The 3D genome during germline development and meiosis

Abstract

Germ cell development involves extensive remodeling of the 3D genome architecture, which is tightly coupled to transcriptional programs, meiotic chromosome dynamics, and re-establishment of totipotency in the next generation. Recent advances in chromosome conformation capture methods have uncovered stage-specific alterations in chromosome organization during spermatogenesis and oogenesis, including germline-specific 3D genome features. These distinctive nuclear configurations orchestrate gene expression programs essential for each developmental stage and meiosis, contribute to epigenetic inheritance, and shape genome evolution. In this review, we synthesize recent progress in understanding 3D genome organization in male and female germlines, and highlight emerging principles, unresolved questions, and innovative approaches that will advance our understanding of germline biology and the principles of genome architecture.

Overview of germline development

The germ cell lineage is responsible for transmitting both genetic and epigenetic information to the next generation. It also undergoes extensive epigenetic reprogramming to restore totipotency after fertilization. In embryos, primordial germ cells (PGCs, see Glossary) – the common precursors of male and female germ cells – are specified early and later receive sex-specific signals from the surrounding somatic niche which lead to the establishment of male or female cell fates by embryonic day 13.5 (E13.5) in mice [1] (Figure 1A).

Figure 1. Developmental timeline and sex-specific differentiation of the mouse germline.

(A) Early stages of germ cell development from primordial germ cell (PGC) specification to sex determination at embryonic day 13.5 (E13.5). PGCs are specified in the proximal epiblast in response to BMP and WNT signaling, migrate through the hindgut, and colonize the bipotential gonad. Upon arrival, germ cells receive sex-specific cues from the somatic environment. In XY gonads, expression of the Y-linked gene Sry in supporting cell precursors induces Sertoli cell differentiation which initiates the male pathway and triggers mitotic arrest of prospermatogonia. In XX gonads, the absence of Sry leads to the differentiation of ovarian supporting cell precursors, resulting in the meiotic entry of oogonia during fetal life. (B) Male germline development and the 3D genome. A few days after birth (around postnatal day 0, P0), prospermatogonia exit quiescence and re-enter the cell cycle to produce undifferentiated spermatogonia (A_single, A_paired, A_aligned), including spermatogonial stem cells (SSCs). These give rise to differentiating spermatogonia (A₁–A₄, intermediate, and type B) which commit to meiosis. Meiotic prophase I progresses through the leptotene, zygotene, and pachytene stages – marked by XY body formation during meiotic sex chromosome inactivation (MSCI) – followed by the diplotene stage and two meiotic divisions, and produces haploid round spermatids. Spermiogenesis then remodels nuclear architecture and replaces most histones with protamines, resulting in highly compacted chromatin in mature sperm. The panel presents a schematic summary of 3D genome organization during male germline development, showing A/B compartment remodeling, including refined A/B compartment emergence, and topologically associating domain (TAD) attenuation. (C) Female germline development and the 3D genome. Female germ cell development progresses from oogonia through meiotic entry, germinal cyst breakdown, and primordial follicle formation, followed by folliculogenesis. Meiotic prophase I substages (leptotene, zygotene, pachytene, diplotene) occur in the fetal ovary, where non-growing oocytes (NGOs) arrest at the dictyate stage. Upon follicular activation, cohorts of NGOs are activated to initiate oocyte growth (growing oocytes, GOs). GOs give rise to fully grown oocytes (FGOs) that resume meiosis I in response to a surge in luteinizing hormone. This leads to an arrest at metaphase II (MII) and subsequent ovulation, releasing the MII oocytes. Fertilization then triggers the completion of MII. The panel presents a schematic summary of 3D genome organization during female germline development, including Polycomb-associating domains (PADs). Abbreviations: Dip, diplotene; Pac, pachytene; Zyg, zygotene.

In males, fetal germ cells differentiate into prospermatogonia (also known as gonocytes) and enter a mitotically quiescent state in the G1/G0 phase of the cell cycle [2]. After birth, these cells resume proliferation and give rise to undifferentiated spermatogonia, which include spermatogonial stem cells (SSCs) that are responsible for long-term maintenance of the male germline in adult testes. In response to retinoic acid signaling, undifferentiated spermatogonia commence differentiation, enter meiosis, and eventually mature into highly condensed spermatozoa (Figure 1B).

In females, fetal germ cells enter meiotic prophase and then transition into meiotically arrested non-growing oocytes (NGOs) which form the ovarian reserve that supports long-term maintenance of the female germline in the adult ovary. During each estrous cycle in mice (or menstrual cycle in humans), cohorts of NGOs are activated, giving rise to fully grown oocytes that resume meiosis I in response to a surge in luteinizing hormone. This leads to arrest at meiosis II, followed by ovulation. Fertilization subsequently triggers the completion of meiosis II (Figure 1C).

These stepwise differentiation processes are accompanied by tightly regulated transcriptional programs which are coupled to profound, stage-specific remodeling of the nuclear architecture [3–6] (Figure 1D). Recent advances in 3D genome-mapping technologies – particularly proximity ligation-based methods such as high-throughput chromatin conformation capture (Hi-C) – have enabled high-resolution profiling of chromatin interactions across the genome, both at the population level and at single-cell resolution. These technologies have revealed a hierarchical organization of the genome at multiple levels [7,8]: (i) chromosome territories (Figure 2A); (ii) A/B compartments, which distinguish between transcriptionally active and inactive regions (Figure 2B); (iii) topologically associating domains (TADs), which constrain regulatory interactions within defined chromatin neighborhoods (Figure 2C); and (iv) chromatin loops, which represent highly frequent interactions between two genomic loci that are typically stabilized by convergent CCCTC-binding factor (CTCF) sites and extruded by the cohesin complex (Figure 2D). 3D genome analysis provides crucial insights into how the nuclear architecture is dynamically reprogrammed during male and female germ cell development and meiosis, and how this reorganization contributes to gene regulation and cell fate decisions.

Figure 2. Hierarchical organization of the mammalian genome.

The mammalian genome is organized into multiple hierarchical layers that regulate gene expression and nuclear architecture. The upper portion of the figure presents conceptual schematics of these genome structures, while the lower portion illustrates how each is visualized using high-throughput chromatin conformation capture (Hi-C) data. This integration links physical genome architecture to its experimental detection and interpretation. (A) At the largest scale, chromosome territories define discrete nuclear domains that are occupied by individual chromosomes. (B) Within each chromosome, the genome is partitioned into A and B compartments that are revealed by Hi-C as large-scale interaction patterns. A compartments are transcriptionally active, gene-rich, and enriched for open chromatin, whereas B compartments are transcriptionally inactive, gene-poor, and enriched for heterochromatic marks. Many B compartment regions coincide with lamina-associated domains (LADs) – genomic regions that are anchored to the nuclear lamina and typically contribute to gene silencing and spatial genome compartmentalization. (C) At the sub-megabase scale, the genome is further partitioned into topologically associating domains (TADs) – self-interacting genomic regions that play key roles in genome architecture and gene regulation by facilitating interactions between regulatory elements and their target genes. (D) At the kilobase to megabase scale, chromatin loops bring distal regulatory elements, such as enhancers and promoters, into close proximity to modulate transcription. Loop anchor sites are often enriched for CTCF and the cohesin complex which extrudes chromatin loops until halted by convergent CTCF binding sites.

In this review, we focus on recent studies employing Hi-C-based approaches to investigate chromatin architecture in male and female germlines during sperm and egg formation. We discuss how 3D genome organization changes during spermatogenesis and oogenesis, how these structural dynamics contribute to germ cell differentiation and meiosis, and how they may influence the development of the next generation. Our primary focus is on gametogenesis, and we provide only a brief introduction to zygotic genome activation and early embryogenesis because these have been extensively reviewed elsewhere [9–13].

The 3D genome in the male germline: from PGCs to spermatogonia

Germ cells preserve their ability to recover totipotency and maintain a flexible organization at the chromatin level. In PGCs, epigenetic marks – including DNA methylation and Polycomb repressive complex (PRC)-mediated silencing – are erased to enable entry into the male or female gametogenesis program and subsequent meiosis [1].

A major challenge in studying the 3D genome in embryonic germ cells is the scarcity and heterogeneity of these cell populations. To address this, a recent study leveraged an in vitro culture system that faithfully recapitulates mouse germ cell development from pluripotent precursors to PGC-like cells (PGCLCs) to characterize features of 3D genome organization in the male germline [14]. The authors characterized key stages of PGCLC differentiation from embryonic stem cells (ESCs) and germline stem cells (GSCs) – in vitro-derived cells from perinatal spermatogenesis that serve as a proxy for undifferentiated spermatogonia [15]. They found that a specific stage of PGCLCs corresponding to E11.5 exhibited enhanced chromatin insulation at TAD boundaries and increased CTCF binding which demarcates TADs [14]. TADs are chromatin regions with high interaction frequencies that appear as discrete domains on population-averaged Hi-C maps, and typically span ~100 kb to ~2 Mb [16]. Because they arise from averaged conformations across many cells, TADs are probably statistical rather than physical structures present in individual cells. CTCF-marked boundaries restrict interactions with neighboring domains [10,17] while allowing frequent contacts among genes and regulatory elements within a TAD. The enhanced chromatin insulation observed in PGCLCs might safeguard transcriptional fidelity at a stage in which DNA methylation is extensively reprogrammed.

Compared to PGCLCs, chromatin insulation is substantially lower in GSCs, suggesting extensive euchromatinization during male germline development, particularly during the transition from PGCs to undifferentiated spermatogonia [18]. This feature may be attributed to a germline-specific mitotic cohesin complex in spermatogonia that contains the STAG3 subunit – that was previously considered to be a meiosis-specific cohesin subunit [19]. Because STAG3–cohesin has a shorter chromatin residence time, it attenuates TAD strength, rewires enhancer–promoter interactions, and strengthens compartmentalization, and thereby establishes unique chromatin features in spermatogonia. Consistent with this, spermatogonia exhibit weaker insulation strength than fibroblasts while maintaining comparable TAD boundary positions [20].

During the transition from ESCs to PGCLCs and subsequent differentiation into GSCs, there is also a major genome-wide shift from the B compartment to the A compartment [14]. Notably, more than one third of the genomic regions classified as B compartment in ESCs are reclassified as the A compartment in GSCs [14]. A/B compartments represent large-scale chromatin domains with distinct transcriptional and spatial properties [21]. A compartments correspond to gene-rich, transcriptionally active regions that are characterized by open chromatin and active histone modifications. These regions are typically localized towards the nuclear interior. By contrast, B compartments are gene-poor and transcriptionally inactive regions that are associated with repressive histone marks and are generally found near the nuclear periphery, and often overlap with lamina-associated domains (LADs) [22]. Consistent with these general features, the transition from PGCLCs to GSCs is associated with a reduction in LADs and reorganization of heterochromatin [14], which further supports the occurrence of euchromatinization during male germline development.

In line with these observations, 3D chromatin reorganization and less well defined A/B compartments have been detected by Hi-C analysis of in vivo-isolated prospermatogonia [23]. This appears to be a unique feature of E17.5 mitotically arrested prospermatogonia, which suggests that a major 3D genome reorganization occurs during this developmental time window. A recent study employing single-cell Hi-C analysis of embryonic germ cells provided a detailed view of 3D genome architecture in mitotically arrested prospermatogonia [24]. At this G1/G0-arrested stage, male germ cells exhibited a special chromatin structure similar to that of mitotic embryonic germ cells in G1 phase, and short-range chromatin interactions were increased [24]. This may be a preparatory stage for postnatal spermatogenesis [5,25].

Epigenetic priming through 3D chromatin

A key feature of spermatogenesis is unidirectional differentiation towards haploid sperm. Notably, many genes required for later stages of spermatogenic differentiation are epigenetically primed in undifferentiated spermatogonia, including the SSC population, before their activation [5] (Box 1). Thus, SSCs are preset for unidirectional differentiation while self-renewing. Recent Hi-C studies have revealed that this priming mechanism involves 3D chromatin contacts. Meiotic super-enhancers (SEs), which drive high gene expression in pachytene spermatocytes, are premarked by CTCF-mediated 3D chromatin contacts in undifferentiated spermatogonia [26]. Using an ESC model that recapitulates meiotic gene expression, deletion of a CTCF binding site adjacent to a meiotic SE results in ectopic expression of the target gene in undifferentiated conditions and failure of upregulation in differentiating conditions [26], which confirms the regulatory role of CTCF in meiotic gene expression. In addition to SEs, a large number of promoters (~12 000) utilized during spermatogenesis are also primed in undifferentiated spermatogonia through 3D chromatin contacts and the binding of transcription factors such as ZBTB16, SALL4, and SOX3 [27]. Thus, epigenetic priming in undifferentiated spermatogonia involves 3D chromatin organization at both meiotic SEs and a broad range of promoters.

Box 1. Epigenetic priming in spermatogenesis.

In undifferentiated spermatogonia, the gene expression program for spermatogenic differentiation is predetermined at the chromatin level before the initiation of differentiation. Within this population, SSCs self-renew but retain unidirectional differentiation potential. This state is established and maintained through the activity of PRC1 and PRC2, two major players in the maintenance and differentiation of SSCs and in epigenetic priming [83]. Upon differentiation, loss of PRC2-mediated histone H3 lysine 27 trimethylation (H3K27me3) activates key spermatogenic genes [83], but these genes are already primed with histone H3 lysine 4 dimethylation (H3K4me2), an active histone modification, and RNA polymerase II (Pol II) in undifferentiated spermatogonia [84,85].

Programmed gene repression during differentiation is also predetermined in undifferentiated spermatogonia by epigenetic repressive marks. SCML2, a spermatogenic-specific component of PRC1, binds to thousands of active genes in spermatogonia whose promoters have low DNA methylation and are enriched for H3K4me3 deposited by the histone methyltransferase KMT2B (also known as MLL2) [86,87]. SCML2 recruits PRC2 to these targets during meiosis, which induces H3K27me3 and establishes extensive bivalent domains that repress these genes later in meiotic prophase I [88].

Moreover, during meiotic prophase I, Pol II is already loaded onto the promoters of target genes of A-MYB, a key transcription factor that drives pachytene transcription, but it remains paused before the pachytene stage [89]. At the later pachytene stage, A-MYB and the testis-specific bromodomain protein BRDT cooperate to release paused Pol II, thereby activating these target genes [89].

The 3D genome during spermatogenic differentiation and meiosis

The transition from mitotic spermatogonia to meiotic spermatocytes is accompanied by major chromatin restructuring [26]. During meiotic prophase I, male germ cells undergo dynamic nuclear reorganization to support homologous chromosome pairing, recombination, and segregation [28] (Box 2). This shift reflects global 3D chromatin reprogramming from the typical interphase configuration of spermatogonia to the condensed chromosomes that are present during meiotic prophase which resemble mitotic chromosomes [29]. However, unlike transcriptionally inert mitotic chromosomes, meiotic prophase chromosomes remain transcriptionally active and establish a distinct 3D chromatin landscape. Hi-C studies have consistently found that TADs are attenuated or absent in pachytene spermatocytes when homologous synapsis is completed [20,26,30–36]. Stage-specific analyses revealed that TADs are already largely attenuated upon meiotic entry [34], and even before meiosis at the differentiating spermatogonia stage [36]. Given that this is a population analysis based on bulk Hi-C, this attenuation may reflect variable chromatin organization among individual spermatocytes.

Box 2. Substages and features of meiotic prophase I.

Meiotic prophase I is a prolonged and tightly regulated phase of meiosis that is characterized by the pairing, synapsis, and recombination of homologous chromosomes. It is traditionally divided into five sequential substages – leptotene, zygotene, pachytene, diplotene, and diakinesis – each of which is defined by distinct chromosomal events that are essential for successful gametogenesis (Figure 3 in the main text).

During the leptotene stage, chromosomes begin condensing and become visible as thin threads. DNA double-strand breaks (DSBs), which initiate homologous recombination, are formed by SPO11, a bacterial topoisomerase VI-like enzyme. Chromosomes also start attaching to the nuclear envelope via telomeres.

The zygotene stage marks the onset of homolog recognition and pairing. Homologous chromosomes initiate pairing (synapsis) through the assembly of the synaptonemal complex (SC), a proteinaceous structure that aligns homologs along their length.

By the pachytene stage, synapsis is complete along the entire length of homologous chromosomes, and crossover events occur. Crossovers are crucial for generating genetic diversity and ensuring proper chromosome segregation.

During the diplotene stage, the synaptonemal complex disassembles, and homologous chromosomes begin to separate, and remain connected only at crossover sites via chiasmata.

Finally, the diakinesis stage involves further chromatin condensation, preparation for metaphase I, and nuclear envelope dissolution. Homologs continue to separate and chiasmata migrate toward chromosome ends.

Together, these coordinated events ensure accurate recombination and segregation of homologous chromosomes, and set the foundation for the formation of haploid gametes.

Reflecting the dynamic transcriptional changes that occur during the transition from mitotic spermatogonia to meiotic spermatocytes, chromatin loops – often visualized as dots in Hi-C maps that are frequently anchored at active regulatory elements [16] – also undergo extensive reorganization [26]. Chromatin loops, which facilitate enhancer–promoter interactions within TADs, are presumably formed through cohesin-mediated loop extrusion and anchored by convergent CTCF binding sites [7,8]. In meiotic prophase I, many meiosis-specific chromatin loops are established through the action of A-MYB (MYBL1), a master transcription factor that drives pachytene transcription [37], while spermatogonia-specific loops are resolved in part by SCML2, a germline-specific Polycomb protein which suppresses the spermatogonial transcription program in meiosis [26]. In addition, the meiotic cohesin subunit RAD21L is required for proper establishment of chromatin compartments and for meiotic gene expression [38]. Chromatin loops are thus tightly associated with transcription.

A hallmark of meiotic prophase I is the formation of a specialized axis–loop chromatin architecture in which axis loops emanate from a proteinaceous axial core composed of meiosis-specific cohesins (e.g., REC8) and axial element proteins such as synaptonemal complex protein 3 (SYCP3) [39]. Upon meiotic entry, loop arrays assemble along chromosome axes (axis loops) to promote homolog pairing and recombination (Figure 3A). However, these axis loops observed by microscopy [40] or inferred from Hi-C contact probability analyses [31,34] appear to be distinct from the chromatin loops that are detected as ‘dots’ in typical Hi-C heatmaps [26]. Therefore, we use the term ‘axis loops’ here to distinguish them from conventional ‘chromatin loops’. Although the relationship between conventional ‘axis loops’ and ‘chromatin loops’ remains unknown, population-based analyses of bulk Hi-C suggest that frequent interactions that appear as ‘dots’ at sites of active transcription may underlie ‘chromatin loops’ [26], whereas the positions of ‘axis loops’ may vary among cells in the population and therefore be detectable only through Hi-C contact probability analyses [31,34].

Figure 3. Meiotic chromosome organization.

(A) Axis–loop organization across the stages of meiotic prophase I in mammals. In leptotene, chromosomal axial elements form along each sister chromatid, supported by meiosis-specific cohesin complexes. Programmed DNA double-strand breaks (DSBs) occur preferentially in loop regions and facilitate homologous recombination. During zygotene, partial synapsis between homologous chromosomes begins and the axis–loop framework supports alignment. The synaptonemal complex (SC) starts to assemble, and axial and lateral elements form the structural backbone that is joined by transverse filaments and central elements to align homologous chromosomes. By pachytene, full synapsis of autosomes and complete formation of SC are achieved, and recombination intermediates are resolved. In diplotene, SC disassembly begins, while loop–axis organization is preserved until homolog separation at the meiotic divisions. (B) At the onset of zygotene, telomeres attach to the inner nuclear envelope and cluster into a bouquet configuration, thereby bringing chromosome ends into proximity to promote efficient homolog pairing and synapsis.

The mechanism underlying the formation of ‘axis loops’ in mammals remains a mystery, although a non-stochastic nature has been suggested [41,42]. Insights from budding and fission yeast Hi-C studies highlight a crucial role for the meiosis-specific cohesin subunit Rec8 in this process. In yeast, Rec8 binding sites are enriched at the anchor points of axis loops, which manifest as dot-like interaction patterns in Hi-C contact maps, and these axis loops are dependent on Rec8 [43,44]. Detecting such features in mice is more challenging because their much larger genome size limits the Hi-C resolution compared to yeast. A recent study employed Micro-C, a Hi-C derivative that uses micrococcal nuclease instead of restriction enzymes to fragment the genome [45,46], to achieve a higher-resolution view of chromosomal interactions in mouse pachytene spermatocytes and found that CTCF is associated with axis–loop anchor sites [36]. Nevertheless, during mammalian meiosis, the positions of these axis–loop anchors may vary considerably between cells, further complicating their detection.

Beyond axis–loop organization, meiotic chromosomes also undergo reprogramming at the compartment level. While some earlier studies did not detect obvious switching of A/B compartments on autosomes during meiotic prophase I [30,31], attenuated compartment strength [30] and remodeling/refinement in the genomic location of A/B compartments were observed in pachytene spermatocytes [20,32–35]. These differences appear to reflect technical issues arising from the subjective nature of A/B compartment calling. According to one study, a subset of regions switches compartments from spermatogonia to spermatocytes, and this coincides with gene expression changes that implicate compartment dynamics in germ cell differentiation [33]. Notably, pachytene PIWI-interacting RNA (piRNA) clusters – genomic regions that are expressed during the mid- to late-pachytene stage – undergo a B-to-A switch in spermatocytes. This is linked to activation by A-MYB [26], which promotes transcription of both piRNA precursors and core piRNA biogenesis factors such as MIWI [47], and thereby couples 3D genome reorganization to germline-specific RNA-based gene regulation. By contrast, another Hi-C study reported compartment reprogramming during meiosis as refined local A compartments, which are smaller than conventional compartments and correlate with pachytene transcription [20]. Together, these findings suggest that compartment remodeling reflects changes in global gene expression during spermatogenesis.

A/B compartments further reflect the behavior of meiotic chromosomes and recombination patterns. Analysis of hybrid mouse strains revealed that homologous pairing initiates in long interspersed nuclear element (LINE)-enriched B compartments and progresses toward short interspersed nuclear element (SINE)-enriched A compartments [35], while meiotic DNA breaks and interhomolog crossovers preferentially occur in the gene-dense A compartments [31]. Contact-probability analyses show that, during early meiotic prophase I, axis loops are shorter in compartment A than in compartment B, but from pachytene onward this difference narrows [34]. Another study, however, reports that compartment A consistently maintains smaller loops than compartment B across all prophase substages [35]. Nevertheless, variations in axis–loop length across compartments may influence recombination machinery accessibility, and thus links 3D chromatin structure to meiotic recombination outcomes.

Other key aspects of meiotic prophase I revealed by Hi-C analyses include telomere attachment to the nuclear envelope where they cluster into a bouquet configuration and promote homolog recognition and synapsis (Figure 3B). Sub-telomeric regions in particular engage in strong interchromosomal interactions [30–32,34]. This telomere-led alignment depends on the LINC (linker of nucleoskeleton and cytoskeleton) complex – a molecular bridge that connects the nucleus to the cytoskeleton – and revealed dynamic chromosome-end organization during early meiosis [34]. Together, 3D genome analysis enables detailed assessment of telomere clustering and bouquet formation – key processes that safeguard accurate homolog recognition and synapsis during meiotic prophase I.

Meiotic sex chromosome inactivation and the XY body

Unlike autosomes, sex chromosomes undergo distinct 3D genome dynamics during male meiosis because of meiotic sex chromosome inactivation (MSCI), a process triggered by the unsynapsed state of the X and Y chromosomes (Box 3). In spermatogonia, the X and Y chromosomes exhibit clear A/B compartmentalization and TAD structures similar to those of autosomes. However, in pachytene spermatocytes, this organization is largely lost [20,30–32]. At this stage, the X and Y chromosomes undergo MSCI and form a specialized structure known as the XY body. It has been proposed that the XY body forms through a phase-separation mechanism which is driven by the DNA damage response (DDR) pathway – a master regulator of MSCI [48,49]. This process results in a highly dynamic chromatin environment in which prominent 3D genome features are not detectable in bulk Hi-C analyses (which lack single-cell resolution). Interestingly, the proteins SCML2 and A-MYB have distinct roles on the sex chromosomes compared to autosomes – they help to establish unique chromatin states on the XY body and promote the spatial segregation of the XY body from autosomes [26]. Since SCML2 and A-MYB function downstream of the DDR pathway on the sex chromosomes, there appears to be a progressive 3D chromatin reorganization of the sex chromosomes during MSCI – which is initiated by the DDR pathway and further remodeled by SCML2 and A-MYB [26].

Box 3. Meiotic sex chromosome inactivation (MSCI).

MSCI is a specialized form of transcriptional silencing that occurs during meiotic prophase I. In male meiosis, MSCI begins at the early pachytene stage when the X and Y chromosomes fail to fully synapse due to their limited homologous region, which is restricted to the pseudoautosomal region. The unsynapsed chromosomal axes are recognized by the DNA damage response (DDR) machinery [48,49]. MSCI initiates when ATR kinase, recruited by proteins such as BRCA1, localizes to unsynapsed axes. ATR phosphorylates histone variant H2AX (γH2AX) along the axis loops of the sex chromosomes and marks them for silencing [90]. γH2AX spreading is mediated by MDC1, which binds to γH2AX and amplifies the DDR signals along the axis loops of the entire sex chromosomes, leading to initiation of MSCI at the early pachytene stage [91]. Additional chromatin modifiers – such as ubiquitin ligase RNF8, SCML2, A-MYB, and components of the SUMO pathway – further establish a unique chromatin state of the sex chromosomes [86,92]. Together, these events lead to the formation of the XY body, a distinct nuclear domain that contains the silenced sex chromosomes at the mid-pachytene stage.

After meiotic prophase I, spermatocytes complete meiosis to generate haploid round spermatids, which then differentiate through spermiogenesis into elongated spermatids and mature sperm (Figure 1B). The silent state of sex chromosomes during pachytene persists into post-meiotic spermatids through the formation of a silent nuclear compartment termed post-meiotic sex chromatin (PMSC) [50]. 3D chromatin analyses reveal that the 3D features established during MSCI persist in PMSC [26,30]. In addition, 3D structural analyses of spermatozoa show that the centralized positioning of PMSC within the nucleus is maintained in spermatozoa [51]. These findings are in line with the idea that the epigenetic states of sex chromosomes are inherited through sperm [52].

The 3D genome in sperm

The 3D chromatin state of postmeiotic round spermatids shares similarity with pachytene spermatocytes. A/B compartments and TADs are partially re-established in round spermatids, and spermatid-specific architectural features are acquired [26]. During the transition from elongated spermatids to mature sperm, most histones are replaced by protamines, and only a small fraction of histones are retained (1–8% in mice and 10–15% in humans) [53,54]. This results in an extremely compacted genome and a chromatin architecture that is markedly different from that of somatic cells and earlier male germ cells. This unique packaging raises important questions about the 3D genome structure in mature sperm and whether epigenetic information carried by the 3D genome contributes to epigenetic inheritance.

The 3D genome features of mouse spermatozoa remain a subject of active debate. While earlier studies detected A/B compartments and TADs in mouse spermatozoa [55,56], others reported that TADs are largely absent, and only A/B compartmentalization was detectable [32,57]. A recent single-cell Hi-C analysis further confirmed the presence of A/B compartments in both mouse and human spermatozoa [51]. The apparent absence of TADs in sperm has also been observed in humans [51,58], zebrafish [59], and Xenopus [60], suggesting that the lack of TADs may be a conserved feature across vertebrate sperm. In place of TADs, mouse sperm exhibits frequent ultra-long-range (>4 Mb) and interchromosomal interactions – that are rare in somatic cells – which emphasizes the fundamentally distinct 3D genome organization [57]. Interestingly, a recent study reported that, in purified mouse spermatozoa, canonical A/B compartments and TADs are largely absent, and the authors argue that earlier reports of these features may have reflected contamination by cell-free chromatin [61]. This challenges the inheritance of 3D chromatin features from sperm to embryos [56,62]. Further investigation will be necessary to resolve these discrepancies.

In Xenopus tropicalis sperm, instead of TADs or A/B compartments, a supersized loop structure (SSL) with a median size of 5 Mb is observed. The anchor sites of SSLs lack CTCF, RNA polymerase II, and modified histones, but almost all are marked by Helitrons, a rolling-circle DNA transposon [60]. Notably, zebrafish sperm display particularly distinctive features that are characterized by ‘hinge-like’ domains of ~150 kb recurring every 1–2 Mb [59]. Unlike in mice and humans, zebrafish sperm chromatin does not undergo histone-to-protamine replacement; instead, the genome is packaged by histone variants [63,64]. This alternative mode of nuclear organization may serve as a means to transmit epigenetic states in sperm.

The 3D genome and paternal epigenetic inheritance

Paternal epigenetic inheritance remains an area of intensive investigation, in part due to the extensive reprogramming of the paternal genome after fertilization [65]. Nevertheless, recent studies have revealed specific mechanisms through which paternal epigenetic information can escape reprogramming and influence the phenotype of the offspring. For example, it was shown that bisphenol A (BPA)-induced obesity is transmitted across generations via CTCF-mediated chromatin remodeling in sperm [66]. BPA exposure during fetal germline reprogramming stabilizes CTCF binding and enhancer–promoter looping at the Fto locus, which correlates with increased food intake. Deletion of the CTCF site disrupts these interactions and prevents obesity inheritance, which underscores the role of 3D genome architecture in epigenetic inheritance. In support, CTCF and cohesin are implicated in maintaining transcriptionally active chromatin marked by paternally derived histone H3 lysine 4 trimethylation (H3K4me3) during reprogramming in preimplantation embryos [67]. Another study using a different transgenerational obesity model also implicates the 3D genome in epigenetic inheritance [68]. Together, these findings highlight sperm 3D genome features as key mediators of paternal epigenetic inheritance.

The 3D genome in the male germline and evolution

A notable feature of 3D genome organization in the male germline is its crucial role in gene regulation and genome evolution. In somatic cells, many TAD boundaries are evolutionarily conserved and are associated with stable gene expression patterns across species [17,69,70]. Comparative Hi-C studies across vertebrates further show that evolutionary genome reshuffling often occurs at TAD boundaries and compartment transitions, which suggests that 3D chromatin folding influences large-scale chromosomal evolution [71]. In spermatogenic cells, these boundaries are particularly important for regulating nearby genes. Disruption of an evolutionarily conserved TAD boundary adjacent to Dmrtb1 and Lrp8 – two genes expressed during spermatogenesis – altered their expression and impaired sperm production [72]. This observation underscores the importance of these structures in maintaining proper transcriptional regulation. Beyond its regulatory role, recent studies have highlighted the involvement of 3D chromatin architecture in shaping the evolutionary plasticity of the germline genome. For example, in post-meiotic stages of spermatogenesis, evolutionary breakpoint regions, – sites where genomic rearrangements have disrupted the conserved gene order – are associated with long-range chromatin interactions that reflect ancestral syntenic associations and ancient chromosome fusions [73]. Furthermore, chromosomal fusions – key drivers of karyotypic evolution – have been shown to alter higher-order chromatin organization in male germ cells and reshape the landscape of meiotic recombination [74]. Together, these findings support the view that 3D genome architecture in the male germline not only regulates precise gene expression but also facilitates evolutionary change [75]. In this context, it is tempting to speculate that the unique 3D genome organization of the sex chromosomes may underlie their distinct evolutionary trajectories compared to autosomes.

The 3D genome in the female germline and maternal epigenetic inheritance

Currently, knowledge of the 3D genome organization in the female germline remains limited. Following sex determination, female germ cells initiate meiosis at ~E14.5 (Figure 1A), in contrast to male germ cells which enter meiosis after birth. As in males, during female meiosis a reduction in the boundary strength of TADs and insulation score is observed from the preleptotene stage onwards [24,35]. In addition, the refined chromatin compartments identified in male pachytene spermatocytes are already apparent in female zygotene oocytes and become further enhanced at the pachytene stage [24]. These findings suggest that male and female meiotic germ cells may share similar mechanisms for maintaining meiosis-specific transcription within the highly condensed chromosomal environment of meiotic prophase.

While autosomes in male and female meiotic germ cells exhibit comparable 3D genome patterns, the X chromosome behaves differently. In female germ cells, the X chromosome adopts an autosome-like pattern in which A/B compartments are still present although they are slightly attenuated, and refined local compartments are emerging [24]. This contrasts with male pachytene spermatocytes where MSCI causes a disappearance of A/B compartments on the X chromosome. Furthermore, female embryonic germ cells display reinforced interactions between B compartments on the X chromosome [24].

After birth, a subset of primordial follicles grow under hormonal stimulation and progress from non-growing oocytes (NGOs), to growing oocytes (GOs), and ultimately to fully grown oocytes (FGOs) (Figure 1C). Notably, A/B compartments disappear after the early GO stage [76]. Instead, FGOs exhibit Polycomb-associating domains (PADs), which are cohesin-independent compartmental domains characterized by broad H3K27me3 signals [76]. Conversely, TADs are largely weakened or lost in FGOs, except in regions where they overlap with PADs. By metaphase II (MII), PADs and TADs are absent, which reflects the typical chromosomal architecture of mitotic chromosomes [76]. However, single-cell Hi-C analyses of oocytes reveal substantial cell-to-cell variability in A/B compartment strength, as well as in the positioning of TADs and loops, suggesting that TADs and loops in bulk Hi-C may primarily reflect frequent chromatin contacts observed in a cell population [77].

Following fertilization, TADs are indistinct in both parental alleles at the zygote and 2-cell stages but are progressively reestablished between the 8- and 64-cell stages [57]. While CTCF remains bound throughout preimplantation development, cohesin gradually associates with chromatin from the 2- to 8-cell stages, and accompanies TAD re-establishment and forms genic cohesin islands that coincide with embryonic hypertranscription [78]. PADs reappear on the maternal allele at the two-cell stage but gradually disappear after the 8-cell stage. On the paternal allele, PADs remain absent [76]. Broad H3K27me3 domains, which are associated with PADs in oocytes, are inherited from mouse oocytes and persist into preimplantation embryos until the blastocyst stage [76,79,80], which suggests that maternal H3K27me3 contributes to PAD formation during early embryogenesis. Of note, maternal H3K27me3 functions as DNA methylation-independent imprinting marks in early mouse embryos [81]. In addition, maternal H3K27me3 inheritance drives the formation of LADs in the zygote [82]. This maternal histone modification-based establishment of distinct 3D genome structures during early development contrasts with the paternal allele where epigenetic states undergo extensive reprogramming [65].

Concluding remarks

High-resolution and single-cell 3D genome mapping has revealed that germ cell development involves extensive reorganization of chromatin compartments, TADs, and chromatin loops which are closely linked to transcriptional regulation, meiotic chromosome dynamics, and epigenetic priming. Distinct architectural features – such as axis–loop organization during meiosis, refined compartments, and ultra-long-range interactions in sperm – underscore the specialized nature of germline nuclear architecture. The nuclear architecture not only plays a key role in the regulation of gene expression but may also facilitate epigenetic inheritance and influence genome evolution. Key challenges ahead include uncovering the molecular mechanisms that regulate the germline-specific 3D genome, clarifying the relationships between different loop types – such as chromatin loops and axis loops – and determining how these features are remodeled post-fertilization (see Outstanding questions). Addressing these questions will reveal broader principles of genome organization and deepen our understanding of germline biology.

Outstanding questions.

What are the characteristics of 3D chromatin organization in germ cells before sex determination? Do these features differ based on genetic sex, or are they consistent across sexes?

What molecular mechanisms underlie the establishment of epigenetic priming mediated by the 3D genome? Is the prospermatogonia stage – that precedes undifferentiated spermatogonia – a critical window for establishing this priming?

How do sex chromosome-specific 3D chromatin structures, including the XY body and PMSC, regulate gene silencing, escape from silencing, and influence post-meiotic gene expression?

To what extent are germline-specific 3D genome features conserved across species, and how do evolutionary variations correlate with differences in reproductive strategies and genome size?

When are PADs established in fully grown oocytes? Beyond the roles of histone modifications such as H3K27me3 and histone H2A lysine 119 ubiquitination (H2AK119ub), how much does 3D chromatin architecture contribute to gene repression within PADs?

How do the dramatic chromatin reorganizations during fertilization and early embryogenesis remodel or erase germline-specific nuclear architecture? Which aspects of the 3D genome are retained to facilitate epigenetic inheritance?

Highlights.

High-resolution and single-cell high-throughput chromatin conformation capture (Hi-C) analyses have uncovered dynamic reorganization of A/B compartments, topologically associated domains (TADs), and chromatin loops throughout germ cell development.

3D genome remodeling during germ cell development is closely coupled to transcriptional regulation, epigenetic priming, and meiotic chromosome dynamics.

The distinct germ cell nuclear architecture underlies epigenetic inheritance and influences genome evolution.

Glossary

CCCTC-binding factor (CTCF)

a DNA-binding protein that defines topologically associating domain (TAD) boundaries, mediates chromatin looping, and insulates regulatory interactions.

Cohesin

a ring-shaped protein complex that mediates sister chromatid cohesion and facilitates loop extrusion, and is essential for chromosome segregation and 3D genome organization.

DNA methylation

an epigenetic modification in which a methyl group is added to cytosine bases in DNA, typically at CpG dinucleotides. The modification leads to transcriptional repression and contributes to genomic stability. In mammalian germ cells, DNA methylation patterns are globally erased and re-established to reset developmental potential.

Histone modifications

covalent post-translational modifications of histone proteins that alter chromatin structure and influence DNA-based processes such as transcription, replication, and repair. Common modifications include methylation, acetylation, phosphorylation, ubiquitination, and SUMOylation, which occur primarily on the N-terminal tails of core histones (H2A, H2B, H3, and H4).

Lamina-associated domains (LADs)

genomic regions in contact with the nuclear lamina that are typically gene-poor, heterochromatic, and transcriptionally repressed.

PIWI-interacting RNA (piRNA)

small noncoding RNAs that interact with PIWI proteins to silence transposable elements and regulate gene expression in the germline.

Polycomb repressive complex (PRC)

multiprotein complexes (PRC1 and PRC2) that deposit repressive histone marks which lead to silencing of developmental or germline genes. PRC1 mediates histone H2AK119 ubiquitination, and PRC2 catalyzes H3K27 trimethylation.

Primordial germ cells (PGCs)

embryonic precursors of sperm and oocytes that undergo extensive epigenetic reprogramming to erase parental epigenetic marks and regain developmental potential.

Protamine

small, arginine-rich nuclear proteins that replace most histones during the final stages of spermatid maturation, leading to chromatin compaction and protection of paternal DNA in the sperm.

REC8

a meiosis-specific cohesin subunit that is essential for axis–loop formation, homolog pairing, and recombination during meiotic prophase I.

Spermatogonialstem cells (SSCs)

a subset of undifferentiated spermatogonia that self-renew and give rise to differentiating germ cells, and thus ensure lifelong spermatogenesis.

Super-enhancers (SEs)

large clusters of enhancers spanning several kilobases that are densely occupied by transcription factors, coactivators, and the Mediator complex. Super-enhancers drive exceptionally high levels of transcription of target genes, often those that define cell identity or regulate developmental programs.

Synaptonemal complex protein 3 (SYCP3)

a component of the synaptonemal complex axial element that organizes meiotic chromosome axes and contributes to the formation of the axis–loop structure.