Introduction
Thailand’s northeast region has the highest rice (Oryza sativa L.) production nationwide, including areas within the cities of Ubon Ratchathani, Surin, Sri Sa Ket, Roi Et, and Nakhon Ratchasima, where most rice production systems are rainfed. The major rice varieties commonly grown in this region are ‘KDML105’ and ‘RD6’ (Thanasilungura et al. Reference Thanasilungura, Kranto, Monkham, Chankaew and Sanitchon2020) due to their cooking quality, aroma, and softness. However, the quantity and quality of rice production in this region is limited by certain constraints, such as blast and bacterial blight diseases, drought, salinity, and weeds, particularly in relation to weedy rice.
Weedy rice, or red rice, is described as a genetically diverse population of cultivated rice that has adapted to coevolve with this crop (Gealy et al. Reference Gealy, Burgos, Yeater and Jackson2015). Kane and Baack (Reference Kane and Baack2007) proposed three possible origins for weedy rice: (1) evolving from wild rice; (2) originating from escaped domesticated rice seeds, which then evolved into weedy traits; and (3) evolving through interbreeding between cultivated and wild rice. Two wild species, brownbeard rice (Oryza rufipogon Griffiths) and Oryza nivara Sharma & Shastry, are perennial and annual ancestors of O. sativa (Khush Reference Khush1997). The wild rice, O. rufipogon, is the ancestor of the Asian cultivated rice (O. sativa). Domesticated rice plants and their wild ancestor, O. rufipogon, share some reproductive barriers, such as reproductive isolation due to differences in their flowering times. However, gene flow can occur between domesticated, weedy, and wild rice (Bah et al. Reference Bah, van der Merwe and Labuschagne2017; Jena Reference Jena2010). Natural hybridization has been reported between wild rice (O. rufipogon), weedy rice (O. sativa f. spontanea), and cultivated rice (O. sativa) (Song et al. Reference Song, Lu, Zhu and Chen2003). Weedy rice populations have adapted to the agroecosystem environment of cultivated rice, developing phenotypic mimicry to cultivated rice (Gressel Reference Gressel2005). Although weedy rice and cultivated rice are both from the O. sativa species, weedy rice retains a red seed pericarp, as well as seed shattering and seed dormancy traits in diverse weedy rice populations (Gealy Reference Gealy and Gressel2005).
Weedy rice is one of the most troublesome weeds (Roma-Burgos et al. Reference Roma-Burgos, San Sudo, Olsen, Werle and Song2021) affecting rice production in several countries such as China, Malaysia, the United States, and Thailand (Cao et al. Reference Cao, Lu, Xia, Rong, Sala, Spada and Grassi2006; Londo and Schaal Reference Londo and Schaal2007; Shivrain et al. Reference Shivrain, Burgos, Agrama, Lawton-Rauh, Lu, Sales, Boyett, Gealy and Moldenhauer2010; Sudianto et al. Reference Sudianto, Neik, Tam, Chuah, Idris, Olsen and Song2016). The occurrence of weedy rice in the same field as cultivated rice causes significant yield losses and reduces the grain quality of crop rice, because it has to compete for nutrients and other resources (Burgos et al. Reference Burgos, Norman, Gealy and Black2006; Dai et al. Reference Dai, Dai, Song, Lu and Qiang2014; Ratnasekera et al. Reference Ratnasekera, Perera, He, Senanayake, Wijesekara and Lu2014). Furthermore, weedy rice adds to production costs, due to it being of the same biological species as cultivated rice (O. sativa) and its removal and control are difficult when it infests rice fields, the removal and control of weedy rice are difficult. The strong seed shattering and seed dormancy cause weedy rice to remain in soil seedbanks in the field for long periods (Chauhan Reference Chauhan2013; Chin Reference Chin2001). In Thailand, an outbreak of weedy rice occurred in 1975 in the Songkhla, Nakhon Si Thammarat, Prachinburi, and Phitsanulok provinces, reducing rice yields by more than 80% (Maneechote et al. Reference Maneechote, Jamjod and Rerkasem2004). A further outbreak was reported in Kanchanaburi Province in 2001. Since 2002, weedy rice has developed into a seriously invasive weed in rice-growing areas, causing losses in yield ranging from 10% to 100%, depending on the level of infestation (Maneechote et al. Reference Maneechote, Jamjod and Rerkasem2004). Outbreaks of weedy rice have been attributed to the unintentional mixing of weedy rice seeds with cultivated rice seeds in farming equipment, especially combine harvesters (Pinglei et al. Reference Pinglei, Zhang, Sun, Yu and Qiang2018). Moreover, the increase in weedy rice infestation in Asia is related to the adoption of direct-seeded rice production, which increases the severity of the infestation (Chauhan Reference Chauhan2013). Such outbreaks are not a significant issue in transplanted rice culture, where weeding can be conducted through hand labor. Hand labor, however, may be limited due to water scarcity, rising labor costs, and the migration of the labor force to urban areas. As a result, planting methods have shifted significantly from transplanting to direct seeding (Chauhan Reference Chauhan2013).
Previous studies on the genetic diversity of weedy rice in Thailand suggest the existence of an outcrossing ability causing gene flow between weedy rice and cultivated rice, further establishing that weedy rice is a genetic admixture of O. rufipogon (Prathepha Reference Prathepha2009a, Reference Prathepha2011; Pusadee et al. Reference Pusadee, Schaal, Rerkasem and Jamjod2013). This indicates that weedy rice is generated by the hybridization between and coexistence of O. rufipogon and cultivated rice. The genetic diversity of weedy rice in Thailand has been reported to range from 0.598 (Prathepha Reference Prathepha2011) to 0.700 (Pusadee et al. Reference Pusadee, Schaal, Rerkasem and Jamjod2013). However, the genetic diversity of weedy rice populations from agricultural areas might be higher (Xia et al. Reference Xia, Wang, Xia, Zhao and Lu2011). Nonetheless, the limited geographic coverage in previous studies makes comparison of population structures difficult, as sample sizes per se are critical in evolutionary studies (Hobas et al. Reference Hobas, Gaggiotti and Bertorelle2013; Nazareno et al. Reference Nazareno, Bemmels, Dick and Lohmann2017). The high number of samples and diverse populations selected in certain areas represent accurate estimates of genetic diversity within weedy rice populations (Rosenberger et al. Reference Rosenberger, Schumacher, Brown and Hoban2021). The aim of this study, therefore, is to determine the genetic diversity and population structure of weedy rice throughout northeast Thailand. The basic information on genetic diversity, population structure, and origin of weed rice in northeast Thailand supports the prevention of weedy rice in future rice production.
Materials and Methods
Plant Materials and DNA extraction
This study used 380 samples of weedy rice collected from northeast Thailand in October 2020 (Figure 1; Supplementary Table S1). The weedy rice was examined during the flowering stage to make it easier to spot and differentiate the characteristics between weedy rice and cultivated rice based on its awn (Sudianto et al. Reference Sudianto, Neik, Tam, Chuah, Idris, Olsen and Song2016). The leaf samples of weedy rice were collected from individual plants in rice fields, abandoned fields, roadsides, canals, and marshes (Figure 2). The number of samples in each population was not equal, being dependent on the size of the weedy rice population and the different morphological characteristics identified. The collected leaves were kept in an icebox for 12 to 24 h and then maintained in a freezer at a temperature of −20 C until DNA extraction. Four rice check varieties were employed in this study: ‘KDML105’, ‘RD6’, ‘RD15’, and ‘RD22’ from the Roi Et Rice Research Center, as well as five wild rice species: Oryza granulata Nees, Oryza officinalis Wall., Oryza ridleyi Hook. f., O. nivara, and O. rufipogon from the Pathum Thani Rice Research Center.
Geographic distribution of the 380 samples of weedy rice collected from northeast Thailand. Number of weedy rice samples collected are shown within circles.
Habitats of weedy rice in northeast Thailand: (A) roadsides, (B) marshes, (C) canals, (D) abandoned fields, (E) paddy fields, and (F) buffalo swamps.
DNA was extracted from the leaves according to the modified cetyltrimethylammonium bromide (CTAB) method of Lodhi et al. (Reference Lodhi, Ye, Weeden and Reisch1994). The DNA concentrations were determined using a Nanodrop and adjusted to 25 ng μl−1 for simple sequence repeat (SSR) marker analysis (Promega, Madison, WI, USA).
SSR Marker Analysis
Initially, 90 SSR markers and five specific markers were chosen from previous studies (Cao et al. Reference Cao, Lu, Xia, Rong, Sala, Spada and Grassi2006; Gealy et al. Reference Gealy, Agrama and Eizenga2009; Jiang et al. Reference Jiang, Xia, Basso and Lu2012; Jing et al. Reference Jing, Jiang, Zhang, Zhai and Wan2008; Prathepha Reference Prathepha2003, Reference Prathepha2009a; Shivrain et al. Reference Shivrain, Burgos, Agrama, Lawton-Rauh, Lu, Sales, Boyett, Gealy and Moldenhauer2010; Sun et al. Reference Sun, Qian, Ma, Xu, Liu, Du and Chen2013; Swain et al. Reference Swain, Mohapatra, Roy, Swain, Singh, Meher, Dash, Rao and Subudhi2017; Vilayheuang et al. Reference Vilayheuang, Machida-Hirano, Bounphanousay and Watanabe2016; C Zhang et al. Reference Zhang, Zhu, Chen, Fan, Li, Lu, Wang, Yu, Yi, Tang, Gu and Liu2019; L Zhang et al. Reference Zhang, Zhu, Wu, Ross-Ibarra, Gaut, Ge and Sang2009, Reference Zhang, Dai, Wu, Song and Qiang2012; Zhu et al. Reference Zhu, Si, Wang, Jingjie, Shangguan, Lu, Fan, Li, Lin, Qian, Sang, Zhou, Minobe and Han2011) (Supplementary Table S2), due to the high level of polymorphism observed in the weedy rice population. The markers were screened for polymorphism in four weedy rice samples, WD032-1, WD046-2, WD060-2, and WD082-2, collected from different regions with diverse characteristics (Supplementary Figure S1). A polymerase chain reaction (PCR) mixture of 15 μl contained 1 ng of genomic DNA, 5 pmol of each forward and reverse primer, 5× buffer, 10 mM dNTPs, 25 mM MgCl2, and 5 U Taq DNA polymerase (Thermo Fisher Scientific, Invitrogen, Waltham, MA USA). The PCR components were conducted via a TProfessional Basic 96 Gradient Thermocycler (Biometra, Analytik Jena, Gmbh, Göttingen, Germany). The thermocycler profile was programmed to 94 C for 5 min, followed by 35 cycles at 94 C for 30 s, 55 C for 30 s, 72 C for 1 min, with the final extension at 72 C for 7 min. The PCR products were separated on 4% polyacrylamide denaturing gel. Electrophoresis was run at 70 W constant power for 1.5 h using the Model S2 Sequencing Gel Electrophoresis Apparatus (Biometra, Analytik Jena, Gmbh, Göttingen, Germany). The ϕX174 DNA/Hinf1 marker (Promega, Madison, WI, USA) was used to represent the standard size of the SSR alleles. The PCR products were visualized through silver staining.
Data Analysis
The genetic diversity and population structure of the samples were calculated based on the SSR allele data. The genetic parameters of the marker, consisting of the number of alleles (N A), observed heterozygosity (H O), and gene diversity (expected heterozygosity, H E) were calculated via FSTAT 2.9.3.2 (Goudet Reference Goudet2002). The polymorphism information content (PIC) of each marker was calculated following Anderson et al. (Reference Anderson, Churchill, Autrique, Tanksley and Sorrells1993). The population structures of the 380 weedy rice accessions were analyzed via STRUCTURE 2.3.4 (Pritchard et al. Reference Pritchard, Stephens and Donnelly2000). Initially, 20 simulation runs were performed with assumed populations (K) ranging from 1 to 10 and a burn-in period of 10,000 to 50,000 replicates of the Bayesian Markov chain Monte Carlo (MCMC) algorithm. Afterward, the simulation outputs were gathered to estimate the optimum K using the ad hoc ΔK method described by Evanno et al. (Reference Evanno, Regnaut and Goudet2005). Finally, STRUCTURE analyses with true K and a burn-in period of 100,000 and 500,000 replicates of the MCMC algorithm were conducted to assign individual weedy rice accessions to clusters. Additionally, the genetic parameters for each subpopulation and region, namely N A , H O, H E, Wright’s fix index (FIS), and allelic richness (AR), were calculated via FSTAT 2.9.3.2 (Goudet Reference Goudet2002). The FIS was utilized to calculate the outcrossing rates (t) through the equation proposed by Weir (Reference Weir1996): t = (1 − FIS)/(1 + FIS). Nei’s genetic distance (D A) was calculated using POPULATIONS 1.2.32 (Langella Reference Langella1999), which assessed the genetic relationships between the 380 weedy rice accessions. The distance matrix was then submitted to XLStat for principal coordinate analysis (PCoA) (Addinsoft, USA) and MEGA 6.0 (Tokyo, Japan) (Tamura et al. Reference Tamura, Stecher, Peterson, Filipski and Kumar2013) for neighbor-joining (NJ) analysis.
Results and Discussion
SSR Polymorphism
Of the 90 SSR markers and five specific markers screened in the four samples of weedy rice, 42 (39 SSR and three specific markers) (43.29%) demonstrated polymorphism (Supplementary Table S3). Of these, 31 markers were selected, including 28 SSR, spread throughout the 12 chromosomes, with three specific markers revealing polymorphism and concise DNA bands. Further analyses were conducted on the DNA in all 380 weedy rice samples (Table 1). The 31 markers detected 213 alleles in total, with an average of 6.87 alleles per locus. RM481 and RM228 showed the highest number of alleles (11). H O ranged from 0.000 (BADH2) to 0.196 (RM21). H E was between 0.255 (RID12) and 0.843 (RM228), with an average of 0.723. The PIC values varied from 0.256 (RID12) to 0.841 (RM228), with an average of 0.721. Most of the markers had high PIC values except RID12, RM7158, and Wx, as shown in Table 1. This established that the primers in this study were effective in distinguishing between the weedy rice samples. Nearly all primers had a PIC value greater than 0.70 (Table 1). The high PIC values of the primers employed in this study were selected from highly polymorphic primers in previous studies that demonstrated the ability to distinguish between samples (Cao et al. Reference Cao, Lu, Xia, Rong, Sala, Spada and Grassi2006; Gealy et al. Reference Gealy, Agrama and Eizenga2009; Jiang et al. Reference Jiang, Xia, Basso and Lu2012; Jing et al. Reference Jing, Jiang, Zhang, Zhai and Wan2008; Prathepha Reference Prathepha2003, Reference Prathepha2009a; Shivrain et al. Reference Shivrain, Burgos, Agrama, Lawton-Rauh, Lu, Sales, Boyett, Gealy and Moldenhauer2010; Sun et al. Reference Sun, Qian, Ma, Xu, Liu, Du and Chen2013; Swain et al. Reference Swain, Mohapatra, Roy, Swain, Singh, Meher, Dash, Rao and Subudhi2017; Vilayheuang et al. Reference Vilayheuang, Machida-Hirano, Bounphanousay and Watanabe2016; C Zhang et al. Reference Zhang, Zhu, Chen, Fan, Li, Lu, Wang, Yu, Yi, Tang, Gu and Liu2019; L Zhang et al. Reference Zhang, Zhu, Wu, Ross-Ibarra, Gaut, Ge and Sang2009, Reference Zhang, Dai, Wu, Song and Qiang2012; Zhu et al. Reference Zhu, Si, Wang, Jingjie, Shangguan, Lu, Fan, Li, Lin, Qian, Sang, Zhou, Minobe and Han2011). Moreover, this study employed both general and specific primers. In a comparison of average PIC values, the general SSR primers produced higher polymorphism than the specific primers (Wx, BADH2, and RID12) (Table 1). Most of the specific primers had two alleles, resulting in low diversity (Table 1). However, the specific primers were able to distinguish between the traits of cultivated rice, wild rice, and weedy rice. Waxy and fragrant genes are present in cultivated rice through natural mutation and artificial selection for improved eating and cooking quality (Kovach et al. Reference Kovach, Calingacion, Fitzgerald and McCouch2009; Prathepha Reference Prathepha2009a; Tian et al. Reference Tian, Qian, Liu, Yan, Liu, Yan, Liu, Gao, Tang, Zeng, Wang, Yu, Gu and Li2009), whereas the red pericarp–colored gene originates from wild rice (Prathepha Reference Prathepha2009b; Sudianto et al. Reference Sudianto, Neik, Tam, Chuah, Idris, Olsen and Song2016). These primers can demonstrate the domestication or gene flow processes of several rice traits in weedy rice.
Number of alleles (N A), observed heterozygosity (H O), genetic diversity (H E), and the polymorphism information content (PIC) of 380 weedy rice samples from northeast Thailand detected by 28 simple sequence repeat (SSR) markers and three specific markers.
This study identifies the combination of alleles (both homozygous and heterozygous) within waxy, fragrant, and red pericarp–colored genes in the weedy rice population. This suggests that the outcrossing ability causes gene flow between weedy and cultivated rice, as well as between weedy rice and wild species coexisting in rice-growing areas (Bourgis et al. Reference Bourgis, Guyot, Gherbi, Tailliez, Amabile, Salse, Lorieux, Delseny and Ghesquière2008; Kovach et al. Reference Kovach, Calingacion, Fitzgerald and McCouch2009). Prathepha (Reference Prathepha2009a) found the badh2 allele of the fragrant gene in the weedy rice of northeast Thailand, which is likely to reflect the introgressions of the derived allele into weedy rice populations growing in the surrounding fields largely planted with fragrant cultivars. Especially in northeast Thailand, most of the rice production area is dominated by the fragrant rice cultivars RD6 and KDML105 (Thanasilungura et al. Reference Thanasilungura, Kranto, Monkham, Chankaew and Sanitchon2020). Furthermore, Zhang et al. (Reference Zhang, Zhu, Chen, Fan, Li, Lu, Wang, Yu, Yi, Tang, Gu and Liu2019) demonstrated that the waxy (Wx) gene in cultivated rice (Wx a, Wx b, and Wx In) explains the artificial selection and domestication of the Wx IV allele, which originated directly from wild rice through differentiation after gene mutation. This allele contributes to the improvement of eating and cooking quality. Therefore, the presence of the waxy allele in the weedy rice also indicates the outcrossing ability of gene flow between weedy rice and cultivated rice in a particular area (Muto et al. Reference Muto, Ishikawa, Olsen, Kawano, Bounphanousay, Matoh and Sato2016).
Cluster Analysis
STRUCTURE analysis was performed to estimate the population structure among weedy rice samples using the Bayesian algorithm. Based on Evanno’s ad hoc ΔK statistic (Evanno et al. Reference Evanno, Regnaut and Goudet2005) (Figure 3A), the weedy rice samples were grouped into two subpopulations (I and II) (Figure 3B). Subpopulation I was the largest, containing 260 samples (49, 132, and 79 samples from Central NE, Northern NE, and Southern NE, respectively). It should be noted that all wild rice samples were included in this subpopulation, except O. nivara. The smaller Subpopulation II comprised 120 samples (15, 17, and 88 samples from Central NE, Northern NE, and Southern NE, respectively). All cultivated rice samples and O. nivara were included in this subpopulation. The results indicated that the weedy rice sample grouping was independent of geographic origin.
STRUCTURE analyses: (A) Delta K values according to the method of Evanno et al. (Reference Evanno, Regnaut and Goudet2005) with modal value detecting a true K of the two populations (k = 2) based on 31 simple sequence repeat (SSR) markers analyzed in 380 samples of weedy rice from northeast Thailand; (B) two subpopulations of the 380 samples of weedy rice from northeast Thailand, four samples of cultivated rice, and five samples of wild rice determined through STRUCTURE analysis.
Molecular Genetic diversity of Weedy Rice
The number of alleles (N A), observed heterozygosity (H O), genetic diversity (H E), allelic richness (AR), fixation index (FIS), and outcrossing rate (t) values for Subpopulations I and II are shown in Table 2. Among the two subpopulations, Subpopulation II had higher H O and t values than Subpopulation I, whereas both subpopulations had similar N A, AR, and H E. The results further indicated that the weedy rice in Subpopulation II had higher heterozygosity than Subpopulation I, as confirmed by the greater outcrossing rate of Suppopulation II. The FIS value of the weedy rice in Subpopulation I was greater than that in Subpopulation II. It can, therefore, be inferred that Subpopulation I contained weedy rice samples with a high degree of self-hybridization.
Number of alleles (N A), allelic richness (AR), observed heterozygosity (H O), genetic diversity (H E), fixation index (FIS), and outcrossing rate (t) in weedy rice Subpopulations I and II from northeast Thailand identified via STRUCTURE.
The weedy rice samples 380 (64+149+167 samples) collected for Thailand’s Central NE, Northern NE, and Southern NE regions totaled 64, 149, and 167, respectively. The number of alleles (N A), observed heterozygosity (H O), genetic diversity (H E), allelic richness (AR), fixation index (FIS), and outcrossing rate (t) values of weedy rice from different regions are shown in Table 3. The N A and AR in the Central NE and Southern NE varied from 177 to 197 and 176.705 to 190.229, with averages of 189 and 184.957, respectively. The H O was between 0.082 in Northern NE and 0.134 in Southern NE, with a total average of 0.101. Genetic diversity was high (0.723) in all regions and highest in Southern NE at 0.714, while the FIS was similar in all regions, with a total value of 0.858. The outcrossing rate was relatively high for self-pollinated crops, ranging from 5.988% in Northern NE to 9.769% in Southern NE, with an average of 7.723%. The results suggest that weedy rice in Southern NE had higher genetic diversity than other regions. Weedy rice in Southern NE and Central NE produced a high outcrossing rate, which is extremely detrimental to the significant commercial rice production in these areas.
Number of alleles (N A), allelic richness (AR), observed heterozygosity (H O), genetic diversity (H E), fixation index (FIS), and outcrossing rate (t) in weedy rice samples from northeast Thailand.
a Northern NE contains Loei, Nong Bua Lam Phu, Udon Thani, Nong Khai, Bueng Kan, Sakon Nakhon, Nakhon Phanom, and Mudahan provinces; Central NE consists of Khon Kaen, Kalasin, Maha Sarakham, and Roi Et provinces; Southern NE contains Chaiyaphum, Nakhon Ratchasima, Buri Ram, Surin, Si Sa Ket, Yasothon, Amnat Charoen, and Ubon Ratchathani provinces.
This research produced relatively high genetic diversity for weedy rice (H E = 0.7), including a particularly high outcrossing rate in Southern NE and Central NE (Table 3). These findings align with those of Pusadee et al. (Reference Pusadee, Schaal, Rerkasem and Jamjod2013), who studied the population structure of the primary gene pool of O. sativa. This gene pool included wild rice (O. rufipogon Griffiths), cultivated rice (O. sativa L.), and weedy rice (O. sativa f. spontanea) using 12 SSR markers. The results demonstrate that weedy rice in these regions exhibits moderate to high genetic diversity (H E = 0.736) and relatively high outcrossing rates. The stigma exsertion of weedy rice, promoting the likelihood of outcrossing in nature between weedy and cultivated rice in the same area (Figure 4), created high t values (Tables 2 and 3). The rice genotype exhibiting this trait, together with male sterile ability, represents a breeding goal for the hybrid rice production system (Bakti and Tanaka Reference Bakti and Tanaka2019; Tan et al. Reference Tan, Wang, Luan, Zheng, Ni, Yang, Yang, Zhu, Zeng, Liu, Wang and Zhang2021, Reference Tan, Bu, Chen, Yan, Chang, Zhu, Yang, Zhan, Lin, Xiong, Chen, Liu, Liu, Wang and Zhang2022; Zou et al. Reference Zou, Zhao, Li, Zheng, Zhang, Sun, He, Pan, Liu and Fu2020). The genetic diversity revealed in this study is higher than that previously reported by Prathepha (Reference Prathepha2011) (H E = 0.598) and Cao et al. (Reference Cao, Lu, Xia, Rong, Sala, Spada and Grassi2006) (H E = 0.313). This may be attributed to the varying amounts of weedy rice in each region. The results demonstrate that the selection of a high number of samples from diverse populations throughout a particular area provides an accurate estimate of genetic diversity within the weedy rice population (Hobas et al. Reference Hobas, Gaggiotti and Bertorelle2013; Nazareno et al. Reference Nazareno, Bemmels, Dick and Lohmann2017; Rosenberger et al. Reference Rosenberger, Schumacher, Brown and Hoban2021).
The stigma exsertion of weedy rice and the likelihood of outcrossing in nature. C and W on arrows indicate cultivated and weedy rice, respectively.
Neighbor-joining Analysis
Nei’s genetic distance was used to construct the neighbor-joining trees of the 389 rice samples (380 weedy rice, five cultivated rice samples, and four wild rice samples) based on the geographic region, as shown in Figure 5A. The 389 rice samples were divided into two major clusters (I and II). Cluster I consisted of three samples from Southern NE (WD099-5, WD099-8, and WD099-10) and four wild rice samples (O. granulata, O. officinalis, O. ridleyi, and O. nivara) (Supplementary Figure S2). In contrast, Cluster II contained five subclusters (II-A, II-B, II-C, II-D, and II-E) and comprised 382 samples (64, 149, and 164 weedy rice + 4 cultivated + 1 wild samples from the Central NE, Northern NE, and Southern NE regions, respectively), including four cultivated rice types (KDML105, RD6, RD15, and RD22) and one wild rice species (O. rufipogon). Additionally, although most of the weedy rice from a single area was clustered together, several samples in Southern NE were separated in all subclusters. When a neighbor-joining tree based on the weedy rice classified by STRUCTURE analysis was constructed, the results revealed that weedy rice from the two subpopulations was distributed into two clusters (Figure 5B). As a result, different geographic regions were grouped into the same cluster and could not be separated geographically. Comparing the neighbor-joining trees of the construct based on geographic region with results using STRUCTURE, most weedy rice in Northern NE was classified in Subpopulation I (red circles). In contrast, most weedy rice in Central NE and Southern NE was classified in Subpopulation II (green circles).
Neighbor-joining tree of the 389 samples of weedy rice from northeast (NE) Thailand, four samples of cultivated rice, and five samples of wild rice based on allelic data of 31 simple sequence repeat (SSR) loci calculated by Nei’s genetic distance: (A) samples based on geographic regions; and (B) samples based on subpopulation through STRUCTURE analysis. (Subcluster I contained I-A, and Subcluster II consisted of II-A, II-B, II-C, II-D, and II-E.)
PCoA
According to the PCoA analysis, PC1 (9.63%), PC2 (5.78%), and PC3 (5.50%) combined accounted for 20.91% of the total variation. A 3D scatter plot of PC1, PC2, and PC3, based on different geographic regions, revealed that the distribution of all weedy rice samples overlapped, particularly those from Southern NE, which overlapped with other regions (Figure 6A). When clustering based on STRUCTURE analysis was combined with the PCoA scatter plot, the distribution of all weedy rice samples was clearly separated into two subpopulations, yet some samples from both subpopulations overlapped in the middle, as shown in Figure 6B. The results indicate that PCoA was unable to clearly separate weedy rice according to geographic region. In addition, the results of the PCoA were similar to those for the STRUCTURE analysis and genetic diversity using neighbor-joining analysis for the clustering of weedy rice (Figures 3, 5, and 6).
Distribution of 389 samples of weedy rice in northeast (NE) Thailand, four samples of cultivated rice, and five samples of wild rice on a scatter plot from principal coordinate analysis (PCoA) using Nei’s genetic distance: (A) samples based on geographic regions; and (B) samples based on subpopulation by STRUCTURE analysis.
Genetic diversity among the weedy rice populations in northeast Thailand is likely caused by a combination of factors: seed migration, seed contamination, introgression, natural hybridization with different rice varieties over time, and ability to adapt to specific geographic isolation. However, the results from the neighbor-joining trees indicate that weedy rice is genetically similar to cultivated rice and O. rufipogon (Figures 3 and 5). Pusadee et al. (Reference Pusadee, Schaal, Rerkasem and Jamjod2013) similarly reported that 12 weedy rice populations (two from the lower north, five from the central region, and five from northeast Thailand) were divided into two major clusters, revealing that weedy rice in the northeast region developed through a genetic admixture of cultivated rice varieties KDML105 and RD6, (and O. rufipogon), which are popular varieties in this region. This implies that weedy rice is generated from the hybridization and coexistence of O. rufipogon and cultivated rice. In contrast, Wongtamee et al. (Reference Wongtamee, Maneechote, Pusadee, Rerkasem and Jamjod2015) grouped Thailand’s weedy rice into three clusters, with weedy rice in the northeast showing a genetic relationship with the area’s popular cultivated rice varieties, KDML105, RD6, and RD15. The Wongtamee study further determined that weedy rice may have originated from the independent hybridization between native wild rice and popular cultivated rice varieties in that region. Interestingly, in some rice-growing areas of North and South America, weedy rice can develop outside the range of O. rufipogon (Londo and Schaal Reference Londo and Schaal2007).
According to genetic distance analysis, the clustering of weedy rice samples in northeast Thailand is not associated with the geographic region (Figures 5 and 6). The variation within populations strengthens the hypothesis of gene flow. Seed migration is one of the possible causes of gene flow. Farmers in northeast Thailand commonly use uncertified seeds of RD6, KDML105, and other cultivars, which are often contaminated with weedy rice. This scenario aligns with that reported by Ives et al. (Reference Ives, Tereza, Valmir and Aldo2014), who speculate that gene flow is related to the seed migration of uncertified seeds frequently contaminated with weedy rice by farmers in southern Brazil. The within-field and between-field dispersal of weedy rice is not only caused by farmers enabling seed migration but through the contamination of weedy rice seed via combine harvesters (Gao et al. Reference Gao, Zhang, Sun, Yu and Qiang2018). The harvesting time of cultivated rice, especially RD6 and KDML105, shows a gradual delay from the Northern to Southern NE in response to the hours of daylight and water status in the field (Sujariya et al. Reference Sujariya, Jongdee and Fukai2023). Therefore, the combine harvester can be rotated accordingly in these areas, and this scenario could explain the wide dispersal of weedy rice between fields during harvesting. According to the results of this study, some weedy rice samples from Central and Southern NE were clustered together with samples from Northern NE (Figures 5 and 6). In northeast Thailand, few rice cultivars are grown by farmers due to the two high-quality rice cultivars, KDML105 and RD6, dominating the market. Sowing takes place mostly from June to July, flowering occurs in mid-October, and the crop is harvested in mid-November (Sujariya et al. Reference Sujariya, Jongdee and Fukai2023). Both cultivars exhibit strong photoperiod sensitivity (Nettuwakul et al. Reference Nettuwakul, Pongtongkam, Thongpan and Peyachoknagul2007), yet the cross-ability with weedy rice (weak photoperiod sensitivity) may be limited due to its reproductive isolation (Li et al. Reference Li, Gui, Yu, Liang, Cui, Zhao, Zhang, Yu, Chen and Sun2022). In the northeast region, farmers have introduced several non–photoperiod sensitive rice cultivars, such as RD15, RD22, ‘RD41’, and ‘San-pah-tawng’. This situation creates sexual compatibility between weedy rice and cultivated rice (Song et al. Reference Song, Lu, Zhu and Chen2002) within the North NE, Central NE, and South NE regions, as confirmed by this research (Figure 7). In Thailand’s North NE region, the synchronization between cultivated and weedy rice was low (Figure 7A–C) because weedy rice, for example, enters the flowering stage much earlier than cultivated rice, and weedy rice flowers at the panicle initiation stage of cultivated rice (Figure 7A–C). However, there were diverse periods in Central NE when weedy rice and cultivated rice were flowering (Figure 7D–F). In contrast, flowering was synchronized for cultivated and weedy rice in the South NE area (Figure 7 G–I), further promoting the cross-ability between cultivated and weedy rice (Song et al. Reference Song, Lu, Zhu and Chen2002). An examination of the outcrossing rates confirmed that South NE produced higher t values than the other areas (Table 3).
The diverse synchronous flowering time of weedy rice and cultivated rice in northeast (NE) Thailand. The flowering stage of weedy rice was much earlier than cultivated rice in North NE (A–C), there were diverse periods of flowering in Central NE (D–F), and flowering was synchronized for cultivated and weedy rice in South NE (G–I) areas of northeast Thailand. C and W on arrows indicated cultivated and weedy rice, respectively.
Based on the stigma exsertion of weedy rice, a strong likelihood exists of outcrossing and high introgression between weedy rice and cultivated rice over time in the same area (Figure 4). These results demonstrate why some weedy rice samples mimic the morphological characteristics of cultivated rice, such as seed morphology and panicle architecture (Figure 8). In hybrid rice production, stigma exsertion is used to improve hybrid rice seed production to overcome the barrier of nonsynchronous flowering between the parents (Lou et al. Reference Lou, Yue, Yang, Mei, Luo and Lu2014; Takano-Kai et al. Reference Takano-Kai, Doi and Yoshimura2011; Yan et al. Reference Yan, Li, Agrama, Luo, Gao, Lu and Ren2009). The exserted stigma can stay viable and accept pollen for 6 d, enhancing each plant’s cross-ability (Yan et al. Reference Yan, Li, Agrama, Luo, Gao, Lu and Ren2009). In this study, the exserted stigma also demonstrated the potential to promote the gene flow between cultivated and weedy rice with a nonsynchronous flowering time (Figure 7). This phenomenon is of great concern, because it allows the escape of cultivated traits into weedy rice populations. Bakti and Tanaka (Reference Bakti and Tanaka2019) demonstrated that the stigma exsertion facilitates outcrossing to maintain the diversity of natural populations of wild and weedy rice, while this trait is reduced during rice domestication, because it disturbs the uniformity and stability of cultivated rice achieved by self-populating. In 2013, in the rice-farming areas of Arkansas, a southern state in the United States, herbicide-resistant rice varieties made up roughly 57% of the cultivated rice grown. Among the 26 fields sampled, weedy rice produced 10% to 60% of resistant offspring (Burgos et al. Reference Burgos, Singh, Tseng, Black, Young, Huang, Hyma, Gealy and Caicedo2014). Regrettably, this situation shows that herbicide resistance can be transferred from cultivated rice to weedy rice. Moreover, the similar morphology of weedy rice and cultivated rice represents a highly limiting factor in weedy rice management and may affect the production of high-quality rice in the future. Various post- and prezygotic barriers can affect the degree of gene flow occurring between cultivated and weedy rice, according to the respective cultivar (Reagon et al. Reference Reagon, Thurber, Gross, Olsen, Jia and Caicedo2010). This activity further minimizes the outcrossing rate or gene flow between weedy rice and cultivated rice. Therefore, it is necessary to consider the selection of rice varieties that are nonsexually compatible with the weedy rice in each region. In addition, farmers should be encouraged to use the certified seeds of rice cultivars without weedy seeds and clean combine harvesters to delay the gene flow of weedy rice (Gao et al. Reference Gao, Zhang, Sun, Yu and Qiang2018). Moreover, planting methods should shift from direct seeding to transplanting, allowing the weedy rice plants to grow between two rows of cultivated rice so they can easily be distinguished and removed by farmers during the vegetative stage (Chauhan Reference Chauhan2013).
The mimicry in morphological characteristics of weedy and cultivated rice, such as plant types (A), seed morphology(B), and panicle architecture (C) in northeast Thailand. C and W on arrows indicate cultivated and weedy rice, respectively.
In summary, the genetic diversity of weedy rice in northeast Thailand is relatively high due to its high outcrossing rate. Genetic diversity among weedy rice populations in northeast Thailand is likely due to a combination of factors. These include seed migration, seed contamination, introgression, natural hybridization with different rice varieties over time, and adaptation to specific geographic isolation. The weedy rice populations analyzed in this study are grouped into two clusters and genetically related to the cultivated rice and O. rufipogon. Notably, genetic clustering does not relate to other geographic regions. These results support the hypothesis that weedy rice in northeast Thailand originates from O. rufipogon, as an ancestor of cultivated rice, and hybridization or gene flow between cultivated rice and wild relatives. The sexual compatibility of weedy rice and cultivated rice must be addressed to prevent weedy rice rapidly developing the characteristics of cultivated rice.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/wsc.2023.60
Acknowledgments
This research was supported by the Research Fund for Supporting Lecturer to Admit High Potential Student to Study and Research on His Expert Program Year 2020, Graduate School, Khon Kaen University, Khon Kaen, Thailand. The authors would also like to thank the Plant Breeding Research Center for Sustainable Agriculture, Khon Kaen University, for providing plant materials and research facilities. The authors declare that no conflicts of interest exist.
