Successful CRISPR/Cas9 mediated homologous recombination in a chicken cell line

Background: CRISPR/Cas9 system is becoming the dominant genome editing tool in a variety of organisms. CRISPR/Cas9 mediated knock out has been demonstrated both in chicken cell lines and in chicken germ cells that served to generate genetically modified birds. However, there is limited data about CRISPR/Cas9 dependent homology directed repair (HDR) for avian, even in cell culture. Few attempts have been made with integrations in safe harbor loci of chicken genome that induces constitutive expression of the inserted gene. Gene expression under an endogenous promoter would be more valuable than under a constitutive exogenous promoter, as it allows the gene expression to be tissue-specific. Methods: Three gRNAs were chosen to target chicken 3’-untranslated region of GAPDH gene. Cas9-mediated activity in the targeted locus for the gRNAs in DF-1 cells was estimated by T7E1 assay. To edit the locus, the HDR cassette was added along with CRISPR/Cas9. The inserted sequence contained eGFP in frame with a GAPDH coding sequence via P2A and Neomycin resistance gene ( neoR) under cytomegalovirus promoter. Correct integration of the cassette was confirmed with fluorescent microscopy, PCR analysis and sequencing. Enrichment of modified cells was done by G418 selection. Efficiency of integration was assessed with fluorescence activated cell sorting (FACS). Results: We have established a CRISPR/Cas9 system to target an endogenous locus and precisely insert a gene under endogenous control. In our system, we used positive and negative selection to enrich modified cells and remove cells with undesirable insertions. The efficiency of CRISPR/Cas9-mediated HDR was increased up to 90% via G418 enrichment. We have successfully inserted eGFP under control of the chicken GAPDH promoter. Conclusions: The approach can be used further to insert genes of interest under control of tissue-specific promoters in primordial germ cells in order to produce genetically modified birds with useful for biotechnological purposes features.

. The CRISPR/Cas9 tool is a novel instrument compared with TALEN and some aspects of successful targeting with the Cas9 nuclease still need to be elucidated for avian species. However, the ability not only to knock-out genes, but also to induce a tissue-specific expression of a gene of interest without destroying its endogenous locus would be beneficial.
Here, we show the system to precisely insert a gene under the control of an endogenous promoter and select the cells with the successful integration. The result indicates that this system can be used as a tool for chicken genome editing.

Construction of expression vectors
We used human codon-optimized Cas9 (hCas9) as it has been previously demonstrated that the optimized Cas9 works in chicken cells and there was no need to synthesize chicken codonoptimized nuclease [

Targeting vector design
The targeting vector was designed based on the plasmid LSL-Cas9-Rosa26TV (#61408, Addgene). Homology regions of 999bp and 3093bp for left and right arms respectively were amplified from the genomic DNA of chicken cell line DF1 by PCR, and cloned using MauBI and PmeI restriction sites for the left arm, and SgrDI and AscI for the right arm respectively (DF-1 genome is yet to be sequenced, common chicken genomic data is available here). Left and right homology regions in the shuttle flank the P2A-eGFP sequence, where eGFP is enhanced green fluorescent protein. Coding sequence of the left arm was in frame with P2A-eGFP in order to provide the gene transcription under the control of endogenous promoter. Neomycin resistance (neo) gene was cloned after eGFP under a constitutive cytomegalovirus (CMV) promoter for positive selection of cells with the desired insertion. The total length of the inserted sequence between two homology arms in the shuttle was 3259bp. Diphteria Toxin Fragment A (DTA) coding sequence was inserted in the shuttle after the right arm under a PGK promoter for negative selection ( Figure 1A). The left homology arm was amplified by PCR with the following primers: 5'-TTGTTGACCTGACCTGCCGTCTGGAG-3' and 5'-CTCCTTGGATGCCATGTGGACCATCAAG-3'; the right homology arm was amplified with primers 5'-CCCTTTGT TGGAGCCCCTGCTCTTC-3' and 5'-GAGCCCTGTATCTT CCTTGCACAGACC-3'. The primer were designed in Primer-BLAST and synthesized by Evrogen.
Length of the right homology arm is longer in order to increase HDR due to high repeat content in 3'URT region. Chicken right and left homology arms are separated by a stretch of 42bp in the beginning of 3'UTR that is targeted by the gRNAs ( Figure 1B). Thus, CRISPR/Cas9 cutting of successfully edited genomic DNA is prevented due to the targeted sequence not being present in the vector for HDR.

gRNA selection
We searched for gRNAs to target the GAPDH locus in the 3'UTR within 50bp near the stop codon of the gene. We sequenced exon 10 of GAPDH and the beginning of the 3'UTR. 100bp region of the chicken GAPDH around the stop-codon was analyzed on the ChopChop server for gRNA design. We selected three gRNAs in the locus (Table 1, Figure 1B).

Amendments from Version 1
The title for Figure S2.III has been modified from "Transgene copies per genome" to "Transgene copies per GAPDH allele". The consistent changes have been done with the figure description.
All guide RNAs or sgRNA were written as gRNA through the paper.
We rounded the number 88% as 90% in the page 7 and the discussion section.
In the Discussion part we have mentioned about an additional advantage to express a transgene from endogenous promoter.  These gRNAs had a few predicted off-target sites (Table S1). None of the off-target sites were present in a known coding sequence of the Gallus gallus genome.

Cell transfection
Cells were co-transfected with 4µg DNA plasmid mix for 6-well plates or 1µg for 24-well plates using TurboFect transfection reagent (ThermoFisher; Waltham MA, USA). The plasmid mix contained the Cas9 plasmid, the gRNA plasmid, the RFP plasmid and the linear HDR shuttle. We used the equimolar ratio of all components: pCas9:pgRNA:pRFP:HDR shuttle. After 72 hours post transfection, cells were analyzed by fluorescence microscopy. Five -seven technical repeats were made for every experiment, which were used for genome analysis, G418 selection and for flow cytometry analysis.

Selection with geneticin
Titration of geneticin (G418, Gibco) on the DF-1 cell line had been performed before selection and the optimal concentration 500ng/µl was chosen. Cells were cultured with geneticin for 10 days with daily medium change. Geneticin-resistant cells were analyzed with PCR to confirm the correct insert. Fluorescence microscopy and FACS were used to visualize and define the percentage of eGFP-positive cells respectively.

Cell viability and cytotoxicity assays
For evaluation of an appropriate concentration of geneticin, CellTiter-Blue® Cell Viability Assay (Promega, Fitchburg WI, USA) was used. The principle of the assay is based on conversion of resazurin to resorufin by metabolically active cells that results in the generation of a fluorescent shift from 605 nm to 573 nm. Thus, the produced fluorescence is proportional to the number of viable cells. 48-well assay plate containing cultured cells with media was set up. Three technical repeats selection after successful modification. The product is only amplified in case of an unmodified genome. This happens due to the reverse primer aligning to the 50bp area of 3'UTR that was targeted by gRNAs, but it does not align to the HDR vector itself.
FACS analysis FACS analysis (Accuri™ C6 Flow Cytometer, BD Biosciences; San Jose CA, USA) was used to estimate the proportion of eGFP-positive cells.
Off-target analysis Potential CRISPR/Cas9 off-target sites for selected guides are presented in Table S1. The off-targets were predicted by ChopChop tool. All off-target sites were located in introns or in intergenic loci.
In vitro Cas9 cleavage Primers for in vitro gRNA synthesis: In order to make dsDNA template for RNA transcription the two following oligonucleotides were annealed.
-CRISPR R: (common primer for all targets, it contains the sequence for RNA scaffold synthesis) -the oligonucleotide was common for synthesis of all DNA templates coding a gRNA: The oligonucleotide synthesis was ordered from Evrogen.

PCR analysis
To confirm HDR cassette integration several pairs of primers were used ( Figure S1D). The primers were designed based on schematic sequence of the edited locus in Figure S1D, checked in BLASTn for specificity and synthesised by Evrogen.
Primers forward: 5'-GACCATTTCGTCAAGCTTGTTTCC-3' and reverse: 5'-GATCAGTTTCTATCAGCCTCTCCCAC-3' amplify the area surrounding the site of insertion ( Figure S1D, primers pointed as 1' and 1''). PCR product in the case of insertion had a length of 3851bp. Amplified product from the endogenous locus without an integration had a length of 635bp.

In vitro Cas9 digestion
In vitro digestion was made by Cas9 Streptococcus pyogenes (S. pyogenes; New England Biolabs) according to the protocol from the manufacturer. PCR product for analysis was amplified with following primers: forward 5'-GACCATTTCGTCAAGC TTGTTTCC-3'and reverse 5'-GATCAGTTTCTATCAGCCTCT CCCAC-3'.

Probes description
We designed three kinds of probes. Probe 1 -a reference probe (VIC) located away from the editing site to count all genomic copies ( Figure S2I). GAPDH was used as a reference, having a single copy per genome [Weiskirchen et al., 1993].
Probe 2 -a (FAM) probe, located in the inserted sequence ( Figure S2I). Probe 3 -a (VIC) probe located in DTA gene of the cassette that does not insert into the locus or kill the cells in case of insertion ( Figure S2I). The nucleotide sequence of the probes and primers are shown in the Table 2.

Samples description.
The following samples were used: -DNA from DF1 cells transfected with linearised cassette only (negative control). The sample was taken in amount 20ng.
-DNA from geneticin enriched DF1 cells, isolated 1 month after HDR. The sample was taken in amount -1ng and 20ng.
-Water was used as a no template control (NTC) to rule out cross-contamination.
We set up two repeats for each amount of DNA.
Cassette probe was used to determine the copy number of transgene in DF1 cells. GAPDH probe was used as the reference.

ddPCR reaction preparation.
The following reagents were mixed in 96 plate to make reaction: ddPCR Supermix for Probes (#186-3026, Bio-Rad); 10 U of EcoRI (#R3101S, New England BioLabs); Genomic DNA (DNA dilutions 1ng and 20ng were selected based on preliminary experiments); The total volume of the reaction was 20µl.
Droplets were generated with 20µl of the premixed reaction and a QX200 Droplet Generator according to the manufacturer's instructions (Bio-Rad) and transferred to a 96-well PCR plate for standard PCR on a CFX96 Touch™ Real-Time PCR Detector system (Bio-Rad).
The following cycling programs were used: 1) 95°C for 10 min; 2) 95°C for 30 s; 3) 59-63 °C (in depends on pair of primers and probe) for 1 min; repeat steps 2 and 3 for 40 times; Optimal annealing temperature was determined empirically for each pair of primers and probe with a temperature gradient. After PCR amplification, each droplet provides an independent fluorescent positive or negative signal indicating the target DNA was present or not. The droplets were analysed with a QX200 Droplet Reader (Bio-Rad) with the selected option "absolute quantification". Positive and negative droplets are counted for each samples, and the software calculates the concentration of target DNA as copies per microliter.
An appropriate threshold between the positive and negative droplets was applied manually based on the NTC wells. ddPCR software reads the positive and negative droplets in each sample and plots the fluorescence droplet by droplet. The fraction of positive droplets determines the concentration of the target in the sample. Software calculates the concentration of target DNA as copies per microliter. Then the copy number of an unknown target is calculated relative to a known reference. In our case we estimated the copy number of inserted sequence relative to GAPDH gene. The confidence intervals for each well are calculated by QuantaSoft based on Poisson distribution.
The formula used for Poisson modeling is: where p = fraction of positive droplets. The primers and probes were designed in Primer-BLAST. Primers were synthesised by Evrogen. Probes were ordered from Syntol (Moscow, Russia)

Results
In the current research we have studied CRISPR/Cas9mediated homology directed repair in an endogen locus for expression of an integrated gene under the control of the endogenous promoter.
The GAPDH locus was chosen as a commonly expressed constitutive gene. In order to make the expression of our inserted gene be controlled by the promoter, we decided to insert the gene just after the coding sequence of GAPDH ( Figure 1A). To target the 3'UTR of chicken GAPDH we selected and designed three gRNAs. Before starting experiments on a chicken cell line, all selected gRNAs were tested in vitro using recombinant Cas9 S. pyogenes. Cas9 in complex with one of the three gRNAs made cuts and produced lengths of cleaved products corresponded to the expected lengths ( Figure S3).

Targeting 3'UTR of chicken GAPDH by Cas9
In order to test the effectiveness of targeting endogenous 3'UTR of GAPDH with the selected gRNAs in chicken cells, Cas9, gRNA plasmids, combined with the RFP plasmid, were co-transfected in DF-1 cells. Cas9 expression vector without a gRNA was used as the negative control. RFP expression was estimated at 72h after transfection ( Figure S4). Figure S2 is represented by seven technical repeats. The average level of transfection efficiency was more than 45%.
Genomic DNA was extracted from the cells, and T7 endonuclease I (T7EI) assay demonstrated that only gRNA2 in complex with Cas9 had activity in DF1 cells with the targeting rate around 1.8% ( Figure 2).

Homologous recombination in GAPDH locus
In order to obtain cells with the insertion we co-transfected DF-1 cells with plasmids encoding Cas9, gRNA, RFP and cassette for homology directed repair. As a negative control we added Cas9, RFP and cassette without any gRNAs. At 72h after transfection we observed 0.5% GFP-positive cells of transfected cells in the experimental group ( Figure 3). Figure 3 is represented by five technical repeats. In the vector for HDR, eGFP does not have its own promoter and can be expressed only in the case of the correct insertion in frame with the GAPDH gene. The genomic DNA was extracted for PCR analysis. The analysis confirmed the presence of the expected integration ( Figure S1A and Figure S6B).

Enrichment of successfully edited cells with geneticin selection
Applying drug selection in combination with CRISPR/Cas9 allowed us to select and grow colonies carrying the modification. Based on the survival curve ( Figure S5) we added 500 ng/µl of geneticin (G418) at 72h after transfection. The appropriate concentration of the drug was selected after titration in the DF-1 cell line ( Figure S5). Single eGFP-positive cells had developed in colonies after 10 days of incubation on G418. RFP fluorescence had vanished due to its transient expression ( Figure 4). Figure 4 is representative of five technical repeats.
One month after the drug selection the enriched cells were analyzed by PCR and FACS. PCR analysis confirmed the absence of cells without modification ( Figure S1C). FACS analysis showed about 90% of eGFP positive cells in the cell population ( Figure S6). The nucleotide composition around the insertion was analyzed by DNA sequencing, additionally confirming the correct integration ( Figure S7). Also we made a ddPCR analysis to measure the copy number of the integrated sequence    . The researches demonstrated activity of mammalian adapted CRISPR/Cas9 in avian cells and the ability to knock out a gene. The CRISPR/Cas9 targeting efficiency in our experiments, according to T7E1 assay analysis, was around 1.8% that is similarly with the results from the previous article [Bai et al., 2016]. The low effectivity in our case can be explained by several reasons: we were restricted by 50-100 bp sequence of genomic DNA for selection of gRNAs; the area of targeting is the beginning of 3'UTR, which usually has a lower GC content and has a lot of repeated elements; also the transfection efficiency of the DF1 cells with multiple plasmids was less than 50%. The effectiveness of targeting could be improved by increasing transfection effectiveness and applying surrogate reporter assay.
Traditionally homology directed recombination with long homology arms was used to insert a desired sequence at a desired place. For example, homology directed recombination at chicken JH segment was performed without CRISPR/Cas9 using ~8-9 kb of total homology arms. The inserted sequence had a size ~2000bp. , that potentially could reduce the opportunity to design long homology regions required for a traditional gene targeting approach, so the CRISPR/Cas9 system is very helpful in the case. The length of homology regions has to correlate with the size of the insertion and we used homology regions 1000 and 3000bp. In practice amplification and cloning of the homology arms of such length was not labour-intensive and the same strategy of targeting a tissue-specific gene can be easily applied to other loci of interest. In our experiments HDR effectiveness at 72 hours after transfection was around 0.5% of RFP-positive cells in the case of targeting with gRNA2. Using drug selection, we achieved up to 90% targeted integration.
Expression from endogenous promoters could be favorable for other applications, for instance, synthesis of pharmaceutical proteins in the egg white under the ovalbumin promoter, or expression of a gene that provides a defense against a pathogen in tissues that are located in contact with the infection. Usage of an endogenous promoter to express an inserted gene also could prevent its epigenetic silencing. This is very important for long-term stable expression of pharmaceutical proteins and other applications. For now, genome modification in chickens has been established using . We are planning to use our approach to insert a gene of interest, in place of eGFP, under the control of the tissue-specific promoter of PGC, and enrich cells after homology directed repair. Cultured PGCs can be transfected and injected into recipient-embryos, where they will produce the germline.
In conclusion, we demonstrated that the CRISPR/Cas9 system along with cassette for HDR can successfully target the 3'UTR of endogenous genes to integrate a gene under endogenous control. Available under a CC0 -"Public Domain Dedication"

Competing interests
No competing interests were disclosed.

Grant information
The work has completed under the financial support of the Russian Science Foundation within project No. 16-16-04094.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Table S1. Predicted off-targets for each gRNA.

Supplementary material
Click here to access the data.  Click here to access the data. Figure S3. In vitro Cas9 digestion, using gRNA1, 2 and 3. Cas9 with non-targeting scramble gRNA and Cas9 without any gRNA were used as negative controls.
Click here to access the data. Click here to access the data.

Figure S5. Results of cell viability assays for G418 using the CellTiter-Blue Reagent.
Click here to access the data.

Figure S6. Geneticin selection of eGFP-positive cells by flow cytometry analysis.
Click here to access the data. Click here to access the data.
The incomplete labeling of Fig. S1A,B,C prevents interpretation of the integration events -are these mostly monoallelic or biallelic events? Derivation of clonal colonies may be necessary to fully address this question.
Many of the figure legends require much more information to interpret: Fig. S1A, B, C -what samples are in each of the lanes? Fig S6 requires labels and a better legend. What cell populations are on the top and on the bottom -before and after selection? Fig. S7 requires a better description/legend. Which components of A) are relevant to the authors statements in the text? What are the sequences shown in B), and which primer sequencing are represented in the chromatographs? Was a single clonal colony raised and selected for sequencing, or is this a PCR product amplified from the bulk selected DNA?
Minor concerns: Fig S2 I: what does the reference at the bottom mean? Fig. 3 shows microscopy, but the text (Results, "Homologous recombination…") gives a measurement of GFP+ cells -0.5% -how was the quantification done? These results should be presented directly. Methods, In vitro cleavage by Cas9 -it would be useful to include the concentrations of gRNA, Cas9 and target DNA, as well as the reaction conditions. Methods, first paragraph -the "Bai Y" citation is incomplete. Methods, "Targeting vector design" -was the DTA sequence inserted or already present in the parental LSLCas9-Rosa26TV plasmid? Methods, "PCR analysis", third paragraph -the authors are likely referring to Fig. S1D for primers pair 3 as well. Results, "Targeting 3' UTR.." - Fig. S2 is likely referenced by mistake. Table 2: Primer and probe sequences for ddPCR probes 1 and 2 are listed, but not probe 3 as described in the text. Results, "Homologous recombination…" - Fig. S6 does not show PCR analysis results.

If applicable, is the statistical analysis and its interpretation appropriate? Not applicable
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed. Competing Interests: 1.

2.
3. The work is a confirmation that the CRISPR/Cas9 system and a template DNA design can be technically used in a chicken cell line for inserting a reporter gene into a genomic locus. The authors demonstrated that insertion of an EGFP gene cassette into the 3'UTR of the endogenous GAPDH gene with an efficiency of 0.5%, and 89% with selection with antibiotic. The authors provided a detailed description of the protocol, which is helpful for other researchers to follow. However, the work remains a proof-of-principle demonstration that the CRISPR+DNA donor can work in a chicken cell line, the efficiency is likely not high enough for applications. The authors didn't spend efforts to improve the efficiency, nor investigate what might be the causes of the low efficiency. The protocol used in the chicken cell line is largely taken from the one that is used in mammalian cells, including the human codon optimized Cas9, and mammalian promoters for expressing Cas9 and gRNA. To make a convincing case that their method is "robust", they should test inserting more genes (not only a reporter gene) into more genomic loci (not only GAPDH). Because of this, it remains a question if the described method is robust enough for inserting any gene of interest into any genomic locus of interest. The suggested next-step experiments include optimizing the method to increase the efficiency, testing more genomic loci, testing gene insertion with different lengths, and testing in the chicken primodial germ cells that will be useful for the applications. Other suggestions for improvement:

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
While the authors demonstrate inserting a promoterless gene into the genomic under the control of an endogenous promoter, they also claimed potential tissue-specific gene expression. However, they didn't prove that their method can result in a cell line that exhibits tissue-specific gene expression.
The efficiency of gene cutting in vitro is lower than expected, only 1.8%. What causes this low efficiency? In general, the authors should provide a detailed discussion of the generally low efficiency observed both in vitro and in vivo, which will be the primary targets for improvement.
The authors should mention one additional advantage of their use of the endogenous promoter to express inserted genes, which is less likely be epigenetically silenced. From their data, they observed a relatively long-term (1 month) expression from the endogenous GAPDH promoter. This can be an advantage for long-term stable expression of pharmaceutical proteins compared to using an exogenous promoter that can be epigenetically silenced over time.
There is not off-target assay to characterize the real off-target sites from the computational ones. Also, what is NHEJ rate in the correctly EGFP inserted cells? Are there alleles also show indel editing via NHEJ? It would be useful to understand the NHEJ vs. HDR landscape in a chicken cell line.
The authors should try additional experiments to improve the efficiency. For example, they can try to reduce the number of plasmids used for transfection, test methods such as RNP delivery or using Cas9 mRNA, use a chicken codon-optimized Cas9 and chicken promoters to express Cas9 and gRNA, and re-design the 5' and 3' homologous arms. Via these experiments, it may become a more useful resource to understand what might lead to better HDR efficiency in a chicken cell line.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Partly
No competing interests were disclosed.

Competing Interests:
Referee Expertise: genetic engineering, synthetic biology I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Author Response 29 May 2018 , Moscow Institute of Physics and Technology, Russian Federation Ekaterina Antonova 1. Yes, we tell that potentially the method can be used for integration in a tissue-specific locus. The next step will be to apply the approach for integration under a germ cell-specific promoter in PGC cells. 2. The low efficiency can be explained by several reasons: we were restricted by 50-100 bp sequence of genomic DNA for selection of gRNAs; the area of targeting is the beginning of 3'UTR, which usually has a lower GC content and has a lot of repeated elements; also the transfection efficiency of the DF1 cells with multiple plasmids was less than 50%. The 1,8% effectiveness was pointed for T7E assay but not for in vitro assay. The following simple steps could be done for improvement of the locus targeting in the cell line: (for increasing CRISPR/Cas9 targeting:) -Targeting the last exon and mutation of PAM in HDR cassette instead of targeting the beginning of 3'UTR in the case when there are no appropriate predicted gRNAs.
-Increasing of transfection effectiveness (usage of all-in one vector or minicircles, usage of chicken 3'UTR in the case when there are no appropriate predicted gRNAs. -Increasing of transfection effectiveness (usage of all-in one vector or minicircles, usage of chicken U6 promoter, usage of RNA/RNP delivery for the instruments, etc). (for HDR improvement:) -Usage of longer or smaller homology arms. The optimal length of arms could vary in depends on targeting locus. -Blocking NHEJ pathway in order to increase HDR events. 3. The advantage was mentioned in the Version 2, the Discussion section: "Usage of an endogenous promoter to express an inserted gene also could prevent its epigenetic silencing. This is very important for long-term stable expression of pharmaceutical proteins and other applications." 4. Yes, however from computational analysis we found that none of the off-target sites were present in a known coding sequence of chicken genome. For instance, gRNA2 targets intragenic locus. Thus experimental off-target estimation is not so actual from point of view of probable harm.
Regarding NHEJ and NHEJ vs.HDR: We did not estimate NHEJ rate in the correctly EGFP inserted cells, we used ddPCR only for HDR calculation. We can not analyse the NHEJ vs. HDR landscape like in the publication [1], due to different design of the experiment. In our study long homology arms and long integrated sequences were used that did not allow us to estimate HDR and NHEJ simultaneously. Different integrations might occur at two genomic alleles in a single cell. Some EGFP+ cells could have the correct knock-in in at least one allele. Another allele can be unchanged or can contain NHEJ. 5. There are a lot of parameters for changing in order to improve gene targeting efficiency. Some of them has been already optimized for chicken cells in literature. For instance, U6 promoter yields 4-fold higher gRNA expression than the human U6 in DF1 cell line [2]. We have tested gRNAs delivery in plasmids or amplicons in the cell line, Cas9 with different nuclear localisation signals (SV40 and nucleoplasmin) and several ratios of the instruments [data not shown]. The best conditions were taken to be presented in the article. The aim of the study was not to optimise Cas9 mediated targeting. We wanted to demonstrate that the approach with negative and positive selection works in chicken cells allowing to obtain modified cell line even when targeting effectiveness is low that is expected in the case of PGC cells. We suppose that the results can be improved. For example, variation of different sizes for the 5' and 3' homologous arms could increase the quantity of desired cells before enrichment etc. We agree with you that our method can not be called as a robust. It is a possible way for a gene insertion under control of a desired endogene promoter in chicken cells. In further researches the methods should be applied in PGC cells.
6. Three separate plasmids were used to optimise several conditions that can influence on gene targeting efficiency. We tested different ratios of Cas9 and gRNA. Also we tested amplicons and plasmids for gRNAs delivery. The best conditions were selected for HDR and were presented in the article. RFP was used for cotransfection. We agreed with you that usage one plasmid with all components could increase the number of transfected molecules of instruments per cell and thereby could increase targeting effectiveness. Alternatively mini-circles can be used for the same purpose. 7. We mentioned about one copy of the gene, because genomes of other animals have different number of GAPDH pseudogenes [Yuen-Jong Liu et al., 2009]. However one copy of GAPDH implies two alleles. PCR analysis does not allow to answer the question bi-allelic or mono-allelic integration happened, the amplicon of the exact size can be amplified in any case of the correct insertion. Based on ddPCR data we calculated the ratio of copies of transgene per one allele of GAPDH gene in total. In the figure S2.III there is a misprint in the title. It should be "Transgene copies per GAPDH allele" instead "Transgene copies per genome". The title was modified in the version 2. EGFP-positive cells and Geneticin resistant cells could have monoallelic or biallelic integration. The ratio of transgene per genome is 1,6±0,4 copies based on ddPCR analysis.
No competing interests were disclosed.

Competing Interests:
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage For pre-submission enquiries, contact research@f1000.com