指环指泛素 E3 连结基因 TaSDIR1-4A 有助于确定普通小麦的粒状大小

Introduction

Given rapid increases in population together with climate change, we could soon witness global food shortages (Boyer, 1982; Godfray et al., 2010; Tilman et al., 2011). Grains are the most important food sources for humans (Zuo and Li, 2014; Li et al., 2018), and grain size is an important com- ponent of grain weight. Therefore, an understanding of the mechanisms underlying seed development is important for

improvement of grain yield in crop species. Recent studies have shown that ubiquitin-proteasome pathways are in- volved in control of seed size (Li and Li, 2016; Nadolska- Orczyk et al., 2017; Kelley, 2018; Li et al., 2019b; Xu and Xue, 2019), and one of these, the DA1 pathway, plays a key role in controlling seed size in Arabidopsis by regulating cell proliferation in the integuments (Disch et al., 2006; Xia

© The Author(s) 2020. Published by Oxford University Press on behalf of the Society for Experimental Biology.

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.

et al., 2013a; Du et al., 2014; Li and Li, 2016; Dong et al., 2017). This pathway of control of seed size is highly con- served among species (Su et al., 2011; Wang et al., 2017; Xie et al., 2018). For example, Grain Width and Weight2 (GW2), a major quantitative trait locus (QTL) in rice, is a homolog of DA2, a RING-type E3 ubiquitin ligase gene (Song et al., 2007; Xia et al., 2013a). GW2 homologs in maize and wheat are also involved in the control of grain size (Li et al., 2010b; Bednarek et al., 2012; Zhang et al., 2018).

Another RING-type E3 ubiquitin ligase, salt and drought- induced RING finger1 (SDIR1), has been shown to primarily participate in stress responses in numerous species, including Arabidopsis (Zhang et al., 2007, 2015), maize (Xia et al., 2012; Liu et al., 2013), rice (Gao et al., 2011), tobacco (Xia et al., 2013b), and grape (Tak and Mhatre, 2013), while others have been identified to not only be involved in stress responses, but also in plant growth and grain development (Sun et al., 2019). In wheat, there are two AtSDIR1 homologs,TaSDIR1-4A and TaSDIR1-4D, and the latter might be involved in abiotic stress responses (Wang et al., 2018). To date, the role of TaSDIR1 in grain size development in wheat is unknown.

Based on gene-sequence polymorphisms, candidate gene as- sociation analysis has been shown to be an efficient approach to reveal relationships between gene/polymorphic loci and traits, and it has been widely used in many crop species, including maize (Wang et al., 2016b) and wheat (Zhang et al., 2017). A series of elite alleles and molecular markers associated with important traits have been discovered using this method (Cao et al., 2020), and pyramiding these elite alleles through marker- assisted selection (MAS) will greatly accelerate the process of wheat breeding. Hence, finding elite alleles associated with im- portant agronomic traits is considered as being fundamental to wheat genetic improvement.

In the present research, we identified the RING-type E3 ubiquitin ligase gene TaSDIR1-4A and performed association analysis using multiple wheat populations, which demonstrated that it was significantly associated with 1000-grain weight.The favorable haplotype/allele has been positively selected for in Chinese wheat breeding programs, and the functional marker that we have developed for this haplotype will be beneficial in future programs.

Materials and methods

Plant materials

The common wheat (Triticum aestivum) cultivar Hanxuan 10 (H10) was used for cloning and structural analysis of the TaSDIR1 genes. A diverse wheat population of 32 accessions previously screened with 209 SSR markers (Zhang et al., 2013) was used for gene sequencing and detection of nucleotide polymorphisms (Supplementary Table S1 at JXB online).

Four wheat populations were used in this study, as follows. Population 1 was a doubled-haploid (DH) population with 150 lines developed from the cross H10 × Lumai 14 (L14) and was used for genetic mapping. H10 was selected from more than 20 000 wheat accessions due to its strong drought tolerance, whilst L14 is a high-yielding cultivar adapted to well- watered and fertile conditions, and was widely grown in northern China during the 1990s. Population 2 (262 accessions) was used for association analysis and consisted of 209 modern varieties, 43 advanced lines, and 10 landraces (Zhang et al., 2013). Population 3 comprised 157 landraces, and

Population 4 comprised 348 modern cultivars mainly from a Chinese wheat mini-core collection that represents more than 70% of the genetic diversity of the entire Chinese germplasm collection (Hao et al., 2011). Populations 3 and 4 were also used for determination of haplotype fre- quencies in 10 different wheat production zones covering the whole of China (Zhang et al., 2002).

Growth conditions and measurement of agronomic traits

Population 1 was planted in 10 environments (year×site×water re- gime combinations) at Shunyi (40°230′N; 116°560′E) and Changping (40°130′N; 116°130′E) in Beijing during the 2015–2016 and 2016– 2017 growing seasons. Population 2 was planted in 10 environments (year×site×water regime×heat stress combinations), including the same sites as Population 1, during 2010–2011, 2011–2012, and 2012–2013 for measurement of the following agronomic traits: plant height (PH), spike length (SL), peduncle length (PLE), length of penultimate node (LPN), number of spikes per plant (NSP), number of spikelets per spike (NSS), number of grains per spike (NGS), and 1000-grain weight (TGW). Two water regimes were applied at each site, namely well-watered (WW) and rain-fed (drought stress, DS). The WW plots were irrigated with 750 m3 ha–1 (75 mm) water at pre-overwintering, jointing, flowering, and grain-filling, whereas the DS plots were rain-fed only. The rainfalls were 131 mm in the 2010–2011 growing season, 180 mm in 2011–2012,

158 mm in 2012–2013, 161 mm in 2015–2016, and 173 mm in 2016–

2017. A heat-stress (HS) treatment was applied 1 week post-anthesis at Shunyi by positioning a plastic film supported by steel frames over the plots in the field (Li et al., 2019a). Populations 3 and 4 were planted in three environments, at Luoyang (34°610′N; 112°450′E) in Henan prov- ince in 2002 and 2005, and at Shunyi in 2010. In all cases, each accession was planted in a 2-m, 4-row plot with a row spacing of 30 cm and 40 seeds per row. Five plants in the middle of each plot were sampled for phenotyping at maturity.

Isolation of TaSDIR1-4A and construction of a phylogenetic tree

The reference sequence of TaSDIR1-4A was obtained from the URGI website (https://wheat-urgi.versailles.inra.fr/). Genome-specific pri- mers were designed based on the gene sequence (all primers are listed in Supplementary Table S2). Genomic DNA and cDNA from cv. H10 were used as templates. The gene structure of TaSDIR1-4A was determined using Lasergene 7.1.0 (DNASTAR, Inc., Madison, WI, USA) by align- ment of the amplified cDNA and genomic DNA sequences.A neighbor- joining phylogenetic tree was constructed using the MEGA software v5.2 (https://www.megasoftware.net/).

E3 ubiquitin ligase activity assays

The full-length TaSDIR1-4A ORF (840 bp) was cloned into a pMAL- C5X vector that has a maltose-binding protein (MBP) tag through seamless DNA cloning using In-Fusion® HD Cloning Plus (638910, TaKaRa). A TaSDIR1-4A mutant (His-244 to Tyr) containing a muta- tion in the RING finger domain was constructed using a Site-directed Mutagenesis Kit (B639281, Sangon Biotech, Beijing) according to the manufacturer’s protocol. MBP, MBP-AtSDIR1, MBP-TaSDIR1-4A, and MBP-TaSDIR1-4AH244Y were prepared and purified in Escherichia coli. MBP was used as a negative control, and MBP-AtSDIR1 was the positive control. In vitro E3 ligase assays were performed as described by Xie et al. (2002). Briefly, E1 (from wheat, 50 ng), E2 (UBCh5b, 100 ng), E3 (1μg), and 6×His tag ubiquitin (Ub, 4 μg) were mixed and incubated at 30 °C for 60 min. The mixture was separated by SDS-PAGE and blotted onto PVDF membranes (Millipore, IPVH00010). Anti-Ub and anti-MBP antibodies were used and bands were detected using a Thermo Pierce ECL (NCI4106) according to the manufacturer’s instructions.

Sequence polymorphisms in TaSDIR1-4A

Nucleotide polymorphisms in TaSDIR1-4A sequences were detected by scanning a small diverse panel of 32 common wheat accessions.

Genomic DNA was extracted from young leaves of 10-d-old seed- lings using the CTAB method. TaSDIR1-4A fragments were amplified using primers specific to the A genome. Target bands were purified and cloned into the pEASY-Blunt vector, and transformed into Trans1-T1 phage-resistant chemically competent cells (CD501-03, Beijing Trans Gen Biotech Co., Ltd.) using the heat-shock method. Twelve positive clones were randomly selected for sequencing using a 3730 XI DNA Analyzer (ABI). TaSDIR1-4A nucleotide se- quence polymorphisms were obtained using the SeqMan program in Lasergene 7.10 (DNASTAR).

Functional marker development

A functional marker was developed based on a SNP (G/A) at position

–395 bp, by mismatching bases in the dCAPS primer to create a re- striction site. The first-round PCR product obtained by using primers specific to the A genome was used as a template for the second round of PCR, which was performed as follows: 95 °C for 5 min, followed by 33 cycles of 95 °C for 30 s, annealing 58 °C for 30 s, and extension at 72 °C for 30 s, with a final extension of 72 °C for 10 min. The PCR products were digested with restriction enzyme at 37 °C for 2 h and then separ- ated by electrophoresis in 4% agarose gels.

Chromosome location and linkage analysis

Diploid and tetraploid wheat relatives with various genomes and a set of nullisomic-tetrasomic lines of the common wheat cultivar Chinese Spring were used to identify the genomic origin and chromosome location of TaSDIR1-4A.The H10 × L14 DH population, and the IciMapping V4.1 software (www.isbreeding.net/software/) were used for gene mapping and linkage analysis.

Population structure and association analysis

The population structure of Population 2 has previously been scrutinized with STRUCTURE v2.3.4 using data from 209 SSR markers evenly distributed on the 21 chromosomes in wheat (Li et al., 2012b). The two subpopulations comprised 110 and 152 accessions. Association mapping was performed using the mixed linear model in TASSEL v2.1 (https:// www.maizegenetics.net/tassel), which accounted for population struc- ture. One-way ANOVA was performed in SPSS 19.0 to identify signifi- cant associations between gene haplotypes and agronomic traits. Tukey tests were used to determine significant differences at P=0.05.

Virus-induced gene silencing (VIGS)

A series of recombinant barley stripe mosaic virus (BSMV) vectors was constructed.A non-conserved TaSDIR1-4A fragment of 253 bp spanning the 3´-terminal sequence and 3´-UTR region was selected to construct the BSMV vector BSMV:TaSDIR1-4A.This target region was compared with the Chinese Spring genome assembly v1.0 sequence (http://plants. ensembl.org/Triticum_aestivum/Tools/Blast) to ensure its specificity. A BSMV vector containing the GFP sequence was used as a control. To construct recombinant BSMV, the BSMV-γ vector was digested at the Not I restriction site, and the TaSDIR1-4A fragment amplified from the cDNA of Chinese Spring was inserted into the vector by seamless DNA cloning. The BSMV was inoculated onto leaves of jointing wheat plants following the procedures described in Wei et al. (2019). Flag leaves from BSMV-infected plants were collected at anthesis for determination of gene expression patterns. At maturity, seeds were collected, dried, and kernel size and weight were determined on samples of at least 200 seeds using an automatic seed measurement device (SC-A1, WSeen). Results were obtained from independent experiments.

Production of transgenic Arabidopsis TaSDIR1-4A-overexpression lines

Seeds of the sdir1-1 (SALK_052702) mutant were obtained from the Arabidopsis Biological Resource Center (https://abrc.osu.edu/. The

full-length cDNA of TaSDIR1-4A with the 35S promoter was inserted into the modified pCAMBIA1300 vector at the Xba I and Kpn I restriction en- zyme sites. Agrobacterium harboring TaSDIR1-4A and empty vectors were transferred into wild-type and sdir1-1 Arabidopsis flowers using the floral dip method to obtain transgenic lines (Clough and Bent, 1998). The seed sizes of homozygous T3 generation plants were measured using a SC-A1 device.

Expression of TaSDIR1-4A in wheat

The H10 cultivar was used for expression pattern analysis of TaSDIR1-4A. At the flowering stage, spikes, flag leaves, stems, nodes, root tissue at the stem base, and root tissues at different depths were sampled for spatial- temporal analysis of expression patterns (Wang et al., 2019).

Twelve accessions of each of the two different haplotypes were ran- domly selected from each of Populations 1 and 2 (Supplementary Table S3). Flag leaves at anthesis were sampled from field-grown plants and used to assess the expression levels.

Based on pilot experiments, 2-week-old wheat seedlings were sprayed with 50 μM ABA solution or treated at 250 mM NaCl for 0.5–72 h. Whole plants were then sampled for detection of stress-induced expres- sion of TaSDIR1-4A.

Yeast one-hybrid assays

Yeast one-hybrid assays were performed to check the binding of the ethylene response factor DNA binding domain (TaERF-BD) to the TaSDIR1-4A promoter region. The pB42AD vector was digested with restriction enzyme Not I in order to insert the TaERF-BD. The TaERF-BD fragment was then cloned into the vector through seamless DNA cloning. Six fragments of the TaSDIR1-4A promoter region were cloned into pLacZi as the reporter gene plasmid as follows. For the H10 construct, a 588-bp (–757 to –170 bp) fragment from the TaSDIR1-4A promoter region of H10 (Hap-4A-1) was amplified and cloned into pLacZi by seamless DNA cloning. For the L14 construct, a 586-bp (–755 to –170 bp) fragment (Hap-4A-2) was amplified and cloned using the same method. For the ACC-H10 Box construct, the target frag- ment contained the triplicates of the ACC Box (32 bp, –418 to –387 bp from H10). For the GCC-H10 Box construct, the same fragment with an A-to-G mutation was used. For the GCC-L14 Box construct, the target fragment contained the triplicates of the GCC Box (32 bp, –418 to

–387 bp from L14). For the ACC-L14 box construct, the same fragment with a G-to-A mutation was used. These four vectors were separately cloned into pLacZi by whole-gene synthesis at the Beijing Genomics Institute (BGI). Plasmids for the pB42AD vector with the TaERF-BD were co-transformed with the six reporter gene constructs into yeast strain EGY48 using standard yeast transformation methods (Lin et al., 2007).The yeast was incubated on media lacking Ura and Leu with x-α- gal. Interactions between TaERF-BD and each promoter fragment were tested by LacZ staining (Wang et al., 2016a).

Purification of the TaERF-BD protein and electrophoretic mobility shift assays (EMSAs)

TaERF-BD cDNA (219 bp) was fused into the pGEX-4T1 vector, which has a glutathione S-transferase (GST) tag, using seamless DNA cloning at an EcoR I site using In-Fusion® HD Cloning Plus. TaERF-BD protein expression in Escherichia coli BL21 cells was induced by 0.5 mM isopropyl- β-d-thiogalactoside (IPTG) at 28 °C for 6 h, and it was purified using glutathione–Sepharose 4B (52-2303-00, GE Healthcare). Unlabeled GCC-L14 and ACC-L14 Boxes, biotin-labelled GCC-L14 and ACC- L14, and GCC-H10 and ACC-H10 Boxes and their reverse comple- mentary sequences were synthesized and annealed as probes. EMSA was carried out using a LightShift® Chemiluminescent EMSA Kit (20148, Thermo Scientific) according to the manufacturer’s instructions.

Dual-luciferase assays of transformed tobacco leaves

The full-length cDNAs of TaERF3 and TaERF115 were ampli- fied using the primers TaERF3-1300-F/R and TaERF115-1300-F/R

(Supplementary Table S2) and separately cloned into the effector vector pCAMBIA1300 under the control of CaMV 35S. The TaSDIR1-4A promoter fragment was amplified from H10 and L14 using the primers Hap-LUC-F/R (Supplementary Table S2) and separately ligated into the reporter vector pGreen II 0800-LUC. The effector and reporter con- structs were transformed into GV3101 cells using pSoup, and transformed into leaves of 4-week-old Nicotiana tabacum plants by co-infiltration.The activities of firefly luciferase (LUC) and Renilla luciferase (REN) were measured using the Dual-Glo® Luciferase Assay System (E2920, Promega) with a multimode reader (TriStar2 S LB942) at 48 h after infiltration. The promoter activity was calculated as the ratio of LUC to REN, and the ratio in leaves transformed with the empty vector (pCAMBIA1300/ pGreen II 0800-LUC) was set to 1.

Results

TaSDIR1-4A is a functional E3 ligase

Two copies of TaSDIR1 were isolated from the chromosome group 4 of subgenomes A and D in wheat, and hence were named as TaSDIR1-4A and TaSDIR1-4D. TaSDIR1-4D is known to be involved in responses to abiotic stress (Wang et al., 2018). The cDNA sequence of TaSDIR1-4A obtained from cv. H10 consisted of eight exons and seven introns and encoded a protein containing 280 amino acids.TaSDIR1-4A, belonging to the RING finger domain family, contained two transmembrane domains in the N-terminal region, and a C3H2C3-type RING domain in the C-terminal region (Supplementary Fig. S1).

Previous studies have shown that the RING finger- containing protein AtSDIR1 functions as an E3 ligase (Zhang et al., 2007). An in vitro E3 ubiquitin ligase activity assay was therefore performed to test whether TaSDIR1-4A had E3 ligase activity, using AtSDIR1 as a positive control. Ubiquitination activity was observed in the presence of 6×His tag ubiquitin (Ub), E1 (from wheat), E2 (UBCh5b) and purified MBP- AtSDIR1 or MBP-TaSDIR1-4A proteins using an anti-Ub antibody (Fig. 1). An anti-MBP blot analysis also indicated that MBP-AtSDIR1 and MBP-TaSDIR1-4A were ubiquitinated. However, with MBP alone, or in the absence of E1, E2, E3, or Ub, no polyubiquitination was detected. The highly conserved RING finger domain of SDIR1, from Cys211 to Cys241 in Arabidopsis, has E3 ubiquitin ligase activity.The equivalent func- tional region in TaSDIR1-4A was from Cys221 to Cys251.The RING motif was essential for the E3 ligase activity of AtSDIR1. Our previous work had generated an allele with a single amino acid substitution by mutagenizing His-234 to Tyr (H234Y) in AtSDIR1, which abolished the E3 ligase activity (Zhang et al., 2007). Presumably the mutation disrupted the RING domain. A similar mutation in TaSDIR1-4A, His-244-Tyr, was produced, and again an in vitro ubiquitination assay indicated that the E3 ligase activity was completely disrupted (Fig. 1). These results indicated that TaSDIR1-4A has E3 ligase activity and that an intact RING domain is required for enzyme activity.

Sequence polymorphism assays, genetic mapping, and association analysis

Sequence polymorphism assays indicated that there was no nucleotide polymorphism in the TaSDIR1-4A coding re- gion, but two nucleotide variation sites were detected in the

Fig. 1. E3 ubiquitin ligase activity of wheat TaSDIR1-4A. The fusion proteins MBP-TaSDIR1-4A and its mutant form MBP-TaSDIR1-4AH244Y were assayed for E3 activity in the presence of E1 (from wheat), E2 (UBCh5b), and 6×His tag ubiquitin (Ub). The molecular masses of the marker proteins are shown (kDa). MBP was used as a negative control and MBP-AtSDIR1 was used as a positive control. Samples were resolved by 10% SDS-PAGE. An anti-Ub antibody was used to detect His tag ubiquitin (top image), and the anti-MBP antibody was used for detection of maltose fusion proteins (bottom image).

TaSDIR1-4A promoter region, a 2-bp InDel at –412 to –411 and a SNP at –395(Supplementary Fig. S2). Two haplotypes, Hap-4A-1 and Hap-4A-2, were identified (Fig. 2A, B) and a functional marker (FM) based on SNP-395 (G/A) was able to distinguish between them (Fig. 2C). The marker contained a mismatch in the downstream primer that produced a recogni- tion site for the restriction enzyme Age I.

Using diploid and tetraploid wheat relatives and a set of Chinese Spring nullisomic-tetrasomic lines, TaSDIR1-4A was found to be located on chromosome 4A (Supplementary Fig. S3A), and the functional marker was used to genotype the doubled-haploid (DH) lines in order to further map its position. Linkage analysis located a QTL for TGW between this functional marker and AX-111611657 (Supplementary Fig. S3B). Using the same DH lines, a QTL for TGW and the

Fig. 2. Nucleotide polymorphisms and development of a functional marker in wheat TaSDIR1-4A. (A) Schematic diagram of the TaSDIR1-4A structure. The ATG start codon was designated as position 1 bp. (B) Polymorphic sites were detected in the promoter regions of TaSDIR1-4A. (C) A dCAPS marker was developed based on the –395 bp single-nucleotide polymorphism (G/A). Digestion of the amplified 200-bp fragment with Age I produced fragments of 179 bp and 21 bp for accessions with the haplotype Hap-4A-1 (A), whereas this fragment was not digested in accessions with the haplotype

Hap-4A-2 (G). Marker, 100-bp DNA ladder.

Table 1. TaSDIR1-4A haplotypes associated with agronomic traits across 10 environments

TaSDIR1-4A is a negative regulator of grain size

To further examine the role of TaSDIR1-4A in wheat we si-

Year planted

Site Conditions Effect on 1000-grain

weight

P-value PVE (%)

lenced its transcription by BSMV-mediated VIGS in the cul-

tivar Chinese Spring. BSMV-associated chlorotic stripe mosaic symptoms appeared in newly emerged leaves 10 d post in-

2010 CP DS 3.56×10–6*** 10.78

CP WW 8.38×10–6*** 8.59

SY DS 4.09×10–5*** 8.20

SY WW 1.74×10–4*** 6.04

2011 SY DS 0.0264* 4.41

SY WW 0.0013*** 4.11

2012 CP DS 4.57×10–6*** 9.48

PVE, phenotypic variation explained. Sites: CP, Changping; SY, Shunyi. Conditions: DS, drought-stressed; WW, well-watered. Significant differences between the TaSDIR1-4A haplotypes were determined using Student’s t-test: *P<0.05, ***P<0.001.

TGW-related trait stem water-soluble carbohydrates had pre- viously been identified in an adjacent interval flanked by the markers Xgwm601 and WMC420 (Yang et al., 2007; Li et al., 2012a).The marker Xgwm601 was 3.56 cM from TaSDIR1-4A. Another QTL for the TGW-related trait chlorophyll content is also present in this region (Li et al., 2010a). This indicated that TaSDIR1-4A might have a connection with grain weight. Population 2 was used to examine the associations be- tween the TaSDIR1-4A haplotypes and agronomic traits. Results of sequence polymorphism analysis of 262 acces- sions (Supplementary Table S4) indicated that variation at the TaSDIR1-4A locus was significantly associated with TGW in all the 10 environments that were examined (Table 1). Hap-4A-1 and Hap-4A-2 accounted for 36.0% and 64.0% of entries in this population, respectively. Accessions with Hap-4A-2 had a higher mean TGW than accessions with Hap-4A-1 (Fig. 3A), and thus it represented a more favorable haplotype. These differences were also found in

Populations 1 and 4 (Fig. 3B, C).

oculation with the virus (Supplementary Fig. S4A). Flag leaves were harvested at anthesis and RT-PCR showed a 60% de- cline in transcripts in the BSMV:TaSDIR1-4A (silenced) lines compared with the BSMV:GFP lines (Fig. 4A). Quantitative real-time PCR was performed to confirm the specificity of knock-down of the expression of TaSDIR1-4A, no significant difference was observed in the expression of TaSDIR1-4D (the gene with the highest sequence similarity to TaSDIR1-4A) between the control and BSMV:TaSDIR1-4A lines (Supplementary Fig. S4B). Measurements after harvest showed that the silenced lines had larger grain size and higher TGW than the BSMV:GFP lines (Fig. 4B–D) whilst the BSMV-free lines had larger grain size and higher TGW than the BSMV- infected lines. Assuming that damage from BSMV infection would lead to smaller grain size and lower TGW, these results indicated that TaSDIR1-4A had a negative role in the deter- mination of TGW.

To further confirm its effect on grain size and TGW,

TaSDIR1-4A was overexpressed in the wild-type and the sdir1-1 mutant of Arabidopsis and relative expression was de- termined by semi-quantitative PCR (Supplementary Fig. S5). The sdir1-1 line had larger seeds whilst the TaSDIR1-4A- overexpression lines in both the wild-type and mutant back- ground had smaller seeds than the non-transgenic controls (Fig. 4E).

TaSDIR1-4A expression patterns

Because sequence polymorphism sites were located in the TaSDIR1-4A promoter region, we detected effects of vari- ation on expression. We first performed real-time PCR on a range of tissues to determine expression patterns in wheat at the flowering stage, and found that TaSDIR1-4A was

Fig. 3. Comparisons of 1000-grain weight (TGW) in the two wheat TaSDIR1-4A haplotypes Hap-4A-1 and Hap-4A-2. (A) Comparisons in Population 2 (262 accessions) grown in 10 different environments: either at Changping (CP) or Shunyi (SY), and under drought-stressed (DS) or well-watered (WW) conditions, planted in 2010, 2011, or 2012. (B) Comparisons in Population 1 (doubled-haploid) grown in 10 different environments: either at CP or SY, and under DS, WW, or heat-stressed (HS) conditions, planted in 2015 or 2016. (C) Comparisons in Population 4 (348 modern cultivars) grown in three different environments: SY planted in 2010 or Luoyang (LY) planted in 2002 or 2005. Data are means (±SE), n=3. Significant differences between the haplotypes were determined using Student’s t-test: *P<0.05, **P<0.01, and ***P<0.001.

Fig. 4. Phenotypic analysis of TaSDIR1-4A-silenced wheat and transgenic Arabidopsis. Virus-induced gene silencing (VIGS) of wheat was undertaken using a series of recombinant barley stripe mosaic virus (BSMV) vectors. (A) Relative expression of TaSDIR1-4A showing that transcription was lower in BSMV:TaSDIR1-4A plants than BSMV-free and BSMV:GFP plants. (B) Thousand-grain weight (TGW), (C) grain size, and (D) representative images of the grains. (E) Seed size in Arabidopsis plants in the Col-0 wild-type and T-DNA insertion mutant sdir1-1 backgrounds overexpressing (OE) TaSDIR1-4A. Vector, empty vector control. Significant differences between means were determined using one-way ANOVA followed by Tukey’s test (P<0.05).

constitutively expressed in all tissues, with highest expression levels in the flag leaves (Fig. 5A). We then randomly selected 12 accessions of each of the two haplotypes from Populations 1 and 2 and planted them in the field (Supplementary Table S3). Flag leaves were collected at anthesis and the expression levels of TaSDIR1-4A were determined. The mean relative expression level of the Hap-4A-1 accessions was 2.21-fold

greater than that of the Hap-4A-2 accessions in Population 1 (Fig. 5B), and 2.77-fold greater in Population 2 (Fig. 5C). In agreement with the results of the association analysis, accessions with Hap-4A-1 had higher gene expression and lower TGW relative to those with Hap-4A-2, indicating that lower expres- sion of TaSDIR1-4A Hap-4A-2 resulted in higher TGW, and that TaSDIR1-4A plays a negative role in regulating TGW.

Fig. 5. Expression patterns of the wheat TaSDIR1-4A haplotypes Hap-4A-1 and Hap-4A-2. (A) Expression patterns of TaSDIR1-4A in different tissues of cv. H10 plants as measured by real-time PCR. S, spikes at flowering; FL, flag leaves; St, stems; N, nodes; RB, root bases; R30, roots from 0–30 cm depth; R50, roots from 30–50 cm; R70, roots from 50– cm; R90, roots from 70–90 cm; R100, roots from 90–100 cm. The values are relative to R90, which was set as 1. (B, C) Expression of Hap-4A-1 and Hap-4A-2 in (B) Population 1 and (C) Population 2. For each haplotype 12 accessions were

randomly selected for measurement. (D) Expression patterns of TaSDIR1-4A following ABA treatment. Seedlings of cv. H10 at 2 weeks old were sprayed with 50 μM ABA solution and were sampled at 0 h (control, CK) and at 0.5–72 h. Expression is relative to the value in the control, which was set as 1. (E) Expression patterns of TaSDIR1-4A following NaCl treatment. Seedlings of cv. H10 at 2 weeks old were treated with 250 mM NaCl and were sampled at 0 h (CK) and at 0.5–72 h. Expression is relative to the value in the control, which was set as 1. In all cases GAPDH was used as the internal control. All data are means (±SE) of three biological replicates.