Mutations in the Arabidopsis H3K4me2/3 demethylase JMJ14 suppress posttranscriptional gene silencing by decreasing transgene transcription

Abstract

Posttranscriptional gene silencing (PTGS) mediated by sense transgenes (S-PTGS) results in RNA degradation and DNA methylation of the transcribed region. Through a forward genetic screen, a mutant defective in the Histone3 Lysine4 di/trimethyl (H3K4me2/3) demethylase Jumonji-C (JmjC) domain-containing protein14 (JMJ14) was identified. This mutant reactivates various transgenes silenced by S-PTGS and shows reduced Histone3 Lysine9 Lysine14 acetylation (H3K9K14Ac) levels, reduced polymerase II occupancy, reduced transgene transcription, and increased DNA methylation in the promoter region, consistent with the hypothesis that high levels of transcription are required to trigger S-PTGS. The jmj14 mutation also reduces the expression of transgenes that do not trigger S-PTGS. Moreover, expression of transgenes that undergo S-PTGS in a wild-type background is reduced in jmj14 sgs3 double mutants compared with PTGS-deficient sgs3 mutants, indicating that JMJ14 is required for high levels of transcription in a PTGS-independent manner. Whereas endogenous loci regulated by JMJ14 exhibit increased H3K4me2 and H3K4me3 levels in the jmj14 mutant, transgene loci exhibit unchanged H3K4me2 and decreased H3K4me3 levels. Because jmj14 mutations impair PTGS of transgenes expressed under various plant or viral promoters, we hypothesize that JMJ14 demethylation activity is prevented by antagonistic epigenetic marks specifically imposed at transgene loci. Removing JMJ14 likely allows other H3K4 demethylases encoded by the Arabidopsis thaliana genome to act on transgenes and reduce transcription levels, thus preventing the triggering of S-PTGS.

Figure 1.

Analysis of S-PTGS and tasiRNA Pathways in jmj14 Mutants. (A) LMW and HMW RNA gel blots of aerial parts of 14-d-old seedlings of the indicated mutant plants were probed with an RNA GUS probe. 25S rRNA and U6 small nuclear RNA hybridizations served as loading controls. GUS activity quantification is in fluorescence units per minute and per microgram of total protein. EtBr, ethidium bromide. (B) LMW and HMW RNA gel blot analyses of mature rosette leaves of the indicated mutant plants. LMW RNA gel blots were probed with DNA oligonucleotides complementary to TAS2 major product (siRNA F). HMW RNA gel blots were probed with DNA complementary to the TAS2 precursor. The expected migration positions of primary TAS RNA precursor and the 5′ and 3′ cleavage products generated after miR173-guided cleavage are indicated. 25S rRNA and U6 small nuclear RNA hybridizations served as loading controls. Col-0, ecotype Columbia. (C) Percentage of 35S:NIA2 and endogenous NIA1 and NIA2 cosuppressed plants at the adult stage in the indicated genotypes. Cosuppression frequency was scored as the number of plants that died from NIA cosuppression. 100 plants were analyzed.

Figure 2.

Developmental Phenotype and Analysis of IR-PTGS in jmj14 Mutants. (A) Pictures of the indicated mutants. L1/sgs8 plants flower earlier than L1. (B) Allelism test between sgs8 and pkdm7b-2 as scored by flowering time. (C) Schematic representation of the SGS8/JMJ14/PKDM7b gene. Black boxes represent exons, gray boxes represent untranslated regions, and thin lines represent introns and intergenic regions. Nucleotide and amino acid changes in sgs8/jmj14-4 are indicated. (D) Allelism test between sgs8 and jmj14-3 as scored by JAP3 IR-PTGS phenotype. [See online article for color version of this figure.]

Figure 3.

Pol II Occupancy in jmj14 Mutants. ChIP experiments using anti-Pol II antibodies of the indicated locus in the indicated genotypes. The mean of qPCR is reported relative to the STM gene control in the same genotype (see Methods). Graphical representation shows the fold change as the mean of the different biological repeats. Error bars represent the sd.

Figure 4.

Expression Analysis in jmj14 Mutants. (A) RT-PCR analysis of the 35S:NIA2 pre-mRNA transcript level from 14-d-old-seedlings in sgs8/jmj14-4 mutant. EF1a served as a standard for RT-PCR. Col-0, ecotype Columbia. (B) HMW RNA gel blot analysis of aerial parts of 14-d-old seedlings of the indicated backgrounds was probed with a GUS probe. 25S rRNA hybridizations served as loading control. (C) GUS activity in the indicated backgrounds. GUS activity quantification is in fluorescence units per minute and per microgram of total protein.

Figure 5.

Histone Acetylation Analysis in jmj14 Mutants. ChIP experiments using anti-H3K9K14-Ac antibodies at the indicated locus in the indicated genotypes. The mean of qPCR is reported relative to the STM control in the same genotype (see Methods). Graphical representation shows the fold change as the mean of the different biological repeats. Error bars represent the sd.

Figure 6.

DNA Methylation Analysis in jmj14 Mutants. (A) Partial map of the 35S:GUS locus and expected digestion fragments. Methylation-insensitive EcoRI, HindIII, and methylation-sensitive HpaII/MspI sites are indicated, as well as the expected digestion fragments detected by hybridizing with the 35S probe. (B) DNA of 15-d-old seedlings was digested with EcoRI, HindIII, and MspI (indicated as M) or EcoRI, HindIII, and HpaII (indicated as H) and probed with a DNA 35S probe.

Figure 7.

Global and Targeted H3K4 Methylation Levels in jmj14 Mutants. (A) Immunoblot analysis of H3K4me1, H3K4me2, and H3K4me3 whole-genome levels using the corresponding antibodies in the indicated genotypes. β-tubulin (b-tub) served as a loading control. (B) ChIP experiments using anti-H3K4me3 antibodies of the indicated locus in the indicated genotypes. The FWA endogenous locus was previously shown to be a JMJ14 target, whereas STM served as an insensitive control.

Figure 8.

ChIP Analysis of H3K4me2 and H3K4me3 Levels in jmj14 Mutants. ChIP experiments using anti-H3K4me3 ([A] and [B]) or anti-H3K4me2 ([C] and [D]) antibodies at the indicated locus in the indicated genotypes. The mean of qPCR is reported relative to the STM gene control in the same genotype (see Methods). Graphical representation shows the fold change as the mean of the different biological repeats. Error bars represent the sd.