Allelic variation of gene expression in maize hybrids

Abstract

Allelic expression variation of nonimprinted autosomal genes has recently been uncovered in mouse hybrids and humans. The allelic expression variation is attributed to differences in noncoding DNA sequences and does not involve epigenetic regulation or gene imprinting. This expression variation is suggested to play important roles in determining phenotypic diversity. Virtually nothing is known about such allele-specific expression variation in a hybrid plant where two alleles are compared in the same genetic context. We examined parental transcript accumulation in maize (Zea mays) hybrids using allele-specific RT-PCR analysis. Among 15 genes analyzed, 11 showed differences at the RNA level, ranging from unequal expression of the two alleles (biallelic) to expression of a single allele (monoallelic). Maternal or paternal transmission had little effect on the allele-specific transcript ratio of nearly all genes analyzed, suggesting that parent-of-origin effect was minimal. We analyzed the allelic difference in genetically contrasting hybrids and hybrids under high planting density and drought stress. Whereas a genetically improved modern hybrid expressed both alleles, a less improved old hybrid frequently showed mono-allelic expression. Furthermore, the two alleles in the hybrid responded differentially to abiotic stresses. The results of allele-specific regulation in different tissues in responding to environment and stress suggest an unequivalent function of the parental alleles in the hybrid, which may have an impact on heterosis.

Figure 1.

Schematic Illustration of Allele Expression Analysis Using the WAVE dHPLC System and Genomic PCR as Control. (A) RT-PCR and allele-specific cDNA quantification. The parental alleles of a gene are cloned and sequenced to find an allelic polymorphism, either SNPs or InDels, shown as the thin line. RT-PCR is then performed with hybrid cDNA using primers designed at the conserved region between the alleles. The RT-PCR products were separated and quantified by the WAVE dHPLC system. The longer DNA fragments corresponding to one parental allele have higher affinity than the shorter DNA fragments and, therefore, take a longer time to be eluted from the WAVE column. (In the case of SNPs, the allele-specific fragments were separated by differential melting temperature.) Chromatogram traces for each PCR were generated by UV detection. When both alleles are expressed, the two types of cDNA sequences could form two types of heteroduplex (on the left of the chromatogram in some hybrids if any) in addition to the two types of homoduplexes (corresponding to two alleles). The two types of heteroduplexes are eluted earlier than the homoduplexes because of the low affinity to the column and shown as one or two peaks depending on the separation condition. Peak areas corresponding to the homoduplexes and heteroduplexes were calculated by WAVEMAKER software. The x axis is the time in minutes when the DNA fragments are eluted out. The y axis is the UV absorbency unit measuring the DNA concentration or expression level. This analysis quantifies the allele-specific transcript in a relative ratio and does not measure the absolute transcript level expressed in the hybrid. (B) Allelic ratio of PCR product from genomic DNA mixture series. The genomic PCR was used to test whether PCR amplification maintained the allelic ratio with different proportions of the allelic genomic DNA. Genomic DNA from parental inbreds S1 and NS1 were mixed according to the ratio of S1:NS1 as 1:1, 1:2, 1:4, and 1:6. We used gene-specific primers (LTP was used because of the available genomic DNA sequences and its dynamic allelic expression differences; see Results) for the PCR analysis. The amplified genomic DNA fragments, a mixture of the two parental alleles, were separated and quantified by the WAVE dHPLC system. Three replicates were done for each mixture, and the means of the genomic allelic ratio (NS1:S1) are shown with standard deviation (error bar). The allelic ratios of the genomic PCR product are not significantly different (P > 0.05) from the corresponding DNA mixture ratio.

Figure 2.

WAVE Chromatograms Showing Allelic Variation in Transcript Accumulation in Hybrids. The x axis is time in minutes when the cDNA fragments are eluted from the column. The y axis is the DNA concentration measured by UV absorbency. (The y axis plots are not shown to allow for better juxtaposition of the allelic traces for comparison.) The parent samples were used as allele references. (A) Allele expression in different tissues and growing environments. Each sample consisted of tissue pooled from three individual plants. A, April 1 planting; M, May 8 planting; J, June 3 planting. 1°, primary ear; 2°, secondary ear. The parent samples were from the May planting and are shown as allele references. Arrows indicate the relative allele expression changes. (B) Expression of five genes is shown as examples in hybrids of reciprocal cross, in which the female parent is written on the left and the male on the right. The first three genes (left) are examples of genes in which the allele expression pattern is not significantly different between the reciprocal hybrids. The two genes on the right exhibited significant difference in allele-specific transcript ratio (P < 0.05) when alleles transmitted maternally versus paternally (Table 3). (C) Allele expression patterns in the S1/NS1 hybrid. (D) Allele expression patterns in the S2/NS2 hybrid. Contrasting expression patterns (biallelic versus monoallelic expression) between the two hybrids are shown in the first four genes (see arrows). The last gene, Histone H2B, which exhibited biallelic expression, is shown as a control to illustrate that the monoallelic expression of other genes in hybrid S2/NS2 was not an artifact of RNA or cDNA quality because the same RNA and cDNA samples were used in all genes analyzed.

Figure 3.

Allele Expression Variation in Response to Density and Drought Stresses. (A) Density treatment. Plants showed symptoms of thin and tall stature and leaf senesces (inset) in the density stressed treatment. (B) Drought treatment. Leaf rolling can be seen in the drought-stressed seedlings as compared with the control. For expression analysis, three plants were pooled as one sample, and three biological replicates are shown. Expression of the two alleles in the hybrid was compared relative to each other. Arrows indicate the relative allele expression changes in Discussion. The seedlings were at the V4 leaf stage. The control plants in the drought treatment experiment appeared to experience some degree of density stress as shown by the lower expression of the inbred S1 PRP allele compared with the control in the density experiment. This could be because the drought experiment was planted at 50 plants per flat, and the control in the density experiment was planted at 20 plants per flat. However, the differential response of the two alleles to the drought stress treatment can still be observed. Parents were used as allele references and not treated with any stress.