Transgene-induced silencing identifies sequences involved in the establishment of paramutation of the maize p1 gene

Abstract

A transgene carrying a distal enhancer element of the maize P1-rr promoter caused silencing of an endogenous P1-rr allele in the progeny of transgenic maize plants. Expression of both the transgene and the endogenous P1-rr allele was reduced in the affected plants. The silenced phenotype was observed in the progeny of seven of eight crosses involving three independent transgenic events tested (average frequency of 19%). This phenotype was associated with an induced epigenetic state of the P1-rr allele, termed P1-rr', which is characterized by increased methylation of the P1-rr flanking regions and decreased levels of P1-rr transcript. The P1-rr' epiallele is highly heritable in the absence of the inducing P1.2b::GUS transgene, and it can impose an equivalent state on a naive P1-rr allele in subsequent crosses (paramutation). In contrast, parallel experiments with two other P::GUS transgenes that contained the same basal P1-rr promoter fragment but different upstream sequences revealed no detectable silencing effect. Thus, transgenes carrying a specific enhancer fragment of the P1-rr gene promoter can trigger a paramutant state (P1-rr') of the endogenous P1-rr gene that is maintained in the absence of the inducing transgene. We discuss the potential role of the P1-rr distal enhancer element in the establishment and propagation of a paramutation system in maize.

Figure 1.

Map of the P1-rr Gene and Transgene Constructs. (Top) The maize P1-rr gene coding sequence is indicated by small black boxes (exons) and thin lines (introns); the transcription start site is indicated by a bent arrow. The two 5.2-kb direct repeats flanking the coding sequence are shown as black bars, and the 1.2-kb direct repeats are shown as hatched boxes. The triangle indicates a 1.6-kb transposon present in standard P1-rr (P1-rr-4B2). (Bottom) The P1.0b::GUS, P1.2b::GUS, and P2.0b::GUS transgene constructs used in this study. All constructs contain the basal fragment of the P1-rr promoter sequence (Pb; horizontally striped box), the AdhI gene intron I (bent line), the GUS gene coding region (dotted box), and the PinII gene transcription terminator (black box). P1.0b::GUS contains a 1-kb proximal enhancer fragment (open box); P2.0b::GUS contains an additional upstream-adjacent 1-kb HindIII fragment. P1.2b::GUS contains a 1.2-kb distal (hatched box) enhancer fragment in reverse orientation relative to its genomic position.

Figure 2.

Representative Results Demonstrating the Silencing of P1-rr in the F₁ Progeny of a Cross with a P1-ww Plant Containing the P1.2b::GUS Transgene (event 29-3). Expression of the GUS transgene is illustrated in the square inset overlying the photograph of the corresponding P1-ww ear. In the F₁ progeny ears, the seven ears on the left with normal P1-rr pigmentation are from plants sensitive to herbicide; the four tagged ears on the right have suppressed P1-rr′ pigmentation and are from plants resistant to herbicide (i.e., plants carrying the transgene).

Figure 3.

Cosilencing of the P1.2b::GUS Transgene and the Endogenous P1-rr Allele Correlates with Increased DNA Methylation. (A) The endogenous P1-rr allele and the P1.2b::GUS transgene (event 29-3) exhibit coordinated phenotypic suppression in the F₁ progeny (ears 3, 4, and 6, asterisks). (B) DNA gel blot analysis of plants shown in (A). The GUS probe is an ∼1.8-kb NcoI-AlwnI fragment of the uidA gene coding sequence and detects transgene sequences only. The P15 probe is P1-rr genomic fragment 15, a 405-bp SacI-SalI fragment from the P1.2 enhancer element, which detects both the P1.2b::GUS transgene and the endogenous p1 genes. Digestion with KpnI and hybridization with GUS and P15 probes indicated that the sibling transgenic plants carried similar transgene insertions and p1 alleles. As expected, lanes containing endogenous P1-rr and P1-pr(TP) did not hybridize to the GUS probe. When DNA from the same plants was digested with the methylation-sensitive enzyme SalI and hybridized with the GUS probe, increased methylation of the P1.2b::GUS transgene was detected. Higher molecular mass bands were detected in lanes 3*, 4*, and 6*, containing DNA from silenced plants, compared with lanes 1, 2, and 5, which contained DNA from nonsilenced plants. Rehybridization of the same blot with the P15 probe indicated increased methylation of the P1.2b::GUS transgene (high molecular mass smears in lanes 3*, 4*, and 6*) as well as increased methylation of the endogenous P1-rr allele. Increased methylation of the P1-rr allele was detected as bands with higher molecular mass (12, 11, 4.6, and 4.2 kb) in silenced plants 3*, 4*, and 6*. Nonsilenced P1-rr plants 1, 2, and 5 exhibited bands with the same molecular mass as a standard P1-rr plant (3.4, 3.0, and 1.2 kb). The P1-rr′ methylation pattern is similar to that of the P1-pr allele described by Das and Messing (1994). The P1-pr(TP) isolate used here exhibited the greatest degree of methylation, with two high molecular mass bands of 12 and 11 kb and none of the smaller bands visible. (C) Map of methylation patterns for SalI restriction sites. The P1-rr gene coding region is shown as black boxes (exons) connected by bent lines (introns); the transcription start site is indicated by a bent arrow. The 5.2-kb direct repeats (black bars) overlap with the 1.2-kb direct repeats (hatched boxes). SalI restriction sites are shown as octagons above the map: sites not methylated in both P1-rr and P1-rr′ are indicated by open octagons, sites methylated in both P1-rr and P1-rr′ are indicated by closed octagons, and sites that are not methylated in P1-rr but are methylated in P1-rr′ are indicated by striped octagons. The P1.2, P1.0, and Pb fragments used for transgene constructs are indicated below the map (not to scale). Small red boxes indicate the locations of sequences homologous with probe P15.

Figure 4.

The P1-rr′ Silenced Phenotype Is Correlated with Reduced P1-rr Transcript Levels. Phenotypes of the tested plants are shown at the top. Blots of total RNA from developing pericarp (20 days after pollination) were hybridized with P1-rr cDNA probe (Grotewold et al., 1991). Total RNA loading was normalized by hybridization of the same blot with soybean 18S rRNA (Shirley and Meagher, 1990). The P1-rr and P1-ww control lanes show samples from homozygous plants, whereas the silenced P1-rr′ plants (lanes 1, 2, and 3) were heterozygous with P1-ww. Thus, the expression level of control P1-rr was assigned a value of 200% to account for the twofold dosage of the P1-rr allele; the transcript levels of the P1-rr′ plants are presented as dosage-compensated values relative to P1-rr and normalized to the rRNA signals. The arrow indicates the major P1-rr transcript.

Figure 5.

Representative Cross Demonstrating the Heritability of the P1-rr′ State. Phenotypes of plants used in crosses are shown in the photographs. The segregation of genotypes for two related families carrying the P1.2b::GUS transgene (event 29-3) is described in the table below the cross. Progeny that exhibited suppressed P1-rr′ phenotype in the absence of the P1.2b::GUS transgene (confirmed by DNA gel blot hybridizations) are highlighted in boldface.

Figure 6.

The P1.2b::GUS Transgene Induces a Paramutant State of the Endogenous P1-rr Allele. (Top) Cross of homozygous P1-rr with a plant heterozygous for P1-rr′/P1-ww and hemizygous for the P1.2b::GUS transgene (event 29-4) resulted in 15 plants; nine were herbicide sensitive, and three of these plants produced ears with the suppressed P1-rr′ phenotype. The genotypes of 10 plants were determined by DNA gel blot analysis and are presented here; the genotypes of the other five plants were not determined precisely and hence are not shown. Phenotypes of plants used in crosses are shown in the photographs. (Bottom) One plant homozygous for P1-rr′ and lacking the P1.2b::GUS transgene was backcrossed with a plant homozygous for naive P1-rr. Most of the resulting progeny exhibit the silenced P1-rr′ phenotype (photograph). See text for details. BC1, backcross 1.

Figure 7.

Paramutagenic Activity of P1-rr′ Is Linked with the p1 Locus. (Top) Diagram of sequential crosses used to test whether the silencing effect of P1-rr′ segregates with P1-rr. In the first cross, P1-rr′ segregates from the P1-ww allele. In the second cross, sibling ears carrying P1-rr′, or P1-ww previously exposed to P1-rr′, are crossed with homozygous naive P1-rr. Expression of the naive P1-rr allele is scored in the following generation. See text for details. (Bottom) Tables show the numbers of suppressed ears (P1-rr′) observed among the total ears in each family. The results show that ability to suppress P1-rr is linked with P1-rr′ and is not transmitted through P1-ww nor through randomly segregating factor(s).