Phylogenetic analysis of AGAMOUS sequences reveals the origin of the diploid and tetraploid forms of self-pollinating wild buckwheat, Fagopyrum homotropicum Ohnishi.

Fagopyrum homotropicum Ohnishi is a self-pollinating wild buckwheat species indigenous to eastern Tibet and the Yunnan and Sichuan Provinces of China. It is useful breeding material for shifting cultivated buckwheat (F. esculentum ssp. esculentum Moench) from out-crossing to self-pollinating. Despite its importance as a genetic resource in buckwheat breeding, the genetic variation of F. homotropicum is poorly understood. In this study, we investigated the genetic variation and phylogenetic relationships of the diploid and tetraploid forms of F. homotropicum based on the nucleotide sequences of a nuclear gene, AGAMOUS (AG). Neighbor-joining analysis revealed that representative individuals clustered into three large groups (Group I, II and III). Each group contained diploid and tetraploid forms of F. homotropicum. We identified tetraploid plants that had two diverged AG sequences; one belonging to Group I and the other belonging to Group II, or one belonging to Group II and the other belonging to Group III. These results suggest that the tetraploid form originated from at least two hybridization events between deeply differentiated diploids. The results also imply that the genetic diversity contributed by tetraploidization of differentiated diploids may have allowed the distribution range of F. homotropicum to expand to the northern areas of China.

Introduction

Fagopyrum homotropicum Ohnishi is a self-pollinating wild species indigenous to eastern Tibet and the Yunnan and Sichuan Provinces of China (Ohnishi 2010). Since the discovery of F. homotropicum (Ohnishi 1998), several researchers have attempted to transfer the self-compatible property of F. homotropicum to cultivated common buckwheat (F. esculentum ssp. esculentum Moench) by conducting interspecific crosses between these two species (Hirose , Matsui , Woo ). These attempts have been quite successful and new self-pollinating varieties have been generated (Campbell 2003, Matsui ). Now, F. homotropicum is widely recognized as one of the most important genetic resources for buckwheat breeding.

Taxonomic studies of the genus Fagopyrum suggested that F. homotropicum probably differentiated from F. esculentum ssp. ancestrale Ohnishi (Ohnishi and Matsuoka 1996, Ohnishi and Asano 1999, Ohsako , Nishimoto , Yasui and Ohnishi 1998). Ohnishi and Asano (1999) revealed that F. homotropicum consists of morphologically indistinguishable diploid and tetraploid forms. Based on the observation of fixed heterozygosity at three allozyme loci, they also suggested that the tetraploid form arose from crosses between diploid F. homotropicum and F. esculentum ssp. ancestrale, which is also distributed in eastern Tibet and the Yunnan and Sichuan Provinces (Ohnishi 2010). However, because of the limited population number for the tetraploid form used in their study (three populations), the origin of tetraploid F. homotropicum is still under dispute. Ohnishi and Tomiyoshi (2005) discovered novel natural populations of F. homotropicum in the Nu river valley and adjacent to the Three River area (Sichuan-Yunnan-Tibet Three-Province connected area in the central part of Hengduan Mountains region) in China and found that diploid and tetraploid forms grow sympatrically in this area. Samples from new populations will provide useful experimental materials for research into the genetic relationships between the two forms of F. homotropicum.

Many nuclear genes are routinely used to study genetic diversity and phylogenetic relationships within a plant species (e.g., ALCOHOL DEHYDROGENASE for Arabidopsis thaliana, Innan and ACETYL-COA CARBOXYLASE, 3-PHOSPHOGLYCERATE KINASE and GRANULE-BOUND STARCH SYNTHASE I for tetraploid wheat, Takenaka ). In Fagopyrum, Nishimoto determined the nucleotide sequences of two nuclear genes, AGAMOUS (AG) and FLORICAULA/LEAFY (FLO/LFY), and indicated that a comparison of these sequences provides useful information about the phylogenetic relationships among Fagopyrum species. In the present study, we aimed to evaluate the amount of genetic diversity within F. homotropicum based on the nucleotide sequences of AG. Furthermore, we investigated the genetic diversity of F. esculentum ssp. ancestrale, which was used as an ancestral outgroup species. Finally, we formulated a hypothesis for the origin of the diploid and tetraploid forms of F. homotropicum.

Materials and Methods

Plant Materials

The samples examined are listed in Table 1 and their geographical locations are plotted in Fig. 1. Plant materials were maintained at Yunnan Agricultural University (Kunming, China), and accession numbers are presented in Table 1. Previous studies suggested that some diploid forms found in the C0129 and C0130 populations in Changbo of Sichuan Province have unique phenotypes, such as red-winged seeds (Ohnishi 2002, Tomiyoshi and Ohnishi 2004). Thus, we selected two samples (red-winged and ordinary types) from each of the two populations. Both diploid and tetraploid forms were selected from the populations in which diploid and tetraploid forms grow sympatrically (C0129, C0130 and C0452). In total, 16 and nine samples of diploid and tetraploid forms, respectively, covering the entire range of F. homotropicum, were considered in this study. Four samples of F. esculentum ssp. ancestrale were selected from four populations. For each sample, total DNA was isolated from the fresh leaf tissue of a representative individual as described by Escaravage .

Table 1

Samples of Fagopyrum species used in the present study

Species and sample name	Accession no.	Ploidy	Locality	Province	Population no. in Fig. 1	Sequence name in Fig. 2
F. homotropicum
9517_hom_2x	C9517	2x	Qiaotou	Yunnan	7	hom-2x-1
9519_hom_2x	C9519	2x	Yonsheng	Yunnan	5	hom-2x-2
9520_hom_2x	C9520	2x	Luding	Sichuan	2	hom-2x-3
9610_hom_2x	C9610	2x	Sigangping	Sichuan	3	hom-2x-4
9619_hom_2x	C9619	2x	Lijian	Yunnan	6	hom-2x-5
9620_hom_2x	C9620	2x	Songgang	Sichuan	1	hom-2x-6
0129_hom_2x	C0129	2x	Changbo	Sichuan	10a	hom-2x-7
0129_hom_2x_rwb	C0129	2x	Changbo	Sichuan	10a	hom-2x-8rw
0130_hom_2x	C0130	2x	Changbo	Sichuan	11a	hom-2x-9
0130_hom_2x_rwb	C0130	2x	Changbo	Sichuan	11a	hom-2x-10rw
0131_hom_2x	C0131	2x	Xuebo	Sichuan	9	hom-2x-11
0212_hom_2x	C0212	2x	Nidon	Yunnan	8	hom-2x-12
0252_hom_2x	C0252	2x	Mianning	Sichuan	4	hom-2x-13
0452_hom_2x	C0452	2x	Quzong	Tibet	13a	hom-2x-14
0445_hom_2x	C0445	2x	Ludebi	Tibet	12	hom-2x-15
9616_hom_4x	C9616	4x	Xiangcheng	Sichuan	16	hom-4x-1a, -1bc
0125_hom_4x	C0125	4x	Moduo	Sichuan	18	hom-4x-2
0129_hom_4x	C0129	4x	Changbo	Sichuan	10a	hom-4x-3a, -3bc
0130_hom_4x	C0130	4x	Changbo	Sichuan	11a	hom-4x-4a, -4bc
0210_hom_4x	C0210	4x	Zhongdian	Yunnan	14	hom-4x-5
0314_hom_4x	C0314	4x	Walong	Sichuan	15	hom-4x-6
0452_hom_4x	C0452	4x	Quzong	Tibet	13a	hom-4x-7a, -7bc
0444_hom_4x	C0444	4x	Ludebi	Tibet	12	hom-4x-8a, -8bc
2017_hom_4x	C2017	4x	Zhubalong	Sichuan	17	hom-4x-9
hom_2x_DDBJd	C9139	2x	Yonsheng	Yunnan	5	hom-2x-DDBJ
F. esculentum ssp. ancestrale
9135_anc	C9135	2x	Jinan	Yunnan	21	anc-1a, -1bc
9922_anc	C9922	2x	Guanmei	Yunnan	22	anc-2a, -2bc
2009_anc	C2009	2x	Yanjing	Tibet	19	anc-3
0203_anc	CK0203	2x	Adong	Yunnan	20	anc-4
anc_DDBJd	C9136	2x	Yonsheng	Yunnan	5	anc-DDBJ
F. cymosum
cym_DDBJd	C9911	4x	Yue Hua	Sichuan	—	cym-DDBJ
F. tataricum ssp. potanini
pot_DDBJd	C9029	2x	Maerkan	Sichuan	—	pot-DDBJ

Populations in which diploid and tetraploid forms grow sympatrically (see also Fig. 1).

Red-winged seed plants.

Two different clones of PCR products from five samples of tetraploid form of F. homotropicum and two samples of F. esculentum ssp. ancestrale were obtained. Thus, these samples have two AG sequences, denoted by “a” and “b” (see also Materials and Methods).

Sequence data obtained from DDBJ/EMBL/GenBank.

Fig. 1

Locations of sampling sites for plants used in the present study. Blue- and red-filled circles indicate sampling sites of diploid and tetraploid Fagopyrum homotropicum, respectively; circles filled in both red and blue indicate sites where diploid and tetraploid F. homotropicum coexist; stars indicate samples of F. esculentum ssp. ancestrale. Sample codes are given in Table 1.

PCR amplification, DNA cloning and sequencing

To amplify the AG gene, polymerase chain reaction (PCR) primers were constructed from the consensus sequences of F. homotropicum, F. esculentum ssp. ancestrale, F. cymosum Meisn. and F. tataricum ssp. potanini Batalin, based on AG sequences obtained from DDBJ/EMBL/GenBank (Nisimoto ). PCR amplification were conducted using Ex Taq Polymerase (Takara, Otsu, Japan). The primer sequences used for PCR amplification were 5′-GTCACCTTCTGCAAACGTAG-3′ and 5′-GATTGTTGTTGTGCAGTTCGATTTC-3′. The PCR conditions were as follows: denaturation at 94°C for 2 min; 30 cycles of 94°C for 30 s, 56°C for 30 s and 72°C for 3 min; and a final extension at 72°C for 5 min. PCR products were cleaned up using Mag-Extractor (Toyobo, Osaka, Japan). The sequences of purified PCR products were directly determined by a 3130 Genetic Analyzer (Applied Biosystems, Foster, California, USA). The two amplification primers and 5′-GAAAATCAACGACCCAGAA-3′ were used as sequencing primers. All samples of diploid F. homotropicum and two of F. esculentum ssp. ancestrale were successfully sequenced. However, in the case of tetraploid F. homotropicum samples and two F. esculentum ssp. ancestrale samples, double peaks were frequently found in the sequencing trace. Then, these PCR products were ligated into pGEM-T Easy vectors using the TA Cloning Kit (Promega, Madison, Wisconsin, USA) and two clones were randomly selected and sequenced for each sample. As F. homotropicum has a self-pollinating nature, it would be expected that tetraploid forms of F. homotropicum have homozygous alleles at the AG locus. Thus, two different sequences obtained from a single tetraploid plant were considered to be derived from homoeologous AG loci.

Data analysis

The nucleotide sequences examined were aligned using CLUSTALW (Thompson ) with visual corrections. Sequence data of F. esculentum ssp. ancestrale, F. cymosum and F. tataricum obtained from DDBJ/EMBL/GenBank were also aligned (Table 1). The gaps and ambiguously aligned regions were removed from the sequence data sets. The DNA sequences alignment matrix is available from the corresponding author upon request. NEXUS files of the aligned sequences were analyzed using DnaSP ver. 4 (Rozas and Rozas 1999) to estimate nucleotide diversity (Nei and Li 1979). Phylogenetic trees were constructed by the neighbor-joining method (Saitou and Nei 1987) using MEGA 4 (Tamura ). Genetic distances were calculated using the method of Tamura and Nei (1993) with the option of pairwise deletion. Bootstrap values were calculated from 1,000 replicates.

Data deposition

The DNA sequences have been deposited in the DDBJ/EMBL/GenBank DNA databases under accession numbers AB689706-AB689740.

Results

DNA polymorphisms in F. homotropicum and F. esculentum ssp. ancestrale

The consensus sequence of the 39 aligned AG sequences from F. homotropicum, F. esculentum ssp. ancestrale, F. tataricum and F. cymosum was 788 bp long, of which 720 and 68 bp were on the first intron and second exon, respectively. Within the all examined region (788bp), the nucleotide diversity of F. esculentum ssp. ancestrale was higher than that of diploid and tetraploid forms of F. homotropicum (Table 2). The high nucleotide diversity of F. esculentum ssp. ancestrale might reflect its obligate outcrossing nature, which is controlled by distylous self-incompatibility. On the other hand, the low nucleotide diversity of the tetraploid form of F. homotropicum might indicate that tetraploidization in F. homotropicum occurred recently (see also the next section).

Table 2

Nucleotide diversity of AG gene surveyed in this study

Species	no. of samples	no. of polymorphic site	Nucleotide diversity	SDa
F. homotropicum (2x)	16	46	0.0170	0.0019
F. homotropicum (4x)	14	33	0.0100	0.0020
F. esculentum ssp. ancestrale	7	38	0.0205	0.0032

Standard deviation of nucleotide diversity.

Phylogenetic relationships of F. homotropicum and F. esculentum ssp. ancestrale

The phylogenetic relationships among AG sequences from diploid and tetraploid forms of F. homotropicum and F. esculentum ssp. ancestrale were analyzed by the neighbor-joining method (Fig. 2). Three large phylogenetic groups (I, II and III) were detected with relatively high bootstrap values (≥ 60%). Within Group I and II, AG sequences from diploid and tetraploid forms of F. homotropicum were included, whereas no sequences from F. esculentum ssp. ancestrale were included. There were cases in Group I and II in which the AG sequence from the diploid form was identical to that from the tetraploid form; i.e., hom-2x-7 is identical to hom-4x-9 in Group I and hom-2x-14 is identical to hom-4x-6 in Group II. Taking into account the difficulty of crosses between diploid and tetraploid forms of F. homotropicum and F1 sterility between them (Wang ), gene flow between diploid and tetraploid forms may be rare events. Thus, these results suggest the recent tetraploidization of F. homotropicum in Group I and II. Group III contains AG sequences from F. esculentum ssp. ancestrale in conjunction with those from diploid and tetraploid forms of F. homotropicum. It is clear that AG sequences from F. esculentum ssp. ancestrale and F. homotropicum cannot be separated into different groups. For example, the AG sequence from F. homotropicum (hom-2x-13) is identical to that from F. esculentum ssp. ancestrale (anc-3), whereas the sequence from F. homotropicum (hom-2x-12) is clustered with those of F. esculentum ssp. ancestrale (anc-1b and anc-2b) with a high bootstrap value (89%). These results reveal the close relationship between the diploid form of F. homotropicum and F. esculentum ssp. ancestrale. Within Group III, only one sequence from tetraploid F. homotropicum (hom-4x-7b) is included. Clustering the sequence (hom-4x-7b) with those of the diploid form (hom-2x-3 and hom-2x-6) with a high bootstrap value (87%) indicates the close relationship among these three samples. Lastly, two sequences (hom-2x-8rw and hom-2x-10rw) from red-winged diploid F. homotropicum, which are locally distributed in the 0129 and 0130 populations (collection site #10 and #11 in Fig. 1) in Changbo of Sichuan Province, were also included in Group III, and clustered with the sequence (hom-2x-9) from the ordinary diploid F. homotropicum in the 0130 population (collection site #11 in Fig. 1) with a bootstrap value of 99%. The origin of the red-winged diploid F. homotropicum will be discussed in detail later.

Fig. 2

The phylogenetic tree of F. homotropicum and its relatives based on the AG gene. Phylogenetic analysis was constructed by the neighbor-joining method. cym-DDBJ, pot-DDBJ, anc-DDBJ and hom-2x-DDBJ were from DDBJ/EMBL/GenBank. Bootstrap values (≥ 50%) are shown above or below branches. Blue and red colored samples indicate diploid and tetraploid forms of F. homotropicum, respectively. The three large phylogenetic groups (I, II and III) are surrounded by a dashed border. Two-headed arrows indicate the AG sequences from one tetraploid plant of F. homotropicum that belong to separate groups.

Discussion

In this study, we aimed to infer phylogenetic relationships among diploid and tetraploid forms of F. homotropicum and F. esculentum ssp. ancestrale based on the comparison of AG sequences. Although only partial regions of the first intron and second exon of the AG gene were analyzed, sequences from these closely related species were sufficiently diversified to provide a well-supported neighbor-joining tree. Since the nucleotide sequences of AG genes of all Fagopyrum species have already been determined (Nishimoto ), a comparison of these sequences can be conducted to infer the phylogenetic relationships among other closely related species in the genus. However, one should be cautious of PCR errors that become apparent when PCR products are cloned. As PCR errors are thought to occur randomly, singleton mutations would be over-counted when conducting a phylogenetic analysis using cloned PCR products. In this study, we used cloned PCR products from the tetraploid forms of F. homotropicum and F. esculentum ssp. ancestrale. Thus, nucleotide diversity within these species would also be overestimated. On the other hand, as phylogenetic clustering largely depends on shared mutations among sequences, there is little chance that clusters supported in the present study were notably influenced by random PCR errors.

The genetic diversity and phylogenetic position of F. homotropicum within Fagopyrum have been analyzed in a number of studies (Nishimoto , Ohnishi and Matsuoka 1996, Ohnishi and Asano 1999, Ohsako , Yasui and Ohnishi 1998). These previous studies suggested that F. homotropicum has evolved from F. esculentum ssp. ancestrale through loss of distylous self-incompatibility. This suggestion has been strengthened by the recent findings on the candidate gene for distylous self-incompatibility in the lineage of F. homotropicum (Yasui ). In this study, we analyzed samples of F. homotropicum obtained from regions throughout its distribution range in China, and detected three large phylogenetic groups of F. homotropicum. Of the three groups, only Group III contains F. esculentum ssp. ancestrale and the diploid form of F. homotropicum. The observation that these two species are phylogenetically nested in Group III suggests the possibility of lineage sorting driven by the recent origin of F. homotropicum from F. esculentum ssp. ancestrale and/or the recent genetic introgression between these two species. The finding that these two species are close relatives will be of great use in buckwheat breeding programs. As mentioned above, it is recognized that F. homotropicum is suitable breeding material for generating self-pollinating lines of common buckwheat. Indeed, a self-pollinating variety of buckwheat (Norin-PL1) has been developed from a hybrid between F. homotropicum and common buckwheat, F. esculentum ssp. esculentum (Matsui ). It is expected that the diploid form of F. homotropicum in phylogenetic Group III will have high cross-compatibility not only with F. esculentum ssp. ancestrale but also common buckwheat which has normal crossability with F. esculentum ssp. ancestrale (Ohnishi 1999). Thus. diploid forms in phylogenetic Group III may permit the generation of a novel variety of self-pollinating buckwheat.

It is curious that none of the F. esculentum ssp. ancestrale samples clustered with Group I and II; i.e., the phylogenetic cluster of F. esculentum ssp. ancestrale is completely included in that of F. homotropicum. This result suggests that F. esculentum ssp. ancestrale evolved from F. homotropicum and conflicts with the previous hypothesis that F. homotropicum evolved from F. esculentum ssp. ancestrale. However, it seems unlikely that a self-incompatiblespecies, F. esculentumssp. ancestrale, would have evolved from a self-compatible species, F. homotropicum. Probably, populations of F. esculentum ssp. ancestrale that would be included in Group I and II have become extinct or have not yet been discovered. Recently, large populations of F. esculentum ssp. ancestrale were discovered around the Taocheng district in Shichuan Province (Ohnishi 2010). Phylogenetic analyses using new populations of F. esculentum ssp. ancestrale will elucidate the evolutionary pathway from F. esculentum ssp. ancestrale to the diploid form of F. homotropicum.

AG sequences from the tetraploid form of F. homotropicum clustered into three groups. It is noteworthy that two AG sequences from one tetraploid plant clustered into different groups; i.e., hom-4x-1b and hom-4x-1a clustered into Group I and II, hom-4x-8a and hom-4x-8b clustered into Group I and II, and hom-4x-7a and hom-4x-7b clustered into Group II and III, respectively (see also the AG sequences indicated by two-head arrows in Fig. 2). As these six sequences are closely clustered with sequences from diploid forms of F. homotropicum (Fig. 2), it is expected that these six sequences were derived from diploid forms of F. homotropicum. Combined, these results indicate that tetraploidization in F. homotropicum occurred after at least two hybridization events between deeply differentiated diploids forms of F. homotropicum; i.e., tetraploidization after hybridization between diploids forms of F. homotropicum belonging to Group I and II, and belonging to Group II and III. Including the tetraploidization between F. homotropicum and F. esculentum revealed by Ohnishi and Asano (1999), the tetraploid form of F. homotropicum has emerged at least three times. The tetraploid form is exclusively distributed in the northern areas of the Jinsha and Lancang river region (Tomiyoshi and Ohnishi 2004). A frequent (at least three times) supply of genetic diversity through tetraploidization between differentiated diploids might be important for expanding the distribution range to the northern areas of China.

In this study, we clarified the phylogenetic relationships among diploid and tetraploid forms of F. homotropicum and F. esculentum ssp. ancestrale based on a comparison of AG sequences. However, the origin of the red-winged diploid form of F. homotropicum remains unexplained. As mentioned above, the close relationships among red-winged and ordinary types of F. homotropicum might suggest that the red-winged type has simply mutated from the ordinary type. However, this hypothesis contradicts the highly differentiated morphological characters observed in the two types, such as flower color, number of nodes and vein pubescence (Tomiyoshi and Ohnishi 2004). We should therefore consider the possibility that hybridization occurred between differentiated diploid forms of F. homotropicum or between F. homotropicum and F. esculentum ssp. ancestrale. Additional phylogenetic analyses based on genomic regions other than the AG gene would provide clues for understanding the origin of the red-winged type of F. homotropicum. They are also required for confirming the phylogenetic relationship among and within F. homotropicum and F. esculentum ssp. ancestrale obtained by the current study, in which only a single nuclear gene (AG gene) was considered. In addition to the molecular phylogenetic studies presented here, cross-breeding experiments and subsequent cytological studies would provide insight into the evolution of F. homotropicum.