Climate change includes the instantaneous and rapid alterations of numerous important environmental parameters that regulate the dynamics of the ecosystem, such as temperature and precipitation. These rapid changes may have direct and indirect effects on plants via alterations in plant secondary metabolism (Christensen and Christensen, 2007; Bita and Gerats, 2013; Rienth et al., 2021). Lignin is one of the most important secondary metabolites and is essential to the formation of cell walls. Changes in lignin biosynthesis can influence plant fitness and their adaptation to environment (Liu et al., 2018). For example, a lignin-deficient mutant of Arabidopsis was found to have a significantly lower seed germination rate compared to the wild type, hypothetically due to an increase of sensitivity to unfavorable environments (Liang et al., 2006). Moreover, previous studies have shown that the lignin biosynthesis is highly responsive to environmental changes (Moura et al., 2010; Cesarino, 2019; Gu et al., 2020). Under drought and salt stresses, increased lignin can reduce plant cell wall water penetration and transpiration, which helps to maintain cell osmotic balance and protect membrane integrity (Monties and Fukushima, 2001; Chun et al., 2019). When plants are exposed to an excess of light, an increased level of lignin content is observed, which may prevent the accumulation of hydrogen peroxide and thus protect the cellular damage from the formation of reactive oxygen species (ROS) (Chen et al., 2002). Similarly, increased lignin content can reduce damage from ultraviolet (UV)-B radiation in plants (Rozema et al., 1997). In addition, changes in lignin content are also found to be related to cold acclimation (Wei et al., 2006; Zeng et al., 2016). One possible reason is that the altered ratio of syringyl (S) to guaiacyl (G) subunits affects cell wall rigidity and water permeability, thus increasing plant resistance to cold temperatures (Zhao et al., 2013). Overall, these previous findings indicate that lignin biosynthesis is tightly associated with environmental conditions. Deciphering the molecular basis underlying this association is thus crucial to understanding plants adaptation and evolution and to promoting plants conservation and breeding.

DNA methylation (5-methylcytosine) is an important regulatory mechanism by which plants respond to environmental fluctuations (McCaw et al., 2020; Sun et al., 2022). In plants, DNA methylation occurs in three sequence contexts: CG, CHG, and CHH, where H is any nucleotide except G. Maintenance DNA methylation in symmetrical CG and CHG contexts is executed by DNA METHYLTRANSFERASE 1 (MET1) and CHROMOMETHYLASE 3 (CMT3), respectively. Asymmetric CHH methylation is maintained by CHROMOMETHYLASE 2 (CMT2) and the 24nt-siRNA involved RNA-directed DNA methylation (RdDM) pathway (Law and Jacobsen, 2010; Zemach et al., 2013; Stroud et al., 2014). Changes in DNA methylation in response to altered environmental conditions may facilitate plant adaptation via their effects on gene expressions. For example, in B. gymnorhiza, genome-wide CHH hypermethylation induced by increased environmental salinity can enhance several salt-responsive genes expression, such as Bg70 and BURP Domain Containing protein 3 (BgBDC3) (Miryeganeh et al., 2022). In maize, genome-wide demethylation enables the activation of stress-responsive genes under cold stresses (Steward et al., 2002). In natural populations, genome-wide CHH methylation variations are associated with the changes of temperature seasonality in Arabidopsis accessions (Shen et al., 2014). CG methylation variations, which affects dehydration-responsive genes in oak, are found to be significantly associated with the mean maximum temperature of the warmest month (Gugger et al., 2016). Together, these results provide evidence that epigenetic variation can be an important molecular mechanism that influences plant adaptations to the changing climates.

Since DNA methylation and lignin production can both be adaptive responses to external environments, we hypothesized that DNA methylation is likely to respond to environmental fluctuations by participating in lignin regulation. This was based on two lines of evidence. First, previous studies have shown that the variations of DNA methylation are, at the same time, associated with the biosynthesis of secondary metabolism and with the environmental gradients. For example, in birch, variations in methylation of the bHLH9 promoter are found to affect the biosynthesis of betulin and be significantly associated with local temperatures and precipitations (Wang et al., 2021). In wheat, DNA methylation of genes involved in carbon metabolism or fatty acid metabolism contributes to male spikelet sterility in a temperature-sensitive genic male sterile (TGMS) line, BS366, in response to cold (Liu et al., 2021). Further, DNA methylation levels involved in metabolic processes are significantly associated with high temperatures in rice (Li et al., 2022). These results provide the evidence that the regulatory role of DNA methylation in plant adaptation can be mediated by its effects on secondary metabolisms.

Second, lignin biosynthesis is a very complex network, involving many enzymes and transcription factors (TFs). Lignin monomer synthesis is mainly composed of phenylpropane metabolic pathways and downstream synthases (Xie et al., 2018). The classic phenylpropane metabolic pathways are primarily composed of phenylalanine ammonia-lyase (PAL), cinnamate-4-hydroxylase (C4H), and 4-coumarate: CoA ligase (4CL), contributing to the conversion of phenylalanine into p-coumaroyl-CoA (Fraser and Chapple, 2011; Wang et al., 2018). In addition to these three enzymes, some other phenylpropanoid biosynthetic enzymes are responsible for the sequential conversion of intermediate/downstream products of the phenylpropane metabolic pathway into lignin monomers (Raes et al., 2003; Yoon et al., 2015). The polymerization of lignin monomers is then carried out by peroxidase (POD) and laccase (LAC) (Liu et al., 2011; Alejandro et al., 2012). Moreover, the transcriptional regulation of lignin biosynthesis genes is mainly well-known for the first-layer NAC-domain TFs (Kubo et al., 2005; Zhong et al., 2007; Fang et al., 2020), and the second-layer MYB-domain TFs (Ko et al., 2014; Nakano et al., 2015; Xiao et al., 2021). This complexity can offer more chances in which DNA methylation could be involved. As an example, we recently found that a transcription factor, CmMYB6, that regulates another downstream branch of phenylpropane metabolic pathways, can have methylation variations in the promoter which regulates anthocyanins biosynthesis (Tang et al., 2022). However, whether DNA methylation is involved in lignin biosynthesis and this relation, if exist, would contribute to environmental adaptation in natural conditions are still unclear.

In this study, to investigate the relation of DNA methylation with lignin and its role in plant adaptation, Chinese birch (Betula platyphylla Suk.), which is widely distributed in the northeast China, was selected as the target species (Geng et al., 2022). Lignin content, methylome and transcriptome profiles of genes involved in the lignin biosynthesis pathway were assessed to uncover a possible molecular regulatory mechanism and to test their relations with climatic factors. We found that the lignin content of natural birch was significantly negatively correlated with local temperature ranges. Nearly all critical lignin biosynthesis genes were expressed at significantly higher levels in birch trees grown at low temperature ranges than at high temperature ranges. This differential expression was likely regulated by two master switches of lignin activators, BpNST1/2 and BpSND1, whose expressions were likely regulated by differential methylation of the promoters between birches grown in different temperature ranges. In summary, our results support the hypothesis that DNA methylation can be involved in responses to climate changes via its regulation on plant secondary metabolites and can further provide useful information about how to manage the white birch to efficiently utilize lignin from natural birch sources.

Lignin is a group of phenolic polymers which is abundant in the woody tissues of vascular plants and has long been considered to play an important role in plant growth and development (Liljegren et al., 2000; Bonawitz et al., 2014; Peng et al., 2014). Recently, a few studies have shown that the lignin biosynthesis is related to plant adaptations to external stresses, such as metal exposure, drought, salinity, and temperature stresses (Mao et al., 2004; Moura-Sobczak et al., 2011; Liu et al., 2018). For example, Gindl et al. (2000) found that lignin deposition in the late terminal trachea of Norway spruce is positively correlated with temperature from early September to late October. Yun et al. (2013) found that high temperature stress can induce lignin deposition in Satsuma mandarin pericarp. Similarly, at 30°C, the accumulation of lignin in the MtCAD1 mutant of Medicago was inhibited and thus the vascular system was damaged, which led to impaired transpiration and overheating and thus the stalled growth (Zhao et al., 2013). In addition, an increased lignin production during cold acclimation has also been observed in several species, suggesting a physiological mechanism of freezing tolerance in trees (Griffith et al., 1985; Marzanna et al., 1999; Wei et al., 2006; Vitasse et al., 2014). Besides environmental factors, tree age is also shown to affect lignin content (Durkovic et al., 2012). However, the general trend is not consistent and somehow controversy with increase (Jahan and Mun, 2005; Rencoret et al., 2011), decrease (Stolarski et al., 2011; Healey et al., 2016) or no responses (Lachowicz et al., 2019) of lignin in wood being reported with tree aging. In this study, we found that the impact of the age of trees on lignin content was not significant ( Supplementary Figure S2 ). This inconsistency among various studies may be ascribed to other factors, including the species, whether the wood came from a plantation or a natural forest, forest habitat type, the part of the tree from which samples were taken, the time of sampling, etc. (Lachowicz et al., 2019). Nevertheless, after controlling for the potential influence of tree age, we found that the lignin content of natural birch was significantly negatively correlated with the temperature ranges (bio4 and bio7) ( Figure 1 and Supplementary Table S6 ). This suggested that temperature regimes can play a significant role in lignin biosynthesis, and that changes in lignin content may contribute to plants adaptations to fluctuations in temperature.

Uncovering the basis underlying the associations between traits and the environment is a central topic in molecular biology. At the molecular level, DNA methylation is thought to be involved in plant plasticity in the face of environmental stresses and/or climate changes, but solid evidence in natural population is few and mostly come from Arabidopsis. For example, the global DNA methylation of Arabidopsis accessions are negatively associated with the temperatures in Europe, in which the accessions with hypomethylomes are enriched in habitats with higher temperatures (Kawakatsu et al., 2016). The methylation level of NMR19, a naturally occurring epiallele controlling leaf senescence, is negatively associated with the Mean Temperature of Driest Quarter (bio9) (He et al., 2018). As one of the studies that focuses on trees, Wang et al. (2021) found that in natural birch the methylation variations of bHLH9 are associated with winter temperature and summer precipitation and affected betulin biosynthesis. Overall, these results suggested that temperature regimes are strong environmental factors that can shape the methylation patterns of key genes along their gradients. To support this, we found that DNA methylation levels of the two NAC-domain TFs (BpNST1/2 and BpSND1) were significantly correlated with the local Temperature Seasonality (bio4) and Temperature Annual Range (bio7) ( Figures 1 and 3 ). These trends were strong, especially in CHH sequence context ( Figure 4A ). This result was consistent with the temperature-dependent characteristics of CHH methylation (Bewick and Schmitz, 2015), which probably caused by the differentiation of CMT2 alleles that participates the maintenance of CHH methylation status along temperature regime gradients, such as temperature (Dubin et al., 2015) and Temperature Seasonality (Shen et al., 2014). Indeed, we found 14 genetic differences in the intron of BpCMT2 between low and high temperature ranges regions ( Figure 4C and Supplementary Table S13 ). These differences may form alleles and lead to different expression patterns of BpCMT2 or alter the splicing of mRNA. Additionally, there were significant expression differences of BpCMT3 and BpDDM1 but no and only 2 genetic variations in intron, respectively, between low and high temperature ranges regions ( Figures 4B, C and Supplementary Table S13 ). Nevertheless, other factors, such as environmentally-induced DNA methylation (Miryeganeh et al., 2022), may also contribute to the differential expressions of BpCMT2, BpDDM1 and BpCMT3.

One of the crucial consequences of DNA methylation is to affect the gene expressions and thus contribute to phenotypic variations (Zemach et al., 2013; Zhang et al., 2018). Several epigenetic variations with phenotypic consequences have been characterized as epialleles, such as plant architecture (Miura et al., 2009), flowering time (Ikeda et al., 2007), root length (Cortijo et al., 2014), and biomass (Ong-Abdullah et al., 2015). Although limited epialleles have been identified so far, transposons in which most epigenetic variations occur may be important carriers for epialleles (Slotkin and Martienssen, 2007). In birch, the content of transposons can be up to 50% with a relatively uniformly distribution in the genome, suggesting that this species is prone to have more epialleles (Salojarvi et al., 2017; Wang et al., 2021). To support this, we found that both differential methylation modifications of BpNST1/2 and BpSND1 occurred on transposons (TIR and Gypsy, respectively) within the gene promoter region and affected corresponding gene expressions ( Figures 3A, B ). Similarly, other genes, such as Bp4CL2, BpCOMT9, and BpMYB103, also differed significantly in their promoter (2 kb upstream of ATG) DNA methylation, which occurred on TEs (Copia, TIR and TIR) ( Supplementary Figure S11 and Supplementary Table S14 ). These methylation variations were also opposite to their genes’ expression pattern ( Figure 2 ). Thus, we cannot exclude the possibility that the DNA methylation modification of these genes’ promoter could be involved in regulating their expression. Yet, we believe that the differentiation of DNA methylation and the resultant gene expression patterns of BpNST1/2 and BpSND1, which are the first-layer switches in regulation of downstream lignin biosynthesis genes and MYB-domain TFs (Zhong et al., 2007; Golfier et al., 2017; Fang et al., 2020), are likely to be the main cause of the changes in these downstream genes’ expression and thus lignin biosynthesis. Further, we found that CHH and part of CHG methylation levels on transposons of BpNST1/2 and BpSND1 were negatively related to their expressions ( Figures 2B , 4A ), similar to the CMT2/3-dependent epiallele (Cnr) that regulates tomato fruit ripening in tomato (Chen et al., 2015). Likely, we recently found that CHH methylation can be involved in anthocyanin biosynthesis (another downstream branch of the phenylpropane metabolic pathway) by affecting the expression of CmMYB6 in chrysanthemum (Tang et al., 2022). We hypothesized that CHH as well as CHG methylation can form a heterochromatin by coupling with the H3K9me2, thereby inhibiting the gene expression by restricting the access of TFs to their binding sites in promoters, which may ultimately result in the changes of lignin content. This positive feedback loop between H3K9me2 and non-CG methylation has also been considered as a major contributing factor to the origins of spontaneous epialleles and that heterochromatin is a quantitative trait that influences epiallele formation (Zhang et al., 2021).

Cellulose and hemicellulose are also the target products regulated by BpNST1/2 and BpSND1. We evaluated the transcript levels of critical cellulose and hemicellulose biosynthesis genes (10 homologs of BpCesA and 11 homologs of BpCsl) in Sl, Fhs, Ame, and Alh ( Supplementary Figure S12 and Supplementary Table S15 ). The gene expression levels of BpCesA04, BpCesA07, and BpCesA08, which function mainly in the secondary plant cell wall (Desprez et al., 2007; Persson et al., 2007), were significantly decreased in Ame and Alh (trees with high temperature ranges) compared with Sl and Fhs (trees with low temperature ranges), and this expression pattern was highly consistent with the expression of BpSND1 ( Figure 2B ). Further, the expression level of BpCslB and BpCslG subfamilies, which are specific in non-grass angiosperms (Keegstra and Walton, 2006; Suzuki et al., 2006), were tightly correlated with that of BpNST1/2 and BpSND1 ( Supplementary Figure S13 and Supplementary Table S15 ). Based on these findings, we hypothesised that these two gene families (CesA and Csl) may be commonly indirectly regulated by NST1/2 and SND1 in trees, and likely play a role in the biosynthesis of cellulose and hemicellulose in second cell wall of woody plants.

To summary, we proposed a DNA methylation-mediated lignin biosynthesis model which responds to environmental factors ( Figure 5 ). Specifically, relatively low temperature ranges may decrease DNA methylation levels of BpNST1/2 and BpSND1, and lead to the enhanced expressions of BpNST1/2 and BpSND1 and the downstream MYB-domain TFs and lignin biosynthesis genes, and ultimately, result in a high level of lignin content. However, manipulated experiments are needed to test whether DNA methylation of BpNST1/2 and BpSND1 can be directly altered by Temperature Seasonality and/or Temperature Annual Range (TAR), and if variation of these methylation is heritable.

The datasets generated for this study are available on request to the corresponding authors. The sequencing data are available through the NCBI Sequence Read Archive under the accession number PRJNA902946.

BC, JW, SS, and QZ conducted the experimental design, data analysis, and prepared the manuscript. YG, XZ, LW, TZ, LC, ZZ and WZ provided help in carrying out the experiments and guidance in designing the research. QZ, CY, SS, JW, LX and BC interpreted the results and modified the manuscript. All authors contributed to the article and approved the submitted version.

This work was supported by the National Key R&D Program of China during the 14th Five-year Plan Period (2021YFD2200304), Heilongjiang Touyan Innovation Team Program (Tree Genetics and Breeding Innovation Team), the National Natural Science Foundation of China (31871220, 62273086, 62001088), the National Nonprofit Institute Research Grant of the Chinese Academy of Forestry (CAFYBB2019ZY003), and the Fundamental Research Funds for the Central Universities (2572022BD04, 2572018AA27).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.