The distribution of Larrea tridentata essentially defines the desert floristic provinces of North America. The Mojave Desert (left) is the northernmost and westernmost of the warm deserts; The Sonoran Desert (center) is considered the wettest of the warm deserts and supports a wide diversity of plants, including columnar cactus; The Chihuahuan Desert (right) is the easternmost of the warm deserts, extending south into central Mexico.
Distributions of the three cytotypes of Larrea tridentata. Diploids (squares) are found in the Chihuahuan Desert (green), tetraploids (circles) are found in the Sonoran Desert (blue), hexaploids (triangles) are found in both the Mojave (red) and Sonoran Deserts. The morphologically distinct tetraploid dune creosote (Larrea tridentata var. arenaria) is found only in the sandy habitat of the Algodones Dunes in southeastern California (crosshatched circles).
An example of one of my permanently marked plants (left). Detail of the leaves of L. tridentata (center). Collecting L. tridentata tissue for later DNA extraction in the lab and genetic analyses (right).
To identify areas where more than one cytotype occurred (mixed sites), I ran sampling transects so I could sample plants randomly (left). Morphological differences are slight between the cytotypes making identification largely dependent on DNA content evaluation through flow cytometry. However, this process was successful and I have identified several mixed-cytotype sites where the plants are permanently marked; a site with both diploids and tetraploids (left) and a site with both L. tridentata var. arenaria and hexaploids (right). The tall plant on the left is tetraploid L. tridentata var. arenaria and the shorter plant on the right is the hexaploid.
Recording flowering phenology of L. tridentata with an undergraduate student, Maria Strangas, during spring break (left). Flowering time is broadly overlapping at all mixed-cytotype sites (right), suggesting that reproductive isolation is largely effected by mechanisms other than the timing of flowering, such as habitat differences, pollinator discrimination, or post-zygotic hybrid lethality.
The native bees of the southwestern deserts are important pollinators for desert plants, including L. tridentata. However, many species of pollinators develop a "search image" for a specific plant type (species or cytotype, for example) during foraging bouts, which may result in pollen isolation between closely related plants. To determine if the pollinators visiting L. tridentata cytotypes differ, I have been collecting pollinators on plants of known cytotype at mixed-cytotype sites with the help of undergraduate students over spring break. Here, Adrian and Chan net bees on creosote bush in a diploid/tetraploid mixed site (left). So far, there does not appear to be a strong difference in pollinator assemblages between diploid and tetraploid L. tridentata (Discriminant Function Analysis, right), but there may be biases in the cytotype of pollen carried by some native bees. Additional sampling is needed and work continues.
Flow cytometry is the principal method for identifying the cytotypes of sampled plants. The technique uses a DNA-specific fluorescent dye and laser illumination to infer the genomic DNA content of individual plants (left). Using molecular phylogenetic analyses of both chloroplast and nuclear DNA sequences of L. tridentata I found that the three North American cytotypes are monophyletic with respect to the closest relatives found in xeric regions of South America, but that there was no differentiation among the North American cytotypes (right).
Population genetic and network analyses of chloroplast haplotype diversity (left) among the North American cytotypes of L. tridentata suggests a relatively recent demographic expansion. No fixed chloroplast DNA differences have been identified, however a single character difference largely differentiates diploid and polyploid haplotypes (right).
In combination with the identified chloroplast DNA (cpDNA) marker, I have developed Amplified Fragment Length Polymorphism (AFLP) markers to assay potential gene flow among the cytoypes. Theory predicts a depression of gene flow between cytotypes. This appears to be the case for L. tridentata, though analysis with the program STRUCTURE (left) suggests that diploids and tetraploids have somewhat admixed nuclear DNA genotypes where they come into contact, and some tetraploids have the diploid cpDNA haplotype. However, the rate of gene flow may differ between diploids/tetraploids and tetraploids/hexaploids. I have only identified two triploid hybrids at the diploid/tetraploid contact zone, but several pentaploid hybrids (right) at the tetraploid/hexaploid contact zone, and STRUCTURE analysis of AFLPs suggests even more genotypic admixture between tetraploids and hexaploids.
Justin Ramsey - Black Hills State University
Robert Minckley - University of Rochester
Julienne Ng - University of Colorado-Boulder
Stacey Smith - University of Colorado-Boulder
Eric Lyons - University of Arizona
Nolan Kane - University of Colorado-Boulder
John DeLong - University of Nebraska-Lincoln
Diana Pilson - University of Nebraska-Lincoln