'Candidatus Rickettsia asemboensis' and Wolbachia spp. in Ctenocephalides felis and Pulex irritans fleas removed from dogs in Ecuador.

BACKGROUND: Flea-borne infections are distributed worldwide. Up to date there are no reports about microorganisms associated to fleas in Ecuador.
METHODS: Seventy-one Pulex irritans and 8 Ctenocephalides felis fleas were removed from dogs in two Ecuadorian areas (Pastaza and Chimborazo Provinces) in December 2012. DNA extracts were tested by polymerase chain reaction (PCR) assays targeting universal 16S rRNA, as well as screened for the presence of Rickettsia spp. (gltA, htrA, ompB, sca4 and ompA genes) and Bartonella spp. (rpoB, gltA and ITS genes).
RESULTS: Our results showed the presence of 'Candidatus Rickettsia asemboensis' (highly similar to R. felis) in C. felis and Wolbachia spp. endosimbionts in P. irritans collected from animals in Ecuador. No fleas were found to be positive for any Bartonella species or Yersinia pestis.
CONCLUSIONS: Clinicians should be aware of the potential risk of this new Candidatus Rickettsia sp. and keep in mind other flea-borne infections since these flea species frequently bite humans.

Background

Flea-borne diseases are worldwide-distributed emerging and re-emerging infections. Among them, plague, which is caused by Yersinia pestis, is the most severe human infection transmitted by fleas [1]. In South America, permanent plague foci exist among native rodent and flea populations in Bolivia, Brazil, Ecuador and Peru [2]. Rats have been the responsible hosts and from them, the disease has spread to other rodents. Ecuador is considered a plague ‘hot-spot’ since its introduction in 1908, and has experienced important outbreaks. Chimborazo Province has historically been a highly endemic area and the last fatal Ecuadorian cases of plague were reported there in 2004 [3,4].

In addition, fleas are vectors of murine typhus (caused by Rickettsia typhi), flea-borne spotted fever (caused by Rickettsia felis) and harbour Bartonella spp. [1,5-7]. Recent evidence of murine typhus in Ecuador is lacking, but the disease may be endemic in localities where commensal rodents (Rattus spp.) are abundant. To the best of our knowledge, data about distribution of R. felis in this country are unknown and there are no reports describing Bartonella spp. in fleas from Ecuador. For these reasons, our interest was focused on the study of flea-borne agents in two Ecuadorian areas (one of them where the last plague outbreak occurred) using molecular biological methods [polymerase chain reaction (PCR) and DNA sequencing].

Methods

In December 2012, a total of 79 fleas were removed from dogs by members of Red Iberoamericana para la Investigación y Control de las Enfermedades Rickettsiales, Programa Iberoamericano de Ciencia y Tecnologías para el Desarrollo (RIICER, CYTED; no. 210RT0403). Fifty-two specimens (44 Pulex irritans and 8 Ctenocephalides felis) were collected in the Estación Experimental Fátima (ESPOCH), Cantón Puyo, in Pastaza Province (01°24’34.6”S; 77°59’57.5”W), and 27 specimens (all P. irritans) were collected in Cantón Guamote, in Chimborazo Province (02°00’25.0”S; 78°47’09.3”W). The study areas had altitudes of 1,034 m and 3,657 m, respectively. After identification at the species level, samples were kept in 70% ethanol at room temperature before being tested. DNA of each arthropod was extracted by lysis with 0.7 M ammonium hydroxide and tested by PCR with the universal primers fD1 and rp2 [8]. This primer pair amplifies the main part of the 16S rRNA gene and has been used for the identification of Y. pestis as the causative agent of plague in India [9]. Samples were also screened for the presence of Rickettsia spp. with PCR assays targeting rickettsial citrate synthase (gltA) and 17 kDa antigen (htrA) genes [10,11]. In accordance with the taxonomic scheme [12], additional rickettsial genes (ompB, sca4 and ompA) were tested to properly identify Rickettsia-positive specimens [13-17]. Moreover, Bartonella spp. was tested using RNA polymerase β-subunit–encoding gene (rpoB), gltA and intergenic spacer region gene (ITS) PCR primers, which amplify fragments of Bartonella genes [18-20].

Each PCR included positive controls consisting of Bartonella henselae DNA extracted from a cat flea (C. felis) from La Rioja - Spain, or Rickettsia slovaca strain S14ab DNA (obtained from Vero cells inoculated in our facility with a Dermacentor marginatus tick from La Rioja - Spain, and known to be infected with R. slovaca). Negative controls (DNA-free water) were included in all assays. Sequences generated by each pair of primers were then compared with those in GenBank using BLAST (www.ncbi.nlm.nih.gov/blast/Blast.cgi).

Results and discussion

PCR assays using universal eubacterial primers for 16S rRNA gene yielded amplicons of different intensity for 69 out of 79 fleas (8/8 C. felis and 35/44 P. irritans from Pastaza and 26/27 P. irritans from Chimborazo). All C. felis (n = 8) were also found to be infected with Rickettsia species using gltA and htrA as rickettsial PCR targets, whereas no evidence of Rickettsia spp. was found in P. irritans. Moreover, no sample was positive for Bartonella species as determined either by rpoB, gltA or ITS PCR assays. Positive and negative controls worked as expected in all cases.

Sequences of rickettsial 16S rRNA gene obtained from 7/8 C. felis (1303–1373 bp) showed the closest identity (99.6-99.9%) with ‘Candidatus Rickettsia asemboensis’, a potentially new Rickettsia species according to the established criteria [12,21] (Table 1). In these samples, the percentage of identity with 16S rRNA gene of a validly published Rickettsia species reached 99.4% for R. felis (accession no. NR074483). The 1387 bp-long sequence of 16S rRNA gene obtained from the remaining C. felis had the highest identity (97.1%) with a sequence from uncultured flea-associated bacterium [22], and showed 93.6% identity with Snodgrasella alvi, betaproteobacteria classified in the family Neisseriaceae and previously isolated from the bee gut [23] (Table 1).

Table 1

Results of nucleotide sequence analysis corresponding to PCR products amplified from the 8 specimens of this study

Gene*	Flea gender 1	Location	Length of sequence (bp)	% identity	Bacteria with closest identity in the BLAST search	GenBank accession no.
16S rRNA	1/2 M; 6/6 F	Pastaza	1303-1373	99.6-99.9	‘Candidatus Rickettsia asemboensis’ (Rickettsia F30)	JN315967
	1/2 M; 0/6 F	Pastaza	1387	97.1	Uncultured flea-associated bacterium	EU137419
				93.6	Snodgrasella alvi	JQ746645
gltA	2/2 M; 6/6 F	Pastaza	350	100	Rickettsia felis-like (Rickettsia RF2125)	AF516333
				99.7	‘Candidatus Rickettsia asemboensis’ (Rickettsia F30)	JN315968
htrA	2/2 M; 6/6 F	Pastaza	394	100	‘Candidatus Rickettsia asemboensis’ (Rickettsia F30)/Rickettsia felis-like (Rickettsia SE313/cf1and5)	JN315969/DQ166937/AY953286
ompB	2/2 M; 6/6 F	Pastaza	464	100	‘Candidatus Rickettsia asemboensis’ (Rickettsia F30)	JN315972
sca4	2/2 M; 6/6 F	Pastaza	352	100	‘Candidatus Rickettsia asemboensis’ (Rickettsia F30)	JN315970

*Inconclusive sequences were obtained for ompA gene.

1Number of positive specimens / Total number of specimens from each flea gender; M: Male; F: Female.

Sequencing of gltA fragments (350 bp) identified R. felis-like rickettsiaceae in C. felis (100% identity with Rickettsia sp. genotype RF2125) [24] and showed 99.7% identity (349/350 bp) with ‘Candidatus R. asemboensis’ (Table 1). The sequences of the 17-kDa amplicons (394 bp) were homologous (100% identical) to each other and to ‘Ca. R. asemboensis’ as well as to other molecular isolates included in the R. felis-like genotype group (Rickettsia sp. SE313 detected in Echidnophaga gallinacea from Egypt and Rickettsia sp. cf1and5 detected in C. felis from USA) (Table 1).

The sequences of ompB and sca4 amplicons (464 and 352 bp, respectively) were also 100% identical to ‘Ca. R. asemboensis’ (Table 1). Unfortunately, ompA PCR primers did not yield amplicons of the expected size, and inconclusive sequences were obtained for this target gene.

Percentages of identity with validated species R. felis (accession no. CP000053) were 98.6, 96.2, 97.9 and 96.6% for gltA, htrA, ompB and sca4 genes, respectively.

In addition, 59 out of 61 sequences of 16S rRNA gene obtained from 61 P. irritans specimens provided evidence of the presence of probable endosymbionts similar to those found within other arthropods and belonging to the genus Wolbachia (Table 2). Unfortunately, in two cases (corresponding to two P. irritans samples from Chimborazo) it was not possible to get a good-quality sequence to identify the bacteria.

Table 2

Results of nucleotide sequence analysis corresponding to 59 PCR products amplified from the 71 of this study (44 specimens from Pastaza and 27 from Chimborazo)

Gene	Flea gender 2	Location	Length of sequence (bp)	% identity	Bacteria with closest identity in the BLAST search	Genbank
16S rRNA	9/16 M; 10/28 F	Pastaza	1287-1386	98.3-100	Wolbachia sp. wRi endosymbiont of Drosophila simulans	NR074437
	11/13 M; 13/14 F	Chimborazo
	4/16 M; 2/28 F	Pastaza	1338-1383	98.4-98.7	Wolbachia sp. endosymbiont of Pseudolynchia canariensis	DQ115538
	0/16 M; 2/28 F	Pastaza	1300-1372	98.3-99.3	Wolbachia sp. endosymbiont of Gryllus crickets	U83094
	0/16 M; 1/28 F	Pastaza	1305	99.3	Wolbachia sp. endosymbiont of Gryllus ovisopis	U83093
	0/16 M; 1/28 F	Pastaza	1374	98.3	Wolbachia sp. endosymbiont of Curculio hachijoensis	AB746399
	0/16 M; 6/28 F	Pastaza	1291-1368	98.3-99	Wolbachia sp. endosymbiont of Kleidocerys resedae	JQ726770

1In two cases we did not obtain enough good-quality sequences to identify the bacteria, and we did not obtain amplicons for ten specimens.

2Number of positive specimens / Total number of specimens from each flea gender; M: Male; F: Female.

In our study ‘Candidatus R. asemboensis’ has been found in fleas that bite humans (C. felis) removed from dogs in Ecuador. This potential new species was previously detected in Ctenocephalides canis from Kenya and whether it is a human pathogen remains unknown [21]. ‘Ca. R. asemboensis’ is highly similar to Rickettsia RF2125, a member of the R. felis-like genotype group that circulates in fleas from Uruguay [25]. Apart from epidemic typhus (caused by Rickettsia prowazekii and transmitted by lice), no data about human diseases associated with Rickettsia species have been published from Ecuador [26]. Nevertheless, in the Pastaza province (one of our sampling areas) cases of acute undifferentiated febrile illness compatible with rickettsioses have been reported [27].

Up to the present study, the presence of R. felis and/or Bartonella spp. has not been demonstrated in fleas from Ecuador. However, R. felis has been found in South American fleas from Brazil, Peru, Uruguay, Chile, Argentina and Colombia [11,25,28-31]. In addition, human infection with R. felis in South America has been confirmed in Brazil by molecular methods [32], and human serological evidence of R. felis infection has been recently reported in Colombia [33]. Moreover, there are limited reports describing Bartonella spp. in fleas from South America. A molecular study conducted in a Pulex specimen found on a Peruvian person evidenced the presence of a potential new Bartonella species [5]. Years later, our research group detected B. rochalimae, B. clarridgeiae, and B. henselae in P. irritans and C. felis collected from cats and dogs in Chile, suggesting the role of fleas as possible vectors of Bartonella spp. [7]

Lastly, Wolbachia spp. are alphaproteobacteria included in the family Anaplasmataceae that were first detected in fleas in 2000 [34]. In this study, the detection rate of Wolbachia spp. in P. irritans was 83% (59/71). On the contrary, no C. felis analysed (0/8) showed evidence of carriage of Wolbachia endosymbionts. Previous studies had identified Wolbachia in around 20% of cat fleas [35,36]. It has been suggested that R. felis infection in fleas might diminish the richness of flea microbiota [37]. According to our data, there is a strong association of C. felis with ‘Ca. R. asemboensis’ whereas P. irritans is associated with Wolbachia spp. Nevertheless, the interaction of Wolbachia with R. felis or other related species, such as ‘Ca. R. asemboensis’ in fleas needs further investigation.

Based on 16S rDNA analysis, the presence of Y. pestis DNA has not been demonstrated in our fleas despite P. irritans having been previously described as vectors of Y. pestis in Ecuador [2] and plague outbreaks have been repeatedly reported in the Chimborazo region [3,4], where some flea specimens were collected. The rodent flea Xenopsilla cheopis, which is the main vector, does not exist in the inter-Andean region of Ecuador (2,500-4,000 m above sea level) possibly, due to very sudden changes in the climatic conditions [38].

Conclusions

In summary, our result confirms the presence of ‘Candidatus R. asemboensis’ and Wolbachia spp. in fleas removed from dogs in Ecuador. Clinicians should be aware of the potential risk of this new Candidatus Rickettsia sp. and keep in mind other flea-borne infections in areas where humans are exposed to fleas.