Allozyme Variability and Pbylogenetic Relationships in the Cultivated Potato (Solanum tuberosum) and Related Species

Gene frequencies at 13 isozyme loci were determined in three South American taxa of cultivated potatoes [the diploid group (gp.) Stenotomum, the diploid subgroups (subgp.) Goniocalyx, and the tetraploid gp. Andigena of S. tuberosum], in the diploid weed species S. sparsipilum, and in most of the main cultivars now raised in the Northern Hemisphere (the tetraploid gp. Tuberosum of S. tuberosum). High levels of genetic variability (mean number of alleles per locus, percentage of polymorphic loci, and mean heterozygosity) were detected, being higher in tetraploid potatoes. An equilibrium among the evolutionary factors which increase genetic variability and artificial selection for maximum yield would explain the high uniformity of heterozygosity values we observed in both Andigena (0.36 * 0.02) and Tuberosum (0.38 f 0.01) cultivars. -The low value of genetic distance (D = 0.044) between Stenotomum and Goniocalyx does not support the status of species for S. goniocalyx.-In most isozyme loci, the electromorphs of gp. Andigena were a combination of those found in both gp. Stenotomum and S. sparsipilum, suggesting an amphidiploid origin of gp. Andigena from that two diploid taxa. The presence in Andigena of unique electromorphs, which were lacking in both gp. Stenotomum and S. sparsipilum, suggests that other diploid species could be also implied in the origin of tetraploid Andean potatoes. Furthermore, since Andigena were more related to Stenotomum (D = 0.052) than to S. sparsipilum (D = 0.241), the autopolyploidization of Stenotomum individuals and the subsequent hybridization with gp. Andigena may also have occurred. Thus, our study suggests a multiple origin (amphidiploidy, autoploidy, and hybridization at tetraploid level) of gp. Andigena. Most of the electromorphs of gp. Tuberosum were also found in gp. Andigena; both the direct derivation of that group from the Andean tetraploid potatoes and the repeated introgression provided by breeding programmes could explain this result. However, the allele c of Pgm-B, present in 30 out of 76 Tuberosum cultivars from Northern Hemisphere as well as in 3 Chilean Tuberosum cultivars, lacks in the 258 Andigena genotypes sampled, suggesting 1 P I . s,vst. E:vol.. T-01. 118. S,r. l-2 2 J. L. OLIVER & J. M. MARTiNE ZAPATER: that Chilean germplasm could have taken part in the origin of at least the 39% of the potato cultivars from Europe and North America analyzed here. -The distance WAGNER procedure provides an estimate of a 30% of heterogeneity in the evolutionary divergence shown by different groups of cultivated potatoes. Diploid groups show a higher (22.5%) evolutionary rate than tetraploids, which can be attributed to both tetrasomic inheritance and facultative autofecundation that exists in Andigena and Tuberosum groups. Thus, artificial selection acting since 10 000 years has not resulted in a higher rate of molecular evolution at the isozyme level in the tetraploids. Despite the large series of historical, biogeographical, morphological, cytogenetical and biochemical studies which have been carried out on Solanum subsect. Potatoe D’ARCY (see SWAMINATHAN & MAGOON 1961, HOWARD 1970, 1978, UGENT 1970, HAWKES 1978a, b for revisions), a wide controversy still remains about the origin, the geographical distribution, and the phylogenetic relationships of the different taxa of cultivated potatoes (Solanum tuberosum L.), reflected in the different systematic classifications of the species. Here we use the terminology of DO D D S (1962), which appears to conform closely with the conclusions of the present study. The complex pattern of genetic variability, in addition to the high degree of phenotypic plasticity, makes it difficult to study biosystematic relationships among potatoes. A number of studies dealing with morphology and chemotaxonomy attempted to resolve this question. The basic problem with most of these traits is the lack of equivalence between phenotype and genotype due to complications introduced by varying heritability, dominance, epistasis, and pleiotropy. Cytogenetic analyses were also used in phylogenetic studies on potatoes. However, since most diploid potato species hybridize readily and produce fertile F l’s, genome differentiation seems to have not progressed very far (HAWKES 1978b), thus rendering cytological studies inconclusive in revealing the phylogenetic history (SWAMINATHAN & MAGOON 1961). Isozymes have been succesfully used in the analysis of animal and plant populations, including some solanaceous groups (RICK & FOBES 1975, RICK & al. 1977, WHALEN 1979, MCLEOD & al. 1983). In order to estimate the genetic variability in potatoes, we have undertaken an electrophoretic analysis of several genetically controlled enzyme systems in the most important cultivated groups of S. tuberosum as well as in one of the most related wild species, S. sparsipilum. Gene frequencies at 13 isozyme loci were used to trace the phylogenetic relationships among these taxa of potatoes. Material and Methods Collections. The cultivated potato S. tuberosum L. (group Stenotomum including the subgroup Goniocalyx group Andigena, and group Tuberosum) and Allozyme Variability in Solanum tuberosum groups 3 the diploid weed S. sparsipilum (BITT.) Juz. et BUK. were analyzed. The collections were of three kinds. The material received from Centro International de la Papa (Lima, Peru) was composed of seeds obtained from different cultivated varieties (Table 1). The material from the Inter-Regional Potato Introduction Station, Sturgeon Bay (Wisconsin) consisted of tuber families (Table 1). With respect to group Tuberosum we analyzed tubers of 76 cultivars from the Northern Hemisphere (listed in MARTINEZ ZAPATER & OLIVER 1984 a) including those of most agronomic interest in Europe and North America, derived from different sources (Estacion de Mejora de la Patata, Vitoria, Spain; Instituto National de Semillas y Plantas de Vivero, Madrid, Spain; Government Institute for Research on Varieties of Cultivated Plants, Wageningen, Holland). In addition, we also analyzed plantlets derived from seeds of 8 Tuberosum cultivars from Chile (Universidad Austral de Chile, Valdivia; see Table 3). Both tubers and seeds were grown in standard conditions in the greenhouse. We included also S. pinnatisectum DUN. in the study (Table 1) because this species is also in the section Petota of the genus Solarium, but is not especially closely related; it thus serves as an “outgroup” in tree construction. Electrophoretic Methods. Four organs of the plant were analyzed: young and mature leaves, tubers, and shoots. Enzyme extraction procedures and horizontal starch gel electrophoresis conditions were as described previously by OLIVER & al. (1983). Samples of cv. ‘Desiree’ were included in all the slab gels as internal markers to determine the electrophoretic mobilities of the different allozyme bands. Different gel slices were assayed for 10 consistently storable enzymes: Esterases (EST, E.C. 3.1.1. l), Alcohol dehydrogenase (ADH, E.C. 1.1.1.1), Glutamate oxaloacetate transaminase (GOT, E.C. 2.6.1. l), Phosphoglucose isomerase (PGI, E.C. 5.3.1.9), Phosphoglucomutase (PGM, E.C. 2.7.5.1), Peroxidases (POX, E.C. 1.11.1.7), Malic enzyme (Me, E.C. 1.1.1.40), Glutamate dehydrogenase (GDH, E.C. 1.4.1.3), Malate dehydrogenase (MDH, E.C. 1.1.1.37), and 6-Phosphogluconate dehydrogenase (PGD, E.C. 1.1.1.43). For most enzymes standard staining methods (BREWER 1970, SHAW & PRASAD 1970) were used; detailed references will be found elsewhere (OLIVER & al. 1983, MARTINEZ ZAPA~ER & OLIVER 1984 b). Estimation of Gene Frequencies. A computer program for general maximum likelihood estimation (MAXLIK) written by T. E. REED, University of Toronto, was used to estimate gene frequencies for each locus. In those accesions of gp. Stenotomum and Andigena grown from seeds (Table l), we directly computed electromorph frequencies in each accession; the genotypes found in the remaining accessions, composed by tuber families (Table l), were pooled by taxa before computing electromorph frequencies. The 76 cultivars of gp. Tuberosum from Northern Hemisphere were used as a whole to estimate gene frequencies in this group. Owing to reasons which we shall see below, the sample of gp. Tuberosum from Chile was only analyzed for the Pgm-B locus; gene frequencies at other isozyme loci were not computed and, therefore, Chilean Tuberosum was not used in computing variability measures, genetic distances of phylogenetic trees. In the tetraploid taxa analyzed here, isozyme loci were expressed in duplicate, originating asymmetrical banding intensities in the heterozygotes due to gene dosage effects (MARTINEZ ZAPATER & OLIVER 1984 b; see also OLIVER & al. 1983 for details about the criteria employed in the determination of duplicate expression of isozyme loci); these asymmetrical bands can be readily detected on the gels, allowing us to compute allelic frequencies in the duplicated loci of tetraploids.