Rickettsia asembonensis belongs to a group of R. felis–like organisms (RFLOs) that are similar, yet distinct, from their closest known relative, R. felis (1,2). Although R. felis causes disease in humans (3), the pathogenicity of RFLOs remains unknown (1,4,5). R. asembonensis was initially identified in domestic fleas from Kenya (1). Subsequently, reports from the Americas, Asia, and Africa established that R. asembonensis is ubiquitous and closely associated with human habitats because of its arthropod hosts (4–7). However, reports of R. asembonensis rarely include robust genomic information needed to establish degrees of genetic diversity. Consequently, many rickettsial infections remain underdiagnosed, even when prevalence is high (8). We recently described R. asembonensis in multiple ectoparasites (Ctenocephalidesfelis fleas and Rhipicephalussanguineus ticks) collected in the Peruvian Amazon (9). Here, we detail multilocus sequence typing of a single molecular isolate using next-generation sequencing data for 5 complete genomic regions, including conserved (17kDa and gltA) and variable (ompA, ompB, and sca4) genes.

The internal review board of the US Naval Medical Research Unit No. 6 and the Institutional Animal Care and Use Committee approved the study protocol in compliance with all applicable regulations. Genomic DNA was mechanically extracted from half of a single C. felis flea as described (9) and fragmented by Bioruptor (Diagenode, Denville, NJ, USA). Fragmented DNA served as template to prepare IonPGM libraries using IonPlus Fragment Library Kits (ThermoFisher, Lima, Peru) according to the manufacturer’s directions. We conducted quality control using Bioanalyzer High Sensitivity chips (Agilent, Lima, Peru). We prepared libraries for sequencing using IonPGM Template OT2 200 Kits (ThermoFisher, Lima, Peru) and conducted sequencing on 318 chips using IonPGM Sequencing 200 Kits v2 (ThermoFisher). We processed raw data by reference mapping against NMRCii from Kenya (10). Of the 20,575,878 shotgun sequencing reads generated, ≈12% matched Rickettsiaceae.

Comparison of the consensus sequences we generated (GenBank accession nos. KY650696–KY650700) with those of strain NMRCii (GenBank accession no. JWSW01000078.1) (10) indicates high identity at the nucleotide (99.8%–100.0%) and amino acid (99.6–100.0%) levels (Table). As expected, conserved genes (17kDa and gltA) showed fewer substitutions than variable genes (ompA, ompB, and sca4). The 17kDa gene exhibited no mutations along its 480-nt open reading frame (ORF), whereas the gltA gene exhibited 3 mutations along its 1,314-nt ORF. Two mutations in gltA encoded silent changes, whereas the third encoded a conservative lysine-to-glutamic acid change at position 290. In the variable group, ompB exhibited no mutations along its 4,947-nt ORF; ompA and sca4 exhibited 5 each. ompA had 2 conservative changes (leucine-to-valine at position 1454 and valine-to-alanine at position 1627) and 2 nonconservative changes (tyrosine-to-aspartic acid at position 162 and arginine-to-glycine at position 1280). sca4 had 1 conservative change (glutamine-to-histidine at position 608) and 3 nonconservative changes (leucine-to-proline at position 128, arginine-to-glycine at position 754, and isoleucine-to-threonine at position 831). On the basis of these data, we conclude that the Peru molecular isolate is distinct from the original Kenya strain.

To further characterize the Peru isolate, we conducted phylogenetic analysis using the conserved gltA gene. Although reference sequences are available for multiple Rickettsiaceae, R. asembonensis sequences are limited in number and length (Technical Appendix Table). Nevertheless, we constructed a phylogenetic tree using almost the complete gltA gene (1,068 [81%] nt of the ORF). As expected, the Peru isolate groups with RFLOs, including other R. asembonensis isolates and R. senegalensis (Technical Appendix Figure ,panel A). Construction of an additional tree using only 348 nt of gltA sequence available for an increased number of isolates (Technical Appendix Table) enabled us to confirm placement and relationship with other strains from the Americas (Technical Appendix Figure, panel B). This tree focuses exclusively on the transitional group and includes partial R. asembonensis references from Brazil, Colombia, and Costa Rica that were not available for inclusion in the 1,068-nt gltA tree. The Peru isolate clearly groups with other American isolates, and this subgroup is distinct from the original Kenya strain.

R. asembonensis is a new species (2) with potential as a ubiquitous human pathogen. Despite worldwide distribution, whether R. asembonensis and other RFLOs are pathogenic to humans, as is their closest relative R. felis, remains unknown. Complete genomic data, which are largely lacking from public repositories, are required to assess genetic diversity. Using next-generation sequencing, we generated complete sequences for 2 conserved (17 kDa and gltA) and 3 variable (ompA, ompB, and sca4) genes of an R. asembonensis molecular isolate from Peru. Although characterization of 1 isolate is not sufficient to evaluate strain diversity within Peru, much less among American strains, these sequences represent a major contribution toward the expansion of availability of much needed genomic information. Our multilocus sequence typing and phylogenetic analyses indicate that the Peru isolate is closer to American strains than to the original strain from Kenya. Characterization of additional isolates, derived from a variety of ectoparasites in which R. asembonensis has been detected, is needed to further validate our findings and to conduct in-depth diversity studies. In turn, these results should help decrease the chronic underdiagnosis of rickettsial diseases throughout the Americas.