Ecology Shapes Microbial Immune Strategy: Temperature and Oxygen as Determinants of the Incidence of CRISPR Adaptive Immunity

Abstract

Bacteria and archaea are locked in a near-constant battle with their viral pathogens. Despite previous mechanistic characterization of numerous prokaryotic defense strategies, the underlying ecological drivers of different strategies remain largely unknown and predicting which species will take which strategies remains a challenge. Here, we focus on the CRISPR immune strategy and develop a phylogenetically-corrected machine learning approach to build a predictive model of CRISPR incidence using data on over 100 traits across over 2600 species. We discover a strong but hithertounknown negative interaction between CRISPR and aerobicity, which we hypothesize may result from interference between CRISPR associated proteins and non-homologous end-joining DNA repair due to oxidative stress. Our predictive model also quantitatively confirms previous observations of an association between CRISPR and temperature. Finally, we contrast the environmental associations of different CRISPR system types (I, II, III) and restriction modification systems, all of which act as intracellular immune systems.

Identification of Natural CRISPR Systems and Targets in the Human Microbiome

“…The spacer load on mucosal surfaces (buccal mucosa, hard palate, keratinized gingiva, palatine tonsils, saliva, throat, tongue dorsum, gut, and vaginal) is nearly one magnitude higher than that on non-mucosal surfaces (skin) but highest on mucosal-adjacent surfaces (supragingival and subgingival plaque). Though environmental factors such as aerobicity correlate with CRISPR incidence (Weissman et al, 2019) across taxa, HMP-1-II-derived spacer load was not directly associated with the oxygen exposure of the body site (median spacer load for high-O 2 sites = 2.71 CPM, 46.96 CPM for mid-O 2 sites, and 18.19 CPM for low-O 2 sites). Instead, functions of the CRISPR system beyond phage targeting, such as control of genes involved in commensalism and virulence and regulation of inter-microbial interactions within the host , could explain the difference in spacer load between these surfaces.…”

The ecology and evolution of microbial CRISPR-Cas adaptive immune systems

“…Moreover, some uncultured bacterial lineages are virtually devoid of CRISPR-Cas immune systems [30]. A recent computer learning approach suggested that abiotic factors such as oxygen levels and temperature are important predictors of whether microorganisms encode CRISPR-Cas immune systems [31]. However, the ecological drivers of CRISPR distribution remain unclear.…”

Exploring the functional composition of the human microbiome using a hand-curated microbial trait database

“…In order to identify traits strongly associated with a particular body site we took a predictive approach that incorporated random forests for prediction with blocked cross validation [14,42] to correct our error estimates for phylogeny. We split each sample into four individual communities for each of its constituent phyla (Actinobacteria, Bacteroidetes, Firmicutes, Proteobacteria) and calculated the mean trait values individually for each of these phyla-samples (i.e., the set of species in a sample from a given phylum).…”