Construction of the ECA-deficient Salmonella mutant strain
In the S. Typhimurium wild type strain UK-1 χ3761 and its attenuated vaccine strain χ9241, the ECA operon consisted of a series of wec genes, ranging from wecA to wecG, as stated in Fig. 1A. We deleted the whole ECA operon of the wild type strain UK-1 χ3761 and the attenuated vaccine strain χ9241 and generated two ECA defective mutants, χ12357 and χ12358, by suicide plasmid-mediated homologous recombination (Fig. 1B). A pair of primers, Doperon-1F/Doperon-2R, were used to amplify the ECA operon from χ 3761 and χ12357, and a PCR product of approximately 800 bp could be observed for χ12357, but no bands were observed for χ3761 in agarose gel, which indicated that mutants with deletion of the ECA operon (from wecA to wecG) were created in χ12357 (Fig. 1C). Furthermore, immunoblotting using the mouse anti-ECA monoclonal antibody mAb898 demonstrated deletion of the EAC operon, showing the ECA-negative phenotype [33] (Fig. 1D). Following the same method, we also deleted the ECA operon in the attenuated vaccine strain χ9241 and created an ECA defective vaccine strain χ12358 (data not shown).
Phenotypic characterization of the ECA deficient mutants
The growth rate of the mutants was evaluated in LB broth and showed that the growth curves of the ECA-deficient mutants χ12357 and χ12358 were similar to those of their parental strains χ3761 and χ9241. Both χ3761 and χ12357 reached the end of the log phase with an OD600 of 1.5 for approximately 5 hours, while the vaccine strain χ9241 and its derivative strain χ12358 had slightly lower rates of growth due to the deletion of a few virulent genes, such as pabA, pabB, araBAD and relA (Fig. 2A). The LPS profile of the ECA-deficient mutant χ12357 was indistinguishable from that of the parental wild type strain χ3761, indicating that ECA deficiency has no effect on the synthesis of LPS in S. Typhimurium (Fig. 2B). To determine the effect of ECA deficiency on bile resistance, drops of serial dilutions of χ3761 and χ12357 strain suspensions were incubated on LB plates with or without 1% deoxycholate (the main active ingredient of bile, DOC). The growth state of χ3761 and χ12357 appeared similar on regular LB plates (plates 2 and 3). A few colonies were observed on the 107 dilution drops of both strain suspensions, suggesting no differences in the growth rate and susceptibility between the two strains on LB plates. Nevertheless, their growth states were totally different on the LB plates containing 1% DOC (plate 1). For the wild type strain χ3761, a few colonies grew on the 107 dilution, and the same growth tendency as that of the regular LB plate was observed, suggested that 1% DOC had no effect on the growth of the wild type strain; however, the 105 dilution of the χ12357 suspension grew a few colonies, suggesting that the wec operon mutants were significantly suppressed by the presence of 1% DOC on plate 4 (Fig. 2C). Moreover, under the high concentration DOC bile condition, 87.6% of the wec operon χ12357 mutants were also killed, but more than 90% of the wild type strains remained viable (Table 3). To assess the relationship between ECA and motility, the ECA-deficient mutant χ12357 and wild type χ3761 strain were dropped onto 0.3% agar plates. Both the ECA-deficient mutants and wild type strains were able to swim on soft agar, with no defects in motility (Table 1).
Virulence and colonization of the mutant strain in mice
The virulence of the ECA-deficient mutant was assessed by determining the oral LD50 of the strains in BALB/c mice. The results showed that χ3761 was highly virulent with an LD50 of 5×104 CFU, whereas the ECA-deficient mutant χ12357 was highly attenuated and no deaths were observed following infection of the mice at the highest tested dose (1.0×109 CFU), which suggested that the wec operon mutation made the wild type S. Typhimurium avirulent (Table 1). Further, the capability of the ECA-deficient mutant χ12357 to efficiently disseminate and colonize mouse tissues was assessed by determining the bacterial loading in the organs of the mice, including the liver, spleen and Peyer’s patch, post-infection (Fig. 3A, 3B, 3C). Consistent with the previous result, all of the mice inoculated with χ3761 succumbed to infection within 9 days post-infection, and all of the mice inoculated with χ12357 survived until 21 days post-infection and most of them appeared health and active. The colonization data showed that the bacterial counts in the organs from the mice inoculated with the wild-type strain rapidly increased, and bacterial loading in the organs of the wild-type-inoculated mice was significantly higher than that of the mice inoculated with the ECA-deficient mutant within 9 days post-infection. Further, the ECA-deficient mutant also colonized well in the mice organs and maintained a moderate level of bacterial loading, with persistent infection throughout the experiment. These data demonstrate that the ECA-deficient mutant was able to efficiently colonize the mice and establish a persistent infection, consistent with the results of a previous study [18].
Table 1 Motility, DOC sensitivity and virulence of the wild-type S. Typhimurium χ3761 and its ECA-deficient mutant
Strain
|
Motility (cm) on soft agar
|
Death rate (%) (DOC)
|
LD50
|
χ3761
|
6.2±0.1
|
9.5
|
2×104
|
χ12357
|
6.4±0.1
|
87.6
|
>109
|
Expression of the pneumococcal antigen PspA in the Salmonella strains
Attenuated Salmonella χ9241 carrying ΔpabA ΔpabB mutations is often used in our lab to evaluate the effects of other mutations on immunogenicity [24]. To assess the effects of ECA deletion on immunogenicity in attenuated Salmonella, we introduced the wec operon mutation into χ9241 to yield strain χ12358. Moreover, the asd recombinant plasmid pYA4088, which carries a recombinant pspA gene fused to the DNA encoding the β-lactamase signal sequence to direct PspA to the periplasmic region, was introduced into both χ9241 and χ12358. Both strains (χ9241 and χ12358) were grown in LB medium with or without 0.1% arabinose, and western blot analyses were used to determinate the synthesis levels of PspA in the above bacterial cells. No PspA was detected in χ9241 (pYA3493, carrying the empty vector, which served as a negative control) regardless of whether it was grown in the presence of arabinose. Compared with the negative control strain χ9241 (pYA3493), both strains (χ9241 and χ12358) carrying pYA4088 produced PspA when arabinose was absent, and instead expressed LacI in the presence of arabinose (Supplementary Fig. 1). This result suggests two points: the regulated delayed expressed system was successfully used to construct the attenuated vaccine vector strains χ9241 (pYA4088) and χ12358 (pYA4088); and the wec operon mutant χ12358 has a similar ability to express heterologous antibodies, such as PspA, as its parent strain χ9241.
Immunogenicity of the wec operon mutants after oral administration
To assess the immunogenicity of the wec operon mutants expressing heterologous antigens as vaccine vectors, three groups of mice were immunized and boosted with 1×109 of χ12358 (pYA4088), χ9241 (pYA4088) or χ9241 (pYA3493). Serum IgG, IgG1 and IgG2a antibody responses to PspA in sera collected at week 4 and week 6 were measured by ELISA. Compared with the negative control strain χ9241 (pYA3493), both χ12358 (pYA4088) and χ9241 (pYA4088) significantly stimulated mice to produce high levels of IgG, IgG1 and IgG2a anti-PspA antibodies, suggesting that the deletion of the wec operon had no effect on the immunogenicity of the vaccine strains. Noticeably, χ12358 (pYA4088) induced levels of anti-PspA IgG1 isotype antibodies that were significantly higher than that of χ9241 (pYA4088) (P<0.001), but the levels of total IgG and IgG2a isotype antibodies in the mice induced by χ12358 (pYA4088) were significantly lower than that induced by χ9241 (pYA4088) (Fig. 4A, 4B, 4C).
Evaluation of immunogenicity against conserved OMPs from different enteric bacteria
To evaluate the cross-reactivity of the serum antibodies raised from mice orally immunized with the mutant χ12358 (pYA4088) against outer membrane proteins (OMPs) purified from several different homologous and heterologous wild-type enteric bacterial strains, including S. Typhimurium (group B), S. Choleraesuis (group C1), E. coli O78 and S. Enteritidis (group D1), we measured the IgG antibody responses to OMPs from the above pathogens in serum collected at week 6 after the mice were immunized with χ12358 (pYA4088) and χ9241(pYA4088). The IgG level of cross-reaction against OMPs from S. Typhimurium (group B), E. coli O78 and S. Enteritidis (group D1) stimulated by χ9241 (pYA4088) and χ12358 (pYA4088) were not different, but the level of IgG against OMPs from S. Choleraesuis stimulated by the wec mutant χ12358 were significantly higher than that induced by χ9241 (pYA4088) (Fig. 5). This result was consistent with a previous study, in which a lack of or the regulated synthesis of dominant surface antigens, such as the O-antigen of LPS, increased the serum antibody cross-reactive to conserved surface OMPs of other enteric bacteria [24, 34].
Comparison of the protective efficacy of live attenuated vaccine strains (RASVs) and their wec operon mutants
To test whether the wec operon mutation affects the protective efficacy of RASVs against pneumococcal infection, BALB/c mice orally immunized and boosted with χ9241 (pYA4088) and χ12358 (pYA4088) were challenged i.p. with 2.0×104 CFU (50 times of LD50) of wild type S. pneumoniae WU2. No mice in the negative control groups immunized with χ9241 (pYA3493) survived. However, both of the χ9241 (pYA4088) and χ12358 (pYA4088) strains provided significant protection against the challenge. Moreover, there was no difference in the level of protection afforded by the RASVs and wec operon mutants, suggesting that deletion of the wec operon had no influence on the protective efficacy (Table 2).
Table 2 Oral immunization with the PspA-expressing S. Typhimurium χ9241 (pYA4088) vaccine and χ12358 (pYA4088) protected BALB/c mice against a challenge with the virulent S. pneumoniae strain WU2
Vaccine
|
PspA expressiona
|
Total mice/group
|
Survival/totalb
|
Percent protectionb
|
χ12358 (pYA4088)
|
+
|
12
|
4/12
|
33%
|
χ9241 (pYA4088)
|
+
|
12
|
4/12
|
33%
|
χ9241 (pYA3493)
|
-
|
12
|
0/12
|
0%
|
a +, PspA expressed; -, PspA not expressed.
b Mice were challenged with 4×104 CFU of S. pneumoniae WU2 (50 times LD50) 4 weeks after the second immunization.