Breeding Science 55 : 7–14 (2005)

Genetic Diversity of Water Pepper ( hydropiper (L.) Spach) as Revealed by RAPD Markers

Kentaro Yasuda*1) and Hirofumi Yamaguchi

Conservation Ecology, Graduate School of Agriculture and Biological Sciences, Osaka Prefecture University, 1-1 Gakuencho, Sakai, Osaka 599-8531, 1) Present address: National Agricultural Research Center for Kyushu Okinawa Region, 496 Izumi, Chikugo, Fukuoka 833-0041, Japan

The genetic diversity and genetic relationships of three cultivated varieties (benitade, aotade and ayutade) and a wild race of water pepper, Persicaria hydropiper, were analyzed using RAPD markers. The proportion of polymorphic bands and Shannon’s index of phenotypic diversity within accession in five benitade and two aotade accessions were distinctly lower than those in six wild water pepper accessions, suggesting the pres- ence of a decline in genetic diversity within accession in the cultivated varieties under cultivation. A neighbor- joining (NJ) tree for nine cultivated and 38 wild accessions indicated the incidence of the independent domes- tication of the three cultivated varieties, benitade, aotade and ayutade, from wild water pepper. The NJ tree of the individual for five benitade accessions revealed the presence of close relationships among the accessions, but also showed a slight divergence in the accessions from different locations.

Key Words: Persicaria hydropiper, domestication, genetic diversity, RAPD, water pepper.

Introduction with salt. A number of “tade” farmers produce for condi- ment use in the suburbs of large cities in Japan Annual, seed-propagating domesticated plants general- (i.e. metropolitan area, Aichi Pref., Osaka Pref., and ly show a low degree of genetic diversity in their populations Fukuoka Pref., etc.). Tade farmers use the seeds generally compared to their wild ancestors. The low degree of genetic supplied by a particular seed company in each region. In diversity in the cultivated populations is generally caused by some cases, the presence of morphological diversity in the a bottleneck, random genetic drift, the limitation of gene ex- benitade variety has been recognized by consumers. For ex- change between populations, and selection under cultivation ample, the benitade plants from Osaka Prefecture display (Ladizinsky 1985, 1989, Frankel et al. 1995). rounder than those from Aichi Prefecture and from The water pepper Persicaria hydropiper (L.) Spach the Tokyo metropolitan area (Miyanohara and Yamaguchi (), one of the weedy and ruderal species com- 1995) although their genetic base is unknown. mon in temperate Eurasia, consists of three edible, cultivated Water pepper is diploid and usually bears chasmoga- varieties, benitade, aotade and ayutade, as well as of some mous and cleistogamous flowers. Although the chasmoga- ornamental varieties (Makino 1986). These three varieties mous flowers are frequently visited by pollinators such as are widely utilized for their strong flavor and for their straight swift and hoverfly (Hiratsuka 1984), the outbreed- deodorant effects on raw fish and similar Japanese dishes, ing rate has not been documented in wild water pepper, or in because they contain unique pungent constituents, tadeonal the cultivated varieties. The cleistogamous flowers set fruits and polygodial, in their leaves and stems (Tsuchiya et al. larger than those from chasmogamous flowers (Tsuchiya et 1989). Sprouts of the variety benitade are used as garnish for al. 1989). dishes of sliced raw fish, called sashimi (Fig. 1). Its sprouts RAPD (random amplified polymorphic DNA) analysis are bright red (“beni” means red in Japanese), and this col- is one of the most convenient methods for characterizing ge- oration is related to the present of anthocyanins and/or netic polymorphisms since it dose not require probe DNA cyaniding-3-galactoside (Miura et al. 1989). The green seed- and detailed information about the genomic and population lings of the other vegetable cultivated variety, aotade, are used polymorphism (Welsh and McClelland 1990, Williams et al. as garnish for dishes of sliced raw fish. In the cultivated vari- 1993, Hartl 1999). RAPD analysis has been used to investi- ety ayutade (Fig. 1), the young plants before flowering are used gate the genetic diversity and phylogenetic relationships as garnish for dishes of sweetfish (ayu in Japanese) broiled among populations in many cultivated species (Arias and Rieseberg 1995, Bonnin et al. 1996, Hormaza et al. Communicated by O. Ohnishi 1994, Mimura et al. 2000). In the present study, RAPD anal- Received September 30, 2003. Accepted September 2, 2004. ysis was used for the evaluation of the degree of genetic di- *Corresponding author (e-mail: [email protected]) versity in the populations of three cultivated water pepper 8 Yasuda and Yamaguchi

Fig. 1. Cultivated and wild water pepper plants. Upper left, Sprouts of benitade in a booth of a department store; Upper right, Benitade sprouts arranged as garnish for dishes of sliced raw fish (sashimi); Lower left, Ayutade in a department store; Lower right, Wild water pepper in the Mine river (Y36, Tsushima Island, Nagasaki Prefecture). varieties, benitade, aotade and ayutade, and its wild race, sample was collected a distance of at least 1 m away from the water pepper. Based on the RAPD data, their genetic rela- other sample. In the paddy fields, the wild water peppers tionship was examined. were collected from areas adjacent to paddy-levees. The samples which were transplanted into vinyl pots (9 cm in di- Materials and Methods ameter) were cultivated in the experimental fields of Osaka Prefecture University. About one month after transplanting, Plant materials and DNA extraction total DNA was extracted from fresh tissues of each indi- Studies on the diversity of cultivated plants generally vidual by the CTAB method (Doyle and Doyle 1987) with must cover a sufficient number of accessions of landraces, minor modifications. The extracted DNA was diluted to 10 local races distributed by seed companies and farmers. How- ng/µl and used for PCR amplification. The number of sam- ever local cultivation of tade is nearly extinct and the tade ples analyzed differed in the two kinds of evaluation de- seeds for crops are usually reproduced and distributed by a scribed in each section. limited number of seed companies in Japan. Moreover, since the disclosure of tade seeds is severely controlled by tade Evaluation of genetic relationships among the accessions of producers and seed companies, we used live materials from the three cultivated varieties and wild water pepper shops in the present study. All the samples of cultivated va- Forty-seven accessions, including 9 cultivated and 38 rieties, except for one accession (Be1), were purchased from wild water pepper accessions were used for this evaluation department stores in large cities (Table 1). In the department (Table 1). In each accession, one individual per accession stores, seedlings of the benitade and aotade varieties and was used except for two accessions, Be3 and Y10. In the Be3 young plants of the ayutade variety were packed in small and Y10 accessions, two individuals per accession were used. cases (Fig. 1) and sold. A small pack of each cultivated vari- Twenty-five RAPD primers (Table 2), which detected many ety was bought, and the authors asked a clerk where it was clear bands among the 60 RAPD primers preliminary tested produced. Samples of the Be1 accession were secured from (OPA1~A20, OPY1~Y20, OPZ1~Z20) were used. The 25 µl a cultivated field in Osaka Prefecture. Samples of 38 wild PCR reaction mixture contained 10 ng of template DNA, accessions were collected in natural habitats on paddy fields 0.5 µM of primer (Operon Technologies, CA, USA), 125 µM or riverbeds throughout Japan. About ten juvenile plants of of each of dATP, dCTP, dGTP and dTTP (TOYOBO, wild water pepper were sampled from each location. Each Osaka, Japan), 2.5 mM of Mg2+, 50 mM of KCl, 10 mM of Genetic diversity of vegetable water pepper 9

Table 1. Plant materials used in this study Location/area of production or natural Cluster group Race Site (obtained) Code population in Fig. 2 Benitade1) Osaka Cultivated field Be1 I Osaka Department store (Osaka, Osaka) Be2 I Aichi Department store (Osaka, Osaka) Be3a, Be3b I Saitama Department store (Chiyoda, Tokyo) Be4 I Hiroshima Department store (Yokohama, Kanagawa) Be5 I Fukuoka Department store (Ibaraki, Osaka) Be6 I Aotade Saitama Department store (Chiyoda, Tokyo) Ao1 IV Fukuoka Department store (Ibaraki, Osaka) Ao2 IV Ayutade Aichi Department store (Chiyoda, Tokyo) Ay II Water pepper Obihiro, Riverbed of Y1 VIII Chitose, Hokkaido Riverbed of Y2 V Aomori, Aomori Paddy field Y3 VIII Hachinohe, Aomori Paddy field Y4 VIII Hanamaki, Iwate Paddy field Y5 IX Kannari, Miyagi Paddy field Y6 IV Natori, Miyagi Paddy field Y7 VIII Date, Fukushima Paddy field Y8 VIII Sugakawa, Fukushima Riverbed of Y9 V Kamo, Niigata Riverbed of Y10a, Y10b IX, IX Mitsuke, Niigata Paddy field Y11 IX Kashiwazaki, Niigata Riverbed of U river Y12 IX Nou, Niigata Paddy field Y13 VIII Toyama, Toyama Riverbed of Jintsu river Y14 V Kaga, Ishikawa Riverbed of Iburihashi river Y15 II Imajo, Fukui Riverbed of Y16 II Utsunomiya, Tochigi Paddy field Y17 VII Ohmiya, Saitama Riverbed of Shiba river Y18 VII Tozuka, Kanagawa Riverbed of Kashiwao river Y19 IV Kutsuki, Shiga Riverbed of Ado river Y20 IX Sonobe, Kyoto Riverbed of Sonobe river Y21 III Kyoto, Kyoto Riverbed of Y22 IX Takatsuki, Osaka Riverbed of Y23 IX Matsubara, Osaka Riverbed of Y24 IV Itami, Hyogo Riverbed of Ina river Y25 VII Sanda, Hyogo Riverbed of Arita river Y26 IX Hashimoto, Wakayama Riverbed of Kino river Y27 IX Gobo, Wakayama Riverbed of Hidaka river Y28 IV Kozagawa, Wakayama Riverbed of Koza river Y29 II Shinjo, Okayama Riverbed of Shinjo river Y30 V Hiroshima, Hiroshima Riverbed of Seno river Y31 V Higashihiroshima, Hiroshima Riverbed of Nyuuno river Y32 III Hikari, Yamaguchi Riverbed of Shimada river Y33 II Sanyo, Yamaguchi Riverbed of Asa river Y34 V Kannonnji, Kagawa Riverbed of Saita river Y35 V Mine, Tsushima, Nagasaki Riverbed of Mine river Y36 V Uchiyama, Izuhara, Tsushima, Nagasaki Paddy field Y37 VI Tsutsu, Izuhara, Tsushima, Nagasaki Paddy field Y38 VI 1) Benitade accessions from Osaka Prefecture show fewer branches on their stems and rounder leaves than those from Aichi, Saitama, Hiroshima and Fukuoka Prefectures.

Tris-HCl (pH 8.3), and 0.5 units of rTaq DNA polymerase denaturation for 2 min at 94°C. The amplified products, (TOYOBO, Osaka, Japan). The reaction mixture was over- which were separated on 2% agarose gels in TAE buffer, laid with mineral oil and heated in a Program Temp Control were stained with ethidium bromide and visualized under System PC-800 (ASTEC Co.) programmed for 40 cycles of UV rays. To confirm accuracy the presence of the amplified 30 sec at 94°C, 30 sec at 37°C and 60 sec at 72°C after initial products, we carried out amplification twice and determined 10 Yasuda and Yamaguchi

Table 2. Number of polymorphic bands detected in the evaluation Evaluation of within-accession genetic diversity in benitade, of genetic relationships among the three cultivated varie- aotade and wild water pepper accessions and genetic rela- ties and wild water pepper by each of the 25 primers tionships among accessions within benitade variety No. of No. of Five benitade, two aotade and six wild water pepper Sequence of primer No. of Primer polymorphic monomorphic accessions were used for this evaluation. In each accession, (5′-3′) bands bands bands six to ten individuals were analyzed (Table 4). The 6 wild OPA021) TGCCGAGCTG 4 4 0 water pepper accessions were selected to cover a wide OPA04 AATCGGGCTG 3 1 2 geographical area extending from Hokkaido to Yamaguchi OPA07 CAATCGGGTG 4 0 4 Prefectures in Japan. Nine RAPD primers (Table 4) that OPA09 GGGTAACGCC 5 4 1 were randomly chosen among the 25 primers described in 1) OPA10 GTGATCGCAG 2 0 2 the previous section were used for the evaluation of diversi- OPA11 CAATCGCCGT 5 1 4 1) ty. Protocols of RAPD analysis and band scorings were the OPA17 GACCGCTTGT 2 2 0 same as those described in the previous section. Because the OPY02 CATCGCCGCA 4 1 3 RAPD bands are dominant markers, Shannon’s index of OPY05 GGCTGCGACA 6 3 3 OPY061) AAGGCTCACC 9 6 3 phenotypic diversity (Hs) was used to evaluate the level of OPY09 AGCAGCGCAC 4 1 3 within-accession diversity (Wolff and Morgan-Richards ∑ OPY111) AGACGATGGG 3 2 1 1999). Hs = − Pi ⋅ log2Pi, where Pi is the frequency of the ith OPY15 AGTCGCCCTT 4 2 2 band. The pattern of genetic relationships among accessions OPY16 GGGCCAATGT 2 0 2 within the benitade variety was investigated by employing OPY171) GACGTGGTGA 3 3 0 NJ cluster analysis based on the genetic distance matrix OPY18 GTGGAGTCAG 5 0 5 from the data for 9 primers in 45 individuals of five benitade 1) OPY20 AGCCGTGGAA 5 3 2 accessions (Be2, 3, 4, 5, 6) and nine individuals of the Y25 1) OPZ06 GTGCCGTTCA 3 1 2 accession which were collected from a site nearest to the OPZ08 GGGTGGGTAA 3 2 1 production site of a benitade variety in Osaka (Fig. 4). OPZ09 CACCCCAGTC 4 3 1 OPZ13 GACTAAGCCC 4 0 4 OPZ14 TCGGAGGTTC 3 0 3 Results OPZ18 AGGGTCTGTG 5 3 2 OPZ19 GTGCGAGCAA 4 1 3 Genetic relationships among three cultivated varieties and OPZ201) ACTTTGGCGG 10 8 2 wild water pepper Total 106 51 (48.1 %) 55 (51.9 %) Twenty-five RAPD primers detected 106 different

1) RAPD primers used in the evaluation of population genetic bands in 49 individuals of 47 accessions (Table2). The num- diversity. ber of bands per primer ranged from 2 to 10, with a mean value of 4.24. Fifty-one bands (48%) were polymorphic in the 49 individuals. The genetic distance between pairs of the the presence of a band, which represented the same band in nine accessions of the three cultivated varieties and 38 ac- all two replications. cessions of wild water pepper was variable, ranging from 0 RAPD profiles were scored based on the presence (1) to 0.126. The genetic distance was longest between the Ao1 and absence (0) of each band amplified in each individual. accession of aotade and Y7 accession of the wild water pep- The genetic distance (D = 1-similarity coefficient (Nei and Li per. A relatively high genetic divergence was observed be- 1979)) for all the pairs of individuals was calculated by the tween the benitade and aotade accessions (0.073–0.096), following formula: D = 1 − 2Nab/(Na + Nb), where Nab is the while intermediate genetic divergence was noted between number of bands shared by individuals “a” and “b”, Na = num- the benitade accessions and the ayutade accession (0.046– ber of bands scored in individual “a”, and Nb = number of 0.080). A neighbor-joining (NJ) tree based on these genetic bands scored in individual “b” using RAPDistance Programs distances enabled to identify nine large clusters, I~IX (Armstrong et al. 1996). A neighbor-joining (NJ) tree (Fig. 2) (Fig. 2), although the groups showed relatively low boot- was constructed based on the genetic distance matrix (Saitou strap values. Two large clusters, II and IV, contained both and Nei 1987) using MEGA software (Kumar et al. 1993). A wild and cultivated accessions, whereas one large cluster, I, bootstrap analysis was performed using 100 resampled data contained only cultivated benitade accessions. The other six sets generated with Seqboot program in the PHYLIP pack- large clusters, III, V, VI, VII, VIII and IX, contained only age (Felsenstein 2002) prior to the calculation of the distance wild water pepper accessions. Cluster I showed a bootstrap matrices. Consense program in the PHYLIP package was value of 45%. Three subclusters were recognized within used to construct the consensus tree. To test the geographical cluster I. The subcluster composed of Be1 and Be2 was sup- distribution pattern of the wild water pepper, Pearson’s cor- ported by a high bootstrap value (100%). Two aotade acces- relation coefficient between the genetic pair-wise distance sions belonged to cluster IV which included wild accessions of wild water pepper individuals and pair-wise geographic from island in Japan. The cluster formed by the two distance of sampling locations was calculated (Fig. 3). aotade accessions was supported by a high bootstrap value Genetic diversity of vegetable water pepper 11

Fig. 2. Neighbor-joining tree of ten cultivated and 39 wild water pepper individuals using 106 RAPD markers. Code of cultivated individuals, see Table 1; Code surrounded by a circle, cultivated va- riety; Scale bar, unit of genetic distance; Numerals below branch denote a bootstrap percentage higher than 40 %.

Genetic diversity of benitade, aotade and wild water pepper accessions Nine RAPD primers detected 40 different bands in 116 individuals of five benitade, two aotade and six wild water pepper accessions. Twenty-six bands (65%) were polymor- phic (Table 3). The total number of polymorphic bands in the six wild water pepper accessions was far larger than that in the five benitade and two aotade accessions, although the face values listed in Table 3 cannot be compared directly be- cause of the difference in sample size in benitade, aotade and wild water pepper. In the benitade accessions, except for the Aichi accession, and in the aotade accessions, the proportion of polymorphic bands and Shannon’s index of diversity was Fig. 3. Correlation between genetic distance and geographic distance low (Table 4). In contrast, the wild water pepper accessions in all the pairs of wild water pepper individuals. showed a high level of polymorphic bands and a high Shannon’s index of diversity except for one accession from (83%). The ayutade accession was included in cluster II with Hokkaido. Most of benitade and aotade accessions showed a the wild accessions from Honshu island in Japan. There was lower genetic diversity within accession than the wild water a very low correlation (r = 0.125, P < 0.001) between the pair- pepper accessions. wise genetic distance of water pepper individuals and the pair-wise geographic distance of their locations (Fig. 3), in- Genetic relationships among accessions within benitade dicating their ambiguous geographical variation pattern. variety A neighbour-joining tree generated from 40 RAPD bands (Table 3) enabled to identify a large cluster of benitade 12 Yasuda and Yamaguchi

Table 3. Number of polymorphic bands detected in 45 benitade individuals from 5 accessions, 16 aotade individuals from 2 accessions and 55 water pepper individuals from 6 accessions Benitade (45 individuals from 5 Aotade (16 individuals from 2 Wild water pepper (55 All samples1) accessions) accessions) individuals from 6 accessions) Primer No. of No. of No. of No. of No. of bands polymorphic No. of bands polymorphic No. of bands polymorphic No. of bands polymorphic bands bands bands bands OPA02 4 4 4 2 4 1 4 4 OPA10 2 0 2 0 2 0 2 0 OPA17 2 2 2 2 2 1 2 2 OPY06 9 4 7 0 7 0 9 4 OPY11 3 1 3 1 3 0 3 1 OPY17 3 3 2 2 1 0 3 3 OPY20 5 4 5 3 4 2 5 4 OPZ0631313031 OPZ2097835085 Total 40 26 (65.0 %) 36 14 (38.9 %) 31 4 (12.9 %) 39 24 (61.5 %) Average 4.44 2.89 4.00 1.56 3.44 0.44 4.33 2.67 1) 116 individuals: a total of 45 individuals of benitade, 16 individuals of aotade and 55 individuals of wild water pepper.

Table 4. Within-accession diversity in benitade, aotade and wild water pepper accessions for 9 primers No. of Shannon’s index of Location (production site Proportion of Race Code Sample size No. of bands polymorphic phenotypic or natural site) polymorphic bands bands diversity Benitade Osaka Be2 10 28 3 0.107 0.606 Aichi Be3 9 33 9 0.273 1.927 Saitama Be4 10 31 5 0.161 1.440 Hiroshima Be5 10 29 3 0.103 0.801 Fukuoka Be6 6 29 0 0.000 0.000

Aotade Saitama Ao1 10 31 2 0.065 0.664 Fukuoka Ao2 6 30 3 0.100 1.040

Wild water pepper Chitose, Hokkaido Y2 8 31 6 0.194 1.342 Kannari, Miyagi Y6 9 35 14 0.400 3.374 Kamo, Niigata Y10 10 33 11 0.333 2.894 Matsubara, Osaka Y24 9 33 9 0.273 2.573 Itami, Hyogo Y25 9 35 12 0.343 3.600 Hikari, Yamaguchi Y33 10 34 9 0.265 2.578 varieties different from the wild water pepper accession high degree of genetic diversity compared to the Fukuoka, from Hyogo Prefecture (Fig. 4). The largest cluster which Hiroshima and Osaka benitade accessions. contained all the benitade individuals was divided into three subclusters, A, B and C. The A subcluster contained all the Discussion individuals of the Saitama, Hiroshima and Fukuoka benitade accessions. The cluster formed by the Fukuoka benitade was The vegetable water pepper varieties were convention- supported by a high bootstrap value (90%). The B subcluster ally treated as infraspecific taxa under P. hydropepper by contained all the Aichi benitade individuals. The C subclus- Makino (1940), although their genetic relationship had not ter contained all the benitade individuals of the Osaka acces- been elucidated. Our results revealed that ca. 52% of the sion and was supported by a high bootstrap value (88%). bands detected by 25 RAPD primers were common to beni- The other group consisting of wild accessions showed a high tade, ayutade, aotade and wild water pepper (Table 2) and variability. All the individuals of the Fukuoka benitade ac- that the three cultivated varieties showed a close affinity to cession showed an identical RAPD phenotype (Fig. 4). Eight one of the wild water pepper accessions with a relatively individuals of each of Hiroshima and Osaka benitade ac- short genetic distance ranging from 0.044 to 0.126 (Fig. 2). cessions also displayed identical RAPD phenotypes. The The genetic distance within species based on RAPD data Saitama and Aichi benitade accessions showed a relatively usually falls into a value ranging from 0 to 0.5 (Pistacia Genetic diversity of vegetable water pepper 13

local populations, following the rapid expansion into Japan after the last glacial period. Further studies should be carried out to identify their direct donors. Although Zohary (1999) and Frankel et al. (1995) suggested simultaneous crop do- mestication at many locations in seed-propagated plant spe- cies can not be easily achieved, parallel and simultaneous domestication is assumed to occur in different areas in this species as well as in common bean (Koenig and Gepts 1989) and foxtail millet (Fukunaga et al. 2002). The NJ trees of Figures 2 and 4 indicated the presence of a slight genetic divergence among the benitade acces- sions. This suggests that specialization of a particular variety for consumer demand is proceeding for the local benitade varieties at both genetic and morphological levels after benitade was brought into cultivation at one particular loca- tion. For example, the benitade variety from Osaka showed fewer branching stems and rounder leaves than the materials from other production areas (e.g. Tokyo, Saitama) (Miyanohara and Yamaguchi 1995). Uniform sprouts with rounder coty- ledon and without adventitious root formation at the cutting edge of the hypocotyl were preferred by a Japanese restau- rant in Osaka. We analyzed the genetic diversity within accession in the cultivated varieties by using the samples purchased from department stores because of the constraint described in the introduction section. Due to the small sample of the cultivat- ed accessions in our results (Table 4), the genetic diversity within each cultivated accession may have been underesti- mated. However, the degree of genetic diversity within ac- cession in the benitade and aotade varieties was significantly lower than that of the wild water pepper accessions (Table Fig. 4. Neighbor-joining tree of 45 individuals of 5 benitade acces- 4). Although the benitade samples from Saitama and Aichi sions and nine individuals of one wild water pepper accession showed a high diversity compared with the uniform samples using 40 RAPD markers. A, B, C, subclusters. Benitade indi- from Fukuoka, Hiroshima and Osaka (Fig. 4 and Table 4), vidual is listed according to the production location; water pep- the levels of diversity of the Saitama and Aichi samples was per individual from Y25 population is denoted by a black cir- still lower than those of wild accessions. Since seeds to be cle. Benitade individual with a code number is the same sold to tade farmers are well mixed in a seed bag, it is ex- sample as that in Figure 2. Scale bar denotes the genetic dis- pected that genetic diversity evaluated here reflects correctly tance. Numerals below a branch denote a bootstrap percentage the level in a cultivated population. higher than 40 %. The benitade, aotade and ayutade seeds are usually pro- duced, released and sold to tade farmers through very closed vera, Hormaza et al. 1994; walnut, Nicese et al. 1998; guild processes from a few seed companies. The benitade Ipomoea batatas, Sagredo et al. 1998; some species in the and aotade farmers generally scatter the seeds in culture-soil genus Passiflora, Fajardo et al. 1998). Our data support the beds for production of young sprouts. The sprouts less than hypothesis (Makino 1986, Aoba 1991) formulated by the 1cm in height packed in a small case are sold at department taxonomists who consider that vegetable water pepper vari- booths and in particular markets. The remand from tade con- eties might have been domesticated from the wild water pep- sumers, such as high-class Japanese style restaurants, for per. sprouts with a uniform shape is very high. The genetic sepa- The NJ tree (Fig. 2) indicated that the clusters of the ration among benitade accessions shown in Figures 2 and 4 three cultivated varieties, benitade, aotade and ayutade, were may reflect the present closed guild process of a few seed genetically different from each other, suggesting that their companies for benitade marketing. For example, it is gener- origin was independent that of different wild ancestral pop- ally suggested that the Osaka benitade (Be2) sample in the ulations. However, the wild water pepper did not show a department store originated from the Be1 farmer’s field geographical structure of genetic variation (Fig. 2, Fig. 3 and (Table 1) or from the same seed lot of a seed company. Ge- Table 1), which may be partly ascribed to its metapopulation netic variation of a particular local variety was significantly structure with frequent extinction and re-colonization of the low, although some local varieties showed a certain degree 14 Yasuda and Yamaguchi of variation (Table 4). Our results strongly indicate that the University, University Park, PA 16802. genetic diversity of a cultivated plant population is affected Ladizinsky, G. (1985) Founder effect in crop-plant evolution. Econ. by a bottleneck, random genetic drift, the limitation of gene Bot. 39: 191–199. exchange between populations, and selection under cultiva- Ladizinsky,G. (1989) Ecological and genetic considerations in collect- tion (Ladizinsky 1985, 1989, Frankel et al. 1995). ing and using wild relatives. In “The Use of Plant Genetic Re- sources” Brown, A.H.D. et al. (eds.), Cambridge Univ. Press, New York. p. 297–308. Acknowledgments Makino, T. (1940) An illustrated flora of Nippon, with the cultivated and naturalized plants. Hokuryukan. p. 617 (in Japanese). The authors thank Dr. Nakayama, Mr. Ito, Mr. Michisita, Makino, T. (1986) Makino’s Illustrated Flora in Colour. Hokuryukan. Mr. Miyagaki and Mr. Sakaino for collecting the wild water p. 53 (in Japanese). pepper accessions. Mimura, M., K. Yasuda and H. Yamaguchi (2000) RAPD variation in wild, weedy and cultivated azuki beans in Asia. Genet. Res. Literature Cited Crop Evol. 47: 603–610. Miura, H., M. Shimizu, A. Tazuke and M. Iwata (1989) Effect of mono- Aoba, T. (1991) Histories of Vegetables in Japan. Yasakashobo, chromatic light on anthocyanin content in seedlings of benitade Tokyo. Japan. p. 249–254 (in Japanese). ( hydropiper L.). J. Jap. Soc. Hort. Sci. 58: 123– Arias, D.M. and L.H. Rieseberg (1995) Genetic relationships among 129. domesticated and wild sunflowers (Helianthus annuus, Aster- Miyanohara, J. and H. Yamaguchi (1995) Domestication syndrome in aceae). Econ. Bot. 49: 239–248. benitade Persicaria hydropiper (L.) Spach. Jpn. J. Breed. 45 Armstrong, J., A. Gibbs, R. Peakall and G. Weiler (1996) The RAPDis- (Suppl. 2): 302 (in Japanese). tance Package. Version 2.00. Australian National University, Nei, M. and W.H. Li (1979) Mathematical model for studying genetic Canberra, Australia. variation in terms of restriction endonucleases. Proc. Natl. Bonnin, I., T. Huguet, M. Gherardi, J.M. Prosperi and I. Olivieri (1996) Acad. Sci. USA 76: 5269–5273. High level of polymorphism and spatial structure in selfing Nicese, F.P., J.I. Hormaza and G.H. McGranahan (1998) Molecular char- plant species, Medicago truncatula (Leguminosae), shown us- acterization and genetic relatedness among walnut (Juglans ing RAPD markers. Am. J. Bot. 83: 843–855. regia L.) genotypes based on RAPD markers. Euphytica 101: Doyle, J.J. and J.L. Doyle (1987) A rapid DNA isolation procedure for 199–206. small quantities of fresh leaf tissue. Phytochem. Bull. 19: 11– Sagredo, B., P. Hinrichsen, H. López, A. Cubillos and C. Muñoz (1998) 15. Genetic variation of sweet potatoes (Ipomoea batatas L.) culti- Fajardo, D., F. Angel, M. Grum, J. Tohme, M. Lobo, W.M. Roca and vated in Chile determined by RAPDs. Euphytica 101: 193– I. Sanchez (1998) Genetic variation analysis of the genus 198. Passiflora L. using RAPD markers. Euphytica 101: 341–347. Saitou, N. and M. Nei (1987) The neighbor-joining method: A new Felsenstein, J. (2002) PHYLIP: Phylogeny Inference Package, Version method for reconstructing phylogenetic trees. Mol. Evol. 4: 3.6. Department of Genome Sciences, University of Washing- 406–425. ton, Seattle, Washington, USA. Tsuchiya, K., K. Hoshikawa, E. Asayama, M. Hotta, A. Nitta, K. Arai, Frankel, O.T., A.H.D. Brown and J.J. Burdon (1995) The Conservation M. Kawashima and K. Takahashi (1989) Polygonum L. In of Plant Biodiversity. Cambridge Univ. Press, New York. 299 p. “Useful Plant of the World” Hotta, M., K. Ogata, A. Nitta, Fukunaga, K., Z. Wang, K. Kato and M. Kawase (2002) Geographical K. Hoshikawa, M. Yanagi and K. Yamazaki (eds.), Heibonsha, variation of nuclear genome RFLPs and genetic differentiation Tokyo. p. 835–841. in foxtail millet, Setaria italica (L.) P. Beauv. Genet. Res. Crop Welsh, J. and M. McClelland (1990) Fingerprinting genomes using Evol. 49: 95–101. PCR with arbitrary primers. Nucleic. Acids Res. 18: 7213– Hartl, D.L. (1999) A Primer of Population Genetics. Sinauer Associ- 7218. ates, Inc., Sunderland (Third edition). 221 p. Williams, J.G.K., M.K. Hanafey, J.A. Rafalski and S.V. Tingey (1993) Hiratsuka, A. (1984) Life history and population dynamic of Polygonum Genetic analysis using random amplified polymorphic DNA plants. In “Life History and Evolution in Plants 1” Kawano, H. markers. Methods in Enzymology 218: 704–740. (ed.), Baifukan, Tokyo. p. 57–80 (in Japanese). Wolff, K. and M. Morgan-Richards (1999) The use of RAPD data in Hormaza, J.I., L. Dollo and V.S. Polito (1994) Determination of related- the analysis of population genetic structure: case studies of ness and geographical movement of Pistacia vera (pistachio; Alkanna (Boraginaceae) and Plantago (Plantaginaceae). In Anacardiaceae) germplasm by RAPD analysis. Econ. Bot. 48: “Molecular Systematics and Plant Evolution” Hollingsworth, 349–358. P.M., R.M. Bateman and R.J. Gornall (eds.), Taylor and Francis Koenig, R. and P. Gepts (1989) Allozyme diversity in wild Phaseolus Limited, London. p. 51–73. vulgaris: further evidence for two major centers of genetic di- Zohary, D. (1999) Monophyletic vs polyphyletic origin of the crops on versity. Theor. Appl. Genet. 78: 809–817. which agriculture was founded in the Near East. Genet. Res. Kumar, S., K. Tamura and M. Nei (1993) Mega: Molecular Evolution- Crop Evol. 46: 133–142. ary Genetics Analysis, version 1.01. The Pennsylvania State