he evolution us weedy 1 - arts & sciences pages · in the evolution of weedy rice in the united...

1737

American Journal of Botany 101 ( 10 ): 1737 – 1747 , 2014 .

American Journal of Botany 101 ( 10 ): 1737 – 1747 , 2014 ; http://www.amjbot.org/ © 2014 Botanical Society of America

Flowering time variation in plants is important for local ad-aptation. The timing of fl owering initiation has a large impact on mating success and on seed production and viability in any given environment ( Barrett, 1998 ; Roux et al., 2006 ; Elzinga et al., 2007 ). Multiple environmental variables such as tempera-ture and day length can act as cues for fl owering time in differ-ent species ( Weber and Schmid, 1998 ; Franks et al., 2007 ). In crop plants, fl owering time has often been further modifi ed by artifi cial selection. For example, crops have often undergone selection for uniformity in fl owering time to facilitate harvest-ing ( Jung and Müller, 2009 ). Since humans have spread crops to locations outside their native range, selection to reduce or elim-inate photoperiod sensitivity has also been common (reviewed

by Olsen and Wendel, 2013 ). Flowering time is thus a trait that has been selected upon in multiple types of plant species and to different degrees and directions.

Agricultural weeds are plants that invade and persist in the ag-ricultural environment, but are not intentionally selected on by humans ( Warwick and Stewart, 2005 ). Agricultural weeds are a leading cause of crop losses, resulting in a 10% reduction in crop productivity worldwide ( Oerke, 2006 ). Flowering time is thought to play an important role in the ability of agricultural weeds to compete with crops in the fi eld ( Ellstrand et al., 2010 ). Some weeds may benefi t from fl owering simultaneously with their local crop because this decreases conspicuousness and seed may be unwittingly collected and replanted. Weeds may also benefi t from earlier fl owering and seed dispersal before crop harvest, thereby escaping into the seed bank. For conspecifi c weeds (weeds related to the crop they invade), many species differ in fl owering phenotype from their cultivated relatives ( Ellstrand et al., 2010 ). Reproductive isolation through differences in fl ow-ering time may act as a temporal, prezygotic mating barrier; thus, in conspecifi c weedy systems, fl owering time can substantially infl uence the amount of gene fl ow that occurs between weeds and their crop relatives and can be a potentially important factor in limiting the spread of crop traits into noncrop species.

One of the most devastating conspecifi c weeds in the United States is weedy red rice ( Oryza sativa L.; Poaceae), which invades

1 Manuscript received 4 April 2014; revision accepted 21 July 2014. The authors thank D. Gealy and S. R. McCouch for seed stocks used in

this study. This project was funded in part by U.S. National Science Foundation Plant Genome Research Program grants IOS-1032023 to A.L.C., K.M.O., and Y.J. and DBI-0638820 to. K.M.O., A.L.C., and Y.J. The USDA is an equal opportunity provider and employer.

7 Present address: School of Science and Math, Abraham Baldwin Agri-cultural College, Tifton, Georgia, 31793 USA.

8 Author for correspondence (e-mail: [email protected]), phone (413) 545-0975

doi:10.3732/ajb.1400154

SPECIAL INVITED PAPER

THE EVOLUTION OF FLOWERING STRATEGIES IN US WEEDY RICE 1

CARRIE S. THURBER 2,7 , MICHAEL REAGON 3 , KENNETH M. OLSEN 4 , YULIN JIA 5 , AND ANA L. CAICEDO 6,8

2 Biology Department, University of Massachusetts, Amherst, Massachusetts 01003 USA; 3 Department of Biology, The Ohio State University–Lima, Lima, Ohio 45804 USA; 4 Department of Biology, Washington University, St. Louis, Missouri 63130

USA; 5 USDA-ARS Dale Bumpers Rice Research Center, Stuttgart, Arkansas 72160 USA; and 6 Biology Department, University of Massachusetts, Amherst, Massachusetts 01003 USA

• Premise of the study: Local adaptation in plants often involves changes in fl owering time in response to day length and tem-perature. Many crops have been selected for uniformity in fl owering time. In contrast, variable fl owering may be important for increased competitiveness in weed species invading the agricultural environment. Given the shared species designation of cultivated rice ( Oryza sativa ) and its the invasive conspecifi c weed, weedy rice, we assessed the extent to which fl owering time differed between these groups. We further assessed whether genes affecting fl owering time variation in rice could play a role in the evolution of weedy rice in the United States.

• Methods: We quantifi ed fl owering time under day-neutral conditions in weedy, cultivated, and wild Oryza groups. We also sequenced two candidate gene regions: Hd1 , a locus involved in promotion of fl owering under short days, and the promoter of Hd3a , a locus encoding the mobile signal that induces fl owering.

• Key results: We found that fl owering time has diverged between two distinct weedy rice groups, such that straw-hull weeds tend to fl ower earlier and black-hull awned weeds tend to fl ower later than cultivated rice. These differences are consistent with weed Hd1 alleles. At both loci, weeds share haplotypes with their cultivated progenitors, despite signifi cantly different fl owering times.

• Conclusions: Our phenotypic data indicate the existence of multiple fl owering strategies in weedy rice. Flowering differences between weeds and ancestors suggest this trait has evolved rapidly. From a weed management standpoint, there is the potential for overlap in fl owering of black-hull awned weeds and crops in the United States, permitting hybridization and the potential escape of genes from crops.

Key words : adaptation; gene fl ow; Oryza sativa ; photoperiodism; Poaceae; prezygotic mating barriers; weed evolution.

1738 AMERICAN JOURNAL OF BOTANY [Vol. 101

whose protein moves to the shoot apical meristem, initiating the changeover from vegetative to reproductive growth ( Kojima et al., 2002 ; Tamaki et al., 2007 ; Yang et al., 2007 ). Under LD, several MADS-box transcription factors act on Ehd1 to pro-mote fl owering through activation of Hd3a and RFT1 , a close paralog of Hd3a ( Tsuji et al., 2011 ). However, concurrent with this, Hd1 acts negatively on Hd3a to repress fl owering. Flower-ing under LD conditions is complex and the cut-off between LD and SD photoperiod is unclear. However, the critical thresh-old for Hd3a expression is about 13–13.5 h of light ( Itoh et al., 2010 ). Importantly, a recent study ( Takahashi et al., 2009 ) has shown that the three major determinants of fl owering time di-versity in cultivated rice are polymorphisms in the Hd1 coding sequence that affect protein function, polymorphisms in the Hd3a promoter sequence that affect the expression of this locus, and expression levels of Ehd1 . Additional research also sug-gests that Hd1 was a target of selection during domestication, leading to selective sweeps in cultivated rice for multiple loss-of-function alleles that decrease or eliminate photoperiod sensi-tivity ( Takahashi and Shimamoto, 2011 ).

Since fl owering time potentially plays a major role in the ex-tent of gene fl ow that can occur between cultivated and weedy rice, here we explore the genetic basis of fl owering time diversity in weedy rice. Gene fl ow between crops and weeds is of great concern because it allows the escape of crop traits into weed populations. In weedy rice, the concern has grown with the recent increase in use of herbicide-resistant cultivated rice varieties in the southern United States; by 2011, 60–65% of cultivated rice in this region was reported to be herbicide-resistant ( Salassi et al., 2012 ). We fi nd that weedy rice populations in the United States have diverged in fl owering phenotype and that these phenotypes are different from the weeds’ cultivated ancestors. Additionally, two major genetic determinants of fl owering time in cultivated rice appear to be only partially responsible for the evolution of differences in fl owering time in weedy rice.

MATERIALS AND METHODS

Plant material — All plant material for this study has been previously described ( Thurber et al., 2010 ; Reagon et al., 2011 ) and includes 58 weedy rice accessions from the southern United States and 87 samples of AA genome Oryza species. Weedy rice accessions include representatives of the major US weedy rice groups, SH and BHA, with BHA comprising a genetically sub-divided population of subgroups identifi ed as BHA1 and BHA2 ( Reagon et al., 2010 ); a few MX accessions, of known hybrid weedy–cultivated rice origin ( Reagon et al., 2010 ) and BRH accessions, of known hybrid SH–BHA origin ( Reagon et al., 2010 ) were also included. Other Oryza accessions include culti-vated Asian O. sativa from the indica , aus , aromatic , tropical japonica , and temperate japonica groups, as well as the wild ancestors of cultivated rice, O. rufi pogon and O. nivara . AA genome outgroup Oryza species consisted of O. glaberrima , O. barthii , O. glumaepatula , and O. meridionalis samples. For further details of origin and collection, see Appendix S1 (see Supplemental Data with the online version of this article).

Phenotyping — A subset of 90 Oryza accessions was grown in Conviron PGW36 growth chambers (Conviron, Winnipeg, Canada) at the University of Massachusetts Amherst under day neutral (12 h light) conditions, and pheno-typed for fl owering time, as described by Thurber et al. (2010) and Reagon et al. (2011) . Rice’s evolutionary origin is in tropical latitudes where day neu-tral conditions are most common ( Khush, 1997 ). The range of changing day lengths experienced in a typical southern US rice fi eld throughout the growing season cannot be easily captured in our growth chamber, but fl owering of both weedy and cultivated rice likely happens after the summer solstice but before the autumnal equinox (D. Gealy, National Rice Research Center, Stuttgart, Arkansas; personal communication). This puts the day length experienced by these plants below 14 h (LD) and closer to day neutral conditions. Thus, we

cultivated rice fi elds in the southern United States. Weedy rice is a worldwide problem in rice agriculture and is the primary constraint to rice production in systems that rely on direct seed-ing ( Chauhan et al., 2013 ). Weedy rice is generally classifi ed as the same species or a close relative of cultivated rice. Because of this, there are concerns about the amount of gene fl ow that may occur between weeds and cultivars.

In the United States, genome-wide single nucleotide poly-morphism (SNP) data suggest that two major genetically dif-ferentiated populations of weedy rice exist; the straw-hull (SH) group is characterized by having straw colored hulls with no awns, and the black-hull awned (BHA) group tends to have black hulls with long black awns ( Londo and Schaal, 2007 ; Reagon et al., 2010 ). Each weed group independently arose from closely related Asian cultivated rice ancestors, SH from indica and BHA from aus ( Reagon et al., 2010 ), suggesting that weeds have undergone a process of de-domestication during their evolutionary origin. Weedy rice possesses many traits that differ from cultivated rice, which are likely adaptive for a weedy life-history, such as the presence of seed shattering, height differences, pericarp coloration, and seed dormancy ( Bres-Patry et al., 2001 ; Gu et al., 2005 ).

The crop ancestors of US weedy rice are highly distinct from the main cultivated variety of rice grown in the United States, tropical japonica . Multiple hybrid incompatibility interactions have been shown to exist between cultivated rice of the japon-ica lineage and varieties of the indica lineage (which encom-pass indica and aus , the US weedy rice ancestors) ( Chen et al., 2008 ; Long et al., 2008 ; Mizuta et al., 2010 ). However, a recent survey of hybrid incompatibility loci in weedy rice has shown fewer than expected postzygotic mating barriers between US weeds and the local crop ( Craig et al., 2014 ). Thus, prezygotic mating barriers are likely to play a larger role in determining the levels of gene fl ow between cultivated and weedy rice. Both weedy and cultivated rice have a tendency to self-pollinate ( Gealy and Gressel, 2005 ), and fl owering time differences be-tween the groups have been observed ( Langevin et al., 1990 ). In rice fi elds in the southern United States, straw-hulled weedy rice typically fl owers earlier than weedy rice with black hulls and also shows some photoperiod sensitivity in some collection sites ( Shivrain et al., 2009 , 2010 ). Flowering time in weedy rice may also overlap with fl owering time in the US crops because there is substantial variation in this trait within weed ecotypes ( Shivrain et al., 2010 ). However, nothing is known about the genetic basis of fl owering time variation in US weedy rice.

In rice, fl owering time is known to be affected by photope-riod (day length), but has also been suggested to be regulated by temperature ( Luan et al., 2009 ). Flowering time is highly vari-able within both cultivated and wild rice, although rice is com-monly referred to as a short-day (SD) plant, with fl owering promoted under SD conditions and delayed under long-day (LD) conditions ( Dung et al., 1998 ; Yano et al., 2000 ). The fl owering time regulatory network has been well characterized in rice, and several genes are known to affect fl owering time variation. For a complete picture of fl owering time regulation, see recent reviews ( Tsuji et al., 2011 , 2013 ; Itoh and Izawa, 2013 ). Briefl y, under SD, OsGI , an ortholog of Arabidopsis GIGANTEA, activates Hd1 , which encodes a B-box zinc fi nger protein and is the ortholog of Arabidopsis CONSTANS , and Ehd1 , a B-type response regulator, which is not orthologous to any Arabidopsis gene. Both genes go on to activate Hd3a , which encodes a phosphatidylethanolamine-binding protein and is the ortholog of Arabidopsis FT and the mobile fl origen,

THURBER ET AL.—EVOLUTION OF FLOWERING STRATEGIES IN WEEDY RICE 1739October 2014]

accession numbers KM043287–KM043383; KM088095–KM088210; KM088211–KM088340; KM088341KM088434; KM088435–KM088567; KM088568–KM088668). Due to high conservation of the Ehd1 promoter and coding sequences in rice, this gene was not investigated despite its expression level being highly cor-related with fl owering time ( Takahashi et al., 2009 ). Primers used in this study can be found in Appendix S2 (see online Supplemental Data).

Genetic diversity analyses — For the Hd1 coding region (exonic sequences) and the Hd3a promoter region, relationships among sequence haplotypes were assessed using maximum likelihood (ML) analyses for each sequenced locus, with the best-fi t model of nucleotide substitution selected in the program jModelTest 2.1.4 ( Guindon and Gascuel, 2003 ; Darriba et al., 2012 ) based on likelihood scores for 88 different models. The GTR model was employed for both sequence data sets as the best available model. Maximum likelihood trees were generated in the program PhyML 3.0 ( Guindon et al., 2010 ) through the ATGC web platform (http://atgc.lirmm.fr/phyml/), with default settings for tree searching and 1000 bootstrap replicates to assess branch support. Silent site nucleotide diversity ( π ), Watterson's estimator of theta ( θ W ) and Tajima’s D were calculated using DNAsp v5 ( Librado and Rozas, 2009 ) for the Hd3a promoter and coding portions of Hd1 .

RESULTS

Flowering time variation in US weedy rice — We had previ-ously shown ( Reagon et al., 2011 ) that, among growth-related traits that may be important for weed fi tness, such as height, tiller number, emergence growth rate, and average growth rate, fl owering time is one of the most divergent traits between the main US weed groups and that the Oryza population of origin has an effect on fl owering time ( H = 45.67, df = 6, P = 1.8 × 10 −13 ). Under a day-neutral growth chamber environment, SH weeds fl owered on average 73 d after emergence, signifi cantly earlier than BHA1 and BHA2 weeds (126 and 111 d on aver-age, respectively) ( Fig. 1 ; pairwise MW with BH correction

believe that the SD path is the one active in the fi eld and growth chamber, although we cannot be sure given that the boundary between what is consid-ered a long and short day has not been well explored. Flowering time was measured as the approximate 50% heading date, between when the fi rst few fl orets began emerging from the boot until anthesis of those fi rst fl orets, on the fi rst emerging panicle of the main tiller for each plant, as described by ( Counce et al., 2000 ). Dates were then transformed into number of days relative to germination date, which was fairly uniform across all individuals. Averages were calculated for each individual and for each major group (Appendix S1) and have been previously reported in Reagon et al. (2011) . Boxplots of the fl owering date data were made using the program R version 3.1.0 (R Develop-ment Core Team, Vienna, Austria; http://www.r-project.org/) ( Fig. 1 ). Signifi cant differences in fl owering time among groups were detected using a nonparamet-ric Kruskal–Wallis (KW) test, followed by pairwise Mann–Whitney (MW) tests using the Benjamini and Hochberg (BH) correction for multiple comparisons ( Reagon et al., 2011 ).

Sequencing — Plants used for DNA sequencing were grown at the University of Massachusetts Amherst, and DNA was extracted and quantifi ed as described in Reagon et al. (2010) . Primers for the Hd1 gene open reading frame (chromosome 6) and Hd3a promoter (chromosome 6) were designed using the program Primer3 ( Rozen and Skaletsky, 1999 ) based on the O. sativa genome (TIGR v. 5 January 2008). Sequencing was performed by either Cogenics (Houston, Texas, USA) or Beckman Coulter Genomics (Danvers, Massachusetts, USA). Sequence alignment, including base pair calls, quality score assignment, and construction of contigs was performed as by Caicedo et al. (2007) using BioLign version 2.09.1 (Tom Hall, North Carolina State University, Raleigh, North Carolina, USA). Approximately 4.5 kb of sequence data were generated for 144 wild, weedy, and cultivated rice accessions at the Hd1 locus (GenBank accession numbers KM063440–KM063573). Because of excessive amounts of polymorphism in the intron and regions immedi-ately fl anking the gene, only 1.25 kb of exonic sequence were retained for analyses. Additionally, six ~500-bp regions of genes increasingly distant from the Hd1 locus were sequenced spanning a region of 4 Mb so as to assess nucleotide diversity in the broader genomic region surrounding this gene (GenBank accession numbers KM077305–KM077432). An additional ~1.7 kb of data were generated for the re-gion immediately upstream of the Hd3a coding sequence, which is likely to contain the promoter region, in 84 accessions of wild, weedy, and cultivated rice (GenBank

Fig. 1. Flowering time phenotype in weedy, cultivated, and wild Oryza . Flowering time, also referred to as “days to heading” or “heading date”, was averaged across two individuals per accession and a boxplot distribution of those averages is shown here. Black line is median, red dot is mean, and white dots represent outliers. Numbers in parenthesis represent sample sizes. Weedy rice groups are as follows: SH (straw-hull), BHA1 (black-hull awned), BHA2 (black-hulled awned), BRH (brown-hull), and MX (mixed ancestry). Boldfaced letters indicate signifi cant differences between groups. Flowering time values were also presented by Reagon et al. (2011) .


Multiple haplotypes (11) were found containing the well-characterized 2-bp deletion ( Fig. 1 ). Most of these haplotypes are unique to cultivars (7), and some are unique to weeds (3), but the most common haplotype, haplotype 3, is shared be-tween weedy, wild, and cultivated groups. Interestingly, haplo-type 3 is present in nearly all BHA weeds examined (BHA1: 81%, BHA2: 66%), indicating that weeds in this group are un-likely to be photoperiod sensitive. Haplotype 3 is also present in most tropical japonica surveyed (86%) and in the great major-ity of US cultivars surveyed (91%; Appendix S1); the sole US cultivar we observed without haplotype 3, carried haplotype 10, which contains the same 2-bp deletion (Appendix S1), suggest-ing that US cultivars also tend to be photoperiod insensitive.

Various other wild and Asian domesticated Oryza carry haplo-type 3, including a majority of aus cultivars surveyed (84%), the progenitor group of BHA weeds (Appendix S1). This fi nding is consistent with the fi nding of Takahashi et al. (2009) that the deletion is common in cultivated rice from both indica ( indica and aus ) and japonica ( tropical japonica , temperate japonica , and aromatic ) lineages. Interestingly, despite largely sharing the same Hd1 haplotypes, and thus photoperiod insensitivity, BHA groups fl owered substantially later than their aus relatives.

The second most commonly observed haplotype in Oryza groups, haplotype 2, formed part of a clade that also contained some haplotypes exclusive to cultivated and/or weedy rice. Haplotype 2, however, was present in all Oryza groups sur-veyed. Haplotypes in this clade did not contain the 2-bp dele-tion, although haplotypes 1, 4, 7, and 43 had additional novel deletions (36 bp, 4 bp, 36 bp, 43 bp, respectively) that may af-fect function. The remaining haplotypes are likely functional, rendering carriers sensitive to day length and leading to early fl owering under SD conditions. Haplotype 2 was found in the majority of SH weeds surveyed (91%; Appendix S1). It was also the only haplotype found in the minor BRH weed group, which is believed to stem from SH–BHA hybridization ( Reagon et al., 2010 ). Haplotype 2 was also present in two in-dica , one aus and one tropical japonica we surveyed.

The indica cultivar group, which is the putative progenitor of the SH weeds, was highly diverse in Hd1 alleles among the ac-cessions we sampled. Seven different haplotypes were found in the group, with no more than 30% of indica cultivars carrying a single shared haplotype. SH weeds shared an Hd1 haplotype with a subset of cultivars from China and Cambodia. The dif-ference in predominant Hd1 allele in SH weeds and indica may explain the differences in fl owering time observed between those groups. However, the majority of Hd1 haplotypes de-tected in the indica group seem to be functional (Appendix S1), suggesting that many indica varieties are photoperiod sensitive like the SH weeds.

Additional nonfunctional alleles of Hd1 were identifi ed by Takahashi et al. (2009) , yet only one was present in our study panel. This allele, haplotype 4, is shared by a single indica cul-tivar and a single weed of mixed ancestry and contains a 4-bp deletion shown to produce a nonfunctional Hd1 allele ( Takahashi et al., 2009 ). Other novel deletions, ranging in size from sin-gle bases to 43 bases, which may reduce or eliminate function of this gene, were present in several haplotypes (haplotypes 1, 7, 27, 37, 40, and 43). Some of these haplotypes (1, 7, and 43) were found in some cultivars or a few SH weeds, while others contained separate deletions unique to wild rice (27, 37 and 40). Additional SNPs that cause amino acid changes were present; however, the extent to which these mutations cause functional changes in the protein is not known.

P = 1.4 × 10 −5 and P = 0.005, respectively; see also [ Reagon et al., 2011 ]). BRH, a minor weed group that is likely descen-dant from SH and BHA hybridization events ( Reagon et al., 2010 ) has a fl owering time range intermediate between its two progenitors (84 d on average; Fig. 1 ), as may be expected if fl owering time is a polygenic trait.

In the fi eld in the United States, weedy rice as a whole has been shown to have a range in fl owering dates in relation to the fl owering date of the local US crop, typically tropical japonica ( Shivrain et al., 2010 ). Our common garden experiment mir-rored what has been observed in the fi eld; SH weeds fl owered signifi cantly earlier than tropical japonica (pairwise MW with BH correction P = 0.003), while BHA1 and BHA2 weeds fl ower concurrently or somewhat after ( Fig. 1 ; see also Reagon et al., 2011 ). Although past evidence of crop–weed hybridiza-tion has been low ( Shivrain et al., 2009 ), fl owering time in indi-viduals known to be derived from SH or BHA hybridization with tropical japonica cultivars (MX weeds) spanned the large range seen in weeds and crops, as expected ( Fig. 1 ).

Interestingly, fl owering time in the main US weedy rice groups also differed substantially from that observed for their cultivated relatives. We found that SH weeds fl owered signifi -cantly earlier than their progenitor indica , ( Fig. 1 , pairwise MW with BH correction P = 0.003; see also Reagon et al., 2011 ). Both BHA groups fl ower later than their progenitor aus , with the BHA1 group doing signifi cantly so ( Fig. 1 , pairwise MW with BH correction P = 0.007; see also Reagon et al., 2011 ). Among cultivars, the aus cultivars fl owered slightly earlier than either indica or tropical japonica ; however, this difference is not statistically signifi cant ( Fig. 1 ). Oryza rufi pogon , the wild ancestor of cultivated rice showed the widest range in fl owering times under our growth chamber conditions, although few tend to fl ower as early as SH weedy rice ( Fig. 1 ). In fact, many wild accessions do not reach fl owering under our conditions.

Sequence diversity in Hd1 and fl anking regions — To try to understand the genetic mechanisms underlying differences in fl owering phenotype among the weed groups and their crop progenitors, we investigated sequence polymorphism at the major fl owering time candidate gene Hd1 . In particular, we wanted to know whether the US weeds had functional or non-functional Hd1 alleles and whether they shared the same Hd1 alleles as their putative cultivated ancestors. Hd1 has been im-plicated in many studies as being the most important locus for photoperiod sensitivity and fl owering time variation in culti-vated rice ( Yano et al., 2000 ; Takahashi et al., 2009 ; Fujino et al., 2010 ). Multiple alleles in cultivated rice harbor deletions or SNPs that render the resulting protein nonfunctional, eliminat-ing day length sensitivity and leading to later fl owering under short days. The most common of these alleles, present in ~43% of rice cultivars, is a 2-bp deletion in the second of the two ex-ons that causes a premature stop codon and is found in both some indica and some japonica varieties ( Takahashi et al., 2009 ).

In our samples, exons of the Hd1 gene displayed extremely high levels of diversity, unlike those typically seen for other loci ( Reagon et al., 2010 ); over 50 haplotypes were detected for the coding region alone. A majority of the haplotypes are unique to wild rice, yet there are several haplotypes unique to cultivated rice and even a few unique to weeds ( Fig. 2 ; Appendix S1). The most common haplotypes (haplotypes 2 and 3) are shared be-tween weeds, cultivars and wild rice ( Fig. 2 ).


Fig. 2. Maximum likelihood tree for Hd1 coding region haplotypes. Numbers next to nodes correspond to bootstraps using 1000 replicates. The tree is rooted by outgroup AA genome rice species (haplotypes 48 and 49). The clade marked with an arrow and labeled −2bp contains haplotypes that share a suite of ~14 SNPs and a causative 2-bp deletion in exon two that truncates the HD1 protein at the C terminal end, leaving the protein nonfunctional. Additional small arrows point to haplotypes containing separate nonfunctional alleles containing different deletions. The two starred haplotypes (haplotype 2-white, haplotype 3-black) are the most common haplotypes in SH and BHA weeds, respectively. Haplotypes are color-coded by the key on the bottom left.


fl owering time variation in cultivated and wild rice ( Yano et al., 2000 ; Fujino et al., 2010 ; Takahashi and Shimamoto, 2011 ), other loci are known to contribute to diversity in fl owering time. We decided to investigate another major fl owering time determinant, Hd3a , the mobile fl origen that initiates the transi-tion to fl owering. Research has shown that the Hd3a promoter type is highly associated with fl owering time in cultivated rice and that there are two main promoter types of Hd3a in culti-vated rice ( Takahashi et al., 2009 ). Type A promoters lead to lower expression of Hd3a and a later fl owering phenotype un-der SD, while type B promoters have higher Hd3a expression and earlier fl owering under SD. These promoter types differ by 11 SNPs and 12-bp indel, any of which may be responsible for the differences in gene expression. Both types of promoter oc-cur in indica and japonica varieties.

In our data set, there was a moderate amount of diversity in the number of Hd3a promoter haplotypes ( Fig. 3 ). The type A promoters (haplotypes 1, 4 and 5) are distinct from the type B promoters (haplotypes 2, 3, 6, 8, 10, 23 and 27). However, the inclusion of weedy rice brings in new recombinant haplo-types, which cannot be classifi ed as A or B types based on se-quence polymorphism, and might even be intermediate in expression level. About 50% of BHA1, 37.5% of BHA2, and 24% of SH weeds had these unique recombinant promoter types. However, the majority of SH weeds (67%) grouped with indica (50%), sharing B type promoters (haplotypes 6 and 10), consistent with early fl owering seen in both the growth chamber and fi eld ( Fig. 3 , Appendix S1). Interestingly, some BHA1 weeds (43.7%) and BHA2 weeds (12.5%) grouped with aus (60%), also sharing a B type promoter (haplotype 2), which is

We examined silent site nucleotide diversity within the Hd1 gene in each of the groups we surveyed. Diversity was fairly high in O. rufi pogon , reduced in the cultivated groups ( indica , aus , and tropical japonica ) and further reduced in some of the weeds (SH and BHA1) ( Table 1 ), following expected patterns brought on by bottlenecks during domestication and weed ori-gins ( Reagon et al., 2010 ). However, levels of nucleotide diver-sity in wild rice were high compared to other genes we have surveyed, mirroring the great number of haplotypes carried by this group, and consistent with diversifying selection for local adaptation (e.g., Caicedo et al., 2007 ; Reagon et al., 2010 ; Thurber et al., 2010 ; Huang et al., 2012 ). In cultivated and weedy groups, on the other hand, there was no clear trend in Hd1 nu-cleotide diversity; the SH, BHA1, and aus groups had lower than genome-wide averages, while BHA2, indica and tropical japonica had higher than genome-wide averages ( Reagon et al., 2010 ) ( Table 1 ). In two weed groups and in the aus varieties, low levels are primarily due to the near fi xation of a single hap-lotype in each group.

We sequenced a sample of loci fl anking Hd1 to determine whether the larger genomic region surrounding this gene showed any signs of selection. Flanking regions had a drastic reduction in diversity in the cultivated and weedy groups, with the great-est in the SH weeds where diversity did not recover within 4 Mb ( Table 1 ). There was an over-abundance of nucleotide diversity in the O. rufi pogon and indica samples at the farthest fl anking region (hd13_011, Table 1 ).

Sequence diversity in the Hd3a promoter — Although Hd1 has been shown to be one of the most important determinants of

TABLE 1. Silent site nucleotide diversity per kilobase (Watterson’s estimator nucleotide variation ( θ W ), the average pairwise nucleotide diversity ( θ π ), and Tajima's D ) for wild Oryza rufi pogon , the main cultivated O. sativa groups, and the main weedy O. sativa groups.

Cultivated Oryza sativa US weedy rice

Region Statistic Oryza rufi pogon/

Oryza nivara indica aus tropical japonica SH BHA1 BHA2

Hd3a promoter θ w 15.05 5.8 3.23 2.37 2.05 4.52 3.06 θ π 13.76 5.3 2.69 1.23 1.68 2.96 4.31

Tajima’s D −0.39 −0.54 −1.21 − 1.85 −0.62 −1.4 2.08 No. of haplotypes 10 6 3 4 9 10 6

Hd1 locus θ w 14.3 3.57 0 0.72 0 0 3 θ π 9.11 3.07 0 1.96 0 0 2.28

Tajima’s D −1.21 −0.51 N/A −1.51 N/A N/A −1.14No. of haplotypes 28 7 2 7 3 2 3

Flanking fragments hd1f_001 θ w 2.7 2.7 3.8 2.4 0 0 2.63 (hd15_004) θ π 3 2.67 3.7 0.83 0 0 1.8 Tajima’s D 0.35 −0.02 −0.31 −2 N/A N/A −1.5 hd1f_002 θ w 1.1 0 0 0 0 0.8 1.03 (hd15_008) θ π 0.5 0 0 0 0 0.35 0.66 Tajima’s D −0.93 N/A N/A N/A N/A −1.16 −1.1 hd1f_003 θ w 7 0 0 0 0 0 0 (hd15_012) θ π 2.9 0 0 0 0 0 0 Tajima’s D −1.6 N/A N/A N/A N/A N/A N/A hd1f_004 θ w 2.3 0 0 0 0 0 0 (hd13_001) θ π 1.4 0 0 0 0 0 0 Tajima’s D −0.97 N/A N/A N/A N/A N/A N/A hd1f_005 θ w 3.95 9.1 0 0.73 0 0.79 0 (hd13_006) θ π 1.7 5.89 0 0.26 0 0.31 0 Tajima’s D −1.5 −1.7 N/A −1.2 N/A N/A N/A hd1f_006 θ w 9.1 15.2 1.3 0.8 0 1.7 1.16 (hd13_011) θ π 16.4 19 1.06 0.6 0 1.1 0.87 Tajima’s D 2.5 1.8 −0.82 −0.5 N/A −0.96 −0.93

Notes : Boldfaced values are statistically signifi cant at P < 0.05.


in contrast to most genes surveyed in this group thus far ( Thurber et al., 2010 ; Reagon et al., 2011 ). This limited sharing is possibly due to under-sampling of aus haplotypes.

Associations between haplotypes and fl owering time — Since the functionality of the HD1 protein is integral to rice fl owering and the common 2-bp deletion has been shown to increase fl owering under SD, we examined whether fl owering time dif-ferences occurred between individuals with or without func-tional alleles. Due to small samples sizes and the close relatedness of weeds and cultivars, we chose to pool these indi-viduals. Thus, our analysis did not account for population struc-ture. When all weeds were pooled into a “weed” group and all cultivars were pooled into a “cultivated” group, there was a sig-nifi cant difference in days to fl owering between individuals with functional Hd1 alleles and those with the 2-bp deletion ( Fig. 4 ). Those with a loss of function allele fl ower signifi cantly

unexpected given their later fl owering. It is important to note that there was one type A haplotype (haplotype 1) that was shared between a subset of both SH (9.5%), BHA1 (6.3%) and BHA2 (37.5%) weeds and was very common in the tropical japonica sampled, both US and Asian cultivars, which is sug-gestive of hybridization between crops and weeds in the fi eld ( Fig. 3 , Appendix S1).

Nucleotide diversity was quite high for all groups for the Hd3a promoter. Although fewer haplotypes were observed, nucleotide diversity was higher than that of the Hd1 coding region ( Table 1 ). As expected, diversity was highest for wild rice. However, nucleo-tide diversity was also surprisingly high in weed groups compared with that of cultivated groups ( Table 1 ). This was particularly no-ticeable in the BHA groups in comparison to their aus relatives, consistent with the great diversity of haplotypes found in these weed groups. The limited sharing of haplotypes between aus and BHA groups as well as the greater levels of nucleotide diversity are

Fig. 3. Maximum likelihood tree for Hd3a promoter haplotypes. Numbers next to nodes correspond to bootstraps using 1000 replicates. The tree is mid-point rooted. The type A promoters and type B promoters are labeled as classifi ed by Takahashi et al. (2009) . Type A promoters typically show lower expression of the Hd3a gene compared with type B promoters. The inclusion of weedy rice and additional wild samples brings in new recombinant haplotypes, which have yet to be classifi ed as A or B types and may be intermediate in expression level (unlabeled haplotypes). The two starred haplotypes (haplotype 2-black, haplotype 6-white) are the most common haplotypes in SH and BHA weeds respectively. Haplotypes are color-coded by the key on the bottom left.


BHA weeds tend to fl ower concurrently or later than the crop ( Fig. 1 ). Weed–weed hybrids and crop–weed hybrids presented a range of fl owering phenologies, as expected from their ances-try ( Fig. 1 ).

Similar behavior has been observed in the fi eld in morpho-logically characterized straw-hull and black-hull accessions ( Shivrain et al., 2010 ). Interestingly, BHA weeds have a highly dissimilar morphology to US cultivated rice, which has straw-colored hulls and lacks awns. Thus, it is unlikely that the later fl owering of BHA is benefi tting the group through crypsis. However, BHA may still benefi t from fl owering concurrently or later than the crop through evasion of early removal attempts and from shattering (dispersal) of seeds before harvest.

Because the ancestry of weedy rice populations is known ( Reagon et al., 2010 ), we were also able to examine the extent to which weedy rice fl owering time resembles that of cultivated progenitor groups. Flowering time differs greatly between weedy rice groups and the respective cultivated progenitors, suggesting that it is a trait that has undergone rapid evolution ( Fig. 1 ). Straw-hull weeds tended to fl ower earlier than their indica relatives, and BHA weeds tended to fl ower later than aus. Given such dramatic differences in fl owering time between these crop groups and their weedy descendants, we attempted to determine whether genetic changes at Hd1 , the most impor-tant determinant of variation in the fl owering time pathway ( Takahashi et al., 2009 ), was responsible for the novel weed phenotypes.

The drastic differences in fl owering phenology displayed by the SH and BHA groups are consistent with and possibly ex-plained by the Hd1 alleles. BHA weeds predominantly carry nonfunctional Hd1 alleles, consistent with loss of day-length sensitivity and later fl owering under short-day conditions, such as those in our growth chamber and presumably in most US rice fi elds. SH weeds predominantly carry functional Hd1 alleles and are day length sensitive, leading to early fl owering under short-day conditions, also consistent with what has been ob-served in rice fi elds ( Shivrain et al., 2010 ). Likewise, the shar-ing of nonfunctional haplotypes between US tropical japonica cultivars and BHA is consistent with the overlap in fl owering time seen in these groups. Additionally, the fact that every sin-gle US cultivar we surveyed carried nonfunctional alleles sug-gests that selection for loss of day-length sensitivity has been an important strategy in the development of US rice agriculture. Note, however, that different fi eld conditions and agricultural practices might increase variation in fl owering time in any of these groups, and thus deviations from the expected patterns could occur.

However, Hd1 haplotype alone cannot explain the differ-ences in fl owering time phenotype between weeds and their progenitors. BHA weeds shared the same Hd1 haplotype as most aus cultivars sampled, suggesting that Hd1 cannot explain the divergence in fl owering time in this weed–crop pair. For SH, the role of Hd1 in fl owering time divergence from indica may be more straightforward. SH shared the Hd1 haplotype only with a subset of indica samples (only two in our sample), and indica cultivars had a great diversity in the number of hap-lotypes and in the nucleotide diversity at this locus ( Table 1 , Fig. 2 ). Although most indica we surveyed possessed func-tional Hd1 alleles, some did not, and it is possible that fl owering is not equivalent among all functional haplotypes. The diversity of Hd1 haplotype we found in indica and dearth of haplotypes in aus is supported by other recent surveys of Hd1 in cultivated rice ( Fujino et al., 2010 ). A recent survey of Hd1 exon-only

later than those with a functional allele ( t = 38.65, df = 30, P = 5.62e −8 ), suggesting that Hd1 allele functionality is a determi-nant of fl owering time behavior in weeds as well as crops.

We also examined whether there was a difference in pheno-type between individuals with type A and type B Hd3a pro-moter types. Again, due to small samples sizes and the removal of intermediate promoter types, we pooled weeds and cultivars. When weeds and cultivars were pooled, there was no signifi -cant difference in days to fl owering between individuals with type A Hd3a promoter and those with type B promoters ( Fig. 4 ; t = 3.93, df = 20, P = 0.05426). However, the type A showed a trend toward later fl owering as shown by cultivated rice.

DISCUSSION

In agricultural fi elds, crops have often undergone strong arti-fi cial selection for uniformity in fl owering time to facilitate har-vesting ( Sawers et al., 2005 ; Jung and Müller, 2009 ). Much less is known about the selective pressures that act on agricultural weeds that invade crop fi elds. Early fl owering may be favored in agricultural weeds to ensure seed release and dispersal prior to crop harvest. Alternatively, fl owering simultaneously with the crop may contribute to crypsis and evasion of targeted re-moval attempts, and in some cases, it maybe advantageous for weed seeds to contaminate the crop seed supply. Late-fl owering weeds may also be able to evade harvest through quicker seed maturation and/or seed dispersal. For weedy rice, a conspecifi c weed of cultivated rice, our data shows that multiple fl owering strategies are successful in the fi eld. SH weeds tend to fl ower signifi cantly earlier than the local tropical japonica crop, while

Fig. 4. Phenotypic differences between major haplotype classes. (A) Differences in heading dates between functional and nonfunctional Hd1 haplotypes in cultivated and weedy rice. Differences are statistically sig-nifi cant ( t = 38.65, df = 30, P = 5.62 × 10 −8 ). (B) Differences in heading dates between type A and type B Hd3a promoters in cultivated and weedy rice. Differences are not statistically signifi cant ( t = 3.93, df = 20, P = 0.054). In both panels, sample sizes of combined weedy and cultivated in-dividuals are in parentheses.


mating barriers are slightly stronger between BHA and tropical japonica , previous results suggest that prezygotic mating barri-ers have been more important in limiting gene fl ow between weeds and crops in US fi elds ( Craig et al., 2014 ). High levels of self-fertilization and differences in fl owering time are likely among the most prevalent prezygotic mating barriers in rice. Our results suggest that gene fl ow between SH weeds and US cultivars is likely to be rare, although the length of the entire fl owering period (not measured in this study) will also affect the likelihood of gene fl ow. In contrast, BHA weeds have a greater probability of hybridization with the crop based on fl owering time and Hd1 haplotype.

If hybridization is selectively favored, as could occur if crops carry alleles favorable for the weed, any weak postzygotic mat-ing barriers are unlikely to prevent gene fl ow. In the past 10 yr, rice farmers in the southern United States have increasingly been planting herbicide resistant rice cultivars. Use of these cultivars has grown to the extent that upward of 65% of fi elds in Arkansas in 2011 were planted with herbicide resistant rice ( Salassi et al., 2012 ). Because resistance to herbicide is likely to be selectively favored in weedy rice, we predict that hybrids between weedy and cultivated rice are likely to be increasingly seen in US rice fi elds ( Shivrain et al., 2009 ). Based on fl ower-ing time and our Hd1 genotyping results, we predict that these weed–crop hybrids are likely to predominantly be due to gene fl ow between BHA weeds and US crops.

LITERATURE CITED

BARRETT , S. C. H. 1998 . The evolution of mating strategies in fl owering plants. Trends in Plant Science 3 : 335 – 341 .

BRES-PATRY , C. , M. LORIEUX , G. CLÉMENT , M. BANGRATZ , AND A. GHESQUIÈRE . 2001 . Heredity and genetic mapping of domestication-related traits in a temperate japonica weedy rice. Theoretical and Applied Genetics 102 : 118 – 126 .

CAI , H.-Y. , S. DIAO , Y.-G. HE , L.-P. ZHANG , S.-J. LIU , Y.-G. ZHU , AND Z.-H. ZHANG . 2012 . Genetic and physical mapping of qHY-8 , a pleiotropic QTL for heading date and yield-related traits in rice. Euphytica 184 : 109 – 118 .

CAICEDO , A. L. , S. H. WILLIAMSON , R. D. HERNANDEZ , A. BOYKO , A. FLEDEL-ALON , T. L. YORK , N. R. POLATO , ET AL . 2007 . Genome-wide patterns of nucleotide polymorphism in domesticated rice. PLOS Genetics 3 : e163 .

CHAUHAN , B. S. , K. SINGH , J. K. LADHA , V. KUMAR , AND Y. S. SAHARAWAT . 2013 . Weedy rice: An emerging threat for direct-seeded rice produc-tion systems in India. Journal of Rice Research 1 : 106 .

CHEN , J. , J. DING , Y. OUYANG , H. DU , J. YANG , K. CHENG , J. ZHAO , ET AL . 2008 . A triallelic system of S5 is a major regulator of the reproductive barrier and compatibility of indica–japonica hybrids in rice. Proceedings of the National Academy of Sciences, USA 105 : 11436 – 11441 .

COUNCE , P. A. , T. C. KEISLING , AND A. J. MITCHELL . 2000 . A uniform, objective, and adaptive system for expressing rice development. Crop Science 40 : 436 – 443 .

CRAIG , S. M. , M. REAGON , L. E. RESNICK , AND A. L. CAICEDO . 2014 . Allele distributions at hybrid incompatibility loci facilitate the potential for gene fl ow between cultivated and weedy rice in the US. PLOS ONE 9 : e86647 .

DARRIBA , D. , G. L. TABOADA , R. DOALLO , AND D. POSADA . 2012 . jModelT-est 2: More models, new heuristics and parallel computing. Nature Methods 9 : 772 .

DUNG , L. V. , T. INUKAI , AND Y. SANO . 1998 . Dissection of a major QTL for photoperiod sensitivity in rice: Its association with a gene expressed in an age-dependent manner. Theoretical and Applied Genetics 97 : 714 – 720 .

polymorphism reported values for silent site nucleotide diver-sity ( π ) in O. rufi pogon (5.2), indica (3.6), and japonica (5) ( Huang et al., 2012 ) that have a similar trend as our values (9.11, 3.07 and 1.96 respectively). However, in our samples, we found slightly elevated diversity in O. rufi pogon , and our ja-ponica values were smaller because they were limited to the tropical japonica variety group, rather than to the japonica lin-eage. The diversity of Hd1 haplotypes in indica and the low levels of diversity in SH suggest that fl owering time divergence between these groups may have occurred in part through the bottleneck that gave rise to the SH weeds and led to the loss of diversity of Hd1 in this group.

In contrast with previous work on cultivated rice, we found little evidence that Hd3a contributed to the variation in fl ower-ing of weedy rice. Most SH weeds carried B type promoters that are consistent with early fl owering under short-day condi-tions ( Fig. 3 , Appendix S1). However, the B type promoter was also common in BHA weeds, which fl owered much later. Ad-ditionally, because of recombination, both weed groups also carried promoters of unknown effect that cannot be classifi ed as the A or B type ( Fig. 3 , Appendix S1). Because diversity in the weed groups was higher at this locus, and there was less strict sharing of haplotypes with progenitor groups, Hd3a could con-tribute some to fl owering time divergence during weed evolu-tion, but it does not seem to be the major contributor to the trends seen. It should be noted that, as for Hd1 , the exclusive occurrence of A type Hd3a promoters in US cultivars is consis-tent with later fl owering under SD seen in US cultivars (Ap-pendix S1).

Our results suggest that, although divergence in fl owering time strategies between weed groups was largely attributable to Hd1 , other loci have probably been involved in the divergence of weed fl owering time from that of their cultivated progeni-tors. A possible candidate is Ehd1 , a B type response regulator, whose expression level has been found to be an important fl ow-ering time regulator in cultivated rice ( Takahashi et al., 2009 ). Additionally, a QTL study conducted on F2 mapping popula-tions created from weed × crop crosses identifi ed at least one region on chromosome 8 that is involved in fl owering time dif-ferences between SH and BHA weeds and an indica cultivar, albeit likely more so under long day conditions ( Thurber et al., 2013 ). Within this region is a very promising candidate gene ( Ghd8 / DTH8 / qHY8 ), encoding a putative histone-like CCAAT-box binding transcription factor ( Wei et al., 2010 ; Yan et al., 2011 ; Cai et al., 2012 ). This gene may function as a regulator of Ehd1 and Hd3a downstream of Hd1 , with nonfunctional alleles conferring weaker photoperiod sensitivity ( Wei et al., 2010 ; Yan et al., 2011 ; Cai et al., 2012 ) and temperate adapted rice in Asia has been shown to possess a large deletion in this gene ( Fujino et al., 2013 ). Many additional fl owering candidate genes on nearly every chromosome of cultivated rice have been identifi ed, so additional fi ne mapping of QTLs specifi c to weedy rice may be more informative than additional candidate gene studies.

The overlap in fl owering time and Hd1 haplotypes between BHA weeds and US crops suggests that some interesting dy-namics are likely to occur between these two groups. In the past, hybridization levels between weedy rice and cultivated rice have been relatively modest in crop fi elds ( Shivrain et al., 2009 ). Our previous work has shown that postzygotic mating barriers (i.e., hybrid incompatibility) are unexpectedly weak between both weedy rice groups and the US crop, given each group’s ancestry ( Craig et al., 2014 ). Although postzygotic


ELLSTRAND , N. C. , S. M. HEREDIA , J. A. LEAK-GARCIA , J. M. HERATY , J. C. BURGER , L. YAO , S. NOHZADEH-MALAKSHAH , AND C. E. RIDLEY . 2010 . Crops gone wild: Evolution of weeds and invasives from domesti-cated ancestors. Evolutionary Applications 3 : 494 – 504 .

ELZINGA , J. A. , A. ATLAN , A. BIERE , L. GIGORD , A. E. WEIS , AND G. BERNASCONI . 2007 . Time after time: Flowering phenology and biotic interactions. Trends in Ecology & Evolution 22 : 432 – 439 .

FRANKS , S. J. , S. SIM , AND A. E. WEIS . 2007 . Rapid evolution of fl ow-ering time by an annual plant in response to a climate fl uctua-tion. Proceedings of the National Academy of Sciences, USA 104 : 1278 – 1282 .

FUJINO , K. , J. WU , H. SEKIGUCHI , T. ITO , T. IZAWA , AND T. MATSUMOTO . 2010 . Multiple introgression events surrounding the Hd1 fl owering-time gene in cultivated rice, Oryza sativa L. Molecular Genetics and Genomics 284 : 137 – 146 .

FUJINO , K. , U. YAMANOUCHI , AND M. YANO . 2013 . Roles of the Hd5 gene controlling heading date for adaptation to the northern limits of rice cultivation . Theoretical and Applied Genetics 126 : 611 – 618 .

GEALY , D. R. , AND J. GRESSEL . 2005 . Gene movement between rice (Oryza sativa) and weedy rice (Oryza sativa): A U.S. temperate rice perspec-tive . In J. Gressel [ed.], Crop ferality and volunteerism, 323–354. CRC Press, Boca Raton, Florida, USA.

GU , X.-Y. , S. F. KIANIAN , AND M. E. FOLEY . 2005 . Seed dormancy im-posed by covering tissues interrelates to shattering and seed morpho-logical characteristics in weedy rice. Crop Science 45 : 948 – 955 .

GUINDON , S. , J.-F. DUFAYARD , V. LEFORT , M. ANISIMOVA , W. HORDIJK , AND O. GASCUEL . 2010 . New algorithms and methods to estimate maxi-mum-likelihood phylogenies: Assessing the performance of PhyML 3.0. Systematic Biology 59 : 307 – 321 .

GUINDON , S. , AND O. GASCUEL . 2003 . A simple, fast, and accurate al-gorithm to estimate large phylogenies by maximum likelihood. Systematic Biology 52 : 696 – 704 .

HUANG , C.-L. , C.-Y. HUNG , Y.-C. CHIANG , C.-C. HWANG , T.-W. HSU , C.-C. HUANG , K.-H. HUNG , ET AL . 2012 . Footprints of natural and artifi cial selection for photoperiod pathway genes in Oryza . Plant Journal 70 : 769 – 782 .

ITOH , H. , AND T. IZAWA . 2013 . The coincidence of critical day length rec-ognition for fl origen gene expression and fl oral transition under long-day conditions in rice. Molecular Plant 6 : 635 – 649 .

ITOH , H. , Y. NONOUE , M. YANO , AND T. IZAWA . 2010 . A pair of fl oral reg-ulators sets critical day length for Hd3a fl origen expression in rice. Nature Genetics 42 : 635 – 638 .

JUNG , C. , AND A. E. MÜLLER . 2009 . Flowering time control and applica-tions in plant breeding. Trends in Plant Science 14 : 563 – 573 .

KHUSH , G. S. 1997 . Origin, dispersal, cultivation and variation of rice . In T. Sasaki and G. Moore [eds.], Oryza : From molecule to plant, 25–34. Springer Netherlands. Available at http://link.springer.com/chapter/10.1007/978-94-011-5794-0_3 [accessed 21 January 2014].

KOJIMA , S. , Y. TAKAHASHI , Y. KOBAYASHI , L. MONNA , T. SASAKI , T. ARAKI , AND M. YANO . 2002 . Hd3a , a rice ortholog of the Arabidopsis FT gene, promotes transition to flowering down-stream of Hd1 under short-day conditions. Plant & Cell Physiology 43 : 1096 – 1105 .

LANGEVIN , S. A. , K. CLAY , AND J. B. GRACE . 1990 . The incidence and ef-fects of hybridization between cultivated rice and its related weed red rice ( Oryza sativa L.). Evolution 44 : 1000 – 1008 .

LIBRADO , P. , AND J. ROZAS . 2009 . DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25 : 1451 – 1452 .

LONDO , J. P. , AND B. A. SCHAAL . 2007 . Origins and population genetics of weedy red rice in the USA. Molecular Ecology 16 : 4523 – 4535 .

LONG , Y. , L. ZHAO , B. NIU , J. SU , H. WU , Y. CHEN , Q. ZHANG , ET AL . 2008 . Hybrid male sterility in rice controlled by interaction between di-vergent alleles of two adjacent genes. Proceedings of the National Academy of Sciences, USA 105 : 18871 – 18876 .

LUAN , W. , H. CHEN , Y. FU , H. SI , W. PENG , S. SONG , W. LIU , ET AL . 2009 . The effect of the crosstalk between photoperiod and temperature on the heading-date in rice. PLOS ONE 4 : e5891 .

MIZUTA , Y. , Y. HARUSHIMA , AND N. KURATA . 2010 . Rice pollen hy-brid incompatibility caused by reciprocal gene loss of duplicated genes. Proceedings of the National Academy of Sciences, USA 107 : 20417 – 20422 .

OERKE , E.-C. 2006 . Crop losses to pests. Journal of Agricultural Science 144 : 31 – 43 .

OLSEN , K. M. , AND J. F. WENDEL . 2013 . Crop plants as models for un-derstanding plant adaptation and diversifi cation. Frontiers in Plant Science 4 : 290 [accessed 19 March 2014] Available at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3729982/ .

REAGON , M. , C. S. THURBER , B. L. GROSS , K. M. OLSEN , Y. JIA , AND A. L. CAICEDO . 2010 . Genomic patterns of nucleotide diversity in divergent populations of US weedy rice. BMC Evolutionary Biology 10 : 180 .

REAGON , M. , C. S. THURBER , K. M. OLSEN , Y. JIA , AND A. L. CAICEDO . 2011 . The long and the short of it: SD1 polymorphism and the evolution of growth trait divergence in U.S. weedy rice. Molecular Ecology 20 : 3743 – 3756 .

ROUX , F. , P. TOUZET , J. CUGUEN , AND V. LE CORRE . 2006 . How to be early fl owering: An evolutionary perspective. Trends in Plant Science 11 : 375 – 381 .

ROZEN , S. , AND H. SKALETSKY . 1999 . Primer3 on the WWW for general users and for biologist programmers . In S. Misener and S. A. Krawetz [eds.], Bioinformatics methods and protocols, Methods in Molecular Biology, 132:365–386. Humana Press, New York, New York, USA. Available at http://link.springer.com/protocol/10.1385/1-59259-192-2%3A365 [accessed 21 January 2014].

SALASSI , M. , C. E. WILSON JR ., AND T. W. WALKER . 2012 . Proceedings of the Thirty-Fourth Rice Technical Working Group, 17 – 26 , Hot Springs, Arkansas, USA. University of Arkansas, Fayetteville, Arkansas, USA.

SAWERS , R. J. , H. MOIRA , J. SHEEHAN , AND T. P. BRUTNELL . 2005 . Cereal phytochromes: Targets of selection, targets for manipulation ? Trends in Plant Science 10 : 138 – 143 .

SHIVRAIN , V. K. , N. R. BURGOS , D. R. GEALY , M. A. SALES , AND K. L. SMITH . 2009 . Gene fl ow from weedy red rice ( Oryza sativa L.) to cultivated rice and fi tness of hybrids. Pest Management Science 65 : 1124 – 1129 .

SHIVRAIN , V. K. , N. R. BURGOS , R. C. SCOTT , E. E. GBUR JR ., L. E. ESTORNINOS JR ., AND M. R. MCCLELLAND . 2010 . Diversity of weedy red rice ( Oryza sativa L.) in Arkansas, U.S.A. in relation to weed management. Crop Protection (Guildford, Surrey) 29 : 721 – 730 .

TAKAHASHI , Y. , AND K. SHIMAMOTO . 2011 . Heading date 1 ( Hd1 ), an ortho-log of Arabidopsis CONSTANS , is a possible target of human selec-tion during domestication to diversify fl owering times of cultivated rice. Genes & Genetic Systems 86 : 175 – 182 .

TAKAHASHI , Y. , K. M. TESHIMA , S. YOKOI , H. INNAN , AND K. SHIMAMOTO . 2009 . Variations in Hd1 proteins, Hd3a promoters, and Ehd1 ex-pression levels contribute to diversity of fl owering time in cultivated rice. Proceedings of the National Academy of Sciences, USA 106 : 4555 – 4560 .

TAMAKI , S. , S. MATSUO , H. L. WONG , S. YOKOI , AND K. SHIMAMOTO . 2007 . Hd3a protein is a mobile fl owering signal in rice. Science 316 : 1033 – 1036 .

THURBER , C. S. , M. H. JIA , Y. JIA , AND A. L. CAICEDO . 2013 . Similar traits, different genes? Examining convergent evolution in related weedy rice populations. Molecular Ecology 22 : 685 – 698 .

THURBER , C. S. , M. REAGON , B. L. GROSS , K. M. OLSEN , Y. JIA , AND A. L. CAICEDO . 2010 . Molecular evolution of shattering loci in U.S. weedy rice. Molecular Ecology 19 : 3271 – 3284 .

TSUJI , H. , K. TAOKA , AND K. SHIMAMOTO . 2011 . Regulation of fl owering in rice: two fl origen genes, a complex gene network, and natural varia-tion. Current Opinion in Plant Biology 14 : 45 – 52 .

TSUJI , H. , K. TAOKA , AND K. SHIMAMOTO . 2013 . Florigen in rice: Complex gene network for fl origen transcription, fl origen activation com-plex, and multiple functions. Current Opinion in Plant Biology 16 : 228 – 235 .


WARWICK , S. I. , AND C. N. STEWART . 2005 . Crops come from wild plants: How domestication, transgenes, and linkage together shape ferality . In Crop ferality and volunteerism, 9–30. CRC Press, Boca Raton, Florida, USA.

WEBER , E. , AND B. SCHMID . 1998 . Latitudinal population differentiation in two species of Solidago (Asteraceae) introduced into Europe. American Journal of Botany 85 : 1110 – 1121 .

WEI , X. , J. XU , H. GUO , L. JIANG , S. CHEN , C. YU , Z. ZHOU , ET AL . 2010 . DTH8 suppresses fl owering in rice, infl uencing plant height and yield potential simultaneously. Plant Physiology 153 : 1747 – 1758 .

YAN , W.-H. , P. WANG , H.-X. CHEN , H.-J. ZHOU , Q.-P. LI , C.-R. WANG , Z.-H. DING , ET AL . 2011 . A major QTL, Ghd8 , plays pleiotropic roles in regulating grain productivity, plant height, and heading date in rice. Molecular Plant 4 : 319 – 330 .

YANG , Y. , J. KLEJNOT , X. YU , X. LIU , AND C. LIN . 2007 . Florigen (II): It is a mobile protein. Journal of Integrative Plant Biology 49 : 1665 – 1669 .

YANO , M. , Y. KATAYOSE , M. ASHIKARI , U. YAMANOUCHI , L. MONNA , T. FUSE , T. BABA , ET AL . 2000 . Hd1 , a major photoperiod sensitivity quantita-tive trait locus in rice, is closely related to the Arabidopsis fl owering time gene CONSTANS . Plant Cell Online 12 : 2473 – 2483 .

he evolution us weedy 1 - arts & sciences pages · in the evolution of weedy rice in the united...

Documents