Wolbachia-Mediated Antibacterial Protection and Immune Gene Regulation in Drosophila

The outcome of microbial infection of insects is dependent not only on interactions between the host and pathogen, but also on the interactions between microbes that co-infect the host. Recently the maternally inherited endosymbiotic bacteria Wolbachia has been shown to protect insects from a range of microbial and eukaryotic pathogens. Mosquitoes experimentally infected with Wolbachia have upregulated immune responses and are protected from a number of pathogens including viruses, bacteria, Plasmodium and filarial nematodes. It has been hypothesised that immune upregulation underpins Wolbachia-mediated protection. Drosophila is a strong model for understanding host-Wolbachia-pathogen interactions. Wolbachia-mediated antiviral protection in Drosophila has been demonstrated for a number of different Wolbachia strains. In this study we investigate whether Wolbachia-infected flies are also protected against pathogenic bacteria. Drosophila simulans lines infected with five different Wolbachia strains were challenged with the pathogenic bacteria Pseudomonas aeruginosa PA01, Serratia marcescens and Erwinia carotovora and mortality compared to paired lines without Wolbachia. No difference in mortality was observed in the flies with or without Wolbachia. Similarly no antibacterial protection was observed for D. melanogaster infected with Wolbachia. Interestingly, D. melanogaster Oregon RC flies which are naturally infected with Wolbachia showed no upregulation of the antibacterial immune genes TepIV, Defensin, Diptericin B, PGRP-SD, Cecropin A1 and Attacin D compared to paired flies without Wolbachia. Taken together these results indicate that Wolbachia-mediated antibacterial protection is not ubiquitous in insects and furthermore that the mechanisms of antibacterial and antiviral protection are independent. We suggest that the immune priming and antibacterial protection observed in Wolbachia-infected mosquitoes may be a consequence of the recent artificial introduction of the symbiont into insects that normally do not carry Wolbachia and that antibacterial protection is unlikely to be found in insects carrying long-term Wolbachia infections.


Introduction
The interaction between two microbes within a host can impact on the outcome of infection for the host. Wolbachia are a maternally transmitted endosymbiotic a-proteobacteria that is predicted to infect up to 70% of insect species [1,2]. Wolbachia can protect insects from infection by a range of microbes and parasites [3][4][5][6][7][8][9][10]. Where the microbes are pathogens of the insect, this protection has the potential to greatly influence the ecology of the host, pathogen and Wolbachia [11][12][13]. In addition, it has been widely suggested that Wolbachia-mediated pathogen protection could be harnessed in biological control programs to interfere with the transmission of human diseases that are vectored by insects, including dengue and malaria. However, the molecular mechanisms involved in protection are yet to be determined.
Mosquitoes that are experimentally infected with Wolbachia are protected from a range of viruses, bacteria and parasites. A number of mosquito species that are important human disease vectors are not naturally found to be infected with Wolbachia, for example Aedes aegypti and Anopheles species. However, utilising transinfection techniques Ae. aegypti has been experimentally infected with Wolbachia strains wMelPop-CLA or wAlbB [14,15]. The mosquitoes stably infected with Wolbachia accumulate and transmit RNA viruses such as Dengue and Chikungunya less readily than Wolbachia-free mosquitoes [3,8]. Infection of Ae. aegypti with the D. melanogaster derived Wolbachia strain wMelPop-CLA also reduced the prevalence of the filarial nematode Brugia pahangi, impaired the ability to transmit the avian malarial parasite Plasmodium gallinaceum. In addition reduced mortality induced by infection with the Gram-negative bacterium Erwinia caratovora but not the Gram-positive bacteria Micrococcus luteus was observed in these mosquitoes [7]. In addition, whilst stable transinfection of Anopheles mosquitoes is yet to be achieved, An. gambiae that were somatically infected with Wolbachia showed reduced accumulation of Plasmodium oocysts [5,6]. Thus in mosquitoes, artificially introduced Wolbachia induces broad ranging antipathogen protection.
The phenomenon of Wolbachia-mediated antiviral protection is well established in the model insect Drosophila. Naturally Wolbachiainfected flies are protected from a diverse range of RNA viruses [4,10]. In the case of the pathogenic viruses Drosophila C virus (DCV), Flock House virus (FHV) and Cricket paralysis virus (CrPV) Wolbachia-infected flies survive upwards of twice as long as their unprotected Wolbachia-free counterparts [4,10]. This protection has been shown to be consistent across the closely related Wolbachia strains that infect D. melanogaster (wMel, wMelCS and wMelPop) and across different host backgrounds [4,10]. Wolbachia antiviral protection has also been demonstrated in the related species D. simulans [9].
Not all Wolbachia:host combinations result in antiviral protection. D. simulans are naturally infected with diverse Wolbachia strains from both supergroup A (wAu, wRi and wHa) and B (wNo) [16,17]. D. simulans lines naturally infected with wAu and wRi (line CO and DSR respectively) are protected from DCV and FHV, whereas those naturally infected with wHa and wNo (line DSH and N7NO respectively) are not protected [9]. In these fly lines antiviral protection correlates with both phylogenetic relatedness to wMel and also high density of Wolbachia in the host [9]. The D. simulans line Me29, which was transinfected with wMel in 1998 [18], is also protected from both DCV and FHV infection.
The mechanism(s) of Wolbachia-mediated protection have not been determined. The correlation between density and distribution of Wolbachia in flies and mosquitoes supports the hypothesis that Wolbachia and pathogens may be in competition for limited host resources [8,9]. Alternatively several studies have demonstrated that antipathogen protection in experimentally infected mosquitoes is concomitant with Wolbachia induced upregulation of a range of host immune genes [3,[5][6][7][8]. Genes involved in the antimicrobial IMD and Toll pathways are upregulated, with the effector genes such as cecropins and other antimicrobial peptides (AMPs) showing the highest upregulation. These observations led to the hypothesis that Wolbachia infection primes the immune system so that when Wolbachia-infected insects are challenged with a pathogen the insect is protected from the pathogen. Little direct evidence is available in support of this hypothesis, although in somatically infected An. gambiae upregulation of Tep1 has been experimentally linked with protection against parasite P. berghei [6] and Dengue virus is somewhat impacted by upregulation of the Toll pathway [19].
In contrast to mosquitoes, it is less clear if Wolbachia stimulates immune priming in Drosophila. Cultured D. melanogaster cells (S2 cell line) showed upregulation of genes from the Toll and IMD pathways as well as AMP effector molecules when experimentally infected with the Wolbachia wRi strain (which naturally infects D. simulans) [20]. In contrast, an early study using Northern blot analysis found no difference in cecropin or diptericin RNA levels in D. simulans line DSR with and without wRi infection and similarly no difference was recorded for Defensin expression in Ae. albopictus with and without Wolbachia infection [21]. It remains to be confirmed whether Wolbachia-mediated immune priming is linked to the antiviral protection that has been documented in Drosophila.
In mosquitoes Wolbachia mediates protection against a range of pathogens, yet it is unclear whether a single molecular mechanism underpins protection against this diverse group of microbes and parasites. In Drosophila not all host:Wolbachia combinations protect against virus infection [9]. If the mechanism that underlies protection were the same for bacteria and viruses we would predict that those Drosophila:Wolbachia combinations with reduced virusinduced mortality would similarly reduce pathogenic bacterial infection.
In this study we investigated whether antibacterial protection occurs in flies infected with Wolbachia. To do this five D. simulans:Wolbachia lines were utilised that we previously used to investigate antiviral protection. Bacterial pathogens Pseudomonas aeruginosa PA01, Serratia marcescens and E. carotovora were used to challenge the flies. P. aeruginosa PA01 and S. marcescens are the opportunistic and natural pathogens of Drosophila respectively [22,23] and Wolbachia mediates protection in Ae. aegypti against mortality induced by E. carotovora [7]. Further, using reverse transcription and quantitative PCR to assay expression of six AMPs and immune genes there was no evidence of antibacterial immune priming in D. melanogaster naturally infected with Wolbachia.

Results
Wolbachia does not protect D. simulans from pathogenic bacteria The impact of Wolbachia on the outcome of virus infection varies in D. simulans lines challenged with RNA viruses. D. simulans lines CO, DSR and Me29 are protected against DCV and FHV infections by Wolbachia strains wAu, wRi and wMel respectively. In contrast, D. simulans lines DSH and N7NO are not protected against DCV and FHV infections by wHa and wNo respectively [9]. In order to investigate whether these Wolbachia strains confer protection to D. simulans lines challenged with pathogenic bacteria, D. simulans CO, DSR, Me29, DSH and N7NO lines with Wolbachia and paired lines that had been cured of Wolbachia infection were challenged with three pathogenic Gram-negative bacteria (P. aeruginosa PA01, S. marcescens and E. carotovora) and mortality recorded for up to 36 hours.
Mortality of CO flies challenged with pathogenic bacteria was similar regardless of Wolbachia infection status ( Figure 1). Flies both with and without Wolbachia challenged with P. aeruginosa PA01 died within 25 hours of infection and there was no significant difference in the survival curves ( To investigate whether lack of protection in D. simulans was limited to this particular host:Wolbachia combination we challenged four other fly lines with and without Wolbachia with the three pathogenic bacteria. DSR, Me29, DSH and N7NO that were challenged with P.aeruginosa PA01 died within 25 to 30 hours post infection (Figure 2 A, 3 A, 4 A and 5 A). D. simulans DSR, Me29, DSH and N7NO when challenged with S. marcescens and E. carotovora, died within 10 to 25 hours post infection (Figure 2 B and C, 3 B and C, 4 B and C, 5 B and C). In each assay there was no difference between the survival curves of flies with and without Wolbachia challenged with each of the pathogenic bacteria (p.0.05). Results shown are representative of at least three independent bioassays, each with similar results obtained. The survival bioassays were also repeated with at least two independent experiments with a lower concentration of bacterial culture (O.D. 600 nm = 0.1-0.5) (data not shown). At this lower concentration of bacteria, 100% mortality was not achieved in most cases, however, there was still no protection against bacterial-induced mortality. Taken together these results give strong evidence that Wolbachia does not confer protection in D. simulans against bacterial-induced mortality.

Analysis of Wolbachia-mediated antibacterial protection in D. melanogaster
In contrast to Ae. aegypti mosquitoes [7] our results show that there is no antibacterial protection mediated by Wolbachia in D. simulans. To investigate whether lack of protection was restricted to this single species we utilised the D. melanogaster line ORC in protection bioassays. This species was chosen as Wolbachia has been shown to protect D. melanogaster from a number of different viral pathogens [4,9,10]. ORC flies with and without wMelCS were challenged with P. aeruginosa PA01, S. marcescens and E. carotovora and survival was monitored ( Figure 6). There was no difference in the survival of flies with and without Wolbachia in response to S. marcescens challenge (p = 0.3). ORC flies with Wolbachia were somewhat more susceptible to P. aeruginosa PA01 and E. carotovora than flies without Wolbachia (p,0.05) in the results shown in Figure 6. However, this small difference was only observed in one out of three experiments, suggesting that if

Regulation of immune genes in D. melanogaster in response to Wolbachia infection
Wolbachia has been shown to stimulate different immune gene responses in different Wolbachia:host combinations [6][7][8]20,21,24,25] and it has been suggested that immune priming stimulated by Wolbachia may be causally linked to Wolbachia-mediated antipathogen protection [3,[6][7][8]. Given no antibacterial protection was mediated in the six host:Wolbachia combinations utilised in this study we wanted to investigate whether antibacterial immune  genes were upregulated by the presence of Wolbachia in our system. The regulation of six immune genes in D. melanogaster ORC flies with and without Wolbachia was investigated using RT-qPCR. These genes were chosen as they were homologous to genes that were upregulated by Wolbachia presence in Ae. aegypti [7]. None of the six immune genes that were investigated in Drosophila (TepIV, Defensin, Diptericin B, PGRP-SD, Cecropin A1 and Attacin D) were differentially regulated in the presence of Wolbachia (p,0.05; Figure 8).

Discussion
The lack of Wolbachia-mediated antibacterial protection in Drosophila differs from studies in mosquitoes where Wolbachia has been shown to mediate broad-spectrum anti-microbe and -parasite protection [3,[5][6][7][8]. In mosquitoes experimentally infected with Wolbachia there is upregulation of a number of genes for immune effector molecules and those involved in antimicrobial pathways [3,7,8]. It has been suggested that this Wolbachia induced ''immune priming'' may be the mechanism underlying protection and some evidence has been presented for this for anti-Plasmodium protection [6]. We reasoned that as Wolbachia-infected Drosophila were not protected from pathogenic Gram-negative bacteria, this may be because Wolbachia was not stimulating an immune response in these flies. To assess this we analysed the expression of six antimicrobial immune genes whose homologues were upregulated in Wolbachia-infected mosquitoes [7]. In D. melanogaster Wolbachia infection did not stimulate expression of the antimicrobial response genes.
It is interesting that Wolbachia differentially stimulates immune responses in different hosts. The host:Wolbachia examples discussed above differ in two ways. Firstly, the hosts are from different insect families: mosquitoes and flies. But perhaps more importantly each analysis of immune regulation has been performed using mosquitoes that have been recently experimentally infected with Wolbachia. In contrast the D. melanogaster ORC line used here, is naturally infected with Wolbachia. As a  maternally inherited endosymbiont, Wolbachia remains in close association with its host from generation to generation. Coevolution of the bacteria and host is postulated to eventuate in commensal or mutualistic associations. Experimental introduction of Wolbachia into a new host can lead to over-replication of the bacteria and pathogenicity, although these effects can be ameliorated in later generations [5,26,27]. This raises the possibility that Wolbachia induced immune priming is a consequence of maladapted interactions following experimental introduction of Wolbachia into a new host. This premise is supported by previous studies that show upregulation of immune genes in D. melanogaster cell line S2 experimentally infected with D. simulans derived Wolbachia strain wRi [20], but not in D. simulans flies which were naturally infected with wRi or Ae. albopictus naturally infected with wAlbB [21]. Further some natural Wolbachia infections can depress antibacterial immunity [25]. It should be noted that the D. simulans Me29 line used in the present study is artificially infected with the D. melanogaster derived Wolbachia strain wMel. We suggest that the lack of antibacterial protection in the Me29 line may be a consequence of adaptation that has occurred during the 13 or more years since this line was established [18]. Thus both the data presented here and previously is consistent with Wolbachia-mediated immune priming being important for antibacterial and antiparasite protection in hosts artificially infected with Wolbachia.
This study investigated Wolbachia-mediated antibacterial protection by utilising three Gram-negative bacterial challenge models. It is possible that the protection response upon challenge with a Gram-positive bacterium may differ from that observed in this study. However, we consider this to be unlikely given the lack of protection against Gram-positive bacteria in mosquitoes [7] and lack of Wolbachia-mediated immune stimulation observed in naturally Wolbachia-infected Drosophila.
The immune pathways Toll, IMD and AMPs showed no evidence of antibacterial immune priming by Wolbachia in D. simulans [21] or D. melanogaster, and no protection against bacterial infection was observed. Given that antiviral protection has been demonstrated using the same Drosophila lines as used in the current study [4,9], our results indicate that stimulation of the Toll, IMD or AMPs pathways are not necessary for Wolbachia stimulated antiviral mechanisms. This leaves other immune pathways, such as Vago and vir-1 [28,29], the major insect viral-defence RNA silencing pathway [30] or competition for host resources as potential Wolbachia-mediated antiviral responses. As Wolbachiamediated antiviral protection occurs in both naturally and experimentally infected hosts [3,4,[8][9][10]31] its likely to occur through a conserved mechanism that is independent of that involved in protection against other pathogens.
Wolbachia-mediated antiviral protection in Drosophila has been shown to be robust. Antiviral protection is observed against many different RNA viruses including both natural pathogens of Drosophila and viruses that normally infect mosquito vectors or other insects [4,9,10,32]. In addition, several different Wolbachia strains have been shown to protect flies from viruses in a number of different lines of both D. simulans and D. melanogaster [4,9]. In stark contrast, we show here that Wolbachia does not mediate protection against pathogenic bacteria in Drosophila. We have demonstrated this using three different Gram-negative bacterial pathogens and six different Wolbachia:-Drosophila combinations, including both D. simulans and D. melanogaster hosts. These results have implications both for the potential mechanisms of and uses for Wolbachia-mediated protection.

Flies and Wolbachia
All fly lines were maintained on standard cornmeal diet at 25uC with 12-hours light/dark cycle and were sourced from the lab fly collection. The D. melanogaster line Oregon RC (ORC) is naturally infected with Wolbachia strain wMelCS [33]. The D. simulans lines CO, DSR, N7NO and DSH are naturally infected with Wolbachia strains wAu [34], wRi [35], wNo [36] and wHa [37] respectively. The D. simulans line Me29 was experimentally infected with wMel over a decade ago by transinfection of NAaTC embryos with wMel from D. melanogaster embryos [18]. Wolbachia-free lines were generated from each of the fly lines as previously described [4,9].

Pathogenic bacteria
Three Gram-negative bacteria that are pathogens of Drosophila were used in challenge experiments. S. marcescens and E. carotovora were grown in LB medium [23]. P. aeruginosa PA01 carries an ampicillin resistance gene and was grown in LB medium supplemented with 100 mg/ml ampicillin [38]. To prepare bacteria for challenge bioassays, LB broth was inoculated with bacteria from single colonies on agar plates and incubated for 16 hours at 37uC. Bacteria were then pelleted by centrifugation at 1,5006 g for 10 minutes at room temperature. Fresh bacterial pellets were prepared for each infection bioassay.

Survival bioassays
To analyse the susceptibility of flies with and without Wolbachia to bacterial induced mortality, 4-7 day old adult male Drosophila were challenged with each of the pathogenic bacteria. Flies were anaesthetised with CO 2 prior to infection. A thin needle (diameter = 0.193 mm) was dipped into the undiluted bacterial pellet and used to prick into the thoracic region of each fly. For each fly line, two groups of flies were challenged with pathogenic bacteria: flies with Wolbachia and flies without Wolbachia. For each group of flies three vials of ten flies were challenged with one of the pathogenic bacteria and one vial of ten flies was mock infected with LB medium. Following challenge flies were maintained in a 25uC incubator and survival of the flies was monitored every 2-6 hours. Mortality within the first 6 hours was deemed to be due to needle injury. At least three independent survival bioassays were done for each bacteria/fly line combination. Survival curves of the two groups of flies were compared using Kaplan-Meier analysis and log-rank statistics reported (GraphPad Prism).

In vivo bacterial growth
To analyse the impact of Wolbachia on the accumulation of pathogenic bacteria in flies, 4-7 day old adult male ORC flies with and without Wolbachia were infected with P. aeruginosa PA01 as described above. At 0 and 12 hpi, three live flies were collected individually into 1.5 ml tubes. After addition of 200 ml of LB medium supplemented with 100 mg/ml ampicillin and two 3 mm glass beads (Sigma-Aldrich) to each individual, flies were homogenised for 90 s using a TissueLyser II (Qiagen). Fly homogenates were serially diluted and spread on LB agar plates containing 100 mg/ml ampicillin. Colony forming units per fly (CFU/fly) were calculated after overnight incubation of plates. The experiment was replicated on two independent cohorts of flies and the data pooled.

Analysis of immune gene regulation
RT-qPCR was used to compare the expression of six immune genes in ORC flies with and without Wolbachia. Genes were chosen on the basis of homology to genes that were upregulated in Ae.
aegypti in the presence of Wolbachia [7]. Primer sets for the target genes (Table S1) were designed with Primer3 software [39]. The primer efficiency across a six 5-fold cDNA dilution series was confirmed [40].
For comparison of gene regulation five individual 4-7 day old male flies from the ORC flies with and without Wolbachia were frozen and homogenised in Ribozol (Amesco) with two 3 mm glass beads using TissueLyser II (Qiagen) for 90 seconds with the frequency of 30 shakes/s. Total RNA was extracted and treated with DNase (Promega) for 30 minutes at 37uC to eliminate DNA contamination. 1 mg of total RNA was reverse transcribed using random primers (Promega) and SuperScript III reverse transcriptase (Invitrogen) as per the manufacturer's protocol. Quantitative PCR was performed in duplicate reactions using Platinum SYBRH green qPCR supermix as per manufacturers instructions (Invitrogen). The temperature profile for the qPCR was 95uC for 2 min, 50uC for 2 min and 40 cycles of 95uC for 10 s, 60uC for 10 s and 72uC for 20 s. qPCR was performed using a Rotor-Gene 6000 (Qiagen). Expression of the target genes was normalised using reference genes Actin 79b (GenBank accession no. NM_079486) and ribosomal protein L 32 (RpL 32) (Genbank accession no. NM_001144656.2) [41]. Target to reference gene ratios were obtained suing QGene4.2 [42] and treatment effects on the expression ratios were assessed using Mann Whitney-U tests in STATISTICA V8 (StatSoft). Table S1 RT-qPCR primers. (DOCX)