Multilocus Variable-Number Tandem-Repeat Analysis Scheme for Chlamydia felis Genotyping: Comparison with Multilocus Sequence Typing

Abstract

Chlamydia felis is an important ocular pathogen in cats worldwide. A multilocus variable-number tandem-repeat analysis (MLVA) system for the detection of tandem repeats across the whole genome of C. felis strain Fe/C-56 was developed. Nine selected genetic loci were tested by MLVA in 17 C. felis isolates, including the C. felis Baker vaccine strain, and 122 clinical samples from different geographic origins. Analysis of the results identified 25 distinct C. felis MLVA patterns. In parallel, a recently described multilocus sequence typing scheme for the typing of Chlamydia was applied to 13 clinical samples with 12 different C. felis MLVA patterns. Rare sequence differences were observed. Thus, the newly developed MLVA system provides a highly sensitive high-resolution test for the differentiation of C. felis isolates from different origins that is suitable for molecular epidemiological studies.

INTRODUCTION

Chlamydiaceae spp. are obligate intracellular Gram-negative bacterial parasites that have a worldwide distribution and cause various diseases in animals, including humans. According to the most recent taxonomy, the family Chlamydiaceae with its single genus, Chlamydia, contains nine species (18). Among them, Chlamydia felis is an important pathogen of cats (10). Conjunctival epithelium appears to be the major target of C. felis, although the microorganism has been detected in a variety of tissues and anatomical locations in cats (reviewed in reference 36). C. felis causes acute-to-chronic conjunctivitis, particularly in young cats, and is the most common cause of conjunctivitis in cats, detected in approximately 30% of clinical cases (7, 11, 28, 34, 35, 41). Feline chlamydiosis is typically characterized by unilateral or bilateral acute conjunctivitis and excretion of high levels of C. felis, peaking approximately 1 to 2 weeks postinfection (5, 23, 37, 40). Thereafter, chronic conjunctivitis and lower levels of C. felis excretion often persist for up to about 2 months postinfection prior to the resolution of clinical signs and cessation of shedding (23, 40). However, some cats may remain persistently infected for longer time periods and may represent a population of asymptomatic carriers (23, 36, 40). Among other clinical signs, C. felis infection has been associated with vaginal discharge and excretion in experimentally infected cats and the organism has been suspected as a cause of abortion and reproductive disease in queens; however, the limited evidence for this is largely circumstantial (36).

Chlamydiosis is relatively common in the domestic cat population, and the infection may become endemic, especially in cat colonies (6). Transmission of C. felis between cats occurs primarily by direct contact with infective ocular secretions. Rescue shelters may therefore represent high-risk environments for the transmission of chlamydiosis, as a result of large numbers of animals being housed in limited spaces and the periodic introduction of new, susceptible cats, particularly kittens.

The zoonotic potential of this bacterium appears to be low; however, exposure to C. felis may occur when handling infected cats. Hence, cat owners and professionals who work with cats (veterinarians, employees at catteries, and cat breeders) are at increased risk, especially where there are insufficient hygienic conditions. The risk of zoonotic transmission is presumably greater for immunocompromised individuals. C. felis infection has been associated with conjunctivitis and/or respiratory tract disease (4, 15), community-acquired pneumonia (21), hepatosplenomegaly, glomerulonephritis, and endocarditis in humans (31).

Until now, genotyping studies conducted with C. felis have shown low genetic diversity among isolates of this species, with a substantial degree of rRNA and ompA gene conservation (29, 33). The 16S rRNA genes of C. felis strains that have been sequenced to date differ by less than 0.6% (8). However, heterogeneity of feline strains related to geographic origin was demonstrated by restriction fragment length polymorphism of the groEL gene (9). Also, analysis by random amplification of polymorphic DNA revealed at least two different genetic fingerprints among the six C. felis strains examined (30).

With the increasing availability of whole bacterial genome sequences, new molecular tools based on genome-wide screens are currently being developed. The typing method termed multiple-locus variable-number tandem-repeat (VNTR) analysis (MLVA) is based on the detection of tandem-repeat polymorphisms (38) and has been used successfully for the typing of many pathogens, including Chlamydia trachomatis (27), Chlamydia psittaci (19), and Chlamydia abortus (20). The typing method termed multilocus sequence typing (MLST) is often based on the sequence analysis of internal fragments of seven housekeeping genes. A previously described MLST scheme has been successfully used to understand the population genetic structures of C. trachomatis, C. pneumonia, and C. psittaci (24, 25).

The aims of the present study were to develop a novel MLVA scheme specific for C. felis genotyping and to compare it with the previously described MLST typing method.

MATERIALS AND METHODS

Bacterial isolates and clinical samples.

DNA extracted from 122 clinical samples (121 conjunctival swabs and 1 placental swab) and 17 C. felis isolates (16 field isolates and the Baker vaccine strain) (Table 1) were obtained from ANSES (Maisons-Alfort, France), Scanelis (Colomiers, France), the Department of Veterinary Medical Sciences, University of Bologna (Ozzano dell' Emilia, Italy), the Clinic for Swine (Giessen, Germany), the National Veterinary Institute of Sweden (Uppsala, Sweden), and the University of Bristol (Bristol, United Kingdom). Samples were collected from both eyes of some Italian cats and analyzed.

Table 1.

C. felis isolates and samples organized by MLVA genotype^a

MLVA pattern	No. of different patterns									Country	Clinical sample(s)/isolate(s)b	No. of samples/isolates
	ChlaFe_115	ChlaFe_191	ChlaFe_257	ChlaFe_500	ChlaFe_532	ChlaFe_619	ChlaFe_699	ChlaFe_715	ChlaFe_721
1	2	2	2	2	3	2	1	4	4	Italy	10_rightd, 10_leftd	2
2	2	2	2	2	3	2	2	3	4	France	11E3-09	4
										Sweden	07-MTB002765, 07-MTB001512, 07-MTB001513
3	2	2	2	2	3	2	2	4	4	France	11E3-05, 11E3-14, 11E3-15, 11E3-23, 11E3-25, 11E3-32, 11E3-35, 11E3-37, 11E3-38c	25
										Italy	2_lefte, 3e, 4_righte, 4_lefte, 8_rightd, 8_leftd, 9_leftd, 14f, 17_rightf, 17_leftf, 20_left, 24, M
										Sweden	07-VIR000369, 07-MTB001076
										United Kingdom	1910
4	2	2	2	2	3	2	2	5	3	Sweden	07-VIR043754	1
5	2	2	2	2	3	2	2	5	4	France	11E3-04, 11E3-07, 11E3-18, 11E3-28, 11E3-29	7
										United Kingdom	12159, Cello
6	2	2	2	3	3	2	2	3	4	France	11E3-08	2
										United Kingdom	6110
7	2	2	2	3	3	2	2	4	4	France	11E3-13	17
										Italy	1e, 2_righte, 5e, 6e, 22, A, G, Mer, P
										Sweden	07-MTB002107
										United Kingdom	1781, 6043, 6225, 6445, 6881, 12114
8	2	2	2	3	3	2	2	5	4	United Kingdom	1497, 9918	2
9	2	2	2	3	3	2	2	6	4	United Kingdom	6910	1
10	2	2	2	4	3	2	2	4	3	Germany	Z 3422/92	1
11	2	2	2	4	3	2	2	4	4	France	11E3-10, 11E3-16, 11E3-17, 11E3-21, Baker	19
										Italy	7_rightd, 7_leftd, 25, 26, H, R, Z
										Sweden	07-MTB002869, 07-MTB002870, 07-MTB001556
										United Kingdom	1534, 6292, 9036, 6689
12	2	2	2	4	3	2	2	5	4	France	11E3-22	5
										Italy	9_rightd
										United Kingdom	1787, 9603, 12285
13	2	2	2	5	3	2	2	3	4	United Kingdom	9888	1
14	2	2	2	5	3	2	2	4	2	France	11E3-02	1
15	2	2	2	5	3	2	2	4	4	France	11E3-03, 11E3-24, 11E3-31, 11E3-34	21
										Germany	Z 3202-I/91, Z 2930/91
										Italy	11f, 12f, 13_rightf, 13_leftf, 15f, 16_rightf, 16_leftf, 18_rightf, 18_leftf, 19_rightf, 19_leftf, 21, L
										United Kingdom	6327, 6218
16	2	2	2	5	3	2	2	5	4	United Kingdom	1993, 1612	3
										Sweden	07-MTB3028
17	2	2	2	6	3	2	2	3	4	France	99-9537	2
										Germany	Z 3405/92
18	2	2	2	6	3	2	2	4	4	France	11E3-06, 11E3-12, 11E3-20, 11E3-26, 11E3-30, 11E3-36	11
										Sweden	07-MTB001055, 07-MTB001056
										United Kingdom	9607, 9710, 9696
19	2	2	2	6			2	5	4	Germany	Z 2095/88	1
20	2	2	2	7	3	2	2	3	4	France	07-1099, 11E3-01	2
21	2	2	2	7	3	2	2	4	3	France	07-399	1
22	2	2	2	7	3	2	2	4	4	France	11E3-11, 11E3-27, 11E3-33	5
										Sweden	07-VIR022949, 07-VIR039188
23	2	2	2	8	3	2	2	4	4	France	11E3-19	3
										Sweden	07-MTB002098
										United Kingdom	6704
24	2	2	2	9	3	2	2	2	4	Sweden	07-MTB000246	1
25	2	2	2	9	3	2	2	3	4	Sweden	07-VIR052311	1

^aGenotypes were based on similarity clustering of PCR product sizes obtained with the nine primer sets used. The genotypes are numbered 1 through 25.

^bIsolates and the vaccine strain are underlined, and samples analyzed by MLST are in bold.

^cPlacental swab (11E3-38).

^dFrom cattery B.

^eFrom cattery A.

^fFrom cattery C.

Italian clinical samples were collected in 2000 and 2001 at three different catteries (6). German, Swedish, and United Kingdom (13) samples were collected at various times from 2005 to 2007, and French samples were collected in 2011 as part of routine diagnostic analysis.

Both clinical samples and C. felis isolates had been previously subjected to DNA extraction, and all of the DNA included in this study tested positive with the C. felis-specific real-time PCR assay described by Pantchev et al. (26). The protocol included primers CpfOMP1-F (5′-TCGGATTGATTGGTCTTGCA-3′) and CpfOMP1-R (5′-GCTCTACAATGCCTTGAGAAATTTC-3′) and probe CpfOMP1-S (6-carboxyfluorescein–5′-ACTGATTTCGCCAATCAGCGTCCAA-3′–6-carboxytetramethylrhodamine). Each reaction mixture (20 μl) contained 2 μl of sample template DNA, 10 μl of 2× Universal Master Mix (Applied Biosystems), a 0.625 μM final concentration of each primer, a 0.1 μM final concentration of the probe, and 5 μl of deionized water. The temperature-time profile was 95°C for 10 min, followed by 45 cycles at 95°C for 15 s and 60°C for 60 s.

MLVA genotyping method.

Tandem-repeat sequences within the whole genome sequence of C. felis strain Fe/C-56 (GenBank accession no. AP006861) (1) were determined using the Tandem Repeats Finder program (http://tandem.bu.edu/trf/trf.html) (3). Nine tandem repeats with a repeat unit length of ≥12 bp were identified, and primers for PCR amplification on both sides of the repeats were chosen (Table 2). VNTR size was determined in order to allow efficient differentiation on agarose gel.

Table 2.

Primers for PCR amplification and MLVA^a

VNTR primer	Sequence (5′–3′)		Tm (°C)	Repeat unit length (bp)	C. felis Fe/C-56 theoretic repeat no.	Size of C. felis Fe/C-56 amplicon (bp)
	Forward	Reverse
ChlaFe_115	TCATCAGCGATGAATTATTCAGAATAG	CAATGCCGCCGTAACATTT	58	23	1.9	220
ChlaFe_191	CCTCCCTGAAGAAGCACTGTTT	ATTTTACGCGAACAACGTTATCC	58	21	2.0	215
ChlaFe_257	TTTTAGAGAAAGATGCGGTAATGAAAG	TCCATGAGGTTGTGATTGCAA	59	21	1.9	143
ChlaFe_500	CGAGGTTCTCTCAATAATGCTCTAAACT	TTCTTCTCATCGGCACCAAAC	59	12	4.0	213
ChlaFe_532	GGCCCCCTCGCAAGG	CCCCACAACCCGACCA	58	17	2.6	160
ChlaFe_619	TTCTGAGTTCCCTTCAATAGCTTTTC	TTTGCGCGTGTTGGCAT	59	18	2.2	167
ChlaFe_699	AAGTAGACAGTGCTGGTTTTGCTACT	ACAAAAGAAACGTAGCAGCAAGA	57	30	2.3	470
ChlaFe_715	AAAATTGACACCAGCACCACC	GGTAGAGTCGTTAGTAATCTCGTGGTAG	58	18/9	2.0/4.0	171
ChlaFe_721	AGAGATTGTAATGCCTGTGGGTTC	TCCCCATTGGCTCCTGTG	59	111	4.0	572

^aThe primers were used to characterize tandem repeats at nine genetic loci.

All 17 isolates (including the Baker vaccine strain) and 122 clinical samples were subjected to MLVA. For VNTR amplification, a PCR was performed in a total volume of 15 μl containing 5 to 10 ng of DNA, 1× PCR buffer, 1 U of HotStarTaq DNA polymerase (Qiagen), a 200 μM final concentration of each deoxynucleotide, and a 0.3 μM final concentration of each flanking primer. The initial denaturation step of 95°C for 15 min was followed by 40 cycles consisting of denaturation at 95°C for 30 s, primer annealing at 58°C for 30 s, and elongation at 72°C for 45 s. The final extension step was 72°C for 10 min. Five microliters of amplification product was loaded onto a 3% agarose gel. Gels stained with ethidium bromide were visualized under UV light and photographed. The size marker used was a 100-bp ladder (MBI-Fermentas–Euromedex, Souffelweyersheim, France).

Amplicon size was determined using the Quantity One software package version 4.6.7 (Bio-Rad Laboratories, Inc., Hercules, CA). The number of motifs in each allele was derived from the amplicon size. For each discriminant VNTR marker, representative patterns were sequenced (MWG Biotech France, Roissy, France).

MLST genotyping method.

Based on previous work (25), we amplified and sequenced seven housekeeping genes (enoA, fumC, gatA, gidA, hemN, hflX, and oppA) from 13 randomly selected French clinical C. felis samples (11E3-2, 11E3-3, 11E3-4, 11E3-5, 11E3-6, 11E3-8, 11E3-9, 11E3-10, 11E3-11, 11E3-13, 11E3-19, 11E3-22, 11E3-38) (Table 1, samples in bold) and from the C. felis Baker vaccine strain.

Cell culture.

For cell culture, yolk sacs of 7-day-old embryonated eggs were inoculated with 0.2 ml of the C. felis suspension per egg and five eggs were used per sample. For each inoculation set, three eggs were inoculated with C. psittaci strain Loth as a positive control and three other eggs were kept separately as noninfected controls. Eggs were incubated at 38°C and observed daily. Vitellus membranes were collected from dead embryos and submitted to 10 serial passages in chicken eggs. DNA was extracted from collected vitellus membranes and analyzed by PCR and then MLVA as described above.

Nucleotide sequence accession numbers.

The sequences determined by MLST have been deposited in the GenBank database under accession numbers HE795868 to HE795965, and those determined by MLVA have been deposited under accession numbers HE795966 to HE795983.

RESULTS

Selection of VNTRs for MLVA.

Fragment size analysis of the amplified PCR products showed that four of the tandem repeats, namely, ChlaFe_500, ChlaFe_699, ChlaFe_715, and ChlaFe_721, were polymorphic, with eight, two, five, and three different patterns, respectively (Fig. 1; Table 1). Based on these four VNTR markers, C. felis isolates and/or samples were subdivided into 25 genotypes termed genotypes 1 through 25 (Table 1). Expected tandem-repeat sizes were observed, except for the marker ChlaFe_715, for which a 9-bp tandem-repeat unit was discovered rather than the 18-bp repeat identified by the software, which was set to detect tandem-repeat units with a minimum size of 12 bp. No geographic link between isolates or samples from different European countries was found (data not shown).

Fig 1.

C. felis MLVA patterns observed with different VNTR polymorphic markers. ChlaFe_500 showed eight patterns (2 [Cello strain], 3 [isolate A], 4 [Baker strain], 5 [isolate L], 6 [sample 07-MTB001055], 7 [sample 07-VIR022949], 8 [sample 07-MTB002098], and 9 [sample 07-MTB000246]), ChlaFe_699 showed two patterns (1.4 [sample 10_right] and 2 [sample 9607]), ChlaFe_715 showed five patterns (2 [sample 07-MTB000246], 3 [sample 07-1059], 4 [Baker strain], 5 [sample 07-VIRO43754], and 6 [sample 6910]), and ChlaFe_721 showed three patterns (2 [sample 11E3-02], 3 [sample 07-VIRO43754], and 4 [Cello strain]).

The stability of the four discriminant markers was assessed by comparison of the MLVA patterns of two strains (the Baker vaccine strain and the 99-9537 isolate) after 10 serial passages in chicken eggs. MLVA typing results were identical to those obtained initially (data not shown).

Focus on samples collected in three catteries.

Clinical samples from four, six, and nine cats belonging to three different Italian catteries were investigated. Samples were harvested during an epidemiological study performed in 2000 and 2001 (6). Most of the cats belonging to the catteries sampled had conjunctivitis for several months, and all of the clinical samples analyzed in the present study were from cats showing ophthalmic problems at the time of sampling.

Up to four different genotypes were identified within the same cattery. In particular, in cattery A, genotypes 3 and 7 were detected, whereas genotypes 1, 3, 11, and 12 were detected in cattery B and genotypes 3 and 15 were identified in cattery C (Tables 1 and 3).

Table 3.

MLVA patterns of C. felis strain samples from three Italian catteries where conjunctivitis was reported^a

Cattery	Sample no.	Corresponding cat no.	Sample no. (corresponding cat no.) with C. felis MLVA pattern:
			1	3	7	11	12	15
A	8	6		4 (3)	4 (4)
B	8	4	2 (1)	3 (2)		2 (1)	1 (1)
C	14	9		3 (2)				11 (7)

^aBoth eyes of some cats were swabbed.

In cattery A, two cats harbored genotype 3, three cats had genotype 7, and one cat had both genotypes 3 and 7. In cattery B, one cat harbored genotype 1, one cat had genotype 3, one cat had genotype 11, and one cat had both genotypes 3 and 12. In cattery C, two cats harbored genotype 3 and seven cats had genotype 15.

From a clinical sample from cattery A, one isolate (M) with MLVA genotype 3 was obtained. This is consistent with the finding of clinical samples with the same MLVA pattern from this cattery (Table 3).

Of the cats from both of whose eyes samples were collected, nine had the same genotype in both samples. In contrast, in two cats from two different catteries, a different genotype was identified in each eye. One of these cats (number 2), in cattery A, had genotype 3 in its left eye and genotype 7 in its right eye. These two genotypes, which differ by only one tandem repeat in marker ChlaFe_500, were also detected in samples from other cats in this cattery: three cats harbored genotype 3, and four cats harbored genotype 7. In this cattery, the same genotype was detected in both eyes of one cat (number 4, genotype 3).

A different genotype was also detected in each eye of cat number 9 (genotypes 3 and 12). Interestingly, whereas genotype 3 was identified in other cats in this cattery, genotype 12 was detected only in cat number 9. Genotypes 3 and 12 differ in two markers (ChlaFe_500 and ChlaFe_715), with two and one tandem-repeat differences, respectively.

Sequence analysis of PCR products.

PCR products generated for each of the four polymorphic markers (Fig. 1) were sequenced. Sequence analysis confirmed the fragment size analysis of the PCR products. Locus ChlaFe_500 is localized in a noncoding sequence. As expected, DNA sequencing of representatives of the eight different patterns identified different numbers of tandem repeats consisting of a 12-bp sequence (CCAACCTAGGAA) (frequencies of tandem repeats in ChlaFe_500: two tandem repeats, 28.1% of the samples analyzed; three tandem repeats, 15.8%; four tandem repeats, 18.0%; five tandem repeats, 19.4%; six tandem repeats, 10.1%; seven tandem repeats, 5.8%; eight tandem repeats, 2.2%; nine tandem repeats, 1.4%). All of the other markers are located in a coding region or in a hypothetical coding region. Locus ChlaFe_715 is located in a hypothetical-protein-encoding gene (CF0611) where the 9-mer GAAGCAGAA coding for the peptide EAE is repeated two to six times. In most (76.3%) of the samples analyzed, this 9-mer was repeated four times. Notably, CF0611 has no homology with other chlamydial proteins described to date. A particular pattern was observed for the locus ChlaFe_699, which was detected only in clinical samples collected from both eyes of a single cat (the same pattern in samples from both eyes). DNA sequence analysis of PCR products revealed a 60-bp deletion outside the tandem-repeat sequence region. The deleted sequence which corresponds to the peptide VVFLTVLRAVVFLATVLRTV is part of the histone H1-like protein Hc2. Sequencing of the locus ChlaFe_721 revealed the presence of two different tandem-repeat sequences in this amplified fragment. These two tandem-repeat sequences are each repeated two times in C. felis reference strain Fe/C-56. Including the reference strain pattern, three new patterns were identified, one with one deletion in the first tandem-repeat sequence and another with one deletion in each of the tandem-repeat sequences. The deletions are contained in a hypothetical-protein-encoding gene (frequencies of tandem repeats in ChlaFe_721: two tandem repeats, 0.7%; three tandem repeats, 2.2%; four tandem repeats, 97.1%).

MLST.

The clinical samples selected from large (multitudinous) MLVA C. felis patterns 2, 3, 5, 6, 7, 11, 12, 14, 15, 18, 22, and 23 were successfully amplified by the MLST scheme. Identical sequences (ST_21) were identified within this clinical sample panel for the seven targeted housekeeping genes, except for one point mutation in the hflX gene in sample 11E3-4. In addition, two nonsynonymous mutations were detected in the enoA gene of the C. felis Baker vaccine strain.

DISCUSSION

Previous studies based on 16S rRNA, ompA, or groEL gene sequences (8, 9, 29, 33) have highlighted the remarkable homogeneity of C. felis strains. Here we present an MLVA-based molecular typing system for the discrimination of C. felis isolates that does not require culture and can therefore be performed directly with DNA extracted from clinical specimens. MLVA typing is a reproducible method which can be standardized and performed at low cost, thus facilitating large-scale molecular epidemiological investigations. Out of nine tandem-repeat loci identified in the C. felis Fe/C-56 genome by the program Tandem Repeat Finder, four were polymorphic. The analysis of about 140 C. felis clinical samples and isolates resulted in a total of 25 distinct patterns designated genotypes 1 through 25. High genotypic diversity was demonstrated in samples from each of the European countries of origin; however, four genotypes, 3, 7, 11, and 15, appear more widespread both in the samples tested and in the countries of origin. As transmission of C. felis relies on close contact between infected and susceptible cats, significant international spread of C. felis is unlikely. The discovery of broadly comparable C. felis genotypic profiles in each country suggests that C. felis populations in different geographic locations undergo similar evolutionary divergence patterns in VNTR loci. Notably, variation arising in the most polymorphic VNTR marker, ChlaFe_500, is responsible for the majority of the genotypic differences, particularly between dominant genotypes. Thus, the potential for homoplasy may be relatively high, which could have an impact on the discriminatory power of C. felis genotyping by MLVA in molecular epidemiology studies (32).

Interestingly, the Baker vaccine strain, which was first isolated from cats with pneumonia (2), was classified as genotype 11. Therefore, MLVA could provide the means to differentiate between vaccinated and naturally infected animals if local strains were of a genotype different from that of the locally available vaccine strains. Isolates and/or samples with genotype 11 were obtained from multiple countries; however, no information is available concerning the vaccination history of the sampled animals included in this study. Thus, it remains undetermined if these field isolates represent the vaccine strain or circulating field strains with a similar MLVA genotype.

Two of the VNTR markers identified in the present study (ChlaFe_715 and ChlaFe_721) were located within open reading frames encoding putative proteins (CF0611 and CF0618, respectively). A further marker (ChlaFe_699) was located in hctB (CF0595), which encodes a histone-like protein involved in the establishment of the nucleoid structure of chlamydial elementary bodies. Of the five polymorphic VNTR markers for C. abortus, one of them was located within this same family of genes (20). In C. trachomatis, Hc2 (encoded by hctB) varies in size between strains due to internal deletions encoding lysine- and alanine-rich pentameric repeats (12). Different numbers of repetitive elements and point mutations were recently detected in this gene in genital clinical samples, leading to high-resolution intraserotype variation in C. trachomatis (17). In the present study, hctB differed in only one clinical sample with a 60-nucleotide deletion corresponding to a 20-mer peptide. Polymorphic sites, including those within hctB, were shown to be responsible for the pathogenic diversity of C. trachomatis trachoma strains (16). The biological significance of the coding tandem repeats identified in two putative proteins (CF0611 and CF0618) and the deletion in the hctB gene (CF0595) is unknown. Coding tandem repeats are believed to enhance specific properties of bacterial virulence factors. Potential differences in pathogenicity between strains/isolates of C. felis remain largely uncharacterized, and few studies have investigated virulence or genotypic differences (22, 14). MLVA genotyping may, however, provide a useful molecular method for identifying strains from different lineages for investigation of phenotypic diversity and differences in clinical behavior. In the present study, it was not possible to determine if C. felis isolates with different MLVA genotypes might be associated with variations in disease pathogenesis or clinical outcomes, as detailed clinical information from individual cases was not available, except for cats observed within the Italian catteries. In the latter cases, cats infected with different C. felis MLVA genotypes, including the cat bearing the C. felis isolate with hctB deleted, showed broadly similar clinical signs typical of C. felis infection.

Investigations conducted in catteries revealed the diversity of C. felis strains that can be harbored by cats within a confined environment. Interestingly, in samples from two cats in two different catteries, where both eyes were sampled, two different genotypes were identified, which in one case differed by only one tandem repeat. These findings provide the first evidence that mixed infections with strains of different C. felis genotypes occur in cats, similar to those described for C. trachomatis in humans (39). The possibility that the different genotypes identified in these two cases may represent divergent populations resulting from instability in the MLVA loci cannot be excluded; however, analysis of two C. felis isolates passaged 10 times through chicken eggs suggests that the MLVA genotype is highly stable. Nevertheless such in vitro methods may not adequately reflect in vivo selective pressures from factors such as host immune responses. Future experiments analyzing the stability of MLVA genotypes during the course of C. felis infection and antibiotic treatment and studies of C. felis superinfection with two or more MLVA genotypes in individual cats and in cat communities would be of considerable interest.

MLST based on the partial sequences of seven housekeeping genes was recently used to understand the population genetic structure of C. trachomatis, C. pneumoniae (24), and C. psittaci (25). Results of cluster analyses of concatenated sequences of the seven housekeeping genes validated the proposed typing system for all chlamydial species. In our study, 12 conjunctival samples with 12 different C. felis MLVA patterns, as well as a unique sample collected from abortive feline placental tissue, and the Baker vaccine strain were submitted to the MLST scheme. Sequences were identical, with only two mutations detected in the enoA gene of the Baker vaccine strain and one point mutation detected in the hflX gene in a clinical sample. Almost all of the clinical samples were of the same sequence type (ST_21) as the C. felis Fe/C-56 reference strain (http://pubmlst.org/chlamydiales/), highlighting the strong conservation of these genes in C. felis, except for one mutation identified only in the Fe/C-56 hemN gene. MLVA is based on differences in the numbers of tandem repeats at multiple loci on the chromosome of bacteria, whereas MLST is based on the sequence analysis of housekeeping genes. In this case, the discriminatory power of MLVA was superior to that of MLST. No specific pattern was identified by either MLVA or MLST in the single C. felis abortion tissue sample examined or in the conjunctival samples collected. However, before any formal conclusion can be drawn, it is necessary to analyze a larger cohort of C. felis samples recovered from cases of feline abortion. Interestingly, by MLST, the Baker strain could be differentiated from others strains by two nonsynonymous mutations in the enoA gene; however, the biological significance of these mutations remains unknown.

In summary, our findings show that MLVA is a novel genotyping method providing enhanced discrimination of C. felis strains. Preliminary work conducted with clinical samples showed that MLVA is more discriminant than MLST for C. felis genotyping.