Evolution of dioecy in the Cucurbitaceae genus Bryonia

Genetic crosses between the dioecious Bryonia dioica Jacq.
(Cucurbitaceae) and the monoecious B. alba L. in 1903 provided the
first clear evidence for Mendelian inheritance of dioecy and made
B. dioica the classic case of XY sex determination in plants. We
use chloroplast (cp) and nuclear (nr) DNA sequences from 129
individuals representing all morphological species to study species
relationships and distribution, sexual system evolution, and
association of ploidy-level with dioecy in Bryonia. Chloroplast and
nuclear trees mostly fit morphological species concepts; there are
seven dioecious and three monoecious species, together ranging from
the Canary Islands to Central Asia. Bryonia verrucosa, the
morphologically most differing species from the Canary Islands is
sister to all other species. Our data argue for the inclusion of
the narrowly endemic Central Asian species B. lappifolia and B.
melanocarpa in B. monoica. Conflicts between cp and nr topologies
imply that the dioecious hexaploid B. cretica arose from
hybridization(s) involving the diploid species B. dioica, B.
syriaca, and/or B. multiflora. The tetraploid B. marmorata likely
originated via autopolyploidy. The nr phylogeny implies at least
two transitions between dioecy and monoecy, but no correlation
between change in sexual system and ploidy level. Fossil-calibrated
molecular clocks using family-wide rbcL data with a
Bryonia-centered sampling suggest that the deepest divergence in
Bryonia occurred ca. ten million years ago and that monoecious and
dioecious species crossed in the classic studies are separated by
several million years of evolution. Traits, such as annual regrowth
from a tuberous rootstock and other adaptations to a seasonal
climate, as well as species and haplotype abundance, point to an
origin of Bryonia in the Middle East. Species and haplotype poverty
north of the Alps together suggest recolonization there after the
last glacial maximum. Most species of Bryonia have 10 chromosomes
(as confirmed by my own counts), and there appears to be no
morphologically distinct pair that would represent the sex
chromosomes. However, we know from the crossings carried out by
Correns and others that in B. dioica, sex shows monofactorial
dominant inheritance, setting up the hypothesis that B. dioica may
have a pair of chromosomes on which key sex-determining gene(s) and
sexlinked genes have accumulated. To gain insight into the possible
presence of such a pair of sex chromosomes in B. dioica, it is
necessary to sequence a fairly long sex-linked region to study its
substitution behavior and to eventually visualize its physical
placement using FISH. As a first step towards this goal, I
developed a sex-linked SCAR marker for B. dioica from AFLP bands
and sequenced it for individuals representing the full distribution
range of the species from Scotland to North Africa. The region
north of the Alps harbours distinct Y and X alleles that differ in
a 197-bp indel, with the Y allele being perfectly linked to the
male sex. In southern Europe, however, the XY system appears to
break down (to an extent that is not clear), and there are signs of
recombination between the Y and X homologues. Population genetic
analyses suggest that the sex-linked region I amplified (i.e., the
SCAR marker) experienced different evolutionary pressures in
northern and southern Europe. These findings fit the evidence from
my phylogenetic and phylogeographic analyses that the XY system in
Bryonia is evolutionarily labile. Overall, my work suggests that
Bryonia may be a good, but very complex, system in which to study
the early steps of plant sex chromosome evolution.