Engineering mouse cell fate controller by rational design – Nature Communications

Engineered cell-fate regulator enhances cell reprogramming

It is conceivable that CiM can be engaged and tested with regulators designed and engineered based on CAD and related considerations. For this purpose, based on the concept of CAD, we compared the differences in CAD between mESCs and MEFs to identify chromatin regions that may require opening during cell reprogramming. We then selected candidate factors by referring to binding sites similar to those in the target cell (Supplementary Fig. 1a). In addition to chromatin remodeling factors like Ss18, Smarca4, Mbd3, Chd7, and Rnf2, we found that pluripotent factors Nanog, Sox2, and Pou5f1 (Oct4) are critical in maintaining open chromatin accessible for transcription in ESC with high GIGGLE score. Considering their importance as core factors for pluripotent maintenance and cell reprogramming, we explored whether fusion the BiD of SS18, a candidate component of CiM previously identified by CRISPR-based screening in a pluripotent-somatic transition system, with these pluripotent factors could achieve cell reprogramming using the GGSGG linker (Fig. 1a). Besides, many reported TFs that could facilitate iPSC induction (Klf4, Esrrb, Prdm14, Zfp296)^22,23 or replace Oct4 (Nr5a2, Jdp2)^24,25 were selected for screening. In addition, Rax and Klf17, which we found they can enhance somatic cell reprogramming (not published results) were also selected. We initially attempted cell reprogramming using a single BiD factor, but only a limited number of clones were produced by Oct4^BiD. Therefore, we hypothesize that introducing additional reprogramming factors alongside the BiD factors that may facilitate reprogramming. Subsequently, we tested these BiD factors in combination with 11 wild-type TFs to convert MEFs to iPSCs and demonstrate that Oct4^BiD with DsRed, Esrrb, or Nanog, as well as Nanog^BiD with Jdp2 or Oct4, could generate iPSCs with varying efficiencies (Fig. 1b, c). As the most efficient combination merged from these attempts is the combination of Nanog^BiD+Oct4, followed by Oct4^BiD+Esrrb (Fig. 1c), we decided to focus on the front runner—Nanog^BiD+Oct4. We further optimized the process by varying the length of culture in iCD3 or 2iL (Fig. 1d). Using the best condition, Nanog+Oct4 (hereafter referred as Nanog^WT) could hardly induce iPSCs, while more than 600 iPSCs colonies can be induced by Oct4+Nanog^BiD (hereafter referred to Nanog^BiD) from 15,000 MEFs (Fig. 1e, f). The reprogramming efficiency is approaching to that of OKS which we have shown previously to be superior to the original OKSM²⁶, the gold standard combination for iPSC generation. Interestingly, when we compare BiD with full-length or IDR of SS18, BiD is the most effective (Supplementary Fig. 1b). As expected, our findings demonstrate that Nanog^BiD iPSCs exhibit transcriptional similarities to mESCs, also can generate chimera with blastocysts injection (Supplementary Fig. 1c, d, f). Moreover, we have collected RNA-seq data from three additional cell reprogramming strategies conducted by other research groups: seven factors-derived reprogramming (GSE127927)²⁵, OKS derived reprogramming (GSE93029)²⁷, and small-molecule induced reprogramming (GSE48252)²⁸. As shown in Supplementary Fig. 1e, we demonstrate that NanogBiD iPSCs exhibit transcriptional similarities to OKS-iPSCs and other types of iPSCs, particularly the small-molecule-induced ones. Genomic PCR detection revealed that all the iPSC is generated by wild type or synthetic Oct4 and Nanog, but not other reprogramming factors (Fig. 1g). Moreover, we isolated mouse tail tip fibroblasts (TTFs) and mouse neonatal fibroblasts (MNFs) from OG2 transgene mice, to investigate the potential of our system in reprogramming other cell types. Our results demonstrate that all these types of cells could be efficiently reprogrammed to Oct4-GFP positive iPSC colonies that can be stably passaged with key pluripotency features (Supplementary Fig. 1g–i), indicating the validity and generalizability of our system.

a Architecture of the synthetic factors. The N-terminus 70aa domain from SS18 was fused to the N-terminus of full-length transcription factors. b Schematic diagram illustrating the process of synthetic BiD factors induced cell reprogramming. c Heatmap of numbers for Oct4-GFP⁺ colonies of MEF cells reprogramming by different combinations. Red rectangles represent mean data for n = 3 biological replicates. d Number of iPSC colonies induced from MEFs infected by Nanog^BiD+Oct4 under different medium switch strategies. Bars are mean with SD, and red plots are individual data points for n = 4 biological replicates. Statistical significance relevant to number of Oct4-GFP⁺ colonies at Day 7 was measured with two-tailed unpaired t-test. P values were performed on graph. NS, nonsignificant. e Number of Oct4-GFP⁺ colonies during the process of Oct4+Sox2+Klf4, Nanog^BiD+Oct4, Nanog+Oct4^BiD, Nanog^BiD+Oct4^BiD, and Nanog+Oct4 mediated reprogramming. Bars are mean ± SD for n = 3 biological replicates. Statistical significance relative to colonies number of Nanog+Oct4 at Day 12 was measured with two-tailed unpaired t-test. P values were performed on graph. f Representative images of different reprogramming strategies induced colonies on day 12. Scale bar, 5 mm. g Detection of plasmid integration by PCR. Gels shown are representative of n = 2 independent experiments. Source data are provided as a Source Data file.

Full size image

SS18 is a component of BAF complex whose components have been shown to enhance reprogramming²⁹. We then compare BAF subunits with Nanog^BiD and show that even the best subunit Smarca4 among them is 40× less efficient (Supplementary Fig. 1j), suggesting that eCR engineering offers a more direct approach to engage the candidate CiM.

Direct engagement of BAF complex by Nanog^BiD

Native NANOG interacts with ~130 proteins including TFs, chromatin-modifying complexes, and basal transcriptional machinery members through classic protein-protein interactions or PPIs³⁰. By fusing BiD to Nanog, we expect a more direct interaction with BAF. To this end, we collected MEF cells undergoing reprogramming with wild type or synthetic Nanog in combination with Oct4 for 1 day and performed IP-MS experiments (Fig. 2a). Candidate Nanog^BiD partners were selected based on detection in at least two independent MS experiments. Initially, we employed Log2 > 1.2 and p-value < 0.05 as the criteria to determine the presence of positive proteins detected by IP-MS. Only positive proteins presented in both replicates of the same treatments would be considered as candidates for the next differential analysis between Nanog^BiD and Nanog^WT. This analysis eventually identified 19 high-confidence interacting partners for Nanog^BiD and 2 for Nanog^WT, and 531 shared by both. Among the 19 partners for Nanog^BiD, 9 are BAF subunits, SMARCA4/BRG1, SMARCC1/2, SMARCD1/3, SMARCE1, ACTL6A, ARID1A/1B (Fig. 2b). However, we did not detect PBAF or ncBAF specific subunits such as BRD7/9, PHF10, GLTSCR1, or PBRM1 as expected as SS18 is not involved in PBAF complex assembly. Those results suggest that Nanog^BiD directly engages cBAF to facilitate chromatin opening. The interaction between Nanog^BiD and BAF subunits was further confirmed by Co-IP experiments (Fig. 2c). Recent report suggests that the interaction between SS18 and BRG1 is largely disrupted when residues A65, L54, and I32 are mutated to glutamic acid, referred as 3M³¹. We then tested this by mutating them in Nanog^BiD and showed that the resulting mutant became ineffective in reprogramming (Fig. 2d, e; Supplementary Fig. 2a, b). These results suggest that the directed engagement of cBAF by Nanog^BiD is critical for cell fate conversion in iPSC generation.

a Proteins of reprogramming samples at day 1 performed in two replicates for FLAG pull-down followed by MS analysis. b Venn diagram shows numbers of protein common and specific identified in Nanog^WT or Nanog^BiD associated reprogramming samples. Proteins are listed as official gene symbols. c Flag Co-IP followed by western blot analysis of SMARCC1, SMARCA4, FLAG, and NANOG. Blots shown are representative of n = 2 independent experiments. d Percentage of homogeneous Oct4-GFP positive colonies induced in Nanog^BiD, Nanog^BiD(3M), and Nanog^WT mediated reprogramming at different time points. Data are presented as mean ± SD for n = 3 biological replicates. Statistical significance relative to colonies number of Nanog^WT at Day 12 was measured with two-tailed unpaired t-test and precise p values are shown on graph. e Representative images of different reprogramming strategies induced colonies on day 12. Scale bar, 5 mm. Source data are provided as a Source Data file.

Full size image

Nanog^BiD-BAF complex accelerates chromatin opening

A previous study has documented that the OKS-Brg1–Baf155 combination yielded 12 times more induced pluripotent stem cells (iPSCs) compared to the OKS combination. This was attributed to the OKS-Brg1–Baf155 combination’s ability to induce a euchromatic chromatin state and facilitate the binding of reprogramming factors to key gene promoters²³. Given the observed physical association between the BAF complex and Nanog^BiD, we then investigated whether cBAF expedites Nanog^BiD reprogramming by deconstructing compact somatic chromatin and establishing an open chromatin state that is conducive to pluripotency. To this end, we initially analyzed the transcriptional changes in Nanog^BiD and Nanog^WT reprogramming by harvesting cells at various time points (day 0, day 1, day 3, day 5, day 7, day 8, day 10, and day 12), along with MEFs and mESCs, as control, for RNA-sequencing. Based on the transcriptional dynamics, all the differentially expressed genes can be categorized into 12 different clusters (Fig. 3a). Notably, three patterns emerge to depict significant dynamics during the reprogramming process and obvious difference between Nanog^BiD and Nanog^WT (Fig. 3b). Principal component analysis shows that there is accelerated transition from somatic state to pluripotent state with Nanog^BiD compared to Nanog^WT (Supplementary Fig. 3a). Particularly, on day 5 of the reprogramming process, discernible divergence in gene expression between Nanog^BiD and Nanog^WT emerged, subsequently exhibiting a progressive expansion. Similar correlations can be observed among specific genes, such as pluripotent genes Sall4 and Sox2, or the somatic specific genes Dab2 and Thbd (Supplementary Fig. 3b). As such, we picked day 5 as the time point for CUT&Tag experiment (Supplementary Fig. 3k). Gene ontology analysis of the genes associated with the cluster 1 revealed genes activated by Nanog^BiD are associated with stem cell differentiation, embryonic development, and pattern specification, whereas the genes associated with the cluster 12 revealed genes repressed by Nanog^BiD are related to extracellular structure organization, and cellular response to growth factor stimulus (Fig. 3c, d).

a Heatmap of the expression dynamics during Nanog^WT and Nanog^BiD reprogramming path as well as MEF and mESC for all differentially expressed genes. These genes were grouped into 12 clusters according to their similarity of expression dynamics. D indicates day. b Line plots show the expression dynamics trends of different gene clusters defined in (a). c, d Functional enrichment analysis of the C1 and C12 genes, respectively. e Heatmap of the chromatin accessibility dynamics during Nanog^WT and Nanog^BiD reprogramming path as well as MEF and mESC. Loci of open chromatin were arranged into groups depending upon the day of Nanog^BiD reprogramming they changed from closed to open (CO) or open to closed (OC). f Motif of transcription factors significantly enriched in each group of chromatin loci of Nanog^BiD reprogramming path (as defined in (e)) and Nanog^WT reprogramming path (as defined in Fig. S3e.). Motifs with at least 3-fold enrichment and less than 1 × 10⁻⁵ p-value are marked with asterisk. g Bar plots show the number of peaks specifically affected by Nanog^BiD (Failed) and both affected by Nanog^BiD and Nanog^WT. h, i The motif enrichments of pluripotency factors and somatic factors in the Nanog^BiD specifically affected CO and OC regions, respectively. The dots were colored by the percentage of each motif in targets. j Functional enrichment analysis of Nanog^BiD specifically affected CO regions demonstrated in heatmap. k Genome track view of the ATAC-seq data for the pluripotency gene cluster miR302a-d. l Venn diagram shows the overlapped gene number of C1 and Nanog^BiD specifically affected CO genes. The 43 transcription factors are listed. m Functional enrichment analysis of the overlapped genes in (l).

Full size image

Given the clear role of BAF complex in reprogramming, we wish to resolve how this complex regulates chromatin dynamics in Nanog^BiD reprogramming by ATAC-seq (Fig. 3e). All the performed samples were collected as same time points as RNA-seq and the close to open (CO) and open to close (OC) peaks were divided into 9 subgroups (Supplementary Fig. 3c). Counting peaks further shows that the number of peaks from OC1-9 has not significant difference, but the number of total peaks from CO1-CO8 in Nanog^BiD (5674) are two times more than Nanog^WT (2735), suggesting that BAF complex may regulate CO but not OC. Besides, more than 80,000 peaks (CO9) remain at closed state at day 12 but open in ESCs. Nearly 40,000 peaks (OC1) are open in MEFs and closed in day 0 (Supplementary Fig. 3d). Later analysis demonstrates that BAF complex regulates chromatin accessibility at early stage and neither CO9 nor OC1 are not the key chromatin loci that determine the successful reprogramming. If these two subgroups were to be included in the heatmap, additional subgroups would be over-compressed and not visible, thus, not included (Fig. 3e).

To gain mechanistic insight into chromatin dynamics, we perform motif analysis for TFs associated with OC and CO peaks (Fig. 3e, Supplementary Fig. 3e). We show that OC loci are enriched with somatic TFs such as TEAD, ATF, and AP-1 family TFs in both systems as expected. However, motifs for SOX and RFX family TFs, such as SOX15, SOX2, SOX17, and RFX2 are enriched in CO peaks at early stage in Nanog^BiD but not Nanog^WT (Fig. 3f). These results suggest that Nanog^BiD recruit BAF complex to open pluripotent chromatin loci. To further identify the differences, we focus on peaks from OC2-9 and CO1-8 in both Nanog^BiD and Nanog^WT and calculated those loci that failed to close and failed to open (Fig. 3g). More than 30% of early CO peaks are sensitive to Nanog^BiD but failed to open in Nanog^WT. Importantly, less than 10% of early OC is different between two systems, suggesting that the primary role of Nanog^BiD—BAF complex is for chromatin opening.

As expected, Nanog^BiD-sensitive CO peaks are dominated by motifs for TFs such as SOXs, and OC peaks enriched for somatic TFs of BACH1/2, FOS, FRA, and ATFs (Fig. 3h, i, Supplementary Fig. 3f). Consistently, CO and OC peaks are associated with loci for mESC and MEF specific enhancers, respectively (Supplementary Fig. 3g). At the chromatin accessibility dynamics level, Nanog^BiD also exhibits quicker and more similar to pluripotency than Nanog^WT (Supplementary Fig. 3h). To validate above analysis, we expressed SOX family TFs with Nanog+Oct4 for reprogramming and show that Sox15, Sox3, and Sox1 significantly promote iPSCs generation (Supplementary Fig. 3i). Consistent with motif enrichment results, gene ontology of Nanog^BiD-sensitive CO1-CO5 show that Nanog^BiD appears to divert cell fate towards to stem cell such as leukemia inhibitory factor (LIF) activation, G1/S transition, and stem cell maturation (Fig. 3j). Among these GO terms, miR-302/367 family responds to LIF signaling and fails to open in Nanog^WT, but promotes iPSCs generation with Nanog+Oct4 as reported (Fig. 3k, Supplementary Fig. 3j). Furthermore, integrating analysis with RNA-seq, we identify 211 genes between cluster 1 and the Nanog^BiD-sensitive CO regions (Fig. 3l). As expected, these genes are related to maintenance of pluripotent stem cell (Fig. 3m). Among them, we can identify 43 TFs, and show that at least Sox15, Foxb1, Sall4, Klf4, Sox2 and Olig3, can facilitate Nanog+Oct4 reprogramming (Supplementary Fig. 3j). Of note, we performed CUT&Tag for Flag-tagged Nanog^BiD and Nanog^WT reprogramming at day 5 using anti-FLAG/BRG1/H3K27ac antibodies and showed that they are more significantly enriched in Nanog^BiD than Nanog^WT (Supplementary Fig. 3k). Together, these results suggest that Nanog^BiD accelerates opening of pluripotent specific chromatin loci.

Nanog^BiD targets BRG1 to pluripotent loci

We then wish to probe the mechanism through which Nanog^BiD promotes chromatin opening by performing CUT&Tag experiments on day 5 reprogramming MEFs for Nanog^BiD and Nanog^WT, BRG1, H3K27ac, and H3K4me1. Quantification of Flag-tagged NANOG signal indicates that Nanog^BiD overlaps significantly with Nanog^WT (Fig. 4a), suggesting that grafting BiD onto Nanog does not change DNA binding specificity of NANOG. In contrast, we show that BRG1 binding sites are doubled in Nanog^BiD compared to Nanog^WT, validating the utility of this engineered factor to open chromatin loci more efficiently (Fig. 4b). We also show higher correlation and more common peaks in Nanog^BiD than that of Nanog^WT (Supplementary Fig. 4a, b). These results reveal that Nanog^BiD targets BRG1 to open closed chromatin.

a, b Venn diagrams show the specific and overlapped NANOG or BRG1 CUT&Tag peak numbers at D5 of the Nanog^WT and Nanog^BiD. c Heatmap of the NANOG and BRG1 binding loci at D5 in Nanog^WT and Nanog^BiD. d, e Scatter plot shows the relationship of NANOG or BRG1 binding change in the specific and overlapped regions defined in (a) promoter regions and enhancer regions, respectively. The Spearman correlation coefficients were calculated for each type of region. f Boxplot of the ATAC-seq signal at each stage of two reprogramming paths in 0101 regions. g Heatmap of the NANOG, BRG1, H3K27ac, and H3K4me1 signal at D5 of two reprogramming paths in 0101 regions. h Boxplot of the H3K27ac and H3K4me1 signal at D5 of two reprogramming paths in 0101 regions. i Enrichment analysis of four classes of regions (0100, 0101, 1100, 1101) in the enhancers and promoters. The dots were colored by the percentage of each motif in targets. j Percentage of specific open loci in the above-mentioned four classes of regions at D3, D5, and D7 during reprogramming. k Heatmap of enhancers regions defined by H3K27ac and H3K4me1 in Nanog^WT and Nanog^BiD. l The NANOG and BRG1 signal in the specific and common regions defined in (k). The box plots (f, h, j, l) indicate the medians (centerlines), first and third quartiles (bounds of boxes), and 1.5 multiplied by interquartile range (whiskers). Statistical analysis was performed using student’s two-sided t-test. ***p < 0.0001, NS nonsignificant. m Genome track view of the ATAC-seq and CUT&Tag data for the C1 genes which have NANOG and BRG1 Co-binding and CO (closed to open) pattern affected by Nanog^BiD. n, o The effects of Sox15, Foxb1, Klf4, Sox2, Sall4, and Olig3, or different combinations in Nanog+Oct4 mediated reprogramming system. Data are presented as mean ± SD (n = 3 biological replicates), p values are determined by two-tailed unpaired t-test. p A model for fibroblasts reprogramming with BiD factors. Precise p values were provided in the Source Data file. WT indicates Nanog^WT. BiD indicates Nanog^BiD.

Full size image

We further classified CUT&Tag peaks into three groups: BiD-specific, WT-specific, and common (Fig. 4c). Among them, BRG1 exhibits distinct distribution with ~35% regions (10,376) occupied by NANOG and BRG1 simultaneously (refer as 0101), ~57% (17,316) by both in a much weaker manner (refer as 0100). We also observed an interesting pattern for BRG1 vs Nanog^WT and Nanog^BiD in the common group, with 14,212 (refer as 1101) regions occupied by BRG1 strongly with NANOG, again suggesting that Nanog^BiD also targets BRG1 to those regions. There are also 6741 regions (refer as 1100) with relatively weak BRG1 binding with Nanog^BiD in the common group. Together, these results appear to suggest that Nanog^BiD targets BRG1 to a much wider portion of the genome compared to Nanog^WT (Fig. 4d). Both BRG1 and NANOG binding changes tend to be at distal regions, i.e., mainly enhancer remodeling (Fig. 4e). Consistently, we show a similar result when we analyzed NANOG binding and H3K27ac changes (Supplementary Fig. 4e). Furthermore, NANOG motif searching from FLAG peaks revealed Nanog^BiD binding regions contain more ESC-specific NANOG motifs (Supplementary Fig. 4c, d). These results suggest that Nanog^BiD and BRG1 have similar genomic binding patterns.

To assess the impact of Nanog^BiD and BRG1 co-occupancy described above, we analyzed the ATAC-seq signals in defined regions. As shown in Fig. 4f, Nanog^BiD occupancy correlates positively with greater chromatin accessibility in the 0101 regions. A similar relationship could also be found in the 0100 regions (Supplementary Fig. 4f). Consistent with chromatin accessibility, H3K27ac and H3K4me1, both marking enhancers, are enriched more in 0101 and 0100 areas in Nanog^BiD than that of Nanog^WT (Fig. 4g, h; Supplementary Fig. 4g, h). In contrast, there is no such positive correlation between ATAC-seq and Nanog^BiD or Nanog^WT in the common group from day 3 to day 12, but an evidently negative one (Supplementary Fig. 4i, l). Similar relationships are observed for H3K27ac and H3K4me1 modifications in similar regions (Supplementary Fig. 4j, k, m, n). For these four regions, we extracted chromatin state annotations for enhancer and promoter and showed that both NANOG and BRG1 modulate epigenetic modification in enhancer, but not promoter regions (Fig. 4i). Of note, both 0101 and 1101 have higher enrichment scores than their counterpart 0100 and 1100, suggesting that BRG1 occupancy promotes enhancer activation.

To investigate the relationship between BRG1 occupancy and chromatin accessibility and histone modification, we analyzed BRG1 binding at regions with different ATAC-seq signals and histone modifications. First, the day 5 ATAC-seq samples were calculated and Nanog^BiD and Nanog^WT specific regions were defined. Then, we computed the numbers of overlap regions between ATAC-seq and CUT&Tag. The percentage of overlap regions is calculated by counting overlap peaks in CUT&Tag regions (Fig. 4j). The distinct ratio between 0101 and 0100 is higher in Nanog^BiD—specific ATAC-seq regions than that of Nanog^WT—specific regions indicating that BRG1 binding increases chromatin accessibility. Then, Nanog^BiD and Nanog^WT—specific enhancers were analyzed with H3K27ac and H3K4me1 modification to show that the levels of BRG1 and NANOG binding correlate well with epigenetic modifications at active enhancers (Fig. 4k, l).

The genome-wide correlation analyses described above can also be validated at specific gene loci, revealing specific genes targeted by Nanog^BiD and BRG1. As shown in Fig. 4m, genes such as Sall4, Esrrb, miR-302, Dppa5a are known to promote iPSC formation^32,33,34, validating the positive role of the Nanog^BiD and BRG1 interaction. We then tested select candidates and showed that they indeed promote iPSC generation (Fig. 4n, o; Supplementary Fig. 3i, j), including those not previously known such as Sox15, Foxb1, and Olig3. It is likely that the sum of these downstream targets may account for the efficiency boost from the engineered factor Nanog^BiD.

• SEO Powered Content & PR Distribution. Get Amplified Today.
• PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
• PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
• PlatoESG. Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
• PlatoHealth. Biotech and Clinical Trials Intelligence. Access Here.
• Source: https://www.nature.com/articles/s41467-024-50551-2