Origins and Evolution of Tracrrnas in CRISPR-Cas Systems

The CRISPR-Cas systems belong to two classes, with multi-subunit effector complexes in Class 1, and single multi-domain protein effectors in Class 2. The latter encompasses types II, V and VI that differ primarily by the domain architectures of the effector proteins. In all type II systems as well as subtype V-B, maturation of the precursor crRNA (pre-crRNA) requires a trans-acting CRISPR (tracr) RNA whereas in subtype V-A and type VI, no tracrRNA has been identified. In Cas9-CRISPR technologies, tracrRNA is artificially fused to crRNA to form a single-guide RNA. The tracrRNA consists of a portion complementary to the direct repeat from the corresponding CRISPR array and a unique portion. The origin and evolution of tracrRNA remain open problems which we sought to investigate by comparative analysis of co-folding between tracrRNA and crRNA. By comparing the available structures of Class 2 effectors complexed with the crRNA, tracrRNA (where involved) and the target, we identified the nexus structure of the tracrRNA in types II-A and II-B and a similar local fold in type V-B, where it involves both tracrRNA and the repeat portion of the crRNA, and in subtype V-A, where it is formed by the repeat alone. This observation suggests an important function of the nexus structure that is common to all type II and type V systems but could have evolved convergently, involving tracrRNA alone, the tracrRNA-repeat hybrid or the repeat alone. Using a larger set of tracrRNA-repeat pairs representative of the diversity of type II systems, we predicted the structures of each crRNA-tracrRNA co-folding and identified the nexus structure in almost all of them. The nexus is almost always located 2 to 3 nucleotides away from the tracrRNA-repeat hybrid region. We also observed that the 5’-terminal base of the repeat is always paired with the tracrRNA. We propose that the role of the nexus structure is to prevent base-pairing between tracrRNA and the spacer whereas the pairing of the 5' base of the repeat prevents interaction between the repeat and the DNA target. Thus, both features appear to ensure full availability of the spacer to interact specifically and completely with the DNA target. We investigated the origins of tracrRNA by tracing the coevolution of the tracrRNA and the corresponding repeats using simulated and ancestral RNA guide sequences, and observed a conservation of optimal hybrid energy that corresponds to partial, rather than complete, base-pairing. Then, we analyzed the tracrRNA in microbial genomes and found that tracrRNAs located close to CRISPR arrays are expressed from the same strand and downstream of the CRISPR array, and can even overlap with the array. We propose an evolutionary model in which tracrRNA emerges from the distal decaying repeat of the CRISPR array.