Pre- and post-association barriers to host switching in sympatric mutualists

Coevolution between mutualists can lead to reciprocal specialization, potentially causing barriers to host switching.  In the present study, we conducted assays to identify pre- and post-association barriers to host switching by endosymbiotic bacteria, both within and between two sympatric nematode clades.  In nature, Steinernema nematodes and Xenorhabdus bacteria form an obligate mutualism.  Free-living juvenile nematodes carry Xenorhabdus in a specialized intestinal receptacle.  When nematodes enter an insect, they release the bacteria into the insect hemocoel.  The bacteria aid in killing the insect and facilitate nematode reproduction.  Prior to dispersing from the insect, juvenile nematodes must form an association with their symbionts; the bacteria must adhere to the intestinal receptacle.  We tested for pre-association barriers by comparing the effects of bacterial strains on native verses non-native nematodes via their virulence towards, nutritional support of, and ability to associate with different nematode species.  We then assessed post-association barriers by measuring the relative fitness of nematodes carrying each strain of bacteria.  We found evidence for both pre- and post-association barriers between nematode clades.  Specifically, some bacteria were highly virulent to nonnative hosts, and some nematode hosts carried fewer cells of nonnative bacteria, leading to pre-association barriers.  In addition, reduced infection success and lower nematode reproduction were identified as post-association barriers.  No barriers to symbiont switching were detected between nematode species within the same clade.  Overall, our study suggests a framework that could be used to generate predictions for the evolution of barriers to host switching in this and other systems.

Methods
Data Collection

In vitro development assays were conducted on two types of media to differentiate bacterial virulence towards (LKA media) or nutritional support (LA media) of the nematodes.  For each media type, three nematode isolates carrying their native symbionts were tested on lawns of 6 strains of bacteria. Five replicates of each combination was observed daily for 8 days.  For in vivo association assays,we infected each Galleria mellonella caterpillar by pipetting 100 nematodes in 500uL of ddH2O on to the dorsum of the caterpillar. For pre-association assays, the nematodes were axenic, and bacteria cultures were injected into the caterpillars 24 hours after nematode infection.  For post-association assays, the nematodes were assumed to be carrying bacteria.  Infection success (Infection_Success) was measured as the proportion of infected caterpillars with any nematode emergence.  Mean nematode emergence (Estimated_IJ_Count) was estimated by volumetric subsampling.  We estimated bacterial carriage (CFUs) by crushing a sample of 1000 IJs for at least 5 collections per treatment group per generation.  If no colonies grew from the crushing, we used the detection limit of 0.003 CFU/IJ as the estimated carriage the in statistical analyses.  

Software and data processing description

All in vitro statistical analyses were performed in SAS version 9.4.  All in vivo statistical analyses were performed in R version 3.6.3.