The Acidaminococcus sp. Cas12a nuclease recognizes GTTV and GCTV as non-canonical PAMs

ABSTRACT

The clustered regularly interspaced short palindromic repeat (CRISPR)-associated (Cas) nuclease Acidaminococcus sp. Cas12a (AsCas12a, also known as AsCpf1) has become a popular alternative to Cas9 for genome editing and other applications. AsCas12a has been associated with a TTTV protospacer-adjacent motif (PAM) as part of target recognition. Using a cell-free transcription-translation (TXTL)-based PAM screen, we discovered that AsCas12a can also recognize GTTV and, to a lesser degree, GCTV motifs. Validation experiments involving DNA cleavage in TXTL, plasmid clearance in Escherichia coli, and indel formation in mammalian cells showed that AsCas12a was able to recognize these motifs, with the GTTV motif resulting in higher cleavage efficiency compared to the GCTV motif. We also observed that the -5 position influenced the activity of DNA cleavage in TXTL and in E. coli, with a C at this position resulting in the lowest activity. Together, these results show that wild-type AsCas12a can recognize non-canonical GTTV and GCTV motifs and exemplify why the range of PAMs recognized by Cas nucleases are poorly captured with a consensus sequence.

The Cas12a nuclease from Acidaminococcus sp. (AsCas12a) can recognize a wider set of protospacer-adjacent motif (PAM) sequences, expanding the targeting range for CRISPR technologies.

INTRODUCTION

Clustered regularly interspaced short palindromic repeat (CRISPR)-CRISPR-associated (Cas) systems have become widespread tools used for various biotechnological applications (Adli 2018; Li et al. 2019). In nature, these systems serve as adaptive immune systems that protect prokaryotes from the invasion of mobile genetic elements by targeting and cleaving foreign DNA or RNA (Barrangou et al. 2007; Brouns et al. 2008; Hale et al. 2009; Garneau et al. 2010). Cleavage is directed by non-coding guide RNAs (gRNAs), which possess a guide sequence that is complementary to a target sequence (known as the protospacer). These gRNAs complex with its associated Cas nuclease and bind to the target sequence, leading to cleavage and/or degradation of these targets (Hale et al. 2009; Garneau et al. 2010; Westra et al. 2012). The ease of reprogramming Cas nucleases and gRNAs has led to advancements in various fields, such as genome editing, gene regulation and diagnostics (Barrangou and Doudna 2016; Chertow 2018).

To successfully utilize Cas nucleases, each nuclease must be thoroughly characterized. This includes characterizing the nuclease's gRNA structure, cleavage pattern (i.e. blunt vs. staggered), ability to tolerate mismatches across the target sequence, and other factors required for gRNA processing and DNA/RNA cleavage (Deltcheva et al. 2011; Semenova et al. 2011; Wiedenheft et al. 2011; Jinek et al. 2012; Jiang et al. 2013; Zetsche et al. 2015). Arguably one of the most important requirements is determining the protospacer-adjacent motif (PAM), a short sequence comprised of 3–8 nucleotides that is initially recognized by the nuclease prior to assessment of base-pairing between the guide and the target (Mojica et al. 2009; Leenay and Beisel 2017). The PAM requirement is key for target recognition and cleavage (Jinek et al. 2012; Jiang et al. 2013) and also helps to prevent the system from targeting its own CRISPR locus which lacks a PAM sequence. Over the past decade, the PAMs of various Cas nucleases have been deciphered. For example, one of the pioneering Cas nucleases, the Type II-A Cas9 from Streptococcus pyogenes (SpCas9), was shown to recognize an NGG PAM directly adjacent to the 3′ end of the displaced strand of the protospacer (Mojica et al. 2009; Jinek et al. 2012; Jiang et al. 2013). While NGG remains the canonical PAM for this nuclease, it has been shown that SpCas9 more weakly recognizes non-consensus PAMs, such as NAG and NGA (Hsu et al. 2013; Jiang et al. 2013; Zhang et al. 2014). As the PAMs of various Cas nucleases have been reported, a common theme has emerged in which different Cas nucleases are associated with less-recognized PAM sequences that deviate from the consensus PAM (Esvelt et al. 2013; Hsu et al. 2013; Zhang et al. 2014; Leenay et al. 2016). These non-canonical PAMs have the potential to broaden the target space of a given nuclease but also increase the potential number of off-target sites associated with these PAMs.

Like SpCas9, the Type V-A Cas12a from Acidaminococcussp. (AsCas12a, also known as AsCpf1) has become a widely-used Cas nuclease. However, unlike SpCas9, AsCas12a is able to cleave target sequences containing a TTTV PAM adjacent to the 5′ end of the displaced strand of the protospacer (Zetsche et al. 2015). AsCas12a has been repurposed for various applications, such as genome editing, gene regulation, and diagnostics (Zetsche et al., 2015, 2017; Tak et al. 2017; Chen et al. 2018). While the TTTV PAM has been reported as the consensus motif for AsCas12a, this nuclease is known to recognize weaker CTTV and TCTV PAMs (Zetsche et al. 2015; Kim et al. 2017). To further broaden the PAM specificity of AsCas12a, others have performed structure-guided mutagenesis to create variants of AsCas12a capable of recognizing TCCV and TATV PAMs (Gao et al. 2017; Kleinstiver et al. 2019). While these variants have increased the targeting range of AsCas12a, they were developed from a wild-type version thought to only recognize the TTTV and, to a lesser extent, CTTV and TCTV motifs.

In this work, we discovered that the wild-type AsCas12a can recognize GYTV motifs. This insight came from a cell-free transcription-translation (TXTL)-based PAM screen (Marshall et al. 2018; Maxwell et al. 2018). DNA cleavage in TXTL, plasmid clearance in Escherichia coli, and indel formation in HEK293T cells confirmed that AsCas12a could recognize GYTV motifs, with targets containing GTTV PAMs being more efficiently recognized compared to GCTV PAMs. We also observed a bias at the -5 position of both motifs, suggesting that AsCas12a could recognize a wider stretch of DNA as part of the PAM. Together, these results indicate that the range of PAM sequences recognized by AsCas12a is broader than originally reported, thus increasing the targeting range of AsCas12a for its various biotechnological applications.

MATERIALS AND METHODS

Strains, plasmids and oligonucleotides

All strains, plasmids, primers and gBlocks used in this work can be found in Table S1 (Supporting Information).

TXTL-based PAM screen and DNA cleavage assay

For the PAM screen and DNA cleavage assays, we used a commercially available cell-free TXTL system developed from an all-E. coli lysate (Arbor Biosciences, Cat: 507096) (Garamella et al. 2016). The materials and methods to perform these assays, prepare the next-generation sequencing (NGS) library and conduct data analysis have been described in thorough detail elsewhere (Marshall et al. 2018; Maxwell et al. 2018). The AsCas12a expression plasmid used for these experiments has been reported previously (Liao et al. 2018). For the DNA cleavage assay, we used a modified version of pCB848 from previous work that expresses the GFP reporter (Maxwell et al. 2018). The PAM sequence of pCB848 was mutated using the Q5 Site-Directed Mutagenesis Kit (NEB, Cat: E0554S). The same target site was used for all PAMs tested. The gRNA was expressed from the gBlock (custom gene fragment) TJ524 to target the GFP reporter plasmid and the 5N-randomized PAM-library. The non-targeting control was gBlock CSM-GB019, which contained a randomized, non-targeting gRNA. Each gRNA was expressed in its processed form (20-nt repeat and 24-nt guide) from the J23119 promoter and terminated using a rho-independent terminator. GFP fluorescence was measured using a Synergy H1 plate reader from Biotek with excitation and emission wavelengths of 488 and 553 nm, respectively. The reported production of GFP was calculated using a linear standard calibration curve developed from recombinant eGFP (Marshall et al. 2018). While this calibration curve will vary from factors such as the plate reader and reagents used, our GFP production was calculated by dividing the raw fluorescence values by a conversion factor 9212.6. The PAM library was previously made using methods described in detail elsewhere (Maxwell et al. 2018). The reaction setup for the TXTL reactions can be found in Table S2 (Supporting Information). The PAMs and target sequences used for the DNA cleavage assay can be found in Table S3 (Supporting Information). The NGS data, including the raw data and post-processing reads, were deposited in the NCBI gene expression omnibus (accession # GSE123443).

Plasmid clearance assay in E. coli

We used CBS-445 for expression of bacterial AsCas12a and CBS-444 for expressing the gRNA. The plasmid containing the gRNA target sequence flanked by NGYTC, TTTC or GGCT PAMs were constructed using Q5 mutagenesis. We first transformed 50 ng of CBS-444 into electrocompetent E. coli cells containing CBS-445 and a plasmid containing the gRNA target sequence. After a 1-h recovery at 37°C with shaking at 250 rpm in SOC, serial-diluted cells were plated on LB agar plates supplemented with ampicillin, kanamycin and chloramphenicol. After a 16-h incubation at 37°C, the colonies were counted for analysis.

Indel formation in DNMT1

The target sites and their PAMs for editing DNMT1 can be found in Table S3 (Supporting Information). The mammalian AsCas12a expression plasmid was obtained from Addgene (Cat: 69982). The gRNA expression plasmids, which encode a processed gRNA under the control of an hU6 promoter, were constructed as described elsewhere (Kim et al. 2016). Briefly, the empty gRNA plasmid was digested with BbsI-HF (NEB, Cat: R3539S), and ligated with phosphorylated and annealed oligos, which contained the target sequence-of-interest. Transfection-grade DNA was prepared using the QIAGEN Plasmid Mini Kit (Qiagen, Cat: 12125). One day prior to the transient transfections, 2 × 10⁵ HEK293T cells were seeded in each well of a 12-well plate with 1 mL of complete media (Dulbecco's Modified Eagle Medium (Invitrogen, Cat: 11965-092) supplemented with 10% fetal bovine serum (Invitrogen, Cat: A3840001) and 1% penicillin-streptomycin (Invitrogen, Cat: 15070063). For each gRNA tested, 160 ng of the gRNA plasmid and 640 ng of the AsCas12a plasmid were transfected using jetPRIME (Genesee Scientific, Cat: 55-132). Cells were then incubated for 20 h at 37°C prior to replacing fresh growth media into each well. After media replacement, the transfected cells were incubated for another 52 h at 37°C prior to genomic DNA isolation.

Tracking of Indels by Decomposition (TIDE) analysis

Genomic DNA was isolated using the GeneJET Genomic DNA Purification Kit (ThermoFisher Scientific, Cat: K0721). An amplicon bridging all tested target sites were amplified by PCR from the genomic DNA using Q5 Hot Start High-Fidelity 2X Master Mix (NEB, Cat: M0494L) and primers TJ719/TJ722. After successful amplification, samples were prepared for Sanger sequencing using a DNA cleanup kit (Zymo Research, Cat: D4013). The primer closest to the predicted cleavage site for each target sequence was chosen for the Sanger sequencing reactions (either TJ719 or TJ722). The chromatograms for each sample were analyzed using TIDE (Brinkman et al. 2014). Genomic DNA isolated from each tested PAM was analyzed against a non-PAM negative control. The target sequences and their PAMs can be found in Table S3 (Supporting Information).

RESULTS

PAM screen of AsCas12a reveals non-canonical motifs

We were interested in the full range of PAM sequences recognized by AsCas12a. We began with a TXTL-based PAM screen that we recently developed (Marshall et al. 2018; Maxwell et al. 2018). As part of the assay, constructs separately encoding AsCas12a, a targeting gRNA or non-targeting gRNA, and a 5N-randomized PAM library flanking the target sequence were added to a TXTL reaction. Each reaction was incubated at 29°C for 16 h, and the uncleaved library members were then amplified by PCR and subjected to NGS. As part of the screen, the most strongly recognized PAMs should exhibit the highest depletion in the reactions with the targeting versus non-targeting gRNAs (Fig.   1A, see Table S2, Supporting Information, for the reaction setup). To determine the extent of cleavage of the PAM library, a GFP expression plasmid containing a TTTC PAM flanking the same target sequence was added to the reaction. GFP levels plateaued after 4 h (Fig. S1, Supporting Information), indicating that the PAM-library likely underwent extensive cleavage.

Figure 1.

A TXTL-based PAM screen with AsCas12a identifies GTTV and GCTV as non-canonical motifs. (A) Schematic of the TXTL-based PAM screen. The AsCas12a expression construct, a targeting or non-targeting gRNA expression construct and the PAM-library plasmid were added into a TXTL reaction and incubated at 29°C for 16 h. Plasmids containing recognized PAM sequences are cleaved, while the non-PAMs remain in the reaction mix. After incubation, the remaining PAM sequences in the library were amplified by PCR and subjected to next-generation sequencing. PAM sequences were then identified based on the extent of their depletion in the reaction with the targeting gRNA versus that with the non-targeting gRNA. (B) Plots showing the fold-change of each nucleotide at the different positions in the PAM library following cleavage by AsCas12a. The vertical axis has been inverted to emphasize depleted nucleotides. (C) List of the top 20 most depleted PAM sequences from the screen in rank order. The bolded, red text indicates a G in the PAM sequence. (D) The fold-change of selected 5-mers from the PAM screen.

Figure 1B shows the fold-change of each nucleotide at each position in the PAM library resulting from AsCas12a cleavage. From this plot, AsCas12a recognized the canonical TTTV PAM, with the T in the -4 position being more flexible compared to the -2 and -3 positions. While the data from Fig.   1B confirmed the consensus motif, the top 20 depleted PAM scores included GGTTA and AGTTA (Fig. 1C, Table S4, Supporting Information). Similarly, the PAM wheel of AsCas12a showed a high depletion of the consensus TTTV motif, though the GYTV can be observed when investigated deeper (Supplementary File 1). To probe the screening results more deeply, we generated dot plot showing the fold-change of 5-mers for the previously reported PAMs of NCTTN, NTCTN and NTTCN (where the last PAM is poorly recognized) (Zetsche et al. 2015; Kim et al. 2017) as well variants of the first two PAMs with G at the -4 position (NGTTN and NGCTN) (Fig. 1D). While the canonical NTTTV PAM was the highest-depleted motif, the depletion of the NGTTV PAM was comparable to that of the NCTTV and NTCTV PAMs. Furthermore, the depletion of the NGCTV motif was less pronounced but comparable to that of the NTTCV motif. Together, these data suggested that AsCas12a could recognize a non-canonical GYTV motif, with the GTTV motif preferred over the GCTV motif.

AsCas12a can recognize the GYTV motif in vitro

To further investigate the output of the PAM screen, we tested the cleavage activity of AsCas12a using the GYTV motif in an in vitro DNA cleavage assay (Fig. 2). We conducted the assay similarly to the TXTL-based PAM screen (Fig. 1A) except that the GFP reporter plasmid was used in place of the PAM-library plasmid. The same target sequence as that in the PAM library was also used. We first tested the cleavage efficiency of AsCas12a using all potential sequences within an AGYTN motif as well as two controls: a canonical TTTC PAM and unrecognized GGCC non-PAM. Performing the DNA cleavage assays, we observed that AGYTV sequences but not AGYTT sequences led to DNA cleavage, where the canonical TTTC PAM resulted in more rapid cleavage than the GTTV motif, while the GTTV motif led to more rapid cleavage than the GCTV motif. These results further confirm the ability of AsCas12a to recognize the GYTV motif, with a preference for GTTV over GCTV.

Figure 2.

AsCas12a can recognize the GYTV motif as part of DNA cleavage in TXTL. (A) Time-courses of GFP expression in TXTL. The AsCas12a nuclease was expressed with a non-targeting gRNA (red line) or a targeting gRNA (blue line) designed to target a site upstream of the constitutive promoter controlling GFP. Cleavage leads to rapid degradation of the plasmid and loss of GFP expression. The PAM sequence is indicated above each time-course. Each reaction was incubated at 29°C for 16 h. The error bars represent the standard deviation from three separate TXTL reactions. (B) Fold reduction in GFP for each PAM sequence. The fold reductions were calculated using the GFP fluorescence data from the 16-h time point from the reactions with the targeting gRNA and the non-targeting gRNA. The dotted red line indicates a fold reduction of one (i.e. no reduction). The error bars represent the standard deviation from three separate TXTL reactions.

Further interrogating the results from the PAM screen revealed that only the AGCTV motifs were present from the top 100 depleted PAMs. We therefore asked if the -5 position of the NGCTV motif contributes to DNA cleavage activity. Testing each possible nucleotide in the DNA cleavage assay resulted in a modest but noticeable bias at the -5 position within the NGCTC motif, with A being the most preferred, G and T being similarly preferred and C being the least preferred (Fig. 2). Therefore, the -5 position of the NGCTC motif can influence the in vitro DNA targeting activity of AsCas12a.

The -5 PAM position influences target recognition in E. coli

We next asked if the GTTV and GCTV motifs allow DNA cleavage by AsCas12a in cells, and how the -5 position influences cleavage activity. To answer these questions, we performed a plasmid clearance assay in E. coli by transforming a targeting or non-targeting gRNA plasmid into cells harboring the AsCas12a plasmid and a separate plasmid containing the gRNA target sequence flanked by various PAM sequences (see Table S3, Supporting Information, for target sequence and PAMs). To assess if the bias at the -5 PAM position extends to multiple target sequences, we targeted a sequence different from the one used in the DNA cleavage assay in TXTL. After counting resistant colonies, the transformation fold-reduction was calculated in relation to that with the non-targeting gRNA control. We found that the NGTTC motif led to plasmid clearance that was at least two orders-of-magnitude lower than that for the canonical TTTC PAM (Fig.   3A). We did notice that the colonies associated with many of the NGTTC sequences were smaller than those for the non-targeting control (Fig. S2, Supporting Information), suggesting that cleavage had only partially cleared the plasmid. Sequences associated with the GCTC motif yielded negligible clearance (Fig. 3A and Fig. S2, Supporting Information). When comparing nucleotides at the -5 position of the NGTTC motif, we observed a bias toward an A and against a C (Fig. 3A), in agreement with the in vitro DNA cleavage results (Fig. 2 ). These results indicate that the GTTC motif is a viable PAM for AsCas12a, while the GCTC motif may not be viable at least for plasmid clearance in E. coli. Also, as different target sequences were used for the TXTL experiments and plasmid clearance assays, nucleotide bias at the -5 PAM position of the NGYTV motif can be observed across multiple target sequences.

Figure 3.

AsCas12a can recognize the GYTV motif in vivo as part of plasmid clearance in E. coli and DNA editing in mammalian cells. (A) The average fold reduction of the number of transformants in a plasmid clearance assay in E. coli. Cells that contain the AsCas12a plasmid and the target plasmid with the protospacer flanked by various PAM sequences were transformed with a plasmid expressing the crRNA. The dotted red line indicates a fold reduction of one (i.e. no reduction). The error bars represent the standard deviation from four independent experiments starting from separate colonies. (B) Section of DNMT1 containing the sequences targeted in the indel formation experiments. The red, green, cyan and yellow highlighted text mark the targets and flanking PAMs, including non-PAMs (CAGGT, AAAGT) and the TTTV, GTTV and GCTV motifs, respectively. The 5-mer PAM for each target is bolded. Note that some of the gRNAs were derived from the bottom strand, where each PAM is the reverse complement of the bolded sequence. (C) Indel frequencies for eight DNMT1 target sites in HEK293T cells quantified using TIDE (Brinkman et al. 2014). As part of the analysis, sequencing data from each target sequence was compared to that of the target sequence flanked by the AAAGT non-PAM in the same amplified region. Using a two-tailed t-test, the indel frequencies of all tested GCTV PAMs were statistically different compared to that of the non-PAM. The P-values for GGCTC, AGCTG and TGCTG compared to the non-PAM were 0.0012, 0.0004 and 0.0002, respectively (n = 3). The error bars represent the standard deviation from three independent transfections.

Indel formation can be achieved with AsCas12a using GYTV PAMs

We next sought to test AsCas12a's ability to recognize the GYTV motif in mammalian cells. We specifically chose to target sites within the DNMT1 gene in HEK293T cells, as this gene had been used previously to assess indel formation with AsCas12a (Zetsche et al., 2015, 2017; Tu et al. 2017; Yamano et al. 2017; Li et al. 2018). We transiently transfected a plasmid constitutively expressing AsCas12a along with a plasmid expressing a gRNA targeting DNMT1 at various locations. After 72 h post-transfection, we assessed the frequency of indel formation using TIDE analysis (Brinkman et al. 2014). As part of the experiments, we targeted two different sites flanked by the TTTC motif, three different sites flanked by the GTTV and GCTV motifs and one site flanked by an AGGT non-PAM (Fig. 3A). An additional target site with an AAGT non-PAM was used as the reference for TIDE.

Targets containing the TTTV motif resulted in the highest indel frequency (Fig. 3B, individual averages of two sites = 16.6, 29.6%), while targets with the GTTV motif resulted in an efficiency either less than or comparable to the TTTV motif (individual averages of three sites = 17.3, 9.2, 8.1%). While using the selected GCTV PAMs resulted in a low indel frequency (individual averages of three sites = 5.1, 3.5, 3.1%), the GCTV PAMs resulted in indel frequencies that were statistically different compared to the AGGT non-PAM (0.63%). We thus conclude that AsCas12a can successfully edit targets flanked by GYTV in mammalian cells, with higher editing efficiencies for targets flanked by the GTTV motif than by the GCTV motif.

DISCUSSION

In this work, we found that AsCas12a recognized GTTV and GCTV as non-canonical PAMs for DNA targeting. DNA targeting with the GTTV motif was more efficient than that of the GCTV motif, as expected given AsCas12a's preference for CTTV over TCTV (Zetsche et al. 2015; Kim et al. 2017). While the in vitro DNA cleavage assay and plasmid clearance indicated that TTTV was preferred over GTTV (Fig. 2 and 3A), the indel formation assays in HEK293T cells yielded at least one instance in which targets flanked by TTTV and GTTV motif exhibited similar editing efficiencies (Fig.   3B). This occurrence is most likely due to the influence of the target site sequence, as reported in prior work (Esvelt et al. 2013). Therefore, GTTV in particular may be a viable PAM sequence for genome-editing with AsCas12a in mammalian cells.

In the DNA cleavage and plasmid clearance assays, we showed that the -5 position of the PAM could affect the cleavage efficiency (Fig. 2 and 3A). This insight specifically came from interrogating the NGCTC motif in the in vitro DNA cleavage assay and the NGTTC motif in the plasmid clearance assay in E. coli, which both favored an A and disfavored a C at this position. Similar biases that extended beyond the standard PAM was observed in previous work for the Cas12a from Francisella novicida (FnCas12a) in its -4 position (consensus PAM of NTTV), as well as Cascade from the Type I-E CRISPR-Cas system (consensus PAM of AWG) (Leenay et al. 2016). This effect could be dependent on the target sequence as well as the sequences extending beyond the PAM.

Shortly following the first reports of genome-editing with SpCas9, progress has been made to increase the targeting range of Cas nucleases and to reduce off-target effects. The former has been achieved in part by the discovery of Cas nucleases that recognize distinct PAMs as well as the engineering of Cas nucleases to increase PAM flexibility of wild-type Cas nucleases (Kleinstiver et al. 2015, 2019; Gao et al. 2017; Hu et al. 2018; Nishimasu et al. 2018). While increasing the targeting range has allowed for the targeting of almost every sequence of interest, this inevitably expands the number of potential off-target sites. Work has been done to increase PAM flexibility while reducing off-target effects, such as engineering high-fidelity Cas nucleases, tightly controlling nuclease levels, and carefully selecting target sequences (Kleinstiver et al. 2016; Lee et al. 2016; Hu et al. 2018; Shen et al. 2018). Although one would assume that the targeting range and off-targeting frequency would go hand-in-hand, one engineered variant of SpCas9 called xCas9 managed to achieve a wider targeting range and lower off-target effects compared to the parental SpCas9 (Hu et al. 2018). Though the frequency of off-target effects can be decreased using available web-based tools (Lee et al. 2016), non-canonical PAMs that can be recognized by a particular Cas nuclease must be identified to better predict potential off-target sites.

Consensus sequences have been one of the most common methods of communicating PAMs (Deveau et al. 2008; Horvath et al. 2008). This approach involves the reporting of a single sequence that captures the best recognized set of PAM sequences. For example, the consensus sequence of two commonly used Cas nucleases, SpCas9 and AsCas12a, have been reported as NGG and TTTV, respectively (Mojica et al. 2009; Jinek et al. 2012; Jiang et al. 2013; Zetsche et al. 2015). While the consensus sequence allows for a simple means to describe the PAM, it fails to reveal other PAM sequences Cas nucleases are able to recognize. Examples include SpCas9 recognizing weaker NAG and NGA PAMs and AsCas12a recognizing CTTV and TCTV PAMs (Hsu et al. 2013; Jiang et al. 2013; Zhang et al. 2014; Zetsche et al. 2015). While this work and others have reported PAMs using methods that include the sequence logo, consensus sequence, PAM wheel and PAM table (Deveau et al. 2008; Horvath et al. 2008; Leenay et al. 2016; Leenay and Beisel 2017; Marshall et al. 2018), there currently is no standard for conveying PAM sequences. Though reporting the consensus sequence has become the norm for communicating the PAM, more thorough methods or a set of methods are needed to fully describe the targeting range of Cas nucleases.

Funding

This work was funded by the National Institutes of Health (1R35GM119561) and the National Science Foundation (MCB-1413044).

Conflict of interest. None declared.