Planta (2007) 226:11–20 DOI 10.1007/s00425-006-0460-4

O RI G I NAL ART I C LE

Origin of seed shattering in rice (Oryza sativa L.) Zhongwei Lin · Megan E. GriYth · Xianran Li · Zuofeng Zhu · Lubing Tan · Yongcai Fu · Wenxu Zhang · Xiangkun Wang · Daoxin Xie · Chuanqing Sun

Received: 10 September 2006 / Accepted: 5 December 2006 / Published online: 10 January 2007 © Springer-Verlag 2006

Abstract A critical evolutionary step during rice domestication was the elimination of seed shattering. Wild rice disperses seeds freely at maturity to guarantee the propagation, while cultivated rice retains seeds on the straws to make easy harvest and decrease the loss of production. The molecular basis for this key event during rice domestication remains to be elucidated. Here we show that the seed shattering is controlled by a single dominant gene, Shattering1 (SHA1), encoding a member of the trihelix family of plantspeciWc transcription factors. SHA1 was mapped to a 5.5 kb genomic fragment, which contains a single open reading frame, using a backcrossed population between cultivated rice Teqing and an introgression Electronic supplementary material The online version of this article (doi:10.1007/s00425-006-0460-4) contains supplementary material, which is available to authorized users.

line IL105 with the seed shattering habit derived from perennial common wild rice, YJCWR. The predicted amino acid sequence of SHA1 in YJCWR and IL105 is distinguished from that in eight domesticated rice cultivars, including Teqing, by only a single amino acid substitution (K79N) caused by a single nucleotide change (g237t). Further sequence veriWcation on the g237t mutation site revealed that the g237t mutation is present in all the domesticated rice cultivars, including 92 indica and 108 japonica cultivars, but not in any of the 24 wild rice accessions examined. Our results demonstrate that the g237t mutation in SHA1 accounts for the elimination of seed shattering, and that all the domesticated rice cultivars harbor the mutant sha1 gene and therefore have lost the ability to shed their seeds at maturity. In addition, our data support the theory that the non-shattering trait selection during rice domestication occurred prior to the indica–japonica diVerentiation in rice evolutionary history.

Z. Lin · X. Li · Z. Zhu · L. Tan · Y. Fu · W. Zhang · X. Wang · C. Sun (&) Department of Plant Genetics and Breeding and State Key Laboratory of Agrobiotechnology, China Agricultural University, Key Laboratory of Crop Heterosis and Utilization of Ministry of Education, Beijing Key Laboratory of Crop Genetic Improvement, Key Laboratory of Crop Genetic Improvement and Genome of Ministry of Agriculture, Beijing 100094, China e-mail: [email protected]

Abbreviations AL Abscission layer AZ Abscission zone SSR Single sequence repeat

M. E. GriYth Institute of Molecular and Cell Biology, 61 Biopolis Drive, Singapore, Singapore 138673,

Introduction

D. Xie (&) MOE Key Laboratory of Bioinformatics, Department of Biological Sciences, Tsinghua University, Beijing 100084, China e-mail: [email protected]

Seed shattering of wild rice (O. ruWpogon GriV.) is one of the most greatly changed traits compared with cultivated rice. In wild rice, seeds disperse immediately by shattering at maturity, protecting them from

Keywords Elimination of seed shattering · Indica– japonica diVerentiation · Oryza · Shattering 1 (SHA1)

123

12

being easily eaten by small animals and guaranteeing the propagation of wild rice. In the course of rice domestication, the artiWcial selection of non-shattering habit makes easy harvest and decreases the loss of production. There is a wild range of forcible shattering in cultivated rice. Generally, seeds of indica rice can shed easier than that of japonica rice by hand after maturity. Shattering is the major cause of crop loss and remains an emphasis on selection for the loss of seed shattering habit in most rice breeding programs today. Shattering habit in rice plant has been shown to be controlled by the formation of abscission layer (AL) (Zee et al. 1979; Oba et al. 1995; Watanabe et al. 2003), which occurs at the juncture between the sterile lemma and pedicel. AL constituted by a band of small cells comes to form before heading and Wnishes cell expansion to maximum size at the stage of heading (Jin and Inouye 1981; Oba et al. 1995). The presence of AL with delay is often found in japonica varieties, even no AL can be viewed after maturity in some japonica rice. A number of transcription factors controlling abscission zones (AZs) formation have been recently described in diVerent plant systems. SHATTERPROOF (SHP1 and SHP2) MADS-box genes act together to specify the valve margin of the silique in Arabidopsis, and are essential for the ligniWcation of the valve margin cells (Lilijegren et al. 2000), further study shows that the two transcription factors FRUITFULL and REPLUMLESS negatively regulate SHP expression in valve margin, resulting in the limit of the valve margin diVerentiation at the valve edges (Ferrándiz et al. 2000; Roeder et al. 2002). Another MADS-box gene JIONTLESS plays a central role in the diVerentiation of the pedicel AZ in tomato plant (Mao et al. 2000). ALCATRAZ (ALC) encodes a myc/bHLH transcription factor and contributes in the formation of a strip of labile nonligniWed cells link with partly ligniWed valve and replum, which provide the tension for pod dehiscence (Rajani and Sundaresan 2001). The mutant alc stops the dehiscence of fruit and SHP, FUL genes normal expression in the alc mutant indicates ALC may work downstream to SHP genes or in another pathway. Several genes have been identiWed to control seed shattering in rice by classical genetic analysis and molecular analysis (Cai and Morishima 2000; Kennard et al. 2002; Thomson et al. 2003). Two dominant genes located on chromosome 1 and 4 respectively, have been speculated to be important for seed shattering in rice (Cai and Morishima 2000). During abscission of seed in rice plant, the gene of chromosome 1 (qsh1)

123

Planta (2007) 226:11–20

controls the formation of AL at the base of sterile lemma, a mutant, g to t occurred in the regulatory region located 12kb away from qsh1, leads to the absence of abscission layer in japonica rice (Konishi et al. 2006). Here we describe the isolation and detailed characterization of the dominant shattering gene, Shattering 1 (SHA1), on chromosome 4 from perennial common wild rice, YJCWR. We show that SHA1 plays an important role in activation of cell separation and encodes a member of the trihelix family of plant-speciWc transcription factors (Nagano 2000; Murata et al. 2002), a mutant, g to t in the trihelix domain, results in the elimination of seed shattering in cultivated rice.

Materials and methods Construction of introgression lines and genetic mapping of SHA1 Teqing, a high-yielding commercial indica rice cultivar of China, was used as the recipient to cross with YJCWR, an accession of perennial common wild rice collected from Yuanjiang County, Yunnan Province, China. The F1 plants were self-pollinated to generate F2 population. A total of 383 plants from the F2 population were examined for seed shattering phenotype. To generate shattering introgression lines, the shattering plants from the above F2 population were used as the donor to backcross with Teqing, which generated backcrossed-one (BC1) plants. The BC1 plants with the seed shattering trait were consecutively backcrossed with Teqing until the BC3 shattering plants were obtained. The BC3 plants were self-pollinated for two generations to produce the BC3F3 introgression lines, among which two lines (IL102 and IL105) were used for further study. IL105 was crossed with Teqing to generate F1 plants, which were self-pollinated to produce F2 population of 13,012 plants. Of the 13,012 F2 plants 9,876 exhibited seed shattering phenotype; Among the 9,876 shattering F2 plants, 376 were used in the preliminary mapping to place SHA1 on Chromosome 4 between markers SP33 (5⬘-cttcagcttggacaggcgat-3⬘; 5⬘-gaaaagagatgtctggacg-3⬘) and RM1113 (5⬘-gggcgcatg-tgtatttcttc-3⬘; 5⬘-tggggaaaa accacaagcc-3⬘). Twenty-two recombinant plants were identiWed by these two makers, their F3 progeny were further examined on segregation of the seed shattering phenotype to determine the genotype at the SHA1 locus. Of the 13,012 F2 plants 3,136 clearly showed nonshattering phenotype and were used in Wne mapping of

Planta (2007) 226:11–20

the sha1 locus with the PCR-based polymorphic markers including RM1113, RM6441 (5⬘-cgaaaggtcggc atctttc-3⬘; 5⬘-atggcaatgatacggaggag-3⬘), 19E (5⬘-gtcgtcc tcttctcttcatc-3⬘; 5⬘-gcgaaacgtgttattgctgc-3⬘), 19K (5⬘-cat tgcacgtaacc-agatcc-3⬘; 5⬘-ccggtatggtcatatggatg-3⬘), 19Q (5⬘-cttgagggatacaaccttac-3⬘; 5⬘-agacgtg-aatgtacggtaac-3⬘), PS01 (5⬘-cacccgagaagaaaaaacgc-3⬘; 5⬘-gtgtgtggtgtgaatg atcg-3⬘) and PS02 (5⬘-aagtgcagccatggctttcc-3⬘; 5⬘-cccagg ttgggtcttatctg-3⬘). Evaluation of seed shattering rate The panicles of each rice plant were bagged at the stage of heading. Shattered seeds and non-shattered ones were gathered at maturity. The shattering rate is expressed by a percentage of shattered seed to the total seed weight (Gu et al. 2005). Plant materials A total of 233 accessions includes 25 accessions of O. ruWpogon, 96 indica varieties, 112 japonica varieties, were used to examine the mutant (g237t) in this study (Tables S1–S3). O. sativa is diVerentiated into two subspecies, indica and japonica, as a result of isolation and selection. DNA sequencing Genomic regions (5.5 kb) containing SHA1 from YJCWR and IL105 were ampliWed and sequenced with six pairs of primers including P5500A (5⬘-atcgatgagtg tgctggttg-3⬘, 5⬘-gatctgcagactgtgagatg-3⬘), P5500B (5⬘-aa ggtgcctatgattctggg-3⬘, 5⬘-ctgggcctgttggtttt-ttc-3⬘), P5500C (5⬘-catatggagtggacctaagcctag-3⬘, 5⬘-atgcaaatgcaaagcgtc cgctag-3⬘), P5500D (5⬘-ctagcggacgctttgcatttgcat-3⬘, 5⬘cgtgtgtttccacccgttttttccctcgtt-3⬘), P5500E (5⬘-aacgagg gaaaaaacgggtggaaacacacg-3⬘, 5⬘-gcaatgtaacgcattacgcat gcttgcagc-3⬘) and P5500F (5⬘-gagagggtggtcgcggtaat aataatgtcc-3⬘, 5⬘-gctactcctcttcttgcagccattccaaac-3⬘). Two pairs of primers, p5gt (5⬘-ccttgtaaatacgggcgacg-3⬘) and p3gt (5⬘-agtc- gcggaccttcttgtag-3⬘), pr5 (5⬘-cagaa ccagtgcaatgacaagtgggacaac-3⬘) and pr3 (5⬘-ataatcc-gatg cctcgatccatgcatctcc-3⬘) were used to amplify and sequence the full length cDNAs from YJCWR and IL105 and from the non-shattering domesticated rice cultivars including four indica cultivars (Teqing, 9311, Guangluai4, Guichao2) and four japonica cultivars (Nipponbare, Zhonghua15, Sewon392 and JX17). Total RNAs were prepared from the rice Xowers using RNeasy Plant-Mini Kit (Qiagen). First-strand cDNA synthesis was performed with BcaBESTTM RNA PCR Kit (DRR023A, TaKaRa).

13

Microscopy Pedicels (containing abscission layer) from IL105 and Teqing plants at the indicated stages were Wxed in FAA (70% ethanol, 5% glacial acetic acid, 3.7% formaldehyde), embedded in LR white acrylic resin (SigmaAldrich; L9774-100G). Thin sections were prepared and stained with 1% toluidine blue. For scanning electron microscopy (Liu et al. 2004), the lemma base of IL105 seeds that naturally fall to ground at 14th day after pollination was viewed.

Results SHA1 doesn’t govern the development of AL In screening for the plant recovers seed shattering in cultivated rice, we performed crosses between YJCWR, a shattering perennial common wild rice in China, and the non-shattering cultivated rice Teqing (O. sativa ssp. indica). An advanced backcrossed line, IL105 (BC3F3), was obtained, which exhibits similar morphological phenotypes with Teqing except for the high rate of seed shattering (Fig. 1a, b). To accurately describe this phenotype of SHA1, we quantitatively compared the shattering rates between IL105 and Teqing at maturity (Fig. 1c). The shattering rate of Teqing plant (0–2.3%) decreased more than 40 times than that of IL105 plants (92–100%), So largely decreasing shattering rate of Teqing plant suggested that the structure of AL may be changed in the sha1 mutant plant. We used scanning electron microscopy to examine the base of lemma of the IL105 seeds, which naturally fall to ground from 14th day after pollination, and found that seed shattering was caused by the complete separation of the AL, with no evidence of mechanical sheering (Fig. 2a–c); observation of AL at a histological level indicates no signiWcant variation in AL development, including the formation of the AL and cell size or cell layer number, between IL105 and Teqing from two days before pollination to 12 days after pollination (Fig. 2d–g; and data not shown). These data imply that, unlike JOINTLESS in tomato and SHP1/SHP2 in Arabidopsis and qsh1 in rice, SHA1 in rice is not involved in regulation of AL development. SHA1 gene may work for the degradation of the cell wall of the AL. High-resolution mapping of SHA1 In an attempt to identify the seed shattering gene, we generated a population of 383 F2 plants from a cross between Teqing as the recipient and YJCWR as the

123

14

Planta (2007) 226:11–20

Fig. 1 a Phenotypes of the introgression line IL105 and the cultivated rice Teqing plants. b Seeds were shed from the panicle of IL105 but retained on Teqing at maturity. c The shattering rate in the introgression lines IL102 and IL105, Teqing, common wild rice YJCWR and TeqingYJCWR F1 hybrid plants (F1) is expressed as a percentage of shattered seeds to the total seed weight. The experiment was repeated four times. Error bars represent SD (n > 15). d, e Graphical genotypes show that three YJCWR chromosomal segments (black bar) were introgressed into the IL102 genome (white bar) on chromosomes (Chr.) 1, 4 and 12 (d), and into the IL105 genome (white bar) on chromosomes 3, 4 and 11 (e). f Graphical genotypes of chromosome 4 of IL102 and IL105 with 10 SSR markers (Tan et al. 2004). Both IL102 and IL105 share a common YJCWR segment (black bar) detected by RM348 (5⬘ccgctactaatagcagagag-3⬘; 5⬘-ggagctttgttcttgcgaac-3⬘), RM131 (5⬘-ggagcagcttctcg-agcatgg-3⬘; 5⬘-ccaaatctcgcctcgtttagcc-3⬘) at 113.2 and 123.8 cm, respectively

donor. All of the F1 progeny and 255 out of the 383 F2 plants exhibited the seed shattering phenotype similar to YJCWR (Fig. 1c, and data not shown), indicating that this shattering trait is controlled by a single dominant gene, referred to as Shattering1 (SHA1). We further backcrossed two shattering plants three times consecutively with Teqing to produce the introgression lines, designated IL102 and IL105, which exhibited similar morphological phenotypes with Teqing except for the high rate of seed shattering (Fig. 1a–c). We then used 117 simple-sequence-repeat (SSR)

123

markers with polymorphisms between YJCWR and Teqing, which distribute evenly throughout 12 chromosomes of rice genome (Tan et al. 2004), to scan genomes of IL102 and IL105 for chromosomal segments introgressed from YJCWR. YJCWR chromosomal segments were detected in the genome of IL102 on chromosomes 1, 4 and 12 (Fig. 1d), and in IL105 on chromosomes 2, 4 and 11 (Fig. 1e). Both IL102 and IL105 shared one common YJCWR segment detected by SSR markers RM348 and RM131 near the southern end of chromosome 4 (Fig. 1f), indicating that this

Planta (2007) 226:11–20

15

Sequence analysis of the 5.5 Kb region from IL105 and YJCWR revealed only a single open reading frame that corresponded to the SHA1 gene (Fig. 3b). The SHA1 cDNA was isolated from YJCWR and its sequence predicted a 1173-nucleotide transcript encoding 390-amino acid protein. Protein sequence analysis shows that SHA1 contains a DNA binding domain belonging to the trihelix family of plant transcription regulators, and a proline-rich region. The trihelix motif shares many features with Myb DNA-binding domains (Nagano 2000; Murata et al. 2002) (Fig. 4a, b). Sequence comparison of SHA1 gene

Fig. 2 a–g Scanning electron micrographs (a, b) of the base of lemma of the IL105 seeds (c), which naturally fall to ground at 14th day after pollination. d–g Histological longitudinal sections of the lemma base of Xowers collected at the day of pollination (0 DAP) and 12 days after pollination (12 DAP). AL Abscission layer; d and e, IL105. f and g, Teqing. Scale bars 100
m

common YJCWR segment probably harbored the SHA1 locus. A population of 13,012 F2 plants was generated from a cross between IL105 (SHA1/SHA1) and Teqing (sha1/sha1) for genetic mapping of SHA1. Tests on 376 shattering plants for genetic linkage between the seed shattering phenotype and SSR markers (Tan et al. 2004) placed SHA1 on chromosome 4 Xanked by SP33 and RM1113 within the region detected by RM348 and RM131 (Fig. 3a). A total of 3,136 non-shattering plants were further used to map the sha1 locus, two markers RM6441 and RM1113 Xanking sha1 identiWed 76 and 27 recombinant plants, respectively (Fig. 3b). DNA sequences of two BACs (http://www.tigr.org/) in this region were used to develop polymorphic markers including 19E, 19K, 19Q, PS01 and PS02 (Fig. 3b). Fine mapping with these markers in the 103 recombinant plants identiWed by RM6441 and RM1113 placed sha1 within 5.5 kb region, between PS01 and PS02 markers.

Sequencing of sha1 cDNAs from domesticated rice cultivars, including four indica and four japonica cultivars, revealed that sha1 cDNA is identical in these indica and japonica cultivars, and yet deviated from SHA1 of YJCWR by a single nucleotide change, g to t, at position +237 relative to the translation start site (g237t), which causes a single amino acid substitution from Lysine residue to Asparagine at position 79 (K79N) (Fig. 4a, b). These data suggest that loss of seed shattering is caused by the g237t mutation leading to the K79N substitution in SHA1. Support for the conclusion that the g237t mutation is responsible for the non-shattering phenotype was further derived from large scale sequence veriWcation on the g237t region from 24 shattering wild rice accessions and 200 non-shattering domesticated rice cultivars. Among these wild rice accessions, O. alata was originally collected from Brazil, O. latifolia from Costa Rica, O. longistaminata from Cameroon, three alleles of O. ruWpogon from India and 18 alleles of O. ruWpogon from various areas in China. These wild rice accessions that represent diverse genetic origins, in particular O. alata and O. latifolia that have tetraploid genomes, all contain SHA1 with the “g” at the g237t site and with the Lysine (K) residue at position 79 in its predicted amino acid sequence (Fig. 4a; Table S1). Among the domesticated rice cultivars, indica and japonica cultivars exhibit many distinguishing morphological and physiological characters (Matsuo et al. 1997); 92 indica and 108 japonica cultivars with distinct backgrounds, which were collected from all the geographic range of O. sativa in South, Southeast, and East Asia, all contain the mutant sha1 gene with the “t” at the g237t site and with the Asparagine (N) residue at position 79 of the predicted amino acid sequence (Fig. 4a; Table S2, S3). These sequence veriWcation data clearly imply that the Lysine residue (K) at position 79 is critical for the SHA1 function.

123

16

Planta (2007) 226:11–20

a

r=18

Marker

r=4

RM348 SP33

SHA1 RM1113 RM131

Chr.4

Phenotype in F3 progeny

r=9/376 r=5/376 r=4/376

Shattering Segregated Segregated

r=1/376

Shattering

r=2/376

Segregated

r=1/376

Segregated

b Marker

r=76

r=30 r=16 r=8

r=4

r=3

r=27

SP33 RM6441 19E 19K 19Q PS01 sha1 PS02

RM1113

Chr.4 BAC

OSJNBB0003A17

OSJNBa0043A12

r=4/3136 r=3/3136 OSJNBa0043A12.36

sha1

OSJNBa0043A12.38

Fig. 3 a Genetic mapping of the SHA1 locus on chromosome 4 within 376 shattering plants from the F2 population of a cross between IL105 and Teqing. The 22 recombinants with chromosomal recombination between SSR markers SP33 and RM1113 are classiWed into six groups, their genotypes are schematically shown together with their phenotypes in their F3 progeny (black bar, homozygous IL105; grey bar, heterozygous; white bar, homozygous Teqing; shattering, homozygous for the shattering phenotype in their F3 progeny; segregated, heterozygous for shattering phenotype in their F3 progeny). Numbers of recombinants (r) for markers or in each recombinant group are indicated at the top of

the marker or on the left side of graphic genotype of each group. b Fine mapping of the sha1 locus within 3,136 non-shattering F2 plants. The sha1 locus was placed within 5.5 kb region Xanked by two markers PS01 and PS02. The genotypes of recombinant plants identiWed by PS01 and PS02 are schematically shown (grey bar, heterozygous; white bar, homozygous Teqing). The DNA sequences of BACs were determined by the rice genome-sequencing project. Arrows indicate position and orientation of the sha1 gene (OSJNBa0043A12.37, LOC_Os04g57530.1) and two adjacent genes (OSJNBa0043A12.36 and OSJNBa0043A12.38)

Phylogenetic analysis

may alter the DNA binding or sequence recognition properties of the protein, aVect association of the protein with promoter elements of its target genes, and therefore result in loss of seed shattering.

Alignment analysis of 53 transcription factors of the trihelix family from various plant species including tobacco, rice, pea, soybean and Arabidopsis shows that each phylogenetic group has a distinctive amino acid at the position in the trihelix domain corresponding to 79 in SHA1 (Fig. 5a). TwentyWve transcription factors including SHA1 fall into the single largest group; the amino acid at this position is completely conserved by the positive charged residue K or R (Arginine) (Fig. 5). Sequence alignment of these 25 members among their helix–loop–helix–loop–helix (trihelix) DNA binding domain shows that K or R corresponds to the 13th position in the conserved region of the Wrst helix, the 70% consensus being WxxxExxxL(I/L)xx(R/K)xxx(D/ E) (Nagano 2000; Murata et al. 2002) (Fig. 5b). This K or R residue may play an essential role in DNA binding or sequence recognition in this group of plant transcription factors. In sha1, this K residue is substituted by the neutral amino acid Asparagine (K79N), which

123

Discussion Simple change in a single gene can cause a dramatic phenotype change during rice domestication A set of archaeological evidences shows crop domestication began about 10,000 years ago, and the morphological diVerences between today’s crop and their progenitor are so vast. Crop genomic projects have revealed that crops own a large number of genes, exempliWed by 37,544 non-transposable-element protein-coding sequences predicted by the map-based sequence of the rice genome. Then it is full of controversy that artiWcial selection can turn progenitors into modern domestic crops during the few thousand years.

Planta (2007) 226:11–20

17

a p5gt

+1

+237

pr5 p3gt

*

pr3

5’

3’

SHA1

acc gcc aag cgt ctg T

sha1

A

1

R

L

acc gcc aat cgt ctg T

b

K

A

N

R

L

at gt cgggct cct ct gccgacccct cgccat ccgcct cgaccgcgggggcggcggt ct cg S G S S A D P S P S A S T A G A A V S H

D

Y

G

H

G

H

L

T

A

T

P

P

P

R

A

T

G

P

A

P

P

P

P

S

P

A

S

G

S

A

R

K

G

N

W

T

L

H

E

T

L

I

L

I

T

A

K R N ct ggacgacgaccgccgcgccggcgt t gggggcgcggcggct ggt ggcggcggcgccggg L D D D R R A G V G G A A A G G G G A G

sha1

241

S

P

P

T

P

R

S

A

E

Q

R

W

K

W

V

E

N

Y

C

W

K

N

G

C

L

R

S

Q

N

Q

C

N

D

K

W

D

N

L

L

R

T

M

A

P

G

A

T

I

G ggagcggcacgagcgc E R H E R

T

P

P

P

P

P

L

A

L

P

L

P

P

P

P

P

P

S

P

P

K

P

L

V

A

Q

Q

Q

H

H

H

H

G

H

H

H

H

P

P

P

P

Q

P

P

S

S

L

Q

L

P

P

A

V

V

A

P

P

P

A

S

V

S

A

P gag E

S

G

S

S

E

S

G

E

E

E

E

G

S

G

G

E

P

S

S

V

V

R

S

A

T

V

E

E

R

T

E

V

R

R

Q

G

S

S

A

I

H

A

L

V

S

L

c

(aat)

Flag leaf

E

E

R

L

I

Culm

R

L

R

L

E

A

A

V

N

S

L

D

S

S

G

R

*

Root

Pedicel

IL105 Teqing IL105 Teqing IL105 Teqing IL105 Teqing IL105 Teqing

1 DBP Fig. 4 a Diagram of the SHA1 gene and the g237t mutation in sha1. The p5gt and p3gt are primers used for large scale sequence veriWcation on the single nucleotide change (g to t) at position +237, all the 25 shattering wild rice alleles including YJCWR contain SHA1 with “g” at position +237, whereas all the 208 nonshattering cultivated rice alleles including Teqing contain sha1 with the g237t mutation (Table S1–S3). Open bars, exons. Thick line, intron. Thin line, non-translation sequences at 5⬘ or 3⬘ end. +1, transcription start site. *, stop codon. Pink area, Trihelix DNA binding domain. Blue area, Proline rich region. b Predicted amino acid sequence of SHA1 from the common wild rice YJCWR. The sha1 cDNA sequences from eight domesticated rice cultivars deviated from that of YJCWR by a single nucleotide change (g to

5 DAP

t) at position +237 (marked in red). Two pairs of primers, including p5gt and p3gt, pr5 and pr3 (a), were used to amplify and sequence the full length cDNA. Pink letter, Trihelix DNA-binding domain. Blue letter, proline-rich region. c Expression of the SHA1 transcript. Total RNA, from the Xag leaf, culm and root tissues, and from the pedicel tissues at the developmental stages of one day before pollination (1 DBP) or 5 days after pollination (5 DAP), was used in RT-PCR to examine the expression of the SHA1 or sha1 transcripts ampliWed by pr5 and pr3 primers (upper panel). The ACTIN transcript was ampliWed by the acitin primers (5⬘-gtgtgtgacaatggaactgg-3⬘; 5⬘-gtgatctccttgctcatacg-3⬘) as a control (low panel)

123

18

Planta (2007) 226:11–20

Fig. 5 a Phylogenetic tree estimation derived from the alignment of entire proteins using the multiple sequence comparison by log-expectation (second iteration) (http:// www.ebi.ac.uk/muscle/; Edgar 2004). The locus name for each protein is preceded with the amino acid corresponding to position K79 of SHA1. The four groups have uniquely conserved amino acids at this position: group I with R or K, group II with W or Y, group III with R, and group IV with V or I. b Multiple alignment of the trihelix DNA binding domains belonging to group I. The secondary structure prediction of the three helices of SHA1 is depicted (top), based on the DSSP EBGHSTL (logo format; Karchin et al. 2003). C coil, H helix, T turn/ loop. The position K79 of SHA1 is depicted by the arrowhead. 70% consensus amino acid residues are highlighted. Basic (RKH) in green, acidic (DE) in blue, aromatic amino acids (WFY) in red, small hydrophobic are purple (AVPMILG) and neutral, polar (STCNQ) residues are grey. The logo (bottom) for the consensus sequence (http://www.weblogo.berkeley.edu/logo.cgi) graphically depicts the sequence conservation at that position (by overall height) and the height of symbols within the stack indicates the relative frequency of each amino acid at that position (Crooks et al. 2004)

But single-nucleotide polymorphism analysis in 774 maize genes showed only a bit of these gene (2–4%) experienced artiWcial selection (Wright et al. 2005), and the mutation within the sd-1 gene responsible for green revolution can change many aspects of plant growth, indicating that artiWcial selection on a small number of

123

genes with mutations of large genetic eVect can carry out this conversion. Our results showed that Lysine residue at position 79 (K79) in SHA1 is critical for shattering. Simultaneously, Li et al. (2006) cloned SH4, from the annual common wild rice O. nivara, which is allelic to SHA1 and also harbor the K79 residue. The

Planta (2007) 226:11–20

residue V at the position 152 in SHA1 is changed to residue A in SH4 and six residues TGGAAA from the position 158 to 163 in SHA1 are deleted in SH4, however, both A152 residue and the TGGAAA deletion in SH4 are not essential for seed shattering, only the K79 residue was found to be crucial for SH4 function in shattering (Li et al. 2006). These data suggest that a simple change in a single gene can cause a dramatic phenotype change during rice domestication; such a transformation was also found in the maize tga1 gene where a single amino acid substitution can caused the liberation of the kernel from the protective casing (Wang et al. 2005), which was the most critical step in maize domestication. The common wild rice was Wrst domesticated before indica-japonica diVerentiation The origin of the indica-japonica diVerentiation of Asian cultivated rice has been proposed in three hypotheses (Oka 1988): The Wrst hypothesis was proposed by Ting (1957, 1961). It assumed that keng (japonica) rice was diVerentiated from hsien (indica) rice, which was developed from the common wild rice in south China; the second by Wang et al. (1984) suggested that the common wild rice domesticated in upland Welds formed keng rice and that in marshy low lands formed hsien rice. The third, a dual origin of the cultivated rice, proposed by Chou (1948) and Second (1982), assumed that the indica-japonica diVerentiation had preexisted in the common wild rice and indica-like type and japonica-like type common wild rice developed into indica and japonica respectively. The origins of the indica and japonica have been studied for decades. Whether common wild rice had diVerentiated into the indica and japonica type like cultivars before it evolved to cultivated rice is still a matter for debate. We found that the sha1 cDNA sequences from four typical indica and four japonica cultivars are identical, and that all 96 indica and 112 japonica cultivars all contain the same g237t mutation. This molecular evidence indicates that the non-shattering trait selection during rice domestication occurred prior to the indica–japonica diVerentiation in rice evolutionary history, in support of the hypothesis that common wild rice was Wrst cultivated before the indica–japonica diVerentiation.

19

pedicel AZ development (Mao et al. 2000). In Arabidopsis, two MADS-box genes SHP1 and SHP2 are essential for seed release from siliques, and both genes are functionally redundant in the control of dehiscence zone diVerentiation and promotion of the ligniWcation of adjacent cells (Liljegren et al. 2000). In rice, seed shattering occurs in the AL where a layer of specialized cells diVerentiate at the base of the lemma (Jin and Inouye 1981; Oba et al. 1995; Konishi et al. 2006). No signiWcant variation in AL development (cell number and cell size) was found between IL105 (SHA1) and Teqing (sha1), the AL has fully developed in both IL105 and Teqing before pollination at 1 DBP (one day before pollination) and 0 DBP (Fig. 2d–g, and data not shown). The SHA1 transcript in pedicel where the AL develops, before pollination (such as 1 DBP), was detected at a very low level, similar to that in Xag leaf, culm and root; whereas the SHA1 transcript was found to highly accumulate in pedicel after pollination (such as 5 DAP) (Fig. 4c, and data not shown). These data indicate that SHA1 is not involved in AL development. Li et al. (2006) recently demonstrated that SH4, an allelic gene of SHA1 (98% amino acid identity), is involved in AL development. It would be interesting to verify whether the A152 residue and the TGGAAA deletion, which distinguish from SHA1, in SH4 could be responsible for its role in AL development. We speculated that SHA1 may activate the important cell separation process within AL in response to pollination. Mutated sha1, with the N residue at position 79, may fail to regulate its target genes essential for this important cell separation process and therefore fail to initiate seed shattering in domesticated rice cultivars. Shedding of leaves, Xowers, seeds or fruits, resulting from cell separation within AZs in response to signals from development, tissue damage and stress, is a general adoptive strategy for plants (Roberts et al. 2002). Further mechanistic understanding of SHA1 action will provide new insights into the molecular basis of this adoptive strategy in plant kingdom. Acknowledgments We thank the International Rice Research Institute and Chinese Rice Research Institute for providing the common wild rice and cultivated rice samples. This work was supported by grants from the Conservation and Utilization of AgroWild Plants of the Ministry of Agriculture of China, the National Basic Research Program of China (2005CB120801).

References SHA1 may activate the important cell separation process within AL In tomato, shattering of the Xower or fruit is controlled by the MADS-box gene JOINTLESS that regulates the

Cai HW, Morishima H (2000) Genomic regions aVecting seed shattering and seed dormancy in rice. Theor Appl Genet 100:840–846 Chou SL (1948) China is the place of origin of rice. J Rice Soc China 7:53–54

123

20 Crooks GE, Hon G, Chandonia JM, Brenner SE (2004) Weblogo: a sequence logo generator. Genome Res 14:1188–1190 Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32:1792–1797 Ferrándiz C, Liljegren SJ, Yanofsky MF (2000) Negative regulation of the SHATTERPROOF genes by FRUITFULL during Arabidopsis fruit development. Science 289:436– 438 Gu XY, Kianian SF, Hareland GA, HoVer BL, Foley ME (2005) Genetic analysis of adaptive syndromes interrelated with seed dormancy in weedy rice (Oryza sativa). Theor Appl Genet 110:1108–1118 Jin I, Inouye J (1981) On the degree of grain shedding of Japonica–Indica hybrid rice bred in Korea. Jpn J Crop Sci 50:181– 185 Karchin R, Cline M, Mandel-Gutfreund Y, Karplus K (2003) Hidden Markov. Models that use predicted local structure for fold recognition: alphabets of backbone geometry. Proteins Struct Funct Genet 51:504–514 Kennard WC, Phillips RL, Porter RA (2002) Genetic dissection of seed shattering, agronomic, and color traits in American wild rice (Zizania palustris var. interior L.) with a comparative map. Theor Appl Genet 105:1075–1086 Konishi S, Izawa T, Lin SY, Ebana K, Fukuta Y, Sasaki T, Yano M (2006) An SNP caused loss of seed shattering during rice domestication. Science 312:1392–1396 Li CB, Zhou AL, Sang T (2006) Rice domestication by reducing shattering. Science 311:1936–1939 Liljegren SJ, Ditta GS, Eshed Y, Savidge B, Bowman JL, Yanofsky MF (2000) SHATTERPROOF MADS-box genes control seed dispersal in Arabidopsis. Nature 404:766–780 Liu F, Ni WM, GriYth ME, Huang ZY, Chang CQ, Peng W, Ma H, Xie DX (2004) The ASK1 and ASK2 genes are essential for Arabidopsis early development. Plant Cell 16:5–20 Mao L, Begum D, Chuang HW, Budiman MA, Szymkowiak EJ, Irish EE, Wing RA (2000) JOINTLESS is a MADS-box gene controlling tomato Xower abscission zone development. Nature 406:910–913 Matsuo T, Futsuhara Y, Kikuchi F, Yamaguchi H (1997) Science of the rice plant. Food and Agriculture Policy Research Center, Tokyo, pp 145–151 Murata J, Takase H, Hiratsuka K (2002) Characterization of a novel GT-box binding protein from Arabidopsis. Plant Biotechnol 19:103–112 Nagano Y (2000) Several features of the GT-factor trihelix domain resemble those of the Myb DNA-binding domain. Plant Physiol 124:491–493

123

Planta (2007) 226:11–20 Oba S, Sumi N, Fujimoto F, Yasue T (1995) Association between grain shattering habit and formation of abscission layer controlled by grain shattering gene sh-2 in rice (Oryza sativa L.). Jpn J Crop Sci 64:607–615 Oka HI (1988) Origin of cultivated rice. Jpn Sci Soc Press, Amsterdam Rajani S, Sundaresan V (2001) The Arabidopsis myc/bHLH gene ALCATRAZ enables cell separation in fruit dehiscence. Curr Biol 11:1914–1922 Roberts JA, Elliott KA, González-Carranza ZH (2002) Abscission, dehiscence, and other cell separation processes. Annu Rev Plant Biol 53:131–158 Roeder AHK, Ferrándiz C, Yanofsky MF (2002) The role of the REPLUMLESS homeodomain protein in patterning the arabidopsis fruit. Curr Biol 13:1630–1635 Second G (1982) Origin of the genetic diversity of cultivated rice (Oryza spp.), study of the polymorphism scored at 40 isozyme loci. Jpn J Genet 57:25–57 Tan LB, Zhang PJ, Fu YC, Liu FX, Wang XK, Sun CQ (2004) IdentiWcation of quantitative trait loci controlling plant height and days to heading from Yuanjiang common wild rice (Oryza ruWpogon GriV.). Acta Genet Sin 31:1123–1128 Thomson MJ, Tai TH, McClum AM, Lai XH, Hinga ME, Lobos KB, Xu Y, Martinez CP, McCouch SR (2003) Mapping quantitative trait loci for yield, yield components and morphological traits in an advanced backcross population between Oryza ruWpogon and Oryza sativa cultiar JeVerson. Theor Appl Genet 107:479–493 Ting Y (1957) The origin and evolution of cultivated rice in China. Acta Agron Sin 8:243–260 Ting Y (1961) Rice crop science in China. China Agri Press, Beijing Wang H, Nussbaum-Wagler T, Li B, Zhao Q, Vigouroux Y, Faller M, Bomblies K, Lukens L, Doebley JF (2005) The origin of the naked grains of maize. Nature 436:714–719 Wang XK, Cheng KS, Lu YX, Luo J, Huang LW, Liu GR (1984) Coordinated studies on genetic resources of rice in Yuannan. III. Glabrous hull rice (guangkedao) in Yunnan. Res Bull Beijing Agri Univ 10:333–344 Watanabe K, Oba S, Horiuchi T (2003) Allelic test of rice shattering genes sh1 and sh2 in an F sub(2) population derived from the cross between Momigaredatsu and Dee-Geo-Woo-Gen (Oryza sativa L.) SABRAO. J Breed Genet 35:57–64 Wright SI, Bi IV, Schroeder SG, Yamaski M, Doebley JF, Mcmullen MD, Gaut BS (2005) The eVects of artiWcial selection on the maize genome. Science 308:1310–1313 Zee SY, Vergara BS, Chu TM (1979) Abscission layer in the rice pedicel. Int Rice Res Notes 4:5–6