Novel nuclear markers inform the systematics and the evolution of serpentine use in Streptanthus and allies (Thelypodieae, Brassicaceae)

[Display omitted] •Jewelflowers (Streptanthus complex) are florally and ecologically diverse mustards.•Many Jewelflowers are edaphic (soil) specialists, 1/3 endemic to serpentine soils.•We present the best-resolved and most-comprehensive Jewelflower phylogeny to date.•Several genera in the group are not reciprocally monophyletic as currently defined.•Serpentine tolerators and endemics have evolved repeatedly and are not ‘dead-ends'. Streptanthus is a genus of ca. 35 species in the tribe Thelypodieae (Brassicaceae) that has remarkable morphological and ecological diversity, a large number of species in the group being edaphic specialists endemic to unusual soils such as serpentine. While ecological research has shed some light on adaptation to serpentine in Streptanthus, there have been few insights on the origins and evolution of serpentine tolerance in this group, largely due to limited success in resolving the phylogenetic relationships among Streptanthus and allied genera of the Thelypodieae (Streptanthoid complex). We present a well-resolved phylogenetic hypothesis for the Streptanthoid complex, based on three newly identified and highly variable single copy nuclear regions (AT4G34700, AT1G61620, and AT1G56590, and three others that are widely used (ITS, phyA, and PEPC). We also include data for two chloroplast regions (trnL and trnH-psbA). Collectively, our new markers provide 75% of the nuclear parsimony informative characters in our data. Taxonomically, our sampling is the most inclusive of any study of the Streptanthoid Complex to date, including 46 out of the 53 species of Streptanthus and Caulanthus, as well as representatives of several closely allied genera in the Thelypodieae. Our results reveal that Streptanthus, Caulanthus, and Thelypodium are not reciprocally monophyletic as currently defined. The species of Streptanthus form two rather distantly related clades. One clade (SC-I) is comprised of species with bilateral flowers and urn-shaped calyces that occur mainly within the California Floristic Province (CFP) hotspot; the other clade (SC-II) is composed of species with extant ranges mainly outside the CFP. Our data indicate that serpentine tolerance has evolved between eight and ten times in this group, of which between four and five have resulted in endemism. While serpentine endemism has been rarely lost, large and diverse clades composed mainly of serpentine endemics indicate that serpentine endemics in this group are more than mere ‘d