Introduction
It is now widely recognized that animals live in a microbial world, and that many aspects of animal biology, ecology and evolution are a product of their symbioses with microorganisms (McFall-Ngai et al., 2013). In invertebrates, these symbioses may be particularly intimate, and involve transmission of the microbe from parent to offspring (Hurst, 2017). The alignment of host reproduction with symbiont transmission produces a correlation between the fitness interests of the parties, reflected in symbionts evolving to play a number of physiological roles within the host, from defence (Jaenike, Unckless, Cockburn, Boelio, & Perlman, 2010; Łukasik, Guo, van Asch, Ferrari, & Godfray, 2013; Teixeira, Ferreira, & Ashburner, 2008) through to core anabolic and digestive functions (Douglas, 2009; Rio, Attardo, & Weiss, 2016). However, the maternal inheritance of these microbes has led to the retention of parasitic phenotypes associated with distortion of reproduction, with symbiont phenotypes including biases towards daughter production and cytoplasmic incompatibility (Hurst & Frost, 2015). These diverse individual impacts alter the ecology and evolution of the host, in terms of diet, dynamics of interaction with natural enemies, sexual selection and speciation.
Heritable symbioses have evolved on multiple occasions amongst microbial taxa. In some cases, the microbial lineage is limited to a single clade of related animal hosts, such as Buchnera in aphids (Munson, Baumann, & Kinsey, 1991). In other cases, particular heritable microbes are found across a wide range of arthropod species. Wolbachiarepresents the most common associate, considered to infect nearly half of all species (Zug & Hammerstein, 2012), and this commonness is a function in part of the ability of Wolbachia to transfer to a broad range of new host species and spread within them (host shift events) (Siozios, Gerth, Griffin, & Hurst, 2018; Turelli et al., 2018). Aside Wolbachia , other microbes are found commonly as heritable symbionts of arthropod hosts (Duron et al., 2008). Cardinium andRickettsia , for instance, have been estimated at being present in 13-55% and 20-42% of species respectively (Weinert, Araujo-Jnr, Ahmed, & Welch, 2015).
In this paper, we address the diversity and commonness of symbioses between Rickettsia and arthropods. The Rickettsia have increasingly been recognized as a genus of bacteria with diverse interactions with arthropods (Perlman, Hunter, & Zchori-Fein, 2006; Weinert, Werren, Aebi, Stone, & Jiggins, 2009). First discovered as the agents underlying several diseases of humans vectored by haematophagous arthropods (da Rocha-Lima, 1916; Ricketts, 1909), our understanding of the group changed in the 1990s with the recognition that Rickettsia were commonly arthropod symbionts (Chen, Campbell, & Purcell, 1996; Fukatsu & Shimada, 1999; Werren et al., 1994).Rickettsia were recognized first as male-killing reproductive parasites (Hurst, Walker, & Majerus, 1996; Werren et al., 1994) and then later as beneficial partners (Hendry, Hunter, & Baltrus, 2014; Himler et al., 2011; Łukasik et al., 2013).
Following this extension of our understanding ofRickettsia -arthropod interactions, a new clade ofRickettsia was discovered from work in Torix leeches (Kikuchi & Fukatsu, 2005; Kikuchi, Sameshima, Kitade, Kojima, & Fukatsu, 2002). This clade was sister to all other Rickettsia genera, with no evidence to date of any strain having a vertebrate pathogen phase. The host range for Torix Rickettsia is broader than that for other members of the genus, going beyond arthropods to include leech and amoeba hosts (Dyková, Veverková, Fiala, Macháčková, & Pecková, 2003; Galindo et al., 2019; Kikuchi & Fukatsu, 2005). Targeted PCR based screening have revealed Torix group Rickettsia as particularly common in three groups with aquatic association: Culicoides biting midges, deronectid beetles and odonates (Küchler, Kehl and Dettner, 2009; Pilgrim et al., 2017; Thongprem et al.2020). However, some previous hypothesis-free PCR screens that aimed to detect Rickettsia in arthropods have likely missed these symbioses, due to divergence of the marker sequence and mismatch with the primers (Weinert, 2015).
During our previous work on Torix Rickettsia in biting midges (Pilgrim et al., 2017), we became aware of the presence ofRickettsia CoxA sequences deposited in Genbank that derived from studies where the intended target of amplification/sequencing was cytochrome c oxidase I (COI ), the mitochondrial equivalent of CoxA . These deposits derived from studies using mtDNA for phylogeographic inference (Lagrue, Joannes, Poulin, & Blasco-Costa, 2016), or in barcoding based species identification approaches (Ceccarelli, Haddad, & Ramírez, 2016; Řezáč, Gasparo, Král, & Heneberg, 2014). Non-target amplification of Rickettsia COI using mitochondrial COI barcoding primers has been reported in spiders (Ceccarelli et al., 2016; Řezáč et al., 2014) and freshwater amphipods (Lagrue et al., 2016; Park & Poulin, 2020). Furthermore, we have noted two cases in our lab where amplicons obtained for mtDNA barcoding of an insect have, on sequence analysis, revealed Rickettsia COIamplification (Belli group Rickettsia from Collembola, and Torix group Rickettsia from Cimex lectularius bedbugs). Previous work had established barcoding approaches may amplify COI fromWolbachia symbionts (Smith et al., 2012), and the data above indicate that non-target Rickettsia COI may be likewise amplified during this PCR amplification for mitochondrialCOI .
In this paper, we use three approaches to reveal the diversity and commonness of Torix Rickettsia in arthropods. First, we probed the contaminant bin of the Barcode of Life Data System (BOLD) forRickettsia sequences and used the template from these projects to define the diversity of Rickettsia observed using a multilocus approach. Second, we screened DNA templates from multiple individuals from 169 arthropod species for Rickettsia presence using PCR assays that function more broadly than previously utilized in screens. Within this, we included a wider range of aquatic taxa, to investigate if the previous work highlighting particular aquatic taxa as hotspots for Rickettsia symbiosis (water beetles, biting midges, damselflies) reflects a wider higher incidence in species from this habitat. Finally, we used bioinformatic approaches to examine the Sequence Read Archive (SRA) depositions for one individual from 1,342 arthropod species for the presence of Rickettsia and used this as a means of estimating the relative balance of Torix group to otherRickettsia within symbioses.