UC Berkeley

How populations adapt to their local environment has been a central question in evolutionary biology for the history of the field. Darwin’s evolution by natural selection provided a means for how pressures such as predation or resource availability, can select for variations of traits. It was not until the rapid growth of the field of genetics that a mechanism for heritable variation could be identified.The intersection of population genetics, developmental, evolutionary, and cell biology, allows researchers to probe the underlying means by which populations have been able to adapt. In countless species regions of the genome and genes have been identified that are responsible for evolution by looking for patterns within population genetic data. For example, regions that are highly divergent between populations may drive adaptation to either of the environments. In laboratory crosses quantitative trait loci (QTL) mapping can identify possible candidate genes in the evolution of a trait of interest. Species that have evolved a trait multiple times like armor plating reduction in the threespine stickleback (Gasterosteus aculeatus) spark the question: is the same gene, or allele of a specific gene, re-used each time, or does each instance of evolution have a unique genetic mechanism? While the answer is undoubtedly somewhere between the two extremes, where exactly the answer falls on the spectrum is constantly being re-evaluated. The threespine stickleback is an ideal system to study how populations adapt. A small, marine species, found throughout the Northern Hemisphere, the stickleback has repeatedly colonized an incredible range of freshwater habitats. Following colonization, sticklebacks tend to evolve certain traits, most famously a loss or reduction of armor plating and pelvic spines. For the majority of traits in which candidate genes have been identified, it appears that changes in gene regulation is what drives the differences in phenotype between the ancestral marine and derived freshwater ecotypes, and often the same gene or even allele is re-used. Overall, very few cases have implicated coding changes in the evolution of specific traits. The rarity of coding mutations underlying examples of stickleback evolution reflects a broader conversation in evolutionary biology. As genetics and molecular biology matured the incredible degree of homology between orthologous proteins across the tree of life led to the hypothesis that evolution can occur not only by mutations in coding regions, thereby changing the proteins, but by modifying where, when, and how genes, and subsequently proteins are expressed. Cis-regulatory elements control gene transcription, causing tissue and time specific expression of the gene, as well as modulating the intensity of expression. Changes in cis-regulatory regions can reduce organism wide, or pleiotropic, effects, compared to coding changes, and target specific aspects of a gene’s activity or function. Genes essential for development are controlled by multiple regulatory elements and so mutations in these elements can lead to evolutionary change. The central focus of this dissertation serves as an example of each of the two previously stated concepts in evolutionary biology: 1) adaptation through the re-use of alleles and 2) adaptation through changes in the cis-regulation of developmental genes. Another trait that can evolve following colonization of freshwater environments by marine fish is an increase in pharyngeal tooth number, and is thought to be driven by cis-regulatory changes. The increase in tooth number is thought to be an adaptation to a newly available niche or resource. When comparing a high toothed population from Paxton Lake and a low toothed marine population, the tooth numbers are similar in young fish but later in development a difference arises. Using QTL mapping, the candidate gene Bmp6 was identified as potentially underlying evolved tooth gain, a result replicated in multiple freshwater populations. Allele specific expression data found differences in Bmp6 expression between the marine and freshwater allele that occurred in a similar time frame as the tooth number difference, suggesting a change in regulation of the gene underlies the tooth number divergence. Comparing genomic sequence of chromosomes that had an effect on tooth number and those that did not yielded a set of six single nucleotide polymorphisms (SNPs) upstream of a tooth enhancer for Bmp6 that co-occur with the presence of the QTL peak. I hypothesize the SNPs modulate the activity of the tooth enhancer, resulting in differences in Bmp6 expression and thereby affecting tooth number. The six SNPs define a high-tooth associated haplotype which was identified in multiple QTL crosses with fish from different, geographically isolated populations and is, therefore, likely standing genetic variation. If the haplotype underlies the evolved tooth gain in multiple populations then it would be another example of evolution “re-using” alleles, and an example of evolution via changes in cis-regulation.

This dissertation can be broken down into three questions:1. Do high and low tooth associated alleles of the Bmp6 tooth enhancer drive different expression patterns? 2. Can the alleles be replaced through editing, exchanging a high-tooth associated allele with a low-tooth associated/marine allele in a freshwater fish? 3. Has the high-tooth associated allele experienced selection in wild populations? The answers from the three questions can begin to address a larger question: Is the haplotype responsible for evolved tooth gain, and if so through what mechanism?

Chapter 1 explains the scope of the dissertation, the two big ideas that form the current model for potential adaptation through the haplotype: evolution via cis-regulatory changes in developmental genes and the use of standing genetic variation to drive local adaptation. In both instances, the threespine stickleback is an ideal system for both the repeated “natural experiments” and prevalence of cis-regulatory examples already characterized in the species. The chapter also details the history of the haplotype in research and the evidence, before this dissertation work, that supports cis-regulatory changes in the gene as driving evolved tooth gain.Chapter 2 details expression pattern differences between a high-tooth haplotype containing and low-tooth haplotype containing alleles of the Bmp6 tooth enhancer. Two methods were used, with independent transgenes for each enhancer/reporter, and a single bicistronic construct that allows comparison of enhancers within a single organism, while controlling for positional insertion effects. Expression differences in both domains of the tooth enhancer, the condensed mesenchyme and the overlying epithelium of pre-eruption teeth, were identified. Patterns were consistent across multiple integrations and methods. Differences in the extent of expression domains, stages of fish, and tooth plates (ventral compared to dorsal), mirrors previous work such as the Bmp6 allele specific expression result, suggesting the haplotypes may have a role in regulating Bmp6 and subsequently, impacting tooth number. In chapter 3, the CRISPR/Cas9 system is used to replace a high tooth associated allele in freshwater stock fish with a low tooth associated/marine allele. CRISPR/Cas9 has had success inducing insertions or deletions for gene knockout experiments and repairing small mutations. More rarely has it been used to replace large stretches of sequence, and in most instances where it has been used for that purpose, the sequences being exchanged are either non-homologous, or divergent orthologues. In this experiment, the sequences have high degrees of similarity making partial replacement a potential outcome. Multiple chimeric alleles were created, in some case containing just portions of replaced haplotype, while in other cases complete replacement appears to have occurred. As the replacement was demonstrated in F0 fish a phenotypic effect of the replacement has not been determined. However, multiple partial or complete replacement alleles have been propagated and subsequent experiments could examine tooth number in the transgenic lineages, determining the impact of different chimeric alleles on the trait, and even potentially testing individual sites within the haplotype. Lastly, in chapter 4, whole genome sequence data from a wild population, Fishtrap Creek, was scanned for signals of natural selection surrounding the haplotype. The creek was formed approximately 10,000 years ago and likely colonized soon after. The haplotype is segregating within the population and was found in other populations in the Pacific northwest, expanding the range of the allele and supporting the model it was present as standing genetic variation. Multiple metrics were calculated for the entire genome of the data set, including pairwise nucleotide diversity , Tajima’s D, iHS, and nSL. A relatively new method, using ancestral recombination graphs with the program Relate, was also performed. Overall, variation existed in the scores for the single site metrics (iHS, nSL, Relate) and consistently resulted in the same site having elevated scores compared to the others within the haplotype. The site creates a binding site for the transcription factor NFATc1 which has a role in maintaining stem cell quiescence in hair follicle stem cells. This result supports further experiments, such as those in chapters 1 and 2 that focus on this specific site as a potential causative mutation for the observations in chapter 1.