University of Florida Thesis Or Dissertation Formatting

DECONSTRUCTING INDEPENDENT EVOLUTION: A DEVELOPMENTAL APPROACH TO THE EVOLUTION OF ANIMAL CARTILAGE AND APPENDAGES

OSCAR ALEJANDRO TARAZONA REY

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2015

To Francisca

ACKNOWLEDGMENTS

I thank Francisca Leal for her support and understanding during the development of this project. I want to thanks to my advisor Martin J. Cohn for everything he did to make me a better scientist, for his constant support, constructive criticism and for giving me the confidence to become a critical scientist. I also thank Barbara Battelle and the H.

J. Brockmann and D. Julian labs (University of Florida) for Limulus eggs, Nadean Brown

(Cincinnati Children's Research Foundation) for sharing protocols and reagents,

GuangJun Zhang (Purdue University Vet School) for sharing data, and Monica Welten for assisting me with OPT scanning and 3D reconstructions. I thank HHMI International

Student Research Fellow for giving me the independence to develop this project.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 7

LIST OF FIGURES ...... 8

LIST OF ABBREVIATIONS ...... 10

ABSTRACT ...... 11

CHAPTER

1 INTRODUCTION ...... 13

2 MATERIALS AND METHODS ...... 16

Embryo Collection and Preparation ...... 16 Alcian Blue and Masson’s Trichrome Stain ...... 16 Gene Cloning and RACE-PCR ...... 17 In situ Hybridization and Immunohistochemistry ...... 18 In situ Hybridization and Phalloidin Staining ...... 18 Optical Projection Tomography (OPT) of Collagen in situ Hybridization...... 19 Molecular Phylogenetic Analysis of Collagen and Sox Genes...... 19 Treatments With Small Molecule Inhibitors ...... 19 BrdU Labeling and BrdU Pulse Chase ...... 20 Luciferase Assay ...... 21

3 CARTILAGE DEVELOPMENT IN TWO PROTOSTOME LINEAGES ...... 26

Cartilage in Metazoans ...... 26 Homology of Cartilage in Metazoans ...... 29 Origin of New Cell Types ...... 29 Cell Identity and Gene Regulatory Networks ...... 30 Histology and Molecular Nature of the ECM ...... 31 Invertebrate Cartilage is Collagen Based and Contains Hyaluronan ...... 32 Expression of Developmental Genes in Sepia chondrogenesis ...... 34 Expression of Developmental Genes in Limulus chondrogenesis ...... 35 Invertebrate SoxE proteins can activate vertebrate Col2a1 cartilage enhancer ...... 36 Late Chondrogenesis and Cartilage Appositional Growth...... 37 Hedgehog and β-catenin signaling pathways have antagonistic functions in Sepia chondrogenesis ...... 38 Invertebrate and Vertebrate Chondrogenesis Share a Gene Regulatory Network . 39 Origin of Cartilage an the Chondrocyte in Bilateria ...... 40 Hypothesis 1 ...... 41

Hypothesis 2 ...... 41 Hypothesis 3 ...... 42

4 EVOLUTION OF CEPHALOPOD APPENDAGES ...... 61

Cephalopod Arms and Tentacles as Molluscan Evolutionary Novelties ...... 61 Deep Homology of Appendage Gene Regulatory Network ...... 62 Description of Limb Morphogenesis in Cuttlefishes ...... 64 Gene Expression and the Control of Axis of Growth in Cuttlefish Limb ...... 65 Anteroposterior Axis ...... 65 Proximodistal Axis ...... 66 Dorsoventral Axis ...... 68 Developmental Basis of a Morphological Novelty ...... 69 A Conserved Appendage Developmental Program ...... 70

5 LIMB RE-EVOLUTION IN SNAKE-LIKE LIZARDS ...... 84

Phylogenetic Patterns of Limb Loss in Squamate Lizards ...... 84 Limb Loss and Re-evolution in Bachia ...... 85 Limb Morphology of Bachia bicolor ...... 88 Embryonic Development of B. bicolor Hindlimbs ...... 89 B. bicolor Has an Autopod ...... 89 Phylogenetic Patterns Hide Ontogenetic Processes ...... 91

6 CONCLUSION ...... 102

LIST OF REFERENCES ...... 104

BIOGRAPHICAL SKETCH ...... 114

LIST OF TABLES

Table page

2-1 Sequence identifiers for the collagen phylogenetic analysis...... 22

2-2 Sequence identifiers for the phylogenetic analysis of Sox genes...... 23

2-3 Sequence identifiers for the phylogenetic analysis of HAS genes...... 24

2-4 S. officinalis embryos used for each drug treatment and DMSO controls...... 25

LIST OF FIGURES

Figure page

3-1 Animal phylogeny depicting the independent evolution of cartilage in the three major lineages of Bilateria ...... 44

3-2 Developmental series of chondrogenesis in Sepia and Limulus...... 45

3-3 Molecular phylogenetic analysis of fibrillar collagen, Sox transcription factors (SoxC, SoxD, SoxE and SoxF) and Hyaluronan synthases in bilateria ...... 46

3-4 Protostome invertebrate cartilage is structurally similar to vertebrate cartilage, is ColA-based, and contains hyaluronan...... 48

3-5 ColAa and ColAb show the same pattern of gene expression in Sepia embryos...... 49

3-6 Deep conservation of gene expression, from induction of chondrogenesis to transcriptional regulation and secretion of a cartilage ECM, during protostome cartilage development ...... 50

3-7 Chondrogenesis of multiple cartilages occurs near Hedgehog-expressing tissues in Sepia ...... 51

3-8 Patterns of gene expression in developing funnel cartilage of Sepia at stage 25 ...... 53

3-9 Cuttlefish chondrogenesis is regulated positively by Hh signaling and negatively by β-catenin ...... 54

3-10 Gill cartilage in Limulus is collagen-based and expresses SoxE during chondrogenesis ...... 56

3-11 Luciferase reporter assay testing the human Col2a1 cartilage specific enhancer with vertebrate and invertebrate SoxE proteins ...... 57

3-12 Cell proliferation during late chondrogenesis in Sepia ...... 58

3-13 Bright field micrographs and immunofluorescence of Sepia embryos before and after treatments with the small molecule inhibitors...... 59

3-14 Upregulation and downregulation of β-catenin signaling has opposing effects on Sepia chondrogenesis ...... 60

4-1 General morphology of S. officinalis, rendered by OPT imaging...... 73

4-2 Bright field images of Sepia officinalis embryos at different stages of limb development ...... 74

4-3 Embryonic gene expression of Hh during limb development ...... 75

4-4 Embryonic gene expression of Wnt5, Wnt1 and Wnt7 during limb development...... 76

4-5 Embryonic gene expression of Fzd, Lrp and Tcf during limb development ...... 77

4-6 Embryonic gene expression of Dll during limb development ...... 78

4-7 Embryonic gene expression of Dac and Sp8 during limb development ...... 79

4-8 Embryonic gene expression of Hth during limb development ...... 80

4-9 Embryonic gene expression of Bmp2/4 during limb development ...... 81

4-10 Embryonic gene expression of the Wnt inhibitor Sfrp during limb development ...... 82

4-11 Summary of pattern of gene expression along the three axes of growth in cephalopod limb buds...... 83

5-1 Phylogenetic analysis of genus Bachia ...... 96

5-2 Morphology of forelimbs and hindlimbs of Bachia bicolor ...... 97

5-3 Hindlimb development of B. bicolor as seen by SEM imaging in an embryonic series ...... 98

5-4 Embryological evidence of posterior AER loss and a transitory digital plate in B. bicolor ontogeny ...... 99

5-5 Autopod development in forelimbs and hindlimbs of B. bicolor...... 100

5-6 A hypothetical scenario proposed for limb re-evolution in Bachia...... 101

LIST OF ABBREVIATIONS

BrdU BromodeoxyUridine

BSA Bovine Serum Albumin

ColA Clade A fibrilar collagen.

DMSO Dimethyl sulfoxide

ECM Extracellular matrix.

GAG Glycosaminoglycan.

GSK3 Glycogen synthase kinase 3.

ISH In situ hybridization

OPT Optical projection tomography

PBS Phosphate buffered saline

PCNA Proliferating nuclear antigen

Sox SRY-related HMG box transcription factor.

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

DECONSTRUCTING INDEPENDENT EVOLUTION: A DEVELOPMENTAL APPROACH TO THE EVOLUTION OF ANIMAL CARTILAGE AND APPENDAGES

Oscar Alejandro Tarazona Rey

August 2015

Chair: Martin J. Cohn Major: Zoology

Similarity in non-homologous characters is an unexpected outcome of evolution, and the ultimate and proximate causes of the independent evolution of similar characters are diverse. Similar but independently acquired traits are artificially classified, according to their phylogenetic distribution and generative processes, as convergences, parallelisms and reversals. The study of independent evolution can lead to important insights into the mechanisms of evolutionary change. Are independently acquired characters genetically controlled by a similar molecular machinery? Are independently acquired characters homologous if generated in parallel from homologous genetic material? Is the evolutionary gain or loss of structures reversible?

Here, I use a developmental approach to study three cases of independent evolution in invertebrates and vertebrates, and I discuss the results in light of the genetic and developmental mechanisms controlling morphological evolution. First, I investigated the evolutionary origin of animal cartilage. I studied cartilage development in two invertebrates from the cephalopods and arthropods. I found that despite their independent origins in vertebrates and invertebrates, cartilaginous tissues conserve a

homologous gene regulatory network for chondrogenesis. Next, I studied the evolution and development of cephalopod limbs and found that despite having evolved independently from arthropod and vertebrate limbs, they conserve a homologous appendage gene regulatory program. Finally I investigated limb development in a snake-like legless lizard that belongs to a lineage in which reversal of limb loss has occurred in derived species. I discovered that despite being legless as adults, as embryos they develop transitory but almost complete limbs with digits but these later degrade. I propose a model to explain how this mode of limb loss could explain the extraordinary case of limb re-evolution in this lizard lineage. These three examples reveal how independent evolution of similar characters involved the parallel recruitment of homologous pre-existing genetic cassettes. The modular regulatory dynamics of animal development then allowed these cassettes to be independently activated, an important mechanisms for the evolution of novel structures or tissue types. Finally, my discovery that these conserved developmental modules can be re-activated provides a mechanistic basis for so-called evolutionary reversals.

CHAPTER 1 INTRODUCTION

The advent of molecular phylogenetic analyses has created unparalleled opportunities for the study of evolutionary morphology [1]. Robust phylogenies are now invaluable tools to understand the patterns of morphological evolution and the processes behind such phenotypic transformations [2]. As organisms evolve, homologous structures diverge, sometimes are eliminated and other times seem to re- appear. Thus, even with robust phylogenies, tracing the evolution of morphological traits over phylogenies is some times quite challenging [1].

An organismal perspective is needed in order to reveal the relationships of certain characters that have diverged so much during evolutionary time, of which no one would predict from a first glance they were homologous in the first place [2-4].

Homology, or similarity due to common ancestry regardless of form and function, is a central piece in our understanding of evolution in general. Although, conceptually homology at the DNA level is some times trivial (due to the semiconservative form of

DNA synthesis), homology at a morphological level is a little harder to understand, due to our limited knowledge on the relationship between genotype and phenotype and the fact that morphological characters are not physically transmitted, constructed de novo in each generation [5]. Nonetheless, homology can be easily understood simply as a byproduct of evolution, molecularly and morphologically. However, an operational definition of morphological homology has been harder to achieve due to the limitations in our understanding of character identity [5].

It is expected that characters diverge over time, but in some instances, characters that are far apart in the phylogeny display astonishing morphological

similarity. However, the phylogenetic position of the species that carry these traits can indicate that homology has to be ruled out, since the characters show no historical continuity, and instead show that they converged on a similar morphology independently [6]. But how can such characters evolve independently ? What are the molecular mechanisms controlling the formation of convergent morphologies? Are they using parallel genetic mechanisms or completely different ones? It was a big surprise for evolutionary biologist to learn that textbook examples of convergent characters in animal evolution, such as locomotory appendages and eyes, were proven to form by the control of homologous regulatory genes despite their clear phylogenetic discontinuity for over 600 millions of years of animal evolution [3, 4, 7, 8]. These regulatory genes were not only responsible for the formation of the character in question, but for many others as well, demonstrating their developmental promiscuity [3-5, 9].

As a result, the concept of deep homology was coined to refer to cases where there was clear independent evolution of the morphological character but the underlying molecular generative mechanisms bear homology, although at a complete different level of biological organization [4]. Therefore, rather than rampant convergence at the level of morphology, function, and gene expression, independent evolution of non-homologous characters can be caused by the parallel co-option of homologous gene regulatory networks [4, 5, 8].

Here I take a developmental genetic approach to address the mechanisms behind three cases of independent evolution, animal cartilage, animal appendages and lizard limbs. In Chapter 2, I address the question of how cartilage has evolved independently in animals belonging to each of the three of the major animal lineages,

Lophotrochozoa, Ecdysozoa and Deuterostomia. We compared the developmental mechanisms that control the histogenesis of this vertebrate cartilage in two invertebrates and demonstrated that they share important gene regulatory properties with vertebrate cartilage, suggesting independent recruitment of a homologous gene regulatory network.

In Chapter 3, I investigate how cephalopod mollusks evolved arms and tentacles.

Within mollusks, cephalopod arms and tentacles have no clear homologous counterparts, representing a true evolutionary novelty. However, I demonstrate that even though cephalopod, arthropod, and vertebrate appendages are not homologous structures, the developmental genetic mechanisms that underlie the evolution of cephalopod appendages are conserved, indicating parallel recruitment of the same gene regulatory network that vertebrate and arthropods use for limb development.

Finally, in Chapter 4, I analyze one of the most interesting cases of complex character re-evolution, the reversal of limb loss in a South American lineage of miniaturized lizards. Dollo’s law states that once a character is lost in evolution, then it will never return in the ancestral state. In this lizard lineage, however, robust molecular phylogenetic analysis has suggested that complete limbs re-appeared from lineages that ancestrally lost them. I show that legless lizards in the genus Bachia conserve a surprisingly complete limb (with digits) during embryonic development, but at later stages, this embryonic limb is completely degraded, indicating that complete hindlimbs are only transitory in their ontogeny. Therefore, by taking an ontogenetic approach, I provide evidence that limb loss is not necessarily unidirectional and that re-acquisition of complete limbs is a plausible evolutionary transformation.

CHAPTER 2 MATERIALS AND METHODS

Embryo Collection and Preparation

Sepia phraonis eggs were obtained from the National Resource Center for

Cephalopods, Galveston TX USA. Sepia bandensis and Sepia officinalis eggs were purchased from commercial suppliers. Upon arrival at our institution, eggs were cultured in artificial seawater (Petco© Real Ocean Water) at 22 degrees Celsius. Embryos were collected by manual removal of egg cases and were staged according to Reference

[10]. Embryos used for in situ hybridization and immunohistochemistry were fixed and processed as previously described [11]. Limulus embryos were kindly provided by B.

Battelle and H. J. Brockmann, Department of Biology, University of Florida, were staged according to Reference [12] ,and were processed for in situ hybridization as previously described [13].

Alcian Blue and Masson’s Trichrome Stain

Alcian blue staining is the gold standard for glycosaminoglycan (GAG) staining, particularly for vertebrate cartilage. Alcian blue can detect highly anionic (GAG), such as hyaluronan and sulfated GAGs. Early biochemical analyses of cartilaginous tissues in cephalopods and horseshoe crabs indicate the presence of highly sulfated GAGs, such as chondroitin sulfate [14-16]. To detect GAGs in Sepia and Limulus cartilage, we used alcian blue/nuclear fast red staining on paraffin sections. Deparaffinized sections were stained for 30 minutes in 1% Alcian blue (in 3% acetic acid) and counter stained with nuclear fast red for 5 minutes. Masson’s trichrome staining was performed on paraffin sections using a Masson’s Trichrome Kit (22-110-648, Richard-Allan Scientific™) following manufacturer instructions.

Gene Cloning and RACE-PCR

RNA extraction from Sepia embryos at stages 24-26 and from Limulus embryos at stages 19-20 was done using Trizol® reagent (Ambion®) following manufacturer instructions. cDNA synthesis was carried by an AMV reverse transcriptase (New

England Biolabs© inc) following manufacturer instructions. PCR amplification was carried using the following primers: SepiaSoxEr, TGCTACCATGTTAGAAGTCATGCCT;

SepiaSoxEf, GATTACCCTGA TTACAAATACCAGCCC; SepiaSoxDf,

CCACTACCAGCTCATAGC AACCATCAG; SepiaSoxDr,

GGGCTTTGAGGGGTCAGGTTTCTCT; SepiaColAaf,

AACGCCCCTGCCCGTTCCTGTCGCGATC; SepiaColAar, TCCCAATTCTATATGGA

AGTCTTGT; Sepiahhf, TAATGTATCGGAAAACACAGTTGGTGCCA; Sepiahhr, GAGG

AAGGCGATGACTTCGCTGTAA; SepiabetacateninF1, TGTGCTGCTGGCATTCTGTC

CAATC; SepiabetacateninR1, GCGACTCCTTCGTTCCTGGAGTGTA; LimulusSoxDF,

CCAAAGAGAACTTGTATTGTGGATGGC; LimulusSoxDR, GGTGTCTGTCTCTCAGC

TTGAAACATACCA; LimulusSoxEF, TTGCATGGACAAACTCG TCAACTCGGT;

LimulusSoxER, GGAAACTGGATACTGATGATATGGAGTATC; LimulusColAaF, ATA

TGATGCAAGTGCTCTTGCTGCTCTCCT; LimulusColAaR, CTCACTGAAGAGTTGTA

GGAAACTAAGCTG; LimulusHhF, GTCTTTAAGCARCAYGTNCCNAA; LimulusHhR,

AAAGTTTGCGTACCARTGDATNCC; LimulusbcatF, TTATGCCATCACTACCTTGCAC

AATCTC; LimulusbcatR, CTTGACAAGTGCAGG AATTCCCCCAGAT.

Full-length cDNA clones were isolated by rapid amplification of cDNA ends using

SMARTer® RACE 5’/3’ Kit (Clontech©) and synthesis of 5’ and 3’ RACE cDNA libraries was performed following manufacturer instructions. The primers used in RACE-PCR experiments were: HSCraceColAbr, GTAAAACGACGGCCAGTCGGCAGTGGTAGGTA

ATATTCTGTACAGC; SepPhRACEcolAbr, GTAAAACGACGGCCAGTCGGCAGTGGT

AGGTAATATTCTGTACAGC; SepPhRACEcolAar, GTAAAACGACGGCCAGTGAGAC

AACCACACATAGGACTCTCCGGCT.

Sequences for -catenin, and for

Limulus -catenin have been submitted to GenBank.

In situ Hybridization and Immunohistochemistry

Whole mount in situ hybridization was performed using digoxigenin- and fluorescein-labeled antisense (or sense control) RNA probes according to protocols previously described for Sepia [11] and Limulus [13]. In situ hybridization on cryosections was performed using previously described protocols for vertebrate tissues

[17]. PCNA, β-catenin and hyaluronan detection was performed on cryosections using mouse anti-PCNA (ab29, abcam®), ant-β-catenin (C2206, Sigma-Aldrich®) and biotinylated hyaluronic acid binding protein (385911, EMD Millipore©). Hyaluronan detection on mouse, Sepia and Limulus cartilages was carried out using streptavin-HRP with Alexa Fluor 488 tyramide signal amplification (TSATM, Molecular Probes®).

In situ Hybridization and Phalloidin Staining

Phalloidin staining was performed in cryosectioned embryos after whole mount in situ hybridization. RNA expression was imaged by detecting the fluorescence generated by the NBT/BCIP precipitate, which emits at over 700nm when exited at a 633nm.

Phalloidin staining was done using Alexa Fluor® 488 phalloidin (Life Technologies™).

Sections were blocked with 1% BSA (A9647, Sigma-Aldrich®) in PBS before 30 minutes incubation with Alexa Fluor® 488 phalloidin at 6.6 µM in blocking solution (1%

BSA in PBS).

Optical Projection Tomography (OPT) of Collagen in situ Hybridization

Sepia and Limulus embryos were fixed in 4% paraformaldehyde in PBS after the completion of whole mount in situ hybridization and the embryos were prepared and scanned following previously described protocols for vertebrate embryos [18, 19]. OPT scanning was performed using a Bioptonics 3001 OPT Scanner. The anatomy and gene expression channels were reconstructed using NRecon software and imported into the Amira® program for 3D visualization, analysis and renderings of 3D images and movies.

Molecular Phylogenetic Analysis of Collagen and Sox Genes

Phylogenetic analyses of Collagen and Sox sequences cloned from Sepia and Limulus cDNA pools were aligned with putative orthologs derived from EST databases (NCBI) by tBlastn searches using mouse Col2a1 (AAH51383) and Haliotis collagen pro-alpha chain (BAA75668) collagens, and mouse Sox9 (NP_035578) and

Sox6 (AAC52263). The retrieved sequences used for the phylogenetic analyses can be found in Table S1 and S2. Amino acid sequences were aligned using MUSCLE [20] and phylogenetic reconstruction was performed with MrBayes 3.2.2 [21] using the WAG model [22] of amino acid substitution, as described previously [23].

Treatments With Small Molecule Inhibitors

S. officinalis embryos were staged inside their egg cases by removing most of the outer layers of the egg case until the remaining inner layers were translucent enough to see the embryo. Embryos were selected for the treatments once

they reached stage 23-24, when they have already developed the funnel cartilage primordia. Treatments were done in 100 ml glass beakers with 50 ml of sterile artificial seawater. Control embryos were treated with DMSO at 0.1%, and experimental embryos were exposed to 10 μM of cyclopamine (C988400, Toronto Research

Chemicals Inc), SANT-1 (S4572, Sigma-Aldrich®), Alsterpaullone (A4847, Sigma-

Aldrich®), BIO (B1686, Sigma-Aldrich®) or PNU-74654 (P0052, Sigma-Aldrich®) by adding 50 μl of 10 mM stock solutions for each of the drugs. DMSO was used as the solvent for all stock solutions.

Embryos were incubated at 22 degrees Celsius in seawater with the corresponding drug or DMSO. Seawater containing the drug at the appropriate final concentration (or DMSO for controls) was replaced every two days, for a total exposure period of 5 days or 10 days. Specimens were then collected and fixed for histology or in situ hybridization, as described above.

BrdU Labeling and BrdU Pulse Chase

S. officinalis embryos (n=8) were incubated in 0.05% 5-bromo-2'-deoxyuridine

(BrdU) in seawater at 22 degrees Celsius for 6hr. Four embryos were processed immediately for immunofluorescence after BrdU incubation, and the rest of the embryos were rinsed several times in seawater free of BrdU, then washed 5 times every 10 minutes, then 5 times every 30 minutes, and were finally incubated for 3 days making complete water changes every 24 hours. Embryos were fixed overnight in 4% paraformaldehyde. BrdU labeling was done by immunofluorescence on cryosections,

antigen retrieval was performed by 30 minutes incubation in 2N HCL. BrdU was detected anti-BrdU (G3G4, DSHB).

Luciferase Assay

We used a firefly luciferase reporter construct controlled by the col2a1 promoter and the Col2a1 chondrocyte specific enhancer [24]. We cloned SoxE and Sox9 full- length from cDNA by RT-PCR and ligated into a pcDNA3.3 expression vector under the control of a CMV promoter (Invitrogen). We used two cell lines, NIH3T3 mouse fibroblast and immortalized rat chondrocytes and culture them as previously described

[24]. Cells were transfected using Lipofectamine3000 (Invitrogen) with the luciferase plasmid and the corresponding expression vector together with a Renilla luciferase control (for well normalization). Cells were incubated after transfection for 48h until cell lysis and the luciferase activity measured using the Dual-Luciferase® Reporter Assay

System (Promega).

Table 2-1. Sequence identifiers for the collagen phylogenetic analysis. Species name Genbank Acyrthosiphon pisum XM_001942728 Alvinella pompejana AAC35289 Anopheles gambiae NT_078267 XM_391942 Apis mellifera XM_393523 EB249446 Aplysia californica GD229044 Arenicola marina AAC47545 Branchiostoma floridae ABG36939 NP_001029004 Ciona intestinalis XP_004226827 Cochliomyia hominivorax FG294109 Daphnia pulex FE335735 Epiperipatus sp. AM499446 DW282906 Euprymna scolopes DW252688 Haematobia irritans FD463616 BAA75668 Haliotis discus BAA75669 EY380843 Helobdella robusta EY365668 Hirudo medicinalis FP591665 Hirudo medicinalis EY479974 AAB69977 AAH86874 Homo sapiens BAD92412 BAD92923 Ixodes scapularis EL516416 Locusta migratoria CO856142 FC620367 Lottia gigantea FC564501 Petrolisthes cinctipes FE776091 Tribolium castaneum XM_966372 Well-annotated reference sequences (Ref-seq) and non-annotated expressed sequence tags (EST) in NCBI used for the collagen phylogenetic analysis.

Table 2-2. Sequence identifiers for the phylogenetic analysis of Sox genes. Species name Genbank XM_003746176 Metaseiulus XM_003741168 occidentalis XP_003746224

XM_006529276 NP_035578 Mus musculus NP_035576 NP_033260

Xenopus laevis NM_001087769

NM_166792 NP_651839 Drosophila NP_523739 melanogaster NP_476894 NP_001014695

NM_214472 Strongylocentrotus XP_786809 purpuratus XP_791983 XP_798084

XM_001603093 Nasonia vitripennis XP_001604932 XP_001603143

XM_002415871 Ixodes scapularis XP_002404127 XP_002415916

Crassostrea gigas BQ426397

XP_001122996 Apis mellifera XP_006562841 XP_006570327

Anopheles gambiae XP_560125 Well-annotated reference sequences (Ref-seq) and non-annotated expressed sequence tags (EST) in NCBI used for the phylogenetic analysis of Sox genes.

Table 2-3. Sequence identifiers for the phylogenetic analysis of HAS genes. Species name Genbank XP_002586939 XP_002586937 XP_002586938 Branchiostoma floridae XP_002589930 XP_002586924 XP_002585658 XP_002589932

NP_001514 Homo sapiens NP_005319 NP_005320

XP_005995754 Latimeria chalumnae XP_006003209 XP_006008738

XP_007884111 Callorhinchus milii XP_007898735 XP_007906128

XP_006641657 Lepisosteus oculatus XP_006641229 XP_006636013

Petromyzon marinus ENSPMAP00000003453 Reference sequences (Ref-seq), predicted genes and non-annotated expressed sequence tags (EST) in NCBI and Ensembl used for the phylogenetic analysis of HAS genes.

Table 2-4. S. officinalis embryos used for each drug treatment and DMSO controls. Drugs Masson’s ColA Hh PCNA/βcatenin Death DMSO 5 4 3 5 1 Alsterpaullone 4 4 3 4 1 SANT-1 4 4 - 4 0 Cyclopamine 4 4 - 4 1 BIO 4 - - - 1 IWR-1 4 - - - 0 PNU 4 - - - 0 Total 29 16 6 17 3 Out of 71 embryos the effect of these small molecule inhibitors on chondrogenesis was analyzed by ColA and Hh in situ hybridizations, PCNA/ βcatenin immunofluorescence and Masson’s trichrome histological staining.

CHAPTER 3 CARTILAGE DEVELOPMENT IN TWO PROTOSTOME LINEAGES

Cartilage in Metazoans

The evolution of different cell types was fundamental in the evolution of organismal complexity of metazoans [25]. Even the earliest branching metazoan lineages already have a significant diversity of cell types, as in the case of ctenophores, sponges and cnidarians [26]. Complex metazoans, such as bilaterians, display a high diversity of cell types, which are the building blocks for the construction of complex organ systems for numerous physiological functions, like the nervous system, muscular system, sensory organs, excretory apparatuses and endocrine regulation [26]. Due to their dissimilar morphologies, finding homologous organs in metazoans, and by extension homologous cell types, has been challenging. This challenge has been perhaps nicely represented by the study of two of the most fundamental animal tissues and cell types, the nervous system and muscles, formed by neurons and myocytes, respectively [27, 28].

Our understanding of the evolutionary origin of nervous system and muscles was violently shaken by the sequencing of the first ctenophore genomes and recent discoveries on the development and physiology of these two tissues in ctenophores [27,

28], all of this fueled by the comprehensive comparative genomic analysis of ctenophore genomes with the ones sequenced in other early branching metazoans, such as sponges and cnidarians, as well as many bilaterians from the three major branches, deterostomes, ecdysozoans and lophotrochozoans.

Classic literature on animal evolutionary morphology had little doubts on the presence of these two tissues in ctenophores and their close phylogenetic position to

cnidarians or at least, if not sister to cnidarians, their almost equivalent phylogenetic position in relation to bilaterians and sponges [26]. The first comparative genomic and phylogenomic analysis of ctenophore genomes showed to everyone’s surprise that despite the shared overall morphology of ctenophores with cnidarians they most likely branched out from the metazoan phylogenetic tree before sponges [27, 28].

This well supported analysis had major implications for the evolution of many morphological similarities among ctenophores and other metazoans, like cnidarians and bilaterians. Going back to the evolution of the nervous system and muscles, these discoveries meant that either those two tissues were simply lost in sponges or they evolved independently in ctenophores and eumetazoans (cnidarians + bilateria) [27,

28].

Our knowledge on the biological nature of neurons and myocytes in early branches of metazoans comes only from histological or ultrastructural studies of cell morphology. Therefore, the unexpected discovery of ctenophore phylogenetic position raised the possibility that despite how unlikely this scenario might seem, neurons and myocytes could have evolved independently in metazoans.

Indeed the comparative genomic analysis of ctenophore and eumetazoan genomes, together with gene expression analysis during embryogenesis, revealed that cnidarian and bilaterian neurons and myocytes are homologous, sharing numerous genes crucial for their characteristic cell differentiation programs, structural molecules and cell physiological behavior. On the other hand ctenophores do not contain the genes that provide the structural and physiological hallmarks of neurons and myocytes,

which imply the two cell types are: first, non-homologous and second, they therefore should have evolved independently in ctenophores and eumetazoans [27, 28].

These example show how challenging it is to establish cell identity in animals based merely on cell morphology, and the game-changing perspective when molecular genetics analysis are considered [25]. Given the proposed homology of these cell types at least across eumetazoans, another important finding is that neurons and myocytes are very ancient cell types. Thus, despite the divergent bodyplans of eumetazoan lineages, they seem to be constructed by similar cell types that evolved very early in animal evolution, raising the possibility of existence of many other cell types conserved among metazoans.

For over a century a similar debate has been disputed in relation to the evolutionary origin of cartilage [29-32]. Cartilage is rarely recognized as such by zoologists outside vertebrates, and invertebrates are generally considered devoid of endoskeletal tissues, where invertebrates usually have either hydrostatic skeletons or exoskeletons [26].

However, in some protostome and deuterostome invertebrate lineages special endoskeletons are present, but they have been considered cartilage-like tissues at the most. From a developmental, genetic and evolutionary perspective the term “cartilage- like” bears very little information about the biological nature of these tissues and how they are related to vertebrate cartilage or connective tissues in general.

Similar to vertebrate cartilage, invertebrate cartilage shares interesting morphological similarities, such as being avascular and composed by non-adjacent cells separated by abundant extracellular matrices, as well as their organization into complex

endoskeletal systems [30, 33]. These morphological similarities, however, are very narrow in order to recognize these tissues as homologous.

Homology of Cartilage in Metazoans

Invertebrate cartilaginous cells are more diverse in shape compared to vertebrate cartilage, they have a fibroblastic or epithelial cell shapes rather than the round shape vertebrate chondrocytes have [34]. Moreover, in some cases invertebrate endoskeletal tissues are formed by acelullar cartilage-like extracellular matrices, like the pharyngeal skeleton of hemichordates [35]. This diversity in tissue and cell morphology questions the equivalence of invertebrate and vertebrate cartilage, making possible the scenario of convergent tissue evolution as was demonstrated for nervous system and muscles of ctenophores compared to eumetazoans.

Cartilage-like tissues are largely absent within bilaterians; therefore, if cartilage is a homologous tissue across Bilateria, then it was lost multiple times during bilaterian evolution since there is no phylogenetic continuity of cartilage in the bilaterian evolutionary tree. This implies that cartilage is non-homologous across bilateria and should have evolved independently, and the tissue similarity of these tissues is only rampant evolutionary convergence.

Origin of New Cell Types

The origin of cartilage and the chondrocyte has been classically centered in the study of the emergence of cartilage in jawed vertebrates (gnathostomes), as a new cell type that correlates with the evolutionary origin of this lineage as well [32, 36, 37].

Nearly a decade ago, however, it was demonstrated that cartilage originated earlier than previously thought. Cohn and collaborators demonstrated that lampreys and hagfishes also have cartilage [23, 38], which despite their unique ECM composition and

low collagen content, is still based on type II fibrilar collagen ECM and developmentally regulated by homolog master cartilage regulatory genes. This discovery pushed back the origin of cartilage to at least the last common ancestor of cyclostomes and gnathostomes and showed that cartilage can come in many flavors [23, 38], which motivated the quest for the cellular precursors of chondrocytes in invertebrate deuterostomes.

Recent discoveries on the molecular composition and regulation of endoskeletal tissues in invertebrate deuterostoms have also shed light on our understanding of cartilage evolution as a new cell type [35, 38-40]. Gene expression analyses during embryological formation of some endoskeletal tissues of cephalochordates, urochordates and hemichordates have pointed out similarities of these tissues with vertebrate cartilage at a molecular level [35, 38-40]. The scenario that emerged from these works essentially suggests that at least part of the cartilage gene regulatory network could predate the evolution of cartilage, but they do not postulate a putative invertebrate cell precursor for the origin of vertebrate cartilage as a new cell type.

Cell Identity and Gene Regulatory Networks

Vertebrate cartilage is a type of connective tissue where cells are embedded in abundant ECM. Its ECM is enriched in two major biomolecules, fibrilar collagen and highly hydrophilic acidic polysaccharides (acidic glycosaminoglycans or GAGs) [41].

Although the structural nature of cartilage has been well established (histologically and molecularly), little was known about what factors regulated cartilage formation from undifferentiated mesenchymal cells until the first studies on Sox transcription factors (20 different genes in humans) [42, 43]. Sox5, Sox6 and Sox9 are transcriptional regulatory factors that play essential roles in chondrogenesis, which is why they are also known as

master regulators of cartilage development [44]. They control the active expression of numerous genes during chondrogenesis and also the repression of genes for other cell lineages, in that way Sox transcription factors control cartilage lineage identity at the expense of other cell differentiation pathways [45, 46]. Sox genes are especially important in the direct control of cartilage structural genes, such as collagens, thrombospondins and proteoglycans [44]. Sox families of transcription factors are critical for controlling cell differentiation programs, acting as gatekeepers of cell differentiation stages [45, 46]. Sox genes are usually located in highly integrated positions of gene regulatory networks where they integrate upstream signals and translate them in downstream cell differentiation programs [46, 47].

Therefore, Sox5,6 and 9 are not only essential for chondrogenesis but also sufficient to experimentally induce ectopic chondrogenesis in competent cell populations in cell culture and in vivo [48].

Histology and Molecular Nature of the ECM

Cartilage forms the embryonic endoskeleton of all vertebrates and has been widely considered to be a vertebrate-specific tissue [31, 36, 39]. Cartilage-like tissues have been recognized in invertebrate species scattered throughout the Protostomia; however, the relationship of these protostome tissues to the bona fide cartilage of vertebrates has long been debated [31, 49], and the evolutionary origin of cartilage and its parent cell type, the chondrocyte, is unknown [29, 30, 34, 37, 38]. Cartilage is an endoskeletal connective tissue formed by cells embedded in abundant extracellular matrix (ECM) that is rich in collagen and acidic glycosaminoglycans (GAGs) [30, 32]. To determine whether invertebrate cartilage-like tissues have structural and/or chemical similarities to vertebrate cartilage, we first investigated the structure and matrix

composition of the these connective tissues in adults of two distantly related protostome invertebrates, the cuttlefish Sepia bandensis from the Lophotrochozoa and the horseshoe crab Limulus polyphemus from the Ecdysozoa, which are among the best- known examples of invertebrate cartilage-like tissues [30, 33]. We found that cartilage- like tissues in both species are composed of cells embedded in abundant ECM that is rich in collagen and acidic GAGs, and that these tissues form conspicuous endoskeletal structures (Figure 3-1). Based on the similar composition of these cartilage-like tissues

(hereafter cartilage) to vertebrate cartilage, we investigated the pattern of chondrogenesis in Sepia and Limulus. In both species, development of the cartilaginous endoskeleton begins during late stages of organogenesis, with formation of prechondrogenic mesenchymal cell condensations before secreting an ECM that is rich in collagen and acidic GAGs, mirroring the process of vertebrate chondrogenesis

(Figure 3-2).

Invertebrate Cartilage is Collagen Based and Contains Hyaluronan

To test whether a common genetic program for cartilage development is conserved across the Bilateria, we asked whether Sepia and Limulus cartilages express the invertebrate pro-orthologs of vertebrate Collagen2a1 (Col2a1), which encodes type II collagen, the most abundant protein of the vertebrate cartilage ECM. Despite the absence of fibril-forming collagens in model ecdysozoan organisms [50], recent genome wide analyses in other protostome and cnidarian genomes suggest that all three families (A, B and C) of fibril forming collagens were likely present at the root of Bilateria [51, 52]. We isolated the C-terminal propeptide of numerous protostome putative Clade A collagens

(ColA) using public EST databases and by RT-PCR amplification of two putative ColA genes in Sepia and one in Limulus. Molecular phylogenetic analysis of these sequences

revealed that Clade A fibrillar collagens are widespread in Protostomia and placed our putative ColA sequences from Sepia and Limulus as Clade A collagens (Figure 3-3). To determine if invertebrate cartilages express ColA during chondrogenesis, we analyzed its expression during cartilage formation in cuttlefish and horseshoe crab embryos. In Sepia embryos at stage 26, ColAa and ColAb expression maps to numerous regions of chondrogenesis, including the funnel, nuchal, fin and cranial cartilages (Figure 3-4 A-C and

Figure 3-5). Similarly, in Limulus embryos at stage 20, we detected ColA expression in the endosternite (Figure 3-4 D-F) and gill cartilages (Figure 3-6 Q arrowhead).

We isolated full-length cDNAs and found that ColA genes in Sepia and Limulus have an extended triple helical domain located between the N-terminal von Willebrand type-C domain and the C-terminal propeptide domain (Figure 3-4 G), which demonstrates that cuttlefish and horseshoe crab ColA genes have conserved the architecture of Clade A fibril-forming collagens [50, 52]. Another essential component of the ECM of vertebrate cartilage is hyaluronan (HA), a very large, non-sulphated and abundant GAG in the cartilage ECM [53, 54]. Although, largely absent outside vertebrates HA has been detected in some protostome tissues[55]. Using a hyaluronan binding peptide, we determined that Sepia and Limulus adult cartilage ECM are positive for hyaluronan (Figure 3-4 H). Intriguingly, hyaluronan synthases (HAS) were not recovered from transcriptomic or genomic databases outside chordates (Figure 3-3), suggesting the existence of yet non-characterized HAS enzymes or the parallel evolution of HAS from chitin synthases (abundant in arthropods, mollusks and other protostomes), as proposed for the origin of vertebrate HAS [56]. Taken together these results demonstrate that key structural molecular components of cartilage are shared

between invertebrate and vertebrate cartilages and that invertebrate cartilage is fibrillar collagen-based.

Expression of Developmental Genes in Sepia chondrogenesis

We next investigated whether cell signaling proteins and transcription factors that function upstream of Col2a1 in the vertebrate chondrogenesis pathway also are expressed during invertebrate chondrogenesis. The hedgehog signaling pathway plays essential roles in early vertebrate chondrogenesis, when Shh regulates transcriptional activation of Sox5/6/9 [57-59], and during later chondrogenesis, when Ihh regulates cartilage proliferation and differentiation [47]. We cloned Hh from Sepia, and analysis of its expression revealed that cartilage differentiation takes place in close proximity to Hh expression domains (Figure 3-6 D, e and Figure 3-7). In Sepia embryos at stage 26,

ColAa and ColAb are expressed in a “U-shaped” domain of mesenchymal pre- cartilaginous cells immediately adjacent to the Hh-expressing cuboidal epithelium, in the region that will later form the funnel cartilage (Figure 3-6 D-G and Figure 3-7 A and B).

Downstream of hedgehog signaling, vertebrate Sox9, Sox5 and Sox6 transcription factors are master regulators of chondrogenesis that function as direct transcriptional activators of Col2a1 [42, 43, 60]. We isolated invertebrate pro-orthologs of Sox9 and Sox5/Sox6, called SoxE and SoxD respectively (Figure 3-3), and analyzed their expression in Sepia embryos. We found that Sepia funnel cartilage condensations express SoxE (Figure 3-6 H and I and Figure 3-8) and SoxD genes (Figure 3-6 J and K and Figure 3-8), mirroring the demarcation of vertebrate cartilage condensations by

Sox9 and Sox5/6 [43]. Comparison of the SoxE and SoxD domains with the expression pattern of ColA indicated that these factors are co-expressed in the developing cartilaginous skeleton of Sepia embryos (Figure 3-9 F-K, Q-T).

Sox9 and β-catenin have opposing functions in vertebrate chondrogenesis; they inhibit each other’s transcriptional activity and β-catenin functions as an anti- chondrogenic transcriptional regulator [61-63]. During cartilage differentiation, reduction of β-catenin transcript and protein levels in Sox9-positive chondroprogenitor cells is necessary for cartilage differentiation [61-63]. To test whether this regulatory relationship is conserved in Sepia cartilage, we cloned and characterized the dynamics of β-catenin expression during funnel cartilage differentiation. β-catenin transcripts were detected in pre-cartilaginous mesenchymal cells and in the overlying epithelium at stage

24-25 (Figure 3-8 E), but by stage 26, β-catenin transcription had been shut down and expression could no longer be detected in funnel pre-cartilaginous cells (Figure 3-6 L).

As in vertebrates, this decrease of β-catenin in Sepia precedes differentiation of chondroblasts and the first appearance of aniline blue-positive cartilage matrix at stage

27 (Figure 3-2).

Expression of Developmental Genes in Limulus chondrogenesis

Having established that components of the core chondrogenic network that regulates ColA are conserved between vertebrates and a lophotrochozoan, we tested whether this conservation extends to Ecdysozoa by cloning and analyzing the expression of Hh, SoxE, SoxD and β-catenin during Limulus chondrogenesis. In

Limulus embryos, the endosternite cartilage forms immediately adjacent to the Hh- expressing ventral nerve cords (Figure 3-6 O-P and Figure 3-2). The pre-chondrogenic condensation of the endosternite can be identified as a thin plate of ColA expressing cells dorsal to the paired nerve cords (Figure 3-6 Q and R). SoxE also is expressed throughout the developing endosternite plate (Figure 3-6 S and T) and gill cartilage

(Figure 3-10 C), although SoxD, was not detectable in these tissues (Figure 3-6 U-V).

Similar to Sepia and vertebrates, Limulus β-catenin is downregulated in SoxE/ColA- expressing cells prior to cartilage differentiation (Figure 3-6 W). Taken together, our analysis of Sepia and Limulus chondrogenesis indicates that a network of structural and regulatory genes required for vertebrate cartilage development has deeply conserved patterns of expression in the three major lineages of Bilateria.

Invertebrate SoxE proteins can activate vertebrate Col2a1 cartilage enhancer

To demonstrate the invertebrate SoxE proteins could play a similar role in invertebrate collagen regulation, we tested if Sepia and Limulus SoxE proteins can transcriptionally activate the vertebrate Col2a1 cartilage specific enhancer, as is the case with vertebrate Sox9. We used a luciferase reporter downstream of the human

Col2a1 enhancer and tested the activity of the invertebrate SoxE proteins, as well as

SoxE proteins from amphioxus (which lacks cartilage), a lamprey and a hagfish, and

Sox9 proteins from a shark and a teleost fish. We found that in the two cell lines used

(mouse fibroblasts and rat chondrocytes) the is activation of the Col2a1 enhancer by the

Sepia and Limulus SoxE proteins, also by the some of the lamprey and one of the hagfish SoxE and as expected by the Sox9 proteins from the shark and teleost fish

(Figure 3-11). Therefore, amphioxus, the only species with no cartilage, was the only species that had no SoxE transactivation function over the cartilage specific Col2a1 enhancer. This suggests that Sepia and Limulus SoxE proteins share similar transactivation functions as vertebrate SoxE and Sox9 proteins, and support the idea of a conserved regulatory network between cartilage bearing vertebrates and invertebrate species.

Late Chondrogenesis and Cartilage Appositional Growth

During embryonic development of the vertebrate long bones, cartilage growth is controlled by the progress of chondrocytes through a series of phases moving from the resting zones at the epiphyseal ends into the zones of proliferation, maturation and hypertrophy towards the center of each element [64]. Indian hedgehog (Ihh) is expressed in prehypertrophic chondrocytes and controls growth of cartilage anlagen by regulating chondrocyte differentiation and stimulating chondrocyte proliferation adjacent to the Ihh domains [65]. In Sepia embryos at stage 29-30 (just before hatching), we observed a thin non-cartilaginous cell layer positioned between the funnel epithelium and the overtly differentiated matrix-embedded chondrocytes situated deeper in the funnel cartilage (Figure 3-9 A). To test whether growth of the Sepia funnel cartilage involves directional proliferation and maturation, similar to vertebrate cartilage, we examined cell proliferation by detecting the proliferating cell nuclear antigen (PCNA) and incorporation of the nucleotide analog 5-bromo-2'-deoxyuridine (BrdU). PCNA immunoreactivity showed that at stages 27-28, chondrocytes are proliferating over the entire element (Figure 3-12) but later, at stages 29-30, PCNA and BrdU show that cell proliferation is restricted to the thin non-cartilaginous layer (Figure 3-9 B and C), indicating that a zone of cell proliferation is maintained immediately under the Hh expression domain (Figure 3-9, compare B to J). Pulse chase BrdU labeling further demonstrate that the cell progenitors of the proliferation zone are incorporated as chondrocytes into the funnel cartilage (Figure 3-9 D-F). SoxE, SoxD, β-catenin, ColAa and ColAb also are expressed in the thin non-cartilaginous layer at the periphery of the funnel cartilage (Figure 3-9 G, H, I and K), but only ColA remains on in chondrocytes as they move deeper into the funnel (Figure 3-9 G). The restriction of proliferating cells to

the perimeter of the funnel cartilage, adjacent to Hh-expressing cells, the expression of

SoxE, -catenin and ColAa/ColAb in this highly-proliferating undifferentiated layer, and the maintenance only of ColA deeper in the cartilage suggests that appositional growth of the Sepia funnel cartilage occurs at the ends of the element

(Figure 3-9 X), similar to vertebrate cartilage growth.

Hedgehog and β-catenin signaling pathways have antagonistic functions in Sepia chondrogenesis

We then tested experimentally whether expression of Hh and degradation of β- catenin are necessary for activation of ColA and differentiation of funnel cartilage in

Sepia, as is the case in vertebrates. Functional studies were carried out using the small molecules cyclopamine and SANT-1 (inhibitors of Smoothened) to block Hh signaling

[11, 66], and alsterpaullone (inhibitor of the GSK-3β destruction complex) to stabilize β- catenin [67]. We reasoned that if these signaling pathways are conserved between protostomes and deuterostomes, then blocking Hh signaling or preventing degradation of β-catenin should inhibit ColA expression and formation of the funnel cartilage in

Sepia. We initiated treatments at stage 23-24, before formation of pre-cartilaginous cell condensations, but after the appearance of the funnel epithelium and associated mesenchyme (Figure 3-13). After 5 days of treatment (~ stage 26), antagonism of Hh signaling by cyclopamine or SANT-1 and stabilization of β-catenin by alsterpaullone resulted in loss of ColAa expression (Figure 3-9 M-O), whereas control embryos treated with DMSO showed normal ColAa expression in the pre-cartilaginous cells (Figure 3-9

L). Hh antagonism also prevent downregulation of β-catenin (Figure 3-9 P-S) but stabilization of β-catenin did not affect Hh expression in the funnel epithelium (Figure 3-

14). Thus, the opposite effects of Hh and β-catenin signaling on ColA expression are conserved between vertebrates and cuttlefishes.

We next asked whether sustained antagonism of ColA by modulation of Hh and

β-catenin signaling would affect differentiation of Sepia cartilage. After 10 days of treatment (~ stage 28), funnel cartilage differentiation was inhibited in embryos treated with cyclopamine, SANT-1 and alsterpaullone (Figure 3-9 U-W) (also by stabilization of

β-catenin with BIO, another GSK-3β inhibitor, Figure 3-14), whereas DMSO-treated controls underwent normal funnel cartilage differentiation, including generation of a conspicuous cartilage ECM (Figure 3-9) (chondrogenesis was not affected by repression of β-catenin signaling with IWR-1 and PNU, Figure 3-14). Therefore, mesenchyme in the funnel region remained undifferentiated and had little to no cartilage

ECM, appearing to have arrested at the condensation stage after antagonism of Hh or stabilization of β-catenin (Figure 3-9 U-W).

Invertebrate and Vertebrate Chondrogenesis Share a Gene Regulatory Network

Our analysis of cartilage formation in two distantly related protostomes supports the hypothesis that a common genetic program for cartilage development exists across the Bilateria (Figure 3-9 Y). These findings indicate that despite the apparent independent evolution of cartilage tissue in vertebrates, arthropods, and mollusks, there is a deep homology of the genetic program for cartilage development. Although alternative hypotheses cannot be ruled out (Figure 3-9 Z), we propose that evolution of cartilage in the different lineages of Ecdysozoa, Lophotrochozoa and Deuterostomia involved redeployment of an ancient core chondrogenic gene network, which probably had an ancestral role in the construction of specialized ECM. As is the case for non- cartilaginous endoskeletal tissues in some invertebrate deuterostoms, such as in

cephalocordates, tunicates and hemichordates where non-cartilaginous endoskeletal tissues, like the notochord and the pharyngeal skeletons, recruited at least part of this gene network [35, 38-40]. This suggests that, similar to other developmental networks that control the architecture of the bilaterian body plan (i.e., eyes [4], appendages [4], forebrain [68], segmented body [66], central nervous system [69]), the core of the chondrogenic gene network necessary for cartilage development, was probably present in the urbilaterian ancestor. Finally, the finding that protostome and vertebrate cartilages share deep homology at the gene regulatory network and cellular levels raises the potential for the emergence of invertebrates as new model systems for the study of chondrogenesis, cartilage physiology, and regeneration.

Origin of Cartilage and the Chondrocyte in Bilateria

Within bilateria the majority of protostome and deuterostome lineages lack cartilage-like endoskeletal tissues [29]. Therefore the lack of historical continuity of this trait in the animal phylogenetic tree implies that cartilage has evolved more than once within bilateria. Alternatively, cartilage could have evolved in the last common ancestor of bilateral animals but it has been lost multiple times, which would explain its scattered phylogenetic pattern within bilateria. The conservation of core cartilage structural building blocks as well as the genetic mechanisms of chondrogenesis in members of all three bilaterian lineages (Lophotrochozoa, Ecdysozoa and Deuterostomia) suggest that the genetic program of cartilage differentiation is homologous within bilateria and was probably already present in the urbilaterian ancestor. We therefore consider three possible scenarios that could explain the phylogenetic pattern of cartilage like tissues scatter within the main three branches of bilaterians, all of them under the light of our current interpretation of a homologous cartilage gene regulatory network (GRN).

Hypothesis 1: Evolution of Cartilage/Chondrocyte Predates the Divergence of Bilaterian Lineages

In this scenario the conservation of this GRN is causally correlated with the origin of chondrocytes. Therefore chondrocytes homologous in all bilaterian cartilage made endoskeletal tissues. In that way the similarities in structure cartilage ECM and in the regulation of chondrogenesis are due to common descent. A significant implication of this model is that, since the majority of bilaterians have no trace of cartilage made endoskeletal tissues, the bilaterian lineages that lack cartilage must have lost them sometime during their evolutionary history. This hypothesis predicts that homology of

GRN of chondrogenesis in bilateria is causally correlated to homology of cartilage and chondrocyte, as a tissue and cell type respectively.

Hypothesis 2: Independent Origins of Cartilage/Chondrocyte From a Homologous Cell Type

The conservation of a homologous GRN does not necessarily imply homology at a higher level of biological organization, in this case homology of chondrocytes. It is plausible that non-chondrocyte made connective tissue could be the source of cartilage forming cells. In this scenario homologous, yet non-cartilaginous fibroblast-like cells, could have been the cell type where chondrocytes evolved in parallel in multiple lineages of bilateria. As such, the independent evolution of chondrocytes in bilateria could have been originated from a homologous fibroblast-like non-chondrocyte cell type by the independent redeployment of this homologous GRN in parallel and multiple times within bilateria. In that way this putative fibroblast-like cell type could still be present in many bilaterian lineages that show no trace of endoskeletal cartilage-like tissues. This model proposes non-homology of cartilage due to the lack of historical continuity of the chondrocyte as a cell type across bilaterians, yet homology of a putative fibroblast-like

cell type that would serve as a recurrent source for independent chondrocyte evolution within bilaterian lineages. A similar phenomena probably underlies the evolution of specialized placental cell/tissue types during the evolution of viviparity in amniote vertebrates where placental tissues have evolved independently in mammals and some reptile lineages, yet these placental tissues often evolve from homologous tissues between mammals and reptiles, extraembryonic membranes (specifically the chorioallantois) [70].

Hypothesis 3: Independent Origins of Cartilage/Chondrocyte From Non- Homologous Cell Types

A third scenario for the origin of bilaterian chondrocyte could be the independent recruitment of a homologous GRN by non-homologus cell types. As opposed to the second scenario where the GRN is recurrently activated to the same cell type, in the third model proposes that the tissue/cell type where the GRN is activated could be potentially different every time cartilage evolves. In this way, the independent evolution of cartilage in this model implies non-homology at high levels of organization but deep homology of GRN of chondrogenesis within bilateria. A frequently cited example of this kind of scenario is the independent evolution of animal appendages, showing no signs of homology at higher levels of organization yet they display deep homology of at the

GRN that controls appendage development.

Our work provided for the first time evidence of deep homology [4] at the GRN controlling chondrogenesis. Any of these three hypotheses about the origin of the chondrocyte are plausible given our data. They will be open to the scientific scrutiny in the light of new data, which will probably favor one model over the others. Nonetheless, we think the third hypothesis, independent origin of the chondrocyte arising from non-

homologous cell types, is moderately better supported by other lines of evidence than hypotheses one and two. The analysis of gene expression of members of this GRN in invertebrate deuterostoms, has demonstrated that a variety of cell/tissue types activate this GRN during the histogenesis of non-cartilaginous endoskeletal tissues, such as the mesodermal vacuolated epithelial cell of the notochord in cephalocordates and urochordates [38, 40], the mesodermal fibroblast cells of the pharyngeal skeleton of cephalocordates [39] and the endodermal epithelial cells pharyngeal skeleton of hemichordates [35]. In that way we propose that the evolution of cartilage in vertebrate deuterostoms, lophotrochozoa and ecdysozoa probably involved the recruitment of a homologous GRN to a variety of cell types (i.e. mesodermal derived connective tissue), this GRN was probably involved ancestrally in the construction of a extracellular matrix type, likely controlling the synthesis of fibrillar collagen and highly hydrophilic proteoglycans.

Figure 3-1. Animal phylogeny depicting the independent evolution of cartilage in the three major lineages of Bilateria. A, Cartilage tissues have evolved in Deuterostomia, Ecdysozoa and Lophotrochozoa. Cartilaginous endoskeletons of mouse (Vertebrata) Limulus (Arthropoda) and Sepia (Mollusca) are shown in blue in B-G. Sections through cartilage of mouse vertebra in B and C, horseshoe crab endosternite in D and E and cuttlefish funnel in F and G showing chondrocytes embedded in an abundant extracellular matrix (ECM). Aniline blue in Masson’s trichrome stain shows high content of collagen, in B, D and F nuclei are red, and alcian blue stain shows high content of acidic GAGs, in C, G and E nuclei are pink, in cartilage of mouse, horseshoe crab and cuttlefish.

Figure 3-2. Developmental series of chondrogenesis in Sepia and Limulus. A-D, In Sepia the green arrowheads mark the pre-cartilaginous cell condensation and the yellow dashed line marks the level of the basal lamina of the funnel epithelium. E-J, In Limulus the green arrowheads mark the pre-cartilaginous cell condensations and the yellow dashed line delineates the mesenchyme from the yolk cavity.

Figure 3-3. Molecular phylogenetic analysis of fibrillar collagen, Sox transcription factors (SoxC, SoxD, SoxE and SoxF) and Hyaluronan synthases in bilateria. A, Molecular phylogeny of putative bilaterian using the c-terminal propeptide Clade A fibrillar collagens (ColA) shows that that ColA genes are represented in all major lineages of bilateria (Deuterostomia-purple; Annelida-green; Mollusca-cyan; Arthropoda-red) and indicates that Sepia and Limulus (orange arrowheads) sequences belong to the ColA family. B, Molecular phylogeny of Sox genes using the HMG DNA binding domain under the WAG amino acid model of evolution. The sequences derived from Sepia and Limulus (in orange) belong to the SoxE and SoxD families (see Supplementary Table S2 for sequence access numbers). C, Molecular phylogeny of Hyaluronan synthases; outside vertebrates only putative cephalochordate HAS were recovered by BLAST searches over bilaterian transcriptomes and genomes. All trees were generated by Bayesian phylogenetic inference under WAG model of amino acid substitution. Branch support shown as percentage of posterior probabilities.

Figure 3-4. Protostome invertebrate cartilage is structurally similar to vertebrate cartilage, is ColA-based, and contains hyaluronan. A-F, ColA in situ hybridization (ISH) of Sepia A-C and Limulus D-F embryos scanned with optical projection tomography (OPT) and reconstructed in 3D using Amira . A- C OPT 3D reconstruction of ColAa ISH in stage 26 Sepia embryos maps the expression of ColAa to numerous cartilages during chondrogenesis. Embryos are shown in A ventral, in B dorsal, and in C lateral views. D-F, ColAa ISH of stage 20 Limulus embryos showing expression in the large endosternite cartilage. Embryos are shown in D ventral, in E dorsal and in F lateral views. G, Shared architecture of ColA pro-peptide between vertebrates and protostome invertebrates. In vertebrates, the von Willebrand type C domain is absent in Col2a1 but present in the other clade A collagens (Col1a1, Col1a2, Col1a3 and Col2a5). H, The cartilage ECM in vertebrates and protostome invertebrates is positive for hyaluronan as detected by fluorescent microscopy using a hyaluronan binding peptide.

Figure 3-5. ColAa and ColAb show the same pattern of gene expression in Sepia embryos. A and B Whole mount in situ hybridization (ISH) for A ColAa and B ColAb. Dorsal views. C, Ventral view of ColAb ISH showing the funnel cartilage precursors, marked by green arrowheads. D, Cryosections of these embryos reveal that ColAb is expressed in pre-chondrogenic mesenchyme (green arrowheads). Funnel epithelium marked with black open arrowhead. E-F, negative control ISH for E SoxD and D SoxE using sense RNA probes, dashed line outlines prechondrogenic cells that form the funnel cartilage.

Figure 3-6. Deep conservation of gene expression, from induction of chondrogenesis to transcriptional regulation and secretion of a cartilage ECM, during protostome cartilage development. A, Schematic drawing of the endoskeleton in Sepia hatchling, dashed line indicates the approximate plane of histological sections in relation to the funnel cartilage (red arrowheads). B-C, Funnel cartilage in hatchlings stained with alcian blue in whole mount in B and in histological section in C. Cartilage is marked by red arrowhead and the funnel epithelium by black open arrowhead. D-L, Pattern of gene expression in stage 26 Sepia embryos during chondrogenesis of funnel cartilage. Red arrowheads mark the funnel cartilage precursors in whole mounts in D, F, H and J and the pre- chondrogenic mesenchyme on histological sections in E, G, J, K and L; open arrowheads mark the epithelium above the pre-chondrogenic mesenchyme on histological sections in E, G, I, K and L. M, Schematic drawing of the endoskeleton in Limulus hatchling. Dashed line indicates the approximate plane of histological sections in relation to the endosternite cartilage (left red bracket). N, Endosternite cartilage of Limulus hatchling stained with alcian blue. Endosternite (red arrows) is located dorsal to the two ventral nerve cords (yellow arrowheads). O-W, Pattern of gene expression during endosternite chondrogenesis in stage 20 Limulus embryos. Yellow arrowheads mark ventral nerve cords in whole mounts in O and in sections in P-W; yellow arrow marks the brain in O. Pre-chondrogenic domains of endosternite chondrogenesis are delineated by red brackets in whole mounts in O, Q, S and U and by red arrows in histological sections in P, R, T, V and W; black arrowhead marks the ectoderm in V.

Figure 3-7. Chondrogenesis of multiple cartilages occurs near Hedgehog-expressing tissues in Sepia. A, Funnel cartilage in a hatchling of Sepia (black arrows) located underneath the funnel epithelium (red arrowhead). B, Double ISH of the funnel cartilage primordium at stage 26, showing the expression of ColAa in pre-cartilaginous cells (green arrowheads) and Hedgehog (Hh) in the funnel epithelium (red arrowhead). C, Fin cartilage in a hatchling (black arrows) located at the base of the fin. D, Double ISH of the fin at stage 26 showing pre-cartilaginous mesenchyme expressing ColAa (green arrowheads) next to a Hh domain (red arrowhead). E, Whole mount alcian blue stained Sepia hatchling. The white dashed outline marks the right nuchal cartilage and the yellow dashed line indicates the approximate plane of the section shown in F, the which is stained with Masson’s trichrome. G, Whole mount ISH of Hh at stage 26 showing its expression on the right and left nuchal cartilage primordia (red arrowheads). A large domain of Hh expression can also be observed in the midline (black open arrowhead) between the nuchal cartilage primordial; yellow dashed line indicates approximate level of H, a cryosection showing the expression of Hh in the epithelium of the nuchal cartilage primordium (red arrowheads) but not in the mesenchyme (green open arrowhead). I, Whole mount ISH of ColAa at stage 26 showing its expression on the nuchal cartilage primordia (green open arrowheads); yellow dashed line indicates approximate level of J, a cryosection showing the expression of ColAa in the mesenchyme (green open arrowhead) on the nuchal cartilage primordium, but not in the epithelium (red arrowheads). K, Histological section stained with Masson’s trichrome at the level of the paired statocyst cavities surrounded by cranial cartilages. L, ISH on cryosections at stage 26 embryos reveal that the brain (marked by a red *) and most of the inner epithelial lining of the statocyst cavities express Hh (red open arrowheads) and M, the pre-cartilaginous cells underneath it express ColAa (marked by green open arrowheads).

Figure 3-8. Patterns of gene expression in developing funnel cartilage of Sepia at stage 25. A, Hh is expressed in the funnel epithelium. B, ColAa is expressed in pre- cartilaginous cells. C, SoxE is expressed in the funnel epithelium as well as in the pre-cartilaginous cells, similar to D, SoxD and E, β-catenin. In all figures, red open arrowhead marks the funnel epithelium and the green open arrowheads the pre-cartilaginous cells.

Figure 3-9. Cuttlefish chondrogenesis is regulated positively by Hh signaling and negatively by β-catenin. A, In the mature funnel cartilage (stage 29-30) a thin non-cartilaginous, undifferentiated cell layer (yellow broken line outline) lies below the funnel epithelium (black arrowhead). B, PCNA immunoreactivity indicates that funnel cartilage cell proliferation is restricted to this undifferentiated cell layer, white arrowhead marks the epithelium overlying the proliferative zone. C-F, Pulse chase assay of BrdU incorporation in C 6hr exposure and in D-F 3 days of incubation after initial exposure. At 6hr PCNA/BrdU are co-localization in the proliferation zone (yellow arrowheads). After 3 days cells descendants in D-E advance into the funnel cartilage, and also F contribute to the cell pool of the proliferation zone. G-K, ISH on funnel cartilage sections shows that the proliferative non-cartilaginous cell layer express G, ColAa, H, SoxE, I, SoxD, J, Hh and K, β-catenin. L-O, Section in situ hybridization of funnel region of embryos treated with small molecule inhibitors of Hh signaling (cyclopamine and SANT-1) and the β-catenin destruction complex (alsterpaullone) after 5 days of treatment. In DMSO controls, pre-cartilaginous mesenchyme of funnel (outlined by dashed lines) shows normal ColAa expression in L, whereas ColAa is undetectable in the funnel region of embryos treated with cyclopamine in M, Sant-1 in N or alsterpaullone in O. P-S, shows the opposite pattern for β-catenin expression. T-W, Masson’s trichrome staining of funnel region after 10 days of development shows that DMSO control embryos undergo normal cartilage differentiation in T, but embryos treated with cyclopamine in Q, Sant-1 in R or alsterpaullone in S do not differentiate into cartilage and lack a collagenous matrix. X, Model of funnel cartilage appositional growth, where new cartilage derives from a proliferative layer of chondrocyte precursors expressing high levels of ColAa, SoxE, and SoxD but low levels of β-catenin, and is regulated upstream by the Hh-expressing signaling epithelium. Y, Conservation of the developmental genetic program of cartilage development between vertebrates and invertebrates. Vertebrate cartilage is represented by a sclerotome-derived vertebra in mouse, and invertebrate cartilage is represented by the funnel cartilage in Sepia. Z, Hypothesis of bilaterian chondrocyte origins, single and independent origins are contrasted in the light of our hypothesis of a homologous gene regulatory network (GRN).

Figure 3-10. Gill cartilage in Limulus is collagen-based and expresses SoxE during chondrogenesis. A, Section through gills of Limulus hatchlings stained with Masson’s trichrome. Gill cartilage is located at the base of the gills (outlined by yellow dashed lines). B, Adult gill cartilage stained with Masson’s trichrome showing a cell-rich tissue with hypertrophic cells (black arrowhead) separated by thin extracellular matrix (green open arrowheads); the gill cartilage ECM shows no aniline blue stain compared to the surrounding connective tissue; however, during embryonic development SoxE in C and ColA in D are expressed in the gill cartilage primordia (green open arrowheads), suggesting that gill cartilage is collagen-based. E, Confocal imagining of phalloidin staining and ColA in situ hybridizations showing the boundary between ColA expressing pre-chondrogenic cells (white arrowheads) and the differentiating muscles cells (white arrows) attaching to the endosternite.

Figure 3-11. Luciferase reporter assay testing the human Col2a1 cartilage specific enhancer with vertebrate and invertebrate SoxE proteins. A, NIH3T3 mouse fibroblast and B, Immortalized rat chondrocytes. All transactivated plasmids showed statistically significant luciferase activity, asterisk (*) denotes the ones that showed a statistically significant increment in luciferase activity. R.L.U. stands for relative luciferase units, relative to the Col2Enh control (Luciferase plasmid with no SoxE expression vector). From left to right: Col2Enh, luciferase control; BfSoxE, amphioxus SoxE; PmSoxE1, lamprey SoxE1; PmSoxE2, lamprey SoxE2; PmSoxE3, lamprey SoxE3; EpSox9a, hagfish Sox9a; EpSox9b, hagfish Sox9b; ScSox9, shark Sox9; DrSox9a, zebrafish sox9a; LpSoxE, Limulus SoxE; SpSoxE, Sepia SoxE.

Figure 3-12. Cell proliferation during late chondrogenesis in Sepia. PCNA immunoreactivity of mature funnel cartilage in stage 28 embryos indicates active proliferation in the chondrocytes over the entire cartilaginous element (white open arrowhead marks the epithelium). Proliferation becomes restricted to the sub-epithelial layer one stage later.

Figure 3-13. Bright field micrographs and immunofluorescence of Sepia embryos before and after treatments with the small molecule inhibitors cyclopamine, SANT-1, and Alsterpaullone, or with DMSO control. A-C, Sepia embryos at the beginning of drug treatments (stage 23-24). D, Histological sections at the beginning of the treatments demonstrate that the primordium of the funnel cartilage is already formed. The cuboidal signaling epithelium (blue arrow) and pre-cartilaginous mesenchyme (green arrow) can be identified. e-t, Sepia embryos after 10 days of treatment. E-G, control DMSO-treated embryos; J- K, cyclopamine-treated embryos; M-O, SANT-1-treated embryos; Q-S, Alsterpaullone-treated embryos. H, I, P and T PCNA staining was used to determine that drug treatments were not causing major global toxicity, compared to H DMSO control, in I cyclopamine and in P SANT-1 shows significant cell proliferation, although in T alsterpaullone treated embryos seem to boost cell proliferation compared to cyclopamine, SANT-1 and DMSO controls.

Figure 3-14. Upregulation and downregulation of β-catenin signaling has opposing effects on Sepia chondrogenesis. A, upregulation of β-catenin signaling using the GSK-3β inhibitor BIO also prevents funnel cartilage development as reveled by Masson’s trichrome. B-C, On the contrary downregulation of β- catenin signaling by inducing Axin stabilization (stabilization of β-catenin destruction complex) with IWR-1 in B and by blocking the interaction of β- catenin and Tcf with PNU in C show no effect on chondrogenesis of funnel cartilage. D-M, Cellular accumulation of β-catenin on alsterpaullone treated embryos compared to DMSO controls. β-catenin nuclear localization is not observed in D-G DMSO control embryos but after alsterpaullone treatment in I-L, β-catenin accumulates in the cytoplasm and the nucleus; two cells marked with arrowheads in J Hoechst can be followed in K β-catenin and in L Hoechst/β-catenin overlay. H and M Nuclear co-localization plots of funnel cartilage cells showing β-catenin intensities in Hoechst positive domains (nuclei). Compared to DMSO in H control, alsterpaullone in M shows higher β-catenin intensities, which demonstrates β-catenin accumulation in the nuclei. N-O, accumulation of β-catenin does not affect Hh expression in the funnel epithelium after alsterpaullone treatments in O compared to DMSO controls in N.

CHAPTER 4 EVOLUTION OF CEPHALOPOD APPENDAGES

Cephalopod Arms and Tentacles as Molluscan Evolutionary Novelties

Among mollusks, cephalopods are perhaps the best model system to understand the sudden emergence of appendages. When comparing the major lineage of mollusks

(monoplacophora, polyplacophora, aplacophora, scaphopoda, bivalve, gastropoda and cephalopoda), at first glance it is hard to imagine what all these disparate bodyplans have in common [1, 71], and only training in invertebrate comparative anatomy and embryology can help us to understand why and how these lineages are grouped into a monophyletic assemblage [1, 71, 72]. The recent advent of molecular phylogenetics has helped us to solve the puzzle of mollusk phylogenetic relationships but has done very little in helping us to understand how these disparate bodyplans evolved from a molluscan common ancestor [73, 74]. Even with our current knowledge of molluscan phylogenetic relationships, trying to imagine the last common ancestor of Mollusca is a challenge given the sudden appearance of all the major molluscan lineages during the

Cambrian explosion [2, 73, 74].

Perhaps the most dramatic macroevolutionary transformation within mollusks is the emergence of the cephalopods [71, 75]. Compared to other mollusk bodyplans, the cephalopod bodyplan shows few obvious resemblances to other mollusks. The cephalopods emerged suddenly during the Cambrian and they are equipped with an astonishing battery of evolutionary novelties [71, 75] (Figure 4-1).

Leaving aside the complex brain, camera type eyes, jet-like propulsion hydrodynamics, one of the most interesting innovations of the cephalopod body plan is the evolution of arms and tentacles. Arms and tentacles are a special type of animal appendage that

bears no counterpart within the bilateria and represent true evolutionary novelties with no clear homologs in other mollusks [71, 75, 76].

Deep Homology of Appendage Gene Regulatory Network

Animal appendages are usually portrayed as a textbook example of non- homologous characters, a byproduct of similar adaptive pressures leading to the evolution of similar morphologies. Appendages in bilaterians come in different forms, with lineage specific architectures and functional morphology [1, 2, 8, 71]. The diverse array of morphologies we can see in bilaterian appendages leaves little room for a serious consideration of the conservation of a bilaterian appendage bauplan, and by extension makes unlikely a scenario where most appendages evolved by the diversification of an ancestral urbilaterian appendage [1, 8]. That was the classic view on appendage evolution taken by comparative zoologists until the first discoveries of the conservation of the molecular genetic machinery that underlies the morphogenesis of fly and vertebrate limbs [3, 7]. It was a big surprise that this clear example of morphological convergences, say a fly leg and a mouse leg, were controlled by homologous genes, but more importantly playing similar roles during embryonic development of fly and mouse limbs [7, 77]. The evidence favoring the conservation of genetic machinery controlling fly and mouse limb morphogenesis was expanded during the last few decades, revealing that this gene regulatory network is also involved in the formation of many types of appendages in arthropods and vertebrates [77-79]. The impressive conservation of these molecular processes between arthropod and vertebrate morphogenesis could have contrasting interpretations. We can understand it as evidence of common ancestry. However, most researchers argue that arthropod and vertebrate limbs are clearly non-homologous and that the data unveils an important

principle in evolutionary biology, which is the emergence of novel traits by the independent recruitment of pre-existing genetic machinery (probably homologous) in two distantly animal lineages [1, 4, 8]. An important point into consideration is the lack of historical continuity in the animal phylogenetic tree, given that external appendages in bilaterians seem to be scattered showing no clear pattern of common ancestry.

Therefore even if the urbilaterian ancestor had appendage-like structures [80], the lack of appendages in many lineages would be a consequence of a widespread loss of these ancestral structures.

Despite the major efforts in understand the developmental genetics underlying animal appendage morphogenesis little is known about appendage development in the third major lineage of bilateria, Lophotrochozoa. In Lophotrochozoa annelids and cephalopods mollusk are emblematic examples of animal lineages bearing appendages that show no clear phylogenetic connection to appendages in other bilaterian lineages, thus they can be considered to some extend morphological novelties. In annelids, polychete worms display serial repetition of lateral appendages called parapodia [81]. It has been proposed that although some expression patterns are conserved, many others are not, suggesting that, the animal genetic programs for appendage development might not be conserved during polychete parapodia development [81].

That interpretation would restrict the animal appendage genetic network to only vertebrates and arthropods.

Here we show for the first time that cephalopod appendages evolved by the co-option of the same gene regulatory network that controls the morphogenesis of vertebrate and arthropod appendages. We used the cuttlefish as a system to

understand the developmental genetic mechanisms behind the appearance of this evolutionary novelty, but given the lack of data about appendage development in mollusks specifically and in lophotrocozoans in general, we also shed some light into the mechanisms involved in the evolution of animal appendages in bilateria. We found that, similar to arthropods and vertebrates, the main axes of cuttlefish appendages

(proximodistal, dorsoventral and anteroposterior) are controlled by the polarized expression of homologous developmental genes, such as signaling proteins, receptors and transcription factors.

Description of Limb Morphogenesis in Cuttlefishes

In contrast to the embryogenesis of most mollusk lineages, cephalopods are direct developers, meaning that instead of hatching as free-living aquatic trochophore larvae, cephalopod embryos fully develop until they hatch as miniature versions of the adult animal. This unique feature of cephalopod life history has important implications in cephalopod early development and embryogenesis in general. Early cleavage of the cephalopod egg is only superficial, at the animal pole, as opposed to the complete cleavage (holoblastic) seen in indirect developer mollusks. As a consequence, the cephalopod embryo develops restricted to the animal pole of the egg, as is the case of vertebrate eggs with high yolk content (the amniote macrolecithal egg).

Early cuttlefish embryos are then a flat sheath of tissue at the animal pole that eventually will engulf the rest of the yolk into the developing embryo (Figure 4-2).

Cuttlefishes have ten appendages, considered serial homologs, eight arms and two tentacles (Figure 4-1 and 4-2). Early during cuttlefish development, arm and tentacle primordia can be observed as a higher density of cell domains that eventually rise up as small protuberances (Figure 4-2A). They develop over the

embryonic foot region (a homologous region across Mollusca, where other mollusks develop their foot), which indicates cephalopod limbs are a lineage specific specialization of the molluscan foot [76]. These protuberances elongate distally, transforming into small arm/tentacle buds (Figure 4-2B). At subsequent stages arm/tentacle buds elongate further and start to develop the sucker buds over their ventral surface. Around stage 25-26, the differentiation between arms and tentacles is clear, with tentacles are longer and only display sucker buds over a small distal patch, rather than uniformly over their whole ventral surface as is the case with the arms

(Figure 4-2C and D).

Gene Expression and the Control of Axis of Growth in Cuttlefish Limb

Anteroposterior Axis

The polarized expression of signaling proteins in vertebrate and arthropod limbs is essential for the proper establishment of the anteroposterior axis of their limbs. One of the better studied of these signaling factors is from the Hedgehog family. In fruitflies

(and other arthropods; also in onychophorans) and mouse, the expression of Hedgehog

(Hh; Sonic Hedgehog or Shh in mouse) is restricted to the posterior side of the limb.

Misexpression of Hh in Drosophila and Shh in mouse and chickens generate a mirror image duplication of the limb along the anteroposterior axis [82-84]. Moreover, genetically engineered reduction or complete elimination of polarized Shh expression in mouse limb buds causes gradual digit loss along the anteroposterior axis to almost complete limblessness, respectively [85, 86].

We found that in the cuttlefish Sepia officinalis, Hh is expressed very early during the first stages of limb budding (Figure 4-3). Surprisingly, its expression is polarized, showing a restricted domain of expression at the anterior region of each limb bud. Later

in development, the pattern of Hh expression is shifted to the center of the limb marking the developing brachial nerves of each limb. The early-polarized expression of Hh in S. officinalis suggests that, similar to vertebrates, arthropods, and onychophorans (sister group of arthropods bearing unsegmented appendages called lobopodia [87]), cephalopod appendages regulate the anteroposterior axis by Hh signaling.

Proximodistal Axis

As arthropod and vertebrate limbs start to growth out from the main body axis, signaling proteins and transcription factors are transcribed at the distal most part of each limb bud. These developmental genes are essential for controlling the directed elongation of limbs distally from the main body axis and specify proximal and distal identities [77-79, 88]. We asked if in S. officinalis a similar set of distal markers during limb development is conserved with arthropod and vertebrate limb development.

Wnt signaling is central to distal outgrowth in many animal appendages and to main body axis elongation in vertebrates, short germ arthropods, and in planarian tail specification during regeneration [89]. Moreover, during lobopodia development in onychophorans, there is also an activation of a whole battery of wnt signaling genes marking the distal tip of the limb buds [90], as is the case in vertebrate and arthropod limb development [89, 91]. We cloned an analyzed the expression of some Wnt ligands,

Wnt receptors and transcription factors that act as effectors of activated Wnt signaling pathway, during S. officinalis limb development. We found that the Wnt ligands Wnt1,

Wnt5 and Wnt7 are expressed in a broad distal domain, with each Wnt reaching different proximodistal levels along the limb (Figure 4-4). Similarly, the expression of the

Wnt receptor Fzd in S. officinalis limbs also shows distal expression (Figure 4-5), reaching a similar proximal level than Wnt7. The Wnt co-receptor Lrp show strong

expression in the limbs as well, but it is more ubiquitous rather than distally restricted

(Figure 4-5). We also assayed the expression of the transcription factor that acts as an effector of the transcriptional response triggered by the activation of Wnt signaling, Tcf, and we found that its expression is restricted to a distal domain of S. officinalis limb buds (Figure 4-5). These data strongly suggest that the distal domain of S. officinalis limb buds activates the Wnt signaling pathway and might be playing a role in proximodistal outgrowth.

In arthropods and vertebrates, limb initiation, distal outgrowth and proximodistal patterning is also controlled by a set of transcription factors that are expressed at different levels in the proximodistal limb axis. The genes Dll (Dlx genes in vertebrates) and Sp8 are expressed in the distal most part of the limb buds and are critical master regulators of limb outgrowth and distal identities. On the other hand, genes like

Homothorax (Hth; Meis in vertebrates) and extradendicle (ext; Pbx in vertebrates) are expressed at the proximal base of the limb buds and are important for the specification of proximal limb identities [77, 79, 92]. Other genes such as Dachshund (Dac) show either distal expression or are expressed between the proximal domain and the distal domain genes [77, 79]. The partitioned expression of these transcription factors is important for the proper layout of limb structures and compartments along the proximodistal axis as well as the regulation of limb outgrowth [77, 88].

We isolated the cuttlefish homologs of genes expressed along these limb compartments over the proximodistal axis and studied their expression during limb development. Our work in S. officinalis reveals that cephalopods have recruited these genes during limb development, partitioning the limb into proximal and distal domains.

We found that Dll and Sp8 are expressed at very early stages of limb budding. Dll is first broadly expressed over the entire embryonic foot region and slightly stronger at the distal tip of each limb bud (Figure 4-6). Later, around stage 19-20, the expression in the interlimb regions is almost completely shutdown but limb the buds show sustained Dll expression until stage 24, when the expression is restricted only to a very small distal spot (Figure 4-6). Sp8 and Dac on the other hand show no broad foot expression

(Figure 4-7), but instead they show a very strong limb specific distal expression at early budding stages, which is sustained latter in development, with an expanded domain proximally in the case of Sp8 (Figure 4-7). We also examined the expression of proximal genes and found thatin cuttlefishes, as in vertebrates and arthropods, Hth shows a very specific proximal expression during limb development (Figure 4-8), marking a proximal domain of the limb.

Taken together, cephalopod limbs are patterned in their proximodistal axis in a similar way as arthropod and vertebrate limbs, with Wnt signaling activity at their distal tip, as well as the expression of other distal markers as Dll, Sp8 and Dac, but also the proximal expression of Hth. This suggests that the appendage program controlling the proximodistal axis is conserved in arthropods, vertebrates and cephalopod limbs.

Dorsoventral Axis

The dorsoventral axis in vertebrate and arthropod limbs is also controlled by the polarized activity of different signaling pathways in the ventral and dorsal part of the limbs. In arthropods and vertebrates, Wnt and Bmp signaling pathways control the dorsoventral axis. In vertebrates, however, Wnt7a controls dorsal patterning and Bmp is involved in the ventral identities of the limb bud (which triggers the expression of engrailed in the ventral ectoderm) [93, 94], in arthropods on the other hand Wnt

signaling controls the ventral side and Dpp (invertebrate homolog of Bmp) regulates ventral identity [95]. Misexpression of Wnt7a or engrailed (downstream of Bmp signaling in the ventral limb ectoderm) in mouse limbs causes a change in the identity of the limb dorsoventral axis [93]. We isolated Bmp2/4, the homolog of Dpp, in S. officinalis and found that its expression is restricted to the dorsal side of the developing limb buds

(Figure 4-9). Although we could not find a Wnt ligand expressed to only over the ventral side, we did find that the extracellular Wnt represor Sfrp (secreted frizzled related protein) is expressed only in the dorsal pole (Figure 4-10). Therefore, the dorsal expression of Sfrp suggests that there is probably a polarized activity of the Wnt signaling pathway ventrally, mediated by the depletion of Wnt ligands in the dorsal domain. Additionally, the slightly stronger expression of Tcf (the effector of Wnt mediated transcriptional responses) ventrally and distally, is in marked contrast to the expression of Sfrp, giving support to the hypothesis of Wnt signaling as having an important role in ventral identity. Taken together, the dorsoventral axis of cephalopod limbs is probably controlled by the polarized activities of Bmp and Wnt signaling, as in vertebrates and arthropods. Furthermore, like arthropods, cephalopods seem to use

Wnt signaling to specify ventral identities but Bmp signaling for dorsal identities (Figure

4-11).

Developmental Basis of a Morphological Novelty

The evolution of cephalopod body plan is a significant departure from the molluscan body plan. Explaining its origin is one of the most challenging but interesting puzzles in evolutionary morphology and zoology. The cephalopod body plan is rich with evolutionary innovations that were critical for their evolutionary success [75]. One of those innovations is their arms and tentacles, which represent a clear evolutionary

novelty within the Mollusca. Other than their proposed embryological origin from the embryonic molluscan foot primordia [76], there are no clear homologous structures that could be regarded as an ancestral homolog in other mollusk as a possible evolutionary precursor of cephalopod limbs. Here we demonstrated for the first time that the origin of cephalopod limbs involved the recruitment of an appendage developmental program that also was responsible for the formation of arthropod and vertebrate limbs.

Our results suggest that cephalopod limb morphogenesis shares important developmental genetic similarities with arthropod and vertebrate limbs despite of their profoundly dissimilar morphologies and clearly independent phylogenetic origins. During

S. officinalis limb formation, polarized patterns of gene expression suggests that cephalopods establish the limb axes coordinate system using the same genetic machinery as arthropods and vertebrates. Furthermore, for the first time there is evidence that the appendage development program is recruited for the evolution of appendages in any lophotrochozoan.

A Conserved Appendage Developmental Program

The conservation of this developmental genetic machinery to pattern many different animal appendages across the three major branches of the bilaterian phylogenetic tree (Deuterostomia [vertebrates], Ecdysozoa and Lophotrochozoa) could potentially be interpreted as evidence of homology of all animal appendages, regardless of their form or function. However, this would suggest that the urbilaterian ancestor was probably a more complex animal than previously conceived, bearing locomotory appendages that along bilaterian evolutionary history were then lost in many of the extant lineages that lack these structures [80]. Despite the attractiveness of this idea, the data presented here suggest that it is over simplistic and highly unparsimonious. We

propose that animal appendages have evolved multiple times and independently within bilateria. Therefore they are not homologous structures, but they evolved by the parallel recruitment of a homologous gene regulatory network [4]. Together, the data presented here support the hypothesis that animal appendages display deep homology at the level of the gene regulatory machinery controlling limb outgrowth but not at the level of the structure [4].

The evolution of limbs in the three major branches of the bilaterian tree by the co-option of a homologous gene regulatory program is surprising. Here I propose the addition of distal activation of wnt signaling as a major regulator of limb proximodistal outgrowth and as an integral part of the appendage developmental program. The involvement of Wnt signaling in the control of distal outgrowths in other developmental context such as the elongation of the main embryonic axis [89], polyp evagination in Hydra [96] and amniote genitalia [97], further support the promiscuity of these regulatory networks and the likelihood to be co-opted to regulate the formation of novel structures in evolution. This gives support to an intriguing hypothesis by Minelli [8] that proposes animal appendages can be regarded as duplications of the primary axis, as is clearly the case with the astonishing parallel of the developmental mechanisms that control of polyp evagination and tentacle formation in Hydra [96]. Finally, an alternative scenario is the co-option of this network from the centralized nervous system. It has been proposed that the origin of this regulatory network dates back to its assemblage for the patterning of the brain of bilaterians as some genes (i.e. Sp8, Dll,

Hth, Dac) partition the embryonic brain of protostome and deuterostoms as they do during the establishment of the proximodistal axis of animal limbs [98].

My results indicate that the developmental basis behind limb evolution in cephalopods is the recruitment of an ancient gene regulatory program. Arms and tentacles in cephalopods are truly evolutionary novelties; however, they share deep homology with the appendages of arthropods and vertebrates in the developmental mechanisms used to pattern their proximodistal, dorsoventral and anteroposterior axes.

The study of cephalopod limb evolution and development further emphasis a common principle in organismal evolution, which is the redeployment of pre-assembled modules

(appendage program) into a new developmental context for the emergence of a morphological innovation.

Figure 4-1. Figure General morphology of S. officinalis, rendered by OPT imaging. A, whole body at different views and at the far left, the internalized tentacles can be visualized when the posterior most arms, funnel and siphon become translucent (in orange). B, ventral view of a typical arm covered with suckers. C, lateral view showing that arms have a clear dorsoventral polarity with smooth dorsal surfaces and ventral surface covered by suckers.

Figure 4-2. Bright field images of Sepia officinalis embryos at different stages of limb development. At early stages S. officinalis is a flat embryo where limb buds start to as increased cell density zones to later for a discernible limb bud. Not all limb buds develop in synchrony. At later stages the embryo rises and other parts of their body plan star to differentiate, such as the mantle, eyes and fins. In all figures red arrowheads mark arms and yellow arrowheads tentacles.

Figure 4-3. Embryonic gene expression of Hh during limb development. A, dorsal view showing Hh expression in arms and tentacle primordia. In B, a posterior high magnification in posterior view. C, a higher magnification show that Hh expression is polarized to an anterior domain of the limb bud (dashed line outline). D, dorsal view showing that Hh is still expressed at an anterior domain in latter stages of budding. E, lateral view. F, high magnification in dorsal view showing the anterior polarization of Hh expression. In all figures black arrowheads indicate arm primordia and orange arrowhead tentacle primordia.

Figure 4-4. Embryonic gene expression of Wnt5, Wnt1 and Wnt7 during limb development. Black arrowheads indicate arm primordia and orange arrowhead tentacle primordia.

Figure 4-5. Embryonic gene expression of Fzd, Lrp and Tcf during limb development. Black arrowheads indicate arm primordia and orange arrowhead tentacle primordia.

Figure 4-6. Embryonic gene expression of Dll during limb development. Black arrowheads indicate arm primordia and orange arrowhead tentacle primordia.

Figure 4-7. Embryonic gene expression of Dac and Sp8 during limb development. Black arrowheads indicate arm primordia and orange arrowhead tentacle primordia (on the top images only the right side is labeled).

Figure 4-8. Embryonic gene expression of Hth during limb development. Black arrowheads mark arms, note the proximal restricted expression in the limbs (dashed line outline).

Figure 4-9. Embryonic gene expression of Bmp2/4 during limb development. A and C show lower magnification, black arrowheads indicate arm primordia and orange arrowheads tentacle primordia. In B and D higher magnification at a ventral view in B and lateral view in D show the dorsal location of Bmp2/4 expression in the developing limb buds (dashed line outlines).

Figure 4-10. Embryonic gene expression of the Wnt inhibitor Sfrp during limb development. A, dorsal view of the embryo. B, high magnification of first arm (dashed line outline) showing the absence of Sfrp expression at the distal tip of developing limb buds, suggestive of active Wnt signaling (green arrowhead). C, shows the absence of expression in the ventral region of the limb buds (dashed line outline).

Figure 4-11. Summary of pattern of gene expression along the three axes of growth in cephalopod limb buds.

CHAPTER 5 LIMB RE-EVOLUTION IN SNAKE-LIKE LIZARDS

Phylogenetic Patterns of Limb Loss in Squamate Lizards

The evolution of the tetrapod limb has been subject to extensive modifications of its architecture, meeting highly adapted functional demands in many lineages and represent one of the most striking examples of the evolution of functional morphology yet conserving a general anatomy [99, 100]. Despite its obvious advantages, limbs have been secondarily lost many times during tetrapod evolution in all the major tetrapod lineages (amphibians, sauropsids and mammals) [99, 100]. Squamate reptiles show one of the most dramatic evolutionary cases of limb loss [100-103], in terms of both the high frequency of independent losses and the diversity of intermediate morphologies, all in a single lineage [100, 103]. Within squamates, the genus Bachia is a stereotypic example of the impressive diversity of limb evolution in squamates as a whole [104,

105]. In Bachia, we can find completely limbed and pentadactyl species together with almost completely limbless species. Additionally, in Bachia there are also species that show different patterns and degrees of limb loss with species that only lack a few digits in their hindlimbs and/or forelimbs and others that are legless but bear forelimbs with variable degrees of digit loss, as well as species where forelimbs are absent but show variable degree of toe number [104-106]. The impressive breath of limb morphological diversity in Bachia can be only paralleled, although to a slightly less degree, by other squamate lizard lineages within the family Scincidae [103]. Although it would be tempting to explain the evolution of limb loss in squamates by assuming the transformation of ancestral limbed forms in an ordered series of forms until reaching complete limbless, by following some kind of morphocline polarity, molecular

phylogenetic analyses rather suggest a more complex phylogenetic pattern in many squamate lineages [107-109]. Molecular phylogenetic analyses not only demonstrate the fundamental logical fallacy of a priori establishing morphocline polarities, but also give support to a surprising evolutionary phenomenon, the re-evolution of complex characters [107-109].

Ancestral character reconstruction in these molecular phylogenies show that complex limb morphologies (complete limbs) are not necessarily always ancestral,. In many cases, well formed limbs are the derived condition [107-109]. In those cases, some well-formed limb forms are deeply nested in lineages where there are only limbless forms. In such cases, ancestral character state reconstruction indicate that it is more likely that well-formed limbs have then been secondarily re-acquired [107-109]. An alternative explanation, however, is that limb loss is unidirectional and well-formed limbs must represent the ancestral state, thus the rest of the phylogenetic landscape has to be explained by multiple cases of limb loss [110, 111].

Indeed, in such cases some researchers have favored the position that well- formed limbs should be regarded as a persistent ancestral state, rather than a secondary re-acquisition [110, 111]. Their position is based on two strong assumptions:

(1) the high probability of limb loss and (2) the implausibility of complex character re- evolution.

Limb Loss and Re-evolution in Bachia

The genus Bachia is a lineage of small and body-elongated burrowing lizards with an impressive degree of variation in limb morphology [104-106]. Their enigmatic pattern of limb evolution has started to be elucidated by recent molecular phylogenetic analyses [108, 109]. The phylogenetic ancestral character state reconstruction analysis

of limb loss in Bachia revealed a very interesting phenomenon digits and entire limbs that were previously lost in Bachia evolution seem to be secondarily re-acquired in some lineages [108, 109]. As stated above, an alternative but less parsimonious explanation to character reversal would be that limb loss happened several times in

Bachia, and therefore those species with completer limbs rather conserved the ancestral state [110].

Interestingly, if we consider that limb loss is likely to occur in squamates, then a paradox emerges in those cases where limb re-acquisition is considered a forbidden morphological transformation: why is the “complete limb state” conserved in the internal braches of the phylogeny for millions of years if the probability of limb loss is so high?

In phylogenetic-based character reconstruction the preference of one type of transformation over the other depends on the ratio of penalty cost between them (e.g. 3 to 1 favoring limb lost), which in turns constrains the reversibility of character transition

[107-109, 112-114]. In Bachia, for instance, it is necessary to use a strongly biased ratio of character transformation in favor of digit/limb lost to preclude character reversals, giving strong support to digit and limb re-evolution [108, 109]. On the other hand,

Brandley and collaborators [107] used a phylogenetic based Bayesian ancestral reconstruction analysis to infer the rates of loss vs gains in a “supertree” of 258 squamate species and they found that there is a strong bias to digit/limb lost over gains.

However, regardless of the strong bias they found strong statistical support for at least 6 events of digit and limb re-evolution of previously lost structures, including the re- evolution of pentadactyl limbs from limbless ancestors (Bipes and Scelotes).

Limb re-acquisition in the genus Bachia and in other squamate reptiles is a violation of Dollo’s Law. Although character reversal in phylogenetic reconstructions has been porposed in discussions of simple characters [108, 109, 115], the situation is different when discussing the plausibility of re-evolution of complex character, such as a complete limb from a limbless ancestor [110, 115]. However, in principle there is little from the study of phylogenetic patterns of character evolution that can help us to understand whether such a morphological evolutionary transformation is plausible or not. One of the main limitations in this kind of discussion is the “adult-centric” approach of character transformations in evolution. Character transformations can only be understood as terminal transformation of a fully developed structure (limbless to limbed). However, all morphological characters are formed de novo each generation by the execution of developmental programs during embryogenesis. Morphological changes in evolution happen by modification of the embryonic morphogenesis of morphological traits. Thus, a developmental approach is needed in order to understand whether such transformation is possible or not, by first understanding the mechanisms behind phenotypic transformations. In the case of Bachia, the adult-centric perspective of limb evolution in Bachia phylogeny would lead us to the conclusion that limb re- evolution is something extremely unlikely to happen. Nonetheless, here we show that the plausibility of limb re-evolution in squamates cold be likely to happen, that when we take into account how limbless species of Bachia have lost their limbs. We found that the limbless species Bachia bicolor does conserve a complete embryonic limb that even starts to develop digits, but this complete embryonic limb is later degraded as development progress.

Therefore Bachia embryonic limbs demonstrated that in order to understand the evolution of limb loss in this lineage, we have to understand organismal evolution as the historical transformation of life cycles (organism development) rather than focusing our explanations of morphological transformations as mere terminal modifications of adult morphologies.

Limb Morphology of Bachia bicolor

To understand the possible mechanisms behind limb re-evolution in the genus

Bachia, we studied the embryonic mechanisms behind limb loss in B. bicolor, a species that has a pivotal phylogenetic position in the Bachia phylogeny [108]. B. bicolor is placed at the base of a lineage that shows secondary hindlimb vestigialization as its ancestral character state [108]. Deep inside this lineage, there are at least three species that underwent complete hindlimb re-evolution from a very likely ancestral limbless state

[108]. Therefore, the study of limbs loss in B. bicolor can be instrumental in our understanding of how complete limbs might re-appear in some species of Bachia

(Figure 5-1).

B. bicolor show miniaturized but well formed forelimbs bearing 4 digits (Figure 5-

2A-C). In contrast, their hindlimbs are vestigial and barely visible externally (Figure 5-

2A, E and F). We did whole mount alcian blue staining of late stage B. bicolor embryos and reconstructed them tridimentionaly using optical projection tomography (Figure 5-

2D and G). We found that hindlimbs show remnants of a small femur and nodular tibia and fibula (Figure 5-2G and H). Distal to the malformed tibia and fibula there is a cartilaginous mass of ambiguous identity (Figure 5-2H).

Embryonic Development of B. bicolor Hindlimbs

We studied a series of B. bicolor embryos using scanning electron microscopy, starting from early stages of hindlimb bud development to stages where the its adult morphology achieved (Figure 5-3). We discovered that limb development in B. bicolor unexpectedly progresses to very late stages of limb development but then displays a marked degradation of the distal tip of the limb (Figure 5-3). A closer analysis of early stages of B. bicolor hindlimb buds shows that they form a normal apical ectodermal ridge (AER) (Figure 5-4A and B), which is a distal columnar epithelium that divides the dorsal ectoderm from the ventral ectoderm of the limb. The AER is an essential signaling center of the developing limb that controls proximodistal outgrowth of the limb by the expression of FGF ligands [88, 116]. Experimental ablation of the AER at different developmental stages of limb development generates significant distal truncation of limb skeletal elements, earlier ablations cause more distally truncated limbs [117, 118]. However, as the hindlimb starts to reach a paddle-shape stage the

AER is lost from the posterior part of the limb (Figure 5-4C to F). The AER is also responsible to sustain expression of Shh, which encodes a signaling protein in the posterior mesenchyme at the zone of polarizing activity (ZPA). SHH, in turns, stimulates expression of Fgfs, which encode FGF ligands that are secreted from the

AER , establishing a positive feedback loop [88]. This finding suggests that the posterior loss of the AER would be an indication of a premature breakdown of the AER-

ZPA feedback loop in B. bicolor hindlimbs.

B. bicolor Has an Autopod

Surprisingly, B. bicolor hindlimb development progresses into the formation of a paddle-shape limb and even far enough to clearly differentiate a digital plate, or an

autopodium (digit forming domain) (Figure 5-4G to K). This result was unexpected taking into account the high degree of limb truncation in B. bicolor and the complete absence of digits in the hindlimb. The digital plate starts to degenerate in a posterior to anterior direction until the autopod is reduced to an sphere-shaped mass attached to the rest of the limb by a stalk (Figure 5-4K). This mass eventually degenerates completely, leaving no trace of an autopod in the late embryo (Figure 5-3G).

We then reasoned that if B. bicolor hindlimbs reached such stage of limb development, then they probably activated atupodial gene expression essential for the development of digits. The distal expression of Hoxd13 and Hoxa13 genes are hallmarks of the digit-forming domain in the tetrapod embryonic limb [119]. Genetic knockouts of these two genes causes significant reduction in digit length and if double knockouts are generated then the complete all digits are eliminated [120]. Their transcriptional regulation during embryogenesis has demonstrated that distal expression of these Hox genes during limb development is controlled by autopod specific enhancers, which position them as master regulators of the autopod [121, 122]. We cloned and analyzed the expression of Hoxd13 in B. bicolor embryos and found that, as it was the case for their forelimbs (Figure 5-5A and B), there was strong distal expression of Hoxd13 in the hindlimb (Figure 5-5F and G). This result unambiguously demonstrates that the limb digital plate in B. bicolor is an autopod or a digit-forming domain.

We next asked if the autopod in B. bicolor has the developmental potential to form digital elements. We sectioned forelimbs (Figure 5-4C and D) and hindlimbs

(Figure 5-4H and I) at digital plate stages and used peanut aglutining (PNA) staining to

reveal the presence of pre-chondrogenic digital condensations before the deposition of cartilage extra cellular matrix. We found that, in B. bicolor hindlimbs, a femur, tibia, and fibula can be clearly identified (Figure 5-4I). To our surprise, the digital plate in hindlimbs of B. bicolor develops at least three pre-chondrogenic digital condensations, putatively identified as toes 3, 4 and 5 (Figure 5-4I) based on the stereotypical organization of skeletal development tetrapods and the relative positions of these condensation in the limb [123]. We performed whole mount alcian blue staining for the detection of cartilage ECM at a later stage and we discovered that at least two of the three digital pre-chondrogenic condensations differentiate into cartilaginous elements

(Figure 5-4J), which later in development leave no trace of their existence. Taken together, our developmental analysis of B. bicolor hindlimbs demonstrates that they not only form a digit-fi B. heteropa trinitalis, there are no other pentadactyl species of

Bachia. Interestingly, despite having only four fingers, we found that n B. bicolor forelimbs, there is a transitory digit one condensation (Figure 5D) that never differentiates further into cartilage (Figure 5E), similar to our findings in B. bicolor hindlimbs. This situation is highly reminiscent to chicken wings, where forelimbs are show five digit condensations but adult wings only differentiate three digits [124].

Therefore, B. bicolor forelimbs are pentadactyl as embryos but conserve only four fingers as adults.

Phylogenetic Patterns Hide Ontogenetic Processes

Our developmental analysis of B. bicolor hindlimb and forelimb development demonstrates for the first time that a legless lizard species as adult develop almost fully formed embryonic limbs transitorily. In the light of this new data the plausibility of complete hindlimb re-appearance seems more likely. The study of hindlimb

development in a legless species demonstrates how insightful this ontogenetic approach can be in our discussion about the directionality of character transformation.

Phylogenetic based character reconstruction usually considers adult character states, thus character transformation in the phylogenetic tree can be easily misinterpreted as terminal transformation of adult character states in evolution.

Therefore, the reversal of adult character states, such as well-developed limbs arising from limbless ancestors, can be misinterpreted as reassembling a complex structure de novo which seems to be a counterintuitive phenomenon, a complex structure rising from the ashes.

However, as evidenced in B. bicolor, the mechanism underling such morphological transformations in evolution are executed during embryonic development. Thus, our understanding of the embryological and genetic basis of character morphogenesis and character degradation/loss are very important in order to consider the plausibility of character re-evolution in a particular lineage. Therefore, in the case of Bachia limb evolution this knowledge will allow us to interpret character reversals as either, (1) de novo re-evolution of a lost structure, thus an unlikely event, or

(2) the rescue of a transitory and cryptic structure in ontogeny. We think our data clearly supports the second scenario, where despite the limblessness in the adult species, the developmental program in limb reduced species could be conserved as a cryptic feature in ontogeny, and thus the developmental and genetic tools can switch back on again, potentially rescuing the transitory limb and then forming an entire hindlimb all over again.

Thus, the re-evolution of hindlimbs in Bachia has broader implications. If the developmental and genetic mechanism underling the morphogenesis of lost characters in evolution can be conserved despite of character lost [115], then the occurrence of character re-evolution would be more likely. At the same time, it could explain the apparent erratic patterns seen in some complex cases of phylogenetic character reconstructions, in which characters seems to be flickering between gain and loss in different branches of the tree [112, 114, 125, 126].

One dramatic example can be observed in the beetles of the genus

Onthophagus. Moczek and collaborators [127] showed that the gain and loss of adult horns in this genus is apparently very complex, requiring multiple loss and reacquisition of adult horns when only the adult morphology is considered. Nonetheless, if early ontogenetic stages are considered, in this case the pupae, then a different picture emerges. In the genus Onthophagus all the hornless adult beetles studied so far develop from fully horned pupae, but pupal horns are resorbed before turning into adults. Therefore, as with the hindlimbs of Bachia, some horned adult beetles probably arise from hornless adult beetles ancestors [127]. Thus, the mechanism behind the reacquisition of the horns in this case is not necessarily a de novo origin of the structure, but the available data indicates that it could involve the prevention of pupal horn resorption. Thus, the conservation of pupal horns bears an important ontogenetic potential, and is probably the underlying mechanism behind the flickering of horned and hornless adult morphologies in an apparent erratic phylogenetic pattern.

Based on the data presented above, I propose that Bachia hindlimb re- evolution is plausible and here I provide evidence for a mechanistic explanation of this

evolutionary transformation. I suggest that early in Bachia evolutionary history, hindlimbs were lost in the adult forms, but as is the case of B. bicolor, hindlimbs were not completely lost during embryonic development and they have persisted as cryptic structures in the ontogeny of legless species during their evolutionary diversification, but later rescued in derived species (Figure 5-6). In those legless species, embryonic hindlimbs are arrested in development, probably as a consequence of a late breakdown of the AER-ZPA feedback loop but are still able to reach an autopod stage. Legless species then conserve an autopod of reduced size but, as in B. bicolor, it is probably undergoes early stages of toe differentiation and is later degraded. This transitory autopod would have the developmental potential to be rescued and give rise to a fully formed limb with terminal toes (Figure 5-6).

The mechanistic explanation of reversibility of complex character transformations presented here eliminates the constraint of unidirectional character modification in evolutionary morphology. Accordingly, limbless or limb reduced ancestors probably could give rise to descendants with less reduced limbs (digit re- acquisition) or pentadactyl limbs (whole-limbs), respectively. My results show that ontogenetic studies can reveal the presence of transitory structures that are frequently disregarded in common phylogenetic analyses that are mainly focused in adult morphologies. These transitory structures can occur in many forms, depending on the nature of the character considered to undergo an evolutionary transformation.

An important factor to consider in Bachia limb re-evolution is that many times when Bachia species display hindlimb vestigialization, they still conserve miniature well- formed forelimbs. Although, we do not understand the possible locomotory selective

pressures or neutrality, miniature limbs can bring to a body of an elongated burrowing lizard well-formed forelimbs, indicating that the limb developmental program is conserved in those legless Bachia species. Furthermore, if we also take into account the evidence that most of the regulatory networks used for limb development are also used in the morphogenesis of many other organs, then the limb developmental program in legless species is, to some extent ,protected from erosion.

Figure 5-1. Phylogenetic analysis of genus Bachia by (modified from Ref [108]). Maximal likelihood reconstruction of ancestral states are illustrated in black for complete hindlimbs with toes and in white hindlimb vestiges. The morphology of forelimbs and hindlimbs of each species can be seen to the left (drawings modified from Ref [104, 106]).

Figure 5-2. Morphology of forelimbs and hindlimbs of Bachia bicolor. A, general body morphology of the legless B. bicolor, showing their elongated body and in B, a higher magnification of the miniaturized head and forelimbs. C, dorsal view of the hand of B. bicolor showing four fingers. D, dorsal OPT view of an alcian blue stained forelimb in a late B. bicolor embryo showing the developing cartilaginous skeleton. E, ventral view of the cloacal region (black bracket) of B. bicolor showing the small hindlimb vestiges (yellow arrowheads). F, lateral view of a hindlimb vestige. G, lateral OPT view of an alcian blue stained late B. bicolor embryo showing the reduced hindlimb skeleton, non-skeletal tissue in green. H, OPT isolated hindlimb skeleton showing the small but elongated femur proximally, nodular shaped tibia and fibula and ambiguous distal elements labeled with a white bracket

Figure 5-3. Hindlimb development of B. bicolor as seen by SEM imaging in an embryonic series. A to G bright field images of a series of embryonic stages (red arrows mark hindlimbs), below each embryo (A’ to G’) a high magnification of the corresponding hindlimbs using SEM imaging can be observed.

Figure 5-4. Embryological evidence of posterior AER loss and a transitory digital plate in B. bicolor ontogeny. A, at early stages of the hindlimb bud there is a clear AER seen by SEM and in B, by hematoxilin and eosin stained histological sections. C, at onset of paddle shape limb stage the AER is lost posteriorly (red zone) but not anteriorly (yellow zone). D, is a higher magnification of the limb bud edge, showing the anterior AER (yellow arrowhead) but its absence posteriorly (red arrowhead), and indicating the approximate level of histological sections seen in E (yellow dashed line) and F (red dashed line) that confirm the SEM observations. G, at later stages B. bicolor develops a clear digital plate, pseudocolored as a red domain. H, the digital plate is starting to degrade at the posterior margin of the limb, in I and J a significant reduction of the digital plate is observed and in K only an small sphere of tissue is still attached to the rest of the hindlimb.

Figure 5-5. Autopod development in forelimbs and hindlimbs of B. bicolor. A, forelimb digital plate seen by SEM imaging and showing autopod specific expression of Hoxd13 in B. In C, a bright image of a late digital plate at the stage used for PNA staining in D. D, PNA staining show that despite having only four fingers as adults, B. bicolor fore limbs are pentadactyl, with a transitory condensation for digit I. E, alcian blue staining show that digit I condensation does not chondrify. F, hindlimbs of B. bicolor as seen by SEM, showing a clear early digital plate show autopod specific Hoxd13 expression in G. H, at a later stage of development the digital plate show at least three digit condensations in the autopod in I. Later, OPT imaging of alcian blue stained embryos in J, shows that only two of these condensations chondrify.

100

Figure 5-6. A hypothetical scenario proposed for limb re-evolution in Bachia, based in the hypothetical rescue of transitory digits and an autopod as seen in B. bicolor. In this hypothetical phylogeny the ancestral state of complete hindlimbs is illustrated in blue and hindlimb vestigialization in red, with the external morphology on the right. Ancestrally hindlimb morphology is complete but later there is an event of hindlimb loss that is reversed in two independent lineages. Our model propose that while the ancestral ontogeny was modified early in the phylogeny, the cryptic existence of a transitory autopod provides to this lineage with the flexibility of rescuing it later in evolution and in that way seamlessly re-acquire the ancestral morphology.

101

CHAPTER 6 CONCLUSION

When compared to molecular evolution, our knowledge on the mechanisms underlying how morphology evolves is still limited [5, 9]. We have accumulated an impressive catalog of the astonishing morphological diversity in animals that, together with robust phylogenetic analyses, have lead to a better appreciation of the patterns of morphological changes during animal evolution. For comparative biologists, this information is fundamental to the progress in understanding how morphology evolves, but it provides very little mechanistic insight on the processes involved behind morphological transitions and in the evolution of morphological novelties [2, 5, 7, 9].

To achieve a better understanding of how morphology evolves, we need to incorporate mechanistic explanations into our phylogenetic frameworks, integrating patterns and processes of evolutionary change. Evolutionary developmental biology is perhaps one of the most exiting fields into that direction, int hat phylogenetics, genomics, embryology and cellular processes are integrated to fully appreciate the evolution of the organism as a whole.

Morphology is created and modified during embryonic development, there is no physical transmission of morphological traits to the descendants, and those traits have to be formed de novo each generation. Thus, given our current knowledge of molecular biology and genetics, it is reasonable to focus our attention in the study of the cellular and developmental genetic parameters that might control morphological transformations in development and by extension in evolution. This mechanistic and organismal perspective has been fruitful, giving us some of the most amazing detailed examples on how organisms change their morphology and life history [128-133].

102

One interesting finding from evolutionary developmental biology is that morphological characters are embryonically regulated by gene regulatory cassettes, such as signaling pathways and transcription factor networks [5, 9]. These cassettes are not only highly promiscuous, as they are not necessarily organ specific, but they are very ancient in animal evolution. Another critical finding concerning our understanding of the role of development in the evolution of animal form is that these gene regulatory cassettes are directed to particular places of the embryo by tissue specific DNA regulatory elements (enhancers). These tissue specific enhancers activate the transcription of master genes, at the top of the gene regulatory cascade, to specific tissues and cause the activation of the entire gene regulatory network [5, 134]. This genetic and epigenetic architecture shape these regulatory cassettes, allowing them to display semiautonomous regulatory dynamics (modularity) and providing enormous flexibility to then be activated in different times and locations in the embryo as a whole package.

Therefore, in many cases, developmental modularity allows morphological modularity and it is an important component for linking patterns and processes of morphological evolution [5, 134]. The co-option of these preassembled gene regulatory networks can then facilitate the (1) independent evolution of similar tissue types, as in the case of animal cartilage, (2) new morphological traits, such as cephalopod arms and tentacles, and (3) the reactivation of whole developmental programs previously shut down in ancestral species, as is the case of limb loss reversal in the lizards genus

Bachia.

103

LIST OF REFERENCES

1. Minelli, A., Perspectives in animal phylogeny and evolution. Oxford biology. 2009, Oxford ; New York: Oxford University Press. xiii, 345 p.

2. Raff, R.A., The shape of life : genes, development, and the evolution of animal form. 1996, Chicago: University of Chicago Press. xxiii, 520 p.

3. Shubin, N., C. Tabin, and S. Carroll, Fossils, genes and the evolution of animal limbs. Nature, 1997. 388(6643): p. 639-48.

4. Shubin, N., C. Tabin, and S. Carroll, Deep homology and the origins of evolutionary novelty. Nature, 2009. 457(7231): p. 818-23.

5. Wagner, G.P., Homology, genes, and evolutionary innovation. 2014, Princeton ; Oxford: Princeton University Press. xiii, 478 pages.

6. Wake, D.B., M.H. Wake, and C.D. Specht, Homoplasy: from detecting pattern to determining process and mechanism of evolution. Science, 2011. 331(6020): p. 1032-5.

7. Gerhart, J. and M. Kirschner, Cells, embryos, and evolution : toward a cellular and developmental understanding of phenotypic variation and evolutionary adaptability. 1997, Malden, Mass.: Blackwell Science. xiii, 642 p.

8. Minelli, A., The development of animal form : ontogeny, morphology, and evolution. 2003, Cambridge, UK ; New York: Cambridge University Press. xviii, 323 p.

9. Wagner, G.P., The developmental genetics of homology. Nat Rev Genet, 2007. 8(6): p. 473-9.

10. Lemaire, J., Table de developpement embryonnaire de Sepia officinalis. L. (Mollusque Cephalopode). Bull Soc Zool France, 1970(95): p. 95:773–782. .

11. Grimaldi, A., et al., A hedgehog homolog is involved in muscle formation and organization of Sepia officinalis (mollusca) mantle. Dev Dyn, 2008. 237(3): p. 659-71.

12. Sekiguchi, K., Y. Yamamichi, and J.D. Costlow, Horseshoe crab developmental studies I. Normal embryonic development of Limulus polyphemus compared with Tachypleus tridentatus. Prog Clin Biol Res, 1982. 81: p. 53-73.

13. Blackburn, D.C., et al., Isolation and expression of Pax6 and atonal homologues in the American horseshoe crab, Limulus polyphemus. Dev Dyn, 2008. 237(8): p. 2209-19.

104

14. Hall, B.K., Chapter 37 - AERs in Limbed and Limbless Tetrapods, in Bones and Cartilage, B.K. Hall, Editor. 2005, Academic Press: San Diego. p. 469-477.

15. Sugahara, K., et al., Novel Sulfated Oligosaccharides Containing 3-O-Sulfated Glucuronic Acid from King Crab Cartilage Chondroitin Sulfate K: Unexpected degradation by chondroitinase ABC. Journal of Biological Chemistry, 1996. 271(43): p. 26745-26754.

16. Kinoshita, A., et al., Novel Tetrasaccharides Isolated from Squid Cartilage Chondroitin Sulfate E Contain Unusual Sulfated Disaccharide Units GlcA(3-O- sulfate)β1–3GalNAc(6-O-sulfate) or GlcA(3-O-sulfate)β1–3GalNAc(4,6-O- disulfate). Journal of Biological Chemistry, 1997. 272(32): p. 19656-19665.

17. Abzhanov, A., Darwin's finches: analysis of beak morphological changes during evolution. Cold Spring Harb Protoc, 2009. 2009(3): p. pdb emo119.

18. Quintana, L. and J. Sharpe, Preparation of mouse embryos for optical projection tomography imaging. Cold Spring Harb Protoc, 2011. 2011(6): p. 664-9.

19. Quintana, L. and J. Sharpe, Optical projection tomography of vertebrate embryo development. Cold Spring Harb Protoc, 2011. 2011(6): p. 586-94.

20. Edgar, R.C., MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res, 2004. 32(5): p. 1792-7.

21. Huelsenbeck, J.P. and F. Ronquist, MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics, 2001. 17(8): p. 754-5.

22. Whelan, S. and N. Goldman, A General Empirical Model of Protein Evolution Derived from Multiple Protein Families Using a Maximum-Likelihood Approach. Molecular Biology and Evolution, 2001. 18(5): p. 691-699.

23. Zhang, G., M.M. Miyamoto, and M.J. Cohn, Lamprey type II collagen and Sox9 reveal an ancient origin of the vertebrate collagenous skeleton. Proc Natl Acad Sci U S A, 2006. 103(9): p. 3180-5.

24. Lefebvre, V., et al., SOX9 is a potent activator of the chondrocyte-specific enhancer of the pro alpha1(II) collagen gene. Mol Cell Biol, 1997. 17(4): p. 2336- 46.

25. Arendt, D., The evolution of cell types in animals: emerging principles from molecular studies. Nat Rev Genet, 2008. 9(11): p. 868-82.

26. Nielsen, C., Animal Evolution: Interrelationships of the Living Phyla. 2012: OUP Oxford.

27. Moroz, L.L., et al., The ctenophore genome and the evolutionary origins of neural systems. Nature, 2014. 510(7503): p. 109-14.

105

28. Ryan, J.F., et al., The genome of the ctenophore Mnemiopsis leidyi and its implications for cell type evolution. Science, 2013. 342(6164): p. 1242592.

29. Cole, A.G. and B.K. Hall, Cartilage is a metazoan tissue; integrating data from nonvertebrate sources. Acta Zoologica, 2004. 85(2): p. 69-80.

30. Person, P. and D.E. Philpott, The nature and significance of invertebrate cartilages. Biological Reviews, 1969. 44(1): p. 1-15.

31. Schaffer, J., Die t t gewebe, in Handbuch der Mikroskopischen Anatomie des Menschen, .v. M llendorf, Editor. 1930, Springer: Berlin. p. pp 1–390.

32. Zhang, G., B.F. Eames, and M.J. Cohn, Chapter 2. Evolution of vertebrate cartilage development. Curr Top Dev Biol, 2009. 86: p. 15-42.

33. Cole, A.G. and B.K. Hall, The nature and significance of invertebrate cartilages revisited: distribution and histology of cartilage and cartilage-like tissues within the Metazoa. Zoology (Jena), 2004. 107(4): p. 261-73.

34. Cole, A.G., A review of diversity in the evolution and development of cartilage: the search for the origin of the chondrocyte. Eur Cell Mater, 2011. 21: p. 122-9.

35. Rychel, A.L., et al., Evolution and development of the chordates: collagen and pharyngeal cartilage. Mol Biol Evol, 2006. 23(3): p. 541-9.

36. Gans, C. and R.G. Northcutt, Neural crest and the origin of vertebrates: a new head. Science, 1983. 220(4594): p. 268-73.

37. Hall, B.K. and J.A. Gillis, Incremental evolution of the neural crest, neural crest cells and neural crest-derived skeletal tissues. J Anat, 2013. 222(1): p. 19-31.

38. Zhang, G. and M.J. Cohn, Hagfish and lancelet fibrillar collagens reveal that type II collagen-based cartilage evolved in stem vertebrates. Proc Natl Acad Sci U S A, 2006. 103(45): p. 16829-33.

39. Meulemans, D. and M. Bronner-Fraser, Insights from amphioxus into the evolution of vertebrate cartilage. PLoS One, 2007. 2(8): p. e787.

40. Wada, H., et al., Molecular evolution of fibrillar collagen in chordates, with implications for the evolution of vertebrate skeletons and chordate phylogeny. Evol Dev, 2006. 8(4): p. 370-7.

41. Hall, B.K., Chapter 3 - Cartilage, in Bones and Cartilage, B.K. Hall, Editor. 2005, Academic Press: San Diego. p. 33-47.

42. Akiyama, H., et al., The transcription factor Sox9 has essential roles in successive steps of the chondrocyte differentiation pathway and is required for expression of Sox5 and Sox6. Genes Dev, 2002. 16(21): p. 2813-28.

106

43. Smits, P., et al., The transcription factors L-Sox5 and Sox6 are essential for cartilage formation. Dev Cell, 2001. 1(2): p. 277-90.

44. Pourquié, O., The Skeletal System. 2009: Cold Spring Harbor Laboratory Press.

45. Kamachi, Y. and H. Kondoh, Sox proteins: regulators of cell fate specification and differentiation. Development, 2013. 140(20): p. 4129-44.

46. Sarkar, A. and K. Hochedlinger, The sox family of transcription factors: versatile regulators of stem and progenitor cell fate. Cell Stem Cell, 2013. 12(1): p. 15-30.

47. Kronenberg, H.M., Developmental regulation of the growth plate. Nature, 2003. 423(6937): p. 332-6.

48. Akiyama, H., et al., Misexpression of Sox9 in mouse limb bud mesenchyme induces polydactyly and rescues hypodactyly mice. Matrix Biol, 2007. 26(4): p. 224-33.

49. Cowden, R., A histochemical study of chondroid tissue in Limulus and Octopus. Histochemie, 1967. 9(2): p. 149-163.

50. Boot-Handford, R.P. and D.S. Tuckwell, Fibrillar collagen: the key to vertebrate evolution? A tale of molecular incest. Bioessays, 2003. 25(2): p. 142-51.

51. Huxley-Jones, J., D.L. Robertson, and R.P. Boot-Handford, On the origins of the extracellular matrix in vertebrates. Matrix Biol, 2007. 26(1): p. 2-11.

52. Exposito, J.Y., et al., Demosponge and sea anemone fibrillar collagen diversity reveals the early emergence of A/C clades and the maintenance of the modular structure of type V/XI collagens from sponge to human. J Biol Chem, 2008. 283(42): p. 28226-35.

53. Roughley, P.J., The structure and function of cartilage proteoglycans. Eur Cell Mater, 2006. 12: p. 92-101.

54. Matsumoto, K., et al., Conditional inactivation of Has2 reveals a crucial role for hyaluronan in skeletal growth, patterning, chondrocyte maturation and joint formation in the developing limb. Development, 2009. 136(16): p. 2825-35.

55. Volpi, N. and F. Maccari, Purification and characterization of hyaluronic acid from the mollusc bivalve Mytilus galloprovincialis. Biochimie, 2003. 85(6): p. 619-25.

56. DeAngelis, P.L., Evolution of glycosaminoglycans and their glycosyltransferases: Implications for the extracellular matrices of animals and the capsules of pathogenic bacteria. Anat Rec, 2002. 268(3): p. 317-26.

107

57. Zeng, L., et al., Shh establishes an Nkx3.2/Sox9 autoregulatory loop that is maintained by BMP signals to induce somitic chondrogenesis. Genes Dev, 2002. 16(15): p. 1990-2005.

58. Park, J., et al., Regulation of Sox9 by Sonic Hedgehog (Shh) is essential for patterning and formation of tracheal cartilage. Dev Dyn, 2010. 239(2): p. 514-26.

59. Abzhanov, A. and C.J. Tabin, Shh and Fgf8 act synergistically to drive cartilage outgrowth during cranial development. Dev Biol, 2004. 273(1): p. 134-48.

60. Lefebvre, V., P. Li, and B. de Crombrugghe, A new long form of Sox5 (L-Sox5), Sox6 and Sox9 are coexpressed in chondrogenesis and cooperatively activate the type II collagen gene. EMBO J, 1998. 17(19): p. 5718-33.

61. Hill, T.P., et al., Canonical Wnt/beta-catenin signaling prevents osteoblasts from differentiating into chondrocytes. Dev Cell, 2005. 8(5): p. 727-38.

62. Day, T.F., et al., Wnt/beta-catenin signaling in mesenchymal progenitors controls osteoblast and chondrocyte differentiation during vertebrate skeletogenesis. Dev Cell, 2005. 8(5): p. 739-50.

63. Akiyama, H., et al., Interactions between Sox9 and beta-catenin control chondrocyte differentiation. Genes Dev, 2004. 18(9): p. 1072-87.

64. Lefebvre, V. and P. Bhattaram, Vertebrate skeletogenesis. Curr Top Dev Biol, 2010. 90: p. 291-317.

65. St-Jacques, B., M. Hammerschmidt, and A.P. McMahon, Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation. Genes Dev, 1999. 13(16): p. 2072-86.

66. Dray, N., et al., Hedgehog signaling regulates segment formation in the annelid Platynereis. Science, 2010. 329(5989): p. 339-42.

67. Broun, M., et al., Formation of the head organizer in hydra involves the canonical Wnt pathway. Development, 2005. 132(12): p. 2907-16.

68. Tosches, M.A. and D. Arendt, The bilaterian forebrain: an evolutionary chimaera. Curr Opin Neurobiol, 2013. 23(6): p. 1080-9.

69. Arendt, D., et al., The evolution of nervous system centralization. Philos Trans R Soc Lond B Biol Sci, 2008. 363(1496): p. 1523-8.

70. Blackburn, D.G., Chorioallantoic placentation in squamate reptiles: Structure, function, development, and evolution. Journal of Experimental Zoology, 1993. 266(5): p. 414-430.

108

71. Nielsen, C., Animal evolution : interrelationships of the living phyla. 3rd ed. 2012, Oxford ; New York: Oxford University Press. x, 402 p.

72. Schmidt-Rhaesa, A., The evolution of organ systems. Oxford biology. 2007, Oxford ; New York: Oxford University Press. x, 385 p.

73. Kocot, K.M., et al., Phylogenomics reveals deep molluscan relationships. Nature, 2011. 477(7365): p. 452-456.

74. Smith, S.A., et al., Resolving the evolutionary relationships of molluscs with phylogenomic tools. Nature, 2011. 480(7377): p. 364-7.

75. Kroger, B., J. Vinther, and D. Fuchs, Cephalopod origin and evolution: A congruent picture emerging from fossils, development and molecules: Extant cephalopods are younger than previously realised and were under major selection to become agile, shell-less predators. Bioessays, 2011. 33(8): p. 602- 13.

76. Shigeno, S., et al., Evolution of the cephalopod head complex by assembly of multiple molluscan body parts: Evidence from Nautilus embryonic development. J Morphol, 2008. 269(1): p. 1-17.

77. Angelini, D.R. and T.C. Kaufman, Insect appendages and comparative ontogenetics. Dev Biol, 2005. 286(1): p. 57-77.

78. Abzhanov, A. and T.C. Kaufman, Homologs of Drosophila appendage genes in the patterning of arthropod limbs. Dev Biol, 2000. 227(2): p. 673-89.

79. Pueyo, J.I. and J.P. Couso, Parallels between the proximal-distal development of vertebrate and arthropod appendages: homology without an ancestor? Curr Opin Genet Dev, 2005. 15(4): p. 439-46.

80. Panganiban, G., et al., The origin and evolution of animal appendages. Proc Natl Acad Sci U S A, 1997. 94(10): p. 5162-6.

81. Winchell, C.J., J.E. Valencia, and D.K. Jacobs, Expression of Distal-less, dachshund, and optomotor blind in Neanthes arenaceodentata (Annelida, Nereididae) does not support homology of appendage-forming mechanisms across the Bilateria. Dev Genes Evol, 2010. 220(9-10): p. 275-95.

82. Ingham, P.W., Y. Nakano, and C. Seger, Mechanisms and functions of Hedgehog signalling across the metazoa. Nat Rev Genet, 2011. 12(6): p. 393- 406.

83. Kojima, T., et al., Induction of a mirror-image duplication of anterior wing structures by localized hedgehog expression in the anterior compartment of Drosophila melanogaster wing imaginal discs. Gene, 1994. 148(2): p. 211-7.

109

84. Riddle, R.D., et al., Sonic hedgehog mediates the polarizing activity of the ZPA. Cell, 1993. 75(7): p. 1401-16.

85. Scherz, P.J., et al., The limb bud Shh-Fgf feedback loop is terminated by expansion of former ZPA cells. Science, 2004. 305(5682): p. 396-9.

86. Zhu, J., et al., Uncoupling Sonic hedgehog control of pattern and expansion of the developing limb bud. Dev Cell, 2008. 14(4): p. 624-32.

87. Janssen, R. and G.E. Budd, Deciphering the onychophoran 'segmentation gene cascade': Gene expression reveals limited involvement of pair rule gene orthologs in segmentation, but a highly conserved segment polarity gene network. Dev Biol, 2013. 382(1): p. 224-34.

88. Zeller, R., J. Lopez-Rios, and A. Zuniga, Vertebrate limb bud development: moving towards integrative analysis of organogenesis. Nat Rev Genet, 2009. 10(12): p. 845-58.

89. Nusse, R. and R. van Amerongen, Wnt Signaling: A Subject Collection from Cold Spring Harbor Perspectives in Biology. 2013: Cold Spring Harbor Laboratory Press.

90. Hogvall, M., et al., Analysis of the Wnt gene repertoire in an onychophoran provides new insights into the evolution of segmentation. Evodevo, 2014. 5(1): p. 14.

91. Murat, S., C. Hopfen, and A.P. McGregor, The function and evolution of Wnt genes in arthropods. Arthropod Struct Dev, 2010. 39(6): p. 446-52.

92. Capellini, T.D., et al., Pbx1/Pbx2 requirement for distal limb patterning is mediated by the hierarchical control of Hox gene spatial distribution and Shh expression. Development, 2006. 133(11): p. 2263-73.

93. Ahn, K., et al., BMPR-IA signaling is required for the formation of the apical ectodermal ridge and dorsal-ventral patterning of the limb. Development, 2001. 128(22): p. 4449-61.

94. Soshnikova, N., et al., Genetic interaction between Wnt/beta-catenin and BMP receptor signaling during formation of the AER and the dorsal-ventral axis in the limb. Genes Dev, 2003. 17(16): p. 1963-8.

95. Brook, W.J. and S.M. Cohen, Antagonistic interactions between wingless and decapentaplegic responsible for dorsal-ventral pattern in the Drosophila Leg. Science, 1996. 273(5280): p. 1373-7.

96. Philipp, I., et al., Wnt/beta-catenin and noncanonical Wnt signaling interact in tissue evagination in the simple eumetazoan Hydra. Proc Natl Acad Sci U S A, 2009. 106(11): p. 4290-5.

110

97. Cohn, M.J., Development of the external genitalia: conserved and divergent mechanisms of appendage patterning. Dev Dyn, 2011. 240(5): p. 1108-15.

98. Lemons, D., et al., Co-option of an anteroposterior head axis patterning system for proximodistal patterning of appendages in early bilaterian evolution. Dev Biol, 2010. 344(1): p. 358-62.

99. Kardong, K.V., Vertebrates: comparative anatomy, function, evolution. 2006: McGraw-Hill Higher Education.

100. Hall, B.K., Fins into Limbs: Evolution, Development, and Transformation. 2008: University of Chicago Press.

101. Zug, G.R., L.J. Vitt, and J.P. Caldwell, Herpetology : an introductory biology of amphibians and reptiles. 2nd ed. 2001, San Diego, Calif.: Academic Press. xiv, 630 p.

102. Pough, F.H., Herpetology. 2nd ed. 2001, Upper Saddle River, N.J.London: Prentice Hall ;Prentice-Hall International. xi, 612p, 4 p of plates. 103. Greer, A.E., Limb Reduction in Squamates: Identification of the Lineages and Discussion of the Trends. Journal of Herpetology, 1991. 25(2): p. 166-173.

104. Dixon, J.R., A systematic review of the Teiid lizards, genus Bachia, with remarks on Heterodactylus and Anotosaura. University of Kansas Museum of Natural History Miscellaneous publication no 57. 1973, Lawrence,: University of Kansas. 47 p.

105. Presch, W., Evolutionary History of the South American Microteiid Lizards (Teiidae: Gymnophthalminae). Copeia, 1980. 1980(1): p. 36-56.

106. Kizirian, D.A. and R.W. McDiarmid, A New Species of Bachia (Squamata: Gymnophthalmidae) with Plesiomorphic Limb Morphology. Herpetologica, 1998. 54(2): p. 245-253.

107. Brandley, M.C., J.P. Huelsenbeck, and J.J. Wiens, Rates and patterns in the evolution of snake-like body form in squamate reptiles: evidence for repeated re- evolution of lost digits and long-term persistence of intermediate body forms. Evolution, 2008. 62(8): p. 2042-64.

108. Kohlsdorf, T., et al., Data and Data Interpretation in the Study of Limb Evolution: A Reply to Galis Et Al. On the Reevolution of Digits in the Lizard Genus Bachia. Evolution, 2010. 64(8): p. 2477-2485.

109. Kohlsdorf, T. and G.P. Wagner, Evidence for the reversibility of digit loss: a phylogenetic study of limb evolution in Bachia (Gymnophthalmidae: Squamata). Evolution, 2006. 60(9): p. 1896-912.

111

110. Galis, F., J.W. Arntzen, and R. Lande, Dollo's law and the irreversibility of digit loss in Bachia. Evolution, 2010. 64(8): p. 2466-76; discussion 2477-85.

111. Kearney, M. and B.L. Stuart, Repeated evolution of limblessness and digging heads in worm lizards revealed by DNA from old bones. Proc Biol Sci, 2004. 271(1549): p. 1677-83.

112. Bonett, R.M., et al., Evolution of paedomorphosis in plethodontid salamanders: ecological correlates and re-evolution of metamorphosis. Evolution, 2014. 68(2): p. 466-82.

113. Wiens, J.J., Re-evolution of lost mandibular teeth in frogs after more than 200 million years, and re-evaluating Dollo's law. Evolution, 2011. 65(5): p. 1283-96.

114. Wiens, J.J., et al., Loss and re-evolution of complex life cycles in marsupial frogs: does ancestral trait reconstruction mislead? Evolution, 2007. 61(8): p. 1886-99.

115. Collin, R. and M.P. Miglietta, Reversing opinions on Dollo's Law. Trends Ecol Evol, 2008. 23(11): p. 602-9.

116. Niswander, L., et al., FGF-4 replaces the apical ectodermal ridge and directs outgrowth and patterning of the limb. Cell, 1993. 75(3): p. 579-87.

117. Saunders, J.W., Jr., The proximo-distal sequence of origin of the parts of the chick wing and the role of the ectoderm. J Exp Zool, 1948. 108(3): p. 363-403.

118. Tickle, C., The contribution of chicken embryology to the understanding of vertebrate limb development. Mechanisms of Development, 2004. 121(9): p. 1019-1029.

119. Zakany, J., M. Kmita, and D. Duboule, A dual role for Hox genes in limb anterior- posterior asymmetry. Science, 2004. 304(5677): p. 1669-72.

120. Fromental-Ramain, C., et al., Hoxa-13 and Hoxd-13 play a crucial role in the patterning of the limb autopod. Development, 1996. 122(10): p. 2997-3011.

121. Spitz, F., et al., Large scale transgenic and cluster deletion analysis of the HoxD complex separate an ancestral regulatory module from evolutionary innovations. Genes & Development, 2001. 15(17): p. 2209-2214.

122. Montavon, T., et al., A regulatory archipelago controls Hox genes transcription in digits. Cell, 2011. 147(5): p. 1132-45.

123. Shubin, N. and P. Alberch, A Morphogenetic Approach to the Origin and Basic Organization of the Tetrapod Limb, in Evolutionary Biology, M. Hecht, B. Wallace, and G. Prance, Editors. 1986, Springer US. p. 319-387.

112

124. Larsson, H.C. and G.P. Wagner, Pentadactyl ground state of the avian wing. J Exp Zool, 2002. 294(2): p. 146-51.

125. Whiting, M.F., S. Bradler, and T. Maxwell, Loss and recovery of wings in stick insects. Nature, 2003. 421(6920): p. 264-7.

126. Collin, R., et al., Molecular phylogenetic and embryological evidence that feeding larvae have been reacquired in a marine gastropod. Biol Bull, 2007. 212(2): p. 83-92.

127. Moczek, A.P., T.E. Cruickshank, and A. Shelby, When ontogeny reveals what phylogeny hides: gain and loss of horns during development and evolution of horned beetles. Evolution, 2006. 60(11): p. 2329-41.

128. Abzhanov, A., et al., The calmodulin pathway and evolution of elongated beak morphology in Darwin's finches. Nature, 2006. 442(7102): p. 563-7.

129. Abzhanov, A., et al., Bmp4 and morphological variation of beaks in Darwin's finches. Science, 2004. 305(5689): p. 1462-5.

130. Chan, Y.F., et al., Adaptive evolution of pelvic reduction in sticklebacks by recurrent deletion of a Pitx1 enhancer. Science, 2010. 327(5963): p. 302-5.

131. Prud'homme, B., et al., Repeated morphological evolution through cis-regulatory changes in a pleiotropic gene. Nature, 2006. 440(7087): p. 1050-3.

132. Khila, A., E. Abouheif, and L. Rowe, Evolution of a novel appendage ground plan in water striders is driven by changes in the Hox gene Ultrabithorax. PLoS Genet, 2009. 5(7): p. e1000583.

133. Hirasawa, T., et al., The evolutionary origin of the turtle shell and its dependence on the axial arrest of the embryonic rib cage. J Exp Zool B Mol Dev Evol, 2014.

134. Stern, D.L., Evolution, Development, & the Predictable Genome. 2011: Roberts and Company Publishers.

113

BIOGRAPHICAL SKETCH

Oscar Tarazona was born in Colombia, South America. Oscar completed his bachelor of science with a major in biology and a minor in vertebrate reproductive biology were he graduated with Cum Laude honors at Universidad Industrial de

Santander in Colombia. Oscar studied skeletal biology and skull development in lizards for his undergraduate dissertation where he started to show interest in genetics and cell and developmental biology of cartilage and bone. His first peer-reviewed research papers were published while working as a research assistant and all were dealing with skull and limb development in lizards and salamanders. In 2008 Oscar pursued his

Ph.D. degree in Zoology in the College of Liberal Arts and Science at the University of

Florida working with Dr. Martin J. Cohn studying the molecular basis of invertebrate cartilage development. He was awarded a Grinter Fellowship by University of Florida from 2008 to 2010 and an International Student Research fellowship by Howard Hughes

Medical Institute from 2011 to 2013. He graduated in August 2015 with a Ph.D. in zoology.

114