Introduction

Arthropods harbor a variety of microorganisms, and Wolbachia are perhaps among the most commonly occurring facultative bacterial endosymbionts [1]. This group of vertically transmitted Gram-negative bacteria attracted more and more attention for their capability of manipulating host reproduction by causing cytoplasmic incompatibility (CI), feminization, male killing, and parthenogenesis induction [2], which is of great significance in pest control and human disease transmission [3].

The fruit fly Drosophila melanogaster uses multiple innate defense strategies to combat bacterial infection, many of which are also used by higher organisms including human beings [4]. These defense strategies include physical barriers and immunity: local immune response in the barrier epithelia by producing antimicrobial peptides (AMPs) and reactive oxygen species, cellular immunity via phagocytosis and encapsulation, and humoral immunity by synthesizing AMPs in the fat body. The signaling pathways regulating the production of AMPs were identified using the Drosophila model [5]. Seven groups of AMPs were characterized, among which Diptericin, Drosocin, and Attacin are highly effective against Gram-negative bacteria [6, 7]. On the other hand, the cellular encapsulation is a dramatic defense response mediated by lamellocytes in Drosophila [8]. Edin (elevated during infection) acted as an important determinant of the encapsulation response in D. melanogaster larvae [9]. In the past decades, remarkable progresses have been made in insect immunity, although the mechanisms underlying the insect-Wolbachia interactions are only partially understood [10]. Wolbachia are very common in Drosophila but they cannot be cultured outside of host cells [11]. Recognition of bacteria by Drosophila is achieved through the sensing of specific forms of peptidoglycan by peptidoglycan recognition proteins (PGRPs). The discovery of PGRP-LE as an intracellular sensor of Gram-negative bacteria may be among the important advances in understanding the immune defense of insects to Wolbachia [12]. It was reported that the PG-associated lipoprotein (PAL) was located on the cell membrane of Wolbachia [13]. PAL was known to specifically bind diaminopimelic acid (DAP) [14]. Therefore, Wolbachia can be recognized by PGRPs which then trigger the Imd pathway and subsequent AMP generation [15]. Nevertheless, up to now, the molecular mechanism of insect-Wolbachia and how the titer is controlled in vivo is poorly understood, particularly when multiple infections occur.

The success of Wolbachia is attributed to efficient maternal transmission and manipulations of host reproduction commonly through CI [16]. CI is affected by both host and Wolbachia [17, 18]. For instance, CI factor A (CifA) encoded by syntenic loci within Wolbachia’s WO prophage region played a key role in the rescue of CI [19], which was further supported by a recent study using two conspecific Wolbachia strains from Drosophila pandora [20]. Moreover, the strength of CI was correlated with the density or titre of Wolbachia [21], which appeared to be influenced by both host- and Wolbachia-intrinsic factors [22, 23]. It can be expected that the titre of Wolbachia should reflect a balanced interaction between host defense (immunity, resistance and tolerance) and Wolbachia anti-defense. Previous studies showed that the native Wolbachia strain did not elicit an AMP-based immune response in the host, while a strong induction of AMP gene expression was observed when Wolbachia were introduced into novel hosts [24,25,26]. Nevertheless, the mechanisms underlying the complex interactions between host insects and co-existing Wolbachia strains are still unclear.

Great advances have been made in Wolbachia genomics. The whole-genome sequence of Wolbachia pipientis wMel strain from D. melanogaster provides an ideal system for studying the Wolbachia–Drosophila interactions [27]. The wMel strain is a typical CI-inducing Wolbachia strain, belonging to Supergroup A based on gene sequencing and MLST typing [28]. It was successfully transferred into Aedes aegypti mosquitoes and blocked transmission of dengue [29, 30]. What’s more, in an experimental transfection by microinjection, the wMel strain established itself in a phylogenetically distant host insect Bemisia tabaci [31]. It is therefore intriguing to explore whether a Wolbachia strain derived from a phylogenetically distant host insect can also establish itself in D. melanogaster. Here we used a previously characterized Wolbachia wCcep strain from the rice moth Corcyra cephalonica [31] to establish a Drosophila/wCcep/wMel system. Our purpose was to investigate the multiple interactions between the host and different Wolbachia strains and analyze the factors influencing the establishment of a Wolbachia strain in a novel insect host. We found that the wCcep strain elicited a significant host immune response from the novel host, supporting the notion that the exogenous bacteria may trigger a robust innate immune response that eliminates the intruders [32]. Furthermore, based on Wolbachia titre measurement using RT-qPCR, the intrusion of wCcep elicited an exclusion reaction from the native wMel strain, inconsistent with the theoretical prediction that multiple infections favor cooperation between co-existing Wolbachia strains [33]. In the present study, we firstly transferred a Wolbachia strain derived from a distantly related host into D. melanogaster, which provides new insights into the multiple associations between the host and co-existing Wolbachia strains.

Materials and Methods

Insect Rearing and Wolbachia Isolation

The rice moth C. cephalonica was maintained on Maize-Rice bran–Sugar medium (25 °C, 65% RH and 14L:10D). The fruit fly D. melanogaster was maintained on Maize-Agarose-Yeast medium (25 °C, 60–70% RH and 14L:10D. The wCcep strain was isolated from two moths using the Percoll density-gradient centrifugation method [31]. The purified bacteria were detected using the primers 81F/522R targeting wsp of Group B Wolbachia [34].

Microinjection

A volume of 46 nl bacterial suspension in SPG buffer (220 mM sucrose, 4 mM KH2PO4, 9 mM Na2HPO4, 5 mM l-glutamate, pH 7.4) was injected into the pupa of D. melanogaster using a glass needle on the platform of Nanoliter 2000 (World Precision Instruments, Sarasota, FL, USA). Approximately 100 pupae were injected, which were then placed in a climate incubator until eclosion (25 °C, 60–70% RH and 14L:10D). The newly emerged adults (G0) were separately maintained in pairs (♀/♂) for establishing isofemale lines.

Quantitative Analysis of Wolbachia Titre

The relative titres of wMel and wCcep were measured using real-time quantitative polymerase chain reaction (qPCR) in D. melanogaster over 8 generations after microinjection. The primers 81F/522R specifically targeting wCcep (B-Wolbachia) [34] and the primers wspQ384/wspQ513 targeting both wMel and wCcep [28] were used in qPCR analysis, with GAPDH as the internal reference (Table S1). The stability of primers was judged by the cycle threshold (Ct). Three adult flies were extracted for one DNA sample (50 ng/µl). The reaction was performed in a total volume of 20 µl containing 10 µl AceQ® qPCR SYBR® Green Master Mix (Vazyme, Nanjing, China), 0.4 µl of each primer (10 µM), 1 µl gDNA (50 ng) and 8.2 µl ddH2O. The thermocycling program was 50 °C for 2 min, 95 °C for 5 min, 40 cycles of 95 °C for 10 s, 60 °C for 30 s. The relative titre was calculated using the 2−ΔΔCt method [35]. All samples were assayed in triplicate on an ABI 7500 (Applied Biosystems, Carlsbad, CA, USA).

Crossing Experiments

The native strain was removed by tetracycline (0.25 mg/ml) for two consecutive generations. The uninfected flies were then injected with wCcep solution (46 nl) and female isolines were constructed. Infected (IN) and uninfected (UN) flies from the 4th generation (G4) were used for reciprocal crossing: UN♀ × UN♂; UN♀ × IN♂; IN♀ × UN♂, and IN♀ × IN♂. The newly emerged adults were used for mating in a tube (Φ2.2 cm) for 48 h, and the inseminated females were then placed individually in a petri dish (Φ3.5 cm). The number of eggs per female, hatching rate and developmental durations were calculated, and the level of CI was assessed according to the hatching rate of eggs.

Transcriptome Sequencing

The sequencing libraries were constructed from the pupae of fruit fly. Total RNA was extracted from approximately 30 pupae for each treatment: 24 h or 48 h post-injection (hpi) with wCcep (46 nl) or the same volume of SPG buffer (negative control), with two repetitions. The cDNA libraries were established by Illumina Truseq RNA Sample Preparation Kit (NEB, San Diego, USA) with 2 μg RNA for each sample. Then, the Illumina MiSeq platform was used to produce 300-bp paired-end sequences. After the high-quality clean data were achieved, the genome sequences of D. melanogaster downloaded from NCBI (https://www.ncbi.nlm.nih.gov/genome/47) were used as the reference for identifying unigenes using Bowtie v2.0.6, TopHat v2.0.9 and HTSeq v0.5.4p3; the DESeq R package was used to characterize the differentially expressed genes (DEGs) (the corrected P value < 0.005; the log2 (fold change) > 1).

Functional Annotation of DEGs

The Gene Ontology (GO) enrichment analysis of DEGs [36] was conducted on the GO seq R package; the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis [37] was carried out using the KEGG orthology‑based annotation system (KOBAS). The significantly enriched GO and KEGG pathway terms were determined using the hypergeometric test (P < 0.05). The DEGs related to host immunity, detoxification and stress responses against Wolbachia invasion were classified [4].

Association Analysis Between Gene Expression and Wolbachia Titre

Six treatments were carried out: 24 hpi with SPG buffer, 48 hpi with SPG buffer, 24 hpi with wCcep, 48 hpi with wCcep, 24 hpi with twice wCcep (48 h apart), and 48 hpi with twice wCcep (48 h apart). The expression of attacin-D and edin and the titre of wCcep were measured using real-time qPCR, but the DNA templates were different: cDNA for the former and gDNA for the latter. Total RNA and gDNA were successively isolated from approximately 30 whole fruit flies using Trizol (TransGen Biotech, Beijing, China) [38]. The cDNA was synthesized using 0.5 μg total RNA and reverse transcriptase (HiScript® II One-Step RT-PCR Kit, Vazyme Biotech, Beijing, China) according to the supplier’s instructions. The primers used (Table S1), the reaction system and thermocycling program for qPCR analysis were the same as described above. A linear regression analysis was performed to identify the association between gene expression and Wolbachia titre.

Data Analysis

The statistical differences were analyzed using One-way AVOVA followed by Student Newman Keuls (SNK) test at 0.05 and 0.01 levels on SPSS v.20.0 (SPSS Inc., Chicago, IL, USA). Linear regression analysis was performed on Microsoft Excel v.1903.

Results

Quantitative Analysis of Wolbachia Titre

The results showed that the reference GAPDH was quite stable over different generations (Fig. S1). Quantitative analysis indicated that wCcep could be transmitted over four generations but it was undetectable after G4 (Fig. 1). Specifically, the titre of wCcep climbed during the early stage after microinjection (24hpi-G0), but then declined rapidly, even undetectable at G3; surprisingly, it showed a sudden rebound at G4, but then returned to an undetectable level (G5 and later). In comparison, wMel dropped immediately (24–48 hpi) and remained at a low level till G4, but then began to rise (G5 and later).

Fig. 1
figure 1

The dynamics of the titres of wMel and wCcep strains in D. melanogaster during different stages after transinfection. Data are represented as means ± SE of three repetitions. WT wildtype, hpi hours post-injection

Full size image

Crossing and CI

The crossing experiments showed that there was no significant difference in the developmental durations among different crossing types (SNK, P = 0.731) (Table S2); no significant difference was observed in the number of eggs laid per female (SNK, P = 0.662). However, a highly significant difference existed in the hatching rate between UN♀ × IN♂ and the other crossing types (SNK, P < 0.001) (Table 1). The significantly lower hatching rate in UN♀ × IN♂ indicated a strong CI induced by wCcep.

Table 1 Fecundity and hatchability in different crossings between wCcep-infected (IN) and antibiotic-treated uninfected (UN) fruit flies
Full size table

Host Responses to Wolbachia Intrusion

Transcriptome sequencing identified 240 DEGs (173 upregulated; 67 downregulated) at 24 hpi; 295 DEGs (183 upregulated; 112 downregulated) at 48 hpi, and 497 DEGs (254 upregulated; 243 downregulated) when comparing 24 hpi with 48 hpi (Fig. S2). KEGG analysis of DEGs identified a variety of induced biological pathways. Interestingly, more pathways were activated at 48 hpi than at 24 hpi (Fig. S3). Functional annotations revealed that wCcep intrusion elicited typical immune reactions, including the Toll and JAK/STAT signaling pathways (Table S3), humoral and cellular immunity (Table S4). The majority of antimicrobial peptides (AMPs) were downregulated, whereas the lysozymes were upregulated. In addition, host detoxification and stress responses were also regulated (Table S5). The raw sequence data are available upon request.

Association Between Expression of Attacin-D/Edin and Wolbachia Titre

The results showed that the gene expression and Wolbachia titre varied considerably among different treatments (Table 2; Fig. S4). A general trend was that the injection of wCcep downregulated gene expression; surprisingly, twice injection of wCcep drastically reduced the expression of attacin-D and edin. Linear regression indicated that a significant negative correlation existed between the titre of wCcep and the expression of attacin-D (R2 = 0.8157; P = 0.00082) and edin (R2 = 0.8825; P = 0.00034) (Fig. 2).

Table 2 The relative expression levels of attacin-D and edin and the relative titres of wCcep strain in D. melanogaster under different treatments
Full size table
Fig. 2
figure 2

Linear regression analysis between the relative titre of wCcep strain and the relative expression level of attacin-D (a) and edin (b) in D. melanogaster

Full size image

Discussion

Our studies suggested that the exogenous wCcep strain from the moth could infect the fruit fly D. melanogaster and induce a high level of CI, but could only be transmitted for four generations (G0 to G4) in the novel host. Moreover, the intruding wCcep strain should have suffered a competitive exclusion from the native wMel strain, as their tires were negatively correlated during the invasion process. Furthermore, the recipient host imposed a remarkable immune suppression against the wCcep strain. All of these reactions caused the failure of the establishment of wCcep in D. melanogaster.

Although Wolbachia are common in Drosophila, the mechanisms underlying the host-Wolbachia interactions are only partially understood due to its unculturability. Microinjection is an ideal method for deciphering the interactions between the host and Wolbachia [39]. The wCcep strain native to the rice moth C. cephalonica had previously been shown to establish itself in the hemipteran pest insect B. tabaci through microinjection [31]. D. melanogaster is known to harbor the wMel strain [27], and thus it is expected that the invading wCcep strain should actively interact with the novel host and native wMel strain [40, 41]. The interactions may be viewed from the change in the titre of Wolbachia and the expression levels of immune genes. Our quantitative analysis of Wolbachia titre showed that wCcep was negatively correlated with wMel in their titres, indicating that there might exist a competitive relationship between the two co-existing strains. Indeed, the existence of a competition between the novel and native strains can partially explain why wCcep could only be transmitted in the new host for a relatively short period of time (four generations). Several previous studies investigating the interactions between co-existing Wolbachia strains by comparing their titres (or densities) achieved mixed results: Wolbachia titre was highly strain-specific and unaffected by the presence of other strains in some parasitoid wasps and moths [42], whereas competition obviously existed between co-occurring strains in the beetle Callosobruchus chinensis and Acromyrmex leafcutter ants [43,44,45]. These results suggest that the interactions between the invading strain, host insect and native strain may be influenced by a complex of factors that need to be identified.

To investigate the effects of wCcep on host developmental duration, fecundity (fitness effects) and CI level, crossing experiments were conducted using flies treated with antibiotics to obviate the effect of the native strain and then injected with or without wCcep. The results suggested that wCcep could induce a strong unidirectional CI in D. melanogaster, confirming the infection capability of wCcep. Crossings also indicated that wCcep infection imposed no significant fitness costs on the host as no obvious changes were observed in the developmental durations and the number of eggs laid per female. This is consistent with our previous results achieved in B. tabaci, where transinfection of wCcep had no significant effect on the fecundity of the whitefly [31].

Transcriptome sequencing via RNA-seq coupled with functional annotations identified a host of genes involved in insect-Wolbachia interactions, including humoral and cellular immune responses, detoxification and stress resistance. One interesting finding is that sampling at 48 hpi identified more DEGs, while no substantial change was detected at 24 hpi. Another finding is that many immune-related DEGs (including the majority of AMPs such as attacin-D) were downregulated in response to wCcep infection. One possible explanation is that the host has shut down these genes to provide protection for the native Wolbachia strain due to unknown fitness-related benefits. Nevertheless, massive doses of exogenous Wolbachia might be a possible factor causing the apparent suppression of many immune-related genes as observed in Aedes albopictus, D. melanogaster, D. simulans and Tetranychus urticae [46,47,48]. It seems that the immune- and stress-related genes played a subtle role in regulating the host insect-Wolbachia associations. This is further supported by our association analysis between the expression of attacin-D/edin and the titre of wCcep, in which the expression of attacin-D/edin was significantly negatively correlated with the titre of wCcep. This finding revealed that attacin-D/edin are two determinants of wCcep titre. Considering Attacin-D and Edin are key components of insect innate immunity, our results suggest that Attacin-D and Edin play important roles in the host defense against the invading Wolbachia strain.

From an evolutionary perspective, coevolution is expected to favor low fitness cost, low level of CI, and high transmission rate. Conversely, the intrusion of an exogenous bacterial strain (e.g., injection of Wolbachia) into a novel host is expected to lead to negative fitness effect, high CI level, and low transmission rate [49]. In the present study, we did not measure the transmission rate, but the neutral fitness effect and high CI level measured for wCcep are partially in agreement with the theoretical prediction. Thus, for future research, the Drosophila/wMel/wCcep system is expected to be useful for investigating the coevolution between Drosophila and wMel, the competitive interaction between wMel and wCcep and the functional genes involved in the defense and anti-defense interactions between Drosophila and wCcep.

In conclusion, the wCcep strain can induce a high level of CI in the phylogenetically distant host D. melanogaster after infection, but can only be transmitted in the novel host for several generations. The invading Wolbachia strain imposed no significant fitness costs on the novel host, but suffered a robust immune response from the host and incurred a competitive exclusion from the native Wolbachia strain, which resulted in the failure of its establishment in D. melanogaster. Our data indicate that D. melanogaster and wMel might have established a symbiotic relationship after a long-term coevolution.