Abstract:
Nemesia is a genus endemic to southern and tropical Africa, containing approximately 65 named species of annual or perennial herbs or sub-shrubs. The species are predominantly diploid with 2n=18 chromosomes. Phylogenies of 26 Nemesia taxa were constructed using nucleotide sequences of the ITS, ETS and the trnL-intron regions. ITS and ETS sequences had a higher rate of evolution and were more informative than the trnL-intron region. In the ITS and ETS based phylogenies, species were grouped into four clades, two composed of annual species, one that contained one annual and one perennial species, and one that was predominantly composed of perennial species. The relationships of three Nemesia species were not fully resolved by this phylogenetic analysis.
The divergence times of Nemesia species were estimated from ITS trees using penalized likelihood. In the absence of fossils of Nemesia of known ages another method of calibrating the tree was required. To achieve this a phylogenetic tree of species from Scrophulariaceae, Plantaginaceae, Calceolariaceae, Gesneriaceae, Gentianaceae and Boraginaceae species was constructed using published sequences of three chloroplast genes, rbcL, ndhF and rps2. This tree was calibrated using the published dates of 77-81 Mya for the separation of the Lamiales from the Boraginales, and 7.1 Mya for the separation of Plantago and Aragoa. Penalised likelihood was used to estimate dates of nodes in this tree. Which suggested that the most recent common ancestor of Nemesia and Alonsoa existed 47.5 million years ago. This allowed the nodes on the ITS trees to be dated using penalised likelihood, calibrating the split of Nemesia and Alonsoa at 47.5 Mya. The results from this indicated that the genus evolved during the Miocene, and that the majority of extant Nemesia species studied radiated during the Pliocene. These dates correlate with the aridification of western southern Africa and a spread of open habitats, such as grassland, desert, scrubland and fynbos. Penalised likelihood analysis of substitution rates of the ITS region indicated that the mean substitution rate of the ITS region of annual species was higher than that of perennial species.
Chromosome evolution in Nemesia was investigated using fluorescent in situ hybridization (FISH) to identify the locations of 45S ribosomal genes, 5S ribosomal genes, telomere related sequences, centromere repeats and transposable sequences in addition to measurements of chromosome length and arm ratios. Conservation was found between Nemesia species in chromosome number, size and centromere position, suggesting that either karyotype orthoselection or karyotype conservation has occurred in the genus. FISH showed that Nemesia contain Arabidopsis-type telomere repeats (TTTAGGG)n at both ends of all their chromosomes. Nemesia species were found to contain either one or two 5S sites and between two and four 45S sites. These were either sub-terminally or interstitially located, and 45S and 5S sites were often located on the same chromosome pair. A large amount of variation was found in both number and position of ribosomal genes in different Nemesia species (21 different arrangements of 45S and 5S rRNA genes were observed in the 28 Nemesia taxa studied). Plants with a higher number of 45S rDNA loci showed a higher rate of ITS substitution, but this did not account for all of the rate variation observed between annual and perennial Nemesia species.
Degenerative PCR primers were used to amplify fragments from retrotransposons and transposons from N. anisocarpa, N. foetans and N. strumosa. These were sub-cloned and sequenced. The majority showed homology to transposable elements isolated from other species. A total of four Tyl-copia-like retrotransposons, five non-LTR retrotransposons and six En/Spm-like transposons were isolated from N. anisocarpa, N. foetans or N. strumosa. A range of sequence divergence was observed in these clones. Some of the transposable elements isolated from different Nemesia species were very similar. This may be the result of sequence divergence during the vertical transmission of these transposable elements in Nemesia. Other sequences were quite diverse and were more closely related to transposable elements isolated from other plants than to the other clones isolated from Nemesia.
Differences were observed in the numbers and the distributions of the different types of transposable elements in the genomes of Nemesia. Tyl-copia-like retrotransposons were present in high numbers and widely distributed in Nemesia chromosomes. The clones pNaty2 and pNfty1 were found along the lengths of both arms of all 18 chromosomes of twelve Nemesia species tested. Variation was seen in the numbers and distribution of different LINEs in Nemesia. Two of the LINEs were present in low numbers in the genomes of Nemesia species and did not show clustering of repeats. A third clone, pNali10, was present in different numbers in different Nemesia species and was found to cluster in sub-terminal regions. Some Nemesia species contained high numbers of this LINE and clusters of signals were seen in the sub-terminal regions of most chromosome arms, other Nemesia species contained fewer copies of this LINE and clusters of signals were seen in the sub-terminal regions of a few chromosome arms. Two En/Spm-like transposons were used as probes for in situ hybridisation these were found to be present in low numbers in the genomes of twelve Nemesia species.
Several Nemesia species were able to sexually hybridise with other Nemesia species. The ability of species to hybridise correlated with the clades determined from DNA sequence. The majority of inter-fertile hybrid combinations occurred between species from the same clade, with only one cross out of several attempted between clades being successful. Hybrids were produced from crosses between the perennial members of Clade II, between N. strumosa and, N. macroceras, which are both members of Clade III, and between N. anisocarpa (Clade I) and N. foetans (Clade II). The analysis of chromosome behaviour in interspecific hybrids indicated that Nemesia chromosomes in different parental species were homeologous. No evidence of chromosome inversions or chromosome translocations was observed during meiosis in interspecific hybrids between the perennial Nemesia species in Clade II. This suggests that the similarities in the karyotypes of perennial Nemesia species within Clade II are due to karyotype conservation rather than karyotype orthoselection. It also suggests that at least some of the changes in the relative positions of 45S and 5S rDNA in Nemesia have occurred through transposition rather than by chromosome rearrangement.
Within clade III, hybrids could be produced between N. macroceras and N. strumosa; a quadrivalent was observed during meiotic metaphase in these hybrids, indicating that these two species differ by a reciprocal translocation. This chromosome rearrangment may have played a role in speciation within this clade by preventing recombination in this region of the genome between populations of Nemesia that differ by this translocation.
Hybrids could not be produced between many different Nemesia combinations. In all of these crosses pollen tubes were observed entering ovaries and ovules suggesting that post-fertilisation barriers rather than pre-fertilisation barriers are preventing sexual hybridisation. Many of these crosses produced shrunken, empty seeds that were inviable, suggesting that endosperm breakdown and embryo abortion prevent interspecific hybridisation in these crosses. The nature of these reproductive barriers suggest that they have developed following speciation perhaps due to a loss of genetic compatibility rather than having been active drivers of the speciation process.