In vitro mutasynthesis of lantibiotic analogues containing nonproteinogenic amino acids.

Lantibiotics are ribosomally synthesized and post-translationally modified peptide antibiotics containing the characteristic thioether cross-links lanthionine and methyllanthionine. To date, no analogues of lantibiotics that contain nonproteinogenic amino acids have been reported. In this study, in vitro-reconstituted lacticin 481 synthetase was used in conjunction with synthetic peptide substrates containing nonproteinogenic amino acids to generate 11 analogues of lacticin 481. These analogues contained sarcosine and aminocyclopropanoic acid in place of Gly5, D-valine at position 6, 4-cyanoaminobutyric acid in place of Glu13, beta(3)-homoarginine at the position of Asn15, N-butylglycine and beta-Ala at Met16, naphthylalanine (Nal) at Trp19, 4-pyridynylalanine (Pal) at Phe21, and homophenylalanine (hPhe) at Phe23. Of these analogues, the Trp19Nal and Phe23hPhe mutants provided zones of inhibition larger than the parent compound in agar diffusion assays against the indicator strains Lactococcus lactis HP and Bacillus subtilis 6633. These two compounds also demonstrated improved MIC values against liquid cultures of L. lactis HP.

The lantibiotics are a class of polycyclic peptide natural products that exhibit antimicrobial activity against a variety of clinically relevant bacterial pathogens.(1),(2) These compounds have a long history of use in food preservation and possess considerable promise as human therapeutics because of their potency and unique modes of action.(3) Lantibiotics are ribosomally synthesized as linear prepeptides and are extensively post-translationally modified to their active forms. These modifications introduce thioether-containing cross-linked amino acids called lanthionine (Lan) and methyllanthionine (MeLan) as well as dehydroalanine (Dha) and dehydrobutyrine (Dhb) residues. The latter motifs result from enzymatic dehydration of Ser or Thr residues, respectively. (Me)Lan cross-links are then installed through enzymatic intramolecular Michael-type addition of Cys thiols onto the unsaturated amino acids. Despite their potential as antimicrobial therapeutics, significant hurdles must be overcome for lantibiotics to reach the clinic, including improving the solubility and stability of the compounds under physiological conditions.

The generation of lantibiotic analogues for structure−activity studies has traditionally been conducted through in vivo mutagenesis of lantibiotic prepeptides. Such studies have provided a wealth of knowledge regarding lantibiotic structure−activity relationships but have yielded only a few lantibiotics with enhanced bioactivity(4) or improved therapeutic properties.(5) The chemical synthesis of lantibiotics has been achieved for two family members.[6,7] In principle, synthesis provides access to lantibiotic analogues with expanded chemical diversity by introduction of nonproteinogenic amino acids. At present, however, no such analogues have been reported. An alternative approach is to use biosynthetic enzymes with synthetic substrate analogues, a strategy we term in vitro mutasynthesis (IVM). We report herein the first preparation of lantibiotic analogues containing nonproteinogenic amino acids using IVM. Several of these analogues possess improved antimicrobial activity.

Lacticin 481 is a lantibiotic produced by Lactococcus lactis CNRZ 481 that contains three Me(Lan) rings and one Dhb residue.(8) Lacticin 481 synthetase (LctM) introduces these post-translational modifications by catalyzing both dehydration and cyclization of its substrate prepeptide, LctA.(9) LctA is a 51 amino acid peptide consisting of a C-terminal structural peptide that undergoes post-translational modification as well as an N-terminal leader peptide that is required for efficient processing by LctM.[10,11] Several recent reports have characterized the promiscuous substrate specificity of LctM toward truncated LctA mutant substrates.[12−14] A significant challenge for the introduction of nonproteinogenic amino acids into lacticin 481 is the preparation of full-length LctA prepeptides. For this purpose, we developed the synthesis of a triazole-linked LctA peptide analogue (3) via Cu(I)-catalyzed 1,3-dipolar cycloaddition of an alkyne-functionalized LctA leader peptide (1) and an azide-modified LctA structural region (2).(14) This strategy was utilized here to prepare triazole-linked LctA substrates containing several nonproteinogenic amino acids in the structural peptide, including β-amino acids, d-amino acids, and N-alkylglycine (peptoid) residues. An additional two mutations, Asn15Arg and Phe21His, were included in substrate 3 to improve solubility.[15,16] On the basis of our previous survey of the types of amino acids that are tolerated by LctM,(14) substrates including the following nonproteinogenic amino acid mutations were prepared (Figure 1): sarcosine (Sar) and aminocyclopropanoic acid (Acpc) in place of Gly5, d-valine at position 6, 4-cyanoaminobutyric acid (Cba) in place of Glu13, β3-homoarginine (β3-Arg) at position 15, N-butylglycine (N-Nle) and β-Ala replacing Met16, naphthylalanine (Nal) at Trp19, 4-pyridynylalanine (Pal) at position 21, and homophenylalanine (hPhe) in place of Phe23. A final analogue contained three concomitant replacements of Gly2, Gly3, and Gly5 with β-Ala. The triazole-linked LctA substrate analogues (0.5−1.5 mg) were incubated with 0.5 μM LctM in the presence of 10 mM Mg2+ and 1 mM ATP. Assay progress was monitored by MALDI-TOF MS. Each of the unnatural LctA substrates was dehydrated four times by LctM, as shown for representative examples in Figure 2 (all of the MS data is provided in the Supporting Information). Typically, a small amount of threefold-dehydrated product was also present and readily removed via reversed-phase HPLC. To produce bioactive lacticin 481 analogues, the leader peptide and triazole linker were removed by proteolysis using commercially available endoproteinase LysC, which efficiently cleaved the modified substrates C-terminal to Lys1.(16) Thus, lacticin 481 analogues were produced in high purity (Figure S4 in the Supporting Information). To ensure that complete cyclization had occurred, the analogues were incubated with a thiol-modifying reagent, demonstrating that no free thiols remained in the LctM-treated peptides.

Figure 1

In vitro mutasynthesis of lacticin 481 analogues. Synthetic substrate analogues were prepared using copper-catalyzed [2 + 3] cycloaddition of two fragment peptides. The resulting LctA analogues were treated with LctM, which dehydrated the underlined Ser and Thr residues and incorporated the thioether rings shown. Mutant analogues generated are indicated with arrows. Lan residues are shown in red, and the MeLan cross-link is shown in blue.

Figure 2

Representative MALDI-MS spectra after incubation of triazole-linked LctA substrate analogues (3, blue) with LctM in the presence of ATP and Mg2+ (red). Asterisks represent threefold-dehydrated products.

The bioactivity of the analogues was evaluated by agar diffusion assays against the indicator strain L. lactis HP (Figure 3). The analogues and the parent compound (4) that did not contain any nonproteinogenic amino acids were spotted against L. lactis HP in a 96-well format. For initial characterization, each compound was spotted and the zone of inhibition measured following overnight incubation. Three of the lacticin 481 analogues, the ones containing Trp19Nal, Phe21Pal, and Phe23hPhe mutations, provided zones of inhibition larger than the parent compound 4, as shown in Figure 3. The Gly5Sar, Gly5Acpc, Met16β-Ala, and Asn15β3-Arg mutant lantibiotics displayed a significant decrease in activity (∼16−34 fold; see Figure S6), while the Glu13Cba, Met16N-Nle, and Val6d-Val mutants as well as the analogue with three concomitant β-Ala mutations displayed no bioactivity within the limits of the assay. The compounds containing Trp19Nal, Phe21Pal, and Phe23hPhe mutations were also tested against Bacillus subtilis ATCC 6633 and provided trends in activity analogous to those observed with L. lactis HP (Figure S7). Minimum inhibitory concentration (MIC) values were determined against L. lactis HP for the three variants with improved activity in the agar diffusion assay. The obtained MIC values were 3.12 μM for Trp19Nal and Phe23hPhe, compared with 6.25 μM for the parent 4 and 12.5 μM for Phe21Pal; the improved agar diffusion activity for the latter compound may have been the result of differential diffusion rates.

Figure 3

Antimicrobial assays for biological activity of selected lacticin 481 analogues against L. lactis HP.

The results presented here provide guidelines for further engineering. For example, the two analogues with improved bioactivity each contained nonproteinogenic amino acid substitutions in the C-ring of lacticin 481, suggesting this region as a site for additional mutation. Furthermore, each of these analogues contained non-natural side-chain substituents, whereas analogues with an altered amide backbone (β-amino acids or peptoids) had diminished activity. Finally, a dramatic decrease in activity was noted with all mutants containing substitutions in the N-terminal residues of lacticin 481.

In summary, we have presented the first preparation of lantibiotic analogues containing nonproteinogenic amino acids, suggesting that in vitro engineering of lantibiotics can be a viable method for identifying improved bioactive compounds.