Genome editing of human pluripotent stem cells (hPSCs) can be done quickly and efficiently. Presented here is a robust experimental procedure to genetically engineer hPSCs as exemplified by editing the AAVS1 safe harbor locus to express EGFP and introduce antibiotic resistance.
Genome-editing of human pluripotent stem cells (hPSCs) provides a genetically controlled and clinically relevant platform from which to understand human development and investigate the pathophysiology of disease. By employing site-specific nucleases (SSNs) for genome editing, the rapid derivation of new hPSC lines harboring specific genetic alterations in an otherwise isogenic setting becomes possible. Zinc finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs) and clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 are the most commonly used SSNs. All of these nucleases function by introducing a double stranded DNA break at a specified site, thereby promoting precise gene editing at a genomic locus. SSN-meditated genome editing exploits two of the cell's endogenous DNA repair mechanisms, non-homologous end joining (NHEJ) and homology directed repair (HDR), to either introduce insertion/deletion mutations or alter the genome using a homologous repair template at the site of the double stranded break. Electroporation of hPSCs is an efficient means of transfecting SSNs and repair templates that incorporate transgenes such as fluorescent reporters and antibiotic resistance cassettes. After electroporation, it is possible to isolate only those hPSCs that incorporated the repair construct by selecting for antibiotic resistance. Mechanically separating hPSC colonies and confirming proper integration at the target site through genotyping allows for the isolation of correctly targeted and genetically homogeneous cell lines. The validity of this protocol is demonstrated here by using all three SSN platforms to incorporate EGFP and a puromycin resistance construct into the AAVS1 safe harbor locus in human pluripotent stem cells.
Genome editing technologies are rapidly evolving into standard tools for molecular and cell biology 1. Genetic engineering of human pluripotent stem cells (hPSCs) is of particular interest as hPSCs represent a self-renewing source of genetically intact primary human cells. hPSCs can be differentiated into various cell types for disease modeling or as a source for transplantation therapies 2,3. Demonstrated here is a protocol that utilizes three different types of site-specific nucleases (SSNs) in conjunction with endogenous DNA repair mechanisms for targeted integration of a reporter construct at the AAVS1 locus. After transfection of SSNs into hPSCs, we demonstrate how to isolate isogenic cell populations harboring the reporter.
The ability to manipulate genomes, specifically pluripotent stem cell genomes, using SSNs is not a new phenomenon as the utility of zinc finger nucleases (ZFNs) and transcription activator-like effector nucleases (TALENs) for gene editing was demonstrated several years ago 4-10. However, with the advent of S. pyogenes CRISPR/Cas9 technology 11-13, gene editing has become widely accessible 14. All SSNs introduce a double stranded DNA break (DSB) at the specified target site 1,4,5,11 that is repaired by endogenous cellular mechanisms using either non-homologous end-joining (NHEJ) or homology directed repair (HDR) 15. NHEJ is error-prone and can introduce frame shift mutations resulting in loss of gene function, while HDR allows for novel elements to be introduced through the co-transfection of a repair template with the SSN. While the underlying principles of DNA repair that facilitate gene editing are thought to be largely the same for each SSN, some differences between the platforms can be noted. De novo design of ZFNs allows flexibility and nuclease optimization 16, however the use of publicly available assembly libraries and screening tools to design individual ZFNs can be time consuming. Once the desired locus for ZFN-mediated targeting is determined, ZFN pairs can be designed with the online tool ZiFit 17. After design, ZFNs can be modularly assembled through several rounds of plasmid cloning 18. Alternatively, there are many commercially available, pre-validated ZFNs 19. TALE nucleases can also be designed using online tools and publically available components 17,20. For example, TALENs can be rapidly assembled from blocks of five TALE repeats, through FLASH assembly 21 or using PCR based hierarchical Golden Gate assembly 22. Ease of SSN design and speed of construction using CRISPR/Cas9 have made genome editing a widely accessible tool. The short guide RNA-mediated targeting of CRISPR/Cas9 also allows for multiplexing of guide RNAs to target several loci with a single construct 14. The design of Cas9 for gene editing requires only the identification of a protospacer adjacent motif (PAM; an NGG trinucleotide for S. pyrogenes Cas9) proximal to the target locus. By inserting an oligonucleotide corresponding to the 20 base pairs 5' of the PAM into the px330 plasmid 14, the construct can be assembled in one cloning step. In addition to S. pyogenes Cas9, Cas9 from N. meningitidis (NmCas9) that recognizes a 5′-NNNNGATT-3′ (PAM) has been shown to allow for efficient gene-editing in hPSCs 23.
In addition to the differences in ease of SSN design, each platform has specific properties. For example, ZFNs and TALENs utilize the FokI nuclease domain, which generates a four nucleotide 5' overhang 24 while Cas9 is thought to generate mostly blunt ended DSBs. ZFNs, TALENs, and Cas9 differ in their protein stabilities, on-off rate on target DNA, and mode of DNA scanning, all of which could theoretically result in small differences in the editing outcome 1. While further studies will be required to fully understand the consequences of these differences, we describe here a protocol that is highly robust across all three platforms and can be used to readily generate genetically modified hPSCs.
Regardless of SSN choice, electroporation is a robust procedure to transfect SSNs and homology repair templates into hPSCs 25. The number of surviving colonies after selection for antibiotic resistance depends on locus-specific parameters and the editing strategy (e.g., size of transgenic insert and mode of selection). The protocol described here usually results in about 150-400 single-cell derived colonies.
Gene-editing at the AAVS1 locus using this protocol has previously been used to demonstrate the effectiveness of SSNs 4,5. The AAV-CAGGS-EGFP repair template uses a gene trap strategy to confer puromycin resistance in a locus specific manner. Briefly, the repair template contains a splice acceptor site upstream of the promoterless puromycin resistance cassette. Upon correct integration into the first intron of the PPP1R12C gene at the AAVS1 locus, the resistance cassette is expressed from the edited gene's promoter. The robustness of this specific AAVS1 assay allows us to compare the efficiency of each SSN platform.
Gene editing using SSNs is powerful given the ability to disrupt and/or alter theoretically any gene. Applying this strategy to hPSCs provides versatility as hPSCs can be subsequently differentiated into a multitude of human cell types such as neurons 26, hepatocytes 27, and cardiomyocytes 28. Additionally, the use of patient-derived induced pluripotent stem cells allows the repair or introduction of known disease-causing mutations in a patient-specific genetic background 29, providing a platform from which to investigate disease mechanisms and test therapeutics using a patient's own cells 30. In summary, gene editing in hPSCs is an efficient and versatile approach for investigating the basic biology of human development and disease 31.
The procedures described in this manuscript were reviewed and approved by the UC Berkeley Stem Cell Research Oversight Committee.
1. Prepare Stem Cells for Editing
2. Editing Pluripotent Stem Cells
3. Selection of Positive Colonies
4. Picking Selected Colonies (Day 12-14)
Here we demonstrate a protocol compatible with three different SSN platforms to create genetically engineered hPSC lines. We targeted WIBR#3 human embryonic stem cells at the AAVS1 locus using previously published ZFNs 4, TALENs 5 and CRISPR/Cas9s 37 using a repair template that introduces an EGFP reporter and a puromycin resistance cassette 4.
We cultured our hPSCs on MEFs for a cell culture workflow permitting the maintenance and expansion of undifferentiated hPSCs (Figure 2), that is also cost effective and scalable. If cells are grown longer than necessary there is a risk of increased differentiation, reducing the number of pluripotent cells that are transfected and consequently the number of correctly targeted hPSC colonies obtained.
We electroporated 5.0 x 106 cells per SSN platform and plated the cells from each targeting onto a single 6 well plate of DR4 MEFs. After selection, each platform resulted in EGFP-positive colonies (Figure 3) and a combination of untargeted, homozygously targeted and heterozygously targeted clones (Figure 4A,B; Table 3). Under the conditions presented here, we find that the AAVS1 TALENs resulted in the most EGFP-positive clones. The repair template used for this experiment consists of a splice acceptor site upstream of the EGFP reporter and puromycin resistance cassette. Using this "gene-trap" strategy (Figure 1), the construct should insert into the first intron of the AAVS1 locus, using the endogenous promoter to drive expression of the puromycin-resistance cassette. The lack of a promoter driving the expression of the puromycin resistance gene within the repair template should prevent expression in the event of random integration.
Therefore, we expect that all puromycin-resistant clones would be targeted at the AAVS1 site. It should be noted that a subset of clones carry aberrant integrations in the AAVS1 locus that are detectable by Southern blot using an internal probe, but not by most PCR strategies (Figure 1; Figure 4C) 4. These integration events are most likely the result of heterologous targeting events leading to multiple integrations of the donor plasmid 4. We picked 24 colonies from each SSN experiment and found that all platforms had very high targeting efficiencies and showed only minimal differences. As interrogated by PCR, CRISPR/Cas9 yielded the most correctly targeted clones while the TALEN platform had the most homozygously targeted clones (Table 3).
Figure 1. Schematic of gene edited AAVS1 locus using the AAV-CAGGS-EGFP repair template. Modified from Hockemeyer et al., 2009. Please click here to view a larger version of this figure.
Figure 2. Colonies of WIBR#3 cells. Representative bright field images of colonies of WIBR#3 human embryonic stem cells prior to targeting. Note the lack of differentiation and the clear separation from the feeder layer in an ideal colony (left) as opposed to a non-ideal one (right). Please click here to view a larger version of this figure.
Figure 3. EGFP-positive WIBR#3 cells. Representative images of WIBR#3 cells targeted with an EGFP-expressing repair template at the AAVS1 locus. Images of representative colonies edited with ZFNs, TALENs and CRISPR/Cas9 are shown. Please click here to view a larger version of this figure.
Figure 4. Genotyping strategies to confirm correct targeting. (A) Representative PCR genotyping results showing untargeted, heterozygous and homozygous targeted clones across the three SSN platforms. WT CTL=wild-type control. (B) Representative Southern blot results showing a heterozygous targeted clone and a homozygous targeted clone detected with a 3' external probe. Fragment sizes: WT-6.5 kb, Edited-6.9 kb. (C) Representative Southern blot results showing a properly edited clone and a heterozygous clone with a non-random double-integration. Fragment sizes: Properly edited-6.9 kb, aberrant additional integration-5 kb. Please click here to view a larger version of this figure.
Study | Gene Targeted | Platform | Repair template type | # of clones picked | Targeting efficiency |
Sexton et al., 2014 | TPP1 | ZFN | GFP-Puro | unreported | unreported |
Sexton et al., 2014 | TERT | ZFN | Hygromycin | unreported | unreported |
Hockemeyer et al., 2009 | POU5F1 | ZFN | GFP-Puro | 31 | 39.0% |
Hockemeyer et al., 2009 | PITX3 | ZFN | GFP-Puro | 74 | 14.9% |
Hockemeyer et al., 2011 | POU5F1 | TALEN | GFP-Puro | 68 | 91.0% |
Hockemeyer et al., 2011 | PITX3 | TALEN | GFP-Puro | 96 | 13.0% |
Merkle et al., 2015 | VASA | Cas9 | Reporter-Geneticin | 139 | 94.0% |
Merkle et al., 2015 | CRH | Cas9 | Reporter-Geneticin | 30 | 93.0% |
Merkle et al., 2015 | HCRT | Cas9 | Reporter-Geneticin | 154 | 92.0% |
Merkle et al., 2015 | HMX2 | Cas9 | Reporter-Geneticin | 11 | 45.0% |
Forster et al., 2014 | LRG5-Nterm | ZFN | GFP-Puro | unreported | 30.0% |
Forster et al., 2014 | LRG5-Cterm | ZFN | GFP-Puro | unreported | 14.0% |
Soldner et al., 2011 | SNCA | ZFN | Puro | 96 | 1.0% |
Table 1. Description of other genes targeted using this method, with corresponding targeting efficiencies taken from previously published studies. 4,5,38-41
Locus | Sequence | Notes |
AAVS1-F primer | CTCTAACGCTGCCGTCTCTC | PCR conditions: Tm = 57 °C, 35 cycles |
AAVS1-WT-R primer | GCTTCTCCTCTTGGGAAGTG | WT band: 1273 bp |
AAVS1-Targeted-R primer | CGTCACCGCATGTTAGAAGA | Targeted band: 992 bp |
T2-Cas9-guide | GGGCCACTAGGGACAGGAT | From Mali et al., 2013 |
AAVS1-ZFN-Right | TAGGGACAGGAT | From Hockemeyer et al., 2009 |
AAVS1-ZFN-Left | TGGGGTGTCACC | From Hockemeyer et al., 2009 |
AAVS1-TALEN-Right | TCCTAACCACTGTCTTT | From Hockemeyer et al., 2011 |
AAVS1-TALEN-Left | CCCCTCCACCCCACAGT | From Hockemeyer et al., 2011 |
Table 2. List of primers and SSN targeting sequences.
Targeting Construct | Number of EGFP+ Colonies | Targeted picked clones (PCR verified) | Previous Reported Correct Targeting Efficiency (by Southern blot) |
ZFN | 150 | 86.9% (73.9% het/ 13.0 % homo) | 56% (50% het/6% homo) |
TALEN | 412 | 91.3% (47.8% het/39.1% homo) | 47% (37.5% het/9.3% homo) |
CRISPR-Cas9 | 235 | 95.7 % (69.5% het/26.3% homo) | unreported |
Table 3. Comparative numbers of EGFP-positive and targeted TALEN, ZFN and CRISPR/Cas9 human stem cell colonies. PCR verified integrations at the AAVS1 locus for this experiment are compared to Southern blot verified proper single integrations in previous experiments 4,5.
The method presented here for isolating homogenous populations of gene-edited human pluripotent stem cells is a powerful approach for generating isogenic hPSC lines that differ only at the targeted locus. These cells are an ideal system for probing the mechanisms of human cellular differentiation and development as well as for understanding the pathophysiology of monogenic diseases in a controlled genetic setting. As demonstrated here, it is possible to use three independent SSN design strategies (ZFNs, TALENs, and CRISPR/Cas9) to achieve targeted integration at the AAVS1 locus. Each of these methodologies has its own advantages and disadvantages. A potential advantage of ZFNs and, to some extent, TALENs is their design flexibility, which allows iterative engineering in order to improve the DNA binding domains of individual nucleases 42. This nuclease optimization could increase the specificity of ZFNs and TALENs beyond what is achievable with the CRISPR/Cas9 system. Such selectivity may be important for clinical applications requiring a high degree of target specificity. The primary advantage of the CRISPR/Cas9 system is its ease of use. Although TALEN and ZFN construction kits have been made available to the public (i.e., through Addgene 21), CRISPR/Cas9-based SSNs are significantly easier to construct, as the only necessary customization is a 20 base pair oligonucleotide (when using the px330 plasmid design 14). This simplicity proves to be advantageous for research labs looking to include genome editing in their studies.
There are alternative transfection techniques, including nucleofection 43, to create gene-edited hPSC lines; however electroporation has been shown to be consistent and cost effective 4,5,25. Nucleofection can be used to directly transfect Cas9-guide RNA ribonucleoprotein complexes into the nucleus, increasing the SSN efficiency and fidelity 44. Growing hPSCs on MEFs is a robust and inexpensive method to maintain hPSCs in a pluripotent state without excessive differentiation. In addition, it allows for the easy isolation of genetically identical colonies. Alternatively, it is possible to culture hPSCs without MEFs, however these culture conditions can be more expensive than feeder based cultures. Furthermore, the whole process is scalable, allowing for the isolation of very rare editing events or the generation of many distinct cell lines in parallel editing experiments.
The protocol described here is robust; however there are several key steps that affect the efficiency with which correctly edited clones can be obtained. The most critical component for this method is having high quality MEFs and drug resistant DR4 MEFs. The survival of single hPSCs is tenuous, and low-quality MEFs will impede the isolation of undifferentiated hPSC lines. Second, the use of Y-27632 is also key to allow single cell survival without generating a selective pressure for cells with an abnormal karyotype 45. Third, picking well-spaced colonies ensures genetic homogeneity of the derived cell lines. Finally, it is important to prepare glass pipettes so that the opening is small enough to break the colony into many smaller pieces, ensuring that multiple colonies will grow in the new well. This permits the picking of well-spaced subclones for a replica plate to have in culture, leaving the original plate of isolated colonies for genotyping. A challenging part in this protocol is the manual pulling of the glass pipettes; this should be practiced beforehand. It should be noted that there are multiple techniques of clone picking that do not require a glass pipette. Experimenters are encouraged to find the one that works best for them.
There are limitations to this protocol that can be overcome by simple modifications. An experiment intended to only disrupt the locus of interest without repair or one whose repair template does not contain a selection cassette, must use another method of enriching for cells that were edited. Note that the number of colonies that must be picked to find a positive clone increases greatly in the absence of selection. One strategy to improve the efficiency is to co-transfect a non-integrating plasmid that expresses a fluorescent protein. After allowing the cells to recover for two days, targeted cells can be sorted on positive fluorescence using fluorescence-activated cell sorting and replated 29. This process enriches for cells that have been transfected with the plasmids by electroporation and thereby increases the probability of a concurrent editing event in the cell.
The techniques described here can be extended to use multiple guide RNAs to simultaneously target several loci 14,46,47. Many protocols have been established to differentiate hPSCs into distinct cell types, allowing various genetic manipulations in cell types of interest 30. Overall, we have demonstrated the potential for efficient genome editing in hPSCs regardless of SSN choice. We propose that this technique can be adapted to create isogenic hPSC lines that have been gene-edited at any genomic locus.
The authors have nothing to disclose.
This work was supported by a Brain Research Foundation Seed Grant (BRFSG-2014-02) to Helen Bateup. Dirk Hockemeyer is a New Scholar in Aging of the Ellison Medical Foundation and is supported by the Glenn Foundation as well as the The Shurl and Kay Curci Foundation. DH is also supported by NIH grant 1R01CA196884-01.
DMEM/F12 | Life Technologies | 11320082 | |
Fetal Bovine Serum (HI) | Life Technologies | 10082-147 | |
Knockout Serum | Life Technologies | 10828-028 | |
Fibroblast Growth Factor – basic | Life Technologies | PHG0261 | |
Pen/Strep | Life Technologies | 15140-122 | |
Glutamine | Life Technologies | 25030-081 | |
MEM NEAA | Life Technologies | 11140-050 | |
2-mercaptoethanol | Life Technologies | 21985-023 | |
Y-27632 | Calbiochem | 688000 | |
6-well plates | Corning | 3506 | |
12-well plates | Corning | 3512 | |
4mm Electroporation cuvettes | Bio-rad | 165-2081 | |
X-cel gene pulser II | Bio-rad | 165-2661 | |
0.25 % Trypsin-EDTA | Life Technologies | 25200-056 | |
10× Phosphate buffered saline (PBS) pH7.4 | Life Technologies | 70011-044 | |
Puromycin | Life Technologies | A11138-02 | |
Pasteur pipettes, plugged | VWR | 14672-412 | |
Tris-HCl | Sigma-Aldrich | S5941 | |
NaCL | Sigma-Aldrich | S9888 | |
SDS | Sigma-Aldrich | L3771 | |
EDTA | Sigma-Aldrich | E5134 | |
Proteinase-K | Life Technologies | AM2544 | |
Ethanol | VWR | TX89125-172SFU | |
Isopropyl Alcohol | VWR | MK303216 | |
TE Buffer | Life Technologies | 12090-015 | |
Dimethyl Sulfoxide (DMSO) | Sigma-Aldrich | D8418 | |
1.8 ml Cryotubes | ThermoScientific | 377267 |