Split Cas12a Protospacer Engineering Enables Ultraspecific Recognition of a PAM-Less DNA Strand via an External PAM-Proximal Duplex

Abstract

CRISPR-Cas12a is a programmable RNA-guided endonuclease whose activation by double-stranded DNA normally depends on the recognition of an adjacent protospacer-adjacent motif (PAM), which constrains target selection in diagnostic applications. Here, we divide the DNA activator into PAM-proximal (Pp) and PAM-distal (Pd) duplexes, spatially separating the PAM requirement from the interrogation of a PAM-less target sequence. FRET, electrophoretic mobility shift, fluorescence polarization, and functional assays indicate that either fragment can associate with Cas12a, but that robust productive activation requires the PAM-containing Pp duplex together with the Pd duplex. Our data further support a model in which PAM-mediated protein–DNA contacts stabilize an assembly-competent complex and facilitate binding of the PAM-less Pd target capture. Weakening hybridization between the guide and Pp with a chimeric DNA/RNA increased the dependence on PAM-mediated stabilization and improved structural selectivity in the Pp region and single-nucleotide specificity in the Pd region. The resulting split-duplex architecture provides an ultraspecific platform for resolving single-nucleotide polymorphisms in the PAM-less DNA strand.

Introduction

Clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein 12 (Cas12) is primarily an RNA-guided, DNA-targeted endonuclease. Like all CRISPR proteins, it is derived from the acquired immune system of prokaryotes. It evolved to record previous infections by integrating sections of the parasitic, exogenous genetic factors (spacers) from invading bacteriophages or plasmids into the bacteria’s genome. These genetic factors are approximately 20 nucleotides in length and are known as protospacers.

In the case of Cas12a, these protospacers are adjacent to T-rich protospacer-adjacent motifs (PAMs) which are recognized by an RNA-guided Cas12a and used to unwind the double-stranded target DNA. , Given that the PAM (5′-TTTV-3′) is required to unwind the first bases of a double-stranded activator sequence, Cas12a does not require a PAM when its activator is single-stranded. , Once the first bases are unwound, Cas12a begins hybridizing with the first ∼5 bases in its crRNA’s programmable section, which are structured in an open and solvent-exposed conformation in a process known as R-loop initiation (Figure S1A,B).

Successful annealing leads to R-loop elongation, a process in which the crRNA displaces the nontarget strand (NTS) in the double-stranded target DNA and anneals to the target strand (TS) (Figure S1C). Once the 20th base in the crRNA’s programmable region anneals to the TS, the Cas12a caps the formation of the R-loop by inserting a large aromatic residue, effectively signaling that the Cas12a has reached its activation state (Figure S1D). , Once in this state, the nuclease site in the RuvC domain is exposed, allowing it to cleave the NTS, which is followed by cleavage of the TS by the same nuclease site, in a process referred to as cis-cleavage. This is followed by the release of the PAM-distal section of the DNA. At this time, Cas12a remains bound to the PAM-proximal DNA, allowing the protein to remain in its active state wherein its nuclease site is exposed to the solvent. This allows suspended ssDNA to enter the nuclease site, resulting in its indiscriminate cleavage in a process known as trans-cleavage.

Given the ease with which the crRNA can be reprogrammed to target new genes, Cas12a has been used with great success in gene editing, genetic screening, gene circuits, transcription regulation, intracellular imaging, and diagnostics. ,

These unique properties have been leveraged to create a new category of low-burden nucleic acid amplification tests , which have significant advantages but have had some long-standing challenges. For example, the requirement for a PAM to unwind dsDNA limits the scope of target sequences. In addition, Cas12a exhibits low fidelity for single-nucleotide polymorphisms, reducing its target specificity. , Furthermore, as a primarily DNA-targeting enzyme, Cas12a generally demonstrates reduced activity on RNA targets, impacting its ability to diagnose this important biomarker. ,

Protein engineering has addressed some of these limitations by either removing or replacing residues that interact with the PAM or stabilize the R-loop, leading to PAM-relaxed and higher-specificity variants, respectively. ,,, Similar results were achieved by modifying the nucleic acid substrates. Modifying the activator to be fully or partially single-stranded has alleviated PAM requirements, , while installing artificial mutations in the guide has improved on-target resolution. ,

Cas12a-based RNA detection has been improved by replacing the crRNA with a novel “CRISPR DNA” to target an RNA- rather than DNA-encoded activator. In other instances, the crRNA has been split such that a miRNA in the solution assembles with the Cas12a to target a preprogrammed DNA activator, signaling the presence of the aforementioned miRNA. , Alternatively, other approaches have modified the activator sequence, retaining a DNA molecule in the PAM-proximal portion, and target RNA in the PAM-distal portion. ,,

The structure of the complementary activator sequence has been modified to include either a “nick” or a “break”, effectively creating a “split activator”. In this scheme, we are left with the PAM-proximal (Pp) and PAM-distal (Pd) activator sequences (Figure A). ,− By maintaining a DNA-based Pp, a single-stranded RNA or RNA/DNA heteroduplex can be targeted in the Pd, greatly increasing Cas12a’s activity in direct RNA detection. ,, However, despite these remarkable advances, the specificity of split activator systems for single-nucleotide polymorphisms is generally only slightly improved compared to the unmodified CRISPR/Cas12a system. , Furthermore, the majority of studies to date have focused on targeting single-stranded DNA and RNA, ,,,− with relatively little attention having been placed on double-stranded DNA, , despite the importance of this format in diagnostic and therapeutic applications.

1.

Position of the double-stranded break modulates Cas12a activity. (A) Cas12a activators are composed of the target (20nt) protospacer, which anneals to the crRNA, and the protospacer-adjacent motif (PAM), which is required to unwind the duplex. Introducing a double-stranded break in the activator yields PAM-proximal and PAM-distal sequences. Breaks were introduced in the ten bases adjacent to the PAM. Trans-cleavage activity of Cas12a was measured over time when providing (B) PAM-proximal (Pp) duplexes only, (C) PAM-distal (Pd) duplexes only, and (D) both the Pp and Pd in combination. Conditions were benchmarked against an intact activator. Background-subtracted fluorescence intensity (n = 3) was measured over 60 min at intervals of 3.6 min. Values represent total area under the curve (RFU.min) ± standard error. (E) Cas12a can associate with PAM-less protospacers (Pd = 20nt) provided at least the last of the first five bases adjacent to the PAM are present (Pd > 15nt). These five bases correspond to the “seed” region in the crRNA, which are preordered and solvent-exposed to initiate hybridization to the target strand. Graphics were created in https://BioRender.com.

It is through this lens that we set out to better understand the assembly of the split Cas12a systems. We began by first analyzing the response of Cas12a to split duplexes of different lengths, allowing us to focus our next steps on designing a system that provides strong activity in the presence of both duplexes and no activity in the absence of either. We then built on a previously proposed assembly model by including new conditions, integrating our findings into the original model, and showing key interactions that had not been previously elucidated. Furthermore, while previous work with split activators suggested that chimeric DNA/RNA guides (crXNA) could not be used in split Cas12a systems, we leveraged our improved split activator length optimizations and structural evidence for the interaction between the guide and the Cas12a to rationally design a new crXNA-guided system. The resulting design was then leveraged to study the function of the PAM in mediating the assembly of the intermediate ternary complex. Finally, using this selectively destabilized ternary complex, we demonstrate an improved base resolution of Cas12a for single-nucleotide polymorphisms, achieving ultraspecific signaling in almost every position of the PAM-less DNA strand. Combined, we anticipate that our optimization for split duplex lengths and the proposed mechanism for the assembly of split activators will help to guide future developments in this field, with implications for diagnostics, logic operations, and DNA nanotechnologies.

Experimental Section

Nucleic Acid Design and Preparation

ssDNA and guide RNAs (crRNA and crXNA) were synthesized by Integrated DNA Technologies (IDT, Coralville, IA, USA). ssDNA from IDT was resuspended in IDT Nuclease-free Duplex Buffer. PAM-proximal (Pp) sequences were annealed with the target strand (TS) and the nontarget strand (NTS) at a molar ratio of 1:1 unless otherwise indicated (50 μM TS, 50 μM NTS). PAM-distal (target sequences) were annealed with TS and NTS at a molar ratio of 1:1.2 (40 μM TS, 48 μM NTS). To anneal them, samples were brought to 98 °C for 3 min, then cooled to 20 °C over 20 min and kept at 4 °C until used. Annealed oligonucleotides were diluted in IDT Nuclease-free Duplex Buffer with a TS concentration of 1 μM and 125 nM for the Pp and Pd, respectively.

In Vitro Trans-Cleavage Assays

Unless otherwise indicated, trans-cleavage assays were performed by incubating 33 μL of a solution composed of 1× NEBuffer r2.1, 1 U/μL Murine RNase inhibitor, 75 nM LbaCas12a, 75 nM guide RNA (crRNA or crXNA), and 45 nM Pp at 37 °C for 30 min. Then, 17 μL of a second solution composed of 1× NEBuffer r2.1, 1 U/μL Murine RNase inhibitor, 375 nM DNase ALERT (IDT), and 30 nM Pd dsDNA was added to the first for 50 μL total. The final master mix was composed of 1× NEBuffer r2.1, 1 U/μL Murine RNase inhibitor, 125 nM DNase ALERT, 30 nM Pp, 50 nM LbaCas12a, 50 nM guide RNA, and 10 nM Pd. 15 μL of this master mix was aliquoted at 4 °C into three wells of a 384-well opaque, black plate for three technical replicates (n = 3). Plates were sealed with a sticky lid, centrifuged at 2000g for 1 min at 4 °C, and read on a TECAN Infinite M1000 with λ_ex: 536 nm and λ_em: 556 nm. Plates were incubated at 37 °C in the TECAN and read over the course of 90–120 min.

Background-subtracted fluorescence intensity was calculated by subtracting the average background fluorescence of a control (with no target sequences) from the raw fluorescence values of the samples of interest. Background controls and samples of interest were prepared at n = 3. Plotted values represent the mean background-subtracted fluorescence ± their standard deviation.

WT-normalized fluorescence calculated from background-subtracted values (Figures , S20, S27, S28, and S30) for mutated targets was calculated by dividing individual background-subtracted fluorescence values of the mutated samples (n = 3) by the mean background-subtracted fluorescence value of the perfectly matched target (WT) (n = 3). Plotted values represent the mean WT-normalized fluorescence ± standard deviation.

5.

Activation of intact vs split activator-based Cas12a reaction schemes on single-nucleotide polymorphisms (SNPs) in the Pd. An intact 20nt protospacer was used to activate (A) crRNA-guided (Gray) and (B) crXNA-guided (Blue) Cas12a. This was compared to a split activator scheme with a Pp/Pd split at 6nt/14nt. When using split activators, the Cas12a was guided by either (C) a crRNA (Blue) or (D) a crXNA (Green). Values represent mean WT-normalized, background-subtracted fluorescence for each target relative to the perfectly matched activator (mutant/WT) ± SD at the indicated times. All experiments were performed at n = 3. Graphics were created in https://BioRender.com.

FRET Assays

The viability of some of the fluorescence resonance energy transfer (FRET) assays was first predicted by simulating the complexes we anticipated forming (Figure S2). Simulations were run on the AlphaFold3 server (https://alphafoldserver.com/) and included sequences listed in the Supporting Information. Structures were deemed viable if the distance between the DNA/RNA extremities of all simulations was less than 10 nm (100 Å). Fluorescent oligonucleotides were synthesized by IDT by adding fluorescein (FAM) to their 5′ extremities. Quenching oligonucleotides were prepared by adding IDT’s proprietary Iowa Black quencher to 3′ extremities.

FRET experiments were performed by preparing a solution comprising 1× NEBuffer r2.1, 1 U/μL Murine RNase inhibitor, 50 nM LbaCas12a, 50 nM synthetic crRNA, 30 nM Pp, and 10 nM Pd. Reaction mixtures were prepared at 68 μL, and three aliquots of 20 μL were added to an untreated black clear-bottom 384-well plate. Plates were incubated at 37 °C in the TECAN M1000 Pro and read with an excitation wavelength of 495 nm and an emission wavelength of 520 nm.

In proximity-format assays (Figure A–G), fluorescence changes were interpreted as changes in the proximity of labeled components and therefore as evidence for compatible assembly states. In cleavage-format assays (Figure H,I), increased fluorescence was interpreted as cleavage-associated separation of the fluorophore and quencher within the labeled Pd duplex. These assays were used to compare productive and nonproductive conditions, but they were not intended to resolve a unique order of assembly or map cleavage sites directly.

2.

Assembly of split protospacer duplexes. To model the association of the RNP (Cas12a + crRNA) to the Pp and Pd when both are required to achieve trans-cleavage, the Pp/Pd split was placed at 6nt/14nt. (A, B) Assembly of Pp-F and Pd-Q. (C) Pp can associate with the RNP, but not the crRNA alone. The strength of association is based on (D) the structure and (E) the sequence of the PAM. Three structural configurations were compared, with the Pp being (I) fully double-stranded; (II) double-stranded PAM, single-stranded TS; and (III) fully single-stranded. (F, G) Pd can associate weakly with the RNP. (H, I) Cleavage of the Pd requires the Pp and is influenced by the sequence of the PAM. FRET readouts are interpreted here as proximity changes between labeled partners and were used together with EMSA, fluorescence polarization, and cleavage assays to infer assembly and productive activation. All samples (n = 3) were incubated over 30 min at 37 °C. Values represent mean fluorescence intensity ± SD. Analyses were performed using a one-way ANOVA with a multiple comparison test, where ns = not significant with p > 0.05 and asterisks (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, and ****p ≤ 0.0001) denote significant differences. Graphics were created in https://BioRender.com.

Fluorescence Polarization Assay

Fluorescence polarization measurements were performed to assess binding of LbaCas12a to fluorescently labeled probes (FAM or Cy5). Recombinant LbaCas12a was prepared in assay buffer (PBS, 10 mM MgCl₂, 0.01% Triton) and serially diluted to the indicated concentrations. Next, fluorescently labeled probes were added to each well to achieve a final concentration of 5 nM. Binding experiments were assembled in a total volume of 20 μL in black, low-binding 96-well plates, and incubated for 30 min at room temperature prior to measurement. Fluorescence polarization was measured using a plate reader equipped for detection (FAM; excitation 485 nm, emission 528 nm, gain = 120 and Cy5; excitation 620 nm, emission 680 nm, gain = 120). Polarization values were recorded in millipolarization (mP) units. Each condition was measured in duplicate or triplicate as indicated. Data were analyzed by plotting polarization as a function of LbaCas12a concentration, and binding curves were generated using GraphPad Prism 10.

EMSA Binding Assay

The EMSA binding assay was performed by preparing a solution composed of 1× NEBuffer r2.1, 1 U/μL Murine RNase inhibitor, 500 nM LbaCas12a, 500 nM synthetic crRNA, and 200 nM each of Pp and/or Pd. Reactions were incubated at 37 °C for 1 h before being placed at 4 °C. 1/6 volume of EMSA loading buffer (1× NEBuffer r2.1, 50% glycerol) was added to the samples. Samples were run on a 5% PAGE gel prepared with 0.5× TBE and 5 mM MgCl₂, with a running buffer composed of 0.5× TBE and 5 mM MgCl₂. Samples were run at 4 °C in the dark.

Results

Splitting a Double-Stranded Protospacer

We begin by investigating how split duplexes impact the activation of Lachnospiraceae bacterium ND2006 Cas12a (herein termed Cas12a). Here, a double-stranded break was introduced immediately following the PAM and extended base by base through the first 10 bases of the protospacer (Figure A) and separated into two parts: the PAM-proximal (Pp) and PAM-distal (Pd). When combined, the Pp (left) and Pd (right) pairs encoded the complete 20nt protospacer. As expected, Pp activators alone failed to initiate trans-cleavage by Cas12a, as their hybridizing sequence was less than the critical minimum of 15 nucleotides required to activate Cas12a (Figure B).

Next, we evaluated whether the PAM, which is used by Cas12a to melt the first bases in the protospacer, was dispensable when targeting truncated DNA duplexes. Here, a PAM-less protospacer complementary to the full crRNA spacer (Pd = 20nt) was placed on the extremity of a double-stranded activator sequence (Figures C and S3B). Under these conditions, the complementary PAM-less Pd activator rapidly activated Cas12a, as measured by the complete trans-cleavage of the reporter. Following recent literature, we hypothesize that DNA fraying, the transient opening of DNA termini, provided access to short ssDNA segments complementary to the crRNA (Figure S1E). , This process would be roughly analogous to the uncoupling of the activator’s first base pairs by Cas12a in the presence of the PAM (Figure S1B). Normally, this uncoupling facilitates the hybridization between the crRNA’s first ∼5nt, which are preordered and solvent-exposed (Figure S1A), and the melted ssDNA immediately adjacent to the PAM (Figure S1B,E). By monitoring trans-cleavage, we found that at least one of these five bases needs to be present in the Pd duplex (Pd > 15nt) for the crRNA to associate to the rest of the duplex, supporting the hypothesis (Figure C,E).

Having established that Cas12a could tolerate serial truncations in its activator (Pd), up to the loss of the seed region, we next investigated the effects of introducing double-stranded breaks in the activator. Here, a Pp and a Pd, which collectively encoded both the PAM and the full 20nt protospacer, were exposed to a crRNA-guided Cas12a, otherwise known as a ribonucleoprotein (RNP). Interestingly, Cas12a’s trans-cleavage activity, which was diminished when exposed to a truncated DNA duplex (Pd = 16nt) (yellow bars, Figure C), could be improved when the missing section of the activator sequence was provided in the form of a Pp (4nt) (green bars) (Figures D and S3F). This trend was even more interesting for Pd = 15nt-13nt, where there is no trans-cleavage activity without both Pp and Pd components (green bars) (Figures D and S3G–I). This effect was not universal and could not be leveraged to significantly restore Cas12a’s activity at all positions (Figure S3J–L). Importantly, the finding that Pp and Pd are both required to induce Cas12a’s activity left open the question as to how the truncated activators associate with Cas12a.

Assembly of the Split Protospacer

To probe how the split activators interact with Cas12a during complex formation, we tracked proximity changes among labeled nucleic acid components using Fluorescence Resonance Energy Transfer (FRET) assays. In these experiments, changes in fluorescence were interpreted as evidence of altered proximity between tagged partners and therefore as assembly-compatible interactions. Because these measurements do not by themselves establish a unique temporal order of assembly or indicate whether a detected intermediate is productive, we interpreted the FRET data together with fluorescence polarization, EMSA, and functional cleavage assays. In short, we labeled the DNA duplexes and the crRNA with a Fluorescein (F) and a Quencher (Q), then measured for changes in the sample’s fluorescence. These changes in FRET allowed us to monitor assembly or dissociation of the nucleic acids and Cas12a, depending on the nature of the experiment. To represent the assembly of double-stranded split activator systems, the DNA break was introduced immediately following the sixth base of the protospacer. When combined, the Pp/Pd split at 6bp/14bp completely restored activity in Cas12a but led to no notable activity in the absence of either (Figures B–D and S3H).

Since blunt-end DNA duplexes can undergo end-to-end stacking in certain conditions, we first tested whether the split protospacer fragments could assemble independent of the crRNA and Cas12a. To test this, the Pp and Pd duplexes were labeled with FAM and the quencher, respectively, so that, upon interaction, a reduction in fluorescence would be observed (Figure A). We found that the baseline fluorescence of the Pp-FAM (Pp-F) was not quenched in the presence of a Pd-Quencher (Pd-Q), indicating that they did not undergo spontaneous end-to-end stacking (Figure B). Further, neither the crRNA with a sequence complementary to Pp and Pd nor Cas12a, which can unwind DNA duplexes, could assemble the duplexes individually. Rather, Pp:Pd assembly required all four components (Pp, Pd, crRNA, and Cas12a), as can be seen with a reduction in fluorescence (Figure B).

Next, we sought to determine which, if either, of the split activators (Pp or Pd) would preferentially associate with Cas12a during the formation of the active complex. Since Cas12a first interacts directly with the PAM to melt the first bases of a canonical target sequence (Figure S1B), we reasoned that the Pp:RNP complex should be favored. To track the formation of this complex, here, we labeled the Pp with a fluorophore and the crRNA with a quencher (Figure C, reduced signal indicates assembly). In the absence of Cas12a, the crRNA-Q did not quench the Pp-F regardless of whether the Pp was single-stranded, indicating that the two nucleic acid structures did not hybridize independently (Figure D). In the presence of Cas12a, the Pp was quenched, indicating that the protein was required to associate with the two (Pp + crRNA).

Interestingly, signal reduction by Cas12a was influenced by the sequence and, to a lesser extent, the structure of the PAM. A strong quenching effect was observed when a canonical PAM was present (5′-TTTG-3′), indicative of assembly (Figure D). In contrast, when the PAM sequence was replaced with an Anti-PAM sequence (5′-GAGA-3′) (Figure E, reduced signal indicates assembly), the fully double-stranded Pp-F did not associate with the crRNA-Q (Configuration I, Figure E), implying that the crRNA-Q could not anneal to the target strand. In contrast, the crRNA-Q could assemble with the partially and fully single-stranded Pp-F Anti-PAM (GAGA) (Figure E, Configurations II and III), implying that the crRNA-Q could anneal to the target strand only when the latter was single-stranded. Taken together, the evidence suggests that Cas12a interacts with the PAM to facilitate annealing by the crRNA. This follows the canonical activation pathway for Cas12a , and indicates that the DNA/protein interactions between the Pp and the Cas12a are important for the initial formation of the Pp:Cas12a:crRNA complex. Furthermore, these interactions are distinct from the subsequent DNA:crRNA interactions.

Recent crystallographic evidence (PDB 8Y04) from a study detailing precursor crRNA processing elucidates the structure of the RNP complexed to a double-stranded Pp with 6nt annealing to the crRNA (Figure S4). This structural evidence supports a theory that associating with the Pp improves the Cas12a’s ability to target the Pd. Indeed, we demonstrated that Cas12a can form productive complexes with target PAM-less Pd sequences, provided the preordered seed (first ∼5nt) in the crRNA can hybridize (Figures C and S3B–G). In this crystal structure (PDB 8Y04), the crRNA bases immediately downstream of the Pp are similarly preordered (bases 6–8) (Figure S5). This would support the observed activation of Cas12a along Pd targets that would otherwise not anneal to the preordered bases (first ∼5nt) in the crRNA. These complementary pieces of evidence suggest that the Pp-bound intermediate complex may preorder the bases in the crRNA, thereby facilitating Pd-target searching.

Having established that the Pp:RNP complex can assemble independently of the Pd, and that this assembly is influenced by the PAM’s structure and sequence, we next explored whether the RNP could assemble with the Pd alone. To this end, a fluorophore and quencher were installed on the TS of the Pd and the crRNA, respectively (Figure F). In this context, fluorescence quenching was indicative of assembly. We found that RNP-Q and Pd–F could associate (Figures G and S6B), despite the lack of significant trans-cleavage under these conditions (Figures C and S3H). To confirm these results, an EMSA binding assay demonstrated that the RNP did indeed bind to the Pd–F without the Pp (Figure S6A).

Importantly, the FRET data support proximity changes consistent with both Pp:RNP and Pd:RNP association but do not by themselves distinguish which route occurs first in all complexes or whether each observed association is on-pathway to activation. To quantitatively evaluate the interactions underlying the assembly pathway, we performed fluorescence polarization assays using representative PAM-proximal (Pp) and PAM-distal (Pd) substrates (Figure S7). For canonical PAM-containing Pp substrates, apparent affinities were strongly influenced by substrate architecture, with K _D values of 410 nM, 33 nM, and 60 nM for fully double-stranded, partially exposed, and single-stranded configurations, respectively. In comparison, anti-PAM controls yielded K _D values of 524 nM, 99 nM, and 33 nM. Notably, the canonical PAM improved binding approximately 3-fold in the partially exposed duplex configuration.

We also observed measurable association of the isolated PAM-distal duplex with Cas12a (K _D ≈ 45 nM), indicating that both Pp and Pd can interact with the protein. Because fluorescence polarization reports total association rather than productive complex formation, we interpret the Pd value as an apparent affinity that may include heterogeneous or nonproductive binding states. This interpretation is consistent with the orthogonal FRET and EMSA experiments, which support measurable Pd:RNP association, while the corresponding functional assays show that Pd alone does not efficiently induce cleavage. We therefore favor a model in which the isolated Pd can transiently sample the RNP, potentially through limited guide:target pairing at an accessible duplex terminus, but does not efficiently progress to the PAM-stabilized activation pathway required for productive cleavage. Together with the EMSA results, these data support assembly routes in which the RNP can encounter either fragment independently. However, Pd:RNP association alone did not produce substantial Pd cis-cleavage or reporter trans-cleavage under our assay conditions, whereas inclusion of Pp supported productive activation. We therefore interpret these data as supporting multiple assembly-compatible encounter states, while distinguishing productive activation from simple association in the working model shown in Figure S5.

Having determined that the crRNA-guided Cas12a associates with both the Pp and the Pd, we next investigated the interaction between the Cas12a and the Pd. Here, we placed the fluorophore on the nontarget strand, and the quencher on the target strand of the Pd-FQ (Figure H,I, increased signal indicates assembly with Cas12a and strand separation). Notably, the fluorescence did not change when the Pd-FQ was in the presence of the crRNA and Cas12a (Figure I), despite the evidence that these components could assemble (Figures G, S6, and S7C). This suggests that association alone did not result in productive activation of Cas12a, which was supported by the trans-cleavage findings (Figures C and S3H, yellow). Once a Pp was present, the Pd-FQ strands were separated (Figure I), supporting the interpretation that the Pp is required for the productive activation of, and likely cleavage by, the Cas12a. This was similarly supported by the trans-cleavage data (Figures D and S3H, green). Interestingly, the structure of the PAM did not significantly affect the cleavage of the Pd-FQ (Figure I, configurations I–III), despite the double-stranded PAM improving Pp binding (Figure D, configurations I and II) relative to the single-stranded PAM (Figure D, configuration III). Instead, we see that only the sequence of the PAM significantly impacted cis-cleavage of the Pd-FQ, with the anti-PAM reducing separation of the Pd-FQ strands (Figure I, yellow bars). This evidence suggests that the association between Pp and crRNA-guided Cas12a must be significantly impaired for Cas12a activation to cease on Pd targets. Notably, since this assay does not discriminate products directly, the data support Pp-dependent activation but do not directly confirm Pd cleavage products.

To summarize the findings so far, these experiments add to previous work with split activators in a few important ways. (I) The first is that the position of the split is a key factor in whether the Pd duplex can activate Cas12a alone (Figure C). (II) We also provide experimental evidence that the PAM structure (or duplex configuration) and sequence influence the assembly of the Pp:RNP complex independent of effects on Pd association (Figure D,E). (III) We also found that the RNP can associate with a short Pd independent of the Pp (Figures G and S6), but this association alone does not result in substantial productive cleavage of the Pd (Figure I). (IV) In contrast, inclusion of the Pp enables productive activation, including cis-cleavage of the Pd (Figure I) and trans-cleavage of the reporter (Figure D). (V) Finally, when we combine these findings with previously published structural data (PDB: 8Y04), the results are consistent with a model in which association to the Pp may preorder crRNA bases immediately downstream of the Pp:crRNA duplex and thereby facilitate Pd capture; however, we present this as a working model compatible with the data rather than a uniquely established sequence of assembly events (Figure S5).

Altering the Stability of the Complexes

Given the importance of the Pp:RNP association in the context of split activators (Figures S5 and A), we hypothesized that improving the stability of this intermediate may improve its ability to target Pd, increasing activity of Cas12a. With this in mind, we first increased the GC content between the Pp and the crRNA (Figures B and S8); however, we did not observe improved Cas12a activity, even when the hybridizing sequence in the Pp was single-stranded to reduce interference by the NTS.

3.

Split-activator-mediated activation of Cas12a can be modulated by altering duplex stability. (A) Since Cas12a interacts most stably with the Pp, the crRNA:Pp duplex stability was first altered. (B) Trans-cleavage activity of Cas12a was measured over time using PAM-proximal hybridizing sequences with a GC content of 33%, 66%, and 83%. Conditions were benchmarked against an intact activator. Values represent mean fluorescence intensity (n = 3) ± SD after incubating at 37 °C over 33 min. (C) Locked nucleic acids (LNAs) were substituted into the Pp duplex’s target strand one at a time to determine Cas12a’s tolerance to the analogs. LNA-modified PAM-proximal sequences were benchmarked against nonmodified (WT) Pp sequences. Background-subtracted fluorescence intensity (n = 3) was measured over 30 min at intervals of 3.5 min. Values represent total area under the curve (RFU.min) ± Standard error. Finally, the trans-cleavage activity of Cas12a was measured when substituting DNA bases into the crRNA to create crXNA. Cas12a was guided by (D) a canonical crRNA, crXNA with DNA bases (DNA bases in red) (E) in the Pp-annealing region (positions 2, 3, 5 and 6), DNA bases in the Pd annealing region (F) (positions 7–13, 15, 16, 19, and 20) and (G) (positions 11–13, 15, 16, 19, and 20). Cas12a was activated using various Pp structures. Values represent mean background-subtracted fluorescence intensity (n = 3) ± SD after incubation at 37 °C over 100 min. Analyses were performed using a one-way ANOVA with a multiple comparison test, where ns = not significant with p > 0.05 and asterisks (****p ≤ 0.0001) denote significant differences. Graphics were created in https://BioRender.com.

We next introduced Locked Nucleic Acids (LNAs) into the Pp to increase the stability of the crRNA:Pp duplex. Since LNAs have yet to be reported in split activators, we first mapped the Cas12a’s tolerance to the nucleotide analogs (Figures C and S9). Measurement of trans-cleavage over time revealed that LNAs in positions five and six had adverse effects on Cas12a’s activity (Figures C and S9H,I). The presence (NTS:TS duplex) or absence (single-stranded TS) of a competing NTS did not significantly impact Cas12a’s activity (Figure S9C–I). Furthermore, while single substitutions in positions 1–4 were tolerated (Figures C and S9C–G), introducing multiple LNAs at once reduced Cas12a’s trans-cleavage activity (Figure S10). This reduction was most evident when the multilocked Pp was fully double-stranded, likely because it was more difficult to displace the NTS and access the TS (Figure S10C,D). Preincubating the complex for longer eliminated the difference in kinetics between the duplexed and overhanging Pp, though neither was as favorable as the unlocked Pp (Figure S10E,F). Furthermore, while Cas12a is roughly as sensitive to trace quantities of Pd as compared to an intact activator, locking the Pp (positions 2–4) negatively impacted the protein’s sensitivity (Figure S11A–C,E). Given that locked nucleotides improve the stability of duplexes by shifting the DNA backbone, we conclude that Cas12a can tolerate individual but not consecutive LNAs.

Next, we wondered if destabilizing the complexes could make Cas12a more specific to alterations in the split activator sequences. Following recent literature, we decreased the binding energy of the Pp:crRNA duplex by introducing DNA bases in the guide to create chimeric DNA/RNA guides (crXNA). To minimize the rejection of the DNA bases by Cas12a, we rationally designed the guides to avoid placing deoxyribose sugars where a previous Cas12a structural analysis indicated direct contact with Cas12a’s residues (Figure S12). To determine the effects of weakening the hybridization between the guide and activator sequences, we compared the activity of an unmodified crRNA to the crXNAs. When guided by a crRNA, the structure of the PAM did not significantly impact the trans-cleavage activity of Cas12a (Figures D, S13, and S14A), despite the PAM’s importance in forming the Pp:RNP intermediate (Figure D), which is in agreement with the cis-cleavage findings (Figure I). In contrast, when the Pp-annealing region of the crRNA was partially replaced with DNA, Cas12a was only significantly activated when the PAM was double-stranded (Figures E and S14B). This implies that weakening the nucleic acid binding energy between the Pp and the crXNA may have emphasized the PAM-Cas12a interaction (Figure D,E), making the Cas12a selective for the structure of the PAM. However, it remains unclear exactly how this was achieved. Consistent with this interpretation, fluorescence polarization measurements using the single-stranded Pp substrate (configuration III, Figure ) showed essentially no guide-dependent change in apparent affinity for the canonical crRNA (KD = 33.85 nM with crRNA versus 34.33 nM without guide), whereas the crXNA produced a clear ∼2-fold shift (KD ≈ 11.43 nM with crXNA versus ≈ 20.61 nM without guide), indicating that altered guide chemistry changes the energetic balance of the Pp-bound intermediate (Figure S7D,E). Finally, DNA substitutions in the Pd annealing region reduced Cas12a activity, making those chimeric guides nonviable (Figures F,G and S14C,D).

Improved Substrate Selectivity with Chimeric Guide

To determine if the crXNA alone made Cas12a more reliant on the PAM, we incubated the crXNA-guided Cas12a with an intact activator sequence (full 20nt), either with or without the PAM (Figure A). Under these conditions, just like with the crRNA (Figures C and S3B), the crXNA was equally active regardless of the PAM. From this, we can infer that the crXNA does not inherently require a PAM to associate with its target DNA and activate Cas12a. This implies that the selectivity for the structure of the PAM in the Pp sequence (Figures E and B) may be an emergent property of a split activator. In light of these findings, we hypothesized that altering the crXNA backbone destabilized the Pp-hybridized intermediate, ultimately requiring the PAM to maintain the Pp-Cas12a association (Figure S15D, configuration III → II).

4.

Improved substrate selectivity with chimeric guide. (A) Trans-cleavage activity of crXNA-guided Cas12a targeting a full-length protospacer with or without the PAM, and split activator with or without the Pp. Background-subtracted fluorescence intensity was measured over 2 h at intervals of 200 s. Values represent total area under the curve (RFU.min) ± standard error. (B) crXNA-mediated activation of Cas12a requires a double-stranded PAM in the context of a split activator. (C–F) crRNA (left column)- and crXNA-mediated (right column) activation of Cas12a was compared when activated with Pp of different structures and PAM sequences as well as double-stranded (top row) or single-stranded (bottom row) Pd activators. Background-subtracted fluorescence intensity was measured over 60 min, with readings every 220 s. Values represent total area under the curve (RFU.min) ± standard error. (G) Trans-cleavage of the reporter was measured after a crXNA-guided Cas12a was provided with a double-stranded Pd activator, various structures of the Pp activator, and incubated over 20 h. Values represent mean background-subtracted fluorescence intensity ± SD. Analyses were performed using a one-way ANOVA with a multiple comparison test, where ns = not significant with p > 0.05 and asterisks (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, and ****p ≤ 0.0001) denote significant differences. Graphical depictions of crXNA-guided Cas12a associating with a Pp and (H) activating when the PAM is properly structured and (I) not activating when the PAM is altered. This is likely because destabilizing hybridization in the Pp-bound intermediate makes split activator complexes prone to collapse when other key interactions are altered. (J) If so, then introducing mismatches between the crXNA and the split activators may similarly destabilize the complexes. All experiments were performed at n = 3. Graphics were created in https://BioRender.com.

To explore the PAM’s relevance in maintaining the Pp-bound complexes (Figure S15B–E), we compared crRNA- (normal hybridization) and crXNA-mediated (altered hybridization) activation using various Pp structures and PAM sequences (PAM = 5′-TTTG-3′ vs Anti-PAM = 5′-AGAG-3′) (Figures C–F, S16, and S17). We first targeted a double-stranded Pd whose nontarget strand would need to be displaced by the Pp-bound Cas12a, making association more challenging. When guided by a crRNA, activity was slightly reduced when the PAM was single-stranded and significantly reduced under an “Anti-PAM” (Figures C and S16A,B), reflecting the importance of the PAM’s structure and sequence in associating the Pp and Cas12a (Figure D,E). When targeting a more accessible, single-stranded Pd, crRNA-mediated activity was drastically enhanced (Figure D). Here, the structure and sequence of the PAM still affected the activation of Cas12a, though the overall activity was very strong (Figure S16C,D). From this, we can infer that under conditions where the crRNA can hybridize more easily to its target (single-stranded Pd target), the PAM is less relevant. This is either because (I) the PAM was involved in unwinding the Pd target, which is now unnecessary with a single-stranded Pd target, or (II) the PAM is involved in stabilizing the Pp-bound state, making it very important when targeting harder-to-access double-stranded Pd targets, but comparatively less relevant when targeting single-stranded targets.

To test this, we used the crXNA to weaken the nucleic acid binding energy between the Pp and the Cas12a. We noted that the crXNA-mediated activity was strong when the PAM was double-stranded, and all but eliminated when the PAM was either single-stranded or contained the incorrect sequence (Anti-PAM = 5′-GAGA-3′) (Figures E and S17A,B). When the Pd-target was single-stranded and therefore easier to hybridize, the crXNA-guided Cas12a was slightly active when using a single-stranded PAM, but did not tolerate an Anti-PAM (Figures F and S17C,D). The crXNA-guided Cas12a was very active on the single-stranded Pd target with the PAM, implying that the PAM is important for Pp binding when hybridization is impaired, but not very active when the PAM was poor. These findings are more consistent with PAM playing a role that stabilizes the Pp-bound state than promoting unwinding of the Pd.

Cas12a’s activity can be modulated through selective truncations in the PAM’s duplex (Figure S18). When guided by a crXNA, the final position of the PAM does not affect Cas12a’s activity (5′-TTTG-3′) (Figure S18C,D). The removal of the adjacent thymine significantly reduces Cas12a activation, with no noticeable activity over baseline when two thymines are removed (Figure S18E,F). Interestingly, the Lys595 residue in the PAM-interaction domain is inserted between these two positions and is thought to be involved in stabilizing the PAM-binding channel. Compromising this interaction significantly reduced Cas12a activity, supporting the hypothesis that the protein relies more heavily on PAM-bound interactions in the context of a split activator (Figure S15D,E).

Having determined that the PAM’s structure could modulate the activity of a split activator system, we wondered how well this effect would persist over a long incubation. To test this, we incubated Cas12a with various DNA substrates over 20 h. When targeting a double-stranded Pd, whose NTS would need to be displaced for the crXNA to hybridize, the double-stranded PAM yielded a strong signal (Figures G and S19A), while the single-stranded PAM was significantly less intense. When targeting a more easily accessible, single-stranded Pd, Cas12a activity was weaker in the presence of the single-stranded PAM (Figure S19B), but still resulted in full cleavage after 20 h (Figure G). When the PAM was absent, the presence or absence of the Pp did not affect crXNA-guided Cas12a activity (Figures G and S19A,D). In comparison, the crRNA that hybridizes normally to the Pp could associate with the PAM-less Pp and activate Cas12a (Figure S19C,D).

Taken together, we can see that altering the hybridization between the Pp and the guide (crXNA) caused Cas12a to rely more heavily on the PAM when associating with the Pp (Figure H). Alterations to the PAM were therefore poorly tolerated by Cas12a, making it unlikely for the protein to associate with both the Pp and Pd, significantly reducing activity (Figure I). In other words, interfering with the stability of the complexes made them prone to collapse if any other key interactions were compromised, effectively making Cas12a selective for the structure of the PAM. This therefore led us to wonder if destabilizing the complexes through mismatches between the crXNA and the split activator would similarly make Cas12a specific for individual single-nucleotide polymorphisms (Figure J).

Improved Specificity with Chimeric Guide

When targeting an intact activator, Cas12a anneals its crRNA to the TS in a process known as R-loop elongation. This dynamic process is primarily driven by the electrostatic forces from DNA:RNA hybridization and DNA:DNA unpairing, with little intervention on the part of Cas12a. , As such, mismatches in the TS increase the dissociation kinetics of the growing R-loop, favoring its collapse and preventing Cas12a activation. However, this is not a perfect process, and outside of its seed region (first ∼5nt), Cas12a is not very specific to individual mismatches between its crRNA and the TS.

To test if the less stable crXNA-guided Cas12a could be more specific to mismatches in the context of a split activator, we first investigated the effects of mutations along the Pp (Figure S20). Individual mutations were installed along all six bases of the Pp and the first six bases of the control (intact) activator. As expected, the crRNA-guided Cas12a demonstrated poor specificity for individual mutations in an intact activator (Figure S20A). Interestingly, neither the crXNA targeting an intact activator (Figure S20B) nor the crRNA targeting a split activator (Figure S20C) improved the specificity of Cas12a. However, when a crXNA was used to target a split activator, all but the fifth position in the crRNA became more specific for individual mutations (Figure S20D). This indicates that mismatches between the Pp and the crXNA reduced the stability of the complex, leading to improved specificity for individual mutations. Notably, while the specificity of Cas12a was improved in the Pp, it was theorized that the dsPAM allowed the Pp:XNP to reassemble from the PAM-bound conformation after R-loop collapse. This would reduce the impact of mismatches in the duplex by facilitating Cas12a’s reentry into the Pp-bound state (Figure S15D).

Since the Pd does not have a PAM, we reasoned that R-loop collapse along this fragment could result in improved specificity. Like before, we substituted individual positions in the Pd and an intact activator (positions 7–20) and used the new duplexes to activate Cas12a. As expected, when targeting an intact activator, the crRNA-guided Cas12a demonstrated poor specificity, with mutant/WT-normalized fluorescence values ranging from 0.19–1.18 and a median of 1.00 after 20 min (Figures A and S21). The crXNA noticeably improved the Cas12a’s specificity, resulting in WT-normalized fluorescence values ranging from 0.04–0.90 and a median of 0.36 after 20 min. After 90 min, the WT-normalized fluorescence ranged from 0.13–1.17, with a median of 0.76 (Figures B and S22). Here, it is possible that the seed-bound conformation (Figure S1B) with the crXNA was less stable than the one formed with the crRNA, allowing the reverse propagation of the R-loop to more easily dissociate the Cas12a from the DNA duplex.

Splitting the activator while using a crRNA (Figures C and S23) similarly improved the specificity of Cas12a in many positions, resulting in WT-normalized fluorescence values of 0.07–0.99 with a median of 0.38 after 20 min, and 0.21–1.09 with a median of 0.90 after 90 min. Here, it is possible that after encountering a mutation in the Pd, the reverse propagation of the R-loop proceeded to the blunt end of the Pd, releasing the Pp:RNP complex.

Interestingly, combining both approaches significantly reduced the WT-normalized fluorescence from mutated targets, with a range of −0.05 to 1.87 and a median of 0.08 after 20 min, and a range of −0.01 to 1.31 with a median of 0.04 after 90 min (Figures D and S24). Remarkably, after incubating the samples for 20 h, the crXNA-guided split activator system remained highly discriminatory, with WT-normalized fluorescence values ranging from 0.01 to 0.95 and a median of 0.08, despite evidence that the Cas12a/crXNA complex remained functional (Figures S25 and S26). Importantly, this enhanced discrimination is not explained solely by slower overall activity of the crXNA-guided system. The perfectly matched crXNA-guided split activator reached strong fluorescence within the experimental window, whereas mismatch-containing Pd targets remained near baseline even after 20 h of incubation (Figure S26). Moreover, addition of a perfectly matched Pd after the 20 h incubation restored reporter cleavage, indicating that the Cas12a/crXNA/Pp complex remained active and that the low signal from mismatched targets reflected failure to productively assemble and/or activate the complex rather than loss of enzyme activity.

Based on these results, we hypothesize that the reduced stability at the Pp:crXNA interface promoted reverse propagation of the R-loop in the Pd under noncomplementary base-pairing conditions. To determine if this improvement in specificity was effective in discriminating between transition and transversion mutations, we next prepared Pd duplexes with all possible combinations of mismatches (e.g., C:G, C:T, C:A, and C:C) for each of the four bases (e.g., A, C, G, and T) in the crRNA (Figure S27A,B). Indeed, the crXNA-guided Cas12a demonstrated remarkable specificity against these substrates (Figure S27C–J), with the interesting exception of a G:T mismatch at position 12 of the crRNA’s spacer (Figure S27K,L). Furthermore, the activity of the crXNA-guided Cas12a on a split activator with two adjacent polymorphisms is absent for all combinations of mutations, aside from the final 19/20 mutation (Figures S28 and S29).

Finally, because the crXNA-guided Cas12a activated relatively slowly in the split activator format, we tested PEG-8000 as a crowding additive to improve reaction kinetics. PEG-8000 increased reporter cleavage over time (Figure S30A). Importantly, when specificity was assessed at a representative PAM-distal mismatch position (position 18), the crXNA-guided system retained substantially stronger discrimination than the corresponding crRNA-guided system (Figure S30B–D). We therefore interpret this experiment as evidence that PEG-8000 can improve reaction kinetics while preserving specificity at the position tested rather than as a comprehensive demonstration of broad SNP specificity under PEG-8000 conditions.

Discussion

In this study, we explored the structural requirements of double-stranded split activator systems and how these components assemble to activate the CRISPR/Cas12a. First, we found that Cas12a can be activated when the interrogated PAM-distal strand (Pd) is physically separated from the PAM-bearing PAM-proximal duplex (Pp). In terminal Pd-only configurations, activation can occur under fraying-permissive conditions even though Pd itself lacks an adjacent PAM; however, in the split-activator architecture used here for specificity engineering, productive activation is more accurately described as externally PAM-supported because the Pp duplex supplies the PAM-dependent interaction that stabilizes Cas12a assembly. This corroborates and builds upon a previous report, which proposed that Cas12a can target split activators using frayed DNA extremities as surrogates for the PAM. We therefore interpret our data as evidence for spatial separation of PAM presentation from Pd interrogation rather than PAM independence of Cas12a.

Next, we thoroughly explored the assembly of the various components in the split activator system. We experimentally demonstrated that the role of the PAM is to mediate the assembly between the truncated PAM-proximal (Pp) and RNP, and that this effect is influenced by the PAM’s sequence and structure. Given the evidence demonstrating the role of the PAM in mediating the colocalization between the Pp and the Cas12a, we propose that the PAM primarily favors the cleavage of the PAM-distal (Pd) by stabilizing the Pp:RNP complex. Contrary to previous literature, we additionally found that a Pd that does not induce notable trans-cleavage activity in the Cas12a can still associate to the protein. We have therefore proposed a revised model detailing the association of the split activator system.

To further demonstrate the importance of the direct DNA:protein interactions in the context of a split activator, we selectively altered the Pp:crRNA interface. By replacing the crRNA with a chimeric DNA/RNA guide, we altered the balance of stabilizing interactions within the split-activator intermediate, thereby increasing the contribution of PAM-supported protein–DNA interactions. This enhances our understanding of Cas12a by further showcasing the role of PAM in immobilizing Cas12a during target interrogation. Additionally, given the recent focus on single-stranded split activators, understanding the importance of the double-stranded PAM in associating the Cas12a and the TS can be leveraged to improve downstream applications such as logic gates and diagnostics, hence, our work.

Additional support for our model was obtained through fluorescence polarization measurements of isolated split activator components (Figure S7A–C). Importantly, these experiments were designed to quantify the intrinsic association of individual Pp and Pd probes with Cas12a and therefore do not directly report on formation of the complete ternary assembly. Nevertheless, canonical PAM-containing Pp substrates consistently exhibited stronger binding than corresponding anti-PAM controls, particularly in partially exposed duplex architectures, supporting the importance of PAM presentation during early assembly events. Interestingly, isolated Pd probes also displayed measurable Cas12a association. This observation suggests that the inability of Pd alone to activate trans-cleavage cannot be attributed solely to an absence of binding. Additionally, the fluorescence polarization measurements further support this interpretation by showing that, unlike the canonical guide, the crXNA produces a measurable guide-dependent shift in apparent affinity for the mode-III Pp substrate (Figure S7d,e), indicating that the chimeric guide does not simply weaken binding uniformly but instead reweights the stabilizing interactions within the split intermediate. When considered alongside the FRET, EMSA, and functional cleavage experiments, these data support a model in which substrate association and productive activation are distinct processes, with efficient activation requiring the PAM-sensitive assembly pathway established by the Pp component.

On top of increasing the stringency for the structural recognition of the Pp, destabilizing the Pp:RNP interaction significantly increased the burden on the complementarity between the Pd and the guide. This was leveraged to achieve greatly enhanced specificity along all but the first and final bases in the Pd even after 20 h. This specificity remained extremely strong regardless of whether the system was used on transition or transversion mutations, with only one exception. This architecture has potential utility for the field of diagnostics, as detecting individual mutations in the Pd is inherently a PAM-less process, overcoming yet another of Cas12a’s major limitations.

Although the split-activator strategy imposes sequence design constraints, its broader significance lies in revealing a previously underexplored mechanism for controlling Cas12a activation through programmable nucleic acid assembly. By decoupling PAM recognition, guide-target hybridization, and activator architecture, we established new design principles for engineering Cas12a specificity and activity. These mechanistic insights provide a foundation for the development of next-generation CRISPR technologies and will guide future efforts to expand target compatibility and translate split activator systems into applications such as programmable genome engineering and diagnostics.

Conclusion

In conclusion, our study provides a detailed breakdown of the activity profiles of a split double-stranded activator when using Cas12a with different truncations. The detailed analysis of how the complexes form will allow future work to improve the types of activity that can be leveraged using such an architecture, a feat we have proven to be possible by already significantly improving the specificity in PAM-less target sequences.

Glossary

Abbreviations

PAM

protospacer-adjacent motif

CRISPR

clustered regularly interspaced short palindromic repeats

Cas12

CRISPR-associated protein 12

Pp

PAM-proximal DNA

Pd

PAM-distal DNA

Pp-F

PAM-proximal DNA labeled with fluorescein

Pd-Q

PAM-distal DNA labeled with a quencher

DNA

DNA

RNA

ribonucleic acid

NTS

nontarget strand

TS

target strand

crXNA

CRISPR chimeric DNA/RNA guide

crRNA

CRISPR RNA guide

crRNA-Q

CRISPR RNA guide labeled with a quencher

FRET

fluorescence resonance energy transfer

FAM

fluorescein

RFU

relative fluorescence units

RNP

ribonucleoprotein

PDB

protein database

LNA

locked nucleic acids

WT

wild-type

SNP

single-nucleotide polymorphism

Correspondence and requests for materials should be addressed to Keith Pardee (keith.pardee@utoronto.ca). Materials are available from Keith Pardee upon request and subject to completion of a materials transfer agreement with the Leslie Dan Faculty of Pharmacy, University of Toronto. Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request. The data underlying this article are available in the article and in its online Supporting Information. Sequences for synthetic nucleic acid oligonucleotides used in this work are detailed in Table S1 of the Supporting Information. All data supporting the findings of this study are available within the article and its online Supporting Information document. Data supporting the plots are available upon reasonable request to the corresponding authors.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.analchem.6c03855.

• Proposed assembly mechanisms, predicted protein/RNA/DNA structures, Pp and Pd length optimizations, crystal structures, EMSA binding assay, FRET results, fluorescence spectra, LNA position, sensitivity data for chemically modified split activators, hydrogen bond map, crXNA position, PAM structure, specificity, chemical additive (PEG-8000), DNA sequences, XNA sequences, RNA sequences, and tagged oligo sequences (PDF)

G.L.: Project administration, conceptualization, formal analysis, methodology, investigation, validation, visualization, writingoriginal draft. F.V.: Conceptualization, formal analysis, methodology, investigation, validation, writingoriginal draft. I.A.I.: Conceptualization, methodology, supervision, writingreview and editing. A.C., S.S., S.S.: Methodology, investigation. C.B., K.G., Y.L., J.R.: Writingreview and editing. K.P.: Supervision, visualization, writingreview and editing. J.P.T.: Resources, supervision, writingreview and editing. All authors have given approval to the final version of the manuscript.

The authors declare the following competing financial interest(s): The authors declare the following competing interests: I.A.I., K.P., and A.C. are listed as inventors on a patent application related to the CRISPR-Cas12 PAM-free nucleic acid detection through target sequence breaks, involving the University of Toronto - Canada, and the Hong Kong University of Science and Technology (US Patent App. 63/737,199). The patent is not directly related to the current work. K.P. is the co-founder of En Carta Diagnostics, Inc. The remaining authors declare no competing interests.