Pluripotency exit is guided by the Peln1-mediated disruption of intrachromosomal architecture

Pluripotent stem cells may become useful reagents in regenerative medicine and cell replacement therapy, primarily due to their indefinite self-renewal properties in vitro and their ability to differentiate into any cell type that is needed (Silva et al., 2019; Yilmaz and Benvenisty, 2019). Pluripotency is precisely regulated by a network of key transcription factor genes, including Oct4, Nanog, and Sox2, as well as signaling pathways, including the Wnt/β-catenin pathway, leukemia inhibitory factor (LIF)/Janus kinase-signal transducer and activator of transcription 3 pathway, phosphoinositide 3 kinase pathway, and Ras/MAPK pathway (Ramazzotti et al., 2019; Rasmussen et al., 2018). Together, this gene regulation network orchestrates molecular processes that ensure the maintenance of pluripotent identity in stem cells. Disruption of this network leads to loss of pluripotency and cellular differentiation (Li and Izpisua Belmonte, 2018; Niwa, 2018). Understanding the molecular mechanisms required for stem cells to maintain a delicate balance between pluripotency and differentiation is critical for advancing stem cell–based therapies to clinical applications.

Mouse embryonic stem cells (ESCs) exist in naive and primed states with distinct transcriptional regulatory networks, epigenotypes, and signaling requirements (Nichols and Smith, 2009; Shiozawa et al., 2020; Weinberger et al., 2016). Mouse ESCs derived from the inner cell mass of blastocysts are considered as being in the naive state. While maintained in a self-renewing state through the LIF and BMP pathways, these cells have the ability to differentiate to produce chimeric animals. In contrast, postimplantation epiblast-derived mouse epiblast stem cells are in a primed pluripotent state; these cells are dependent on the FGF–Activin pathways and rarely form chimera. However, primed cells can be reverted to naive status by KLF4, MYC, or Nanog in the presence of LIF. Conversely, naive ESCs can be induced into the primed state by activating the FGF2/Activin pathway. Furthermore, naive mESCs can become homogeneous “ground-state” pluripotent stem cells under the influence of two small-molecule MEK/ERK and GSK inhibitors plus LIF (2i/LIF; Panepucci and de Souza Lima, 2019; Papp and Plath, 2011; Papp and Plath, 2013; Shiozawa et al., 2020). Despite the fact that a number of pathways are known to be involved in the interconversion of these states, specific underlying mechanisms remain to be characterized.

Stem cell fate is dynamically regulated by balanced self-renewal and differentiation signals. To initiate the exit from pluripotentcy toward differentiation, cells need to disable pluripotency-specific transcriptional networks and activate lineage-specific pathways (Betschinger et al., 2013; Kalkan and Smith, 2014). In response to differentiation signals, such as withdrawal of LIF from the culture medium or treatment with retinoic acid, pluripotent stem cells dismantle the naive pluripotency network and shut down the expression of the stem cell core transcription factors. The molecular components that coordinate pluripotent exit are poorly understood.

Long noncoding RNAs (lncRNAs) have versatile biological functions in a variety of biological processes (Huang et al., 2019; Jae et al., 2019; Kopp, 2019; Krishnan and Mishra, 2014). Recent studies suggest that lncRNAs may play important roles in regulating the network that contributes to ESC self-renewal and differentiation (Fico et al., 2019; Mirzadeh Azad et al., 2021; Perry and Ulitsky, 2016). Multiple pluripotency-associated lncRNAs have been identified embedded in the chromatin regulatory network using CRISPR Cas9 screening, RNA transcriptome sequencing (RNA-seq), and RNA reverse transcription-associated trap sequencing (RAT-seq; Chen et al., 2019, Chen et al., 2020; Du et al., 2018; Zhu et al., 2020b). These lncRNAs regulate the pluripotency network using a variety of epigenetic mechanisms. After binding to pluripotency-associated gene promoters, like Oct4 and Sox2, these lncRNAs coordinate the construction of a pluripotency-specific intrachromosomal loop, modulate promoter methylation by recruiting demethylase TET1, and activate the eRNA pathway of stemness genes (Cao et al., 2021; Du et al., 2021; Zhu et al., 2020a).

However, little information is available about the role of lncRNAs in controlling the exit of stem cells from the pluripotent status. To better understand the underlying mechanisms leading to pluripotent exit and initiation of lineage specification, we explored a novel strategy to identify lncRNAs that interact with the promoter of Oct4, a master transcription factor that maintains the ground pluripotent state. By using chromatin RNA in situ reverse transcription sequencing (CRIST-seq; Zhang et al., 2019), we identified Peln1 as an important lncRNA component in the circuitry that controls the exit from pluripotency in stem cells. Notably, we showed that Peln1 exerted this function by regulating the activity of the Oct4 promoter in trans. These findings may shed light on our understanding of the molecular basis of lncRNAs as new chromatin players in the regulatory network of stem cells.

We used a novel strategy to map functional lncRNAs related to pluripotency exit in stem cells (Fig. 1 A). First, a CRIST-seq approach (Zhang et al., 2019) was used to identify functional lncRNAs that interact with the promoter of Oct4, a critical master factor that is required for the maintenance of stem cell pluripotency (Fig. S1 A). Un-reprogrammed fibroblasts were used for the initial CRIST-seq screening of lncRNAs that negatively regulate Oct4. Cells were transfected to express a lentiviral dCas9 Oct4 gRNA cassette. After membrane lysis, the nuclei were collected and RNAs were reverse transcribed in situ with biotin–dCTP. The Oct4-lncRNA-cDNA chromatin complex was pulled down by Cas9-FLAG immunoprecipitation. The biotin–cDNA complexes were then purified with streptavidin beads. Using quantitative PCR (Q-PCR), we showed that the Oct4 gRNA guided Cas9 specifically to the targeting site (Site I) at the Oct4 locus. There was a very low binding signal at the control site (Site II). Only slight background signals were detected in the vector control (Vector) and the Cas9 random gRNA control (gCT; Fig. S1 B). After confirming specific Cas9 Oct4 gRNA targeting, the CRIST library was constructed for Illumina sequencing (Zhang et al., 2019).

Conventional RNA-seq was performed to identify differentially expressed lncRNAs related to the induction of pluripotent stem cells (Du et al., 2018). The RNA-seq dataset was then integrated with the CRIST-seq dataset to identify potential lncRNAs that are not only silenced during pluripotency induction, but also have the ability to regulate the Oct4 gene (Fig. 1 A).

Using this strategy, we identified lncRNA NONMMUT056388 as a novel RNA molecule that is associated with exit from pluripotency in stem cells. We refer to this lncRNA as Peln1 (pluripotency exit lncRNA 1). Quantitative CRIST-PCR confirmed that Peln1 was enriched in the Cas9 Oct4 gRNA library, but not in the Cas9 vector control and Cas9 random gRNA (gCT) libraries (Fig. S1 C). The other two control lncRNAs, Prdl4 and Prdl5 (pluripotent reprograming-downregulated lncRNA 4, 5) were not detected in the Oct4 promoter complex. The RNA-seq data indicated that Peln1 was expressed in fibroblasts but was silenced when cells completed reprogramming into iPSCs.

Finally, we used RAT-seq to identify the genome-wide interaction targets for Peln1 (Fig. S1 D; Chen et al., 2018; Sun et al., 2014; Wang et al., 2014). Cells were cross-linked to fix the lncRNA–chromatin structure, and chromatin RNAs were reverse transcribed in situ using three Peln1-specific complementary primers (Table S1) containing biotin–dCTP. The Peln1 cDNA-interacting chromatin complex was pulled down with streptavidin beads and the target gene DNAs were purified for library sequencing. A control RAT library was constructed using random oligonucleotide primers. The Peln1 RAT-seq data were normalized over the peaks of the control RAT-seq data to map the lncRNA target gene interaction network. Using RAT-seq analysis, we found that Peln1 interacted with numerous pluripotency-associated genes (Fig. 1 B). Specifically, Peln1 interacted with the promoter and enhancer of the stem cell core factor Oct4 (Fig. 1 C) and the regulatory elements of Sox2 and Nanog (Fig. S1, E and F).

We then characterized the gene structure of Peln1. Using 5′-rapid amplification of cDNA ends (RACE) and 3′-RACE, we obtained the full-length cDNA for Peln1 (Fig. S2, A–D). We also cloned the 2 kb upstream fragment of Peln1 as its putative promoter and inserted it into the pGL3 vector to generate a luciferase reporter construct (Fig. S2 E). The luciferase reporter assay showed that this 2 kb upstream fragment contained strong promoter activity. These data suggest that Peln1 may be transcribed from an intergenic promoter that is located in the Ptn (Pleiotrophin) gene locus.

We also examined whether the status of DNA methylation in the Peln1 promoter was related to differential expression of Peln1 between fibroblasts and E14 cells. We found that one of the CpGs was completely unmethylated in E14 cells but was completely methylated in fibroblasts and mouse embryonic fibroblasts (MEFs; Fig. S2 F), suggesting that the Peln1 promoter may be regulated by transcriptional repressors that bind to the unmethylated CpG sites.

We then conducted a series of assays to characterize the role of Peln1 in stem cells. We first quantitated the expression of Peln1 during pluripotent reprograming, showing that the expression of Peln1 was negatively correlated with pluripotency. Peln1 was highly expressed in fibroblasts as well as un-reprogrammed cells (non-iPSCs), which expressed four viral reprogramming factors (OSKM), but failed to complete reprogramming (Zhang et al., 2013). However, Peln1 became silenced once cells were fully reprogrammed into iPSCs (Fig. 2 A). The gene was also silenced in embryonic stem cells (E14). Q-PCR confirmed that Peln1 was highly expressed in fibroblasts and non-iPSCs (Fig. 2 B) and expressed at variable levels in adult tissues (Fig. S2 G). Peln1 was upregulated in E14 cells after LIF withdrawal (Fig. S2 H), but there was little difference in Peln1 expression between the cultured “primed” cells and “naive” cells (Fig. S2 I).

The cytoplasmic and nuclear fraction assay showed that Peln1 was predominantly located in the nucleus (Fig. 2 C). RNA-FISH confirmed the nuclear localization (Fig. 2 D). During embryoid body differentiation, Peln1 was activated in parallel with decreased expression of Oct4, Sox2, and Nanog (Fig. 2 E). These data suggest that Peln1 is a critical nuclear factor that is negatively associated with stem cell pluripotency.

To examine the function of Peln1 in pluripotency, we cloned the full length of Peln1 cDNA into a lentiviral expression vector and overexpressed it in E14 cells. The empty lentiviral vector was used as a negative control (Fig. S3 A). The success of lentiviral transfection was tracked using the DsRed fluorescence marker in the lentiviral vector. After partial puromycin selection, Peln1-expressing E14 cells (red) showed obvious morphological changes and lost the stem cell marker OCT4 (Fig. 3 A, top panel, red arrows). In those cells that did not express Peln1, OCT4 was detected by immunostaining, suggesting that the untargeted cells remained pluripotent. After complete puromycin selection, all E14 cells became differentiated (middle panel). As expected, normal cell morphology and the stem cell marker OCT4 were preserved in the vector control cells that expressed the DsRed tracking maker (bottom panel).

We also collected DsRed-expressing cells and used Q-PCR to measure the expression of the stem cell core factor genes Oct4, Sox2, and Nanog. As compared with the vector control and the untreated cells, cells that overexpressed Peln1 had decreased expression of these three pluripotent-associated factor genes (Fig. 3 B). We also quantitated three lineage markers in these cells and found that they expressed relatively high markers in the endoderm lineage (Fig. 3 C). The colony formation assay also showed reduced self-renewal capacity in Peln1-overexpressing E14 cells (Fig. S3 B).

Since Peln1 is negatively associated with stem cell pluripotency, we examined whether knockdown of Peln1 was able to enhance pluripotent reprograming. Four Peln1 shRNAs (shRNA-1, shRNA-2, shRNA-3, and shRNA-4) were cloned into two separate lentiviral vectors (shPeln1-1 and shPeln1-2; Fig. S3 C). The pGreenPuro Peln1–shRNA lentiviruses were transduced into MEFs that carried the doxycycline (DOX)-inducible Oct4-Sox2-Kilf4-c-Myc (Carey et al., 2010). Two control lentiviral groups were used in parallel with the shPeln1 group, including a mock virus that carries the empty vector (Vector) and the control virus that carries random shRNAs (shCT). The copGFP reporter gene in the lentiviral vector was used as the green fluorescence tracking marker for lentiviral infection.

After lentiviral transfection, MEFs were selected by puromycin. Pluripotency was induced by the addition of the reprograming inducer DOX. On day 12 of reprograming, iPSC colonies started to form. The efficiency of iPSC colony formation was assessed by immunostaining of the pluripotency biomarker NANOG (Fig. 4 A, red). After Peln1 knockdown (Fig. 4 B), we observed significant increase in the number of NANOG-positive colonies as compared with the two control groups (Fig. 4 C). On day 3 after lentiviral infection, cells were collected to confirm the Peln1 knockdown by Q-PCR. Peln1 knockdown was accompanied with the upregulation of the stem cell core factors Oct4, Sox2, and Nanog (Fig. 4 D). These data suggest that downregulation of Peln1 increases the efficiency of pluripotent reprogramming. However, downregulation of Peln1 alone was not sufficient to induce reprogramming in fibroblasts. The OSKM cocktail factors were still needed to initiate the pluripotency.

To gain insights into the mechanistic basis for Peln1 action, we focused on the lncRNA-mediated regulation of the Oct4 gene, one of the most important factors in the maintenance of the pluripotent state. Since Peln1 interacts with the Oct4 locus, we used Q-PCR to quantitate the Peln1 interaction using primers covering the whole Oct4 locus, including the promoter (sites E–F) and enhancers (sites A–B and sites M–N; Fig. 5 A, top panel). Peln1 primarily interacted with the Oct4 promoter and enhancer at E–F and M–N sites (bottom panel). Relatively few abundant background interaction signals were detected in the random control RAT group (Ctl).

Previous work has demonstrated the importance of specific intrachromosomal enhancer–promoter architecture for activation of endogenous pluripotency genes during reprograming (Zhang et al., 2013). As Peln1 binds to the Oct4 promoter and the 3′-enhancer, we hypothesized that this lncRNA might function by altering this intrachromosomal structure. We used chromatin conformation capture (3C; Dekker et al., 2002) to assess intrachromosomal looping. E14 cells were transfected with full-length Peln1. The empty vector served as the mock control. The interacting chromatin complex DNAs were detected by Q-PCR using the 3C primers located in the promoter and the enhancers of Oct4 (Fig. 5 B, top panel). We detected five intrachromosomal interaction products in untreated E14 cells and mock-treated E14 cells: JH4782/JH4783, JH4780/JH4783, JH4780/JH4785, JH4785/JH4794, and JH4788/JH4785. Overexpression of Peln1 abolished the intrachromosomal interaction signals at three sites (bottom panel), in parallel with the exit of E14 cells from pluripotency. We sequenced the 3C products and confirmed the presence of the ligated Mbo1 sites, flanked by the sequences from the promoter and enhancers of Oct4 (Fig. 5 C). These data suggest that, after binding to the regulatory elements of Oct4, Peln1 abolished these vital intrachromosomal interactions, leading to exit from pluripotency.

To further delineate the role of Peln1 in this abrogation of intrachromosomal looping, we examined the Hoogsteen base-pairing binding between the Peln1 lncRNA and the Oct4 DNA. The lncRNA-DNA interaction prediction indicated that nearly all the predicted Peln1–Oct4 base-pairing binding sites (TFO1–TFO10) were aggregated in a short fragment of the 5′-Peln1 (Fig. S3 E). We deleted a 95 bp fragment containing the putative base-pairing region. Interestingly, the base-pairing fragment-deficient mutant lost its role in the regulation of stemness (Fig. S3 F), suggesting the involvement of a base-pairing pathway in the Peln1–Oct4 interaction.

The Oct4 promoter is epigenetically regulated by DNA methylation of CpG islands. We compared the methylation status of two CpG islands in the Oct4 promoter. Genomic DNAs were isolated from the untreated iPSC control, vector control, and Peln1-overexpressing E14 cells. After sodium bisulfite treatment using an EZ DNA Methylation-Gold Kit, genomic DNAs in the Oct4 promoter were amplified and digested with restriction enzymes to separate the methylated and unmethylated CpGs. Overexpression of Peln1 induced DNA methylation at CpG1 and CpG2 sites in the Oct4 promoter, in parallel with the downregulation of the stem cell marker gene (Fig. 6 A). These data suggest that the lentivirus-induced overexpression of Peln1 bound to the Oct4 promoter and silenced it by altering the promoter epigenotype. In fibroblasts, however, silencing of Peln1 only induced a slight change in CpG methylation at the Oct4 promoter (Fig. S3 D), suggesting that other factors, in addition to Peln1, may be also required for the maintenance of DNA methylation in fibroblasts.

The status of DNA methylation in the Oct4 promoter is controlled by the three active DNA methyltransferases, DNMT1, DNMT3A and DNMT3B. We found that Dnmt3a was significantly upregulated in Peln1-overexpressing E14 cells (Fig. S4 A), although it did not result in a significant change in EB differentiation (Fig. S4 B). We then used RNA immunoprecipitation (RIP) to determine if the binding of Peln1 to the Oct4 promoter might recruit this DNA methyltransferase, thereby accounting for the induced DNA demethylation. The chromatin complex in Peln1-overexpressing E14 cells was immunoprecipitated with an anti-DNMT3A antibody. After RNA purification and reverse transfection, we quantitated the presence of Peln1 lncRNA in the immunoprecipitated products using primers that cover the different locations of Peln1 lncRNA and showed that DNMT3A interacted primarily with the 3′-fragment of Peln1 (Fig. 6 B). However, expression of Peln1 did not recruit DNMT3B and DNMT1 to the Oct4 promoter (Fig. S4 C).

We confirmed this interaction by incubating DNMT3A recombinant protein with different biotin-labeled Peln1 lncRNA fragments. After streptavidin bead pulldown, Western blotting was used to detect the interaction with DNMT3A. In agreement with the above RIP assay, we found that Peln1 primarily utilized its 3′-terminal 100 bp fragment to recruit DNMT3A (Fig. 6 C).

We also used chromatin immunoprecipitation to validate the binding of DNMT3A at the Oct4 locus. Peln1-overexpressing E14 cells and control cells were immunoprecipitated with a DNMT3A antibody. The immunoprecipitated chromatin DNAs were purified and examined for binding of DNMT3A. No significant enrichment signals were detected in the untreated and vector control cells. However, we detected binding of DNMT3A at the Oct4 promoter in a pattern similar to the binding of Peln1 (Fig. 6 D). Expression of the Peln1 3′-terminal–deleted mutant failed to alter CpG methylation (Fig. S5 A), recruit DNMT3A (Fig. S5 B), or downregulate the Oct4 in E14 cells (Fig. 6 E). These data suggest that Peln1 plays a dual role in the epigenetic regulation of Oct4, upregulating DNMT3A at the transcriptional level and recruiting DNMT3A enzyme to the Oct4 promoter.

As a new class of epigenetic regulators, lncRNAs play a critical role in the maintenance of stem cell pluripotency. Depending on the intracellular localization, an lncRNA may regulate pluripotency maintenance using different mechanism (Chen et al., 2020). Nuclear lncRNAs usually regulate pluripotency-associated marker genes by interacting with chromatin modification factors, RNA binding proteins, or transcription factors. In contrast, cytoplasmic lncRNAs are functionally different from nuclear lncRNAs and mainly act as posttranscriptional regulators by modulating mRNA degradation and protein translation.

However, the molecular components of the epigenetic regulatory circuity in pluripotency exit have not been fully characterized. Exit from pluripotency and stem cell differentiation require precise regulation of this pluripotency network. In this study, we profiled pluripotency exit–associated lncRNAs by combining three sequencing datasets: the lncRNA interaction network in the stem cell core factor gene Oct4, the differential transcriptome during pluripotent reprogramming, and the lncRNA genome-wide gene target interactome. Using this approach, we identified Peln1 as a critical lncRNA that determines the exit of E14 cells from pluripotency. Peln1 is actively transcribed in fibroblasts and tissues and becomes silenced in pluripotent reprograming. Overexpression of this lncRNA induced the exit of E14 cells from pluripotency. In contrast, knockdown of Peln1 enhances reprogramming efficiency. Peln1 triggers exit from pluripotency through the dual mechanism of upregulating Dnmt3a and recruiting it to the Oct4 promoter, thereby altering the promoter epigenotype and abrogating pluripotency-specific intrachromosomal architecture. This study provides the first evidence that a lncRNA can epigenetically control the exit from stem cell pluripotency by trans targeting the Oct4 promoter (Fig. 6 E).

In stem cells, the transcriptional regulatory circuity that controls the ground pluripotent state is maintained by coordinated expression of the core transcription factors Oct4, Sox2, and Nanog. The exit from pluripotency requires the dismantling of this gene expression circuity. As a critical stem cell core transcription factor, Oct4 is actively transcribed in pluripotent stem cells, but is transcriptionally silenced in somatic cells. We hypothesized that a pluripotency exit–associated lncRNA would have the ability to negatively regulate the Oct4 promoter. We demonstrate that Peln1 interacts with the Oct4 promoter, recruiting the DNA methyltransferase DNMT3A, which induces de novo DNA methylation and thereby inhibits promoter activity. Decreased expression of Oct4 initiates the exit of stem cells from pluripotency.

Currently, it is not clear how this lncRNA recognizes its target genes. Using LongTarget lncRNA–DNA interaction prediction software (He et al., 2015), we found that nearly all the predicted Peln1–Oct4 binding sites were aggregated in a short fragment of 5′-Peln1. Deletion of this short fragment caused the loss of this Peln1 function. Together, these data suggest that this lncRNA may use a Hoogsteen base-pairing mechanism to bind to the Oct4 promoter. However, other chromatin factors, like CTCF and Cohesin, might be also involved in this lncRNA–chromatin DNA interaction.

Exit from pluripotency and priming for differentiation into somatic lineages is associated with genome-wide de novo DNA methylation. This is a dynamic regulatory system, which is critical for coordinating pluripotent exit with early lineage specification. Exit from the ESC state requires dissolution of the core pluripotency transcription-factor circuit, but how the differentiation machinery breaks down this network is unclear. In response to differentiation cues, the cell can activate distinct lineage-specific gene networks while turning off or rewiring pluripotency networks. The molecular components that coordinate this pluripotent exit are poorly understood.

The maintenance of the pluripotent identity also requires a precisely regulated 3D chromatin interaction network composed of intra- and interchromosomal loops and bridges (Denholtz et al., 2013; Ji et al., 2016; Sexton and Cavalli, 2013). By comparing the “non-iPSCs” that failed to complete pluripotent reprogramming, we found that the virally expressed OSKM factors bound to their target genes, but could not activate them to achieve pluripotency due to the lack of a promoter-enhancer intrachromosomal loop architecture (Zhang et al., 2013). Formation of this intrachromosomal loop is a critical epigenetic barrier that must be overcome for induction of pluripotency, as it juxtaposes the distal enhancer in close proximity with the promoter for activation. Notably, the CRISPR Cas9-enforced formation of functional intrachromosomal looping alters the epigenotypes in the Oct4 promoter, thereby activating the pluripotent gene (Morgan et al., 2017). Using the same dCas9-ABA–inducible system, Wang et al. (2021) recently showed that this artificial chromatin linking enhanced reprogramming efficiency. However, it is not clear how DNA methylation is involved with the pluripotency chromatin loop in Oct4. Peln1 actively participates in the regulation of this pluripotency-specific 3D interaction, acting as a regulatory chromatin factor. After binding to the Oct4 promoter, Peln1 disrupts this intrachromosomal interaction, leading to exit from pluripotency. These results suggest a master regulatory role of Peln1 in maintaining pluripotency of stem cells.

In summary, the ability of a stem cell to self-renew or to exit from pluripotency is controlled by a tightly regulatory circuity consisting of transcriptional factors, chromatin architecture, and epigenetic modifications. In this study, we have provided the first evidence that lncRNA Peln1 acts as a new epigenetic player in this regulatory circuity. By binding to numerous pathway genes, particularly Oct4, Peln1 plays a “fine-tuning” role in shutting down the regulatory network conferring pluripotency and ultimately induces dissolution of the pluripotent state in E14 cells. The findings presented here may have broad implications for other chromatin-associated lncRNAs in the epigenetic regulation of stem cell pluripotency. Most notably, by quantitating three lineage markers in Peln1-expressing E14 cells, we noticed that Peln1 induced cell differentiation in favor of the endoderm lineage by expressing higher endoderm markers, including PDX1 and NKX6-1 (Fig. 3 C). This useful finding may suggest its application in regenerative medicine. Future studies are encouraged to determine if Peln1 can promote the formation of endoderm-related organs, such as pancreatic β cells in the treatment of diabetes.

Mouse fibroblasts and 293T cells were cultured in DMEM containing 10% (vol/vol) FBS (Sigma-Aldrich), 1% (vol/vol) of penicillin–streptomycin (Sigma-Aldrich) at 37°C in a 5% CO₂ air atmosphere. E14 and IPS cells were cultured in knockout DMEM containing 15% (vol/vol) KnockOut SR (Gibco), 1% (vol/vol) of penicillin–streptomycin (Sigma-Aldrich), 200 µMβ mercaptoethanol, 1,000 U/ml lif (Millipore), 1XMEM- NEAA at 37°C in a 5% CO₂ air atmosphere.

E14 mouse ESCs in the naive status were purchased from ATCC. Pluripotent cells were induced into the primed status by culturing E14 cells under culture conditions including FGF2 (12 µg/ml) and Activin A (20 µg/ml) in the absence of LIF (Tosolini and Jouneau, 2016).

To profile the Oct4 promoter-interacting lncRNAs, we constructed the Cas9–Oct4 gRNA vector by cloning two Oct4 promoter gRNAs into the lenti Cas9 gRNA vector that contains the catalytically inactive Cas9 (dCAs9). Oct4-gRNA1: 5′-GAA​CAT​TCA​ATG​GAT​GTT​TT-3′ and Oct4-gRNA2: 5′-GTG​TGA​GGG​GAT​TGG​GGC​TC-3′ (Table S1). For CRIST-seq control, we performed a CRIST assay using a random gRNA (gCT5′-GAA​GTG​GGA​TGA​TCC​TCT​GA-3′; Table S1). The dCas9–gRNA lentiviruses were packaged in 293T cells.

Fibroblasts were transfected with dCas9–gRNAs lentiviruses produced by 293T cells. After transfection, cells were selected with puromycin and collected for immunoprecipitation. Cells were cross-linked with 2% formaldehyde and lysed with cell lysis buffer (10 mM Tris [pH 8.0], 10 mM NaCl, 0.2% NP-40, and 1× protease inhibitors). After centrifugation, DNA was reverse transcribed at 37°C for 30 min in a 20-μl reaction (1 μl random hexamer, 1 μl 10 mM dNTP, 1 μl 0.4 mM biotin–dCTP, 1 μl RT enzyme, 0.5 μl RNase inhibitors, 1 μl 0.1 M DTT, 4 μl 5× cDNA synthesis buffer, 1 μl Maxima Reverse Transcriptase, and RNase-free water to 20 μl). The reaction was stopped by adding 4 μl 0.5 M EDTA. The chromatin complex was subjected to sonication for 180 s (10 s on and 10 s off) on ice with a Branson sonicator with a 2-mm microtip at 40% output control and 90% duty cycle settings. The biotin-labeled cDNA/Cas9 complex was immunoprecipitated with anti-FLAG antibodies (#MA1-91878; Thermo Fisher Scientific, IL, or #F7425; Sigma-Aldrich). The DNAs were released after cross-linking reversal and proteinase K treatment. The biotin-labeled cDNAs were further purified from genomic DNAs with M-280 streptavidin beads (Invitrogen). The Oct4 promoter–interacting cDNAs were subjected to Illumina sequencing (Zhang et al., 2019). The lncRNA–Oct4 interaction was quantitated by Q-PCR using target gene primers (Table S1).

For RNA-seq, total RNA was isolated from fibroblasts and iPSCs (Zhai et al., 2015; Zhang et al., 2013) using TRIzol (Invitrogen). RNA-seq was performed by Shanghai Biotechnology using a HiSeq4000 (Illumina). Differentially expressed RNAs in fibroblasts and iPSCs were defined using a cutoff threshold of more than twofold, P < 0.05 (Du et al., 2018).

A RAT assay (Du et al., 2018) was modified to identify the interacting target genes of Peln1. Specifically, cells were cross-linked with 2% formaldehyde and lysed with cell lysis buffer (10 mM Tris [pH 8.0], 10 mM NaCl, 0.2% NP-40, and 1× protease inhibitors). Gene strand–specific reverse transcription was performed in the presence of biotin–dCTP. After 30 min of reverse transcription of Peln1 lncRNA with Maxima Reverse Transcriptase (Thermo Fisher Scientific) at 62°C, the reaction was stopped by adding 4 μl 0.5 M EDTA. The chromatin complex was then subjected to sonication for 180 s (10 s on and 10 s off) on ice with a Branson sonicator with a 2-mm microtip at 40% output control and 90% duty cycle settings. The biotinylated–lncRNA–cDNA/chromatin DNA complex was pulled down with biotin–streptavidin magic beads (Invitrogen). After reversing the cross-links and treatment with 10 mg/ml proteinase K at 65°C for 2 h, the genomic DNA that interacted with the lncRNA was extracted, digested by MboI, and ligated with the NEBNext adaptors (NEBNext ChIP-Seq Library Prep Master Mix Set for Illumina) to construct the library. The library DNAs were subjected to Illumina sequencing (Shanghai Biotechnology). We performed a RAT assay by replacing Peln1 complementary primers with random primers for RAT-seq control, and constructed a control library for sequencing.

After RAT sequencing, the low-quality reads were filtered. Clean reads were mapped to the mouse genome (genome version: mm10). By comparing the RAT-seq peaks with input samples, enriched regions of the genome were identified using MACS2 (version: 2.1.1). The upstream 2 kb of the transcription start sites and the downstream 5 kb of the transcription termination region were defined as the gene regions. The significant GO terms of biological processes with a P value < 0.05 were selected. To reduce the background, the RAT-seq data were further normalized over the peaks of the control RAT-seq data that were generated by using random oligonucleotide primers in the RAT assay.

Reprogramming-associated RNA candidates were identified by RNA-seq using a cutoff threshold of more than twofold, P < 0.05. Peln1 target genes were identified using a RAT. The top 40 genes, sorted by expression differences from high to low in iPSC, were selected for further study.

The full-length Peln1 lncRNA was characterized by cDNA 5′- and 3′-RACE (Sun et al., 2014). For 3′-end RACE, we used the 3′-RACE system for rapid amplification of cDNA ends (1446137; Invitrogen). For 5′-RACE, the Fib RNAs were reverse transcribed using Maxima Reverse Transcriptase with random hexamer oligonucleotides for 30 min. The poly G primer J526: 5′-CCA​GAT​TCA​GGA​CTG​TCG​ACA​TCG​AAT​TCG​GGG-3′ was added to the mixture and the reaction was allowed to continue for another 30 min. After 10-fold dilution, the first RACE PCR was performed using a 3′-RACE primer J527: 5′-GAT​TCA​GGA​CTG​TCG​ACA​TCG​A-3′ and a 5′-Peln1-specific primer JH4980: 5′-CTC​CTA​TTA​TGT​ACT​ACA​GAG​TTC-3′. DNAs were diluted 500-fold and were used for the second nested PCR using a 3′-RACE primer J527: 5′-GAT​TCA​GGA​CTG​TCG​ACA​TCG​A-3′ and 5′- Peln1-specific primer JH5149: 5′-AGC​CCC​TAC​TGG​TAC​TAT​GGT​A-3′. The RACE PCR bands were cut out and cloned into a pJet vector for sequencing.

Cells were spun down in Falcon tubes for 5 min at 3,000 rpm in a Beckman tabletop centrifuge, washed once with PBS, and spun again for 5 min at 3,000 rpm. After completely aspirating the PBS, 800 μl of hypotonic buffer (10 mM Hepes, pH 7.9, 1.5 mM MgCl₂, and 10 mM KCl) was added to each sample, the pellet was gently resuspended, transferred to a microfuge tube, and placed in ice for 2 min. Through the remaining steps of the protocol, samples were kept on ice and spun at 4°C. 10% Nonidet P-40 was added to a final concentration of 0.4% (35 μl). Samples were inverted a few times and spun at 3,000 rpm for 7 min. Supernatants (450 μl; cytoplasmic fractions) were collected for processing, and the remaining supernatants were discarded. Then, the pellet (nuclear fraction) was gently resuspended in 500 μl of hypotonic buffer and spun at 3,000 rpm for 2 min. This washing step was repeated three to four times. After removing the buffer following the last spin, the samples were briefly spun once again to remove the remaining supernatant from the cells. Both the cytoplasmic and nuclear fractions were processed following the RNA extraction protocol as recommended by the manufacturer.

The full length of Peln1 lncRNA was amplified by PCR and ligated into the EcoRV/EcoRI site in 1× Spanch2 vector. Peln1 lncRNA was knocked down by two separate shRNA lentiviral vectors (shPeln1-1 and shPeln1-2). The shRNA vectors were constructed by cloning four shRNAs into two separate pGreenPuro vectors (#SI505A-1, SBI), each containing two shRNAs. For cloning, shRNAs were linked to the H1 and U6 promoters using PCR and were ligated into the EcoR1/BamH1 site in pGreenPuro vector (SBI). The copGFP reporter in the vector was used to track the lentivirus transfection in E14 cells.

MEFs were derived from transgenic mice that carried DOX-inducible OSKM genes. The four reprogramming factors, OSKM, were expressed from the collagen type 1 gene locus upon induction with DOX (Carey et al., 2010). MEFs were cultured in MEF medium (DMEM supplemented with 10% FBS, 1% P/S, 1% GlutaMAX, and 1% NEAA), according to standard methods at 5% CO₂ and 37°C (Kleger et al., 2012). The Peln1 shRNA1-1 and shRNA1-2 and control shRNA lentiviruses were packaged in 293T cells using lipofectamine 3000 (Invitrogen). The virus-containing supernatants were collected and concentrated with Centrifugal Filter Units (Amicon Ultra-15; Millipore). MEFs in 6-well plates were infected with lentiviruses. 3 d after infection, the cells were digested and transferred to a p12 plate at 15,000–20,000 cells/well. On the following morning, the medium was changed into serum-KSR iPS medium (10% FBS, 10% KSR, 1% P/S, 1% GlutaMAX, 1% NEAA, 1% sodium pyruvate, 0.1% β-mercaptoethanol, 1,000 U/ml LIF) containing 2 μg/DOX. iPSC medium was changed every other day for the first 3–5 d before cells become confluent and every day thereafter. On day 12–15, immunostaining was performed, and the Nanog positive iPSC colonies were counted.

Fibroblasts were collected for immunoprecipitation. In order to map the DNMT3A-lncRNA interactome, the biotinylated lncRNA-cDNA/Dnmt3a complex was immunoprecipitated with anti-DNMT3A antibodies (ab88470; Abcam). The biotinylated lncRNA–cDNAs were further purified from genomic DNAs using M-280 streptavidin beads (Invitrogen; Zhang et al., 2014b). After purification, the DNMT3A-interacting lncRNA–cDNAs were quantitated by PCR using target gene primers.

The 2 kb of Oct4 promoter was amplified and then inserted into the pGL3vector (Promega) to create a luciferase reporter construct, which was named pGL3-OCT4. E14 cells (1 × 10⁶) were cotransfected with 500 ng of pGL3-OCT4, 2 µg of Peln1-overexpressing plasmid, and 100 ng of a Renilla luciferase expression construct pRL-TK (Promega) using Lipofectamine 3000. Luciferase activity was then measured with the dual luciferase assay system (Promega) after 48 h of transfection. Renilla luciferase activity was used as the internal control for normalization.

Immunofluorescence staining was used to examine stem cell markers in iPSC colonies (Chen et al., 2012). Briefly, cells were fixed by 4% paraformaldehyde/PBS for 10–15 min and rinsed with PBS, then permeabilized and blocked with 0.1% Triton X-100/PBS containing 3% BSA for 30 min. After incubation with primary antibodies for 1 h at room temperature or overnight at 4°C, the samples were washed three times with PBS, and then incubated with secondary antibody for 1 h. The following antibodies were used in the immunostaining: rabbit anti-NANOG (sc-33759, 1:100 dilution; Santa Cruz Biotechnology) and rabbit anti-OCT4 (AB3029, 1:100 dilution; Millipore). The cell samples were subsequently incubated with Cy3- or Alexa Fluor 488–labeled secondary antibodies for 1 h. After washing three times with PBS, samples were counterstained with Hoechst 33258 (Invitrogen). Alternatively, the pluripotency of stem cells was examined using a Fluorescent Mouse ES/iPS Cell Characterization kit (Cat. no. SCR077; Millipore) following the protocol provided by the manufacturer. Fluorescence images were acquired with a Zeiss AxioCam Camera.

For embryoid body (EB) formation, iPSCs or E14 cells were trypsinized by collagenase IV (Invitrogen), and cell clumps were transferred to 60-mm dishes in ES medium without LIF. After being maintained in floating culture for 3 d, the EB was seeded in 0.1% gelatin-coated 6-well plates in DMEM/F12 containing 20% FBS to allow for spontaneous differentiation. Cells were collected at different time points and used for gene expression analysis using Q-PCR.

Genomic DNAs were extracted from E14 cells, E14+control vector cells, and E14+Peln1-overexpressing cells and treated with a EZ DNA Methylation-Gold Kit (Cat. no. D5005). Methylation PCR was performed, and PCR DNAs were cloned into pJet vector for sequencing (Zhang et al., 2014a). After treatment with the Methylation Kit, unmethylated cytosines were converted to uracils, which can be distinguished from the methylated cytosines by sequencing (Ma et al., 2014; Zhao et al., 2016).

The chromosome conformation capture (3C) assay was performed to determine intrachromosomal interactions (Du et al., 2021). Briefly, fibroblasts and iPSCs were cross-linked with 2% formaldehyde and lysed with Hi-C Lysis buffer (10 mM Tris-HCL, pH 8.0, 10 mM NaCl, 0.2% NP-40, 1× protease inhibitors). An aliquot of nuclei (2 × 10⁶) was digested with 100 U Mbo1 at 37°C 2 h. Chromatin DNA was diluted with NEB ligation buffer and ligated with 2,000 U of T4 DNA ligase. After reversing the cross-links, DNA was purified and used for PCR amplification using primers that are derived from different regions of the Oct4 locus. The 3C PCR products were cloned and sequenced to validate the intrachromosomal interaction by determining the presence of the Mbo1 ligation site (Li et al., 2008).

Total RNA was extracted with TRIzol reagent (Sigma-Aldrich) from cells and stored at −80°C. RT-PCR was performed with a Bio-Rad Thermol Cycler. The target amplification was performed by PCR of 1 cycle at 95°C for 5 min, 33 cycles at 95°C for 20 s, 62°C for 15 s and 72°C for 15 s, and 1 cycle at 72°C for 10 min. Quantitative RT-PCR was performed using SYBR GREEN PCR Master (Applied Biosystems). The threshold cycle (Ct) values of target genes were assessed by Q-PCR in triplicate using a sequence detector (ABI Prism 7900HT; Applied Biosystems) and were normalized over the Ct of the β-actin control.

The DNMT3A–Oct4 interaction was measured by chromatin immunoprecipitation (ChIP; Schulz and Haussler, 2014). One million E14 cells, E14 cells with 1xSpanch2 vector, and E14 cells with Peln1 overexpression were harvested. The DNA/Dnmt3a complex was immunoprecipitated with anti-Dnmt3a antibodies (ab88470; Abcam). After purification, the Dnmt3a-interacting DNAs were quantitated by PCR using target gene primers (Table S1).

Peln1 lncRNA was synthesized using a HiScribeTM T7 Quick High Yield RNA Synthesis Kit (NEB) with biotin–CTP. The biotinylated Peln1 lncRNA was purified for the following RNA–protein binding assay. Briefly, The EP tubes containing prepared streptavidin agarose are refilled with 500 μl of streptavidin washing buffer (20 mM Tris–HCl, pH 7.5, 100 mM KCl, 2 mM EDTA, 0.5 mM DTT, and 0.5 mM PMSF) and gently pipetted 4–5 times with an end-cut blue tip. 1 μl (100 pmol) of biotin-labeled RNA is applied to the EP tubes except for the EP tubes that served as negative controls. EP tubes are incubated at 4°C for 1–4 h under 100 RPM rotation in a rotator. After washing twice with 500 μl of the RNA–streptavidin interaction buffer (20 mM Tris–HCl, pH 7.5, 300 mM KCl, 0.2 mM EDTA, 0.5 mM DTT, and 0.5 mM PMSF), 1 μg of Dnmt3a protein is applied to each of above EP tubes. The reaction was maintained at 4°C overnight (about 12 h) under 100 RPM rotation. On the following day, the preparation is washed twice with RNA–streptavidin interaction buffer. 16 μl of 1× protein loading dye is applied to each EP tube obtained from above procedure. The EP tubes are heated at 80°C for 5 min and then the Peln1-interacting DNMT3a was detected by Western blotting.

In both CRIST and RAT assays, quantitative RT-PCR was used to compare the enrichment signals between treatment groups. Q-PCR was performed using SYBR GREEN PCR Master (Applied Biosystems) in triplicate using a sequence detector (ABI Prism 7900HT; Applied Biosystems; Bao et al., 2015; Li et al., 2014). The enrichment signals were calculated using threshold cycle (Ct) values standardized over the input, applying the 2^−(ΔCt) method.

All the experiments were repeated three times and were analyzed using SPSS software (version 16.0, Inc.). The Kolmogorov-Smirnov test was used to test the data normality. One-way ANOVA with Bonferroni’s multiple comparisons test was used to compare statistical differences for selected variables among treatment groups. Differences were considered significant for P values < 0.05 (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

Table S1 shows the oligonucleotide primers used for PCR and RAT assays. Fig. S1 maps the pluripotency exit–associated lncRNAs by CRIST-seq. Fig. S2 characterizes the Peln1 promoter and Peln1 expression in tissues. Fig. S3 examines the role of Peln1 and its mutant in the Oct4 promoter methylation and colony formation. Fig. S4 shows the data of DNA methyltransferases expression and ChIP–Oct4 assay. Fig. S5 shows that the Peln1 3′-mutant failed to induce DNA methylation in the Oct4 promoter.

As previously reported (Du et al., 2018), the CRIST-seq, RAT-seq, and RNA-seq data generated in this study have been deposited in NIH GEO databases under accession nos. GSE137769, GSE107945, GSE101765, and GSE116605, respectively.

This work was supported by the National Key R&D Program of China (2018YFA0106902, 2020YFA0707704), the Innovative Program of National Natural Science Foundation of China (82050003), the National Natural Science Foundation of China (31871297, 81874052, 81900701, 81900327, 82001670, 82101675, 32000431), the Fund of Jilin Provincial Science and Technology Department (20210101311JC, 20190303146SF, 20200602032ZP, 20200201390JC), the Provincial Science Fund of Jilin Province Development and Reform Commission (2021C10 and 2020C038-4), the Natural Science Fund of Jilin Provincial Finance Department (JLSWSRCZX2020-023, JLSWSRCZX2020-100), the Youth Fund of Jilin Provincial Health Commission (2016Q035), the Fund of Changchun City Science and Technology Bureau (21ZGY28), China Guanghua Fund and the Youth Fund of First Hospital of Jilin University (2020-CXM-01, JDYYGH2019004, JDYY 102019002, JDYY 102019034), California Institute of Regenerative Medicine (CIRM) grant (RT2-01942), and the Department of Veterans Affairs (BX002905).

The authors declare no competing financial interests.

Author contributions: Y. Wang and L. Jia designed and performed assays for the study. C. Wang and X. Wen performed cell reprogramming and cell culture studies. Z. Du, S. Zhang, and L. Zhou performed the CRIST-seq assay and RNA-seq. H. Li, H. Chen, Y. Nie, D. Li, and S. Liu conducted some Q-PCR and RT-PCR studies. D.S. Figueroa and F. Ay contributed to bioinformatics analyses of RNA-seq, CRIST-seq, and RAT-seq data. J.-F. Hu, A.R. Hoffman, W. Xu, S. Zhang, W. Li, J. Cui, and H. Guo contributed to the study design and supported the study. J.-F. Hu wrote the manuscript and A.R. Hoffman edited the manuscript. The authors read and approved the final manuscript.