University of Massachusetts Medical School eScholarship@UMMS Open Access Articles Open Access Publications by UMMS Authors 2019-04-16 Single-Cell RNA-Sequencing-Based CRISPRi Screening Resolves Molecular Drivers of Early Human Endoderm Development Ryan M. Genga University of Massachusetts Medical School Et al. Let us know how access to this document benefits ou.y Follow this and additional works at: https://escholarship.umassmed.edu/oapubs Part of the Amino Acids, Peptides, and Proteins Commons, Cell Biology Commons, Cells Commons, Developmental Biology Commons, Embryonic Structures Commons, Genetic Phenomena Commons, and the Nucleic Acids, Nucleotides, and Nucleosides Commons Repository Citation Genga RM, Kernfeld EM, Parsi KM, Parsons TJ, Ziller MJ, Maehr R. (2019). Single-Cell RNA-Sequencing- Based CRISPRi Screening Resolves Molecular Drivers of Early Human Endoderm Development. Open Access Articles. https://doi.org/10.1016/j.celrep.2019.03.076. Retrieved from https://escholarship.umassmed.edu/oapubs/3818 Creative Commons License This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License. This material is brought to you by eScholarship@UMMS. It has been accepted for inclusion in Open Access Articles by an authorized administrator of eScholarship@UMMS. For more information, please contact [email protected]. Report Single-Cell RNA-Sequencing-Based CRISPRi Screening Resolves Molecular Drivers of Early Human Endoderm Development Graphical Abstract Authors Ryan M.J. Genga, Eric M. Kernfeld, Krishna M. Parsi, Teagan J. Parsons, Michael J. Ziller, Rene´ Maehr Correspondence [email protected] In Brief Genga et al. utilize a single-cell RNA- sequencing-based CRISPR interference approach to screen transcription factors predicted to have a role in human definitive endoderm differentiation. The perturbation screen identifies an important role of TGFb signaling-related factors. Follow-up of FOXA2 reveals genome-wide molecular changes and altered differentiation competency in endoderm. Highlights d atacTFAP analysis predicts key factors during human ESC differentiation to endoderm d scRNA-seq CRISPRi identifies factors important for human endoderm differentiation d Targeting of the TGFb pathway affects differentiation in a target-specific manner d FOXA2 knockdown causes genome-wide changes and impairs differentiation Genga et al., 2019, Cell Reports 27, 708–718 April 16, 2019 ª 2019 The Author(s). https://doi.org/10.1016/j.celrep.2019.03.076 Cell Reports Report Single-Cell RNA-Sequencing-Based CRISPRi Screening Resolves Molecular Drivers of Early Human Endoderm Development Ryan M.J. Genga,1,3 Eric M. Kernfeld,1,3 Krishna M. Parsi,1 Teagan J. Parsons,1 Michael J. Ziller,2 and Rene´ Maehr1,4,* 1Program in Molecular Medicine, Diabetes Center of Excellence, University of Massachusetts Medical School, Worcester, MA 01605, USA 2Department of Translational Psychiatry, Max Planck Institute of Psychiatry, 80804 Munich, Germany 3These authors contributed equally 4Lead Contact *Correspondence: [email protected] https://doi.org/10.1016/j.celrep.2019.03.076 SUMMARY CRISPR interference (CRISPRi) systems can effectively disrupt gene function in human pluripotent stem cells (Kearns Studies in vertebrates have outlined conserved mo- et al., 2014; Mandegar et al., 2016), with low off-target effects lecular control of definitive endoderm (END) develop- (Gilbert et al., 2014). CRISPRi can also be combined with ment. However, recent work also shows that key droplet-based, single-cell RNA-sequencing (scRNA-seq) read- molecular aspects of human END regulation differ outs (Adamson et al., 2016; Xie et al., 2017), allowing functional even from rodents. Differentiation of human embry- analysis of molecular pathways guiding differentiation while onic stem cells (ESCs) to END offers a tractable sys- balancing resolution and throughput. Here, we predicted candi- date molecular drivers of END differentiation (END-Diff) by tem to study the molecular basis of normal and computationally integrating dynamics of chromatin accessibility defective human-specific END development. Here, and transcriptome-wide changes. To delineate candidate roles, we interrogated dynamics in chromatin accessibility we conducted a parallel scRNA-seq CRISPRi screen to perturb during differentiation of ESCs to END, predicting the factors during END-Diff. We uncover distinct blocks in early DNA-binding proteins that may drive this cell fate human END development mediated by loss of TFs involved in transition. We then combined single-cell RNA-seq transforming growth factor b (TGFb) signaling, while perturbation with parallel CRISPR perturbations to comprehen- of the TF FOXA2 results in an altered differentiation competency sively define the loss-of-function phenotype of those at later stages. factors in END development. Following a few candi- dates, we revealed distinct impairments in the differ- RESULTS entiation trajectories for mediators of TGFb signaling and expose a role for the FOXA2 transcription factor Chromatin Accessibility and Transcriptome Dynamics of in priming human END competence for human fore- END-Diff Using an efficient ESC differentiation platform (Figures S1A and gut and hepatic END specification. Together, this S1B), we compare ESC and END by RNA-seq and assay for single-cell functional genomics study provides transposase-accessible chromatin using sequencing (ATAC- high-resolution insight on human END development. seq) (Figure 1A) revealing 2,905 differentially expressed tran- scripts (Figure S1C; Table S2; false-discovery rate [FDR] < INTRODUCTION 0.01; log fold change R 1.0) and differential chromatin accessi- bility at 34,025 sites (Figures 1B and S1D; Table S2; FDR < 0.05; Human embryonic stem cell (ESC) differentiation strategies to log fold change R 1.0), respectively. Analysis by ATAC-seq tran- generate definitive endoderm (END) allow for interrogation of dif- scription factor activity prediction (atacTFAP) of ESC, END, and ferentiation-associated signaling requirements and chromatin pancreatic beta cells was applied to reveal putative molecular states (D’Amour et al., 2005; Gifford et al., 2013; Loh et al., drivers of END-Diff. While many of the predicted DNA-binding 2014). While various transcription factors (TFs) have been evalu- proteins have been associated with mesendoderm and END for- ated for their role in vertebrate END formation (Zorn and Wells, mation (e.g., GATA4, GATA6, GSC, SOX17, FOXH1, FOXA2, 2009), there are notable species differences in TF requirements OTX2, EOMES, SMAD2, SMAD4 , and MIXL1)(Zorn and Wells, (Shi et al., 2017; Tiyaboonchai et al., 2017; Zhu and Huangfu, 2009), other candidates have not been directly implicated in early 2013). For example, recent loss-of-function analyses revealed END-Diff (e.g., ZBTB33, ESRRA, ZNF410, and E2F6). In total, 50 important roles of TFs, including GATA6 and KLF8, specifically TF candidates were selected for functional follow-up, covering a in human END (Allison et al., 2018; Chu et al., 2016; Tiyaboonchai spectrum of potential facilitators and repressors of human END- et al., 2017), highlighting the need for increased throughput in Diff. Among the 50 candidates, two factors were included that functional analyses of TF dependencies in human. are implicated in structural aspects of chromatin organization 708 Cell Reports 27, 708–718, April 16, 2019 ª 2019 The Author(s). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). A B 56kb 41kb ATAC-seq Transcription Factor Activity Prediction (atacTFAP) Analysis [0 - 150] [0 - 30] RNA-seq [0 - 6.00] [0 - 6.00] hESC ATAC-seq hESC Signal Motif TF candidate differential enriched TF expressed TFs [0 - 150] [0 - 30] ATAC peaks binding motifs with motif specificity RNA-seq [0 - 6.00] [0 - 6.00] hESC GATA1 GATA4 ATAC-seq hEND hEND hEND GATA2 GATA6 n=34,025 GATA3 DBA score ATAC-Seq 0 10 OCT4 SOX17 C delta atacTFAP atacTFAP −0.04 0 0.04 log2 FC RNAdiff −4.0 0 4.0 YY1 SP1 MNT MAX ZIC3 GSC ATF7 ATF3 E2F6 ELF3 ETV5 USF1 HSF2 PBX2 OTX2 HBP1 CTCF JUND NFYB ARNT PITX2 RARA TGIF2 MIXL1 MLXIP GATA6 GATA4 FOXJ3 SOX17 SOX11 FOXA2 FOXA3 NR5A2 FOXH1 MYBL2 CDC5L FOXQ1 RREB1 CREB3 FOXM1 SMAD2 SMAD4 ESRRA EOMES ZNF263 ZNF410 ZNF143 ZBTB33 NFE2L2 HMBOX1 D EF Scramble gRNA cells in END Deliver SNN clusters (END) Induce Differentiate Run scRNA-seq SNN (percent of cluster) gRNAs Pool dCas9-KRAB to END with gRNA detection cluster x160 0 4% 1 2 3% p < 2.2x10-16 3 versus random 2% allocation mRNA 1% gRNA ESC-AAVS1- 0 TetOn-dCas9-KRAB G 0 END Cluster 123 H Selected genes in END LEFTY2 MIXL1 LEFTY1 SOX17 CXCR4 FZD5 NODAL SOX17 POU5F1 Ln Normalized Expression Max ID1 ID3 POU5F1 ID1 Genes EPCAM 3 0 LHX1 0 MIXL1 CER1 -3 Standardized ln expression Cells (n=16,110) Figure 1. scRNA-Seq CRISPRi Screen Identifies Molecular Drivers of Human END-Diff (A) Representative integrative genomics viewer (IGV) tracks at the OCT4 and SOX17 loci. RNA-seq and ATAC-seq datasets for H1 ESC or END highlight dynamic transcriptome and chromatin changes. (legend continued on next page) Cell Reports 27, 708–718, April 16, 2019 709 (CTCF and YY1) with described roles in the exit of pluripotency to be regulated by TGFb signaling (SOX17)(Alexander and Stain- (Balakrishnan et al., 2012; Weintraub et al., 2017). ier, 1999)(Figures 2A and 2B; Table S3). FOXH1-specific gRNAs are significantly enriched in cluster 1, SMAD2- and SMAD4-spe- Single-Cell CRISPRi Screening Reveals Candidate cific gRNAs are significantly enriched in cluster 2, and SOX17- Regulators of END-Diff specific gRNAs are enriched in cluster 3 (Figure 2B; Table S3). We utilized a lentiviral guide RNA (gRNA) delivery system (Datlin- To see whether any cluster could be interpreted as a differen- ger et al., 2017) together with a gene-targeted H1-AAVS1- tiation block, we characterized a bulk RNA-seq END-Diff time TetOn-dCas9-KRAB ESC line (Figures S1E–S1H) to assay the course of control ESCs (Figure 2C; transcript list curated from transcriptomic effects of atacTFAP candidate repression on literature; Chu et al., 2016; Loh et al., 2014).