|Home | About | Journals | Submit | Contact Us | Français|
The twin-arginine translocation (Tat) system mediates the secretion of folded proteins that are identified via an N-terminal signal peptide in bacteria, plants, and archaea. Tat systems are associated with virulence in many bacterial pathogens, and our previous studies revealed that Tat-deficient Yersinia pseudotuberculosis was severely attenuated for virulence. Aiming to identify Tat-dependent pathways and phenotypes of relevance for in vivo infection, we analyzed the global transcriptome of parental and ΔtatC mutant strains of Y. pseudotuberculosis during exponential and stationary growth at 26°C and 37°C. The most significant changes in the transcriptome of the ΔtatC mutant were seen at 26°C during stationary-phase growth, and these included the altered expression of genes related to virulence, stress responses, and metabolism. Subsequent phenotypic analysis based on these transcriptome changes revealed several novel Tat-dependent phenotypes, including decreased YadA expression, impaired growth under iron-limited and high-copper conditions, as well as acidic pH and SDS. Several functionally related Tat substrates were also verified to contribute to these phenotypes. Interestingly, the phenotypic defects observed in the Tat-deficient strain were generally more pronounced than those in mutants lacking the Tat substrate predicted to contribute to that specific function. Altogether, this provides new insight into the impact of Tat deficiency on in vivo fitness and survival/replication of Y. pseudotuberculosis during infection.
IMPORTANCE In addition to its established role in mediating the secretion of housekeeping enzymes, the Tat system has been recognized as being involved in infection. In some clinically relevant bacteria, such as Pseudomonas spp., several key virulence determinants can readily be identified among the Tat substrates. In enteropathogens, such as Yersinia spp., there are no obvious virulence determinants among the Tat substrates. Tat mutants show no growth defect in vitro but are highly attenuated in in vivo. This makes Tat an attractive target for the development of novel antimicrobials. Therefore, it is important to establish the causes of the attenuation. Here, we show that the attenuation is likely due to synergistic effects of different Tat-dependent phenotypes that each contributes to lowered in vivo fitness.
Bacteria have evolved several specialized secretion systems for protein export as part of their successful strategies to colonize niches where they encounter various environmental conditions. Gram-negative bacteria possess different mechanisms to transport/export proteins, including virulence-related factors from the cytoplasm across the inner membrane and/or the outer membrane, either in a single step or acting together with other secretion systems (1). Two pathways, the secretory (Sec) and the twin-arginine translocation (Tat) pathways, are involved in the translocation of proteins from the cytoplasm to the periplasm. In Gram-negative bacteria, they can cooperate with other secretion systems to promote the secretion of virulence effectors across the outer membrane to the external environment (2).
The Tat system was first discovered in plant chloroplasts and designated the Cp-Tat pathway. Later, homologues of the pathway were found in bacteria and archaea. The Tat translocation complex consists of three inner membrane proteins, TatA, TatB, and TatC. Of these, TatC is the most highly conserved protein among bacteria, plants, and in archaea. One unique hallmark of the Tat pathway is the ability to translocate fully folded proteins across the cytoplasmic membrane. The translocation process relies on the proton motive force (PMF). Proteins that are translocated via the Tat pathway have an N-terminal signal peptide ([S/T]-R-R-X-F-L-K) with a distinctive “twin-arginine” motif (3, 4). Even though the Tat-dependent proteins are relatively few compared to those secreted by the Sec system, they still have a variety of important functions with impact on bacterial cell physiology, such as respiratory energy metabolism, iron acquisition, stress response, and cell division. The Tat pathway also has an important role in the transport of virulence factors in pathogenic bacteria, such as phospholipase toxins, PlcC and PlcH in Pseudomonas aeruginosa, and Shiga toxin A1-B1 subunits in Escherichia coli O157:H7 (5, 6). However, Tat mutants of some other pathogenic bacteria not known to harbor any Tat-dependent virulence factors show pleiotropic phenotypes, such as reduced or loss of motility, impaired biofilm formation, delayed cell division, and reduced survival under different stress conditions. These phenotypes linked to Tat functions result in virulence attenuation, reduced in vivo fitness, and survival in the infected host (7,–12).
The three human-pathogenic Yersinia species are Yersinia pestis, the causative agent of bubonic and pneumonic plague, and the enteric pathogens Y. enterocolitica and Y. pseudotuberculosis. The most studied secretion system in Yersinia is the Ysc-Yop type three secretion system (T3SS), which is encoded by a 70-kb virulence plasmid (pYV). The T3SS enables targeting of virulence effector proteins denoted Yersinia outer proteins (Yops) to different immune cells where they interfere with host immune response to promote infection (13, 14). We have previously shown that a loss of Tat function in Y. pseudotuberculosis results in strong virulence attenuation. The level of attenuation was similar to that of T3SS knockout mutants without an apparent effect on function or expression of the T3SS (15). In addition, no in silico-predicted Tat substrates could be linked directly to virulence (15). To investigate the reason for the avirulent phenotype of the Y. pseudotuberculosis Tat mutant, we compared the global gene expression profiles of wild-type and ΔtatC mutant strains of Y. pseudotuberculosis by microarray analysis. We found that the transcription of genes involved in virulence, as well as stress adaptation and central metabolic pathways, was reprogrammed. This suggested that several important virulence-related functions could contribute to the pleiotropic phenotype of a Tat-deficient strain. Some phenotypes were confirmed by direct functional assays, and these included decreased YadA expression, impaired growth under iron-limited and high-copper conditions, sensitivity to acidic pH, and SDS. This provides novel information for the function of Tat pathway on in vivo survival of Y. pseudotuberculosis during infection.
The bacterial strains and plasmids used in this study are listed in Table 1. Bacteria were routinely cultured in Luria-Bertani (LB) broth, Yersinia selective agar (YSA), or liquid modified Higuchi's (TMH) medium (16) at 26°C or 37°C for Y. pseudotuberculosis IP32953 and at 37°C only for E. coli with aeration. For culturing of Y. pseudotuberculosis at 37°C, the LB medium was supplemented with 2.5 mM CaCl2, and strains were pregrown at 26°C for 1 h before shifted to 37°C. For the virulence plasmid-cured strains (pYV−), cultures were grown without CaCl2 supplementation at 37°C. The final antibiotic concentrations used for selection were 25 μg/ml chloramphenicol and 100 μg/ml carbenicillin. Isopropyl-β-d-1-thiogalactopyranoside (IPTG) was added to the medium at a final concentration of 0.4 mM to induce gene expression for genes cloned under control of the tac promoter.
In-frame deletion of the tatC and cynT genes was performed as previously described (17). Insertion mutants were constructed by cloning internal fragments of 300 to 350 bp of the genes encoding fhuD, amiC, cueO, and sufI into plasmid pNQ705 by the HD In-Fusion kit (Clontech, USA), according to the manufacturer's instructions, and then introduced directly into the conjugal donor strain E. coli S17-1 λpir. The resulting donor strains were then conjugated with the recipient Y. pseudotuberculosis IP32953 strain to generate insertion mutants, and the correct insertions were confirmed by PCR. For trans complementation of the tatC deletion mutant, plasmid pML31 carrying the tatC gene (15) was transferred into the recipient Yersinia mutant strain by conjugation.
RNA isolation for microarray analysis was performed as previously described (18). Briefly, overnight cultures of Y. pseudotuberculosis IP32953 (wild type [WT]) and the ΔtatC mutant were diluted to an optical density at 600 nm (OD600) of 0.1 and grown until mid-exponential phase (OD600, 0.8) and early stationary phase (OD600, 2.0) at both 26°C and 37°C. The sampling was based on established growth curves for the conditions used, where exponential growth was seen between OD600 of 0.2 and 1.6, and a final OD600 reached at late-stationary phase was 4.0. Sixteen independent cultures were grown for each of the four growth conditions. One milliliter of two independent cultures of each condition and strain were pooled and mixed with a 0.2 volume of stop solution (5% water-saturated phenol in 95% ethanol) and snap-frozen in liquid nitrogen. The resulting 8 samples for each strain and growth conditions were thawed on ice, centrifuged (2 min at 14,000 rpm and 4°C), and RNA was isolated using the SV total RNA purification kit (Promega), according to the manufacturer's instructions. RNA concentration and quality were determined by measurement of A260 and A280, and samples were stored at −80°C.
Sequences used for the design of the microarrays (8 × 15 K format; Agilent), containing three different 60-nucleotide (nt) oligonucleotides for all 4,172 chromosomal genes (open reading frames [ORFs], >30 codons) of the Y. pseudotuberculosis YPIII genome and six probes for the 92 genes of the virulence plasmid pYV of Y. pseudotuberculosis strain IP32953, were obtained from the NCBI GenBank database (accession numbers NC_010465 and NC_006153). The ORF-specific oligonucleotides were designed using the Web design application eArray from Agilent. In total, four RNA replicates were obtained by pooling two replicates for each strain and growth condition, and 1 μg of each was used in hybridization experiments, as previously described (18). After washing and drying of the microarray slide, data were scanned using Axon GenePix personal 4100A scanner, and array images were captured using the software package GenePix Pro 6.015.
Total RNA was prepared from three independent cultures of the wild-type and tatC mutant strains grown to logarithmic and stationary phase at 26°C or 37°C using the same protocol as for the microarray analysis. Five hundred nanograms of total RNA was converted to cDNA with the Revert-Aid first strand cDNA synthesis kit (Thermo Scientific). The gene-specific primers designed to give a 200- to 250-bp PCR product are listed in Table S1 in the supplemental material. Real-time quantitative PCR analysis was performed in triplicate, and dilutions of cDNA on each run were performed to control PCR efficiency, by using Kapa SYBR Fast Bio-Rad iCycler 2× qPCR master mix (Kapa Biosystems) on a Bio-Rad i5 light cycler. The expression levels were normalized to levels of the rpoA gene as a reference, and the relative expression levels of each gene were calculated as previously described (19).
Phenotypic assays were performed at 26°C and 37°C. For the assays performed at 37°C, the virulence plasmid-cured variants (pYV−) of the same strains were used to omit the effect of CaCl2. For growth under iron-limited conditions, the glassware used in the assay was first rinsed with 8% HCl and then with double-distilled water. The different strains were grown overnight at 26°C in TMH medium where FeCl3 was omitted (TMH −Fe). The cultures were then diluted in TMH −Fe medium supplemented with 50 μM 2-2 dipyridyl (VWR, Sweden) or with 10 μM ferrichrome (Sigma) to an OD600 of 0.05, and growth at 26°C or 37°C was followed for 10 h. In order to monitor the acidification of the medium, overnight cultures were diluted to an OD600 of 0.1, and 5 μl was spotted on LB agar that was supplemented with 0.2% glucose and 0.04 g/liter phenol red and incubated at 26°C. A copper susceptibility assay was carried out with overnight cultures that were pelleted and resuspended in 2 ml of LB medium to an OD600 of 0.05 containing different concentrations of CuCl2 (Merck). Bacterial growth at either 26°C or 37°C was determined by the OD600 after 6 h. Percent survival was calculated by the OD600 value of bacteria grown in LB with copper divided by the OD600 value of bacteria grown in LB and multiplied with 100. Survival at low pH and susceptibility to the hydrophobic detergent SDS were determined as described previously (20) but with some minor modifications. For survival at low pH, overnight cultures grown at 26°C were diluted to an OD600 of 0.1 in 5 ml and grown at 37°C until they reached mid-exponential phase (OD600, 0.5). The cultures were centrifuged at 4,000 rpm for 10 min and resuspended both in LB broth as a control and LB broth with pH adjusted to 3 by the addition of 100 mM citrate, followed by incubation at 37°C for 15 min. After the low-pH exposure, the cultures were serially diluted and plated on Luria agar (LA) or LA containing 100 μg/ml carbenicillin. CFU were counted after 48 h. To test the susceptibility to SDS, overnight cultures of pYV− and pYV+ strains were diluted to an OD600 of 0.1 and grown for 3 h at 37°C and 26°C, respectively. Viable bacterial numbers were enumerated after serial dilution and plating onto LA plates with 0.0125% (for 26°C) and 0.00625% (for 37°C) SDS, followed by incubation at 26°C and 37°C for 48 h.
To monitor YadA expression, cell extracts were prepared as described previously (21) from different strains grown to an OD600 of 0.8 in LB medium with 2.5 mM CaCl2 (T3SS noninducing) or 5 mM EGTA and 20 mM MgCl2 (T3SS inducing) at 37°C. One milliliter of the culture was centrifuged for 1 min at top speed, and the pellet resuspended in 1× SDS sample buffer. After electrophoresis on SDS–7.5% polyacrylamide gel, proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore) and incubated in Tris-buffered saline with Tween 20 (TBST) with polyclonal anti-YadA (1:1,000) antibody or anti-DnaK antibody (1:5,000) as a primary antibody and horseradish peroxidase-conjugated donkey anti-rabbit antibody (1:10,000) (Amersham) in TBST as a secondary antibody. The membrane was developed using a chemiluminescence method (Millipore) and analyzed in an LAS4000 (Fuji, Inc.) instrument.
Overnight cultures of bacterial strains were diluted to an OD600 of 0.1 and incubated at 26°C until they reached an OD600 of 1.0. Cultures were then diluted to an OD600 of 0.05 in 3 ml of TMH medium supplemented with 0.4% glucose and 1 μM Na2MoO4·(H2O)2 in the presence of alternative electron acceptors (0.4% dimethyl sulfoxide [DMSO], 50 mM nitrite, 50 mM nitrate, 50 mM formate, 50 mM fumarate). Two-milliliter screwcap tubes were filled on top with bacterial cultures to create a microaerophilic environment. OD600 values were determined after static incubation for 20 h at 26°C.
Overnight cultures of the different strains were diluted to an OD600 of 0.1 and incubated at 26°C until they reached to an OD600 of 2.0. Cultures with equal CFU calculations were centrifuged at 4,500 rpm for 15 min, and pellets were resuspended in assay buffer. Bacteria were lysed by sonication (10-s pulse intervals for 2 min, 30% amplitude), and the fumarate concentrations in the cell lysate were quantified with the fumarate assay kit (Sigma), according to the manufacturer's instructions.
Microarray data analysis was performed as previously described in reference 22 using the Limma package from the R/Bioconductor framework (23). All array data generated in this study were deposited in the Gene Expression Omnibus database and are available under accession number GSE80532.
To investigate the global changes and physiological consequences of the loss of Tat function, a microarray study was performed. The transcriptomes of wild-type and ΔtatC mutant strains of Y. pseudotuberculosis IP32953 were compared under four different culture conditions: exponential and early stationary-growth phase at environmental and host temperatures (26°C and 37°C, respectively). We determined the genes that were differentially regulated (fold change, ≥1.8; P < 0.05) in the ΔtatC mutant strain. During growth at 37°C, 41 genes (15 upregulated and 26 downregulated) and 18 genes (14 upregulated and 4 downregulated) were differentially expressed in early stationary-phase and exponential growth, respectively. During growth at 26°C, 389 genes (200 upregulated and 189 downregulated) and 56 genes (23 upregulated and 23 downregulated) were differentially expressed in stationary-phase and exponential growth, respectively. This highlights that a loss of Tat function results in a reprogramming of transcription under all growth conditions, and changes are especially significant at environmental temperature during stationary phase. Using PRED-TAT, Y. pseudotuberculosis is predicted to encode 27 Tat substrates that function in cell cycle, respiration, iron acquisition, and transport (24). Loss of functional targeting of these substrates in the tatC mutant is likely to have an impact on the physiology and the transcriptome of the bacteria, which is also reflected in the large number of differentially expressed genes seen in the tatC mutant.
To clarify which metabolic and/or physiological pathways were influenced by this global change in the transcriptome, we performed a functional gene clustering analysis of differentially expressed genes. KEGG database and BLAST searches were used to identify the functional categories of the affected genes (Fig. 1A; see also Table S2 in the supplemental material). During stationary-phase growth at 26°C, a large number of genes predicted to be involved in virulence, stress adaptation, metabolism, and transport were differentially regulated in the absence of TatC (Fig. 1A; see also Table S2). The majority of the differentially regulated genes at 37°C at stationary phase encoded proteins that are involved in metabolism, and transport. Moreover, during growth at exponential phase at both temperatures, the differentially expressed genes were related to stress responses, metabolism, and virulence genes (Fig. 1A; see also Table S2). The differential expression of a subset of genes identified in the microarray analysis was also validated by real-time quantitative PCR (qRT-PCR) (see Fig. S1 in the supplemental material).
In total, 31 differentially expressed genes were common during exponential and stationary growth at 26°C in the ΔtatC mutant. These included genes involved in virulence, stress response, metabolism, and transport. In contrast, only 6 differentially regulated genes in the ΔtatC mutant were in common at 37°C in the exponential and stationary phase, indicating that more pathways are under Tat control at moderate temperatures (Fig. 1B). Based on this initial characterization of the changes in global transcription, the TatC-deficient strain would be predicted to exhibit multiple phenotypes associated with altered expression of genes related to certain virulence, stress response, and metabolic functions. To learn more about Tat function and with the view of identifying novel Tat substrates, we decided to perform a number of functional studies to experimentally verify several of the phenotypes that could be predicted based on the transcriptome analysis.
Efficient iron acquisition is a prerequisite for pathogenic bacteria to replicate in the host. Interestingly, several genes encoding iron transport and storage proteins were differentially expressed in the ΔtatC mutant. The bfr gene encoding bacterioferritin was upregulated under all conditions tested (see Table S2 in the supplemental material). The role of bacterioferritin is to store and/or supply iron and reduce its toxicity (25). As such, bacterioferritin has been shown to be important for the virulence of Mycobacterium tuberculosis (26). Although less is known about its function in the Yersinia species, a 19-kDa cytoplasmic bacterioferritin-like protein in Y. pestis is suggested to be involved in iron storage (27). In addition, expression levels of three genes, fbpC, fbpB, and fbpA, in Y. pestis proposed to be the proteins of an iron uptake system were also upregulated in the ΔtatC mutant (see Table S2). Therefore, we decided to investigate growth of the ΔtatC mutant under iron-limited conditions. A mutant lacking the FhuD protein (ΔfhuD), a predicted Tat substrate, known to be involved in iron-hydroxamate siderophore uptake in Yersinia (28), was also evaluated for growth under iron-limited conditions. Compared to the wild-type strain, the ΔtatC mutant showed a severe growth defect in iron-limited media at both 26°C and at 37°C (pYV− derivatives of the same strains were used), which could be complemented with a wild-type copy of the tatC gene supplied in trans (Fig. 2A and andB).B). The ΔfhuD mutant also showed impaired growth at host temperature, but the defect was less severe than for the ΔtatC mutant (Fig. 2B). On the other hand, at 26°C, the ΔfhuD mutant showed a growth defect almost as severe as that with the ΔtatC mutant (Fig. 2A). Hence, FhuD function is more important at 26°C, when changes in the ΔtatC mutant transcriptome are most dramatic. This also suggests that other iron acquisition systems, such as the Ybt system that is known to be expressed mainly at host temperature, could have a preferred role in iron uptake at 37°C, and that a functional Tat system is required for the function of one or more iron uptake systems (29). To investigate the involvement of FhuD in the growth defect of the tatC mutant at 37°C and to look into the Tat dependency of FhuD export, we also investigated the growth of these strains in iron-limited medium containing ferrichrome. FhuD functions in the uptake of ferric-hydroxamate-type ferrichrome, which is an iron-chelating siderophore (28). As a control, we also included a ΔybtP mutant, as ybtP is a gene required for another independent iron uptake system. The ΔfhuD and the ΔtatC mutant strains showed similar and highly impaired growth phenotypes, arguing that the growth defect observed for the tatC mutant was caused by the failure to translocate FhuD to periplasm and, as a consequence, to take up ferrichrome (Fig. 2C). This phenotype was readily complemented by expression of an intact tatC gene in trans. As expected, the addition of ferrichrome did not influence growth of the ybtP mutant strain, as this mutation affects another nonrelated iron uptake system.
Overall, our findings show that a functional Tat system is required for iron uptake and that this at least in part is due to a loss of FhuD translocation across the inner membrane. At host temperature, the growth defect under iron limitation is more severe in the ΔtatC mutant strain than in the fhuD mutant. This indicates that the defect cannot be solely due to a loss of FhuD function. Finally, the impaired growth under iron-limited conditions would be expected to significantly impact in vivo fitness and virulence.
Several of the genes encoding T3SS components, including secreted effectors (yopE, yopK, and yopM), translocators and regulators (lcrV, lcrG, and yopD), and the secretome components (yscN and yscO), were downregulated at 26°C in stationary phase (see Table S2 in the supplemental material). However, under these conditions, the expression of T3SS genes is overall low, and none of the genes encoding T3SS components were differentially regulated at 37°C. This finding was in line with our previous study that a loss of Tat function did not have any impact on secretion and expression of Yops- or YopE-mediated cytotoxicity in an in vitro cell infection model (15).
Y. pseudotuberculosis YPIII harbors four type VI secretion system clusters (T6SS-1 to -4). Interestingly, we observed that 11 out of 18 genes in a homologous T6SS-4 cluster in the IP32953 strain were upregulated in the ΔtatC mutant background at 26°C in stationary phase (see Table S2 in the supplemental material). Previous reports have shown the T6SS-4 cluster to be highly expressed at 26°C during stationary phase but downregulated at 37°C (30). The T6SS-4 cluster is important for Y. pseudotuberculosis survival under acidic conditions, high osmolarity, and exposure to the detergents, such as deoxycholate, that impair outer membrane integrity. Consistent with this, T6SS-4 is known to be regulated by RpoS and might function as a part of the general stress response (31,–33). Hence, the observed upregulation in the TatC-deficient strain at 26°C could be a response to alterations in the outer membrane integrity (see also below) and the consequent upregulation of one or more stress response genes involved in the maintenance and repair of this structure.
Finally, expression of yadA encoding the multifunctional adhesin was highly downregulated at 26°C, and in contrast to the other pYV-carried T3SS genes, yadA was also downregulated during exponential growth at 37°C (see Table S2 in the supplemental material). We could also demonstrate the downregulation of YadA production significantly in the tatC mutant by immunoblotting using an anti-YadA antibody, when bacteria were grown in both the presence and absence of Ca2+ at 37°C (Fig. 3). YadA is encoded by the common pYV virulence plasmid of human-pathogenic Yersinia spp., and expression is induced by a temperature shift to 37°C (34). Significantly, YadA has been shown to be crucial for the virulence of Y. enterocolitica (35). There is no obvious direct link between a functional Tat system and expression of YadA, suggesting that the observed effect on expression is likely to be mediated through another unidentified pathway. In a mouse model, Y. pseudotuberculosis yadA mutants can still cause a systemic infection but with significantly lower numbers of bacteria in systemic organs (36). A role for YadA in persistence and systemic dissemination has been shown to occur in cooperation with the Ail adhesion (37) and involves an interaction with neutrophils (38). Based on this finding, we speculate that a downregulation of YadA expression could constitute an impediment during the initial in vivo colonization by the ΔtatC mutant.
Functional annotation analysis revealed that under all tested conditions, a significant fraction (23%) of the differentially regulated genes are involved in metabolism. The most notable effect occurs during stationary-phase growth at 26°C, where 40% of the 91 differentially regulated genes are implicated in carbon metabolism, while others are involved in amino acid, nucleotide, energy, cofactor, vitamin, and lipid metabolism (Fig. 4A). Upregulation of genes encoding enzymes involved in glycolysis, such as pgi, pgk, eno, pykF, pfkA, and gpmA, indicates that the glycolytic pathway is induced in response to the lack of a functional Tat system. To confirm that there was an upregulation of glycolysis, we compared the growth of the WT and ΔtatC mutant strains in LB agar containing glucose and phenol red. Phenol red is a pH indicator; if the pH is acidic, the medium turns from orange to yellow, or it turns pink if the pH is alkaline. After growth at 26°C, there was a very clear yellow color around the colony of the ΔtatC mutant (Fig. 4C). The acidification of medium by the ΔtatC mutant is in line with the observed upregulation of glycolysis that would result in higher levels of acidic end products. However, the medium around colonies of the WT and the trans-complemented ΔtatC mutant strain both turned pink, which indicated that they both fermented the peptones in the medium and secreted alkaline end products (Fig. 4C). Both the dld and lld genes that encode lactate hydrogenases were also upregulated in the ΔtatC mutant, indicating the excess pyruvate had been oxidized to lactic acid, confirming the intense secretion of acidic end products. It is not clear why the ΔtatC mutant uses this pathway more actively, since glycolysis yields low energy and that would render a low metabolic state, especially in stationary phase. Interestingly, genes encoding enzymes involved in succinate biogenesis through the oxidative branch of the tricarboxylic acid (TCA) cycle (sdhABCD and sucBC) were downregulated, while the expression of the frdABCD operon that is involved in the reductive branch of the TCA cycle was upregulated (Fig. 4B; see also Table S2 in the supplemental material). These changes in central carbon metabolism are similar to those previously described for E. coli and Shigella flexneri upon entry into stationary phase (39, 40), and it is noteworthy that these changes were far more pronounced in a Tat-deficient strain than in the wild type. Additionally, the expression of argH encoding argininosuccinate lyase that produces fumarate from aspartate and the downstream urease operon (ureABDCEG) was also upregulated (Fig. 4B; see also Table S2), which could cause the upregulation of fumarate reduction. These findings were further supported, as they also correlated with increased intracellular fumarate concentration in the ΔtatC mutant. The ΔtatC mutant had an average of 29.5 ng/μl fumarate, whereas the parental strain contained 18.2 ng/μl (Fig. 4D). The phenotype was partially complemented by expressing ptatC in trans with fumarate levels of 22.1 ng/μl (Fig. 4D). Furthermore, upregulation of the reductive branch of the TCA cycle (frdABCD) suggests implication in anaerobic respiration pathways in response to lower oxygen availability. It is known that bacteria repress aerobic metabolism in late-stationary phase (41), and also there are 6 predicted Tat substrates in Y. pseudotuberculosis with possible function in anaerobic respiration and that could result in the upregulation of anaerobic-like metabolism in the ΔtatC mutant. Importantly, however, we found the Tat-deficient strain was not impaired in in vitro growth under anaerobic conditions (data not shown), similar to what was reported for Salmonella species (8).
Approximately 26 genes encoding ribosomal proteins and various tRNA synthetases were found to be downregulated at 26°C in stationary phase in the ΔtatC mutant, which is again the hallmark of bacterial stationary phase (42). Altogether, the changes in central carbon and energy metabolism in the ΔtatC mutant are a clear indication that Tat-deficient strains are more stressed in harsh environments, such as those with nutrient starvation, and this would be expected to negatively impact their in vivo fitness.
High copper levels are toxic, and bacteria have evolved copper resistance and translocation systems that function to equilibrate the intracellular copper levels. In E. coli, the CopA-CueO and CusCFBA systems are involved in protecting the bacterium during exposure to moderate and high copper concentrations under aerobic and anaerobic conditions, respectively (43). Our transcriptomic analysis showed that the Cop system is differentially regulated in the ΔtatC mutant. At 26°C in stationary phase, the copA gene encoding a copper resistance protein was upregulated 2.4-fold, whereas the copCD genes were downregulated (see Table S2 in the supplemental material). It is notable that the periplasmic multicopper oxidase CueO involved in copper tolerance is an in silico-predicted Tat substrate in Y. pseudotuberculosis (15). Moreover, CueO homologous proteins of E. coli, Salmonella enterica serovar Typhimurium, and M. tuberculosis are established Tat substrates (8, 44, 45). This prompted us to compare the copper tolerance of Y. pseudotuberculosis wild-type, ΔtatC mutant, and ΔcueO mutant strains at different temperatures. When strains were grown in media with increasing copper concentrations at either 26°C or 37°C, the ΔtatC mutant was significantly more sensitive to copper than the wild-type strain, and this sensitivity could be complemented with the ptatC strain in most of the tested concentrations at both temperatures (Fig. 5A and andB).B). The lowered complementation level at elevated copper concentrations (>2 mM CuCl2) might be due to growth difficulties that the parental strain also encounters. Additionally, a ΔcueO mutant showed an intermediate sensitivity, although the putative Tat substrate CueO seemed to be more important for copper tolerance at elevated temperature. Nevertheless, these results suggest that the mislocalization of CueO in the absence of a functional Tat pathway is only partly responsible for the marked copper sensitivity of the Y. pseudotuberculosis ΔtatC mutant. Similar to the finding in E. coli, our data support the hypothesis that Tat deficiency causes a more general effect that cannot be attributed to the mislocalization of any given Tat substrate (46). However, since pathogenic bacteria often must cope with copper stress generated upon their internalization by phagocytes (47), increased copper sensitivity could impair the in vivo fitness of Tat-deficient bacteria.
Several studies indicate that a functional Tat system is required for many pathogenic bacteria, such as Pseudomonas syringae pv. tomato DC3000, S. enterica, Campylobacter jejuni, and Ralstonia solanacearum (5, 48,–51), to efficiently respond to diverse forms of physicochemical stresses. Indeed, our transcriptomic analysis of the Y. pseudotuberculosis ΔtatC mutant revealed that a number of prominent stress-responsive genes were induced at both temperatures, and this number was higher during stationary-phase growth. During growth at 26°C in exponential phase, genes encoding osmotically inducible protein (osmY), osmotic response regulator (ompR), superoxide dismutases (sodCB), sigma 54-modulation protein (yfiA), universal stress protein (uspA), and peptidase (pepT) were upregulated (see Table S2 in the supplemental material). Strikingly, at 26°C during stationary-phase growth, these genes and several additional genes (28 in total) that encode most types of stress responses ranging from acid, heat shock, and stationary phase survival to cold shock, oxidative stress, and high osmolarity were found to be upregulated. Significantly, the rpoS gene encoding the sigma factor S, known to regulate responsiveness to high osmolarity, high temperature, and oxidative stress during entry into stationary phase (52), was also upregulated 4.1-fold at 26°C during stationary growth (see Table S2). The RpoS regulon in Yersinia spp. has not been investigated extensively but was shown to be crucial for adaptation to various stresses as well as biofilm formation and motility in both Y. pseudotuberculosis and Y. enterocolitica (33, 53). Hence, our data provide further evidence that the stationary-phase environment is much harsher for the ΔtatC mutant than for the wild type and results in the induction of many different stress responses.
The acid stress response genes are of particular interest, since survival at low pH is crucial for the enteric route of infection. Interestingly, genes encoding proteins involved in several different acid tolerance systems were upregulated in the ΔtatC mutant. These included the urease operon (ureABCDEG) encoding urease accessory proteins and adiA and adiC encoding amino acid decarboxylase and antiporter systems, respectively. Finally, the hdeB and hdeD genes, encoding the acid-stress-activated periplasmic chaperone HdeB and membrane protein HdeD, respectively, were upregulated in the ΔtatC mutant (see Table S2 in the supplemental material). In addition to its role in metabolism, the urease operon is known to be essential for virulence and acid resistance of enteric pathogens, including Y. enterocolitica (54). It has been suggested that the expression of urease genes is regulated by OmpR in Y. pseudotuberculosis (55), but there is no direct evidence for a role of urease in virulence (56). The adiAC genes, encoding arginine decarboxylase and the arginine-agmatine antiporter, respectively, were highly upregulated (fold change, 7.4 and 6.8, respectively; Table S2). This system consumes cytoplasmic protons and promotes acid survival in E. coli (57). Bacteria that are exposed to acid challenge often upregulate the expression of the respiratory chain complexes that function to pump out protons, while ATP synthases that transport protons to the inside are downregulated (58). We also observed a downregulation of atpHBI encoding the ATP synthases, and that could be an indication of severe acid stress in the ΔtatC mutant, even though we did not observe a corresponding upregulation of specific proton-pumping cytochromes. Nevertheless, these transcription data corroborate previous results where we showed that Tat deficiency impaired the bacterial survival in acidic pH at moderate temperature (15). An interesting observation from our previous study was that the putative Tat substrate carbonic anhydrase (CynT) may play a role in the survival of Y. pseudotuberculosis at low pH (15). This is also in line with the expression levels of cynT, which were upregulated at both tested growth temperatures (see Table S2). For these reasons, we examined a ΔcynT mutant in this study and also performed the experiment at 37°C. We found that the wild-type and ΔcynTc (pYV− derivative) mutant strains showed similar survival rates, whereas the survival rate of the ΔtatCc (pYV− derivative) mutant was much lower in LB medium at pH 3.0 (Fig. 5C). The trans complementation in the strain expressing tatC in trans was incomplete and most likely due to the reasons discussed above. This verifies that the ΔtatC mutant is indeed sensitive to acidic conditions at 37°C but that the putative Tat substrate CynT is actually dispensable for survival under acidic conditions. The attenuation of the tatC mutant in the oral infection mouse model is likely to be caused by the lower survival under acid conditions encountered during the passage through the stomach or upon internalization by host cells.
A pleiotropic defect in the outer membrane relating to Tat function was first seen in E. coli, where Tat mutants were characterized by sensitivity to SDS and hydrophobic drugs (59). These effects were found to be the result of the mislocalization of two Tat-dependent periplasmic amidases, AmiC and AmiA (60). The same was also reported for Salmonella spp., but in this case, the sensitivity in addition involved the mislocalization of the periplasmic protein SufI. These defects were suggested to cause a severe attenuation of the Tat mutant in a mouse model (8). Moreover, a transcriptomic study of E. coli revealed that the rcs envelope stress regulon genes were upregulated in the Tat mutant in response to pleiotropic defects in the outer membrane (46). However, our transcriptomic analysis revealed no differences in the expression of the rcs regulon of Yersinia species (61). In addition, no other known genes encoding envelope and periplasmic stress response regulators, such as rpoE, bae, or inner membrane stress response system psp, were upregulated in the tatC mutant. Based on the results from E. coli and Salmonella, we targeted the role of AmiC and SufI especially in envelope stress, since both are predicted to be Tat substrates in Yersinia (15). Especially, we tested a ΔamiC and a ΔsufI mutant together with a ΔtatC mutant and the wild-type strain for SDS sensitivity at 26°C and 37°C. Y. pseudotuberculosis was found to be highly sensitive to envelope stress at 37°C, which necessitated a decrease in the SDS concentration used compared to that with assays at 26°C. This indicates that envelope integrity and/or permeability are different at host temperature than under environmental conditions. The wild-type, ΔsufI mutant, and ΔamiC mutant all grew equally well on LA plates alone and with SDS, while the ΔtatC mutant was highly impaired for growth (10,000-fold, compared to wild type) on LA plates containing SDS at both growth temperatures (Fig. 6). This clearly shows that Tat deficiency drastically impairs cell envelope integrity, and that this phenotype is not caused solely by the loss of AmiC and SufI. Additionally, these results indicate that the relation of outer membrane integrity to the Tat pathway in Yersinia is different from that in E. coli and Salmonella. Since AmiA has a clear Sec signal peptide in Yersinia, the mislocalization of AmiC alone in the Tat-deficient strain cannot be directly attributed to the defect in outer membrane integrity. However, the downregulation of msbB that is responsible for lipid A biosynthesis and upregulation of the mechanosensitive channel protein encoded by genes mscL and mscS that function in sensing and release of turgor pressure during hypo-osmotic shock could be linked to increased outer membrane permeability changes (62, 63). Therefore, it is that the changes in the outer membrane permeability explain the induced stress response and effects on overall cellular physiology in a more stressful environment, such as stationary phase.
Twin-arginine translocation pathways enable the translocation of fully folded proteins across the cytoplasmic membrane of bacteria. Tat-dependent proteins are involved in highly versatile functions, ranging from respiration to different metal ion homeostasis and transport in the periplasm and inner membrane of bacteria (64). Moreover, the Tat pathway also directly or indirectly plays a role in the virulence of many pathogenic bacteria. In some cases, pathogenic bacteria exhibit decreased virulence in their particular host organism, even though they do not encode readily identifiable Tat-dependent virulence factors. In our previous study, we verified that Tat deficiency in Y. pseudotuberculosis IP32953 resulted in severe virulence attenuation that was similar to that in strains lacking a functional T3SS (15). This was surprising, since none of the in silico-predicted Tat substrates were an obvious virulence factor candidate. In an attempt to resolve this issue, we revealed the global changes in the transcriptional profile resulting from Tat deficiency. This enabled us to predict a number of novel phenotypes linked to virulence or in vivo fitness that could be coupled to Tat deficiency. These included genes encoding proteins with roles in virulence, central metabolism, and stress adaptation. Another important result from this study was that bacteria harboring the engineered loss of an individual Tat substrate predicted to be important for a specific function displayed no phenotypic or only a modest intermediate phenotype, whereas bacteria with an isogenic tatC mutant displayed a much more severe phenotypic defect. This indirect effect could be related to the accumulation of fully folded Tat substrates in the cytoplasm and/or in the inner membrane. It is known that some of the Tat substrates are guided to the inner membrane translocase complex by their cognate chaperone (such as DmsA-DmsD or TorA-TorD) naturally (65). Other studies have demonstrated an interaction of precursor Tat substrates with the inner membrane prior to translocation (66, 67). Since this might also occur in the absence of TatC, it might result in the accumulation and insertion of folded substrates in the inner membrane. It was also suggested that diffusion of the accumulated substrates in the membrane in a Tat-deficient strain of E. coli affects the membrane permeability and inhibits the action of the inner membrane protease FtsH, which is important for lipid biosynthesis. In line with this, the same authors demonstrated that the proton motive force was affected in E. coli Tat mutants due to the leakiness of the membrane (68). Based on these observations, we propose that Tat deficiency in Yersinia could result in outer and inner membrane integrity defects that may compromise cellular physiology and in vivo fitness.
We thank Matthew S. Francis for critical reading of the manuscript.
This work was funded by the Swedish Research Council (grant 2011-3439 to Å.F.) and support from the J. C. Kempe Foundation to U.A.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/JB.00352-16.