Y. pseudotuberculosis toggles pYV plasmid copy number in response to temperature

To determine the average PCN per chromosome equivalents at the bacterial population level, we used droplet digital (dd)PCR to measure the number of pYV molecules relative to chromosomes in wildtype Y. pseudotuberculosis IP2666pIB1 [22]. Cells were either grown at 26°C to represent growth in environmental conditions outside of the host or at 37°C in low calcium media to represent host cell contact during mammalian infection. The pYV-encoded T3SS is expressed at 37°C and becomes active for secretion following host cell membrane contact or calcium chelation [53–56]. Consistent with previous studies [21,22], we observed an increase in pYV PCN from an average of ~1.5 pYV copies per chromosome equivalents at 26°C to ~3 copies at 37°C/low calcium (Fig 1A). In addition, using a YPIII/pIBX strain of Y. pseudotuberculosis harboring a pYV-encoded luciferase reporter, we found that relative pYV-encoded luciferase activity normalized to cell density increases one hour after shifting the cultures to 37°C/low calcium conditions (Fig 1B). As luciferase gene dosage correlates with bioluminescence [21], these data support the conclusion that pYV PCN increases at 37˚C.

Download:

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

Fig 1. Y. pseudotuberculosis pYV plasmid copy number (PCN) dynamically changes in response to temperature.

(A) pYV PCN per chromosome was determined for wildtype Y. pseudotuberculosis IP2666pIB1 grown at 26°C or 37°C/low calcium using droplet digital PCR (dd)PCR, as the ratio between concentrations of pYV and chromosome DNA. (B) Relative pYV plasmid copy number of wildtype Y. pseudotuberculosis YPIII/pIBX was determined following growth at 26°C or 37°C/low calcium using a luciferase reporter stably incorporated in the pIBX plasmid, normalized to bacterial cell density (OD₆₀₀). ODU, optical density units. (A-B) Averages of three independent experiments are shown ± standard error of the mean (Student t-test, *** p < 0.001, * p < 0.05). (C) Representative phase contrast (left) and fluorescence (right) images of Y. pseudotuberculosis YPIII/pIBX cells where the pYV plasmid is visualized using a ParB-msfGFP fluorescent marker. Images are shown for cells grown at 26°C or under the 37°C/low calcium condition. (D) Histogram depicting the number of ParB-msfGFP-labeled puncta of pYV quantified from data illustrated in panel C. Three biological replicates per condition were imaged and analyzed; for each strain and condition, between 790–4428 cells were analyzed per replicate.

https://doi.org/10.1371/journal.ppat.1012655.g001

To examine pYV PCN at the single-cell level, we employed a microscopy-based approach. To visualize and quantify pYV puncta in living Y. pseudotuberculosis cells, we generated a YPIII/pIBX Y. pseudotuberculosis strain that harbors a parB-msfGFP translational fusion in the native parB pYV locus. This approach has been used to visualize plasmids in other bacterial species [57,58], as ParB binds to plasmid-encoded parS sequences but has never been used to study pYV in Y. pseudotuberculosis. At 26°C, cells were found to have an average of about four pYV puncta per cell (Fig 1C-D). This number increased to an average of about seven puncta per cell by 1.5 hours after the shift to 37°C (Fig 1C-D). The higher number of pYV per cell than per chromosome equivalents (Figs 1D vs. 1A) is consistent with the expectation that Yersinia has more than one chromosome per cell on average under nutrient-rich (fast) growth conditions, similar to E. coli [59,60]. Regardless of the normalization method (per cell or chromosome), Yersinia pYV PCN consistently increased by ~1.5-fold upon shifting bacterial cultures from 26°C to 37°C/low calcium conditions. The pYV plasmid was highly stable, with the vast majority of cells retaining pYV in Y. pseudotuberculosis (Fig 1D).

Disruption of the pcnB gene encoding the polyadenylase PAP I decreases pYV PCN and T3SS expression

In a screen for genes that regulate the Yersinia T3SS, we identified a novel factor that appeared to regulate pYV replication. Y. pseudotuberculosis with a functional T3SS, but lacking T3SS Yop effectors (∆yop6), induces activation of the transcription factor NFκB in mammalian cells [61]. The LPS metabolite ADP heptose is delivered to the host cell cytoplasm through the T3SS, stimulating Alpk1/TIFA-dependent NFκB activation [62], although YopJ inhibits NFκB signaling [63,64]. We used human embryonic fibroblast cells, HEK293T, expressing an NFκB luciferase reporter to identify a Y. pseudotuberculosis IP2666pIB1 ∆yop6 transposon (Tn) mutant that induced elevated bioluminescence compared to the parental strain, ∆yop6/pil::Tn (S1A Fig) [61]. This mutant was found to harbor a Tn insertion in a previously unannotated locus adjacent to the highly conserved pYV cop-rep locus. We will refer to this locus as the plasmid inhibitory locus (pil) (S1B Fig). Given the location of the pil locus relative to cop-rep, we hypothesized that this transposon insertion affects the pYV PCN. To test this, we introduced the pil::Tn insertion mutation into a clean wildtype Y. pseudotuberculosis IP2666pIB1 strain background with intact T3SS effector Yop-encoding genes (referred to as the pil::Tn mutant). Consistent with our hypothesis, the pil::Tn mutant displayed elevated secretion of T3SS cargo proteins compared to wildtype Yersinia (S1C Fig). In addition, we constructed a mutant in the YPIII/pIBX background with a plasmid insertion into the same locus in pil (pil::pNQ). As expected, the pil::Tn and pil::pNQ strains had elevated pYV PCN at 26°C and early after shifting to 37°C relative to the parental strain (S1D, E Fig), which is likely the underlying cause of the increased T3SS activity.

Insertion into the pil locus was associated with poor growth compared to the parental strain under T3SS-inducing conditions, and suppressor mutants could be readily isolated (S1F Fig). We hypothesized that constitutively elevated pYV PCN contributed to this growth defect, resulting in selective pressure favoring suppressor mutations that would alleviate high pYV PCN. Therefore, we used the pil::pNQ mutation as a tool to identify genes important for pYV PCN homeostasis. We carried out whole genome sequencing of a subset of pil::pNQ suppressor mutants and identified several categories of mutations affecting growth, pYV PCN, and the T3SS (S1 Table). Two clones harbored a single point mutation that was associated with reduced pYV PCN and T3SS activity relative to the pil::pNQ parental strain. This mutation was mapped to the pcnB gene encoding the polyadenylase PAP I, and was predicted to confer a leucine to arginine change at position 291 (L291R) in the PAP I neck domain linked to RNA binding (S2A Fig). The L291 residue is conserved in PAP I from E. coli, suggesting that the residue might be important for PAP I structure and/or function (S2A Fig). We will use the term pcnB when referring to the gene, and PAP I when referring to the pcnB-encoded protein.

To confirm that the pcnB gene encoding PAP I affects Y. pseudotuberculosis pYV PCN, we generated a full in-frame deletion of the pcnB gene (∆pcnB) in the wildtype and pil::Tn IP2666pIB1 strain backgrounds. We used ddPCR to measure relative pYV PCN and found that compared to wildtype, the ∆pcnB mutant had lower pYV PCN following growth at 26°C and 37°C/low calcium (Fig 2A). In keeping with our previous findings, the pil::Tn mutant displayed significantly higher pYV PCN at 26°C compared to wildtype. Furthermore, loss of PAP I in the pil::Tn/∆pcnB double mutant significantly reduced pYV PCN at 26°C compared to the pil::Tn parental strain. These results show that loss of pcnB leads to decreased pYV PCN. Despite this, loss of pcnB did not prevent Yersinia from increasing pYV PCN upon shift from 26°C to 37°C, although not to the same high PCN as in wildtype (Fig 2A). Consistent with this reduction in T3SS gene dosage, the ∆pcnB mutant displayed two-fold less expression and secretion of the T3SS effector protein YopE compared to wildtype Y. pseudotuberculosis (Fig 2B, C). Similarly, pcnB deletion reduced YopE expression and secretion in the pil::Tn strain background (Fig 2B, C). Taken together, these data suggest that pcnB contributes to maintaining pYV PCN, which in turn promotes robust T3SS expression under T3SS-inducing conditions.

Download:

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

Fig 2. Disruption of the pcnB gene encoding poly(A) polymerase PAP I decreases pYV PCN as well as T3SS expression and activity.

(A) pYV plasmid copy number per chromosome was determined for Y. pseudotuberculosis IP2666pIB1 strains grown at 26°C or 37°C/low calcium using ddPCR. Averages of three independent experiments are shown ± standard error of the mean. Statistical significance was performed using a one-way ANOVA with Tukey’s multiple comparisons test (**** p < 0.0001, *** p < 0.001, ** p < 0.01, * p < 0.05). (B) Y. pseudotuberculosis IP2666pIB1 strains were grown in low calcium media at 37°C before preparing samples for western blotting. Bacterial pellets (P) were collected and secreted proteins (Yops) were precipitated from the supernatant (S). Images shown represent a single blot that was independently probed for YopE and RpoA, which serves as a loading control. (C) YopE pellet protein levels were quantified by densitometry relative to RpoA. Averages ± standard error of the mean are shown for three independent experiments. Statistical analysis was performed using a one-way ANOVA with Tukey’s multiple comparisons test (** p < 0.01, ** p < 0.001).

https://doi.org/10.1371/journal.ppat.1012655.g002

To determine how the pcnB gene and the pil::Tn disruption broadly impact expression of pYV-encoded genes, we carried out RNA-seq analyses on wildtype, ∆pcnB, pil::Tn, and pil::Tn/∆pcnB Y. pseudotuberculosis IP2666pIB1 grown at 26°C and 37°C/low calcium (S1 Data, GEO series accession number GSE292107). Most pYV-encoded gene transcripts were present at higher steady-state level in the pil::Tn mutant compared to wildtype Y. pseudotuberculosis at both temperatures (S3 Fig), consistent with an increase in pYV gene dosage. However, T3SS genes are not translated at 26°C due to the action of several RNA thermometers in Yersinia [53], and the pil::Tn strain does not display T3SS activity at 26˚C despite elevated T3SS gene transcripts at this temperature (S4 Fig). Interestingly, the pil::Tn strain displayed an altered RNA-Seq read pattern mapping to the pil-copB-repA locus at both temperatures (S5 Fig). We observed no difference in pYV PCN or T3SS expression when we generated a full in-frame deletion of the predicted ORF found in the pil locus (S1C, S1E Fig). Therefore, we propose that Tn or pNQ insertion into pil leads to derepressed transcription of the cop-rep locus leading to elevated pYV replication through an as-yet unknown mechanism. In contrast, deletion of pcnB resulted in fewer transcripts of most pYV-encoded genes compared to wildtype Yersinia at both temperatures (S3 Fig), consistent with a decrease in pYV gene dosage. In addition, pcnB deletion in the pil::Tn background also led to a decrease in the steady-state levels of most T3SS genes compared to the pil::Tn parental strain (S3 Fig). We conclude that loss of pcnB leads to a decrease in pYV PCN in both the wildtype and pil::Tn backgrounds, leading to decreased transcription of pYV-encoded genes and mitigation of T3SS activity-induced growth arrest.

PAP I is required for pYV stability

When working with the ∆pcnB strain, we noted that cells lost the pYV plasmid at a higher rate than observed with wildtype Y. pseudotuberculosis (Table 1). To determine whether PAP I affects pYV stability in addition to pYV PCN, we generated a ∆pcnB strain in the ParB-msfGFP YPIII/pIBX Y. pseudotuberculosis background. We observed ∆pcnB cells with lower cellular GFP fluorescence and lacking clear fluorescent foci, consistent with plasmid loss (Fig 3A). The presence of these plasmid-devoid cells in the ∆pcnB strain resulted in a clear bimodal distribution of cellular fluorescence (Fig 3B), showing that ~34% of ∆pcnB cells displayed low cellular fluorescence compared to only ~3% for the parental strain population at the time of imaging at 26ºC (p < 10^-100, Kolmogorov-Smirnov (KS) test; n = 7,033 cells per strain). The ∆pcnB mutant also had a high percentage (24%) of cells with low cellular fluorescence under the 37°C/low calcium condition (Fig 3B). The majority of cells with low cellular fluorescence had no detectable pYV spots (S6 Fig). By analyzing only those cells that retained at least one copy of pYV, we found that the ∆pcnB mutant showed the expected increase in ParB-msfGFP puncta per cell from 26°C to 37°C/low calcium (Fig 3C). However, the ∆pcnB distributions were significantly shifted toward lower values relative to the parental strain (p < 10^-100; Fig 3D, E). While the parental strain had, on average, 4.5 and 7.9 pYV spots per cell at 26˚C and 37˚C/low calcium, respectively, the ∆pcnB strain had 3.9 and 6.3. Thus, the pcnB gene does not affect the ability of the cells to respond to the T3SS-inducing environmental shift, but its loss leads to lower average PCN as well as increased frequency of plasmidless Yersinia cells. It is likely that the T3SS expression defect seen in the ∆pcnB mutant (Fig 2B, C) is due to an overall decrease in pYV burden in the bacterial population.

Download:

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

Table 1. PAP I is required for pYV maintenance in Y. pseudotuberculosis.

https://doi.org/10.1371/journal.ppat.1012655.t001

Download:

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

Fig 3. The pcnB gene is required for maintenance of the pYV plasmid.

(A) Representative phase contrast (left) and fluorescence (right) images of Y. pseudotuberculosis YPIII/pIBX ∆pcnB cells where the pYV plasmid is visualized using a ParB-msfGFP fluorescent marker. Images are shown for cells grown at 26°C or under the 37°C/low calcium condition. White arrowheads indicate cells with low fluorescence signal and no puncta, suggesting that they lack plasmids. (B) Histograms comparing the fluorescence intensity of ParB-msfGFP in the parental strain and ∆pcnB cells under 26°C or 37°C/low calcium conditions. Mean fluorescence intensity values are plotted on the x-axis (log₁₀). Shaded error bars represent the mean ± standard error of the mean from bootstrapping the data 100 times. Both histograms were randomly sampled without replacement to match the lowest n value of either condition (7,033 and 10,003 for the 26ºC and 37ºC/low calcium conditions, respectively). Both distributions are significantly different (p < 10^-100) based on the Kolmogorov-Smirnov test. A.U. indicates arbitrary units. (C) Histograms depicting the number of ParB-msfGFP-labeled puncta of pYV quantified from data illustrated in panel A. Only cells with at least one detected pYV spot were considered. Three biological replicates were imaged and analyzed per condition. For each strain and condition, between 2063 to 14,479 cells were analyzed per replicate. 26˚C, pink; 37˚C/low calcium, blue. (D) Same as (C) except that the comparison is between the parent and ∆pcnB strains grown at 26°C. For each strain and condition, between 760 to 4179 cells were analyzed per replicate. (E) Same as (D) except that it is for the 37°C/low calcium condition. For each strain and condition, between 2046 to 14,479 cells were analyzed per replicate. Distributions in panels D and E are significantly different (p < 10^-80) by Kolmogorov-Smirnov tests.

https://doi.org/10.1371/journal.ppat.1012655.g003

To further examine the extent of the pYV plasmid stability defect in the ∆pcnB mutant, we passaged bacteria on solid media for several days at 26°C and then assessed pYV retention for each clone by patching onto low calcium media containing the dye Congo red, and incubating at 37˚C. Congo red only binds to bacterial colonies that are expressing a functional T3SS [65]. Wildtype cells retained the pYV plasmid at 100% while the ∆pcnB mutant cells only retained pYV in 82% of cells following incubation at 26˚C for 16 hours (Table 1). When cells were exposed to T3SS-inducing conditions at 37°C/low calcium and returned to 26°C to mimic the Y. pseudotuberculosis facultative lifecycle, we found that only 58% of ∆pcnB cells retained pYV compared to 99% for wildtype (Table 1). This observed increase in pYV loss likely stems from the growth arrest of pYV⁺ clones specifically under T3SS-inducing conditions at 37˚C [15,54,55], which provides a selective growth advantage for ∆pcnB clones that have lost pYV. Altogether, these data support a role for PAP I in regulation of pYV PCN and plasmid stability during the Y. pseudotuberculosis facultative lifecycle.

The PAP I^L291R mutation affects PAP I protein levels

The original pcnB suppressor mutation identified in our pil::Tn suppressor screen was a single point mutation that changed leucine in position 291 to an arginine (L291R) (S2 Fig). Although the corresponding L291 residue is conserved in E. coli, it has not yet been identified as a critical PAP I residue. Thus, as a first step in investigating the role of this residue, we assessed whether the L291R mutation affects PAP I expression. We introduced the L291R missense mutation into the native pcnB locus of both Y. pseudotuberculosis IP2666pIB1 (with or without a PAP I C-terminal FLAG tag) and YPIII/pIBX to generate PAP I^L291R strains. We found that the PAP I^L291R protein was poorly expressed, particularly at 37°C/low calcium (Figs 4A and S7). These data suggest that the PAP I^L291R allele may be unstable, leading to low PAP I expression at least at 37°C/low calcium. Interestingly, a strain engineered with a more conservative leucine to alanine mutation at position 291, PAP I^L291A, did not significantly affect PAP I expression levels (Figs 4A and S7). Consistent with these data, the PAP I^L291R mutant showed enhanced loss of pYV compared to the wildtype strain or the PAP I^L291A mutant, as indicated by the emergence of white, fast-growing colonies following growth under 37°C/low calcium conditions on Congo red agar (Fig 4B). Interestingly, the PAP I^L291R strain displayed a greater loss of pYV following long passage at 37°C/low calcium compared to when bacteria were passaged at 26°C (Table 2), reflective of the normal PAP I^L291R protein steady-state levels at 26˚C compared to 37˚C. Consistently, the PAP I^L291R strain had significantly lower average pYV PCN compared to the wildtype and PAP I^L291A strains at 37˚C/low calcium (Fig 4C), and this correlated with lower relative T3SS activity (Fig 4D). Structure modeling of Yersinia PAP I suggests that the L291 residue is not surface exposed (S2C Fig). Our data suggest that mutation of the L291 residue to an arginine, but not an alanine, disrupts PAP I protein levels, explaining why we identified the L291R mutation in our suppressor screen.

Download:

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

Table 2. Predicted PAP I catalytic residues are required for stability of the pYV plasmid.

https://doi.org/10.1371/journal.ppat.1012655.t002

Download:

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

Fig 4. PAP I point mutations lead to a decrease in pYV PCN, plasmid stability, and T3SS activity.

(A) Y. pseudotuberculosis IP2666pIB1 strains harboring various PAP I^FLAG alleles grown at 26°C or 37°C/low calcium were pelleted and prepared for western blot analyses. Images shown represent a single blot that was cut after transferring and independently probed for expression of PAP I^FLAG (top) and RpoA (loading control, bottom). Data shown is representative of three biological replicates, with the averaged data shown in S7 Fig. (B) Y. pseudotuberculosis IP2666pIB1 strains were grown at 26°C before diluting and spotting onto either LB or low calcium plates containing Congo red dye and incubated at 26°C or 37°C, respectively, prior to imaging. The black arrow points to an example white colony indicative of plasmid loss and concomitant faster growth. Black box shows an example lawn of slow-growing bacteria harboring pYV and so therefore taking up the Congo red dye. (C) Relative pYV plasmid copy number was estimated for strains in the Y. pseudotuberculosis pIBX/YPIII background. For each timepoint, luminescence was measured and normalized to cell density (OD₆₀₀). Averages of three independent experiments are shown ± standard error of the mean. Each strain per condition was compared to the wildtype strain in a one-way ANOVA with Dunnett’s multiple comparisons test (*** p < 0.005, ** p < 0.01, * p < 0.05). (D) Y. pseudotuberculosis IP2666pIB1 strains were grown at 37°C/low calcium and secreted proteins in the supernatant fractions were precipitated and visualized after separation on as SDS-PAGE gel by Coomassie blue staining. Bovine serum albumin (BSA) was used as a protein precipitation control. The YopE T3SS effector protein band from four independent experiments was visualized, quantified, and normalized to BSA from the same samples. Averages ± standard error of the mean are shown. Statistical significance was determined using an unpaired t test (** p < 0.005, * p < 0.05).

https://doi.org/10.1371/journal.ppat.1012655.g004

Mutation of residues predicted to affect polyadenylase activity decreases pYV PCN, plasmid stability, and T3SS activity

PAP I D80 and D82 have been shown to be essential for PAP I polyadenylase catalytic activity in E. coli [66,67]. These residues are also present in Y. pseudotuberculosis PAP I (S2 Fig). To determine whether these catalytic residues are important for normal pYV PCN and stability, we mutated the two corresponding aspartic acid residues (D114 and D116) to alanine to produce a PAP I^D2A catalytic mutant. As expected, the D2A mutation did not affect PAP I steady state protein levels (Figs 4A, and S7) yet displayed the same level of reduced pYV PCN as the ∆pcnB strain that lacks PAP I completely (Fig 4C). Spot test serial dilution analyses found that PAP I^D2A grows similarly to wildtype and ∆pcnB at 26°C (Fig 4B). Furthermore, at 37°C/low calcium, PAP I^D2A and ∆pcnB exhibited about 10-fold more growth compared to wildtype (Fig 4B), consistent with a reduction in growth arrest due to decreased T3SS activity. Similarly, the strain expressing PAP I^D2A, like the ∆pcnB mutant, displayed an increased number of plasmidless, rapidly-growing colonies (Fig 4B) and reduced pYV stability (Table 2). Lastly, the PAP I^D2A allele did not support normal T3SS activity (Fig 4D). Taken together, these data suggest that Yersinia PAP I residues predicted to be required for polyadenylase activity are necessary to maintain normal pYV PCN, plasmid stability, and T3SS gene expression and activity.

Yersinia PAP I is required for plasmid stability and plasmid-mediated antibiotic resistance from a broad range of plasmid families

To determine if Yersinia PAP I contributes to regulation of PCN and retention of other plasmid types, we introduced a ColE1-like pGFP-uv plasmid into wildtype and ∆pcnB Y. pseudotuberculosis. This plasmid lacks active partitioning systems such as ParABS, but can be maintained within the population by antibiotic resistance selection [68]. We observed a decrease in fluorescence intensity of ∆pcnB colonies, as well as ~10-fold less plating efficiency of the ∆pcnB strain, when grown on carbenicillin-containing media compared to wildtype (Fig 5A). In addition, we found that over 70% of ∆pcnB cells lack the pGFP-uv plasmid within 10 hours of removing carbenicillin and the plasmid is completely lost from the population by ~36 hours. In contrast, 100% of wildtype cells retained pGFP-uv over the same 36 hour time course (Fig 5B). These data suggest that loss of PAP I activity attenuates GFP expression, plasmid stability, and plasmid-encoded β-lactam resistance in Yersinia harboring the ColE-like pGFPuv plasmid.

Download:

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

Fig 5. The pcnB gene is required for plasmid-mediated antibiotic resistance in Y. pseudotuberculosis.

(A) Wildtype and ∆pcnB Y. pseudotuberculosis IP2666pIB1 harboring pGFP-uv were spotted simultaneously onto either plain LB plates or LB plates containing carbenicillin to assess retention of pGFP-uv. Plates were incubated at 26°C for ~16 hrs before imaging spots under a GFP filter. (B) Retention of the pGFP-uv plasmid by wildtype and ∆pcnB Y. pseudotuberculosis IP2666pIB1 was assessed at 26°C over 40 hrs following the removal of carbenicillin. (C) Wildtype and ∆pcnB Y. pseudotuberculosis IP2666pIB1 harboring either no vector, a pTrc99::empty vector control, or a pTrc99::PAP I recombinant plasmid were spotted onto either plain LB plates or LB plates containing carbenicillin to select for pTrc99 and 0.2 mM IPTG to induce PAP I expression. Plates were incubated at 26°C for ~16 hrs prior to imaging. (D) Wildtype and ∆pcnB Y. pseudotuberculosis IP2666pIB1 harboring either no vector or a pNF06 mini-F plasmid with a ccdAB TA system (pNF06::ccdAB) were spotted onto either plain LB plates or LB plates containing 25 µg/mL kanamycin to select for pNF06. Plates were incubated at 26°C for ~16 hrs prior to imaging.

https://doi.org/10.1371/journal.ppat.1012655.g005

To ensure that the reduced plating efficiency on antibiotics seen for ∆pcnB harboring pGFP-uv was due to loss of PAP I specifically, we next constructed a PAP I overexpression plasmid in the pTrc99 vector backbone. Importantly, pTrc99 is also a ColE1-like plasmid that lacks partitioning systems [69]. We found that ∆pcnB harboring pTrc99::empty displays 10-fold less plating efficiency when spotted onto LB-carbenicillin but not LB-only plates (Fig 5C), consistent with what we previously observed with the pGFP-uv plasmid. Importantly, this decreased plating efficiency was nearly completely abolished by expression of PAP I from the pTrc99 plasmid (Fig 5C). In addition to ColE1 plasmids, we found that PAP I is also required for robust antibiotic resistance conferred by plasmids with p15A-like origins of replication (pACYC184, pBAD33) (S8 Fig). Interestingly, replication of plasmids containing ColE-like and p15A-like origins of replication, as well as IncFII plasmids like pYV, is regulated by plasmid-specific sRNA [70–72], suggesting a general role for PAP I in regulating plasmids with replicons regulated by a sRNA. To test if loss of pcnB would affect antibiotic resistance conferred by a plasmid whose replication is not known to be regulated by a sRNA, we introduced the mini-F plasmid pNF06::ccdAB into wildtype and ∆pcnB Y. pseudotuberculosis [30,48,73,74]. The ∆pcnB strain displayed equal plating efficiency to the wildtype strain in the presence of antibiotic selection, suggesting that pNF06::ccdAB maintenance is likely not affected by PAP I (Fig 5D). These data suggest that PAP I is important in Yersinia for normal maintenance of plasmids whose replication is regulated by a sRNA.

PAP I is required for Y. pseudotuberculosis virulence in mice

It is well established that the pYV plasmid and T3SS are critical for Y. pseudotuberculosis virulence in mouse infection models [6,75]. To determine if loss of PAP I contributes to Yersinia virulence, we infected mice via the intraperitoneal (IP) route and monitored colonization of the spleen and liver. Four days post-inoculation, significantly fewer ∆pcnB-infected mice exhibited robust colonization of the spleen and liver compared to wildtype-infected mice (Fig 6A and 6B). Interestingly, ∆pcnB-infected mice formed two distinct populations: one population where the spleen and liver were colonized to the same levels as wildtype-infected mice, and a second population that had completely cleared the infection (Fig 6A and 6B). We isolated several ∆pcnB clones from the livers of colonized mice and found that their pYV stability and T3SS activity phenotypes were similar to the inoculum strain (S9 Fig). In addition, we sequenced the genomes of these isolates and could not identify any significant mutations, suggesting that host tissue colonization by the ∆pcnB mutant in a subset of mice does not stem from suppressor mutations that stabilize pYV. Taken together, these data suggest that PAP I increases the likelihood of Yersinia colonization of host tissues.

Download:

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

Fig 6. The pcnB gene is required for normal Y. pseudotuberculosis virulence in a mouse IP infection model.

Six- to eight-week-old C57Bl/6 mice were infected with 1-3x10³ Y. pseudotuberculosis IP2666pIB1 via intraperitoneal (IP) injection. Tissues were collected, homogenized, and plated to determine CFU per gram (CFU/g) tissue 4 days post-inoculation. The log₁₀ CFU/g tissue from the spleens (A) and livers (B) of infected mice are shown. Each circle represents one mouse organ; lines represent geometric means.

https://doi.org/10.1371/journal.ppat.1012655.g006

PAP I is required for maintenance of plasmid-encoded T3SS expression and antibiotic resistance in Shigella flexneri

The human pathogen Shigella flexneri harbors a virulence plasmid, pINV, with a RepFIIA origin similar to the one found on Yersinia pYV [1]. To test whether PAP I also regulates Shigella pINV, we deleted the pcnB gene in the S. flexneri M90T ∆ipaH2.5 (∆ipaH2.5::tet^R) strain. We chose the ipaH2.5 deletion strain to introduce an antibiotic marker on the pINV virulence plasmid to force pINV retention if necessary. IpaH2.5 was the optimal target locus because IpaH2.5 and IpaH1.4 share 98% identity and are functionally redundant. Indeed IpaH2.5 is a pseudo gene in 2457T and is poorly expressed in M90T [76]. The resulting ∆ipaH2.5∆pcnB mutant was defective in Congo red binding compared to its parental strain (Fig 7A), suggesting a T3SS defect. Indeed, the ∆ipaH2.5∆pcnB strain exhibited a stark defect in expression and secretion of the T3SS translocon proteins IpaB, IpaC, and IpaD compared to the parental strain (Fig 7B and Fig 7C). Lastly, we introduced pTRC99 that confers IPTG-inducible expression of Yersinia PAP I into the ∆ipaH2.5 and ∆ipaH2.5∆pcnB strains. The Yersinia PAP I allele was able to partially complement the Shigella ∆pcnB defect in T3SS expression (Fig 7B) and secretion (Fig 7C). Consistent with our previous data using Y. pseudotuberculosis strains (Fig 5C), deletion of pcnB led to a > 10-fold decrease in plating efficiency in the presence of antibiotic that could be complemented by PAP I expression (Fig 7D). Taken together, these data suggest that PAP I is essential for maintaining robust plasmid replication, resulting in increased expression of plasmid-encoded virulence and antibiotic resistance genes in pathogenic bacteria.

Download:

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

Fig 7. The pcnB gene is required for plasmid-encoded expression of T3SS and antibiotic resistance genes in Shigella flexneri.

(A) S. flexneri M90T ∆ipaH2.5 and ∆ipaH2.5∆pcnB, as well as a virulence plasmid cured (VP^-) BS103 strain, were incubated on Congo red media at 37˚C. (B) S. flexneri M90T ∆ipaH2.5 and ∆ipaH2.5∆pcnB carrying pTRC99 with and without IPTG-inducible Yersinia PAP I were incubated at 37˚C, and secreted T3SS effector proteins were visualized after separation on as SDS-PAGE gel using anti-IpaB, anti-IpaC, and anti-IpaD antibodies. (C) S. flexneri M90T ∆ipaH2.5 and ∆ipaH2.5∆pcnB carrying pTRC99 with and without IPTG-inducible Yersinia PAP I were grown on either plain TCS media or TCS containing ampicillin and IPTG. Shown is one of two independent experiments.

https://doi.org/10.1371/journal.ppat.1012655.g007

Ethics statement

All experiments were approved by the University of California Santa Cruz Institutional Animal Care and Use Committee and were performed in accordance with National Institutes of Health guidelines.

Bacterial strains, plasmids, and growth conditions

All strains used in this study are listed in S2 Table. Y. pseudotuberculosis strains were grown in LB or 2xYT media (pH = 7) at 26°C or 37°C shaking at 250 rpm on a New Brunswick Scientific Innova 2300 platform shaker, unless otherwise indicated. S. flexneri were grown in tryptic soy broth media at 30°C or 37°C on a roller.

Construction of Y. pseudotuberculosis and S. flexneri mutant strains

All primers and plasmids used to generate Y. pseudotuberculosis strains for this study are listed in S3 and S4 Tables, respectively. Inserts were amplified by PCR using primers designed for Gibson Assembly and final recombinant plasmids were all made using the pCVD442 suicide vector backbone (λpir-dependent replicon, carbenicillin resistant [Carb^R], sacB gene conferring sucrose sensitivity) (Addgene) following digestion with NdeI and BbsI. All ligations were performed using NEBuilder HiFi DNA Assembly kit (New England Biolabs). Recombinant pCVD442 plasmids were transformed into E. coli S17-1 λpir competent cells and introduced into Y. pseudotuberculosis IP2666pIB1 or YPIII/pIBX via conjugation. Carb^R and irgansan-resistant (Yersinia selective antibiotic) clones were grown in the absence of antibiotics and plated onto media containing 10% sucrose to induce SacB-mediate toxicity for pCVD442 loss. Carb sensitive, sucrose-resistant, and Congo red-positive (pYV+) clones were screened by PCR and sequence verified by Sangar sequencing. pil insertion strains: A pil::Tn Y. pseudotuberculosis IP2666pIB1 ∆yop6 strain was originally isolated from a TnHimar1 library screened for altered NFκB induction compared to the parental strain [61]. The Tn insertion was then transferred to the wildtype IP2666pIB1 strain using the pSR47s suicide plasmid, as described above. The pil::pNQ Y. pseudotuberculosis YPIII/pIBX mutant was generated using the synthesized pNQ705 suicide plasmid carrying a truncated fragment of pil (Genescript) [89]. ∆pcnB: ~ 500 bp upstream of pcnB and ~500 bp downstream of pcnB was PCR amplified using genomic DNA (gDNA) from Y. pseudotuberculosis IP2666pIB1. Amplified fragments were cloned into pCVD442 to produce pCVD442::∆pcnB. ParB-GFP: To make ParB-GFP YPIII/pIBX strains, three fragments were amplified for insertion into pCVD442: (1) the last ~500 bp of the parB gene excluding the stop codon was amplified from Y. pseudotuberculosis gDNA, (2) msfGFP was amplified from pDK13 and (3) 500 bp downstream of the parB gene was amplified from Y. pseudotuberculosis gDNA. The resulting fragments were cloned into pCVD442 to generate a parB-msfGFP construct that includes 500 bp of downstream homology after msfGFP. PAPI-C-6xHis and PAPI-C-3xFLAG: To make PAPI-6xHis strains, pcnB was amplified from Y. pseudotuberculosis gDNA and cloned into pET28 that was previously digested with Xhol and NcoI to produce PAPI-C-6xHis. PAPI-C-6xHis was then amplified from pET28 and 500 bp downstream of pcnB was amplified from gDNA for insertion into pCVD442 to make pCVD442::PAPI-c-6xHis containing 500 bp of homology flanking the 6xHis tag. We subsequently found that the 6xHis tag was not sufficient to visualize the PAPI protein, and so alternative PAPI-C-3xFLAG strains were made in order to carry out western blot analyses. To make PAPI-3xFLAG, three fragments were amplified for insertion into pCVD442: (1) the pcnB gene was amplified from Y. pseudotuberculosis gDNA (2) 3xFLAG was amplified from pET28 and (3) 300 bp downstream of pcnB was amplified from Y. pseudotuberculosis gDNA. The fragments were then cloned into pCVD442 to make pCVD442::PAPI-c-3xFLAG containing 300 bp of downstream homology following the 3xFLAG tag. PAP 1 point mutations: To make PAP I^D2A, PAP I^L291R, and PAP I^L291A mutants, we used the Q5 site-directed mutagenesis kit (New England Biolabs) to introduce base pair changes at the corresponding residues using the pCVD442::PAPI-c-6xHis or pCVD442::PAPI-c-3xFLAG recombinant plasmids as templates. ∆ipaH2.5∆pcnB Shigella: A fragment containing 200 bp of homology upstream and 240 bp of downstream pcnB flanking sequence was amplified from ECK0142 (∆pcnB E. coli K-12 BW25113) using pcnb_F and pcnb_R [90]. This fragment was introduced into the chromosome of ∆ipaH2.5::tetR Shigella using lambda red recombination [91,92]. The insertion was confirmed using pcnB_R_op and k2_wanner.

In vitro suppressor screen

LB cultures of the pil::pNQ YPIII/pIBX mutant strain were grown at 26ºC shaking overnight before diluting in fresh media and spotting for single colonies onto low calcium plates (LB supplemented with 20 mM sodium oxalate and 20 mM MgCl₂). Plates were incubated at 37ºC for ~16 hrs to induce T3SS expression and selective pressure for growth of suppressor mutants. Clones that broke growth arrest and thus formed large colonies were selected and screened for changes in pYV PCN using the luciferase assay (see below). Suppressor mutants were also screened for changes in T3SS activity using an in vitro secretion assay (see below).

Identification of mutations arising from pil::pNQ suppressor screen

A subset of pil::pNQ YPIII/pIBX suppressor mutants were selected for whole genome sequencing. In brief, cells were grown shaking at in 2 mL LB at 26°C overnight and total genomic DNA was extracted using the Blood & Tissue DNeasy kit (Qiagen) according to manufacturer recommendations. Qubit quantification, library preparation, and Illumina paired-end sequencing with 150 bp read length was carried out by the UC Davis DNA Technologies Core. Adapters and low-quality reads were trimmed by Trimmomatic (Version 0.39) [93]. The trimmed sequences from each sample were aligned to the reference genome using BWA [94] The reference sequence was composed of the YPIII chromosome (CP009792.1) [95] and its plasmid (CP032567.1) [96]. Samtools [97] was used to sort and index the aligned reads. Duplicate reads originated from a single DNA fragment during sample preparation were located and tagged by MarkDuplicates in Picard (http://broadinstitute.github.io/picard/). Single nucleotide polymorphisms were detected and joined by GATK (Version 4.1.6) [98].

Droplet digital PCR (ddPCR)

ddPCR to determine pYV plasmid copy number was carried out as previously described [99]. Total DNA was extracted from bacterial cells using GeneJet Genomic DNA Purification kit (Thermo Scientific) according to the manufacturer’s recommendations for Gram-negative bacteria. Each genomic DNA sample was split into two reactions and the primer pair F_chrom_ddPCR/R_chrom_ddPCR was used to amplify chromosomal DNA and F_pYV_ddPCR/R_pYV_ddPCR was used to amplify pYV DNA. Individual ddPCR samples were prepared as follows: 10.5 µL of total genomic DNA was combined with 12.5 µL of 2xEVA-Green master mix for unlabeled probes (Bio-Rad), 1 µL of HindIII-HF (NEB), and 1 µL of respective primer pairs (200 nM final concentration) in 8-strip PCR tubes. PCR tubes were incubated at room temperature for 10 mins to allow HindIII to digest genomic DNA. Droplet generation was carried out using an Automated Droplet Generator (Bio-Rad) and droplets were subsequently transferred to a deep-dish qPCR plate (Bio-Rad) for PCR amplification. PCR was carried out with hot-start activation at 95°C for 5 mins following by denaturation at 94°C for 30s, amplification at 58°C for 1 min over 40 cycles, and signal stabilization at 4°C for 10 mins and 90°C for 5 mins. A ramp rate of 2C/s was used for all PCR steps. Droplets were analyzed using a QX200 droplet reader (Bio-Rad) and data was analyzed using QuantaSoft Analysis Standard Edition (version 1.2) (Bio-Rad). Statistical significance was calculated using a one-way ANOVA with Tukey’s multiple-comparisons test (**** P < 0.0001, *** P < 0.001, ** P < 0.01, * P < 0.05). All graphs and statistical analyses were generated using GraphPad Prism (10.2.0).

Luciferase pYV plasmid copy number assay

Overnight cultures for strains in the YPIII/pIBX background were grown shaking at 26°C before subculturing to an OD₆₀₀ of 0.2 in 2xYT low calcium media (containing 20mM sodium oxalate and 20mM MgCl₂). Cultures were then grown shaking for 1.5 hrs at 26°C before shifting to 37°C. Luminescence and OD₆₀₀ were measured after 1.5 hrs at 26°C just before shifting to 37°C and every hour after shifting to 37°C. Samples were collected at each timepoint in white walled, clear bottom 96 well plates and luminescence (using lum aperture for 96 well plate with measurement time = 0.1s) and OD₆₀₀ (using Photometric 600 excitation filter) was measured using a PerkinElmer Multimode Plate Reader EnVision system. Relative plasmid copy number was determined by dividing luminescence by OD₆₀₀ for the same sample. Statistical analysis was determined using a one-way ANOVA with Dunnett’s multiple comparisons test (**** P < 0.0001, *** P < 0.001, ** P < 0.01, * P < 0.05) where applicable. All graphs were made using GraphPad Prism (10.2.0).

ParB-msfGFP microscopy and image analysis

Wildtype and ∆pcnB YPIII/pIBX cells expressing ParB-msfGFP from LB plates supplemented with kanamycin were inoculated into EZ Rich Defined Medium (EZRDM) (Teknova) without antibiotic supplementation. EZRDM medium was chosen to reduce medium autofluorescence for downstream imaging purposes. Cultures were grown shaking at 200 rpm overnight at 26°C in a New Brunswick Scientific Excella E24 shaker to a final OD₆₀₀ of ~2–4. A secretion analysis was performed to confirm that Y. pseudotuberculosis displays normal T3SS activity when grown in EZRDM under T3SS-inducing conditions (37°C, supplemented with 20 mM sodium oxalate and 20 mM MgCl₂), referred to as low calcium EZRDM (S12 Fig). In addition, following incubation at 26˚C or under 37°C/low calcium conditions as in Table 1, 99% and 97% of ParB-msfGFP clones retained pYV, respectively, indicating that the ParB-msfGFP construct does not disturb plasmid partitioning. For non-T3SS inducing conditions, overnight cultures were diluted to an OD₆₀₀ of ~0.2 in fresh EZRDM and grown shaking in a New Brunswick Innova 44 incubator shaker for 3hrs at 26°C. For T3SS-inducing conditions, overnight cultures were diluted to an OD₆₀₀ of ~0.2 in low calcium EZRDM and grown shaking in a New Brunswick Innova 44 incubator shaker for 1.5 hrs at 26°C before shifting to 37°C for an additional 1.5 hrs. Cells were prepared for imaging by spotting onto a 1% agarose-low calcium EZRDM pad and covering with a no. 1.5 coverslip. Imaging was carried out using a Nikon Eclipse Ti microscope equipped with a 100X Plan Apo 1.45 NA phase-contrast oil objective, a Hamamatsu Orca-Flash4.0 V2 CMOS camera, and CoolLED pE-400 (Lumencor). The microscopes were controlled by the Nikon Elements software. The following Chroma filter cubes were used to acquire the fluorescence images: GFP: excitation ET470/40x, dichroic T495lpxr, emission ET525/50 m. Images were acquired with 50 ms of exposure (50% power) for GFP and 100 ms for phase contrast. For each strain and condition, three independent biological replicates (inoculated with different colonies) were grown and imaged.

For image analysis, cells were segmented and cell outlines were generated using the open-source Oufti software [100] using the following parameters: Edgemode, 1; Dilate, 1; openNum, 0; InvertImage, 0; ThreshFactorM, 0.86254; ThreshMinLevel, 0.3; EdgeSigmaL, 0.5; LogThresh, 0. Cell meshes were further curated using the SVMCuratedCellDetection.m script to identify and remove misdetected cells [101]. To quantify the number of pYV copies per cell, detection of pYV spots was performed using the ImageJ plug-in package ThunderSTORM [102,103]. For cells grown at 26°C, image filtering used a wavelet filter (B-spline) with a scale of 2.0 and order of 3.0. Under non-T3SS-inducing conditions, a local maximum using 5.0*std(Wave.F1) and a connectivity of 8-neighborhood was used to localize molecules based on peak intensity thresholds, and sub-pixel localization used a Gaussian point spread function (PSF) with a fitting radius of four pixels and an initial sigma of one pixel. For the 37°C/low calcium condition, the parameters were adjusted where the local maximum instead used 2.0*std (Wave.F1) and a connectivity of 8-neighborhood, and sub-pixel localization used a Gaussian PSF with a fitting radius of three pixels and an initial sigma of one pixel. Once parameter optimization was completed, the custom-made ImageJ macro BatchProcess_ThunderSTORM,ijm was used to iterate ThunderSTORM spot detection over all image files, applying the same parameters for a given biological replicate. Detected spots were linked with cell meshes using the custom MATLAB code analyze_ensemble_spot_intensity.m, as detailed further below. Code developed as part of this study has been deposited in the publicly accessible Github code repository (https://github.com/JacobsWagnerLab/published/tree/master/Schubert_et_al_2025). Three independent experiments were performed and analyzed per strain and condition.

Elaborating on the MATLAB script analyze_ensemble_spot_intensity.m, each cell was quantified by averaging whole cell intensity, recording the number of ThunderSTORM-detected spots, and noting the cell length. Only cells with length of at least 1 µm were kept to rule out incorrect cell segmentations. Each cell was identified and measured from its original Oufti mesh. There, the fluorescent GFP images were background-subtracted, then cell contours were used to obtain the boundaries of the cell area. Cellular fluorescence was determined by Oufti [100]. To link ThunderSTORM spots, (x,y) localizations, to Oufti meshes, the cell meshes were first converted to MATLAB polygons using the function polyshape. The function inpolygon was then used to feed the (x,y) localizations and polygon shape to find if any spots occurred inside the cell. For Fig 3 C-E, cells with zero spot were excluded from the analysis.

To assess the fraction of cells devoid of GFP signal, strains were grouped by condition (inducing vs. non-inducing as 37°C/low calcium vs. 26°C, respectively). Each strain in the condition was subsetted by random sampling without replacement to match the lowest N value (7,033 cells in 26°C and 10,003 cells in 37°C/low calcium) using the function datasample. These data were then bootstrapped 100 times to recalculate histograms and determine the error for each bin (using the custom function Hist_Errors). Since the ∆pcnB conditions had clear bimodal distributions for low/high fluorescent states, an Otsu threshold was calculated using the function otsuthresh to find the split between them [104]. Cell with intensity values below this threshold were deemed cells without plasmid content. The background spot detection of cells below the Otsu threshold were plotted in S5 Fig. To compare these two distribution shapes statistically, a Kolmogorov-Smirnov (KS) test was used via the function kstest2.

Images were visualized using NIS Elements and Fiji. Image pixel intensity values were autoscaled for display using Fiji, and all images were converted to RGB format for display. Data were plotted using GraphPad Prism 10.2.0, and Adobe Illustrator was used to generate figures involving images.

In vitro pYV plasmid stability assay

Cells were struck for single colonies onto LB plates and grown at 26°C for 1.5-2 days. To test plasmid stability after 26°C-only passage, single colonies were then patched onto LB plates for ~16 hrs at 26°C before heavily streaking for single colonies onto fresh LB plates at 26°C for 1.5-2 days. To test plasmid stability after exposure to T3SS-inducing conditions (37°C), the same colonies were also patched onto LB plates supplemented with 20 mM sodium oxalate to induce T3SS activity for ~16 hrs at 37°C before streaking for single colonies onto LB plates and returning to 26°C to grow for 1.5-2 days. Final plasmid stability was determined for both conditions by patching single colonies from the final 26°C streak on LB plates supplemented with 20 mM sodium oxalate and containing 1% Congo red. The percentage of cells that retained the pYV plasmid was calculated by dividing the number of red (pYV+) CFU by the total number of CFU. Single colony PCR was used on a subset of patches to validate that red CFU retain pYV and while white CFU do not.

In vitro T3SS activity assay

Quantification of secreted Yersinia T3SS cargo was carried out as previously described [105]. Y. pseudotuberculosis cultures were grown shaking overnight at 26°C before subculturing in LB low calcium media (LB containing 20 mM sodium oxalate and 20 mM MgCl₂). Cultures were grown shaking at 26°C for 1.5 hrs before shifting to 37°C/shaking for an additional 1.5 hrs. Cultures were normalized to the same number of cells by normalizing OD₆₀₀ values and pelleted at 13,200 rpm (~19,000 xg) for 15 mins at room temperature. Supernatants were transferred to new tubes and proteins precipitated with 10% trichloroacetic acid (TCA). A bovine serum albumin (BSA) protein control was spiked into each sample immediately before the addition of TCA. Samples were incubated in TCA on ice for at least 1 hr before pelleting at 13,200 rpm (19,000 xg) for 15 mins at 4°C. Pellets were then washed twice with ice-cold 100% acetone and resuspended in final sample buffer (FSB) containing 0.2M dithiothreitol (DTT). Samples were boiled for 10 mins at 95°C before loading and separating on a 12.5% SDS-PAGE gel. Gels were stained with Coomassie blue to visualize proteins and imaged using Bio-Rad Image Lab Software Quantity and Analysis tools.

Shigella T3S secretion assay

S. flexneri cultures were grown with shaking overnight in tryptic soy broth (TCS) media with 15 µg/mL tetracycline and 50 µg/mL kanamycin at 37°C. In the morning, the cultures were sub-cultured in TCS and grown for 2.5 hours with shaking at 37°C to reach OD₆₀₀ ~ 1. Cultures were normalized to the same number of cells based on the OD₆₀₀ and pelleted at 13,000 rpm (~16,000 xg) for 1 min before resuspending in PBS and 10 µM Congo red and incubated for 30 mins shaking at 37°C. Pellets and supernatants were separated by centrifugation and supernatants were transferred to new tubes and precipitated with 10% TCA. Supernatant samples were incubated on ice and subsequently pelleted at 15000 rpm (~21,000 xg) for 15 mins at 4°C. Pellets were then washed with ice-cold 100% ethanol and resuspended in 1x loading dye buffer with 5% β-mercaptoethanol. Samples were boiled for 10 mins at 95°C before loading on a 12% SDS-PAGE gel for western blot analysis.

RNA isolation and RNA-seq analysis

Y. pseudotuberculosis IP2666pIB1 strains were grown in 3 mL LB overnight at 26°C and diluted the following morning to an OD₆₀₀ of 0.2 in 5 mL of fresh LB. Each strain was diluted into two separate cultures (one for each temperature). For 26°C samples, cultures were diluted in plain LB media and grown shaking at 26°C for 3 hrs before extracting RNA. For 37°C samples, cultures were diluted into LB low calcium media (containing 20 mM sodium oxalate and 20 mM MgCl₂) and grown shaking at 26°C for 1.5 hours before shifting to 37°C to growing shaking for an additional 1.5 hours. After incubating, 1–2 mL of each culture was collected and pelleted for 5 mins at 4,000 rpm (~200 xg). Supernatants were removed and cell pellets were resuspended in 1mL RNA Protect (Qiagen) and total RNA was isolated using the RNeasy minikit (Qiagen) according to the manufacturer’s protocol. DNA was removed from samples using the TURBO DNA-free kit (Life Technologies/Thermo Fischer) and RNA quality was determined using an Agilent 2000 Bioanalyzer. Samples were DNase treated with Invitrogen DNase. Library preparation was performed using Illumina’s Stranded Total RNA Prep Ligation with Ribo-Zero Plus kit and 10 bp IDT for Illumina indices. Sequencing was done on a NovaSeq 6000 giving 2x51bp reads. Demultiplexing, quality control, and adapter trimming was performed with bcl-convert (v4.0.3). Read mapping was performed with HISAT2 [106]. Read quantification was performed using Subread’s featureCounts functionality [107]. Read counts were normalized using edgeR’s Trimmed Mean of M values (TMM) algorithm (S1 Data) [108]. Subsequent values were then converted to counts per million (cpm). Differential expression analysis was performed using edgeR’s exact test for differences between two groups of negative-binomial counts with an estimated dispersion value of 0.1. Genes for which the |logFC| > 1 and p < 0.05 were considered differentially expressed (S1 Data).

Heatmaps for pYV-encoded genes were generated using average TMM values via ClustVis [109]. RNA-Seq data are available for viewing, and the track hub data can be found at http://lowelab.ucsc.edu/hubs/yersinia_pcnB_pil_expression/hub.txt on the UCSC Microbial Genome Browser. For these data, chromosomal reads were normalized separately from pYV-encoded reads due to differences in pYV PCN among samples. For this analysis, raw paired reads were aligned to the Y. pseudotuberculosis IP2666pIB1 genome (ASM381434v1) using Bowtie2 (v.2.3.5.1) with default options. Bam files were further processed to remove PCR duplicates using samtools (v1.15.1) as follows, samtools sort -n {input_bam} | samtools fixmate -m - - | samtools sort - | samtools markdup -r - - | samtools view -bq 10 -f 0x1 -> {output_bam}. Finally, bamCoverage-deeptools (v3.5.4) was used to normalize the chromosomal read counts using the following options: --normalizeUsing CPM --ignoreForNormalization NZ_CP032567.1 --outFileFormat bedgraph --binSize 1 --region NZ_CP032566.1. For the non-normalized pYV reads we used bamCoverage-deeptools with options --ignoreForNormalization NZ_CP032566.1 --outFileFormat bedgraph --binSize 1 --region NZ_CP032567.1. Finally, the separate bedgraph files were concatenated and converted to bigwig format using bedGraphToBigWig (v4) with default options.

Western blot analysis

Cells were grown at the indicated temperature before preparing samples for western blotting. Samples were normalized to the same number of cells using OD₆₀₀ at the time of sample collection. Bacterial pellets were then resuspended in FSB containing 0.2M DTT and boiled at 95°C for 15 mins. Secreted proteins were precipitated using TCA in the same way as described for the in vitro secretion assay above. The same volume of each sample was separated on a 12.5% SDS-PAGE gel and proteins were transferred to a blotting membrane (Immunobilon-P) using a wet mini trans-blot cell (Bio-Rad) at 4°C. Blots were blocked for 1 hr in Tris-buffered saline containing Tween 20 and 5% milk. Blots were probed with anti-RpoA (gift from Melanie Marketon), goat anti-YopE (Santa Cruz Biotech), mouse M2 anti-FLAG (Sigma), and horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotech). Bands were visualized using a BioRad ChemiDoc and quantification analysis was performed using Image Lab software (Bio-Rad).

For S. flexneri strains, samples were separated by a 12% SDS-PAGE gel and transferred to a nitrocellulose membrane. Blots were probed with anti-IpaB, anti-IpaC, and anti-IpaD antibodies, and anti-rabbit secondary antibodies. Bands were visualized using SuperSignal West Pico PLUS Chemiluminescent Substrate on an Azure 300 Imager.

Mouse Infections

C57BL/6 mice were purchased from Jackson Laboratory. Mice were housed in specific-pathogen free conditions in microisolator cages, and provided with food and water ad libitum. Six- to eight-week-old C57Bl/6 mice were infected with 1-3x10³ Y. pseudotuberculosis via intraperitoneal (IP) injection. LB overnight cultures of Y. pseudotuberculosis were grown shaking at 26°C before spinning down at 5000 rpm (~300 xg) and resuspending in sterile PBS (pH 7.4). CFU used for the inoculum cultures were confirmed to contain pYV prior to infection using single colony PCR with primers specific to the pYV plasmid. Samples were then diluted in PBS such that 100 µL contained 1-3x10³ cells. Inoculums were plated to confirm the inoculation dose and to estimate the percentage of cells that were pYV+ by patching onto LB plates supplemented with 20 mM sodium oxalate and containing 1% Congo red dye. The plasmid retention of the starting inoculums was found to be 100% for wild type and 81% on average for ∆pcnB. Mice were injected with 100 µL of inoculum and spleens and livers harvested four days post-inoculation. Organs were homogenized and serial dilutions of the homogenate plated to determine CFU per gram (CFU/g) of tissue.

Construction of PAP I expression vector

The pcnB ORF was PCR amplified from Y. pseudotuberculosis genomic DNA using the primer pair FpcnB_pTrc/RpcnB_pTrc and ligated into pTrc99a previously digested with NcoI and Xbal using NEBuilder HiFi DNA Assembly kit (New England Biolabs, Inc.). Constructs were transformed into DH5α E. coli for propagation and sequence verified via Sangar sequencing. The resulting pTrc99a::PAPI recombinant expression vector was introduced to wildtype and ∆pcnB Y. pseudotuberculosis by electroporation and selection on LB plates containing carbenicillin. Expression of PAPI was induced by plating onto LB plates containing 0.2 mM IPTG.

Spot tests to analyze antibiotic resistance conferred by pGFPuv, pTrc99, and pNF06

Wildtype and ∆pcnB Y. pseudotuberculosis IP2666pIB1 were electroporated with 1 µL of pGFPuv [Carb^R] (Gifted by Dr. Manel Camps), pTrc99a, or pNF06::ccdAB [Kan^R] (Gifted by Dr. Laurence Van Melderen) and single colonies were selected by plating on LB plates containing carbenicillin (pGFPuv, pTrc99a) or kanamycin (pNF06). For spot test dilution series analyses, overnight cultures were grown at 26°C in the presence of antibiotics before diluting in PBS (pH 7.4) and making a series of 1:10 dilutions. Serial dilutions were spotted onto LB plates containing no antibiotic or LB plates containing the selective antibiotic and plates were visualized after 1-1.5 days of growth at 26°C.

∆ipa2.5::tet^R and ∆ipa2.5∆pcnB S. flexneri M90Twere electroporated with pTrc99a, and single colonies were selected by plating on LB plates containing ampicillin. For spot tests, overnight cultures were grown at 30°C in the presence of antibiotics before diluting in PBS (pH 7.4) and making a series of 1:10 dilutions. Serial dilutions were spotted onto LB plates containing no antibiotic or LB plates containing the selective antibiotic and plates were visualized after 1 day of growth at 30°C.

pGFPuv plasmid loss curve

Wildtype and ∆pcnB Y. pseudotuberculosis IP2666pIB1 lacking pYV (pYV-) were electroporated with pGFPuv as described above. Overnight cultures were grown at 26˚C in the presence of carbenicillin. Overnight cultures were diluted 1:1000 into LB media without antibiotics to allow loss of the plasmid. To assess plasmid retention at each timepoint, aliquots of the cultures were plated onto LB agar and incubated at 26˚C. Fifty single colonies from each sample were patched onto LB agar containing carbenicillin. Percentage of plasmid retention was calculated as the number of patched colonies able to grow in the presence of carbenicillin relative to the total number of patched colonies.