UC Berkeley

Evolution - the great tinkerer - has produced the astounding diversity of form within

and between existing species. It is a fundamental goal of evolutionary biology to understand

the origin of such diversity. What types of genes underlie evolved changes in morphology?

Are certain types of mutations (notably changes within regulatory regions) more likely to be used to produce adaptive changes in form? When distinct populations evolve similar morphological changes, are the underlying genetic bases changes to the same genes, the same genetic pathways, or largely independent? Are changes in form modular, or are their concerted changes to multiple developmentally similar organs? The ever cheapening cost of sequencing, coupled the availability of high-quality reference genomes, allows high-throughput approaches to identifying the loci of evolution. The emergence of a robust genome engineering system, CRISPR/Cas9, allows for efficient and direct testing of a gene's phenotype. Combining both of these techniques with a model system with naturally evolved phenotypic variation, the threespine stickleback, allows for systems-level answers to the many evolutionary questions.

Chapter one outlines the field of evolutionary developmental biology. It proposes two

alternative viewpoints for thinking about the evolution of form. The first is the view of the

`Modern Synthesis', linking Mendelian inheritance with Darwinian natural selection, which

explains evolution as the change in allele frequencies over time. The second views evolution

through the lens of deep homology, focusing on changes to developmental programs over

time, even across related organs within the same animal. It then introduces key concepts

within evolutionary and developmental biology, including cis-regulation of gene expression,

and gene regulatory networks. It then provides examples of evolution reusing similar gene

regulatory networks, including Hox genes, Pax6 dependent eye initiation, and ectodermal

placode development. Teeth use highly conserved signaling pathways, during both their

initiation and replacement. Threespine sticklebacks Gasterosteus aculeatus have repeatedly

adapted following a shift from marine to freshwater environments, with many independently

derived populations sharing common morphological traits, including a gain in tooth number.

The following chapters investigate this gain in tooth number in multiple distinct populations

of sticklebacks.

Chapter two describes the discovery and mapping of a spontaneous stickleback albino

mutation, named casper. casper is a sex-linked recessive mutation that results in oculocutaneous albinism, defective swim bladders, and blood clotting defects. Bulked segregant mapping of casper mutants revealed a strong genetic signal on chromosome 19, the stickleback X chromosome, proximal to the gene Hps5. casper mutants had a unique insertion of a G in the 6th exon on Hps5. As mutants in the human orthologue of Hps5 resulted in similar albino and blood clotting phenotypes, Hps5 is a strong candidate underlying the casper phenotype. Further supporting this model, genome editing of Hps5 phenocopied casper. Lastly, we show that casper is an excellent tool for visualizing the activity of uorescent transgenes at late developmental stages due to the near-translucent nature of the mutant animals.

Chapter three details the fine mapping of a quantitative trail locus (QTL) on chromosome 21 controlling increases in tooth number in a Canadian freshwater stickleback population. Recombinant mapping reduced the QTL-containing region to an 884kb window. Repeated QTL mapping experiments showed the presence of this QTL on multiple, but not all, wild derived chromosomes from the Canadian population. Comparative genome sequencing revealed the perfect correlation with genetic data of ten variants, spanning 4.4kb, all within the 4th intron of the gene Bmp6. Transgenic analysis of this intronic region uncovered its role as a robust tooth enhancer. TALEN induced mutations in Bmp6 revealed required roles for the gene in stickleback tooth development. Finally, comparative RNA-seq between Bmp6 wild-type and mutant dental tissue showed a loss of mouse hair stem cell genes in Bmp6 mutant fish teeth, suggesting deep homology of the regeneration of these two organs.

Chapter four investigates the evolved changes in gene expression that accompany evolved increases in tooth number in two distinct freshwater populations. Independently derived stickleback populations from California and Canada have both evolved increases in tooth number, and previous work suggested that these populations used distinct genetic changes during their shared morphological changes. RNA-seq analysis of dental tissue from both freshwater populations compared to marine revealed a gain in critical regulators of tooth development in both freshwater populations. These evolved changes in gene expression can be partitioned in cis changes (mutations within regulatory elements of a gene) and trans changes (changes to the overall regulatory environment) using phased RNA-seq data from marine-freshwater F1 hybrids. Many genes show evidence for stabilizing selection of expression levels, with cis and trans changes in opposing directions. Most evolved changes in gene expression are due to changes in the trans environment, and these trans changes are more likely to be shared among the high-toothed freshwater populations. Thus, Californian and Canadian sticklebacks have convergently evolved similar trans regulatory environments through distinct cis regulatory changes.

Chapter five identifies candidate genes underlying evolved tooth gain in multiple geographically distinct freshwater populations. Many populations of freshwater sticklebacks have evolved increases in both oral and pharyngeal tooth number. QTL mapping of this evolved gain in pharyngeal tooth number revealed that a 438bp regulatory haplotype of Bmp6 is associated with increased tooth number in five distinct Pacific Northwest populations, though not in the high-toothed California population. QTL mapping of evolved oral tooth gain in California reveals the surprisingly modular nature of evolved changes in dentition. Correlation analysis of gene expression data from 33 separate samples across multiple populations and genotypes revealed Plod2 and Pitx2 as dentally expressed candidate genes underlying evolved tooth gain. CRISPR/Cas9 genome editing of Plod2 resulted in mutants displaying increases in pharyngeal but decreases in oral tooth number. Mutations in Pitx2 are homozygous lethal and show a recessive near-complete loss of dentition across all tooth fields. The pleitropic effects of the coding mutations and the lack of evolved coding changes suggest that modular regulatory changes to Plod2 and Pitx2 underlie increases in tooth number.

Combined, these results make significant contributions to our understanding of the evolutionary genetics underlying an adaptive change in morphology. Modular cis-regulatory alleles appear to play critical roles during the evolution of increased tooth number. Some alleles, such as the regulatory haplotype of Bmp6, are repeatedly used by multiple independently derived freshwater populations, suggesting both that the haplotype is adaptive and that evolution is partially repeatable. The Californian specific use of Plod2 and Pitx2 shows that evolution is not entirely predictable, and that there are many ways to modify teeth. Additionally, the use of high-throughput expression assays and genome sequencing, combined with genome editing with CRISPR/Cas9, allowed for rapid identification and testing of candidate genes underlying evolved changes in morphology. Additional studies could use these approaches to further identify the loci of evolved changes in morphology.