A screen for mutants defective in B. anthracis sporulation

To screen for sporulation-defective mutants by transposon-insertion sequencing (Tn-seq) [24], we built a plasmid-based transposon delivery system for B. anthracis modeled after one used previously for B. subtilis [25]. This plasmid harbors a temperature-sensitive replicon, a hyperactive allele of the mariner-Himar1 transposase, and a spectinomycin resistance cassette flanked by inverted repeats recognized by the transposase (see Materials and methods). An MmeI restriction site was engineered into one of the inverted repeats to aid in the preparation of the sequencing library. Using this system, we generated a library with >10⁶ transposants in the plasmid-less Sterne strain 9131 and mapped the insertion sites by next generation sequencing (see Materials and methods). The library contained >170,000 unique insertions that were well distributed across the genome (Fig 1A), with a Tn insertion, on average, every 35 bp. Using EL-ARTIST [26] to analyze the results of this Tn-seq experiment, we identified essential genes in the parental strain. Of the 5,456 genes in the B. anthracis genome, 400 were predicted to be essential for growth in BHI medium (S1 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. A Tn-seq screen identifies a comprehensive set of genes required for sporulation in B. anthracis.

(A) Transposon-insertion profile across the entire B. anthracis genome of the initial transposon library. Each line indicates a transposon-insertion site in the genome and its height reflects the relative number of insertions mapped to that site. The maximum read height for panels (A) and (B) was set to the number shown on the right. (B) Schematic of the sporulation screen. The B. anthracis transposon library was grown in PA liquid medium. At the onset of starvation, a sample was removed, and the remaining culture was sporulated. The sporulated culture was incubated at 65 °C for 30 min to kill vegetative cells and sporulation-defective mutants. The surviving spores were germinated and outgrown on BHI agar plates and the resulting colonies pooled. Transposon-insertion sites and their abundance were determined by Tn-seq. Transposon-insertion profiles from three regions of the genome are depicted. Transposon-insertions in sigG (spoIIIG), sigE (spoIIGB), spoIIGA, yyaC, and ipdA were underrepresented after sporulation, germination, and outgrowth. Insertions in the essential ftsZ gene were not tolerated under either condition. (C) Scatterplot highlighting all the potential B. anthracis sporulation genes identified in the Tn-seq screen. Every gene in the B. anthracis genome is plotted on the graph. The Y-axis is the log fold-reduction in transposon-insertions at the onset of starvation relative to the insertions after sporulation, heat treatment, germination, and outgrowth. The X-axis is the –log of the Mann–Whitney U p-value. The statistical significance thresholds (p ≤ 0.05, fold-reduction ≥3) are indicated by dashed red lines. B. anthracis genes that encode B. subtilis homologs that are required for efficient sporulation in B. subtilis are shown in magenta. B. anthracis genes that meet the statistical significance thresholds but are not required for sporulation in B. subtilis are highlighted in yellow. B. anthracis genes that meet the statistical thresholds for which there are no B. subtilis homologs are highlight in cyan. All candidate sporulation genes in B. anthracis, including those outside the statistical significance thresholds, are listed in S2–S6 Tables. The data underlying panel C can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003521.g001

We used the Tn library to screen for genes required for spore formation, heat-resistance, germination, and outgrowth. The library was grown in Phage Assay (PA) medium and at the onset of starvation, an input sample was collected and the remaining culture was sporulated by nutrient exhaustion (Fig 1B). The culture was harvested seven hours after starvation, a time when Sterne strain 9131 produced the maximum number of heat-resistant spores (see Materials and methods). The culture was then incubated at 65 °C for 30 min to kill vegetative cells and sporulation-defective mutants. The spores were plated on BHI agar plates and ~10⁶ colonies from the germination-proficient spores were pooled. Finally, the transposon-insertions at the onset of starvation and after sporulation, heat treatment, germination, and outgrowth (for simplicity, referred to as sporulation) were mapped by next generation sequencing. The location and number of reads for each transposon-insertion were compared between the input and output samples (Fig 1B and 1C). We identified >200 genes in which transposon-insertions were statistically (Mann–Whitney U; p ≤ 0.05) underrepresented by ≥3-fold after sporulation compared to the onset of starvation (Fig 1C and S2–S6 Tables). Visual inspection of the transposon-insertion profiles revealed that ~22% of these genes had relatively few insertions at the onset of starvation and most of these were not considered further. In total, our screen identified 159 putative sporulation genes in B. anthracis.

Similarities between B. anthracis and B. subtilis sporulation

In B. subtilis, ~142 genes are required for efficient sporulation [17]. Our Tn-seq screen identified 79 of these as putative sporulation genes in B. anthracis (magenta circles in Fig 1C and S2–S6 Tables). Among the 142 B. subtilis sporulation genes, 74 have “spo” names and 56 of these display strong sporulation defects (≤5% sporulation efficiency compared to wild-type (WT)) when mutated. Our B. anthracis sporulation Tn-seq screen identified 38 of these (S2 Table). Fig 1B shows the Tn-seq profile from a small region of the B. anthracis genome that contains three of these spo genes, spoIIGA, spoIIGB (sigE), and spoIIIG (sigG). Fig A in S1 Text contains Tn-seq profiles that highlight 11 additional B. anthracis spo genes. Of the 56 B. subtilis spo genes with strong sporulation defects, 18 were not identified in our B. anthracis Tn-seq screen. Two barely missed our cutoff (spoIIAB, spoVM) and two (spo0A, spo0H) had almost no insertions at the onset of sporulation. Eleven have paralogs in B. anthracis (e.g., spoIIP, spoIIIJ, and spoIVH) and might require the loss of both genes to impair sporulation. In addition, the sigK gene that encodes pro-SigmaK, is not interrupted by an intervening (skin) element in B. anthracis [27]. Accordingly, B. anthracis has a single sigK gene rather two gene fragments (spoIVCB, spoIIIC) and lacks spoIVCA that encodes the recombinase that excises the skin element during sporulation. This accounts for two more missing genes. The final gene was spoIISB which encodes an anti-toxin that inhibits the SpoIISA toxin during sporulation in B. subtilis [28]. Thus, SpoIISA is unlikely to impair sporulation in B. anthracis. We conclude that most of the core sporulation factors in B. subtilis are required for spore formation in B. anthracis.

B. subtilis has an additional 19 genes that do not have spo names but, when mutated, have strong sporulation defects (≤5% sporulation efficiency compared to WT). Among these, 10 were hits in our B. anthracis screen (S3 Table). Of the remaining 9, one barely missed the cutoff (gerM); one (ricT) had insertions exclusively at the 3′ end of the gene after sporulation; two (ricA, ricF) had virtually no insertions at the onset of sporulation; and three (asnO, yqhT, and ccdA) have putative paralogs. The remaining two genes (fin and putB) were unambiguously not underrepresented in the B. anthracis screen, suggestive of real differences in sporulation between B. anthracis and B. subtilis. Finally, similar to B. subtilis, a subset of the genes in the TCA cycle were hits in the B. anthracis screen, while several others were essential (S4 Table).

Altogether, these findings suggest that the core machinery that drives spore formation is hard-wired in both the model and pathogen. We note that 63 of the B. subtilis sporulation genes with more mild mutant phenotypes were not identified in our screen. Although some of these genes are not present in B. anthracis (yerC, skfE, and skfF), the failure to identify the others (e.g., racA, sirA, fisB, spoVIGA, and spoVIGB) suggests that our screen was not sensitive enough to detect them or that they are not critical for sporulation in B. anthracis.

Differences between B. anthracis and B. subtilis sporulation

Our Tn-seq screen identified 58 putative sporulation genes in B. anthracis that are present in B. subtilis but have not been reported to be required for efficient sporulation in the model (yellow circles in Fig 1C and S5 Table). None of these genes were identified as hits in a previous B. subtilis sporulation Tn-seq screen [17]. Among the 58 genes, 25 of them are induced during B. subtilis sporulation but have modest or no reported phenotypes [16,29,30]. The identification of the B. anthracis homologs in our screen, underscores species-specific differences and highlights the importance of investigating sporulation in the pathogen. Fig 1B shows the Tn-seq profile from a small region of the B. anthracis genome that contains one of these B. anthracis genes, yyaC. Fig B in S1 Text compares Tn-seq profiles from regions of the B. anthracis and B. subtilis genomes that contain five hits (yyaC, yqfT, yhfN, ytxC, and ysxE) in the B. anthracis screen but not the B. subtilis screen.

Finally, our Tn-seq screen uncovered 22 putative sporulation gene in B. anthracis that are not present in the B. subtilis genome (cyan circles in Fig 1C and S6 Table). One such gene, locus 04176 (renamed ipdA), is illustrated in Fig 1B and two additional examples are shown in Fig C in S1 Text.

Validation of hits from the screen

To experimentally validate hits from the sporulation screen, we constructed an ordered Tn mutant library using the Knockout Sudoku method [31] (See Materials and methods). Briefly, ~52,000 transposants from our Tn library were arrayed in 96-well plates and saturated cultures from the arrayed mutants were distributed combinatorially into 47 distinct pools. These pools were sequenced and individual Tn mutants mapped to their unique well in the arrayed library based on the pools in which the mutant was identified. We then cherry-picked a mutant from all represented genes in the library to generate a condensed collection. This nonredundant quality-controlled collection contains 3,826 Tn mutants that represent ~70% of the 5,456 genes in the B. anthracis genome.

To test the hits from our screen, we transduced the Tn mutations into Sterne strain NR-9401. This strain sporulates asynchronously, but the total number of heat-resistant spores that can be recovered after 30 hours of nutrient exhaustion is approximately 10-fold higher and more reproducible than the plasmid-less strain we used to build the ordered mutant library (see Materials and methods). In total, we tested the sporulation efficiency of 71 hits from our screen. Of these, 62 were validated, with sporulation efficiencies ranging from <0.0001% to 40%. Forty-three of these had sporulation efficiencies of ≤5% compared to WT. The complete dataset can be found in S2–S6 Tables.

Analysis of conserved sporulation genes

Our data indicate that homologs of many of the well-characterized B. subtilis sporulation genes are important for efficient sporulation in B. anthracis (S2 Table). To investigate whether these factors play similar roles in the morphological process of sporulation, we optimized conditions to visualize B. anthracis sporulation by fluorescence microscopy. Fig 2A shows a comparison of a sporulation time course of WT B. subtilis and the plasmid-less B. anthracis Sterne strain. Other than the striking difference in cell size, the morphological process of sporulation was similar between model and pathogen. To monitor gene expression in the mother cell and forespore compartments, we generated transcriptional reporters in which the spoIIQ promoter was fused to yfp and the spoIID promoter was fused to cfp. In B. subtilis, spoIIQ is expressed under the control of the first forespore-specific transcription factor σ^F and spoIID is controlled by the mother cell-specific transcription factor σ^E [32,33]. As can be seen in Fig 2B and Fig D in S1 Text, the B. anthracis promoters have similar compartment-specific expression patterns to those in B. subtilis. We selected five genes (spoIIE, spoIIR, spoIID, spoIIQ, and spoIIIAE) that have been extensively characterized in B. subtilis and analyzed the B. anthracis Tn mutants during sporulation by fluorescence microscopy.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Conserved B. anthracis sporulation genes have mutant phenotypes that are similar, but not identical, to B. subtilis mutants.

(A) Representative fluorescence and phase-contrast micrographs of sporulating B. anthracis and B. subtilis. Wild-type strains were induced to sporulate by nutrient exhaustion at 37 °C in PA (B. anthracis) or DSM (B. subtilis). Time after the onset of sporulation is indicated in the top left corner of the B. anthracis images. The membranes were stained (0–5 hour) with TMA-DPH. (B) Representative fluorescence images of the indicated B. anthracis mutants at hour 4 of sporulation. All strains harbor forespore (P_spoIIQ-yfp) and mother cell (P_spoIID-cfp) transcriptional reporters. σ^F activity in the forespore (SigF) and σ^E activity in the mother cell (SigE) were visualized with YFP (false-colored green) and CFP (false-colored red), respectively. The membranes were stained with TMA-DPH. Disporic cells in spoIIR::Tn (yellow carets), impaired engulfment in spoIID::Tn and spoIIQ::Tn (white carets), and small forespores in spoIIIAE::Tn (light blue carets) are highlighted. Scale bars indicate 5 μm. Images from a complete sporulation time course for each mutant can be found in Figs D and E in S1 Text.

https://doi.org/10.1371/journal.pbio.3003521.g002

Sporulation genes that are important in B. anthracis but are not critical in B. subtilis

Fifty-eight of the B. anthracis genes that were hits in our sporulation screen have B. subtilis homologs that are not required for efficient sporulation in the model. We chose seven (yyaC, yhfN, yqfT, yuzB, ysxE, ytxC, and murAB) to characterize further. We focused on genes that are induced during sporulation in B. subtilis [16,30] but have not been reported to have mutant phenotypes. We anticipated that the B. anthracis Tn mutants might have interesting morphological defects, and we liked the idea that studies on the pathogen could inform the model.

The B. subtilis yyaC gene is expressed under the control of σ^F in the forespore and encodes a soluble protein that is predicted to be in the peptidase HybD-like domain superfamily (IPR023430). Foldseek [49] predicts that YyaC is structurally similar to the B. subtilis germination protease Gpr that is expressed at a later stage in sporulation [50]. B. anthracis has a true Gpr homolog, suggesting that, if YyaC is a protease, it has a different function. Analysis of the yyaC::Tn mutant by fluorescence microscopy revealed that at a late stage in sporulation the forespores become shrunken with large membrane invaginations (Fig 3 and Fig Fa in S1 Text). These defects are similar to the “collapsed forespore” phenotype in B. subtilis spoIIQ and spoIIIA mutants [48] and raise the possibility that YyaC-mediated degradation of forespore proteins is critical to maintain metabolic potential of the developing spore. We also observed increased cell lysis at late stages of sporulation in the yyaC::Tn mutant (Fig 3), suggesting that loss of forespore metabolic potential leads to mother cell lysis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Genes that are important for sporulation in B. anthracis but are not critical for spore formation in B. subtilis.

Representative fluorescence and phase-contrast images of the indicated B. anthracis mutants at hour 4 of sporulation. All strains have forespore and mother cell transcriptional reporters. The membranes were visualized with TMA-DPH. Phase-contrast images were from hour 5.75 of sporulation. The sporulation efficiency of each mutant, determined in the BaR6 background, is shown in the bottom right corner of the phase-contrast images. Collapsed or distorted forespores (yellow carets), spherical forespores (light blue carets) are highlighted. White carets highlight sporangia that have CFP and YFP in both compartments. Scale bars indicate 5 μm.

https://doi.org/10.1371/journal.pbio.3003521.g003

A second mutant, yqfT::Tn, also had collapsed forespores. However, these forespores had abnormal and heterogeneous morphologies and appeared to lose the ability to maintain compartmentalization (Fig 3 and Fig Fa in S1 Text). At late stages of differentiation, the σ^F and σ^E fluorescent reporters were present in both compartments of the sporangia. The B. subtilis yqfT gene is expressed under σ^E control and encodes a soluble protein of unknown function. How this mother cell protein supports forespore development in B. anthracis remains to be discovered.

One of the B. anthracis mutants, yuzB::Tn, was impaired in entry into sporulation (Fig Fa in S1 Text) and 4 hours after nutrient exhaustion only a small subset of cells had polar septa (Fig 3). yuzB is not expressed during sporulation in B. subtilis and encodes a small soluble protein. Foldseek predicts that YuzB has a thioredoxin fold, suggesting this protein could be involved in maintaining the cellular redox state as B. anthracis exhausts its nutrients. Two of the mutants we analyzed, yhfN::Tn and ytxC::Tn, produce dormant spores that were more spherical than WT (Fig 3 and Fb in S1 Text). Aside from this subtle phenotype, the mutants appeared to undergo differentiation normally. Several of the mutants, exemplified by ysxE::Tn, produced normal-looking spores yet had a strong sporulation defect (Fig 3 and Fb in S1 Text). These mutant spores fail to mature and could have defects in DPA accumulation, core dehydration, or germination.

Finally, we analyzed the B. anthracis murAB::Tn mutant. MurAB and its essential paralog MurAA catalyze the first committing step in PG precursor synthesis [51]. A recent B. subtilis study, using a sensitized background that is partially impaired in engulfment, reported a more pronounced engulfment defect when murAB was also deleted [52]. However, B. subtilis sporulating cells that only lack murAB had no phenotype. Interestingly, the B. anthracis murAB::Tn mutant was severely impaired in sporulation (Fig G in S1 Text). The nutrient-exhausted cells had morphological defects at early stages of sporulation, followed by impaired polar division, defects in engulfment, and spore lysis (Fig G in S1 Text). These data suggest that MurAB is more critical for cell wall synthesis during starvation and sporulation and could reflect a greater division of labor between the two paralogs in B. anthracis compared to B. subtilis.

Altogether, our initial characterization of the B. anthracis sporulation genes that have B. subtilis homologs with weak or undetectable mutant phenotypes indicates that there are several additional factors required for sporulation in B. anthracis. Since 25 of these genes are induced during sporulation in B. subtilis, our findings suggest that some of these factors may play similar roles in B. subtilis, but alternative genes or pathways exist that compensate for their absence. Synthetic lethal sporulation screens with the B. subtilis mutants could be a fruitful path to revealing their function. Alternatively, enrichment screens in B. anthracis (see below) could guide functional characterization.

04176 (IpdA) is required for the morphological process of engulfment

Finally, we turned out attention to a putative sporulation factor 04176 that is unique to B. anthracis. Based on the experiments described below, we have renamed this protein IpdA for Inhibitor of polysaccharide deacetylases. The ipdA::Tn mutant is severely impaired in sporulation (<0.01% of WT) (Fig 4B and S6 Table) and fluorescence microscopy indicates that ipdA::Tn mutant cells are engulfment-defective and have septal membrane bulges that protrude into the mother cell cytosol (Fig 4A and Fig H in S1 Text). This phenotype is reminiscent of B. subtilis mutants lacking the engulfment cell wall hydrolases SpoIID (Fig 4C) or SpoIIP [41,42,53]. Importantly, expression of ipdA in trans restored efficient sporulation and normal engulfment to the ipdA::Tn mutant, confirming that the defects were attributable to the loss of IpdA (Fig 4A and 4B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. IpdA (04176) is required for the morphological process of engulfment.

(A) Representative fluorescence images of the indicated B. anthracis strains at hour 4 of sporulation. Wild-type (WT) and ipdA::Tn harbor a σ^F-responsive promoter (P_spoIIQ) fused to yfp (false-colored green). The complementation strain (ipdA::Tn + ipdA) lacks the transcriptional reporter. The membranes were stained with TMA-DPH. Yellow carets highlight bulged septal membranes in the ipdA::Tn mutant. (B) Sporulation efficiency of the indicated strains in the BaR6 background. (C) Representative fluorescence images of the indicated B. subtilis strains at hour 4 of sporulation. Bulged septal membranes in the ∆spoIID mutant are highlighted (yellow carets). A complete sporulation time course of the B. subtilis strains can be found in Fig I in S1 Text. (D) Superimposition of the AlphaFold-predicted structure of IpdA (white) and the experimentally determined structure of IseA (magenta) (PDB:2RSX). The root mean squared displacement (RMSD) is 3.9 Å. (E) Representative fluorescence images of sporulating B. anthracis cells harboring an ipdA transcriptional reporter (P_ipdA-yfp). Time after the initiation of sporulation is indicated in the bottom left corners. Membranes were stained with TMA-DPA. White carets highlight cells with YFP fluorescence prior to polar division. YFP images were scaled identically. A complete time course can be found in Fig J in S1 Text. (F) Representative images of the indicated B. anthracis strains at hour 4 of sporulation. Yellow carets highlight septal bulges in the ∆ipdA mutant. A complete time course can be found in Fig H in S1 Text. Scale bars indicate 5 µm. The data underlying panel B can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003521.g004

IpdA is a small soluble protein with a predicted signal peptide. It does not have a B. subtilis homolog based on a BLAST homology search. However, Foldseek predicts that IpdA is structurally similar to B. subtilis IseA (Fig 4D), a secreted inhibitor of cell wall hydrolases that contain NlpC/P60 catalytic domains [54–56]. Neither IseA nor the NlpC/P60 peptidoglycan hydrolases have been implicated in sporulation in B. subtilis (Fig Ia in S1 Text) and our analysis of an iseA mutant indicates that sporulation is unaffected in its absence (Fig 4C and Fig Ib in S1 Text).

To investigate whether B. anthracis ipdA is induced during sporulation, we fused its promoter to yfp and monitored fluorescence during a sporulation time course. As can be seen in Fig 4E, ipdA was induced at an early stage of sporulation, prior to asymmetric division (Fig 4E and Fig J in S1 Text). Consistent with the idea that ipdA is controlled by the response-regulator Spo0A, its promoter region contains four near-consensus Spo0A binding motifs [57]. By contrast, analysis of an iseA transcriptional reporter revealed that it is not specifically induced during B. subtilis sporulation (Fig J in S1 Text), consistent with our findings that it is not required for successful sporulation.

B. subtilis IseA inhibits NlpC/P60 cell wall hydrolases by binding the catalytic groove of these enzymes (Fig Ka in S1 Text) [54]. Importantly, several of these enzymes are required for cell separation after cytokinesis. Based on the structural similarity between IseA and IpdA, our initial hypothesis was that IpdA also inhibits PG hydrolases that function in cell separation during vegetative growth. In this model, the absence of IpdA leads to inappropriate septal PG cleavage after polar division resulting in septal bulges. To investigate this idea, we analyzed sporulating cells lacking the mother cell transcription factor σ^E in the presence and absence of IpdA. In B. subtilis, σ^E controls the expression of the engulfment cell wall hydrolases SpoIID and SpoIIP [43]. If IpdA inhibits vegetative cell wall hydrolases then sporulating cells lacking σ^E and IpdA should have septal bulges, while the matched ipdA+ cells should not. As can be seen in Fig 4F and Fig H in S1 Text, no septal bulges were observed in the sigE::Tn mutant in the presence or absence of IpdA. These findings suggest that IpdA does not inhibit the vegetative PG hydrolases and instead is required to promote or enable the cell wall degrading activities of the engulfment enzymes. Consistent with the idea that IpdA does not inhibit NlpC/P60 cell separases, AlphaFold-multimer (AF-M) did not predict interactions between IpdA and any of the B. anthracis NlpC/P60-containing PG hydrolases (Fig Kb in S1 Text).

Evidence that IpdA inhibits the polysaccharide deacetylase PdaN to promote engulfment

To investigate IpdA function, we performed a one-by-all AF-M screen, in which IpdA was folded pairwise with each protein in the B. anthracis proteome. Five independent models were generated for each prediction and the confidence metrics were analyzed using a published pipeline [58]. An average model score was assigned to each prediction based on how well the predicted contacts agreed across the five models, and the average predicted alignment error (pAE) across the contact region in the best model was measured (S7 Table). Hits with an average model score ≥3 and an average pAE ≤ 5 Å were considered strong candidates. Forty proteins met these criteria (Fig 5A) of which 12 were predicted to be secreted or membrane anchored (S7 Table). In a complementary approach, we performed a classical genetic enrichment for suppressors of the ∆ipdA mutant. Four independent cultures of the mutant were subjected to cycles of sporulation, heat-kill, and reinoculation into sporulation medium. After 3–5 cycles, the sporulation efficiency of each culture increased >1000-fold. A single colony from each was re-tested (Fig 5B) and the suppressor mutations identified by whole–genome sequencing. Three of the suppressors mapped to gene 04184 and one mapped to gene 02960 (Fig 5C). The 04184 protein is a member of the YkyA protein family (PF10368) that is annotated as a putative cell wall binding lipoprotein, although we could not find experimental evidence for PG binding. Based on our characterization of 04184, we have renamed the gene pdcF for polysaccharide deacetylase co-factor. 02690 is homologous to a putative secreted polysaccharide deacetylase YheN (or PdaN) in B. subtilis [59] and we used the PdaN nomenclature for 02690 (Fig 5C). Previous biochemical studies of B. anthracis PdaN indicate that it can deacetylate N-acetyl-glucosamine (GlcNac) in vitro [60]. Importantly, PdaN was among the strongest hits in our AF-M screen (Fig 5A). Furthermore, three other polysaccharide deacetylases (00373, 05406, and 02021) [61] were also strong hits (Fig 5A). We validated pdcF and pdaN as suppressors of ∆ipdA using the Tn-insertion mutants (Fig 5B). Both loss-of-function mutants improved sporulation efficiency by >1000-fold.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. IpdA inhibits the polysaccharide deacetylase PdaN (02960).

(A) Scatterplot of all potential hits (circles) from a one-by-all AlphaFold2-Multimer screen using IpdA as the bait. Five models were predicted for each protein in the B. anthracis proteome. All proteins with an average model score ≥1 are plotted against the average pAE score from the best model. Dashed blue lines show arbitrarily assigned significance thresholds (pAE ≤ 5, average models ≥3). Forty-four proteins met these criteria. 02960 (PdaN) is shown in cyan. Three other polysaccharide deacetylases are also indicated. (B) Sporulation efficiency of suppressors of ∆ipdA (top) identified through an enrichment screen (see Materials and methods). Validation of the hits from the screen using transposon-insertion mutants (bottom). (C) Schematics of the 02960 (pdaN) and 04184 (pdcF) genes with suppressor mutations highlighted. (D) AlphaFold3-predicted complex of IpdA (white) and the catalytic domain of PdaN (cyan). PdaN’s catalytic residues (magenta) and Zn²⁺ ligand (purple sphere) are shown. Residues in IpdA predicted to interact with PdaN are shown in dark blue. (E) AlphaFold3-predicted complex of PdaN (cyan) and PdcF (pink). The predicted lipidated N-terminus of PdcF is shown in dark blue. (F) AlphaFold3-predicted complex of IpdA (white), PdaN (cyan), and PdcF (pink). The pAE plots for Fig 5D–5F can be found in Figs L and M in S1 Text. (G) Representative fluorescence images of the indicated B. subtilis strains with cytoplasmic mCherry (red) 3 hour after addition of IPTG (500 µM, final) to induce expression of B. anthracis pdaN and pdcF. Short, chubby cells (yellow carets) that ultimately lyse (white carets) are highlighted. A complete time course can be found in Fig P in S1 Text. (H) Representative fluorescence images of the indicated B. anthracis strains at hour 4 of sporulation. Yellow carets highlight sporulating cells with septal bulges. pdaN::Tn modestly suppresses the morphological defects associated with ∆ipdA despite the 500-fold suppression of the sporulation defect. pdcF::Tn suppresses the septal bulging to a greater extent. (I) Sporulation efficiencies of the indicated strains. An IpdA variant with three amino acid substitutions in residues predicted to interact with PdaN fails to complement the ∆ipdA mutant. (J) Representative anti-His immunoblot of the indicated strains over a sporulation time course. Time in hours after the onset of sporulation is indicated above the blots. Black carets highlight the bands specific to IpdA-His. The lower band (highlighted with an asterisk) is lysozyme, used to generate the lysates. A Coomassie-stained gel was used to control for loading. (K) Representative fluorescence images of the indicated B. subtilis strains at hour 4 of sporulation. IPTG (500 µM, final) was added to the cultures at the onset of sporulation. Yellow carets highlight septal bulges. A complete time course can be found in Fig R in S1 Text. Scale bars indicate 5 µm. The data underlying panels A, B, and I can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003521.g005

Fig 5D shows the high-confidence AlphaFold3-predicted IpdA-PdaN complex. Importantly, residues in a short α-helix of IpdA are predicted to bind and block PdaN’s catalytic pocket. Similar complexes are predicted for the other three deacetylases (Fig L in S1 Text). PdcF (04184) is predicted to be a lipoprotein with a four-helix bundle. Strikingly, AlphaFold3 (AF3) predicts a high-confidence interaction between PdaN and PdcF (Fig 5E and Fig Ma and N in S1 Text), suggesting the two proteins function as a complex and may require each other in vivo for proper deacetylase activity. Importantly, AF3 predicts that IpdA can bind and inhibit PdaN in the context of the PdcF-PdaN complex (Fig 5F and Fig Mb in S1 Text). Transcriptional reporters in which the pdaN and pdcF promoters were fused to yfp indicate that both factors are expressed during vegetative growth (Fig O in S1 Text), consistent with the model that they act together to modify the envelope during growth. In support of this idea, expression of both genes, but not either one alone, in B. subtilis caused impaired growth (Fig Pa in S1 Text). Furthermore, fluorescence imaging revealed that co-expression of pdaN and pdcF specifically impaired cell elongation, resulting in short chubby cells (Fig 5G and Fig Pb in S1 Text). These morphological defects are strikingly similar to the phenotypes associated with B. subtilis cells depleted of the PG hydrolases CwlO and LytE required for cell wall elongation [62,63]. Collectively, these findings argue that PdaN and PdcF act together to deacetylate peptidoglycan, blocking cell wall cleavage.

Although the pdaN::Tn mutant significantly suppressed the sporulation defect of ∆ipdA (Fig 5B), it only modestly suppressed the engulfment defect (Fig 5H). By contrast, pdcF::Tn was more effective at suppressing the septal bulges in the ∆ipdA mutant. These data raise the possibility that additional deacetylases are inhibited by IpdA and these enzymes might impair engulfment in the ∆ipdA pdaN::Tn background (see Discussion). Similarly, our findings also suggest that PdcF is required for the activity or membrane localization of one or several additional deacetylases. In support of this latter model, vegetative B. anthracis cells lacking pdcF but not pdaN are more susceptible to lysozyme than WT (Fig Q in S1 Text). Since deacetylation provides high-level lysozyme resistance in B. anthracis [64], this result argues that PdcF functions as a co-factor for several deacetylases and in its absence the cell wall is more acetylated and susceptible to lysozyme.

To investigate whether IpdA inhibits PdaN as predicted by the structural prediction, we substituted alanines for the three residues (N73, Q76, and D80) predicted to bind and occlude PdaN’s catalytic pocket. Consistent with the AF3 prediction, this protein variant failed to rescue the ∆ipdA mutant (Fig 5I). Immunoblot analysis of IpdA-His during a sporulation time course revealed that the levels of the IpdA variant were reduced relative to the WT protein (Fig 5J). The reduced protein levels could be due to instability caused by the amino acid substitutions or could result from a failure to interact with PdaN, leading to release of the mutant protein into the medium. We tentatively conclude that IpdA binds and inhibits PdaN in a manner predicted by AlphaFold.

Altogether, our data are consistent with a model in which induction of IpdA at the onset of sporulation inhibits PdaN and possibly other vegetative deacetylases. The acetylated PG sugars ensure that the sporulation-specific cell wall hydrolases can efficiently cleave the septal PG and promote engulfment. To further investigate this model, we co-expressed PdaN and PdcF in B. subtilis cells at the onset of sporulation and visualized the engulfment process over time. As can be seen in Fig 5K and Fig R in S1 Text, engulfment was impaired and many cells had septal bulges similar to those observed in the B. anthracis ∆ipdA mutant. Importantly, the engulfment defects required expression of both proteins (Fig R in S1 Text). We observed similar septal bulges in sporulating B. subtilis cell when we induced the expression of B. subtilis pdaC at the onset of sporulation (Fig R in S1 Text). PdaC is a membrane anchored MurNAc deacetylase that is not normally produced during sporulation [25,65]. We conclude that one or both of the engulfment hydrolases in B. subtilis require acetylated peptidoglycan to degrade the septal cell wall. We infer that the B. anthracis enzymes require a similar substrate and that induction of IpdA at the onset of sporulation ensures the septal PG remains sufficiently acetylated (Fig 6) (See Discussion).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Schematic model for IpdA function during sporulation in B. anthracis.

(A) During vegetative growth, PdaN, 02021, and 05406 are proposed to deacetylate the N-linked acetyl groups from GlcNAc (G) and MurNAc (M) sugars of the peptidoglycan (PG). (B) At the onset of sporulation but prior to polar division, IpdA is induced and inhibits deacetylase activity, resulting in acetylated septal PG (left). In the ∆ipdA mutant, the septal PG is deacetylated (right). (C) The cell wall hydrolases SpoIID and SpoIIP can cleave acetylated PG and promote engulfment (left). SpoIID and SpoIIP are unable to efficiently cleave deacetylated PG in the ∆ipdA mutant, resulting in septal membrane bulges (right). For simplicity, the deacetylase co-factor PdcF is not included in the schematic. (D) Comparison of IseA homologs from B. subtilis, Ralstonia pickettii and B. anthracis. The region on each protein that interacts with its cognate enzyme active site is colored dark blue. Structures of IseA-LytE (PDB: 8I2E) and Tai3:Tae3 (PDB: 4HZ9) are shown on the right. Only the catalytic domains of LytE and Tae3 are depicted.

https://doi.org/10.1371/journal.pbio.3003521.g006

Bacterial strains and media

B. anthracis strains were cultured in brain-heart infusion (BHI) broth or agar. When required, BHI medium was supplemented with chloramphenicol (5 µg/mL), erythromycin (5 µg/mL), kanamycin (25 µg/mL), or spectinomycin (300 µg/mL). Sporulation of B. anthracis was induced at 37 °C by nutrient exhaustion in PA medium [85]. B. subtilis strains were grown in Luria–Bertani (LB) broth or agar. When required, LB medium was supplemented with chloramphenicol (5 µg/mL), erythromycin (1 µg/mL) plus lincomycin (25 µg/mL), or spectinomycin (100 µg/mL). Sporulation of B. subtilis was induced at 37 °C by nutrient exhaustion in supplemented Difco Sporulation medium (DSM) [86] or by resuspension according to the method of Sterlini–Mandelstam [87]. Sporulation efficiency was determined in 24–30-hour cultures as the total number of heat-resistant (B. anthracis: 65 °C for 30 min, or B. subtilis: 80 °C for 20 min) colony forming units (CFUs) compared to WT heat-resistant CFUs. The transposon-sequencing screen and the transposon-mutant collection were constructed in BaR1, the plasmid-free Sterne derivative of B. anthracis 9131 [88]. For sporulation efficiency assays, transposon mutants were transduced into BaR6 (B. anthracis Sterne NR-9401). B. anthracis 9131 (BaR1) spores are more adherent to glass and plastic tubes reducing the accuracy of CFU quantification. B. subtilis strains were derived from the prototrophic strain PY79 [89]. B. subtilis insertion-deletion mutants were from the Bacillus knock-out (BKE) collection [90] or were generated by direct transformation of plasmids into B. subtilis. All strains, plasmids, and oligonucleotide primers can be found in Tables A–C in S1 Text, and a description of plasmid constructions can be found in Supplemental Methods in S1 Text.

CP51-mediated transduction

A temperature-sensitive mutant of the generalized transducing phage CP51 was used to transduce antibiotic-marked alleles [85]. Briefly, the donor strain was grown in BHI supplemented with 0.5% glycerol, then mixed with phage CP51 and PA soft agar. The mixture was poured onto nutrient broth/yeast extract/glycerol (NBYG) agar plates and incubated overnight at 30 °C. The transducing lysate was prepared by collecting the soft agar and resuspending it in BHI supplemented with 0.5% glycerol, 10% dimethyl sulfoxide, and 20 mM MgSO₄, followed by filtration through a 0.45 μm filter. The recipient strain was grown in BHI supplemented with 0.5% glycerol, mixed with the transducing lysate, and incubated with constant agitation at 37 °C for 45 min. The mixture was then plated on BHI agar plates supplemented with the appropriate antibiotic and incubated overnight at 37 °C. Transductants were confirmed by PCR.

Electroporation

Transposon and allelic exchange plasmids, prepared from the Escherichia coli strain SC110 (dam−, dcm−), were transformed into B. anthracis by electroporation. Electrocompetent cells were prepared by growing B. anthracis in BHI with 0.1% glucose (BHIG) and 0.005 µg/mL tunicamycin. Low concentrations of tunicamycin increase transformation efficiency ~30-fold, likely due to the partial inhibition of TagO, the committing enzyme for the synthesis of secondary cell wall polysaccharides [91,92]. Cells were harvested at an OD₆₀₀ of 0.3, chilled on ice, and washed 4 times in cold electroporation buffer (0.5 M sucrose, 10% glycerol). 100 µL of cells (concentrated 1000-fold) were mixed with ~2 µg plasmid DNA and transferred to a cold 1 mm electroporation cuvette. The electroporator settings were 25 µF, 200 Ω, 1.85 kV. Immediately after electroporation, cells were resuspended in 1 mL recovery medium (BHIG supplemented with 20 mM MgCl₂ and 2 mM CaCl₂) and incubated with constant agitation for 2 hour at 30 °C. The culture was plated on BHI agar supplemented with the appropriate antibiotic and incubated overnight at 30 °C.

Allelic exchange

Deletions and integration of transgenes at neutral loci in the genome were generated by allelic exchange using pFR50, derived from pMAD [93]. This modified vector carries a thermosensitive origin of replication, an erythromycin resistance cassette for Bacilli, and mCherry fused to the B. subtilis veg promoter. 1 kb DNA fragments flanking the deletion or insertion site were subcloned into this vector with a second antibiotic resistance cassette between them. The plasmids were prepared in E. coli strain SC110 and electroporated into B. anthracis followed by selection on BHI agar supplemented with erythromycin. A streak-purified pink colony was then grown in 5 mL BHI-Erm at 30 °C. After 10 hour of growth, the culture was serially diluted and plated on BHI-Erm and incubated overnight at 42 °C to select for a single crossover integration (loop-in). A single colony was then grown without selection in 5 mL BHI at 30 °C for 10 hour. The culture was serially diluted and plated on BHI agar supplemented with the second antibiotic. Plates were incubated overnight at 37 °C. Colonies were screened for loop-out by the loss of pink color and sensitivity to erythromycin. Allelic exchange was confirmed by PCR.

Transposon library construction and Transposon-insertion sequencing

Tn-seq was performed as described previously [17,24,25] with modifications described below. The B. anthracis transposon library was generated using the transposon delivery plasmid pFR38. This E. coli-Bacillus shuttle vector contains a temperature-sensitive replicon for Bacilli, an erythromycin resistance cassette, the mariner-Himar1 transposase C9, and a spectinomycin resistance cassette flanked by inverted repeats recognized by the transposase. One of the two inverted repeats contains an MmeI endonuclease restriction site. There is enough readthrough of the spec gene to provide expression of downstream essential genes in operons when co-oriented. pFR38 was introduced into the B. anthracis WT strain BaR1 via electroporation. Transformants were selected at 30 °C on BHI agar supplemented with erythromycin. Ten test tubes with 4 mL of BHI supplemented with spectinomycin were each inoculated with 4 transformants and grown at 22 °C for 16 hour. The cultures were pooled and transposants were selected on BHI agar plates supplemented with spectinomycin at 42 °C. Approximately 1,000,000 transposant colonies were pooled, diluted to an OD₆₀₀ of 5, distributed into 500 µL aliquots, and frozen in 15% glycerol at −80 °C. One aliquot was used to prepare genomic DNA (gDNA) to determine the depth of the library. The gDNA was digested with MmeI, followed by barcoded adaptor ligation. The transposon-chromosome junctions were amplified in 18 PCR cycles using a universal primer that binds the adaptor and a primer that binds within the transposon. The PCR product was gel-purified and sequenced on the Illumina HiSeq platform using TruSeq reagents (Tufts University Core Facility [TUCF]). The reads were mapped to the B. anthracis BaR1 genome.

Gene essentiality

Essential genes were determined from the Tn-seq described above using EL-ARTIST [26]. 0.03 was used as the p-value cutoff for calling a region significantly underrepresented in reads and a sliding window size of 7 was used in the analysis.

Sporulation Tn-seq screen

A frozen aliquot (500 µL) of the Tn-library was thawed, washed twice in PA medium, and diluted into 50 mL PA medium to an OD₆₀₀ of 0.05. The culture was grown at 37 °C and growth was monitored by OD₆₀₀ over time. An aliquot was harvested at the onset of starvation (input) and the rest of the culture was sporulated for 7 hour. Spores from the BaR1 strain are less adherent after 7 hour of sporulation and the maximum number of heat-resistant CFUs can be recovered. The sporulated sample was incubated at 65 °C for 30 min to kill vegetative cells and sporulation-defective mutants and plated on BHI agar. Approximately 1,000,000 heat-resistant colonies were pooled (output). gDNA was isolated from the input and output samples and processed for transposon-sequencing as described above. Reads were mapped to the B. anthracis BaR1 genome, tallied at each TA site, and genes in which reads were statistically underrepresented were identified using the Mann–Whitney U test. Visual inspection of transposon-insertion profiles was performed with the Sanger Artemis Genome Browser and Annotation tool [94]. Some of the genes listed in S2–S6 Tables that were chosen for further characterization based on visual inspection of the insertion profiles despite having a P-value >0.05 and/or a fold-enrichment of less than 3.

Construction of the ordered transposon-mutant library by Knock-out Sudoku

The ordered library was constructed based on the protocols in Baym and colleagues 2016 [31]. Briefly, ~52,000 B. anthracis transposon-mutant colonies were picked using a RapidPick MP robot and transferred into 182 96-well plates containing 140 µL BHI-Spec per well. Construction of this progenitor collection was divided into three rounds with ~17,472 transposon mutants picked per round. The arrayed mutants were grown overnight at 30 °C with agitation at 600 rpm. The next morning, 4 × 20 µL of culture from each well was removed using a 96-channel Gilson PlateMaster P220 and deposited into four pools (rows, columns, plate rows, and plate columns) using 8-channel and 12-channel reservoir plates, and two single-well plates. In total, we generated 47 address pools: 8 row pools, 12 column pools, 13 plate rows, and 14 plate columns (using 13 × 14 plate grid). 50 µL of 20% glycerol was added to the remaining culture in each well of the progenitor collection and cryopreserved at −80 °C. The transposon-chromosome junctions of the transposons in each pool were determined by barcoded Tn-seq in a single NextSeq run.

To locate a Tn mutant in each gene within the progenitor collection, the sequencing dataset was parsed to identify the address pools in which each Tn mutant was found. A set of four pools containing a single Tn insertion dictates the coordinates of the well. In the cases where multiple distinct Tn insertion sites were identified for a given gene, we chose the Tn that had inserted co-directionally with the gene and had inserted closest to a position 1/5th of the gene length. When these criteria could not be satisfied, we used the next best Tn insertion (prioritizing co-directional insertion and then closeness to the 1/5th position). In total, we identified Tn mutants in 3,826 genes in singly occupied wells in the progenitor collection. We identified Tn mutants in 40 additional genes in doubly occupied wells. Tn mutants in 86 genes had insufficient address data to map. No mutants were identified for 1,504 genes in the B. anthracis genome. In many cases, this was due to gene essentiality and/or small size.

After transposon mapping, we used the colony picking robot to cherry-pick the Tn mutants from the progenitor collection and inoculate BHI supplemented with spectinomycin. These mutants were arrayed in 40 96-well plates. The 40 plate collection was pooled using the same strategy as the original progenitor collection with a 6 × 7 plate grid. The Tn-insertion sites in the 33 pools were determined using barcoded adaptors and Tn-sequencing. The insertions were mapped and their location in the condensed collection was validated. The nonredundant quality-controlled collection contains 3,826 Tn mutants that represent 70% of the 5,456 genes in the B. anthracis genome. We preserved two plates (D41 and D42) in which the 40 doubly occupied wells were struck for single colonies and single colonies inoculated into four wells in the final collection. Although these wells are not concretely mapped, they can be checked for the presence of either Tn mutant. The direction and position of the Tn insertions chosen for the final collection can be found in S8 Table. The ordered B. anthracis Tn-mutant collection has been submitted to NIAID for consideration to be included in the BEI Resources program.

Enrichment screen for ∆ipdA suppressors

Four colonies of ∆04176::kan (∆ipdA::kan) in the BaR6 background were separately grown in 3 mL PA-Kan for 30 hour at 37 °C. The sporulated cultures were heat-killed (65 °C for 30 min), and 30 µL of the heat-treated material was used to inoculate 3 mL of PA-Kan medium for another round of sporulation. Heat-resistant CFUs were assessed after each cycle. After the fifth enrichment cycle, individual heat-resistant colonies from each culture were re-tested and gDNA from a representative suppressor from each culture was prepared and subjected to whole–genome sequencing (SeqCenter).

Fluorescence and phase-contrast microscopy

Vegetative and sporulating cells and dormant spores were collected by centrifugation at 8,000 × g for 1 min and immobilized on 2% agarose pads. Phase-contrast and fluorescence microscopy were performed using an Olympus BX61 microscope equipped with an UplanF1 100× phase-contrast objective lens and a monochrome CoolSnapHQ digital camera (Photometrics). Membranes were stained with 1-(4-trimethylammoniumphenyl)-6-phenyl-1,3,5-hexatriene p-toluenesulfonate (TMA-DPH; Molecular Probes) at a final concentration of 50 μM. Exposure times were typically 100, 200, and 500 ms for YFP, CFP, and TMA-DPH, respectively. Image analysis and processing were performed using MetaMorph software (Molecular Devices; version 7.7).

Immunoblot analysis

B. anthracis sporulating cells were collected by centrifugation at 16,000 × g for 2 min. Whole cell extracts were prepared by resuspending cell pellets in 50 µL lysis buffer (20 mM Tris pH7.5, 1 mM EDTA) containing 10 mg/mL lysozyme, 10 mM MgCl_2, 25 units DNase I, 0.08 mg/mL RNase A, and 1 mM phenylmethylsulfonyl fluoride) and incubating at 37 °C for 15 min. The samples were then mixed with an equal volume of 2× SDS-PAGE sample buffer containing 10% BME and heated at 50 °C for 15 min prior loading. Equivalent loading was based on OD₆₀₀ at the time of harvest. Proteins were separated by SDS-PAGE on a 20% polyacrylamide gel. Proteins were transferred to a PVDF membrane and blocked for 1 hour in 1× PBS-0.5% Tween-20 containing 5% nonfat milk. The blocked membrane was washed 4 times for 5 min with PBS-0.05% Tween-20 (PBST) and probed overnight at 4 °C with mouse anti-His tag antibody (GenScript) at a 1:4,000 dilution in PBST containing 3% BSA. The next morning the membrane was washed 4 times for 5 min with PBST and then incubated for 1 hour at room temperature with horseradish peroxidase-conjugated goat anti-mouse immunoglobulin G (Bio-Rad) at 1:10,000 dilution in PBST containing 3% BSA. The membrane was washed 4 times for 5 min with PBST and then incubated with Western Lightning reagent (PerkinElmer) and detected using a ChemiDoc MP, Imagin System (BIO-RAD). To control for loading, the lysates were separately resolved by SDS-PAGE and stained with Coomassie Brilliant Blue. Two biological replicates were analyzed, and a representative immunoblot and a Coomassie-stained gel are shown.

AlphaFold2 Multimer screen

A one-by-all AF-M screen was performed using local Colabfold version 1.5.2 in January of 2025 [95,96]. IpdA (A0A6L7GZZ9) was folded pairwise against the proteome of B. anthracis (UP000000594). Each prediction was run with all five AF-M models, using three recycles, no dropout, no templates. These runs were performed using a local installation of Colabfold running on NVIDIA A6000 GPUs managed by the HMS O2 computing cluster. MSAs were generated using MMseqs2 API server [97]. All predictions used paired and unpaired MSAs, and structures were unrelaxed. To analyze the results of the pairwise screen, we utilized a published analysis pipeline [58]. Predictions were considered a hit when the average models score was ≥3 and the average pAE ≤ 5 Å across the contact region in the best model. The metrics for each prediction are listed in S7 Table.

Genomic neighborhood and domain fusion analysis

Gene neighborhood analysis was performed as previously described [98] using EFI-EST v2.0 from the Enzyme Function Initiative (available at: https://efi.igb.illinois.edu/efi-est/). Alignment score cutoff of 35% was used for the analysis. Gene neighborhood diagrams were generated to visualize the 10 nearest genes surrounding the pdcF (ykyA) gene. Domain fusions were identified for the YkyA protein family (Pfam ID: PF10368) on the InterPro website, available at: https://www.ebi.ac.uk/interpro/entry/pfam/PF10368/.

Spot-dilution assays

B. subtilis strains were grown at 37 °C in LB medium until mid-exponential phase. Cultures were normalized to OD₆₀₀ = 0.6 and 10-fold serial dilutions were generated. 5 μL of each dilution was spotted onto LB agar supplemented with or without indicated concentrations of IPTG. Plates were incubated at 37 °C overnight and photographed the next day.

Lysozyme sensitivity assays

B. anthracis and B. subtilis strains were grown at 37 °C in sporulation media until mid-exponential phase and then back-diluted 50-fold and distributed into 96-well microtiter plates containing lysozyme. The OD₆₀₀ was recorded every 3 min for 5 hour using an Infinite M Plex plate reader (Tecan). The 96-well plate was maintained at 37 °C with constant agitation between measurements. Growth kinetics were analyzed in technical triplicate, and the values were averaged. The results reported are representative of experiments that were performed on at least two biological replicates.