Correspondence: Jason A. Rosenzweig, Department of Biology, Center for Bionanotechnology and Environmental Research (CBER), Texas Southern University, 3100 Cleburne Street, Houston, TX, USA. Tel.: +1 713 313 1033; fax: +1 713 313 7932; e-mail: firstname.lastname@example.org
Yersinia polynucleotide phosphorylase (PNPase), a 3′–5′ exoribonuclease, has been shown to affect growth during several stress responses. In Escherichia coli, PNPase is one of the subunits of a multiprotein complex known as the degradosome, but also has degradosome-independent functions. The carboxy-terminus of E. coli ribonuclease E (RNase E) serves as the scaffold upon which PNPase, enolase (a glycolytic enzyme), and RhlB helicase all have been shown to bind. In the yersiniae, only PNPase has thus far been shown to physically interact with RNase E. We show by bacterial two-hybrid and co-immunoprecipitation assays that RhlB and enolase also interact with RNase E. Interestingly, although PNPase is required for normal growth at cold temperatures, assembly of the yersiniae degradosome was not required. However, degradosome assembly was required for growth in the presence of reactive oxygen species. These data suggest that while the Yersinia pseudotuberculosis PNPase plays a role in the oxidative stress response through a degradosome-dependent mechanism, PNPase's role during cold stress is degradosome-independent.
Like other closely related Gram-negative enteric pathogens, Yersinia pseudotuberculosis employs a type III secretion system (T3SS) to infect host cells, and polynucleotide phosphorylase (PNPase), a phosphorolytic 3′–5′ exoribonuclease involved in RNA decay, is required for its optimal functioning (Rosenzweig et al., 2005, 2007). Furthermore, we (and others) have observed that PNPase is required for the cold-shock response and/or acclimation for a number of organisms including Yersinia pestis and Y. pseudotuberculosis (Rosenzweig et al., 2005, 2007), Escherichia coli (Jones et al., 1987; Mathy et al., 2001; Yamanak & Inouye, 2001; Polissi et al., 2003), and Yersinia enterocolitica (Goverde et al., 1998; Neuhaus et al., 2000; Neuhaus et al., 2003). Intriguingly, PNPase has been shown to physically interact with an essential endoribonuclease, RNase E, in both Escherichia coli (Carpousis et al., 1994; Vanzo et al., 1998; Khemici & Carpousis, 2004) and Y. pseudotuberculosis (Yang et al., 2008) forming a large multiprotein RNA surveillance/quality control complex termed the degradosome. However, the role of the degradosome in various yersiniae stress responses has not been well studied.
RNase E, PNPase, RhlB RNA helicase and enolase have all been identified as components of the E. coli degradosome (Carpousis, 2002; Khemici & Carpousis, 2004; Lawal et al., 2010). PNPase, enolase, and RhlB each bind to the carboxy-terminal domain (CTD) of RNase E that serves as the scaffold. Interestingly, the CTD of RNase E is not well conserved and varies widely in various bacterial species (Erce et al., 2010). Typically, a degradosome consists of both an exo- and endoribonuclease (e.g. PNPase and RNase E), and they are thought to work together in concert producing a synergistic effect that optimizes RNA decay of unwanted transcripts. However, a degradosome consisting of both RNase R, a cold-inducible exoribonuclease in E. coli (Cairrão et al., 2003; Chen & Deutscher, 2010) required for the maturation of SsrA/tmRNA (Cairrão et al., 2003), and RNase E has also been identified in the psychrotrophic Pseudomonas syringae, possibly suggesting the existence of a specialized cold-adapted degradosome (Purusharth et al., 2005). What remains uncertain within the field of RNA biology is the exact contribution the degradosome plays in RNA decay/maturation relative to its individual components. In fact, an inability of E. coli to assemble a degradosome resulted in affecting only some RNA decay outcomes and had only a modest impact on growth kinetics (Kido et al., 1996; Jiang et al., 2000). Perhaps, the degradosome specifically degrades subsets of transcripts following periods of induced gene expression (e.g. stress response), while other stress-induced transcripts are degraded by degradosome-independent mechanisms. In support of this, an E. coli PNPase-deficient mutant was found to be more sensitive to oxidative stress in the form of H2O2, and this PNPase requirement for tolerating oxidative stress was independent of degradosome association (Wu et al., 2009). However, a dominant-negative, carboxy-truncated RNase E variant in E. coli (unable to form a degradosome) resulted in poor autoregulation of the rne transcript, suggesting that the degradosome might be required for the degradation of specific transcripts in E. coli (Briegel et al., 2006). When a similar carboxy-truncated RNase E variant was expressed in Y. pseudotuberculosis, increased sensitivity to host cell–induced stress (HCIS), prompted by macrophage challenge, ensued (Yang et al., 2008). In addition to degradosome constituents' physical interactions being demonstrated by co-immunoprecipitation (Co-IP) (Coburn et al., 1999; Yang et al., 2008), several bacterial 2 hybrid, B2H, (Karimova et al., 1998) assay studies have supported earlier Co-IP findings. More specifically, the B2H demonstrated an interaction between E. coli-derived PNPase and RhlB helicase (Liou et al., 2002). Additionally, the B2H assay demonstrated interactions between full-length PNPase, enolase, RhlB, and RNase E CTD as well as interactions between microdomains of RNase E's CTD and the aforementioned full-length binding partners derived from Vibrio angustum S14 (Erce et al., 2009, 2010). Therefore, we sought to characterize the Y. pseudotuberculosis degradosome further because only PNPase has been shown to physically interact with RNase E (Yang et al., 2008) and to determine whether it is required for various abiotic stress responses.
Materials and methods
Strains and plasmids
DHM1 is a cya-deleted E. coli, slow growing, temperature sensitive mutant strain that is used for the B2H screening. For the immunoprecipitations, Y. pseudotuberculosis ATCC® 6902™ was used.
YPT YPIII pIB102 (Bölin & Wolf-Watz, 1984) (WT) and YPIII pIB100Δpnp (Rosenzweig et al., 2005) were used for the cold growth and H2O2 plate-based assays.
The arabinose-inducible promoter containing pBAD24 (Guzman et al., 1995) plasmid was used as a cloning vector into which a carboxy-truncated RNase E (encoding only the first 465 amino acid residues in the amino terminus) was cloned (Yang et al., 2008). For all inductions, 0.02% arbinose was used unless otherwise noted. Ampicillin working concentrations were 100 μg mL−1.
Yersinia pseudotuberculosis primers for B2H cloning
Enolase: Forward: tcaggatcctatgtccaaaattgttaaag
Yersinia enterocolitica primers for B2H cloning
Cloning of proteins for B2H
RNase E CTD was cloned into the plasmid pKT25, while full-length enolase, PNPase, and RhlB were cloned into plasmid pUT18C. PCR products were generated using a 2X PCR master mix (New England Biolabs), and all cloning was performed using BamH1 and EcoR1 high-fidelity enzymes (New England Biolabs). All constructs were sequenced to confirm that they were correct.
DHM1 E. coli transformations and B2H assay
After DHM1 were harvested at OD600 nm of 0.5–0.6, pellets were washed twice in 1.0 mL of ice-cold water, washed once in ice-cold 10% polyethylene glycol (PEG), and resuspended in ~ 750 μL 10% PEG. To introduce plasmids, the cells were electroporated at 1700 V (BioRad Inc.). Following a 1-h recovery at 30 °C with agitation, transformations were plated on LB agar plates containing 40 μg mL−1 X-Gal (5-bromo-4-chloro-3-indolyl-β-D-galactopyranosid), IPTG (0.5 mM), 100 μg mL−1 of ampicillin, and 50 μg mL−1 of kanamycin. Plates were placed at 30 °C, and colonies were observed between 48–72 h later (Euromedex Inc.).
YPT was grown in 100 mL of LB medium to OD620 nm of 0.7. Cells were harvested by centrifugation at 5000 g for 15 min at 4 °C. Pellets were then resuspended in 5 mL 1X IP Buffer as part of a commercially available Protein G immunoprecipiation kit (Sigam Aldrich IP50). Complete EDTA-free protease inhibitor cocktail (Roche) (one tablet per 5 mL of solution) was added to the resuspended cells. Cells were lysed via sonication, and 600 μL of sonicate/lysate was used for downstream IP reaction (Sigma IP kit protcol).
Polyacrylamide gel electrophoresis (PAGE) was carried out using precast, 10 well, 4–12% gradient BIS-TRIS gels (Invitrogen Life Technologies), and proteins were transferred unto nitrocellulose membranes (Biorad) and blocked with nonfat milk. Polyclonal rabbit anti-PNPase, anti-RNase E, and/or anti-RhlB helicase antibodies were used to probe RNase E complexes or whole-cell extracts (at a dilution of 1 : 3000) for 1 h at room temperature.
A previously published protocol (Rosenzweig et al., 2005) was used with several modifications. In short, 10-fold serial dilutions of saturated bacterial cultures were spotted in duplicate in ~ 2 μL volumes (using a pronger) on 2 LB agar (Difco) plates containing 100 μg mL−1 ampicillin (Sigma) and 0.02% arabinose (Sigma). One plate was placed at 30 °C, while the other was placed at 4 °C and monitored for 11-day period. Alternatively, cultures were streaked out on the aforementioned plates and monitored for their growth over 11-day period.
H2O2 plate and liquid-based assays
Previously published protocols (Wu et al., 2009) were employed. In short, saturated cultures were diluted, and subcultures of OD600 nm ~ 0.2 were established in triplicate 100 μL volumes of LB medium (Difco) in 96-well plates. Following static growth at 30 °C for 1.0 h (with the appropriate antibiotic added and arabinose at 0.02%), a stock 0.88 M H2O2 was added to the various cultures yielding H2O2 concentrations of either 0, 20, 50, or 100 mM, respectively. Growth in the liquid cultures was monitored every 30 min over a 12-h period with continuous agitation. Growth curves were plotted, and the Student's t-test was used to determine statistical significance with P values < 0.05 considered significant.
For plate-based H2O2 assays, 10-fold serial dilutions of saturated bacterial cultures were spotted in duplicate (using a pronger) in ~ 2 μL volumes on 2 LB agar (Difco) plates containing 100 μg mL−1 Ampicillin (Sigma) and 0.02% Arabinose (Sigma). Plate H2O2 concentrations were 0, 0.4, 1, 2, 4, and 100 mM.
The RhlB RNA helicase and enolase are subunits of the Yersinia degradosome
In an attempt to further identify Y. pseudotuberculosis degradosome constituents, we employed the B2H assay (Karimova et al., 1998) to determine whether RhlB and enolase also associate with the RNase E CTD. In this B2H assay, interaction between two proteins results in transcription of the Lac operon and thus blue color on plates containing X-gal. Our data indicated that the RNase E CTD interacted very strongly with full-length RhlB helicase as evidenced by intensely blue colonies (Fig. 1c). In fact, the intensity of blue mirrored that of the positive control Zip–Zip (compare c to b). Blue colonies also appeared when PNPase interacted with RNase E CTD (d); however, the overall intensity of blue was less than that of an RhlB–RNase E CTD interaction (compare d to b). Little interaction occurred between enolase and the RNase E CTD, as evidenced by weekly blue colonies (e). All experimental colonies observed appeared bluer than the empty vector negative control, pKT25RNE-CTD vs. pUT18Cempty vector (compare all to a).
In addition to evaluating degradosome interaction of Y. pseudotuberculosis proteins, we also evaluated degradosome interaction of closely related Y. enterocolitica proteins and tested whether the Y. enterocolitica RNase E CTD interacted with both the Y. pseudotuberculosis and Y. enterocolitica RhlB degradosome-associated proteins. We chose looking at RhlB because it was the strongest binding partner for the Y. pseudotuberculosis RNase E CTD tested earlier (Fig. 1). Interestingly, the Y. enterocolitica RNase E CTD appeared to bind as well to the Y. enterocolitica RhlB protein as it did to the Y. pseudotuberculosis RhlB protein (Fig. 2). As was observed earlier with the Y. pseudotuberculosis RNase E CTD vs. Y. pseudotuberculosis enolase (Fig. 1), the Y. enterocolitica-derived RNase E CTD also interacted poorly with the Y. pseudotuberculosis derived enolase (Fig. 2). The positive control Zip–Zip appeared blue (as expected), while the two empty vector negative controls were white (as expected), pKT25RNE vs. pUT18Cempty and pKT25empty vs. pUT18CRhlB (Fig. 2).
To validate our B2H findings (Figs 1 and 2), co-immunoprecipitation (Co-IP) assays, utilizing polyclonal anti-RNase E antibodies fused to Protein G agarose beads, were employed. Immunoprecipitated complexes were resolved by SDS-PAGE and probed with polyclonal anti-RhlB or anti-PNPase antibodies. In agreement with our B2H results, RhlB clearly co-immunoprecipitated with RNase E (Fig. 3). PNPase also appeared to co-immunoprecipitate with RNase E (Fig. 3) as was demonstrated in earlier work (Yang et al., 2008). These B2H and co-IP experiments indicate that the RhlB and enolase are conserved subunits of the degradosome in Yersiniae.
Degradosome involvement in Y. pseudotuberculosis stress responses
The degradosome and PNPase have previously been implicated in various stress responses, including macrophage-induced stress, and cold stress (see 'Discussion'). To more completely understand their role during stress, we exposed a Δpnp mutant and a strain over-expressing an rne truncation to a variety of stresses. This rne truncation removed the CTD responsible for interaction with the other degradosome subunits, and its over-expression has previously been shown to interfere with degradosome assembly (Briegel et al., 2006; Yang et al., 2008).
As the ability of Y. pseudotuberculosis to respond to HCIS was previously shown to be dependent upon PNPase (Rosenzweig et al., 2005, 2007) as well as upon degradosome assembly (Yang et al., 2008), we were curious as to whether degradosome assembly was required for growth under oxidative stress which would be experienced during macrophage encounters. To test this directly, H2O2 liquid- and plate-based experiments were carried out. For plate-based assays, 0, 0.1, 0.2, 0.3, 0.4, 0.5, 1, 2, 4, 5, 10, 20, 50 and 100 mM H2O2 plate concentrations were all evaluated. The Δpnp mutant formed smaller colonies on plates, which was exacerbated by 0.1–0.4 mM H2O2 (Fig. 4). In a manner similar to how E. coli did not require degradosome assembly during oxidative stress (Wu et al., 2009), interfering with degradosome assembly did not affect growth on H2O2-containing plates (Fig. 4b).
To better quantify H2O2 sensitivity, a liquid-based H2O2 stress test was carried out in which subcultures were optically monitored for growth following the addition of either 0, 20, 50, or 100 mM H2O2 (Fig. 5). Remarkably, the more sensitive liquid-based assay revealed two significant effects. First, as indicated by the change in the slope of the graphs in Fig. 6, the Δpnp mutant had a longer doubling time in H2O2-containing media, but not in control media. In addition, interfering with degradosome assembly caused a reduced culture density as cultures entered stationary phase. Both of these differences were statistically significant. For reasons not well understood, interfering with degradosome assembly in the Δpnp mutant mirrored the phenotype of the Δpnp mutant strain and suppressed the early stationary phenotype when only degradosome assembly was disrupted (Fig. 5).
We also tested growth of these same strains at 4 °C (Fig. 6). Not surprisingly, and in agreement with previously published data (Rosenzweig et al., 2005, 2007), the Δpnp mutant was unable to grow at 4 °C (Fig. 6b) despite relatively normal growth at 28 °C (Fig. 6a). When RNE1-465 was expressed, there was no effect on the cold-sensitive phenotype (Fig. 6). These data strongly suggest that the psychrotropic yersiniae's ability to grow in the cold depends on PNPase in a degradosome-independent manner. To further evaluate the role that degradosome assembly might be playing in yersiniae stress responses, we challenged the strains with several antibiotics that target protein translation, membrane integrity, and cell wall integrity and found that neither the presence of PNPase nor the ability of the yersiniae degradosome to assemble altered antibiotic susceptibility profiles (data not shown).
As we observed that over-expression of RNE1-465 led to a significant reduction in biomass during oxidative stress, but that there was no similar reduction in biomass when expressed in the Δpnp background (Fig. 7), we hypothesized that perhaps PNPase affected expression of the plasmid-encoded RNE1-465. Following a 1.5-h induction of RNE1-465 in both strains and Western blot analysis, we concluded that the truncated RNE1-465 was expressed similarly in both strains and that PNPase was not modulating RNE1-465 expression levels. More specifically, the Y. pseudotuberculosis + empty vector pBAD24 (WT) and Y. pseudotuberculosis Δpnp + empty vector pBAD24 (pnp) controls did not express RNE1-465 when either induced with 0.02% arabinose or not (lanes 1, 2, 5, and 6). However, the Y. pseudotuberculosis + pBAD-RNE1-465 (RNE) and the Y. pseudotuberculosis Δpnp + pBAD-RNE1-465 (pnp/RNE) both expressed the ~ 52 KDa RNE1-465 when induced with 0.02% arabinose (lanes 3 and 7).
Yersinia pseudotuberculosis is a very close relative of the etiological agent of plague, Y. pestis, which diverged from Y. pseudotuberculosis between 15 000–20 000 years ago (Achtman et al., 1999). In fact, their RNase E, PNPase, RhlB and enolase proteins are 97–100% identical. Unlike Y. pestis (which has caused three major human pandemics), Y. pseudotuberculosis (like the more distantly related Y. enterocolitica) causes a relatively benign self-limiting gastrointestinal disease in humans (Galindo et al., 2011). Being psychrotropic and a human pathogen, a better understanding of Y. pseudotuberculosis stress responses could result in the discovery of novel targets for chemotherapeutic design. Both temperature (i.e. cold) and oxidative stress responses have been characterized in this manuscript, the former potentially experienced by Y. pseudotuberculosis or Y. enterocolitica during food processing and shipping and the latter experienced when attacked by host innate immune cells during an infection. Knowing that the exoribonuclease, PNPase, is required for cold growth of several organisms (Jones et al., 1987; Goverde et al., 1998) including Y. pseudotuberculosis (Rosenzweig et al., 2005), we strove to evaluate whether the PNPase requirement for cold growth of Y. pseudotuberculosis was degradosome-dependent. Similarly, we chose to characterize the Y. pseudotuberculosis oxidative stress response because PNPase had already been implicated in the E. coli H2O2 stress response in a degradosome-independent manner (Wu et al., 2009). In fact, PNPase has already been shown to promote yersiniae virulence and is required for optimal T3SS function (Rosenzweig et al., 2005, 2007), so identifying the exact constituents of the Y. pseudotuberculosis degradosome improves our understanding of how RNA metabolism impacts bacterial virulence as well.
Our data have identified RhlB, PNPase, and RNase E as components of the Y. pseudotuberculosis degradosome which previously had been shown to only include PNPase and RNase E (Yang et al., 2008). Furthermore, using the B2H assay, we demonstrated how the carboxy-terminus of a Y. enterocolitica-derived RNase E protein can also interact with Y. pseudotuberculosis RhlB helicase strongly supporting the notion that all pathogenic yersiniae can assemble a degradosome. We further characterized the role the Y. pseudotuberculosis degradosome plays in various stress responses and surprisingly found that the Y. pseudotuberculosis degradosome is not implicated in all stress responses that require PNPase involvement. More specifically, we determined that the Y. pseudotuberculosis cold-growth requirement for PNPase (Rosenzweig et al., 2005, 2007) is degradosome-independent. However, Y. pseudotuberculosis degradosome assembly was required for the oxidative stress response. Degradosome involvement with oxidative stress is in agreement with a previously published report of its requirement for macrophage-induced stress (Yang et al., 2008) and in contrast to its dispensability in the E. coli oxidative stress response (Wu et al., 2009). This is a shining example of how even closely related Gram-negative, enteric bacteria, for example, E. coli and Y. pseudotuberculosis, might behave differently to various stressors and employ different coping mechanisms to overcome the stress itself.
Unexpectedly, PNPase and the degradosome affect growth during H2O2 stress in different phases of growth. PNPase appeared important during log-phase growth of Y. pseudotuberculosis, while degradosome assembly affected biomass accumulation resulting in an early stationary phase. Even more unexpected was that the absence of PNPase suppressed the H2O2-sensitive phenotype of RNE1-465. Furthermore, the deletion of the PNPase-encoding gene did not diminish expression levels of RNE1-465, so the observation remains both intriguing and unexplained. In one scenario, PNPase responds to oxidative stress in Y. pseudotuberculosis independently during early growth; however, during later growth, PNPase associates with the degradosome to overcome the stress and enter into an acclimation phase. Of course, such a scenario fails to explain the surprising and unexplained phenomenon in which the absence of PNPase suppressed the H2O2-sensitive phenotype of RNE1-465. Perhaps a global evaluation of transcript abundance in each strain during oxidative stress is warranted and could reveal clues to help explain why PNPase and the degradosome affect growth during H2O2 stress differently despite PNPase not diminishing expression levels of RNE1-465.
Taken together, these data have expanded our understanding of the Y. pseudotuberculosis degradosome by clearly identifying RhlB helicase as a member of the multiprotein complex. Additionally, these data have delineated the role of the Y. pseudotuberculosis degradosome in various stress responses. Whereas PNPase seemingly affects growth at 4 °C in a degradosome-independent manner, the Y. pseudotuberculosis oxidative stress response clearly requires degradosome assembly to achieve optimal biomass during late log-phase growth. Realizing the unique contributions made by the degradosome during various stress responses could enable us to uncover novel chemotherapeutic targets more specifically aimed at disarming pathogens and making them more vulnerable/susceptible to those agents.
We gratefully acknowledge the generosity of W. Margolin for B2H strains and plasmids, K. Morano for use of a 96-well plate reader for the growth curves, K. Schesser for the YPT strains and pBAD-RNE1-465 and A.J. Carpousis for anti-RNase E, -PNPase, and -RhlB polyclonal antibodies used for IPs and immunoblotting. We would also like to thank M. Erce for her helpful suggestions and A.K. Chopra for stimulating discussion. We would also like to acknowledge our funding from NASA Cooperative Agreement NNXO8B4A47A (JAR) and NSF Research Opportunity Award MCB-1020739 001 (AVH).
A.H and J.S. contributed equally as first authors on this manuscript.