Pleiotropism of the Photoperiod-Insensitive Allele of Hd1 on Heading Date, Plant Height and Yield Traits in Rice

Five populations segregated in isogenic backgrounds and three sets of near isogenic lines (NILs) overlapping in a 362.3-kb region covering heading date gene Hd1 were developed from the indica rice cross Zhenshan97 (ZS97)/Milyang 46 (MY46). They were used to analyze the effects of Hd1 on heading date, plant height and yield traits. In a background of the parental mixtures, the photoperiod-sensitive allele derived from ZS97 functioned in promoting and delaying flowering in the natural short-day and long-day conditions, respectively. In the background of ZS97, no response to the photoperiod was observed, whereas the photoperiod-insensitive allele derived from MY46 functioned in delaying flowering, increasing plant height, and enhancing grain productivity. The additive effects estimated in two NIL sets were 6.14 and 6.14 d for heading date, 4.46 and 5.55 cm for plant height, 10.82 and 11.54 for the number of spikelets per panicle, 6.82 and 8.00 for the number of grains per panicle, and 2.16 and 2.23 g for grain yield per plant, which explained 94.1% and 96.3%, 70.5% and 84.8%, 52.4% and 55.2%, 28.9% and 39.2%, and 36.5% and 26.9% of the phenotypic variances, respectively. Since the photoperiod-insensitive allele of Hd1 confers a long vegetative phase, it is a good candidate for breeding rice varieties with high yielding potential for low latitudes.


Introduction
Heading date (HD) is the most crucial factor determining the regional and seasonal adaptation of rice (Oryza sativa L.). This trait is decided by basic vegetative phase (BVP), photoperiod sensitivity (PS) and temperature sensitivity (TS), among which PS plays the leading role as shown in molecular studies of quantitative trait loci (QTLs) underlying natural variation of flowering in rice. Of the eight QTLs that have been cloned, seven are involved in the photoperiodic control. Hd3a is the rice florigen for the PS pathway, with the expression being promoted in short-day (SD) conditions and suppressed in long-day (LD) conditions [1]. Six other QTLs involved in this pathway are Hd1 and Ehd1 that regulate Hd3a [2,3], and Hd6 [4], Ghd7 [5], DTH8/Ghd8 [6,7] and Hd17 [8] that regulate Hd1 and/or Ehd1.
Nonetheless, the remaining QTL cloned for HD in rice, DTH3, appears to be not involved in the PS pathway, showing similar effects in SD and LD conditions [9]. Two of the QTLs involved in the PS pathway, Hd1 and Ehd1, are also known to control the BVP. At the Hd1 locus, the photoperiod-sensitive allele (PS allele) confers a short BVP and the non-sensitive allele (non-PS allele) confers a long BVP. At the Ehd1 locus, the dominant allele Ef1 confers a short BVP and the recessive alleles ef1 and ef-h confer long BVP [10].
Increasing grain yield has been the most important objective of rice production. The duration of accumulating assimilation products are largely determined by heading date, while plant height (PH) is a key morphological trait related to yield potential.
In populations that were used in QTL mapping for HD, PH and yield traits, it is not uncommon that QTLs for PH and/or yield traits were detected in regions where genes or QTLs for HD were located [11]. Among the eight QTLs cloned, Ghd7 and DTH8/ Ghd8 were reported to have major effects on HD, PH and yield [5][6][7]. On the one hand, the association of enhancing grain yield with prolonging heading date might limit the regional and seasonal adaption of rice varieties [10]. On the other hand, long BVP is desirable for increasing the yield potential of rice in low-latitude regions and efforts have been made to identify genes responsible for long BVP [12,13]. Investigation of the pleiotropism of genes controlling BVP in rice would help to establish an optimum breeding strategy for a specific ecological area.
In a previous study using populations derived from a residual heterozygote identified from a recombinant inbred population of the indica rice cross Zhenshan 97 (ZS97)/Milyang 46 (MY46), we found that a 1.90-Mb genomic region covering Hd1 had significant effects on heading date and yield traits [14]. In the present study, five populations segregated in isogenic backgrounds and three sets of near isogenic lines (NILs) overlapping in a 362.3kb region covering Hd1 were developed and used to analyze the multiple effects of Hd1 on heading date, plant height and yield traits.

Development of the Rice Materials
Five populations and three sets of near isogenic lines (NILs) were used in this study. Each of them was segregated in a region on the short arm of rice chromosome 6 that cover Hd1 or is closely linked to Hd1. They were derived from the indica rice cross ZS97/ MY46 as described below and summarized in Fig. 1. At the Hd1 locus, ZS97 has the PS genotype of Se-1 u Se-1 u and MY46 has the non-PS genotype of Se-1 e Se-1 e [15].
An F 9 plant of ZS97/MY46 carrying a heterozygous segment covering the Hd1 region was selected. From selfed seeds of this plant, 413 F 10 individual were produced and assayed using SSR markers RM19715 and RM19784 that flank the Hd1. Nonrecombinant homozygotes were selected and an F 11 population consisting of 168 lines was developed.
Another F 9 plant that carried a MY46 segment covering the Hd1 region was selected and backcrossed to ZS97 for two times, followed by four generations of selfing. A BC 2 F 5 plant carrying a 362.3-kb heterozygous segment covering the Hd1 gene was identified. This plant was also found to be homozygous at 138 SSR marker loci that are polymorphic between the parental lines and located in background regions distributed on all the 12 chromosomes of rice. From selfed seeds of the BC 2 F 5 plant, a BC 2 F 6 population was derived and assayed with RM19784 and six new markers ( Table 1) that are located in the 362.3-kb region. Three plants carrying overlapped heterozygous segments were identified, from which three BC 2 F 7 populations were developed. They consisted of 468, 312 and 325 individuals and designated G7001, G7002 and G7003, respectively. While G7001 and G7002 segregated in regions Si9337-Si9369 and Si9337-RM19784 covering Hd1, respectively, G7003 was homozygous at the Hd1 locus (Fig. 2).
The three BC 2 F 7 populations were each assayed with two markers flanking their segregation regions, respectively, i.e., Si9337 and Si9369 for G7001, Si9337 and RM19784 for G7002, and RM19784 and Si9575 for G7003. Plants carrying non-recombinant homozygotes were identified and selfed to produce homozygous lines. Three sets of NILs were developed and named ZH1, ZH2 and ZH3 (Fig. 2). The numbers of ZS97 and MY46 homozygous lines included were 37 and 35 in ZH1, 34 and 34 in ZH2, and 36 and 44 in ZH3, respectively.

Field Experiments
The rice populations were tested in the experiment stations of the China National Rice Research Institute located in either Lingshui of the Hainan Province or Hangzhou of the Zhejiang Province, China (Table 2). A planting density of 16.7 cm626.7 cm was used for all the trials. During the floral transition period of the rice materials, the day length was shorter than 12.5 h in Lingshui and longer than 13.5 in Hangzhou ( Figure  S1) (Data was collected from www.timeanddate.com). Since the critical day length for triggering heading in rice is known to be 13.5 h [16,17], day-length in Lingshui and Hangzhou are corresponding to natural SD and LD conditions, respectively.
The F 10 , F 11 and three BC 2 F 7 populations were measured for HD only, among which the F 10 and three BC 2 F 7 populations were scored on a single-plant basis, whereas the F 11 lines were scored as the mean value over two replications. For the three NIL sets, a randomized complete block design was applied with two replications of eight plants per line. HD and PH were scored for each of the plants. At maturity, five middle plants of each line were harvested in bulk and measured for four yield traits ( Table 2).

DNA Marker Analysis
Total DNA was extracted following the method of Zheng et al [18]. PCR amplification was performed according to Chen et al [19]. The products of the two SSR and the three InDel markers were visualized on 6% non-denaturing polyacrylamide gels using silver staining, and that of the three CAPS markers were visualized on 2% agrose gels using Gelred staining. All of the SSR markers were selected from the Gramene database (www.gramene.org). The CAPS and InDel markers were designed using Oligo Primer Analysis Software Version 7.0 [20] based on SNPs and InDels between ZS97 and MY46 detected by the whole-genome resequencing. While Si9337 detected a 4-bp InDel in the exon 2 of the Hd1 gene ( Figure S2), other markers were located outside the Hd1 locus.

Data Analysis
Linkage map construction and QTL analysis were performed for the F 10 , F 11 and BC 2 F 7 populations. The maps were constructed using Mapmaker/Exp 3.0 [21]. Distances between markers were presented in centiMorgan (cM) derived using Kosambi function. QTLs were determined with the composite interval mapping of Windows QTL Cartographer 2.5 [22].
Two-way ANOVA was conducted to test the phenotypic differences between the two genotypic groups in each of the three NIL sets. The analysis was performed with SAS procedure GLM [23] as described previously [24]. Given the detection of significant difference (P,0.01), the same modal was applied to estimate the additive effect and the proportion of phenotypic variance explained.

The Effect of Hd1 on Heading Date
The F 10 plants grown in Lingshui displayed a discontinuous distribution of 312 early-heading (87-91 d) and 99 late-heading (97-104 d), fitting the 3:1 ratio for a single dominance gene (P = 0.67). The major effect of the Hd1 region on heading date was confirmed by QTL analysis using the segmental linkage map spanning 5.8 cM. It accounted for 98.2% of the phenotypic variance, with the ZS97 allele promoting heading date by 6.99 d ( Table 3).  The major effect was also detected in the F 11 population grown in Hangzhou, but the allele for promoting heading date was derived from MY46. The effect explained 97.6% of the phenotypic variance, with the ZS97 allele delayed heading date by 6.98 d ( Table 3).
In the two BC 2 F 7 populations that segregated the Hd1 locus and grown in Lingshui, G7001 and G7002, continuous distribution for heading date was observed although the ZS97 homozygotes tend to flower earlier than MY46 homozygotes. The additive effect of Hd1 and its contribution to the phenotypic variance were estimated to be 4.64 d and 67.8% in G7001, and 5.34 d and 63.4% in G7002, respectively (Table 3), which are much lower than the effects detected in the F 10 and F 11 populations. As expected, no significant effect on heading date was detected in G7003 that was segregated in a region excluding the Hd1 locus.
Among the three NIL sets grown in Hangzhou, significant variations on heading date were detected in ZH1 and ZH2 that were heterogenous at the Hd1 locus. The additive effect of Hd1 and its contribution to the phenotypic variance were estimated to be 6.14 d and 94.1% in ZH1, and 6.14 d and 96.3% in ZH2, respectively ( Table 3). As expected, no significant effect on heading date was detected in ZH3 that was homogenous at the Hd1 locus.
Worthy of note, the allele for promoting heading date were all derived from ZS97 in the two BC 2 C 7 populations and two NIL sets, although the BC 2 C 7 populations were grown in Lingshui and the NIL sets in Hangzhou. This is obviously different from the results obtained from the F 10 and F 11 populations.

The Effect of Hd1 on Plant Height and Yield Traits
As described above, the effect of the Hd1 was detected in the NIL sets ZH1 and ZH2. As shown in Fig. 2, the common heterogenous region of the two NIL sets extended from markers Si9337 to Si9369. As referred to the physical positions in the Nipponbare genome (www.gramene.org), Si9337 is located at 9,337,119-9,337,269 bp and Si9369 at 9,369,434-9,369,808 bp of the rice chromosome 6. In addition to Hd1 (LOC_Os06g16370), four annotated genes, LOC_Os06g16380, LOC_Os06g16390, LO-C_Os06g16400, and LOC_Os06g16410, were located in this region. None of the four annotated genes have been shown to affect heading date, plant height and yield traits in rice. Additionally, no gene for plant height and yield traits has been cloned or finemapped in this region, thus the effects on plant height and yield traits detected in the two NIL sets would provide an evidence for the pleiotropism of the Hd1 gene.
Result of the two-way ANOVA on the phenotypic differences between the two genotypic groups in each of the three NIL sets were presented in Table 4. Significant variations were found in the NIL sets ZH1 and ZH2 for all the traits analyzed except for TGW in ZH1. Both the frequency distribution ( Figure S3) and the results of ANOVA indicated that the Hd1 region had major effects on PH, NSP, NGP and GY. Moreover, phenotypic variations and the genetic effects detected for each of the traits were similar over the two NIL sets. In ZH1 and ZH2, the contribution to the phenotypic variance were 70.5% and 84.8% for PH, 52.4% and 55.2% for NSP, 28.9% and 39.2% for NGP, and 36.5% and 26.9% for GY, with the MY46 allele increasing PH by 4.46 and

Discussion
Although many regions covering QTLs for HD exhibits significant effects on PH and yield traits in primary mapping, the pleiotropism has only been confirmed for Ghd7 and DTH8/ Ghd8. In the present study, we demonstrated that the key HD gene Hd1 has major effects on HD, PH and yield traits in the genetic background of the indica rice variety ZS97. Similar to Ghd7 and DTH8/Ghd8 [5][6][7], Hd1 affects grain yield primarily because of its influence on NSP and NGP. Nevertheless, our study showed that Hd1 might also affect TGW although the effects are smaller and less consistent.
It has been shown that Hd1 exhibits dual functions on the flowering of rice depending on the day length, which promotes heading under SD conditions but is converted to represses heading in LD conditions [25]. In the F 10 and F 11 populations of ZS97/ MY46, the ZS97 allele at Hd1 promoted heading in natural SD condition (Lingshui) and delayed heading in natural LD condition (Hangzhou), which is in accordance with the understanding that ZS97 and MY46 carries PS and non-PS alleles at the Hd1 locus, respectively [15].
Breeding rice varieties with long BVP and weak PS has been consider an important strategy in low latitudes where short photoperiods persist throughout the year. In the BC 2 F 7 and BC 2 F 8 populations of ZS97/MY46, the non-PS allele at Hd1 which was derived from MY46 delayed heading in both the natural SD and LD conditions, which is in accordance with the understanding that non-PS allele of the Hd1 confers a long BVP [10]. Moreover, the non-PS allele from MY46 was found to be associated with increases in grain number, providing a good candidate for breeding rice varieties with high yielding potential for low latitudes.
It is noteworthy that the sensitivity to photoperiod of the ZS97 allele of Hd1 observed in the F 10:11 population disappeared in the BC 2 F 7:8 populations. It was found that the F 10:11 individuals are similar to MY46 with 73.1% of 207 polymorphic markers in the background showing MY46 genotype, while the BC 2 F 7:8 individuals are largely identical to ZS97 with 95.7% of 138 polymorphic markers in the background showing ZS97 genotype (data not shown). This suggests that the effect of the Hd1 is not only determined by the environmental conditions but also depends on the genetic background. The similar phenomenon was observed in se5 mutant, in which the null allele of Hd1 consistently promoted heading in both SD and LD conditions [26]. Such a phenomenon was also observed in phyB mutant, in which delaying HD caused by over-expression of the Hd1 under SD conditions was eliminated [17]. It has been reported that the Hd1 gene functions in the evolutionarily conserved OsGI-Hd1-Hd3a pathway for the photoperiodic control of flowering in rice [27]. Based on the genotypes of DNA markers used for background detection in our studies, it was found that the F 10:11 populations carried MY46 homozygotes in the region covering OsGI on the short arm of rice chromosome 1, whereas the BC 2 F 7:8 populations carried ZS97 homozygotes in this region. Since ZS97 is an early season rice cultivars which is insensitive to photoperiod, our results suggest that ZS97 carries a non-functional allele at OsGI, thus the PS-allele carried by ZS97 does not response to photoperiod in the BC 2 F 7:8 populations.