Comparative transcriptional profiling of three super-hybrid rice combinations

Abstract

Utilization of heterosis has significantly increased rice yields. However, its mechanism remains unclear. In this study, comparative transcriptional profiles of three super-hybrid rice combinations, LY2163, LY2186 and LYP9, at the flowering and filling stages, were created using rice whole-genome oligonucleotide microarray. The LY2163, LY2186 and LYP9 hybrids yielded 1193, 1630 and 1046 differentially expressed genes (DGs), accounting for 3.2%, 4.4% and 2.8% of the total number of genes (36,926), respectively, after using the z-test (p < 0.01). Functional category analysis showed that the DGs in each hybrid combination were mainly classified into the carbohydrate metabolism and energy metabolism categories. Further analysis of the metabolic pathways showed that DGs were significantly enriched in the carbon fixation pathway (p < 0.01) for all three combinations. Over 80% of the DGs were located in rice quantitative trait loci (QTLs) of the Gramene database, of which more than 90% were located in the yield related QTLs in all three combinations, which suggested that there was a correlation between DGs and rice heterosis. Pathway Studio analysis showed the presence of DGs in the circadian regulatory network of all three hybrid combinations, which suggested that the circadian clock had a role in rice heterosis. Our results provide information that can help to elucidate the molecular mechanism underlying rice heterosis.

Figure 1.

Clustering analysis of expressed genes among three hybrid combinations. Hierarchical clustering was done by GeneSpring 12.6 using normalized data. (A) LY2163 hybrid combination; (B) LY2186 hybrid combination; (C) LYP9 hybrid combination. Fw, flowering; Fi, filling.

Figure 2.

Venn diagram of differentially expressed genes (DGs) shared by three hybrid combinations. (A) DGs at the flowering (Fw) stage; (B) DGs at the filling (Fi) stage; (C) DGs at either the Fw or Fi stage.

Figure 3.

Pie graph representing expression patterns of DGs.

Figure 4.

Graphical representation for the function classification of DGs in the F₁ hybrid of three hybrid combinations.

Figure 5.

Gene regulatory network in three hybrid combinations. DGs in three hybrid combinations were used for direct interaction analysis by Pathway Studio 9.0. Interaction searching includes promoter binding, expression, regulation and binding. LHY, LATE ELONGATED HYPOCOTYL; PRR7, PSEUDO-RESPONSE REGULATOR 7; SVP, SHORT VEGETATIVE PHASE; TIC, TIME FOR COFFEE; GI, GIGANTEA; ZTL, ZEITLUPE; APRR3, PSEUDO-RESPONSE REGULATOR 3; FKF1, FLAVIN-BINDING, KELCH REPEAT, F-BOX 1; TSF, TWIN SISTER OF FT; COP1, CONSTITUTIVELY PHOTOMORPHOGENIC 1; TT4, TRANSPARENT TESTA 4. Black, dark green, navy refers to protein are encoded by DGs in LY2163, LY2186 and LYP9 respectively; dark magenta refers to protein are encoded by DGs in LY2163 and LYP9; orange refers to protein are encoded by DGs in LY2186 and LYP9; red refers to protein are encoded by DGs in LY2163, LY2186 and LYP9.

Figure 6.

Distribution of DGs among QTLs of a small interval. QTLs in Gramene (number of harbored genes ≤ 100) and harbored DGs were aligned to the Michigan State University (MSU) Rice Genome Release 6.1. The long horizontal line represents 12 chromosomes of the rice genome; the short horizontal lines represent the QTL span, and short vertical lines represent DGs.