Antigenic Classification of Rickettsia felis by Using Monoclonal and Polyclonal Antibodies

Abstract

Rickettsia felis is a flea-transmitted rickettsia. There is a discrepancy between its reported phylogenic and phenotypic identifications. Following the first report of R. felis, it was considered by tests with serologic reagents to be closely related to another recognized flea-transmitted rickettia, R. typhi. Subsequently, it appeared to be more closely related to spotted fever group (SFG) rickettsiae by genetic analysis. In the present work, R. felis was studied by microimmunofluorescence (MIF) serologic typing and with monoclonal antibodies (MAbs). Mouse polyclonal antisera to R. felis cross-reacted only with SFG rickettsiae. A neighbor-joining analysis based on MIF indicated that R. felis is actually related to SFG rickettsiae antigenically, clustering with R. australis, R. akari, and R. montanensis. A panel of 21 MAbs was raised against a 120-kDa protein antigen or a 17-kDa polypeptide of R. felis. They cross-reacted with most members of the SFG rickettsiae but not with R. prowazekii, R. typhi, or R. canadensis of the typhus group (TG) rickettsiae. Sixty-four MAbs previously generated to seven other ricketttsial species were tested with R. felis. Three MAbs reacted with the 120-kDa antigen and were generated by R. africae, R. conorii, and R. akari, respectively. They exhibited cross-reactivities with R. felis. All our data show that R. felis harbors the antigenic profile of an SFG rickettsia.

Rickettsia spp. are gram-negative and obligate intracellular bacteria (24). They are bacilli of 0.3 to 0.5 μm in diameter and 0.8 to 0.2 μm in length which retain basic fuchsin when they are stained by the method of Gimenez. Pathogenic rickettsiae are transmitted to humans by arthropods and cause clinical diseases that manifest typically as fever, rash, and vasculitis. At present, rickettsiae are divided into two groups, the spotted fever group (SFG) and the typhus group (TG), on the basis of their intracellular positions, optimal growth temperatures, percent G+C DNA contents, clinical features, epidemiological aspects, and antigenic characteristics (24, 30).

Recently, a novel rickettsia-like organism was observed in the midgut epithelial cells of cat fleas (Ctenocephalides felis) in California by electron microscopy (1). It was described as the ELB agent, for the EL Laboratory (Soquel, Calif.), where it was originally described (1, 23). In 1996, this flea-borne bacterium was proposed as a distinct species, “Rickettsia felis” (12), and was characterized as a typhus-like rickettsia. Monoclonal antibodies (MAbs) specific for R. typhi reacted with this organism (6). Its ultrastructure and tissue distribution in fleas resembled those of R. typhi (1, 6). However, molecular data, which were obtained by sequencing and phylogenetic analysis of a 17-kDa protein-encoding gene, a citrate synthase-encoding gene, a 155-kDa protein-encoding gene, a 120-kDa protein-encoding gene, and the metK, ftsY, polA, and dnaE genes, classified R. felis into the SFG rickettsiae (2, 6, 9, 26, 27, 29).

Although advanced genetic techniques have been extensively used to classify rickettsial species, serotyping by indirect microimmunofluorescence (MIF) with mouse antisera is still considered valuable for its general applicability (20). Serotyping of R. felis organisms among rickettsial species has never been completed, most likely owing to its resistance to cultivation (21). Recent isolation of this organism (strain Marseille-URRWFXCal₂^T) from cat fleas provided by the EL Laboratory and cultivation by the shell vial cell culture procedure with XTC-2 cells have made it available for subsequent work (15, 23). In order to provide complementary data for serological classification of R. felis, we generated a series of murine polyclonal antisera against SFG and TG rickettsial species to serotype this new agent. Moreover, MAbs to R. felis, which were produced in this work, and MAbs to R. africae, R. massiliae, R. akari, R. conorii, R. slovaca, R. prowazekii, and R. typhi, which were developed in previous studies (10, 32-35), were used to further ensure its antigenic position.

MATERIALS AND METHODS

Rickettsial strains.

Strain Marseille-URRWFXCal₂^T of R. felis was isolated and reliably propagated (23). Thirteen strains of Rickettsia species, including R. rickettsii, R. australis, R. akari, R. montanensis, R. honei, R. japonica, R. typhi, R. canadensis, R. prowazekii, R. massiliae, R. bellii, R. conorii Seven, and R. conorii Moroccan, were used as antigens for MIF serotyping (Table 1). R. felis was cultivated in XTC-2 cells (23), and the other rickettsial strains were grown in L929 cell monolayers (ATCC CCL 1 NCTC clone 929) at 37°C with 5% CO₂. After 7 to 10 days, heavily infected cells were monitored by Gimenez staining (11) and then harvested with glass beads. After cultivation, purification, and quantification as reported previously (32, 33), the antigens were stored at −70°C until use for mouse immunization, MIF, sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE), and Western blotting.

TABLE 1.

Rickettsial strains studied

Rickettsial strain	Source	Geographic origin	Pathogenicity (disease)	Standard strain no.a
R. rickettsii Sheila Smith	Human	Montana	Rocky Mountain spotted fever	ATCC VR149
R. sibirica 246	Dermacentor nuttalli	Former USSR	North Asian tick typhus	ATCC VR151
R. conorii Malish	Human	South Africa	Mediterranean spotted fever	ATCC VR613
R. conorii Moroccan	Unknown	Morocco	Mediterranean spotted fever	ATCC VR141
R. conorii Astrakhan	Human	Russia	Astrakhan fever
R. conorii Israel ISTT CDC1	Human	Israel	Israeli spotted fever
R. conorii Indian (Indian tick typhus)	Rhipicephalus sanguineus	Indian	Indian tick typhus
Thai tick typhus rickettsia TT-118	Ixodes or Rhipicephatus spp.	Thailand	Spotted fever
R. australis Philips	Human	Australia	Queensland tick typhus
R. akari MK (Kaplan)	Human	New York State	Rickettsialpox	ATCC VR148
R. montanensis tick strain	Dermacentor andersoni	Montana		ATCC VR611
R. rhipicephali 3-7-6	Rhipicephalus sanguineus	Mississippi
R. parkeri Maculatum 20	Amblyomma maculatum	Mississippi
R. africae Z9-Hu	Human	Zimbabwe	African tick-bite fever
R. helvetica C9P9	Ixodes ricinus	Switzerland
R. massiliae Mtu 1	Rhipicephalus turanicus	France		ATCC VR1375
“R. mongolotimonae” HA-91	Hyalomma asiaticum kozlovi	China	Spotted fever
Bar 29	Rhipicephalus sanguineus	Spain
R. slovaca 13-B	Dermacentor marginatus	Slovakia	Tick-borne lymphadenopathy
R. honei RB	Human	Australia	Flinders Island spotted fever
R. bellii 369L42-1	Dermacentor andersion	Ohio
R. felis Marseille	Ctenocephalides felis	United States	Flea-borne spotted fever
R. japonica YH	Human	Japan	Japanese spotted fever	ATCC VR1363
R. aeschlimannii MC16	Hyalomma marginatum	Morocco
Strain S	Rhipicephalus sanguineus	Former USSR
R. typhi Wilmington	Human	North Carolina	Murine typhus	ATCC VR144
R. canadensis 2678	Haemophysalis leporispalustris	Canada		ATCC VR610
R. prowazekii Breinl	Human	Poland	Epidemic typhus	ATCC VR142

^aATCC, American Type Culture Collection.

Mouse immunization.

Murine polyclonal antisera were produced by a previously described method (7). It differs from that of Philip et al. (20) in terms of the mouse strain, the number of mice per antiserum pool, the number of immunizations, the timing of the immunizations, and the route of the first three immunizations used. In brief, for each species, three BALB/c mice were immunized intraperitoneally three times at 7-day intervals with 0.5 ml of phosphate-buffered saline containing ∼4 × 10⁴ PFU (16). One week after the last injection, the mice were boosted intravenously in the tail vein with 0.1 ml of the same suspension. In order to avoid host variations in antibody responses, three mice were immunized with each strain. Three days later, the animals were killed and the samples were pooled.

Production of MAbs.

Six-week-old female BALB/c mice were immunized with R. felis as described above. Splenocytes from two mice were subjected to fusion with SP2/0-Ag14 myeloma cells by using 50% (wt/vol) polyethylene glycol (molecular weight, 1,300 to 1,600; Sigma Chemical Co., St. Louis, Mo.) (32). The hybridoma cells were cultivated in selective hybridoma medium (Gibco BRL, Life Technologies Ltd., Paisley, Scotland) containing 20% fetal bovine serum (Gibco BRL), and 1× hypoxanthine-aminopterin-thymidine (Gibco BRL), plated in 96-well plates, and cultivated for 2 weeks. The fused cells were then grown in 1× hypoxanthine-thymidine selective medium (Sigma) for 5 days. The viable hybridoma clones were screened for antibodies against R. felis by the MIF assay. Positive clones were subcloned twice by limiting dilution. The immunoglobulin class and subclass of each MAb were determined with an ImmunoType mouse monoclonal antibody isotyping kit (Sigma).

MIF assay.

The MIF assay was performed by a previously described method (20, 32-34). Antigens were deposited onto the slides with a pen nib and dried. The slides were fixed in methanol for 10 min at room temperature. Polyclonal antisera or supernatants of hybridoma clones were tested by using twofold dilutions. Binding of sera was detected by using fluorescein (fluorescein isothiocyanate)-conjugated AffiniPure goat anti-mouse immunoglobulin G (IgG) plus IgM (heavy and light chains; Jackson ImmunoResearch Laboratories, Inc., West Grove, Pa.) diluted 1/100 in phosphate-buffered saline containing 0.2% Evans blue (BioMerieux, Marcy l'Etoile, France). The sera of immunized and normal mice were used as positive and negative controls, respectively, in each assay.

Sixty-four MAbs to rickettsial species were selected to analyze the antigenic cross-reactions with R. felis (Table 2). MAbs to R. slovaca (13-B strain) and R. akari [strain MK(Kaplan)] were kindly provided by D. H. Walker. Other MAbs raised against R. prowazekii (Breinl strain), R. typhi (Wilmington strain), R. africae (Z9-Hu strain), R. massiliae (Mtu 1 strain), and R. conorii (Malish 7 strain) were previously developed in our laboratory (10, 32-35). They have been demonstrated to react with different antigenic epitopes distributed on a 155-kDa protein, a 120-kDa protein, or a lipopolysaccharide (LPS)-like antigen.

TABLE 2.

MAbs to other species tested in this study

Reference MAb	Immunogen	Specificity (kDa)	Reactivities with other speciesa	Reactivity with R. felis
AF1-B9	R. africae	155	5, 9, 10, 13, 16, 17, 20, 24, 27, 28	Negative
AF1-D12	R. africae	155	5, 9, 10, 13, 16, 17, 20, 24, 27, 28	Negative
AF4-C8	R. africae	155	2, 5, 9, 10, 13, 14, 16, 17, 18, 19, 20, 24, 27, 29	Negative
AF6-D6	R. africae	120	2, 5, 7, 9, 10, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 24, 25, 26, 27, 28, 29, 30	Positive
AF6-E10	R. africae	155	2, 5, 9, 10, 13, 14, 15, 16, 17, 18, 23, 26, 28, 29	Negative
AF8-F3	R. africae	120	2, 5, 9, 10, 24, 27, 28	Negative
RC1-C7	R. conorii	155	13, 16, 17, 20,	Negative
RC1-H10	R. conorii	120	1, 2, 3, 5, 7, 9, 10, 12, 13, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30,	Negative
RC2-F2	R. conorii	120	1, 2, 5, 9, 10, 13, 14, 15, 16, 17, 18, 19, 20, 20, 24, 25, 26, 27, 28	Negative
RC2-D9	R. conorii	120	7, 9, 10, 13, 14, 15, 16, 17, 18, 19, 20, 28, 29	Negative
RC2-E3	R. conorii	155	13, 16, 17, 18, 20	Negative
RC2-G12	R. conorii	155	5, 13, 16, 17, 18, 19, 20, 29	Negative
RC3-G4	R. conorii	155	1, 2, 5, 9, 10, 12, 13, 14, 16, 17, 20, 24, 26, 27, 28, 30	Negative
RC3-G7	R. conorii	120	1, 2, 5, 13, 14, 16, 17, 18, 19, 20, 24, 27	Negative
RC4-C1	R. conorii	155	5, 13, 16, 17, 18, 19, 20,	Negative
RC5-A5	R. conorii	LPS-like antigen	1, 2, 3, 4, 5, 6, 7, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30	Negative
RC7-G11	R. conorii	120	5, 13, 14, 16, 17, 18, 19, 20,	Negative
RC8-D1	R. conorii	120	1, 2, 5, 9, 10, 13, 14, 16, 17, 18, 19, 20, 23, 24, 27, 28, 29	Negative
RC8-C9	R. conorii	120	13, 16, 17, 18, 19, 20,	Negative
RC9-B8	R. conorii	120	1, 2, 5, 9, 10, 13, 14, 15, 16, 17, 18, 19, 20, 24, 25, 26, 27, 28	Negative
RC9-G11	R. conorii	120	13, 16, 17, 18, 19, 20	Negative
RC10-A3	R. conorii	120	1, 2, 3, 5, 7, 9, 10, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30	Positive
RC10-B5	R. conorii	120	13, 16, 17, 18, 19, 20	Negative
MA1-D2	R. massiliae	Not identified	1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30	Negative
MA4-C11	R. massiliae	120	7, 21, 22	Negative
MA6-F3	R. massiliae	120	21, 22	Negative
MA7-A11	R. massiliae	120	3, 5, 15, 21, 22, 25	Negative
MA8-F6	R. massiliae	120	15, 21, 22	Negative
AK1-C12	R. akari	120	4	Negative
AK2-B41	R. akari	120	4, 6	Positive
SV10-D2	R. slovaca	120	1, 2, 5, 13, 14, 16, 17, 18, 19, 20, 21, 22, 29	Negative
SV10-E8	R. slovaca	120	1, 2, 9, 10, 13, 16, 17, 18, 19, 20, 24, 27, 28, 29	Negative
P11A12	R. prowazekii	120	31	Negative
P9H3	R. prowazekii	120	31	Negative
P5G11	R. prowazekii	120	31	Negative
P12A12	R. prowazekii	120	31, 33	Negative
P2G9	R. prowazekii	120	31	Negative
P12E11	R. prowazekii	120	31, 33	Negative
P2H12	R. prowazekii	120	31	Negative
P8H11	R. prowazekii	120	31, 32, 33	Negative
P11C8	R. prowazekii	LPS-like antigen	31, 32, 33	Negative
P10B8	R. prowazekii	LPS-like antigen	31, 32, 33	Negative
P6F12	R. prowazekii	LPS-like antigen	31, 32, 33	Negative
P12E12	R. prowazekii	LPS-like antigen	31, 32, 33	Negative
P7F12	R. prowazekii	LPS-like antigen	31, 32, 33	Negative
P9G12	R. prowazekii	LPS-like antigen	31, 32, 33	Negative
P10F12	R. prowazekii	LPS-like antigen	31, 32, 33	Negative
P5G7	R. prowazekii	LPS-like antigen	31, 32, 33	Negative
P12H10	R. prowazekii	LPS-like antigen	31, 32, 33	Negative
T7A5 E10E10	R. typhi	120	32	Negative
T6A10G3F3	R. typhi	120	32	Negative
T4D2C9H7	R. typhi	120	32	Negative
T7F5B12G8	R. typhi	120	5, 31, 32	Negative
T7F5H8H9	R. typhi	120	26, 32	Negative
T7F2H8E10	R. typhi	120	32	Negative
T7F12F11H10	R. typhi	120	26, 32	Negative
T7G10F7G12	R. typhi	120	30, 32	Negative
T7G10F7E11	R. typhi	120	30, 31, 32	Negative
T7G2F8G11	R. typhi	120	26, 30, 12, 15, 33, 32	Negative
T7G2G9H12	R. typhi	120	26, 30, 12, 15, 32	Negative
T7G11A8G12	R. typhi	120	31, 32	Negative
T1F3F9E9	R. typhi	120	31, 32, 33	Negative
T2G7F11A12	R. typhi	LPS-like antigen	31, 32, 33	Negative
T6F1E10	R. typhi	LPS-like antigen	31, 32, 33	Negative

^a1, R. africae (Z9-Hu); 2, R. africae (Ethiopian), 3, R. aeschlimannii (MC16); 4, R. akari [MK(Kaplan)]; 5, Astrakhan fever rickettsia (A167); 6, R. australis (Philips); 7, strain Bar 29; 8, R. bellii (369L42-1); 9, strain BJ90; 10, “R. mongololimonae” (HA-91); 11, R. helvetica (C9P9); 12, R. honei (RB); 13, R. conorii Indian; 14, R. conorii Israeli (ISTT CDC1); 15, R. japonica (YH); 16, R. conorii Kenya; 17, R. conorii Manuel; 18, R. conorii M-1; 19, R. conorii Moroccan; 20, R. conorii Seven; 21, R. massiliae (Mtu 1); 22, R. massiliae (GS); 23, R. montanensis (tick stain); 24, R. parkeri (Maculatum 20); 25, R. rhipicephali (3-7-6); 26, R. rickettsii (Sheila Smith); 27, strain S; 28, R. sibirica (246); 29, R. slovaca (13-B); 30, Thai tick typhus rickettsia (TT-118); 31, R. prowazekii (Breinl); 32, R. typhi (Wilmington); 33, R. canadensis (2678); 34, R. felis (Marseille).

SDS-PAGE and Western immunoblotting.

SDS-PAGE was performed by a modification of the method of Laemmli (14), as described previously (32). The Renografin-purified preparations of R. felis and other rickettsial reference strains were electrophoretically separated in a 10% resolving gel at a constant current of 16 mA for 2 h in an electrophoretic cell (Mini-Protean II; Bio-Rad Laboratories, Richmond, Calif.). Prestained SDS-PAGE standards (Bio-Rad) were used as a reference.

The separated polypeptides were stained with Coomassie brilliant blue or were transferred to a nitrocellulose sheet (pore size, 0.45 μm; Trans-Blot transfer medium; Bio-Rad) at 100 V for 1 h. Then, the nitrocellulose membrane was cut into strips and blocked overnight at 4°C with Tampon saturation buffer (0.121% [wt/vol] Tris base, 150 mM NaCl, 0.05%[vol/vol] Tween 20) containing 5% nonfat powdered milk. After the strips were washed three times in Tampon saturation buffer for 10 min each time, the strips were incubated in MAbs or polyclonal antisera to R. felis diluted 1:6 in Tampon saturation buffer supplemented with 0.5% nonfat dry milk and were then incubated at 4°C overnight on a rocker. The strips were washed as described above and then incubated for 2 h at room temperature with the peroxidase-conjugated F(ab′)₂ fragment of goat anti-mouse IgG (heavy and light chains; AffiniPure; Jackson Immuno Research) diluted at 1:400 in Tampon saturation buffer containing 3% nonfat dry milk. After the strips were washed, the bound conjugate was detected in a solution which contained 0.015% 4-chloro-1-naphthol (Sigma) in 16.7% methanol in Tampon saturation buffer containing 0.015% hydrogen peroxide. The purified preparations of R. felis were suspended in an equal volume of sample buffer (0.0625 M Tris hydrochloride [pH 8.0], 2% SDS, 5% 2-mercaptoethanol, 10% glycerol, 0.02% bromophenol blue) and either used as native antigen, boiled for 10 min, or digested with 0.2 mg of proteinase K per ml for 2 h at 37°C. The sera of the nonimmunized mice were used as a negative control.

Serotyping and numerical taxonomic analysis.

The MIF test was performed and the specificity difference (SPD) was calculated by use of the formulae of Philip et al. (20). Briefly, as described above, polyclonal mouse antisera (twofold dilutions) were incubated with both homologous and heterologous antigens (infected cells). The highest serum dilutions that gave positive reactions were recorded as endpoint titers: SPD = (Aa + Bb) − (Ab + Ba), where Aa and Bb are −log₂ of the endpoint titer between serum samples A and B and homologous antigens a and b, respectively, and Ab and Ba are −log₂ of the endpoint titer of serum samples A and B against heterologous antigens b and a, respectively. If the SPD was less than 3, the two strains were considered to belong to the same species; if the SPD was 3 or greater, the strains were considered to belong to different species.

The SPDs were used as a measure of the antigenic relationship of R. felis with selected Rickettsia species and to construct a dendrogram tree from the matrix by using the neighbor-joining method available in version 2.1 of the MEGA software package (13) according to the instructions of the manufacturer.

RESULTS

SDS-PAGE and Western blotting with mouse polyclonal antisera.

On SDS-PAGE, the protein of R. felis produced electrophoretic mobility patterns distinctive from those of the other rickettsiae including the SFG and TG rickettsiae (Fig. 1). A predominant heat-resistant protein with a molecular mass of 60 kDa (arrow in Fig. 1) and another heat-resistant 30-kDa protein were noticed, especially for R. felis. Mouse sera with antibodies to R. felis showed reactivities with several major bands, corresponding to 120-, 60-, 30-, and 17-kDa proteins and an LPS-like antigen (Fig. 2). After digestion with proteinase K, the bands reactive with proteins of 120, 60, 30, and 17 kDa were destroyed. The reactivity of the LPS-like antigen was not affected by digestion with proteinase K and heating.

FIG. 1.

Protein profiles of R. felis and representative rickettsiae on Coomassie brilliant blue-stained SDS-polyacrylamide gels. Lane markers on the left, molecular mass markers (catalog no. 161-0324, control 92577; Bio-Rad) were loaded in the order 205, 133, 83, 39.7, 31.4, 19.8, and 7.7 kDa; lane 1, R. felis; lane 2, R. australis; lane 3, R. akari; lane 4, R. bellii; lane 5, R. honei; lane 6, R. montanensis; lane7, R. typhi; lane 8, R. prowazekii; lane 9, R. canadensis. The predominant heat-stable 60-kDa protein (arrow) of R. felis was distinctive compared with the proteins detected among the other rickettsial species examined.

FIG. 2.

R. felis antigen treated in three different ways: staining with Coomassie blue following SDS-PAGE and detection by immunoblotting with murine polyclonal antisera to R. felis. Molecular markers (in kilodaltons) are indicated on the left for SDS-PAGE (catalog no. 161-0324, control 91722; Bio-Rad) and were loaded for immunobloting (catalog no. 161-0324, control 92577; Bio-Rad) in the order 204, 124, 80, 42.6, 31.7, 17.9, and 7.9 kDa. Lane NP, native antigen treated with polyclonal antisera to R. felis; lane NG, Coomassie blue-stained native antigen after SDS-PAGE; lane HP, heated antigen treated with polyclonal antisera to R. felis; lane HG, stained heated antigen; lane PG, stained proteinase K-treated antigen; lane PP, proteinase K-digested antigen treated with polyclonal antisera to R. felis. Note the presence of two reactive proteins between 80 and 120 kDa and one at 40 kDa. Reactions with the LPS-like antigen are observed in the range of 15 to 32 kDa (lane PP).

Serotyping by MIF assay and numerical taxonomic analysis.

The results (reciprocal titers and SPDs) of serotyping by the MIF assay with mouse polyclonal antisera are shown in Table 3. The strains tested in our work (except for strains R. conorii Seven and R. conorii Moroccan) are different serotypes (3 ≤ SPDs ≤ 24) and represent different species on the basis of classic antigenic taxonomic criteria (20). The SPDs between R. felis and other strains were always greater than 5, for example, 9 for R. rickettsii, 14 for R. conorii, and 23 for R. typhi. Antisera to R. felis cross-reacted with all SFG species evaluated in this study but not with TG species. It is noteworthy that the R. canadensis antigen showed weak cross-reactivities with mouse sera with antibodies against R. prowazekii, R. typhi, R. felis, R. australis, and R. rickettsii; these cross-reactivities were not observed in the previous study (20). Differences in cross-reactivity were presumably due to differences in the protocol for stimulation of the mouse antisera used in the method of Philip et al. (20)

TABLE 3.

MIF antibody titers and SPDs obtained from reciprocal cross-reactions of mouse antisera with SFG and TG rickettsiae^a

Species studied	MIF antibody titer (SPD)b
	R. fel.	R. ric.	R. aus.	R. mon.	R. hon.	R. jap.	R. typ.	R. can.	R. pro.	R. mas.	R. bel.	R. con. 7	R. con. M	R. aka.
R. felis	4,096 (0)	64 (9)	512 (5)	256 (5)	256 (15)	128 (7)	0 (23)	0 (20)	0 (19)	32 (9)	64 (9)	16 (14)	16 (14)	128 (7)
R. rickettsii	512 (9)	4,096 (0)	512 (6)	256 (6)	1,024 (4)	256 (5)	64 (12)	64 (11)	4 (16)	128 (7)	64 (9)	128 (8)	32 (9)	1,024 (8)
R. australis	2,048 (5)	1,024 (6)	8,192 (0)	2,048 (13)	256 (15)	128 (17)	16 (20)	16 (14)	16 (18)	256 (10)	1,024 (8)	64 (19)	64 (15)	512 (15)
R. montanensis	1,024 (5)	512 (6)	0 (13)	2,048 (0)	128 (16)	256 (7)	0 (22)	0 (23)	0 (20)	1,024 (6)	0 (15)	1,024 (11)	512 (11)	0 (22)
R. honei	0 (15)	1,024 (4)	4 (15)	0 (16)	4,096 (0)	128 (6)	0 (23)	0 (23)	0 (17)	2,048 (4)	0 (17)	512 (6)	256 (8)	128 (8)
R. japonica	0 (7)	1,024 (5)	0 (17)	128 (7)	1,024 (6)	2,048 (0)	0 (22)	0 (23)	0 (22)	256 (6)	0 (20)	0 (17)	0 (17)	0 (17)
R. typhi	0 (23)	32 (12)	0 (20)	0 (22)	0 (23)	0 (22)	2,048 (0)	32 (10)	1,024 (3)	0 (17)	0 (17)	0 (19)	0 (22)	0 (22)
R. canadensis	16 (20)	128 (11)	128 (14)	0 (23)	0 (23)	0 (23)	256 (10)	4,096 (0)	256 (15)	0 (23)	0 (21)	0 (20)	0 (23)	256 (15)
R. prowazekii	64 (19)	128 (16)	4 (18)	16 (20)	256 (17)	4 (22)	2,048 (3)	4 (15)	8,192 (0)	128 (14)	0 (21)	128 (14)	32 (19)	32 (19)
R. massiliae	2,048 (9)	2,048 (7)	256 (10)	256 (6)	1,024 (4)	1,024 (6)	128 (17)	4 (23)	32 (14)	8,192 (0)	128 (19)	1,024 (7)	1,024 (5)	256 (6)
R. bellii	1,024 (9)	1,024 (9)	256 (8)	512 (15)	128 (17)	16 (20)	128 (17)	16 (21)	32 (21)	0 (19)	8,192 (0)	0 (22)	0 (24)	1,024 (14)
R. conorii 7	8 (14)	512 (8)	0 (19)	4 (11)	512 (6)	64 (17)	16 (19)	16 (20)	4 (14)	256 (7)	8 (22)	4,096 (0)	2,048 (2)	16 (12)
R. conorii M	32 (14)	512 (9)	8 (15)	4 (11)	128 (8)	32 (17)	0 (22)	0 (23)	0 (19)	512 (5)	0 (24)	1,024 (2)	2,048 (0)	8 (12)
R. akari	128 (7)	32 (8)	0 (15)	0 (22)	256 (8)	32 (17)	0 (22)	0 (15)	0 (19)	1,024 (6)	0 (14)	128 (12)	128 (12)	2,048 (0)

^aTiters are the reciprocals of the highest dilution of antisera that gave a positive reaction. SPDs are calculated according to the formula given in the text.

^bR. fel., R. felis; R. ric., R. rickettsii; R. aus., R. australis; R. mon., R. montanensis; R. hon., R. honei; R. jap., R. japonica; R. typ., R. typhi; R. can., R. canadensis; R. pro., R. prowazekii; R. mas., R. massiliae; R. bel., R. bellii; R. con. 7, R. conorii Seven; R. con. M, R. conorii Moroccan; R. aka., R. akari.

The dendrogram inferred from comparison of the SPD matrices by the neighbor-joining method is presented in Fig. 3. It clearly demonstrated that the rickettsial species tested could be divided into two distinctive groups: the SFG and the TG rickettsiae. R. felis was clustered with SFG species and was especially closely related to R. australis, R. akari, and R. montanensis.

FIG. 3.

Taxonomic position of R. felis among Rickettsia species on the basis of antigenic analysis inferred from comparison of the SPDs by the neighbor-joining method. The scale bar represents a 2% difference.

Production and characterization of MAbs to R. felis.

A total of 415 wells containing viable, visible hybridoma clones after spleen cells from two mice fused with SP2/0 myeloma cells were observed. By screening by the MIF assay, the hybridomas in 286 wells were found to secrete antibodies to R. felis, and 65 were selected for subcloning by limiting dilution. Of these hybridomas, 37 reacted specifically with R. felis and 28 recognized some epitopes shared by other rickettsiae of the SFG species but not the TG species.

Among 21 MAbs against R. felis that were finally selected, 8 were species specific, 13 cross-reacted with several SFG rickettsiae, and none cross-reacted with TG rickettsiae (Table 4). Twenty MAbs showed reactivities with the 120-kDa protein band, and nine of these reacted with heat-resistant epitopes, since their reactivities were not affected by boiling (Fig. 4). One MAb, named F18C4, was reactive with the 17-kDa protein antigen and was heat sensitive. No MAbs against the LPS like-antigen were obtained, and none retained their reactivities after digestion with proteinase K. Three MAbs revealed that R. felis shares one or more epitopes located on the 120-kDa protein antigen with most SFG rickettsiae, excluding R. helvetica, R. bellii, and R. aeschlimannii. MAb F18C4 demonstrated that an epitope distributed on the 17-kDa protein antigen of R. felis was shared by SFG rickettsiae including R. rickettsii, R. australis, R. akari, R. japonica, R. sibirica, R. rhipicephali, R. massiliae, R. helvetica, R. conorii Israeli, R. bellii, R. aeschlimannii, R. conorii Indian, R. parkeri, R. africae, R. honei, Thai tick typhus rickettsia, and R. conorii strains Moroccan and S. The isotypes of the MAbs are presented in Table 4.

TABLE 4.

Reactivities of a set of MAbs to R. felis with SFG and TG rickettsiae^a

MAab	Isotype	Specificity (kDa)	Reactivity (reciprocal titer)
			1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22	23	24	25	26	27	28
F20A9H5G8	IgG1	120	41	512	8	2							64						256
F14C9H5H5	IgG2b	120		1,024										16			64
F19F8A4A8	IgG2a	120		2,048
F11E9A2G1	IgG2a	120		4,096
F11A3H1B11	IgG1	120	2	8,192	512	1,024		1,024	1,024				4,096	4,096	4,096		4,096	4,096			4,096	4,096	2,048	2,048	1,024	4,096	2,048	2,048	2,048	4,096
F23H2B4B12	IgG1	120	2	4,096	512	1,024		1,024	2,048				1,024	2,048	2,048		1,024	2,048			2,048	2,048	1,024	2,048	2,048	2,048	1,024	1,024	1,024	1,024
F18C4G3H1	IgM	17	8	256	32	8			4				128	128	32	2		32	32	32		32	128	64				32	16	4
F14C9E7F5	IgG2b	120		2,048											2
F20A9H5F10	IgG1	120	4	2,048	4								64		4				1,024			2	8							4
F23H2H4	IgG1	120		8,192	512	1,024		1,024	1,024				4,096	4,096	4,096		2,048	4,096			4,096	2,048	2,048	1,024	2,048	4,096	2,048	1,024	4,096	4,096
F19F8H3H4	IgG2a	120		4,096
F24cH3C3C	IgG2b	120		512
F20F9F3D12	IgG3	120		1,024
F24A2D1F9	IgG1	120		4,096
F24cD9E2G1	IgG1	120		4,096
F24cH5H2G	IgG2a	120		2,048
F24cH5G4	IgG2a	120		1,024
F10H6C1	NTb	120		256				16	16
F24cH12D1E	NT	120		128
F20A9B7B10	IgG1	120		256									4						128							2				4
F11E9B1H2	IgG2a	120		512	4			32	32

^aThe numbers in the subheads refer to the indicated species: 1, R. rickettsii; 2, R. felis; 3, R. australis; 4, R. akari; 5, R. montanensis; 6, R. honei; 7, R. japonica; 8, R. typhi; 9, R. canadensis; 10, R . prowazekii; R. sibirica; 12, R. rhipicephali; 13, R. massiliae; 14, R. helvetica; 15, Bar 29; 16, R. conorii Israeli; 17, R. bellii; 18, R. aeschlimannii (MC16); 19, R. conorii 7; 20, R. conorii Indian; 21, R. parkeri; 22, R. africae; 23, “R. mongolotimonae”; 24, R. conorii Astrakhan; 25, R. slovaca; 26, Thai tick typhus rickettsia; 27, R. conorii Moroccan; 28, strain S.

^bNT, not typeable.

FIG. 4.

Immunoblotting of boiled R. felis antigen treated with 11 representative MAbs. Lane 1, F14C9H5H5; lane 2, F20A9H5G8; lane 3, F19F8A4A8; lane 4, F11E9A2G1; lane 5, F11A3H1B11; lane 6, F24cH3C2C3; lane 7, F24cD9; lane 8, F24A2; lane 9, F20F9; lane 10 F23H2H4; and lane 11, F18C4.

Antigenic relationship revealed by reactivities with other MAbs.

Three of 64 MAbs, which were derived from the other seven species tested in this study, displayed cross-reactivities with R. felis (Table 2) (10, 32-35). Use of an MAb to R. conorii Malish 7, MAb RC10-A3, demonstrated that one epitope on the 120-kDa protein was shared by R. conorii and R. felis, as well as most SFG rickettsiae except R. akari, R. australis, and R. helvetica. MAbs to R. africae and R. akari, named AF6-D6 and K2-B41, respectively, both showed weak reactivities with R. felis. An MAb generated by immunization with R. africae, MAb AF6-D6, reacted with most of the SFG rickettsiae but did not react with R. akari, R. australis, R. aeschlimannii, R. bellii, R. helvetica, or R. montanensis. MAb K2-B41 was reactive with an epitope of R. akari and was also detected in R. felis and R. australis. MAbs to R. prowazekii and R. typhi did not react with R. felis.

DISCUSSION

R. felis, a recently described rickettsial species, has been demonstrated to be a flea-borne bacterium pathogenic for humans that is transmitted in fleas horizontally and vertically (3-6, 19, 23, 36). This agent has been observed in several species of fleas including C. felis and Pulex irritans in the United States, Ethiopia, Spain, and Brazil (5, 6, 17, 19, 28, 31) and France (unpublished data). The pathogenic role of R. felis in humans was demonstrated in studies with patients from Texas (28), Mexico (36), Brazil (23), France (23), and Germany (25) and by PCR and/or serologic tests (36). Although the prevalence of this infection is not established, it is suspected that R. felis has a worldwide distribution in fleas (5, 23). In this study, as reported previously (15), antisera to R. felis have low cross-reactivities with R. rickettsii, R. conorii, and R. typhi, which are the only commercially available antigens. It seems unlikely that R. felis infection could be detected by use of these antigens, and therefore, it is suggested that for the diagnosis of R. felis infections, it may be necessary to include a specific antigen for serologic assays.

Until now, 29 Rickettsia species or serotypes had been identified, and these were divided into two groups, namely, the SFG and the TG (20). Attempts to define the classification of R. felis have proved problematic (8, 15). Following the first isolation attempt it was considered a typhus-like rickettsiae that was not distinguished from R. typhi by serologic reagents. The Coomassie blue-stained SDS-PAGE profiles of this isolate and two R. typhi strains were identical (5, 12, 21, 22). Immunoblotting analysis of this bacterium with typing sera against R. typhi and its reactivity with MAbs previously thought to be specific for R. typhi (6) indicated that it is closely related to R. typhi (12, 18). Rat polyclonal antisera against this isolate reacted at a titer of 1:4,096 to cultured “R. felis” and had lower reactivities to R. typhi Wilmington (1:1,024), R. akari MK(Kaplan) (1:512), and R. australis JC (1:64) (21, 22). The sequence of the 17-kDa PCR product of the rickettsia-like agent resembled those of TG rickettsiae more than those of SFG rickettsiae (6). The data presented above support the evidence that R. felis clustered into TG. Subsequently, the same isolate was considered contaminated with R. typhi (21).

In the present study, SDS-PAGE and immunoblotting with murine polyclonal antisera against this agent supported the hypothesis that R. felis is phenotypically closely related to the SFG rickttsiae. Serotyping by the MIF assay demonstrated its intensive antigenic relations with the SFG rickettsiae rather than the TG rickettsiae. The use of murine antisera to R. felis showed that it clustered with R. australis, R. akari, and R. montanensis. Antisera to the aforementioned SFG species also cross-reacted with R. felis. The numerical taxonomic analysis, based on the traditional phenotypic criteria, clearly separated Rickettsia into two groups, the TG and the SFG rickettsiae, and placed R. felis into the SFG species.

MAbs against some rickettsial antigens have been used to study the antigenic relationships among the rickettsiae (33, 35). Antigenic analysis of MAbs against R. felis and other seven Rickettsia species examined in the present work provided further evidence that this newly described species belongs to the SFG and clusters with R. australis, R. akari, and R. montanensis.

Moreover, in our experience R. felis is a temperature-dependent rickettsia and cannot not be cultivated at 37°C, which is the optimal temperature for most bacterial organisms (15, 23). On the basis of these data, we believe that the first phenotypic description of this species, which is related to R. typhi, is not reliable and may reflect contamination with R. typhi in the earlier studies.

In conclusion, we have reported that R. felis is antigenically related to the SFG rickettsiae but has a low level of cross-reactivity, which possibly makes it necessary to evaluate specimens for R. felis when they are tested for rickettsial diseases. We did not find cross-reactions involving LPS between R. felis and other rickettsiae. We were not able to generate an anti-LPS MAb to R. felis. This is very usual and may indicate that R. felis is responsible for a specific immune response that is different from those of the other rickettsiae.