The diversification of a host organism can be influenced by both the external environment and its assemblage of microbes. Here, we use a young lineage of spiders, coupled with a chronologically arranged series of volcanic mountains, to determine the evolutionary history of a host and its associated microbial communities, altogether forming the “holobiont”. Using the stick spider Ariamnes waikula (Araneae, Theridiidae) on the island of Hawaiʻi, and outgroup taxa on older islands, we tested whether the host spiders and their microbial constituents have responded in similar ways to the dynamic abiotic environment of the volcanic archipelago. The expectation was that each component of the holobiont (the spider hosts, intracellular endosymbionts, and gut microbiota) should show a similar pattern of sequential colonization from older to younger volcanoes. In order to investigate this, we generated ddRAD data for the host spiders and 16S rRNA gene amplicon data from their microbiota. Results showed that the host A. waikula is strongly structured by isolation, suggesting sequential colonization from older to younger volcanoes. The endosymbiont communities were markedly different between Ariamnes species on different islands, but more homogenized among A. waikula populations. In contrast, the gut microbiota was largely conserved across all populations and species, and probably mostly environmentally derived.  Our results highlight the different evolutionary trajectories of the distinct components of the holobiont, showing the necessity of understanding the interplay between components in order to assess any role of the microbial communities in host diversification.

To examine the population structure of the spider host, we used ddRAD to obtain reduced representation genome-wide SNP data. Genomic DNA was extracted from spider legs with several modifications to the Qiagen DNeasy kit protocol. Legs were first removed from each specimen using sterile tweezers so that the abdomen remained intact for the microbial DNA analysis. DNA was then extracted by placing the tissue in Proteinase K and lysis buffer and grinding them with a sterile pestle to break up the exoskeleton. We then added 4uL RNase A (100 mg/ml) and incubated the extractions for two minutes at room temperature. Tubes with tissue and extraction solution were then placed overnight in a heat block at 56°C. The remainder of the extraction protocol was performed following the manufacturer’s instructions. We built ddRAD libraries following an adapted protocol of Peterson et al. (2012) (Saarman & Pogson, 2015; see Maas et al. 2018 for protocol optimization steps). Briefly, we started the ddRAD protocol with a total of 100 nanograms of DNA per sample. The DNA was digested using SphI‐HF (rare‐cutting) and MlucI (frequent‐cutting) restriction enzymes. We assessed fragmentation with a Bioanalyzer High Sensitivity chip (Agilent). We multiplexed 15-20 individuals per library for a total of eight ddRAD libraries. We used a Sage Science Pippen Prep to size select 451-551bp (including internal adapters) fragments, and confirmed the sizes using a Bioanalyzer. Ten indexing polymerase chain reaction cycles (PCRs) were run on each library to enrich for double‐digested fragments and to incorporate a unique external index for each library pool. The eight libraries were sequenced using 100bp paired-end sequencing on one Illumina HiSeq 2500 lane at the Vincent J. Coates Genomic Sequencing Facility at UC Berkeley.

To characterize the microbial community within the A. waikula hosts, 71 individuals from eight populations of three Hawaiian Islands were selected for analysis. We focused on the mid and hindgut, both located in the spider’s opisthosoma. The preservation in ethanol led to considerable shrinkage of the opisthosoma and thus did not allow us to separately dissect out the gut. Instead, we used the whole opisthosoma to extract DNA. Specimens which did not have the opisthosoma intact were not used. The digestive tract comprises the majority of the opisthosoma’s cavity. In addition, it contains silk glands, the heart, lungs, and gonads. The opisthosoma was removed with a sterile razor blade and then washed in ethanol to remove external bacteria (Hammer et al. 2015). We considered it to be representative of the “gut microbiota”, even if it technically consists of the “opisthosoma microbiota”, but previous studies have shown that the gut microbiota dominates in the opisthosoma (Sheffer et al. 2020, Kennedy et al. 2020). The tissue was then transferred into lysis buffer and finely ground with a sterile pestle. DNA was extracted using the Gentra Puregene Tissue Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. Spider abdominal tissue can contain PCR inhibitors (Schrader et al. 2012), thus we cleaned the DNA extract with 0.9X AmPure Beads XP.

We next amplified a ~300 bp fragment of the V1-V2 region of the bacterial 16S rRNA using the Qiagen Multiplex PCR kit according to the manufacturer’s protocols and using the primer pair MS-27F (AGAGTTTGATCCTGGCTCAG) and MS-338R (TGCTGCCTCCCGTAGGAGT) (Gibson et al. 2014). PCRs were run with 20ng of template DNA and 30 cycles at an annealing temperature of 55°C. PCR products were separated from leftover primer by 1X AmPure Beads XP. A six-cycle indexing PCR was performed on the cleaned products, adding dual indexes to every sample using the Qiagen Multiplex PCR kit. Indexing was performed according to (Lange et al. 2014). The dual indexed libraries were isolated from leftover primer as described above, quantified using a Qubit fluorometer, and pooled in equal amounts into a single tube. The library was sequenced on an Illumina MiSeq using V3 chemistry and 300 bp paired reads. We also performed blank extraction controls and negative PCR controls (without DNA template) in order to discard contaminants from our final dataset.