Characterization of an Immunogenic Outer Membrane Autotransporter Protein, Arp, of Bartonella henselae

Abstract

Bartonella henselae is a recently recognized pathogenic bacterium associated with cat scratch disease, bacillary angiomatosis, and bacillary peliosis. This study describes the cloning, sequencing, and characterization of an antigenic autotransporter gene from B. henselae. A cloned 6.0-kb BclI-EcoRI DNA fragment expresses a 120-kDa B. henselae protein immunoreactive with 21.2% of sera from patients positive for B. henselae immunoglobulin G antibodies by indirect immunofluorescence, with 97.3% specificity and no cross-reactivity with antibodies against various other organisms. DNA sequencing of the clone revealed one open reading frame of 4,320 bp with a deduced amino acid sequence that shows homology to the family of autotransporters. The autotransporters are a group of proteins that mediate their own export through the outer membrane and consist of a passenger region, the α-domain, and an outer membrane transporter region, the β-domain. The passenger domain shows homology to a family of pertactin-like adhesion proteins and contains seven, nearly identical 48-amino-acid repeats not found in any other bacterial or Bartonella DNA sequences. The passenger α-domain has a calculated molecular mass of 117 kDa, and the transporter β-domain has a calculated molecular mass of 36 kDa. The clone expresses a 120-kDa protein and a protein that migrates at approximately 38 kDa exclusively in the outer membrane protein fraction, suggesting that the 120-kDa passenger protein remains associated with the outer membrane after cleavage from the 36-kDa transporter.

Bartonella henselae is a recently recognized extremely fastidious, gram-negative rod that has been found to be associated with cat scratch disease (CSD), bacillary angiomatosis, and bacillary peliosis (10, 12, 25, 37, 43). CSD is the most common syndrome caused by B. henselae and affects an estimated 25,000 people in the United States annually, many of whom are children (23). Severe and prolonged cases of CSD often require extensive clinical and laboratory workups, including lymph node biopsy. Serological tests specific for B. henselae may help in avoiding invasive surgical procedures in many of these patients.

Current immunologic assays for determining the presence of Bartonella antibodies include the indirect immunofluorescence assay (IFA) and enzyme immunoassay (EIA). A wide range of sensitivities have been reported for the IFA from 32 (6) to 100% (39) in various studies. Recently, an EIA based on a recombinant B. henselae 17-kDa antigen showed 71.1% sensitivity and 93.0% specificity compared to the IFA (34). A high degree of cross-reactivity of Bartonella serology with other intracellular pathogens such as Mycoplasma, Chlamydia, and the Q fever pathogen Coxiella burnetii has been described and has made the antibody-based diagnosis of CSD problematic (26, 35, 39). Therefore, characterization and cloning of the individual antigens involved in eliciting an immune response specific for B. henselae in humans should aid in the development of more sensitive and specific diagnostic serologic tests.

Not much is known about the immunoreactive proteins associated with B. henselae infections, especially in humans. Only a few immunoreactive B. henselae antigens important in the human humoral response have been cloned and characterized: the 17-kDa antigen (3), the HtrA stress response protein (2), and the 45-kDa SucB protein (17, 32). A number of other proteins, however, have been found to be important in the feline humoral response to B. henselae, including outer membrane proteins OMP89, GroEL, IalB, and OMP43 (9) and the p26 protein (41).

In this article, we describe the cloning and characterization of a gene encoding an immunoreactive protein of B. henselae, Arp (for acidic repeat protein), homologous to a family of autotransporter proteins. Autotransporters are a diverse family of proteins that facilitate their own secretion through the outer membrane by means of an outer membrane protein pore. They contain an N-terminal passenger region, the α-domain, and a C-terminal transporter region, the β-domain. Specifically, our study characterizes the immunoreactivity of the secreted passenger protein (α-domain) of Arp in patients with serologic evidence of B. henselae infection.

MATERIALS AND METHODS

Bacterial strains, growth conditions, and media.

Escherichia coli strains XL1-Blue MR (Stratagene, La Jolla, CA), DH5α, and BL21(DE3) (Invitrogen, Carlsbad, CA) were grown at 37°C in Luria-Bertani (LB) broth or on LB agar supplemented, where appropriate, with kanamycin (50 μg/ml). B. henselae was grown on freshly prepared Columbia 5% sheep blood agar plates for 7 days at 35°C under 5% carbon dioxide. All strains were maintained at −70°C in LB media containing 15% glycerol. Characteristics of the plasmids used in this study are diagramed in Fig. 1. Construction of these plasmids is described in the sections below.

FIG. 1.

(A) Restriction map of the region near the arp gene and positions of clones. The arrow represents the arp gene, with the arrowhead showing the direction of transcription. The numbers indicate the amino acid sequence positions. The diagonal lines in the arrowhead represent the area of homology with the β-domain region of proteins in the family of autotransporters. The gray area in the stem of the arrow represents a pertactin-like passenger region predicted by a search of the Pfam database. The vertical bars represent seven tandem nearly identical acidic amino acid repeats. (B) Homology of Arp with other Bartonella autotransporter proteins and various microbial antigens. The black, gray, and patterned bars represent regions of homology within the Arp protein with Bartonella autotransporter proteins and microbial antigens of P. falciparum and U. urealyticum. (C) Sequence of the 48-amino-acid repeat. The boldface letters joined by a slash indicate the variant amino acids. The seventh repeat does not contain the last serine and proline of this 48-amino-acid sequence.

Construction of B. henselae cosmid library.

A cosmid library of B. henselae ATCC 49793 was constructed using a SuperCos1 cosmid vector kit (Stratagene) as previously described (32).

Clinical samples.

One hundred forty-eight serum samples positive for B. henselae immunoglobulin G (IgG) antibodies by IFA (titers of 1:256 to 1:4,096) but negative for IgM antibodies by IFA, 111 samples positive for both B. henselae IgM and IgG antibodies by IFA (titers of 1:64 to 1:4,096 for IgG and 1:16 to 1:256 for IgM), 50 samples negative for both IgM and IgG (titers of <1:16 for IgM and <1:64 for IgG), and samples from 100 random healthy donors in the Salt Lake City, UT, area were used to determine sensitivity and specificity compared to IFA. The procedures followed were in accordance with the ethical standards established by the University of Utah and are in accordance with the Helsinki Declaration of 1975 (42). Specimens were collected under approval by the University of Utah Institutional Review Board (IRB#11343). Specimens were stored at −70°C until testing commenced and were then stored at 2 to 8°C while the evaluations were performed.

B. henselae IFA.

B. henselae-coated slides were obtained from Focus Technologies. Testing was performed according to the manufacturer's recommendation. Patient samples were considered equivocal at a dilution of 1:64 to 1:128; titers of 1:256 and above were considered positive.

Sera used for cross-reactivity analysis.

A panel of sera from patients with IgG antibodies against the following were used in cross-reactivity studies: Treponema pallidum (titer, ≥1:5), Coxiella burnetii (titer, ≥1:16), herpes simplex virus types 1 and 2 (an index value [IV] of ≥1.1 is positive), Brucella melitensis (IV, ≥1.1), Chlamydia pneumoniae (titer, ≥1:64), Mycoplasma pneumoniae (≥0.95 U/liter), Ehrlichia chaffeensis (titer, ≥1:16), Francisella tularensis (titer, ≥1:80), Rickettsia typhi (titer, ≥1:256), Rickettsia rickettsii (IV, ≥1.1), Legionella pneumophila (titer, ≥1:128), and Bordetella pertussis (IV, ≥1.1). The B. melitensis, M. pneumoniae, F. tularensis, C. burnetii, and R. typhi samples were previously tested and known to cross-react with the previously characterized SucB immunogenic protein of B. henselae (17, 32).

IFA was used for serum testing for antibodies against T. pallidum (SCIMEDX Corporation, Denville, NJ), C. burnetii (Focus Technologies, Cypress, CA), E. chaffeensis (PANBIO, Inc., Columbia, MD), R. typhi (Focus Technologies), and L. pneumophila (MARDX, Inc., Carlsbad, CA). An enzyme-linked immunosorbent assay (ELISA) was used for serum testing for antibodies to B. melitensis (PANBIO, Inc.), M. pneumoniae (GenBio, San Diego, CA), R. rickettsii (PANBIO, Inc.), herpes simplex virus types 1 and 2 (Focus Technologies), and B. pertussis (PANBIO, Inc.). A microimmunofluorescent assay (Focus Technologies) was used for C. pneumoniae, and agglutination (Germaine Laboratories, Inc., San Antonio, TX) was used for F. tularensis. A single positive representative sample was used for each organism. All tests were performed according to the manufacturers' recommendation.

Preparation and analysis of outer membrane, cytoplasmic, and secreted proteins and whole-cell proteins of pCML78.

Enriched outer membrane proteins and cytoplasmic proteins were prepared by modification of the procedure by Hantke (20). Although pCML78 contains a putative promoter region, expression was greatly enhanced with isopropyl-β-d-thiogalactopyranoside (IPTG) induction. After exponentially grown cells were exposed to 1 mM IPTG for 1 h, the cells were pelleted for 10 min at 4°C at 10,000 rpm (8,000 × g) in a Beckman JA-20 rotor in a 50-ml tube. The pellet was resuspended in 1 ml of 0.2 M Tris (pH 8.0) and transferred to a microcentrifuge tube. The bacteria were pelleted at 16,000 × g for 2 min at 4°C. After decanting of the supernatant, the pellets were resuspended in 50 μl of 0.2 M Tris (pH 8.0) on ice. In sequence on ice, 100 μl of 0.2 M Tris (pH 8.0), 1 M sucrose, 10 μl of 10 mM EDTA (pH 8.0), 10 μl of lysozyme (2 mg/ml), and 320 μl of distilled water were added with gentle mixing after each addition. The mixture was incubated for 10 min at room temperature. Then 10 μl of DNase (bovine pancreatic DNase I [Sigma] at 1 mg/ml in 0.15 M NaCl plus 50% glycerol) and 500 μl of a mixture of 2% Triton X-100, 10 mM MgCl₂, and 50 mM Tris (pH 8.0) were added. The solution was centrifuged at 16,000 × g for 30 min at 4°C. The supernatants (cytoplasmic proteins) were suspended in 2× Laemmli's sample buffer. After removal of the supernatants, the pellets were washed four times with ice-cold distilled water and then resuspended in 100 μl of 2× electrophoresis sample buffer.

For purification of supernatant (secreted proteins), 100-ml cultures containing kanamycin and 1 mM IPTG were incubated for 16 to 18 h at 37°C on a rotary shaker at 225 rpm. Culture supernatants were recovered by centrifugation and filtered. Ammonium sulfate was dissolved to a final concentration of 60% saturation. The precipitate was centrifuged and dissolved in 3 ml of 50 mM Tris-HCl (pH 7.5), and this preparation was dialyzed against 50 mM Tris-HCl (pH 7.5). This resulted in an approximately 10-fold concentration of the original supernatant that was suspended in 2× Laemmli's sample buffer.

For whole-cell protein preparations, exponentially grown cells were exposed to 1 mM IPTG for 1 h (optical density at 600 nm [OD₆₀₀] of approximately 1.0). Five milliliters of pelleted cells was washed with 1 ml 0.2 M Tris (pH 8.0) and centrifuged again. The pellets were resuspended in 100 μl of distilled water and 100 μl of 2× Laemmli's sample buffer.

The protein preparations were analyzed by separation on a 4 to 15% sodium dodecyl sulfate (SDS)-polyacrylamide gel (Ready Gel; Bio-Rad Laboratories, Inc., Hercules, CA) and stained with Coomassie blue.

Heat extraction.

We attempted to extract proteins from bacterial cells by using the procedure used by Charbonneau et al. (7) to extract cleaved mature AIDA-I autotransporter protein from E. coli. Overnight cultures from E. coli BL21(DE3) containing pCML78 induced with 1 mM IPTG after exponential growth, or control vector pET28b, were normalized to the same OD₆₀₀ in 40 ml of LB broth. Bacteria were harvested and resuspended in 375 μl of 10 mM sodium phosphate buffer (pH 7.0). In order to release the cleaved mature Arp, the samples were heated at 60°C for 20 min or resuspended on ice as a control. The samples were centrifuged for 5 min at 12,000 × g in microcentrifuge tubes. Ten-microliter aliquots from the supernatants were submitted to SDS-polyacrylamide gel electrophoresis (PAGE) and stained with Coomassie blue, as described above.

Liquid chromatography-tandem mass spectrometry.

For liquid chromatography-tandem mass spectrometry, the proteins of interest were resolved by SDS-PAGE, stained with Coomassie blue, and excised for analysis. Mass spectrometry was performed by the University of Utah Mass Spectrometry and Proteomics Core Facility using an LTQ-FT instrument (Thermo Electron, Bremen, Germany), which is a linear ion-trap/FTMS hybrid mass spectrometer. Proteins were identified by automated database searching (Matrix Science, London, United Kingdom) of all tandem mass spectrometry spectra using the NCBI nr database. The probability-based Mowse ion score is −10 · log(P), where P is the probability that the observed match is a random event. Ion scores greater than 33 are significant. Protein scores are derived from ion scores as a nonprobabalistic basis for ranking protein hits.

DNA manipulations and cloning.

Standard methods were followed for molecular biological techniques (38).

DNA sequencing.

The DNA sequence was determined by the AB 3700 96-capillary DNA analyzer from Applied Biosystems. Synthetic oligonucleotides used as primers for DNA sequencing were synthesized by the Huntsman Cancer Center DNA peptide facility, University of Utah.

Cloning and expression of Arp and portions of Arp.

Plasmid pCML78 was constructed by ligating gel-purified 6.0-kb BclI-EcoRI containing the entire arp gene and promoter region to BamHI/EcoRI-digested pET28b (Novagen, San Diego, CA) and transformed into E. coli DH5α. Plasmid pCML79 was constructed by ligating a gel-purified 3.2-kb BglII/SacI DNA fragment from pCML78 containing the entire passenger region into the BamHI/SacI site of pET28b and confirmed to translate in frame with the N-terminal histidine-tagged (HisTag)/thrombin/11-amino-acid fusion tag (T7Tag; Novagen) configuration in the pET28b expression vector.

Control plasmid pCML81 was constructed by ligating the same 3.2-kb BglII/SacI fragment into pET28c (subtracting 1 bp from pET28b), resulting in a frameshift so that the HisTag sequence is out of frame with the cloned protein sequence. Plasmid pCML80 was constructed by cutting pCML79 with SacI/SacII to delete the 1.2-kb DNA fragment encoding the portion of the passenger region containing a repeat region. The restriction enzyme overhangs were blunt ended with Klenow fragment and ligated, resulting in a HisTag fusion protein of the pertactin-like portion of the passenger region.

Plasmid pCML82 contains a 1.1-kb HindIII/SacI that encompasses the repeat region of arp (Fig. 1). Plasmid pCML82 was constructed by ligating a gel-purified 1.1-kb HindIII DNA fragment from pCML79 (containing the repeat region from HindIII to SacI of arp and the HindIII vector restriction site from pET28b) into the HindIII site of pET28b and confirmed to translate in frame with the N-terminal HisTag/thrombin/T7Tag configuration in the pET28b expression vector. All plasmids were transformed into E. coli strain BL21(DE3) for protein expression by heat shock according to the manufacturer's instructions.

Purification of the HisTag recombinant Arp proteins.

Cells of E. coli strain BL21(DE3) containing pCML79, pCML80, pCML82, or control plasmid pCML81 were grown overnight in 5 ml LB medium containing kanamycin. One milliliter of the culture was added to 100 ml of fresh LB broth containing kanamycin. At log phase, IPTG was added to a final concentration of 1 mM to induce expression of the protein for at least 2 h.

Cells from the 100-ml culture with a wet weight of about 200 mg were resuspended in 5 ml of BugBuster lysing reagent (Novagen), and inclusion bodies were purified according to the manufacturer's instructions. The inclusion bodies were further purified on a nickel HisBind affinity column (Novagen) according to the manufacturer's instructions. Eluted fractions were pooled and concentrated by ultrafiltration with a YM100 filter (Millipore, Bedford, MA).

Western blot.

Ten microliters of proteins (approximately 6.5 mg/ml) was loaded into each well of a 4 to 15% SDS-PAGE gel. Electrophoresis was carried out at 200 V for 36 min. The separated proteins were then transferred to a nitrocellulose membrane at 100 V for 1 h with the Mini-Trans-Blot electrophoretic transfer cell (Bio-Rad Laboratories). The membranes were then blocked overnight at 4°C with 3% dry milk in Tris-buffered saline (TBS) solution.

Serum samples were diluted 1:50 in a diluent/wash (3% dry milk in TBS) and were reacted with the membrane containing the protein samples for 3 h on a rocking platform. The membrane was then washed three times, changing the diluent wash every 5 min. A 1:10,000 dilution of alkaline phosphatase-conjugated goat anti-human IgG (γ-chain specific) (Sigma) was then added to the membranes. The membrane was then incubated for 1 h and washed as before, before the addition of 5-bromo-4-chloro-3-indolyl phosphate-nitroblue tetrazolium (BCIP/NBT) substrate (Sigma). This reaction was carried out for 5 to 10 min and then stopped by the addition of distilled water.

Serum activity.

Individual sera positive and negative for B. henselae antibodies by IFA, sera positive for antibodies to various other organisms listed above, and sera from healthy blood donors were analyzed for reactivity against the recombinant proteins. Recombinant proteins were resolved on a single preparative well of a 4 to 15% gradient SDS-PAGE minigel. Proteins were transferred to nitrocellulose, and the resulting membrane was cut into individual 3-mm strips and reacted with individual serum samples. Serum samples were diluted 1:50 in a diluent/wash (3% dry milk in TBS), and Western blot analysis was performed on the individual 3-mm strips by the Western blot procedure described above.

Statistical analyses.

The sensitivity, specificity, and positive predictive value of the outer membrane immunoblot compared with the B. henselae IFA were calculated with 95% confidence intervals (CI) using two-way contingency table analysis with Yates-corrected chi-square test (14).

Protein structure analysis.

Protein secondary structure analysis and antigenic index was carried out using the Garnier-Robson (16) and Jameson-Wolf (24) algorithms contained in a commercially available software package (Lasergene; DNAStar, Inc., Madison, WI).

DNA and protein database searches.

The National Center for Biotechnology Information Services was used to consult the SwissPROT, GenBank, and EMBL databases with the BLAST algorithm and conserved domain search (1, 19). Prediction of the signal peptide cleavage site was performed using the SignalP 3.0 Server at http://www.cbs.dtu.dk/services/SignalP/ (4).

Nucleotide sequence accession number.

The GenBank accession number for the nucleotide sequence of the 6.0-kb BclI-EcoRI DNA fragment presented in this article is EF540991.

RESULTS

Cloning of an immunogenic autotransporter from B. henselae ATCC 49793.

We previously used a pool of sera positive for antibodies to B. henselae to screen a B. henselae cosmid library for immunoreactive outer membrane proteins (32). One cosmid clone expressing an approximately 120-kDa antigen identified on a Western blot was sequenced and found to contain a 41-kb insert consisting of an uninterrupted sequence of the B. henselae genome (GenBank accession no. YP034040) from a portion of locus tag BH13030 (Also known as cfa, previously characterized by our group [31]; GenBank accession no. AY695890) to locus tag BH13210.

Sequence analysis of the arp gene.

After analyzing potential open reading frames (ORF) in the 41-kb insert that would theoretically express a 120-kDa protein, we focused on examining the protein expression of locus tag BH13120 (a hypothetical protein). A 6.0-kb BclI-EcoRI DNA fragment containing this locus tag was subcloned into a pET28b expression vector (pCML78). DNA sequencing of the 6.0-kb insert revealed one ORF. The coding sequence consists of 4,320 bp with a deduced amino acid sequence of 1,441 residues with a calculated molecular mass of 156 kDa. The nucleotide sequence is identical to the nucleotide sequence at locus tag BH13120 of the sequenced B. henselae Houston strain genome (GenBank accession no. YP034040). The gene was named arp. The ORF begins 793 bp downstream of a BclI restriction site. A putative Shine-Dalgarno sequence (nucleotides −13 to −10) and a promoter sequence (nucleotides −118 to −80) are identified upstream of the putative initiation codon. The predicted transcriptional start site is at nucleotide −70 (Fig. 2).

FIG. 2.

Partial nucleotide sequence of the B. henselae arp gene and its promoter region. The sequence starts 631 bases downstream of the BclI site and ends 87 bases within arp. The deduced amino acid sequence of the first 29 amino acids of B. henselae Arp is shown below the arp sequence. The approximate start site of transcription is indicated by an asterisk. The −35 region, −10 region, and Shine-Dalgarno sequence (SD) are underlined and labeled. A vertical arrow marks the signal peptidase cleavage site.

Homology of Arp with pertactin-like passenger proteins, AAA family of ATPases, and antigens of Plasmodium falciparum and Ureaplasma urealyticum.

A signal sequence of 26 amino acids is predicted using the Signal P program (Fig. 2). When the deduced amino acid was subjected to a Pfam, BLAST, and conserved domain search, a C-terminal 287-amino-acid β-barrel domain characteristic of an autotransporter protein was predicted (expect [E] value, 8e-14) (Fig. 1A). This homology suggests the protein has three main domains: a signal peptide, a secreted α-domain, and a transporting β-domain similar to other autotransporter proteins (22). A conserved domain search also predicts the presence of two other protein domains within the passenger region of the autotransporter, a pertactin-like passenger domain (E value, 0.008) and an AAA ATPase domain (MDN1) (COG database: 4,873 position-specific scoring matrices; E value, 0.006) (Fig. 1A). The presumed secreted α-domain has a calculated molecular weight of 116,880 and a calculated isoelectric point of 4.66. The predicted outer membrane transporting β-domain has a calculated molecular weight of 36,039 and an isoelectric point of 8.88.

The β-domains of autotransporters are conserved in secondary structure, with a high content in β-strands forming a transmembrane β-barrel. The β-barrel region is preceded by an N-terminal α-helix which is transported through the barrel pore (36). A search of the Arp protein for α- and β-regions using the Garnier-Robson algorithm also shows that the Arp β-domain has this β-strand secondary structure (residues 1168 to 1441) preceded by an α-helix (residues 1132 to 1167).

Recently, Dautin et al. (11) described a cleavage site within the α-helix of the β-domain (Asn-Asn). The same motif (residues 1142 to 1143 of Arp) is apparently present in Arp sequence within the α-helix region immediately preceding the β-strand secondary structure. This suggests that Arp may also have a similar processing mechanism. Although analysis of secondary structure of Arp reveals that the pertactin-like region appears to be prone to β-helix formation, the acidic repeat region was not predicted to form a β-helix. Using the Jameson-Wolf algorithm for antigenic profiles, the acidic repeat region showed a high antigenic index of 1.7.

The N-terminal sequence of the passenger region from the first amino acid to residue 780 shows a high degree of homology to a Bartonella quintana strain Toulouse putative autotransporter hypothetical protein (BQ10410), with 72% identity (percentage of identical amino acids) and 82% similarity (percentage of identical and conserved amino acids) (Fig. 1B). No other protein in the GenBank database shows as high a degree of homology with the N-terminal portion of the protein from residues 1 to 514.

However, the pertactin-like domain of Arp (residues 515 to 780 [Fig. 1B]) shows significant homology (low E values) to two other putative B. henselae autotransporters (hypothetical protein BH13160 and probable surface protein BH13140) and one other putative B. quintana autotransporter (hypothetical protein BQ10380), with a range of 44 to 54% identity (Fig. 1B). The β-domains of the two B. henselae proteins (BH13160 and BH13140) are nearly identical (96 to 97% identity) to the β-domain of Arp (Fig. 1B). The B. quintana autotransporters, BQ10410 and BQ10380, are similar to a lesser extent in the β-domain with 67 and 54% identity, respectively (Fig. 1B).

In the AAA ATPase domain, from residues 775 to 1108, there is a highly repetitive region, rich in acidic amino acids (Fig. 1A). There are seven, nearly identical, tandem glutamic acid-, threonine-, and serine-rich amino acid repeats comprised of 48 amino acids (Fig. 1C).

The acidic repeat region of Arp shows no significant homology to any other B. henselae or other Bartonella species proteins and appears to be a unique protein for B. henselae. A BLAST search shows only 31% identity and 51% similarity to a tryptophan/threonine-rich antigen of the parasite Plasmodium falciparum (GenBank accession no. AAK11835), and 30% identity and 39% similarity to the multiple-banded antigen of Ureaplasma urealyticum (GenBank accession no. AAT79411) (Fig. 1B).

Protein analysis of a secreted 117-kDa protein, a 155-kDa uncleaved protein, and the 36-kDa β-domain of the B. henselae arp gene expressed in E. coli.

We examined the expression of Arp in E. coli by SDS-PAGE in various cellular fractions to observe whether the predicted 117-kDa autotransporter passenger protein is secreted and whether the predicted 36-kDa β-domain is expressed in the outer membrane fraction in an E. coli background. Western blotting was also performed on the individual cellular fractions, to further examine whether the 117-kDa passenger protein of Arp is antigenic and recognized by sera from patients with B. henselae infection.

After induction with IPTG, concentrated culture supernatants, outer membrane proteins, cytoplasmic proteins, and whole-cell proteins were purified from E. coli containing the plasmid pCML78 (arp clone) and the pET28b cloning vector as a control. These proteins were subjected to SDS-PAGE and Western blot analysis (Fig. 3). Two new protein bands with apparent molecular masses of 120 kDa and approximately 38 kDa were observed in the outer membrane protein fraction and the whole-cell protein preparation of E. coli cells containing the arp clone (pCML78; Fig. 3A, lanes 4 and 8) and were not seen in the control lanes (Fig. 3A, lanes 3 and 7). The observed 120-kDa band is consistent with the molecular mass of the predicted mature secreted protein (α-domain) of Arp. A much fainter band is also identified at approximately 155 kDa in the outer membrane fraction of the E. coli cells containing the arp clone (Fig. 3A, lane 4). The size of the 155-kDa band is consistent with the molecular mass of the uncleaved protein. Mass spectrometry was performed on the 155-, 120-, and 38-kDa proteins and confirmed the identity of all of these proteins as Arp. The protein scores for the 155-, 120-, and 38-kDa proteins were very high at 1,093, 1,356, and 3,036, respectively, indicating a greater 95% likelihood of an identical match.

FIG. 3.

(A) SDS-PAGE of secreted, outer membrane, and cytoplasmic proteins and whole-cell protein preparations stained with Coomassie blue. Lane 1, 10× concentrated supernatant of E. coli containing pET28b; lane 2, 10× concentrated supernatant of E. coli containing pCML78 (entire arp clone); lane 3, outer membrane protein from E. coli containing pET28b; lane 4, outer membrane protein from E. coli containing pCML78; lane 5, cytoplasmic protein from E. coli containing pET28b; lane 6, cytoplasmic protein from E. coli containing pCML78; lane 7, whole-cell protein from E. coli containing pET28b; lane 8, whole-cell protein from E. coli containing pCML78. (B) Western blot transfer of panel A. The Western blot was probed with serum from a patient with suspected CSD (IgG IFA titer of 1:512 and IgM IFA titer of 1:32, both for B. henselae). The black arrows indicate the positions of the 120-kDa immunogenic passenger protein in the SDS-PAGE gel and Western blot. The smaller black arrow with an asterisk indicates the 155-kDa uncleaved protein. The white arrows indicate the positions of the 38-kDa transport protein in the SDS-PAGE gel. No corresponding 38-kDa band is identified on the Western blot. The numbers on the left indicate the positions of protein standards (in kDa).

Other groups have shown that the cleaved mature autotransporter AIDA-I remains associated with the outer membrane by noncovalent interactions and can be released by treatment at 60°C (5, 7). We performed heat extraction of the E. coli strain at 60°C with the same procedure previously used for AIDA-I (7). We could not demonstrate Arp in the supernatant of the heat-treated E. coli cells containing the arp clone.

The data suggest that the 117-kDa passenger region of the protein is cleaved from the 36-kDa portion of the transport protein as it exits the transport pore but remains associated with the outer membrane protein instead of being secreted into the media. The 36-kDa transport protein is associated with the outer membrane protein fraction, as would be predicted for a β-barrel pore protein.

Western blot analysis.

Western blot analysis showed that serum from a patient with CSD reacted with the 120-kDa passenger protein of Arp (Fig. 3B). No reactivity was seen against the 38-kDa protein. Antibodies with reactivity against other size bands are presumed to be naturally occurring antibodies against E. coli proteins.

The passenger region of Arp is antigenic.

To assess whether Arp is immunogenic in patients with suspected Bartonella infections, outer membrane preparations of the arp clone (pCML78) in DE3(BL21) were probed by immunoblotting with individual sera from patients with serologic evidence of CSD that were positive for B. henselae antibodies by IFA, as well as sera negative for B. henselae antibodies by IFA and sera from healthy normal control donors (Fig. 4A). Sera from 111 patients positive for both IgM and IgG antibodies by IFA, 148 patients positive for only IgG antibodies by IFA, 50 patients negative by IFA, and 100 healthy controls were individually tested. Three-millimeter-wide nitrocellulose strips containing transferred outer membrane preps of the arp clone were used to test the samples. By Western blotting, reactivity with Arp was seen with 26.2% of IgM and IgG IFA-positive sera (29 of 111 sera) and 17.3% of IgG-only IFA-positive sera (26 of 148 sera), for an overall sensitivity of 21.2% (95% CI, 19.2 to 22.2%). Two percent of IFA-negative sera (1 of 50 sera) and 3% of healthy control sera (3 of 100 sera) reacted with Arp, for an overall specificity of 97.3% (CI, 93.8 to 98.9%). The positive predictive value of the Western blot was 93.2% (CI, 84.4 to 97.3%). No reactivity with the 38-kDa protein was seen in these samples.

FIG. 4.

Western blot of an outer membrane preparation of E. coli cells containing the cloned arp gene reacted with individual serum samples. Outer membrane preparations expressing the induced 120- and 38-kDa proteins of arp cloned on pCML78 were electrophoresed through a 4 to 15% gradient SDS-PAGE gel and transferred to nitrocellulose. The nitrocellulose was cut into 3-mm strips and reacted with the individual serum samples diluted 1:50. (A) Individual serum samples positive and negative for B. henselae antibodies by IFA and healthy controls. Lanes: 1, IFA-positive titers for IgG and IgM of 1:64 and 1:16, respectively; 2, IFA-positive titers for IgG and IgM of 1:2,048 and 1:32, respectively; 3, IFA-positive titers for IgG and IgM of 1:256 and <1:16, respectively; 4, IFA-negative titers for IgG and IgM of <1:64 and <1:16, respectively; 5 and 6, healthy controls. (B) Study of cross-reactivity of B. henselae 120-kDa Arp with individual serum samples positive for antibodies against various infectious agents. Lanes: 1, B. henselae IgG IFA titer of 1:512; 2, B. henselae, IgG IFA titer of <1:64; 3, T. pallidum fluorescent treponemal antibody positive; 4, Q fever (C. burnetii), titer phase I and II titers of 1:16 and 1:32, respectively; 5, herpes simplex virus type 1 and 2 IV of 5.74 and 2.39, respectively; 6, B. melitensis titer of 1:1,280; 7, C. pneumoniae titer of >1:1,024; 8, M. pneumoniae IV of 3.67; 9, E. chaffeensis titer of 1:4,096; 10, F. tularensis titer of 1:1,280; 11, R. typhi titer of 1:256; 12, R. rickettsii IV of 10.5; lane 13, L. pneumophila titer of 1:128; 14, B. pertussis IV of 3.1. Arrows indicate the position of the 120-kDa Arp protein. Molecular mass standards (in kDa) are shown on the left.

There was no significant difference between the B. henselae IgM and IgG IFA titers for the positive samples reactive with the 120-kDa band (median IgM, 1:32; median IgG, 1:512) compared to the total of all IFA-positive samples, regardless of reactivity with the 120-kDa band.

Cross-reactivity studies.

Reactivity to the 120-kDa passenger protein of Arp was tested against sera from patients with antibodies against other bacterial species (Fig. 4B). No significant cross-reactivity was seen with sera positive for IgG antibodies to T. pallidum, C. burnetii, herpes simplex virus types 1 and 2, B. melitensis, C. pneumoniae, M. pneumoniae, E. chaffeensis, F. tularensis, R. typhi, R. rickettsii, L. pneumophila, and B. pertussis. However, the serum sample containing antibodies against M. pneumoniae appeared to react with a protein the same size as the 38-kDa protein (Fig. 4B, lane 8).

Cloning, expression, and purification of two separate regions of the Arp passenger protein, the pertactin-like region and the acidic repeat region.

To examine whether the pertactin-like region and/or the acidic repeat region was responsible for the immunogenicity of the Arp passenger protein, the coding regions of the entire passenger region (pCML79), the pertactin-like region (pCML80), and the acidic repeat region (pCML82) were directionally cloned into pET28b and confirmed to translate in frame with the N-terminal HisTag/thrombin/T7Tag configuration in the pET28b expression vector, inducible by IPTG. Control plasmid pCML81 was constructed by cloning the passenger region into pET28c (subtracting 1 bp from pET28b) resulting in a frameshift so that the HisTag sequence is out of frame with the cloned protein sequence.

After induction with IPTG, the soluble and insoluble (inclusion bodies) protein fractions of the E. coli strains containing the cloned HisTag proteins were examined for induced proteins. Most of the induced proteins were expressed exclusively as inclusion bodies for the HisTag fusion proteins of the entire passenger region (pCML79), the pertactin-like region (pCML80), and the acidic repeat region (pCML82). No induced protein was identified in the control plasmid (pCML81). The insoluble protein fractions were further purified using a nickel HisBind affinity column.

These HisTag recombinant proteins were subjected to SDS-PAGE and Western blot analysis (Fig. 5). Lanes 2 to 4 in Fig. 5A show major protein bands at approximately 150, 80, and 45 kDa, respectively. These major protein bands are consistent with the predicted molecular masses of the HisTag recombinant proteins of the entire passenger region (130 kDa), the pertactin-like region of the passenger region (81 kDa), and the repeat region (45 kDa). Minor lower protein bands are also identified in lanes 2 to 4 that may represent protease-cleaved proteins.

FIG. 5.

(A) SDS-PAGE of purified HisTag recombinant proteins of Arp. Lane 1, HisBind-purified protein from E. coli containing control plasmid pCML81; lane 2, 130-kDa HisTag passenger protein of Arp, purified from E. coli containing pCML79; lane 3, 80-kDa HisTag pertactin-like region of Arp, purified from E. coli containing pCML80; lane 4, 45-kDa HisTag repeat region of Arp, purified from E. coli containing pCML82. (B) Western blot transfer of panel A probed with a serum sample positive for IgM and IgG antibodies to B. henselae by IFA. The Western blot was probed with serum from a patient with suspected CSD (IgG and IgM IFA titers of 1:512 and 1:32, respectively, both for B. henselae). (C) Western blot transfer of panel A probed with a serum sample positive only for IgG antibodies to B. henselae by IFA. The Western blot was probed with serum from a patient with suspected CSD (IgG IFA titer of 1:1,024 and IgM IFA-negative titer of <1:16, both for B. henselae). Arrows indicate positions of the HisTag recombinant proteins by both SDS-PAGE and Western blotting. Molecular mass standards (in kDa) are shown on the left.

All 55 patient sera, 1 IFA-negative serum, and three healthy control sera, all originally reactive with the 120-kDa secreted passenger region, and 10 IFA-negative sera not reactive with the 120-kDa protein were individually reacted with the proteins in Fig 5A by Western blotting. Two major patterns of reactivity with the recombinant proteins were observed. Eleven of the 55 patient sera (20%) showed reactivity with the entire HisTag passenger region and the 45-kDa repeat region protein but no reactivity with the 81-kDa pertactin-like portion of the passenger region. Figure 5B shows a representative Western blot with the first pattern of reactivity. The remaining 44 samples (80%) showed reactivity with all three recombinant proteins (130-, 81-, and 45-kDa proteins). Figure 5C shows a representative Western blot with the second pattern of reactivity. This suggests that the repeat region may be somewhat more immunogenic than the pertactin-like region, since reactivity to Arp protein is restricted to the repeat region in 20% of the cases.

DISCUSSION

The β-domains of autotransporters are highly conserved and serve a common function to translocate the passenger protein to the bacterial cell surface where they may remain associated with the cell surface or released into the outside environment (22). The homology of C-terminal end of Arp with the conserved β-barrel pores of autotransporters predicts that one of the components of the Arp protein would be an outer membrane protein of approximately 36 kDa. Protein electrophoresis of the outer membrane protein preparations of the arp clone confirms the presence of an approximately 38-kDa protein (Fig. 3A, lane 4) not seen in the control lane (Fig. 3A, lane 3). An additional approximately 120-kDa mature protein and 155-kDa uncleaved protein were also identified in the outer membrane protein preparation of the arp clone not seen in the control lane. The data suggest that the passenger protein remains associated with the outer membrane after secretion through the β-barrel pore. Preparations of supernatant (secreted proteins) of the arp clone (Fig. 3A, lane 2) did not reveal any proteins near 120 kDa different from the control, suggesting that at least the majority of the protein remains associated with the bacterium and is not released into the environment.

In contrast to the conserved β-domains, the passenger α-domains of autotransporters are very divergent and many different functions have been described for these proteins (21). Numerous diverse virulence factors have been classified as autotransporter proteins, including adhesion proteins, toxins, proteases, lipases, and hemagglutinins (21). A search of the Pfam database with the amino acid sequence of the Arp passenger region predicts the presence of a pertactin-like passenger domain. The pertactin autotransporter proteins have been best characterized for the three Bordetella species, B. pertussis (8), B. parapertussis (30), and B. bronchiseptica (29), producing 69-, 70-, and 68-kDa proteins, respectively. The processed mature pertactin proteins (passenger domains) of these proteins remain noncovalently associated with the autotransporter β-domains on the outer membrane of the bacterium. The experimental evidence presented here suggests that the passenger protein Arp is associated with the outer membrane fraction of the bacterium, consistent with what has been seen with other pertactin proteins.

The difference in sizes between the different Bordetella species' pertactin proteins can be accounted for by the number of internal repeats contained within the passenger domain. However, the repeats of Arp within the passenger domain show no homology to the Bordetella pertactin proteins. In fact, the repeat region of Arp shows no significant homology to any other B. henselae or other Bartonella species' proteins and appears to be a unique protein for B. henselae. Recently, Gilmore et al. (18) cloned and characterized a Bartonella vinsonii subsp. arupensis gene encoding a 382-kDa immunodominant surface protein that is part of a gene family encoding large proteins, each containing multiple regions of repetitive segments. Though the Arp protein is not in the same family of BrpA proteins and not as large a protein, the role of repetitive segments in conferring antigenicity to these surface proteins could be an important subject of investigation for Bartonella species in general.

On the Western blot, the secreted α-domain of Arp was recognized by 21.2% of sera in patients with suspected CSD positive for B. henselae antibodies by IFA. A slightly higher percentage of positive sera (26.2%) were noted in patients with evidence of more recent infections, as defined by those with both IgM and IgG titers to B. henselae by IFA. When the reactivity of the sera was further examined by assaying the reactivity of the sera with the purified individual pertactin-like region and the repeat region, 20% of the positives were reactive with only the repeat region, not the pertactin-like region. The data suggest that overall both portions of the protein are immunogenic, though the repeat region may be a more important antigen in some patients. However, the apparent difference in reactivity may also be due to a conformational change in the HisTag pertactin-like recombinant protein that could change the epitopes recognized by the antibodies in the patient sera.

Although reactivity to Arp was seen in only 21.2% of sera from patients with suspected CSD, the specificity was very high at 97%. There was no cross-reactivity with a panel of sera that included antibodies with known cross-reactivity with the B. henselae IFA test (21, 29, 32). Therefore, Arp, in combination with other highly specific B. henselae antigens, could be a potential candidate for an immunodiagnostic test for the identification of individuals with B. henselae antibodies. In one of our previous studies, 14 of 25 sera from patients with CSD positive by IFA had antibodies against a 116-kDa B. henselae outer membrane protein by Western blotting (33). The discordance between percent positivity for the B. henselae outer membrane Western blot (56%) versus the purified expressed Arp (21%) may be due to additional antigenic proteins migrating at the same molecular weight in the B. henselae outer membrane preparation.

The functional role of the B. henselae Arp autotransporter remains to be established. Since virulence determinants are often either secreted to the bacterial cell surface or released into the external environment, Arp may be a candidate virulence factor. The pertactin molecule of Bordetella species has been shown to confer bacterial adhesion to CHO and HeLa cells (13, 27, 28) and may also be involved in the cytotoxicity of this organism for mononuclear phagocytic cells (15), possibly by promoting stable adhesion of the organism to the macrophages. The homology of Arp to pertactin-like proteins and evidence that it associates with the outer membrane proteins suggest that it may also be an adhesion protein. The repeat region shows similarity to the AAA family of ATPases and to several antigens of P. falciparum and U. urealyticum. Members of the AAA family of ATPases have diverse functions, such as membrane fusion, proteolysis, DNA replication and recombination, microtubule organization, and intracellular motility (40). The role of the repeat regions and the possible biologic functions of Arp, such as adherence to various cell types, will be some of the topics of future studies.