Subgenome Dominance in an Interspecific Hybrid, Synthetic Allopolyploid, and a 140-Year-Old Naturally Established Neo-Allopolyploid Monkeyflower

Abstract

Recent studies have shown that one of the parental subgenomes in ancient polyploids is generally more dominant, having retained more genes and being more highly expressed, a phenomenon termed subgenome dominance. The genomic features that determine how quickly and which subgenome dominates within a newly formed polyploid remain poorly understood. To investigate the rate of emergence of subgenome dominance, we examined gene expression, gene methylation, and transposable element (TE) methylation in a natural, <140-year-old allopolyploid (Mimulus peregrinus), a resynthesized interspecies triploid hybrid (M. robertsii), a resynthesized allopolyploid (M. peregrinus), and progenitor species (M. guttatus and M. luteus). We show that subgenome expression dominance occurs instantly following the hybridization of divergent genomes and significantly increases over generations. Additionally, CHH methylation levels are reduced in regions near genes and within TEs in the first-generation hybrid, intermediate in the resynthesized allopolyploid, and are repatterned differently between the dominant and recessive subgenomes in the natural allopolyploid. Subgenome differences in levels of TE methylation mirror the increase in expression bias observed over the generations following hybridization. These findings provide important insights into genomic and epigenomic shock that occurs following hybridization and polyploid events and may also contribute to uncovering the mechanistic basis of heterosis and subgenome dominance.

Figure 1.

Whole-Genome Duplications in Mimulus and Related Species. (A) Tree showing locations of whole-genome duplications (stars) on Lamiales phylogeny. Coalescence-based phylogeny of 96 single copy loci estimated in ASTRAL. Node labels represent bootstrap values for 100 replicates. Nodes with bootstrap values less than 80 were collapsed. Green stars are published WGD events not identified in this study. Red stars indicate events identified in this study. Blue stars represent uncertainty in the nature of either a single event with varying support for the timing of paralog coalescence or two individual events. (B) to (E) Mimulus species used in this study. M. guttatus (2x) hybridized with M. luteus (4x) (C) to produce a sterile triploid M. robertsii (3x) (D), which underwent a subsequent whole-genome duplication giving rise to fertile natural allopolyploid M. peregrinus (6x) (E). (F) Graphic showing chromosome complement of individuals ([B] to [E]) in the middle panel. The allotetraploid M. luteus has two distinct subgenomes; L_a and L_b represent the two distinct subgenomes.

Figure 2.

Expression Bias of Homoeologs Resulting from M. luteus-Specific WGD Event in M. luteus, F1 Hybrid, Synthetic Allopolyploid, and Natural Allopolyploid. Gray histograms show distribution of expression bias () for all testable homoeolog pairs. Testable homoeolog pairs (N) are those that could clearly be identified as homologous and had at least 1 read in each tissue sampled. Homoeolog pairs significantly biased toward the M. guttatus-like homoeolog are crosshatched, while pairs significantly biased toward the “other” homoeolog are shown in solid blue. Across all three hybrid individuals (F1, synthetic, and natural allopolyploid) the “other” subgenome dominates the M. guttatus-like subgenome either by the number of homoeologs biased toward it (NL_b > NL_a) or on average, , where and are averages over all homoeolog pairs that were biased toward L_b or L_a, respectively.

Figure 3.

Homoeolog Expression Bias in Hybrid and Allopolyploids, Comparing the M. guttatus Homoeolog to the Weighted Average Expression of Its Pair of M. luteus Homoeologs. The weighted average of expression of the two M. luteus homoeologs was calculated by dividing the sum of read count of the two M. luteus homoeologs by the sum of their individual gene lengths. Gray histograms show distribution of expression bias () for all testable homoeolog pairs. Only genes that had a clear 2 to 1 (M. luteus to M. guttatus) homology were considered. Homoeolog pairs significantly biased toward the M. guttatus homoeolog are shown in yellow, while pairs significantly biased toward the M. luteus homoeolog are shown in blue. Across all three hybrid individuals (F1, synthetic, and natural allopolyploid) the pair of M. luteus homoeologs, when added together, dominates the M. guttatus homoeolog (i.e., and , where and are averages over all homoeolog pairs).

Figure 4.

Homoeolog Expression Bias in Hybrid and Allopolyploids, Comparing the M. guttatus Homoeolog to Each of Its M. luteus Homoeologs Separately. Gray histograms show distribution of expression bias () for all testable homoeolog pairs. Homoeolog pairs significantly biased toward the M. guttatus homoeolog are shown in yellow, while pairs significantly biased toward the M. luteus homoeolog are shown in blue. Across all three hybrid individuals (F1, synthetic, and natural allopolyploid) the M. luteus homoeolog dominates the M. guttatus homoeolog (i.e., and , where and are averages over all homoeolog pairs).

Figure 5.

Expression Bias in Three Separate Hybrid Lineages. (A) Venn diagram of the number of biased homoeolog pairs across hybrid lineages (1532 homoeolog pairs were biased in all three lineages). (B) to (D) Scatterplots of expression bias () for these 1532 homoeolog pairs comparing hybrid to synthetic allopolyploid, hybrid to natural allopolyploid, and synthetic to natural allopolyploid (red line is linear regression; thin blue line is identity).

Figure 6.

TE Density in a Window Spanning 10 kb Upstream to 10 kb Downstream of a Gene Is Negatively Related to Gene Expression. The vertical axis is gene expression in RPKM. The horizontal axis is transposon density, binned into 10 windows with width proportional to the number of data points it contains. Horizontal gray line indicates the mean of the response, log(RPKM). TE density is negatively related to gene expression in M. guttatus (A) and M. luteus (B).

Figure 7.

Subgenome-Specific Methylation Repatterning in Hybrid and Allopolyploid Mimulus. M. guttatus and M. luteus subgenome-specific patterns of gene (top two rows) and transposon (bottom two rows) methylation. The y axis is the weighted methylation level. The x axis shows the gene body (TSS = transcription start site and TTS = transcription termination site) or TE body and 1 kb upstream and downstream. CG, CHG, and CHH methylation levels are shown in the first, second, and third column, respectively. Methylation levels of each individual are shown in unique colors (parents = red; F1 hybrid = light blue; synthetic allopolyploid = dark green; natural allopolyploid = yellow).