latter phase of adaptation is mediated by highly conserved Cas1:Cas2 adaptation complex. However, some CRISPR systems also require additional proteins, such as Csn2 and Cas9 of S. pyogenes type II system (Heler et al. 2015; Wei et al. 2015), or Cas4 of type I and type V systems (Lee et al. 2018). Furthermore, in certain species with type III and type VI systems, fusions of cas1 and reverse transcriptase (RT) genes have been discovered (Silas et al. 2016; Gonzalez‐Delgado et al. 2019; Toro et al. 2019), and then shown to be able to insert spacers originating from RNA and DNA, where the integration of RNA is followed by cDNA synthesis by the RT domain (Silas et al. 2016; Gonzalez‐Delgado et al. 2019).
The adaptation complex is composed of the central Cas2 dimer flanked on either side by a Cas1 dimer (Figure 3.6d). The captured DNA fragments (termed prespacer DNA) are bound by the central Cas2 dimer, with the Tyr22 of the E. coli Cas1 acting as a protein wedge to splay open the bound dsDNA fragment (Nunez et al. 2014; Nunez et al. 2015). The splayed ends of the bound prespacer DNA can be trimmed by host’s DnaQ 3’‐5’ exonuclease‐domain containing enzymes (such as DNA polymerase III, exonuclease T, DnaQ, or in some species by Cas2‐DnaQ fusion proteins) into fragments with 5 nt overhangs (Kim et al. 2020; Ramachandran et al. 2020), which are optimal for integration into CRISPR locus (Nuñez et al. 2015). In some CRISPR systems, trimming can be performed by Cas4, a RecB nuclease domain‐containing protein (Lee et al. 2018). Either way, the length of the prespacer DNA is maintained at the fixed length by the distance between two Cas1 subunits flanking Cas2 dimer, ensuring a uniform length of spacers within the CRISPR array. Finally, the 3’ ends are positioned into the active sites of Cas1, making them poised for catalysis at CRISPR array.
While the interactions between Cas1:Cas2 complex and protospacer DNA are through phosphate backbone rather than base‐specific interactions, the selection of specific protospacers by the adaptation machinery is often nonrandom. Spacers are preferentially acquired from sequences proximal to PAM sites (Savitskaya et al. 2013), with the adaptation complex in E. coli type I system able to select functional prespacers by directly recognizing the PAM sequence (Datsenko et al. 2012; Swarts et al. 2012; Wang et al. 2015). Similarly, the aforementioned Cas4 also seems to have a key role in selecting prespacers (Rollie et al. 2018). Furthermore, recognizing the PAM sequence is also used to orient the prespacer into CRISPR array so that the crRNA ultimately contains the correct sequence necessary for recognition (Shiimori et al. 2018).
Crucially, while PAM is used to identify functional prespacers and position them in the correct orientation, it must be removed prior to integration into CRISPR array; otherwise, it will induce self‐immunity. This is achieved by removing the PAM sequence immediately prior to integration, either by the exonucleases (Ramachandran et al. 2020) or Cas4 (Lee et al. 2019). By coupling recognition of functional PAM sites with the prespacer processing increases the chance of integrating a functional spacer that will be compatible with future interference phases.
The integration into the CRISPR array occurs more frequently at the 5’ end of the locus, adjacent to the AT‐rich leader sequence (Figure 3.6d). To perform an efficient and on‐target integration, a major distortion of the target DNA is required. In E. coli, this is performed by the architectural protein integration host factor (IHF), which tightly binds to the leader sequence of the CRISPR array (Nunez et al. 2016), inducing a major bend in the DNA (Wright et al. 2017). Active integration complex binds to the bent DNA, catalyzing the nucleophilic attack of the 3’OH group of the prespacer DNA at each end of the repeat sequence in the CRISPR array (Nuñez et al. 2015), resulting in a gapped intermediate that is ultimately resolved by the host’s DNA repair and transcription machinery (Ivančić‐Baće et al. 2015; Budhathoki et al. 2020). In species lacking IHF (most of Gram‐positive bacteria), recognition of the first repeat instead requires a short 5 base pair‐long motif termed leader‐anchoring site (LAS) (McGinn and Marraffini 2016), which is directly recognized by the adaptation complex.
Figure 3.6 Adaptation phase(s) in CRISPR system. Depending on whether invading genomes have been encounter before, spacers can be acquired without (naïve) and with (primed) the assistance of CRISPR interference machinery. Fragments of the invading DNA generated by the RecBCD/AadAB restriction machinery can be captured by the Cas1:Cas2 adaptation complex (a). If a host is reinfected by previously encountered genetic element, interference machinery can recognize the pathogen and digest its genome. Successful targeting by type II effectors (b) leads to recruitment of Cas1:Cas2:Csn2 complex which acquires prespacer sequences proximal to the double‐strand break. Type I systems (c) recruit Cas1:Cas2 tetramer when efficient on‐target cleavage is inhibited, followed by helicase Cas3, allowing the adaptation complex to acquire prespacers from the nontarget strand. In any of the cases, the mechanism of integration into the CRISPR array is similar (d). Prespacer ends are processed and then integrated by two nucleophilic attacks by the 3’OH end at the first repeat sequence, with the gapped DNA filled by the host machinery. Sites of DNA cleavage activity are depicted by orange arrows.
In order to rapidly respond to the ever‐changing gallery of infectious predators, CRISPR systems have evolved a further adaptation pathway that allows rapid acquisition of new spacers through the action of already existing CRISPR‐mediated immunity. This process is termed primed spacer acquisition or priming, and is at least 1000‐fold more efficient than naïve adaptation (Staals et al. 2016; Stringer et al. 2020). A reinfecting strain against which the host has already been immunized will be efficiently targeted and cleaved, allowing new spacers from proximal sequences to be acquired, updating the CRISPR array with more spacers. This mechanism is particularly important in preventing infection by new phage strains that arise by mutations in the target sequence, be it PAM or the protospacer, which would make CRISPR‐targeting ineffective (van Houte et al. 2016).
Priming mechanistically differs from naïve adaptation in early stages, i.e. the origin of prespacer DNA. While in naïve adaptation the prespacers originate from intermediates of nucleolytic cleavage at compromised replication forks, in primed adaptation they are generated by the effector complex. The simplest mechanism is observed in type II systems (Figure 3.6b), where primed adaptation initiates with Cas9‐crRNA‐tracrRNA complexes recognizing and then cleaving the invading genome’s protospacer. Generated double‐stranded breaks allow recruitment of the adaptation complex Cas1:Cas2:Csn2 which together with Cas9 leads to nucleolytic degradation of the bound DNA, albeit many mechanistic details are still missing (Wilkinson et al. 2019). The degradation of the invading DNA leads to the acquisition of spacers from both strands of DNA, and most frequently proximal to the DSB (Nussenzweig et al. 2019). Here, the priming is absolutely dependent on on‐target cleavage by Cas9, and intuitively mutations in the seed regions or PAM abrogate spacer acquisition. As a result of this mechanism, the efficiency of spacer acquisition directly depends on the cutting efficiency mediated by a pre‐acquired spacer. This means that once immunized, the host can rapidly acquire new spacers, forming a feed‐forward immunization response.
In contrast, type I priming is promoted by failure to recognize a perfect target. Mutations in the protospacer will nominally decrease cleavage efficiency; however, Cascade can bind to such targets nonetheless (Blosser et al. 2015). Failure to establish a full‐length R‐loop, in essence failing to recognize a previously encountered target, is thought to abrogate potential off‐target cleavage and instead promote Cascade complex into primed adaptation mode (Xue et al. 2016). This is mechanistically achieved by delaying the recruitment of Cas3 nuclease (Figure 3.6c), permitting recruitment of Cas1:Cas2 complex first (Redding et al. 2015). Upon subsequent binding, Cas1:Cas2:Cas3 complex is able
