Serologic Evidence for Effective Production of Cytolysin A in Salmonella enterica Serovars Typhi and Paratyphi A during Human Infection

Abstract

ClyA_STy and ClyA_SPaA are closely related pore-forming cytolysins of Salmonella enterica serovars Typhi and Paratyphi A whose expression is strongly repressed under standard in vitro growth conditions. We show here that human infections by these pathogens cause a specific antibody response to ClyA, indicating effective toxin production during infection.

Pore-forming cytolysins represent important virulence factors of bacterial pathogens. Recently, a novel family of this type of toxins has been detected in some organisms belonging to the Enterobacteriaceae. The prototype member of this toxin family is cytolysin A (ClyA) from Escherichia coli, a hemolytic and cytotoxic 34-kDa protein that forms stable pores in target membranes by assembling into ring-shaped toxin oligomers (1, 4, 5, 7, 12, 17, 18, 24, 25). ClyA is encoded by the chromosomal clyA gene (also referred to as hlyE and sheA), which is found in many E. coli strains causing enteric/diarrheal diseases as well as in E. coli K-12 (4, 5, 11, 12, 17). This gene is under the control of several transcriptional regulators; under standard in vitro cultivation conditions, its expression is strongly repressed (5, 10-12, 15, 17, 26, 27).

Salmonella enterica serovar Typhi, the etiologic agent of typhoid fever, and Salmonella enterica serovar Paratyphi A, which causes a similar systemic disease referred to as paratyphoid fever, were recently shown to harbor functional clyA homologues, termed here clyA_STy and clyA_SPaA, while various other Salmonella enterica subsp. enterica serovars (serovars Paratyphi B, Paratyphi C, Typhimurium, Enteritidis, and others) were found to lack such a gene (16; C. von Rhein, S. Bauer, E. J. López Sanjurjo, R. Benz, W. Goebel, and A. Ludwig, submitted for publication). The proteins encoded by clyA_STy and clyA_SPaA, ClyA_STy and ClyA_SPaA, are 90 to 91% identical in amino acid sequence to ClyA from E. coli K-12 (ClyA_K-12); they show hemolytic and cytotoxic activities and generate membrane pores that have properties very similar to those of pores formed by ClyA_K-12. Expression of clyA_STy and clyA_SPaA, like that of E. coli clyA, proved to be strongly down-regulated under standard laboratory growth conditions. Thus, when grown in or on a rich medium, wild-type Salmonella serovar Typhi and serovar Paratyphi A strains produce ClyA only in very small, basal amounts that are detectable only within the bacterial cells and do not show a clyA-dependent hemolytic phenotype (16; C. von Rhein et al., submitted).

The question whether Salmonella serovar Typhi and serovar Paratyphi A strains synthesize ClyA_STy and ClyA_SPaA in significant amounts during infection has not yet been investigated. In this study, we addressed this issue by analyzing whether infections by these pathogens specifically induce the production of ClyA-reactive antibodies in the human host. For this purpose, serum samples from patients with Salmonella serovar Typhi and/or serovar Paratyphi A infections and from individuals with various other physical conditions were tested by Western blot analysis for the presence of such antibodies.

Serum samples.

The sera tested in this study were obtained as follows. (i) Serum samples were taken from individuals with Salmonella serovar Typhi and/or serovar Paratyphi A infections. This group comprised eight patients with active or recently contracted enteric fever (typhoid or paratyphoid fever) caused by Salmonella serovar Typhi (n = 5), serovar Paratyphi A (n = 2), or both serovars (n = 1) and one chronic carrier of Salmonella serovar Typhi. The nine individuals were hospitalized between May 2001 and April 2005 in Frankfurt am Main, Germany, and had contracted the Salmonella infections abroad at different times and at different places. The causative serovar Typhi and serovar Paratyphi A strains were isolated from blood or stool specimens of these patients. All these strains were shown by PCR and DNA sequencing to harbor a functional clyA_STy (serovar Typhi strains) or clyA_SPaA (serovar Paratyphi A strains) gene, but none of them exhibited a clyA-dependent hemolytic phenotype on normal yeast extract-tryptone (YT) agar plates (11) supplemented with horse blood. Western blot analyses revealed that these strains produce only basal amounts of ClyA (detectable only within the bacteria) when grown under aerobic conditions in YT medium. (ii) Sera were collected from persons who had suffered infection with nontyphoid Salmonella serovars. This group comprised six individuals with culture-proven Salmonella serovar Enteritidis infections, one individual with a recent Salmonella serovar Typhimurium infection, and eight individuals showing serological diagnostic data (Widal test results) indicative of recent or former infections with other nontyphoid Salmonella serovars. (iii) Serum samples were taken from 124 persons with disorders other than Salmonella infections. This control group included 19 individuals with culture-proven Yersinia infections, 34 individuals showing clinical and serological signs of Lyme borreliosis, 20 individuals with active or recent syphilis, 20 individuals with active or recent Epstein-Barr virus infection, 20 individuals showing positive results for antinuclear antibodies, and 11 individuals showing increased levels of rheumatoid factors (6). (iv) Sera were collected from 128 healthy blood donors. In principle, a single serum sample from each individual was tested in this study.

Western blot analysis of human serum samples for reactivity with ClyA.

To prepare ClyA antigen, ClyA_K-12 and ClyA_SPaA were overexpressed in E. coli DH5α from plasmid pAL201 (12) and plasmid pAL207 (C. von Rhein et al., submitted), respectively, isolated from the E. coli cells by osmotic shock, and purified by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as previously described (12). For each Western blot assay, several identical aliquots of purified ClyA (either ClyA_K-12 or ClyA_SPaA) were separated by SDS-PAGE (about 5 to 10 ng per lane) and blotted onto a polyvinylidene difluoride membrane as described recently (11). The membrane was then cut into strips corresponding to the lanes containing ClyA, and after blocking, each strip was incubated overnight with a human serum sample at a final dilution of 1:5,000. The presence of membrane-bound ClyA-reactive antibodies was subsequently analyzed using either horseradish peroxidase (HRP)-conjugated secondary antibodies to human immunoglobulin G (IgG), IgA, and IgM (IgG/IgA/IgM) or HRP-conjugated affinity-purified anti-human IgM antibodies (Acris antibodies), both at a final dilution of 1:50,000. The binding of the secondary antibodies was visualized using the ECL Plus Western blotting detection system (Amersham Biosciences).

Table 1 summarizes the data of the Western blot assays, and Fig. 1 shows several of the test results for illustrative purposes. In principle, the results obtained were virtually the same no matter whether purified ClyA_K-12 or ClyA_SPaA was used as the antigen. The serum samples from all eight patients with active or recently contracted enteric fever caused by Salmonella serovar Typhi and/or serovar Paratyphi A exhibited rather strong polyvalent (IgG/IgA/IgM) reactivity and weaker IgM reactivity against ClyA. For the chronic carrier of Salmonella serovar Typhi, we found moderate polyvalent reactivity against ClyA while IgM reactivity was not detectable, in line with the presumption that the pathogen persists in niches where it is largely shielded from the host immune system. The serum samples from the 15 individuals with infections by nontyphoid Salmonella serovars showed no ClyA reactivity under the same conditions, in agreement with the finding that the clyA gene is absent in all hitherto tested S. enterica strains belonging to nontyphoid serovars.

TABLE 1.

Presence of ClyA-reactive antibodies in human serum samples as determined by Western blot analysis

Disorder or physical conditiona	No. of individuals tested	No. showing detectable levels of ClyA-reactive antibodiesb
		IgG/IgA/IgM	IgM
Active or recent infection by serovar Typhi and/or serovar Paratyphi A	8c	8 (++)	8 (+)
Chronic carriage of serovar Typhi	1	1 (+)	0
Infection by nontyphoid Salmonella serovars	15d	0	0
Other disorders
Yersinia infection	19	1 (±)	0
Lyme borreliosis	34	2 (±)	0
Syphilis	20	2 (±)	0
EBV infection	20	1 (±)	0
Presence of ANA	20	0	0
Increased levels of RF	11	0	0
Healthy blood donors	128	4 (±)	0

^aEBV, Epstein-Barr virus; ANA, antinuclear antibodies; RF, rheumatoid factors.

^bTo confirm the data, all serum samples were assayed at least twice in independent experiments for polyvalent (IgG/IgA/IgM) and IgM reactivity to ClyA. The strength of the reactivity is indicated in parentheses: ++, strong reactivity; +, weak to moderate reactivity; ±, very weak reactivity at the detection limit.

^cThese individuals had suffered infection with either Salmonella serovar Typhi (n = 5) or Salmonella serovar Paratyphi A (n = 2), or with both serovars (n = 1).

^dThese individuals had suffered infection with Salmonella serovar Enteritidis (n = 6), Salmonella serovar Typhimurium (n = 1), or other nontyphoid Salmonella serovars (n = 8).

FIG. 1.

Western blot analysis of human serum samples for the presence of ClyA-reactive antibodies. The sera tested here were drawn from individuals showing the following conditions: lanes 1 to 5, patients with active or recently contracted typhoid fever caused by the Salmonella serovar Typhi (S. Typhi) strains ST2757A/01, ST423/03, ST1989/03, ST4457/03, and ST1617/05, respectively; lanes 6 and 7, patients with active paratyphoid fever caused by the Salmonella serovar Paratyphi A (S. Paratyphi A) strains ST4042B/03 and ST463/05, respectively; lane 8, patient with active enteric fever caused by double infection with serovar Typhi strain ST2116A/03 and serovar Paratyphi A strain ST2116B/03; lane 9, chronic carrier of Salmonella serovar Typhi strain ST6941/01; lane 10, individual with recent infection by Salmonella serovar Typhimurium (S. Typhimurium); lane 11, healthy blood donor (typical result); lane 12, healthy blood donor showing very weak ClyA reactivity (false-positive result; rare). To test the sera for ClyA reactivity, polyvinylidene difluoride membrane strips carrying identical amounts (5 to 10 ng) of purified and electrophoretically separated ClyA_K-12 were incubated overnight with the different serum samples. ClyA-bound antibodies were subsequently detected on the membrane by using HRP-conjugated anti-human IgG/IgA/IgM secondary antibodies.

Regarding the other control groups, only 6 of the 124 individuals with disorders other than Salmonella infections (i.e., 4.8%) and only 4 of the 128 healthy blood donors (3.1%) showed a certain polyvalent serum reactivity against ClyA, which was, however, generally much weaker than that observed for the patients with Salmonella serovar Typhi and/or serovar Paratyphi A infections. In addition, IgM reactivity against ClyA was not found for any of these sera. Agglutination assays (Widal test assays) performed with the serum samples from individuals with non-Salmonella diseases and healthy blood donors showing weak reactivity against ClyA and defined Salmonella suspensions (Dade Behring, Marburg, Germany) containing O and/or H antigens of different Salmonella serovars (Typhi, Paratyphi A, Paratyphi B, Paratyphi C, Typhimurium, Enteritidis) revealed that these sera do not contain significant levels of agglutinating antibodies to Salmonella O or H antigens. Thus, there were no indications of former Salmonella infections in these cases, suggesting that the marginal ClyA reactivity of the corresponding sera was due to cross-reacting antibodies.

Taken together, the presence of readily detectable amounts of anti-ClyA antibodies in human sera obviously correlated with Salmonella serovar Typhi and serovar Paratyphi A infections. The differences in the detectability of ClyA-reactive serum antibodies observed between the group of patients with Salmonella serovar Typhi and/or serovar Paratyphi A infections and the control groups (taken both separately and as a whole) were statistically analyzed with the two-sided Fisher exact test by using BiAS software, version 8.1. P values of <0.05 were considered statistically significant. In fact, P was 0.0000 in all cases, which confirmed that the serologic differences detected are significant.

Conclusions.

The data presented here show that infections with Salmonella serovar Typhi and serovar Paratyphi A specifically trigger the production of substantial levels of ClyA-reactive antibodies in patients. To our knowledge, this is the first evidence that these pathogens synthesize effective amounts of ClyA (ClyA_STy and ClyA_SPaA, respectively) during infection, i.e., when they encounter the environmental conditions prevailing in their human hosts. Theoretically, ClyA_STy and ClyA_SPaA might hence play a role in the pathogenesis of typhoid and paratyphoid fever. The actual importance of these toxins in Salmonella serovar Typhi and serovar Paratyphi A infections is, however, unclear at present, since classical Salmonella serovar Paratyphi B and serovar Paratyphi C strains, which are also able to cause systemic, typhoid-like diseases (paratyphoid fever) in humans, do not encode ClyA (C. von Rhein et al., submitted).

It is not clear how strongly and in what way clyA_STy and clyA_SPaA are induced in vivo. Nevertheless, we have recently observed that clyA of typhoid Salmonella is positively controlled by the Salmonella transcriptional regulator SlyA (C. von Rhein et al., submitted). Previous studies have shown that SlyA is required for the virulence of Salmonella serovar Typhimurium (8), that it controls the expression of various virulence-associated genes of this pathogen (9, 13, 20, 22, 23), and that the expression of the slyA gene is activated in Salmonella serovar Typhimurium following internalization of the bacteria by macrophages (i.e., most likely in the phagosomal vacuoles of macrophages, where Salmonella replicates) (2, 14). A corresponding induction of slyA in Salmonella serovar Typhi and serovar Paratyphi A during infection might thus contribute to activation of the expression of clyA_STy and clyA_SPaA in the human host.

The results of the current study might also be of practical relevance, since they suggest that ClyA could be employed as a serologic marker for Salmonella serovar Typhi and serovar Paratyphi A infections. Clearly, typhoid and paratyphoid fever remain important diseases to be considered in the differential diagnosis of febrile diseases in the tropics and subtropics (3, 19, 21). Timely diagnosis of such diseases is crucial, but reliable direct laboratory confirmation of the infectious agents by culture or microscopy is difficult under the limited conditions of a field laboratory. Moreover, the detection of antibodies against typhoid salmonellae by conventional serological methods like the Widal test is error-prone and unsatisfactory due to low sensitivity and specificity. The development of sensitive serologic bedside tests based on specific antigens of typhoid Salmonella certainly would advance the rapid and reliable diagnosis of typhoid and paratyphoid fever. With regard to the detection of Salmonella serovar Typhi and serovar Paratyphi A infections, the sensitivity and specificity of the Western blot analysis described above, using ClyA as an antigen, were 100% and more than 95%, respectively (sensitivity indicates how many of the infected individuals are reliably detected with a test system, and specificity indicates how many of the noninfected individuals are reliably recognized as not infected). ClyA thus appears to be a very promising candidate antigen that is suitable for the detection of specific antibodies in patients with Salmonella serovar Typhi and serovar Paratyphi A infections.