A tripartite survey of CRISPR-Cas diversity: Assessing the contours and limits of natural variation in CRISPR-Cas systems by associated genes and spacer acquisition biases

Explorations of anti-phage defence systems, with a focus on CRISPR-Cas system diversity, in the past decade has enabled many significant advances in biotechnology especially in programmable gene editing platforms. Many previous investigations have employed guilt by association based data mining approaches to discover new anti-phage defence systems. In light of these endeavors, it remains unclear to what extent undiscovered functional diversity linked to CRISPR mediated immunity remains to be explored. In this study, I employed 3 interrelated avenues of inquiry to assess different aspects of CRISPR-Cas system diversity. I emulated past data mining approaches and constructed a computational pipeline which screened known and unknown putative CRISPR-associated genes from a large volume of assembled sequencing data. I then conducted a census of the landscape of genes co-encoded in proximity to array containing CRISPR-Cas subtypes. I found that a significant fraction of co-encoded genes possessed homology to anti-phage defence genes and mobile genetic elements even with low CRISPRicity scores, which traditionally been used as the gold standard for proof of co-association with CRISPR-Cas subtypes. A small number of associated antiphage defence genes such as HicAB and DrmB were also found in novel association with Type VI CRISPR-Cas systems. The second avenue I employed was to analyse the intra-subtype diversity using multi-gene gene-genome network based taxonomic representations. Focusing this approach on Type VI CRISPR-Cas systems, I observed a high degree of segregation between local clusters, with relatively few shared genes. Conversely, many additional genes predicted to be anti-phage defence related were found co-encoded at the level of local clusters. I then performed spacer mapping using spacers derived from the type VI systems CRISPR arrays. This revealed that local clusters of Type VI systems target a mixture of plasmids and phages, even in cases where the type VI CRISPR-array was found to be prophage encoded. As a consequence of red-queen competition between host-encoding CRISPR-Cas systems and mobile genetic elements (MGEs) such as phages or plasmids, the taxonomic diversity of CRISPR associated genes and mapped sequences of their corresponding MGEs are inextricably linked. This justified constructing an analogous gene sharing network to model and analyse the diversity of these sequences. In contrast to the sequences of host-encoded Type VI systems, these sequences formed a contiguous and extensive network of many exchanged and inter-related genes. Finally, a bioinformatic technique called "spacer distribution analysis" was employed to predict priming-like effects which have been shown to result from the coupling of interference and adaptation in certain CRISPR-Cas subtypes. This uncovered evidence of priming in type I-A and type V-F CRISPR-Cas systems. I also extended the sensitivity of spacer distribution analysis by including spacers with partial matches which increased sensitivity of spacer distribution analysis in certain subtypes. However, this increase was accompanied by differences in the spacer distribution observed. In Summary, my investigation discovered several novel gene co-associations and priming-like effects in several CRISPR-Cas subtypes which have gone unreported in previous studies. This suggests that pockets of functional diversity among known and unknown CRISPR-Cas systems remains to be explored.

Author dc:creator

• Mckay, Alexander

Language dc:language.iso

en_AU

Handle dc:identifier.uri

https://hdl.handle.net/1885/733747446

OAI identifier oai:identifier

oai:openresearch-repository.anu.edu.au:1885/733747446