Identification of a Glycosylated Ehrlichia canis 19-Kilodalton Major Immunoreactive Protein with a Species-Specific Serine-Rich Glycopeptide Epitope

Abstract

Ehrlichia canis has a small subset of major immunoreactive proteins that includes a 19-kDa protein that elicits an early Ehrlichia-specific antibody response in infected dogs. We report herein the identification and molecular characterization of this highly conserved 19-kDa major immunoreactive glycoprotein (gp19) ortholog of the Ehrlichia chaffeensis variable-length PCR target (VLPT) protein. E. canis gp19 has substantial carboxyl-terminal amino acid homology (59%) with E. chaffeensis VLPT and the same chromosomal location; however, the E. chaffeensis VLPT gene (594 bp) has tandem repeats that are not present in the E. canis gp19 gene (414 bp). Consistent with other ehrlichial glycoproteins, the gp19 protein exhibited a larger-than-predicted mass (∼3 kDa), O-linked glycosylation sites were predicted in an amino-terminal serine/threonine/glutamate (STE)-rich patch (26 amino acids), carbohydrate was detected on the recombinant gp19 protein, and the neutral sugars glucose and galactose were detected on the recombinant amino-terminal polypeptide. E. canis gp19 composition consists of five predominant amino acids, cysteine, glutamate, tyrosine, serine, and threonine, concentrated in the STE-rich patch and a carboxyl-terminal domain predominated by cysteine and tyrosine (55%). The amino-terminal STE-rich patch contained a major species-specific antibody epitope strongly recognized by serum from an E. canis-infected dog. The recombinant glycopeptide epitope was substantially more reactive with antibody than the synthetic (nonglycosylated) peptide, and periodate treatment of the recombinant glycopeptide epitope reduced its immunoreactivity, demonstrating the importance of a carbohydrate immunodeterminant(s). The gp19 protein was present on reticulate and dense-cored cells, and it was found extracellularly in the fibrillar matrix and associated with the morula membrane, the host cell cytoplasm, and the nucleus.

Ehrlichia canis is a tick-transmitted obligately intracellular bacterium that causes moderate to severe and sometimes fatal disease in wild and domestic canids. The genomes of E. canis and other organisms in the genus, including Ehrlichia chaffeensis and Ehrlichia ruminantium, exhibit a high degree of genomic synteny, paralogous protein families, a large proportion of proteins with transmembrane helices and/or signal sequences, tandem repeats and ankyrin domains in proteins associated with host-pathogen interactions, and a unique serine-threonine bias associated with a potential for O glycosylation and phosphorylation (6, 10, 11, 18). A small subset of the approximately 1,000 proteins (including hypothetical proteins) encoded by each of these genomes is recognized by antibody (8, 20, 25, 32). Several of the major immunoreactive proteins identified and molecularly characterized are serine-rich glycoproteins that are secreted. Many of these glycoproteins have tandem repeats; however, one has numerous eukaryote-like ankyrin domains (7, 20, 25, 27, 31, 35).

Numerous proteins that contain tandem repeats have been identified in E. canis (n = 12) and E. chaffeensis and E. ruminantium (n = 31) strains (6, 18). Notably, three immunoreactive proteins with tandem repeats have been identified and molecularly characterized in E. chaffeensis (gp120, gp47, and variable-length PCR target [VLPT]) as well as two orthologs in E. canis (gp140 and gp36). The ortholog of the E. chaffeensis VLPT gene has not been identified in E. canis or other ehrlichial genomes (10). Extensive variability in the number and/or sequence of tandem repeats in the E. chaffeensis immunoreactive proteins (gp120, gp47, and VLPT) as well as E. canis gp36 is well documented (5, 7, 8, 32). The presence of tandem repeats in both coding and noncoding regions of the genome has been linked to an active process of expansion and reduction of ehrlichial genomes (11) and is considered a major source of genomic change and instability (4).

Although E. chaffeensis VLPT is immunoreactive, little is known regarding its cellular location, function, and role in the development of protective immunity. The E. chaffeensis VLPT gene exhibits variations in the number of 90-bp tandem repeats (three to five) and has been utilized as a molecular diagnostic target and for the differentiation of isolates (32, 33). The VLPT protein of E. chaffeensis Arkansas is a 198-amino-acid protein that has four repeats (30 amino acids) and has a molecular mass approximately double that predicted by its amino acid sequence (32). The E. chaffeensis VLPT protein appears to have a posttranslational modification consistent with those of other described ehrlichial glycoproteins, but the presence of carbohydrate on VLPT has not been demonstrated.

In this investigation, we report the identification and characterization of a conserved 19-kDa E. canis major immunoreactive glycoprotein (gp19), the ortholog of E. chaffeensis VLPT. The E. canis gp19 protein lacks tandem repeats of E. chaffeensis VLPT, but the two proteins exhibit substantial amino acid similarity (59%) in a cysteine/tyrosine-rich carboxyl-terminal domain, and both genes have the same relative chromosomal location. Carbohydrate was detected on the recombinant gp19 protein, and an antibody epitope-containing region was mapped to a serine/threonine/glutamate (STE)-rich patch. Antibody recognition of the epitope-containing region was reduced by periodate treatment, and the recombinant peptide was substantially more immunoreactive than the corresponding synthetic peptide, demonstrating the presence of a carbohydrate or carbohydrate-dependent immunodeterminant. The gp19 protein was found on both reticulate and dense-cored cells and was observed in the extracellular matrix and associated with the morula membrane.

MATERIALS AND METHODS

Culture and purification of ehrlichiae.

E. canis strains (Jake, DJ, Demon, Louisiana, Florida, and Sao Paulo strains) were propogated as previously described (21). Ehrlichiae were purified by size exclusion chromatography over Sephacryl S-1000 (Amersham Biosciences, Piscataway, NJ) as previously described (30). The fractions containing bacteria were frozen and utilized as antigen and DNA sources.

Construction and screening of the E. canis genomic library.

An E. canis Jake strain genomic library was constructed using an HpaII restriction digest and screened as previously described (21).

DNA sequencing.

Library inserts, plasmids, and PCR products were sequenced with an ABI Prism 377XL DNA sequencer (PerkinElmer Applied Biosystems, Foster City, CA) at the University of Texas Medical Branch Protein Chemistry Core Laboratory.

Glycoprotein analysis.

Nucleic acid and amino acid alignments were performed with MegAlign (Lasergene v5.08; DNAStar, Madison, WI). The E. canis gp19 and E. chaffeensis VLPT protein sequences were evaluated for potential O-linked glycosylation and phosphorylation with the computational algorithms YinOYang 1.2 (http://www.cbs.dtu.dk/services/YinOYang) and NetOGlyc v3.1 (http://www.cbs.dtu.dk/services/NetOGlyc/) (13). Tandem Repeats Finder (http://tandem.bu.edu/trf/trf.html) (3) was used to analyze the tandem repeats of the genes encoding E. canis gp19. Potential signal sequences were identified with the computational algorithm SignalP trained on gram-negative bacteria (http://www.cbs.dtu.dk/services/SignalP-2.0/) (28). Amino acid alignments were performed using the protein-protein Basic Local Alignment Search Tool (BLAST) (http://www.ncbi.nlm.nih.gov/BLAST/).

PCR amplification of the E. canis gp19 gene fragments.

Oligonucleotide primers for the amplification of the E. canis gp19 gene fragments (gp19N-terminal, gp19N1, gp19N2, gp19N1-C, and gp19C-terminal) were designed manually or by using Primer Select (Lasergene v5.08l DNAStar, Madison WI) (Table 1). The E. canis gp19 gene fragments were amplified using a PCR master mix (F. Hoffmann-La Roche Ltd., Basel, Switzerland) and E. canis (Jake strain) genomic DNA as the template.

TABLE 1.

Oligonucleotide primers for amplification of the E. canis gp19 gene

Recombinant protein	Forward primer	Sequence (5′ to 3′)	Reverse primer	Sequence (5′ to 3′)	Amplicon size (bp)
gp19	P16N-F	CACGTTCAAAATCATGTTGA	P16C-R	CGCACAATCACAACAGTTGT	405
gp19N	P16N-F	CACGTTCAAAATCATGTTGA	P16N-R	GCATACTGGTCTTTCCT	222
gp19N1	P16N-F	CACGTTCAAAATCATGTTGA	p19 132-R	AGATACTTCTTGTAACTCCATT	126
gp19N1-C	Small FOR	CATTTTACTGGTCCTACT	p19 132-R	AGATACTTCTTGTAACTCCATT	72
gp19N2	p19 133-F	TCTATTGATAGTGTAGGATGC	P16N-R	GCATACTGGTCTTTCCT	96
gp19C	P16C-F	GCAGGTTTAGAGAGCTT	P16C-R	CGCACAATCACAACAGTTGT	180

Cloning and expression of the recombinant E. canis gp19 gene.

The amplified PCR products were cloned directly into the pBAD/TOPO ThioFusion or pCR T7/NT TOPO expression vector (Invitrogen, Carlsbad, CA). Escherichia coli cells (TOP10; Invitrogen) were transformed with the plasmid containing the E. canis gp19 gene fragments, and positive transformants were screened by PCR for the presence of the insert and its orientation and were sequenced to confirm the reading frames of the genes. Recombinant protein expression was induced with 0.2% arabinose (pBAD/TOPO ThioFusion) or IPTG (isopropyl-β-d-thiogalactopyranoside) (pCR T7/NT) using Overnight Express autoinduction system 1 (Novagen, Madison, WI). Bacteria were pelleted (5,000 × g for 20 min) and resuspended in phosphate-buffered saline, and recombinant proteins were purified under native conditions and quantitated by a bicinchoninic protein assay (Pierce Biotechnology, Rockford, IL) as previously described (9).

Gel electrophoresis and Western immunoblotting.

Purified E. canis or E. chaffeensis whole-cell lysates or recombinant proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose, and Western blot analyses were performed as previously described (19), except that primary antibodies were diluted (1:500). Anti-E. canis and anti-E. chaffeensis dog sera were derived from experimentally infected dogs (no. 2995 and 2551, respectively).

Carbohydrate detection and glycosyl composition.

Glycan detection on the recombinant protein gp19 was performed with a digoxigenin glycan detection kit (Roche, Indianapolis, IN) as previously described (25). The glycosyl composition was determined by alditol acetate analysis at the University of Georgia Complex Carbohydrate Research Center (UGA-CCRC). The glycoprotein was hydrolyzed using 2 M trifluoroacetic acid (2 h in a sealed tube at 121°C), reduced with NaBD₄, and acetylated using acetic anhydride-trifluoroacetic acid. The resulting alditol acetate was analyzed on a Hewlett Packard 5890 gas chromatograph interfaced to a 5970 mass selective detector (electron impact ionization mode), and separation was performed on a 30-m Supelco 2330 bonded-phase fused silica capillary column.

Mouse anti-gp19 antibody production.

Five BALB/c mice (Jackson Laboratories, Bar Harbor, ME) were immunized with the recombinant E. canis gp19 protein (ThioFusion; amino acids 4 to 137). Recombinant protein (100 μg) in 0.1 ml was mixed with an equal volume of Freund's complete adjuvant (Sigma, St. Louis, MO) for the first intraperitoneal injection and with Freund's incomplete adjuvant for the subsequent injections. The mice were given injections twice at 2-week intervals.

E. canis gp19 synthetic peptide antibody epitope.

A 24-amino-acid peptide (N1-C; HFTGPTSFEVNLSEEEKMELQEVS) corresponding to the E. canis gp19 epitope-containing region was synthesized by Bio-Synthesis, Inc. (Lewisville, TX). The peptide (lyophilized) was resuspended in molecular biology grade water (1 mg/ml).

ELISA.

Enzyme-linked immunosorbent assay (ELISA) plates (Nunc-Immuno plates with MaxiSorp surface; NUNC, Roskilde, Denmark) were coated (1.25 μg/well, 100 μl) with recombinant protein (gp19 N1-C thioredoxin fusion) or peptide (wt/vol) in phosphate-buffered saline. Antigen was adsorbed to the ELISA plates overnight at 4°C with gentle agitation, subsequently washed three times with 200 μl Tris-buffered saline with Tween 20 (0.2%) (TBST) blocked with 3% bovine serum albumin (BSA) in TBST for 1 h at room temperature with agitation, and washed again. Convalescent anti-E. canis canine serum (1:4,000) diluted in 3% BSA-TBST was added to each well (100 μl) and incubated at room temperature for 1 h with gentle agitation. The plates were washed four times, and an alkaline phosphatase-labeled goat anti-dog immunoglobulin G (IgG) (heavy plus light chains [H+L]) secondary antibody (1:2,500) (Kirkegaard & Perry Laboratories) in 3% BSA-TBST was added and incubated for 1 h. The plates were washed four times, and substrate (BluePhos; 100 μl; Kirkegaard & Perry Laboratories) was added to each well. The plates were incubated for 30 min in the dark with agitation, color development was read on a microplate reader (VersaMax; Molecular Devices, Sunnyvale, CA) at A₆₅₀, and data were analyzed by SoftmaxPro v4.0 (Molecular Devices). Optical density (OD) readings represent the means for three wells (± standard deviations), with the OD of the buffer-only wells subtracted (see Fig. 5). Periodate treatment of the recombinant gp19 protein was carried out for 20 min in 100 mM sodium acetate buffer with 100 mM sodium metaperiodate. A Sham-treated control protein was incubated in the same buffers in the absence of periodate. The ELISA was carried out as described above for E. canis antigens, with the exception that the plate was blocked with milk diluent-blocking solution (Kirkegaard & Perry Laboratories, Gaithersburg, MD).

FIG. 5.

Immunoreactivity of recombinant gp19 (N1-C epitope; glycosylated) with anti-E. canis dog (no. 2995) serum compared to that of the synthetic peptide (aglycosylated) by ELISA (top). Immunoreactivity of E. canis gp19 (N1-C epitope) with anti-E. canis dog serum after treatment with periodate as determined by ELISA (bottom). The OD readings of the bar graph represent the means for three wells (± standard deviations), with the OD of the buffer-only wells subtracted.

Immunoelectron microscopy.

Immunogold electron microscopy was performed on infected DH82 cells as previously described (9), except that primary mouse anti-E. canis gp19 antibody was diluted 1:10,000. Uninfected DH82 cells were reacted with mouse anti-E. canis gp19 as a negative control.

Fluorescence confocal microscopy.

Antigen slides were prepared from DH82 cells infected with E. canis (Jake strain) as described previously (21). The rabbit anti-E. canis disulfide bond formation protein (DsbA) (22) diluted 1:100 was added to each well (15 μl) and allowed to incubate for 30 min. Slides were washed, and mouse anti-gp19 (1:100 dilution) was added and incubated for 30 min. Alexa Fluor 488 goat anti-rabbit IgG (H+L) secondary antibody (Molecular Probes, Eugene, OR) diluted 1:100 was added and incubated for 30 min, followed by washing and subsequent addition and incubation of rhodamine-labeled goat anti-mouse IgG (H+L) secondary antibody (Kirkegaard & Perry Laboratories). Mounting medium (ProLong Gold; Molecular Probes) was added, and the slides were viewed with an Olympus FV-1000 laser confocal microscope and FluoView software.

PCR amplification of E. canis gp19 from geographically dispersed isolates.

E. canis DNA from the United States (Jake, Demon, DJ, Louisiana, and Florida strains), from South America (Sao Paulo strain; kindly provided by Marcelo Labruna), from Israel (611 strain; DNA kindly provided by Avi Keysary), and from an infected dog from Mexico (Yucatan; kindly provided by Carlos Perez) was used as template to amplify the entire gp19 gene using flanking primers (forward-Ecanis p19 FOR, 5′-AAAATTAGTGTTGTGGTTATG-3′, and reverse-Ecanis p19 REV, 5′-TTTTACGCTTGCTGAAT-3′). The amplicons were cloned into a TA cloning vector (pCR 2.1; Invitrogen), and plasmids were transformed into E. coli (TOP10). Plasmids containing the gp19 gene were purified with a plasmid purification kit (Roche) and sequenced.

Nucleotide sequence accession numbers.

The gp19 gene sequences from E. canis isolates (Jake, DJ, Demon, Louisiana, Florida, Sao Paulo, and Mexico) were deposited into GenBank and assigned the following respective accession numbers: DQ858221, DQ858222, DQ858223, DQ858224, DQ858225, DQ860145, and DQ858226.

RESULTS

Molecular identification of the E. canis gp19 major immunoreactive protein.

Screening of an E. canis genomic expression library identified a clone that reacted strongly with antibody and contained an ∼3-kb insert. This clone was partially sequenced (∼900 bp) to reveal an incomplete 5′ open reading frame, which was aligned with the available E. canis genome sequence to fully identify the genes present within the 3-kb clone. The clone contained a complete 1,086-bp gene encoding a riboflavin biosynthesis protein (RibD) and a downstream 414-bp open reading frame encoding a protein of 137 amino acids with a predicted mass of 15.8 kDa with unknown function. The protein had a 37-amino-acid (QLGLLLGGFLSAMNYISYSYPCYYYDCCDRNYYDCCH) C-terminal region with >53% identity and ∼60% overall homology to E. chaffeensis VLPT, a known immunoreactive protein; therefore, this gene was considered for further investigation. The E. canis protein had substantial C-terminal region homology (60%) with E. chaffeensis VLPT, but it lacked the characteristic tandem repeats. The E. canis protein did have several predicted O-glycan attachment sites and one amino acid (serine 44) that was a predicted yin-yang site (glycosylation/phosphorylation). Further analysis of the gene position in the chromosome revealed the same adjacent genes for the 414-bp E. canis gene and the E. chaffeensis VLPT gene (Fig. 1). The ortholog of the E. canis gp19 gene or the E. chaffeensis VLPT gene was not identifiable in the E. ruminantium genome. However, a hypothetical protein of similar size (no homology) with high serine content and low cysteine and tyrosine content was found in the same relative chromosomal position as those of the gp19 gene and the VLPT gene, with the same adjacent genes as shown in Fig. 1.

FIG. 1.

Schematic of the relative genomic locations of the E. canis gp19 gene (Jake strain) (top) and the E. chaffeensis VLPT gene (Arkansas [Ark] strain) (bottom) and adjacent genes (sizes of genes and intergenic regions in base pairs) and enlarged E. canis gp19 and E. chaffeensis VLPT proteins illustrating the relevant domains (leader [L], STE-rich, S-rich, EC-rich, and CY-rich domains; number of amino acids in parentheses). The amino acid sequences of the comparable STE-rich and S-rich domains (similar amino acid usages and sizes; no homology) and homologous (59%) CY-rich domains (underlined) of gp19 and VLPT are shown. A BLASTp alignment of the CY-rich domains of E. canis and E. chaffeensis is shown in the box below (+, conserved substitution; −, gap).

Protein composition and characteristics.

Cysteine (14 residues; 10.2% of entire amino acid content), serine and threonine (13; 9.5% combined), glutamate (13; 9.5%), and tyrosine (13; 9.5%) were the most frequently occurring amino acids in the E. canis gp19 protein, accounting for more 38% of the entire amino acid content. Cysteine residues were not present in the first 50 amino acids, but the carboxyl-terminal domain of the protein (last 28 amino acids) was dominated by cysteine and tyrosine (55%). Serine, threonine (seven each; 27%), and glutamate (six; 23%) residues were the most frequently occurring amino acids in a small central domain (STE-rich patch; 26 amino acids) and accounted for 50% of the amino acid content, and a glutamate and cysteine (28% total) were dominant in an adjacent downstream 40-amino-acid domain (Fig. 1).

Conservation of E. canis gp19.

The E. canis gp19 gene was examined in geographically dispersed North American (Jake, DJ, Demon, Louisiana, and Florida) and South American (Brazil; Sao Paulo) isolates and was completely conserved. The gp19 sequence amplified from E. canis-infected dog blood from Mexico (Yucatan; kindly provided by Carlos Perez) and DNA from Israel (strain 611; kindly provided by Avi Keysary) had single nucleotide substitutions (positions 71 and 104, respectively) that resulted in single amino acid changes (amino acid 24-Gly to Asp and amino acid 35-Glu to Gly, respectively) that were located in the epitope-containing region.

Molecular mass and immunoreactivity of gp19.

The mass of the gp19 fusion recombinant protein was ∼35 kDa and was larger (∼3 kDa) than the predicted (32 kDa) mass which included the fusion tags (16 kDa) but was consistent with the ∼3-kDa larger-than-predicted (16 kDa) mass of the native gp19 protein (19 kDa). The recombinant gp19 protein reacted strongly with serum from a dog (no. 2995) experimentally infected with E. canis (data not shown).

Carbohydrate detection.

Carbohydrate was detected on the recombinant gp19 protein (N terminal) (see Fig. 4 for orientation), which contained the STE-rich patch (Fig. 2 ). Furthermore, glycosyl composition analysis of the N-terminal fragment by the UGA-CCRC using alditol acetate analysis revealed the presence of glucose (65%) and galactose (35%).

FIG. 4.

(Top) Schematic of E. canis recombinant gp19 fragments, including the epitope-containing region N1-C. (A) SDS-PAGE of E. canis recombinant gp19 fragments (N1, lane 1; N2, lane 2; N terminal, lane 3; C terminal, lane 4; and thioredoxin control, ctrl) and (B) corresponding Western immunoblot probed with anti-E. canis dog serum. (C) Comparison of E. canis gp19N1 and gp19N1-C reacted with anti-E. canis dog serum. aa, amino acids; M, Precision Protein standard (Bio-Rad).

FIG. 2.

Carbohydrate detection of E. canis gp19 (gp19N polypeptide) (lane 2) and E. canis Dsb (negative control) (lane 1). M, Precision Protein standard (Bio-Rad); CCM, CandyCane glycoprotein molecular mass standard (Molecular Probes) containing a mixture of glycosylated and nonglycosylated proteins (glycosylated proteins, 42 and 18 kDa; nonglycosylated proteins, 29 and 14 kDa).

Identification of native gp19 and species-specific reactivity.

Anti-recombinant gp19 antiserum reacted strongly with a 19-kDa protein in E. canis whole-cell lysates, and this protein was similarly recognized by anti-E. canis dog serum (Fig. 3A). The anti-recombinant gp19 serum also appeared to react weakly with another well-characterized E. canis glycoprotein, gp36, suggesting a minor cross-reactivity between these two proteins. The anti-recombinant gp19 serum did not recognize antigens in E. chaffeensis whole-cell lysates (Fig. 3B).

FIG. 3.

(A) Western immunoblot of E. canis whole-cell lysates probed with anti-E. canis gp19 serum (lane 1) and anti-E. canis dog (no. 2995) serum (lane 2) and uninfected DH82 cell lysates reacted with anti-E. canis dog serum (lane 3). (B) E. chaffeensis whole-cell lysates probed with anti-E. canis gp19 serum (lane 1) and anti-E. chaffeensis dog (no. 2551) serum (lane 2). M, Precision Protein standard (Bio-Rad).

Single STE-rich major epitope.

Epitope determinants of other glycoproteins have been determined, including E. chaffeensis gp47 and E. canis gp36 (8). The E. canis gp19 protein is strongly recognized by antibody of infected dogs, and it elicits an early antibody response (19). In order to identify the epitope-containing region, the E. canis gp19 gene fragments (N-terminal, C-terminal, N₁, N₂, and N_1C gene fragments) (Fig. 4) were amplified with primers (Table 1) to create overlapping recombinant fusion proteins. The expressed thioredoxin fusion gp19 recombinant fragments (N₁, N₂, and N terminal) exhibited larger-than-predicted (2- to 6-kDa) masses by SDS-PAGE, but the C-terminal fragment migrated at the predicted mass (Fig. 4A). Notably, some N-terminal recombinant proteins (N, N₁, and N_1-c) that were also expressed without a thioredoxin fusion partner (pCR T7/NT vector) exhibited an even larger (6 to 10 kDa) differential between predicted and observed masses (data not shown).

Antibody reacted strongly with the N-terminal recombinant fragment but did not react with the C-terminal fragment, indicating that an epitope(s) was located in the N-terminal region of the protein (Fig. 4B). Further localization of the epitope-containing region was determined with fragments N₁ and N₂. Antibody strongly reacted with N₁ (42 amino acids), and N₂ was weakly recognized (Fig. 4B). A region within N₁ that had a high Ser/Thr/Glu content (N_1C; 24 amino acids) consistent with those of other epitopes identified in other ehrlichial proteins reacted strongly with antibody consistent with that of the larger N₁ fragment, demonstrating that an epitope(s) was located in the 24-amino-acid region of N_1C (Fig. 4C).

Carbohydrate immunodeterminant.

We have previously shown carbohydrate to be an important immunodeterminant on major immunoreactive glycoproteins (8). Carbohydrate was detected on the N-terminal region of gp19, and the epitope localized to the STE-rich patch. Glycan attachment sites were also predicted within the STE-rich patch. To determine the role of carbohydrate determinants in antibody recognition, we compared the immunoreactivity of recombinant N_1C with that of the corresponding synthetic peptide. By ELISA, the synthetic peptide was substantially less immunoreactive with anti-E. canis dog serum (no. 2995) than the recombinant version (Fig. 5). Similarly, N_1C treated with periodate to alter glycan structure was less immunoreactive than sham-treated N_1C protein (Fig. 5).

Cellular and extracellular localization of gp19.

Several characterized ehrlichial glycoproteins are differentially expressed on dense-cored ehrlichiae (gp120, gp36, and gp47). However, by immunoelectron microscopy, the E. canis gp19 protein was observed within the cytoplasm of both reticulate and dense-cored ehrlichiae but was also detected on the morula fibrillar matrix and associated with the morula membrane (Fig. 6A). Some gp19 was observed in the host cell (DH82) cytoplasm and nucleus that was not observed in uninfected cells stained with anti-gp19 antisera (Fig. 6A and B). The localization of gp19 by immunogold labeling was consistent with observations for infected DH82 cells by confocal immunofluorescence microscopy using anti-gp19 (Fig. 7A) and anti-Dsb (present on ehrlichiae but not in the morula matrix) (Fig. 7B) antibodies, showing both Dsb and gp19 colocalization on ehrlichiae and border staining of the morula membrane by anti-gp19 only (Fig. 7C). Stained morulae were not observed in uninfected DH82 cells adjacent to infected cells in the same field (data not shown).

FIG. 6.

Electron photomicrograph of an ultrathin section of E. canis-infected DH82 cells demonstrating E. canis gp19 localization in a morula containing both reticulate and dense-cored ehrlichiae (A) and a corresponding ultrathin section containing uninfected DH82 cells (negative control) (B). Cells in both panels were reacted with mouse anti-gp19 antibody (1:10,000). Bar = 1 μm.

FIG. 7.

Confocal immunofluorescence photomicrograph of E. canis gp19 expression. E. canis-infected DH82 cells were dually stained with anti-E. canis gp19 (red) (A) and anti-ehrlichial Dsb (green) (B), and a merged image is shown in panel C.

DISCUSSION

The kinetics of antibody responses to major immunoreactive antigens of E. canis during experimental infection has been well established in a previous study (19). Two E. canis antigens (37 and 19 kDa) were consistently recognized early in the acute immune response. In a more a recent study, we described the identification and molecular characterization of the 37-kDa protein (gp36), which is a differentially expressed glycoprotein on dense-cored ehrlichiae and is secreted (7). As more-major immunoreactive proteins have been molecularly characterized in E. canis and E. chaffeensis, it has become apparent that many exhibit high serine/threonine content, contain tandem repeats, and are glycosylated (7, 25, 34, 35).

Although others have reported that orthologs of the E. chaffeensis VLPT gene were not identified in related genomes (E. canis and E. ruminantium) (10), we provide evidence herein that the 19-kDa protein identified in this study is the ortholog of the previously described VLPT protein in E. chaffeensis (32). E. chaffeensis VLPT is immunoreactive and has nonidentical serine-rich tandem repeats. Although carbohydrate has not been reported on E. chaffeensis VLPT, the protein also exhibits a mass double that predicted by its amino acid content, similar to other described ehrlichial glycoproteins (32). Interestingly, the VLPT ortholog that we identified in E. canis in this study lacks the tandem repeats found in the E. chaffeensis VLPT gene, but it has a Ser/Thr/Glu-rich patch that is similar in size and composition to that of a single VLPT repeat unit. The addition and deletion of tandem repeats are considered a major source of change and instability in ehrlichial genomes (11). The fact that the E. canis gp19 protein lacks tandem repeats while the E. chaffeensis VLPT protein has variable numbers is indicative of these genes being affected by this process. Additionally, these genes share the same chromosomal location and have substantial amino acid homology (∼60%) in a carboxyl-terminal domain.

Another major immunoreactive protein (MAP2) related to Anaplasma marginale MSP5 has been identified and molecularly characterized in E. canis, E. chaffeensis, and E. ruminantium with a molecular mass (∼21 kDa) similar to that of the gp19 protein identified in this study (1, 2, 17). However, there is no amino acid homology between MAP2 and gp19, and thus, these proteins are molecularly and immunologically distinct. Unlike gp19, MAP2 appears to have a mass consistent with that predicted by its amino acid sequence and does not have any serine-rich domains. There is substantial homology among MAP2 orthologs in Ehrlichia spp., and cross-reactions among heterologous MAP2 proteins have been reported (14, 17). In contrast, antibodies generated to E. canis gp19 were not cross-reactive with E. chaffeensis VLPT, and therefore, these proteins are species-specific orthologs. Other notable differences between MAP2 and gp19 include a major serine-rich linear epitope of gp19 that is strongly recognized by antibodies by Western immunoblotting, while antibodies to MAP2 of E. canis and E. chaffeensis appear to be directed primarily at a conformational epitope (1, 2, 14). In a previous study, we suggested that the 19-kDa major immunoreactive protein that we identified may be MAP2 (19); however, the data presented in this study support our conclusion that this protein is not MAP2 but rather gp19.

Consistent with numerous other major immunoreactive proteins that we have characterized, carbohydrate was present on the N-terminal region of the E. canis gp19 protein, and glucose and galactose were detected on this fragment. We have reported the presence of glucose and galactose as sugars attached to E. chaffeensis gp120 and E. canis gp140. Although E. canis gp19 did not have serine-rich tandem repeats, it contained a STE-rich patch within the N-terminal region, similar to the amino acid composition of tandem repeats found in other ehrlichial glycoproteins. Therefore, it is likely that O-linked glycans are attached to amino acids (serine/threonine) in this STE-rich patch. Furthermore, by using the prediction server YinOYang, serine residues within this region were identified as potential glycosylation/phosphorylation sites. Since this prediction server is trained on eukaryotic glycoproteins, the identification of specific residues that are glycosylated may not be reliable; however, it is worth noting that there is a consistent positive correlation between our experimental data and the prediction generated by this eukaryote-based prediction algorithm.

The amino acid composition of E. canis gp19 consisted predominately of five amino acids, cysteine, glutamate, tyrosine, serine, and threonine. Interestingly, these amino acids were concentrated in two specific domains, the epitope-containing region and the carboxyl-terminal region. The high Ser/Thr/Glu content of the epitope-containing region has been reported in other ehrlichial glycoproteins where epitopes have been mapped (8), and high serine and threonine content has been found in other ehrlichial glycoproteins, particularly in tandem repeat regions (7, 34, 35).

Another novel feature of gp19 is a carboxyl-terminal tail dominated by tyrosine and cysteine (55%). This carboxyl-terminal tail was also present on the E. chaffeensis VLPT protein downstream of the repeat region, suggesting that it is an important conserved domain in these proteins. Overall, cysteine was present more than any other amino acid, and because of this, gp19 is a member of a small group of proteins (n = 36) within the E. canis genome with high cysteine content (18). Cysteine is essential for intra- and intermolecular disulfide bond formation, and its high content in gp19 suggests that this protein has the potential to be linked with other cysteine-containing proteins by disulfide bonds or that cysteine residues are important for intramolecular bonding necessary for maintaining gp19 structure.

Tyrosine and serine are commonly phosphorylated. The high proportion of tyrosine residues in the carboxyl-terminal region of gp19 suggests a high potential for this domain of the protein to be phosphorylated. This condition also raises the possibility that gp19 is involved in protein signaling. The presence of phosphoproteins in E. chaffeensis has been reported (31), and more Ser/Thr/Tyr kinases and phosphoproteins are being identified in bacteria (12, 15, 16, 29). Nevertheless, tyrosine residues in this C-terminal region were not identified as sites of phosphorylation by NetPhos, which is trained on eukaryote proteins. Therefore, further studies will be necessary to determine whether any of the tyrosine residues are phosphorylated and how this possibility relates to protein function.

A major epitope-containing region was identified in E. canis gp19 in the STE-rich domain. This region elicits an early antibody response in dogs experimentally infected with E. canis (19). There appears to be a single carbohydrate-dependent epitope in this region, but it is also possible that multiple epitopes (peptide and carbohydrate) are present within the 24-amino-acid STE patch. Other epitopes that we have characterized within ehrlichial glycoproteins were mapped to serine-rich tandem repeats. Hence, finding a major epitope within the STE-rich region of gp19 is consistent with previous studies and demonstrates the importance of serine-rich regions and carbohydrate immunodeterminants for Ehrlichia species. Carbohydrate was detected on the N-terminal region of gp19 containing the STE-region. The recombinant gp19 epitope was more immunoreactive than the corresponding synthetic peptide, indicating that a posttranslational modification was present on this epitope. In addition, treatment of the recombinant epitope-containing peptide with periodate reduced its immunoreactivity, further supporting a role for carbohydrate as an immunodeterminant. These findings are consistent with our previous demonstration of carbohydrate as an immunodeterminant on the epitopes that were mapped in the serine-rich tandem repeat regions of E. chaffeensis gp47 and E. canis gp36 (8). Notably, this epitope is species specific, as the anti-gp19 antibody did not cross-react with E. chaffeensis antigens, similar to other species-specific major immunoreactive antigens that we have identified, including gp36 (8, 34, 35). Therefore, the potential exists for the development of sensitive species-specific immunodiagnostics utilizing E. canis gp19 alone or in combination with other antigens, such as gp36.

E. canis gp19 was found on both reticulate and dense-cored cells and appeared to be localized predominantly in the cytoplasm of ehrlichiae. The localization of gp19 is in contrast to that of another E. canis glycoprotein (gp36) that we reported to be differentially expressed primarily on the surface of dense-cored cells (8). However, similar to gp36, gp19 was also observed extracellularly in the morula fibrillar matrix and associated with the morula membrane. The presence of gp19 on ehrlichiae, the fibrillar matrix, and the morula membrane was further corroborated with immunofluorescence using dual staining with Dsb, which is not secreted and is present on both reticulate and dense-cored organisms (22). Some small morulae appeared to have less gp19, suggesting that the expression of gp19 becomes more predominant as the morula matures. E. canis gp19 does not have an amino-terminal signal sequence; therefore, the export of this protein probably involves a sec-independent secretion system (type I or type III). The gp19 protein also appears to be present outside the morula within the cell cytoplasm and nucleus, which suggests that gp19 may be directly involved host cell interactions during infection. Other ehrlichial proteins (gp200) are translocated to the host cell nucleus (27).

E. canis gp19 was highly conserved in E. canis isolates from the United States, Mexico, Brazil, and Israel. Single amino acid changes were identified in the Israeli and Mexican strains that were located in the mapped epitope, suggesting that immune pressure may be responsible for these changes. The effect of these mutations on immune recognition of gp19 remains to be determined. However, the conservation of major immunoreactive genes (p28, gp140, and gp36 genes) in geographically separated E. canis isolates has consistently been reported (7, 23, 24, 26, 35) and supports the conclusion that globally or regionally effective vaccines and reliable immunodiagnostics for E. canis based on major immunoreactive proteins such as gp19 are feasible.