V.7 Evolution of Gene Expression Patricia J.Wittkopp

OUTLINE of a gene and that has allele-specific effects on gene expression. 1. The importance of regulatory evolution: a Co-option. Using existing functional parts of a genome historical perspective for new purposes. 2. Finding expression differences within and Ectopic Expression. Expression in cells that do not usu- between species ally express the gene of interest. 3. Genomic sources of regulatory evolution Orthologous Genes. Homologous genes that diverged 4. Enhancer evolution following a speciation event. 5. Evolution of transcription factors and Pleiotropy. Occurs when a mutation or gene affects more transcription factor binding than one phenotype. 6. Evolutionary forces responsible for expression Quantitative Trait Locus (QTL). Aregionofthegenome divergence shown to influence a (quantitative) phenotype of interest. Genetic changes affecting either the function or regu- RNA Interference (RNAi). A technique in which short RNAs lation of a gene product can contribute to phenotypic are used to interfere with the successful production of evolution. Studies of evolutionary mechanisms have his- proteins for a gene of interest. torically focused on changes in protein-coding sequen- Transcription Factor. A protein that binds to DNA in a ces, but during the last decade, multiple lines of evidence sequence-specific manner and affects transcription. have shown that changes in gene expression are at least equally important. The last few years have brought great progress in understanding the genetic basis of expression 1. THE IMPORTANCE OF REGULATORY EVOLUTION: differences within and between species. From a growing A HISTORICAL PERSPECTIVE collection of single-gene case studies and comparative analyses of gene expression on a genomic scale, common For most of the twentieth century, conventional wis- themes and patterns in regulatory evolution have begun dom among biologists was that, as François Jacob to emerge. described it, “cows had cow molecules and goats had goat molecules and snakes had snake molecules, and it GLOSSARY was because they were made of cow molecules that a cow was a cow.” Near the end of the twentieth century, Chromatin. The higher-order complex of DNA, histones, however, it became clear that this was not the case. and other proteins that packages nuclear DNA within Species-specific genes exist (see chapter V.6), but they a eukaryotic cell. are the exception rather than the rule; much of the Chromatin Immunoprecipitation. A technique in which biological diversity seen in nature is produced by genes transcription factors are cross-linked to DNA, the whose functions are highly conserved among species. DNA is sheared, and fragments binding to a specific Discovering this conservation was a boon to the med- transcription factor of interest are isolated using an ical genetics community, because it justified the use of antibody. Identity of the isolated DNA fragments model organisms such as fruit flies and mice to in- can be assessed by PCR, microarrays, or sequencing. vestigate human disease, but also presented a paradox: cis-regulatory Element. A DNA sequence (such as an how can divergent traits be constructed using con- enhancer or promoter) located near the coding region served genes? 414 Genes, Genomes, Phenotypes

The answer to this question is, in part, by modifying still are) critical for establishing links between divergent the regulation of gene expression. Expression of a gene is gene expression and divergence of a particular pheno- necessary before it can impact the phenotype of an or- type; however, they are not suitable for obtaining the ganism; that is, the DNA sequence encoding a gene prod- genomic measures of expression required to identify uct must be transcribed into RNA and then (usually, but global trends in the evolution of gene expression. Ra- not always) translated into a protein before the gene can ther, microarrays, which are short DNA sequences com- function in a cell. Each cell expresses only a subset of the plementary to transcribed sequences from a particular genes in its genome, and the specific genes expressed species arrayed onto a filter or a microchip, have been determine a cell’s fate (see chapter V.11). In 1969, before used to quantitatively compare the abundance of RNA the molecular details of gene regulation were known, from hundreds to thousands of expressed genes in the Roy J. Brittan and Eric H. Davidson proposed a theory genome simultaneously. Today, microarrays are largely for the regulation of gene expression in eukaryotic cells. being replaced by a method known as RNA-seq that uses They viewed gene regulation as integral to evolution and massively parallel sequencing to obtain quantitative mea- suggested that differences among species could be at- sures of gene expression (i.e., RNA abundance). Tech- tributable to changes in the regulation of gene expres- niques for measuring protein abundance (which is not sion. Six years later, Mary-Claire King and A. C. Wilson always highly correlated with RNA abundance) on a ge- published a seminal paper showing that the amino acid nomic scale are also available (e.g., two-dimensional gel sequences of homologous human and chimpanzee pro- electrophoresis, mass spectrometry), but thus far they teins appeared to be more than 99 percent identical. have not been used to compare protein expression gen- Based on this result, they argued that the degree of pro- ome-wide in an evolutionary context. tein divergence was insufficient to account for the ex- By contrast, the transcriptome (i.e., the collection of tensive morphological, physiological, and behavioral all RNAs expressed in a biological sample) has been differences between these two species. analyzed in a wide variety of taxa, including human, Despite these (and similar) predictions more than 35 mice, fishes, flies, yeast, and plants. Comparing tran- years ago, the idea that changes in gene expression might scriptomes has shown that differences in RNA abun- be a common source of phenotypic divergence did not dance are common both within and between species and gain mainstream acceptance among evolutionary biol- that the number of genes showing expression differences ogists until after the turn of the twenty-first century. between a pair of species is often proportional to their Seeds of this acceptance were sown when developmental divergence time. For example, in one of the first pub- biologists, using newly developed tools for visualizing lished transcriptome comparisons between species, mi- gene expression, began comparing expression among croarrays containing sequences complementary to ap- species. This approach catalyzed the expansion of evolu- proximately 12,000 human genes were used to measure tionary developmental biology, a field of research known mRNA abundance in white blood cells, liver, and brain today as evo-devo. Within a few years, researchers ac- of humans, chimpanzees, orangutans, and macaques. quired many examples of cases in which divergent RNA Comparing expression in samples from three humans, and/or protein expression of genes known to be im- three chimpanzees, and one orangutan showed exten- portant for development correlated with morphological sive variation within both humans and chimpanzees. divergence between species. Such correlations suggest The extent of expression divergence between humans that the genetic changes responsible for altered gene ex- and chimpanzees was smaller than the divergence ob- pression might be the same changes responsible for al- served when either of these species was compared to the tered phenotypes. In parallel, quantitative geneticists orangutan, suggesting that expression divergence cor- mapping the mutations responsible for phenotypic dif- relates with phylogenetic distance. In the samples de- ferences among individuals of the same species or (less rived from brains, one human was found to differ more commonly) different species were finding that changes in from another human than from a chimpanzee, but this protein sequence could not always account for the phe- type of relationship is rare: polymorphic gene expression notypic effect of a quantitative trait locus (QTL) (see within a species is typically less extensive than divergent chapter V.12). gene expression between species. In a slightly different experiment, macaques were 2. FINDING EXPRESSION DIFFERENCES WITHIN used as an outgroup, and gene expression in humans was AND BETWEEN SPECIES found to have evolved faster in the brain than in the liver or blood. Although it is tempting to speculate that this Early comparative studies of gene expression focused on apparently accelerated evolution of gene expression in one or a small number of genes within or between spe- the human brain may have contributed to the evolu- cies. These low-throughput types of studies were (and tion of human-specific cognitive abilities, a reanalysis of Evolution of Gene Expression 415 these data that more completely modeled the sources of a particular place, time, or environment. Because of their variance in the experiment found more genes with dif- more limited effects on an organism (i.e., lower pleio- ferential expression in the liver than in the brain between tropy), enhancers are commonly thought to be more humans and chimpanzees. This example illustrates the likely to harbor mutations that survive in natural po- potential tremendous impact of statistical analysis me- pulations and give rise to polymorphism and divergence thods on the conclusions drawn from this type of work. than mutations in basal promoters or coding sequences Particularly problematic in this case (and in other cases of transcription factors. where a microarray with sequences from one species Chromatin can also have cis-regulatory effects on gene is used to compare expression between species) is ac- expression. Like the rest of the genome, cis-regulatory counting for the effects of sequence divergence between sequences are wrapped around histones and packaged the microarray probes and the heterologous species. The into nucleosomes that form chromatin structure. The newer RNA-seq method of quantifying and comparing state of chromatin influences interactions between cis- RNA abundance among species circumvents this pro- regulatory sequences and trans-regulatory factors, thus blem, but presents its own set of challenges for proper it is also an important component of transcriptional reg- data analysis and interpretation. ulation. Methods suitable for comparing chromatin structure within and between species have recently become 3. GENOMIC SOURCES OF REGULATORY EVOLUTION available and researchers are investigating how chromatin structure evolves, as well as how this evolution im- Heritable differences in the distribution of RNA or pro- pacts gene expression. Many transcription factors are tein within or between species often result from changes known to modify chromatin, for example by acetylating in the sequence of genomic DNA. To understand the or deacetylating histones, so changes in cis-regulatory types of sequences in the genome that can be mutated to sequences affecting binding of such transcription fac- alter gene expression, one must consider the molecular tors could be responsible (at least in part) for differ- mechanisms controlling transcriptional and posttran- ences in chromatin structure when they are observed (see scriptional regulation of gene expression. Within pro- chapter V.8). karyotes and eukaryotes, these mechanisms are highly Determining whether an expression difference be- conserved, but they differ significantly between the two tween two genotypes is caused by genetic changes in cis- groups. The remainder of this chapter focuses solely on or trans-regulation can be done using transgenic analy- transcriptional regulation in eukaryotes because it has sis, allele-specific expression, or genetic mapping. In the been studied most extensively in an evolutionary con- first two cases, activity of homologous cis-regulatory text. Also, the term gene expression is used synony- sequences controlling a divergent expression pattern of mously with transcription from this point forward. interest are compared in the same cellular environment When, where, and how much mRNA is produced (i.e., when regulated by the same set of trans-acting from a particular gene is determined by its cis-regulatory factors). If a difference in the activity of the two cis- DNA sequences as well as the trans-regulatory tran- regulatory sequences is observed, this indicates that scription factor proteins and noncoding RNAs present there has been functional cis-regulatory divergence. This in a cell. These cis-regulatory DNA sequences include test can be performed by using transgenes to move cis- the basal promoter that binds to RNA polymerase and regulatory sequences from species A into the trans- its associated cofactors as well as one or more enhancers acting genetic background of species B (and vice versa) that encode instructions for spatiotemporal expression (figure 2A) or by simply crossing the two genotypes and and the amount of mRNA to produce (figure 1). Basal testing for differences in allele-specific expression when promoter sequences are located near the transcriptional the two cis-regulatory alleles are in the same hetero- start site and are more highly conserved than enhancer zygous trans-acting genetic background (figure 2B). sequences, because they bind to transcription factors Putatively cis- and trans-acting changes can also be in- such as the TATA-binding protein required for tran- ferred from genetic mapping, in which regions of the scription of most genes. Enhancer sequences typically genome contributing to the expression difference of in- comprise a few hundred base pairs, can be located up- terest are identified. If such a region is located close to the stream (5’), downstream (3’), or in an intron of the as- affected gene, it is assumed to act in cis; if such a region is sociated gene (figure 1), and are bound by transcription located far from the affected gene, it is assumed to act factors that activate expression from the basal promoter in trans. in a subset of cells or under a subset of environmental As a group, studies using transgenes to investigate conditions. In multicellular eukaryotes, expression of regulatory evolution provide evidence for both cis- and a gene tends to be controlled by multiple enhancers, trans-regulatory changes underlying expression diver- each acting independently and controlling expression in gence, with cis-regulatory divergence detected most often. 416 Genes, Genomes, Phenotypes

DNA Enhancer EnhancerExon Exon Enhancer Exon

Basal Intron Intron TF binding site promoter Figure 1. Basic eukaryotic gene structure. Cis-regulatory sequences compose the RNA polymerase II complex bind to sequences in the include enhancers and the basal promoter. Most transcription factors basal promoter (neither are shown). (TFs) bind to sequences in enhancers, and transcription factors that

Allele-specific expression has been used to examine sour- be produced by multiple arrangements of transcription ces of polymorphic and divergent expression genome- factor binding sites, and most transcription factors bind- wide in flies, yeast, and plants, and these data suggest ing sites are degenerate, meaning that the same tran- that trans-acting variation is the predominant source of scription factor can bind to multiple sequences. expression differences among individuals of the same This complex relationship between enhancer sequence species, whereas cis-regulatory changes play a larger role and function has been nicely illustrated by comparative in expression divergence between individuals of differ- studies of two Drosophila enhancers whose activity ap- ent species. Genetic mapping of expression differences pears to be conserved between species. In the first case, the has thus far been limited to variation within a species, DNA sequence and transcription factor binding sites of but it also shows an abundance of variants with apparent an early embryonic enhancer (controlling “stripe 2” ex- trans-acting effects on gene expression segregating with- pression of the even-skipped gene) have been extensively in a species. Both genetic mapping and allele-specific changed between species, yet the function of the elements tests show that although cis-regulatory polymorphisms remains the same. Orthologous cis-regulatory elements within a species are more rare than trans-regulatory from D. melanogaster and D. pseudoobscura had similar polymorphisms, they tend to have larger effects on ex- activities in transgenic D. melanogaster, whereas chi- pression and are less likely to be recessive than trans- meric enhancers containing the 5’ half from one species regulatory variants. The additivity of cis-regulatory mu- and the 3? half from the other showed abnormal activity. tations, combined with the expected lower levels of Similarly, extensive rearrangement of transcription factor pleiotropy relative to trans-acting mutations, may also binding sites was found in an enhancer driving con- make them more likely to contribute to phenotypic served expression in the developing eye of Drosophila. evolution. The D. melanogaster allele of this enhancer has been extensively analyzed, allowing predictions to be made 4. ENHANCER EVOLUTION about the consequences of some observed changes. These types of analyses provide insight not only into evolu- As described above, enhancer sequences are an im- tionary processes, but also into enhancer architecture portant source of evolutionary change. This class of cis- in general. regulatory sequences has been studied in the most detail If enhancer sequences can change extensively and during the last decade, and these studies have revealed a still retain their original function, how much does an complex relationship between the DNA sequence and enhancer need to change to acquire new activities? A function of an enhancer. The evolution of enhancer number of studies have been published during the last sequences that are functionally conserved, functionally few years, most notably from the laboratories of Sean B. divergent, and those that have acquired novel activities Carroll, David M. Kingsley, and David L. Stern, that are are discussed below. suitable for addressing this question (see chapter V.12). Because enhancer sequences are critical for proper In some cases, as little as a single nucleotide change is development and physiology of an organism, most mu- sufficient to account for the divergent activity of an en- tations that alter their activity are expected to be dele- hancer, whereas, in others, multiple mutations (on the terious and removed from a population by purifying order of 10 or fewer) are responsible for expression dif- selection. Consequently, enhancer sequences should be ferences. In addition to single nucleotide changes, larger more highly conserved than surrounding nonfunctional lesions also contribute to divergent activity. For ex- DNA. In fact, they are, and this conservation is a helpful ample, in the threespine stickleback, recurrent deletions tool for finding enhancers within a genome (see chapter that disrupt the activity of an enhancer contribute to the V.3). The degree of sequence conservation in an en- repeated loss of pelvic structures in freshwater popula- hancer is typically lower than that of protein-coding se- tions. In Drosophila, deletions in an enhancer of the quences, however, because of the structure-function re- desatF gene have been shown to contribute to expression lationship of enhancers: the same enhancer activity can divergence by (surprisingly) creating novel binding sites A Using transgenes to compare spatiotemporal expression

Transgenic host species Species A Species B

Native expression

Species A or from species A from A

CRE Control cis trans divergence divergenceof factor of CREs regulating CREA

Species B CRE donor species

or from species B from B

CRE cis trans Control divergence divergence of factor of CREB regulating CREB

B Using allele-specific expression to compare RNA abundance

F1 hybrid Genotype 1 Genotype 2 (geno 1/geno 2) If change is in: cis trans # RNAs # RNAs # RNAs # RNAs 10 20 10 15 10 20 20 15

cis-acting trans-acting DNA protein Figure 2. Determining whether divergent expression is due to a ch- the promoter location indicated by an arrow. The solid black line ange in cis-and/ortrans-regulation. (A) Transgenic analysis can represents DNA, including the CRE, as indicated. Circles and trian- distinguish between cis-andtrans-regulatory divergence by com- gles represent two different transcription factor proteins, each of paring the activity of orthologous cis-regulatory elements (CREs) in which is present in multiple copies per cell. Hypothetical numbers of the presence of the same set of transcription factors. This can be RNA molecules produced by each allele in each cell (# RNA) is also done by creating a pair of artificial genes, each with a CRE controlling shown. The F1 hybrid contains a CRE allele and transcription factors expression of a protein that is easy to detect. These so-called reporter from each of its parental genotypes. If the expression difference ob- genes are then introduced into the genomes of the two species from served between genotypes 1 and 2 is due solely to cis-regulatory which the CREs were derived. Different patterns of expression are changes (i.e., the trans-acting transcription factors are equivalent expected if cis-ortrans-regulatory changes occur between species; a between genotypes), each allele produces the same number of RNA hypothetical example of this is shown in which each box represents molecules in the F1 hybrid as it did when homozygous in genotype a region of tissue, and gray represents either native expression in 1 or 2. If, on the other hand, the cis-regulatory sequences are func- species A and species B or the expression of the CRE tested in each tionally equivalent between alleles and the difference in RNA abun- species using a reporter gene. (B) Measures of allele-specificRNA dance observed between genotypes 1 and 2 results from differences abundance can also be used to distinguish between cis-andtrans- in trans-acting factors between genotypes, the two CRE alleles in the regulatory changes in diploid organisms. Schematic representations F1 hybrid will produce an equal number of RNA molecules, with of cells from two different (homozygous) genotypes (two different the precise number (15, in this example) determined by the specific species or two different genotypes from the same species) are shown. type of trans-regulatory divergence. Combinations of cis-andtrans- A schematic cell from an F1 hybrid produced by crossing genotype 1 regulatory changes are also possible, with the cis-regulatory differ- and genotype 2 is also shown. In each cell, two copies of a gene are ence always reflected in relative expression between the two alleles shown with the transcribed region indicated by a grey rectangle and in the F1 hybrid. 418 Genes, Genomes, Phenotypes for an unknown transcription factor that activates ex- sequences. Recently, techniques for monitoring binding of pression. In the few cases where multiple changes have a particular transcription factor genome-wide have been been implicated in expression divergence and their ef- developed, and comparative studies show that the gain fects tested individually, the substations have been found and loss of TF binding sites is very common among spe- to interact in a nonadditive (i.e., epistatic) fashion. cies. Between closely related species, changes in the The majority of work on enhancer evolution has fo- quantitative binding of a TF to a particular site rather than cused on cases in which enhancer activity is either con- the gain or loss of individual binding sites appears more served or divergent. But what about new enhancers? How prevalent. In the next few years, these types of experi- do they evolve? Simulations suggest that new point muta- ments, which rely on chromatin immunoprecipitation, tions could frequently generate novel transcription factor will likely be combined with genomic measures of gene binding sites and that they could fix over microevolu- expression and cis-regulatory sequence divergence to tionary timescales, even in the absence of selection. This provide a more complete understanding of how changes suggests that new enhancers driving novel expression pat- in DNA sequence impact TF binding, and how this in terns might frequently arise de novo. Despite this finding, turn affects gene expression. all the cases of (putatively) novel enhancers characterized As described above, many of the TFs functionally to date appear to have evolved using other mechanisms tested in vivo show conserved functions between species, (i.e., duplication and divergence, transposition, or co- but this is not always the case—even for highly pleio- option), with co-option (i.e., repurposing) of existing reg- tropic regulators of development. For example, the func- ulatory elements the most common mechanism—in both tion of the HoxA-11 protein has acquired a novel func- fruit flies and primates, cis-regulatory sequences control- tion required for pregnancy in placental mammals, and ling novel expression patterns have been shown to include the Hox genes fuzi tarazu and Ultrabithorax have sites required for one or more preexisting enhancers. diverged between Drosophila melanogaster and other insects. In each case, the proteins seem to have retained 5. EVOLUTION OF TRANSCRIPTION FACTORS AND some ancestral functions while gaining and losing others. TRANSCRIPTION FACTOR BINDING 6. EVOLUTIONARY FORCES RESPONSIBLE FOR To function, cis-regulatory sequences must be bound by EXPRESSION DIVERGENCE transcription factors (TFs), which are proteins that bind to specific DNA sequences and influence (i.e., either acti- With differences in mRNA expression cataloged for a vate or repress) transcription. Molecularly, TFs typically variety of species, researchers are now faced with the contain a DNA binding domain, one or more protein- daunting task of figuring out what these expression dif- protein interaction domains, a transcriptional activation ferences mean for organismal phenotypes, especially fit- or repression domain, and sometimes a chromatin mod- ness. Classic genetic mutants and reverse genetic tech- ification domain. As a group, genes encoding TFs are niques such as RNA interference (RNAi) can be used to among the most highly conserved in eukaryotic gen- assess the function of individual genes, but these tech- omes, especially in their DNA binding domains. This niques are rarely able to predict the consequences of the high degree of similarity among species is seen not only quantitative changes in expression commonly found in in terms of protein sequence, but also with functional nature. To complicate matters further, mRNA levels do tests. Perhaps the most seminal of these tests showed that not always correlate with protein abundance, and si- the Drosophila eyeless and mouse Pax-6 genes are or- milar changes in expression of different genes will al- thologous genes, and that ectopically expressing either most certainly have different effects; for example, a 10 of them in developing Drosophila wings or legs was suf- percent change in expression of one gene might have a ficient to transform cells into ectopic eyes. Importantly, larger effect on the phenotype than a 1000 percent change both the Drosophila and mouse alleles of this gene in- in expression of another gene. Connecting changes in duced similar morphological transformations, with cell gene expression to specific phenotypes is currently best types and organizational structures resembling the nor- done by studying one gene and one phenotype at a time; mal Drosophila eye. This study, and others like it that however, high-throughput phenotyping strategies cur- followed, demonstrated that development is often con- rently being developed should soon make it possible to trolled by highly conserved master regulatory proteins. address this question more systematically. Conserved master regulatory proteins such as Pax-6 Knowing the impact of a change in gene expression can create divergent structures by regulating different sets on fitness can help determine the likelihood that the of target genes in different species. Changes in the identity change resulted from natural selection. Assessing the of target genes are mediated by the evolution of TF bi- relative roles of neutral and non-neutral processes is a nding, resulting primarily from changes in cis-regulatory major challenge for evolutionary biology in general. Evolution of Gene Expression 419

To date, three main strategies have been used to in- provided some of the earliest experimental evidence of the vestigate the role of natural selection in the evolution of importance of changes in gene expression for evolution, gene expression: the comparative method, tests of neu- one of the current leaders of the evo-devo field takes a look trality, and empirical patterns (see chapter V.14). In the at the experimental evidence supporting this assertion comparative method, evidence of natural selection is today. Carroll, S. B., J. K. Grenier, and S. D. Weatherbee. 2004. inferred when a change in expression is found to corre- From DNA to Diversity: Molecular Genetics and the late with an environmental or other biological factor in Evolution of Animal Design. 2nd ed. New York: Wiley- a manner that exceeds the correlation expected simply Blackwell. This book describes the central role of gene because of shared ancestry. To use tests of neutrality, regulation in development and evolution with accessible patterns of regulatory evolution expected from neutral discussions of specific case studies beautifully illustrated. processes must be specified or inferred from the data Davidson, E. H. 2006. The Regulatory Genome: Gene Reg- available. Studies of mutational variance for gene ex- ulatory Networks in Development and Evolution. San pression provide a starting point for developing these Diego, CA: Academic. This book provides a comprehen- neutral models, but much more remains to be learned sive summary of the data and logic behind the assertion about the neutral expectations for regulatory evolution. that gene regulatory networks are essential for development and play a critical role in evolution. Finally, empirical patterns, especially comparisons be- Halder, G., P. Callaerts, and W. J. Gehring. 1995. Induction tween polymorphism and divergence for expression of a of ectopic eyes by targeted expression of the eyeless gene particular gene, capture elements of regulatory variation in Drosophila. Science 267: 1788–1792. This landmark that cannot easily be incorporated into neutral models. study demonstrated the remarkable conservation of reg- With this approach, one or more representative “base- ulatory proteins by showing that homologous genes from line” genes assumed to be evolving neutrally are used as species as diverse as fruit flies, mice, and humans all had references to test for selection, but it is generally not clear similar effects on development when introduced into the which genes should be considered to be evolving neu- fruit fly. trally. Presently, there is no consensus about the relative Stern, D. L. and V. Orgogozo. 2008. The loci of evolution: roles of selection and drift in shaping regulatory evolu- How predictable is genetic evolution? Evolution 62: 2155–2177. After clearly describing the rationale and al- tion, although most species show a strong signal of sta- ternative models for regulatory evolution, this review bilizing selection within a species, indicating that ex- provides the most thorough summary to date of studies pression levels do matter for fitness. identifying the types of genetic changes responsible for Regardless of the evolutionary forces underlying the a divergent phenotype, contrasting the relative fre- evolution of gene expression, understanding how this quency of changes attributable to coding and to regulatory important molecular phenotype evolves is a critical com- mutations. ponent of understanding how the evolutionary process Wittkopp, P. J., and G. Kalay. 2012. Cis-regulatory elements: works. The pressing question is no longer whether chan- Molecular mechanisms and evolutionary processes un- ges in gene expression contribute to phenotypic evolu- derlying divergence. Nature Reviews Genetics 13: 59–69. tion, but rather when and how they do. The develop- This review uses data from case studies revealing the genetic and molecular changes responsible for divergent cis- ment of many new tools for studying gene expression regulatory sequences to examine how these types of se- combined with the recent rapid accumulation of expres- quences evolve and to gain insights into more general sion and transcription factor binding data, suggest that questions about the mechanisms of evolution. researchers may be able to answer these questions soon. Wray, G. A., M. W. Hahn, E. Abouheif, J. P. Balhoff, M. Pizer, M. V. Rockman, and L. A. Ramano. 2003. The evolution of FURTHER READING transcriptional regulation in eukaryotes. Molecular Biology and Evolution 20: 1377–1419. Despite being nearly a dec- Carroll, S. B. 2005. Evolution at two levels: On genes and ade old, this review remains one of the most complete dis- form. PLoS Biology 3: e245. In honor of the thirtieth cussions of the mechanics of gene regulation and its re- anniversary of the seminal work of King and Wilson, who lationship to evolution.