Lawrence Berkeley National Laboratory

Viruses are the most ubiquitous biological entities on Earth. Even so, elucidating the impact of viruses on microbial communities and associated ecosystem processes often requires identification of unambiguous host-virus linkages-an undeniable challenge in many ecosystems. Subsurface fractured shales present a unique opportunity to first make these strong linkages via spacers in CRISPR-Cas arrays and subsequently reveal complex long-term host-virus dynamics. Here, we sampled two replicated sets of fractured shale wells for nearly 800 days, resulting in 78 metagenomes from temporal sampling of six wells in the Denver-Julesburg Basin (Colorado, USA). At the community level, there was strong evidence for CRISPR-Cas defense systems being used through time and likely in response to viral interactions. Within our host genomes, represented by 202 unique MAGs, we also saw that CRISPR-Cas systems were widely encoded. Together, spacers from host CRISPR loci facilitated 2,110 CRISPR-based viral linkages across 90 host MAGs spanning 25 phyla. We observed less redundancy in host-viral linkages and fewer spacers associated with hosts from the older, more established wells, possibly reflecting enrichment of more beneficial spacers through time. Leveraging temporal patterns of host-virus linkages across differing well ages, we report how host-virus co-existence dynamics develop and converge through time, possibly reflecting selection for viruses that can evade host CRISPR-Cas systems. Together, our findings shed light on the complexities of host-virus interactions as well as long-term dynamics of CRISPR-Cas defense among diverse microbial populations.