Two MADS-box transcription factors mediate epigenetic control of tomato fruit ripening

doi:10.21203/rs.3.rs-2003863/v1

Download PDF

Article

Two MADS-box transcription factors mediate epigenetic control of tomato fruit ripening

https://doi.org/10.21203/rs.3.rs-2003863/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

DNA methylation is an important epigenetic mark involved in many biological processes in eukaryotes. It was recently proposed to be one of the most important factors controlling fruit ripening. In tomato, dysfunction of DML2, a DEMETER-like DNA demethylase, causes genome-wide DNA hypermethylation and dramatically delays fruit ripening. However, the link between the change in DNA methylation and ripening phenotype is unclear. In this study, we revealed a novel function of RIN and TDR4 (two well-known ripening-related transcription factors) in mediating DML2-dependent epigenetic control of fruit ripening. We found that double mutations in RIN and TDR4, which are both silenced in dml2-3, are sufficient to mimic the ripening phenotype of dml2 fruits. Restoration of RIN expression in dml2-3 largely rescued ripening phenotype of dml2 fruits, and majority of RIN binding peaks were also restored in dml2 fruits. Interestingly, we found that in addition to regulating RIN and TDR4 expression, DML2 also regulates RIN binding in the genome to control fruit ripening. At a subset of RIN targets, RIN binding is blocked in dml2-3. DNA methylation is known to directly or indirectly influence TF binding, however, the sphere of influence of DNA methylation on TF binding has not been determined in any organism. Here, we found that the loss of RIN binding is highly associated with DNA methylation increase within ~ 100 bp of the binding site, and is also associated with an enrichment of histone H3, a core protein in the nucleosome. In summary, our findings filled the missing link between epigenetic change and its regulation of fruit ripening through identifying two important downstream factors of SlDML2, and we determined for the first time the sphere of influence of DNA methylation on TF binding, thus furthering the understanding of the effect of DNA methylation on TF binding in vivo.

DNA methylation

demethylase

DML2

RIN

TDR4

Fruit ripening is a highly coordinated and regulated process that involves a set of physiological and biochemical changes, such as softening, coloration, and aroma biosynthesis^1–3. The control of fruit ripening depends in part on phytohormones and transcription factors (TFs)^4–6. In tomato, the model plant for study of fleshy fruit development and ripening, ripening is accompanied by ethylene production and ethylene signaling transduction^5,7. In addition, some ripening-related TFs, including RIN, TDR4, MBP7, NOR, TAGL1, and AP2a, have been characterized as a core set of ripening regulators^8–13, and these TFs function in ethylene-dependent and ethylene-independent pathways in the control of fruit ripening^11,14,15. RIN is one of the most important TFs in tomato fruit ripening. A recent study showed that RIN targets the promoters of many ripening-related genes, such as ACS2, ACS4, PSY1, and PG2a, and that knockout of RIN can delay ripening⁸. RIN homologs MaMADS1 and MaMADS2 are also important for ripening in banana¹⁶. Besides phytohormones and TFs, DNA methylation was recently proposed to be the third important factor in the regulation of fruit ripening^17–19. DNA methylation, an epigenetic modification in mammals and plants, is involved in many biological processes, such as human aging, gene imprinting, and genome stability^18,20,21, and is usually associated with inactive transcription in the genome^20,22,23. Global change of DNA methylation has been observed during ripening of several fruits, such as tomato and strawberry^17,24−26. In tomato, ripening-induced expression of SlDML2, a DNA demethylase gene, causes a global loss of DNA methylation during fruit ripening¹⁷. Knockout or knockdown of SlDML2 dramatically delays fruit ripening and preferentially increases DNA methylation at promoters of ripening-related genes^17,18. However, the link between the change in DNA methylation and ripening phenotype remains unclear, and whether and how the SlDML2-mediated DNA hypomethylation modulates TF binding at promoters is not known.

SlDML2 prefers to target promoter regions in the tomato genome¹⁷, indicating that SlDML2 may influence binding of ripening-related TFs and thereby regulate fruit ripening. DNA methylation could affect TF binding directly or indirectly. In mammals, the sensitivity of TFs to DNA methylation was first investigated in vitro in the late 1980s, and an EMSA assay showed that DNA methylation prevents TF binding²⁷. TF-binding motifs enriched with CpGs, such as those of NRF1 (nuclear respiratory factor 1), E2F, and ETS, might be directly sensitive to DNA methylation due to steric hindrance^27–31. TF binding could also be indirectly modulated by DNA methylation through recruitment of methyl-CpG-binding domain proteins, nucleosome positioning, and histone modifications^32,33. A genome-wide analysis of DNA methylation and TF binding profiles in mouse embryonic stem cells recently demonstrated that NRF1 gained new binding sites in hypomethylated regions in DNMT (DNA methyltransferase) triple knockout cells, and that NRF1 was the first TF shown to be sensitive to DNA methylation genome-wide in vivo²⁸. Most studies about the effects of DNA methylation on TF binding have been conducted with mammals. To date, how the change of DNA methylation affects genome-wide TF binding profiles in vivo has not been investigated in any plant system.

Here, we report that the DML2-dependent epigenetic control of tomato fruit ripening is mediated by two MAD-box TFs, RIN and TDR4. To determine the effects of RIN and TDR4, we generated and studied single mutants of SlDML2, RIN, TDR4, and NOR and double mutants of RIN and TDR4 in the tomato cv. Ailsa Craig (AC) background using CRISPR/Cas9-mediated gene editing. Our results demonstrate that RIN and TDR4, both of which are hypermethylated and silenced in sldml2-3, function redundantly in regulating fruit ripening. We show that double mutations in RIN and TDR4 are sufficient to mimic the ripening phenotype of dml2 fruits. Transcriptome analysis of rin/tdr4-1 and dml2-3 revealed that a > 70% change in the gene expression in dml2-3 could be explained by the simultaneous silencing of RIN and TDR2 in dml2-3. To investigate the importance of RIN as a downstream factor of SlDML2, we used the CBC promoter, which is not regulated by DNA methylation in dml2-3, to drive RIN in the dml2-3 mutant. We found that the restoration of RIN expression largely rescued the ripening phenotype of dml2-3 and recovered ethylene production and ACC concentration, demonstrating that RIN is a key downstream factor mediating SlDML2-dependent fruit ripening. To study whether and how RIN binding is regulated by SlDML2-mediated DNA hypomethylation during ripening, we then performed chromatin immunoprecipitation followed by high throughput sequencing (ChIP-seq) in AC and dml2-3 fruits that carry pCBC-RIN-3xFLAG. We show that majority of RIN binding peaks are restored in dml2, which is consistent with the largely rescued ripening phenotype of dml2. However, at a subset of RIN targets, RIN binding is blocked by dml2-3 and this loss of RIN binding is highly correlated with the increase in DNA methylation in dml2-3. Further analysis showed that RIN binding is most likely blocked by a > 20% increase in DNA methylation within 100 bp of the binding site. In addition, we found that the loss of RIN binding in dml2-3 is also associated with histone H3 enrichment, suggesting that RIN binding might be indirectly regulated by SlDML2. In summary, our findings provide new insight into the epigenetic regulation of fruit ripening, and also increase the understanding of the effect of DNA methylation on TF binding in plants.

Two MADS-box transcription factors function redundantly in downstream of DML2 in regulating fruit ripening

We first used CRISPR/Cas9-mediated gene editing to generate single mutants of SlDML2, i.e., dml2-3 and dml2-4, in the tomato cv. Ailsa Craig (AC) background (Fig. 1a and Supplementary Fig. 1). We then found that dml2-3 and dml2-4 had delayed fruit ripening compared with the wild type (WT) AC (Fig. 1a). That finding was consistent with previous results showing that dysfunction of SlDML2 in tomato significantly delayed fruit ripening^17,18. Thousands of genes are known to be differentially expressed in dml2-1 relative to the WT^17,18. However, key factors that mediate SlDML2-dependent fruit ripening had not been previously identified. To identify these key factors, we performed whole-genomic bisulfite sequencing and RNA-seq in the fruits of dml2-3 and AC at 25 dpa (days post anthesis) and 34 dpa (Breaker stage), respectively. By analyzing transcriptomes and DNA methylomes, we noticed that several well-known ripening-related genes are hypermethylated and silenced in dml2-3 vs. AC. Four MADS-box transcription factors, including RIN, TAGL1, TDR4, and MBP7, have been reported to be important for fruit ripening^11,13,14^,³⁴. The MADS-box transcription factors usually function as multimers³⁵. Previous studies and our results showed that TAGL1, TDR4, and MBP7, can physically interact with RIN^11,13 (Supplementary Fig. 2). Among these four transcription factors, we found that RIN and TDR4 but not MBP7 or TAGL1 were hypermethylated and silenced in dml2-3 vs. the WT (Fig. 1b-1d and Supplementary Fig. 3a and 3b).

Given that RIN and TDR4 are simultaneously silenced in dml2, we determined whether dysfunction of both RIN and TDR4 is sufficient to mimic the ripening phenotype of dml2. We generated rin-1, rin-2, tdr4-1, tdr4-2, rin/tdr4-1, and rin/tdr4-2 mutants using the CRISPR/Cas9 gene editing system and found that the ripening phenotypes of rin/tdr4 double mutants are more severe than those of rin and tdr4 single mutants, suggesting a redundant function of RIN and TDR4 in regulating fruit ripening (Fig. 1a and Supplementary Fig. 1). In addition, fruits of rin/tdr4-1 and rin/tdr4-2 turned yellow at 45 dpa rather than at 34 dpa, which is similar with the ripening of dml2 mutants (Fig. 1a). These results demonstrated that RIN and TDR4 function redundantly in regulating fruit ripening, and that the rin/tdr4 mutant can mimic the ripening phenotype of dml2 mutants. TDR4 was previously identified as a direct binding target of RIN ^13,36. Consistent with those previous results, we found that the TDR4 was partially but not fully repressed in rin-1 (Supplementary Fig. 3c), which explained why rin/tdr4 double mutant has a more severe phenotype than rin single mutant. On the other hand, we showed that the expression of RIN was not affected in tdr4-1 during ripening (Supplementary Fig. 3c). These results indicated that even though RIN functions upstream of TDR4, TDR4 function is not fully dependent on RIN since TDR4 is genetically redundant with RIN in regulating ripening (Fig. 1a).

Silencing of RIN and TDR4 in dml2 significantly contributes to the change in gene expression in dml2

As noted in the previous section, the dml2 mutation alters the expression of thousands of genes¹⁷. Considering that RIN and TDR4 are transcription factors, we determined how much of the dml2 transcriptome can be explained by the silencing of RIN and TDR4 in dml2. We compared transcriptomes of fruits at 34 dpa (the breaker stage) in AC, dml2-3, and rin-1/tdr4-1 with three biological replicates. Compared to AC, we identified 6,576 upregulated differentially expressed genes (up-DEGs) and 5,955 downregulated differentially expressed genes (down-DEGs) in dml2-3, and 6,374 up-DEGs and 5,972 down-DEGs in rin/tdr4-1 (Fig. 2a and Supplementary Table 1). We found that 4,389 dml2-3 down-DEGs overlapped with rin/tdr4-1 down-DEGs, and that 4390 dml2-3 up-DEGs overlapped with rin/tdr4-1 up-DEGs (Fig. 2b). These results demonstrated that about 70% ((4389+4390) / (5955+6576)) of the change in gene expression in dml2-3 vs. AC could be explained by the simultaneous silencing of RIN and TDR4 in dml2-3. We also assessed the transcript levels of these shared DEGs in AC fruits at 25 dpa and 34 dpa, and found that the majority of these shared DEGs are ripening-induced DEGs (DEGs in AC 34 dpa vs. AC 25 dpa) (Fig. 2c).

On the other hand, we noticed that at 34 dpa, AC fruits were at the breaker stage, while rin/tdr4-1 and dml2-3 fruits remained green. To determine whether the similarity between rin/tdr4-1 and dml2-3 transcriptomes was due to the similar developmental stage, we generated a nor knockout mutant using gene editing (Fig. 1a and Supplementary Fig. 1), and found that nor-1 fruits remained green at 34 dpa, which is similar to the fruits of rin/tdr4-1 and dml2-3 (Fig. 1a). We then analyzed the transcriptomes of AC, rin-1, tdr4-1, rin/tdr4-1, dml2-3, and nor-1. According to a correlation analysis, rin/tdr4-1 and dml2-3 have the most similar transcriptomes and were clustered into one subgroup that did not include nor-1 (Fig. 2d). This indicated that similarity between rin/tdr4-1 and dml2-3 transcriptomes was due to silencing of RIN and TDR4 in dml2.

Restored RIN expression rescues the dml2 ripening phenotype

To investigate the importance of RIN in mediating DML2-dependent fruit ripening, we wanted to restore RIN expression in the dml2-3 mutant. Considering that both the E8 promoter, which is commonly used in fruit-specific expression, and the native promoter of RIN are hypermethylated and repressed in the dml2 mutant (Fig. 1b, 1c and Supplementary Fig. 4a), we used the CBC promoter to drive RIN (pCBC::RIN-3xFLAG). pCBC, which is a previously reported fruit-specific promoter³⁷, was not regulated by SlDML2 according to our DNA methylome and transcriptome data (Supplementary Fig. 4b). We transformed pCBC::RIN-3xFLAG into the AC WT, and crossed the plants carrying pCBC::RIN-3xFLAG with dml2-3 to generate pCBC::RIN-3xFLAG/dml2-3 plants. The fruits of pCBC::RIN-3xFLAG/dml2-3 turned light orange at 38 dpa and turned red at 45 dpa, while dml2-3 fruits remained green at 38 dpa and turned light yellow at 45 dpa (Fig. 3a). An examination of the carotenoid concentration in the mutants and in pCBC::RIN-3xFLAG /dml2-3 revealed that the contents of β-carotene and lycopene were significantly lower in the fruits of rin-1/tdr4-1 and dml2-3 mutants than in AC fruits. β-carotene and lycopene contents were substantially but not fully rescued in pCBC::RIN-3xFLAG /dml2-3 fruits at 45 dpa (Fig. 3b). These results demonstrated that the restored RIN expression largely rescued the ripening phenotype of dml2.

Although ethylene is known to be important for tomato fruit ripening, it was still unknown at the outset of our study whether and how SlDML2 interacts with RIN or TDR4 to affect ethylene production or signaling during tomato fruit ripening. We examined ethylene production in the fruits of AC, dml2-3, and dml2-4 at different stages, i.e., at 25, 34, 38, and 45 dpa. As shown in Figure 3c, a characteristic burst of ethylene synthesis was observed after the onset of AC fruit ripening, whereas ethylene production was substantially inhibited in dml2-3 and dml2-4 mutants, demonstrating that SlDML2 is required for ethylene production during ripening. We also examined ethylene production in rin-1, tdr4-1, and rin-1/tdr4-1, and found that ethylene production was slightly inhibited in tdr4-1 but substantially inhibited in rin-1 and rin-1/tdr4-1 mutants (Fig. 3c). Consistent with ethylene production results, the ethylene precursor 1-aminocyclopropane-1-carboxylic acid (ACC) level was high in AC fruits, a little lower in tdr4-1 fruits, and much lower in rin-1, rin-1/tdr4-1, and dml2-3 mutant fruits at 38 dpa (Fig. 3d).

Based on these results, we hypothesized that SlDML2-dependent ethylene production is mainly mediated by RIN. To test this hypothesis, we examined the ethylene production and ACC level in the fruits of pCBC::RIN-3xFLAG /dml2-3. In pCBC::RIN-3xFLAG /dml2-3, the ACC content and ethylene production were both largely restored compared with dml2-3 (Fig. 3c and 3d). These results indicated that DML2-dependent ethylene production is largely mediated by RIN. The dml2 mutants notably repressed the expression of both RIN and TDR4, while SlDML2 expression was not affected in the rin-1, tdr4-1, or rin-1/tdr4-1 mutants (Fig. 3e), suggesting that RIN and TDR4 do not regulate SlDML2 expression. Together, our results revealed that SlDML2 functions upstream of RIN and TDR4 in regulation of ethylene production and tomato fruit ripening.

We then investigated the ability of AC, dml2-3, rin-1, tdr4-1, and rin-1/tdr4-1 fruits to undergo ripening in response to ethylene treatment. The detached green fruits at 30 dpa were treated with external ethephon (2 mM). External ethylene accelerated the onset of the red color in AC and of the orange color in tdr4-1 and rin-1 (Fig. 4). However, dml2-3 and rin-1/tdr4-1 fruits only turned light yellow rather than red or orange after the external ethylene treatment. These results suggested that the influence of dml2-3 or rin-1/tdr4-1 on fruit ripening not only results from the deficiency in ethylene production in the mutants but might also involve a deficiency of factors downstream or independent of ethylene.

DML2 is required for RIN binding at a subset RIN targets

DML2 preferentially targets gene promoters, and DML2-demethylated regions are close to RIN binding sites^17,19. Besides silencing RIN and TDR4 gene expression, does DML2-regulated fruit ripening involve regulation of RIN binding? In this study, we determined whether and how dml2 mutation affects RIN binding in genome-wide. To compare RIN-binding profiles in AC and dml2 backgrounds, we performed chromatin immunoprecipitation followed by high throughput sequencing (ChIP-seq) in AC and dml2-3 fruits that carry pCBC::RIN-3xFLAG. In AC, we defined a total of 14,129 RIN-binding loci, including promoters of TDR4 and CNR, which were previously identified as direct binding targets of RIN (Fig. 5a and Supplementary Table 2). We found that the majority of RIN binding was maintained (“maintained” binding sites) in dml2-3 compared to the signal in AC (Fig. 5a). However, some of RIN binding was lost (“blocked” binding sites) in dml2-3 compared to the signal in AC (Fig. 5a). The ChIP-seq signals of RIN enrichment at several representative maintained binding sites and representative blocked binding sites are shown (Fig. 5b and 5c).

To confirm that the difference of RIN binding in AC and dml2 was not due to sequencing quality, we validated RIN binding using ChIP-qPCR at several loci. At promoters of AP2a and PG2a genes, which were identified as maintained binding loci, the ChIP-qPCR results showed RIN enrichment in AC and dml2-3 (Fig. 5d). In contrast, at promoters of NOR and ZISO genes, which were identified as blocked binding loci, we observed RIN enrichment in AC but not in dml2-3 (Fig 5d). These results demonstrated that, besides regulating RIN and TDR4 expression, dml2-3 mutation also affected RIN binding at a subset of RIN-binding targets in the genome, which is consistent with the observation that restoration of RIN can partially rather than fully rescue the dml2-3 phenotype (Fig. 3a).

Loss of RIN binding is positively correlated with the increase of DNA methylation in dml2

To assess the influence of DNA hypermethylation on RIN binding in dml2, we first examined whether the hypermethylation in dml2-3 is maintained in pCBC::RIN-3xFLAG /dml2-3. In a comparison of DNA methylomes of fruits of AC, dml2-3, and pCBC::RIN-3xFLAG /dml2-3 at 34 dpa, we identified 68,230 hyper differentially methylated regions (hyper DMR) in dml2-3 relative to AC at 34 dpa (Supplementary Table 3), and found that these DMRs were hypermethylated in both dml2-3 and pCBC-RIN-3xFLAG /dml2-3 in all three sequence contexts, including CG, CHG, and CHH (H is C, A, or T) (Fig. 6a); this suggested that the hypermethylation caused by dml2 was not compromised by the restored expression of RIN. We then divided a total of 14,129 RIN-binding loci into three groups based on their change in DNA methylation level in dml2-3 relative to AC at 34 dpa (Supplementary Fig. 5). The three groups have similar ChIP-seq signals in AC but not in the dml2-3 background (Fig. 6b). Among the three groups in dml2-3, the RIN binding decreases as DNA methylation increased (Fig. 6b). In particular, the group of peaks with a > 20% increase in DNA methylation had a dramatically lower RIN binding signal in dml2-3 (Fig. 6b).

To further study the relationship between DNA methylation and RIN binding, we calculated the change of DNA methylation and change of RIN enrichment in dml2-3 relative to AC for each binding peak. Subsequent correlation analysis showed that the increase in DNA methylation is highly correlated with loss of RIN binding in dml2-3 (correlation coefficients were -0.78 and -0.79 for two replicates) (Fig. 6c). Randomly selected regions were used as negative controls (Fig. 6c). In addition, we examined DNA methylation levels at several individual RIN-binding peaks. Consistent with our statistical results, we observed an increase in DNA methylation at blocked RIN-binding loci but not at maintained RIN-binding loci (Fig. 6d, Supplementary Fig. 6 and 7). These results demonstrated that the lost of RIN binding is highly correlated with the increase in DNA methylation in dml2-3, suggesting that SlDML2-dependent DNA demethylation is required for RIN binding at a subset of RIN targets during fruit ripening.

Sphere of influence of DNA methylation on RIN binding

DNA methylation can directly or indirectly influence TF binding, however, the sphere of influence of DNA methylation on TF binding has not been investigated in any organism. Previous studies revealed that RIN-binding sites are adjacent to ripening-induced hypo DMRs^17,19. Consistent with those previous reports, we found that our identified RIN-binding sites are located near dml2-3 hyper DMRs (Supplementary Fig. 8a). To further investigate how the DNA hypermethylation influences RIN binding in dml2, among the 14,129 RIN-binding peaks, we identified 8,980 peaks that have dml2-3 hyper DMRs within 200 bp. However, we found that among the 8,980 hyper DMR-associated binding sites, RIN enrichment was only affected at 1,384 in dml2 (Fig. 7a and Supplementary Table 4). For the other 7,596 RIN-binding sites, although they have dml2-3 hyper DMRs within 200 bp, their binding was not affected by dml2-3 (Fig. 7a and Supplementary Table 4). We then compared the two groups of binding sites to find their differences. The upstream and downstream 1-kb regions surrounding binding peaks were used and divided into 20 continuous 100-bp bins for the following analysis. For each 100-bp bin, the differences in DNA methylation between dml2-3 and AC were calculated and plotted (Fig. 7b and Supplementary Fig. 8b). Interestingly, we observed different DNA methylation patterns in the two groups of peaks. For the 1,384 blocked peaks, the bins surrounding the binding peak summits [+ 100 bp] were significantly hypermethylated in dml2-3 relative to the WT; however, for the 7,596 maintained RIN-binding peaks, bins at the same position were not differentially methylated in dml2-3 vs. the WT (Fig. 7b). Among the 8,980 peaks, the change in DNA methylation of bins flanking the binding summit [+ 100 bp] gradually increased with the loss of RIN binding in dml2-3 (Fig. 7c). Similar patterns, however, were not evident for the methylation levels of the bins 100 bp away from the peak summit [bins from -100 to -200 bp and from +100 to +200 bp] (Fig. 7c and Supplementary Fig. 8b). These results suggested that RIN binding is highly associated with the DNA methylation change in regions within 100 bp of the binding site. Considering Figure 6b, our results suggest that RIN binding is likely blocked by a > 20% increase in DNA methylation that occurs within 100 bp of the binding site.

A recent study demonstrated that in DNA methylation mutants of Arabidopsis, a change in the DNA methylation pattern may substantially influence chromatin architecture³⁸. In AC and dml2-3, we performed ChIP-seq for histone H3, a core protein in the nucleosome. For the bins flanking the peak summit [+ 100 bp], H3 enrichment gradually increased with the loss of RIN binding in dml2-3 (Fig. 7c). H3 levels of bins located > 100 bp away from the binding summit (Fig. 7c) served as controls. These results suggested that DNA methylation may block RIN binding by influencing nearby chromatin architecture.

Fruit ripening is a highly coordinated process under hormonal, transcriptional, and epigenetic control. In tomato, RIN and TDR4 are two well-known transcription factors that target and regulate many ripening-related genes, such as TAGL1 and AP2a^36,39. In this study, we revealed a novel function of RIN and TDR4 in mediating epigenetic control of fruit ripening. Previous studies reported that dysfunction of demethylase SlDML2 causes genome-wide DNA hypermethylation and dramatically delayed ripening^17,18. However, the link between the change in DNA methylation and the ripening phenotype had not been elucidated. The current results suggest that RIN and TDR4 could be the missing link. We found that RIN and TDR4 were both silenced by dml2-3, and that the rin/tdr4 double mutant is sufficient to mimic the ripening phenotype of dml2 fruit (Fig. 1). In addition, we found that restoration of RIN expression rescued the ripening phenotype, ethylene production, and carotenoid accumulation of dml2 fruits (Fig. 2). Our genetic evidence, together with our comprehensive analysis of transcriptomes and DNA methylomes, indicates that RIN and TDR4 are key downstream factors of SlDML2 in the regulation of fruit ripening.

DNA methylation can directly or indirectly regulate TF binding, however, it is still not clear how far the influence of DNA methylation on TF binding may reach. During fruit ripening, SlDML2 preferentially targets promoter regions of ripening-related genes ^17,18. Before the current study,whether and how TF binding (e.g. RIN binding) is affected by the SlDML2-mediated loss of DNA methylation during ripening was unknown. Previous study indicated that DML2 may broadly influence RIN binding in the genome, since DML2-demethylated regions are close (~ 200 bp) to RIN binding sites ^17,19. We generated and compared genome-wide RIN-binding profiles in AC and dml2-3 backgrounds. However, our data show that majority of RIN binding (at ~ 7000 peaks) is not blocked in dml2, although the distances between these maintained RIN binding sites and their nearby DML2-demethylated regions are not more than 200 bp. Further analysis shows that RIN binding is blocked at a subset of binding peaks (1384/8980) in dml2, and these DNA-methylation sensitive RIN binding sites were highly associated with a > 20% increase in DNA methylation within 100 bp of the binding site in dml2. To our knowledge, this is the first time the sphere of influence of DNA methylation on TF binding is determined in vivo in a higher organism.

In plants, inhibition of TF binding by DNA methylation is poorly understood, because most studies about DNA methylation have used Arabidopsis, in which gene promoters are barely regulated by DNA methylation⁴⁰. In tomato, in contrast, ripening-related genes are extensively regulated by SlDML2-mediated DNA demethylation^17,18. In this study, our results demonstrate that binding of RIN, one of the most important ripening-related TFs, is regulated by DNA methylation at a subset of targets during ripening, and that genome-wide loss of RIN binding is associated with the increase in DNA methylation in dml2-3 (Fig. 6). In mammals, most studies have found that DNA methylation inhibits TF binding due to steric hindrance²⁹, but several studies have found that DNA methylation might modulate TF binding through crosstalk with methyl-CpG-binding domain proteins, nucleosome positioning, and histone modifications^32,33. Researchers recently reported that a change in the DNA methylation pattern in DNA methylation mutants could alter chromatin accessibility in Arabidopsis³⁸. Here, we found that genome-wide loss of RIN binding was associated with an increase in histone H3, a core component of the nucleosome, in addition to being associated with an increase in DNA methylation in dml2-3, suggesting that RIN binding might be indirectly influenced by DNA methylation. In summary, our study demonstrated that two TFs, RIN, and TDR4, play key roles in mediating SlDML2-related epigenetic regulation of fruit ripening (Fig. 7d). In addition, by comparing and characterizing RIN-binding profiles in plants with different methylation patterns, our research has increased the understanding of how DNA methylation affects TF binding in plants.

Plant materials and growth conditions

The following were grown in a condition-controlled greenhouse (light:16 h at 25℃; dark: 8 h, 22℃; 80% relative humidity): WT tomato AC (Solanum lycopersicum cultivar Ailsa Craig) plants; dml2-3, dml2-4, rin-1, rin-2, tdr4-1, tdr4-2, rin-1/tdr4-1,rin-1/tdr4-2, and nor-1 CRISPR knock-out mutant plants; and pCBC::RIN-3xFLAG, pCBC::TDR4-3xFLAG, pCBC::RIN-3xFLAG/dml2-3, pDML2::gDML2-3xFLAG/dml2-3 transgenic plants. CRISPR/Cas9 construction and tomato transformation were performed as previously described⁴¹, and oligonucleotide sequences for guide RNAs are listed in Supplementary Table 5. Three independent biological replicates with nine fruits each were collected at different ripening stages for analysis. All information concerning the mutations is presented Figure S1.

Whole-genome bisulfite sequencing and data analyses

Genomic DNA was extracted from fruits at different stages using Plant DNAzol (Invitrogen), purified with a DNA mini kit (Qiagen), and then sent to the Core Facility for Genomics at the Shanghai Center for Plant Stress Biology (PSC) for whole-genome bisulfite sequencing (WGBS).

For data analysis, reads containing adapter and low-quality reads (q < 20) were trimmed using cutadatp⁴² and Trimmomatic⁴³, respectively, and clean reads that were shorter than 45 nt were discarded. The clean reads were then mapped to Tomato Genome version SL4.0 using BSMAP with default parameters⁴⁴. When methratio.py script was used to extract the methylation ratio from BSMAP mapping results, the option -r was used to remove potential PCR duplicates. DNA methylation levels in every 200-bp window with a step size of 50 bp were compared between the WT and mutant plants using the R package “methylkit”⁴⁵, which considered the variation among biological replicates. The windows with a q-value < 0.01 and at least 10 for the absolute value of methylation percentage change were considered for further analysis, and windows within 100 bp of each other were condensed to larger regions.

Ethylene content and response

To measure ethylene content, transgenic CRRISPR mutant and WT fruits were harvested at different ripening stages. After the fruits were kept for approximately 2 h in a 300-ml gas-tight jar, a 1-mL gas sample was taken and injected into an Agilent 6890 N gas chromatograph equipped with a flame ionization detector (Agilent, Palo Alto, CA, USA). Ethylene production was calculated as previously described¹⁴.

Fruits harvested at the mature green stage (30 dpa) were dipped into a 2 mM fresh ethephon (Sangon Biotech Co., Ltd) aqueous solution for 5 min. Each treatment was represented by three independent biological replicates with six fruits each. Tomato fruits the mature green stage (30 dpa) were placed into an airtight 5-liter plastic container with 5 ppm of 1-MCP that was generated by dissolving of 1-MCP–releasing powder in water. The 1-MCP were replenished every 48 h.

ACC Content

Fruits of the WT and mutants were harvested at 4 days after the initiation of ripening (38 dpa). Frozen tomato fruit tissue was ground to a powder in liquid nitrogen. A 100-mg sample of the powder was dissolved in 1 ml of water in 5-ml screw-cap tubes. After ultrasonic treatment for 30 min, the supernatant was collected and purified with a Sep-Pak^TM C18 reverse-phase extraction cartridge. The column was washed once with 2 ml of water and then with 2 ml of methanol. A 3-ml volume of 80% methanol-water (with 5% ammonia) was applied to the column for elution. The pooled extracts were dried under nitrogen gas. The residue was resuspended in 50% methanol in water (500 µL). The AB SCIEX QTRAP 5500 system features a Accucore PFP column (100 mm X 2.1 mm, 2.6 µm particle size, Thermo Fisher). The mobile phases, consisting of 0.1% FA in water (solvent A) and 0.1% FA in methanol (solvent B), were pumped at a flow rate of 0.3 ml min^-1 with a volume ratio (A : B) of 50:50. A 5-µL volume of suspension (500 µL)was injected into the column at a column temperature of 30 °C and was separated for 10 min.

Carotenoid content

Carotenoids of fruits were extracted and then quantified by HPLC as described previously⁴⁶.

RNA-seq

Total RNA was extracted and genomic DNA was removed with RNase-free DNase (RNeasy mini kit; Qiagen). For RNA-seq, the cDNA library was constructed and sequenced as previously described⁴⁷.

For processing RNA-seq data, adapter and low-quality reads were trimmed using cutadatp⁴² and Trimmomatic⁴³, respectively, before alignment. The trimmed reads were aligned to Tomato Genome version SL4.0 using STAR(2.5.4b)⁴⁸ using parameters --quantMode TranscriptomeSam GeneCounts. The ITAG4.0 annotation gff3 file was used. The column 4 (count of antisense reads) of STAR outputs file ReadsPerGene.out.tab) was used for analysis of DEGs. rsem-calculate-expression in RSEM was used to estimate gene expression⁴⁹. DESeq2⁵⁰ was used to compute the DEGs (fpkm>1, padj<0.01, and fold-change>1.2 )

ChIP-seq

WT fruits and pCBC::RIN-3xFLAG, pCBC::RIN-3xFLAG/dml2-3 tagline fruits were harvested at break stage (34 dpa). Protein G dynabeads (Invitrogen, cat# 10003D) were used for antibody binding. The antibody for FLAG was anti-Flag M2 (Sigma; F3165). The ChIP assay was performed as previously described⁴⁷.

For ChIP-seq data, adapter and 3’-end of reads with a low-quality score (q<20) were trimmed using cutadatp⁴² and Trimmomatic⁴³, respectively, and clean reads with < 50 nt were discarded. Clean paired-end reads were mapped to Tomato Genome version SL4.0 with Bowtie2 (version 2.3.4.1) with default parameters⁵¹, and samtools⁵² was used to remove potential PCR duplicates; only uniquely aligned reads were kept for the downstream analysis. Model-based analysis for ChIP-Seq (Macs2: v2.1.1)⁵³ was then used to identify the peaks with the “callpeak” function and the parameters --format BAMPE -q 0.01. The IDR (irreproducible discovery rate) was used to measure the reproducibility of peaks identiﬁed from replicate experiments⁵⁴. The peaks identified from the WT without FLAG were then removed. bigWig files were generated by bamCompare (RIN-Flag vs. input) in DeepTools (v3.0.1) using parameters --binSize 10 --numberOfProcessors 10 --normalizeUsing CPM –ignoreDuplicates --operation subtract⁵⁵. The bigWig files were then visualized with Integrative Genomics Browser (IGB) software⁵⁶. The scores over peaks were calculated using computeMatrix in DeepTools (v3.0.1)⁵⁰. The fold-change of RIN-Flag signal in the dml2 mutant vs input was generated by bamCompare in DeepTools (v3.0.1) using parameters --binSize 10 --normalizeUsing CPM –ignoreDuplicates --operation ratio. Negative control regions were then generated using the ‘shuffle’ command in BEDTools to simulate the length distribution of RIN-binding peaks.

ChIP qRT-PCR

For ChIP-qPCR, the enrichment of DNA fragments was measured by quantitative real-time PCR using the BioRad CFX96 Touch Real-Time PCR System. The purified solution of DNA fragments was diluted five times, and 2 µl was used as the template in a 15-µl PCR reaction with SYBR Green mix (Takara). The actin gene was used as an internal control. Primer sequences are listed in Supplementary Table 5.

Real-time RT–PCR analysis

For real-time RT-PCR analysis, total RNA was extracted and contaminating DNA was removed with RNase-free DNase (RNeasy mini kit; Qiagen). mRNA (1 mg) was used for the first-strand cDNA synthesis with the Takara RT-PCR System following the manufacturer’s instructions. The cDNA reaction mixture was then diluted five times, and 1 µl was used as template in a 15-µl PCR reaction with SYBR Green mix (Takara). The EF1α gene was used as an internal control. Primer sequence information is listed in Supplementary Table 5.

Data Availability

The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus and are accessible through GEO Series accession number GSE210590.

Acknowledgements

This work was supported by the National Key R&D Program of China (2021YFD1000200 to Z.L.). The Strategic Priority Research Program of the Chinese Academy of Sciences (XDB27040000 to Z.L.) and the National Nature Science Foundation of China (31900394).

Author Contributions

Q.N. and Y.X. performed experiments. H.H. and Y.M. did the bioinformatics analysis. Q.N., Y.X., D.T., P.L. and R.L. contributed to the Agrobacterium-mediated transformation and sample collection. L.L. and S.W. were involved in Y2H and Split-LUC assays. J.N. and B. Z. were involved in ethylene detection and the discussion of the research. Q.N., H.H. and Z.L. designed the study, interpreted the data, and wrote the manuscript.

Competing Interests

The authors declare no competing interests.

Seymour, G.B., Østergaard, L., Chapman, N.H., Knapp, S. & Martin, C. Fruit development and ripening. Annu Rev Plant Biol 64, 219–41 (2013).
Forlani, S., Masiero, S. & Mizzotti, C. Fruit ripening: the role of hormones, cell wall modifications, and their relationship with pathogens. J Exp Bot 70, 2993–3006 (2019).
Saladié, M. et al. A reevaluation of the key factors that influence tomato fruit softening and integrity. Plant Physiol 144, 1012–28 (2007).
Giovannoni, J., Nguyen, C., Ampofo, B., Zhong, S. & Fei, Z. The Epigenome and Transcriptional Dynamics of Fruit Ripening. Annu Rev Plant Biol 68, 61–84 (2017).
Klee, H.J. & Giovannoni, J.J. Genetics and control of tomato fruit ripening and quality attributes. Annu Rev Genet 45, 41–59 (2011).
Fenn, M.A. & Giovannoni, J.J. Phytohormones in fruit development and maturation. Plant J 105, 446–458 (2021).
Alba, R. et al. Transcriptome and selected metabolite analyses reveal multiple points of ethylene control during tomato fruit development. Plant Cell 17, 2954–65 (2005).
Ito, Y. et al. Re-evaluation of the rin mutation and the role of RIN in the induction of tomato ripening. Nat Plants 3, 866–874 (2017).
Gao, Y. et al. Re-evaluation of the nor mutation and the role of the NAC-NOR transcription factor in tomato fruit ripening. J Exp Bot 71, 3560–3574 (2020).
Wang, R. et al. Re-evaluation of transcription factor function in tomato fruit development and ripening with CRISPR/Cas9-mutagenesis. Sci Rep 9, 1696 (2019).
Vrebalov, J. et al. Fleshy fruit expansion and ripening are regulated by the Tomato SHATTERPROOF gene TAGL1. Plant Cell 21, 3041–62 (2009).
Chung, M.Y. et al. A tomato (Solanum lycopersicum) APETALA2/ERF gene, SlAP2a, is a negative regulator of fruit ripening. Plant J 64, 936–47 (2010).
Shima, Y. et al. Tomato FRUITFULL homologues act in fruit ripening via forming MADS-box transcription factor complexes with RIN. Plant Mol Biol 82, 427–38 (2013).
Bemer, M. et al. The tomato FRUITFULL homologs TDR4/FUL1 and MBP7/FUL2 regulate ethylene-independent aspects of fruit ripening. Plant Cell 24, 4437–51 (2012).
Karlova, R. et al. Transcriptome and metabolite profiling show that APETALA2a is a major regulator of tomato fruit ripening. Plant Cell 23, 923–41 (2011).
Elitzur, T. et al. Banana MaMADS Transcription Factors Are Necessary for Fruit Ripening and Molecular Tools to Promote Shelf-Life and Food Security. Plant Physiol 171, 380–91 (2016).
Lang, Z. et al. Critical roles of DNA demethylation in the activation of ripening-induced genes and inhibition of ripening-repressed genes in tomato fruit. Proc Natl Acad Sci U S A 114, E4511-E4519 (2017).
Liu, R. et al. A DEMETER-like DNA demethylase governs tomato fruit ripening. Proc Natl Acad Sci U S A 112, 10804–9 (2015).
Zhong, S. et al. Single-base resolution methylomes of tomato fruit development reveal epigenome modifications associated with ripening. Nat Biotechnol 31, 154–9 (2013).
Matzke, M.A. & Mosher, R.A. RNA-directed DNA methylation: an epigenetic pathway of increasing complexity. Nat Rev Genet 15, 394–408 (2014).
Tang, D., Gallusci, P. & Lang, Z. Fruit development and epigenetic modifications. New Phytol 228, 839–844 (2020).
Zhang, H., Lang, Z. & Zhu, J.K. Dynamics and function of DNA methylation in plants. Nat Rev Mol Cell Biol 19, 489–506 (2018).
Liu, R. & Lang, Z. The mechanism and function of active DNA demethylation in plants. J Integr Plant Biol 62, 148–159 (2020).
Cheng, J. et al. Downregulation of RdDM during strawberry fruit ripening. Genome Biol 19, 212 (2018).
Lü, P. et al. Genome encode analyses reveal the basis of convergent evolution of fleshy fruit ripening. Nat Plants 4, 784–791 (2018).
Huang, H. et al. Global increase in DNA methylation during orange fruit development and ripening. Proc Natl Acad Sci U S A 116, 1430–1436 (2019).
Tate, P.H. & Bird, A.P. Effects of DNA methylation on DNA-binding proteins and gene expression. Curr Opin Genet Dev 3, 226–31 (1993).
Domcke, S. et al. Competition between DNA methylation and transcription factors determines binding of NRF1. Nature 528, 575–9 (2015).
Yin, Y. et al. Impact of cytosine methylation on DNA binding specificities of human transcription factors. Science 356(2017).
Gaston, K. & Fried, M. CpG methylation has differential effects on the binding of YY1 and ETS proteins to the bi-directional promoter of the Surf-1 and Surf-2 genes. Nucleic Acids Res 23, 901–9 (1995).
Campanero, M.R., Armstrong, M.I. & Flemington, E.K. CpG methylation as a mechanism for the regulation of E2F activity. Proc Natl Acad Sci U S A 97, 6481–6 (2000).
Zhu, H., Wang, G. & Qian, J. Transcription factors as readers and effectors of DNA methylation. Nat Rev Genet 17, 551–65 (2016).
Bell, O., Tiwari, V.K., Thomä, N.H. & Schübeler, D. Determinants and dynamics of genome accessibility. Nat Rev Genet 12, 554–64 (2011).
Vrebalov, J. et al. A MADS-box gene necessary for fruit ripening at the tomato ripening-inhibitor (rin) locus. Science 296, 343–6 (2002).
Honma, T. & Goto, K. Complexes of MADS-box proteins are sufficient to convert leaves into floral organs. Nature 409, 525–9 (2001).
Fujisawa, M., Nakano, T., Shima, Y. & Ito, Y. A large-scale identification of direct targets of the tomato MADS box transcription factor RIPENING INHIBITOR reveals the regulation of fruit ripening. Plant Cell 25, 371–86 (2013).
Hiwasa-Tanase, K. et al. Novel promoters that induce specific transgene expression during the green to ripening stages of tomato fruit development. Plant Cell Rep 31, 1415–24 (2012).
Zhong, Z. et al. DNA methylation-linked chromatin accessibility affects genomic architecture in Arabidopsis. Proc Natl Acad Sci U S A 118(2021).
Fujisawa, M. et al. Transcriptional regulation of fruit ripening by tomato FRUITFULL homologs and associated MADS box proteins. Plant Cell 26, 89–101 (2014).
Zhang, X. et al. Genome-wide High-Resolution Mapping and Functional Analysis of DNA Methylation in Arabidopsis. Cell 126, 1189–1201 (2006).
Niu, Q. et al. Expanding the scope of CRISPR/Cas9-mediated genome editing in plants using an xCas9 and Cas9-NG hybrid. J Integr Plant Biol 62, 398–402 (2020).
Kechin, A., Boyarskikh, U., Kel, A. & Filipenko, M. cutPrimers: A New Tool for Accurate Cutting of Primers from Reads of Targeted Next Generation Sequencing. J Comput Biol 24, 1138–1143 (2017).
Bolger, A.M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–20 (2014).
Xi, Y. & Li, W. BSMAP: whole genome bisulfite sequence MAPping program. BMC Bioinformatics 10, 232 (2009).
Akalin, A. et al. methylKit: a comprehensive R package for the analysis of genome-wide DNA methylation profiles. Genome Biol 13, R87 (2012).
Fraser, P.D. et al. Manipulation of phytoene levels in tomato fruit: effects on isoprenoids, plastids, and intermediary metabolism. Plant Cell 19, 3194–211 (2007).
Niu, Q. et al. A histone H3K4me1-specific binding protein is required for siRNA accumulation and DNA methylation at a subset of loci targeted by RNA-directed DNA methylation. Nat Commun 12, 3367 (2021).
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Li, B. & Dewey, C.N. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics 12, 323 (2011).
Love, M.I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15, 550 (2014).
Langmead, B. & Salzberg, S.L. Fast gapped-read alignment with Bowtie 2. Nat Methods 9, 357–9 (2012).
Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–9 (2009).
Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol 9, R137 (2008).
Li, Q., Brown, J.B., Huang, H. & Bickel, P.J. Measuring reproducibility of high-throughput experiments. The Annals of Applied Statistics 5(2011).
Ramirez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res 44, W160-5 (2016).
Freese, N.H., Norris, D.C. & Loraine, A.E. Integrated genome browser: visual analytics platform for genomics. Bioinformatics 32, 2089–95 (2016).

There is NO Competing Interest.

supplementaryFigures.pdf
Supplementary Fig. 1.Diagrams showing mutations in rin-1, rin-2, rdr4-1, tdr4-2, rin/tdr4-1, rin/tdr4-2, dml2-3, dml2-4, and nor-1. The sgRNAs used in CRISPR-Cas9 gene editing system were indicated. Supplementary Fig. 2. RIN physically interacts with TDR4, MBP7 and TAGL1. a, Yeast-two-Hybird assays showing protein interactions among RIN, TDR4, MBP7, and TAGL1. (DDO: SD-Leu/-Trp medium; QDO:SD-Ade/-His/-Leu/-Trp medium) b, Split luciferase complementation assays showing protein interactions among RIN, TDR4, MBP7, and TAGL1 in tobacco leaves. Circles indicate leaf regions that were infiltrated with Agrobacterium strains containing the indicated constructs. (Scale bar = 1 cm). Supplementary Fig. 3. MBP7 and TAGL1 were not silenced in dml2-3. a&b, Methylation levels of promoters of MBP7 (a) and TAGL1(b) in AC and dml2-3 fruits at 25 dpa and 34 dpa (left). Transcript profiles of MBP7 (a) and TAGL1 (b) in AC and dml2-3 fruits (right). The whole-genome bisulfite sequencing data were displayed with Integrative Genome Browser software (IGB). Transcriptome data of AC and dml2-3 fruits at 25 dpa and 34 dpa were shown. c, qRT-PCR analysis showing the relative expression of RIN, TDR4, MBP7in different mutant fruits at 34 dpa. Values are means ± SD. One-tailed Student’s t-test was used to determine significance, ***p < 0.001 Supplementary Fig. 4. CBC promoter was not silenced in dml2-3. a, Methylation levels and transcript profiles of E8 in AC and dml2-3fruits at 34 dpa. b, Methylation levels and transcript profiles of CBCin AC and dml2-3 fruits at 34 dpa. The whole-genome bisulfite sequencing data were displayed with IGB (Top), where each vertical bar represents an mC, and the height of the bar indicates the methylation level. Transcriptome data of AC and dml2-3 fruits at 34 dpa were used to show transcript profiles (Bottom). Supplementary Fig. 5. Analysis of RIN-binding peaks. RIN-binding peaks were divided into three groups, where the change in DNA methylation was not increased (difference < 0.05), little increased (difference > 0.05 and < 0.2), or greatly increased (difference > 0.2) in dml2-3 relative to the WT. The boxplots show the change of DNA methylation levels in dml2-3relative to the WT in the three groups. Bar plot show the numbers of RIN-binding peaks in the three groups. Supplementary Fig. 6. RIN ChIP-seq signal and DNA methylation level at several representative maintained RIN binding peaks. RIN ChIP signals in pCBC::RIN-3xFLAG and pCBC::RIN-3xFLAG/dml2-3 were shown (Top). The ChIP signal in AC served as negative control. The DNA methylation levels in AC, dml2-3, and pCBC::RIN-3xFLAG/dml2-3 were displayed with IGB (Bottom). Supplementary Fig. 7. RIN ChIP-seq signal and DNA methylation level at several representative blocked RIN binding peaks. RIN ChIP signals in pCBC::RIN-3xFLAG and pCBC::RIN-3xFLAG/dml2-3 were shown (Top). The ChIP signal in AC served as negative control. The DNA methylation levels in AC, dml2-3, and pCBC::RIN-3xFLAG/dml2-3 were displayed with IGB (Bottom). Supplementary Fig. 8. RIN-binding was not affected by change of DNA methylation that occurs 100bp away from binding site. a, dml2-3hyper DMR density around RIN binding peaks. b, Boxplots showing the change of mCG, mCHG and mCHH in dml2-3 relative to the WT in the two groups of RIN binding peaks. The maintained and blocked RIN-binding peaks were used in this analysis. DNA methylation levels of 100-bp bins within ± 2 kb of binding peak summits were used.
SupplementaryTable1.xlsx
Supplementary Table 1. List of Differentially expressed genes (DEGs) in dml2-3 and rin/tdr4-1 compared to AC.
SupplementaryTable2.xlsx
Supplementary Table 2. List of RIN-binding peaks in AC and dml2-3.
supplementaryTable3.xlsx
Supplementary Table 3. List of Differentially methylated regions (DMRs) in dml2-3 relative to AC fruits at 34 dpa.
SupplementaryTable4.xlsx
Supplementary Table 4. List of maintained and blocked hyperDMR-associated RIN binding peaks.
SupplementaryTable5.xlsx
Supplementary Table 5. Primers and oligonucleotide sequences.

Download PDF

Version 1

posted

You are reading this latest preprint version

Two MADS-box transcription factors mediate epigenetic control of tomato fruit ripening

Status:

Version 1

Abstract

Figures

Main

Results

Discussion

Materials And Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1