The role of CRWN nuclear proteins in chromatin-based regulation of stress response genes

ABSTRACT

The periphery in animal nuclei is generally considered to be a transcriptionally repressive environment. Recent studies indicate that chromatin-based mechanisms establish a similar situation in plant nuclei. We demonstrated recently that the loss of CRWN nuclear lamina proteins in Arabidopsis leads to the misregulation of a group of genes involved in plant defense. How this defense response is triggered is largely unknown. Here, we briefly review recent findings that identify several layers of chromatin-based regulation responsible for this response. Further, we introduce new data suggesting that histone H3 lysine 27 tri-methylation levels are reduced in the absence of CRWNs near genes encoding transcription factors regulating SA biosynthesis, providing an explanation for SA induction. These discoveries begin to uncover the interplay between nuclear architecture and stress response in plants.

The animal nuclear lamina (NL) and its associated components have been studied intensively, and interest in this field has heightened with the understanding that mutations and polymorphism in genes encoding NL proteins can elicit lethal genetic diseases in humans.¹ The study of the nuclear lamina and related proteins in plants, however, has a relatively short history, and the phenotypic effects of changing nuclear structure on plant growth and development are poorly understood. Recent publications about plant nuclear envelope proteins, and their association with various phenotypes in corresponding mutants, begin to fill in this gap and shed light on the impact of altering the plant NL and nuclear envelope.

Disruption of CRWN genes in Arabidopsis, which encode components of a putative plant nuclear lamina, is associated with the induction of salicylic acid (SA) responses and ectopic defense signaling.² A key question remains – what is the first spark that ignites SA responses in these mutants? In our recent report, we suggest that plant NL proteins might directly affect transcriptional machinery or epigenetic modifications to activate pathogen signaling through the SA pathway. We also consider an alternative ‘cell-damage’ model in which nuclear dysfunction or age-dependent damage leads to cell death and subsequent induction of SA responses. Here, we discuss some recently published work and additional observations from our group that provide insight into these models and suggest possible refinements.

Among these publications is a study by Hu et al.³ that mapped CRWN1 interacting sites throughout the genome. The authors equate these sites with lamina-associated domains (LADs) in metazoan cell nuclei. These so-called plant-LAD or PLAD domains are characterized by transcriptional repression and reduced accessibility, suggesting that the nuclear periphery facilitates a transcriptionally quiescent environment. Similar to LADs, PLADs are characterized by genes transcribed at low levels, as well as transcriptionally-silent transposons. Hu et al. also show that the repressive histone modification mark H3K27me3 is enriched in PLADs.

The connection between CRWN1 and histone H3K27me3 is reinforced by two recent papers from the Schubert group, who demonstrate an indirect physical interaction between CRWN1 and the PRC2 complexes responsible for histone H3 K27 methylation. Specifically, Hohenstatt et al.⁴ report that PWO1 can bind multiple components of PRC2, including CLF, SWN and MEA, as well as the core histone H3. In addition, the Schubert group⁵ show that PWO1 interacts with CRWN1. That work also demonstrates that upregulated genes in pwo1 and crwn1 crwn2 mutants overlap with H3K27me3 targets.

Consistent with these studies, we find that up-regulated genes in crwn1 crwn2 mutants are enriched in chromatin state 2, which is characterized by both active histone marks and repressive H3K27me3.² These results suggest a model wherein H3K27me3 levels are decreased in crwn1 crwn2 mutants, leading to the loss of transcriptional repression of stress genes – although other explanations exist, such as a relative enrichment in active chromatin marks at these loci. The simpler scenario that the H3K27me3 mark is depleted at over-expressed loci is supported by chromatin immunoprecipitation followed by quantitative PCR (ChIP-qPCR) experiments showing that H3K27me3 levels are reduced in the PR1 locus in crwn1 crwn2 mutants relative to wild-type.²

One complication to this straightforward model and interpretation is that H3K27me3 levels on PR1, and other stress genes, might be reduced as a secondary consequence of activation of these genes via other mechanisms. In the case of SA-regulated genes, such as PR1, it is possible to remove the effect of SA in a sid2 mutant background, in which inducible SA biosynthesis is blocked.⁶ Thus, we performed ChIP-qPCR on sid2 and sid2 crwn1 crwn2 mutants. We found that H3K27me3 levels at PR1 were elevated in sid2 mutants relative to wild type (figure 1a). It is unclear why H3K27me3 abundance is higher in the sid2 mutant, but this pattern was also seen at other loci (see below). Against this higher baseline of repressive marks in the sid2 background, we observed a reduction in H3K27me3 levels at PR1 by the addition of crwn1 and crwn2 mutations. These results demonstrate that CRWN1 and CRWN2, acting redundantly, are required for normal levels of H3K27me3 deposition on the PR1 locus. We observed a similar pattern when examining H3K27me3 levels associated with the AGAMOUS (AG) locus, which is not known to be regulated by SA. The reduction in H3K27me3 at the PR1 locus in crwn1 crwn2 mutants might explain the elevated expression of this locus, but similar changes at the AG locus are not tied to elevated transcript abundance in the crwn1 crwn2 double mutant.² These findings suggest that the chromatin changes at PR1 might play a role, but could be insufficient on their own to explain the boost in expression of this locus in crwn1 crwn2 mutants.

Figure 1.

Chromatin immunoprecipitation/PCR examining H3K27me3 levels at specific genomic loci in crwn mutants. Chromatin was precipitated using anti-H3K27me3 antibodies and the recovery of specific genomic loci was quantified using quantitative PCR as previously described Choi et al.² The data shown here represent a subset of the samples presented in Supplemental Figure S13 of Choi et al.² The genotype of the input samples are shown in the legend at the base of both panels. (a) The recovery of two loci associated with H3K27me3, AGAMOUS (AG) and PR1, is shown. The euchromatic ACTIN7 locus (ACT7), lacking H3K27me3, served as a negative control. (b) The same set of templates was subsequently interrogated for the SARD1 and CBP60g loci using the primers indicated below. The error bars indicate standard deviation (n = 3). Black asterisks indicate the results of a Student’s t test against WT for the crwn1, crwn1/2 and sid2 samples. Green asterisks represent a test of the sid2 crwn1/2 sample relative to the sid2 sample [*, P < .05 and **, P < .01).

SARD1 primers: 5’-GAGAAACGGCGTTGCAACAA-3’ and 5’-CTCCCTTTCGCAACTTTCGC-3’ CBP60g primers: 5’-TGCCCCAGTGATGAGGTTTG-3’ and 5’-AACGGCGAAAATCCTTGACG-3’ Primers for the other loci are provided in [2], Supplemental Table S5.

To investigate other possible contributors, we examined two genes encoding transcription factors, SARD1 and CBP60g, which stimulate the expression of the SA-biosynthesis gene, SID2. The expression of these genes increases markedly in crwn1 crwn2 mutants.² Here, we show that H3K27me3 levels were reduced on the chromatin near CBP60g in crwn1 crwn2 mutants relative to wild type (figure 1b). A similar trend is seen for the SARD1 locus. Further, H3K27me3 levels at the SARD1 locus were lower in sid2 crwn1 crwn2 mutants compared to sid2 mutants (figure 1b). These chromatin modification changes provide an explanation for the increase in SA levels in crwn1 crwn2 mutants and the attendant changes in gene expression and pathogen response.²

Our studies, and related work from other groups, begin to decipher the multi-layered regulation that ties nuclear architecture to transcriptional control of SA-pathway pathogen response genes. On one level, CRWN1 acts as a transcriptional co-repressor of PR1 in conjunction with NTL9.⁷ Second, loss of CRWN1 and its paralog CRWN2 lead to reduced levels of H3K27me3 at the PR1 locus, which might cause or potentiate its elevated expression. Third, a reduction of H3K27me3 at other loci in the genome in crwn mutants induces the expression of genes encoding transcription factors that stimulate SID2 and thus elevate SA levels. This regulatory small molecule then sets off a cascade of signaling events, some of which lead to the amplification of the stimulatory effect on the expression of stress response genes.

The integration of gene regulatory roles in architectural proteins involved in nuclear organization is also seen in other examples in both plants and animals.^8-12 The next challenge is to understand how this integration coordinates cellular responses to stress at both the level of gene expression and organelle structure.

Funding Statement

This work was supported by the Boyce Thompson Institute.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.