Examining patterns of CRISPR immunity in natural populations of Sulfolobus islandicus

Virus-host interactions are important for shaping microbial diversity in natural environments. Viruses impact their hosts through predation and also through gene flow. The recently discovered microbial CRISPR immune system is a mechanism that modulates interactions between hosts and viruses, and in the process, records a sequence-based history of interactions in the host’s DNA. We investigate virus-host interactions in the natural environment by assessing their patterns of diversity at the CRISPR locus within and between geographically isolated populations of Sulfolobus islandicus. S. islandicus is a thermoacidophilic crenarchaeon that has been isolated from acidic hot springs in the Northern Hemisphere. In this dissertation, a biogeographic pattern of isolation by distance, previously established for S. islandicus, is shown for viruses and plasmids that infect Sulfolobus. Core genes from a provirus show that strains that cluster genetically also cluster geographically. Additionally, CRISPR spacer sequences show higher identity to each other and sequenced viruses and plasmids from the same geographic location than to those from more distant locations. Because we have identified a biogeographic pattern for both host and virus, we can investigate local population dynamics. We examine a population of S. islandicus strains from a single time and place, and demonstrate an even and diverse population with a defining lack of strain dominance in the CRISPR arrays. This suggested a conceptual model in which clonal competition with many hosts having independently acquired different CRISPR spacers to the same virus persist in natural populations. Through a collaborative effort, we tested this model using simulations of density-dependent ecological dynamics with evolutionary changes associated with the CRISPR system in both host and virus. We see three types of virus-host dynamics emerge over the course of the simulations: dominance of a new strain with CRISPR immunity to dominant viruses, dominance of a rare strain that had been dominant in the population previously, and coalitions of multiple strains that have similar immunity to dominant viruses. Testing the model’s predictions requires natural population data from multiple time points, so we sampled the Mutnovsky population a second time, ten years later. Surprisingly, we see that the CRISPR system evolves largely through recombination of both the CRISPR spacer arrays and the cas genes. This allows immunity to be shuffled throughout the population such that every member of the population does not have to acquire immunity independently. This dissertation demonstrates a role for the CRISPR system in shaping virus-host interactions in natural populations of S. islandicus