Mapping of a major QTL associated with protein content on chromosome 2B in hard red winter wheat (Triticum aestivum L.).

A quantitative trait locus (QTL) controlling wheat grain protein content (GPC) and flour protein content (FPC) was identified using doubled haploid (DH) lines developed from a cross between the hard red winter wheat variety 'Yumechikara' with a high protein content used for bread making, and the soft red winter wheat 'Kitahonami' with a low protein content used for Japanese white salted noodles. A single major QTL, QGpc.2B-yume, was identified on the short arm of wheat chromosome 2B for both the GPC and FPC over 3 years of testing. QGpc.2B-yume was mapped on the flanking region of microsatellite marker Xgpw4382. The DH lines grouped by the haplotype of the closest flanking microsatellite marker Xgpw4382 showed differences of 1.0% and 1.1% in mean GPC and FPC, respectively. Yield-component-related traits were not affected by the haplotype of QGpc.2B-yume, and major North American hard red winter wheat varieties showed the high-protein haplotype. Unlike Gpc-B1 derived from tetraploid wheat, QGpc.2B-yume has no negative effects on yield-component-related traits and should be useful for wheat breeding to increase GPC and FPC.

Intoroduction

Protein content is a major factor determining end-use quality in wheat, and genetic improvement to achieve higher protein content is a key issue worldwide in wheat breeding. In Japan, wheat varieties with a low to moderate protein content have been preferred for white salted noodles (Udon); however, increasing demand for domestic bread and yellow alkaline Chinese noodles (Ramen) require wheat varieties with a high protein content.

Screening for high protein content in wheat is subject to interference from environmental effects, and a negative correlation with yield has been reported (Bogard ). Furthermore, it has been suggested that improved nitrogen remobilization efficiency or nitrogen uptake after flowering would increase GPC without reducing the yield, and that there are genes with independent effects on the yield and GPC (Bogard ).

Marker-assisted selection (MAS) is a powerful tool used in breeding programs for less laborious and more efficient screening under varied environmental conditions. In hexaploid wheat, four stable quantitative trait loci (QTLs) were identified for GPC on chromosomes 2A, 3A, 4D, and 7D (Groos ), and recent studies also found that protein content is determined mostly by multiple quantitative genes (Simons , Tsilo ).

Successful MAS for high protein content in wheat was first achieved with the introgression of Gpc-B1, a major QTL for GPC mapped on chromosome 6BS derived from the wild tetraploid wheat T. turgidum var. dicoccoides (Joppa , Olmos ). Brevis and Dubcovsky (2010) compared near-isogenic lines (NILs) with contrasting Gpc-B1 alleles in tetraploid and hexaploid wheat, and they showed that functional Gpc-B1 is associated with an increase in GPC. Kumar and Tabbita showed that GPC was increased in Indian and Argentine hexaploid wheat carrying Gpc-B1. Besides protein content, Gpc-B1 is also associated with smaller grain size (Tabbita , Uauy ) and reduced grain yield (Cater ), which reduce wheat production.

Yumechikara is a hard red winter wheat with wheat yellow mosaic disease resistance and a high protein content, and of the preferred quality for bread making and yellow alkaline Chinese noodle production (Tabiki ). Recently a major QTL responsible for wheat yellow mosaic disease resistance was mapped on chromosome 2D in Yumechikara (Kojima ), however, the genes responsible for the high GPC and FPC of Yumechikara have not been identified. We accordingly sought to identify a novel gene that increases protein content without pleiotropically reducing wheat production. To identify the QTL and screen for markers associated with high GPC and FPC, we performed mapping and QTL analysis using a doubled haploid (DH) population derived from Yumechikara.

Materials and Methods

Plant materials

Ninety-four DH lines derived from a cross between two Japanese elite winter wheat varieties, Kitami 81 (the former name of Kitahonami, a low-protein-content parent) and Kachikei 63 (the former name of Yumechikara, a high-protein-content parent) were used (Ito , Kojima ). The pedigree of Kitahonami is Kitami 72 (Kitamoe)/Kitakei 1660 (Yanagisawa ) and that of Yumechikara is Satsukei 159/KS 831957 (F1) and Tsukikei 9509 (Kitanokaori) (Tabiki ). The DH lines were produced by the wheat × maize method (Inagaki and Tahir 1990). To identify the genotypes of the major winter wheat varieties in Hokkaido and the pedigrees of the parental lines, six wheat lines were used: Kitamoe, Hokushin, Kitanokaori, Touhoku 140, Horoshirikomugi, and KS 831957. The Kitahonami pedigree includes Kitamoe and Hokushin (Yanagisawa ). The Yumechikara pedigree includes Kitanokaori, Touhoku 140, Horoshirikomugi, and KS 831957 (Tabiki ). To identify the genotypes of North American winter wheat varieties, we used nine representative varieties, Betty (PI-612578), CDC Kestrel (Fowler 1997a), CDC Osprey (Fowler 1997b), CDC Claire (Fowler 1997c), Heyne (PI-612577), Jagger (PI-593688), Karl (PI-527480), Karl 92 (PI-564245), and Palo Duro (Wilson and Salm1970). These varieties were obtained from USDA and University of Saskatchewan.

Field experiments

The DH and parental lines were cultivated in a research field at NARO/HARC, Memuro, Hokkaido (42.5°N, 140.9°E) for a 3-year period between 2006 and 2009 (2006–2007; experiment 2007, 2007–2008; experiment 2008, 2008–2009; experiment 2009). The experiment was divided into different field rotations and different field conditions were used in each year. The 2007, 2008, and 2009 field rotation areas were 126a, 109a, and 205a, respectively. The preceding crop was soybean. Planting dates for the experiment were September 19, 2006 (experiment 2007), September 18, 2007 (experiment 2008), and September 15, 2008 (experiment 2009). Cultivation methods consisted of the use of a plot with 2-m long rows with 72-cm widths, basal nitrogen (N) fertilization treatment (4 kg/10 a) and additional N fertilization treatment (6 kg/10 a) in April (Ito ). The total amounts of fertilization treatment were 10 kg/10 a (N), 8 kg/10 a (P2O5), and 4.8 kg/10 a (K2O) for every year.

The daily mean air temperatures during the grain-filling period in 2007, 2008, and 2009 were 16.9°C, 17.0°C and 16.0°C, respectively. Total precipitation during the grain-filling period was 104 mm (2007), 183 mm (2008), and 348 mm (2009). Weather data were obtained from the Japan Meteorological Agency (http://www.data.jma.go.jp/obd/stats/etrn/select/prefecture00.php).

Yield components

Four yield-component-related traits: culm length (CL), spike length (SL), thousand kernel weight (TKW), test weight (TW), and grain yield per unit area (GY) were measured in two years (2008 and 2009). TW was determined using the test weight module of the Infratec 1241 grain analyzer (Foss, Denmark). Single-kernel hardness (SKH), single-kernel weight (SKW), and single-kernel diameter (SKD) were determined with a Single-Kernel Characterization System (SKCS) (SKCS 4100, Perten Instruments, USA) as the mean measurement of 50 kernels. All yield component data are shown in Supplemental Table 1. Pairwise Pearson’s correlation coefficients were calculated for yield-component-related traits (SKH, SKW, SKD, TW, TKW, CL, SL and GY) and protein content (GPC and FPC) (Table 2). Continuous variables (GPCs and FPCs in three years) were compared by one-way analysis of variance (ANOVA) (Table 1). Other statistical comparisons of traits were performed by t-test.

Table 2

Correlation coefficients among yield-component-related traits and protein contents

	SKH	SKW	SKD	TW	TKW	CL	SL	GY	GPC	FPC
SKH	–
SKW	−0.01	–
SKD	0.08	0.92**	–
TW	−0.08	0.43**	0.39**	–
TKW	0.06	0.88**	0.83**	0.47**	–
CL	−0.06	0.27**	0.27**	0.26*	0.31**	–
SL	0.05	0.35**	0.32**	0.10	0.30**	0.35**	–
GY	−0.11	0.44**	0.43**	0.44**	0.46**	0.63**	0.27**	–
GPC	0.10	−0.02	0.10	−0.22*	0.02	−0.13	0.21*	−0.35**	–
FPC	0.54**	0.03	0.16	−0.18	0.09	−0.17	0.21*	−0.34**	0.83**	–

Significant at P < 0.05 and P < 0.01.

Table 1

Mean values and standard deviations of yield-component-related traits and protein content of the doubled haploid population and parents evaluated in 2007, 2008, and 2009

Trait	Year	Parental lines			DH population (n = 94)
		Yumechikara	Kitahonami	P-value1	Mean	SD	P-value2
SKH	2007	90.7	35.9	***	73.2	18.0
	2008	85.8	38.6	***	65.4	19.9
	Two-year mean	88.3	37.2	***	69.3	18.7	*
SKW	2007	42.5	41.2	*	41.8	4.7
	2008	38.9	38.1	*	39.7	4.1
	Two-year mean	40.7	39.6	*	40.8	4.1	*
SKD	2007	2.7	2.4	*	2.7	0.2
	2008	2.6	2.4	*	2.6	0.2
	Two-year mean	2.6	2.4	*	2.6	0.2	NS
TW	2008	793.0	791.0	*	786.9	28.7
	2009	797.0	788.0	*	750.6	31.9
	Two-year mean	795.0	789.5	*	768.7	24.6	***
TKW	2008	40.9	38.6	*	40.6	4.2
	2009	39.6	37.9	*	39.7	4.1
	Two-year mean	39.9	38.8	*	40.1	3.8	NS
CL	2008	72.8	79.9	*	79.2	6.9
	2009	78.1	86.0	**	84.4	6.0
	Two-year mean	75.5	82.9	*	81.8	6.0	***
SL	2008	11.3	9.9	NS	11.0	0.8
	2009	10.4	9.1	NS	9.9	0.7
	Two-year mean	10.6	9.7	NS	10.4	0.7	***
GY	2008	463.4	465.7	NS	450.0	11.0
	2009	418.4	437.3	NS	371.7	7.3
	Two-year mean	440.9	451.5	NS	410.8	8.6	***
GPC	2007	12.5	9.4	***	11.5	1.0
	2008	15.2	11.9	***	14.0	0.9
	2009	16.1	12.4	***	14.4	0.9
	Three-year mean	14.6	11.1	***	13.2	0.8	***
FPC	2007	12.8	8.3	***	10.7	1.2
	2008	13.5	9.9	***	12.1	1.2
	2009	13.6	10.1	***	12.7	1.3
	Three-year mean	13.3	9.4	***	11.8	1.1	***

SKH single-kernel hardness determined by single kernel characterization system (SKCS); SKW single-kernel weight determined from SKCS; SKD single-kernel diameter determined from SKCS; TW test weight (g/L); TKW thousand-kernel weight (g); CL culm length (cm); SL spike length (cm); GY grain yield per unit area (kg/10 a); GPC grain protein content (%); FPC flour protein content (%).

*,**,*** Parental means were significantly different at the level of P < 0.05, P < 0.01 and P < 0.001, respectively; NS not significant at P > 0.05.

*,*** Year means were significantly different at the level of P < 0.05 and P < 0.001, respectively; NS not significant at P > 0.05.

Protein content analysis

Wheat grain samples were milled with a test mill MLU-202 (Brabender Inc., Germany). GPC for the 3-year period (2007–2009) was determined using a near-infrared reflectance instrument Infratec 1241 grain analyzer. FPC for the 3-year period was also determined with a near-infrared reflectance instrument Inframatic 8120 (Percon Corporation, Germany). All GPC and FPC data are shown in Supplemental Table 1.

Microsatellite markers

A total of 1261 SSR markers (Somers , Song , Sourdille ) were tested for polymorphism in order to construct a genetic linkage map. Polymerase chain reaction (PCR) was performed in a total volume of 20 μl, which contained 0.4 μM of each primer, GoTaq Master Mix (Promega), and 100 ng of total DNA from leaves and seeds of the DH population. A GeneAmp 9700 PCR System (Applied Biosystems) and a T100 Thermal Cycler (Bio-Rad) were used for the reactions according to the following protocol: denaturation at 94°C for 5 min; 35 or 42 cycles of 94°C for 30 s or 1 min, 51°C, 55°C, or 60°C for 60 s, and 72°C for 60 s; and a final extension step at 72°C for 5 min. The PCR products were electrophoresed in a 3% Agarose 21 (Nippon Gene) gel in TAE buffer or a 10% polyacrylamide gel in TBE buffer.

Mapping and QTL analysis

A total of 275 polymorphic microsatellite markers assigned to the 21 wheat chromosomes were used for this analysis. A genetic map was constructed with MAPMAKER/Exp version 3.0b (Lander ), antmap Ver. 1.2. (Iwata and Ninomiya 2006) and additional microsatellite markers potentially targeting the QTL interval were obtained from the wheat microsatellite consensus map (Somers , Song , Sourdille ). The Kosambi mapping function was used to convert recombination fractions into map distances (Kosambi 1943). QTLs were detected by composite interval mapping using the software Windows QTL Cartographer Version 2.5 (Wang ). A log-likelihood (LOD) score threshold for detection of QTLs were determined by computing 1,000 permutations. The total phenotypic variance of multiple QTLs was estimated by the multiple-interval mapping model of the software.

Results

Mean values of the yield component-related traits are shown in Table 1. Significant differences between Kitahonami and Yumechikara were found for 8 traits (SKH, SKW, SKD, TW, TKW, CL, GPC, and FPC). For all traits analyzed, significant variations were due to the effect of each year in p-value by the t-test. There were significant differences for six traits (SKH, SKW, TW, CL, SL, and GY) between 2007 and 2008 or 2008 and 2009. The p-values of TW, CL, SL, and GY were less than 0.001. The results of the one-way ANOVA indicated p-values of GPC and FPC were also less than 0.001 for the 3-year period. Pairwise correlations were estimated among all traits measured in Table 2. Strong positive correlations were observed among TKW, SKW, and SKD, and the correlation between GPC and FPC was also high. Significant negative correlation was also observed between GY and protein content (GPC and FPC).

GPC and FPC variation

The distributions of GPC and FPC of the DH lines are shown in Fig. 1 and Table 1, respectively. In 2007, the GPC of the DH population ranged from 9.1% to 14.4% (mean, 11.5%) and the FPC ranged from 8.2% to 13.4% (mean, 10.7%). These protein contents were the lowest in the three-year period. In 2009, the GPC ranged from 12.4% to 17.2% (mean, 14.4%) and FPC ranged from 10.0% to 16.4% (mean, 12.7%), values that were the highest in the 3-year period. The GPCs of the parents Kitahonami and Yumechikara were significantly different. The 3-year means of GPC were 11.1% for Kitahonami and 14.6% for Yumechikara. The 3-year mean of the GPCs in the DH population ranged from 11.6% to 15.6% (mean, 13.2%); this range was higher than the value for Kitahonami and extended beyond that for Yumechikara. The results for the GPC in each parent showed the same tendency as that for the mean GPC of the DH population among the 3 yearly tests. These results suggested that the different protein contents in the 3-year period were due to differences in the environment among the growing seasons. In particular, total precipitation during grain-filling period in 2007 (104 mm) was lower than that of the other two years (2008: 183 mm and 2009: 348 mm).

Fig. 1

(A) Frequency distribution of the grain protein content (GPC) in 94 DH-lines and their parental lines in a 3-year period. Kitahonami (K), Yumechikara (Y) and the mean GPC of the DH population (M) are indicated by arrows. (B) Frequency distribution of the flour protein content (FPC) in 94 DH-lines and their parental lines in a 3-year period. The arrow description and labels are the same as in (A).

The 3-year means of the FPC were 9.4% for Kitahonami and 13.3% for Yumechikara. The 3-year means of the FPC in the DH population ranged from 9.6% to 14.8% (mean, 11.8%); this range was higher than the mean value for Kitahonami and extended beyond the mean for Yumechikara as well as GPC. The results for FPC in each of the three years showed the same tendency as that for the 3-year mean FPC. Both the GPC and FPC of individual DH lines showed a good fit to the normal distribution, and the values of a few DH lines were higher than those of Yumechikara in each year (Fig. 1).

QTL analysis

A total of 1261 SSR markers were used for detection of polymorphism between the parental genotypes (Kitahonami and Yumechikara) of the DH mapping population. Of the markers tested, 275 polymorphic PCR products were produced and subsequently used for genotyping of all DH lines. The resulting linkage map contained 36 linkage groups covering a total genetic distance of 2257.1 cM in the DH population. QTL analysis revealed a GPC and FPC QTL on the short arm of chromosome 2B (temporarily named QGpc.2B-yume) was significantly associated with GPC and FPC (Fig. 2A). Genotyping data of chromosome 2B are shown in Supplemental Table 2. The highest LOD scores were located just at the position of Xgpw4382 for both GPC and FPC in all years (2007, 2008 and 2009) and for the 3-year mean (Fig. 2A). These LOD scores were ranged from 2.60 in FPC-2009 to 9.58 in GPC-2009 and all LOD peak points were well above the significant LOD score threshold (Table 3). All LOD scores of Xgpw4382 were higher than those of both Xgpw3215 and Xwmc245. The Xgpw4382 region explained 32.1% and 16.5% of the phenotypic variance explained (PVE) in the 3-year means of the GPC and FPC, respectively.

Fig. 2

(A) QTL analysis of GPC and FPC on wheat chromosome 2B. Line 1: GPC-2007; line 2: GPC-2008; line 3: GPC-2009; line 4: GPC-3-year mean; line 5: FPC-2007; line 6: FPC-2008; line 7: FPC-2009; line 8: FPC 3-year mean. Genetic map of the arm of chromosome 2B with candidate markers. The highest and lowest LOD score threshold is indicated by dashed line. Arrows indicates the approximately position of Xgwm271 and Xgpw1249 SSR marker. (B) Genotypes of a subset of eight DH-lines. Five additional markers were mapped between the flanking markers Xhbg246 and Xwmc245 (‘K’ white cells: Kitahonami alleles, ‘Y’ gray cells: Yumechikara alleles).

Table 3

QTLs identified for Gpc and Fpc using 94 DH lines in 2007, 2008, 2009 and Three-year mean

Year	Vicinity Markler	LOD score	LOD score threshold	PVE (%)	Additive effect
GPC-2007	Xgpw4382	4.31		18.9	0.44
	Xgpw3215	3.76	2.4	16.6	0.41
	Xwmc245	3.72		17.8	0.42
GPC-2008	Xgpw4382	4.16		18.3	0.37
	Xgpw3215	3.89	2.2	16.1	0.35
	Xwmc245	3.36		16.2	0.35
GPC-2009	Xgpw4382	9.58		37.9	0.57
	Xgpw3215	8.72	2.2	34.4	0.54
	Xwmc245	7.91		34.4	0.55
GPC three-year mean	Xgpw4382	7.73		32.1	0.44
	Xgpw3215	7.00	2.4	28.5	0.44
	Xwmc245	6.62		29.0	0.44
FPC-2007	Xgpw4382	3.74		16.7	0.47
	Xgpw3215	3.53	2.4	15.6	0.45
	Xwmc245	3.38		15.4	0.46
FPC-2008	Xgpw4382	2.90		13.2	0.39
	Xgpw3215	2.42	2.5	10.9	0.37
	Xwmc245	2.11		10.1	0.38
FPC-2009	Xgpw4382	2.60		12.3	0.47
	Xgpw3215	2.14	2.4	9.6	0.41
	Xwmc245	1.71		8.1	0.39
FPC three-year mean	Xgpw4382	3.62		16.5	0.42
	Xgpw3215	3.13	2.6	13.0	0.41
	Xwmc245	2.77		12.9	0.40

Fine mapping and efficiency of DNA markers

To perform fine mapping of the region of the Xgpw4382, we used 24 microsatellite-markers on the short arm of chromosome 2B associated with this region by reference to the ITMI map (Somers ), the Ta-Synthetic/Opata-GPW map (Sourdille ), genetic and physical maps of wheat Xbarc SSR loci (Song ) and an SSR-based linkage map (Torada ). As a result, polymorphic PCR-products between the parental lines were confirmed for two makers (Xgpw3215 and Xgpw2225) and were subsequently used for genotyping all DH lines. Graphical genotypes of eight DH lines are shown in Fig. 2B. White cells labeled with ‘K’ indicate Kitahonami alleles, whereas gray cells labeled with ‘Y’ indicate Yumechikara alleles. DH line 71 was of particular interest, given that the closest flanking crossovers in this line delimited a region between Xgpw4382 and Xgpw2225-Xwmc3215. This QTL region between Xhbg246 and Xwmc245 was delimited by seven markers based on seven lines (DH 6, 10, 36, 40, 43, 71 and 90). Fig. 3A shows the PCR amplification products of the eight DH and parental lines in Fig. 2B obtained using the Xgpw4382-2B primers (Xgpw4382 F: 5′-TGTTAGCAGAATAAGCTGGGTG-3′ and Xgpw4382 R: 5′-GTTGTTCAATGTTGTAGGTGCC-3′; amplification conditions: Tm, 60°C), as reported by Sourdille . Polymorphic bands (indicated by arrows in Fig. 3A) at 130 bp were associated with the genotypes of Kitahonami and Yumechikara.

Fig. 3

(A) PCR amplification products using Xgpw4382 primers. Lane 1: 100-bp DNA marker (WAKO). Kitahonami–Yumechikara DH lines (lanes 2–9) are shown in Fig. 2B. (B) PCR amplification products of major Hokkaido winter wheat varieties using the Xgpw4382 marker. (C) PCR amplification products of representative North American winter wheat varieties using the Xgpw4382 marker. The arrow indicates PCR amplification products for Xgpw4382 (approximately 130 bp). Kitahonami- and Yumechikara-type amplification products are labeled as ‘K’ and ‘Y,’ respectively.

Efficiency of DNA markers

The DH lines used for genotyping with the Xgpw4382 marker were classified into two types: Kitahonami (K) and Yumechikara (Y). As a result, all of the DH lines were divided into 41 K-type and 53 Y-type lines. The mean values of the GPC and FPC for the three years and the results of a significance test between the K-type and Y-type groups are shown in Table 4. The 3-year mean and standard error of the mean of the GPC of the K type were 12.7 ± 0.11%, whereas that of the Y type was 13.7 ± 0.09%. The 3-year mean FPCs of the K and Y types were 11.1 ± 0.15% and 12.1 ± 0.15%, respectively. In all three years, significant differences were found between the Kitahonami and Yumechikara types. These results suggested that this marker will be useful for selecting lines with high or low level protein content.

Table 4

Mean values and standard error of the mean of the GPC and FPC in the DH population grouped by Yumechikara type and Kitahonami type for Xgpw4382 on chromosome 2B

	Yumechikara type (n = 53)			Kitahonami type (n = 41)			p-value
	Mean ± SEM	Max.	Min.	Mean ± SEM	Max.	Min.
GPC-2007	11.7 ± 0.13	14.4	9.8	10.8 ± 0.13	12.4	9.1	***
GPC-2008	14.3 ± 0.13	16.5	12.9	13.5 ± 0.13	15.9	11.8	***
GPC-2009	15.1 ± 0.09	17.2	13.6	13.9 ± 0.13	16.3	12.4	***
GPC three-year mean	13.7 ± 0.09	15.6	12.4	12.7 ± 0.11	14.7	11.6	***
FPC-2007	10.9 ± 0.15	13.4	8.6	9.8 ± 0.15	12.2	8.2	***
FPC-2008	12.5 ± 0.16	15.3	10.3	11.6 ± 0.17	14.5	9.5	***
FPC-2009	12.9 ± 0.18	16.4	10.8	11.9 ± 0.19	15.0	10.0	***
FPC three-year mean	12.1 ± 0.15	14.8	9.9	11.1 ± 0.15	13.7	9.6	***

Significantly different at P < 0.001.

The PCR products of the major winter wheat varieties in Hokkaido using Xgpw4382 are shown in Fig. 3B. The varieties Hokushin and Kitamoe are soft red winter wheat varieties in Hokkaido and are suitable for Japanese Udon noodles (Yanagisawa , 2002), and both are pedigrees of Kitahonami (Yanagisawa ). Kitanokaori is a hard red winter wheat in Hokkaido that is suitable for bread (Tabiki ). Kitanokaori, Touhoku 140, Horoshiri komugi, and KS 831957 are pedigrees of Yumechikara (Tabiki ). The DNA fragments of the former major Japanese noodle-type soft red wheat Kitamoe and Hokushin were assigned to the K (Kitahonami) type. In contrast, the DNA fragments of the former bread-type hard red wheat Kitanokaori and all pedigrees of Yumechikara were all assigned to the Y (Yumechikara) type. Thus, the discrimination between Japanese noodle-type wheat with lower protein content and bread-type wheat with higher protein content would be possible using this marker among winter wheat varieties in Hokkaido.

The PCR products of representative North American hard red winter wheat varieties amplified using the Xgpw4382 marker are shown in Fig. 3C. The DNA fragments of all the North American lines were assigned to the Y type.

As with the GPC and FPC (Fig. 3A), the DH lines used for genotyping with Xgpw4382 were classified into two types, Kitahonami and Yumechikara, for yield-component-related traits. The mean values of the yield-component-related traits in the 2-year period (2008 and 2009) and the results of a statistical test between the DH lines of Kitahonami-type and Yumechikara-type groups are shown in Table 5. The mean values of the traits of each parental line are shown in Table 1.

Table 5

The mean values and standard error of mean of yield component related traits of the DH population grouped by ‘Yumechikara’ type and the ‘Kitahonami’ type for Xgpw4382 on chromosome 2B

Trait	Year	DH lines		p-value
		Yumechikara type (n=53)	Kitahonami type (n=41)
		Mean ± SEM
SKH	2007	72.0 ± 2.5	74.8 ± 2.7	NS
	2008	63.9 ± 2.7	67.3 ± 3.0	NS
	Two-year mean	67.9 ± 2.5	71.1 ± 2.8	NS
SKW	2007	42.7 ± 0.7	40.6 ± 0.5	NS
	2008	39.8 ± 0.6	39.6 ± 0.6	NS
	Two-year mean	41.2 ± 0.6	40.1 ± 0.5	NS
SKD	2007	2.7 ± 0.03	2.6 ± 0.03	NS
	2008	2.6 ± 0.03	2.6 ± 0.02	NS
	Two-year mean	2.7 ± 0.03	2.6 ± 0.02	NS
BD	2008	786.5 ± 4.2	787.5 ± 3.9	NS
	2009	749.6 ± 5.1	751.9 ± 3.4	NS
	Two-year mean	768.0 ± 3.7	769.7 ± 3.2	NS
TKW	2008	40.7 ± 0.7	40.3 ± 0.4	NS
	2009	40.4 ± 0.6	39.1 ± 0.5	NS
	Two-year mean	40.4 ± 0.6	39.7 ± 0.4	NS
CL	2008	79.6 ± 1.0	78.6 ± 1.0	NS
	2009	84.4 ± 0.9	84.3 ± 0.8	NS
	Two-year mean	82.0 ± 0.9	81.3 ± 0.9	NS
SL	2008	11.0 ± 0.1	11.0 ± 0.1	NS
	2009	9.9 ± 0.1	9.6 ± 0.1	NS
	Two-year mean	10.5 ± 0.1	10.3 ± 0.1	NS
GY	2008	445.1 ± 16.6	456.4 ± 12.9	NS
	2009	366.3 ± 11.1	378.7 ± 8.6	NS
	Two-year mean	405.7 ± 13.1	417.6 ± 9.7	NS

NS not significant at P > 0.05.

There were significant differences between the parental lines in the 2-year period. The mean values of SKH, SKW, SKD, TW, and TKW of Yumechikara were higher than those of Kitahonami in the 2-year period, however, SL and GY were not significantly different between the parents (Table 1). In contrast, the means of all yield-component-related traits of the DH lines grouped by Kitahonami or Yumechikara haplotypes in Xgpw4382 showed similar values, and there was no significant difference for any trait between the two groups in the 2-year period including GY (Table 5).

Discussion

Wheat varieties with high protein content are valuable for preferred market class and end-use quality, and high protein content is one of the most important targets of wheat breeding. However, it has been difficult to screen for protein content because the trait is determined by both genetic and environmental (G × E) factors (Chope ). In this study, a single major QTL (QGpc.2B-yume) for high protein content was mapped between Xgwm37 and Xwmc179 on chromosome 2B for all field trials, despite large differences in the distribution of protein content in the DH population during the tested years (Fig. 1). The SSR marker Xgpw4382 was selected by fine mapping and permitted discrimination between low and high-level protein content groups in the DH population (Table 4). We infer that this marker is closely associated with protein-content genes on chromosome 2B.

Previous studies found protein-content QTLs on chromosome 2B (Prasad , Tsilo ). However, QTLs in these previous studies and the novel QTL (QGpc.2B-yume) found in the present study are mapped to different positions on chromosome 2B by the wheat consensus map (Peleg , Sommers ). Prasad reported that the QTL (Gpc.ccsu-2B) was located near the marker Xgwm1249 on the long arm of chromosome 2B and Tsilo reported a QTL (QGpc.mna-2B) for GPC between the microsatellite markers Xwmc245-Xgwm271 on chromosome 2B (Fig. 2A). In GPC-2008, GPC-2007, FPC-2007 and FPC-2009, a minor QTL-peak was detected at the position of Xwmc245 (Fig. 2A, Table 3). This QTL might be related to the Gpc.ccsu-2B and QGpc.mna-2B.

Olmos and Uauy found Gpc-B1 in T. turgidum, and this gene has been reported to increase the GPC in hexaploid wheat (Brevis and Dubcovsky 2010). Uauy reported high GPC from Gpc-B1 in T. turgidum and showed that the gene is an NAC transcriptional factor, NAM1. Carter and Tabbita reported that the TKW was decreased in Gpc-B1 introgression lines with high GPCs. Gpc-B1 also accelerated senescence during the grain-filling stage (Uauy ).

We confirmed that Yumechikara does not have Gpc-B1 by PCR-based experiments reported by Distelfeld (data not shown). The TKW of the high-protein-content parent Yumechikara was higher than that of the low-protein-content variety, Kitahonami (Tabiki , Yanagisawa ) (Table 1). Although significant correlations were observed between GY and protein contents (GPC and FPC) in the DH lines (Table 2), GY and the yield-component-associated traits were not significantly different between the DH lines grouped by the haplotype of Xgpw4382 at the QTL (QGpc.2B-yume) on chromosome 2B (Table 5). It is also suggested that Yumechikara might have higher nitrogen uptake capacity compare to previous cultivars (Tabiki ). These results suggest that unlike Gpc-B1, the protein-content QTL on chromosome 2B exerts no significant negative effects on GY or other yield-component traits. The Xgpw4382 maker should be useful for practical wheat breeding for the purpose of increasing the GPC and FPC.

The haplotype of Xgpw4382 on chromosome 2B was shown to be useful for discrimination of Japanese Udon noodle-type wheats and bread-type wheats among Hokkaido varieties (Fig. 3B). The mean GPCs of the Japanese noodle-type varieties Kitamoe and Hokushin categorized low protein varieties as Kitahonami, were 9.5% and 9.2%, respectively, during a 5-year period (Yanagisawa ). In contrast, the mean GPC of bread-type Kitanokaori categorized high protein variety as Yumechikara, was 12.8% during a 4-year period (Tabiki ). Major soft red winter varieties for Japanese Udon noodles and hard red winter varieties for bread making in Hokkaido were clearly discriminated by the marker.

Yumechikara and four Yumechikara pedigrees showed the same haplotype for Xgpw4382 (Fig. 3B), and the origin of the QTL in Yumechikara needs further investigation. Haplotype analyses of North American hard red winter wheat varieties indicated that all varieties were assigned to the Yumechikara type. These varieties belong to the hard red winter market class (HRW) and are required to have a defined minimum protein content. This result suggests that the North American HRW varieties analyzed in this study carry the the Yumechikara’s allele at the QGpc.2B-yume on chromosome 2B.

Uauy showed that a gene paralogous to Gpc-B1 is present on chromosome 2B (Gpc-B2) of hexaploid wheat. Cantu reported high similarity to NAC2, a rice gene related to the GPC gene. Distelfeld indicated that the wheat gene Gpc-B2 and its rice ortholog have divergent functions. However, these studies did not identify genes on chromosome 2B directly associated with protein of hexaploid wheat. Further study may reveal whether the QTL is related to Gpc-B2 and affects GPC and FPC in diverse genetic backgrounds.