UC Berkeley

The biosphere contains an incredible level of natural morphological diversity, and most differences within and between species can be explained by evolved differences in their genetic code. While traditional genetics has made great strides to connect genes to phenotypes in laboratory strains of model organisms, understanding the link between genotype and phenotype in natural populations is one of the greatest challenges of modern biology. Acquiring the genome sequences of organisms has been facilitated by rapidly advancing technologies, but connecting genetic variants to evolved differences remains elusive. What types of genetic changes underlie adaptive differences in morphology? How predictable is the path of evolution? Do individual mutations control multiple adaptive phenotypes? Answering these questions requires harnessing the power of modern genetics in a system with naturally evolved phenotypic variation.

Chapter one outlines key questions in the fields of evolutionary developmental biology and adaptation genetics. It describes the cis-regulatory hypothesis for the genetic basis of morphological evolution and occurrence of supergenes that control multiple evolved phenotypes. It describes the natural history of the threespine stickleback fish and explains why the stickleback system is an outstanding model to tackle these questions. Gasterosteus aculeatus has adapted to unique habitats across the Northern Hemisphere and is amenable to both forward and reverse genetic studies. Marine and freshwater sticklebacks eat different foods and have different adaptations in the skeletal elements used to process food. Quantitative trait locus (QTL) mapping has demonstrated that a large number of genomic regions control the evolution of these skeletal traits, but a few key regions control a disproportionate number of traits. The following chapters investigate the developmental and genetic bases of two evolved skeletal changes that are controlled by the same genomic regions.

Chapter two explores the genetic and developmental basis of the elongation of the branchial bones of the throat in freshwater sticklebacks. Elongation of these bones expands the buccal cavity, likely enabling freshwater fish to consume larger prey items. This increase in bone length is found in both wild and lab-reared fish from two populations, suggesting heritable convergent evolution in freshwater environments. In one population, an early increase in cartilage size contributes to increased bone length, and in both populations the bones grow faster throughout development. In both freshwater populations, the increase in bone length maps to two chromosomes: 4 and 21, with distinct effects of these two chromosomes on individual bones over the course of development, but similar effects in each cross. Collectively, these results suggest a largely parallel genetic and developmental basis of evolved bone length gain in two populations.

Chapter three describes further mapping and functional testing of the chromosome 21 bone length QTL. While pharyngeal tooth gain maps to a regulatory haplotype of the gene Bone Morphogenetic Protein 6 (Bmp6), bone length gain maps to a nearby region containing the gene Tfap2a in two freshwater populations of sticklebacks. Therefore, evolved pharyngeal tooth gain and bone length gain are controlled by separate loci. Tfap2a is an important transcriptional regulator of craniofacial development and produces severe craniofacial phenotypes when mutated in vertebrates. In sticklebacks, the freshwater allele of Tfap2a is downregulated in the developing branchial skeleton of hybrid animals and deletion of Tfap2a causes a nearly complete absence of pharyngeal arch-derived skeletal elements. Heterozygous loss of Tfap2a alters branchial bone length, suggesting that dosage of this gene is important to determining bone patterning. Combined with previous findings in the lab, these results suggest that closely linked regulatory changes to two key developmental patterning genes produce skeletal gain phenotypes.

Chapter four investigates the extent of genetic parallelism for repeated phenotypic evolution. In British Columbia, several lakes have independently evolved two freshwater stickleback ecotypes: a bottom-dwelling benthic form and an open-water limnetic form. Using crosses of benthic populations from three lakes, this study tests whether the genetic architecture underlying skeletal differences between benthic and marine individuals is repeatable across lakes. The majority of genomic regions underlying skeletal differences are unique to an individual lake, but there is more parallelism of QTL than expected by chance in simulations. Furthermore, the chromosome 21 QTL controlling bone length and tooth number were identified in multiple lakes, suggesting that these loci may be adaptive in the benthic habitat. These findings suggest that benthic evolution in three lakes has a significantly parallel but largely nonparallel basis.

Chapter five examines the regulation of the gene Bmp6, which, like the bone length QTL, is found on chromosome 21 and likely underlies evolved tooth gain in sticklebacks via a cis-regulatory down-regulation of the freshwater allele. A short conserved regulatory element upstream of Bmp6 drives robust reporter gene expression during tooth development in both sticklebacks and distantly related zebrafish. This enhancer responds to TGFß signaling, likely via SMAD3 binding, and the enhancer is required for normal expression of Bmp6. Therefore, changes to additional regulatory loci controlling Bmp6 and interacting with this enhancer may underlie pharyngeal tooth number evolution.

Finally, the future of stickleback molecular genetics will rely on functional genetic manipulations that will be facilitated by the emerging genome-editing revolution. The Appendix outlines a protocol for generating transgenic sticklebacks (carrying both transgenes and genome-edited alleles) using techniques developed and optimized over the course of the experiments described in chapters three and five. This protocol is intended to serve as a resource for the fish evolution and development community.

Combined, the results described here offer several insights towards the molecular genetic and developmental basis of evolved skeletal change. Two adaptive alleles controlling related traits (tooth number and bone length gain) are found tightly linked in the genome, indicating that linkage of the QTL controlling these phenotypes may be adaptive for rapid colonization of freshwater habitats. Both QTL are associated with cis-regulatory down-regulation of candidate genes with highly pleiotropic roles during development. This finding suggests that skeletal gain traits may be readily accomplished by a loss of gene expression. Future studies will attempt to identify the causative mutations responsible for each trait and examine their frequencies and evolutionary histories in natural stickleback populations. Additional studies will attempt to identify the precise developmental effects of the regulatory mutations underlying the evolved differences.