Generation of Sex-Reversed Female Clonal Mice via CRISPR-Cas9-Mediated Y Chromosome Deletion in Male Embryonic Stem Cells

Abstract

Mice derived entirely from embryonic stem (ES) cells can be generated through tetraploid complementation. Although XY male ES cell lines are commonly used in this system, occasionally, monosomic XO female mice are produced through spontaneous Y chromosome loss. Here, we describe an efficient method to obtain monosomic XO ES cells by CRISPR-Cas9-mediated deletion of the Y chromosome, allowing generation of female clonal mice by tetraploid complementation. The monosomic XO female mice are viable and able to produce normal male and female offspring. Direct generation of clonal mice in both sexes can significantly accelerate the production of complex genetically modified mouse models.

Introduction

Genetically modified (GM) animals are essential tools for the study of both fundamental biology and human diseases. The production of GM animals relies on two critical features: (1) stable genome modifications and (2) germline transmission of the mutations into a model system. A typical approach for creation of complex GM mice involves the generation of tetra-parental chimeras from normal embryos and GM embryonic stem (ES) cells, followed by multiple rounds of breeding to obtain both male and female homozygotes for germline propagation of the mutations. This process is time-consuming, laborious, and costly, particularly if the final objective requires many independent germline manipulations in the same animal.

Mouse ES cells derived from the inner cell mass of blastocysts have unlimited self-renewal and differentiation capacity if maintained in their ground-state pluripotency.^1–3 Pure ES cell-derived mice (all-ES mice) can be directly and efficiently generated through tetraploid complementation, in which ground-state ES cells are injected into tetraploid blastocysts such that the host 4n cells can only contribute to the extra-embryonic tissue but not somatic tissues.^4–6 In this system, by design, most viable animals are male. However, fertile female all-ES mice (39 chromosome, XO) are occasionally produced from the male ES cell lines (∼2%) through spontaneous Y chromosome loss.⁷ Although the monosomic XO female (39, XO) mice have been proposed for the use of GM mice production to avoid mutant allele segregation during outcrossing,⁷ the observed low frequency makes it impractical for routine use in transgenic facilities.

Here, we present a novel CRISPR-Cas9-mediated approach for directed elimination of the Y chromosome from mouse ES cells permitting efficient generation of monosomic XO female clonal mice by tetraploid complementation. The obtained monosomic XO female clonal mice are viable and fertile, and produce offspring of both sexes when crossed to male clonal mice from the same ES cells.

Methods

Mice and embryos

Animals were housed and prepared according to the protocol approved by the Institutional Animal Care and Use Committee of Weill Cornell Medical College (protocol number: 2014-0061). Wild-type ICR mice were purchased from Taconic Farms (Germantown, NY). Females were superovulated at 6–8 weeks with 0.1 mL CARD HyperOva (Cosmo Bio Co., Ltd., Tokyo, Japan; cat. no. KYD-010-EX) and 5 IU human chorionic gonadotrophin (hCG; Sigma–Aldrich, St. Louis, MO) at intervals of 48 h. The females were mated individually to males and checked for the presence of a vaginal plug the following morning. Plugged females were sacrificed at 1.5 days post hCG injection to collect two-cell embryos. Embryos were flushed from the oviducts with KSOM + AA (Specialty Media) and subjected to electrofusion to induce tetraploidy. Fused embryos were moved to new KSOM + AA micro drops covered with mineral oil and cultured further in an incubator under 5% CO₂ at 37°C until blastocyst stage for ES cell injection.

Blastocyst injection

ES cells were trypsinized, re-suspended in ES medium, and kept on ice. A flat-tip microinjection pipette was used for ES cell injection. ES cells were picked up in the end of the injection pipette, and 10–15 of them were injected into each blastocyst. The injected blastocysts were kept in KSOM + AA until embryo transfer. Ten injected blastocysts were transferred into each uterine horn of 2.5 dpc pseudo-pregnant ICR females.

ES cell-line derivation

ES cell lines were derived from hybrid F1 fertilized embryos of crossing between C57BL/6J females and 129S1 males. The ES cell derivation medium comprises 75 mL KnockOut™ Dulbecco's modified Eagle's medium (Gibco, Grand Island, NY; cat. no. 10829-018), 20 mL KnockOut™ Serum Replacement (SR; Gibco; cat. no. 10828), 1 mL penicillin/streptomycin (Specialty Media, cat. no. TMS-AB-2C), 1 mL l-glutamine (Specialty Media, cat. no. TMS-001-C), 1 mL nonessential amino acids (Specialty Media; cat. no. TMS-001-C), 1 mL nucleosides for ES cells (Specialty Media; cat. no. ES-008-D), 1 mL β-mercaptoethanol (Specialty Media, cat. no. ES-007-E), 250 μL PD98059 (Promega, Madison, WI; cat. no. V1191), and 20 μL recombinant mouse LIF (Chemicon International, Temecula, CA; cat. no. ESG1107). The procedure to derive ES cell lines has been described previously.⁴ Briefly, cell clumps originated from the blastocysts were trypsinized in 20 μL 0.025% trypsin and 0.75 mM EDTA (Specialty Media; cat. no. SM-2004-C) for 5 min, and 200 μL ES medium was added to each well to stop the reaction. Colony expansion of ES cells proceeded from 48-well plates to six-well plates with feeder cells in ES medium, and then to gelatinized 25 cm² flasks for routine culture in regular ES culture medium. Cell aliquots were cryopreserved using Cell Culture Freezing Medium (Specialty Media; cat. no. ES-002-D) and stored in liquid nitrogen.

CRISPR Cas9, gRNA, and ES cell electroporation

Two crRNAs (IDT, Coralville, IA) targeting the Rbmy1a1 gene at sequences TTCAAGTGATGATGGTCTCCTGG and TCCTTCATGTGAAGGGAACTTGG (including 3′ “NGG” protospacer adjacent motif)⁸ were annealed to a tracrRNA (IDT, cat# 1072533) at a 1:1:2 molar ratio to form dual duplex gRNAs by heating at 95°C for 5 min and then cooled to room temperature. Duplex gRNAs were then incubated with recombinant Cas9 protein (IDT; HiFi Cas9 nuclease V3, cat. no. 1081060) at room temperature for 20 min to form gRNA-Cas9 ribonucleoproteins (RNPs), followed by co-delivery with a green fluorescent protein (GFP) plasmid (Addgene, Watertown, MA; cat. no. 42028). All RNAs were in Duplex Buffer (30 mM HEPES, pH 7.5; 100 mM potassium acetate) and DNA in TE buffer (10 mM Tris-HCl, 0.1 mM EDTA, pH 7.5). The final concentration for each electroporation is 1.8 μM gRNA and 1.5 μM Cas9 nuclease.

To prepare ES cells for electroporation with the Neon Transfection System (Thermo Fisher Scientific, Waltham, MA), the cells were collected by trypsinization from culture, washed twice with phosphate-buffered saline (without Ca2+ and Mg2+), and re-suspended in supplied R buffer at 100,000 cells/10 μL. The cell suspension (10 μL) was mixed with 0.5 μL GFP plasmid DNA (27–270 mM) and 0.5 μL gRNA-Cas9 RNPs, and 10 μL of the mix was loaded to a Neon 10 μL tip for electroporation. Our optimized program is #14 (1,200 V, 20 s, two pulses). Treated cells were placed in a gelatin-coated 24-well with pre-warmed ES medium and returned to regular culture conditions.

Polymerase chain reaction genotyping

Polymerase chain reaction (PCR) was performed on genomic DNA extracted from cell pellets or mouse tails to determine the loss of the Y chromosomal genes Uba1y and Ssty1 by utilizing the KAPA Mouse Genotyping Kit (Roche, Basel, Switzerland; cat. no. KK7302). Specific procedures were followed according to the reagent instructions. PCR products were separated on 2–3% agarose gel in TBE buffer, and inverted ethidium bromide stained images are shown. Cycling conditions were as follows: 95°C for 3 min followed by 40 cycles of 95°C for 15 s, 60°C for 15 s, and 72°C for 120 s.

Uba1/Uba1y primers: forward, TGGATGGTGTGGCCAATG; reverse, CACCTGCACGTTGCCCTT (335 bp product for Y-linked Uba1y, 253 bp for X-linked Uba1).⁹

Ssty1 primers: forward, GCCACTATAGCTGGATTATGAG; reverse, GTCTTCACATCAGAGGTTCTAC (1,444 bp product).⁸

Quantitative PCR of genomic DNA

To distinguish XX and XO female mice, the X chromosome dosage was determined by measuring the abundance of the X chromosome resident Bcor gene relative to Actb in mouse tail genomic DNA using quantitative PCR with PowerUp SYBR Green Master Mix (Thermo Fisher Scientific; cat. no. A25778). The relative abundance of Bcor is presented as 1000*2^{(Ct(Bcor)-Ct(Actb))}, where Ct is cycle threshold. Among several X-linked genes tested with this assay, only Bcor abundance in XX genome is consistently twice that in XY or karyotype-confirmed XO genome.

Bcor primers: forward, TTTCCCACTCCATCCCCGACTAGTT; reverse, TCCCAAATAAACACCAGAGGCGACA.

Actb primers: forward, GATATCGCTGCGCTGGTCGT; reverse, CCCACGATGGAGGGGAATACAG

Results

Elimination of Y chromosome on ES cells

To eliminate the mouse Y chromosome, we used synthetic gRNAs to target the Rbmy1a1 gene sequences (Fig. 1A). We included a circular GFP reporter plasmid with the Cas9 cocktail to allow for validation of successful electroporation. After ES cell colonies formed 48 h post electroporation, they were trypsin digested, and single GFP+ cells were manually picked with the aid of a micromanipulator under a fluorescent microscope (Fig. 1B and Supplementary Fig. S1A). The GFP+ cells were individually plated in 96-well dishes for clonal expansion over feeder layers (Fig. 1B and Supplementary Fig. S1A). Cell colonies usually emerged after a week in culture, and were expanded in gelatin-coated cultures for one to two passages before genotyping (Supplementary Fig. S1B).

FIG. 1.

Efficient elimination of the Y chromosome in mouse embryonic stem (ES) cells using CRISPR-Cas9. (A) Mouse Y chromosome and indicated relevant genes. (B) Scheme illustration of CRISPR-mediated Y chromosome elimination in mouse ES cells. (C) Y chromosome deletion rates as determined by Uba1y loss in targeted cells with varying concentrations of co-transfection green fluorescent protein (GFP) plasmid. BS: blinded selection.

We first optimized our electroporation parameters by including a high dose of GFP plasmid DNA (123 nM, i.e., 500 ng/reaction) in the reaction, along with gRNA-Cas9 RNPs at the supplier-recommended concentration (1.5 μM). Under optimal electroporation conditions, >20% of the cells expressed GFP with minimal cell death. The GFP+ cell population did not vary in GFP plasmid concentrations ranging from 300 to 500 ng, although lower plasmid concentrations resulted in significantly decreased electroporation efficiency (Supplementary Fig. S1C).

We initially determined the state of the Y chromosome by genomic DNA PCR analyses for the presence or loss of the Uba1y gene located on the short arm on the Y chromosome and distal to the targeted Rbmy1a1 gene (Fig. 1A). Of 19 clones, three (16%) showed loss of the Uba1y gene in the 500 ng electroporation group (Fig. 1C and Supplementary Fig. S1D). Similarly, 5/27 (19%) clones showed Uba1y gene loss in the 250 ng plasmid electroporation (Fig. 1C and Supplementary Fig. S1E).

As Cas9 concentration is fixed in our system, next we lowered the GFP plasmid concentration in an effort to improve the targeting efficiency. Indeed, at a GFP plasmid concentration of 150 ng, we found that 7/8 (87%) GFP+ subclones had lost the Uba1y gene (Fig. 1C and Fig. 2A), and at the lowest GFP plasmid concentration (50 ng), 100% (20/20) of the subclones exhibited Uba1y gene loss (Fig. 1C and Fig. 2D).

FIG. 2.

Genotyping and karyotyping analyses of Y chromosome–deleted ES cells. (A) Genomic polymerase chain reaction (PCR) analyses of Uba1 and Uba1y genes in GFP+ subclones with 150 ng GFP plasmid (36.9 nM). Female (F) and male (M) control DNAs were included. (B) Genomic PCR analyses of Uba1, Uba1y, and Ssty1 genes in GFP+ subclones with 50 ng GFP plasmid (12.3 nM). (C) Genomic PCR analyses of Uba1 and Uba1y genes in GFP– subclones with 150 ng of GFP plasmid. (D) Karyotype analysis shows absence of the Y chromosome in a targeted ES cell (39, XO).

We also examined the status of GFP– subclones in the conditions when essentially ∼90% of GFP+ clones demonstrated efficient CRISPR-Cas9 targeting to assess whether the GFP reporter had value as a surrogate marker. While 7/8 GFP+ cell subclones yielded Uba1y gene loss, 0/11 GFP– subclones had Uba1y gene loss in the 150 ng group (Fig. 2A and C). In addition, of the 24 subclones derived from the cells blindly selected in the 50 ng group, only one had Uba1y gene loss (Fig. 1C). These data confirm the values of the GFP plasmid in the appropriate ratio as a co-electroporation surrogate marker for successful gene transfer and likely successful gene targeting.

To assess further whether loss of the Uba1y gene effectively reflects deletion of the entire Y chromosome, we examined for retention of the Ssty1 gene (>35 copies) that is located on the long arm of the Y chromosome (Fig. 1A). PCR analysis indicated the concomitant loss of the Y chromosome long-arm gene, Ssty1, in each GFP+ subclone demonstrated to lack the Uba1y gene (Fig. 2B). Thus, loss of both Uba1y and Ssty1 genes indicates that a large DNA segment, including the centromere of the Y chromosome, has been deleted. We further karyotyped the cell subclones with confirmed loss of both Uba1y and Ssty1 genes using metaphase analyses. In the 20 metaphase chromosome spreads analyzed from clone 1, all cells displayed complete loss of Y chromosome, with 18 cells having the expected 39 chromosome, XO karyotype (90%), while two cells were found with a segmental loss of 1q on chromosome 1 (Fig. 2D and Supplementary Fig. S1E). In a second clone (clone 2), all 22 cells examined showed XO karyotype, with three cells showing additional abnormalities, including an additional loss of 15q, chromosome 13, and an extra chromosome 14 (Supplementary Fig. S1F). These are likely de novo mutations arising during clonal cell expansion. In aggregate, these results confirm an efficient methodology for the physical elimination of the Y chromosome by targeting double-strand breaks (DSBs) in the Rbmy1a1 gene in ES cells with minimal additional karyotypic perturbation.

Generation of fertile female mice from monosomic XO ES cells

The next critical step was to assess the potential and efficiency of CRISPR-Cas9 monosomic XO ES cells to generate all-ES mice by tetraploid complementation. ES cells with confirmed loss of the Uba1y and Ssty1 genes were cultured and expanded for three to eight passages (total passages 16–21). We selected 6 of the 20 subclones (ES-1) with confirmed loss of the Uba1y gene (Fig. 2B) for blastocyst injection. Single-cell clones from the parental ES cells without targeting were used as control. Of the six subclones tested, two clones gave rise to live pups, with efficiency ranging from 20% to 25% of the embryos transferred. A similar frequency (4/7 subclones) and efficiency (11–12%) giving rise to all-ES mice were obtained for single-cell clones of the parental ES cells (Fig. 3A, Supplementary Table S1, and Supplementary Table S2). These data indicate that the preceding CRISPR-Cas9 manipulation of the ES cells did not adversely affect their pluripotency. In another test of four subclones generated from ES lines 2 and 3 (two subclones for each cell line; Fig. 1C), three were able to generate live pups with similar efficiency to their parental ES cells by tetraploid complementation (Fig. 3A). As expected, pups obtained from the Y-deletion subclones were all of monosomic XO (39, XO) genotype and developed a morphologically normal female external genital anatomy (Fig. 3B and C). Meanwhile, pups produced from the parental ES cells were all exclusively males (Fig. 3C).

FIG. 3.

Viable and fertile monosomic XO female clonal mice generated from 39, XO ES cells. (A) Developmental potential of Y chromosome deletion ES cells by tetraploid complementation. (B) A litter of five XO all-ES pups from targeted subclone (ES-2). (C) XO all-ES mice show a typical female external genital anatomy at 2 weeks and 2 months, while all-ES mice from the parental ES cells show male external genital morphology. (D) XO female clonal mice produce normal offspring when mated with clonal male all-ES mice. (E) Quantitative PCR quantification of genomic abundance of the X chromosome Bcor gene in indicated genomes of the offspring from litter 1. (F) Normal male and female as well as XO offspring are produced by XO female clonal mice with a smaller litter size.

We further investigated monosomic XO female clonal mouse fertility by breeding with clonal males from the same parental ES cell lines. All XO female clonal mice were fertile and delivered normal male and female offspring but with fewer littermates (4–8 pups; Fig. 3D–F and Supplementary Table S3). The XO mice generate two types of oocytes—X oocytes (19, X) and O oocytes (19, O)—that result in four different genotypes following fertilization (XX, XO, XY, and OY). The X chromosome harbors essential developmental genes, and therefore OY embryos are expected to fail during embryonic development. To distinguish XO females from XX females in the offspring, we used quantitative PCR to analyze copy number of an X chromosome single copy gene, Bcor (Fig. 3E). Monosomic XO females were also present among the offspring from the following generation with a frequency of one of three or four females (Fig. 3F and Supplementary Table S3). These numbers are lower than the expected Mendelian frequency of XO mice among females, given that half of the females are expected to be XO. These results suggest that monosomic XO embryos may be less robust than XX embryos during development. Partial loss of XO embryos and embryonic lethality of OY embryos during the developmental stage would therefore account for the smaller litter size from XO mice. More importantly, when bred with clonal males, the XO female clonal mice give rise to both male and female offspring with normal genotypes (XY and XX; Fig. 3E and F).

Discussion

We derived new ES cell lines from hybrid F1 embryos by crossing C57BL/6J females with 129S1 males. The resultant male ES cell lines used in this study were all confirmed to produce normal all-ES mice by tetraploid complementation. Previous studies have demonstrated that targeted chromosomal generation of multiple DNA DSBs using CRISPR-Cas9 can induce directed chromosomal deletion.^8,10 Thus, to eliminate the mouse Y chromosome, we targeted the RNA-binding motif gene Rbmy1a1, which has more than 50 copies exclusively clustered on the short arm of the Y chromosome.¹¹ We used synthetic gRNAs to target the Rbmy1a1 gene sequences that have been successfully targeted to eliminate the Y chromosome in mouse embryos and ES cells.⁸ Here, we used purified Cas9 proteins with nucleus localization signals that can form functional gRNA–Cas9 ribonucleoprotein complexes (RNPs) in vitro. The use of preassembled gRNA–Cas9 RNPs allows for more accurate control of RNP composition and doses and has been shown to reduce off-target effects and cytotoxicity effectively in mammalian cells.^12,13

Electroporation is widely used to deliver RNPs due to the simplicity and large capacity.¹⁴ We included a circular GFP reporter plasmid with the Cas9 cocktail to allow for validation of successful electroporation. GFP expression could also serve as a surrogate albeit indirect marker for Cas9-induced Y chromosome elimination—indirect because plasmids and RNPs have differing modes of cell entry. The approximately 20% efficiency likely reflects conditions in which either: (1) CRISPR-induced chromosome loss occurs only in certain highly electroporation-receptive cells where excessive genomic breaks overwhelm cellular DNA repair capability, or (2) the excess uptake of GFP plasmids over Cas9 proteins could possibly generate a GFP+ cell insufficient Cas9 to induce enough DSBs for chromosome elimination, resulting in a GFP “blank-reporter.” Our data confirm the values of the GFP plasmid in the appropriate ratio as a co-electroporation surrogate marker for successful gene transfer and likely successful gene targeting.

When ES cells with confirmed loss of the Uba1y and Ssty1 genes were used to perform tetraploid complementation, we obtained all-ES mice. Our data indicate that the preceding CRISPR-Cas9 manipulation of the ES cells did not adversely affect their pluripotency. In agreement with other studies,^7,8 our monosomic XO female pups developed and matured normally to adulthood without noticeable defects. These results demonstrate the feasibility of efficient deletion of the Y chromosome from mouse ES cells using CRISPR-Cas9 technology, allowing generation of male and female clonal mice from the same ES cell line.

Finally, we confirmed that the production of normal XX and XY offspring from XO clonal mice indicates the feasibility to use XO all-ES mice for germline propagation of the complex genetic mutations in mouse models.

Conclusion

We demonstrate that targeting the Y chromosome single multi-copy Rbmy1a1 gene in XY male ES cells using CRISPR-Cas9 technology can efficiently eliminate the Y chromosome. Importantly, the resultant 39 chromosome XO ES cells retain pluripotency and can generate viable and fertile all-ES mice that are phenotypically sex-reversed from male to female. This system provides a practical strategy to manipulate sex in mice via ES cells, making it possible to expedite the production of complex multi-transgene GM mouse models that are a frequent necessity in current biomedical research, avoiding the time-consuming chimera development step as well as the complex breeding process (Fig. 4A and B).

FIG. 4.

Proposed approach to expedite the production of complex genetically modified (GM) mouse models. (A) Expedite production of GM mice through the generation of XO female GM mice. (B) Overview of the time needed for the proposed approach versus the conventional chimera approach.

Author Disclosure Statement

The authors declare no competing interests.

Funding Information

This work was funded by grants 1 R01 GM129380-01 from the National Institutes of Health and New York State Stem Cell Science Program (NYSTEM Contract C32581GG) to D.C.W.

Supplementary Material

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Supplementary Figure S1