Identification of the Major Antigenic Protein of Helicobacter cinaedi and Its Immunogenicity in Humans with H. cinaedi Infections

Abstract

Helicobacter cinaedi infection is now recognized as an increasingly important emerging disease. Its pathogenesis and epidemiological features are not fully understood, however. Here, we investigated the antigenic protein of H. cinaedi and the immunological response to it in H. cinaedi-infected patients. We constructed a genomic library of H. cinaedi from an H. cinaedi clinical isolate, and various H. cinaedi recombinant proteins were expressed. We identified the 30-kDa protein, encoded in an 822-bp H. cinaedi genome, as a major antigen, which was specifically recognized by serum from an H. cinaedi-immunized rabbit and H. cinaedi-infected patients. The gene encoding this 30-kDa antigen had high sequence similarity with genes encoding putative membrane proteins of bacteria. To evaluate whether the 30-kDa protein can be applied in serological testing for H. cinaedi infections, the recombinant protein was expressed in Escherichia coli as a His-tagged fusion protein and purified by Ni²⁺ affinity chromatography. Western blot analysis revealed strong immunoreactivity of the 31-kDa fusion protein with serum antibody from patients infected with H. cinaedi, but such an immunoreaction was absent or was very weak with uninfected control serum. An enzyme-linked immunosorbent assay using this H. cinaedi major antigen showed significantly high antibody titers for H. cinaedi-infected subjects compared with those of various control groups. We therefore conclude that the 30-kDa putative membrane protein is a major antigen of H. cinaedi and is useful for immunological and serological testing for clinical diagnosis and for further epidemiological study of H. cinaedi infection in humans.

Helicobacter cinaedi, which is classified as an enterohepatic Helicobacter species, causes bacteremia complicated by cellulitis or arthritis in humans (1, 12-14, 20, 21, 26-28). Although H. cinaedi reportedly affects mostly immunocompromised patients (12, 17, 18, 25, 32), we recently observed a series of 11 cases of H. cinaedi bacteremia and cellulitis that occurred after orthopedic surgery in a hospital setting in apparently immunocompetent patients (13). H. cinaedi has also been isolated previously from the blood of immunocompetent children and adults (30), and an immunocompetent patient of H. cinaedi myopericarditis was reported recently (15). H. cinaedi is normal intestinal flora of hamsters (7), and thus zoonotic (hamster-to-human) transmission has been suggested for some cases of neonatal infection (19). The clinical and epidemiological characteristics of H. cinaedi infection are not fully clarified, however. For example, although our recent report (13) suggested that an outbreak of H. cinaedi infection may have been caused by direct person-to-person spread, a question about the current level of H. cinaedi infection in the human community in general remained. Because of the fastidious nature of this organism, which we discovered during bacterial isolation and identification, optimizing culture and isolation methods is a prerequisite to better understanding of the epidemiology and clinical significance of this infection. Blood culture methods utilized in the clinical microbiology laboratory, including microaerobic culture using commercial gas generator envelopes, seem to overlook an appreciable number of H. cinaedi infections. In fact, we have had experience with several cases in which this bacterial species was not isolated but the infection was highly suspected because of marked antibody responses as assessed by an enzyme-linked immunosorbent assay (ELISA) using whole-cell antigens from H. cinaedi (13).

In this context, the development of immunological assays to analyze the specific immune responses of hosts infected with H. cinaedi is important. For example, serological tests such as ELISA, which can be applied exclusively to screening for these infections, are needed and H. cinaedi antigens related to immunological responses of H. cinaedi-infected hosts should be identified. Several H. cinaedi antigens were previously detected via Western blotting with serum obtained from H. cinaedi-infected patients, but these antigens did not show consistent antibody response profiles (5). We thus first examined the production of immunoreactive bands that reacted specifically with serum samples obtained from H. cinaedi-infected patients, and we identified, among several immunogenic proteins, a 30-kDa protein with the greatest immunoreactivity. Subsequently, we constructed the H. cinaedi genomic library, which allowed screening for genes encoding the major antigenic proteins, with use of serum specimens from an H. cinaedi-immunized rabbit and H. cinaedi-infected patients. In this study, we identified via rigorous methods the 30-kDa putative membrane protein as a major antigen and reported the human antibody responses to this antigen during H. cinaedi infection.

MATERIALS AND METHODS

Bacteria.

As the standard strain for the present genomic and protein analyses, we used a clinical isolate of H. cinaedi (gene accession number of the strain, AB275324 [13]), which we obtained during an outbreak of H. cinaedi bacteremia and cellulitis that we recently described (13). We cultured the isolate via routine methods with Campylobacter agar (Becton Dickinson, Franklin Lakes, NJ) containing 10% sheep blood and incubation at 37°C for 3 to 5 days under microaerobic conditions (CampyPak microaerophilic system; Becton Dickinson) with high humidity. For Western blot analysis of major antigens from H. cinaedi organisms, whole-cell lysate was prepared by sonication of bacterial cells that had been cultured on and collected from Campylobacter agar, as described earlier (8). Genomic H. cinaedi DNA was extracted from bacteria grown in brain heart infusion broth (Becton Dickinson) at 37°C under controlled microaerobic conditions in an atmosphere of 10% hydrogen for 3 to 5 days. Other bacteria used in this study were Helicobacter pylori ATCC 43504, Helicobacter hepaticus PAGU604 (LMG 16316), and Campylobacter jejuni clinical isolates. These strains were cultured by standard procedures and used mainly for Western blot analysis.

Anti-H. cinaedi antiserum and antibody.

Immune serum to H. cinaedi was obtained experimentally with a male Japanese White rabbit. To raise the antibody in the rabbit, the bacteria were washed, suspended in 10 mM phosphate-buffered saline (PBS) (pH 7.4), and sonicated to yield the H. cinaedi whole-cell lysate as described elsewhere (13). The rabbit was immunized subcutaneously with the lysate, 500 μg of protein in PBS with Freund's complete adjuvant, followed by two booster doses of 50 μg of lysate protein plus Freund's incomplete adjuvant at a 2-week interval. Six weeks after the initial immunization, blood was collected to obtain the immune serum. Similarly, antibody was developed for recombinant H. cinaedi protein by means of subcutaneous injection of 480 μg of recombinant protein with Freund's complete adjuvant and then three booster doses of 480 μg of protein each plus Freund's incomplete adjuvant.

Serum obtained from patients and other control subjects.

Samples of immune serum were collected from eight apparently immunocompetent patients (seven women and one man; ages ranged from 58 to 79 years, with an average of 69.6 years) with bacteremia and cellulitis caused by H. cinaedi who were admitted to Kumamoto Orthopedic Hospital, as we reported previously (13). Details of the clinical characteristics of these patients are reported in our previous paper: patients 1 to 8 (lanes 1 to 8) in Fig. 1A correspond to cases 1, 3, 5, 6, 7, 9, 10, and 11, respectively, documented in Table 1 of our previous paper (13). Another set of serum samples, which were obtained from one patient at various time points after the onset of fever due to H. cinaedi infection in the Kumamoto Orthopedic Hospital, was used for the time profile analysis of anti-H. cinaedi antibody production during infection. This patient (an 82-year-old woman without apparent underlying immunocompromising conditions [patient 5 in Fig. 8 below]) developed fever, with a positive blood culture of H. cinaedi, at 7 days after surgery for pyogenic spondylitis (for which an etiologic agent could not be identified) and was readily improved by the antibiotic treatment. In addition to the patient sera just mentioned, sera from four patients in the same hospital (all women of ages ranging from 71 to 77 years, with an average of 75.0 years [patients 1 to 4 in Fig. 8 below]) for whom we observed typical clinical symptoms for H. cinaedi infections such as cellulitis and fever without immunocompromising diseases were examined for their immune response to H. cinaedi. Similar to other H. cinaedi-infected patients, all four subjects had orthopedic operations for osteoarthritis (three patients) and bone fracture (one patient) before the onset of clinical symptoms suspected for H. cinaedi infection and responded well to the antibiotic chemotherapy. For control subjects, serum samples were obtained from various groups, including 9 age- and sex-matched patients (8 women and 1 man; ages ranged from 57 to 81 years, with an average of 70.6 years) at Kumamoto Orthopedic Hospital who had no apparent H. cinaedi infection, 10 H. pylori-infected subjects (2 women and 8 men; ages ranged from 37 to 69 years, with an average of 48.2 years) at Oita University Hospital, 5 infants younger than 1 year of age (Oita University Hospital), and 16 healthy volunteers (13 men and 3 women; ages ranged from 21 to 40 years, with an average of 27.4 years) working in the Department of Microbiology at Kumamoto University.

FIG. 1.

Antibody responses of H. cinaedi-infected patients (lanes 1 to 8) and an immunized rabbit (A) or control groups (B to D) to proteins in H. cinaedi whole-cell lysate as obtained via Western blotting. Each lane contained 6.5 μg of protein. Rabbit serum was diluted 1:10,000; human serum, 1:1,500. The predominant 30-kDa immunoreactive protein (arrowhead) and other high-molecular-mass proteins (bracket) are indicated. Immunized rabbit and H. cinaedi-infected patient profiles were similar. Normal rabbit serum (B) was preimmunized with H. cinaedi, and human controls (B to D) had no apparent H. cinaedi infection. In general, compared with responses shown in panel A, low immunoreactivity was observed, but lane 1 of the H. pylori-infected group (B), lane 8 of the age- and sex-matched control group (C), and lane 5 of the healthy volunteer group (D) showed weak to moderate 30-kDa bands (arrowheads). High-molecular-mass bands (50 to 100 kDa) were seen more frequently. Coomassie brilliant blue staining (CBB) is illustrated at the left. Molecular masses are given at the left in kilodaltons.

FIG. 8.

Time course of serum response in patients with suspected or true H. cinaedi infection and analyzed via recombinant His-MAP30_Hc-based ELISA. Day 0 indicates the time that cellulitis (patients 1 to 4) and fever (patient 5) were first recognized in the patients. Patients 1 to 4 showed typical symptoms of H. cinaedi infection though the bacteria were not isolated from them, while H. cinaedi was isolated from the blood of patient 5.

Ethical compliance.

Procedure for obtaining human serum samples used in this study complied with those recommended by the Regional Ethical Committee on Human Experimentation of Kumamoto University and Kumamoto Orthopedic Hospital. Written informed consent was obtained from each subject.

Construction of H. cinaedi genomic library.

Genomic DNA was extracted from cultured bacteria by using the DNeasy tissue kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol. Genomic DNA was partially digested with HindIII (Takara Bio Inc., Shiga, Japan), DNA was separated on an agarose gel (Invitrogen Corporation, Carlsbad, CA), and fragments ranging from 4 to 12 kbp were purified with the Qiaex II gel extraction kit (Qiagen). Partially digested genomic DNA was then ligated into the HindIII site of the pBluescript SK(+) vector (Stratagene Inc., La Jolla, CA). The ligation mixture was used to transform Escherichia coli ElectroMax DH10B T1 phage-resistant cells (Invitrogen) by electroporation, and then transformed bacteria were plated on Luria-Bertani agar plates containing 100 μg/ml ampicillin and 50 μg/ml 5-bromo-4-chloro-3-indoxyl-β-d-galactopyranoside (X-Gal), after which incubation proceeded for 14 to 16 h at 37°C.

Screening of the genomic library.

Colonies of transformed bacteria were adsorbed onto nitrocellulose membranes (Hybond-C Extra; GE Healthcare Bio-Sciences Corporation, Piscataway, NJ) which were pretreated with filter paper saturated with a lysis buffer (100 mM Tris-HCl, pH 7.8, containing 150 mM NaCl, 5 mM MgCl₂, 1.5% bovine serum albumin, 1 μg/ml pancreatic DNase I, and 40 μg/ml lysozyme) (22). Colony-adsorbed membranes were incubated for 1 h at 37°C with anti-H. cinaedi rabbit serum diluted 1:10,000 in 0.1% Tween 20-Tris-buffered saline (T-TBS) (10 mM Tris-HCl, 0.15 M NaCl, pH 7.6) containing 5% skim milk (T-TBS-skim milk); membranes were then washed three times with T-TBS and incubated with horseradish peroxidase (HRP)-conjugated anti-rabbit immunoglobulin G (IgG) antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) diluted 1:5,000 in T-TBS-skim milk. Immunoreactive colonies on membranes were visualized by using an ECL Plus detection system (GE Healthcare). Clones of transformants producing immunoreactive colonies were further cultured in 3 ml of Luria-Bertani broth containing 100 μg/ml ampicillin, and production of antigenic proteins was confirmed with a whole-cell lysate of each clone by using Western blotting, as described below. Four clones, designated 3-1, 8-2, 9-5, and 10-2, each containing corresponding plasmids, were identified and isolated.

DNA sequencing and analysis.

Plasmids from positive clones were isolated by using the QIAprep spin miniprep kit (Qiagen). Plasmid p9-5 was chosen for further study and was sequenced in several clonal batches by the dideoxy chain termination method via an ABI Prism 310 genetic analyzer (Applied Biosystems, Foster City, CA). DNA sequences were assembled and analyzed for open reading frames (ORFs) by inspection. A search for sequence similarity and conserved domains was conducted with National Center for Biotechnology Information (NCBI) BLAST programs (http://www.ncbi.nlm.nih.gov/BLAST/). LipoP (http://www.cbs.dtu.dk/services/LipoP/) was utilized for prediction of lipoprotein signal peptides in gram-negative bacteria (10).

Construction of the protein expression vector.

One of the ORFs found in clone 9-5, which was presumed to encode a 30-kDa major antigenic protein of H. cinaedi, henceforth tentatively called MAP30_Hc, was amplified by PCR with a thermal cycler (GeneAmp PCR system 2400; PerkinElmer Inc., Waltham, MA) by using the Expand high-fidelity PCR system (Roche, Hoffman-La Roche Ltd., Basel, Switzerland). The template was 200 ng of clone 9-5 plasmid; thermal cycling conditions were 94°C for 2 min and 40 cycles of 94°C for 15 s, 55°C for 30 s, and 68°C for 1 min. The forward and reverse MAP30_Hc ORF oligonucleotide primers used for this PCR were 5′-gggccatgggccatcatcatcatcatcatcatcatcatcacATGAAAAAGTTACTTGCTTTA-3′ (forward) and 5′-ccggatccTTAAAATGCAGGAACTATTTC-3′ (reverse); lowercase letters in each primer sequence were added to create sites for restriction enzymes NcoI and BamHI, respectively. In addition, the forward primer had the sequence for 10 His residues. The amplified MAP30_Hc ORF was digested with NcoI and BamHI and was directionally cloned, in frame, into an expression vector, pET3d (Novagen, Inc., Madison, WI), also digested with NcoI and BamHI, to create a recombinant plasmid, pET3d-10×His-HcMAP30.

Expression and purification of the recombinant protein.

E. coli BL21 Star (DE3)pLysS (Invitrogen) transformed with plasmid pET3d-10×His-HcMAP30 produced an H. cinaedi-specific recombinant 10-His-tagged fusion protein, His-MAP30_Hc, which was chosen for further study. Proteins expressed in the pET system were induced by IPTG (isopropyl-β-d-thiogalactopyranoside) at 37°C, and the recombinant fusion protein produced was purified via Ni²⁺ affinity column chromatography with use of Ni-nitrilotriacetic acid agarose (Qiagen). The protein eluted from Ni-nitrilotriacetic acid agarose was dialyzed overnight at 4°C against PBS through the 5,000-Da cutoff membrane, and protein content was determined via the bicinchoninic acid protein assay kit (Pierce Biotechnology Inc., Rockford, IL).

Western blot analysis.

Whole-cell lysates (6.5 μg of protein/lane) of H. cinaedi and its related bacterial species or E. coli transformants harbored plasmids for the H. cinaedi genomic library, and the purified antigenic fusion protein (His-MAP30_Hc) (0.6 μg/lane) was subjected to sodium dodecyl sulfate-15% polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions. After SDS-PAGE, proteins were electrophoretically transferred to Immobilon-P membranes (Millipore Corporation, Bedford, MA) and blocked with T-TBS-skim milk. The membranes were then incubated with various serum specimens (diluted from 1:1,000 to 1:625,000 in T-TBS-skim milk) and an antibody: sera were anti-H. cinaedi rabbit serum and, as a control, its preimmune serum, along with human serum from different groups, including H. cinaedi-infected patients, and the antibody was anti-MAP30_Hc antibody. The immunoreacted membranes were washed three times with T-TBS and were then incubated further with HRP-conjugated anti-rabbit IgG antibody (Santa Cruz Biotechnology) diluted 1:3,000 to 1:10,000 or HRP-conjugated anti-human IgG antibody (Sigma-Aldrich Corporation, St. Louis, MO) in T-TBS-skim milk. Immunoreactive protein bands were visualized by using an ECL Plus detection system (GE Healthcare) with a luminescent image analyzer (LAS1000UV mini; Fujifilm, Tokyo, Japan). In some experiments, to compare the immunoreactivities of protein bands under a consistent immunoblotting condition throughout the study, a membrane after transfer of the same protein preparation (e.g., bacterial cell lysate and recombinant protein) was cut into respective membrane strips for each lane of the electrophoresis gel, followed by incubation with different serum samples. After further reaction with the HRP-labeled secondary antibody, the membrane strips were sorted again into each group of H. cinaedi-infected and other control sera and were then subjected to the image analyzer to visualize the immunoreactive bands on the electrophoretic lanes under the same condition. Fujifilm Multi Gauge V3.0 image analysis software served for analysis of band intensity on Western blots. For the reprobing study, each membrane was incubated in a WB stripping solution (Nacalai Tesque, Inc., Kyoto, Japan) for 15 min at room temperature, followed by Western blotting as just described.

Immunoprecipitation.

H. cinaedi whole-cell lysate was prepared by sonicating bacterial cells in lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, and 1% Triton X-100) containing a protease inhibitor cocktail (Roche) and 1 mM phenylmethylsulfonyl fluoride. After centrifugation at 12,000 × g for 10 min, lysate supernatant was pretreated with protein A/G PLUS-agarose (Santa Cruz Biotechnology) for 30 min at 4°C to eliminate nonspecific binding to the agarose gel. Then, the lysate supernatant (1 ml; 65 μg/ml) was incubated with 5 μg of rabbit normal IgG (Santa Cruz Biotechnology) as a control or with 5 μg of anti-His-MAP30_Hc antibody for 5 h at 4°C, and immune complexes were precipitated by the addition of protein A/G PLUS-agarose for 2 h. The anti-His-MAP30_Hc antibody was purified by use of Sepharose CL-4B (GE Healthcare) chromatography and His-MAP30_Hc affinity column (HiTrap NHS-activated HP column; GE Healthcare) chromatography. Nonspecific binding was removed from the immunoprecipitated protein A/G PLUS-agarose by four washes with lysis buffer. The immunoprecipitated proteins that were isolated were then analyzed via SDS-PAGE and Western blotting with various serum samples obtained from patients.

ELISA for determination of the immune response to H. cinaedi.

Recombinant His-MAP30_Hc was used as an H. cinaedi antigen for ELISA. Each well of a 96-well microtiter plate was coated with 100 μl of His-MAP30_Hc (0.2 μg of protein/well) in 0.1 M carbonate buffer (pH 9.6), blocked with 1% bovine serum albumin, and washed three times with PBS containing 0.05% Tween 20 (washing buffer). Samples in the wells (human serum diluted 200-fold) were incubated for 1 h at 37°C. Wells were then washed with washing buffer and were reacted with HRP-conjugated anti-human IgG antibody (Sigma-Aldrich), followed by reaction with 1,2-phenylenediamine dihydrochloride. The reaction was terminated by the addition of 50 μl of 2.0 M sulfuric acid, and absorbance readings at 490 nm were obtained by means of a micro-ELISA plate reader. To examine the host immune response to H. cinaedi infection, serum levels of antibody against His-MAP30_Hc in H. cinaedi-infected and infection-suspected groups, and four different control groups, i.e., H. pylori-infected subjects, age- and sex-matched controls without apparent H. cinaedi infection, infant subjects younger than 1 year old, and volunteers, were compared.

Statistical analysis.

Statistical significance between groups was determined via the two-tailed unpaired Student t test. A P value of <0.05 was considered statistically significant.

Nucleotide sequence accession number.

The sequence data for MAP30_Hc gene determined in this work were deposited in the DDBJ/GenBank/EMBL databases under accession number AB365645.

RESULTS

Antibody responses to H. cinaedi whole-cell antigens in patients with bacteremia.

Production of antibodies against H. cinaedi during infections was examined by use of Western blotting. As illustrated in Fig. 1A, among eight patients, five had sera showing a marked immunoreaction with the 30-kDa antigen (lanes 1, 2, 4, 5, and 6), while three showed a moderate or weak reaction (lanes 3, 7, and 8). The times of serum sampling are 249 days, 226 days, 365 days, 120 days, 47 days, 20 days, 18 days, and 7 days after clinical onsets for each patient (lanes 1 to 8, respectively). The intensity of immunoreactive bands therefore seems to correlate with the timing of serum sampling for each patient. For example, much strong immunoreaction was observed with lanes 4 to 6, in which the serum was obtained around a few months after infection. Relatively strong immunoreactive bands with higher molecular masses, ranging from 50 to 100 kDa, were also found. The antibody response for these large antigens appeared to vary among subjects, so a consistent trend in the immune response was not observed. An important finding was the similarity of the immunoreactive pattern for the anti-H. cinaedi rabbit serum (leftmost Western blotting lane in Fig. 1A) and the most prominent reaction (lane 5 in Fig. 1A) to the 30-kDa antigen of the cell lysate proteins.

However, control serum from human subjects (the H. pylori infected, the age and sex matched, and volunteers in the laboratory) showed no appreciable or very weak immunoreactivity with the 30-kDa antigen; weak to moderate reaction was observed for 3 of 33 control subjects (Fig. 1B to D).

To verify that the 30-kDa protein is indeed a dominant antigen associated with these bacterial cells, Western blotting was performed with whole-cell lysates serially diluted with anti-H. cinaedi rabbit serum (Fig. 2A) and with anti-H. cinaedi rabbit serum serially diluted against a certain amount of cell lysates (data not shown). Densitometric analysis of immunoreactive protein bands on Western blots clearly indicated that this 30-kDa protein produced the strongest immune response of the proteins (Fig. 2B). We thus focused our work on identification and characterization of the 30-kDa H. cinaedi major antigenic protein.

FIG. 2.

Western blotting assessment of immunoreactivity of H. cinaedi proteins. (A) Serially diluted (dilution, 1-, 5-, 25-, and 125-fold) H. cinaedi whole-cell lysates (left; 1.3 μg of protein) were electrophoresed and reacted with anti-H. cinaedi rabbit serum (1:1,000 dilution). Molecular masses are given at the right in kilodaltons. (B) Densitometric analysis of the intensity of four immunoreactive bands shown in panel A. The 30-kDa protein (P30) was most dominant. P60, P65, and P90 indicate: 60-kDa, 65-kDa, and 90-kDa proteins, respectively.

Identification of antigenic proteins of H. cinaedi.

Several clones encoding immunoreactive proteins were isolated via immunoscreening of colonies of E. coli transformants expressing H. cinaedi genomic DNA with use of anti-H. cinaedi rabbit serum. These clones were shown to express H. cinaedi proteins of different sizes, i.e., of 29, 30, 45, 50, and 80 kDa (Fig. 3). Almost the same immunoreactive profiles were observed with serum from an H. cinaedi-infected patient (Fig. 3). Clone 9-5, which contained the corresponding plasmid p9-5, produced a protein with the same electrophoretic mobility as MAP30_Hc and showed strong immunoreactivity with both rabbit and human anti-H. cinaedi serum samples, as evidenced by Western blotting (Fig. 3). A protein produced by clone 8-2 showed a weak response to human anti-H. cinaedi serum even though it strongly responded to rabbit antiserum and was different from the protein produced by clone 9-5. We therefore chose the plasmid 9-5 for further characterization of MAP30_Hc.

FIG. 3.

Western blot analysis of proteins encoded by plasmid clones obtained by immunoscreening via anti-H. cinaedi rabbit serum. Cell lysates (6.5 μg of protein/lane) of E. coli clones or H. cinaedi were subjected to Western blotting using anti-H. cinaedi rabbit serum (1:10,000) and serum from an H. cinaedi-infected patient (1:1,500). Several immunoreactive proteins of different sizes were observed (arrowheads). A protein produced by clone 9-5 was almost the same size as MAP30_Hc and reacted strongly with both H. cinaedi-immunized rabbit serum and H. cinaedi-infected patient serum. Vector indicates a negative control E. coli clone that had an empty vector. Molecular masses are given at the right in kilodaltons.

Sequencing and characterization of the MAP30_Hc gene.

The DNA sequence of clone 9-5 was determined, and ORFs therein were predicted. One of the ORFs, which encodes the approximately 30-kDa protein, was presumed to be the MAP30_Hc gene. The MAP30_Hc gene (GenBank accession no. AB365645) had an ORF of 822 nucleotides, a putative ribosome-binding site (24), a TAA stop codon, and a G+C content of 40%. The ORF showed an identity of 74% in the nucleotide sequence with an ORF of the H. hepaticus gene encoding a 29-kDa conserved hypothetical protein (GenBank accession no. AE017125). The predicted translated protein of the MAP30_Hc ORF contained 273 amino acids and had relatively high homology with two other bacterial proteins, according to the NCBI BLAST search (Fig. 4). For example, as just mentioned for the nucleotide sequence homology, the predicted MAP30_Hc protein had an identity of 77% with a 29-kDa conserved hypothetical protein of H. hepaticus (GenBank accession no. AAP77310, named 0713_Hh). It also had a 57% identity with a 30-kDa outer membrane protein of H. pylori (GenBank accession no. AAD08604, named 1564_Hp). It is interesting that the NCBI conserved domain search showed that the MAP30_Hc domain was similar to the domain related to ABC-type metal ion transport systems that has wide distribution among bacteria.

FIG. 4.

Comparison of predicted amino acid sequences between MAP30_Hc and other helicobacter proteins. The areas enclosed by squares show the perfect matching of the three sequences. HH0713, H. hepaticus protein 0713_Hh (GenBank accession number AAP77310); HP1564, H. pylori protein 1564_Hp (GenBank accession number AAD08604); HcMAP30, MAP30_Hc. The numbers shown at the right of each alignment indicate the amino acid numbers from the N termini of the proteins.

In fact, the N-terminal 18-amino-acid peptide fragment of MAP30_Hc was presumed to be a probable membrane lipoprotein signal sequence on the basis of the database LipoP, used to predict lipoprotein signal peptides in gram-negative bacteria (http://www.cbs.dtu.dk/services/LipoP/). Searching the same database suggested that the amino acids at positions 17 and 18 serve as a putative cleavage site for a specific lipoprotein signal peptidase, signal peptidase II (9, 23), and that the Cys at position 18 may be a site of attachment to a membrane lipid (9). It is thus highly plausible that MAP30_Hc is a membrane protein related to the bacterial ABC-type metal ion transport system. As described above, Fig. 4 compares predicted amino acid sequences of the MAP30_Hc protein with H. hepaticus protein 0713_Hh and H. pylori protein 1564_Hp, illustrating a relatively high degree of homology.

Characterization of recombinant MAP30_Hc as a major antigen of H. cinaedi.

The MAP30_Hc ORF was amplified by PCR, and amplicons were directionally cloned in frame into vector pET3d to obtain plasmid pET3d-10×His-HcMAP30 for use in the recombinant protein expression system. His-MAP30_Hc recombinant protein expressed in and purified from E. coli served as an antigen for analysis of the immunoreaction of antibodies in rabbit and human anti-H. cinaedi sera (Fig. 5A and B). The 31-kDa His-MAP30_Hc fusion protein strongly reacted not only with serum from an H. cinaedi-immunized rabbit but also with serum from an H. cinaedi-infected patient. A few immunoreactive bands of relatively high molecular mass were found for serum samples from rabbits immunized and patients infected with H. cinaedi that were reacted with whole-cell lysates of H. pylori, H. hepaticus, and C. jejuni (Fig. 5A and B). However, protein bands similar to those for MAP30_Hc were not observed for other bacteria tested, except for the weak signal of an approximately 29-kDa protein detected for H. hepaticus.

FIG. 5.

Western blot analysis, after electrophoresis, showing reactivity of sera from an H. cinaedi-immunized rabbit (1:15,000) (A), an H. cinaedi-infected patient (1:1,500) (B), and a His-MAP30_Hc-immunized rabbit (1:15,000) (C) to His-MAP30_Hc (0.6 μg) and whole-cell lysates (6.5 μg) of four bacterial species (H. cinaedi, H. pylori, H. hepaticus, and C. jejuni). A membrane blotted in panel A was stripped and then used for blotting in panel C. Black arrowheads and open arrowheads in panels A to C indicate MAP30_Hc and His-MAP30_Hc, respectively. SDS-PAGE of recombinant His-MAP30_Hc (Coomassie brilliant blue [CBB] staining) is illustrated at the left of panel A. (D) H. cinaedi whole-cell lysate (6.5 μg) was incubated with anti-His-MAP30_Hc antibody or normal rabbit IgG (0.5 μg) for 5 h at 4°C. Proteins recognized by antibodies were precipitated with protein A/G PLUS-agarose beads and then applied to Western blots with H. cinaedi-infected patient serum. A protein immunoprecipitated with anti-His-MAP30_Hc antibody reacted with the H. cinaedi-infected patient serum (arrowhead) and was the same size as MAP30_Hc in H. cinaedi whole-cell lysate (6.5 μg/lane) (right). Molecular masses are given at the left in kilodaltons. HcMAP30, MAP30_Hc.

A slight difference in electrophoretic mobilities of MAP30_Hc immunoreactive bands was noted for the recombinant His-MAP30_Hc and the H. cinaedi whole-cell lysate, which was most likely due to the His repeat peptides added to the fusion protein. That the recombinant MAP30_Hc protein and the naturally occurring 30-kDa major antigen in H. cinaedi were identical was confirmed by a reprobing study (Fig. 5C) via a specific anti-MAP30_Hc antibody for the membrane immunoblotted with anti-H. cinaedi serum shown in Fig. 5A. An immunoprecipitation study of MAP30_Hc with H. cinaedi lysates, as illustrated in Fig. 5D, further elucidated these findings. An immunoprecipitated protein recovered from H. cinaedi lysate by means of anti-His-MAP30_Hc antibody was clearly identified via serum from H. cinaedi-infected patients, which was consistent with the immunoreactive protein band evident on the Western blot for the H. cinaedi lysate.

Furthermore, Fig. 6 provides results of Western blotting analysis with recombinant MAP30_Hc, which revealed immunoreactive profiles of MAP30_Hc that were in fact similar to those of H. cinaedi whole-cell lysate antigens in terms of reactivity with human serum samples, as shown in Fig. 1. Specifically, serum from six of eight patients produced moderate to strong signals, two samples reacted weakly, and almost all controls had no significant reactivity, or rarely just a weak to moderate reaction.

FIG. 6.

Western blot analysis of responses to recombinant His-MAP30_Hc in serum from H. cinaedi-infected patients (A) and controls (B to D). Sample (lane) numbers above the image correlate with those in the respective groups in Fig. 1. Each lane contained 0.6 μg of His-MAP30_Hc. All serum samples were diluted 1:1,500. Responses for the H. cinaedi-infected patients (A) and controls (B to D) were similar to those for H. cinaedi whole-cell lysate (Fig. 1). With a few exceptions, e.g., lane 1 of the H. pylori-infected group (B), lane 8 of the age- and sex-matched control group (C), and lane 5 of the healthy volunteer group (D), control sera did not react to this protein (only a few representative lanes for control negative reactions are shown in panels B to D). The arrowhead indicates His-MAP30_Hc. Molecular masses are given at the left in kilodaltons.

Recombinant protein-based ELISA.

His-MAP30_Hc recombinant protein was used as an antigen for ELISA to detect anti-H. cinaedi-specific antibodies in serum from infected patients. The specific host immune response to His-MAP30_Hc was examined by determining antibody titers. Figure 7 demonstrates that patients' antibody titers were significantly higher (P < 0.05) than those of all control groups. The levels of ELISA titers correlated well with the intensities of immunoreactive bands illustrated in Fig. 1 and 6.

FIG. 7.

Recombinant His-MAP30_Hc-based ELISA results. Serum antibody levels of H. cinaedi-infected patients (n = 8) were compared with those of control groups: H. pylori-infected patients (n = 10), controls 1 (age- and sex-matched subjects) (n = 9), controls 2 (healthy volunteers) (n = 16), and infants (n = 5). Numbers for the H. cinaedi-infected patients and control groups correlate with the numbers of respective groups in Fig. 1 and 6. Black circles indicate data for individuals in the patient and control groups; open boxes indicate averages, with standard deviations indicated. *, P < 0.05.

More importantly, as shown in Fig. 8, we also examined the time course of antibody responses for four patients (no. 1 to 4) for whom infection was strongly suspected because of typical clinical symptoms of H. cinaedi infection but from whom this bacterial species was not isolated. One patient (a 71-year-old woman) showed a drastic elevation of antibody titer, and three patients showed a slow but substantial rise in the response to His-MAP30_Hc antigen (Fig. 8). Similarly, the antibody titers of patient 5 (an 82-year-old woman), who was truly infected with H. cinaedi, showed a clear increase in antibody production after H. cinaedi infection.

DISCUSSION

In our previous report, we suggested that H. cinaedi infection may be more widespread in the general world community than previously believed (13). Optimal conditions for isolation of this species from clinical specimens have not been established, and specific conventional approaches such as serological assays are required to better define the pathogenic and epidemiological features of this emerging pathogen.

Our present study revealed that a 30-kDa immunogenic protein is the most potent of various H. cinaedi proteins for evoking an immune response in hosts. Flores et al. (5) described several immunogenic H. cinaedi antigens to serum from H. cinaedi-immunized rabbits and infected patients: they identified 96-, 64-, 59-, 47-, 41-, 29-, and 12.4-kDa immunoreactive bands via Western blotting. We believe that the 29-kDa antigen that they reported is identical to the 30-kDa antigen (MAP30_Hc) that we demonstrated in our current study. They also indicated that the 59- and 64-kDa antigens consistently exhibited the strongest reactions. In fact, we found relatively strong immunoreactivity for the 60- and 65-kDa proteins in our Western blot analysis (Fig. 1A and 2). However, because we observed similar bands with controls to some degree, these 60- and 65-kDa proteins may not be the major antigens functioning during H. cinaedi infections (Fig. 1B to D). Our dilution experiments with antigen and antisera confirmed that the 30-kDa antigen, identified as MAP30_Hc, had the strongest immune response among the different proteins of whole-cell H. cinaedi lysates (Fig. 2).

One of eight H. cinaedi-infected patients showed very little reaction to the 30-kDa protein (Fig. 1A, lane 7), and two had only a weak reaction (Fig. 1A, lanes 3 and 8), findings that correlated well with results of recombinant MAP30_Hc-based Western blotting (Fig. 6A, lanes 3, 7, and 8) and ELISA (Fig. 7). A similar correlation was observed for the immunoreaction of various control groups between Western blotting (Fig. 1B to D and 6B to D) and ELISA (Fig. 7). The variability in the host immune reactions to H. cinaedi antigens is thought to be due to differences in the time courses of infection and antibody development at the time of the serum sampling. In fact, the strongest immune response was observed with the serum obtained a few months after the onset of clinical symptoms of H. cinaedi infection, as evidenced by the Western blotting and ELISA data from lanes 4 to 6 (patients 4 to 6) in Fig. 1A, 6A, and 7. Such a time-dependent antibody production might be reasonably understood because of the well-known time profile of the host's immune response against various pathogens in general. Alternatively, it may reflect distinct individual immune responses to the specific antigen. As just discussed, in H. cinaedi-infected patients (seven women and one man), high antibody titers were seen with two women and one man (patients 4 to 6 in Fig. 7). The average age of all patients was 69.6 years, and that of three patients with high titers was 69.7 years. Thus, there is no particular tendency in age or gender for prominent immune response to MAP30_Hc. Details of the clinical characteristics of the present patients are described in our previous paper (13), which indicate again that no substantial difference was found in their clinical characteristics among the patients showing varied antibody titers. Similarly, we could not find any clear factors (e.g., age, gender, and clinical background) associated with a high antibody production for MAP30_Hc among the control groups tested in our current study.

This 30-kDa protein, which we identified here as a major immunogenic protein (MAP30_Hc), has high homology with a 29-kDa protein of H. hepaticus (0713_Hh; GenBank accession no. AAP77310). Although the function of this 29-kDa H. hepaticus protein is not clear, database information is available for a similar protein, an H. pylori putative outer membrane protein (1564_Hp; GenBank accession no. AAD08604). In addition, via an NCBI conserved domain search, we found a similarity to a domain of the ABC-type metal ion transport system, periplasmic component/surface antigen. Thus, it is highly plausible that MAP30_Hc is a membrane lipoprotein related to the bacterial ABC-type metal ion transport system.

Western blotting with H. hepaticus whole-cell lysates demonstrated that rabbit anti-H. cinaedi serum strongly reacted with antigens of high molecular mass (50 to 100 kDa) and weakly reacted with the 29-kDa antigen (Fig. 5A); rabbit anti-His-MAP30_Hc antibody showed a moderate reaction to the 29-kDa antigen of H. hepaticus (Fig. 5C). These results suggest that MAP30_Hc and the 29-kDa protein of H. hepaticus are antigenically cross-reactive. More-marked cross-reactions were observed between high-molecular-mass lysate proteins of H. cinaedi and those of other related bacteria such as H. hepaticus, H. pylori, and C. jejuni (Fig. 5A and B). Several investigators have suggested a possible association of H. hepaticus with human diseases (16, 31, 33). Indirect support for this idea comes from the detection of H. hepaticus 16S rRNA gene sequences in DNA extracted from colonoscopy samples obtained from children with inflammatory bowel disease (33) and from the finding that an H. hepaticus cell extract-based ELISA showed elevated antibody levels in patients with chronic liver diseases compared with controls (31). However, H. hepaticus is generally recognized as a pathogen that causes hepatitis in mice (6) rather than as a human pathogen: no confirmed evidence exists of isolation of H. hepaticus from human specimens, which would prove pathogenesis in humans. Moreover, the cross-reactivity between H. cinaedi and H. hepaticus antigens, as revealed in the present work, suggests that false-positive ELISA results for human infections may have been attributed to H. hepaticus but may have actually been caused by H. cinaedi. In contrast, human anti-H. cinaedi serum produced no measurable immunoreaction for the 29-kDa protein of H. hepaticus (Fig. 5B). Thus, the MAP30_Hc protein that we identified here may prove useful as a major antigen in serological assays for H. cinaedi, with high specificity in screening patients with various infections caused by Helicobacter and related bacteria.

Several studies have reported immunogenic proteins from Helicobacter or Campylobacter species and examined recombinant protein-based ELISAs (2-4, 11, 16, 29). The specificity of these ELISAs seems to be greater than that of ELISAs using whole-cell antigens. Similarly, the His-MAP30_Hc-based ELISA used here has high specificity, because the nonspecific ELISA titer obtained for control subjects was much lower than that in the whole-cell protein-based ELISA (13). Although the host antibody response seems to vary depending on the time of serum sampling (for H. cinaedi-infected patients, from 7 days to 1 year from the onset of clinical symptoms), the present MAP30_Hc-based ELISA revealed a significantly higher antibody titer for H. cinaedi-infected patients than for controls (P < 0.05) (Fig. 7). The relatively high antibody level detected for a few control subjects (Fig. 7) suggests the presence of subclinical H. cinaedi infection in the general population. All these ELISA patterns in H. cinaedi-infected and control groups were consistent with profiles of the MAP30_Hc immunoreactive band found via Western blotting, as described above (Fig. 1 and 6). We therefore believe that the ELISA using MAP30_Hc as an antigen may become a potent serological tool for screening humans for H. cinaedi infections.

The time course analysis of antibody responses in the group with suspected H. cinaedi infection and its true infection indicated that these patients had different degrees of increase in antibody titers (Fig. 8). One explanation for this variation may be different individual immune responses to specific antigens of this microorganism. Also, a determinant may be whether hosts had been previously immunized by the antigen during subclinical infections, as mentioned earlier for control subjects with relatively high antibody titers. Nevertheless, for typical clinical cases, H. cinaedi infection may be diagnosed by means of this serological procedure that utilizes the ELISA described here and does not involve bacterial culture.

In conclusion, we identified in our current study a 30-kDa H. cinaedi putative membrane protein as a major antigen (MAP30_Hc) and successfully constructed the expression system of the MAP30_Hc His-tagged fusion recombinant protein. This recombinant protein has great potential for use in serological assays, including Western blotting and ELISA, for clinical diagnosis of H. cinaedi infection. More important, the present identification of MAP30_Hc not only aids better understanding of the etiology of H. cinaedi infections but also provides evidence to warrant further epidemiological study of these infections in various hosts in nature.