Geographical distribution of GmTfl1 alleles in Chinese soybean varieties

Guifeng Liu，Lin Zhao，Benjamin J.Averitt，Ying Liu，Bo Zhang，Ruzhen Chang，Yansong Ma，Xiaoyan Luan，Rongxia Guan，*，Lijuan Qiua，，**

aSoybean Research Institute，Northeast Agricultural University，Harbin 150030，ChinabThe National Key Facility for Crop Gene Resources and Genetic Improvement（NFCRI），Institute of Crop Science，Chinese Academy of Agricultural Sciences，Beijing 100081，ChinacDepartment of Crop and Soil Environmental Sciences，VA Polytechnic Institute and State University，Blacksburg 24061，USAdHeilongjiang Academy of Agricultural Sciences，Harbin 150086，China

?

Geographical distribution of GmTfl1 alleles in Chinese soybean varieties

Guifeng Liua，1，Lin Zhaoa，1，Benjamin J.Averittc，Ying Liub，Bo Zhangc，Ruzhen Changb，Yansong Mad，1，Xiaoyan Luand，1，Rongxia Guanb，*，Lijuan Qiua，b，**

aSoybean Research Institute，Northeast Agricultural University，Harbin 150030，China
bThe National Key Facility for Crop Gene Resources and Genetic Improvement（NFCRI），Institute of Crop Science，
Chinese Academy of Agricultural Sciences，Beijing 100081，China
cDepartment of Crop and Soil Environmental Sciences，VA Polytechnic Institute and State University，Blacksburg 24061，USAdHeilongjiang Academy of Agricultural Sciences，Harbin 150086，China

A R T I C L E I N F O

Article history：

Received in revised form

24 May 2015

Accepted 23 June 2015

Available online 2 July 2015

Soybean

Stem growth habit GmTfl1

Distribution

A B S T R A C T

Stem growth habit is an important agronomic trait in soybean and is subject to artificial selection.This study aimed to provide a theory for genotypic selection of stem growth habit for breeding purposes by analyzing the alleles of GmTfl1 gene in Chinese soybean varieties and establishing a database of GmTfl1 variation.Using knowledge of insertion and deletion（Indel）in the non-coding region and four single-nucleotide polymorphisms （SNPs）in the coding sequences of the GmTfl1 gene，four CAPS and one Indel markers were developed and used to test 1120 Chinese soybean varieties.We found that the dominant GmTfl1 allele was prevalent in accessions from the Northern ecoregion，whereas the recessive allele，Gmtfl1，was more common in the Southern ecoregion，and the proportions of GmTfl1 and recessive alleles were respectively 40.1%and 59.9%in the Huang-Huai ecoregion.The proportion ofGmTfl1decreased and thatof Gmtfl1increased，gradually from north to south.Allele GmTfl1-a was present in higher proportions in the Huang-Huai spring，Huang-Huai summer，and Northern spring sub-ecoregions than that in the other sub-ecoregions.GmTfl1-b was common in the Northeast spring，Northern spring and Southern summer sub-ecoregions.Gmtfl1-ta was found mainly in the Huang-Huai spring，Huang-Huai summer and Southern spring sub-ecoregions.The Gmtfl1-ab allele was distributed in all six soybean sub-ecoregions.The Gmtfl1-bb allele was distributed mainly in the Huang-Huai spring and summer and Southern spring and summer sub-ecoregions，but the Gmtfl1-tb allele was detected only in the Huang-Huai summer sub-ecoregion.The distributions of GmTfl1 and Gmtfl1 have shown no large changes in nearly 60 years of breeding，but the frequency of the recessive genotype Gmtfl1 has shown a rising trend in the last 20 years.This study provides a theoretical foundation for breeding new soybean varieties for different ecoregions.

1.Introduction

The stem growth habit of soybean［Glycine max（L.）Merr.］，also known as the stem termination podding habit，refers to the flowering and podding habit of soybean.It is not only important as a morphological and ecological trait［1］but can be used to study theories of soybean origin and evolution［2］.Generally，stem growth habit is divided into three types:determinate，semi-determinate，and indeterminate.These types are closely associatedwithseveralotherimportanttraitsincluding height，flowering period，node number，growth period，water use efficiency，and yield［3，4］.Under high water and nutrient conditions，semi-determinate varieties are most likely to develop higher yield potentials［5，6］.Varieties with different stem growth habits also differ in dry matter accumulation，yield formation period，and other physiological traits that can affect the cultural practices that maximize yield［7］.As mentioned above，the accurate identification of stem growth habit is important for producing commercially viable soybean varieties. However，changing environments can markedly influence the accurate identification of a particular variety's determinacy［8］. Previously，researchers have identified the stem growth habit mainly by observing the difference of inflorescence and compound leaf development at the top of the stem during the floweringperiod［9］，thelength+widthratioofthemiddleleaflet of the top leaf to that of the biggest leaf of plants，dry matter accumulation with or without inflorescence，and the number of nodes above a given point［10—12］.However，each of these criteria can be greatly influenced either directly or indirectly by environment and human error.

Stem growth habit is genetically controlled by two loci，Dt1 and Dt2［8，13］.Of these，Dt1 is the primary locus.The indeterminate stem growth habit controlled by the Dt1/Dt1 genotypeexhibitsadominantorcompletelydominantroleover the determinate stem growth habit controlled by dt1/dt1［14，15］. Dt2 has a nearly dominant role over the recessive dt2.In the Dt1/Dt1 background，the Dt2/Dt2 genotype shows a determinate stem growth habit and dt2/dt2 an indeterminate one.Moreover，dt1 exerts an epistatic effect over Dt2 and dt2，explaining the determinate growth of soybeans with genotype Dt2/Dt2 and dt2/dt2inadt1/dt1geneticbackground［8］.BothLiuetal.［16］and Tian et al.［17］cloned the Dt1 gene in soybean and named it GmTfl1.They also showed that the change from indeterminacy to determinacy is controlled by the substitution of four nonsynonymous amino acids.The two dominant alleles are named respectively GmTfl1-a and GmTfl1-b and the four recessive alleles are named Gmtfl1-ta，Gmtfl1-ab，Gmtfl1-bb，and Gmtfl1-tb［17］.By sequencing，Tian et al.［17］identified the stem growth genotype of 195 soybean varieties from a minicore collection and someAmerican ancestralparents and bred lines.

In recent years，with the rapid development of functional genomics，a new type of functional molecular marker based on genetic variation has been developed.The development of functional molecular markers is based on single-nucleotide polymorphisms（SNPs）and insertion/deletion（Indel）sites in functional gene sequences.They can be used not only to identify and select the alleles in natural or breeding populations but to assist in pyramiding different genes with related markers［18］.Functional markers have been used in wheat ［19］，barley［20］，rye［21］，maize［22］，and rice［23］，and have been developed for efficiently identifying genes including GmNARK［24］，Gy4，mutant α［25］，and Gmbadh2-2［26］.

In this study，four cleaved amplified polymorphic sequences（CAPS）markers were developed on the basis of four nonsynonymous SNPs and one insertion/deletion（Indel）marker developed from intron variation.The GmTfl1 genotype was identified among 1120 varieties bred in China over the past 60 years.The distribution of different alleles was analyzed and the results provide a theoretical foundation for breeding new varieties adapted to different ecoregions.

2.Materials and methods

2.1.Materials

A total of 1120 soybean varieties bred from 1954 to 2012 were provided by the Chinese National Soybean Genebank at the Institute of Crop Science，Chinese Academy of Agricultural Sciences.Phenotypic data of stem growth habit were obtained from the records of the National Germplasm Evaluation Program of China（NGEPC）.The soybean varieties were from three main ecological regions:the Northern（NR），Huang-Huai （HR），and Southern（SR）ecoregions，which are distinguished by cropping system and subclassified into 10 subregions［17，27］. Varieties used in this study were from six of the 10 subregions，referred to as Northeast spring，North spring，Huang-huai spring，Huang-huai summer，South spring，and South summer.

2.2.Leaf tissue collection and genomic DNA extraction

The materials were planted at the Shunyi Research Station of Chinese Academy of Agricultural Sciences in Beijing in the spring of 2011 and Heilongjiang.Each variety was planted in a 4-row，3-m long plot with 0.45-m spacing.Leaves were collected from 8 to 10 plants of each variety and frozen in liquid nitrogen.DNA was isolated with a Genomic DNA Purification Kit（Thermo Fisher Scientific）.

2.3.Marker development and genotyping

Gene functional markers were developed based on the sequence variation of different GmTfl1 alleles.Indel marker TFL1ab（with primers5′-GGTTGAGATTGAGGGTGGTG-3′and5′-GGGGGACATT TGCTTAATGA-3′）was developed based on AA and AAGC insertion at positions 285 and 306 in the first intron of GmTfl1-b and used to distinguish GmTfl1-a and GmTfl1-b.The cleaved amplified polymorphic sequence（CAPS）marker tfl1ta（5′-CCAC ACTCCCTTCTTCCTCA-3′and 5′-TCACTAGGGCCAGGAACATC-3′）was developed based on the G-＞T variation at position 185of the first exon of Gmtfl1-ta.CAPS marker tfl1b（5′-TGCCACATT TGGTAGG TTCA-3′and 5′-CATAGCATACACACTAGTCA-3′）was developed based on three single-nucleotide variations（C-＞T，G-＞A，and A-＞T）at positions 1130，1181，and 1288 of exon 4.

PCR was conducted in the T100 Thermal Cycler（Bio-Rad，Hercules，CA，USA）.PCR products of the Indel marker were separated with polyacrylamide gel electrophoresis.For the CAPS markers，PCR products were digested with appropriate restriction enzymes and separated on 1.8%agarose gel.

2.4.Statistical analysis

Statistical analysis and allelic frequency distribution were performed with Microsoft Excel 2010.The Shannon—Weaver genetic diversity index computation formula was used to calculate the Dt1 loci genetic diversity of different classified groups by the following formula［28］:

where Piis the fraction of i allelic mutation occurrences among all Dt1 alleles.

3.Results

3.1.Stem growth habits of soybean varieties in different ecoregions

We analyzed the distribution of growth habit of 1120 soybean varieties in the three large soybean growingecoregions in China. In Northern ecoregion，a majority of varieties（51.4%）exhibited a semi-determinate growth habit.A high proportion（35.5%）of indeterminate growth habit was also found in this subregion，whereasthedeterminategrowthhabitaccountedforonly13.0%. Of the varieties from the Huang-Huai ecoregion，56.6%were determinate，32.6%ofsemi-determinate，andonly10.8% indeterminate.Nearly all（93.0%）varieties from the Southern ecoregion were determinate，with only 5.3%semi-determinate and 1.8%indeterminate（Fig.1）.

Fig.1-Distribution of three stem growth habits in three ecoregions.NR：Northern ecoregion，HR：Huang-Huai ecoregion，SR：Southern ecoregion.

3.2.Development of molecular markers for different Dt1 genotypes

The Indel marker of TFL1ab was designed to distinguish GmTfl1-a and GmTfl1-b according to the AA and AAGC deletion in GmTfl1-b（Fig.2-a）.Varieties containing the insertions yield fragments of 245 bp corresponding to the GmTfl1-a allele，whereas those lacking the insertions yield 239-bp fragments corresponding to the GmTfl1-b allele（Fig.2-b）.

A CAPS marker was designed as tfl1ta based on the G to T mutation in the first exon of GmTfl1.The PCR product of Gmtfl1-ta was 537 bp and could not be digested by Ava II，but the other genotypes could be digested into two fragments of 271 bp and 266 bp，respectively.

The 511 bp fragment amplified by CAPS marker tfl1b includes three SNP mutations corresponding to Gmtfl1-bb，Gmtfl1-tb，and Gmtfl1-ab，respectively.The PCR product of Gmtfl1-bb was digested into two fragments with（140 and 371 bp）by Hind III，whereas the products of other genotypes could not be cleaved.The products of Gmtfl1-tb were digested by Ac c I into two fragments with lengths 280 and 231 bp，whereas the other genotypes were cut into three fragments of 89，191，and 231 bp.The PCR products for allele Gmtfl1-ab could be digested by HpyCH4III into three fragments with lengths of 102，194，and 215 bp，whereas the other genotypes were cut into two fragments of 194 and 317 bp（Fig.2-b）.

3.3.DistributionpatternsofDt1genotypeindifferent ecoregions

To determine the distributions of different alleles of Dt1 in soybean varieties，the Indel and CAPS markers were used to genotype 1120 soybean varieties.Among the 1120 varieties tested，68.6%carried the dominant allele GmTfl1，with GmTfl1-a and GmTfl1-b accounting for 30.7%and 37.9%，respectively.The four recessive alleles Gmtfl1-ta，Gmtfl1-ab，Gmtfl1-bb，and Gmtfl1-tb accounted for 12.9%，12.8%，5.7%，and 0.1%，respectively.

In the Northern subregion，the respective proportions of GmTfl1 and Gmtfl1 were 86.1%and 13.9%，（Fig.3），whereas in the Huang-Huai and Southern subregions，they were 40.1%，59.9%and 10.5%，89.5%.Thus，the indeterminate genotype was the most common one in Northern ecoregion，and the proportion of the determinate genotype was slightly higher than that of the indeterminate genotype in the Huang-Huai ecoregion.However，determinate genotypes were most common in the Southern ecoregion.

In the Northern ecoregion，five of the six alleles for GmTfl1，GmTfl1-a，GmTfl1-b，Gmtfl1-ta，Gmtfl1-bb，and Gmtfl1-ab were observed and Gmtfl1-tb was absent.In the Huang-Huai ecoregion，all six alleles were found，but GmTfl1-a，Gmtfl1-ta，Gmtfl1-ab，and Gmtfl1-bb accounted for the largest number. Five alleles could be detected in the Southern ecoregion，with GmTfl1-a，Gmtfl1-ab，and Gmtfl1-bb the most common alleles and Gmtfl1-tb absent.

Further analysis showed that the proportions of the variousallelesdifferednotonlybyecoregionbutby sub-ecoregion.GmTfl1-a was present in relatively high proportion in three sub-ecoregions:Huang-Huai spring（HSp），Huang-Huai summer（HSu），and North spring（NSp）；GmTfl1-b，in contrast，had a relatively high proportion in the Northeastspring（NESp），Northspring，andSouthsummer（SSu）sub-ecoregions.The four recessive alleles were distributed as follows:Gmtfl1-ab was ubiquitous with no particular distribution trend，Gmtfl1-ta was found mostly in the Huang-Huai spring and summer and South spring（SSp）sub-ecoregions，Gmtfl1-bbwasdistributedmainlyacrossthefoursubecoregions in the Huanghai and Southern areas，and Gmtfl1-tb was found only in the Huang-Huai summer sub-ecoregion （Fig.4）.

The numbers of alleles present in the different subecoregions were in the following order:Huang-Huai summer＞Huang-Huai spring and South summer＞Northeast spring，North spring，South spring.The order of the genetic diversity index for each sub-ecoregion was Huang-Huai summer＞Huang-Huaispring＞Northspring＞Southspring＞South summer＞Northeast spring（Table 1）.This result shows that varieties in the Huang-Huai ecoregions had the highest variation at Dt1，followed by that of the North spring sub-ecoregion， and that the lowest diversity of GmTfl1 was found in the Northeast spring sub-ecoregion.

Fig.2-Molecular markers developed from GmTfl1 gene for genotyping Chinese soybean varieties.（A）Diagrammatic representation of polymorphic sites used for development of Indel and CAPS markers.Grey rectangles represent 5′and 3′UTRs. Blue rectangles represent exons and black rectangles represent introns.Arrows mark the positions of insertions/deletions and single-nucleotide polymorphic sites for different alleles of the GmTfl1 gene.（B）DNA markers distinguishing each allele from the other.The Tfl1ta+Ava II marker distinguished Gmtfl1-ta from other alleles；Tfl1b+HpyCh4III distinguished Gmtfl1-ab from other alleles；Tfl1b+Hind III distinguished Gmtfl1-bb from other alleles；Tfl1b+Acc I distinguished Gmtfl1-tb from other alleles；TFL1ab distinguished GmTfl1-a and its derived recessive alleles Gmtfl1-ta from GmTfl1-b and its derived recessive alleles Gmtfl1-ab，Gmtfl1-bb，and Gmtfl1-tb.

3.4.Changes in allele frequency of GmTfl1 during the soybean breeding history of China

Of the 1120 soybean varieties used，940 had records of the year in which they were released.Using this information，the characteristics of varieties released in the last 60 years were analyzed.The varieties were divided into subgroups by 5-year release intervals（Fig.5）.Of varieties released in the Northern eco-region during the 60-year period，77.7—91.5%carried the GmTfl1 allele.After the 1970s，recessive alleles accounted for 50.0—68.8%ofvarieties from theHuang-Huai ecoregion. Althoughthere were no obvious frequency changes of recessive and dominate genotypes in either the Northern or Huang-Huai ecoregions，the relative frequency of the recessive Gmtfl1 increased from 8.5%to 22.3%and 57.1%to 68.8%after 1996 in the two eco-regions.The genetic diversity index ofGmTfl1showedsimilartrendsintheNorthernand Huang-Huai eco-regions，increasing in the former during 1954—1980，decreasing over five years，and then gradually increasingto1.2；andincreasinginthelatterduring 1961—1985，decreasing over 10 years，and then increasing to 1.4（Fig.6）.These trends suggest that the artificial selection of different genotypes of GmTfl1 during the breeding process increased the genetic diversity of growth habit.The Southern varieties were not included in this analysis，owing to their relatively low numbers.

Fig.3-Distributions of six different alleles at the GmTfl1 locus in Northern，Huang-Huai and Southern ecoregions. HR：Huang-Huai ecoregion；NR：Northern ecoregion；SR：Southern ecoregion.

Fig.4-Proportions of six different alleles at GmTfl1 locus in six sub-ecoregions.NEs：Northeast spring sub-ecoregion；NSp：North spring sub-ecoregion；HSp：Huang-Huai spring sub-ecoregion；HSu：Huang-Huai summer sub-ecoregion；SSp：South spring sub-ecoregion；SSu：South summer sub-ecoregion.

3.5.Consistency of stem growth habit and GmTfl1 genotype

According to a consistency analysis of stem growth habit phenotype and genotype in this experiment（Table 2），297 of 337 soybean varieties with determinate stem growth habit had the Gmtfl1（including Gmtfl1-ta，Gmtfl1-ab，Gmtfl1-bb，and Gmtfl1-tb）genotype，for a consistency of 88.1%.Among the 487 soybean varieties with semi-determinate stem growth habit，440 varieties（90.4%）had genotype GmTfl1（including Gmtfl1-a and Gmtfl1-b）.Among the 296 soybean breeding varieties with indeterminate stem growth habit，288 varieties had genotypeGmTfl1（including GmTfl1-a and GmTfl1-b）for a consistency of 97.3%.

Table 1-Genetic diversity index of different GmTfl1 alleles in six sub-ecoregions.

4.Discussion

4.1.Relationship between stem growth habit and ecological environment and geographic distributions of GmTfl1 alleles

Chang et al.［29］found that，in the north of the Northern ecoregion，determinate varieties were very rare，and from north to south the frequency of determinate varieties gradually increased.The stem growth habit in the Huang-Huai ecoregion included both indeterminate and determinate，and in the Southern area，determinate types were predominant. Zhang found that areas with high temperature and humidity are more suitable for determinate varieties，areas with limited rain and water are suitable for indeterminate varieties，and semi-humid regions are most often planted with determinate ones［30］.Sun and Song［31］found that in some northern areas in China with fertile soil，sufficient rainfall and heat energy，determinate varieties are favored under summer and autumn sowing conditions，whereas in the spring soybean region in the northeast and northwest，Inner Mongolia，and part of Shaanxi province that have low temperatures and limited rainfall and fertility，the climate is more suitable for growing indeterminate varieties，given that these varieties have a long period between vegetative and reproductive growth and scattered flowering times，thereby displaying stronger adaptive capacity under imperfect growth conditions.Our results were consistent with these previous results.

In the present study，the distribution of dominant and recessive genotypes of GmTfl1 is consistent with that found in soybean landraces of the minicore collection［17］.Unlike the minicore collection，in which Gmtfl1-bb was not observed in the Northern ecoregion，it accounted for 0.4%of varieties in the Northeast spring sub-ecoregion in our study.Although GmTfl1-b and Gmtfl1-ta were not observed in the Huang-Huai spring sub-ecoregion in the minicore collection［17］，these two have proportions of 4.3%and 26.1%in Huang-Huai spring breeding varieties.Gmtfl1-tb was found in the Northern spring，Huang-Huai summer，and Southern summer sub-ecoregions in the minicore collection［17］.However，in this study，only one varietywith Gmtfl1-tb genotype wasfound inthe Huang-Huai summer sub-ecoregion.The difference in the distribution of these alleles among landraces and varieties shows that this locus was also under selection during the adaptation of soybean to various ecological and cultural conditions.Among the American ancestor varieties，no Gmtfl1-ta and Gmtfl1-tb alleles were found，with both Gmtfl1-bb and GmTfl1-a having been eliminated during modern American soybean breeding［17］.However，all four alleles can be found in Chinese varieties，although Gmtfl1-tb is extremely rare，indicating a broader genetic base of Chinese varieties based on the higher level of genetic diversity in the Chinese soybean germplasm collection.

Fig.5-Genotype frequencies of GmTfl1 and Gmtfl1 over 60 years in soybean varieties released in the Northern and Huang-Huai ecoregions.HR：Huang-Huai ecoregion；NR：Northern ecoregion.

4.2.Molecular markers for different Gmtfl1 alleles and their application

In this experiment，functional molecular markers associated with alleles controlling soybean stem growth habit were developed，and the identification efficiencies for the determinate，semi-determinate，and indeterminate growth habits in 1120 soybean varieties were 88.1%，90.3%，and 97.3%，respectively.The marker-assisted selection efficiency for determinate and semi-determinate varieties was lower than that for indeterminate varieties.The possible reason may be thedifficulty of distinguishing the determinate from the semideterminate phenotype in diverse soybean germplasm ［32］. Classical genetic analyses have shown that soybean growth habit is regulated by both Dt1 and Dt2［8］.Soybean with Dt2/Dt2 and dt2/dt2 genotypes all show determinate phenotype in dt1/ dt1 background，but Dt2/Dt2 and dt2/dt2 in the Dt1/Dt1 backgroundshowsemi-determinateandindeterminatephenotypes，respectively.Although Dt2，a MADS domain factor gene，has recently been cloned，no clearly causative variation of Dt2 was found，making it difficult to evaluate the genotypes of Dt2 in soybean germplasm with molecular markers［8，32］.The genotypic database of stem growth habits established in the present experiment may be useful for future selection of parental lines using molecular markers.

The proportions of the various alleles in this experiment suggest that the individual alleles may have differing advantages within and between the different ecoregions.The ubiquity of GmTfl1 in the Northeast spring and Northern spring sub-ecoregions suggests that this allele is especially advantageous in those ecoregions，but that there is variation among the alleles present，such that the most common allele inNortheastspring varieties is GmTfl1-b and those in Northern spring varieties are GmTfl1-a and GmTfl1-b.The main alleles in Huang-Huai spring varieties are GmTfl1-a and Gmtfl1-ta.In the Southern soybean sub-ecoregions，the main allelesareGmtfl1-bbforSouthernspringvarietiesand Gmtfl1-ab and Gmtfl1-bb for Southern summer varieties.The widespread presence of Gmtfl1-ab across all ecoregions may be due to its relatively high adaptability.Thus，knowledge of stem growth habits in different ecoregions and the distribution proportions of the alleles for gene GmTfl1 could provide theoretical guidance for directional breeding.

Fig.6-Changes in genetic diversity index in varieties from Northern and Huang-Huai ecoregions released during 1954-2012. HR：Huang-Huai ecoregion，NR：Northern ecoregion.

Table 2-Genetic diversity index of six different alleles in three stem habits.

Acknowledgments

This work was supported by the National Natural Science Foundation of China（31271753），the National Key Technology R&D Program of China（2012AA101106），the Agricultural ScienceandTechnologyInnovationProgram（ASTIP），Developmentand Application ofMolecularMarkersin Crops，the Agricultural Science and Technology Innovation Program，the National Key Basic Research Program of China （2009CB118400），the Crop Germplasm Resources Protection （2014NWB030，2015NWB030-05）and the Platform of National Crop Germplasm Resources of China（2014-004，2015-004）.

R E F E R E N C E S

［1］F.J.Gao，S.L.Zhang，T.K.Zhao，Research advance and breeding prospect of pod bearing habit on soybean，Soybean Sci.Technol.4（2010）12—16.

［2］W.B.Wang，Q.Y.He，H.Y.Yang，S.H.Xiang，N.G.Xing，T.J. Zhao，J.Y.Gai，Identification of wild segments associated with stem termination，pod color，and seed coat color in soybean，Acta Agron.Sin.39（2013）1155—1163（in Chinese with English abstract）.

［3］L.G.Heatherly，J.R.Smith，Effect of soybean stem growth habit on height and node number after beginning bloom in the midsouthern USA，Crop Sci.44（2004）1855—1858.

［4］J.E.Specht，K.Chase，M.Macrander，G.L.Graef，J.Chung，J.P. Markwell，M.Germann，J.H.Orf，K.G.Lark，Soybean response to water，Crop Sci.41（2001）493—509.

［5］S.H.Song，R.M.Lin，X.L.Liu，L.Zhang，K.L.Wei，Evaluation of adaptability of soybean cultivars of different podding habits to different ecotypes，Sci.Agric.Sin.32（1999）77—80（in Chinese with English abstract）.

［6］S.H.Song，F(xiàn).T.Xie，Z.M.He，Research on the relationship between soybean podding habit and yield potential，Liaoning Agric.Sci.（6）（2004）24—25（in Chinese）.

［7］P.L.Sun，Z.H.Song，T.Feng，Study on the dynamic dry matter accumulation in the different growth stages of soybean podding propensity，Soybean Sci.Technol.（4）（2009）42—45.

［8］R.L.Bernard，Two genes affecting stem termination in soybeans，Crop Sci.12（1972）235—239.

［9］D.M.Cao，Studies on the growth habit of soybean the main differences of soybean with different pod-setting habit and their identification，Acta Agron.Sin.8（1982）81—86（in Chinese with English abstract）.

［10］Q.C.Zhu，Studies on the growth habit of soybean:I.The substantial difference of different habits of soybean and their classification，Soybean Sci.3（1984）318—326（in Chinese with English abstract）.

［11］S.H.Song，Z.Dong，Comparative study on blooming sequence and podding habit of soybeans，Sci.Agric.Sin.35（2002）1420—1423（in Chinese with English abstract）.

［12］S.H.Liu，J.H.Wang，M.C.Zhang，J.Y.Gai，A study on identification procedure for soybean stem growth habit type，Soybean Sci.24（2005）81—89（in Chinese with English abstract）.

［13］J.L.Wang，Soybean，Popular Science Press，Beijing，1966. 140—144（in Chinese）.

［14］C.M.Woodworth，Genetics and breeding in the improvement of the soybean，University of Illinois Agricultural Experiment Station Bulletin 3841932.297—404.

［15］L.F.Williams，Structure and genetic characteristics of the soybean，in:K.S.Markley（Ed.）Soybeans and Soybean Products，vol.1，Interscience Publishers，New York 1950，pp.111—134.

［16］B.Liu，S.Watanabe，T.Uchiyama，F(xiàn).Kong，A.Kanazawa，Z. Xia，A.Nagamatsu，M.Arai，T.Yamada，K.Kitarmura，C. Hyuya K.Masuta，J.Abe，The soybean stem growth habit gene Dt1 is an ortholog of Arabidopsis TERMINAL FLOWER1，Plant Physiol.153（2010）198—210.

［17］Z.Tian，X.Wang，R.Lee，Y.Li，J.E.Specht，R.L.Nelson，P.E. McClean，L.Qiu，J.Ma，Artificial selection for determinate growth habit in soybean，Proc.Natl.Acad.Sci.U.S.A.107 （2010）8563—8568.

［18］J.H.Yang，S.W.Wang，X.Y.Liu，J.F.Yang，M.F.Zhang，Development and application of functional markers in higher plants，Sci.Agric.Sin.41（2008）3429—3436（in Chinese with English abstract）.

［19］I.Eujayl，M.E.Sorrells，M.Baum，P.Wolters，W.Powell，Isolation of EST-derived microsatellite markers for genotyping the A and B genomes of wheat，Theor.Appl. Genet.104（2002）399—407.

［20］T.Thiel，W.Michalek，R.Varshney，A.Graner，Exploiting EST databases for the development and characterization of gene-derived SSR-markers in barley（Hordeum vulgare L.），Theor.Appl.Genet.106（2003）411—422.

［21］B.Hackauf，P.Wehling，Identification of microsatellite polymorphisms in an expressed portion of the rye genome，Plant Breed.121（2002）17—25.

［22］T.F.Zang，M.Ge，Y.C.Wei，H.Zhao，Discovery for maize function insertion/deletion（Indel）polymorphic marker and its implication in purity identification of maize hybrid seeds，J.Maize Sci.20（2012）64—68（in Chinese with English abstract）.

［23］F.Wang，J.H.Li，W.G.Liu，Y.L.Liao，M.S.Zhu，Z.R.Liu，H.J. Huang，D.J.Huang，Development of a functional marker for fragrance gene in rice，Chinese J.Rice Sci.22（2008）347—352（in Chinese with English abstract）.

［24］M.Y.Kim，K.Van，P.Lestari，J.K.Moon，S.H.Lee，SNP identification and SNAP marker development for a GmNARK gene controlling supernodulation in soybean，Theor.Appl. Genet.110（2005）1003—1010.

［25］M.U.Rayhan，K.Van，D.H.Kim，S.I.Kim，M.Y.Kim，Y.H.Lee，S.H.Lee，Identification of Gy4 nulls and development of multiplex PCR-based co-dominant marker for Gy4 and α' subunit of β-conglycinin in soybean，Genes Genomics 33 （2011）383—390.

［26］R.Juwattanasomran，P.Somta，A.Kaga，S.Chankaew，T. Shimizu，W.Sorajjapinun，P.Srinive，Identification of a new fragrance allele in soybean and development of its functional marker，Mol.Breed.29（2012）13—21.

［27］L.Qiu，R.Chang，G.Singh，The origin and history of soybean，in:G.Singh（Ed.），The Soybean:Botany，Production and Uses，CABI，Oxfordshire 2010，pp.1—23.

［28］K.P.Robert，The measurement of species diversity，Annu. Rev.Ecol.Syst.5（1974）285—307.

［29］R.Z.Chang，Analysis of Chinese soybean genetic resources: III.Distribution characteristics of soybean reproductive and podding habits，Crop Variety Resour（2）（1990）1—2（in Chinese）.

［30］J.Z.Zhang，Breeding and Cultivation of Soybean in China，China Agriculture Press，Beijing，1993.pp 164—165，（in Chinese）.

［31］P.L.Sun，Z.H.Song，Study on growth characters of different podding propensity soybean varieties，Soybean Sci.Technol. （5）（2008）17—20.

［32］J.Ping，Y.Liu，L.Sun，M.Zhao，Y.Li，M.She，Y.Sui，F(xiàn).Lin，X. Liu，Z.Tang，H.Nguyen，Z.Tian，L.Qiu，R.Nelson，T. Clemente，J.Specht，J.Ma，Dt2 is a gain-of-function MADS-domain factor gene that specifies semideterminacy in soybean，Plant Cell 26（2014）2831—2842.

*Corresponding author.Tel.:+86 10 82105843；fax:+86 10 82105840.

**Correspondence to：L.Qiu，Soybean Research Institute，Northeast Agricultural University，Harbin 150030，China.Tel.:+86 10 82105843；fax:+86 10 82105840.

E-mail addresses:guanrongxia@caas.cn（R.Guan），qiulijuan@caas.cn（L.Qiu）.

Peer review under responsibility of Crop Science Society of China and Institute of Crop Science，CAAS.1These authors contributed equally to this work.

http://dx.doi.org/10.1016/j.cj.2015.05.004

2214-5141/?2015 Crop Science Society of China and Institute of Crop Science，CAAS.Production and hosting by Elsevier B.V.This is an open access article under the CC BY-NC-ND license（http://creativecommons.org/licenses/by-nc-nd/4.0/）.

8 January 2015