{"paper_id":"213a041f-af83-4cfe-abcd-b1b4c0cb1580","body_text":"ONE-STEP tagging: a versatile method for rapid site -specific \nintegration by simultaneous reagent delivery.  \n \nValentina Migliori*1, Michaela B. Bruntraeger$1, Ivan S. Guylev$1, Thomas Burgold1, Florence Lichou1,2, \nAndrew L. Trinh1, Sam J. Washer1, Carla P Jones1, Gosia Trynka1,2 and Andrew R. Bassett*1. \n \n$These authors contributed equally \n*Corresponding authors emails: vm14@sanger.ac.uk; ab42@sanger.ac.uk \n \nAffiliations       \n1 Wellcome Sanger Institute, Wellcome Genome Campus, Hinxton, Cambridge, CB10 1SA, UK \n2Open Targets, Wellcome Genome Campus, Cambridge, UK  \n \n \n \nAbstract  \nSite-specific integration of DNA sequences into the genome is an important tool in fundamental \nresearch, synthetic biology and cell therapeutic applications. It can be used for protein tagging to \ninvestigate expression, localisation, and interactions as well as for expression of transgenes either \nunder endogenous regulatory elements or at consistent safe harbour loci. Here we develop and \noptimise a simple and effective method for site specific integration in a single step that combines \nCRISPR-Cas9 mediated homology directed repair using single stranded oligonucleotide templates \nwith the site-specific recombinase Bxb1 to allow large cargos to be integrated at any location in the \ngenome. Our technology requires off the shelf Cas9 and oligonuc leotide reagents combined with a \nset of cargo plasmids that are universal to any integration site. We demonstrate the method s \nadaptability by tagging at multiple sites and in multiple cell types including induced pluripotent stem \ncells and primary T cells. We show that our method can integrate large (up to 14 kb) cargos and that \nit is possible to simultaneously tag two genes or edit two sites with combination of integration and \nCas9-mediated knockouts or other HDR events.   \n \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nIntroduction  \nRecent advancements in genome -editing technologies have provided efficient tools for specific \ngenome modifications across various cell types and organisms. An important component of genome \nengineering is site -specific integration of DNA sequences to genomi cally tag particular proteins to \ninvestigate their function and allow expression of transgenes under endogenous regulatory elements \nor at specific safe harbour loci. These are important for studying the localisation, temporal \ndynamics  and protein interact ions of genes for understanding cellular function and for controlling \ntransgene expression in cell therapeutics such as CAR-T therapies (T. Li et al. 2023).  \nCRISPR-Cas9 enhanced homology-directed repair (HDR) has become a key technology for transgene \nintegration (Pacesa, Pelea, and Jinek 2024) . This involves creating a double -strand break (DSB), and \nsupplying an excess of a homologous template DNA which is used to elicit a highly precise repair (Ran \net al. 2013).  Short synthetic single-stranded DNA (ssDNA) oligonucleotides with ~100 nt of homology \nhave been highly effective in many cell types (Chen et al. 2011; Wang et al. 2014) but they are limited \nin cargo capacity to around 100 nt making insertion of large transgenes impossible. Longer plasmid \nDNA with around 500-1000 nt of homology can be used to integrate larger fragments (Friedel et al. \n2005), but there have been observations of complex, multimeric integration events at the on target \nsite (Norris et al. 2020). Others have used long ssDNA (Roth et al. 2018) or linear double-strand DNA \n(dsDNA) to circumvent these problems (F. Song and Stieger 2017) , but the former is difficult to \nproduce, especially with longer cargos, and the latter is prone to random integration at off-target sites \nin the genome (Zelensky et al. 2017; A. Song et al. 2017) . Protection of linear dsDNA with DNA \nstructures or chemical modifications such as biotin has been shown to reduce non-specific integration \nand achieve tagging in some cell types (Gutierrez-Triana et al. 2018; Shy et al. 2023) . However, with \nall of these methods, it is still necessary to produce a HDR template DNA of thousands of bases that \nis different for every targeted site. Also, the efficiency of integration drops rapidly with increasing \ninsertion size, making it difficult to insert inserts of more than 10 kb (K. Li et al. 2014).  \nTo overcome the size limitation, some groups have successfully combined HDR-mediated integration \nof a landing pad at a specific genomic locus followed by a second step of site-specific recombination \n(Feng et al. 1999; Low et al. 2023; Xu et al. 2013; Mulholland et al. 2015) . The serine integrase Bxb1 \nis particularly useful for this, as it efficiently and specifically recombines heterologous attB and attP \nsites without known pseudo -sites in the human genome  (Russell et al. 2006) . Its recombination is \ndirectional and irreversible  (Singh, Ghosh, and Hatfull 2013) . However, such integration typically \ninvolves a two-step process, with clonal selection after the HDR event, making it quite lengthy and \ndifficult to scale to multiple sites.  \nAn alternative method for integration employs the non -homologous end joining (NHEJ) repair \nmechanism that ligates two dsDNA ends together (Suzuki et al. 2016; Zeng et al. 2020) . By \nsimultaneously cutting the genome and a donor DNA within the cell, this can be exploited to insert \nthe donor DNA into any desired genomic site. This allows common donor plasmids to be used, \navoiding the need for cloning, and making scaling of this method possible. It also has less of a length \nlimitation than HDR-based methods, and tens of kilobases can be integrated using these methods. \nHowever, efficiency is variable between cell types, there is no control over orientation of the insertion, \nin some cases the whole plasmid will be integrated, and the NHEJ repair mechanism can sometimes \nintroduce small insertions and deletions around the genomic cut site.  \nPrime editing (PE) combines a modified single-guide RNA (pegRNA) containing the template for the \ndesired edit, with a reverse transcriptase (RT) fusion to Cas9 (Anzalone et al. 2019). This makes a nick \nat a genomic locus and extends the genomic DNA by reverse transcription of the pegRNA to introduce \nthe edit, and through manipulation of mismatch repair (MMR) pathways can be biassed towards \nincorporation of the newly edited strand (Ferreira da Silva et al. 2022). PE offers advantages over HDR \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nsuch as fewer mutagenic DSBs, but it also has limitations, including fewer targetable sites and limited \ninsertion length (<50 bp) (Fichter, Setayesh, and Malik 2023) . Recently, insertion of a site -specific \nrecombinase site using PE, and simultaneous delivery of the cognate recombinase resulted in efficient \nintegration of large cargos into specific genomic sites in a single-delivery reaction (PASTE technology) \n(Yarnall et al. 2023) . Recombination efficiency has recently been improved using evolved Bxb1 \nintegrase (eeBxb1 and evoBxb1) (Pandey et al. 2024) . Similarly, template-jumping prime editing (TJ \nPE) has allowed insertions up to 800 bp (Zheng et al. 2023) . However, despite progress with \ncomputational prediction tools, it is still difficult to design effective PE guides without testing several \npermutations (Wu et al. 2023). Also, the target sites are somewhat limited by directionality of the PE \nprocess and availability of Cas9 cut sites (Durrant et al. 2024). \nAll of these methodologies rely on the endogenous DNA repair pathways of HDR, NHEJ or MMR. The \nefficiency of these repair pathways varies significantly between cell types. Embryonic stem cells and \ninduced pluripotent stem cells (iPSCs) tend to favour HDR pathways (Guo et al. 2018), but many cancer \ncell lines and terminally differentiated cells preferentially repair through NHEJ (Srivastava and \nRaghavan 2015; Dharanipragada et al. 2023). Thus, the choice of the method will depend on the cell \ntype and respective repair pathways that are active.  \nWe present ONE STEP tagging, a technology allowing simple, efficient, and directional integration of \nlarge transgenes at any genomic site using a single -step delivery protocol that employs single -\nstranded oligodeoxynucleotides (ssODN)-templated HDR combined with Bxb1-mediated integration. \nWe show its utility in tagging at multiple genomic locations in multiple cell types including pluripotent \nstem cells, cancer cell lines and primary T cells. We further demonstrate that large cargos of up to 14 \nkb can be integrated. Tagging plus other Cas9-mediated knockout or HDR events can be performed \nsimultaneously at two different sites. We also optimised the use of heterotypic Bxb1 recombinase sites \nto avoid the integration of plasmid backbones and enable directional dual tagging. Our system allows \nthe use of completely off -the-shelf reagents, namely commercially available Cas9 protein, synthetic \nsgRNAs and ssODNs to define the genomic location and a common set of cargo vectors to define the \ninserted fragment which will make this methodology scalable to a large number of target sites.  \nResults \nONE STEP tagging combines CRISPR -Cas9 editing with Bxb1 site specific integration  \nWe have devised a versatile integration system that combines the precision of CRISPR -Cas9-based \nediting with the efficient, and less size-dependent integration of DNA cargos by Bxb1 serine integrase. \nBxb1 is functional in mammalian cells and efficiently ca talyses unidirectional (non -reversible) \nrecombination between dsDNA sequences containing an attP and their complementary attB \nattachment site. By utilising CRISPR -Cas9 to position the integrase attP sites at specific genomic \nlocations, we can direct Bxb1 (delivered in trans) to act at the chosen sites. By simultaneously providing \na circular double-stranded DNA template containing the attB attachment site, we aim to achieve direct \nintegration in a single step (Fig 1a).  \n \nDue to its specificity for dsDNA over ssDNA, Bxb1 cannot recombine attP attachment sites in a ssODN \ndonor template until it is inserted at the chosen genomic location  and made double stranded , and \nthus recombination with the dsDNA cargo donor will only occur after attP integration. Since Bxb1 \nleaves residual sequences in the genome (termed attL and attR) after recombination, we utilise these \ngenomic scars as protein linkers by strategically positioning the attB site on the cargo which we initially \ndelivered on a minicircle. We designed an experiment aimed at endogenously tagging the N-terminus \nof the constitutively expressed ACTR10 protein with a fluorescent protein (mNeonGreen) in hiPS Cs. \nWe combined all necessary reagents and delivered them simultaneously in a single nucleofection. This \nincluded the CRISPR-Cas9 ribonucleoprotein complex which carried an sgRNA targeting our site of \ninterest, along with a short (200 nt) single -stranded HDR template containing the attP site and \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \napproximately 70 bp of homology arms each side of the specific site. Additionally, we included a \nplasmid for the mammalian expression of Bxb1 and a circular double-stranded DNA cargo containing \nthe attB attachment site and the sequence encoding for mNeonGr een (Fig 1a). To avoid integration \nof plasmid backbone sequence, we generated this cargo as a minicircle by self-circularising a restriction \nfragment. \nWe assessed the outcome by flow cytometry analysis of mNeonGreen fluorescence (Fig S1a). By \ntitrating the amount of each component, we found that delivering 24 pmol (4 ug) Cas9, 45 pmol \nsgRNA, 50 pmol (3.1 ug) of ssODN HDR template, 127 fmol (500 ng) of Bx b1 integrase plasmid and \n0.8 pmol (380 ng) of circularised donor yielded optimal efficiency. Under these conditions, 11.3% of \ncells exhibited mNeonGreen fluorescence indicating in -frame tagging (Fig 1b). Correct and precise \nintegration was further confirmed by PCR-based genotyping (Fig S1b) and Sanger sequencing of the \nproducts. In conclusion, our results demonstrate that  we can achieve site -specific integration of \nmNeonGreen at the N -terminus of ACTR10 in a single step, hence the designation of ONE STEP \ntagging. By comparing integration rate of mNeonGreen in the presence or absence of the ssDNA \ndonor containing the attP site, we observed that random integration of the tag was minimal.  Tagging \nACTR10 with increasing concentrations of mNeonGreen dsDNA cargo donor significantly enhanced \ntagging efficiency, thus demonstrating that an excess of cargo donor alone is adequate to improve \nresults. \n \nOptimisation of ONE STEP tagging in hiPSCs  \nNHEJ and HDR are the two major branches of DNA damage response pathways that process DSBs. \nNHEJ is characterised by the modification and ligation of blunt DNA ends, and acts throughout the \ncell cycle. It functions independently of sequence homology, is kin etically faster than  HDR-related \nmechanisms and is the predominant DSB repair pathway in many cell types. By comparison, HDR is \nlimited to the S- and G2-phases of the cell cycle, where the presence of a sister chromatid allows for \nfaithful and potentially error-free repair.  \n \nEfforts to enhance repair by template -based pathways include regulation of key DNA repair factors, \nmodulation of the CRISPR-Cas9 components and alterations of the intracellular environment around \nDSBs (Charpentier et al. 2018; Rees et al. 2019; Aird et al. 2018) . Recent work by Wimberger et al. \nconfirms that inhibition of DNA-PK, a key player in the NHEJ pathway, is most effective at improving \nCRISPR-mediated insertions(*). To that end, we tested four commercial DNA -PK inhibitors (NU7441, \nAZD-7648, M3814 and IDT HDR Enhancer) with the aim of promoting repair via HDR and ultimately \nimproving tagging efficie ncy. Optimal concentrations of inhibitors were established using a hiPSC \nreporter line composed of a single copy of blue fluorescent protein (BFP) integrated at the hROSA26 \nlocus of the Kolf2.1s cell line which can be specifically targeted at the fluoropho re binding site by \nCRISPR-Cas9 generating a DSB. Subsequent repair by NHEJ will predominantly result in indel \nformation and loss of BFP fluorescence. However, if a template for HDR introducing a 2 -nucleotide \nmutation is utilised in the repair process, fluo rescence is altered from blue to green, indicating \nmeasurable levels of repair by HDR (BFP-GFP reporter assay). In this system, AZD-7648 proved to be \nthe most effective at increasing rates of repair by HDR with 1.6 fold increase relative to controls (Fig \nS2).  \n \nApplication of DNA -PK inhibitors in the ONE STEP tagging system produced similar results, with \ntagging of ACTR10 with mNeonGreen improving 5 -fold and 8 -fold in A1ATD and Kolf2.1s cells \nrespectively (Fig 1c). Tagging of three further sites (MAP4, LMNA and FBL) and subsequent analysis of \nattP integration by ICE (Conant D, et al. CRISPR J. 2022 Feb;5(1):123-130) confirmed improved rates \nof HDR with the use of AZD-7648 (Fig S3). In all subsequent experiments, we added AZD-7648 (0.5µM) \nafter nucleofection and changed the media 24 hours later. \nUtilising the BFP reporter line, we additionally assessed the optimal concentration of ssODN oligo by \ntitrating the amount of HDR template employed to introduce the 2 -nucleotide mutation in the BFP \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nreporter gene, resulting in subsequent conversion to GFP. The addition of 100 pmol of HDR template \nyielded HDR efficiency above 60%, whereas 25 and 12.5 pmol of the same template still resulted in \napproximately 40% HDR efficiency (refer to Fig S4). \nIn all subsequent experiments, we kept the oligo amount at 25 pmol.  \nTagging ACTR10 with increasing concentrations of mNeonGreen dsDNA cargo donor significantly \nenhanced tagging efficiency, even in the absence of AZD-7648, thus demonstrating that an excess of \ncargo donor alone is adequate to improve results (Fig1d).  In all subsequent experiments, we \nmaintained the donor concentration at 6 fold to maintain optimal efficiency without affecting cell \nviability. \n \nDirectional ONE STEP tagging with variant recombinase sites  \nThe central dinucleotide within the attP and attB sites of Bxb1 plays a crucial role in the integration \nprocess by facilitating the association of these attachment sites (Ghosh, P., Kim, A. I. & Hatfull, G. F., \n2003). Since one bottleneck of the technology  is the efficient and reproducible generation of the \nminicircle DNA cargo, we decided to investigate the use of these variant recombination sites to allow \nspecific integration of a defined part of a plasmid.   We decided to test mNeonGreen integration at \nthe ACTR10 site by using the GA attP, which has been reported to have greater efficiency than the WT \nGT attP sequence (Jusiak, B. et al., 2019), and explored the specificity of matched and unmatched \nattB/attP dinucleotide interactions. We found that both GA and GT attP variant sites efficiently \nintegrated cargo only when paired with the corresponding attB/attP pair, with minimal integration \nacross mismatched combinations, with 10 fold less integration of GA donor into GT site, indicating low \nlevels of crosstalk between these variants (Fig S5).  \n \nBased on this observation, we designed a directional tagging strategy by adding two variant attP sites \n(attP-GA and attP -GT) onto the 200 nt ssDNA used as an HDR template still preserving ~50 bp \nhomology arms on each side. This is combined with a plasmid containing the cargo flanked by the two \nvariant sites as a donor (Fig S6). We compared the original ONE STEP technology with the directional \ntagging strategy and achieved similar tagging efficiency (Fig S7) with the advantage that now we can \ngenerate a more scalable, pure and reproducible plasmid library of donor cargos.  \n \nVersatile tagging at multiple loci in different cell types and with large cargo sizes  \nSince our technology does not require the assembly of long ssDNA HDR templates but instead uses a \nshort ssODN (200 nt) with approximately 50 bp of homology arms surrounding the recombination \ncassettes, integration of the cargo can be easily scaled across different loci. We tested and confirmed \nintegration of mNeonGreen at three other genomic loca tions (LMNA, FBL and MAP4).  To assess \ncorrect gene tagging, we used fluorescence imaging to compare the subcellular localisation of \nmNeonGreen with the reported location of the tagged protein. For all four targeted loci, mNeonGreen \nlocalised as expected, indicating successful tagging (Fig1e).  \nWe also expanded ONE STEP tagging to additional cell types beyond hIPSC lines (A1ATD and \nKolf2.1s) and tested it in the K562 lymphoblast line and the HAP1 near -haploid human myeloid \nleukaemia cell line. We observed varying efficiencies in the integration o f a 0.8 kb fluorescent tag at \nthe MAP4 locus, with integration rates ranging from 1.4% in HAP1 cells to nearly 17% in K562 cells \n(Fig. 1f). Despite the lower integration efficiency in some cell lines, we mitigated this issue by sorting \nand expanding the cells, thereby enriching for a pure population. \nAnother advantage of the recombinase tagging methodology is that integration efficiency is less \ndependent on length than with HDR. We therefore tried integration of larger cargo plasmids, and we \nwere also able to achieve precise integration of cargos as large as ~14kb into the ROSA26-BFP locus \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nin Kolf2.1s (Fig S8/Fig 1g), with ~14.5% complete integration of the full cargo confirmed by \nFACS/genotyping. \nMultiplexed tagging and simultaneous CRISPR editing  \nMultiplexed gene integration is a valuable technique for labelling different proteins, enabling the \nvisualisation of their intracellular localization and interactions within the same cell. We employed \ndirectional tagging to simultaneously tag ACTR10 with mNeonGreen and MAP4 with mCherry in the \nsame cell (Fig. 2a). Using FACS, we observed successful simultaneous tagging of MAP4-mCherry and \nACTR10-mNeonGreen (Fig S9). Interestingly, there was a greater percentage of double tagging events \n(2,41% mChr-mNg+) than you would expect from the combination of the efficiency at each individual \nsite (1,81% mCh+, 8.92% mNg+), suggesting that successful tagging at one site might enrich for a \nsecond tagging event. The double-positive cells were sorted, and the tagged genes were confirmed \nto be visible in their respective cellular compartments, consistent with the known subcellular \nlocalizations of ACTR10 and MAP4 protein products by imaging (Fig S9). \nAnother valuable application of the ONE STEP tagging technology is the simultaneous tagging of a \nprimary gene along with a knockout (KO) or homology-directed repair (HDR) editing event at a second \ngene). We delivered reagents designed to tag MAP4 with mChe rry into the BFP reporter line. The \npresence of the mCherry signal, combined with the BFP -GFP reporter assay, enabled us to monitor \nKO and HDR rates by detecting the loss of BFP or the gain of GFP signals, respectively. We observed \nthat 70% of the cells successfully tagged with mCherry also exhibited either a BFP KO (Fig2c) or GFP \nHDR event (Fig2d). \n \nONE STEP tagging can be used to integrate transgenes in primary T cells  \nWe also expanded the ONE STEP tagging system to edit more clinically relevant cells, such as primary \nhuman T cells, at therapeutically relevant loci, such as the TCR. To achieve best efficiency of tagging, \nwe changed nucleofection conditions. Higher concen trations of Cas9, gRNA and ssODN HDR \ntemplate were used. The P3 buffer (Lonza) was replaced by an optimised buffer (B1mix) shown to \nimprove HDR efficiency notably when integrating large DNA fragments (An et al. 2023).  The evo Bxb1 \nrecombinase was used instead of the WT version (Pandey et al. 2024) . We successfully integrated a \ncargo of 4.4kb (expressing EGFP) at the TCR locus (Fig 2d), using both a sense and an antisense ssODN \nas HDR templates containing the attP site, targeting the TCR locus (Fig S10). We observed a tagging \nefficiency of ~6% using the antisense ssODN.  suggesting the potential to adapt this technology for \nthe introduction of chimeric antigen receptor (CAR) transgenes at the TCR locus in primary T cells. This \nadvancement could enhance current strategies for cancer immunotherapy through adoptive transfer. \nAdditionally, the ability to simultaneously tag the TCR locus while performing B2M and PD1 knockouts \nopens the possibility of generating off-the-shelf CAR-T cells. \n \nConclusion  \nOur study demonstrates the successful development and application of the ONE STEP tagging system, \na streamlined and efficient method for site -specific integration and tagging of genes. This system \nleverages the precision of CRISPR-Cas9-mediated editing and the unique capabilities of Bxb1 serine \nintegrase, enabling the integration of DNA cargos in a single step. The use of completely off-the-shelf \nreagents, from commercial suppliers and a common set of cargo plasmids which we made available \nthrough Addgene, highlights the accessibility and practicality of this approach for a wide range of \nresearch and clinical applications. \nONE STEP tagging has proven effective in different cell types from hiPSCs to cancer cells and primary \nhuman T cells. This system allows for large DNA insertions and the tagging of multiple genomic sites \nsimultaneously. We demonstrated the integration of a 0.8 kb fluorescent tag at the MAP4 locus with \nvarying efficiencies across different cell lines, achieving up to 17% in K562 cells. Furthermore, the \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nsystem facilitated the precise insertion of larger cargos, up to ~14 kb, at specific loci in hiPSCs, \nunderscoring its versatility and efficiency. \nThe ability to combine tagging with knockout (KO) and homology-directed repair (HDR) events further \nenhances the utility of the ONE STEP system. For instance, we achieved simultaneous tagging of a \nprimary gene and a KO or HDR event at a secondary gene, with high rates of co-occurrence. This dual \nfunctionality is particularly valuable for complex genetic modifications and studies of gene interactions. \nOur results also demonstrate the applicability of ONE STEP tagging in clinically relevant cells. We \nsuccessfully integrated a cargo of 4.4kb (expressing EGFP) at the TCR locus in primary human T cells, \npaving the way for potential cell therapeutics adaptations, for example introduction of chimeric antigen \nreceptor (CAR) transgenes at the TCR locus. Additionally, the use of Cas9 nuclease offers simultaneous \ntagging of the TCR locus and performing B2M and PD -1 knockouts that opens the possibility of \ngenerating off-the-shelf CAR-T cells, which are able to resist alloresponsive T cell attacks and with \nreduced T cell exhaustion (Eyquem et al. 2017), offering new avenues for cancer treatment. \nAn important comparison between recently published PE -based systems and our HDR -based \nrecombinase-mediated system is the limitation of insert size when integrating recombination sites. \nssODNs, which can now be routinely synthesised up to 200 bp (with a cargo capacity of ~100-120 bp), \ncan insert dual recombination sites, whereas PE is limited to a single recombination site. This poses a \ndrawback for recombination donor production in PE -based methods. In the dual cassette version of \nour system, recombination donors are readily produced and do not need modification prior to \ntransfection, maintaining insert specificity. This quality makes them truly off -the-shelf integration \nvectors, as only the ssODN template and the sgRNA need to be designed. \nIn conclusion, the ONE STEP tagging system represents a significant advancement in genetic \nengineering, providing a fast, efficient, and versatile method for precise site-specific gene integration. \nIts ability to facilitate large insertions, multiple site tagging, and combination with other genetic \nmodifications makes it a powerful tool for both research and clinical applications. The system's success \nin hiPSCs and primary T cells underscores its potential for advancing gene therapy and regenerative \nmedicine. \nReferences  \nAn, Jing, Chuan-Ping Zhang, Hou-Yuan Qiu, Hong-Xia Zhang, Qiu-Bing Chen, Yu-Ming Zhang, Xin-Lin \nLei, Cai -Xiang Zhang, Hao Yin, and Ying Zhang. 2023. “Enhancement of the Viability of T Cells \nElectroporated with DNA via Osmotic Dampening of the DNA -Sensing c GAS–STING Pathway.” \nNature Biomedical Engineering 8 (2): 149–64. \nAird, Eric J., Klaus N. Lovendahl, Amber St Martin, Reuben S. Harris, and Wendy R. Gordon. 2018. \n“Increasing Cas9-Mediated Homology-Directed Repair Efficiency through Covalent Tethering of DNA \nRepair Template.” Communications Biology 1 (May):54. \nAnzalone, Andrew V., Peyton B. Randolph, Jessie R. Davis, Alexander A. Sousa, Luke W. Koblan, \nJonathan M. Levy, Peter J. Chen, et al. 2019. “Search-and-Replace Genome Editing without Double-\nStrand Breaks or Donor DNA.” Nature 576 (7785): 149–57. \nCharpentier, M., A. H. Y. Khedher, S. Menoret, A. Brion, K. Lamribet, E. Dardillac, C. Boix, et al. 2018. \n“CtIP Fusion to Cas9 Enhances Transgene Integration by Homology -Dependent Repair.” Nature \nCommunications 9 (1): 1133. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nChen, Fuqiang, Shondra M. Pruett -Miller, Yuping Huang, Monika Gjoka, Katarzyna Duda, Jack \nTaunton, Trevor N. Collingwood, Morten Frodin, and Gregory D. Davis. 2011. “High -Frequency \nGenome Editing Using ssDNA Oligonucleotides with Zinc-Finger Nucleases.” Nature Methods 8 (9): \n753–55. \nDharanipragada, Prashanthi, Xiao Zhang, Sixue Liu, Shirley H. Lomeli, Aayoung Hong, Yan Wang, \nZhentao Yang, et al. 2023. “Blocking Genomic Instability Prevents Acquired Resistance to MAPK \nInhibitor Therapy in Melanoma.” Cancer Discovery 13 (4): 880–909. \nDurrant, Matthew G., Nicholas T. Perry, James J. Pai, Aditya R. Jangid, Januka S. Athukoralage, \nMasahiro Hiraizumi, John P. McSpedon, et al. 2024. “Bridge RNAs Direct Programmable \nRecombination of Target and Donor DNA.” Nature 630 (8018): 984–93. \nEyquem, Justin, Jorge Mansilla -Soto, Theodoros Giavridis, Sjoukje J. C. van der Stegen, Mohamad \nHamieh, Kristen M. Cunanan, Ashlesha Odak, Mithat Gönen, and Michel Sadelain. 2017. “Targeting a \nCAR to the TRAC Locus with CRISPR/Cas9 Enhances Tumour Rejection.” Nature 543 (7643): 113–17. \nFeng, Y. Q., J. Seibler, R. Alami, A. Eisen, K. A. Westerman, P. Leboulch, S. Fiering, and E. E. \nBouhassira. 1999. “Site-Specific Chromosomal Integration in Mammalian Cells: Highly Efficient CRE \nRecombinase-Mediated Cassette Exchange.” Journal of Molecular Biology 292 (4): 779–85. \nFerreira da Silva, J., G. P. Oliveira, E. A. Arasa-Verge, C. Kagiou, A. Moretton, G. Timelthaler, J. Jiricny, \nand J. I. Loizou. 2022. “Prime Editing Efficiency and Fidelity Are Enhanced in the Absence of Mismatch \nRepair.” Nature Communications 13 (1): 760. \nFichter, Katye M., Tahereh Setayesh, and Punam Malik. 2023. “Strategies for Precise Gene Edits in \nMammalian Cells.” Molecular Therapy. Nucleic Acids 32 (June): 536–52. \nFriedel, Roland H., Andrew Plump, Xiaowei Lu, Kerri Spilker, Christine Jolicoeur, Karen Wong, Tadmiri \nR. Venkatesh, et al. 2005. “Gene Targeting Using a Promoterless Gene Trap Vector (‘targeted \nTrapping’) Is an Efficient Method to Mutate a Large Fraction of Genes.” Proceedings of the National \nAcademy of Sciences 102 (37): 13188–93. \nGuo, Q., G. Mintier, M. Ma-Edmonds, D. Storton, X. Wang, X. Xiao, B. Kienzle, D. Zhao, and John N. \nFeder. 2018. “‘Cold Shock’ Increases the Frequency of Homology Directed Repair Gene Editing in \nInduced Pluripotent Stem Cells.” Scientific Reports 8 (1): 1–11. \nGutierrez-Triana, Jose Arturo, Tinatini Tavhelidse, Thomas Thumberger, Isabelle Thomas, Beate \nWittbrodt, Tanja Kellner, Kerim Anlas, Erika Tsingos, and Joachim Wittbrodt. 2018. “Efficient Single-\nCopy HDR by 5’ Modified Long dsDNA Donors.” eLife 7 (August). \nhttps://doi.org/10.7554/eLife.39468. \nLi, Kai, Gang Wang, Troels Andersen, Pingzhu Zhou, and William T. Pu. 2014. “Optimization of \nGenome Engineering Approaches with the CRISPR/Cas9 System.” PloS One 9 (8): e105779. \nLi, Tianxiang, Yanyan Yang, Hongzhao Qi, Weigang Cui, Lin Zhang, Xiuxiu Fu, Xiangqin He, Meixin Liu, \nPei-Feng Li, and Tao Yu. 2023. “CRISPR/Cas9 Therapeutics: Progress and Prospects.” Signal \nTransduction and Targeted Therapy 8 (1): 36. \nLow, Benjamin E., Vishnu Hosur, Simon Lesbirel, and Michael V. Wiles. 2023. “Author Correction: \nEfficient Targeted Transgenesis of Large Donor DNA into Multiple Mouse Genetic Backgrounds Using \nBacteriophage Bxb1 Integrase.” Scientific Reports 13 (1): 11756. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nMulholland, Christopher B., Martha Smets, Elisabeth Schmidtmann, Susanne Leidescher, Yolanda \nMarkaki, Mario Hofweber, Weihua Qin, et al. 2015. “A Modular Open Platform for Systematic \nFunctional Studies under Physiological Conditions.” Nucleic Acids Research 43 (17): e112. \nNorris, Alexis L., Stella S. Lee, Kevin J. Greenlees, Daniel A. Tadesse, Mayumi F. Miller, and Heather \nA. Lombardi. 2020. “Template Plasmid Integration in Germline Genome -Edited Cattle.” Nature \nBiotechnology 38 (2): 163–64. \nPacesa, Martin, Oana Pelea, and Martin Jinek. 2024. “Past, Present, and Future of CRISPR Genome \nEditing Technologies.” Cell 187 (5): 1076–1100. \nPandey, Smriti, Xin D. Gao, Nicholas A. Krasnow, Amber McElroy, Y. Allen Tao, Jordyn E. Duby, \nBenjamin J. Steinbeck, et al. 2024. “Efficient Site -Specific Integration of Large Genes in Mammalian \nCells via Continuously Evolved Recombinases and Prime Editing .” Nature Biomedical Engineering, \nJune. https://doi.org/10.1038/s41551-024-01227-1. \nRan, F. Ann, Patrick D. Hsu, Jason Wright, Vineeta Agarwala, David A. Scott, and Feng Zhang. 2013. \n“Genome Engineering Using the CRISPR-Cas9 System.” Nature Protocols 8 (11): 2281–2308. \nRees, Holly A., Wei-Hsi Yeh, and David R. Liu. 2019. “Development of hRad51-Cas9 Nickase Fusions \nThat Mediate HDR without Double-Stranded Breaks.” Nature Communications 10 (1): 2212. \nRoth, Theodore L., Cristina Puig -Saus, Ruby Yu, Eric Shifrut, Julia Carnevale, P. Jonathan Li, Joseph \nHiatt, et al. 2018. “Reprogramming Human T Cell Function and Specificity with Non -Viral Genome \nTargeting.” Nature 559 (7714): 405–9. \nRussell, John P., David W. Chang, Anna Tretiakova, and Malla Padidam. 2006. “Phage Bxb1 Integrase \nMediates Highly Efficient Site-Specific Recombination in Mammalian Cells.” BioTechniques 40 (4): 460, \n462, 464. \nShy, Brian R., Vivasvan S. Vykunta, Alvin Ha, Alexis Talbot, Theodore L. Roth, David N. Nguyen, \nWolfgang G. Pfeifer, et al. 2023. “High -Yield Genome Engineering in Primary Cells Using a Hybrid \nssDNA Repair Template and Small-Molecule Cocktails.” Nature Biotechnology 41 (4): 521–31. \nSingh, Shweta, Pallavi Ghosh, and Graham F. Hatfull. 2013. “Attachment Site Selection and Identity in \nBxb1 Serine Integrase-Mediated Site-Specific Recombination.” PLoS Genetics 9 (5): e1003490. \nSong, Anren, Yujin Zhang, Leng Han, Gennady G. Yegutkin, Hong Liu, Kaiqi Sun, Angelo D’Alessandro, \net al. 2017. “Erythrocytes Retain Hypoxic Adenosine Response for Faster Acclimatization upon Re -\nAscent.” Nature Communications 8 (February): 14108. \nSong, Fei, and Knut Stieger. 2017. “Optimizing the DNA Donor Template for Homology -Directed \nRepair of Double-Strand Breaks.” Molecular Therapy. Nucleic Acids 7 (June): 53–60. \nSrivastava, Mrinal, and Sathees C. Raghavan. 2015. “DNA Double -Strand Break Repair Inhibitors as \nCancer Therapeutics.” Chemistry & Biology 22 (1): 17–29. \nSun, Wenli, Hui Liu, Wenhao Yin, Jie Qiao, Xueke Zhao, and Yi Liu. 2022. “Strategies for Enhancing \nthe Homology-Directed Repair Efficiency of CRISPR-Cas Systems.” The CRISPR Journal 5 (1): 7–18. \nSuzuki, Keiichiro, Yuji Tsunekawa, Reyna Hernandez-Benitez, Jun Wu, Jie Zhu, Euiseok J. Kim, Fumiyuki \nHatanaka, et al. 2016. “In Vivo Genome Editing via CRISPR/Cas9 Mediated Homology -Independent \nTargeted Integration.” Nature 540 (7631): 144–49. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nWang, Xiaoling, Yingjia Wang, He Huang, Buyuan Chen, Xinji Chen, Jianda Hu, Tammy Chang, Ren-\nJang Lin, and Jiing -Kuan Yee. 2014. “Precise Gene Modification Mediated by TALEN and Single -\nStranded Oligodeoxynucleotides in Human Cells.” PloS One 9 (4): e93575. \nWimberger, Sandra, Nina Akrap, Mike Firth, Johan Brengdahl, Susanna Engberg, Marie K. Schwinn, \nMichael R. Slater, et al. 2023. “Simultaneous Inhibition of DNA -PK and Pol ϴ Improves Integration \nEfficiency and Precision of Genome Editing.” Nature Communications 14 (1): 4761. \nWu, Youjun, Mega Sidharta, Aaron Zhong, Benjamin Persily, Mu Li, and Ting Zhou. 2023. “Protocol for \nthe Design, Conduct, and Evaluation of Prime Editing in Human Pluripotent Stem Cells.” STAR \nProtocols 4 (4): 102583. \nXu, Zhengyao, Louise Thomas, Ben Davies, Ronald Chalmers, Maggie Smith, and William Brown. 2013. \n“Accuracy and Efficiency Define Bxb1 Integrase as the Best of Fifteen Candidate Serine Recombinases \nfor the Integration of DNA into the Human Genome.” BMC Biotechnology 13 (October): 87. \nYarnall, Matthew T. N., Eleonora I. Ioannidi, Cian Schmitt -Ulms, Rohan N. Krajeski, Justin Lim, Lukas \nVilliger, Wenyuan Zhou, et al. 2023. “Drag -and-Drop Genome Insertion of Large Sequences without \nDouble-Strand DNA Cleavage Using CRISPR-Directed Integrases.” Nature Biotechnology 41 (4): 500–\n512. \nZelensky, Alex N., Joost Schimmel, Hanneke Kool, Roland Kanaar, and Marcel Tijsterman. 2017. \n“Inactivation of Pol θ and C -NHEJ Eliminates off -Target Integration of Exogenous DNA.” Nature \nCommunications 8 (1): 66. \nZeng, Fanning, Valerie Beck, Sven Schuierer, Isabelle Garnier, Carole Manneville, Claudia Agarinis, \nLapo Morelli, et al. 2020. “A Simple and Efficient CRISPR Technique for Protein Tagging.” Cells  9 (12). \nhttps://doi.org/10.3390/cells9122618. \nZheng, Chunwei, Bin Liu, Xiaolong Dong, Nicholas Gaston, Erik J. Sontheimer, and Wen Xue. 2023. \n“Template-Jumping Prime Editing Enables Large Insertion and Exon Rewriting in Vivo.” Nature \nCommunications 14 (1): 3369.     \n       \nAcknowledgements  \n       \nWe would like to thank the flow cytometry facility in Cellular Operations for analysis support  and to \nBen Davies for supplying the Bxb1 and attachment site sequences. We are grateful to members of the \nBassett lab for helpful comments on the manuscript, advice, discussions and support.    \n    \n     \nConflicts of interest  \n       \nAB is a founder of and consultant for Ensocell therapeutics.  \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nFigure 1  \n \n \nFigure 1 - Principle and optimisation of ONE -STEP tagging in different genomic locations, cell \nlines and cargo types. a, Schematic of ONE-STEP tagging technology. The system involves insertion \nof landing site via Crispr-Cas9, followed by landing site recognition and integration of cargo by Bxb1 \nintegrase. All reagents are delivered simultaneously. b, ONE-STEP tagging efficiency at the 5’ end of \nthe ACTR10 locus using mNeonGreen as cargo donor. ‘Negative control’ omits Bxb1 and circular \ndonor; ‘Random integration control’ omits the ssDNA donor indicating unspecific donor integration of \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \n<1% and ‘Tagging efficiency’ includes delivery of all components, resulting in on -target tagging \nefficiency of 11%. c, AZD-7648 is a potent DNA -PK inhibitor that improves gene tagging in \niPSCs. mNeonGreen integration efficiency at the 5’end of ACTR10 increases to 47.31% (A1ATD) and \n19.36% (Kolf2.1s) from 5.98% and 2.30% respectively using 0.5uM of DNA -PK inhibitor AZD -7648. \nd,Testing of cargo donor concentration to achieve optimal tagging efficiency.  Using higher \nconcentrations of mNeonGreen dsDNA cargo donor significantly enhanced tagging efficiency at the \nACTR10 locus, even in the absence of AZD -7648, thus demonstrating that an excess of cargo donor \nalone is adequate to improve results. e, Endogenous protein tagging with mNeonGreen via ONE STEP \ntagging at four loci (ACTR10, FBL, LMNA and MAP4). Fluorescence images of representative cells are \nshown. Cells have nuclear Hoechst staining. Endogenous expression of mNeonGreen localise \naccording to known, literature based, location of the four tagged genes. Scale bar 25μm. f, ONE-step \nTagging across multiple cell types. Percentage of integration of a 0.8kb tag at the MAP4 locus across \ndifferent human cell lines (Kolf2.1s, A1ATD, K562, HAP1). g, FACS plot showing the integration of a \n14.4 kb cargo into the BFP locus of the Kolf21s BFP reporter line. Cells that undergo integration are \ncharacterised by the loss of BFP signal and the gain of RFP signal, indicating successful integration. \nThe tagging efficiency is approximately 14.5%. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nFigure 2  \n \n \n \n \n \nFigure 2 - ONE-step tagging can be used for double tagging, simultaneous tagging and \nendogenous gene editing and in primary T cells. a, ONE-step Directional Tagging allows for \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nsimultaneous tagging of multiple loci with different cargo vectors. Schematic of experimental design \non the left. FACS analysis showing efficiency of multiplexed tagging insertion of combinations of \nfluorophores (mNeonGreen and mCherry) at two different loci (ACTR10 and MAP4), on the right. b, \nSimultaneous tagging of a primary gene (MAP4) along with a knockout (KO) event at a secondary gene \n(BFP). Schematic of experimental design on the left. Reagents designed to tag MAP4 with mCherry \nwere delivered simultaneously with a BFP sgRNA into a BFP reporter line. The presence of the mCherry \nsignal, combined with the BFP-GFP reporter assay, enabled us to monitor KO rates by detecting the \nloss of BFP. Roughly 76% of the cells successfully tagged with mCherry also exhibited BFP KO. c, \nSimultaneous tagging of a primary gene (MAP4) along with a HDR event at a secondary gene (BFP to \nGFP). Schematic of experimental design on the left. Reagents designed to tag MAP4 with mCherry \nwere delivered simultaneously with a HDR template employed to introduce the 2-nucleotide mutation \nin the BFP repo rter gene, resulting in subsequent conversion to GFP into a BFP reporter line. The \npresence of the mCherry signal, combined with the BFP -GFP reporter assay, enabled us to monitor \nHDR rates by detecting the gain of GFP. Roughly 72% of the cells successfully  tagged with mCherry \nalso exhibited GFP HDR events. d, ONE-STEP tagging integration efficiency of a 4.4kb insert into the \nTCR locus (EGFP expression) in primary human T cells, using sense and antisense ssODNs.  \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nMaterials and Methods  \n \nssODN design  \nssODNs that contain a single recombination site use an attP variant sequence that is 58bp long.  \nssODNs that contain two recombination sites use core attP variant sequences that are 48bp long \n(derived from the 58bp variant). When convenient, the length of one of the two 48bp long sites in the \ndual recombination cassette is extended by including an additional 5bp from the 58bp variant as \ndescribed below. \nIn this study, the recombination sites used in each part of the doublet pair of attP -attB sites contain \ndifferent central dinucleotide mutants (heterotypic sites) in order to ensure the efficient integration of \nonly the insert portion and not the backbone of the donor vector. This theoretically improves \nrecombination efficiency by 50%.  Table S1 contains a breakdown of the recombination sites used \nthroughout this work - attP, attB, resulting attR and attL sites and the lengths of DNA spacers used to \nkeep insertions in-frame in the case of gene tagging. \nDue to the order of insertion of recombination sites in the genome - attP(s) being inserted in the \ngenome first, and subsequently being recombined with a donor molecule containing attB(s) - the \nresulting recombinant sites and cassette are of the following structure: attR-insert-attL (this is the case \nfor both the single and double pair of attP-attB site systems). In the case of 5’ end tagging (where an \nN-terminal protein fusion is desired), either the single or dual attP cassettes are designed to replace \nthe start codon of the targeted CDS. In the case of 3’ end tagging (where a C-terminal protein fusion \nis desired), either the single or dual attP cassettes are designed to replace the stop codon of the \ntargeted CDS.The single attP -attB pair results in attR and attL sites that are 52bp long po st-\nrecombination. The double attP -attB pair results in attR and attL sites which are 47bp long post -\nrecombination. This difference necessitates a slightly different approach in the design of either ssODN \noligos, dsDNA donor cargos or both when keeping inserts in-frame with a native gene is of concern. \nIn this study, we address this difference by implementing the necessary changes in the ssODN design \nand keep the dsDNA donor plasmids similar.  For information on how single pair N -terminal tagging \nand C-terminal constructs are kept in-frame see the “dsDNA donor design” section. To keep double \npair N-terminal tagging constructs in -frame, a DNA spacer is included after the second attP of the \nssODN in order to influence the length of the resulting attL. The preferred sequence used in this study \nis the final 5bp of the 58bp-long attP - effectively the ssODN contains a single 48bp attP followed by \nanother 53bp attP. However, an alternative spacer can be used as long as it fulfils th e requirements \noutlined in the “dsDNA donor design” section. To keep double pair C-terminal tagging constructs in-\nframe, a DNA spacer is included before the first attP of the ssODN in order to influence the length of \nthe resulting attR. The sequence used i n this study is the first 5bp of the 58bp attP variant. This \neffectively results in a ssODN that contains a single attP of 53bp, followed by another 48bp attP. \nHowever, an alternative spacer can be used as long as it fulfils the requirements outlined in the “dsDNA \ndonor design” section.  \nsgRNA sequences for CRISPR cuts are designed to cut as close to the insertion site as possible, whilst \nminimising off-target effects, and ideally with the sgRNA sequence spanning the insertion site to block \nrecutting after correct integration. If this is not possible, synonymous or non-coding mutations should \nbe introduced in the PAM site within the homology arms of the ssODN. sgRNAs are synthesised as \nchemically modified RNAs (Synthego) to minimise toxicity and maximise editing efficiency.  \nAll sgRNA sequences are listed in table S2, and ssODN sequences used for single attP, dual attP and \nBFP targeting are listed in tables S3, S4 and S5.   \n \ndsDNA donor design  \nRecombination donor constructs used in this study are circular dsDNA molecules and contain attB sites \nwhich are 46bp long (plasmid maps used in this study in Table S6). In the case of the single attP-attB \npair system, a circular plasmid is cut and self-ligated to create a circular donor (containing a single attB \nsite) - the entire length of the circularised cargo is integrated into the genome (the circularised cargo \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \ndoes not contain any bacterial vector backbone sequence). The parent attB cargo vector also contains \nsingle attB site and inward facing Type IIS (BbsI) restriction sites that flank the attB-insert sequence. \nThe double attP-attB pair system contains two attB sites flanking the insert. It can be used without in \nvitro circularisation as upon completion of the recombination the insert cassette is inserted and the \nbacterial backbone sequence is excised. \nN-terminal tagging constructs are designed to be in -frame with the gene -of-interest post-\nrecombination. The length of the attR and attL sites that result from recombination is 52bp in the case \nof the single attP-attB pair system and 47bp in the case of the dual attP-attB system. \nTo keep N-terminal tagging constructs in-frame a DNA spacer is included in the 3’ end of the dsDNA \ndonor construct. Importantly, the spacer shouldn’t contain stop or start codons and its length should \ncomplement the attL to a length divisible by 3. A quick way to check that this is possible is to perform \na modulo operation with a divisor of 3 on the sequence length - it should return a remainder of 2. This \nis because 52 (the length of attR and attL resulting from recombination of the single attP-attB system) \nhas a remainder of 1 when divided by 3, their sum is 54, which is divisible by 3 and hence would be \nin-frame in the absence of stop codons.  \nTo keep C-terminal tagging constructs in-frame a DNA spacer is included in the 3’ end of the dsDNA \ndonor construct. As in the case of N -terminal tagging, the spacer’s length is designed to return a \nremainder of 2 when divided by 3 and not to avoid stop codons. \n \ndsDNA donor circularisation  \nThe dsDNA donors used in the single attP-attB pair system were generated from producer plasmids. \nThe producer plasmids were purified from bacterial cultures. The plasmids were then digested with \nBbsI and the fragment corresponding to the insert was gel purified. The purified insert fragment was \nself-ligated using T4 DNA Ligase. Low molar concentrations of DNA substrate and high concentrations \nof ligase improve the yield of the self-ligated circularised monomer. High DNA concentrations increase \nthe accumulation of intermolecular ligation and the formation of linear and circular dimers, trimers and \nhigher multimers. The ligation reaction was then purified using silica columns. In some experiments, \nthe ligation reactions were treated with T5 exonuclease to remo ve carryover linear monomers and \nligated linear multimers as well as open circles (nicked DNA) and ssDNA. \n \nEndotoxin removal from DNA preparations  \nDNA extractions were purified using the TXS method(Ma et al. 2012) prior to nucleofections. Briefly, a \n0.25 volume of TXS solution was added to 1 volume DNA solution and mixed thoroughly by inverting. \nThe solution was then incubated at room temperature for 5-10 minutes or for as long as overnight at \n4°C. \nA 0.25 volume of 5M NaCl was then added and mixed thoroughly by inverting. The sample was then \ncentrifuged for at least 10 minutes at maximum speed (>15,000 x g) at 4 °C. The clear upper layer, \nwhich contains the purified plasmid, was aspirated into a clean tube (care was taken not to take up the \nred tinted solution at the bottom of the tube). The clear supernatant was then precipitated using 100% \nisopropanol, washed with 70% v/v ethanol and dried. The DNA pellet was then resuspended with \nnuclease-free water or TE buffer (pH 8). \n  \nMammalian cell culture  \nKolf2.1s is an edited induced pluripotent stem cell   (iPSC) line, corrected for a 19bp deletion in one \ncopy of ARID2, which has undergone extensive characterisation and is commonly used for editing. \nKolf2.1s are derived by the Human Induced Pluripotent Ste m Cell Initiative (HipSci) consortium (ref). \nKolf2.1s BFP/GFP reporter line contains a BFP reporter inserted in the ROSA26 locus under a EF1a \npromoter (Bassett’s lab). \niPSC lines were cultured under feeder free condition in Stemflex medium (combo kit, Gibco TM \nA3349401) on Vitronectin substrate. Vitronectin is used at 1:100 dilution of a 1mg/ml stock solution in \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \nPBS, using 1ml per 6w well for coating. Incubate at RT for 1 hour before aspirating and replacing with \nculture media immediately . After initial thaw, cells were clump -passaged 1:10 every 4 -5 days and \ncultured in a humidified incubator at 37°C and 5% CO2. \nK562 and HAP1 were cultured respectively in RPMI and DMEM/F12 medium both supplemented with \n10% (vol/vol) fetal bovine serum (FBS).  \n  \nNucleofection  \nNucleofection of Cas9 RNP, ssODN containing the attP site, Bxb1 plasmid and cargo donor containing \nthe attB site in iPSCs (200000 cells) were carried out in P3 Primary Cell buffer (program CA137) in 16-\nwell cuvettes (Lonza) using an Amaxa 4D-Nucleofector (Lonza). Synthetic sgRNA (Synthego) and end-\nblocked ssODN (ultramer; IDT) were diluted in IDT duplex buffer to a concentration of 200mM and \n100mM, respectively.  eSpCas9 protein was in house produced and diluted in PBS to a final \nconcentration of 4mg/ml, but equivalent results can be obtained with commercial Cas9 proteins (Cas9 \nHiFi, IDT). \n  \nSmall molecule titration  \nAll small molecule inhibitors tested are commercially available. AZD -7648 (HY-111783, MedChem \nExpress), M3814 (HY-101570, MedChem Express) and NU7441 (S2638, Selleckchem) were dissolved \nin DMSO (Thermofischer Scientific) at a concentration of 5mM. IDT HDR Enhancer v2 (10007910, IDT) \nwas purchased as a 0.69 mM concentrated solution in DMSO. IDT HDR Enhancer v1 is no longer \ncommercially available.  \nOptimal concentrations of small molecule inhibitors were determined using the BFP -GFP reporter \nassay. Nucleofection of sgRNA (Synthego), Cas9 protein (produced in -house) and ssODN (Ultramer \nDNA oligo, IDT) into iPSC BFP reporter line were carried out in 100uL cuvettes (Lonza) using an Amaxa \n4D-Nucleofector (Lonza), P3 Primary Cell buffer and program CA137. Final amounts per nucleofection: \n1x106 cells in 100 μL P3 solution, 20 μg Cas9 protein, 20 μg sgRNA and 500pmol ssODN. Post -\nnucleofection, cells were maintained in small molecule inhibitor supplemented culture media at various \nconcentrations (0, 0.5, 1, 2, 4, 10, 20, 30 and 50uM) for 24 hours.  \n3 days post-nucleofection, cells were analysed for presence or absence of GFP and BFP respectively \nby FACS (CytoFLEX, Beckman Coulter). \n \nGenomic DNA extraction and characterisation by PCR  \nDNA was extracted from nucleofected cells using the DNAeasy Bood and Tissue kit from Qiagen (Cat. \nNo. / ID: 69504) according to manufacturer’s instructions. After purification, genomic DNA was eluted \nin 50μl of water. To confirm correct integration, target  regions were PCR amplified (primers in table \nS7) and analysed by gel electrophoresis. 10ng of gDNA was used to set up 50µl PCR reactions using \nKAPA HiFi HotStart ReadyMix (2x) (KAPA Biosystems, KK2601). The thermal cycling profile of the PCR \nwas: 95°C 3 min; 35x (98°C 20 s, 65°C 15s, 72°C 15s); 72°C 1 min. \n \nGenome-editing characterisation by ICE analysis  \nQuantification of ssODN integration \nA few days post -electroporation, edited cells were pelleted by centrifugation and the DNA isolated \nusing PureLink™ Genomic DNA Mini Kit according to manufacturer’s instructions. \nAmplicons were generated using indicated PCR primers (Table S7), designed to amplify an \napproximately 1000 bp region of genomic DNA surrounding the target site. 10ng of gDNA was used \nto set up 50µl PCR reactions using KAPA HiFi HotStart ReadyMix (2x) (KAPA Biosystems, KK2601). The \nthermal cycling profile of the PCR was: 95°C 3 min; 35x (98°C 20 s, 65°C 15s, 72°C 15s); 72°C 1 min. \nResulting PCR product was purified using Monarch® PCR & DNA Cleanup Kit (5 μg) according to \nmanufacturer’s instructions and shipped to Genewiz (Leipzig, Germany) for Sanger sequencing. Trace \nsequencing files were uploaded to ICE v2 (https://ice.synthego.com/, Synthego) to quantify efficiency \nof ssODN integration. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint \n\n \n  \nImaging \nCells for imaging were plated on PhenoPlate 96-well black walled microplates (Revvity, cat. #6005182). \nOnce confluent cells were washed with 100µl PBS per well and fixed at room temperature for 10 \nminutes with 4% PFA supplemented with Hoechst at 1 µg/ml for nuclei staining (40µl/well). Post-fixing, \ncells were washed 3 times then stored in 100µl PBS. Cells were imaged on an Opera Phenix (Perkin \nElmer) using a 40x/1.1 NA water lens, or as specified in the figure legends. Hoechst and mNeonGreen \nwere excited at 375nm and 488nm, respectively. Hoechst fluorescence was collected at 435 -480nm \nwith an exposure time of 60ms, and mNeonGreen fluorescence was collected at 500 -525nm with an \nexposure time of 100ms. \n \nPrimary T cell isolation and stimulation  \nHuman biological samples were sourced ethically, and their research use was in accord with the terms \nof informed consent under an institutional review board/ethics committee -approved protocol \n(15/NW/0282). Peripheral blood mononuclear cells (PBMCs) were is olated from fresh leukapheresis \nproducts from human healthy donors (Leukopaks, BioIVT) using a Ficoll-Paque PLUS (GE Healthcare, \ncat. #GE17-1440-03) density gradient centrifugation. Cells were cryopreserved in freezing media (RPMI \n1640 (Gibco, cat. #524000 25),  10% DMSO, 50% Fetal Bovine Serum (FBS, SIGMA -ALDRICH, cat. \n#F9665)) and stored in liquid nitrogen. PBMCs were thawed a day before T cell isolation, resuspended \nin complete RPMI media at 20e6 cells/mL (RPMI 1640, 10% FBS, 100 U/mL Penicillin -Streptomycin \n(Gibco, cat. #15140122), 2 mM L -Glutamine (Merck, cat. #G7513)) and incubated at 37°C 5% CO2 \novernight. PBMCs were collected and washed twice with DPBS without calcium or magnesium (Gibco, \ncat. #14190144). Total CD4+ T cells were isolated by immunomagn etic negative selection using the \nEasySep™ Human CD4+ T Cell Isolation Kit (STEMCELL Technologies, cat. #17952) according to the \nmanufacturer’s instructions. After isolation, cells were cultured in media consisting of StemPro ™-34 \nSFM (Gibco, cat. #10639011), 10% FBS, 100 U/mL Penicillin -Streptomycin, 2 mM L -Glutamine, \nrecombinant Human IL -2 at 40U/mL 10 ng/mL (PeproTech, cat. #200 -02BiolegendR&D systems) \n(named as complete StemPro) at 1e6 cells/mL. Cells were then stimulated wi th 12 µL/mL \nImmunoCult™(TM) Human CD3/CD28 T Cell Activator (STEMCELL Technologies, cat. #10971). \n \nNucleofection of primary T cells  \nLyophilized sgRNAs (Synthego) and ssODNs (IDT) were resuspended in water to a stock concentration \nof 100µM and stored at -20°C until use. RNPs were produced by mixing sgRNAs (180 pmol) and Cas9 \n(Alt-R™ S.p. Cas9 Nuclease V3, IDT, 61 pmol) at a 3:1 sgRNA:Cas9 molar ratio. The following ONE \nSTEP components were added to the RNP complexes: ssODN GA GT donor template (100 pmol), \nBxb1 plasmid (500 ng), and GA mNeonGreen GT plasmid (1.5µg). Primary T cells were spun down for \n5min at 400g and washed twice with DPBS without calcium or magnesium, before being resuspended \nin B1mix buffer (buffer B1 (5mM KCl, 15mM MgCl2, 120mM Na2HPO4/NaH2PO4 pH 7.2, 10mM sodium \nsuccinate and 25mM mannitol) were combined with mix solution (0.5mM sodium pyruvate, 0.8mM \nCa(NO3)2, 0.26mM i-inositol (myo-inositol), 4mM GlutaMAX (L-alanyl-L-glutamine dipeptide in 0.85% \nNaCl), 20mM d-glucose) in a v/v ratio of 72.3 to 27.7 ) at 7e5 cells per 20µL and added to the ONE \nSTEP pre-mix. The cells in the buffer were then transferred to a 16-well Nucleocuvette™ Strip (Lonza, \n4D-Nucleofector™ X Kit) for nucleofection using the pulse code EH -115. Immediately after \nnucleofection, 80µL of pre-warmed complete StemPro media was added to each well and incubated \nat 37°C with 5% CO2 for 15 min. The cells were then transferred to a 96 -well round-bottom plate \ncontaining 145µL of media with 0.5µM AZD-7648 and incubated at 37°C with 5% CO2 for 24 hours. \nAZD-7648 was then washed out by removing supernatants without disturbing the pellets and \nresuspending the cells in 250µL of pre-warmed media. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 29, 2024. ; https://doi.org/10.1101/2024.08.29.610246doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}