Craniofacial Divergence and Ongoing Adaptation Via the Hedgehog Pathway Reade B

Craniofacial divergence and ongoing adaptation via the hedgehog pathway Reade B. Robertsa, Yinan Hub, R. Craig Albertsonb,c, and Thomas D. Kochera,1 aDepartment of Biology, University of Maryland, College Park, MD 20742; bDepartment of Biology, Syracuse University, Syracuse, NY 13244; and cDepartment of Biology, University of Massachusetts Amherst, Amherst, MA 01003

Edited by Clifford J. Tabin, Harvard Medical School, Boston, MA, and approved June 27, 2011 (received for review December 8, 2010) Adaptive variation in craniofacial structure contributes to resource riaclima is a generalist with gracile jaws better adapted for suc- specialization and speciation, but the genetic loci that underlie tion feeding (Fig. 1 A and B). These two genera represent craniofacial adaptation remain unknown. Here we show that opposite ends of the benthic/limnetic ecomorphological contin- alleles of the hedgehog pathway receptor Patched1 (Ptch1) gene uum that characterizes East African cichlid radiations (10) as are responsible for adaptive variation in the shape of the lower well as many other notable divergences among teleosts at both jaw both within and among genera of Lake Malawi cichlid fish. the population and species level (14–19). Identifying the mo- The evolutionarily derived allele of Ptch1 reduces the length of the lecular genetic basis for these ecomorphologic shifts therefore retroarticular (RA) process of the lower jaw, a change predicted to has the potential to inform a more comprehensive understanding increase speed of jaw rotation for improved suction-feeding. The of teleost diversity in general. alternate allele is associated with a longer RA and a more robustly A recent genome scan study identified single nucleotide mineralized jaw, typical of species that use a biting mode of feed- polymorphisms (SNPs) exhibiting unusually high differentiation ing. Genera with the most divergent feeding morphologies are (FST) between Labeotropheus and Metriaclima (20). The high FST nearly fixed for different Ptch1 alleles, whereas species with in- of these SNPs suggests they may be linked to genetic loci asso- termediate morphologies still segregate variation at Ptch1. Thus, ciated with evolutionary divergence between the two genera. fi the same alleles that help to de ne macroevolutionary divergence Here we compared the genomic locations of the high FST SNPs among genera also contribute to microevolutionary fine-tuning of fi

and previously identi ed QTL (8, 12) as a starting point to EVOLUTION adaptive traits within some species. Variability of craniofacial mor- identify the speciﬁc genes contributing to morphological di- phology mediated by Ptch1 polymorphism has likely contributed vergence between the genera. We pinpoint alleles of the Ptch1 to niche partitioning and ecological speciation of these ﬁshes. gene as contributing to morphological divergence between the two genera and further show that the same alleles of Ptch1 adaptive radiation | development | genetics | quantitative trait loci | segregate within other species, producing continuing adaptive trophic morphology effects on craniofacial structure.

he astounding diversity of vertebrate craniofacial morphology Results and Discussion Treflects adaptation to a wide variety of resources and envi- In a genome scan comparison of Labeotropheus and Metriaclima ronments. Evolution of craniofacial structure has been key to (20), SNP Aln100281_1741 had an FST of 0.86 between the two niche specialization and speciation in vertebrates (1), most fa- genera, and a comparative BLAST search placed the SNP in the fi fi mously exemplified by the varied beak morphology of Darwin’s rst intron of the Ptch1 homolog. We identi ed a sequence finches (2). The radiation of finches demonstrates both species- scaffold containing Aln100281 and Ptch1 from the draft genome fi defining craniofacial divergence (2) and rapid adaptation of assembly of another cichlid sh, the Nile tilapia (Oreochromis craniofacial morphology within species over the course of a few niloticus). With additional comparative genomics we placed this scaffold on linkage group (LG) 12 of the cichlid genetic map, years in response to changes in resource availability (3). Al- fi though shifts in the expression of key craniofacial genes have within a previously identi ed QTL interval associated with 8.6% of variance in length of the retroarticular process (RA) (8). This been correlated with differences in trophic morphology among fi fi species of birds and cichlid fishes (4–8), the specific genetic loci QTL was one of ve identi ed for RA length (8). Functionally producing these differences remain uncharacterized. As a result, the inlever for the action of lower jaw depression (Fig. 1C), RA it is not clear whether interspecific divergence and intraspecific length modulates the functional tradeoff between force and ve- adaptation share the same molecular genetic basis. locity during jaw rotation (21, 22). No association was found Diversification of trophic morphology in the adaptive radia- between the LG12 QTL and outlever length of the lower jaw (8). Shortening of the RA while outlever length remains uniform is tion of cichlids in Lake Malawi (East Africa) is also correlated predicted to increase the speed of jaw opening, and the LG12 with specialized modes of feeding and resource partitioning and QTL also explains 11.3% of the functional phenotype, the me- has likely contributed to their rapid speciation (9, 10). Because chanical advantage of jaw opening (MA ; Fig. 1C). We refer to craniofacial structure arises from a complex and dynamic de- O the Labeotropheus genotype at the LG12 locus as the “long”-RA velopmental program with both pleiotropic and modular effects allele, and that of Metriaclima as the “short”-RA allele, to reflect (11), the expectation is that it should evolve via continuous fine- tuning steps. In accordance with this prediction, we previously demonstrated that morphological differences between closely Author contributions: R.B.R., R.C.A., and T.D.K. designed research; R.B.R., Y.H., and R.C.A. related cichlid genera are the result of directional selection on performed research; R.B.R., Y.H., R.C.A., and T.D.K. analyzed data; and R.B.R., R.C.A., and numerous genetic loci of small to moderate effect (8, 12). These T.D.K. wrote the paper. studies identified numerous quantitative trait loci (QTL) for The authors declare no conflict of interest. craniofacial morphology segregrating in the F2 hybrid progeny This article is a PNAS Direct Submission. of the genera Labeotropheus and Metriaclima (8, 12). Although Data deposition: The sequences reported in this paper have been deposited in the Gen- Labeotropheus and Metriaclima are closely related rock-dwelling Bank database (accession nos. JN037665–JN037693 and JN116727). cichlid genera, they are divergent in craniofacial morphology and 1To whom correspondence should be addressed: E-mail: [email protected]. microhabitat utilization (10, 13). Labeotropheus is a specialist This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. algal scraper with robust jaws adapted for biting, whereas Met- 1073/pnas.1018456108/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1018456108 PNAS Early Edition | 1of6 A BB

C D 4 10 cM

LOD 2

0 Ptch1 GM284 UNH874 UNH2141 UNH2133 UNH2104 UNH2084 UNH2162 UNH1009 E UNH2012 0.8 )

0.6 Labeotropheus vs. vs. 0.4 Metriaclima

( 0.2 ST F

0 0 40 80 120 160 location on Oreochromis genome assembly F 6

O Rad26l Ptch1 Fancc

4 (log(1/p))

2 Tropheops Tropheops association with MA in

0 04080120160 location on Oreochromis genome assembly

Fig. 1. Ptch1 locus characterization in families, genera, and populations. Craniofacial appearance of (A) Labeotropheus trewavasae (image courtesy of Justin

Marshall) and (B) Metriaclima mbenjii.(C) Lower jaw phenotype measures; RA and inlever length is equivalent. (D) QTL mapping interval for MAO on cichlid LG12. (E) Population differentiation (FST)betweenLabeotropheus and Metriaclima (n = 24 each) for SNPs at the Ptch1 locus. Dashed lines indicate two experimental measures of mean FST; the lower line from global comparison of Labeotropheus vs. Metriaclima across many populations (20), and the upper line from comparisons of randomized Labeotropheus and Metriaclima population pairs from distinct sites (48). (F)Signiﬁcance of association between SNPs at the

Ptch1 locus and MAO in the genus Tropheops (Wald test, n = 59); dashed line indicates a P value of 0.001. Genes in the region annotated with bold arrows. The SNP used to indicate long and short alleles of Ptch1 is circled in E and F. the morphological phenotypes associated with each. The long a marker adjacent to Ptch1 conﬁrmed its location on LG12 and allele acts dominantly to increase RA length and MAO (8). centers the QTL interval squarely over this locus (Fig. 1D). We then used a DNA panel consisting of four Labeotropheus and Population Genetic Mapping of the Causative Interval. We de- four Metriaclima wild-caught populations to characterize pat- veloped additional markers across the Ptch1 genome scaffold by terns of sequence divergence in the region (Table S2). The SNP resequencing a panel of Malawi cichlid DNA to identify poly- Aln100281_1741 continued to show high FST between genera in morphisms (Table S1). Genotyping the F2 mapping family for the populations included in the panel. Several markers located as

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1018456108 Roberts et al. far as 20 kb upstream of Ptch1 also exhibited extremely high FST, where bone development has or will be initiated (Fig. 3 B–D and F– approaching alternate fixation (Fig. 1E). The consistency of ge- H). Notably, species exhibit differences in Ptch1 expression that netic divergence between Labeotropheus and Metriaclima across correlate with biting and suction feeding jaw morphologies. In the species and populations suggests that variation in this region long RA species Labeotropheus fuelleborni (LF) there is marked underlies a fundamental phenotypic difference between the expression of Ptch1 surrounding the lower jaw at this stage, with two genera. nodes of robust expression in or adjacent to areas of skeletal dif- Although genetic differentiation at Ptch1 is coincident with the ferentiation, including the regions where the dentary and RA will QTL for RA length, the Labeotropheus and Metriaclima panel form (Fig. 3 B and D). In contrast, levels of Ptch1 are much lower in does not allow correlation of genotype to a specific phenotype the short RA species Metriaclima zebra (MZ) in the same context within populations because alternate alleles are essentially fixed (Fig. 3 F and H). Unlike expression at the developing lower jaw, within each genus. However, these alleles are segregating within levels of Ptch1 expression are qualitatively similar between LF and populations of a third genus, Tropheops, which exhibits a range MZ in the developing fin-ray elements of the tail (Fig. 3 C and G). of trophic morphology roughly intermediate between Metria- To more accurately describe regional differences in levels of clima and Labeotropheus (23, 24). Thus, we were able to directly expression of Ptch1, as well as its downstream target Gli1, we measure association between individual MAO phenotypes and performed a pixel density analysis on stage-matched LF and MZ fi fi markers across the high FST region, using a panel of four natural larvae as previously described (27, 28). We nd that a signi - populations of Tropheops (Table S2). The association peaks at an cantly greater proportion of pixels were labeled in the jaws of LF SNP roughly 15 kb upstream of Ptch1 and falls below significance compared with MZ for both Ptch1 and Gli1 (Fig. S1 and Table before the Ptch1 start codon (Fig. 1F). The genotype–phenotype S3). No difference was observed for either factor in the caudal association within and among Tropheops species provides strong fin(Fig. S1 and Table S3). Additionally, a quantitative PCR evidence that the Ptch1 locus is modulating RA length and MAO assay for Ptch1 transcript in microdissected tissues from 6.5-dpf and narrows the interval of interest to the region upstream of the larvae demonstrated higher levels of Ptch1 transcript in the lower gene (Figs. 1F and 2). The upstream SNPs serve to label long and jaw of LF relative to MZ, but there was no difference in Ptch1 short alleles at the locus; as in the F2 mapping family, the long transcript levels in the trunk and tail by species (Table S4). To- allele seems to be dominant over the short allele in natural gether these data strongly support the assertion that PTCH1- Tropheops populations (Fig. 2). mediated hedgehog signaling is up-regulated in the jaws of LF fi compared with MZ, speci cally at the time and within the area EVOLUTION Ptch1 Expression Differences Correlate with Developmental Diver- where the RA develops. gence. We did not find any nonsynonymous polymorphisms in Ptch1 coding sequence. The lack of structural differences in the Manipulation of Hedgehog Pathway Signaling Affects Bone Devel- protein is not surprising given the fundamental importance of Ptch1 opment and Recapitulates Natural Interspecific Variation in Bone to a variety of developmental processes (25). Because signatures of Shape. Our expression results are consistent with recent work genetic divergence immediately upstream of Ptch1 suggest adap- demonstrating a role for hedgehog signaling in dermal bone tation via cis-regulatory elements, we examined Ptch1 expression by development (29, 30) and predict that modulation of this path- in situ hybridization. Dimensionality of the lower jaw is determined way can lead to variation in jaw shape. To test this prediction we early, and in comparisons of Labeotropheus and Metriaclima dif- treated LF larvae with cyclopamine, an antagonist of hedgehog ferences in the shape of the lower jaw precursor (Meckel’s carti- pathway signaling (31), at a critical stage of craniofacial bone lage) are evident as early as 5 d postfertilization (dpf), stage 10 (8, development and Ptch1 expression (stage 12). During normal 26). By the following day (stage 12) many of the bones that con- development, both Ptch1 and its downstream target Gli1 are stitute the feeding apparatus have begun to condense, and at this expressed in areas adjacent to or coincident with the bone dif- early stage we observed discrete nodes of Ptch1 expression in areas ferentiation marker Col1a1 (Fig. S1 A–L). In treated larvae, we

a a a b c c oe power 0.4

0.3

mechanical advantage, opening mechanical advantage, speed 0.2

Genotype: CC CC CT TT TT TT Genus: Labeotropheus Tropheops Metriaclima Cynotilapia Feeding: biting suction

Fig. 2. Morphology and function of the lower jaw by genus and Ptch1 genotype. MAO values of wild individuals (Table S2) indicated by box-plot, with letters above designating statistical groups by one-way ANOVA and Tukey’shonestsigniﬁcance test (P < 0.05). Lower jaw images from representative individuals at top; each jaw has an MAO matching the mean of the corresponding genus and genotype. Genotype refers to high-association SNP (circled in Fig. 1 E and F). Box boundaries indicate quartiles, and whiskers indicate range of MAO data.

Roberts et al. PNAS Early Edition | 3of6 fi A B treated larvae exhibit signi cantly reduced RA length (and MAO), whereas the dentary and overall length of the lower jaw, which is determined primarily by outgrowth of Meckel’s cartilage that occurs earlier than the stages examined here, remains roughly the same (Fig. 4 A–C and Table S5). In terms of gene expression patterns (Fig. S1 and Table S3), relative bone development (Fig. S1 and Table S3), and MAO (Fig. 4 A–C and Table S5), cyclopamine-treated LF recapitulate a suction-feeding, MZ-like lower jaw phenotype. Although rates of overall C D development do not differ between LF and MZ during early larval stages, we suggest that key differences in lower jaw shape are due at least in part to PTCH1-mediated differences in the rate of bone development within the RA. The exact cellular mechanism(s) through which hedgehog signaling influences differences in bone development remain to be characterized. We also noted that treated animals exhibited bifurcated expression of Col1a1 within the developing branchiostegal rays (asterisk in Fig. S1Z), as well as aberrant expression within the dermal fin ray elements of the caudal fin (arrow in Fig. S1AA). These E F patterns are also consistent with craniofacial defects after cyclopamine treatment, in which fusion of the branchiostegal rays and truncation of the caudal fin ray elements were observed. These data extend previously documented roles for hedgehog signaling in dermal bone development (29, 30) and suggest that this pathway plays an important role in polarizing dermal bone development along a proximal–distal axis. It is well established that hedgehog signaling is critical for the proper patterning and polarization of several organs in various animal taxa (32–37), but here a similar G H role has been documented for bone development. The extent to which hedgehog-mediated outgrowth of other dermal bones (e.g., RA) has influenced species-specific differences in craniofacial Metriaclima zebra Metriaclima zebra Labeotropheus fuelleborni shape would be a fruitful area of future investigation. Our data suggest the presence of at least two functionally distinct alleles of Ptch1 that produce different levels of transcript in a spatially and temporally restricted manner. The long allele produces relatively higher levels of Ptch1, resulting in greater bone deposition, a longer RA, and higher MAO. The short allele fi has relatively lower expression, resulting in less bone and Fig. 3. Interspeci c differences in Ptch1 expression during cichlid cranio- a shorter RA that increases jaw opening speed. Although the facial development. Craniofacial outcomes in (A) LF and (E) MZ larvae at 13 association between Ptch1 expression and bone deposition is dpf; LF larvae exhibit accelerated bone (pink) development compared with fi fi MZ; cartilage stains blue. Ptch1 in situ labeling in representative 6-dpf (stage interesting, we nd no signi cant association between Ptch1 ge- 12) (B–D) LF and (F–H) MZ larvae: (B and F) Lateral whole-mount view of notype and length or thickness of the jaw, suggesting that dif- lower jaw with nodes of Ptch1 expression, particularly in mesenchymal cells ferences in developmental outcomes of the Ptch1 alleles are at the distal (arrow) and proximal (arrowhead) ends of the lower jaw pre- largely restricted to the RA. Whether the early effects of Ptch1- cursor where the dentary and RA process will form, respectively. (C and G) mediated bone deposition represent an adaptation for the timing Ptch1 labeling is qualitatively similar in developing fin-ray elements of tail and/or rate of bone development among cichlids remains an (arrows). (D and H) Flat-mount preparation of the jaw joint in lateral view, open question. Certainly the results of our expression and showing the cartilaginous precursor of the RA process (outlined) relative to cyclopamine experiments support a role for hedgehog signaling node of Ptch1 expression. dnt, dentary; mx, maxilla; pap, posterior articu- in modulating bone development. lation process; pmx, premaxilla. (Scale bars, 200 μm in A, B, E, and F;100μm in C and G;and10μm in D and H.) Ptch1 Alleles Contribute to Microevolutionary and Macroevolutionary Processes. Using the SNP in highest association with MAO in Tro- find that expression of both Ptch1 and Gli1 is dramatically and pheops, we genotyped the Ptch1 locus in additional wild-caught fi globally reduced (Fig. S1 P–W and Table S3). We also find that individuals of four genera (Fig. 2). The signi cant correlation to bone development is delayed in areas where Ptch1 is normally MAO phenotype remains consistent across all taxa measured, expressed (Fig. S1 X–AA and Table S3). Specifically, less Col1a1 suggesting extremely tight linkage of the SNP with the causative expression was observed around the developing RA, whereas polymorphism. Obligate biters (Labeotropheus) and suction feed- ers (Cynotilapia and Metriaclima) are nearly fixed for long and more cells were expressing Col1a1 around the branchiostegal short alleles, respectively. The genus Tropheops is particularly in- rays and within the caudal fin(Fig. S1 Y–AA and Table S3), teresting, because segregating Ptch1 alleles contribute to “shallow” suggesting that these structures were in a more undifferentiated and “deep” ecomorphs that correlate with resource use and depth state after cyclopamine treatment relative to control animals. distribution (24, 38). In deep ecomorph Tropheops species, the Expression of Col1a1 around the dentary was relatively un- short Ptch1 allele seems to be fixed, contributing to a morphology affected by cyclopamine treatment (Fig. S1X); however, this is adapted to sifting and suction feeding of loose material. In the likely because development of this structure was well underway Tropheops populations examined from shallow water, the two at the stage when animals were treated. Overall, these patterns alleles segregate and are associated with different morphologies of expression are consistent with the phenotypic outcome of within species. Individuals carrying a long allele at Ptch1 have a jaw cyclopamine treatment. Specifically, we find that cyclopamine- morphology similar to that of Labeotropheus, well adapted for

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1018456108 Roberts et al. A C p = 0.0003

0.3 p = 0.001

EtOH control 0.2 B

0.1 LF LF LF MZ EtOH cyc Mechanical advantage, opening Mechanical advantage,

cyclopamine 0

Fig. 4. Treatment of biting species larvae with a hedgehog pathway inhibitor recapitulates a suction-feeding jaw phenotype. LF larvae were treated with (A) 0.5% ethanol (EtOH control, n =7)or(B)50μM cyclopamine (cyc, n = 10) for 6 h at stage 12; RA length measured at 12 dpf (black arrow in A and B). (C) Relative to the length of the lower jaw (measured as the distance between the center of the jaw joint and the tip of the dentary), RA length (measured as the distance from the posterior tip of the posterior articulation process to the ventral tip of the retroarticular) was significantly reduced in cyclopamine treatment group compared with the control LF group but was not significantly different from MZ control larvae (n = 7). Larvae treated with ethanol were in- distinguishable from untreated siblings (n = 6). P values, one-way ANOVA. (Scale bar, 100 μm.) feeding on attached algae found at shallow depth in the lake. to produce adaptive changes in craniofacial phenotype. Notably, Individuals homozygous for the short allele of Ptch1 have a trophic we demonstrate that the same Ptch1 alleles underlying macro- morphology that more closely resembles that of the deep eco- evolutionary divergence between genera also form the basis of EVOLUTION morphs, better suited for sifting and suction feeding on loose ma- microevolutionary adaptation among species and even within terial. Thus, allelic variation at the Ptch1 locus is associated with the populations. Where Ptch1 alleles segregate, a range of ecologically continuum in lower jaw morphology both within and among Tro- relevant morphologies are produced within populations. Impor- pheops species. Although these data are consistent with recent work tantly, the resources used by segregating ecomorphs are distrib- showing that early Sonic hedgehog signaling from the embryonic uted by depth, providing an opportunity for spatial segregation of forebrain can lead to continuous craniofacial variation in mouse individuals by genotype within populations, possibly contributing embryos (39) and jaw width differences in the Mexican tetra (40), to character displacement and even the development of incipient our results posit a unique role for hedgehog signaling in regulating species (43). Thus, diversification via adaptive polymorphism fine-scale, continuous differences in feeding morphology via bone could support iterative niche partitioning in complex communi- deposition. ties. Additionally, such adaptive polymorphism segregating within Overtly, the genus Labeotropheus has a highly specialized populations would permit rapid morphological evolution to take trophic apparatus, whereas the genus Metriaclima has a general- place in response to sudden environmental change. Segregating ist morphology, more similar to that of the ancestral lineages that fine-tuning alleles of the kind we describe here have likely been invaded Lake Malawi (10). It is therefore surprising that, a key component in both adaptation and speciation in the ongoing according to a Ptch1 gene phylogeny including other African explosive radiation of Lake Malawi cichlids. cichlid genera from Lake Malawi and beyond (Fig. S2), the short Ptch1 allele fixed in Metriaclima is more recently derived, Materials and Methods whereas Labeotropheus carries the ancestral allele. The SNP Animals. Fish for breeding and association studies were collected directly from haplotype that labels the short allele is found in only a subset of Lake Malawi. Fish were maintained and euthanized according to Institutional the rock-dwelling cichlids in Malawi. The short allele also shows Animal Care and Use Committee guidelines. Larval fish were obtained by signatures of recent selection, including relatively higher ex- natural matings within facilities at Syracuse University and the University of tended haplotype homozygosity, reduced nucleotide diversity, Maryland, with developmental staging based on de Jong et al. (26). A hybrid and significantly low values for Tajima’s D (Fig. S3). These data cross between LF and MZ generated an F2 mapping population (n = 173), fi which was phenotyped for jaw morphology and used as the basis of QTL suggest that selection has acted recently to ne-tune the RA to analysis as previously described (8, 12). improve suction feeding within the rock-dwelling cichlids. East African rift-valley cichlid radiations are characterized by Genetic Analysis. Draft sequence of the O. niloticus genome was assembled rapid and repeated bentho-pelagic divergence following the from Illumina sequence reads totaling 60× sequence coverage. The assembly colonization of a lacustrine environment by riverine ancestors has an N50 scaffold size of 138 kb and was aligned to stickleback genome (10). In Lake Malawi this divergence has led to the establishment sequence using BLAST (available at http://BouillaBase.org). Shotgun genome of the genetically distinct rock- and sand-dwelling clades (41, 42). sequence of Malawi cichlid species (20) was aligned to the Oreochromis On average, rock-dwelling species possess trophic morphologies genome assembly using BLASTN. Other polymorphisms were identified consistent with a benthic lifestyle, whereas most sand-dwellers (Table S1) and genotyped by resequencing aligned sequences using Applied are pelagic in terms of their feeding ecology (10). We propose Biosystems Big Dye Terminator chemistry and an ABI3730 sequencer. The “ ” that, after an initial colonization of the rock niche by cichlids marker Ptch1Loc10_SNP1 was added to the genetic map as Ptch1 by genotyping the original F2 mapping family (12); the resulting map was used with specialized biting morphologies, selection for the short al- for QTL analysis. Population genetic and association analyses were per- lele of Ptch1 contributed to the reevolution of an ancestral-like formed in DnaSP and PLINK, respectively (44, 45), using polymorphisms with feeding strategy and concomitant expansion of the rock-dwelling aminorallelefrequency>0.05. niche into the water column. Despite the major and conserved roles the hedgehog pathway In Situ Hybridization. Whole-mount in situ hybridization (WISH) was per- plays throughout development, we show that it continues to evolve formed as previously described (8). Embryos used for the reported experi-

Roberts et al. PNAS Early Edition | 5of6 ments were raised, staged, fixed, and processed in parallel. Sense and an- 50 μM (46). Cyclopamine (10 mM, in ethanol) or ethanol alone (control) was tisense Ptch1 riboprobes were made from cichlid clones (sequence identical added to larval fish water. Animals were treated for 6 h in the dark because for LF, accession no. JN037690, and MZ, accession no. JN037691), corre- cyclopamine is light sensitive. After treatment, larvae were washed several sponding to exons 1–7and7–17. These yielded identical WISH results; data times in fish water before returning to circulating culture flasks to be grown derived from the exons 1–7 riboprobe are reported. Accession numbers for out to 12 dpf. Cartilage and bone staining was performed as previously de- Gli1 and Col1a1 probes are JN037689 and JN116727, respectively. To facili- scribed (47). Larvae were photographed with a Zeiss Axiocam digital imaging tate probe penetration, alkaline hydrolysis was used to fragment probes to system mounted to an M2 Bio stereomicroscope (Zeiss). Image processing and pixel density analysis was performed in Adobe Photoshop CS4 (Fig. S4). ≈500 bp. A solution of 20 mL RNA probe, 12 mL H20, 4 mL 0.4 M sodium bicarbonate, and 4 mL 0.6 M sodium carbonate was incubated at 60 °C for a period based on the following: time (min) = (starting kb − desired kb)/ ACKNOWLEDGMENTS. We thank Matthew Conte for bioinformatics support; Claire O’Quin and Jennifer Ser for fish husbandry support; Rolf Karl- (0.11 × starting kb × desired kb). strom for advice on cyclopamine treatment in fish; Justin Marshall for the Labeotropheus photograph used in Fig. 1; and J. Todd Streelman, William Pharmacological Treatment. On the basis of previously published work in Jeffery, Karen Carleton, and Kelly O’Quin for their comments on previous cichlids, we used cyclopamine (LC Laboratories) at a working concentration of drafts of the manuscript.

1. Gans C, Northcutt RG (1983) Neural crest and the origin of vertebrates: A new head. 26. de Jong IM, Witte F, Richardson MK (2009) Developmental stages until hatching of Science 220:268–273. the Lake Victoria cichlid Haplochromis piceatus (Teleostei: Cichlidae). JMorphol270: 2. Darwin CR (1845) Journal of Researches into the Geology and Natural History of the 519–535. Various Countries Visited During the Voyage of H.M.S. Beagle, Under the Command 27. Albertson RC, Yelick PC (2007) Fgf8 haploinsufficiency results in distinct craniofacial of Captain FitzRoy, R.N (John Murray, London), 2nd Ed. defects in adult zebrafish. Dev Biol 306:505–515. 3. Grant PR, Grant BR (2002) Unpredictable evolution in a 30-year study of Darwin’s 28. Cooper WJ, Albertson RC (2008) Quantification and variation in experimental studies finches. Science 296:707–711. of morphogenesis. Dev Biol 321:295–302. 4. Abzhanov A, Protas M, Grant BR, Grant PR, Tabin CJ (2004) Bmp4 and morphological 29. Abzhanov A, Rodda SJ, McMahon AP, Tabin CJ (2007) Regulation of skeletogenic variation of beaks in Darwin’s finches. Science 305:1462–1465. differentiation in cranial dermal bone. Development 134:3133–3144. 5. Abzhanov A, et al. (2006) The calmodulin pathway and evolution of elongated beak 30. Hammond CL, Schulte-Merker S (2009) Two populations of endochondral osteoblasts morphology in Darwin’s finches. Nature 442:563–567. with differential sensitivity to Hedgehog signalling. Development 136:3991–4000. 6. Wu P, Jiang TX, Suksaweang S, Widelitz RB, Chuong CM (2004) Molecular shaping of 31. Chen JK, Taipale J, Cooper MK, Beachy PA (2002) Inhibition of Hedgehog signaling by the beak. Science 305:1465–1466. direct binding of cyclopamine to Smoothened. Genes Dev 16:2743–2748. 7. Wu P, Jiang TX, Shen JY, Widelitz RB, Chuong CM (2006) Morphoregulation of avian 32. Chuong CM, Patel N, Lin J, Jung HS, Widelitz RB (2000) Sonic hedgehog signaling beaks: Comparative mapping of growth zone activities and morphological evolution. pathway in vertebrate epithelial appendage morphogenesis: Perspectives in de- Dev Dyn 235:1400–1412. velopment and evolution. Cell Mol Life Sci 57:1672–1681. 8. Albertson RC, Streelman JT, Kocher TD, Yelick PC (2005) Integration and evolution of 33. Koyama E, et al. (1996) Polarizing activity, Sonic hedgehog, and tooth development in the cichlid mandible: the molecular basis of alternate feeding strategies. Proc Natl embryonic and postnatal mouse. Dev Dyn 206:59–72. Acad Sci USA 102:16287–16292. 34. Ingham PW, Fietz MJ (1995) Quantitative effects of hedgehog and decapentaplegic 9. Kocher TD (2004) Adaptive evolution and explosive speciation: The cichlid fish model. activity on the patterning of the Drosophila wing. Curr Biol 5:432–440. Nat Rev Genet 5:288–298. 35. Krauss S, Concordet JP, Ingham PW (1993) A functionally conserved homolog of the 10. Cooper WJ, et al. (2010) Bentho-pelagic divergence of cichlid feeding architecture Drosophila segment polarity gene hh is expressed in tissues with polarizing activity in was prodigious and consistent during multiple adaptive radiations within African rift- zebrafish embryos. Cell 75:1431–1444. lakes. PLoS ONE 5:e9551. 36. Tanaka M, et al. (2000) Distribution of polarizing activity and potential for limb 11. Helms JA, Schneider RA (2003) Cranial skeletal biology. Nature 423:326–331. formation in mouse and chick embryos and possible relationships to polydactyly. 12. Albertson RC, Streelman JT, Kocher TD (2003) Directional selection has shaped the Development 127:4011–4021. oral jaws of Lake Malawi cichlid fishes. Proc Natl Acad Sci USA 100:5252–5257. 37. Harris MP, Williamson S, Fallon JF, Meinhardt H, Prum RO (2005) Molecular evidence 13. Albertson RC, Kocher TD (2001) Assessing morphological differences in an adaptive for an activator-inhibitor mechanism in development of embryonic feather branch- trait: A landmark-based morphometric approach. JExpZool289:385–403. ing. Proc Natl Acad Sci USA 102:11734–11739. 14. Skulason S, Noakes DLG, Snorrason SS (1989) Ontogeny of trophic morphology in four 38. Ribbink AJ, Marsh BA, March AC, Ribbink AC, Sharp BJ (1983) A preliminary survey of sympatric morphs of arctic charr Salvelinus alpinus in Thingvallavatn, Iceland. Biol J the cichlid fishes of rocky habitats in Lake Malawi. SAfrJZool18:149–310. Linn Soc Lond 38:281–301. 39. Young NM, Chong HJ, Hu D, Hallgrímsson B, Marcucio RS (2010) Quantitative analyses 15. Walker JA (1997) Ecological morphology of lacustrine threespine stickleback Gaster- link modulation of sonic hedgehog signaling to continuous variation in facial growth osteus aculeatus L. (Gasterosteidae) body shape. Biol J Linn Soc Lond 61:3–50. and shape. Development 137:3405–3409. 16. Adams CE, Huntingford FA (2002) The functional significance of inherited differences 40. Yamamoto Y, Byerly MS, Jackman WR, Jeffery WR (2009) Pleiotropic functions of in feeding morphology in a sympatric polymorphic population of Arctic charr. Evol embryonic sonic hedgehog expression link jaw and taste bud amplification with eye Ecol 16:15–25. loss during cavefish evolution. Dev Biol 330:200–211. 17. Parsons KJ, Robinson BW (2006) Replicated evolution of integrated plastic responses 41. Albertson RC, Markert JA, Danley PD, Kocher TD (1999) Phylogeny of a rapidly during early adaptive divergence. Evolution 60:801–813. evolving clade: the cichlid fishes of Lake Malawi, East Africa. Proc Natl Acad Sci USA 18. Riopel C, Robinson BW, Parsons KJ (2008) Analyzing nested variation in the body form 96:5107–5110. of Lepomid sunfishes. Environ Biol Fishes 82:409–420. 42. Danley PD, Kocher TD (2001) Speciation in rapidly diverging systems: Lessons from 19. Cooper WJ, Westneat MW (2009) Form and function of damselfish skulls: Rapid and Lake Malawi. Mol Ecol 10:1075–1086. repeated evolution into a limited number of trophic niches. BMC Evol Biol 9:24. 43. Mayr E (1963) Animal Species and Evolution (Harvard Univ Press, Cambridge, MA). 20. Loh YH, et al. (2008) Comparative analysis reveals signatures of differentiation amid 44. Librado P, Rozas J (2009) DnaSP v5: A software for comprehensive analysis of DNA genomic polymorphism in Lake Malawi cichlids. Genome Biol 9:R113. polymorphism data. Bioinformatics 25:1451–1452. 21. Liem KF (1978) Modulatory multiplicity in the functional repertoire of the feeding 45. Purcell S, et al. (2007) PLINK: A tool set for whole-genome association and pop- mechanism in cichlid fishes. JMorphol158:323–360. ulation-based linkage analyses. Am J Hum Genet 81:559–575. 22. Westneat MW (1990) Feeding mechanics of teleost fishes (Labridae; Perciformes): A 46. Fraser GJ, Bloomquist RF, Streelman JT (2008) A periodic pattern generator for dental test of four-bar linkage models. JMorphol205:269–295. diversity. BMC Biol 6:32. 23. Albertson RC, Kocher TD (2005) Genetic architecture sets limits on transgressive 47. Walker MB, Kimmel CB (2007) A two-color acid-free cartilage and bone stain for segregation in hybrid cichlid fishes. Evolution 59:686–690. zebrafish larvae. Biotech Histochem 82:23–28. 24. Albertson RC (2008) Morphological divergence predicts habitat partitioning in a Lake 48. Mims MC, Darrin Hulsey C, Fitzpatrick BM, Streelman JT (2010) Geography disen- Malawi cichlid species complex. Copeia 2008:689–698. tangles introgression from ancestral polymorphism in Lake Malawi cichlids. Mol Ecol 25. Carroll SB (2005) Evolution at two levels: On genes and form. PLoS Biol 3:e245. 19:940–951.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1018456108 Roberts et al. Supporting Information

Roberts et al. 10.1073/pnas.1018456108 SI Materials and Methods Quantitative PCR Analysis. Larval fish were staged as described in the Pixel Density Analysis (PDA). To determine and compare the relative main text, and tissue microdissected at 6.5 days postfertilization amounts of Ptch1 transcript within the region of the pharyngeal (dpf) and preserved in RNAlater (Ambion). Lower jaw cartilage skeleton that contains the retroarticular process, PDA was per- back to and including the first four ventral pharyngeal arch tissues formed as described previously (1, 2). In brief, after whole-mount in (i.e., the mandibular, hyoid, and branchial arches 1 and 2) were situ hybridization (WISH), specimens were mounted in 80% glycerol taken, as were the tail and trunk posterior of the yolk sac. RNA was and photographed with a Zeiss Axiocam digital imaging system isolated using the RNeasy kit (Qiagen) and quantified by absor- mounted to an M2 Bio stereomicroscope (Zeiss). Homologous areas bance. Normalized RNA was used as template for reverse tran- within the jaw were selected using the lasso tool in Adobe Photoshop, scription by SuperScript III (Invitrogen). Primer-probe sets were and the total number of pixels within that region was determined designed across exon junctions using Primer Express (Applied using the histogram tool (Fig. S4). Next, the relative area in which Biosystems) software (Ptch1: forward, 5′-TTCTGATGCTGG- Ptch1 was expressed was measured by selecting for pixels that were CCTATGCA-3′; reverse, 5′-CCCCTGAGACTTGGCACAGT- the color of the in situ reaction (i.e., purple–dark purple) using the 3 , probe, 5 -6FAM-CCTGACCATGCTGCGAT-TAMRA-3 ; “select color range” option within Photoshop. For Ptch1 and Gli1 ′ ′ ′ expression this analysis was performed on the area ventral to the eye Rad26l: forward, 5′-CGAGTCCAGATCGTCAGAGACTT-3′; (as shown in Fig. S4). Because Col1a1 expression is more restrictive, reverse, 5′-CACCTGCCATGGTGGAAAC-3′, probe, 5′-6FAM- PDA was performed on the region immediately adjacent to the RA AACAGCTCCTCTCACATC-TAMRA-3′). Real-time PCR re- (i.e., Fig. 3D,maintext).Finally,asapointofcomparison,similar actions used Taqman Universal PCR Master Mix (Applied Bio- analyses were performed using all three probes on the caudal fin. systems) and were run on a Roche Lightcycler 480. Relative (GenBank accession nos.: Gli1,JN037689;Ptch1,JN0376990and expression was calculated using the 2-ΔΔCT method (3), normaliz- JN037691; Rad26l,JN037692andJN037693;Col1a1,JN116727). ing to β-actin expression as an endogenous control.

1. Albertson RC, Yelick PC (2007) Fgf8 haploinsufficiency results in distinct craniofacial 3. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real- defects in adult zebrafish. Dev Biol 306:505–515. time quantitative PCR and the 2(-ΔΔC(T)) Method. Methods 25:402–408. 2. Cooper WJ, Albertson RC (2008) Quantification and variation in experimental studies of morphogenesis. Dev Biol 321:295–302.

Roberts et al. www.pnas.org/cgi/content/short/1018456108 1of8 Fig. S1. Hedgehog pathway is necessary for proper craniofacial bone development. Expression of the hedgehog receptor Ptch1 (A–D), its downstream target Gli1 (E–H), and the bone differentiation marker Col1a1 (I–L) is shown in 6-dpf (stage 12) Labeotropheus fuelleborni (LF) larvae. Colocalized expression was observed around the developing dentary (A, E, I), retroarticular (B, F, J), branchiostegal rays (C, G, K; note the discrete node of Ptch1 expression at the distal end of these bones in C,arrow;althoughGli1 expression was not observed around these structures, G), and within the dermal fin ray elements of the caudal fin (arrowheads D, H, and L). Mineralized structures are shown as a reference in older fish (12 dpf, M–O). Treatment with a hedgehog pathway inhibitor resulted in the down-regulation of hedgehog signaling and aberrant craniofacial bone development. LF larvae were treated with 50 μMofcyclopamineat5.5dpf(stage 11) for 6 h. Expression of Ptch1 (P-S)anditsdownstreamtargetGli1 (T–W) were drastically reduced. Although expression of the osteoblast differentiation marker Col1a1 was relatively unaffected in the dentary (X), its expression in other structures suggests an attenuation and/or delay in bone development. Specifically, Col1a1 expression was reduced around the retroarticular process (Y), whereas expanded expression was observed around the branchiostegal rays (Z) and within the caudal fin(AA), suggesting that these structures are in a more undifferentiated state relative to control animals. We also noticed bifurcated expression of Col1a1 in the branchiostegal rays of cyclopamine-treated animals (asterisk in Z), as well as disorganized expression within the developing caudal fin ray elements (arrow in AA). The phenotypic outcome of this treatment is consistent with altered patterns of gene expression. Although the dentary is relatively unaffected in cyclopamine-treated animals, the length of the retroarticular process is reduced in treated animals (AB). In addition, the branchiostegal rays are bifurcated and fused in treated animal (AC), and dermal fin ray elements are dramatically reduced in the caudal fin(AD). bsr, branchiostegal rays; cfr, caudal fin rays; dnt, dentary; M, Meckel’s cartilage; pap, posterior articulation process; ra, retroarticular. (Scale bars, 10 μm in A–C, E–G, I–K, P–R, T–V, and X–Z; and 100 μm in D, H, L–O, S, W, and AA–AD.)

Roberts et al. www.pnas.org/cgi/content/short/1018456108 2of8 88 Serranochromis robustus Riverine piscivore Various Lake Tanganyika cichlids Rhamphochromis esox Malawi pelagic piscivore Ptch1 44 Aulonocara hansbaenschi (sand) insectivore

88 Various Malawi (sand) specialists “long-RA” allele Labeotropheus fuelleborni Mankjila (rock) biter 48 Labeotropheus trewavasae Nakantenga (rock) biter 81 Labeotropheus trewavasae Thumbi West (rock) biter Labeotropheus fuelleborni Nankoma (rock) biter Asatotilapia calliptera Riverine/Malawi generalist Pseudotropheus williamsi (rock) generalist 88 Tropheops sp. Chinyamwezi (rock) generalist, shallow ecomorph 97 Tropheops sp. Chinyankwazi (rock) generalist, shallow ecomorph Petrotilapia tridentiger (rock) aufwuchs comber, generalist 89 90 Cynotilapia afra Higga Reef (rock) suction-feeder Metriclima benetos (rock) suction-feeding generalist

Metriclima zebra (rock) suction-feeding generalist Ptch1 Metriclima pyrsonotus (rock) suction-feeding generalist

Cynotilapia afra Lupingu (rock) suction-feeder “short-RA” allele Cynotilapia afra Chinyamwezi (rock) suction-feeder Metriclima mbenjii (rock) suction-feeding generalist Tropheops sp. Chinyamwezi (rock) generalist, deep ecomorph Cynotilapia afra Mbenjii (rock) suction-feeder Tropheops intermedius (rock) generalist Melanochromis auratus (rock) generalist 88 Tropheops sp. Chinyankwazi (rock) generalist, deep ecomorph Tropheops gracilior (rock) generalist

Fig. S2. Ptch1 gene tree. Neighbor-joining tree of representative Ptch1 locus genotypes by species and site, based on 32 polymorphic sites. Haplotypes with the “short” Ptch1 allele SNP genotype at the high MAO association SNP (Fig. 1F, main text) form a single derived clade found in Malawi rock-dwelling cichlids. The gene tree can be contrasted to previously published species phylogenies of the three genera analyzed in the main text (1) and East African cichlids in general (2). Gene tree created with MAFFT (3).

1. Albertson RC, Markert JA, Danley PD, Kocher TD (1999) Phylogeny of a rapidly evolving clade: The cichlid ﬁshes of Lake Malawi, East Africa. Proc Natl Acad Sci USA 96:5107–5110. 2. O’Quin KE, Hofmann CM, Hofmann HA, Carleton KL (2010) Parallel evolution of opsin gene expression in African cichlid ﬁshes. Mol Biol Evol 27:2839–2854. 3. Katoh K, Misawa K, Kuma K, Miyata T (2002) MAFFT: A novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acids Res 30:3059–3066.

Roberts et al. www.pnas.org/cgi/content/short/1018456108 3of8 Fig. S3. Signatures of selection upstream of the Ptch1 gene. (A)AssociationbetweenSNPgenotypeandthequantitativephenotypeMAO in natural Tro- pheops populations, as reference for B–D; all share the same x axis for comparison. (B) Extended haplotype homozygosity (EHH) from the high MAO association SNP. Ptch1 short alleles (dashed line) show greater relative EHH than long alleles (solid line). Genotypes were phased in FastPhase (1) and the EHH plot made in Sweep (2). (C) Nucleotide diversity (π)and(D)Tajima’s D values for comparison of Labeotropheus (closed circles, solid lines) and Metriaclima (open circles, dashed lines), calculated in DnaSP (3). Lines indicate sliding average of three adjacent values. Metriaclima show consistently reduced π in the region surrounding the high MAO association SNP, as well as consistently low Tajima’s D, indicative of recent positive selection. Dotted horizontal lines in D indicate signiﬁcance thresholds determined by coalescent simulations in DnaSP (3).

1. Scheet P, Stephens M (2006) A fast and ﬂexible statistical model for large-scale population genotype data: Applications to inferring missing genotypes and haplotypic phase. Am J Hum Genet 78:629–644. 2. Sabeti PC, et al. (2002) Detecting recent positive selection in the human genome from haplotype structure. Nature 419:832–837. 3. Librado P, Rozas J (2009) DnaSP v5: A software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25:1451–1452.

Roberts et al. www.pnas.org/cgi/content/short/1018456108 4of8 Fig. S4. Pixel density analysis of Ptch1 WISH results. After WISH, specimens were mounted in 80% glycerol and photographed (A). (B)Homologousareas within the jaw were selected using the lasso tool in Adobe Photoshop, and (C) the “select color range” option within Photoshop was used to select pixels that were the color of the in situ reaction. (Scale bar, 200 μm.)

Roberts et al. www.pnas.org/cgi/content/short/1018456108 5of8 oet tal. et Roberts

Table S1. Ptch1 locus markers † ‡ ﬂ § ¶ www.pnas.org/cgi/content/short/1018456108 Amplimer* Primer 1 Primer 2 S Loc Marker Sequence anking polymorphism SML ST

Ptch1Loc1 CCACTTAATTGCCTCGTCTCA TTCTCGGCTACAAAGCCACT 5 179.1 Loc1_SNP1 AGAAAAAAAGAGAGAGAGGGGGAAC(A/G)GGTCAGTGAGCATTTTCTGATGGGC + Loc1_SNP2 GAGCATTTTCTGATGGGCTCAGGAT(C/G)GTCTCTACTCGATGTGGAGGCATTT + Ptch1Loc2 ACTGGGTACAGGTGGCTGTC TTGGAGGAGATGGTGGAGTC 7 164.4 Loc2_SNP1 TGAGATGGTGATCCCTT(C/T)CTGTGGCTGTTGATATTCTACCACT + Loc2_SNP2 ACAGAGCTTCACAGCTGGTGAGTTT(G/T)TTTGTTTTTTGCATAAATCATTTTA + Loc2_SNP3 TTATTTGCAGGTCTGTTGTTGTTGA(A/C)AGCAAAAGAAACAATACAGAGGGGC + Loc2_SNP4 TTTACCTCTCTGTCTCTTGCTTGTC(C/T)GCCTCCCATCATAACCCTTGGAGAC + Ptch1Loc3 CAATCCCGGAGCTTTTTACA TGCTAGAGCCCCATGGTAGT 6 106.5 Loc3_SNP1 AGCTCAGTCATGCACAAATGCACAC(A/G)TATTGCAGTGGTACAGTAAAGTCTA + Loc3_SNP2 GTATGTTTGGCTGTACCTTACAGTA(A/G)CATTTCAAATTAATGTGGAGCTGGA + Ptch1Loc4 GGATCATCAGCTGAGGGGTA CTGCGATGTCTGTGGGTCT 6 84.1 Loc4_SNP1 AGTCTGTGTTTTTCTGACACAAAGG(A/C)ACAAATATCTGCAAAATATAGACAC ++ Ptch1Loc5 TCGTGTGTGGGGAGACAGTA TGCCTGTTTATGTTTCATTTGG 5 71.4 Loc5_SNP1 TACAATCTGCCCTAAATAGAAACCA(C/T)GGTCACTTTGATGTTTTCTAAGTGG + Loc5_SNP2 TCAGAAAGCAGGTGACGACCGACAC(A/G)ACACGAGATCGCAAACGTGAATGTT + Loc5_SNP3 GCTGAGAGATCTGATGGGAGCTGGG(A/C)TGTATTGTATGAGCGCACGCATCCA + Ptch1Loc6 GTAACCTGCACGCACACATC AAAATGTGGCCAAACTGTCC 5 67.4 Loc6_SNP1║ CTGCGAGAGTAAAAGCTGCATGTTT(A/T)TTGCCTGGTGCGGGACTGCTCTAGG ++ Ptch1Loc7 CGTGATTGGCACAGTTTACG TTGCTGCACAATAAGCAAGC 5 65.7 Loc7_SNP1 CCAAGAGTGGCTGCTGGAGTTAATT(G/T)TTTTGCTTATTATAACCTTATTATT ++ Loc7_SNP2 TCAGAATAAAGTCAATTATAAAAAT(G/T)ATGCCTGCAGCACGGATAGATAAAT + Ptch1Loc8 GCACACTAGACACTTAACCA GCATCAGCTGTGCTGGAAT 5 55.3 Loc8_INDEL1 TATATGTCCCGGATTTAACCGTGTT(-/TGGGCCGCTT)AAGCTATGACTTCATGGTATTGTG + AAACA Loc8_SNP2 TGGGCCGCTTAAGCTATGACTTCAT(A/G)GTATTGTGAAGAAAATAAGCAGACG ++ Loc8_SNP3 AGAAATGTGACTTGAAAAAGAGCAG(C/T)GAATGAGAAAATTCCCTTGTTGTGT + Loc8_SNP4 TTTTTTCTTAATGTGTTCTGATGGT(C/T)AAAACAATCCAGAATCAAATGTGGC + Ptch1Loc9 GGCCTTACTGTCGTCGTTTC ATAGGGTTAGGCTCGGGTTG 5 52.8 Loc9_SNP1 TGTCTTCCAAGCTACTAAAACTCCG(C/T)ACATTATTACAAAATATGGTGCCCT + Loc9_SNP2 GGGGGATTCGGAGTGTGTGTGTTTG(A/G)AATGTTTGTGAAGCCAGTCTCTGAT + Ptch1Loc10 CAATGGTTTTAAACACAAC TCATCCGATCATCCACTTGA 6 51.8 Loc10_SNP1** TTCCAGAAATACTAGGAAAAACACA(C/T)GCAGTAGTTACTAATGCCTTGAATG ++ TCTGG Loc10_SNP2 ATGAGTAATGGCAAATTTAAAAAAA(A/C)CCTCACTGTAGTATGTTTGTTTTAC + Ptch1Loc11 AGCATCTGGTTCAGGGTCAT CAGGGCTGAAGATGTGACAA 1 48.3 Loc11_SNP1 ACACTGCTCACAAAGACTCTAGTCT(A/G)GTGTCTTTGTCACATCTTCCAGCCC + Ptch1Loc12 CCCCAGGAAATAGATAGAC CACAGTGCACACTGCTCACA 3 48.4 Loc12_SNP1 GGATTTAACAAGTGCATGTAACAGC(C/T)TCACAGCCATGCCTTTTATGGCCTT ++ TCC Loc12_SNP2 CTTGTAAATTATCATAATCTAAAAC(A/C)AGTTTGCTCACTACGTCAGGTGGCG + Loc12_SNP3 AGGGCTGAAGATGTGACAAAGACAC(C/T)AGACTAGAGTCTTTGTGAGCAGTGT + Ptch1Loc13 TTTGTCTGTTTTGCACATCCTC CCGTTCTTCCTGGGTGAAT 1 45.3 Loc13_SNP1 AGACGGTATGAAGGGAGGAGAGGAA(A/G)GAGAAAAAGAAAGAGAGTCATGGAG + Ptch1Loc14 CTCCTTTCATCCATCCCTCA CCCTTCTCTGTCACCACCAT 2 44.9 Loc14_SNP1 TCTCTTTTACCTGGATATTCCAGAA(C/G)TGACTACTATGCAAGCTATAGCGTT + Ptch1Loc15 ATCAGTCGAGCTGCTGTACG TGCAGCCTCATAGTGGAATG 4 22.4 Loc15_SNP1 TCCTAAAAGTCCTATAATTACGCTA(A/T)ATCATAATGAACAGCATTTATTTAC ++ Loc15_SNP2 TAAAATTAATTTCACTGATCAACTC(G/T)GTTTCAAATCAGCACAAACTGCCCA ++ Ptch1Loc16 CGATTTTGCTTGGCGTAGTC TCCCATTTCCCATGTACACC 4 6.6 Loc16_SNP1 AGATCTTAAGTGGAATAGTCTGCAG(A/C)AACCAGCCCTGTACACACATGCACA + Loc16_SNP2 TGCTTCCATTTGTTTTGTCTCCAGC(A/G)CTGACGTACTGCAGTGAATATTTTA + Loc16_SNP3 AAGTTAGTTTAAAGAAAATAATTGC(C/T)TTCGTTTCATAAAACTTGAGATGAT +

*GenBank accession nos.: JN037665-JN037688. †Total number of polymorphic sites identiﬁed in amplimer; analyzed for Fig. S3 B–D. ‡Location (kb) on genomic scaffold from Tilapia (Oreochromis niloticus)draftgenomeassembly. §Polymorphic sites with a minor allele frequency >0.05 in Metriaclima and Labeotropheus populations; analyzed for Fig. 1E, main text. ¶Polymorphic sites with a minor allele frequency >0.05 in Tropheops populations; analyzed for Fig. 1F, main text. ║Ptch1Loc6_SNP1 is also the SNP Aln100281_1741 discussed in main text. **Ptch1Loc10_SNP1 was used as the “Ptch1” marker for quantitative trait loci analysis (Fig. 1D, main text) and is the high association SNP circled in Fig. 1 E and F (main text) and genotyped for Fig. 2 (main text). 6of8 Table S2. Species and populations used in study Species Location Niche n Trophic mode Comparison

Astatotilapia calliptera Riverine (Malawi) — 1 — Fig. S1 Aulonocara hansbaenschi Otter Point Rock/Sand 1 Insectivore Fig. S1 Benthochromis tricoti Lake Tanganyika — 1 — Fig. S1 Cynotilapia afra Thumbi West Rock 2 Suction feeding Fig. 2 Cynotilapia afra Chiofu Bay Rock 1 Suction feeding Fig. 2 Cynotilapia afra Chinyamwezi Rock 22 Suction feeding Fig. 2 Fossorochromis rostratus Chembe Sand 1 Sifter Fig. S1 Hemitilapia oxyrhynchus Masasa Sand 1 Leaf cleaner Fig. S1 Labeotropheus fuelleborni Makanjila Point Rock 6 Biting Fig. 1E Labeotropheus fuelleborni Nankoma Rib.G Rock 6 Biting Fig. 1E Labeotropheus fuelleborni Makanjila Point Rock 13 Biting Fig. 2 Labeotropheus fuelleborni Chiofu Bay Rock 13 Biting Fig. 2 Labeotropheus fuelleborni Eccles Reef Rock 5 Biting Fig. 2 Labeotropheus fuelleborni Chinyamwezi Rock 14 Biting Fig. 2 Labeotropheus fuelleborni Chinyankwazi Rock 11 Biting Fig. 2 Labeotropheus trewavasae Thumbi West Rock 6 Biting Fig. 1E Labeotropheus trewavasae Nakantenga Rock 6 Biting Fig. 1E Lethrinops auritus Otter Island Sand 1 Sifter Fig. S1 Melanochromis auratus Otter Point Rock 1 Omnivore Fig. S1 Metriaclima aurora Thumbi West Rock 1 Suction feeding generalist Fig. 2 Metriaclima aurora Otter Point Rock 7 Suction feeding generalist Fig. 2 Metriaclima benetos Mazinzi Reef Rock 6 Suction feeding generalist Fig. 1E Metriaclima benetos Mazinzi Reef Rock 9 Suction feeding generalist Fig. 2 Metriaclima mbenjii Mbenji Island Rock 6 Suction feeding generalist Fig. 1E Metriaclima pyrsonotus Nakantenga Rock 6 Suction feeding generalist Fig. 1E Metriaclima zebra Thumbi West Rock 6 Suction feeding generalist Fig. 1E Ophthalmotilapia ventralis Lake Tanganyika — 1 — Fig. S1 Paracyprichromis nigrapinnis Lake Tanganyika — 1 — Fig. S1 Petrochromis famula Lake Tanganyika — 1 — Fig. S1 Petrotilapia tridentiger Ikambe Sanga Rock 1 Aufwuchs comber Fig. S1 Protomelas taeniolatus Thumbi West Sand 1 Suction feeding Fig. S1 Pseudotropheus williamsi Thumbi West Rock 1 Insectivore Fig. S1 Rhamphochromis esox Otter Point Pelagic 1 Piscivore Fig. S1 Serranochromis robustus Riverine (Malawi) — 1 — Fig. S1 Tanganicodus irsacae Lake Tanganyika — 1 — Fig. S1 Tropheops gracilior Otter Island Rock 12 Biting/suction Fig. 1F, Fig. 2 Tropheops intermedius Thumbi West Rock 12 Biting/suction Fig. 1F, Fig. 2 Tropheops sp. Chinyankwazi Rock 15 Biting/suction Fig. 1F, Fig. 2 Tropheops sp. Chinyamwezi Rock 20 Biting/suction Fig. 1F, Fig. 2

Roberts et al. www.pnas.org/cgi/content/short/1018456108 7of8 Table S3. Pixel density analysis of WISH Species Treatment* Gene Tissue† n Proportion labeled pixels (mean ± SE) t DF P (t test)‡

Metriaclima zebra None Ptch1 Lower jaw 7 0.260 ± 0.032 4.285 13 0.0009 Labeotropheus fuelleborni None Ptch1 Lower jaw 8 0.448 ± 0.030 Metriaclima zebra None Ptch1 Tail 7 0.134 ± 0.006 0.906 13 0.39 Labeotropheus fuelleborni None Ptch1 Tail 8 0.142 ± 0.006 Metriaclima zebra None Gli1 Lower jaw 5 0.216 ± 0.020 4.604 8 0.002 Labeotropheus fuelleborni None Gli1 Lower jaw 5 0.357 ± 0.023 Metriaclima zebra None Gli1 Tail 5 0.115 ± 0.010 0.335 8 0.74 Labeotropheus fuelleborni None Gli1 Tail 5 0.119 ± 0.005 Metriaclima zebra None Col1a1 RA 3 0.212 ± 0.012 3.434 6 0.027 Labeotropheus fuelleborni None Col1a1 RA 3 0.283 ± 0.017 Metriaclima zebra None Col1a1 Tail 3 0.327 ± 0.005 3.877 6 0.018 Labeotropheus fuelleborni None Col1a1 Tail 3 0.286 ± 0.010 Labeotropheus fuelleborni Cyc Ptch1 Lower jaw 5 0.122 ± 0.019 5.926 8 0.0004 Labeotropheus fuelleborni Control Ptch1 Lower jaw 5 0.417 ± 0.048 Labeotropheus fuelleborni Cyc Ptch1 Tail 5 0.027 ± 0.005 15.844 8 0.0001 Labeotropheus fuelleborni Control Ptch1 Tail 5 0.118 ± 0.009 Labeotropheus fuelleborni Cyc Gli1 Lower jaw 5 0.208 ± 0.022 5.011 8 0.001 Labeotropheus fuelleborni Control Gli1 Lower jaw 5 0.373 ± 0.024 Labeotropheus fuelleborni Cyc Gli1 Tail 5 0.033 ± 0.002 13.815 8 < 0.0001 Labeotropheus fuelleborni Control Gli1 Tail 5 0.127 ± 0.006 Labeotropheus fuelleborni Cyc Col1a1 RA 3 0.162 ± 0.012 5.325 6 0.006 Labeotropheus fuelleborni Control Col1a1 RA 3 0.294 ± 0.021 Labeotropheus fuelleborni Cyc Col1a1 Tail 3 0.339 ± 0.009 4.062 6 0.015 Labeotropheus fuelleborni Control Col1a1 Tail 3 0.289 ± 0.009

*See Materials and Methods in main text (Cyc, 50 μM cyclopamine in ethanol; Control, ethanol control). †See SI Materials and Methods for exact regions analyzed. ‡Boldface indicates signiﬁcance based on a threshold of P < 0.05.

Table S4. Quantitative PCR gene expression analysis in microdissected tissue Species Gene Tissue n Expression (mean ± SE)* t DF P (t test)†

Metriaclima zebra Ptch1 Lower jaw 6 1.033 ± 0.050 −3.177 9 0.011 Labeotropheus trewavasae Ptch1 Lower jaw 5 1.268 ± 0.055 Metriaclima zebra Ptch1 Trunk/tail 6 1.023 ± 0.085 0.729 9 0.48 Labeotropheus trewavasae Ptch1 Trunk/tail 5 0.931 ± 0.093 Metriaclima zebra Rad26l Lower jaw 6 1.015 ± 0.036 −0.944 9 0.37 Labeotropheus trewavasae Rad26l Lower jaw 5 1.066 ± 0.040 Metriaclima zebra Rad26l Trunk/tail 6 1.036 ± 0.076 −1.095 9 0.30 Labeotropheus trewavasae Rad26l Trunk/tail 5 1.160 ± 0.084

*Relative transcript level vs. M. zebra average for each gene and tissue, normalized to β-actin transcript level as endogenous control (SI Materials and Methods). †Boldface indicates signiﬁcance based on a threshold of P < 0.05.

Table S5. Lower jaw measurements in larval ﬁsh

Species Treatment* n Outlever (μm) (mean ± SD) Inlever (μm) (mean ± SD) MAO (mean ± SD) Labeotropheus fuelleborni untreated control 6 1147.9 ± 83.0 288.2 ± 22.0 0.251 ± 0.019 Labeotropheus fuelleborni ethanol control 7 1176.9 ± 339.4 296.5 ± 94.2 0.250 ± 0.035 Labeotropheus fuelleborni 50 μM cyclopamine 10 1127.2 ± 131.1 212.3 ± 24.4 0.190 ± 0.023 Metriaclima zebra untreated control 7 1516.1 ± 70.2 311.6 ± 33.6 0.206 ± 0.026

MAO,mechanicaladvantageofopening. *See Materials and Methods in main text.

Roberts et al. www.pnas.org/cgi/content/short/1018456108 8of8