Annals of Botany 82 : 523–530, 1998 Article No. bo980715

Genetic Variability of Lepidium meyenii and other Andean Lepidium Species (Brassicaceae) Assessed by Molecular Markers J. T O L E D O*†, P. D E H A L*, F. J A R R IN*‡, J. H U*§, M. H E R M A N Ns, I. A L - S H E H B A Z¶ and C. F. Q U I R OS** * Department of Vegetable Crops, UniŠersity of California, DaŠis, CA 95616, USA, s International Potato Center, POB 1558, Lima 12, Peru and ¶ Missouri Botanical Garden, POB 299, St Louis, MO 63166-0299, USA Received : 10 March 1998

Accepted : 19 June 1998

A phylogenetic survey based on similarity levels was performed for 29 cultivated accessions of maca (Lepidium meyenii Walp.) and 27 accessions of wild species of Lepidium from Ecuador, Peru and Bolivia, with RAPD markers. Chromosome counts for each accession were also performed. The similarity tree matrix separated in two main branches : cultivated and wild species. The similarity level among cultivated accessions was high (0±952 or higher) indicating a low level of polymorphism. Within the wild species, two main secondary branches could be resolved, of which one was subdivided into two tertiary branches. Morphological evaluation of the wild species accessions within each main group identified three wild species : (1) L. bipinnatifidum, consisting mostly of tetraploids and a single octoploid accession ; (2) L. kalenbornii, consisting only of tetraploid accessions ; and (3) L. chichicara, consisting mostly of octoploid and a few tetraploid accessions. Clustering by principal coordinates analysis supported the results obtained by the similarity tree matrix. These results indicate that none of the three wild species is related enough to be considered ancestral to the cultivated L. meyenii. Three accessions of intermediate position may be of hybrid origin. None of the wild species was found to be diploid, which suggests that polyploidy has been an important adaptation to high altitude habitats in these species. # 1998 Annals of Botany Company Key words : Lepidum meyenii, Lepidium peruŠianum, maca, DNA markers, phylogeny.

INTRODUCTION The genus Lepidium L. is widely distributed throughout the world on all continents except Antarctica. It consists of approximately 175 species (Mummerhoff et al., 1992) being one of the largest genera in the Brassicaceae (Al-Shehbaz, 1984). The genus probably originates in the Mediterranean basin where most of the diploid species are found (Thellung, 1906 ; Mummerhoff, Hurka and Bandelt, 1992). The species are classified in three large cosmopolitan sections, Dileptium, Monoploca and Lepidium, and three smaller sections restricted to the Old World, Lepia, Lepiocardamon and Cardamon (Thellung, 1906 ; Mummerhoff et al., 1995). Little is known about the time of origin of the genus and the mechanisms responsible for its worldwide distribution. However, most existing evidence indicates that long-distance dispersal taking place in the late Tertiary or Quaternary, rather than continental drift, was responsible for its colonization of the Americas and Australia. This seems to be the prevalent mechanism of distribution of other genera in the family such as Capsella L. and Cardamine L. (Mummerhoff et al., 1992). Common genetic features observed in the immigrant species of Lepidium are autogamy and polyploidy which probably helped their establishment in new habitats. Although there are extensive recent Present addresses : † International Potato Center, POB 1558, Lima 12, Peru, ‡ International Potato Center Research Station, Quito, Ecuador, § Lipton, 2029 E. Harding Way, Stockton, CA 95205. ** For correspondence. Fax 530 752 9659, e-mail cfquiros!ucdavis.edu

0305-7364}98}100523­08 $30.00}0

taxonomic treatments of Lepidium species of Australia (Hewson, 1981) and North America (Al-Shehbaz 1986 a, b ; Rollins, 1993), as well as an outdated general monograph on the genus (Thellung, 1906), South American species need further study following Hitchcock’s (1945) account. The Andean Lepidium species belong mostly to the sections Dileptium and Monoploca (Thellung, 1906). These species are interesting because they grow at high altitudes, up to 4500 m above sea level and include the cultivated species maca, Lepidium meyenii Walp. Maca is an Andean crop with limited distribution restricted today to the suni and puna ecosystems (Bonnier, 1986) of the Departments of Junin and Cerro de Pasco of Peru, where it grows at elevations above 3500 and often reaching 4450 m (Leo! n, 1964 ; Tello, Hermann and Caldero! n, 1992). The taxonomic status of the cultivated species has been questioned by Chaco! n (1990), who proposed changing the name of cultivated maca to L. peruŠianum G. Chaco! n de Popovici based on morphological observations. According to Rea (1992), maca was domesticated at least 2000 years ago in San Blas, Junin. It has one of the highest frost tolerances among other native cultivated plants, since maca is able to grow in the puna, where only alpine grasses and bitter potatoes thrive (Bonnier, 1986). Maca is an octoploid (2n ¯ 8x ¯ 64) plant cultivated for consumption of the root and hypocotyl, and is used extensively for medicinal purposes (Quiros et al., 1996). At least six other Andean species in Peru have been reported by Brako and Zarucchi (1993) from the departments of Ancash south into Puno. Some of these, however, are also found in Ecuador, Bolivia and # 1998 Annals of Botany Company

524

Toledo et al.—Genetic Variability of Lepidium meyenii and Related Species T     1. Accessions of Lepidium species used in the study ID

Original number

Species

Ploidy

11 12 14 15 371 373 374 470 471 472* 473* 474 475* 476* 477* 478 479* 480 481* 482 483 484* 485 486 487 488* 489 490 491* 492* 493* 494 495* 496† 497* 498* 507* 508 509* 511* 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591* 592

CQI-5 CQII-6 CQIV-1 CQIV-2 JTA-116 JTA-140 n}a JTA-038 JTA-039 JTA-040 JTA-041 JTA-075 JTA-082 JTA-085 JTA-086 JTA-090 JTA-092 JTA-097 JTA-106 JTA-108 JTA-114 JTA-115 JTA-117 JTA-118 JTA-121 JTA-123 JTA-125 JTA-126 JTA-127 JTA-138 JTA-139 JTA-192 JTA-193 JTA-195 CIMA-20 CIMA-24 MH-0692 MH-0707 MH-0709 MH-0761 AVC-1 AVC-2 AVC-3 AVC-4 AVC-5 AVC-6 CIMA-34 CIMA-42 CIMA-43 CIMA-48 CIMA-51 CIMA-52 CIMA-53 CIMA-56 CIMA-58 MH-0743 MH-0732

L. kalenbornii L. kalenbornii L. chichicara Hybrid ? L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. meyenii L. kalenbornii L. kalenbornii L. bipinnatifidum L. chichicara L. chichicara L. chichicara L. kalenbornii L. bipinnatifidum L. bipinnatifidum L. bipinnatifidum L. chichicara L. kalenbornii L. chichicara L. chichicara L. kalenbornii L. bipinnatifidum L. bipinnatifidum L. chichicara L. chichicara L. chichicara L. chichicara Hybrid ? Hybrid ?

4x 4x 8x 4x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 8x 4x 4x 4x 8x 8x 8x 4x 4x 8x 4x 4x 4x 8x 8x 4x 4x 4x 8x 8x 8x 8x 4x 8x

* Accessions included in RFLP survey. † Accession not included in RAPD survey.

Origin Colquepata, Peru Sicuani, Peru Apurimac, Peru Apurimac, Peru Huayre, Junin, Peru Junin, Peru Cerro de Pasco, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Huayre, Junin, Peru Carhuamayo, Junin, Peru Carhuamayo, Junin, Peru Carhuamayo, Junin, Peru Carhuamayo, Junin, Peru Carhuamayo, Junin, Peru Carhuamayo, Junin, Peru Carhuamayo, Junin, Peru Junin, Peru Junin, Peru Yanacancha, Huancayo, Peru Yanacancha, Huancayo, Peru Yanacancha, Huancayo, Peru Huayre, Junin, Peru Carhuamayo, Junin, Peru Pichincha, Ecuador La Paz, Bolivia La Paz, Bolivia Cuzco, Peru Cerro de Pasco, Peru Huayre, Junin, Peru Achipampa, Hunacayo, Peru Nununhuayo, Huancayo, Peru Achipampa, Huancayo, Peru Cantara, Junin, Peru La Oroya, Peru La Oroya, Peru Junin, Peru La Oroya, Peru La Oroya, Peru La Oroya, Peru La Oroya, Peru La Oroya, Peru La Oroya, Peru Cuzco, Peru Puno, Peru

Toledo et al.—Genetic Variability of Lepidium meyenii and Related Species Argentina (Hermman, unpubl. res.). Little is known about the origin of these species and even less about their possible relationship to the cultivated species. MATERIALS AND METHODS A survey was carried out of 29 cultivated accessions of maca and 27 wild species of Lepidium from Ecuador, Peru and Bolivia, with RAPD markers. Accessions from wild species found sympatrically to the cultivated species in the region of Lake Junin and Huancayo, Peru were also included. A subsample of 14 cultivated and five wild species accessions were surveyed for RFLP markers (Table 1). Five plants from each of ten cultivated accessions and from ten accessions of the wild species were surveyed for intra-accession variability. Leaves of five plants of each accession were pooled for DNA extraction to provide a representative marker profile of each accession. The exception was accession 592 which had fewer individuals. RFLP analysis DNA extraction, digestion and agarose electrophoresis were carried out as described earlier (Kianian and Quiros, 1992). BamHI, EcoRI, EcoRV and HindIII restriction endonucleases were used for DNA digestion. Alkaline transfer onto Zeta probe GT membranes and hybridization were carried out following manufacturer’s protocol (BioRad). After hybridization, membranes were washed at moderate stringency : twice in 5 % SDS, 20 m Na HPO at # % 65 °C for 15 min and twice in 2 % SDS, 20 m Na HPO at # % 65 °C for 5 min. Probes for hybridization were prepared by random hexamer labelling (Feinberg and Vogelstein, 1983). The following DNA clones identifying known genes were used as probes : self-incompatibility from Brassica oleracea (pBOS5 ; Nasrallah et al., 1985) ; isocitrate lyase and malate synthase from B. napus (IL9 and MS1 ; Comai et al., 1989), napin and cruciferin from B. napus (N2 and C1 ; Crouch et al., 1983), an anonymous clone from B. napus K01-900, associated to linolenic acid content (Hu et al., 1995), rDNA 18S-25S genes from wheat (TA71) and their intergenic spacer (Delseny et al., 1990). Arabidopsis thaliana homologues to Bradyrhizobium Cyc J (At1), to rat p67 translation factor, to yeast Sac1 (At5). Also from A. thaliana, an Emlike coding protein gene Em-1 (Gaubier et al., 1993) (At3), and chlorophyll synthase (Gaubier et al., 1995) (At4). RAPD markers The protocol of Hu and Quiros (1991) was used for DNA amplification with the following modifications : the reaction mix consisted of 10 ng DNA, 25 m of each primer (Operon Technologies, kits A to M), 200 m of each of dATP, dCTP, dGTP and dTTP (Boehringer Mannheim), 1±9 m MgCl , 50 m KCl, 10 m Tris-HCl, (pH 9±0 at 25 °C), # 0±1 % Triton X-100 and one unit of Taq DNA polymerase (Promega). The final volume of the reactions was 10 µl, covered with two drops of mineral oil in 0±5 ml microfuge tubes. Primer combinations were chosen at random. A

525

Perkin-Elmer Cetus DNA thermal cycler was used to amplify the DNA fragments. The cycler was programmed as follows : one cycle at 94 °C for 30 sec followed by 45 cycles of 30 sec at 92 °C, 1 min at 35 °C and 1 min at 72 °C for denaturing, primer annealing and primer extension, respectively. A final 5 min extension at 72 °C was carried out after the cycles were completed. The amplified products were separated by electrophoresis on 2 % agarose gel in 1¬TAE buffer, stained with 10 ppm ethidium bromide and photographed under UV light with Polaroid black and white film or an image analysis system. Operon (Alameda, CA) primers from kits A, B, C, G and H were used for the amplifications. Each sample was run at least twice to select the primers yielding consistent profiles and reproducible bands. Multiple runs were done side by side for accessions sharing the same bands to make sure that these indeed corresponded to the same marker. The SAHN clustering program in the NTSYS-pc package version 1.8 (Rohlf, 1990) was used to construct a tree matrix of nested clusters according to accession similarity levels based on RAPD markers. A principal coordinate analysis was also performed with this package by similarity index of the accession based on marker correlations.

Chromosome counts Chromosome numbers were determined for the wild species of Lepidium in pollen mother cells, following the procedure of Quiros et al. (1996).

Species identification Plants from the wild species were grown from seed in 10 cm pots in our glasshouses. Pressed mature plants were identified by Al-Shehbaz.

RESULTS Polymorphisms of the cultiŠated accessions Very little polymorphism was disclosed among the cultivated accessions of maca for the RFLP markers. The EcoRI profile for 18S–25S rDNA was fixed for the cultivated species. It consisted of two strong signal intensity fragments of approx. 4±2 and 1±6 kb and a light intensity fragment of 5±6 kb. The wild species conserved the first two fragments, but were polymorphic for the third fragment (data not shown). The general profile observed for all the species tested corresponded to that of L. draba L. reported by Delseny et al. (1990). Lack of polymorphism in the cultivated accessions was the rule for the rest of the probes, including napin and cruciferin, which often show intraspecific variation in other Brassicaceae (Cruciferae), such as Brassica oleracea (Kianian and Quiros, 1992). The exception was probe At2, where polymorphism was observed in EcoRI digests for presence or absence of a 7±7 kb band (data not shown). On the other hand, differences were evident for all the probes between cultivated and wild species, separating readily both groups. Also, each of the wild species tested

526

Toledo et al.—Genetic Variability of Lepidium meyenii and Related Species

F. 1. EcoRI profile for napin. Lanes 1 to 4 correspond to four different cultivated accessions (492, 493, 495, 496), the rest are wild species : L. kalenbornii 497 (lane 5), 498 (lane 6) ; 591 (lane 9) of possible hybrid origin ; L. bipinnatifidum 507 (lane 7) ; L. chichicara 509 (lane 8), 511 (lane 10).

F. 2. EcoRV profile for anonymous probe KOL-900 (same accession numbers as Fig. 1. L. bipinnatifidum is not represented in this photo). Lanes 1 to 4 correspond to four different cultivated accessions, the rest are wild species : L. kalenbornii (lanes 5, 6 and 8) and L. chichicara (lanes 7 and 9).

had a unique profile, although some of the fragments were conserved (Figs 1 and 2). The same low level of polymorphism revealed by RFLP markers in the cultivated accessions was detected with RAPD markers (Fig. 3). Primers A4, B6, C18, G10, G11, G19, H2 and H14 were selected for constructing the species relationship tree because of the high reproducibility of their products. These produced a total of 89 bands, of which only 16 were polymorphic in 29 cultivated accessions. On the other hand, polymorphism was higher among the wild species which had different profiles to those observed for the

F. 3. RAPD profiles disclosed with primer B07 for four Lepidium species : L. meyenii ; 493 (lane 1), 494 (lane 2). 495 (lane 3) ; L. bipinnatifidum ; 577 (lane 4), 585 (lane 5), 586 (lane 6), L. kalenbornii ; 584 (lane 7), 497 (lane 8), 489 (lane 9) ; L. chichicara 580 (lane 10), 582 (lane 11), 583 (lane 12).

cultivated species and also a characteristic general profile for each of them (Fig. 3). Seventy nine out of 89 bands were polymorphic in the wild species. Within species polymorphism for each wild accession was not assessed ; however, those tested rarely showed marker polymorphism. When the tree matrix was constructed, similarity levels among cultivated accessions was very high. The lowest level was 0±952, which represents less than three band differences between two accessions. The only appreciable phenotypic variation among accessions was root colour, which ranged from purple to yellow. Only a single representative of accessions sharing the same profile for all the primers tested was included in the final tree matrix. This was done to construct a more concise tree including the cultivated and wild species. Accessions sharing the same marker profile in most cases were morphologically similar. In a few instances, they differed in root pigmentation. Intra-accession variability was rare. Most of the plants within each accession shared the same bands with a few exceptions. The tree separated the cultivated and the wild species into two main branches. Within the wild species, two main secondary branches could be resolved. The top one was subdivided into two tertiary branches. Separating the two main branches of the tree there was a single octoploid accession of a wild species, 592, from Puno, Peru (Table 1). Accession 15 from Apurimac located at the base of the tree, showed low similarity to the adjacent branch of the wild species (Fig. 4). Morphological evaluation of the wild species accessions within each group identified three species : (1) L.

Toledo et al.—Genetic Variability of Lepidium meyenii and Related Species 0.6

0.7

0.8

0.9

527

1.0 371 483 374 486 495 480 471 487 474

L. meyenii

488 490 476 485 489 492 493 494 592

Hybrid?

577 586 585 578

L. bipinnatifidum

507 579 576 581 584 497

L. kalenbornii

498 11 12 591

Hybrid?

580 583 589 590 582 587

L. chichicara

588 511 508 14 509 15

Hybrid?

F. 4. Phylogenetic tree constructed with Tree G (NTSYS), based on 89 RAPD markers, showing similarity indexes for 17 accessions of cultivated L. meyenii and 27 accessions of wild species. (See Table 1 for list of accessions).

528

Toledo et al.—Genetic Variability of Lepidium meyenii and Related Species accessions. Accession 591 from Cusco, Peru separated L. kalenbornii from L. chichicara. Three accessions of L. chichicara at the end of the tree showed lower similarity to the rest of the accessions of this species. Two of them, 15 and 509 were tetraploid forms from Apurimac (Peru) and La Paz (Bolivia), respectively. The other one, 14 was an octoploid from Apurimac. The accession clustering was supported by principal coordinate analysis, which resolved three main clusters without overlaps (Fig. 6) : the accessions of the cultivated species located at one end and the wild species in two opposite clusters at the other end. One of the clusters of wild species included L. chichicara and the other included L. bipinnatifidum and L. kalenbornii, with overlaps. In between were accessions 509 close to the L. chichicara cluster, and 591, close to the L. kalenbornii cluster. Accessions 592 and 15 occupied intermediate positions between the wild and the cultivated species, but at opposite positions. DISCUSSION

F. 5. Metaphase I in pollen mother cells of tetraploid accession 92B498 showing 16 bivalents.

bipinnatifidum Desvaux, consisting mostly of tetraploids (Fig. 5) and a single octoploid accession from Huancayo, Peru ; (2) L. kalenbornii L. C. Hitchcock, consisting only of tetraploid accessions ; and (3) L. chichicara Desvaux, consisting mostly of octoploid and a few tetraploid 55 15 582 582 582

The separation of cultivated and wild species into two main branches based on RAPD markers indicates that none of the wild species screened so far is closely related to maca. Therefore, none of the species of Andean Lepidium tested can be considered an immediate ancestor of the cultivated species. The closest accession to the cultivated ones was 592, but its level of similarity was only approx. 0±6. Accession 592, as well as 591 and 15, which also display intermediate

509

15

14

585 586 507 577 585 584

494 494 493 476 407476 406

592

498 40711

591

374

12

F. 6. Clusters produced by principal coordinate analysis (NTSYS) based on 89 RAPD markers, showing similarity indexes for 17 accessions of cultivated L. meyenii and 27 accessions of wild species (see Table 1 for list of accessions).

Toledo et al.—Genetic Variability of Lepidium meyenii and Related Species positions in the tree and cluster graph may be of hybrid origin. RFLP data support the RAPD data in the sense that distinct fragment profiles were obtained between wild and cultivated species. Although RAPD markers are not ideal for phylogenetic inferences at the species level (Quiros et al., 1995 a ; Quiros, Truco and Hu, 1995 b) in the present study the results were confirmed by other data sets, such as the RFLP survey, morphological description of the species and chromosome number determination. Furthermore, RAPD markers were used more to cluster accessions intraspecifically than to establish phylogenetic linkages among species, which in any case seem quite removed from each other, except for L. bipinnatifidum and L. kalenbornii. It was interesting to discover tetraploid as well as octoploid forms for two species, L. bipinnatifidum and L. chichicara. In most cases, the similarity of 4x and 8x forms was as high as that observed between accessions of the same ploidy. No diploid accessions were found for any species among those tested in this experiment. RFLP profiles observed for all species were typical of diploid crucifers of complex genomes, such as Brassica. The probes used in this experiment, when hybridized to Brassica diploids, generate a similar number of bands to that observed for Lepidium species in spite of being tetraploid and octoploid (Kianian and Quiros, 1992 ; Sadowski et al., 1996). On the other hand, when these probes are hybridized to Brassica amphidiploids, such as B. napus L., B. carinata A. Br. and B. juncea (L.) Czern., twice as many bands are observed corresponding to the two different genomes included in these species (Hosaka et al., 1990). Therefore, it is unlikely that Lepidium meyenii and the wild species surveyed here are allopolyploids. These species most likely contain a single genome or closely related genomes monomorphic for duplicated loci. However, their chromosomes show strict disomic pairing resulting in high pollen fertility. Therefore, these species are probably autopolyploids which have evolved through time to this type of chromosome behaviour assuring their reproduction and successful colonization of the harsh environment found at high altitudes. The low level of polymorphism observed for the different cultivars of L. meyenii agrees with its breeding system, obligate autogamy by cleistogamy (Quiros et al., 1996). The accessions of L. meyenii sampled in the present study may represent approx. 80 % of the existing maca cultivars. The present cultivated area may not cover more than 500 km# from Cerro de Pasco to the Jauja and Huancayo regions. Therefore, maca has a very narrow genetic base. Cross hybridization of maca accessions must be a rare event because of the cleistogamous nature of the flowers and presence of few pollinating insects at high altitudes. Our artificial hybridization attempts have failed due to the small flower size and poor tolerance to injury caused by emasculation. One can speculate that interspecific hybridization may occur to some extent. We know very little, however, about the floral biology of the wild species to assess whether cleistogamy is also widespread in related wild species. The morphological differences between Lepidium meyenii and each of L. kalenbornii, L. bipinnatifidum and L. chichicara are substantial, and the last three fall into section

529

Dileptium, whereas L. meyenii is a member of section Monoploca. Cultivated L. meyenii is the only species in the entire genus that produces fleshy roots. Lepidium bipinnatifidum is readily distinguished from the other species included in this study by having pinnately divided stem leaves with distinctly auriculate bases. Both L. chichicara and L. kalenbornii have toothed or entire stems leaves that are not auriculate at base, but they can be easily separated from each other by their fruits, which are 2–2±5 mm long in L. kalenbornii and are longer (3–4 mm) in L. chichicara. The calyx in L. kalenbornii is persistent and remains attached well after fruit maturity, while in L. chichicara it is often deciduous before the fruit is fully mature. Furthermore, the latter has a minute nectar gland hardly reaching 0±1 mm, whereas L. kalenbornii has distinct nectar glands 0±3 mm or longer. In summary, none of the three wild species sympatric to the cultivated one appears closely related to cultivated L. meyenii. The closest was accession 592, but it still was too distant from the cultivated species to be considered ancestral. None of the wild species was found to be diploid suggesting that polyploidy has been an important adaptation to high altitudes. Two questions still remain to be answered : what is the origin of maca and how did it become established as a new crop that does not exist in the Old World ? Undoubtedly, a more extensive collection of Andean Lepidium species and their characterization will be necessary before attempting to answer these questions. The next step would be to study species of Lepidium that occur in the same section as L. meyenii. These include L. solomonii AlShehbaz (Bolivia), L. jujuyanum Al-Shehbaz (Argentina) and L. weddellii O. E. Schulz (Peru). These species are very rare ; the first two are each known from a single type collection, whereas L. weddellii is known from two or three old collections. Attempts should be made to re-collect these rare species, especially the latter, which produces thickened roots. A C K N O W L E D G E M E N TS We are indebted to Vincent D’Antonio for technical assistance, to Rolando Aliaga for supplying seeds of some of the cultivated accessions and to Miguel Holle for useful discussion and comments on the manuscript, to Carlos Arbizu, Ramiro Ortega and Andres Valladolid for participation in collection trips of the wild species. Research was supported in part by a grant from the Swiss Development Corporation to the Collaborative Program of Andean Root and Tuber Biodiversity at the International Potato Center (CIP-COI-SDC).

LITERATURE CITED Al-Shehbaz IA. 1984. The tribes of Cruciferae (Brassicaceae) in the southeastern United States. Journal of Arnold Arboretum 65 : 343–373. Al-Shehbaz IA. 1986 a. New wool-alien Cruciferae (Brassicaceae) in North America : Lepidium and Sisymbrium. Rhodora 88 : 347–356. Al-Shehbaz IA. 1986 b. The genera of Lepidieae (Cruciferae ; Brassicaceae) in southeastern United States. Journal of the Arnold Arboretum 67 : 265–311.

530

Toledo et al.—Genetic Variability of Lepidium meyenii and Related Species

Bonnier E. 1986. Utilization du sol a l’epoque prehispanique : le cas archeologique du Shaka-Placamayo. Cahiers des Sciences Humaines 22 : 97–113. Brako L, Zarucchi JL. 1993. Catalogue of the flowering plants and gymnosperms of Peru. Monograph in Systematic Botany from the Missouri Botanic Garden 45 : 1–1286. Chaco! n G. 1990. La maca (Lepidium peruŠianum Chaco! n sp. nov. y su habitat. ReŠista Peruana de Biologia 3 : 171–272. Comai L, Deitrich RA, Maslyar DJ, Baden CS, Harada JJ. 1989. Coordinate expression of transcriptionally regulated isocitrate lyases and malate synthase genes in Brassica napus L. The Plant Cell 1 : 293–300. Crouch ML, Tenbarge KM, Simon AE, Ferl R. 1983. cDNA clones from Brassica napus seed storage proteins : evidence from nucleotide sequence analysis for both subunits of napin are cleaved from a precursor polypeptide. Journal of Molecular and Applied Genetics 2 : 273–283. Delseny M, McGrath M, This C, Chevre A, Quiros CF. 1990. Ribosomal RNA genes in diploid and amphidiploid Brassica and related species : organization, polymorphism and evolution. Genome 33 : 733–744. Feinberg AP, Vogelstein B. 1983. A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Analytical Biochemistry 132 : 6–13. Gaubier P, Wu J, Laudie M, Delseny M, Grellet F. 1995. A chlorophyll synthase gene from Arabidopsis thaliana. Molecular and General Genetics 249 : 58–64. Gaubier P, Raynal M, Huestis G, Hull G, Grellet F, Arenas C, Pages M, Delseny M. 1993. Two different Em-like genes expressed in Arabidopsis thaliana seeds during maturation. Molecular and General Genetics 238 : 409–418. Hewson HJ. 1981. The Genus Lepidium L. (Brassicaceae) in Australia. Brunonia 4 : 217–308. Hitchcock LC. 1945. The South American species of Lepidium. Liloa 11 : 75–134. Hosaka K, Kianian SF, McGrath JM. Quiros CF. 1990. Development and chromosomal localization of genome specific DNA markers of Brassica and the evolution of amphidiploid and n ¯ 9 diploid species. Genome 33 : 131–142. Hu J, Quiros CF. 1991. Identification of broccoli and cauliflower cultivars with RAPD markers. Plant Cell Reports 10 : 505–511. Hu J, Quiros C, Arus P, Struss D, Robbelen G. 1995. Mapping of a gene determining linolenic acid concentration in rapeseed with DNAbased markers. Theoretical and Applied Genetics 90 : 258–262.

Kianian SF, Quiros CF. 1992. Genetic analysis of major multigene families of Brassica oleracea and related species. Genome 35 : 516–527. Leo! n J. 1964. The ‘ maca ’ (Lepidium meyenii), a little-known food plant of Peru. Economic Botany 18 : 122–127. Mummerhoff K. 1994. Subunit polypeptide composition of RUBISCO as a taxonomic tool in separating critical Lepidium species. Phytochemistry 35 : 683–684. Mummerhoff K, Hurka H, Bandelt H-J. 1992. Systematics of Australian Lepidium species (Brassicaceae) and implications for their origin : evidence from IEF analysis of Rubisco. Plant Systematics and EŠolution 183 : 99–112. Mummerhoff K, Kuhnt E, Koch M, Zunk K. 1995. Systematic implications of chloroplast DNA variation in Lepidium sections Cardamon, Lepiocardamon and Lepia (Brassicacea). Plant Systematics and EŠolution 196 : 75–88. Nasrallah JB, Kao TH, Goldberg ML, Nasrallah ME. 1985. A cDNA clone encoding an S-locus specific glycoprotein from Brassica oleracea. Nature 318 : 263–267. Quiros CF, Truco MJ, Hu J. 1995 b. Sequence comparison of two codominant RAPD markers in Brassica nigra : deletions, substitutions and microsatellites. Plant Cell Reports 15 : 268–270. Quiros CF, Epperson A, Hu J, Holle M. 1996. Physiological studies and determination of chromosome number in maca, Lepidium meyenii (Brassicaceae). Economic Botany 50 : 216–223. Quiros CF, This P, Laudi M, Benet A, Chevre AM, Delseny M. 1995 a. Analysis of a set of RAPD markers by hybridization and sequencing in Brassica : a note of caution. Plant Cell Reports 14 : 630–634. Rea J. 1992. Cultivos marginados, otra perspectiva de 1492. FAO Rome : JE Hernandez-Bermejo, J Leo! n, 163–166. Rohlf FJ. 1990. NTSYS-pc. Numerical taxonomy and multiŠariate analysis system. New York : Exeter Software. Rollins RC. 1993. The cruciferae of Continental North America. Stanford : Stanford University Press. Sadowski J, Gaubier P, Delseny M, Quiros CF. 1996. Genetic and physical mapping in Brassica diploid species of a gene cluster defined in Arabidopsis thaliana. Molecular and General Genetics 251 : 298–306. Tello J, Hermann J, Caldero! n A. 1992. La Maca (Lepidium meyenii Walp) : Cultivo Alimenticio Potencial para las Zonas Altoandinas. BoletıU b de Lima 81 : 59–66. Thellung A. 1906. Die Gattung Lepidium (L.) R. Br. Eine monographische Studie. Neue Denkschrift Allgemeinen Schweizerischen Naturforscher Gesellschaft 41 : 1–340.