|Home | About | Journals | Submit | Contact Us | Français|
Antibody responses for 882 of the 1,039 proteins in the proteome of Treponema pallidum were examined. Sera collected from infected rabbits were used to systematically identify 106 antigenic proteins, including 22 previously identified antigens and 84 novel antigens. Additionally, sera collected from rabbits throughout the course of infection demonstrated a progression in the breadth and intensity of humoral immunoreactivity against a representative panel of T. pallidum antigens.
Syphilis is a multistage sexually transmitted disease that remains a public health problem, with an estimated 12 million new cases per year worldwide (13). If untreated, syphilis can evolve from localized primary lesions (chancres) to disseminated, chronic infections, including secondary, latent, and tertiary forms (32). Primary, latent, and occasional secondary manifestations can be reproduced in experimental animal models, including the rabbit (25). Although no vaccine for the prevention of syphilis is currently available (20, 34), the immunization of rabbits with recombinantly expressed antigens prior to infection has resulted in attenuated lesion development (4, 6, 7, 35). Therefore, a combination of recombinantly expressed antigens may provide protection in rabbits and, hopefully, humans. Systematic screening of antigens for reactivity with antibodies from Treponema pallidum-infected animals represents a logical approach for identifying new, potentially protective antigens.
An immunoassay that utilizes a protein array was employed to identify antigenic proteins (21, 28). Of the 1,039 predicted genes in the T. pallidum genome, 882 were fused to glutathione S-transferase (GST) by using the Univector Cre-loxP recombination system (19), and the fusion proteins were tested for antibody binding. Protein arrays were generated by using the GST fusion protein set deposited into 96-well enzyme-linked immunosorbent assay plates coated with glutathione. Construction of the GST clone set and the techniques used to generate the protein arrays have been described previously by McKevitt et al. (21). Rabbit sera were derived from three rabbits that were each infected with 4 × 108 Treponema pallidum Nichols strain organisms by intratesticular injection. Sera were collected 2 days before infection, as well as 7, 14, 28, 56, and 84 days postinfection. All animal procedures were reviewed and approved by the University of Texas Health Science Center Animal Welfare Committee. Sera from the three rabbits were pooled for use in the enzyme-linked immunosorbent assay experiments to detect antigenic proteins. Rabbit serum (45 μl) that had been preabsorbed to Escherichia coli protein lysates (0.70 ml) (21) was diluted in 12 ml phosphate-buffered saline (pH 7.4) and 0.05% Tween 20 (buffer A) for a final serum dilution of 1:282. In order to systematically identify antigenic proteins, 110 μl of the preabsorbed, diluted sera pooled from three infected rabbits was incubated in each well of the protein arrays and washed five times with buffer A. Proteins reactive with antibodies in the sera were identified by chemiluminescence with an anti-rabbit immunoglobulin G antibody conjugated to horseradish peroxidase as described previously (21) (Fig. (Fig.1).1). Light emission from each plate well was monitored at 5- and 15-minute intervals with a Genios plate reader (Tecan, Durham, North Carolina) for 200 milliseconds; the results at these two time points were comparable and were averaged. When the reactivity of the fusion protein was significantly greater than the background reactivity of the immunoassay, as judged by comparing the experimental fusion protein data with the GST protein-only negative control data by using the Student t test, the ability of a protein to elicit an immune response was inferred. The experiment was repeated twice, and the results were combined and used for statistical analysis. Analysis of the immunoassay data collected from 882 arrayed proteins indicated that 106 proteins exhibited reactivity with antibodies at higher-than-background levels (Fig. (Fig.1;1; Table Table11).
Examination of T. pallidum literature generated a list of 29 proteins known to elicit an immune response in humans or rabbits. Of these 29 proteins, 27 were included in the array, and 22 of these were detected as antigens (Table (Table1).1). Reactivity of 81% of the known immunogenic proteins indicates that the assay provides a representative view of the immunoproteome. The identified proteins, however, are likely to be an incomplete set of all the antigens that react with serum antibodies from infected rabbits, for several reasons. First, only 882 of the 1,039 predicted proteins in the T. pallidum proteome were assayed. Second, some of the T. pallidum proteins may be poorly expressed in the system. For example, the TprK protein is known to elicit a strong humoral and cell-mediated immune response (8, 22, 23) but was not identified in the antigenic screen (Table (Table1).1). Examination of expression of the GST-TprK fusion protein by immunoblotting with anti-GST antibody indicated that the fusion is very poorly expressed in E. coli, suggesting that the lack of reactivity with rabbit sera is due to the absence of the protein in the lysate. Thus, TprK is a false negative in these experiments. Finally, DNA sequence information for 106 clones and a total of 72,187 bp in the clone set indicated one error per 2,673 bp (generated during PCR amplification of the genes prior to insertion into the Univector plasmid); therefore, some genes are likely to be false negatives due to PCR-generated mutations (21). The occurrence of false negatives, however, does not invalidate the positive identification of the 106 antigens identified in Table Table11.
Functional classifications have been assigned to 56 of the 106 antigens identified (Table (Table1).1). Proteins associated with the cell envelope are the most represented class (35%). Twenty-four of the 35 proteins (69%) exhibiting the highest reactivity were predicted to encode signal peptides or an N-terminal transmembrane helix (18); 10 of these were predicted or experimentally verified lipoproteins, consistent with prior observations that lipoproteins are among the immunogenic proteins of T. pallidum (Table (Table11).
The GST fusion proteins were next used to monitor the development of the humoral immune responses in rabbits against T. pallidum infection (Table (Table2).2). Sera from three rabbits were collected and pooled prior to the inoculation of T. pallidum and again at 7, 14, 28, 56, and 84 days after intratesticular inoculation. For this analysis, 74 proteins that are representative of the 106 reactive proteins identified in the global screen were selected (Table (Table22).
With the exception of the protein encoded by TP0329 (Table (Table2),2), little reactivity between serum antibodies from uninfected rabbits and T. pallidum proteins was observed. At 7 days postinfection, however, antibody binding at levels twofold higher than the level of binding to the negative control, GST protein alone, was observed for the following proteins: TP0319 membrane lipoprotein TmpC, TP0971 membrane antigen TpD, TP0574 47-kDa carboxypeptidase, and TP0684 methylgalactoside ABC transporter. These proteins bound high levels of antibody at all postinfection time points (Table (Table2).2). The early and continuous antibody response is a characteristic that would make these proteins useful as diagnostics of T. pallidum infection.
The humoral response broadened at 14 days postinfection in that antibodies bound to 19 of the 74 proteins arrayed at levels twofold higher than the binding level of GST alone (Table (Table2).2). The most reactive proteins at this time point corresponded with those seen at 7 days postinfection, including MglB-2, the 47-kDa carboxypeptidase, and TmpC. By 28 days postinfection, the immune response was robust, with significant antibody binding to 47 of the 74 proteins tested. The response continued to expand, however, and by 84 days postinfection, 70 of the 74 proteins assayed exhibited twofold-higher levels of antibody binding than the GST-only control.
In general, proteins with relatively low reactivity at 84 days did not elicit a humoral response until late in the infection, while those proteins that exhibited high levels of antibody binding at 84 days elicited a response early that continued to increase up to the 84-day level (Table (Table2).2). An exception is the hypothetical protein encoded by TP0956 that exhibited a 30-fold increase in antibody binding between 56 and 84 days. The significance of this finding is that rabbits typically exhibit only partial immunity to reinfection 1 to 2 months after intratesticular inoculation of T. pallidum. However, by 3 months postinoculation, rabbits exhibit chancre immunity (31). The increase in antibodies against the TP0956 protein correlates with the timing of chancre immunity, and therefore, it is an interesting candidate as a potential protective antigen.
This study utilized a comprehensive clone set consisting of ~90% of T. pallidum open reading frames (ORFs) (21) to conduct the first systematic search for antigenic proteins in this organism. A large set of antigens which were further characterized with respect to the timing of the immune response was identified. Proteins that elicit an early response that continues throughout experimental rabbit infection may be useful for immunodiagnosis of humans. Additional studies with sera from individual rabbits and humans to distinguish antigens that consistently elicit an antibody response in patients with syphilis or other treponemal infections are needed. Several of the novel antigens identified in these experiments are currently being tested in rabbits for the ability to provide protective immunity.
This work was supported by National Institutes of Health grants AI45842 (to T.P.) and AI49557 (to S.J.N.).
We thank Mary Mosher for technical assistance. We also thank Joseph Petrosino for comments on the manuscript.