis of lassomycin (1) are the isopeptide bond formed between the N-terminus and the carboxylate side chain of Asp-8, and the methylated Cterminus ( Fig. 1). We envisaged two possible synthetic approaches to access lassomycin (1) and these are outlined in Scheme 1. Route A involved the preparation of a linear peptide and the formation of the lactam ring on resin using an allyl ester orthogonal protecting group strategy for the side chain of Asp-8. The C-terminal methyl ester would be installed using 2-chlorotrityl chloride resin and the approach reported by Turner et al. (Scheme 1, Route A). 10 The alternative approach (Scheme 1, Route B), involved assembling the linear peptide on resin via side chain anchoring. In this approach a di-peptide fragment would be loaded initially onto the resin allowing the C-terminal methyl ester functionality to be installed at this point. Lactam ring formation would be carried out in solution. Initially, Route A (Scheme 1) was investigated and the linear peptide 7 was successfully assembled via standard Fmoc solid phase peptide synthesis (SPPS). Next, we attempted the on-resin deprotection of the Asp-8 side chain. In our hands all attempts to remove the allyl ester and generate peptide 6 were unsuccessful, despite the use of stoichiometric amounts of tetrakis(triphenylphosphine)palladium(0) catalyst and triphenylsilane, and trialling a number of reaction solvents. The lack of success in allyl ester removal was assumed to arise from a difficulty in coordinating the allyl moiety to the catalyst, presumably due to steric hindrance around the metal centre. This effect will depend upon peptide sequence length and folding of the peptide, and as such is likely to be highly sequence dependent. In light of these difficulties alternative orthogonal protecting groups were considered (i.e. highly acid sensitive, 11 photolabile 12 or silicon-based 13 protection), however the requirement for more specialized deprotection conditions, or the lack of available commercial building blocks, led us to investigate Route B (Scheme 1).
The synthesis of the required orthogonally protected dipeptide (10) for Route B was carried out in solution. Removal of the t-butyl side chain from 10 under standard acidic conditions followed by attachment to Rink amide resin gave the methyl ester protected C-terminal fragment of lassomycin (Scheme 2). The peptide chain was then extended on resin using Fmoc SPPS to give a complete linear version of lassomycin (12). Cleavage from the resin afforded the linear peptide (3) which was ready for lactam ring formation.
While peptide macrocyclizations are often best performed under high dilution to minimize unwanted oligomer formation, 14 the 'pseudo-high dilution' conditions described by  Ion mobility (IM) mass spectrometry was also carried out on the synthetic peptide. A number of peaks with distinct collisional cross sections were resolved belonging to either lassomycin or a trace amount of the C-terminal acid hydrolysis product (14) (Supporting Information, SI Figure 7). Only a single peak was observed for each species suggesting that only one conformation is present, as opposed to distinct populations of threaded and unthreaded peptide (or that the collisional cross section for threaded/unthreaded ions is identical). It was also possible to access lassomycin-amide (2) a Cterminal amide version of the original lassomycin (1) peptide. Synthesis of the required linear peptide (15) was carried out on Rink Amide resin using Fmoc SPPS, and cleavage from the resin afforded peptide 16 (Scheme 3). Formation of the lactam ring in solution and HPLC purification to yield lassomycinamide (2) were carried out as previously described for the synthesis of 1. Purity and structure were confirmed by analytical HPLC (Supporting Information, SI Figure 2), MALDI-TOF MS (Supporting Information, SI Figure 2) and tandem mass spectrometry (Supporting Information, SI Figure 1).
Peptides 1 and 2, as well as truncated analogue lassomycin(1-9) (17) representing the peptide without the tail were tested against Mycobacterium tuberculosis H37Rv in MIC assays utilizing both solid-agar based and liquid-media based methods. None of the peptides showed inhibitory activity when tested, up to concentrations of 100 μg/mL. These results were surprising in particular for peptide 1, given the previously reported biological activity profile. 4 Given the unlikely nature of spontaneous threading during the synthesis of 1 (and 2) it is logical to assume that the peptides reported herein are unthreaded. This, in combination with the biological data, leads us to speculate that the un-threaded structure previously reported may not be correct. The structural discrepancy from the original report 4 could possibly be due to instability of lassomycin in its threaded conformation. Several examples have been documented of lasso peptides unthreading upon heat treatment. Caulosegnin I unthreads completely after 4 hours at 95 °C, while caulosegnin III unthreads and decomposes at elevated temperatures. 17 Similarly, astexin-1 unthreads when heated to 50 °C for 4 hours 18 and astexin-2 unthreads and decomposes after heating at 95 °C for 2 hours. 19 In contrast, caulosegnin II and astexin-3 show minimal to no unthreading when heated at 95 °C for prolonged periods of time. 17,19 In order to further elucidate the effect of temperature on the conformational state of lassomycin, structural characterization of synthetic peptide 1 was carried out using two-dimensional NMR. A comparison of 1 H-13 C HSQC spectra for 1 at two different temperatures is shown in Fig. 3. The NMR data gatherered is also compared in Fig. 3 with the naturally isolated labelled [ 13 C, 15 N]lassomycin. 4 No noticeable changes in chemical shift values are observed between the spectrum of 1 at 25 °C and at 40 °C. Furthermore, clear differences in the 1 H-13 C HSQC spectrum can be seen between the synthetic peptide (1) and the naturally isolated lassomycin at 40 o C (Fig. 3). These results strongly suggest that the conformation of 1 does not change appreciably upon heating, and they highlight that there are significant differences between the conformation of synthetically prepared lassomycin (1) and that of naturally isolated lassomycin. In light of this evidence, we propose that synthetic lassomycin (1) either exists in the unconstrained unthreaded form, or as a threaded lasso, but that there is no interconversion between these structures at temperatures upto 40 °C. Given the clear differences observed in the NMR between the synthetic and naturally isolated lassomycin, it seems likely that one peptide is threaded while the other is not. This idea of two distinct conformations being adopted by the synthetic and natural peptides is supported further by the biological data obtained where clear differences in activity are seen.
We propose that i) neither peptide has undergone unthreading (or threading) upon heating during NMR data collection, or ii) one peptide has undergone threading from an initial unthreaded state whilst the other has undergone unthreading. For case ii) the peptide undergoing unthreading cannot be the naturally isolated lassomycin, as this would imply that the synthetic peptide (1) has undergone threading and that it is threaded at the point of NMR data collection, thus paradoxically implying (given the NMR study of temperature dependence already discussed) that a threaded lassomycin cannot unthread. Neither scenario therefore supports the hypothesis that the naturally isolated lassomycin has undergone unthreading upon heating to 40 °C. Furthermore, given the unlikely nature of scenario ii), taken together with the previously dicussed likelihood that synthetic lassomycin is in fact unthreaded, we propose that the previously published naturally isolated lassomycin actually exists as a threaded lasso that does not unthread upon heating to 40 °C.

Conclusions
We have developed a practical synthetic route to an unthreaded version of the lasso peptide lassomycin (1).
Lassomycin-amide (2), a C-terminal amide derivative of 1 was also prepared. The biological evaluation of peptides 1 and 2 against Mycobacterium tuberculosis revealed that neither had any activity. A lack of biological activity and our NMR analysis of synthetic lassomycin (1) suggest that naturally occurring lassomycin exhibits a standard lasso peptide threaded conformation rather than the reported unthreaded structure. Further conformational studies to gain additional insight in this area are ongoing.

Experimental
All reagents were purchased from Sigma-Aldrich unless otherwise specified. Peptide synthesis grade DMF was purchased from AGTC Bioproducts (Hessle, UK) and amino acid derivatives were purchased from CEM, Novabiochem (Merck) or AGTC. PyBOP® was purchased from Apollo Scientific (Stockport, UK). All resins were purchased from Novabiochem. Peptide molecular weight calculations and mass assignments were carried out using Pep-Calc.com. 20

Peptide synthesis
Automated SPPS was carried out at 0.10 mmol scale on a CEM Liberty1 single-channel microwave peptide synthesizer equipped with a Discover microwave unit. All reactions were carried out at room temperature in DMF using 2 × 1 h couplings for all residues except Arg, which was triple coupled (3 × 1 h). Fmoc-protected amino acids were used (5 equiv), with PyBOP® (5 equiv) as the activator in the presence of DIPEA (10 equiv, 2 M solution in NMP). Amino acid side chain functionality was protected as follows: Fmoc-His(Trt)-OH, Fmoc-Ser(OtBu)-OH and Fmoc-Thr(OtBu)-OH. The Fmoc group was removed by two successive treatments with 20% (v/v) piperidine solution in DMF (5 + 10 min). An additional 10 min deprotection was carried out for the penultimate residue (Leu-2). Bubbling with nitrogen gas was used to ensure efficient agitation of the reaction mixture during each step. Preswelling of dry resin was carried out in DMF for a minimum of 1 h.

Cleavage from solid support
Peptide-resin was shrunk in diethyl ether and treated with 2.85 mL TFA, 0.15 mL deionized water and 0.15 mL TIPS for 4 h at room temperature. The resin was then removed by filtration and the filtrate concentrated in vacuo before precipitation using ether and decanting of the liquid (followed by subsequent ether washes). The resulting solid peptide was dissolved in deionized water containing 0.1% TFA and lyophilized.

Solution-phase cyclization
The cyclization was carried out using the pseudo-high dilution conditions described by Malesevic et al. 15 Fully deprotected peptide (113 μmol) and HATU (339 μmol) were dissolved separately in 11.3 mL each of DMF which had been dried over molecular sieves. Both solutions were added simultaneously using two syringe pumps at a rate of 0.01 mL/min each to a stirred solution of HATU (11.3 μmol) and DIPEA (678 μmol) in dry DMF (11.3 mL) over a period of ~20 h. After the addition was complete, the reaction was stirred for a further 30 min and the solvent was removed under vaccum. The resulting oil was kept at -20 °C until needed for purification.

High-performance liquid chromatography
The crude cyclization reaction product was dissolved in 2.5 mL deionized water/acetonitrile, filtered, centrifuged and injected in 0. 5

Analytical liquid chromatography mass spectrometry (LCMS)
Analytical LCMS was carried out using an Acquity UPLC system (Waters Ltd, UK) equipped with a photodiode array detector. Samples were injected onto an Acquity UPLC BEH C18 column (1.7 μm, 2.1 × 50 mm) and a gradient of 5-95% B (solvent A = H 2 O, 0.1% formic acid; B = MeCN) was run over 3.8 min with a flow rate of 0.6 mL/min. The flow of solvent from the UPLC system was introduced into the electrospray ion source of an Aquity TQD or QToF Premier mass spectrometer, and positive ions were measured.

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS)
MALDI-TOF mass data was collected using an Autoflex II ToF/ToF mass spectrometer (Bruker Daltonik GmBH) equipped with a 337 nm nitrogen laser. Peptides were dissolved in 1:1 deionized water/MeCN for MS analysis. Sample solution (1 mg/mL) was mixed with matrix solution (cyano-4hydroxycinnamic acid, ~50 mg/mL) in a ratio of 1:9, and 1 μL of the resulting solution spotted onto a metal target and placed into the MALDI ion source. MS data was processed using FlexAnalysis 2.0 (Bruker Daltonik GmBH). MALDI MS/MS was performed using LIFT technology, which enables detection of product ions that result from elevating the laser power.

Ion-mobility (IM) mass spectrometry
Ion-mobility mass spectrometry was carried out using a Synapt G2S mass spectrometer equipped with an Acquity UPLC (Waters Ltd, UK). 1 μL of sample was injected onto a BEH C18 column (1.

MIC testing of compounds against mycobacteria
The MICs of the compounds were determined using standard liquid-broth and solid-agar based methods. Briefly, serial dilutions of each compound were made between 100 mg/L and 0.78 mg/L with MB7H9/OADC (BD Biosciences) media in 96-well U-bottom plates. Mycobacterium tuberculosis H37Rv was grown to log phase (OD 600 = 1.0) and 10 4 cells were added to all wells. Solid-agar based method was carried out as previously described 21 using a 48-well plate format. Compounds (see above) were tested in triplicate by adding serial dilutions into separate wells overlaid with MB7H11/OADC (BD Biosciences) agar based media. ).

Lassomycin-amide
(2). Linear peptide GLRRLFADQLVGRRNI-amide was synthesized via automated SPPS at room temperature (as described) on Rink amide AM resin (0.79 mmol/g substitution, 0.10 mmol scale). Resin cleavage, cyclization and purification by high-performance liquid chromatography were carried out as detailed above to afford lassomycin-amide (