A prevalent alpha-proteobacterium Paracoccus sp. in a population of the Cayenne ticks (Amblyomma cajennense) from Rio de Janeiro, Brazil

As Rocky Mountain Spotted Fever is the most common tick-borne disease in South America, the presence of Rickettsia sp. in Amblyomma ticks is a possible indication of its endemicity in certain geographic regions. In the present work, bacterial DNA sequences related to Rickettsia amblyommii genes in A. dubitatum ticks, collected in the Brazilian state of Mato Grosso, were discovered. Simultaneously, Paracoccus sp. was detected in aproximately 77% of A. cajennense specimens collected in Rio de Janeiro, Brazil. This is the first report of Paracoccus sp. infection in a specific tick population, and raises the possibility of these bacteria being maintained and/or transmitted by ticks. Whether Paracoccus sp. represents another group of pathogenic Rhodobacteraceae or simply plays a role in A. cajennense physiology, is unknown. The data also demonstrate that the rickettsial 16S rRNA specific primers used forRickettsia spp. screening can also detect Paracoccus alpha-proteobacteria infection in biological samples. Hence, a PCR-RFLP strategy is presented to distinguish between these two groups of bacteria.

Ticks are blood-feeding arthropods harboring several associated bacteria, viruses, fungi and protozoans. Although some of the bacteria are classified as either symbionts, or human pathogens, the biological functions of most of these, only recently described prokaryotes, are still unknown (Niebylski et al., 1997;Noda et al., 1997;Marin and Schmidtman, 1998;Marquez et al., 1998;Yparraguirre et al., 2007;Stromdhal et al., 2008;Machado-Ferreira et al., 2009. In South America, Amblyomma cajennense is considered the most important tick species, representing the main vector of the Rhodobacteraceae Rickettsia rickettsii, the etiological agent of Rocky Mountain Spotted Fever (Sexton et al., 1993;Lemos 2000;Galvão et al., 2003;Guedes et al., 2005). A. cajennense may also harbor several other bacteria, such as Coxiella, Francisella and other Rickettsia species, which may represent new agents of human diseases or symbionts (Sexton et al., 1993;Parola et al., 1998;Machado-Ferreira et al., 2009. Emerging infectious disease (EID) agents are comprised of pathogens that recently entered human popula-tions, most of these with still unknown reservoirs or associated arthropod vectors (Salyers and Whitt, 2002). In this scenario, the CDC Group EO-2 Paracoccus yeeii, recently isolated from patients in the US, Canada and Europe, could be classified as another Rhodobacteraceae EID agent, with a currently unidentified source or transmission route (Daneshvar et al., 2003;Funke et al., 2004).
In the present study, an attempt was made to detect Rickettsia spp. in Amblyomma ticks, collected in the Brazilian states of Rio de Janeiro, Minas Gerais and Mato Grosso. We showed that the use of 16S rRNA specific primers for screening Rickettsia DNA in biological samples, also proved to be functional in detecting tick-associated Paracoccus sp.. This species, by being highly prevalent in A. cajennense ticks collected in a sampling site in Rio de Janeiro, possibly represents another group of tick-borne alpha-proteobacteria.
A total of 213 actively feeding adult A. cajennense and 20 A. dubitatum ticks were collected from horses and a domestic dog in the cities of Sambaetiba, Rio do Ouro and Três Rios (Rio de Janeiro), Pouso Alto (Minas Gerais) and São José do Xingu (Amazonia region of Mato Grosso) between May and November, 2006 (Table 1). After washing engorged females four times in 70% (v/v) ethanol, individuals and pools of ten ticks were snap-frozen in liquid nitro-gen. After homogenization in 1 mL of a lysis solution (sucrose 5%, SDS 1%, 100 mM Tris pH 7.5, 600 mM NaCl and 100 mM EDTA), samples were incubated at 56°C for 3 h with 0.4 mg/mL of proteinase K and 0.1 mg/mL of ribonuclease A. The resultant DNA was purified by phenol-chloroform extraction and re-suspended in 200 mL of 1 mM Tris pH 7.4 and 0.1 mM EDTA. DNA integrity was analyzed by agarose gel electrophoresis. Tick morphological classification was confirmed by PCR amplification of mitochondrial 16S rRNA genes, using the primer set 16S+1/16S-1 (Black and Piesman, 1994) followed by sequencing and BLAST analyses.
The Rickettsia genus-specific primer set fD1-Rc16S.452n (Marquez et al., 1998;Simser et al., 2002) was used to amplify the 16S rRNA gene (16S rDNA). Presumptive Rickettsia positive samples were then characterized by PCR amplification by using the Spotted Fever group primer sets, Rr190.70p/Rr190.602n for ompA (Regnery et al., 1991;Simser et al., 2002), the Rickettsia-specific primer set named Rr120 for ompB (Gage et al., 1992;Noda et al., 1997), Rr17.61p/Rr17.491n for htrA (Williams et al., 1992;Simser et al., 2002), and RpCS.877p/RpCS.1258n for gltA (Regnery et al., 1991;Simser et al., 2002). PCR temperature/time cycling conditions were 35 cycles of 94°C for 1 min, 55°C for 30 s and 72°C for 1 min. PCR products from four independent reactions were then merged and directly purified with a GFX PCR DNA and Gel band purification kit (GE Healthcare, Buckinghamshire, UK), according to manufacturer's instructions. In the case of suspect spurious PCR products, bands of expected size were cut from agarose gels and used as PCR templates with the same primer set. Purified PCR products were sequenced using the BigDye Terminator DNA sequencing kit (Applied Biosystems, Foster City, Calif., U.S.A.), and analyzed in a Megabace 1000 DNA sequencer (Amersham Biosciences). Sequences were edited using the SeqMan program (DNASTARinc package for the Windows platform, 1989Windows platform, -1999, and gene identities obtained by BLAST analyses (blastn).
In the Major-94 group, DNA samples from the 55 A. cajennense ticks collected in Sambaetiba (RJ) yielded numerous spurious PCR bands. According to initial-replicon analysis, re-amplification and sequencing, they were not identical to either Rickettsia sp., or to any other bacterial DNA. From these apparently positive samples, 17 tick-bulk DNA samples were also PCR analyzed, in an attempt to detect ompA, ompB, htrA and gltA genes. No amplification was obtained, confirming that they were in fact free of Rickettsia spp. (data not shown). Four randomly selected DNA samples of A. cajennense collected at Três Rios (RJ), and pertaining to the same group, were also sequenced and analyzed. Although these nucleotide sequences (GenBank EF531228) were all identical, there was no similarity to any Rickettsia spp. sequence. Even so, there was a close relationship to soil and invertebrate-associated alphaproteobacteria Paracoccus spp., the latter mostly of the Paracoccus sp. SA5 strain (GenBank AY864654) and the soil P. alcaliphilus KTC002 (GenBank AJ294415), with 99% and 98% identity, respectively ( Figure 1A). Thus, there is every indication that an A. cajennense population at Três Rios is infected with a specific Paracoccus sp. population, denoted here as "Paracoccus sp-Cayenne". This bac- Most of the ticks were individually analyzed. Samples analyzed by pools are indicated in the parentheses (No. of pools X pool size).

3
The minimum infection rate was obtained by assuming that only one tick was infected. 4,5 Minimum number of Paracoccus sp. 4 or Rickettsia sp. 5 positive ticks / total number of ticks analyzed.
terial population shows 96% 16S rRNA gene identity to the clinical P. yeeii strain G3060 (GenBank AY014179). PCR products from a total of 38 Três Rios tick samples, positive for the~430 bp 16S rDNA amplicon were subjected to additional endonuclease digestion. A RFLP strategy was designed to specifically assign these amplicons to either the Rickettsia or the Paracoccus 16S rDNA group. After analyzing the 16S rRNA gene sequences presented in Figure 1A, we observed that every Rickettsia 16S rRNA gene sequence included in the primer set fD1-Rc16S.452n amplicon generated the same HinfI digestion profile, with the presence of 117 bp, 124 bp and 195 bp bands, whereas all the Paracoccus sequences generated a profile with bands of 25 bp, 100 bp and 297 bp, this including P. sp-Cayenne. All 38 samples from Três Rios showed the typical restriction enzyme digestion profile expected for the Paracoccus group (Figure 2). The 14 Três Rios Cay-enne-tick DNA samples were all negative for additional PCR reactions when using primer sets for ompA, ompB, gltA and htrA genes (data not shown), thereby confirming the absence of Rickettsia sp.
The genus Paracoccus comprises metabolically versatile bacteria, which are either aquatic or terrestrial freeliving microbes or are associated with other organisms, such as plants, humans and invertebrates, including insects (Baker et al., 1998;Baj, 2000;Daneshvar et al., 2003;Funke et al., 2004;Amarasekare et al., 2008;Huang et al., 2012;Kämpfer et al., 2012;Singh et al., 2012). Interestingly, even though Paracoccus spp. have never been identified in ticks, the application of primers designed for Rickettsia 16S rDNA, followed by sequence analysis and RFLP, revealed the presence of P. sp-Cayenne in~77% of the cayenne ticks collected at Três Rios (Table 1). Phylogenetic inference, using partial 16S rDNA sequences, 864 Alphaproteobacteria in Cayenne ticks places P. sp-Cayenne into a soil-associated Paracoccus group, closely related to the Paracoccus sp. SA5 strain ( Figure 1A). Due to the proximity of P. sp-Cayenne and free-living soil isolated Paracoccus, the possibility of an accidental contamination was assessed, in an attempt to detect Paracoccus sp-Cayenne in soil samples from the same area at Três Rios. Soil samples were collected from four distinct areas used for horse pasture. A mixture of 7 g of each soil sample with 6 mL of TE was shaken, and after particle decantation, DNA was purified by phenolchloroform or DNeasy Kit (Qiagen, Valencia, CA) extraction. After PCR reactions using Eubacteria universal primers 27f/1492R (Lane, 1991) and a Rickettsia sp. 16S rDNA specific primer set, all samples proved to be positive for Eubacteria, although only one presented positive amplification using the Rickettsia primer set. Subsequent sequence analysis of this DNA sample indicated sequence relationship to a soil Acidobacteria (data not shown). Nevertheless, this geographically localized analysis does not rule out the soil origin of this tick-associated Paracoccus sp.
Over the last two decades, several new members of the Rickettsia genus were identified in association with ticks. Most were classified as tick symbionts, while others were considered EID agents (Parola et al., 1998;Fournier et al., 2000;Beninati et al., 2002;Nilsson et al., 2002). Many of the symbionts were further reclassified as pathogenic agents (Fournier et al., 2000;Nilsson et al., 2002). Within this scenario, R. amblyommii, which was identified in Amblyomma ticks from Brazil and the United States, could possibly represent an unkown ricketsiosis agent (Stothard and Fuerst, 1995;Labruna et al., 2004;Stromdahl et al., 2008), due to a suggestive association of R. amblyommii infections with several cases of a Rocky Mountain spotted-fever-like disease, ascertained by serological assays (Apperson et al., 2008), and the fact that R. amblyommii antigens cross-react with antibodies against R. rickettsii (Walker et al., 2008). In the present study, the presence of Rickettsia sp. A45 DNA in A. dubitatum collected in the Amazon region of São Jose do Xingu (MS), was confirmed. Sequence analysis showed that Rickettsia sp. A45 is closely related to R. amblyommii (Figure 1), which was also recently described in Amblyomma ticks collected in the Amazon (Labruna et al., 2004). Although the pathogenic potential of Rickettsia sp. A45 was not assessed, ompA phylogenetic analysis ( Figure 1C) indicated certain similarities to R. amblyommii, through the OmpA protein being directly associated with host-cell adherence and rickettsial pathogenicity (Blanc et al., 2005).
Primer sets for the Rickettsia genes ompA, ompB, htrA, gltA and 16S rDNA have been collectively and successfully applied to diagnose rickettsial infection (Regnery et al., 1991;Fournier et al., 2000;Simser et al., 2002;Pacheco et al., 2007;Phan et al., 2011;Radulovic et al., 2011;Zou et al., 2011). However, our results suggest a limitation in the use of the 16S rDNA primer set fD1-Rc16S, since PCR reactions using these primers also revealed the presence of P. sp-Cayenne in approximately 77% of A. cajennense ticks collected at Três Rios ( Figures 1A and 2). This suggests certain limitations when rickettsial 16S rRNA specific primers are solely used to assess Rickettsia infection in a biological sample.
Phylogenetic inference clustered P. sp-Cayenne 16S rDNA sequences with soil Paracoccus sp. Interestingly, the phylogenetic reconstruction did not keep soil, invertebrate associated, and clinical Paracoccus clusters robustly apart (bootstrap values < 80%) ( Figure 1A), thereby reflecting an intimate phylogenetic relatedness grouping all these organisms (Baker et al., 1998). A robust Paracoccus cluster was observed, keeping the marine group with the isolates P. sp. JLT1284, P. homiensis and P. sp. RA19 apart. It is important to note that the human-infectious P. yeeii cluster is not only closely related to the insect-associated Paracoccus marcusii cluster, but is often detected in patients with leg or foot wounds (Daneshvar et al., 2003;Funke et al., 2004), thus, a possible indication of vector bites. Recently a P. yeeii infection, associated with a peritonitis in a habitual horse rider (Wallet et al., 2010), could be correlated with the presence of P. sp-Cayenne encountered in horse-fed ticks at Três Rios. Also recently, several uncultured bacteria phylotypes with 16S rRNA gene sequences 99% identical to P. sp-Cayenne, were identified as composing the skin microbiota associated with human dermatitis (Grice et al., 2009;Kong et al., 2012) and infecting gastro-  Figure 2 -Agarose gel of 16S rDNA PCR products digested with HinfI. PCR reactions were achieved using the Rickettsia spp. specific primer set, as described in the text. Lanes include amplicon digestions for Rickettsia sp. A45 (Lane 1) and Paracoccus sp-Cayenne (Lane 2), as well as the tick-total DNA samples Cav22F and Cav49F (Lanes 3-4). The nondigested PCR product for Cav49F was run in parallel (Lane 5). The molecular marker is presented in base pairs (bp). The expected bands for the Rickettsia sample include the 195 bp, 117 bp and 124 bp bands, the last two appearing as a single band in the gels. The Paracoccus sp-Cayenne expected profile includes the bands of 297 bp and 100 bp, as well as the faint band of 25 bp, not observed in the gels.
intestinal tissues (Frank et al., 2007) (data not shown). It is not clear whether P. sp-Cayenne represents a potential pathogen, or is capable of competing with, or even exacerbating, other human tick-borne infectious agents. Even though possibly tick symbionts, other recognized symbiotic bacteria in Ixodidae are also potentially pathogenic (Burgdorfer et al., 1973). Although Paracoccus sp. was not specifically isolated in this study, this is the first report of Paracoccus sp. detection in ticks, thereby indicating that this bacterial group is harbored in Brazilian tick populations. Future studies should focus on isolating this Paracoccus sp. and on infected tick metagenomics, to clearly define both its pathogenic potential and/or its role in promoting aspects of A. cajennense biophysiology.