Present address: Jason A. Rosenzweig, Division of Math Science and Technology, Nova Southeastern University, Ft. Lauderdale, FL, USA.
Polynucleotide phosphorylase independently controls virulence factor expression levels and export in Yersinia spp.
Article first published online: 28 MAR 2007
FEMS Microbiology Letters
Volume 270, Issue 2, pages 255–264, May 2007
How to Cite
Rosenzweig, J. A., Chromy, B., Echeverry, A., Yang, J., Adkins, B., Plano, G. V., McCutchen-Maloney, S. and Schesser, K. (2007), Polynucleotide phosphorylase independently controls virulence factor expression levels and export in Yersinia spp. FEMS Microbiology Letters, 270: 255–264. doi: 10.1111/j.1574-6968.2007.00689.x
Editor: Mark Schembri
- Issue published online: 28 MAR 2007
- Article first published online: 28 MAR 2007
- Received 8 December 2006; revised 24 January 2007; accepted 30 January 2007.First published online 28 March 2007.
Previously, it was shown that optimal functioning of the Yersinia type III secretion system (T3SS) in cell culture infection assays requires the exoribonuclease polynucleotide phosphorylase (PNPase) and that normal T3SS activity could be restored in the Δpnp strains by expressing just the ∼70-aa S1 RNA-binding domain of PNPase. Here, it is shown that the YersiniaΔpnp strain is less virulent in the mouse compared with the isogenic wild-type strain. To begin to understand what could be limiting T3SS activity in the absence of PNPase, T3SS-encoding transcripts and proteins in the YersiniaΔpnp strains were analyzed. Surprisingly, it was found that the Δpnp Yersinia strains possessed enhanced levels of T3SS-encoding transcripts and proteins compared with the wild-type strains. We then found that an S1 variant containing a disruption in its RNA-binding subdomain was inactive in terms of restoring normal T3SS activity. However, T3SS expression levels did not differ between Δpnp strains expressing active and inactive S1 proteins, further showing that T3SS activity and expression levels, at least as related to PNPase and its S1 domain, are not linked. The results suggest that PNPase affects the expression and activity of the T3SS by distinct mechanisms and that the S1-dependent effect on T3SS activity involves an RNA intermediate.
Host–microorganism interactions typically involve cell surface components and soluble factors, both of which must be exported from the cell. Gram-negative bacteria exhibit a unique variation to this general theme, an export pathway that mediates the direct transfer of bacterially expressed proteins into the host cell. This ‘protein injection system’ (designated as type III secretion) is used to modulate a number of host processes including bacterially induced responses in both plant and animal cells (Galán & Wolf-Watz, 2006). Yersinia and Salmonella pathogenic strains use type III secretion systems (T3SS) to limit host cellular responses that are normally activated following bacterial cell contact. Although both species utilize a T3SS, the proteins they inject (referred to as ‘effectors’) are largely unique and drive contrasting cellular responses. The cellular activity of the yersinae T3S effectors YopE, YopT, and YpkA(YopO) results in actin depolymerization that essentially paralyzes the host cell (Viboud & Bliska, 2005). The cellular activity of the salmonellae T3S effectors SopC, SopE, and SopA also affects actin dynamics, although, in this case, these proteins promote the nucleation, extension, and stabilization of actin filaments (Zhou & Galan, 2001). These disparate effects serve either to block phagocytosis (Yersinia) or sabotage this otherwise normal cellular response in order to invade the cell (salmonellae). In each case, however, their respective T3SSs must rapidly inject these effectors immediately following host cell contact in order to outpace cellular responses that are normally activated following bacterial contact.
As quickly shaping the host cell response is vital to bacterial survival, considerable effort has been invested in determining how T3SSs are regulated. In the Yersinia, many of the constituents that regulate the T3SS have been identified, and the process can be divided into two stages. ‘Step 1’ regulation involves the temperature-dependent expression of the transcriptional activator LcrF(VirF) that directs the expression of the T3SS regulon up to a certain level in the presence of Ca2+ (Lambert de Rouvroit et al., 1992; Hoe & Goguen, 1993). Under these conditions, further expression of the T3SS regulon is repressed by cytosolic LcrQ/SycH and YopD/SycD complexes (Williams & Straley, 1998; Francis et al., 2001; Anderson et al., 2002). The secretion conduit opens up as extracellular Ca2+ levels decline to the low millimolar range, leading to the export of the negative regulators LcrQ and YopD that in turn results in the full derepression of T3SS expression (‘Step 2’) (Pettersson et al., 1996). A Step 2-like derepression of T3SS expression also occurs when the Yersinia interact with cultured vertebrate cells (Pettersson et al., 1996; Bartra et al., 2001).
Several chromosomally encoded gene products have been identified as being necessary for T3SS regulation in the Yersinia. One of the first such factors identified was the small histone-like protein YmoA, which, under noninducing conditions for T3SS gene expression, is present at relatively high levels in the bacterial cell (Mikulskis & Cornelis, 1994). Upon inductive conditions, YmoA is rapidly degraded and consequently its repressive effect on T3SS gene expression is lifted (Jackson et al., 2004). Recently, another such factor, the exoribonuclease polynucleotide phosphorylase (PNPase), was identified as being required for normal T3SS functioning in the pathogenic Yersinia (Rosenzweig et al., 2005). Surprisingly, the effect of PNPase on T3SS activity was independent of its catalytic activity, in contrast to the well-established role that PNPase plays in low-temperature growth. Normal T3SS activity could be restored to the Δpnp Yersinia strains by expressing just the ∼70-aa RNA-binding S1 domain of PNPase as well as the S1 domains found in a number of other proteins including RNase R, RNase II, RNase E, and the ribosomal S1 protein (Rosenzweig et al., 2005 and unpublished observations).
Unique among the ribonucleases, the 5-domain structure of PNPase is remarkably well conserved between prokaryotes and eukaryotes. Structural studies of the Streptomyces antibioticus PNPase showed that PNPase forms a homotrimer with a central cavity comprising the catalytic core and six PH domains that are flanked by extensions that contain the KH and S1 RNA-binding domains (Symmons et al., 2000). These key functional domains display a high degree of sequence conservation in mammalian and plant PNPases, suggesting that similar higher-order complexes exist in eukaryotic cells. In bacteria and plant chloroplasts, PNPase regulates RNA stability by functioning as an exoribonuclease and poly(A) polymerase (Deutscher, 2006). In mammalian cells, PNPase has been associated with cellular differentiation and oncoprotein-mediated transformation (Leszczyniecka et al., 2002; French et al., 2007).
S1 domains were originally identified in the ribosomal protein S1 and are classically defined as a subclass of the oligonucleotide/oligosaccharide-binding (OB-fold) family. This fold is characterized by a distinctive β-barrel core and is found in a variety of proteins that bind nucleic acids and carbohydrates. The S1 domain of PNPase plays a clear role in substrate binding (Stickney et al., 2005). The versatility of S1 domains has recently been extended to include their involvement in protein–protein interactions, most notably in the autoregulation of the NusA protein of Escherichia coli involved in template-specific transcriptional pausing and termination (Mah et al., 2000). The McIntosh group determined the structure of RNase E's S1 domain using both NMR spectroscopy and X-ray crystallography and found that this isolated S1 domain possesses a dimerization interface that is distinct from its nucleic-acid binding region (Schubert et al., 2004). This group bolstered their ‘distinct domain’ claims by developing a very elegant NMR-based assay in which they could identify residues whose amide shifts were sensitive to either S1 protein or oligonucleotide concentrations, implying involvement in dimerization and nucleic acid binding, respectively. The oligonucleotide-binding surface of this same S1 domain has been further refined recently by a structural determination of the S1–oligonucleotide complex (Callaghan et al., 2005).
In the majority of cases cited above, it is unknown how PNPase exerts its cellular effects. Here, the relationship between PNPase and the functioning of the T3SS will be characterized. It is first examined whether the reduced T3SS activity of the YersiniaΔpnp strains observed in secretion and cell culture infection assays is accompanied by a similar reduction in virulence in the mouse model of infection. Then, one possible mechanism by which PNPase and S1 could control T3SS activity is investigated and then the structural attributes of the S1 domain that are required for it to affect T3SS activity positively are analyzed.
Materials and methods
Bacterial strains and plasmids
The parental strains Yersinia pestis (YP) KIM5/3001 and Yersinia pseudotuberculosis (YPT) YPIII/pIB100 yopE::gfp, as well as the construction of their Δpnp derivatives, were described previously (Lindler et al., 1990; Bartra et al., 2001; Rosenzweig et al., 2005). The DNA fragments encoding the S135–125, S150–125, and S165–125 portions of RNase E were generated by PCR using an E. coli K12 chromosomal preparation as a template. The fragments were cloned into the pCR 8/GW/TOPO TA Cloning vector (Invitrogen) that features an arabinose-inducible promoter as well as an amino-terminal Thio-His Patch tag and carboxyl-terminal V5/6XHis epitope tags. The PNP- and PNPF638G-encoding plasmids have been previously described (Jarrige et al., 2002). All cloned DNA was verified by sequencing. Oligonucleotide primer sequences used for the generation of the plasmids described above will be supplied by the authors upon request.
For mouse virulence testing, bacterial stocks of YPT and YPT/Δpnp were prepared by first propagating the strains in Luria–Bertani (LB) media at 27°C with aeration, diluting the saturated cultures 1 : 20 with LB medium, and continuing to propagate the cultures as before until they reached saturation. Bacterial cells were then collected, washed twice with Hank's balanced salt solution (HBSS), resuspended in HBSS+20% glycerol, and stored at −80°C. Prior to infection, frozen bacterial stocks were thawed, washed twice with HBSS, and then diluted serially with HBSS. The actual administered doses were determined by viable plating. Seven-week-old female BALB/c mice, purchased from Harlan Sprague Dawley (Indianapolis, Indiana), were rested for 1 week, during which time they were given food and water ad libitum. Mice were inoculated orogastrically with 100 μL of diluted bacterial suspensions using a syringe attached to a 22-gauge plastic feeding needle (Monack et al., 1998). Mice were monitored for signs of distress twice daily for 30 days and animals that had become unable to move were euthanized and their deaths were included in the analysis. All virulence testing was conducted with the approval of the University of Miami School of Medicine Animal Care Committee.
Cell culture infection assays
For cell culture infection assays, bacterial cultures were propagated overnight in 2 mL of heart infusion broth (HIB-Difco) with the appropriate antibiotics (if necessary) at 26°C. Subcultures were begun the following day at an OD at λ=620 nm (OD620 nm) of ∼0.3. After 1 h of growth at 26°C, 0.2% of l-arabinose (Sigma) was added for induced samples, followed by 1.5 h of additional growth at 26°C. HeLa cells seeded in six-well plates at densities of 1.5 × 105 to 3.0 × 105 were infected with the subcultures at multiplicities of infection (MOIs) between 30 and 60. The severity of the cytotoxicty was quantified and presented as described previously (Rosenzweig et al., 2005).
T3SS expression and secretion assays
For T3SS transcript analysis, cultures of YP and YP/Δpnp were grown to saturation in 2 mL Thoroughly Modified Higuchi's (TMH) media (Zahorchak & Brubaker, 1982) containing 2.5 mM Ca2+ at 37°C. Subcultures were then diluted to an OD620 nm of 0.3 in 3 mL of TMH containing 2.5 mM Ca2+ and were cultured for 3 h at 37°C. Total RNA was extracted using the Ribopure bacterial RNA extraction kit (Ambion) and labeled cDNA synthesized by standard methods. Yersinia pestis-specific microarray slides were provided by the Pathogen Functional Genomics Resource Center (NIAID/NIH), and hybridization and analysis were performed at the University of Miami DNA Microarray Core Facility. For T3SS protein analysis, cultures were propagated as described above and subsequently resuspended in a buffer containing 30 mM Tris, 2 M thiourea, 7 M urea, and 4% Chaps (pH 8.5). A Bradford-based protein assay was performed to determine the initial protein concentration of each sample and samples were prepared and analyzed by two-dimensional differential gel electrophoresis (2D-DIGE) as described (Chromy et al., 2005). Briefly, samples were cleaned using a trichloroacetic acid (TCA)/acetone precipitation/wash and then labeled with CyDye and fractionated by 2-D gel electrophoresis. In the first dimension, separation was according to pI using isoelectric focusing, and in the second dimension according to size using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Two-dimensional gels were imaged and analyzed using decyder software to detect differential protein expression. Proteins were subsequently identified by MS as described (Chromy et al., 2005). For secretion assays, cultures were grown to saturation in TMH media containing 2.5 mM Ca2+ and the appropriate antibiotic (if necessary) at 37°C. Subcultures were then diluted to an OD620 nm of 0.3 in a 3 mL final volume of TMH containing 2.5 mM Ca2+ and 0.2%l-arabinose. After 1.5 h at 37°C, the calcium chealator EGTA (5 mM final concentration) was added to half of the cultures. At the indicated time-points, 1 mL aliquots were removed and placed in prechilled microfuge tubes on ice. After centrifugation at 12 000 g for 7 min, 500 μL was removed from each tube for supernatant fraction analysis, while the bacterial cell pellets were resuspended in sample buffer. One-tenth the volume of TCA was added to the supernatant fractions and left on ice for ∼18 h. TCA-precipitated proteins were collected by centrifugation and resuspended in sample buffer, and together with the pellet fractions, were resolved by SDS-PAGE and analyzed by immunoblotting with the indicated antisera. All samples were normalized based on OD620 nm readings.
Virulence testing in the mouse
The YersiniaΔpnp strains are clearly deficient in T3SS activity in both secretion and cell culture infection assays. To test whether these differences extended to the mouse model of infection, the virulence of YPT and its Δpnp derivative strain (YPT/Δpnp) was compared. BALB/c mice orogastrically with various doses of the two strains and scored for lethality following a 30-day infection period were inoculated. Under these conditions, the lethal dose 50% (LD50) of the wild-type strain was 1.2 × 108 and that of the Δpnp strain was 5.7 × 109 (Table 1). In addition to differences in mortality, there were clear differences in the morbidity between wild-type and Δpnp-infected mice. For example, all four mice in the 2.9 × 108-dosed wild-type group displayed several signs of severe disease (including ruffled fur, hunched backs, body tilting to one side, and general wasting) for several days following their inoculation. In contrast, three of the four mice in the 1.8 × 109-dosed Δpnp group displayed slightly ruffled fur ∼2.5 days following their inoculation but by ∼5 days were in generally good health. As the T3SS plays a well-documented role in yersiniosis in the mouse, the nearly 50-fold increase in the LD50 of the Δpnp strain is consistent with this strain having reduced T3SS activity.
|% Death (day succumbed)|
|2.5 × 1010||–||75 (3, 7, 7)|
|2.5 × 109||100‡ (9, 9)||25 (6)|
|2.5 × 108||50 (12, 18)||0|
|2.5 × 107||25 (17)||0|
|2.5 × 106||25 (30)||0|
|2.5 × 105||0||–|
|LD50§||1.2 × 108||5.7 × 109|
T3SS expression levels in the Δpnp strains
One mechanism by which PNPase could affect T3SS activity is by regulating its expression. In the Yersinia, the majority of the structural and regulative components of the T3SS are encoded by three operons, each of which contains 7–10 ORFs. We used gene arrays to measure T3SS-encoding transcript levels in the Yersinia strains and found that sequences representing all three operons, yscC, yscO, and lcrG, were 2.4-, 1.8-, and 2.2-fold higher, respectively, in the Δpnp strain compared with the wild-type strain (data not shown). Several of the transcripts encoding T3SS effector proteins were also at elevated levels in the Δpnp strain compared with the wild-type strain (e.g. yopE, 3.9-fold; ypkA, 2.2-fold; yopH, 2.7-fold). Together with the previous findings that yopE promoter activity was indistinguishable between the wild-type and Δpnp strains (Rosenzweig et al., 2005), these data support a scenario in which the enhanced steady-state levels of T3SS-encoding transcripts in the Δpnp strain are likely a posttranscriptional effect.
2-D DIGE was used to compare the relative levels of T3SS proteins between the wild-type and Δpnp mutant strains. Although no gross-level differences in YopD and YopE protein levels were observed between the wild-type and Δpnp Yersinia strains in a previous study (Rosenzweig et al., 2005), 2-D DIGE is a markedly more sensitive approach to address this issue and has been successfully applied to analyze differences in the YP proteome between blood- and flea-like conditions (Chromy et al., 2005). A total of ∼2300 protein spots were reproducibly observed following fractioning of whole-cell extracts prepared from the wild type and Δpnp strains, of which 69 were present at different levels between the two strains. To validate the approach, two excised spots were analyzed that were present in the wild-type strain but not the Δpnp strain that were predicted to be PNPase (based on its calculated pI and MW) as this was the only protein that was certain to differ between the two strains. These spots were definitely identified as PNPase by matrix-assisted laser desorption/ionization (MALDI) although from the present analysis, the basis for their differential migration could not be determined (Table 2). Then, 14 spots were examined that were at elevated levels in the Δpnp strain compared with the wild-type strain. Of the 13 spots that were positively identified, six spots were constituents of the T3SS regulon: YpkA, YopD, LcrV, and three separate spots of YscC (Table 2). The identities of the other proteins displaying elevated levels in the Δpnp strain are shown in Table 2 (see below for further comments). Of this latter group, phage shock protein A (PspA) is of special interest as its levels have been shown to be positively associated with levels of outer membrane porin YscC (Darwin & Miller, 2001). Collectively, these data show that the steady-state levels of T3SS transcripts and proteins are higher in the Δpnp strain compared with the wild-type strain, suggesting that PNPase negatively regulates the expression of the T3SS regulon.
|Protein/gene||Accession number||Fold- change†||t-test‡|
|PNP/pnp||NP_668031||−3.33||1.60 × 10−6|
|PNP/pnp||NP_668031||−2.63||1.40 × 10−5|
|YpkA/ypkA||NP_857776||+1.87||1.10 × 10−7|
|YopD/yopD||NP_857754||+1.53||8.00 × 10−8|
|LcrV/lcrV||NP_857751||+1.86||7.80 × 10−10|
|YscC/yscC||NP_857925||+2.30||2.20 × 10−10|
|YscC/yscC||NP_857925||+2.29||7.40 × 10−10|
|YscC/yscC||NP_857925||+1.90||2.10 × 10−5|
|PEP-protein phosphotransferase system enzyme I/ptsI||NP_668807||+1.54||3.60 × 10−5|
|Putative aminotransferase||NP_668827||+1.55||6.80 × 10−7|
|Thiosulfate-binding protein/cysP||NP_668786||+1.85||1.20 × 10−10|
|Plasminogen activator/pla||NP_857784||+1.85||1.20 × 10−10|
|Aspartate carbamoyltransferase catalytic subunit/pyrB||NP_667504||+1.85||1.20 × 10−10|
|NADH-quinone oxireductase chain B/nuoB||NP_668949||+2.04||7.40 × 10−9|
|Phage shock protein A/pspA||NP_669295||+2.04||7.40 × 10−9|
S1 effects on T3SS activity and expression levels
As S1 domains themselves have well-characterized subdomains (see ‘Introduction’), the aim was to determine what attributes of S1 were required for its T3SS-promoting effect and whether S1-mediated restoration of T3SS activity in the pnp strain was accompanied by altered T3SS expression levels. Therefore, the S1-encoding sequences from RNase E (corresponding to RNase E residues 35–125) were cloned into an arabinose-inducible vector that generated a His-patch thioredoxin S1 hybrid protein that hereafter will be designated as ‘t-S135–125’. The S1 domain of RNase E was chosen because, as discussed in more detail below, it is well defined structurally (Callaghan et al., 2005). Levels of the t-S135–125 hybrid protein, as well as amino-terminal-deleted variants, in the YP/Δpnp mutant strain increased after a 1.5-h induction period as determined by immunoblotting (Fig. 1a).
One consequence of Yop cellular activity observed in infected cultured cells is a severe ‘cytotoxicity’ characterized by a collapse of the actin-based cytoskeleton (Rosqvist et al., 1990). The level of cytotoxicity is markedly reduced (but not abolished) in cells infected with the YP/Δpnp strain compared with the isogenic wild-type strain (Rosenzweig et al., 2005; Fig. 1b). A moderate increase in the level of cytotoxicity was observed in cells infected with the ‘uninduced’ t-S135–125-expressing YP/Δpnp strain relative to the empty vector control strain, likely indicating ‘leakiness’ of the arabinose-inducible promoter (Fig. 1b). If the t-S135–125-expressing YP/Δpnp strain had been cultured in arabinose-containing media before infection, the levels of cytotoxicity became comparable to the level of cytotoxicity observed in cells infected with the isogenic wild-type strain (Fig. 1b). These data show that the thioredoxin epitope tag does not interfere with S1-mediated cytotoxicity.
There have been a number of recent structural and functional reports concerning the S1 domain of RNase E. Two different amino-terminal truncated His-patch thioredoxin S1 hybrid variants were constructed: t-S150–125 is deleted for residues along the S1 dimerization interface and t-S165–125 is additionally deleted for residues that form the oligonucleotide-binding surface (Schubert et al., 2004; Callaghan et al., 2005). These t-S1 variants were stably expressed in the YP/Δpnp strain (Fig. 1a) and tested in the infection assay. There was no discernable difference in the levels of cytotoxicity in cells infected with the YP/Δpnp strain expressing the t-S150–125 variant compared with cells infected with the YP/Δpnp strain expressing the t-S135–125 variant (not shown). However, there was an obvious reduction in the level of cytotoxicity in cells infected with the t-S165–125-expressing YP/Δpnp strain compared with cells infected with the t-S150–125-expressing YP/Δpnp strain (Fig. 2). In fact, the level of cytotoxicity in cells infected with the t-S165–125-expressing Δpnp strain did not noticeably differ from that of cells infected with the empty vector control strain (see Fig. 1b). These data indicate that S1 residues 50–65 are critical for S1-mediated enhancement of cytotoxicity.
To test more directly the effects of S1 on T3SS activity, Yop secretion was measured in the Δpnp strain expressing either t-S135–125, t-S150–125, or t-S165–125. YPT is more suitable for secretion assays as YP possesses the unique extracellular protease Pla that rapidly degrades secreted Yops (Sodeinde et al., 1988). Similar to YP, all three variants were stably expressed in the YPT/Δpnp strain (not shown), and, again, there was a clear reduction in the level of cytotoxicity observed in cells infected with the empty vector control and t-S165–125-expressing YPT/Δpnp strains compared with cells infected with either the t-S135–125- or t-S150–125-expressing YPT/Δpnp strains (Fig. 3a). To measure secretion, the same strains were propagated at 37°C in calcium-containing media, and Yop levels were determined in the pellet and supernatant fractions following the removal of calcium from the medium (removing free Ca2+ fully induces the T3SS). There were no appreciable differences in the steady-state YopD and YopE protein levels in the pellet fractions between these four strains (Fig. 3b, lower blot; data for YopE not shown). Additionally, there were no discernable differences in YopD and YopE protein levels in the pellet fraction within each strain following a 5-min induction period. In contrast, in the supernatant fractions there was a several-fold difference in YopD levels between the empty vector control and t-S135–125-expressing YPT/Δpnp strains (Fig. 3b, upper blot, lanes 1–4). The YopD levels in the supernatant fraction of the t-S150–125-expressing YPT/Δpnp strain were somewhat reduced compared with the YopD levels in the supernatant fraction of the t-S135–125-expressing (lanes 3–6). However, akin to the empty vector control strain, YopD levels were below the level of detection in the supernatant fraction of the t-S165–125-expressing YPT/Δpnp strain (lanes 7–8). These data show that the amino-terminal S1 region, and especially residues 50–65, play an important role in controlling Yop secretion. The markedly reduced level of Yop secretion in the t-S165–125-expressing YPT/Δpnp strain almost certainly accounts for its reduced level of cytotoxicity observed in the infection assays shown in Figs 2 and 3a.
As there was such a clear difference in activities between S150–125 and S165–125 in terms of cytotoxicity and secretion, it was investigated whether there was also a difference in T3SS expression levels between strains expressing these different S1 variants. Therefore, the proteomes of YP/Δpnp strains expressing either t-S135–125- or t-S165–125 were compared using 2-D DIGE. The proteomes of these two strains had no detectable differences in T3SS protein levels when analyzed by 2-D DIGE using conditions similar to those described above (not shown). In fact, the proteomes of the t-S150–125- and t-S165–125-expressing YP/Δpnp strains were, for practical purposes, identical at least in the pI and MW ranges that were examined. This latter finding indicates that S1 affects T3SS activity via a mechanism that does not involve altering T3SS expression levels. Furthermore, this analysis also indicates that the pleiotropic effects associated with deletion of the PNPase-encoding locus (e.g. see Table 2) are likely not responsible for reduced T3SS activity.
PNPF638G and T3SS activity
The F57 residue of the S1 domain of RNase E is conserved in almost all S1 domains and plays an integral role in RNA binding (Schubert et al., 2004; Callaghan et al., 2005). Replacement of the homologous residue in the S1 domain of PNPase, F638, with glycine results in reduced catalytic activity (Jarrige et al., 2002). The PNPF638G variant in YP/Δpnp was expressed and the resulting strain for T3SS activity was tested in an infection assay. When under the control of the native pnp promoter, the steady-state levels of PNP and PNPF638G were comparable in YP/Δpnp (not shown), indicating that, similar to what had been shown in E. coli, the F638G replacement does not affect PNPase autoregulatory activity (Jarrige et al., 2002). In cells infected with PNP- and PNPF638G-expressing YP/Δpnp strains, there were distinct differences in the levels of cytotoxicity (Fig. 4). There was a clear reduction in the level of cytotoxicity in cells infected with the PNPF638G-expressing YP/Δpnp compared with that observed in cells infected with PNP-expressing YP/Δpnp. In fact, the level of cytotoxicity observed in cells infected with PNPF638G-expressing YP/Δpnp was indistinguishable from that observed in cells infected with the empty vector control YP/Δpnp strain (Fig. 4). Similar results were obtained when PNP and PNPF638G were expressed from the arabinose-inducible promoter used in S1-expression studies (not shown). The PNPF638G-expressing YP/Δpnp strain was also defective in growing at low temperature compared with the YP/Δpnp strain expressing wild-type PNP (Fig. 4c). In conclusion, it is believed that the relative reduction in T3SS activity in the t-S165–125- and PNPF638G-expressing strains, compared with their respective full-length or wild-type counterparts, indicates that S1 and PNP affect the T3SS through a common mechanism.
A link between PNPase and T3SS activity in Salmonella enterica has also been recently established by Rhen and colleagues. Interestingly, there appear to be both similarities and differences between the Yersinia and salmonellae with respect to how their respective PNPases relate to their T3SSs. The Δpnp strains of both species have reduced virulence in the mouse (Table 1; Clements et al., 2002). (The situation for the salmonellae, however, is complicated by the fact that despite having a reduced ability to cause acute disease in the mouse, the Δpnp strain has an enhanced ability to cause persistent infections.) Also, T3SS-encoding transcripts and proteins are at relatively higher levels in the Δpnp Yersinia and salmonellae strains compared with their respective wild-type counterparts. At this point, however, the similarities end. In the salmonellae Δpnp strain, increased T3SS expression levels are matched with a corresponding increase in T3SS functional activity as measured in a cell culture infection assay (using bacterial invasion as a readout). In contrast, it is shown that despite increased T3SS expression levels, there is a clear reduction in T3SS functional activity (using cytotoxicity as a readout) in the Δpnp Yersinia strains compared with the wild-type strains. This reduced T3SS functional activity is certainly due to the fact that there is a reduction in the Yop secretion rate in the Δpnp strain compared with the wild-type strain. As suggested recently (Ygberg et al., 2006), the accumulative data indicate that, depending on the species and/or stress (e.g. macrophages, low temperature, etc.), PNPase can play a variety of roles in enhancing bacterial survival.
The seemingly inverse relationship between T3SS expression and secretion levels in the Δpnp Yersinia strains greatly contrasts with that of the ΔymoA strain in which enhanced T3SS expression is matched by enhanced T3SS secretion and cytotoxicity in an infection assay (Jackson et al., 2004). The contention that the PNPase effects on T3SS activity and expression are distinct is further supported by the findings that S1 greatly affects T3SS activity without altering T3SS expression levels. Both functional- and structural-based studies have shown that residues 35–50 of the S1 domain of RNase E are involved in protein–protein interactions and are located far from the oligonucleotide-binding cleft (Diwa et al., 2002; Schubert et al., 2004); an S1 variant lacking these residues remained fully active in the secretion and cytotoxicity assays. In contrast, deleting residues 50–65 resulted in a substantial reduction in T3SS activity. This region of the S1 domain of RNase E forms the conserved oligonucleotide-binding cleft (Bycroft et al., 1997; Schubert et al., 2004; Callaghan et al., 2005). These analyses suggest that oligonucleotide-binding properties of S1 play a key role in regulating T3SS activity. A PNPase variant containing an F638G substitution within its S1 domain was also tested(F638 of PNPase is the structural equivalent of F57 of RNase E). PNPF638G possesses reduced catalytic activity, but retains wild type-like levels of autoregulation (Jarrige et al., 2002). PNPF638G did not restore normal T3SS activity to the Δpnp strain when assayed in the infection assay. The findings presented here and previously (Rosenzweig et al., 2005), together with recently published structural information (see ‘Introduction’), could indicate that the S1 domain of PNPase binds an mRNA or sRNA that in some way negatively regulates T3SS activity. In summary, it is believed that PNPase modulates the Yersinia T3SS by affecting the steady-state levels of T3SS-encoding transcripts and, through a separate mechanism that requires an intact RNA-binding surface on its S1 domain, by controlling the secretion rate.
The authors thank David Wiley, Murray Deutscher, George Mackie, and Hans Wolf-Watz for advice and reagents. This work was supported in part by grants from the Miller School of Medicine, University of Miami, and from Public Health Service grants AI53459 (KS) and AI50552 (GVP) from the National Institute of Allergy and Infectious Diseases. Additional reagents and technical support were provided by the Pathogen Functional Genomics Resource Center (NIAID/NIH) in consortium with The Institute of Genomic Research (Rockville, MD).
- 2002) YopD and LcrH regulate expression of Yersinia enterocolitica YopQ by a posttranscriptional mechanism and bind to yopQ RNA. J Bacteriol 184: 1287–1295. , , & (
- 2001) The Yersinia YopE and YopH type III effector proteins enhance bacterial proliferation following contact with eukaryotic cells. BMC Microbiol 1: e22–32. , , & (
- 1997) The solution structure of the S1 RNA binding domain: a member of an ancient nucleic acid-binding fold. Cell 88: 235–242. , , , & (
- 2005) Structure of Escherichia coli RNase E catalytic domain and implications for RNA turnover. Nature 437: 1187–1191. , , , , & (
- 2005) Proteomic characterization of Yersinia pestis virulence. J Bacteriol 187: 8172–8180. , , , , , , & (
- 2002) Polynucleotide phosphorylase is a global regulator of virulence and persistency in Salmonella enterica. Proc Natl Acad Sci USA 99: 8784–8789. , , , , , & (
- 2001) The psp locus of Yersinia enterocolitica is required for virulence and for growth in vitro when the Ysc type III secretion system is produced. Mol Microbiol 39: 429–444. & (
- 2006) Degradation of RNA in bacteria: comparison of mRNA and stable RNA. Nucleic Acids Res 34: 659–666. (
- 2002) Two distinct regions on the surface of an RNA-binding domain are crucial for RNase E function. Mol Microbiol 46: 959–969. , , & (
- 2001) The type III secretion chaperone LcrH co-operates with YopD to establish a negative, regulatory loop for control of Yop synthesis in Yersinia pseudotuberculosis. Mol Microbiol 42: 1075–1093. , & (
- 2007) The TCL1 oncoprotein binds the RNase PH domains of the PNPase exoribonuclease without affecting its RNA degrading activity. Cancer Lett 248: 198–210. , , et al. (
- 2006) Protein delivery into eukaryotic cells by type III secretion machines. Nature 444: 567–573. & (
- 1993) Temperature sensing in Yersinia pestis: translation of the LcrF activator protein is thermally regulated. J Bacteriol 175: 7901–7909. & (
- 2004) The ATP-dependent ClpXP and Lon proteases regulate expression of the Yersinia pestis type III secretion system via regulated proteolysis of YmoA, a small histone-like protein. Mol Microbiol 54: 1364–1378. , & (
- 2002) Mutational analysis of polynucleotide phosphorylase from Escherichia coli. J Mol Biol 321: 397–409. , , , & (
- 1992) Role of the transcriptional activator, VirF, and temperature in the expression of the pYV plasmid genes of Yersinia enterocolitica. Mol Microbiol 6: 395–409. , & (
- 2002) Identification and cloning of human polynucleotide phosphorylase hPNPase old-35, in the context of terminaldifferentiation and cellular secenscence. Proc Natl Acad Sci USA 99: 16636–16641. , , , , , & (
- 1990) Yersinia pestis pH 6 antigen: genetic, biochemical, and virulence characterization of a protein involved in the pathogenesis of bubonic plague. Infect Immun 58: 2569–2577. , & (
- 2000) The alpha subunit of E. coli RNA polymerase activates RNA binding by NusA. Genes Dev 15: 2664–2675. , , , & (
- 1994) A new class of proteins regulating gene expression in enterobacteria. Mol Microbiol 11: 77–86. & (
- 1998) Yersinia-induced apoptosis in vivo aids in the establishment of a systemic infection of mice. J Exp Med 188: 2127–2137. , , & (
- 1996) Modulation of virulence factor expression by pathogen target cell contact. Science 273: 1231–1233. , , , , , & (
- 1938) Simple method of estimating fifty per cent endpoints. Am J Hyg 27: 493–497. & (
- 2005) Modulation of yersinia type three secretion system by the S1 domain of polynucleotide phosphorylase. J Biol Chem 280: 156–163. , , & (
- 1990) The cytotoxic protein YopE of Yersinia obstructs the primary host defence. Mol Microbiol 4: 657–667. , , , & (
- 2004) Structural characterization of the RNase E S1 domain and identification of its oligonucleotide-binding and dimerization interfaces. J Mol Biol 341: 37–54. , , , , , & (
- 1988) Plasminogen activator/coagulase gene of Yersinia pestis is responsible for degradation of plasmid-encoded outer membrane proteins. Infect Immun 56: 2749–2752. , , & (
- 2005) Function of the conserved S1 and KH domains in polynucleotide phosphorylase. J Bacteriol 187: 7214–7221. , , & (
- 2000) A duplicated fold is the structural basis for polynucleotide phosphorylase catalytic activity, processivity, and regulation. Structure 8: 1215–1226. , & (
- 2005) Yersinia outer proteins: role in modulation of host cell signaling responses and pathogenesis. Annu Rev Microbiol 59: 69–89. & (
- 1998) YopD of Yersinia pestis plays a role in negative regulation of the low-calcium response in addition to its role in translocation of Yops. J Bacteriol 180: 350–358. & (
- 2006) Polynucleotide phosphorylase negatively controls spv virulence gene expression in Salmonella enterica. Infect Immun 74: 1243–1254. , , , , , & (
- 1982) Effect of exogenous nucleotides on Ca2+ dependence and V antigen synthesis in Yersinia pestis. Infect Immun 38: 953–959. & (
- 2001) Salmonella entry into host cells: the work in concert of type III secreted effector proteins. Microbes Infect 3: 1293–1298. & (