The roles of epigenetic regulators in plant regeneration: Exploring patterns amidst complex conditions

Abstract Plants possess remarkable capability to regenerate upon tissue damage or optimal environmental stimuli. This ability not only serves as a crucial strategy for immobile plants to survive through harsh environments, but also made numerous modern plant improvements techniques possible. At the cellular level, this biological process involves dynamic changes in gene expression that redirect cell fate transitions. It is increasingly recognized that chromatin epigenetic modifications, both activating and repressive, intricately interact to regulate this process. Moreover, the outcomes of epigenetic regulation on regeneration are influenced by factors such as the differences in regenerative plant species and donor tissue types, as well as the concentration and timing of hormone treatments. In this review, we focus on several well-characterized epigenetic modifications and their regulatory roles in the expression of widely studied morphogenic regulators, aiming to enhance our understanding of the mechanisms by which epigenetic modifications govern plant regeneration.


Introduction
Plants possess a remarkable capacity for regeneration to repair their damaged tissues.This feature has become an agronomically important trait, serving as the foundation for key plant improvement procedures, such as grafting, haplotype breeding, and genetic transformation.Plants can regenerate via a diverse range of developmental processes (Box 1).At the cellular level, regardless of the type of plant regeneration, a common key feature is the acquisition of cell pluripotency through a transition in cell fate from terminally differentiated somatic cells.This process is associated with an intensive reprogramming of gene expression profiles but involves limited, if any, changes in genetic information.Instead, dynamic epigenetic changes, both globally and in key developmental genes, are believed to be critical for this process.
In this review, we summarize advances in studying some of the most well-characterized epigenetic modifications during plant regeneration, including DNA methylation, trimethylation at lysine 27 of histone H3 (H3K27me3), trimethylation at lysine 4 of histone H3 (H3K4me3), and histone H3/H4 acetylation (H3/H4 acetylation).We then focus on key morphogenetic gene families to summarize and illustrate the possible mechanisms of epigenetic regulation associated with plant regeneration.For additional details, we recommend excellent recent reviews on this topic (Lee and Seo 2018;Aflaki et al. 2022;Chen et al. 2023b;Liu et al. 2023b).

DNA methylation and plant regeneration
DNA methylation in plants occurs in 3 primary sequence contexts, CG, CHG, and CHH (where H represents A, T, or C), which are catalyzed by different families of DNA methyltransferases.In Arabidopsis (Arabidopsis thaliana), METHYLTRANSFERASE 1 (MET1) maintains CG DNA methylation during DNA replication, while CHROMOMETHYLASE 3 (CMT3) is primarily responsible for CHG DNA methylation.CHH DNA methylation is either de novo deposited in euchromatic regions by the RNA-directed DNA METHYLATION (RdDM) pathway enzyme DOMAIN REARRANGED METHYLTRANSFERASE 2 (DRM2) or maintained in heterochromatic regions by CMT2 (Law and Jacobsen 2009;Matzke and Mosher 2014).Additionally, DNA methylation in all sequence contexts can be actively removed by 5methylcytosine DNA glycosylases, including REPRESSOR OF SILENCING 1 (ROS1) and its homologous proteins TRANSCRIPTIONAL ACTIVATOR DEMETER (DME), DEMETER-LIKE 2 (DML2), and DML3 (Zhang et al. 2018).

Variation in DNA methylation and regeneration efficiency
Dynamic changes in global DNA methylation levels and local distribution patterns are recognized as common features of plant regeneration in various plant species (Chakrabarty et al. 2003;Tanurdzic et al. 2008;Stroud et al. 2013;Zakrzewski et al. 2017;Grzybkowska et al. 2018).The specific patterns of DNA methylation can be influenced by several factors, including genetic variation, the source of the explant, and experimental conditions (Loschiavo et al. 1989;Santos and Fevereiro 2002;Chakrabarty et al. 2003;Tanurdzic et al. 2008;Shibukawa et al. 2009;Stroud et al. 2013;Gao et al. 2014;Zakrzewski et al. 2017;Grzybkowska et al. 2018).For example, genotype-dependent DNA methylomes have been observed in different cultivars, such as japonica vs. indica rice (Oryza sativa) and Lumian vs. Yuzao cotton (Gossypium hirsutum), which are also associated with differential regenerative capabilities (Abe and Futsuhara 1986;Altpeter et al. 2016;Hsu et al. 2018;Guo et al. 2020).Mechanistically, since DNA methylation occurs in various sequence contexts (CG, CHG, and CHH) and is catalyzed by distinct families of DNA methyltransferases (Wendte and Schmitz 2018), changes in DNA sequences can directly affect the target recognition of these enzymes.Alternatively, genetic variations in genes encoding DNA methyltransferases or demethylases may lead to differences in gene expression patterns and altered functions of these enzymes, with pleiotropic effects on the DNA methylome.Similarly, developmental stage-dependent changes in DNA methylation and plant regeneration efficiency have also been reported in various plant organs and species (Becerra et al. 2004;Cardoza and Stewart 2004;Naing et al. 2015;Zhang et al. 2018;Sun et al. 2020;Tang et al. 2020).
However, interpreting the causal relationship between DNA methylation and plant regeneration efficiency is a complex matter.First, at a global scale, a linear relationship between changes in DNA methylation and gene expression is often lacking (Berdasco et al. 2008;Stroud et al. 2013;Guo et al. 2020;Sun et al. 2020;Liu et al. 2022;Zheng et al. 2022).Second, DNA methylation undergoes dynamic changes during plant regeneration.For instance, during cotton regeneration, there is a decrease in CHH methylation at the embryonic calli stage but not at the nonembryonic calli stage (Guo et al. 2020).Similarly, the total level of DNA methylation in rapeseed (Brassica napus) calli appears to fluctuate across different stages of regeneration, displaying 2 peaks at Day 12 and Day 30 on callus induction medium (CIM; Gao et al. 2014).

DNA methylation and somaclonal variation
It has long been observed that plants regenerated through tissue culture often exhibit diverse phenotypic traits that are distinct from each other and from the donor plants.This phenomenon is known as somaclonal variation (Bairu et al. 2010).Similarly, only a limited number of loci showed common changes in DNA methylation in different rice regenerants (Stroud et al. 2013).One possible underlying mechanism involves stochastic changes in the genetic material of somatic cells due to random transposable element insertions (Ozeki et al. 1997;Ong-Abdullah et al. 2015).These changes may also be attributed to the dysregulated functions of DNA methyltransferases and the genome-wide perturbation of DNA methylation (Tanurdzic et al. 2008).Supporting this notion, research has confirmed that all major Arabidopsis methyltransferase and demethylase genes, including MET1, CMT2, CMT3, DRM1, and ROS1, undergo dynamic changes in

ADVANCES
• Time-scale profiling using single-cell sequencing techniques has begun to offer cellular-level details of cell type composition and cell fate transitions during plant regeneration.
• Multiomics profiling techniques offer a genome-wide perspective on the transition of chromatin states and gene expression patterns that underlie changes in plant cell fate.
• Central elements in various plant signaling pathways, including TOR kinase and PIF transcription factors, directly modify or interact with epigenetic complexes, such as PRC2 and INO80, to regulate plant development and organogenesis.

Box 1. Different modes of plant regeneration
• Somatic embryogenesis: A single somatic cell can be induced by auxin to form an embryo, which develops into a whole plant.Somatic embryos can either be generated directly from individual somatic cells or indirectly from embryonic callus.
• De novo organogenesis: This includes de novo shoot organogenesis and de novo root organogenesis.In the direct de novo organogenesis pathway, shoots or roots are directly produced from somatic tissues and organs.In the indirect pathway, pluripotent callus is produced from detached plant explants, from which shoots and roots are subsequently generated.expression during callus proliferation (Grzybkowska et al. 2018;Liu et al. 2018aLiu et al. , 2022;;Sun et al. 2020).
The stochastic changes in DNA methylation may also occur at both the cellular and locus levels.DNA methylation patterns have been profiled in plants using maize microspores at single-cell resolution through bisulfite-converted randomly integrated fragments sequencing (BRIF-Seq) and using soybean single-cell root hairs (Hossain et al. 2017;Li et al. 2019).Alternative methods for profiling cell type-specific epigenetic information have also been reported, such as fluorescence-activated cell sorting (FACS) coupled with whole-genome bisulfite sequencing (WGBS) or Chromatin Immunoprecipitation sequencing (Kawakatsu et al. 2016;Lee et al. 2019b;Zhu et al. 2023) and single-cell Assay for Transposase-Accessible Chromatin with sequencing (Marand et al. 2021a(Marand et al. , 2021b)).However, despite these advances, investigating DNA methylation at single-cell resolution in the context of plant regeneration has posed technical challenges.
Analysis at the whole-genome scale has uncovered distinct patterns of changes in DNA methylation in various genomic features during plant regeneration (Tanurdzic et al. 2008;Stroud et al. 2013;Shim et al. 2022).For instance, during the transition from leaves to calli in Arabidopsis, changes in DNA methylation predominantly occur in transposable element regions, sparing genic regions (Shim et al. 2022).Additionally, dedifferentiated Arabidopsis suspension cells exhibit hyper-DNA methylation in euchromatin and hypo-DNA methylation in heterochromatin (Tanurdzic et al. 2008).In regenerated rice plants, the loss of DNA methylation is enriched in gene promoters (Stroud et al. 2013).As different plant species possess varying genomic compositions, including differences in the types and enrichment of repetitive sequences (IWGSC 2018;Shang et al. 2023;Chen et al. 2023a), the response of DNA methylation at the genome-wide scale may significantly differ across these diverse plant species.
A deficiency in PRC2 function leads to irregular organ development and homeotic transformations in plants, such as changes from hypocotyls and roots to flower-like organs (Goodrich et al. 1997;Kinoshita et al. 2001;Holec and Berger 2012).The disrupted maintenance of cell fate can be attributed to disturbances in the expression patterns of cell identity genes that are normally maintained by PRC2-mediated H3K27me3 deposition.The critical role of PRC2 in regulating cell fate transitions is further underscored by the substantially disrupted progress of plant regeneration.
Numerous lines of evidence support the crucial function of PRC2 in preserving cell identities in a differentiated state and its overall inhibitory effects on callus formation and somatic embryogenesis (Mozgová et al. 2017).For instance, an analysis using Arabidopsis mutants with impaired PRC2 function demonstrated that the terminally differentiated root hair cells undergo additional rounds of mitotic divisions, leading to the development of callus-like tissue, which can eventually give rise to somatic embryos in these mutants (Ikeuchi et al. 2015).Moreover, the expression of a nonmodifiable histone variant, H3 at residue K27 with an alanine (H3.3 K27A ), in Arabidopsis resulted in enhanced callus formation (Fal et al. 2023).Similarly, in response to wounding, the expression of H3.15, an Arabidopsis H3 variant lacking lysine residue 27, is induced.This induction promotes cell proliferation and callus formation by reducing the levels of H3K27me3 (Yan et al. 2020).The presence and sequence conservation of H3.15 in multiple eudicots species suggest that the wounding-induced incorporation of H3.15 and its promotion of callus formation may serve as a conserved wounding response mechanism in plants (Yan et al. 2020).
In addition to genetic evidence, the use of GSK-J4, a small molecule that hinders the activity of H3K27me3 demethylases, markedly diminished callus development in peach (Prunus persica) leaf explants (Zheng et al. 2022).Taken together, all of the studies support the idea that decreased H3K27me3 levels facilitate the initial phases of plant regeneration, where cells undergo dedifferentiation and proliferate to attain pluripotency.
While deficient PRC2 function can lead to the spontaneous formation of callus from root hair cells (Chanvivattana et al. 2004;Ikeuchi et al. 2015), Arabidopsis clf-50 swn-1 mutants lost the ability to generate callus from leaf blades on CIM (He et al. 2012).This implies that other factors also influence plant regeneration in PRC2-deficient mutants.Notably, callus induced from leaf blades lacks the embryonic features observed in callus spontaneously derived from root hairs and cotyledons, as evidenced by the absence of lipid accumulation (He et al. 2012;Ikeuchi et al. 2015).Therefore, the size and proliferation rate of callus do not necessarily reflect the efficiency of PRC2 in establishing cell pluripotency.In support of this notion, even though the incorporation of the H3.3 K27A histone variant improved callus production, these calli encountered difficulties in generating functional shoots (Fal et al. 2023).
As mentioned above, in mutants defective in PRC2 function, the seemingly contradictory observation of spontaneous callus formation from root hairs and impaired callus formation in leaf blades may also be related to the specific mutant alleles used.
These studies have uncovered the involvement of different alleles and genetic backgrounds, including hypomorphic alleles such as swn-1 in the Ws background (He et al. 2012) and swn-7 in the Col-0 background (Ikeuchi et al. 2015), as well as the null swn-3 allele (Chanvivattana et al. 2004).Additionally, callus formation closely mirrors the developmental process of lateral roots (Sugimoto et al. 2010).Therefore, the transition mode of cell fate must differ when using different types of explant tissues, such as root segments, cotyledons, and leaf blades.At the cellular level, this necessitates distinct reprogramming patterns mediated by PRC2, resembling the creation of "newspaper blackout poems," as recently proposed (Bieluszewski et al. 2021).
Upon wounding, the immediate activation of many genes that regulate callus formation was observed, such as WOUND-INDUCED DEDIFFERENTIATION 3 (WIND3), ETHYLENE RESPONSE FACTOR 115 (ERF115), PLETHORA 3 (PLT3), PLT4, and PLT5 (Ikeuchi et al. 2017).Since a number of these genes were previously silenced and marked with H3K27me3 before wounding, their rapid activation upon wounding indicates that the active demethylation of H3K27me3 is required (Ikeuchi et al. 2017).Supporting this concept, JMJ30 was demonstrated to promote callus proliferation by binding to and activating a subset of LATERAL ORGAN BOUNDARIES DOMAIN (LBD) genes (Lee et al. 2018b).Interestingly, in this context, JMJ30 was found to primarily demethylate trimethylation at lysine 9 of histone H3 (H3K9me3) at LBD16 and LBD29, rather than H3K27me3 (Lee et al. 2018b).Therefore, further research is needed to determine whether JMJ30-mediated H3K27me3 demethylation also contributes to its role in regulating plant regeneration.Alternatively, JMJ30 may exhibit distinct substrate specificity in a locus-or cell type-dependent manner, as recently shown for JMJ16 (Wang et al. 2020b).
Apart from JMJ30, the roles of other H3K27me3 demethylases in regulating plant regeneration remain largely uncharacterized.Indirect evidence suggests that these additional H3K27me3 demethylases may also play a part in regulating plant regeneration.For instance, CUP-SHAPED COTYLEDON 1 (CUC1), a gene with a crucial role in regulating shoot formation, has been identified as a target of REF6 (Daimon et al. 2003;Cui et al. 2016;Cao et al. 2023).In rice, WUSCHEL-RELATED HOMEOBOX 11 (WOX11) recruits the H3K27me3 demethylase JMJ705 to genes involved in shoot development, resulting in reduced H3K27me3 accumulation and increased gene expression (Cheng et al. 2018).Therefore, further research is necessary to gain a comprehensive understanding of the roles of other H3K27me3 demethylases in plant regeneration.

Passive dilution of H3K27me3 during plant regeneration
In addition to active H3K27me3 demethylation, the passive dilution of H3K27me3 marks through mitotic cell division may also contribute to H3K27me3 reprogramming during plant regeneration (Hugues et al. 2020;Stewart-Morgan et al. 2020).The early phase of plant regeneration is characterized by rapid cell division and the upregulation of many cell division marker genes, including CYCB1;1, CYCLIN D3;1-3 (CYCD3;1-3), and E2F TRANSCRIPTION FACTOR 3a (E2Fa; He et al. 2012;Ikeuchi et al. 2017;Liu et al. 2018a).Treatment with olomoucine, a cyclindependent kinase inhibitor, not only inhibited cell cycle progression but also delayed the reduction of H3K27me3 at the WUSCHEL (WUS) locus and the induction of WUS expression (Zhang et al. 2017).
Moreover, genome-wide profiling across multiple stages of regeneration revealed a gradual decrease in H3K27me3 modification and sequential changes in the expression of many key morphogenic genes in both Arabidopsis and wheat (Triticum aestivum; Figs. 1 and 2; Wu et al. 2022;Liu et al. 2023a).These findings may help explain why directly placing explants on shoot induction medium (SIM) is ineffective in triggering shoot formation.Perhaps the dilution of H3K27me3 via cell divisions on CIM is required to derepress critical regeneration-related genes for the acquisition of pluripotency.

The role of H3K27me3 in mediating environmental signaling
During the process of plant regeneration, PRC2 mediates H3K27me3 deposition in response to not only developmental signals but also environmental stimuli, such as nutrient and light conditions.The impact of carbon source type and concentration on plant regeneration efficiency has long been recognized (Yaseen et al. 2013).However, the exact underlying mechanism has remained elusive.Recent research has shed light on this matter by revealing the role of PRC2 in mediating the influence of environmental nutrient signals on plant organogenesis (Li et al. 2017;Ye et al. 2022).This mechanism involves the evolutionarily conserved target of rapamycin (TOR) kinase signaling pathway, as FIE, a key component of Arabidopsis PRC2, serves as a direct target of TOR kinase.The phosphorylation status of FIE in response to environmental sugar signals dictates the subcellular localization of FIE and the global pattern of H3K27me3 modification.Consequently, this regulatory process influences stem cell fate and cell proliferation at the shoot and root meristems (Li et al. 2017;Ye et al. 2022).
In addition to carbon nutrients, the TOR-PRC2 pathway also transmits light signals, thereby activating cell proliferation at the shoot apical meristem (Li et al. 2017).Similarly, spatial transcriptome profiling of tomato (Solanum lycopersicum) callus during shoot regeneration revealed increased expression of SlTORs and genes related to sucrose metabolism in the shoot primordium (Song et al. 2023).Alternatively, light signals may directly regulate plant regeneration via PRC2.Notably, phytochrome B interacts with the PRC2 component VERNALIZATION INSENSITIVE 3-LIKE1/VRN5 (VIL1/VRN5) in a light-dependent manner (Kim et al. 2021).This interaction plays a role in fine-tuning H3K27me3 modification and the expression of photomorphogenesis genes (Kim et al. 2021).Such a mechanism may also be involved in regulating the light responses of genes associated with plant regeneration (Nameth et al. 2013;Chen et al. 2016;Blair Nameth et al. 2018;Dai et al. 2022).

Effects of H3K4me3 deposition
During plant regeneration, changes in the expression of key developmental genes, such as LEAFY COTYLEDON 2 (LEC2), BABY BOOM (BBM), and WOX5, coincide with dynamic changes in H3K4me3 modification, pointing to the involvement of H3K4me3 histone modification in regulating the transition of cell identities during plant regeneration (Liu et al. 2023a).Notably, the rise in H3K4me3 modification across pivotal regeneration-promoting genes often coincides with a decrease in H3K27me3 modification (Zhao et al. 2020;Liu et al. 2023a).This suggests that the function of PRC2 and Trithorax (Trx) chromatin modifiers is dynamically regulated to ensure proper cell fate transition.In Arabidopsis, H3K4 methylation is catalyzed by 7 proteins that share homology with Drosophila Trx, including ARABIDOPSIS TRITHORAX (ATX) 1 to 5, ARABIDOPSIS TRITHORAX-RELATED 3 (ATXR3), and ATXR7 (Zhou et al. 2020).ATXR3, also known as SET DOMAIN GROUP 2 (SDG2), serves as the primary H3K4me3 methyltransferase (Guo et al. 2010).A deficiency in ATXR3 results in reduced H3K4me3 deposition at nearly half of all H3K4me3 sites in the genome, leading to pleiotropic developmental defects in Arabidopsis (Guo et al. 2010; Yao Genome-wide profiling revealed that ATX4 is responsible for H3K4me3 deposition and the active expression of shoot identity genes in Arabidopsis (Lee et al. 2019a).As a result, the deficiency in ATX4 function in leaf explants led to increased callus formation but a reduced capacity for shoot regeneration (Lee et al. 2019a).This outcome corresponds with the observation that Arabidopsis clf-50 swn-1 mutants, which are characterized by deficiencies in the silencing of leaf identity genes, fail to initiate callus formation (He et al. 2012).In addition to ATX4, other H3K4me3 methyltransferases may also contribute to the regulation of plant regeneration.For instance, ATXR3 is essential for maintaining the root stem cell niche (Yao et al. 2013).

Effects of histone acetylation
Acetylation of histone lysine residues neutralizes the positive charges of nucleosomes, reducing their interactions with negatively charged DNA, thereby enhancing the accessibility of transcription factors and the transcriptional machinery (Kumar et al. 2021).Plants contain 4 distinct families of histone acetyltransferases, including the General Control Non-depressible 5 (GCN5) family acetyltransferases (HAG), the Moz-Ybf2/Sas3-Sas2/Tip60 (MYST) family acetyltransferases (HAM), the cAMP-Responsive Element-Binding Protein-binding protein (CBP) family acetyltransferases (HAC), and the TATA-binding Protein-Associated Factor (TAFII250) family acetyltransferases (HAF; Pandey et al. 2002).Histone acetylation is typically linked with gene activation and plays a pivotal role in regulating various biological processes in plants, including the response to wounding and plant regeneration (Kim et al. 2018;Rymen et al. 2019).Notably, genes that respond to wounding and exhibit early induction are often predeposited with H3K9/14Ac modifications (Rymen et al. 2019).
Studies have demonstrated the importance of histone acetylation in plant regeneration.Both Arabidopsis and rice hag1 mutants exhibit abnormalities in the root stem cell niche (Kornet and Scheres 2009;Zhou et al. 2017).In line with this observation, the formation of callus was compromised in Arabidopsis hypocotyl explants lacking HAG1 and HAG3 (Rymen et al. 2019).Interestingly, the Arabidopsis hag1 mutant displayed accelerated callus development on CIM (Kim et al. 2018).The differing effects of HAG1 on callus formation suggest that HAG1-mediated histone acetylation plays distinct roles in regulating plant responses to wounding and plant hormones, which might be orchestrated by different downstream target genes (Kim et al. 2018).Alternatively, these varying effects might be related to the different types of assays used and tissues examined.While 1 study focused on spontaneous callus formation at the wounding site in hypocotyls (Rymen et al. 2019), the other study analyzed callus formation from root explants placed on CIM (Kim et al. 2018).It is also conceivable that cell proliferation and pluripotency are differentially influenced by HAG1.In support of this notion, even though the hag1 mutant showed accelerated callus formation on CIM, this led to severe deficiencies in shoot regeneration, indicating that enhanced cell proliferation did not result in the efficient acquisition of cell pluripotency (Kim et al. 2018).Meanwhile, mutant plants deficient in HAG2, HAM1, and HAM3 did not exhibit substantially impaired plant regeneration (Rymen et al. 2019), which could be partially attributed to the potential functional redundancy of these histone acetyltransferases.To address this matter, experiments using chemical inhibitors targeting different histone acetyltransferase families suggested that the GNAT-MYST family, but not CBP family histone acetyltransferases, functions in wounding-induced callus formation (Rymen et al. 2019).

The role of histone deacetylation
The Arabidopsis genome encodes 18 histone deacetylases (HDACs) categorized into 3 subfamilies: the Reduced Potassium Dependency 3/Histone Deacetylase 1 (RPD3/ HDA1) family, the Silent Information Regulator 2 (SIR2) family, and the plant-specific Histone Deacetylase 2 (HD2) family (Pandey et al. 2002).Several studies have indicated that callus formation is compromised in Arabidopsis mutants such as hda9 and hdt1, as well as rice Oshda710 (HDA1 homolog) mutants (Lee et al. 2016;Zhang et al. 2020).A different investigation evaluated de novo shoot formation in a group of Arabidopsis HDAC mutants, including hda6, hda9, hda10, hda14, hda17, hda18, hda19, and hd2b (Temman et al. 2023).Among these mutants, only hda19 exhibited reduced shoot formation, and further analysis revealed that hyperacetylation and overexpression of ENHANCER OF SHOOT REGENERATION 1 (ESR1) and CUC2 were responsible for this effect in the hda19 mutant (Temman et al. 2023).Interestingly, HDA19 was found to negatively regulate somatic embryogenesis (Morończyk et al. 2022).In this case, the observed phenotypic defects in hda19 mutant plants were likely linked to the dysregulation of embryogenesisrelated transcription factors, including LEC1, LEC2, and BBM (Morończyk et al. 2022).These findings underscore the complex and distinct roles of various HDAC proteins in plant regeneration.

The effects of HDAC inhibitors
In addition to genetic studies, the application of chemical inhibitors targeting HDACs has also contributed to our understanding of the role of histone acetylation in regulating plant regeneration.Among these inhibitors, Trichostatin A (TSA) is one of the most commonly used.TSA treatment induces somatic embryogenesis by upregulating auxin biosynthesis (Wójcikowska et al. 2018), a phenomenon also observed in the hda19 mutant (Morończyk et al. 2022).Furthermore, TSA treatment affects the cytokinin response (Furuta et al. 2011), leading to inhibited auxin-induced root primordium formation from hypocotyl segments and redirecting their development toward callus formation (Furuta et al. 2011).However, in other instances, TSA had inhibitory effects on callus formation in Arabidopsis leaf explants and mature rice embryos (Lee et al. 2016;Zhang et al. 2020).
Several factors may explain the contrasting effects of TSA treatment on callus formation.As discussed earlier for PRC2, the routes of the cell fate transition may depend on the types of cells and explants involved, which can be differentially affected by TSA treatment.In support of this idea, while TSA impaired callus formation in mature rice embryos, it did not influence callus formation when using rice roots and shoots as explants (Zhang et al. 2020).Furthermore, concentration-dependent effects of TSA on plant regeneration have been observed in rice and wheat (Bie et al. 2020;Zhang et al. 2020).While 0.5 μM TSA treatment promoted embryonic callus formation and shoot induction in wheat, increasing the concentration to 2.5 μM had an inhibitory effect (Bie et al. 2020).A similar concentrationdependent effect on plant regeneration has also been noted for nicotinamide, another HDAC inhibitor (Wang et al. 2023).In addition, TSA treatment is known to alter endogenous phytohormone production and perception (Furuta et al. 2011;Wójcikowska et al. 2018).Consequently, variations in the levels of endogenous phytohormones resulting from TSA treatment may lead to differential responses of explants to exogenously applied phytohormones.

Epigenetic regulation of key morphogenic genes
Plant regeneration is governed by a complex gene network that encompasses multiple morphogenic transcription factors.For instance, during the initial phase of organogenesis on CIM, auxin signals stimulate callus formation via the AUXIN RESPONSE FACTOR (ARF)-mediated activation of LBDs (Fan et al. 2012).Auxin also promotes the establishment of cellular pluripotency through 2 parallel pathways: 1 involving WOX11 and LBD16 (Liu et al. 2018b) and the other mediated by PLTs and CUC2 (Kareem et al. 2015).Following the establishment of cell pluripotency, cytokinin-induced shoot formation on SIM is facilitated by the ARABIDOPSIS RESPONSE REGULATOR (ARR)-mediated activation of WUS expression (Meng et al. 2017).Wounding-induced shoot regeneration is dependent on WIND1, which activates ESR1 expression (Iwase et al. 2017).During somatic embryogenesis, a positive feedback network has been proposed, which involves multiple morphogenic regulators, including BBM, AGAMOUS-LIKE15 (AGL15), LEC1, and LEC2 (Zheng et al. 2009;Horstman et al. 2017).These embryonic regulators stimulate the expression of YUCCAs (YUCs; Stone et al. 2008) and INDOLEACETIC ACID-INDUCED PROTEIN30 (IAA30; Zheng et al. 2009) to modulate auxin biosynthesis and signaling.For a more comprehensive summary of the gene network underlying plant regeneration, we recommend excellent reviews from other authors (Ikeuchi et al. 2019;Ince and Sugimoto 2023;Liu et al. 2023b).We have chosen several key morphogenic gene families as examples to elucidate potential modes and patterns of epigenetic regulation observed during plant regeneration (Fig. 3), which we describe below.

WUS
WUS is a homeodomain transcription factor crucial for various types of plant regeneration, including somatic embryogenesis, mesophyll regeneration, and de novo shoot organogenesis (Jha et al. 2020;Xu et al. 2021).During plant regeneration, the expression of WUS undergoes intricate epigenetic regulation involving both DNA methylation and histone modification.In Arabidopsis root explants, the promoter sequences of WUS are methylated in both the CG and CHG contexts (Li et al. 2011;Shemer et al. 2015).Incubating root explants on CIM triggers DNA demethylation, which enhances WUS activation on SIM (Li et al. 2011;Shemer et al. 2015).The met1 and drm1 drm2 cmt3 mutants show the loss or reduction of DNA methylation in the WUS promoter, enhancing WUS expression and shoot formation (Shemer et al. 2015;Liu et al. 2018a).Upon shoot induction, the expression of WUS is activated via 2 consecutive steps: first, the repressive histone mark H3K27me3 is removed from WUS; second, WUS is activated through the binding of the B-type ARR-HD-ZIP III complex (Zhang et al. 2017).Intriguingly, B-type ARR1 also interacts with ATXR2 and temporally promotes the expression of A-type ARR5/7 via H3K36me3 deposition, which represses cytokinin signaling and WUS expression prior to shoot formation to ensure a balanced cell fate transition (Lee et al. 2021).Extensive epigenetic reprogramming involving other histone modifications was also shown to regulate the expression of WUS.Mutations of the H3K9 methyltransferase gene KRYPTONITE (KYP) and the H3K4 demethylase gene JMJ14 greatly enhance WUS transcription, while mutations of the histone acetyltransferase gene HAC1 reduce WUS transcription (Li et al. 2011).

WOX genes
The WOX proteins constitute a group of plant-specific homeodomain transcription factors.Several members of this family have been demonstrated to play essential roles in various biological processes related to plant organogenesis, such as meristem maintenance, cell proliferation, and the regulation of differentiation (Tvorogova et al. 2021;Wan et al. 2023).
In rice, OsWOX11 recruits the ADA2-GCN5 histone acetyltransferase complex and the H3K27me3 demethylase JMJ705 to activate downstream target genes associated with root and shoot development (Zhou et al. 2017;Cheng et al. 2018).In Arabidopsis, the downstream targets of WOX11 and WOX12 include LBD16 and LBD29 (Liu et al. 2014).As ATXR2 also targets LBD16 and LBD29 for trimethylation at lysine 36 of histone H3 (H3K36me3) deposition to promote their expression during callus formation (Lee et al. 2017(Lee et al. , 2018b)), it has been suggested that WOX11/12 may form a complex with ATXR2 for the transcriptional activation of LBD genes (Lee et al. 2018a).Conversely, the expression of WOX11 is itself epigenetically regulated.During callus formation, the deposition of the histone 3 variant H3.15 at the WOX11 locus facilitates the removal of H3K27me1/3 and leads to the transcriptional derepression of WOX11 (Yan et al. 2020).
In addition to the activation of LBD16 and LBD29, the expression WOX5 is also upregulated by WOX11/2 via the direct binding of WOX11/12 to its promoter (Liu et al. 2014;Hu and Xu 2016).This transcriptional activation of WOX5 is required for callus to acquire the competence needed for de novo shoot regeneration (Kim et al. 2018).During callus formation, the expression of WOX5 is epigenetically induced by SDG8 through H3K36me3 modification (Ma et al. 2022) and by HAG1 through histone H3 acetylation (H3Ac) modification (Kim et al. 2018).Similarly, the expression of SHORT-ROOT (SHR) and SCARECROW (SCR), encoding upstream positive regulators of WOX5, is also regulated by histone acetylation mediated by HAG1 and elongator complex subunit 2 (ELP2; Jia et al. 2015;Kim et al. 2018).As expected, the shoot regeneration defect of the hag1 mutant was successfully rescued by the ectopic activation of WOX5 and SCR (Kim et al. 2018).The expression of WOX5 is also responsible for specifying the stem cell niche during root development, which involves developmental regulation via H3K27me3 and H3K4me3 modifications (Takatsuka and Umeda 2015;Zhai et al. 2020).

PLTs
The PLT transcription factors are members of the AINTEGUMENTA-LIKE (AIL) APETALA2/ETHYLENE RESPONSIVE FACTOR (AP2/ERF), many of which are known for their roles in regulating stem cell identity in embryonic and meristem tissues (Galinha et al. 2007;Horstman et al. 2017;Xu et al. 2023a).PLT genes are expressed in actively dividing tissues and contribute to various regenerative responses (Santuari et al. 2016;Ikeuchi et al. 2019).Overexpressing Arabidopsis PLT genes induced pluripotency and stimulated cell proliferation in somatic tissues (Nole-Wilson et al. 2005;Galinha et al. 2007;Tsuwamoto et al. 2010;Krizek and Eaddy 2012).
Several PLT family genes were shown to be epigenetically regulated.The PRC1 homolog BMI1A/B proteins are responsible for mediating H2A monoubiquitination in Arabidopsis.This mark, in conjunction with PRC2-mediated H3K27 trimethylation, has been demonstrated to repress the expression of BBM (PLT4; Bratzel et al. 2010;Mozgová et al. 2017).In addition, the misregulated expression of BBM was observed in Arabidopsis explants lacking the histone deacetylase HD-TUINS PROTEIN 1 (HDT1), HDT4, and HDA19 during callus formation and somatic embryogenesis (Lee et al. 2016;Morończyk et al. 2022).Similarly, during the regeneration of Coffea canephora, the transition of explants from proembryogenic masses to globular embryos was found to be epigenetically regulated by DNA methylation and H3K27me3 histone modification of BBM (Nic-Can et al. 2013).Treatment with the DNA methylation inhibitor 5-azacytidine inhibited somatic embryogenesis by reducing the expression of BBM (Nic-Can et al. 2013).Furthermore, the level of DNA methylation at the BBM locus was influenced by treatment with phytohormones, such as 2,4-D, in a dosagedependent manner (Grzybkowska et al. 2020).
Regarding other PLT genes, H3Ac at the Arabidopsis PLT1 and PLT2 genes is regulated by HAG1.Explants deficient in HAG1, regardless of their organ source, exhibited much reduced expression of PLT1 and PLT2, along with impaired competency for shoot formation (Kim et al. 2018).Consistent with this finding, the acetylation and transcriptional activation of PLTs are essential for plant regeneration from Arabidopsis leaf mesophyll protoplasts by activating auxin biosynthesis genes, including YUC1 (Sakamoto et al. 2022).Apart from histone acetylation, other epigenetic factors might also contribute to the regulation of PLT expression, including changes in chromatin accessibility and H3K27me3 modification (Nakamura et al. 2020;Wang et al. 2020a).

LBD genes
LBD proteins constitute a family of plant-specific transcription factors distinguished by their highly conserved N-terminal LATERAL ORGAN BOUNDARIES (LOB) domain (Husbands et al. 2007;Majer and Hochholdinger 2011).The functions of LBDs include defining lateral organ boundaries, xylem differentiation, secondary growth in woody plants, and so on (Uchida et al. 2007;Soyano et al. 2008;Yordanov et al. 2010).
LBD proteins, such as LBD16, LBD17, LBD18, and LBD29, operate downstream of auxin signaling and serve as crucial promoters of callus formation (Fan et al. 2012;Xu et al. 2018).In response to the auxin signal, the expression of LBDs is directly activated by the sequential action of the ARF-JMJ30 complex, which mediates H3K9me3 removal, and ATXR2, which mediates H3K36me3 deposition (Lee et al. 2017(Lee et al. , 2018b)).Similar to the mechanism observed for WOX11, the induction of expression and incorporation of the histone 3 variant H3.15 upon wounding led to a reduction in H3K27me3 deposition and enhanced transcription of LBD16, LBD18, and LBD29, promoting callus formation (Yan et al. 2020).A comparable mechanism involving LBD activation via reduced H3K27me3 deposition has also been reported in peach during the leaf-to-callus transition (Zheng et al. 2022).A recent study revealed that strong expression of TaLBD17 in wheat is associated with increased chromatin accessibility, the removal of H3K27me3 modification, and the increase in H3K4me3 modification during the late stage of callus formation from immature embryos (Liu et al. 2023a).

WINDs
WIND1 and its homologs are AP2/ERF transcription factors that promote cell dedifferentiation and proliferation in response to wounding (Iwase et al. 2011a).The ectopic expression of WINDs can induce spontaneous callus formation even in the absence of exogenous phytohormones (Iwase et al. 2011a(Iwase et al. , 2011b)).
Damage triggers the reprogramming of differentiated cells into stem cells (Gu et al. 2020).The rapid induction of WIND1 following wounding was found to be orchestrated by the high levels of H3K9/14ac and H3K4me3 that were predeposited at the WIND1 locus prior to tissue damage (Rymen et al. 2019).Further investigation revealed that the GNAT-MYST family of histone acetyltransferases, such as HAG1/GCN5 and HAG3, but not the CBP family of histone acetyltransferases, is involved in wounding-induced epigenetic reprogramming and the promotion of callus formation (Rymen et al. 2019).In addition to histone acetylation, the deposition of H3K27me3 by PRC2 also regulates the expression of WIND3.A deficiency in PRC2 function, as observed in the Arabidopsis emf2 vrn2 mutant, results in the activation of WIND family genes and dedifferentiation of terminally differentiated root hair cells (Ikeuchi et al. 2015).Other epigenetic modifications, such as H3K36me3 deposited by ASH1-RELATED 3 (ASHR3), also activate the expression of WIND3 (Lee et al. 2020).which represents a pivotal and rate-limiting step in the biosynthesis of this major auxin in plants (Won et al. 2011).In Arabidopsis, the YUC gene family encompasses 11 homologous genes, many of which are subjected to epigenetic regulation during plant regeneration (He et al. 2012;Gyula et al. 2018;Lambolez et al. 2022).

YUCs
During the transformation from leaf tissue to callus, the repressive histone mark H3K27me3 at the YUC4 locus diminishes, leading to the upregulation of YUC4 expression (He et al. 2012).Similarly, when de novo root organogenesis occurs from detached leaf tissue, the activation of both YUC1 and YUC4 expression is associated with the removal of H3K27me3 (Chen et al. 2016).Besides YUC genes, the expression of another auxin biosynthesis-related gene, ANTHRANILATE SYNTHASE α1 (ASA1), was shown to be epigenetically regulated by SDG8-mediated H3K36me3 deposition during de novo root regeneration from a detached leaf (Zhang et al. 2019).
In addition to H3K27me3, YUC genes can also be epigenetically regulated by H2A.Z during plant regeneration.Elevated temperatures (such as 27 °C) enhance shoot regeneration in Arabidopsis (Lambolez et al. 2022).This effect is partially due to the induction of YUC4 expression by warmer temperatures, which is accompanied by a loss of H2A.Z modification (Lambolez et al. 2022).The loss of H2A.Z in hta9 hta11 mutants promotes de novo shoot organogenesis and causes the dysregulated expression of YUC1 and YUC4 in a temperaturedependent manner (Lambolez et al. 2022).
Furthermore, the participation of YUC genes in diverse developmental processes is regulated by various epigenetic mechanisms, including histone acetylation, RdDM, and nucleosome remodeling (Gyula et al. 2018;Peng et al. 2018;Yamaguchi et al. 2018).However, whether similar mechanisms also participate in regulating YUC gene expression during plant regeneration remains a subject for further investigation.

Conclusions and future perspectives
The impact of epigenetic changes on plant regeneration is complex and cannot be attributed to a single locus or a single type of epigenetic modification.Instead, it involves intricate gene regulatory networks with dynamic interactions and the balance among various epigenetic modifications (Figs. 1 and  2).Moreover, the vast diversity of plant species, the sources of explants, and the experimental treatments add a layer of complexity.Using the examples provided above, our aim was to illustrate scenarios in which epigenetic regulation is intricately integrated with the dynamic reprogramming of gene expression during plant regeneration.As specific genes are often regulated by multiple types of epigenetic modifications, the exploration of additional epigenetic regulatory circuits presents new avenues for future research (see OUTSTANDING QUESTIONS).Nevertheless, questions regarding the causal relationship between changes in gene expression and epigenetic modifications often lack simple, straightforward answers due to the presence of reciprocal feedback between transcription status and chromatin state.
Nonetheless, the continuous advancements in molecular technology hold the promise of enhancing our understanding of the role of epigenetic regulation in plant regeneration.Key developments on the horizon include gaining insights into cell type-specific epigenetic features and their impact on cell fate transitions using single-cell sequencing platforms.Additionally, we anticipate the development of advanced epigenetic engineering tools for precise regulation of morphogenic genes to break the genotype-dependent barriers of plant regeneration and transformation.These innovations are set to illuminate the intricate realm of plant regeneration.Simultaneously, plant regeneration will keep presenting us with a distinctive and ideal opportunity to develop a profound understanding of epigenetic regulatory mechanisms.

Figure 1 .
Figure 1.Dot plots depicting the expression levels of Arabidopsis genes and the density of H3K4me3, H3K27me3, and H3K27ac within 1.5 kb upstream and downstream of the genes.The size of the dots represents the absolute expression level of the genes or the level of each epigenetic modification on corresponding gene, normalized by Z-score.As described in the original study, hypocotyls of 7-d-old Arabidopsis seedlings were cut, followed by 7 d of CIM treatment and 24 d of SIM treatment.Data were collected at 3 stages of regeneration: CIM7, SIM3, and SIM8.The information was retrieved from the Beijing Institute of Genomics Data Center (accession number PRJCA005872; Wu et al. 2022).

Figure 2 .
Figure 2. Dot plots display gene expression levels and the density of H3K4me3, H3K27me3, and H3K27ac in wheat variety Fielder at different stages of regeneration.The size of the dots represent the absolute expression level of the genes or the level of each epigenetic modification within 1.5 kb upstream and downstream of the corresponding genes, normalized by Z-score.The scutellum of the immature embryo, after excision of the embryo axis without tissue culture, served as the control.Wheat materials were sampled from the surface of the callus at 3, 6, and 9 d after induction (DAI).The analysis followed the procedures described in Fig. 1.Data were downloaded from the National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA008502 and CRA010204; Liu et al. 2023a).

Figure 3 .
Figure 3.A schematic diagram of the key genes and epigenetic regulators involved in plant regeneration.Orange rounded rectangles indicate major morphogenic genes that regulate different types and stages of regeneration.Ellipses represent various types of epigenetic modifiers, each color signifying a specific type of epigenetic modification.Lines with arrowheads indicate positive regulation, while lines with bars indicate negative regulation.The solid and dotted lines represent direct and indirect regulation, respectively.