Heterochromatin in plant meiosis

ABSTRACT

Heterochromatin is an organizational property of eukaryotic chromosomes, characterized by extensive DNA and histone modifications, that is associated with the silencing of transposable elements and repetitive sequences. Maintaining heterochromatin is crucial for ensuring genomic integrity and stability during the cell cycle. During meiosis, heterochromatin is important for homologous chromosome synapsis, recombination, and segregation, but our understanding of meiotic heterochromatin formation and condensation is limited. In this review, we focus on the dynamics and features of heterochromatin and how it condenses during meiosis in plants. We also discuss how meiotic heterochromatin influences the interaction and recombination of homologous chromosomes during prophase I.

KEYWORDS:

• Centromere
• heterochromatin
• meiosis
• recombination
• telomere

Introduction

Eukaryotic chromosomes are composed of DNA, RNA, and proteins which are collectively called chromatin. The basic structural component of chromatin is the nucleosome, which typically consists of two copies of each of the core histones H2A, H2B, H3, and H4 wrapped with ~146 bp of DNA [1]. The core nucleosome can also be associated with the linker histone H1 which aids in nucleosome stabilization and compaction [2]. Nucleosomes dynamically regulate the accessibility of the genetic information through varied chromatin states. Chromatin can be subdivided into heterochromatin and euchromatin, which were classically defined by differential staining and are now identified by sequence features, nucleosome density, histone, and DNA modifications, and transcriptional states. Euchromatin is characterized by actively transcribed genes with low density of nucleosomes, whereas heterochromatin comprises repetitive sequences and transposable elements (TEs) with highly condensed nucleosome array. Heterochromatin is transcriptionally silenced and sub-classified into two types: facultative heterochromatin and constitutive heterochromatin. The facultative type is defined as regions that are condensed and transcriptionally silent, but can become ‘open’ and transcriptionally active in specific circumstances, such as mammalian X chromosomes during embryo development [3]. Most heterochromatin is constitutive and remains silenced, including pericentromeric regions and telomeres [4,5]. The establishment of heterochromatin restrains the activity of TEs, isolates damaged repetitive DNA, and ensures proper chromosome segregation, which are essential functions for maintaining genomic stability in both mitosis and meiosis [3,6].

Meiosis is essential for sexual reproduction in eukaryotes. In meiosis, germ cells divide twice after one round of DNA replication, producing haploid gametes. Meiotic chromatin and nuclear organization differ dramatically from somatic or mitotic cells, with specific properties and dynamics [7–9]. During mitosis, chromatin fibers rapidly coil into discrete rod-shaped chromosomes from prophase to metaphase [9]. In contrast, the process of chromosome compaction during meiosis prophase I is a long-duration event, with homologous chromosomes interacting and condensing from thin thread-like structures (leptotene) to short rod-shaped bivalents (metaphase I) [10]. During this period, the core events of meiosis, including pairing, synapsis, and recombination between homologous chromosomes, are essential for the exchange of genetic information and segregation of homologs [11]. The properties and dynamics of highly condensed meiotic heterochromatin are important for meiotic homologous recombination and chromosome segregation, but the specific mechanisms of heterochromatin condensation during meiosis are not well understood. In this review, we will use examples from plants supplemented with relevant studies from other model organisms to discuss the features of meiotic heterochromatin with special attention paid to condensation, and summarize how heterochromatin influences meiotic recombination and genome stability.

The features and epigenetics of pericentromeric heterochromatin

Constitutive heterochromatin is mainly present in pericentromeric regions, centromeres, and telomeres [4,5]. The organization of pericentromeric heterochromatin in multicellular eukaryotes appears to use shared principles [3], but many mechanistic details remain elusive, especially in plants. Heterochromatin and euchromatin are distinguished by epigenetic features, including DNA methylation, histone modifications, and the presence of histone variants (Figure 1). In most eukaryotes, DNA in pericentromeric heterochromatin is densely methylated (Figure 1), which contributes to the heterochromatin structure and silencing of transposons and other repeats [3,12]. In plants, the majority of DNA methylation occurs in the repeat-rich pericentromeric regions where cytosines are methylated in three sequence contexts: CG, CHG, and CHH (H=A, T or C) [12,13]. CG methylation is catalyzed by DNA Methyltransferase 1 (MET1), which interacts with PCNA (proliferating cell nuclear antigen) and methylates the hemi-methylated CG dinucleotides during DNA replication in a process known as maintenance DNA methylation [14,15]. CHG methylation is mainly catalyzed by methyltransferase Chromomethylase 3 (CMT3), which positively correlates with H3K9me2 [12,16]. CHH methylation is generated by the Domains Rearranged Methylase 2 (DRM2) methyltransferase and regulated by the RNA-directed DNA methylation (RdDM) pathway [12,17]. CHH methylation is also promoted by Decreased DNA Methylation 1 (DDM1) and CMT2 at histone H1-containing heterochromatin independent of RdDM [18].

Figure 1. Structure and characteristics of chromosomes in arabidopsis. (a) The model of chromosome, which consists of centromere (purple), pericentromeric heterochromatin (red), telomere (orange) and euchromatin (blue). Pericentromeric heterochromatin and euchromatin are classified by sequence features, transcriptional states, epigenetic modifications and nucleosome density. The euchromatin is enriched with actively transcribed genes that possess loosely organized H3.3-nucleosomes marked by H3K4me3. The heterochromatin is composed of repetitive sequences and transposons with highly methylated DNA and condensed H3.1- or H2A.W-nucleosomes modified by H3K9me2 and H3K27me1. (b) Immunostaining of heterochromatic markers (5-mC, H3.1 and H3K27me1) at pachytene of meiosis, mitotic prophase and interphase of somatic cells of arabidopsis. The 5-mC, H3.1, and H3K27me1 signals show co-localization with darkly staining pericentromeric regions. SYN1/REC8 is meiosis-specific cohesion labeling chromatin. Scale bars: 5 μm.

Heterochromatin is characterized by specific histone variants and modifications. In plants, the two main histone H3 variants are H3.1 and H3.3 [19]. Histone variant H3.1 is incorporated into newly synthesized chromatin through the Chromatin Assembly Factor-1 (CAF1) complex in a DNA replication-dependent manner [20,21]. Histone variant H3.3 deposition correlates with active euchromatic genes by replacing H3.1, resulting in the enrichment of H3.1 in heterochromatin [22–24]. Deficiency of CAF1 leads to impaired enrichment of H3.1, derepression of transposable elements (TEs), and pericentromeric heterochromatin deficiency in plants [20,25]. The plant-specific H2A variant H2A.W also promotes heterochromatin formation [26], and linker histone H1 mediates heterochromatin condensation and TEs silencing [27]. Interestingly, H3.1, H2A.W, and H1 are all associated with DNA methylation pathways that maintain the suppression of TEs [27,28].

The histone modifications H3K9me1/2 and H3K27me1 are usually present in the tails of H3.1 and co-occur with H2A.W in heterochromatic regions [21,24,26,29]. In Arabidopsis, the histone methyltransferases SU(VAR)3–9 HOMOLOG 4 (SUVH4), SUVH5, and SUVH6 mediate H3K9me1/2 modification, which primarily occurs in pericentromeric heterochromatin [30,31]. H3K9me2 and CHG DNA methylation form a reinforcing loop [32]. From S. pombe to mammals, heterochromatin protein 1 (HP1) is an evolutionarily conserved protein that serves as a hub for heterochromatin formation [33,34]. HP1 localizes at pericentromeres by recognizing methylated H3 lysine 9 (H3K9me) and recruits histone modifiers, RNAi components, chromatin remodelers, and cohesins to facilitate heterochromatin condensation [35,36]. HP1 can also promote heterochromatin compaction via phase separation [34]. In plants, Like Heterochromatin Protein 1 (LHP1), an ortholog of HP1, recognizes H3K27me3 (H3 lysine 27 trimethylation) on euchromatin and is required for transcriptional regulation [37,38], and a multivalent H3K9me reader ADCP1 has been found to play a role in heterochromatin formation and phase separation by modulating H3K9me and DNA methylation [39,40]. Polycomb Repressive Complex 2 (PRC2) has conserved functions in animals and plants in catalyzing H3K27me [41,42]. In plants, H3K27me1 is preferentially located in constitutive heterochromatin, while H3K27me3 is mainly restricted to gene-rich euchromatin (Figure 1) [43,44]. Intriguingly, plant PRC2 is primarily responsible for H3K27me2/3, while more recently evolved Arabidopsis Trithorax-Related proteins 5 and 6 (ATXR5 and 6) catalyze H3K27me1 of H3.1 preferentially on heterochromatin [29,45]. ATXR5 and 6 are essential for chromatin condensation, genomic stability and transcriptional repression within pericentromeric heterochromatin [46–48].

Other factors also contribute to heterochromatin formation. For example, two members of the conserved Microrchidia (MORC) adenosine triphosphatase (ATPase) protein family, MORC1 and MORC6, form a hetero-dimer that represses TE expression and promotes heterochromatin condensation in Arabidopsis [49]. The dimer is proposed to interact with RdDM components and the SWI/SNF chromatin-remodeling complex to mediate TE silencing and chromatin condensation [50,51]. Additionally, Arabidopsis MORC4 and MORC7 can form homodimers that have redundant roles in repressing gene expression and establishing RdDM [52,53]. In addition to their classical functions in DNA replication, DNA replication factors also contribute to the formation and inheritance of heterochromatin [54]. PCNA not only plays a role in CG DNA methylation maintenance but also is required for H3.1 deposition and H3K27me1 within heterochromatin during DNA replication [12,55,56]. In S. pombe, DNA polymerase epsilon (POL ε) collaborates with RNAi machinery and histone methyltransferase of H3K9me2 to recruit Swi6 (the fission yeast HP1 orthologue), thereby promoting heterochromatin assembly [57,58]. In S. cerevisiae, POL ε is required for the inheritance of silenced telomeres and rDNA after replication [59,60]. Similarly, Arabidopsis POL ε also functions in heterochromatin condensation and TE silencing in somatic cells, and POL ε can prevent CHG DNA hypermethylation of heterochromatin thereby safeguarding DNA replication [61].

Heterochromatin condensation in meiosis

In mitosis, chromatin rapidly condenses into rod-shaped chromosomes from prophase to metaphase, whereas meiotic chromatin undergoes relatively slow condensation from thread-like structures to short rod-shaped bivalents during prophase I, which is divided into 5 stages (Figures 1b and 2) [9,62]. In early meiotic prophase, homologs, comprised of two sister chromatids, are organized as co-oriented linear arrays DNA-loops, that are highly packed and joined by a proteinaceous structure known as the chromosome axis [7,63]. The structural organization of the mitotic prophase axis is directly comparable to that observed in meiosis. The axes in both processes consist of a complex structural component, including cohesins, condensins, and topoisomerases [7,63]. Meiotic chromosome axes contain meiotic specific HORMA domain proteins, such as ASY1, ASY3 and ASY3 in Arabidopsis [64–66], which are important for early chromosome architecture and homolog synapsis during prophase I [7,11]. Cohesin and condensin are related complexes that consist of SMC (Structural Maintenance of Chromosomes) proteins and are essential for chromatin axis formation and condensation during cell division [67–70]. Some subunits of cohesin and condensin are used in both meiosis and mitosis, but other subunits are specifically involved in meiotic condensation and chromosome segregation, including the meiosis-specific cohesin subunit SYN1/REC8 [68,71,72]. The condensation of meiotic chromosomes in meiosis I is characterized by two distinct processes. The first process occurs during early prophase I, where chromosomes perform elongated structures mainly organized around the central axis composed of cohesins. The second process resembles mitotic-like longitudinal compaction and takes place after recombination, as meiotic cells prepare for division. This compaction primarily relies on condensins. Moreover, cohesin has different localization patterns and functions in meiosis I and II. In meiosis I, cohesin holds sister chromatids together and localizes along the whole chromosome as part of the chromosome axes and has been proposed to regulate chromosome loop size, which is required for homolog pairing and synapsis [7,72]. Prior to anaphase I cohesin is released from the chromosome arms to facilitate segregation of homologs, but is retained near centromeres to keep the sister chromatids paired until anaphase II [67,73,74].

Figure 2. Chromosome morphology during meiotic prophase I. (a) Diagrams showing the process of chromosome condensation from leptonemato metaphase I. (b) The DAPI (dark) and centromere (red) signals are shown from leptonema to metaphase I of Arabidopsis by chromosome spread. Centromeres are signed via FISH assay using 180-bp repeat DNA probes. Pericentromeric heterochromatin is visualized as dark-staining chromatin adjacent to centromeres. Scale bars: 5 μm.

Normal heterochromatin architecture with centromeric cohesion is required for faithful chromosome segregation during both mitosis and meiosis [62,67,75]. In Arabidopsis, SCC2 (Sister Chromatid Cohesion protein 2) is indispensable for loading cohesin during both mitosis and meiosis. Consequently, scc2 null alleles result in embryonic lethality, while weak alleles of scc2 exhibit meiotic defects in meiosis-specific cohesion loading, synapsis, recombination, and chromosome condensation [76,77]. Meiosis-specific cohesin subunit SYN1/REC8 is conserved in a wide range of eukaryotes and is required for chromosome condensation, segregation, and DSB repair [75,78]. The REC8-cohesin complex exhibits significant accumulation in the pericentromeric heterochromatin of meiosis in yeast and plants [72,79,80]. In S. pombe, Swi6 binds H3K9me to recruit meiosis-specific cohesin complexes, further demonstrating the importance of cohesin for meiotic heterochromatin architecture [81]. In addition, Arabidopsis Male Meiocyte Death 1 (MMD1), a PHD-finger protein, specifically regulates meiotic chromatin condensation, including heterochromatin, by interacting with the histone demethylase JMJ16 to promote the expression of condensin CAP-D3 [82–84].

Similar epigenetic features are associated with heterochromatin have been detected using immunofluorescence in both meiosis and mitosis, including DNA methylation, histone modifications, and histone variants (Figure 1) [85,86]. A recent study reveals that the catalytic subunit of POL ε (POL2A) recognizes histone H3.1-H4 and recruits ATPase MORC1-MORC6 to function in meiotic heterochromatin formation and condensation in Arabidopsis [86]. POL2A selectively recognizes histone variant H3.1-H4 via a ZF1 (Zinc finger 1) domain in its C terminus and probably helps mediate H3.1-H4 deposition with CAF1 during replication [86]. Subsequently, POL2A recruits MORC1-MORC6 via its N terminus to function in heterochromatin condensation during meiosis [86]. Although POL2A is also involved in heterochromatin condensation in somatic cells, the mechanism may differ from that in meiotic cells [61]. In somatic cells, POL2A is essential for heterochromatin organization and TE silencing by maintaining proper DNA replication [61]. Heterochromatin condensation and silencing are compromised in three pol2a mutants with mutations in its DNA polymerase domain. These condensation defects are accompanied by compensatory CHG hyper-methylation, while exhibiting minimal effects on H3K27me1, H2A.W, and siRNA [61]. By comparison, mutations in N terminus of POL2A result in decondensed structure, decreased nucleosome density and H3K27me1 modification of meiotic heterochromatin [86]. Interestingly, the N terminus of POL2A interacts not only with MORC1 but also SU(VAR)3–9 homologs SUVH2 and SUVH9 (two RdDM components), and POL2A does not affect TE silencing but is required for CHH methylation in meiosis [87]. These findings suggest that POL2A plays a crucial role in the organization of heterochromatin in both meiotic and somatic cells in plants, but the mechanisms regulated by POL2A may differ from each other [25,49,86]. MORC1 was initially identified in mice and its mutation causes arrest of spermatogenesis in early prophase I [88]. Further study showed that MORC1 is responsible for TEs repression and DNA methylation in the mouse male germline [89]. In the C. elegans germline, CeMORC1 plays an important role in heterochromatin silencing and siRNA-dependent chromatin condensation by preventing the invasion of H3K36me3 into heterochromatin [90]. Deficiency of CeMORC1 and nuclear RNAi leads to decondensation of the germline chromatin and germline mortality [90]. These results indicate the important role of MORC proteins in heterochromatin condensation during germline development. These examples show that meiosis and mitosis share several epigenetic and genetic mechanisms to regulate heterochromatin condensation, albeit with some specific factors in meiosis.

Meiotic recombination rarely occurs within heterochromatin

Recombination, as a key event during meiosis, is initiated by the formation of double-strand breaks (DSBs) generated by an evolutionarily conserved DNA topoisomerase-VI-like complex including SPO11 (SPORULATION 11) [91]. After excising short DNA segments from one end of the break to generate 3’-OH single-stranded DNA, this DNA can invade its allelic partner on one homolog chromatid, forming a D-loop. Subsequently, a few recombination intermediates through DNA synthesis templated from the homologous chromosome and ligation are designated to become crossovers (COs). The remaining majority of intermediates are destined to mature without the exchange of flanking regions, referred to as noncrossovers (NCO) [10,11,92]. Recombination ensures the proper segregation of homologs after metaphase I by facilitating homolog pairing and producing COs to link them during prophase I (Figure 2) [11]. Recombination also generates new allelic combinations and phenotypic diversity among gametes and progeny.

COs are not randomly distributed, but instead occur preferentially in regions known as recombination hotspots [93]. Chromosome organization has a significant influence on the formation and distribution of DSBs and COs, which predominantly occur in euchromatin regions that are enriched with H3K4me3 and H2A.Z, have low nucleosome density (LND) and low DNA methylation [93–103]. H3K4me3 methyltransferases and readers of their activity play a crucial role in determining the localization of DSBs in euchromatin in yeast and animal [104,105]. In mouse and human, a SET domain protein with the zinc finger PRDM9 recognizes the specific DNA motifs and catalyzes H3K4me3 to initiate the DSB formation [106,107]. Although PRDM9 subfamily is lost in plants, in Arabidopsis, DSBs tend to occur within promoter or terminator regions of euchromatin genes that are associated with a series of chromatin marks that promote RNA polymerase II transcription, such as proximity to H3K4me3, with hypomethylated DNA and LND [93,108]. Furthermore, Arabidopsis DSB hotspots exhibit a significant association with certain transposable elements, namely the Helitron, Pogo/Tc1/Mariner and MuDR families that are enriched in proximal regions of the chromosome arms. Conversely, RNA transposons including LTR (long terminal repeat), Copia LTR and LINE-1 are rarely found to overlap with DSBs and tend to be enriched in proximity to the centromeres. These observations indicate that these features may contribute to the chromatin reorganization in euchromatin regions, which is required for the DSB formation. In contrast, due to the repression of DSBs formation in heterochromatic regions with contrasting characteristics, crossover events (COs) are typically suppressed in heterochromatic regions, particularly in the pericentromeric regions [109,110]. This is thought to be due to the inaccessibility of highly condensed heterochromatin and selective pressures against the negative consequences of COs within repetitive sequences that pose a risk to chromosome segregation and genome integrity [111]. The occurrence of COs in pericentromeric or centromeric regions can disrupt cohesin loading and prevent the proper association between kinetochores and spindles during segregation [111,112]. For instance, mis-segregation of chromosomes due to COs near the centromere is correlated with birth defects in human [113]. Pericentromeric COs are associated with premature separation of sister chromatids and lead to sporulation defects in S. cerevisiae [114]. Genome-wide sequencing has enabled high-resolution maps of meiotic recombination in Arabidopsis [95,97,102,109], rice [115,116], tomato [117,118], wheat [119–121], maize [122,123], barley [124] and soybean [125,126], confirming that centromeric and pericentromeric CO suppression is a common feature of plant species.

In addition to the physical exclusion of recombination proteins by highly condensed heterochromatin, epigenetic modifications also play important roles in regulating COs near plant centromeres. DNA methylation commonly functions in gene silencing and correlates with heterochromatin formation [12]. In plants, MET1 is the methyltransferase for CG DNA methylation, which represents the predominant methylation pattern in plants [12]. In met1, the total number of COs is not altered, but they are redistributed along the chromosomes. The number of COs decreases in pericentromeric heterochromatin and increases on euchromatin and centromere-proximal regions [97]. Furthermore, several investigations show that loss of DNA methylation leads to redistribution of meiotic recombination along chromosomes with increased rates in chromosome arms and reinforced or maintained suppression within pericentromeric regions [98,99,127]. These results are surprising given prior hypotheses that DNA methylation is a suppressor of recombination within the heterochromatin near centromeres. Another study provides a possible explanation by showing that while MET1 influences CG DNA methylation, recombination is also suppressed by RdDM (RNA directed DNA methylation) inducing non-CG (CHH and CHG) DNA methylation, H3K9me2, or increased nucleosome occupancy [95]. Consistent with this observation, both CG and non-CG DNA methylation repress meiotic DSBs formation, whereas non-CG methylation and/or H3K9me2 directly inhibit COs in pericentromeric regions [94,108]. Similarly, the RNAi pathway and H3K9 methyltransferase Clr4-Rik1 function in suppressing meiotic recombination within centromeric and pericentromeric heterochromatin in S. pombe [128]. Swi6 can recruit meiosis-specific cohesin by recognizing H3K9me, thereby repressing DSB formation in meiotic pericentromeric heterochromatin [81]. Furthermore, in mice, the dnmt3l (required for de novo DNA methylation) and miwi2 (functions in small RNA-dependent TE silencing) mutants exhibit loss of DNA methylation and H3K9me2, as well as an increase in deleterious DSBs and recombination within retrotransposons, resulting in meiotic arrest [129]. Therefore, repression of DSBs formation likely makes a significant contribution to the inhibition of recombination; however, other factors such as specific histone and DNA modifications play a role in the maturation of COs.

In some species, CO frequency increases from the centromere toward the telomeres, and is high close to the telomeres [121,124,130–134]. In S. cerevisiae, the telomeres and centromeres are both identified as ‘DSB cold’ regions [135], and the recombination frequency in subtelomeric regions is much lower than in the adjacent unique gene-rich regions or the genome-average frequency [136–138]. It has been suggested that a change in the homeostatic balance between CO and NCO with a switch from interhomolog to intersister repair in telomeres is responsible for this pattern [138]. In maize, hexaploid wheat and barley, the distribution of COs exhibits a gradual decline from telomere to centromere, with a notable concentration in the telomeric-near regions [133]. In contrast, CO frequencies show a slight increase from telomere to centromere in Arabidopsis [102,139]. Sex-specific patterns are also observed near telomeres in some species; for example, elevated CO frequencies have been observed near telomeres in human and mouse males [140]. In Arabidopsis, female CO rates are lower in telomeric regions and higher in proximal regions; however, male COs remain high in distal regions including telomeres [132]. The distribution of meiotic recombination is probably correlated with the structural and functional features along the chromosomes, such as GC content, gene density, gene types, transcription, and chromosome modifications, but these data do not reflect a general pattern for all species [108,121,124,134,141].

Centromere coupling during meiosis

Centromeric DNA ranges from discrete motifs of a few hundred base-pairs, as seen in some fungal species such as S. cerevisiae, to megabase-sized arrays of satellite repeats and retrotransposons typical in most multicellular eukaryotes [109,142]. Centromeres function in orchestrating the assembly of the kinetochore complex, facilitating precise attachment of the kinetochore to the spindle and ensuring accurate chromosome segregation [142,143]. Pairing of homologous chromosomes is critical for meiotic recombination. In S. cerevisiae and most plants, centromeres cluster during early prophase I of meiosis [144–150]. Centromeres of non-homologous chromosomes are associated with the leptotene–zygotene transition stage. Subsequently, centromeres pair which aids in the synapsis of homologous chromosomes from zygotene to pachytene [148–151]. During this process, recombination-independent non-homologous centromere clustering draws the meiotic chromosomes into close proximity, promoting recognition and pairing of homologous chromosomes (Figure 2) [148].

Heterochromatin, cohesin, and the synaptonemal complex (SC) help promote centromere associations during meiotic prophase I. In mouse, centromeres do not appear to cluster in early prophase I, but pair in late prophase I [152,153]. However the meiosis-specific cohesin component STAG3 is important for the formation of pericentromeric heterochromatin clustering at leptotene and zygotene [154]. Furthermore, the SC is required for the homologous pericentromeric heterochromatin pairing at pachytene, which further mediates homologous centromere connections from heterologous centromere clusters [144,153]. Interestingly, following the dissolution of SC-induced centromere pairing during diplotene, the pericentromeric heterochromatin with centromeres of the homologs remains persistently associated, thereby maintaining connections between homologous chromosomes even without crossover tethering [144,153]. Therefore, SC plays a role in the formation but not maintenance of homologous heterochromatin/centromere connections. However, it is still unknown what maintains heterochromatin connections with centromeres between homologs without SC. In S. cerevisiae, the central SC element Zip1 and Rec8 play an important role in the interaction between nonhomologous centromeres independently of Spo11 [151,155,156], and the zip1 mutants show increased recombination rate in pericentromeric regions [138].

In plants, ZYP1 (the ortholog of yeast Zip1) does not participate in meiotic centromere coupling, whereas REC8 is still required [157,158]. In maize, centromere clustering is independent of ZYP1, but REC8 and the SC component Structural Maintenance of Chromosome 6 (SMC6) have effects on centromere pairing [158]. Furthermore, the transition from non-homologous centromere clustering to homologous centromere pairing seems to rely on the initiation of meiotic recombination by SPO11 [145,148,155]. In Arabidopsis, homologous centromere pairing partially depends on the meiosis-specific recombinase DMC1 and SPO11, while RAD51 is required for subsequent pairing and synapsis of chromosome arms [145].

Non-homologous centromere clustering leads to homologous pairing and bivalent formation after pachytene [86,148,149]. Intriguingly, deficiency of POL2A or MORC1 not only compromises meiotic heterochromatin condensation, but also leads to defects in disassembly of non-homologous centromere clustering after pachytene [86]. In pol2a-1 and morc1, meiotic chromosomes show centromere coupling at late leptotene, but aberrantly maintain centromere clustering with heterochromatin association from zygotene to diakinesis, which may lead to meiotic genome instability [86]. In polyploid wheat, the Ph1 (pairing homoeologous 1) locus plays an important role in the switch from centromere clustering to homologous pairing upon entry into meiosis [62,159,160]. Ph1 functions in homologous pairing by regulating the remodeling of heterochromatic regions. In absence of Ph1, heterochromatin assumes a more open state, which leads to the abnormal association of repetitive regions on homoeologous chromosomes, which can lead to ectopic recombination [62,161]. The mechanisms that facilitate clustering and separation of meiotic centromeres with pericentromeric heterochromatin in prophase I remain to be fully characterized.

Telomere bouquet in meiosis

Telomeres are composed of repeated sequences and function to protect the ends of chromosomes from degradation and loss due to the end replication problem [162]. Telomeres also participate in chromosome organization, pairing, synapses, and movement during meiotic prophase [163,164]. In many organisms, the telomeres cluster together to form a bouquet that facilitates homologous interactions in early prophase I of meiosis. Telomere bouquets are required for homologous pairing, synapsis, and recombination [165,166]. Surprisingly, telomere bouquets also play a role in centromere assembly and spindle attachment that promotes proper chromosome segregation in S. pombe [167], but similar phenotype is not observed in other organisms.

In many fungi and animals, telomeres cluster on the nuclear envelope (NE) and form a bouquet microtubule organizing center (MTOC) in meiotic prophase I, which is regulated by several telomere-binding proteins, including the Sad1/UNC-84 (SUN) homology and Klarsicht/ANC-1/Syne (KASH) homology (SUN/KASH) complexes [166]. In S. pombe, a SUN domain protein Sad1 and a putative KASH domain protein Kms1 form a transmembrane complex linking telomeres to microtubules and cytoplasmic dynein for telomere clustering [168]. Bouquet1 and − 2 forming dimers function as a bridge to link Sad1 with telomere protein Rap1, thereby tethering telomeres to the spindle pole body (SPB, the MTOC in yeast) [168,169]. Additionally, Bouquet3 and − 4 also target telomeres to the SPB to promote bouquet formation [170]. In mammals, telomere-bound SUN/KASH complexes with dynein and microtubules to link the nucleoskeleton and cytoskeleton thereby facilitating the movement of telomeres on the NE [171–173]. SUN1 and SUN2 are components of the linker of nucleoplasm and cytoplasm (LINC) complexes and function, with partial redundancy, in recruiting telomeres to the NE and promoting bouquet formation in meiosis [167,173–175]. The meiosis-specific KASH protein, KASH5, interacts with SUN1 and regulates localization of telomeres in a SUN1-dependent manner [172,176]. Deficiency of SUN1 or KASH5 prevents telomere association with NE, leading to impaired homolog synapsis and sterility [172,173]. Recent reports found that the transmembrane protein KASH5 functions as a dynein activating adaptor in driving telomere dynamics in meiotic prophase I, which binds and converts dynein into a processive motor [177,178]. In summary, the LINC complexes, consisting of SUN on inner nuclear membrane and KASH on outer nuclear membrane, play a crucial role in connecting the nucleoplasm and cytoplasm across the NE, thus facilitating the transmission of force from the cytoskeleton to the nuclei. The force is finally transduced onto telomeres via telomere-binding proteins and promotes the movement of telomeres along the NE to transiently cluster in a ‘bouquet’ that brings homologs into close proximity for homologous pairing, synapsis and recombination [166,172,175–178].

In plants, telomere-led movements are also conserved and function in facilitating homolog pairing and recombination [179]. In Arabidopsis, the homologous telomeres pair in early prophase and move to the nuclear membrane to promote synapsis (Figure 2). However, telomere clustering is very loose and forms a transitory bouquet because of the absence of a MTOC [163,180]. In Arabidopsis and rice, SUN1 and SUN2 (SUN domain proteins) function in telomere bouquet formation independent of DSB formation and affect homologous pairing and synapsis [181,182]. In wheat, subtelomeric regions participate in homolog recognition and pairing [183]. In barley, both chromosome axes and synapsis initiate at the telomeric regions, which is required for synapsis and recombination of homologs [184]. In maize, PAM1 (plural abnormalities of meiosis 1) was identified to function in telomere clustering and movement on NE, and deficiency of PAM1 leads to impaired homologous synapsis and delayed DSB repair [185]. However, the biochemical function and mechanism of PAM1 in telomere clustering are still elusive. Further investigation is required to understand the formation of telomere bouquets and related proteins in plants.

Discussion and perspectives

An important challenge in studying meiotic heterochromatin in plants is that epigenomic profiling of meiocytes is relatively difficult compared to somatic or mitotic cells. Plant meiocytes are typically low in abundance, buried within somatic tissue, and are often found as a heterogeneous mixture of stages. In recent years, advances in germ cell isolation by microdissection [87,186–188] or flow cytometry [189,190] and low‐input epigenomic sequencing techniques [191,192] have enabled the analysis of the DNA methylation landscape, small RNA and transcription dynamics in germline cells, including meiocytes of plants [87,190,193–196]. In animals, epigenetic reprograming occurs during meiosis, including DNA methylation and histone modifications in heterochromatin [197,198]. As mentioned above, CG, CHG and CHH methylation are the three main types of DNA methylation in plants, which are enriched in heterochromatin. Interestingly, in Arabidopsis, reprograming of DNA methylation in the male germline is driven by tapetal 24‐nt siRNAs [189]. Furthermore, it is observed that CHH methylation levels are relatively low in heterochromatin regions compared to somatic cells, while CG or CHG methylation largely remains unchanged [87,190]. However, the mechanism for establishing CHH methylation and the effects of the CHH reprograming on meiotic heterochromatin are still elusive. Moreover, the dynamics of histone modifications are also unclear in plant meiocytes. Although significant progress in meiotic epigenetics has been made in the past 10 years, the association between histone modifications and meiotic recombination in plants is predominantly inferred from epigenomic sequencing data obtained from somatic cells [72,93,94,102]. Thus, the landscapes of meiotic histone modifications need to be further defined in order to understand the specific characteristics and regulatory mechanisms that govern meiotic heterochromatin condensation.

The roles of heterochromatin during meiosis are largely unknown in most species. Emerging evidence support the idea that recombination suppression in heterochromatin is a common feature in most eukaryotes [11,93,100,103,137,140], which could be mediated by several potential mechanisms. First, epigenetic factors enriched in heterochromatin involved in DNA methylation and H3K9me2 were demonstrated to function in repressing COs. Second, meiotic recombination initiates from DSBs, whose formation is commonly repressed in heterochromatin [108,141], likely due to the compact chromatin. Third, selection of the meiotic DSB repair pathways may be different in heterochromatin and euchromatin, providing a possible explanation why DSBs are observed in pericentromeric regions, but are rarely repaired as COs [102,108,109]. Alternatively, the underlying mechanisms could also be different across species, due to differences in chromatin architecture. In S. pombe, the histone reader Swi6 specifically binds to H3K9me on pericentromeric heterochromatin, which impedes the recruitment of the machinery for meiotic DSB formation, including Spo11 [81]. In mammalian, PRDM9 (a SET domain protein with the zinc finger) recognizes the specific DNA motifs and catalyzes H3K4me3 to promote DSB formation on euchromatin [106,107]. In the model plant Arabidopsis, DSB or recombination hotspots are also closely associated with H3K4me3 and specific motifs on euchromatin [102,199]. However, plants have neither a functionally conserved Swi6 ortholog nor an ortholog of PRDM9. Taking advantage of newly developed technologies such as ultramicro sample sequencing and super-resolution microscopy, will aid in the investigation of commonalities and differences in the roles of heterochromatin in meiosis across species.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

CW and YW were supported by the National Natural Science Foundation of China [31925005, 32000246 and 32370350], and Guangdong Laboratory for Lingnan Modern Agriculture [NG2022002] and Double first-class discipline promotion project of SCAU [2023B10564004]. GPC was supported by UNC Chapel Hill.