Distribution and origin of the symbionts in the lice microbiomes
Since the process of genome degradation starts once the bacterium becomes an obligate vertically-transmitted symbiont, we should expect, at least during the initial phase of the symbiogenesis, a correlation between the degree of genome degradation and the duration of host-symbiont coevolution (Moran, 1996). For example, among the symbionts of sucking lice, the high degree of genome degradation ofR. pediculicola and L. polyplacis indicates a relatively long and intimate association with the host. In correspondence with this presumption, the Riesia lineage has been found in several louse species of two different genera, Pediculus and Phthirus,and L. polyplacis is hosted at least by two louse species,P. serrata and P. spinulosa (Hypsa & Krizek, 2007). Moreover, our extensive amplicon screening shows that L. polyplacis is consistently present in a broad geographic and phylogenetic sample of P. serrata as a dominant bacterium (Figure 3). When compared to these two well documented examples of established P-symbionts, the Neisseria -related symbionts, with an intermediate degree of the genome degeneration, show more restricted and patchy distribution (Figure 3). Only in H. acanthopus were they consistently present (the Neisseriaceae OTU) as the most dominant bacterium (with the exception of two specimens from one population; Figure 2), but were not found in the two examined specimens of the related species (H. edentula) . The overall variability of the microbiomes was apparently correlated with the lice genetic background: in Bulgarian samples the Neisseriaceae OTU was the only present bacterium, in other populations of H. acanthopus it was accompanied by an unknown bacterium which the blast search affiliated with Blochmania , and in the two samples of H. edentula the only present OTU corresponded to the genus Arsenophonus. SinceP. serrata is known to harbor the typical obligate P-symbiontLegionella polyplacis (Rihova et al., 2017), we screened this louse more extensively across several populations and genetic lineages. The results confirmed a ubiquitous presence of L. polyplacis , which in most cases was the most abundant OTU, and only occasional co-occurrence of the Neisseriaceae OTU (Figure 3). The split of L. polyplacis into two different OTUs correlated to the host’s phylogeny, reflects evolutionary changes during the evolution of the symbiont in distant host lineages, but certainly does not suggest the presence of two independent symbiotic lineages (in fact phylogenetic and genomic analyses confirm that Polyplax -Legionella co-evolution crosses the host species boundaries and the same symbiont is also present in the related louse species P. spinulosa ; (Hypsa & Krizek, 2007). Three additional OTUs in P. serrata microbiomes show affinity to known insect symbionts, the Buchnera OTU and twoArsenophonus OTUs. Based on their genetic divergence and the differences in their GC content, the two Arsenophonus OTUs seem to represent two different lineages.
In respect to the general concept of symbiont acquisition, loss, and replacement within insects, and the high dynamism of louse microbiomes, two OTUs are of particular interest. Both the Buchnera OTU fromP. serrata and the Blochmania OTU from H. acanthopus seem to represent strongly derived symbiotic genomes (Figure 2 & 3), which blast-assigned taxonomy reflects the low GC content rather than real phylogenetic relationships. Since our metagenomic data did not yield any reliable information on either of these bacteria, it is difficult to hypothesize about their phylogenetic origin and function in the host. However, strong genome reduction, deduced from the GC content of the 16S rRNA gene amplicon, suggests that they may represent the scattered remains of ancient symbionts, now retreating from the host’s population and replaced with more recent acquisitions. Interestingly, the FISH analysis shows that apart from the “Neisseria -H” symbiont, the bacteriocytes of H. acanthopus harbor another bacterium (Figure 4). Since the metagenomic assembly did not contain any other bacterial contigs, we were not able to identify the origin of this second symbiotic bacterium.
The diversity of microbiomes and the rapid process of symbionts acquisitions/replacements make sucking lice an interesting model for studying the conditions and processes in early stage of symbiogenesis. Considering the distribution patterns and the low degree of genome modifications in the two Neisseria­- related symbionts, it is unlikely that their occurrence in two different lice lineages is due to a common symbiotic origin in the Hoplopleura -Polyplaxancestor. The most parsimonious explanation is thus an independent origin of the symbiosis in each louse genus. This poses an interesting question on the source of these symbionts and the mechanisms underlying their acquisition and symbiogenesis. Co-occurrences of closely related symbiotic bacteria in related insect hosts are usually consequences of either co-speciation or a tendency of specific bacterial lineages to frequently establish symbiosis with specific insect hosts (e.g.Arsenophonus , Wolbachia ). However, neither of these explanations can be applied to the lice-Neisseria association. Members of the family Neisseriaceae are only rarely found in symbiotic association with insects. The only well documented case of obligate symbiosis is the genus Snodgrassella found in several species of bees and bumblebees (Kwong & Moran, 2013). Based on the 16S rRNA gene phylogeny, the closest relative of the louse-associated Neisseriais an uncultured bacterium described from a flea Oropsylla hirsuta (Jones, McCormick, & Martin, 2008), for which no other information is currently available. It is interesting to note that, similar to Legionella polyplacis , the Neisseria -like symbionts originate from a bacterial lineage which is rarely found in insects and is a well-known vertebrate pathogen. Phylogenetic correlation between the vertebrate pathogens and symbionts of blood-feeding arthropods was previously reported in ticks (Ahantarig, Trinachartvanit, Baimai, & Grubhoffer, 2013; Felsheim, Kurtti, & Munderloh, 2009; Guizzo et al., 2017; Niebylski, Peacock, Fischer, Porcella, & Schwan, 1997; Noda, Munderloh, & Kurtti, 1997). For one of the tick symbionts, Francisella -like bacterium, the origin from mammalian pathogen was recently suggested by Gerhart, Moses and Raghavan (2016). Based on the data available for the few louse genera examined so far, their microbiomes contain both vertebrate pathogens and typical insect-associated bacteria (e.g. Arsenophonus , Sodalis ). A more detailed high throughput screening of different groups of sucking lice is now needed to assess significance of these two ecological groups of bacteria as a source of nutritional symbionts, and to study the processes during the early symbiogenesis.