Paramutation at the maize pl1 locus is associated with RdDM activity at distal tandem repeats

Abstract

Exceptions to Mendelian inheritance often highlight novel chromosomal behaviors. The maize Pl1-Rhoades allele conferring plant pigmentation can display inheritance patterns deviating from Mendelian expectations in a behavior known as paramutation. However, the chromosome features mediating such exceptions remain unknown. Here we show that small RNA production reflecting RNA polymerase IV function within a distal downstream set of five tandem repeats is coincident with meiotically-heritable repression of the Pl1-Rhoades transcription unit. A related pl1 haplotype with three, but not one with two, repeat units also displays the trans-homolog silencing typifying paramutations. 4C interactions, CHD3a-dependent small RNA profiles, nuclease sensitivity, and polyadenylated RNA levels highlight a repeat subregion having regulatory potential. Our comparative and mutant analyses show that transcriptional repression of Pl1-Rhoades correlates with 24-nucleotide RNA production and cytosine methylation at this subregion indicating the action of a specific DNA-dependent RNA polymerase complex. These findings support a working model in which pl1 paramutation depends on trans-chromosomal RNA-directed DNA methylation operating at a discrete cis-linked and copy-number-dependent transcriptional regulatory element.

Fig 1. pl1 haplotypes.

(A) BAC clone contigs containing Pl1-Rhoades assembled from long-read sequences. (B) Pl1-Rhoades haplotype representation with arrows identifying annotated B73 AGPv4 gene models. Bracketed line represents the region used for GEvo alignments with pl1-B73 (D), Pl1-CML52 (E), and pl1-NC350 (F). (C) Images of anther phenotypes conferred by the respective active and repressed states of the Pl1-Rhoades and Pl1-CML52 alleles. pl1-CO159 is a recessive non-functional allele used here to illustrate the pigmentation potential of Pl1-CML52. (D-F) Shaded areas represent regions of synteny, black boxes represent gene models, and magenta boxes immediately upstream of the pl1 gene models represent a doppia transposon fragment. Green shading and black arrows represent sequences found as a penta-repeat in Pl1-Rhoades.

Fig 2. Pl1-Rhoades haplotype 4C-based interactions.

Uniquely-mapping reads per million (rpm) 4C tags in 2kb bins across the Pl1-Rhoades haplotype: schematic combined datasets (A), inner seedling leaves (S) (B), and inner husk leaves (H) (C). (D) Schematic identifying positions and ranks of the five most abundant 4C tags (blue bars) ligated to the bait sequence (magenta bar) in S and H samples excluding 2kb on either side of the bait. Black arrows represent annotated gene models. A single copy of the penta-repeat (green triangles) is displayed in (E) and shows 4C tag positions and ranks (as above), repetitive index (heat map) per 24nt sliding windows, and annotated TE features. Cyan box represents a unique subregion (USR). Arrows represent DNA transposons (light gray), Helitrons (black), and LTR retrotransposons (dark gray).

Fig 3. Chd3a-dependent penta-repeat sRNA profiles.

Alignments of uniquely-mapping 18-30nt reads from libraries representing single Chd3a / Chd3a (A-B), and chd3a-3 / chd3a-3 (C-E) Pl´ 8-day seedlings across a single repeat unit (F) in reads per million (rpm). Arrows represent DNA transposons (light gray), Helitrons (black), and LTR retrotransposons (dark gray).

Fig 4. USR nucleosome profiles.

(A) Relative locations of amplicons (black lines) used in (B). (B) Fold DNA relative to Pl´ samples of mono- and dinucleosome bound DNA after light MNase digestion measured by qPCR in triplicate Pl-Rh and Pl´ seedlings. * p < 0.05. Error bars are s.e.m.

Fig 5. USR poly A+ RNA levels.

Mean fold RNA (2^-ΔΔCt)(±s.e.m) across the USR measured by qRT-PCR normalized to gapdh levels in biological triplicate Pl-Rh and Pl´ male flowers. Error bars are s.e.m.

Fig 6. Penta-repeat sRNA profiles.

Alignments of uniquely-mapping 18-30nt reads from libraries representing single Pl-Rh (A), Pl´ (B), and Pl-Rh / Pl´ (C) immature cob and seedling (D-F) across a single repeat unit (G) in reads per million (rpm). Arrows represent DNA transposons (light gray), Helitrons (black), and LTR retrotransposons (dark gray). Mean rpm over the USR in 10nt bins is shown in stacked bar graphs for cob Pl-Rh (H), Pl´ (I), and Pl-Rh / Pl´ (J), and seedling Pl-Rh (K), Pl´ (L), and Pl-Rh / Pl´ (M) genotypes. Reads colored by length mapping to the plus or minus stands are shown above and below the line respectively.

Fig 7. Unique penta-repeat sRNA profiles dependent on RMR functions.

Alignments of uniquely-mapping 18-30nt reads from libraries representing Rpd1 / rpd1-1 (two combined) (A), rpd1-1 / rpd1-1 (B), Rmr1 / rmr1-1 (C), rmr1-1 / rmr1-1 (D), Rmr2 / rmr2-1 (E), rmr2-1 / rmr2-1 (F), Dcl3 / dcl3-2 (G), and dcl3-2 / dcl3-2 (H) immature cob across a single repeat unit (I) in reads per million (rpm). Arrows represent DNA transposons (light gray), Helitrons (black), and LTR retrotransposons (dark gray).

Fig 8. USR 5mC profiles.

Cytosine methylation profiles of individual PCR amplicons of genomic DNA from Pl-Rh (A), Pl´ (B) and Pl-Rh / Pl´ (C) day-8 seedlings representing the bracketed region of the USR shown in (D). Circles (open: unmethylated, closed: methylated) represent cytosines in CG (red), CHG (blue), and CHH (green) contexts. Black arrows in (D) represent PCR primers used to generate the sequenced amplicons.