Yulin Jia,David Gealy

USDA Agricultural Research Service,Dale Bumpers National Rice Research Center,Stuttgart,AR 72160,USA

Keywords:Weedy red rice Oryza sativa Geng Aus Blast disease Sheath blight disease

A B S T R A C T Weedy red rice(Oryza sativa;WRR),a close relative of cultivated rice,is a highly competitive weed that commonly infests rice fields and can also naturally interbreed with rice.Useful genes for biotic stress have been maintained in WRR and can be explored for breeding.Here we describe genetic and physiological traits of WRR that can be beneficial in preventing major rice diseases.Rice blast,caused by the hemibiotrophic fungal pathogen Magnaporthe oryzae,and sheath blight disease,caused by the necrotrophic pathogen Rhizoctonia solani,are the two most damaging biotic stresses of rice.Many major and minor resistance genes and QTL have been identified in cultivated and wild rice relatives.However,novel QTL were recently found in the two major U.S.biotypes of WRR,blackhull-awned(BH)and strawhullawnless(SH),suggesting that WRR has evolved novel genetic mechanisms to cope with these biotic stresses.Twenty-eight accessions of WRR(PI 653412–PI 653439)from the southern USA were characterized and placed in the National Small Grains Collection,and are available for identification of novel genetic factors to prevent biotic stress.

1.Introduction

Rice is one of the most important food crops.It feeds half of all human beings and has been grown in diverse ecological systems for more than a thousand years[1,2].The United Nations estimates that by 2030,the world population will grow by an additional 1.2 billion people (http://www.un.org/en/sections/issues-depth/population/index.html).Thus,a demand for additional rice needed to meet increased human consumption is recognized worldwide.Long before there were organized rice breeding efforts,rice plants that survived after each disease epidemic were selected and used as parents for subsequent seed production (a process known as “selection” breeding). Such selection after epidemics over the history of rice domestication and cultivation has created a genetic bottleneck for genetic resistance to other strains or other diseases in cultivated rice.In contrast,a highly competitive weedy relative of rice,weedy red rice(WRR,Oryza sativa)that commonly infests rice fields has not been selected by humans,but has survived through stochastic introduction and de-domestication[3].

Rice,a monocot,is a self-pollinated species,and its genetic diversity is often limited in comparison with that of outcrossing crop species such as maize[1].During rice domestication,some resistance(R)genes were introgressed into cultivated rice,and some of them were lost owing to the absence of selection pressures from pathogens and environments.In theory,genes for resistance to all races of rice pathogens are present in different rice varieties.Temporal and spatial mismatching of host R genes to pathogen races is a common cause of disease epidemics.Oryza and its pathogens represent one of the best models to elucidate genetic and physiological mechanisms of host–pathogen interactions,not only because rice is an essential food for humanity,but also because the rice genome is one of the smallest among cereal crops,with a complete genome sequence available since 2002[4]as well as genome sequences of rice pathogens[5].There are two cultivated rice species,Asian rice(O.sativa)and African rice(O.glaberrima).Cultivated Asian rice is thought to have originated from O.rufipogon and O.nivara in Asia and cultivated Africa rice from O.barthii in Africa,respectively.Asian rice has two major subgroups,Geng(japonica)and Xian(indica)[6].The Geng subgroup consists of tropical Geng,temperate Geng,and aromatic rice,and the Xian subgroup consists of Aus and Xian[7].

WRR is one of the major rice weeds,reducing both the quantity and quality of the rice crop and adapted to a wide range of environments.WRR is a troublesome weed in rice in the southern USA(where primarily tropical Geng cultivars are grown),and throughout the world,especially where direct seeding methods are employed(Fig.1)[8].The diversity of biotypes of red rice creates a challenge for management of this weed,but also creates a unique opportunity for detailed biological and genetic studies of adaptive traits that have been lost in domesticated rice and are not found in wild rice relatives(Fig.2).

Fig.1–Photographs of red rice in a commercial rice field in Arkansas(A)and in an experimental field(B),Stuttgart,Arkansas,showing the diverse weedy red rice lines submitted to the GRIN(Germplasm Resources Information Network)collection.In(A),the taller plants with light green-colored leaves are red rice and the others are cultivated rice.In(B),there are 28 nine-row plots containing individual weedy red rice biotypes.Plots are arranged four wide by seven deep,but individual plots are less clearly differentiated toward the top of the photograph.

Fig.2–Photographs of seeds of weedy rice and a closely related cultivated rice.A.Blackhull-awned(PI 653419);B.Straw hullawnless(PI 653435);C.DGWG,a Geng variety,a donor of SD1.A full-length view of the awns of the blackhull weedy rice biotype in(A)is shown in Fig.4-B.

2.Major biotic stress

In the USA and worldwide,major biotic stresses of rice are rice blast and sheath blight disease. Cultivated rice has been constantly challenged by the ascomycete filamentous fungus Magnaporthe oryzae(synonym of Pyricularia oryzae)causing rice blast disease. At present, rice blast disease is the most damaging rice disease worldwide,responsible for an estimated annual 30%crop loss corresponding to food for 60 million people[9].In the USA,where approximately 2%of the world's rice is produced,rice farmers spend US$69 million per year to prevent crop loss due to blast disease[10].

M.oryzae,a highly adaptive fungus,is a polycyclic pathogen that can reproduce three to five times during a single crop season [11]. For each life cycle, asexual spores begin to germinate in water within minutes,and germinated tubes commence growing within hours. Penetration into rice directly through the cell wall,enabled by high fungal turgor pressures, starts when the germinated tubes encounter inductive signals from rice[12].Subsequent biotrophic and necrotrophic invasion is accelerated by highly diverse effectors and pathogenicity factors[12].After approximately two days post-pathogen attachment,M.oryzae feeds on rice saprophytically,similarly to a necrotrophic pathogen.At the end of the life cycle,millions of infectious asexual spores can be produced and disseminated by wind,rain,insects,birds,and stochastic human activities[11].After rice is harvested,the infectious spores can reproduce in all rice debris,including seeds.Magnaportha oryzae can infect all parts of rice,including roots,although it causes injury primarily above ground.

Genetic studies of cultivated rice and wild rice relatives have identified both major and minor resistance(R)genes that prevent blast infection.To date,blast R genes,most of them from Geng rice types,have been found on all 12 rice chromosomes,and over a dozen have been cloned. Most of these R genes encode a cytoplasmic protein with a nucleotide binding site–leucine richrepeat domain(NLR),indicating their roles as receptors and/or in protein–protein interaction[13].

Fig.3–Demonstration of sheath blight infection with sclerotia under controlled environmental conditions.A.RR20(PI 653419)was inoculated with sclerotia of Rhizoctonia solani(Rs),and wrapped with Parafilm for one week;B.Rice cultivar Lemont was inoculated with mycelia of Rs for one week as a susceptible control.

The second most damaging disease of rice,sheath blight,is caused by a soilborne necrotrophic pathogen,Rhizoctonia solani,and often surpasses blast disease in severity and economic damage.Rhizoctonia solani is a monocyclic pathogen,reproducing only once in a single crop season. It infects diverse monocots and dicots in 13 anastomosis groups(AGs)that define their host specificity.The AG1-IA of R.solani infects rice above ground and initiates infection from sclerotia once in contact with the stem of rice after permanent flooding is established in flood-irrigated rice production systems(Fig.3).Sheath blight is an old disease,but a new problem,because of worldwide deployment of high-density,high-yielding,semidwarf rice production systems.Yield losses of 5%–15%from sheath blight are still fairly common on many farms in the southern USA and worldwide.However,sheath blight disease often damages the rice stem,sheath,and leaf,and is not as dramatic as neck blast,which may cause loss of the entire panicle.Genetic resistance to R.solani has been a subject of intensive study.To date,no major R genes for R.solani have been identified in cultivated and wild rice relatives. It is believed that plant architecture,plant height,and heading date play major roles in reducing damage due to R.solani.The major effect of resistance QTL on chromosome 9 was identified as originating in cultivated rice[14],and should be useful for improving resistance.Despite the wealth of candidate genes that have been identified,the extent of phenotypic contributions to resistance has not been determined.

3.Genetic advantages of weedy red rice

3.1.Population structure in the USA

The phenotypic descriptions of southern USA WRR populations have been largely grouped into strawhull-awnless(SH)and blackhull-awned(BH)groups,with SH typically comprising the bulk of the accessions(60%–72%),BH comprising the second-largest fraction(22%–40%),and other less common biotypes comprising the rest[15].Interestingly,the phenotypic population structure of the two major southern U.S.WRR biotypes corresponds to two major genetic groupings representing different phylogenetic origins for SH(related to Geng cultivated rice)and BH(related to Aus cultivated rice)[16–20].Polymorphism information about diversity in the SH and BH biotypes based on SNP data from genotyping-bysequencing analysis is available in Qi et al.[23](data can be accessed at doi:https://doi.org/10.5061/dryad.566 h9, reference#23 in this review).There is good evidence[3,16,18,21]that the two major biotypes of U.S.WRR were initially derived from de-domestication of different lines of Asian cultivated rice in different events in the distant past,and not directly from wild rice(O.rufipogon).A recent analysis[3],however,suggests that the BH biotypes descended more directly from O.rufipogon than previously thought,given their relatively high component of private SNPs belonging to wild rice(relative to U.S.SH or China WRR biotypes),and that their divergence from O.rufipogon occurred earliest among all U.S.WRR and after the Aus–Geng varietal divergence.It further suggests that the weediness adaptations of SH and BH WRR employed different genetic mechanisms.

In contrast to rice production systems in the southern USA,California production systems show relatively rare WRR infestation,owing in large part to many decades of aggressive efforts to eradicate the weed in California[22].California weedy rice is generally morphologically similar to mediumgrain or gourmet cultivars of the area,but also has colored pericarp,seed shattering,and awns typical of wild relatives.Based on genetic characteristics, it appears to have very recently evolved from O.sativa temperate Geng cultivars in California,with a genetic background very different from that of southern U.S.weedy rice[22].

In a genotyping-by-sequencing analysis of mapping populations derived from distinct southern U.S.WRR populations of SH and BH[23],QTL mapping and sequencing of candidate genes indicated that trait variation for certain weediness traits including awn length,hull pigmentation,and pericarp pigmentation,was generally attributable to individual loci(Fig.2).However,more complex quantitative traits such as heading date, panicle length, and seed shattering were controlled by multiple QTL,which did not appear to be shared between the SH and BH populations. Thus, the genetic variation in O. sativa apparently has allowed agricultural weedy traits to evolve in multiple ways[23].

3.2.Gene flow

WRR has used multiple methods to increase genetic diversity and adaptive ability. One of these is outcrossing with cultivated rice and other congeners where available.Phenotypes(such as very short awns or intermediate hull colors)of some relatively rare biotypes in the southern USA have been shown to be genetically consistent with past crossing among different major WRR biotypes, while other rare biotypes resulted from crossing between WRR and rice cultivars[16,18, 24, 25]. Other relatively rare, but swiftly increasing,populations of red rice biotypes of WRR have resulted from recent introgression of cultivated(often herbicide-resistant)rice genes into red rice[26].Burgos et al.[26]found that herbicide-resistant WRR biotypes recently derived from herbicide-resistant rice×WRR crosses appear to be derived predominantly from BH biotypes.Genes involved in flowering time may have helped WRR evolve as a weed and there may have been multiple flowering strategies in WRR in the USA[27].Early maturity has been a trend among modern cultivars,including those resistant to imidazolinone herbicides,so that these cultivars are more likely to have flowering periods that are synchronized with SH WRR biotypes,which generally flower earlier than BH biotypes. Thus, new “weed-like”crosses between these WRR and rice are likely increasing in farm fields,but a biological bottleneck imposed by extremely delayed flowering in F1crosses between SH WRR and rice probably has helped minimize the number of such weed-like plant biotypes in these fields[28].

3.3.Physiological advantages of weedy red rice in resisting biotic stress

Although abundant information on weedy traits is available,information about disease-resistance traits in weedy rice is limited.Tall plants can avoid damage by sheath blight.Most successful WRR plant biotypes in the USA are taller than modern rice cultivars(～95–105 cm)[15,24,25].In an extensive study of 215 U.S.WRR biotypes in Arkansas,Shivrain et al.[15]found that BH biotypes were tallest,ranging 75–190 cm with an average of 139 cm.Most of these BH biotypes were taller than rice cultivars and possibly better equipped to avoid sheath blight damage.We recently demonstrated this disease damage-avoidance phenomenon by showing that PI 653419(RR20)BH WRR is susceptible to artificial inoculation under greenhouse conditions,even though it has tolerance under field conditions,suggesting that plant height is one of the major confounding factors for mapping sheath blight resistance genes(Fig.3,D.Goad and Y.Jia et al.,unpublished data).Because of this height advantage,many U.S.BH WRR biotypes may be able to avoid damage from R. solani under field conditions.

Shivrain et al.[15]reported that BH WRR produced more tillers (102 plant?1), produced narrower flag leaves, and flowered later than SH WRR,which produced 85 tillers plant?1.Tiller angles of less than 45 degrees from vertical can reduce the formation of dew and reduce both blast and sheath blight infections.Tiller angles of WRR are usually greater(more open or spreading)than those for rice.In Arkansas,88%and 93%of BH and SH WRR biotypes,respectively,were“spreading”or“open”plant types[14,29].Although most of these biotypes may be disadvantaged for tolerance to these diseases based on tiller angle,12%of BH and 7%of SH biotypes had upright,erect plant biotypes similar to those of cultivated rice,and thus might acquire some disease tolerance due to this trait.Hybrid progeny from WRR×cultivated rice crosses are also diverse,and can produce a range of plant biotypes(D.Gealy,unpublished data),some of which are erect and might show tolerance to these diseases.

Elevated silicon (Si) levels can reduce blast in rice.Although we are unaware of any publications that report Si levels in WRR,it is interesting to consider the relative Si content of Aus and Geng rice (closely related to southern U.S. WRR) compared with rice cultivars. In a study of Si content among genetically diverse Oryza species,differences in Si concentrations between tropical Geng and Geng rice were not consistent[30],and thus it is uncertain whether Si concentration contributes markedly to blast tolerance or susceptibility among WRR biotypes or populations. This might be a fruitful area for future research.

Interference studies have demonstrated the destructive effects of WRR against rice or rice against WRR.In the southern USA,Shivrain et al.[31]showed that yield losses due to interference from red rice biotypes ranged from 6%to 45% in herbicide-resistant rice and that these losses generally increased with later planting dates. In the Philippines,Chauhan and Johnson[32]showed that IR72 rice interference greatly reduced tiller and leaf numbers,leaf area,and shoot biomass of all weedy rice variants from five Asian countries,and at eight weeks after sowing,shoot biomass of weedy rice in competition with 12 IR72 plants ranged from 13%to 30%of that where plants were grown alone.

Shorter duration of growth can reduce the probability of blast infection.Interestingly,the predominant SH WRR plant biotypes in the southern USA typically have shorter growing seasons than BH biotypes and most historical cultivars [15, 28]. Rice cultivars released in the past several decades, however, often flower as early as SH WRR.

4.Novel resistance genes in weedy red rice

Fig.4–Weedy red rice and leaf blast disease reaction.A.Plants of strawhull-awnless(SH)and blackhull-awned(BH)grown in a greenhouse;B.Seeds of BH;C.seeds of SH;D.M.oryzae spores;E.Disease reactions of rice seedlings of DGWG and weedy red rice-RR20(PI 653419)and RR9(PI 653435)with different blast races(modified from[34]).

We speculate that absence of human selection favors the development of genetic diversity that may enable better survival of WRR in natural settings.Convergent evolution is one of the adaptive strategies of weedy species of rice.Convergent traits include rapid growth,seed shattering,and dormancy[33].Liu et al.[34]identified 28 new genetic loci associated with blast resistance(Fig.4).Among them,six QTL were identified in SH WRR and five in BH WRR.Four of the 11 QTL were found in the same genomic region on chromosome 4 and another four were found in the same genomic region on chromosome 12,suggesting control by the same genes.The other three resistance QTL were mapped to different chromosome regions. These results suggest that weedy rice evolved different genetic mechanisms to prevent blast disease.Of the 100 major and minor blast R genes that have been identified in cultivated rice and wild rice relatives[13],none has been identified in either SH or BH red rice.Sequence analysis of WRR indeed revealed several new Pi-ta haplotypes,raising the prospect that these new Pi-ta haplotypes and nearby genomic regions harbor novel blast R genes[35,36].Considering the totality of these findings,we suggest that WRR has evolved novel genetic mechanisms to prevent rice diseases that can be used for the improvement of cultivated rice.

5. Establishment of public resources for weedy species of rice in the USA

The National Small Grains Collection(NSGC)houses germplasm of rice and other small grains for distribution through the Germplasm Resources Information Network(GRIN)for scientific uses.For rice,it emphasizes maintenance and distribution of crop rice germplasm and genetic stocks,and historically,it has not included WRR biotypes of O. sativa. With the aim of establishing a collection of typical weedy rice biotypes available for research similar to that for crop rice cultivars,diverse weedy rice accessions were acquired from locations in Arkansas,Mississippi,Louisiana,Missouri,and Texas from 1994 to 2000 before the deployment of herbicide-resistant rice cultivars,and observed and characterized in single row or single hill seedincrease plots at Stuttgart,Arkansas for several seasons.In 2007,28 representative accessions were grown for a major seed increase in multiple-row field plots at Dale Bumpers National Rice Research Center(Fig.1-B and Fig.5).In 2008,the WRR seed was deposited in the NSGC for distribution,and the accessions are now identified as PI 653412 through PI 653439.Half of the accessions(PI 653426–PI 653439)were developed by single-seed descent and thus have particularly high genetic purity.Phenotypes are typically BH or SH and awned or awnless.Plant heights range from 149 to 182 cm,although these values may vary substantially from year to year.Kernel lengths range from 5.4 to 6.7 mm,kernel length/width ratios range from 2.5 to 3.3(similar to those of“medium-grain”rice),and kernel widths range from 1.8 to 2.4 mm.Some SH and BH red rice lines have now been sequenced[23],including the SH line PI 653435 and BH line PI 653419 described in Liu et al.[34](Fig.5),as well as most of the rest of the 28 lines in GRIN(excluding 5A,PI 653412;11D,PI 653417;MS4,PI 653421;StgB,PI 653422;1994–8,PI 653425;1196,PI 653431;1118,PI 653433;and 1134,PI 653434).Seed samples are available in small quantities from the NSGC through GRIN(http://www.ars-grin.gov/npgs/acc/acc_queries.html).

Fig.5–Plot close-ups of selected weedy red rice biotypes in GRIN.Strawhull biotypes on left;blackhull biotypes on right.PI 653435 and PI 653419 were previously[34]used as parents to create mapping populations for mapping novel blast resistance genes.The lines sampled to create PI 653423 and PI 653422 have frequently been used as weedy rice standards in previous agronomic studies.

Acknowledgments

The authors thank Michael Lin,Howard Black,and Tracy Bianco of the United States Department of Agriculture-Agricultural Research Service for excellent technical assistance.The USDA is an equal opportunity provider and employer.