Generation of Mouse Haploid Somatic Cells by Small Molecules for Genome-wide Genetic Screening

Resource

Generation of Mouse Haploid Somatic Cells by Small Molecules for Genome-wide Genetic Screening Graphical Abstract

Authors Zheng-Quan He, Bao-Long Xia, Yu-Kai Wang, ..., Liu Wang, Wei Li, Qi Zhou

Correspondence [email protected]

In Brief He et al. show that CDK1 and ROCK inhibition can suppress the spontaneous diploidization of haploid embryonic stem cells (haESCs) and generate haploid somatic cells of all three germ layers for genome-wide genetic screening.

Highlights d

Mitotic slippage causes spontaneous diploidization of mouse haESCs

d

Inhibition of CDK1 and ROCK suppresses the diploidization of mouse haESCs

d

ROCK inhibition can generate haploid somatic cells of all three germ layers

d

Haploid neural stem-cell-like cells can be used for genomewide genetic screening

He et al., 2017, Cell Reports 20, 2227–2237 August 29, 2017 ª 2017 The Authors. http://dx.doi.org/10.1016/j.celrep.2017.07.081

Accession Numbers GSE101795

Cell Reports

Resource Generation of Mouse Haploid Somatic Cells by Small Molecules for Genome-wide Genetic Screening Zheng-Quan He,1,2,5 Bao-Long Xia,1,5 Yu-Kai Wang,1,5 Jing Li,1,3 Gui-Hai Feng,1,2 Lin-Lin Zhang,1,2 Yu-Huan Li,1,2 Hai-Feng Wan,1 Tian-Da Li,1 Kai Xu,1,2 Xue-Wei Yuan,1,4 Yu-Fei Li,1,2 Xin-Xin Zhang,1,4 Ying Zhang,1 Liu Wang,1,2 Wei Li,1,2 and Qi Zhou1,2,6,* 1State

Key Laboratory of Stem Cell and Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, Beijing 100101, China of Chinese Academy of Sciences, Beijing 100049, China 3School of Life Sciences, University of Science and Technology of China, Hefei 230026, China 4College of Life Science, Northeast Agricultural University, Harbin 150030, China 5These authors contributed equally 6Lead Contact *Correspondence: [email protected] http://dx.doi.org/10.1016/j.celrep.2017.07.081 2University

SUMMARY

The recent success of derivation of mammalian haploid embryonic stem cells (haESCs) has provided a powerful tool for large-scale functional analysis of the mammalian genome. However, haESCs rapidly become diploidized after differentiation, posing challenges for genetic analysis. Here, we show that the spontaneous diploidization of haESCs happens in metaphase due to mitotic slippage. Diploidization can be suppressed by smallmolecule-mediated inhibition of CDK1 and ROCK. Through ROCK inhibition, we can generate haploid somatic cells of all three germ layers from haESCs, including terminally differentiated neurons. Using piggyBac transposon-based insertional mutagenesis, we generated a haploid neural cell library harboring genome-wide mutations for genetic screening. As a proof of concept, we screened for Mn2+-mediated toxicity and identified the Park2 gene. Our findings expand the applications of mouse haploid cell technology to somatic cell types and may also shed light on the mechanisms of ploidy maintenance. INTRODUCTION Haploid cells are valuable resources for high-throughput genetic analysis, as they lack the second gene allele and, thus, can show the recessive phenotypes effectively (Yi et al., 2009). However, unlike in yeast and plants, haploids are very rare in mammals. Near-haploid cell lines were established from human tumors with massive chromosome loss; however, such cancer cell lines were genetically very unstable €rckstu €mmer et al., 2013). Recently, haploid embryonic (Bu stem cells (haESCs) with pluripotency and self-renewal ability have been derived from parthenogenetic and androgenetic haploid embryos in various mammalian species, including the

mouse (Elling et al., 2011; Leeb et al., 2012; Leeb and Wutz, 2011), rat (Li et al., 2014), monkey (Yang et al., 2013), and human (Sagi et al., 2016; Zhong et al., 2016). Although experiencing spontaneous diploidization during culture, the haESCs can maintain the haploidy indefinitely in culture through continuous purification of haploid cells using fluorescence-activated cell sorting (FACS). These unique characteristics make mammalian haESCs a powerful tool for large-scale genetic analysis; for instance, genome-wide screening for genes regulating pluripotency maintenance (Leeb et al., 2014) and X chromosome inactivation (Monfort et al., 2015). However, the haESCs with an embryo origin are mainly applied for studies related to early embryonic development, whereas the haploid somatic cells are more suitable for genetic analysis of development and disease of adult tissues. Nevertheless, the diploidization occurs very rapidly after differentiation of mouse and rat haESCs, which cannot be blocked using FACS-assisted purification (Leeb and Wutz, 2011; Li et al., 2014). Recent studies have reported the derivation of human haploid somatic cells of three germ layer (Sagi et al., 2016) and mouse haploid neurons (Xu et al., 2017) from the haESCs. However, it remains elusive whether the rodent haESCs can differentiate in a haploid state to generate haploid somatic cells of all three germ layers and to serve as an alternative approach for genetic screening. Mitotic slippage was a biological process through which the cells at mitotic phase eventually exit mitosis and re-enter the next G1 phase without undergoing chromosome segregation or cytokinesis (Lanni and Jacks, 1998; Minn et al., 1996). Here, we found that the mitotic slippage was the primary cause of the diploidization of mouse haESCs through cell-cycle analysis, using live cell imaging, and inhibition of the CDK1 or ROCK pathway could efficiently suppress the diploidization during culture and differentiation. Importantly, the mouse haESCs could differentiate into haploid somatic cell types representative of all three germ layers, including terminally differentiated haploid neurons. As a proof-of-concept, we also demonstrated the feasibility of using the haploid neural cells for a forward genetic screen of neurotoxin resistance, by which we identified park2 regulating intracellular accumulation of Mn2+.

Cell Reports 20, 2227–2237, August 29, 2017 ª 2017 The Authors. 2227 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Figure 1. Live-Cell Imaging Reveals Mitotic Slippage as a Cause of Diploidization of haESCs (A) Representation of DNA content analysis of spontaneous diploidization of haESCs during culture (for 6 passages) and differentiation by FACS. There were at least 10 independent repeated experiments for haESCs cultured in vitro and at least 6 independent repeated experiments for haploid ESCs differentiation in vitro. (B) Live-cell imaging of a dividing diploid ESC (top) and a diploidized haploid ESC (bottom). The cells were labeled with an H2B-GFP reporter. Metaphase was taken as the starting point. Scale bars, 5 mm. (C) Quantitative analysis of the ratios of cells experiencing mitotic slippage in control diploid ESCs (n = 152; 2 different diploid ESC lines, H2B-GFP-1 and H2B-GFP-2, in 2 independent experiments) and haploid ESCs (n = 163; 2 difference haESC lines, AH-H2B-GFP-1 and AHH2B-GFP-2, in 3 independent experiments) as observed in (B) during 20 hr. Error bars represent mean ± SD. See also Figure S1 and Tables S1 and S2.

RESULTS Spontaneous Diploidization of Mouse haESCs Is Induced by Mitotic Slippage DNA content analysis using flow cytometry showed that around 90% of mouse parthenogenetic and androgenetic haESCs became diploidized after 6 passages, and nearly all the haESCs became diploidized after differentiation into neural cells (Figure 1A). To investigate at which stage of the cell cycle the autodiploidization happened, we used real-time live-cell imaging to trace the division of haploid and diploid embryonic stem cells (ESCs) expressing the histone H2B-GFP fusion protein. While all the diploid ESCs could rapidly enter the anaphase and divide into two cells (n = 152), a small ratio of haESCs (
9%; n = 163) were blocked at metaphase and re-entered G1 phase without division, indicating the occurrence of mitotic slippage due to the failed transition from metaphase to anaphase (Figures 1B, 1C, and S1; Tables S1 and S2). These results suggested that the diploidization of haESCs was caused by mitotic slippage. Inhibitors of CDK1 and ROCK Suppress Spontaneous Diploidization of haESCs Next, we aimed to screen for small molecules that could promote maintenance of haploidy of haESCs. We started screening with

2228 Cell Reports 20, 2227–2237, August 29, 2017

10 chemical inhibitors targeting genes that regulated the spindle assembly and mitotic division (Figure 2A; Table S3). They were added into the culture medium, and their effect on haploidy maintenance was tested after 4 passages. Most of these small molecules induced massive cell death, polyploidization, or accelerated the diploidization of haESCs (Figures 2A and S2A). Only CDK1 inhibitor RO-3306 and ROCK inhibitor Y-27632 could inhibit haESC diploidization rate during cell culture (Figure 2B). Consistently, the diploidization rate was accelerated when CDK1-cyclin B1 was overexpressed in haESCs (Figures S2B–S2D). We also determined the effects of different concentrations of RO-3306 and Y-27632 on haploidy maintenance. The results showed that 3 mM and 4.5 mM RO-3306, and 30 mM and 40 mM Y-27632, could significantly suppress diploidization (Figure 2C, p < 0.001; Figure 2D, p < 0.05 and p < 0.001). The haESCs cultured with RO-3306 and Y-27632 showed normal morphology, karyotype, and pluripotency gene expression (Figures S3A–S3C). Furthermore, RO-3306 or Y-27632 treatment has no significant effect on cell growth curve (Figure S4B), proliferation (Figures S4D and S4E), or apoptosis (Figures S4F and S4G) of haESCs within 4 passages. DNA content analysis revealed that RO-3306 and Y-27632 did not increase the proportion of diploid cells in G0 or G1 phase compared with the control, indicating that their effects were not caused by G0/G1 arrest (Figure S4C). We also tested the effects of RO-3306 and Y-27632 on ploidy maintenance on independent haESC lines, and both chemicals could exert consistent and significant inhibition of diploidization in all tested cell lines, suggesting that their effects were not cell line dependent (Figures 2E and 2F). To test the effects of RO-3306

Figure 2. Inhibitors of CDK1 and ROCK Suppress the Diploidization of haESCs (A) Effects of different inhibitors on ploidy maintenance of haESCs (haESC lines AH129-1 and PHGFP-1); more than 2 independent experiments for each molecule. (B) DNA content analysis showed that, after 4 passages (16 days), the percentages of haploidy in haESCs treated with CDK1 and ROCK inhibitor (PHGFP-1 and AH-2; n = 3 repeated experiments) were much higher than that of control haESCs (DMSO treated). (C and D) The effects of different concentrations of RO-3306 (C; PHGFP-1; n = 3 repeated experiments) and Y-27632 (D; AH129-1; n = 3 repeated experiments) on haploidy maintenance of haESCs after a 20-day culture; 3 repeated experiments. Student’s t test. Error bars represent mean ± SD. (E and F) The effects of 4.5 mM RO-3306 (E; n = 6 different haESC lines [AH-1, AH-H2B-GFP-2, AH129-1, AH129-2, PHGFP-1, and PH129-1] with 16 independent experiments) and 30 mM Y-27632 (F; n = 3 different haESC lines [AH129-1, AH129-2, and PHGFP-1] with 6 independent experiments) treatment on haESC lines. *p < 0.05; **p < 0.01; ***p < 0.001, Student’s t test. Error bars represent mean ± SD. See also Figures S2–S4 and Tables S1 and S3.

and Y-27632 on mitotic slippage rates, the division of H2B-GFP haESCs under DMSO, Ro-3306, and Y-27632 treatment was traced by live-cell imaging. The results showed that RO-3306 (n = 460 dividing cells) and Y-27632 (n = 1,707 dividing cells) can significantly reduce the mitotic slippage (Figure S4A). Induction and Characterization of Haploid NSCLCs from Mouse haESCs Next, we examined whether haploid neural stem-cell-like cells (NSCLCs) could be differentiated from mouse haESCs. To induce NSCLCs, mouse haESCs were differentiated into embryoid bodies (EBs) in N2B27 medium with Y-27632 for 3 days, and then the EBs were re-plated on PDL/laminin-coated dishes for about 30 days (Figure 3A). The results showed that 20 mM Y-27632 treatment (adopted for the following differentiation experiments) was enough to generate haploid somatic cells (Figure 3B). DNA content analysis at day 10 after differentiation showed that only 2.3% of total cells were still haploid in the control group, consistent with previous reports that differentiation could result in rapid diploidization (Elling et al., 2011; Leeb and Wutz, 2011) (Figure 3C). However, around 91.5% of the differentiated cells maintained a haploid state by Y-27632 treatment, suggesting that Y-27632 could suppress the differentiation-induced diploid-

ization (Figure 3C). These differentiated cells could proliferate in a haploid state in neural stem cell (NSC) culture medium containing fibroblast growth factor 2 (FGF2) and epidermal growth factor (EGF) (Figure 3D), and exhibited NSC-like morphology (Figure 3E). Moreover, they expressed typical NSC marker genes Nestin, Sox2, N-cadherin, Zfp521, and Pax6, as revealed by qPCR analysis and immunostaining analysis (Figures 3F and 3G). These results collectively suggested that the haESCs could be differentiated into haploid NSCLCs by ROCK inhibitor Y-27632 treatment. Haploid NSCLCs Can Differentiate into Haploid Neurons and Astrocytes To examine whether terminally differentiated haploid cells could be generated, we further differentiated haploid NSCLCs into neurons and astrocytes using Y-27632. Ten days after withdrawal of FGF2 and EGF, approximately 8.56% of differentiated NSCLCs expressed neuron marker Tuj1, and 19% expressed astrocyte marker GFAP. DNA content analysis of Tuj1-positive neurons and GFAP-positive astrocytes showed that a high proportion of cells were still haploid (Figures 4A–4D), indicating the generation of haploid neurons and astrocytes from haESCs. Single-cell genome sequencing analysis of the haploid Tuj1-positive neurons reveals no large genomic deletion, insertion, or consistent copy number variation, indicating the maintenance of genome integrity during haESC differentiation in a haploid state (Figure 4E). After 3-week in vitro maturation, the differentiated cells developed into mature neurons expressing microtubule-associated protein 2 (MAP2) (Figure 4F). Notably, the

Cell Reports 20, 2227–2237, August 29, 2017 2229

Figure 3. Generation and Characterization of Haploid NSCLCs from haESCs with Y-27632 Treatment (A) Strategy overview for differentiation of haploid NSCLCs from haploid ESCs. (B) The effects of different concentrations of Y-27632 on haploidy maintenance of differentiated haESCs (AH129-1; n = 3 independent repeated experiments). Student’s t test. Error bars represent mean ± SD. **p < 0.01; ***p < 0.001. (C) Representative of FACS analysis of the proportion of haploid cells with (right) or without (left) Y-27632 treatment on day 10 of differentiation (AH129-1, AH129-2, and PH129-1; n = 4 independent repeated experiments). (D) Representative chromosome spreads of haploid NSCLCs (AH129-1; n = 3 independent repeated experiments). Scale bar, 100 mm. (E) Morphology of haploid NSCLCs cultured in dishes (AH129-1, n = 3 independent repeated experiments). (F) Gene expression analysis for key NSC markers Nestin, Sox2, N-Cadherin, and Zfp521. Diploid ESC line F1_45 and diploid NSC line C_C12 were used as controls (n = 3 repeated experiments). Error bars represent mean ± SD. (G) Immunostaining of NSC markers Nestin, Sox2, and Pax6 in haploid NSCLCs (AH-NSCLCs1, AH-NSCLCs2, and PH-NSCLCs1; n = 3 repeated experiments). Scale bars, 100 mm. See also Table S1.

MAP2-positive cells remained haploid by DNA content analysis (Figure 4G). These neurons expressed Synapsin, a marker for synaptic formation in vitro (Figure 4H), and could generate Na+/K+ currents and single-evoked action potential by electrophysiological tests using the whole-cell patch-clamp technique (Figures 4I and 4J), indicating the successful generation of haploid terminally differentiated mature neurons. haESCs Can Differentiate into Haploid Cardiomyocytes and Pancreatic Progenitors Next, we explored whether the haESCs could differentiate into haploid somatic cells of mesoderm and endoderm lineages by

2230 Cell Reports 20, 2227–2237, August 29, 2017

treatment with Y-27632. First, we differentiated haESCs into cardiomyocytes using a protocol modified from a previous report (Kattman et al., 2011) (Figure 5A). Haploid cardiac precursor cells were observed by day 6 (Figure S5). Beating cardiomyocytes emerged after 12-day differentiation, which expressed the marker gene cardiac troponin-T (cTnT) (Figures 5B and 5C; Movie S1). DNA content analysis by FACS revealed that nearly all cTnT-positive cells maintained a haploid genome, indicating the generation of haploid cardiomyocytes from haESCs (Figure 5D). We next differentiated haESCs into pancreatic lineage cells using differentiation medium containing Y-27632 (Figure 5E). After 16 days of differentiation, the haESCs could

generate a portion of Pdx1-positive pancreatic progenitors (Figures 5F and 5G). DNA content analysis showed that these Pdx1-positive cells were haploid (Figure 5H). Taken together, these results suggested that haESCs were able to differentiate into haploid somatic cells of mesoderm and endoderm lineages. Haploid NSCLCs as a Tool for High-Throughput Genetic Screening To demonstrate the feasibility of using haESC-derived haploid somatic cells for genetic screening, we transfected the haploid NSCLCs with a piggyBac transposon-based gene-trap vector (Ding et al., 2005) carrying a neor gene to generate insertional mutations. After 6 days of selection, we identified the insertion sites by splinkerette PCR-based sequence analysis. The results showed that the insertion sites evenly spread in all chromosomes, suggesting the genome-wide coverage of mutations (Figures S6A and S6B). We randomly picked 8 NSCLC colonies and amplified the insertional sites with site-specific primers. Only the mutant allele, but not the wild-type allele, could be amplified, verifying the efficient homozygous mutation in haploid NSCLCs (Figure 6A). It has been reported that neural cells are more susceptible to Mn2+-induced toxicity, compared with other cell types (Chen et al., 2015b). We then performed a genetic screen to identify genes regulating the neural toxicity of MnCl2 by using the mutant library of haploid NSCLCs (Figure 6B). After MnCl2 treatment for 4 days, all wild-type haploid NSCLCs died; however, several cells from the mutant library were still alive (Figure 6C). We then collected the live cells and identified the mutated genes in them. The sequencing results showed that the mutated genes in these live cells were enriched in neuralactivity-related pathways such as synaptic transmission, membrane potential, and ion transporter (Figure 6D). Notably, Park2 was among the most frequently mutated genes in two independent screens (Figure 6E), and many insertion sites in the Park2 gene locus were detected in the surviving cells (Figure 6F). To confirm the function of the Park2 mutation in resistance of MnCl2 toxicity, we directly knocked out the Park2 gene in wildtype NSCs by co-transfection with the Cas9-expressing vector and two single-guide RNAs (sgRNAs) targeting Park2. After MnCl2 treatment for 3 days, almost all NSCs died in the control group; however, more than 30% of cells were still alive in the Park2-targeting group (Figures 6G and 6H). Genotyping analysis of the surviving cells from MnCl2 treatment confirmed the Park2 gene mutation, indicating that Park2 was involved in the toxicity effect of MnCl2 on neural cells (Figure S6C). DISCUSSION The haESCs derived in multiple mammalian species have provided valuable resources for genetic screening and modification in mammals. However, as the in vitro counterparts of early embryonic cells, the haESCs are not suitable for genetic analysis of phenotypes of adult cells and tissues, but the rapid diploidization during haESC differentiation hindered the generation of haploid somatic cells. Here, we showed that the diploidization of the mouse haESCs was caused by the mitotic slippage. After screening of a set of inhibitors of genes related to cell-cycle regu-

lation, we found that CDK1 and ROCK inhibitors could efficiently suppress diploidization during the culture and differentiation of haESCs. Moreover, haploid somatic cells representative of all three germ layers, including the terminally differentiated neurons, could be differentiated from haESCs by inhibitor treatment. We also used the haploid NSCLCs for genetic screening of genes involved in Mn2+ neurotoxin resistance and identified multiple candidates including the Park2 gene. That was consistent with the previous reports that Park2 deletion could decrease Mn2+ accumulation in human neuron precursor cells (Aboud et al., 2012) and that Park2 was involved in apoptotic activity (Fujiwara et al., 2008; Gong et al., 2017; Konovalova et al., 2015; Lee et al., 2012; Zhang et al., 2014). Hence, our findings can expand the resources of mammalian haploid cells and demonstrate the feasibility of using mouse haploid somatic cells for efficient genetic screening. Although the recently emerged CRISPR/Cas technology has enable efficient genetic screening in diploid cells, the haploid-cell-based genome-wide genetic screening still has advantages. For instance, the haploid cells can be used to screen the SNP effect induced by chemical mutagenesis (Forment et al., 2017). Hence, genetic screening with haploid somatic cells can serve as a valuable alternative approach in addition to the CRISPR-based genetic screening. It remains elusive how the mammalian haploid cells sense the haploidy and regulate the diploidization. Ploidy duplication can be induced by the successive replication in S phase of the cell cycle, or by direct entering into G1/S phase of the next cell cycle without division, which is known as mitotic slippage (Brito and Rieder, 2006). We found that mitotic slippage was the primary cause of diploidization of mouse haESCs and that CDK1 and ROCK chemical inhibitors could suppress the diploidization. Previous studies have shown that inhibition of CDK1 activity could attenuate G2/M transition (Enomoto et al., 2009; Pagano, 2004), while cyclin B1 degradation-induced decrease of CDK1 activity could promote metaphase-to-anaphase transition (Va´zquez-Novelle et al., 2014). Therefore, inhibition of CDK1 may help to release the arrest of metaphase-to-anaphase transition during haESC diploidization. CDK1 inactivation during mitosis is achieved through degradation of cyclin B by the anaphase-promoting complex/cyclosome (APC/C) (Murray et al., 1989; Peters, 2006). Hence, the suppression effect of CDK1 inhibition on diploidization suggests that the APC/C activity may be insufficient in haESCs, which further leads to diploidization. Since ROCK has an important role in regulating cell contractility by acting on the cytoskeleton (Ramanathan et al., 2015), the effect of ROCK inhibition on diploidization is probably due to the direct alteration of the mechanical force required for cell division. Studies on the role of CDK1 and ROCK in diploidization may shed light on ploidy maintenance mechanisms in the future. EXPERIMENTAL PROCEDURES Live-Cell Fluorescent Imaging The androgenetic haploid cell lines carrying H2B-GFP were established from haploid androgenetic embryos, as described previously (Li et al., 2012). The H2B-GFP transgenic mice were obtained from Beijing Vital River Laboratory Animal Technology. The fluorogenic, cell-permeable reagent SiR-Tubulin (Cytoskeleton, CY-SC002) was used to image tubulin in living cells.

Cell Reports 20, 2227–2237, August 29, 2017 2231

Figure 4. Characterization of Terminally Differentiated Haploid Neurons (A) Immunostaining of neuron marker Tuj1 of the haESC-differentiated cells. Scale bar, 100 mm. AH-NSCLCs1 was used for differentiation experiments; n = 3 repeated experiments. (B) FACS analysis of the proportion of haploid cells in Tuj1-positive neurons. Diploid ESCs stained with the same antibody were used as a staining control (shown as gray area). AH-NSCLCs1 and PH-NSCLCs1 were used for differentiation experiments; n = 3 repeated experiments. (C) Immunostaining of astrocyte marker GFAP of the haESC-differentiated cells. Scale bar, 100 mm.

(legend continued on next page)

2232 Cell Reports 20, 2227–2237, August 29, 2017

Haploid ESC Culture The mouse parthenogenetic and androgenetic haploid ESC lines were cultured on a feeder layer in N2B27 medium plus 2i + mLIF, as described previously (Ying et al., 2008). The N2B27 medium consists of DMEM/F12 (GIBCO, 11320-033) and Neurobasal (GIBCO, 21103-049) at a ratio of 1:1 and was supplemented with 1% N-2 supplement (GIBCO, 17502048), 0.5% B-27 supplement (GIBCO, 17504044), 20 ng/mL BSA (Sigma, A3803), 10 mg/mL insulin (Roche Applied Science, 11376497001), GlutaMAX (1003; GIBCO, 10565-018), 5% knockout serum replacement (KOSR; GIBCO, 12618013), b-mercaptoethanol (10003, GIBCO, 21985023), and penicillinstreptomycin (1003; GIBCO, 15140163). 2i + mLIF consisted of the following: 1 mM PD0325901 (MEK inhibitor; Stemgent, 04-0006), 3 mM CHIR99021 (GSK3b inhibitor; Stemgent, 04-0004), and 1,000 U/mL mLIF (Millipore, ESG1107). The medium was changed daily, and the haESCs were disaggregated and passaged by 0.25% Trypsin-EDTA (GIBCO, 25200-072) every 3 to 5 days. Cells were maintained in an incubator at 37 C and 5% CO2. Purification of Haploid Cells and DNA Content Analysis Haploid cells were purified and analyzed by FACS using a protocol modified from a previous report (Leeb and Wutz, 2011). Briefly, single-cell suspensions were obtained by trypsin-EDTA digestion and repetitive pipetting and sieved through a 40-mm cell strainer. Haploid cells (1C) were incubated with 10 mg/mL Hoechst 33342 (Invitrogen, H3570) for 15–20 min at 37 C before analysis. Data were collected with a BD FACS Aria II cell sorter (BD Biosciences) and MoFlo XDP cell sorter (Beckman-Coulter). Diploid ESCs (2C) were used as a control. The percentage of haploid cells was calculated based on the percentages of 1C peak (including haploid cells at G1 phase), 2C peak (including haploid cells at G2/M phase and diploid cells at G1 phase), and 4C peak (including diploid cells at G2/M phase). The formula is shown below, according to the previous literature (Li et al., 2012; Xu et al., 2017). Haploid G1 phaseð%1CÞ Diploid G1 phaseð%2C  %xÞ = : Haploid G2=M phaseð%xÞ Diploid G2=M phaseð%4CÞ The percentage of total haploid = %1C + %x: EdU and Phospho-Histone H3 (Ser10) Assay ESCs were treated with DMSO, Ro-3306 (4.5 mM; Biovision, 2039-1), and Y-27632 (30 mM; Selleck, S1049) for 4 passages, respectively. The Click-iT EdU Alexa Fluor 488 Imaging Kit (Thermo Fisher Scientific, C10337) was used to detect cell DNA replication. Briefly, ESCs were labeled with EdU for 35 min followed by cell fixation, permeabilization, EdU detection, and DNA staining were performed according to the manual posted on the website (https://www.thermofisher.com/order/catalog/product/C10337). An anti-phospho-histone H3 (Ser10) antibody (Cell Signaling Technology, 9708) was used to assay mitotic phase by immunofluorescence microscopy. Apoptosis/Necrosis Measurement The Annexin V-FITC Apoptosis Detection Kit (Abcam, ab14085) was used to detect cell apoptosis. Briefly, (1) haESCs were treated with DMSO, Ro-3306 (4.5 mM), and Y-27632(30 mM) for 4 passages, respectively; (2) cells were gently trypsinized and washed once with serum-containing media before incubation with annexin V-FITC (fluorescein isothiocyanate); (3) 2 3 105 cells were re-suspended in 500 mL of 13 Binding Buffer added with 5 mL annexin V-FITC and incubated at room temperature for 5 min in the dark; (4) 5 mL propidium

iodide (PI; 50 mg/mL) was added before incubation for another 5 min. Cells were then analyzed by FACS. As described by the manufacturer’s instruction, annexin V/PI cells were viable; annexin V+/PI cells were in early apoptosis, and annexin V+/PI+ cells were in late apoptosis or necrotic. Immunofluorescence Staining and Somatic Cell DNA Content Analysis For immunostaining, the samples were fixed with 4% paraformaldehyde (PFA) for 20 min at room temperature, washed 3 times with PBS, and blocked with 2% BSA (diluted with 0.3% Triton X-100 in PBS) for 1 hr at room temperature. The samples were then incubated with the primary antibodies overnight at 4 C. On the second day, samples were washed for 3 times with PBS followed by incubating with appropriate Alexa-Fluor-conjugated secondary antibodies (diluted in 2% BSA) at 37 C for 1 hr. After incubation, the samples were washed for 3 times with PBS and counterstained with DAPI (Sigma, D9542) for 5 min at room temperature. Images were taken using a confocal microscope (ZEISS, LSM 780 META). The primary antibodies used were as follows: anti-Nestin (Millipore, MAB353), anti-Sox2 (Millipore, AB5603), anti-Pax6 (Abcam, ab5790), anti-Tuj1 (Millipore, MAB1637; Covance, MRB-435P), anti-cTnT (Abcam, ab8295), anti-Pdx1 (Abcam, ab47383), anti-Map2 (Abcam, ab32454), and Anti-Synapsin I (Mllipore, AB1543P). In a DNA content analysis for haploid differentiated cells, the differentiated cells were gently pipetted to single-cell suspension with a pipette (the appropriate trypsin digestion may be alternative) after immunofluorescence staining with indicated antibody and DAPI (DNA, Sigma), and flow analysis was performed after the cells were filtered with a 400-mesh cell sieve. Differentiation of Mouse haESCs into NSCLCs The method used to differentiate the mouse haESCs into NSCLCs was schematically summarized in Figure 3A. Briefly, the mouse haESCs were cultured in suspension for EB formation for 3 days. After 3 days, the EBs were plated onto dishes coated with 0.2% gelatin (Sigma, G1890), and the medium was changed to NSCLC differentiation medium. The NSCLC differentiation medium consists of N2B27 medium supplemented with 20 ng/mL FGF2 (R&D Systems, 233-FB-001MG/CF), 20 ng/mL EGF (R&D, 2028-EG-200), 300 nM LDN193189 (MedChemExpress [MCE], HY-12071A), 10 mM SB431542 (Stemgent, 04-0010-10), and a ROCK inhibitor (20 mM Y-27632 or 4 mM Thiazovivin [MCE, HY-13257] or 10 mM Y-27632 plus 2 mM Thiazovivin). Five days later, when the cells became confluent, cells were dissociated into single cells, plated on dishes coated with poly-D-lysine (10 mg/mL, Sigma, P6407) and laminin (10 mg/mL; Invitrogen, 23017-015), and cultured in NSCLC medium for another 4 days. At day 10 after initial induction, the differentiated cells were dissociated for DNA content analysis. Differentiation for Haploid Neurons and Haploid Astrocytes For haploid neuron differentiation, the haploid NSCLCs were attached on dishes coated with poly-D-lysine (10 mg/mL) and laminin (10 mg/mL) and were cultured with N2B27 medium supplemented with 20 ng/mL NT-3 (Peprotech, 450-03; 10 mg), 20 ng/mL brain-derived neurotrophic factor (BDNF; Peprotech, 450-02; 10 mg), and 20 mM ROCK inhibitor Y-27632 (MCE) for more than 10 days for neuronal maturation. For haploid astrocyte differentiation, the haploid NSCLCs were attached on dishes coated with poly-D-lysine (10 mg/mL) and laminin (10 mg/mL) and were cultured with N2B27 medium supplemented with 2% FBS (GIBCO, 10099141) and 20 mM ROCK inhibitor Y-27632 (MCE) for 14 days. FACS was used to analyze the ploidy of the Tuj1positive neurons or GFAP (anti-GFAP, Thermo, A21282)-positive astrocytes.

(D) FACS analysis of the proportion of haploid cells in GFAP-positive astrocytes. Diploid ESCs stained with the same antibody were used as a staining control (shown as gray area). AH-NSCLCs1 was used for differentiation experiments; n = 2 repeated experiments. (E) Analysis of genome integrity of haploid neurons by single-cell genome sequencing. (F–H) In (F) and (H), immunostaining is shown of mature neuron markers MAP2 (F) and Synapsin (H). (G) FACS analysis of the proportion of haploid cells of MAP2positive neurons. Diploid ESCs stained with the same antibody were used as a staining control (shown as gray area). AH-NSCLCs1 was used for differentiation experiments; n = 3 repeated experiments. Scale bars, 100 mm. (I and J) Detection of action potentials (I) and Na+ and K+ currents (J) in haploid neurons derived from haESCs. AH-NSCLCs1 was used for differentiation experiments; n = 3 repeated experiments. See also Table S1.

Cell Reports 20, 2227–2237, August 29, 2017 2233

Figure 5. Generation and Characterization of Haploid Cardiomyocytes and Pancreatic Progenitor Cells from haESCs with Y-27632 Treatment (A–H) In (A) and (E), strategy overviews are shown for differentiation of cardiomyocytes (A) and pancreatic cells (E) from haploid ESCs. (B and F) Immunostaining of cardiomyocyte marker cTnT (B) and pancreatic marker Pdx1 (F) of the haESC-differentiated cells. Scale bars, 100 mm. (C and D) FACS analysis of the proportion (C) and the haploidy ratio (D) of cTnT-positive cells. Diploid ESCs stained with the same antibody were used as a staining control (shown as gray area). (G and H) FACS analysis of the proportion (G) and the haploidy ratio (H) of Pdx1-positive cells. Diploid ESCs stained with the same antibody were used as a staining control (shown as gray area). AH129-1 was used for all differentiation experiments; n = 2 repeated experiments. See also Figure S5, Table S1, and Movie S1. Differentiation for Haploid Cardiomyocytes The method used to differentiate the mouse haESCs into cardiomyocytes was schematically summarized in Figure 5A. The mouse haploid ESCs were induced to form embryoid bodies (EBs) by suspended culture in N2B27 medium supplemented with 10% KOSR (GIBCO) for 2 days. EBs were plated on dishes coated with fibronectin (20 mg/mL; Millipore, fc010) and cultured in cardiomyocyte differentiation medium for another 5 days. The cardiomyocyte differentiation medium consists of DMEM/F12 (GIBCO) with 10% KOSR (GIBCO), 10 ng/mL Bmp4 (Peprotech, 120-05ET), 10 ng/mL activin A (R&D, 338-AC-050/CF), 6 mM CHIR99021 (Stemgent),

2234 Cell Reports 20, 2227–2237, August 29, 2017

and 20 mM Y-27632 (MCE). At day 5 after initial induction, CHIR99021 was replaced by 2 mM IWR1 (Sigma, I0161). At day 7, the differentiation medium was changed into DMEM/F12 (GIBCO) with 10% KOSR (GIBCO) and 20 mM Y-27632 (MCE) for another 14-day culture. At day 6 of differentiation, the DNA content of Sca1 (anti-Sca1; Abcam, ab25031)-positive, CD90 (anti-CD90; Abcam, ab226)-positive, and c-kit (anti-c-kit; Abcam, ab25495)-positive cardiac precursor cells were each analyzed by FACS, and cTnT (anti-cTnT, Abcam)-positive cardiomyocytes were identified by immunofluorescence staining and analyzed for DNA content by FACS after 12 days of differentiation.

Figure 6. Genetic Screening with Haploid NSCLCs (A) PCR verification of the piggyBac insertion sites in 8 different colonies. Red arrows represent location of PCR primers. The left panel shows the detected wild-type gene loci, and the right panel shows the detected insertion sites. Genes on the left are the identified gene loci in each colony. M, marker; WT, wild-type cells without piggyBac insertion; H2O, PCR without DNA template. (B) Schematic diagram of genetic screening in haploid NSCLCs to identify genes involved in MnCl2 neurotoxicity. (C) Cells after MnCl2 treatment for 4 days. The white arrow indicates cells that survived after treatment. (D) Gene enrichment analysis of mutant genes in the cells that survived after MnCl2 treatment. (E) Mutation frequency in two independent screening experiments. Genes are ranked by the mutation reads. The red point represents Park2 in each experiment. Note that Park2 is among the most frequently mutated genes in the two experiments. (F) Insertion sites and mutation frequency in the Park2 locus in two independent screening experiments (AH-NSCLCs-1, n = 2 repeated experiments). (G) WT NSCs transfected with Park2 sgRNAs after MnCl2 treatment for 3 days. WT, wild-type control. (NSC_12, n = 3 repeated experiments). (H) Survival rate of cells transfected with Park2 sgRNAs after MnCl2 treatment for 3 days, compared with the wild-type control. ***p < 0.001. (NSC_12, n = 3 repeated experiments). Scale bars, 100 mm. Error bars represent mean ± SD. See also Figure S6 and Table S1.

the medium was changed into DMEM/F12 (GIBCO) with 10% KOSR (GIBCO) and 20 mM Y-27632 (MCE). Pdx1 (anti-Pdx1, Abcam)-positive endodermal somatic cells were identified by immunofluorescence staining and analyzed for DNA content by FACS after 16 days of differentiation.

Differentiation for Haploid Endodermal Somatic Cells The method used to differentiate the mouse haESCs into endodermal somatic cells was schematically summarized in Figure 5E. Briefly, the mouse haploid ESCs were induced to form EBs by suspension in N2B27 medium with 10% KOSR for 2 days. Then, the EBs were plated on dishes coated with fibronectin (20 mg/mL) and cultured in medium consisting of DMEM/ F12 (GIBCO) with 10% KOSR (GIBCO), 100 ng/mL activin A (R&D), 3 mM CHIR99021 (Stemgent), and 20 mM Y-27632 (MCE) for 5 days. At day 7,

Karyotype Analysis of Haploid NSCLCs NSCLCs were incubated with 0.2 mg/mL nocodazole (Sigma, M1404) for 3 hr. After trypsinization, the NSCLCs were suspended in 0.075 M KCl at 37 C for 30 min. Then, the cells were fixed with solution consisting of methanol and acetic acid (3:1 in volume) for 30 min and then were dropped onto pre-cleaned slides. The cells were stained with Giemsa stain (Sigma, GS500ML) or Hoechst 33342 for 15 min after being incubated in 5 M HCl. More than 20 metaphase spreads were analyzed. Generation of a Mutant Library of Haploid NSCLCs and Inverse PCR To generate the mutant library, 107 of haploid NSCLCs were transfected using Lipofectamine LTX and PLUS reagents (Invitrogen, 15338100) with 20 mg piggyBac gene-trapping plasmid and 10 mg codon-optimized

Cell Reports 20, 2227–2237, August 29, 2017 2235

piggyBac transposase. G418 was used to select gene-trap insertions. An adapted inverse PCR protocol was based on a previous report (Chen et al., 2015a). Splinkerette PCR-Based Sequence Analysis To identify the trapping sites of a piggyBac transposon in a pool of cells, the splinkerette PCR protocol was performed as previously described (Leeb et al., 2014; Li et al., 2014; Potter and Luo, 2010). In brief, the integration sites in genome were amplified using nested PCR after ligation with splinkerette adapters, and the PCR products were sequenced on the HiSeq 2500. Only the reads containing the piggyBac sequences were used for the alignment analysis, and after trimming the piggyBac and adaptor sequences by the perl scripts, the reads were aligned to the mouse reference genome (mm9) by the BWA (Li and Durbin, 2009) with default parameters. The alignments with unique genomic location used for the next analysis. As previously described (Li et al., 2014), the alignments were used to calculate distribution mode in different genomic regions, such as the UTR, CDS, and intergenic regions, and the alignments genome-wide were shown by the ggbio package (Yin et al., 2012). Gene Ontology (GO) enrichment analysis were performed by GOEAST (Zheng and Wang, 2008). The alignments distributed in Park2 gene region were shown by an IGV browser (Robinson et al., 2011). Manganese (II) Neurotoxicity Screen A mutated library of haploid NSCLCs was exposed to a lethal dose (1 mM) of manganese (II) chloride for 4 days, and the wild-type haploid NSCLCs were used as control. DNA of the survived cells was enriched and then subjected to inverse PCR for insertion site analysis by deep sequencing. Statistical Analysis All statistical data were analyzed and are presented as means ± SD. The unpaired two-tailed Student’s t test was used to calculate statistical significance with p values. The p values < 0.05 were considered as significantly different. ***p < 0.001, **p < 0.01, *p < 0.05, and p > 0.05 were considered non-significant. ACCESSION NUMBERS The accession number for the Mn2+ screening data reported in this paper is GEO: GSE101795. SUPPLEMENTAL INFORMATION Supplemental Information includes six figures, three tables, and one movie and can be found with this article online at http://dx.doi.org/10.1016/j. celrep.2017.07.081. AUTHOR CONTRIBUTIONS Q.Z. conceived the project. Q.Z. and W.L. supervised the experiments and data analysis. Q.Z., W.L., Z.-Q.H., B.-L.X., and Y.-K.W. wrote the manuscript with help from all of the authors. Z.-Q.H., B.-L.X., Y.-K.W., J.L., G.-H.F., L.-L.Z., Y.-H.L., H.-F.W., T.-D.L., K.X., X.-W.Y., Y.-F.L., X.-X.Z., Y.-Z., and L.W. participated in the experiments and data analysis. ACKNOWLEDGMENTS This work was supported by grants from the National Basic Research Program of China (2014CB964800), the National Natural Science of China (31422038, 31471395, and 31601173), and The National Key Technology R&D Program (2014BAI02B01). We thank Qing Meng, Ting Li, Hua Qin, Li-Juan Wang, and Shi-Wen Li for their help with FACS and confocal laser-scanning microscopy. Received: March 30, 2017 Revised: July 5, 2017 Accepted: July 28, 2017 Published: August 29, 2017

2236 Cell Reports 20, 2227–2237, August 29, 2017

REFERENCES Aboud, A.A., Tidball, A.M., Kumar, K.K., Neely, M.D., Ess, K.C., Erikson, K.M., and Bowman, A.B. (2012). Genetic risk for Parkinson’s disease correlates with alterations in neuronal manganese sensitivity between two human subjects. Neurotoxicology 33, 1443–1449. Brito, D.A., and Rieder, C.L. (2006). Mitotic checkpoint slippage in humans occurs via cyclin B destruction in the presence of an active checkpoint. Curr. Biol. 16, 1194–1200. €rckstu €mmer, T., Banning, C., Hainzl, P., Schobesberger, R., Kerzendorfer, Bu C., Pauler, F.M., Chen, D., Them, N., Schischlik, F., Rebsamen, M., et al. (2013). A reversible gene trap collection empowers haploid genetics in human cells. Nat. Methods 10, 965–971. Chen, P., Chakraborty, S., Mukhopadhyay, S., Lee, E., Paoliello, M.M., Bowman, A.B., and Aschner, M. (2015a). Manganese homeostasis in the nervous system. J. Neurochem. 134, 601–610. Chen, P., Chakraborty, S., Peres, T.V., Bowman, A.B., and Aschner, M. (2015b). Manganese-induced neurotoxicity: from C. elegans to humans. Toxicol. Res. (Camb.) 4, 191–202. Ding, S., Wu, X., Li, G., Han, M., Zhuang, Y., and Xu, T. (2005). Efficient transposition of the piggyBac (PB) transposon in mammalian cells and mice. Cell 122, 473–483. Elling, U., Taubenschmid, J., Wirnsberger, G., O’Malley, R., Demers, S.P., Vanhaelen, Q., Shukalyuk, A.I., Schmauss, G., Schramek, D., Schnuetgen, F., et al. (2011). Forward and reverse genetics through derivation of haploid mouse embryonic stem cells. Cell Stem Cell 9, 563–574. Enomoto, M., Goto, H., Tomono, Y., Kasahara, K., Tsujimura, K., Kiyono, T., and Inagaki, M. (2009). Novel positive feedback loop between Cdk1 and Chk1 in the nucleus during G2/M transition. J. Biol. Chem. 284, 34223–34230. Forment, J.V., Herzog, M., Coates, J., Konopka, T., Gapp, B.V., Nijman, S.M., Adams, D.J., Keane, T.M., and Jackson, S.P. (2017). Genome-wide genetic screening with chemically mutagenized haploid embryonic stem cells. Nat. Chem. Biol. 13, 12–14. Fujiwara, M., Marusawa, H., Wang, H.Q., Iwai, A., Ikeuchi, K., Imai, Y., Kataoka, A., Nukina, N., Takahashi, R., and Chiba, T. (2008). Parkin as a tumor suppressor gene for hepatocellular carcinoma. Oncogene 27, 6002–6011. Gong, Y., Schumacher, S.E., Wu, W.H., Tang, F., Beroukhim, R., and Chan, T.A. (2017). Pan-cancer analysis links PARK2 to BCL-XL-dependent control of apoptosis. Neoplasia 19, 75–83. Kattman, S.J., Witty, A.D., Gagliardi, M., Dubois, N.C., Niapour, M., Hotta, A., Ellis, J., and Keller, G. (2011). Stage-specific optimization of activin/nodal and BMP signaling promotes cardiac differentiation of mouse and human pluripotent stem cell lines. Cell Stem Cell 8, 228–240. Konovalova, E.V., Lopacheva, O.M., Grivennikov, I.A., Lebedeva, O.S., Dashinimaev, E.B., Khaspekov, L.G., Fedotova, E.Y., and Illarioshkin, S.N. (2015). Mutations in the Parkinson’s disease-associated PARK2 gene are accompanied by imbalance in programmed cell death systems. Acta Naturae. 7, 146–149. Lanni, J.S., and Jacks, T. (1998). Characterization of the p53-dependent postmitotic checkpoint following spindle disruption. Mol. Cell. Biol. 18, 1055–1064. Lee, K., Lee, M.H., Kang, Y.W., Rhee, K.J., Kim, T.U., and Kim, Y.S. (2012). Parkin induces apoptotic cell death in TNF-a-treated cervical cancer cells. BMB Rep. 45, 526–531. Leeb, M., and Wutz, A. (2011). Derivation of haploid embryonic stem cells from mouse embryos. Nature 479, 131–134. Leeb, M., Walker, R., Mansfield, B., Nichols, J., Smith, A., and Wutz, A. (2012). Germline potential of parthenogenetic haploid mouse embryonic stem cells. Development 139, 3301–3305. Leeb, M., Dietmann, S., Paramor, M., Niwa, H., and Smith, A. (2014). Genetic exploration of the exit from self-renewal using haploid embryonic stem cells. Cell Stem Cell 14, 385–393. Li, H., and Durbin, R. (2009). Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25, 1754–1760.

Li, W., Shuai, L., Wan, H., Dong, M., Wang, M., Sang, L., Feng, C., Luo, G.-Z., Li, T., Li, X., et al. (2012). Androgenetic haploid embryonic stem cells produce live transgenic mice. Nature 490, 407–411.

Sagi, I., Chia, G., Golan-Lev, T., Peretz, M., Weissbein, U., Sui, L., Sauer, M.V., Yanuka, O., Egli, D., and Benvenisty, N. (2016). Derivation and differentiation of haploid human embryonic stem cells. Nature 532, 107–111.

Li, W., Li, X., Li, T., Jiang, M.G., Wan, H., Luo, G.Z., Feng, C., Cui, X., Teng, F., Yuan, Y., et al. (2014). Genetic modification and screening in rat using haploid embryonic stem cells. Cell Stem Cell 14, 404–414.

Va´zquez-Novelle, M.D., Sansregret, L., Dick, A.E., Smith, C.A., McAinsh, A.D., Gerlich, D.W., and Petronczki, M. (2014). Cdk1 inactivation terminates mitotic checkpoint surveillance and stabilizes kinetochore attachments in anaphase. Curr. Biol. 24, 638–645.

Minn, A.J., Boise, L.H., and Thompson, C.B. (1996). Expression of Bcl-xL and loss of p53 can cooperate to overcome a cell cycle checkpoint induced by mitotic spindle damage. Genes Dev. 10, 2621–2631. Monfort, A., Di Minin, G., Postlmayr, A., Freimann, R., Arieti, F., Thore, S., and Wutz, A. (2015). Identification of Spen as a crucial factor for Xist function through forward genetic screening in haploid embryonic stem cells. Cell Rep. 12, 554–561. Murray, A.W., Solomon, M.J., and Kirschner, M.W. (1989). The role of cyclin synthesis and degradation in the control of maturation promoting factor activity. Nature 339, 280–286. Pagano, M. (2004). Control of DNA synthesis and mitosis by the Skp2-p27Cdk1/2 axis. Mol. Cell 14, 414–416. Peters, J.M. (2006). The anaphase promoting complex/cyclosome: a machine designed to destroy. Nat. Rev. Mol. Cell Biol. 7, 644–656. Potter, C.J., and Luo, L. (2010). Splinkerette PCR for mapping transposable elements in Drosophila. PLoS ONE 5, e10168. Ramanathan, S.P., Helenius, J., Stewart, M.P., Cattin, C.J., Hyman, A.A., and Muller, D.J. (2015). Cdk1-dependent mitotic enrichment of cortical myosin II promotes cell rounding against confinement. Nat. Cell Biol. 17, 148–159. Robinson, J.T., Thorvaldsdo´ttir, H., Winckler, W., Guttman, M., Lander, E.S., Getz, G., and Mesirov, J.P. (2011). Integrative genomics viewer. Nat. Biotechnol. 29, 24–26.

Xu, H., Yue, C., Zhang, T., Li, Y., Guo, A., Liao, J., Pei, G., Li, J., and Jing, N. (2017). Derivation of haploid neurons from mouse androgenetic haploid embryonic stem cells. Neurosci. Bull. 33, 361–364. Yang, H., Liu, Z., Ma, Y., Zhong, C., Yin, Q., Zhou, C., Shi, L., Cai, Y., Zhao, H., Wang, H., et al. (2013). Generation of haploid embryonic stem cells from Macaca fascicularis monkey parthenotes. Cell Res. 23, 1187–1200. Yi, M., Hong, N., and Hong, Y. (2009). Generation of medaka fish haploid embryonic stem cells. Science 326, 430–433. Yin, T., Cook, D., and Lawrence, M. (2012). ggbio: an R package for extending the grammar of graphics for genomic data. Genome Biol. 13, R77. Ying, Q.L., Wray, J., Nichols, J., Batlle-Morera, L., Doble, B., Woodgett, J., Cohen, P., and Smith, A. (2008). The ground state of embryonic stem cell selfrenewal. Nature 453, 519–523. Zhang, C., Lee, S., Peng, Y., Bunker, E., Giaime, E., Shen, J., Zhou, Z., and Liu, X. (2014). PINK1 triggers autocatalytic activation of Parkin to specify cell fate decisions. Curr. Biol. 24, 1854–1865. Zheng, Q., and Wang, X.J. (2008). GOEAST: a web-based software toolkit for Gene Ontology enrichment analysis. Nucleic Acids Res. 36, W358–W363. Zhong, C., Zhang, M., Yin, Q., Zhao, H., Wang, Y., Huang, S., Tao, W., Wu, K., Chen, Z.J., and Li, J. (2016). Generation of human haploid embryonic stem cells from parthenogenetic embryos obtained by microsurgical removal of male pronucleus. Cell Res. 26, 743–746.

Cell Reports 20, 2227–2237, August 29, 2017 2237