Unusual Class I Lanthipeptides from the Marine Bacteria Thalassomonas viridans

Abstract

A novel class I lanthipeptide produced by the marine bacterium Thalassomonas viridans XOM25^T was identified using genome mining. The putative lanthipeptides were heterologously coexpressed in Escherichia coli as GFP–prepeptide fusions along with the operon-encoded class I lanthipeptide modification machinery VdsCB. The core peptides, VdsA1 and VdsA2, were liberated from GFP using the NisP protease, purified, and analyzed by collision-induced tandem mass spectrometry. The operon-encoded cyclase and dehydratase, VdsCB, exhibited lanthipeptide synthetase activity via post-translational modification of the VdsA1 and VdsA2 core peptides. Modifications were directed by the conserved double glycine leader containing prepeptides of VdsA1 and VdsA2.

The majority of natural products (NPs) produced by microbes have been discovered using top-down screening methods, but there are two major limitations to this approach: novel NPs may not be produced under laboratory culturing conditions, and there is a growing risk of rediscovery of a known NP. Employing genome-guided approaches to identify biosynthetic gene clusters (BGCs) that potentially encode novel NPs before screening assays are performed provides several advantages.¹⁻³ Research efforts can easily be focused on uncharacterized BGCs originating from microorganisms habituating understudied environments instead of using unguided top-down screenings. However, most microorganisms are unculturable, and for those that can be cultured, expression of an interesting BGC often requires stimuli that have not yet been described.¹⁻³ Therefore, the significant quantities required to characterize novel compounds remain a bottleneck.

Improving expression via stimulation of a silent BGC in the native producer can be a challenging task, let alone purifying the NPs from complex growth media. This renders testing for the expression of an NP with unconfirmed or novel bioactivity an impossible task. Therefore, NP discovery using next-generation sequencing information is driving the development of heterologous or culture-independent synthesis approaches. Chemical synthesis is a complementary bottom-up approach used to characterize bioinformatically mined peptides that may have useful bioactivities.^2,3 However, peptide synthesis is not always feasible for novel ribosomally produced and post-translationally modified peptides (RiPPs), as the installation of post-translational modifications is often unpredictable.^2,3

Lanthipeptides are a class of RiPPs characterized by the post-translational installation of thioether-cross-linked bis-amino acids known as lanthionine (Lan) and methyllanthionine (MeLan). Thioether cross-linkages are introduced by operon-encoded modification enzymes that dehydrate serine and threonine to form dehydroalanine (Dha) and dehydrobutyrine (Dhb), respectively. Ring structures are formed by enzymatic cyclization of the linear peptide via Michael addition, which binds the thiol side group from a cysteine residue to a cognate Dha or Dhb residue. Thioether cross-linkages between Dha or Dhb residues and cysteine form Lan and MeLan, respectively. Post-translational processing of the core peptide is directed by an N-terminal leader that is recognized by the operon-encoded modification machinery. Upon secretion, the leader peptide is cleaved from the core peptide, which results in a mature and active core peptide.⁴

Synthetase-mediated ring installation makes the location of lanthionine rings difficult to predict in silico, and furthermore, Lan and MeLan residues exist as stereoisomers.⁵ Lanthionine residues exist as dl-Lan (2S,6R) or ll-Lan (2R,6R) stereoisomers, with the prefixes dl and ll referring to the configurations of the former serine/threonine and cysteine α-carbons, respectively.⁶ Methyllanthionine has an additional chiral center that produces four possible stereochemical configurations. MeLan cross-linkages with the dl-MeLan (2S,3S,6R) and ll-MeLan (2R,3R,6R) conformations are commonly reported, while the d-allo-l-MeLan (2S,3R,6R) conformation has only recently been observed.⁷ The l-allo-l-MeLan (2R,3S,6R) configuration has not yet been reported.⁷ It has been demonstrated that the stereochemical compositions of Lan and MeLan residues may affect bioactivity, as reported for cytolysin.⁵ Therefore, chemical synthesis of a previously unobserved and significantly different lanthipeptide would need to consider all possible thioether bond locations as well as the stereoisomer configuration at each Lan and MeLan cross-linkage.

Despite all of the difficulties in isolating novel lanthipeptides, they are still sought because of the broad range of bioactivities already displayed by the group.^4,8 To date, class I lanthipeptides with characterized bioactivities have predominantly been isolated from terrestrial microbial isolates, while the bioactivities of class I lanthipeptides from the marine environment are still largely undescribed. Interest in the lanthipeptide producers from Bacteroidota and Gammaproteobacteria (Gram-negative) phyla is growing and currently includes the first antifungal lanthipeptide, pinensin, from Actinophage pinensis DSM 28390.⁹ More recently, two lanthipeptide operons with unknown peptide bioactivities have been described via heterologous expression in Escherichia coli. Ped15.1 and Ped15.2 originate from Pedobacter lusitanus NL19, while thalassomonasins A and B originate from Thalassomonas actiniarum A5K-106^T. Unlike other lanthipeptide producers, which are generally Gram-positives isolated from terrestrial environments, P. lusitanus NL19 was isolated from deactivated uranium mine sludge, and T. actiniarum A5K-106T was isolated from a sea anemone.

Lanthipeptides are classified into five classes based on amino acid sequences of the lanthionine synthetases instead of the peptide amino acid sequences. Currently, the five classes of lanthipeptides are class I (LanBCs), class II (LanMs), class III (LanKC), class IV (LanL), and class V (LanYK).^10,11 The conserved domains within these lanthionine synthetase classes allow for rapid genome mining and classification of putative operons from otherwise understudied genomes.

Thalassomonas viridans XOM25^T is a recently described marine bacterium within the Gammaproteobacteria.¹² Members of the genus Thalassomonas are rod-shaped, motile, halophilic, strictly aerobic, and chemo-organotrophic and are represented by three species: T. viridans, T. actiniatium, and T. haliotis.¹² AntiSMASH (https://antismash.secondarymetabolites.org) was used to mine the T. viridans XOM25T draft genome in search for novel secondary metabolites. Putative class I lanthipeptide cyclase and dehydratase genes were identified along with three downstream lanthipeptides containing a conserved double-glycine leader peptide (Figure 1a and b). Similarly, for the Ped15 and thallasomonasin operons, the putative viridisin operon, Vds, contained multiple putative lanthipeptides that had N-terminal double-glycine leaders. These lanthipeptides have not been isolated from the native producers T. viridans XOM25T and T. actiniatium A5K-106(T) or from P. lusitanus NL19.^13,14 These findings have led to a case study demonstrating that genomes from Gram-negative bacteria are a rich source of novel but seemingly cryptic lanthipeptides with largely uncharacterized but clearly important biological roles.¹⁵

Figure 1.

(a) The T. viridans class I lanthipeptide operon consisting of the vdsC cyclase (green), vdsB dehydratase (green), and vdsA1, vdsA2, and vdsA3 lanthipeptides (pink) along with flanking genes (yellow). (b) Amino acid sequences of VdsA1, VdsA2, and VdsA3 precursor peptides. Conserved residues in the leader peptides are highlighted in black, the double-glycine leader in green, threonine residues in orange, serine residues in red, and cysteine residues in blue.

Phylogenetically, the VdsC and VdsB proteins are classified as class I lanthipeptides due to their independent nature and homology (Figures 1a and 2). Within the conserved LanC cyclases domain (CD04793) and characterized cyclases, VdsC displays the highest sequence identity to thallasomonasin (TlnC) and pinensin (PinC) cyclases, followed by PedC cyclase of Ped15.1/2 peptides (Figure 2a). Within the LanB dehydratase Pfams pf04738 and pf14028 and characterized lanthipeptide dehydratases, VdsB is most strongly associated with TlnB and then PedB and is only distantly related to PinB (Figure 2b). The cluster in which VdsC and VdsB are located is significantly different from other lanthipeptide operons from the Bacillota phylum with antimicrobial activity against species like Listeria, Enterococcus, and Clostridium spp. (Figure 2). No significant sequence similarity was observed between the core peptides of VdsA1 and VdsA2 with Ped15.1 and Ped15.2, which has antifungal activity (Figure S1). Pinensin was natively isolated from Chitinophaga pinensis DSM 2588, and although the local PinC cyclase and PinB1/2 dehydratase are thought to post-translationally modify the prepeptide PinA, this has not yet been confirmed. Two additional lanthipeptide dehydratase and cyclase pairs are present in the C. pinensis DSM 2588 genome (Figure 2, red arrows). The additional dehydratase, ACU58933.1, shows a close association with the characterized PedA15, Thalassomonasin and Viridisin dehydratase, unlike the split Pinensin dehydratase PinB1/B2.

Figure 2.

Phylogenetic placement of the VdsC and VdsB amino acid sequences within the known class I lanthipeptide synthetases. Light green branches highlight the VdsC and VdsB proteins. Red is epidermin-like, pink is epicidin-like, purple is nisin-like, dark blue is subtilin-like, light blue is paenicidin-like, cyan is microbisporicin-like, and dark green is pinensin. Accession numbers represent putative lanthipeptide LanC cyclases in the conserved domain CD04793 in (a) and LanB dehydratases in pfam pf04738 (N-terminal) and pfam pf14028 (C-terminal) combined in (b). Solid line branches belong to lanthipeptide modification machinery from Bacillota (Gram-positive hosts), while dotted lines represent lanthipeptide operons from Bacteroidota and Gammaproteobacteria (Gram-negative) hosts. Vertical blue lines represent the phylogenetic location of modification machinery that produces lanthipeptides with antibacterial activity against some other Gram-positive bacteria like Listeria, Enterococcus, and Clostridium spp. Red arrows indicate additional lanthipeptide synthetic machinery (LanC and LanB, respectively) found in C. pinensis DSM 2588.

The putative T. viridans lanthipeptides VdsA1 and VdsA2 were heterologously coexpressed with the VdsCB modification machinery in E. coli as GFP fusion proteins with an N-terminal polyhistidine tag (Figure 3). The double-glycine cleavage sequence and two upstream amino acids in the VdsA1 and VdsA2 prepeptides were substituted with the Ala-Ser-Pro-Arg (ASPR) NisP cleavage site (Figures 1b and S2). Modified prepeptide gene sequences were synthesized and fused to the C-terminus of GFP in the MCS1 of pRSF-Duet (Figures 3 and S2). The vdsCB genes were cloned into MCS2 of pRSF-Duet and transcribed polycistronically (Figure 3). After expression, the His₆-tagged GFP-fused lanthipeptides were extracted under native conditions and purified using IMAC and anion-exchange chromatography along with recombinantly expressed NisP as previously described.^16,17

Figure 3.

(a) Heterologous coexpression of the VdsA-1 and -2 lanthipeptides with the VdsCB modification machinery in E. coli using the pRSF-Duet 1 system. (b) The native configurations of the vdsCB genes were cloned into MCS2 and expressed as a polycistronic unit by the T7 promoter along with GFP prepeptide in MCS1. (c) Serine and threonine residues in the prepeptide (fused to GFP) are dehydrated by the VdsB dehydratase. (d) Cyclization of the dehydrated prepeptide by way of thioether bond formation between Dha or Dhb residues and cysteine residues (Michael-type additions) is performed by the VdsC cyclase. (e) Two stereochemical configurations of the Lan cross-linkage (dl- and ll-Lan) are possible and have been observed in nature. Four stereochemical configurations are possible for the MeLan cross-linkage, with dl- and ll-MeLan being commonly reported whereas d-allo-l-MeLan has only recently been observed and l-allo-l-MeLan has not yet been reported. (f) The GFP-fused lanthipeptide is then extracted and purified by IMAC and anion-exchange chromatography. (g) The mature lanthipeptide was liberated via proteolytic cleavage using the NisP protease. (h) The liberated lanthipeptides are then isolated from the cleavage mixture by reversed-phase HPLC and analyzed by time-of-flight mass spectrometry (LC–TOFMS).

The VdsA1 and VdsA2 peptides were liberated from GFP via proteolytic cleavage using the NisP protease as previously described and isolated by C8 reversed-phase high-performance liquid chromatography (RP-HPLC).^16,17 The VdsA1 and VdsA2 samples each produced two peaks of interest and were analyzed by UHPLC-MSMS (Figure 4a, P1A and P1B; Figure 5a, P2A and P2B). Prior to UHPLC-MSMS, all samples representing VdsA1 and VdsA2 were reduced with TCEP and alkylated with methyl methanethiosulfonate (MMTS) to identify free cysteine residues that were not involved in a thioether cross-linkage (lanthionine ring). All mass measurements presented have less than a 3 ppm error relative to the theoretical mass of the discussed species (Tables S1 and S2).

Figure 4.

(a) HPLC-C8-purified VdsA1 liberated from GFP using NisP was separated into samples P1A and P1B, which were collected for further analysis with UHPLC–MSMS. Samples P1A and P1B produced the TICs in (b) and (c), respectively. Accurate mass measurements indicated that the peaks observed in (b) and (c) both represented quadruply dehydrated VdsA1 (Table S1). However, the most abundant species in peaks (b) P1A-I and (c) P1B-1 carry an additional oxygen, those in P1A-III and P1B-III have undergone deamination, and those in (b) P1A-II and (b) P1B-II have a monoisotopic mass of VdsA1. All mass measurements correspond to their theoretical values to ≤2.47 ppm (Table S1). (d) Fragmentation spectra for the (d) P1A-II and P(e) 1A-III species from (b) indicating (Me)Lan cross-linkages as regions without fragment ions, although overlapping ring topologies cannot be elucidated by mass spectrometry alone (red rings). (f, g) Fragmentation spectra for the (f) P1B-II and (g) P1B-III species from (c) indicating the (Me)Lan cross-linkage topology. Amino acids of interest have been color-coded: Ser (dark red), Ala/Dha from Ser (red), Abu/Dhb from Thr (yellow), and Ala from Cys (blue). P1A-III and P1B-III lost an ammonia group from the N-terminal glycine, which likely produced a pyroglutamine (pGlu) residue (purple).

Figure 5.

(a) HPLC-C8-purified VdsA2 liberated from GFP using NisP was separated into samples P2A and P2B, which were collected for further analysis with UHPLC–MSMS. Samples P2A and P2B produced the TICs shown in (b) and (c), respectively. Accurate mass measurements indicated that the most abundant species in P2B-II observed in (c) represented VdsA2 with five dehydrations ([M – 5H₂O]), while P2B-I carried an additional oxygen ([M – 5H₂O + O]) (Table S2). The most abundant species in P2A-II carried an additional 307.08 Da which likely corresponds to a glutathione (GSH) adduct ([M – 5H₂O + GSH]), while P2A-I carried an additional GSH and oxidation ([M – 5H₂O + GSH + O]). All of the mass measurements correspond to their theoretical values to ≤1.95 ppm (Table S2). (d) Fragmentation spectra for the P2A-I and P2A-II species in (b) indicating MeLan cross-linkages as regions without fragment ions, while GSH (green) binds to either Dhb-10 or -11. (e) Fragmentation spectra for the P2B-II and P2B-III species in (c) indicating the MeLan cross-linkage topology. Amino acids of interest have been color-coded: Dhb/Abu from Thr (yellow) and Ala from Cys (blue). P2A-I and P2B-I gained an oxygen at the C-terminal methionine to form methionine sulfoxide (MetO) (cyan).

Samples P1A and P1B from VdsA1 were further resolved into the peaks P1A-I, P1A-II, and P1A-III (Figure 4b) and P1B-I, P1B-II, P1B-III, and P1B-IV (Figure 4c), respectively. Peaks P1B-I, P1B-II, and P1B-IV elute about 2.5 min later than P1A-I, P1A-II, and P1A-IV, respectively (Figure S2). Samples P2A and P2B from VdsA2 were further resolved into the peaks P2A-I and P2A-II (Figure 5b) and P2B-I and P2B-II (Figure 5c), respectively. Accurate mass measurements showed that the most abundant species in P1A-II corresponds to the mass of quadruply dehydrated VdsA1 peptide, [M – 4H₂O + nH]ⁿ⁺ (n = 2, 3, 4) (Figures 4 and S4–S7 and Table S1). Peptide fragmentation of the m/z 684.082 parent ion in peak P1A-II corresponds to the VdsA1 amino acid sequence with four dehydrated amino acids, [M – 4H₂O + 4H]⁴⁺ (Figures 4d and S16). These dehydrations occur at Thr-4 to yield Dhb-4 and at Ser-11 ,-15, and -21 to yield Dha-11, -15, and -21, respectively (Figures 4d and S16). Ser-26 escapes enzymatic dehydration by VdsB, but a neutral loss of 18 Da may be found in the y1, y2, and y7 fragment ions (Figure S16). Nonoverlapping MeLan and Lan ring locations can be assigned between Dhb-4 and Cys-8 as well as Dha-21 and Cys-25, respectively, due to the absence of fragment ions in the CID spectra (Figures 4d and S16). The fragmentation spectrum for the m/z 911.773 parent ion for the [M – 4H₂O + 3H]³⁺ species of VdsA1 agrees with the discussed m/z 684.082 [M – 4H₂O + 4H]⁴⁺ parent ion (Figure S17).

The most abundant species in peak P1A-I, at m/z 688.3298, corresponds to the mass of dehydrated VdsA1 peptide ([M – 4H₂O + 4H]⁴⁺) with an additional 16.00 Da that corresponds to an amino acid oxidation (Table S1 and Figures S8–S11). Peak P1A-III corresponds to VdsA1 with a loss of 17.03 Da at m/z 906.096 ([M – 4H₂O – NH₃ + 3H]³⁺) (Table S1 and Figures S12–S15). Peptide fragmentation of the m/z 906.096 parent ion in P1A-III indicated that deamination had occurred at the N-terminal Glu-1, likely forming a pyroglutamate residue (Figures 4e and S15).

Similar mass fragmentation patterns were observed for the P1B sample compared to the P1A sample derived from the VdsA1 expression. The accurate mass measurements showed that the most abundant species in P1B-II corresponds to the mass of quadruply dehydrated VdsA1, [M – 4H₂O + nH]ⁿ⁺ (n = 3, 4) (Table S1 and Figures S18–S20). Peptide fragmentation of the m/z 684.081 parent ion in peak P1B-II, [M – 4H₂O + 3H]³⁺, corresponds to the VdsA1 amino acid sequence with four dehydrated amino acids, as seen in sample P1A (Figures 5f and S31). These dehydrations occur at Thr-4 to yield Dhb-4 and at Ser-11,-15, and -21 to yield Dha-11, -15, and -21 respectively (Figures 5f and S32). Similarly, Ser-26 escapes enzymatic dehydration by VdsB, but a neutral loss of 18 Da may be found in the y1, y2, and y7 fragment ions (Figure S31). Nonoverlapping MeLan and Lan rings can be unambiguously assigned between Dhb-4 and Cys-8 as well as between Dha-21 and Cys-25, respectively (Figure 5e).

The most abundant species in peak P1B-I, at m/z 688.3303, corresponds to VdsA1 ([M – 4H₂O + 4H]⁴⁺) carrying an additional 16.00 Da, which indicates an amino acid oxidation (Table S1 and Figures S21–S23). Peaks P1B-III and P1B-IV correspond to VdsA1 with a loss of 17.03 Da indicating deamination of the N-terminal glycine in VdsA1 [M – 4H₂O – NH₃ + 3H]³⁺ (Figures 4g and S24–S30 and Table S1).

For the VdsA2 expression, accurate mass measurement showed that the most abundant species in P2A-II (Figure 5) corresponds to the mass of fully dehydrated VdsA2 with an additional 307.08 Da corresponding to glutathione (GSH), [M – 5H₂O + GSH] (Table S2 and Figures S33–S36). Peptide fragmentation of the m/z 749.305 parent ion indicated that the GSH is bound to either the Dha-10 or Dhb-11 residue in VdsA2 (Figures 5e, S37, and S41). Reaction of GSH with Dha and Dhb in lanthipeptides has previously been reported for nisin and cytolysin.^5,18 Nisin loses its antimicrobial activity upon reaction with GSH, while addition of GSH to cytolysin inhibited the formation of cytolysin’s ring C which lowered the antimicrobial activity and removed the hemolytic activity.^5,18 Interestingly, when cytolysin’s C-ring was prevented through alanine substitution, the antimicrobial activity was completely lost.⁵ Although GSH was bound to sample P2A for the majority of expressed VdsA2, it did not inhibit lanthionine ring formation. The absence of fragment ions indicates the presence of a MeLan ring between Dhb3 and Cys6 as well as between Dhb15 and Cys21. The effect that this GSH adduct may have on the bioactivity is not yet known, as the functions of these lanthipeptides have not yet been described.

The most abundant species in peak P2A-I, at m/z 1004.4012, corresponds to the fully dehydrated VdsA2 peptide with a GSH adduct and an additional oxidation giving a 16.00 Da mass increase ([M – 5H₂O + GSH + O]) (Figures 5d, S39, and S40 and Table S2). This oxidation occurs at the C-terminal methionine (Figures 5d, S41, and S42).

Accurate mass measurement showed that the most abundant species in P2B-II corresponds to the VdsA2 peptide with five dehydrations, [M – 5H₂O + nH]ⁿ⁺ (n = 2, 3) (Table S2 and Figures S44 and S45). Peptide fragmentation of the m/z 896.376 parent ion in peak P2B-II indicates that two MeLan bonds occur, between Dhb3 and Cys6 and between Dhb15 and Cys21 (Figures 5g and S46). Peak P2B-I contains VdsA2 with five dehydrations and an additional oxidation, [M – 5H₂O + O + nH]ⁿ⁺ (n = 2, 3) (Figure 5f and S47–S49 and Table S2). As in P2A-I, this oxidation in P2B-I occurs at the C-terminal methionine (Figures 5e and S50).

In addition to the dehydration of Ser and Thr residues observed in the VdsA1 and -A2 samples, the absence of MMTS adducts indicates that all cysteine residues are involved in a thioether cross-linkage. However, the unambiguous ring topologies in VdsA1, between Dha-11 and Dha-15 with Cys-16 and Cys-18, cannot be assigned with mass spectrometry alone. The formation of lanthionine rings between neighboring Dha/Dhb residues and a cognate cysteine residue has been reported in mersacidin, and therefore, a lanthionine bond between Dha15 and C16 in VdsA1 cannot be ruled out.^19,20 The most likely topology has been depicted based on the thallasomonasin A structure elucidated recently via NMR spectroscopy.¹³ VdsA1 must undergo peak splitting caused by some changes to its tertiary structure, as samples P1A and P1B have identical fragmentation spectra. Samples P1A and P1B may contain different (Me)Lan stereoisomer compositions or arrangements at each thioether cross-linkage. However, proline isomers have also been reported to cause peak splitting in peptides and lanthipeptides.²¹

It is difficult to anticipate the bioactivity of VdsA1 and VdsA2 peptides because there are no similar peptides with a characterized bioactivity. However, the VdsCB biosynthetic machinery provides more insight into the VdsA1 and VdsA2 lanthipeptides’ evolutionary history.^8,10 Pinensin, thalassomonasin, Ped15.1/2, and now viridisin seem to have a more closely shared origin within class I lanthipeptides and are likely used by their producers for a related function in their environments. Pinensin has antifungal activity, which is interesting considering that the genus is known to degrade chitin—an essential component in the cell wall and septa of pathogenic fungi. However, the multiple lanthipeptide dehydratases and cyclases present in the genome, which associate strongly with thalassomonasin and viridisin synthetases, may indicate that the biosynthetic pathway for these peptides is more complex than previously thought. Alternatively, thalassomonasin, Ped15.1/2, and now viridisin may induce a novel bioactivity altogether. Either way, new and exciting insights are on the horizon for this group.

Methods

Bacterial Strains and Culture Conditions

The marine isolate T. viridans was cultured in Marine broth at 37 °C. Escherichia coli BL21 (DE3) was cultured on Brain Heart Infusion or Luria–Bertani medium with agitation at 37 °C. For heterologous expression, E. coli BL21 was grown in Terrific broth with kanamycin added to a final concentration of 50 μg/mL to maintain the pRSF-Duet-derived plasmids.²² All growth media were supplied by Merck-Millipore (USA).

Phylogenetic Analysis

Representatives of the LanC cyclases conserved domain (CD04793) and LanB dehydratase Pfams pf04738 (C-terminus) with pf14028 (N-terminus) were downloaded from NCBI (https://www.ncbi.nlm.nih.gov/). These representatives were combined with characterized LanC and LanB, respectively.^2,3 Multiple sequence alignments were created for VdsC with the LanC (CD04793) library, and VdsB was aligned with the LanB (pf04738 and pf14028) library using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). The resulting alignments were analyzed by maximum likelihood phylogenetic tree generation using RaxMLGUI2.0 and visualized using ITOL.^23,24

Molecular Techniques

DNA analysis, manipulation, and plasmid cloning were performed according standard protocols described by Sambrook et al.(25) Genomic DNA was isolated from T. viridans according to Ausubel et al.(20) Plasmid DNA extractions were performed using the QIAGEN Plasmid Mini Kit (cat. no./ID:12123, Qiagen, Valencia, CA, USA).

Restriction enzymes (REs) and T4 DNA ligase were purchased from New England Biolabs (NEB) (Ipswich, MA, USA) and used according to the manufacturer’s instructions. Amplification of DNA via polymerase chain reaction (PCR) was performed using Phusion Plus DNA Polymerase (Thermo Scientific) according to the manufacturer’s instructions in a GeneAmp PCR system 9700 (ABI, Foster City, CA, USA).

Oligonucleotides were designed using the CLC main workbench program (CLC bio, Aarhus, Denmark) and purchased from Inqaba Biotechnical Industries (Pretoria, South Africa). DNA sequencing was performed by the Central Analytical Facilities (CAF) at the University of Stellenbosch, South Africa. DNA synthesis of the putative lanthipeptide gene sequences was performed by Biomatik (Cambridge, Ontario, Canada).

Agarose gel electrophoresis was used for the analysis and purification of RE-digested DNA fragments in TBE buffer at 10 V/cm using an Ephortec 3000 V power supply (Triad Scientific, Manasquan, NJ, USA).²⁶ Excised gel DNA fragments were purified using the QIAquick Gel Extraction Kit (cat. no./ID: 28706x4, Qiagen).

Construction of the pRSF–VdsA1 and pRSF–VdsA2 Expression Systems

The pRSF–VdsCB backbone vector was constructed by cloning the vdsCB modification genes from T. viridans into the MCS2 of pRSF-Duet for T7-controlled expression using IPTG. The VdsCB fragment was amplified as a polycistronic unit using the VdsCB_BgI_Fwd/VdsCB_Xho_Rev primers. These primers installed the 5′ BglII and 3′ XhoI restriction sites and placed the vdsCB in frame with the start codon in the MCS2 of pRSF-Duet (Figure S2).

Three putative lanthipeptides were identified upstream of the vdsCB modification genes by way of a conserved leader peptide with a double-glycine cleavage sequence (Figure 1). The double-glycine cleavage motif and two upstream amino acids were substituted with the NisP cleavage sequence (ASPR). The modified peptide gene sequences were synthesized on a single fragment with restriction sites flanking each peptide gene for cloning of each peptide into pBluescript, respectively. After each peptide was cloned into pBluescript, the respective peptide sequences were modified with flanking HindIII and SalI restriction sites during PCR amplification, which allowed cloning into pRSF–VdsCB. The primer sets used for each respective peptide HindIII/SalI modification can be found in Table S1. The GFP gene, mgfp5, was then amplified using the GFP_EcoRI_Fwd/GFP_Hind_Rev primer sets and recovered.

The pRSF–VdsCB pDNA was digested with EcoRI/SalI, and mgfp5 was digested with EcoRI/HindIII; each peptide amplicon was digested with HindIII/SalI. Each digested peptide fragment was respectively ligated into digested pRSF–VdsCB along with the digested mgfp5 fragment and used to transform chemically competent E. coli BL21. Green-fluorescent single colonies from each respective transformation reaction were subcultured for pDNA extraction and DNA sequencing to confirm construction.

Heterologous Expression and Purification of the GFP–VdsA1 and GFP–VdsA2 Fusion Proteins

The Minifors 5L bioreactor was filled with 3.5 L of Terrific broth and inoculated with overnight cultures of E. coli pRSF–VdsA1 and −VdsA2, respectively. Respective cultures were incubated at 37 °C until an OD₆₀₀ of 0.6 was reached, at which point they were induced with 0.1 mM IPTG for pRSF–VdsA1 and 0.05 mM for pRSF–VdsA2. The incubation temperature was lowered to 20 °C for 48 h before harvesting by centrifugation at 10000g. Cell pellets from each expression were resuspended in 15 mL of SB (50 mM Tris, 500 mM NaCl, pH 8.0) per gram of wet cell mass. Cell resuspensions were frozen at −20 °C and then thawed, and lysozyme, DNase, and RNase were added to final concentrations of 1 mg/mL, 5 μg/mL and 10 μg/mL, respectively, followed by sonication for 5 min at 50% pulse. The cell lysate was centrifuged at 15000g for 45 min, whereafter the cell-free supernatant was removed.

The GFP–VdsA1 and −VdsA2 fusions were purified from their respective cell-free lysates by IMAC using the His-Bind resin according to the manufacturer’s instructions. The IMAC elution, which contained the respective target proteins in SB500 buffer (50 mM Tris, 500 mM NaCl, 500 mM imidazole, pH 8.0), was diluted 30× with AB buffer (50 mM Tris, pH 8.3). The diluted eluents were respectively loaded onto a 5 mL DEAE Sepharose anion-exchange resin and washed with 2 column volumes of AB50 buffer (50 mM Tris, 50 mM NaCl, pH 8.3). The GFP–VdsA1 and −VdsA2 fusion proteins were respectively eluted in AB150 (50 mM Tris, 150 mM NaCl, pH 8.3).

HPLC Purification

The HPLC purification was performed using an Agilent 1260 Infinity HPLC system with a ZORBAX 300SB-C8 column (4.6 mm × 150 mm, 5 μm particle size) (Agilent, Santa Clara, CA, USA). Sample separation was achieved using linear gradient elution from 10% solvent B (acetonitrile + 0.1% TFA) to 60% solvent B over 25 min against solvent A (analytically pure water + 0.1% TFA). Elution profiles were monitored at 230 and 214 nm, respectively.

Nano-HPLC MS/MS Analysis

A Thermo Scientific Ultimate 3000 RSLC system equipped with a C18 trap column (2 cm × 100 μm) and a C18 analytical column (Luna C18, 5 μm, Phenomenex) was used for further analysis of HPLC-purified P1A, P1B, P2A, and P2B samples. Solvent A (2% acetonitrile in water with 0.1% formic acid) and solvent B (100% acetonitrile) were used for sample separation over a 55 min Chromeleon nonlinear gradient 5 from 5% to 50% solvent B at 0.3 μL/min. The samples were loaded onto the trap column in solvent A at 2 μL/min from a temperature-controlled autosampler set to 7 °C. Loading was performed for 5 min before the sample was eluted onto the analytical column. Chromatography was performed at 45 °C, and the outflow was delivered to the mass spectrometer through a stainless steel nanobore emitter.

Mass spectrometry was performed using a Thermo Scientific Fusion mass spectrometer equipped with a Nanospray Flex ionization source. Data were collected in positive mode with the spray voltage set to 1.9 kV and the ion transfer capillary set to 275 °C. Spectra were internally calibrated using polysiloxane ions at m/z 445.12. The MS1 scans were performed using the Orbitrap detector set at 120000 resolution over the scan range of m/z 400–1800 with automatic gain control (AGC) target at 4 × 10⁵ and a maximum injection time of 50 ms. Data were acquired in profile mode. The MS2 acquisitions were performed using monoisotopic precursor selection for ions with charges 2+ to 9+ with the error tolerance set to ±10 ppm. Precursor ions were excluded from fragmentation once for a period of 60 s. Precursor ions were selected for fragmentation in higher-energy collisional dissociation (HCD) mode using the quadruple mass analyzer with the HCD energy set to 30%. Fragment ions were detected in the Orbitrap mass analyzer set to 30000 resolution. The AGC target was set to 5 × 10⁴ and the maximum injection time to 100 ms. The data were acquired in centroid mode.

Software

Peptide fragmentation data were analyzed with MZmine²⁷ (http://mzmine.github.io/), MSconvert²⁸ (https://proteowizard.sourceforge.io/download.html), and pLabel (http://i.pfind.org/). Images were created with BioRender.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acssynbio.2c00480.

• All identified monoisotopic mass measurements with the expected mass and ppm error calculations (Tables S1 and S2); oligonucleotide primer sequences (Table S3); multiple sequence alignments for all characterized lanthipeptides against VdsA1 and VdsA2 (Figure S1); expression plasmid maps (Figure S2); UHPLC-MSMS chromatograms (Figure S3); VdsA1 P1A monoisotopic mass measurements and fragmentation spectra (Figure S4–S17); VdsA1 P1B monoisotopic mass measurements and fragmentation spectra (Figure S18–S32); VdsA2 P2A monoisotopic mass measurements and fragmentation spectra (Figure S33–S37); VdsA2 P2B monoisotopic mass measurements and fragmentation spectra (Figure S38–S50) (PDF)

The authors declare no competing financial interest.