1. Chromosomes and Gene Expression
  2. Plant Biology
Download icon

Natural depletion of histone H1 in sex cells causes DNA demethylation, heterochromatin decondensation and transposon activation

  1. Shengbo He
  2. Martin Vickers
  3. Jingyi Zhang
  4. Xiaoqi Feng  Is a corresponding author
  1. John Innes Centre, United Kingdom
Research Article
  • Cited 2
  • Views 1,818
  • Annotations
Cite this article as: eLife 2019;8:e42530 doi: 10.7554/eLife.42530

Abstract

Transposable elements (TEs), the movement of which can damage the genome, are epigenetically silenced in eukaryotes. Intriguingly, TEs are activated in the sperm companion cell – vegetative cell (VC) – of the flowering plant Arabidopsis thaliana. However, the extent and mechanism of this activation are unknown. Here we show that about 100 heterochromatic TEs are activated in VCs, mostly by DEMETER-catalyzed DNA demethylation. We further demonstrate that DEMETER access to some of these TEs is permitted by the natural depletion of linker histone H1 in VCs. Ectopically expressed H1 suppresses TEs in VCs by reducing DNA demethylation and via a methylation-independent mechanism. We demonstrate that H1 is required for heterochromatin condensation in plant cells and show that H1 overexpression creates heterochromatic foci in the VC progenitor cell. Taken together, our results demonstrate that the natural depletion of H1 during male gametogenesis facilitates DEMETER-directed DNA demethylation, heterochromatin relaxation, and TE activation.

https://doi.org/10.7554/eLife.42530.001

eLife digest

In most organisms, the genetic information contains DNA segments called transposable elements which are able to move around in the genome. When transposable elements insert themselves in a new location, this can lead to negative outcomes such as cell death or cancer.

Animals, plants and other organisms have evolved sophisticated mechanisms to ‘lock in’ transposable elements and prevent them from jumping from place to place in the genome. For instance, adding small molecules called methyl groups onto these sequences tightly packages the DNA, which wraps itself around proteins known as histones. The resulting structure is known as heterochromatin, and it limits the movement of the transposable elements.

In certain situations, cells may ‘reactivate’ some of their transposable elements: this is for example the case in plant sperm companion cells, which protect the sperm and deliver them to the egg cell. However, it was not clear how many transposable elements are reactivated in these cells, or how this process works.

Here, He et al. investigate this process in the sperm companion cells of a small weed known as Arabidopsis thaliana. The experiments showed that around 100 transposable elements were reactivated, most of them when an enzyme called DEMETER removed the methyl groups found in heterochromatin. However, this enzyme alone was not enough. It could only access the methyl groups if the tightly packed structure of the heterochromatin had relaxed following the removal of a histone protein called H1.

Taken together, these results indicate that histone H1 and DEMETER cooperate to regulate the activity of transposable elements in the genome. In addition, H1 is known to prevent the addition of methyl groups onto DNA; that it also impedes their removal suggests that this protein plays a complex role in controlling the way genetic information is interpreted. The next step would now be to investigate the impact of the reactivation of transposable elements on the next generation of plants and during evolution.

https://doi.org/10.7554/eLife.42530.002

Introduction

Large proportions of most eukaryotic genomes are comprised of transposable elements (TEs), mobile genetic fragments that can jump from one location to another. For example, TEs comprise approximately 50% of the human genome (International Human Genome Sequencing Consortium, 2001; Venter et al., 2001), and more than 85% of the genomes in crops such as wheat and maize (Schnable et al., 2009; Wicker et al., 2018). Regarded as selfish and parasitic, activities of TEs compromise genome stability, disrupt functional genes, and are often associated with severe diseases including cancers in animals (Anwar et al., 2017). To safeguard genome integrity, eukaryotic hosts have evolved efficient epigenetic mechanisms, including DNA methylation, to suppress TEs (He et al., 2011; Law and Jacobsen, 2010). Curiously, recent studies point to episodes of TE activation that occur in specific cell types and/or particular developmental stages (Garcia-Perez et al., 2016; Martínez and Slotkin, 2012). These TE activation events provide unique opportunities to understand epigenetic silencing mechanisms, and the co-evolution between TEs and their hosts.

Developmental TE activation has been shown in mammalian embryos, germlines and brain cells. In pre-implantation embryos and the fetal germline, LINE-1 retrotransposons are highly expressed despite relatively low levels of transposition (Fadloun et al., 2013; Kano et al., 2009; Percharde et al., 2017; Richardson et al., 2017). Recently, LINE-1 RNA was shown to play a key regulatory role in promoting pre-implantation embryo development in mice (Percharde et al., 2018). LINE-1 elements have also been shown to transcribe and mobilize in neuronal precursor cells in mice and human (Coufal et al., 2009; Muotri et al., 2005). The underlying mechanism of such cell-specific TE activation is still unclear. Hypomethylation at LINE-1 promoters in neurons has been proposed to contribute (Coufal et al., 2009), and possibly the availability of transcription factors (Muotri et al., 2005; Richardson et al., 2014). The frequency of LINE-1 retrotransposition in mammalian brain is still under debate; however, it has been speculated that LINE-1 activities might serve to promote genetic diversity among cells of a highly complex organ like the brain (Garcia-Perez et al., 2016; Richardson et al., 2014; Singer et al., 2010).

One of the best demonstrated cases of developmental TE activation occurs in the male gametophyte of flowering plants, pollen grains. Pollen are products of male gametogenesis, which initiates from haploid meiotic products called microspores. Each microspore undergoes an asymmetric mitosis to generate a bicellular pollen comprised of a large vegetative cell (VC) and a small generative cell engulfed by the VC (Berger and Twell, 2011). Subsequently the generative cell divides again mitotically to produce two sperm. Upon pollination, the VC develops into a pollen tube to deliver the sperm to meet the female cells, and subsequently degenerates. In the mature tricellular pollen of Arabidopsis thaliana, several TEs were found activated and transposed (Slotkin et al., 2009). Enhancer/gene trap insertions into TEs showed specific reporter activity in the VC, and TE transpositions detected in pollen were absent in progeny (Slotkin et al., 2009). These results demonstrated TE activation in pollen is confined to the VC. TE expression in the short-lived VC has been proposed to promote the production of small RNAs, which may be transported into sperm to reinforce the silencing of cognate TEs (Calarco et al., 2012; Ibarra et al., 2012; Martínez et al., 2016; Slotkin et al., 2009). However, TE transcription in the VC has not been comprehensively investigated, hindering our understanding of this phenomenon.

The mechanisms underlying TE reactivation in the VC are also unknown. One proposed mechanism is the absence of the Snf2 family nucleosome remodeler DDM1 (Slotkin et al., 2009). DDM1 functions to overcome the impediment of nucleosomes and linker histone H1 to DNA methyltransferases (Lyons and Zilberman, 2017; Zemach et al., 2013). Loss of DDM1 leads to DNA hypomethylation and massive TE derepression in somatic tissues (Jeddeloh et al., 1999; Lippman et al., 2004; Tsukahara et al., 2009; Zemach et al., 2013). However, global DNA methylation in the VC is comparable to that of microspores and substantially higher than in somatic tissues (Calarco et al., 2012; Hsieh et al., 2016; Ibarra et al., 2012). This suggests that DDM1 is present during the first pollen mitosis that produces the VC, so its later absence is unlikely to cause TE activation.

A plausible mechanism underlying TE activation in the VC is active DNA demethylation. DNA methylation in plants occurs on cytosines in three sequence contexts: CG, CHG and CHH (H = A, C or T). Approximately ten thousand loci – predominantly TEs – are hypomethylated in the VC, primarily in the CG context and to a lesser extent in the CHG/H contexts (Calarco et al., 2012; Ibarra et al., 2012). Hypomethylation in the VC is caused by a DNA glycosylase called DEMETER (DME) (Ibarra et al., 2012). DME demethylates DNA via direct excision of methylated cytosine, and its expression is confined to the VC and its female counterpart, the central cell, during sexual reproduction (Choi et al., 2002; Schoft et al., 2011). DME demethylation may therefore cause TE transcription in the VC; however, this hypothesis has not been tested.

Another plausible mechanism for epigenetic TE activation is chromatin decondensation (Feng et al., 2013). Drastic reprogramming of histone variants and histone modifications occurs during both male and female gametogenesis, rendering the gametes and companion cells with radically different chromatin states (Baroux et al., 2011; Borg and Berger, 2015). For example, centromeric repeats, which are condensed in sperm and other cell types, are decondensed in the VC, accompanied by the depletion of centromeric histone H3 (Ingouff et al., 2010; Mérai et al., 2014; Schoft et al., 2009). Chromocenters, which are comprised of condensed pericentromeric heterochromatin and rDNA repeats (Chandrasekhara et al., 2016; Fransz et al., 2002; Tessadori et al., 2004), are observed in sperm nuclei but absent in the VC nucleus, suggesting that pericentromeric heterochromatin is decondensed in the VC (Baroux et al., 2011; Ingouff et al., 2010; Schoft et al., 2009). Heterochromatin decondensation in the VC is proposed to promote rDNA transcription that empowers pollen tube growth (Mérai et al., 2014). However, the cause of VC heterochromatin decondensation remains unclear.

Our previous work showed that histone H1, which binds to the nucleosome surface and the linker DNA between two adjacent nucleosomes (Fyodorov et al., 2018), is depleted in Arabidopsis VC nuclei (Hsieh et al., 2016). H1 depletion in the VC has also been observed in a distantly related lily species (Tanaka et al., 1998), suggesting a conserved phenomenon in flowering plants. In Drosophila and mouse embryonic stem cells, H1 has been shown to contribute to heterochromatin condensation (Cao et al., 2013; Lu et al., 2009). H1 is also more abundant in heterochromatin than euchromatin in Arabidopsis (Ascenzi and Gantt, 1999; Rutowicz et al., 2015). However, it is unknown whether H1 participates in heterochromatin condensation in plant cells, and specifically whether the lack of H1 contributes to heterochromatin decondensation in the VC.

Whether and how the depletion of H1 in the VC contributes to TE derepression is also unclear. A recent study pointed to an intriguing link between H1 and DME. In the central cell, the histone chaperone FACT (facilitates chromatin transactions) is required for DME-directed DNA demethylation in heterochromatic TEs, and this requirement is dependent on H1 (Frost et al., 2018). However, DME activity in the VC is independent of FACT (Frost et al., 2018). One attractive hypothesis is that the lack of H1 in the VC causes heterochromatin decondensation and thereby contributes to the independence of DME from FACT. H1 depletion may therefore participate in VC TE activation by promoting DME-directed demethylation. Additionally, H1 depletion may activate TE transcription independently of DNA methylation, as shown in Drosophila where DNA methylation is absent (Iwasaki et al., 2016; Lu et al., 2013; Vujatovic et al., 2012; Zemach et al., 2010; Zhang et al., 2015).

In this study, we identify heterochromatic TEs that are epigenetically activated in Arabidopsis VCs. We demonstrate that these TEs are typically subject to DME-directed demethylation at the transcriptional start site (TSS), which is at least partially permitted by the depletion of H1. However, we find that loss of H1 activates some TEs without altering DNA methylation. We also show that developmental depletion of H1 decondenses heterochromatin in late microspores and is important for pollen fertility. Our results demonstrate that H1 condenses heterochromatin in plants and maintains genome stability by silencing TEs via methylation-dependent and -independent mechanisms.

Results

Heterochromatic transposons are preferentially expressed in the vegetative cell

To measure the extent of TE activation in the VC, we performed RNA-seq using mature pollen grains, followed by the annotation of gene and TE transcripts using Mikado and the TAIR10 annotation (Venturini et al., 2018). We identified 114 TEs that are transcribed at significantly higher levels in pollen than rosette leaves (fold change >2; p<0.05, likelihood ratio test), and hence likely to be specifically activated in the VC (Figure 1—source data 1) (Slotkin et al., 2009).

The VC-activated TEs are primarily located in pericentromeric regions and exhibit features of heterochromatic TEs, such as being long and GC rich (Frost et al., 2018) (Figure 1A,B, Figure 1—figure supplement 1A). As is typical of heterochromatic TEs (Zemach et al., 2013), VC-activated TEs are significantly enriched in dimethylation of histone H3 on lysine 9 (H3K9me2) in somatic tissues, and are significantly depleted of euchromatin-associated modifications (Figure 1B, Figure 1—figure supplement 1B). VC-activated TEs encompass diverse TE families, among which MuDR DNA transposons and Gypsy LTR-retrotransposons are significantly overrepresented (p<10−9 and 0.01, respectively, Fisher’s exact test; Figure 1C).

Figure 1 with 1 supplement see all
VC-activated TEs are heterochromatic and demethylated by DME.

(A) Expression and locations of VC-activated TEs along Chromosome 5. The purple ellipse and red bars indicate the centromere and borders of pericentromeric regions, respectively. (B) Box plots showing the length, GC content, and H3K9me2 level of TEs. Each box encloses the middle 50% of the distribution, with the horizontal line marking the median and vertical lines marking the minimum and maximum values that fall within 1.5 times the height of the box. Difference between the two datasets compared for each feature is significant (Kolmogorov-Smirnov test p<0.001). (C) Percentages of TEs classified by superfamily. (D) VC-activated TEs were aligned at the TSS and TTS (dashed lines), respectively, and average CG methylation levels for each 100 bp interval were plotted (referred to as ends analysis). (E) Snapshots demonstrating the expression (Log2RPKM), absolute and differential CG methylation at two example VC-activated TEs. Red lines under TE annotations indicate VC DME targets. Spm, sperm. (F) Pie charts illustrating percentages of VC-activated TEs with TSS overlapping (top) or within 500 bp (bottom) of VC DME targets. (G) Box plot showing the expression level of VC-activated TEs in pollen from WT and dme/+ heterozygous mutant. WT/2 shows half of the WT expression level. Wilcoxon matched-pairs signed-rank test is used. n.s., no significance.

https://doi.org/10.7554/eLife.42530.003

Transposon derepression in the VC is caused by DME-directed DNA demethylation

To assess whether TE activation in the VC is caused by DME-mediated DNA demethylation, we examined DNA methylation in VC and sperm at the 114 activated TEs. We found that these TEs have substantially lower CG methylation in the VC than in sperm at and near the TSS (Figure 1D,E), indicative of DME activity. Because TEs tend to be flanked by repeats (Joly-Lopez and Bureau, 2018), the transcriptional termination site (TTS) regions of activated TEs also tend to be hypomethylated in the VC (Figure 1D,E, Figure 1—figure supplement 1C). Examination of DNA methylation in VCs from dme/+ heterozygous plants (dme homozygous mutants are embryonic lethal), which produce a 50:50 ratio of dme mutant and WT pollen, revealed an intermediate level of methylation at TSS and TTS of VC-activated TEs (Figure 1D,E). CHG and CHH methylation is also substantially increased at the TSS (and TTS) of VC-activated TEs in dme/+ VC (Figure 1—figure supplement 1C), consistent with the knowledge that DME demethylates all sequence contexts (Gehring et al., 2006; Ibarra et al., 2012).

Consistent with the above results, 71 of the 114 (62%) VC-activated TEs overlap VC DME targets at their TSSs (Figure 1F, Figure 1—source datas 1 and 2). 92 out of the 114 TEs (81%) have VC DME targets within 500 bp of the TSS (Figure 1F, Figure 1—figure supplement 1D, Figure 1—source data 1). As DNA methylation at/near the TSS has been well-demonstrated to suppress the transcription of genes and TEs in plants and animals (Barau et al., 2016; Eichten et al., 2012; Hollister and Gaut, 2009; Manakov et al., 2015; Meng et al., 2016), our results indicate that DME-directed demethylation is a major mechanism of TE activation in the VC. Consistently, RNA-seq of pollen from dme/+ heterozygous plants showed significantly reduced expression at the 114 VC-activated TEs (Figure 1G, Figure 1—source data 1). As dme/+ heterozygous plants produce half dme mutant and half WT pollen, we expect the transcription of VC-activated TEs to be reduced to roughly half in dme/+ pollen. The 114 VC-activated TEs are transcribed at levels close to this expectation (Figure 1G), with the median ratio of their transcription in dme/+ versus WT pollen being 0.43 (Figure 1—source data 1).

Vegetative-cell-expressed H1 impedes DME from accessing heterochromatic transposons

We next tested our hypothesis that the lack of histone H1 in the VC (Hsieh et al., 2016) allows heterochromatin to be accessible by DME. We first examined the developmental timing of H1 depletion during microspore and pollen development using GFP translational fusion lines (Hsieh et al., 2016; She et al., 2013). There are three H1 homologs in Arabidopsis, with H1.1 and H1.2 encoding the canonical H1 proteins, and H1.3 expressed at a much lower level and induced by stress (Rutowicz et al., 2015). H1.1- and H1.2- GFP reporters exhibit the same expression pattern: present in early microspore nucleus but absent in the late microspore stage, and remaining absent in the VC nucleus while present in the generative cell and subsequent sperm nuclei (Figure 2A). H1.3 is not detectable in either microspore or pollen (Figure 2A). These results are consistent with our previous observations, confirming that H1 is absent in the VC (Hsieh et al., 2016), and demonstrating that H1 depletion begins at the late microspore stage.

Figure 2 with 2 supplements see all
Ectopic H1 expression in the vegetative cell impedes DME at the most heterochromatic loci.

(A–B) Confocal images showing H1 localization under native promoter (A) and VC-specific promoter (pVC, B) during male gametogenesis. Msp, microspore; BCP, bicellular pollen; TCP, tricellular pollen. Bars, 5 μm. All pVC::H1.1-mRFP (short as pVC::H1) refers to line #2. (C) Ends analysis of all TEs or genes in VCs from pVC::H1 (line #2) and WT. (D–E) Kernel density plots illustrating frequency distribution of methylation differences in 50 bp windows between VCs from pVC::H1 and WT (D), and between WT sperm (Spm) and VC (E). (F) Snapshots showing CG methylation difference between the indicated cell types. Arrows point to DME targets that are hypermethylated by pVC::H1. (G) Snapshots demonstrating CG methylation in sperm and VCs at single-nucleotide resolution, with the cytosine most hypomethylated by DME marked in red. VC DME targets are underlined in black. (H) Scatter plot illustrating CG methylation differences between the indicated cell types at H1 hyperDMRs. 82.25% of H1 hyperDMRs show significant increase in sperm in comparison to VCs. (I) Box plot illustrating H3K9me2 level at VC DME targets that are significantly hypermethylated in pVC::H1 (H1-inhibited) or not (H1-independent), respectively. Difference between the two groups is significant (Kolmogorov-Smirnov test p<0.001). (J) VC DME targets were grouped according to H3K9me2 levels, aligned at the most demethylated cytosine (dashed lines), and plotted for average CG methylation difference as indicated in each 10 bp interval (left). Similarly, CG methylation in pVC::H1 and WT VCs was plotted for the group with the lowest and highest H3K9me2, respectively. Spm, sperm.

https://doi.org/10.7554/eLife.42530.007

To understand how H1 affects DME activity, we ectopically expressed H1 in the VC. To ensure H1 incorporation into VC chromatin, we used the pLAT52 promoter, which is expressed from the late microspore stage immediately prior to Pollen Mitosis 1, and is progressively upregulated in VC during later stages of pollen development (Eady et al., 1994; Grant-Downton et al., 2013). Using pLAT52 to drive the expression of H1.1 tagged with mRFP (simplified as pVC::H1), we observed continuous H1-mRFP signal in the VC at the bicellular and tricellular pollen stages, while the signal was undetectable in the generative cell and sperm (Figure 2B). H1-mRFP signal was also undetectable in late microspores (Figure 2B), probably due to the low activity of pLAT52 at this stage (Eady et al., 1994). Notably, we found H1 expression in VC leads to shortened siliques, a substantial proportion of malformed pollen, and reduced pollen germination rate (Figure 2—figure supplement 1A–D), suggesting the depletion of H1 in the VC is important for pollen fertility.

To evaluate the effect of VC-expressed H1 on DNA methylation, we obtained genome-wide methylation profiles for VC nuclei from a strong pVC::H1 line (#2; Figure 2B) and WT via fluorescence-activated cell sorting (FACS) followed by bisulfite sequencing (Supplementary file 1). CG methylation in the VC of pVC::H1 plants is largely similar to that of WT, except for a slight increase in TE methylation (Figure 2C, Figure 2—figure supplement 2A). Consistently, the frequency distribution of CG methylation differences between VCs of pVC::H1 and WT at loci that are not DME targets peaks near zero, showing almost no global difference (Figure 2D). However, a substantial proportion of loci that are targeted by DME show hypermethylation in pVC::H1 VC (Figure 2D, Figure 2—figure supplement 2B). DME targets also show preferential hypermethylation in CHG and CHH contexts in the VC of pVC::H1 (Figure 2—figure supplement 2C–D). These results indicate that H1 expression in the VC specifically impedes DME activity.

Across the genome, we found 2964 differentially methylated regions (DMRs) that are significantly CG hypermethylated in the VC of pVC::H1 plants (referred to as H1 hyperDMRs hereafter; ranging from 101 to 2155 nt in length, 280 nt on average; Figure 2—source data 1). Most of the H1 hyperDMRs (1618, 55%) overlap DME targets in the VC (Figure 2—source data 1), and H1 hyperDMRs exhibit strong hypomethylation in WT VCs, with 81.4% (2412 sites) having significantly more CG methylation in sperm than VC (p<0.001, Fisher’s exact test), indicating that most H1 hyperDMRs are DME targets (Figure 2E–H).

Our results demonstrate that H1 hyperDMRs are primarily caused by the inhibition of DME. However, only 3066 out of 11896 (26%) VC DME targets have significantly more CG methylation in the VC of pVC::H1 than WT (p<0.001, Fisher’s exact test; Figure 1—source data 2), indicating that VC-expressed H1 impedes DME at a minority of its genomic targets. These H1-impeded DME targets are heterochromatic and significantly enriched in H3K9me2 compared with H1-independent DME targets (Figure 2I). To further examine the link with heterochromatin, we aligned all VC DME target loci at the most hypomethylated cytosine, and separated them into five groups by H3K9me2 levels (Figure 2J). pVC::H1-induced hypermethylation peaks where DME-mediated hypomethylation peaks, but is apparent only in the most heterochromatic group (highest H3K9me2) of DME target loci (Figure 2J). Taken together, our results demonstrate that developmental removal of H1 from the VC allows DME to access heterochromatin.

H1 represses transposons via methylation-dependent and independent mechanisms

Given the importance of H1 removal for DME-directed DNA demethylation and the well-demonstrated role of DME demethylation in regulating gene expression (Choi et al., 2002; Ibarra et al., 2012; Schoft et al., 2011), we investigated the contribution of H1 to gene regulation in pollen. RNA-seq was performed using pollen from the pVC::H1 line (#2), which showed strong H1 expression in VC (Figures 2B and 3A). Only a small fraction of pollen-expressed genes (3%; 89 out of 2845) is differentially expressed (fold change >2; p<0.05, likelihood ratio test) between pVC::H1 and WT (Figure 3B, Figure 3—source data 1). Among these 89 genes, 58 (65%) are suppressed by H1 expression in the VC, and 31 (35%) are activated (Figure 3B, Figure 3—source data 1). 85 out of these 89 genes (96%) do not overlap DME targets within 500 bp of the TSS, hence the effect of H1 on their expression is probably not mediated by DME. Among the four genes that overlap DME targets within 500 bp of the TSS, two genes gain a small amount of methylation at the overlapping DME targets and are suppressed in pVC::H1 pollen (Figure 3—figure supplement 1, Figure 3—source data 1), hence are possibly suppressed by H1 via the inhibition of DME.

Figure 3 with 1 supplement see all
Vegetative-cell-expressed H1 represses heterochromatic TEs in a dosage-dependent manner.

pVC::H1 refers to line #2 except as specified in A) and F). (A,F) quantitative RT-PCR demonstrating H1.1 (A) or TE (F) expression in pollen from WT and two independent pVC::H1 transgenic lines. Relative expression is calculated by normalizing to WT (WT = 1). Student’s t test *p<0.05, **p<0.01, ***p<0.001; ns, not significant; n = 3; mean ± SD are shown. (B,D) Scatter plot illustrating the expression (Log2TPM) of TEs or genes in WT and pVC::H1 pollen. Red and green dots indicate significant down- and up-regulation in pVC::H1 compared to WT (|fold change| > 2, marked by dashed lines; likelihood ratio test p<0.05), respectively. (C) Box plots illustrating H3K9me2 enrichment and H1 occupancy at endosperm DME targets which are within 500 bp of the TSS of imprinted genes, or VC DME targets inhibited by pVC::H1 (refer to Figure 2I). Difference between the two datasets compared for each feature is significant (Kolmogorov-Smirnov test p<0.001). (E) Snapshots showing the expression (Log2RPKM) of 3 example H1-repressed TEs in WT and pVC::H1 pollen. Rep, biological replicate. (G) Chromosomal view of H1-repressed TEs, similar to Figure 1A. (H) Box plots illustrating the length, H3K9me2 enrichment, and H1 occupancy at two groups of VC-activated TEs. Difference between the two datasets compared for each feature is significant (Kolmogorov-Smirnov test p<0.05 for length, and <0.001 for others). (I) Percentages of TEs classified by superfamily.

https://doi.org/10.7554/eLife.42530.011

Because DME regulates the expression of imprinted genes in the endosperm (Choi et al., 2002; Hsieh et al., 2011; Ibarra et al., 2012), we specifically examined the imprinted genes to further investigate the role of VC H1 removal in regulating pollen gene transcription. Out of the 640 known imprinted genes (Gehring et al., 2011; Hsieh et al., 2011; Pignatta et al., 2014; Wolff et al., 2011), 85 overlap endosperm DME targets within 500 bp of their TSS (Figure 3—source data 2). Because DME targets different loci in the VC and CC (Ibarra et al., 2012), we subsequently examined if these 85 putative DME-regulated imprinted genes are also subject to DME demethylation in the VC. 63 of these genes also overlap VC DME targets within 500 bp of their TSS, none of which is differentially expressed in pVC::H1 pollen compared with WT pollen (Figure 3—source data 2), showing that transcription of these genes in pollen is unlikely regulated by H1. This is not surprising, because only 8 out of these 63 genes are expressed in WT pollen (Figure 3—source data 2), and H1 preferentially regulates heterochromatic DME targets (Figures 2I,J and 3G), whereas DME targets involved in gene regulation are typically euchromatic sites next to genes (Figure 3C). Consistently, DME targets likely regulating imprinted genes are associated with significantly less H1 in somatic tissues than the VC DME targets that are dependent on H1 (Figure 3C).

In contrast to the small effect of H1 on gene transcription, a substantial proportion of VC-activated TEs (41%; 47 out of 114) show significant differential expression (fold change >2; p<0.05, likelihood ratio test) due to H1 expression in VC (Figure 3D). Among these differentially expressed TEs, the overwhelming majority (46; 98%) are repressed (Figure 3D,E, Figure 1—source data 1). These data indicate that ectopic expression of H1 in the VC preferentially represses TEs. Quantitative RT-PCR validated our RNA-seq results and confirmed the strong suppression of TEs in pVC::H1 (Figure 3F). Taking advantage of a pVC::H1 line #7 with weaker H1 expression in pollen (Figure 3A), we found H1 represses TE expression in a dosage-dependent manner: TEs are suppressed to a lesser extent in line #7 compared to the strong line #2 (Figure 3F). H1-repressed TEs in the VC are predominantly localized to pericentromeric regions and are overrepresented for LTR retrotransposons, including Gypsy and Copia elements (Figure 3G–I). Compared to other VC-activated TEs, the H1-repressed TEs are significantly longer and enriched for H3K9me2 and H1 in somatic tissues (Figure 3H), consistent with the observation that H1 precludes DME access to heterochromatin.

In support of the hypothesis that H1 represses VC TE expression by blocking DME, 20 of 46 H1-repressed TEs show significant increase of DNA methylation in at least one sequence context within 500 bp of the TSS in pVC::H1 (p<0.001, Fisher’s exact test; Figure 4A,B). Four more TEs overlap a DME target, which is hypermethylated in pVC::H1, within 500 bp of the TSS, and hence may also be suppressed by DME inhibition. However, 21 out of the rest of 22 TEs do not overlap any H1 hyperDMRs within 500 bp of the TSS (Figure 4A, marked by asterisks in the lower panel), indicating that their suppression by H1 is not mediated by DNA methylation. Of these 21 TEs, 15 TEs overlap DME targets within 500 bp of TSS. DME maintains access to these TEs in the presence of H1, suggesting their VC demethylation does not rely on the depletion of H1 and their repression in pVC::H1 is DME-independent as exemplified by AT3TE60310 (Figure 4C). Our results demonstrate that H1 overexpression in the VC represses heterochromatic TEs via both DNA methylation-dependent and independent mechanisms.

H1 suppresses TEs in the vegetative cell via two mechanisms.

(A) Heat map demonstrating DNA methylation differences between pVC::H1 and WT VCs within 500 bp of the TSS of H1-repressed TEs. Asterisks mark TEs whose suppression is not caused by changes in DNA methylation. Data are sorted in descending order based on CG methylation difference for upper and lower panels, respectively. (B,C) Snapshots showing the expression and DNA methylation of representative TEs suppressed by pVC::H1 via methylation-dependent (B) and -independent (C) mechanisms, respectively. Spm, sperm.

https://doi.org/10.7554/eLife.42530.015

Depletion of H1 decondenses heterochromatin during male gametogenesis

H1 depletion and TE activation in the VC are accompanied by loss of cytologically detectable heterochromatin (Baroux et al., 2011; Ingouff et al., 2010; Schoft et al., 2009). We therefore tested whether H1 contributes to heterochromatin condensation in plant cells. Immunostaining of leaf nuclei showed that H1 co-localizes with H3K9me2 in highly-compacted heterochromatic foci, known as chromocenters (Figure 5A). Furthermore, we found that chromocenters become dispersed in the nuclei of h1 mutant rosette leaves (Figure 5B). These observations demonstrate that H1 is required for heterochromatin condensation in plants.

Depletion of H1 decondenses heterochromatin in leaves and late microspores.

(A) Immunostaining with GFP and H3K9me2 antibodies showing the co-localization of H1 and H3K9me2-enriched chromocenters. (B) Examples of leaf nuclei with condensed, intermediate or decondensed chromocenters, and their percentages in WT or the h1 mutant. (C) Gradual decondensation of heterochromatin during male gametogenesis in Arabidopsis. Micrographs of Hoechst-stained microspores (Msp) and pollen (BCP, bicellular pollen; TCP, tricellular pollen) demonstrate a gradual dispersion of chromocenters in late microspores and subsequently the vegetative nucleus (VN) in pollen. Chromocenters are not detected in the VN of BCP and TCP. GN, generative nucleus. (D) Percentages of late microspores with no, weak or strong chromocenters (CCs; examples on the left) in WT and pMSP1::H1.1-mRFP, in which H1 co-localizes with the strong CCs. (E) H1 is induced only in late microspores in pMSP1::H1.1-mRFP, and co-localizes with strong CCs. All bars, 5 μm.

https://doi.org/10.7554/eLife.42530.016

We then examined whether ectopic H1 expression can condense the heterochromatin in VC nuclei. Consistent with previous observations (Baroux et al., 2011; Ingouff et al., 2010; Schoft et al., 2009), no condensed chromocenters were detected in WT VC (Figure 5C). pVC::H1 VC also showed no obvious chromocenters (n > 500; Figure 2B). This suggests either that H1 expression is not strong enough in pVC::H1, or other factors are involved in heterochromatin decondensation in the VC.

Heterochromatin decondensation during male gametogenesis seems to be gradual: chromocenters are observed at early microspore stage, but become dispersed in late microspore stage, when H1 is depleted (Figures 2A and 5C). We observed strong and weak chromocenters, respectively, in 27% and 59% of late microspore nuclei, whereas no chromocenters were observed in the VC at either bicellular or tricellular pollen stage (Figure 5C,D). The further decondensation of VC heterochromatin after H1 depletion during the late microspore stage suggests the involvement of other factors in the VC. To test whether H1 is sufficient to induce chromatin condensation in microspores, we used the late-microspore-specific MSP1 promoter (Honys et al., 2006) to drive H1 expression (pMSP1::H1.1-mRFP, short as pMSP1::H1). In pMSP1::H1, we observed strong chromocenters in the majority (68%) of late microspores (Figure 5D). H1 expression in pMSP1::H1 is specific to late microspores, and co-localizes with induced chromocenters (Figure 5E). These results show that H1 is sufficient to promote heterochromatic foci in late microspores, thus demonstrating the causal relationship between H1 depletion and the decondensation of heterochromatin.

Discussion

Epigenetic reactivation of TEs in the VC of flowering plants is an intriguing phenomenon, which is important not only for understanding sexual reproduction, but also for elucidating epigenetic silencing mechanisms. Here we show that Arabidopsis VC-activated TEs are heterochromatic, and mostly subject to DME-directed demethylation at their TSS (Figure 1F). Given the well-demonstrated role of DNA methylation at the TSS for transcriptional suppression (Barau et al., 2016; Eichten et al., 2012; Hollister and Gaut, 2009; Manakov et al., 2015; Meng et al., 2016), our data demonstrate that DME-mediated demethylation in the VC is the primary cause of TE activation. As DNA demethylation of TEs during reproduction also occurs in rice and maize (Park et al., 2016; Rodrigues et al., 2013; Zhang et al., 2014), species that diverged from Arabidopsis more than 150 million years ago (Chaw et al., 2004), our results suggest that TE activation in the VC is prevalent among flowering plants.

DME demethylates about ten thousand loci in the VC and central cell, respectively, however, only half of these loci overlap (Ibarra et al., 2012). It was unclear why DME targets differ in these cell types, but differences in chromatin configuration have been postulated to contribute (Feng et al., 2013). Our finding that the access of DME to heterochromatic TEs in the VC is permitted by the lack of H1 supports this idea. H1 is presumably present in the central cell (Frost et al., 2018) but is absent in the VC (Hsieh et al., 2016), thus rendering heterochromatic TEs more accessible in the VC. Differential distribution of other factors in the VC and central cell, such as histone variant H3.1 (Borg and Berger, 2015; Ingouff et al., 2010), may also affect DME targeting. Consistently, FACT is required for DME activity in the central cell at many loci even in the absence of H1, whereas DME is entirely independent of FACT in the VC (Frost et al., 2018), suggesting the presence of impeding factor(s) other than H1 in the central cell. With distinct chromatin architectures, the vegetative and central cells are excellent systems for understanding how chromatin regulates DNA demethylation.

Our finding that histone H1 affects DME activity adds to the emerging picture of H1 as an important and complex regulator of eukaryotic DNA methylation. H1 depletion causes local hypomethylation in mouse cells (Fan et al., 2005) and extensive hypermethylation in the fungi Ascobolus immersus (Barra et al., 2000) and Neurospora crassa (Seymour et al., 2016). In Arabidopsis, loss of H1 causes global heterochromatic hypermethylation in all sequence contexts by allowing greater access of DNA methyltransferases (Lyons and Zilberman, 2017; Zemach et al., 2013). Our results suggest that H1 may also influence DME-homologous demethylases that control methylation in somatic tissues (He et al., 2011). By regulating both methylation and demethylation, H1 may serve as an integrator of methylation pathways that tunes methylation up or down depending on the locus.

Our data also indicate that the regulatory functions of H1 extend beyond DNA methylation in plants. Activated TEs in the VC can be categorized into four groups, based on the mechanism of their activation (Figure 6). TEs in Group I are the least heterochromatic and their activation is dependent on DME but not H1 (Figures 3H and 6). Group II comprises TEs in which H1 absence is required for DME demethylation and activation (Figure 6). For TEs in Group III, H1 depletion and DME demethylation are both required for activation, but DME activity is not affected by H1 (Figure 6). Group IV TEs are activated by H1 depletion and are not targeted by DME (Figure 6). Groups III and IV demonstrate that H1 can silence TEs independently of DNA methylation. Group III also demonstrates that DNA methylation and H1 cooperate to suppress TE expression in plants. Thus, H1 regulates TEs via DNA methylation-dependent and -independent mechanisms.

Model depicting four mechanisms underlying TE activation in the VC.

The number of TEs in each group is shown on the top. Significantly less heterochromatic than TEs in other groups (Figure 3H), Group I TEs are activated by DME-directed DNA demethylation. Group II TEs rely on H1 depletion to allow DME demethylation and activation. Group III TEs are demethylated by DME but require H1 depletion to allow transcription (ie. pVC::H1 represses these TEs without affecting DME). Group IV TEs are not demethylated by DME; their activation is solely dependent on the depletion of H1. TEs belong to each group are listed in Figure 1—source data 1. Red lollipops denote DNA methylation.

https://doi.org/10.7554/eLife.42530.017

During the ongoing arms race between TEs and their hosts, it may be difficult to determine whether TE expression represents temporary TE triumphs or is domesticated by the host to serve a function. TE activation in the VC – a cell that engulfs the male plant gametes – has been proposed as a defense strategy, which generates small RNAs that enhance TE silencing in sperm (Calarco et al., 2012; Ibarra et al., 2012; Martínez et al., 2016; Slotkin et al., 2009). However, TEs can also use companion cells as staging grounds for invasion of the gametes (Wang et al., 2018). Our demonstration that programmed DME demethylation, which is facilitated by developmental heterochromatin decondensation, is the predominant cause of VC TE activation is consistent with a defensive, host-beneficial model. Nonetheless, the alternative TE-driven model is also plausible. DME demethylates about ten thousand loci in the VC, most of which are small and euchromatic TEs (Ibarra et al., 2012). However, only a hundred heterochromatic TEs are activated, at least partially permitted by developmental H1 depletion. As small euchromatic TEs tend to be next to genes, DME demethylation regulates genes and is important for pollen fertility (Choi et al., 2002; Ibarra et al., 2012; Schoft et al., 2011). Our data show that developmental H1 depletion is also important for pollen fertility. Therefore, at least some TEs may be hijacking an essential epigenetic reprogramming process. TE activation in the VC may facilitate both host defense and transposition, with the balance specific to each TE family and changing over evolutionary time. The effects of VC TE activation on TE proliferation in the progeny may warrant investigation, particularly in out-crossing species with aggressive TEs and in natural populations.

Materials and methods

Key resources table
Reagent type
(species) or
resource
DesignationSource or referenceIdentifiersAdditional information
Gene
(Arabidopsis thaliana)
H1.1NAAT1G06760
Gene (A. thaliana)H1.2NAAT2G30620
Gene (A. thaliana)DEMETERNAAT5G04560
Genetic
reagent (A. thaliana)
h1.1–1Nottingham
Arabidopsis Stock
Centre
SALK_128430C
Genetic
reagent (A. thaliana)
h1.2–1GABI-KatGABI_406H11
Genetic
reagent (A. thaliana)
dme-7Nottingham
Arabidopsis Stock
Centre
SALK_107538
Genetic
reagent (A. thaliana)
pLAT52::H1.1-mRFPthis paperLAT52 promoter from
tomato is used
Genetic
reagent (A. thaliana)
pMSP1::H1.1-mRFPthis paperMSP1 (AT5G59040)
promoter
is used
Biological
sample (A. thaliana)
sperm nuclei
from Col-0
this paperbisulfite-sequencing
Biological
sample (A. thaliana)
vegetative nuclei
from Col-0
this paperbisulfite-sequencing
Biological
sample (A. thaliana)
vegetative nuclei
from pLAT52::H1.1-mRFP
this paperbisulfite-sequencing
Biological
sample (A. thaliana)
pollen from Col-0this paperRNA-sequencing
Biological
sample (A. thaliana)
pollen from
pLAT52::H1.1-mRFP
this paperRNA-sequencing
Biological
sample (A. thaliana)
pollen from dme-7/+this paperRNA-sequencing
AntibodyGFPAbcamCat# ab290,
RRID:AB_303395
(1:100)
AntibodyH3K9me2AbcamCat# ab1220,
RRID:AB_449854
(1:100)
AntibodyAlexa 488- secondariesThermoFisher
Scientific
Cat# A-11034,
RRID:AB_2576217
(1:200)
AntibodyAlexa 555- secondariesThermoFisher
Scientific
Cat# A28180,
RRID:AB_2536164
(1:200)
Software,
algorithm
MikadoVenturini et al., 2018RRID:SCR_016159
Software,
algorithm
KallistoBray et al., 2016RRID:SCR_016582
Software,
algorithm
SleuthPimentel et al., 2017RRID:SCR_016883
OtherHoechst 33342ThermoFisher
Scientific
Cat# H3570(1:1000)
OtherDAPIThermoFisher
Scientific
Cat# D1306,
RRID:AB_2629482
(100 ng/mL)

Plant materials and growth conditions

Request a detailed protocol

A. thaliana plants were grown under 16 hr light/8 hr dark in a growth chamber (20°C, 80% humidity). All plants used are of the Col-0 ecotype. pH1.1::H1.1-eGFP, pH1.2::H1.2-eGFP, dme-7, and the h1 (h1.1 h1.2 double) mutant lines were described previously (Schoft et al., 2011; She et al., 2013; Zemach et al., 2013). pLAT52::H1.1-mRFP and pMSP1::H1.1-mRFP were constructed with MultiSite Gateway System into the destination vector pK7m34GW (Invitrogen). The BP clones pDONR-P4-P1R-pLAT52 and pDONR-P2R-P3-mRFP were kindly provided by Prof. David Twell (Leicester University, UK) (Eady et al., 1994). MSP1 promoter was cloned into pDONR-P4-P1R as described previously (Honys et al., 2006). WT plants were transformed via floral dip (Clough and Bent, 1998), and T2 or T3 plants homozygous for the transgene were used in this study.

Pollen extraction, RNA sequencing and quantitative RT-PCR

Request a detailed protocol

Open flowers were collected for pollen isolation in Galbraith buffer (45 mM MgCl2, 30 mM sodium citrate, 20 mM MOPS, 1% Triton-X-100, pH 7.0) by vortexing at 2000 rpm for 3 min. The crude fraction was filtered through a 40 μm cell strainer to remove flower parts, and subsequently centrifuged at 2600 g for 5 min to obtain pollen grains. RNA was extracted from pollen grains with RNeasy Micro Kit (Qiagen) following manufacturer’s instructions. RNA-sequencing libraries were prepared using Ovation RNA-seq Systems 1–16 for Model Organisms (Nugen Technologies), and sequenced on the Hiseq 2500 (Illumina) instrument at the UC Berkeley Vincent J. Coates Genomics Sequencing Laboratory or on the Nextseq 500 (Illumina) at the John Innes Centre. Quantitative RT-PCR (qRT-PCR) was performed as described previously (Walker et al., 2018), and TUA2 was used as an internal control. Primers for qRT-PCR are listed in Supplementary file 2.

In vitro pollen germination

Request a detailed protocol

The experiment was performed as described previously with some modifications (Rodriguez-Enriquez et al., 2013). Pollen from three newly opened flowers for individual plants were brushed on cellulose membrane sitting on germination medium (18% sucrose, 0.01% boric acid, 1 mM CaCl2, 1 mM Ca(NO3)2, 1 mM KCl, 0.25 mM spermidine, pH 8.0 with KOH adjusted, 0.5% agarose) in small petri dishes. Petri dishes were placed in a box with a piece of wet tissue at the bottom to keep humidity. The boxes were incubated in Sanyo cabinet (MLR-351H) at 21°C for 4 hr. 300 pollen grains were counted from each cellulose membrane and eight replicates for each genotype. Pollen germination was counted by using ImageJ.

RNA-seq analysis

Request a detailed protocol

TE transcript annotation was created using RNA-seq data from four biological replicates of pollen. Tophat2, Hisat, and STAR were used to align RNA-seq reads to the TAIR10 genome, and transcripts were assembled using CLASS2, StringTie, and Cufflinks, respectively. Assembled transcripts were selected by Mikado using default options except that the BLAST and Transdecoder steps were disabled (Venturini et al., 2018). As a result, 21381 transcripts (called superloci; GSE120519) were identified.

To identify VC-activated TEs, we first refined the list of superloci by selecting those overlapping with TAIR10 TE annotation. Subsequently to eliminate TE-like genes from the refined list, superloci with CG methylation less than 0.7 in rosette leaves (Stroud et al., 2014; Stroud et al., 2013) were excluded. This gave rise to an annotation of pollen TE transcripts, which was combined with TAIR10 gene annotation for Kallisto analysis (Bray et al., 2016). RNA-seq data from WT and dme/+ pollen (this study) and rosette leaves (Walker et al., 2018), each including three biological replicates, were processed using Kallisto and Sleuth (Bray et al., 2016; Pimentel et al., 2017). TEs that are transcribed at least five times more in WT pollen than leaves (with p<0.05, likelihood ratio test) are considered as activated in the VC (refer to Figure 1—source data 1 for the list of VC-activated TEs). A total of 2845 genes were found to be expressed in pollen with TPM (transcripts per million) more than five in the Kallisto output (Figure 3—source data 1; data used in Figure 3B).

To identify TEs and genes that are suppressed by H1 in the VC, we analyzed RNA-seq data from WT and pLAT52::H1.1-mRFP line #2 (short as pVC::H1 unless specified otherwise) pollen using Kallisto and Sleuth as described above. Significant differential expression was defined with a fold change at least two and a p-value less than 0.05. H1-repressed TEs were listed in Figure 1—source data 1.

Whole-genome bisulfite sequencing and analysis

Request a detailed protocol

Vegetative and sperm nuclei were isolated via FACS as described previously (Ibarra et al., 2012). Bisulfite-sequencing libraries were prepared as previously described (Walker et al., 2018). Sequencing was performed on Hiseq 2500 (Illumina) at the UC Berkeley Vincent J. Coates Genomics Sequencing Laboratory, Hiseq 4000 (Illumina) at Novogene Ltd. and Harvard University, and Nextseq 500 (Illumina) at Cambridge University Biochemistry Department and the John Innes Centre. Sequenced reads (100, 75, or 50 nt single-end) were mapped to the TAIR10 reference genome, and cytosine methylation analysis was performed as previously described (Ibarra et al., 2012).

Identification of DME targets and H1 hyperDMRs in the VC

Request a detailed protocol

As all CG hypomethylation in the VC in comparison to sperm is caused by DME (Ibarra et al., 2012), we identified VC DME targets via detecting CG differentially methylated regions (DMRs) that are hypermethylated in sperm in comparison to the VC. DMRs were identified first by using MethPipe (settings: p=0.05 and bin = 100) (Song et al., 2013), and subsequently retained if the fractional CG methylation across the whole DMR was at least 0.2 higher in sperm than the VC. The refined DMRs were merged to generate larger DMRs if they occurred within 300 bp. Finally, merged DMRs were retained if they cover at least 100 bp, and the fractional CG methylation across the whole DMR was significantly (Fisher’s exact test p<0.01) and substantially (>0.2) higher in sperm than the VC. This resulted in the identification of 11896 VC DME targets (Figure 1—source data 2).

H1 hyperDMRs were identified using the same criteria, except comparing CG methylation in VCs from pVC::H1 and WT. In total, 2964 H1 hyperDMRs were identified (Figure 2—source data 1).

Box plots

Request a detailed protocol

Box plots compare the enrichment of genomic or chromatin features among TEs (Figures 1B and 3G, Figure 1—figure supplement 1B) or VC DME targets (Figure 2I) as described in corresponding figure legends. ChIP-seq data for H3K9me2 (Stroud et al., 2014), and ChIP-chip data for H1 (Rutowicz et al., 2015), H3K27me3 (Kim et al., 2012), and other histone modifications (Roudier et al., 2011) were used.

Density plots

Request a detailed protocol

All DNA methylation kernel density plots compare fractional methylation within 50 bp windows. We used windows with at least 20 informative sequenced cytosines and fractional methylation of at least 0.5 (Figure 2D, Figure 2—figure supplement 2) or 0.7 (Figure 2E) for CG context, and 0.4 and 0.1 for CHG and CHH context, respectively, in at least one of the samples being compared.

Meta analysis (ends analysis)

Request a detailed protocol

Ends analysis for TEs and genes was performed as described previously (Ibarra et al., 2012). Similarly, ends analysis of TE transcripts was performed using the annotation of VC-activated TEs described above (Figure 1—source data 1). DNA methylation data from Ibarra et al. (2012) was used.

In Figure 2J, DME sites were aligned at the most demethylated cytosine, and average CG methylation levels for each 10 bp interval at both sides were plotted. To identify individual hypomethylation sites created by DME, we first obtained the 50 bp windows with a CG methylation difference larger than 0.5 between sperm and VC (sperm – VC >0.5 and Fisher’s exact test p<0.001). Windows were then merged if they occurred within 200 bp. Merged windows were retained for further analysis if the fractional CG methylation across the whole site was 0.2 greater in sperm than VC (sperm – VC >0.2 and Fisher’s exact test p<0.0001). This resulted in 13610 DME sites, which were separated into five groups according to H3K9me2 level (Stroud et al., 2014):<2.5, 2.5–4.3, 4.3–6.5, 6.5–10.5, and >10.5 (Figure 2J). The most demethylated cytosine within each site was identified if it had the greatest differential methylation in sperm than VC among cytosines in the CG context (sperm – VC >0.2, and Fisher’s exact test p<0.001) and was sequenced at least 10 times.

DNA methylation analysis of H1-repressed TEs

Request a detailed protocol

Differential methylation at a 1000 bp region centered upon the TSS of H1-repressed TEs was calculated between VCs of pVC::H1 and WT (Figure 4A). TEs whose differential methylation is significant (Fisher’s exact test p<0.001) and larger than 0.2 (in CG context), 0.1 (in CHG context), or 0.05 (in CHH context) are illustrated in the upper panel in Figure 4A.

Confocal and scanning electron microscopy

Request a detailed protocol

Microspores and pollen were isolated as described previously (Borges et al., 2012), stained with Hoechst or DAPI, and examined under a Leica SP8 confocal microscope. Scanning electron microscopy was performed on a Zeiss Supra 55 VP FEG.

Immunofluorescence

Request a detailed protocol

Immunofluorescence was performed as described previously with small modifications (Yelagandula et al., 2014). Rosette leaves from 3-week-old plants were fixed in TRIS buffer with 4% paraformaldehyde (10 mM Tris-HCl pH 7.5, 10 mM EDTA, 100 mM NaCl) for 20 min. After being washed with TRIS buffer twice, the fixed leaves were chopped with razor blades in 1 mL of lysis buffer (15 mM Tris pH 7.5, 2 mM EDTA, 0.5 mM spermine, 80 mM KCl, 20 mM NaCl, 0.1% Triton X-100) and filtered through a 35 μm cell strainer. Nuclei were pelleted via centrifugation at 500 g for 3 min and resuspended in 100 μL of lysis buffer. Next, 10 μL was spotted onto coverslips, air-dried, and post-fixed in PBS with 4% paraformaldehyde for 30 min. After being washed with PBS twice, coverslips were incubated in blocking buffer (PBS with 1% BSA) at 37°C for 30 min and then incubated in blocking buffer with primary antibodies at 4°C overnight (Mouse anti-H3K9me2 Abcam ab1220, 1:100; Rabbit anti-GFP Abcam ab290, 1:100). After being washed with PBS three times, coverslips were incubated in PBS with secondary antibodies at 37°C for 30 min, and then washed with PBS three times again before being counterstained and mounted in Vectashield mounting media with DAPI (Vector H-1200).

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64
    The B73 maize genome: complexity, diversity, and dynamics
    1. PS Schnable
    2. D Ware
    3. RS Fulton
    4. JC Stein
    5. F Wei
    6. S Pasternak
    7. C Liang
    8. J Zhang
    9. L Fulton
    10. TA Graves
    11. P Minx
    12. AD Reily
    13. L Courtney
    14. SS Kruchowski
    15. C Tomlinson
    16. C Strong
    17. K Delehaunty
    18. C Fronick
    19. B Courtney
    20. SM Rock
    21. E Belter
    22. F Du
    23. K Kim
    24. RM Abbott
    25. M Cotton
    26. A Levy
    27. P Marchetto
    28. K Ochoa
    29. SM Jackson
    30. B Gillam
    31. W Chen
    32. L Yan
    33. J Higginbotham
    34. M Cardenas
    35. J Waligorski
    36. E Applebaum
    37. L Phelps
    38. J Falcone
    39. K Kanchi
    40. T Thane
    41. A Scimone
    42. N Thane
    43. J Henke
    44. T Wang
    45. J Ruppert
    46. N Shah
    47. K Rotter
    48. J Hodges
    49. E Ingenthron
    50. M Cordes
    51. S Kohlberg
    52. J Sgro
    53. B Delgado
    54. K Mead
    55. A Chinwalla
    56. S Leonard
    57. K Crouse
    58. K Collura
    59. D Kudrna
    60. J Currie
    61. R He
    62. A Angelova
    63. S Rajasekar
    64. T Mueller
    65. R Lomeli
    66. G Scara
    67. A Ko
    68. K Delaney
    69. M Wissotski
    70. G Lopez
    71. D Campos
    72. M Braidotti
    73. E Ashley
    74. W Golser
    75. H Kim
    76. S Lee
    77. J Lin
    78. Z Dujmic
    79. W Kim
    80. J Talag
    81. A Zuccolo
    82. C Fan
    83. A Sebastian
    84. M Kramer
    85. L Spiegel
    86. L Nascimento
    87. T Zutavern
    88. B Miller
    89. C Ambroise
    90. S Muller
    91. W Spooner
    92. A Narechania
    93. L Ren
    94. S Wei
    95. S Kumari
    96. B Faga
    97. MJ Levy
    98. L McMahan
    99. P Van Buren
    100. MW Vaughn
    101. K Ying
    102. CT Yeh
    103. SJ Emrich
    104. Y Jia
    105. A Kalyanaraman
    106. AP Hsia
    107. WB Barbazuk
    108. RS Baucom
    109. TP Brutnell
    110. NC Carpita
    111. C Chaparro
    112. JM Chia
    113. JM Deragon
    114. JC Estill
    115. Y Fu
    116. JA Jeddeloh
    117. Y Han
    118. H Lee
    119. P Li
    120. DR Lisch
    121. S Liu
    122. Z Liu
    123. DH Nagel
    124. MC McCann
    125. P SanMiguel
    126. AM Myers
    127. D Nettleton
    128. J Nguyen
    129. BW Penning
    130. L Ponnala
    131. KL Schneider
    132. DC Schwartz
    133. A Sharma
    134. C Soderlund
    135. NM Springer
    136. Q Sun
    137. H Wang
    138. M Waterman
    139. R Westerman
    140. TK Wolfgruber
    141. L Yang
    142. Y Yu
    143. L Zhang
    144. S Zhou
    145. Q Zhu
    146. JL Bennetzen
    147. RK Dawe
    148. J Jiang
    149. N Jiang
    150. GG Presting
    151. SR Wessler
    152. S Aluru
    153. RA Martienssen
    154. SW Clifton
    155. WR McCombie
    156. RA Wing
    157. RK Wilson
    (2009)
    Science 326:1112–1115.
    https://doi.org/10.1126/science.1178534
  65. 65
  66. 66
  67. 67
  68. 68
  69. 69
  70. 70
  71. 71
  72. 72
  73. 73
  74. 74
  75. 75
  76. 76
  77. 77
    The sequence of the human genome
    1. JC Venter
    2. MD Adams
    3. EW Myers
    4. PW Li
    5. RJ Mural
    6. GG Sutton
    7. HO Smith
    8. M Yandell
    9. CA Evans
    10. RA Holt
    11. JD Gocayne
    12. P Amanatides
    13. RM Ballew
    14. DH Huson
    15. JR Wortman
    16. Q Zhang
    17. CD Kodira
    18. XH Zheng
    19. L Chen
    20. M Skupski
    21. G Subramanian
    22. PD Thomas
    23. J Zhang
    24. GL Gabor Miklos
    25. C Nelson
    26. S Broder
    27. AG Clark
    28. J Nadeau
    29. VA McKusick
    30. N Zinder
    31. AJ Levine
    32. RJ Roberts
    33. M Simon
    34. C Slayman
    35. M Hunkapiller
    36. R Bolanos
    37. A Delcher
    38. I Dew
    39. D Fasulo
    40. M Flanigan
    41. L Florea
    42. A Halpern
    43. S Hannenhalli
    44. S Kravitz
    45. S Levy
    46. C Mobarry
    47. K Reinert
    48. K Remington
    49. J Abu-Threideh
    50. E Beasley
    51. K Biddick
    52. V Bonazzi
    53. R Brandon
    54. M Cargill
    55. I Chandramouliswaran
    56. R Charlab
    57. K Chaturvedi
    58. Z Deng
    59. V Di Francesco
    60. P Dunn
    61. K Eilbeck
    62. C Evangelista
    63. AE Gabrielian
    64. W Gan
    65. W Ge
    66. F Gong
    67. Z Gu
    68. P Guan
    69. TJ Heiman
    70. ME Higgins
    71. RR Ji
    72. Z Ke
    73. KA Ketchum
    74. Z Lai
    75. Y Lei
    76. Z Li
    77. J Li
    78. Y Liang
    79. X Lin
    80. F Lu
    81. GV Merkulov
    82. N Milshina
    83. HM Moore
    84. AK Naik
    85. VA Narayan
    86. B Neelam
    87. D Nusskern
    88. DB Rusch
    89. S Salzberg
    90. W Shao
    91. B Shue
    92. J Sun
    93. Z Wang
    94. A Wang
    95. X Wang
    96. J Wang
    97. M Wei
    98. R Wides
    99. C Xiao
    100. C Yan
    101. A Yao
    102. J Ye
    103. M Zhan
    104. W Zhang
    105. H Zhang
    106. Q Zhao
    107. L Zheng
    108. F Zhong
    109. W Zhong
    110. S Zhu
    111. S Zhao
    112. D Gilbert
    113. S Baumhueter
    114. G Spier
    115. C Carter
    116. A Cravchik
    117. T Woodage
    118. F Ali
    119. H An
    120. A Awe
    121. D Baldwin
    122. H Baden
    123. M Barnstead
    124. I Barrow
    125. K Beeson
    126. D Busam
    127. A Carver
    128. A Center
    129. ML Cheng
    130. L Curry
    131. S Danaher
    132. L Davenport
    133. R Desilets
    134. S Dietz
    135. K Dodson
    136. L Doup
    137. S Ferriera
    138. N Garg
    139. A Gluecksmann
    140. B Hart
    141. J Haynes
    142. C Haynes
    143. C Heiner
    144. S Hladun
    145. D Hostin
    146. J Houck
    147. T Howland
    148. C Ibegwam
    149. J Johnson
    150. F Kalush
    151. L Kline
    152. S Koduru
    153. A Love
    154. F Mann
    155. D May
    156. S McCawley
    157. T McIntosh
    158. I McMullen
    159. M Moy
    160. L Moy
    161. B Murphy
    162. K Nelson
    163. C Pfannkoch
    164. E Pratts
    165. V Puri
    166. H Qureshi
    167. M Reardon
    168. R Rodriguez
    169. YH Rogers
    170. D Romblad
    171. B Ruhfel
    172. R Scott
    173. C Sitter
    174. M Smallwood
    175. E Stewart
    176. R Strong
    177. E Suh
    178. R Thomas
    179. NN Tint
    180. S Tse
    181. C Vech
    182. G Wang
    183. J Wetter
    184. S Williams
    185. M Williams
    186. S Windsor
    187. E Winn-Deen
    188. K Wolfe
    189. J Zaveri
    190. K Zaveri
    191. JF Abril
    192. R Guigó
    193. MJ Campbell
    194. KV Sjolander
    195. B Karlak
    196. A Kejariwal
    197. H Mi
    198. B Lazareva
    199. T Hatton
    200. A Narechania
    201. K Diemer
    202. A Muruganujan
    203. N Guo
    204. S Sato
    205. V Bafna
    206. S Istrail
    207. R Lippert
    208. R Schwartz
    209. B Walenz
    210. S Yooseph
    211. D Allen
    212. A Basu
    213. J Baxendale
    214. L Blick
    215. M Caminha
    216. J Carnes-Stine
    217. P Caulk
    218. YH Chiang
    219. M Coyne
    220. C Dahlke
    221. A Mays
    222. M Dombroski
    223. M Donnelly
    224. D Ely
    225. S Esparham
    226. C Fosler
    227. H Gire
    228. S Glanowski
    229. K Glasser
    230. A Glodek
    231. M Gorokhov
    232. K Graham
    233. B Gropman
    234. M Harris
    235. J Heil
    236. S Henderson
    237. J Hoover
    238. D Jennings
    239. C Jordan
    240. J Jordan
    241. J Kasha
    242. L Kagan
    243. C Kraft
    244. A Levitsky
    245. M Lewis
    246. X Liu
    247. J Lopez
    248. D Ma
    249. W Majoros
    250. J McDaniel
    251. S Murphy
    252. M Newman
    253. T Nguyen
    254. N Nguyen
    255. M Nodell
    256. S Pan
    257. J Peck
    258. M Peterson
    259. W Rowe
    260. R Sanders
    261. J Scott
    262. M Simpson
    263. T Smith
    264. A Sprague
    265. T Stockwell
    266. R Turner
    267. E Venter
    268. M Wang
    269. M Wen
    270. D Wu
    271. M Wu
    272. A Xia
    273. A Zandieh
    274. X Zhu
    (2001)
    Science 291:1304–1351.
    https://doi.org/10.1126/science.1058040
  78. 78
  79. 79
  80. 80
  81. 81
  82. 82
  83. 83
  84. 84
  85. 85
  86. 86
  87. 87
  88. 88

Decision letter

  1. Richard Amasino
    Reviewing Editor; University of Wisconsin Madison, United States
  2. Christian S Hardtke
    Senior Editor; University of Lausanne, Switzerland
  3. Richard Amasino
    Reviewer; University of Wisconsin Madison, United States
  4. Germán Martínez
    Reviewer; Swedish University of Agricultural Sciences | SLU · Department of Plant Biology and Forest Genetics, Sweden

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Natural depletion of H1 in sex cells causes DNA demethylation, heterochromatin decondensation and transposon activation" for consideration by eLife. Your article has been reviewed by two peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Christian Hardtke as the Senior Editor.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

For a long time, it has been known that there is a reprogramming of epigenetic marks in both the vegetative nucleus and the sperm nuclei. For example, it was known that in the vegetative nuclei there is a loss of H3K9me2, lack of activity of the chromatin remodeler DDM1 and high activity of the DNA demethylase DME and of the de novo methyltransferase DRM2. This manuscript adds new knowledge about the intricacy of the mechanisms leading to TE reactivation in the vegetative nucleus. The manuscript is well written, concise, and sincere in the presentation of the results including pointing to the limits of the analysis. If the issues raised below can be properly addressed, the manuscript should be suitable for publication in eLife.

Essential revisions:

1) TE reactivation observed by the authors is very modest compared to the analysis of DNA methylation for the VC that has been published before (1781 TEs in Calarco et al., 2012, Cell vs. 114 TEs on this manuscript). I think that this difference must be discussed. At the same time, in previous manuscripts it has been shown that companion cell demethylation mediated by DME takes place at loci that tend to be within smaller AT-rich TEs that are enriched for euchromatin-associated histone modifications. This is quite the opposite of what the authors find on the reactivated TEs and should be discussed in order to connect better with DME function.

2) Another point regarding the 114 reactivated TEs. In terms of proximity to genes these TEs are located in pericentromeric regions where genes and TEs coexist. Is this a feature of these TEs? Could they be involved in gene regulation? Do they give rise to more or a different profile of small RNAs compared to other TEs? Are these sRNAs more mobile to the SCs? This analysis could easily be added since datasets are already available and will help to dig into the characteristics of theses TEs and why they are targets of DME.

3) Because a very important function of DME is the regulation of imprinted genes I think it would be interesting to the overall community for the authors to also include an analysis (or at least comment) on the effects of H1 overexpression on imprinted genes. In principle DME and related glycosylases preferentially demethylate gene-adjacent sequences so this might be an important part of how it regulates gene expression.

4) It is intriguing why the authors use two different transgenic lines for H1 overexpression: one with the LAT52 promoter and a different one with the MSP1 promoter. Since the LAT52 promoter is active at later stages of pollen development, are chromocenters also seen at later stages of pollen development in the pLAT52::H1.1-RFP line? Can this be quantified and added to the manuscript along with a more detail about why the two different promoters were used?

Two relatively major issues in writing:

1)"Examination of DNA methylation in VCs from dme/+ heterozygous 155 plants (dme homozygous mutants are embryonic lethal), which produce a 50:50 ratio of dme 156 mutant and WT pollen, REVEALED PARTIAL RESTORATION of methylation at TSS and TTS of VC – 157 activated TEs (Figure 1D,E)."

Do you really mean "restoration"? Restoration would mean there was demethylation in the VC and then DNA methylation was added back. If in the dme VCs, there was simply not as much loss of DNA methylation, do not use the word restoration.

2) The following sentence (subsection “Depletion of H1 decondenses heterochromatin during male gametogenesis”) does not "make sense."

"We observed strong and weak chromocenters in 27% and 59%, respectively, of late microspore nuclei, whereas no chromocenters were observed in the VC at either bicellular or tricellular pollen stage (Figure 5C, Figure 5—figure supplement 1).”

"respectively" is used to match A and B with C and D. In the sentence there is no D (see below).

“we observed strong and weak chromocenters in 27% (A) and 59% (B), respectively, of late microspore nuclei (C).”

Something is missing in this sentence.

https://doi.org/10.7554/eLife.42530.034

Author response

For a long time, it has been known that there is a reprogramming of epigenetic marks in both the vegetative nucleus and the sperm nuclei. For example, it was known that in the vegetative nuclei there is a loss of H3K9me2, lack of activity of the chromatin remodeler DDM1 and high activity of the DNA demethylase DME and of the de novo methyltransferase DRM2. This manuscript adds new knowledge about the intricacy of the mechanisms leading to TE reactivation in the vegetative nucleus. The manuscript is well written, concise, and sincere in the presentation of the results including pointing to the limits of the analysis. If the issues raised below can be properly addressed, the manuscript should be suitable for publication in eLife.

We are delighted for the positive valuation from the reviewers and the reviewing editor, and grateful for their constructive comments and suggestions. In accordance to these valuable comments, we have performed additional experiments and analyses, and revised the manuscript.

Beyond the comments we received, we have further validated the causal relationship between DME demethylation and TE activation in the VC through performing pollen RNA-seq from dme/+ heterozygous plants (dme homozygous mutant is embryonic lethal). This additional data has been added as Figure 1G, and to the Figure 1—source data 1 and the text: “Consistently, RNA-seq of pollen from dme/+ heterozygous plants showed significantly reduced expression at the 114 VC-activated TEs (Figure 1G, Figure 1—source data 1). As dme/+ heterozygous plants produce half dme mutant and half WT pollen, we expect the transcription of VC-activated TEs to be reduced to roughly half in dme/+ pollen. The 114 VC-activated TEs are transcribed at levels close to this expectation (Figure 1G), with the median ratio of their transcription in dme/+ versus WT pollen being 0.43 (Figure 1—source data 1).”

We believe this result strengthens one of the major conclusions of our manuscript, i.e. TE activation in the VC is caused by DME-directed DNA demethylation.

Essential revisions:

1) TE reactivation observed by the authors is very modest compared to the analysis of DNA methylation for the VC that has been published before (1781 TEs in Calarco et al., 2012, Cell vs. 114 TEs on this manuscript). I think that this difference must be discussed. At the same time, in previous manuscripts it has been shown that companion cell demethylation mediated by DME takes place at loci that tend to be within smaller AT-rich TEs that are enriched for euchromatin-associated histone modifications. This is quite the opposite of what the authors find on the reactivated TEs and should be discussed in order to connect better with DME function.

We thank the reviewers and editor for the helpful comment. The 1781 TEs were identified as overlapping with 2270 VC-sperm CHH DMRs, almost all of which are hypermethylated in the VC in comparison to sperm (Calarco et al., 2012). Therefore, these TEs/DMRs should not be related to DME-directed demethylation or TE activation in the VC. Indeed, only 6 of the 2270 CHH DMRs overlap DME targets, and only 6 out of the 1781 TEs are activated in pollen.

However, there is indeed a contrast between the large number of small euchromatic TEs demethylated by DME in the VC and the small number of heterochromatic TEs we found actually activated. We thank the reviewers for pointing this out and agree that this is indeed worth more discussion, because one of the significant contributions of our manuscript is revealing the actual modest scale of TE activation in the VC. However, first of all, it is worth to note that DME demethylates both small euchromatic TEs and the edges of long heterochromatic TEs (Supplementary Figure 9G and H in Ibarra et al. 2012; Frost et al. 2018), although the latter are less in their number. For transcriptional regulation, the methylation at TE edges is most important as they tend to overlap the TSS.

Our results demonstrate that only 114 heterochromatic TEs in pericentromeric regions are activated in the VC, as a result of DME demethylation and developmental H1 depletion. This supports the “TE-driven model” we proposed in the last paragraph of the Discussion, which we have now expanded: “DME demethylates about ten thousand loci in the VC, most of which are small and euchromatic TEs (Ibarra et al., 2012). However, only a hundred heterochromatic TEs are activated, at least partially permitted by developmental H1 depletion. As small euchromatic TEs tend to be next to genes, DME demethylation regulates genes and is important for pollen fertility (Choi et al., 2002; Ibarra et al., 2012; Schoft et al., 2011). Our data show that developmental H1 depletion is also important for pollen fertility. Therefore, at least some TEs may be hijacking an essential epigenetic reprogramming process […]”

2) Another point regarding the 114 reactivated TEs. In terms of proximity to genes these TEs are located in pericentromeric regions where genes and TEs coexist. Is this a feature of these TEs?

Yes. As shown in Figure 1A and B and Figure 1–figure supplement 1A and B (subsection “Heterochromatic transposons are preferentially expressed in the vegetative cell”), being heterochromatic and pericentromeric is a significant feature of these reactivated TEs.

Could they be involved in gene regulation?

To understand whether nearby genes are regulated/affected by the activation of these TEs, we identified 70 genes within 1 kb of the 114 reactivated TEs. Analysis of our newly obtained dme/+ pollen RNA-seq data shows that the transcription of none of the 70 genes is significantly affected in dme/+ pollen, demonstrating that TE activation in the VC is unlikely involved in gene regulation. This is consistent with previous findings that DME regulates gene expression through the demethylation of small euchromatic TEs that are adjacent to genes, and is now elaborated in the Discussion section as quoted in response to Comment 1, specifically: “As small euchromatic TEs tend to be next to genes, DME demethylation regulates genes and is important for pollen fertility (Choi et al., 2002; Ibarra et al., 2012; Schoft et al., 2011)”.

Do they give rise to more or a different profile of small RNAs compared to other TEs? Are these sRNAs more mobile to the SCs? This analysis could easily be added since datasets are already available and will help to dig into the characteristics of theses TEs and why they are targets of DME.

We are grateful for the thoughtful suggestion. Using published pollen and sperm sRNA-seq data (Slotkin et al., 2009; GSE61028), we find that the 114 reactivated TEs are associated with more sRNAs of 21-24 nt lengths than other TEs in both pollen and sperm (see box plots Author response image 1; all P < 0.01, Kolmogorov–Smirnov test). This result is consistent with the proposal that TE transcription in the VC gives rise to sRNAs that can move into sperm (Slotkin et al., 2009). These 114 activated TEs are associated with more 24 nt sRNAs in sperm than in pollen (see the pie charts Author response image 1), suggesting 24 nt sRNAs might be more mobile. This is different from the previous proposal that the mobile sRNAs are 21 nt sRNAs produced from transcribed TEs (Slotkin et al., 2009), and an alternative explanation also cannot be excluded, i.e. sperm may produce more 24 nt sRNAs from these TEs than the VC. Furthermore, because the differences in the abundance of sRNAs between the reactivated and other TEs are small, and the published data involve no replicates, we do not feel these analyses are conclusive. Neither this manuscript aims to elucidate on sRNA movement between VC and sperm, which we feel require elaborate studies. Therefore, we did not add these analyses to the manuscript.

Author response image 1

3) Because a very important function of DME is the regulation of imprinted genes I think it would be interesting to the overall community for the authors to also include an analysis (or at least comment) on the effects of H1 overexpression on imprinted genes. In principle DME and related glycosylases preferentially demethylate gene-adjacent sequences so this might be an important part of how it regulates gene expression.

We appreciate the excellent suggestion, and have performed the requested analysis and revised the manuscript by adding the following to the Results:

“Given the importance of H1 removal for DME-directed DNA demethylation and the well demonstrated role of DME demethylation in regulating gene expression (Choi et al., 2002; Ibarra et al., 2012; Schoft et al., 2011), we investigated the contribution of H1 to gene regulation in pollen. RNA-seq was performed using pollen from the pVC::H1 line (#2), which showed strong H1 expression in VC (Figures 2B and 3A). […] Consistently, DME targets likely regulating imprinted genes are associated with significantly less H1 in somatic tissues than the VC DME targets that are dependent on H1 (Figure 3C).” (Subsection “H1 represses transposons via methylation-dependent and independent mechanisms”).

4) It is intriguing why the authors use two different transgenic lines for H1 overexpression: one with the LAT52 promoter and a different one with the MSP1 promoter. Since the LAT52 promoter is active at later stages of pollen development, are chromocenters also seen at later stages of pollen development in the pLAT52::H1.1-RFP line? Can this be quantified and added to the manuscript along with a more detail about why the two different promoters were used?

These are excellent questions. As we explained in the manuscript, “To understand how H1 affects DME activity, we ectopically expressed H1 in the VC. To ensure H1 incorporation into VC chromatin, we used the pLAT52 promoter, which is expressed from the late microspore stage immediately prior to Pollen Mitosis 1, and is progressively upregulated in VC during later stages of pollen development (Eady, Lindsey and Twell, 1994; Grant-Downton et al., 2013). Using pLAT52 to drive the expression of H1.1 tagged with mRFP (simplified as pVC::H1), we observed continuous H1-mRFP signal in the VC at the bicellular and tricellular pollen stages, while the signal was undetectable in the generative cell and sperm (Figure 2B). H1-mRFP signal was also undetectable in late microspores (Figure 2B), probably due to the low activity of pLAT52 at this stage (Eady et al., 1994).” (Subsection “Vegetative-cell-expressed H1 impedes DME from accessing heterochromatic transposons”)

pVC[LAT52]::H1 VC also showed no obvious chromocenters (n > 500; Figure 2B). This suggests either that H1 expression is not strong enough in pVC[LAT52]::H1, or other factors are involved in heterochromatin decondensation in the VC. […] These results show that H1 is sufficient to promote heterochromatic foci in late microspores, thus demonstrating the causal relationship between H1 depletion and the decondensation of heterochromatin.” (Subsection “Depletion of H1 decondenses heterochromatin during male gametogenesis”)

The quantification of chromocenters in the pLAT52::H1.1-RFP line has been added to the text (bolded and underlined in the quotation).

Two relatively major issues in writing:

1)"Examination of DNA methylation in VCs from dme/+ heterozygous 155 plants (dme homozygous mutants are embryonic lethal), which produce a 50:50 ratio of dme 156 mutant and WT pollen, REVEALED PARTIAL RESTORATION of methylation at TSS and TTS of VC – 157 activated TEs (Figure 1D,E)."

Do you really mean "restoration"? Restoration would mean there was demethylation in the VC and then DNA methylation was added back. If in the dme VCs, there was simply not as much loss of DNA methylation, do not use the word restoration.

In accordance, the text has been changed to “revealed an intermediate level of methylation at TSS and TTS of VC-activated TEs”.

2) The following sentence (subsection “Depletion of H1 decondenses heterochromatin during male gametogenesis”) does not "make sense."

"We observed strong and weak chromocenters in 27% and 59%, respectively, of late microspore nuclei, whereas no chromocenters were observed in the VC at either bicellular or tricellular pollen stage (Figure 5C, Figure 5—figure supplement 1).”

"respectively" is used to match A and B with C and D. In the sentence there is no D (see below).

“we observed strong and weak chromocenters in 27% (A) and 59% (B), respectively, of late microspore nuclei (C).”

Something is missing in this sentence.

We apologize for the confusion caused by the ambiguous expression, and have revised the text into “we observed strong and weak chromocenters, respectively, in 27% and 59% of late microspore nuclei […]”.

https://doi.org/10.7554/eLife.42530.035

Article and author information

Author details

  1. Shengbo He

    Department of Cell and Developmental Biology, John Innes Centre, Norwich, United Kingdom
    Contribution
    Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing—original draft
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3773-9995
  2. Martin Vickers

    Department of Cell and Developmental Biology, John Innes Centre, Norwich, United Kingdom
    Contribution
    Data curation, Data analysis
    Competing interests
    No competing interests declared
  3. Jingyi Zhang

    Department of Cell and Developmental Biology, John Innes Centre, Norwich, United Kingdom
    Contribution
    Methodology, Acquisition of data
    Competing interests
    No competing interests declared
  4. Xiaoqi Feng

    Department of Cell and Developmental Biology, John Innes Centre, Norwich, United Kingdom
    Contribution
    Conceptualization, Resources, Supervision, Funding acquisition, Investigation, Methodology, Project administration, Writing—review and editing
    For correspondence
    xiaoqi.feng@jic.ac.uk
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4008-1234

Funding

Biotechnology and Biological Sciences Research Council (BBL0250431)

  • Jingyi Zhang
  • Xiaoqi Feng

Gatsby Charitable Foundation

  • Shengbo He
  • Xiaoqi Feng

H2020 European Research Council (804981 SexMeth)

  • Xiaoqi Feng

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank David Twell for the pDONR-P4-P1R-pLAT52 and pDONR-P2R-P3-mRFP vectors, the John Innes Centre Bioimaging Facility (Elaine Barclay and Grant Calder) for their assistance with microscopy, and the Norwich BioScience Institute Partnership Computing infrastructure for Science Group for High Performance Computing resources. This work was funded by a Biotechnology and Biological Sciences Research Council (BBSRC) David Phillips Fellowship (BB/L025043/1; SH, JZ and XF), a European Research Council Starting Grant ('SexMeth' 804981; XF) and a Grant to Exceptional Researchers by the Gatsby Charitable Foundation (SH and XF).

Senior Editor

  1. Christian S Hardtke, University of Lausanne, Switzerland

Reviewing Editor

  1. Richard Amasino, University of Wisconsin Madison, United States

Reviewers

  1. Richard Amasino, University of Wisconsin Madison, United States
  2. Germán Martínez, Swedish University of Agricultural Sciences | SLU · Department of Plant Biology and Forest Genetics, Sweden

Publication history

  1. Received: October 8, 2018
  2. Accepted: May 26, 2019
  3. Accepted Manuscript published: May 28, 2019 (version 1)
  4. Version of Record published: June 26, 2019 (version 2)

Copyright

© 2019, He et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,818
    Page views
  • 385
    Downloads
  • 2
    Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Crossref, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Chromosomes and Gene Expression
    2. Evolutionary Biology
    Ryan Bracewell et al.
    Research Article Updated
    1. Cell Biology
    2. Chromosomes and Gene Expression
    Andrey Poleshko et al.
    Research Article Updated