The antimicrobial activity of the glycocin sublancin is dependent on an active phosphoenolpyruvate-sugar phosphotransferase system

Abstract

Antimicrobial resistance is a global challenge that is compounded by the limited number of available targets. Glycocins are antimicrobial glycopeptides that are believed to have novel targets. Previous studies have shown that the mechanism of action of the glycocin sublancin 168 involves the glucose uptake system. The phosphoenolpyruvate:sugar phosphotransferase system (PTS) phosphorylates the C6 hydroxyl group on glucose during import. Since sublancin carries a glucose on a Cys on an exposed loop, we investigated whether phosphorylation of this glucose might be involved in its mechanism of action by replacement with xylose. Surprisingly, the xylose analog was more active than wild-type sublancin and still required the glucose PTS for activity. Overexpression of the individual components of the PTS rendered cells more sensitive to sublancin and their resistance frequency was considerably decreased. These observations suggest that sublancin is activated in some form by the glucose PTS or that sublancin imparts a deleterious gain-of-function on the PTS. Superresolution microscopy studies with fluorescent sublancin and fluorescently labeled PTS proteins revealed localization of both at the poles of cells. Resistance mutants raised under conditions that would minimize mutation of the PTS revealed mutations in FliQ, a protein involved in the flagellar protein export process. Overexpression of FliQ lead to decreased sensitivity of cells to sublancin. Collectively, these findings enforce a model in which the PTS is required for sublancin activity, either by inducing a deleterious gain-of-function or by activating or transporting sublancin.

Graphical Abstract

Glycocins (glycopeptide bacteriocins) are members of a family of natural products called RiPPs (Ribosomally synthesized and posttranslationally modified peptides).^{1, 2} Currently characterized glycocins carry one or two sugar moieties that are required for their antimicrobial activities.^3–8 Genome mining studies have rapidly expanded the glycocin family and suggest that their structural diversity is much larger than originally anticipated.^9–13 Type I glycocins¹³ are characterized by the presence of disulfides that tether two helices. In addition, these glycocins carry a sugar moiety that is present on Cys, Ser, or Thr residues located in a loop that connects the two helices (Figure 1).¹ The type I glycocin sublancin 168 has antimicrobial activity against drug resistant pathogens such as methicillin resistant Staphylococcus aureus (MRSA) and vancomycin resistant Enterococcus sp. (VRE) (Figure 1).^{3, 14} The sublancin biosynthetic gene cluster is present on the spβ prophage in the genome of Bacillus subtilis 168. Sublancin is generated by posttranslational modification of the precursor peptide SunA.¹⁴ The peptide is glucosylated on Cys22 by the glycosyltransferase SunS,³ followed by disulfide formation by the thiol disulfide oxidoreductases BdbA and BdbB.¹⁵ The mature peptide is generated via proteolytic cleavage of a leader peptide and export of the bioactive natural product by the bifunctional enzyme SunT. The resistance gene sunI imparts immunity (Figure 1).¹⁶ B. subtilis 168 Δspβ lacks the sunI gene and is highly sensitive to sublancin.¹⁷

Figure 1.

Biosynthetic gene cluster (BGC), biosynthetic pathway, and structure of sublancin 168.

The mechanism by which sublancin elicits its bioactivity is still mostly unknown. The NMR-structure of sublancin has been solved and structure-activity relationship studies have indicated the requirement of a sugar,^18–20 but the identity of the carbohydrate seemed less critical as galactose, mannose and N-acetylglucosamine (GlcNAc) attached to Cys22 still resulted in an active peptide. Several factors have been implicated in the bactericidal activity of sublancin. The mscL gene, which encodes a large mechanosensitive channel in bacteria, has been linked to the sensitivity to sublancin.²¹ Some bacteria are resistant to sublancin due to the alternative sigma factor σ^W, which is responsible for the regulation of the yqeZ-yqfAB operon.^{16, 22} While the precise functions of these proteins are unknown, yqeZ is hypothesized to encode a transmembrane protease and both YqfA and YqfB are membrane-anchored proteins.²² Additionally, the phosphoenolpyruvate:sugar phosphotransferase system (PTS) (Figure 2) is intricately involved with the bioactivity of sublancin. Deletion of the ptsGHI operon confers resistance to sublancin, and spontaneous resistance mutants in Bacillus halodurans trace back to the pts locus.¹⁷ The PTS is responsible for the transport of sugars to the cytoplasm of bacterial cells (Figure 2).²³ Site-directed mutagenesis of PtsH/HPr has demonstrated that only the phosphorylation branch of the PTS is involved in the bioactivity of sublancin and not the catabolite repression branch.¹⁷ Finally, microarray analysis of the gene expression profiles of sensitive strains demonstrated downregulation of the PTS components upon exposure to sublancin.¹⁷

Figure 2.

(A) The glucose PTS in B. subtilis and an overview of the regulatory network. (B) Structures of sublancin-glucose and sublancinxylose.

These findings suggested that the PTS might be the target of sublancin, but a simple blockage of glucose import would be expected to be bacteriostatic and not bactericidal. Other bactericidal natural products that target sugar PTSs typically have additional activities such as pore formation,^24–26 but such activity is not seen with sublancin (Figure S1).¹⁷ Furthermore, the presence of PTS sugars in the media decreases the sensitivity to sublancin,²⁰ whereas the activity of several other bacteriocins that target the PTS is increased in the presence of added PTS sugars including glucose.^27–29 These findings suggested that the PTS may be needed for activity but may not constitute the target of sublancin.

It seems more than coincidental that the glucose PTS is important for sublancin activity and that sublancin is modified with a glucose moiety. In this study, we investigated the possibility that the sugar on sublancin might undergo phosphorylation at C6, similar to phosphorylation upon sugar import (Figure 2). Furthermore, we probed the effects of overexpressing or deleting the individual genes of the ptsGHI operon in sensitive B. subtilis Δspβ cells and used microscopy with cells expressing fusions of the PTS components with genetically encoded fluorescent proteins in combination with fluorescently labeled sublancin. Finally, we raised resistant mutants to sublancin under conditions where the glucose PTS is important for growth and performed whole genome sequencing on these mutants. These experiments resulted in mutations in proteins other than the PTS. Collectively, the results of these studies indicate that sublancin requires a functional glucose PTS to exert downstream deleterious effects.

RESULTS AND DISCUSSION

The C6 hydroxyl group of glucose on sublancin is not required for activity

In the process of glucose import, PtsG phosphorylates the 6-OH as the sugar crosses the cell membrane, forming glucose-6-phosphate.²³ Previous results indicated sublancin competes with glucose for the PTS.²⁰ Thus, we hypothesized that phosphorylation of the glucose moiety on the glycocin might be a novel mechanism of activation of sublancin. Variants carrying galactose, mannose or GlcNAc had antimicrobial activity, albeit diminished, but they also all contain a C6 hydroxyl group. Therefore, in this study, we prepared xylose-modified sublancin by taking advantage of the tolerance of SunS for alternative nucleotide sugars including UDP-xylose that was reported previously;³ xylose lacks the C6 hydroxyl group (Figure 2B). The bioactivity of the xylose variant was evaluated using the broth dilution method against B. subtilis Δspβ. The minimal inhibitory concentration (MIC) of sublancin-xylose was 78 nM, surprisingly indicating it was about 4-fold more active than sublancin itself. This observation argues against phosphorylation of the glucose on sublancin as a required mechanism of activation. As shown below, the activity of sublancin-xylose still involves the glucose PTS.

Effects of overexpression of the PTS proteins on sublancin sensitivity

To further investigate the role of the glucose PTS in sublancin activity, we overexpressed the pts genes in B. subtilis Δspβ using the xylose inducible pHCMC04 plasmid (Figures 3, S3, Tables S1–2), which has a copy number of 2–6.^{30, 31} Overexpression of the genes was induced with 2% xylose after two hours of growth, and one hour later sublancin was added to the media at the MIC for B. subtilis Δspβ, which is named wild-type (WT) from here onwards. Whereas the WT strain overcame the growth inhibition imposed by sublancin after six hours of incubation due to rapid development of resistance as reported previously,¹⁷ growth restoration was much slower for the strains overexpressing the PTS proteins (Figure 3). The overexpression of ptsH and ptsG was accompanied with a considerable decrease in MIC, with a smaller effect upon expression of ptsI (Table 1). These results show that expression of the PTS proteins does not lead to decrease in sensitivity as would be expected if sublancin inhibited the PTS function, but instead increases sensitivity and represses development of resistance.

Figure 3.

Growth curves obtained upon overexpression of various PTS proteins in B. subtilis Δspβ in LB medium and treatment with sublancin. The data shown is representative of three independent experiments. For clarity, individual datapoints are not shown. For examples of growth curves showing individual datapoints, see Figure S10.

Table 1.

MICs of sublancin against B. subtilis Δspβ overexpressing individual PTS genes in refined LB medium.³²

Strains	MIC (nM)
WT	312
pHCMC04_ptsG	78
pHCMC04_ptsH	39
pHCMC04_ptsI	156

Resistance frequency upon overexpressing PTS proteins

We next determined the resistance frequency of the WT and overexpression strains by plating bacterial culture onto refined LB agar plates containing sublancin at 4xMIC and lacking sublancin to determine the number of viable bacterial cells. For all strains overexpressing pts genes, a considerable reduction was observed in the resistance frequency compared to B. subtilis Δspβ (Table 2). These observations are consistent with the growth curves depicted in Figure 3. Sublancin requires functional copies of all three PTS proteins (PtsGHI) for antimicrobial activity.¹⁷ Thus, cells overexpressing one of the pts genes (e.g. ptsH) from a multi-copy plasmid are less likely to accumulate mutations in all copies of that gene.³³ The resistance frequency may therefore be reduced because only spontaneous mutations in the two chromosomal pts genes (e.g. ptsGI) would result in resistance.

Table 2.

Resistance frequency of B. subtilis Δspβ overexpressing different PTS genes in refined LB medium.

Strains	Resistance frequency
WT	10−6
pHCMC04	10−6
pHCMC04_ptsG	10−8
pHCMC04_ptsH	10−9
pHCMC04_ptsI	10−7

Deletion of PTS genes and complementation

Using gene editing technology based on the CRISPR-Cas9 system (Figure S4, Tables S1–2), knockouts were generated of ptsG, ptsH, ptsI, and ptsGHI which were confirmed via Sanger sequencing (Table S3). As anticipated, the four deletion strains were resistant to sublancin. In addition, deletion of ptsG, ptsH, or ptsI resulted in resistance against sublancin-xylose (Table S4). These data resemble previous findings that showed that spontaneous resistant mutants against sublancin were also resistant to sublancin variants carrying GlcNAc, mannose, and galactose.²⁰ Xylose and galactose are non-PTS sugars in B. subtilis. Furthermore, glucose protected B. subtilis from the activity of sublancin-xylose (Figure S5). Collectively, these data strongly suggest that sublancin-xylose exerts its activity using a very similar mechanism as WT sublancin that still requires the glucose PTS.

In control experiments, the PTS deletion strains were then individually transformed with the pHCMC04 plasmid carrying one of the PTS genes to test for complementation. Transformed cells were plated on agar plates containing 2% xylose and a spot-on-lawn agar diffusion assay was performed with sublancin. Sensitivity to sublancin was only observed when complementing the deleted genes, and MIC values were determined (Table 3). The observation that only complementation with the deleted gene restored sensitivity to sublancin and the resistance phenotype displayed by the ΔptsGHI mutant under all transformation and over expression conditions further demonstrates the necessity of a fully functional glucose PTS for the bioactivity of sublancin.

Table 3.

MIC values of different deletion mutants of B. subtilis Δspβ overexpressing PTS genes in refined LB medium.

Bacterial strainsa	MIC (nM)
WT	312
ΔptsG_ptsG	156
ΔptsG_ptsH	Resistant
ΔptsG_ptsI	Resistant
ΔptsH_ptsG	Resistant
ΔptsH_ptsH	78
ΔptsH_ptsI	Resistant
ΔptsI_ptsG	Resistant
ΔptsI_ptsH	Resistant
ΔptsI_ptsI	156
ΔptsGHI_ptsG	Resistant
ΔptsGHI_ptsH	Resistant
ΔptsGHI_ptsI	Resistant

^aThe nomenclature used lists the overexpressed protein after the underscore symbol in the deletion mutant of interest.

Sublancin localization by superresolution microscopy

To gain further insights, we performed super-resolution microscopy studies. We first coupled cyanine-5-amine to the C-terminus of sublancin, which is the only carboxylate in the molecule (Figure S2; N-terminally labelled sublancin was inactive). Sublancin-Cy5 was active with an MIC of 625 nM against B. subtilis Δspβ. These cells were then transformed with pHCMC04 encoding YFP-labeled PtsG, GFP-labeled PtsH, CFP-labeled PtsI or RFP (mCherry) labeled SunI. The genetically encoded fluorophores were fused at the C-terminus of the proteins separated by a six amino acid flexible linker. Such fusions have been previously demonstrated to retain the function of PTS proteins.^{34, 35} Following overnight growth with 2% xylose, cells were diluted 1:100, grown for an additional 2 h, treated with Cy5-labeled sublancin and fixed on slides, which were imaged with super-resolution Airyscan confocal laser-scanning microscopy. Sublancin localized predominantly at the poles of the cells and the pattern of localization was similar to that of the PTS proteins or the immunity protein SunI (Figure 4, Figure S6).

Figure 4.

Super-resolution microscopy of sublancin-Cy5 with B. subtilis Δspβ overexpressing fluorescently labeled PTS or SunI proteins. Representative images of cells are shown with the bar graph showing the statistical analysis for the total number of randomly chosen cells that showed fluorescence. (A) Sublancin-Cy5 localization in B. subtilis Δspβ. (B) Localization of sublancin-Cy5 and PtsG-YFP in B. subtilis Δspβ/pHCMC04-ptsGYFP. (C) Localization of sublancin-Cy5 and PtsH-GFP in B. subtilis Δspβ/pHCMC04-ptsHGFP. (D) Localization of sublancin-Cy5 and PtsI-CFP in B. subtilis Δspβ/pHCMC04-ptsICFP. (E) Localization of sublancin-Cy5 and SunI-RFP in B. subtilis Δspβ/pHCMC04-sunIRFP. Statistical analyses were performed in each panel. N, the total number of cells counted.

The results are in agreement with the localization pattern observed for a subset of PTS proteins (PtsHI) investigated in a previous study.³⁴ Pre-incubation of cells with unlabeled sublancin abolished binding of Cy5-labeled sublancin (Figure S7) suggesting that the fluorescently labelled analog binds to the same target(s) as WT sublancin. In addition, WT sublancin added after labelling cells with sublancin-Cy5 was not able to completely compete off the fluorescent variant suggesting that the interaction at the poles of the cells is relatively tight.

Resistance mutants grown on glucose

All data in the current study point to a deleterious gain of function of the glucose PTS²⁰ or a role for the PTS to facilitate sublancin to reach its molecular target. Our previous spontaneous resistant mutants in B. halodurans were generated in rich medium (LB) where mutations in the PTS were tolerated, and hence readily accessible. In this study, we obtained resistant mutants of sublancin under conditions designed to minimize mutations in the PTS by selecting sublancin-resistant mutants in chemically defined medium (CDM) with glucose as the sole carbon-source. The PTS is the main system for glucose uptake in B. subtilis, but a hexose/H⁺ symporter (GlcP) can also be used.³⁶ However, GlcP appears to also require a functional PTS.³⁷ A third uptake system may exist but is not well characterized.^{36, 38} For these experiments, we used B. subtilis Δ6 cells that carry the chloramphenicol resistance gene marker to avoid contamination during the long growth times required for resistance to develop. For comparison, we also generated resistance mutants of B. subtilis Δ6 in LB. We then performed whole genome sequencing of seven selected mutant strains as well as wt B. subtilis Δ6. The three mutants raised in glucose CDM had no mutations in PTS proteins, as we had hoped by using glucose as the only carbon source. It appears that under these conditions there is a selection against accumulation of mutations in the PTS. Instead, the three mutants sequenced all contained the same single nucleotide deletion in the gene that encodes FliQ, an 88-amino acid protein that is part of the Bacillus type III secretion system (T3SS) essential for cell motility and flagellar protein export (Table S5).³⁹ The deletion occurs at Ile66. FliQ plays a key role in the formation of the export channel and serves as the nucleation point for subsequent T3SS basal body assembly.^{40, 41} FliQ, FliP and FliR form a helix-turn-helix complex with 5:4:1 stoichiometry to form the core of the T3SS secretion channel.⁴² FliQ has been shown to be essential in Caulobacter crescentus as FliQ mutants exhibit cell division defects.⁴³ Overexpression of FliQ in B. subtilis Δspβ resulted in a more than ten-fold increase in the MIC of sublancin (>20 μM) compared to wt (312 nM), suggesting that this protein may be tied to sublancin’s mode of action. Expression of FliQ in Δpts mutants did not lead to restoration of sublancin sensitivity, indicating that if FliQ is a target, the PTS is required for sublancin to reach this target (Table S6) As expected, all four sequenced B. subtilis Δ6 resistance strains that were grown in LB accumulated mutations in at least one of the glucose PTS components (Table S5), as observed previously for B. halodurans.¹⁷

Comparison with other bacteriocins targeting the PTS

A number of bacteriocins have been reported that target a sugar PTS. Pediocin PA-1 and garvicin Q use the mannose PTS^{25, 44, 45} as a receptor as do lactococcin A, B, and Z. The lactococcins are bactericidal with lactococcin A inducing pore formation,^{25, 46} whereas lactococcin Z does not.²⁷ Hence, lactococcin Z, like sublancin, must use another, currently unknown, mechanism of killing after interaction with the PTS. Dysgalacticin targets mannose and glucose PTSs and inhibits glucose uptake. Its bactericidal activity is the result of dissipation of the membrane potential.²⁶ Perhaps the closest phenotype to the observations with sublancin was reported for the circular bacteriocin⁴⁷ garvicin ML. The maltose ABC transporter MalEFG was shown to be essential for the antimicrobial activity of garvicin ML and the sensitivity increased with higher expression levels of malEFG.²⁸ Although the mechanism of cytotoxicity was not determined, the authors suggested that interaction with the bacteriocin might result in an open state of the permease of the transporter causing efflux of intracellular metabolites. Such a mechanism is also a possibility for sublancin 168.

Comparisons with other glycocins are also informative. Glycocin F carries a GlcNAc on a loop connecting two helices that is essential for bioactivity,^{4, 48, 49} much like the glucose on Cys22 of sublancin. Glycocin F also has an additional S-linked GlcNAc on a C-terminal tail. The compound interacts with the GlcNAc PTS but its activity is bacteriostatic possibly through growth inhibition by preventing GlcNAc uptake.^{5, 50} It is tempting to speculate that during evolution, sublancin may initially have had similar bacteriostatic activity and may later have acquired its bactericidal activity. This hypothesis suggests that most likely its killing activity is associated with its interaction with the glucose PTS, consistent with the observed similar localization of the PTS proteins and sublancin. A divergent evolutionary relationship between glycocin F and sublancin is also suggested from phylogenetic analysis.¹³

DISCUSSION

Sublancin induces a deleterious gain of function of the glucose PTS that is magnified by overexpression of either PtsG, PtsH, or PtsI. Overexpression of these proteins renders the cells more sensitive and makes resistance development more difficult. Sublancin is not likely to be phosphorylated on its glucose group for activity, and the compound appears to co-localize with the PTS proteins at the poles of the cells, suggesting it may potentially remain bound to the membrane-bound PTS complex. The current data illustrating the requirement of a functional glucose PTS complement previous data showing that mutation of His15 in PtsH, which is required for phosphoryl transfer to PtsG (Figure 2), is sufficient to impart sublancin resistance.¹⁷ The observation that sublancin-xylose is active and requires the glucose PTS despite xylose not being a PTS sugar, suggests that the peptide sequence and fold may be more critical for the interaction with the PTS than the sugar.

When cells are grown under conditions where a functional PTS is beneficial, presumably reducing the probability of mutations in the pts locus, spontaneous resistance mutations accumulated in FliQ, which may therefore also be involved in sublancin toxicity. How the compound eventually kills cells is at present not clear and requires further investigation.

METHODS

Polymerase chain reaction (PCR) amplifications were carried out using an automated thermocycler (C1000, Bio-Rad). DNA sequencing was performed using appropriate primers by ACGT, Inc. MALDI-TOF MS analyses were conducted at the Mass Spectrometry Facility at UIUC using an UltrafleXtreme spectrometer (Bruker Daltonics). For MALDI–TOF MS analysis, samples were desalted using ZipTipC18 (Millipore), and spotted onto a MALDI target plate with a matrix solution usually consisting of a saturated aqueous solution of super DHB (2,5-dihydroxy benzoic acid; Sigma Aldrich). Peptides obtained from expression in E. coli were purified by preparative reversed-phase high performance liquid chromatography (RP–HPLC) on an Agilent 1260 Infinity II instrument equipped with a Phenomenex C5 column at a flow rate of 8 mL/min or with a Macherey Nagel C18 (MN_C18) column at a flow rate of 4 mL/min. For RP–HPLC, solvent A was 0.1% TFA in H₂O, and solvent B was pure MeCN containing 0.1% TFA. An elution gradient from 0% solvent B to 100% solvent B over 30 min was used unless specified otherwise.

Materials

All chemicals and enzymes used were purchased from Sigma-Aldrich or Fisher Scientific unless noted otherwise. Endoproteinase Glu-C was purchased from Worthington Biosciences. UDP-xylose was obtained from Complex Carbohydrate Research Center. Oligonucleotide primers used for molecular cloning were purchased from Integrated DNA Technologies. Phusion High-Fidelity DNA polymerase, Q5 DNA polymerase, Taq ligase, dNTP solutions, T4 DNA ligase, and all restriction endonucleases were purchased from New England BioLabs. Gel extraction, plasmid miniprep, and PCR purification kits were purchased from QIAGEN. Protein Calibration Standard I and Peptide Calibration Standard II for MALDI–TOF MS were purchased from Bruker. E. coli NEB Turbo was used as host for cloning and plasmid propagation, and E. coli BL21(DE3) SHuffle was used as a host for expression of deglycosublancin.¹⁹ Nickel resin used for gravity-based purification of His-tagged peptides or proteins was purchased from Takara Biosciences. Gradient gels used for checking protein expression were purchased from Biorad Inc. Plasmids pHCMC04 and pJOE8999 for molecular cloning in B. subtilis Δspβ were purchased from the Bacillus Genetics Stock Center (BGSC). B. subtilis Δ6 and plasmids encoding fluorescent proteins in Bacillus were also purchased from BGSC. Cyanine-5-amine was purchased from Lumiprobe.

Isolation of authentic sublancin and MIC determination of sublancin and analogues

Isolation of authentic sublancin was performed as described previously.³ We obtained 6 mg of wild type sublancin per L of culture. The structure and the purity of the peptide were confirmed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) and high-performance liquid chromatography (HPLC) using a Shimadzu Prep-instrument. For MIC determination, a 96-well microtiter plate was prepared as mentioned previously.³² B. subtilis Δspβ was grown in refined LB at 37°C overnight. Refined LB was prepared as described previously.³² Sublancin is more active against B. subtilis Δspβ grown in refined LB and B. subtilis Δspβ does not develop resistance as quickly in refined LB such that MIC determinations can be done at 18 h of growth. The culture was diluted to 10⁵ CFU/mL in refined LB media in each well. Sublancin or its analogue was dissolved in water and serially diluted to obtain working concentrations of 200, 100, 50, 25, 12.5, 6.25, 3.125, 1.56, 0.78, 0.39, 0.19, and 0.09 μM. OD600 readings were obtained after 18 h using a Synergy H4 Hybrid Multi-Mode Microplate BioTeK plate reader. Cell-free untreated refined LB medium, and untreated cell culture were used as negative controls.

Purification of SunS for in vitro assays

SunS used to add xylose to SunA was obtained and purified according to previous protocols with a few modifications.²⁰ Briefly, the plasmid encoding SunS in pET28b was used to transform E. coli Rosetta BL21 (DE3) cells. Bacterial cultures were grown in TB medium incubated at 37 °C with shaking until an OD₆₀₀ of between 1.0 or 1.4 was reached. Protein expression was induced with 0.5 mM IPTG overnight at 18 °C. After harvesting, the cell pellets were suspended at 1 g/mL in buffer A (20 mM Tris–HCl, 100 mM NaCl, 10% glycerol, pH 7.4) supplemented with Pierce Protease Inhibitor tablets (Thermo Fisher Scientific, 1 tablet per 20 mL), benzonase (1 U for every 20 mL of lysate) and lysozyme (100 μg/mL), and the cells were incubated on ice for 30 min. The lysate was sonicated (12 cycles of 10 s on followed by 10 s off) on ice to ensure complete lysis. Insoluble cell debris was pelleted via centrifugation at 50,000 × g for 1 h at 4 °C. The supernatant was filtered using 0.45 μm syringe filters. The filtrate was loaded onto a column containing settled Ni resin and cell lysate was applied three times to it under the action of gravity (5 mL of settled resin for 2 L of cell culture). The column was washed with 50 mL of buffer A and then a manual gradient from 0% to 100% buffer B (500 mM imidazole, 20 mM Tris–HCl, 100 mM NaCl, pH 7.4) was applied for a total of 15 column volumes. UV absorbance at 280 nm was monitored and fractions were collected and analyzed by SDS–PAGE. Fractions containing the desired protein were pooled together and concentrated and used for in vitro assays; 20 mg of SunS was obtained per liter of culture.

Preparation and purification of sublancin-xylose

Sublancin-xylose was prepared according to a previously described protocol^{3, 20} with minor modifications, as detailed here. Instead of SunA-core peptide, HalA2_LP-SunA_CP was used¹⁹ as the peptide in the glycosylation and oxidative folding reactions. The in vitro glycosylation reaction contained 50 mM Tris buffer, 2 mM HalA2_LP-SunAcore_CP, 20 μM SunS, 1 mM TCEP and 1 mM UDP-xylose in the presence of 2 mM MgCl₂ in a total reaction volume of 200 μL. After glycosylation was complete as shown by MALDI-TOF MS analysis following incubation for 48 h, the disulfides were installed in the peptide via incubation with cystamine and cysteamine to final concentrations of 10 mM each in 50 mM Tris Buffer which contained dimethyl sulfoxide (DMSO) as an additive to a final concentration of 2%. The reaction mixture was digested with endoproteinase Glu-C to remove the leader peptide and filtered prior to HPLC purification which was performed as described previously.¹⁹ A total of 2 mg of sublancin-xylose was obtained after HPLC purification from 20 mg of purified HalA2_LP-SunA_CP peptide used as the substrate for the in vitro glycosylation, oxidative folding and leader peptide removal.

Molecular cloning of PTS and FliQ genes for overexpression in B. subtilis Δspβ

The plasmid pHCMC04 was obtained from BGSC and it was linearized using the appropriate primers (Table S2, pHCMC04_BB_R and pHCMC04_BB_F). This shuttle plasmid allows for controlled induction of the inserted genes under a xylose promoter and encodes for resistance to chloramphenicol for selection in Bacillus and an ampicillin resistance marker for selection in E. coli. The genes ptsG, ptsH and ptsI were amplified from B. subtilis Δspβ using colony PCR and the primer pairs pHCMC04_ptsX_F and pHCMC04_ptsX_R (X=G, H or I, Table S2). For transformation of a plasmid in Bacillus, a recA+ strain of E. coli is required, which enables the isolation of multimeric plasmid (the exact oligomerization state of the plasmid has not been determined, but usually they are dimers or tetramers). Hence we used the NEB Turbo strain for all our cloning purposes where plasmids would be used to transform Bacillus. A Gibson assembly was performed to ligate each gene with the plasmid. Following the isolation of the plasmid from E. coli and verifying the sequence via Sanger sequencing, the plasmids were used to transform chemically competent B. subtilis Δspβ following standard protocols.⁵¹ The same procedure was used to obtain pHCMC04_FliQ plasmids by using primer pair pHCMC04_FliQ_F and pHCMC04_ FliQ_R (Table S2).

Determination of MIC and resistance frequency of overexpression strains

MICs for the strains overexpressing the PTS genes were determined as described previously except that the overnight incubation mixture contained 2% xylose and chloramphenicol and the same components were present in the 96-well plate in addition to sublancin. Additionally, sublancin was prepared at working concentrations of 50, 25, 12.5, 6.25, 3.125, 1.56, 0.78, 0.39, 0.19, 0.09, 0.045 and 0.0225 μM such that the effective concentration in the wells would be 10-fold lower after addition to the wells.

For determination of resistance frequency, cells transformed with the empty pHCMC04 plasmid or containing PTS genes were plated on agar containing chloramphenicol at 5 μg/mL in the presence or absence of sublancin at 4x MIC of the respective strain. Additionally, 2% xylose was added to all plates that were overexpressing the PTS genes. Serial dilutions were plated and the ratio of the number of colonies on plates that contained sublancin and plates without sublancin was determined to obtain the resistance frequency of the strains containing either the empty plasmid or overexpressing the PTS genes. The resistance frequency of the cells with the empty plasmid towards sublancin was identical to that of cells without the plasmid. The resistance frequencies of the cells containing plasmids carrying pts genes is shown in Table 2.

Membrane potential studies with DiOC2 dye

Cultures of B. subtilis Δspβ were grown to OD600 1.0 and diluted to OD600 0.1 in fresh refined LB. The culture was then combined with 30 μM DiOC2 dye (ThermoFisher D14730) and incubated for 20 min. Then sublancin (0.5xMIC, 1xMIC and 4xMIC; 1xMIC 0.31 mM) or nisin (0.5xMIC, 1xMIC and 4xMIC; 1xMIC 0.65 mM) were added and incubated for another 20 min at room temperature before analysis. Another two tubes were prepared without antibiotic treatment as controls. The culture was analyzed with a BD Biosciences LSR II flow cytometer using excitation at 488 nm with an argon laser and measurement of emission through a band-pass filter at 616/200 nm. A minimum of 30,000 events were detected for each sample. The experiment was repeated three times and the mean red/green fluorescence ratio was determined.

Construction of CRISPR-Cas9 based knockouts in B. subtilis Δspβ

For performing CRISPR-Cas9 deletions of ptsG, ptsH, ptsI and ptsGHI, the CutSPR software was used which is designed for editing with the pJOE plasmid for Bacillus sp.^{52, 53} The exact loci of the gene of interest were specified in the software, which then provides a choice of different protospacer sequences for fusion to the sgRNA in the primer following an NGG sequence in the gene of interest and scores them. This allows for the design of a specific sgRNA to target deletion of only the gene of interest. The sequence with the least predicted off-target hits in the genome was used in this experiment. Once the guide RNA had been chosen (Table S2, ptsX_pJOE_n, where X=G, H, I or operon, n =1 or 2) the software designed six primers and numbered them sequentially (see below). It also allowed the specification of the length of flanking residues on each side of the chosen target sequence. Ideally 700–1000 base pairs are required on either side of the gene of interest. Primers 1 and 2 (Table S2) were hybridized via an incubation reaction at room temperature and served as the protospacer sequence that was fused to the sgRNA for the plasmid. Reaction volume was 50 μL and each primer was used at a concentration of 5 μM. The pJOE8999 plasmid (obtained form BGSC) was linearized with BsaI and the protospacer was ligated using T4 DNA ligase. The ligation mixture was used to transform E. coli. Several plasmids were isolated from E. coli, digested with BsaI and EcoRI to confirm the insertion, and sequenced. Plasmids were sequenced with primer RH003 (Table S2).

Following verification of the presence of the specific protospacer fused to the sgRNA in the pJOE8999 plasmid, the editing template was constructed. The flanking sites were amplified using genomic DNA as the template and primers 3 and 4 and primers 5 and 6 (Table S2, the numbering was used for each gene), respectively. The PCR products were purified on a 1.5% agarose gel and the concentration was adjusted to 0.1 ng/μL. Then, 10 μL of each PCR reaction was combined in an eppendorf tube and mixed thoroughly and 1 μL of this mixture was used as a template for the next PCR reaction which used primers 3 and 6 as these were the external primers flanking the editing template. The PCR revealed a band size corresponding to the combined size of the two flanking sequences.

The PCR product from above and the sgRNA containing plasmid were both digested with SfiI and following purification via gel electrophoresis, the two fragments were ligated, and the resulting reaction mixture was used directly to transform E. coli cells. Several plasmids were isolated from E. coli, digested with SfiI and sequenced using primers RH001_SNB and RH002_SNB (Table S2).

The plasmids were used to transform B. subtilis Δspβ and transformants were plated on agar plates supplemented with 0.4% mannose and kanamycin at 30 °C and incubated for 2 d. Mannose induces the Cas9 gene in the plasmid that is under the control of a mannose promoter. Only the large colonies were incubated on a new plate at 45–50 °C without kanamycin. This incubation was performed to cure the plasmid from the cells. The cells were then grown at 42 °C in liquid medium overnight and the culture was diluted and successively inoculated into a fresh culture the following morning. This process was repeated three times. Cells were then plated on agar plates without kanamycin and incubated at 42 °C. Individual colonies were then grown and streaked on plates with and without kanamycin to confirm loss of resistance to the antibiotic. Positive constructs were further checked for resistance to sublancin. Finally, colony PCR was performed using specific primers to check the absence of the gene [ptsX_cPCR _F and ptsX_cPCR_R (X = G, H, I or GHI)] (Table S2) and then the PCR product was sequenced using appropriate primers [ptsXdelSP_F or ptsXdelSP _R (X = G, H or I)] (Table S2) and the new strain was saved as a glycerol stock. For a successful deletion mutant, the length of the PCR product from the mutant was shorter than the length of the PCR product obtained from the wild type as the control. All sequencing results are summarized in Table S3.

Complementation studies in B. subtilis Δspβ and MIC determination

The deletion strains obtained as described above were transformed with appropriate plasmids and sensitivity to sublancin was determined. Transformation of B. subtilis Δspβ ΔptsG with the plasmid encoding ptsG and induction with xylose restored sensitivity to sublancin but transformation with plasmids encoding ptsH or ptsI followed by xylose induction had no effects on the activity of sublancin. The B. subtilis Δspβ ΔptsGHI mutant was resistant to sublancin under all the tested conditions. MICs for these strains was determined as described above.

Glucose protection assay

Cultures of B. subtilis Δspβ were grown to OD₆₀₀ 1.0 and diluted to OD₆₀₀ 0.2 in fresh refined LB. Then 100 μL of culture was dispensed into each well of a 96-well plate. Sublancin or sublancin-xylose was added to each well at 1x MIC and incubated at 37 °C for 20 min. Then, glucose (final concentration of 2%) or sterile water was added to each well. OD₆₀₀ readings were obtained after 7 h using a Synergy H4 Hybrid Multi-Mode Microplate BioTeK plate reader. Untreated cell culture was used as the negative control. The same experiment was repeated three times with qualitatively similar results.

Sublancin-resistant mutants sample preparation and DNA sequencing

B. subtilis Δ6 strain was used in this study to generate resistant mutants under conditions where the PTS is required because it carries the chloramphenicol resistance gene marker. This marker avoids contamination during the growth of resistant strains, especially in glucose M9 media where it takes a long time for resistance mutants to grow. An overnight culture of B. subtilis Δ6 was inoculated either in LB containing 4x MIC sublancin (2.5 μM) and 5 μg/mL chloramphenicol, or in glucose minimal media (M9 minimal salts 1x, 2% glucose, 0.2% potassium glutamate, 11 μg/mL iron ammonium citrate, 3 mM MgCl2 and tryptophan at a final concentration of 50 μg/ml) containing 4x MIC sublancin (2.5 μM) and 5 μg/mL chloramphenicol. Each culture was grown at 37 °C until it turned cloudy, and the same procedure was repeated three times to ensure the resistance of the liquid culture. The last batch of liquid culture was streaked on agar plates containing LB or glucose minimal media with 4x MIC sublancin (2.5 μM) and 5 μg/mL chloramphenicol. Four single colonies from the LB plates and three single colonies from the glucose minimal media plate were picked and grown in liquid media with antibiotics, and their genomes were sequenced.

For obtaining resistant mutants in strains overexpressing PTS proteins, these strains were treated with sublancin at 4x MIC following xylose induction as described above and growth was continued for 48 h when cultures showed turbidity. Cultures were allowed to grow for an additional 24 h before the cells were plated on LB agar plates supplemented with 5 μg/mL chloramphenicol, 2% xylose and 4x MIC sublancin. Two random colonies were picked from each plate and grown in liquid media with xylose, appropriate antibiotics and sublancin.

The genomic DNA from each culture was purified using PureLink^™ Genomic DNA Mini Kit. The purity of the extracted genomic DNA was determined by running a DNA gel. NovaSeq 6000 shotgun sequencing was performed on seven genomic DNA samples at the Roy J. Carver Biotechnology Center at UIUC. Subsequent bioinformatic analysis was performed by the Roy J. Carver Biotechnology Center at UIUC.

Overexpression of FliQ in Bacillus and MIC determination.

The pHCMC04-His6FliQ plasmid was used to transform B. subtilis Δspβ or the strains in which pts genes were deleted. Transformants were plated on a LB agar plate containing 5 μg/mL chloramphenicol. The culture was grown in LB with 5 μg/mL chloramphenicol for 4 h and induced with 2% xylose when OD600 reached 0.5. After shaking at 37 °C for additional 3 h, the culture was collected. A 96-well microtiter plate was prepared as described previously to determine the MIC values.

Strain preparation for microscopy

pRSF Duet was linearized using the primer pairs pRSF_BB_ptsX/SunI_F and pRSF_BB_ptsX/SunI_R ((Table S2, X=G, H or I) and the genes ptsG, ptsH, ptsI and sunI were amplified from B. subtilis Δspβ using colony PCR and the primer pairs, ptsX/SunI_pRSF_F and ptsX/SunI_pRSF_F (Table S2). The amplified fragments were joined together using Gibson assembly to encode N-terminally histidine-tagged Pts proteins or SunI in the plasmid. Using these initial plasmids as template, the genes ptsG, ptsH, ptsI and sunI were amplified again with primer pairs pHCMC04_ptsX/SunI_T_F and pHCMC04_ptsX/SunI_T_R (Table S2) and the PCR fragments were assembled with linearized pHCMC04 (pHCMC04_BB_T_F and pHCMC04_BB_T_F; Table S2) via Gibson assembly to obtain plasmids encoding proteins with hexa-histidine tag for expression in Bacillus. These plasmids were subsequently linearized with primer pairs (color_protein_BB_T_R and color_protein_BB_T_F; color =yellow, green, cyan or red, protein = PtsG/PtsH/PtsI or SunI). All genes encoding fluorescent proteins were PCR amplified from appropriate backbones using the primers color_T_F and color_T_R.

The primers were designed such that pHCMC04 plasmid retained the TrpAt downstream of the MCS and each N-terminally His-tagged protein carried a GSSGSS linker between the PTS protein or SunI and the corresponding fluorescent fusion. Gibson assembly was performed to ligate each gene with the reporters and the plasmid backbone. Following the isolation of the plasmid from E. coli and verifying the sequence via Sanger sequencing, the plasmids were used to transform chemically competent B. subtilis Δspβ following standard protocols as described above.

Sublancin treatment of B. subtilis Δspβ cells overexpressing PtsG, PtsH, or PtsI

After the transformation of B. subtilis Δspβ with the desired plasmid, cells were plated on agar plates containing chloramphenicol at a concentration of 5 μg/mL and incubated for 2 d at 30 °C. Only single large and amorphous colonies were selected for overnight incubation in LB medium at 37 °C. Cells were suspended the following morning in refined LB medium supplemented with 5 μg/mL chloramphenicol and xylose was added at a concentration of 2% after an incubation period of 2–3 h to induce the expression of the appropriate pts gene (20 μL of 20% xylose was added to 180 μL of culture). After 1 h, sublancin was added to the cells at a concentration of 312 nM. OD₆₀₀ readings were obtained using a Synergy H4 Hybrid Multi-Mode Microplate Biotek plate reader. Readings were recorded at an interval of 15 min for the first 6 h, following which readings were taken at 1 h intervals. Growth curves were plotted and analyzed using Origin Pro 2018 software. Experiments were performed three independent times.

Expression of the PTS proteins was checked on a 4–20% gradient gel (Figure S8). Since the proteins were not tagged with any epitope tags, it was difficult to ascertain the overexpression of the PTS genes from the gel with confidence. To overcome this problem, the PTS genes were cloned into pHCMC04 such that they carried a histidine tag and western blot analysis was performed with anti-His-tag antibodies to confirm overexpression of the proteins. Western blot analysis was performed as described elsewhere.⁵⁴ Blotting was performed after 1 h and 2 h of overexpression using mouse monoclonal anti-His antibody from Genscript (1:4000) (THE^™ His Tag mAb; Cat. No. A00186) and then goat anti mouse HRP (1:20000) (ab205719) from Abcam (Figure S9). The His-tagged PTS genes were still functional as they complemented the respective deletion mutants with respect to sublancin sensitiviy (Table S7). As control experiments, cells carrying the empty plasmid and cells transformed with plasmids encoding the PTS genes were grown with and without inducer and no growth defects were observed in the absence of sublancin (Figure S10).

Expression and complementation of fluorescent proteins in B. subtilis Δspβ

Expression of the various fluorescent proteins was checked using a Promega Glomax Plate reader equipped with appropriate filters and lasers for monitoring fluorescence from GFP and its derivatives as well as RFPs such as mCherry. Briefly cells from an overnight culture were diluted 1:100 in liquid media and induced with 2% xylose when the OD reached 0.3. Following induction, cells were spun down and resuspended in PBS, washed three times with PBS and a 10-fold dilution of the cells was prepared in PBS and the fluorescence intensity was measured using a plate reader. Uninduced samples served as controls. All readings were recorded in triplicates (Figure S11). Functionality of the fluorescent protein fusions was established using complementation assays as previously described (Table S8).

Synthesis of sublancin-cyanine 5 (Cy5)

HPLC-purified sublancin was dissolved in DMF at an initial concentration of 2 mM and 20 mM propyl phosphonic anhydride (T3P), 20 mM hydroxy benzotriazole (HOBt) and 20 mM of cyanine-5-amine (Cy5-amine) were added and the reaction was incubated for 18 h at 25 °C on an orbital shaker in amber tubes. After drying the reaction on a speed-vac, the solid residue left behind was dissolved in water and filtered using a 0.22 μm syringe filter, and the supernatant was purified on a preparative Agilent 1260 Infinity II HPLC system equipped with a Macherey-Nagel NUCLEODUR® 100–5 μM C18 ec column (250 mm × 4.6 mm; catalog no. 760002.46). Water containing 0.1% TFA was used as solvent A while solvent B was acetonitrile containing 0.1% TFA. An elution gradient from 0% solvent B to 100% solvent B over 30 min at a flow rate of 4 mL/min was used, and sublancin labeled with Cy5-amine was eluted at 60–62% solvent B. Fluorescent fractions were pooled and after lyophilization, 2 mg of purified sublancin-Cy5 (Figure S2) was obtained, which was used in microscopy experiments.

Fluorescence microscopy

Slide preparation: Cultures of B. subtilis Δspβ transformed with pHCMC04-PtsG_YFP, pHCMC04-PtsH_GFP, pHCMC04-PtsI_CFP, or pHCMC04-SunI_RFP plasmids were induced with 2% xylose and grown in LB containing 5 μg/mL chloramphenicol overnight. Standard microscope slides coated with poly-D-lysine (R&D Systems, 34-392-000101) overnight were washed with water and sterilized in ethanol. Overnight bacterial culture was centrifuged at 4,500×g and the pellet was resuspended gently three times in fresh PBS to remove the media background. Then 20 μL of cell culture in PBS and 180 μL of PBS was loaded on each slide and incubated at room temperature for 30 min. Next, 1 mL of PBS was used to wash the slides at least five times to remove cells. The final concentration of 200 nM sublancin-Cy5 in 200 μL of PBS was applied on each slide and incubated at room temperature for 30 min. PBS was used to wash the slides at least five times to wash away unbound sublancin. A final concentration of 10% formaldehyde (Thermo Scientific, PI28906) and 2.5% glutaraldehyde (Fisher Chemical, O2957–1) in PBS was applied to each slide and incubated for 30 min. Three times 1 mL of PBS was applied to wash away excess fixing solution. A drop of ProLong^™ Glass Antifade Mountant (Invitrogen, P36982) was applied to each slide, covered with a glass coverslip (Ibidi, NC0601315) and dried overnight protected from light.

For competition assays (Figure S7), cultures of B. subtilis Δspβ were grown in refined LB overnight. Standard microscope slides coated with poly-D-lysine (R&D Systems, 34-392-000101) overnight were washed with water and sterilized in ethanol. Overnight bacterial culture was centrifuged at 4,500×g and the pellet was resuspended gently three times in fresh PBS to remove the media background. Then 20 μL of cell culture in PBS and 180 μL of PBS was loaded on each slide and incubated at room temperature for 30 min. Next, 1 mL of PBS was used to wash the slides at least five times to remove cells. A final concentration of 200 nM sublancin or sublancin-Cy5 in 200 μL of PBS was applied on each slide and incubated at room temperature for 15 min. PBS was used to wash the slides at least ten times to wash away unbound sublancin. The same process was then repeated but with 200 nM sublancin-Cy5 for the slides that were initially incubated with wild type sublancin. Similarly, the process was repeated but with type sublancin for slides that were initially treated with sublancin-Cy5. The slides were again incubated for 15 min followed by ten washing steps. A final concentration of 10% formaldehyde (Thermo Scientific, PI28906) and 2.5% glutaraldehyde (Fisher Chemical, O2957–1) in PBS was applied to each slide and incubated for 30 min. Three times 1 mL of PBS was applied to wash away excess fixing solution. A drop of ProLong^™ Glass Antifade Mountant (Invitrogen, P36982) was applied to each slide, covered with a glass coverslip (Ibidi, NC0601315) and dried overnight.

All micrographs were captured using a Zeiss LSM 880 Airyscan microscope (Carl R. Woese Institute for Genomic Biology, UIUC). The 63× phase-contrast objective (oil immersion) was used for image capturing, in combination with ZEN 2.6 software (Zeiss) to control the microscope setup and to perform the imaging of cells. The following visible excitation lines were used: 488 nm to visualize GFP, 514 nm to visualize YFP, 458 nm to visualize CFP, 561 nm to visualize RFP (mCherry), and 633 nm to visualize Cy5. Images were processed with ZEN 2.6 software (Zeiss).

ABBREVIATIONS

PTS

phosphoenolpyruvate:sugar phosphotransferase system

RiPPs

ribosomally synthesized and post-translationally modified peptides

MRSA

methicillin resistant Staphylococcus aureus

VRE

vancomycin resistant Enterococcus sp.

BGC

Biosynthetic gene cluster

LB

lysogeny broth

CDM

chemically defined medium

WT

wild type

Cy5

cyanine 5

CFP

cyan fluorescent protein

GFP

green fluorescent protein

YFP

yellow fluorescent protein

RFP

red fluorescent protein