The CRISPR-Cas RNA-guided nucleases, Cas9 and Cas12, of class II CRISPR-Cas systems provide bacteria and archaea with acquired immunity using guide RNAs encoded by CRISPR loci to recognize and degrade nucleic acids of invaders such as plasmids and viruses.1, 2, 3, 4 These natural defense systems have been successfully repurposed for modifying genes across a wide range of cells and organisms, including bacteria, yeast, plants and animals, and for the development of molecular diagnostics.5, 6, 7

The Streptococcus pyogenes CRISPR-Cas9 system (SpCas9) has been the most widely used genome-editing tool.5, 6, 7 Paradoxically, however, multiple studies have raised concerns about unintended on-target mutations, genomic rearrangements, chromosome loss, and activation of p53-dependent apoptotic pathway following on-target CRISPR-Cas9-mediated genome editing in mammalian cells,8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 yet the mechanistic basis underlying these deleterious errors remain undescribed. Although engineered SpCas9 variants display improved target specificity, they still induce target and off-target mutations.20, 21, 22, 23, 24, 25 In part, this could be due to the ability of SpCas9 to tolerate some level of gRNA:DNA mispairing.9, 11 By comparison, Cas12a (previously designated as Cpf1) exhibits several notable differences, including size, simplicity and relatively high target selectivity.5 Besides, Cas12a is highly sensitive to mismatches between the target DNA and crRNA, requires a shorter guide RNA, and exhibits genome editing with reduced off-target activity in human cell lines.25, 26, 27, 28 Furthermore, Cas9 effectors employ two guide RNA and dgRNA (or a single fused crRNA and tracrRNA) molecules to recognize and cleave protospacer adjacent-motifs (PAMs), including purine-rich, pyrimidine-rich, and mixed purine and pyrimidine PAMs,29 while a single RNA molecule is sufficient for Cas12a effectors to bind and cleave T-rich PAM sequences.30, 31 Thus, the CRISPR-Cas12a system has emerged as a powerful tool for genome editing in a wide variety of cells.5, 25, 30, 31, 32 However, a better understanding of the substrate specificity of CRISPR-Cas12a effector may pave the path towards developing approaches to mitigate its low-frequency off-target events.26, 27, 29, 31

Whereas a full discussion of the adaptive immune system is outside the scope of this work, we note that structure-function studies of Cas9 and Cas12a effectors have provided valuable insights into the mechanism of crRNA-guided target DNA cleavage.33, 34, 35 Briefly, central to the function of Cas9 and Cas12a nucleases is their capacity to inflict a double-strand break (DSB) in the target DNA. While Cas9 cleaves the two DNA strands with HNH and RuvC nuclease domains,36, 37, 38 Cas12a uses a single RuvC nuclease domain to cleave sequentially the non-target strand followed by target strand.39, 40, 41, 42 As such, the RuvC-like nuclease domain in both Cas9 and Cas12a effectors is assembled from three discontinuous RuvC-like segments (RuvC-I to RuvC-III).33, 34, 35 Unlike Cas9 nuclease, which produces either blunt-ended or 1–2 bp staggered DSB ends,43, 44 Cas12a nuclease generates staggered DSBs with 5-8 nucleotide 5' overhangs.26, 30, 33, 34, 35 Although crRNA-guided, PAM-dependent target DNA degradation by Cas9 and Cas12a nucleases has long been appreciated, compelling evidence suggests that Cas9 orthologs catalyze robust, nonspecific cleavage of single-stranded DNA (ssDNA) in a crRNA-dependent manner, but independently of both PAM and tracrRNA.45, 46, 47 Furthermore, Cas12a orthologs exhibit crRNA-independent DNA nicking/cleavage activity.48, 49, 50 This is akin to S. pyogenes and Francisella tularensis novicida Cas9 endonucleases, suggesting that both crRNA and PAM are dispensable for non-specific degradation of ssDNA and double-stranded DNA (dsDNA).48 Yet, several questions remain - for example, how Cas9 and Cas12a nucleases discriminate between different DNA structures is particularly noteworthy.

Despite tremendous advances in our understanding of structure-function relationships of CRISPR-Cas9 and CRISPR-Cas12a nucleases,33, 34, 35 considerably less is known about how they facilitate DNA degradation to nucleotides. It is generally thought that adaptive immunity hinges on the RNA-guided DNA endonucleolytic activity of CRISPR-Cas12a; however, it is unclear whether this enzymatic mechanism alone is sufficient to degrade invading genetic elements into nucleotides. Moreover, recent focus on CRISPR-Cas12a system as an alternative genome engineering tool presents a compelling reason for in-depth understanding of its substrate specificity. In this setting, we report that Acidaminococcus sp. apo-Cas12a (hereafter denoted AsCas12a) has a broad range of substrates, binds preferentially to branched DNA structures, independently of crRNA and divalent cations. While seeking to elucidate the mechanism, we found that AsCas12a binds to the Holliday junction (HJ), specifically at the crossover region, distorts the junction structure prior to its resolution into nonligatable intermediates and, subsequently, their complete degradation in a Mn2+-dependent manner. We further demonstrate that crRNA impedes HJ cleavage by AsCas12a, and that of Lachnospiraceae bacterium (LbCas12a), without affecting their DNA-binding capacity. Of significance, we present robust evidence that AsCas12a endonuclease also exhibits 3'-to-5' and 5'-to-3' exonuclease activity. Collectively, our findings provide unanticipated insights into the substrate specificity of AsCas12a and important insights into the degradation of different types of DNA substrates.