Introduction:
Maternally-inherited endosymbionts infecting arthropods are one of the most diverse and abundant of all bacteria infecting them. About two-third of terrestrial arthropods are infected with at least one maternally-inherited endosymbiont (Hilgenboecker, Hammerstein, Schlattmann, Telschow, & Werren, 2008; Weinert, Araujo-Jnr, Ahmed, & Welch, 2015). These endosymbionts play crucial role in the ecology and evolution of their hosts (Gebiola et al., 2017; Semiatizki et al., 2020). The most abundant of these are Wolbachia, Cardinium, Arsenophonus, Rickettsia and Spiroplama. Out of these,Wolbachia remains the most widely distributed endosymbiont with an incidence rates of 16-66% (Hilgenboecker et al., 2008; Werren, Windsor, & Guo, 1995) and 18 different clades (supergroup A to R) reported in different hosts across the world (Landmann, 2019). Incidence of the other endosymbionts varies from 4-10% (Duron et al., 2008; Zchori-Fein & Perlman, 2004).
The key factor explaining this abundance has been their ability to jump from one host to the other, in spite of being vertically transmitted from mother to offspring (Werren, Baldo, & Clark, 2008). As a result, these endosymbionts rarely show congruence with the host phylogeny, indicating substantial horizontal transfer across evolutionary timescales to taxonomically unrelated hosts (Shoemaker et al., 2002; Werren, Zhang, & Guo, 1995). This is also evident from the occurrence of similar endosymbiont strains in taxonomically unrelated hosts and conversely, the presence of divergent strains in closely related hosts (Vavre, Fleury, Lepetit, Fouillet, & Bouletreau, 1999). Individual arthropods can harbor multiple strains of one endosymbiont as well as multiple strains of different endosymbionts (Zélé et al., 2018) which perhaps indicates that different endosymbionts can use the same host to spread across different arthropod communities.
Another key feature of endosymbionts is the pervasive recombination seen in their genomes (Ellegaard, Klasson, Naslund, Bourtzis, & Andersson, 2013). This has been particularly well documented in Wolbachia(Malloch & Fenton, 2005) as well as other endosymbionts (Mouton et al., 2012). The level of recombination is so extensive that single gene sequences are unable to properly reflect the evolutionary history of a strain. Unsurprisingly, this has necessitated the development of multi locus strain typing (MLST) system (Maiden et al., 1998). The results of such MLST surveys revealed the extent of recombination to be similar to those of pathogenic free-living bacteria (Yahara et al., 2016). This is surprising because, unlike free-living bacteria, most endosymbionts cannot survive outside the host. Therefore, this extensive recombination must be an outcome of the horizontal transfer of strains across arthropod communities, since, for recombination to happen, two endosymbionts must come in contact within one host cytoplasm. As new strains are horizontally transferred to novel hosts; they encounter resident endosymbionts and thereby increase opportunities of recombination between them. Evidence for such recombination is also well documented. The parasitoid wasp Nasonia and its hostProtocalliphora show the presence of a very similar recombinantWolbachia (Werren & Bartos, 2001). Recombinant Wolbachiastrains have also been reported in Anastrepha fruit flies and their parasitoid braconid wasps (Mascarenhas, Prezotto, Perondini, Marino, & Selivon, 2016).
Horizontal transfer, therefore, can explain at least two major characteristics of endosymbionts, their wide distribution as well as the recombinant nature of their genomes. A major question that emerge from this is, at what level of biological organization are these horizontal transfers taking place? A relatively simple way to uncover this level is to enumerate specific ecological interactions where close associations between the two interacting arthropods have been implicated in horizontal transfer. These include host-parasite, host-parasitoid, prey-predator and other ecological relationships. Examples where host-parasitoid interactions have been implicated for such transfer include the presence of similar Wolbachia strains among frugivorous Drosophila and their hymenopteran parasitoid (Vavre et al., 1999), Nasonia vitripennis and Muscidifuraxuniraptor sharing similar Wolbachia with their fly hostProtocalliphora (Baudry, Bartos, Emerson, Whitworth, & Werren, 2003), transmission of Wolbachia into whitefly via parasitoid wasps (Ahmed, Breinholt, & Kawahara, 2016). Another such ecological association which can lead to endosymbiont transfer is prey-predator relationships like the predatory mite Metaseiulus occidentalisand its prey Tetranychus urticae (spider mite) sharing similar endosymbionts (Hoy & Jeyaprakash, 2005). Parasites like mites can also facilitate the transfer of Wolbachia to Drosophila host populations (Brown & Lloyd, 2015). These can also be host plant mediated transfer of Cardinium to different leaf hopper species (Gonella et al., 2015) as well as horizontal transfer ofWolbachia in whitefly via cotton leaves (Li et al., 2017).
It is clear from these examples that these horizontal transfers are taking place when two hosts are coming together to perform a particular ecological function. The endosymbiont present within these hosts are then serendipitously getting transferred from one host to the other. Therefore, to understand the dynamics of the spread of endosymbionts through horizontal transfer, one needs to look at the level where most of these ecological associations are taking place, which is within a particular ecological community. A well-defined ecological community will have several diverse host taxa interacting with each other, thereby, facilitating horizontal transfer. Moreover, many host taxa can belong to many different ecological communities (Morrow, Frommer, Shearman, & Riegler, 2014). This cosmopolitan nature of a few host taxa will further facilitate the spread of endosymbionts from one ecological community to another, almost like spreading through a metacommunity (Brown, Mihaljevic, Des Marteaux, & Hrček, 2020). Therefore, investigating endosymbiont diversity and horizontal transfer within specific ecological communities seems logical. Yet, there are very few studies that have taken this approach and instead focus mainly on endosymbiont spread within a particular habitat (Stahlhut et al., 2010), or in a specific genus (Baldo et al., 2008; Raychoudhury, Baldo, Oliveira, & Werren, 2009) or within specific taxa (Ahmed et al., 2016). Amongst community-wide surveys, Kittayapong, Jamnongluk, Thipaksorn, Milne, and Sindhusake (2003), demonstrated Wolbachia strain diversity within rice field arthropod community. Sintupachee, Milne, Poonchaisri, Baimai, and Kittayapong (2006), reported plant-mediated horizontal transfer among arthropod community found on pumpkin leaves. Most of these studies are based on single gene phylogenies which makes identification of recombination difficult. An important corollary of this view of within-community horizontal transfer of endosymbionts can lead to another important hypothesis about sequence diversity of the endosymbionts themselves. If endosymbionts are rapidly undergoing horizontal transfer within a particular ecological community then very similar bacterial strains would be found among the arthropod hosts of that community. This would make these bacteria more closely related to each other, than expected, resulting in lower than expected pairwise sequence divergence among them. This lower than expected levels of sequence divergence can serve as a signature of recent and relatively rapid community-wide horizontal transfer of resident endosymbionts.
In the present study, we try to answer whether such relatively rapid horizontal transfer and resulting recombination are happening within the endosymbionts of a diverse soil arthropod community. Three major endosymbionts, Wolbachia, Cardinium and Arsenophonus , were selected and screened across arthropod hosts. We investigatedWolbachia sequence diversity using the well-established MLST scheme (Baldo et al., 2006) and also identified specific recombination events. We also investigated Cardinium and Arsenophonusincidence but with single gene sequences. A statistical model was then used to test whether the endosymbiont found within this community are more closely related to each other than expected.