RNA-dependent RNA polymerase is required for enhancer-mediated transcriptional silencing associated with paramutation at the maize p1 gene

Abstract

Paramutation is the ability of an endogenous gene or a transgene to heritably silence another closely related allele or gene. At the maize p1 (pericarp color1) gene, paramutation is associated with decreases in transcript levels and reduced pigmentation of the endogenous allele that normally specifies red seed coat (pericarp) and cob pigmentation. Herein we demonstrate that this silencing occurs at the transcriptional level and that a specific enhancer fragment from p1 is sufficient to induce all aspects of paramutation. Further, we demonstrate that a mutation in the mop1 gene (mediator of paramutation1), which encodes a RNA-dependent RNA polymerase, is absolutely required for establishing the silencing associated with p1 paramutation. In contrast to its effects on other paramutation loci, the mop1 mutation does not immediately reactivate a previously silenced allele; several generations in the presence of the mop1 mutation are required. In addition, the mop1 mutation was also able to release tissue-specific silencing of another p1 allele that does not participate in paramutation, but does contain a tandem repeated structure and is likely regulated through epigenetic mechanisms. These results demonstrate that RNA-mediated gene-silencing mechanisms play key roles in p1 paramutation and the spectrum of roles for MOP1 is broadened to include tissue-specific expression patterns.

Figure 1.—

Phenotypes of the endogenous p1 alleles and structure of P1-rr. (A) The phenotypes of endogenous alleles showing cob and pericarp pigmentation. (B) Silencing phenotypes of P1-rr′, ranging from orange (weakly silenced) to colorless (completely silenced). (C) The P1-rr coding sequence is flanked by long 5.2-kb direct repeats (solid rectangles) that overlap with smaller 1.2-kb direct repeats (hatched boxes). The 1.6-kb transposon interrupting the upstream-most 1.2-kb repeat is shown as an open triangle. Structure of P1-rr mRNA is shown above the locus map with shaded boxes denoting exons 1–3; the open box indicates alternatively spliced exon 4, and the solid line represents introns. Downstream copies of the P1.2 fragment overlaps with P1-rr exons 3 and 4. The bent arrow indicates the P1-rr transcription start site. The SalI restriction sites within 1.2-kb direct repeats delimit the long-distance enhancer fragment (P1.2). Positions of the P1.2 fragment, as well as the proximal fragments (P1.0 and P2.0, which also have enhancer activities, but show no paramutation activity as transgenes) and the basal P1-rr promoter and 5′-UTR (Pb), are indicated below the map.

Figure 2.—

Nuclear run-on analysis for the P1-rr allele and the P1.2b∷GUS silenced P1-rr′ epialleles in immature pericarp tissue. Nuclei isolated from immature pericarp tissue (18 days after pollination) from homozygous P1-rr and P1-rr′ were used for analysis. The transcription levels for P1-rr and two flavonoid biosynthetic genes directly regulated by p1, a1, and c2 were normalized to the maize Ubiquitin2 gene probe.

Figure 3.—

The P1.2 fragment is sufficient to silence the P1-rr gene. (Top) The crossing scheme. Sequences contained within each construct are diagrammed to the left with constant construct components marked above. The Pb promoter and 35S promoter deletions are variable in the constructs and are shown as solid and shaded rectangles, respectively. Silencing efficiency results for transgenic lines containing each construct are shown to the right.

Figure 4.—

Heritability of silencing induced by promoter replacement constructs. (Top) The crossing scheme and predicted genotypes. In all cases, the female parents used in the crosses demonstrated strong P1-rr′ silencing phenotypes. Sequences contained within each construct are diagrammed to the left with constant construct components marked above. The Pb promoter and 35S promoter deletions are variable in the constructs and are shown as solid and shaded rectangles, respectively. Silencing efficiency results for transgenic lines containing each construct are shown on the right.

Figure 5.—

Secondary paramutagenicity of promoter replacement constructs. The crossing scheme and predicted genotypes are indicated on the top. In all cases, the female parents used in the crosses demonstrated strong P1-rr′ silencing phenotypes. Sequences contained within each construct are diagrammed to the left with constant construct components marked above. The Pb promoter and 35S promoter deletions are variable in the constructs and are shown as solid and shaded rectangles, respectively. Silencing efficiency results for transgenic lines containing each construct are shown on the right.

Figure 6.—

Ear phenotypes of mop1-1/Mop1 and mop1-1 plants. Three representative ears are shown for each group to demonstrate the range of variation observed in pericarp phlobaphene pigmentation in the mop1-1 maintenance test. (A) The P1-rr′/P1-wr; mop1-1/Mop1 plants have a codominant phenotype; dark red cob is specified by P1-wr and light stripes and pigmentation at the silk scar are contributed by P1-rr′. (B) After one generation of exposure to homozygous mop1-1, the P1-rr′/P1-wr ears remain lightly pigmented. (C) After two generations of exposure to mop1-1/mop1-1, the P1-rr′/P1-wr ears exhibit sporadic increases in pericarp pigmentation: light orange blush throughout the pericarp and increased pigmentation near the silk scar. (D) After three generations of exposure to homozygous mop1-1, dark orange and red pericarp pigmentation can be observed in P1-rr′/P1-wr. Ears of the P1-rr′/P1-rr′ ears (not shown) were indistinguishable from P1-rr′/P1-wr. (E) P1-wr/P1-wr ears in mop1-1/Mop1 background have colorless pericarp and red cob typical for P1-wr phenotype. (F) After one generation of exposure of both P1-wr alleles to at least one generation of mop1-1/mop1-1, some P1-wr/P1-wr ears exhibit increased pericarp pigment that we term “blushed.” (G) After two generations of exposure to mop1-1/mop1-1, further increases in P1-wr pericarp pigmentation can be observed.