Genetic Analysis of Sugar Composition and Its Relationship with Protein, Oil, and Fiber in Soybean

Published August 30, 2018

RESEARCH Genetic Analysis of Sugar Composition and Its Relationship with Protein, Oil, and Fiber in Soybean

Guo-Liang Jiang,* Pengyin Chen, Jiaoping Zhang, Liliana Florez-Palacios, Ailan Zeng, Xianzhi Wang, Ronald A. Bowen, Amanda Miller, and Haley Berry

G.-L. Jiang, R.A. Bowen, A. Miller, and H. Berry, Agricultural Research ABSTRACT Station, Virginia State Univ., PO Box 9061, Petersburg, VA 23806; P. Soybean [Glycine max (L.) Merr.] is one of the Chen, Dep. of Crop, Soil and Environmental Sciences, Univ. of Arkansas, most important crops in the world. It is a major Fayetteville, AR 72701, current address, Univ. of Missouri, Fisher Delta source of vegetable oil for consumption and Research Center, Portville, MO 63873; J. Zhang, Plant Science Dep., protein meal for animal feeds and has also been South Dakota State Univ., Brookings, SD 57007, current address, Dep. widely used in human food industries because of Agronomy, Iowa State Univ., Ames, IA; L. Florez-Palacios and A. of its nutritive and health benefits. To provide Zeng, Dep. of Crop, Soil and Environmental Sciences, Univ. of Arkansas, useful information for soybean quality improve- Fayetteville, AR 72701; X. Wang, Plant Science Dep., South Dakota State ment, seed individual sugars, total sugar, Univ., Brookings, SD 57007, current address, College of Agriculture, protein, oil, and dietary fiber were genetically Yunnan Univ., Kunming, China. Received 11 Mar. 2018. Accepted 10 analyzed in replicated trials with 323 germplasm July 2018. *Corresponding author ([email protected], gljiang99@yahoo. lines grown in South Dakota and 137 cultivars com). Assigned to Associate Editor Owen Hoekenga. and breeding lines grown in Virginia. The results Abbreviations: HPAEC-PAD, high-performance anion-exchange indicated significant differences among the chromatography coupled with pulsed amperometric detection; HPLC, genotypes for all traits investigated. Environment high-performance liquid chromatography; NIFA, National Institute effect and genotype ´ environment interaction of Food and Agriculture; NIR, near-infrared; PI, plant introduction; were also significant in most cases. Heritability RIL, recombinant inbred line. estimates were high (94.45–97.79%) for all traits in the germplasm population, and higher in the population of breeding lines for most traits. High oybean [Glycine max (L.) Merr.] is a leading crop grown world- genotypic correlation existed between sucrose Swide for production of vegetable oil for human consumption and and total sugar, which helps improvement of provision of protein meal for animal feeds. It has also been widely digestible sugars and sweetness in soybean used in human food industries because more and more people food. However, attention should be paid to the have become aware of its health benefits. In general, soybean seeds lines with higher sucrose but lower oligosaccha- consist of approximately 40% protein, 20% oil, 35% carbohydrate, rides, since stachyose was positively associated and 5% ashes (Liu, 1999; Karr-Lilienthal et al., 2005). Soybean with total sugar. Genotypic correlations between as a dietary protein source contains 18 amino acids, including all seed sugars and protein were insignificant or essential amino acids that cannot be synthesized in human body very low in most cases, implying that altera- and conditionally essential amino acids except glutamine, the tion of seed sugars might not necessarily affect synthesis of which can be limited under given conditions. Soybean protein. In some cases, however, there might be negative correlations between seed sugars and oil is composed of five fatty acids: palmitic, stearic, oleic, linoleic, oil or dietary fiber in soybean. This study also and linolenic acids. Of the carbohydrates in soybean, important identified some unique germplasm lines with components include dietary fiber (or nonstarch polysaccharides) a desired level of a specific seed composition: one with high sucrose, five with low raffinose, Published in Crop Sci. 58:2413–2421 (2018). 15 with high total sugar, seven with high protein, doi: 10.2135/cropsci2018.03.0173 and four high in both sucrose and total sugar. © Crop Science Society of America | 5585 Guilford Rd., Madison, WI 53711 USA This is an open access article distributed under the CC BY license (https:// creativecommons.org/licenses/by/4.0/). crop science, vol. 58, november–december 2018 www.crops.org 2413 and sugars such as sucrose, raffinose, and stachyose (Choct was positively associated with protein (Hymowitz et al., et al., 2010). The proportion of seed composition deter- 1972). Wilcox and Shibles (2001) reported that concentra- mines the uses of soybean. For instance, cultivars high in tions of carbohydrates were not associated with seed yield, oil are preferred by vegetable oil and soy-diesel industries, but increased protein was coupled with decreases in oil, whereas soy food products usually need lower oil but higher total carbohydrates, and sucrose. In an evaluation of 23 protein and sugar contents. conventional and food-grade cultivars, Geater and Fehr Soybean seed protein and oil contents have been (2000) suggested that total sugar was highly correlated extensively investigated, particularly in plant breeding with the sum of protein and oil. In a study with 30 vege- and genetics from quantitative genetics to molecular table soybean genotypes, negative correlations were found mapping and candidate gene identification (Wang et al., between protein and total sugar (r = −0.52) or sucrose (r = 2014; Hwang et al., 2014; Zhang et al., 2018). Relatively, −0.43), whereas no significant correlations were detected studies on sugar and fiber content in soybean are less between oil and total sugar or sucrose (Li et al., 2012). Yu reported. Sucrose, a disaccharide and the most important et al. (2016) evaluated 35 soybean germplasm lines (mostly component of total sugar in soybean, is a free or digestible from China) for five individual and total sugars, as well as sugar and is very important for food soybean (Kumar et protein. They found that protein content was positively al., 2010; Song et al., 2013). Similarly, monosaccharides correlated with total sugar and sucrose contents but nega- glucose and fructose are also easily digested, and thus they tively correlated with fructose and glucose contents. should be worthy of exploring, in particular for food use There have historically been fewer carbohydrate- of soybeans like edamame (Song et al., 2013). However, focused studies than studies focused on protein and/ two oligosaccharides, raffinose and stachyose, cannot be or oil content, and the studies that considered all prox- digested in monogastric animals and cause flatulence imate nutrients including sugars, protein, oil, and fiber (Choct et al., 2010). A decreased concentration of both were even more limited. In most of the previous studies raffinose and stachyose is preferred in soy food industries on seed sugar content in soybean, the number of geno- such as soymilk and tofu production (Kumar et al., 2010; types used was relatively small compared with the studies Saldivar et al., 2011). on protein and oil research. Inconsistencies between the Using high-performance anion-exchange chroma- studies existed to some extent. In addition, there is lack tography coupled with pulsed amperometric detection of understanding of genetic variability of fiber content (HPAEC-PAD), Bainy et al. (2008) reported varietal in soybean, although it is an important trait in vegetable differences in carbohydrates in defatted soybean flour and soybeans that are consumed directly for fresh market and soy protein isolate byproducts among 12 soybean lines. other food-grade soybeans that are used by food product Hou et al. (2009) analyzed five individual sugars and manufacturers (Redondo-Cuenca et al., 2007). Genotypic total sugar in 241 germplasm accessions of three maturity correlation between sugars and other traits has rarely been groups using high-performance liquid chromatography discussed. Therefore, there is a need to further investigate (HPLC). They identified some plant introductions (PIs) the genetic feature of sugars and their relationships with considerably low or high in individual sugars. Cicek et other seed composition in soybean. To provide useful al. (2006) reported a high heritability for sucrose content, information for quality improvement and related research, but relatively low heritabilities for stachyose and raffinose using two populations, one consisting of 323 soybean contents in a recombinant inbred line (RIL) population germplasm accessions grown in South Dakota and one of derived from an interspecific cross. Quantitative trait loci 137 cultivars and breeding lines grown in Virginia, we associated with sucrose and oligosaccharide contents were characterized seed individual sugar, total sugar, protein, also explored (Maughan et al., 2000; Kim et al., 2005). oil, and dietary fiber content in the present study. We also Strong correlations between sugars were previously analyzed the genotypic correlations between the traits. reported (Cicek et al., 2006; Hou et al., 2009). In the study by Hou et al. (2009), the absolute values of simple MATERIALS AND METHODS correlation coefficients among individual sugar and total Genotypes and Field Trials sugar contents varied from 0.59 to 0.999, except between Experiment 1 total sugar and glucose (r = −0.27) or fructose (r = −0.24). In total, 323 soybean germplasm accessions or PIs were obtained Hymowitz et al. (1972) evaluated 60 soybean lines from from the US Soybean Germplasm Collection without specific maturity groups 00 through IV for protein, oil, total sugar, criteria of selection but mainly limited to early maturity groups, and individual sugar content. Their results suggested that since soybean is highly photoperiod sensitive and a given trial total sugar content and oil content were positively asso- cannot accommodate a wide range of maturities. Approximately ciated, and each was negatively correlated with protein 91% of the PIs are maturity group 0 and 9% belong to maturity content. Sucrose and raffinose content were positively group 00; 91% originated from China, and the remaining from correlated with oil content, whereas stachyose content other countries or unknown origins (Supplemental Table S1).

2414 www.crops.org crop science, vol. 58, november–december 2018 These PIs are all cultivated soybeans (G. max) and represent automated sampler with a 25-mL injection loop, and a Chrome- landraces and obsolete and improved cultivars. All the PIs were leon Chromatography Management System (Dionex), was used planted in a randomized complete block design with three to identify and quantify sugars. Sugar separation was done replications at three locations: Aurora in 2011, Brookings in with an analytical CarboPac PA10 pellicular anion-exchange 2012, and Watertown in 2012, all in South Dakota. Plots were resin column (25 ´ 4 mm) preceded by a CarboPac PA10 composed of a single 3-m-long row with 0.76-m row spacing, guard column (50 ´ 4 mm) and an AminoaTrap column (30 and 80 seeds were planted per plot. The field management was ´ 3 mm, Dionex). Before HPLC injection, a 24-mL aliquot of similar to general soybean production. After full maturity (R8), sugar sample was diluted with 576 mL of distilled water. Sugars all plots were individually bulk harvested. Then the seeds were were eluted with 90 mM NaOH at a flow rate of 1.0 mL min−1 dried in an air-drying chamber for a week. under isocratic conditions. Glucose, fructose, sucrose, raffinose, and stachyose in the extracts were determined and quantified Experiment 2 based on the retention time and the regression curve established One hundred and thirty-seven cultivars and breeding lines of for each of five standard sugars (Hou et al., 2009). maturity groups IV, V, and VI were grown at the Virginia State University Randolph Research Farm in Ettrick, VA, in 2014 and Experiment 2 2015, respectively. Of them, 130 were breeding lines developed About 50 g of seeds per plot was randomly taken with a small by Virginia State University soybean and edamame program cup and analyzed for protein, oil, dietary fiber, sucrose, raffi- and derived from 12 different crosses each with 10 to 12 F3:4 or nose, and stachyose content using NIR spectroscopy in a F4:5 lines evaluated (Supplemental Table S2), and the remaining DA 7250 NIR analyzer (Perten Instruments) in the Soybean were released cultivars. The breeding lines were selected on the Breeding and Genetics Laboratory at Virginia State University. basis of agronomic performance in plant rows and preliminary Glucose and fructose were not evaluated because no appro- yield trials, with connections to some extent among those lines priate calibration was available in the system. derived from similar crosses. The released cultivars (Asmara, Mooncake, N6202-8, Osage, Owens, Randolph, and UA Statistical Analysis 4805) were used as the checks of edamame or general-purpose Data of the seed composition were presented as milligrams per soybean. The breeding lines are mostly unrelated to the seven gram on a dry-weight basis. Analysis of variance was performed released cultivars and the early-maturing germplasm lines used using PROC GLM in SAS version 9.4 (SAS Institute, 2013), in Exp. 1. Four-row plots with a 4.8-m length and 0.76-m row and frequency distribution was computed in Microsoft Excel spacing were planted in a three-replicate randomized complete 2013. Heritability was estimated on a genotype mean basis block design, and ?100 plants were planted per row. After 2 2 2 2 2 (Fehr, 1987) as h = sg /(sg + se /l) for Exp. 1, where sg is full maturity, one row per replication was intended to be bulk 2 genotypic variance, se is environmental variance, and l is the harvested in 2014, but only one replication was harvested due number of environments or locations. Phenotypic and geno- to severe deer and weed damages. In 2015, two central rows typic correlations between traits were computed as described were combine harvested for each replication. by Holland (2006). Genotypic and phenotypic correlation coefficients were tested for significance using a t test as suggested Sampling and Seed Composition Analysis by Robertson (1959) and Sharma (1988). For Exp. 2, herita- 2 2 2 2 2 Experiment 1 bility was computed as h = sg /[sg + sge /l + se /(rl)], where 2 Approximately 200 g of dried seeds per plot was randomly taken sge is genotype´ environment interaction variance and r is the and ground with Perten Laboratory Mill 3600 to prepare the number of replications (Fehr, 1987). In addition, heritability for flour samples for seed composition analysis. Concentrations of protein, oil, and fiber content in Exp. 1 was also estimated in protein, oil and dietary fiber were determined by near-infrared the same way with individual plot data. (NIR) spectroscopy in a DA 7200 NIR analyzer (Perten Instruments) in the Soybean Breeding and Genetics Labora- RESULTS AND DISCUSSION tory at South Dakota State University. To be consistent with Genetic Variation and Heritability the design for sugar analysis below, the individual-plot values in Soybean Germplasm of three replications estimated per genotype were averaged for Analysis of variance showed that there were highly protein, oil, and fiber, and then the means were used as data significant differencesP ( < 0. 01) among the germplasm units for statistical analysis in a randomized complete block accessions for all traits investigated. Overall means and design with three locations used as replications. To reduce the number of samples and save resources in variation of the traits evaluated across three environments sugar analysis, one mixed flour sample per genotype was are shown in Table 1. Comparatively, glucose and fructose prepared by mixing a similar amount (5 g) of soy flour for all exhibited larger relative variations, whereas the smallest three replications in each environment. Sugar sample prepara- relative variations were observed in protein and fiber. Of tion, extraction, and analysis were conducted in the Soybean total sugar that averaged 103.61 mg g−1, sucrose (50.65 mg Breeding and Genetics Laboratory at the University of Arkansas g−1) was the largest component (48.90%), followed by following Hou et al. (2009) using HPLC. Briefly, a Dionex stachyose (39.26%) and raffinose (9.38%). Glucose and DX500 HPAEC-PAD system, which was equipped with a fructose together contributed a very small part (2.46%) GS50 pump, an ED40 pulsed amperometric detector, an AS40 to total sugar content. The result was consistent with the

crop science, vol. 58, november–december 2018 www.crops.org 2415 Table 1. Mean, variation, and estimate of heritability of seed composition trait in 323 soybean germplasm accessions across three environments. Seed composition 2 Trait Mean Range CV FGen† FEnv† h ————————— m g g −1 ————————— % % Glucose 1.68 ± 0.26 0.14 – 5.22 15.50 20.99** 2.60 95.23 Fructose 0.88 ± 0.20 0.00–2.93 23.12 21.29** 2.78 95.30 Sucrose 50.65 ± 2.20 33.63–72.82 4.34 19.88** 10.18** 94.97 Raffinose 9.72 ± 0.86 0.57–24.68 8.90 45.32** 8.22** 97.79 Stachyose 40.68 ± 2.03 14.74– 58.28 5.00 19.88** 2.72 94.97 Total sugar 103.61 ± 3.48 69.99–140.32 3.36 27.0 5* * 4.81** 96.30 Protein 418.31 ± 8.01 352.40–499.70 1.91 18.01** 762.94** 94.45 Oil 197.8 8 ± 4.60 129.70–233.00 2.23 29.90** 302.93** 96.66 Fiber 62.10 ± 1.11 55.30–80.40 1.80 21.26** 17.8 6* * 95.29 ** Significant at the 0.01 probability level.

† FGen, and FEnv are F values of significance tests for genotype and environment effects, respectively.

previously reported results (Karr-Lilienthal et al., 2005; three environments, in spite of being coupled with a lower Hou et al., 2009; Yu et al., 2016). This might be the reason oil concentration (130.9–178.9 mg g−1). Of them, PI 603712 that glucose and fructose were not often included in studies also possesses high resistance to soybean aphids (Aphis glycines on sugars in soybean (Maughan et al., 2000; Kim et al., Matsumur; Bhusal et al., 2017). 2005). Overall across three environments, concentrations As shown by ANOVA (Table 1), environmental of protein, oil, and dietary fiber averaged 418.31, 197.88, effects were also significant in six traits but insignifi- and 62.10 mg g−1, respectively (Table 1). This is similar to cant for glucose, fructose, and stachyose. In other words, previous reports (Liu, 1999; Wang et al., 2014). locations and/or years did not significantly affect the The variability of seed sugar, protein, oil, and dietary concentrations of glucose, fructose, and stachyose. By fiber observed in this study falls within a normal range of evaluating seven genotypes grown at different geograph- variation (Liu, 1999; Karr-Lilienthal et al., 2005). Frequency ical locations, Kumar et al. (2010) found that sucrose distribution was continuous for all the seed compositions content was significantly higher at a cooler location, but (Fig. 1). Except for fiber content, the frequency distributions the differences in raffinose and stachyose contents across approximated to a normal distribution, with some skewness growing locations were genotype dependent. Since no for fructose, stachyose, and oil (Fig. 1). This is consistent with data of individual plots were available for sugars in this the results reported by Cicek et al. (2006) using a population study, variance of genotype ´ environment interaction of 303 RILs derived from an interspecific cross. Hou et al. could not be tested for its significance but was used as (2009) identified many germplasm accessions with extremely the residual or error variance in the significance test for high or low concentrations in individual sugars from a diverse genotypic and environmental or locational effect. This population of 206 PIs originating from multiple countries might affect the results. However, for protein, oil, and and regions. In the present study, no germplasm acces- dietary fiber, ANOVAs with and without individual plot sion with that high of glucose and/or fructose content was data exhibited similar results to the significance tests for found. However, some unique germplasm lines with desired genotypic and environmental effects. Analysis of variance concentrations in specific seed composition were identified based on individual plot data also showed a highly signifi- (Table 2) that could be used as parents in soybean breeding. cant genotype ´ environment interaction (P < 0. 01) These lines include five low-raffinose PIs (PI 603443B, PI for these traits (data not shown). In addition, correlation 358323, PI 603426E, PI 603429A, and PI 612753A; raffinose coefficients between environments or locations were high content < 3 mg g−1), one high-sucrose PI (PI 291329; sucrose for all the traits, averaging 0.882 and ranging from 0.809 content > 64 mg g−1), 15 PIs with >120 mg g−1 in total sugar for protein between 2012 Watertown and 2011 Aurora content, and four PIs high in both sucrose and total sugar to 0.944 for raffinose between 2012 Watertown and 2011 (PI 538395, PI 458827, PI 597651, and PI 597652). These PIs Aurora. Therefore, the results presented without indi- have not been previously reported to possess such specific vidual plot data are reliable and should be informative. sugar features, and no genetic connection between them and Heritability of protein and oil content in soybean seed was previously reported lines has been evidenced by molecular estimated as high in a recent study with two RIL populations marker and/or pedigree information (Hou et al., 2009; Jo (Wang et al., 2014). In the present study, high heritability was et al., 2018). We also confirmed seven high-protein PIs (PI also found for all traits, from 94.45% for protein to 97.79% for 468909, PI 612758A, PI 468910, PI 319536B, PI 612759B, PI raffinose (Table 1). Interestingly, the estimates of heritability 603712, and PI 597467) with an average of >462 mg g−1 over for protein, oil, and dietary fiber with individual plot data

2416 www.crops.org crop science, vol. 58, november–december 2018 Fig. 1. Frequency distribution of 323 soybean germplasm accessions in seed composition based on average over three environments. not used but the means of three replications were the same (Table 3). Among the individual sugars, low significant as or very close to those values (94.50, 96.53, and 95.05% correlations were found between glucose and fructose (r = for protein, oil and fiber, respectively) estimated when indi- 0.239), sucrose and raffinoser ( = 0.264), and sucrose and vidual plot data were used in the computation. This indicated stachyose (r = 0.411). No correlations with |r| > 0.16 existed that genotypic effects dominated in phenotypic variation, between other pairs of individual sugars. This is different and selection would play an important role in trait improve- from the results of Hou et al. (2009). They reported that ment. In a study with 31 vegetable soybean genotypes, there were higher simple correlations among individual Mebrahtu and Mohamed (2006) reported lower estimates sugars, either positive or negative with |r| = 0.590 to 0.989. of heritability for individual sugars and total sugar, with h2 The reason for the inconsistency might be that some germ- = 26.6 to 56.1%. Cicek et al. (2006) reported that sucrose plasm lines with extremely high or low individual sugars content exhibited a higher heritability, but stachyose and were included in their study (Hou et al., 2009). However, raffinose content had relatively low heritabilities in a popula- significant correlations between total sugar and sucrose, tion of 308 RILs derived from an interspecific cross. The raffinose, or stachyose were detected in this study, with r = difference might be mainly attributed to different types and 0.828, 0.541, and 0.759 (P < 0.01), respectively. numbers of soybean germplasm lines used. No heritability To further elucidate the relationships among traits, we was previously reported for fiber content in soybean. Further performed a genotypic correlation analysis. The results investigations will be of interest. indicated that both phenotypic and genotypic correlations were ignorable or very low among individual Phenotypic and Genotypic Correlations sugars in most cases, with an average of 0.141 ranging between Traits in Soybean Germplasm from 0.012 to 0.392 and 0.156 ranging from 0.004 to Pearson correlation analysis showed that the simple corre- 0.422 for absolute coefficients of phenotypic and geno- lations between traits investigated were not or slightly typic correlations, respectively (Table 4). Similar to simple significant (|r| < 0.18) for two-thirds of 36 pairs of traits correlations described above, there were weak or low

crop science, vol. 58, november–december 2018 www.crops.org 2417 Table 2. Means of seed composition traits in soybean germplasm accessions with low raffinose, high sucrose, high total sugar, or high protein content over three environments. PI no. Glucose Fructose Sucrose Raffinose Stachyose Total sugar Protein Oil Fiber ———————————————————————————————————————— m g g −1 ———————————————————————————————————————— PI 603443B 1.23 0.15 52.58 0.73† 40.10 94.80 440.45 169.64 68.60 PI 603429A 1.38 0.68 50.29 2.25 40.34 94.94 416.10 193.07 66.88 PI 603426E 1.57 0.08 52.46 2.63 42.20 98.91 421.67 187.3 0 63.86 PI 612753A 1.84 0.49 39.70 2.85 36.52 81.40 440.55 154.77 70.41 PI 358323 2.03 1.85 50.30 2.95 39.68 96.81 397.31 211.10 61.50 PI 291329 2.34 1.61 65.45 7.38 42.10 118.89 429.77 180.92 61.06 PI 538395 2.41 1.03 71.76 16.32 44.42 135.94 400.74 192.74 60.31 PI 597651 1.23 0.78 65.58 20.47 45.95 134.01 416.07 189.14 63.04 PI 458827 1.82 1.94 64.42 14.53 44.65 127.36 402.04 192.52 61.14 PI 597652 1.61 0.85 64.01 12.39 41.83 120.70 421.28 185.57 61.78 PI 639559A 1.34 0.50 63.89 23.20 39.21 128.14 422.95 190.44 68.41 PI 603437C 1.10 1.69 60.70 13.28 48.55 125.32 403.20 197.19 62.86 PI 603301A 1.59 0.47 61.47 16.76 45.48 125.77 425.91 197.0 0 59.73 PI 603306 1.34 1.57 60.60 17.3 4 42.23 123.07 426.66 196.87 60.21 PI 603292 2.29 0.91 55.57 15.60 46.96 121.33 424.91 202.71 59.54 PI 578372 2.53 0.63 55.95 12.51 49.52 121.15 405.97 212.00 62.12 PI 561332 2.61 1.07 48.95 19.66 50.32 122.61 448.30 190.87 59.90 PI 511867 2.70 0.37 61.50 17.0 4 46.57 128.17 385.57 212.39 60.34 PI 452432 1.05 0.55 60.62 11.39 47.14 120.75 395.38 201.47 62.17 PI 464886 0.17 0.58 61.27 11.30 48.89 122.21 382.36 208.50 63.00 PI 467323A 2.84 0.51 54.33 13.04 53.25 123.96 386.06 209.79 62.81 PI 467345 1.42 1.18 62.42 15.58 51.98 132.58 401.12 197.77 62.20 PI 437982 0.64 0.63 58.10 13.77 50.14 123.28 417.78 196.79 63.73 PI 291312 1.97 1.81 60.72 10.79 48.68 123.97 417.51 199.93 60.53 PI 291313 2.54 2.21 62.30 11.73 43.67 122.45 421.34 194.82 61.14 PI 319536B 2.20 1.55 41.69 9.25 35.96 90.65 462.61 187.91 60.03 PI 468909 1.42 0.60 46.82 9.92 40.16 98.93 488.36 130.91 75.12 PI 468910 1.39 0.32 46.89 10.91 31.69 91.19 471.74 135.35 78.03 PI 597467 2.60 0.91 52.00 8.73 46.78 111.03 462.41 157.38 59.38 PI 603712 0.86 1.45 49.38 6.85 39.38 97.91 462.90 157.97 61.10 PI 612758A 2.10 1.22 46.19 5.08 40.51 95.09 480.66 147.47 71.19 PI 612759B 3.31 1.65 55.73 16.31 40.67 117.67 463.56 153.64 72.31

LSD0.05 0.42 0.33 3.54 1.39 3.27 5.61 12.93 7.42 1.79 † Bolding indicates that the value fell within the range of desired concentration for a given seed constituent (i.e., raffinose content < 3 mg g−1, sucrose content > 64 mg g−1, total sugar content > 120 mg g−1, and protein content > 462 mg g−1).

Table 3. Coefficients of simple correlations between seed composition traits based on 323 soybean germplasm accessions grown in three environments. Trait Fructose Sucrose Raffinose Stachyose Total sugar Protein Oil Fiber Glucose 0.239** −0.049 0.023 −0.117* 0.000 0.065 −0.051 0.003 Fructose 0.008 −0.071 −0.151** −0.027 0.182** −0.082 −0.144* Sucrose 0.264** 0.411** 0.828** −0.249** 0.104 −0.066 Raffinose 0.159** 0.541** −0.104 0.081 −0.109 Stachyose 0.759** −0.139* 0.093 −0.024 Total sugar −0.223** 0.121* −0.090 Protein −0.779** 0.250** Oil −0.678** *,** Significant at the 0.05 and 0.01 probability levels, respectively. genotypic correlations between glucose and fructose, and High genotypic correlation between sucrose and total between sucrose and raffinose or stachyose. In addition, sugar helps improvement of digestible sugar and sweetness total sugar content exhibited a moderate genotypic corre- in food-grade soybean, whereas increases in total sugar lation with raffinose and a higher genotypic correlation might be accompanied by increased indigestible oligosac- with sucrose or stachyose, but no noticeable genotypic charides (raffinose and stachyose) to some extent. Thus, correlation was found between total sugar and glucose or more attention should be paid to the lines with higher fructose. This is consistent with the proportion of indi- sucrose but lower oligosaccharides, such as PI 597652 and vidual sugars contributing to the total sugar content. PI 291313 (Table 2).

2418 www.crops.org crop science, vol. 58, november–december 2018 Table 4. Coefficients of phenotypic (above diagonal) and genotypic (below diagonal) correlations between seed composition traits based on 323 soybean germplasm accessions grown in three environments. Trait Glucose Fructose Sucrose Raffinose Stachyose Total sugar Protein Oil Fiber Glucose 0.219** −0.049 0.021 −0.110* 0.001 0.068 −0.053 −0.001 Fructose 0.252** 0.012 −0.068 −0.134** −0.017 0.175** −0.087 −0.131* Sucrose −0.050 0.004 0.250** 0.392** 0.822** −0.225** 0.095 −0.060 Raffinose 0.024 −0.072 0.271** 0.154** 0.529** −0.100 0.080 −0.104* Stachyose −0.122* −0.168** 0.422** 0.163** 0.756** −0.122* 0.085 −0.022 Total sugar −0.001 −0.039 0.832** 0.548** 0.760** −0.203** 0.113* −0.084 Protein 0.071 0.220** −0.263** −0.104 −0.144* −0.231** −0.764** 0.189** Oil −0.065 −0.105** 0.107 0.082 0.095 0.123* −0.781** −0.637** Fiber 0.006 −0.151** −0.067 −0.115* −0.022 −0.092 0.270** −0.696** *,** Significant at the 0.05 and 0.01 probability levels, respectively.

Fructose, sucrose, stachyose, and total sugar exhibited (P < 0. 01) and a relatively large range of variation (Table 5), low but different phenotypic and genotypic correlations in spite of smaller values compared with those of soybean with protein (Table 4). No noticeable genotypic corre- germplasms in Exp. 1. Similar to Exp. 1, coefficients of vari- lations were found between other sugars and protein. ation for raffinose, sucrose, and stachyose were larger than In a trial with 20 high-oil and 20 high-protein soybean others. Year or environment effects and genotype ´ envi- genotypes, Hartwig et al. (1997) also revealed that the ronment interaction were significant for all the traits as well. correlation between protein and stachyose + raffinose was Higher estimates of heritability were found for protein, oil, negative but nonsignificant. In our study, the genotypic sucrose, stachyose, and total sugar content than for raffinose correlations between the concentrations of sugars and oil, and fiber content (Table 5). This indicated that further selec- as well as dietary fiber, were ignorable or very low in most tion for the traits in current breeding populations would be cases. This indicated that the improvement of oil did not effective, although raffinose and dietary fiber content were obviously affect sugar content in soybean, and alteration of more likely affected by environments. Basically, the results protein or dietary fiber would have no or limited impact on were consistent with those in Exp. 1 in most cases, although sugars as well. Similarly, selection for sugars might not have the estimates of heritability in this experiment were smaller a greatly negative impact on protein, oil, or dietary fiber. than the values in Exp. 1. For raffinose, low heritability esti- There were higher negative phenotypic and genotypic mated in Exp. 2 might be largely attributed to a small range correlations between protein and oil content (Table 4). of variation, compared with Exp. 1. This is consistent with previously reported results (Wang et Correlation analysis indicated that no or very weak al., 2014). Dietary fiber content exhibited higher negative phenotypic and genotypic correlations existed between phenotypic and genotypic correlations with oil concentra- sucrose and stachyose, protein, or oil and between raffi- tion, but lower positive correlations with protein content. nose and protein, oil, or fiber, with an R2 ranging from It implied that compared with protein, alteration of oil 0.002 to 0.044 for phenotypic correlation and from 0.008 had a larger impact on dietary fiber content in soybean. In to 0.067 for genotypic correlation (Table 6). Similar to addition, the phenotypic and genotypic correlation coef- Exp. 1, total sugar content was positively associated with ficients were very similar, and they were also close to the sucrose and/or stachyose content but had no noticeable simple correlation coefficients (Table 3), suggesting that association with raffinose. Different from Exp. 1, however, the effects of environment on relationships among the there were moderately negative phenotypic and genotypic traits might not matter much in similar situations. correlations between total sugar or stachyose and oil or dietary fiber, but a weak positive genotypic correlation Trait Genetic Characterization and was found between total sugar and protein. It could be Relationships in Soybean Breeding Lines assumed that this phenomenon is mainly due to the differ- To effectively perform selection and further improvement ence in materials used in the two experiments. Therefore, for quality traits in breeding materials, we evaluated 130 the potential reverse impact of selection for total sugar breeding lines selected from different crosses of maturity concentration on oil should be taken into consideration in groups IV, V, and VI and seven released cultivars for the further improvement of the breeding materials. Negative concentrations of seed sugars, protein, oil, and dietary fiber. genetic association was also found between protein and As discussed above, these breeding lines are not genetically oil in this experiment with breeding lines and released related to the seven released cultivars and the PIs evaluated cultivars, which was consistent with Exp. 1 and previously in Exp. 1, although some lines were derived from similar reported results (Wang et al., 2014). Dietary fiber exhib- crosses. Among the breeding lines and released cultivars, ited a negative correlation with protein but a positive all traits evaluated exhibited a highly significant difference correlation with oil, which was inconsistent with Exp. 1.

crop science, vol. 58, november–december 2018 www.crops.org 2419 Table 5. Mean, variation, and estimate of heritability of seed composition traits in 137 soybean lines in two environments. Seed composition 2 Trait Mean Range CV FGen† FGen/Env† FEnv† h ———————— m g g −1 ———————— % % Sucrose 48.57 ± 3.14 26.30–63.00 6.46 3.63** 1.37* 479.98** 64.51 Raffinose 7.6 5 ± 0.72 5.70–12.50 9.47 2.03** 1.39* 12.27** 33.60 Stachyose 35.57 ± 1.88 25.90–56.90 5.27 10.79** 2.79** 7.3 4* * 78.50 Total sugar 91.79 ± 3.79 72.60 –119.00 4.12 5.64** 1.81** 305.78** 71.36 Protein 429.75 ± 11.00 330.20–493.83 2.56 7.91* * 1.63** 126.28** 81.51 Oil 194.35 ± 6.80 150.50–230.40 3.50 9.29** 1.87** 50.83** 82.40 Fiber 59.74 ± 2.11 51.00–71.90 3.53 4.77** 2.41** 49.67** 54.96 *,** Significant at the 0.05 and 0.01 probability levels, respectively.

† FGen, FGen/Env, and FEnv are F values of significance tests for genotype, genotype ´ environment interaction, and environment effects, respectively. Table 6. Coefficients of phenotypic (above diagonal) and genotypic (below diagonal) correlations between seed composition traits based on 137 soybean lines grown in two environments. Trait Sucrose Raffinose Stachyose Total sugar Protein Oil Fiber Sucrose −0.292** 0.174** 0.768** −0.046 −0.155* −0.466** Raffinose −0.481* −0.012 −0.061 0.209** 0.092 0.134* Stachyose 0.089 0.479* 0.753** 0.436** −0.565** −0.584** Total sugar 0.672** 0.110 0.803** 0.270** −0.458** −0.675** Protein −0.116 0.245 0.431** 0.239* −0.517** −0.464** Oil −0.250* 0.198 −0.605** −0.580** −0.421** 0.408** Fiber −0.510** −0.131 −0.493** −0.699** −0.792** 0.698** *,** Significant at the 0.05 and 0.01 probability levels, respectively. CONCLUSION Acknowledgments Significant differences in concentrations of seed indi- This study was supported in part by the United Soybean Board, vidual sugars, total sugar, protein, oil, and dietary fiber the USDA National Institute of Food and Agriculture (NIFA) existed among the soybean germplasm accessions or Evans-Allen Research Program, and the USDA-NIFA Capac- released cultivars and breeding lines. All the traits exhib- ity Building Grant Program (funding awarded to G.-L. Jiang). This article is a contribution of the Virginia State University, ited a moderate to high heritability except for raffinose Agricultural Research Station (Journal Series no. 350). in breeding materials, implying that selection for desired phenotypes would be effective. Unique germplasm lines References with desired concentration of a specific seed composition, Bainy, E.M., S.M. Tosh, M. Corredig, V. Poysa, and L. Wood- such as high sucrose, low raffinose, high total sugar, and row. 2008. Varietal differences of carbohydrates in defatted high protein, were identified and could be used as parents soybean flour and soy protein isolate by-products. Carbohydr. in soybean breeding. Higher positive genotypic correla- Polym. 72:664–672. doi:10.1016/j.carbpol.2007.10.008 tion existed between sucrose and total sugar, helping with Bhusal, S., G.-L. Jiang, Q. Song, P.B. Cregan, D. Wright, and J.L. Gonzalez-Hernandez. 2017. Genome-wide detection of improvement of digestible sugar and sweetness in soybean genetic loci associated with soybean aphid resistance in soy- food, whereas increases in total sugar might be accompa- bean germplasm PI 603712. Euphytica 213:144. doi:10.1007/ nied by increased indigestible oligosaccharides (especially s10681-017-1933-1 stachyose) to some extent. Genotypic correlations between Cicek, M.S., P. Chen, M.A. Saghai Maroof, and G.R. Buss. 2006. seed sugars and protein were ignorable or relatively low in Interrelationships among agronomic and seed quality traits most cases. Therefore, improvement of seed sugars may in an interspecific soybean recombinant inbred population. not necessarily affect protein. In some cases, however, Crop Sci. 46:1253–1259. doi:10.2135/cropsci2005.06-0162 there could be a negative correlation between seed sugars Choct, M., Y. Dersjant-Li, J. McLeish, and M. Peisker. 2010. Soy oligosaccharides and soluble non-starch polysaccharides: A and oil or dietary fiber in soybean. review of digestion, nutritive and anti-nutritive effects in pigs and poultry. Asian-Australas. J. Anim. Sci. 23:1386–1398. Conflict of Interest doi:10.5713/ajas.2010.90222 The authors declare that there is no conflict of interest. Fehr, W.R. 1987. Principles of cultivar development Vol. 1. The- ory and technique. MacMillian Publ. Co., New York. Supplemental Material Available Geater, C.W., and W.R. Fehr. 2000. Association of total sugar Supplemental material for this article is available online. content with other seed traits of diverse soybean cultivars. Crop Sci. 40:1552–1555. doi:10.2135/cropsci2000.4061552x Hartwig, E.E., T.M. Kuo, and M.M. Kenty. 1997. Seed protein and its relationship to soluble sugars in soybean. Crop Sci. 37:770– 773. doi:10.2135/cropsci1997.0011183X003700030013x

2420 www.crops.org crop science, vol. 58, november–december 2018 Holland, J.B. 2006. Estimating genotypic correlations and their Mebrahtu, T., and A. Mohamed. 2006. Genetic variation for standard errors using multivariate restricted maximum likeli- green pod yield and quality among vegetable soybean hood estimation with SAS Proc MIXED. Crop Sci. 46:642– genotypes. J. Crop Improv. 16:113–130. doi:10.1300/ 654. doi:10.2135/cropsci2005.0191 J411v16n01_08 Hou, A., P. Chen, J. Alloatti, D. Li, L. Mozzoni, B. Zhang, and A. Redondo-Cuenca, A., M.J. Villanueva-Suarez, M.D. Rodriguez- Shi. 2009. Genetic variability of seed sugar content in world- Sevilla, and I. Mateos-Aparicio. 2007. Chemical composition wide soybean germplasm collections. Crop Sci. 49:903–912. and dietary fiber of yellow and green commercial soybeans doi:10.2135/cropsci2008.05.0256 (Glycine max). Food Chem. 101:1216–1222. doi:10.1016/j. Hwang, E.Y., Q.J. Song, G.F. Jia, J.E. Specht, D.L. Hyten, J. Costa, foodchem.2006.03.025 and P.B. Cregan. 2014. A genome-wide association study of Robertson, A. 1959. The sampling variance of the genetic correla- seed protein and oil content in soybean. BMC Genomics 15:1. tion coefficient. Biometrics 15:469–485.doi:10.2307/2527750 doi:10.1186/1471-2164-15-1 Saldivar, X., Y.-J. Wang, P. Chen, and A. Hou. 2011. Changes Hymowitz, T., F.I. Collins, J. Panczner, and W.M. Walker. 1972. in chemical composition during soybean seed develop- Relationship between the content of oil, protein, and sugar in ment. Food Chem. 124:1369–1375. doi:10.1016/j.food- soybean seed. Agron. J. 64:613–616. doi:10.2134/agronj1972. chem.2010.07.091 00021962006400050019x SAS Institute. 2013. The SAS system for Windows. Release 9.4. Jo, H., J.-D. Lee, and K.D. Bilyeu. 2018. Environmental stability SAS Inst., Cary, NC. of carbohydrate profiles in different soybean genotypes. Crop Sharma, J.R. 1988. Statistical and biometrical techniques in plant Sci. 58:773–782. doi:10.2135/cropsci2017.08.0497 breeding. New Age Int., New Delhi. Karr-Lilienthal, L.K., C.T. Kadzere, C.M. Grieshop, and G.C. Song, J., C. Liu, D. Li, and Z. Gu. 2013. Evaluation of sugar, Fahey, Jr. 2005. Chemical and nutritional properties of soy- free amino acid, and organic acid compositions of dif- bean carbohydrates as related to nonruminants: A review. Liv- ferent varieties of vegetable soybean (Glycine max [L.] est. Prod. Sci. 97:1–12. doi:10.1016/j.livprodsci.2005.01.015 Merr). Ind. Crops Prod. 50:743–749. doi:10.1016/j.ind- Kim, H.-K., S.-T. Kang, J.-H. Cho, M.-G. Choung, and D.-Y. crop.2013.08.064 Suh. 2005. Quantitative trait loci associated with oligosaccha- Wang, X., G.-L. Jiang, M. Green, R.A. Scott, Q. Song, D.L. ride and sucrose contents in soybean (Glycine max L.). J. Plant Hyten, and P.B. Cregan. 2014. Identification and valida- Biol. 48:106–112. doi:10.1007/BF03030569 tion of quantitative trait loci for seed yield, oil and protein Kumar, V., A. Rani, L. Goyal, A.K. Dixit, J.G. Manjaya, J. Dev, contents in two recombinant inbred populations of soybean. and M. Swamy. 2010. Sucrose and raffinose family oligosac- Mol. Genet. Genomics 289:935–949. doi:10.1007/s00438- charides (RFOs) in soybean seeds as influenced by genotype 014-0865-x and growing location. J. Agric. Food Chem. 58:5081–5085. Wilcox, J.R., and R.M. Shibles. 2001. Interrelationships among doi:10.1021/jf903141s seed quality attributes in soybean. Crop Sci. 41:11–14. Li, Y.-S., M. Du, Q.-Y. Zhang, G.-H. Wang, M. Hashemi, doi:10.2135/cropsci2001.41111x and X.-B. Liu. 2012. Greater differences exist in seed pro- Yu, X., F. Yuan, X. Fu, and D. Zhu. 2016. Profiling and rela- tein, oil, total soluble sugar and sucrose content of veg- tionship of water-soluble sugar and protein compositions in etable soybean genotypes [‘Glycine max’ (L.) Merrill] in soybean seeds. Food Chem. 196:776–782. doi:10.1016/j.food- Northeast China. Aust. J. Crop Sci. 6:1681–1686. chem.2015.09.092 Liu, K.S. 1999. Soybeans: Chemistry, technology, and utilization. Zhang, J., X. Wang, Y. Lu, S. Bhusal, Q. Song, P.B. Cregan, Aspen Publ., Gaithersburg, MD. Y. Yen, M. Brown, and G.-L. Jiang. 2018. Genome-wide Maughan, P.J., M.A. Saghai Maroof, and G.R. Buss. 2000. scan for seed composition provides insights into soybean Identification of quantitative trait loci controlling sucrose quality improvement and the impacts of domestication content in soybean (Glycine max). Mol. Breed. 6:105–111. and breeding. Mol. Plant 11L460–472. doi:10.1016/j. doi:10.1023/A:1009628614988 molp.2017.12.016

crop science, vol. 58, november–december 2018 www.crops.org 2421