New Results

Loss of flight in the Galapagos cormorant mirrors human skeletal ciliopathies

Alejandro Burga, Weiguang Wang, Paul C. Wolf, Andrew M. Ramey, Claudio Verdugo, Karen Lyons, Patricia G. Parker, Leonid Kruglyak
doi: https://doi.org/10.1101/061432

Abstract

Changes in the size and proportion of limbs and other structures have played a key role in the adaptive evolution of species. However, despite the ubiquity of these modifications, we have a very limited idea of how these changes occur on the genetic and molecular levels. To fill this gap, we studied a recent and extreme case of wing and pectoral skeleton size reduction leading to flightlessness in the Galapagos cormorant (Phalacrocorax harrisi). We sequenced and de novo assembled the genomes of four closely related cormorant species and applied a joint predictive and comparative genomics approach to find candidate variants. Here we report that function-altering variants in genes necessary for both the correct transcriptional regulation and function of the primary cilium contributed to the evolution of loss of flight in P. harrisi. Cilia are essential for Hedgehog signaling, and humans affected by skeletal ciliopathies suffer from premature arrest of bone growth, mirroring the skeletal features associated with loss of flight.

Introduction

The evolution of loss of flight is one the most recurrent limb modifications encountered in nature. In fact, Darwin used the occurrence of flightless birds as an argument in favor of his theory of natural selection (Darwin, 1859). He proposed that loss of flight could evolve by two mechanisms: 1) selection in favor of larger bodies and 2) relaxed selection due to the absence of predators. Loss of flight has evolved repeatedly, and cases are found among 26 families of birds in 17 different orders (Roff, 1994). Moreover, recent studies strongly suggest that the ratites (ostriches, emus, rheas, cassowaries and kiwis), long thought to derive from a single flightless ancestor, actually constitute a polyphyletic group characterized by multiple independent instances of loss of flight and convergent evolution (Baker et al., 2014; Harshman et al., 2008; Mitchell et al., 2014). However, despite the ubiquity and evolutionary importance of loss of flight (Wright et al., 2016), the underlying genetic and molecular mechanisms remain largely unknown.

The Galapagos cormorant (Phalacrocorax harrisi) is the only flightless cormorant among approximately 40 extant species (Livezey, 1992). The entire population is distributed along the coastlines of Isabela and Fernandina Islands in the Galapagos archipelago. P. harrisi has a pair of short but perfectly shaped wings, which are smaller than those of any other cormorant (Fig. 1a). This results in a significant deviation from the well-established allometric relationship between wing length and body mass (Livezey, 1992). In addition, the Galapagos cormorant has a highly reduced keel and underdeveloped pectoral muscles. The keel or carina is an extension of the sternum that runs along its midline and provides an attachment surface for the flight muscles, the largest muscles in birds. Flightless taxa, such as ratites and Cretaceous Hesperornis, typically evolve flat sternums in which the keel has been largely reduced or lost (Feduccia, 1999). Surprisingly, despite its striking phenotype, the Galapagos Cormorant escaped Darwins attention during his discovery voyage on the Beagle. In contrast to ratites and penguins, which became flightless over 50 million years ago (MYA) (Mitchell et al., 2014; Slack et al., 2006), the Galapagos cormorant and its flighted relatives are estimated to share a common ancestor 2 MYA (Kennedy et al., 2009). This recent and extreme modification of wing size and pectoral skeleton makes P. harrisi an attractive model to study loss of flight.

To investigate the genetics of flightlessness evolution, we applied a joint predictive and comparative genomics approach to first identify all coding variants between the Galapagos cormorant and its flighted relatives, and to then classify these variants according to their probability of affecting protein function. Among the strongest function-altering variants present in P. harrisi, we found a significant enrichment for genes mutated in human skeletal ciliopathies. The primary cilium is essential for Hedgehog (Hh) signaling in vertebrates, and individuals affected by ciliopathies have small limbs and rib cages, mirroring the phenotype of P. harrisi. Our results indicate that the combined effect of variants in genes necessary for both the correct transcriptional regulation and function of the primary cilium contributed to the evolution of highly reduced wings and other skeletal adaptations associated with loss of flight in P. harrisi.

High quality genome sequences of four cormorant species

To identify variants associated with loss of flight, we sequenced and de novo assembled the 1.2 Gb genomes of the Galapagos cormorant (Galapagos Islands, Ecuador) and three flighted cormorant species: the Doublecrested cormorant (Phalacrocorax auritus; Minnesota, USA), the Neotropical cormorant (Phalacrocorax brasilianus; Valdivia, Chile), and the Pelagic cormorant (Phalacrocorax pelagicus; Alaska, USA) (Supplementary Fig. 1). P. auritus and P. brasilianus are the closest relatives of P. harrisi (Kennedy et al., 2009), and P. pelagicus is part of a sister clade and served as an outgroup. Genomes were assembled from a combination of short insert and mate-pair Illumina libraries using SOAPdenovo2 (Luo et al., 2012). Details about the sequenced individuals and library statistics can be found in Supplementary Table 1a and Supplementary Table 2. Among these four genomes, the Galapagos cormorants assembly had the longest contig and scaffold N50 metrics (contig N50: 103 kb; scaffold N50: 4.6 Mb; Supplementary Table 1b). We evaluated the completeness of the cormorants genomes by estimating the total number of uniquely annotated proteins in each assembly and by using the CEGMA pipeline (Parra et al., 2009)(see Methods). Overall, we found good agreement between these two independent metrics in a dataset including the four cormorant genomes and 17 recently published bird genomes (r2 = 0.75 P = 4.3×10-07; Fig. 1b and Supplementary Table 3). Interestingly, commonly used metrics of assembly quality such as contig and scaffold N50 were very poor predictors of the total number of proteins present in each assembly (r2 = 0.13 P = 0.15 and r2 = 0.06 P = 0.81; Supplementary Fig. 2 Supplementary Table 3).Three of the four cormorant genomes (P. harrisi, textitP. auritus and textitP. pelagicus) obtained the highest CEGMA scores and number of uniquely annotated genes of all bird genomes (red triangles, Fig. 1d). For instance, the following CEGMA scores were obtained for the cormorants: P. harrisi, 90.3%; P. auritus, 91.3%; P. brasilianus, 72.6%; P. pelagicus, 87.1%. In contrast, Sanger and PacBio genomes obtained lower scores: Gallus gallus (Sanger assembly) (Consortium, 2004), 80.7%; Taeniopygia guttata (Sanger assembly)(Warren et al., 2010), 71.4%; and Melopsittacus undulatus (PacBio assembly) (Ganapathy et al., 2014), 79.0%. Thus, our cormorant genomes are the most complete bird genomes published to date from a gene-centric perspective, performing even better than genomes assembled from Sanger sequences and PacBio long reads (complete statistics in Supplementary Table 3).

Figure 1:
Figure 1:

The Galapagos Cormorant, a model to study flightlessness evolution. a, The Galapagos cormorant has evolved very short wings. The average wing length of an adult male is 19 cm (3.6 kg. body mass), whereas the wing length of its closest relative, the doublecrested cormorant, is 31.5 cm. (2.2 kg. body mass). b, CEGMA score is a good predictor of genome completeness from a gene-centric perspective. Blue line is the linear regression model (r2 = 0.75 P = 4.3×10-07). Genomes reported in this study are red triangles and other published avian genomes are black circles (see Supplementary Table 3 for full details). c, Phylogenetic tree reconstructed using fourfold degenerate sites from whole genome sequences. In orange, time span between approximate origin of proto-Galapagos archipelago (9 MYA) and the origin of the oldest extant island, San Cristobal (4 MYA). Nodes represent median divergence ages. Blue bars indicate the 95% Highest Posterior Density (HPD) Interval. d, Heterozygosity levels inferred from whole genome sequences. Birds are not drawn to scale.

Phylogeny and genetic diversity

We reconstructed the cormorant phylogeny using a Bayesian framework (see Methods). Our analysis confirmed the previously reported phylogenetic relationships among the four sequenced species (Fig. 1c). Moreover, our results indicate that P. harrisi last shared a common ancestor with P. auritus and P. brasilianus 2.37 MYA, in agreement with a previous estimate based on mitochondrial DNA (Kennedy et al., 2009) (Fig. 1c). The oldest extant island in the Galapagos archipelago, Espaola, emerged at most 4 MYA, and proto-Galapagos islands existed at least 9 MYA (Harpp, 2014). Our results are consistent with the view that P. harrisi lost the ability to fly while inhabiting the archipelago. We calculated the proportion of single nucleotide polymorphism (SNP) heterozygous sites for each sequenced individual to estimate the levels of intra-specific genetic diversity (Fig. 1d). P. harrisi showed the lowest proportion of heterozygous SNPs among the sequenced cormorants (0.00685%; Fig. 1d). The heterozygosity of P. harrisi is even lower than that of the Crested Ibis, Nipponia nippon, a highly endangered bird that nearly went extinct in the 1980s (Wang and Li, 2008) (0.0172%; Fig. 1d). The low level of heterozygosity found in the Galapagos cormorant is most likely due to its small population size ( 1,500 individuals) and multiple population bottlenecks (Valle and Coulter, 1987).

Discovery and characterization of function-altering variants in P. harrisi

For our variant discovery approach to be effective, it was imperative to interrogate most of the Galapagos cormorants genes. To increase our power to do so, we annotated genes using two different strategies: homology-based and transcriptome-based gene annotations (Supplementary Fig. 3). The latter was derived using mRNA expression data from the developing wing of a double-crested cormorant embryo (Supplementary Fig. 5c, see Methods). We then predicted all missense, deletion, and insertion variants in ortholog pairs between P. harrisi and each of its three flighted relatives (see Methods and Supplementary Fig. 2). We used PROVEAN (Choi et al., 2012), a phylogeny-corrected variant effect predictor, to evaluate the impact on protein function of each of the Galapagos cormorants variants on a genome-wide scale. A PROVEAN score is calculated for each variant; the more negative the score, the more likely a given variant is to affect protein function in the Galapagos cormorant. To illustrate our approach, Fig. 2a shows the distribution of PROVEAN scores obtained when comparing 12,442 ortholog pairs between P. harrisi and P. auritus. Of these 12,442 ortholog pairs, 4,959 (40%) did not contain coding variants; the remaining 7,483 pairs contained a total of 23,402 coding variants. These variants were further classified as: 22,643 single amino acid substitutions, 456 deletions, and 303 insertions (Fig. 2b). Most variants were predicted to be neutral (distribution is centered around zero). As expected, deletion and insertions were enriched in the tails of the distributions (Fig. 2b). Very similar numbers of variants and PROVEAN score distributions were obtained for the other homology-based (Supplementary Fig. 4) and transcriptome-based annotations (Supplementary Fig. 5).

Enrichment for genes mutated in skeletal ciliopathies

We reasoned that variants responsible for loss of flight should be enriched among those affecting protein function in P. harrisi. To find those variants, we applied a stringent threshold to our four prediction datasets: PROVEAN score < −5, two times the threshold for human disease variants discovery (Choi et al., 2012) (Fig. 2a, see Methods). Variants with a PROVEAN score < −5 are typically at residues that have been perfectly conserved at least since mammals and birds last shared a common ancestor (300 MYA), and consequently, they very are likely to be important for protein function.

Gene enrichment analysis of function-altering variants in the Galapagos cormorant revealed a significant overrepresentation of human developmental disorders (see Methods and Supplementary Table 4a). Strikingly, 8 out of the 19 significant categories were related to syndromes affecting limb development, such as polydactyly, syndactyly, and duplication of limb bones. Control enrichment analysis didnt reveal any limb syndrome disorder in the other flighted cormorants (see Methods, Supplementary Table 4b and 4c).

The genes underlying the enrichment for limb syndromes were, upon examination, all mutated in a family of human disorders known as ciliopathies. Cilia are hair-like microtubule based structures that are nucleated by the basal body (centriole and associated proteins) and project from the surface of cells. They are essential for mediating Hedgehog (Hh) signaling, serving as antennae for morphogens during development (Goetz and Anderson, 2010). We confirmed by Sanger sequencing the presence of function-altering variants in Ofd1, Evc, TalpidS, Dync2h1, Ift122, Wdr34, and Kif7, all of which are necessary for the assembly or functioning of the primary cilium and are mutated in human ciliopathies, particularly those affecting the skeleton (Table 1). Individuals affected by skeletal ciliopathies have small limbs and rib cages, mirroring the main features of the Galapagos cormorant. Furthermore, our enrichment analysis also revealed the presence of a function-altering variant in Gli2, a transcription factor essential for Hh signaling (Haycraft et al., 2005) (Table 1).

Figure 2:
Figure 2:

Distribution of the effect of variants betweenP. auritus. andP. harrisi. a, We used PROVEAN to predict the effect on protein function of 23,402 variants contained in 12,449 orthologous pairs betweenP. auritus. andP. harrisi. 4,966 pairs contained no variants. The more negative the score; the more likely the variant affects protein function. PROVEAN score thresholds used in this study are drawn as vertical dashed lines. Number of proteins and variants found for each threshold are presented in inset table. b, Density of PROVEAN scores for each class of variant. The same variants presented in (a) were classified as single amino acid substitutions, deletions, and insertions. Number of variants in each class is indicated in the legend.

View this table:
Table 1:

Function-altering variants inP. harrisi are enriched for genes that cause skeletal ciliopathies in humans. Sanger validated examples of function-altering variants (PROVEAN score < −5) in P.harrisi. Cilia/Hh related genes were found based on functional enrichment for human syndromes. PCP (Planar cell polarity) genes were selected based on literature evidence linking cilia and PCP. These variant are fixed in the population. (*) Based on phenotypic similarity to mutant mouse model

The Galapagos cormorants ortholog of human OFD1 (mutated in Oral-facial-digital syndrome 1) is affected by three function-altering variants with PROVEAN score < −5 (R325C −6.913, K517T −5.673 and E889G −5.068; Supplementary Fig. 6a and 6b). Ofd1 mutant mice display polydactyly and shortened long bones (Bimonte et al., 2011). Also, a function-altering missense variant (Q691L −5.491) was found in Ift122, a component of the IFT complex that controls the ciliary localization of Shh pathway regulators (Ocbina et al., 2011). The mutated glutamine in Ift122 is virtually invariant among eukaryotes ranging from green algae to vertebrates (Supplementary Fig. 6c).

There is a well-established genetic link between the Planar Cell Polarity (PCP) pathway and cilia (Goetz and Anderson, 2010). For instance, intu and fuz are core downstream effectors of the PCP pathway in Drosophila, and their respective Xenopus (Park et al., 2006) and mouse (Gray et al., 2009; Zeng et al., 2010) mutants are characterized by short cilia and disrupted Hh signaling. We searched our P. harrisi data set (PROVEAN score < −5) for members of the PCP pathway and found function-altering variants in Fat1 atypical cadherin (Fat1), Dachsous cadherin-related 1 (Dchs1) and Disheveled-1 (Dvl1) (Supplementary Fig. 6d; Table 1). The Galapagos Cormorants FAT1 contains two function-altering variants (S1717L and Y2462C, Table 1) and Fat1 mouse mutants show very selective defects in muscles of the upper body (Caruso et al., 2013). In addition, Dvl1 is mutated in patients with Robinow syndrome, characterized by limb shortening (Bunn et al., 2015; White et al., 2015).

Additional Sanger sequencing of 20 Galapagos cormorant individuals from two different populations (Cabo Hammond and Caones Sur (Duffie et al., 2009)) revealed only homozygous carriers for all of the variants in Table 1, indicating that these variants are most likely fixed in the Galapagos Cormorant. Overall, we found an overrepresentation of function-altering variants in genes that, when mutated in humans, cause skeletal ciliopathies and bone growth defects.

Cuxl is mutated in P. harrisi.

To identify the strongest function-altering variants inP. harrisi, we applied a more stringent PROVEAN score threshold: −12.5 delta alignment score, five times the threshold used for human disease variants discovery (Choi et al., 2012). This strategy narrowed our search to 23 proteins. These represent 0.16% of annotated proteins inP. harrisi (Supplementary Table 5). We manually curated these 23 proteins and performed additional Sanger sequencing reducing the list of proteins with confirmed or putative variants to 12. (Supplementary Table 5; see Supplementary Note for a detailed analysis). These variants were exclusively small deletions. Among these 12 proteins, two stood out based on their known role in development: LGALS-3 and CUX1. LGALS-3 is affected by a 7 amino acid deletion inP. harrisi (PROVEAN score −26.319). LGALS-3 (Galectin-3) is localized at the base of the primary cilium and is necessary for correct ciliogenesis in mice (Koch et al., 2010), yet it has not been implicated in human ciliopathies. Moreover, LGALS-3 physically interacts with SUFU, an important regulator of mammalian Hh signaling (Paces-Fessy et al., 2004), and knockout mice show pleiotropic defects in chondrocyte differentiation (Colnot et al., 2001).

In addition, we found a 4 amino acid deletion (PROVEAN score −15.704) in CUX1. CUX1 (cut-like homeobox 1), also known as CDP, is a highly conserved transcription factor with diverse developmental roles that contains four DNA binding domains: three CUT domains (CR1-3) and one homeodomain (HD) (Fig. 3a) (Sansregret and Nepveu, 2008). It is noteworthy that cut, the Drosophila ortholog of Cux1, is necessary for proper development of wings and flight muscles in flies(Sudarsan et al., 2001). Moreover, the expression pattern of cut is lost in the wing imaginal disc of flightless ant populations compared to their flighted relatives (Fave et al., 2015). In chicken, Cux1 mRNA expression in the limb is restricted to the ectoderm bordering the Apical Ectodermal Ridge (AER) at embryonic stage 23 (Tavares et al., 2000). The AER is one of the key signaling centres that drive limb development. Furthermore, adenovirus-mediated overexpression of a dominant negative form of CUX1 in the developing chicken wing results in AER disruption and limb truncations most strongly affecting distal skeletal elements (digits, radius and ulna) (Tavares et al., 2000). Interestingly, inP. harrisi, the radius and ulna are disproportionately small compared to the humerus (Livezey, 1992).

We Sanger-sequenced and confirmed the predicted Cux1 12 base pair (bp) deletion in P. harrisi. We also confirmed that this variant was fixed in the population and absent in the other cormorant species (Supplementary Fig. 7a). The 12 bp deletion in Cux1 removes four amino acids, AGSQ, immediately adjacent to the C-terminal end of the homeodomain (Fig. 3b). We will refer to this variant as CUX1-Δ4aa. Alignment of CUX1 orthologs from available vertebrate genomes revealed that the 4 missing residues are extremely conserved among tetrapods (Fig. 3b). The deleted serine is phosphorylated in human cells (Matsuoka et al., 2007), but the molecular consequences of this modification are unknown. The Cux1 deletion does not include any of the predicted residues responsible for DNA contact and recognition (Iyaguchi et al., 2007), but given its close proximity to the homeodomain, we decided to test whether the DNA binding activity of CUX1 was affected. We performed Electrophoretic Mobility Shift Assay (EMSA) with purified CR3HD CUX1-Ancestral and CUX1-A4aa protein variants (Supplementary Fig. 7b), as previously described (Moon et al., 2000), and found that the DNA binding activity was not abolished in the deletion variant (Supplementary Fig. 7c). CUX1 is able to both directly repress and activate gene expression through its C-terminal tail (Mailly et al., 1996; Truscott et al., 2007). We performed a luciferase reporter assay (Mailly et al., 1996; Nishio and Walsh, 2004) and found that both variants were equally capable of repressing the expression of a UAS/tk luciferase reporter (Fig. 3d). Thus, the Galapagos cormorant CUX1-Δ4aa variant affects neither the DNA binding activity in vitro nor the C-terminal mediated repression activity in COS-7 cells.

CUX1 regulates the expression of cilia and PCP genes

We hypothesized that the Cux1 deletion variant is mechanistically related to the enrichment of function-altering variants in ciliopathy-related genes based on the following observations: First, transgenic mice overexpressing the CUX1-CR3HD isoform develop polycystic kidneys. Cilia from the cystic epithelial cells from these animals were twice as long as the ones derived from control epithelial cells (Cadieux et al., 2008). Second, the CUX1-CR2CR3HD isoform has been shown to directly up-regulate the expression of Ftm, a component of the cilia basal body that is mutated in human ciliopathies and involved in Shh signaling (Stratigopoulos et al., 2011). Third, Cux1 knockout mice show deregulation of SHH expression in hair follicles (Ellis et al., 2001).

To test whether Cux1 could globally regulate the expression of cilia genes, we analyzed expression array data from a study that profiled Hs578t cells stably expressing an shRNA against Cux1, as well as cells overexpressing the human CUX1-CR2CR3HD isoform (Vadnais et al., 2013). In concordance with the well-established role of Cux1 as a regulator of cell growth and proliferation (Ramdzan and Nepveu, 2014), genes significantly up-or down-regulated in both conditions (p < 0.05 and > 1.1 fold change) were enriched for pathways such as Cell cycle and Mitotic G1-G1/S phases (P = 3.99×10-5 and 0.016, respectively: Supplementary Table 6). Importantly, we also found and enrichment for cilia-related categories such as Assembly of the primary cilium and Intraflagellar transport (P = 1.2×10-3 and 5.7×10-3 respectively; Supplementary Table 6). These results suggest that cilia-related genes are indeed enriched among Cux1 target genes.

To further test whether Cux1 can regulate ciliary genes in an appropriate cellular context, we generated ATDC5 stable lines expressing N-terminal HIS-tagged versions of CR3HD CUX1-Ancestral and CUX1-Δ4aa variants. ATDC5 is a well-characterized mouse chondro-genic cell line that largely recapitulates in vitro the differentiation landmarks of chondrocytes (Shukunami et al., 1996). We chose to express the CR3HD isoform because western blot analysis revealed that this was the most abundant CUX1 isoform expressed in the developing wing of mallard embryos ( 50 kDa; Fig. 3a). We performed quantitative reverse transcription PCR (RT-qPCR) on a selected number of genes containing predicted strong function-altering variants inP. harrisi (Table 1) and showing detectable levels of expression in ATDC5 cells. In addition, we measured the expression of Ptch1, the receptor of the Hh pathway. Our experiments revealed that the CUX1-Ancestral variant transcriptionally up-regulated the expression of Ofd1 (1.7 fold, P = 1.2×10-6; Fig. 3c) and Fat1 (1.8 fold, P = 0.029; Fig 3c) and downregulated the expression of Ift122 (0.77 fold, P=2. 5x10-3; Fig 3c) and Ptch1 (0.53 fold, P=1.4×10-2;Fig 3c) compared to the control line. In contrast, neither Dync2h1 nor Wdr34 expression levels were changed by CUX1-Ancestral overexpression (Fig. 3c). These results confirmed that cilia and Hh related genes are indeed transcriptional targets of CUX1 in chondrocytes.

Figure 3:
Figure 3:

The Galapagos Cormorants Cux1 is a transcriptional activation hypomorph. a, Western blot showing the expression of CUX1 isoforms in the developing wing of a mallard embryo (22 days). The most abundant band corresponds to the predicted size of the CR3HD CUX1 isoform. b, Protein alignment showing the deleted AGSQ-residues in the Galapagos Cormorant CUX1 and their high degree of conservation among vertebrates. c, Differential up-regulation of genes by CUX1-Ancestral and CUX1-Δ4aa variants in ATDC5 cells. Pooled stable lines carrying CR3HD CUX1-Ancestral or CUX1-Δ4aa variants were generated by lentiviral transduction and puromycin selection. Controls cells were transduced with an empty vector. Gene expression levels were measured by reverse transcriptase quantitative PCR (n=5 biological replicates, each compromising 3 technical replicates). d, Luciferase based assay to test the repression activity of CUX1 C-terminal domain lacking CR3 and HD domains. GAL4 DNA binding domain was fused to CUX1-Ancestral or CUX1-Δ4aa variants. Both constructs equally repressed a promoter containing UAS binding sites in COS-7 cells (n=3 biological replicates, each compromising 3 technical replicates). Gene expression levels were measured by RT-qPCR (n=5 biological replicates, each compromising 3 technical replicates). Error bars indicate standard errors. We used ANOVA and Tukeys HSD to test for statistical significance. Black stars indicate a significant difference between a CUX1 variant and the control. Red starts indicate a significant difference between CUX1-Ancestral and CUX1-Δ4aa variants. Absence of stars indicates no significant difference. (*) P<0.05, (**) P<0.01, and (***) P<0.001

Impaired transcriptional activity of the Galapagos cormorant CUX1

Remarkably, we found that the Galapagos cormorants CUX1 showed impaired transcriptional activity compared to the ancestral variant. Ofdl upregulation was significantly reduced in CUX1-Δ4aa cells compared to CUXl-Ancestral cells (1.7 vs. 1.2 fold, P = 6×10-5; Fig. 3c). Similarly, Fatl expression levels in CUX1-Δ4aa cells were not significantly different from control lines (1.2 fold, P=0.17; Fig 3c). However, downregulated genes (Ift122 and Ptchl) showed no significant differences between CUX1-Ancestral and CUX1-Δ4aa cells (0.78 fold vs. 0.78 fold; P = 0.99 and 0.53 fold vs. 0.56 fold, P=0.96; Fig. 3c). These results indicate that the four amino acids deletion in the Galapagos Cormorant CUX1 exclusively affects its ability to activate gene expression and are consistent with our luciferase reporter assays, which show no effect on repression (Fig. 3d). Furthermore, these results indicate that both the transcriptional activator (Cuxl) and its target genes (Ofdl and Fatl) are affected by strong function-altering variants in the Galapagos Cormorant.

CR3HD-CUX1 promotes chondrogenesis

Chondrocytes are the main engine of bone growth. The growth of skeletal elements depends on the precise regulation of chondrocyte proliferation and hypertrophy. Mutations that affect cilia result in the premature arrest of bone growth due defects in Indian Hedgehog (IHH) signaling in chondrocytes (Kronenberg, 2003). To test the role of CUX1-CR3HD in chondrogenesis, we differentiated control, CUX1-Ancestral and CUX1-Δ4aa ATDC5 cell lines and quantified the expression of Ihh and Sox9, two well-established markers of chondrocyte differentiation in vitro and in vivo (Kronenberg, 2003). We found that overexpression of both CR3HD CUX1-Ancestral and CUX1-Δ4aa variants significantly promoted chondrogenic differentiation of ATDC5 cells after 7 and 12 days of differentiation (Fig. 4a). However, the CUX1-Δ4aa variant was not as efficient as the Ancestral variant, showing significant differences from CUX1-Ancestral in Ihh expression after 7 days of differentiation ( 50% decrease, P = 5.9×10-4; Fig. 4a) and for Sox9 after 12 days ( 15% decrease, P = 1.6×10-2; Fig. 4a).

These results indicate that the Galapagos cormorant CUX1 is not as effective as the ortholog from its flighted relatives in promoting chondrogenic differentiation, and that mutations in Cux1 are sufficient to affect the dynamics of chondrogenesis. This observation is further supported by the fact that CUX1 is expressed in the hypertrophic chondrocytes of developing bones in mice, and that the bones of Cux1 mutant mice are thin and flaky (Sinclair et al., 2001)

Possible evolutionary scenarios

Loss of flight has traditionally been attributed to relaxed selection. In such a scenario, the first cormorants that inhabited the Galapagos Islands found a unique environment that lacked predators and provided food year long, drastically reducing the need to migrate. However, we found no evidence for pseudogenization of developmental genes inP. harrisi (see Supplementary Note for a detailed analysis on missing genes and premature stop codons). On the other hand, loss of flight in the Galapagos Cormorant is thought to confer an advantage for diving by decreasing buoyancy (shorter wings) and indirectly allowing an increase in body size (larger oxygen storage) (Watanabe et al., 2011). This advantage could make flightlessness a target of positive selection. To evaluate if any of our candidate genes (Table 1) showed signatures of positive selection in the Galapagos Cormorant lineage, we estimated their ratio of non-synonymous to synonymous substitutions (ω = dN/dS). This is a very stringent selection test because it assumes all sites in a protein are evolving under the same selective pressure, a condition that is rarely met in highly conserved regulatory genes (Yang, 1997). We found that three out of eleven tested genes showed signs of positive selection (ω >1) in the Galapagos cormorant lineage compared to a background phylogeny of 35 Taxa (Ofdl ω=1.92, Evc ω=1.93, Gli2 ω=1.10; Supplementary Table 7). One of these three genes, Gli2, showed a statistically significant difference (ω=1.10 (Galapagos branch) vs. ω=0.11 (Background branch), P = 2.4×10-3; Supplementary Table 7). In contrast, GU2 showed no signs of selection in the Galapagos Cormorants sister group (P. auritus and P. brasilianus) (ω=0.11 (Galapagos branch) vs. ω=1×10-4 (Background branch), P = 0.46). As a control, we also analyzed Gli3, the partially redundant paralog of Gli2, which also mediates Hh signaling but has no predicted function-altering variants inP. harrisi and found no evidence for positive selection (ω=0.04 (Galapagos branch) vs. ω=0.15 (Background branch), P = 0.11; Supplementary Table 7). These results suggest that positive selection could be partially responsible for the flightless phenotype of P. harrisi.

Figure 4:
Figure 4:

CR3HD CUX1 promotes chondrogenesis. a, ATDC5 control cells and cells carrying CR3HD CUX1-Ancestral or CUX1-Δ4aa variants were differentiated into chondrocytes. Gene expression levels were measured by RT-qPCR (n=4 biological replicates, each compromising 3 technical replicates) after 7 and 12 days. Error bars indicate standard errors. We used ANOVA and Tukeys HSD to test for statistical significance. Black stars indicate a significant difference between a CUX1 variant and the control. Red starts indicate a significant difference between CUX1-Ancestral and CUX1-Δ4aa variants. Absence of stars indicates no significant difference. (*) P<0.05, (**) P<0.01, and (***) P<0.001. b, Proposed mechanism for the reduction of wing size inP. harrisi. The left panel depicts the normal functioning of IHH signaling pathway in vertebrates. A selected number of proteins are shown for simplicity. OFD1 and TALPID3 are part of the basal body from which cilia are nucleated. textitOfd1 and talpid3 mutants are defective in cilia formation. In the presence of IHH secreted by hypertrophic chondrocytes, PTCH1 exits the cilium, where it cannot longer inhibit SMO. SMO is then activated and enriched in the cilium in association with EVC. KIF7 mediates the movement of GLI2/3-SUFU complex to the cilium. Then a series of molecular events leads GLI2/3 activation, transcription of IHH target genes, and chondrocyte proliferation. CUX1 promotes the expression of cilia related genes such as textitOfd1 and promoters chondrogenesis. The right panel depicts the state of the IHH pathway in P. harrisi. Proteins in red are affected by function-altering variants in P. harrisi. We predict these variants will affect both cilia formation and functioning leading to a reduction in IHH pathway activity. As a result, the pool of proliferating chondrocytes would decrease in wing bones and the number of hypertrophic chondrocytes would increase resulting in impaired bone growth.

Discussion

The study of flightlessness evolution in the Rallidae family led Olson to propose in 1973 that flightlessness could be a fast-evolving heterochronic condition (Feduccia, 1999; Olson, 1973). Heterochrony, the relative change in the rate or timing of developmental events among species, is thought to be an important factor contributing to macroevolutionary change(Gould, 1977). Yet, virtually nothing is known about its genetic and molecular mechanisms. Diverse myological, osteological, and developmental observations strongly suggest that flightlessness in the Galapagos cormorant is caused by the retention into adulthood of juvenile characteristics affecting pectoral and forelimb development (a class of heterochrony known as paedomorphosis) (Livezey, 1992).

In this study we provide a genetic and molecular model that accounts for this heterochronic condition. We show that Cux1, a highly conserved transcription factor, has a four amino acid deletion in its regulatory domain. We molecularly dissected the consequences of this deletion and demonstrated that this variant impairs the ability of CUX1 to transcriptionally up-regulate cilia related genes (some of which are affected by variants themselves like Ofd1 and Fat1) and promote chondrogenic differentiation. The wing and sternum of birds are some of the last structures to grow and ossify, and most of this growth occurs after hatching (Dunn, 1975; Olson, 1973). A premature arrest of bone growth due to reduced IHH signaling would mainly affect these organs. Thus, perturbations of cilia/Ihh signaling could be responsible for the reduction in growth of both wings and keel in the Galapagos cormorant. Of special interest is the gene Fatl, a target of Cux1 (Fig. 3d), which contains two function-altering variants (Table 1). Fat1 −/- mouse mutants show selective defects in face, pectoral and shoulder muscles but not in in hindlimb muscles (Caruso et al., 2013). Thus, variants in Fat1 could explain, in addition, the underdeveloped pectoral muscles ofP. harrisi.

Although we have identified multiple variants that very likely contribute to the flightless phenotype ofP. harrisi, we cannot exclude the contribution of other genes and pathways that our analysis may have missed, nor the contribution of non-coding regulatory variants (see Supplementary Note). Further characterization of the individual and joint contributions of the variants found in this study will help us to reconstruct the chain of events leading to flightlessness and to genetically dissect macroevolutionary change. We hypothesize that mutations in cilia or functionally related genes could be responsible for limb and other skeletal heterochronic transformations in birds and diverse organisms, including humans.

Methods

Library preparation and sequencing DNA fromP. harrisi was extracted from blood samples preserved in lysis buffer using the phenol-chloroform method. DNA fromP. auritus., P. brasilianus and P. pelagicus was extracted from fresh tissues using the Qiagen DNA extraction kit and genomic tips (500/G). Before preparing DNA libraries, we checked the integrity of the samples by running 200 ng of DNA on a 0.6% agarose gel and observing a single high molecular weight band without smear. We generated libraries using three different protocols depending on the insert size. See Supplementary Table 2. For the smallest insert libraries (160bp) we used the Nextera protocol (Illumina). For mid-sized inserts (380, 430, and 670bp), we used the Truseq Nano protocol (Illumina). For the mate-pair libraries (2, 3, 4, 6, and 9 kb) we used the Nextera Mate-pair protocol (Illumina). We followed standard protocols with the following exceptions. For Nextera and Truseq protocols, we performed agarose size selection to guarantee a small variance in insert size. For Nextera Mate-pair protocols we used the Gel-plus protocol starting from 4 μg of genomic DNA. To shear circularized DNA we used 130μL microTube AFA Snap-cap vials and the Covaris M220 sonicator. All libraries were quantified using Qubit HS and the qPCR KAPA library quantification kit. Small insert and mate-pair libraries were sequenced using 100bp paired-end reads on an Illumina HiSeq 2000 at the UCLA Broad Stem Cell Center. The 170bp, 380bp insert, and mate-pair libraries were sequenced using a loading concentration of 12pM. 430bp insert and 670bp insert libraries were sequenced using 15pM and 18pM.

Genome assembly Sequence reads were filtered to meet genome assembly standards. We removed all duplicated reads. We used Trimmomatic (Bolger et al., 2014) to remove reads containing adapter sequences. We filtered out those reads having 40% of bases with Illumina Q-score 7 (Nextera and Truseq Nano libraries) and 60% of bases with Illumina Q-score 7 (Nextera mate-pair libraries). We used NextClip (v0.8)(Leggett et al., 2014) to remove the junction adapter from mate-pair libraries. We corrected errors (with the exception of mate-pair libraries) using SOAPec (v2.01), an error correction tool based on k-mer frequency. We assembled genomes using SOAPdenovo2 (Luo et al., 2012). We assembled the four genomes were using a k-mer size of 35. Small and mid-sized insert libraries were used for contig and scaffold assembly, and mate-pair libraries were used only for scaffolding. Once the assemblies were completed, we used GapCloser (v1.12) to fill the gaps (poly-N stretches) in the final assembly.

Genome completeness evaluation We ran the CEGMA pipeline (v2.4)(Parra et al., 2009) using: geneid (v1.4), genewise (v2.4.1), HMMER (v3.0) and NCBI Blast (v2.2.25+). This pipeline provides an unbiased estimate of assembly completeness based on its ability to retrieve a pre-defined subset of 248 highly conserved eukaryotic genes (CEGs). All genomes were analyzed with default parameters and the option-vrt (optimization for vertebrate genomes). The number of uniquely annotated proteins was calculated using the same pipeline used for the annotation of the genomes. To make a fair comparison between our genomes and those in the literature, we re-annotated 17 previously published bird genomes using exactly the same pipeline. The CEGMA score was highly correlated with the total number of uniquely annotated proteins (Fig. 2d).

Homology-based annotation We annotated the genomes using a homology-based approach from a set of 14,052 high confidence chicken proteins, for which there is evidence of expression in chicken and/or conservation in other species (Ensembl release 78, Galgal4). This allowed us to annotate 13,677 (97.3%), 13,652 (97.2%), 13,652 (97.2%), and 13,116 (93.3%) unique proteins inP. harrisi,P. auritus,P. pelagians and P. brasilianus, respectively. We also used this pipeline to predict orthologous proteins in the recently published Great cormorant (Phalaaroaorax carbo) genome (Jarvis et al., 2014) and found only 11,994 (85.0%) of them. The lower completeness of the P. carbo genome is most likely due to the lower coverage and lack of mate-pair libraries sequences. First, we used genBlastA (v1.38)(She et al., 2009) to define high-scoring segment pairs (HSPs) for each protein with options-c 0.3-e 0.00001-gff-pid-r 1. Then, we used exonerate (v2.2.0) (Slater and Birney, 2005) to build a gene model for every chicken protein using the genomic region defined by the HSP and 100kb of additional sequence surrounding it. We selected the model with the best score for each protein. If more than one model had the same score, we choose one at random. Our final datasets contain only one putative ortholog per chicken protein. A protein is defined as present in a given assembly if at least 30% of the length of the orthologous chicken protein can be mapped.

Transcriptome-based annotation We isolated the developing wings from aP. auritus embryo (Supplementary Fig. 5c). We estimated the embryo to be 15 days based on a photographic guide(Powell et al., 1998). This is the same individual whose DNA was used to generate the reference genome. Total RNA was extracted from the wings using the RNeasy Isolation kit (QIAGEN). RNA quality (RIN: 8.8) was determined using the RNA ScreenTape on an Agilent 2200 TapeStation (Agilent Technologies). We used the TruSeq Stranded mRNA kit (Illumina) to prepare the library and sequenced it using 100bp paired-end reads on an Illumina Hiseq 2000. We used 120 million paired-end reads (without removing duplicates) to de novo assemble the transcriptome using Trinity (v2.0140717) (Grabherr et al., 2011). We identified 63,500 candidate coding regions within 211,607 contigs using Transdecoder. 50.305 out of the 63,500 protein isoforms were further selected because they retrieved a blast hit against a database of the 14,052 high quality Chicken proteins (e =1×10-6). These 50.305 isoforms correspond to 11,753 of the 14,052 unique chicken proteins. We finally annotated the genome ofP. harrisi and P. pelagicus (control for gene enrichment) using the 50.305 predicted protein isoforms from P. auritus. in an analogous way to the homology-based approach using genBlastA and exonerate.

Phylogenetic tree and divergence estimates We obtained common orthologs from the whole genome sequences of five different avian species: G. gallus (outgroup), N. nippon,P. harrisi,P. auritus, P. brasilianus, P. pelagicus, and P. carbo. We used BLAST to identify wrongly assigned orthologs and removed those genes from the dataset. Then, we aligned orthologs with MUSCLE(Edgar, 2004) and extracted all the fourfold degenerate sites. Our final dataset consisted of 9,641 orthologs containing a total of 1,717,714 fourfold degenerate sites. We used MrBayes (v3.2.5)(Ronquist et al., 2012) to reconstruct the phylogeny and estimate species divergence time using the following parameters: lset nst=6 rates=invgamma; prset brlenspr=clock:uniform, prset clockvarpr=igr, prset igrvarpr = exponential(2), prset clockratepr = normal(0.002,0.001), prset treeagepr = offsetexponential(0,500), calibrate tree = truncatednormal(70, 88,3). We used a GTR + I + Γ nucleotide substitution model. An estimate for substitution rate of fourfold degenerate sites in birds (0.002 substitutions per site per million years) was obtained from Green et al.(Green et al., 2014). We used the independent gamma rates (IGR) relaxed clock model. Due to the lack of an appropriate fossil for this phylogeny, the only calibration applied was the split between Galloanseres and Neoaves based on the estimate by Jarvis et al.(Jarvis et al., 2014). The analysis was run for 5 million generations.

Heterozygosity calculation For each species, we aligned using bwa (v0.6.2)(Li and Durbin, 2009) 20X of reads, originally used to generate the assembly. We then use samtools (v0.1.18)(Li et al., 2009) to call variants and generate vcf files. We used R to extract the number of heterozygous site. We restricted our counts to only single nucleotide polymorphisms (SNPs) without ambiguous bases, with a quality score >150, and depth < 100X. The heterozygosity in each genome is calculated as the number of heterozygous sites divided by total size of the assembly.

Variant effect predictions We ran PROVEAN (v1.1.5)(Choi et al., 2012) using: CD-HIT (v4.5.8), BLAST 2.2.30+ and NCBI nr database (downloaded February 2015). PROVEAN is based on evolutionary conservation and it has been validated with known human disease causing mutations and outperforms other prediction tools in experimental evolution studies that mimic the process of gradual accumulation of mutations encountered in nature(Rockah-Shmuel et al., 2015). PROVEAN needs two inputs: the reference protein sequence in fasta format and an amino acid variant file. In order to generate the variant file, the predicted reference and alternative proteins were first aligned using MUSCLE (v3.8.31) and then a custom Python script was used to call substitutions, deletions and insertions following the HGVS (Human Genome Variation Society) recommendations. We curated our predictions to minimize the presence of spurious variants. First, we re-blasted all the predicted proteins for each cormorant species against a database consisting of the initial 14,052 high quality chicken proteins. We removed those instances when a paralog was wrongly assigned as an ortholog of the chicken protein. Second, for each cormorant genome, we aligned 20X of the reads used to generate the assemblies back to the assemblies using bwa (v0.6.2). Then we called variants using samtools (v0.1.18). We searched for sites where 100% of the original reads supported a different nucleotide than the one in the final assembly. We assumed those SNPs corresponded to errors in the assembly. We removed genes that contained assembly errors within their coding regions from the final analyses.

Gene enrichment analysis We used the g:Profiler (Reimand et al., 2011) package for R. For the analysis of function-altering variants in P. harrisi. we set the list of 14,052 chicken proteins used to annotate the genomes as the background list. Our dataset consisted of three homology-based (P. harrisi vs. P. auritus,P. harrisi vs. P. brasilianus, and P. harrisi vs. P. pelagicus) and one transcriptome-based (P. harrisi vs. P. auritus) variant prediction datasets. We defined a set of proteins containing function-altering variants (PROVEAN score < −5) in at least one of our four prediction datasets (1129 proteins of which 1101 have a known ontology). This strategy was conservative and guaranteed that variants derived exclusively from the transcriptome prediction were included in the analysis. The enrichment for human disease categories was derived from annotations from the Human Phenotype Ontology (HPO), a standardized vocabulary of phenotypic abnormalities encountered in human disease. To check for the specificity of enrichments for syndromes affecting limb development inP. harrisi, we performed gene enrichment analysis in the union set of variants predicted to be function-altering in P. auritus, P. brasilianus and P. pelagicus compared to P. harrisi (PROVEAN score > 5) in our four prediction datasets. Categories on this set did not show any limb related syndrome or overlap with the previous categories, (Supplementary Table 4b). As a second specificity control, we repeated our prediction pipeline, but instead ofP. harrisi, we used P. pelagicus as a reference species. Enriched human disease categories in P. pelagicus did not show any overlap with the ones fromP. harrisi. (Supplementary Table 4c). For the enrichment analysis of CUX1 target genes in Hs578t cells (breast cancer cell line), we used the list of all genes contained in the array probes as the background. We applied the g:SCS method for estimating significance and correcting for multiple testing.

Confirmation of function-altering variants We performed Sanger sequencing to confirm the presence of predicted function-altering variants inP. harrisi and its absence in the three closely related cormorants. A list of the primers used can be found in Supplementary Table 10. To determine if a variant was fixed in the Galapagos Cormorant population, we genotyped by Sanger sequencing another 20 different individuals from two of the most divergent populations: Cabo Hammond (10) and Caones Sur (10) (Duffie et al., 2009). Given that the total population of the Galapagos Cormorant is 1,500 individuals, we considered that if 20/20 genotyped individuals were homozygous, then the variant was most likely fixed in the population.

Electrophoretic Mobility Shift Assay (EMSA) We digested the plasmid pMAL-c5X-His (NEB) with NcoI and SbfI-restriction enzymes and performed Gibson assembly to insert gBlocks (IDT) coding for E. coli codon optimized CR3HD CUX1-Ancestral or CUX1-Δ4aa variants. Expression of this plasmid results in a fusion protein containing the Maltose binding protein (MBP) on the N-terminal to increase solubility, the CR3HD domain, and a 5X His tag on the C-terminal for protein purification. Proteins were expressed in E. coli and purified at the Protein Expression Technology Center, UCLA. After protein elution, we performed MWCO 50 kDa dialysis using the Pur-A-Lyzer Maxi 50000 kit (Sigma) and 1X Binding Buffer. Purified protein was quantified using the Bradford Protein Assay Kit (Thermo-Scientific) and kept at −80C. Expected size and protein purity was confirmed running 10 ng on a 10% Bis-Tris polyacrylamide gel under denaturing conditions and performing silver staining with Pierce Silver Stain kit (ThermoFisher). In addition, purified proteins were recognized by MBP (NEB) and (M-222) CUX1 antibodies, as expected (not shown). To perform gel-shifts we used the LightShift Chemiluminescent EMSA kit from Thermo Scientific. The experimental conditions for CR3HD CUX1 binding were adapted from a previous study(Moon et al., 2000). Purified CR3HD CUX1-Ancestral or CUX1-Δ4aa protein was pre-incubated for 5 minutes with 1μg of poly(dI-dC) and 3.3μg of BSA in 25mM NaCl, 10mM Tris, pH 7.5, 5nM EDTA, 5% Glycerol, 1mM DTT, 1mM MgCl2. 20fmol of Biotin end-labeled target dsDNA (TCGAGACGGTATCGATAAGCTTCTTTTC) were added to the mix and incubated at RT for 20 minutes.

Luciferase assay We adapted a previously validated chloramphenicol acetyltransferase (CAT) reporter assay to test the repressor activity of the C-terminal tail of CUX1(Mailly et al., 1996). We generated a luciferase reporter plasmid by cloning a gBlock (IDT) containing 7xUAS binding elements followed by the thymidine kinase (tk) promoter using Gibson assembly into the pGL4.15[luc2P/Hygro] plasmid (Promega) previously digested with XhoI and HindIII restriction enzymes. We then generated GAL4DBD protein fusion plasmids following two steps. First, we cloned a gBlock (IDT) coding for a human codon optimized GAL4 DBD (1:147 aa) into a pcDNA3.1(+) (Thermo-Fisher) mammalian vector previously digested with HindIII and EcoRI using Gibson assembly. We then used this plasmid as a backbone to introduce the CUX1 C-terminal domains containing the repressor activity. We then digested the previously modified a pcDNA3.1 plasmid with XhoI and XbaI, and used gBlocks (IDT) to clone the CUX1 C-terminal domain Ancestral or Δ4aa variants using Gibson assembly. We transfected COS-7 cells growing in 12 well plates under standard conditions with Lipofectamine 3000 (ThermoFisher) using 600ng of GAL4DBD:CUX1-Cterm plasmid, 200ng of 7xUAS/tk:luc2 reporter, and 2ng of pGL4.75[h/Rluc/CMV] as a normalization control. We measured reporter activity 48 hours after transfection. All transfections and measurements were performed in triplicate.

Western blot We incubated commercial mallard eggs (Metzer Farms, CA) and extracted the wings from developing mallard embryo after 22 days of incubation at 37.5C and 55% relative humidity. We incubated the tissue in RIPA lysis buffer supplemented with Halt Protease Inhibitor Cocktail (Thermo Scientific). The lysate was further subjected to mechanical lysis and sonication at 4C. Protein concentration was determined using the Bradford Protein Assay Kit (Thermo-Scientific). 20μg of protein were loaded into a 10% Bis-Tris polyacrylamide gel under denaturing conditions. Sample was transferred to a PVDF membrane and blocked for 2 hours in 5% BSA PBST buffer at RT. The membrane was incubated overnight at 4C with 5% BSA PBST and a 1:1000 dilution of rabbit polyclonal M-222 CUX1 antibody (200 μg/mL) (Santa Cruz Biotechnology). After 3 washes with PBST buffer, the membrane was incubated with a 1:10,000 dilution of Peroxidase-conjugated Goat Anti-Rabbit IgG (H+L) (Jackson ImmunoResearch) in PBST for 1h at RT.

ATDC5 cell line generation and differentiation CR3HD CUX1-Ancestral and CR3HD CUX1-Δ4aa mouse codon optimized genes were cloned into a lentiviral vector to produce N-terminal HIS-tag fusion proteins under a CMV promoter. ATDC5 cells were transduced with lentiviral particles and stable cell lines and were generated using puromycin selection. A control ATDC5 line was generated under identical conditions but using lentiviruses carrying an empty vector. Stables cell lines were generated by Cellomics Technology, MD, USA. Immunostaining against the HIS tag revealed that both variants have strong nuclear localization, indicating that the Galapagos Cormorant variant does not affect the subcellular localization of CUX. Cells were plated at 1.5105 cells/well in 24-well plates, and were cultured using DMEM+10% FBS and 3 μg/mL puromycin. To differentiate cells we cultured cells in chondrogenic differentiation medium:-MEM containing 5%FBS, 200 μg/ml ascorbic acid, 60 nm Na2SeO3, 10 μg/ml transferrin, 1% antibiotic. After 7 and 12 days in culture, cells were lysed and total RNA was isolated using the RNeasy Isolation kit (QIAGEN).

Real time quantitative PCR RNA from ATDC5 cell lines was extracted using the RNeasy Isolation kit (QIAGEN). cDNA was prepared using Superscript III Reverse Transcriptase (ThermoFisher) using Poly-T primers. All gene specific primers were designed to span exon-exon junctions and/or exon-intron boundaries when possible. A list of the primers used can be found in Supplementary Table 11. We validated primers by constructing standard curves and checking for amplification efficiency. All measurements were performed in at least three biological replicates per condition. As a reference gene, we used Hprt, which was previously validated in ATDC5 cells (Zhai et al., 2013). Importantly, Hprt levels were not affected in Hs578t cells overexpressing CUX1 nor expressing CUX1 shRNA and have been previously used to validate Cux1 target genes. The use of another validated reference gene, Ppia, gave identical results to Hprt. All measurements were performed using the SYBR Green I master mix (Roche) in a LightCycler 480 Real-time PCR System (Roche)

Statistical test for positive selection Ortholog coding sequences were obtained from our assemblies and the NCBI database. Nucleotides sequences were aligned using protein alignments as a reference with PAL2NAL (Suyama et al., 2006). Alignments were visually inspected and taxa containing errors were discarded if low quality. We estimated the ratio of non-synonymous to synonymous substitutions (ω = dN/dS) using a maximum likelihood framework and a branch model of adaptive evolution implemented in CODEML (PAML v4.8) (Yang, 2007). We used the likelihood ratio test to compare a null model (M0) that estimates an average for a given gene across all branches in the phylogeny and an alternative two-ratio branch model that estimates one for the branch of interest and another for the background branches. Our phylogeny consisted of 36 taxa: 28 birds (Columba livia, Acanthisitta chlo-ris, Manacus vitellinus, Corvus cornix, Taeniopygia guttata, Nestor notabilis, Calypte anna, Meleagris gallopavo, Gallus gallus, Cuculus canorus, Mesitornis unicolor, Struthio camelus, Anas platyrhynchos, Falco peregrinus, Falco cherrug, Egretta garzetta, Nipponia nippon, Charadrius vociferous, Phalacrocorax pelagicus, Phalacrocorax carbo, Phalacrocorax auritus, Phalacrocorax harrisi, Phalacrocorax brasilianus, Fulmarus glacialis, Aquila chrysaetos, Hali-aeetus leucocephalus, Aptenodytes forsteri, and Pygoscelis adeliae), 6 other reptiles (Python bivittatus, Anolis carolinensis, Gekko japonicus, Chelonia mydas, Alligator sinensis, Alligator mississippiensis) and 2 mammals (Homo sapiens and Mus musculus).

Author contributions

A.B. and L.K. conceived the study. A.B. coordinated the collection of samples, prepared libraries, assembled and annotated genomes, and performed all the analyses and experiments. P.W., A.R., C.V., and P.P. provided DNA or tissue samples. W.W. and A.B. carried out ATDC5 cell line experiments supervised by K.L. AB. and L.K. wrote the manuscript. All authors discussed and agreed on the final version of the manuscript.

Acknowledgement

A.B. is funded by the Jane Coffin Childs Memorial Fund for Medical Research. L.K. is funded by the Howard Hughes Medical Institute (HHMI). A.R. is funded by the U.S. Geological Survey through the Wildlife Program of the Ecosystems Mission area. Any use of trade names is for descriptive purposes only and does not imply endorsement by the U.S. Government. C.V. is funded by FONDECYT (11130305). We thank Christopher Koo and Dr. Mark Arb-ing at the UCLA-DOE Protein Expression Laboratory Core Facility for protein purification and Dr. Suhua Feng at the Broad Stem Cell Research Center High Throughput Sequencing Core for assistance. We thank Dr. Carlos Valle (USFQ) for advice. We thank Dr. Kimball Garrett (Natural History Museum of Los Angeles), Dr. Kevin Burns (San Diego Natural History Museum and SDSU Museum of Biodiversity), and Dr. Rebecca Duerr (International Bird Rescue) for providing samples used in preliminary stages of this study.

Footnotes

  • ↵* E-mail: aburga{at}mednet.ucla.edu; lkruglyak{at}mednet.ucla.edu

References

  1. ↵
    Baker, Allan J., Oliver Haddrath, John D. Mcpherson, and Alison Cloutier. 2014. ‘Genomic Support for a Moa-Tinamou Clade and Adaptive Morphological Convergence in Flightless Ratites’. Molecular Biology and Evolution 31 (7): 1686–96. doi:10.1093/molbev/msu153.
    OpenUrlCrossRefPubMedWeb of Science
  2. ↵
    Bimonte, Sabrina, Amalia De Angelis, Luca Quagliata, Fabiola Giusti, Roberta Tammaro, Romano Dallai, Maria-Grazia Ascenzi, Graciana Diez-Roux, and Brunella Franco. 2011. ‘Ofd1 Is Required in Limb Bud Patterning and Endochondral Bone Development’. Developmental Biology 349 (2). Elsevier Inc.: 179–91. doi:10.1016/j.ydbio.2010.09.020.
    OpenUrlCrossRefPubMed
  3. ↵
    Bolger, Anthony M., Marc Lohse, and Bjoern Usadel. 2014. ‘Trimmomatic: A Flexible Trimmer for Illumina Sequence Data’. Bioinformatics 30 (15): 2114–20. doi:10.1093/bioinformatics/btu170.
    OpenUrlCrossRefPubMedWeb of Science
  4. ↵
    Bunn, Kieran J., Phil Daniel, Heleen S. Rosken, Adam C. O’Neill, Sophia R. Cameron-Christie, Tim Morgan, Han G. Brunner, et al. 2015. ‘Mutations in DVL1 Cause an Osteosclerotic Form of Robinow Syndrome’. The American Journal of Human Genetics 96 (4): 623–30. doi:10.1016/j.ajhg.2015.02.010.
    OpenUrlCrossRefPubMed
  5. ↵
    Cadieux, Chantal, Ryoko Harada, Marilene Paquet, Olivier Cote, Marie Trudel, Alain Nepveu, and Maxime Bouchard. 2008. ‘Polycystic Kidneys Caused by Sustained Expression of Cux1 Isoform p75’. Journal of Biological Chemistry 283 (20): 13817–24. doi:10.1074/jbc.M709332200.
    OpenUrlAbstract/FREE Full Text
  6. ↵
    1. Gregory A. Cox
    Caruso, Nathalie, Balazs Herberth, Marc Bartoli, Francesca Puppo, Julie Dumonceaux, Angela Zimmermann, Simon Denadai, et al. 2013. ‘Deregulation of the Protocadherin Gene FAT1 Alters Muscle Shapes: Implications for the Pathogenesis of Facioscapulohumeral Dystrophy’. Edited by Gregory A. Cox. PLoS Genetics 9 (6): e1003550. doi:10.1371/journal.pgen.1003550.
    OpenUrlCrossRefPubMed
  7. ↵
    Choi, Yongwook, Gregory E. Sims, Sean Murphy, Jason R. Miller, and Agnes P. Chan. 2012. ‘Predicting the Functional Effect of Amino Acid Substitutions and Indels’. PLoS ONE 7 (10). doi:10.1371/journal.pone.0046688.
    OpenUrlCrossRefPubMed
  8. ↵
    Colnot, C., S.S. Sidhu, N. Balmain, and F. Poirier. 2001. ‘Uncoupling of Chondrocyte Death and Vascular Invasion in Mouse Galectin 3 Null Mutant Bones’. Developmental Biology 229 (1): 203–14. doi:10.1006/dbio.2000.9933.
    OpenUrlCrossRefPubMedWeb of Science
  9. Consortium, International Chicken Genome Sequencing. 2004. ‘Sequence and Comparative Analysis of the Chicken Genome Provide Unique Perspectives on Vertebrate Evolution.’ Nature 432 (7018): 695–716. doi:10.1038/nature03394.
    OpenUrlCrossRefPubMedWeb of Science
  10. ↵
    Darwin, Charles. 1859. On the Origin of Species by Means of Natural Selection or the Preservation of Favoured Races in the Struggle for Life. London: John Murray.
  11. ↵
    Duffie, Caroline V., Travis C. Glenn, F. Hernan Vargas, and Patricia G. Parker. 2009. ‘Genetic Structure within and between Island Populations of the Flightless Cormorant (Phalacrocorax Harrisi)’. Molecular Ecology 18 (10): 2103–11. doi:10.1111/j.1365-294X.2009.04179.x.
    OpenUrlCrossRefPubMedWeb of Science
  12. ↵
    Dunn, Erica H. 1975. ‘Growth, Body Components and Energy Content of Nestling Double-Crested Cormorants’. The Condor 77 (4): 431. doi:10.2307/1366090.
    OpenUrlCrossRefWeb of Science
  13. ↵
    Edgar, Robert C. 2004. ‘MUSCLE: Multiple Sequence Alignment with High Accuracy and High Throughput’. Nucleic Acids Research 32 (5): 1792–97. doi:10.1093/nar/gkh340.
    OpenUrlCrossRefPubMedWeb of Science
  14. ↵
    Ellis, Tammy, Laure Gambardella, Markus Horcher, Stefan Tschanz, Janine Capol, Paula Bertram, Wolfram Jochum, Yann Barrandon, and Meinrad Busslinger. 2001. ‘The Transcriptional Repressor CDP (Cutl1) Is Essential for Epithelial Cell Differentiation of the Lung and the Hair Follicle’. Genes and Development 15 (17): 2307–19. doi:10.1101/gad.200101.
    OpenUrlAbstract/FREE Full Text
  15. ↵
    Fave, Marie-Julie, Robert A. Johnson, Stefan Cover, Stephan Handschuh, Brian D. Metscher, Gerd B. Muller, Shyamalika Gopalan, and Ehab Abouheif. 2015. ‘Past Climate Change on Sky Islands Drives Novelty in a Core Developmental Gene Network and Its Phenotype’. BMC Evolutionary Biology 15 (1). BMC Evolutionary Biology: 183. doi:10.1186/s12862-015-0448-4.
    OpenUrlCrossRefPubMed
  16. ↵
    Feduccia, Alan. 1999. The Origin and Evolution of Birds. 2nd edition. New Haven: Yale University Press.
  17. ↵
    Ganapathy, Ganeshkumar, Jason T Howard, James M Ward, Jianwen Li, Bo Li, Yingrui Li, Yingqi Xiong, et al. 2014. ‘High-Coverage Sequencing and Annotated Assemblies of the Budgerigar Genome’. GigaScience 3 (1): 11. doi:10.1186/2047-217X-3-11.
    OpenUrlCrossRefPubMed
  18. ↵
    Goetz, Sarah C., and Kathryn V. Anderson. 2010. ‘The Primary Cilium: A Signalling Centre during Vertebrate Development’. Nature Reviews Genetics 11 (5). Nature Publishing Group: 331–44. doi:10.1038/nrg2774.
    OpenUrlCrossRefPubMedWeb of Science
  19. ↵
    Gould, Stephen Jay. 1977. Ontogeny and Phylogeny. Cambridge, Massachusetts: The Belknap Press of Harvard Universiry Press.
  20. ↵
    Grabherr, Manfred G, Brian J Haas, Moran Yassour, Joshua Z Levin, Dawn A Thompson, Ido Amit, Xian Adiconis, et al. 2011. ‘Full-Length Transcriptome Assembly from RNA-Seq Data without a Reference Genome’. Nature Biotechnology 29 (7): 644–52. doi:10.1038/nbt.1883.
    OpenUrlCrossRefPubMed
  21. ↵
    Gray, Ryan S, Philip B Abitua, Bogdan J Wlodarczyk, Heather L Szabo-Rogers, Otis Blanchard, Insuk Lee, Greg S Weiss, et al. 2009. ‘The Planar Cell Polarity Effector Fuz Is Essential for Targeted Membrane Trafficking, Ciliogenesis and Mouse Embryonic Development.’ Nature Cell Biology 11 (10). Nature Publishing Group: 1225–32. doi:10.1038/ncb1966.
    OpenUrlCrossRefPubMedWeb of Science
  22. ↵
    Green, Richard E, Edward L Braun, Joel Armstrong, Dent Earl, Ngan Nguyen, Glenn Hickey, Michael W Vandewege, et al. 2014. ‘Three Crocodilian Genomes Reveal Ancestral Patterns of Evolution among Archosaurs’. Science 346 (6215): 1355 –. doi:10.1126/science.1254449.
    OpenUrlCrossRef
  23. ↵
    1. Harpp, Karen S,
    2. Eric Mittelstaedt,
    3. Noemi D’Ozouville, and
    4. David W. Graham, eds.
    2014. ‘The Galapagos’. In The Galapagos: A Natural Laboratory for the Earth Sciences, 145–66. Geophysical Monograph Series. Hoboken, New Jersey: John Wiley & Sons, Inc.
  24. ↵
    Harshman, John, Edward L Braun, Michael J Braun, Christopher J Huddleston, Rauri C K Bowie, Jena L Chojnowski, Shannon J Hackett, et al. 2008. ‘Phylogenomic Evidence for Multiple Losses of Flight in Ratite Birds.’ Proceedings of the National Academy of Sciences of the United States of America 105 (36): 13462–67. doi:10.1073/pnas.0803242105.
    OpenUrlAbstract/FREE Full Text
  25. ↵
    Haycraft, Courtney J, Boglarka Banizs, Yesim Aydin-Son, Qihong Zhang, Edward J Michaud, and Bradley K Yoder. 2005. ‘Gli2 and Gli3 Localize to Cilia and Require the Intraflagellar Transport Protein Polaris for Processing and Function.’ PLoS Genetics 1 (4): e53. doi:10.1371/journal.pgen.0010053.
    OpenUrlCrossRefPubMed
  26. ↵
    Iyaguchi, Daisuke, Min Yao, Nobuhisa Watanabe, Jun Nishihira, and Isao Tanaka. 2007. ‘DNA Recognition Mechanism of the ONECUT Homeodomain of Transcription Factor HNF-6’. Structure 15 (1): 75–83. doi:10.1016/j.str.2006.11.004.
    OpenUrlCrossRefPubMed
  27. ↵
    Jarvis, E. D., S. Mirarab, A. J. Aberer, B. Li, P. Houde, C. Li, S. Y. W. Ho, et al. 2014. ‘Whole-Genome Analyses Resolve Early Branches in the Tree of Life of Modern Birds’. Science 346 (6215): 1320–31. doi:10.1126/science.1253451.
    OpenUrlAbstract/FREE Full Text
  28. ↵
    Kennedy, Martyn, Carlos a. Valle, and Hamish G. Spencer. 2009. ‘The Phylogenetic Position of the Galapagos Cormorant’. Molecular Phylogenetics and Evolution 53 (1). Elsevier Inc.: 94–98. doi:10.1016/j.ympev.2009.06.002.
    OpenUrlCrossRefPubMed
  29. ↵
    Koch, Annett, Francoise Poirier, Ralf Jacob, and Delphine Delacour. 2010. ‘Galectin-3, a Novel Centrosome-Associated Protein, Required for Epithelial Morphogenesis.’ Molecular Biology of the Cell 21 (2): 219–31. doi:10.1091/mbc.E09-03-0193.
    OpenUrlAbstract/FREE Full Text
  30. ↵
    Kronenberg, Henry M. 2003. ‘Developmental Regulation of the Growth Plate.’ Nature 423 (6937): 332–36. doi:10.1038/nature01657.
    OpenUrlCrossRefPubMedWeb of Science
  31. ↵
    Leggett, Richard M., Bernardo J. Clavijo, Leah Clissold, Matthew D. Clark, and Mario Caccamo. 2014. ‘Next Clip: An Analysis and Read Preparation Tool for Nextera Long Mate Pair Libraries’. Bioinformatics 30 (4): 566–68. doi:10.1093/bioinformatics/btt702.
    OpenUrlCrossRefPubMed
  32. ↵
    Li, Heng, and Richard Durbin. 2009. ‘Fast and Accurate Short Read Alignment with Burrows-Wheeler Transform’. Bioinformatics 25 (14): 1754–60. doi:10.1093/bioinformatics/btp324.
    OpenUrlCrossRefPubMedWeb of Science
  33. ↵
    Li, Heng, Bob Handsaker, Alec Wysoker, Tim Fennell, Jue Ruan, Nils Homer, Gabor Marth, Goncalo Abecasis, and Richard Durbin. 2009. ‘The Sequence Alignment/Map Format and SAMtools’. Bioinformatics 25 (16): 2078–79. doi: 10.1093/bioinformatics/btp352.
    OpenUrlCrossRefPubMedWeb of Science
  34. ↵
    Livezey, Bradley C. 1992. ‘Flightlessness in the Galapagos Cormorant (Compsohalieus [Nannopterum] Harrisi): Heterochrony, Gigantism, and Specialization’. Zoological Journal of the Linnean Society 105: 155–224. doi: 10.1111/j.1096-3642.1992.tb01229.x.
    OpenUrlCrossRef
  35. ↵
    Luo, Ruibang, Binghang Liu, Yinlong Xie, Zhenyu Li, Weihua Huang, Jianying Yuan, Guangzhu He, et al. 2012. ‘SOAPdenovo2: An Empirically Improved Memory-Efficient Short-Read de Novo Assembler.’ GigaScience 1 (1): 18. doi:10.1186/2047-217X-1-18.
    OpenUrlCrossRefPubMed
  36. ↵
    Mailly, F, G Berube, R Harada, P L Mao, S Phillips, and A Nepveu. 1996. ‘The Human Cut Homeodomain Protein Can Repress Gene Expression by Two Distinct Mechanisms: Active Repression and Competition for Binding Site Occupancy.’ Molecular and Cellular Biology 16 (10): 5346–57.
    OpenUrlAbstract/FREE Full Text
  37. ↵
    Matsuoka, Shuhei, Bryan A Ballif, Agata Smogorzewska, E Robert McDonald, Kristen E Hurov, Ji Luo, Corey E Bakalarski, et al. 2007. ‘ATM and ATR Substrate Analysis Reveals Extensive Protein Networks Responsive to DNA Damage.’ Science 316 (May): 1160–66. doi:10.1126/science.1140321.
    OpenUrlAbstract/FREE Full Text
  38. ↵
    Mitchell, Kieren J, Bastien Llamas, Julien Soubrier, Nicolas J Rawlence, Trevor H Worthy, Jamie Wood, Michael S Y Lee, and Alan Cooper. 2014. ‘Ancient DNA Reveals Elephant Birds and Kiwi Are Sister Taxa and Clarifies Ratite Bird Evolution.’ Science 344 (6186): 898–900. doi:10.1126/science.1251981.
    OpenUrlAbstract/FREE Full Text
  39. ↵
    Moon, Nam Sung, Ginette Berube, and Alain Nepveu. 2000. ‘CCAAT Displacement Activity Involves CUT Repeats 1 and 2, Not the CUT Homeodomain’. Journal of Biological Chemistry 275 (40): 31325–34. doi:10.1074/jbc.M002912200.
    OpenUrlAbstract/FREE Full Text
  40. ↵
    Nishio, Hitomi, and Martin J Walsh. 2004. ‘CCAAT Displacement Protein/cut Homolog Recruits G9a Histone Lysine Methyltransferase to Repress Transcription.’ Proceedings of the National Academy of Sciences of the United States of America 101 (31): 11257–62. doi:10.1073/pnas.0401343101.
    OpenUrlAbstract/FREE Full Text
  41. ↵
    Ocbina, Polloneal Jymmiel R, Jonathan T Eggenschwiler, Ivan Moskowitz, and Kathryn V Anderson. 2011. ‘Complex Interactions between Genes Controlling Trafficking in Primary Cilia.’ Nature Genetics 43 (6). Nature Publishing Group: 547–53. doi:10.1038/ng.832.
    OpenUrlCrossRefPubMed
  42. ↵
    Olson, Storrs L. 1973. ‘Evolution of the Rails of the South Atlantic Islands (Aves: Rallidae)’. Smithsonian Contributions to Zoology 152: 1–53. doi:10.5479/si.00810282.152.
    OpenUrlCrossRef
  43. ↵
    Paces-Fessy, Melanie, Dominique Boucher, Emile Petit, Sandrine Paute-Briand, and Marie-Franoise Blanchet-Tournier. 2004. ‘The Negative Regulator of Gli, Suppressor of Fused (Sufu), Interacts with SAP18, Galectin3 and Other Nuclear Proteins.’ The Biochemical Journal 378 (Pt 2): 353–62. doi:10.1042/BJ20030786.
    OpenUrlCrossRefPubMedWeb of Science
  44. ↵
    Park, Tae Joo, Saori L Haigo, and John B Wallingford. 2006. ‘Ciliogenesis Defects in Embryos Lacking Inturned or Fuzzy Function Are Associated with Failure of Planar Cell Polarity and Hedgehog Signaling.’ Nature Genetics 38 (3): 303–11. doi:10.1038/ng1753.
    OpenUrlCrossRefPubMedWeb of Science
  45. ↵
    Parra, Genis, Keith Bradnam, Zemin Ning, Thomas Keane, and Ian Korf. 2009. ‘Assessing the Gene Space in Draft Genomes’. Nucleic Acids Research 37 (1): 289–97. doi:10.1093/nar/gkn916.
    OpenUrlCrossRefPubMedWeb of Science
  46. ↵
    Powell, Author D C, R J Aulerich, R J Balander, K L Stromborg, and S J Bursian. 1998. ‘A Photographic Guide to the Development of Double-Crested Cormorant Embryos.’ Colonial Waterbirds 21 (3): 348–55. doi:10.2307/1521647.
    OpenUrlCrossRef
  47. ↵
    Ramdzan, Zubaidah M., and Alain Nepveu. 2014. ‘CUX1, a Haploinsufficient Tumour Suppressor Gene Overexpressed in Advanced Cancers’. Nature Reviews Cancer 14 (10). Nature Publishing Group: 673–82. doi:10.1038/nrc3805.
    OpenUrlCrossRefPubMed
  48. ↵
    Reimand, J., T. Arak, and J. Vilo. 2011. ‘g:Profiler––a Web Server for Functional Interpretation of Gene Lists (2011 Update)’. Nucleic Acids Research 39 (suppl): W307–15. doi:10.1093/nar/gkr378.
    OpenUrlCrossRefPubMedWeb of Science
  49. ↵
    Rockah-Shmuel, Liat, Agnes Toth-Petroczy, and Dan S. Tawfik. 2015. ‘Systematic Mapping of Protein Mutational Space by Prolonged Drift Reveals the Deleterious Effects of Seemingly Neutral Mutations’. PLOS Computational Biology 11 (8): e1004421. doi: 10.1371/journal.pcbi.1004421.
    OpenUrlCrossRef
  50. ↵
    Roff, Derek A. 1994. ‘The Evolution of Flightlessness: Is History Important?’ Evolutionary Ecology 8 (6): 639–57. doi:10.1007/BF01237847.
    OpenUrlCrossRef
  51. ↵
    Ronquist, Fredrik, Maxim Teslenko, Paul Van Der Mark, Daniel L. Ayres, Aaron Darling, Sebastian Hohna, Bret Larget, Liang Liu, Marc A. Suchard, and John P. Huelsenbeck. 2012. ‘Mrbayes 3.2: Efficient Bayesian Phylogenetic Inference and Model Choice across a Large Model Space’. Systematic Biology 61 (3): 539–42. doi:10.1093/sysbio/sys029.
    OpenUrlCrossRefPubMed
  52. ↵
    Sansregret, Laurent, and Alain Nepveu. 2008. ‘The Multiple Roles of CUX1: Insights from Mouse Models and Cell-Based Assays’. Gene 412 (1-2): 84–94. doi:10.1016/j.gene.2008.01.017.
    OpenUrlCrossRefPubMedWeb of Science
  53. ↵
    She, Rong, Jeffrey S C Chu, Ke Wang, Jian Pei, and Nansheng Chen. 2009. ‘genBlastA: Enabling BLAST to Identify Homologous Gene Sequences’. Genome Research 19 (1): 143–49. doi:10.1101/gr.082081.108.
    OpenUrlAbstract/FREE Full Text
  54. ↵
    Shukunami, Chisa, Chohei Shigeno, Tadao Atsumi, Kiyoto Ishizeki, Fujio Suzuki, and Yuji Hiraki. 1996. ‘Chondrogenic Differentiation of Clonal Mouse Embryonic Cell Line ATDC5 in Vitro: Differentiation-Dependent Gene Expression of Parathyroid Hormone (PTH)/PTH-Related Peptide Receptor’. Journal of Cell Biology 133 (2): 457–68. doi:10.1083/jcb.133.2.457.
    OpenUrlAbstract/FREE Full Text
  55. ↵
    Sinclair, Angus M., Jamie a. Lee, Adrian Goldstein, Dongxia Xing, Shengxi Liu, Ruzeng Ju, Philip W. Tucker, Ellis J. Neufeld, and Richard H. Scheuermann. 2001. ‘Lymphoid Apoptosis and Myeloid Hyperplasia in CCAAT Displacement Protein Mutant Mice’. Blood 98 (13): 3658–67. doi:10.1182/blood.V98.13.3658.
    OpenUrlAbstract/FREE Full Text
  56. ↵
    Slack, Kerryn E., Craig M. Jones, Tatsuro Ando, G. L. Harrison, R. Ewan Fordyce, Ulfur Arnason, and David Penny. 2006. ‘Early Penguin Fossils, plus Mitochondrial Genomes, Calibrate Avian Evolution’. Molecular Biology and Evolution 23 (6): 1144–55. doi:10.1093/molbev/msj124.
    OpenUrlCrossRefGeoRefPubMedWeb of Science
  57. ↵
    Slater, Guy St C, and Ewan Birney. 2005. ‘Automated Generation of Heuristics for Biological Sequence Comparison.’ BMC Bioinformatics 6 (1): 31. doi: 10.1186/1471-2105-6-31.
    OpenUrlCrossRefPubMed
  58. ↵
    Stratigopoulos, G., C. A. LeDuc, M. L. Cremona, W. K. Chung, and R. L. Leibel. 2011. ‘Cut-like Homeobox 1 (CUX1) Regulates Expression of the Fat Mass and Obesity-Associated and Retinitis Pigmentosa GTPase Regulator-Interacting Protein-1-like (RPGRIP1L) Genes and Coordinates Leptin Receptor Signaling’. Journal of Biological Chemistry 286 (3): 2155–70. doi:10.1074/jbc.M110.188482.
    OpenUrlAbstract/FREE Full Text
  59. ↵
    Sudarsan, Vikram, Sumita Anant, Preeta Guptan, K. Vijayraghavan, and Helen Skaer. 2001. ‘Myoblast Diversification and Ectodermal Signaling in Drosophila’. Developmental Cell 1 (6): 829–39. doi:10.1016/S1534-5807(01)00089-2.
    OpenUrlCrossRefPubMed
  60. ↵
    Suyama, Mikita, David Torrents, and Peer Bork. 2006. ‘PAL2NAL: Robust Conversion of Protein Sequence Alignments into the Corresponding Codon Alignments’. Nucleic Acids Research 34 (WEB. SERV. ISS.): 609–12. doi:10.1093/nar/gkl315.
    OpenUrlCrossRef
  61. ↵
    Tavares, a T, T Tsukui, and J C Izpisua Belmonte. 2000. ‘Evidence That Members of the Cut/Cux/CDP Family May Be Involved in AER Positioning and Polarizing Activity during Chick Limb Development.’ Development 127 (23): 5133–44.
    OpenUrlAbstract
  62. ↵
    Truscott, Mary, Jean Bernard Denault, Brigitte Goulet, Lam Leduy, Guy S. Salvesen, and Alain Nepveu. 2007. ‘Carboxyl-Terminal Proteolytic Processing of CUX1 by a Caspase Enables Transcriptional Activation in Proliferating Cells’. Journal of Biological Chemistry 282 (41): 30216–26. doi:10.1074/jbc.M702328200.
    OpenUrlAbstract/FREE Full Text
  63. ↵
    Vadnais, Charles, Arif A Awan, Ryoko Harada, Pier-Luc Clermont, Lam Leduy, Ginette Berube, and Alain Nepveu. 2013. ‘Long-Range Transcriptional Regulation by the p110 CUX1 Homeodomain Protein on the ENCODE Array.’ BMC Genomics 14: 258. doi:10.1186/1471-2164-14-258.
    OpenUrlCrossRefPubMed
  64. ↵
    Valle, Carlos A, and Malcolm C MC Coulter. 1987. ‘Present Status of the Flightless Cormorant, Galapagos Penguin and Greater Flamingo Populations in the Galapagos Islands, Ecuador, after the 1982-83 El Nino’. The Condor 89 (January 1986): 276–81. doi:10.2307/1368480.
    OpenUrlCrossRef
  65. ↵
    Wang, Guiming, and Xinhai Li. 2008. ‘Population Dynamics and Recovery of Endangered Crested Ibis (Nipponia Nippon) in Central China’. Waterbirds. doi:10.1675/1524-4695-31.3.489.
    OpenUrlCrossRef
  66. ↵
    Warren, Wesley C, David F Clayton, Hans Ellegren, Arthur P Arnold, LaDeana W Hillier, Axel Kunstner, Steve Searle, et al. 2010. ‘The Genome of a Songbird’. Nature 464 (7289): 757–62. doi:10.1038/nature08819.
    OpenUrlCrossRefPubMedWeb of Science
  67. ↵
    Watanabe, Yuuki Y, Akinori Takahashi, Katsufumi Sato, Morgane Viviant, and Charles-Andre Bost. 2011. ‘Poor Flight Performance in Deep-Diving Cormorants.’ The Journal of Experimental Biology 214 (Pt 3): 412–21. doi:10.1242/jeb.050161.
    OpenUrlAbstract/FREE Full Text
  68. ↵
    White, Janson, Juliana F. Mazzeu, Alexander Hoischen, Shalini N. Jhangiani, Tomasz Gambin, Michele Calijorne Alcino, Samantha Penney, et al. 2015. ‘DVL1 Frameshift Mutations Clustering in the Penultimate Exon Cause Autosomal-Dominant Robinow Syndrome’. The American Journal of Human Genetics 96 (4): 612–22. doi:10.1016/j.ajhg.2015.02.015.
    OpenUrlCrossRefPubMed
  69. ↵
    Wright, Natalie A., David W. Steadman, and Christopher C. Witt. 2016. ‘Predictable Evolution toward Flightlessness in Volant Island Birds’. Proceedings of the National Academy of Sciences 113 (17): 201522931. doi:10.1073/pnas.1522931113.
    OpenUrlAbstract/FREE Full Text
  70. ↵
    Yang, Ziheng. 1997. ‘Likelihood Ratio Tests for Detecting Positive Selection and Application to Primate Lysozyme Evolution’, 568–73.
  71. ↵
    Yang, Ziheng. 2007. ‘PAML 4: Phylogenetic Analysis by Maximum Likelihood’. Molecular Biology and Evolution 24 (8): 1586–91. doi:10.1093/molbev/msm088.
    OpenUrlCrossRefPubMedWeb of Science
  72. ↵
    Zeng, Huiqing, Amber N Hoover, and Aimin Liu. 2010. ‘PCP Effector Gene Inturned Is an Important Regulator of Cilia Formation and Embryonic Development in Mammals.’ Developmental Biology 339 (2). Elsevier Inc.: 418–28. doi:10.1016/j.ydbio.2010.01.003.
    OpenUrlCrossRefPubMedWeb of Science
  73. ↵
    Zhai, Zhichen, Yongchang Yao, and Yingjun Wang. 2013. ‘Importance of Suitable Reference Gene Selection for Quantitative RT-PCR during ATDC5 Cells Chondrocyte Differentiation’. PLoS ONE 8 (5). doi:10.1371/journal.pone.0064786.
    OpenUrlCrossRefPubMed
Back to top
Posted July 02, 2016.
Download PDF

Supplementary Material

Email
Loss of flight in the Galapagos cormorant mirrors human skeletal ciliopathies
Alejandro Burga, Weiguang Wang, Paul C. Wolf, Andrew M. Ramey, Claudio Verdugo, Karen Lyons, Patricia G. Parker, Leonid Kruglyak
bioRxiv 061432; doi: https://doi.org/10.1101/061432
Reddit logo Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area