In the present study, the microbial community of various populations of two predators of the mirid genus Macrolophus was investigated. The bacterial diversity of Macrolophus spp. was explored by cloning 16S rRNA sequences and PCR-DGGE. The cloning experiment was executed on the laboratory strain of M. pygmaeus, revealing the presence of bacteria from the Alpha-proteobacteria, Beta-proteobacteria, Gamma-proteobacteria and Firmicutes classes (Table 3). Three bacteria -R. limoniae, R. bellii and Wolbachia- can be considered as endosymbionts. The presence of these endosymbionts was confirmed using a PCR-DGGE profile of the hypervariable V3 region of the 16S rRNA gene. The PCR-DGGE was carried out using a semi-nested approach, as the bacterial primers targeting the V3-region are known to amplify eukaryotic DNA . Three bands corresponding to these three endosymbionts recurred in all studied M. pygmaeus populations. The DGGE-profile of bacteria in the M. caliginosus populations were similar to those of M. pygmaeus, confirming the presence of Wolbachia and the Rickettsia strain from the ‘Limoniae’ group, but the bellii-like Rickettsia was not found (Fig. 2). A PCR using specific primers for each endosymbiont confirmed this result.
The bands with lower density present in some populations corresponded to the Gamma-proteobacteria and Firmicutes. Most of these bands were attributed to Serratia species of the Enterobacteriaceae family, which have been found in the gut of various insect orders, including Hymenoptera, Lepidoptera, Neuroptera and Hemiptera [53–56]. One band however (Fig. 2, no. 7), has been amplified in five wild Macrolophus populations. This band corresponded to an uncultured Gamma-proteobacterium, the role of which is unknown. The low bacterial diversity in the gut of M. pygmaeus may be attributed to its natural diet. A more diverse bacterial community is mostly detected in insects that consume nutritionally poor diets , whereas the main food of Macrolophus bugs consists of nutrient-rich arthropod prey. Also, the microbial diversity of the investigated Macrolophus spp. may have been underestimated by the dominance of the endosymbionts in its host. Samples of the wild Macrolophus populations were collected in ethanol and DNA-extraction was performed on whole adults; gut dissections were thus only feasible for the two laboratory reared populations. The faint bands in the DGGE-profile of the wild populations of Macrolophus may originate from prey remnants in the gut. A PCR-DGGE profile of the gut of the laboratory populations of M. pymaeus and M. caliginosus established the presence of the Gamma-proteobacteria and the Rickettsia endosymbionts in M. pygmaeus (Fig. 3), whereas the gut of M. caliginosus was only infected by R. limoniae. In both species, Wolbachia was virtually absent in the gastro-intestinal tract.
The DGGE profile of the ovaries only indicated an infection by the Wolbachia and Rickettsia endosymbionts, suggesting that no other bacteria infected the reproductive tissues. A diagnostic PCR on adults and ovaries of M. pygmaeus and M. caliginosus confirmed that all individuals are multiple infected and that the endosymbionts are vertically transmitted, implying that the infections are fixed. A FISH analysis confirmed high densities of both Wolbachia and Rickettsia in the ovarioles of M. pygmaeus (Fig. 4 and 5), suggesting a high rate of vertical transmission to the progeny .
Wolbachia is the only endosymbiont infecting the studied Macrolophus spp. which is known to cause CI in its insect host . As there is no evidence in the literature that Rickettsia is involved in causing CI effects in insects, the strong CI observed in M. pygmaeus  is more likely related to the presence of Wolbachia rather than the Rickettsia species. The impact of the Rickettsia species on the biology of Macrolophus bugs is as yet unclear. A bio-assay was performed to examine differences in development and fecundity between an endosymbiont-infected and a cured population of M. pygmaeus. In accordance with the findings of Chiel el al.  on the tobacco whitefly B. tabaci, nymphal development of infected individuals was faster (albeit in the current study only for males), but fecundity was not affected. On the other hand, Himler et al.  demonstrated the rapid spread and fixation of a southwest American whitefly population infected with Rickettsia bellii. This population dominated all other populations by large fitness advantages and a higher proportion of females. Although the proportion of females was also higher in the infected M. pygmaeus population in our study (Table 4), the observed effects do not allow to explain the Rickettsia fixation in Macrolophus.. The Rickettsia symbiont of the booklouse L. bostrychophila is essential for the development of the embryos . Conversely, cured M. pygmaeus adults produce normal progeny, confirming the facultative secondary character of Rickettsia in this host. Theoretically, the Rickettsia endosymbionts could have invaded its Macrolophus host by ‘hitchhiking’ with the CI-inducing Wolbachia endosymbiont, as CI promotes females with multiple infections .
Besides influencing developmental and reproductive parameters, microbial endosymbionts can affect their host in various other ways, e.g. by being nutritional mutualists. Recently, Wolbachia has been shown to provide a positive fitness effect in iron-restricted diets . Also, the so-called ‘symbiont-mediated protection’ is an emerging topic [2, 3, 59]: here, insects are protected against pathogens (including viruses [51, 63] and fungi ) or parasitoids (e.g. the braconid wasp Aphidius in aphids ) by vertically transmitted symbionts (reviewed in ). This protection could be a potential system for endosymbionts to preserve their infection.
To clarify the impact of the individual endosymbiont species, their hosts can be partially cured, yielding singly infected individuals. White et al.  used low dose antibiotics to partially cure the doubly infected parasitoid wasp Encarsia inaron. This wasp needed to be cured of Wolbachia and Cardinium, two endosymbionts belonging to two different classes, the Alpha-proteobacteria and Bacteroidetes respectively. However, Rickettsia and Wolbachia belong to the same family (Rickettsiaceae), which would complicate partial curing in Macrolophus. The role of Wolbachia and Rickettsia in M. caliginosus has not been demonstrated. Establishing an endosymbiont-free population and performing crossing experiments can be a first step to investigate possible reproductive effects also in the latter Macrolophus species.
A Rickettsia-specific phylogenetic tree elucidated that one M. pygmaeus Rickettsia endosymbiont belonged to the ‘Limoniae’ group, whereas the other is a member of the ‘Bellii’ group (Fig. 1). The M. pygmaeus Rickettsia endosymbiont belonging to the ‘Bellii’ group was phylogenetically closely related to the symbionts of natural prey species of the mirid predator, including the two-spotted spider mite T. urticae, the pea aphid A. pisum and the tobacco whitefly B. tabaci. This finding may indicate a possible horizontal transfer between predator and prey. The horizontal transfer of an endosymbiont has, however, currently only been established in an arthropod parasitoid-host system. Chiel et al.  investigated the interspecies horizontal transfer of Rickettsia from B. tabaci (belonging to the ‘Bellii’ group) to its aphelinid parasitoids Eretmocerus emericus and E. emiratus. This Rickettsia infection reached the reproductive tissues of its host, but was not transmitted to its progeny.
Sharing the same habitat and using the same plant tissues may also constitute a transmission route for bacterial endosymbionts. Macrolophus spp. are facultatively phytophagous predators with piercing-sucking mouthparts and may inoculate plant tissues with micro-organisms. Other species, feeding on the same host plant may then take up these micro-organisms. Furthermore, the PCR-DGGE profile showed the presence of R. limoniae and R. bellii in the gut, suggesting that an infection of the faeces is likely. However, more research is needed to confirm these hypothetical horizontal transmission routes.