Characterization of Myanmar Paw San Hmwe Accessions Using Functional Genetic Markers

1, 2, 3 4 2, 5 1 1 Kyaw Swar OO , Alisa KONGJAIMUN , Srisawat KHANTHONG , Myint YI , Tin Tin MYINT , 2 2 6 2, 3, 7 Siriporn KORINSAK , Jonaliza Lanceras SIANGLIW , Khin Myo MYINT , Apichart VANAVICHIT , 3 2 Chanate MALUMPONG , Theerayut TOOJINDA (1Rice Division, Department of Agricultural Research, Yezin, Nay Pyi Taw 15013, Myanmar; 2Rice Gene Discovery Unit, National Centre for Genetic Engineering and Biotechnology, National Science and Technology Development Agency, Kasetsart University, Kamphaeng Saen, Nakhon Pathom 73140, Thailand; 3Agronomy Department, Faculty of Agriculture at Kamphaeng Saen, Kasetsart University, Kamphaeng Saen, Nakhon Pathom 73140, Thailand; 4Center of Excellence on Agricultural Biotechnology, Office of Higher Education Commission, Ministry of Education, Bangkok 10900, Thailand; 5Plant Breeding, Faculty of Agriculture at Kamphaeng Saen, Kasetsart University, Kamphaeng Saen, Nakhon Pathom 73140, Thailand; 6Plant Biotechnology Center, Myanmar Agriculture Service, Yangon 11021, Myanmar; 7Rice Science Center, Kasetsart University, Kamphaeng Saen, Nakhon Pathom 73140, Thailand)

Abstract: Paw San Hmwe (PSM) rice has been cultivated in many areas of Myanmar for a long time. Strong aroma, good taste and its elongation during cooking are the key characteristics of PSM rice. Thirty-one PSM accessions were genotypically characterized, and their physical grain and cooking quality traits were studied. We used specific gene markers associated with aroma, apparent amylose content (AAC) and alkali spreading value to determine the alleles carried by different PSM accessions. The results revealed that six PSM accessions (PSM10, PSM12, PSM13, PSM21, PSM22 and PSM30) had a 3-bp insertion in Os2AP gene. Gel consistency (GC) allele was predominant among the PSM accessions for gelatinization temperature (GT), however, the phenotype observed was between low and intermediate GT because of the combination of the GC allele with the presence of low GT allele at heterozygous state from the other loci of the SSIIa gene. Intermediate to high AAC was observed among the PSM accessions corresponding to the haplotype identified for the single nucleotide

polymorphism G/T and the (CT)n repeat in the Wx gene. The characterization and grouping data of PSM accessions posted benefits to Myanmar seed banks, and our results will help in maintaining the integrity of PSM rice variety. Key words: Paw San Hmwe; rice; aroma; apparent amylose content; gel consistency; gelatinization temperature; alkali spreading value; functional genetic marker; grain quality; haplotype analysis

Quality traits are important characteristics of rice characteristics include apprent amylose content (AAC) affecting the market value and consumer acceptance. (Webb, 1980; Juliano, 1985), gelatinization temperature The demand for high-quality rice varieties has been (GT) (Little et al, 1958) and gel consistency (GC) increasing because of the recent changes in consumer (Cagampang et al, 1973). Good fragrance is another preferences and market requirements. The quality of important feature of high-quality rice. Amylose and rice grain is determined by many factors, such as grain amylopectin contents and their structures determine appearance, nutritional value, cooking and taste the cooking, processing and culinary characteristics of properties (Juliano et al, 1990). Physical properties of the rice grain. Generally, rice grains contain 16%– the grain, including size, shape, color, uniformity, 30% of amylose content in the storage starch. Starch general appearance, chalkiness, milling and head rice properties of rice grains are determined by AAC, GC recovery, are the most important traits determining the and GT (Juliano, 1998; Bergmann et al, 2004). Amylose overall quality (Dela Cruz and Khush, 2000). Cooking content is primarily controlled by waxy gene (Wx) located on chromosome 6. Wx encodes the enzyme granule-bound starch synthase, which plays a key role Received: 28 August 2014; Accepted: 29 December 2014 in amylose synthesis in rice (Tan et al, 1999; Fan et al, Corresponding author: Theerayut TOOJINDA ([email protected]) 2005). GT of the rice flour is controlled by alkali

54 Rice Science, Vol. 22, No. 2, 2015 degeneration locus (alk) encoding a soluble starch from Aeyawady division, 7 from Pago division, 4 synthase IIa (SSIIa) isoform (Fan et al, 2005; Bao et al, from Yangon division, and 1 each from Chin, Kayin, 2006). GC is reportedly associated with Wx gene (Tan Mon, Rakhine and Sagaing divisions (Table 1). Three et al, 1999; Fan et al, 2005; Fitzgerald et al, 2009a). rice varieties, Basmati 370, KDML105 and IR57514, Fragrance is very important in high-quality rice because were used as standard controls to compare the allele it plays a considerable role in determining the price forms of the specific gene markers in this study. The and market. Fragrance of rice grains depends on the plants were grown 25 cm apart with 25 cm between combination of many biochemical compounds (Petrov rows. One plant in the inner portion of the plot was et al, 1996). The most important compound is 2-acetyl- selected for leaf and grain samples for DNA and grain 1-pyrroline (2AP), whose production is controlled by quality analysis. a single recessive gene located on chromosome 8. DNA markers BADH2 (Niu et al, 2008) and Os2AP (Vanavichit et al, 2008) have been identified as genes responsible for Seven specific gene markers for SSIIa, Wx and Os2AP the synthesis of 2AP (Bradbury et al, 2005). loci were used to genotype all plant materials (Table 2). Paw San Hmwe (PSM) from Myanmar, Basmati DNA markers SNP2340-41 and SNP2209, located 370 from India, Jasmine rice (KDML105) from Thailand within SSIIa gene on the short arm of chromosome 6 and Koshihikari from Japan are the best rated among (Lanceras et al, 2000; Bao et al, 2006) and considered the high-quality rice varieties. These varieties are as the most informative single nucleotide polymorphisms priced higher because of their flavor and some other (SNPs), were used to distinguish between high and quality characteristics. These varieties constitute important low GT. These markers were developed by the Rice genetic resources for improving cooking properties of Gene Discovery Unit (Katengam et al, 2008). rice in several rice breeding programs. PSM is one of Waxy (CT)n and Waxy (G-T) markers were used to Paw San rice varieties, grown in the rain-fed lowland areas of Ayeyarwady Division in Myanmar. It is well adapted to the particular local environmental conditions Table 1. Information of Paw San Hmwe (PSM) accessions used in this study. especially in delta region of Myanmar, and known as Pearl Paw San in the international market because of Code Accession No. Local name Region Type its strong aroma, good cooking quality and substantial PSM1 807 Paw San Bay Kyar Aeyawady Indica PSM2 870 Mee Don Aeyawady Japonica grain elongation during cooking (Thein, 2011). With PSM3 923 NgaKywe Yangon Japonica such exceptionally good qualities, 100% of this variety PSM4 930 Paw San Hmwe Pago Indica is exported in super-vacuum packages under the brand PSM5 962 Paw San Hmwe Pago Indica ‘Longevity’. In 2011, PSM was awarded as the ‘World PSM6 998 Mee Don Saba Kayin Indica PSM7 1 045 ZawGyiPyan (Mee don) Aeyawady Indica Best Rice 2011’ and is considered to be the top-grade PSM8 1 105 NgaKywe (U To) Aeyawady Japonica rice in Europe. The culinary traits and genetic and PSM9 1 139 Paw San Hmwe Pago Indica phenotypic characteristics of many varieties have been PSM10 1 207 Paw San Hmwe Aeyawady Japonica PSM11 1 495 Paw San Hmwe Yangon Indica extensively studied. However, the molecular marker- PSM12 1 627 Paw San Hmwe Mon Japonica based analysis of these properties in the Paw San rice PSM13 1 641 Paw San Hmwe Aeyawady Japonica has not been undertaken. The objective of this study PSM14 1 981 NgaKywe Yangon Indica was to investigate the genetic characterization of PSM PSM15 2 117 Mee Don Hmwe Aeyawady Indica PSM16 2 176 Hnan War Mee Don Rakhine Indica accessions obtained from the Myanmar germplasm PSM17 2 447 NgaKywe Pago Indica bank, based on the alleles of known genes associated PSM18 2 460 Paw San Bay Kyar Aeyawady Indica with grain quality and their corresponding phenotypes. PSM19 2 499 Paw San Hmwe Pago Indica Our results may encourage the utilization of PSM rice PSM20 2 500 Paw San Hmwe Pago Japonica PSM21 2 501 Paw San Hmwe Yangon Japonica as a genetic resource in rice breeding programs. PSM22 2 502 Paw San Hmwe Aeyawady Japonica PSM23 2 522 Paw San Hmwe Aeyawady Japonica MATERIALS AND METHODS PSM24 2 578 Paw San Shwe War Aeyawady Indica PSM25 2 620 Paw San Hmwe Aeyawady Indica PSM26 2 878 Hnan War Mee Don Sagaing Indica Rice materials PSM27 2 946 Paw San Hmwe Pago Indica PSM28 3 225 Mee Don Yoe Sein Aeyawady Indica We obtained 31 PSM accessions from the gene bank PSM29 3 530 Bay Kyar Lay Aeyawady Japonica of Department of Agricultural Research (DAR), PSM30 5 802 Paw San Hmwe Chin Japonica Myanmar. Of these 31 accessions, 15 were collected PSM31 11 800 Paw San Hmwe Aeyawady Indica

Kyaw Swar OO, et al. Grain Quality Characterization of Paw San Hmwe Accessions 55

Table 2. Gene specific markers for grain qualities used in clustering 31 Paw San Hmwe accessions.

Trait Gene Chromosome Marker Reference Amylose content Wx 6 Waxy (G-T) Ayres et al, 1997

6 Waxy (CT)n Bligh et al, 1995 Fragrance Os2AP 8 Aromarker Vanavichit et al, 2008 8 3In2AP Myint et al, 2012a 8 FMbadh2-E2A Shi et al, 2008 Gelatinization temperature SSIIa 6 SNP2340-41 Katengam et al, 2008 6 SNP2209 Katengam et al, 2008

characterize the alleles of the Wx gene. Waxy (CT)n, a by denaturation at 94 ºC for 3 min, followed by 35 microsatellite marker, is used to amplify the fragment cycles of 94 ºC for 30 s, 55 ºC for 30 s, and 72 ºC for containing a (CT)n repeat in the 5′ untranslated region 2 min. Final 7 min incubation at 72 ºC completed the of Waxy exon 1 (Bligh et al, 1995) and putative 5′ primer extension. splice site of the leader intron in the Wx gene (Bligh To identify the waxy allele (Wxa/Wxb), PCR et al, 1995; Ayres et al, 1997). When the fragment is products obtained using Waxy markers were digested cut by AccI restriction enzyme, an SNP in the putative with AccI restriction enzyme in 15 µL solution 5′ splice site of the leader intron, located on the short containing 10 µL of PCR products with 1 × NE buffer arm of chromosome 6 (Ayres et al, 1997; Chen et al, and 1 U of restriction enzyme AccI (New England Bio 2008), can be detected. It can differentiate between Labs) for 1 h at 37 ºC. high (Wxa) and low (Wxb) amylose content. PCR reactions for agarose-based primers SNP2209 Three DNA markers, Aromarker (Vanavichit et al, and SNP2340-41 were performed in a total volume of 2008), FMbadh2-E2A (Shi et al, 2008) and 3In2AP 20 µL containing 50 ng of DNA template, 1 × PCR (Myint et al, 2012a), were used to characterize Os2AP buffer, 1.5 mmol/L MgCl2, 0.2 mmol/L dNTPs, 0.25 alleles. Aromarker has been developed for the detection µmol/L each of primers P and Q, 0.125 µmol/L each of the 8-bp deletion (creating a stop codon in fragrant of primers A/G and GG/TT, and 1 U of Taq polymerase. rice) in the seventh exon of Os2AP (Vanavichit et al, PCR reaction was initiated by denaturation at 95 ºC 2008), and FMbadh2-E2A for a region with 7-bp for 3 min, followed by 35 cycles of 95 ºC for 30 s, 62 ºC deletion in the second exon of the Os2AP gene on for 30 s and 72 ºC for 30 s. Final 5-min incubation at chromosome 8, which is discovered in Wuxiangjing 9 72 ºC was used for the completion of primer extension. (japonica) (Shi et al, 2008). The 3In2AP marker has PCR product electrophoresis been developed using the region with a 3-bp insertion in the thirteenth exon of Os2AP in Myanmar rice The amplified PCR products of Aromarker, FMbadh2- Yangon Saba (Myint et al, 2012a). These markers can E2B, 3In2AP and Waxy (CT)n were loaded onto 4.5% distinguish between the alleles associated with aromatic polyacrylamide gels and separated by electrophoresis and non-aromatic rice. in 1 × TBE buffer at 50 ºC and 60 W. The PCR products were visualized using silver standing. The DNA extraction and PCR amplification amplified PCR markers Waxy (G-T), SNP2209 and Young leaf samples were collected from each accession, SNP2340-41 were loaded onto 1.5% agarose gels and placed in plastic bags, and stored on ice for DNA separated by electrophoresis in 0.5 × TBE buffer at 90 extraction. Total genomic DNA was isolated from 0.5 V for 60 min. The PCR products were visualized g of leaf tissue (from PSM accessions and controls) using ethidium bromide staining. Polymorphisms in the using the DNA Trap® kit developed in the DNA DNA profiles were scored using visual comparisons Technology Laboratory (http://dnatec.kps.ku.ac.th), with standard controls and standard DNA gene ladder. Kasetsart University, Kamphaeng Saen Campus, Thailand. Evaluation of physical and chemical properties of PCR reactions for polyacrylamide gel electrophoresis rice grains (PAGE) analysis contained 50 ng of DNA template, 1 × PCR buffer, 2 mmol/L MgCl2, 0.2 mmol/L dNTPs, Seeds were collected from the plants whose leaves 0.2 µmol/L each of forward and reverse primers, and 1 were used for DNA extraction. The harvested seeds U of Taq DNA polymerase. The volume was adjusted were sun-dried until the moisture content was ≤ 14%, to 10 µL with distilled water. PCR reaction was initiated and stored for two months. Before analysis, the seeds

56 Rice Science, Vol. 22, No. 2, 2015 were milled and whole grains were selected and oven- incubated at 65 ºC for 3 h and then cooled. The lids dried. were opened one by one, and samples were sniffed by three panelists and scored as strongly scented (2), mildly Brown rice length and width (BRL and BRW) scented (1) and non-scented (0) (Wanchana et al, 2003). Ten de-husked entire brown rice grains were measured using vernier calipers. Using the length/width ratio, Statistical analysis the grains were classified as long slender (LS), short Analysis of variance slender (SS), medium slender (MS), long bold (LB) All grain quality traits were subjected to statistical and short bold (SB) (Dela Cruz and Khush, 2000; analysis using CropStat Version 7.2. Least significant Anonymous, 2004). difference (LSD) test was used to compare the means Alkali spreading value (ASV) and clearing test of grain quality traits in the different accessions. Six milled rice grains were placed in Petri plates, 15 Simple linear regression and analysis of variance were mL of 1.7% KOH was added, and the plates were kept used to analyze the genetic associations. at room temperature for 23 h. Then, the ASV was Analysis of genetic dissimilarity evaluated and scored. The values 1–3 were classified The data were scored using the bands of the quality as high GT (more than 74 ºC ), 4–5 as intermediate GT markers for 31 PSM accessions and controls (KDML105, (70 ºC –74 ºC), and 6–7 as low GT (less than 70 ºC) Basmati 370 and IR57514). Nineteen simple sequence (Perez and Juliano, 1978). repeat (SSR) markers (RM1, RM5, RM237 and Apparent amylose content (AAC) RM431 on chromosome 1, RM154 on chromosome 2, AAC was measured using the procedure of Juliano RM338 on chromosome 3, RM124 on chromosome 4, (1971, 1985) with minor modification. Rice flour (100 RM538 on chromosome 5, RM510 on chromosome 6, mg) was mixed with 1 mL of 95% ethanol and 9 mL RM11 on chromosome 7, RM25, RM44 and RM447 of 1.0 mol/L NaOH, and incubated at room temperature on chromosome 8, RM215 and RM316 on chromosome overnight to gelatinize the starch. The samples were 9, RM271 and RM474 on chromosome 10, RM287 on diluted to 100 mL with distilled water. A 5 mL aliquot chromosome 11 and RM1227 on chromosome 12) of this suspension was mixed with 1 mL acetic acid (1 (Myint et al, 2012b) were also used to genotype the mol/L) and 1.5 mL iodine solution (0.2% iodine and PSM accessions and controls, which were used in 2% potassium iodide), and the volume was adjusted to grouping the accessions. Khin Myo Win, who initially 100 mL with distilled water. The samples were incubated established the grouping of several Myanmar accessions, at room temperature for 20 min, and then the absorbance kindly supplied the genotypic data of IR8, IR64, at 620 nm was measured. NaOH solution was used as Basmati 217, Aus 143, Baran Boro, Indane, Azucena, a control. The AACs of different accessions were obtained Nipponbare, Leikkalay Meedon, Paw San Taung Pyan using a standard graph (Perez and Juliano, 1978). Hmwe and Nga Bya Yin Kauk Hnyin Net rice (Myint GC et al, 2012b). We used them as controls to help in the In 13 mm × 150 mm test tubes, 200 µL xylene solution group assignments of the PSM accessions. To evaluate was added to each 100 mg rice powder sample. The the genetic dissimilarities between the genotypes, the tubes were vortexed, 2.0 mL of 0.2 mol/L KOH was polymorphic bands were used to construct a binary added and the tubes were vortexed again for 1 min. matrix. The gene-specific markers were obtained from The tubes were covered with glass marbles and placed the dissimilarity matrix computed by a shared allele in a boiling water bath for 8 min, then removed from index using DARwin software (Perrier and Jacquemoud- the bath and kept at room temperature for 5 min. The Collet, 2006). An unweighted neighbor-joining (NJ) tubes were cooled in ice-cold water for 15 min. Finally, tree was built using this dissimilarity matrix. the tubes were laid down horizontally over a graphing paper with mm-scale. The gel was classified as soft RESULTS for more than 60 mm spread, medium for 40–60 mm, and hard for less than 40 mm (Bhattacharya, 1979). Grain quality and agronomic characteristics Aroma PSM rice had an average aroma sensory score of 0.9 Five seeds of the brown rice from each accession were (Table 3). There were nine accessions with aroma added to 200 µL distilled water. The tubes were score ≥1.0 (2.4 was the highest score), which were clustered

Kyaw Swar OO, et al. Grain Quality Characterization of Paw San Hmwe Accessions 57

58 Rice Science, Vol. 22, No. 2, 2015 together. In terms of cooking properties, the average hill (27.5 g), grain weight per panicle (3.8 g), effective ASV, GC and AAC of PSM were 5.4, 30.8 mm and tiller number per hill (9.9), filled grain number per 23.6%, respectively. We also found that most PSM panicle (122.5), seed-setting rate (82.3%), and 100- accessions were non-aromatic, with low-to- grain weight (3.3 g) are presented in Table 3. intermediate ASV, hard GC and intermediate-to-high Allele analysis using grain quality markers AAC. Our results showed that almost all of the aromatic PSM accessions had low ASV. There were Amylose content some variants with intermediate ASV, important for G/T in the putative 5′ splice site of the leader intron of maintaining the desired characteristics of PSM rice the Wx gene can be used to distinguish between high (Table 3). and low amylose content alleles. PCR product of Most of the PSM accessions had high AAC. Most amplification was digested with AccI and yielded 233 accessions had GC less than 40 mm (hard consistency) bp and 77 bp for G allele while 310 bp for T allele. except for PSM17, which belonged to the medium Twenty-six PSM accessions had G allele suggesting consistency group. We also examined physical grain high amylose content similar to the control varieties properties such as the length and width of paddy and Basmati 370 and IR57514 (Table 4). Two PSM brown rice grains and obtained the length/width ratios. accessions showed T allele suggesting low amylose The average paddy length, width, and length/width content, like KDML105. Three accessions were ratio were 7.8 mm, 3.4 mm, and 2.3, respectively; the heterozygous for these alleles. same parameters for the average brown rice grains Waxy (CT)n microsatellite was used to screen the were 5.5 mm, 3.0 mm, and 1.8 (Table 3 and Fig. 1). PSM germplasm. We detected five alleles of (CT)n Brown rice seeds of most PSM accessions were short (Table 4): the first allele of 330 bp, the second of 326 or medium length, with a previously reported bp, the third of 310 bp and two heterozygous alleles characteristic of PSM (Calingacion et al, 2014). (330/310 bp and 326/310b bp). The 330 bp allele was Agronomic characteristics such as average days to detected in almost all accessions. There were some flowering (143.0 d), plant height (144.5 cm), panicle exceptions. PSM6 had the third allele (310 bp), PSM3, length (26.6 cm), yield (4.7 t/hm2), grain weight per PSM19 and PSM21 had the heterozygous allele

PSM6 PSM16 PSM26 PSM12

PSM17 PSM15 PSM18 PSM11

PSM31 PSM19 PSM14 PSM24

PSM9 PSM22 PSM5 PSM3

Fig. 1. Morphology of paddy and milled rice grains of some selected Paw San Hmwe (PSM) accessions.

Kyaw Swar OO, et al. Grain Quality Characterization of Paw San Hmwe Accessions 59

TTable 4. Thirty one Paw San Hmwe (PSM) accessions and controls assayed with seven rice quality markers.

Code Waxy (G-T) Waxy (CT)n SNP2340-41 SNP2209 3In2AP FMbadh2-E2A Aromarker PSM1 G 19 High GT High/low GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM2 G 19 High GT High/low GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM3 T 19/11 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM4 G 19 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM5 G 19 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM6 G 11 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM7 G 19 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM8 G 19 High GT High GT Non aroma (197/194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM9 G 19 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM10 G 19 High GT High GT Aroma (197 bp) Non aroma (207 bp) Non aroma (400 bp) PSM11 G 19 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM12 G 19 High GT High GT Aroma (197 bp) Non aroma (207 bp) Non aroma (400 bp) PSM13 G 19 High GT High GT Aroma (197 bp) Non aroma (207 bp) Non aroma (400 bp) PSM14 G/T 17/11 High/low GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM15 G 19 High GT High/low GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM16 G 19 High GT High/low GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM17 G 17 High GT High/low GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM18 G/T 19 High GT High/low GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM19 G 19/11 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM20 G 19 High GT High/low GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM21 T 19/11 High GT High GT Aroma (197 bp) Non aroma (207 bp) Non aroma (400 bp) PSM22 G 19 High GT High GT Aroma (197 bp) Non aroma (207 bp) Non aroma (400 bp) PSM23 G 19 High GT High GT Non aroma (197/194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM24 G 19 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM25 G 19 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM26 G 17/11 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM27 G 19 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM28 G 19 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM29 G/T 19 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) PSM30 G 19 High GT High GT Aroma (197 bp) Non aroma (207 bp) Non aroma (400 bp) PSM31 G 19 High GT High/low GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) KDML105 T 17 Low GT High GT Non aroma (194 bp) Non aroma (207 bp) Aroma (392 bp) Basmati 370 G 17 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Aroma (392 bp) IR57514 G 11 High GT High GT Non aroma (194 bp) Non aroma (207 bp) Non aroma (400 bp) G means 233/77 bp Wxa, high amylose content; T means 310 bp Wxb, low amylose content.

Waxy (CT)n repeat was considered by comparison between allele and CT repeat in previous reports. The values in this column are the number of CT repeats, and 19, 17 and 11 correspond to 330, 326 and 310 bp alleles, respectively.

(330/310 bp), PSM14 and PSM26 had the heterozygous (G/A). The control rice variety KDML105 showed G allele (326/310 bp), and PSM17 had the second allele. allele associated with high GT (Table 4). Considering KDML105 and Basmati 370 also had the second allele, the combination of SNP2340-41 and SNP2209, the indicative of low amylose content due to 17 CT haplotype 2209G-2340-41GC was associated with repeats (Prathepha, 2003), while IR57514 had 11 CT optimal SSIIa activity and high GT, while the repeats. haplotype 2209G-2340-41TT and 2209A-2340-41GC had no SSIIa activity and was low GT (Katengam et al, GT 2008). Since the bases at the location 2340-41 of We used SNP2340-41 marker to determine the KDML105 are TT, GT of this variety is low although distribution of GC and TT alleles. All PSM accessions the allele at SNP2209 is G. For the PSM accessions, used had the GC allele associated with high GT (ASV three sets of allele combinations were found for = 1–3) except for PSM14, which was heterozygous SNP2340-41 and SNP2209: GC-G/A (heterozygous at (GC/TT). The control variety KDML105 with low GT SNP2209), GC/TT-G (heterozygous at SNP2340-41) had TT allele, while GC allele (high GT) was found in and GC-G. Most of the accessions had GC-G Basmati 370 and IR75714 (Table 4). SNP2209 were combination indicating high GT (Table 4). used to determine the distribution of G and A alleles, and 23 accessions had G allele associated with high Aroma GT, while the remaining accessions were heterozygous Three functional DNA markers were used to

60 Rice Science, Vol. 22, No. 2, 2015 distinguish fragrant from non-fragrant grain alleles of Cluster analysis and genetic relationships gene Os2AP on chromosome 8: Aromarker (Vanavichit Cluster analysis and binary matrix were obtained et al, 2008), FMbadh2-E2A (Shi et al, 2008) and based on genetic dissimilarity among PSM accessions 3In2AP (Myint et al, 2012a). and reference controls. Those dissimilarities were We found that 25 accessions were non-aromatic clearly reflected in the alleles of grain quality markers with allele sizes of 194 bp or 197/194 bp when and microsatellite markers [standard markers used in screeming with 3In2AP, while six accessions (PSM10, grouping rice accessions reported by Myint et al PSM13 and PSM22 from Aeyawady, PSM12 from (2012b)]. Grouping of rice PSM accessions and Mon, PSM21 from Yangon and PSM30 from Chin) controls is presented in Fig. 2. G1 included the indica were aromatic with 197 bp allele. Aromatic rice Yangon group varieties with the indica controls such as Saba had 197 bp allele (Myint et al, 2012b), and KDML105 and IR57514. Surprisingly, two accessions aromatic rice KDML105 had 194 bp allele (Table 4). PSM6 and PSM14 were included in this group. G2 The results of screening with FMbadh2-E2A included the Boro/Aus types, similar to the isozyme showed that all of the PSM accessions had non aroma grouping. The Basmati group belonged to G5A, while allele similar to those of IR57514, Basmati 370 and the japonicas were in G6. Most PSM accessions KDML105 (Table 4). belonged to G5B, as reported by Myint et al (2012b). No aromatic alleles of Os2AP were found in the Three sub-groups of PSM accessions could be PSM accessions using Aromarker. KDML105 and determined in this group, based on the grain quality Basmati 370 rice had 392 bp segments indicating that markers. It is clear that aromatic accessions (PSM10, these controls were from aromatic rice varieties, while PSM22, PSM30, PSM12 and PSM21) form SG1. a non-aromatic control IR57514 had 400 bp allele. Surprisingly, SG1 had low GT values and These results suggest that the aroma rice of different intermediate AAC. SG2 and SG3 contained the non- origins may be controlled by different alleles of the aromatic accessions except for PSM8, which was Os2AP gene. aromatic and heterozygous for 3In2AP marker. PSM8

Fig. 2. Dendrogram of 31 Paw San Hmwe (PSM) accessions (blue), reference checks (violet) and standard checks (red) generated by cluster analysis using unweighted neighbor-joining of DARwin software based on 7 rice quality markers and 19 SSR markers. G1, G2, G5A, G5B and G6 represent the cluster groups; SG1, SG2 and SG3 represent the sub-group in G5B based on the genotype and phenotype of the PSM accessions.

Kyaw Swar OO, et al. Grain Quality Characterization of Paw San Hmwe Accessions 61 belonged to SG3, which had low-to-intermediate GT genetic diversity in the germplasm. Only eight and intermediate-to-high AAC. Gelatinization in this accessions with 3-bp insertion were aromatic, group cannot be well separated using the SNP markers indicating that PSM accessions from the gene bank of for GT. SG2 had low-to-intermediate GT and DAR, which should be aromatic, may have mutated or intermediate AAC. are contaminated and need cleaning up. Intermediate-to-high AAC and GT are preferred in DISCUSSION Myanmar (Calingacon et al, 2014). Thus, the PSM accessions are probably selected by the farmers to have PSM rice can easily maintain its popularity due to its such desired properties. GC, commonly determined by superior texture, taste and good milling recovery. measuring the spread of cooled gel made from flour Grains of PSM elongate up to three times during the cooked in 0.2 mol/L KOH (Cagampang et al, 1973), is cooking and have a strong aroma. In this study, we a measure of firmness of rice after cooking and is used provide information on culinary and agronomic to classify rice varieties with the same AAC, particularly properties of PSM germplasm, which may be useful in in the high AAC class, into hard, medium and soft rice breeding programs and would help to maintain the consistency types (Cagampang et al, 1973; Kohlwey, market position of PSM rice (Calingacion et al, 2014). 1994). The gene coding for GC is located within the Phenotypic and genotypic assessment of grain Wx locus (Tang et al, 1991; Tian et al, 2005). Quantitative quality of PSM accessions trait loci of AAC have been also mapped onto this locus (Lanceras et al, 2000; He et al, 2006; Zheng et al, The aromatic PSM accessions have a 3-bp insertion in 2008), indicating a relationship between AAC and GC. exon 13 of chromosome 8. No aromatic accession Varieties with waxy and very low AAC tended to with a 8-bp deletion in exon 7 or a 7-bp deletion in have soft GC (high GC values), while in the varieties exon 2 was found. The 3-bp insertion in the exon 13 from other AAC classes, GCs ranged from hard to soft. of Os2AP is common in Myanmar aromatic accessions, Thus, other factors apart from amylose content must particularly in the popular and widely grown aromatic be contributing to the strength of the cooling gel, and varieties such as PSM and its relatives (Paw San Yin elucidating these mechanisms will help in breeding and Paw San Bay Kyar). Myint et al (2012b) suggests programs searching for the superior texture traits that the 3-bp insertion allele has originated in (Cuevas and Fitzgerald, 2012). Myanmar region. However, a fragment with the 8-bp Among the markers developed to determine amylose deletion in exon 7, another aromatic allele, can be content in rice, the G/T alleles explains 80%–90% of found in most aromatic accessions grown in South and the variation observed in non-waxy rice accessions Southeast Asia, including the most popular aromatic (Aryes et al, 1997; Bao et al, 2006). Consequently, rice accessions, KDML105 and Basmati 370. The 7-bp there are rice accessions with T allele showing low or deletion in exon 2 is characteristic for Chinese intermediate AAC, and some with G allele showing accessions (Shi et al, 2008), while the 3-bp insertion is high or intermediate AAC. This observation suggests more restricted and mainly found in Myanmar. The that there could be other positions in the Waxy gene localization of the 3-bp mutation in Myanmar reflects that could be involved in the separation of AAC the confinement of the PSM accessions in the country. classes. According to Hoai et al (2014), the low AAC Myint et al (2012b) have reported the 8-bp deletion is associated with the C allele in the exon 10 of the found in some of the indica aromatic accessions from Waxy gene rather than the change from G to T at the Myanmar. The authors believe that the presence of splicing donor sites of the first intron. This allele this deletion is the result of preference for indica change in the exon 10 may explain the intermediate varieties in the neighboring regions or preferences AACs of PSM3 and PSM21, which had T allele at the associated with changes in aroma caused by different splice site of the leader intron. The work of Biselli mutations. However, the results of this study differed et al (2014) complements the findings of Hoai et al from those of Myint et al (2012a), in which only the (2014) since accessions with AAC between 21% and 3-bp insertion was found in PSM accessions and the 22% possess T alleles in the splice site of the leader 8-bp deletion was not detected. Our study focused on intron. In our study, C allele at the exon 10 was not relatively small number of PSM accessions that included in the analysis to predict the corresponding happened not to have the 8-bp deletion, with a low AAC. It is clear that the allele change from T to C in

62 Rice Science, Vol. 22, No. 2, 2015 the exon 10 is associated with changes in AAC. of Myint, some of the Myanmar accessions, like Phenotypically, almost all PSM accessions showed NgaKywe and Meedon Yoyo, are grouped together intermediate-to-high AAC corresponding to the with the indica accessions. NgaKywe and Mee Don genotype. Accessions that were heterozygous for G/T Saba used in our study were also grouped with the alleles had high AAC, which could have been caused indica varieties because of grain size and seed color by the additive and dominant nature of the genes characteristics (Fig. 1). However, a study by Thien et al controlling AAC (Pooni et al, 1993; Kuo et al, 1997). (2012) has clearly separated Mee Don Saba and The PSM accessions contained 11, 17 or 19 CT NgaKywe from an indica variety KDML105. These repeats, and most of them had 19 CT repeats. PSM3 results show that the molecular characterization of and PSM21 had the T allele, which is normally these accessions can help in their grouping and associated with low AAC. However, in these two demonstrating the possible contaminations in the gene accessions, the results did not correspond to their bank. As mentioned by Glaszmann (1987), isozyme AAC phenotypes. Chen et al (2008) have analyzed group V includes rice in the Indian subcontinent from several rice accessions and shown that (CT)11 group Iran to Myanmar. The clustering of Myanmar PSM has high AAC while (CT)17 and (CT)19 groups include accessions in G5B corresponds with the study of both intermediate and low AAC. It might due to the Glaszmann (1987). combination of the G/T alleles, which corresponds to Our study presents a useful characterization and (CT)n repeat, was heterozygous allele 19/11 that grouping of PSM accessions. The results are of great caused the intermediate AAC observed in those PSM benefit to gene banks, such as the bank of the DAR in accessions, although this combination has not been Myanmar. In some accessions, we found heterozygous detected in several haplotyping studies (Bao et al, allele for some markers, which may be due to 2006; Wan et al, 2007; Chen et al, 2008; Prathepha, contamination. Therefore, it has to be taken into 2008; Biselli et al, 2014). account in maintaining the germplasm. We believe For PSM accessions, three sets of allele combinations that the existing PSM accessions must be carefully were determined for SNP2340-41 and SNP2209: GC- checked and the contaminants should be removed to G/A (heterozygous at SNP2209), GC/TT-G (heterozygous maintain the integrity of the PSM varieties in the at SNP2340-41) and GC-G. Although most PSM international market. The rising popularity of these accessions had GC-G allele (Table 4) suggesting high rice types in the world market can be maintained by GT, the GT phenotypes or ASV scores were mostly distributing rice seed from controlled sources of with intermediate GT (Table 3). It is possible that national repute such as DAR. those SNPs do not correspond precisely to the intermediate GT phenotypes. Cuevas et al (2010) have ACKNOWLEDGEMENTS suggested that intermediate GT is caused by the presence of DP24–35 amylopectin chains that can be This study was supported by the National Center for synthesized by other enzymes. The genetic control of Genetic Engineering and Biotechnology, Thailand GT has not been fully elucidated. Nevertheless, the (Grant No. P1020028), and Generation Challenge available GT markers can be used in varietal Program, Mexico (Grant No. P0900835). The authors differentiation (with high and intermediate versus low are grateful to the Center of Excellence on Agricultural GT) as an indicator of the cooking time of rice Biotechnology, Science and Technology Postgraduate (Cuevas et al, 2010). It is an economically important Education and Research Development Office, Office indicator of quality because shorter cooking time leads of Higher Education Commission, Ministry of to significant savings in fuel costs (Fitzgerald et al, Education, Thailand for the advice in phenotyping 2009b). Consequently, the intermediate GT of PSM of traits. This research was conducted at the Rice Gene accessions may be one of the features to be selected in Discovery Unit, Kasetsart University, Thailand, and the rice breeding programs. materials were provided by Department of Agricultural Cluster analysis and genetic relationships Research, Yezin, Nay Pyi Taw, Myanmar. Based on the dendrogram created for PSM accessions, REFERENCES standard controls and reference controls (Myint et al, 2012b), five groups were identified following the Anonymous. 2004. Laboratory Manual on Rice Grain Quality grouping reported by Myint et al (2012b). In the study Procedure. Directorate of Rice Research, Rajendranagar, Hyderabad:

Kyaw Swar OO, et al. Grain Quality Characterization of Paw San Hmwe Accessions 63

1–20. Fan C C, Yu X Q, Xing Y Z, Xu C G, Luo L J, Zhang Q F. 2005. Ayres N M, McClung A M, Larkin P D, Bligh H F J, Jones C A, The main effects, epistatic effects and environmental Park W D. 1997. Microsatellites and a single-nucleotide interactions of QTLs on the cooking and eating quality of rice in polymorphism differentiate apparent amylose classes in an a doubled-haploid line population. Theor Appl Genet, 110(8): extended pedigree of US rice germplasm. Theor Appl Genet, 1445–1452. 94(6): 773–781. Fitzgerald M A, McCouch S R, Hall R D. 2009a. Not just a grain Bao J S, Corke H, Sun M. 2006. Nucleotide diversity in starch of rice: The quest for quality. Trends Plant Sci, 14(3): 133–139. synthase IIa and validation of single nucleotide polymorphisms Fitzgerald M A, Bergman C J, Resurreccion A P, Möller J, Jimenez in relation to starch gelatinization temperature and other R, Reinke R F, Martin M, Blanco P, Molina F, Chen M H, Kuri physicochemical properties in rice (Oryza sativa L.). Theor Appl V, Romero M V, Habibi F, Umemoto T, Jongdee S, Graterol E, Genet, 113(7): 1171–1183. Reddy K R, Bassinello P Z, Sivakami R, Rani N S, Das S, Wang Bergman C J, Bhattacharya K R, Ohtsubo K. 2004. Rice end-use Y J, Indrasari S D, Ramli A, Ahmad R, Dipti S S, Xie L H, Lang quality analysis. In: Champagne E T. Rice Chemistry and N T, Singh P, Toro D C, Tavasoli F, Mestres C. 2009b. Technology. AACC, St Paul, Minnesota: 415–472. Addressing the dilemmas of measuring amylose in rice. Cereal Bhattacharya K R. 1979. Gelatinization temperature of rice starch Chem, 86(5): 492–498. and its determination. Proc Workshop Chem Asp Rice Grain Glaszmann J C. 1987. Isozymes and classification of Asian rice Qual, 24(2): 116–118. varieties. Theor Appl Genet, 74(1): 21–30. Biselli C, Cavalluzzo D, Perrini R, Gianinetti A, Bagnaresi P, Urso He Y, Han Y P, Jiang L, Xu C W, Lu J F, Xu M L. 2006. S, Orasen G, Desiderio F, Lupotto E, Cattivelli L, Valè G. 2014. Functional analysis of starch-synthesis genes in determining rice Improvement of marker-based predictability of apparent eating and cooking qualities. Mol Breeding, 18(4): 277–290. amylose content in japonica rice through GBSSI allele mining. Hoai T T T, Matsusaka H, Toyosawa Y, Suu T D, Satoh H, Rice, 7(1): 1. Kumamaru T. 2014. Influence of single-nucleotide polymorphisms Bligh H F J, Till R I, Jones C A. 1995. A microsatellite sequence in the gene encoding granule-bound starch synthase I on closely linked to the Waxy gene of Oryza sativa. Euphytica, amylose content in Vietnamese rice cultivars. Breeding Sci, 86(2): 83–85. 64(2): 142–148. Bradbury L M T, Fitzgerald T L, Henry R J, Jin Q S, Waters D L E. Juliano B O. 1971. A simplified assay for milled rice amylose. 2005. The gene for fragrance in rice. Plant Biotech J, 3(3): 363– Cereal Sci Today, 16(4): 334–360. 370. Juliano B O. 1985. Rice: Chemistry and Technology. St. Paul, M N: Cagampang G B, Perez C M, Juliano B O. 1973. A gel consistency American Association of Cereal Chemists: 774. test for eating quality in rice. J Sci Food Agric, 24(12): 1589– Juliano B O, Perez C M, Kaosa-Ard M. 1990. Grain quality 1594. characteristics of export rices in selected markets. Cereal Chem, Calingacion M, Laborte A, Nelson A, Resurreccion A, Concepcion 67(2): 192–197. J C, Daygon V D, Mumm R, Reinke R, Dipti S, Bassinello P Z, Juliano B O. 1998. Varietal impact on rice quality. Cereal Foods Manful J, Sophany S, Lara K C, Bao J S, Xie L H, Loaiza K, El- World, 43(4): 207–222. hissewy A, Gayin J, Sharma N, Rajeswari S, Manonmani S, Katengam S, Kimchaiyong W, Wanchana S, Toojinda T. 2008. Rani N S, Kota S, Indrasari S D, Habibi F, Hosseini M, Tavasoli Association analysis and functional marker developmant of F, Suzuki K, Umemoto T, Boualaphanh C, Lee H H, Hung Y P, soluble starch synthase IIa (SSIIa) and gelationization properties Ramli A, Aung P P, Ahmad R, Wattoo J I, Bandonill E, Romero in Thai rice. In: Proceeding of the 5th International Crop Science M, Brites C M, Hafeel R, Lur H S, Cheaupun K, Jongdee S, Congress & Exhibition, Jeju: 1–4. Blanco P, Bryant R, Lang N T, Hall R D, Fitzgerald M. 2014. Kohlwey D E. 1994. New methods for the evaluation of rice Diversity of global rice markets and the science required for quality and related terminology. In: Marshall W E, Wadsworth J consumer-targeted rice breeding. PLoS One, 9(1): e85106. I. Rice Science and Technology. New York, USA: Marcel Dekker: Chen M H, Bergman C, Pinson S, Fjellstrom R. 2008. Waxy gene 113–137. haplotypes: Associations with apparent amylose content and the Kuo Y C, Webb B D, Stansel J W. 1997. Genetic studies on effect by the environment in an international rice germplasm amylose contenct and amylographic breakdown viscosity of collection. J Cereal Sci, 47(3): 536–545. milled rice by means of generation mean analysis. J Agric Res Cuevas R P, Daygon V D, Corpuz H M, Reinke R F, Waters D L E, China, 46(2): 99–115. Fitzgerald M A. 2010. Melting the secrets of gelatinization Lanceras J C, Huang Z L, Naivikul O, Vanavichit A, Ruanjaichon temperature in rice. Funct Plant Biol, 37(5): 439–447. V, Tragoonrung S. 2000. Mapping of genes for cooking and Cuevas R P, Fitzgerald M A. 2012. Genetic diversity of rice grain eating qualities in Thai Jasmine rice (KDML105). DNA Res, quality. In: Caliskan M. Genetic Diversity in Plants. InTech: 7(2): 93–101. Rijeka: 285–310. Little R R, Hilder G B, Dawson E H. 1958. Differential effect of Dela Cruz N, Khush G S. 2000. Rice grain quality evaluation dilute alkali on 25 varieties of milled white rice. Cereal Chem, procedures. In: Singh R K, Singh U S, Khush G S. Aromatic 35: 111–126. Rices. New Delhi, Calcutta: Oxford and IBH Publishing: 15–28. Myint K M, Arikit S, Wanchana S, Yoshihashi T, Choowongkomon

64 Rice Science, Vol. 22, No. 2, 2015

K, Vanavichit A. 2012a. A PCR-based marker for a locus Tan Y F, Li J X, Yu S B, Xing Y Z, Xu C G, Zhang Q F. 1999. conferring the aroma in Myanmar rice (Oryza sativa L.). Theor The three important traits for cooking and eating quality of rice Appl Genet, 125(5): 887–896. grains are controlled by a single locus in an elite rice hybrid, Myint K M, Courtois B, Risterucci A M, Frouin J, Soe K, Thet K Shanyou 63. Theor Appl Genet, 99(3/4): 642–648. M, Vanavichit A, Glaszmann J C. 2012b. Specific patterns of Tang S X, Khush G S, Juliano B O. 1991. Genetics of gel genetic diversity among aromatic rice varieties in Myanmar. consistency in rice (Oryza sativa L.). J Genet, 70(2): 69–78. Rice, 5: 20. Thein M S. 2011. Genetic Diversity of Meedon Rice (Oryza sativa Niu X L, Tang W, Huang W Z, Ren G J, Wang Q L, Luo D, Xiao L.) Germplasm in Myanmar. [PhD Thesis]. Yezin: Yezin Y Y, Yang S M, Wang F, Lu B R, Gao F Y, Lu T G, Liu Y S. Agricultural University, Myanmar. 2008. RNAi-directed down regulation of OsBADH2 results in Thein M S, Lee G A, Cho G T, Sung J S, Jeong J W, Park J H, aroma (2-acetyl-1-pyrroline) production in rice (Oryza sativa L.). Baek H J. 2012. Assessment of genetic diversity in Meedon rice BMC Plant Biol, 8: 100. (Oryza sativa L.) germplasm. J Korean Soc Intl Agric, 24(2): Perez M C, Juliano B O. 1978. Modification of the simplified 232–240. amylose test for milled rice. Starch, 30(12): 424–426. Tian R, Jiang G H, Shen L H, Wang L Q, He Y Q. 2005. Mapping Perrier X, Jacquemoud-Collet J P. 2006. DARwin software. quantitative trait loci underlying the cooking and eating quality http://darwin cirad fr/Darwin. of rice using a DH population. Mol Breeding, 15(2): 117–124. Petrov M, Danzart M, Giampaoli P, Faure J, Richard H. 1996. Rice Vanavichit A, Tragoonrung S, Toojinda T, Wanchana S, aroma analysis: Discrimination between a scented and a non- Kamolsukyunyong W. 2008. Transgenic Rice Plants with scented rice. Sci Alim, 16(4): 347–360. Reduced Expression of Os2AP and Elevated Levels of 2-acetyl- Pooni H S, Kumar I, Khush G S. 1993. Genetical control of amylose 1-pyrroline. US Patent: 7319181. content in selected crosses of indica rice. Heredity, 70(3): 269– Wan Y X, Deng Q M, Wang S Q, Liu M W, Zhou H Q, Li P. 2007. 280. Genetic polymorphism of Wx gene and its correlation with main Prathepha P. 2003. Characterization of Waxy microsatellite classes grain quality characteristics in rice. Rice Sci, 14(2): 85–93. that are closely linked to the rice Waxy gene and amylose Wanchana S, Tragoonrung S, Toojinda T, Vanavichit A. 2003. content in Thai rice germplasm. Songklan J Sci Technol, 25(1): Duplicated coding sequence in the waxy allele of tropical 1–8. glutinous rice (Oryza sativa L.). Plant Sci, 165(6): 1193–1199. Prathepha P. 2008. Variation of the Waxy microsatellite allele and Webb B D. 1980. Rice quality and grades. In: Luh B S. Rice: its relation to amylase content in wild rice (Oryza rufipugon Production and Utilization. Westport, Connecticut: 543–565. Griff.). Asian J Plant Sci, 7(2): 156–162. Zheng X, Wu J G, Lou X Y, Xu H M, Shi C H. 2008. The QTL Shi W W, Yang Y, Chen S H, Xu M L. 2008. Discovery of a new analysis on maternal and endosperm genome and their fragrance allele and the development of functional markers for environmental interactions for characters of cooking quality in the breeding of fragrant rice varieties. Mol Breeding, 22(2): 185– rice (Oryza sativa L.). Theor Appl Genet, 116(3): 335–342. 192.