Allosteric activation of the CRISPR-associated transcription factor Csa 3 by cyclic tetra-adenylate (cA 4)

Abstract

The CRISPR-Cas immune system provides adaptive and heritable immunity to archaea and bacteria to combat viral infection, and is a source of biochemical tools to researchers. This work combines structural biology and biochemical approaches to provide insight into mechanisms prokaryotes use to control the CRISPR-Cas immune system, linking subsystems into a coordinated response. The first structure of S. solfataricus Csa3 determined by Lintner et al. revealed a dimer with a C-terminal wHTH DNA-binding domain and an N-terminal CARF domain with a putative ligand binding site predicted to bind a two-fold symmetric molecule with both negatively charged and hydrophobic/aromatic moieties, such as dinucleoside polyphosphates or nucleic acid molecules. 1 Later, analysis by Topuzlu et al. of the A. fulgidis Csx3 structure containing a 4 base RNA molecule in a binding pocket revealed similarities between Csx3 and the Csa3 CARF domain, and suggested CARF proteins could bind cyclic or pseudosymmetric linear RNA tetranucleotides represented by the ring-shaped RNA density in the Csx3 binding pocket. 2 Functional studies with S. islandicus Csa3 identified that SiCsa3 regulates transcription of acquisition genes (cas1, cas2, and cas4) and several CRISPR loci. 3,4 Additionally, two groups simultaneously showed that the Type III surveillance complexes produce cyclic oligoadenylate messengers, including cyclic tetra-adenylate (cA 4), which allosterically regulate the RNase activity of Csm6 and Csx1, other CARF proteins. 5,6 These advances support the original predictions by Lintner et al. and suggest that Csa3 binds a cyclic RNA as proposed by Topuzlu et al. Binding of cA 4 likely allosterically causes conformation changes in the wHTH, and regulates the protein's transcriptional regulation. We present the crystal structure of S. solfataricus Csa3 complexed with cyclic tetraadenylate (cA 4). cA 4, as predicted, 1 is bound in the CARF domain 2-fold symmetric pocket, which stimulates conformational changes in the C-terminal domain. Additionally, we identify the presence of a palindromic predicted binding motif upstream of the Type I-A(2) acquisition cassette and CRISPR loci C and D, and reveal through EMSA analysis that Csa3 binds dsDNA nonspecifically with high affinity. Finally, ring nuclease activity is not detected in Csa3, suggesting longer term potentiation of the cA 4 in Csa3 than observed for Csx1/Csm6. 5,6