Parental genome separation and asynchronous centromere division in interspecific F-1 hybrids in Lathyrus

Chromosomes were studied in root-tip metaphase cells of several F-1 interspecific Lathyrus hybrids including: L. hirsutus x L. cassius (H x C), L. cassius x L. hirsutus (C x H), L. cassius x L. odoratus (C x O), and their parents, all with 2n = 2x = 14. Two types of morphologically distinct centromeres were identified in the hybrids on the basis of the degree of contraction of the primary constriction. At least 12 well-defined centromeres were seen in all cells of L. hirsutus, L. cassius and L. odoratus, and about 80% of cells had 14. The hybrids were more variable than the species. H x O contained between six and 14 well-defined centromeres, while cells of H x C, C x H and C x O all had seven well-defined and seven weakly defined centromeres. These were used as markers to plot their spatial disposition in two dimensions on metaphase spreads. In H x C, C x H and C x O the two types of centromeres showed a significant tendency to occupy two spatially distinct and concentrically arranged domains on the metaphase plate (P<0.005). Owing to shortage of material subsequent work was restricted to H x C and C x H. Six or seven chromosomes of one parental genome were selectively labelled by in situ hybridization using biotinylated total genomic DNA from either parent as a probe. Moreover, there was a very strong correlation between centromere type and genomic origin (P<0.001). Clearly the tendency for spatial segregation of parental genomes into concentrically arranged domains previously noted in several F-1 interspecific sexual hybrids between grasses can occur strongly in dicots. The present work on Lathyrus also included the first comparative study of the spatial arrangement of parental genomes in reciprocal hybrids. Significantly, L. cassius chromosomes tended to surround L. hirsutus chromosomes in both H x C and C x H. Thus, the polarity of concentric parental genome separation was not determined as a maternal effect.