Mechanisms and function of de novo DNA methylation in placental development reveals an essential role for DNMT3B

Andrews, Simon; Krueger, Christel; Mellado-Lopez, Maravillas; Hemberger, Myriam; Dean, Wendy; Perez-Garcia, Vicente; Hanna, Courtney W.

doi:10.1038/s41467-023-36019-9

Download PDF

Article
Open access
Published: 23 January 2023

Mechanisms and function of de novo DNA methylation in placental development reveals an essential role for DNMT3B

Nature Communications volume 14, Article number: 371 (2023) Cite this article

8193 Accesses
11 Citations
28 Altmetric
Metrics details

Subjects

Abstract

DNA methylation is a repressive epigenetic modification that is essential for development, exemplified by the embryonic and perinatal lethality observed in mice lacking de novo DNA methyltransferases (DNMTs). Here we characterise the role for DNMT3A, 3B and 3L in gene regulation and development of the mouse placenta. We find that each DNMT establishes unique aspects of the placental methylome through targeting to distinct chromatin features. Loss of Dnmt3b results in de-repression of germline genes in trophoblast lineages and impaired formation of the maternal-foetal interface in the placental labyrinth. Using Sox2-Cre to delete Dnmt3b in the embryo, leaving expression intact in placental cells, the placental phenotype was rescued and, consequently, the embryonic lethality, as Dnmt3b null embryos could now survive to birth. We conclude that de novo DNA methylation by DNMT3B during embryogenesis is principally required to regulate placental development and function, which in turn is critical for embryo survival.

Mapping genotypes to chromatin accessibility profiles in single cells

Article 08 May 2024

Systematic epigenome editing captures the context-dependent instructive function of chromatin modifications

Article Open access 09 May 2024

On the genetic basis of tail-loss evolution in humans and apes

Article Open access 28 February 2024

Introduction

DNA methylation typically functions as a repressive epigenetic modification and has been extensively studied for its roles in development, health and disease. Functional investigations into the role of DNA methylation have taken place since the 1990’s, yet establishing its role comprehensively in the mammalian placenta has not been undertaken. DNA methylation exerts its repressive function by sterically interfering with access for transcription factor-DNA interactions or by recruiting DNA methylation-binding repressor complexes to chromatin^1,2. The gene silencing function of DNA methylation has been demonstrated for key regulatory regions conferring genomic imprinting and X chromosome inactivation, and at specific genomic targets such as CpG-rich repetitive elements and germline-specific genes³. Despite these specific regulatory roles, the majority of mammalian cells have high levels of DNA methylation across the entire genome, with the exception of CpG islands (CGIs) and active regulatory elements. There are two phases of development when widespread de novo DNA methylation occurs, during gametogenesis and post-implantation embryonic development. In both contexts, DNA methylation is conferred by de novo DNA methyltransferase (DNMT) 3A and 3B enzymes⁴, with contributions in gametogenesis from catalytically inactive cofactor DNMT3L and rodent-specific DNMT3C^5,6. Early knockout (KO) studies in mouse models have demonstrated that de novo DNA methylation is critical for development as Dnmt3a, Dnmt3b and Dnmt3a/b DKO embryos die in mid-gestation or shortly after birth⁷. However, recent mapping of embryonic lineage specification in de novo Dnmt KO embryos in exquisite detail using single-cell technologies did not identify dramatic effects on lineage commitment or cell identity in gastrulation-staged embryos⁸. Furthermore, transcriptional de-repression in de novo Dnmt KO post-gastrulation embryos was largely restricted to germline genes and specific subclasses of repetitive elements⁹. Thus, the molecular consequences underpinning the role for de novo DNA methylation in embryo survival remains an outstanding question.

Placental trophoblast cells have a unique genome-wide DNA methylation pattern, with the majority of the genome being spanned by large partially methylated domains (PMDs), unlike the highly methylated genomes of somatic cells. This enigmatic feature appears to be conserved among placental mammals and is maintained throughout pregnancy^10,11,12, and remarkably, is shared with many cancers¹³. While the importance of imprinted DNA methylation in placentation is well-established^14,15, the role for the genome-wide DNA methylation landscape acquired during placentation has yet to be established. In this study, we investigate the impact of de novo DNMTs on the trophoblast lineage during embryogenesis, revealing the importance of DNA methylation in genome regulation and placental development beyond genomic imprinting. Our findings show that the deposition of DNA methylation by DNMTs is influenced by the chromatin landscape in trophoblast cells and is not only essential for the formation of the maternal-foetal interface during placentation, but is critical for embryo survival.

Results

DNMT3A, 3B and 3L establish the trophoblast methylome

DNMT3A, 3B and 3L have been reported to contribute to DNA methylation at CGIs (defined as CpG-dense regions that comprise about 1% of the genome) in trophoblast cells¹⁶. To date, the role of de novo DNMTs in establishing the genome-wide DNA methylation landscape in trophoblast remains unknown. Studies in the mouse epiblast and visceral endoderm have shown that de novo DNA methylation occurs between embryonic days (E)4.5 and 7.5;¹⁷ hence, we evaluated Dnmt expression and DNA methylation in E7.5 extra-embryonic ectoderm (ExE) comprising trophoblast cells, and contrasted these to E7.5 epiblast. Assessing the DNA methylation landscape genome-wide is necessary to understand which genomic elements are targeted by specific DNMTs and how this DNA methylation may regulate the underlying genomic features.

Using published RNA-seq data¹⁸, we found that Dnmt3a, Dnmt3b and Dnmt3l are all highly expressed in ExE, with Dnmt3a and Dnmt3b showing approximately two-fold lower expression than in the epiblast and Dnmt3l showing 20-fold higher expression (Supplementary Fig. 1a). Using low-input post bisulphite adaptor tagging (PBAT), we assayed DNA methylation in E7.5 ExE and epiblast from Dnmt3a KO, Dnmt3b KO, Dnmt3a/b double KO (DKO), Dnmt3l KO and wildtype (WT) controls (Fig. 1a–c, Supplementary Data 1). When comparing genome-wide profiles of DNA methylation in E7.5 ExE using 100-CpG windows, all KOs cluster distinctly from WT ExE (Fig. 1b, c). Differentially methylated regions (DMRs) were identified using logistic regression and >20% difference in DNA methylation, demonstrating that there was a loss of DNA methylation in the ExE of Dnmt3a KO, Dnmt3l KO and Dnmt3b KO at 0.7%, 34.6% and 58.9% of autosomal 100-CpG windows, respectively (Supplementary Fig. 1b). Notably, these differences were markedly more pronounced than those observed in the corresponding KO epiblast tissues (0.2%, 0.1%, and 25.3% DMRs, respectively) (Fig. 1b, c, Supplementary Fig. 1c). Paradoxically, DNMT3A appears to be simultaneously dispensable and yet remarkably competent to de novo methylate; in other words, there was almost no loss of DNA methylation in the absence of DNMT3A, but, in the absence of DNMT3B, DNMT3A can still establish a significant amount of genome-wide DNA methylation (Fig. 1b, c). DNMT3L shows an ExE-specific role in facilitating de novo DNA methylation, as Dnmt3l KO epiblast clustered with WT while Dnmt3l KO ExE clustered near Dnmt3b KO by PCA (Fig. 1b). In Dnmt3a/b DKOs, both ExE and epiblast fail to establish tissue-specific DNA methylation patterns and hence cluster together with publicly available data from WT E3.5 inner cell mass and trophectoderm samples¹⁶ (Fig. 1b, c). These data support that all DNMTs are required for the trophoblast methylome, albeit at highly discrepant levels.

**Fig. 1: DNMT3A, 3B and 3L mediate de novo methylation in trophoblast cells.**

Targeting of DNMTs is linked to the chromatin landscape

The de novo DNMT enzymes have been shown to interact with modified histone tails, which modulates their catalytic activity and genome localisation in vitro^19,20,21, although evidence that histone modifications may influence DNMT targeting in vivo remains limited. Thus, we generated ultra-low input ChIP-seq data for H3K4me1 and H3K27ac in E6.5 ExE to combine with published datasets for H3K4me3, H3K27me3 and H3K36me3¹⁸, immediately preceding the completion of de novo DNA methylation (Fig. 2a, Supplementary Data 1). Using UMAP dimensionality reduction and clustering, we identified chromatin features genome-wide for 100-CpG windows, which were then annotated based on abundance of histone marks within each cluster (Fig. 2b, c). Functional genomic features have well-defined epigenetics signatures²² and, consistently, our annotated chromatin features robustly represent the expected underlying genomic state. H3K36me3-marked gene bodies were the most highly DNA methylated, whereas active promoters were predominantly unmethylated in E7.5 WT ExE (Supplementary Fig. 2a). Chromatin features indicative of a promoter state were most highly associated with annotated gene promoters, and the corresponding gene expression levels were consistent with the active or repressed state of each chromatin feature (Supplementary Fig. 2b-c).

**Fig. 2: Targeting of de novo DNMTs in ExE depends on underlying chromatin landscape.**

We then assessed whether 100-CpG windows that lost the most DNA methylation in Dnmt3a, Dnmt3b and Dnm3l KOs were enriched at distinct chromatin features (Supplementary Fig. 1b, Fig. 2d, Supplementary Data 2). Highlighted on each UMAP plot are each set of Dnmt KO DMRs, demonstrating whether underlying chromatin signatures are linked to the activity of DNMTs to these genomic regions (Fig. 2e). This analysis identified that regions that lose DNA methylation preferentially in the Dnmt3a KO were near inactive/active enhancers and active/bivalent promoters (Fig. 2d, e). In the Dnmt3b KO, H3K36me3-marked gene bodies, genic/active enhancers, and unmarked inter-genic regions (background) were affected, while in Dnmt3l KOs, it was almost exclusively unmarked inter-genic regions that were lost (Fig. 2d, e). Given the common enrichment for unmarked inter-genic regions, we directly compared DNA methylation between the Dnmt3l and Dnmt3b KOs, which were highly correlated (R = 0.92) (Supplementary Fig. 2d), suggesting that DNMT3L may act to modulate the localisation of DNMT3B to unmarked inter-genic regions in trophoblast cells. These findings support that DNMT activities are directed by chromatin states in trophoblast development, consistent with binding preferences seen in vitro^20,23.

DNA methylation represses germline genes in trophoblast

To determine the impact of loss of DNA methylation on gene expression, we then performed ultra-low input stranded RNA-seq in E7.5 ExE (Supplementary Data 1). The transcriptomes of Dnmt3a KO, Dnmt3b KO, Dnmt3a/b DKO, Dnmt3l KO and matched WT control ExE were remarkably similar, with no distinct clustering by PCA (Fig. 3a) and correlation coefficients >0.99 compared to WTs (Fig. 3b, Supplementary Fig. 3a-c). Using DESeq2 and >1.5-fold change in expression, differentially expressed genes (DEGs) were identified, with the Dnmt3a/b DKO being most affected with 124 DEGs (83 up-, 41 down-regulated) (Fig. 3b, Supplementary Data 3). Gene ontology analysis found that upregulated DEGs were highly enriched for germline genes, typically expressed during spermatogenesis and oogenesis, including piRNA pathway (e.g., Piwil2, Asz1), meiotic components (e.g., Sycp1) and gametic transcriptional regulators (e.g., Dazl, Sohlh2) (Fig. 3b, c). While there were markedly less DEGs identified in the other Dnmt KOs, Dnmt3b KO had 17 DEGs (14 up- and 3 down-regulated), of which all up-regulated were also present in the Dnmt3a/b DKO up-regulated DEGs (Supplementary Fig. 3d) and similarly showed significant enrichment for germline genes in gene ontology analysis (Supplementary Fig. 3e). Hence, the gene regulatory role for de novo DNA methylation in ExE appears to be predominantly to silence germline genes, consistent with findings in the embryo^4,9. We then assessed whether the de-repression of these germline genes is correlated with the relative loss of promoter DNA methylation. We plotted fold-change in gene expression against the difference in promoter DNA methylation for the 83 Dnmt3a/b DKO up-DEGs in each of the Dnmt KOs. This analysis shows that the extent of DNA methylation loss at these gene promoters was significantly linearly correlated with the up-regulation of gene expression (Fig. 3d).

**Fig. 3: Loss of DNA methylation in ExE results in de-repression of germline genes.**

DNA methylation is important for repression of repetitive elements in many cell contexts^4,6,7 and given the pervasive use of endogenous retroviruses (ERVs) in the trophoblast gene regulatory landscape^24,25, we evaluated differential transcription of ERVs in Dnmt KOs. RNA-seq libraries are only informative for a fraction of ERVs genome-wide, so we first filtered for those with at least one same-stranded read in at least one replicate and excluded those overlapping an annotated gene, generating a list of 8142 ‘informative autosomal ERVs.’ Using the LIMMA statistical package, we identified that a small subset of ERVs (N = 23) are significantly upregulated in the Dnmt3a/b DKO (Supplementary Fig. 3f). While there was no enrichment for a specific class of ERV (p = 0.1, Chi-square), those up-regulated ERVs were significantly longer in sequence length than the set of informative autosomal ERVs (p < 0.001, t-test) (Supplementary Fig. 3g-h), suggestive that up-regulated ERVs are relatively intact. Upregulation of ERVs in the Dnmt3a/b DKO was associated with substantial losses of DNA methylation (median difference=53.4%), which was significantly greater than observed in the Dnmt3a, Dnmt3b, and Dnmt3l KOs (median difference=2.3%, 15.7% and 13.0%, respectively) (Supplementary Fig. 3i) (ANOVA p < 0.0001, post hoc Tukey tests p < 0.01). We observed no significant difference in ERV de-repression between the Dnmt3a, Dnmt3b, and Dnmt3l KOs (post hoc Tukey test not significant for all comparisons), suggesting that substantial loss of DNA methylation is required for ERV de-repression.

Taken together, we find that de novo DNA methylation is important for repression of a remarkably small subset of the genome in E7.5 ExE, including germline genes and a set of ERVs.

Impaired placental labyrinth formation in Dnmt3b KOs

We next determined whether loss of DNA methylation in trophoblast cells impacted subsequent placental development. We evaluated phenotypic changes in Dnmt3a, Dnmt3l and Dnmt3b KOs compared to littermate WT control E12.5 placentas, a time point where the placenta is still developing but the labyrinth and junctional zone are formed. Dnmt3a/b DKOs were not assessed, as these embryos die by E10.5 and hence never form a functional placenta. We first evaluated placental morphology using haematoxylin and eosin staining, with no dramatic effects or apparent differences in size seen in Dnmt3a or Dnmt3l KOs compared to WTs (Supplementary Fig. 4a-b). However, Dnmt3b KO placentas were smaller, which we linked to an underdeveloped labyrinth zone (Fig. 4a, Supplementary Fig. 4c), the main nutrient and gas-exchange surface in the mouse placenta. Immunohistochemistry for E-Cadherin (CDH1) confirmed impaired development of the syncytiotrophoblast (SynT) within the labyrinth of Dnmt3b KO placentas (Supplementary Fig. 4d). To further explore the labyrinth defect, immunofluorescence analysis for monocarboxylate transporters 1 (MCT1) and 4 (MCT4) was used to demarcate the SynTI and SynTII layers, respectively, that separate the maternal and foetal blood circulation. We observed a substantial reduction in the SynT surface available for nutrient transport in Dnmt3b KO placentas, with the SynTII layer being most affected (Fig. 4b). The mis-formation of the maternal-foetal interface of the placental labyrinth was further corroborated by isolectin B4 staining, highlighting the overall foetal vasculature organisation, which confirmed that Dnmt3b KO placentas have an impaired vascularisation (Fig. 4c). The defects in SynT development and vascularisation indicate impaired foetal-maternal exchange²⁶ in Dnmt3b KO placentas, likely compromising foetal growth and survival.

**Fig. 4: *Dnmt3b* KOs show impaired formation of the placental labyrinth.**

To investigate the placental phenotype of Dnmt3b KO at the molecular level, we generated single-nuclei RNA sequencing data from E12.5 placentas (Supplementary Fig. 5a-b). Using published gene signatures for mouse placental cell types²⁷, we were able to identify all placental cell types in both the Dnmt3b KO and WT (Fig. 4d, Supplementary Fig. 5c-d). Consistent with the reduced size of the labyrinth compartment seen in Dnmt3b KO placentas, we observed a subtle decrease in the abundance of labyrinth trophoblasts, with a concomitant increase in junctional zone trophoblast and glycogen cells (Fig. 4e). However, the cell composition of Dnmt3b KO and WT placentas was highly correlated (R = 0.92) (Fig. 4e).

We then assessed whether there was evidence for dysregulated gene expression in Dnmt3b KO SynT cells compared to WT (Supplementary Data 4). Gene ontology analysis of upregulated DEGs (N = 36) showed similar results to those seen for Dnmt3a/b DKO and Dnmt3b KO E7.5 ExE, with a significant de-repression of germline genes in E12.5 SynT cells (Supplementary Fig. 5e). Notably, one significant gene ontology category was enriched among down-regulated DEGs (N = 18) in Dnmt3b KO SynT cells (Supplementary Data 4), the tissue category “placenta,” which included several genes that are highly expressed placenta (e.g., Cts3, Ascl4, Fn1, Ceacam14), but no specific pathway or processes was identified (Supplementary Fig. 5e). To determine whether de-repression of germline genes was specific to SynT cells or was observed across placental compartments, we measured the expression of the 83 upregulated DEGs identified in the Dnmt3a/b DKO ExE in E12.5 junctional zone and labyrinth, which both showed significantly upregulated expression in the Dnmt3b KO (Fig. 4f, Supplementary Fig. 5f). These data suggest that while differentiation does not appear to be affected in Dnmt3b KO placentas, there is widespread de-repression of germline genes and compromised expression levels of some placental genes.

In summary, loss of DNMT3B results in a severe impairment in the formation of the placental labyrinth, but this effect does not appear to be underpinned by a failure of trophoblast differentiation or cell identity.

Dnmt3b KO mid-gestation embryonic lethality is rescued with a functional placenta

The formation of the placental labyrinth is initiated when the allantois, a projection of extra-embryonic mesodermal cells derived from the epiblast, fuses with the chorionic ectoderm at E8.5²⁶. Predominantly thought to be mediated through WNT, FGF, BMP and HGF signalling, the invading mesoderm induces branching and differentiation of the chorionic trophoblast cells to form the two SynT layers²⁸. Therefore, defects in labyrinth formation may either originate from impaired allantoic signals, or from deficiencies intrinsic to the trophoblast compartment itself. To distinguish between these possibilities, we generated conditional Dnmt3b KO (cKO) conceptuses using Sox2-Cre, a transgenic mouse line that uses the endogenous Sox2 promoter to confer epiblast-specific Cre expression, which is widely used for the generation of embryo-specific KOs²⁹ (Supplementary Fig. 6a-b). Consequently, the Dnmt3b cKO results in conditional deletion of Dnmt3b in all embryonic lineages before post-implantation de novo DNA methylation (Fig. 5a). We confirmed the specificity of the cKO with PBAT of Dnmt3b cKO E7.5 epiblast and ExE and PCR genotyping of the Dnmt3b allele in E8.5 whole embryos and placentas (Supplementary Fig. 6c-d). Immunofluorescence analysis of MCT1 and MCT4 in Dnmt3b cKO and WT E12.5 placentas showed that the labyrinth defects seen in the Dnmt3b KO placentas were fully rescued in the Dnmt3b cKO (Fig. 5b, Supplementary Fig. 6e). We performed single-nuclei RNA-sequencing on Dnmt3b cKO and WT E12.5 placentas, demonstrating that gene expression profiles in the Dnmt3b cKO resemble WT (Fig. 5c). Therefore, we conclude that DNMT3B is critical in trophoblast cells, rather than the allantois, in mediating appropriate formation of the placental labyrinth.

**Fig. 5: *Dnmt3b* cKOs show rescued placental labyrinth phenotype and consequently survive to birth.**

We then assessed whether the presence of a functional placenta in the Dnmt3b cKO could rescue embryo survival. At E12.5, Dnmt3b cKO embryos showed a remarkably improved morphology compared to Dnmt3b KO embryos (Fig. 5d). Dnmt3b KO embryos die at approximately E13.5⁷, and consistently we recovered no Dnmt3b KO embryos in late gestation; yet, Dnmt3b cKO embryos survived until just before birth at E18.5, at expected Mendelian ratios (Fig. 5e, Supplementary Fig. 6f-g). Immunofluorescence of E18.5 placentas confirmed phenotypically normal placental labyrinth development in Dnmt3b cKOs throughout pregnancy (Supplementary Fig. 6h). Nevertheless, at E18.5, Dnmt3b cKO embryos showed significantly reduced foetal weight and no Dnmt3b cKO embryos survived to postnatal day 10 (Fig. 5e-f, Supplementary Fig. 6g). Hence, DNMT3B has a late-gestational role in the embryo proper impacting perinatal survival; the phenotype underpinning this perinatal lethality has not been studied here. Our findings demonstrate that impaired formation of the maternal-foetal interface in the placental labyrinth underlies the embryonic lethality observed in Dnmt3b KOs.

Discussion

In this study, we demonstrate that DNMT3A, DNMT3B and DNMT3L are all integral for the establishment of the placental methylome during post-implantation development. We also show that each DNMT preferentially targets a set of highly specific genomic features demarcated by characteristic chromatin states. Importantly, de novo DNA methylation plays a critical role in healthy placental development, with loss of DNMT3B resulting in impaired formation of the placental labyrinth causing mid-gestation embryonic lethality. Despite the striking phenotype of Dnmt3b knockout placentas, we observed no impairment of specification or differentiation of placental cell types. Rather, we find that the prominent transcriptional change in both undifferentiated and differentiated trophoblast cells was up-regulation of germline genes normally expressed in oogenesis and spermatogenesis. This finding mirrors observations made in embryonic cell types^4,8,9, and supports the notion that one of the primary roles of DNA methylation is to silence germline genes³⁰.

The DNMT proteins contain three histone-interacting domains: DNMT3A contains UDR, ADD and PWWP domains, DNMT3B contains ADD and PWWP domains and DNMT3L contains an ADD domain. Furthermore, the DNMTs can form hetero-dimers and oligomers, making the patterns of recruitment in biological contexts challenging to disentangle³¹. Indeed, studies investigating the targeting of de novo DNMTs in the embryo observed relatively uniform kinetics throughout the genome⁴, with later HiC studies showing that A compartments methylate before B compartments³². However, during epigenetic programming in trophoblast cells, we observe distinct genomic targeting preferences for DNMT3A, DNMT3B and DNMT3L activity, linked to histone modifications. DNMT3B-targeted sites show enrichment for H3K36me3-marked gene bodies in trophoblast cells, consistent with its localisation in vitro via its PWWP domain²⁰. Conversely, DNMT3L-dependent regions appear to be largely devoid of H3K4me and H3K36me3, consistent with ADD domain-directed localisation of DNMT3L to unmethylated H3K4¹⁹. DNMT3L enhances the catalytic activity of de novo DNMTs^33,34, thus one could speculate that its trophoblast-specific function may be necessary due to the lower expression levels of Dnmt3a and Dnmt3b compared to epiblast, enabling recruitment of DNMTs to low-affinity inter-genic sites.

While DNMT3A can bind to unmethylated H3K4 and H3K36me2/3 through its ADD and PWWP domains, respectively, we find that DNMT3A-targeted sites are most prominent at the boundaries of active and bivalent domains in trophoblast cells. Recent in vitro work has shown that the localisation of DNMT3A to bivalent regions of the genome is underpinned by an interaction of its N-terminal UDR domain with H2AK119ub in the absence of PWWP recruitment to H3K36me2 domains³⁵. This mechanism may be relevant in trophoblast cells, as it has been reported that H3K27me3-marked CpG islands, which often co-localises with H2AK119ub, are atypically methylated in ExE¹⁶. Interestingly, a recent study reported almost wildtype levels of DNA methylation and gene expression in E11.5 embryos with a catalytically-inactivated DNMT3B³⁶, suggesting that catalytically-inactive DNMT3B can heterodimerise with DNMT3A to restore DNA methylation patterning. While placental cells were not assessed in this study, the rescue of embryonic lethality in this model suggests that similar mechanisms may occur in trophoblast cells. Future studies will be valuable in understanding whether these various mechanisms of DNMT3A-recruitment differ across cell types. To date, the link between genome-wide de novo DNMT activity and chromatin landscape in vivo has only been well-documented in the context of epigenetic programming in gametogenesis^37,38,39. Our findings reveal that similar mechanisms may influence DNA methylation programming during trophoblast development, with chromatin-mediated recruitment of DNMTs likely explaining at least some of the unique features of the placental methylome.

The accumulation of DNA methylation in a lineage-specific manner in E7.5 epiblast and ExE is consistent with its suggested role in re-enforcing transcriptional repression associated with lineage commitment⁴⁰. However, we do not observe precocious activation of development genes in Dnmt KO ExE, nor impaired differentiation or tissue composition in the Dnmt3b KO placenta when assessed by single-nuclei RNA-sequencing. This finding was surprising given the markedly impaired formation of the SynTII layer seen in Dnmt3b KO placental sections. SynTII cells were defined in single nuclei analysis using published markers, including Gcm1, Synb, Gcgr and Vegfa²⁷, while the identification of SynTII cells by immunofluorescence was done using the widely adopted approach of staining for lactate proton-coupled monocarboxylate transporter protein MCT4 (Slc16a3)^41,42. Transcription factor, GCM1, becomes expressed in the chorionic trophoblast at the branchpoints of the allantoic mesoderm and embryos that lack Gcm1 fail to initiate differentiation and formation of the placental labyrinth²⁸. MCT4 is a lactate transporter expressed in differentiated SynTII cells and is essential in its function in maternal-foetal exchange⁴¹. Thus, we interpret our findings to show that specification of SynTII can occur in Dnmt3b KO placentas, but that the morphogenesis and function is severely compromised, resulting in a failure to support foetal development beyond E13.5. Our data suggests that the de-repression of germline genes is interfering with these processes in SynTII formation and function. However, an important limitation of single cell RNA-seq is its inability to detect low and intermediately expressed genes, thus while germline gene de-repression is the promint differential signature of Dnmt3b KO placental cells, further relatively subtle changes in the gene regulatory landscape will likely be overlooked. Intriguingly, impairment in branching morphogenesis and transcriptional de-repression of germline genes has also been observed in the lungs of Dnmt1 cKO mice⁴³, suggesting that failure to repress germline genes may interfere with branching morphogenesis processes across developmental contexts. Future studies will be valuable in gaining further insights into the molecular mechanisms underpinning these phenotypes.

DNA methylation is studied extensively in the context of human pregnancy for association with placental insufficiency and pregnancy complications, such as preeclampsia, growth restriction and preterm birth⁴⁴. Yet, there remain few examples of a functional role for DNA methylation in placentation outside the context of genomic imprinting and X chromosome inactivation. Our study reveals that germline genes and a subset of endogenous retroviruses are susceptible to de-repression upon loss of DNA methylation. Further, we identify formation of the placental labyrinth as a developmental process critically reliant on correct establishment of DNA methylation in trophoblast cells. Pregnancy complications affect as many as one in four women and maternal health during pregnancy has lasting impacts on the lifelong health of offspring. Our findings provide invaluable context for studies investigating environmental exposures in pregnancy and developmental origins of health and disease and will help direct future studies of epigenetic mechanisms underpinning human placental development.

Methods

Sample collections

The use of animals in this study was performed in accordance with the European regulation in Animals (Scientific Procedures) Act 1986 with all protocols approved by the Animal Welfare and Ethical Review Body at the Babraham Institute under licenses issued by the Home Office (UK).

Natural timed matings were used to collect embryos on embryonic days 6.5 (E6.5), E7.5, E12.5, and E18.5. We used published mouse models for Dnmt3l⁵, Dnmt3a⁷, and Dnmt3b⁷, and Sox2-Cre²⁹, which have been bred into the C57BL6/Babr background for this study. To generate KO and wildtype embryos, heterozygous females were bred to heterozygous males for Dnmt3l, Dnmt3a, Dnmt3b and Dnmt3a/3b strains. To generate Dnmt3b cKO embryos, homozygous Dnmt3b floxed females (Dnmt3b fl/fl) were bred to heterozygous Dnmt3b floxed males carrying the Sox2-Cre (Dnmt3b fl/+, Sox2-Cre +ve). For all strains used in this study, heterozygous animals are phenotypically normal.

At E6.5 and E7.5, the epiblast (Epi), extra-embryonic ectoderm (ExE) and ectoplacental cone (EPC) for each embryo were manually separated. Single E7.5 epiblast and ExE samples were individually frozen in 10 µL of buffer RLT Plus (Qiagen) for RNA-seq or PBAT; the corresponding EPC was used for genotyping. For ChIP-seq, pools of C57BL6/CAST E6.5 ExE (N = 8) samples were washed in PBS and flash frozen in 10 µL of nuclear lysis buffer (Sigma), matching the strain and collection approach for incorporated published datasets¹⁸. For placental phenotyping, E12.5 (and E18.5 for the Dnmt3b cKO) placentas were collected and fixed overnight in 4% paraformaldehyde at 4 °C, washed with PBS and stored at 4 °C until paraffin embedded. For E12.5 and E18.5 embryos, a tail clip was used for genotyping. For all molecular and phenotyping experiments, wherever possible, matched littermate wildtype controls were used for comparisons to KOs.

EPC and tail clip samples were lysed in 50 mM Tris pH8.5, 1 mM EDTA, 0.5% Tween-20, and 0.2 µg/µl proteinase K at 55 °C for 3 h, followed by 95 °C 10 min incubation. MyTaq Red Mix (Meridian Bioscience) was used to amplify 1 µL DNA in a 20 µL reaction using a 60 °C annealing temperature and 35 amplification cycles. The following primers were used for each locus: Dnmt3a-F CTGTGGCATCTCAGGGTGATGAGCA, R1 GCAAACAGACCCAACATGGAACCCT, R2 TGAGTGGTGAGGCCC-AGCTTATCGA (Approximate sizes: WT-400bp, KO-250bp); Dnmt3b-F1 GAACTTGGTCTGCAGGACGAT-CGCT, F2 AGAGCACTGCACCACTACTGCTGGA, R CAGGTCAGACCTCTCTGGTGACAAG (Approximate sizes: WT-270bp, floxed-400bp, KO-700bp); Dnmt3l-F1 GGTCCTTAGGGGTTCTGGAC, F2 GTTGGAGGA-TTGGGAAGACA, R1 TAGCTACCCGTGGCCAATAC, R2 CCATGGCATTGATCCTCTCT (Approximate sizes: WT-250bp, KO-170bp); Sox2-Cre-F GCAGAACCTGAAGATGTTCGCGAT, R AGGTATCTCTGACCAGAGTC-ATCC (Approximate size: 700 bp). PCR products were run on a 2% agarose gel and imaged using a Gel Doc XR + System (Bio-Rad).

Immunohistochemistry and immunofluorescence

For histological analysis, at least 4 mutant and 4 wild-type or heterozygous placentas were collected from at least two independent litters (with mutant and wild-type placentas recovered from each litter) for Dnmt3a KO, Dnmt3b KO, Dnmt3l KO and Dnmt3b cKO. Placentas from each litter were processed for routine paraffin histology and embedded side-by-side within a paraffin block. No other randomization is applicable for this study. In all cases, tissue appearance and cellular architecture of the placentas was analysed to confirm they were in viable condition. Consecutive 10μm sections were produced. A series of sections per block was processed for haematoxylin and eosin (H&E) staining, using a standard protocol⁴². Sections through the sagittal midline were chosen for imaging, indicated at E12.5 and E18.5 by the site of insertion of the umbilical cord. Slides were scanned on a Nanozoomer (Hamamatsu). Phenotypes of placentas were assessed for each strain, blinded and recorded by at least two independent investigators.

For immunostaining, sections were deparaffinised in xylene and processed through an ethanol series to PBS. Antigen retrieval was performed by boiling in 1 mM EDTA pH 7.2, 0.05% Tween-20 or in 10 mM Na-citrate pH 6.0 buffer followed by blocking in PBS, 0.5% BSA, 0.1% Tween-20. Antibodies used were anti-Cdh1 (1:100; BD Biosciences 610181), anti-MCT1 (1:100; Merck Millipore AB1286I), anti-MCT4 (1:100; Merck Millipore AB3314P) and biotin-conjugated isolectin from Bandeiraea simplicifolia BSI-B4 (1:100; Sigma L2140). Primary antibodies were detected with appropriate fluorescence or horseradish peroxidase-conjugated secondary antibodies; BSI-B4 was detected with horseradish peroxidase-conjugated streptavidin (1:400 Alexa594 (Goat anti-chicken) Invitrogen Cat. A110442, 1:400 Alexa488 (Donkey anti-rabbit) Invitrogen Cat. A21206, 1:200 Anti-Mouse IgG (H + L)-HRP (goat polyclonal) Cat. 170–6516 Bio-Rad, 1:100 Horseradish peroxidase-conjugated Streptavidin, Vector Laboratories Cat. SA5004). Nuclei were counterstained with haematoxylin or DAPI. For all immunostainings, samples were repeated at least twice in independent experiments. Images were taken with an Olympus BX61 epifluorescence microscope and/or Arperio scanner (Leica). The signal for MCT4 and MCT1 staining was quantified using ImageJ and Student’s t-test analysis was performed to calculate statistical significance of the staining signal differences (p < 0.05) using GraphPad Prism 9.

RNA-seq library preparation

Ultra-low input RNA-sequencing libraries were generated as previously described¹⁸, with the adaptation of using only a single round of oligo(dT)25 capture of mRNA to improve yields. In brief, samples were homogenized in 400 µL of TRIzol (Invitrogen) and phase separated with 80 µL of chloroform:isoamyl alcohol (Sigma), centrifuging at 4 °C for 15 min. The aqueous phase was combined with 1 µL GlycoBlue and 300 µL of cold isopropanol. The RNA was pelleted by centrifugation for 10 min at 4 °C, washed once with 75% ethanol, dried and then resuspended in 5 µL of ultra-pure water. To each RNA sample, 20 µL of lysis/binding buffer was immediately added, proceeding directly to oligo(dT)25 capture of mRNA, using Dynabeads mRNA DIRECT kit (Life Technologies), using previously described adaptations for low starting material¹⁸. Using the SMARTer Stranded RNA-seq kit (Clontech), mRNA libraries were prepared as per the manufacturers’ instructions with 15 amplification cycles. Libraries were quantified using the High DNA Sensitivity Bioanalyzer 2500 (Agilent) and Illumina library quantification (KAPA). Libraries were multiplexed and sequenced using 50 bp single-end on the Illumina HiSeq 2500 RapidRun. Four replicates were excluded due to poor amplification and consequently high duplication, including 2 KO and 2 WT replicates. In total, the number of replicates analysed per group were Dnmt3l KO (N = 4), Dnmt3l WT (N = 3), Dnmt3a/b DKO (N = 3), Dnmt3b KO (N = 4), Dnmt3a KO (N = 3), Dnmt3a/b WT (N = 5).

PBAT library preparation

PBAT libraries were generated as previously described⁴⁵. In brief, cells were lysed with 0.5% SDS at 37 °C for 1 h and bisulphite converted using the Imprint DNA Modification kit (Sigma). Bisulphite-converted DNA was purified using the EZ DNA Methylation Direct kit, as directed (Zymo Research). Using a biotin-conjugated adaptor containing standard Illumina adaptor sequences and 9 bp of random sequences (9 N), first-strand synthesis was performed using Klenow Exo- enzyme (New England Biolabs). Following exonuclease (New England Biolabs) treatment and binding to Dynabeads M-280 Streptavidin beads (Thermo Fisher Scientific), second-strand synthesis was performed. Libraries were amplified and indexed using 10 PCR cycles with Phusion High-Fidelity DNA polymerase (New England Biolabs). Libraries were multiplexed for 75-bp paired-end sequencing on an Illumina NextSeq500. Three replicates of Epi and ExE were sequenced and analysed per group, as detailed in Supplementary Data 1.

ChIP-seq library preparation

Ultra-low input ChIP-seq was performed as previously described⁴⁵, incorporating several adaptions from the original protocol⁴⁶. In brief, replicates of ExE samples were permeabilized in nuclei EZ lysis buffer (Sigma) with 0.1% Triton-X-100/0.1% deoxycholate. Chromatin digestion was completed with 200 U of micrococcal nuclease (New England Biolabs) at 21 °C for 7.5 min. Chromatin was then precleared in complete immunoprecipitation buffer with Protein A/G beads, for 2 h rotating at 4 °C. Chromatin was then divided into three aliquots: one for each immunoprecipitation and one 10% input control. Antibodies for H3K4me1 (250 ng, Active Motif, 39298) or H3K27ac (125 ng, Abcam, ab4729) were bound to Protein A/G beads in complete immunoprecipitation buffer for 3 h rotating at 4 °C. Chromatin was added to antibody-bound beads and rotated overnight at 4 °C. Chromatin-bound beads were washed with two low-salt washes and one high-salt wash, followed by DNA elution at 65 °C for 90 mins. Eluted DNA was purified with Sera-Mag carboxylate-modified Magnetic SpeedBeads (Fisher Scientific) at a 1.8:1 ratio. Library preparation and indexing was performed using the MicroPlex Library Preparation kit v2 (Diagenode), as per the manufacturer’s instructions. Libraries were multiplexed for 75-bp paired-end sequencing on an Illumina NextSeq500.

Single-nuclei RNA-seq library preparation

Single nuclei were isolated as described in the ‘Frankenstein’ protocol, available through 10X Genomics (https://www.10xgenomics.com/resources/customer-developed-protocols) and published adaptations for mouse placenta²⁷. Following collection and genotyping, Dnmt3b KO and WT E12.5 placentas or Dnmt3b cKO and WT E12.5 placentas were processed in parallel immediately for single nuclei isolation and 10X Genomics RNA-seq preparation. Placentas were first chopped with a razor blade in 1 mL nuclei EZ lysis buffer (Sigma) for 2 mins, followed by homogenisation in a Dounce homogenizer. Nuclei were filtered through a 30 µm filter, and then pelleted by centrifugation at 500 g for 5 mins at 4 °C. Nuclei were resuspended in fresh nuclei EZ lysis buffer, pelleted and then resuspended in nuclei wash and resuspension buffer (1% BSA, 0.2U/µL RNase Inhibitor, PBS). Nuclei were gently resuspended, pelleted and resuspended in fresh nuclei wash and resuspension buffer. Nuclei were passed again through a 30 µm filter, before visual inspection of nuclei with a haemocytometer. Nuclei were then stained with DAPI (adding 1 µL of 1 mg/mL DAPI) before a total of 10,000 nuclei per sample were flow sorted using the BD FACSAria Fusion sorter with 70 µm nozzle into RT reaction mix (10X Genomics). Samples were immediately processed to cDNA using the Chromium Controller (10X Genomics), followed by 3’ single-nuclei RNA-seq library preparation (10X Genomics). Libraries were multiplexed for sequencing on an Illumina NextSeq500.

Public datasets

Publicly available datasets, including RNA-seq from E7.5 reciprocal hybrid C57BL6/CAST Epi and ExE, and H3K4me3, H3K36me3, and H3K27me3 ChIP-seq in E6.5 C57BL6/CAST ExE¹⁸, were used in this study (available in Gene Expression Omnibus series GSE124216). Bisulphite-sequencing datasets for E3.5 inner cell mass and trophectoderm were also used in this study¹⁶ (available in series GSE84236). All raw data files were mapped and trimmed using the pipelines described below. Gene expression patterns of Sox2 in pre-implantation embryos were evaluated using publicly available database https://endoderm-explorer.com/⁴⁷.

Data mapping and QC

Fastq sequence files were quality and adaptor trimmed with trim galore v0.4.2 using default parameters. For PBAT libraries, the -clip option was used to remove the bases covered by the PBAT primer. Mapping of ChIP-seq data was performed with Bowtie v2.2.9 against the Mus musculus (GRCm38) derived genome, where SNPs between hybrid strains (C57BL6 and CAST/Ei) had been masked by the ambiguity nucleobase N (N-masked genome). The resulting hits were filtered to remove mappings with a MAPQ scores < 20. RNAseq data was subjected to trimming with Trim Galore (v0.6.6, options ‘--clip_r1 8’ to remove a nucleotide bias at the 5’ end), and aligned to the GRCm38 mouse genome build, to which a ribosomal DNA sequence had been added (BK000964.1), using HISAT2 (v2.1.0, guided by gene models from Ensembl annotation release 90, options: --dta –no-softclip). Hits were again filtered to remove mappings with MAPQ scores <20. Sequencing depths for all libraries are provided in Supplementary Data 1.

Mapping and methylation calling of bisulphite-seq data was performed using Bismark v0.16.3 in PBAT mode against the mouse GRCm38 genome assembly. Trimmed reads were first aligned to the genome in paired-end mode to be able to detect and discard overlapping parts of the reads while writing out unmapped singleton reads; in a second step remaining singleton reads were aligned in single-end mode. Alignments were carried out with Bismark⁴⁸ with the following set of parameters: a) paired-end mode: --pbat; b) single-end mode for Read 1: --pbat; c) single-end mode for Read 2: defaults. Reads were then deduplicated with deduplicate_bismark selecting a random alignment for positions that were covered more than once. Following methylation extraction, CpG context files from PE and SE runs were used to generate a single coverage file (the “Dirty Harry” procedure). Based on DNA methylation levels at CHH sequences, bisulphite conversion efficiencies were estimated to be average of 96.5% (+/−1.9%).

Mapping and quantitation of 10X single nuclei RNA-Seq data was performed with CellRanger v6.0.0 against the gex-mm10-2020-A transcriptome using the --include-introns option.

Bisulphite-seq analysis

DNA methylation data was evaluated using 100-CpG windows, excluding the mitochondria (MT), X and Y chromosomes. Percentage DNA methylation was calculated per base and then averaged per window in each replicate with a minimum of 10 informative CpGs required to quantitate each window. Methylation values were then averaged between replicates for each group. Differentially methylated regions (DMRs) were identified using two methods in SeqMonk v1.48.0, logistic regression with a minimum 20% difference in DNA methylation between KO and WT was used to identify DMRs in each KO. To identify genomic regions most affected in each KO model, a binomial test was used to identify the subset of domains that showed the largest differences in DNA methylation in each of the Dnmt3a, Dnmt3b and Dnmt3l KOs; these DMRs were used for the UMAP analysis, described below.

UMAP dimensionality reduction and clustering

Epigenetic data, including DNA methylation, H3K4me3, H3K27ac, H3K4me1, H3K36me3, and H3K27me3, for autosomal 100-CpG windows were used in UMAP dimensionality reduction and clustering. Regions were quantitated as RPKM for WT samples with coverage outliers excluded. To reduce the impact of outliers and correct for variable ChIP enrichment, capped, scaled and centred values were used for randomly initialised UMAP (n_neighbors=10, min_dist=0.05)⁴⁹. The same input was used for Expectation Maximisation Clustering in Weka⁵⁰, setting the number of clusters to 15. The resulting clusters were summarised into the functional annotations shown in Fig. 2 based on similarity of histone modifications. Enrichment or depletion of the most hypomethylated DMRs at chromatin features in Dnmt3a, Dnmt3b and Dnmt3l KOs were statistically compared to a random set of 100-CpG windows using one-proportion Z test with Yates’ continuity correction (results presented in Supplementary Data 2).

RNA-seq analysis

Gene expression was quantitated using the RNA-seq quantitation pipeline in SeqMonk. Differentially expressed genes were identified using DESeq2 in SeqMonk, with a minimum 1.5-fold-change between KO and matched WT samples. Gene ontology analyses were done using DAVID Bioinformatics Resource (https://david.ncifcrf.gov/), using the default settings and the addition of the UP-TISSUE category. The GRCm38 repeatmasker annotations were used to evaluate expression of all ERV repeat classes. Informative ERV LTRs were defined as those with one same-stranded read in at least one RNA-seq replicate (N = 8142). ERV expression was quantitated as RPKM using only reads that mapped to the same strand as the ERV. LIMMA statistic (p < 0.05 correcting for multiple comparisons) in SeqMonk v1.48.0 was used to identify ERVs that showed significantly different expression between Dnmt3a/b DKO and WT controls.

Single-nuclei RNA-seq analysis

Filtered count matrices from CellRanger were imported and analysed using Seurat v3. Cells containing more than 0.5% mitochrondrial reads were removed, then standard CLR normalisation, VST variable gene selection (1000 genes), scaling, PCA and tSNE (15 PCs, perplexity = 100). Cell clusters were defined using Louvain clustering (resolution 0.3), and clusters deriving from maternal cells were identified based on their Xist expression, and discarded.

Matches to known placental cell types were predicted using SCINA (sensitivity_cutoff=0.9) with the 50 most significant markers per cell type from a previous study²⁵.

Statistical analysis

Statistical tests were performed in Excel, R v4.2.0, http://vassarstats.net/, and GraphPad Prism 9. For linear correlations, Pearson Correlation Coefficients are provided as R-values and linear correlation statistical test was used to calculate p-values. Pairwise comparisons of Dnmt gene expression between E7.5 Epi and ExE, MCT4 and MCT1 fluorescence intensity, foetal weights, placental and labyrinth size and proportions between KOs and WT controls were done using a standard two-tailed t-test statistic, correcting the significance threshold using Bonferroni correction for multiple comparisons within each analysis. Comparisons of frequency data were done using Chi-square statistic (Supplementary Fig. 3g) and Fisher’s Exact test (Fig. 4f). ERV lengths across subclasses were compared using a non-parametric Mann Whitney test (Supplementary Fig. 3h). Comparison of fold change expression of Dnmt3a/b up-regulated ERVs (N = 23) between Dnmt KOs was done using an ANOVA statistic, with post hoc Tukey Test to determine significant pairwise comparisons. Figures were generated using Excel, SeqMonk v1.48.0, R Studio v1.1.456, R v4.2.0 (R package ggplot2), GraphPad Prism 9, CellRanger v6.0.0, and https://endoderm-explorer.com/.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The data that support this study are available from the corresponding author upon reasonable request. The sequencing data generated in this study has been deposited in Gene Expression Omnibus (GEO) database under accession number GSE203462, including raw and processed data files. Publicly available RNA-seq and ChIP-seq datasets (GSE124216) and bisulphite-seq data (GSE84236) were used in this study. Publicly available single-cell RNA-seq data was accessed through https://endoderm-explorer.com/. Source data are provided with this paper.

References

Yin, Y. et al. Impact of cytosine methylation on DNA binding specificities of human transcription factors. Science 356, eaaj2239 (2017).
Article Google Scholar
Bird, A. The dinucleotide CG as a genomic signalling module. J. Mol. Biol. 409, 47–53 (2011).
Article CAS Google Scholar
Smith, Z. D. & Meissner, A. DNA methylation: roles in mammalian development. Nat. Rev. Genet. 14, 204–220 (2013).
Article CAS Google Scholar
Auclair, G., Guibert, S., Bender, A. & Weber, M. Ontogeny of CpG island methylation and specificity of DNMT3 methyltransferases during embryonic development in the mouse. Genome Biol. 15, 545-014–0545-5 (2014).
Article Google Scholar
Bourc’his, D., Xu, G. L., Lin, C. S., Bollman, B. & Bestor, T. H. Dnmt3L and the establishment of maternal genomic imprints. Science 294, 2536–2539 (2001).
Article ADS Google Scholar
Barau, J. et al. The DNA methyltransferase DNMT3C protects male germ cells from transposon activity. Science 354, 909–912 (2016).
Article ADS CAS Google Scholar
Okano, M., Bell, D. W., Haber, D. A. & Li, E. DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99, 247–257 (1999).
Article CAS Google Scholar
Grosswendt, S. et al. Epigenetic regulator function through mouse gastrulation. Nature 584, 102–108 (2020).
Article ADS CAS Google Scholar
Dahlet, T. et al. Genome-wide analysis in the mouse embryo reveals the importance of DNA methylation for transcription integrity. Nat. Commun. 11, 3153 (2020).
Article ADS CAS Google Scholar
Schroeder, D. I. et al. The human placenta methylome. Proc. Natl Acad. Sci. U. S. A. 110, 6037–6042 (2013).
Article ADS CAS Google Scholar
Decato, B. E., Lopez-Tello, J., Sferruzzi-Perri, A. N., Smith, A. D. & Dean, M. D. DNA Methylation Divergence and Tissue Specialization in the Developing Mouse Placenta. Mol. Biol. Evol. 34, 1702–1712 (2017).
Article CAS Google Scholar
Hamada, H. et al. Allele-Specific Methylome and Transcriptome Analysis Reveals Widespread Imprinting in the Human Placenta. Am. J. Hum. Genet. 99, 1045–1058 (2016).
Article CAS Google Scholar
Novakovic, B. & Saffery, R. Placental pseudo-malignancy from a DNA methylation perspective: unanswered questions and future directions. Front. Genet. 4, 285 (2013).
Article Google Scholar
Branco, M. R. et al. Maternal DNA Methylation Regulates Early Trophoblast Development. Dev. Cell. 36, 152–163 (2016).
Article CAS Google Scholar
Surani, M. A., Barton, S. C. & Norris, M. L. Development of reconstituted mouse eggs suggests imprinting of the genome during gametogenesis. Nature 308, 548–550 (1984).
Article ADS CAS Google Scholar
Smith, Z. D. et al. Epigenetic restriction of extraembryonic lineages mirrors the somatic transition to cancer. Nature 549, 543–547 (2017).
Article ADS Google Scholar
Wang, L. et al. Programming and inheritance of parental DNA methylomes in mammals. Cell 157, 979–991 (2014).
Article CAS Google Scholar
Hanna, C. W. et al. Endogenous retroviral insertions drive non-canonical imprinting in extra-embryonic tissues. Genome Biol. 20, 225 (2019).
Article Google Scholar
Ooi, S. K. et al. DNMT3L connects unmethylated lysine 4 of histone H3 to de novo methylation of DNA. Nature 448, 714–717 (2007).
Article ADS CAS Google Scholar
Baubec, T. et al. Genomic profiling of DNA methyltransferases reveals a role for DNMT3B in genic methylation. Nature 520, 243–247 (2015).
Article ADS CAS Google Scholar
Jeltsch, A. & Jurkowska, R. Z. Allosteric control of mammalian DNA methyltransferases - a new regulatory paradigm. Nucleic Acids Res 44, 8556–8575 (2016).
Article CAS Google Scholar
Ernst, J. & Kellis, M. Discovery and characterization of chromatin states for systematic annotation of the human genome. Nat. Biotechnol. 28, 817–825 (2010).
Article CAS Google Scholar
Manzo, M. et al. Isoform-specific localization of DNMT3A regulates DNA methylation fidelity at bivalent CpG islands. EMBO J. 36, 3421–3434 (2017).
Article CAS Google Scholar
Chuong, E. B., Rumi, M. A., Soares, M. J. & Baker, J. C. Endogenous retroviruses function as species-specific enhancer elements in the placenta. Nat. Genet. 45, 325–329 (2013).
Article CAS Google Scholar
Todd, C. D., Deniz, O., Taylor, D. & Branco, M. R. Functional evaluation of transposable elements as enhancers in mouse embryonic and trophoblast stem cells. Elife 8, https://doi.org/10.7554/eLife.44344 (2019).
Rossant, J. & Cross, J. C. Placental development: lessons from mouse mutants. Nat. Rev. Genet. 2, 538–548 (2001).
Article CAS Google Scholar
Marsh, B. & Blelloch, R. Single nuclei RNA-seq of mouse placental labyrinth development. Elife 9, https://doi.org/10.7554/eLife.60266 (2020).
Watson, E. D. & Cross, J. C. Development of structures and transport functions in the mouse placenta. Physiol. (Bethesda) 20, 180–193 (2005).
CAS Google Scholar
Hayashi, S., Lewis, P., Pevny, L. & McMahon, A. P. Efficient gene modulation in mouse epiblast using a Sox2Cre transgenic mouse strain. Mech. Dev. 119, S97–S101 (2002).
Article Google Scholar
Hill, P. W. S. et al. Epigenetic reprogramming enables the transition from primordial germ cell to gonocyte. Nature 555, 392–396 (2018).
Article ADS CAS Google Scholar
Chédin, F. The DNMT3 family of mammalian de novo DNA methyltransferases. Prog. Mol. Biol. Transl. Sci. 101, 255–285 (2011).
Article Google Scholar
Zhang, Y. et al. Dynamic epigenomic landscapes during early lineage specification in mouse embryos. Nat. Genet. 50, 96–105 (2018).
Article CAS Google Scholar
Chedin, F., Lieber, M. R. & Hsieh, C. L. The DNA methyltransferase-like protein DNMT3L stimulates de novo methylation by Dnmt3a. Proc. Natl Acad. Sci. U. S. A. 99, 16916–16921 (2002).
Article ADS CAS Google Scholar
Gowher, H., Liebert, K., Hermann, A., Xu, G. & Jeltsch, A. Mechanism of stimulation of catalytic activity of Dnmt3A and Dnmt3B DNA-(cytosine-C5)-methyltransferases by Dnmt3L. J. Biol. Chem. 280, 13341–13348 (2005).
Article CAS Google Scholar
Weinberg, D. N. et al. Two competing mechanisms of DNMT3A recruitment regulate the dynamics of de novo DNA methylation at PRC1-targeted CpG islands. Nat. Genet. 53, 794–800 (2021).
Article CAS Google Scholar
Nowialis, P. et al. Catalytically inactive Dnmt3b rescues mouse embryonic development by accessory and repressive functions. Nat. Commun. 10, 4374 (2019). 019-12355-7.
Article ADS Google Scholar
Shirane, K., Miura, F., Ito, T. & Lorincz, M. C. NSD1-deposited H3K36me2 directs de novo methylation in the mouse male germline and counteracts Polycomb-associated silencing. Nat. Genet. 52, 1088–1098 (2020).
Article CAS Google Scholar
Stewart, K. R. et al. Dynamic changes in histone modifications precede de novo DNA methylation in oocytes. Genes Dev. 29, 2449–2462 (2015).
Article CAS Google Scholar
Veselovska, L. et al. Deep sequencing and de novo assembly of the mouse oocyte transcriptome define the contribution of transcription to the DNA methylation landscape. Genome Biol. 16, 209 (2015).
Article Google Scholar
Ng, R. K. et al. Epigenetic restriction of embryonic cell lineage fate by methylation of Elf5. Nat. Cell Biol. 10, 1280–1290 (2008).
Article CAS Google Scholar
Nagai, A., Takebe, K., Nio-Kobayashi, J., Takahashi-Iwanaga, H. & Iwanaga, T. Cellular expression of the monocarboxylate transporter (MCT) family in the placenta of mice. Placenta 31, 126–133 (2010).
Article CAS Google Scholar
Perez-Garcia, V. et al. Placentation defects are highly prevalent in embryonic lethal mouse mutants. Nature 555, 463–468 (2018).
Article ADS CAS Google Scholar
Liberti, D. C. et al. Dnmt1 is required for proximal-distal patterning of the lung endoderm and for restraining alveolar type 2 cell fate. Dev. Biol. 454, 108–117 (2019).
Article CAS Google Scholar
Januar, V., Desoye, G., Novakovic, B., Cvitic, S. & Saffery, R. Epigenetic regulation of human placental function and pregnancy outcome: considerations for causal inference. Am. J. Obstet. Gynecol. 213, S182–S196 (2015).
Article CAS Google Scholar
Hanna, C. W. et al. MLL2 conveys transcription-independent H3K4 trimethylation in oocytes. Nat. Struct. Mol. Biol. 25, 73–82 (2018).
Article CAS Google Scholar
Brind’Amour, J. et al. An ultra-low-input native ChIP-seq protocol for genome-wide profiling of rare cell populations. Nat. Commun. 6, 6033 (2015).
Article ADS Google Scholar
Nowotschin, S. et al. The emergent landscape of the mouse gut endoderm at single-cell resolution. Nature 569, 361–367 (2019).
Article ADS CAS Google Scholar
Krueger, F. & Andrews, S. R. Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications. Bioinformatics 27, 1571–1572 (2011).
Article CAS Google Scholar
McInnes, L., Healy, J. & Melville, J. UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction. arXiv https://doi.org/10.48550/arXiv.1802.03426 (2018).
Witten, I. H., Frank, E., Hall, M. A. & Pal, C. J. in Data Mining, Fourth Edition: Practical Machine Learning Tools and Techniques 654 (Morgan Kaufmann Publishers Inc., Sand Francisco, CA, United States, 2016).

Download references

Acknowledgements

Thank you to the Biological Support Unit, Flow Cytometry Facility, Dr. Paula Kokko-Gonzales and Dr. Amelia Edwards at the Sequencing Facility, Dr. Simon Walker at the Imaging Facility, and Dr. Felix Krueger and Dr. Laura Biggins in the Bioinformatics Facility at the Babraham Institute. Thank you to Dr. Tim Lohoff and Dr. Stephen Clark for advice in establishing the single-nuclei isolation protocol. Thank you to Prof. Kathy Niakan, Director of the Centre for Trophoblast Research, and Dr. Georgia Lea, postdoctoral researcher in the Department of Physiology, Development and Neuroscience, at the University of Cambridge, for their valuable comments on the manuscript. This work was funded by a Next Generation Fellowship from the Centre for Trophoblast Research, Career Support Fund from the University of Cambridge, and a Sir Henry Dale Fellowship from the Royal Society and Wellcome Trust (222582/Z/21/Z), awarded to CH. VPG is supported by the Ministerio de Ciencia e Innovación (RYC-2019-026956 and PID2020-114459RA-I00). Part of the equipment employed in this work was funded by Generalitat Valenciana and co-financed with ERDF funds (OP ERDF of Comunitat Valenciana 2014–2020). MH is supported by a Tier I Canada Research Chair in Developmental Genetics and Epigenetics.

Author information

Christel Krueger
Present address: Bioinformatics Innovation Hub, Altos Labs Cambridge Institute, Cambridge, UK

Authors and Affiliations

Bioinformatics Programme, Babraham Institute, Cambridge, UK
Simon Andrews & Christel Krueger
Epigenetics Programme, Babraham Institute, Cambridge, UK
Christel Krueger & Courtney W. Hanna
Centro de Investigación Príncipe Felipe, Valencia, Spain
Maravillas Mellado-Lopez & Vicente Perez-Garcia
Department of Biochemistry and Molecular Biology, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada
Myriam Hemberger
Department of Medical Genetics, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada
Myriam Hemberger
Alberta Children’s Hospital Research Institute, University of Calgary, Calgary, AB, Canada
Myriam Hemberger & Wendy Dean
Department of Cell Biology and Anatomy, University of Calgary, Calgary, AB, Canada
Wendy Dean
Department of Physiology, Development and Neuroscience, University of Cambridge, Cambridge, UK
Courtney W. Hanna
Centre for Trophoblast Research, University of Cambridge, Cambridge, UK
Courtney W. Hanna

Authors

Simon Andrews
View author publications
You can also search for this author in PubMed Google Scholar
Christel Krueger
View author publications
You can also search for this author in PubMed Google Scholar
Maravillas Mellado-Lopez
View author publications
You can also search for this author in PubMed Google Scholar
Myriam Hemberger
View author publications
You can also search for this author in PubMed Google Scholar
Wendy Dean
View author publications
You can also search for this author in PubMed Google Scholar
Vicente Perez-Garcia
View author publications
You can also search for this author in PubMed Google Scholar
Courtney W. Hanna
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

C.H. conceptualised the project with valuable input from M.H. and W.D. C.H., M.M.L. and V.P.G. collected data and performed experiments. S.A., C.K., V.P.G., and C.H. performed data analysis and generated figures. S.A., C.K., W.D., M.H., V.P.G., and C.H. contributed to data interpretation. C.H. wrote the manuscript with contributions from S.A., C.K., W.D., M.H. and V.P.G.

Corresponding author

Correspondence to Courtney W. Hanna.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Maxim Greenberg and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Description of Additional Supplementary Files

Supplementary Data 1-4

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Andrews, S., Krueger, C., Mellado-Lopez, M. et al. Mechanisms and function of de novo DNA methylation in placental development reveals an essential role for DNMT3B. Nat Commun 14, 371 (2023). https://doi.org/10.1038/s41467-023-36019-9

Download citation

Received: 07 June 2022
Accepted: 11 January 2023
Published: 23 January 2023
DOI: https://doi.org/10.1038/s41467-023-36019-9

This article is cited by

Epigenetic role of LINE-1 methylation and key genes in pregnancy maintenance
- Veronica Tisato
- Juliana A. Silva
- Donato Gemmati
Scientific Reports (2024)
Important role of DNA methylation hints at significant potential in tuberculosis
- Yuexuan Qin
- Tianyue Li
- Bikui Tang
Archives of Microbiology (2024)
Expression analysis suggests that DNMT3L is required for oocyte de novo DNA methylation only in Muridae and Cricetidae rodents
- Lirik Behluli
- Alyssa M. Fontanilla
- Lenka Gahurova
Epigenetics & Chromatin (2023)
DNA methylation in poultry: a review
- Xing Ju
- Zhijun Wang
- Qinghua Nie
Journal of Animal Science and Biotechnology (2023)
Base editing-mediated one-step inactivation of the Dnmt gene family reveals critical roles of DNA methylation during mouse gastrulation
- Qing Li
- Jiansen Lu
- Jinsong Li
Nature Communications (2023)

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.