Dissertation Approved by

/ Major Advisor

r?<-^

Dean A PHARMACOLOGICAL STUDY OF CALCITONIN GENE-RELATED PEPTIDE

RECEPTOR SUBTYPES

By

Boyd R. Rorabaugh

A DISSERTATION

Submitted to the faculty of the Graduate School of Creighton University in partial

tulEllment of the requirements for the degree of Doctor of Philosophy in the

Department of Pharmacology

Omaha, Nebraska, May 2002 ABSTRACT

Calcitonin gene-related peptide (CGRP) receptors are divided into CGRP] and

CGRP2 receptor subtypes based on their affinity for CGRPg.37 in functional assays.

CGRP] receptors are blocked with high affinity (Kg < 100 nM) by CGRPg.37, and CGRP2 receptors are blocked with low affinity (Kg > 100 nM) by this antagonist. The CGRP] receptor has been cloned and well characterized. In contrast, the CGRP2 receptor has not been well characterized and has only been identified by its low affinity for CGRPs.37 in functional assays. Furthermore, previous investigators reported that radioligand binding assays using [*^I]CGRP do not support the existence of the CGRP2 receptor. The goals of my work were to further characterize the CGRP2 receptor and to determine why this receptor cannot be identified by radioligand binding.

Porcine coronary arteries were used to study CGRPi receptors because CGRPgg? blocks CGRP-induced relaxation of this tissue with low affinity. I used RT-PCR, radioligand binding, and affinity values from previously reported isolated tissue studies to compare the CGRP receptor in coronary arteries to the cloned CGRP] receptor I identified mRNA encoding calcitonin receptor-like receptor (CRLR) and receptor activity modifying protein (RAMP) 1 in porcine coronary arteries. In addition, the ligand binding characteristics of CGRP receptors in coronary arteries and in CGRP] receptor-transfected

HER 293 cells were similar. There was also a high correlation between antagonist affinities determined by radioligand binding and functional relaxation assays. My data support the conclusion that CGRPg.37 blocks CGRP] receptors with low affinity in functional relaxation assays using porcine coronary arteries.

iii SV40LT-SMC cells were also used to study CGRP2 receptors. In contrast to coronary arteries, the putative CGRP? receptor in this cell line was identified as an adrenomedullin receptor that has low affinity for CGRPg.37. The inability of ['^1]CGRP to label adrenomedullin receptors is consistent with the fact that the CGRP2 receptor has not been previously identified by radioligand binding.

In summary, my data do not support the widely accepted view that CGRP- induced responses are mediated by two different CGRP receptor subtypes. Rather these data support the conclusions that the putative CGRP2 receptor is an adrenomedullin receptor and that CGRPg.37 blocks CGRP] receptor mediated responses with low affinity in some isolated tissues. I anticipate that this work will have a significant impact on the currently accepted classification of CGRP receptors.

iv PREFACE

Portions of the work described in this dissertation have been published in:

Rorabaugh BR, Abel PW, Smith DD, Scofield MA (2002) Putative CGRP2 Receptors are Identified as Adrenomedullin Receptors in SV40LT-SMC Cells. E/f-SEE Joarna/ 16:A576.

Rorabaugh BR. Scofield MA, Smith DD, Jeffries WB, Abel PW (2001) Functional Calcitonin Gene-Related Peptide Subtype 2 Receptors are Identified as Calcitonin Gene-Related Peptide 1 Receptors by Radioligand Binding and Reverse Transcription-Polymerase Chain Reaction. Joaraa/ o/*P/zarTMaco/ogy azzJ Eyperzwe/ztn/ P/zerapeatzc^. 299:1086-1094.

Rorabaugh BR, Abel PW, Smith DD, Scofield MA (2001) CGRP Receptor Signaling Pathways in SK-N-MC and HER 293 Cells. ETfEEP Joarrza/ 15:A931.

Rorabaugh BR, Abel PW, Smith DD, Scofield MA (2000) Evidence for the CGRP] Receptor in Porcine Coronary Artery. P4EEP Joaraaf 14:A1404.

v D eJicoteJ to /?;y /oon'/y.

vi ACKNOWLEDGMENTS Many people have contributed to the successful completion of this work. This project could not have been completed without the encouragement and support of my mentor, Dr. Margaret Scofield. I also gratefully acknowledge the members of my research committee, Dr. Peter Abel, Dr. Frank Dowd, Dr. Joe Knezetic, and Dr. David Smith for their insight and guidance throughout this project. I also thank Dr. William Jeffries, Dr. Michael Bradley, and Dr. Charles Bockman for allowing me to use their laboratory equipment and for their contributions to my professional development. Joe Haun of J& J Quality Meats (Elkhom, Nebraska) also contributed significantly to my research. Joe went out of his way to provide the pig tissues that 1 needed and always made me feel welcome at the slaughterhouse. I enjoyed the mornings that I spent at his business, and I learned some things about butchering too. 1 am also thankful for the support of my friends at Omaha Central Church of the Nazarene. They helped keep me focused on the things that are really important in life. This work could not have been completed without the support of my family, i am particularly grateful for the encouragement of my parents, grandparents, my sister, and my in-laws. These people were constant reminders that there is life after graduate school. I especially appreciate the moral support of Tom and Karen, f have enjoyed our weekends together in Kansas City, and we have become good friends. Finally, I am most deeply indebted to my wife for tolerating my late nights and weekends in the lab, for her constant encouragement, and for making all the cookies that helped keep my presence welcome at the slaughterhouse. I could not have done it without you!

vii TABLE OF CONTENTS

ABSTRACT...... iii

PREFACE...... v

DEDICATION...... vi

ACKNOWLEDGEMENTS...... vii

LIST OF TABLES...... xi

LIST OF FIGURES...... xii

ABBREVIATIONS...... xiv

CHAPTER 1. BACKGROUND AND OBJECTIVES...... 1 Intercellular Communication...... 1 Calcitonin Gene-Related Peptide...... 1 CGRP Receptors and Accessory Proteins...... 6 Objectives...... 12

CHAPTER 2. FUNCTIONAL CGRPi RECEPTORS IN PORCINE CORONARY ARTERIES ARE IDENTIFIED AS CGRP, RECEPTORS BY RADIOLIGAND BINDING AND REVERSE TRANSCRIPTION-POLYMERASE CHAIN REACTION...... 13 Abstract...... 13 Introduction...... 14 Materials and Methods...... 16 Chemicals and Reagents...... 16 Peptide Synthesis...... 16 RNA Isolation...... 17 RT-PCR...... 17 Cell Culture...... 20 Membrane Preparations...... 20 ['*^I]ha.-CGRP Binding Kinetics...... 21 Competition Binding Assay...... 22 Measurement of Intracellular cAMP...... 22 Data Analysis...... 23 Results...... 24 Identification of CRLR and RAMP 1 mRNA in Porcine Coronary Arteries...... 24 Kinetics of ['^I]ha-CGRP Binding to CGRP Receptors...... 26 Binding of CGRP Receptor Ligands to CGRP, Receptors in HEK 293 Cells ,29

viii Binding of CGRP Receptor Ligands to CGRP Receptors in Porcine Coronary Arteries...... 33 Comparison of Ligand Affinities for CGRP Receptors in Porcine Coronary Arteries and HER 293 Cells...... 34 Activation of CGRP] Receptors by a Putative CGRPi Receptor Selective Agonist...... 35 Discussion...... 39

CHAPTER 3. IDENTIFICATION OF AN ADRENOMEDULLIN RECEPTOR AS THE PUTATIVE CGRP2 RECEPTOR IN SV40LT-SMC CELLS...... 45 Abstract...... 45 Introduction...... 46 Materials and Methods...... 48 Chemicals and Reagents...... 48 Cell Culture...... 49 Measurement of Intracellular cAMP...... 49 Membrane Preparation...... 50 P.adioligand Binding Assay...... 50 RNA Isolation...... 51 Identification of mRNA Encoding CRLR and RAMP by RT-PCR...... 51 Data Analysis...... 54 Results...... 55 Identification of Cells Expressing the Putative CGRP2 Receptor...... 55 Radioligand Binding in SV40LT-SMC and DDT Membranes...... 58 Identification of mRNA Encoding CRLR and RAMPs in SV40LTSMC cells...... 61 Radioligand Binding in SK-N-MC Membranes...... 65 Identification of Adrenomedullin Receptors in Rat Vas Deferens...... 73 Discussion...... 76

CHAPTER 4. TISSUE DEPENDENT FACTORS THAT INFLUENCE THE AFFINITY OF ha-CGRPg^?...... 84 Introduction...... 84 Enzymatic Degradation of ha-CGRPg.3 7 ...... 8 6 Release of Endogenous CGRP...... 92 Conformational States of the CGRP Receptor...... 93 Conclusion...... 96

CHAPTER 5. COMPARISON OF LIGAND BINDING CHARACTERISTICS OF PORCINE AND HUMAN CGRP, RECEPTORS...... 97 Introduction...... 97 Materials and Methods...... 98 Cell Culture, Membrane Preparation, and Radioligand Binding...... 98 Data Analysis...... 98 Results...... 100 Ligand Binding to CGRP, Receptors in Human SK-N-MC Cells...... 100

rx Comparison of Ligand Affinities for CGRP Receptors in Human SK-N-MC CeUs and Porcine CGRP] Receptor-Transfected HEK 293 Cells...... 100 Discussion...... 102

CHAPTER 6 . SUMMARY AND IMPLICATIONS FOR THE DEVELOPMENT OF THERAPEUTIC AGENTS...... 104 Summary...... 104 Implications for the Development of Therapeutic Agents...... 107

APPENDIX A. INFLUENCE OF CGRP ON INTRACELLULAR CALCIUM IN SK-N-MC CELLS AND PORCINE CGRP) RECEPTOR-TRANSFECTED HEK 293 CELLS...... 112 Introduction...... 112 Materials and Methods...... 112 Measurement of Intracellular cAMP...... 112 Measurement of Intracellular Calcium...... 113 Data Analysis,...... 114 Results...... 114 CGRP-Induced cAMP Synthesis...... 114 CGRP-Induced Changes in Intracellular Calcium...... 115 Discussion...... 118

APPENDIX B. Partial Mapping of the Calcitonin Receptor-Like Receptor Gene...... 120 Introduction...... 120 Methods...... 120 Results and Future Directions...... 121

REFERENCES...... 124

x LIST OF TABLES

Table 1. Examples of peptides and proteins that are involved in intercellular communication...... 2

Table 2. Classification criteria for CGRP receptor subtypes...... 9

Table 3. Ki values of CGRP analogs in pig coronary arteries and pig CGRP} receptor-transfected HEK 293 cells...... 32

Table 4. Comparison of ligand affinities determined by competition binding (K]) or isolated tissue experiments (Ks) in pig coronary arteries...... 37

Table 5. Primers used to identify mRNA encoding CRLR and RAMPs by RT-PCR ...52

Table 6. Cell lines screened for CGRP2 receptors...... 56

Table 7. Affinity of CGRP and adrenomedullin receptor ligands in SV40LT-SMC and SK-N-MC membranes...... 63

Table 8 . Comparison of buffers known to influence the affinityof ha.-CGRP8 .3 7 ...... 95

Table 9. Comparison of human and porcine CGRP; receptor binding Characteristics...... 101

X) LIST OF FIGURES

Figure 1. Comparison of the amino acid sequences of human adrenomeduiiin, a-CGRP, p-CGRP, amyiin, and caicitonin...... 4

Figure 2. Primary structure of human a-CGRP...... 5

Figure 3. Tissue-dependent splicing leads to the synthesis of calcitonin or a-CGRP...... 7

Figure 4. Coexpression of calcitonin receptor like receptor (CRLR) with different receptor activity modifying proteins (RAMP) influences the ligand selectivity and glycosylation of CRLR...... 11

Figure 5. Primer sequences and their relative positions on the coding region of the CRLR mRNA...... 18

Figure 6 . Identification of CRLR and RAMP 1 mRNA in porcine coronary artery by RT-PCR...... 25

Figure 7. Comparison of porcine, human, and rat CRLR nucleotide sequences...... 27

Figure 8. Kinetics of ['^I]ha-CGRP binding to the cloned porcine CGRP) receptor...... 30

Figure 9. Mean competition binding curves for selected ligands at porcine CGRP receptors...... 31

Figure 10. Correlation plots of affinity values determined by radioligand binding (pK[) or isolated tissue experiments (pKs)...... 36

Figure 11. Mean concentration response curves for [Pro'^J ha-CGRP and ha-CGRP in HEK 293 cells stably expressing the porcine CGRP, receptor...... ,38

Figure 12. Mean concentration response curves for ha-CGRP and adrenomedullin in SV40LT-SMC cells...... 57

Figure 13. Concentration response curves for ha-CGRP and adrenomedullin in the presence and absence of 10 pM ha-CGRPg.37...... 59

Figure 14. Lack of CGRP receptors in SV40LT-SMC and DDT cells...... 60

Figure 15. Mean competition binding curves using SV40LT-SMC membranes labeled with ['^Ij-adrenomedullin...... 62

xii Figure 16. Identification of CRLR mRNA in SV40LT-SMC cells by RT-PCR...... 64

Figure 17. Absence of RAMP 1 mRNA in SV40LT-SMC cells...... 6 6

Figure 18. Identification of RAMP 2 and RAMP 2 B mRNA in SV40LT-SMC cells...... 67

Figure 19. Comparison of partial nucleotide sequences of RAMP 2 and RAMP 2 B...... 6 8

Figure 20. Identification of RAMP 3 mRNA in SV40LT-SMC cells...... 69

Figure 21. Identification of mRNA encoding CRLR, RAMP 1, and RAMP 2 in SK-N-MC cells...... 70

Figure 22. Mean competition curves for selected ligands in SK-N-MC membranes labeled with ['^I]-adrenomedullin...... 72

Figure 23. Mean competition curves for selected ligands in SK-N-MC membranes labeled with ['^I]ha-CGRP...... 74

Figure 24. Identification of specific ['*^I]-adrenomeduHin binding sites in membranes isolated from the rat vas deferens...... 75

Figure 25. Identification of CRLR and RAMP mRNA in the rat vas deferens...... 77

Figure 26. Distribution of affinity values reported for ha.-CGRP8.37...... 85

Figure 27. Proteolytic cleavage sites in ha-CGRP...... 8 8

Figure 28. Time-dependent decrease in the affinity of ha-CGRPg.37 in functional assays using isolated tissues...... 89

Figure 29. Comparison of human and porcine CRLR amino acid sequences...... 99

Figure 30. Impact of this work on the classification of receptors that are stimulated by CGRP...... 106

Figure 31. CGRP-induced cAMP synthesis in SK-N-MC and porcine CGRP] receptor-transfected HEK 293 cells...... 116

Figure 32. CGRP does not influence intracellular calcium concentrations in SK-N-MC or porcine CGRP] receptor-transfected HEK 293 cells...... 117

Figure 33. Partial mapping of the human CRLR gene...... 122

xiii ABBREVIATIONS cDNA: complementary deoxyribonucleic acid CGRP: calcitonin gene-related peptide CGRP} receptor: calcitonin gene-related peptide receptor subtype 1

CGRP2 receptor: calcitonin gene-related peptide receptor subtype 2 Cha: cyclohexylalanine CRLR: calcitonin receptor-like receptor ha-CGRP: human a-calcitonin gene-related peptide ['^l]ha-CGRP: 2-[*^I]Iodohistidyl'%a-CGRP

IC50: ligand concentration that inhibits 50 % of the radioligand bound at equilibrium Kassoc association rate constant of the radioligand Kg: equilibrium dissociation constant determined in functional assays using a single antagonist concentration

K

Kdissoc- dissociation rate constant of the radioligand K,[: equilibrium dissociation constant determined by competition binding Kobs: observed association rate constant of the radioligand PCR: polymerase chain reaction

PA2: -logio equilibrium dissociation constant determined from a Schild regression analysis of functional assays using multiple antagonist concentrations pKe: -logio l^B pKLi: -logio Ki RAMP: receptor activity modifying protein RCP: receptor component protein RT-PCR: reverse transcription-polymerase chain reaction

xrv CHAPTER 1

BACKGROUND AND OBJECTIVES

Intercellular Communication

Intercellular communication is a vital process for the survival of multicellular organisms. Cells communicate with each other through the release of peptides, proteins, amino acids, nucleotides, catecholamines, steroids, and fatty acid derivatives. These extracellular signaling molecules evoke responses in recipient cells by binding to specific receptors that are located within the cell or on the cell surface. This process is required for the regulation of metabolism, cell growth, immune responses, reproduction, regulation of vascular tone, sensory perception, motor activity, and many other biological functions.

Peptides and proteins are the most diverse signaling molecules m terms of their structure and the physiological processes that they regulate. These extracellular messengers vary in size, amino acid composition, and structure. Posttranslational modifications, such as glycosylation, sulphation, and the amidation of the carboxyl terminal of some peptides, further contribute to their diversity. In addition, some of these molecules are monomers while others exert their biological activity as dimers. Table 1 demonstrates the diverse size, biological effects, and target cells of selected peptides and proteins that are involved in intercellular communication.

Calcitonin Gene-Related Peptide

Calcitonin gene-related peptide (CGRP) is a member of a family of peptides that is involved in intercellular communication. This family of peptides includes a-CGRP,

1 Table 1. Examples of peptides and proteins that are involved in intercellular communication'. PEPTIDE OR #OF AN11NO CELL OF TARGET CELL BIOLOGICAL PROTEIN ACIDS ORIGIN OR TISSUE EFFECT

Thyrotropin 3 Neurons in Pituitary Thyrotropes Release of Thyroid Releasing Hormone Hypothalamus Stimulating Hormone

Oxytocin 9 Neurons in Uterine Myocytes, Uterine Contraction, Hypothalamus Myoepithelial Cells of Milk Ejection Mammary Gland

Catcitonin 32 Thyroid C cells Osteoclasts Bone Mineralization

Amy i in 37 P cells of Islets of Brain, p cells of Islets Inhibit Insulin Langerhans of Langerhans Secretion, Satiety

CGRP 37 Neurons in Brain, Myocytes, Endothelial Smooth Muscle Spinal Cord, and Cells, Neurons Relaxation, Periphery Nociception

AdrenomeduHin 52 Neurons in Adrenal Vascular Endothelial Vasodilation, Cell Medulla, Vascular Cells, Proliferation Endothelial Cells Smooth Muscle

Insulin 51 P cells in Islets of Hepatocytes, Skeletal Clearance of glucose Langerhans Myocytes, Adipocytes from blood

Giucagon 29 a cells in Islets of Hepatocytes, Gluconeogenesis, Langerhans Adipocytes Glyogenolysis, Fatty Acid Lipolysis

Proiactin 198 Pituitary Lactotropes Alveolar Cells of Milk Synthesis Mammary Gland

Substance P 11 Neurons in Brain Neurons, Vascular Nociception, and Spinal Cord Smooth Muscle Vasodilation

P-Endorphin 31 Neurons in Brain Neurons in Brain and Analgesia and Spinal Cord Spinal Cord

Parathyroid 84 Chief Cells of Osteoblasts, Renal Bone Resorption, Hormone Parathyroid Gland Proximal Tubule Calcium Retention Epithelial Cells

Thrombopoietin 332 Hepatocytes Megakaryocytes Platelet Production 'This is not an exhaustive list of peptide or protein extracellular signaling molecules, target tissues, or biological effects.

2 [3-CGRP, amylin, adrenomedullin, and calcitonin. Human a-CGRP (ha-CGRP) shares

92 %, 43 %, 19 %, and 17 % sequence similarity with the human form of each of these peptides, respectively (Figure 1). These peptides are related by several common structural features including a disulfide bond, an amidated carboxyl terminal, an a helix, and a putative tum structure (Figure 2). Furthermore, calcitonin and a-CGRP are products of the same gene.

CGRP was first identified following the discovery that the primary transcript of the calcitonin gene is spliced in a tissue-specific manner to form two distinct messenger ribonucleic acid (mRNA) transcripts (Amara et al., 1982). The first three exons of the primary transcript are constitutively spliced and are common to mRNAs that encode either preprocalcitonin or preproCGRP (Figure 3). Thyroid C cells splice exon III to exon IV, and a polyadenylation signal at the 3' end of exon IV is used to generate mature mRNA that encodes preprocalcitonin. In contrast, neurons splice exon III to exons V and

VI. A polyadenylation signal at the 3' end of exon VI results in the production of mature mRNA that encodes preproCGRP. The mechanisms that control tissue-dependent splicing of calcitonin/a-CGRP mRNA are not completely understood. However, several noncanonical splicing signals that surround exon IV are thought to be involved.

Furthermore, a 6 6 kDa RNA binding protein that is required for the inclusion of exon IV has been identified (Cote et al., 1992). Thus, splicing may be regulated by specific nucleotide sequences in the RNA and by the tissue-specific expression of specific RNA

binding proteins (Lou and Gagel, 1998).

The first 76 amino acids of preprocalcitonin and preproCGRP are identical and

include a 25 amino acid signal sequence that targets procalcitonin and proCGRP to the

3 Adrenomeduiiin YRQSMNNFQGLRS FGCRFG - TCTVQKLAHQIYQFTDKDKDNVAPRS K IS PQGY a-CGRP ACDTA-TCVTHRLAGLLSRSGGWKNNFVPTNVGSKA-F P-CGRP ACNTA- TCVTHRIAGLLSRSGGMVKSNFVPTNVGS KA- F Amyiin KCNTA - TCATQRLAiJFLVHS SNNFGAILS STNVGSNT - Y Caicitonin CGNLSTCMLGTYTQD FNK FHTF PQTAIGVGAP

Figure 1. Comparison of the amino acid sequences of human adrenomedullin, a-CGRP, P-CGRP, amylin, and calcitonin. The amino acid sequence of each peptide is represented by single letter abbreviations for each amino acid residue. Conserved residues are shaded. Dashes represent positions where there is no corresponding amino acid residue.

4 Figure 2. Primary structure of human a-CGRP. CGRP and other members of this peptide famiiy are characterized by a disulfide bond (boid solid tine), an amidated carboxyl terminal, an a-helix (shaded) and a putative turn that precedes a region where the secondary structure is less well characterized. The amino acid sequence of human a-CGRP is represented by circles containing the single letter abbreviation for each amino acid residue.

5 endoplasmic reticulum. Preprocalcitonin and preproCGRP are cleaved by endoproteases in the endoplasmic reticulum to form procalcitonin and proCGRP. Further processing by endoproteases, the formation of a disulfide bond, and amidation of the carboxyl terminal leads to the formation of mature calcitonin and CGRP (Figure 3).

CGRP is produced in neurons of the brain and spinal cord, motor neurons, and capsaicin-sensitive neurons that innervate many peripheral tissues (see reviews by

Wimalawansa, 1996 and Poyner, 1993). /u vivo and o? vitro studies have implicated

CGRP in a wide variety of biological processes including nociception (Li et al., 2001), regulation of vascular tone (Gangula et al., 2000), decreased intestinal motility (Fargeas et al., 1985), decreased gastric acid secretion (Lenz et al., 1985), decreased appetite

(Dennis et al., 1990), wound healing (Engin 1998), proliferation of vascular smooth muscle cells (Connat. et al., 2001), synthesis of acetylcholine receptors (New and Mudge,

1986), inhibition of insulin-stimulated glycogen synthesis (Leighton and Foot, 1995), and the maintenance of transmembrane ion gradients (Andersen and Clausen, 1993).

However, the physiological or potential pathological roles of CGRP are not well understood.

CGRP Receptors and Accessory Proteins

CGRP produces its effects by activating specific G-protein coupled receptors that are located on the cell surface. Based on functional studies using the isolated guinea pig

atrium and the isolated rat vas deferens, Dennis et al. (1990) proposed that CGRP-

induced responses are mediated by CGRP, receptors that are blocked by CGRP^.^ and by

CGRP2 receptors that are insensitive to CGRPg.^. Dennis et al. (1989) also reported that

6 CALCITONIN/CGRP GENE Signal 5' UTR Sequence N terminal Calcitonin CGRP 3' UTR ; : ; : M n H"n!

PRIMARY TRANSCRIPT 3*-l } ) rt ) Tii T"Tv vi I -3' mRNA Splicing

Thyroid C Cell Neuron 5'-l i ) M ) IH I tv polyA-3' 5'-l " i li I !!i M M : V C ]Po!yA-3'

PREPROPEPTIDE Preprocaicitonin PreproCGRP

Figure 3. Tissue-dependent splicing leads to the synthesis of calcitonin or a-CGRP. Exons I - VI (represented by rectangles) are transcribed in neurons and in thyroid C cells. The primary transcript is subsequently spliced and translated to produce two prepropeptides that are processed to form calcitonin and CGRP. DNA and RNA coding sequences and amino acid sequences that are common to both pathways are shown in gray. Sequences that are specific for calcitonin or CGRP are represented in white or black, respectively. Exons that are transcribed but not translated (UTR) are represented by speckled rectangles. Arrows represent proteolytic cleavage points.

7 CGRP receptor subtypes could be distinguished by [Cys(ACM)^ ]CGRP, a CGRP2 receptor-selective agonist. Functional studies in which CGRPg.37 is used to block CGRP- induced responses in isolated coronary arteries, isolated basilar arteries, SK-N-MC cells,

Col 29 cells, and other isolated tissues and cell lines have also identified variable affinities for CGRPg.37 (Waugh et al., 1999, Foulkes et al., 1991; Semark et al., 1992;

Cox, 1995). This has led to the widely accepted view that CGRP-induced responses are mediated by CGRP, receptors that are blocked with high affinity (Kg < 100 nM) by

CGRPg.37 and CGRP2 receptors that are blocked with low affinity (Kg > 100 nM) by

CGRPg.37 (Table 2).

The CGRP, receptor has been cloned and well characterized (Aiyar et al., 1996;

Elshourbagy et al., 1998). Aiyar et al. (1996) reported that the calcitonin receptor-like receptor (CRLR), previously identified as an orphan receptor (Fluhuman et al., 1995), was the CGRP, receptor. However, McLatchie et al. (1998) found that CRLR could only bind CGRP if CRLR was coexpressed with receptor activity modifying protein (RAMP)

1. This novel accessory protein contains 148 amino acid residues with an extracellular amino terminus, a single transmembrane spanning domain, and an intracellular carboxyl terminus. RAMP 1 is required for the terminal glycosylaticn of CRLR and for translocation of CRLR from the endoplasmic reticulum to the cell surface. RAMP 1 also colocalizes with CRLR at the cell membrane. However, it is not known whether this

protein directly interacts with CGRP. Two additional members of the RAMP family

have also been identified (McLatchie et al., 1998). RAMP 2 and RAMP 3 share 55 %

and 58 % amino acid sequence identity, respectively, to RAMP 1 and are also involved in

the glycosylation of CRLR and the translocation of CRLR to the cell surface. However,

8 Table 2. Classification criteria for CGRP receptor subtypes.

CGRP, CGRP, RECEPTOR RECEPTOR

^ of CGRPg.37 < 100 nM > 100 nM

[Cys(ACM)^]CORP Inactive Agonist

9 RAMP 2 and RAMP 3 cause CRLR to be core glycosylated rather than terminally glycosylated. This suggests that RAMPs may act as chaperones that lead CRLR through different posttranslational maturation processes (McLatchie et al., 1998). More importantly, coexpression of CRLR with RAMP 2 or RAMP 3 results in the formation of adrenomedullin receptors rather than CGRP receptors. Thus, the ligand selectivity of

CRLR is determined by the coexpression of different RAMPs (Figure 4). The RAMP family of proteins may also regulate the ligand selectivity of other G protein-coupled receptors. For example, Christopoulos et al. (1999) reported that coexpression of the human calcitonin receptor with RAMP 1 or RAMP 3 results in the formation of two different arnylin receptors. However, the influence of RAMPs on receptors for ligands other than CGRP, calcitonin, adrenomedullin, and arnylin has not been investigated.

Evans et al. (2001) reported that an additional accessory protein, receptor component protein (RCP), is also required for the formation of mature CGRP, receptors and adrenomedullin receptors. RCP contains 148 amino acid residues and is peripherally associated with the cytoplasmic surface of the cell membrane (Luebke et al., 1996; Evans et al., 2001). RCP is not required for agonists to bind CGRP, and adrenomedullin receptors. However, RCP is required for these receptors to mediate an agonist induced increase in intracellular 3', 5' cyclic adenosine monophosphate (cAMP), the primary intracellular messenger of CGRP and adrenomedullin receptors (Evans et al., 2001).

The mechanism by which RCP interacts with G proteins, adenylate cyclase, or other proteins that are involved in this biochemical pathway is unknown. However, the discovery of RCP and the RAMP family of proteins has opened new paradigms for the

10 A CGRP, RECEPTOR

^ Terminai glycosylation ^ Core glycosylation B ADRENOMEDULLIN RECEPTOR

C CoreGiycosyiation Examples of Terminal Glycosylation }y y ^ v v v v { $ $ # # ^ V

A Glucose # Mannose n N-Acetylglucosamine Galactose A Sialic Acid

Figure 4. Coexpression of calcitonin receptor-like receptor (CRLR) with different receptor activity modifying proteins (RAMP) influences the ligand selectivity of CRLR. CRLR forms the CGRP, receptor when coexpressed with receptor activity modifying protein (RAMP) 1 (panel A) and forms an adrenomedullin receptor when coexpressed with RAMP 2 or RAMP 3 (panel B) (McLatchie et al., 1998). Receptor component protein (RCP) is required for the activation of adenylate cyclase by CGRP and adrenomedullin receptors (Evans et al., 2001). RAMPs also cause CRLR to undergo core or terminal glycosylation (panel C). CRLR is core glycosylated by the transfer of a core carbohydrate (containing N-acetylglucosamine, mannose, and glucose) to asparagine residues. Some sugars are removed from this core group and replaced with galactose, sialic acid, N-acetylgalactosamine, or other sugars to produce a terminally glycosylated protein. (Panel C from K. Drickamer and M.E. Taylor (1998) Evolving views of protein glycosylation. Trends in Th'ocAenu'cu/ Sciences 23:321-324).

11 ligand selectivity, signal transduction pathways, and potential regulatory mechanisms of

G-protein coupled receptors.

In contrast to the CG R P, receptor, the CGRP2 receptor has not been cloned or well characterized. CGRP2 receptors are distinguished from C G R P , receptors by their low affinity (Kg > 100 nM) for CGRPg.37 in functional assays (Dennis et al., 1990;

Mimeault et al., 1991). In addition, some investigators have reported that radioligand

binding studies with CGRP 3 . 3 7 do not support the existence of multiple C G R P receptor subtypes (Dennis et al., 1990; Mimeault et al., 1991). Consequently, the ability of

CGRPg.37 to discriminate between CGRP, and CGRP2 receptor subtypes in functional

assays but not in radioligand binding assays has been an enigma.

Objectives

The overall goals of my research were to further characterize the CGRP2 receptor

and to determine why CGRPg.37 discriminates between CGRP receptor subtypes in

functional assays but not in radioligand binding assays. This was accomplished through

the following objectives:

1. ) Compare the CGRP2 receptor that has been previously identified in porcine

coronary arteries to the porcine CGRP, receptor that has been cloned and

expressed in HEK 293 cells.

2. ) Test the hypothesis that the putative CGRP2 receptor is an adrenomedullin

receptor.

3. ) Examine the hypothesis that the variable affinity of CGRPg.37 in isolated tissues is

influenced by tissue-dependent factors including the proteolytic degradation of

CGRPg.37 and the release of endogenous C G R P from neuronal storage sites.

12 CHAPTER 2

FUNCTIONAL CGRP: RECEPTORS IN PORCINE CORONARY ARTERIES

ARE IDENTIFIED AS CGRP, RECEPTORS BY RADIOLIGAND BINDING AND

REVERSE TRANSCRIPTION-POLYMERASE CHAIN REACTION

Abstract

Caicitonin gene-reiated peptide (CGRP) receptors are classified into CGRP, and

CGRP2 receptor subtypes based on their affinity for the antagonist, ha-CGRPg.37. ha.-CGRPg.37 antagonizes CGRP, receptor-mediated responses with high affinity

(Kg < 100 nM) and antagonizes CGRP2 receptor-mediated responses with low affinity

(K.B > 100 nM). CGRP2 receptors have been previously reported to mediate relaxation of large porcine coronary arteries because this action is antagonized with low affinity by ha-CGRPg.37. In the present study, I used RT-PCR, radioligand binding, and values from previously reported isolated tissue experiments to compare the CGRP receptor in porcine coronary arteries to the porcine CGRP, receptor stably expressed in HEK 293 cells. I identified calcitonin receptor like receptor (CRLR) and receptor activity modifying protein 1 (RAMP 1) mRNA in coronary arteries. 1 also found that the ligand binding characteristics of the CGRP receptor in coronary arteries and the cloned CGRP, receptor were similar. K, values for ha-CGRPg^ were 6 .6 nM and 5.7 nM in porcine coronary arteries and the cloned CGRP, receptor, respectively. The affinity (Ke) of ha-CGRPg.37 and 5 other antagonists was 22- to 707-fold lower in functional experiments measuring relaxation of coronary arteries than in radioligand binding experiments. Despite this difference in absolute affinity values, there was a high correlation of the rank order of affinity for the antagonists determined by the two methods. These data support the

13 conclusion that CGRP-induced relaxation of porcine coronary arteries is mediated by the

CGRPi receptor.

Introduction

Dennis et al. (1990) were first to propose that two different CGRP receptor subtypes mediate CGRP-induced responses. This conclusion was based on the observation that human a-CGRPs.37 (ha-CGRPg.37), a CGRP receptor antagonist, blocks

CGRP-induced tachycardia in the isolated guinea pig atrium with greater affinity than it blocks CGRP-induced relaxation of the field stimulated rat vas deferens. Studies using isolated coronary arteries, isolated basilar arteries, SK-N-MC cells, Col 29 cells, and other isolated tissues and cultured cell lines have confirmed that CGRP-induced responses are blocked by ha-CGRPg.37 with different affinities (Waugh et al., 1999;

Foulkes et al., 1991; Semark et al., 1992; Cox, 1995). This has led to the widely accepted view that CGRP-induced responses are mediated by CGRP] receptors that are blocked with high affinity (Ks < 100 nM) by ha-CGRPg^? and CGRP2 receptors that are blocked with low affinity (Ks > 100 nM) by ha-CGRPg^? (Dennis et al. 1990; Mimeault et al.

1991; Wisskirchen et al., 1998).

The CGRP] receptor is formed by the coexpression of the calcitonin receptor like receptor (CRLR) and receptor activity modifying protein 1 (RAMP 1) (McLatchie et al.,

1998). In nearly all tissues and cells that contain CGRP] receptors, this receptor is coupled to an increase in intracellular 3',5' cyclic adenosine monophosphate (cAMP).

Activation of the CGRP] receptor is also reported to cause intracellular calcium

14 mobilization (Aiyar et al., 1999) and the activation of extracellular regulated kinase, P-38 mitogen activated protein kinase, and Jun kinase in some systems (Disa et al., 2000;

Parameswaran et al., 2000). These signal transduction pathways have been studied in

CGEJ*) receptor-transfected HEK 293 cells and in cells that endogenously express the

CGRP} receptor (Aiyar et al., 1999; Disa et al., 2000; Parameswaran et al., 2000;

Rorabaugh et al., 2001).

In contrast to the CGRP] receptor, the putative CGRP2 receptor has not been cloned or well characterized. The CGRP2 receptor has only been identified by its low affinity (Ks > 1 pM) for ha-CGRPg.37 in functional studies where a response to CGRP is measured (Dennis et al., 1990). Foulkes et al. (1991) and Waugh et al. (1999) reported that CGRP-induced dilation of porcine coronary arteries is blocked with low affinity (Ks

= 5 pM) by ha-CGRPs-37. In addition, the putative CGRP2 receptor-selective agonrst,

[Cys(ACM)^]ha-CGRP, induces dilation of these arteries (Waugh et al., 1999). These studies establish that porcine coronary arteries have the prototypical characteristics of a tissue containing the putative CGRP2 receptor.

In the present investigation, I used RT-PCR, radioligand binding, and data from previously reported isolated tissue experiments to compare the CGRP2 receptor that is expressed in porcine coronary arteries to the porcine CGRP] receptor that has been previously cloned and expressed in HEK 293 cells. Ligand affinities determined in

isolated tissue experiments (Ks) and by radioligand binding (K[) were also compared. I

found that the CGRP2 receptors that have been previously reported to mediate relaxation

of isolated porcine coronary arteries are identified as CGRP: receptors by radioligand

15 binding. The presence of CRLR mRNA and RAMP 1 mRNA in this tissue further support the conclusion that porcine coronary arteries express the CGRPi receptor.

Materials and Methods

Chemicais and Reagents. Tug DNA polymerase, 10 X PCR buffer (200 mM

Tris-HCl, pH 8.4, $00 mM RC1), amplification grade DNAse 1, 100 base pair DNA ladder, TRlzol, Dulbecco's modified Eagle medium (DMEM), fetal bovine scrum, and antibiotic/antimycotic (containing 10,000 units/ml penicillin G, 10,000 pg/ml streptomycin sulfate, and 25 pg/ml amphotericin B) were purchased from Gibco (Grand

Island, NY). Moloney murine leukemia virus reverse transcriptase was purchased from

Perkin Elmer (Foster City, CA). The pCRH cloning vector was purchased from

Invitrogen (Carlsbad, CA). The radioassay kit for measuring intracellular cAMP was purchased from Diagnostic Products Corporation (Los Angeles, CA) and ['^l]ha-CGRP was purchased from Amersham Pharmacia Biotech (Piscataway, NJ).

Na2Ca-ethylenediaminetetracetic acid, Tris(hydroxymethyl)aminomethane, Sigmacote, and other chemicals were obtained from Sigma (St. Louis, MO). HER 293 cells stably expressing the porcine CGRP] receptor were a generous gift from Dr. Allan R. Shatzman of SmithRline Beecham Pharmaceuticals (Ring of Prussia, PA).

Peptide Synthesis. Adrenomedullin, calcitonin, hot.-CGRPg.37, ha- CGRP, and

[Cys(ACM)^]ha-CGRP were purchased from Peninsula Laboratories (San Carlos, CA).

All other peptides were synthesized by solid phase methods and purified by reversed phase high performance liquid chromatography as previously described (Li et al., 1997;

Smith and Hanley, 1997; Saha et al., 1998; Rist et al., 1999). The structure of these

16 peptides was verified by amino acid analysis and electrospray ionization-mass spectrometry.

RNA Isolation. A fresh porcine heart was obtained from a local slaughterhouse and transported to the laboratory in ice-cold phosphate buffered saline (137 mM NaCl,

2.7 mM KC1, 4.3 mM NazHPC^, 1.4 mM KH2PO4, pH = 7.3). The left circumflex coronary artery was isolated, cleaned of fat and connective tissue with the aide of a dissecting microscope, and the endothelium was removed by gentle scraping with a number 22 scalpel blade. The artery was wrapped in aluminum foil, frozen at -70°C and pulverized with a hammer. TRIzol reagent was used to isolate total RNA from approximately 1 0 0 mg of pulverized artery according to the manufacturer's protocol.

The RNA was dissolved in 20 ul of RNAse-free water containing DNAse 1 buffer and 5 units of amplification grade DNAse 1. The DNAse I was removed by adding an equal volume of TRIzol and repeating the RNA isolation procedure. The RNA was dissolved in RNAse-free water and stored at -70°C.

RT-PCR. RT-PCR was used to identify CRLR mRNA in porcine coronary arteries with the primers shown in Figure 5. These primers were designed based on the porcine CRLR complementary DNA (cDNA) sequence provided by Dr. Allan R.

Shatzman of SmithKline Beecham Pharmaceuticals (King of Prussia, PA). CRLR mRNA was initially detected in porcine coronary artery using primers 2 and 4 (Figure 5). cDNA was synthesized by reverse transcription in a 10 ul reaction volume containing 1

X PCR buffer, 3 pg RNA, 5 mM MgC^, 1 mM dNTP mix, 25 pmoles antisense primer, and 25 units Moloney murine leukemia virus reverse transcriptase. The reaction was incubated in a Perkin Elmer 2400 thermocycler at 42°C for 50 minutes followed by a

17 CN m xt* Primer 5 Primer

m r * — n 223 base pairs

Primer 1 5' -ATG GAG AAA AAG TAT ATC GTG - 3' Primer 2 5' -CTC CTC TAC ATT ATC CAT GG -3' Primer 3 5' -GGT AAT GAA GCG TAA CTT A - 3' Primer 4 5' -CCT CCT CTG GAA TCT TTC C -3' Primer 5 5' -TCA RTC AT A TAA YTT TTC TGG TTT -3'

Figure 5. Primer sequences and their relative positions on the coding region of the CRLR mRNA. The soiid line represents the porcine CRLR mRNA coding region and the horizontal open bar represents the position of the 223 base pair RT-PCR product shown in Figure 6 A. Sense primers are designated by vertical open boxes, and antisense primers are designated by vertical solid boxes. Roman numerals indicate regions of the mRNA that code for the transmembrane regions of the CRLR. A = adenosine, T = thymine, G = guanine, C - cytosine, R = adenosine/guanine degeneracy, and Y = thymine/cytosine degeneracy.

18 five minute incubation at 99°C. PCR was conducted in a 100 pi reaction volume containing 1 X PCR buffer, 3 mM MgC^, 0.2 mM dNTP mix, 50 pmoles sense primer,

50 pmoles antisense primer, 10 pi cDNA, and 2.5 units DNA polymerase. PCR conditions included an initial cDNA denaturation step at 95°C for 5 minutes followed by

30 cycles (95°C denaturation for 30 sec, 55°C annealing for 30 sec, and 72°C extension for 30 sec) of PCR and a final extension period of 7 minutes at 72°C. Primers 2 and 5

(Figure 5) were used to amplify the 3' end of the CRLR mRNA coding region using the

PCR conditions described above. Primers 1 and 3 (Figure 5) were used to amplify the 5'

end of the CRLR mRNA coding region. PCR conditions for this pair of primers were the

same as those described above except that the annealing temperature was changed to

50°C.

Sense (5'-GAC CAT CAG GAG CTA TAA AGA CC-3') and antisense (5'-TGC

CAG ACC ACC AGT GCG GTC-3') primers were designed based upon the porcine

RAMP 1 cDNA sequence (Genbank accession # AF312385). These primers were used to

detect RAMP 1 mRNA in coronary arteries using the method described above. However,

the annealing temperature was adjusted to 54°C, and 40 cycles of PCR were used. The

products of all RT-PCR reactions were visualized on ethidium bromide-stained 1.5 %

agarose gels and subcloned into the pCRII vector. Both DNA strands of each RT-PCR

product were sequenced using an Applied Biosystems 373 DNA sequencer. DNA

sequences were analyzed using the Wisconsin Package Version 10.1 (Genetics Computer

Group, Madison, WI) software.

19 Cell Culture. HEK 293 ceils stably expressing the porcine CGRP) receptor were

grown in T-175 culture flasks in Dulbecco's modified Eagle medium (DMEM) that was

supplemented with fetal bovine serum (10 %), penicillin G (100 units/ml), streptomycin

(100 pg/ml), and amphotericin B (0.25 pg/ml). The flasks were placed m a humidified

incubator in an atmosphere of 5 % CO2 /9 5 % air and maintained at 37°C. The cells were grown to confluence and then harvested for membrane preparations as described below.

Membrane Preparations. Culture media was removed from confluent cells, and the cells were rinsed three times with 25 ml of ice-cold phosphate buffer ed saline. Cells were dislodged from the flask with a cell scraper in the presence of 10 ml ice-cold phosphate buffered saline and centrifuged at 4°C for 5 minutes at 1000 x g. The pellet was suspended in 25 ml buffer A (50 mM Tris-HCl, 5 mM Na2Ca-

ethylenediaminetetracetic acid, pH 7.4) by vortexing and then homogenized with a glass- teflon homogenizer. The homogenate was centrifuged at 100,000 x g for 30 minutes in a

Beckman L5-50 ultracentrifuge. The pellet was washed twice by homogenization in 25

ml buffer B (50 mM Tris-HCl, 100 mM NaCl, 5 mM MgClz, pH 7.4) followed by

centrifugation at 4°C for 30 minutes at 100,000 x g. The supernatant was removed and

the dry pellet was stored for up to 1 month at -70°C. Protein content of the final pellet

was determined by the method of Lowry (Lowry et al., 1951).

Fresh pig hearts were obtained from a local slaughterhouse and transported to the

laboratory in ice-cold phosphate buffered saline. The left circumflex, right circumflex,

and anterior descending coronary arteries were removed and cleaned of fat and

connective tissue with the aid of a dissecting microscope. The outside diameter of all

coronary arteries used in this investigation was > 1 mm. The arteries were cut open, and

20 the endothelium was removed by gentle scraping with a number 22 scalpel blade. The arteries were cut into small pieces with scissors and homogenized in 23 ml buffer A with an Ultra-Turrax T25 tissue homogenizer for 3 minutes at 24,000 rpm. The homogenate was centrifuged at 4°C for 10 minutes at 1,000 x g to remove particulate debris. The supernatant was centrifuged at 4°C for 30 minutes at 100,000 x g, and the resulting pellet was washed twice in buffer B as described above for membranes from HEK 293 cells.

Protein content of the final pellet was determined by the method of Lowry (Lowry et al.,

1951). Membrane pellets were stored for up to 1 week at -70°C.

i'^llha-CGRP Binding Kinetics. The association and dissociation rates of

[*"^I]ha-CGRP binding to CGRP receptors were determined in membranes prepared from either porcine coronary arteries or from HEK 293 cells expressing the porcine CGRP] receptor. Frozen membrane pellets were rehomogenized in ice-cold binding buffer

(50 mM Tris-HCl, 5 mM MgCl2, 100 mM NaCl, 0.2 % bovine serum albumin, 0.1 % bacitracin, pH 7.4) to a concentration of 50-100 pg membrane protein/150 pi. Membrane protein homogenate (150 pi) was added to 13 x 100 mm glass test tubes pretreated with

Sigmacote. 50 pi of ice-cold binding buffer were added to each test tube followed by

50 pi of 200 pM ['*^I]ha-CGRP. To measure the association rate, tubes were quickly vortexed and incubated at 37°C for various times. Bound and free ['* I]ha-CGRP were separated by vacuum filtration by pouring the tube contents through Whatman

(Maidstone, England) GF/B glass microfiber Liters which were presoaked in 0.2 % polyethyleneimine for 30 minutes. Each tube was rinsed 3 times with 5 ml of buffer B, and this buffer was also poured through the filter. To measure the dissociation rate, tubes were vortexed and incubated at 37°C for 30 minutes. 50 pi of 5 pM non-radiolabeled

21 ha-CGRP (final concentration = 0.83 pM) was added to each tube, and the tubes were incubated at 37°C for various times. Bound and free ['^I]ha-CGRP were separated by vacuum filtration as described above. The filters were carefully transferred into 12 x 75 mm polyethylene tubes and bound ['^I]ha-CGRP was measured in a Wallac

Gammamaster 1277 (Gaithersburg, MD) gamma counter.

Competition Binding Assay. The radioligand binding assay used in this study has been previously described in detail (Abel et at., 1997). Frozen membrane pellets were rehomogenized in ice-cold binding buffer to a concentration of 50-100 pg membrane protein/150 pi. Membrane protein (150 pi) homogenate was added to 13 x

100 mm glass test tubes preheated with Sigmacote. The tubes were incubated in a 37°C

shaking water bath for 50 minutes in the presence of 40 pM ['"*i]ha-CGRP and various

concentrations of nonlabeled ligands. The total incubation volume was 250 pi.

Nonspecific binding was determined using lpM ha-CGRP. Whatman GF/B glass

microfiber filters were soaked in 0.2 % polyethyleneimine for 30 minutes prior to their

use. Bound and free ['^i]ha-CGRP were separated by trapping the membranes on the

filters and washing with 15 ml buffer B using a Brandel MB-48R cell harvester

(Gaithersburg, MD). Bound ['^I]ha-CGRP was measured as described above.

Measurement of Intracellular cAMP. HEK 293 cells stably expressing the

porcine CGRPt receptor were grown in 12 well plates under the conditions described

above. Culture medium was removed from the confluent cells, and the cells were gently

rinsed with 2 ml of prewarmed (37 °C) HEPES-buffered Krebs solution that contained 20

mM HEPES, 4 mM NaHCCb, 11 mM dextrose, 1.2 mM NaHzPC^, 5.5 mM KC1, 2.5 mM

CaCl2, 1.2 mM MgCl2, and 0.5 mM isobutylmethylxanthine. This solution was also used

22 to dissolve ha-CGRP and [Pro^]ha-CGRP. Cells were incubated with HEPES-buffered

Krebs solution for 10 minutes at 37°C before various concentrations of either ha-CGRP or [Pro'^jha-CGRP were added to each well. The solution was removed after 30 minutes, and the cells were lysed by adding 100 pi of 90 % ethanol / 10 % water. The ethanol was allowed to evaporate in a 37°C incubator. The cAMP present in the dried lysate was measured using a radioimmunoassay provided by Diagnostic Products

Corporation (Los Angeles, CA) according to the manufacturer's instructions.

Data Analysis. Association and dissociation rates of ['^I]ha-CGRP binding to

CGRP receptors were determined by nonlinear regression analysis using the equations for exponential association and exponential decay. These calculations were performed using

GraphPad Prism (SanDiego, CA).

Three to Eve competition binding curves were performed in duplicate for each ligand. Specific binding was determined by subtracting nonspecific binding (defined using 1 pM ha-CGRP) from total binding, and the IC50 for each competition curve was determined by nonlinear regression analysis using GraphPad Prism. The data were Et to both a one site and a two site binding model and the best Et model was determined using an F test. Hill slopes were calculated from nonlinear regression analysis using a sigmoid curve Et model. In kinetic studies, the K

prepared Eom porcine coronary arteries. Therefore, these values were used to convert

IC50 values to K[ values by the Cheng-Prusoff equation. Mean pK] values for each ligand

were compared using a two-tailed student's t-test to determine if the pK] values in

23 HEK 293 cells were significantly different (p < 0.05) from the pKi values in porcine coronary arteries.

Correlation plots were used to compare antagonist affinities (pKi) determined by competition binding experiments to antagonist affinities determined by their ability to inhibit CGRP-induced relaxation of isolated coronary arteries (pKs). GraphPad Prism was used to perform linear regression analysis, determine the confidence interval of the correlation, and to calculate the slope and correlation coefficient of these data.

The ability of [Pro^]ha-CGRP and ha-CGRP to stimulate cAMP synthesis was determined in HEK 293 cells stably transfected with the porcine CGRP] receptor.

GraphPad Prism was used to analyze these data by nonlinear regression and to determine the EC50 value of these ligands.

Results

Identification of CRLR and RAMP 1 mRNA in Porcine Coronary Arteries.

Primers that spanned a 223 nucleotide segment of the CRLR mRNA (primers 2 and 4 in

Figure 5) were initially used to search for CRLR mRNA in porcine coronary artery.

Porcine lung, the tissue from which this cDNA was originally cloned (Elshourbagy et al.,

1998), was used as a positive control. A 223 base pair RT-PCR product was identified m

both the porcine coronary artery and lung (Figure 6 A), and DNA sequence analysis

demonstrated that this product encoded a portion of the porcine CRLR. No RT-PCR

product was observed in control reactions that lacked reverse transcriptase. Thus, the

RNA was not contaminated with DNA. To confirm that the entire coding region of the

CRLR mRNA was present in the coronary artery, I used primers 1 and 3, and primers 2

24 coronary lung artery E-

+ < +

B

Figure 6. Identification of CRLR and RAMP 1 mRNA in porcine coronary artery by RT-PCR. Porcine lung was used as a positive control for the detection of CRLR mRNA. The inclusion/exclusion of reverse transcriptase in the reverse transcription reaction is indicated by +/-RT. Marker bands are 200 and 300 base pair size markers, and arrows indicate the 223 and 236 base pair RT-PCR products obtained using primers specific for CRLR (panel A) and RAMP 1 (panel B), respectively.

25 and 5 (Figure 5). RT-PCR products from each primer pair were subcloned and sequenced. The amino acid sequence encoded by this mRNA is identical to that previously reported by Elshourbagy et al. (1998). The nucleotide coding sequence is

1389 nucleotides long and shares 92 % and 85 % sequence identity with its human and rat orthologs, respectively (Figure 7).

RAMP 1 is an accessory protein that is reportedly required for intracellular trafficking and maturation of the CRLR into the CGRP] receptor (McLatchie et al. 1998).

RAMP 1 mRNA has been previously identified in several human tissues and cell lines, including HEK 293 cells (McLathie et al., 1998). However, RAMP 1 has not been previously identified in coronary arteries. Since antibodies for this protein were unavailable, I used RT-PCR to identify RAMP 1 mRNA in porcine coronary arteries with the primers described above (Figure 6 B). Porcine RAMP 1 shares 78 %, 78 %, and

82 % nucleotide sequence identity with its rat, mouse, and human orthologs, respectively

(Genbank accession numbers AJ001014, AF146522, and AF181550).

Kinetics of r'^Ilha-CGRP Binding to CGRP Receptors. The K^ of

['^I]ha-CGRP was determined in membranes prepared from porcine coronary arteries and from porcine CGRP] receptor-transfected HEK 293 cells by independently measuring the association and dissociation rates of ['"^Ijha-CGRP binding to CGRP receptors. The Kd of this radioligand was used to calculate the affinity (K j of other ligands that were examined in competition binding assays. The procedure for calculating

Kd values from kinetic experiments has been described in detail by Limbird (1996).

['^I]ha-CGRP dissociated from membranes prepared from porcine CGRP] receptor- transfected HEK 293 cells and from porcine coronary arteries with dissociation rate

26 HUMAN ATGGAGAAAA AGTGTACCCT GTATTTTCTG GTTCTCTTGC CTTTTTTTAT GATTCTTGTT PIG ATGGAGAAAA AGTATATCCT GTATTTTCTG TTTCTCTTAC CTTTTTTTAT GATTCTTGTC 60 RAT ATGGATAAAA AGTGTACACT ATGTTTTCTG TTTCTCTTGC TTCTTAATAT GGCTCTCATC

ACAGCAGA.. .ATTAGAAGA GAGTCCTGAG GACTCAATTC AGTTGGGAGT TACTAGAAAT 61 ATAGCAGAAA CAGAAGAAGA AAACCCTGAT GACTTAATTC AGTTGACTGT TACTAGAAAT 120 GCAGCAGAGT CGGAAGAAG...... GCGCG AACCAAACAG ACTTGGGAGT CACTAGGAAC

AAAATCATGA CAGCTCAATA TGAATGTTAC CAAAAGATTA TGCAAGACCC CATTCAACAA 121 AAAATCATGA CAGCTCAATA TGAATGTTAT CAAAAAATTA TGCAAGACCC TATTCAACAA 180 AAGATCATGA CGGCTCAGTA TGAATGTTAC CAAAAGATCA TGCAGGATCC CATTCAACAA

GCAGAAGGCG TTTACTGCAA CAGAACCTGG GATGGATGGC TCTGCTGGAA CGATGTTGCA 181 ACAGAAGGCA TTTACTGCAA CAGAACCTGG GACGGATGGC TATGCTGGAA TGATGTTGCT 240 GGAGAAGGCC TTTACTGCAA CAGAACCTGG GACGGATGGC TATGCTGGAA TGACGTTGCA

GCAGGAACTG AATCAATGCA GCTCTGCCCT GATTACTTTC AGGACTTTGA TCCATCAGAA 241 GCAGGAACAG AATCAATGCA GCATTGCCCT GATTACTTTC AGGACTTTGA TCCTTCAGAA 300 GCAGGAACCG AGTCAATGCA GTACTGCCCT GATTACTTTC AAGATTTTGA TCCTTCAGAG

AAAGTTACAA AGATCTGTGA CCAAGATGGA AACTGGTTTA GACATCCAGC AAGCAACAGA 301 AAAGTTACGA AAATCTGTGA CCAAGATGGA AACTGGTTTA GACATCCAGA AAGCAACAGA 360 AAGGTTACAA AGATCTGTGA CCAAGATGGA AACTGGTTCA GACATCCAGA TAGTAACAGG

ACATGGACAA ATTATACCCA GTGTAATGTT AACACCCACG AGAAAGTGAA GACTGCACTA 361 ACATGGACAA ATTATACCCA GTGTAATATT AACACACATG AGAAAGTACA GACTGCACTG 420 ACATGGACAA ACTACACCTT GTGTAACAAC AGCACGCATG AGAAAGTGAA GACAGCACTG

AATTTGTTTT ACCTGACCAT AATTGGACAC GGATTGTCTA TTGCATCACT GCTTATCTCG 421 AATTTATTTT ACCTGACTAT AATTGGACAT GGATTATCTA TTGCATCATT GCTTATCTCA 480 AATTTGTTCT ACCTAACTAT AATTGGACAT GGATTATCTA TTGCCTCTCT GATCATCTCA

CTTGGCATAT TCTTTTATTT CAAGAGCCTA AGTTGCCAAA GGATTACCTT ACACAAAAAT 481 CTTGGTATTT TCTTTTATTT CAAGAGCCTA AGTTGCCAAA GGATTACCTT GCACAAAAAT 540 CTCATCATAT TTTTTTATTT CAAGAGCCTA AGTTGCCAAC GGATTACATT GCATAAAAAC

CTGTTCTTCT CATTTGTTTG TAACTCTGTT GTAACAATCA TTCACCTCAC TGCAGTGGCC 541 CTGTTCTTCT CTTTTGTTTG TAACTCCATT GTAACAATCA TTCATCTCAC TGCAGTGGCC 600 CTGTTCTTTT CATTTGTTTG TAATTCGATT GTGACAATCA TTCACCTCAC GGCAGTGGCC

AACAACCAGG CCTTAGTAGC CACAAATCCT GTTAGTTGCA AAGTGTCCCA GTTCATTCAT 601 AACAACCAGG CCTTAGTGGC CACAAATCCT GTTAGTTGTA AAGTGTTCCA GTTCATTCAT 660 AATAACCAGG CCTTAGTGGC CACAAATCCT GTGAGCTGCA AGGTGTCCCA GTTCATTCAT

CTTTACCTGA TGGGCTGTAA TTACTTTTGG ATGCTCTGTG AAGGCATTTA CCTACACACA 661 CTTTACCTGA TGGGATGTAA CTACTTTTGG ATGCTTTGTG AAGGCATTTA CCTACACACA 720 CTTTACCTGA TGGGCTGTAA CTACTTTTGG ATGCTCTGTG AAGGCATTTA CCTGCACACA

CTCATTGTGG TGGCCGTGTT TGCAGAGAAG CAACATTTAA TGTGGTATTA TTTTCTTGGC 721 CTTATTGTGG TGGCCGTGTT TGCAGAGAAG CAACACTTGA TGTGGTATTA TTTTCTTGGC 770 CTCATTGTGG TGGCTGTGTT TGCAGAGAAG CAGCACTTGA TGTGGTATTA TTTTCTTGGC

TGGGGATTTC CACTGATTCC TGCTTGTATA CATGCCATTG CTAGAAGCTT ATATTACAAT 771 TGGGGATTTC CACTGATTCC TGCTTGTATA CATGCTGTTG CCAGACGCTT GTACTACAAT 840 TGGGGGTTTC CTCTGCTTCC TGCCTGCATC CATGCCATCG CCAGAAGCTT GTATTACAAT

Figure 7. Comparison of porcine, human, and rat CRLR nucleotide sequences. Genbank accession numbers AF419317, L76380, and X70658 were used for this comparison. Conserved nucleotides are shaded.

27 Figure 7. (continued)

GACAATTGCT GGATCAGTTC TGATACCCAT CTCCTCTACA TTATCCATGG CCCAATTTGT 84 1 GACAACTGCT GGATCAGTTC TGATACCCAT CTCCTCTACA TTATCCATGG CCCGATTTGT 900 GACAACTGCT GGATCAGCTC AGACACTCAT CTCCTCTACA TCATCCATGG TCCCATTTGT

GCTGCTTTAC TGGTGAATCT TTTTTTCTTG TTAAATATTG TACGCGTTCT CATCACCAAG 901 GCTGCTTTAT TGGTGAATCT TTTTTTCCTT TTAAATATTG TACGAGTTCT TATC'ACCAAG 960 GCTGCTTTAC TGGTAAATCT CTTTTTCCTA TTAAATATTG TACGTGTTCT CATCACCAAG

TTAAAAGTTA CACACCAAGC GGAATCCAAT CTGTACATGA AAGCTGTGAG AGCTACTCTT 961 TTAAAAGTTA CTCACCAAGC AGAGTCCAAT CTCTACATGA AAGCTGTGAG AGCTACACTT 1020 TTGAAAGTTA CACACCAAGC AGAATCCAAT CTCTACATGA AAGCTGTAAG AGCCACTCTC

ATCTTGGTGC CATTGCTTGG CATTGAATTT GTGCTGATTC CATGGCGACC TGAAGGAAAG 10 2 1 ATCTTGGTGC CATTACTTGG CATTGAATTT GTGCTGATTC CATGGCGACC TGAAGGAAAG 1080 ATCTTGGTAC CACTACTTGG CATTGAATTT GTGCTTTTTC CATGGCGGCC TGAAGGAAAG

ATTGCAGAGG AGGTATATGA CTACATCATG CACATCCTTA TGCACTTCCA GGGTCTTTTG 1 0 8 1 ATTGCAGAGG AGGTATACGA TTACATCATG CACATCCTTG TGCACTATCA GGGTCTGTTG 1140 GTTGCTGAGG AGGTGTATGA CTATGTCATG CACATTCTCA TGCACTATCA GGGTCTTTTG

GTCTCTACCA TTTTC'TGCTT CTTTAATGGA GAGGTTCAAG CAATTCTGAG AAGAAACTGG 11 4 1 GTATCTACAA TTTATTGCTT CTTTAATGGA GAGGTTCAAG CAATTCTGAG AAGAAACTGG 1200 GTGTCTACAA TTTTCTGCTT CTTTAACGGA GAGGTTCAAG CAATTCTGAG AAGAAATTGG

AATCAATACA AAATCCAATT TGGAAACAGC TTTTCCAACT CAGAAGCTCT TCGTAGTGCG 1 2 0 1 AATCAATACA AAATCCAGTT TGGAAACAGC TTTTCCCACT CAGATGCTCT CCGAAGTGCA 1260 AACCAGTATA AAATCCAATT TGGCAATGGC TTTTCCCACT CTGATGCTCT CCGCAGCGCA

TCTTACACAG TGTCAACAAT CAGTGATGGT CCAGGTTATA GTCATGACTG TCCTAGTGAA 12 6 1 TCTTACACTG TGTCAACCAT CAGTGATGGT GCAGGTTACA GTCATGACTA TCCAAGCGAA 1320 TCCTATACGG TGTCAACAAT CA.GCGATGTG CAGGGGTACA GCCACGACTG CCCCACTGAA

CACTTAAATG GAAAAAGCAT CCATGATATT GAAAATGTTC TCTTAAAACC AGAAAATTTA 13 2 1 CACTTAAATG GAAAAAGCAT CCATGATATG GAAAATATTG TCATAAAACC GGAAAAGTTA 1380 CACTTAAATG GAAAAAGCAT CCAGGATATT GAAAATGTTG CCTTAAAACC AGAAAAAATG

TATAATTGA 13 8 1 TACGACTGA 1389 TATGATCTAG TGATGTGA

28 constants of 0.38/minute and 0.46/minute, respectively. A representative dissociation curve is shown for CGRP] receptor-transfected HEK 293 cells in Figure 8 A. The association rate constant of ['^I]ha-CGRP was also determined, and a representative association curve is shown in Figure 8 B. K.bs values (0.76/minute and 1.8/minute in

HEK 293 cells and coronary arteries, respectively) were converted to association rate constants (9.5 x ! (f/M/minutc and 3.4 x 10'^/M/minute for CGRP] receptor transfected

HEK 293 cells and coronary arteries, respectively) by the formula

Kassoc= (Robs - K(],ssoc)/[radioligand]. The calculated Kr value (Kdissot/K assoc) was 40 pM in HER 293 cells expressing the porcine CGRP] receptor and 14 pM in coronary arteries.

These values are similar to R^ values previously reported for [*''l]ha-CGRP in saturation binding experiments using HER 293 cells stably transfected with the porcine (R^ =38 pM) or human (K^ = 19 pM) CGRP] receptor (Elshourbagy et al., 1998; Aiyar et a!.,

1996).

Binding of CGRP Receptor Ligands to CGRPj Receptors in HER 293 Cells.

CGRP] receptor-transfected HER 293 cells were used to characterize the ligand binding properties of the CGRP] receptor. ['^I]ha-CGRP \/as used to label CGRP receptors as previously described (Abel et al., 1997). Specific binding was > 90 % in membranes from CGRP] receptor-transfected HER 293 cells, and maximal inhibition of

['^Ijha-CGRP binding for each ligand (except calcitonin) was not different from the maximal inhibition caused by 1 pM ha-CGRP. ha-CGRP and the CGRP] receptor selective ligand, ha-CGRPg.37, bound with high affinity to membranes from HEK 293 cells stably expressing the CGRP] receptor (Figure 9 A, Table 3). ha-CGRPg.37 derivatives that have been previously shown to have a broad range of affinities for CGRP

29 A

B

Figure 8. Kinetics of ['"'l]ha-CGRP binding to the cioned porcine CGRP] receptor. Representative dissociation (panel A) and association (panel B) curves are shown for ['^I]ha-CGRP binding to the porcine CGRP] receptor expressed in HEK 293 cells. The association equilibrium constant (Kassoc) was calculated by the equation Kassoc " (K.bs - Kdissoc)/[radioligand]. The binding equilibrium dissociation constant (K^) was calculated by the equation: K^ = Kdissoc/Kassoc-

30 A

- ha-CGRP v ha-CGRPg_37 A Adrenomedullin a [Cys(ACM)2'7]ha-CGRP * Calcitonin

Figure 9. Mean competition binding curves for selected ligands at porcine CGRP receptors. Panel A shows ligand inhibition of ['""l]ha-CGRP binding in membranes from HEK 293 cells expressing the cloned porcine CGRPi receptor. Panel B shows ligand inhibition of ['^l]ha-CGRP binding in membranes from porcine coronary arteries. Curves represent the mean of three to Eve individual experiments each using cells grown in different cell culture Basks or using porcine coronary arteries Eom different animals.

31 TABLE 3. K] values of CGRP analogs in pig coronary arteries and in pig CGRP] receptor-transfected HEK 293 cells.______CGRP, RECEPTOR- PIG CORONARY TRANSFECTED HEK 293 ARTERIES CELLS

Ligand K, (nM) Hill Slope" K, (nM) Hill Slope"

ha-CGRP 0 .1 1 + .08 -0.94 . *0.08 + 0 .0 1 -0.85

[N-benzoyl]ha-CGRPs.37 0.27* 0.072 + 0.006 -0.96

[N-acetyljha-CGRPs.37 1.3+ 0 .8 -0.82 1.3 + 0.05 -0.82

ha-CGRPg.37 6 .6 + 6.3 -0.90 5.7 + 1.7 -0.84

[Pro"]ha-CGRP 9.0+ 6.3 -0.70 5.5 + 1.3 -0.80

[D-Pen^]ha-CGRP 21.0 + 5.9 -0.93 15.2 + 2.7 -0.95

adrenomedullin 19.9+11.6 -0.97 ' 19.9+12.9 -0.80

[Cys(ACM)^]ha-CGRP 38.4 + 25.2 -0.81 32.0 + 5.8 -0.82

[Cha^]ha-CGRPg.37 32.7+13.6 -0.96 37.9 + 4.3 -0.90

[Pro^,Phe^]ha-CGRP27-37 50.7+16.7 -1.05 118.2 + 40.2 - 1 .1 0

[Ala^]ha-CGRP8-37 499.9 + 265 - 1 .0 0 1 0 2 1 + 2 0 0 -0.53

calcitonin no binding no binding

Each value represents the mean + S.E.M. of three to five separate membrane preparations, each from a different pig or from a different group of cells. A two-tailed t-test of pKi values found no significant differences (p < 0.05) for any of these peptides when comparing their affinities for membranes from pig coronary arteries to their affinities for membranes from porcine CGRP] receptor-transfected HEK 293 cells. "The 95 % confidence intervals of the Hill slope included the value of -1.0 for each ligand except for h-a.CGRP3.37, [Pro'^]h-aCGRP, [Ala^Jh-aCGRPs^, and adrenomedullin in the pig CGRP] receptor clone. *From Smith et al., (2001).

32 receptors in functional assays (Smith et al. 2001) were also examined in this study. The affinity of ha.-CGRPg.37 was increased 4-fold by acetylation ([N-acetyl]ha-CGRPg.37) and 79-fold by benzoylation ([N-benzoyl]ha-CGRPg.37) of the amino terminus. In contrast, the affinity of ha-CGRPg-37 was dramatically decreased by replacing the phenylalanine at position 37 with either alanine ([Ala^jha-CGRPg^) or cyclohexylalanine ([Cha^]ha-CGRP8.37). [Pro'^]ha-CGRP, a putative CGRP? receptor selective agonist, bound with relatively high affinity. The prototypical CGRP2 selective agonist, [Cys(ACM)^]ha-CGRP, bound to the CGRPt receptor with 400-fold lower affinity than ha-CGRP. Adrenomedullin, another member of the CGRP peptide family, competed for ['^I]ha-CGRP binding sites with 248-fold lower affinity than ha-CGRP, and calcitonin (1 nM - 1 pM) did not compete at all. Competition binding curves for several of these ligands are shown in Figure 9 A, and mean Kj values for all ligands are listed in Table 3. Competition curves for each ligand (except calcitonin) fit best to a single site binding model.

Binding of CGRP Receptor Ligands to CGRP Receptors in Porcine Coronary

Arteries. Competition binding was also used to characterize CGRP receptors in porcine coronary arteries. Specific binding was > 70% in all experiments using membranes from porcine coronary arteries, and maximal inhibition of ['^I]ha-CGRP binding for each ligand (except calcitonin) was not different from the maximal inhibition caused by 1 pM ha-CGRP. K] values were determined for the prototypical CGRPt and CGRP2 receptor selective ligands (ha-CGRPg-37 and [Cys(ACM)^]ha-CGRP, respectively) as well as several other peptides. Previous studies have demonstrated that the ha-CGRPg.37 derivatives used in this investigation inhibit CGRP-induced relaxation of porcine

33 coronary arteries with Kg values ranging from 29 nM to > 200 pM (Smith et al., 2001;

Saha et al., 1998; Waugh, personal communication). Therefore, I characterized the

CGRP receptors in coronary arteries with ligands that were predicted to bind with a broad range of affinities. ha-CGRP and ha-CGRPs.37 bound to membranes from porcine coronary arteries with high affinity (0.11 nM and 6 .6 nM, respectively). Consistent with previous studies (Smith et al., 2001), the affinity of ha.-CGRPg.37 was increased 5-fold by acetylation and 24-fold by benzoylation of the amino terminal. In contrast, the affinity of ha-CGRPg-37 was reduced by replacing the phenylalanine at position 37 with alanine or cyclohexylalanine (Table 3). The CGRP2 receptor selective peptide,

[Cys(ACM)^]ha-CGRP, bound with a 349-fold lower affinity than ha-CGRP and a

6 -fold lower affinity compared to ha-CGRPs.37. The high affinity binding of ha-CGRPs.37 and lower affinity binding of [Cys(ACM)^'']ha-CGRP is consistent with the presence of the CGRPt receptor in this tissue. Adrenomedullin competed for

['*^I]ha-CGRP binding sites with 180-fold lower affinity than ha-CGRP, and calcitonin

(1 nM - 1 pM) did not compete for ['^l]ha-CGRP binding sites at all. Competition binding curves for several of these ligands are shown in Figure 9 B, and mean K[ values for all ligands are listed in Table 3. Competition curves for all ligands (except calcitonin) fit best to a single binding site model.

Comparison of Ligand Affinities for CGRP Receptors in Porcine Coronary

Arteries and HEK 293 cells. A comparison of ligand affinities in membranes from porcine coronary arteries and HEK 293 cells expressing the porcine CGRP] receptor is shown in Table 3. There were no significant differences (p < 0.05) for any of these peptides when comparing their affinities for the CGRP receptors in porcine coronary

34 arteries to their affinities for porcine CGRPi receptors transfected into HEK 293 celis. In addition, pKi values in porcine coronary arteries and porcine CGRPi receptor-transfected

HEK 293 cells demonstrated a strong correlation (C - 0.99) (Figure 10 A).

Antagonist affinities (Ki) determined by competition binding were 22- to 707-fold higher than the affinities that have previously been observed for these ligands in functional assays that measure their inhibition of CGRP-induced relaxation of isolated porcine coronary arteries (Table 4). Therefore, 1 examined the correlation between the affinity values determined by functional assays with coronary arteries (pKg) and the affinities determined by competition binding using membranes (pKi) from the same tissue. The affinity of each antagonist was higher when measured by radioligand binding than when measured in functional assays. Thus the linear regression lines correlating these data are not superimposed with the line of identity (Figure 10 B). However, there was a high correlation (r^ = 0 .8 8 ) between affinity values determined by the two methods.

There was also a high correlation (r^ - 0.86) between radioligand binding affinities in

CGRPi receptor-transfected HEK 293 cells and affinities determined by functional experiments with isolated coronary arteries (Figure 10 C). For both correlations (Figure

10 B and Figure 10 C), the 95 % confidence interval of the slope of the regression line included the value of 1 .0 .

Activation of CGRP^ Receptors by a Putative CGRP? Receptor Selective Agonist.

[Pro'^]ha-CGRP has been reported to be a CGRP2 receptor agonist that has no intrinsic activity at CGRPi receptors (Li et al., 1997). However, incubation of this agonist with

HEK 293 cells expressing the porcine CGRP] receptor caused a dose-dependent increase in intracellular cAMP (EC50 = 158 + 57 nM) (Figure 11). Furthermore, the potency of

35 PIG CORONARY ARTERY (pK,) c

PIG CGRPi CLONE (pK,)

Figure 10. Correlation plots of affinity values determined by radioligand binding (pK[) or isolated tissue experiments (pKs). Panel A shows the correlation between antagonist affinities determined by radioligand binding in coronary arteries and HEK 293 cells expressing the porcine CGRPi receptor. Panel B illustrates the correlation between antagonist affinities determined by radioligand binding or isolated tissue experiments with porcine coronary arteries. Panel C shows the correlation between antagonist affinities determined by radioligand binding in HEK 293 cells expressing the porcine CGRP] receptor and antagonist affinities determined by isolated tissue experiments in porcine coronary arteries. The solid line is the linear regression line calculated from the data points. The dotted lines represent the 95 % confidence interval of the linear regression, and the dashed line represents the line of identity. The slope and the correlation coefficient (r^) of the linear regression of the data points are also indicated.

36 TABLE 4. Comparison of ligand affinities determined by competition binding (Ki) or isolated tissue experiments (Ks) in pig coronary arteries. ______ANTAGONIST Ki (nM) Hu (nM) Ke/Kt

[N-benzoyl]ha-CGRPg.37 0.27" 40.36" 149

[N-acetyl]ha-CGRPs.37 1.3+ 0 .8 29.2" 2 2

hor-CGRPg.37 6 .6 + 6.3 4,670^ 707

[D-Pen^]ha-CGRP 21.0 + 5.9 629 + 94.3" 30

[Cha^]ha-CGRPg.37 32.7+13.6 9,528 + 1,156^ 291

[Ala^]ha-CGRPg.37 499.9 + 265 205,589 + 34,500^ 411

Unless indicated otherwise, Ki values were from the data in Table 3. Values from Smith et al. (2001)", Waugh et al. (1999)^, Saha et al. (1998)", or Waugh^, personal communication.

37 iog [AGONIST]

Figure 11. Mean concentration response curves for [Pro'^]ha-CGRP and ha-CGRP in HEK 293 cells stably expressing the porcine CGRPt receptor. Curves represent the mean of 4 experiments performed in separate 12 well plates. These data are normalized to the maximal response produced by [Pro'^]ha-CGRP or the maximal response produced by ha-CGRP.

38 this putative CGRP2 receptor selective agonist was only 3-fold lower than that of ha-CGRP (EC50 = 51+27 nM).

Discussion

CGRP receptors have been classified mto CGRP] and CGRP2 receptor subtypes based upon their affinity for ha-CGRPs.37 in isolated tissue experiments. ha-CGRPs-37 is an antagonist that inhibits CGRP] receptor-mediated responses with high affinity

(Rs <100 nM) and inhibits putative CGRP2 receptor-mediated responses with low affinity (Ks > 1 pM) (Dennis et al., 1990). fn addition, [Cys(ACM)*^]ha-CGRP has been proposed to be an agonist at CGRP2 receptors and inactive at CGRP] receptors

(Dennis et al., 1989). In isolated tissue studies, Foulkes et al. (1991) have previously reported that CGRP receptors in large porcine coronary arteries (outside diameter > 1

mm) have low affinity for ha-CGRPg^ (Ks > 1 pM). Waugh et al. (1999) reported that

ha-CGRPg-37 has low affinity (Ks > 1 pM) in large porcine coronary arteries and that

[Cys(ACM)^]ha-CGRP, causes relaxation of this tissue (Waugh et al., 1999). These

functional studies have established large coronary arteries as a model for studying the

CGRP2 receptor.

In contrast to previous studies, I provide two independent lines of evidence to

show that porcine coronary arteries express the CGRP] receptor. First, I have identified

mRNA encoding CRLR and RAMP 1 in porcine coronary arteries. These proteins have

been previously shown to form the CGRP] receptor (McLatchie et al., 1998). Second, I

have found that CGRP] and CGRP2 receptor-selective ligands do not discriminate

between CGRP receptors in porcine coronary arteries and porcine CGRP] receptors that

39 have been transfected into HEK 293 celis. Furthermore, correlations between affinity values determined by radioligand binding and by isolated tissue experiments suggest that the CGRP receptor that has low affinity for ha-CGRPs-37 in functional studies with coronary arteries is the same CGRP receptor that has been cloned and expressed in HEK

293 cells.

I used radioligand binding to compare the affinity of several ligands for CGRP receptors expressed in coronary arteries to their affinities for CGRP] receptors expressed in HEK 293 cells. Stable expression of the porcine CGRP] receptor in HEK 293 cells provides a system that is free of other receptors that might bind CGRP and its analogs.

This also allowed the comparison of CGRP receptor subtypes from the same species.

This is important because many other comparisons of CGRP] and CGRP? receptors have been complicated by the use of tissues from different species to represent the putative

CGRP receptor subtypes. I found that the affinity of ha-CGRPg-37, the prototypical

CGRP] receptor-selective antagonist, and of [Cys(ACM)^]ha-CGRP, the putative

CGRP2 receptor-selective agonist, were nearly identical in membranes prepared from porcine coronary arteries and from porcine CGRP] receptor-transfected HEK 293 cells.

In fact, each of the twelve ligands used in the radioligand binding assay bound to a single binding site in the porcine coronary artery, and none of the ligands were capable of discriminating the cloned CGRP] receptor from the CGRP receptor that is found in porcine coronary arteries. These data provide evidence that porcine coronary arteries express only the CGRPt receptor.

In contrast to previous functional studies with isolated coronary arteries, the radioligand binding data in the present investigation demonstrate that ha-CGRPg-37 has a

40 CGRP] receptor-iike affinity for CGRP receptors in porcine coronary arteries. This raised the possibility that the CGRP receptor identified by radioligand binding is not the same CGRP receptor that has been characterized by isolated tissue experiments.

Therefore, I used correlation plots to compare the antagonist affinities determined by isolated tissue experiments (Ks) to their affinities determined by radiohgand binding (K<).

Each antagonist demonstrated a lower affinity in functional experiments with isolated coronary arteries than in radioligand binding experiments with membranes prepared from the same tissue. However, the rank order of antagonist affinities was the same in both assays. Furthermore, correlation plots demonstrated that the difference between K.s and

Kj values was consistent for each ligand, and that there was a high correlation between affinity values determined by the two different methods. Regardless of the method used to measure ligand affinities in porcine coronary arteries, there was also a high correlation between affinity values in this tissue and in membranes from CGRP] receptor-transfected

FIEK 293 cells. These data support the conclusion that isolated tissue experiments and competition binding experiments with porcine coronary arteries both identify the same

CGRP receptor.

These radioligand binding data raise an important question: Why does ha-CGRPg-37 appear to identify a low affinity CGRP? receptor in functional assays with isolated tissue but not in competition binding experiments? One explanation for the low affinity of ha-CGRPg-37 in isolated porcine coronary arteries, rat vas deferens, and other tissues is that this ligand may be degraded by proteases, causing the peptide to appear to have a lower affinity in these tissues than in tissues that lack these enzymes. Femandez-

Patron et al. (2000) reported that matrix metalloprotease-2, a protease present in vascular

41 smooth muscle and endothelium, specifically cleaves h(3-CGRP into h(3-CGRP].]4 and hp-CGRP]5-37. In addition, Smith et al. (2001) found that acetylation of the amino terminal of ha-CGRPs.37, a modification that has been demonstrated to protect other peptides from degradation (Drapeau et al., 1993), caused a 160-fold increase in the affinity (Ks) of this ligand in functional relaxation assays with porcine coronary arteries.

In contrast, acetylation of ha-CGRPg-37 caused only a 5-fold increase in its binding affinity (K[) for membranes from the same tissue. These data suggest that ha-CGRPg-37 may be more susceptible to proteolytic degradation in whole coronary arteries than in membranes. Three nonpeptide CGRP receptor ligands (SB-273779, BIBN4096BS, and

Compound 1) have recently been developed (Doods et al., 2000; Aiyar et al., 2001a;

Edvinsson et al., 2001) and may be useful for avoiding ligand degradation while studying

CGRP receptors.

A disadvantage of using membranes for radioligand binding experiments is that the receptor is removed from its native environment and placed under conditions that may cause receptor accessory proteins to be lost. ha-CGRPg.37 has a low, CGRPi receptor-like, affinity in functional studies using isolated porcine coronary arteries and a high, CGRP] receptor-like affinity in membrane binding assays. One explanation for this difference is that the binding characteristics of the CGRP] receptor may be modified during the membrane preparation procedure. McLatchie et al. (1998) and Evans et al.

(2000) have reported that RAMP 1 and receptor component protein (RCP) are required to

form a functional CGRP] receptor. RAMP 1 is an integral membrane protein with a

membrane spanning domain that presumably protects it from being lost during the

42 membrane preparation procedure. However, RCP is a peripheral membrane protein that can be dissociated from the membrane (Evans et al., 2000). Although the affinity of ha-CGRP is unaffected by the presence or absence of RCP (Evans et al., 2000), the effect of this protein on ha-CGRPs^ has not been examined. It is possible that low affinity ha-CGRP8-37 binding is conferred by RCP in intact tissues and that ha-CGRPR-37 does not have low affinity in competition binding experiments because RCP is lost during the membrane preparation procedure.

Agonists were also used to characterize CGRP receptors in this study. Previous investigators have reported that [Cys(ACM)^]ha-CGRP and [Pro'^]ha-CGRP are selective for the putative CGRP 2 receptor (Dennis et al., 1989; Li et af, 1997). In contrast, these ligands demonstrated no selectivity for CGRP receptors in coronary arteries over CGRPt receptors in HER 293 cells in my competition binding experiments.

Furthermore, I found that the putative CGRP2 receptor agonist, [Pro'^ha-CGRP, stimulates cAMP production in porcine CGRPt receptor-transfected HEK 293 cells. The ability of [Cys(ACM)^ '']ha-CGRP and [Pro'^]ha-CGRP to induce responses in some

(but not all) tissues that express CGRP] receptors suggests that the tissue selectivity of these ligands is caused by something other than expression of different CGRP receptor subtypes. Waugh et al. (1999) demonstrated that [Cys(ACM)*']ha-CGRP is a partial agonist in porcine coronary arteries. Since the efficacy of partial agonists is limited by receptor reserve (Kenakin, 1993), the ability of [Cys(ACM)^]ha-CGRP and

[Pro^]ha-CGRP to activate CGRPt receptors in porcine coronary arteries and CGRP]

receptor-transfected HEK 293 cells (Figure 11; Waugh et al., 1999) but not in other

tissues that express CGRP] receptors (Li et al., 1997; Dennis et al., 1989) may reflect

43 tissue-dependent amounts of CGRP] receptor reserve rather than the existence of two different CGRP receptor subtypes. Therefore, the existence of a second CGRP receptor subtype is not necessary to explain the ability of [Cys(ACM)^]ha-CGRP and

[Pro'**]ha-CGRP to be agonists in some isolated tissues and inactive in others.

In summary, CGRP receptors in porcine coronary arteries have been previously classified as the CGRP2 receptor subtype because ha-CGRPg-37 antagonizes CGRP- induced relaxation of these arteries with low affinity. However, I have demonstrated that porcine coronary arteries have CRLR and RAMP 1 mRNA and that the ligand binding characteristics of CGRP receptors in porcine coronary arteries are identical to those of the cloned porcine CGRP] receptor. Furthermore, the correlation between Kg and K[ values are consistent with the conclusion that the low affinity (Kg > 1 pM) of ha-CGRPg.37 in functional studies using isolated porcine coronary arteries occurs at the CGRP] receptor.

44 CHAPTER 3

IDENTIFICATION OF AN ADRENOMEDULLIN RECEPTOR AS THE

PUTATIVE CGRP, RECEPTOR IN SV40LT-SMC CELLS

Abstract

CGRP receptors are subdivided into CGRP, and CGRP, receptor subtypes based on their affinity for ha-CGRPg_3y in functional assays. CGRP, receptor-mediated responses are blocked with high affinity (Kg < 100 nM) by ha-CGRPg„, and CGRP, receptor-mediated responses are blocked with low affinity (Kg > 100 nM) by this antagonist. The CGRP, receptor has been cloned and well characterized by functional assays using isolated tissues and by radioligand binding. In contrast, radioligand binding studies using ['*T]ha-CGRP do not support the existence of the putative CGRP, receptor.

This receptor has only been identified by its low affinity (Kg > 100 nM) for ha-CGRPg,, in functional assays. CGRP also stimulates adrenomedullin receptors. However, the affinity with which adrenomedullin receptor-mediated responses are blocked by ha-CGRPg.37 has not been previously examined. In this study, I tested the hypothesis that the putative CGRP, receptor is an adrenomedullin receptor. ha-CGRPg_3, blocks CGRP- induced cAMP synthesis in SV40LT-SMC cells with a CGRP, receptor-like affinity.

Therefore, these cells were used to further characterize the putative CGRP, receptor.

Adrenomedullin stimulated cAMP synthesis with 25-fold greater potency than CGRP.

However, ha-CGRP^? blocked CGRP-induced cAMP synthesis and adrenomedullin- induced cAMP synthesis with similar affinities (Kg - 4630 nM and 3333 nM for CGRP and adrenomedullin, respectively), suggesting that the effects of both agonists were

45 mediated through a common receptor. The presence of ['^IJ-adrenomedullin binding sites and the absence of ['^I]ha-CGRP binding sites suggest that SV40LT-SMC cehs express adrenomedullin receptors but lack CGRP receptors. Taken together, these data support the conclusion that the putative CGRP 2 receptor is an adrenomedullin receptor.

Radioligand binding demonstrated that the adrenomedullin receptor expressed in

SV40LT-SMC cells is identical to an adrenomedullin receptor expressed in SK-N-MC cells (rank order of affinity: adrenomedullin > ha-CGRP > ha-CGRP = adrenomedullin22.32)- An additional adrenomedullin receptor (rank order of ligand affinity: ha-CGRP = adrenomedullin > ha-CGRPg.^ > adrenomedullb^^) that was not blocked by adrenomedullh^^ was also identified in SK-N-MC cells. Finally, this investigation identified mRNA encoding the novel RAMP 2 B accessor)' protein in

SV40LT-SMC cells, rat lung, and rat vas deferens. This transcript encodes a protein that is identical to RAMP 2 except that amino acid residues Ala"* - Gly^' are deleted. The functional significance of this transcript is currently unknown.

Introduction

CGRP receptors are divided into two subtypes on the basis of their affinity for ha-CGRPg_37 in functional assays using isolated tissues. CGRP, receptor-mediated responses are blocked with high affinity (Kg < 100 nM) by ha-CGRPg.^, and CORP2 receptor-mediated responses are blocked with low affinity (Kg > 100 nM) by this antagonist. The CGRP, receptor has been cloned and well characterized using both functional assays with isolated tissues and radioligand binding assays with membranes prepared from isolated tissues. In contrast, the putative CGRP2 receptor has not been

46 cloned and has only been identified by its low affinity for ha-CGRP in functional assays. The CGRP2 receptor has been identified in rat and guinea pig vas deferens, porcine coronary arteries, rat internal anal sphincter muscle, and other isolated tissues

(Dennis et al., 1990; Foulkes et al., 1991; Wisskirchen et al., 2000). However, radioligand binding assays using membranes prepared from these tissues indicate that they express only the CGRP, receptor (Dennis et al., 1990; Mimeault et al., 1991;

Rorabaugh et al., 2001). Consequently, a significant challenge in this field has been to understand why two different CGRP receptor subtypes are identified by ha-CCRP^; in functional assays but only one CGRP receptor is identified in radioligand binding assays.

The adrenomedullin receptor is structurally similar to the CGRP, receptor

(McLatchie et al., 1998). Both of these G protein coupled receptors are composed of a heptahelical protein called calcitonin receptor like receptor (CRLR) and two accessory proteins. One accessory protein, receptor component protein (RCP), does not influence ligand binding but is required for CGRP, and adrenomedullin receptors to be coupled to adenylate cyclase (Evans et al., 2000). In addition, a group of receptor activity modifying proteins (RAMPs) is involved in the intracellular trafficking and glycosylation of CRLR (McLatchie et al., 1998). When coexpressed with RAMP 1, CRLR undergoes terminal glycosylation and forms the CGRP, receptor. When coexpressed with RAMP 2 or RAMP 3, CRLR undergoes core glycosylation and forms an adrenomedullin receptor.

Thus the CGRP, receptor and adrenomedullin receptor differ only by their glycosylation state and the presence of different RAMPs (Figure 4).

The structural similarity of CGRP, receptors and adrenomedullin receptors allows these receptors to be activated by both CGRP and adrenomedullin (Fraser et al., 1999).

47 Consequently, some ha-CGRP-induced responses may be mediated by adrenomedullin receptors that are blocked with low affinity by ha-CGRPg,^. In the present investigation,

I examined the hypothesis that the putative CGRP 2 receptor that has been previously identified m functional assays is actually an adrenomedullin receptor. Several different cell types were screened to identify a cell line in which CGRP-induced cAMP synthesis is blocked with low (Kg > 100 nM) affinity by ha.-CGRPg.37. The receptors mediating this response were further characterized by radioligand binding and RT-PCR. My data support the conclusion that the putative CGRP^ receptor is an adrenomedullin receptor.

Materials and Methods

Chemicals and Reagents. Tag DNA polymerase, 10 X PCR buffer (200 mM

Tris-HCl, pH 8.4, 500 mM KC1), amplification grade DNAse I, 100 base pair DNA ladder, TRIzol, Dulbecco's modified Eagle medium, minimum essential medium, fetal bovine serum, and antibiotic/antimycotic (containing 10,000 units/ml penicillin G,

10,000 pg/ml streptomycin sulfate, and 25 pg/ml amphotericin B) were purchased from

Gibco (Grand Island, NY). RNasin was purchased from Promega (Madison, WI), and

Moloney murine leukemia virus reverse transcriptase was purchased from Perkin Elmer

(Foster City, CA). The pCRII cloning vector was purchased from Invitrogen (Carlsbad,

CA). ['^I]ha-CGRP and ['^I]rat adrenomedullin were purchased from Amersham

Pharmacia Biotech (Piscataway, NJ). Reagents for measuring intracellular cAMP were provided as a kit from the Diagnostic Products Corporation (Los Angeles, CA), and peptide ligands were purchased from Peninsula Laboratories (San Carlos, CA).

48 Na2Ca-ethylenediaminetetracetic acid, Tris(hydroxymethyl)aminomethane, and other chemicals were obtained from Sigma (St. Louis, MO). A rat aortic smooth muscle cell line (SV40LT-SMC cells) was purchased from the American Type Culture Collection

(Manassas, VA), and human SK-N-MC cells were a generous gift from Dr. Myron

Toews of the University of Nebraska Medical Center (Omaha, NE).

Cell Culture. SK-N-MC cells used for radioligand binding were grown in T-175 culture flasks containing Dulbecco's modified Eagle medium that was supplemented with fetal bovine serum (10 %), penicillin G (100 units/ml), streptomycin (100 pg/ml), and amphotericin B (0.25 pg/ml). SV40LT-SMC cells used for radioligand binding and all cells used for cAMP assays were grown in minimum essential medium that was supplemented as described above. The flasks were placed in a humidified incubator with an atmosphere of 5 % CO2 /9 5 % air and maintained at 37°C.

Measurement of Intracellular cAMP. Cells were grown in 24 well culture plates under the conditions described above. Culture medium was removed from confluent cells and the cells were washed twice with HEPES buffered Krebs solution containing isobutylmethylxanthine (HKI) (20 mM HEPES, 4 mM NaHCOg, 11 mM dextrose, 1.25 mM Na^POi, 120 mM NaCl, 5.5 mM KC1, 2.5 mM CaC^, 1.2 mM MgC^, and 0.5 mM isobutylmethylxanthine). Cells were preincubated for 10 minutes with 450 pi HKI in the presence or absence of 10 pM ha-CGRPg ^ before 50 pi of ha-CGRP or adrenomedullin was added. The appropriate agonist solutions also contained 10 pM ha-CGRPg.^ so that a constant antagonist concentration was retained throughout the experiment. Following a

30 minute exposure to ha-CGRP or adrenomedullin, the solution was removed and cells were lysed with 100 pi 90 % ethanol. The lysate was dried by incubating the cells at

49 37°C for 45 minutes, and cAMP in the dried lysate was measured using a radioimmunoassay from Diagnostic Products Corporation (Los Angeles, CA).

Membrane Preparation. Culture medium was removed from confluent cells, and the cells were rinsed 3 times with 25 ml of ice-cold phosphate buffered saline. Cells were dislodged from the flask with a cell scraper in the presence of 1 0 ml ice-cold phosphate buffered saline and centrifuged at 4°C for 5 minutes at 1000 x g. The pellet was suspended in 25 ml buffer A (50 mM Tris-HCl, 5 mM l^ C a - ethylenediaminetetracetic acid, pH 7.4) by vortexing and then homogenized with a glass- teflon homogenizer. The homogenate was centrifuged at 100,000 x g for 30 minutes in a

Beckman L5--50 ultracentrifuge. The pellet was washed twice by homogenization in 25 ml buffer B (50 mM Tris-HCl, 100 mM NaCl, 5 mM MgCl^, pH 7.4) followed by centrifugation at 4°C for 30 minutes at 100,000 x g. The supernatant was removed and the dry pellet was stored for up to 1 month at -70°C. Protein content of the final pellet was determined by the method of Lowry (Lowry et al., 1951).

Radioligand Binding Assay. The radioligand binding assay used in this study has been previously described in detail (Abel et al., 1997). Frozen membrane pellets were rehomogenized in ice-cold binding buffer (50 mM Tris-HCl, 5 mM MgC^, 100 mM

NaCl, 0.2 % bovine serum albumin, 0.1 % bacitracin, pH 7.4) to a concentration of

50-100 pg membrane protein/150 pi. Membrane protein (150 pi) homogenate was added to 12 x 100 mm polyethylene tubes. The tubes were incubated in a 37°C shaking water bath for 50 minutes in the presence of 40 pM ['^l]ha-CGRP or 40 pM

['^I]rat adrenomedullin and various concentrations of non-radiolabeled ligands. The total incubation volume was 250 pi. Nonspecific binding was determined using lpM

50 ha-CGRP or 1 pM human adrenomedullin. Whatman GF/B glass microfiber filters were

soaked in 0.2 % polyethyleneimine for 30 minutes prior to their use. Bound and free

radioligand were separated by trapping the membranes on the filters and washing with 15

ml buffer B using a Brandel MB-48R cell harvester (Gaithersburg, MD). The filters were

transferred into separated x 75 mm polyethylene tubes, and bound ['^I]ha-CGRP was

measured in a Wallac Gammamaster 1277 (Gaithersburg, MD) gamma counter.

RNA Isolation. Culture medium was removed from SV40LT-SMC or SK-N-MC

cells grown in T-175 cell culture flasks, and the cells were rinsed 4 times with phosphate

buffered saline (137 mM NaCl, 2.7 mM KC1, 4.3 mM Na;HP0 4 , 1.4 mM KH2PO4, pH =

7.3). 15 ml of TRIzol reagent was added to the flask and the cells were dislodged from

the flask with a cell scraper. The TRIzol reagent containing the cells was divided into 1

ml aliquots, and RNA was isolated from each aliquot according to the manufacturer's

protocol. Isolated RNA was dissolved in RNAse free water, and stored at -70°C.

Male Sprague Dawley rats (weighing approximately 200g) were euthanized with

60 mg/kg sodium pentobarbital. Vasa deferentia were isolated and cleaned of connective

tissue with the aide of a dissecting microscope. The tissue was wrapped in aluminum

foil, frozen at -70°C, and pulverized with a hammer. TRIzol reagent was used to isolate

total RNA from approximately 100 mg of pulverized tissue according to the

manufacturer's protocol. The RNA was dissolved in RNAse-free water and stored at -

70°C.

Identification of mRNA Encoding CRLR and RAMP by RT-PCR. RT-PCR was

used to identify mRNA encoding CRLR, RAMP 1, RAMP 2, and RAMP 3 in rat

SV40LT- SMC cells and the rat vas deferens. Primers (Table 5) were designed based on

51 Tabfe 5. Primers used to identify mRNA encoding CRLR and RAMPs by RT-PC R. Transcript Annealing (species) Primer Temp (°C) CRLR Sense: 5'-GGATCCGGATCCGCCACCATGGAGAAAAAGTGTA-3' 55 (human) Antisense: 5'-CTCGAGCTCGAGTCAATTATATAAATTTTCT-3' CRLR (rat) Sense: 5'-CTCCTCTACATTATCCATGG-3' 52 Antisense: 5'-CAATTCTCATTTCATTTGAGGGC-3' RAMP 1 Sense: 5'-CGGCGCGGCCTCTGGCTGC-3' 54 Antisense: 5 ATGCCCTCTGTGCGCTTGCTC-3' RAMP 2 Sense: 5'-CGCTCCGGGTAGAGCGCGCC-3' 58 Antisense: 5'-CCAAGGAGCAGTTGGCAAAGTG-3' RAMP 3 Sense: 5'-CGTCTGGAAGTGGTGCAACCTG-3' 58 Antisense: 5'-CACGACCAGGTAGGGTGTGGC-3' Cyclophiiin Sense: 5'-GTCTCCTTCGAGCTGTTTGC-3' 52, 54, Antisense: 5'-ATCTTCTTGCTGGTCTTGCC-3' 55, or 58

52 the complementary DNA (cDNA) sequences of the rat CRLR (accession # L27487), rat

RAMP 1 (accession # AF181550), rat RAMP 2 (accession # AF181551), and rat RAMP

3 (accession # AF181552). CRLR mRNA was also detected in SK-N-MC cells using primers designed from the human CRLR sequence (accession # L76380).

Immediately prior to RT-PCR, 2 pg of RNA were incubated for 30 - 60 minutes with 1 unit of amplification grade DNase I in DNase I buffer. This incubation was performed at room temperature in the presence of 10 units of RNasin. DNase 1 was inactivated by heating the RNA to 75°C for 10 minutes in the presence of 2.5 mM

EDTA. cDNA encoding CRLR, RAMP 1, or RAMP 2 was synthesized by reverse transcription in a 10 pi reaction volume containing 1 X PCR buffer, 2 pg RNA, 5 mM

MgC^, 1 mM dNTP mix, 25 pmoles antisense primer, and 25 units Moloney murine leukemia virus reverse transcriptase. cDNA encoding RAMP 3 was synthesized under identical conditions except that 50 pmoles of oligo dT,y primers were used instead of the gene specific antisense primer. Each reaction was incubated at 42°C for 50 minutes followed by a five minute incubation at 98 °C. Control reactions were performed in the absence of reverse transcriptase to verify that the RNA was not contaminated with DNA.

PCR was conducted in a 100 pi reaction volume containing 1 X PCR buffer, 3 mM

MgC^, 0.2 mM dNTP mix, 50 pmoles sense primer, 50 pmoles antisense primer, 10 pi cDNA, and 2.5 units 7bg DNA polymerase. PCR conditions included an initial cDNA denaturation step at 95°C for 3 minutes followed by 35 cycles (95°C denaturation for 30 sec, 52 - 58°C annealing for 30 sec, and 72°C extension for 30 sec) of PCR and a final extension period of 7 minutes at 72°C. The exact annealing temperature for each pair of primers is listed in Table 5. RT-PCR reactions for CRLR mRNA were performed in a

53 Perkin Eimer 480 thermocycler. AH other reactions were performed in a Stratagene

Robocycler. The products of all RT-PCR reactions were visualized on ethidium bromide-stained 1.5 % agarose gels and subcloned into the pCRII vector. Each RT-PCR product was sequenced using an Applied Biosystems 373 DNA sequencer. DNA sequences were analyzed using the Wisconsin Package Version 10.1 (Genetics Computer

Group, Madison, Wl) software.

Data Analysis. Three to six competition binding curves were performed in duplicate for each ligand. Specific binding was determined by subtracting nonspecific binding (defined using 1 pM ha-CGRP or 1 pM human adrenomedullin) from total binding, and the IC^ for each competition curve was determined by nonlinear regression analysis using GraphPad Prism (SanDiego, CA). The data were fit to both one site and two site binding models and an F test was used to determine which model provided the best fit. Mean pIC^ values for each ligand were compared using a two-tailed student's t-test to determine if the pIC^ values for [''H]-adrenomedullin binding sites in SK-N-MC cells were significantly different (p < 0.05) from pIC^ values at ['^IJ-adrenomedullin binding sites in SV40LT-SMC cells or ['*H]ha-CGRP binding sites in SK-N-MC cells.

Dose response curves for CGRP and adrenomedullin were also analyzed by nonlinear regression analysis. EC^ values could not be determined from these data because the highest agonist concentrations used in this investigation did not achieve a maximal response. Therefore, the dose ratio of ha-CGRPg.^ was calculated by determining the concentration of agonist required to achieve a 4.7-fold increase in

CGRP-induced cAMP synthesis or a 9-fold increase in adrenomedullin-induced cAMP synthesis in the presence and absence of ha-CGRPg.^. 4.7-fold and 9-fold increases in

54 cAMP were used because these values represented approximately half of the highest response that was observed in these experiments. Consequently, it was anticipated that they would provide the most accurate dose ratios. The Schild equation

(Kg = [antagonist]/[dose ratio - 1]) was used to convert dose ratios to Kg values.

Results

Identification of Cells Expressing the Putative CGRPi Receptor. A variety of cell types were screened to determine whether incubation with 1 nM - 3 pM ha-CGRP caused a dose-dependent increase in intracellular cAMP. A dose dependent increase in cAMP was detected in SV40LT-SMC and DDT cells but no response was observed in the other cell types that were examined (Table 6 ). Stimulating SV40LT-SMC cells with

3 pM CGRP caused a 10.5 + 2.6-fold increase in intracellular cAMP but the lack of a plateau in the dose response curve suggests that this was not a maximal response. 1 pM adrenomedullin caused a 20.5 + 6.2-fold increase in intracellular cAMP. However, failure of this dose response curve to reach a plateau suggests that this was also a submaximal response. The inability to achieve a maximal response with the agonist concentrations used in this study prohibited the determination of EC^ values from these data. Therefore, the potency of CGRP and adrenomedullin were compared by determining the agonist concentration required to achieve a 7.2-fold increase in basal cAMP. This level of stimulation was chosen because it was within the linear portion of the dose-response curve for both agonists. A 7.2-fold increase was achieved with 794 nM

CGRP and 32 nM adrenomedullin indicating that adrenomedullin was 25-fold more potent than CGRP in SV40LT-SMC cells (Figure 12).

55 Table 6. Ceil lines screened for CGRP? receptors. Cell Type Tissue of Origin Response to CGRP SV40LT-SMC rat aortic smooth muscle + DDT guinea pig vas deferens + smooth muscle

PARC5 rat parotid gland -

1321N1 human neuroblastoma -

Cos African green monkey kidney -

HEK human embryonic kidney -

SAOS-2 human osteosarcoma - porcine aortic endotheiiai cells porcine aorta -

56 20-

-10 -9 -8 -7 -6 -5 !og [agonist]

Figure 12. Mean concentration response curves for ha-CGRP and adrenomeduHin in SV40LT-SMC ceiis. These data represent the mean of five or six experiments each performed in a separate 24 well plate.

57 The affinity with which ha-CG RP^ blocked CGRP and adrenomedullin-

induced cAMP synthesis in SV40LT-SMC cells was measured by determining the dose ratio of the agonist-induced response in the presence and absence of 10 pM ha-CGRP^.

ha-CGRPg_37 caused a rightward shift in the CGRP and adrenomedullin dose response

curves (Figure 13). The dose ratio was determined for CGRP when intracellular cAMP

was 4.7-fold over the basal level and for adrenomedullin when intracellular cAMP was

9-fold over the basal level. These levels of stimulation were chosen because they were

within the linear portion of the CGRP and adrenomedullin dose-response curves,

respectively. The dose ratio of each agonist was converted to a IQ value by the Schild

equation (Kg = [antagonist]/[dose ratio -1]). ha-CGRPg.^ blocked CGRP-induced cAMP

synthesis with a Kg of 4630 nM (pKg = 5.33) and blocked adrenomedullin-induced

cAMP synthesis with a Kg of 3333 nM (pKg = 5.48). Both of these affinity values are

consistent with the presence of the putative CGRP2 receptor in this cell line.

Furthermore, the similarity of these values suggests that CGRP and adrenomedullin

stimulate cAMP synthesis through the same receptor.

Radioligand Binding in SV40LT-SMC and DDT Membranes. Since ha-CGRP g.^

blocked CGRP and adrenomedullin-induced responses with an affinity that was typical of

the putative CGRP2 receptor, this receptor was further characterized by radioligand

binding. Specific binding sites for ['"^Ijha-CGRP were not detected in membranes

prepared from SV40LT-SMC cells or DDT cells (Figure 14). This suggests that CGRP

receptors are not expressed in either cell line. The relatively high potency of

adrenomedullin compared to CGRP in SV40LT-SMC cells, and the lack of specific

58 A

Figure 13. Concentration response curves for ha-CGRP (pane! A) and adrenomedullin (panel B) in the presence and absence of 10 pM ha-CGRP^?. Curves represent the mean of six or seven experiments each performed in a separate 24 well plate.

59 A

-14 -13 -12 -11 -10 -9 -8 -7 -6 -5 log [ha-CGRPg.37]

B

2000-i

O! =5 ^ 1300-1 C

O M1000- ^ Oo "1— 500

0-1------!------I------I------1------1------!------1----—1------1 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 log [ha*CGRPg.3y]

Figure 14. Lack of CGRP receptors in SV40LT-SMC and DDT cells. 40 pM ['^l]ha- CGRP was incubated with 200 pg SV40LT-SMC membrane (panel A) or 400 pg DDT (panel B) membrane in the presence of different ha.-CGRPg.37 concentrations. Specific binding (determined with 1 pM ha-CGRP) was not detectable in membranes from either cell line. These data represent the mean of two or three experiments each using cells grown in different culture flasks.

60 ['^I]ha-CGRP binding sites in SV40LT-SMC membranes, suggests that CGRP and adrenomeduHin stimulate cAMP synthesis through an adrenomedullin receptor rather than through a CGRP receptor. Therefore, ['^Ij-adrenomedullin was used to characterize adrenomedullin receptors in this cell line. Specific binding was > 70 % in all experiments using SV40LT-SMC membranes. Adrenomedullin inhibited

['^IJ-adrenomedullin binding with the highest affinity of all ligands examined, and ha-CGRP inhibited ['^l]-adrenomedullin binding with 71-fold lower affinity than adrenomedullin (Figure 15). ha-CGRPg.^^ and adrenomedullin^ (an adrenomedullin receptor antagonist) inhibited ['^IJ-adrenomedullin binding with 160-fold and 295-fold lower affinity than adrenomedullin, respectively. These experiments demonstrated that

SV40LT-SMC cells express an adrenomedullin receptor that has the following rank order of ligand affinity: adrenomedullin > ha-CGRP = ha-CGRPg.^ = adrenomeduHin^^-

Competition curves for each ligand fit best to a single-site binding model. The IC^ value for each ligand is listed in Table 7.

Identification of CRLR and RAMPs in SV40LT-SMC Cells. Radioligand binding experiments indicated that SV40LT-SMC cells express adrenomedullin receptors but lack CGRP receptors. Therefore, I used RT-PCR to determine whether mRNA encoding

CRLR and the RAMPs that are required to form CGRP, and adrenomeduHin receptors were present in this cell line. Primers spanning a 647 nucleotide region of the rat CRLR mRNA were used to identify CRLR transcripts in SV40LT-SMC cells (Figure 16).

RT-PCR was also used to look for RAMP 1 mRNA in this cell line. RNA isolated from

SK-N-MC cells and from the rat vas deferens was used to ensure that the RAMP 1 primers worked, and cyclophilin primers were used to verify the integrity of the RNA in

61 tog (M) [tigand]

Figure 15. Mean competition binding curves using SV40LT-SMC membranes labeled with ['"dj-adrenomedullin. Curves represent the mean of three experiments using cells grown in separate culture flasks.

62 Tabfe 7. Affinity of CGRP and adrenomeduHin receptor ligands in SV40LT-SMC and SK-N-MC membranes. SV40LT-SMC SK-N-MC

[' ^ 1]- adrenomedul lin ['^I]-adrenomedullin ['^llha-CGRP

Ligand IQ . (nM) IQ.(nM) KQ(nM)

AdrenomeduHin 1.0 ±.07 0.53 + .01" 1.8+ .25

ha-CGRP 71 ± 10 0.07+ .0 1 " (high) 0 . 1 2 + .01

202 + 44' (low)

ha-CGRP^? 160 + 27 41 + 10" 17 + 1.5

Adrenomedullin22_52 295 + 37 190 + 59 not determined

These values represent the mean + S.E.M. of three to six experiments each using membranes prepared from separate cell culture flasks. " A two-tailed t-test found the pICso value of this ligand to be significantly different (p < 0.05) from the pIC^ value determined in SV40LT-SMC membranes labeled with QHj-adrenomedullin. 'A two- tailed t-test found the pIQ. value of this ligand to be significantly different from the pICso value determined in SK-N-MC membranes labeled with ['^I]ha-CGRP.

63 M +

Figure 16. Identification of CRLR mRNA in SV40LT-SMC cells by RT-PCR. The inclusion/exclusion of reverse transcriptase in the reverse transcription reaction is indicated by +/-. Marker bands (M) are 500, 600, 700, and 800 base pair size markers. The size of the RT-PCR product was 647 base pairs.

64 SV40LT-SMC cells. Although RAMP 1 mRNA was identified in SK-N-MC cells and the rat vas deferens, there was no evidence for this transcript in rat SV40LT-SMC cells

(Figure 17). This is consistent with the absence of specific ['^I]ha-CGRP binding sites in the SV40LT-SMC cell line.

When coexpressed with RAMP 2 or RAMP 3, CRLR forms an adrenomedullin receptor (McLatchie et al., 1998). Therefore, I used RT-PCR to determine whether mRNA that encodes RAMP 2 or RAMP 3 was present in SV40LT-SMC cells. Primers that span a 414 nucleotide segment of RAMP 2 mRN A revealed the presence of RAMP 2 mRNA and a smaller (336 nucleotide) RT-PCR product that 1 have named RAMP 2 B

(Figure 18). The nucleotide sequence of RAMP 2 B is identical to the RAMP 2 sequence

(accession # AF181551) except that the 78 nucleotides corresponding to amino acids

Ala"* - Gly*" are deleted (Figure 19). Rat lung RNA was used as a positive control for the identification of RAMP 2 transcripts. RAMP 2 and RAMP 2 B transcripts were also identified in RNA isoiated from this tissue. mRNA encoding RAMP 3 was also identified in SV40LT-SMC cells but was not found in SK-N-MC cells (Figure 20). The presence of mRNA encoding CRLR, RAMP 2 and RAMP 3 in SV40LT-SMC cells is consistent with the presence of adrenomedullin receptors in SV40LT-SMC membranes.

In addition, the absence of RAMP 3 mRNA in SK-N-MC cells is consistent with the northern analysis reported by McLatchie et al. (1998) (Figure 21).

Radioligand Binding in SK-N-MC Membranes. SK-N-MC cells express CRLR,

RAMP 1, and RAMP 2 (Figure 21) and are commonly used to study adrenomedullin and

CGRP, receptors (McLatchie et al., 1998; Poyner et al., 1998). Adrenomedullin receptors in SK-N-MC cells were radiolabeled with ['^Ij-adrenomedullin, and

63 Rat SK-N-MC SV40LT-SMC Vas Deferens

M RAMP 1 cyclophilin RAMP 1 cyclophiiin RAMP 1

Figure 17. Absence of RAMP 1 mRNA in SV40LT-SMC ceiis. RNA isolated from SK-N-MC cells and the rat vas deferens was used to ensure that the RAMP 1 primers worked. Cyclophilin primers were used to verily the integrity of the RNA. The inclusion/exclusion of reverse transcriptase in the reverse transcription reaction is indicated by +/-. Marker bands (M) are 300, 400, and 500 base pair size markers. The RAMP 1 and cyclophilin RT-PCR products contain 413 and 410 base pairs, respectively.

66 rat lung SV40LT-SMC M ^

- -j- -

4 1 4 ------^ 336 -----*

Figure !8 . Identification of R\M P 2 and RAMP 2 B mRNA in SV40LT-SMC cells. Rat lung was used as a positive control for the detection of RAMP 2 mRNA. The inclusion/exclusion of reverse transcriptase in the reverse transcription reaction is indicated by +/-. The 414 base pair product was identified as RAMP 2, and the 336 base pair product was called RAMP 2 B. Marker bands (M) are 300, 400, 500, and 600 base pair size markers.

67 RAMP 2 25 CGCTCCGGGTAGAGCGCGCCCCGGGTGGATCACAGCTCGCTGTGAC H!H!i!!!!!l!HI!!miHHH!l!l!!li RAMP 2 B CGCTCCGGGTAGAGCGCGCCCCGGGTGGATCACAGCT------

CAGCGCCCAGCGGCCCGCAGCGCTCCGCCTCCCTCCGCTGTTACTGCTGCTGTTGCTGC

TGCTGCTGGGCGCTGTCTCAACCTCTCCGGAGTCCCTGAATCAATCTCATCCTACTGAG iiHHHiHumtmmiimmmimiHmn!!! ------CGCTGTCTCAACCTCTCCGGAGTCCCTGAATCAATCTCATCCTACTGAG

GACAGCCTTCTGTCAAAAGGGAAGATGGAGGACTACGAAACAAATGTCCTACCTTGCTG ttH[IIHtmm]H)H!tm]Himt!H!)!)HHI!IHtm!)l! GACAGCCTTCTGTCAAAAGGGAAGATGGAGGACTACGAAACAAATGTCCTACCTTGCTG

GTATTATTACAAGACTTCCATGGACTCTGTCAAGGACTGGTGCAACTGGACTTTGATTA t]!)m]]HIHtHH)!i!tl!H!!!H!lt!!Hmm)l))!ttH)!!! GTATTATTACAAGACTTCCATGGACTCTGTCAAGGACTGGTGCAACTGGACTTTGATTA

GCAGGTATTACAGCAACCTGCGGTATTGCTTGGAGTACGAGGCAGACAAGTTTGGCCTG

GCAGGTATTACAGCAACCTGCGGTATTGCTTGGAGTACGAGGCAGACAAGTTTGGCCTG

GGCTTCCCAAATCCCTTGGCAGAAAGTATCATCCTTGAGGCTCACCTGATACACTTTGC

CAACTGCTCCTTGG 439 HHUHHHH CAACTGCTCCTTGG

Figure 19. Comparison of partial nucleotide sequences of RAMP 2 and RAMP 2 B. Nucleotides 25-439 of rat RAMP 2 (Accession # AF181551) are compared to the RAMP 2 B nucleotide sequence. Dashes in the RAMP 2 B sequence indicate nucleotides that are deleted in this transcript. Underlined nucleotides were used as primers for RT-PCR.

68 SK-N-MC SV40LT-SMC RAMP 3 cyclophilin RAMP 3 cyclophilin M 7" + _ "+ I + ^ i#

^

Figure 20. Identification of RAMP 3 mRNA in SV40LT-SMC cells. The inclusion/exclusion of reverse transcriptase in the reverse transcription reaction is indicated by +/-. The 409 base pair RT-PCR product identified in SV40LT-SMC cells was identified as RAMP 3. Cyclophilin primers were used as a positive control to verify the integrity of the RNA. The absence of RAMP 3 mRNA in SK-N-MC cells is consistent with the work of McLatchie et al. (1998). Marker band (M) are 300, 400, 500, and 600 base pair size markers.

69 A human SK-N-MC lung

M^ *f - -i-

Celt ttnas B

RAMP1

RAMP2

t RAMP3

" ^

Figure 21. Identification of mRNA encoding CRLR, RAMP 1, and RAMP 2 in SK-N-MC cells. CRLR mRNA was identified in SK-N-MC cells by RT-PCR (panel A). Human lung RNA was used as a positive control, and the inclusion/exclusion of reverse transcriptase in the reverse transcription reaction is indicated by +/-. The 1367 base pair RT-PCR product was identified as CRLR, but the sequence of the 600 base pair product was not determined. McLatchie et al. (1998) identified RAMP 1 and RAMP 2 mRNA in SK-N-MC cells by northern blotting (panel B).

70 competition binding was used to compare the adrenomeduHin receptor that is expressed in SK-N-MC cells to the adrenomedullin receptor identified in SV40LT-SMC cells.

Specific binding of [*T]-adrenomedullin was > 70 % in all experiments.

Adrenomedullin inhibited ['^I]-adrenomedu!lin binding with high affinity at a single binding site in experiments using SK-N-MC membranes (Figure 22). In contrast, competition binding with ha-CGRP revealed the presence of two [^Tj-adrenomedullin binding sites that demonstrated a 2886-fold difference in their affinity for ha-CGRP.

71 % of the radioligand bound to the high affinity (IC^ = 0.07 nM) binding site, and

29 % bound to the low affinity (IC^ = 202 nM) binding site. ha-CGRP^y inhibited

['*T]-adrenomedullin binding with 77-fold lower affinity than adrenomeduHin. In

addition, adrenomedullin2 ^ 2 bound to ['^I]-adrenomedullin binding sites with 358-fold

lower affinity than adrenomeduHin and inhibited only 40 % of ['^I]-adrenomedullin

binding. These data suggest that SK-N-MC cells have two ['^I]-adrenomedullin receptors. One receptor had a rank order of ligand affinity that was identical to that of

the adrenomeduHin receptor expressed in SV40LT-SMC cells (adrenomeduHin >

ha-CGRP — ha-CGRPg_37 = adrenomedullin22_s2)- The second receptor had similar

affinities for ha-CGRP and adrenomeduHin but did not bind adrenomedullin22_s2 at

concentrations up to 3 pM (rank order of affinity: ha-CGRP - adrenomedullin >

ha-CGRPg.37). The ICso values for these ligands are listed in Table 7.

SK-N-MC cells also express CGRP, receptors. Previous investigators have found

that ['^I]-adrenomeduHin does not label CGRP, receptors at concentrations up to 120 pM

(Husmann et al., 2000). Therefore, it was unlikely that the second ['*4]-adrcnomcdu!I in

binding site was the CGRP, receptor. However, this possibility was examined by

71 ha-CGRP AdrenomeduHin

ha-CGRPg_3? Adrenoniedui!in22-52

!og (M) [tigartd]

Figure 22. Mean competition curves for selected ligands in SK-N-MC membranes labeled with ['^I]-adrenomedullin. Curves represent the mean of three to six individual experiments each using cells grown in different culture flasks.

72 determining the rank order of affinity for adrenomeduHin, ha-CGRP, and ha-CGRPg^ at

['*H]ha-CGRP binding sites. Nonspecific binding was determined with 1 pM ha-CGRP, and specific binding was > 90 % for each experiment. ha-CGRP bound with the highest affinity of all ligands tested (Figure 23). AdrenomeduHin bound with 15-foid lower affinity, and the CGRP, receptor-selective antagonist, ha-CGRPg.^, bound with 142-fold lower affinity than ha-CGRP. Competition binding curves for each ligand fit best to a single -site binding model. The IC^ values for these ligands are listed m Table 7.

Identification of AdrenomeduHin Receptors in Rat Vas Deferens. The data in the present study demonstrate that ha-CGPPg.g? blocks CGRP-induced cAMP synthesis with a CGRP2 receptor-like affinity (Kg = 4630 nM) in SV40LT-SMC cells. However, these data indicate that the receptor that mediates this response is an adrenomedullin receptor rather than a CGRP receptor. The putative CGRP , receptor was first identified in the rat vas deferens by Dennis et al. (1990). Consequently, this tissue has become a commonly used model for studying the CGRP 2 receptor (Dennis et al., 1990; Mimeault et al.. 1991;

Wisskirchen et al., 1998). In order to determine whether the rat vas deferens expresses adrenomeduHin receptors, ['^I]-adrenomedullin was used to radiolabel adrenomeduHin receptors in membranes prepared from the rat vas deferens. Specific

['^Ij-adrenomedullin binding sites (89 % specific binding) were identified in rat vas

deferens membranes. ha-CGRP, the only unlabeled ligand used for competition binding

in this tissue, inhibited ['^Ij-adrenomedullin binding at a single binding site with an IC^

of 79 nM (Figure 24). These data demonstrate that adrenomeduHin receptors that bind

ha-CGRP are present in the rat vas deferens.

73 tog (M) [iigand]

Figure 23. Mean competition curves for selected ligands in SK-N-MC membranes labeled with ['^I]ha-CGRP. Curves represent the mean of three to live individual experiments each using cells grown in different culture flasks.

74 On r t t t ]----- r* -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 log [ha-CGRP]

Figure 24. Identification of specific ['^I]-adrenomeduHin binding sites in membranes isolated from the rat vas deferens. These data represent a single experiment.

75 The adrenomeduHin receptor formed by the coexpression of CRLR with RAMP 2 is pharmacologically indistinguishable from the receptor that is formed by the coexpression of CRLR with RAMP 3 (McLatchie et al., 1998). Therefore, I used

RT-PCR to determine whether CRLR and RAMPs are present in the rat vas deferens. mRNA encoding CRLR, RAMP 1, RAMP 2, and RAMP 3 were identified in RNA isolated from the rat vas deferens (Figure 25). RAMP 2 B mRNA was also present in this tissue. These data are consistent with the presence of both CGRP, receptors and adrenomedullin receptors in the rat vas deferens.

Discussion

CGRP receptors have been subdivided into CGRP, and CGRP2 receptor subtypes based on their affinity for ha-CGRP 337 in functional studies using isolated tissues.

CGRP, receptor-mediated responses are blocked with high affinity (Kg < 100 nM) by ha-CGRPs_37, and CGRP2 receptor-mediated responses are blocked with low affinity

(Kg > 100) by this antagonist (Dennis et al., 1990). In contrast to functional studies, radioligand binding assays have identified only the CGRP, receptor (Rorabaugh et al.,

2001; Mimeault et al., 1991; Dennis et al., 1990). Although the view that there are two

CGRP receptor subtypes is widely accepted, the observation that ha-CGRP3.37 identifies the putative CGRP2 receptor in functional assays but not in radioligand binding assays has been an enigma.

Functional assays, radioligand binding, and RT-PCR support the conclusion that the putative CGRP2 receptor expressed in SV40LT-SMC cells is an adrenomedullin receptor. The fact that ha-CGRP and adrenomedullin-induced cAMP synthesis are

76 + A CRLR

647

C RAMP 2 / RAMP 2 B 414 ---- **

336 ---- ^

D RAMP 3 409 ---- ^

Figure 25. Identification of CRLR and RAMP mRNA in the rat vas deferens. RT-PCR was used to identify CRLR (panel A), RAMP 1 (panel B), RAMP 2 (panel C), RAMP 2 B (panel C) and RAMP 3 (panel D) mRNA in the rat vas deferens. The inclusion/exclusion of reverse transcriptase in the reverse transcription reaction is indicated by +/-. The size (base pairs) of each RT-PCR product is shown.

77 blocked by ha.-CGRPg.37 with similar affinities indicates that the effects of both agonists are mediated by the same receptor. Furthermore, the relatively high potency and efficacy of adrenomedullin (compared to CGRP) suggests that this response is mediated by an adrenomedullin receptor. Radioligand binding data also support the conclusion that

CGRP and adrenomedullin-induced cAMP synthesis is mediated by an adrenomedullin receptor in SV40LT-SMC cells. The absence of detectable ['^I]ba-CGRP binding sites and the presence of ['^I]-adrenomedullin binding sites in SV40LT-SMC membranes is consistent with the absence of CGRP receptors and the presence of adrenomedullin receptors. The presence of adrenomedullin receptors is also supported by the identification of mRNA encoding CRLR, RAMP 2, and RAMP 3 in SV40LT-SMC ceils.

Thus, several lines of evidence support the conclusion that the putative CGRJP2 receptor that is expressed in these cells is an adrenomedullin receptor.

Although the present study focused on characterizing the putative CGRP2 receptor in SV40LT-SMC cells, these results may also extend to some other systems that have been used to investigate the CGRP2 receptor. For example, Col 29 cells have been used to study the CGRP2 receptor because CGRP-induced cAMP synthesis is blocked with low affinity (K^ > 100 nM) by ha-CGRF^y in these cells (Cox and Tough 1994; Poyner et al., 1998). However, Poyner et al. (1998) reported that Col 29 cells lack detectable

['^1-ha-CGRP] binding sites and that adrenomedullin stimulates cAMP synthesis with at least 100-fold greater potency than CGRP in these cells (Poyner et al., 1999). These results are similar to those observed in the present study, and they are consistent with the conclusion that the putative CGRP, receptor that has been studied in Col 29 cells is an adrenomedullin receptor. The rat and guinea pig vas deferens have also been used to

78 study the CGRP2 receptor. However, the observation that CGRP-induced relaxation of the guinea pig vas deferens is blocked by adrenomedullin^^.^ (Poyner et al., 1999) suggests that this response is mediated by an adrenomedullin receptor. Although the

CGRP, receptor has been identified in the vas deferens by radioligand binding and

RT-PCR (Figure 25; Poyner et al., 1999; Dennis et al., 1990; Mimeault et al., 1991), the low affinity with which ha-CGRP ^ blocks CGRP-induced relaxation of this tissue suggests that the CGRP, receptor does not mediate this response.

Despite evidence that the putative CGRP, receptor is an adrenomedullin receptor in SV40LT-SMC cells, Col 29 cells, and the vas deferens, the involvement of adrenomeduHin receptors in CGRP-induced responses of other tissues is unclear. For example, isolated porcine coronary arteries have been used to study CGRP2 and adrenomedullin receptors (Waugh et al., 1999; Foulkes et al., 1991; Hasbak et al., 2001).

Hasbak et al. (2001) reported that 1 pM adrenomcdulii^z^' an adrenomedullin receptor antagonist, did not block CGRP or adrenomedullin-induced relaxation of porcine coronary arteries. However, radioligand binding and functional studies with this antagonist have demonstrated that adrenomedullin22_32 concentrations in excess of 1 pM are required to block some adrenomedullin receptors (Figure 22; Eguchi et al., 1994).

These data suggest that the adrenom edull^^ concentration used by Hasbak et al.

(2001) was too low to block adrenomeduHin receptors. Antagonists that block

adrenomeduHin receptors with higher affinity than adrenom edull^^ are clearly needed.

Other tissues, including the porcine basilar artery and rat internal anal sphincter muscle

have also been reported to express CGRP2 receptors (Waugh et al., 1997; Wisskirchen

and Marshall, 2000). However, the affect of adrenomeduHin receptor antagonists has not

79 been examined in these tissues. Such experiments would be useful for determining

whether the putative CGRP2 receptor is an adrenomedullin receptor in these model

systems.

Although SK-N-MC cells express both CGRP, and adrenomeduHin receptors

(Figure 21; Table 7), CGRP-induced cAMP synthesis is blocked by ha-CGRP,^ with a

CGRP, receptor-like affinity (Kg < 100 nM) in these cells (Longmore et al., 1994; Poyner

et al., 1998; Edvinson et al., 2001). The vas deferens also expresses both CGRP, and

adrenomedullin receptors (Figure 25; Poyner et al., 1999). However, CGRP-induced

relaxation of this tissue is blocked by ha-CGRPg.37 with a CGRP2 receptor-like affinity

(Kg > 100 nM) (Dennis et al., 1990; Wisskirchen et al., 1998; Mimeault et al., 1991).

This raises an important question concerning tissues that express both CGRP, and

adrenomedullin receptors: why does ha.-CGRPg.37 block CGRP-induced responses with

high affinity in some tissues and with low affinity in other tissues? This issue can be

resolved by the fact that CGRP-induced responses can be mediated by either CGRP,

receptors or adrenomedullin receptors (Fraser et al., 1999). ha-CGRPg.37 blocks CGRP-

induced cAMP synthesis in SK-N-MC cells with high affinity because this response is

mediated by CGRP, receptors. Since CGRP is a more potent agonist at CGRP, receptors

than at adrenomedullin receptors (Fraser et al., 1999), a maximal cAMP response may be

reached at concentrations that are too low to activate adrenomeduHin receptors. In

contrast, CGRP-induced relaxation of the vas deferens is blocked with low affinity

(Kg > 100 nM) by ha-CGRPg.37. The observation that this response is also blocked by

adrenomedullin22_52 (Poyner et al., 1999) suggests that adrenomeduHin receptors (rather

than CGRP, receptors) mediate this response. Thus, CGRP, receptors that are present in

80 a tissue may be uninvoived in mediating some responses that are analyzed in functional assays. Consequently, it is important to choose an appropriate CGRP-induced response when functional assays are used to characterize CGRP receptors.

Radioligand binding experiments revealed the presence of two adrenomedullin receptors in SK-N-MC cells. One receptor had ligand binding characteristics that were identical to the adrenomedullin receptor that was characterized in SV40LT-SMC cells

(rank order of affinity: adrenomedullin > CGRP - CGRP^y - adrcnomeduUm^ ^)- The ligand binding characteristics of the second adrenomedullin receptor (rank order of affinity: adrenomedullin = ha-CGRP > ha-CGRPg,^ > adrenomedullir^^) were distinct from either the first adrenomedullin receptor or the CGRP, receptor (rank order of affinity: ha-CGRP > adrenomedullin > ha-CGRP^). One feature of the second adrenomedullin receptor was its inability to bind adrenomedullin?.^ at concentrations up to 3 pM (Figure 22). This adrenomedulli^m'insensitive receptor has not been previously reported in SK-N-MC cells. However, previous investigators have found that adrenomedullin-induced responses in cultured rat cardiac cells and in the vascular bed of the cat hindlimb are not blocked by adrenomeduHin^^ (Nishikimi et al., 1998;

Champion et al., 1997). In contrast, adrenomedullin-induced responses are blocked by this antagonist in rabbit aortic endothelial cells and rat cerebral arteries (Muff et al.,

1998; Dogan et al., 1997). These data support the conclusion that there are 2 different adrenomedullin receptor subtypes that can be distinguished by their affinity for adrenomedullin22.32- SK-N-MC cells express both adrenomedullin receptor subtypes and may be useful for the development of additional ligands that discriminate between these receptors.

81 McLatchie et ah (1998) were first to report that CRLR forms an adrenomeduHin receptor when coexpressed with RAMP 2. RAMP 2 is characterized by a long (136 amino acids) extracellular amino terminal, a single transmembrane domain (21 amino acids), and a relatively short (9 amino acids) intracellular carboxyl terminal. Studies using chimeric RAMPs have demonstrated that the region of this accessory protein that determines the ligand selectivity of CRLR is located in the extracellular domain (Fraser et al., 1999). Amino acid residues Trp^-Pro'^ in RAMP 2 have been specifically identified as residues that influence the ligand selectivity of CRLR (Kuwasako et al.,

2001). In the present investigation, I identified RAMP 2 B mRNA. This transcript was identical to the RAMP 2 transcript except that the nucleotides encoding Ala'^-Gly^ were deleted. The significance of these specific residues has not been investigated. However, the lack of ['^I]ha-CGRP binding sites in SV40LT-SMC cells suggests that RAMP 2 B does not lead to the formation of a CGRP receptor, f urthermore, the fact that each competition curve in SV40LT-SMC membranes fit best to a single site binding model suggests either that the receptor formed by the coexpression of CRLR with RAMP 2 B does not bind ['^!]-adrenomedullin or that the ligands used in this investigation do not discriminate between this receptor and the adrenomedullin receptors that are formed by the coexpression of CRLR with RAMP 2 or RAMP 3. Alternatively, it is also possible that the RAMP 2 B transcript is not translated into protein. Further work is needed to determine whether RAMP 2 B mRNA is translated and whether this putative protein has functional significance.

In summary, the data in the present investigation demonstrate that SV40LT-SMC

cells express an adrenomeduHin receptor that is activated by CGRP and blocked with a

82 CGRP2 receptor-iike affinity by ha-CGRP^.^. These data support the conclusion that the

putative CGRP2 receptor is an adrenomedullin receptor. There is also evidence that the

CGRP2 receptor that has been studied in Col 29 cells and the vas deferens is an

adrenomedullin receptor. However, further experiments are needed to determine whether

this conclusion is valid for porcine coronary arteries and other tissues that have been

reported to express the putative CGRP2 receptor. This investigation also revealed that

SK-N-MC cells express two different adrenomedullin receptors and that one of these

receptors is insensitive to adrenom edull^^- RAMP 2 B mRNA was also identified m

this study. The functional significance of this transcript is currently unknown.

83 CHAPTER 4 TISSUE-DEPENDENT FACTORS THAT INFLUENCE THE

AFFINITY OF ha-CGRP^

Introduction

The previous chapter provided evidence that the putative CGRP2 receptor that is expressed in SV40LT-SMC cells is an adrenomeduHin receptor. This adrenomedullin receptor is activated by CGRP and blocked with low affinity (K^ = 4630 nM) by ha-CGRPg.37 (Figure 13). Furthermore, this receptor has the same radioligand binding properties as one of the adrenomeduHin receptors that is present in SK-N-MC cells

(Table 7). Evidence was also provided to support the conclusion that the putative CGRP2

receptor that has been studied in Col 29 cells and the rat vas deferens is also an

adrenomeduHin receptor. Thus, CGRP-induced activation of both CGRP, and

adrenomeduHin receptors provides one explanation for the variable affinity of ha-CGRPg.37 in functional assays.

Despite evidence that CGRP activates both CGRP, and adrenomeduHin receptors,

additional tissue dependent factors may also contribute to the variable affinity of ha-CGRPg.37. I have conducted an extensive literature search to identify ha-CGRPg.37

affinity values that have been determined by functional assays and by radioligand

binding in a variety of tissues (Figure 26 A). Each of the 42 K, values determined by

radioligand binding was less than 100 nM (pK, > 7). This is consistent with the

conclusion that CGRP receptor subtypes cannot be identified by radioligand binding

assays using ['^I]ha-CGRP. In contrast, functional assays with ha-CGRPg.37

84 A

vas deferens -V—V ------—*—** ------atrium coronary artery SK-N-MC CGRP, clone rat brain ^ * L 6 myocytes ventricle * iung spieen renai cortex renai medulla iiver 5 6 7 8 9 10 Affinity (pK.,) B

vas deferens coronary arteries Coi 29 ceMs basilar artery aorta * SV4Q-LT-SMC cells internal anal sphincter atrium SK-N-MC cells L 6 myocytes mesenteric artery pulmonary artery glomerular mesengial cells ventricular cardiomyocytes pancreatic acinar cells cerebral artery 5 6 7 * 8 9 10 Affinity (pA2/pKg)

Figure 26. Distribution of affinity values reported for ha-CGRPg.37. Affinity values determined for ha-CGRPg.gy using functional assays (panel A) or radioligand binding (panel B) were obtained through a comprehensive literature search. PubMed was used to identify as many published affinity values as possible for ha-CGRPg 37 regardless of the tissue type or species. Affinity values were obtained from Maggi et al. (1991), Wu et al., (2000), Boulanger et al., (1996/, Longmore et al. (1994), Giuliani et al. (1992), Mimeault et al. (1991), Dennis et al. (1990), Zimmerman et al. (1995/, Poyner (1993), Poyner et al. (1998), Edvinson et al. (2001), Wang et al. (2001), Semark et al. (1992), Muff et al. (1992), Waugh et al. (1991), Wisskirchen et al. (1998), Wisskirchen et al. (1999b)*, Wisskirchen et al. (2000), Foulkes et al. (1991), Yoshimoto et al. (1998), Lam (2000), Sheykhzade et al. (1998), Cox (1995), Lei et al. (1994), and Bell and McDermott (1994). *exact value unknown, but < 6.0. ^ pfQ value calculated by the Schild equation (pKg = log [antagonist concentration / (dose ratio - 1 )] based upon EC^ values reported in the presence and absence of ha-CGRPg.37.

85 demonstrated both high (pKLg > 7) and low (pKg < 7) affinities for this antagonist (Figure

26 3). Furthermore, the 54 Kg values determined in functional studies were not distributed into two distinct groups as predicted by the view that CGRP-induced responses are mediated by two different types of receptors. Rather, these values were scattered over a broad range of affinities that spans 4 orders of magnitude (pKg = 5.3 -

9.3) (Figure 26 B). Furthermore, the affinity of ha-CGRPg.37 was highly variable within individual tissues. These data suggest that factors other than CGRP-induced activation of both CGRP, and adrenomedullin receptors contribute to the variable affinity of ha-CGRPg.37 The goals of this study were to identify tissue-dependent factors that may influence the affinity of ha-CGRPg.37 in functional assays

Enzymatic Degradation of ha-CGRPg 3 7

The biological actions of neuropeptides are terminated by peptidases that metabolize them into biologically inactive products. Some investigators have proposed that the variable affinity with which ha-CGRPg.37 blocks CGRP-induced responses reflects various levels of enzymatic degradation of the antagonist in different tissues rather than the existence of multiple CGRP receptor subtypes (Longmore et al., 1994).

The affinity of ha-CGRPg_37 is measured in functional studies by equilibrating this antagonist with an isolated tissue for 5 to 90 minutes before a CGRP-induced response is measured (Hasbak et al., 2001; Sams et al., 2000; Waugh et al., 1999). Equilibration is necessary for accurate affinity measurements. However, this also provides an opportunity for ha-CGRPg.37 to be degraded by enzymes that are present in the tissue.

Neutral endopeptidase (EC 3.4.24.11) and matrix metalloprotease-2 (MMP-2)

86 (EC 3.4.24.24) are enzymes that have been demonstrated to degrade CGRP m vitro

(Figure 27). These enzymes are present on the extracellular surface of neurons, smooth muscle cells, and endothelial cells (Mentlein and Roos, 1996; Puyraimond et al., 2001).

Since cleav age sites of neutral endopeptidase and MMP-2 are located within the ha-CGRPg.37 region of ha-CGRP, it is likely that these enzymes also degrade ha-CGRPg_3?. Consequently, the concentration of ha-CGRPg_37 may be decreased in tissues that contain large amounts of neutral endopeptidase or MMP-2 activity. As a result, ha-CGRPg_37 may block CGRP-induced responses in these tissues with decreased affinity.

This hypothesis is supported by functional assays in which tissues have been preincubated with ha-CGRPg.37 for various times before generating a dose-response curve for CGRP. An analysis of pA2 and pK„ values reported for ha-CGRPg.37 01 tissues commonly used to study the CGRP2 receptor (rat vas deferens, guinea pig vas deferens, and pig coronary artery) revealed that the affinity of this antagonist decreased more than

1.5 orders of magnitude with ha-CGRPg.37 preincubation times ranging from 5 to 60 minutes (Figure 28 A). Furthermore, linear regression analysis of these data revealed that the slope of the regression line was negative and significantly different from zero

(p - 0.0074). in contrast, the affinity of ha-CGRPg 37 changed only slightly with increasing preincubation times in tissues commonly used to study the CGRP, receptor

(guinea pig atrium, rat pulmonary artery, rat mesenteric artery, SK-N-MC cells, porcine

CGRP, receptor-transfected F1EK 293 cells), and the slope of this linear regression line was not significantly different from 0 (p = 0.25) (Figure 28 B). Furthermore, the slope of

87 Figure 27. Proteolytic cleavage sites in ha-CGRP. Neutral endopeptidase (EC 3.4.24.11) and matrix metalloprotease-2 (EC 3.4.24.24) hydrolyze several peptide bonds in CGRP m vitro assays (Katayama et al., 1991; Fernandez Patron et at., 2000; Ruchon et al., 2000). The amino acid sequence of CGRP is represented by circles containing the single letter abbreviation for each amino acid residue. The line between C^ and C^ represents a disulfide bond, and the carboxyl terminal amide is indicated on F".

88 A B

Figure 28. Time-dependent decrease in the affinity of ha-CGRPg.37 in functional assays using isolated tissues. pKg values and their corresponding preincubation times were obtained through a comprehensive literature search for tissues commonly used as models for the putative CGRP2 receptor (panel A) or the putative CGRP, receptor (panel B). Fifteen pKg values were obtained for prototypical CGRP, receptor tissues (rat pulmonary artery, guinea pig atrium, rat mesenteric artery, SK-N-MC cells, and HEK 293 cells stably transfected with the pig CGRP, receptor), and 23 pK„ values were obtained for prototypical CGRP2 receptor tissues (rat vas deferens, guinea pig vas deferens, and pig coronary artery) at the ha-CGRPg.37 preincubation times shown. Mean (+ S.E.M.) pKg values for each preincubation time were plotted and analyzed by linear regression (dashed lines). The slope of the regression line (-0.010) was not significantly different from 0 (p = 0.2549) for prototypical CGRP, receptor tissues. The slope of the regression line (-0.033) was significantly different from 0 (p - 0.0074) in prototypical CGRP2 receptor tissues. pKg values for this analysis were obtained from Hasbak et al. (2001), Wu et al. (2000), Mimeault et al. (1991), Longmore et al. (1994), Dennis et al. (1990), Wisskirchen et al. (1998), Wisskirchen et al. (1999a), Wisskirchen et al. (1999b), Wisskirchen et al. (2000), Foulkes et al. (1991), Sams et al. (2000), Maggi et al. (1991), Boulanger et al. (1996), Giulaini et al. (1992), Tomlinson and Poyner (1996), Lei et al. (1994), Edvinson et al. (2001), and Rorabaugh et al. (unpublished data). Ail studies that did not report ha-CGRP3.37 preincubation times were excluded from this analysis.

89 the regression iine was 3.3-fold greater in tissues commonly used to study the putative

CGRP2 receptor than in tissues used to investigate the CGRP, receptor. Thus, the affinity of ho.-CGRPg_37 decreases more rapidly in tissues expressing the putative CGRP2 receptor than in tissues expressing the CGRP, receptor.

The effect of peptidase inhibitors on the affinity of ha-CGRPg.37 has been studied with mixed results. Longmore et al. (1994) reported that ha-CGRPg.37 blocked CGRP- induced relaxation with higher affinity (pA2 - 6 .6 + 0.07) in the presence of 1 pM thiorphan than in the absence of this peptidase inhibitor (pA2 < 6 ). Other investigators have reported that amastatin, bestatin, captopril, phosphoramide, and thiorphan have no effect on the affinity of ha-CGRPg.37 for the putative CGRP2 receptor in the rat vas deferens, porcine coronary arteries, and other tissues (Yoshiomoto et al., 1998;

Wisskirchen and Marshall, 2000; Wisskirchen et al., 1998). These compounds inhibit aminopeptidase and neutral endopeptidase (Chakir et al., 1995). However, they have not been found to inhibit MMP-2, an enzyme that cleaves the Gly^-Lcu' peptide bond in ha-CGRPg.37 (Figure 27). Therefore, MMP-2, or other enzymes that are not inhibited by these compounds, may degrade ha-CGRP3.37 and cause it to block CGRP-induced responses with lower affinity in tissues containing large amounts of MMP-2 activity than in tissues that lack this enzyme.

The influence of peptidases on the affinity of ha-CGRPg.37 Fas also been investigated by chemically modifying ha-CGRPg.37 so that it is resistant to aminopeptidase. Smith et al. (2001) reported that acetylation of ha-CGRPg.37, a modification previously shown to protect peptides from aminopeptidase (Drapeau et al.,

1993), caused this antagonist to block CGRP-induced relaxation of porcine coronary

90 arteries with 33-fo!d higher affinity. Similarly, benzoylation of ha-CGRPg.37 increased the affinity of this antagonist 24-fold (Smith et al., 2001). Although these modifications

could increase the affinity of ha-CGRPg^ by directly altering the interaction of this

peptide with the binding pocket of the CGRP receptor, these data are also consistent with

the conclusion that aminopeptidase degrades ha-CGRPg.g? during functional assays with

this tissue. Three nonpeptide CGRP receptor antagonists (compound 1, SB-273779, and

B1BN4096BS) have recently been reported (Edvinsson et al., 2001; Aiyar et al., 2001a;

Doods et al., 2000). The use of these and other nonpeptide antagonists will be useful for

studying CGRP receptors while avoiding proteolytic degradation.

In contrast to functional studies with isolated tissues, ha-CGRPg.37 does not have

low affinity for the putative CGRPi receptor in radioligand binding assays. This may be

caused by decreased protease activity m membrane preparations compared to whole

tissues. Ethylenediaminetetraacetic acid (EDTA) and ethylene glycol tetraacetic acid

(EGTA) are often used in radioligand binding studies (Rorabaugh et al., 2001; Dang et

al., 1999; McLatchie et al., 1998; Elshourbagy et al., 1998). These compounds chelate

metal ions and inhibit metalloenzymes such as MMP-2 and neutral endopeptidase. In

addition, radioligand binding assays are often performed in the presence of bovine serum

albumin (Rorabaugh et al., 2001; Dennis et al., 1990; Mimeault et al., 1991; Poyner et al.,

1999). Bovine serum albumin may protect ha-C G R P g g? from proteolytic degradation by

acting as a substrate that saturates the activity of proteolytic enzymes that would

otherwise degrade ha-CGRPg.37. Consequently, the chemical environment in which

radioligand binding assays are performed may protect ha-CGRPg.37 &om proteolytic

degradation.

91 Release of Endogenous CGRP

Endogenous C G R P is released from capsaicin-sensitive neurons that innervate

Mood vessels and other tissues (Kruger et al., 1989). Waugh et al. (1997) proposed that the release of endogenous CGRP from these neurons may increase the CGRP concentration at the CGRP receptor and cause the affinity of ha-CGRPg.37 to be underestimated in functional assays. This is particularly a concern in the vas deferens where a commonly measured response to CGRP is inhibition of the field stimulated twitch response. Field stimulation causes the vas deferens to contract by depolarizing smooth muscle cells. However, field stimulation may also depolarize neurons and confound measurements of ha-CGRPg.37 affinity by stimulating the release of endogenous CGRP from neuronal storage vesicles.

The impact of endogenous CGRP on the affinity of ha-CGRPg.37 has been

addressed in pig and rat tissue models of the putative CGRP, receptor. Wisskirchen et al.

(1998) reported that ha-CGRPg.37 blocks CGRP-induced relaxation of the vas deferens with low affinity regardless of whether the tissue is contracted by field stimulation or by norepinephrine (pA2 - 6.0 and 5.8 for field stimulated and norepinephrine stimulated vas

deferens, respectively). This suggests that field stimulation does not cause the affinity of

ha-CGRPg_37 to be underestimated in this tissue. In contrast, Waugh et al. (1999)

reported that CGRP-induced relaxation of porcine coronary arteries was blocked by

ha-CGRPg.37 with 5-fold greater affinity if endogenous CGRP was depleted by

pretreating the arteries with capsaicin. This suggests that the release of endogenous

CGRP may contribute to the variable affinity of ha-CGRPg_37 that is observed in

functional assays using coronary arteries. However, the magnitude of this effect is

92 insufficient to account for the large variation in ha-CGRPg.37 affinities that is presented in

Figure 26 B. The influence of endogenous CGRP may be tissue-dependent with a greater effect in tissues that are extensively innervated by neurons that contain CGRP.

Conformational States of the CGRP Receptor

The CGRP2 receptor is not the only receptor that has been identified in functional

assays but not by radioligand binding. Similar results have also been observed in tissues

expressing the putative a,^ adrenergic receptor. Three subtypes of a,-adrenergic

receptors (a,A, a,g, and a,„) have been cloned and characterized (Bylund et al., 1994). A

fourth subtype, the putative a,,_ adrenergic receptor, has been identified in functional

assays but has not been cloned or identified by radioligand binding (Flavahan and

Vanhoutte, 1986; Muramatsu et al., 1990; Ford et af, 1997). Ford et al. (1997) used

radioligand binding and functional assays to measure the affinity of several antagonists

for the a,A adrenergic receptor expressed in CHO cells. Radioligand binding yielded

affinities that were similar to previously published affinities for the a,A adrenergic

receptor. In contrast, affinities determined in functional assays were similar to those

reported for the a,^ adrenergic receptor. Ford et al., (1997) proposed that the a,A and a,^

adrenoceptors represent a single protein that adopts a conformation in isolated

membranes that is different from its conformation in intact cells. These conformations

can be distinguished by antagonists (prazosin, RS 17053, WB 4101, 5-methylurapidil,

and S-niguldipine) that discriminate between the a,A adrenergic receptor and the putative

a,^ adrenergic receptor.

93 In light of the work of Ford et al. (1997), I speculate that the conformation of the

CGRP, receptor may also be influenced by the isolation of membranes from intact tissues. Perhaps some tissues contain a factor that stabilizes the CGRP receptor in a conformation that has a low, CGRP2 receptor-like, affinity for ha-CGRPg,^. The loss of this factor during the membrane preparation procedure could cause CGRP receptors to have a high, CGRP, receptor-like, affinity for ha-CGRPg.37 in radioligand binding assays.

Alternatively, the conformation of the CGRP, receptor could be influenced by the use of different buffers. Deupree et al. (1996) reported that the affinity of some 0.3 adrenergic receptor antagonists is dependent upon the buffer that is used in radioligand binding assays. In addition, Emsberger and UTrichard (1987) reported that sodium increased the affinity of prazosin and WB-4101 for a, adrenergic receptors. The influence of buffer composition on the affinity of CGRP, receptor ligands has not been systematically examined. However, in preliminary studies 1 found that ha-CGRP^y and

[Cys(ACM)* ^]ha-CGRP inhibited ['^I]ha-CGRP binding to the cloned porcine CGRP, receptor with 27-fbld and 8 -fold lower affinity, respectively, when I used the radioligand binding protocol of Elshourbagy et al. (1998) rather than the protocol of Abel et al.

(1997). The composition of the buffers used in these assays differed with respect to the presence of HEPES buffer, Tris buffer, NaCl, EGT A, and bovine serum albumin (Table

8 ). Thus, it is possible that the conformation of ha-CGRPg^y or the CGRP, receptor is influenced by buffer components in a way that changes the binding affinity of this ligand for the CGRP, receptor.

94 Tabie 8 . Comparison of buffers known to influence the affinity (K,) of^ ha-CGRPg^. g.37.______Radioligand Binding Radioligand Binding Protocol of Eishourbagy Protocoi of et a!. (1998)" Abel et al., (1997)" Ligand Affinities (K,) ha-CGRPg.^ 0.21 nM 5.7+ 1.7 nM [Cys(ACM)^]ha-CGRP 4.2 nM 32.0 + 5.8 nM

Radioligand Binding Buffer 20 mM HEPES, pH 7.4 50 mM Tris, pH 7.4 Composition 5 mM MgCl^ 5 mM MgCl; 2 mM EGTA 100 mMNaCl 0 .1% bacitracin 0 . 1% bacitracin 0 .2 % bovine serum albumin Radioligand binding was performed using membranes isolated from HER 293 cells stably transfected with the porcine CGRP,. "Radioligand binding was performed using the membrane preparation and radioligand binding protocols described by Elshourbagy et al. (1998). Briefly, culture medium was removed from the cells, and the cells were detached from the culture flasks with phosphate buffered saline containing 1 mM Na^EDTA. The cells were centrifuged at 300 x g and stored at - 70°C. Cells were resuspended and homogenized in ice cold Tris buffer (10 mM Tris-HCl (pH = 7.4), 5 mM Na^EDTA, 0.1 mM phenylmethylsulfonylfluoride, 1 mg/ml bacitracin, and 0.1 mg/ml aprotmin). The homogenate was centrifuged at 47,000 x g for 20 minutes at 4°C. The resulting pellet was washed twice by homogenization and centrifugation in a second buffer containing 20 mM HEPES (pH - 7.4), 5 mM MgCl^, 2 mM Na„EGTA, and 0.1 mg/ml bacitracin. Radioligand binding experiments were performed in the buffer shown in the table above. Membrane protein (50 - 100 pg/ml) was incubated with 125 - 150 pM ['^I]ha-CGRP and 1 pM - 1 pM competing ligand in a total volume of 500 pi for 30 minutes at 25°C. Bound and free radioligand were separated by washing the membranes with 2 ml of 0.9 % NaCl using a Brandel MB-48R cell harvester. "Radioligand binding was performed using the membrane preparation and radioligand binding protocols described by Abel et al. (1997). This procedure is described in detail in chapter 2 under the subheadings Membrane Preparation and Competition Binding Assay.

95 Conclusion

In conctusion, ha-CGRPg.37 blocks CGRP-induced responses with a broad range of affinities. CGRP-induced stimulation of CGRP, and adrenomedullin receptors provides one explanation for the variable affinity of this antagonist. However, the distribution of ha-CGRP3.3? affinities that have been reported in the literature suggests that other factors may also be involved. The release of endogenous CGRP, enzymatic degradation of ha-CGRPg.37, changes in CGRP receptor conformation, or a combination of these tissue-dependent factors may influence the ability of ha-CGRPg.37 to block

CGRP-induced responses in isolated tissues.

96 CHAPTER 5

COMPARISON OF LIGAND IMNDING CHARACTERISTICS OF PORCINE

AND HUMAN CGRP, RECEPTORS

!ntroduction

identification of amino acid residues that form the binding pocket of the CGRP, receptor could provide useful information for the development of new ligands that are more selective or bind with higher affinity than those that are currently available. Studies using mutant forms of CRLR would provide an ideal system to identify amino acid residues in the CGRP, receptor that are directly involved in ligand binding. However, others and I have found this receptor to be difficult to express (Jim Porter, personal communication). In addition, the identification of receptor activity modifying proteins that influence ligand binding to CRLR adds an additional level of complexity to this task

(McLatchie et al., 1998).

An alternative to expressing mutant forms of CRLR is to use cells that express

CGRP, receptors from different animal species. This method limits the conclusions that can be drawn regarding which amino acid residues are responsible for differences that are observed in ligand binding because each residue cannot be systematically investigated.

However, it does avoid the difficulties associated with expressing wild type and mutant

CGRP, receptors and may provide some preliminary information. The present study took advantage of the fact that human and porcine CRLR share 93 % amino acid sequence identity, and 17 of the amino acid differences between these species are located on the

97 extracellular surface of the cell membrane where they could directly interact with ligands

(Figure 29). Radioligand binding experiments were conducted on membranes from

SK-N-MC cells that endogenously express human CRLR and human HEK 293 cells that have been transfected with porcine CRLR. The goal of this work was to determine whether differences in the ligand binding characteristics of these receptors could be detected using a series of peptide ligands.

Materials and Methods

Cell Culture, Membrane Preparation, and Radioligand Binding. The methods used for cell culturing, membrane preparation, and radioligand binding are described in chapters 2 and 3. No deviations were made from these protocols.

Data analysis. Three to five competition binding curves were performed in duplicate for each ligand. Specific binding was determined by subtracting nonspecific binding (defined using 1 pM ha-CGRP) from total binding, and the IC^ for each competition curve was determined by nonlinear regression analysis using GraphPad

Prism. Data were fit to one site and two site binding models and the best fit model was determined using an F test. K, values were calculated by the Cheng-Prusoff equation using the Kp value (40 pM) determined for ['^l]ha-CGRP for porcine CGRP, receptor- transfected HEK 293 cells. Mean pK, values for each ligand were compared using a two-tailed student's t-test to determine if the pK, values in HEK 293 cells were significantly different (p < 0.05) from the pK, values in SK-N-MC cells.

98 Figure 29. Comparison of human and porcine CRLR amino acid sequences. Differences between the porcine and human CRLR amino acid sequences are shown using standard amino acid abbreviations. The first and second letters designate the residues present in porcine and human CRLR, respectively. The dash at position 25 indicates that human CRLR does not have an amino acid residue that corresponds to this position.

99 Results

Ligand Binding to CGRP, Receptors in Human SK-N-MC Cells. Specific binding was > 90 % in ail experiments using SK-N-MC membranes, and maximal inhibition of

['^I]ha-CGRP binding for each ligand was not different from the maximal inhibition caused by 1 pM ha-CGRP. ha-CGRP displaced ['^Ijha-CGRP from SK-N-MC membranes with the highest affinity of all ligands used in this study (Figure 23; Table 9).

AdrenomeduHin displaced the radioligand with 15-fold lower affinity than ha-CGRP,

and calcitonin did not displace ['^l]ha-CGRP at all. [Pro'^]ha-CGRP and

[Cys(ACM^ )ha-CGRP, putative CGRP2 receptor-selective agonists, bound to SK-N-MC

membranes with 65-fold and 707-fold lower affinity than ha-CGRP, respectively. CGRP

receptor antagonists were also used in this study. ha-CGRP^.^ displaced the radioligand

with a K, of 8.5 nM, and the affinity of this ligand was increased 3-fold by benzoylation

of the amino terminal ([benzoyljha-CGRPg^y). [P^F^Jha-CGIH^y displaced

['^I]ha-CGRP with the lowest affinity of all antagonists used in this investigation.

Competition curves for each ligand fit best to a single site binding model.

Comparison of Ligand Affinities for CGRP Receptors in Human SK-N-MC Cells

and Porcine CGRP, Receptor-Transfected HEK 293 Cells . Table 9 shows a comparison

of the affinity determined for each ligand in human SK-N-MC cells to its affinity

determined for porcine CGRP, receptor-transfected HEK 293 cells as reported in chapter

2. Benzoylation of ha-CGRPg.37 ([benzoyljha-CGRP^.^y) caused a 79-fold increase in the

affinity of this ligand (p < 0.0001 comparing the affinities of ha-CGRP3.37 and

[benzoyl]ha-CGRPg_37 in porcine CGRP, receptor-transfected HEK 293 cells) for the

100 Table 9. Comparison of human and porcine CGRP, receptor binding characteristics. PORCINE CGRP, HUMAN RECEPTOR-TRANSFECTED SK-N-MC CELLS HEK 293 CELLS ha-CGRP 0.06 + .007 0.08 + . 0 1 Adrenomedullin 0.90+ .12" 19.9+12.9 [Benzoyl]ha-CGRP 2.69+1.95 0.072 + .006 [Pro"]ha-CCRP 3.9 + 1.3 5.5 + 1.3 ha-CGRPg^ 8.5+ 2 .6 5.7+ 1.7 [Cys(ACM)^]ha-CGRP 42.4 + 6.0 32.0 + 5.8 [P^F^ha-CGRP^ 109.6 + 24.7 118.2 + 40.2 Calcitonin no binding no binding Each value represents the mean + S.E.M of three to Eve separate membrane preparations, each from a different group of cells. K, values listed for porcine CGRP, receptor- transfected HEK 293 cells are from Table 3. "The K, of this ligand is significantly different (p < 0.05) from its K, determined in HEK 293 cells expressing the porcine CGRP, receptor.

101 porcine CGRP, receptor but did not significantly influence the affinity of ha-CGRP^y for human SK-N-MC membranes (p > 0.05). Adrenomedullin bound to SK-N-MC membranes with 22-fold higher affinity than to membranes from porcine CGRP, receptor-transfected HEK 293 cells. No significant differences were detected for other ligands when comparing their affinities for SK-N-MC membranes to their affinities for porcine CGRP, receptor-transfected HEK 293 membranes.

Discussion

Previous radioligand binding studies using adrenomedullin and a nonpeptide

CGRP, receptor antagonist, SB273779, found that human and porcine CGRP, receptors have different ligand binding characteristics when transfected into human HEK 293 cells

(Aiyar et al., 2001a; Aiyar et al. 2001b). Aiyar et al. (2001b) reported that adrenomedullin binds to the porcine CGRP, receptor with 22-fold higher affinity than to the human CGRP, receptor when these receptors are expressed in HEK 293 cells

(K,= 148.98 nM and 6.79 nM for the human and porcine CGRP, receptor, respectively).

In contrast, I found that the affinity of adrenomedullin was 22-fold higher for the human

CGRP, receptor than for the porcine CGRP, receptor. In addition, benzoylation of ha-CGRPg.37 ([benzoyljha.-CGRPg.37) had a much greater effect on the affinity of this antagonist for the porcine than for the human CGRP, receptor. One difference between my study and that of Aiyar et al. (2001b) was that I used human SK-N-MC cells that endogenously express human CGRP, receptors and HEK 293 cells that were transfected with the porcine CGRP, receptor. Thus, my affinity values could have been influenced

102 by the use of two different cell types. SK-N-MC ceiis also express adrenomeduHin receptors (McLatchie et ah, 1998). However, Coppock et al. (1999) demonstrated that

55 pM ['^I]ha-CGRP does not radiolabel adrenomedullin receptors. Therefore, it is unlikely that adrenomeduHin receptors were radiolabled by 40 pM ['^I]ha-CGRP in my experiments. Although the affinities determined for adrenomeduHin in this study are not. identical to those reported by Aiyar et al. (2001b), both studies support the conclusion that human and porcine CGRP, receptors have detectable differences in their ligand binding characteristics.

A comparison of the porcine and human CRLR amino acid sequences revealed 17 differences in the extracellular regions of this protein. Most of these substitutions are conservative. For example, residue 64 is an isoleucine (nonpolar, hydrophobic) in porcine CRLR and a valine (nonpolar, hydrophobic) in human CRLR (Figure 29).

Likewise, residue 30 is a glutamate (acidic) in porcine CRLR and an aspartate (acidic) in human CRLR. However, porcine CRLR contains an acidic side chain (glutamate^) that is not present in human CRLR, and human CRLR contains a basic side chain (lysine'^) that is not present in porcine CRLR. The differences in ligand binding characteristics discovered in this study and by Aiyar (Aiyar et al., 2001a; Aiyar et al., 2001b) suggest that one or more of these 17 amino acid residues may be involved in ligand binding.

In conclusion, the experimental approach used in this study cannot replace the use

of mutant forms of CRLR for identifying amino acid residues that are involved in ligand

binding. However, this study has identified 17 amino acid residues in CRLR that may

interact with adrenomeduHin (Table 9; Aiyar et al., 2001b), [benzoyljha-CGRP^y

(Table 9), and SB 273779 (Aiyar et al., 2001a).

103 CHAPTER 6

SUMMARY AND IMPLICATIONS FOR THE DEVELOPMENT OF

THERAPEUTIC AGENTS

Summary

The overall goal of my research was to characterize the putative CGRP2 receptor.

1 was especially interested in understanding why the CGRP, receptor demonstrates low affinity for ha.-CGRP3.37 in functional assays but not in radioligand binding assays using

['^Ijha-CGRP This was accomplished using pharmacological and molecular approaches to compare the putative CGRP, receptor that has been identified in porcine coronary arteries, SV40LT-SMC cells, and other tissues to the cloned CGRP, receptor and to the adrenomedullin receptor that is endogenously expressed in SK-N-MC cells.

In Chapter 2, a variety of CGRP analogs were used to compare the binding characteristics of the putative CGRP, receptor that has been identified m porcine coronary arteries (Foulkes et al., 1991; Waugh et al., 1997) to the porcine CGRP, receptor that has been cloned and expressed in HEK 293 cells. Radioligand binding experiments with 12 different ligands demonstrated that the CGRP receptor in porcine coronary arteries had the same ligand binding characteristics as the cloned CGRP, receptor. mRNA encoding CRLR and RAMP 1, the proteins required to form the CGRP,

receptor, were also identified in this tissue. Furthermore, correlation plots that compared

antagonist affinities determined by radioligand binding and by functional assays

demonstrated that CGRP-induced relaxation of porcine coronary arteries is mediated by

104 the CGRP, receptor despite the fact that this response is blocked with low affinity by ha-CGRPg.37.

CGRP, and adrenomeduHin receptors are structurally similar (McLatchie et al.,

1998). Furthermore, previous investigators have reported that adrenomeduHin receptors can be stimulated by ha-CGRP (Fraser et al., 1999). Therefore, the objective of Chapter

3 was to determine whether some CGRP-induced responses that are blocked with low

affinity (Kg > 100 nM) by ha-CGRPg.37 could be mediated by adrenomeduHin receptors.

Radioligand binding, RT-PCR, and functional assays were used to investigate the

putative CGRP, receptor identified in SV40LT-SMC cells. CGRP and adrenomedullin-

induced cAMP synthesis was blocked with a CGRP2 receptor-like affinity by

ha-CGRPg.37, and the relative potency of these agonists suggested that this response was

mediated by an adrenomeduHin receptor. This conclusion was also supported by the

identification of [^Tj-adrenomedullin binding sites but not ['^I]ha-CGRP binding sites

in SV40LT-SMC membranes and by the identification of mRNA encoding CRLR,

RAMP 2, and RAMP 3 in these cells. The ligand binding characteristics of the putative

CGRP, receptor in SV40LT-SMC cells were identical to an adrenomeduHin receptor that

is endogenously expressed in SK-N-MC cells.

I anticipate that the most significant impact of my work will be its implications

for the classification of CGRP receptor subtypes. It is widely accepted that CGRP-

mduced responses are mediated by CGRP, and CGRP2 receptors (Figure 30 A). The

CGRP, receptor and its accessory proteins have been cloned and characterized at the

molecular level. However, prior to the present studies, the CGRP2 receptor had not been

well characterized. My data support the conclusion that the putative CGRP2 receptor that

105 A Receptors Stimulated by CGRP

CGRP, CGRP. Receptor Receptor

Moiecuiar CRLR + ? Composition RAMP 1

high low ha-CGRP g_37 (K.B< 100 nM) (Ke > 100 nM) affinity in functional assays

['"I]ha-CGRP binds ['^I]ha-CGRP does not bind ['"l]ha-CGRP binding

B Receptors Stimulated by CGRP

CGRP, AdrenomeduHin Receptor Receptor

Molecular CRLR + CRLR + RAMP 2 Composition RAMP 1 CRLR + RAMP 3

high low ha-CGRPg.^ (IQ < 100 nM) (IQ> 100 nM) affinity in functional assays

['"I]ha-CGRP binds ['^I]ha-CGRP does not bind ['^I]ha-CGRP binding

Figure 30. Impact of this work on the classification of receptors that are stimulated by CGRP. The currently accepted view is that CGRP-induced responses are mediated by CGRP, and CGRP 2 receptor subtypes (panel A). My data indicate that the putative CGRP2 receptor is an adrenomedullin receptor that has low affinity for ha-CGRPg.37 (panel B).

106 has been identified in SV40LT-SMC cells, Col 29 cells, and the rat vas deferens is an adrenomedullin receptor that is blocked with low affinity by ha-CGRPg.37 (Figure 30 B).

The inability of ['^I]ha-CGRP to bind adrenomedullin receptors at the concentrations used in competition binding assays is consistent with the fact that the putative CGRP^ receptor has not been previously identified by radioligand binding.

Despite evidence that the putative CGRP2 receptor is an adrenomedullin receptor, the broad distribution o f ha-CGRPg.37 affinities that have been determined in functional assays suggests that additional tissue-dependent factors also influence the affinity of this antagonist. The goal o f chapter 4 was to identify tissue-dependent variables that may influence the ability o f ho.-CGRPg.37 to block CGRP-induced responses in isolated tissues. A review o f previously published data suggests that proteolytic degradation of ha-CGRP,,.37 and the release of endogenous CGRP from neuronal storage vesicles are two factors that may influence the affinity of ha-CGRPg.37 in functional studies. In addition, my own radioligand binding data indicates that the affinity o f ha-CGRPg.37 and

[Cys(ACM )' ']ha-CGRP is dependent upon experimental conditions. Thus, several factors (proteolytic degradation of ha-CGRPg.37, release of endogenous CGRP, and buffer conditions) in addition to CGRP-induced activation o f CGRP, and adrenomedullin receptors contribute to the variable affinity o f ha-CGRPg.37.

Implications for the Development of Therapeutic Agents

This work may also have implications for the development of therapeutically useful drugs. Previous investigators have proposed that CGRP is involved in a variety of biological processes including vasodilation, angiogenesis, nociception, uterine relaxation,

107 giucose metaboiism, and wound heaiing (Bel! and McDermott, 1996; Wimalawansa,

1996). This suggests that CGRP and its analogs may be useful therapeutic agents for the treatment of preeclampsia, pain, type II diabetes, migraine headache, and a variety of conditions involving the cardiovascular system (Bell and McDermott, 1996;

Wimalawansa, 1996). However, the involvement of CGRP in such diverse processes also suggests that CGRP receptor agonists and antagonists may cause many side effects.

Consequently, there is great interest in the development of CGRP receptor ligands that target specific tissues. Most work toward this goal has focused on the development of ligands that are selective for the CGRP, receptor or the putative CGRP2 receptor.

However, my results suggest that alternative strategies are needed to develop ligands that have tissue-selective effects.

Ligands that specifically interact with either CGRP, or adrenomedullin receptors could have tissue-selective effects. ha-CGRPg.^ has been widely regarded as a CGRP, receptor-selective antagonist. However, my data demonstrate that adrenomedullin receptors are also blocked by this ligand. Adrenomedull^.^ is the only adrenomedullin receptor antagonist that is currently available. However, this ligand blocks adrenomeduHin receptors with very low affinity. Furthermore, it is likely that the therapeutic use of ha-CGRPg.^ and adrenomedullin^.^ would be complicated by proteolytic degradation, poor absorption, and other problems that are inherent to peptides. B1BN4096BS and SB273779 are nonpeptide antagonists that have been reported to block CGRP receptors with high affinity (Aiyar et al., 2001a; Doods et al.,

2000). The selectivity of these ligands for CGRP receptors compared to adrenomedullin

108 receptors has not been reported. However, their nonpeptide nature and their high affinity for CGRP receptors makes these drugs attractive candidates for therapeutic use.

The development of partial agonists may provide another mechanism to achieve tissue-selective effects through CGRP receptors. Partial agonists are not dependent upon receptor subtypes for their tissue selectivity. Rather, they activate receptors in different tissues on the basis of tissue-dependent differences in receptor reserve and the efficiency with which the receptors are coupled to agonist-induced responses (Nickerson, 1956).

[Pro'^jha-CGRP stimulates cAMP synthesis in CGRP, receptor-transfected HER 293 cells (Figure 11) but does not stimulate CGRP, receptors in guinea pig pancreatic acinar cells (Li et al., 1997). Similar effects have been observed with [Cys(ACM)* ']CGRP.

This partial agonist (Waugh et al., 1997) binds membranes prepared from the rat vas deferens and the guinea pig atrium with equal affinities (Van Rossum et al., 1994) but only induces a biological response in the vas deferens (Dennis et al., 1989). Thus,

[Cys(ACM)" ]ha-CGRP and [Pro'^ha-CGRP can selectively induce responses in tissues that express the same CGRP, receptor. The tissue selectivity of [Cys(ACM)^]CGRP and

[Pro' *]ha-CGRP have not been tested in whole animals. However, the selectivity demonstrated by these ligands in isolated tissues suggests that partial agonists could be useful for targeting drug therapy to specific tissues.

There is an increasing interest in the development of viral based systems for tissue-specific drug delivery. Consequently, adenoviral, lentiviral, and herpes virus vectors have been investigated for their ability to deliver genes to specific tissues (see reviews by Lever, 2000 and Cotter and Robertson, 1999). Several investigators have reported that adenoviral based systems are effective for delivering DNA encoding CGRP

109 to seiected tissues. Champion et al. (2000) reported that a recombinant adenovirus expressing CGRP effectively relieved pulmonary hypertension in mice without influencing the heart rate or systemic blood pressure. An adenoviral vector has also been used to express CGRP in the cerebral arteries of rabbits (Toyoda et al., 2000a). This method of CGRP delivery causes dilation of cerebral arteries without influencing systemic arterial pressure and may be a useful therapeutic strategy for the treatment of cerebral vasospasm (Toyoda et al., 2000b). In another study, a vector containing the

CGRP gene controlled by the osteocalcin promoter was used to selectively express

CGRP in mouse osteoclasts (Ballica et al., 1999). Bone densities of mice expressing this construct were 23-29 % greater than mice expressing a control vector that did not encode

CGRP. In addition the serum of these mice did not contain high concentrations of

CGRP, and no adverse effects were reported. These studies indicate that the development of viral based drug delivery strategies may result in greater tissue selectivity for CGRP and its peptide analogs than can be achieved by traditional routes of drug administration.

The greatest challenge to the use of CGRP receptors as therapeutic targets is our lack of knowledge of their physiological roles under normal or pathological conditions.

Studies using live rats have shown that CGRP is involved in nociception, inflammation, and wound healing (Yu et al., 1998; Engin, 1998). However, a great deal of our understanding of the involvement of CGRP in other physiological processes is limited to speculation that is based upon studies with isolated tissues. Although these studies have provided useful information concerning the pharmacological characteristics of CGRP receptors, the physiological roles of these receptors cannot be predicted based upon

110 studies with isoiated tissues alone. This has been most recently emphasized by studies investigating the influence of endogenous CGRP on the cardiovascular system. CGRP is a potent vasodilator and chronotropic agent when applied to isolated tissues. However, studies using mice that do not express CGRP have provided conflicting results regarding the physiological role of this peptide in the regulation of blood pressure, cardiac output, and other cardiovascular parameters (Gangula et al., 2000; Shen et al., 2001). These studies emphasize the importance of integrating data collected in isolated tissues with data obtained from whole animal experiments. Future emphasis in this field should include the identification of the physiological roles of CGRP in vivo and the subsequent identification of disease states in which CGRP receptors might provide a rational therapeutic target.

I l l APPENDIX A

INFLUENCE OF CGRP ON INTRACELLULAR CALCIUM IN SK-N-MC CELLS

AND PORCINE CGRP, RECEPTOR-TRANSFECTED HEK 293 CELLS

Introduction

Aiyar et ai. (1999) reported that human and porcine CGRP, receptors are coupled to an increase in intracellular cAMP and intracellular calcium when stably transfected into HEK 293 cells. In addition, Huang et al. (1999) reported that stimulation of CGRP receptors in mouse and rat cardiomyocytes increases intracellular calcium through cAMP-dependent activation of L-type potassium channels. SK-N-MC cells are commonly used to study CGRP, receptors (Longmore et al., 1994; Semark et al., 1992;

McLatchie et al., 1998). Although CGRP stimulates cAMP synthesis in this cell line, it is unknown whether CGRP influences intracellular calcium in these cells. Therefore, the goal of this study was to determine whether the intracellular calcium concentration is iniluenced by CGRP in SK-N-MC cells that endogenously express the human CGRP, receptor.

Materials and Methods

Measurement of Intracellular cAMP. ha-CGRP-induced increases in cAMP were measured to verify that the CGRP, receptors expressed in SK-N-MC and HEK 293 ceils were functional. SK-N-MC cells that endogenously express the human CGP.P, receptor and HEK 293 cells stably expressing the porcine CGRP, receptor were grown in 24 well plates in minimum essential medium containing fetal bovine serum (10 %), penicillin G

112 (100 units/ml), streptomycin (100 pg/ml) and amphotericin B (0.25 pg/ml). The plates were placed in a humidified incubator in an atmosphere of 5 % €0^/95 % air and maintained at 37°C. Culture medium was removed from the confluent cells, and the cells were gently rinsed with 2 ml of warm (37 °C) HEPES-buffered Krebs solution that contained 20 mM HEPES, 4 mM NaHCO^, 11 mM dextrose, 1.2 mM NaPI^PO^, 5.5 mM

KC1, 2.5 mM CaCl^, 1.2 mM MgC^, and 0.5 mM isobutylmethylxanthine. Cells were incubated with 450 pi HEPES-buffered Krebs solution for 10 minutes at 37°C before an additional 50 pi of HEPES-buffered Krebs solution that contained ha-CGRP was added.

The solution was removed after 30 minutes, and the cells were lysed with 100 pi of 90 % ethanol / 10 % water. The ethanol was evaporated in a 37 °C incubator, and cAMP present in the dried lysate was measured using a radioimmunoassay provided by

Diagnostic Products Corporation (Los Angeles, CA) according to the manufacturer's

instructions.

Measurement of Intracellular Calcium. SK-N-MC and porcine CGRP, receptor-

transfected HEK 293 cells were grown in 35 mm dishes modified with glass coverslip

bottoms. Cells were grown to approximately 30 % confluence on the glass coverslips

under the conditions described above. Cell culture medium was removed and the cells

were rinsed twice with buffered saline solution (120 mM NaCl, 5 mM KC1, 1 mM

MgC^, 1 mM CaC^, 6 mM dextrose, 10 mM MOPS, 5 mM NaHCO^). Cells were

incubated with 1 ml fura-2 AM loading solution (buffered saline solution supplemented

with 0.1 mg/ml bovine serum albumin, 0.02 % cremephor, and 2 pM fura-2 AM) for 40

minutes at 37 °C. Fura-2 AM loading solution was removed and the cells were rinsed

with 2 ml of HEPES/THAM-buffered Krebs solution (120 mM NaCl, 5 mM KC1, 587

113 pM KH2PO4, 5 9 8 pM Na2HP0 4 , 2.5 mM MgC^, 20 mM dextrose, 10 mM 1.8 mM

CaC^, 10 mM HEPES, and 10 mM THAM) and incubated at 37 °C for 10 minutes.

A Nikon inverted fluorescence microscope connected to a PT1 spectrofluorometer

(Lawrenccville, NJ) was used to measure intracellular calcium in single patches of 15-20 cells. Cells grown on glass coverslips were placed on a 37 °C heated microscope stage and viewed at 1000 X magnification. Changes in intracellular calcium were detected by stimulating fura-2 with light at alternating wavelengths of 340 nm and 380 nm. Fura-2 fluorescence emission was measured at 510 nm, and changes in intracellular calcium were detected as changes in the fura-2 340/380 emission ratio. The cells were stimulated with ha-CGRP dissolved in HEPES/THAM-buffered Krebs solution. If they did not respond after approximately 1 minute, the cells were rinsed with HEPES/THAM- buffered Krebs solution and a higher concentration of ha-CGRP was added.

Acetylcholine stimulates an increase in intracellular calcium in SK-N-MC cells and was used as a positive control.

Data Analysis. The ability ha-CGRP to stimulate cAMP synthesis was determined in SK-N-MC cells and porcine CGRP, receptor-transfected HEK 293 cells as described above. GraphPad Prism was used to analyze these data by nonlinear regression and to determine the EC^ value of these dose response curves. No attempt was made to quantify changes in intracellular calcium.

Results

CGRP-induced cAMP Synthesis. CGRP-induced cAMP synthesis was used as a positive control to verify that the CGRP, receptors expressed in SK-N-MC and HEK 293

114 cells were functional. CGRP dose-dependently increased cAMP in both cell lines

(Figure 31). ha-CGRP was 3.3-fold more potent in SK-N-MC cells (ECjQ = 4.8 + 1.4 nM) than in porcine CGRP, receptor-transfected HEK 293 cells (EC,o = 16.0 + 6.2 nM).

However, a larger maximum response was observed in HEK 293 ceils (21.1 +4.5 fold over basal) than in SK-N-MC cells (3.7 + 0.4 fold over basal).

CGRP-induced Changes in Intracellular Calcium. Fura-2 microfluorometry was used to determine if ha-CGRP influences the intracellular calcium concentration in

SK-N-MC cells that endogenously express the human CGRP, receptor No changes in intracellular calcium were observed in SK-N-MC cells stimulated with 1 pM - 1 pM ha-CGRP (Figure 32 A). However, 50 pM acetylcholine caused a rapid increase in intracellular calcium. Thus, SK-N-MC cells possess the required cellular machinery to increase intracellular calcium.

Fura-2 microfluorometry was also used to investigate the effect of ha-CGRP on intracellular calcium in HEK 293 cells transfected with the porcine CGRP, receptor

(Figure 32 B). No changes in intracellular calcium were observed when these cells were stimulated with 1 pM - 0.1 pM ha-CGRP. In fact 5 patches of cells from different 35 mm dishes were stimulated with 1 pM ha-CGRP before a single group of cells was found that did respond to this agonist (Figure 32 B). The effect of 50 pM acetylcholine on these cell patches was similar to that observed in SK-N-MC cells (not shown).

115 A

Figure 31. CGRP-induced cAMP synthesis in SK-N-MC (pane! A) and porcine CGRP, receptor-transfected HEK 293 ceils. These data represent the mean of 5 (panel A) or 3 (panel B) experiments, respectively.

116 A

W

CGRP

Figure 32. CGRP does not infiuence intracellular calcium concentrations in SK-N-MC or porcine CGRP, receptor-transfected HEK 293 cells. Single patches of 15-20 SK-N- MC cells (panel A) were stimulated with different concentrations of ha-CGRP, and intracellular calcium was measured using fura-2 microfluorometry. Acetylcholine was used as a positive control. This tracing is representative of 5 experiments performed using cells grown in different 35 mm dishes. Single patches of 15-20 HEK 293 cells transfected with the porcine CGRP, receptor were stimulated with CGRP. Five different cell patches were analyzed, and this was the only tracing in which a response to ha- CGRP was observed.

117 Discussion

Aiyar et a). (1999) reported that human and porcine CGRP, receptors are coupled to an increase in intracellular calcium when expressed in HEK 293 cells (Aiyar et al.,

1999). In contrast, I found that 1 pM ha-CGRP increased intracellular calcium in only 1 out of 5 patches of HEK 293 cells that were examined. One explanation for these results could be that the HEK 293 cultures used in this study may consist of one subpopulation of cells that express the CGRP, receptor and another subpopulation that does not.

Perhaps the first 4 cell patches that 1 analyzed did not respond to ha-CGRP because they were from a subpopulation of cells that does not express the CGRP, receptor.

The microfluorometry assay that I used to measure intracellular calcium analyzes small groups of cells. In contrast, Aiyar et al. (1999) measured intracellular calcium using a cuvet-based technique that analyzes the response of a large group of cells

(2,000,000 cells) simultaneously. It is possible that the simultaneous analysis of many cells makes the cuvet-based technique less sensitive than microfluorometry to mixed populations of cells that differ in their expression of the CGRP, receptor. Thus, the conflicting conclusions of these studies may result from the different abilities of microfluorometry and the cuvet-based assays to detect responses in the presence of mixed cell populations.

The goal of this study was to determine whether ha-CGRP stimulates an increase

in intracellular calcium in SK-N-MC cells. The ha-CGRP-induced increase in

intracellular cAMP in SK-N-MC and HEK 293 cells verifies that the CGRP, receptors

present in both cell lines are functional. I was unable to demonstrate that CGRP,

receptors are coupled to changes in intracellular calcium in SK-N-MC cells. However,

118 the fact that this response could not be repeatedly demonstrated in the positive control cell line (porcine CGRP, receptor-transfected HEK 293 cells) suggests that the microfluorometry data in SK-N-MC cells are inconclusive.

119 APPENDIX B

PARTIAL MAPPING OF THE CALITONIN RECEPTOR-LIKE

RECEPTOR GENE

Introduction

Caicitonin receptor like receptor (CRLR) cDNA has been cloned from several species including the human, pig, rat, and flounder (Aiyar et al., 1996; Elshourbagy et al.,

1998; Suzuki et al., 2000; Chang et al., 1993). However, the sequence for the entire

CRLR gene has not been reported. In this study, 1 compared the cDNA sequence of human CRLR to other human nucleotide sequences that have been submitted to

Genbank. This data provides preliminary information that could be useful for the further characterization of the CRLR gene.

Methods

I identified the CRLR gene by searching the human genome for DNA sequences that matched the human CRLR cDNA sequence (accession # L76380). This was done using the FASTA function of the Wisconsin Package Version 10.1 (Genetics Computer

Group, Madison, WI) software. This initial search identified a DNA sequence that contained 26 unordered pieces of DNA from human chromosome 2. The total length of these pieces was 198,846 nucleotides. This sequence was too long to compare directly to

L76380. Therefore, this DNA sequence (accession # AC046198) was divided into 40 sections that were 5000 nucleotides long. Each section was individually compared to

L76380 using the DotPlot function of the Genetics Computer Group program to identify

120 sections of AC046198 that were identical to regions of L76380. Identical regions were also compared to L76380 using the Pairwise BLAST function on the web site of the

National Center for Biotechnology and Information (http:www.ncbi.nlm.nih.gov) and by performing a Best-Fit comparison using the Genetics Computer Group software. Five segments of AC046198 contained multiple regions with 100 % identity to regions of

L76380. However, these identical regions were separated on AC046198 by DNA sequences with no homology to L76380 (introns). These data were used to construct a partial map of the human CRLR gene. I also found an additional clone of human chromosome 2 (accession # AC074020). It was analyzed in the same way as AC046198, but this clone provided no additional information.

Results and Future Directions

Figure 33 shows the features of the CRLR gene that were identified in this study.

In some cases, it was unknown whether exons were separated by introns because some of the exons were found on separate unordered pieces of DNA in AC046198. These regions were labeled with a question mark to indicate that the presence or absence of an intron could not be determined in that location. I identified 8 introns ranging in size from 83 to

2751 nucleotides. Exon sizes ranged from 60 tol293 nucleotides. The sum of all intron sequences (9780 nucleotides) and all exon sequences (2941 nucleotides) identified in this study indicates that this gene is at least 12,721 nucleotides long. Nakazawa et al. (2001) recently reported that they sequenced a bacterial artificial chromosome that contains the human CRLR gene. They found that this gene contains 14 introns and 15 exons and spans 103,145 nucleotides (including 1 intron that spans more than 60,000 nucleotides).

121 AC046198 ci 3 3! 31 O' VI r- 3 vi 31 3 00 oo o (genomic sequence) 3 kO C3 00 'O O' O' 00 00 O' Cl 00 00 'C kO 5'- oo 00 C l Cl C l Cl Cl C l G9.Q220 A3ss-G"s y519_Q605 L76380 (cDNA) j C^'-A3'3 1698 bo 2239 bp ? ? ? Intron j intron translation start codon (555)

*3* 3 - VI 3 C l 3 3 1 3 1 C l 3 'c 3 3 8 0 3 3 1 vi O' O' O' O' 3 3 31 3 1 3 1 3 1 C l C l ci ci ^606,(^726 A73..G849 AS50.G962 ^963-A* 054 1 98 bp 1436 bp intron intron

'O 3 1 o o 3 ! 3 ! O' C l 3 - V l O' 0 3 'C 3 o r - 3 'C 'O 'C O' O' O' O' O' Qi055.yH8i Ql!82_Gi333 G1336_A*396 TN397.Q1463 2751 bo 1324 bo 83 bo intron intron intron

3 31 (3 VI 'O 3 O' VI D' 3 O' D' 3 'Cci -^D3) G!464.Q!682 Q1683_Q!724 Q 1725.^3018 151 bn intron transtationt stop codon (1922)

Figure 33. Partial mapping of the human CRLR gene. The horizontal line represents the sequence of the CRLR gene. Numbers in the vertical position refer to nucleotide numbers on the chromosomal DNA (accession # AC046198). The cDNA sequences that correspond to each region are shown in boxes and are numbered according to their position in the cDNA sequence (accession # L76380). Question marks designate areas where the presence or absence of introns could not be determined, and the position of translation start and stop codons are indicated by arrows.

122 The nucleotide sequence investigated by Nakazawa et al. (2001) was not available for comparison to the sequence represented by Figure 33 because they did not publish the sequence or submit it to Genbank. However, their data indicate that this gene is much larger than the sequence that I have identified.

This study provides a starting point to characterize the human CRLR gene.

Future investigators may be interested in determining whether CRLR transcripts are alternatively spliced. Future projects may also involve the identification of the CRLR gene promoter and the identification of processes that regulate transcription of this gene.

It would be particularly interesting to determine if the expression of CRLR and the expression of receptor activity modifying proteins are regulated by similar or different mechanisms.

123 REFERENCES

Abel PW, Waugh D, and Jeffries WB (1997) Radioligand binding with ['^Ij-labeled peptides, in MetAoJx in Mo/ecM/or E/o/ogy (Irvine GB and Williams CH eds) pp 323-330, Humana Press, Totona, New Jersey.

Aiyar N, Daines RA, Disa J, Chambers PA, Sauermelch CF, Quiniou MJ, Khandoudi N, Gout B, Douglas SA, and Willette RN (2001a) Pharmacology of SB-273779, a nonpeptide calcitonin gene-related peptide 1 receptor antagonist. JPAorwoco/ Exp 7%er 296:768-775.

Aiyar N, Disa J, Pullen M, and Nambi P (2001b) Receptor activity modifying proteins interaction with human and porcine calcitonin receptor-like receptor (CRLR) in HEK-293 cells. Mo/ Ce// E/ocAew 224:123-133.

Aiyar N, Disa J, Stadel JM, and Lysko PG (1999) Calcitonin gene-related peptide receptor independently stimulates 3',5'-cyc!ic adenosine monophosphate and Ca*^ signaling pathways. Mo/ Ce// Eioc/mw 197:179-185.

Aryar N, Rand K, Elshourbagy NA, Zeng Z, Adamou JE, Bergsma DJ, and Li Y (1996) A cDNA encoding the calcitonin gene-related peptide type 1 receptor. JEio/ CAem 271:11325-11329.

Amara SG, Jonas V, Rosenfeld MG, Ong ES, and Evans RM (1982) Alternative RNA processing in calcitonin gene expression generates mRNAs encoding different polypeptide products. /Vornre 296:240-244.

Andersen SL and Clausen T (1993) Calcitonin gene-related peptide stimulates sodium-potassium transport in rat soleus muscle. Aw JPAyx/o/ 264:C419-C429.

Ballica R, Valentijn K, Khachatryan A, Guerder S, Kpaadia S, Gundberg C, Gilligan J, Flavell RA, and Vignery A (1999) Targeted expression of calcitonin gene-related peptide to osteoblasts increases bone density in mice. J Rone Miner Rex 14:1067-1074.

Bell D and McDermott BJ (1994) Calcitonin gene-related peptide stimulates a positive contractile response in rat ventricular cardiomyocytes. J CarJiovaxc P/mrwoco/ 23:1011-1021.

Bell D and McDermott BJ (1996) Calcitonin gene-related peptide in the cardiovascular system: characterization of receptor populations and their (patho)physiological significance. PAorwoco/ Rev 48:253-288.

Boulanger Y, Khiat A, Larocque A, Foumier A, and St Pierre S (1996) Structural comparison of alanine-substituted analogues of the calcitonin gene-related peptide 8-37. 7ni JPepiiJe Protein Rex 47:477-483.

Bylund DB, Eikenberg DC, Hieble JP, Langer SZ, Lefkowitz RJ, Minneman KP, Molinoff PB, Ruffolo RR Jr, Trendelenburg U (1994) International union of pharmacology nomenclature of adrenoceptors. PAorwoco/ Rev 46:121-136.

124 Chakir M, D'Orleans-Juste P, and Piante GE (1995) Neutral endopeptidase inhibition, a new approach in the exploration of diabetic vasculopathy in rats. Pur JPAar/uaco/ 285:11-18.

Champion HC, Bivaiacqua TJ, Toyoda K, Heistad DD, Hyman AL, and Kadowitz PJ (2000) In vivo gene transfer of prepro-calcitonin gene-related peptide to the lung attenuates chronic hypoxia-induced pulmonary hypertension in the mouse. C/rcM/a/ion 101:923-930.

Champion HC, Santiago JA, Murphy WA, Coy DH, and Kadowitz PJ (1997) Adrenomedullin22-52 antagonizes vasodilator responses to CGRP but not adrenomedullin in the cat. /hu J PAys/o/ 272:R234-R242.

Chang CP, Pearse RV, O'Connell S, and Rosenfeld MG (1993) Identification of a seven transmembrane helix receptor for corticotropin-releasing factor and sauvagine in mammalian brain. A/euron 11:1187-1195.

Christopoulos G, Perry KJ, Morfis M, Tilakaratne N, Gao Y, Fraser NJ, Main MJ, Foord SM, and Sexton PM (1999) Multiple arnylin receptors arise from receptor activity-modifying protein interaction with the calcitonin receptor gene product. Mo/ PAarwaco/ 56:235-242.

Connat JL, Schnuriger V, Zanone R, Schaeffer C, Gaillard M, Faivre B, and Rochette F (2001) The neuropeptide calcitonin gene-related peptide differently modulates proliferation and differentiation of smooth muscle cells in culture depending on the cell type. Pegu/ Pep/ 101:169-178.

Coppock HA, Owji AA, Austin C, Upton PD, Jackson ML, Gardiner JV, Ghatei MA, Bloom SR, and Smith DM (1999) Rat-2 fibroblasts express specific adrenomedullin receptors, but not calcitonin-gene-related-peptide receptors, which mediate increased intracellular cAMP and inhibit mitogen-activated protein kinase activity. P/ocAew J 338:15-22.

Cote GJ, Stolow DT, Peleg S, Berget SM, and Gagel RF (1992) Identification of exon sequences and an exon binding protein involved in alternative RNA splicing of calcitonin/CGRP. /Vac/ ^4cA/.y Pay 20:2361-2366.

Cotter MA and Robertson ES (1999) Molecular genetic analysis of herpesviruses and their potential use as vectors for gene therapy applications. Carr Gp/a Mo/ PAer 1:633-644.

Cox H (1995) Receptors for calcitonin gene related peptide (CGRP) in gastrointestinal epithelia. Can JPAyy/o/ PAarraaco/ 73:974-980.

Cox HM and Tough IR (1994) Calcitonin gene-related peptide receptors in human gastrointestinal epithelia. Pr JPAarvaaco/ 113:1243-1248.

Dang K, Disa J, Gout B, and Aiyar N (1999) Comparative affinities of adrenomedullin (AM) and calcitonin gene-related peptide (CGRP) for ['^1] AM and [*^1] CGRP specific binding sites in porcine tissues. JPecep/ P/gaa/ PraayJPey 19:803-817.

125 Dennis T, Fournier A, Cadieux A, Pomerleau F, Jolicoeur FB, St. Pierre S, and Quirion R (1990) hCGRPg.37, a calcitonin gene-related peptide antagonist revealing calcitonin gene-related peptide receptor heterogeneity in brain and periphery. 7E/zar/nuco/ Exp 7%er 254:123-128.

Dennis T, Fournier A, St. Pierre S, and Quirion R (1989) Structure-activity profile of calcitonin gene-related peptide in peripheral and brain tissues. Evidence for receptor multiplicity J P/mrwaco/ Exp 7%er 251:718-725.

Deupree JD, Hinton KA, Cerutis DR, and Bylund DB (1996) Buffers differentially alter the binding of [^Hjrauwolscine and ['fl]RX821002 to the alpha-2 adrenergic receptor subtypes. 7 PAarwaco/Exp 7%

Disa J, Parameswaran N, Nambi P, and Aiyar N (2000) Involvement of cAMP-dependent protein kinase and pertussis toxin-sensitive G-proteins in CGRP mediated JNK activation in human neuroblastoma cell line. /VearopepaWay 34:229-233.

Dogan A, Xuzuki Y, Koketsu N, Osuka K, Saito K, Takayasu M, Shibuya M, and Yoshida J (1997) Intra venous infusion of adrenomedullin and increase in regional cerebral blood flow and prevention of ischemic brain injury after middle cerebral artery occlusion in rats. 7 Cercb P/dd7 E/owA/ftaZ? 17:19-25.

Doods H, Hallermayer G, Wu D, Entxerotk M, Rudolf K, Engel W, and Eberlein W (2000) Pharmacological profile of BIBN4096BS, the first selective small molecule CGRP antagonist. Er7E/zarwaco/ 129:420-423.

Drapeau G, Audet R, Levesque L, Godin D, and Marceau F (1993) Development and in vivo evaluation of metabolically resistant antagonists of B1 receptors for kinins. 7 PAnrwncd/ Exp 7Eer 266:192-199.

Edvinsson L, Sams A, Jansen-Olesen I, Tajti J, Kane SA, Rutledge RZ, Koblan KS, Hill RG, and Longmore J (2001) Characterization of the effects of a non-peptide CGRP receptor antagonist in SK-N-MC cells and isolated human cerebral arteries. EMr7E/zarw<2co/ 415:39- 44.

Eguchi S, Hirata Y, Iwasaki H, Sato K, Watanabe TX, Inui T, Nakajima K, Sakakibara S, and Marumo F (1994) Structure-activity relationship of adrenomedullin, a novel vasodilatory peptide in cultured rat vascular smooth muscle cells. En7ocrmo/ogy 135:2454-2458.

Elshourbagy NA, Adamou JE, Swift AM, Disa J, Mao J, Ganguly S, Bergsma DJ, and Aiyar N (1998) Molecular cloning and characterization of the porcine calcitonin gene-related peptide receptor. En7ocrmo/ogy 139:1678-1683.

Engin C (1998) Effects of calcitonin gene-related peptide in wound contraction in denervated and normal rat skin: A preliminary report. E/a.st Eccon.str 101:1887-1890.

126 Emsberger P and UTrichard DC (1987) Modulation of agonist and antagonist interactions at kidney alpha 1-adrenoceptors by nucleotides and metal ions. JPAarwaco/ 133:165-176.

Evans BN, Rosenblatt MI, Mnayer LO, Oliver KR, and Dickerson 1M (2000) CGRP-RCP, a novel protein required for signal transduction at calcitonin gene-related peptide and adrenomedullin receptors. VE/c/ CAew 275:31438-31443.

Fargeas MJ, Fioramonti J, and Bueno L (1985) Calcitonin gene-related peptides: brain and spinal action on intestinal motility. PepEr/a? 6:1167-1171.

Femandez-Patron C, Stewart KG, Zhang Y, Koivunen E, Radomski MW, and Davidge ST (2000) Vascular matrix metalloproteinase-2-dependent cleavage of calcitonin gene-related peptide promotes vasoconstriction. C/rcPas 87:670-676.

Flavahan NA and Vanhoutte PM (1986) ct]-Adrenoceptor subclassification in vascular smooth muscle. Pharmaco/ 7:347-349.

Fluhmann B, Muff R, Hunziker W, Fischer JA, and Bom W (1995) A human calcitonin receptor­ like structure. EmcArv?; Pas- Cowwan 206:341-347.

Ford A, Daniels DV, Chang DJ, Gever GR, Jasper JR, Lesnick JD, and Clarke DE (1997) Pharmacological pleiotropism of the human recombinant ai,\ adrenoceptor: implications for ai-adrenoceptor classification. Er JPAarwaco/ 121:1 127-1135.

Foulkes R, Shaw N, Bose C, and Hughes B (1991) Differential vasodilator profile of calcitonin gene-related peptide in porcine large and small diameter coronary artery rings. Ear J PAarwaca/ 201:1453-149.

Fraser NJ, Wise A, Brown J, McLatchie LM, Main MJ, and Foord SM (1999) The amino terminus of receptor activity modifying proteins is a critical determinant of glycosylation state and ligand binding of calcitonin receptor-like receptor. Mo/P/rarwaco/ 55:1054-1059.

Gangula PRR, Zhao H, Supowit SC, Wimalawansa SJ, Dipette DJ, Westlund KN, Gage! RF, and Yallampalli C (2000) Increased blood pressure in a-calcitonin gene-related peptide/calcitonin gene knockout mice. TT^perteayfoa 35:470-475.

Giuliani S, Wimalawansa SJ, and Maggi CA (1992) Involvement of multiple receptors in the biological effects of calcitonin gene-related peptide and arnylin in rat and guinea-pig preparations. ErJP/;ar7?!aco/107:510-514.

Hasbak P, Sams A, Schifter S, Longmore J, and Edvinsson L (2001) CGRP receptors mediating CGRP-, adrenomeduHin- and amylin-induced relaxation in porcine coronary arteries. Characterization with 'Compound l'(W098/l 1128), a non-peptide antagonist. E rJ P/:ar/aaao/ 133:1405-1413.

127 Huang MH, Knight PR, and Izzo JL (1999) Ca^-induced Ca^ release involved in positive inotropic effect mediated by CGRP in myocytes. Tw JPAyy/o/ 276:R259-R264.

Hussman K, Sexton PM, Fischer JA, and Bom W (2000) Mouse receptor-activity-modifying proteins 1,2, and 3: amino acid sequence, expression, and function. Mo/ Co// Pwr/ocrmo/ 162:35-43.

Katayama M, Nadel JA, Bunnett NW, Di Maria GU, Haxhiu M, and Borson DB (1991) Catabolism of calcitonin gene-related peptide and substance P by neutral endopeptidase. PqpPPay 12.563-567.

Kenakin TP (1993) Methods of drug and receptor classification, in PAormoco/og/c ^oo/yyily o/" Drog-Poceptorfnteroct/ooj pp 344-384, Raven Press, New York.

Kruger L, Silverman JD, Mantyh PW, Stemini C, and Brecha NC (1989) Peripheral patterns of calcitonin-gene-related peptide general somatic sensory innervation: cutaneous and deep terminations. J Comp Aeoro/ 280:291-302.

Kuwasako K, Kitamura K, Ito K, Uemura T, Yanagita Y, Kato J, Sakata T, and Eto T (2001) The seven amino acids of human RAMP2 (86-92) and RAMP3 (59-65) are critical for agonist binding to human adrenomedullin receptors. 7P/o/ CAem 276:49459-49465.

Lam F (2000) Pharmacological char acterization of receptor types mediating coronary vasodilator actions of sensory neuropeptides in the guinea pig. J CorJ/ovo^c PAormooo/ 35:646-652.

Lei S, Mulvany MJ, and Nyborg NCB (1994) Characterization of the CGRP receptor and mechanisms of action in rat mesenteric small arteries. PAormoco/ Pox/co/ 74:130-135

Leighton B and Foot EA (1995) The role of the sensory peptide calcitonin-gene related peptide(s) in skeletal muscle carbohydrate metabolism: effects of capsaicin and resiniferatoxin. P/ocAom 7307:707-712.

Lenz HJ, Mortrud MT, Rivier JE, and Brown MR (1985) Calcitonin gene related peptide inhibits basal, pentagastrin, histamine, and bethanecol stimulated gastric acid secretion. Gat 26:550- 555.

Lever AM (2000) Lentivira) vectors: progress and potential. Corr Cpm Mo/ PAer 2:488-496.

Li J, Matsuura JE, Waugh DJJ, Adrian TE, Abel PW, Manning MC, and Smith DD (1997) Structure-activity studies on position 14 of human a-calcitonin gene-related peptide. 7Mo7 CAom 40:3071-3076.

Li N, Lundeberg T, and Yu LC (2001) Involvement of CGRP and CGRPI receptor in nociception in the nucleus accumbens of rats. Prom Pay 901:161-166.

128 Limbird LE (1996) Identification of receptors using direct radioligand binding techniques, in Ce// Sur/uce Receptors.* ,4 SAort Course on Theory awr/MetAor/x, PecouJ Pr//t/on pp 61-122, Kluwer Academic Publishers, Norwell Massachusetts.

Longmore J, Hogg JE, Hutson PH, and Hill RG (1994) Effects of two truncated forms of human calcitonin-gene related peptide: implications for receptor classification. PurJPAarnMco/ 265:53-59.

Lou H and Gagel RF (1998) Alternative RNA processing-its role in regulating expression of calcitonin/calcitonm gene-related peptide. ./Pnr/ocr/no/156:401-405.

Lowry OH, Rosebrough NJ, Farr AL, and Randall RJ (1951) Protein measurement with the fblin phenol reagent. JP /o/ CAew 193:265-275.

Luebke AE, Dahl GP, Roos BA, and Dickerson IM (1996) Identification of a protein that confers calcitonin gene-related peptide responsiveness to oocytes by using a cystic fibrosis transmembrane conductance regulator assay. ProcNat/v4ca<7 &7 C&4 93:3455- 3460.

Maggi CA, Chiba T, and Giuliani S (1991) Human a-calcitonin gene-related peptide-(8-37) as an antagonist of exogenous and endogenous calcitonin gene related peptide. Pur JPAarnruco/ 192:85-88.

McLatchie, LM, Fraser NJ, Main MJ, Wise A, Brown J, Thompson N, Solari R, Lee MG, and Foord SM (1998) RAMPs regulate the transport and ligand specificity of the calcitonin- receptor-like receptor. Nature 393:333-339.

Mentlein R and Roos T (1996) Proteases involved in the metabolism of angiotensin 11, bradykinin, calcitonin gene-related peptide (CGRP), and neuropeptide Y by vascular smooth muscle cells. Pept/Pe^ 17:709-720.

Mimeault M, Fournier A, Dumont Y, St. Pierre S, and Quirion R (1991) Comparative affinities and antagonist potencies of various human calcitonin gene-related peptide fragments on calcitonin gene-related peptide receptors in brain and periphery. J PAarmueo/ Cxp PAer 258:1084-1090.

Muff R, Leuthauser K, Buhlmann N, Foord, SM, Fischer JA, and Bom W (1998) Receptor activity modifying proteins regulate the activity of a calcitonin gene-related peptide receptor in rabbit aortic endothelial cells. PEAS'Letters 441:366-368.

Muff R, Stangl D, Bom W, and Fischer JA (1992) Comparison of calcitonin gene-related peptide receptor in a human neuroblastoma cell line (SK-N-MC) and a calcitonin receptor in a human breast carcinoma cell line (T47D). Tun NT Tear/ Pc/ 657:106-116.

Muramatsu I, Ohmura T, Kigoshi S, Hashimoto S, and Oshita M (1990) Pharmacological subclassification of oq-adrenoceptors I vascular smooth muscle. PrJPAarTuaco/ 99:197-201.

129 Nakazawa I, Nakajima T, Harada H, Ishigami T, Umemura S, and Emi M (2001) Human calcitonin receptor-like receptor lor adrenomeduHin: genomic structure, eight single-nucleotide polymorphisms, and haplotype analysis. J77MTH Gene/ 46:132-136.

New HV and Mudge AW (1986) Calcitonin gene-related peptide regulates muscle acetylcholine receptor synthesis. Nature 323:809-811.

Nickerson M (1956) Receptor occupancy and tissue response. Nature 178:697-698

Nishikimi T, Horio T, Yoshihara F, Nagaya N, Matsuo H, and Kangawa K (1998) Effect of adrenomeduHin on cAMP and cGMP levels in rat cardiac myocytes and nonmyocytes. Far J PAararaca/ 353:337-344.

Parameswaran N, Disa J, Spielman WS, Brooks DP, Nambi P, and Aiyar N (2000) Activation of multiple mitogen-activated protein kinases by recombinant calcitonin gene-related peptide receptor. F w JPAnrTTMCo/ 389:125-130.

Poyner DR (1993) Calcitonin gene-related peptide: multiple actions, multiple receptors. PAarnmco/ PAer 56:23-51.

Poyner DR, Soomets U, Howitt SG, and Langel U (1998) Structural determinants for binding to CGRP receptors expressed by human SK-N- MC and Col 29 cells: studies with chimeric and other peptides. Fr JPAnrwnco/124:1659-1666.

Poyner DR, Taylor GM, Tomlinson AE, Richardson AG, and Smith DM (1999) Characterization of receptors for calcitonin gene-related peptide and adrenomeduHin on the guinea-pig vas deferens. PrJPAnrTanco/ 126:1276-1282.

Puyraimond A, Fridman R, Lemesle M, Arbeille B, and Menashi S (2001) MMP-2 colocalizes with caveolae on the surface of endothelial cells. F.xp Ce// Pe.s 262:28-36.

Rist B, Lacroix JS, Entzeroth M, Doods HN, and Beck-Sickinger AG (1999) CGRP27-37 analogues with high affinity to the CGRP) receptor show antagonistic properties in a rat blood flow assay. Fegn/Pept 79:153-158.

Rorabaugh BR, Abel PA, Smith DD, and Scofield MA (2001) CGRP receptor signaling pathways in SK-N-MC and HEK 293 cells. FTFFF J 15:A931.

Rorabaugh BR, Scofield MA, Smith DD, Jeffries WB, and Abel PA (2001) Functional calcitonin gene-related peptide subtype 2 receptors in porcine coronary arteries are identified as calcitonin gene-related peptide subtype 1 receptors by radioligand binding and reverse transcription-polymerase chain reaction. JPAnrwnco/Frp 77; er 299:1086-1094.

Saha S, Waugh DJJ, Zhao P, Abel PW, and Smith DD (1998) Role of conformational constraints of position 7 of the disulphide bridge of h-a-CGRP derivatives in their agonist versus antagonist properties. JPepttr/e /?e.s 52:112-120.

130 Sams A, Knykhar-Csiilik E, Engberg J, Szok D, Tajti J, Bodi I, Edvinsson L, Vecsei L, and Jansen- Oiesen I (2000) CGRP and adrenomeduHin receptor populations in human cerebral arteries: in vitro pharmacological and molecular investigations in different artery sizes. Par J P^arwaco/ 408:183-193.

Semark JE, Middlemiss DN, and Hutson PH (1992) Comparison of calcitonin gene-related peptide receptors in rat brain and a human neuroblastoma cell line, SK-N-MC Mo/ VoaropAarnraco/ 2:311-317.

Shen YT, PittmanTJ, Buie PS, Bolduc DL, Kane SA, Koblan KS, Gould RJ, and Lynch .11 (2001) Functional role of a-calcitonin gene-related peptide in the regulation of the cardiovascular system. JP/zarwaco/ Expor 77zor 298:551-558.

Sheykhzade M and Nybor NCB (1998) Characterization of calcitonin gene-related peptide (CGRP) receptors in intramural coronary arteries from male and female Sprague Dawley rats. Pr J P/mrTuaco/ 123:1464-1470.

Smith DD and Hanly AM (1997) Purification of synthetic peptides by high performance liquid chromatography, in Mot/zoA z'zz Mo/oca/ar Pz'o/ogy (Irvme GB and Williams CH eds) pp 75- 87, Humana Press, Totona, New Jersey.

Smith DD, Saha S, Waugh DJJ, and Abel PW (2001) Blocking the N-terminus of calcitonin gene- related peptide (8-37) produces high affinity antagonists, in PoptzJoy 2000 (Martinez J and Fehrentz JA eds) pp741-742.

Suzuki N, Suzuki T, and Kurokawa T (2000) Cloning of a calcitonin gene-related peptide receptor and a novel calcitonin receptor-like receptor from the gill of flounder, Para/zcM/zyy o/zTacei/y. Gone 244:81-88.

Tomlinson AE and Poyner DR (1996) Multiple receptors for calcitonin gene-related peptide and arnylin on guinea-pig ileum and vas deferens. Br J Pharmacol 117:1362-1368.

Toyoda K, Faraci FM, Russo AF, Davidson BL, and Heistad DD (2000a) Gene transfer of calcitonin gene-related peptide to cerebral arteries, rfm JP/zgszo/ 278:H586-H594

Toyoda K, Faraci FM, Watanabe Y, Ueda T, Andresen JJ, Chu Y, Otake S, and Heistad DD (2000) Gene transferof calcitonin gene-related peptide prevents vasoconstriction after subarachnoid hemorrhage. Czrc Pey 87:818-824.

Van Rossum D, Menard DP, and Quirion R (1993) Effect of guanine nucleotides and temperature on calcitonin gene-related peptide receptor binding sites in brain and peripheral tissues. Prazn Poy 617:249-257.

Wang Y, Howitt SB, and Poyner DR (2001) A comparison of the antagonist actions of ["''^Ser]CGRP8-37 with CGRP8.37 on the CGRP receptors expressed on L6 , SK-N-MC, and Col 29 cells. PrJP/zarnraco/133 (Proceedings Supplement): 149P.

131 Waugh DJJ, Bockman CS, Smith DD, and Abei PW (1999) Limitations in using peptide drugs to characterize calcitonin gene-related peptide receptors. JPAarwaco/ Exp 289:1419- 1426.

Wimalawansa SJ (1996) Calcitonin gene-related peptide and its receptors: molecular genetics, physiology, pathophysiology, and therapeutic potentials. EnJocr Eev 17:533-584.

Wisskirchen FM, Burt RP, and Marshall I (1998) Pharmacological characterization of CGRP receptors mediating relaxation of the rat pulmonary artery and inhibition of twitch responses of the rat vas deferens. Er JPA^rwuco/ 123:1673-1678.

Wisskirchen FM, Doyle PM, Gough SL, Harris CJ, and Marshall 1 (1999a) Conformational restraints revealing bioactive P-bend structures for ha-CGRPs.37 at the CGRP 2 receptor of the rat vas deferens. Er JP/zarnzaco/ 126:1163-1170.

Wisskirchen FM, Gray DW, and Marshall I (1999b) Receptors mediating CGRP-induced relaxation in the rat isolated thoracic aorta and porcine isolated coronary artery differentiated by ha-CGRPs.37. Er JP/zrrrwaco/128:283-292.

Wisskirchen FM, Doyle PM, Gough SL, Harris CJ, and Marshall 1 (2000) Bioactive P-bend structures for the antagonist ha-CGRPs-37 at the CGRPi receptor of the rat pulmonary artery. E /.7 P/zarwaco/129:1049-1055.

Wisskirchen FM and Marshall I (2000) CGRP2 receptor in the internal anal sphincter muscle of the rat: implications for CGRP receptor classification. Er JP/zarnzaeoF 130:464-470.

Wu D, Eberlein W, Rudolf K, Engel W, Hallermayer G, and Doods H (2000) Characterization of calcitonin gene-related peptide receptors in rat atrium and vas deferens: evidence for a [Cys(Et)' ^]hCGRP-preferring receptor. Ear./P/iar7?zaco/ 400:313-319.

Yoshimoto K, Mitsui-Saito M, Ozaki H, and Karaki H (1998) Effects of adrenomedullin and calcitonin gene-related peptide on contractions of the rat aorta and porcine coronary artery. E? JP/mrTwacaF 123:1645-1654.

Yu L, Hansson P, Lundenberg S, and Lundenberg T (1998) Effects of calcitonin gene-related peptide (8-37) on withdrawal responses in rats with inflammation. Ezzr .7 P/zazwaco/ 347:275-282.

Zimmerman U, Fischer JA, and Muff R (1995) Adrenomedullin and calcitonin gene-related peptide interact with the same receptor in cultured human neuroblastoma SK-N-MC cells. P

132