Epigenetic Modulation of Short Interspersed Nuclear Elements Activity Influences Neural Precursor Cell Proliferation

doi:10.21203/rs.3.rs-3647230/v1

Download PDF

Article

Epigenetic Modulation of Short Interspersed Nuclear Elements Activity Influences Neural Precursor Cell Proliferation

https://doi.org/10.21203/rs.3.rs-3647230/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Transposable elements (TEs) play a critical role in modulating gene expression during neurodevelopment. Short Interspersed Nuclear Elements (SINEs), a significant subset of TEs, contribute to gene regulation by generating non-coding transcripts and functioning as enhancers. Moreover, SINEs harbor binding sites for the CCCTC-binding factor (CTCF), pivotal in orchestrating chromatin organization. However, the exact function of SINEs in neurodevelopment remains elusive. In our study, we conducted a comprehensive genome-wide analysis using ATAC-seq, ChIP-seq, WGBS, in situ Hi-C, and RNA-seq. We elucidated the intricate epigenetic regulations governing a relatively conserved subset of SINEs in mouse neural precursor cells (NPCs). Our findings revealed that the SET domain bifurcated histone lysine methyltransferase 1 (SETDB1) orchestrates H3K9me3, in conjunction with DNA methylation, to restrict SINEs' chromatin accessibility in NPCs. Loss of SETDB1 granted CTCF access to previously restricted SINE elements, facilitating novel chromatin loop formation and influencing cell cycle genes. Disruptions in cell proliferation were notably observed both in vitro and in vivo following genetic ablation of SETDB1 in NPCs. In summary, our study sheds light on the comprehensive epigenetic regulation of SINEs, suggesting their role in maintaining chromatin integrity and stemness in NPCs.

Biological sciences/Neuroscience/Epigenetics in the nervous system

Biological sciences/Molecular biology/Epigenetics

Transposable elements

SETDB1

H3K9me3

DNA methylation

chromatin interactions

neurodevelopment

Transposable Elements (TEs) make up a substantial portion of the mammalian genome, comprising approximately 40–50% in both human and mouse genomes^1,2. Their historical association with genetic diversity, evolutionary changes, and potential regulatory roles has been extensively documented^3–6. Meanwhile, their roles in the central nervous system, particularly their impact on the diversity of neurons, neuronal communication, and brain development and disorders, represent an essential area of ongoing research^7–9.

Short Interspersed Nuclear Elements (SINEs) represent a distinctive subclass among TEs due to their relatively concise size, typically spanning just a few hundred base pairs¹⁰. While SINEs themselves do not encode proteins, they play a role in transcribing functional RNAs¹¹ or acting as enhancers¹¹, delicately fine-tuning the intricate genetic expression vital for neurogenesis, neuronal maturation, and the establishment of neural networks^12,13. SINEs have been implicated in neuronal plasticity¹⁴, actively contributing to the brain's adaptive responses in experiences, learning and memory processes¹⁴. For instance, the B2 subset, a distinct group within SINEs, is prevalent in the genome and widely distributed in gene-rich regions¹⁵. Studies have indicated that noncoding SINE_B2 RNA can modulate gene expression under stressful conditions^16,17. Additionally, B2 transcription is activated in the mature hippocampus in response to novel stimuli¹⁴. Nevertheless, our understanding of SINEs' role in the developing brain and their potential impact on neurodevelopment remains limited.

Epigenetic regulation of TEs plays a pivotal role in governing neurodevelopment, contributing to the establishment of a sophisticated framework for managing precise spatiotemporal gene expression patterns essential in the intricate process of neurodevelopment^13,18,19. Notably, SINEs significantly contribute by serving as repositories of various epigenetic marks, including DNA methylation and histone modifications, exerting substantial influence on local chromatin configurations. DNA methylation patterns within SINEs wield significant influence over neighboring gene accessibility, modulating their expression and contributing significantly to vital developmental processes²⁰. An integral repressive epigenetic mark, H3K9me3, holds particular importance in TE regulation during evolution. Although its specific role in SINE regulation within the brain remains elusive, existing reports suggest its involvement in silencing SINEs in microphages and influencing gene expression associated with inflammatory responses²¹. Adding another layer to their epigenetic influence, SINEs maintain the capacity to act as binding sites for CCCTC-binding factor (CTCF) and participate in chromatin folding, and this process is under the regulation of chromatin remodeling complex ChAHP (CHD4/ADNP/HP1) in mouse embryonic stem cells (ESCs)^22,23. These features potentially impact the establishment and maintenance of the critical epigenetic landscape required for proper brain development.

SET Domain Bifurcated 1 (SETDB1), a histone methyltransferase targeting H3K9me3, plays a pivotal evolutionary role in TEs^24,25. Researches, including ours, underscore its critical involvement in brain development by regulating endogenous retroviruses (ERVs)^19,26. Analogous to SINEs' influence on gene expression, the absence of SETDB1 in neural precursor cells derepresses ERVs, impacting adjacent gene transcription²⁶, or serving as enhancer elements¹⁹. Beyond its role as an H3K9me3 ‘writer’, SETDB1 orchestrates 3D chromatin conformation within the brain²⁷. Our previous study revealed its ability to maintain the topologically associating domain (TAD) conformation covering the protocadherin gene cluster, through H3K9me3 deposition, DNA methylation, and long-range repressive chromatin interactions²⁸. Notably, SETDB1 loss in mature neurons increases genome-wide CTCF binding²⁸, guiding our ongoing study on SINEs' epigenetic regulation.

Our current study investigated the pivotal role of SETDB1-mediated H3K9me3, complemented by DNA methylation, in regulating a specific subset of SINE B2 elements. Despite their ancient origin, these elements exhibit functional conservation in evolution, particularly enriched in CTCF binding motifs. Our extensive epigenomic profiling, including ATAC-seq, anti-H3K9me3 ChIP-seq, WGBS, anti-CTCF ChIP-seq, in situ Hi-C, and RNA-seq analyses, deciphered intricate epigenetic mechanisms governing these elements in mouse NPCs. Our research also sheds light on their impact on gene expression patterns through CTCF-mediated higher-order chromatin organizations, further enhancing our understanding of the regulation of the cell proliferation of NPCs.

SETDB1 selectively inhibited the chromatin accessibility on SINE_B2 elements

To investigate the regulatory mechanisms of TEs in the developing brain by SETDB1, we dissected out ganglionic eminences (GEs) from both wildtype (WT) and SETDB1-knockout (KO) mouse brains at embryonic day 15.5 (E15.5), and conducted neurosphere culture following an established protocol¹⁹. The purified NPCs were then harvested and subjected to various epigenomic profiling assays, including the evaluation of genome-wide chromatin accessibility (ATAC-seq), H3K9me3 and CTCF occupancy (ChIP-seq), DNA methylation (WGBS), 3D genome organization (in situ Hi-C), and gene transcription (RNA-seq) (Fig. 1A). It is worth noting that ATAC-seq and RNA-seq data were derived from our previously published work¹⁹. However, in this study, we undertook a novel analysis of these datasets from a completely different perspective, shifting our emphasis onto the investigation of TEs instead of individual gene loci.

Differential analysis (KO/WT) of ATAC-seq peaks revealed genome-wide alterations in chromatin accessibility, resulting in 15,868 up-regulated and 9,868 down-regulated peaks (Fig. 1B). Notably, a significant finding was that more than 65% (10,713) of the up-regulated ATAC-seq peaks (“ATAC_up”) were located on TEs. Furthermore, nearly half of these peaks (4,556) were specifically annotated to the B2 family of SINE class (SINE_B2). In contrast, less than 25% of the down-regulated ATAC peaks (2,442) were associated with TEs, and only a limited number of 88 peaks were annotated as SINE_B2 (Fig. 1B, S1A). Further subtype analysis of the overlapping TEs revealed that only SINE and LTR showed significant enrichment within the ATAC_up peaks (Fig. 1C-D). The enrichment on LTR was expected, as previous studies have reported that SETDB1 plays critical roles in repressing LTRs, particularly ERVs, in the brain and other organs^{19,25,29–31}. In our study, within the LTR class, ERVKs were notably enriched in ATAC_up peaks (Fig. 1C). However, the robust enrichment within the SINE class was unexpected. Among the SINE class, we found that B2, as opposed to Alu and other subclasses, exhibited highly enrichment in the ATAC_up peaks. Moreover, among the five B2 subtypes (B3, B3A, B2_Mm2, B2_Mm1t, B2_Mm1a), B3 and B3A were the main contributors to the observed effects (Fig. 1C-D). Considering the relatively short length of SINE (average length of 159 bp) compared to other types of TEs, we recalculated the enrichment by accumulative length and obtained consistent results (Fig. S1B-C).

We defined the group of B2 elements that overlapped ATAC_up peaks as “ATAC_up_B2”. When compared to all B2 elements in the genome (“All_B2”), and the group of B2 that did not overlap with ATAC_up peaks (“ATAC_not_up_B2”), we observed a notable increase in chromatin accessibility for ATAC_up_B2, even in the WT, which was further augmented after SETDB1-KO (Fig. 1E). This suggested that this specific set of B2 elements exhibits a baseline level of activity and potential functionality. Genome annotation analysis revealed that ATAC_up_B2, as well as ATAC_up_B3 and ATAC_up_B3A, were significantly enriched in proximal intergenic regions and gene bodies (Fig. 1F). Interestingly, although these B2 elements were not enriched in promoters (Fig. 1F), they were significantly closer to transcription start sites (TSS) when compared to the distribution of B2 elements throughout genome (Fig. S1D). In contrast, ATAC_not_up_B2 elements were enriched in distal intergenic regions (Fig. 1F), and located significantly farther from TSS (Fig. S1D). Visualizations on the map tracks clearly demonstrated that ATAC_up_B2 elements were highly concentrated in gene-rich regions (Fig. 1G). These all suggest a high potential of ATAC_up_B2 elements to influence gene transcription. Furthermore, we conducted Homer motif analysis and revealed that only ATAC_up_B2 elements were highly enriched with motifs for CTCF, in contrast to ATAC_not_up_B2 or All_B2 (Fig. S1E), underscoring the distinct functional role of ATAC_up_B2 elements. In summary, these findings indicate that the chromatin accessibility of a specific population of SINE_B2 elements is controlled by SETDB1 in NPCs.

SETDB1 suppressed chromatin accessibility of SINE_B2 via H3K9me3 deposition

SETDB1 is one of the most important histone methyltransferases with specificity towards H3K9me3. We, therefore, investigated whether SETDB1 represses the chromatin accessibility on SINE_B2 elements in NPC through direct H3K9me3 deposition. To explore this, we conducted H3K9me3 ChIP-seq in NPCs from both WT and SETDB1-KO. The correlation analysis showed a clear distinction between WT and KO samples (Fig. 2A). Subsequently, we called H3K9me3 narrow peaks and performed a differential analysis between KO and WT, which disclosed a substantial downregulation of H3K9me3 narrow peaks (8,063) (“K9_down”) in KO, with only 6 peaks showing upregulation (Fig. 2B). It is worth noting that H3K9me3 signals predominantly manifest as broad domains in heterochromatic regions, whereas the identified narrow peaks typically represent localized H3K9me3 signals in open chromatin regions. Indeed, a majority (62.2%) of these K9_down peaks were located in genic regions (Fig. S2A). Consistent with the ATAC-seq data (Fig. 1), we observed enrichments of SINE (by TE class), B2 (by SINE family) and B3/B3A (by B2 subtype) elements within the K9_down peaks (Fig. 2C-D, S2B). While LTR elements constituted a substantial portion of the K9_down peaks, they did not reach statistical significance in the enrichment tests (Fig. 2C-D). In addition, when considering enrichment by cumulative length, SINE_B2 remained highly enriched in K9_down peaks, while LTR displayed a milder level of enrichment (Fig. S2C). Furthermore, we detected enrichment of the CTCF motif in the K9_down peaks (Fig. S2D).

Next, we assessed the signal correlation between differential peaks from the ATAC-seq and H3K9me3 ChIP-seq datasets. The H3K9me3 signal was almost completely erased on ATAC_up_B2 elements in KO compared to WT (Fig. 2E), and vice versa, the ATAC-seq signal was increased on K9_down_B2 elements (Fig. 2F). These observations were visually evident in the H3K9me3 track maps, particularly on the ATAC_up_B2 elements (Fig. 2G). In summary, these findings strongly suggested that genetic ablation of SETDB1 in NPCs results in the loss of H3K9me3 signal in the euchromatin regions, which likely contributes to the increased chromatin accessibility observed on the SINE_B2 elements.

DNA methylation participated in SETDB1-mediated repression of SINE_B2

Our previous work demonstrated the interplay between SETDB1-mediated H3K9me3 and DNA methylation, another well-established repressive epigenetic modification²⁸. Given this, in the current study, we examined DNA methylation by conducting WGBS in NPCs derived from both WT and KO. The high efficiency of bisulfite conversion was confirmed using spiked-in lambda DNA (Fig. S3A). While the global DNA methylation levels showed only mild changes (Fig. S3B), differential analysis (KO/WT) revealed over 30,000 significant differentially methylated regions (DMRs) across the genome (Fig. 3A), evenly distributed on promoters, gene bodies and distal intergenic regions (Fig. S3C). As expected, most TEs exhibited high levels of DNA methylation, while non-TE regions, especially gene promoters, displayed lower levels of methylation (Fig. 3B, left panel). Meanwhile, the average methylation level on SINE and its subfamilies including B2 remained largely unchanged (KO/WT) at a global level (Fig. 3B, left panel). However, when examining the significant DMRs, we observed massive downregulation (KO/WT) of DNA methylation on TEs, with the exception of low-complexity regions. Among all TEs, the significant DMRs in SINE were the most hypomethylated, with B2 being the most hypomethylated within the SINE class (Fig. 3B, right panel). Interestingly, gene promoters, but not those active promoters, exhibited increased DNA methylation in KO compared to WT (Fig. 3B, right panel). Moreover, enrichment analysis, both by TE number and accumulative length, revealed that significantly hypomethylated DMRs (“DMR_hypo”) were enriched for SINE (by TE class) and B2 (by SINE family), but not for LTR (Fig. 3C-D, S3D-E). Interestingly, within the subtypes of B2, DMR_hypo were more enriched for Mm1t and Mm1a (Fig. S3D-E), contrasting with B3/B3A that were enriched for the up-regulated ATAC-seq peaks (ATAC_up) (Fig. 1C-D), and down-regulated H3K9me3 narrow peaks (K9_down) (Fig. 2C-D). Nevertheless, the CTCF motif was also enriched in DMR_hypo regions (Fig. S3G).

Next, we evaluated the intersection between DMRs and ATAC-seq peaks (ATAC). We found that for significant DMRs and ATAC peaks overlapping B2 elements (sig_B2), a substantial fraction (514/2497) was situated in the 1st quadrant (DMR_hypo and ATAC_up), suggesting an increase in chromatin accessibility in the DNA hypomethylation regions on B2 elements in KO (Fig. 3E). Notably, there was a near complete depletion of the H3K9me3 signal in this set of B2 elements (Fig. 3F-G). Altogether, these findings imply that DNA methylation and H3K9me3 act synergistically on SETDB1-mediated suppression of B2 elements in NPCs.

SETDB1 constrained excessive CTCF binding at SINE_B2

SINE_B2 elements are well known for their potential to harbor CTCF binding sites^21,22. In our current study, we discovered an enrichment of CTCF motifs in B2 elements with ATAC_up, K9_down, and DMR_hypo events (Fig. S1E, S2D & S3F). Therefore, we performed anti-CTCF ChIP-seq to characterize the CTCF binding landscape in both WT and KO. Correlation analysis showed a clear distinction between WT & KO samples (Fig. S4A). Moreover, differential analysis (KO/WT) revealed a substantial global increase in CTCF binding, with 8,597 significantly up-regulated and 615 down-regulated CTCF peaks in KO (Fig. 4A). Remarkably, more than 70% of these upregulated CTCF peaks (“CTCF_up”) overlapped with SINE_B2 elements (Fig. 4B), and showed exclusive enrichment (both in terms of TE number and cumulative length) for SINE (by TE class), B2 (by SINE family), and B3A/ B3 (by B2 subtype) elements (Fig. 4B, S4B-D). Notably, despite multiple evidence indicating the activation of LTR elements in KO (Fig. 1C-D, S2B-C), LTR was not enriched in the CTCF_up peaks when calculated by TE number (Fig. S4B) and showed only very mild enrichment when calculated by cumulative length (Fig. S4D). Furthermore, CTCF motifs were only enriched in the CTCF_up peaks overlapped with B2 elements (“CTCF_up_B2”), not in those that did not overlap with B2 (Fig. S4E). These findings highlight the specificity of the increased CTCF binding towards SINE_B2 elements following the loss of SETDB1 in NPCs.

We proceeded to conduct an integrated analysis of the CTCF ChIP-seq and ATAC-seq datasets. Permutation tests revealed a significantly higher number of overlaps and a shorter average distance between CTCF_up peaks and ATAC_up_B2, relative to the background of all B2 loci (Fig. S4F). Approximately half of the CTCF_up peaks overlapped with ATAC_up peaks, and over 70% (3,429/4,683) of ATAC_up_B2 peaks were associated with CTCF_up peaks (Fig. S4G). Furthermore, the distribution plot showed that almost all the B2 elements (4,005/4,007) with significant changes in ATAC-seq and CTCF ChIP-seq exhibited increases of both signals in KO compared to WT (Fig. S4H). This suggests that the increase in CTCF bindings in KO may be a result of the alleviation of chromatin accessibility repression on B2 elements. Following these observations, we looked deeper into the accumulative signal of CTCF, ATAC-seq, H3K9me3, and DNA methylation on CTCF_up_B2 elements in both WT and KO (Fig. 4C-G). These B2 elements exhibited moderate enrichment of CTCF occupancy in WT, which increased in KO as anticipated (Fig. 4C). In parallel, chromatin accessibility mirrored this trend (Fig. 4D). Interestingly, the repressive epigenetic marks showed a similar trend but exhibited distinct patterns. For H3K9me3, a notable enrichment (“peak”) was observed on CTCF_up_B2 in WT, which was almost completely diminished in KO (Fig. 4E). In contrast, DNA methylation was lower than the background (“valley”) on CTCF_up_B2 in WT, and this was further reduced in KO (Fig. 4F). These findings suggest that H3K9me3 and DNA methylation play together, but with unique ways in controlling the activity of this population of SINE_B2 elements.

CTCF is a highly conserved zinc finger protein and plays an important role in the regulation of gene transcription^32–34. CTCF occupancy is broadly distributed across the genome. Density plot illustrated that a considerable proportion of CTCF peaks were located on gene promoters (around TSS), while the remaining peaks were concentrated around 15 Kb away from the TSS (Fig. 4H). Interestingly, upregulated CTCF peaks (CTCF_up, “Up” for short), as well as those on the B2 elements (CTCF_up_B2, “Up_B2” for short), were almost completely absent from the promoter regions (Fig. 4H). Further genome annotation revealed that these CTCF_up peaks were primarily enriched in genic regions. Particularly, CTCF_up_B2 showed a significant enrichment in gene body regions (Fig. S4I), suggesting an indirect regulatory role of these peaks in gene expression.

Chromatin loop formation due to increased CTCF occupancy at SINE_B2 elements

In recent years, CTCF has been extensively studied for its pivotal role in maintaining 3D genome architecture through the mediation of chromatin folding^32–34. Therefore, we explored the consequences of a genome-wide increase in CTCF binding on chromatin organization in both WT and KO NPCs. We performed in situ Hi-C and generated Hi-C contact maps with a resolution of < 15 Kb for individual samples, and < 5 Kb for merged samples by genotype (Fig. S5A). There were no significant alterations in normalized contact maps (Fig. S5B) or genome-wide relative contact probability between WT and KO (Fig. S5C). Moreover, although CTCF has been recognized as an insulator in the formation of boundaries for TADs^35,36, our data did not reveal any significant changes in the number of TADs, the width, the insulation score of TAD boundaries, or the interaction frequency between neighboring TADs (Fig. 5A-B & S5D).

We then turned our focus to chromatin loops. While the number and span of loops remained largely unchanged (Fig. S5E-F), we observed significant variations in both the strength and position of chromatin loops (Fig. 5C-D). The differential (KO/WT) contact intensity decreased noticeably in WT loops, while it increased in KO loops (Fig. 5C). Additionally, we detected a significant redistribution of loop anchors, resulting in the emergence of 6,444 “new loop anchors” (NLA), the disappearance of 5,515 “lost loop anchors” (LLA) in KO, and the conservation of 10,240 “unchanged loop anchors” (ULA) between WT and KO (Fig. 5D). Notably, a significant portion of NLAs (p < 0.0001) overlapped with CTCF_up_B2 elements (Fig. S5G). This overlap exceeded expectations (Fig. 5E, left), while the overlaps between LLA (Fig. 5E, middle) and CTCF_up_B2, or ULA and CTCF_up_B2 (Fig. 5E, right), significantly fell below anticipated levels. Visualization of the chromatin loops from WT and KO showed the emergence of new loop anchors corresponding to CTCF_up_B2 elements (Fig. 5F).

In conclusion, these findings suggest that while increased CTCF binding did not cause major structural alterations in the 3D genome, it led to a reorganization of chromatin interactions in the SETDB1-KO NPCs.

SETDB1 maintained cell cycle in NPCs by restraining aberrant loop formation

Chromatin loops, which bring cis-regulatory elements into close proximity with gene promoters, are known to regulate gene transcription. To understand the transcriptional implications of reorganized chromatin loop formation following SETDB1 loss in NPCs, we analyzed RNA-seq data. Interestingly, we found that the promoters of differentially expressed genes (DEGs) significantly overlapped with the anchors of new loops (those with at least one NLA), but not with lost loops (those with at least one LLA) or unchanged loops (Fig. 6A), indicating that these novel chromatin loops might influence gene transcription. Despite a genome-wide increase in chromatin accessibility in the SETDB1-knockout NPCs, particularly on TEs (Fig. 1), we observed a decrease in the ATAC-seq signal on all gene promoters in the KO (Fig. S6A). This finding was unexpected, given SETDB1's known role as a transcriptional repressor. Detailed analysis revealed that the decrease in the ATAC-seq signal primarily occurred on the promoters of downregulated genes, but not on those of upregulated genes (Fig. 6B). In addition, protein-protein interaction analysis of the DEGs specifically overlapping with NLAs on CTCF_up_B2 elements identified enriched functional clusters associated with cell division (Fig. 6C). Indeed, we observed CTCF_up_B2 mediated NLAs on these genes (Fig. S6B). Meanwhile, RNA-seq data indicated a significant downregulation of their transcription levels in KO NPCs (Fig. 6D). This suggests that the newly formed chromatin loops in KO mediated downregulation of gene transcription in NPCs, especially for those genes implicated in the cell cycle.

We subsequently assessed cell proliferation capabilities of both WT and KO NPCs. Our in vitro analysis of primarily cultured neurospheres showed a significant reduction in diameter for the KO group compared to the WT (Fig. 6E). Cell cycle analysis further revealed fewer cells at the G2/M phase in the KO group compared to the WT (Fig. 6F, S6C). For in vivo studies, we administered intraperitoneal injections of BrdU to pregnant mice at E15.5, and then sacrificed them at E16.5. We examined the fetal brains for BrdU labeling and phospho-histone H3 (PH3) expression using immunofluorescence staining. Upon genotyping, we observed a significant decrease in the number of PH3 + and PH3+/BrdU + cells within the ganglionic eminence (GE) region of KO brains (Fig. 6G-I). Therefore, our findings from both in vitro and in vivo studies suggest impaired cell proliferation and dysregulated cell cycle in the NPCs due to the loss of SETDB1.

In this study, we have elucidated that during brain development, SETDB1 selectively represses the accessibility of SINE_B2 elements, particularly the B3 and B3A subtypes. This is primarily achieved through the deposition of H3K9me3, with additional assistance from DNA methylation. In the absence of SETDB1, there is a marked increase in CTCF occupancy at these B2 loci, attributable to the increased availability of CTCF binding sites harbored by B2 elements. This, in turn, mediates the formation of novel chromatin loops. Such structural reorganization contributes to transcriptional dysregulation, with a significant impact on the genes involved in the mitotic cell cycle. Consequently, this disruption leads to an altered cell cycle and diminished proliferation of NPCs.

SETDB1 is well recognized as a master regulator of the LTR class of TEs, particularly as a suppressor of ERV transcription in ESCs^37,38, and in differentiated brain cells, including NPCs and mature neurons^19,25,26. However, less is known about the regulatory roles of SETDB1 on other classes of TEs within the brain. One study has uncovered the regulation of SETDB1 on SINE in bone marrow-derived macrophages²¹. Our study extended its scope into the field of neurodevelopment and uncovered different regulatory mechanisms. We found that SETDB1 is vital for repressing the non-LTR SINE class, particularly the B2 family of SINE. Interestingly, the affected B2 elements, compared to unaffected ones, were enriched with ATAC-seq and CTCF ChIP-seq signals and exhibited lower DNA methylation levels in WT. This suggests that this population of B2 elements exhibits regulatory activity in NPCs at the baseline level. Furthermore, in our NPC system, the loss of SETDB1 only triggered an increase in chromatin accessibility, while the transcription of SINE_B2 remained silenced. These findings suggest that this population of B2 elements plays a role in regulating NPCs via cis-regulation pathways, rather than through trans-regulation avenues.

The interplay between different epigenetic mechanisms, especially the crosstalk between two crucial repressive epigenetic modifications - DNA methylation and H3K9me3 modification - is a well-studied topic. A complex interdependency exists between these two mechanisms; the loss of a critical methyltransferase in either one can impair the stability of the other^39,40. DNA methylation is often perceived as more stable and enduring, while H3K9me3 is considered more dynamic. Previous studies have demonstrated that these two marks can function either synergistically or sequentially to suppress the transcription of genes and ERVs during development^41,42. Conversely, these two marks have been reported to regulate different genes and ERVs in embryonic stem cells (ESCs)⁴³. Our study further elucidates the cooperation between DNA methylation and H3K9me3 in the developing brain. We found that SETDB1-mediated H3K9me3 is critical for maintaining genome-wide DNA methylation homeostasis in NPCs, and both contribute to the repression of SINE_B2. However, this cooperative "double-lock" mechanism may be specific to NPCs, as research in macrophages has ruled out the involvement of DNA methylation in the regulation of SINE_B2²¹.

SINE_B2 elements are classified into five subtypes: B2_Mm1a, B2_Mm1t, B2_Mm2, B3, and B3A, listed from evolutionarily younger to older²². An interesting finding in our study was the divergence in the regulation of SINE_B2 subtypes by DNA methylation and H3K9me3. We discovered that the down-regulation of H3K9me3 was prevalent in the evolutionarily older B3/B3A, while DNA hypomethylation loci were more common in the evolutionarily younger B2 elements, B2_Mm1a and B2_Mm1t. This was unexpected, as DNA methylation, one of the most stable repressive epigenetic mechanisms, was anticipated to be primarily in charge of relatively ancient TEs during evolution. On the other hand, our findings support the notion that the B3/B3A loci, which are selectively controlled by SETDB1-mediated H3K9me3, a reversible histone modification, are more open to regulation. Indeed, our study provided evidence that the loss of SETDB1 selectively opened the B3/B3A loci, enabling CTCF binding. This aligns with previous research indicating that CTCF motifs were enriched in the evolutionarily older B3/B3A, rather than their younger counterparts⁴⁴. Furthermore, a recent research confirmed that B3/B3A are also enriched for CTCF binding sites that mediate conserved chromatin loops between mouse and human⁴⁵. This suggests that these loci are not merely "ancient rubbish DNA" that must be suppressed from activation. Instead, they might be functionally critical, having survived natural selection, evolved into bona fide CTCF binding sites, and contributed to the regulatory flexibility of the genome.

SETDB1, a mammalian histone methyltransferase specific to H3K9me3, is conventionally recognized for its role as a transcriptional repressor. However, our previous study has indicated that its function extends beyond merely facilitating H3K9me3 in gene regulation. In mature neurons, we have demonstrated that SETDB1 influences the transcription of clustered Protocadherin genes by preserving the integrity of the chromatin domain at this locus²⁸. Our current investigation further substantiates this role, providing additional mechanistic insight into the involvement of SETDB1 in regulating higher-order chromatin conformation. We discovered that the absence of SETDB1 in NPCs leads to the emergence of newly accessible SINE_B2 sites. These sites allow for CTCF binding, serving as chromatin loop anchors, and triggering a reorganization of chromatin folding. Furthermore, our results suggest that the formation of chromatin loops primarily contributes to the adaptability of gene expression, rather than directly affecting transcriptional activity. It appears that the ultimate control over changes of gene expression is determined by transcription factors⁴⁶. As neural differentiation progresses, SETDB1 expression declines, which coincides with NPCs exiting the cell cycle and ceasing proliferation²⁶. Our study indicates that certain newly formed chromatin loops could enhance the regulatory flexibility of cell cycle genes, thereby offering more opportunities for repressive transcription factors to inhibit their expression.

Despite the aforementioned insights, limitations exist within our current research. We still lack evidence for causal effects between regulatory elements and gene expression, as our emphasis was laid on genome-wide effects, rather than select single gene or locus. On the other hand, more evidence is needed to clarify the mechanistic conservation between mice and human beings.

Collectively, our data collectively unravel the epigenetic mechanisms regulating a select population of SINE_B2 elements in mouse NPCs. These SINE_B2 elements, under the control of SETDB1-mediated H3K9me3, together with DNA methylation, finely tune CTCF binding and chromatin interactions. Moreover, they participate in regulating the cell cycle of NPCs during brain development. Our findings provide novel insights into the complex mechanisms of epigenetic regulation in neurodevelopment.

Animals

Setdb1^2lox/2lox mice were a gift from Dr. Akbarian, Icahn School of Medicine at Mount Sinai, New York. Nestin-Cre mice were kindly provided by Dr. Mo, Xiamen University. All mice were housed 5 per cage, at 21–26 ℃ under 12/12-hour day/night cycle, and were allowed sterile food and water ad libitum. Mouse usage and experiments were authorized by Animal Care and Use Committee of Shanghai Medical College, Fudan University.

Neurosphere culture

Neurosphere culture of neural precursor cells (NPCs) was conducted according to established protocol¹⁹. Female Setdb1^2lox/2lox mice were crossed with male Setdb1^2lox/+, Nestin-Cre mice and checked for vaginal plug. At gestational day 15.5, mouse fetuses were dissected, genotyped and classified into wildtype (WT, Setdb1^2lox/+ or Setdb1^2lox/2lox) and knockout (KO, Setdb1^2lox/2lox, Nestin-Cre). Ganglionic eminences of fetal brains were harvested, dissociated with Accutase solution (Sigma A6964), and cultured in ultra-low attachment plates (Corning 3471) with advanced DMEM/F-12 (Gibco 12634-010) supplemented with N-2 (Gibco 17502048), B-27 (Gibco 12587010), GlutaMAX (Gibco 35050061), P/S (Sigma V900929), heparin (5 ug/ml), EGF (20 ng/ml) and FGF (20 ng/ml). Fresh medium was added every 3 days, and neurospheres were dissociated with Accutase and passed every 5 days.

ChIP-seq

ChIP was conducted with published protocol^19,28, with minor modifications. ~ 3 million NPCs were harvested, washed in PBS and lysed with NP40 lysis buffer (0.32 M sucrose, 5 mM CaCl₂, 3 mM Mg(Ace)₂, 0.1 mM EDTA, 10 mM Tris pH8, 0.1% NP40). For native ChIP (nChIP, for H3K9me3), nuclei were collected through centrifugation, resuspended in douncing buffer (10 mM Tris pH8, 4 mM MgCl₂, 1 mM CaCl₂) and chromatin was fragmentized with micrococcal nuclease (Sigma N3755). Hypotonic solution (0.2 mM EDTA, cOmplete inhibitor (Roche 11697498001), 0.1% NP40) was added for nuclei lysis and chromatin release. For fixed ChIP (xChIP, for CTCF), nuclei were collected through centrifugation, fixed in 1% formaldehyde (Sigma 252549) at room temperature (RT) for 5 minutes with rotation, and quenched in 0.125 M glycine at RT for 10 minutes with rotation. After collected with centrifugation, the fixed nuclei were resuspended in FSB solution (5 mM EDTA, 20 mM Tris pH8, 500 mM NaCl) supplemented with 0.1% SDS and cOmplete inhibitor, and sonicated with Covaris E220 ultrasonicator system to 300 ~ 500 bp. Next, for both nChIP and xChIP, samples were incubated with 4 µg antibody (anti-H3K9me3: Abcam ab8898, anti-CTCF: Millipore 07-729) in FSB solution (5 mM EDTA, 20 mM Tris pH8, 500 mM NaCl) supplemented with 0.1% NP40 and cOmplete inhibitor at 4 ℃ with overnight rotation. Protein A/G magnetic beads (Thermo 88803) were added to capture antibody-conjugated chromatin, collected on magnetic rack, washed sequentially in low salt buffer (0.1% SDS, 1% Trition X-100, 2 mM EDTA, 20 mM Tris pH8, 150mM NaCl), high salt buffer (0.1% SDS, 1% Trition X-100, 2 mM EDTA, 20 mM Tris pH8, 500mM NaCl), lithium chloride buffer (1% NP40, 1% deoxycholic acid, 1 mM EDTA, 10 mM Tris pH8, 0.25 M LiCl), twice in TE buffer (1 mM EDTA, 10 mM Tris pH8), and were eluted in 0.1 M NaHCO₃ and 1% SDS. The eluate was treated with 1 µl of 10 mg/ml RNase A (Sigma R6513) and 2 µl of 10 mg/ml proteinase K (Sigma P2308), and ChIP DNA was finally collected by ethanol precipitation. For library preparation, ChIP DNA was blunted with End-it DNA Repair Kit (Epicentre ER0720), A-tailed with Klenow Exo-minus (Epicentre KL06041K), ligated with adapter (Vazyme N802-01) using Fast-Link kit (Epicentre LK11025) and amplified with VAHTS HiFi Universal Amplification Mix for Illumina (Vazyme N618-02). The library was size-selected with SPRIselect beads (Beckman B23318) and sent for sequencing on Illumina NovaSeq 6000 (paired-end, 150 bp).

Whole-genome bisulfite sequencing (WGBS)

WGBS was conducted according to published protocol⁴⁷, with minor modifications. ~ 2 million cultured cells were harvested, washed in PBS and lysed with NP40 lysis buffer (0.32 M sucrose, 5 mM CaCl₂, 3 mM Mg(Ace)₂, 0.1 mM EDTA, 10 mM Tris pH8, 0.1% NP40). Nuclei were collected with centrifugation, incubated with 5 µl of 10 mg/ml RNase A (Sigma R6513) in 500 µl PK lysis buffer (1 M Tris pH8, 50 mM EDTA, 2% SDS, 2 M NaCl) at 37 ℃ for 15 minutes, and then with 10 µl of 20 mg/ml proteinase K (Sigma P2308) at 52 ℃ overnight with rotation. The samples were then added with phenol/chloroform (1:1) (Solarbio P1021), mixed with vortex, and centrifugated with 15000 g at RT for 10 minutes. The supernatant was taken and precipitated with isopropanol. The precipitated genomic DNA (gDNA) was washed in 70% ethanol, air dried and dissolved in buffer EB (Qiagen 19086). The gDNA was then spiked in with 1% of unmethylated lambda DNA (Promega D1521) for efficiency validation of bisulfite conversion. 500 ng of gDNA was taken and sonicated to ~ 250 bp with Covaris M220 ultrasonicator. The DNA fragments were end-repaired and A-tailed with NEBNext Ultra II End Repair/dA-Tailing Module (NEB E7546), and then ligated with NEBNext Multiplex Oligos for Illumina (Methylated Adaptor, Index Primers Set 1, NEB E7535). After size-selected to 300 ~ 400 bp with Ampure XP beads (Beckman A63881), DNA was bisulfite converted with EZ DNA methylation kit (Zymo Research D5001), PCR amplified with KAPA HiFi HotStart Uracil + ReadyMix (Kapa Biosystems KK2801), and the library was finally size-selected to 300 ~ 500 bp with Ampure XP beads. After analysis with Agilent Tapestation 4200, the library was sent for deep sequencing on Illumina NovaSeq 6000 (paired-end, 150bp).

Hi-C

In situ Hi-C was conducted according to published protocol²⁸, with minor modifications. ~ 3 million cells were lysed in NP40 lysis buffer (0.32 M sucrose, 5 mM CaCl₂, 3 mM Mg(Ace)₂, 0.1 mM EDTA, 10 mM Tris pH8, 0.1% NP40), fixed in 1% formaldehyde (Sigma 252549) at RT for 5 minutes and quenched with 125 mM glycine at RT for 10 minutes. Nuclei were incubated in 0.5% SDS at 62 ℃ for 10 minutes, and quenched with 1.25% Triton X-100 (Sigma 93443) at 37 ℃ for 15 minutes. Chromatin was digested with 100 U of MboI restriction enzyme (NEB R0147) in 1X NEBuffer2 (NEB B7002S) at 37 ℃ overnight with rotation. Enzyme digestion was inactivated with incubation at 65 ℃ for 20 minutes, and nuclei were collected with centrifugation. Restriction fragment overhangs were filled and DNA ends were marked with biotin using End-it DNA Repair Kit (Epicentre ER0720) supplemented with 0.256 mM biotin-14-dATP (Life Technologies 19524-016), 0.4 mM dCTP, 0.4 mM dGTP and 0.4 mM dTTP (NEB N0446S). For proximity ligation, DNA fragments were ligated with 2000 U of T4 DNA ligase (NEB M0202) in 1200 µl of ligation master mix (T4 DNA ligase buffer (NEB B0202), 0.83% Triton X-100, 0.1 mg/ml bovine serum albumin (NEB B9200S)) at RT for 4 hours with slow rotation. After ligase inactivation at 65 ℃ for 5 minutes, nuclei were collected with centrifugation, and the chromatin was reverse-crosslinked in PK lysis buffer (1 M Tris pH8, 50 mM EDTA, 2% SDS, 2 M NaCl) at 65 ℃ for 4 hours, followed by RNase A (Sigma R6513) and proteinase K (Sigma P2308) digestion. DNA was further purified by phenol/chloroform (1:1) and ethanol precipitation, dissolved in buffer EB (Qiagen 19086) and sonicated to ~ 300 base pairs by Covaris E220 ultrasonicator system. For biotin pull-down, 100 µl per sample of 10 mg/ml Dynabeads MyOne Streptavidin T1 beads (Life technologies 65602) was washed with 400 µl per sample of 1 X TWB (5 mM Tris-HCl pH7.5, 0.5 mM EDTA, 1 M NaCl, 0.05% Tween-20), resuspended in 100 µl per sample of 2X binding buffer (10 mM Tris-HCl pH7.5, 1 mM EDTA, 2 M NaCl) and incubated with samples at RT for 15 minutes with rotation. The beads were separated on magnetic stand, washed three times with 1 X TWB (pre-warmed to 55 ℃) and once in 100 µl buffer EB (Qiagen 19086). For library preparation, the beads were collected for DNA end repair with End-it DNA Repair Kit, A-tailing with Klenow Exo-minus (Epicentre KL06041K), adapter (Vazyme N802-01) ligation with Fast-Link kit (Epicentre LK11025) and PCR amplification with VAHTS HiFi Universal Amplification Mix for Illumina (Vazyme N618-02). Supernatant was taken as the library, which was further size-selected with SPRIselect beads (Beckman B23318) and sent for deep sequencing on Illumina NovaSeq 6000 (paired-end, 150 bp).

Cell cycle analysis

Cell cycle analysis was conducted according to published protocol⁴⁸, with minor modifications. In brief, ~ 1 million NPCs were dissociated with Accutase solution (Sigma A6964), collected by centrifugation and resuspended in 0.5 mL of PBS (Biosharp BL302A), followed by gentle pipetting. For fixation, the resuspended cells were transferred into 4.5 mL of 70% ethanol and fixed overnight. After centrifugation, ethanol was decanted, and fixed cells were washed for 30 seconds in 5 mL of PBS, followed by centrifugation. Fixed cell pellet was resuspended in 1 mL of PI staining buffer (0.1% Triton X-100 (Sigma 93443), 10 µg/ml propidium iodide (Sigma S7109), 100 µg/ml DNase-Free RNaseA (Sigma R6513) in PBS), incubated for 30 minutes at room temperature in darkness and transferred to flow cytometer (Beckman Cytoflex). PI fluorescence was excited by 488 nm laser, and emission was measured with 600 nm filter.

BrdU labeling & immunofluorescence staining

BrdU labeling and immunofluorescence staining was conducted according to published protocol^19,49, with minor modifications. Female Setdb1^2lox/2lox mice were crossed with male Setdb1^2lox/+, Nestin-Cre mice and checked for vaginal plug. At gestational day 15.5, dams were injected with 50 mg/kg BrdU (Sigma B5002) intraperitoneally and sacrificed 24 hours post-injection. Fetus mice were dissected, genotyped, and classified into WT and KO. Fetus mice brains were fixed in 4% paraformaldehyde (Sigma P6148) at 4 ℃ for 6 hours, dehydrated in 30% sucrose (Sinopharm 10021463) at 4 ℃ overnight and embedded in optimal cutting temperature compound (Sakura 4583). Brains were sliced for 15 µm on cryostat (Leica CM1950), dried and preserved at -80 ℃. For immunofluorescence staining, brain slices were dried at RT for 30 minutes, washed in PBS (Solarbio P1022) for 3 X 5 minutes, treated with 2 M HCl at RT for 30 minutes, washed in 0.1 M boric acid buffer (80 mM boric acid (Sinopharm 10004818) and 20 mM sodium borate (Sangon A500832)) for 3 X 5 minutes and in PBS for 5 minutes. For blocking, brain slices were incubated in blocking solution (5% goat serum (Thermo 16210064) + 0.5% Triton X-100 (Sigma 93443) in PBS) at RT for 30 minutes. For first antibody conjugation, brain slices were incubated in Anti-BrdU (Abcam ab6326, 1:1000 dilution) and Anti-PH3 (Merck 06-570, 1:2000 dilution) diluted in blocking solution at 4 ℃ overnight, and then washed in PBS for 3 X 5minutes. For second antibody conjugation and DAPI counterstaining, brain slices were incubated in Goat Anti-Rat 488 (Jackson Immunoresearch 112-545-003, 1:1000 dilution), Goat Anti-Rabbit 594 (Jackson Immunoresearch 111-585-003, 1:1000 dilution) and DAPI (Beyotime C1002, 1:1000 dilution) diluted in PBS at RT for 1 hour, and then washed in PBS for 3 X 5 minutes in darkness. Finally, brain slices were covered in mounting medium (Sigma F4680), and images were captured with Olympus VS120 virtual slide microscope. Manual cell counting was conducted with Imaris 9.5.

Data analysis

All statistical plots were generated with Graphpad Prism 9 and R package “ggplot2”⁵⁰ (version 3.4.2).

ATAC-seq & ChIP-seq

ATAC-seq and ChIP-seq were analyzed according to published protocol¹⁹, with minor modifications. ATAC-seq data were downloaded from GEO (GSE186806)¹⁹. Quality control of ChIP-seq reads were conducted with FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) (version 0.11.9). Reads were trimmed with Trim Galore (https://github.com/FelixKrueger/TrimGalore) (version 0.6.7, --quality 20 --phred33 --stringency 1 --length 20 --fastqc --paired) and aligned to mm10 with Bowtie2⁵¹ (version 2.4.1) in default settings. Reads were further binarized, sorted, deduplicated and indexed with SAMtools⁵² (version 1.6). Bigwig files were generated with “bamCoverage” function of Deeptools⁵³ (version 3.5.1, --binSize 10 --normalizeUsing RPKM --effectiveGenomeSize 2652783500 --ignoreForNormalization chrX chrM). Narrow peaks were called with MACS2⁵⁴ (version 2.2.7.1) (ATAC-seq: --shift − 100 --extsize 200 –nomodel; ChIP-seq: default settings). Differential analysis was conducted with R package “DiffBind”⁵⁵ (version 3.4.11). Peaks were counted with summits = FALSE, bUseSummarizeOverlaps = TRUE. Difference was reported with method = DBA_DESEQ2, contrast = 1, th = 1. Differential peaks were defined as FDR < 0.05 & Log2FoldChange > = 0.585 (up) or <= -0.585 (down). Genomic annotation of peaks was conducted with R package “ChIPseeker”⁵⁶ and “clusterProfiler”⁵⁷, with promoters defined as transcription start site (TSS) ± 3 kb. For DNA repeat analysis, Bedtools⁵⁸ (version 2.30.0) was used to intersect differential peaks and RepeatMasker reference (http://repeatmasker.org/) (-f 0.5 -F 0.5 -e). Profiles and heatmaps of ATAC-seq and ChIP-seq signals were generated with “computeMatrix scale-regions” (-a 1000 -b 1000 --skipZeros) and “plotHeatmap” (--zMin 3) functions of Deeptools⁵³ (version 3.5.1). Motif analysis was conducted with “findMotifsGenome.pl” function of Homer (http://homer.ucsd.edu/homer/) (version 4.11.1), with all peaks from the WT group as background (-bg). Permutation tests were conducted with “permTest” function (ntimes = 200, randomize.function = resampleRegions, universe = “all_loop_anchors”, evaluate.function = numOverlap or MeanDistance) of R package “regioneR”⁵⁹ (version 1.26.1).

WGBS

WGBS reads were analyzed according to published protocol⁴⁷, with minor modifications. Briefly, WGBS reads were filtered and trimmed with “fastp”⁶⁰ (version 0.23.4, -q 20 -u 50 -n 5), and the quality control was conducted by FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) (version 0.11.9). High quality clean reads were aligned to mouse genome mm10 and lambda phage genome with Bowtie2⁵¹ (version 2.2.3) in default settings, and then deduplicated with “deduplicate_bismark” function of Bismark⁶¹ (version 0.22.1). Methylation measurement was extracted with “bismark_methylation_extractor” function of Bismark (--no_overlap --ignore_r2 2 --comprehensive), and the bisulfite conversion rate was calculated according to the methylation level of spiked-in lambda DNA. R package “bsseq”⁶² (version 1.36.0) was used to identify differential methylated regions (DMRs) (ns = 70, h = 1000), and only CpGs covered by at least 2 reads in at least 1 sample per genotype group were retained. The final definition of DMRs are as follows: (1) t-statistic with a qcutoff between (0.01,0.990); (2) maxGap = 300; (3) contain at least 3 CpGs; (4) absolute mean methylation difference > = 0.1; (5) located on autosome.

Hi-C

Hi-C data were analyzed according to published protocol²⁸, with minor modifications. Briefly, quality control was conducted by FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) (version 0.11.9), and reads were trimmed with Trim Galore (https://github.com/FelixKrueger/TrimGalore) (version 0.6.7, --quality 20 --phred33 --stringency 1 --length 20 --fastqc --paired). Restriction sites of MboI on mm10 were generated with “hiccup_digester” function of HiCUP (https://stevenwingett.github.io/HiCUP/) (version 0.8.0). High quality clean reads were aligned to mm10 with Bowtie2 program⁵¹ (version 2.4.1) in HiCUP (max di-tag length: 700, min di-tag length: 50). “.ValidPairs” were generated with the “mapped_2hic_fragments.py” script of HiC-Pro⁶³ (version 2.11.4) in default settings. “.allValidPairs” were merged with “.ValidPairs” of samples with the same genotype (WT or KO), sorted, deduplicated and down-sampled to 1.038 billion for downstream comparison. “.hic” files were generated with “hicpro2juicebox.sh” utility of HiC-Pro⁶³ and JuicerTools (https://github.com/aidenlab/juicertools) (version 1.22.01). Contact maps were generated with Juicebox (https://aidenlab.org/juicebox/) (version 1.11.08). Raw and ICE-normalized matrixes (iced_matrix) were built with HiC-Pro⁶³ in default settings (bin_size: 10000 20000 100000). Iced_matrixes were converted into “.h5” and “.cool” files with “hicConvertFormat” tool of HiCExplorer⁶⁴ (version 3.7.1). Compartments were called in bin_size of 100 kb by “hicPCA” tool of HiCExplorer⁶⁴ in default settings, using bigwig file of WT ATAC-seq¹⁹ to decide whether the eigenvector needs a sign flip. Insulation scores were calculated in bin_size of 10 kb with “insulation_score” function (window = 32) of R package “GENOVA”⁶⁵ (version 1.0.0.9000), and topologically-associated domains (TADs) were called with “call_TAD_insulation” function of GENOVA⁶⁵ in default settings. Insulation heatmaps were generated with “tornado_insulation” function (bed_pos = ‘center’) of GENOVA⁶⁵. Aggregate TAD analysis was conducted with “ATA” function of GENOVA⁶⁵ (dist_thres = c(10000, Inf)). Loops were called in bin_size of 10 kb and 20 kb with “hicDetectLoops” tool of HiCExplorer⁶⁴ in default settings, and were merged with “hicMergeLoops” of HiCExplorer⁶⁴. Aggregate peak analysis was conducted with “APA” function of GENOVA⁶⁵ in default settings. Permutation tests were conducted with “permTest” function (ntimes = 200, randomize.function = resampleRegions, universe = “all_loop_anchors”, evaluate.function = numOverlap) of R package “regioneR”⁵⁹ (version 1.26.1). Bedtools⁵⁸ (version 2.30.0) was used to intersect CTCF_up_B2 and New_loop_anchor in default settings.

RNA-seq

RNA-seq data was analyzed according to published protocol¹⁹, with minor modifications. Raw data were downloaded from GEO (GSE186806)¹⁹. Quality control was conducted by FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) (version 0.11.9), and reads were trimmed with Trim Galore (https://github.com/FelixKrueger/TrimGalore) (version 0.6.7, --quality 20 --phred33 --stringency 1 --length 20 –fastqc --paired). High quality clean reads were aligned to mm10 with HISAT2⁶⁶ (version 2.2.1) in default settings. Reads were further binarized, sorted, and indexed with SAMtools⁵² (version 1.6). Reads on exons were counted with featureCounts⁶⁷ (version 2.0.3) (-p -t exon -g gene_id -a gencode.vM20.annotation.gtf). Differential gene expression analysis was conducted with DESeq2⁶⁸ (version 1.34.0) in default settings, and differentially-expressed genes (diff_genes) were defined as P-value < 0.05. Gene annotations were conducted with R package “biomaRt”⁶⁹ (version 2.50.3) (biomart = "ensembl", version = 102, dataset = "mmusculus_gene_ensembl"), and promoters were defined as TSS ± 3 kb. Permutation tests were conducted with “permTest” function (ntimes = 200, randomize.function = resampleRegions, universe = “all_gene_promoters”, evaluate.function = numOverlap) of R package “regioneR”⁵⁹ (version 1.26.1). Bedtools⁵⁸ (version 2.30.0) was used to intersect diff_gene_promoters and B2_new_loop in default settings. Protein network analysis was conducted by STRING (https://string-db.org/) (version 11.5) with K-means clustering (K = 2).

Data availability

All raw and processed sequencing data have been submitted to the Gene Expression Omnibus (GEO) under the accession number GSE247620. Reviewers can access the data using the secure token 'mvadmcyydturfmz'.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 81971272, No.32170601) (Y.J.), Science and Technology Innovation 2030 - Major Project (No. 2021ZD0203000) (Y.J.).

Contributions

Y.J. and W.M. conceived the ideas and supervised the research; D.S. and Y.Z. designed and performed the experiments; W.P. conducted most of the bioinformatic analysis; all the other authors contributed to experiment performance and data analysis; all authors contributed to scientific discussion and manuscript preparation.

Ethics declarations

The authors declare no competing interests.

Lander, E. S. et al. Initial sequencing and analysis of the human genome. Nature 409, 860-921 (2001).
Waterston, R. H. et al. Initial sequencing and comparative analysis of the mouse genome. Nature 420, 520-562 (2002).
Wicker, T. et al. A unified classification system for eukaryotic transposable elements. Nat Rev Genet 8, 973-982 (2007).
Chuong, E. B., Elde, N. C. & Feschotte, C. Regulatory activities of transposable elements: from conflicts to benefits. Nat Rev Genet 18, 71-86, doi:10.1038/nrg.2016.139 (2017).
Anania, C. Transposable element evolution in mammals. Nat Genet 55, 904, doi:10.1038/s41588-023-01430-x (2023).
Osmanski, A. B. et al. Insights into mammalian TE diversity through the curation of 248 genome assemblies. Science 380, eabn1430, doi:10.1126/science.abn1430 (2023).
Reilly, M. T., Faulkner, G. J., Dubnau, J., Ponomarev, I. & Gage, F. H. The role of transposable elements in health and diseases of the central nervous system. J Neurosci 33, 17577-17586, doi:10.1523/JNEUROSCI.3369-13.2013 (2013).
Chesnokova, E., Beletskiy, A. & Kolosov, P. The Role of Transposable Elements of the Human Genome in Neuronal Function and Pathology. Int J Mol Sci 23, doi:10.3390/ijms23105847 (2022).
Jonsson, M. E., Garza, R., Johansson, P. A. & Jakobsson, J. Transposable Elements: A Common Feature of Neurodevelopmental and Neurodegenerative Disorders. Trends Genet 36, 610-623, doi:10.1016/j.tig.2020.05.004 (2020).
Kramerov, D. A. & Vassetzky, N. S. SINEs. Wiley Interdiscip Rev RNA 2, 772-786, doi:10.1002/wrna.91 (2011).
Elbarbary, R. A., Lucas, B. A. & Maquat, L. E. Retrotransposons as regulators of gene expression. Science 351, aac7247, doi:10.1126/science.aac7247 (2016).
Chesnokova, E., Beletskiy, A. & Kolosov, P. The Role of Transposable Elements of the Human Genome in Neuronal Function and Pathology. Int J Mol Sci 23, doi:10.3390/ijms23105847 (2022).
Ahmadi, A., De Toma, I., Vilor-Tejedor, N., Eftekhariyan Ghamsari, M. R. & Sadeghi, I. Transposable elements in brain health and disease. Ageing Res Rev 64, 101153, doi:10.1016/j.arr.2020.101153 (2020).
Linker, S. B. et al. Identification of bona fide B2 SINE retrotransposon transcription through single-nucleus RNA-seq of the mouse hippocampus. Genome Res 30, 1643-1654, doi:10.1101/gr.262196.120 (2020).
Ferrigno, O. et al. Transposable B2 SINE elements can provide mobile RNA polymerase II promoters. Nat Genet 28, 77-81 (2001).
Bartlett, A. A., Guffanti, G. & Hunter, R. G. B2 SINE RNA as a novel regulator of glucocorticoid receptor transcriptional activity. Neurobiol Stress 23, 100522, doi:10.1016/j.ynstr.2023.100522 (2023).
Zovoilis, A., Cifuentes-Rojas, C., Chu, H.-P., Hernandez, A. J. & Lee, J. T. Destabilization of B2 RNA by EZH2 Activates the Stress Response. Cell 167, doi:10.1016/j.cell.2016.11.041 (2016).
Luo, X. et al. 3D Genome of macaque fetal brain reveals evolutionary innovations during primate corticogenesis. Cell 184, doi:10.1016/j.cell.2021.01.001 (2021).
Li, J. et al. Histone Methyltransferase SETDB1 Regulates the Development of Cortical Htr3a-Positive Interneurons and Mood Behaviors. Biological psychiatry 93, 279-290, doi:10.1016/j.biopsych.2022.08.021 (2023).
Estécio, M. R. H. et al. SINE retrotransposons cause epigenetic reprogramming of adjacent gene promoters. Mol Cancer Res 10, 1332-1342, doi:10.1158/1541-7786.MCR-12-0351 (2012).
Gualdrini, F. et al. H3K9 trimethylation in active chromatin restricts the usage of functional CTCF sites in SINE B2 repeats. Genes Dev 36, 414-432, doi:10.1101/gad.349282.121 (2022).
Kaaij, L. J. T., Mohn, F., van der Weide, R. H., de Wit, E. & Bühler, M. The ChAHP Complex Counteracts Chromatin Looping at CTCF Sites that Emerged from SINE Expansions in Mouse. Cell 178, doi:10.1016/j.cell.2019.08.007 (2019).
Ostapcuk, V. et al. Activity-dependent neuroprotective protein recruits HP1 and CHD4 to control lineage-specifying genes. Nature 557, 739-743, doi:10.1038/s41586-018-0153-8 (2018).
Zhu, Y., Sun, D., Jakovcevski, M. & Jiang, Y. Epigenetic mechanism of SETDB1 in brain: implications for neuropsychiatric disorders. Transl Psychiatry 10, 115, doi:10.1038/s41398-020-0797-7 (2020).
Chandrasekaran, S. et al. Neuron-specific chromosomal megadomain organization is adaptive to recent retrotransposon expansions. Nat Commun 12, 7243, doi:10.1038/s41467-021-26862-z (2021).
Tan, S.-L. et al. Essential roles of the histone methyltransferase ESET in the epigenetic control of neural progenitor cells during development. Development 139, 3806-3816 (2012).
Sun, D., Weng, J., Dong, Y. & Jiang, Y. 3D genome organization in the central nervous system, implications for neuropsychological disorders. J Genet Genomics 48, 1045-1056, doi:10.1016/j.jgg.2021.06.017 (2021).
Jiang, Y. et al. The methyltransferase SETDB1 regulates a large neuron-specific topological chromatin domain. Nature genetics 49, 1239-1250, doi:10.1038/ng.3906 (2017).
Wang, Z. et al. Dominant role of DNA methylation over H3K9me3 for IAP silencing in endoderm. Nat Commun 13, 5447, doi:10.1038/s41467-022-32978-7 (2022).
Kato, M., Takemoto, K. & Shinkai, Y. A somatic role for the histone methyltransferase Setdb1 in endogenous retrovirus silencing. Nat Commun 9, 1683, doi:10.1038/s41467-018-04132-9 (2018).
Adoue, V. et al. The Histone Methyltransferase SETDB1 Controls T Helper Cell Lineage Integrity by Repressing Endogenous Retroviruses. Immunity 50, doi:10.1016/j.immuni.2019.01.003 (2019).
Zheng, H. & Xie, W. The role of 3D genome organization in development and cell differentiation. Nat Rev Mol Cell Biol 20, 535-550, doi:10.1038/s41580-019-0132-4 (2019).
Zuin, J. et al. Nonlinear control of transcription through enhancer-promoter interactions. Nature 604, 571-577, doi:10.1038/s41586-022-04570-y (2022).
Tang, Z. et al. CTCF-Mediated Human 3D Genome Architecture Reveals Chromatin Topology for Transcription. Cell 163, 1611-1627, doi:10.1016/j.cell.2015.11.024 (2015).
Nora, E. P. et al. Targeted Degradation of CTCF Decouples Local Insulation of Chromosome Domains from Genomic Compartmentalization. Cell 169, doi:10.1016/j.cell.2017.05.004 (2017).
Barski, A. et al. High-resolution profiling of histone methylations in the human genome. Cell 129, 823-837 (2007).
Deniz, Ö., de la Rica, L., Cheng, K. C. L., Spensberger, D. & Branco, M. R. SETDB1 prevents TET2-dependent activation of IAP retroelements in naïve embryonic stem cells. Genome Biol 19, 6, doi:10.1186/s13059-017-1376-y (2018).
Rowe, H. M. et al. KAP1 controls endogenous retroviruses in embryonic stem cells. Nature 463, 237-240, doi:10.1038/nature08674 (2010).
Sharif, J. et al. Activation of Endogenous Retroviruses in Dnmt1(-/-) ESCs Involves Disruption of SETDB1-Mediated Repression by NP95 Binding to Hemimethylated DNA. Cell Stem Cell 19, 81-94, doi:10.1016/j.stem.2016.03.013 (2016).
Leung, D. et al. Regulation of DNA methylation turnover at LTR retrotransposons and imprinted loci by the histone methyltransferase Setdb1. Proc Natl Acad Sci U S A 111, 6690-6695, doi:10.1073/pnas.1322273111 (2014).
Mochizuki, K. et al. Repression of germline genes by PRC1.6 and SETDB1 in the early embryo precedes DNA methylation-mediated silencing. Nat Commun 12, 7020, doi:10.1038/s41467-021-27345-x (2021).
Liu, S. et al. Setdb1 is required for germline development and silencing of H3K9me3-marked endogenous retroviruses in primordial germ cells. Genes Dev 28, 2041-2055, doi:10.1101/gad.244848.114 (2014).
Karimi, M. M. et al. DNA methylation and SETDB1/H3K9me3 regulate predominantly distinct sets of genes, retroelements, and chimeric transcripts in mESCs. Cell Stem Cell 8, 676-687, doi:10.1016/j.stem.2011.04.004 (2011).
Bourque, G. et al. Evolution of the mammalian transcription factor binding repertoire via transposable elements. Genome Res 18, 1752-1762, doi:10.1101/gr.080663.108 (2008).
Choudhary, M. N. K., Quaid, K., Xing, X., Schmidt, H. & Wang, T. Widespread contribution of transposable elements to the rewiring of mammalian 3D genomes. Nat Commun 14, 634, doi:10.1038/s41467-023-36364-9 (2023).
Stadhouders, R., Filion, G. J. & Graf, T. Transcription factors and 3D genome conformation in cell-fate decisions. Nature 569, 345-354, doi:10.1038/s41586-019-1182-7 (2019).
Liu, P. et al. Altered DNA methylation pattern reveals epigenetic regulation of Hox genes in thoracic aortic dissection and serves as a biomarker in disease diagnosis. Clin Epigenetics 13, 124, doi:10.1186/s13148-021-01110-9 (2021).
Hang, H. & Fox, M. H. Analysis of the mammalian cell cycle by flow cytometry. Methods Mol Biol 241, 23-35 (2004).
Ligasová, A., Frydrych, I. & Koberna, K. Basic Methods of Cell Cycle Analysis. Int J Mol Sci 24, doi:10.3390/ijms24043674 (2023).
Wickham, H. ggplot2: Elegant Graphics for Data Analysis. (Springer-Verlag New York, 2016).
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat Methods 9, 357-359, doi:10.1038/nmeth.1923 (2012).
Danecek, P. et al. Twelve years of SAMtools and BCFtools. Gigascience 10, doi:10.1093/gigascience/giab008 (2021).
Ramírez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res 44, W160-W165, doi:10.1093/nar/gkw257 (2016).
Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol 9, R137, doi:10.1186/gb-2008-9-9-r137 (2008).
Ross-Innes, C. S. et al. Differential oestrogen receptor binding is associated with clinical outcome in breast cancer. Nature 481, 389-393, doi:10.1038/nature10730 (2012).
Yu, G., Wang, L.-G. & He, Q.-Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics 31, 2382-2383, doi:10.1093/bioinformatics/btv145 (2015).
Yu, G., Wang, L.-G., Han, Y. & He, Q.-Y. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 16, 284-287, doi:10.1089/omi.2011.0118 (2012).
Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841-842, doi:10.1093/bioinformatics/btq033 (2010).
Gel, B. et al. regioneR: an R/Bioconductor package for the association analysis of genomic regions based on permutation tests. Bioinformatics 32, 289-291, doi:10.1093/bioinformatics/btv562 (2016).
Chen, S., Zhou, Y., Chen, Y. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34, i884-i890, doi:10.1093/bioinformatics/bty560 (2018).
Krueger, F. & Andrews, S. R. Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications. Bioinformatics 27, 1571-1572, doi:10.1093/bioinformatics/btr167 (2011).
Hansen, K. D., Langmead, B. & Irizarry, R. A. BSmooth: from whole genome bisulfite sequencing reads to differentially methylated regions. Genome Biol 13, R83, doi:10.1186/gb-2012-13-10-r83 (2012).
Servant, N. et al. HiC-Pro: an optimized and flexible pipeline for Hi-C data processing. Genome Biol 16, 259, doi:10.1186/s13059-015-0831-x (2015).
Wolff, J., Backofen, R. & Grüning, B. Loop detection using Hi-C data with HiCExplorer. Gigascience 11, doi:10.1093/gigascience/giac061 (2022).
van der Weide, R. H. et al. Hi-C analyses with GENOVA: a case study with cohesin variants. NAR Genom Bioinform 3, lqab040, doi:10.1093/nargab/lqab040 (2021).
Kim, D., Paggi, J. M., Park, C., Bennett, C. & Salzberg, S. L. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat Biotechnol 37, 907-915, doi:10.1038/s41587-019-0201-4 (2019).
Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923-930, doi:10.1093/bioinformatics/btt656 (2014).
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15, 550 (2014).
Kinsella, R. J. et al. Ensembl BioMarts: a hub for data retrieval across taxonomic space. Database (Oxford) 2011, bar030, doi:10.1093/database/bar030 (2011).

There is NO Competing Interest.

DSetal.SupplementaryFigures.docx
Supplementary Figures
DSetal.SupplementaryTables.xlsx
Supplementary Tables

Download PDF

Version 1

posted

You are reading this latest preprint version

Epigenetic Modulation of Short Interspersed Nuclear Elements Activity Influences Neural Precursor Cell Proliferation

Status:

Version 1

Abstract

Figures

Introduction

Results

SETDB1 selectively inhibited the chromatin accessibility on SINE_B2 elements

SETDB1 suppressed chromatin accessibility of SINE_B2 via H3K9me3 deposition

DNA methylation participated in SETDB1-mediated repression of SINE_B2

SETDB1 constrained excessive CTCF binding at SINE_B2

Chromatin loop formation due to increased CTCF occupancy at SINE_B2 elements

SETDB1 maintained cell cycle in NPCs by restraining aberrant loop formation

Discussion

Methods

Animals

Neurosphere culture

ChIP-seq

Whole-genome bisulfite sequencing (WGBS)

Hi-C

Cell cycle analysis

BrdU labeling & immunofluorescence staining

Data analysis

ATAC-seq & ChIP-seq

WGBS

Hi-C

RNA-seq

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1