SCIENCE CHINA Life Sciences

• RESEARCH PAPER • doi: 10.1007/s11427-017-9155-4

Function of FEZF1 during early neural differentiation of human embryonic stem cells

Xin Liu1,2, Pei Su1,2, Lisha Lu1,2, Zicen Feng1,2, Hongtao Wang1,2* & Jiaxi Zhou1,2*

1State Key Laboratory of Experimental Hematology, Institute of Hematology & Blood Diseases Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin 300020, China; 2Center for Stem Cell Medicine, Chinese Academy of Medical Sciences & Department of Stem Cells and Regenerative Medicine, Peking Union Medical College, Tianjin 300000, China

Received September 13, 2017; accepted November 19, 2017; published online January 2, 2018

The understanding of the mechanism underlying human neural development has been hampered due to lack of a cellular system and complicated ethical issues. Human embryonic stem cells (hESCs) provide an invaluable model for dissecting human development because of unlimited self-renewal and the capacity to differentiate into nearly all cell types in the human body. In this study, using a chemical defined neural induction protocol and molecular profiling, we identified Fez family zinc finger 1(FEZF1)asa potential regulator of early human neural development. FEZF1 is rapidly up-regulated during neural differentiation in hESCs and expressed before PAX6, a well-established marker of early human neural induction. We generated FEZF1-knockout H1 hESC lines using CRISPR-CAS9 technology and found that depletion of FEZF1 abrogates neural differentiation of hESCs. Moreover, loss of FEZF1 impairs the pluripotency exit of hESCs during neural specification, which partially explains the neural induction defect caused by FEZF1 deletion. However, enforced expression of FEZF1 itself fails to drive neural differentiation in hESCs, suggesting that FEZF1 is necessary but not sufficient for neural differentiation from hESCs. Taken together, our findings identify one of the earliest regulators expressed upon neural induction and provide insight into early neural development in human.

FEZF1, hESCs, CRISPR/Cas9, neural differentiation

Citation: Liu, X., Su, P., Lu, L., Feng, Z., Wang, H., and Zhou, J. (2018). Function of FEZF1 during early neural differentiation of human embryonic stem cells . Sci China Life Sci 61, 1–11. doi: 10.1007/s11427-017-9155-4

INTRODUCTION 2009). They therefore offer an invaluable research tool and unprecedented opportunities for studying human neural lin- Lack of a cellular system and complicated ethical issues have eage specification. To date, many procedures deriving neu- hampered the understanding of the molecular basis of human ral progenitor cell and functional neurons from hPSCs under neural development. Human pluripotent stem cells (hPSCs), chemically defined culture conditions in vitro have been re- including human embryonic stem cells (hESCs) and human ported (Lu et al., 2016; Yao et al., 2006; Zhang et al., 2016; induced pluripotent stem cells (hiPSCs), can be capable of Zhou et al., 2010). Using these neural induction methods, the unlimited proliferation and differentiate into all three embry- molecular mechanisms controlling early human neural differ- onic germ layers in vitro (Thomson et al., 1998; Yamanaka, entiation have begun to be unearthed. The neural commitment from hPSCs is controlled by the *Corresponding authors (Hongtao Wang, email: [email protected]; Jiaxi interplay between extrinsic signaling factors and intrinsic Zhou, email: [email protected]) transcription factors. The extrinsic growth factors regulating

© Science China Press and Springer-Verlag GmbH Germany 2018 life.scichina.com link.springer.com 2 Liu, X., et al. Sci China Life Sci neural induction from hPSCs have been well studied. Ac- hESCs were cultured to confluency in mTeSR medium and tivin/Nodal signaling is critical for pluripotency maintenance subsequently differentiated the cells in DMEM/F12 medium and endoderm induction (D’Amour et al., 2005; Vallier et containing KSR and compound C, which induces high-effi- al., 2005), while BMP signaling drives mesendoderm or ciency neural conversion by blocking both the Activin and the trophoectoderm differentiation (Chadwick et al., 2003). The BMP signaling pathway in hESCs. To identify novel essen- inhibition of either Activin/Nodal signaling, BMP signaling, tial regulators during this neural induction process, we used or both has been reported to promote neural differentiation this method to induce neural differentiation from hESCs and (Chambers et al., 2009; Pera et al., 2004; Smith et al., 2008). performed molecular profiling to progenitor cells through- FGF signaling inhibits neural conversion through suppress- out the differentiation time course during the early stages of ing the exit from hPSCs self-renewal (Vallier et al., 2005). neural induction. After 3 days of differentiation, 24 Retinoic acid signaling, involved in neural development in were found up-regulated by more than 1.5 folds (Figure 1A Xenopus and mice, induces neural induction and patterning and Table 1). Several previously reported genes critical for in hPSCs (Gamse and Sive, 2000; Niederreither et al., 2000). neural development such as ARHGAP15, FLRT3, ROR2 and WNT signaling plays an important role in specifying regional MEIS2 were included, thereby validating our screening strat- identity of hPSC-derived neural progenitor cells (Moya et al., egy (Agoston et al., 2014; Paganoni et al., 2010; Robinson 2014). In contrast to the understanding of extrinsic signaling et al., 2004; Zamboni et al., 2016). In these genes, FEZF1 factors that control early human neural differentiation, the draw our attention due to its previously documented roles in intrinsic transcription factors driving neural induction from neural development in animals (Eckler et al., 2011; Hirata et hPSCs are still poorly defined. It has been suggested that al., 2006; Shimizu et al., 2010; Watanabe et al., 2009). Using PAX6, OTX2, and SOX2 play an important role in neural real-time PCR and Western blot, we confirmed the up-regu- specificaiton from hPSCs and could be used as early neural lation of FEZF1 during hESC early neural differentiation, ac- differentiation markers (Greber et al., 2011; Zhang and Cui, companied by the increase of neural progenitor specific mark- 2014; Zhang et al., 2010). ers (PAX6, OTX2 and SOX2) and the decrease of pluripo- FEZF1 (Fez family zinc finger 1) belongs to the forebrain tency-associated markers (OCT-4 and NANOG) (Figure 1B embryonic zinc finger (Fez) family and first identified inthe and C). Importantly, the expression of FEZF1 was preceded anterior neuroepitheulim of Xenopus and zebrafish embryos the induction of PAX6, suggesting that FEZF1 may function (Matsuo-Takasaki et al., 2000). During mouse embryogene- as a crucial regulator to hESC neural specification. Given that sis, Fezf1 is first detectable in the head fold at E7.5, expresses FEZF2 play vital roles in mouse neural development (Chen et in the forebrain at E8.5 and exhibits expression in olfactory al., 2008; Guo et al., 2013; McKenna et al., 2011), we also ex- system at E10.5. At E15.5, Fezf1 is broadly expressed in pro- amined the expression pattern of FEZF2 during hESC neural genitor cells and neurons of main olfactory epithelium and induction. Compared with the significant increase of FEZF1 shows expression in the developing amygdala and hypothal- expression, FEZF2 only showed a moderate increase after amus. Genetic loss-of-function studies indicate that Fezf1 is 4 days of differentiation, indicating that FEZF1 may play a implicated in forebrain neurogenesis, olfactory system devel- more important role than FEZF2 in hESC neural commitment opment and patterning of the diencephalon, suggesting the es- (Figure 1B and C). Together, these results pointed to FEZF1 sential role of Fezf1 in mouse neural development (Eckler et as a potential regulator of early human neural differentiation. al., 2011; Hirata et al., 2006; Shimizu et al., 2010; Watanabe et al., 2009). However, the role of FEZF1 in human early FEZF1 silencing impairs neural differentiation of hESCs neurogenesis is still undefined. To assess the role of FEZF1 during neural induction, we In this study, using a chemical defined human neural in- stably silenced endogenous FEZF1 expression in H1 and duction model and molecular profiling we identified FEZF1 H9 cells with shFEZF1-expressing lentiviral vectors. The as a potential regulator of human early neurogenesis. FEZF1 expression level of FEZF1 was significantly decreased knock-down or knock-out impaired hESC early neural differ- in shFEZF1-expressing H1 and H9 cells, as revealed by entiation. Our results indicate FEZF1 is a key transcription real-time PCR and Western blot (Figure 2A and B). After 5 factor in hESC neural commitment, providing insight into the days of neural differentiation, shFEZF1-expressing H1 and endogenous regulation of human neural development. H9 cells exhibited a significant reduction in neural mark- ers (PAX6, OTX2 and SOX2) compared with the control, RESULTS suggesting FEZF1 knockdown impaired neural induction of hESCs (Figure 2C and D). Because the exit of pluripotency FEZF1 identified as an essential regulator of hESC early is required for neural lineage commitment of hESCs, we neural differentiation subsequently detected the effect of FEZF1 silencing on We recently developed a single-step method for neural in- pluripotency exit during neural differentiation of hESCs. As duction of hESCs in adherent cultures (Zhou et al., 2010). shown in Figure 2E and F, after 5 days of neural differentia- Table 1 The up-regulated genes during neural differentiation ID Gene sybom Description H1 control 12 h 24 h 36 h 48 h 72 h Homo sapiens tumor necrosis factor receptor 55504 TNFRSF19 superfamily, member 19 (TNFRSF19), 6.830849797 7.214806054 7.472773008 9.08939763 10.20932917 11.50938675 transcript variant 2, mRNA Homo sapiens Rho GTPase activating 55843 ARHGAP15 6.17857567 6.216531767 6.237494906 6.92954502 8.095552273 10.98011909 15 (ARHGAP15), mRNA Homo sapiens G protein-coupled receptor 177 79971 GPR177 6.986815196 7.260690458 7.504402284 7.631032473 8.792017237 10.96460516 (GPR177), transcript variant 1, mRNA Homo sapiens FEZ family zinc finger 1 389549 FEZF1 3.972339086 5.358360747 5.443897218 8.43485166 9.549402363 10.59274282 (FEZF1), mRNA Homo sapiens 6 open reading 135398 C6orf141 6.503771753 7.556295859 7.636410195 9.177139714 9.822380213 10.02537153 frame 141 (C6orf141), mRNA Homo sapiens odz, odd Oz/ten-m homolog 4 26011 ODZ4 6.598836662 6.626606051 6.711838711 7.877950278 8.751594784 9.922999157 (Drosophila) (ODZ4), mRNA Homo sapiens fibronectin leucine rich transmembrane 23767 FLRT3 3.626651121 3.833963324 4.818916747 6.398488074 8.155490718 9.664250325 protein 3 (FLRT3), transcript variant 2, mRNA PREDICTED: Homo sapiens hypothetical protein 154860 LOC154860 5.356789132 5.678161684 5.917150226 6.916955886 8.24676915 9.602810864 LOC154860 (LOC154860), mRNA Liu, X., et al. Sci China Life Sci Homo sapiens lysophosphatidic acid receptor 1 1902 LPAR1 5.86996711 6.406413681 6.79864415 7.194754276 8.587147236 9.55627538 (LPAR1), transcript variant 2, mRNA Homo sapiens receptor tyrosine kinase-like orphan 4920 ROR2 6.059217133 6.211531847 6.564120006 7.208737268 8.222128997 9.278422908 receptor 2 (ROR2), mRNA 80000 KIAA1772 Homo sapiens KIAA1772 (KIAA1772), mRNA 1.572525113 3.155319589 5.455055566 7.343787281 8.511317586 9.250167241 167410 LIX1 Homo sapiens Lix1 homolog (mouse) (LIX1), mRNA 4.44535219 4.514146891 4.617019362 5.740085107 6.632118464 9.066661448 Homo sapiens transmembrane protein 88 92162 TMEM88 5.620700621 5.689882533 6.531574126 7.226391388 8.061386561 8.961945841 (TMEM88), mRNA Homo sapiens glutamate decarboxylase 1 (brain, 2571 GAD1 0.77604603 1.782283371 2.222315993 6.397986612 7.352340637 8.880402206 67 kD) (GAD1), transcript variant GAD67, mRNA Homo sapiens Meis homeobox 2 (MEIS2), 4212 MEIS2 0.978805652 3.496438923 4.918093931 7.215552666 8.003551879 8.541070891 transcript variant g, mRNA Homo sapiens zinc finger protein 521 25925 ZNF521 5.411837991 5.535498066 5.684552989 6.632614985 7.380407004 8.510631833 (ZNF521), mRNA Homo sapiens ADP-ribosylation factor-like 10123 ARL4C 5.039876895 5.495476425 5.835146491 6.541084639 7.201772367 8.125652124 4C (ARL4C), mRNA Homo sapiens low density lipoprotein-related 4036 LRP2 4.503674852 4.85454438 4.976941965 5.947083947 6.3239745 7.821831198 protein 2 (LRP2), mRNA Homo sapiens F-box and leucine-rich repeat 222235 FBXL13 4.413503397 5.502357416 5.654677672 6.221045141 7.136857665 7.654617629 protein 13 (FBXL13), mRNA PREDICTED: Homo sapiens hypothetical 388279 LOC388279 3.131115874 3.358491481 4.944194414 6.040172251 6.770683589 6.80798743 LOC388279 (LOC388279), mRNA Homo sapiens gasdermin B (GSDMB), transcript 55876 GSDMB 3.653411767 3.696176209 3.882669218 4.964604624 5.226769184 6.246861741 variant 2, mRNA Homo sapiens doublecortin domain containing 196296 DCDC5 0.254830356 1.036734589 1.728886426 2.709195994 4.279110981 5.22409331 5 (DCDC5), mRNA 57577 KIAA1407 Homo sapiens KIAA1407 (KIAA1407), mRNA 1.815636073 1.908543054 3.196017148 3.783685202 4.700357891 4.813203358 LOC10013 PREDICTED: Homo sapiens hypothetical 100132649 0.242585697 2.458472694 3.135436201 3.708588737 4.161074932 4.31509408 2649 LOC100132649 (LOC100132649), mRNA 3 4 Liu, X., et al. Sci China Life Sci

Figure 1 FEZF1 as a potential regulator of early human neural differentiation. A, Heat map of gradually increased transcriptional factors during early neural differentiation from H1 micro-array. Fold of changes in log2 scale. B and C, FEZF family (FEZF1 and FEZF2), neural progenitor specific markers (PAX6, OTX2, SOX2) and pluripotency-associated markers (OCT-4 and NANOG) dynamic expression levels during the 5 days of neural differentiation were analyzed by real-time PCR (B) or Western blot (C) analysis. Relative expression is shown normalized to the level (=1) of mRNA extracted from H1 cells cultured in mTeSR1 before KSR and compound C treated. GAPDH was used as loading control. Data represent mean±standard error (SE) of the mean for three independent experiments. tion, shFEZF1-expressing H1 and H9 cells presented higher Real-time PCR and Western blot analysis showed that FEZF1 expression levels of the pluripotent markers (OCT-4 and deletion did not alter the expression of pluripotency-associ- NANOG) than control cells, suggesting FEZF1 knockdown ated markers (NANOG and OCT4) (Figure 3E and F), indi- impairs hESCs neural differentiation partially by maintaining cating that FEZF1 deletion had no effect on hESC pluripo- the pluripotency program. tency.

Establishment of FEZF1-knockout H1 hESC lines using FEZF1 deletion abrogates neural differentiation of CRISPR-CAS9 technology hESCs To further study the function of FEZF1 in hESC early neural We next determined the effect of FEZF1 deletion on hESC specification, we ruled out the potential ambiguity brought neural specification. Consistent with the result of FEZF1 si- by shRNA, and knockouted FEZF1 gene in hESCs using lencing, FEZF1 deletion significantly reduced the expression CRISPR-CAS9 technology. We designed single guide RNA levels of PAX6, SOX2 and OTX2 mRNA and protein after 5 (sgRNA) sequences targeting exon1 of FEZF1 human gene days of neural differentiation (Figure 4A and B), as assessed (Figure 3A) and confirmed its efficacy at the target locus us- with real-time PCR and Western blot analysis. Immunofluo- ing the Surveyor assay (Figure 3B). After picking and ex- rescence analysis further confirmed the suppression of PAX6 panding individual clones, two homozygous H1 hESC lines induction after FEZF1 deletion (Figure 4C). In agreement with FEZF1 deletion were established through Western blot with the results of FEZF1 silencing, FEZF1−/− H1 cells also and gene sequencing analysis (Figure 3C and D). FEZF1−/− exhibited higher expression levels of the pluripotent markers H1 hESCs grew as compact morphologically normal clonies. (OCT-4 and NANOG) than WT H1 cells (Figure 4D and E), Liu, X., et al. Sci China Life Sci 5

Figure 2 Knockdown FEZF1 inhibited neural induction from hESCs. A and B, The expression level of FEZF1 in Scramble-expressing H1, shFEZF1 645- expressing H1 cells and sh-FEZF1 1683-expressing H1 cells was analyzed by realtime-PCR (A), or Western blot (B) analysis. Relative expression is shown normalized to the level (=1) of mRNA extracted from Scramble-expressing H1 after neural induction. α-tubulin was used as loading control. Data represent mean±standard error of the mean for three independent experiments. C and D, The expression level of neural progenitor specific markers (PAX6, SOX2 and OTX2) in wild-type H1 and shFEZF1-expressing H1 cells was analysis by real-time PCR (C), or Western blot (D) analysis. All values are normalized to the level (=1) of mRNA extracted from Scramble-expressing H1 cells after neural differentiation was induced. GAPDH was used as loading control. Data represent mean±standard error of the mean for three independent experiments. E and F, The expression level of pluripotency-associated markers (OCT-4 and NANOG) was show by real-time PCR (E), or Western blot (D) analysis. Relative expression is shown normalized to the level (=1) of mRNA extracted from Scramble-expressing H1 cells after neural differentiation was induced. α-tubulin was used as loading control. Data represent mean±SE of the mean for three independent experiments. *, P<0.05; **, P<0.01; ***, P<0.001; NS, not significant. confirming that the impairment of pluripotency exit partially differentiation of hESCs, we next asked whether FEZF1 over- contributes to the suppression of hESC neural induction me- expression is sufficient to drive hESC neural specification un- diated by FEZF1 deletion. der pluripotency-maintaining conditions. To address this hy- pothesis, we established H1 hESCs that ectopically overex- Forced expression of FEZF1 is not sufficient to drive neu- pressed FEZF1 by using a previously described doxcycline ral differentiation in hESCs (DOX) inducible lentiviral expression system. Specifically, Above results suggest that FEZF1 deletion abrogates neural we generated a GFP-FEZF1 fusion gene and monitored the 6 Liu, X., et al. Sci China Life Sci

Figure 3 Deletion of FEZF1 in H1 cells using CRISPR/Cas9 technology. A, Schematic representation of the Cas9/sgRNA-targeting sequences at the human FEZF1 locus. Target site in exon 1 is shown below in large scale and the protospacer-adjacent motif (PAM) is labeled in larger font. B, Quantification of CRISPR targeting at exon1 by the surveyor assay. C, Western blot analysis of FEZF1 in H1, H1 FEZF1−/− 5# and FEZF1−/− 6# cells. Cells were treated with KSR and compond C for 5 days. α-tubulin was used as loading control. D, DNA sequencing of mutant lines. Sequence alignments revealed breakpoints and mismatch repair in H1 FEZF1−/− 5# and FEZF1−/− 6# cells. E and F, The expression level of pluripotency-associated markers (OCT-4 and NANOG) in H1, H1 FEZF1−/− 5# and FEZF1−/− 6# cells by real-time PCR (E) , or Western blot (F) analysis. Relative expression is shown normalized to the level (=1) of mRNA extracted from wild-type H1 cells in the differentiation system. α-tubulin was used as loading control. Data represent mean±SE of the mean for three independent experiments. *, P<0.05; **, P<0.01; ***, P<0.001. expression of FEZF1 in hESCs through GFP expression. Af- pression was undetectable in the absence of DOX (Figure ter 24 h of DOX addition, the high level of GFP expression 5A). The forced expression of FEZF1 was confirmed through was seen under the fluorescence micro-scopy, while GFP ex- real-time PCR and Western blot analysis (Figure 5B and C). Liu, X., et al. Sci China Life Sci 7

Figure 4 FEZF1 deletion impairs early neural differentiation of hESCs. A and B, The expression level of neural progenitor markers (PAX6, SOX2 and OTX-2) in H1, H1 FEZF1−/− 5# and FEZF1−/− 6# cells was shown by real-time PCR (A), or Western blot (B) analysis. Relative expression is shown normalized to the level (=1) of mRNA extracted from wild-type H1 cells in the differentiation system. α-tubulin was used as loading control. C, Immunofluorescence analysis of the expression of PAX6 in H1, H1 FEZF1−/− 5# and FEZF1−/− 6# cells. Nuclei were labeled with DAPI (blue). Scale bars, 100 μm. D and E, The expression level of pluripotency-associated markers (OCT-4 and NANOG) in H1, H1 FEZF1−/− 5# and FEZF1−/− 6# cells was analyzed by real-time PCR (D), or Western blot (E) analysis. Relative expression is shown normalized to the level (=1) of mRNA extracted from wild-type H1 cells in the differentiation system. α-tubulin was used as loading control. Data represent mean±SE of the mean for three independent experiments. *, P<0.05; **, P<0.01; ***, P<0.001.

We determined the expression of PAX6 5 days after DOX sults, we ectopically overexpressed FEZF1 in H9 cells by us- addition to mTeSR, the conditions of pluripotency-maintain- ing FEZF1-expressing lentiviral vector and GFP-expressing ing. As shown in Figure 5D and E, DOX addition could not H9 cells were used as negative control. The enforced expres- induce the expression of PAX6. To further confirm these re- sion of FEZF1 was confirmed by real-time PCR and Western 8 Liu, X., et al. Sci China Life Sci

Figure 5 Forced expression of FEZF1 is not sufficient to drive neural differentiation in hESCs. A, Fluorescence images of inducible FEZF1-expressing H1 colonies with or without DOX. Scale bar, 50 μm. B, realtime-PCR analysis showing FEZF1 transcript over-expression in hESCs. C, Western blot analysis showing ectopic FEZF1 protein expression in hESCs. D and E, The expression level of PAX6 in H1 cells under self-renewal condition with or without GFP- FEZF1 over-expression and H1 cells treated with KSR and compound C was shown by real-time PCR (D), or Western blot (E) analysis. All values were normalized to the level (=1) of mRNA in H1 cells cultured in mTeSR1 medium without GFP-FEZF1 over-expression. GAPDH was used as a loading control. Results from three independent experiments are shown as mean±SE. blot analysis (Figure S1A and B in Supporting Information). development and patterning of the diencephalon (Eckler et Consistent with the results in H1 cells, the forced expression al., 2011; Hirata et al., 2006; Shimizu et al., 2010; Watanabe of FEZF1 had minimal effect on the expression of PAX6 5 et al., 2009). However, little is known about its role in early days after neural differentiation (Figure S1C and D in Sup- neural induction. Microarray and real-time PCR show that porting Information). Together, these results indicate that the expression of FEZF1 was induced after H1 hESC neural forced expression of FEZF1 is not sufficient to drive neural induction, much more early than PAX6, one well-known differentiation in hESCs. neural progenitor marker (Zhang et al., 2010), suggesting that FEZF1 may play an important role in human early DISCUSSION neural specification. We accessed the function of FEZF1 in neural induction through generating the hESC cell line with Although the role of extrinsic signaling factors in hESC FEZF1 silencing and knock-out and found that both FEZF1 neural differentiation has been well documented (Chambers silencing and knock-out abrogates neural differentiation of et al., 2009; Gamse and Sive, 2000; Moya et al., 2014; hESCs, supporting FEZF1 as a key player in human early Niederreither et al., 2000; Pera et al., 2004; Smith et al., neural induction. Interestingly, although FEZF1 deletion 2008; Vallier et al., 2005), the transcriptional factors impli- has no effect on the expression of pluripotency-associated cated in this process remain largely unknown. We previously markers in pluripotency-maintaining conditions, loss of established single-step procedures for neural induction of FEZF1 abrogates the exit of hESCs from the pluripotent hESC by use of a single chemical compound, compound state in neural induction conditions. This is attributed to C, and identified MSX2 as a negative regulator of neural the fact that FEZF1 shows no or very low expression levels specification (Wu et al., 2015). To further discover novel in pluripotency-maintaining conditions while it is quickly regulators of neural differentiation, we screened the up-reg- induced in neural differentiation conditions and its high ulated genes in early stage of hESC neural induction by expression promotes pluripotency exit, which accelerates applying microarray gene expression analysis and identified neural differentiation of hESCs due to the well-established FEZF1 as a potential regulator of human early neurogen- role of NANOG in suppressing neural differentiation (Xu esis. Animal studies have demonstrated that FEZF1 plays et al., 2008). Previous animal studies also demonstrate that crucial roles in forebrain neurogenesis, olfactory system FEZF2 plays an important role in neural development and al- Liu, X., et al. Sci China Life Sci 9 ways functions redundantly with FEZF1 (Hirata et al., 2006; described (Zhou et al., 2010). Briefly, H1 cells were disso- Shimizu et al., 2010). However, FEZF2 showed significantly ciated into single-cell suspension using 1 mg mL−1 Accutase low expression compared with FEZF1 in early human neural (Gibco, USA) for 3–5 min and plated on Matrigel-coated induction, indicating that FEZF1 is likely more potent than dishes at a density of 105 cells cm−2 in mTeSR1 medium FEZF2 in inducing neural specification in hESCs. with 10 μmol L−1 Y-27632 (Calbiochem, USA). After The inhibitory effect of FEZF1 deletion on neural induction 24 h, Y-27632 was withdrawn and the H1 cells were cultured led us to investigate whether enforced expression of FEZF1 to confluency. To initiate neural induction, hESCs were culti- suffice to induce neural specification in hESCs under pluripo- vated in DMEM/F12 containing 20% KSR, 1% non-essential tency-maintaining conditions. However, PAX6 expression amino acid, β-mercaptoethanol (0.1 mmol L−1), glutamine failed to be induced 120 h after DOX addition in mTeSR, (1 mmol L−1) and 1 mmol L−1 compound C (EMD Bio- indicating that FEZF1 alone is not sufficient to drive neu- sciences, USA) and medium were replaced every day. ral differentiation in hESCs. Recently, various transcriptional factor combinations have been used to directly reprogram fi- Establishing GFP-FEZF1 inducible overexpression stable broblast into neural progenitors. Han et al. have reported hESC lines that a combination of transcription factors (Brn4, Sox2, Klf4, To establish GFP-FEZF1 inducible overexpression stable H1 c-Myc and E47) induced mouse fibroblasts to neural stem cell hESC lines, we used Lenti-X™ Tet-On® Advanced Inducible (Han et al., 2012). A recent study demonstrated that human Expression System (Clontech, Japan). Firstly, we cloned fibroblasts could be directly converted to neural progenitors FEZF1 coding sequences into pLVX-Tight-GFP-Puro vector by six combinations (CBX2, HES1, ID1, which could express GFP to monitor FEZF1 expression. TFAP2A, ZFP42, or ZNF423) (Hou et al., 2017). Although Primers used are listed in Table S1 in Supporting Informa- FEZF1 alone is insufficient to convert hESCs into neural pro- tion. To infect H1 cells, we packaged lentiviral particles genitors, it deserves investigation whether FEZF1 is able to using Viralpower Lentivirus Packaging System (Invitrogen, replace one or more transcriptional factors reported previ- USA) according to the manufacturer’s instructions in 293T ously during reprogramming of human neural stem cells or cells. H1 cells were infected with TetR and GFP-FEZF1 enhance reprograming efficiency in combination with other in sequence and selected with G418 (300 μg mL−1, Sigma, reprogram transcriptional factors. USA) and puromycin (1 μg mL−1, Sigma). After several In conclusion, we have identified FEZF1 as a critical tran- rounds of selection, we dissociated cells with Accutase scriptional factor of neural lineage commitment from hESCs. (Gibco) into single cells and picked colonies derived from Either FEZF1 knockdown or knockout impaired neural in- single cells into independent well. Colonies with high ef- duction in hESCs and the suppressed exit of pluripotency ficiency of GFP expression were identified through adding could partially account for neural induction defect. We show doxycycline (2 μg mL−1, Sigma) and testing by fluorescence that FEZF1 is one of the early genes during hESC neural dif- microscopy and flow cytometry analysis. After several ferentiation, before PAX6, suggesting that it may be used as rounds of selection, the cell lines homogenously expressing neural progenitor marker to predict the neural lineage com- GFP in the presence of doxycycline were expanded and mitment in hESCs. Our study provides mechanistic insight cryopreserved. into the endogenous regulators of early human neural devel- Generation of FEZF1 knockout hESC lines by opment and can be proved invaluable for manipulating cell lenti-CRISPR-Cas9 system fate during neural differentiation and for large-scale neural progenitor cells generation for future clinical applications. Lenti-V2 vector expressing Cas9 and sgRNA was di- gested with Bsmb I and linearized with gel purified. A pair of oligos targeting the exon 1 of the human FEZF1 MATERIALS AND METHODS gene, which were designed using the CRISPR Design Tool (http://tools.genome-engineering.org), were an- hESC culture nealed and ligated to linearized vector. The sequence of H1 and H9 hESC lines (WiCell Research Institute, USA) Lenti-V2-SgFEZF1-E1G1 plasmid was verified by In- were maintained on Matrigel (BD Biosciences, USA)-coated vitrogen sequencing service and packaged into lentiviral plates in mTeSR1 medium (StemCell Technologies, Canada). particles using Viralpower Lentivirus Packaging Sys- For maintenance of self-renewal, cells were fed daily and pas- tem (Invitrogen). H1 cells were infected with Lentivirus saged using dispase (2 U mL−1, StemCell Technologies) at a containing SgFEZF1-E1G1 and selected with puromycin dilution of 1:6 every 3–5 days. (1 μg mL−1, Sigma). Surveyor assay was used to access the genome editing efficiency. H1-SgFEZF1-E1g1 cells hESC neural differentiation with high cleavage efficiency were dissociated into single Neural differentiation of hESCs was performed as previously cell using Accutase. Once the small colonies emerged, we 10 Liu, X., et al. Sci China Life Sci picked and expanded. Western blot assays verified FEZF1 Compliance and ethics The author(s) declare that they have no conflict knockout. hESCs with FEZF1 deletion were expanded and of interest. cryopreserved. The sgRNA sequences and the primers used for Surveyor assay are shown in Table S2 in Supporting Information. Acknowledgements This work was supported by National Ba- sic Research Program of China (2015CB964902 to Jiaxi Zhou and Real-time PCR SQ2016ZY05002105 to Hongtao Wang), CAMS Initiative for Innovative Medicine (2016-I2M-1-018, 2016-I2M-3-002), National Natural Science We extracted total mRNA from cells using TRIzol (Invitro- Foundation of China (81530008, 31671541 to Jiaxi Zhou, 31500949 to gen) following the manufacturer’s instructions, and synthe- Hongtao Wang), Tianjin Natural Science Foundation (16JCZDJC33100 to Jiaxi Zhou), and 3332015128 supported by PUMC Youth Fund and sized complementary DNA using Reverse Transcription Sys- Fundamental Research Funds for the Central Universities to Dr. Hongtao tem (Promega, USA). Real-time PCR was performed using Wang. Quanti-Tech SYBR Green PCR kit (Qiagen) following the manufacturer’s protocol on the ABI 7000 real-time PCR ma- Agoston, Z., Heine, P., Brill, M.S., Grebbin, B.M., Hau, A.C., chine (Applied Biosystems). Primers used for real-time PCR Kallenborn-Gerhardt, W., Schramm, J., Götz, M., and Schulte, D. are listed in Table S1 in Supporting Information. The genes (2014). Meis2 is a Pax6 co-factor in neurogenesis and dopaminergic periglomerular fate specification in the adult olfactory bulb. Develop- relative expression level was analyzed using the comparative ment 141, 28–38. Ct method and normalized to ACTIN gene levels. Real-time Chadwick, K., Wang, L., Li, L., Menendez, P., Murdoch, B., Rouleau, A., PCR data were presented as the mean of three independent and Bhatia, M. (2003). Cytokines and BMP-4 promote hematopoietic experiments and error bars indicated SD. differentiation of human embryonic stem cells. Blood 102, 906–915. Chambers, S.M., Fasano, C.A., Papapetrou, E.P., Tomishima, M., Sadelain, Western blot M., and Studer, L. (2009). Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD signaling. Nat Biotechnol The hESC cultures were lysed using mixture containing 27, 275–280. laemmli sample buffer (BioRad, USA) and β-ME at ratio Chen, B., Wang, S.S., Hattox, A.M., Rayburn, H., Nelson, S.B., and 19:1. The cell lysates were resolved on 10% sodium dodecyl McConnell, S.K. (2008). The Fezf2-Ctip2 genetic pathway regulates the fate choice of subcortical projection neurons in the developing cerebral sulfate polyacrylamide gel and transferred onto nitrocellulose cortex. Proc Natl Acad Sci USA 105, 11382–11387. transfer membranes (Whatman, UK). The membranes were D’Amour, K.A., Agulnick, A.D., Eliazer, S., Kelly, O.G., Kroon, E., and blocked for 1 h at room temperature with 5% nonfat milk in Baetge, E.E. (2005). Efficient differentiation of human embryonic stem PBS containing 0.1% Tween 20. For targeted protein detec- cells to definitive endoderm. Nat Biotechnol 23, 1534–1541. Eckler, M.J., McKenna, W.L., Taghvaei, S., McConnell, S.K., and Chen, tion, the membranes were incubated with primary antibodies B. (2011). Fezf1 and Fezf2 are required for olfactory development and overnight at 4°C, and then incubated with HRP-coupled sensory neuron identity. J Comp Neurol 519, 1829–1846. secondary antibodies for 2 h at room temperature. Targeted Gamse, J., and Sive, H. (2000). Vertebrate anteroposterior patterning: the were detected using the ECL Detection Reagent. Xenopus neurectoderm as a paradigm. Bioessays 22, 976–986. Greber, B., Coulon, P., Zhang, M., Moritz, S., Frank, S., Müller-Molina, A.J., α-Tubulin was used as a loading control. Antibodies and Araúzo-Bravo, M.J., Han, D.W., Pape, H.C., and Schöler, H.R. (2011). dilutions are listed in Table S3 in Supporting Information. FGF signalling inhibits neural induction in human embryonic stem cells. EMBO J 30, 4874–4884. Immunofluorescence Guo, C., Eckler, M.J., McKenna, W.L., McKinsey, G.L., Rubenstein, J.L.R., Culture adherent cells were fixed in 4% paraformaldehyde and Chen, B. (2013). Fezf2 expression identifies a multipotent progeni- tor for neocortical projection neurons, astrocytes, and oligodendrocytes. and permeabilized with 0.2% Triton X-100 in PBS. Non-spe- Neuron 80, 1167–1174. cific proteins were blocked by 2% BSA in PBS. ForPAX6 Han, D.W., Tapia, N., Hermann, A., Hemmer, K., Höing, S., Araúzo-Bravo, detection, cells were incubated overnight with anti-PAX6 M.J., Zaehres, H., Wu, G., Frank, S., Moritz, S., Greber, B., Yang, J.H., antibody (1:100 dilution, Biolegend, USA) and then sub- Lee, H.T., Schwamborn, J.C., Storch, A., and Schöler, H.R. (2012). Di- rect reprogramming of fibroblasts into neural stem cells by defined fac- sequently visualized with fluorescent secondary antibodies. tors. Cell Stem Cell 10, 465–472. The nuclei were labeled by DAPI (Sigma) for 15 min at Hirata, T., Nakazawa, M., Muraoka, O., Nakayama, R., Suda, Y., and Hibi, room temperature. Images were collected with Leica SP2 M. (2006). Zinc-finger genes Fez and Fez-like function in the establish- microscope. Antibodies and dilutions are listed in Table S3 ment of diencephalon subdivisions. Development 133, 3993–4004. Hou, P.S., Chuang, C.Y., Yeh, C.H., Chiang, W., Liu, H.J., Lin, T.N., and in Supporting Information. Kuo, H.C. (2017). Direct conversion of human fibroblasts into neural progenitors using transcription factors enriched in human ESC-derived Statistical analysis neural progenitors. Stem Cell Rep 8, 54–68. All data are shown as mean±SD. Data were analyzed using Lu, P., Chen, X., Feng, Y., Zeng, Q., Jiang, C., Zhu, X., Fan, G., and Xue, GraphPad Prism software. Student’s t-test was used for sta- Z. (2016). Integrated transcriptome analysis of human iPS cells derived from a fragile X syndrome patient during neuronal differentiation. Sci tistical evaluation. P<0.05 was considered as statistically sig- China Life Sci 59, 1093–1105. nificant. Matsuo-Takasaki, M., Lim, J.H., Beanan, M.J., Sato, S.M., and Sargent, T.D. Liu, X., et al. Sci China Life Sci 11

(2000). Cloning and expression of a novel zinc finger gene, Fez, tran- J., Nibu, K.I., Sato, N., Iiboshi, Y., Yusa, K., Kondoh, G., Takeda, J., scribed in the forebrain of Xenopus and mouse embryos. Mech Dev 93, Terashima, T., and Takumi, T. (2009). Fezf1 is required for penetration 201–204. of the basal lamina by olfactory axons to promote olfactory development. McKenna, W.L., Betancourt, J., Larkin, K.A., Abrams, B., Guo, C., J Comp Neurol 515, 565–584. Rubenstein, J.L.R., and Chen, B. (2011). Tbr1 and Fezf2 regulate alter- Wu, Q., Zhang, L., Su, P., Lei, X., Liu, X., Wang, H., Lu, L., Bai, Y., Xiong, nate corticofugal neuronal identities during neocortical development. J T., Li, D., Zhu, Z., Duan, E., Jiang, E., Feng, S., Han, M., Xu, Y., Wang, Neurosci 31, 549–564. F., and Zhou, J. (2015). MSX2 mediates entry of human pluripotent stem Moya, N., Cutts, J., Gaasterland, T., Willert, K., and Brafman, D.A. (2014). cells into mesendoderm by simultaneously suppressing SOX2 and acti- Endogenous WNT signaling regulates hPSC-derived neural progenitor vating NODAL signaling. Cell Res 25, 1314–1332. cell heterogeneity and specifies their regional identity. Stem Cell Rep 3, Xu, R.H., Sampsell-Barron, T.L., Gu, F., Root, S., Peck, R.M., Pan, G., Yu, 1015–1028. J., Antosiewicz-Bourget, J., Tian, S., Stewart, R., and Thomson, J.A. Niederreither, K., Vermot, J., Schuhbaur, B., Chambon, P., and Dolle, P. (2008). NANOG is a direct target of TGFβ/activin-mediated SMAD sig- (2000). Retinoic acid synthesis and hindbrain patterning in the mouse naling in human ESCs. Cell Stem Cell 3, 196–206. embryo. Development 127, 75–85. Yamanaka, S. (2009). A fresh look at iPS cells. Cell 137, 13–17. Paganoni, S., Bernstein, J., and Ferreira, A. (2010). Ror1-Ror2 complexes Yao, S., Chen, S., Clark, J., Hao, E., Beattie, G.M., Hayek, A., and Ding, modulate synapse formation in hippocampal neurons. Neuroscience 165, S. (2006). Long-term self-renewal and directed differentiation of human 1261–1274. embryonic stem cells in chemically defined conditions. Proc Natl Acad Pera, M.F., Andrade, J., Houssami, S., Reubinoff, B., Trounson, A., Stanley, Sci USA 103, 6907–6912. E.G., Ward-van, O.D., and Mummery, C. (2004). Regulation of human Zamboni, V., Armentano, M., Sarò, G., Ciraolo, E., Ghigo, A., Germena, G., embryonic stem cell differentiation by BMP-2 and its antagonist noggin. Umbach, A., Valnegri, P., Passafaro, M., Carabelli, V., Gavello, D., Bian- J Cell Sci 117, 1269–1280. chi, V., D’Adamo, P., de Curtis, I., El-Assawi, N., Mauro, A., Priano, L., Robinson, M., Parsons Perez, M.C., Tébar, L., Palmer, J., Patel, A., Marks, Ferri, N., Hirsch, E., and Merlo, G.R. (2016). Disruption of ArhGAP15 D., Sheasby, A., De Felipe, C., Coffin, R., Livesey, F.J., and Hunt, S.P. results in hyperactive Rac1, affects the architecture and function of hip- (2004). FLRT3 is expressed in sensory neurons after peripheral nerve pocampal inhibitory neurons and causes cognitive deficits. Sci Rep 6, injury and regulates neurite outgrowth. Mol Cell Neurosci 27, 202–214. 34877. Shimizu, T., Nakazawa, M., Kani, S., Bae, Y.K., Shimizu, T., Kageyama, R., Zhang, C., Cao, J., Li, X., Xu, H., Wang, W., Wang, L., Zhao, X., Li, W., and Hibi, M. (2010). Zinc finger genes Fezf1 and Fezf2 control neuronal Jiao, J., Hu, B., Zhou, Q., and Zhao, T. (2016). Treatment of multiple differentiation by repressing Hes5 expression in the forebrain. Develop- sclerosis by transplantation of neural stem cells derived from induced ment 137, 1875–1885. pluripotent stem cells. Sci China Life Sci 59, 950–957. Smith, J.R., Vallier, L., Lupo, G., Alexander, M., Harris, W.A., and Pedersen, Zhang, S., and Cui, W. (2014). Sox2, a key factor in the regulation of pluripo- R.A. (2008). Inhibition of Activin/Nodal signaling promotes specifica- tency and neural differentiation. World J Stem Cells 6, 305–311. tion of human embryonic stem cells into neuroectoderm. Dev Biol 313, Zhang, X., Huang, C.T., Chen, J., Pankratz, M.T., Xi, J., Li, J., Yang, Y., 107–117. LaVaute, T.M., Li, X.J., Ayala, M., Bondarenko, G.I., Du, Z.W., Jin, Y., Thomson, J.A., Itskovitz-Eldor, J., Shapiro, S.S., Waknitz, M.A., Swiergiel, Golos, T.G., and Zhang, S.C. (2010). Pax6 is a human neuroectoderm J.J., Marshall, V.S., and Jones, J.M. (1998). Embryonic stem cell lines cell fate determinant. Cell Stem Cell 7, 90–100. derived from human blastocysts. Science 282, 1145–1147. Zhou, J., Su, P., Li, D., Tsang, S., Duan, E., and Wang, F. (2010). High-ef- Vallier, L., Alexander, M., and Pedersen, R.A. (2005). Activin/Nodal and ficiency induction of neural conversion in human ESCs and human FGF pathways cooperate to maintain pluripotency of human embryonic induced pluripotent stem cells with a single chemical inhibitor of stem cells. J Cell Sci 118, 4495–4509. transforming growth factor beta superfamily receptors. Stem Cells 28, Watanabe, Y., Inoue, K., Okuyama-Yamamoto, A., Nakai, N., Nakatani, 1741–1750.

SUPPORTING INFORMATION

Figure S1 Knockdown FEZF1 inhibited neural induction from hESCs. Table S1 The primers used for real-time PCR Table S2 The primers used for CRISPR sgRNA guide sequences and the genotyping Table S3 The sources and dilutions of the antibodies

The supporting information is available online at http://life.scichina.com and https://link.springer.com. The supporting materials are published as submitted, without typesetting or editing. The responsibility for scientific accuracy and content remains entirely with the authors.