The phylogenetic relationships among 14 Mangifera L. species including three economically important species, i.e., common mango (M. indica L.), horse mango (M. foetida Lour.) and kwini (M. odorata Griff.), were analyzed by comparing 217 amplified fragment length polymorphism (AFLP) markers. The unweighted pair grouping method using arithmetic averages (UPGMA) and neighbor-joining (NJ) method were used and two outgroup taxa, cashew nut (Anacardium occidentale L.) and gandaria (Bouea macrophylla Griff.), were added to both analyses. The common mango was closely related to banana mango (M. sylvatica Roxb.), M. laurina Bl., and M. oblongifolia Hook.f. Intraspecific variation among seven cultivars of common mango was much smaller than interspecific variation and these cultivars were classified into one M. indica group using both methods. Mangifera macrocarpa Bl., M. foetida, and M. odorata were also related to M. indica in both UPGMA and NJ trees, although these three species are classified into a different subgenus (subgenus Limus) from the subgenus Mangifera to which M. indica belongs. Also, in both UPGMA and NJ trees, M. gedebe Miq. and M. griffithii Hk.f. were placed in distant positions among the Mangifera species tested, indicating these two species are related distantly to M. indica. The AFLP technique was confirmed to be useful for phylogenetic analysis.
This report demonstrates the presence of S-ribonucleases (S-RNases), which are associated with gametophytic self-incompatibility (SI) in Prunus L., in styles of self-incompatible and self-compatible (SC) selections of tetraploid sour cherry (Prunus cerasus L.). Based on self-pollen tube growth in the styles of 13 sour cherry selections, seven selections were SC, while six selections were SI. In the SI selections, the swelling of pollen tube tips, which is typical of SI pollen tube growth in gametophytic SI, was observed. Stylar extracts of these selections were evaluated by two-dimensional polyacrylamide gel electrophoresis. Glycoproteins which had molecular weights and isoelectric points similar to those of S-RNases in other Prunus sp. were detected in all selections tested. These proteins had immunological characteristics and N-terminal amino acid sequences consistent with the S-RNases in other Prunus sp. Two cDNAs encoding glycoproteins from `Erdi Botermo' were cloned. One of them had the same nucleotide sequence as that of S4-RNase of sweet cherry (Prunus avium L.), while the amino acid sequence from the other cDNA encoded a novel S-RNase (named Sa-RNase in this study). This novel RNase contained two active sites of T2/S type RNases and five regions conserved among other Prunus S-RNases. Genomic DNA blot analysis using cDNAs encoding S-RNases of sweet cherry as probes indicated that three or four S-RNase alleles are present in the genome of each selection regardless of SI. All of the selections tested seemed to have at least one S-allele that is also found in sweet cherry. Genetic control of SI/SC in tetraploid sour cherry is discussed based on the results obtained from restriction fragment length polymorphism analysis.