Wolbachia are endosymbiotic bacteria that infect a wide range of invertebrates including up to 66% of insect species, as well as spiders, mites and nematodes [1, 2]. These intracellular, maternally transmitted bacteria have evolved several strategies such as male killing, feminization, parthenogenesis and cytoplasmic incompatibility (CI) to manipulate insect host reproduction. CI is the most common phenotype in insects and results in the developmental arrest of embryos when Wolbachia-infected males mate with uninfected females or with the females infected by a different Wolbachia strain. Intriguingly, a female infected with the same strain of Wolbachia can rescue embryonic lethality associated with CI. Although several models have been described to explain Wolbachia-induced CI [3–5], the molecular mechanisms of CI are still unclear.

Several studies in Drosophila melanogaster and Drosophila simulans have revealed that extensive chromosome bridging appears during early nuclear division in CI embryos, which implied that the chromosomes did not replicate completely, thus resulting in the defects in chromatin condensation and segregation [6–8]. Therefore, sperm derived from Wolbachia-infected males may be modified during spermatogenesis. In addition, although Wolbachia are present within Drosophila testes in developing spermatocytes and spermatids, Wolbachia are not present in mature sperm [9]. Hence, sperm modification by Wolbachia is likely to occur in the early stages of spermatogenesis. However, studies in other insect species infected with Wolbachia suggest that this may not occur for all strains of Wolbachia that induce CI. For example, Wolbachia-induced CI is close to 100% in the parasitic wasp Nasonia vitripennis but Wolbachia were found in only 28% of developing sperm [10]. An alternative explanation involves the presence of a product produced by Wolbachia that can spread through the testes to Wolbachia-uninfected immature sperm [10, 11]. Landmann et al. showed that a delay of histone H3.3 and H4 deposition occurs on the sperm nuclei during formation of male pronuclei in CI embryos [12], which may have resulted in defects of paternal chromosomal replication. Recently, we reported that in both D. melanogaster and D. simulans the expression level of Hira, a gene encoding a chaperone of histone H3.3, is significantly decreased in Wolbachia-infected adult males [13]. In addition, Hira-mutated male flies mimic the CI phenotype resulting in a decreased embryo hatch rate when mated to uninfected females. As there were significantly less female offspring from crosses of Hira-mutated males with Wolbachia-uninfected females [13], we speculated that it was the sperm-carrying X chromosome with mutated Hira that caused embryonic lethality. These results suggest that histone modification in the sperm from Wolbachia-infected males could be involved in embryonic lethality in CI embryos.

Reynolds and Hoffmann previously demonstrated that CI levels dramatically decreased with increasing male age in Drosophila flies infected with Wolbachia[14]. We also recently reported that in both D. melanogaster and D. simulans, 1-day-old males could induce very strong CI, while 5-day-old males were not able to induce CI [13]. These results would suggest that sperm resulting in strong CI is modified by Wolbachia during the early stages of spermatogenesis. An additional study reported that Wolbachia numbers in D. simulans progressively decreased within germ cells from the second larval stage up until adult eclosion [11]. This decrease in Wolbachia numbers may be due to a reduced preference with age for the germ line, resulting in a decreased Wolbachia numbers in spermatocytes and spermatids [11]. Ultimately Wolbachia were eliminated from the mature sperm [11]. Yamada et al. demonstrated that Drosophila male development time could greatly influence the strength of Wolbachia-induced CI. Males that undergo the shortest developmental time in larval stages express strong CI, whereas the males that take a longer time for larval development quickly lose their ability to express CI as adult flies [15]. These results suggest that the regulation of gene expression in spermatogenesis in larval stages may be crucial for the sperm to express the CI phenotype in adult flies. Therefore, in order to investigate the mechanism of CI associated with early modification of sperm by Wolbachia, we selected 3^rd instar larval testes for gene expression analysis comparing transcriptional profiles between Wolbachia-infected (Dmel w Mel) and uninfected D. melanogaster (Dmel T) larval testes. Differential expression was observed in 296 genes related to reproduction, immunity, metabolism, and other functions. Expression profiles for 12 genes were confirmed by Quantitative Reverse Transcriptase PCR (qRT-PCR) and in situ hybridization was carried out for four genes. Our data provide a comprehensive analysis of differential gene expression between Wolbachia-infected and uninfected larval testes of Drosophila. These results provide important information to help determine the molecular mechanisms involved in Wolbachia's interaction with their hosts and, in particular, the role of Wolbachia on spermatogenesis leading to induction of CI.

Previous studies revealed that the strength of CI in Wolbachia-infected Drosophila was dramatically reduced with both male age [13, 14] and increased larval stage development [15]. Therefore, Wolbachia action during spermatogenesis in larval stages of Drosophila may be important for induction of CI. In this study, we selected 3^rd instar Drosophila larval testes to investigate the effect of Wolbachia on early spermatogenesis and provide a basis to understand the underlying mechanisms of CI. A genome-wide analysis of the Wolbachia/host interaction in Drosophila larval testes was conducted and a number of genes showing differentially expression between Wolbachia infected and uninfected larval testes were identified. These genes are involved in diverse functions including reproduction, oxidation-reduction, immunity, transportation and metabolism.

The majority of genes involved in reproduction were down-regulated in the presence of Wolbachia, including CG8827-RB (Ance) and CG12052-RP (Lola). Some of these genes have previously been reported to be associated with the interaction between Wolbachia and their insect hosts. For example, angiotensin converting enzyme (Ance) gene has been demonstrated to be down-regulated in the adult testes of Wolbachia-infected flies relative to uninfected flies in both D. melanogaster and D. simulans[17]. This suggests that Ance gene may play a potential role in the interaction between Wolbachia and their hosts. There is also further evidence that Ance is involved in spermatid differentiation [18]. A potential role of Ance in CI was also shown with mutation affecting CI levels [17]. The Longitudinals lacking (Lola) gene is associated with cellular processes involved in reproduction. Lola mutant flies showed a block in developmental nurse cell death and abnormal nuclear organization [19]. During programmed cell death in Drosophila ovaries, chromatin condensation is not complete in Lola-mutant flies [19]. In the present study, Wolbachia infection results in significant down-regulation of Lola gene expression in larval testes, which in turn may induce abnormal chromatin condensation during spermatogenesis.

An additional reproduction gene identified to be differentially expressed, CG32491-RP, is an alternative splice product of mod (mdg4) previously observed in vitro[17]. Soltani-Bejnood et al. found a mutant in the common region of mod (mdg4) expressed in Drosophila spermatocytes but confined to cytoplasm during prophase I and demonstrated that the common region of mod (mdg4) plays an important role in homolog segregation during Drosophila male meiosis [20]. Therefore, we can speculate that Wolbachia may affect Drosophila spermatogenesis by regulating the expression of mod (mdg4). Another gene, CG17934-RA (Mst84Db), encodes for a male specific sperm protein. Kuhn et al. previously demonstrated that Mst84Db is expressed exclusively in the male germ line and associated with sperm motility [21]. Further studies revealed that there was not only a translational repression element but also a transcriptional control element within the Mst84Db gene [22], suggesting that Mst84Db could also be regulated at the transcriptional level. Our data obtained from microarray, qRT-PCR and in situ hybridization showed that the Mst84Db gene was down-regulated in Wolbachia-infected larval testes relative to uninfected testes, suggesting that Mst84Db could be inhibited at the transcriptional level by Wolbachia directly or indirectly in testes. Recent work revealed that there are multiple altered sperm structures during the late stage of spermatogenesis in Wolbachia-infected males [11]. Therefore it is likely that Wolbachia infection inhibits the transcription of Mst84Db during spermatogenesis resulting in modified sperm.

In addition to genes involved in reproduction, microarray analyses revealed differential expression of genes involved in insect immunity between Wolbachia-infected and uninfected D. melanogaster larval testes. One of the first lines of defence of insects against invading microbes is the generation of reactive oxygen species (ROS). However, high concentration of ROS may lead to oxidative stress and potentially damage lipids, nucleic acids and proteins, thus resulting in a reduction in insect lifespan [23]. Wolbachia has been shown to disturb the cellular physiology of its insect host especially via the generation of oxidative stress [24]. Correspondingly, an increase of antioxidant expression in mosquito cells is induced by Wolbachia, which could be an adaptation to symbiosis [24]. Moreover, Wolbachia that infects Asobara tabida interferes with iron, which limits oxidative stress and cell death, thus promoting its survival within host cells [25]. Recently fecundity of Wolbachia-uninfected A. tabida females was correlated with variable expression of genes regulating iron homeostasis and oxidative stress [26]. Our microarray data also show elevated expression levels in Wolbachia-infected larval testes for multiple genes involved in oxidation-reduction. Up-regulated gene expression was observed for CG9081-RA (Cyp4s3), which is involved in the oxidation-reduction process, CG6770-RA which is associated with response to oxidative stress, and CG12262-RA which is correlated with fatty acid beta-oxidation. This suggests the possibility that Wolbachia might regulate the redox reaction by various pathways to neutralize the potentially deadly ROS and thus maintain the Wolbachia/host symbiotic relationship.

Up-regulation of immune genes may represent 'detection' of Wolbachia as occurs in the relationship between the primary endosymbiont of the weevil Sitophilus zeamais through up-regulation of three local immune genes in the bacteriome [27]. We identified multiple up-regulated genes involved in immunity in Wolbachia-infected larval testes relative to uninfected ones, including two genes encoding antimicrobial peptides (Lysozyme E and Drosomycin). In addition, the CG16910-RA (kenny) gene was also up-regulated in Wolbachia-infected testes and located in specific regions of larval testes (Figure 3F). Previous studies have shown that IKKβ (encoded by ird5) and IKKγ (encoded by kenny) constitute the Drosophila IKK complex which directly phosphorylates Relish in the Imd immune signalling pathway [28]. The phosphorylated Relish then recruits RNA polymerase II and induces the expression of antimicrobial peptides. Furthermore, ird5 has been shown to be down-regulated while Relish and some antimicrobial peptides (attacin A, B, C, D and diptericin B) are up-regulated in Wolbachia- infected Drosophila S2 cells [17]. These experiments suggest that high levels of kenny gene expression in Wolbachia-infected testes probably contributes to phosphorylate Relish and thus induces the production of antimicrobial peptides in testes. Earlier work demonstrated that Wolbachia had no effects on the transcription of three antimicrobial peptide marker genes in adults of two insect species, D. simulans and Aedes albopictus, naturally infected with Wolbachia[29]. Therefore, induction of the insect immune system appears to not occur with naturally occurring Wolbachia infections. However, several lines of evidence have shown that Wolbachia infection does increase resistance against pathogens, such as viruses and filarial nematodes in both naturally infected hosts and artificially transinfected hosts [30–33]. Moreover, Wolbachia infection has been demonstrated in both its original host D. melanogaster and a novel mosquito host (Aedes aegypti) to be able to increase the levels of melanization which is a major component of the insect immune system [34]. Cook and McGraw raised two alternate explanations that do not involve an immune response: Wolbachia might mediate modification of the membrane, thus prevent entry of pathogens into host cells, or there may be direct competition between Wolbachia and pathogens for an intracellular resource [35]. Clearly the mechanism by which Wolbachia is able to increase pathogen resistance of insect hosts needs to be further investigated. Further experiments are required to determine if naturally occurring avirulent or low-density Wolbachia strains can activate the insect host immune response.

Our microarray analyses also highlight a large number of differentially expressed genes that code for proteins involved in metabolism including CG32954-RF (coding for alcohol dehydrogenase), CG8627-RA (involved in cellular acyl-CoA homeostasis), and CG8782-RA (coding Ornithine aminotransferase precursor). CG4988-RA, which codes for aldose 1-epimerase, was found to exhibit a lower expression level in Wolbachia-infected larval testes suggesting Wolbachia infection could decrease hexose metabolism. Up-regulation expression of CG2718-RB, which codes for glutamine synthetase 1, suggests that glutamate synthesis is increased in Wolbachia-infected testes. As glutamate importers were identified in the sequenced genome of w Mel [36], it is reasonable to suggest that Wolbachia may use host glutamate as an important component for a variety of metabolic pathways. Indeed, the genome sequence of the w Mel strain revealed that Wolbachia does not contain the complete set of metabolic pathways present in free-living bacteria [36]. Wolbachia probably only use a limited number of substrates and synthesizes very few metabolic intermediates. The successful survival and proliferation of endoysmbiotic Wolbachia in many host species may be due to the effect of Wolbachia on host metabolism to obtain most of the energy by importation of amino acids and other metabolites. It is likely that Wolbachia affects the expression of its host genes involved in metabolic pathway indirectly, namely, Wolbachia presumably consumes metabolites from the host, and then the host has to up-regulate the expression of metabolic related genes to increase the biosynthesis of that metabolite. In addition, the w MelPop strain of Wolbachia was shown to increase both locomotor activity and metabolic rate in Aedes aegypti[37], suggesting Wolbachia can manipulate host metabolism by inducing changes in expression levels of host metabolic genes.

Microarray data also reveals numerous differently expressed genes that are involved in transportation, including CG4450-RA (Shawl), which is involved in transmembrane transport and potassium ion transport, and CG13795-RA, which is related to neurotransmitter transport. As discussed previously, the transportation of host metabolites appears to be critical to the survival of Wolbachia in insect hosts. Interestingly, studies in Nasonia vitripennis wasps have revealed although Wolbachia are found in only ~28% of developing sperm, all sperm are modified. In the beetle Chelymorpha alternans, Wolbachia can modify up to 90% of sperm, but are never observed within the developing sperm or within the surrounding cyst cells, though they are abundant within the outer testis sheath [10]. These observations suggest Wolbachia may produce some factors that can cross multiple tissue membrane barriers to effect developing spermatids. Wolbachia is known to possess a type IV secretion system, which is likely used for exporting molecules into host cells [38]. Although the molecules that Wolbachia secretes into host cells are currently unknown, it is reasonable to suggest that various transporters may be required to establish molecules in locations in which they interact with the insect host.

This interaction between Wolbachia and the insect host may also be influenced by differential expression of other genes with unknown functions. For example, Juvenile hormone (JH) has been shown to play a key role in regulating both development and reproduction in insects [39, 40]. In D. melanogaster, Dubrovsky et al. demonstrated that JH could induce the gene CG3767-RA (JhI-26) rapidly and specifically [41]. High expression levels of JhI-26 were observed in adult male accessory glands, although it was absent during the period from late third instar larvae to eclosion [41]. Recently JhI-26 has been identified as a sperm protein of D. melanogaster[16], implying that it could be associated with sperm function. In our study, we found that Wolbachia infection results in ~10 fold increase of JhI-26 transcription in later larval testes. This high level of JhI-26 gene expression may inhibit testes development in the later larval stages. Therefore, sperm produced by Wolbachia-infected young males that develop fastest in larval stage may be mostly immature. Yamada et al. observed that male Drosophila developmental time influences the strength of Wolbachia-induced CI. Male flies that have the shortest development time in larval stages express very strong CI, while males that spend a longer time in larval development quickly lose their ability to express CI [15]. A longer larval development time may allow the testes to develop completely resulting in young males with fully mature sperm not able to induce CI. Furthermore, studies in Nasonia CI embryos have shown a delay in nuclear envelope breakdown and Cdk1 (cyclin-dependent kinase 1) activation in the male pronucleus relative to the female pronucleus [42]. During the first mitotic division, when the maternal chromosomes enter the metaphase/anaphase transition, the improperly condensed paternal chromosome remnants fail to segregate and remain arrested in metaphase [7]. This is probably due to a delay in recruiting the replication-independent histone H3.3/H4 complex to the male pronucleus [12]. The delay in development of male pronucleus or paternal chromosomes may be involved in inhibiting the production of fully mature sperm. Therefore, it is possible that Wolbachia may mimic the function of JH to induce the expression of JhI-26. As high expression of JhI-26 may inhibit testes development in the larval stages, male flies that develop fastest may produce incompletely matured sperm. Immature sperm may then result in the delayed development of the male pronucleus and high embryonic lethality associated with CI upon fertilization of eggs from Wolbachia-uninfected females.

In this study, microarray, qRT-PCR and in situ hybridization were used to show numerous differentially expressed genes between testes of Wolbachia-infected and uninfected Drosophila melanogaster larvae. The genes showing variable expression levels are involved in diverse functions including metabolism, immunity and reproduction. The data presented here provides a significant panel of candidate genes to investigate the underlying mechanisms of the Wolbachia/host interaction. In particular, the genetic basis of CI and the role of Wolbachia in spermatogenesis in Drosophila. We are currently undertaking additional experiments with several candidate genes using overexpression or RNA interference to determine their function to provide further insight into this complex endosymbiotic relationship.

Drosophila melanogaster flies were reared on a standard cornmeal/yeast diet at 25°C with a photoperiod of 8L:16D (light:dark) and under non-crowded conditions (200 ± 10 eggs per 50 ml vial of media in 150 ml conical flask) [15]. The Wolbachia-infected Drosophila strain designated Dmel w Mel (D. melanogaster Brisbane nuclear background with introgressed w Mel from YW) was kindly provided by Prof. O'Neill S (Monash University, Australia). Wolbachia-uninfected Dmel w Mel T was subsequently generated by tetracycline treatment following established protocols [43] and confirmed to be Wolbachia-free by PCR with Wolbachia surface protein gene (wsp) primers (Additional file 4) (data not shown).