The pathology of embryo death caused by the male-killing Spiroplasma bacterium in Drosophila nebulosa

Background Inherited bacteria that kill male offspring, male-killers, are known to be common in insects, but little is understood about the mechanisms used by male-killing bacteria to kill males. In this paper we describe the tempo and changes that occur during male-killing by Spiroplasma bacteria in the host Drosophila nebulosa. Results Spiroplasma infected D. nebulosa males were developmentally retarded from 6–8 h into embryonic development at 25°C, and arrested at between stages 12 and 13 of embryogenesis (10–12 h). Dying males were characterized by a failure to form segments, and ultimately disintegration of the normal oval embryonic shape. Prior to death, dying males exhibited widespread apoptosis, as testified by TUNEL staining. Conclusion The Spiroplasma kills male Drosophila in a narrow developmental period, shortly after the formation of the host dosage compensation complex that is required for male-killing. Male death is preceded by widespread apoptosis, but it is uncertain if this is primary or secondary apoptosis.


Background
Male-killing bacteria, bacteria that are passed vertically from a female to both her male and female progeny and kill the males during embryogenesis, are known to occur in a wide variety of insect taxa [1]. Much research has been conducted on their frequency in natural populations, the identity of the bacteria concerned, their potential impact. This has revealed that they vary in frequency within and between host species, with prevalence varying from 1 to 99% of female hosts [2,3]. The bacteria are diverse, deriving from the genera Spiroplasma, Wolbachia, Rickettsia, Arsenophonus, and an unnamed representative in the Cytophaga-Flavobacteria group [4][5][6][7][8]. The impact on the host can be profound, with high male-killer prevalence associated with sex role reversal in the butterfly Acraea encedon [9], and the evolution of suppressor genes in Hypolimnas bolina [10].
In contrast, almost nothing is known about the timing or mechanism of male-killing with the exception of a few studies in the Spiroplasma-Drosophila system. Recent studies of this interaction have demonstrated that male-killing requires a functional host dosage compensation complex [11]. However, there is little precise knowledge regarding the timing of male death, and no information on the cellular pathology induced by the Spiroplasma. Past studies have described abnormalities occurring anywhere between cleavage divisions and 8 h after oviposition, with death occurring even in larval or pupal phase [12][13][14]. However, these previous studies are less than satisfactory because researchers were unable to identify the sex of embryos, so unhealthy embryos were automatically considered male. In addition, the focus of three of the four studies was an unnatural association: D. melanogaster transinfected with Spiroplasma from D. nebulosa. Whilst this transinfection does function with a good degree of penetrance, it is somewhat unstable in the laboratory, and these studies did not assess whether the infections functioned in the same way as a natural infection.
In this study, we examine the process of male death in a natural association, that between Drosophila nebulosa and its male-killing Spiroplasma. This strain of Spiroplasma is closely related to S. poulsonii, the male-killer discovered in D. willistoni [15]. This strain shows very high rates of natural vertical transmission, making the infection very stable and amenable to laboratory study. An antibody against SXL [16], the key switch peptide in Drosophila sex determination and present only in females, is used to allow the embryos to be sexed precisely. We first conducted a detailed study of the timing of male death caused by S. poulsonii, and the abnormalities observed. We also investigated the role of apoptosis in cellular death within male embryos at the point of male death. Previous work on bacteria-host interactions has demonstrated that bacteria are able to induce and repress apoptosis, both of which could result in the death of an embryo [17,17]. We tested whether apoptosis of host cells occurred during male death in the Drosophila-Spiroplasma system.

Timing and morphology of male death
We first examined whetherthe SXL antibody raised from D. melanogaster SXL could be used to sex D. nebulosa embryos as it does in D. melanogaster. A comparison of the ratio of SXL-positive to SXL-negative D. nebulosa embryos in uninfected lines shows a 1:1 ratio of SXL+:SXLembryos, as expected for a protein only present in females (see Table 1). Thus we conclude the antibody does allow differentiation of male from female D. nebulosa. There was no difference in the ratio of SXL-positive to SXL-negative D. nebulosa embryos between infected and uninfected lines (χ 2 = 0.0538, NS) indicating that this SXL expression is not altered by S. poulsonii, and can be used in infected as well as uninfected embryos.
We then examined the timing of male embryo death by tracking the developmental stage attained by 2 h cohorts of eggs, infected and uninfected, allowed to develop between 4 and 24 h. Embryos in each cohort were placed into developmental stages delineated by Campos-Ortega and Hartenstein [18]. Infected males behind infected females in terms of development 6-8 h after egg laying (AEL), and by 10-12 h AEL there is a clear arrest of male development ( Figure 1; data summarised in Table 2). Infected males arrested at or before stage 13 of development, whereas infected females continue through embryogenesis. Uninfected male and female embryos developed broadly in parallel, and at a rate equivalent to that observed for infected female embryos ( Table 3).
The morphology of Spiroplasma-killed male embryos was very characteristic. Figure 2 shows a well-developed infected female embryo (stage 16, 16-18 h AEL) and an infected male of the same age. The only normal structure that is recognisable in the Spiroplasma-killed male is the cephalic furrow. There is an abnormal protruding 'sac' from the ventrum of the male embryo and the structure of the infected male becomes grainier and less regular from the point at which abnormalities are first noticeable. By this age, all nuclei in the embryo have broken down, and the embryo gives little hint of any internal structure under DAPI stain ( Figure 2D). Male embryos arrest before segmentation (which is never observed), and around the point of germ band retraction.

Are cells in male embryos dying by apoptosis?
We examined the pattern of apoptosis in Spiroplasmainfected and uninfected embryos using the TUNEL procedure [19]. Infected D. nebulosa female embryos and uninfected embryos of both sexes were found to have apoptotic cells in the same pattern found in D. melanogaster [20], with TUNEL positive cells around the cephalic furrow, and sporadic TUNEL positive cells in the posterior of the embryo, from stage 10 onwards (about 7 h AEL). In contrast, infected male embryos of the same age show extensive apoptosis throughout the embryo, with vast numbers of TUNEL positive cells ( Figure 3). The pattern observed in Figure 2 was seen in all male embryos derived from infected females (n > 200). It is notable that these infected male embryos become TUNEL positive before development completely arrests, as indicated by the widespread apoptosis in the stage 10/11 male embryo in Figure 3.

Discussion and Conclusion
This paper investigated the development of male D. nebulosa embryos infected with male-killing Spiroplasma. It was observed that the infected male embryos become retarded The developmental stage of male and female D. nebulosa embryos, infected with S. poulsonii, aged 4-20 h AEL, in 2-h cohorts What was notable from this study is that the events involved with Spiroplasma induced male-killing proceed very rapidly. Over a 2-h period, relatively early in embryogenesis in absolute terms, the male embryos go from appearing phenotypically normal to showing widespread abnormality. The tight window over which this degradation occurs contrasts with previous studies, and probably reflects that male and female embryos could be discriminated precisely, allowing random deaths to be more precisely partitioned and a clear pattern observed. It may also reflect our study of a natural association where male-killing is being seen as a strong phenotype.
This study also examined the type of cell death occurring in dying males. It was observed that male-killed embryos were characterized by massive apoptotic death of cells within the embryo that preceded the complete arrest of development. Thus, we can conclude that the processes leading to male death are active at stage 10. It is notable also that male death occurs at the point at which natural apoptosis begins [20]. It is therefore tempting to hypothesize that male-killing operated through some subversion of the apoptosis machinery, inducing its uncoordinated expression of the apoptosis pathway across the embryo, that ultimately resulted in the arrest of development. However, it should be noted that the association of apoptosis with male death is not necessarily causal. There are many cases where apoptosis is induced by the host in response to infection rather than as a manipulation by the infection. In these cases of secondary apoptosis, death  would have occurred with or without the apoptotic processes and cell death is merely a response to the pathology induced by the infection.
What remains totally unknown are the molecules secreted by the Spiroplasma that create the changes observed. At the most elementary level, we do not know whether there is a previously secreted factor that becomes active only at 6-10 h into male embryo development, or whether a factor is secreted ab initio in males only, from 6-10 h into development. Future work will need to focus on the nature of this factor and its interaction with host cellular and developmental systems.

Methods
The male-killing infections used were from two D. nebulosa females (G37, G87) collected in Guadeloupe, during July 2001, previously described in Bentley et al. [15]. This infection was maintained in outbred D. nebulosa through mating of the females in each generation to males from an outbred population of D. nebulosa collected at the same time, and maintained at large population size. This population was additionally the source of the uninfected control population.

Timing and morphology of male death
Large groups (>100 individuals) of Spiroplasma infected line G37, Spiroplasma infected line G87, and uninfected D. nebulosa aged between 2-14 days old, were placed onto separate grape juice agar laying plates. Following a prelay to avoid oviposition of retained embryos, the flies were transferred onto new plates for 2 h. In order to obtain a time series of embryos, the eggs from the 2-h collections were then aged for the required time (between 4-22 h) at 25°C. This was conducted for each of the outbred stocks of the S. poulsonii infected lines and the uninfected outbred stock. The flies were aged to produce near-complete vertical transmission efficiency, which is observed in this stock under these conditions.
Following development, embryos were collected and fixed by standard methods as described in Sullivan et al. [21]. In brief, the eggs were dechorionated by shaking in 3-5% bleach followed by washing in PBS. They were then  [16]. Anti-MLE is used as a positive control for antibody staining, visible in both males and females [22] where the fixing and staining procedure has occurred corresecondary ctly. Unbound antibody was removed by washing in PBT, and 1 μl of FITC-conjugated rabbit anti-mouse IgG secondary antibody, and TRITCconjugated goat anti-rabbit IgG (both Molecular probes) added, followed by washing. Unbound secondary antibody was removed by washing in PBT. Embryos of a given age and infection status were mounted together on a slide in Vectashield containing DAPI, and visualised on a fluorescence microscope.
On slides where embryos were positive for MLE binding, embryos were scored for sex using SXL fluorescence (FITC positive = female), and developmental stage using DAPI staining to reveal structure. Developmental stage was categorised as per Campos-Ortega and Hartenstein's description of D. melanogaster development that segregates embryogenesis into 17 stages [18], from newly fertilized zygote through to formed larva. Note that not all embryos were scored: inequalities sex ratio in the observations are associated with differential scoring of embryos by sex in certain cases, rather than biological differences.

Are cells in male embryos dying via apoptosis?
The TUNEL procedure was used to label the 180 bp DNA fragments found in apoptotic cells and make them visible under a fluorescence microscope. In D. melanogaster, apoptosis is seen from stage 11 onwards (7 h AEL), in a dynamic yet characteristic pattern within the embryo [20]. With this is mind D. nebulosa embryos were tested in cohorts 3-5 h, 5-7 h, and 7-9 h AEL for the presence of apoptosis in either Spiroplasma infected or uninfected embryos. Embryos were also stained with anti-SXL as previously to identify their sex. Embryos were collected and fixed as above and the In Situ Cell Death Detection Kit, TMR red (Roche) was used to detect apoptosis. The protocol was adapted slightly from the manufacturer's instructions for detecting apoptosis in cell suspension material. The embryos were washed twice in PBS and then the PBS was removed and 50 μl of the TUNEL reaction mixture was added. The embryos in TUNEL reaction mixture were left for 1 h at 37°C in the dark before being washed three times for five minutes in PBS. Embryos were then stained for anti-SXL and mounted and examined as described above.
During this process, a negative control was also completed as per the manufacturer's instructions. No positive control was necessary, as from 7 h onwards D. melanogaster embryos naturally have some apoptosis visible [20], and this paper has already demonstrated that D. nebulosa embryos should also have apoptotic cells naturally occurring in that timeframe (as they will have also reached stage 11). When TUNEL staining was being performed on embryos younger than 7-9 h AEL an older embryo sample was subjected to the test at the same time to act as a positive control.

Authors' contributions
JB and GH conceived the project. JB and JH conducted the observations of male death pattern, with advice from ZV and GH. JB conducted the TUNEL assays with advice from ZV and GH. JB, ZV, JH and GH wrote the paper.