MOLECULAR MECHANISMS OF ZEBRAFISH MOTONEURON DEVELOPMENT by LAURA ANN HALE A DISSERTATION Presented to the Department of Biology and the Graduate School of the University of Oregon in partial fulfillment of the requirements for the degree of Doctor of Philosophy December 2009 11 University of Oregon Graduate School Confirmation of Approval and Acceptance of Dissertation prepared by: Laura Hale Title: "Molecular Mechanisms of Zebrafish Motoneuron Development" This dissertation has been accepted and approved in partial fulfillment of the requirements for the Doctor of Philosophy degree in the Department of Biology by: Monte Westerfield, Chairperson, Biology Judith Eisen, Advisor, Biology Victoria Herman, Member, Biology John Postlethwait, Member, Biology Clifford Kentros, Outside Member, Psychology and Richard Linton, Vice President for Research and Graduate Studies/Dean of the Graduate School for the University of Oregon. December 12, 2009 Original approval signatures are on file with the Graduate School and the University of Oregon Libraries. © 2009 Laura Ann Hale iii IV An Abstract of the Dissertation of Laura Ann Hale in the Department of Biology for the degree of to be taken Doctor ofPhilosophy December 2009 Title: MOLECULAR MECHANISMS OF ZEBRAFISH MOTONEURON DEVELOPMENT Approved: _ Dr. Judith Eisen This dissertation describes research to identifY genes involved in specification, patterning and development of zebrafish primary motoneurons. We first examined the spatiotemporal expression patterns of retinoic acid and retinoid X receptor mRNAs to determine whether particular ones might be involved in motoneuron specification or patterning. Retinoic acid and retinoid X receptor mRNAs are expressed at the right time to pattern motoneurons, but the expression patterns did not suggest roles for particular receptors. In contrast, netrin mRNAs are expressed in specific motoneuron intermediate targets and knockdown experiments revealed an important role in development of YaP motoneurons. Two identified motoneurons, CaP and YaP, initially form an equivalence pair. CaPs extend long axons that innervate ventral muscle. YaPs extend short axons that stop at muscle fibers called muscle pioneers; YaPs later typically die. Previous work showed that during extension, CaP axons pause at several intermediate targets, including muscle pioneers, and that both CaP and muscle pioneers are required for YaP formation. We found that mRNAs for different Netrins are expressed in intermediate targets before CaP axon vcontact: netrin fa in muscle pioneers, netrin fb in hypochord, and netrin 2 in ventral somite. We show that Netrins are unnecessary to guide CaP axons but are necessary to prevent VaP axons from extending into ventral muscle. Netrin la is necessary to stop YaP axons at muscle pioneers, Netrin la and Netrin 2 together are necessary to stop YaP axons near the hypochord, and Netrin Ib appears dispensable for CaP and YaP development. We also identitY Deleted in colorectal carcinoma as a Netrin receptor that mediates the ability of Netrin la to cause YaP axons to stop at muscle pioneers. Our results suggest Netrins refme axon morphology to ensure fmal cell-appropriate axon arborization. To learn whether Netrin proteins diffuse away from their sources of synthesis to function at a distance, we are developing Netrin antibodies. If successful, the antibodies will provide the research community at large with a new tool for understanding in vivo Netrin function. This dissertation includes both my previously published and unpublished coauthored material. CURRICULUM VITAE NAME OF AUTHOR: Laura Ann Hale PLACE OF BIRTH: Misawa Air Force Base, JAPAN DATE OF BIRTH: June 16,1978 GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon, Eugene University of California, Berkeley DEGREES AWARDED: Doctor of Philosophy, Biology, 2009, University ofOregon, Eugene Bachelor ofArts, Molecular and Cell Biology and Psychology, 2001, University of California, Berkeley AREAS OF SPECIAL INTEREST: Neuroscience Developmental Biology PROFESSIONAL EXPERIENCE: Graduate Teaching Fellow, UO Biology Dept., 2004 Research Assistant, UCB Psychology Dept., Dr. R.T. Knight, 2001-03 Research Assistant, UCB Mol. & Cell Biology Dept., Dr. F.H. Wilt, 2000-01 Undergraduate Student Instructor, UCB Integrative Biology Dept., 1998 Tutor, UCB College Writing Dept., 1997 GRANTS, AWARDS AND HONORS: Northwest Regional Developmental Biology Conference, 2nd place, Student Oral Presentation, 2009 UO Clarence and Lucille Dunbar Scholarship Recipient, 2007-08 Caswell Grave Scholarship, Embryology, Marine Biological Laboratory, Wood Hole, MA, 2007 John & Madeleine Trinkaus Endowed Scholarship, Embryology, Marine Biological Laboratory, Wood Hole, MA, 2007 American Heart Association Predoctoral Fellow, 2006-08 NIH Genetics Training Grant Appointee, 2004-05, 08-09 UCB Biology Fellows Program Research Grant Recipient, 2001 VI PUBLICAnONS: Solbakk, AX., Alpert, G.F., Furst, A.J., Hale, L.A., Oga, T., Chetty, S., Pickard, N., Knight R.T. (2008) Altered prefrontal function with aging: insights into age-associated performance decline. Brain Res. 1232:30-47. Hutchinson, S.A., Cheesman, S.E., Hale, L.A., Boone, J.Q., Eisen, J.S. (2007) Nkx6 proteins specify one zebrafish primary motoneuron subtype by regulating late islet1 expression. Development 134(9):1671-7. Padilla, M.L., Wood, R.A., Hale, L.A., Knight, R.T. (2006) Lapses in a prefronta1- extrastriate preparatory attention network predict mistakes. J. Cogn. Neurosci. 18(9):1477-87. Hale, L.A.*, Tallafuss, A.*, Yan, Y.L., Dudley, L., Eisen, J.S., Postlethwait, J.H. (2006) Characterization of the retinoic acid receptor genes raraa, rarab and rarg during zebrafish development. Gene Expr. Patterns 6(5):546-55. Tallafuss, A.*, Hale, L.A.*, Yan, Y.L., Dudley, L., Eisen, J.S., Postlethwait, J.H. (2006) Characterization of retinoid-X receptor genes, rxra, rxrba, rxrbb and rxrg, during zebrafish development Gene Expr. Patterns 6(5):556-6). Tanaka, M., Hale, L.A., Amores, A., Yan, Y.L., Cresko, W.A., Suzuki, T., Postlethwait, J.H. (2005) Developmental genetic basis for the evolution of pelvic fin loss in the pufferfish Takifugu rubripes. Dev BioI. 281(2):227-39. Yamaguchi, S., Hale, L.A., D'Esposito, M., Knight, R.T. (2004) Rapid prefronta1- hippocampal habituation to novel events. J. Neurosci. 24: 5356-5363. *Authors contributed equally to this work vii viii ACKNOWLEDGEMENTS The collaborative and nurturing environment of the University of Oregon's Institute of Neuroscience enabled me to successfully complete my dissertation research. I am fortunate to be able to consider my colleagues friends. I thank Dr. Judith Eisen for sharing her gift for story-telling; I hope to convey the complexities of developmental neurobiology as compellingly and simply. I am also grateful for her generous open office door policy. Dr. John Postlethwait patiently supported and guided me while I chose a dissertation lab and throughout my graduate career. I appreciate Dr. Monte Westerfield's insightful questions that helped me to delve deeper into my research. Similarly, Dr. Tory Herman's questions provided a more critical perspective of my work. I thank Dr. Cliff Kentros for providing an alternate viewpoint of my work. I also wish to thank Dr. Yasuko Bonjo for her patient training and prodding me to think more critically. I am grateful to the Institute of Neuroscience staff, especially Ellen McCumsey, Peg Morrow, Mike McBorse, Don Pate and Mikel Rhodes, and the biology staff, especially Donna Overall, for their excellent administrative and technical support. I thank the University of Oregon Zebrafish Facility for fish husbandry. Dr. Chi-Bin Chien and his lab at the University of Utah generously provided reagents and advice for the Netrin project. This work was supported by AHA 0610097Z, NIH GM07413 and NIH NS23915. Finally, I am deeply indebted to Bill Gillis for his unwavering support. For Chen Hale and Johnathan Hale ix xTABLE OF CONTENTS Chapter Page I. INTRODUCTION 1 II. CHARACTERIZATION OF THE RETINOIC ACID RECEPTOR GENES RARAA, RARAB AND RARG DURING ZEBRAFISH DEVELOPMENT 4 1. Results and Discussion............................................................................ 4 1.1. Orthologies of Zebrafish rar Genes........................................................ 5 1.2. Expression of rar Genes During Zebrafish Embryonic Development.. ..... 9 1.2.1. Maternal Expression of rar Genes....................................................... 9 1.2.2. raraa Expression 12 1.2.3. rarab Expression................................................................................ 16 1.2.4. rarg Expression 18 1.3. Conclusions 20 2. Experimental Procedures......................................................................... 20 2.1. Cloning ofrarGenes....................................................................... ...... 20 2.2. Whole Mount In Situ Hybridization...................................................... 21 2.3. Sources ofAdditional Gene Expression Data......................................... 21 III. CHARACTERIZATION OF RETINOID-X RECEPTOR GENES RXRA, RXRBA, RXRBB AND RXRG DURING ZEBRAFISH DEVELOPMENT 23 1. Results and Discussion............................................................................ 23 1.1. Orthologies of Zebrafish rxr Genes........................................................ 24 Xl Chapter Page 1.2. Expression of rxr Genes During Zebrafish Embryonic Development 28 1.2.1. Maternal Expression ofrxr Genes....................................................... 31 1.2.2. rxra Expression.................................................................................. 32 1.2.3. rxrba and rxrbb Expression.. 32 1.2.4. rxrg Expression 36 1.2.5. Comparison of Zebrafish rxr Genes 37 1.2.6. Comparison to Rxr Gene Expression in Mouse and Chick................... 39 1.3. Conclusions 40 2. Experimental Procedures..... 41 2.1. Cloning of rxr Genes............................................................................. 41 2.2. Whole Mount RNA In Situ Hybridization................ 41 2.3. Sources ofAdditional Gene Expression Data......................................... 42 IV. NETRIN SIGNALING IS REQUIRED FOR DEVELOPMENT OF AN IDENTIFIED ZEBRAFISH MOTONEURON...................................................... 44 Summary.................................................................................................... 44 Introduction 44 Materials and Methods................................................................................ 47 Results 51 Discussion.................................................................................................. 62 Conclusion 67 xii Chapter Page V. CONCLUSION................................................................................................. 69 REFERENCES...................................................................................................... 71 X111 LIST OF FIGURES Figure Page Chapterll 1. Phylogeny of Retinoic Acid Receptor Proteins..................................................... 7 2. Conserved Syntenies for Rar Genes...................................................................... 10 3. raraa, rarab and rarg Expression Visualized at 1.5-2, 4.7, 9 and 10.5-11 hpf 12 4. raraa, rarab and rarg Transcript Expression Visualized at 12, 24 and 48 hpf 14 5. Distinctive Expression of raraa, rarab and rarg 15 Chapter ill 1. Phylogenetic Analysis ofRetinoid-X Receptor Proteins 27 2. Genomic Analysis of Conserved Syntenies for Zebrafish rxra and rxrg Genes......... 29 3. Genomic Analysis ofConserved Syntenies for Zebrafish RXRB Co-Orthologs 31 4. Expression of rxra, rxrba, rxrbb and rxrg Visualized at 1.5-2, 4.7, 8-9 and 10.5-11 hpf.............................................................. 33 5. Expression of rxra, rxrbb and rxrg at 12, 24 and 48 hpf........................................... 34 6. Distinctive Expression of rxra, rxrbb and rxrg . 35 Chapter IV 1. Netrin mRNAs Are Expressed at Intermediate Targets for the CaP Axon.............. 52 2. Netrins Are Unnecessary for CaP Axons to Extend into Ventral Muscle 54 3. Ntn1a Is Necessary for VaP Axons to Stop at the Horizontal Myoseptum 57 xiv Figure Page 4. Netrins Act Together to Restrict Extension of a Second CaP Axon into Ventral Muscle................................................................................................................... 60 5. Dcc Is Necessary for VaP Axons to Stop at Muscle Pioneers................................. 61 xv LIST OF TABLES Table Page Chapterll 1. Survey of Recent and Previous Rar Gene Names.................................................. 6 Chapter ill 1. Survey of Recent and Previous Rxr Gene Names.................................................. 26 2. Primer Sequences Used to Obtain 1Xr Clones........................................................ 42 Chapter IV 1. MO Sequences Used to Knock Down Netrins and Dcc and RT Primer Sequences Used to Confirm MO Efficacy................................................................ 49 2. Netrin Signaling Is Necessary to Prevent a Second CaP Axon from Extending into Ventral Muscle. 64 CHAPTER I INTRODUCTION Analyzing neurodevelopment to understand the origins ofbehavior What are the origins ofbehavior? Approaches for answering that question range from analyzing macroscopic observations ofpeople behaving under various conditions to analyzing microscopic observations of individual neurons believed to contribute to the neural circuitry of readily identifiable behaviors. Collectively, each approach informs the other. For example, studying people as they behave while monitoring their brain activity with various neuroimaging techniques enables researchers to identify brain regions involved in generating behaviors. Researchers interested in determining neural circuitry underlying behavior can then study neurons in brain regions identified by macroscopic approaches to confIrm whether they contribute to behavior. Molecular approaches to understanding behavior also inform macroscopic approaches to behavior. Knockdown analysis of genes that lead to behavioral defIcits can help identify specific neurons and brain regions that generate behavior. In this dissertation, I describe work that contributes to determining genes involved in the development of a neural circuit responsible for motility. Patterning the vertebrate nervous system I studied genes involved in patterning zebraflsh spinal cord during development as a model for understanding how neuronal identity might be conferred in vertebrates with larger central nervous systems; the relatively small size ofthe zebrafish nervous system provides an excellent system for understanding neurodevelopment at the level of individual neurons. If 1 2neuronal subpopulations are not correctly specified then these components of a neural circuit wi11likely not function correctly to generate behavior. Patterning of the nervous system begins along the rostrocaudal axis and then extends along the dorsoventral axis. Within the developing vertebrate spinal cord, Sonic hedgehog protein released from the notochord serves as a ventralizing signal, diffusing from its source to create a gradient of expression. Cells exposed to different concentrations of Sonic hedgehog adopt different neuronal fates. For example, cells adjacent to the notochord adopt floorplate identity and later, more distant cells adopt motoneuron fates [reviewed by (Concordet and Ingham, 1995)]. Retinoic acid signaling also acts to pattern the nervous system along the rostrocaudal and dorsoventral axes (Maden, 2002). Retinoic acid signaling is also implicated in specification and differentiation of motoneurons (Maden, 2002; Appel and Eisen, 2003; Sockanathan et a1., 2003; Begemann et al., 2004; Linville, 2004; Goncalves et al., 2005). Two distinct populations of motoneurons, primary motoneurons and secondary motoneurons, arise from the motoneuron progenitor domain in zebrafish. Primary motoneurons are born earlier and are less numerous than secondary motoneurons. I studied the role of retinoic acid signaling in patterning motoneurons by analyzing expression of retinoic acid receptor and their potential binding partners, retinoid X receptors. I describe this work in Chapters II and III; both chapters include material coauthored with A. Ta1lafuss, Y.L. Yan, L. Dudley, J.H. Postlethwait, and J.S. Eisen. Connecting neurons to their targets I also studied how individual motoneurons connect to their muscle targets. Studying which genes are required for appropriate motoneuron and muscle connectivity in the 3peripheral nervous systems provides insight into how neurons within the central nervous system may be wired together during development. Building connections between neurons and an inappropriate target would likely generate an improperly wired circuit incapable of producing behavior. Attempts to answer how connections between neurons and their often distant targets are established during development have led to the now familiar axon guidance molecule, Netrin-l. The axonal pathway development of commissural neurons, an interneuron subpopulation whose cell bodies lie within the dorsal spinal cord and whose axons extend across the ventral midline, has been used as a model for studying axon guidance. In zebrafish, Netrins may function to guide a class ofventrally projecting motoneurons to their muscle targets. I discuss the results of this work in Chapter IV. Finally, I conclude in Chapter V with a brief discussion about the implications of my dissertation research. CHAPTERll CHARACTERIZATION OF THE RETINOIC ACID RECEPTOR GENES RARAA, RARAB AND RARG DURING ZEBRAFISH DEVELOPMENT The work described in this chapter was previously published in "Gene Expression Patterns," Vo1. 6. I share first authorship with A. Tallafuss. We were responsible for the majority of data collection, data analysis and writing. YL. Yan and L. Dudley cloned rarg. J.H. Postlethwait completed the syntenic analysis and contributed to writing. J.S. Eisen also contributed to writing. 1. Results and discussion Retinoic acid receptors (Rars) and retinoid-X receptors (Rxrs) form retinoic acid (RA) activated heterodimers that bind retinoic acid response elements (RAREs) and modulate transcription of target genes [reviewed by (Bastien and Rochette-Egly, 2004)]. In zebrafish as well as in tetrapod vertebrates, RA controls patterning of the central nervous system (CNS), paired appendages, and other organs (Gavalas and Krumlauf, 2000; Grandel et aI., 2002; Jiang et a1., 2002), as demonstrated by the characterization of embryos homozygous for mutations altering Raldh2, the enzyme required for RA synthesis (Begemann et a1., 2001; Niederreither et a1., 1999). Furthermore, mutations in RARs can lead to cancers such as acute promyelocytic leukemia or to heart malformations, and aberrant RA signaling may contribute to Parkinson's disease and schizophrenia (Krezel et aI., 1998; Soprano and Soprano, 2002; Lane and Bailey, 2005; Goodman, 2005; Rioux and Arnold, 2005). RA is important for somite formation (Appel and Eisen, 2003) and also plays 4 5a role in patterning both the anterioposterior and dorsoventral axes of the eNS (Maden, 2002) with primary sites of action in the hindbrain and anterior spinal cord (Maden, 2002). RA is also implicated in specification and differentiation of motoneurons (Maden, 2002; Appel and Eisen, 2003; Sockanathan et al., 2003; Begemann et al., 2004; Linville, 2004; Goncalves et al., 2005) and interneurons (Maden, 2002). Three Rar genes (Rara, Rarb, and Rarg) have been isolated from tetrapods (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene). In cultured neural progenitor cells, the order and combination of Rar activation determines the type of neural cells that are generated (Goncalves et al., 2005), suggesting that different Rars may play distinct roles during development. Although three rar genes have been isolated in zebrafish (White et al., 1994; Joore et aI., 1994), there has yet to be a thorough study of their orthologies, genomic arrangements, and embryonic expression patterns. As a necessary precursor to understanding the roles ofRars in cell specification during zebrafish development, we have investigated the orthologous relationships and expression patterns of three zebrafish rar genes. 1.1. Orthologies of zebrafish rar genes To make meaningful comparisons of zebrafish gene expression patterns to those of mouse and other tetrapods, it is essential to evaluate orthologous genes. Phylogenetic analysis of zebrafish Rars (Fig. 1) showed that sequences L03398 and L03399 (formerly called rara2a and rara2b (Mangelsdorf et al., 1995) grouped as co-orthologs of tetrapod Rara proteins with high bootstrap support. Thus, these genes should be renamed raraa and rarab according to zebrafish nomenclature rules (http://zfm.org/zCinfo/nomen.html); refer also to Table 1. We found an ortholog of raraa in the pufferfish genome, but not an ortholog of 6rarab (Fig. 1). Although no rarb-like sequence has been described in zebrafish, or located in the Zv4 or Zv5 assembly of the zebrafish genome (http://www.ensemb1.org/Danio_rerio/index.html), we identified a sequence in the pufferfish genome that fell in the tree as expected for an ortholog of mammalian Rarb (Fig. 1). We do not know whether the zebrafish genome lacks a rarb gene or if we were unable to identify it. Mice lacking Rarb function grow slowly, but are otherwise apparently normal (Ghyselinck et aI, 1997). The tree further showed that zebrafish and pufferfish both have an ortholog of human RARG (Fig. 1). Phylogenetic analysis indicated that RAR genes are more closely related to thyroid hormone receptor genes (THRs), which are their near neighbors on human chromosomes, than to the genes encoding the RXRs with which they functionally interact. Rt dfT bl 1 S thIS paper 1Joore et a1., 1994 2Strausberg et a1., 2002 3ZFIN (2003; http://zfin.org/cgi-bin/webdriver?Mlval=aa-pubview2.apg&OID=ZDB- PUB-030508-1) 4HUGO (2005; http://www.gene.ucl.ac.uk/nomenclature/index.html) 5MGI (2005; http://www.informatics.jax.org/) a e urvey 0 recen an .preVIOUS ar gene names. Zebrafish Gene Accession Previous Zebrafish Human Mouse number Gene Ortholog ortholog raraa* L03398 zRARa/; rara2ti RARA4 Raras rarab* L03399 rarall; rara2bJ none identified* none identified R4RJ34 Rarabs rarg* 874156 zRARy/; rari,J RAR4 Rarg * 71000 1000 Rara_Gaga Rara Xela rarab Dare raraa Teni raraa Dare 991 987 996 RARB_Hosa 998 Rarb Mumu Rarb_Gaga rarb_Paol rarb Teni 998 RARA_HoSa 961 Rara Mumu 889 917 828 1000 996 889 494 983 THRB_Hosa 1000 Thrb_Mumu Thrb_Gaga RXRA Rosa Rxra Mumu Rxra_Gaga Rxra_Xela rxra Dare Thra_Gaga '--- thra_Dare Rarg_Mumu Rarg_Gaa Rarg Xela 270 - rarg Dare10~ rar;_Teni 1000 THRA_Hosa 745 Thra Mumu 875 0.1 1000 1000 Fig. 1. Phylogeny ofretinoic acid receptor proteins. Amino acid sequences were aligned by Clustal-X, sequences were trimmed to include unambiguously aligned regions, and phylogenie analysis was by the neighbor-joining method (Saitou, 1987). Numbers are bootstraps values out of 1000 runs. Alignments are available on request. Species abbreviations: Hosa, Homo sapiens (human); Mumu, Mus musculus (mouse); Gaga, Gallus gallus (chicken); Xela, Xenopus laevis (frog); Dara, Danio ren'o (zebrafish); Paol, Paralichthys olivaceus (flounder); Teni, Tetraodon nigroviridis (pufferfish); Caau, Carassius auratus (goldfish). Sequence accession numbers: RARA Hosa AAB19602; Rara Mumu AAH10216; Rara Gaga CAA55134; Rara Xela A56558; rarab Dare AAH49301; raraa Teni GSTENT00024106001; raraa Dare NP 571481; RARB Hosa AAH60794; Rarb Mumu NP 035373; Rarb Gaga CAA39997; rarb Paol BAB71757; rarb Teni CAG11671; RARG Hosa CAB60726; Rarg Mumu P20787; Rarg Gaga CAA52153; Rarg Xe1a P28699; rarg Dare Q91392; rar Teni GSTENT00021064001; THRA Hosa 135702; Thra Mumu 586089; Thra Gaga NP_990644; thra Dare 1314773; THRB Hosa P37243; Thrb Mumu NP_033406; Thrb Gaga CAA90566; RXRA Hosa NP_002948; Rxra Mumu AAB36778; Rxra Gaga XP_415426; Rxra Xela P51128; rxraa Dare ENSDARP00000003080; Rxra Caau AA022211. 8Analysis of conserved syntenies can provide evidence oforthologies independent of recent phylogenies and is necessary to determine whether the duplicated genes raraa and rarab are tandem duplicates or if they arose in the ancient teleost genome duplication (Amores et aI, 1998; Postlethwait et a1., 1998; Woods et a1., 2000; Taylor et a1., 2003). Zebrafish co-orthologs ofRARA reside in conserved chromosome segments on the duplicated linkage groups LG3 and LG12 (Fig. 2Al, 2A6), with raraalRARA and top2al TOP2A near neighbors in zebrafish and human genomes (Fig. 2A2-4), and rarab just three genes from an ortholog ofNBRI which resides about 2Mb from RARA (Fig. 2A3, 2AS). (Note that Zv4 has raraa incorrectly positioned on Chromosome 18). The analysis of orthologous relationships confirms the decision to rename the zebrafish sequences raraa and rarab based on phylogenetic analyses and shows that raraa and rarab arose in the teleost genome duplication event. Although we found in Zv4 the zebrafish orthologs of TOP2B and THRB, which flank RARE in human, these genes were dispersed in the zebrafish genome and are not located near any RAR~like sequence. In contrast, we found the ortholog ofRARE (GSTENT00033708001) in the genome of the pufferfish Tetraodon nigroviridis next to the pufferfish ortholog of TOP2B (Fig. 2B), confirming orthologies shown in the tree (Fig. 1). We suggest that chromosome rearrangements that dispersed the neighbors of rarb may have led to its loss from the zebrafish genome. The zebrafish sequence Q91392 mapped to LG23 (Fig. 2Cl) and it resides in the genome near several other loci whose human orthologs are near RARGon human (Homo sapiens, Hosa) chromosome Hosal2q13 (Fig. 2C3, 2C4). These comparative synteny data provide strong evidence that rarg is the ortholog of tetrapod RARG. The close linkage ofparalogs ofRAR, THR, TOP2, and HOX genes in the human genome is consistent with the origin of the three human RAR genes by chromosome (or genome) duplication at the base of vertebrate evolution, and the origin of rara paralogs in zebrafish by an additional genome duplication event in ray fm fish evolution. 1.2. Expression of rar genes during zebrafish embryonic development We assayed the temporal and spatial distribution of raraa, rarab and rarg by whole mount RNA in situ hybridization beginning in the cleavage period (0.75-2 hours post fertilization, hpf), extending through the blastula (2.25-4.66 hpf), gastrula (5.25-10 hpf), segmentation (10.33 -22 hpf), and pharyngula (24-42 hpf) stages and concluding in the hatching period (48-72 hpf). For the purpose of comparison, Fig. 3 and 4 contain overviews of rar expression and Fig. 5 details unique rar expression patterns in the head and tail. 1.2.1. Maternal expression of rar genes In situ hybridization experiments revealed rarab and rarg transcripts during cleavage (1.5 hpf; Fig. 3B) and blastula stages (4.7 hpf; 3C). Because early rarab expression was similar to rarg expression, we show only one example of expression for each gene at early developmental stages (Fig. 3B, 3C). Consistent with previous results (Joore et a1., 1994) and in comparison to embryos treated with a raraa sense probe, we did not observe raraa expression at 1.5 hpf (Fig. 3A) and 4.7 hpf (data not shown); expression of raraa was not observed until 8 hpf. In addition, we did not observe localized rar expression patterns (described below) until later gastrula stages. 9 10 Fig. 2. Conserved syntenies for Rar genes. A. Mapping and local conserved syntenies for zebrafish raraa and rarab show that these genes reside in duplicated chromosome segments. B. A Rarb-like sequence in the genome of the pufferfish Tetraodon nigroviridis resides immediately adjacent to orthologs of genes that are adjacent to RARB in human, but orthologous sequences were not found in the zebrafish genome. C. Mapping and local conserved syntenies for zebrafish rarg shows that it resides near genes that are near RARG in human. In A2, raraa is ENSDART00000026235 and top2a is ENSDART00000050143. In A5, rarab is ENSDART00000044677; g6pe is ENSDART000000148l2; myst2 is ENSDART00000049095; nbrl is ENSDART00000018647; and aoe2 is ENSDART00000027258. In B3, Rarb is GSTENT00033708001; Top2b is GSTENT0003370700l; Nglyl is GSTENT00033706001; and Ks is GSTENT0003370500I. In C2, rarg is ENSDARESTT00000015370; kiaal536 is ENSDART00000040596; slc39a5 is ENSDART00000029724; flj34236 is ENSDART00000042557; zge:55389 is ENSDARTOOOOO0l69l2; and hoxel3 is ENSDARTOOOOO0l7685. 11 6, LG3 Zl183 AI584238 8C047806 }. - AW203116 rarab BC056805 j_. AW826550 ,... dnml1 ~ AI658332 - AW281208123A I 0.0 13.3 I 256ill 385QJ---" ,cr \.......51,2637 809 90A 104.6 5. LG33. Hosa17 o.~--UJ 2. LG12 • ~ w p~ '""," ~ ~:~'= ~ I ;;li::J=l ~ 0 -~/~" I?~\~i;n~ ~-......... :::::M- 1 : ~ 17·'123 17124 _ 17'125 --._' A. 1 LG12 25.7 78.5 DOli Z7576 ~ 13.6 11 AI959758 ~ I I BC059465~-~ ~~~~2611I I AI626476 I i BC063953 H- Z11124 ill I i I· AW154240 ~ , Zl141 37.9 48.3 58.5 689 904 1028 I - ~ ~ 0)- 0) i I • w 0"1 b- ill en s: 3 Hosa12q13 l ::r ·0 ~ -> W 2. LG23 ~--- 1:::.,t4 12'Q15 _ Id".r:1 4. Hosa12 AI618627 11l21: Ic'ln t;.:,:>4.1 _ 12:12"'.2 Z3852 rar[J Z1348 Iwisl2 Z22095 81673606 CD777575 AI666933 Z13781 s: M495001 • AI601307 I H II1371 0.0 13.6 22.9 336 48.5 585 68.5 80.7 930 107.2 1·198 C. 1 LG23 B. 2.l-bsa3p24 3. TeniSO\F15041 , f~~I~ I~.~..~sa3 /~~I ...' -"S~rJ~ ro cr I~~r~1- ~ (~ cr I r~a I~"U~I o:r."§I r=~ '<- 0 0(J) ""'" 'i: II)U> S '0 cr ;0; cr 12 1. 5-2h A B III 01 ~~ ~ ~ III C1:l C1:l ~ ~ ~ I 9h II 10.5-11h 0 E ~ F C1:l III ~ C1:l ~ G H ,. ~ III ~ K L Fig. 3. raraa (A, D-F), rarab, (C, G-I), and rarg (B, J-L) expression visualized at 1.5-2 hpf (A,B), 4.7 hpf(C), 9 hpf(D,G,J) and 10.5-11 hpf(E,F,H,I,K,L). Dorsal view (A,B,F,I,L). Lateral view; anterior up and dorsal to the left (D,E,G,H,J,K). (E) Grey arrowheads mark narrow anterior bands. Open black arrowhead marks posterior broad band. (H) Grey arrowhead marks single anterior band. Black arrowhead marks faint broad band. Open black arrowhead marks prominent broad band. Abbreviations: h, presumptive head; 1m, lateral mesoderm; nt, notochord; tb, tailbud; v, vegetal pole; yp, yolk plug. Scale bar represents 118 ~m in 2A, 2B, 135 ~m in 2C, and 121 ~m in 2D-L. 1.2.2. raraa expression At 9 hpf, raraa transcripts were expressed throughout the dorsal epiblast (Fig. 3D), except in the dorsal midline (data not shown), and were expressed ventrally, in prospective head and tail regions (Fig. 3D). Expression of raraa was further restricted during the segmentation period. 13 At the beginning of segmentation (lO-llhpf), raraa was expressed in the presumptive hindbrain, posterior neural plate and tail region (Fig. 3E). Expression of raraa within the neural plate extended in a gradient along the anterioposterior axis (Fig. 3F); the strongest expression occurred anteriorly in the hindbrain and posteriorly in the tailbud. In the hindbrain (Fig. 3F), raraa was expressed in three stripes perpendicular to the anterioposterior axis: two faint, narrow, anterior bands (grey arrowheads in Fig. 3E) and a strong, broad, posterior band that faded posteriorly (open arrowhead in Fig. 3E). Intense raraa staining also appeared in the tail bud and in lateral mesoderm (Fig. 3E, 3F). RNA in situ hybridization experiments with egr2b (krox20), a gene expressed in hindbrain rhombomeres 3 and 5, revealed that the two narrow bands of raraa expression (grey arrowheads in Fig. 3E) were anterior to rhombomere 3 (data not shown), whereas the anterior border of the single broad band (open arrowhead in Fig. 3E) lay within rhombomere(s) 6 and/or 7 (Fig. 5A, 5B). At 12 hpf, expression of raraa in the head and tail remained similar to that observed in earlier segmentation stages (Fig. 4A). Expression lateral to the presumptive spinal cord present earlier had disappeared by 11 hpf. We observed additional raraa expression in the eyes and in tissue adjacent to the two faint narrow stripes described above (Fig. 4A, see also Joore et a1., 1994). The distribution ofRara transcripts during neural tube formation in mouse embryos (Ruberte et a1., 1993) is similar to the expression of raraa in zebrafish embryos. Both mouse Rara (Ruberte et a1., 1993) and zebrafish raraa are expressed strongly anteriorly. 14 l2h II 24h I; 48h 21 A B C 0 C v ry hb \ hb. Ie •h~oc ~I ph.$l; 'b ..th h.b hh.,Ic • ovt . "ill F H I < hb, Ie J hb .- ~ D"hb, ..c ~I ph. •tb • tb hb.r:c{b tit! Cb ov poe If L ,~ N 0nc hb.:o:c;: hb,l1Chb,flC tbtb ph. /:.0fb hb hb.lc e pea Fig. 4. raraa (A-E), rarab (F-J) and rarg (K-O) transcript expression visualized at 12 hpf (A,F,K), 24 hpf(B,C,G,H,L,M) and 48 hpf(D,E,I,J,N,O). Refer to text for detailed descriptions. Flat-mount embryos seen from a dorsal view (A,C,E,F,H,J,K,M,O) or a lateral view (B,D,G,I,L,N). Anterior is to the left. (A) Asterisk marks expression in two stripes lateral to the neural tube. (A-C) Open black arrowhead marks broad stripe of raraa hindbrain expression. (F,G) Open black arrowhead marks broad stripe of rarab hindbrain expression. (1) Inset shows a magnified view of rarab expression in the pharyngeal region. (K) Asterisk marks anterior lateral mesenchyme. (N) Grey arrowhead marks staining around eye. Abbreviations: c, cerebellum; e, eye; fb, forebrain; hb, hindbrain; nc, neural crest; ov, otic vesicle; pee, pectoral fin bud; pa, pharyngeal arches; phe, pharyngeal endoderm; SC, spinal cord; t, tectum; tb, tailbud. Scale bar represents 1151lm in 4A,F,K; lOOllm in 4B,G; 90llm in 4L; 42~Lm in 4C,H; 351lm in 4D,E,I,J,N,O; 331lm in 4M. During the segmentation period, raraa was expressed in both neural and underlying non-neural tissue (Fig. 4A-C), but was lost in the tailbud by 24 hpf (Fig. 4B). At 24 hpf, raraa expression was mostly restricted to the hindbrain and anterior spinal cord (Fig. 4B, 4C). Neural expression of raraa was localized in the hindbrain using mRNA expression of egr2b and the otic vesicle as positional markers. A faint band of raraa expression was located within rhombomere 3 and/or 4 (black anowheads in Fig. 5C, 5D) and a sharp anterior border of raraa expression at the boundary of rhombomeres 6 and 7 (open black arrowheads in Fig. 5C, 5D). In the spinal cord, raraa expression at 24 hpfwas graded along the anterioposterior and dorsoventral axes (Fig. 4B; see also Joore et al., 1994). In non-neural tissue, we found raraa expression in the pharyngeal endoderm and posterior neural crest population (Fig. 4B, 4C). Fig. 5. Distinctive expression of raraa (A-D), rarab (E-H) and rarg (I-M). Refer to text for detailed descriptions. Anterior is to the left. Whole-mount embryos viewed dorsally (A,B,E,F). Flat-mount embryos seen from a dorsal view (C,D,G,H,J,K) or a lateral view (I,L,M). raraa expression and egr2b expression in the hindbrain at II hpf(A,B) and 24 hpf (C,D). Open black anowhead marks anterior boundary of raraa expression in the posterior hindbrain (A,C,D). rarab and egr2b expression in the hindbrain at 11 hpf(E,F) and 24 hpf (G,H). Open black arrowhead marks anterior boundary rarab expression in the posterior hindbrain (E,F). Arrowhead marks the anterior boundary rarab expression in medial hindbrain (G,H). Magnified view of rarg expression in the neurocranium and arches (I) and hindbrain striping (J) at 48 hpf. rarg tail expression at 10 hpf (K), 24 hpf (L) and 48 hpf (M). Abbreviations: aner, anterior neurocranium; ba, branchial arches; c, cerebellum; nt, notochord; ov, otic vesicle; pal, pharyngeal arch 1; pa2, pharyngeal arch 2; r3, rhombomere 3; r5, rhombomere 5; tb, tailbud. Scale bar represents 356 /-lm in 5A,B,E,F; 321 /-lm in 5C,D,G,H; 51 /-lm in 51; 73/-lm in 5J; 16 /-lm in 5K; 88/-lm in 5L,M. At 48 hpf, the posterior raraa hindbrain expression was the same as that observed at earlier stages (Fig. 4D, 4E; the anterior border of hindbrain expression was marked by an open black arrowhead). Expression in non-neural tissue was diffuse in the pharyngeal region. We detected restricted raraa expression in a lateral and ventral region posterior to the otic vesicle and anterior to the segmented mesoderm (Fig. 4D, 4E). 16 1.2.3. rarab expression In late gastrula embryos (9 hpf), rarab expression was present in the dorsal epiblast and presumptive tail regions but expression was otherwise absent in the ventral epiblast (Fig. 3G). In contrast to raraa expression, which was broader and was excluded from only the ventral-most epiblast, rarab expression was limited to the dorsal region of the epiblast. By about 10 hpf, three distinct bands of rarab expression perpendicular to the anterioposterior axis appeared in the head: a narrow anterior band (grey arrowhead in Fig. 3H) and two broad posterior bands (black and open black arrowheads in Fig. 3H, 31). The narrow anterior band (grey arrowhead in Fig. 3H) was located ventrally in the presumptive diencephalon. To determine the position of the rarab expression within the presumptive hindbrain, we performed one-color-double in situ hybridization experiments with egr2b. The boundary of the more anterior broad band (black arrowhead in Fig. 3H, 31) was likely located between rhombomere(s) 3 and/or 5 (data not shown), whereas the anterior border of the strong posterior broad band of rarab expression (open black arrowhead in Fig. 3H, 31, 5E, 5F) was located at the posterior border of rhombomere 5 or perhaps within rhombomere 5 (Fig. 5E, 5F). We observed additional rarab expression in tissue lateral to the head which persisted until at least 12 hpf (data not shown). Strong rarab expression also occurred in the posterior neural plate, in the tail bud and adjacent mesoderm (Fig. 3H). Like raraa, expression of rarab was absent from the notochord (Fig. 31). During somitogenesis (12 hpf), the expression of rarab in the head and tail remained the same as that observed in earlier segmentation stages (Fig. 4F). Comparison of rarab and raraa expression patterns in the presumptive spinal cord revealed that rarab expression was distributed throughout the spinal cord, whereas raraa expression was graded along the anterioposterior axis. 17 At 24 hpf, a low level of rarab transcripts appeared ubiquitously distributed. Stronger expression of rarab occurred in the eyes, hindbrain and throughOl,lt the spinal cord. Transcripts from rarab were detected in a cluster of cells within the dorsal diencephalon (data not shown). In the hindbrain, rarab at 24 hpf exhibited only a single band of expression in contrast to the multiple bands observed earlier. Using egr2b mRNA expression and the otic vesicle as positional markers, we located the anterior border of rarab expression in the hindbrain (black arrowhead in Fig. 5G, 5H) between rhombomeres 3 and 4 (Fig. 5G, 5H). Expression of rarab was also observed in non-neural tissues, including the ventral endodermal and mesodermal derivatives such as the pharyngeal endoderm, mesenchyme and tailbud (4G, 4H). At 48 hpf, rarab expression was maintained in the hindbrain, eyes and diencephalon. Neuronal expression of rarab had expanded to more anterior brain regions with faint expression domains in the forebrain and midbrain. Expression of rarab was now prominent in the tectum (Fig. 41, 4J) but had become diffuse in the hindbrain. In the pharyngeal region, rarab expression was still present (Fig. 41). We found additional expression in two patches in ventral and lateral positions posterior to the otic vesicle, extending into the trunk region (41, inset). We also detected expression in the pectoral fm bud (Fig. 4J). Striped hindbrain expression of rarab faded and became diffuse during development, whereas raraa hindbrain expression continued in a strong band until at least 48 hpf. Earlier, at 11 hpf, both raraa and rarab were expressed in prominent bands with well-defined borders in the hindbrain. Within the posterior hindbrain, rarab transcripts appeared to be located more anteriorly than raraa transcripts. The anterior border of the broad band of raraa expression (open black arrowhead in Fig. 4E, 4F, 5A; 5B) lay within rhombomere(s) 6 and/or 7 while the anterior border ofthe two broad bands of rarab expression lay between 18 rhombomeres 3 and 5 (black arrowhead in Fig. 3H, 31, 5G, 5H) and around rhombomere 5 (open black arrowhead in Fig. 3H, 31, 5E, 5F). By 24 hpf the two broad bands of rarab expression fused into a single band of expression whose anterior border lay within rhombomeres 3 or 4 (Fig. 5H), while expression of raraa in the hindbrain remained striped until at least 48 hpf. 1.2.4. rarg expression At 9 hpf, rarg mRNA expression in the epiblast was ubiquitous and extended to the marginal region (Fig. 3J), a pattern which differed from the more dorsally-restricted raraa and rarab expression patterns. Similarly, RARG/Rarg is the RAR/Rar predominately expressed in human and mouse skin (Krust et a1., 1989). At about 10 hpf, rarg expression was limited to the anterior and posterior regions of the embryo. Unlike raraa and rarab, rarg was not expressed in the presumptive hindbrain at bud stage; instead anteriorly, rarg expression was limited to mesoderm adjacent to the head (Fig. 3K, 3L). Mouse Rarg transcripts are also expressed in mesoderm during late gastrulation (Ruberte et a1., 1990). At 12 hpf, the head and tail expression persisted, but expression in the head was now restricted to two stripes parallel to the anterioposterior axis adjacent to the hindbrain (Fig. 4K). Joore et a1. (1994) describes the striped staining as rarg expression in the anterior lateral mesenchyme. One- color-double in situ hybridization experiments with egr2b revealed that the anterior lateral mesenchyme stripes (Joore et a1., 1994) described above are located in mesoderm flanking rhombomeres 3 and 5 (data not shown). At 24 hpf, we began to detect rarg transcripts in the hindbrain, neural crest cells (Fig 4L, 4M; see also Joore et a1., 1994), and in the tai1. Within head mesenchyme, rarg expression occurred in neural crest-derived mesenchyme that will occupy the anterior 19 pharyngeal pouches (Fig. 4M; compare to dlx2 expression, Akimenko et a1., 1994). All three neural crest streams expressed rarg, whereas only posterior neural crest populations expressed raraa. Transcripts of rarg were not detected in the spinal cord (Fig. 4L). Non- neural rarg expression appeared diffuse. At 48 hpf, hindbrain expression increased and showed sharp borders within the hindbrain, anteriorly at rhombomere 3 and posteriorly at the hindbrain/spinal cord border (Fig. 4N, 5J). In addition, a defmed stripe appeared in the more anterior hindbrain, leaving a one-rhombomere-wide gap between the cerebellum and rhombomere 3 (Fig. 4N, 5J). In the dorsal hindbrain, rarg was expressed in narrow stripes at intervals one rhombomere wide (Fig. 5J), suggesting expression was located at rhombomere boundaries. Head expression of rarg at 48 hpfwas more similar to raraa expression which is limited to the hindbrain, than rarab expression which was located in anterior brain regions, in clusters in the forebrain and the midbrain- hindbrain domain. We found prominent rarg expression in pharyngeal arches I and 2, in the more posterior arches, arches 3 to 7, and in the anterior neurocranium (Fig. 51). Expression in the anterior pharyngeal arches was unique to rarg; raraa and rarab expression appeared in more posterior branchial regions. In addition at 48 hpf, expression of both raraa (Fig. 4D, 4E) and rarab (Fig. 41, 4J) in the pharyngeal region was restricted to a lateral and ventral region bounded by the otic vesicle and the segmented mesoderm, whereas expression of rarg (Fig. 4N, 40) in the anterior pharyngeal region was widely distributed. We also detected rarg expression in cells, probably mesenchymal cells, separating the eyes and forebrain (grey arrowhead in Fig. 4N) and in the pectoral fin buds (Fig. 4N, 40). Although raraa and rarab are more closely related phylogenetically, expression of rars in the fm and tail bud is more similar between rarg and rarab than raraa and rarab. Tailbud expression of rarg was still visible at 48 hpf (Fig. 5K-M), in contrast to tailbud expression of raraa which was 20 absent by 24 hpf (Fig. 4B) and expression of rarab, which was lost from the tailbud after 24 hpf(data not shown). 1.3. Conclusions In general, rar expression patterns coincide with the mRNA expression patterns of genes involved in the synthesis, transport, and degradation ofRA, including retinaldehyde dehydrogenase 2 (raldh2), cellular retinoic acid-binding protein 2 (crabp2s) and cyp26. In particular, expression of raraa (Fig. 4E, 4F) and rarab (Fig. 4H, 41) in the paraxial mesoderm overlaps with raldh2 expression (Grande! et a1., 2002) and crabp2a and crapb2b expression (Sharma et a1., 2005) at 11 hpf. During segmentation, expression of cyp26bl, which encodes the retinoic acid degrading enzyme Cyp26b1, overlaps raraa and rarab in the hindbrain (Zhao et a1., 2005). Later, at 48 hpf, rarg and cyp26bl hindbrain expression overlap instead (Zhao et a1., 2005). It may seem curious that the RA-degrading enzyme and RA receptor proteins are expressed at the same time and place. This coexpression may reflect the fme regulation of RA signaling in the developing embryo. These studies of the syntenic and phylogenetic relationships between zebrafish and human RAR genes, as well as the early expression patterns of the zebrafish genes, will provide the basis for future studies of rar gene function during development. 2. Experimental procedures 2.1. Cloning of rar genes Clones of raraa and rarab were donated by P. Kushner and S. E. Stache1. We cloned a 486-basepair fragment of rarg (Accession number: S74156) using the following primers: 21 5'-CACCCGCCCTGCTCACGA-3' and 5'- GAACCCGTTGAAAGTACACTGTTAAAAG-3' . 2.2. Whole mount in situ hybridization Zebrafish embryos (AB strain) were raised as described by Westerfield (1995) and staged by hours post fertilization (hpt) at 28.5°C according to Kimmel et al. (1995). To prevent pigment formation, embryos were treated with 0.003% 1-phenyl-2-thiourea (PTU) in embryo medium at 12 hpf. RNA in situ hybridization was carried out according to standard protocols described in Hauptmann and Gerster (1994). The following RNA probes were used: krox20(Oxtoby and Jowett, 1993; renamed early growth response 2b (egr2b), refer also to ZFIN (see http://zfm.org/cgi-bin/webdriver?MIval=aa-markerview.apg&OID=ZDB-GENE-980526- 283; Sprague et al., 2002), raraa (Accession number: L03398; also referred to as ram by Joore et al., 1994), rarab (Accession number: L03399) and rarg (Accession number: 574156; also referred to as rary by Joore et al., 1994). Embryos were viewed with Leica MZ6 or MZ9 stereomicroscopes or with a Zeiss Axioplan microscope, and were photographed using a Nikon Coolpix 990 or 995 digital camera. 2.3. Sources of additional gene expression data Some expression data for zebrafish genes cited here were retrieved from the Zebrafish Information Network (ZFIN), the Zebrafish International Resource Center, University of Oregon, Eugene, OR 97403-5274; World Wide Web (URL: http://zfm.org/), August, 2005 and NCBI Entrez Gene (Zhang et al., 2005; World Wide Web (URL: http://www.ncbi.nlm.nih.gov/ entrez/query.fcgi?db=gene). 22 Some mouse gene expression data cited in this paper were retrieved from the Gene Expression Database CGXD), Mouse Genome Informatics Web Site, The Jackson Laboratory, Bar Harbor, Maine. World Wide Web CURL: http://www.informatics.jax.org), August, 2005. Some human gene expression data cited in this paper were retrieved from The Human Genome Organisation (HUGO), London. World Wide Web CURL: http://www.gene.uc1.ac.uk/nomenc1ature/index.html). September, 2005. 23 CHAPTER ill CHARACTERIZATION OF RETINOID-X RECEPTOR GENES RXRA, RXR13AJ RXR13B AND RXRG DURING ZEBRAFISH DEVELOPMENT The work described in this chapter was previously published in "Gene Expression Patterns," Vol. 6. I share first authorship with A. Tallafuss. We were responsible for the majority of data collection, data analysis and writing. A. Tallafuss cloned rxrg. Y.L. Yan and L. Dudley cloned rxra, rxrba, and rxrbb. J.R. Postlethwait completed the syntenic analysis and contributed to writing. J.S. Eisen also contributed to writing. 1. Results and discussion Retinoic acid (RA) signaling has been implicated in a variety of developmental processes including patterning of the central nervous system (CNS), paired appendages and other organs (Gavalas and Krumlauf, 2000; Grandel et al., 2002; Jiang et a1., 2002). RA acts by binding to heterodimeric receptors composed of a retinoic acid receptor (Rar) and a retinoid-X receptor (Rxr); these receptors bind to retinoic acid response elements (RAREs) and modulate transcription of target genes [reviewed by (Bastien and Rochette-Egly, 2004)]. Rxrs also form heterodimers with other nuclear receptors, including thyroid-hormone receptors (TRs), peroxisome proliferator-activated receptors (PPARs), Vitamin D receptor (VDR), and liver X receptor (LXR) (Umesono and Evans, 1989; Willy and Mangelsdorf, 1999), indicating Rxrs mediate expression ofa large variety of hormone-responsive genes. Malfunction of RA receptors has been linked to several cancers (Li et a1. 1998; Soprano et al., 2004). Currently, RA and its analogs are under investigation as chemopreventative 24 agents in cancer therapy (Okuno et a1., 2004) and Rxrs as potential targets for cancer chemotherapies (Crowe et a1., 2004). In this paper, we focus on retinoid-X receptors in zebrafish, a model for molecular and genetic studies of vertebrate development. Previous studies of rxr transcript accumulation during zebrafish development have provided a foundation for understanding when and where these genes function (White et a1., 1994; Jones et aL, 1995; Thisse et a1., 2004; Kawakami et a1., 2005), but knowledge of expression during more developmental stages is needed and a stronger comparison of various receptor expression patterns is necessary for a complete understanding of rxr gene expression. Incomplete knowledge ofgene expression patterns, in addition to recent availability ofgenomic information, prompted us to revisit the expression of rxr genes in concert with understanding their phylogenetic relationships to human orthologs. Based on conserved syntenies between zebrafish and human chromosomes, we show that some of the zebrafish rxr genes were previously assigned inappropriate orthologies, and thus incorrect names. We provide detailed information about the phylogenies of the four known zebrafish rxr genes as well as more detailed descriptions of their expression patterns. This information is crucial for future studies of the functions of rxr genes during development and for understanding how malfunction of these genes may contribute to diseases such as cancer (Okuno et a1., 2004). 1.1. Orthologies of zebrafish rxr genes To properly compare expression dynamics among species, it is essential to evaluate orthologous genes. Two types of data help establish orthologies: phylogenetic analysis and conserved syntenies. 25 Phylogenetic analysis of amino acid sequences rooted on related genes from protostomes and non~vertebratechordates revealed three clades of vertebrate Rxrgenes that arose after the divergence ofvertebrate and non-vertebrate chordate lineages (Fig. 1). These relationships are consistent with the explanation that these genes appeared in the genome amplification events around the origin of vertebrates (Holland and Garcia~Femandez,1994). Our data suggest that rxra (AAH59576; Strausberg et aL, 2002) and rxrg (Accession number NP_571228; Jones et aL, 1995) are incorrectly named in the literature. Our results showed that the zebrafish gene formerly called rxrg (Accession number NP_571228; Jones et aL, 1995) groups with high bootstrap support in the tetrapod Rxra clade, suggesting that NP_571228 is an ortholog ofRxra, rather than an ortholog ofRxrg as previously thought; thus, we have renamed NP_571228 as rxra (Table 1). Reciprocally, the gene previously named rxra (AAH59576; Strausberg et aI., 2002) clusters with the tetrapod Rxrg clade, but with low bootstrap support, so additional data, described below, are required to confum whether AAH59576 is the ortholog ofRxrg. Zebrafish have at least two additional rxr genes: rxrb (AAH54649, Strausberg et aL, 2002) and rxrd (NP_571313, Jones et al., 1995). The tree shows with good bootstrap support that both genes are orthologs ofRxrb (Fig. 1), thus we have renamed them rxrba (formerly rxrb) and rxrbb (formerly rxrd; see also Table 1). Human RXRGmaps to Hosa1q24 (Fig. 2Bl), andAAH59576 (formerly called rxra but which clusters with Rxrg in the phylogeny of Fig. 1) is on Zv5_scaffold1017 in the zebrafish genome. The orthologs of several genes located nearby in Zv5 are also located near RXRG in the human genome (Fig. 2B2 and 2B3). Furthermore, the orthologs of additional genes near RXRG in the central part ofHosa1q are located with AAH59576 on LG2 (Fig. 2B2-2B4). Of 165 protein-coding loci mapped to LG2 in the HS mapping panel (Woods et al., 2005), 22 have orthologs on Hosal (the location ofRXRG), and none from Hosa9 (the 26 location ofRXRA). Because analysis of conserved syntenies provides strong support for the orthology ofAAH59576 to RXRG and corroborates the phylogenetic analysis, we have renamed this sequence rxrg rather than its previous name of rxra. The Zv5 assembly of the zebrafish genome shows the rxrg scaffold on LG20, but we mapped this sequence to LG2 (Postlethwait et aI., 1998) and this location was later confrrmed on four independent mapping panels (see http://zfm.org/cgi-bin/webdriver?Mlval=aa- markerview.apg&OID=ZDB-GENE-980526-36), so it is likely that rxrg is assigned to the wrong linkage group in Zv5. RxdfT bl 1 S this paper; IJones et aI., 1995; 2Strausberg et al., 2002; 3ZFIN (2003; http://zfin.org/cgi-bin/webdriver?Mlval=aa-pubview2.apg&OID=ZDB- PUB-030508-1); 4HUGO (2005; http://www.gene.uc1.ac.uk/nomenc1ature/index.html); 5MGI (2005; http://www.informatics.jax.org/) a e urvey 0 recent an preVIOUS r gene names. Zebrafish Accession number Previous Zebrafish Human MouseGene Gene Ortholog ortholog * NP 571228 RXRv1;rxri RXRA4 Rxreirxra I RXR61; rxrb2,3; rxrba* AAH54649 RXRF/; rxrr! RXR.J34 Rxrb5 rxrbb* NP 571313 RXR61; rxrd3 * AAH59576 RXRa1; rxra2,3 RXRG4 Rxrirxrg * 27 RXRA Hosa Rxra Mumu Rxra_Gaga Rxra Xela ~----- Rxra Caau '--------- rxra Dare (rxrg) RXRG Rosa Rxrg_Mumu '----- Rxrg_Gaga 1-- Rxrg_XeIa r------ rxrg_Dare (rxra) '------------ Rxrg_Paol (Rxra) 1000 RXRB Hosa Rxrb Mumu '-------Rxrb Xela ~ rxrba_Dare (rxrb) ...----- Rxrb_Paol (Rxrg) Rxrba Oria (Rxrb) ...-- rxrbb Dare (rxrd) '-- Rxrbb Teni (Rxr) r-------------- Rxr pomi '- Rxr Brfl Rxr Thel Rxr_Lyst 967 438 994 594 990 585 997 899 831 669 1000 911 651 1000 Usp_Apme '-- Rxrl Lomi 0.02 ~ Fig. 1. Phylogenetic analysis of retinoid-X receptor proteins. Amino acid sequences were aligned by Clustal-X and trimmed to include unambiguously aligned regions; phylogenetic analysis used the neighbor-joining method (Saitou and Nei, 1987). Numbers at nodes are bootstrap values out of 1000 runs. Alignments are available on request. Species abbreviations: Apme, Apis melli/era (honey bee); Brfl, Branchiostomafloridae (amphioxus); (Caau, Carassius auratus (goldfish); Dare, Danio rerio (zebrafish); Gaga, Gallus gallus (chicken); Rosa, Homo sapiens (human); Lomi, Locusta migratoria (locust); Lyst, Lymnaea stagnalis (great pond snail); Mumu, Mus musculus (mouse); OrIa, Oryzias latipes (medaka); Paol, Paralichthys olivaceus (flounder); Pomi, Polyandrocarpa misakiensis (an ascidian Urochordate); Teni, Tetraodon nigroviridis (pufferfish); Thcl, Thais clavigera (rock shell snail); Xela, Xenopus laevis (frog). Sequence accession numbers: RXRA Rosa AAH63827; Rxra Mumu NP_035435; Rxra Gaga XP_415426; Rxra Xela P51128; Rxra Caau AA022211; rxra Dare NP_571228; RXRG Rosa NP_008848; Rxrg Mumu NP_033133; Rxrg Gaga NP_990625; Rxrg Xela P51129; Rxrg Dare AAH59576; Rxra Paol BAB71758; Rosa CAI95622; Rxrb Mumu NP_035436; Rxrb Xe1a AAH72132; Rxrb Dare AAH54649; Rxrg Paol BAB71759; Rxrb OrIa BAD93255; rxrd Dare NP_571313; Rxr Teni CAG12025; Rxr Pomi BAA82618; Rxr Brfl AAM46151; Rxr Thc1 AAU12572; Rxr Lyst AAW34268; Usp Apme NP_001011634; Rxr1 Lomi AAQ55293. 28 Phylogenetic analysis (Fig. 1) showed that rxrba and rxrbb are co-orthologs ofRxrb, but did not reveal whether they arose by a recent tandem duplication event or in the ancient teleost genome duplication (Amores et al., 1998; Postlethwait et al., 1998; Taylor et a1., 2003). Analysis of conserved syntenies showed that rxrba and rxrbb are on LG19 and LG16, respectively (Fig. 3), which are duplicated chromosomes arising from the teleost genome duplication (Amores et a1., 1998; Naruse et al., 2004; Woods et al., 2005). In summary, analysis ofphylogenies and conserved syntenies radically changes the nomenclature of zebraflSh rxr genes (see also Table 2), with the old rxrg becoming rxra, the old rxra becoming rxrg, the old rxrb becoming rxrba, and the old rxrdbecoming rxrbb. These revisions will help us to compare the expression patterns of these genes with their true orthologs in tetrapods. 1.2. Expression of rxr genes during zebrafish embryonic development Having determined the orthologous relationships of zebrafish rxr genes, we next assayed their spatial and temporal expression patterns by whole mount RNA in situ hybridization during embryonic development. Although some aspects of rxr gene expression are available on ZFIN (http://zfm.org/; Sprague et a1., 2001), current descriptions are incomplete and differ slightly from our results, discussed below. Our observations began in the cleavage period (0.75-2 hours post fertilization, hpf), extended through the blastula (2.25- 4.66 hpf), gastrula (5.25-10 hpf), segmentation (10.33 -22 hpf), and pharyngula (24-42 hpf) periods, and concluded in the hatching period (48-72 hpf). Overall expression of rxr genes at representative time points is summarized in Fig. 4 and 5. Fig. 6 displays magnified views of rxr gene expression in specific tissues. 29 Fig. 2. Genomic analysis of conserved syntenies for zebrafish rxr genes. A. rxra (NP_S71228). Al shows human chromosome 9 with regions relevant to zebrafish rxra expanded in A2. The location of human genes and gene names are from NCBI Map Viewer (http://www.ncbi.nlm.nih.gov/mapview/map_search.cgi?%20). with genes transcribed downwards shown on the right of the central line indicating the chromosome and genes transcribed upwards to the left of the line. A3 shows the rxra-containing region of the Zv4 zebrafish genome assembly (http://www.ensemb1.org/Danio_rerio/), and A4 shows zebrafish linkage group 5 with loci orthologous to Hosa9 genes shown in brown (Woods et a1., 2000). B. rxrg (AAH59576). BI shows Hosal, and B2 shows relevant portions expanded. B3 is the portion of ZvS containing rxrg, and B4 is LG2 with Hosal orthologs shown in blue. C. RXRB co-orthologs. CI and CS show the duplicated zebrafish linkage groups containing rxrba and rxrbb, with Hosa6 orthologs indicated in yellow. C2 and C3 show the human chromosome and the region surrounding RXRB. C4 shows the portion of Zv4 containing rxrba. The rxrbb sequence was not assembled in Zv4 or Zv5. 30 Ai A2 ~ A3 A4 a s:Q Zv4CIlr5 LGS 0 OJ. ~:f; IIosa9 I~ 0.0' .. 1'1 _ -;:H "Jl'l =:= 14.7 ,I> i'.lf - ?3.1 -::'1.' , 't. 31A'II) g 35,5 "t"l 490 h\l -'I)." N : I~: (J:) Gl.? ') .. ,~ ~ e, r> 0\ --.j G8.9 0 Jm-;: 70.GOJ wZ .. ffifJ) OJ 8GJ~ mO 88.80 Z:p (f) b_ fJ);U0 ). tJ-G.4 ~ 0-1:po • ;ug 1114 -1 02~ - 00 11!l.' 0- gg -,- 12GO Sl.1pm0- g 0> 1474 GOF22 - "'"m~ ~o·'TI~ • .,. ~S 81 82 83 84 LG2 :!l 00 l_l nB 0=r. 7.5;;: - . .~ (j) 12.3 ,f1osa1p ~ ~ l~i~;~ 0 23U - .U If h~ )< ,Jt1 j, F'"~\j, 40A Ip22 i,,~J m 512IW- '" • 588q ! ,n 'XIt~. ;,:l~~~ GG4 t'i2~ c 71.41124 () Iq2S ~ 70.7 ,,"g?h1-1'31 l'1~L 1.:t41 89 ?1-,42: I~l~ __. 9·1.0 i 1021 copll209.9 'P-HI '" " iw 0 117.G Q$rOlll '" G) s: A I125.4 flCC4D"13~ :r~::y]~ 05.3 I. AIG5't378 145.G flvW020D7 1!)2 P. ISO.1 1GO.3 I .. 1\W202015 31 C1 C2 C3 C4 ~ lGlD C5 HosaG Fr lG 16 7'G~7. lOU lGI9 ,,"' ."" ~ " -I· ~ 00 i- l1271I J{dlll! I 12 ~\ ell," I ~ 4.9 baleHosa6 Tuoa ;-!'V _ Nro • 19 q RI t 20: /tCGOfII!il co ::l Ij;) 71 9 l6275lubb5 v:u1O noll 21 G ~i'C ,con to I. ~p22 j VARSZ --.U~k.:l,1 .~ JS" f~~I:~ "_flXI-HI =!> ~7.7 i IlCrh.=t mhclLlc I J2. -VEGF ~ pSmbOn daJlx t.'.J1U4 ~8 2 r~-- 0-6~ I - hrt i 55.3 '''''~ 5J2 l\51151t.u - BCKDHB 3 "''') 11~ :TfiG 0 .,1(' ~ POIJJ!'2 ~- ~ 1;66 nilme-12d 717. .,21 COQ3 c::l ro 74.R -l ck2b',~13 fy.m47 79 1 (,·~~2 - Slr,,11 l\> 797 ISOx4u-0 '"r.6111tl51 8fifi 61'::) - • '" 3t.12'4 - • 3- 921 ,- 210002clcJn12 940 0- CT"1~S • 0> «> I b~~ _ n CT 102.0 - (lIx6"Irk I I 101.fi • 0;i 0,,It,- • u 110.0 - BC070307 r;u.J21 112.0 ,-,r- _ IV N 'J~Z rio~ .- ~ ::r0" ;;= 1223 _ 1-5183n(:0763/,7 .- !21.i' ~~ '"IV ~p.IIl' I l:lG.l -, <0 I --.p~lf IY•.3 w CT ll9992'" 0> 1298 i~~ Fig. 3. Genomic analysis of conserved syntenies for zebrafish RXRB co-orthologs. C 1 and C5 show the duplicated zebrafish linkage groups containing rxrba and rxrbb, with Hosa6 orthologs indicated in yellow. C2 and C3 show the human chromosome and the region surrounding RXRB. C4 shows the portion of Zv4 containing rxrba. The rxrbb sequence was not assembled in Zv4 or Zv5. I.Z.I. Maternal expression of YXY genes Expression ofmRNA transcripts of Rxra, Rxrba, Rxrbb and Rxrg was observed in early cleavage stages at 1.5 hpf (Fig. 4A), showing that these genes are maternally expressed. Cleavage and blastula period expression patterns of all four zebrafish rxr genes were similar, so Fig. 4A and 4B show only rxrba and rxra expression. We detected rxra, rxrba, rxrbb and rxrg transcripts from the blastula period through at least 48 hpf (Fig. 4), confirming previous results from Northern blots (Jones et aL, 1995). During the blastula period, RNA in situ hybridization revealed that all rxr gene expression patterns are similarly uniform throughout the embryo. Distinctive rxr expression patterns arising during the gastrula period and 32 persisting through later developmental stages are described below. Expression patterns of rxrba and rxrbb were similar, so they are described in a single paragraph. 1.2.2. rxra expression During mid-gastrulation (9hpf), rxra transcripts appeared throughout the embryo (Fig. 4C). Stronger expression was observed in the animal pole and to a lesser extent in the dorsal midline (data not shown). Between 10-11 hpf (late gastrulation), rxra expression remained ubiquitous (Fig. 4D, 4E). At 12 hpf, rxra was strongly expressed in the tail, but in the rest of the embryo, rxra expression was weak and ubiquitous (Fig. 5A). At 24 hpf, faint rxra expression was found in the forebrain and eyes, and the pharyngeal endoderm (Fig. 5B, 5C); expression in tail mesoderm was also maintained (Fig. 5B). At later stages, expression remained in the eye and the pharyngeal endoderm, but was not detected in the brain (Fig. 6A). At 48 hpf, anterior expression was limited to the ventral-most cell rows underlying the head (Fig. 5D). Additionally, rxra expression was detected medially within the mesoderm of the pectoral fin bud (Fig. 5E, 6B), placing Rxra in the appropriate position to transduce retinoic acid signaling for limb bud outgrowth (Stratford et a1., 1996). Expression of rxra in the tail was maintained until at least 48 hpf (Fig. 6C). Currently, ZFIN has no RNA in situ hybridization data available for rxra (formerly rxrg). 1.2.3. rxrba and rxrbb expression Despite significant differences in sequence, the spatiotemporal expression patterns of rxrba and rxrbb appeared similar. Weak expression, probably the result oflow transcript levels, prevented detailed analysis, so we cannot exclude the possibility of subtle differences 33 in rxrba and rxrbb expression. Because rxrba and rxrbb expression patterns appeared similar, our description of rxrbb expression represents both rxrba and rxrbb expression. Spatially restricted rxrbb expression (formerly rxrd) has not been described in ZFIN (see http://zfin.org/cgi-bin/webdriver?Mlval=aa-fxfigureview.apg&OID=ZDB-FIG-050630- 3780), and there are no RNA in situ hybridization data available in ZFIN for rxrba (formerly rxrb). __----=1-'--,s"------=2.:..:.:h'-----__1 1 4_,_7_h _ K B J D G A c F 8-9h II_1....._---- _ Fig. 4. Expression of rxra (B, C-E), rxrba, (A), rxrbb (F-H) and rxrg (I-K) visualized at 1.5-2 hpf(A), 4.7 hpf(B), 8-9 hpf(C,F,I) and 10.5-11 hpf(D,E,G,H,J,K). Dorsal view (A,E,H,K). Lateral view; anterior up and dorsal to the left (B,C,D,F,G,I,J). Abbreviations: v, vegetal pole; yp, yolk plug; tb, tailbud. Scale bar represents 93 J.l.m in A; 106 J.l.m in B, C-K. 34 During gastrulation, between 8-10.5 hpf, rxrbb expression remained ubiquitous (Fig. 4F). At 12 hpf, rxrbb (Fig. 5F) expression was similar to that in earlier periods but staining appeared fainter. I 12h II 24h II B c D W phe pbe ~ • 'b tb tb fb • .t vph ~' G H lib hb.Q 'b... phe~ fb 'b ill 'b . .t vph ~' L hb,BC M Nhb,sc hb, Be 48h ov pee rob hb 'b o Fig. 5. Expression ofrxra(A-E), rxrbb(F-J) and rxrg(K-O) at 12 hpf(A,F,K), 24 hpf (B,C,G,H,L,M) and 48 hpf(D,E,I,J,N,O). Refer to text for detailed descriptions. Flat-mount embryos seen from a dorsal view (A,C,E,F,H,J,K,M,O) or a lateral view (B,D,G,I,L,N). Anterior is to the left. Abbreviations: e, eye; fb, forebrain; hb, hindbrain; mb, midbrain; ncr, neural crest; phe, pharyngeal endoderm; ov, otic vesicle; sc, spinal cord; st, stomadeum; tb, tailbud; vph, ventral tissue in pharyngeal region. Scale bar represents l15)lm in 4A,F,K; lOO)lm in 4B,G; 108)lm in 4L; 40)lm in 4C,M; 35)lm in 4H; 32)lm in 5D,E,I,J,N,O. At 24 hpf, rxrbb expression became spatially restricted; in the neural tube, we detected expression mainly in the ventral diencephalon (Fig. 5G, 5H). We also detected expression in non-neural tissues: diffusely in pharyngeal endoderm and faintly in trunk and tail mesoderm (Fig. 5G, 5H). In the mesoderm, rxrbb expression occurred in the medial cell rows of each somite, along the dorsoventral axis (Fig. 6D, 6E). We describe these "somite stripes" of expression in this section as it was most evident in embryos stained for rxrbb expression (Fig. 5D, 5E), but weak expression in the medial somite stripes was evident for all rxr genes (data not shown). The timing and location of rxr expression in somites is interesting because it appears to be localized to the region traversed by motor axons, neural crest cells and sclerotome cells (Lewis and Eisen, 2004), suggesting that these receptors could be 35 involved in establishing or maintaining that pathway. At 48 hpf, rxrba and rxrbb were expressed in broad domains within the forebrain, eyes, midbrain and anterior hindbrain (Fig. 51, 5J). Expression in the anterior endoderm described above was eventually restricted to the ventral-most cell layers (Fig. 6F). I 32h II 48h 48h A B C pee