Cohabitation and roommate bias of symbiotic bacteria in insect hosts

Symbiotic interactions between insects and bacteria have long fascinated ecologists. Aphids have emerged as the model system on which to study the effect of endosymbiotic bacteria on their hosts. Aphid‐symbiont interactions are ecologically interesting as aphids host multiple secondary symbionts that can provide broad benefits, such as protection against heat stress or specialist natural enemies (parasitic wasps and entomopathogenic fungi). There are nine common aphid secondary symbionts and individual aphids host on average 1–2 symbionts. A cost‐benefit trade‐off for hosting symbionts is thought to explain why not all aphids host every possible symbiont in a population. Both positive and negative associations between various symbionts occur, and this could happen due to increased costs when cohosting certain combinations or as a consequence of competitive interactions between the symbionts within a host. In this issue of Molecular Ecology, Mathé‐Hubert, Kaech, Hertaeg, Jaenike, and Vorburger (2019) use data on the symbiont status of field‐collected aphids to inform a model on the evolution of symbiont co‐occurrence. They vary the effective female population size as well as the rate of horizontal and maternal transmission to infer the relative impact of symbiont‐symbiont interactions versus random drift. Additional data analysis revisits an association between two symbionts in a fruit fly species using a long‐term data set to highlight that such interactions are not limited to aphids.


Funding information
Biotechnology and Biological Sciences Research Council, Grant/Award Number: BB/S010556/1 Symbiotic interactions between insects and bacteria have long fascinated ecologists.
Aphids have emerged as the model system on which to study the effect of endosymbiotic bacteria on their hosts. Aphid-symbiont interactions are ecologically interesting as aphids host multiple secondary symbionts that can provide broad benefits, such as protection against heat stress or specialist natural enemies (parasitic wasps and entomopathogenic fungi). There are nine common aphid secondary symbionts and individual aphids host on average 1-2 symbionts. A cost-benefit trade-off for hosting symbionts is thought to explain why not all aphids host every possible symbiont in a population. Both positive and negative associations between various symbionts occur, and this could happen due to increased costs when cohosting certain combinations or as a consequence of competitive interactions between the symbionts within a host. In this issue of Molecular Ecology, Mathé-Hubert, Kaech, Hertaeg, Jaenike, and Vorburger (2019) use data on the symbiont status of field-collected aphids to inform a model on the evolution of symbiont co-occurrence. They vary the effective female population size as well as the rate of horizontal and maternal transmission to infer the relative impact of symbiont-symbiont interactions versus random drift. Additional data analysis revisits an association between two symbionts in a fruit fly species using a long-term data set to highlight that such interactions are not limited to aphids.

K E Y W O R D S
adaptation, community ecology, insects, species interactions, symbionts 2018). Horizontal transfer of symbionts among aphids can also occur during sexual reproduction, by parasitic wasps when ovipositing eggs into the aphids, or even through infected honeydew (reviewed in Zytynska & Weisser, 2016). To maintain coexistence of symbiont-infected and uninfected aphids, a recent mathematical model showed that horizontal transmission rates via parasitic wasps must be low, otherwise all aphids become infected (Zytynska & Venturino, 2019); this model assumed 100% maternal transmission rates and no migration. A new meta-analysis supports the theory that symbionts are costly to host in aphids and other sap-feeding insects (Zytynska, Thighiouart, & Frago, 2019). However, the effect sizes associated with these interactions varied widely depending on the species and genotype of both symbiont and aphid. Much of this work has been achieved through controlled laboratory experiments, which compared clones of aphids that harboured a symbiont compared to those that were cured of these symbionts. Continued experiments to further understand the impact of cohosting multiple symbionts is encouraged.
Recent interest in aphid symbionts has developed beyond the initial "what do they do?" question towards more community ecology questions based on their role in natural systems. Currently, this is rather theoretical with a number of reviews describing potential effects within multispecies networks (McLean, Parker, Hrček, Henry, & Godfray, 2016), the role of plant and natural enemy diversity (Zytynska & Meyer, 2019), and the challenges for biocontrol (Vorburger, 2018). The building evidence that host-symbiont dynam- F I G U R E 1 Pea aphids (Acyrthosiphon pisum) are found in two common colour morphs (pink and green), and also form host-associated races that feed on different host plant species. A universal host-plant of these aphids is the broad bean plant (Vicia faba), on which the pictured aphids are feeding. Photo credit: Hugo Mathé-Hubert et al The second part of the study by Mathé-Hubert et al. (2019) focused on the within-species variation of the Spiroplasma symbiont.
They identified three main genetic clades (clusters) to show that while there was no variation in strain frequency across host plant species, the frequency of these clades varied strongly due to the presence of other symbionts. For example, Spiroplasma from clade 2 was more frequent in aphids that cohosted other symbionts, and clade 3 Spiroplasma was less frequent if the aphid cohosted H. defensa. This highlights the necessity to consider the co-occurrence of aphid and symbiont genotypes as well as variation across host species, and can also inform on rates of horizontal transmission as suggested by Mathé-Hubert et al. (2019). While this is not a new idea (Ferrari & Vavre, 2011), we now have the knowledge and ability to design field experiments to transfer theoretical and controlled experimental work to realistic systems. Mathé-Hubert et al. (2019) also revisited an interaction between Wolbachia and Spiroplasma in Drosophila neotestacea (Figure 2) as the third and final analysis in the paper. Based on a long-term data set detailing this association, Mathé-Hubert et al. (2019) applied their model to the data and found that a decline in frequency of this association was not due to drift.
In fact, when accounting for drift a strong symbiont-symbiont interaction was still detected. A previous study by Fromont, Adair, and Douglas (2019) showed that while Wolbachia has a positive effect on Spiroplasma density, this does not happen vice versa-yet Wolbachia does benefit from the pathogen-fighting function of Spiroplasma.
Overall, these set of analyses reveal that symbiont-symbiont interactions are abundant and not limited to one particular insect group.
With increasing efforts in detecting symbiosis across multiple insect groups, future work could uncover such effects in a wider set of species.
Research on insect-symbiont interactions is currently taking an exciting leap forward. We are beginning to use experimental and field survey data to inform models on which to create future experimental hypotheses as here by Mathé-Hubert et al. (2019). One important step is to incorporate the cohosting of multiple symbionts into our models and experiments. It follows that there is great potential for uncovering novel microbe-mediated functions in ecological systems, with the term "holobiont" often used to describe an individual and all its interacting species (as an ecological unit). The societal impact of this knowledge is also high, with natural enemies often used as a main biocontrol method in glasshouses -and where the spread of protective symbionts in pest populations can have major consequences for biocontrol efficiency (Vorburger, 2018).

F I G U R E 2
The fruit fly Drosophila neotestacea breeds on mushrooms and is often studied for their interactions with Howardula nematode parasites. Here, they are studied for their interactions with endosymbiotic bacteria Spiroplasma and Wolbachia. Photo credit: Hugo Mathé-Hubert et al