| Maize seed development, including structure formation and grain filling, directly determines the weight and quality, but there was less genetic research to related traits of grain structure in maize. RIL population developed from the cross of the large grain maize inbred line Dan232 and the small grain maize inbred line N04 was used in this study, 23 different traits for grain structure were studied by observing paraffin section of maize kernel using microscope, including pericarp, aleurone layer and different location of endosperm at two stages (20 and 30 days after pollination, DAP). Our objiective was to study the relationship between those traits with grain weight, and quality traits, and to estimate the number of genes controlling each trait.QTLs and their effect were analyzed for each trait by ICIM mapping and composite interval mapping method; meta-analysis was used in this study to analyze consistency with previous QTL detected for grain weight, grain quality traits and grain filling rate using the same population, and to reflect genetic basis for grain structure traits at different developmental stages and their association with grain weight and grain quality traits. Major QTL detected for grain structure traits could provide the basis for further fine mapping, cloning and candidate genes exploiting.The main results were as follows:1. There were significant difference for kernel structure traits between parent lines Dan232 (big kernel) and N04 (small kernel), and between two RIL family lines with extreme kernel weight (high grain weight line R125 and low grain weight line R225) at two development stages. Firstly, R125 was greater than R225 for pericarp thickness (PT), pericarp cell length (PCL), pericarp cell width (PCW), aleurone layer thickness (ALT), cell length (ACL), cell width (ACW), cell area (ACA), middle endosperm cell width (MCW), base endosperm cell width (BCW), base endosperm cell density (BCD) and 30DAP pericarp thickness (PT), the number of cell layer (NCL), top endosperm cell width (TCW), cell area (TCA), middle endosperm cell length (MCL), cell width (MCW) and cell area (MCA) at 20DAP. R125 was less than R225 in other traits. Secondly, there were the great differences between parents. N04 is less than Dan232 in the traits of pericarp cell layers (NCL), aleurone layer cell length (ACL), middle endosperm cell length (MCL), cell width (MCW), cell area ( MCA), base endosperm cell length (BCL), base cell width (BCW), cell area (BCA), 20DAP pericarp thickness (PT), aleurone layer thickness (ALT), cell width (ACW), top endosperm cell density (TCD) and pericarp cell layers of 30DAP (NCL), pericarp cell density (PCD), cell number of aleurone per unit length (NUL), top endosperm cell length (TCL), cell width (TCW ), cell area (TCA). And N04 is larger than Dan232 in other traits.2. All traits among RIL family lines were significant or very significant difference except 20DAP ALT and ACW. Variation coefficient varyed from 18.67% to 116.03%. All the characters showed super-parent separation. Only pericarp cell length 20DAP (PCL) showed a low affinity separation, and pericarp cell density (PCD) showed a high parent separation. Aleurone cells length (ACL), aleurone cell width (ACW), top endosperm cell width (TCW), middle endosperm cell length (MCL), cell width (MCW), base endosperm cell width (BCW), 20DAP aleurone layer per unit length of cell number (NUL), top endosperm cell length (TCL) and base endosperm cell length (BCL) and 30DAP aleurone layer thickness (ALT) showed continuous normal distribution. Pericarp cell width (PCW) and pericarp cell density (PCD) were showed multi-peak distribution. Other traits showed skewed distribution.3. The numbers of genes controlled each trait were great different, some characters were not consistent at two time periods. PCW, TCD, TCL, MCD, MCA, BCL and BCA at 30DAP and TCD and SGD at 20DAP possesed more genes, which were 16, 17, 22, 18, 31, 23, 33, 18 and 19, respectively. The gene numbers controlling NCL, PCD, ALT, ACL, ACW, ACA, NUL, BCW at 20DAP, and PT, PCD, ALT, ACA, NUL, and SGD at 30DAP were 4, 3, 2, 3, 3, 4, 3, 4 and 4, 4, 3, 2, 2, 4, 3, 4, respectively. The gene numbers for other characters ranged from 4-16. There were large differences in PT, MCW, SGD, PCW, TCL, TCA, MCD, MCA, BCL, BCW and the BCA across two stages, and NCL, PCD, ALT, ACL, ACW, ACA, NUL, TCD, TCW, MCL and BCD were consistent across two stages.4. There were different relationships among different traits for kernel structure, and with grain weight and quality traits. The correlations were consistent across 20DAP and 30DAP. Except insignificant phenotypic correlations of ALT and PT, NCL, PCL, PCD, NCL and PCL, PCD, ACL, ACW, ACA, NUL, PCW and ACL,ACW, ACA, the correlations among other traits were significant or highly significant. ACL, TCL, TCW, and TCA at 20DAP showed significant or highly significant positive correlations with one hundred fresh weight (GFW) at the same stage. There were significant negative correlations between central endosperm cell length (MCL), per unit area of starch grains (SGD) and one hundred grain dry weight (GDW). There was significantly positive correlation between TCL and one hundred dry weight (GW) at maturity,while there was significantly negatively correlated between TCL and the SGD. Tthere was significant positive correlation between ACL and GFW at 30DAP. There were significant positive correlations among TCD, MCD, BCD and GDW, there was significantly negative correlations between TCD and BCW, MCD and BCW. GW at maturity also was significantly negatively related to BCW.5. Ninety-five QTL for 23 grain structure traits at two periods were detected using the complete range of mapping and composite interval mapping methods. These QTL distributed throughout 10 chromosomes of the whole genome. There were much more QTL detected on chromosomes 1 and 5, 16 and 15, respectively. Only one QTL were mapped on chromosome 7. Using the WinQTLcart software only 20 QTL were obtained, and 75 QTL were detected using QTLicimapping. Nine QTL were commonly detected using the two methods. The contribution of a single QTL was 3.46% -31.34%. The contributions were greater than 10% for 18 QTL. The favorable alleles of 47 QTL were contributed by dent corn inbred Dan232, while other 48 QTL came from popcorn inbred N04. QTL of qNCL20-1-1, qNCL20-1-2, qNCL20-9-1, qPCD30-5-1/2, qNUL20-8-1/2, qNUL 30-2-1/2 could be further studied as the main objective QTL . 6. Meta-analysis were used to integrate 95 QTL for grain structure traits in this research , and 62 QTL for grain weight at different developmental stages, and 68 QTL for grain filling rate and 147 QTL of grain quality traits using the same population in our previous studies. Meta QTL (mQTL) obtained were 20 for grain structure traits, 5 for grain structure traits and grain filling rate and grain weight, 9 for grain structure traits and grain quality traits. mQTL for grain structure traits were located on 8 chromosomes except chromosomes 3 and 7. A total of 70 QTL located in those mQTL regions, which was 73.7% of the total number of detected QTL. Each mQTL contained 2-5 QTL related with 1-4 traits. Five mQTL for grain structure traits and grain weight and grain filling rate were located on chromosomes 3, 4, 8, 9 and 10. 47 QTL located in those mQTL region, which was 19.42% of the total numbers of detected QTL. Each mQTL contained 2-6 QTL related with 2-5 traits. Nine mQTL for grain structure traits and grain quality traits located on chromosomes 2, 3, 4, 5, 8, 9 and 10. 25 QTL located in the mQTL regions, which was 11.11% of the total numbers of detected QTL. Each mQTL contained 2-10 QTL related with 2-6 traits. |
PDF Full Text Request