Epigenetic modifiers are enzymes that catalyze covalent chemical modifications on chromatin and thus play important roles in setting relatively stable states for gene expression and cell identity. It has been widely observed that mutations in epigenetic modifiers cause various types of human diseases, such as developmental defects and cancers. Therefore, deciphering how the mutations affect functions of the epigenetic modifiers in developmental and pathological conditions is crucial for understanding the underlying mechanisms and for identifying potential therapeutic targets.
Although catalytic activities are usually considered to be essentially important for epigenetic modifiers, recent studies suggest that some of them may exert their functions independently of catalytic activities (i.e., noncatalytic functions) [
1]. These findings raise several interesting questions to promote our understanding of epigenetic regulations: (1) what are the determinants of epigenetic modifiers for their dependency or independency of their major functions on catalytic activity? (2) do these noncatalytic functions invalidate the importance of the corresponding chromatin modifications? and (3) technically, what kind of experiments are most suitable to determine the catalytic versus noncatalytic functions? In this issue of
Frontiers of Medicine, Chen
et al. [
2] report a study of the catalytic and noncatalytic functions of the histone methyltransferase SETD2, which provides valuable clues to address these questions.
SETD2 is responsible for catalyzing co-transcriptional histone H3 lysine 36 trimethylation (H3K36me3) as it directly interacts with the hyperphosphorylated, elongating RNA polymerase II (Pol II) [
3] (Fig.1). Mutations in SETD2 were first identified in clear cell renal cell carcinoma (ccRCC) [
4] and subsequently observed in virtually all types of cancers [
5]. It has previously been shown that knockout (KO) of
Setd2 in mice causes embryonic lethality due to impaired vascular remodeling [
6]. In the present study, Chen
et al. first performed a screening of
SETD2 mutations in ccRCC and identified a single-nucleotide mutation (C1685F in human and C1659F in mouse) to produce a catalytically dead (CD) SETD2 protein.
In vitro biochemical and cell-based experiments confirmed that this mutation only abolishes catalytic activity but retains Pol II-interaction. Chen
et al. went on to generate a
Setd2-CD knockin mouse model by introducing this mutation through homologous recombination approaches and performed a side-by-side comparative study between the
Setd2-CD and
Setd2-KO mouse models.
First of all, the homozygous
Setd2-CD and
Setd2-KO mice showed similar embryonic lethality with clear blood vessel defects (Fig.1). Western blot analysis and gene expression profiling showed comparable levels of decrease of H3K36me3 and gene expression alterations. Compared with the previously published cDNA microarray-based gene expression profiles of
Setd2-KO yolk sacs [
6], the herein performed RNA-seq analysis led to identification of more precise gene expression changes. Among the newly identified putative target genes, many collagen genes were found to be consistently downregulated in the homozygous
Setd2-CD and
Setd2-KO yolk sacs (Fig.1). Given that most of collagen genes are very long and highly interrupted (i.e., consist of many exons), this finding is probably relevant to the notion that the Setd2-mediated H3K36me3 is more required for longer genes [
7] and is enriched more in the exons [
8]. Therefore, the dysregulated collagen genes may serve not only as a mechanistic explanation of the vascular remodeling defects but also as potential model genes for studying the mechanisms of gene transcriptional regulation by Setd2/H3K36me3.
Meanwhile, the side-by-side comparative analyses of the phenotypes and gene expression patterns led to the finding that the
Setd2-CD embryos exhibited slightly milder developmental defects compared with the
Setd2-KO embryos, especially considering their completeness of chorioallantoic attachment, an important milestone event in mammalian embryonic development. To further clarify this point, Chen
et al. performed a single-cell transcriptomic analysis of the embryos at an earlier stage. The results showed that, while the cell populations were closely comparable, the 5′
Hoxa cluster genes, which had been known to be specifically expressed and functionally important in allantois [
9], were downregulated in the homozygous
Setd2-KO, but not
Setd2-CD, allantois cells (Fig.1). Given that these clustered genes require specific chromatin modifications for their temporal and spatial expression [
10], this finding opens perspectives in clarifying the mechanism of how Setd2/H3K36me3 regulates these clustered genes in specific developmental circumstance.
Notably, compared with the recently published studies showing that several epigenetic modifiers’ major functions are independent of their catalytic activities [
11–
13], this study suggests that Setd2 represents a different example, because the physiologic functions of Setd2 in mouse embryonic development definitely requires its catalytic activity. Like Setd2, Dnmt1 has also been shown to require its catalytic activity for its functions in mouse embryonic development [
14]. Therefore, a comparison among these different epigenetic modifiers may arrive at an answer to the above-mentioned first question: those epigenetic modifiers whose major functions are independent of catalytic activities may be involved in multi-protein complexes through protein–protein interactions, and the functions of their corresponding chromatin modifications may be mutually redundant; in contrast, the epigenetic modifiers whose major functions require their catalytic activities may act more uniquely, and their corresponding chromatin modifications may be irreplaceable (Fig.1). Nevertheless, it should be noted that the involvement into multi-protein complexes and seemingly redundant functions of the epigenetic modifiers indicate a “fail-safe” system developed by living beings through evolution, therefore their noncatalytic functions do not invalidate the importance of the epigenetic modifiers and their corresponding chromatin modifications.
Lastly, to precisely determine the catalytic versus noncatalytic functions of an epigenetic modifier, the herein performed approach of generation of point mutation CD knockin animal model and directly comparing with the KO model represents an objective standard, although this approach at present is still time consuming and technically difficult. In this regard, development of advanced, more precise and efficient single-nucleotide genome editing methodology [
15] would be helpful to enable comprehensive studies of many disease-causing mutations in human genome. What is more, compelling evidence has suggested that the significance of epigenetic modifiers might not be conserved across species. In yeast,
Set2, the homolog of
Setd2, is dispensable unless under nutrition stress [
16]; in contrast, the depletion of
Set2 in
Drosophila causes late larval lethality [
17]; as for zebra fish, knocking out
Setd2 only leads to developmental delay without affecting viability and fertility [
18]; while
Setd2 KO mice die around E10.5 [
6]. To address such discrepancy and its connection with the catalytic or noncatalytic functions of epigenetic modifiers, and to unveil the increasing significance of epigenetic modifications during evolution, cross-species comparative studies are also anticipated.
PDF
(1226KB)
