In the present study, we analyzed the QTLs underlying flowering time in the sorghum cultivars Kikuchi Zairai and SC112 and their F2 plants grown under conditions of both natural and 12 h day lengths. A wide variation in flowering time was noted among the parental cultivars and their F2 plants under natural day length. The F2 population demonstrated transgressive segregation for flowering time, which can be caused by both of the parental cultivars contributing favorable or unfavorable alleles for flowering time or a breakage of the linkage between favorable and unfavorable alleles, in addition to the failure to declare small QTLs statistically. The normal distribution signified the continuous genetic variation that exists between F2 plants.
Although a smaller range of variation in flowering time under 12 h day length was found for F2
plants, all of the plants flowered earlier under 12 h day length than when grown under natural day length. The decrease in days to flowering under 12 h day length suggested that sorghum is a short-day plant and flowers most rapidly when illuminated for fewer hours per day (Craufurd et al. 1999
). These results were also reported previously by Garner and Allard (1923)
who showed that flowering in sorghum was accelerated by a daily reduction of the day length. In the present study, flowering in a larger number of F2
plants was accelerated under 12 h day length when compared to the flowering time of the early-flowering Ethiopian cultivar. Accordingly, the Japanese cultivar allele appeared to delay the flowering time under natural day length, whereas the Ethiopian cultivar allele suppressed the delayed effect on flowering by the Japanese cultivar allele and accelerated flowering under 12 h day length. Under 12 h day length, the Ethiopian cultivar flowered nine days earlier than under natural day length and the Japanese cultivar flowered 30 days earlier.
The linkage maps constructed in this study are most likely among the rare sorghum genetic linkage maps constructed entirely of SSR markers. In contrast, the available sorghum genetic linkage maps are based mainly on RFLPs or a combination of different markers types, especially RFLPs with other marker types, such as SSRs (Chanterau et al. 2001
, Menz et al. 2002
), AFLPs, RAPDs (Haussmann et al. 2002
) and DArTs (Mace et al. 2009
). However, under both of the day length conditions, the total map length was larger than the range previously reported: the distances between the adjacent markers are larger in our map than in previously published maps. This result may be due to the segregation pattern of the genotypic data and the type of SSR markers used in this study; most of the markers were highly distorted and skewed. The SSR markers used were more affected by distortion than the other markers used in previous studies. Most of the markers showed 3 : 1 segregation ratio, and markers with unclear polymorphism were excluded to minimize scoring errors; however, the physical distance between the selected markers was relatively large compared with previous maps.
Five QTLs controlling flowering time were detected under both natural and 12 h day lengths, whereas qFT1-1 on chr 1, qFT7 on chr 7, qFT8 on chr 8 and qFT8b on chr 8b were detected only under the natural day length. These four QTLs were considered to be sensitive to the photoperiod due to the response to the change in day length. These QTLs explained 27.1% of the total phenotypic variation and controlled the photoperiodic sensitivity, as a discrepancy in the day length or photoperiod was required for their expression. Conversely, qFT5 on chr 5 and qFT6b on chr 6b were identified only under 12 h day length and were expressed under a fixed day length, suggesting that their expression was not affected by the change in day length and that they were insensitive to the photoperiod.
The 9 QTLs identified under natural day length explained 60% of the variation of the flowering time. The 7 QTLs identified under 12 h day length explained 46.6% of the variation of the flowering time, which explains the complex genetic nature of the flowering time in sorghum and the possibility of environmental influences on this trait.
In this study, positive additive effects suggested that the alleles of SC112 contributed to the earliness of the flowering time of F2 plants. Furthermore, the small additive effects of individual QTLs indicated the complexity of the genetic control of flowering time in sorghum.
These results are similar to the finding of a study conducted by Srinivas et al. (2009)
in which nine QTLs controlling flowering time were identified in sorghum, with very small additive effects ranging from 1.24 to 1.96. These results are also similar to the finding of Mace et al. (2012)
who described that small additive effects of QTLs controlling morphological traits can be explained by the smaller heritability of flowering time.
Similarly, Buckler et al. (2009)
studied the variation in flowering time with a set of 5000 recombinant inbred lines (maize Nested Association Mapping population, NAM) and explained that one million plants were assayed in eight environments but showed no evidence of any single large-effect QTLs. Indeed, the authors identified 36 QTLs that explained 89% of the total variance of the flowering time in maize. Buckler et al. (2009)
identified evidence of numerous small-effect QTLs shared among families; however, allelic effects differ across founder lines. In their study, no major QTLs were identified at which allelic effects were determined by the geographic origin or large effects of epistasis or environmental interactions. On the basis of these results, Buckler et al. (2009)
suggested that in outcrossing species maize, the genetic architecture of flowering time is dominated by small, additive QTLs, concluding that a simple additive model accurately predicts flowering time in maize, in contrast to the genetic architecture observed in rice and Arabidopsis
These findings in maize described by Buckler et al. (2009)
strongly support the results of the present study because Buckler et al. (2009)
concluded that there were two different types of genetic architecture of flowering time in plants: one based on numerous small-effect QTLs controlling flowering time in outcrossing species, (maize) and another type based on a single large-effect QTL in rice and Arabidopsis
Numerous QTLs controlling flowering time in sorghum have been identified in previous studies (Crasta et al. 1999
, Dufour 1996
, Feltus et al. 2006
, Hart et al. 2001
, Lin et al. 1995
, Paterson et al. 1995
, Srinivas et al. 2009
); however, no QTL controlling flowering time or sensitivity to photo-periodic changes with a major effect was identified in previous studies of sorghum. Moreover, it is expected that new recombinations will help to identify new QTLs; therefore, we compared our results with previous studies on flowering time and photoperiodic responses in sorghum to account for possible new QTLs in addition to the QTLs previously identified ().
List of QTLs controlling flowering time in sorghum detected in previous studies and the present study
on chr 2 was mapped to a position adjacent to that mapped by Srinivas et al. (2009)
, as shown in and qFT3
(101.7–123.1 cM) was mapped to a position adjacent to the QTL mapped on chr 3 by Srinivas et al. (2009)
. The QTLs identified on chr 5 in this study (qFT5
) were located at the same physical positions as the QTLs reported by Srinivas et al. (2009)
However, no QTLs were mapped to the same genomic regions as qFT7
(34.7–53.0 cM) and qFT10
(134.4–152.9 cM) in previous studies. In addition, no QTL controlling flowering time in sorghum was reported in previous studies on chr 8 at the same position as qFT8
delimited by SB4292 and SB4327 on chr 8 in this study. Therefore, qFT7
8 and qFT10
mapped in the present study on chr 7, chr 8 and chr 10, respectively, are considered to have been newly mapped, as they were not reported in previous studies (). In addition, qFT8b
was previously mapped by Srinivas et al. (2009)
. The map location of genes involved in the photo-periodic response in sorghum will be discussed in comparison with rice genes involved in photoperiodic responses. The region on chr 8 of sorghum, which carries a photoperiod QTL, aligns with a region on chr 6 of rice between SSR marker locus RZ144 and isozyme pgi-
2, which is linked to Se
-1, a major photoperiod sensitivity gene in rice (Yano et al. 1997
Recently, Murphy et al. (2011)
reported that Ma1
has the largest impact on flowering time in sorghum. Thus, we can suggest that the Ethiopian cultivar might promote flowering time via the effect of Ma1
or its homologs. In addition, Lin et al. (1995)
mapped the QTL (FlrAvgD1 = QMa1.ugaD) linked to SBI06 (31–59 cM) and suggested that this QTL corresponded to Ma1
. Using genotypes known to segregate for Ma1
, Klein et al. (2008)
showed that Ma1
mapped to an adjacent region on SBI-06 (approx. 11–21 cM). In the present study, qFT6b
was mapped in the region delimited by SB3392 and SB3733 (0.0–25.2 cM) on chr 6b under 12 h day length and could correspond to the Ma1
allele because it was mapped to a region adjacent to SBI-06 (Klein et al. 2008
, Lin et al. 1995
Childs et al. (1997)
mapped the Ma3
maturity gene to SBI-01 (115.5–125.7 cM) and determined that the ma3
R mutation of this gene causes a phenotype similar to plants known to lack phytochrome B. In the present study, qFT1-1
was mapped to the region delimited by SB105 and SB258 (112.0–120.3 cM) on chr 1 under natural condition, corresponding to the region adjacent to the Ma3
allele as reported by Child et al. (1997)
. Consequently, qFT1-1
could correspond to the Ma3
allele, as it was mapped on a region adjacent to SBI-01.
As the data in Lin et al. (1995)
were inconsistent with the assigned map location of QMa1.ugaD in Feltus et al. (2006)
, further studies are suggested to confirm these results. Furthermore, the correspondence between the QTLs that modulate flowering time identified in genetic studies and Ma1
is not entirely clear because the location of Ma2
on the linkage map is not known.
The present study indicated that the flowering time in sorghum was controlled by a large number of QTLs with small effects, suggesting that the genetic architecture of the flowering time in sorghum was similar to maize. This study represents a preliminary and basic study of the QTLs controlling flowering time in sorghum, and the results of this study emphasize the investigation of the genetic architecture of flowering time in sorghum, comprising the scope of our future research. Finally, the interaction of the QTLs controlling flowering time in sorghum with the photoperiod appears to be fundamental to the improvement of this crop and to feed the world’s expanding populations, especially because sorghum is particularly adapted at low levels of input and is suited to hot and dry agro-ecologies in which it is difficult to grow other food crops.