|Home | About | Journals | Submit | Contact Us | Français|
The bacterial SmpB-SsrA system is a highly conserved translational quality control mechanism that helps maintain the translational machinery at full capacity. Here we present evidence to demonstrate that the smpB-ssrA genes are required for pathogenesis of Yersinia pestis, the causative agent of plague. We found that disruption of the smpB-ssrA genes leads to reduction in secretion of the type III secretion-related proteins YopB, YopD, and LcrV, which are essential for virulence. Consistent with these observations, the smpB-ssrA mutant of Y. pestis was severely attenuated in a mouse model of infection via both the intranasal and intravenous routes. Most significantly, intranasal vaccination of mice with the smpB-ssrA mutant strain of Y. pestis induced a strong antibody response. The vaccinated animals were well protected against subsequent lethal intranasal challenges with virulent Y. pestis. Taken together, our results indicate that the smpB-ssrA mutant of Y. pestis possesses the desired qualities for a live attenuated cell-based vaccine against pneumonic plague.
Plague, a dangerous and often deadly disease, is caused by a Gram-negative bacterium, Yersinia pestis (38, 39). Depending on the route of entry, the disease can develop into a variety of forms, such as bubonic, pneumonic, or septicemic plague. Pneumonic plague is considered the most dangerous form of the disease since the organism can disseminate through aerosol droplets, resulting in high mortality. In fact, these features have led to the classification of Y. pestis as a category A agent of bioterrorism (24). Antibiotic therapy can be effective upon early diagnosis of plague. However, the appearance of multidrug-resistant Y. pestis strains in recent years presents a challenge for currently available antibiotic therapy (39). Therefore, there is a need for a safe and effective plague vaccine, which is currently not available.
Animal infection studies have identified several antigens that could be used as recombinant subunit vaccines. These include the F1 antigen and the LcrV protein. Active or passive immunization of experimental animals with these antigens was shown to be protective against pneumonic plague (1-3, 18, 23). However, F1− mutants of Y. pestis have been reported to retain full virulence in animal infection studies (15, 41, 52). Also, animals immunized with the LcrV protein can still be susceptible to Yersinia infections due to the variations in LcrV protein (44). Such strains could circumvent the effectiveness of subunit vaccines. Therefore, inclusion of additional elements, such as additional antigens or a library of antigens, could provide better protection against genetically engineered, fully virulent Y. pestis strains. One way to present many antigens at once is to utilize killed or live attenuated Y. pestis organisms. The use of heat-killed or formalin-fixed Y. pestis has a long history as a plague vaccine, and they were shown to be effective against bubonic plague (46). However, these vaccines have also caused significant adverse reactions, such as fever, malaise, headache, and lymphadenopathy. In addition, immunization with heat- or formalin-killed bacteria has generally failed to protect experimental animals against pneumonic plague (46). On the other hand, live attenuated plague vaccines, such as one based on the Y. pestis EV76 strain, appeared to be protective against pneumonic plague (46, 49, 53). Such genetically undefined strains can be unstable and retain significant virulence. Therefore, there is still a need to identify novel attenuated Y. pestis strains that can be used in production of safe and effective vaccines against all forms of plague.
SsrA is a unique RNA molecule that performs an important quality control function in collaboration with its protein partner, SmpB (17). SsrA RNA functions as both tRNA and mRNA through its unique sequence and structural properties. The SmpB-SsrA function is required to deal with ribosomes stalled on defective mRNAs (27, 28). The smpB and ssrA genes are present in all bacteria examined to date (21, 28, 51). The SmpB-SsrA system is important for maintaining cellular homeostasis and for survival of bacteria under adverse conditions. Unfortunately, there are only a few studies examining the contribution of this system to bacterial pathogenesis. Previous reports showed that the SmpB-SsrA system plays a critical role in Salmonella pathogenesis through controlling the expression of virulence factors and improving the ability of this organism to survive within macrophages (6, 26). More recently we showed that the smpB-ssrA mutant of Yersinia pseudotuberculosis was avirulent in a mouse infection model (34). Based on this evidence, we investigated the importance of smpB-ssrA in Y. pestis pathogenesis and the possibility of using its mutants as a live cell-based plague vaccine. Our results show that the smpB-ssrA mutant of Y. pestis is severely attenuated in a mouse model of infection. Most importantly, mice vaccinated with this mutant are protected against pulmonary Y. pestis infection.
Escherichia coli cells were grown at 37°C on Luria-Bertani (LB) agar or broth (Difco). Y. pseudotuberculosis and Y. pestis were routinely cultured at 26°C on heart infusion (HI) agar or broth (Difco) in the presence of antibiotics, kanamycin (25 μg/ml), chloramphenicol (30 μg/ml), and tetracycline (20 μg/ml), when appropriate.
The smpB and ssrA genes are located adjacent to each other in the genomes of Y. pseudotuberculosis and Y. pestis. Deletion of smpB and ssrA in Y. pseudotuberculosis IP2666 (Y. pseudotuberculosis ΔBA) was described elsewhere (34). A new construct was prepared to delete smpB-ssrA in a conditionally virulent Y. pestis CO92Δpgm strain (32). To avoid any polar effect, the deletion construct was designed to leave the first 150 nucleotides of smpB and the last 100 nucleotides of ssrA unchanged. Briefly, a set of primers, BA-del-fw (5′-CCGGATCCTTGCACTGCAAGGTTGGGAAG TAAAATCACTGCGTGCAGGCAAAGCAAATTGTAGGCTGGAGCTG CTTCG-3′) and BA-del-rev (5′-CCGGATCCGCACTACATGCTTAGTCCAATCATTACATTCGCCGGCCAGCTGCGGATGGATGGGAATTAGCCATGGTCC-3′), carrying a BamHI restriction endonuclease (RE) site, was used to amplify the kanamycin resistance cassette from the pKD4 plasmid (14) by PCR. Following DNA restriction digestion with BamHI, the DNA fragment was cloned into the suicide vector pSB890 (35) and transformed into E. coli S17 λpir cells. This strain was used for conjugal transfer of the suicide plasmid into Y. pestis. Tetracycline-resistant transconjugants were selected on Yersinia-specific agar (Oxoid, Basingstoke, United Kingdom). Recombinant colonies were selected on LB agar plates supplemented with 5% sucrose (34). Y. pestis smpB-ssrA (ΔBA) mutant cells were identified by using colony PCR analysis and confirmed by Northern blot analysis. ΔBA mutant cells were also transformed with the previously described psmpBssrA plasmid (34) to complement the mutant phenotypes. This plasmid is a low-copy-number pBR322-based plasmid that carries the smpB and ssrA genes under their native promoters.
The caf1 gene is 513 nucleotides and encodes the F1 capsular antigen subunit. In order to create the Δcaf1 mutation in the CO92Δpgm strain, a 2-kb DNA fragment including 1,200 nucleotides upstream of caf1, the caf1 gene itself, and 388 nucleotides downstream of caf1 was PCR amplified and cloned into the SmaI RE site of pSB890 plasmid. A DNA fragment encoding a chloramphenicol resistance gene was DNA restriction digested from the previously cloned pUC18-based pCAM plasmid by using the XbaI and SmaI enzymes and was inserted into the XbaI and HpaI RE sites that are present within the caf1 gene. This final construct was transformed into E. coli S17 λpir cells and utilized for insertional inactivation of the caf1 gene as described above. All mutations were confirmed by DNA sequencing.
Seven-week-old female C57BL/6 mice (Charles River Laboratories) were used in all infections. IACUC institutional guidelines were strictly followed for all animal infections. Mice infected with the Y. pseudotuberculosis IP2666 ΔBA strain were subjected to fasting for 20 h prior to orogastric infection. Cultures of Y. pseudotuberculosis cells were grown overnight in LB broth at 26°C, diluted to an optical density at 600 nm (OD600) of 0.1 in LB broth, and further grown to logarithmic phase at 26°C. Bacterial cells were harvested, washed, and suspended in phosphate-buffered saline (PBS). Mice were infected via the orogastric route with 0.2 ml of bacteria in PBS.
Infection of mice with Y. pestis was performed either intravenously or intranasally. Bacterial cells were grown for 16 to 18 h in HI broth at 26°C. Saturated cultures were diluted to an OD600 of 0.02 in HI broth and further grown for 13 to 15 h at 26°C. Bacterial cells were harvested, washed once, and suspended in PBS. Y. pestis (0.1 ml) was injected into the lateral tail vein. For intranasal infection, mice were first anesthetized with a mixture of ketamine HCl (Fort Doge Animal Health Laboratories) (100 mg/ml) and xylazine HCl (Ben Venue Laboratories) (20 mg/ml) mixed 2:1, which was further diluted 3:20 in PBS before a dose of 6 ml/kg of body weight was delivered by intraperitoneal injection. Lightly anesthetized mice were inoculated by the intranasal route with 25 μl of Y. pestis in PBS (5 μl at a time for each nare). To enumerate the actual number of bacteria that were given to each mouse, serial dilutions of inocula were spread on HI agar plates and grown at 28°C for 2 days.
To determine the 50% lethal dose (LD50), mice were infected intranasally or intravenously with 10-fold serial dilutions of suspensions of Y. pestis. All infected mice were monitored twice daily for 15 days. Mice exhibiting severe signs of disease (hunched posture and unresponsiveness to handling) were euthanized by CO2 asphyxiation. The LD50 was calculated by the method of Reed and Muench (42).
For tissue colonization experiments, mice were infected with Y. pestis at an intravenous dose of 103 CFU or an intranasal dose of 2 × 105 CFU. Infected mice were sacrificed at 1, 4, or 5 days postinfection. Spleens, lungs, livers, cervical lymph nodes, and blood were aseptically removed and mechanically homogenized in PBS using a Stomacher 80 blender (Seward Ltd.). Serial dilutions of homogenates were spread onto HI plates and incubated at 28°C for 2 days to enumerate the bacterial CFU.
Blood samples were collected from each mouse by tail bleeding 7 days before and 28 days after active immunization. Serum was isolated from blood samples using serum gel separator tubes (Sarstedt) according to the manufacturer's guidelines. Several immunization protocols were carried out. (i) Groups of mice were vaccinated with Y. pseudotuberculosis ΔBA via the orogastric route with either 5 × 106 or 2 × 107 CFU bacteria. The immunized mice were challenged intranasally with Y. pestis CO92Δpgm 48 days postinfection. (ii) Mice were vaccinated intravenously with 10 or 104 CFU of the CO92Δpgm ΔBA strain. Thirty-three days postinfection, they were challenged intranasally with Y. pestis CO92Δpgm. (iii) Separate groups of mice were immunized intranasally with 103, 104, 104 heat killed(heat killed at 56°C for 1 h), 106, and 107 CFU of the CO92Δpgm ΔBA strain. They were challenged intranasally with Y. pestis CO92Δpgm at 33 days postinfection. (iv) For passive immunization studies, immune sera of mice that had been infected with either 104 CFU of CO92Δpgm ΔBA or the PBS control via the intranasal route were collected 28 days postimmunization. The serum samples were mixed 1:1 with PBS and injected (200 μl) intraperitoneally into naïve mice 18 h after challenge infection with Y. pestis. Each set of immunization protocols also included a control group of mice that were given PBS only. All challenges were carried out at the indicated time points with either the CO92Δpgm strain (2 × 105 CFU = 10 LD50) intranasally or the KIM5 strain (103 CFU) intravenously.
Antibodies of the IgG class that are specific to Y. pestis were detected by an enzyme-linked immunosorbent assay (ELISA) and by Western blot analysis. Briefly, CO92Δpgm cells were spread on LB plates and grown at 37°C for 36 h. Bacteria were harvested from the plates, resuspended in cold PBS containing protease inhibitors (Roche Diagnostics), and sonicated. Cell sonicates (0.5 μg/well) in an equilibration buffer (0.3 N sodium carbonate buffer [pH 9.6]) were used to coat 96-well ELISA plates overnight at 4°C. Following the removal of the cell lysate, the wells were first blocked with PBS containing 1% casein and 0.05% Tween 20 for 1 h at 37°C and later incubated with 2-fold serial dilutions of immune sera for 1 h. Alkaline phosphatase-conjugated secondary goat anti-mouse IgG antibody (Southern Biotech) was added to each well and incubated for 1 h at 37°C. Following washes to remove unbound antibodies, the chromogenic substrate p-nitrophenyl phosphate (Sigma) was added to each well and incubated at 37°C for 30 min. Absorbance was measured at OD405. The antibody titer is the lowest serum dilution that has an OD405 reading of >0.1. For Western blot analysis, cell sonicates were mixed with Laemmli sample buffer and resolved on an SDS-12.5% polyacrylamide gel via electrophoresis. Proteins were transferred to nitrocellulose membranes, and immunoblotting was performed with immune sera (diluted 1:500) or rabbit anti-LcrV antibody (diluted 1:50,000) (25). Secondary IRDye 800CW anti-mouse or anti-rabbit antibodies (LI-COR) were used according to the manufacturer's recommendation. Membranes were scanned by an Odyssey infrared imaging system (LI-COR).
In order to determine the endogenous protein levels of Yops, Y. pestis cells were grown under Yop secretion-inducing conditions. Briefly, overnight cultures of bacteria were diluted to an OD600 of 0.1 in LB supplemented with 20 mM MgCl2 and 20 mM sodium oxalate. Cultures were incubated at 26°C for 1 h and then switched to 37°C and incubated for an additional 3 h. Optical densities of cultures were measured to harvest the same number of bacteria. Protein samples were resuspended with Laemmli sample buffer and resolved on an SDS-12.5% polyacrylamide gel via electrophoresis. Following the transfer of proteins to nitrocellulose membranes, immunoblotting was performed using the following antibodies: rabbit anti-LcrV (1:50,000), mouse anti-YopB (1:1,000), and mouse anti-YopD (1:1,000) (25). Secondary IRDye 800CW anti-mouse or anti-rabbit antibodies (LI-COR) were used according to the manufacturer's recommendation. Membranes were scanned by an Odyssey infrared imaging system (LI-COR), which was also used to quantify the protein bands.
Mouse survival studies were analyzed using the Prism software program (GraphPad), using the Mantel-Haetszel log rank test. The Mann-Whitney test was utilized to analyze bacterial colonization in organs.
To investigate the role of the smpB-ssrA genes in Y. pestis pathogenesis and explore the possibility of using a mutant strain as a live cell vaccine against plague, we constructed a nonpolar deletion mutation of the smpB-ssrA genes in Y. pestis CO92Δpgm (Table (Table1).1). For complementation, we also transformed the ΔBA mutant strain with the psmpBssrA plasmid (34), a pBR322-based plasmid that expresses the smpB and ssrA genes under their native promoters. Initial analysis revealed that the smpB-ssrA (ΔBA) mutant and its parental strain have similar growth rates when grown in HI broth at 28°C (Fig. (Fig.1A).1A). However, ΔBA cells had a slower growth rate in HI broth supplemented with 2.5 mM CaCl2 at 37°C. We have previously shown that the ΔBA mutant of Y. pseudotuberculosis has a defect in the type III secretion system (T3SS), most notably a decrease in expression of yopB, yopD, and lcrV. A similar phenotype exists in the Y. pestis ΔBA mutant. Western blot analysis of the whole cell lysates showed that the protein levels of YopB, YopD, and LcrV were lower in the ΔBA mutant cells than in the isogenic parental strain when the T3SS was induced in a low-calcium environment at 37°C (Fig. (Fig.1B1B).
To assess the impact of the smpB-ssrA mutation on Y. pestis virulence and pathogenesis, we evaluated the ability of the ΔBA mutant to cause a lethal mouse infection. To this end, C57BL/6 mice were infected intravenously with 103 CFU of Y. pestis CO92Δpgm or the ΔBA mutant and monitored for 15 days. Mice infected with CO92Δpgm appeared sick as early as 3 days postinfection and eventually succumbed to infection within 6 days (Fig. (Fig.1C).1C). In contrast, all mice infected with the ΔBA mutant survived and showed no signs of infection (hunchback posture and unresponsiveness to handling) throughout the 15-day observation period. When mice were infected with ΔBA mutant cells carrying the psmpBssrA plasmid, the virulence phenotype was restored to a significant level (P < 0.02). This indicated that the observed difference in virulence between ΔBA and its parental strain was due to the mutation in the smpB-ssrA genes and not a result of a secondary mutation.
Next, we sought to investigate the level of attenuation. LD50 analysis was conducted with CO92Δpgm and its ΔBA mutant by intravenous and intranasal routes. We found that the LD50 of CO92Δpgm is 5 CFU when it is administered intravenously (Fig. (Fig.2A),2A), which is similar to previously published reports (4). No mortality was observed among mice infected with up to 106 CFU ΔBA cells (Fig. (Fig.2B).2B). This represents a more than 105-fold difference in LD50 values between CO92Δpgm and its ΔBA mutant. The LD50 of CO92Δpgm was 2 × 104 CFU for the intranasal route (Fig. (Fig.3A).3A). Intranasal administration of ΔBA mutant cells up to 108 CFU caused no mortality (Fig. (Fig.3B).3B). It is also important to note that mice infected with the highest dose of the ΔBA strain via either route showed no signs of sickness and appeared healthy throughout the observation period. Taken together, these data strongly suggest that the smpB-ssrA genes are required for both intranasal and intravenous infection of mice by Y. pestis and that the ΔBA mutant is highly attenuated.
To gain further insight, we evaluated the ability of the mutant bacteria to disseminate and colonize target organs. To this end, we infected mice via the intravenous or intranasal route with Y. pestis ΔBA or the parental CO92Δpgm strain. Infected mice were sacrificed at 1, 4, and 5 days after infection to determine the bacterial load in spleen, lung, liver, cervical lymph nodes, and blood. One day after intravenous inoculation, CO92Δpgm cells were detected in the spleen, liver, and lung (Fig. (Fig.4A).4A). Bacterial growth continued to progress by day 4, and bacteremia could be detected at this point (Fig. (Fig.4B).4B). Every organ examined contained large amounts of bacteria, up to 108 CFU/g of organ. In sharp contrast, Y. pestis ΔBA mutant cells were less efficient in colonization via the intravenous route (Fig. 4A and B). Although they were detected in spleen and liver tissue, the growth of mutant bacteria was severely impaired. At 4 day postinfection, ~104-fold fewer mutant bacteria were recovered in the spleen and ~3 × 102-fold fewer mutant bacteria were recovered in the liver. Unlike CO92Δpgm, ΔBA mutant infection did not lead to bacteremia. There were no detectable levels of mutant bacteria recovered from the blood.
The effect of the smpB-ssrA mutation in bacterial colonization was more dramatic in pulmonary Y. pestis infection. When administered intranasally, CO92Δpgm and its ΔBA mutant could initially be detected in the lungs at comparable CFU 1 day postinfection (Fig. (Fig.4C).4C). By day 5, CO92Δpgm bacteria had successfully disseminated to other organs and were present at high concentrations (Fig. (Fig.4D).4D). In addition, CO92Δpgm could be detected in large quantities in the blood. In contrast, only 3 of 8 Y. pestis ΔBA mutant-infected mice carried bacteria in their spleens and livers. Additionally, the bacterial burden in organs of the ΔBA mutant-infected mice was ~104-fold lower than that in the parental CO92Δpgm strain (Fig. (Fig.4D).4D). Most of the infected mice cleared the mutant bacteria from their lungs, and it appeared that ΔBA cells were never able to disseminate and colonize the cervical lymph nodes and the blood.
Furthermore, to evaluate potential bacterial clearance, a group of mice infected intravenously with 103 CFU of Y. pestis ΔBA mutant cells was sacrificed 10 days after infection. We did not find any bacteria in the liver or the spleen, indicating that ΔBA mutant bacteria were effectively cleared from the infected host organs (not shown). It is also important to note that mice infected with the CO92Δpgm strain lost about 15% of their body weight at the time of sacrifice, a sign of a severe infection. In stark contrast, mice infected with Y. pestis ΔBA bacteria did not show any weight loss and appeared healthy. Taken together, these results indicate that the SmpB-SsrA system is important for bacterial dissemination from infected lungs and that ΔBA mutant cells are severely impaired in colonization of target organs.
Data presented here and in the previously published study (34) suggest that the smpB-ssrA mutants of Y. pestis and Y. pseudotuberculosis are highly attenuated, and infection of mice with such strains is not associated with telltale sings of severe infection and mortality. These phenotypes share desired characteristics for an effective vaccine strain. Another important feature required for an effective vaccine strain is the ability to mount long-lasting protective humoral and cellular immune responses against the target pathogen. Therefore, we sought to investigate the ability of each of the Y. pestis and Y. pseudotuberculosis ΔBA mutant strains to induce protective immunity against pulmonary Y. pestis infection.
Groups of C57BL/6 mice were immunized with Y. pseudotuberculosis ΔBA via the orogastric route, the natural route of infection for this enteric pathogen. Each group of mice received one of the following: 5 × 106 CFU ΔBA, 2 × 107 CFU ΔBA, or PBS alone. Following immunization, mice were challenged with a lethal dose (2 × 105 CFU, or 10 LD50) of the Y. pestis CO92Δpgm strain via intranasal inoculation and monitored closely. All control mice that were inoculated with PBS alone succumbed to infection within 9 days (Table (Table2).2). On the other hand, mice immunized with either 5 × 106 CFU or 2 × 107 CFU of Y. pseudotuberculosis ΔBA showed partial protection (50% and 55% survival, respectively) when challenged with a lethal dose of Y. pestis.
Having observed a partial protective response with the Y. pseudotuberculosis ΔBA mutant, we wished to determine if a similar or better protective response could be achieved by vaccination of mice with the Y. pestis ΔBA mutant strain. To this end, groups of C57BL/6 mice were immunized with Y. pestis ΔBA cells via the intravenous or intranasal route at the following doses: 10, 103, 104, 106, or 107 CFU live bacteria or 104 CFU heat-killed bacteria. Similar to previous lethal challenge assays, immunized mice were challenged with the Y. pestis CO92Δpgm strain via the intranasal route at a dose of 2 × 105 CFU at 33 days postvaccination. As expected, all control mice that received PBS alone succumbed to the lethal challenge (Table (Table2).2). Similarly, none of the mice that were immunized with heat-killed bacteria survived the lethal challenge. Most strikingly, mice immunized with the Y. pestis ΔBA mutant at a vaccination dose of 104 CFU or higher via either route were completely protected (Table (Table2).2). This experiment clearly demonstrated the effectiveness of intranasal immunization against intranasal Y. pestis challenge and suggested that the observed protective immunity depended on immunization with live bacteria. Additionally, in order to determine whether the intranasal immunization is also protective against lethal intravenous challenge, a group of 5 mice were immunized intranasally with 104 CFU of the Y. pestis ΔBA strain and later challenged via the intravenous route with Y. pestis CO92Δpgm at a dose of 103 CFU. All of the immunized mice were protected (not shown). Thus, we conclude that intranasal delivery of the Y. pestis ΔBA mutant at a dose of 104 CFU is sufficient to provide full protection against pulmonary Y. pestis infection in a mouse infection model.
To gain further insights into the mechanism of protection conferred by the Y. pestis ΔBA strain, we investigated the immune sera for the presence of protective antibodies against LcrV and the F1 antigen. To this end, we first evaluated mouse immune sera that were collected at 28 days postimmunization for total Y. pestis-specific IgGs by ELISA. The results indicate that intravenous administration of bacteria is usually associated with a strong antibody response. Immunization with 103 CFU of the Y. pestis ΔBA mutant via the intravenous route yielded titers ranging from 1:1,600 to 1:102,400 (Fig. (Fig.5A).5A). On the other hand, higher CFU doses were necessary to get a high antibody titer when bacteria were administered intranasally, since 3 of 5 mice that were immunized via the intranasal route with 103 CFU of the Y. pestis ΔBA mutant did not possess detectable levels of antibodies (not shown). However, immunization with 104 CFU bacteria increased the titer in the range of 1:800 to 1:6,400. Immunization with 106 CFU or higher doses yielded higher titers, ranging from 1:25,600 to 1:102,400. In addition, bacterial lysates were analyzed by Western blotting to reveal major antigens recognized by mouse immune sera. Results showed that administration of bacteria via the intravenous and intranasal routes produced similar antibody profiles (Fig. (Fig.5B).5B). Interestingly, mice immunized with the Y. pestis ΔBA mutant via either the intranasal or intravenous route produced no detectable level of antibodies against LcrV. However, we did observe a strong antibody response against the F1 capsular antigen, independent of the route of infection. Our analysis also revealed that the observed antibody responses depended on live cells, since intranasal administration of heat-killed bacteria did not yield significant levels of antibodies that recognized major Y. pestis antigens (Fig. (Fig.5C,5C, lane 7).
We also analyzed the antibody response triggered by orogastric immunization with the Y. pseudotuberculosis ΔBA strain. As expected, since Y. pseudotuberculosis does not possess the caf1 gene, sera from mice immunized with the ΔBA mutant of this pathogen did not show F1-specific antibodies (Fig. (Fig.5C,5C, lanes 5 and 6). Immunization with a dose of 5 × 106 CFU yielded very low levels of antibodies that cross-reacted with Y. pestis antigens. However, it was possible to detect Y. pestis-reactive antibodies in sera from mice immunized with a higher dose (2 × 107 CFU) of the Y. pseudotuberculosis ΔBA strain (lane 6), albeit significantly less than for mice immunized with Y. pestis ΔBA cells.
Next, we performed passive immunization studies to investigate whether the protection observed with Y. pestis ΔBA cells depended solely on anti-F1 antibodies. To this end, sera were isolated from mice that had been immunized intranasally with CO92Δpgm ΔBA (104 CFU) or the PBS control. One hundred μl of the immune sera were diluted 1:1 with PBS and injected intraperitoneally into naïve mice 18 h after infection with the CO92Δpgm, KIM5, or KIM5caf1A strain. The KIM5caf1A mutant cannot assemble F1 capsule on its surface due to the lack of the Caf1A usher, which is required for secretion of the F1 subunit (25, 48). Since there is no established intranasal infection protocol for KIM5caf1A, mice were infected intravenously with ~103 CFU (~100 LD50) bacteria as described by Ivanov et al. (25). As expected, all of the control mice that were passively immunized with PBS serum succumbed to infection with CO92Δpgm, KIM5, or KIM5caf1A (Table (Table3).3). On the other hand, passive immunization of mice with the CO92Δpgm ΔBA serum protected them against lethal challenge with CO92Δpgm or KIM5. Interestingly, transfer of the CO92Δpgm ΔBA serum did not protect naïve mice against the KIM5caf1A strain, confirming the role of F1 as a major protective antigen. In contrast, when mice were actively immunized with the CO92Δpgm ΔBA strain (104 CFU) and challenged with the KIM5caf1A strain, we observed that 3 of 6 mice survived the lethal challenge (0 of 3 mice survived after PBS control immunization). Taken together, these results suggest that vaccination with the CO92Δpgm ΔBA strain triggers a robust protective response against Y. pestis infections, which includes a strong antibody response against the F1 antigen.
In this study, we demonstrated that the smpB-ssrA mutant of Y. pestis CO92Δpgm was severely attenuated in a mouse model of infection. ΔBA mutant cells were defective in colonization and unable to establish a lethal infection when administered intravenously or intranasally, even at very high doses. Additionally, mice infected with ΔBA cells did not show any signs of infection. Most importantly, our analysis showed that mice immunized via the intranasal route with a low dose of 104 CFU ΔBA bacteria were completely protected against lethal pulmonary Y. pestis challenges. This is a promising finding toward the goal of designing an effective plague vaccine that is highly protective against pneumonic and bubonic plague with minimal side effects.
The SmpB-SsrA system fulfills an important task in maintaining the translational machinery at full capacity by dealing with aberrant mRNAs that stall ribosomes. Although smpB-ssrA mutations in most bacterial species may not be associated with a growth defect under normal laboratory conditions, it is apparent that mutant cells become compromised in their ability to adapt and survive in hostile environments. For example, it is well documented that smpB-ssrA mutations render cells more sensitive to translation-specific inhibitors as well as oxidative and nitrosative stress (16, 27, 34). The versatility to adapt to different environments can be critical for pathogenic bacteria, since they have to deal with a variety of host defense mechanisms. Phagocytosis is one such mechanism, which is essential in guarding against and disposing of facultative intracellular pathogens like Yersinia (9, 22, 33). smpB-ssrA mutants of Y. pseudotuberculosis and Salmonella enterica serovar Typhimurium were shown to be impaired in intracellular survival and replication within macrophages (6, 34). Additional functions of the SmpB-SsrA system in bacterial pathogenesis can be illustrated by alterations in the expression of virulence factors. smpB-ssrA mutants were reported to have deregulated expression of Salmonella ivi genes and Yersinia yop genes (26, 34). The overall combination of defects suffered by the loss of SmpB-SsrA function contributes to a high level of attenuation. Therefore, such mutant strains should be strongly considered a candidate for live cell vaccines.
A number of studies have utilized recombinant LcrV and the F1 capsular antigen as potential vaccine candidates. Although these subunit vaccines look promising, some shortcomings have been reported (1-3, 5, 13, 18, 23). An alternative strategy has been to design vaccines based on live attenuated cells. One theoretical advantage of such a vaccine strain is the potential to stimulate both humoral and cellular immunity while simultaneously priming the host immune system against many antigens (7, 30, 36). Another advantage of using attenuated live cells is that bacteria can easily be directed to express additional antigens or variants of antigens. Furthermore, any desired modifications to subunit vaccines can be costly, since the procedure will likely require new sets of protein purification and optimization protocols. However, desired modifications could be achieved with minimal cost with live-cell vaccines.
Use of a less-virulent and highly related surrogate pathogen as a vaccine strain has been a known strategy in the field of vaccine research and can historically be exemplified by smallpox vaccines. Similarly, using Y. pseudotuberculosis as a live plague vaccine has remained an intriguing alternative. Y. pseudotuberculosis is very closely related to Y. pestis and possess more than 95% DNA sequence homology. Despite sharing many characteristics, Y. pseudotuberculosis causes a less dangerous enteric course of illness, which is rarely lethal (12). Previous studies by Wake et al. (50) and Simonet et al. (45) showed that intravenous or subcutaneous administration of Y. pseudotuberculosis provided 50% protection against bubonic plague in a mouse model. More recently other researchers have shown that oral delivery of a Y. pseudotuberculosis dam mutant or the naturally attenuated IP32680 strain provides significant protection against bubonic plague (8, 47). One theoretical advantage of oral delivery of Y. pseudotuberculosis cells is the possible induction of gut-associated mucosal immunity. Such immunity is usually associated with secretory IgAs and could potentially be protective against infection of other mucosal surfaces, such as lungs. Consistent with these findings, our results indicate that oral immunization of mice with the Y. pseudotuberculosis ΔBA strain provided ~50% protection against pulmonary Y. pestis infection (Table (Table2).2). The aforementioned studies similarly reported that oral immunization with Y. pseudotuberculosis did not significantly protect mice against pneumonic plague. To explain the lack of protection, several factors have to be considered: Y. pseudotuberculosis does not possess the F1 capsular protein, an immunodominant antigen. It is possible that oral immunization is unable to trigger the robust IgA and IgG responses that are necessary to protect mice against Y. pestis. In fact, analysis of serum IgGs from mice immunized with IP2666 ΔBA showed no significant level of antibodies that could cross-react with Y. pestis (Fig. (Fig.5C,5C, lane 5). Therefore, oral vaccines based solely on Y. pseudotuberculosis may not be effective against pneumonic plague.
Y. pestis EV76 is an attenuated strain of Y. pestis that lacks the pigmentation locus (Δpgm) in addition to other genetically undefined modifications. The pigmentation (pgm) locus is a 102-kb chromosomal DNA region that can be spontaneously lost at low frequencies (11, 29). The pgm locus is associated with several virulence factors, such as Ybt and hms, which are related to iron uptake and biofilm formation, respectively, and ripA, which is required for survival in activated macrophages (20, 31, 37, 40). The EV76 strain has been used as a live-cell vaccine, and its effectiveness is strongly supported by field observations (46, 49). However, the highly reactogenic nature of the EV76 strain is known to cause severe side effects. Also, the possibility of it regaining its virulence highlights the reluctance to use the EV76 strain as a widely accepted vaccine strain. Recent studies identified several promising Y. pestis mutations that could be protective in animal studies. These mutations include pcm, dam, and yopH (10, 19, 43). In theory, incorporating additional genetically defined mutations into a Y. pestis pgm-negative strain may eliminate these drawbacks while maintaining the desired efficiency as a live-cell vaccine. Therefore, it is important to identify additional mutations and use the best combination among them to design the safest and the most effective live attenuated plague vaccine strain.
Our studies showed that both intravenous and intranasal delivery of Y. pestis ΔBA cells induced a strong IgG response against several Yersinia proteins (Fig. (Fig.5),5), including the F1 antigen. Although passive immunization studies suggest that antibodies against F1 play a major role, innate immunity may also contribute to protection induced by immunization with ΔBA cells. Indeed, mice immunized with the Y. pestis ΔBA mutant provided partial protection (50% protection) against lethal challenges with F1− Y. pestis cells. Consistent with these observations, studies with Y. pseudotuberculosis ΔBA cells, which lack the F1 antigen, showed that immunization of mice with this bacterium provided 50% protection against a lethal challenge with Y. pestis, despite the absence of IgG antibodies that recognize major Y. pestis antigens (Fig. (Fig.5C,5C, lane 5). Furthermore, it is possible that any potentially protective IgA response acquired by intranasal immunization may not be utilized in countering a lethal challenge through the intravenous route. Therefore, it is possible that intranasal immunization with the Y. pestis ΔBA strain manifests its full potential against pulmonary Y. pestis infections.
In summary, the smpB-ssrA mutant of Y. pestis is highly attenuated, and this defined mutation could be an excellent candidate for the foundation of a live-cell vaccine.
We thank Celine Pujol for the construction of the pgm-negative Y. pestis CO92 strain and Jens P. Grabenstein for the construction of the pSB890-Δcaf1 plasmid.
This work was supported by the USPHS grants PO1-AI-055621 (to A.W.K. and J.L.B.), RO1-AI-043389 (to J.B.B.), and RO1-GM-065319 (to A.W.K.) and Northeast Biodefense Center grant no. U54-AI057158-Lipkin (to J.L.B. and J.B.B.).
Editor: A. J. Bäumler
Published ahead of print on 11 January 2010.