Haplotype analysis of the genes encoding glutamine synthetase plastic isoforms and their association with nitrogen-use- and yield-related traits in bread wheat

Authors

  • Xin-Peng Li,

    1. The State Key Laboratory for Plant Cell and Chromosome Engineering, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Xue-Qiang Zhao,

    1. The State Key Laboratory for Plant Cell and Chromosome Engineering, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Xue He,

    1. The State Key Laboratory for Plant Cell and Chromosome Engineering, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, China
    Search for more papers by this author
  • Guang-Yao Zhao,

    1. Institute of Crop Science, Chinese Academy of Agricultural Sciences, Beijing 100081, China
    Search for more papers by this author
  • Bin Li,

    1. The State Key Laboratory for Plant Cell and Chromosome Engineering, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, China
    Search for more papers by this author
  • Dong-Cheng Liu,

    1. The State Key Laboratory for Plant Cell and Chromosome Engineering, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, China
    Search for more papers by this author
  • Ai-Min Zhang,

    1. The State Key Laboratory for Plant Cell and Chromosome Engineering, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, China
    Search for more papers by this author
  • Xue-Yong Zhang,

    1. Institute of Crop Science, Chinese Academy of Agricultural Sciences, Beijing 100081, China
    Search for more papers by this author
  • Yi-Ping Tong,

    1. The State Key Laboratory for Plant Cell and Chromosome Engineering, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, China
    Search for more papers by this author
  • Zhen-Sheng Li

    1. The State Key Laboratory for Plant Cell and Chromosome Engineering, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing 100101, China
    Search for more papers by this author

Authors for correspondence:
Yi-Ping Tong
Tel: +86 10 64844889
Email:
yptong@genetics.ac.cn

Zhen-Sheng Li
Tel: +86 10 64844889
Email:
zsli@genetics.ac.cn

Summary

  • Glutamine synthetase (GS) plays a key role in the growth, nitrogen (N) use and yield potential of cereal crops. Investigating the haplotype variation of GS genes and its association with agronomic traits may provide useful information for improving wheat N-use efficiency and yield.
  • We isolated the promoter and coding region sequences of the plastic glutamine synthetase isoform (GS2) genes located on chromosomes 2A, 2B and 2D in bread wheat. By analyzing nucleotide sequence variations of the coding region, two, six and two haplotypes were distinguished for TaGS2-A1 (a and b), TaGS2-B1 (af) and TaGS2-D1 (a and b), respectively.
  • By analyzing the frequency data of different haplotypes and their association with N use and agronomic traits, four major and favorable TaGS2 haplotypes (A1b, B1a, B1b, D1a) were revealed. These favorable haplotypes may confer better seedling growth, better agronomic performance, and improved N uptake during vegetative growth or grain N concentration.
  • Our data suggest that certain TaGS2 haplotypes may be valuable in breeding wheat varieties with improved agronomic performance and N-use efficiency.

Introduction

Glutamine synthetase (GS, EC 6.3.1.2) plays an essential role in the metabolism of nitrogen (N) by catalyzing the condensation of glutamate and ammonia to form glutamine (Miflin & Habash, 2002; Coruzzi, 2003). Higher plants have two types of GS – cytosolic (GS1) and plastic (GS2) isoforms – which perform nonoverlapping metabolic functions (Edwards et al., 1990). In diploid plants, GS1 is usually encoded by three to five genes, predominantly expressed in the vascular tissues, and involved in generating glutamine for intercellular N transport (Edwards et al., 1990; Pereira et al., 1995; Sakurai et al., 1996; Brugière et al., 1999; Tabuchi et al., 2005; Canovas et al., 2007; Bernard et al., 2008). GS2 is often encoded by a single gene in diploid species, and is located mainly in the chloroplasts and mitochondria of green tissues (Tingey et al., 1988; Taira et al., 2004). GS2 is the major isozyme in leaf mesophyll cells during vegetative growth, and assimilates ammonia originated from nitrite reduction and photorespiration (Wallsgrove et al., 1987; Tobin & Yamaya, 2001).

Accumulating evidence demonstrates that N assimilation catalyzed by GS is instrumental in determining the N status and yield potential of cereal crops (Hirel et al., 2001, 2005;Obara et al., 2001; Yamaya et al., 2002; Kichey et al., 2006; Martin et al., 2006). For example, in maize, coincidences between quantitative trait locus (QTL) for yield and the genes encoding GS1 and the corresponding enzyme activity were detected (Hirel et al., 2001; Galais & Hirel, 2004). Mutation and overexpression analyses of GS1 in maize further confirmed the major role of GS1 in controlling kernel yield and the remobilization of N to developing kernels (Martin et al., 2006). In rice, QTLs for leaf GS1 protein content have been found to co-locate with the chromosomal regions for various biochemical and physiological traits involving N recycling, and OsGS1.1 coding for GS1 was mapped to a QTL for spikelet weight (Yamaya et al., 2002; Obara et al., 2004). A recent study showed that knockout of OsGS1.1 severely retarded growth rate and grain filling of rice plants under normal N conditions (Tabuchi et al., 2005). In wheat leaves, GS activity was strongly correlated to soluble protein and N contents (Kichey et al., 2006; Habash et al., 2007). The leaf GS activity and protein content were found to exhibit genetic variations in wheat germplasm (Kichey et al., 2006; Habash et al., 2007; Bernard et al., 2008). Using a mapping population from the cross of Chinese Spring (CS) × SQ1, Habash et al. (2007) found that QTLs for leaf GS activity co-localized with those for leaf protein and N contents and yield components. Importantly, two QTL clusters for leaf GS activity were found to associate with mapped GS genes (a GS2 gene on chromosome 2A, and a GS1 gene on 4A) (Habash et al., 2007).

Although previous studies have revealed genetic variations of leaf GS activity in wheat germplasm, the underlying molecular mechanism remains unclear. One possibility is that the nucleotide sequence variations in GS genes may be responsible for the GS activity differences among different wheat materials. As the first step to study this hypothesis, we chose to focus on GS2 genes of wheat because these genes are comparatively the major contributors to green leaf GS activity in wheat (Habash et al., 2001). The objectives of this work were to characterize the GS2 genes in bread wheat; to analyze the haplotype variations of bread wheat GS2 genes; and to investigate the associations between GS2 haplotypes and the traits related to yield and N use.

Materials and Methods

Wheat genetic stocks

Three populations of Triticum aestivum L. were used in this study. The first population was the mini core collection (MCC) of Chinese wheat varieties. The MCC consisted of 231 accessions with an estimated 70% representation of genetic variation of Chinese wheat (Hao et al., 2008). In this study, the collection was expanded to 260 accessions. The second population was the double haploid (DH) population derived from two Chinese winter wheat varieties, Hanxuan 10 and Lumai 14, and has previously been evaluated the performance under different soil N concentrations (An et al., 2006). The last population consisted of 142 recombinant inbred lines (RILs) derived from two Chinese winter wheat varieties, Xiaoyan 54 and Jing 411.

Isolation of TaGS2 genes

All the PCR primers described in this paper were listed in the Supporting Information Table S1. In order to isolate the full-length genomic sequences of all the TaGS2 genes in the wheat genome, a bacterial artificial chromosome (BAC) library of the common wheat variety Xiaoyan 54 was screened with the consensus primer pair GS2F2/WLR1 designed according to the expressed sequence tags (ESTs) of TaGS2 in NCBI GenBank. The candidate BAC clones contained three different TaGS2 genes, namely TaGS2-A1, TaGD2-B1 and TaGS2-D1. Four primer pairs (GS2F2/WLR1, GS2F5/GS2R4, WLF1/F1spAR1 and F1spAF1/GS2R6) were used to clone the fragments of TaGS2-A1; and three primer pairs (GS2F2/WLR1, GS2F5/GS2R4 and WLF/GS2R6) were used to amplify the fragments of TaGS2-B1 and TaGS2-D1 from the candidate BAC clones. These fragments were then assembled to get the full-length sequences of the three TaGS2 genes in Xiaoyan 54. The amplified fragments from CS by the three primer pairs, GS2F2/WLR1, WLF/R6 and GS2F5/GS2R4, were used for TaGS2 full-length sequence assembly. The 3′-untranslated regions (3′-UTRs) of TaGS2s were amplified with primer pair GS2TF/WLR5. All these PCR products were amplified using Pfu DNA polymerase (Promega) and sequenced by ABI3730 DNA Analyzer (Applied Biosystems Inc., Foster City, CA, USA).

The 5′-flanking sequence of TaGS2-D1 was first cloned through inverse PCR from the BAC clone containing TaGS2-D1. The BAC plastid was digested by HindIII, and then self-ligated by T4 DNA ligase (New England BioLabs, Ipswich, MA, USA), and amplified by the primer pair GS2F8/WLFR. The amplified products were sequenced and rearranged according to HindIII sites. The forward primers P1, P2, P4.1 and P4.2 designed according to the 5′ upstream sequence of TaGS2-D1 were paired with reverse primer WLFR or TS2 to amplify homologous sequences from the BAC clones containing TaGS2 of Xiaoyan 54 and the genomic DNA of CS.

Analysis of TaGS2 haplotypes

Eco-TILLING (targeting induced local lesions in genomes) was employed to analyze the haplotypes of TaGS2-A1. Celery juice extract (CJE) used in Eco-TILLING was prepared as described by Till et al. (2004). The c. 780-bp 3′-end fragment of TaGS2-A1 amplified with primer pair GS2sp5/WLR1 was used in the Eco-TILLING for single nucleotide polymorphism (SNP) discovery in TaGS2-A1 in the MCC. DNA templates were amplified with Ex Taq™ DNA Polymerase (Takara, Dalian, China) in a reaction volume of 10 μl. The PCR product of each accession was mixed with 3 μl of the product of CS or Xiaoyan 54, respectively. The mixed PCR products were held at 99°C for 10 min and then decreased to 40°C gradually to form hybrid molecules. To cleave mismatches in the hybrid molecules, each tube containing the hybrid molecules was added with 0.3 μl CJE, 0.2 μl water and 1.5 μl 10 × buffer (100 mM MgSO4, 100 mM Hepes, pH 7.5, 100 mM KCl, 0.02% Triton X-100, 2 μg ml−1 BSA). After incubation at 45°C for 15 min, the reaction was terminated by adding 2 μl of 0.225 M EDTA. The cleaved fragments were detected by electrophoresis in 1.5% agarose gel.

A 239-bp insertion/deletion in the second intron of TaGS2-A1 was detected through PCR with primer pair WLF1/SPAR1. Primer pair GS2TF/WLR1 was employed to detect the Miniature Inverted Transposable Element (MITE) insertion in intron 12 of TaGS2-D1. These amplified fragments were electrophoresed on 1% (w/v) agarose gel and stained with ethidium bromide.

Primer pairs In1BF1/In1BR and In2F1/F6R were used for detecting the simple sequence repeats (SSRs) in intron 1 and the 33-bp indel insertion in intron 2 of TaGS2-B1, respectively. These amplified fragments were separated on 6% (w/v) denaturing polyacrylamide gels and were detected by silver staining.

Association analysis between TaGS2 haplotypes and growth and related traits

The Chinese wheat MCC and two biparental populations derived from bred wheat varieties were used to discover the associations between TaGS2 and growth and related traits. In all, 244 accessions of the MCC were evaluated and the shoot dry weight (SDW) and root dry weight (RDW) per plant under high N and low N conditions were deteremined in a hydroponic culture at seedling stage. The seeds of these MCC accessions were germinated at 25°C for 7 d, and then the germinated seedlings with residual endosperm removed were transferred to a tank (2.5 m long × 1.5 m wide × 0.5 m high) containing 1800 l of nutrient solution. The nutrient solution contained (mM) KH2PO4, 0.2; MgSO4, 0.5; KCl, 1.5; CaCl2, 1.5; H3BO3, 1 × 10−3; (NH4)6Mo7 O24, 5 × 10−5; CuSO4, 5 × 10−4; ZnSO4, 1 × 10−3; MnSO4, 1 × 10−3; Fe(III)-EDTA, 1 × 10−1. The high- and low-N nutrient solutions contained 1.0 and 0.05 mM Ca(NO3)2, respectively. To the low-N nutrient solution an additional 0.95 mM CaCl2 was added to balance the calcium concentration in the high- and low-N nutrient solutions. Each treatment had three replicates, and each replicate contained two plants per wheat accession. The wheat plants were grown in a glasshouse from 22 December 2005 to 10 March 2006. During the experimental period, the lowest daytime temperature in the glasshouse ranged from 0 to 6°C, and the highest daytime temperature varied from 10 to 30°C.

The DH population of ‘Hanxuan 10 × Lumai 14’ and the RIL population of ‘Xiaoyan 54 × Jing 411’ were used to discover agronomic and N-use-related traits associated with TaGS2 haplotypes. Phenotype evaluations of the DH population of ‘Hanxuan 10 × Lumai 14’ have been described previously (An et al., 2006).

The RIL population of ‘Xiaoyan 54 × Jing 411’ was evaluated in two consecutive years (harvested in 2006 and 2007) in the experimental stations of our institute located in Luancheng, Hebei province, and Changping, Beijing. The trials in Beijing and the trial in 2007 in Hebei province involved both low- (nil N) and high-N (180 kg N ha−1) treatments, and the field trial in 2006 in Hebei province involved only a low-N treatment. Each treatment in all the trials had three replicates. For each RIL in each replicate, 60 seeds were sown in two 1.5-m-long rows, and rows were spaced 23 cm apart. RILs in each replicate were randomly positioned. Biomass and grain yields were recorded for 30 representative plants. Total N concentration in straw and grain tissues in the 2006 trials were measured using a semiautomated Kjeldahl method (Tecator Kjeltec Auto 1030 Analyzer; Tecator, Hogänäs, Sweden).

Association analysis between TaGS2 and the investigated traits was performed using the General Line Model in SPSS11.5 for Windows (SPSS Inc., Chicago, IL, USA).

Bioinformatics

Sequence alignment, secondary structure analysis, and protein translation were performed using DNAMAN 6.0 for Windows (Lynnon Biosoft, Quebec, Canada). The transcription start site (TSS) of TaGS2, and the poly(A) splicing site of the upstream gene in the 5′-upstream region of TaGS2 were predicted by PlantProm database (Shahmuradov et al., 2003).

Results

Isolation of TaGS2 genes

We screened a BAC library (with 4.5-fold genome coverage) of the wheat cv Xiaoyan 54 using a pair of conserved primers for TaGS2, and obtained eight BAC clones. Three different TaGS2 sequences (namely TaGS2-A1, B1 and D1), each covering the 5′-flanking region and the open reading frame (ORF), were isolated from the eight BAC clones. Using genomic PCR, the orthologs of the three genes were also isolated from CS. The coding regions of the TaGS2 genomic sequences in Xiaoyan 54 and CS were 3.6–3.9 kb long, and each of them contained 13 exons and 12 introns (Fig. 1). The protein sequences deduced from the three genes were very similar (with > 99% identity). Subsequently, the three GS2 genes were mapped to the terminal deletion bins of 2AL1-0.85-1.00, 2BL6-0.89-1.00 and 2DL9-0.76-1.00 of CS, respectively (Fig. S1a). Owing to the great similarities in their exon/intron structures, predicted proteins and chromosomal locations, we suggest that TaGS2-A1, B1 and D1 are homoeologous GS2 genes located on wheat group two chromosomes.

Figure 1.

 Structure of TaGS2. ATG and TGA are the predicted start and stop codons, respectively.

Using two biparental populations, we further genetically localized TaGS2-A1 between the SSR markers Xgwm294 and Xwmc181 on chromosome 2A, and TaGS2-B1 linked with the SSR marker Xlhq259 on chromosome 2B in the ‘Xiaoyan 54 × Jing 411’ genetic map. TaGS2-D1 was mapped between the SSR markers WMC170 and CWM96.2 on chromosome 2D in the ‘Hanxuan 10 × Lumai 14’ genetic map (Fig. S1b).

Nucleotide sequence comparisons between allelic GS2 genes in CS and Xiaoyan 54

The genomic nucleotide sequences of allelic GS2 genes of CS and Xiaoyan 54 showed different levels of identity, 93.28% between the TaGS2-A1 alleles, 99.56% between the B1 alleles, and 85.74% between the D1 alleles. To facilitate further analysis, the three GS2 genes in CS were named TaGS2-A1a (accession no. GQ169684), TaGS2-B1a (accession no. GQ169687) and TaGS2-D1a (accession no. GQ169688), respectively, whereas their orthologous alleles in Xiaoyan 54 were designated as TaGS2-A1b (accession no. GQ169685), TaGS2-B1b (accession no. GQ169686) and TaGS2-D1b (accession no. GQ169689), respectively. In the coding regions of TaGS2 mRNA sequences of the two varieties, there were two SNPs between TaGS2-A1 haplotypes, no SNP between TaGS2-B1 haplotypes, and 29 SNPs between TaGS2-D1 haplotypes (Table S2). The deduced TaGS2 proteins of the two varieties were highly conserved, with the largest difference (two amino acid substitutions) found between the TaGS2-D1 proteins of the two varieties.

The intronic sequences of allelic TaGS2 genes in CS and Xiaoyan 54 showed substantial differences. For A1a and A1b intronic sequences, the differences included eight SNPs, two single nucleotide indels, one triple nucleotide (TTC) indel, and one 239-bp MITE insertion in A1b allele. The B1a and B1b intronic sequences differed by six SNPs and a deletion of the sequence element 5′-GATTAGATTA-3′ in intron 1. Surprisingly, the intronic sequences of D1a and D1b exhibited very large nucleotide sequence variations, which included three large indels and numerous SNPs, multiple nucleotide substitutions and small indels.

We also cloned the 5′-flanking sequences of TaGS2 genes from Xiaoyan 54 and CS. A gene encoding a JmjC domain-containing protein belonging to the cupin superfamily was found immediately upstream of TaGS2, with the distance between the predicted poly(A) site of the cupin gene and the putative start codon of TaGS2 varying from approx. 930 to 1280 bp. The 5′-flanking regions of A1a and A1b alleles differed by five SNPs and a 10-bp indel. However, no difference was found between the 5′-flanking regions of B1a and B1b alleles. The corresponding regions in D1a and D1b alleles differed from each other to a greater degree; they shared 80.43% identity, with the largest variation in the 5′ leader intron. A 347-bp insertion was observed in the 5′-flanking regions of B1a and B1b alleles but not in those of the A1a and A1b or D1a and D1b alleles. All the sequenced TaGS2 genes contained a leader intron between the predicted transcription start site and the start codon. The size (114–115 bp) and nucleotide sequence of the leader intron were highly conserved among TaGS2-A1a, A1b, B1c, B1b and D1a. However, the size of this intron in D1b (135 bp) was much greater (Fig. S2).

Bioinformatic analysis indicated that the three MITEs in the coding regions of A1a, D1a or D1b were members of the Stowaway family. In the promoter region of B1a and B1b there was a 347-bp inverted repeat insertion.

Haplotype analysis of TaGS2 genes in the MCC of Chinese wheat

The MCC of Chinese wheat was employed to discover more haplotypes of TaGS2 genes using the sequence information of CS and Xiaoyan 54 as references, and the detailed results are listed in Table S3. For TaGS2-A1, no new haplotypes (differing from A1a or A1b) were discovered based on the surveys of nucleotide information in two subregions of the genomic coding sequence, one containing exons 1 and 2 and the other covering exons 11, 12 and 13. Among the 254 accessions analyzed, 90 (35%) contained A1a, whereas 164 (65%) possessed A1b (Table 1). The frequency of A1a was significantly higher in historical landraces than in modern varieties. By contrast, A1b had a significantly higher frequency in modern varieties than in landraces.

Table 1.   Distribution frequencies of TaGS2 haplotypes in the MCC accessions
Wheat regionCategoryNo. of accessions
TaGS2-A1TaGS2-B1TaGS2-D1
ababcdefab
  1. Wheat regions: I, northern winter wheat region; II, Yellow and Huai River valley winter wheat region; III, Middle and Low Yangtze River valley winter wheat region; IV, southwestern winter wheat region; V, southern winter wheat region; VI, northeastern spring wheat region; VII, northern spring wheat region; VIII, northwestern spring wheat region; IX, Qinghai-Tibet spring-winter wheat region; X, Xinjiang winter-spring wheat region.

China ILandrace109910500181
Bred variety112371100121
IILandrace211215011700330
Bred variety12161870000254
IIILandrace4181630100203
Bred variety0842002080
IVLandrace11112000120220
Bred variety1624000161
VLandrace0540000050
Bred variety1321000040
VILandrace0632100051
Bred variety0750200032
VIILandrace2770020081
Bred variety2241000050
VIIILandrace681030001121
Bred variety3543000062
IXLandrace491200000130
Bred variety1202000012
XLandrace55710200100
Bred variety2120011040
ChinaLandrace639010310228211467
Bred variety2362442732317412
Introduced variety 412950010124
Total 90164156425306223223

In the case of TaGS2-B1, four new haplotypes (named as B1c, B1d, B1e and B1f, respectively) were identified by analyzing nucleotide variation in introns 1 and 2. The variation of intron 1 was caused by the differential presence of the repeated sequence element GATTA (Fig. S3). In intron 2, a 33-bp element was directly repeated in some accessions. In the 241 accessions analyzed, B1a, b and d were the major haplotypes, while B1c, e and f were the rare haplotypes in the MCC (Table 1). The frequency of B1b was significantly higher in modern varieties than in landraces. However, both B1a and B1d exhibited lower frequency in bred varieties than in landraces.

For TaGS2-D1, no additional haplotypes were found in the MCC by investigating the nucleotide sequence variation in intron 12. Among the 255 accessions analyzed, the great majority (232) contained D1a, whereas only 23 lines possessed D1b. The frequency of D1a was very high (above 85%) in both landraces and modern varieties.

Association analysis between TaGS2 haplotypes and growth-related traits

The MCC accessions were grown in hydroponic culture with either high- or low-N supplies, and SDW and RDW per plant were evaluated at the seedling stage (Table 2). Among the landraces, the accessions with TaGS2-A1b had, on average, significantly higher SDW and RDW than those with TaGS2-A1a under either high- or low-N conditions, whereas the lines containing TaGS2-B1a showed, on average, significantly higher SDW and RDW than those with TaGS2-B1d under both N supply conditions (Table 2). Among the bred varieties, the accessions possessing TaGS2-A1b displayed, on average, significantly higher SDW than those with TaGS2-A1a under low-N conditions, while the lines with TaGS2-B1a exhibited significantly higher SDW than those with TaGS2-B1b under high-N conditions (Table 2).

Table 2.   Traits associated with TaGS2-A1or B1 haplotypes detected in the mini core collection (MCC) accessions grown under low- and high-N conditions in hydroponic culture at seedling stage
TraitTreatmentLandraceBred variety
TaGS2-A1 haplotypeTaGS2-B1 haplotypeTaGS2-A1 haplotypeTaGS2-B1 haplotype
a (n = 59)b (n = 83)a (n = 92)d (n = 27)a (n = 22)b (n = 61)a (n = 44)b (n = 26)
  1. SDW, shoot DW; RDW, root DW.

  2. The MCC accessions were subdivided into two categories (landrace and bred variety) for association analysis. Statistical difference between TaGS2-A1 haplotypes (a, b) detected in the landrace or bred variety accessions is indicated by different letters after the means. The statistical differences between TaGS2-B1 haplotypes detected in the landrace (a, d) or bred variety (a, b) accessions are indicated in a similar way. Capital and small letters designate significance at P < 0.01 and P < 0.05, respectively.

  3. n, number of accession.

  4. Values are given as means ± SD.

SDW (g per plant)LowN0.354 ± 0.212B0.488 ± 0.229A0.487 ± 0.225A0.235 ± 0.133B0.428 ± 0.162B0.548 ± 0.254A0.524 ± 0.2410.495 ± 0.222
HighN0.858 ± 0.534B1.135 ± 0.515A1.165 ± 0.472A0.501 ± 0.274B1.009 ± 0.3721.141 ± 0.6171.235 ± 0.613a0.914 ± 0.412b
RDW (g per plant)LowN0.233 ± 0.104B0.297 ± 0.104A0.289 ± 0.104A0.196 ± 0.091B0.328 ± 0.2840.317 ± 0.1110.306 ± 0.1010.351 ± 0.271
HighN0.162 ± 0.098B0.216 ± 0.108A0.217 ± 0.096A0.107 ± 0.063B0.209 ± 0.0790.228 ± 0.1170.238 ± 0.1040.190 ± 0.093

The MCC accessions were collected from 10 ecological regions in China (Table 1). Differences in the regions’ growth habit (spring and winter), photoperiodic response, and freezing tolerance may exert significant influence on agronomic performance and N use under field conditions. Therefore, the RIL population of ‘Xiaoyan 54 × Jing 411’ and the DH population of ‘Hanxuan 10 × Lumai 14’ were used to discover agronomic and N-use-related traits associated with TaGS2 haplotypes under field conditions. The wheat variety Jing 411 had TaGS2-A1a and B1a haplotypes, while Xiaoyan 54 harbored A1b and B1b haplotypes, so the RIL population of ‘Xiaoyan 54 × Jing 411’ was employed to discover traits associated with A1 and B1 haplo-types. Among the investigated traits of biomass yield per plant (BY), grain weight per plant (GW), ear number per plant (EN), 1000-grain weight (TGW), grain number per ear (GNE), and straw and grain N concentrations in the 2006 and 2007 field trials in Hebei and Beijing, EN, TGW and GNC were found to associate with TaGS2-A1, and BY, GW, EN and TGW were found to associate with TaGS2-B1 (Table 3). For TaGS2-A1 haplotypes, A1b associated with higher TGW under low-N conditions in three trials and higher grain N concentration in two trials, while A1a associated with higher EN under low-N conditions in two trials. For TaGS2-B1 haplotypes, B1b associated with higher BY under low- or high-N conditions in three trials, higher GW under low- or high-N conditions in two trials, and higher EN and TGW under low-N conditions in one trial.

Table 3.   Traits associated with TaGS2-A1 and B1 haplotypes in the recombinant inbred line (RIL) population of ‘Xiaoyan 54 × Jing 411’
TraitTrialTreatmentTaGS2-A1 haplotypeTaGS2-B1 haplotype
a (n = 44)b (n = 87)a (n = 49)b (n = 89)
  1. n, number of accessions; BY, biomass yield per plant; GW, grain weight per plant; EN, ear number per plant; TGW, 1000-grain weight; BJ, Beijing; HB, Hebei province.

  2. The RIL lines were grouped according to their TaGS2-A1 (a, b) and B1 (a, b) haplotypes. Statistical difference between the two haplotypes is indicated by different letters after the means. Capital and small letters designate significance at P < 0.01 and P < 0.05, respectively. Biomass yield was not recorded in the trial in 2007 in Hebei province, and grain N concentration was determined only in 2006 field trials.

BY (g per plant)2006 BJHigh N31.4 ± 5.031.9 ± 6.130.2 ± 5.0b32.5 ± 6.0a
Low N28.7 ± 5.827.4 ± 4.926.9 ± 5.928.3 ± 5.0
2006 HBLow N20.2 ± 3.420.0 ± 3.719.3 ± 3.7b20.7 ± 3.6a
2007 BJHigh N20.8 ± 3.221.7 ± 3.420.5 ± 3.2b21.8 ± 3.3a
Low N18.5 ± 2.718.9 ± 2.818.0 ± 2.7b19.2 ± 2.7a
GW (g per plant)2006 BJHigh N12.3 ± 2.612.5 ± 2.711.9 ± 2.312.5 ± 2.8
Low N11.9 ± 3.311.5 ± 2.611.2 ± 3.211.8 ± 2.7
2006 HBLow N10.4 ± 1.610.1 ± 1.99.8 ± 1.8b10.5 ± 1.8a
2007 BJHigh N9.2 ± 1.59.3 ± 1.88.7 ± 1.6b9.4 ± 1.8a
Low N8.1 ± 1.28.0 ± 1.47.7 ± 1.3b8.2 ± 1.4a
2007 HBHigh N10.6 ± 1.910.0 ± 2.19.9 ± 1.810.4 ± 2.2
Low N8.4 ± 2.07.8 ± 2.17.5 ± 2.18.3 ± 2.1
EN (ear per plant)2006 BJHigh N12.7 ± 2.212.7 ± 2.512.4 ± 2.313.0 ± 2.5
Low N11.3 ± 1.9A10.1 ± 1.5B10.4 ± 1.910.6 ± 1.7
2006 HBLow N5.9 ± 1.25.8 ± 1.15.6 ± 1.16.0 ± 1.1
2007 BJHigh N7.7 ± 1.17.8 ± 1.27.6 ± 1.27.9 ± 1.1
Low N6.6 ± 1.06.6 ± 0.86.4 ± 0.7b6.7 ± 0.9a
2007 HBHigh N7.5 ± 1.57.4 ± 1.17.4 ± 1.17.5 ± 1.3
Low N6.3 ± 1.2a5.8 ± 1.2b5.8 ± 1.46.1 ± 1.1
TGW (g)2006 BJHigh N36.3 ± 4.438.0 ± 4.537.3 ± 4.137.7 ± 4.9
Low N37.9 ± 5.6B40.9 ± 3.9A38.8 ± 5.440.4 ± 4.3
2006 HBLow N42.8 ± 3.7b44.3 ± 4.0a42.9 ± 3.7b44.3 ± 4.0a
2007 BJHigh N41.7 ± 4.143.0 ± 4.142.0 ± 4.442.9 ± 3.9
Low N42.2 ± 3.5b43.9 ± 3.8a42.6 ± 3.643.7 ± 3.7
2007 HBHigh N41.4 ± 3.642.5 ± 3.741.9 ± 3.642.4 ± 3.9
Low N43.1 ± 3.944.3 ± 4.443.9 ± 4.744.0 ± 3.9
Grain N concentration (%)2006 BJHigh N2.34 ± 0.212.43 ± 0.18a2.38 ± 0.172.41 ± 0.21
Low N2.40 ± 0.162.51 ± 0.20A2.48 ± 0.192.48 ± 0.20
2006 HBLow N2.28 ± 0.172.38 ± 0.23a2.38 ± 0.232.34 ± 0.22

The wheat varieties Hanxuan 10 and Lumai 14 had different haplotypes for TaGS2-D1. The performance of the DH population derived from these two varieties has been evaluated previously (An et al., 2006). In this study, performances of this population were measured in two field trials harvested in 2003 and 2004 under low- and high-N conditions in the experiment station of our institute located in Changping, Beijing. The following traits were used in the association analysis: BY, GW, EN, TGW, GNE, N uptake per plant (NUP) before and after flowering, and straw and grain N concentrations. Association analysis showed that TaGS2-D1a was associated with significantly higher TGW under both low- and high-N conditions, and with the amount of N uptake before flowering under high-N conditions in both years. By contrast, no statistically significant difference was found between the two haplotypes in the amount of N uptake after flowering under either set of N conditions (Table 4). Neither of the two TaGS2-D1 haplotypes showed any significant association with BY, EN, GNE, and straw and grain N concentrations (data not shown).

Table 4.   Traits associated with TaGS2-D1 haplotypes in the double haploid (DH) population of ‘Hanxuan 10 × Lumai 14’
TraitYearTreatmentTaGS2-D1a (n = 73)TaGS2-D1b (n = 47)
  1. n, number of accession; TGW, 1000-grain weight.

  2. The DH lines were grouped according to their TaGS2-D1 haplotypes (a, b). Statistical difference between the two haplotypes is indicated by different letters after the means. Capital and small letters designate significance at P < 0.01 and P < 0.05, respectively.

  3. Values are given as means ± SD.

TGW (g)2003Low N40.3 ± 3.2A37.1 ± 3.4B
2003High N40.3 ± 3.5A37.9 ± 3.4B
2004Low N41.9 ± 3.6A39.1 ± 3.7B
2004High N41.9 ± 3.8A38.9 ± 3.3B
N uptake before flowering (mg per plant)2003Low N88.3 ± 14.787.5 ± 15.8
2003High N123.9 ± 19.8a114.6 ± 20.4
2004Low N70.3 ± 14.967.9 ± 14.4
2004High N105.3 ± 17.5a98.2 ± 20.1
N uptake after flowering (mg per plant)2003Low N19.2 ± 13.818.5 ± 12.1
2003High N19.5 ± 16.722.1 ± 13.2
2004Low N8.7 ± 10.810.2 ± 10.7
2004High N15.3 ± 15.815.9 ± 15.7

Discussion

Chromosomal location and gene structure of TaGS2 genes

The present study identified three TaGS2 genes (TaGS2-A1, B1 and D1) in two Chinese bread wheat varieties (Xiaoyan 54 and CS). These three genes were physically and genetically localized to the distal ends of chromosome arms 2AL, 2BL, and 2DL (Fig. S1), respectively. Previous studies have identified three TaGS2 genes (TaGS2a, b and c) in European wheat and assigned them to the distal bins of chromosome arms 2AL, 2DL and 2BL, respectively (Habash et al., 2007; Bernard et al., 2008). The high similarity in gene sequence and chromosomal location indicates that TaGS2-A1, B1 and D1 were allelic to TaGS2a, c and b, respectively. Together, these data suggest the presence of three TaGS2 genes on the long arms of group two chromosomes in a wide range of bread wheat varieties. The present research also revealed strong collinearity between the GS2 loci of wheat and rice. The chromosomal locations of wheat and rice GS2 genes were syntenic. Moreover, in both species, the GS2 gene was located immediately downstream of a cupin gene (Fig. S1a).

The three TaGS2 genes all had 12 introns and 13 exons in their coding region (Fig. 1), sharing a similar gene structure with rice GS2. The 5′-flanking sequence of TaGS2-A1, B1 and D1 genes contained a 5′-UTR leader intron (Fig. S2). A 5′-UTR leader intron also exists in the GS2 genes of rice, Arabidopsis and pea (Tjaden et al., 1995), indicating that it may be a highly conserved element in higher plant GS2 genes. It has been reported that 12–17% of plant 5′-UTRs carry introns (Pesole et al., 2001; Hong et al., 2006). Furthermore, it is known that 5′-UTR introns can enhance gene expression several- to 100-fold (Fu et al., 1995; Wang & Oard, 2003; Kim et al., 2006; Sivamani & Qu, 2006), and regulate tissue-specific gene expression pattern (Gadea et al., 1999; Plesse et al., 2001). Consequently, it will be interesting to investigate if the 5′-UTR intron affects GS2 gene expression in wheat in future research.

We identified three MITE insertions in the introns of the TaGS2-A1 and D1 genes (Fig. 1) and an unknown inverted repeat element in the promoter of B1. The former observation is consistent with the overwhelming finding that MITEs are generally situated close or within genes (Casacuberta & Santiago, 2003; Benjak et al., 2009). Because it has been demonstrated that MITE insertions in intronic regions can affect gene expression through influencing RNA splicing (Casacuberta & Santiago, 2003; Rostoks et al., 2006), further work is needed to verify whether the MITEs located in TaGS2 introns may affect the spicing of their host genes. Interestingly, the 347-bp unknown inverted repeat was observed only in the 5′-flanking region of TaGS2-B1 but not those of TaGS2-A1 or D1. This element might be useful for investigating the unique features of TaGS2-B1 evolution and/or expression in future.

TaGS2 haplotypes and their association with growth- and yield-related traits

Nucleotide sequence comparisons of TaGS2 genes from CS and Xiaoyan 54 revealed many types of variation between allelic TaGS2 genes, including SNPs, multiple nucleotide substitutions, and small and large indels. These variations mainly existed in the 5′-flanking and intronic regions, while in the exonic regions, only SNPs existed between allelic TaGS2-A1 or D1 genes. However, only the SNPs between the allelic D1 genes caused a low amount of amino acid substitutions, indicating that the protein sequences of allelic TaGS2 enzymes may be highly conserved in bread wheat.

By utilizing nucleotide sequence polymorphisms in the intronic regions, this work distinguished two, six and two haplotypes for TaGS2-A1, B1 and D1, respectively, in the MCC accessions of Chinese wheat (Table 1). The MCC was estimated to represent approx. 70% genetic diversity of Chinese wheat varieties (Hao et al., 2008). Hence the diversities of TaGS2 genes revealed in the MCC may reflect the main variations of TaGS2 genes in Chinese wheat germplasm.

From the association data obtained in this work, it appears that TaGS2-A1b may play a more important role in early growth performance than A1a in landraces, irrespective of N supply levels; however, the beneficial effect of this haplotype in bred varieties was only observed under low-N conditions in the MCC population (Table 2). Interestingly, in the RIL population derived from two bred varieties, Xiaoyan 54 and Jing 411, significant trait–TaGS2-A1 associations were predominantly detected under low-N conditions (Table 3). These results indicate that TaGS2-A1 is a candidate locus for breeding wheat with improved agronomic performance and N use under low-N conditions. Among the three TaGS2-B1 haplotypes investigated, B1a was found to benefit early growth performance in landraces in the MCC population (Table 2), and B1b was linked with better agronomic performance in the RIL population of ‘Xiaoyan 54 × Jing 411’ (Table 3). The latter result may partially explain the increasing frequency of B1b in bred varieties from the northern winter wheat region (in which Beijing is located) and the Yellow and Huai River valley winter wheat regions (in which Hebei province is located) (Table 1). TaGS2-D1a shows a significant association with higher TGW and NUP before flowering in the DH population of ‘Hanxuan 10 × Lumai 14’ (Table 4). However, neither D1a nor D1b exhibits clear association with the NUP after flowering. This finding is consistent with a previous investigation showing that GS2 is the major isozyme in the leaf tissues during vegetative growth in wheat (Bernard et al., 2008).

In the literature, a number of QTLs for agronomic traits related to N use and yield have been mapped to the chromosomal regions containing GS2 in both wheat and rice. For example, QTLs for TGW have been reported to link with the markers near rice GS2 (Zhuang et al., 1997), and TaGS2-D1 in wheat (Yang et al., 2007). QTLs for wheat grain protein content are linked to the markers near TaGS2-A1 (Laperche et al., 2007) and TaGS2-D1 (Prasad et al., 1999). QTLs for wheat leaf GS activity and soluble protein content co-localize with TaGS2-A1. QTLs for rice leaf soluble protein content (Obara et al., 2001; Yamaya et al., 2002) and chlorophyll content (Wu & Luo, 1996) were linked to the marker near GS2. Collectively, the above data suggest that the genomic region surrounding GS2 may be valuable for breeding rice and wheat varieties with improved agronomic performance and N-use efficiency. The association analysis data generated in this work also support the importance of GS2 genes in wheat growth, N uptake before flowering, and yield traits. Furthermore, we revealed four functionally superior GS2 haplotypes (A1b, B1a, B1b, D1a) using the MCC population of Chinese wheat or genetic populations derived from bred varieties. Importantly, according to the frequency data in Table 1, A1b prevails over A1a; B1a and B1b are dominant over the remaining four B1 haplotypes; and D1a is more widespread than D1b. Together, our data indicate that A1b, B1a, B1b and D1a have been the favorable GS2 haplotypes in wheat breeding in China.

In summary, this study provides more detailed information on TaGS2 genes in terms of genomic sequence and physical and genetic position. The molecular diversity of TaGS2-A1, B1 and D1 in the MCC population of Chinese wheat was revealed, and favorable TaGS2 haplotypes were uncovered. These data may provide useful information for further defining the physiological function of TaGS2 genes in bread wheat, and for the breeding of wheat varieties with improved N-use efficiency and yield potential.

Acknowledgements

We thank Professors Xueyong Zhang and Jizeng Jia (Institute of Crop Science, Chinese Academy of Agricultural Sciences) for providing the wheat micro-core collections, and Prof. Hongqing Ling (Institute of Genetics and Developmental Biology, Chinese Academy of Sciences) for providing the BAC library of Xiaoyan 54. This research was supported by the Ministry of Science and Technology of China (2010CB125900, 2005CB120904, 2006AA02Z142, 2006AA10A105 and KSCX2-SW-304), the National Natural Science Foundation of China (30521001), and the Chinese Academy of Sciences (KSCX1-YW-03 and KSCX2-YW-N-001).

Ancillary