Wolbachia are widespread obligate intracellular Alphapro-teobacteria, estimated to infect 40-65% of insect species (Hilgenboecker et al., 2008; Werren et al., 2008; Zug and Hammerstein, 2012). While being primarily maternally transmitted, horizontal transfers have frequently occurred over evolutionary time-scales (O’Neill et al., 1992; Rousset et al., 1992; Werren et al., 1995). To date, Wolbachia strains are divided into 16 clades, referred to as “supergroups” A-Q (Lo et al., 2007; Ramírez-Puebla et al., 2015).

The Nasonia species complex has been a major model system for Wolbachia-insect symbioses for more than 25 years (Breeuwer and Werren, 1990). This young species complex has enabled scientists to reconstruct the history of Wolbachia acquisitions and transmission routes across the Nasonia clade (van Opijnen et al., 2005; Raychoudhury et al., 2009). Moreover, the ability to produce interspecies hybrids has been exploited to introgress the cytotype (including the Wolbachia) of a given species into the nuclear genotype of another, thereby providing insights into Wolbachia-host genotype interactions, different modes of CI and the role of Wolbachia in speciation (Breeuwer and Werren, 1990, 1993b, Bordenstein and Werren, 1998; Bordenstein et al., 2001, 2003; Chafee et al., 2011; Raychoudhury and Werren, 2012).

The emerging picture of the Nasonia-Wolbachia association is as follows: The four Nasonia species together harbor 11 different Wolbachia strains from the A and B supergroups (Figure 3A) (Raychoudhury et al., 2009), two major arthropod-Wolbachia clades that diverged about 60 million years ago (Werren et al., 1995). N. vitripennis carries two Wolbachia strains (one from each supergroup), while the three younger species are all triple infected: N. giraulti and N. oneida both harbor two supergroup A strains and one supergroup B strain, whereas N. longicornis harbors one supergroup A strain and two supergroup B strains (Figure 3A) (Raychoudhury et al., 2009). Comparing the phylogenetic relationships between these Wolbachia strains with host phylogenies based on nuclear and mitochondrial genes revealed that several Wolbachia strains were most likely acquired independently via horizontal transfers from other insects, including Drosophila spp. (supergroup A strains of N. giraulti and N. longicornis), the parasitoid wasp Muscidifurax uniraptor (supergroup A strain of N. vitripennis) as well as the blowfly Protocalliphora sialia (supergroup B strain of N. vitripennis) (van Opijnen et al., 2005; Raychoudhury et al., 2009). The latter cases point towards an ecological interaction as the source of the Wolbachia transfers to N. vitripennis, since both Nasonia and M. uniraptor parasitize blowflies. Indeed, horizontal transfers between parasitoids and their fly hosts as well as between different parasitoid species infecting the same host are known to occur occasionally (Heath et al., 1999; Vavre et al., 1999; Huigens et al., 2004).

The supergroup B Wolbachia of N. longicornis and N. giraulti were acquired prior to the speciation of the two species and subsequently co-diverged with their hosts (Figure 3A) (van Opijnen et al., 2005; Raychoudhury et al., 2009). Hence, the supergroup B strain from N. giraulti is nearly identical to one of the B strains from N. longicornis and the estimated divergence time of the two Wolbachia strains coincides with the divergence of their host species, i.e., about 0.4-0.5 mya (van Opijnen et al., 2005; Raychoudhury et al., 2009). In addition, the second B strain of N. longicornis is estimated to have diverged from the other B strains about 1.5 mya. Considering that this time point was long before the speciation of the two host species, it is likely that the common ancestor of N. longicornis and N. giraulti harbored two B strains, one of which was lost in N. giraulti after the speciation event (Raychoudhury et al., 2009). A similar co-divergence event between the A strains of N. longicornis and N. giraulti is possible, but the similarity of these strains with Wolbachia from several Drosophila species currently makes it impossible to rule out independent horizontal transfer events (Raychoudhury et al., 2009). The three Wolbachia strains in the recently discovered species N. oneida are identical (for 5 house-keeping genes and the wsp gene) to those of the closely related N. giraulti and have likely been acquired via hybridisation between the two species, resulting in a mitochondrial-Wolbachia sweep from N. giraulti to N. oneida (Figure 3A) (Raychoudhury et al., 2009). Future phylogenomic comparisons based on the entire genomes of the different Wolbachia strains will be needed to obtain a higher resolution. Nonetheless, these findings illustrate a high Wolbachia diversity in the Nasonia species complex, along with various patterns of Wolbachia transfers within a single insect genus.

One of the most prominent aspects of Wolbachia is undoubtedly its ability to manipulate host reproduction in various ways (Werren et al., 2008). All Nasonia-associated Wolbachia induce CI, a reproductive incompatibility consisting of two components: A symbiont-induced modification of the paternal chromosomes during spermatogenesis that needs to be ‘rescued’ by the same symbiont being present in the fertilized egg (Tram and Sullivan, 2002; Serbus et al., 2008). If the female is uninfected (unidirectional CI) or harbors a different bacterial strain (bidirectional CI), the modification may not be rescued, resulting in CI (Figure 3B). Consequently, infected females have a fitness advantage over uninfected females, since they can reproduce successfully with all available males, regardless of male infection status. Similarly, bidirectional CI results in a fitness benefit for multiply infected females since they are at lower risk to suffer from CI (Mouton et al., 2003, 2004). The fact that all Nasonia species generally harbor double or triple infections that are mutually incompatible reinforces bidirectional incompatibility between all species pairs, with the notable exception of N. giraulti and N. oneida, whose Wolbachia strains are identical (Breeuwer and Werren, 1990; Breeuwer et al., 1992; Bordenstein and Werren, 1998, 2007; Bordenstein et al., 2001, 2003; Raychoudhury et al., 2010a). While the exact molecular mechanisms of CI are still not understood, it is evident that the paternal chromosomes fail to condense correctly and may be lost during the first mitotic division, causing embryo mortality in diploid organisms (Tram and Sullivan, 2002). In contrast, in haplodiploid insects such as Nasonia, CI can be manifested in different ways. The most extreme phenotypes are Male Development and Female Mortality (Vavre et al., 2000; Bordenstein et al., 2003; Vavre et al., 2009). In the first case, the paternal genome is completely lost, which restores haploidy in fertilized eggs and results in the conversion of female into male offspring. Hence, this type of CI is characterized by all-male broods (Figure 3B) with little or no embryonic mortality compared to compatible crosses. In contrast, an incomplete loss of the paternal chromosomes would instead cause aneuploidy and a high mortality of fertilized eggs (i.e., female offspring), resulting in smaller broods with male-biased sex-ratios (Bordenstein et al., 2003). It has been hypothesized that paternal chromosome loss was mediated by the intensity of Wolbachia-induced sperm modification, with less efficient modifications leading to only partial chromosome loss and aneuploidy in fertilized eggs (Vavre et al., 2009). While CI-induced mortality seems to be common in haplodiploid species, including the younger Nasonia species N. longicornis and N. giraulti (Bordenstein et al., 2003; Dedeine et al., 2004; Vavre et al., 2009), N. vitripennis is an exception from the rule in that the Male Development type is the predominant CI phenotype in this species (Breeuwer and Werren, 1990; Bordenstein et al., 2003). Interestingly, interspecies crosses and introgression experiments between N. vitripennis and N. giraulti revealed that this rare CI phenotype is determined by the N. vitripennis nuclear genotype rather than Wolbachia-related effects, since it is even observed in incompatible crosses between infected N. giraulti males and uninfected N. vitripennis females (Bordenstein et al., 2003). These results show that the host genotype may determine the fate of the paternal chromosomes in fertilized eggs, independent of Wolbachia (Bordenstein et al., 2003). On an evolutionary scale, the conversion of incompatible fertilized eggs into viable haploid males may represent a selective advantage for the more widely distributed N. vitripennis, in that it prevents embryo mortality as a consequence of incompatible matings with its microsympatric sister species (Bordenstein et al., 2003). Taken together, the above illustrates the preeminent role of Nasonia as a model to understand host-Wolbachia interactions, notably in terms of diverse acquisition routes, Wolbachia-host genotype interactions and modes of CI.

The genus Arsenophonus lies in the Gammaproteobacteria and is most closely related to the genera Proteus, Providencia, and Photorhabdus. Members of this genus establish diverse symbiotic interactions with insect hosts, with around 5% of insect species estimated to carry the symbiont (Duron et al., 2008). Many, but not all interactions are based on heritable symbiosis – passage from female to her progeny. However, unlike many heritable symbionts (e.g., Wolbachia), there is substantial diversity in the transmission biology of this microbe. Some strains maintain substantial ‘infectious’ transmissions through the environment, others show vertical transmission via maternal transfer of the microbe to the egg surface followed by ingestion of the symbiont by the hatching larva, and others show classical maternal inheritance through the egg cytoplasm (see Table 1 for the diversity of Arsenophonus-host interactions).

Arsenophonus nasoniae, the symbiont of Nasonia wasps, represents the type species of the genus (Gherna et al., 1991). The discovery of the symbiont followed observations of maternally inherited sex ratio biases in certain N. vitripennis isofemale lines. These lineages were typified by the presence of all (or near all) female broods associated with the death of male progeny, referred to as the ‘son-killer’ trait (Figure 3C) (Skinner, 1985). The trait was heritable through the female line. Subsequent microscopical examination recorded diffuse microbial infections throughout the soma of affected females (Huger et al., 1985). Unusual for heritable microbes, the infection was relatively easily isolated into pure culture, and reinjection into fly pupae alongside ovipositing Nasonia led to establishment of the son-killer trait, fulfilling Koch’s postulates for this microbe (Werren et al., 1986). Later, male-killing was observed to be confined to unfertilized (rather than simply haploid) eggs, and associated with destruction of the maternal centrosome upon which unfertilized eggs depend for development, while diploids, which also inherit a paternal centrosome, do not (Ferree et al., 2008).

The early experiments of Skinner (1985) demonstrated that unlike other heritable microbes known at the time, A. nasoniae possessed the capacity for infectious transmission in addition to maternal inheritance. He observed that when an infected and uninfected female used the same host fly pupa (superparasitism), the progeny from the uninfected female acquired the infection, which would then be passed on in turn by them to their progeny. The infectious transmission is a result of the microbe not being inherited within eggs, but passed through calyx fluid deposited in the fly pupa during oviposition. The microbe grows saprophytically inside the deceased fly puparium, and is ingested by the developing Nasonia larvae, whether they are the progeny of the infected female or are derived from co-parasitising individuals. Following ingestion, the microbe enters the wasp through the gut epithelium to produce the diffuse infection seen in the adult soma. This infection includes presence in the calyx gland, thus ensuring A. nasoniae onward transmission. Arsenophonus thus has a saprophytic stage, a stage where it is a component of the gut microbiota, and a systemic infection stage.

This unusual life cycle is distinct from many other heritable microbe-host interactions and has important biological consequences in terms of the population dynamics of A. nasoniae infections in natural populations. First, superparasitism (and the infectious transmission that follows from it) apparently represents a necessary condition for the maintenance of the microbe (Parratt et al., 2016). In laboratory emulation, A. nasoniae was lost from wasp populations where females were forced to oviposit alone, because vertical transmission is leaky (10% of progeny do not inherit the microbe). In contrast, where females are forced to oviposit in patches with other females and superparasitism rates are high, the infection spreads to fixation in just 3-5 generations. In experiments where the opportunity for superparasitism varies between these global absence/presence extremes, the prevalence achieved is a function of the superparasitism opportunity. Thus, in contrast to other heritable reproductive parasitic microbes, infectious transmission is necessary for A. nasoniae maintenance, and son-killing appears to represent a secondary benefit to the microbe.

A likely consequence of this transmission biology is also the presence of multiple A. nasoniae strains within an individual. The presence of co-infections is likely to erode the correlation between host and microbe fitness that selects for benign and beneficial infections, as the ability to compete (and outcompete) other strains is also a fitness-related trait for the microbe. In this context, it is notable that the genome carries colicin elements whose canonical function is to disable cells that do not carry the element (Wilkes et al., 2010). Further, A. nasoniae represents an antagonist of Nasonia (male-killing makes the element parasitic) that may drive host evolution to prevent its acquisition; as such it may be a driver of immune system evolution that may impact upon other microbiota members.

The transmission pathway also has an important consequence for the movement of the symbiont within chalcid wasp communities. (Duron et al., 2010) noted that uninfected individuals of a different species acquired infection during multiparasitism with an A. nasoniae infected N. vitripennis. This exposure resulted in the efficient transfer of the symbiont into the second species with efficiency comparable to that seen during superparasitism. Using ecological data on multiparasitism rates, the authors concluded that A. nasoniae in a female N. vitripennis had a 12% chance of transfer to N. giraulti in nature. Further to this, the ability of A. nasoniae to be maintained in different species of chalcid wasps in the laboratory was (as above) related to their tendency to superparasitize, with A. nasoniae being lost in species where females were reluctant to utilize already parasitized pupae (Parratt et al., 2016). These results collectively indicate this heritable microbe passes readily between species through multiparasitism, and may pass freely amongst sympatric Nasonia species. Closely related A. nasoniae strains have been retrieved from other chalcid parasitoids (Taylor et al., 2011; Bohacsova et al., 2016), indicating this symbiont infects a wide range of parasitic wasps.

Aside from these biological consequences, the unusual biology of A. nasoniae makes the system more manipulable than most heritable microbe-host interactions. The saprophytic life style stage is almost certainly the reason this microbe grows readily in cell free culture, a property distinct from other heritable microbes (Werren et al., 1986). Further, the strains can easily be reintroduced into the wasp by injection into the host pupa allowing targeted and controlled study of the interaction of the microbe with different wasp genotypes and other elements of the microbiome. Finally, the microbe has intact systems for recombination, and is likely to be genetically manipulable, which would make this one of the few heritable microbes in which functional genetics, both forward and reverse, are possible. However, this ability comes with the caveat that the microbe is not necessarily a reflective model for heritable symbioses in general – it is genetically and biologically more similar to pathogens and gut ‘commensals’ than the classically considered heritable microbiota.

The previous sections have illustrated our growing knowledge regarding binary interactions of Nasonia with either Wolbachia or Arsenophonus. This leads us to an as yet unexplored question: Do the two reproductive parasites also interact with each other in the same host? And if so, what is the outcome? The effects of co-infections studied to date vary immensely. For instance, Wolbachia and Asaia occupy different niches when co-infecting mosquitoes (Hughes et al., 2014; Rossi et al., 2015). In contrast, male-killing Spiroplasma (Mollicutes) have been observed to negatively affect Wolbachia densities when co-infecting D. melanogaster, while Wolbachia had no impact on Spiroplasma (Goto et al., 2006). Phenotypically, Wolbachia has been shown to interfere with Cardinium-induced CI in the spider mite Bryobia sarothamni, although little is known about the underlying mechanisms (Ros and Breeuwer, 2009). If Arsenophonus also had a negative impact on Wolbachia, co-infection could directly affect CI (and thereby population dynamics) in Nasonia, since the strength of CI is dependent on Wolbachia density (Breeuwer and Werren, 1993a). How Arsenophonus and Wolbachia interact remains to be determined, but the symbionts may well represent important drivers of each other’s biology.

The co-existence of different symbionts in the same host environment also creates conditions that are permissive for the exchange of genetic information between symbionts. Hence, the detection of a lateral gene transfer from Wolbachia to the genome of the A. nasoniae strain infecting N. vitripennis (Darby et al., 2010) is interesting in several ways: First, its presence indicates that cross-talk between the two symbionts can occur. Second, the transferred gene is highly similar to a Wolbachia surface protein and likely has a functional role at the host-symbiont interface (Darby et al., 2010). Bacteriophages associated with either bacterium could be potential agents for lateral gene transfers between these two symbionts (and potentially other members of the Nasonia microbiome). Indeed, both Wolbachia and Arsenophonus are known to have associated phages, a rare feature in bacterial endosymbionts and a potential source of genomic innovation mediating the adaptive plasticity of both symbionts (Kent and Bordenstein, 2010; Duron, 2014). Bacteriophage WO is extremely widespread in arthropod-Wolbachia (Masui et al., 2000; Bordenstein and Wernegreen, 2004; Wu et al., 2004; Braquart-Varnier et al., 2005; Gavotte et al., 2007) and renowned for its ability to jump between different Wolbachia strains, especially when infecting the same host (Gavotte et al., 2004; Gavotte et al., 2007; Chafee et al., 2010; Kent et al., 2011). This has led to the “intracellular arena hypothesis” whereby genetic material can be exchanged between different bacterial endosymbionts co-occurring in the same intracellular environment (Bordenstein and Wernegreen, 2004). Indeed, the discovery of a Wolbachia gene in the genome of Arsenophonus (Darby et al., 2010), a prophage-flanked region of the wMel genome on a plasmid of a Rickettsia strain infecting the tick Ixodes scapularis (Ishmael et al., 2009), and numerous lateral gene transfers from other bacteria associated with phage regions in the wBol1 genome (Duplouy et al., 2013) strongly indicate that WO phages may also vehicle genetic material between Wolbachia and other bacterial taxa. Moreover, sequence similarities between phage WO and eukaryotic genes suggest a history of lateral genetic transfers between the two entities (Bordenstein and Bordenstein, 2016). Interestingly, active lytic phages might also be implicated in Wolbachia regulation and reduce the strength of CI via a phage-mediated reduction of Wolbachia titers (Bordenstein et al., 2006; Bordenstein and Bordenstein, 2011). Regarding Arsenophonus, it has recently been shown that the majority of strains harbor the phage APSE (Duron, 2014), previously known from the protective aphid secondary symbiont Hamiltonella defensa (Oliver et al., 2003, 2009), suggesting a transfer of phage elements between Arsenophonus and Hamiltonella. Such an exchange would be all the more relevant as this phage encodes toxicity genes mediating defense against natural enemies of aphids, which happen to be – parasitoid wasps (Oliver et al., 2009). The role of these genes in Arsenophonus, a symbiont of parasitoid wasps, remains to be elucidated. Taken together, this section illustrates the complexity and diversity of potential interactions between only two bacterial symbionts infecting the same host organism, without even considering the complexity of the microbiome as a whole.

The bacterial diversity of Nasonia has been described in lab-reared larvae, pupae and adult males for the three Nasonia species N. vitripennis, N. giraulti, and N. longicornis (Figure 3) (Brucker and Bordenstein, 2012b, 2013). Microbial diversity in these strains ranged from 44 to 83 OTUs at a 97% identity cutoff and varies between host species and developmental stages (Figure 4). Overall bacterial diversity in Nasonia is similar to other Hymenoptera, such as honey bees (Apis mellifera, 82-116 OTUs), bumblebees (Bombus sp., 33-47 OTUs), and fungus farming ants (Mycocepurus smithii, an average of 52 OTUs) (Martinson et al., 2011; Cariveau et al., 2014; Corby-Harris et al., 2014; Kellner et al., 2015). Like most insects, the Nasonia microbiota is dominated by members of the Proteobacteria phylum. The average Nasonia microbiota in adult males is composed of 74.4% Proteobacteria, 15.7% Actinobacteria, and 9.5% Firmicutes (Figure 4) (Brucker and Bordenstein, 2013). Interestingly, at the bacterial genus level, there are three major taxa (Gammaproteobacteria) that account for up to 75% of the male microbiota: Providencia, Proteus, and Morganella (Figure 4). Alone, Providencia sp. compose 59, 68, and 41% of the microbiota in N. vitripennis, N. giraulti, and N. longicornis, respectively. Comparatively, Nasonia is a more tractable laboratory model for controlled experiments and is consistently comprised of 4-5 OTUs that make up the majority of all bacterial sequences.

The two genera Providencia and Proteus are often the most dominant OTUs observed in the three wasp species throughout their development. These same two OTUs are frequently found in the fly host as well, which could represent a natural reservoir of the bacteria for Nasonia. Notably, Nasonia undergoes bacterial community successions throughout its development: The microbial community remains relatively simple when the developing larvae are feeding on the likewise relatively simple microbiota of the fly pupa. Then, microbiota composition shifts during pupation, a time when the wasp is no longer feeding, and again before emergence as adult wasps (Figure 4). As such, Providencia and Proteus represent 95-100% of the microbiota in larvae (Brucker and Bordenstein, 2012b, 2013). Although the microbiota of pupae is less diverse than the microbiota of adults, both tend to exhibit a reduction in the dominance of Proteobacteria (Brucker and Bordenstein, 2012b, 2013).

While little is known about the specific functional roles of the microbiota in Nasonia, several of the major bacterial genera have been previously studied in other insect models. For instance, Proteus has been shown to control the gut microenvironment in blowflies from overgrowth by other bacteria (Erdmann et al., 1984). This colonization resistance could be important for Nasonia, which feed on a decaying pupal fly host. Another major taxon, Providencia, has been implicated in two symbiotic roles: (i) Providing vitamin B to the blood-feeding leech Haementeria officinalis (Manzano-Marin et al., 2015) and (ii) acting as a natural control against the insect pathogen Paenibacillus in the Japanese honeybee Apis cerana japonica (Yoshiyama and Kimura, 2009).

Ongoing studies are now testing the functional significance of the microbiota in different species of Nasonia to determine their role in host development, e.g., through immune regulation, nutrition, and other mechanisms. In addition, transplantations of Nasonia microbiotas between host species will elucidate whether interspecific microbiotas alter host development traits such as larval size, larval and pupal development time or adult viability in comparison to intraspecific microbiotas. In this context, studies in D. melanogaster have demonstrated that axenic individuals suffer from developmental defects along with smaller body size and an altered nutrient metabolism (Shin et al., 2011; Newell and Douglas, 2014). These defects can be rescued by the acetic acid bacterium Acetobacter pomorum, which promotes larval growth and reduces lipid and sugar levels by modulating insulin signaling (Shin et al., 2011). In addition, Lactobacillus plantarum promotes larval growth in conditions of nutrient scarcity by enhancing protein assimilation and TOR-dependent hormonal growth signals (Storelli et al., 2011). In turn, the Drosophila innate immune response is fine-tuned to maintain gut microbiota homeostasis and responds to bacterial pathogens via ROS-production triggered by bacteria-derived uracil, which is released by various opportunistic pathogens but not autochthonous gut microbes (Ryu et al., 2008; Lee et al., 2013).

The microbial community is not limited to bacteria. Nasonia also harbors a diverse set of viruses (Bordenstein and Bordenstein, personal communication) and fungi, and their functional effects on the holobiont await further investigation. While no studies to date have investigated Nasonia’s fungal microbiota, the original draft of the Nasonia transcriptome revealed three novel single-stranded RNA viruses: NvitV-1, -2, -3 (Oliveira et al., 2010). These viruses were not previously found in other insect hosts, though they are related to the Picornavirales—a known order of insect pathogens. The observation of novel viruses in the system is interesting from the perspective that viruses are known to influence the biology of other parasitoid wasps. For instance, the virus Leptopilina boulardi Filamentous Virus (LbFV) manipulates the foraging behavior of its solitary parasitoid wasp host, Leptopilina boulardi, by inducing superparasitism (Varaldi et al., 2003, 2005, 2006). The virus is injected into the fly host together with the parasitoid eggs, allowing it to spread horizontally to uninfected individuals. In contrast to this infectious virus, polydnavirus-like particles have been integrated into the genomes of braconid and ichneumonid wasps and encode particles that contain wasp DNA and proteins which, when injected into the host with the parasitoids’ eggs, enable evasion or direct suppression of the host’s immune response against the parasitoid, thereby contributing significantly to parasitoid fitness (reviewed in Federici and Bigot, 2003).

The changes in the bacterial community throughout development raise questions as to how the Nasonia microbial community assembles through metamorphosis. The answer is not yet clear in any animal system but the patterns exhibited by Nasonia offer an opportunity to better understand how animals change developmentally and anatomically with their microbiota. Since Nasonia embryos are directly deposited within fly host pupae via oviposition, both maternal and fly host microbes could contribute to the initial microbiota assembly of Nasonia larvae. Based on the transmission of microbes in Drosophila (Bakula, 1969), it is possible that Nasonia acquire their first non-endosymbiotic bacteria through ingestion of the chorion during hatching. Alternatively, the microbial community could be passaged via maternal deposition of calyx fluid and venom – using the same process of transmission as Arsenophonus (Huger et al., 1985; Werren et al., 1986). During this event, rare microbes could be introduced into the Nasonia microbiota. Subsequent colonization of the microbiota would then occur through feeding on the fly host. However, the excretion of the larval gut content prior to pupation presents a marked bottleneck for the microbiota. As larvae and pupae develop, it is possible that Nasonia species-specific innate immune genes regulate this community, which would parallel species-specific antimicrobial regulation of the microbiota in Hydra (Franzenburg et al., 2013). On the other hand, the innate immune response of honey bees has been shown to be strongly reduced during the pupal stage compared to larvae and adults (Gatschenberger et al., 2013). If this pattern is consistent across the order Hymenoptera, then a weaker immune regulation during the pupal stage could be influential in the mechanisms that establish the new host species-specific microbiota.

An important aspect that is often overlooked is that microbiota composition may not be regulated solely by host mechanisms, but also through interactions between the microbes themselves. From the microbial perspective, a host organism represents an ecosystem consisting of different microhabitats (i.e., niches) (Sicard et al., 2014), and microbes can be expected to differ in their preference for particular niches. Given that the Nasonia microbiota consists at least partly of bacteria acquired from its fly host during larval development, one might ask whether the transfer to Nasonia as a new host results in fitness consequences for the microbes. While some might be opportunistic and able to find suitable niches, Nasonia may represent a dead-end for other microbes, either due to host factors or competition with other bacteria. The latter may be due to competition for a shared resource/niche and/or by direct interference, for instance via the production of bacteriocidal toxins. Moreover, there may be indirect interactions, mediated through host mechanisms. A particular bacterium may, for instance, activate or suppress the host innate immune system, which then affects the proliferation of other bacteria.

An as yet unexplored but highly relevant aspect is the role of heritable symbionts in the establishment and composition of the Nasonia microbiota. Wolbachia, for instance, are generally highly abundant in various host tissues, thereby limiting available niches and resources for other bacteria (Dittmer et al., 2014, 2016). It is also known to influence other aspects of its host environment, such as immunity, apoptosis and iron homeostasis (Braquart-Varnier et al., 2008; Brownlie et al., 2009; Kremer et al., 2009, 2012; Kambris et al., 2010; Pan et al., 2012). Considering that Wolbachia is ubiquitous in Nasonia under natural conditions, Wolbachia infection represents the natural infection status of Nasonia. Arsenophonus, on the other hand, is a more variable heritable symbiont in this system and has the ability to efficiently infect uninfected individuals via horizontal/environmental transmission within the same fly host (Duron et al., 2010; Parratt et al., 2016). Investigating the impact of Arsenophonus infection on establishment and composition of the wider Nasonia microbiota therefore constitutes a promising line of future research.

Microbiota composition is shaped by both host and environmental factors (e.g., immunity and diet, respectively (Ley et al., 2008; Ochman et al., 2010; Chandler et al., 2011; Colman et al., 2012; Ridley et al., 2012). While the first can be considered deterministic, the latter would be rather stochastic. However, it can be challenging to disentangle these two components and to precisely determine the relative roles of the host versus the environment on the establishment of a species’ microbial community. Controlled conditions can provide evidence for host-microbiota interactions by removing confounding variations in diet, age and gender, for instance. Indeed, under a controlled experimental design, three Nasonia species were found to harbor distinguishable microbiotas whose beta-diversity relationships parallel host phylogeny at all developmental stages (Figure 3A) (Brucker and Bordenstein, 2012b,c). The congruence of host phylogeny and dendrograms reflecting relationships in microbiota composition has since been dubbed “phylosymbiosis” (Brucker and Bordenstein, 2013). For a particular set of closely related species, phylosymbiosis predicts that intraspecific microbial communities are more similar than interspecific communities (Bordenstein and Theis, 2015). Based on that, one could hypothesize that (i) microbiota-based models should predict host species origin with high accuracy, and (ii) various topological congruence analyses of host phylogeny and microbiota dendrograms will reveal significant degrees of phylosymbiosis. Furthermore, if phylosymbiosis were driven by both evolutionary and ecological forces, we might also observe that experimental transplants of autochthonous (intraspecific) versus allochthonous (interspecific) microbiota will drive reductions in host survival and fitness. In addition to Nasonia, the pattern of phylosymbiosis is evident in Hominidae (Ochman et al., 2010; Moeller et al., 2016), Hydra (Fraune and Bosch, 2007; Franzenburg et al., 2013), sponges (Easson and Thacker, 2014), ants (Sanders et al., 2014), and bats (Phillips et al., 2012). One future area of investigation will be to understand the factors influencing phylosymbiosis, e.g., fine-tuned host immune mechanisms and/or different transmission modes (i.e., through maternal transmission or environmental acquisition).

Our growing knowledge of the many ways in which microbial symbionts can induce changes in host phenotypic traits raises the question - to what extent do the microbiota contribute to host diversification, reproductive isolation barriers, and speciation (see Brucker and Bordenstein, 2012c; Vavre and Kremer, 2014; Shropshire and Bordenstein, 2016 for recent reviews)? Isolation barriers can be either pre-mating or post-mating. Pre-mating reproductive barriers may be driven by ecological or behavioral isolation. For instance, particular bacterial symbionts can confer novel traits (e.g., increased thermal tolerance or adaptation to new host plants), allowing their insect host to exploit new ecological niches (Montllor et al., 2002; Ferrari et al., 2004; Tsuchida et al., 2004). Niche expansions such as these can result in geographically or sympatrically isolated populations and, given enough time, lead to speciation. In addition, the microbiota has been implicated in behavioral changes related to mate choice, which may result in symbiont-driven behavioral isolation due to differences in courtship or mate discrimination (reviewed in Shropshire and Bordenstein, 2016). For example, Wolbachia plays a crucial role in driving pre-mating isolation between semispecies of the Drosophila paulistorum species complex (Miller et al., 2010). In addition, the gut microbiota influences kin recognition and mating investment in several Drosophila species (Lize et al., 2014). Specifically, both D. bifasciata and D. melanogaster are able to distinguish between mates that have a more similar or dissimilar microbiota to themselves (Sharon et al., 2010; Lize et al., 2014). This results in a tendency for assortative mating in D. melanogaster after feeding on different food sources (Sharon et al., 2010), although this behavior was replicated only in inbred laboratory lines (Najarro et al., 2015). Similarly, mate selection in scarab beetles is dependent upon immune competence that the females sense in the bacterial-derived male pheromones secretions (Leal, 1998; Vasanthakumar et al., 2008; Andert et al., 2010).

In contrast, post-mating reproductive barriers may be driven by genetic conflicts between host and microbes (i.e., Wolbachia) or a breakdown in holobiont complexes. In Nasonia, both types occur. Wolbachia-induced CI in this system is a pre-eminent case of symbiont-assisted isolation in which nearly complete CI levels (Figure 2) between the Nasonia species cause F1 hybrid lethality that is reversible by curing the Wolbachia infections. In other words, the interspecific F1 isolation is essentially undone with antibiotics that restore production of viable F1 hybrids (Breeuwer and Werren, 1990, 1995; Bordenstein et al., 2001; Raychoudhury et al., 2010a). The study system is notable in that it provided the opportunity to investigate whether Wolbachia-induced CI evolved early or late in the speciation process, i.e., before or after other interspecific pre- or post-mating barriers. While the “older” species pair, N. vitripennis and N. giraulti, diverged ∼1 million years ago and evolved other post-mating barriers such as high F2 hybrid mortality and abnormal courtship behavior (Breeuwer and Werren, 1995; Bordenstein et al., 2001), the very young species pair, N. giraulti and N. longicornis, diverged only ∼400,000 years ago and produce viable and fertile hybrids (Bordenstein et al., 2001). This observation indicates that Wolbachia-induced reproductive isolation via CI preceded the evolution of other post-mating barriers in the younger species pair, and therefore is the first major step in the speciation process (Bordenstein et al., 2001). The even younger species pair, N. giraulti and N. oneida, represents an interesting case in this context: N. oneida females show strong mate discrimination against N. giraulti males, but not vice versa, resulting in strong but incomplete and asymmetrical pre-mating isolation (Raychoudhury et al., 2010a). Moreover, the mate discrimination phenotype in N. oneida is recessive and lost in F1 hybrid females (Raychoudhury et al., 2010a). The impact of Wolbachia on this speciation event is unfortunately blurred by the recent Wolbachia-mitochondrial sweep from N. giraulti into N. oneida (Raychoudhury et al., 2009), which eliminated any Wolbachia-induced incompatibilities that may have existed previously. Therefore, pre-mating isolation is the only barrier currently preventing hybridisation between the two species.

An additional microbiota-mediated reproductive barrier has recently been uncovered in Nasonia, manifested as strong F2 hybrid lethality in interspecies crosses after curing of Wolbachia (Brucker and Bordenstein, 2013). Specifically, hybrid lethality between Nasonia vitripennis and Nasonia giraulti is reversed through removal of the Nasonia gut microbiota, and can be reinstated by inoculating germ-free hybrids with Nasonia-derived bacterial cultures (Brucker and Bordenstein, 2013). Nasonia hybrid lethality was also characterized by an altered gut microbiota in which a rare microbial genus became abundant in hybrids. The change in bacterial community structure was coupled with aberrant host immune gene expression (specifically differential regulation of serine proteases, antimicrobial peptides, and several signaling molecules from the IMD and Toll pathways) compared to the parental species (Brucker and Bordenstein, 2013). In this case, hybrid breakdown at the holobiont level led to severe mortality. This is the first study, to our knowledge, in which the microbial community contributes to hybrid mortality (Figure 2).

Changes in the microbiota could also result in other microbe-dependent reproductive barriers, similar to phenotypes observed in various animal systems, e.g., in terms of development time, behavior, and fecundity (Brucker and Bordenstein, 2012c). For example, species-specific cuticular hydrocarbons help in mate discrimination in Nasonia (Buellesbach et al., 2013), but the impact of the microbiota on mate-choice is unknown. The behavioral issues that arise in hybrids (Clark et al., 2010) may therefore have microbial underpinnings.

A powerful aspect of the Nasonia model lies in the ability to selectively rear (or co-rear) Nasonia hosts in germ-free (Brucker and Bordenstein, 2012a; Shropshire et al., 2016), gnotobiotic (harboring a known, controlled microbiota) and transbiotic (harboring the microbiota of a different species) conditions. The ability to inoculate germ-free Nasonia larvae with monocultures or whole microbial communities will enable high precision studies that deconstruct the effects of specific microbial functions in Nasonia. These studies have the benefit of being implemented at any stage throughout the Nasonia developmental process, which is important to understand the assembly and regulation of the Nasonia microbiota. Specific host genes could also be knocked down to observe their direct effects on microbiota assembly and host-microbe interactions in the four different Nasonia species. With rising interest in utilizing CRISPR genome editing in Nasonia (Lynch, 2015), host gene addition and removal could soon be incorporated into the toolbox available for deciphering hologenomic interactions. With the unique environmental controls afforded by the Nasonia rearing system, there is ample opportunity to study microbiota influences on Nasonia development and fitness. With these tools at hand, the Nasonia model could also be used to experimentally test hologenomic evolution, for instance by exposing a wasp population to a selective pressure (e.g., an environmental stressor) and subsequently monitor (i) whether changes in the microbiota correlate with changes in host life history traits or behavior and (ii) whether this shift in the microbial community persists over multiple generations, as long as the selective pressure persists.