Developing gene-tagged molecular markers for functional analysis of starch-synthesizing genes in rice ( Oryza sativa L.)

Euphytica 135: 345–353, 2004.© 2004 Kluwer Academic Publishers. Printed in the Netherlands.

345

Developing gene-tagged molecular markers for functional analysis ofstarch-synthesizing genes in rice (Oryza sativa L.)

Xinyan Liu1, Minghong Gu1, Yuepeng Han1, Qing Ji2, Jufei Lu1, Shiliang Gu1, Rong Zhang1,Xin Li1, Jianmin Chen2, Schuyler S. Korban3 & Mingliang Xu2,∗1Agriculture College, Yangzhou University, 12 East Wenhui Road, Yangzhou, Jiangsu 225009, P. R. China;2College of Bioscience and Biotechnology, Yangzhou University, 12 East Wenhui Road, Yangzhou, Jiangsu225009, P. R. China; 3Department of Natural Resources and Environmental Sciences, University of Illinois, 310Madigan Building, 1201 West Gregory Drive, Urbana, IL 61801, U.S.A.; (∗Author for correspondence; e-mail:[email protected])

Received 4 September 2003; accepted 20 November 2003

Key words: amylose content, gene-tagged marker, granule-bound starch synthase, rice, starch branching enzyme

Summary

The granule-bound starch synthase (GBSS), starch branching enzymes 1 (SBE1) and 3 (SBE3) are major enzymesinvolved in starch biosynthesis in rice endosperm. Available sequences of Sbe1 and Sbe3 genes encoding corres-ponding SBE1 and SBE3 have been used to identify homologous regions from genome databases of both the indicarice 93-11 and the japonica rice Nipponbare. Sequence diversities were exploited to develop gene-tagged markersto distinguish an indica allele from a japonica allele for both Sbe1 and Sbe3 loci. With these newly developedgene-tagged markers and available Wx gene markers, the roles of these starch-synthesizing genes (Sbe1, Sbe3,and Wx) in determining amylose content (AC) in the rice endosperm were evaluated using a double-haploid (DH)population derived from a cross between the indica rice cv. Nanjing11 and the japonica rice cv. Balilla. Only theWx and Sbe3 loci had significant effects on the AC variation. On average, indica Wxa genotypes showed ∼9.1%higher AC than japonica Wxb genotypes, while indica Sbe3a genotypes showed ∼1.0% lower AC than japonicaSbe3b genotypes. A significant interaction was also observed between Wx and Sbe3 loci whereby the amylosecontent was 0.3% higher in Sbe3a than Sbe3b genotypes in the presence of the Wxa allele, but it was lower by2.3% in the presence of the Wxb allele. Overall, polymorphisms at the Wx and Sbe3 loci together could accountfor 79.1% of the observed AC variation in the DH population. The use of gene-tagged markers in marker-assistedselection and gene functional analysis was also discussed.Abbreviations: AC – amylose content; GBSS – granule bound starch synthase; SBE1 – starch branching enzyme 1;SBE3 – starch branching enzyme 3; Waxy – granule bound starch synthase locus; Sbe1 – starch branching enzyme1 gene; Sbe3 – starch branching enzyme 3 gene

Introduction

Rice (Oryza sativa L.) is one of the most importantcrops in the world that provides staple food for almosthalf of the world’s population. However, consumer’spreference for cooking, processing, and eating proper-ties of rice varies around the world. Hence, it is criticalthat breeders develop rice cultivars with diverse grainqualities to meet the consumer’s demand.

Starch comprises ∼90% of milled rice, and it iscomposed of linear amylose and branching amylo-pectin (Blakeney, 1984; Smith et al., 1997). Thegranule-bound starch synthase (GBSS) encoded bythe Waxy (Wx) gene controls the synthesis of amyl-ose (Smith et al., 1997). Starch branching enzymes(SBEs) are involved in determining amylopectin syn-thesis by introducing α-1,6-glucosidic linkages intoα-polyglucans (Martin & Smith, 1995; Okagaki &

346

Wessler, 1988). Amylose content has been recognizedas one of the pivotal determinants of eating and cook-ing qualities of rice. Low amylose content is usuallyassociated with tender, cohesive, and glossy cookedrice; while high amylose content results in dry, fluffy,and granular cooked rice (Juliano, 1971). The apparentamylose level is also an important factor for the marketvalue of rice (Larkin et al., 2003). Although amyl-ose content is a key-determining factor of rice quality,there are minor differences in the texture of cookedrice among cultivars with similar amylose contents(Juliano, 1985). It has been verified that both structureand characteristics of amylopectin are also determin-ants of rice texture (Takeda & Hizukuri, 1987; Ong &Blanshard, 1995).

The Wx locus encoding GBSS, or waxy protein,harbors two alleles, Wxa and Wxb, in non-waxy ricecultivars (Sano et al., 1986). The Wxa allele contrib-utes to a higher level of waxy protein than the Wxb

allele, and thus results in high amylose content in thegrain (Mikami et al., 2000). TheWxa allele is predom-inant in the indica subspecies, while the Wxb allele ispredominant in the japonica subspecies (Sano et al.,1986). Wang et al. (1995) have reported that the mo-lecular basis for the observed differences in amylosecontent in the rice endosperm is attributed to post-transcriptional regulation. The Wxa allele, containingthe AGGTATA sequence in the putative 5’leader intronsplice-junction, shows a high efficiency in excisingthe leader intron. This results in high accumulation ofmature wx transcripts, and thus high amylose content.Whereas the Wxb allele, containing the AGTTATA se-quence, exhibits incomplete splicing, leading to lowlevels of mature wx transcripts, and those low amylosecontent. These observations have been later confirmedby various studies (Bligh et al., 1998; Cai et al.,1998; Isshiki et al., 1998). The G/T polymorphismis one factor involved in AC variation, and can ac-count for 79.7% of the observed AC variation in 89non-glutinous rice cultivars tested (Ayres et al., 1997).In addition, polymorphism of microsatellites or (CT)nrepeats in the Wx leader sequence is yet another majorfactor in determining AC variation, and can accountfor >82% of the AC variation observed in non-waxyrice cultivars (Bligh et al., 1995; Ayres et al., 1997;Tan & Zhang, 2001).

Multiple isoforms of starch branching enzymeshave been detected in higher plants. These can begrouped into two distinct families, A and B, basedon primary structure and their functions in starch syn-thesis (Burton et al., 1995; Martin & Smith, 1995).

Three isoforms of rice starch branching enzymes, in-cluding SBE1 (B family), SBE3 (A family), and SBE4(A family), are present in filling grains (Mizuno etal., 1992; Mizuno et al., 1993). Each isoform may beresponsible for a unique starch structure in the amyl-opectin biosynthesis pathway (Mizuno et al., 2001).The two major starch branching enzymes SBE1 andSBE3 account for ∼70% and ∼30%, respectively,of the total amylopectin synthesized in the rice en-dosperm with SBE3 transferring shorter chains thanSBE1 (Yamanouchi & Nakamura, 1992; Mizuno et al.,1993). Moreover, cDNA sequences for SBE1, SBE3,and SBE4 have been published and characterized(Mizuno et al., 1992, 1993, and 2001).

The completed genome sequencings of both theindica rice line 93-11 and the japonica rice cv. Nip-ponbare provides unique opportunities for isolationand functional analysis of economically importantgenes (Sasaki, 2002). Apart from all rice markers cur-rently available, high numbers of molecular markers,especially gene-tagged markers, can be developed byexploiting sequence diversities between these two ricesubspecies. These markers will be particular usefulin map-based gene cloning, functional analysis, andmarker-assisted selection.

In the present study, the development of gene-tagged markers for the Sbe1 and Sbe3 loci weredemonstrated. Moreover, these gene-tagged mark-ers were utilized to analyze the effects of starch-synthesizing genes on amylose content in the riceendosperm.

Materials and methods

Plant material

A total of 16 rice cultivars, including eight indicaand eight japonica rice subspecies, were used to de-velop gene-tagged molecular markers (Table 1). Adouble-haploid (DH) population consisting of 127lines, derived from a cross between the indica ricecv. Nanjing11 and the japonica rice cv. Balilla, wasused to investigate the effects of starch-synthesizingenzymes on amylose content.

All rice cultivars/lines were grown at the Exper-imental Farm of Yangzhou University during 2002.Leaves of young seedlings were harvested, and groundinto a fine powder in liquid nitrogen using a mortarand pestle. DNA extraction was conducted accordingto the protocol of Liu et al. (1997).

347

Table 1. A listing of all rice cultivars used todevelop gene-tagged markers

Rice cultivars Origin Subspecies

Balilla Italy japonica

Wuxiangjing China japonica

9516 China japonica

Nanjing11 China indica

Nongken57 China japonica

9311 China indica

Zhonghua11 China japonica

3037 China indica

Lemont USA indica

Guichao2 China indica

Suyu Nuo China japonica

Qiu guang Japan japonica

Diantuan502 China indica

909 China indica

Wuxiangxian China indica

02428 China japonica

Sequence diversities in Sbe1 and Sbe3 between tworice subspecies

The cDNA or gene sequences of both Sbe1 andSbe3 genes were obtained from the following website:http: //www.ncbi.nlm.nih.gov/entrez/query.fcgi. These sequences were derived from the japon-ica rice Nipponbare. A BLAST search was conductedto identify corresponding structural gene sequencesfrom genome databases of the indica rice 93-11(http: //www.btn.genomic.org.cn/rice) and thejaponica rice Nipponbare (http: //www.gramene.org/gramene/searches/blast). For both Sbe1 andSbe3 loci, sequence diversities, ranging from pro-moter to 3’-untranslated regions (UTRs), betweenline 93-11 and cv. Nipponbare were identified.The indels (inserts and deletions) were amenablefor developing sequence-tagged site (STS) markers,while those single nucleotide polymorphisms (SNPs)present within some restriction recognition sites wereuseful for developing cleaved-amplified polymorphicsequence (CAPS) markers.

Primer design

Primers for candidate STS or CAPS markers were de-signed based on either the flanking regions of indelsor SNPs of both Sbe1 and Sbe3 using the PRIMER0.5software (http: //www-genome.wi.mit.edu/ftp/

pub/software/primer.0.5). The criteria set up fordesigning primers were as follows: 1) length of 20–25bp; 2) GC content of 50∼60%; 3) Tm of 60–75 ◦C; 4)No secondary structure and no run of three identicalnucleotides; and 5) preference for either G or C at the3’-end of the primer.

Development of gene-tagged STS and CAPS markers

Each pair of primers was first assayed using PCRto determine whether or not an expected and unam-biguous PCR band was present. Then, PCR bands ofdifferent sizes must be present for those STS mark-ers useful for distinguishing indica from japonica ricecultivars, while polymorphic bands must be present indigested PCR products for those CAPS markers.

The PCR reaction mixture was adjusted to avolume of 10 µl containing 30 ng of template DNA,10 mM Tris-HCL, 50 mM KCL, 2.5 mM MgCL2,5 mM dNTP, 50 pmol primer, and 1 U Taq polymerase.The reaction was performed as follows: denaturationat 94 ◦C for 2 min, followed by 35 cycles of 94 ◦C for45s, 60 ◦C for 60s, 72 ◦C for 2 min, and with a finalextension at 72 ◦C for 10 min. For CAPS markers, thePCR product was then digested with a correspondingrestriction enzyme. The reaction was carried out in 20µl, consisting of 2 µl of 10×buffer, 5 U restrictionendonuclease, and 5 µl PCR product, at 37 ◦C for 2 h.

PCR products (STS markers) or digested PCRfragments (CAPS markers) were subjected to elec-trophoresis on either 1% agarose gel or 6% polyac-rylamide gel, depending on sizes and differences ofthe polymorphic bands. If the difference between anytwo polymorphic bands was large; e.g. > 50 bp, then1% agarose gel was used; whereas, if the differencewas small; e.g. < 10 bp, then 6% polyacrylam-ide gel was used instead. Following electrophoresis,the agarose gel was stained with ethidium bromide,while the polyacrylamide gel was silver-stained forvisualization.

Data collection

The gene-tagged STS or CAPS markers for both theSbe1 and Sbe3 genes as well as the marker 484/W2Rfor the Wx gene (Ayres et al., 1997) were used to in-vestigate a total of 127 lines in the DH population. Thegenotype for each line, at the Sbe1, Sbe3, and Wx loci,was deduced by analyzing the performance of all fivemarkers tested. Meanwhile, the apparent amylose con-tent for each line was determined using the proceduredescribed by Tan et al. (1999).

348

Figure 1. Developing gene-tagged STS and CAPS markers for both Sbe1 and Sbe3 genes using 16 rice cultivars. Panel A: Sbe1 5’-end STSmarkers, polymorphic bands were observed with a band of 882 bp in size for all japonica varieties and the other band of 548 bp for allindica cultivars. Panel B: Sbe3 5’-end STS markers, PCR products of 603bp for all japonica cultivarss and 594 bp for all indica cultivars wereobtained. Panel C: Sbe1 3’-end CAPS marker, PCR products were subjected to AccII digestion and then separation on agarose gel. Two digestedfragments (182 bp and 274 bp) were observed in all japonica cultivars; whereas, only a single undigested band (456bp) was present in indicacultivars. Panel D: Sbe3 3’-end CAPS marker, all japonica cultivars gave rise to two digested bands, 215 bp and a 295 bp, whereas a singleundigested band of 510 bp in size was present in eight indica cultivars. Lanes 1 – Balilla (japonica), 2 – Wuxiangjing (japonica), 3 – 9516(japonica), 4 – Nanjing11 (indica), 5 – Nongken57 (japonica), 6 – 9311 (indica), 7 – Zhonghua11 (japonica), 8 – 3037 (indica), 9 – Lemont(indica), 10 – Guichao2 (indica), 11 – Suyu Nuo (japonica), 12 – Qiu guang (japonica), 13 – Diantuan502 (indica), 14 – 909 (indica), 15 –Wuxiangxian (indica), and 16 – 02428 (japonica). ∼ M (panels A, C, and D), 1 kb DNA ladder; M (panel B), pBR322/HaeIII DNA marker.

Statistical analysis

Contributions of the three starch-synthesizing genes,Sbe1, Sbe3, and Wx, to AC variations in the DH pop-ulation were estimated using a multiple linear regres-sion analysis in SAS statistical package (SAS instituteInc., 2000).

Results

Diversities of Sbe1 and Sbe3 genes in the two ricesubspecies

The Sbe1 sequence (accession no. D10838) was ob-tained from Genbank. Its corresponding structuralgene sequences from both ‘Nipponbare’ (BAC cloneAP0046) and ‘93-11’ (contig 677) databases wereidentified using the Blast search. Likewise, the Sbe3cDNA sequence (from ‘Nipponbare’, accession no.D16201) was obtained from Genbank, and used toidentify its corresponding structural gene sequence.As a result, the two contigs 6831 and 8157 wererecovered from the ‘93-11’ genome databases.

The size of the Sbe1 coding sequence has been es-timated at 2460 bp, and consisting of 14 exons. The

full-length Sbe1 structural gene in line ‘93-11’ (7265bp) is different from that in ‘Nipponbare’ (7242 bp).Only two polymorphic nucleotides, located in exons6 and 14, were observed in the Sbe1 coding regionsbetween ‘93-11’ and ‘Nipponbare’; however, these donot alter the corresponding amino acids. A transposonof 334 bp in length was present at 522 bp upstream ofthe putative start codon in the ‘Nipponbare’, but not inthe ‘93-11’.

The size of the Sbe3 cDNA was 2475 bp, and con-sisting of 22 exons. The full-length Sbe3 structuralgene was estimated at 10,944 bp in the ‘93-11’. Ninepolymorphic nucleotides were present in the Sbe3coding region between ‘93-11’ and ‘Nipponbare’, res-ulting in five polymorphic amino acids. Moreover, anumber of indels were observed in the promoter regionof Sbe3 between ‘93-11’ and ‘Nipponbare’.

Development of gene-tagged STS and CAPS markers

Sequence alignment of ‘93-11’ and ‘Nipponbare’ re-vealed diversities in both Sbe1 and Sbe3 regions. Todifferentiate and characterize a given gene, both up-stream and downstream gene-tagged markers must beused; otherwise occasional crossovers within the genemay lead to ambiguous results. Accordingly, sequence

349

Table 2. Sequences and sizes of gene-tagged STS and CAPS markers for both Sbe1 and Sbe3 genes

Genes Type of Primer sequences Sizesa

marker

Sbe1 3’-end CAPS Forward: CCGAGGGAATGCCAGGAGTACCAG 456bp (indica)

Reverse: GAACCACAACCAAGTCCAAGGCAA 182bp, 274bp (japonica)

5’-end STS Forward: GAGTTGAGTTGCGTCAGATC 548bp (indica)

Reverse: CAGCAGCAAGCAACCTCATT 882bp (japonica)

Sbe3 3’-end CAPS Forward: GTCTTGGACTCAGATGCTGGACTC 510bp (indica)

Reverse: ATGTATAACTGGCAGTTCGAACGG 215bp, 295bp (japonica)

5’-end STS Forward: TCGGTCAATTCGGTTAGTCTCCTC 594bp (indica)

Reverse: ACATCCTCTAGCATACTGGCGACTC 603bp (japonica)

a Sizes of STS markers directly correspond to sizes of PCR products; whereas, sizes of CAPS markers correspondto sizes of digested PCR products.

diversities in the promoter and 3’-end or -UTR regionsare used to develop gene-tagged markers.

In the promoter region, two pairs of primers weredesigned based on sequences flanking the transposonat the Sbe1 locus, while several pairs of primers weredesigned flanking indels at the Sbe3 locus. A seriesof PCR assays was carried out using a number ofdifferent indica and japonica cultivars to search forthe best primers for developing gene-tagged markers.The upstream STS marker developed for Sbe1 yiel-ded two unambiguous and polymorphic bands, oneband of 882 bp in size for all japonica cultivars andthe other band of 548 bp for all indica cultivars, andthese bands were readily separated on the agarose gel(Figure 1A, Table 2). The upstream STS marker de-veloped for Sbe3 yielded PCR products of 603 bp and594 bp in size for all japonica and all indica cultivars,respectively, and these were readily resolved on 6%polyacrylamide gel (Figure 1B, Table 2).

For 3’-end or -UTR regions, a T/C SNP at the lastexon (exon 14) of Sbe1 was detected between ‘93-11’ and ‘Nipponbare’. The cytosine nucleotide waspresent within the AccII recognition site (CGCG) inthe ‘Nipponbare’, but not in the ‘93-11’ (CGTG).A pair of primers, flanking the T/C SNP site, wasdesigned, and used to amplify japonica and indicacultivars. PCR products were subjected to AccII di-gestion, and then separated on an agarose gel. Twodigested fragments, 182 bp and 274 bp, were observedin all japonica cultivars; whereas, a single undigestedband, 456 bp, was present in all indica cultivars (Fig-ure 1C, Table 2). Likewise, a C/G SNP, located in the3’-UTR of Sbe3, was linked to the SpeI recognitionsite (ACTAGT) in the ‘Nipponbare’, but not in the‘93-11’ (ACTACT). A pair of primers was designed to

amplify all japonica and indica cultivars analyzed, andPCR products were subjected to SpeI digestion. Alljaponica cultivars gave rise to two digested bands, 215bp and 295 bp, whereas, a single undigested band, 510bp, was present in eight indica cultivars (Figure 1D,Table 2). Thus, two CAPS markers were developedfrom the last exon of Sbe1 and the 3’-UTR region ofSbe3. All gene-tagged STS and CAPS markers de-veloped in this study can be readily used to distinguishan indica allele from a japonica allele for both Sbe1and Sbe3 loci.

Impacts of starch-synthesizing genes on amylasecontent variation in the DH population

The Wx, Sbe1, and Sbe3 genes play key roles in de-termining amylose content in the rice endosperm. ADH rice population is ideal for studying the func-tions of these starch-synthesizing genes by minimizingthe interference of the genetic background. The fournewly developed gene-tagged markers for Sbe1 andSbe3 as well as an available CAPS marker for the Wxgene were used to analyze all lines of the DH popu-lation (Figure 2). All five markers were polymorphicbetween the two parental cultivars, ‘Nanjing11’ and‘Balilla’, and therefore were used to differentiatethe indica allele, namely Wxa, Sbe1a, and Sbe3a

(present in the ‘Nanjing11’), from the japonica al-lele, namely Wxb, Sbe1b, and Sbe3b (present in the‘Balilla’) (Figure 2). The genotype for each line inthe DH population at these three loci (Wx, Sbe1, andSbe3) could be inferred based on the performanceof each gene-tagged marker. Eight genotypes werepresent in the DH population including the follow-ing: an indica genotype, Wxa/Sbe1a/Sbe3a; a japon-

350

Figure 2. Gene-tagged STS and CAPS markers were used to detect Sbe1 and Sbe3 alleles in 30 randomly selected lines from the DH population.Panel A: Sbe1 gene 5’- end STS marker; Panel B: Sbe3 gene 5’- end STS marker; Panel C: Sbe1 gene 3’- end CAPS marker; Panel D: Sbe3gene 3’- end CAPS marker. Lanes 1–30 correspond to 30 randomly selected lines from the DH population. M (panels A, C, and D): DL2000DNA marker; M (panel B): pBR322/HaeIII DNA marker.

ica genotype, Wxb/Sbe1b/Sbe3b; and six recombin-ant genotypes, Wxa/Sbe1a/Sbe3b, Wxa/Sbe1b/Sbe3a,Wxa/Sbe1b/Sbe3b, Wxb/Sbe1a/Sbe3a, Wxb/Sbe1a/Sbe3b,and Wxb/Sbe1b/Sbe3a (Table 3). Of all eight gen-otypes, the indica genotype (Wxa/Sbe1a/Sbe3a) wasmost frequently observed, in 41 out of 127 lines,and this was followed by one of the recombin-ant genotypes, Wxb/Sbe1a/Sbe3a, observed in 22lines. The least frequent was the japonica genotype(Wxb/Sbe1b/Sbe3b), which was observed in only fivelines. These results confirmed a commonly observedphenomenon whereby the phenotypic performance ofa DH population derived from a cross between in-dica and japonica parents was usually biased towardsindica genotype.

The amylose content for each of the 127 lines wasmeasured, and large variations in AC were observedamong the eight different genotypes (Table 3). Basedon the multiple linear regression analysis, the roles ofthe three starch-synthesizing genes in determining ACin the rice endosperm were delineated. The Wx geneexerted the highest effect on the observed variation inAC in the DH population, and this was followed by theSbe3 gene, whereas, the Sbe1 gene had no significanteffect on the AC variation. On average, indica Wxa

genotypes in the DH population had ∼9.1% higher AC

Table 3. Mean values of amylose contents ofeight genotypes at the Wx, Sbe1, and Sbe3loci in the DH population containing 127lines

Genotypes Number AC%

Wxa/Sbe3a /Sbe1a 41 20.6A

Wxa/Sbe3a /Sbe1b 7 19.2A

Wxa/Sbe3b /Sbe1a 16 19.3A

Wxa/Sbe3b /Sbe1b 8 21.2A

Wxb/Sbe3a /Sbe1a 22 2.0A

Wxb/Sbe3a /Sbe1b 16 10.4BC

Wxb/Sbe3b /Sbe1a 12 12.6B

Wxb/Sbe3b /Sbe1b 5 11.3B

‘a’ Indicates an indica allele, while ‘b’ cor-responds to a japonica allele.Mean amylose content values with a differ-ent letter are significantly different (p <0.01)using Fisher’s Protected LSD.

than japonica Wxb genotypes (Table 4). This was con-sistent with previous observations that the Wxa allelewas more efficiency than the Wxb allele in producingwaxy protein, thus leading to high AC (Wang et al.1995; Bligh et al. 1998; Cai et al. 1998; Isshiki et al.1998). The indica Sbe3a genotypes had 1.0% lowerAC than japonica Sbe3b genotypes (Table 4), indic-

351

Table 4. Major effects of both Wx and Sbe loci, along with theirinteraction, on amylose content in rice

Sbe loci

Wx loci Sbe3a Sbe3b Sbe3a – Sbe3b Average

Wxa 20.3 20.0 0.3 –1.0

Wxb 9.9 12.2 –2.3

Wxa – Wxb 10.4 7.8

Average 9.1

‘a’ Indicates an indica allele, while ‘b’ indicates a japonica allele.

ating that the indica Sbe3a allele was more powerfulthan the japonica Sbe3b allele in starch branchingactivity. A significant interaction between the Wx andSbe3 loci was also observed. The amylose content inSbe3a genotypes exceeded Sbe3b genotypes by 0.3%in the presence of the Wxa allele, but lagged behind by2.3% in the presence of the Wxb allele (Table 4). Over-all, polymorphisms at both Wx and Sbe3 loci couldaccount for 79.1% (adjusted R2) of the AC variationobserved among all 127 lines in the DH population.

Discussion

Up till now, as many as 5000 DNA markers havebeen mapped, with an average of one marker for every90 kb, within the rice genome (Sasaki, 2002). Thesemarkers are valuable for mapping target regions forgene isolation and marker-assisted selection. How-ever, they are frequently insufficient to saturate thetarget region or to tag a gene due to their unbal-anced distribution within the genome. The completionof the genome sequencings of both indica and ja-ponica rice cultivars allows for generation of largenumbers of molecular markers based on sequence di-versities between two rice subspecies. These markerscan be used to saturate target regions for map-basedgene cloning, conduct functional analysis of genesof interest, or marker-assisted selection. For this pur-pose, establishing a segregant population, derivedfrom a cross between indica and japonica cultivars,is deemed necessary in obtaining high frequenciesof polymorphisms within DNA sequences (Sasaki,2002). In the present study, STS and CAPS markerslocated at both upstream and downstream of Sbe geneshave been developed based on sequence diversities oftwo rice subspecies. In fact, such gene-tagged markerscan be developed for almost all genes, as sequencedivergence suited for molecular marker development

is present along either upstream or downstream se-quences of the target gene. These gene-tagged markersare useful for tracking the origin of a target gene, andthen to analyzing its function. Moreover, these mark-ers are of particular importance in marker-assistedselection due to the following: 1) flanking regions canbe narrowed down to either reduce or eliminate thelinkage drag, and 2) escapes can be excluded due toco-segregation between markers and the target locus.

Pervious investigations of the genetic basis of theapparent amylose content in rice endosperm focusedmainly on the Wx gene encoding the granule-boundstarch synthase (Smith et al., 1997). Polymorphismsat the 5’ splice-junction site of the putative first in-tron and the presence of (CT)n repeats in the leadersequence of the Wx locus accounted for most of theobserved AC variation (Bligh et al., 1995; Ayres etal., 1997; Tan & Zhang, 2001). The starch branchenzymes SBE1 and SBE3 encoded by Sbe1 and Sbe3genes account for ∼70% and ∼30% of the total amyl-opectin in the rice endosperm, respectively, with SBE3transferring shorter chains than SBE1 (Yamanouchi &Nakamura, 1992; Mizuno et al., 1993; Guan & Preiss,1993). Our results confirmed that the Wx gene was amajor contributing factor to the observed AC variationin the DH population. In addition, the Sbe3 gene alsoplayed an important role in AC variation. Unexpec-tedly, the Sbe1 gene seemed to have no significanteffect on AC variation. Polymorphisms at the Wx andSbe3 loci together could account for 79.1% of the ACvariation in the DH population.

Comparative analysis of coding sequences of Sbe1gene between indica and japonica cultivars indicatedthat there was no amino acid difference within theSBE1 protein. This lack of difference in amino acidsequence and activity of SBE1 suggested that inser-tion of the transposon in the indica Sbe1 promoter hadno significant effect on gene expression. As for SBE3,the observed differences in five amino acids betweenindica and japonica subspecies might account for thedifferences in its activity, and contributing to the sig-nificant AC variation in the DH population. This ofcourse presents an interesting evolutionary questionconcerning the issue of whether or not the Sbe1 geneis more conservative than the Sbe3 gene. As it is com-monly known that SBE3 transfers shorter chains thanSBE1, it could be concluded that these short branching(A) chains in starch are expected to be more variablethan longer branching (B) chains.

Generally, the indica Sbe3a allele was more power-ful than the japonica Sbe3b allele in starch branching

352

activity, as Sbe3a genotype show ∼1.0% lower amyl-ose content than Sbe3b genotype, whereas, the differ-ences between the Sbe3a and Sbe3b genotypes werelikely modified depending on the presence of the Wxa

allele. The amylose content in Sbe3a genotypes ex-ceeded Sbe3b genotypes by 0.3% in the presence of theWxa allele, but lagged behind by 2.3% in the presenceof the Wxb allele. As a result, the amylose contents forthe four genotypesWxa/Sbe3a, Wxa/Sbe3b, Wxb/Sbe3a,and Wxb/Sbe3b were 20.3%, 20.0%, 9.9%, and 12.2%in the DH population, respectively (Table 4). Since theGBSS dosage in the Wxb genotype was much lowerthan that in the Wxa genotype, therefore it was sug-gested that the SBE3 activity could only function fullyunder either low or no GBSS enzyme levels. This con-clusion could be further confirmed by another study(published elsewhere) using 40 waxy rice cultivarswhereby both Sbe1 and Sbe3 loci played key rolesin determining starch viscosity characteristics undercondition of no GBSS activity, and polymorphisms atthese two loci could account for ∼70% of the observedvariations in both hot and cool viscosities.

Wide variations in amylose content were observedin the DH population. The AC in the highest line wasestimated at 25.7%, slightly higher than the indica par-ent ‘Nanjing 11’ (24.1), whereas, that of the lowestline was only 6.7%, much lower than the japonicaparent ‘Balilla’ (16.6%). The estimated AC of eachline in the DH population was mainly determined byits genotype. In recent years, major advances havebeen made in our understanding of the roles of en-zymes involved in the cereal starch synthesis (Jameset al., 2003). Apart from GBSS, three major groupsof enzymes along with their isoforms are known tofunction in starch synthesis in the cereal endosperm,including the starch synthase (isoforms, SSI, SSIIa,SSIIb, and SSIII), starch branching enzymes (SBE1,SBE3, and SBE4), and starch de-branching enzymes(isoamylase and pullulanase). With the completion ofthe rice genome-sequencing project, genome-basedapproaches can be used to investigate the influenceof each starch-synthesizing gene on all properties ofrice grains. This can be achieved by utilizing up-stream and downstream gene-tagged markers for eachstarch-synthesizing gene as demonstrated in this study.Using a large segregant population and gene-taggedmarkers, the subtle role(s) of each of the starch-synthesizing genes can then be determined. Moreover,in a marker-assisted selection program, gene-taggedmarkers will be very useful in identifying genotypes

with the best combination of these starch-synthesizinggenes in order to improve rice grain qualities.

Acknowledgements

This study was financially supported by the NationalHigh-Tech ‘863’ Program of P. R. China, grant No.2002AA224041 and the Natural Science Foundation,Jiangsu, P. R. China, grant No. BK2001212.

References

Ayres, N.M., A.M. McClung, P.D. Larkin, H.F.J. Bligh, C.A. Jones& W.D. Park, 1997. Microsatellites and a single-nucleotide poly-morphism differentiate apparent amylose classes in an extendedpedigree of US rice germplasm. Theor Appl Genet 94: 773–781.

Blakeney, A., 1984. Rice grain quality. In: A. Currey (Ed.), RiceGrowing in New South Wales, pp. 1–5. Department of Agricul-ture New South Wales and the Rice Research Committee, Yanco,Australia.

Bligh, H.F.J., P.D. Larkin, P.S. Roach, C.A. Jones, Y.H. Fu & W.D.Park, 1998. Use of alternate splice sites in granule-bound starchsynthase mRNA from low-amylose rice varieties. Plant Mol Biol38: 407–415.

Bligh, H.F.J., R.I. Till & C.A. Jones, 1995. A microsatellitesequence closely linked to the waxy gene of Oryza sativa.Euphytica 86: 83–85.

Burton, R.A., J.D. Bewley, A.M. Smith, M.K. Bhattacharyya, H.Tatge, S. Ring, V. Bull, W.D.O. Hamilton & C. Martin, 1995.Starch branching enzymes belonging to distinct enzyme famil-ies are differentially expressed during pea embryo development.Plant J 7: 3–15.

Cai, X.L., Z.Y. Wang, Y. Xing, J.L. Zhang & M.M. Hong, 1998.Aberrant splicing of intro1 leads to the heterogeneous 5’UTRand decreased expression of the waxy gene in rice cultivars ofintermediate amylose content. Plant J 14: 459–465.

Guan, H.P. & J. Preiss, 1993. Differentiation of the properties of thebranching isozymes from maize (Zea mays). Plant Physiol 102:1269–1273.

Isshiki, M., K. Morino, M. Nakajima, R.O. Okagaki, S.R. Wessler,T. Izawa & K. Shimamoto, 1998. A naturally occurring func-tional allele of he waxy locus has a GT to TT mutation at the 5’splice site of the first intron. Plant J 15: 133–138.

James, M.G., K. Denyer & A.M. Myers, 2003. Starch synthesis inthe cereal endosperm. Curr Opin Plant Biol 6: 1–8.

Juliano, B.O., 1971. A simplified assay for milled-rice amylose.Cereal Sci Today 16: 334–340.

Juliano, B.O., 1985. Criteria and tests for rice grain qualities.In: B.O. Juliano (Ed.), Rice Chemistry and Technology, pp.443–513. American Association of Cereal Chemists, Saint Paul,Minnesota.

Larkin, P.D., A.M. McClung, N.M. Ayres & W.D. Park, 2003. Theeffect of the Waxy locus (Granule Bound Starch Synthase) onpasting curve characteristics in specialty rice (Oryza sativa L.).Euphytica 134: 1–11.

Liu, K.D., J. Wang, H.B. Li, C.G. Xu, A.M. Liu, X.H. Li & Q.F.Zhang, 1997. A genome-wide analysis of wide compatibility inrice and the precise location of the S5 locus in the molecular map.Theor Appl Genet 95: 809–814.

353

Martin, C. & A.M. Smith, 1995. Starch biosynthesis. Plant cell 7:971–985.

Mikami, L., L.V. Dung, H.Y. Hirano & Y. Sano, 2000. Effects of thetwo most common Wx alleles on different genetic background inrice. Plant Breed 119: 505–508.

Mizuno, K., K. Kimura, Y. Arai, T. Kawasaki, H. Shimada &T. Baba, 1992. Starch branching enzymes from immature riceseeds. J Biochem (Tokyo) 112: 643–651.

Mizuno, K., T. Kawasaki, H. Shimada, H. Satoh, E. Kobayashi,S. Okumura, Y. Arai & T. Baba, 1993. Alteration of the struc-tural properties of starch components by the lack of an isoformof starch branching enzyme in rice seeds. J Biol Chem 286:19084–19091.

Mizuno, K., E. Kobayashi, M. Tachibana, T. Kawasaki, T. Fujimura,K. Funane, M. Kobayashi & T. Baba, 2001. Characteriza-tion of an isoform of rice starch branching enzyme, RBE4, indeveloping seeds. Plant Cell Physiol 42(4): 349–357.

Okagaki, R.J. & S.R. Wessler, 1988. Comparison of non-mutant andmutant waxy genes in rice and maize. Genetics 120: 1137–1143.

Ong, M.H. & J.M.V. Blanshard, 1995. Texture determinants ofcooked, parboiled rice. 2. Physicochemical properties and leach-ing behavior of rice. J Cer Sci 21: 261–269.

Sano, Y., M. Katsumata & K. Omura, 1986. Genetic studies of spe-ciation in cultivated rice. Inter- and intraspecific differentiationin the waxy gene expression of rice. Euphytica 35: 1–9.

SAS Institute Inc., 2000. SAS/STAT User’s Guide, Cary, NC, USA.Sasaki, T., 2002. Rice genomics to understand rice plant as an

assembly of genetic codes. Curr Sci 83: 834–839.Smith, A.M., K. Denyer & C. Martin, 1997. The synthesis of the

starch granule. Annu Rev Plant Physiol Plant Mol Biol 48: 67–87.

Takeda, Y., S. Hizukuri & B.O. Juliano, 1987. Structures of amyl-opectins with low and high affinities for iodine. Carbohydr Res168: 79–89.

Tan, Y.F., J.X. Li, S.B. Yu, Y.Z. Xing, C.G. Xu & Q.F. Zhang, 1999.The three important traits for cooking and eating quality of ricegrains are controlled by a single locus in an elite rice hybrid,Shanyou 63. Thero Appl Genet 99: 642–648.

Tan, Y.F. & Q.F. Zhang, 2001. Correlation of simple sequence repeatvariants in the leader sequence of the waxy gene with amylosecontent of the grain in rice. Acta Botanica Sinica 43: 146–150.

Wang, Z.Y., F.Q. Zheng, G.Z. Shen, J.P. Gao, D.P. Snustad, M.G.Li, J.L. Zhang & M.M. Hong, 1995. The amylose content in riceendosperm is related to the post-transcriptional regulation of thewaxy gene. Plant J 7: 613–622.

Yamanouchi, H. & Y. Nakamura, 1992. Organ specificity of iso-forms of starch branching enzyme (Q-enzyme) in rice. Plant CellPhysiol 33: 985–991.