Fleas (Insecta: Siphonaptera) are recognized vectors of bacteria that affect human and other animal health, whose reservoirs are in the majority mammals. Among these, some species of the genera Rickettsia (Rickettsiales: Rickettsiaceae) and Bartonella (Rhizobiales: Bartonellaceae) are emerging and re-emerging throughout the world; however, their circulation across vast regions of Argentina and numerous animal species, particularly wild species, remains largely unknown. The study of wild animal roadkill provides valuable insights into parasitic associations and the presence of pathogenic microorganisms, allowing the generation of a health alert in certain ecosystems. The aim of this study was to describe the diversity of fleas associated with roadkilled wild native meso-mammals in the extreme northeast of Argentinian Patagonia and to detect the presence of Rickettsia and Bartonella bacteria. Five host species were examined, including Chaetophractus villosus Desmarest (Cingulata: Chlamyphoridae); Didelphis albiventris Lund (Didelphimorphia: Didelphidae); Lagostomus maximus Desmarest (Rodentia: Chinchillidae); Leopardus geoffroyi d'Orbigny & Gervais (Carnivora: Felidae) and Lycalopex gymnocercus Fischer (Carnivora:  Canidae). A total of 248 fleas were recovered, identified as Hectopsyllidae: Hectopsylla broscus Jordan & Rothschild and Hectopsylla cypha Jordan; Malacopsyllidae: Phthiropsylla agenoris Rothschild and Malacopsylla grossiventris Weyenbergh; Pulicidae: Ctenocephalides felis Bouché and Pulex irritans Linnaeus. Molecular analysis detected two flea-borne pathogenic bacteria: Rickettsia felis (Bouyer et al.), found in C. felis from D. albiventris, and Bartonella rochalimae (Eremeeva et al.), reported here for the first time in Argentina, detected in P. agenoris from C. villosus, and in P. irritans from L. maximus and L. gymnocercus. The results contribute to knowledge of flea diversity in Argentinian Patagonia and provide new information about flea-borne pathogens circulating in the wildlife of this region. Furthermore, this study is the first in Argentina to investigate ectoparasites and their associated bacteria in roadkilled animals, making a pioneering contribution to the field. The interesting findings highlight the importance of implementing and expanding road ecology studies, which could easily be replicated in other regions where information gaps on flea and flea-borne bacteria diversity still exist.

2.1 Study area

The study was conducted in the extreme northeast of Río Negro province, Argentinian Patagonia, on the coast of the Atlantic Ocean (Fig. 1; Supplemental Material 1). The natural ecosystem corresponds to the Monte Desert biome (Abraham et al. 2009). The climate is semiarid to arid, with high evaporation enhanced by windy conditions, and mean annual rainfall varies between 100 and 450 mm. The vegetation is rather uniform in terms of physiognomy and flora composition (Abraham et al. 2009; Oyarzabal et al. 2018).

The area covers Provincial Route n°1 between two cities: Viedma (-40.177288 S; -62.9988395 W) with 80,000 inhabitants, and Balneario El Cóndor (-41,0437078 S; -62,8288118 W), an urban entity with less than 2,500 inhabitants, totalling 28 km. Between both cities, xerophytic native flora alternates with semi-extensive livestock (mainly cattle) and farming areas. Also, the progress of urbanization is evident, with the consequent loss of the natural ecosystem, generating a transition area where wild species, domestic species and humans coexist.

2.2. Host and flea collection and identification

The study received approval from the Secretary of Environment and Sustainable Development of Río Negro province (085206SAYDS/218/2015). Carcasses of wild native meso-mammals were then collected along Provincial Route n°1 and urban roads in the seaside town of Balneario El Cóndor between January 2020 and May 2024.  The animals were found randomly by the authors, dead due to vehicular collisions. Each host found was immediately placed in a white polyethylene bag and then transferred to a plastic container to prevent fleas from escaping and to facilitate their collection.

Afterward, the mammals were taken to the Universidad Nacional de Río Negro laboratory, where the host species were identified. The native mammal species recorded were: Chaetophractus villosus (Desmarest) “Big hairy armadillo” (Cingulata: Chlamyphoridae); Didelphis albiventris Lund “White-eared opossum” (Didelphimorphia: Didelphidae); Lagostomus maximus (Desmarest) “Plains vizcacha” (Rodentia: Chinchillidae); Leopardus geoffroyi d'Orbigny & Gervais “Geoffroy's cat” (Carnivora:  Felidae) and Lycalopex gymnocercus (Fischer) “South American grey fox” (Carnivora: Canidae). Identification of the mammals was performed according to Chebez et al. (2014).

Fleas were collected using toothbrushes and forceps, and stored in 96% ethyl alcohol until molecular analyses was carried out at the Centro de Bioinvestigaciones (CeBio). The parasitological examination and flea identification methods are based on the protocols outlined in Sanchez (2013) and Sanchez and Lareschi (2019). First, the fleas were identified at the genus level using a stereoscopic microscope, followed by DNA extraction (see section 2.3). Afterward, species identification was performed using the methodology described below.

Fleas were cleared and softened in aqueous solution of potassium hydroxide 10% (KOH), dehydrated in an increasing series of ethanol (80-100%), further diaphanized in eugenol, and mounted on permanent slides with Canadian balsam and studied under light microscopy. Identification keys used included those of Hopkins and Rothschild (1953), Johnson (1957), and Smit (1987).

2.3. DNA extraction

For DNA extraction, one flea per host was randomly selected, washed, and cut between the third and fourth abdominal tergites using a sterile scalpel. When more than one flea species was present on the same host, one individual of each species was randomly selected and analysed separately. All materials used to handle the fleas, including brushes for moving the ectoparasites, Petri dishes for holding them, and the scalpel for making incisions, were sterilized between each sample. Genomic DNA extraction was performed from individual fleas per host, using Chelex^®^-100 (Bio-Rad Laboratories, CA, USA) described by Acosta et al. (2023). Following the DNA extraction, the fleas exoskeletons were recovered and stored in 96% ethanol and subsequently mounted for species identification (see section 2.2).

2.4. PCR amplification of genes from Bartonella spp. and Rickettsia spp.

The presence of Bartonella was screened using the citrate synthase (gltA) and RNA polymerase beta-subunit (rpoB) genes, while the presence of Rickettsia was screened using the citrate synthase (gltA), outer membrane protein A (ompA) and outer membrane protein B (ompB) genes (Table 1). For the amplification, the polymerase chain reaction (PCR) program started with an initial denaturation for 5 min at 95 °C, followed by 40 cycles of 95 °C for 30 s, gene-specific annealing °C for 30 s, and 72 °C for 30 s, and a final extension step at 72 °C for 5 min (Table 1). PCR reaction was set to a final volume of 20 μL, containing: 25-100 ng of template DNA, 1.5 mM MgCl₂, 0.2 μM of each primer, 0.2 mM of each dNTP, 1X reaction buffer, 0.5U of Taq Pegasus DNA polymerase and ultrapure sterile water to final volume. All amplifications were performed in conjunction with a negative (distilled water) and positive (DNA of Bartonella henselae Regnery et al., provided by “ANLIS Malbrán”, Buenos Aires, Argentina, and DNA of Rickettsia parkeri, provided by Instituto Nacional de Enfermedades Virales Humanas “Dr. Julio I. Maiztegui”, Buenos Aires, Argentina) controls. DNA fragment amplification was confirmed by electrophoresis on 1% w/v agarose gel, stained with ethidium bromide (10 mg/μL) and visualized under UV light.

In the samples with positive PCR for the genes analysed, we proceeded to purification using 10U of Exonuclease I (Thermo Fisher Scientific) and 1U of FastAp thermosensible alkaline phosphatase (Thermo Fisher Scientific), incubating at 37°C for 15 min and a subsequent 15 min at 85°C to stop the reaction. The purified samples were sequenced by Macrogen Co. Ltd. (South Korea).

2.5. Sequencing, nBLAST, and phylogenetic analysis

The obtained sequences for the genes were analysed and manually edited using the BioEdit program (Hall 2004). To assign identity to each sequence with its respective statistical significance, it was subjected to a homology comparison against the GenBank nucleotide database (https://www.ncbi.nlm.nih.gov/genbank/), making use of the nBLAST algorithm (https:// blast.ncbi.nlm.nih.gov/Blast.cgi). The sequences have been deposited in the GenBank nucleotide database (https://www.ncbi.nlm.nih.gov/genbank/) under accession numbers PP985429 - PP985431 for Rickettsia, and PP985432 - PP985440 for Bartonella.

The complete set of the** gene sequences was employed for a multiple alignment performed with the ClustalW algorithm and the MEGA v.6 software (Tamura et al. 2013), together with sequences taken from the GenBank database. The resulted alignment was checked and manually corrected. Moreover, phylogenetic trees were built using the Maximun-Likelihood (ML) clustering methods, both for individual genes and for concatenated sequences. In the case of gene concatenation for both bacteria, the Farris test (Farris et al. 1994) was initially performed using Phylogenetic Analysis Using Parsimony (PAUP*) methods based on parsimony inference (Swofford and Sullivan, 2003) to establish whether these genes could be used. Subsequently, the Mesquite program (Maddison and Maddison 2021) was utilized to concatenate these sequences. The nucleotide substitution model that best fit the data was calculated using JModelTest software v2.1.4 (Darriba et al. 2012). The evolutionary history was inferred using the ML method based on the Tamura 3-parameter (I + G) model with 10,000 replicates of random-addition taxa and tree bisection and reconnection branch swapping. All positions were weighted equally.