The Oxygen- and Iron-Dependent Sigma Factor pvdS of Pseudomonas aevuginosa Is an Important Virulence Factor in Experimental Infective Endocarditis

Journal of Infectious Diseases, Mar 2000

In Pseudomonas aeruginosa, pvdS, a key oxygen (O2)-dependent locus, regulates expression of a number of virulence genes, including toxA (which encodes exotoxin A production). To define the in vivo role of differing O2 tensions on pseudomonal virulence, 2 knockout mutants, ΔpvdS and ΔtoxA, were compared with their parental strain, PA01, in rabbit aortic and tricuspid endocarditis models (representing aerobic vs. microaerobic conditions in vivo, respectively). In aortic endocarditis, ΔpvdS densities were significantly less than those of PA01 in vegetations, kidneys, and spleen (P < .01). In contrast, in tricuspid endocarditis, there were no significant differences between ΔpvdS and PA01 tissue densities in these same target tissues. The ΔtoxA mutant proliferated within target tissues to the same extent as the parental strain. Thus, pvdS (but not toxA) appears to be required for optimal virulence of P. aeruginosa, particularly in tissues preferentially exposed to high O2 tensions (e.g., aortic vegetations).

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://jid.oxfordjournals.org/content/181/3/1020.full.pdf

The Oxygen- and Iron-Dependent Sigma Factor pvdS of Pseudomonas aevuginosa Is an Important Virulence Factor in Experimental Infective Endocarditis

Yan-Qiong Xiong () 2 3 Michael L. Vasil 0 2 Zaiga Johnson 0 2 Urs A. Ochsner 0 2 Arnold S. Bayer 1 2 3 0 Department of Microbiology, University of Colorado Health Science Center , Denver 1 UCLA School of Medicine , Los Angeles, California 2 Received 23 August 1999; revised 10 November 1999; electronically pub- lished 20 March 2000. Presented in part: American Society for Microbiology 98th general meet- ing, Atlanta, May 1998 (abstract B-137). Animal experimentation guidelines of the US Department of Health and Human Services and those of the authors' institutions were followed in the conduct of the studies. Grant support: NIH (AI-15940 to M.L.V.; AI-39108 to A.S.B.). Division of Infectious Diseases, St. John's Cardiovascular Research Center , Bldg. RB-2 , LAC-Harbor UCLA Medical Center , 1000 W. Carson St., Tor- rance, CA 90509 3 Department of Medicine, Division of Infectious Diseases, Research and Education Institute, St. John's Cardiovascular Research Center, LAC-Harbor UCLA Medical Center In Pseudomonas aeruginosa, pvdS, a key oxygen (O2)-dependent locus, regulates expression of a number of virulence genes, including toxA (which encodes exotoxin A production). To define the in vivo role of differing O2 tensions on pseudomonal virulence, 2 knockout mutants, DpvdS and DtoxA, were compared with their parental strain, PA01, in rabbit aortic and tricuspid endocarditis models (representing aerobic vs. microaerobic conditions in vivo, respectively). In aortic endocarditis, DpvdS densities were significantly less than those of PA01 in vegetations, kidneys, and spleen (P ! .01). In contrast, in tricuspid endocarditis, there were no significant differences between DpvdS and PA01 tissue densities in these same target tissues. The DtoxA mutant proliferated within target tissues to the same extent as the parental strain. Thus, pvdS (but not toxA) appears to be required for optimal virulence of P. aeruginosa, particularly in tissues preferentially exposed to high O2 tensions (e.g., aortic vegetations). - Pseudomonas aeruginosa is an opportunistic pathogen that frequently causes severe systemic infections, particularly in patients with burns, infective endocarditis, organ transplants, cystic fibrosis, and cancer chemotherapy [15]. Invasive infections caused by P. aeruginosa are often refractory to treatment by many common antibiotics, both because of intrinsic resistance manifested by the organism and because of the emergence of antimicrobial resistance during therapy [614]. Therefore, it is critical to better understand the virulence factors that affect the induction and progression of serious P. aeruginosa infections in order to define novel targets for therapeutic and prophylactic strategies against this pathogen. P. aeruginosa produces a large cadre of extracellular products that play a role in the virulence of the organism [9, 13, 1517]. Among them, exotoxin A has been implicated as a significant virulence factor in several different infection models [9, 17, 18]. The expression of exotoxin A is coordinately influenced by several specific O2- and iron-regulated genes; these regulatory genes respond to specific microenvironmental conditions in vitro (and possibly in vivo), such as O2 tensions, iron limitation, and temperature [13, 19, 20]. One such gene, pvdS, encodes an alternative sigma factor for RNA polymerase and is optimally expressed under iron-limiting conditions in an aerobic (20% O2) environment [21] and is 110-fold repressed in an iron-limited, microaerobic (8%10% O2) environment. The pvdS locus also regulates the expression of toxA, the structural gene encoding exotoxin A [21]. However, the overall in vivo effects of pvdS and toxA on the virulence of P. aeruginosa are largely undefined. Consequently, it is pivotal to precisely delineate the in vivo influence of selected O2- and iron-defined host environments on the pvdS and toxA loci and their subsequent effects on the pathogenesis of pseudomonal infections. Herein we describe a study in which we utilized experimental rabbit models of aortic (left-sided) and tricuspid (right-sided) endocarditis to define the in vivo effect of differing O2 tensions, in the presence of a constant iron milieu, on the virulence of parental versus isogenic mutants of P. aeruginosa. The latter mutant strains featured genetic inactivation of either the pvdS (DpvdS) or toxA (DtoxA) locus. These rabbit models were selected because they represent well-established and reproducible in vivo systems to examine and characterize microbial virulence factors [8, 22, 23]. Moreover, the prevailing O2 tensions in the right-side and left-side heart chambers differ by >50 mm Hg (40 vs. 90 mm Hg, respectively), allowing evaluation of the in vivo effects of O2-dependent loci in 2 distinct O2 microenvironments. Furthermore, in both the aortic and tricuspid endocarditis models, the infecting organism will undergo hematogenous dissemination from the infected valvular tissues to 2 visceral organs that receive a disproportionate supply of oxygenated blood (the kidney and spleen). Thus, one can contemporaneously evaluate the influence of O2-regulated pseudomonal genes on in vivo virulence in at least 3 target tissues [2, 22]. Materials and Methods Strains and construction of deletion mutants. We used the P. aeruginosa parental strain (PA01) and isogenic mutant strains (DpvdS and DtoxA) derived from this wild-type strain. The PA01 strain is a prototypical parental strain used in numerous previous investigations related to the genetics and virulence of P. aeruginosa [9, 13, 16, 17, 24, 25]. The DpvdS mutant has a deletion of 460 bp of the pvdS gene, with insertion of a gentamicin-resistance cartridge in place of the deleted pvdS locus in the PA01 chromosome [21]. Furthermore, the pvdS gene is required for optimal production of exotoxin A. Consequently, the DpvdS mutant produces 20-fold less toxin than the PA01 parental strain [21]. The DtoxA strain has not been described. In brief, it was constructed by deletion of a 1554-bp NruI-BamHI DNA fragment from the toxA gene, which extends from 6 bp 50 to the translational initiation codon of toxA to 1549 bp into the toxA gene, which is 1914 bp in size. A 1.8-kbp gentamicin-resistance cartridge was inserted into a BamHI site into an allelic exchange vector that carried the toxA deletion described above [21]. This allelic exchange vector was used in the replacement of the toxA gene in the PA01, as previously described [21]. The insertion of the toxA deletion mutation containing the gentamicinresistance cartridge, the replacement of the wild-type toxA gene, and the loss of the vector sequences were confirmed by analysis of genomic DNA of this mutant by Southern blot hybridization studies, as described elsewhere [21]. The DtoxA mutant does not produce any detectable cell-associated or extracellular exotoxin A, as confirmed by reaction with affinity-purified polyclonal antibody against exotoxin A in a Western blot assay (Z.J. and M.L.V., unpublished data). RNase protection and Western blot assays. We previously reported that the expression of pvdS in vitro is differentially affected by varying O2 tensions. Substantially higher levels of the pvdSspecific transcript are expressed in cells grown in artificial media under aerobic conditions than in cells grown under microaerobic conditions [21]. However, it is possible that in vivo factors in plasma, other than the differential levels of O2, could alter the expression pattern of pvdS. To address this issue, we grew PA01 parental cells in pooled rabbit plasma under aerobic (20% O2) or microaerobic (8%10% O2) conditions (approximating O2 tensions on the left and right sides of the heart, respectively). The relative expression of pvdS under these distinct growth conditions was evaluated by the RNase protection assay for specific pvdS mRNA and by Western blot analyses of the relative levels of the encoded PvdS protein. The mRNA extractions were done by standard techniques. Total RNA was isolated by the acid lysishot phenol method and treated with DNase I [19, 26, 27]. Bacterial mRNA samples were obtained from PA01 parental cells grown either in chelated and dialyzed tryptic soy broth (DTSB) or in pooled heparinized rabbit plasma under aerobic or microaerobic conditions. Cells grown in DTSB were grown only under aerobic conditions. RNA was extracted during late logarithmic growth in all media (i.e., 6 h). There were no differences in the growth kinetics of pseudomonal cells grown in synthetic media or rabbit plasma (data not shown). For RNase protection studies, 20 mg of RNA was used for each sample. The riboprobes required for RNase protection assays were synthesized with suitable oligonucleotide (23 mer) primers. The primers were used to amplify certain portions of the omlA [26] and the pvdS gene by polymerase chain reaction (PCR). The probe for the constitutively expressed omlA gene, which encodes an outer membrane lipoprotein (used as a control for these studies) has been described elsewhere [26]. The 364-bp probe for the pvdS gene covers a region that is completely within this gene. It begins 117 bp 50 to the initiation codon of pvdS (GenBank accession no. U12891). Both probes contain additional sequences from the T7 promoter region in the vector plasmid pCRII (Invitrogen, Carlsbad, CA) that are not homologous to either the omlA or the pvdS gene. The amplified DNA fragments were cloned into the vector (pCRII), and their sequences were verified. RNA probes were then generated from these fragments by run-off transcription from the T7 promoter, using the Riboprobe kit (Promega, Madison, WI). The RNase protection technique was utilized as described elsewhere [19, 26, 27]. Autoradiographs of the dried gels were scanned and imported into Photoshop (version 4.0; Adobe Systems, San Jose, CA) for adjustment of brightness and contrast. Quantitative analysis of the autoradiographs was performed with Image software (version 1.55; NIH, Bethesda, MD). The measured intensities of the bands are relative to each other. For Western blot assays, whole-cell extracts were obtained from P. aeruginosa PA01 parental cells grown either in DTSB or in pooled heparinized rabbit plasma (under different O2 conditions), as above. Samples were obtained during both logarithmic (6 h) and stationary (18 h) phases of the growth cycle. Prior to wholecell extractions, bacterial cells under each growth condition were adjusted to an A590 value of 0.175. The cells (6 mL) were centrifuged, and the entire cell pellets were dissolved in 200 mL of lysis buffer (100 mM Tris-HCl, pH 6.8; 200 mM dithiothreitol; 4% SDS; 0.2% bromophenol blue; and 20% glycerol). The cells in lysis buffer were then boiled for 5 min and spun in a microcentrifuge at 10,000 g for 10 min; 10 mL of the supernatant from these extracts was then run on an SDS-PAGE gel, blotted onto nitrocellulose, and probed with a monoclonal antibody against purified, histidinetagged PvdS (provided by I. Lamont and M. Wilson, University of Otago, Dunedin, New Zealand). The monoclonal antibodies against PvdS were produced and isolated at the Cancer Center Monoclonal Antibody Core University of Colorado Health Sciences Center, Denver. This monoclonal antibody does not react with cell extracts of the DpvdS mutant on Western blot analysis (I. Lamont, M. Wilson, M. L. Vasil, Z. Johnson, and U. Ochsner, unpublished data). As with autoradiographs in the RNase protection assay, we measured the intensities of the bands on Western blotting by using NIH Image software and quantitated the intensities relative to each other. Endocarditis model. Endocarditis on either the aortic or the tricuspid valve was induced as described elsewhere [68, 22]. In brief, rabbits were anesthetized by intramuscular injection of ketamine chloride (Aveco, Fort Dodge, IA) at 35 mg/kg of body weight and xylazine (Mobay, Shawnee, KA) at 5 mg/kg of body weight. For the aortic valve endocarditis model, a polyethylene catheter with an internal diameter of 0.86 mm (Becton Dickinson, Cockeysville, MD) was introduced retrograde via the carotid artery across the aortic valve into the left ventricle to produce sterile thrombotic vegetations on the aortic valve. For the tricuspid valve endocarditis model, the right jugular vein was cannulated, and the catheter was placed retrograde across the tricuspid valve into the right ventricle. The catheters remained in place for the duration of the experiments. Establishment of the ID95. Because in vivo microbial virulence is determined by factors enabling the organism to colonize and proliferate within host tissues, measurement of properties such as infectivity is an important part of in vivo virulence studies. The ID95, the number of bacteria necessary to infect 95% of the animals exposed to the bacterium, is widely used as an important virulence parameter in animal models. To define the ID95 inoculum of the PA01 parental and the DpvdS and DtoxA mutant strains in the endocarditis model, we performed preliminary studies in our laboratory in which each strain was individually injected intravenously into animals with indwelling transvalvular catheters. For these studies, an inoculum range of 106108 cfu per catheterized rabbit was utilized for the PA01 parental and DpvdS mutant strains. This inoculum range encompasses the ID95 for P. aeruginosa in most prior experimental endocarditis models. The ID95 defined for the DpvdS mutant was then utilized for the DtoxA mutant in subsequent studies. Simultaneous challenge (PA01 parental plus either Dpvd or DtoxA mutant) in aortic and tricuspid valve endocarditis. After defining the ID95 inoculum for each study strain, the relative capabilities of these strains to induce experimental endocarditis and cause metastatic abscesses in target tissues were evaluated in either aortic or tricuspid valve endocarditis models. Rabbits were simultaneously challenged intravenously with the ID95 inoculum, as defined above, for the PA01 parental strain (108 cfu/rabbit) plus the ID95 of either the DpvdS or DtoxA mutant strain (108 cfu/rabbit for each mutant construct) 24 h after catheterization. The simultaneous challenge strategy provides a direct competitive analysis of the capacities of the 2 isogenic strains to survive and proliferate in diverse target tissues over a defined time course [28]. Early colonization of valvular vegetation. To optimize the interpretation of the relative capacities of the PA01 parental and DpvdS mutant strains to proliferate within target tissues, we compared the capacity of these strains to initially colonize the sterile vegetative lesion. Rabbits underwent aortic catheterization as above and were then simultaneously challenged with 108 cfu of the PA01 parental strain and 108 cfu of the DpvdS mutant strain. Rabbits were killed 30 or 60 min after challenge, and vegetations were quantitatively cultured (see below). Postmortem analyses. For the ID95 and the simultaneous challenge studies, rabbits were euthanized 24 h after infection by rapid intravenous injection of 150 mg of sodium pentobarbital. Vegetations, kidneys, and spleen were then removed from the animals with proper transvalvular catheter placements and macroscopic evidence of vegetative endocarditis. All vegetations from individual rabbits were separately excised, pooled, weighed, and homogenized in 1.0 mL of sterile normal saline solution. Similarly, the spleens from individual animals were excised, weighed, and homogenized. For kidneys, overt abscesses were excised and similarly processed. If no macroscopic renal abscesses were observed, a 1-cm3 wedge biopsy sample was taken randomly from the corticomedullary area (the site of hematogenous renal seeding) and processed for quantitative culture. All 3 target tissue homogenates were then serially diluted and parallel plated onto both antibiotic-free trypticase soy agar (TSA) plates and TSA plates containing 75 mg/mL gentamicin. Since both the DpvdS and DtoxA mutant strains carry the gentamicin-resistance cartridge in place of the pvdS and the toxA genes, this plating technique allows selective quantitation of both the parental and mutant strains in target tissues that are dually infected. Plates were then incubated for 24 h at 377C, and the P. aeruginosa densities in the target tissue infected with the different pseudomonal strains were calculated. P. aeruginosa densities in the target tissues obtained 24 h after infection were expressed as median values and mean 5 SD log10 cfu/g of tissue, with a lower limit of detection of 20 cfu/mL of homogenate. For the purposes of statistical comparison, all culture-negative tissues were assigned a value of !1.33.7 log10 cfu/g of tissue, based on this lower detection limit and the actual tissue weight. Regardless of the calculated bacterial densities in a specific target tissue, any culture-positive tissues were considered to have been infected by the intravenous challenge inoculum. Statistical analyses. To compare the median values of P. aeruginosa densities in vegetations, kidneys, and spleen in the different infection groups, we used the Mann-Whitney U test. We used Fishers exact test to compare the proportions of culture-negative tissues in the different infection groups. P ! .05 was considered significant. For adequate statistical power, >8 animals were included in each group analyzed. Results Expression of pvdS in plasma under aerobic and microaerobic conditions. To assess the expression of the pvdS locus under physiologic conditions that mimic those that exist on the right side versus the left side of the heart during infection, the PA01 parental strain was grown in parallel in pooled, heparinized rabbit plasma under aerobic or microaerobic conditions. As a control, we assessed the expression of a gene (omlA) encoding an outer membrane lipoprotein that is constitutively expressed under all in vitro growth conditions thus far examined [26]. As shown in figure 1A, the level of the omlA-specific message was virtually constant whether the cells were grown under aerobic or microaerobic conditions in rabbit plasma or grown in synthetic media. In contrast, there was a substantial increase in the expression of pvdS-specific mRNA and in the PvdS protein in cells grown under aerobic conditions in rabbit plasma (and in synthetic medium), compared with cells grown in parallel under microaerobic conditions in rabbit plasma (figure 1). The differences in expression of the encoded PvdS protein were most notable during the logarithmic phase of growth (figure 1B). Establishing the ID95 of the study strains in experimental endocarditis. In these studies, animals with aortic catheters were individually challenged with either the PA01 parental or the DpvdS mutant strain over an inoculum range of 106108 cfu per rabbit (figure 2). For the PA01 parental strain, the ID95 for inducing infective endocarditis with concomitant metastatic abscesses in kidney and spleen was 107 cfu. At the same inoculum, the DpvdS mutant caused aortic endocarditis in all animals. However, at both 107 and 108 cfu inocula challenges, substantially fewer animals had metastatic infection in kidneys and spleen. For the DtoxA mutant strain, the ID95 for inducing infective endocarditis was 108 cfu per rabbit; at this inoculum, 90% of kidneys and spleen were infected (data not shown). On the basis of these data and to ensure high rates of endocarditis and metastatic target tissue infection, for all subsequent simultaneous challenge experiments in catheterized rabbits we used 108 cfu for all study strains. Relative virulence of the PA01 parental versus DpvdS mutant strain in experimental endocarditis. In aortic endocarditis, the densities of the DpvdS mutant were significantly lower in all 3 target tissues than those of the parental strain (table 1; P ! .01, !.005, and !.0005 for vegetation, kidney, and spleen, respectively). Furthermore, the frequency of inducing metastatic infection in kidney and spleen was substantially lower in animals challenged with the DpvdS mutant strain than that in those challenged with the parental PA01 strain (figure 3A). For example, the DpvdS mutant caused metastatic splenic infection in only 30% of animals, whereas the parental strain caused splenic infection in 100% of animals (P ! .05). In tricuspid endocarditis, the differences in bacterial densities in vegetation, kidney, and spleen between the PA01 parental and DpvdS mutant strains did not reach statistical significance, although there was a trend toward lower bacterial densities in animals infected with the DpvdS mutant (table 1). In addition, the frequencies with which the parental and mutant strains induced endocarditis or caused metastatic infection in kidneys and spleen did not differ statistically (figure 3B). Early colonization of valvular vegetations by the PA01 parental versus DpvdS mutant strain. We performed additional studies to ensure that the differences in the relative virulence of PA01 and DpvdS strains in aortic endocarditis were not related to differences in early colonization of vegetations. There were no significant differences between the parental and DpvdS mutant strains in their capacities to colonize sterile aortic vegetations early after intravenous challenge. The mean 5 SD log10 cfu/g of vegetation for the DpvdS mutant and the PA01 parental strain were 2.53 5 0.14 versus 2.79 5 0.19, respectively, 30 min after challenge and 2.59 5 0.23 versus 2.64 5 0.10, respectively, 60 min after challenge. Virulence of PA01 parental versus DtoxA mutant strains in experimental endocarditis. Since the pvdS locus regulates the expression of the toxA gene in P. aeruginosa, we investigated whether the lowered virulence noted above for the DpvdS mutant in aortic endocarditis was explicable on the basis of altered DtoxA expression in vivo. No significant differences were observed between the parental and DtoxA mutant strains in terms of achievable bacterial densities in vegetation, kidney, and spleen (table 2). Furthermore, there were no significant differences in the frequencies of endocarditis induction (i.e., the parental and DtoxA mutant each induced infective endocarditis in 100% of animals) or metastatic infection in kidneys and spleens (the parental strain induced metastatic infection in 87% of kidneys and 100% of spleens; the DtoxA mutant caused metastatic infection in 100% of kidneys and spleens). Discussion P. aeruginosa is a leading bacterial etiology in a broad spectrum of severe human infections [1, 3]. However, its precise pathogenetic mechanisms and specific virulence factors are incompletely understood. P. aeruginosa is endowed with a large ensemble of potential virulence determinants that appear to promote its survival and replication in diverse target tissues [13, 15, 17]. These include cell surfaceassociated factors such as alginate, pili, and lipopolysaccharide and extracellular factors including exotoxin A, exoenzyme S (ExoS) and T (ExoT), proteases, cytotoxin, phospholipases, heat-stable hemolysins, pyocyanin, and siderophores [9, 13, 29]. A type III secretion system identified as a novel virulence regulon in P. aeruginosa is responsible for transport of 2 of these proteins, ExoS and ExoT [30, 31]. Several in vivo studies have implicated exotoxin A as a major pathogenic factor in P. aeruginosa infections [17, 18, 25]. However, the production of exotoxin A is regulated in response to specific physiologic and microenviromental signals (e.g., temperature, pH, O2 tensions, microbial cell density, iron limitation) [13, 15, 19, 21]. This is accomplished by the transcriptional regulation of toxA by an increasingly recognized number of regulatory genes, which themselves are regulated by these same environmental factors [21]. One such gene that is central to the present study is the pvdS locus, which encodes an alternative sigma factor for RNA polymerase. The pvdS regulon belongs to a class of sigma factors known as extracytoplasmic function (ECF). ECF appears to control the expression of genes encoding proteins that are secreted from the cytoplasm [21]. Another study showed that ECF sigma Figure 3. Infection frequencies of vegetations, kidneys, and spleens after simultaneous challenge with 108 cfu/rabbit of the PA01 parental strain (open bars) and DpvdS mutant strain (hatched bars) in aortic (A) and tricuspid (B) endocarditis. *P ! .05. Median (mean 5 SD, log10 cfu/g tissue) 8.6 (8.1 5 0.9) 7.4 (7.1 5 1.1)a 7.4 (6.7 5 1.5) 6.9 (6.1 5 1.7) 3.4 (3.3 5 0.9) 1.7 (1.9 5 0.3)b 2.6 (2.9 5 1.1) 2.8 (2.4 5 0.8) 3.1 (3.1 5 0.5) 2.0 (1.9 5 0.3)c 3.4 (3.4 5 0.8) 3.2 (2.9 5 0.9) factors probably play a role in production of OprF, a major outer membrane protein in type I Pseudomonas species [32]. Furthermore, Ochsner et al. [21] recently reported that the expression of pvdS is repressed 110-fold under microaerobic conditions (8%10% O2) but derepressed under aerobic conditions (20% O2). In this context, in the present study, we demonstrated a substantial increase in the expression of pvdS-specific mRNA as well as in PvdS protein under aerobic conditions, compared with microaerobic conditions in rabbit plasma. Moreover, our data showed that the expression of PvdS is growth phasedependent both in synthetic media and in rabbit plasma (i.e., maximal expression at logarithmic vs. stationary phases of growth). Collectively, these data indicate that the expression of pvdS is specifically up-regulated by O2 levels and suggest that this expression is the most critical during the initial stages of an infection. Our in vivo studies were designed to examine the interplay of pvdS and the pvdS-regulated toxA gene in an experimental endocarditis model in which the organism would be exposed to differing O2 tensions in the host. The in vivo studies produced several findings of interest. First, the inactivation of the pvdS gene resulted in significantly decreased P. aeruginosa virulence in aortic (but not in tricuspid) endocarditis. These differences could not be explained by differences in the extent of initial colonization of the sterile vegetation by the 2 strains. Second, unlike the DpvdS mutant, infection with the DtoxA mutant yielded no significant differences in vegetation, kidney, and spleen densities in comparison with the parental strain. Thus, although the pvdS regulon controls expression of the toxA gene [21], our data suggest that the lowered virulence of the DpvdS mutant in aortic endocarditis may be mediated by a toxAindependent pathway. Alternatively, toxA may contribute to virulence, but its contribution may not be large enough to be detected by the methods used in this study. It is possible that another pvdS-regulated gene makes a significant contribution to virulence or that the concerted action of several pvdS-regulated genes is necessary to account for the differences between the pvdS mutant and the parental strain. Accordingly, we recently identified a pvdS-regulated gene encoding an extracellular endoprotease that could play a role in endocarditis alone or in concert with toxA or other, as yet unidentified pvdSregulated genes [33]. It was conceivable that the pvdS deletion caused by the gentamicin cassette insertion might have introduced a polar effect on downstream genes. However, extensive RNase protection analyses show that the pvdS gene is in a monocistronic operon and that it contains a strong transcriptional termination signal at its 30 end. Consequently, there are no cis polar effects of inserting the gentamicin-resistant cassette into the pvdS gene on genes 30 to pvdS [33]. In both human [2, 4] and experimental [68] P. aeruginosa endocarditis, clinical cure rates are significantly lower in aortic than in tricuspid infection (25% vs. 75%). Moreover, in experimental endocarditis models, final achievable pseudomonal densities in left-side vegetations are substantially higher than those in right-side vegetations [7, 8]. Data from this study support the latter observation on the positive effect of higher O2 tensions on in vivo pseudomonal proliferation in vegetations. In the parental strain, the mean achievable intravegetation microbial densities were significantly higher in aortic than in tricuspid vegetations (8.1 and 6.7 log10 cfu/g, respectively; P ! .01). At least 3 distinct mechanisms have been postulated to explain this disparity in intravegetation proliferation and responses to antimicrobial therapy in aortic versus tricuspid pseudomonal endocarditis. First, since P. aeruginosa growth flourishes aerobically, the higher O2 tensions extant would favor the growth of the organism [34, 35]. Second, the polymorphonuclear (PMNL) influx throughout the vegetation lesion, which might limit microbial proliferation, is quantitatively less in Pseudomonas-infected aortic lesions than in tricuspid lesions [2]. Third, alginate production is substantially increased at higher versus lower O2 tensions (80 vs. 40 mm Hg, respectively) in vitro and in aortic lesions in the endocarditis model [3436]. Pseudomonal organisms expressing higher alginate levels in vitro are less susceptible to PMNL-mediated phagocytic killing [34, 35] and to numerous antimicrobial effects, including killing and postantibiotic growth inhibition and postantibioticmediated leukocyte enhancement effects [35]. The level of alginate production by PA01 is too low to measure in vitro by standard assays. Furthermore, we found no in vitro differences in alginate production between a mucoid parental strain, FRD1, and its DpvdS mutant (authors unpublished data). Taken together, the above findings strongly suggest a survival benefit for this organism at higher O2 tensions. Moreover, the differential O2 tension between the left and right sides of the Table 2. Comparative virulence of the PA01 parental and DtoxA mutant strains in experimental aortic endocarditis. PA01 parent DtoxA mutant Median (mean 5 SD, log10 cfu/g tissue) 7.6 (7.6 5 0.6) 6.9 (6.9 5 1.0) 3.4 (3.1 5 1.1) 2.7 (2.7 5 1.1) 2.6 (2.5 5 0.3) 2.5 (2.5 5 0.4) heart is unlikely to be the sole contributing factor to the reduced virulence of the DpvdS strain in this model. In addition, P. aeruginosa possesses a number of genetic loci, similar to pvdS, that are also substantially affected by iron. Thus, additional studies will be required to evaluate each of these regulons individually and in combination to better delineate their hierarchy of effect upon pseudomonal virulence and also to distinguish the specific influences of differential O2 tensions from iron effects in this model. Nevertheless, our data provide new insights into the potential in vivo role of O2-regulated genes of P. aeruginosa.


This is a preview of a remote PDF: http://jid.oxfordjournals.org/content/181/3/1020.full.pdf

Yan-Qiong Xiong, Michael L. Vasil, Zaiga Johnson, Urs A. Ochsner, Arnold S. Bayer. The Oxygen- and Iron-Dependent Sigma Factor pvdS of Pseudomonas aevuginosa Is an Important Virulence Factor in Experimental Infective Endocarditis, Journal of Infectious Diseases, 2000, 1020-1026, DOI: 10.1086/315338