Identification of candidate genes for fiber length quantitative trait loci through RNA-Seq and linkage and physical mapping in cotton

Abstract

Background: Cotton (Gossypium spp.) fibers are single-celled elongated trichomes, the molecular aspects of genetic variation in fiber length (FL) among genotypes are currently unknown. In this study, two backcross inbred lines (BILs), i.e., NMGA-062 ("Long") and NMGA-105 ("Short") with 32.1 vs. 27.2 mm in FL, respectively, were chosen to perform RNA-Seq on developing fibers at 10 days post anthesis (DPA). The two BILs differed in 4 quantitative trait loci (QTL) for FL and were developed from backcrosses between G. hirsutum as the recurrent parent and G. barbadense.

Results: In total, 51.7 and 54.3 million reads were obtained and assembled to 49,508 and 49,448 transcripts in the two genotypes, respectively. Of 1551 differentially expressed genes (DEGs) between the two BILs, 678 were up-regulated and 873 down-regulated in "Long"; and 703 SNPs were identified in 339 DEGs. Further physical mapping showed that 8 DEGs were co-localized with the 4 FL QTL identified in the BIL population containing the two BILs. Four SNP markers in 3 DEGs that showed significant correlations with FL were developed. Among the three candidate genes encoding for proline-rich protein, D-cysteine desulfhydrase, and thaumatin-like protein, a SNP of thaumatin-like protein gene showed consistent correlations with FL across all testing environments.

Conclusions: This study represents one of the first investigations of positional candidate gene approach of QTL in cotton in integrating transcriptome and SNP identification based on RNA-Seq with linkage and physical mapping of QTL and genes, which will facilitate eventual cloning and identification of genes responsible for FL QTL. The candidate genes may serve as the foundation for further in-depth studies of the molecular mechanism of natural variation in fiber elongation.

Keywords: Backcross inbred lines (BILs); Fiber elongation; G. barbadense; Gossypium hirsutum; Quantitative trait loci (QTL); RNA-Seq; Single nucleotide polymorphism (SNP).

Fig. 1

The dynamic change in fiber length during development. a “Long” and “Short” cotton fiber lengths at different developmental stages. b “Long” and “Short” cotton fiber elongation rates at different developmental stages. Error bars show the standard errors calculated from three replicates. * indicates a significant difference at P = 0.01 between “Long” and “Short” at a given DPA

Fig. 2

A comparison of sequences between genes from the TM-1 reference genome and the “Long” and “Short” genotypes from Sanger sequencing using PCR products. Vector sequences have been removed, only the sequences with SNP loci between the designed primers were retained. CotAD genes are from the CDS of the AD genome. a CotAD_05094; b CotAD_49847; c CotAD_40792

Fig. 3

Comparative distributions of FL QTL hotspots or a QTL in cotton genome. c5-FL-mQTL-Hotspots refers to hotspots for FL QTL. qFL-07X-c5–1 refers to a FL QTL

Fig. 4

A HRM analysis to confirm the presence of single nucleotide polymorphisms of CotAD_28189 at position 879 nt in a subset of a backcross inbred line population. a Original melting curves. b Melting curves after logarithm calculations. Blue and red curves correspond to “Long”, “Short” genotypes, respectively

Fig. 5

qRT-PCR expression levels of 3 FL candidate genes. a-c Expression levels of gene CotAD_02556, CotAD_28189, and CotAD_02795, respectively. The x-axis represents developmental stages (5, 10, 15, and 20 DPA), and the y-axis indicates the relative expression levels as determined by qRT-PCR. The error bars shown are the means of three biological replicates. * indicates a significant difference at P = 0.01 between “Long” and “Short” at a given DPA