Impaired immune function accompanies social evolution in spiders

An efficient immune system is essential to the survival of many animals. Sociality increases risk of pathogen transmission, which should select for enhanced immune function. However, two hypotheses instead predict a weakened immune function: relaxed selection caused by social immunity/protection, and reduced efficacy of selection due to inbreeding, reproductive skew, and female bias in social species that reduces effective population size and accelerate genetic drift. We assessed the effect of social evolution on immune function in a comparative study of two social spider species and their closely related subsocial sister species (genus Stegodyphus). The haemolymph of social species was less efficient in inhibiting the growth of potentially pathogenic bacteria than that of subsocial species. Reduced efficacy of selection in social species was supported by comparative genomic analysis of immune genes showing substantially elevated non-synonymous substitutions in one of the social species. We propose that this impaired immune function is likely to be a result of reduced efficacy of selection because the evolution of sociality in spiders is accompanied by demographic processes that elevate genetic drift. There may also be positive feedback between pathogen-induced local extinctions and the resulting elevation of genetic drift which further weakens responses to selection by pathogens.Funding provided by: Danmarks Frie ForskningsfondCrossref Funder Registry ID: http://dx.doi.org/10.13039/501100004836Award Number: 6108-00565Two assays were conducted, one with the gram-positive Bacillus subtilis DSM10 and one with the gram-negative Escherichia coli OP50, both in NB medium (Scharlau, Spain). Samples of haemolymph were obtained from fully hydrated, sterilised and anaesthetized female spiders by piercing a book lung and collecting 2 μl from the emerging droplet. The extracted haemolymph was then randomly distributed between 100 μl samples of bacterial medium in 96-well plates. Plates were incubated at 27°C except during measurements at room temperature (~20°C). Each plate contained two treatment controls; one sample of NB medium without haemolymph and one sample of NB medium with Ringer solution, the latter to mimic the haemolymph salt composition. Sample sizes for inhibition assays were: B. subtilis: nafricanus = 37, nmimosarum = 51, ntentoriicola = 14, ndumicola = 28, nNB = 27, nNB+Ringer = 27, and E. coli: nafricanus = 35, nmimosarum = 52, ntentoriicola = 14, ndumicola = 24, nNB = 24, nNB+Ringer = 24. Before the addition of haemolymph, bacterial concentration in each well was measured as light absorbance using an Epoch microplate spectrophotometer to standardize start concentrations. After addition of haemolymph the light absorbance was measured every 2nd minute from 0 to 16 min, every 5th minute from 20 to 50 min, and then at 60, 70, 80, 100 and 120 min. To investigate if selection at the molecular level differs between social and subsocial species, we used 23 immune genes annotated in the S. mimosarum genome. Alignments consisting of ortholog sequences from S. mimosarum, S. africanus and S. lineatus were obtained from Bechsgaard et al. Stegodyphus dumicola sequences were extracted from the genome assembly published by Liu et al. and S. tentoriicola sequences were extracted from an unpublished genome assembly. We used PAML 4.6 to estimate dN/dS ratios for each branch separately, first for individual loci, and secondly for all loci concatenated