Cpf1 nucleases demonstrate robust activity to induce DNA modification by exploiting homology directed repair pathways in mammalian cells
- Eszter Tóth†1, 2,
- Nóra Weinhardt†1, 2, 3,
- Petra Bencsura1,
- Krisztina Huszár1, 2,
- Péter I. Kulcsár1, 2, 3,
- András Tálas1, 4,
- Elfrieda Fodor2 and
- Ervin Welker1, 2Email authorView ORCID ID profile
© The Author(s). 2016
Received: 24 May 2016
Accepted: 31 August 2016
Published: 14 September 2016
Cpf1 nucleases have recently been repurposed for site-specific genome modification. Two members of the Cpf1 family, the AsCpf1 from Acidaminococcus sp. and the LbCpf1 from Lachnospiraceae bacterium were shown to induce higher indel frequencies than SpCas9 when examining four randomly-selected target sequences for each type of nuclease. Whether they are a real match for Cas9 nucleases, however, remains to be verified.
Here, we used AsCpf1 and LbCpf1 to induce homology directed repair, either single strand annealing (SSA) or homologous recombination (HR), in N2a mouse neuroblastoma cells. Exploiting a plasmid that contains two GFP halves with overlapping sequences and exploring 20 targets, on all but one both nucleases consistently performed with above 10 % efficiency. Several Cas9 nucleases have been previously characterised in order to find an orthogonal counterpart for the most widely used promiscuous SpCas9. Here, we found that AsCpf1 and LbCpf1 might be better candidates than three of the best such counterparts: Cas9 from Staphylococcus aureus, from Streptococcus thermophilus and from Neisseria meningitidis, when assessed for inducing efficient SSA mediated repair in N2a cells. When tested on genomic targets exploiting HR, both nucleases were able to induce the integration of a donor cassette with 1000 bp-long homologous arms. We also generated plasmids that express these Cpf1 nucleases together with their cognate crRNAs and that are equipped with type IIS restriction enzyme sites to facilitate spacer cloning.
Our results suggest that employing As- or LbCpf1 nuclease to induce homology directed repair in N2a cells, although is less effective at present than employing SpCas9, it is an equally or more effective tool than the most frequently used orthogonal Cas9 counterparts of SpCas9. These findings support the position of Cpf1 nucleases on the side of SpCas9 on the palette of effective genome engineering tools.
This article was reviewed by Eugene Koonin, Haruhiko Siomi and Jean-Yves Masson.
KeywordsCpf1 nuclease AsCpf1 LbCpf1 CRISPR Cas9 nuclease GFxFP assay SpCas9 SaCas9 NmCas9 StCas9 Genome engineering RNA DNA
Although recent studies suggest that clustered regularly interspaced short palindromic repeat (CRISPR)-associated protein (Cas) systems are also used for non-defense roles in the host [1, 2], they are best known as part of the adaptive immune systems of prokaryotes that have recently been exploited for generating extremely powerful genetic devices [3–6]. The CRISPR-Cas system of the host grants specific resistance against phages and plasmids by keeping “memories” of past infections through acquiring short DNA fragments and destroying the reinvading agents that are identified based on these short, 20-30 nucleotide-long sequences [7, 8]. The adaptation modules that acquire and deposit appropriate sequence motifs from the invading agent to specific locations called CRISPR arrays of the microbial genome, consist of two proteins: Cas1 and Cas2 . The effector complexes show much greater diversity in terms of the participating Cas proteins, across various prokaryotic CRISPR-Cas systems, and are being divided into two classes: Class 1 systems that utilize multiple subunit effector complexes, and Class 2 systems that work with a single large effector Cas protein . Within the two classes five types of complexes are being distinguished (type I, III and IV of Class 1 and type II and V of Class 2) according to the presence of one of the signature proteins: Cas3, Cas10, Csf1, Cas9, and Cpf1, respectively [11–13]. Cas9 proteins, effectors of the type II Class 2 systems have attracted particular attention in the past few years due to their possible harnessing for genome editing that quickly revolutionized the way we are capable to execute site-specific modification of complex vertebrate genomes today [3–5, 13, 14]. Cas9 proteins are also exploited for the purpose to regulate transcription and epigenetic states at specific genomic locations [15–22]. The power of Cas9 as a genome manipulating tool lays in its easy programming. Altering the so called spacer sequence of its associated RNA molecule – that determines the specificity of the nuclease – the Cas9 protein can target different DNA sequences: i.e., those that contain complementary sequences (protospacers) to the chosen spacer sequence and also are immediately followed by a short DNA motif called protospacer adjacent motif (PAM). The PAM is recognized by the protein’s PAM-binding domain and its sequence may differ for Cas9 proteins from different species [22, 23].
Recently, Cpf1 (CRISPR from Prevotella and Francisella 1) nucleases – single effector proteins of the type V Class 2 systems – have been investigated for their possible applicability for genome engineering tasks. Zetsche et al. demonstrated that the Cpf1-containing CRISPR-Cas locus of Francisella novicida U112 encodes a functional defense system that is capable of mediating plasmid interference in bacterial cells . This work revealed three features of Cpf1 of Francisella novicida (FnCpf1) that seem to be general among Cpf1 nucleases but are distinct from those of Cas9s. First, FnCpf1 works with a single CRISPR RNA (crRNA) in contrast to Cas9 that also requires a trans-activating CRISPR RNA (tracrRNA) for crRNA processing and target recognition activity. Second, the PAM of FnCpf1 is rather T-rich in contrast to the more G-rich Cas9 PAMs. Third, while Cas9 nucleases produce blunt-end termini at double strand breaks, FnCpf1 produces 4- or 5-nt-long 5’ overhangs . Initially, Cpf1 nucleases were proposed to act as dimers based on the fact that they contain only one known nuclease domain (RuvC) and the inactivation of this domain leads to a completely inactive Cpf1 in contrast to Cas9 where the inactivation of one nuclease domain generates a nickase . However, it is not clear how the concerted action of two, dimerized Cpf1 proteins might result in a double strand break. Recent studies examining the crystal structure of one of the Cpf1 nucleases, the Lachnospiraceae bacterium ND2006 Cpf1 (LbCpf1), and carrying out biophysical assays on FnCpf1 showed that they display no oligomerisation upon binding to crRNS and/or DNA [25, 26]. The crystal structure of Acidaminococcus sp. BV3L6 Cpf1 (AsCpf1) reveals a nuclease domain with a new fold responsible for the cleavage of the target strand. Mutation of the conserved residues of this domain abolishes the double strand breaking activity of the nuclease, leading to a nickase . These distinct features of Cpf1 nucleases might render them a very useful alternative genome manipulating tool to Cas9s. This contention was further strengthened by recent studies on the activity of Cpf1s on off-target sequences [28–30]. Genome-wide off-target analyses and targeted deep sequencing suggest that Cpf1 nucleases tolerate only one or two mismatches in contrast to SpCas9 that has been reported to tolerate 5–6 mismatches [28–30].
Zetsche et al. examined 16 members of the Cpf1 nuclease family. Two of them AsCpf1 and LbCpf1 were proved to mediate efficient genome editing in HEK293FT cells. In these experiments, exploring a limited set of four targets and non-homologous end joining (NHEJ) repair, the indel frequencies caused by AsCpf1 and LbCpf1 were actually higher than that of SpCas9 .
Here, we investigated the activity of AsCpf1 and LbCpf1 nucleases on a number of various targets in N2a mouse neuroblastoma cells exploiting two homology directed repair (HDR) pathways: single strand annealing (SSA) and homologous recombination (HR), in contrast to NHEJ repair, and compared their activity to various Cas9 proteins.
In order to monitor the activity of AsCpf1 and LbCpf1 nucleases we employed a fluorescence reporter assay using an interrupted-GFP expression cassette in a plasmid where the two fragments (GFP “halves”) of the GFP sequence are separated in a way to contain an overlapping 480 base-pairs long region (Additional file 1) . The target site for the nuclease is placed between the two halves and upon nuclease cleavage the generated double strand DNA break is repaired by single strand annealing (SSA) directed by the overlapping homologous sequences (Additional file 1). A presumably small fraction of the breaks might also be repaired by homologous recombination between two plasmids. The advantage of this assay, that we refer to as “GFxFP assay”, is that it allows measuring the ability of a nuclease to mediate HDR by cleaving a specific target while the cleavage efficiency is not influenced by factors like the epigenetic state of the targeted locus or the relative distance of the target from the homologous arms.
In order to test this assay system we used SpCas9 with two targets that had been tested earlier in either mammalian (PrP10) or bacterial (Sp1) cells (unpublished results). The interrupted GFP plasmids containing either the Sp1 or the PrP10 target (see Additional file 2: Table S1 for the target sequences) were cotransfected into N2a mouse neuroblastoma cells with a vector expressing SpCas9, the corresponding guideRNA (gRNA) and the fluorescent protein iRFP670 . The latter facilitated monitoring of transfection efficiency. As control, a vector expressing an inactive SpCas9 (dead, dSpCas9) was used for transfecting the cells. The number of fluorescent cells was counted two days after transfection. The Sp1 target resulted in 73 %, while the PrP10 target resulted in 54 % fluorescent cells in the transfected populations, respectively (Additional file 1). Surprisingly, the control population also showed a remarkable percentage, about 30 % GFP positive cells (Additional file 1 – EGxxFP). Apparently, in a fraction of the plasmids recombination between the homologous sequences occurred without nuclease cleavage.
Since we observed that the detected fluorescence varied from experiment to experiment (possibly reflecting the actual condition of the culture) but remained relatively stable within the parallel transfections, here and in further experiments we handled the samples to be compared in parallel. Also, we subtracted the corresponding background fluorescence from each sample. We found similar results with the pGF-ori-FP plasmid (data not shown) and hereafter used this plasmid in further experiments.
AsCpf1 and LbCpf1 nucleases are more efficient when the crRNAs are expressed from a plasmid rather than a PCR-based template
To facilitate an easier use of these Cpf1 nucleases against various targets, we cloned the As- and Lb-crRNAs driven by a human U6 promoter into the pY010 and pY016 plasmids , respectively, in such a way that the 3’ spacer can be precisely inserted between two sites of a type IIS restriction endonuclease (Esp3I) (Additional file 4). The created plasmids (pTE4396 for AsCpf1 and pTE4398 for LbCpf1) are deposited at Addgene (#74041 and #74042, respectively).
As- and LbCpf1 induce efficient repair of the GFP cassette using homologous sequences
Although the utility of As- and LbCpf1 nucleases may be less than that of SpCas9 they do demonstrate comparable efficiencies to three other Cas9 nucleases
Genomic DNA cleavage by As- and LbCpf1 nucleases can induce efficient HDR
In order to check whether the As- and LbCpf1 nucleases can effectively induce homology directed repair on some of the tested targets when these are in their genomic contexts, we exploited a homologous recombination donor molecule (pHRdonor-PrP1000) (Tálas et al., submitted manuscript). The donor molecule contains a promoterless GFP open reading frame (ORF) flanked by 1000 bp-long homologous arms to the PRNP gene. Upon targeted integration, the PrP ORF is replaced by the GFP ORF, where GFP expression will be driven by the promoter of the PRNP gene. The As- and LbCpf1 plasmids containing the four crRNAs targeting PRNP used in the previous experiment (Fig. 3a) were cotransfected with the donor plasmid and the number of GFP expressing cells were monitored. Interestingly, transfection of the promoterless donor plasmid with the inactive dSpCas9 expressing plasmid resulted in some GFP fluorescence when measured two days after transfection. We used this transient fluorescent signal hereafter in the experiments to compare the transfection efficiencies of different samples using identical donor plasmid. These transient signals decayed in about six to nine days, after that allowing the measurement of the GFP signal originating from the integrated GFP cassettes.
The results presented here significantly extend our knowledge on Cpf1 nucleases: these are the first demonstration that Cpf1 nucleases induce efficient DNA recombination in mammalian cells based on homologous sequences and exploring N2a cells – a different cell line – as compared to human HEK293FT cells used in the pioneering studies of Zetsche et al. Double strand DNA breaks are generally repaired by one of the main repair pathways, NHEJ or HR. The factors affecting the choice between the two pathways are not fully understood; however, they include both the types of the DNA break and of the host cells . SpCas9 nuclease has been reported to remain bound to the DNA target sites after cleavage for an extended period of time  that is likely to significantly affect the choice of repair pathways. It is not known if Cpf1 nucleases also remain bound to their DNA substrates after cleavage or not. Based on the available X-ray structure [25, 27, 40] and on their cleavage position, it is likely that a Cpf1 nuclease would not keep the cleaved DNA ends together after cleavage, even if it remains bound to the DNA. How this distinct feature of Cpf1 nucleases would affect the repair pathway choice, it is not known. Nevertheless, our data show that Cpf1 nucleases are able to induce efficient HR repair in mammalian cells.
During DNA break repair, SSA and HR repair pathways are initiated by identical mechanism and are being separated only after the step of extensive 3’ end resections of the DNA that also blocks the NHEJ repair pathways [38, 41]. Although the HR repair pathway, exploiting homology arms on a donor plasmid is employed more often, here, the SSA pathways are exploited to interrogate the activities of Cpf1 nucleases for the following reasons. HR repair is heavily influenced by the distances of the targets from the homology arm. To avoid this effect when a number of nucleases are compared, targets with the corresponding PAM sequences need to be found at nearly identical genomic positions. However, this is far from being a trivial task because several of these nucleases require at least three-nucleotide-long PAM sequences (As-, LbCpf1, Nm-, Sa-, StCas9). In addition, carrying out a HR assay takes about nine to fourteen days while conducting a GFxFP assay requires only four days that presents advantages.
Our results not only show that the two Cpf1 nucleases are as or more effective than the most widely used Cas9 counterparts of SpCas9, but also confirm the superior utility of SpCas9 and the somewhat lower activity of NmCas9 to that of Sa- and StCas9s. These results are consistent with our previous characterization of these Cas9 nucleases (data not shown).
Even more significant is the demonstration of the robust activity of Cpf1 nucleases in mammalian cells. In the experiments here (Fig. 3 and Additional file 2: Table S3), 19 sequences out of the targeted 20 were cleaved efficiently by both Cpf1 nucleases inducing SSA repair. The activities of Cas9 nucleases that are characterized so far vary more widely among various target sequences. A thorough characterization of Cpf1 nucleases’ target specificity, like in the case of SpCas9, may further lead to improve their effectiveness.
While examining only a limited number of target sequences, our results suggest that As- and LbCpf1 nucleases, although being less utile than SpCas9 in these experiments, are able to induce homology directed repair (SSA and HR) in N2a cells. They demonstrate especially robust activity to induce SSA in the plasmid-based GFxFP assay and are equally or more effective than the most frequently used orthogonal Cas9 counterparts of SpCas9. These results suggest an important position for the Cpf1 nucleases on the palette of effective genome engineering tools on the side of SpCas9.
Restriction enzymes, Klenow polymerase, T4 ligase and Pfu polymerase were purchased from Thermo Fischer Scientific. DNA oligonucleotides were acquired from Microsynth AG  or Sigma-Aldrich Co. All DNA constructs made were verified by sequencing by Microsynth AG. Plasmids were purified with GenElute HP Plasmid Miniprep kit (Sigma-Aldrich). Q5 polymerase was from New England BioLabs Inc. Dulbecco’s modified Eagle’s Medium, foetal bovine serum, Turbofect, penicillin and streptomycin were acquired from Thermo Fisher Scientific. The following plasmids were acquired from Addgene: piRFP670-N1 #45457 ; pX330-U6-Chimeric_BB-CBh-hSpCas9 #42230 ; pCAG-EGxxFP #50716 ; pY010 (pcDNA3.1-hAsCpf1) #69982 ; pY016 (pcDNA3.1-hLbCpf1) #69988 ; M-ST1cas #48669 ; and pSimpleII-U6-tracr-U6-BsmBI-NLS-NmCas9-HA-NLS(s) #47868, .
Vectors were constructed using standard molecular biology techniques. For detailed information see Additional file 7. The sequences of DNA oligonucleotides used in these studies are listed in Additional file 2.
Cell culturing and transfection
N2a cells (Neuro-2a mouse neuroblastoma cells, ATCC – CCL-131) were grown at 37 °C in a humidified atmosphere of 5 % CO2 in high glucose Dulbecco’s Modified Eagle medium (DMEM) supplemented with 10 % heat inactivated fetal bovine serum, 4 mM L-glutamine (Gibco), 100 units/ml penicillin and 100 μg/ml streptomycin.
Details of GFxFP assay
Cells cultured on 48-well plates, were seeded a day before transfection at a density of 3 × 104 cells/well. The next day, at around 40 % confluence, cells were transfected with plasmid constructs using Turbofect reagent, briefly as follows: 250 ng total plasmid DNA and 1 μl Turbofect were mixed in 50 μl serum free DMEM and the mixture was incubated for 30 min at room temperature prior adding to cells. Transfection medium was changed on the cells to fresh supplemented DMEM after 24 h of incubation. Three parallel transfections were made from each sample. A mCherry expression plasmid was cotransfected with each sample to monitor the transfection efficiency. Cells were analyzed by flow cytometry two days after transfection.
Details of genomic HDR assay
Cells cultured on 6-well plates were seeded a day before transfection at a density of 105 cells/well. The next day, at around 30 % confluence, cells were transfected with plasmid constructs using Turbofect reagent as follows: 4000 ng total plasmid DNA and 4 μl Turbofect was mixed in 400 μl serum free DMEM and the mixture was incubated for 30 min at room temperature prior adding to cells. Transfection medium on the cells was changed to fresh supplemented DMEM after 24 h of incubation. Three parallel transfections were made for each sample. Two days after transfection, cells were trypsinized and divided as follows: 10 % of the cells were seeded on new 6-well plates and 90 % of the cells were analyzed by flow cytometry. Seven days after transfection samples were measured again by flow cytometry.
Flow cytometry analysis was carried out on Attune Acoustic Focusing Cytometer (Applied Biosystems by Life Technologies) and on BD FACSCanto II [Becton Dickinson Immunocytometry Systems]. For data analysis Attune Cytometric Software was used. In all experiments, a total of 10,000 viable single cells were acquired and were gated based on side and forward light-scatter parameters. Cells expressing GFP and mCherry from a control plasmid were used to identify GFP and mCherry positive cells in the samples. The GFP signal was detected using the 488 nm diode laser for excitation and the 530/30 nm filter for emission. The mCherry signal was detected using the 488 nm diode laser for excitation and a 640LP filter for emission. The signal of iRFP670 (near-infrared fluorescent protein with emission maxima at 670 nm) was detected using the 633 nm diode laser for excitation and the 660/20 nm filter for emission.
Author’s response : We thank all three reviewers for accepting reviewing our manuscript and for their comments/recommendations that helped to improve the manuscript. We also thank all three reviewers for the thorough reviewing of the text; for spotting typos, omissions and inaccuracies, that helped us in editing and improving the manuscript.
Reviewer’s report 1
Eugene Koonin, NCBI, NLM, NIH, USA
In the article titled “Cpf1 nucleases demonstrate robust activity to induce DNA modification by exploiting homology directed repair pathways in mammalian cells”, Toth et al describe target cleavage and induction of homology-directed repair in human cells by two RNA-guided nucleases of the Cpf1 family. The results suggest that both Cpf1 proteins are capable of efficient target cleavage, being apparently less active than the most commonly used SpCas9 but about as active as other characterized Cas9 species. The work extends the existing knowledge on Cpf1, potentially an important, novel genome editing tool.
The work is somewhat preliminary in the sense that only two Cpf1 species have been studied, off-target effects were not assessed, and the comparison to SpCas9 cannot be considered reliable because for SpCas9, the guide RNA were designed whereas for the rest of the studied proteins, the guides were chosen randomly. It is hard to tell why the experiments were done in this non-uniform manner. It should not be difficult to use random guides with SpCas9. Without that simple experiment, the conclusion that Cpf1 is less effective than SpCas9 does not seem to be quite justified.
Author’s response: The reviewer is right that the work would be more comprehensive by studying more Cpf1 nucleases. However, Zetsche et al. showed that only these two Cpf1 species have considerable activity in mammalian (HEK293FT) cells. We agree that it is likely interesting to extend these studies on other Cpf1 species and by using more kinds of mammalian systems and we plan to do so.
He is also right in his comment on the differential ways we selected our targets using prediction tools only for SpCas9, which makes less fair the comparison of their natural activities. However, our attempt was different here. Although the sequence specificity of SpCas9 is not yet fully understood and the prediction of spacer efficiency is yet at an immature stage at present, SpCas9 is the only RNA-guided nuclease for which such prediction tools are available. Contrary, Cpf1 nucleases and the other Cas9s are much less studied and so far we know very little about their specificities as compared to SpCas9. Thus, from a practical point of view when one needs to decide which nucleases to choose to efficiently carry out genome modification experiments, currently, SpCas9 being supported with better spacer prediction may stand against the use of a Cpf1 nuclease where spacer sequence efficiency selection is not supported yet by spacer efficiency prediction tools, leaving the only option of random selection from possible targets. It is this situation where our experiments might provide valuable guidance. Nevertheless, when more knowledge is gained on Cpf1 nucleases based on which spacer efficiency prediction tools can be built, this issue may need to be revisited. We discuss and emphasize more this issue, highlighted by the reviewer, in our revised manuscript.
Reviewer’s report 2
Haruhiko Siomi, Keio University School of Medicine, Japan
Recently Cpf1 has been established as a class 2 CRISPR-Cas system that includes an effective single RNA-guided endonuclease (Zetsche et al. Cell 2015). This study compares the single strand annealing (SSA) activity of two members of the Cpf1 family, the AsCpf1 and the LbCpf1 with that of various Cas9 proteins. The authors find that both AsCpf1 and LbCpf1 appear to mediate SSA repair activity more efficiently than three of Cas9 counterparts do; SaCas9, StCas9 and NmCas9. However, SpCas9 mediates SSA repair more efficiently than the two Cpf1 proteins. These Cpf1 proteins are potentially powerful tools to edit genomes, but the paper falls short of clearly demonstrating the usefulness of the SSA activity mediated by Cpf1 proteins.
1. The authors should, at least, describe potential useful features of the SSA mediated by Cpf1 proteins for genome editing.”
Author’s response : SSA mediated repair can be useful when direct repeats are to be edited exploiting designer nucleases, among them Cpf1s. However, our point using SSA repair here is different. We aim to compare the ability of Cpf1 and Cas9 nucleases to induce homology directed repair (HDR), i.e. repair mechanisms based on homologous sequences involving either homologous recombination (HR) or single strand annealing (SSA).
During DNA break repair, SSA and HR repair pathways are initiated by identical mechanism and are being separated only after the step of extensive 3’ end resections of the DNA ends that also block the NHEJ repair pathway. Thus, the differing ability of Cas9 and Cpf1 nucleases (emerging from differences in the nature of the double strand breaks and in their interaction with the target DNA) to allow the engagement of the repair system for 3’ end resections may be well revealed by monitoring either the SSA or the HR repair pathways.
Although the HR repair pathway, exploiting homology arms on a donor plasmid is employed more often, here, the SSA pathway is also exploited to interrogate the activities of Cpf1 nucleases to induce HDR for the following reason: HR repair is heavily influenced by the distances of the targets from the homology arm. To avoid the contribution of this effect when two or more nucleases (with different PAM sequences) are compared, the targets need to be found at nearly identical genomic positions in relation to the homology arms. For this, the corresponding PAM sequences need to be found at nearly identical positions. However, this is far from being trivial because several of these nucleases require at least three to five nucleotide-long PAM sequences (As- and LbCpf1 three, Nm-, Sa-, StCas9 four to five), which are rather infrequent.
Nevertheless, we also monitored here how Cpf1 nucleases induce HR repair, although a very strict comparison cannot be carried out with Cas9s from the above discussed reasons.
Author’s response : In respect to Figs. 3 and 4 an active Cpf1 nuclease without crRNA, whereas on Fig. 5 a dead SpCas9 was used as negative control. We did generate the corresponding dead As- and LbCpf1 nucleases; however, at the time the experiments were carried out we had not characterized them whether they are really fully inactive or not. Now, we detect no difference between these three types of controls, dead Cpf1, SpCas9 and active As- and LbCpf1s without crRNAs, therefore, they can be used in an interchangeable manner.
3. It seems the authors take the increase in the number of GFP positive cells as read-out of the increase in the efficiency of cleavage. But the expression of GFP in their assay involves cleavage and homologous recombination. These are in essence totally different steps.
Author’s response: We attempted to discern the ability of these nucleases to “induce homology directed repair”. There is no experimental tool currently available to assess directly the in-cell cleavage activity of a nuclease. The closest equivalent may be the BLESS (direct in situ breaks labeling, enrichment on streptavidin and next-generation sequencing) method (N Crosetto et al., Nature Methods, 2013) that captures the double strand breaks that are present at a given moment in the cells. The “efficiency” of a nuclease is generally inferred based on detecting the outcome, generally the alteration in the DNA sequence as a direct result of the cell DNA repair mechanism that is measured by various readouts. As such, the cleavage activities may be different than assessed by relaying on the outcome of different repair mechanisms. Plausibly, this is the point the reviewer refers to - and it is the point of the manuscript as well. We try to assess here the ability of these nucleases to induce homology directed repair, SSA and HR, contrary of to induce NHEJ-mediated repair. To avoid confusions, as pointed by the reviewer, we have changed the wording of corresponding lines in the manuscript to clarify these points.
Reviewer’s report 3
Jean-Yves Masson, Universitè Laval, Canada
In this manuscript, E. Toth and colleagues aim to characterize the efficiency of two Cpf1 family members (AsCpf1 and LbCpf1) in genome engineering. These nucleases have been characterized previously for indel frequencies by non-homologous end joining (Zetsche et al., Cell 163, 759-771). However, their activities on homology directed repair and single-strand annealing remained unclear. This is an important question in a fast growing field where the most efficient tools must be used for genome engineering. They also used another cell line, N2a cells, which is different from the studies of Zetsche et al.
Overall, this nice study suggests that AsCpf1 and LbCpf1 are alternative tools to Cas9. In fact, they are more efficient to induce SSA than Cas9 homologs. Moreover, these nucleases are very efficient, showing cleavage of 19 out of 20 loci.
Figure 1 and Additional file 1: Figure S1B. It is surprising that ~ 30 % of the control cells display GFP fluorescence without any nuclease cleavage. What is the level of endogenous DSBs in the N2a cells? This can be monitored with immunofluorescence staining with g-H2AX and 53BP1.
Author’s response: This value is actually 10 to 20 % when the GF-chl-FP constructs are used (Fig. 1 b) where recombination events that possibly occur in Escherichia coli during plasmid-generation is minimized. Such background fluorescence obtained in N2a cells can also be observed at a similar extent in HEK293T and at a lower extent in HeLa cells (Additional file 3 ). The higher fluorescence apparent also after nuclease cleavage may indicate that the complex process starting from the SSA repair of the DNA break and extending to the detection of the fluorescent signal of the expressed GFP protein, may somehow more efficiently take place in N2a cells. However, a contribution from a somewhat higher level of endogenous DSBs neither can be ruled out, as also evidenced in Additional file 3 .
What is the percentage of cells that are transfected with mCherry? It is not clear whether SSA is low because of the transfection efficiency. The transfection efficiency for each experiment should be shown.
Author’s response: The results that are shown on Fig. 2 c are normalized to the corresponding mCherry positive population in case of each sample. In the revised version of the manuscript we added this information, previously present in the materials and methods section, to the main text too in order to clarify this issue. We did a few attempts to increase the efficiency of the Cpf1 nucleases using the PCR products, however, without much success. The transfection efficiencies were relatively similar in these experiments: 55.9 % +/- 10.5 % when the plasmid was used, whereas 45.0 % +/- 8.4 %, when the PCR product. Now we indicate these values in the corresponding figure legend for clarity.
The authors should make a clear distinction between SSA and HDR. SSA is a RAD52-dependent pathway and HDR relies mostly on BRCA1-PALB2-BRCA2 and RAD51. For instance, SSA is monitored in Fig. 4, but the authors mention in the text that « Cpf1 nucleases were equally or more effective in inducing HDR on average than the three SpCas9 counterparts ». This is confusing with Fig. 5. In Fig. 5, a typical HDR assay is employed (with an homologous recombination donor). The authors should carefully edit the manuscript to distinguish SSA assays from HDR assays.
Author’s response: There is some inconsistency in the literature regarding the terminologies used. We refer as homology directed repair (HDR) to the repair pathways that are based on homologous sequences, involving both homologous recombination (HR) and single strand annealing (SSA), as it is frequently used in the literature (A Ciccia and SJ Elledge, Mol Cell, 2010; G Soria and G Almouzni, Cell Cycle, 2013; F Larminat et al., Oncogene, 2002; M Keimling and L Wiesmüller, Carcinogenesis, 2009). We also noticed papers that use HR and HDR in the opposite sense [HR involving both HDR and SSA pathways (SM Howard, DA Yanez and JM Stark, PLoS Genet, 2015)] or which use other terms, such as conservative or inter molecular HDR and homology repair [also denoted as HR, (L Cong et al., Science, 2013)] for the pathway relying mostly on BRCA1-PALB2-BRCA2 and RAD51 [as opposed to homologous recombination between direct repeats (RAD52-dependent]). This is clearly indicated now at the end of the background section.
Also, in this particular experiment, SaCas9, StCas9 and NmCas9 should be used as controls. This would help the reviewer to conclude whether they are indeed better tools for HDR.
Author’s response: We had performed HR experiments with Sa- and StCas9s (results now presented in Additional File 6 ); however, the target positions of the Cas9s generally do not match those of the Cpf1s, which in our judgment precludes their use for a fair comparison (relative distance from the cleavage position to the homology arms might affect the efficiency of homologous repair-based integration). This was one of the reasons we preferred to use the GFxFP plasmid system where the infrequent presence of the different PAM sequences doesn’t impede the comparison of the efficiencies of different nucleases.
To make a more comprehensive study, the authors should use also integrated reporter cassettes.
Author’s response: We also considered the approach of using an integrated GFP reporter cassette. Unfortunately, in contrast to SpCas9 having a two nucleotide-long PAM restriction (GG), Lb- and As-Cpf1 and Sa-, St-and NmCas9 nucleases have longer, 3 to 5 nucleotide-long PAM sequences and thus, much less targetable sequences are available. As a result, Lb- and AsCpf1 (PAM: TTT) nucleases have only two targets, while StCas9 (PAM: AGAAW) and NmCas9 (PAM: GATT) have no target in the GFP reporter construct. Considering mCherry, it has only one position that has more than two consecutive T nucleotides. The construction of a new cell line harboring an integrated reporter cassette of which sequences are altered in order to contain a reasonable number of PAM sequences in question was out of the time frame of this study.
They should also make an attempt to discuss off-target events in light of the two recent studies below:
Daesik Kim, Jungeun Kim, Junho K Hur, Kyung Wook Been, Sun-heui Yoon & Jin-Soo Kim. Genome-wide analysis reveals specificities of Cpf1. Nature Biotechnology, June 2016 doi:10.1038/nbt.3609.
Junho K Hur, Kyoungmi Kim, Kyung Wook Been, Gayoung Baek, Sunghyeok Ye, Junseok W Hur, Seuk-Min Ryu, Youn Su Lee & Jin-Soo Kim. Targeted mutagenesis in mice by electroporation of Cpf1 ribonucleoproteins. Nature Biotechnology, June 2016 doi:10.1038/nbt.3596.
Author’s response: We included a short discussion about the fidelity of these Cpf1 nucleases based on the more recent publications in the Background section.
I acknowledge the spirit of the authors, for deposition of AsCpf1 and LbCpf1 plasmids at Addgene.
Acidaminococcus sp. BV3L6
CMV early enhancer/chicken β actin promoter
Chloramphenicol resistance gene expression cassette
CRISPR from Prevotella and Francisella 1
Clustered regularly interspaced short palindromic repeat
Dead or nuclease inactive Streptococcus pyogenes Cas9
Dulbecco’s modified Eagle Medium
Francisella novicida U112
Green fluorescent protein
Homology directed repair
Near-infrared fluorescent protein with emission maxima at 670 nm
Lachnospiraceae bacterium ND2006
Multiple cloning site
Mouse doppel protein
Mouse prion protein
Mouse shadoo protein
Neuro-2a mouse neuroblastoma cell line
Non-homologous end joining
Open reading frame
Protospacer adjacent motif
- PRND :
Gene of doppel protein
- PRNP :
Gene of prion protein
- SPRN :
Gene of shadoo protein
Single strand annealing
Simian vacuolating virus 40
Transactivating CRISPR RNA
We thank Ildikó Szücsné Pulinka for her excellent laboratory assistance, Barbara Vodicska, Sarah Krausz and Antal Nyeste for contribution to plasmid construction.
This study was supported by the Hungarian Scientific Research Fund [K-82090].
Availability of data and material
The pAsCpf1-crRNA (pTE4396) and pLbCpf1-crRNA(pTE4398) plasmids are deposited at Addgene (#74041 and #74042, respectively).
EW designed the study, ET and NW carried out plasmid constructions, cell culturing and transfections. KH and PIK contributed to plasmid construction and AT participated in flow cytometry analysis. PB contributed to cell culturing, transfection and flow cytometry analysis. All authors contributed to the evaluation of the data. ET, NW, EF and EW wrote the paper. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Consent for publication
Ethical approval and consent to participate
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Vercoe RB, Chang JT, Dy RL, Taylor C, Gristwood T, Clulow JS, Richter C, Przybilski R, Pitman AR, Fineran PC. Cytotoxic chromosomal targeting by CRISPR/Cas systems can reshape bacterial genomes and expel or remodel pathogenicity islands. PLoS Genet. 2013;9:e1003454.View ArticlePubMedPubMed CentralGoogle Scholar
- Westra ER, Buckling A, Fineran PC. CRISPR-Cas systems: beyond adaptive immunity. Nat Rev Microbiol. 2014;12:317–26.View ArticlePubMedGoogle Scholar
- Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E. A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science. 2012;337:816–21.View ArticlePubMedGoogle Scholar
- Cong L, Ran FA, Cox D, Lin S, Barretto R, Habib N, Hsu PD, Wu X, Jiang W, Marraffini LA, et al. Multiplex genome engineering using CRISPR/Cas systems. Science. 2013;339:819–23.View ArticlePubMedPubMed CentralGoogle Scholar
- Mali P, Yang L, Esvelt KM, Aach J, Guell M, DiCarlo JE, Norville JE, Church GM. RNA-guided human genome engineering via Cas9. Science. 2013;339:823–6.View ArticlePubMedPubMed CentralGoogle Scholar
- Jinek M, East A, Cheng A, Lin S, Ma E, Doudna J. RNA-programmed genome editing in human cells. Elife. 2013;2:e00471.View ArticlePubMedPubMed CentralGoogle Scholar
- Wiedenheft B, Sternberg SH, Doudna JA. RNA-guided genetic silencing systems in bacteria and archaea. Nature. 2012;482:331–8.View ArticlePubMedGoogle Scholar
- Fineran PC, Charpentier E. Memory of viral infections by CRISPR-Cas adaptive immune systems: acquisition of new information. Virology. 2012;434:202–9.View ArticlePubMedGoogle Scholar
- Sashital DG, Wiedenheft B, Doudna JA. Mechanism of foreign DNA selection in a bacterial adaptive immune system. Mol Cell. 2012;46:606–15.View ArticlePubMedPubMed CentralGoogle Scholar
- Makarova KS, Wolf YI, Alkhnbashi OS, Costa F, Shah SA, Saunders SJ, Barrangou R, Brouns SJ, Charpentier E, Haft DH, et al. An updated evolutionary classification of CRISPR-Cas systems. Nat Rev Microbiol. 2015;13:722–36.View ArticlePubMedGoogle Scholar
- Makarova KS, Aravind L, Wolf YI, Koonin EV. Unification of Cas protein families and a simple scenario for the origin and evolution of CRISPR-Cas systems. Biol Direct. 2011;6:38.View ArticlePubMedPubMed CentralGoogle Scholar
- Makarova KS, Wolf YI, Koonin EV. The basic building blocks and evolution of CRISPR-CAS systems. Biochem Soc Trans. 2013;41:1392–400.View ArticlePubMedGoogle Scholar
- Barrangou R, Marraffini LA. CRISPR-Cas systems: Prokaryotes upgrade to adaptive immunity. Mol Cell. 2014;54:234–44.View ArticlePubMedPubMed CentralGoogle Scholar
- Shalem O, Sanjana NE, Hartenian E, Shi X, Scott DA, Mikkelsen TS, Heckl D, Ebert BL, Root DE, Doench JG, et al. Genome-scale CRISPR-Cas9 knockout screening in human cells. Science. 2014;343:84–7.View ArticlePubMedGoogle Scholar
- Fineran PC, Dy RL. Gene regulation by engineered CRISPR-Cas systems. Curr Opin Microbiol. 2014;18:83–9.View ArticlePubMedGoogle Scholar
- Dominguez AA, Lim WA, Qi LS. Beyond editing: repurposing CRISPR-Cas9 for precision genome regulation and interrogation. Nat Rev Mol Cell Biol. 2015;17:5–15.View ArticlePubMedPubMed CentralGoogle Scholar
- Gilbert LA, Larson MH, Morsut L, Liu Z, Brar GA, Torres SE, Stern-Ginossar N, Brandman O, Whitehead EH, Doudna JA, et al. CRISPR-mediated modular RNA-guided regulation of transcription in eukaryotes. Cell. 2013;154:442–51.View ArticlePubMedPubMed CentralGoogle Scholar
- Qi LS, Larson MH, Gilbert LA, Doudna JA, Weissman JS, Arkin AP, Lim WA. Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell. 2013;152:1173–83.View ArticlePubMedPubMed CentralGoogle Scholar
- Mali P, Aach J, Stranges PB, Esvelt KM, Moosburner M, Kosuri S, Yang L, Church GM. CAS9 transcriptional activators for target specificity screening and paired nickases for cooperative genome engineering. Nat Biotechnol. 2013;31:833–8.View ArticlePubMedPubMed CentralGoogle Scholar
- Zalatan JG, Lee ME, Almeida R, Gilbert LA, Whitehead EH, La Russa M, Tsai JC, Weissman JS, Dueber JE, Qi LS, et al. Engineering complex synthetic transcriptional programs with CRISPR RNA scaffolds. Cell. 2015;160:339–50.View ArticlePubMedGoogle Scholar
- Chavez A, Scheiman J, Vora S, Pruitt BW, Tuttle M, P R Iyer E, Lin S, Kiani S, Guzman CD, Wiegand DJ, et al. Highly efficient Cas9-mediated transcriptional programming. Nat Methods. 2015;12:326–8.View ArticlePubMedPubMed CentralGoogle Scholar
- Anders C, Niewoehner O, Duerst A, Jinek M. Structural basis of PAM-dependent target DNA recognition by the Cas9 endonuclease. Nature. 2014;513:569–73.View ArticlePubMedPubMed CentralGoogle Scholar
- Esvelt KM, Mali P, Braff JL, Moosburner M, Yaung SJ, Church GM. Orthogonal Cas9 proteins for RNA-guided gene regulation and editing. Nat Methods. 2013;10:1116–21.View ArticlePubMedPubMed CentralGoogle Scholar
- Zetsche B, Gootenberg JS, Abudayyeh OO, Slaymaker IM, Makarova KS, Essletzbichler P, Volz SE, Joung J, van der Oost J, Regev A, et al. Cpf1 is a single RNA-guided endonuclease of a class 2 CRISPR-Cas system. Cell. 2015;163:759–71.View ArticlePubMedGoogle Scholar
- Dong D, Ren K, Qiu X, Zheng J, Guo M, Guan X, Liu H, Li N, Zhang B, Yang D, et al. The crystal structure of Cpf1 in complex with CRISPR RNA. Nature. 2016;532:522–26.
- Fonfara I, Richter H, Bratovic M, Le Rhun A, Charpentier E. The CRISPR-associated DNA-cleaving enzyme Cpf1 also processes precursor CRISPR RNA. Nature. 2016;532:517–21.View ArticlePubMedGoogle Scholar
- Yamano T, Nishimasu H, Zetsche B, Hirano H, Slaymaker IM, Li Y, Fedorova I, Nakane T, Makarova KS, Koonin EV, et al. Crystal Structure of Cpf1 in Complex with Guide RNA and Target DNA. Cell. 2016;165:949–62.
- Kim D, Kim J, Hur JK, Been KW, Yoon SH, Kim JS. Genome-wide analysis reveals specificities of Cpf1 endonucleases in human cells. Nat Biotechnol. 2016;34:863–68.
- Kleinstiver BP, Tsai SQ, Prew MS, Nguyen NT, Welch MM, Lopez JM, McCaw ZR, Aryee MJ, Joung JK. Genome-wide specificities of CRISPR-Cas Cpf1 nucleases in human cells. Nat Biotechnol. 2016;34:869–74.
- Hur JK, Kim K, Been KW, Baek G, Ye S, Hur JW, Ryu SM, Lee YS, Kim JS. Targeted mutagenesis in mice by electroporation of Cpf1 ribonucleoproteins. Nat Biotechnol. 2016;34:807–8.PubMedGoogle Scholar
- Mashiko D, Fujihara Y, Satouh Y, Miyata H, Isotani A, Ikawa M. Generation of mutant mice by pronuclear injection of circular plasmid expressing Cas9 and single guided RNA. Sci Rep. 2013;3:3355.View ArticlePubMedPubMed CentralGoogle Scholar
- Shcherbakova DM, Verkhusha VV. Near-infrared fluorescent proteins for multicolor in vivo imaging. Nat Methods. 2013;10:751–4.View ArticlePubMedPubMed CentralGoogle Scholar
- Ran FA, Cong L, Yan WX, Scott DA, Gootenberg JS, Kriz AJ, Zetsche B, Shalem O, Wu X, Makarova KS, et al. In vivo genome editing using Staphylococcus aureus Cas9. Nature. 2015;520:186–91.View ArticlePubMedPubMed CentralGoogle Scholar
- Xu K, Ren C, Liu Z, Zhang T, Li D, Wang L, Yan Q, Guo L, Shen J, Zhang Z. Efficient genome engineering in eukaryotes using Cas9 from Streptococcus thermophilus. Cell Mol Life Sci. 2014;72:383–99.View ArticlePubMedGoogle Scholar
- Hou Z, Zhang Y, Propson NE, Howden SE, Chu LF, Sontheimer EJ, Thomson JA. Efficient genome engineering in human pluripotent stem cells using Cas9 from Neisseria meningitidis. Proc Natl Acad Sci U S A. 2013;110:15644–9.View ArticlePubMedPubMed CentralGoogle Scholar
- Doench JG, Fusi N, Sullender M, Hegde M, Vaimberg EW, Donovan KF, Smith I, Tothova Z, Wilen C, Orchard R, et al. Optimized sgRNA design to maximize activity and minimize off-target effects of CRISPR-Cas9. Nat Biotechnol. 2016;34:184–91.
- Chari R, Mali P, Moosburner M, Church GM. Unraveling CRISPR-Cas9 genome engineering parameters via a library-on-library approach. Nat Methods. 2015;12:823–6.View ArticlePubMedGoogle Scholar
- Daley JM, Sung P. 53BP1, BRCA1, and the choice between recombination and end joining at DNA double-strand breaks. Mol Cell Biol. 2014;34:1380–8.View ArticlePubMedPubMed CentralGoogle Scholar
- Richardson CD, Ray GJ, DeWitt MA, Curie GL, Corn JE. Enhancing homology-directed genome editing by catalytically active and inactive CRISPR-Cas9 using asymmetric donor DNA. Nat Biotechnol. 2016;34:339–44.View ArticlePubMedGoogle Scholar
- Gao P, Yang H, Rajashankar KR, Huang Z, Patel DJ. Type V CRISPR-Cas Cpf1 endonuclease employs a unique mechanism for crRNA-mediated target DNA recognition. Cell Res. 2016;26:901–13.View ArticlePubMedGoogle Scholar
- Shrivastav M, De Haro LP, Nickoloff JA. Regulation of DNA double-strand break repair pathway choice. Cell Res. 2008;18:134–47.View ArticlePubMedGoogle Scholar
- Microsynth, A. http://www.microsynth.ch/. Accessed 5 Sept 2016.