A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection

Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic-acid editing and detection Ylenia Jabalera1,8, Igor Tascón2,3,8, Sara Samperio1, Jorge P. López-Alonso3,4, Monika Gonzalez-Lopez1, Ana M. Aransay1,5, Guillermo Abascal-Palacios2,3, Chase L. Beisel6,7, Iban Ubarretxena-Belandia2,3* & Raul Perez-Jimenez1,2* 1CICbioGUNE, Basque Research & Technology Alliance (BRTA), Bizkaia Technology Park, Building 800, 48160 Derio, Bizkaia, Spain. 2Ikerbasque Foundation for Science, Bilbao, Spain. 3Instituto Biofisika (UPV/EHU, CSIC), University of the Basque Country, Leioa, Spain. 4Basque Resource for Electron Microscopy, Leioa, Spain. 5CIBERehd, ISCIII, 28029 Madrid, Spain. 6Helmholtz Institute for RNA-based Infection Research, Helmholtz Centre for Infection Research, Würzburg, Germany. 7Medical Faculty, University of Würzburg, Würzburg, Germany 8These authors contributed equally to the work. *Corresponding authors: ivan.ubarrechen[email protected] (I. U-B) and [email protected] (R.P-J). Summary Cas12a nucleases offer powerful tools for genome editing and molecular diagnostics, yet their existing properties constrict the range of accessible targets and applications. Here, we apply ancestral sequence reconstruction (ASR) to a set of Cas12a orthologs from hydrobacteria to reconstruct a common ancestor (ReChb) that yields near-PAMless targeting and the recognition of diverse nucleic-acid activators and collateral substrates. ReChb shares 53% sequence identity with the closest Cas12a ortholog; however, ReChb no longer requires a T-rich PAM and can achieve genome editing in human cells at sites inaccessible to the natural FnCas12a or the engineered and PAM-flexible EnAsCas12a. Furthermore, ReChb can be triggered not only by double-stranded DNA but also by single-stranded RNA and DNA targets, which leads to nonspecific collateral cleavage of all three nucleic-acid substrates with similar efficiencies. Finally, tertiary and quaternary structures of ReChb obtained by cryo-EM reveal the molecular details underlying its expanded biophysical activities. Overall, ReChb expands the application space of Cas12a nucleases and underscores the potential of ASR for enhancing CRISPR technologies. Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Introduction Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) and their CRISPRassociated (Cas) nucleases confer bacteria and archaea with adaptive immunity against foreign mobile genetic elements1-4. These CRISPR-Cas systems rely on the ability of the nucleases to use a CRISPR RNA (crRNA) as a guide, with the additional requirement of a non-self sequence (i.e., protospaceradjacent motif (PAM) in DNA, protospacer-flanking sequence (PFS) in RNA) flanking the complementary nucleic-acid target. Target recognition then enacts a myriad of immune activities, from cutting target nucleic acids to inducing cellular dormancy or cell death through widespread RNA or DNA degradation. The adeptness of Cas nucleases at RNA-guided targeting of nucleic acids has catapulted their repurposing for genome editing and diagnostics among many applications5. Founded on Cas9 from Streptococcus pyogenes (SpCas9) 6,7, the catalogue of Cas nucleases has been expanded with the discovery of variants of the compact CRISPR class 2 system, including orthologs of Cas9, Cas12 and Cas13 nucleases2,5,8-11. In particular, Cas12a nucleases offer a promising alternative to Cas9 due to relying on T-rich PAMs, generating staggered double-strand breaks in their DNA target, and using a short crRNA with no need for a transactivating-crRNA (tracrRNA). Also, Cas12a nucleases can process their own crRNA lending to multiplexing and exert fewer off-targets 12,13. Finally, activated Cas12a collaterally cleave single-stranded (ss)DNA, which has been used for signal amplification as part of CRISPR-based molecular diagnostics14. Despite the successful implementation of Cas nucleases as technologies, these natural counterparts and ideal technological components remain far apart9. To bridge this gap, molecular engineering has been the basis for the development of improved CRISPR tools with expanded capabilities15-18. In the case of Cas12a, engineered versions relying on a few mutations to natural nucleases such as Cas12a from Acidaminococcus sp (AsCas12a), Lachnospiraceae bacterium (LbCas2a) and Francisella novicida (FnCas12a) have broadened PAM preferences17,19-21. Other mutations have been reported to improve editing efficiencies22, and efforts have been made to expand substrate recognition, such as the recognition of RNA substrates by the PAM distal end of the crRNA23. Nevertheless, even these engineered forms of Cas12a still exhibit properties that at most partially deviate from their natural counterparts. To overcome these limitations, we adopted a distinct approach based on Ancestral Sequence Reconstruction (ASR). Using ASR, we previously resurrected active Cas9 nucleases that no longer exist in nature15. These proteins, derived from type II-A Cas9 orthologs from Clostridia and Bacilli classes, retraced the evolution of Cas9 up until modern Streptococcus pyogenes Cas9 and displayed PAMflexible target recognition, gRNA and substrate promiscuity, and even collateral cleavage. While the ancient Cas9 orthologs lost the ability to cleave dsDNA and exhibited limited genome editing activity as dual nickases, the orthologs suggested that ancestral reconstruction could unlock expanded properties of functions of other Cas nucleases. Here, we apply ASR to engineer a fully functional, resurrected Cas nuclease derived from orthologs of Cas12a from hydrobacteria phyla14,24. The resulting resurrected Cas12a ancestor, which we call ReChb, yielded PAM-flexible editing in human cells across three cell lines, seven genes and 16 target sites that outperformed natural and engineered Cas12a nucleases EnAsCas12a17 and could both target and collaterally cleave RNA, ssDNA and dsDNA. ReChb thus offers a unique and versatile tool for a multitude of CRISPR-based biotechnological applications, from gene therapy to molecular diagnostics. Results Applying Ancestral Sequence Resurrection of Cas12a nucleases to generate ReChb To reconstruct the sequence of ReChb using ASR, we first collected sixty-three homologs of Cas12a sequences from species belonging to the phyla Pseudomonadota, Planctomycetota, Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Spirochaetota, Bacteroidota, belonging to the clade hydrobacteria, and Bacillota (as outgroup) (Fig. 1a and Extended Data Fig. 1). A phylogeny using Bayesian inference was constructed following previously described methods15. The internal node representing a common ancestor of hydrobacteria (excluding the outgroup) was selected for resurrection, resulting in an amino-acid sequence with 53% identity shared with the Francisella novicida Cas12a (FnCas12a)24 and a posterior probability average of 0.92. This ancestor is estimated to have diverged around 3,000 Mya, CI: 2978.5-3030.4 Mya (ref.25). The alignment of the reconstructed ReChb against representative sequences from the other phyla shows substantial differences in the REC1 domain (Fig. 1b) with a K-rich region unique to ReChb. Also, of note is the presence of a tryptophan (W1041) in a highly conserved region of the RuvC-II domain (Fig. 1b), which is conserved in uncharacterized Cas12a sequences from Planctomycetota and Spirochaetota but is absent in the other orthologs (Supplementary Fig. 1). The reconstructed sequence was synthesized and expressed in Escherichia coli with higher expression yields than FnCas12a (Supplementary Fig. 2). ReChb efficiently edits human cells FnCas12a, the query sequence for ReChb, has been shown to display variable genome editing activity in human cells12,24,26, while the ancestral Cas9 nucleases did not exhibit any measurable editing activity when targeting a single site. To evaluate the editing activity of ReChb, we quantified the formation of insertions or deletions (indels) via the generation of dsDNA breaks in HEK293 cells using crRNA guides from FnCas12a targeting sites flanked by a canonical 5′-TTTV-3′ PAM in the genes DNMT1 (5′-TTTC-3′), AAVS1 (5′-TTTG-3′) and EMX1 (5′-TTTG-3′). After transiently transfecting ReChb and crRNA expression constructs, we employed a T7 endonuclease mismatch assay to quantify indel formation. Compared to FnCas12a, which only exhibited detectable indel formation (1%) at AAVS1, ReChb displayed robust indel formation (15-35%) against the target in each of the three genes (Extended Data Fig. 2a-b). Considering that only a limited number of type II and V orthologs have been successfully employed for genome editing in mammalian cells, these results highlight the potential of ReChb as a gene-editing tool. ReChb exhibits robust PAM-flexible editing in human cells Given the PAM flexibility of the ancestral version of SpCas9 (ref.15), we next explored the PAM requirements for ReChb. We applied an in vitro cleavage assay in which purified ReChb loaded with an in vitro-transcribed crRNA complex was incubated with a linear DNA target flanked seven random nucleotides. The PAM-containing cut DNA fragment was then gel purified and submitted for amplicon sequencing (Fig. 2a). The resulting normalized read counts revealed a modest preference for C/T at the –2 and –3 positions (Fig. 2b,c). However, all sequences were identified in the gel-extracted DNA (Source of Data 1), suggesting that the resurrected nuclease has PAM biases but no obvious requirements. This PAM-flexible activity was verified in vitro using both the non-T-rich PAM sequences 5′-ATTG-3′, 5′-ATTT-3′, 5′-ACCC-3′, 5′-ATCA-3′ and 5′-ATGG-3′ on a supercoiled plasmid, as well as 5′-TAAA-3′, 5′-TCCC-3′, and 5′-TGGG-3′ on a FAM-labelled dsDNA substrate (Supplementary Fig. 3). Given the parallels to the PAM flexibility of the ancestral form of SpCas915, flexible PAM recognition could be a general feature of resurrected Cas nucleases. After demonstrating the ability of ReChb to target non-canonical sites in vitro, we evaluated its targeting flexibility in human cells in comparison to the engineered Cas12a variant enAsCas12a (ref.17). enAsCas12a nuclease was selected because it has been associated with the broadest PAM recognition of Cas12a nucleases to-date, with recognition of canonical numerous non-canonical PAMs17. To provide a more comprehensive analysis, we targeted 13 sites in four human genes (DNMT1, CFTR, FANC and RUNX) and analysed the frequency of indels by Sanger-based indel detection methodology, which provides precise results for indels with a frequency above 1-2%27. These experiments yielded robust and editing activities (ranges 30-37%) at two target sites with a canonical PAM (5′-TTTC-3′, 5′-TTTG- Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 3′) as well as two target sites with T/C-rich PAM (5′-TCCT-3′, 5′-TCCC-3′) PAMs (ranges 22-29%), but also at sites with PAMs not recognized efficiently by enAsCas12a such as 5′-TCGG-3′ (40%), 5′- TATT-3′ (38%), 5′-AGTT-3′ (17%) or 5′-TGTT-3′ (17%) (Fig. 2d). At four sites with a 5′-TTAC-3′, 5′- TCGC-3′, 5′-CGGT-3′ or 5′-TAAC-3′ PAM, both nucleases yielded poor indel frequencies (< 5%), either due to limited PAM recognition or other factors such as poor sgRNA performance or site accessibility. Across the other sites, ReChb efficiently edited those targeted by enAsCas12a and, more importantly, demonstrated robust genome editing at sites not recognized by enAsCas12a (Fig. 2e). These data show that ReChb enables robust editing of sites with non-canonical PAMs, with a tested PAM preference towards 5′-NYYN-3′, 5′-NRYN-3′ and 5′-NYRN-3′ in eukaryotic cells, extending beyond those recognized by enAsCas12a. Additionally, ReChb could drive indel formation when transfected as a ribonucleoprotein (RNP) complex in two other cell lines (HeLa and Hs27 fibroblast cells) (Supplementary Fig. 4). Across the four target sites in four genes (DNMT1, SOD1, FANCF, CFTR) representing both canonical and non-canonical PAMs, ReChb showed robust editing between 10-60%. Finally, off-target analysis of five on-target sites revealed that the indel frequencies at the predicted/potential off-target sites28 were below 1% (Source of Data 2), much lower than those at corresponding on-target sites. Thus, ReChb exhibits PAM-flexible editing in human cells that outperforms enAnCas12a in a head-to-head comparison. ReChb accepts RNA, ssDNA and dsDNA targets and collateral substrates as well as altered crRNAs Given the broadened PAM recognition of ReChb as one biochemical feature of this nuclease, we explored whether it can expand beyond dsDNA targets and ssDNA collateral substrates traditionally recognized by Cas12a nucleases. Beginning with target recognition, we found that ReChb cleaved complementary dsDNA (Fig. 3a), ssDNA (Fig. 3b) and ssRNA (Fig. 3c) targets containing the reverse complement of the signature T-rich PAM sequence of canonical Cas12a substrates under in vitro conditions24. While the dsDNA target encoded within a supercoiled plasmid was cleaved into a linearized form, the ssDNA and RNA targets were completely degraded. In contrast, FnCas12a completely cleaved the dsDNA and ssDNA targets and marginally cleaved the RNA target under the same conditions (Extended Data Fig. 3), in line with prior work14. We next turned to collateral substrate cleavage, now with the knowledge that ReChb could be activated by dsDNA, ssDNA or ssRNA. For all activating substrates, ReChb degraded non-target dsDNA (Fig. 3d), ssDNA (Fig. 3e), and ssRNA (Fig. 3f). FnCas12a degraded non-target ssDNA, but only in the presence of ssDNA and dsDNA targets (Extended Data Fig. 3). Interestingly, the closely related Cas12a2 (that shares 26% of sequence identity with ReChb) nucleases recognize RNA but not ssDNA targets and collaterally cleave ssRNA, ssDNA and dsDNA29,30, suggesting that the ancestrally resurrected nuclease captures the targeting properties of both nucleases despite Cas12a and Cas12a2 respectively specializing in DNA and ssRNA target recognition and ASR only drawing from the Cas12a branch. Critically, the flexible targeting properties exhibited by ReChb extend beyond those of any individual characterized Cas nuclease found in modern organisms14,30-32. Based on the observed collateral cleavage activity of ReChb, we assessed ReChb as a tool for molecular diagnostics. Cas12a has been extensively used to for the detection of specific dsDNA and ssDNA sequences, with cleavage of ssDNA beacons for signal production33. However, RNA remains a poor substrate for Cas12a, requiring workarounds such as supplying a partial dsDNA target23. To explore the potential of ReChb, we incubated a ReChb-crRNA RNP complex with a dsDNA, ssDNA or ssRNA target as well as short ssRNA (Fig. 3h) and ssDNA (Fig. 3i) beacons fusing a fluorophore and quencher (Fig. 3g). All combinations of targets and molecular beacons resulted in increased fluorescence compared to no target. In contrast, an FnCas12a-crRNA RNP only yielded increased fluorescence when combining a dsDNA or ssDNA target with the ssDNA molecular beacon. Therefore, Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 ReChb could expand existing capabilities for CRISPR-based molecular diagnostics with a single nuclease capable of accepting different types of nucleic-acid biomarkers and molecular beacons. Traditional CRISPR diagnostics systems based on collateral cleavage by Cas12 nucleases are restricted by the presence of PAM sequence before the target sequence, compromising the universal recognition of any target. Based on PAM-flexible recognition of ReChb, we envisioned that combining the collateral cleavage of ReChb next to its PAM-flexible recognition, we could overcome this main limitation of this technology. To test this hypothesis, we first compared the collateral cleavage of ReChb with those exposed by wild-type LbCas12a14, a widely used Cas12a nuclease for molecular diagnostics33, after the recognition of a dsDNA target with a T-rich PAM (Fig 4a). The activity exhibited by ReChb (0.0045 AU/s) was lower to that of LbCas12a (0.0051 AU/s) under equivalent conditions (Fig. 4b,d) at 37ºC. Reducing the temperature from 37ºC to 25ºC (room temperature, RT, Supplementary Fig. 5) significantly decreases the activity of both nucleases (Fig. 4b). However, at RT, ReChb retains its ability to recognize target substrates at low concentrations (2 pM, Fig. 4c) compared to LbCas12a which shows a higher detection limit (10 nM, Fig. 4c). This is in line with previously reported results, where higher temperatures yielded optimal cleavage activity for Ca12a nucleases34,35. Thus, ReChb exhibits catalytic properties comparable to LbCas12a, a leading nuclease for CRISPR-based diagnostics while also offering PAM flexibility and measurable activity at room temperature well suited for point-of-care molecular diagnostics. Given the ability of ReChb to recognize non-canonical PAMs with similar activities to that of LbCas12a, we applied ReChb in application requiring dsDNA recognition and PAM flexibility: detection of a short (21-bp) variable region of a single amplicon of 16S rDNA36 (Fig. 4e). By applying this approach, it is possible to amplify the target region using universal primers that anneal to the conserved region of 16S rDNA and thus avoid difficulties associated with multiplexed preamplification. To evaluate this approach, we designed four different crRNAs that hybridizes to four variable regions (from hypervariable region V1) of bacterial 16S rDNA with a non-canonical PAM (GTCG). Importantly, this region does not possess a T-rich stretch and thus cannot be recognized by traditional Cas12a-based diagnostics. Specifically, we selected four bacterial pathogens, Acinetobacter baumannii, Escherichia coli, Klebsiella pneumoniae and Staphylococcus aureus, relevant pathogens associated with sepsis in newborns36. Then, we evaluated the collateral cleavage activity for each 16S dsDNA fragment and each crRNA. We found that ReChb triggered fluorescence release only when there was a match between the dsDNA 16S fragment and the crRNA (Fig. 4f). As expected, LbCas12a was not able to recognize any of these DNA targets (Supplementary Fig. 6). We noticed crRNA from K. pneumoniae also triggered fluorescence release when paired with the other three pathogens, likely due to gRNA sequence similarity with seed region as previously reported37. In fact, the mismatch profile analysis of ReChb showed that the protein did not tolerate single or double mismatches at positions 120, only triggering a reduced fluorescence signal to single mismatches at positions 7-20 nt or double mismatches at the end of the PAM distal region (Supplementary Fig. 7). However, this signal was significantly reduced compared to that displayed by ReChb with a target dsDNA substrate. These results demonstrate the specificity and flexibility of ReChb for molecular diagnostics. Given ReChb’s flexibility in target and collateral substrate recognition, we reasoned that ReChb could also exhibit flexibility in crRNA recognition. To determine the crRNA scaffold recognition determinants of ReChb, in vitro dsDNA cleavage assays were performed by incubating the enzyme with crRNAs from various CRISPR types and Cas12a species (Extended Data Fig. 4a-c). ReChb linearized plasmid DNA using targeting crRNAs from not only different Cas12a nucleases but also from Cas9 and Cas12j. The use of crRNAs with distinct repeats from different types and subtypes is not associated with modern CRISPR-Cas systems and at most was observed with the ancestral Cas915. ReChb could also process its own precursor (pre-)crRNA comprising multiple crRNAs in the same transcript (Extended Data Fig. 4d). The pattern of cleavage products (65 nts) against a single customized pre- Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 crRNA transcript13 with four repeats matched with the expected length obtained from cutting the 5′ end of a direct repeat hairpins24. Thus, ReChb can process its own crRNA guides similar to type V effector nucleases38,39, and it can share crRNAs with different types of Cas nucleases. This capability of ReChb should be relevant for multiplexed genome editing applications that employ customized CRISPR arrays for multi-site targeting13. Structure of ReChb ternary complex To shed light on the mechanism behind the flexibility of PAM and substrate recognition exerted by ReChb, we solved the cryo-EM structure of the nuclease effector bound to a crRNA and different substrates. The 3.1 Å resolution cryo-EM structure (PDB: 8QWE) captures a ternary complex of ReChb bound to crRNA and target dsDNA containing a T-rich PAM (Fig. 5, Supplementary Figs. 8 and 9, Extended Data Table 1). Overall ReChb displays a bilobed architecture resembling the characteristic oval “sea conch” shape described for Cas12a40. The recognition (REC) lobe comprises both REC1 and REC2 domains, and the nuclease (NUC) lobe comprises the PAM-interacting (PI), wedge (WED), RuvC, Nuc and bridge helix (BH) domains (Fig. 5a). The crRNA-target DNA heteroduplex (R-loop, Fig. 5b, Extended Data Fig. 5) threads through the positively charged central channel between the two lobes. The target strand (TS) hybridizes with the crRNA, while the dissociated non-target strand (NTS) traces a straight path through the groove of the DNA nuclease site. There is unambiguous cryo-EM density downstream of the cleaved end of the TS, which we attribute to a post-cleavage product (Figs. 5c,d). As in Cas12a, the nuclease active site of ReChb lies in a pocket at the interface between Nuc and RuvC. The RuvC domain carries the three (D878, E967 and D1216) highly conserved active site residues (Extended Data Fig. 6). In contrast to other Cas12a nucleases31,41,42, the nuclease site of ReChb embraces the dissociated NTS through stacking interactions with W1041 and F971 residues. The tertiary structure also presented an opportunity to understand the basis of flexible PAM recognition. Both WED and PI domains grip the PAM region of the target dsDNA in a manner analogous to FnCas12a but with notable differences (Fig. 5e,f)40. FnCas12a predominantly recognizes the PAM through multiple interactions between K671 and the minor-groove edge of the PAM duplex, employing a mechanism of base and shape readout from both the minor groove and major groove41. Notably, in the case of FnCas12a, only K671 interacts with the bases in the PAM region, indicating that shape may play a more pivotal role than direct base readout in PAM recognition40,41. In contrast, ReChb achieves altered PAM recognition through newly formed electrostatic interactions between residues K124, K125, and K130 from the REC1 domain with the phosphate backbone of the PAM bases through the minor groove. A similar mechanism has already been described for engineered versions of AsCas12 (RVR and RR), which recognized TATV and TYCV, respectively20,43. The structure of these engineered forms revealed that altered PAM recognition by these variants relies on newly formed interactions between substituted residues and the altered PAM-complementary nucleotides41,43. Consequently, these findings could provide a structural framework for the flexible PAM readout exhibited by ReChb, reinforcing the notion that amino acid substitutions strengthening the PAM-binding channel could contribute to the alteration of Cas12a nucleases PAM specificity. Continuing with the catalytic cycle of ReChb, Cas12a structural studies have suggested a reaction cycle coupled to conformational changes40,42,44,45. The cycle begins with an apo state at equilibrium between closed and open conformations, where crRNA binding shifts the equilibrium to the closed form and target dsDNA causes a further structural tuning that includes the rearrangement of the bridge helix and helix 1 of the RuvC II sub-domain40,42,44,45. These latter changes in RuvC are absent in the cryo-EM structure of the ReChb ternary complex, likely because the structure represents a post-cleavage form, and the nuclease might have reverted to a non-activated state. We note that the 3.3 Å resolution structure of apo ReChb (Supplementary Note 1, Supplementary Figs. 10 and 11, Extended Table 1) without a crRNA displays a ring-like architecture distinct from the closed oval “sea conch” shape of the ternary Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 complex (Extended Data Fig. 7) and from the open and closed conformations suggested for Cas12a in the apo state. The REC2 domain closes the ring by connecting with the RuvC domain in a way that hinders access to the nuclease active site (Supplementary Fig. 11), and the central channel between the recognition and nuclease lobes, where the crRNA-target DNA heteroduplex threads in the ternary complex, is not formed. Structure of ReChb quaternary complex The 3.0 Å resolution cryo-EM structure (PDB: 8QWF) (Fig. 6, Supplementary Figs. 8 and 9, Extended Data Table 1) of a quaternary complex of ReChb bound to crRNA, target dsDNA containing a T-rich PAM, and a non-specific collateral dsDNA with both strands contain non-hydrolysable phosphorothioate modifications (Fig. 6a), provides a framework to understand its collateral nuclease activity. As shown for Cas12a nucleases31, after cleavage, the target strand (TS) remains hybridized to the crRNA. However, the cryo-EM density for most of the dissociated non-target strand (NTS) is absent, indicating its displacement from the ternary complex pathway (Fig. 6b and 6c), maintaining the ReChbcrRNA complex in the catalytically activated conformation. Specifically, the structure captures RuvC in a pre-hydrolysis state, during which the non-hydrolysable collateral substrate, visible only at low density thresholds, is attempting to enter its active site (Fig. 6d). This weak density for the collateral substrate is consistent with a constant motion of the substrates, aligning with biochemical data showing dsDNA degradation occurring through multiple-turnover DNA nicking (Fig. 3d). Therefore, the conformational arrangements of the complex and the PAM-distal target DNA displacement increase accessibility to the RuvC active site, enabling rapid substrate capture and cleavage in trans as measured for ReChb (Fig. 3). In fact, with the notable exception of Cas12a2, other natural Cas12 nucleases have difficulties collaterally cleaving dsDNA substrates at short incubation times29,31,46. This limitation has been attributed to the RuvC active site not being able to accommodate duplex DNA properly. In addition, the Nuc domain might also act as a physical barrier limiting cleavage in trans29. Analogously to Cas12a229, ReChb can nick, linearize and degrade supercoiled non-specific plasmid DNA. Using FnCas12a as a reference, structures of FnCas12a post-cleavage state (PDB: 5MGA)40 reveal that the presence of the crRNA-TS complex hinders the path to the catalytic site creating a narrow pathway to reach the RuvC active site, which is incompatible with rigid dsDNA substrates (Extended Data Fig. 8). This contrasts with the accessible ReChb RuvC active site in the post-cleavage state, offering a structural basis for the wide range of collateral substrates degraded by ReChb (Extended Data Fig. 8). In fact, this mechanism resembles that of Cas12a2, where the absence of Nuc domain and the presence of an exposed RuvC active site afford high substrate accessibility, enabling the cleavage a wide range of substrates in trans29. Overall, the structural determination together with the biochemical assays highlight the versatility of ReChb for substrate recognition as well as nucleic-acid editing. Discussion Protein design techniques using computational methods offer an opportunity to improve and even design catalysts with properties not found in natural enzymes47. The increasing number of protein sequences in databases and the advent of new methods for sequence alterations and design, including deep-learning methods and language models48-50, have the potential to expand the universe of biomolecules and revolutionize the fields of biotechnology and synthetic biology. ASR is rapidly gaining prominence in this context, as it not only provides important evolutionary information, but also is able to generate novel protein sequences not found today51-55. ASR is arguably the only technique that can handle non-natural sequences with many residues (>1,000) and with substantial identity alterations relative to natural proteins. We previously established that ASR can be utilized to uncover important aspects of the evolution of CRISPR-Cas systems15. What we present here goes a step further, demonstrating that it is possible Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 to obtain a nuclease with a set of properties that have not been found yet in any existing natural Cas nuclease, surpassing the functionality and versatility of any known Cas nuclease. As shown before15, ancestral proteins display a functional promiscuity that is clearly present in ReChb perhaps as its most prominent feature. ReChb can recognize any nucleic-acid form as a target and is able to carry out specific or non-specific cis/trans cleavage. This can be accomplished without the need for the recognition of a strict PAM and with a variety of crRNA-guide sequences. ReChb can process its own crRNA and is proficient at genome editing in human cells at a wide range of sites. Another aspect relevant to note is that the promiscuity of ReChb towards nucleic acids could lead to cell toxicity when it is applied as genome editing tool. In this case, as observed with other Cas12a nucleases56, no signs of cell toxicity have been detected during cell experiments. This suggests that these promiscuous collateral activities are likely only observed in vitro and do not negatively impact the protein's on-target activity in vivo. Although ReChb is capable of collateral cleavage of ssRNA, our hypothesis is that, unlike Cas13, ReChb binds to its DNA target, which limits its ability to freely diffuse and affect surrounding RNA. The cryo-EM structures provide a framework to understand the differences between synthetic ReChb and natural Cas nucleases, and for further finetuning as a biotechnology tool. Importantly, all of these new features have been readily achieved using ASR, unlike other protein design techniques based on rational design where only one feature is improved at a time. This is clearly exemplified, for example, in the effort made to achieve relaxed recognition of PAM using protein engineering techniques such as that performed for enAsCas12a rather than the ease of ASR. On the other hand, in terms of evolution, the different promiscuous features found in the ancestors of Cas915 and Cas12 reveal that the exact connections between them are unknown and hidden, and surely interconnected mechanisms are yet to be discovered. In practical terms, the versatility of ReChb makes it unique and versatile, as it can be used for both genome editing and diagnostics based on nucleic-acid detection. In recent years, assays based on the collateral activity of Cas12a and Cas13 have been developed as diagnostic tools33. Cas12a and Cas13 can only be efficiently activated by DNA and RNA, and thus depending on the sequence to be detected one enzyme or the other normally must be used. ReChb is not limited in this way, as it can be activated by any nucleic acid and is thus able to identify any type of genetic target. In addition, we demonstrate that the PAM-flexible recognition ability of ReChb allows overcoming one of the main limitations of Cas12-based diagnostic, i.e., recognition of dsDNA targets without the PAM restriction. To the best of our knowledge, ReChb is the only reported nuclease with such expanded activities (Supplementary Table 1). A nuclease with these features might well exist in nature, but searching the natural enzyme space could prove arduous, costly and time-consuming. Similarly, machine learning-based computational methods are still in their infancy and are not yet capable of designing enzymes with complex and controlled functions such as large-scale conformational changes. The recent developed deep-learning methods (OpenCRISPR-157), while promising, have not yet demonstrated the ability to design proteins with new functionalities. These limitations highlight the importance of ASR to generate complex synthetic enzymes with multiple and improved properties opening new avenues for its inclusion and combination with deep-learning and language methods. Methods Engineering of ReChb sequence ReChb was engineered through computational methodologies, specifically, ancestral sequence resurrection (ASR)15. We use the BLAST tool with custom parameters and criteria—that is, a maximum of 1,000 hits and minimum identity 35% to ensure the selection of Cas12a sequences (individually inspected) and BLOSUM62 scoring matrix. E-values were virtually zero for all sequences. Sixty-three sequences were selected following similar proportions of sequences in each phylum as in the database Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 (Extended Figure 1). Sequences belong to five bacterial phyla: Pseudomonadota, Planctomycetota, Spirochaetota, Bacteroidota and Bacillota, as outgroup. Alignment of sequences was performed using MUSCLE software on the MEGA platform and manually edited to eliminate gaps, poorly aligned sites and divergent regions. We inferred the best evolutionary model using MEGA, resulting in the JTT with gamma distribution model (eight categories), Yule model for speciation and length chain of 100 million generations, sampling every 1,000 generations. Phylogeny was carried out using BEAST v.2.6.6 package software (https://beast.community/) including the BEAGLE library for parallel processing and based on Bayesian inference using MCMC. Divergence times were estimated using the Reltime method58 implemented in MEGA with discrete eight-category Gamma distribution for evolutionary rates. We set calibration times using information from the TToL25,59 in three major clades with 95% confidence interval (CI). Finally, ancestral sequence reconstruction was performed by maximum likelihood using PAML v.4.9 (http://abacus.gene.ucl.ac.uk/software/paml.html) with a gamma distribution of eight categories for variable replacement rates across sites. Posterior probabilities were calculated for all amino acids, and the residue with highest posterior probability was chosen for each site. The reconstructed sequence displays average posterior probability of 0.92 and amino acid sequence identities of 52 % with respect to FnCas12a and 26% with respect to SuCas12a2. In vitro characterization Expression and purification of ReChb. The ReChb gene was synthesized and codon-optimized for E. coli cell expression. ReChb was cloned in pET-28a(+) expression vector and transformed in E. coli BL21 (DE3) (Life Technologies) for protein expression. Cells were incubated in LB medium at 30 °C at 160 rpm until OD600 reached 0.6 and IPTG (400 µM) was added for protein induction overnight at 18 ºC. Cells were pelleted by centrifugation at 5000 g. Pellets were resuspended in extraction buffer (Tris 50 mM, NaCl 500 mM, pH 8, Imidazole 10 mM) supplemented with EDTA Free protease inhibitor (Thermo). Then, the pellet was sonicated for 3 cycles for 10 min at 30% amplitude. Cell debris was separated by ultracentrifugation at 33,000 G for 1 h. For purification, the supernatants were mixed with Ni-NTA agarose beads (Thermo) and incubated for 1 hour. Then, the beads were washed with 50 x column volumes with 40 mM imidazole. After that, imidazole concentration was reduced to 10 mM and aliquot of protease 3C was added. After 1 hour of incubation, protein was eluted in elution buffer (TrisHCl 50 mM, NaCl 500 mM, pH 8, Imidazole 10 mM). The protein was further purified by size exclusion chromatography using a Superose 6 10/300 GL column (GE Healthcare) and eluted in 50 mM Tris pH 7.5, 150 mM NaCl, 2 mM MgCl2. For protein purification verification, sodium dodecyl sulphate– polyacrylamide gel electrophoresis (SDS–PAGE) was used with 8% gels. The protein concentration was calculated by measuring the absorbance at 280 nm in Nanodrop 2000C. crRNA synthesis. crRNA (Supplementary Table 2) was synthesized using a HiScribe T7 High Yield RNA Synthesis Kit (NEB). The DNA sequences includes the T7 promoter at the 5′ end and the sequence from crRNA with the target sequence at the 3′ end. ssDNA oligos were hybridized and the reaction was incubated overnight. Then the crRNA was purified following the protocol of the Monarch RNA Purification Columns Kit. Nucleic acid cleavage assays. For analysis of targeted cleavage (Supplementary Table 3) on supercoiled DNA, 30 µL reactions of 170 nM of ReChb-crRNA and 100 nM of dsDNA target in 1 x NEB 3.1 buffer (50 mM Tris-HCl pH 7.9, 100 mM NaCl, 10 mM MgCl2, 100 µg ml−1 BSA) were incubated at 37 ºC for varying incubation times. In the case of ssDNA substrate, 30 µL reactions of 80 nM of ReChbcrRNA and 30 nM of ssDNA target in 1 x NEB 3.1 buffer were incubated at 37 ºC for varying incubation times. Reactions were stopped by adding 6X loading dye (NEB) with EDTA and running 1-2% agarose gel. Gels were dyed with SYBR gold (ThermoFisher) and imaged with ChemiDoc XRS + System (BioRad). Cleavage was quantified by ImageJ. Finally, for ssRNA targeted cleavage, 30 µL reactions of 250 nM of ReChb-crRNA and 120 nM of ssRNA target in 1 x NEB 3.1 buffer were incubated at 37 ºC for Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 39 Swarts, D. C., van der Oost, J. & Jinek, M. Structural Basis for Guide RNA Processing and SeedDependent DNA Targeting by CRISPR-Cas12a. Mol Cell 66, 221-233 e224, doi:10.1016/j.molcel.2017.03.016 (2017). 40 Stella, S., Alcon, P. & Montoya, G. Structure of the Cpf1 endonuclease R-loop complex after target DNA cleavage. Nature 546, 559-563, doi:10.1038/nature22398 (2017). 41 Yamano, T. et al. Crystal Structure of Cpf1 in Complex with Guide RNA and Target DNA. Cell 165, 949-962, doi:10.1016/j.cell.2016.04.003 (2016). 42 Stella, S. et al. Conformational Activation Promotes CRISPR-Cas12a Catalysis and Resetting of the Endonuclease Activity. Cell 175, 1856-1871 e1821, doi:10.1016/j.cell.2018.10.045 (2018). 43 Nishimasu, H. et al. Structural Basis for the Altered PAM Recognition by Engineered CRISPRCpf1. Mol Cell 67, 139-147 e132, doi:10.1016/j.molcel.2017.04.019 (2017). 44 Strohkendl, I., Moy, C., Nguyen, A.-H., Russell, R. & Taylor, D. W. Structural basis of Cas12a Rloop propagation on pathway to DNA cleavage. bioRxiv, 2023.2003.2013.532460, doi:10.1101/2023.03.13.532460 (2023). 45 Worle, E., Newman, A., D'Silva, J., Burgio, G. & Grohmann, D. Allosteric activation of CRISPRCas12a requires the concerted movement of the bridge helix and helix 1 of the RuvC II domain. Nucleic Acids Res 50, 10153-10168, doi:10.1093/nar/gkac767 (2022). 46 Murugan, K., Seetharam, A. S., Severin, A. J. & Sashital, D. G. CRISPR-Cas12a has widespread off-target and dsDNA-nicking effects. J Biol Chem 295, 5538-5553, doi:10.1074/jbc.RA120.012933 (2020). 47 Hanreich, S., Bonandi, E. & Drienovska, I. Design of Artificial Enzymes: Insights into Protein Scaffolds. Chembiochem 24, e202200566, doi:10.1002/cbic.202200566 (2023). 48 Yeh, A. H. et al. De novo design of luciferases using deep learning. Nature 614, 774-780, doi:10.1038/s41586-023-05696-3 (2023). 49 Lovelock, S. L. et al. The road to fully programmable protein catalysis. Nature 606, 49-58, doi:10.1038/s41586-022-04456-z (2022). 50 Madani, A. et al. Large language models generate functional protein sequences across diverse families. Nat Biotechnol 41, 1099-1106, doi:10.1038/s41587-022-01618-2 (2023). 51 Manteca, A. et al. Mechanochemical evolution of the giant muscle protein titin as inferred from resurrected proteins. Nat Struct Mol Biol 24, 652-657, doi:10.1038/nsmb.3426 (2017). 52 Perez-Jimenez, R. et al. Single-molecule paleoenzymology probes the chemistry of resurrected enzymes. Nat Struct Mol Biol 18, 592-596, doi:10.1038/nsmb.2020 (2011). 53 Zakas, P. M. et al. Enhancing the pharmaceutical properties of protein drugs by ancestral sequence reconstruction. Nat Biotechnol 35, 35-37, doi:10.1038/nbt.3677 (2017). 54 Risso, V. A., Gavira, J. A., Mejia-Carmona, D. F., Gaucher, E. A. & Sanchez-Ruiz, J. M. Hyperstability and substrate promiscuity in laboratory resurrections of Precambrian betalactamases. J Am Chem Soc 135, 2899-2902, doi:10.1021/ja311630a (2013). 55 Risso, V. A. et al. De novo active sites for resurrected Precambrian enzymes. Nat Commun 8, 16113, doi:10.1038/ncomms16113 (2017). 56 Marino, N. D., Pinilla-Redondo, R. & Bondy-Denomy, J. CRISPR-Cas12a targeting of ssDNA plays no detectable role in immunity. Nucleic Acids Res 50, 6414-6422, doi:10.1093/nar/gkac462 (2022). 57 Ruffolo, J. A. et al. Design of highly functional genome editors by modeling the universe of CRISPR-Cas sequences. doi:10.1101/2024.04.22.590591 (2024). 58 Tamura, K. et al. Estimating divergence times in large molecular phylogenies. Proc Natl Acad Sci U S A 109, 19333-19338, doi:10.1073/pnas.1213199109 (2012). 59 Kumar, S., Stecher, G., Suleski, M. & Hedges, S. B. TimeTree: A Resource for Timelines, Timetrees, and Divergence Times. Mol Biol Evol 34, 1812-1819, doi:10.1093/molbev/msx116 (2017). Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 60 Leenay, R. T. et al. Identifying and Visualizing Functional PAM Diversity across CRISPR-Cas Systems. Mol Cell 62, 137-147, doi:10.1016/j.molcel.2016.02.031 (2016). 61 Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Meth. 14, 290-296, doi:10.1038/nmeth.4169 (2017). 62 Jamali, K. et al. Automated model building and protein identification in cryo-EM maps. bioRxiv, doi:10.1101/2023.05.16.541002 (2023). 63 Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr 60, 2126-2132, doi:10.1107/S0907444904019158 (2004). 64 Afonine, P. V. et al. Real-space refinement in PHENIX for cryo-EM and crystallography. Acta Crystallogr D Struct Biol 74, 531-544, doi:10.1107/S2059798318006551 (2018). Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Figures Figure 1. Ancestral sequence resurrection of Cas12a yields ReChb. (a) Phylogenetic tree of sequences used for ASR of ReChb. (b) The domain architecture of ReChb. REC, recognition; PI, protospacer adjacent motif (PAM) interacting; WED, wedge; BH, bridge helix; Nuc, nuclease. Both the WED and RuvC domains are formed by three discontinuous segments of the protein sequence. (Top) Aligned sequences associated with ReChb and representative Cas12a nucleases from REC1. (Bottom) Aligned sequences associated with ReChb and representative Cas12a nucleases from RuvC-II domain. Highlighted sequences indicated conserved positions mutated in ReChb. Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Figure 2. ReChb exhibits broad PAM recognition activity in human cells. (a) Scheme of in vitro determination of PAM preference using a substrate library encoding random 7 nt PAMs with NGS. (b) 7nt-Weblogo and (c) 3-nt PAM wheels (Krona plots) recognized by ReChb. (d) On-target activities of ReChb and enAsCas12a across sites encoding canonical and non-canonical PAM sequences, quantified using TIDE (ref.27). Error bars represent the mean ± SEM, where n = 3. (e) Assessment of the gene editing activities of ReChb and enAsCas12a on-target sites harbouring canonical and non-canonical PAM sequences (n = 13). NS (non-significant differences), P > 0.05; *, P < 0.05 (two-tailed Welch’s t- Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 tests). NS difference for those target sites with PAMs recognized by enAsCas12 (i.e., 5′-TTTC-3′, 5′- TTTG-3′, 5′-TCCT-3′, 5′-TCCC-3′, 5′-TTAC-3′) indicated as empty symbols. Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Figure 3. Characterization ReChb activity in vitro. (a) In vitro DSB activity of ReChb against a 4,007 bp supercoiled plasmid DNA as a function of time. Exponential fit of the data expressed as % of cleaved product. (b) In vitro activity of ReChb against M13 phage derived ssDNA (M13 phage) as a function of time. Exponential fit of the data expressed as % of cleaved product. (c) In vitro activity of α -synCas against ssRNA as a function of time. Exponential fit of the data expressed as % of cleaved product. ReChb exhibits collateral activity against non-target nucleic acids in vitro. Time-course of collateral activity against non-target (d) dsDNA, (e) ssDNA, (f) ssRNA by an ReChb/crRNA complex activated with target dsDNA, ssDNA, and ssRNA substrates. Please refer to Extended Data Fig. 3 for the same characterization using FnCas12a as nuclease. (g) Nucleic acid detection assays. Quantification of maximum fluorescence signal generated after incubating ReChb-crRNA-activator with a custom (h) ssRNA-beacon or (i) ssDNA beacon for 2 hours and 30 min at 37ºC, respectively. For FnCas12a sample, nuclease was target activated with a dsDNA substrate. The negative control was prepared using nuclease-free water instead of the nucleic acid target activator. Error bars represent the mean ± SD, where n = 3. NS, P > 0.05; ***P < 0.001 (two-tailed Welch’s t-tests). Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Figure 4. ReChb as molecular diagnostic tool for PAM-free dsDNA target recognition. (a) Schematic of in vitro collateral cleavage assay for dsDNA recognition by ReChb. The ReChb-gRNA complex recognizes and cleavage the dsDNA target, which triggers non-specific collateral cleavage on ssDNA. (b) Sensitivity of ReChb and LbCas12a using a dsDNA with canonical PAM (5′-TTTC-3′) as target substrate and ssDNA beacon as collateral substrate. Velocities were obtained by linear regression analysis of the linear regions of the progress curves. (c) Velocities for dsDNA target recognition and posterior collateral cleavage of ssDNA beacon by ReChb and LbCas12a at RT. (d) Velocities for dsDNA target recognition and posterior collateral cleavage of ssDNA beacon by ReChb and LbCas12a at 37ºC. Error bars represent the mean ± SD, where n = 3. NS: not significant, P > 0.05; ***P < 0.001 (twotailed Welch’s t-tests). (e) 16S dsDNA from four different pathogens including both conserved and variable region used for dsDNA target detection with non-canonical PAM sequence (5′-GTCG-3′). (f) Specific detection of pathogens based on 16S rDNA with ReChb and each specific crRNA using universal primers for specific amplification of variable region (yellow, V1) using conserved regions of 16S rDNA (blue). Ab: Acinetobacter baumannii; Ec: Escherichia coli; Kp: Klebsiella pneumoniae; Sa: Staphylococcus aureus. Error bars represent the mean ± SD, where n = 3. Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Figure 5. Ternary complex between ReChb, crRNA and target dsDNA. (a) Domain organization of ReChb. (b) Sequences of the crRNA guide and target strand and non-target strand DNA used to form the ternary complex. Nucleotides with coloured background are visible in the cryo-EM map, while uncoloured nucleotides are disordered. (c) Unsharpened cryo-EM map of the ternary complex coloured by nucleic acid and protein domain. Unassigned density attributed to the cleaved target strand is highlighted. (d) Cryo-EM structure of the ReChb ternary complex coloured and oriented as in c. A circle marks the putative location of the cleaved target strand. (e) Close-up view of the PAM region of the dsDNA highlighting the interactions with selected residues, shown as sticks, of the PI, WED, and REC1 domains. (f) Schematic of the interactions of target and non-target DNA strands of the PAM with FnCas12a (left) and ReChb (right). Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Figure 6. Quaternary complex between ReChb, crRNA, target dsDNA and collateral dsDNA. (a) Sequences of the crRNA guide, target dsDNA and collateral dsDNA used to form the quaternary complex. Nucleotides with coloured background are visible in the cryo-EM map, while uncoloured nucleotides are disordered. The central region of the non-hydrolysable collateral dsDNA includes the following phosphothiates: F, A-Phosphorothioate; O, C-Phosphorothioate; E, G-Phosphorothioate; Z, T-Phosphorothioate. (b) Unsharpened cryo-EM map obtained of the quaternary complex coloured by nucleic acid and protein domains. Unassigned density attributed to part of the collateral dsDNA is highlighted. (c) Cryo-EM structure of the quaternary complex coloured and oriented as in b. An oval mark the putative location of the collateral dsDNA. (d) Close-up views of front (left) and back (right) of the -synCas active site showing the cryo-EM density attributed to the collateral dsDNA fitted with an idealised four-nucleotide dsDNA fragment. Continuous to the collateral dsDNA density is the density for Trp 1041 shown in transparent green. Selected active site and lid-loop residues are shown as sticks. Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Extended Figure 7. Cryo-EM structure of apo ReChb. Comparison of apo (top) and ternary complex (bottom) structures of ReChb coloured by nucleic acid and protein domain. Arrows depict the direction of movement of the REC1 and REC2 domains upon RNA/DNA binding. Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Extended Figure 8. Post-cleavage conformation of ReChb (left) and FnCas12a (right; PDB: 5MGA). Nucleotides are coloured as in Figure 5b. Jabalera et al. Jabalera, Y., Tascón, I., Samperio, S. et al. A resurrected ancestor of Cas12a expands target access and substrate recognition for nucleic acid editing and detection. Nat Biotechnol 43, 1663–1672 (2025). https://doi.org/10.1038/s41587-024-02461-3 Extended Table 1. Cryo-EM data collection, refinement and validation statistics. Apo ReChb (EMD-18691) (PDB 8QWD) Ternary complex ReChb -crRNAtarget dsDNA (EMD-18693) (PDB 8QWE) Quaternary complex ReChb -crRNAtarget dsDNAcollateral dsDNA (EMD-18694) (PDB 8QWF) Data collection and processing Magnification 105000 105000 105000 Voltage (kV) 300 300 300 Electron exposure (e–/Å2) 50.979 50.979 50.084 Defocus range (μm) 1-1.6 1-1.6 1-1.6 Pixel size (Å) 0.8238 0.8238 0.8238 Symmetry imposed C1 C1 C1 Initial particle images (no.) 3324256 5463714 7589433 Final particle images (no.) 254316 137913 300509 Map resolution (Å) FSC threshold 3.33 0.143 3.14 0.143 3.08 0.143 Map resolution range (Å) 2.9-6.5 2.9-6.5 2.9-6.5 Refinement Initial model used (PDB code) - - - Model resolution (Å) FSC threshold 3.3 0.143 3.1 0.143 3.0 0.143 Model resolution range (Å) 2.9-6.5 2.9-6.5 2.9-6.5 Map sharpening B factor (Å2) -137.6 -103.2 -124.0 Model composition Non-hydrogen atoms Protein residues Ligands 9204 1103 2 11713 1185 5 11499 1186 6 B factors (Å2) Protein Nucleotide Ligand 127.16 - 69.59 101.79 93.27 28.99 31.38 35.04 21.31 R.m.s. deviations Bond lengths (Å) Bond angles (°) 0.002 0.507 0.005 0.645 0.002 0.499 Validation MolProbity score Clashscore Poor rotamers (%) 1.79 8.29 0.20 1.63 6.94 0.00 1.69 7.59 0.00 Ramachandran plot Favored (%) Allowed (%) Disallowed (%) 95.02 4.98 0.00 96.33 3.67 0.00 95.99 4.01 0.00