Abstract

Symbiotic microorganisms have had a large impact on eukaryotic evolution, with effects ranging from parasitic to mutualistic. Mitochondria and chloroplasts are prime examples of symbiotic microorganisms that have become obligate for their hosts, allowing for a dramatic extension of suitable habitats for life. Out of the extraordinary diversity of bacterial endosymbionts in insects, most are facultative for their hosts, such as the ubiquitousWolbachia, which manipulates host reproduction. Some endosymbionts, however, have become obligatory for host reproduction and/or survival. In the parasitoid waspAsobara tabida the presence ofWolbachia is necessary for host oogenesis, but the mechanism involved is yet unknown. We show thatWolbachia influences programmed cell death processes (a host regulatory feature typically targeted by pathogens) inA. tabida, making its presence essential for the wasps' oocytes to mature. This suggests that parasite strategies, such as bacterial regulation of host apoptosis, can drive the evolution of host dependence, allowing for a swift transition from parasitism to mutualism.

Results and Discussion

We observed pyknotic nuclei (Fig. 1A andB), one of the hallmarks of apoptosis, in the ovarioles ofA. tabida females whereWolbachia was antibiotically removed (so-called aposymbiotic females). Furthermore, both acridine orange staining (Fig. 1C andD) and the apoptosis-specific TUNEL assay (Fig. 1E–H) show that absence ofWolbachia bacteria in aposymbiotic females results in copious apoptosis of the nurse cells of mid-stage egg chambers, while at this stage apoptosis does not occur in the ovarioles of symbiotic control females (Fig. 2) (Mann–WhitneyU test:W = 464;n = 45;P ≪ 0.001). Apoptosis is an essential component of insect oogenesis and occurs at several stages during oogenesis in response to both developmental and environmental stimuli (11). InDrosophila there are checkpoints during early and mid-oogenesis at which various environmental stress factors can trigger cell death in egg chambers (11). Our results show that, inA. tabida, the presence ofWolbachia is essential to progress egg chambers past the mid-oogenesis checkpoint by preventing apoptosis of nurse cells and, as such, allows for oocyte maturation. Whether this modification of programmed cell death is controlled byWolbachia or is host-derived requires further investigation. Nevertheless,Wolbachia clearly interacts with apoptotic pathways during oogenesis.

Immune responses against intracellular bacteria frequently involve apoptosis of infected cells (12), and, in turn, intracellular bacteria can inhibit host cell apoptosis to circumvent the immune system (13). Several other well known intracellular pathogens are able to control apoptotic pathways in their host. WhereasSalmonella sp. and the causative agent of Legionnaire's disease,Legionella pneumophila, induce apoptosis in host cells, the sexually transmitted human disease agentChlamydia and the causative agent of Rocky Mountain spotted fever,Rickettsia ricketsii (a close relative ofWolbachia), are both able to either block or induce apoptosis of host cells depending on the stage of the infection (14). The necessary alterations to the cell physiology ofA. tabida most likely occurred after initial invasion of theWolbachia, because (i) they would have been destructive forA. tabida withoutWolbachia's presence and (ii) earlier comparisons within the genusAsobara show thatWolbachia-dependent oogenesis is a derived character (15).

Bacterial inhibition of apoptosis can result in a severe reduction in functionality of the infected host tissues. This is particularly relevant in the ovaries, where the highest densities ofWolbachia bacteria are found (16) and where apoptosis is an essential component of oogenesis (11). Selection is expected to favor a swift recovery of reduced ovarian functionality by coevolution between the two partners. Theoretically, at least three scenarios can lead to restoration of ovarian functionality: (i)Wolbachia ceases to manipulate apoptosis and evades the host immune system by hiding within membrane-bound compartments inside insect cells (13). (ii)Wolbachia permanently takes over control of the host apoptotic pathway. Permanent control becomes possible when in infected hosts the nuclear genes controlling apoptosis are no longer under selection, thus allowing for the accumulation of mutations in these genes. An example of this is found inDrosophila melanogaster, whereWolbachia restores several host mutations, among which is the mutantSxl^f4, which closely resembles the phenotype observed inA. tabida (17,18). (iii) Host gene expression modifies to take into account the bacterial manipulation of the apoptotic pathway. The first scenario does not directly result in the observed complete dependence of the host onWolbachia. Both latter scenarios, however, although indistinguishable inA. tabida without further genetic information, undoubtedly lead to complete host dependence. As such, regulation of host programmed cell death is a likely mechanism in this system to explain the evolutionary shift from facultative parasitism toward obligate mutualism.

Future work could address the generality of apoptosis manipulation in host–Wolbachia associations and the molecular mechanisms underlying these manipulations. Recent findings show that parasites and mutualists often share common strategies and molecular tools, e.g., secretion systems, that have long been exclusively described as virulence factors (13,19). Because the host apoptotic pathway is targeted both by parasitic and mutualistic bacteria (12,20), apoptosis can be a general mechanism of molecular and cellular interaction in host–bacteria associations; initially to evade the host immune system and later to manipulate host development. InA. tabida, however, manipulation of the apoptotic pathway appears to be at the origin of the host dependence and the evolution of obligate mutualism. It has been proposed that endosymbionts have been slaved by their hosts (21); however, our results show that endosymbionts can also hijack their host by becoming obligatory for reproduction without providing any further advantage.

Materials and Methods

Acknowledgments

We thank P. Mavingui for stimulating discussions and M. E. Huigens, R. Stouthamer, P. M. Brakefield, S. A. West, and three anonymous referees for their critical comments. We thank the Centre National de la Recherche Scientifique for funding.

Footnotes

References