Microbiome evolution plays a secondary role in host rapid adaptation

Abstract

Understanding how populations adapt to environmental change is a central goal in evolutionary biology. Microbiomes have been proposed as a source of heritable variation that is central to rapid adaptation in hosts, yet empirical evidence supporting this remains limited, particularly in naturalistic settings. We combined a field evolution experiment in Drosophila melanogaster exposed to an insecticide with microbiome manipulations to disentangle the contributions of host standing genetic variation and microbiome evolution to adaptation. Within three generations, independent populations rapidly and repeatedly evolved increased survivorship, a defining feature of resistance evolution. Adaptive changes in sub-lethal traits such as reproductive output, stress tolerance, and body size occurred with a delayed response following the evolution of resistance. Core microbiome taxa declined following insecticide exposure, and resistant populations evolved to house lower microbial abundances. Axenic rearing and microbiome transplant experiments demonstrated that adaptation via host standing genetic variation was the mechanism for resistance evolution. Microbiome evolution played a secondary and cryptic role in host adaptation by masking slowed development rates that evolved in resistant populations. Together, these results reinforce the primacy of adaptation occurring through selection on host standing genetic variation while also demonstrating the contributions of microbiome evolution in host adaptation.

Keywords: Evolution; Experimental Evolution; Holobiont; Host-Microbe Interactions; Microbiome; Rapid Adaptation; Resistance Evolution.

Figure 1.

Replicate outdoor mesocosms containing Drosophila melanogaster populations were exposed to non-toxic (no insecticide) and toxic (insecticide) conditions from mid-summer to late-fall to: (1) Test whether D. melanogaster populations evolve resistance to insecticide exposure, including both lethal and sub-lethal traits. In October, eggs from each population were collected and reared in common garden conditions for two generations to: (2) Test for the effects of insecticide exposure on microbiome composition; (3) Test for microbiome evolution associated with host resistance evolution by rearing descendants from each field cage in a common garden environment and examining differences in microbiome composition; and (4) Test for the contribution of microbiome evolution to host adaptation through two complementary approaches: (top panel) axenic rearing (microbiome removal) compared to xenic rearing to assess microbiomes contribution to resistance evolution; and (bottom panel) microbiome transplant experiments, where microbiota from control and resistant donor populations were introduced into germ-free hosts.

Figure 2.

Phenotypic evolution of life-history traits in populations exposed (‘resistant’) and unexposed (‘control’) to the insecticide. Phenotypic trajectories were measured at each time point following two generations of common garden rearing, with all assays conducted on insects reared on toxic (insecticide) media to assess resistance. (A) Mean survival, measured as the percentage of eggs (30 per vial) that survived to adulthood. (B) Mean fecundity, measured as the number of eggs produced per female per day. (C) Mean starvation tolerance, measured as the time to death by starvation. (D) Mean adult weight, measured as the average dry weight of female flies. Black triangles (▲) represent the mean phenotype of the founding population (outdoor cages initiated July 16), measured under toxic assay conditions. Thin colored lines represent the mean phenotypic trajectory of each individual population, and thick colored lines show the treatment means (averaged across cages). Mean development rate, calculated as the fraction of development time completed per day [1/(total hours/24)], can be found in the Supporting Information (Figure S1).

Figure 3.

Effects of insecticides on the microbiome measured both through direct exposure and evolutionary divergence associated with the evolution of resistance. Mean abundance (±1 SE) of the two most dominant microbial families, Acetobacteraceae and Lactobacillaceae, in Drosophila melanogaster populations exposed to insecticide stress. (A) To evaluate the microbiome’s response to insecticide exposure, microbiomes of the founding population were assayed after within-generation exposure to non-toxic (no insecticide) and toxic (insecticide) conditions, after two generations of common garden rearing. (B) To evaluate microbiome evolution, microbiomes of control and resistant populations were compared after two generations of common garden rearing on toxic media. Significant differences in abundance between populations are designated by an asterisk (*). The full list of microbial families can be found in the Supporting Information Table S2.

Figure 4.

Testing whether microbiome evolution contributed to host adaptation. Panels A-C show trait values for resistant and control populations reared on a toxic diet under either xenic (microbiome present) or axenic (microbiome removed) conditions. Points represent cage-level means, and lines connect cages that were paired based on spatial proximity in the outdoor experimental system. Bold points and error bars indicate the mean ± SE across cages. (A) Mean survival, measured as the percentage of eggs (30 per vial) that survived to adulthood. (B) Mean adult weight, measured as the average dry weight of female flies. (C) Mean development rate, calculated as the fraction of development time completed per day [1/(total hours/24)]. Axenic flies were generated by sterilization and reared in bioreactor tubes as gnotobiotic flies.

Figure 5.

Testing for the contribution of microbiome donors to host adaptation. Panels (A-D) show the results of microbiome transplants in which microbiomes from each of the 10 outdoor populations (five resistant, 5 control) were transplanted into subsets of the ‘founder’ population. Founder flies received gut microbiomes of either a pooled homogenate from common garden reared control or resistant populations. These recipient flies were measured to determine whether microbiome transplants could recover patterns of phenotypic adaptation. (A) Principal coordinate analysis (PCoA) of Bray-Curtis distances showing microbiome composition of recipient flies after transplants from control (blue) or resistant (red) donor populations; microbiome composition did not differ significantly by donor type. (B) Mean survival as the proportion of eggs (30 eggs per vial) that survived to adulthood; no significant difference by microbiome donor. (C) Mean development rate as the time at which flies eclosed (30 eggs per vial); no significant difference by microbiome donor. (D) Adult weight as the average dry weight of female flies. All flies were sterilized and reared in bioreactor tubes as gnotobiotic flies to complete all assays (B-D).