Abstract
CRISPR adaptation immunizes bacteria and archaea against viruses. During adaptation, the Cas1-Cas2 complex integrates fragments of invader DNA as spacers in the CRISPR array. Recently, an additional protein Cas4 has been implicated in selection and processing of prespacer substrates for Cas1-Cas2, although this mechanism remains unclear. We show that Cas4 interacts directly with Cas1-Cas2 forming a Cas4-Cas1-Cas2 complex that captures and processes prespacers prior to integration. Structural analysis of the Cas4-Cas1-Cas2 complex reveals two copies of Cas4 that closely interact with the two integrase active sites of Cas1, suggesting a mechanism for substrate handoff following processing. We also find that the Cas4-Cas1-Cas2 complex processes single-stranded DNA provided in cis or in trans with a double-stranded DNA duplex. Cas4 cleaves precisely upstream of PAM sequences, ensuring the acquisition of functional spacers. Our results explain how Cas4 cleavage coordinates with Cas1-Cas2 integration and defines the exact cleavage sites and specificity of Cas4.