Social transmission of bacterial symbionts homogenizes the microbiome within and across generations of group-living spiders

Host–symbiont relationships range from obligate, tightly integrated associations to facultative, environmentally determined ones. Understanding temporal and spatial variation in host-associated microbiomes requires identifying how symbionts are transmitted within and across generations. Transmission can be vertical (from parents to offspring) or horizontal (among co-occurring hosts or from the environment), and many systems exhibit mixed-mode transmission. The timing of acquisition can be critical, with several taxa showing narrow developmental windows for beneficial colonization. Social interactions can mediate symbiont transfer; group living increases contact rates and opportunities for repeated inoculations via behaviors such as grooming, trophallaxis, and communal feeding, promoting homogenization within and between generations. In social spiders (e.g., Stegodyphus dumicola), group living, communal prey feeding, regurgitation feeding of offspring, and matriphagy create multiple opportunities for social transmission. Prior studies report low-diversity microbiomes in S. dumicola dominated by a few endosymbionts (often Diplorickettsia), with strong within-nest similarity but between-nest variation. Given that many females do not reproduce, selection may favor social transmission to avoid dead-ends for strictly vertical (germline) transmission. The authors hypothesized that (1) horizontal transmission mediated by social interactions contributes to microbiome assembly and (2) social transmission homogenizes microbiomes within nests. They tested these hypotheses via life cycle sampling, cross-fostering, and adult mixing experiments using 16S rRNA gene amplicon sequencing and qPCR.

The paper situates its research within a broad literature on symbiont transmission modes and their evolutionary consequences. It highlights that stable host–symbiont associations may arise via vertical, horizontal, or mixed transmission, with the timing of acquisition often constrained to critical developmental windows (e.g., corals, stinkbugs, earthworms, and social ants). Social behaviors in group-living animals (primates, mice, honey bees, bumble bees, termites) promote microbial exchange and can maintain both intracellular and extracellular symbionts. In social insects, social transmission can complement or replace vertical transmission, especially where non-reproductive workers would otherwise be dead-ends for vertically transmitted symbionts. In Stegodyphus spiders, prior work documents low-diversity, nest-specific, temporally stable microbiomes dominated by few taxa, suggesting controlled transmission, but the role of social interactions had remained untested. This study addresses that gap by experimentally dissecting timing and routes of transmission in a social spider.

Field collection and husbandry: Communal nests of Stegodyphus dumicola were collected in Namibia and Botswana between April 2017 and November 2019 and transported to a laboratory. Spiders were maintained under a 1:1:1 light regime with fluctuating temperatures from approximately 20°C (night) to 29°C (day) and were watered daily and fed twice weekly with crickets (Gryllus bimaculatus) and blowfly larvae (Calliphora vomitoria). Based on prior evidence of strong within-nest similarity, 2–3 spiders were sampled per nest or life stage to represent nest microbiomes. Life cycle experiment: Four nests with distinct bacterial symbiont compositions were chosen. From each nest, eggs and spiderlings across instars (I1–I9, where possible) were sampled to determine the timing of microbiome acquisition. Instars were identified by morphological traits and behavior. Samples were frozen at −80°C prior to processing. DNA was extracted for 16S rRNA gene amplicon sequencing and qPCR. Cross-fostering experiment: Conducted in spring 2020 using seven nests collected in November 2019. Egg sacs were either left with natal females (control) or transferred to foster nests whose adult females harbored different microbiome compositions. Offspring microbiomes were compared to those of their natal mothers/female helpers and their foster mothers using Bray–Curtis dissimilarity. Adult mixing experiments: Individuals from nests with markedly different microbiomes were combined in shared containers for 39 days. Mixing treatments included equal ratios (e.g., 1:1 per source) and unequal ratios (2:8 or 8:2). Microbiome compositions before and after mixing were profiled to assess horizontal transmission and homogenization. Molecular methods: Each spider was homogenized in liquid nitrogen; DNA was extracted with the Qiagen DNeasy Blood & Tissue Kit (animal tissue protocol). Extraction blanks were included as contamination controls. 16S rRNA gene libraries targeted the V3–V4 regions using primers Bac341F/Bac805R, prepared per Illumina’s 16S Metagenomic Sequencing Library Preparation guide, and sequenced on an Illumina MiSeq with 2×300 bp chemistry, including negative controls in each run. qPCR quantified bacterial load as the ratio of 16S rRNA gene copies to copies of a conserved Stegodyphus gene. Bioinformatics and statistics: Reads were processed with DADA2; chimeras were removed, and ASVs were taxonomically classified using the SILVA SSU database (nr. 132). Analyses were performed in R using phyloseq, microbiome, vegan, ggplot2, and custom scripts. Only ASVs classified as Bacteria were retained. Putative contaminants were removed using the prevalence method (threshold 0.3) via the Decipher package, incorporating extraction and PCR negatives. Community similarity was assessed with Bray–Curtis dissimilarity. Diversity metrics (Shannon index, McNaughton’s dominance index) were calculated. Statistical tests included Wilcoxon rank-sum tests with FDR correction. Relationships between bacterial load and spider body size were evaluated by linear modeling (reporting adjusted R² and p-values).

- Adult nests harbored low-diversity microbiomes strongly dominated by one or two endosymbionts (Shannon index range ≈ 0.210–1.874; McNaughton’s dominance index ≈ 0.762–1; dominant taxa could range from 53–100% relative abundance). - Timing of acquisition: No bacterial 16S rRNA genes were detected in eggs or first instar (within closed egg sacs) by either qPCR or amplicon PCR. Bacteria were first detected at Instar 2, coinciding with emergence from the egg sac and onset of social feeding interactions (regurgitation feeding). At this stage, offspring already carried all main ASVs present in the parental generation; additionally, each nest sometimes showed one extra low-abundance ASV in some offspring and stages. - Bacterial load increased with host size: A linear relationship was observed between bacterial load (16S copies per host gene) and spider body size (Adjusted R² = 0.4244, p < 0.0001), with load dynamics varying by ASV and nest (e.g., Mycoplasma increased rapidly; Ca. Arachinospira more gradually; Diplorickettsia remained low in early ontogeny). - Cross-fostering: Offspring reared by natal females had microbiomes similar to their caretakers (mean Bray–Curtis dissimilarity ≈ 0.316; Wilcoxon rank-sum, FDR-adjusted p < 0.001). Offspring hatched from egg sacs transferred to foster nests acquired microbiomes similar to foster mothers (mean Bray–Curtis ≈ 0.384; FDR-adjusted p < 0.001) and dissimilar to their biological mothers/female helpers (mean Bray–Curtis ≈ 0.724; FDR-adjusted p < 0.001). - Adult mixing experiments (39 days): When individuals from two nests were mixed at equal ratios, the resulting microbiomes of all individuals converged to a combined, homogenized composition. With unequal ratios (2:8 or 8:2), transmission success depended on symbiont identity: Ca. Arachinospira transmitted effectively even from minority donors; Diplorickettsia remained similar in original hosts but established at lower abundance in newly infected individuals; Mycoplasma transfer was ASV-dependent and context-dependent (e.g., one ASV transferred well alone but poorly when co-occurring with another ASV). Some ASVs not initially detected in the source nests appeared after mixing, while Acariomes became rare in mixed groups. - Overall, offspring microbiomes consistently matched those of the adult females with whom they shared the nest, and adult co-housing promoted horizontal exchange and within-group homogenization.

The study resolves the timing and routes of symbiont transmission in a social spider. Offspring hatch symbiont-free, ruling out transovarial transmission under the conditions studied. Acquisition occurs shortly after hatching, coincident with regurgitation feeding and close social contact, indicating vertical transmission across generations mediated by social behavior. Cross-fostering demonstrates that the nest social environment, rather than natal origin, determines offspring microbiome assembly. Adult mixing experiments show that social contact also drives horizontal exchange among group members, homogenizing microbiomes within groups. These processes explain the strong within-nest similarity and temporal stability of microbiomes, alongside distinct between-nest differences. Transmission fidelity is high but not absolute: success varies by symbiont identity and community context (e.g., apparent incompatibility between certain Mycoplasma ASVs), suggesting inter-symbiont interactions and priority effects. Diplorickettsia increases later in development, potentially reflecting resource-driven dynamics or interactions with co-occurring symbionts as spiders begin communal feeding. Environmental reservoirs (e.g., silk nests) appear to contribute little based on prior work showing limited overlap with spider microbiomes, though vectors or prey could occasionally introduce symbionts, especially during colony founding or repeated contacts. Overall, mixed-mode, socially mediated transmission provides a mechanistic basis for the stability of host–symbiont associations in group-living spiders and aligns with patterns seen in other social taxa.

This work shows that in Stegodyphus dumicola, high-fidelity microbiome transmission is achieved through socially mediated vertical transfer early in life (via regurgitation feeding) and horizontal exchange among group members, producing homogeneous within-nest microbiomes across generations. Eggs and first instars lack detectable bacteria, indicating no transovarial transmission. Cross-fostering and mixing experiments reveal that the social environment dictates microbiome assembly and can homogenize divergent communities, with transmission efficiency modulated by symbiont identity and community context. These findings highlight social transmission as a key mechanism maintaining stable host–symbiont associations in group-living species. Future directions include: elucidating the precise behavioral and physiological mechanisms of transfer (e.g., quantifying contributions of regurgitation feeding vs. other contacts), assessing the roles of environmental reservoirs and prey in rare introductions, dissecting inter-symbiont interactions and compatibility, evaluating long-term dynamics across colony life cycles and during nest founding, and testing functional consequences of microbiome variation for host fitness and social behavior.

- Mechanistic pathways of transfer were inferred from timing and experimental designs but not directly visualized or quantified at behavioral or tissue levels. - Potential environmental or prey-mediated introductions cannot be fully excluded; prior studies suggest limited nest reservoir overlap, but this study did not experimentally manipulate environmental sources. - Adult mixing experiments were of limited duration (39 days) and involved specific mixing ratios, which may not capture long-term or natural colony dynamics. - Small within-nest sample sizes (2–3 individuals as representatives) assume strong within-nest homogeneity; rare within-nest variation may be underdetected. - 16S rRNA amplicon data provide taxonomic resolution to ASVs but limited functional inference; qPCR quantifies load but not activity. - Geographic and temporal sampling, while spanning multiple years and sites, may not capture the full range of natural variability.