1 Design and Synthesis of Neurologically Active

Design and Synthesis of Neurologically Active Glycopeptides for Neuroprotection and Antinociception

Item Type text; Electronic Dissertation

Authors Jones, Evan Matthew

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 09/10/2021 09:44:06

Link to Item http://hdl.handle.net/10150/594654

DESIGN AND SYNTHESIS OF NEUROLOGICALLY ACTIVE GLYCOPEPTIDES FOR NEUROPROTECTION AND ANTINOCICEPTION

Evan Matthew Jones

A Dissertation Submitted to the Faculty of the

DEPARTMENT OF CHEMISTRY AND BIOCHEMISTRY

In Partial Fulfillment of the Requirements

For the Degree of

DOCTOR OF PHILOSOPHY

WITH A MAJOR IN CHEMISTRY

In the Graduate College

THE UNIVERSITY OF ARIZONA

2015

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Dissertation Committee, we certify that we have read the dissertation prepared by Evan Matthew Jones, titled Design and Synthesis of Neurologically Active Glycopeptides for Neuroprotection and Antinociception and recommend that it be accepted as fulfilling the dissertation requirement for the Degree of Doctor of Philosophy.

______Date: 12 October 2015 Robin Polt

______Date: 12 October 2015 Victor J. Hruby

______Date: 12 October 2015 Eugene A. Mash

______Date: 12 October 2015 Jeanne E. Pemberton

Final approval and acceptance of this dissertation is contingent upon the candidate’s submission of the final copies of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.

______Date: 12 October 2015 Dissertation Director: Robin Polt

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of the requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that an accurate acknowledgement of the source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the copyright holder.

SIGNED: Evan Matthew Jones

ACKNOWLEDGEMENTS

The path to this point has not always been easy, and I am indebted to any number of people for words of wisdom, encouragement, common sense (which seems to be the first thing to go when the going gets tough), and kindness throughout this process. I owe debts of gratitude to many individuals and groups, including to my advisor, Dr. Robin

Polt, for taking me into the group and giving me the keys to all of the glycopeptide projects; hopefully the scrapes and dings are only cosmetic. To my coworkers (and friends) in the Polt lab, who were always available to discuss synthetic strategies or just lend a sympathetic ear, in particular Dr. Bobbi Anglin, Dr. Lajos Szabo, Dr. Cliff Coss,

Ricardo Palos-Pacheco, and Dillon Hanrahan. To my friends and colleagues in the department for stimulating scientific (and unscientific) discussions, especially Stephanie

Jensen, Robb Bagge, Phil Dirlam, Dr. Jared Griebel, Dr. Dallas Matz, and Andrew

Vanschoiack. To my committee members and other members of the faculty and staff of

CBC for guidance and belief in my abilities. To my collaborators, including Dr. Jean .

Bidlack and Brian I. Knapp in the Department of Pharmacology and Physiology at the

University of Rochester for performing the opioid receptor binding studies of the SAA glycopeptides; Dr. Edward J. Bilsky and Lindsay St. Louis in the Department of

Biomedical Sciences at the University of New England for the antinociception studies of the SAA glycopeptides; Dr. John Streicher and Katie Edwards in the Department of

Biomedical Sciences the University of New England and the Department of

Pharmacology at the University of Arizona for the secretin family receptor activity data;

Dr. Michael Heien, Dr. Nicholas Laude, and Catherine Kramer in Chemistry and

Biochemistry at the University of Arizona for the tandem MS analysis of peptides and

4 glycopeptides, serum degradation experiments, and BBB penetration studies; Dr. John

Konhilas, Dr. Meredith Hay, and Eleni Constantopoulos in the Department of Physiology at the University of Arizona for the preliminary Angiotensin-(1-7) NO release/Mas receptor activation and murine memory experiments; and Dr. Todd Vanderah and Dr.

Tally Largent-Milnes in the Department of Pharmacology at the University of Arizona for the experiments testing the effects of Ang-(1-7) and derivatives in a cancer-induced bone pain model; and to all for giving any number of compounds a chance and having the patience to explain their respective fields to an outsider. And to my family for unequivocal support and belief regardless of what I decided to do.

Thank you.

DEDICATION

To Ernest Hibbard, my friend and mentor for instilling in me a confidence in my ideas, and teaching me that the best goals are slightly unrealistic. Your battle with ALS was as inspiring as it was heart-breaking.

Contents

List of Figures ...... 12

List of Schemes ...... 20

List of Tables ...... 21

List of Abbreviations ...... 23

Abstract ...... 26

1. Introduction ...... 27

1.1. Pain ...... 27

1.2. Neuroprotection ...... 28

1.3. Peptides as neurodrugs ...... 30

1.3.1. Challenges facing the use of peptides as neurodrugs ...... 30

1.4. G Protein-coupled receptors ...... 32

1.4.1. Rhodopsin-class GPCRs ...... 32

1.4.2. Opioid GPCRs ...... 32

1.4.3. Secretin family receptors ...... 35

1.5. Glycosylation of endogenous peptides to improve drug-like activity & the role of

amphipathicity ...... 39

1.5.1. Methods to reduce enzymatic degradation ...... 39

1.5.2. Glycosylation ...... 40 7

1.6. Vasoactive intestinal peptide ...... 45

1.6.1. History and functions...... 45

1.6.2. Receptors and binding ...... 46

1.6.3. VIP in neuroprotection ...... 46

1.6.4. Modifications to improve drug-like activity ...... 49

1.7. Pituitary adenylate cyclase-activating peptide ...... 52

1.7.1. History and functions...... 52

1.7.2. Receptors and binding ...... 52

1.7.3. PACAP in neuroprotection ...... 53

1.7.4. Modifications to improve drug-like properties ...... 54

1.8. Angiotensin-(1-7) ...... 54

1.8.1. History and functions...... 54

1.8.2. Ang-(1-7) as a treatment for postoperative cognitive impairment and stroke . 57

2. Development of novel sugar-amino acids ...... 59

2.1. Design rationale...... 59

2.2. Synthesis of novel SAAs ...... 63

2.2.1. 1-(Methylcarboxyl)-2,3-O-isopropylidene-5-amino-α-D-ribofuranose ...... 64

2.2.2. 1-Carboxyl-2,3-O-isopropylidene-6-amino-α-D-sorbofuranose ...... 65

2.3. Synthesis of DAMGO-SAA glycopeptides...... 66

2.3.1. Solid-phase peptide synthesis of glycopeptides ...... 66

2.4. Biological testing of DAMGO-SAA glycopeptides ...... 69

2.4.1. Binding assays to opioid receptors ...... 69

2.4.2. Antinociceptive effects using murine tail-flick test ...... 71

2.4. Discussion ...... 72

2.5. Materials and methods ...... 74

2.5.1. Sugar-amino acid synthesis ...... 74

2.5.2. Peptide synthesis...... 82

2.5.3. Radioligand-binding studies ...... 84

2.5.4. Animal and antinociceptive testing ...... 85

3. Design and synthesis of glycopeptide analogs of the secretin peptides VIP and

PACAP ...... 87

3.1. Molecular modeling of secretin peptides ...... 87

3.2. Initial syntheses of VIP and PACAP glycopeptides ...... 88

3.3. Oxidation of methionine residues ...... 90

3.4. Aspartimide formation ...... 92

3.5. Materials and methods ...... 95

3.5.1. Peptide synthesis...... 95

4. Secretin glycopeptide analysis and biological activity ...... 99

4.1. Receptor activation analysis ...... 99

4.1.1. Background ...... 99

4.1.2. Results: PAC1R agonists and antagonists ...... 100

4.1.3. Results: VPAC1 and VPAC2 agonists ...... 103

4.2. PACAP and VIP analog degradation & BBB penetration using MSn analysis ... 106

4.2.1. Background ...... 106

4.2.2. Collision-Induced Decay ...... 106

4.2.3. Serum degradation ...... 110

4.3. Discussion ...... 111

4.3.1. PAC1R, VPAC1, and VPAC2 receptor agonism and antagonism ...... 111

4.3.2. Peptide fragmentation and degradation ...... 115

4.4. Materials and methods ...... 119

4.4.1. Secretin receptor cell lines and cell culture ...... 119

4.4.2. Fluorescent Imaging Plate Reader (FLIPR) experiments ...... 119

4.4.3. Peptide degradation ...... 121

5. Design, synthesis, and analysis of angiotensin(1-7) glycopeptide analogs ...... 123

5.1. Design rationale and synthesis ...... 123

5.2. In vitro and in vivo results – cognition model ...... 126

5.2.1. NO production via Mas receptor activation ...... 126

5.2.2. GPCR activation via cell impedence measurements ...... 128

5.2.3. CID and BBB penetration experiments ...... 129

5.2.4. Memory and object recognition experiments ...... 134

5.3. Ang-(1-7) in peripheral neuropathy ...... 136

5.4. The pseudo-proline motif ...... 138

5.4. Materials and methods ...... 143

5.4.1. Peptide synthesis...... 143

5.4.2. Mas receptor activation ...... 146

5.4.3. Cell impedence studies ...... 147

5.4.4. Congestive heart failure model of memory loss in mice ...... 147

5.4.5. Blood-brain barrier penetration ...... 153

5.4.6. Cancer-induced bone pain model ...... 154

6. Summary and future directions...... 157

6.1. Synthesis of cysteine thioglycosides ...... 157

6.2. SAA incorporation into selective MOR agonist DAMGO ...... 159

6.3. Glycosylation of the secretin family peptides VIP and PACAP ...... 159

6.4. Glycosylation of Ang-(1-7) ...... 162

6.5. Peptide and glycopeptide detection and degradation analysis by MSn ...... 162

7. Appendix A. Calculation of the Connolly surface area of glycopeptides ...... 164

8. Appendix B. Supplementary Data ...... 167

9. Appendix C. Copyright Permissions ...... 189

Bibliography ...... 193

List of Figures

Figure 1-1. Overview of pain pathophysiology and pain targets.4 CB: cannabinoid; COX: cyclooxygenase; NMDA: N-Methyl-D-aspartate...... 28

Figure 1-2. SWOT chart highlighting benefits and drawbacks of peptide drugs.18 ...... 31

Figure 1-3: Representation of peptide ligand-GPCR interactions. First, a peptide ligand must associate with the membrane, adopting a helical conformation if energetically favorable. It then performs a two-dimensional search until it is able to interact with the receptor, shown here by insertion of the message into the TM domain...... 34

Figure 1-4: Small molecule agonist bound to MOR. The grey surface map represents potential sites of ligand-receptor interactions (PDB: 5C1M)35,36...... 34

MOR (PDB: 4DKL). 39-43 The grey surface around PACAP6-38 represents the putative receptor surface area necessary for ligand-receptor interactions. Model generated using

MOE software package...... 37

Figure 1-6: (A) Representation of the effects of peptide stapling, both in stabilizing α- helical conformation and decreasing the accessibility of the peptide backbone, and (B) a common amino acid used in peptide stapling...... 40

Figure 1-7. Proposed model of mechanism for peptide drugs to cross membranes.39,85

Courtesy Dr. Bobbi Anglin...... 42

Figure 1-8: The proposed membrane hopping mechanism for amphipathic peptides.

Peptides which are largely lipophilic will associate most strongly with a membrane. As

12 the size of the hydrophilic moiety increases, a peptide will have additional affinity for the aqueous surrounding environment, spending more time (and distance) away from the membrane, leading to longer hops...... 43

Figure 1-9: Roles of VIP in neuroprotection.90 VIP affects neuroimmunomodulation in a variety of ways. (A) Delivery by the peripheral nervous system (PNS) to primary and secondary immune organs affects cells through vasoactivation, and the chemotaxis of resting T cells. In addition, it acts as an immunosuppressive by reducing cytokine production and apoptosis. (B) Central delivery to a range of tissues (MALT: mucosa associated lymphoid tissue) increases activation of VPAC1 signaling (leading to chemotactic and immunosuppressive responses) and VPAC2 signaling (producing vasoactivation) to promote the localization of naïve T cells in organized lymphoid modules, including Peyer’s patches...... 48

Figure 1-10: Protective effects of Ang-(1-7) on stroke.147 ...... 58

Figure 2-1: DAMGO analogs (red) with increasing size of hydrophilic moiety (Gly-ol,

Ser(OH), Ser(OXyl), Ser(OGlc), Ser(OLac).1 ...... 59

Figure 2-2: Proposed SAAs (1, 2) and DAMGO-SAA derivatives (3-5) using Fmoc-based

SPPS synthesis conditions. Glycopeptide 3′ was not observed...... 62

Figure 2-3: HPLC trace of crude (A) and pure (B) DAMGO-SAA analog 3...... 67

Figure 2-4: HPLC traces of crude mixture of DAMGO SAA analogs 4 and 5 (A), and purified 5 (B) and 4 (C)...... 67

Figure 2-5: Conversion of 4 to 5 using (A) 4:1 H2NNH2•H2O:MeOH, (B) 95% AcOH/5%

H2O, 9h, (C) 50% AcOH/50% H2O, 16h. Note that (B) uses a different HPLC method, resulting in a shifted retention time...... 68

Figure 3-1: Example of the difference in ligand-receptor interactions (hatched grey surface) between class A (left) and class B (right) GPCRs. The latter requires a much larger area of the receptor surface to lead to receptor activation, so small molecule agonists like those used in the class A receptors will not suffice. (PDBs: 5C1M, 2JOD,

4DKL 35,36,39-43) ...... 88

Figure 3-2: HPLC trace of native VIP28 (12) prepared by automated synthesis. The arrow indicates the desired peak...... 90

Figure 3-3: Crude HPLC traces of compounds 22 (A), 23 (B), and 24 (C) prepared via automated synthesis. Amino acids 1-9 were deprotected using piperazine/HOBt. The arrow indicates the desired peak in each trace...... 92

Figure 3-4: Additives to reduce aspartimide formation. pKa represents accepted values for the conjugate acids of DBU, piperidine, and piperazine...... 95

Figure 4-1: Agonist activity curves of PACAP derivatives at the PAC1R, measured by

FLIPR. The native and glucosylated versions show good agonism, while the truncated forms show no activity. Courtesy of Katie Edwards...... 101

Figure 4-2: Antagonist activity of PACAP derivatives at the PAC1R, measured by

FLIPR. A. Variable concentration antagonism experiments with all proposed antagonists. B. Fixed concentration antagonism with PACAP6-38. C. Fixed concentration antagonism with PACAP6-27. Courtesy of Katie Edwards...... 102

Figure 4-3: Agonist activity of VIP28 and derivatives at the VPAC1 receptor using

FLIPR assay. Courtesy Katie Edwards...... 104

Figure 4-4: Agonist activity of VIP28 and derivatives at the VPAC2 receptor using

FLIPR assay. Courtesy Katie Edwards...... 105

Figure 4-5: CID fragmentation data for VIP28 (M17L) (22). (A) shows a schematic of the observed major fragmentation locations of the peptide. (B) has the major fragmentation peaks of the peptide at +4 and +5 charge states on the left side, and relative presence of various fragments (where the major fragment at each energy level is set at 100) on the right. Courtesy Catherine Kramer...... 107

Figure 4-6: CID fragmentation data for VIP28-S* (M17L) (23). (A) shows fragmentation occurs exclusively around the C-terminus. (B) shows the fragmentation states at +4 and

+5 in terms of overall signal (left) and relative quantities (right). Courtesy Catherine

Kramer...... 108

Figure 4-7: CID fragmentation data for VIP28-S** (M17L) (24). (A) shows fragmentation locations along the glycopeptide. (B) shows the fragmentation states at +4 and +5 in terms of overall signal (left) and relative quantities (right). Courtesy Catherine

Kramer...... 109

Figure 4-8: Schematic representing the procedures to determine degradation rates (top) and BBB penetration (bottom) of peptides...... 110

Figure 4-9: Serum degradation of PACAP27 and two glycosylated analogs (A). Half-life of each is shown in B...... 112

Figure 4-10: Representation of the hydrophilic (blue) and hydrophobic (red) areas of both

VIP28-S* and SDS. Note that the red portion of the peptide adopts a helical conformation...... 118

Figure 5-1: Ang-(1-7) derivative crude (left) and pure (right) HPLC traces. Red arrows show desired peak in crude, and percentages represent approximate purity based on UV

15 absorbance. Note that traces for 39 are from a separate experiment. S* = Ser(OGlc), S**

= Ser(OLac)...... 125

Figure 5-2: Structure of DAF-FM diacetate...... 126

Figure 5-3: (A) Relative fluorescence produced by the exposure of HUVEC cells to Ang-

(1-7) and derivatives. (B) Concurrent introduction of selective Mas antagonist A779 with

A5 (39) eliminates NO production, showing that 39 is acting at the Mas receptor. Also compared with NO donor SNAP (5 µM). Courtesy Dr. John Konhilas and Dr. Meredith

Hay...... 127

Figure 5-4: (A) HUVEC expression of the Mas receptor, shown by Western blot. (B)

Administration of vehicle and three concentrations of 39 to HUVEC cells, measured by

RTCA. The 10 pM dose shows 3-fold CI response compared with the 1 pM dose. (C)

Graphic comparison of CI response of 1 pM and 10 pM administrations of 39 after 10 minute incubation. Courtesy Dr. John Konhilas and Dr. Meredith Hay...... 129

Figure 5-5: CID fragmentation data for Ang-(1-7)-OH (34). (A) shows a schematic of the observed major fragmentation locations of the peptide. (B) has the major fragmentation peaks of the peptide at the +2 charge state on the left side, and relative presence of various fragments (where the major fragment at each energy level is set at 100) on the right. Courtesy Catherine Kramer...... 130

Figure 5-6: CID fragmentation data for Ang-(1-7)-NH2 (35). (A) shows a schematic of the observed major fragmentation locations of the peptide. (B) shows the fragmentation states at +2 in terms of overall signal (left) and relative quantities (right). Courtesy

Catherine Kramer...... 131

Figure 5-7: CID fragmentation data for Ang-(1-6)-S(OGlc) (39). (A) shows a schematic of the observed major fragmentation locations of the peptide. (B) shows the fragmentation states at +2 in terms of overall signal (left) and relative quantities (right).

Courtesy Catherine Kramer...... 132

Figure 5-8: Concentration of 39 in rats (n = 8, 4 per experiment) 20 minutes after s.c. administration of glycopeptide. The red (left) bars represent a 1.0 mg/Kg dosage, and the blue (right) represent a 10.0 mg/Kg dosage. Concentration of the drug is >10 times higher in CSF than serum (log scale), suggesting that glycopeptide penetration of the BBB leads to slower in vitro clearance. Courtesy Catherine Kramer and Dr. Michael Heien...... 133

Figure 5-9: Results of a Novel Object Test (A) and Morris Water Maze (B) on sham mice, and those subjected to induced congestive heart failure (CHF) and dosed with either saline or glycopeptide 39 (PN-A5). In (A), a negative score represents memory impairment, and a positive score intact memory recognition. In (B), cumulative integrated path length (CIPL) represents total distance a mouse travels to reach the end of the maze, with lower values representing better spatial memory. Courtesy Dr. John Konhilas and

Dr. Meredith Hay...... 135

Figure 5-10: Comparisons of sham and myocardial infarction (MI) mouse hearts were performed. (A) images of a normal and CHF mouse heart. (B) M-mode electrocardiography images, from which dimensions of ventricular chambers were calculated. (C) Ejection fraction, or cardiac activity for normal and MI mice. 39 (PNA5) does not affect heart function. Courtesy Dr. John Konhilas and Dr. Meredith Hay...... 136

Figure 5-11: The effect of Ang-(1-7) and 39 (AT5/ATn5) on cancer-induced bone pain based on measurements of (A) flinching and (B) guarding. (C) shows ability of peptides

17 to treat mechanical allodynia. In general, Ang-(1-7) appears to have a longer duration of effect than 39...... 138

Figure 5-12: Location of the proposed pseudo-proline motif in Ang(1-6)-S*. The interaction of the proton on the internal amide with the oxygen in the glycosidic linkage provides stabilization, and mimics the shape of a proline residue...... 139

Figure 5-13: Proposed conformation of both an α- and β-serine glycoside pseudoproline

(internal hydrogen bond is dashed). Note that the β carbohydrate has free, unhindered rotation around the indicated bond, while rotation of the α-glycoside is constrained by the peptide...... 141

Figure 5-14: Optimal angles of a carbohydrate conjugated to a peptide associated with a membrane prior to receptor docking. The identity of the amino acid the carbohydrate is bound to will affect the angle between carbohydrate and membrane. Courtesy Dr. Robin

Polt...... 142

Figure 5-15: Membrane-bound conformation of glycopeptide 39 at a membrane, based on constraints from NMR data in the presence of d25-SDS micelles. Modelling was performed using MOE® software and an AMBER-99 force field. Lipophilic residues are colored blue, and hydrophilic residues red, and illustrate the amphipathicity of the folded structure. Note the carbohydrate is held almost perpendicular from the membrane.

Courtesy Dr. Robin Polt...... 143

Figure 7-1: Energy-minimized schematic of DAMGO-Ser(OGlc) in MOE, with the peptide assigned as lipophilic (blue) and the Ser(OGlc) as hydrophilic (red). The ASA was calculated for each region independently using a full molecule surface area...... 165

Figure 7-2: Energy-minimized schematic of DAMGO-S(Glc) in MOE, with the peptide assigned as lipophilic (blue) and the S(OGlc) as hydrophilic (red). The ASA of each is assigned and calculated separately...... 165

Figure 8-1: Full normalized data for PACAP degradation studies. Experiments (x#) are presented in the same color, and each compound in the same shape. PACAP27 (6, ♦),

PACAP26-S(OGlc) (7,■), PACAP26-S(OLac) (8,▲)...... 167

Figure 8-2: VIP28 (12) and PACAP27 (6) calculated agonist activity at the VPAC1 receptor, measured by FLIPR. Similar activity is noted, although n = 1 for this experiment. Courtesy Katie Edwards...... 168

List of Schemes

Scheme 1-1: Proposed reaction scheme for Lewis acid (L.A.) catalysis of nucleophile

(Nu) glycosylation with a peracetylated carbohydrate. First, the L.A. stabilizes loss of the anomeric acetate, resulting in the neighboring-group stabilized glycosyl cation. Finally, the nucleophile (Nu) attacks the activated anomeric position...... 45

Scheme 1-2. In vivo production of Ang-(1-7) and its interaction with the MAS1 receptor.

...... 56

Scheme 2-1: Synthetic scheme for the formation of SAA 1...... 65

Scheme 2-2: Synthetic scheme for the formation of SAA 2...... 66

Scheme 3-1: Formation of an aspartimide can occur under acidic or basic conditions.

Once formed, the ring can be opened by a nucleophile at either carbonyl, leading to a) the peptide chain converting to an isoaspartyl-β-peptide, or b) to the nucleophile replacing the X group with the desired peptide backbone...... 93

Scheme 4-1: Proposed degradation of PACAP26-S** into PACAP26-S* via galactosidase action...... 117

Scheme 6-1: Synthesis of four cysteine peracetate glycosides using catalytic InBr3. Glc = glucose, Mal = maltose, Lac = lactose, Cel = cellobiose...... 158

Scheme 6-2: Substitution of classic Asp protecting group tert-butyl to the newly reported

5-n-butyl-5-nonyl group235 is expected to help minimize Asi formation...... 161

List of Tables

Table 1-1. Receptors and ligands for selected secretin-family GPCRs ...... 38

Table 2-1: Binding affinities of the DAMGO-SAA glycopeptides at the MOR, DOR, and

KOR...... 70

Table 2-2: I.v. and i.c.v. antinociceptive A50 values in mice for the DAMGO-SAA glycopeptides, acquired via tail-flick test...... 72

Table 3-1: Identity of initial PACAP (6-11, synthesized by Dr. Bobbi Anglin) and VIP analogs, where the full sequences were designed as receptor agonists, and the N-terminal truncated sequences as antagonists. S* = Ser(OGlc), S** = Ser(OLac)...... 89

Table 3-2: Second generation secretin family glycopeptides, with Leu17 or Nle17

(underlined) in place of the oxidizable methionine residue. S* = Ser(OGlc), S** =

Ser(OLac), £ = Nle...... 91

Table 4-1: Calculated agonist activity for PACAP derivatives at the PAC1R. NC = not converged, meaning no activity was observed...... 101

Table 4-2: Calculated antagonist activity at the PAC1R using variable or fixed concentration experiments...... 103

Table 4-3: Calculated agonist activity of VIP28 and analogs at the VPAC1 receptor using

FLIPR assay...... 104

Table 4-4: Calculated agonist activity of VIP28 and analogs at the VPAC2 receptor using

FLIPR assay...... 105

Table 5-1: Initial batch of Ang(1-7) analogs proposed and synthesized. S* = Ser(OGlc),

S** = Ser(OLac)...... 124

Table 7-1: Comparison of calculated Connolly amphipathicity values using published method (LYM Method) and my modified method (EMJ Method). Amphipathicity for compounds 3-5 was not determined using the LYM Method...... 166

List of Abbreviations

ACE angiotensin-converting enzyme ACE2 angiotensin-converting enzyme II AcOH acetic acid AD Alzheimer’s disease ALS amyotrophic lateral sclerosis

Ang(1-7) angiotensin(1-7): Asp-Arg-Val-Tyr-Ile-His-Pro-NH2 ASA water-accessible surface area Asi aspartimide Asp aspartic acid

AT1 angiotensin subtype I receptor

AT2 angiotensin subtype II receptor BBB blood-brain barrier βGal β-D-galactose βGlcNac β-N-acetyl-D-glucosamine cAMP cyclic adenosine monophosphate CHF congestive heart failure CHO Chinese hamster ovary CI cell impedance CIBP cancer-induced bone pain CID collision-induced decay CMC critical micelle concentration CNS central nervous system COX cyclooxygenase CRF corticotropin-releasing factor CSF cerebrospinal fluid DAMGO Tyr-ala-Gly-(NMe)Phe-Glycinol DBU 1,8-diazabicyclo[5.4.0]undec-7-ene DMPK distribution, metabolism, and pharmacokinetic effects DOR delta (δ)-opioid receptor

DPP-IV dipeptidyl peptidase-IV DPPC dipalmitoylphosphatidylcholine ECD extracellular NT (N-terminal) domain FLIPR Fluorescence Imaging Plate Reader Fmoc fluorenylmethyloxycarbonyl, a protecting group in SPPS g.u. glucose units GPCR G protein-coupled receptor HD Huntington’s disease HOBt hydroxybenzotriazole HUVEC human umbilical vein endothelial cell i.c.v. intracerebroventricular i.p. intraperitoneal i.v. intravenous KOR kappa (κ)-opioid receptor Leu leucine Met methionine MI myocardial infarction MOE Molecular Operating Environment™ MOR mu (µ)-opioid receptor MS mass spectrometry MSn tandem mass spectrometry MWM Morris Water Maze test Nle norleucine NMDA N-methyl-D-aspartate NMR nuclear magnetic resonance NO nitric oxide NOR Novel Object Recognition test NT N-terminal NVU neurovascular unit PAC1R adenylate cyclase-activating polypeptide 1 receptor 1a PACAP pituitary adenylate cyclase-activating peptide

PD Parkinson’s disease RAS renin-angiotensin system ROS reactive oxygen species RTCA Real-Time Cell Analyzer S* Ser(OGlc) – serine O-glucoside S** Ser(OLac) – serine O-lactoside s.c. sub cutaneous s.n. substantia nigra SAA sugar-amino acid SPPS solid-phase peptide synthesis TBI traumatic brain injury TFA trifluoroacetic acid TM transmembrane VIP vasoactive intestinal peptide VPAC1 VIP and PACAP receptor 1 VPAC2 VIP and PACAP receptor 2

Abstract

Endogenous peptides modulate a wide range of physiological conditions in the central and peripheral nervous systems, but have not been harnessed to perform similar functions in pharmaceutical roles due to their ease of degradation and difficulty in introducing into the neurovascular unit. We report herein advances that evidence the wide applicability that glycosylation provides as a pathway for improving the drug-like properties of peptides. This is demonstrated by utilizing novel sugar-amino acids to modify the potent mu opioid receptor agonist DAMGO to provide antinociception, and serine glycosides to modify secretin family peptides for neuroprotection and angiotensin-(1-7) to both reduce cognitive impedance following myocardial infarction and as a treatment for peripheral neuropathy. Evidence is presented via a series of in vitro and in vivo models and assays, and demonstrates the advantageous effects of glycosylation through increased persistence in serum, greatly improved blood-brain barrier penetration, and the tolerance of receptor interactions to the addition of a carbohydrate.

1. Introduction

1.1. Pain

Pain is a major component of the human experience. Improved diets, medical care, and understanding of human physiology have produced a concurrent increase in global life expectancy, and ever more people experiencing some form of long-term, or chronic, pain. According to a recent study, around 40% of the global adult population experiences chronic pain caused by diseases, disorders, or injuries.2 In the United States of America alone, in 2011 the National Institute of Neurological Disorders and Stroke

(NINDS) estimates that 100 million adults currently experience chronic pain, and $560 billion to $635 billion is spent or lost annually dealing with pain.3

Treatment of pain is varied, due to the number of ways in which the perception of pain can be triggered in the human body (Figure 1-1).4 Methods to treat chronic pain include a variety of therapies, including physical therapy and exercise, chiropractic care, electrical stimulation, and surgery, but by far the most common way is the use of pharmaceuticals.4 Of these, opioids are some of the most widely prescribed drugs, of which morphine is the best known. However, the negative side effects of many opioids, ranging from sedation and drowsiness to nausea and vomiting to addiction and potentially dangerous withdrawal symptoms, have prevented their widespread adoption for daily use.

A number of non-opiate targets, including ion channels and cyclooxygenase (COX) receptors and N-methyl-D-aspartate (NMDA) receptors have also emerged as targets for

“non-narcotic” approaches to the treatment of pain.

Figure 1-1. Overview of pain pathophysiology and pain targets.4 CB: cannabinoid; COX: cyclooxygenase; NMDA: N-Methyl-D-aspartate.

The central nervous system (CNS), which includes the brain and spinal cord, controls and regulates messages throughout the human body.5 When this system is damaged, which can be caused by injuries such as stroke, multiple sclerosis, and spinal cord trauma, many common pain treatments become ineffective. For example, in one study adults, who suffered from damage to the CNS and experienced neuropathic pain, were treated with the morphine-like opioid agonist levorphanol. Levorphanol provided an orally-administered pathway for palliative pain relief.6 However, it was noted that a greater proportion of patients receiving higher doses of levorphanol dropped out of the program before its completion due to the drug’s severe side-effects.

1.2. Neuroprotection

Neurodegeneration, or the gradual loss of neuronal structure or function, affects many fewer people than chronic pain, but is still a significant obstacle for many people as 28 they age.7,8 Neurodegeneration can be caused by a wide range of pathologies, including

Alzheimer’s disease (AD), Parkinson’s disease (PD), amyotrophic lateral sclerosis (ALS, or Lou Gehrig’s disease), and Huntington’s disease (HD).9 It can also be caused by specific insults to the brain, including stroke, concussion, or traumatic brain injury (TBI).

In each of these instances, neuronal death, or apoptosis, is the root problem for the varied symptoms that the diseases elicit.9,10

In contrast, neuroprotection can maintain neuronal structure and function. In the case of neurodegenerative insults, neuroprotection typically refers to the relative limitation or complete elimination of the loss of brain function.11 One of the principle reasons why neuroprotection is viewed as an attractive target for treating neurodegeneration is because most neuronal aptosis occurs by one of a few common mechanisms.9 These include increases in oxidative stress through the release of reactive oxygen species (ROS),12 inflammation, and mitochondrial disturbances,13 among others.

For instance, although PD is symptomatically characterized through a sufferer’s impairment of motor skills, posture, speech patterns, and bradykinesia, it is clinically identified as the loss of dopamine neurons in the substantia nigra section of the brain.14

However, small molecules that target oxidative stress, inflammation, and mitochondrial dysfunction have all been shown to provide neuroprotective benefits.14

In strokes, the infarct location is quickly put into an anaerobic state, eventually resulting in the formation of an ischaemic penumbra, where blood flow and oxygen transport is greatly reduced.8,15 This, in turn, induces an ischaemic cascade, which includes the production of ROS and oxygen radicals, as well as inflammatory responses, leading to further neuronal apoptosis.16

1.3. Peptides as neurodrugs

Endogenously derived peptides modulate a wide range of activities and signaling in the brain. In addition, they have potent and highly selective interactions with their associated receptors, and the human body possesses numerous pathways for peptide degradation into relatively non-toxic metabolites, minimizing residual receptor activation and long-term in vivo drug persistence. For these reasons, they are an appealing potential source of drugs. Recently, these advantages have been recognized. Currently there are over 60 US Food and Drug Administration (FDA)-approved peptide therapeutics available to the public, in addition to around 140 in clinical trials, and more than 500 peptides in preclinical development.17,18

1.3.1. Challenges facing the use of peptides as neurodrugs

However, the use of peptides or peptide derivatives as neurodrugs has some significant obstacles related to their distribution, metabolism, and pharmacokinetic effects (DMPK).19-21 As summarized by Hoffman and Fosgerau (Figure 1-2), peptides are susceptible to chemical and physical degradation, oxidation, and aggregation.18 In addition, the multitude of chemical degradation pathways means that peptides generally have short half-lives in vivo.

Figure 1-2. SWOT chart highlighting benefits and drawbacks of peptide drugs.18

An additional concern is penetration of the blood-brain barrier (BBB), a membrane at the interface of the brain and circulatory system.22 The BBB protects the brain from the passive diffusion of hydrophilic molecules from the bloodstream while allowing vital nutrients passage and there are few, if any, diagnostic tests for determining likelihood of BBB penetration for individual compounds. Typically, pharmaceutical companies design neurodrugs for BBB penetration using the observations made by

Lipinksi in his well-known “rules of 5”,23 although an earlier hypothesis that linked lipophilicity measurements with BBB penetration has also been shown to be useful in some instances.24 Most endogenous neuropeptides “break” every one of the rules of 5, yet are still highly efficacious even when produced outside the neurovascular unit (NVU).

Penetration of the BBB is essential for drugs targeted for the CNS; without some

31 mechanism of passage, the only way to administer drugs is direct intracerebroventricular

(i.c.v.) injection, which limits their usage to life-threatening situations.

1.4. G Protein-coupled receptors

G protein-coupled receptors (GPCRs) are cell surface molecules, and the target of choice for an estimated 30-40% of all drugs.17,25 Over 800 human GPCRs have been identified to date, and their accessibility and wide range of physiomodulatory effects have made them attractive pharmaceutical targets.26

1.4.1. Rhodopsin-class GPCRs

The rhodopsin family of GPCRs, or GPCR-A family, are by far the most prevalent: according to the GPCR Network at Scripps College they represent almost 90% of the GPCRs identified to date.27 The structure of rhodopsin when in its G protein-active configuration is typical of most GPCRs, with a single polypeptide folded into seven transmembrane helices that form a pocket in which agonists can bind. However, the N- terminal (NT) extracellular segment is fairly short compared with other types of GPCRs, and does not appear to be involved in ligand binding.

1.4.2. Opioid GPCRs

The rhodopsin family also contains the three opioid receptors: δ-opioid receptor

(DOR), µ-opioid receptor (MOR), and κ-opioid receptor (KOR). The opioid receptors have been utilized for millennia for their well-known antinociceptive role in pain relief, with the most familiar agonists being opiates derived from opium poppies and morphine.

They were first differentiated in 1977,28 and have been a popular system to study in the years since then. In the intervening years, various peptide agonists have been prepared for

32 each of the opioid receptors, including the enkephalins (MOR and DOR ligands), dynorphins (KOR, MOR, and DOR ligands), and endorphins (MOR, DOR, and KOR ligands).29 Interestingly, all of these endogenous ligands possess identical N-terminus sequences, as each begins with the tetrapeptide YGGF.30,31 This has been termed the

“message” section of the peptide, as it triggers receptor response.31 The remainder of each peptide is considered the “address” segment, as it determines receptor selectivity.

For instance, Leu-enkephalin (YGGFL) and Met-enkephalin (YGGFM) are nearly identical, yet the former has higher activity at the DOR and the latter at the MOR. Thus, the leucine and methionine residues can be thought of as directing each peptide for specific receptors.

In addition, the “address” sequence in many of the opioid peptides adopts a helical conformation in cell membrane-like environments.32-34 It is hypothesized that this stabilizing structural change leads peptides to more efficient localization of receptors; instead of randomly floating through the environment surrounding a cell, performing a

3D search for a receptor, the peptide is able to perform a more effectual 2D search of the membrane (Figure 1-3).30

Structurally, the opioid receptors closely match that of rhodopsin (Figure 1-4,

PDB: 5C1M).35,36 For receptor activation binding, a small tripeptide sequence (E/DRY) has been identified as essential at the junction of the third transmembrane (TM) domain and its intracellular loop.37 As this is a relatively small area, many small molecule agonists have been developed for the morphine receptors.

Figure 1-4: Small molecule agonist bound to MOR. The grey surface map represents potential sites of ligand-receptor interactions (PDB: 5C1M)35,36.

1.4.3. Secretin family receptors

The secretin family receptors and their respective ligands include glucagon, which has been well-studied and is known for increasing the concentration of glucose in the blood, as well as secretin, vasoactive intestinal peptide, pituitary adenylate cyclase- activating peptide, gastric inhibitory peptide, and growth hormone releasing factor.

Compared with the opioids, secretin (GPCR-B, or class II) family receptors are much less well understood. Fifteen secretin family receptors have been identified (for selected examples (see Table 1-1), whose myriad functions include regulating neuronal survival.38

GPCR-Bs maintain the same basic GPCR structure, with seven TM helices, but the secretins have a much larger extracellular NT domain (ECD), as represented by a homology model of the PAC1 receptor (PAC1R) (Figure 1-5, PDB: 2JOD, 4DKL).39-43

Interestingly, the ECD domain contains three pairs of cysteine residues that form disulfide bonds that, in turn, give shape to the ECD, and these cysteines are completely conserved in all GPCR-B receptors.38 Mutation of any one of the cysteines involved in disulfide bond formation in the VPAC1 receptor leads to inactivation, both through ligand binding and cyclic AMP (cAMP) production.44 In addition, their peptide ligands are all at least 27 residues long.

Based on binding analysis, it appears that most ligands interact with both the EC and the TM binding pocket,45 and that the ECD is responsible for ligand receptor selectivity.46 For instance, vasoactive intestinal peptide (VIP), when bound to the VPAC1 receptor, has two residues that are in contact with the receptor,47,48 and overall contact with the ECD represents 29% of the peptide’s total surface area.45 In addition, point mutation analysis has shown that the active site of the receptor includes at least four of

35 the seven TM domains.49 Therefore, a synthetic ligand for a class B GPCR must be large enough to simultaneously interact with both the ECD and TM domains. Further complicating development of novel ligands is the lack of any published complete, agonist-bound GPCR-B crystal structures, so the majority of hypothesized binding orientations between ligand and receptor have been performed using photoaffinity labeling, mutational assays, molecular modeling, or NMR.43 Recently, a crystal structure of glucagon bound to a truncated analog of its receptor was reported.50,51

Figure 1-5: Representation of the PAC1R using a crystal structure of the N-terminal extracellular domain with a PACAP6-38 (green) antagonist bound (PDB: 2JOD), aligned with a homology model of the PAC1R TM domain based on the crystal structure of a MOR (PDB: 4DKL). 39-43 The grey surface around PACAP6-38 represents the putative receptor surface area necessary for ligand-receptor interactions. Model generated using MOE software package.

Receptor Ligand Peptide Ligand Sequence (gene)

PAC1R PACAP HSDGI FTDSY SRYRK QMAVK KYLAA VLGKR YKQRV KNK VPAC1 (VIPR1) / VIP HSDAV FTDNY TRLRK QMAVK KYLNS ILN VPAC2 (VIPR2) Secretin Secretin HSDGT FTSEL SRLRE GARLQ RLLQG LV (SCTR) Glucagon Glucagon HSQGT FTSDY SKYLD SRRAQ DFVQW LMNT (GCGR)

Corticotrophin- CRF releasing factor SEEPP ISLDL TFHLL REVLE MARAE QLAQQ AHSNR KLMEI (CRH) (CRF) Table 1-1. Receptors and ligands for selected secretin-family GPCRs

The mechanism of GPCR-B activation by a ligand is not fully understood. It has previously been shown that multiple ligands, including pituitary adenylate cyclase- activating peptide (PACAP)43,52 and corticotropin-releasing factor (CRF)53 adopt an α- helical conformation prior to receptor binding. This suggests that, similar to the peptidic opioids, the secretins associate with the cell membrane before performing a 2D search of their local environment for suitable receptors. It is hypothesized that a two-step activation mechanism is present for the secretin ligands: first, the C-terminal “address” sequence binds to the EC domain, after which the NT “message” enters the TM binding pocket and activates the receptor.45 Additional selectivity is found in the binding pocket, as the active site for VIP was mapped to four amino acids that all need to be interacted with in order to induce the necessary conformational changes to provide activation.54

1.5. Glycosylation of endogenous peptides to improve drug-like activity & the role of amphipathicity

As previously discussed, utilizing neurologically active peptides as drugs faces a host of challenges, including rapid degradation by a range of peptidases and rapid renal clearance leading to short plasma half-life times, and difficulty penetrating the BBB.

1.5.1. Methods to reduce enzymatic degradation

Amide bonds in peptides are susceptible to enzymatic inactivation from a number of sources in the human body, including through conjugation to ubiquitin55 and degradation by enzymes including dipeptidyl peptidase-IV (DPP-IV) and other aminopeptidases, carboxypeptidases, and exopeptidases.56-59 While concurrent administration of an enzyme inhibitor with a therapeutic peptide is widely-utilized option,60 a recent meta-analysis of DPP-IV inhibitor use showed a statistically significant increase in the likelihood of heart failure, in addition to other inhibitor-dependent adverse side-effects.61

Common chemical modifications to improve the plasma stability of peptides include N-terminus acetylation, C-terminus amidation, cyclization, and the substitution of labile amino acids.59 Effort has also been made to affix bulky water-soluble groups, such as polyethylene glycol (PEG) or lipid-soluble fatty acids, which interfere with enzyme binding.59,62 However, such modifications must be performed with some understanding of how a peptide interacts with its desired receptor; in many cases simple transformations lead to large decreases in binding affinity and/or receptor activation.

Another recent development is peptide “stapling”.63 Similar to cyclization, which often involves the formation of a disulfide bond between two cysteine residues,

“stapling” involves the substitution of two or more residues with amino acids containing allyl-terminated R-groups, such as (S)-N-Fmoc-α-(4-Pentenyl)alanine, then performing a

Grubbs-type alkene metathesis on the resulting diallyl peptide (Figure 1-6). Often, the staple fixes part of the peptide in a specific conformation, such as a helix, that is not amenable to enzyme binding, thereby increasing the peptide’s i.v. half-life.

A B

1.5.2. Glycosylation

Appending a carbohydrate to a peptide has shown much promise for the improvement of DMPK for peptide neurodrugs.64 One of the earliest observations of the benefits of glycosylation for improving a peripherally administered drug’s central activity was via the modifications of morphine to morphine-6-glucuronide and morphine-3- glucuronide.65,66 Glycosylation has been demonstrated to increase peptide resistance to enzymatic degradation in a wide range of peptides, particularly if installed near a linkage susceptible to cleavage in a variety of systems, including small opioids,67-70 as well as larger natural peptides.71-75 In the intervening years, glycosylation has been shown to increase membrane penetration, and glycosylation has been used to improve peptide passage through the intestinal tract76 and BBB.34,39,77-81

The initially hypothesized mechanism of transport across membranes was via piggybacking onto carbohydrate transporters such as the glucose transporters GLUT-1 or

GLUT-2.77 While unpublished work performed in collaboration with the Polt group by

Dr. Thomas P. Davis (Pharmacology, The University of Arizona), later tested in mice by

Drs. Hruby and Davis82 seemed to rule out this hypothesis, as increasing the concentration of free glucose at a membrane did not affect the rate of transport,30 Toth and coworkers reported that glycosylated analogs of gonadotrophin-releasing hormone, a decamer peptide GPCR agonist, crossed a membrane model much more rapidly when

GLUT-1 and GLUT-2 inhibitors were not present.83 It is known that polycationic peptides traverse membranes via adsorptive transcytosis (

Figure 1-7),39 and it seems reasonable to assume that glycopepties may utilize a similar system. Some short peptides (4-5 residues) are transported by specific transporters, such as the Peptide Transport System-1 receptor.84 In adsorptive transcytosis, first the positively charged peptide associates with the negatively charged head groups in the membrane, then a vesicle forms around the peptide as it enters the intramembrane space, and finally the peptide is delivered to the other side.85 As a carbohydrate is hydrophilic, it seems reasonable to assume that glycosylation may enhance vesicle formation due to natural disruption of the hydrophobic bilayer. Interestingly, based on the range of peptide lengths that show an improvement to BBB penetration after the addition of a glucose moiety (from four amino acid DAMGO to 27 residue PACAP),1,39 it appears that the presence of the carbohydrate is more important than the relative size of the carbohydrate to the peptide.

Figure 1-7. Proposed model of mechanism for peptide drugs to cross membranes.39,85 Courtesy Dr. Bobbi Anglin.

The inclusion of a carbohydrate with a peptide may also modulate a peptide’s ability to find and bind a receptor, by affecting amphipathicity. As shown in Figure 1-3, the initial process necessary for receptor activation is for the ligand peptide to associate with the membrane. Generally, the membrane is considered relatively hydrophobic and the surround milieu polar, so one could model a peptide’s chances of finding a receptor looking at its relative affinity for either the membrane-bound or solution state. This can be modeled using Connolly surface area calculations based on the computationally determined surface areas of both the hydrophilic (Awater) and total (Atotal) surface area, where amphipathicity:

퐴 (− 푤푎푡푒푟) 퐴푚푝ℎ𝑖푝푎푡ℎ𝑖푐𝑖푡푦 = 푒 퐴푡표푡푎푙

In general, as amphipathicity approaches zero, the peptide’s hydrophilicity outweighs the beneficial energetics of association with the membrane (although it should be noted that

42 the minimum value for any compound using this equation is 0.37). At the other extreme, as amphipathicity approaches one, the hydrophobic effect becomes so important that the peptide is never likely to lift off of the membrane. This forces the peptide to have a relatively narrow 2D search for a receptor along a small membrane surface area.

Therefore, the theory of “membrane hopping” (Figure 1-8) was proposed, where the ability of a peptide to rapidly associate and dissociate with a membrane is essential to its function; if it associates too strongly the searchable surface area of membrane is limited, and if it dissociates too rapidly it will not spend enough time on or near a membrane to bind to the desired receptor.1,30 On a given lipophilic peptide, switching from a monosaccharide to a disaccharide would theoretically give you a lower A, and from a disaccharide to a monosaccharide would increase the A. Additionally, changing the linkages between disaccharides (from β to α, or from 1,4- to 1,6-linkages) can affect the shape, and therefore accessible surface area, for a given carbohydrate.

Figure 1-8: The proposed membrane hopping mechanism for amphipathic peptides. Peptides which are largely lipophilic will associate most strongly with a membrane. As the size of the hydrophilic moiety increases, a peptide will have additional affinity for the aqueous surrounding environment, spending more time (and distance) away from the membrane, leading to longer hops.

Analysis of these trends, termed “biousian” (bi – two, and ousia – essence) by

Polt and coworkers, has been explored.1,34 Their observations produced a U-shaped curve when i.v. activity and amphipathicity were plotted against each other for a series of

DAMGO (Tyr-ala-Gly-(NMe)Phe-Xxx) glycosides (See Chart 2-1 in 2.1. Design rationale for full explanation and for additional details).1 A U-shape fits the hypothesized activity nicely, and it shows a correlation between the size of the carbohydrate and its activity; too small of a carbohydrate presumably does not lift the peptide off of the membrane, and too large of one keeps it in the aqueous environment surrounding the targeted brain cells. However, this plot does not account for differences in binding or resistance to enzymatic degradation. The former was tested at the MOR, KOR, and DOR, and all the DAMGO analogs exhibited similar binding affinities within one order of magnitude of native DAMGO (and there is no recognizable correlation between binding affinity and activity, likely because the binding occurs at much lower concentration than that of activation), while differences attributable to the latter were not directly probed using this system.

In addition, the Polt lab has developed novel methods of amino acid glycosylation using weak Lewis acids.30,86 The basic premise behind this reaction is that the Lewis acid is strong enough to allow the anomeric ester to leave and for the desired nucleophile, such as serine, to attack that position, but weak enough that it does not form a bond with the leaving group, and can be reused (Scheme 1-10).

Scheme 1-1: Proposed reaction scheme for Lewis acid (L.A.) catalysis of nucleophile (Nu) glycosylation with a peracetylated carbohydrate. First, the L.A. stabilizes loss of the anomeric acetate, resulting in the neighboring-group stabilized glycosyl cation. Finally, the nucleophile (Nu) attacks the activated anomeric position. 1.6. Vasoactive intestinal peptide

1.6.1. History and functions

VIP is a 28-residue secretin-family endogenous peptide, first isolated from porcine lung tissue, with actions throughout the central and peripheral nervous system.87

VIP has been shown to modulate a wide range of pathologies, including diabetes, asthma, male impotence, the immune system, and neurological disorders.72,87-90 Measurements of

VIP in plasma has recently been reported to act as an early marker for arthritis,91 and the presence of receptors for a number of cancers.90,92

Elevated levels of VIP in the gastrointestinal tract, caused by a rare endocrine tumor known as a VIPoma, results in a condition known as Verner Morrison syndrome or pancreatic cholera.93,94 The exact basal level of VIP in human plasma is unclear,95,96 although i.v. infusions of VIP into rats over 7 days did not lead to observed detrimental side effects.96 Additionally, although plasma levels of VIP are hypothesized to be in the fM range, localized concentrations around receptors can be orders of magnitude higher without negative side effects.95,97

The structure of VIP has been highly conserved across a wide range of species, with complete sequence identity in all mammals except the guinea pig, and homologous sequences to animals including frogs, alligators, trout, and chickens.96,98

1.6.2. Receptors and binding

VIP has two endogenous receptors: VPAC1 and VPAC2 (VIP and PACAP receptor 1 and 2, respectively),99,100 although it also exhibits some activity towards the

PAC1 receptor.101 As with the other secretin family peptides, VIP is hypothesized to exhibit a two-stage binding mechanism, in which the C-terminal address binds to the receptor ECD, followed by the N-terminus inserting into the TM domain and producing receptor activation.

Both the full peptide (VIP1-28) as well as the N-terminal truncated VIP2-28 act as full agonists in human intestinal epithelial membranes, 102 and VIP1-28 has been widely observed as an agonist at both the VPAC1 and VPAC2 receptors. However, the only reported truncated antagonist that has been reported is the fragment VIP10-28, which showed antagonism in the human cancer cell line HT29.102,103 Significant antagonism was not noticed for VIP10-28 in other systems. Gozes and coworkers have developed an antagonist that combines the N-terminus of neurotensin with the C- terminus of VIP, and shown that it affects neuronal function in cells.104

1.6.3. VIP in neuroprotection

Early studies of VIP recognized the role that it plays in neuronal survival. 105,106

This led to groups showing that VIP exhibits neuroprotective properties88,107-109 through immunomodulatory effects, including by affecting the release of immunomodulatory cytokines IL-10110,111 and IL-6.112,113 In turn, this leads to decreased cerebral inflammation, which plays a major role in a number of diseases.98,109,114 The ability of

VIP to reduce microglial activation (through modulation of the small molecules that cause it) has provided initial evidence of mitigation of Parkinson’s disease in a murine

46 model.115 In addition, the expression and signaling of VIP is naturally reduced in a number of neurological disorders, which suggests that a disruption of the homeostatic effects provided by VIP contribute to the deleterious symptoms of the disorders.116 A later observation that the overexpression of α-synuclein leads to upregulation of VIP &

PACAP suggests that those secretin peptides are endogenous protective factors.117

Finally, recent work has indicated that VIP affects neuroprotection by reducing the activity of reactive oxygen species (ROS)-forming compounds, including NADPH oxidase.118 VIP also reduces ROS in kidney by promoting activity of the oncogene B-cell

CLL/lymphoma 2 (Bcl-2) protein.119 While this is not a neurological benefit, it suggests that VIP still has undiscovered pathways of neuroprotection. Due to the prevalence of

VIP in the brain and the variety of ways in which it affects neuroprotection (Figure 1-9),

VIP exists as an attractive therapeutic for a number of neurological targets.

1.6.4. Modifications to improve drug-like activity

As is true with many potential peptide drugs, development of treatments based on

VIP have been hampered by its rapid in vivo degradation,72 although not by DPP-IV due to the N-terminal His-Ser motif.120 However, due to its wide scope of potential uses, efforts have been made to improve VIP’s DMPK.

1.6.4.1. Glycosylation

Rocchi and coworkers performed a “glycoscan” on VIP, in which they synthesized eight analogs of VIP with either β-D-galactose (βGal) or β-N-acetyl-D- glucosamine (βGlcNac) installed at various positions along the backbone.72 Their observed results were mixed; all glycosylation sites reduced receptor affinity 4-2000 fold

(Chart 1-1), but one of the analogs ([11Glyc]VIP) was significantly more stable against trypsin degradation. In addition, they showed that the major trypsin degradation sites along the backbone were at the C-terminal side of Arg12, the C-terminal side of Arg14, and the C-terminal side of Lys20.

Still, this study lends credence to the binding hypothesis of VIP with its receptors; because the decreases in receptor affinity and potency were of a lower magnitude for modifications near the C-terminus of the peptide – opposite the “message” – this suggests that these analogs had fewer barriers to receptor binding. Although in membrane- mimicking environments the helical portion of VIP appears to extend to the C- terminus,121 work has shown that the majority of residues important for receptor ECD binding are closer to the center and N-terminus of the peptide.47,48,99,100

Effect of Glycosylation on VIP Receptor Interactions 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 Relative affinity 0.2 Relative potency 0.1 0

Chart 1-1: Rocchi and coworkers72synthesized a number of glycoside analogs of VIP, and determined their affinity in HT29 cells versus radiolabeled VIP to determine IC50 values, and potency by measuring cAMP flux to get EC50. Values in the chart above are normalized to the full affinity and potency of native VIP.

1.6.4.2. Peptide cyclization

Another series of modifications involves the cyclization of peptides. Giordanetto and coworkers performed two kinds of cyclizations of VIP: lactamization and olefin- methathesis stapling.122 They initially analyzed the sequence of VIP with relation to residues believed to have important receptor binding interactions, and predicted locations to install residues to form cyclic systems without interfering with binding. Overall, they reported mixed results. Some of the lactamized peptides had very similar activities at the

VPAC2 receptor, but none of the most active ones showed significant resistance against degradation by trypsin over 15 minutes. One of the stapled peptides was much more

50 active, but it, too, showed only a minimal improvement in resistance to trypsin degradation. Still, their preliminary results show that this may be a reasonable approach to improving peptide stability.

1.6.4.3. Incorporation into nanoparticles

A study by Rubinstein and coworkers reported that VIP spontaneously forms micelles with a critical micelle concentration (CMC) (the surface concentration at which aggregation of amphipathic monomers into micelles occurs spontaneously) at 0.4 µM in vitro, and that these micelles had a diameter under 10 nm.123 They also showed that when exposed to a membrane-mimicking monolayer of dipalmitoylphosphatidylcholine

(DPPC), VIP rapidly associated with the DPPC and eventually penetrated the membrane, leading to monolayer decomposition. From this work, they hypothesized that VIP’s activity in the human body is a mix of receptor activation and lipid bilayer disruption, although further work to probe this was not performed.

Other studies have attempted to link VIP to variously sized nanoparticles. For instance, Vukovic and coworkers attempted to non-covalently bind VIP to sterically- stabilized liposomes,124 but were stymied by rapid dissociation. However, covalent bonds between VIP and modified, polyethylene glycol (PEG)-conjugated phospholipids in micelles was successful.124,125 Although these VIP-micelle conjugates served their purpose (targeting breast cancer), this approach is limited for drug delivery as the peptide cannot readily dissociate from the micelle in order to cross membranes. To develop this further, VIP would need to be associated using chemically labile or enzymatically cleavable bonds.62

1.7. Pituitary adenylate cyclase-activating peptide

1.7.1. History and functions

Much like VIP, PACAP is an endogenous neuropeptide with a variety of functions found throughout the central and peripheral nervous systems, including modulation of bronchodilation, neurotransmitter release, insulin secretion, the immune system, and cell proliferation and/or differentiation.126

1.7.2. Receptors and binding

The 38-amino acid peptide PACAP was originally isolated from the brain of sheep, and identified due to its ability to stimulate cAMP production in vitro in rat pituitary cells.127 The following year, an analog of PACAP corresponding to the 27 N- terminal amino acids with an amidated C-terminus was reported.128 The researchers noted that this truncated form of PACAP displayed comparable activity to the longer version, and that it appeared closely related to VIP. The primary receptor for PACAP (27 or 38 residues) is the PAC1 receptor,129 although it also shows strong affinity and activity at the

VPAC1 and VPAC2 receptors.101 Interestingly, PACAP may bind to a different binding groove on the PAC1 receptor than the other secretin peptides do on their respective receptors.52 The ability of the peptide to adopt binding conformations for both the unique

PAC1 receptor groove, and the groove that is found on other secretin receptors, such as the VPAC1 receptor, may explain why PACAP is more promiscuous than VIP.

Shirakawa and coworkers proposed that the Gly4 residue in PACAP stabilizes a β-turn that allows activation of the PAC1 TM domain active site, while the more conformationally constrained Ala4 in VIP prevents a similar conformational adjustment.52

The importance of the first five N-terminal residues of PACAP for receptor selectivity and activation has been shown by multiple groups,130-132 although Robberecht and coworkers did note slight agonist activity of PACAP6-38 in their system. Robberecht also suggests that PACAP6-27 is an antagonist, although at a concentration about 40 times higher than PACAP6-38. Other truncations (e.g. PACAP2-38, PACAP3-38) did not display antagonism.132 Interestingly, the group also noted that, in a human neuroblastoma cell membrane model, the stimulation of adenylate cyclase production restarts again with more highly truncated analogs (PACAP10-27 through PACAP14-27, and PACAP10-38 through PACAP14-38), though no further work was performed on elucidating the reasons behind this observation.133 This may be due to the observation that PACAP10-27 is 6000 times less active than PACAP1-27.

1.7.3. PACAP in neuroprotection

PAC1 is found throughout the central and peripheral nervous systems, and performs similar immunomodulatory roles to VIP,88,109,134 including modulation of the activity of cytokine IL-6,112 and its upregulation following overexpression α-synuclein.117

In particular, the PAC1 is highly expressed in the pituitary gland in humans,135 and the substantia nigra (s.n.) in mice, the latter of which is of particular interest due to PD being characterized by the death of dopaminergic neurons in the s.n..136 PACAP27 has been shown to be effective in providing neuroprotection in general models,137 PD models,138,139 and against neurotoxic agents in neuronal differentiated PC12 cells.140

However, agonism of an as yet unidentified receptor by PACAP38 appears to contribute to migraines, while VIP does not affect migraines.141 Therefore, the development of a

PACAP antagonist may lead to a treatment for migraines.

1.7.4. Modifications to improve drug-like properties

In contrast with VIP, relatively little has been done to improve the DMPK of

PACAP.134 Like VIP, PACAP38 was allowed to associate with lipid-based micelles to determine if that could be used as a drug delivery system.125 Yamada and coworkers reported that substituting all of the lysine residues for asparagine, as well as methionine for leucine, yielded PACAP38 [R15,20,21, L17], an analog that was more stable than the native peptide while maintaining its efficacy.142 Finally, Dr. Bobbi Anglin reported the initial synthesis of glycosylated PACAP27, showing that the glycosylated peptide was able to penetrate the BBB, while the native PACAP27 was not observed in quantifiable amounts.39

1.8. Angiotensin-(1-7)

1.8.1. History and functions

Angiotensinogen is a naturally occurring peptide hormone principally produced in the liver. It is metabolized to 10-residue angiotensin I by the enzyme renin, which is secreted by the kidney, and further cleaved by angiotensin-converting enzyme (ACE) to the 8-residue peptide angiotensin II (Scheme 1-2), which primarily activates the

143 angiotensin subtype 1 (AT1) receptor. This is known as the renin-angiotensin system

(RAS). Over-activation of AT1 can lead to numerous deleterious effects, including hypertension, atherosclerosis, and cardio fibrosis, which are all risk factors for heart failure and stroke.15 Therefore, a popular method for reducing the likelihood of heart failure in high-risk individuals is the prescription of ACE inhibitors and AT1 blockers, in effect shutting down this pathway of action. Using AT1 blockers may also improve

54 stimulation of angiotensin subtype II (AT2) receptors as its ligand, angiotensin II, no longer participates in AT1 receptor binding.

Angiotensin-(1-7) (Ang-(1-7)) is a heptapeptide that contains the seven N- terminal residues of angiotensinogen (and angiotensin I and II), and was initially isolated from a canine brain.144 Soon after, it was shown to be as effective as angiotensin II at stimulating the release of vasopressin in vitro, despite lacking the hydrophobic phenylalanine residue in the 8-position previously thought to be essential for receptor activation.145 Its synthesis in vivo is primarily catalyzed by angiotensin-converting enzyme II (ACE2),146 though other enzymes have been demonstrated to convert angiotensin I to Ang-(1-7) through related intermediates. Like many peptides, Ang-(1-7) has a short in vivo half-life in humans of around 30 minutes, and 9 seconds in rats.147-149

To date, the only reported method of improving Ang-(1-7)’s DMPK profile is demonstrated in a treatment for pulmonary hypertension, through its incorporation into plant cells.150 Interestingly, inhibition of ACE has also been shown to increase the in vivo half-life of Ang-(1-7) by reducing that enzyme’s ability to convert it to angiotensin-(1-

5).148

Ang-(1-7) has been shown to have a range of physiological effects in the human body, including cardiovascular benefits such as vasodilation, antiarrhythmic, and antihypertensive actions, and peripheral benefits including as a diuretic in the kidney.147,151 However, some studies have shown that Ang-(1-7) can also have detrimental effects in the same locations, suggesting that its activity is dose-dependent.

Scheme 1-2. In vivo production of Ang-(1-7) and its interaction with the MAS1 receptor.

Ang-(1-7) binds to the Mas receptor, a GPCR found exclusively in vascular cells that regulate the effects of angiotensin II binding to the angiotensin receptor.15,152 It is found primarily in the brain, heart, testis, and kidney.147 Its primary regulatory function is moderation of blood pressure,153 though the Ang-(1-7)/Mas activation axis has been shown to act in an anti-inflammatory role in the mitigation of diverse inflammatory diseases,15 and has recently been reported to act as a peripheral antinociceptive when activated with Ang-(1-7).154 The metabolites of angiotensin play important roles in a variety of processes, but Ang-(1-7) shows modulatory effects that are detailed further in the following section.

1.8.2. Ang-(1-7) as a treatment for postoperative cognitive impairment and stroke

Postoperative cognitive impairment and stroke are common in patients recovering from heart failure (also called a myocardial infarction), especially after coronary artery bypass surgery.155-158 Ischemia, or lack of blood flow to the brain, leads to an increase in

ROS production and increased inflammation, and increases the likelihood of neuronal dysfunction and cognitive impairment.159

Ang-(1-7) has been shown to reduce the damage caused by strokes or in stroke models in both mice and rats.160-163 Initially, researchers were unable to attribute the benefits of Ang-(1-7) administration in stroke-model mice to either vasoconstriction or hypertension, the two primarily observed results of Ang-(1-7) introduction. However, further observations determined that much of the damage of strokes caused by intracerebral pro-inflammatory responses was reduced.162 Coupling this with the connection between the Mas receptor and inflammatory pain164 led to the hypothesis that

Ang-(1-7) plays a role in modulating immune response. Further experiments showed a

57 strong reduction in microglial activation around the brain parenchyma, which indicates a reduction of inflammation, after Ang-(1-7) administration.163 Interestingly, this beneficial therapeutic effect is in direct opposition to the effects that angiotensin II activation (or

15 overactivation) of AT1 show in the same model. Finally, Ang-(1-7) has been shown to play a role in modulating the activity of ROS in the brain, where the activation of the Mas receptor results in a decrease of ROS.143,165 See Figure 1-10 for a graphic representation of these effects.147

Figure 1-10: Protective effects of Ang-(1-7) on stroke.147

2. Development of novel sugar-amino acids

2.1. Design rationale

The incorporation of glycosides into biologically active peptides has led to a number of improvements of the DMPK properties of peptide drugs.30,64 Explanations of the mechanisms behind these improvements have centered on the modulation of amphipathicity that hydrophilic carbohydrates provide for relatively hydrophobic peptides (the so-called “biousian hypothesis”).34 This hypothesis was explicitly tested by affixing glycosides of varying size to a constant peptide scaffold (Figure 2-1).1,166

Increasing hydrophilicity

Figure 2-1: DAMGO analogs (red) with increasing size of hydrophilic moiety (Gly-ol, Ser(OH), Ser(OXyl), Ser(OGlc), Ser(OLac).1

In this case, the peptide was DAMGO, a potent and selective MOR agonist.

Amphipathicity was originally calculated by creating a surface area map of each molecule using Molecular Operating Environment™ (MOE) software (Chemical

Computing Group, Canada). The calculated Connolly surface area of the message segment, YaG(NMe)F, Alipid, was classified as lipophilic, while the Connolly surface area of the residue or carbohydrate affixed to the C-terminus, Awater, was assigned as

59 hydrophilic. This was then entered into an equation in order to estimate the amphipathicity,167 where:

퐴 (− 푤푎푡푒푟) 퐴푚푝ℎ𝑖푝푎푡ℎ𝑖푐𝑖푡푦 = 푒 퐴푡표푡푎푙

When these values were plotted against a measure of in vivo potency, in this case a mouse tail-flick test (Chart 2-1), a clear correlation between potency (both from i.v. and i.c.v. administration) and amphipathicity was observed with the glycopeptides.1

2.4

DAMGO

1.6

DAMGO-Ser(OLac)

1.2

i.v., µmol/kgi.v.,

50 0.8 DAMGO-Ser(OGlc) A

0.4 DAMGO-Ser(OXyl) DAMGO-Ser 0 0.65 0.7 0.75 0.8 0.85 0.9 0.95 Amphipathicity (Connolly-derived)

Chart 2-1: Plot of the efficacy of glycosylation in improving i.v. activity of the MOR agonist DAMGO, as measured in mouse tail-flick studies.1

The result was a U-shaped curve, in which lipophilic DAMGO analogs did not perform as well as those containing a carbohydrate and more hydrophilic character.

However, the most hydrophilic analog, that which included the disaccharide lactose, evidenced decreased activity when compared to analogs with smaller conjugated sugars.

These data closely matched estimated glucose units (g.u.) assignments of the relative

60 carbohydrate size conjugated to each peptide (data not shown). Based on these results, it was hypothesized that there might exist an ideal amphipathicity profile that lay between the amphipathicities of DAMGO-Ser(OH) and DAMGO-Ser(OXyl).

Two novel, modified furanoid carbohydrates were proposed to exist within that desired amphipathicity profile based on their calculated surface areas. Each is based on the scaffold of an existing carbohydrate, but with the addition of amine and carboxylic acid functionalities; these belong to a family of compounds known as sugar-amino acids

(SAAs).168-171 This methodology, as opposed to the traditional O-linked glycopeptides from the Polt group, was chosen in order to both minimize the additional hydrophilicity attributed to serine and to ease the SAAs’ incorporation into traditional fluorenylethyloxycarbonyl (Fmoc) solid-phase peptide synthesis (SPPS) conditions

(Figure 2-2).

Exact comparisons of the amphipathicity of the proposed DAMGO-SAA derivatives to literature values were unavailable, but a suitable alternative was developed

(for a discussion of the modified approach, see Appendix A. Calculation of the Connolly surface area of glycopeptides). However, the predicted amphipathicity values for the novel DAMGO-SAA compounds (e-w = 0.730-0.777) are very close to those of the

DAMGO-S(OGlc) (e-w = 0.742), and further away from the space between DAMGO-

Ser(OH) (e-w = 0.890) and DAMGO-Ser(OXyl) (e-w = 0.765) that we were attempting to probe.

Figure 2-2: Proposed SAAs (1, 2) and DAMGO-SAA derivatives (3-5) using Fmoc-based SPPS synthesis conditions. Glycopeptide 3′ was not observed.

During these calculations it was also noted that the calculated amphipathicities of the two SAAs were almost identical, despite 2 having one fewer methylene and one more alcohol than 1. This serves to illustrate that using surface areas to calculate amphipathicity is a crude estimate, and does not differentiate between areas that are more or less hydrophilic. Other methods, such as the use of g.u.s, can be experimentally

62 determined or can be estimated. Based on estimates from the Polt lab, looking at the total number of free hydroxyls and the presence of an amide in the SAAs, we were able to assign the carbohydrate end of 3 as 1.5 g.u., and 5 as 1.75 g.u. These values are either equivalent to or between those of of Ser(OH) (1.25 g.u.) and Ser(OXyl) (1.75 g.u.), and therefore seemed reasonable.

In addition to amphipathicity, we hypothesized that the precise identity of the carbohydrate moiety was unimportant for BBB penetration, e.g. that BBB penetration was determined by biophysical interactions with membrane targets, rather than the product of interaction with a membrane transporter system. To date, all biologically active glycopeptides that had been reported in literature had been modified using naturally occurring carbohydrate glycosides. If amphipathicity was the important feature, one would not expect different results if the installed carbohydrate was non-natural.

Conveniently, this, along with the observation that phosphate derivatives of enkephalins have potent antinociceptive effects following i.v. administration,166 would also suggest the hypothesis that transcytosis of glycopeptides occurs via a receptor-mediated pathway is either incomplete or inaccurate, as no receptor should exist for non-natural carbohydrates.

2.2. Synthesis of novel SAAs

In order to synthesize the desired SAAs 1 and 2, we wanted to start with readily available and inexpensive starting materials, and to perform a series of transformations that were both high-yielding and used inexpensive and non-toxic reagents. Molecules similar to each of these had been previously synthesized by other groups,172-175 but the schemes either used toxic reagents or were incompatible with Fmoc-based SPPS

63 conditions. The two Fmoc-protected furanoid SAAs were derived from ribose and xylose, and were synthesized from their respective sugars.

2.2.1. 1-(Methylcarboxyl)-2,3-O-isopropylidene-5-amino-α-D-ribofuranose

To synthesize novel carbohydrate analog 1, commercially available D-ribose (6) was treated with 2,2-dimethoxypropane to give the kinetically favored acetonide- protected sugar 2,3-O-(isopropylidene)-beta-D-ribose (7).176,177 This was then allowed to react with benzyl (triphenylphosphoranylidene)acetate to selectively provide the corresponding benzyl ester at the 1 position in place of the 2⁰ alcohol, while concurrently providing the desired stereochemistry at the 1 position (8).178-180 Utilization of the benzyl ester in place of the methyl ester found in literature served two purposes: it provided a group whose removal is orthogonal to the eventual Fmoc group at the N-terminus; and it affixed a UV-active moiety that allowed for facile automated column chromatography when necessary in the subsequent steps. Alcohol 8 was then treated with tosyl chloride to yield the tosylate (9), which then underwent substitution after addition of sodium azide to provide the corresponding azide (10).181,182 After concurrent reduction of the azide and

181,182 deprotection of the benzyl group from the carboxylic acid via H2 and Pd/C (11), the resulting amine was protected with fluorenylmethyloxycarbonyl succinimide (Fmoc-

OSu) to afford the desired SAA 1. From isopropylidene-ribose, the total yield to 1 was

16% over six steps.

Scheme 2-1: Synthetic scheme for the formation of SAA 1.

2.2.2. 1-Carboxyl-2,3-O-isopropylidene-6-amino-α-D-sorbofuranose

To synthesize carbohydrate 2, commercially available (-)-2,3:4,6-Di-O- isopropylidene-2-keto-L-gulonic acid (12) was allowed to react with benzyl chloride to afford benzyl ester 13. Selective hydrolysis of the 4,6-isopropylidene was performed with

183,184 1:1 AcOH:H2O over 2 h at 50 ⁰C (14). The primary alcohol was then selectively tosylated (15) and substituted to the corresponding azide (16).173,185 Reduction of the azide and deprotection of the carboxylic acid using a Pd/C catalyst in H2 atmosphere gave the SAA 17. The amine was then protected with Fmoc-OSu (2). The overall synthesis was performed with 48% yield over six steps.

Scheme 2-2: Synthetic scheme for the formation of SAA 2. 2.3. Synthesis of DAMGO-SAA glycopeptides

2.3.1. Solid-phase peptide synthesis of glycopeptides

The SAAs 1 and 2 were loaded onto Rink amide-MBHA resin, and the DAMGO amino acids (Tyr-ala-Gly-(NMe)Phe) were coupled without complication as previously described.1 Resin cleavage and sidechain deprotection of the resin-bound, protected analog of 3 proceeded as expected (Figure 2-3). However, after a typical two hour cleavage of DAMGO-SAA 2 in ~ 90% trifluoroacetic acid (TFA), we observed roughly a

3:1 ratio of the less to more polar SAA products via HPLC (Figure 2-4). MS analysis suggested that we had two glycopeptides, 4 and 5. We were excited by these results, as they provided an unexpected, additional DAMGO analog to test. In addition, this would provide a much more lipophilic SAA moiety, which could be used to further extricate the effects of amphipathicity on glycopeptide activity on receptors in the brain and, by extension, on BBB penetration.

A B

Figure 2-3: HPLC trace of crude (A) and pure (B) DAMGO-SAA analog 3. A

B C

Figure 2-4: HPLC traces of crude mixture of DAMGO SAA analogs 4 and 5 (A), and purified 5 (B) and 4 (C).

2.3.1.1. Removal of the isopropylidene moiety in SAA 4.

The partial cleavage of the 1,2-isopropylidene from glycopeptide 4 can be attributed to the increased steric interactions around that group, as compared to the complete cleavage of the 2,3-isopropylidene in glycopeptide 3. We hypothesized that ketal protonation was the rate-limiting step due to sterics, and that a strong, basic

67 nucleophile would provide the desired increase in the rate of deprotection when compared with a strong acid.

Figure 2-5: Conversion of 4 to 5 using (A) 4:1 H2NNH2•H2O:MeOH, (B) 95% AcOH/5% H2O, 9h, (C) 50% AcOH/50% H2O, 16h. Note that (B) uses a different HPLC method, resulting in a shifted retention time.

To isolated samples of 4 were added a) 4:1 H2NNH2•H2O:MeOH, b) 95% acetic acid (AcOH)/5% H2O, c) 50% AcOH/50% H2O, and d) 50% AcOH/50% H2NNH2•H2O

(Figure 2-5). After allowing the glycopeptides to react overnight, it was observed that the acidic conditions (c) afforded the best deprotection, while use of hydrazine led to

68 glycopeptide degradation (unisolated byproducts). The mixed acid-base conditions (d) led to salt formation, and neither peptide could be isolated. Interestingly, although standard

SPPS cleavage conditions (~ 90% TFA) only led to 50% isopropylidene cleavage over two hours, tuning the acid to AcOH was more successful.

2.4. Biological testing of DAMGO-SAA glycopeptides

In order to fully compare the newly synthesized DAMGO glycopeptides 3-5 to the previously reported analogs utilizing naturally occurring carbohydrates, we analyzed their receptor binding to all three opioid receptors, and tested their in vivo efficacy using a murine antinociception assay in collaboration with other research labs.

2.4.1. Binding assays to opioid receptors

Binding studies were performed by radioligand displacement at the University of

Rochester by the lab of Dr. Jean M. Bidlack as previously described.1,166 Briefly, a radiolabeled ligand and the DAMGO analogs at varying concentrations were incubated with Chinese hamster ovary (CHO) cells expressing either human µ-, δ-, or κ-opioid receptors. The displaced radioligands were [3H]DAMGO at the MOR, [3H]Naltrindole at the DOR, and [3H]U69,593 at the KOR. Binding affinity was calculated based on the remaining quantity of radioligand after the cells were washed and using their previously established Kd values.

The observed binding affinities for the DAMGO-SAA glycopeptides, along with that of unmodified DAMGO, can be found in Table 2-1.

[3H]DAMGO [3H]U69,593 [3H]Naltrindole (Mu Opioid (Kappa Opioid (Delta Opioid Receptor) Receptor) Receptor) % Max.

Peptide Ki SEM n Ki SE n Inhib. at SEM n (nM) (nM) M 10 mM 3 0.85 0.021 3 660 33 3 52 0.49 3 4 1.3 0.17 3 380 19 3 66 1.3 3 5 0.88 0.083 3 500 48 3 64 1.8 3 DAMGO 0.56 0.006 270 9.3 100 Table 2-1: Binding affinities of the DAMGO-SAA glycopeptides at the MOR, DOR, and KOR.

As expected with the DAMGO-based peptides, all three DAMGO analogs are potent MOR agonists, and are much less potent at the DOR and KOR. Based on the similarity of binding between the two free glycopeptides, 3 and 5, it appears that the difference in the size of the carbohydrate does not significantly affect the two-step binding mechanism. Encouragingly, these two also have lower experimental Ki values than DAMGO-Ser(OXyl) and DAMGO-Ser(OGlc) (Ki = 1.30 nM for each), but they are higher than those for DAMGO-Ser(OH) (Ki = 0.68 nM) and DAMGO-Ser(OLac) (Ki =

0.55 nM).1 While it is unwise to draw too many conclusions from these few data points, it is encouraging that the DAMGO-SAAs do not appear to negatively affect receptor binding.

In contrast with the two free DAMGO-SAAs is glycopeptide 4, which has Ki =

1.3, about 50% greater than either of the free glycopeptides. This can be explained using the amphipathicity hypothesis; by masking the 1,2-dihydroxyl with an isopropylidene the lipophilicity of the glycopeptide was greatly increased (from 0.74 to 0.78 using surface

70 areas, but from 1.75 to 1.25 g.u.). This in turn decreased the ability of the glycopeptide to engage in a hopping mechanism (discussed in greater detail in 1.5.2. Glycosylation) that may allow for increased opportunity to encounter the appropriate receptor. Although not included in the previous hypothesis, I propose that the steric bulk of the isopropylidene causes the carbohydrate to be rotationally hindered, which affects how the glycopeptide interacts with the binding pocket of the MOR.

2.4.2. Antinociceptive effects using murine tail-flick test

Once it was confirmed that the DAMGO-SAA glycopeptides were highly potent

MOR agonists, we sent the samples to Dr. Edward Bilsky and coworkers at the

University of New England to perform experiments testing their antinociceptive activity in vivo using mice as previously described.1,166 Briefly, the tails of untreated mice were placed in a 55 ⁰C water bath, and the latencies of the removal of their tails from the water was measured and recorded as a baseline. Each mouse was then given a test compound, and antinociception was measured by again using the hot water bath. Mice which did not remove their tails within 10 seconds were considered to have full antinociception (100%).

Antinociception was calculated using the following equation:

(푡푒푠푡 푙푎푡푒푛푐푦 − 푐표푛푡푟표푙 푙푎푡푒푛푐푦) % 푎푛푡𝑖푛표푐𝑖푐푒푝푡𝑖표푛 = 100 푥 (10 − 푐표푛푡푟표푙 푙푎푡푒푛푐푦)

The results of that experiment can be found in Table 2-2.

Activity

Peptide i.v. A50 (µmol/kg) n i.c.v. A50 (pmol/mouse) n

3 1.76 (1.42-2.17) 4 31 (16.0-59.0) 5 4 10.68 (8.08-14.12) 4 82 (58.5-114.9) 4 5 4.86 (3.03-7.81) 4 36.9 (25.0-56.0) 4

Table 2-2: I.v. and i.c.v. antinociceptive A50 values in mice for the DAMGO-SAA glycopeptides, acquired via tail-flick test.

All three glycopeptides show excellent activity in mice, both when administered directly into the NVU (i.c.v.), and when administered peripherally (i.v.). As predicted by amphipathicity, glycopeptide 4 was less active via both administration methods than the more hydrophilic glycopeptides 3 and 5. This is due in part to the reasons presented in the previous section; namely, increased lipophilicity and steric bulk of the carbohydrate.

These experiments were unable to conclusively inform the hypothesis regarding carbohydrates increasing membrane transport, including BBB penetration, as the decrease in activity could be due to that, lipophilicity, or sterics interrupting receptor interactions.

2.4. Discussion

Comparing the activity data of 3-5 with that of the DAMGO analogs previously reported1 does not help to clarify the amphipathicity hypothesis, as the activity of the other analogs is, on average, an order of magnitude less than our observed activities by concentration. It is important to note, however, that we cannot easily compare the two data sets for two reasons.

First, the two series of experiments were conducted years apart, with different researchers. As research has shown, mice react differently in antinociception studies depending on the sex of the researcher performing the experiment,186 as well as based on a wide range of other environmental factors.187 Because of this, direct comparisons of data sets collected at different times cannot be reliably compared on their own. One method of allowing a comparison is through the introduction of a standard (e.g. morphine) within every set of data collected. As this shortcoming was not immediately identified, no standard was used, and therefore each glycopeptide can only reliably be compared to other compounds tested within the same series of experiments.

Despite this limitation, several conclusions can be drawn from the research described above. First, addition of a non-naturally occurring carbohydrate to a neurologically active peptide provides similar benefits to those of natural sugars, namely improved BBB penetration. While improved resistance to enzymatic degradation was not tested on this series of peptides, it is reasonable to expect that SAAs will also provide this advantage. Second, SAAs can be installed in a way that does not disrupt ligand-receptor interactions if installed away from the “message” portion of the peptide. Third, it appears that the hydroxyl groups must be unprotected to afford the greatest benefits. Whether this is due to the increased steric bulk of the isopropylidene-protected glycopeptide (4), a product of the latter’s potential to inhibit rotation, or its decreased hydrophilicity and associated decrease in ability to perform the hypothesized membrane hopping mechanism is not apparent based on the results of the current research.

2.5. Materials and methods

2.5.1. Sugar-amino acid synthesis

Synthesis of Benzyl 3,6-anhydro-2-deoxy-4,5-O-isopropylidene-D-allo-heptulonate (8)

7 was prepared as described previously.176,177 35.97 g (189 mmol) was dissolved in 300 mL acetonitrile while stirring, then benzyl (triphenylphosphoranylidene)acetate (88.35 g,

215 mmol) was added. The reaction was heated under reflux for 2.5 hours and followed by TLC. The reaction mixture was concentrated in vacuo, then dissolved in 300 mL of

MeOH. The pH was checked, and if necessary NaOMe was added to raise the pH to >8.

The mixture was stirred for 10 minutes, then neutralized with AcOH to pH 7. The solution was concentrated in vacuo, and the resulting syrup (≈120 g) was dissolved in

360 mL CH2Cl2. The solution was divided into thirds, and each fraction was purified using an automated Biotage column (65i, hexanes:EtOAc 8:2 to 3:7 v:v, 4000 mL total, no prefraction, flow rate 47 mL/min, 18 mL fractions, threshold 0.01, product appears in fractions 24-60). Fractions containing pure product were combined and evaporated in vacuo, giving 34.2 g (106 mmol, 56% yield) white crystals. TLC (Hex:EtOAc, 1:1, Rf =

0.47) 1H NMR (400 MHz, Chloroform-d) δ 7.41 – 7.29 (m, 5H), 5.15 (d, J = 3.23 Hz,

2H), 4.71 (dd, J = 3.90, 6.65 Hz, 1H), 4.54 (dd, J = 4.79, 6.73 Hz, 1H), 4.29 (dt, J = 4.93,

6.63 Hz, 1H), 4.07 (dt, J = 3.03, 3.65 Hz, 1H), 3.78 (ddd, J = 2.95, 4.20, 12.07 Hz, 1H),

3.64 (ddd, J = 3.72, 8.14, 12.01 Hz, 1H), 2.79 (dd, J = 4.99, 15.79 Hz, 1H), 2.68 (dd, J =

6.65, 15.81 Hz, 1H), 2.57 (dd, J = 4.43, 8.21 Hz, 1H), 1.54 (s, 3H), 1.34 (s, 3H).

Synthesis of tosylate (9)

8 (2.1 g, 6.5 mmol) was dissolved in 6.5 mL of pyridine and the mixture was cooled to 0

⁰C. Tosyl chloride (1.43 g, 7.5 mmol) was added, and the mixture was stirred. After 3 hours, the reaction mixture was allowed to warm to room temperature, where it was stirred for additional 16 hours. The reaction was monitored by TLC. 0.3 mL of H2O and

0.3 mL of MeOH were added, and the reaction mixture was stirred for 2 hours. The solution was then diluted with 15 mL toluene and evaporated in vacuo to half the original volume. The wash and evaporation with toluene was repeated three times. The resulting residue was dissolved in 30 mL EtOAc, then washed with H2O (x3), 0.1 M CuSO4 (x3), saturated NaHCO3 (x2), and H2O (x1). The organic phase was dried, and evaporated in vacuo to give 3.1 g (6.5 mmol) syrup. Syrup was purified using an automated Biotage column (4x4 cm column, hexanes:EtOAc 100:0 to 40:60 gradient, 1000 mL, no prefraction, flow rate 21 mL/min, fraction size 6 mL, threshold 0.01, pure product appears in fractions 16-37). Fractions containing pure product were combined and evaporated in vacuo, giving 2.4 g (5.0 mmol, 77% yield). TLC (Hex:EtOAc , 6:4, Rf =

0.5) 1H NMR (400 MHz, Chloroform-d) δ 7.83 – 7.74 (d, J = 8.04 Hz, 2H), 7.42 – 7.28

(m, 7H), 5.15 (s, 2H), 4.58 (ddd, J = 1.71, 2.53, 6.69 Hz, 1H), 4.50 (dd, J = 4.32, 6.64

Hz, 1H), 4.35 – 4.25 (m, 1H), 4.17 – 4.04 (m, 3H), 2.63 (qd, J = 15.42, 5.22 Hz, 2H),

2.43 (s, 3H), 1.51 (s, 3H), 1.31 (s, 3H).

Synthesis of azide (10)

9 (2.4 g, 5.0 mmol) and sodium azide (1.63 g, 25 mmol) were combined in a flask, and 5 mL of DMF was added. The suspension was stirred in an oil bath at 70 ⁰C for 2.5 hours, at which point a fine, white solid was present in the reaction mixture. After cooling the mixture to room temperature, 30 mL H2O was added, then the solution was extracted with EtOAc (4x 10 mL). The combined organic layers were washed with H2O (2x), saturated NaHCO3 (x1), H2O (x1), and brine (x1), dried, and concentrated in vacuo to 1.4

1 g (4.0 mmol, 80% yield) to give 10 as an oil. TLC (Hex:EtOAc, 9:1, Rf = 0.5) H NMR

(400 MHz, Chloroform-d) δ 7.49 – 7.29 (m, 5H), 5.17 (s, 2H), 4.65 – 4.53 (m, 2H), 3.52

(dd, J = 3.88, 13.06 Hz, 1H), 3.32 (dd, J = 4.84, 13.06 Hz, 1H), 2.98 – 2.94 (m, 1H), 2.85

– 2.66 (m, 2H), 1.55 (s, 3H), 1.35 (s, 3H).

Synthesis of SAA (11)

10 (1.45 g, 4.2 mmol) crude was dissolved in 50 mL MeOH and 0.5 mL H2O. 250 µL of

AcOH was then added, followed by 100 mg 10% Pd/C. The resulting suspension was stirred under H2 atmosphere for 2 hours. TLC showed reaction was incomplete, so an additional 75 mg 10% Pd/C was added, and stirring was resumed for 1 hours. The mixture was filtered through diatomaceous earth, and the solvent was evaporated in vacuo to provide 930 mg (4.0 mmol, 95% yield) crude, solid, product 11. TLC

(Hex:EtOAc, 7:3) 1H NMR (400 MHz, Deuterium Oxide) δ 4.69 – 4.65 (m, 2H), 4.38 –

4.28 (m, 1H), 4.15 (m, 1H), 3.29 (dd, J = 3.55, 13.44 Hz, 1H), 3.05 (dd, J = 9.93, 13.45

Hz, 1H), 2.60 – 2.52 (m, 1H), 2.44 – 2.35 (m, 1H), 1.50 (s, 3H), 1.31 (s, 3H).

Synthesis of Fmoc-SAA (1)

To a stirred solution of 11 (19.0 g, 82.1 mmol) and NaHCO3 (15.2 g, 180 mmol) in 155 mL H2O and 155 mL acetone was added 27.8 g (82.4 mmol) Fmoc-O-Su. After stirring overnight the pH of the reaction was 8.5. Acetone was evaporated in vacuo to a final volume of 150 mL, and an additional 150 mL H2O was added to the reaction. The white solid was filtered off, and the aqueous phase was extracted with Et2O (x3). The pH of the aqueous phase was then set to 3 using 3 M HCl. The resulting white slurry was extracted with EtOAc, then the organic phase was washed with 0.1 M HCl, dried, and evaporated in vacuo to give 35 g of crude product. The solid was dissolved in 50 mL EtOAc, then 50 mL cyclohexane and a seed crystal was added slowly. Crystal formation began

77 immediately, and was allowed to progress overnight. The resulting crystals were filtered off to give 13.5 g pure crystals, while the mother liquor was evaporated in vacuo and the crystallization process was repeated. Final mass pure 1 was 18.6 g (41.0 mmol, 50% yield). 1H NMR (400 MHz, Chloroform-d) δ 7.78 (d, J = 7.63 Hz, 2H), 7.62 (d, J = 7.55

Hz, 2H), 7.41 (tdd, J = 0.63, 1.17, 7.51 Hz, 2H), 7.33 (td, J = 1.19, 7.43 Hz, 2H), 4.49

(m, 1H), 4.44 (m, 2H), 4.24 (m, 1H), 3.49 – 3.43 (m, 2H), 2.77 (dd, J = 4.71, 16.10 Hz,

1H), 2.62 (dd, J = 6.78, 16.16 Hz, 1H), 1.54 (s, 3H), 1.34 (s, 3H).

Synthesis of benzyl ester (13)

12 (2.74 g, 10 mmol) was dissolved in 10 mL CH2Cl2. While stirring, 5.2 mL (3.87 g, 30 mmol) diisopropylethyl amine was added, and the solution was cooled to 0 °C. 1.8 mL

(2.56 g, 15 mmol) benzyl bromide was added dropwise over 5 minutes with stirring, and the reaction was followed by TLC. After 16 hours, 0.5 mL MeOH was added, and the solution was stirred for an additional 3 hours. The solution was evaporated in vacuo, and the resulting residue was dissolved in 50 mL EtOAc, then washed with saturated

NaHCO3 (x3) and H2O (x3), then dried and evaporated in vacuo. Typically this crude residue was used without purification in the synthesis of 14. Purification was performed using an automated column, first eluting with CH2Cl2, then CH2Cl2:EtOAc (9:1) once any initial impurities had been washed off the column. Fractions containing the desired compound were combined and evaporated in vacuo to give 2.6 g (7.1 mmol, 71%) 13. 1H

NMR (400 MHz, Chloroform-d) δ 7.45 – 7.22 (m, 5H), 5.29 (s, 2H), 4.84 (s, 1H), 4.29

(d, J = 2.50 Hz, 1H), 4.16 (q, J = 2.22 Hz, 1H), 4.08 (dd, J = 1.97, 5.04 Hz, 2H), 1.51 (s,

3H), 1.40 (s, 3H), 1.37 (s, 3H), 1.25 (s, 3H).

Synthesis of monoisopropylidene compound (14)

Crude 13 (0.5 g, 1.4 mmol) was stirred in 4 mL 1:1 AcOH:H2O at 50 °C, and the reaction was followed by TLC. After 2-3 hours, the solution was diluted with 10 mL toluene, then concentrated in vacuo to a total volume of 5 mL under reduced pressure at 35 °C. The toluene addition and evaporation was repeated an additional four times, until only

1 colorless syrup remained (0.44 g, 1.4 mmol, 100%). TLC (CH2Cl2:EtOAc, 9:1) H NMR

(400 MHz, Chloroform-d) δ 7.39 – 7.31 (m, 5H), 5.27 (s, 2H), 4.72 (s, 1H), 4.44 – 4.33

(m, 2H), 4.13 (dd, J = 3.59, 12.67 Hz, 1H), 3.99 (dd, J = 2.73, 12.70 Hz, 1H), 1.53 (s,

3H), 1.39 (s, 3H).

Synthesis of tosylate (15)

14 (324 mg, 1 mmol) was stirred into 3 mL pyridine at 0 °C, followed by tosyl chloride

(219 mg, 1.15 mmol). After 3 hours the reaction was allowed to warm to room temperature, and stirred for an additional 16 hours. 200 µL MeOH was added, and the reaction was stirred for an additional 3 hours. The solution was then diluted with 5 mL

79 toluene and evaporated to half volume. The dilution and evaporation with toluene was repeated twice more, until only a residue remained. The residue was dissolved in 50 mL

EtOAc, then washed with H2O (x1), saturated NaHCO3 (x1), and H2O (x1). The organic phase was dried and evaporated in vacuo to afford 460 mg crude white crystals. The crude material was then recrystallized from EtOAc/hexanes, giving 418 mg (0.87 mmol,

1 87% yield) white crystals. TLC (CH2Cl2:EtOAc, 9:1, Rf = 0.7) H NMR (400 MHz,

Chloroform-d) δ 7.86 – 7.79 (m, 2H), 7.43 – 7.33 (m, 7H), 5.30 (d, J = 12.21 Hz, 1H),

5.25 (d, J = 12.21 Hz, 1H), 4.68 (s, 1H), 4.50 (td, J = 2.66, 6.15 Hz, 1H), 4.39 (dd, J =

6.31, 10.62 Hz, 1H), 4.28 (ddd, J = 0.75, 2.63, 9.17 Hz, 1H), 4.21 (dd, J = 6.00, 10.60

Hz, 1H), 2.79 (d, J = 9.17 Hz, 1H), 2.47 (s, 3H), 1.53 (d, J = 0.70 Hz, 3H), 1.35 (d, J =

0.72 Hz, 3H).

Synthesis of azide (16)

15 (4.8 g, 10 mmol) and NaN3 (3.25 g, 50 mmol) were combined in a flask, then 10 mL of DMF was added. The suspension was stirred in an oil bath at 70 ⁰C. After 48 hours, a fine, white solid was present in the reaction mixture. The reaction was checked by TLC, then cooled to room temperature. The contents of the reaction flask were added to 60 mL

H2O, then extracted with EtOAc (4x 20 mL). The combined organic layers were washed with H2O (2x), saturated NaHCO3 (1x), and H2O (1x), then dried and concentrated in vacuo to an oil under reduced pressure. The crude oil was purified on a Biotage automated column (Hex:EtOAc, 9:1, Rf = 0.2) to give 2.8 g (8.0 mmol, 80%) pure

1 product. TLC (Hex:EtOAc, 9:1, Rf = 0.2) H NMR (400 MHz, Chloroform-d) δ 7.42 –

7.33 (m, 5H), 5.34 (d, J = 12.20 Hz, 1H), 5.29 (d, J = 12.26 Hz, 1H), 4.70 (s, 1H), 4.44

(td, J = 2.84, 6.18 Hz, 1H), 4.25 (dd, J = 2.75, 9.15 Hz, 1H), 3.69 – 3.55 (m, 2H) [dd + dd], 2.93 (d, J = 9.37 Hz, 1H), 1.57 (d, J = 0.52 Hz, 3H), 1.38 (d, J = 0.64 Hz, 3H).

Reduction to SAA (17)

Crude product 16 was used without further purification. 16 (30 g, 86 mmol) was dissolved in 750 mL MeOH, 18 mL H2O, and 6 mL AcOH. Into this solution 1000 mg

10% Pd/C was added and stirred under H2 atmosphere. After 1.5 hours, additional 500 mg 10% Pd/C was added, and the reaction was left overnight (small scale [≤1 g] completes in 1.5 hours total). The mixture was filtered through diatomaceous earth, evaporated in vacuo, then co-evaporated with acetonitrile. The residue was dissolved in

50 mL H2O and 50 mL acetonitrile, then lyophilized to provide 20.1 g (86 mmol, 99.5%

1 yield) solid, white product. TLC (Hex:EtOAc, 1:1, Rf = 0.0 [baseline]). H NMR (400

MHz, Deuterium Oxide) δ 4.68 (s, 1H), 4.55 – 4.51 (m, 1H), 4.31 (dd, J = 0.74, 3.18 Hz,

1H), 3.33 (d, J = 4.79 Hz, 2H), 1.46 (d, J = 0.74 Hz, 3H), 1.36 (d, J = 0.76 Hz, 3H).

Synthesis of Fmoc-SAA (2)

To a stirred solution of 17 (800 mg, 3.43 mmol) and NaHCO3 (634 mg, 7.5 mmol) in 10 mL H2O and 10 mL acetone was added 1.16 g (3.43 mmol) Fmoc-O-Su. After stirring overnight the pH of the reaction was 8.5. Acetone was evaporated under reduced pressure to a final volume of 150 mL, and an additional 20 mL H2O was added to the reaction.

The white solid was filtered off, and the aqueous phase was extracted with Et2O (x3). The pH of the aqueous phase was then set to 3 using 3 M HCl. The resulting white slurry was extracted with EtOAc, then the organic phase was washed with 0.1 M HCl, dried, and evaporated in vacuo to give 1.52 g (3.34 mmol, 97% yield). 1H NMR (400 MHz,

Chloroform-d) δ 7.79 (dd, J = 0.82, 7.62 Hz, 2H), 7.62 – 7.56 (m, 2H), 7.43 (t, J = 7.48

Hz, 2H), 7.34 (t, J = 7.44 Hz, 2H), 5.24 (t, J = 6.77 Hz, 1H), 4.75 (s, 1H), 4.58 – 4.46 (m,

2H), 4.30 – 4.24 (m, 1H), 4.21 (t, J = 6.40 Hz, 1H), 4.14 (s, 1H), 3.64 (m, 1H), 3.39 (dt, J

= 5.60, 14.38 Hz, 1H), 1.57 (s, 3H), 1.47 (s, 3H).

2.5.2. Peptide synthesis

Amino acids, coupling reagents, and Rink amide resin were purchased from

Advanced ChemTech (Louisville, KY, USA), and used without further purification. The fluorenylmethoxycarbonyl (Fmoc) pseudo-glycosides (AIXA and AIRA) were synthesized as described above, and a general resin loading procedure follows. The SAA was loaded onto Fmoc-Rink resin at 0.83 mmol/g resin loading in a 25 mL fritted syringe. Initially, the resin was swelled using DMF (~5 mL solvent per gram resin), agitating at RT for two minutes (x2). A solution of 2% DBU and 3% piperidine in DMF

(v:v) was introduced and agitated for 5 minutes, refreshed, and agitated an additional 10 minutes. The resin was then washed with DMF (x5) and NMP (x1). In a separate vial,

SAA 1 (0.12 mmol, 1.2 eq in relation to desired loading) was dissolved in 5 mL NMP,

82 and HOBt•H2O (0.13 mmol, 1.3 eq) was added and allowed to mix for 5 minutes.

Condensing agent DIC (0.26 mmol, 2.6 eq) was then added, and mixed for 5 minutes.

This solution was added to the resin and agitated for 10 minutes. Next, the syringe was placed in a microwave (Emerson 900W Microwave – MW9338SB) set to power level 1 and irradiated for 10 minutes, stopping to shake the syringe every 90 seconds. The syringe was then agitated at RT for an additional 30 minutes. The resin was washed with

NMP (x1), DMF (x5), and CH2Cl2 (x5), and dried in vacuo overnight.

Glycopeptides were synthesized manually using standard Fmoc chemistry on Rink amide resin using previously reported conditions.1 The side-chain protecting group for

Tyr was chosen to be orthogonal to coupling and Fmoc deprotection conditions, and acid- labile for concurrent deprotection at the end of the synthesis. The side-chain protected amino acid used in these syntheses was Fmoc-Tyr(But)-OH. Amide couplings were performed using DIC/HOBt in individual synthesis vessels using microwave assistance as described above. The coupling was monitored using the Kaiser ninhydrin test. Fmoc removal was afforded using 2% DBU/3% piperidine in DMF. After complete glycopeptide synthesis, and subsequent Fmoc removal, the glycopeptide was cleaved from the resin using F3CCOOH:Et3SiH:H2O:CH2Cl2:Ph-OCH3 (v:v:v:v:v,

9:0.3:0.2:1:0.05), agitating at RT for 2 hours, which concurrently removed the Boc protecting group from Tyr. The resulting solution was expelled into a 15 mL centrifuge tube, evaporated under argon, precipitated in ice-cold Et2O, decanted, and rewashed with

Et2O, then dissolved in H2O and lyophilized to afford fluffy, white solid.

Purification of the crude glycopeptides was accomplished by Reversed Phase HPLC

(RP-HPLC) with a preparative RP (C-18) Phenomenex (250 X 22 mm) column using a

CH3CN-H2O gradient solvent system containing 0.1% F3CCOOH. Homogeneity of the purified glycopeptides was confirmed by analytical RP-HPLC and high resolution mass spectrometry.

2.5.3. Radioligand-binding studies

Binding studies were performed in the laboratory of Dr. Jean M. Bidlack at the

University of Rochester by Brian I. Knapp. Chinese hamster ovary (CHO) cell membranes that stably expressed the human µ-, δ-, or κ-opioid receptors (MOR, DOR, and KOR) were used to determine binding as previously described.1,166 Cells membranes were incubated with the radiolabeled ligands in a final volume of 1 mL of 50 mM Tris-

HCl, pH 7.5 at 25 °C; incubation times of 60 minutes were used for the MOR-selective peptide [3H]DAMGO and KOR-selective ligand [3H]U69,593, while a 3 hour incubation was used for the DOR-selective antagonist [3H]naltrindole, at concentrations of 0.25, 0.2, and 1 nM respectively. concentrations of glycopeptide and the appropriate radioligand

(Table 2-1). Non-specific binding was determined by adding 10 µM naloxone to cells for

MOR and KOR binding, and 100 mM for DOR binding. Binding was terminated by filtering the samples through Schleicher & Scheull No. 32 glass fiber filters using a

Brandel 48-well cell harvester. The filters were washed three times with 3 mL of cold 50 mM Tris-HCl, pH 7.5, and then counted in 2 mL of ScintiSafe 30% scintillation fluid

(Fisher Scientific, Fair Lawn, NJ, USA). For the 3H]U69,593 binding, filters were soaked in 0.1% polyethylenimine for a minimum of 30 minutes before use. IC50 values were calculated by a least squares fit to a logarithm-probit analysis. Ki values of unlabeled compounds were calculated using the equation Ki = (IC50)/1 + S, where S =

(concentration of radioligand) (Kd of radioligand).

2.5.4. Animal and antinociceptive testing

All animal studies were performed in the laboratory of Dr. Edward J. Bilsky at the

University of New England by Lindsay St. Louis as described previously.1,166 In vivo studies used adult male ICR mice (25-30 g; Harlan Industries, Cleveland, OH, USA) maintained on 12 hour light/dark cycles (lights on at 07:00 hours) in a temperature- and humidity-controlled animal colony. Testing was carried out between 10:00 and 15:00 hours. Studies were carried out as recommended by the Guide for the Care and Use of

Laboratory Animals as adopted and promulgated by the National Institutes of Health.

Intracerebroventricular (i.c.v.) injections were carried by lightly anesthetized mice with ether, and making a small incision in the scalp. Injections were performed using a 10

µL Hamilton microsyringe (Hamilton Company, Reno, NV, USA) at a point 2 mm caudal and 2 mm lateral from bregma. Compounds were injected at a depth of 3 mm in a volume of 5 µL. Intravenous (i.v.) injections were performed by restraining the mouse in a Plexiglass holder, dipping the tail in 40 °C warm water to dilate the tail vein, and injecting the compound into the vein using a 30-gauge needle. All compounds were dissolved in either distilled water (i.c.v. injections) or physiologic saline (i.v. injections).

Antinociception was assessed using a 55 °C warm water tail-flick assay. Baseline latencies were determined for each mouse by immersing the distal third of the tail into the water. Typical latencies were 1-2 seconds, and mice with latencies over 5 seconds were excluded from further procedures. The test compound was administered at t = 0 minutes, and tail-flick latencies were measured at various time points following injection. A maximum score of 100% was given if the mouse did not respond within 10 seconds to avoid tissue damage. Antinociception was calculated using the formula:

(푡푒푠푡 푙푎푡푒푛푐푦 − 푏푎푠푒푙𝑖푛푒 푙푎푡푒푛푐푦) % 푎푛푡𝑖푛표푐𝑖푐푒푝푡𝑖표푛 = 푥 100 (10 − 푏푎푠푒푙𝑖푛푒 푙푎푡푒푛푐푦)

3. Design and synthesis of glycopeptide analogs of the secretin peptides

VIP and PACAP

Modifying secretin family peptides require an understanding of how each ligand interacts with its respective receptor. In particular, class B, or secretin family, GPCRs undergo a two-stage receptor activation mechanism, in which the α-helical C-terminus

(“address”) of the ligand binds to the ECD, after which the N-terminus (“message”) enters the TM domain. The complexity of this two-stage mechanism is further amplified by the requirement that the address segment of the peptide remains in a helical conformation to fit in the binding groove on the ECD. The total surface area needed for activation when compared to class A GPCRs highlights this challenge (Figure 3-1); whereas the class A receptor only requires a small peptide or small molecule in the NT domain, class B receptors have that same NT surface (not pictured), as well as the much larger ECD surface.

3.1. Molecular modeling of secretin peptides

A major issue with modeling secretin family ligand-receptor interactions is that the exact mechanism and structure of those interactions is ambiguous, as only inactive- state (antagonist bound, for instance) crystal structure and NMR data are available. To address this challenge, Dr. Bobbi Anglin designed a homology model of the PAC1R based on the crystal structure of secretin family receptor CRF1R (PDB 4K5Y).39 Because evidence points to PACAP binding to a different binding groove on the receptor ECD than the rest of the secretin peptides, we designed a separate model for VIP in receptor.

The preliminary secretin family glycopeptide analogs were prepared by synthesizing the native, amidated peptide and glycosylated forms of each (Table 3-1). A series of N-terminus truncated peptides, proposed to act as antagonists at the PAC1R,

VPAC1, and VPAC2 were also prepared. These peptides were synthesized in fairly low yield due to the formation of numerous side products (Figure 3-2), although the majority of side product formation occurs during the addition of the N-terminal five amino acids.

Identities were confirmed using MS analysis.

# Peptide Sequence

6 PACAP27 HSDGI FTDSY SRYRK QMAVK KYLAA VL 7 PACAP26- HSDGI FTDSY SRYRK QMAVK KYLAA VS* S(OGlc) 8 PACAP26- HSDGI FTDSY SRYRK QMAVK KYLAA VS** S(OLac) 9 PACAP6-27 FTDSY SRYRK QMAVK KYLAA VL 10 PACAP6-26- FTDSY SRYRK QMAVK KYLAA VS* S(OGlc) 11 PACAP6-26- FTDSY SRYRK QMAVK KYLAA VS** S(OLac) 12 VIP28 HSDAV FTDNY TRLRK QMAVK KYLNS ILN 13 VIP28-G HSDAV FTDNY TRLRK QMAVK KYLNS ILNG 14 VIP28-S(OGlc) HSDAV FTDNY TRLRK QMAVK KYLNS ILNS* 15 VIP28-S(OLac) HSDAV FTDNY TRLRK QMAVK KYLNS ILNS** 16 VIP6-28-G FTDNY TRLRK QMAVK KYLNS ILNG 17 VIP6-28-S(OGlc) FTDNY TRLRK QMAVK KYLNS ILNS* 18 VIP6-28-S(OLac) FTDNY TRLRK QMAVK KYLNS ILNS** Table 3-1: Identity of initial PACAP (6-11, synthesized by Dr. Bobbi Anglin) and VIP analogs, where the full sequences were designed as receptor agonists, and the N-terminal truncated sequences as antagonists. S* = Ser(OGlc), S** = Ser(OLac).

VIP28

Figure 3-2: HPLC trace of native VIP28 (12) prepared by automated synthesis. The arrow indicates the desired peak. 3.3. Oxidation of methionine residues

As peptide syntheses have become more complex, some residues and sequences have been recognized to potentially have deleterious post-cleavage modifications. One of these, oxidation, occurs primarily with the sulfur-containing amino acids cysteine and methionine, and tryptophan.

After isolation of the product peptide peaks by HPLC, a second peak slowly formed for both PACAP and VIP analogs, despite storage of the lyophilized samples in a

-80 ⁰C freezer. MS analysis showed the desired parent peaks (M + H+), as well as (M +

17) and (M + 33) peaks. The sequences of the peptides only contain one readily oxidizable residue, and thus the impurity peaks were assigned as oxidation states of methionine (Met) to the corresponding sulfoxide and sulfone.188 While it is possible to enzymatically reduce the methionine residues found in the secretin peptides after cleavage from resin,189 multiple research groups have shown that substitution of Met17 with a sterically similar, all-hydrocarbon residue such as leucine (Leu) or norleucine 90

(Nle) does not significantly affect receptor binding or activity.189-192 Indeed, it has been suggested that Leu17-containing PACAP may be more active at the PAC1R.142 Therefore, all subsequent secretin-family peptide analogs synthesized contain either Leu17 or Nle17.

These modifications led to the second generation secretin family glycopeptides listed in

Table 3-2. N-terminal truncated analogs were not prepared (see 4.1.2. Results: PAC1R agonists and antagonists for a detailed discussion). Identities of the second generation glycopeptides (HPLCs: Figure 3-3) were confirmed by MS.

# Peptide Sequence

19 PACAP27-L17 HSDGI FTDSY SRYRK QLAVK KYLAA VL

20 PACAP26-S(OGlc)-L17 HSDGI FTDSY SRYRK QLAVK KYLAA VS*

21 PACAP26-S(OLac)-L17 HSDGI FTDSY SRYRK QLAVK KYLAA VS** 22 VIP28-G-L17 HSDAV FTDNY TRLRK QLAVK KYLNS ILNG 23 VIP28-S(OGlc)-L17 HSDAV FTDNY TRLRK QLAVK KYLNS ILNS* 24 VIP28-S(OLac)-L17 HSDAV FTDNY TRLRK QLAVK KYLNS ILNS** 25 VIP28-G-£17 HSDAV FTDNY TRLRK Q£AVK KYLNS ILNG 26 VIP28-S(OGlc)-£17 HSDAV FTDNY TRLRK Q£AVK KYLNS ILNS* 27 VIP28-S(OLac)-£17 HSDAV FTDNY TRLRK Q£AVK KYLNS ILNS** 28 VIP6-28-G-L17 HSDAV FTDNY TRLRK QLAVK KYLNS ILNG 29 VIP6-28-S(OGlc)-L17 HSDAV FTDNY TRLRK QLAVK KYLNS ILNS* 30 VIP6-28-S(OLac)-L17 HSDAV FTDNY TRLRK QLAVK KYLNS ILNS** 31 VIP6-28-G-£17 HSDAV FTDNY TRLRK Q£AVK KYLNS ILNG 32 VIP6-28-S(OGlc)-£17 HSDAV FTDNY TRLRK Q£AVK KYLNS ILNS* 33 VIP6-28-S(OLac)-£17 HSDAV FTDNY TRLRK Q£AVK KYLNS ILNS** Table 3-2: Second generation secretin family glycopeptides, with Leu17 or Nle17 (underlined) in place of the oxidizable methionine residue. S* = Ser(OGlc), S** = Ser(OLac), £ = Nle.

A VIP28-G-L17

B VIP28-S(OGlc)-L17

C VIP28-S(OLac)-L17

A particularly problematic series of unwanted side reactions in peptide synthesis can occur when an aspartic acid (Asp) residue is present the peptide sequence. In many conditions, the neighboring amide nitrogen can undergo a nucleophilic attack on the protected side-chain ester of the Asp (Scheme 3-1), leading to aspartimide (Asi, also

“Asu” in literature) ring formation.193

Scheme 3-1: Formation of an aspartimide can occur under acidic or basic conditions. Once formed, the ring can be opened by a nucleophile at either carbonyl, leading to a) the peptide chain converting to an isoaspartyl-β-peptide, or b) to the nucleophile replacing the X group with the desired peptide backbone.

Formation of Asi is catalyzed under both acidic and basic conditions (Scheme

3-1), which are present during Fmoc-based SPPS during Fmoc deprotection (piperidine and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU)), amino acid coupling, and resin/side chain protection cleavage. In this reaction, the nitrogen of the amino acid on the C- terminal side of Asp performs a nucleophilic attack on the protected side-chain ester of

Asp (as Asp is typically protected as a tert-butyl ester). This leads to formation of the

Asi. When the Asi is present, nucleophiles in the reaction can open the ring via attack at either succinimidic carbonyl. Attack at the β-carbonyl results in either the original α- peptide with an unmodified side-chain (in aqueous conditions), or with the nucleophile bound to the Asp carbonyl (for instance, a piperidide when using piperidine in Fmoc deprotection steps). More pernicious is attack at the α-carbonyl, which results in a modified peptide backbone, an isoaspartyl-β-peptide. This product has identical mass

93 with the desired peptide, and can be very difficult to separate. In addition, if hydrolysis opens the succinimide at either position, the result is a free carboxylic acid, which is susceptible to further chain elongation amino acid coupling steps. This can rapidly lead to a wide range of undesired branched peptides.

The most susceptible residue sequences in basic conditions are those with Gly,

Ala, and Asn.193 As PACAP has the Asp3-Gly4 sequence (as well as Asp8-Ser9), and VIP contains both Asp3-Ala4 and Asp8-Asn9, Asi formation is a significant concern in secretin peptide synthesis and was identified as a major reason for the low yields observed with the PACAP and VIP glycopeptide analogs.

One way to address this is by installing the Asp as part of a dimer with the amide nitrogen protected, for instance as the commercially available Fmoc-Asp(OtBu)-

HmbGly-OH, which blocks the nucleophilic site. However, this is the only readily available protected dimer, and it is fairly expensive (~$500/g). Therefore, an alternative approach was taken.

Other approaches to reduction of Asi formation in peptides include modifying the protecting group on the aspartic acid and forming a pseudoproline between Asp and either Ser, Thr, or Cys,194 but those have been shown to be minimally effective or only beneficial in select circumstances. An approach that appeared promising was through modification of the bases used during the Fmoc deprotection step. Bodanszky and coworkers reported that the use of N-hydroxylamines such as hydroxybenzotriazole

(HOBt) decreased Asi formation.195,196 It was reported that during the synthesis of CRF, use of a weaker base, piperazine, in place of piperidine led to decreased Asi formation.197

In each of these cases, the modifications are proposed to decrease the nucleophilicity and basicity of the nitrogen adjacent to the aspartic ester (Figure 3-4).

Figure 3-4: Additives to reduce aspartimide formation. pKa represents accepted values for the conjugate acids of DBU, piperidine, and piperazine.

Initial results from these modifications appeared more pure than the first series of glycopeptides (Figure 3-3), so the second generation glycopeptides and all subsequent secretin-class glycopeptide analogs were synthesized using HOBt/piperazine deprotection conditions.

3.5. Materials and methods

3.5.1. Peptide synthesis

Unless otherwise noted, all solvents were obtained from EMD Chemicals

(Gibbstown, NJ), and used without further purification. Fmoc-protected amino acids

Fmoc-Ala-OH, Fmoc-Lys(Boc)-OH, Fmoc-Leu-OH, Fmoc-Met-OH, Fmoc-Tyr(tBu)-

OH, and condensing agent N,N’-diisopropylcarbodiimide (DIC) were acquired from

Advanced ChemTech (Louisville, KY, USA). Fmoc-Asp(tBu)-OH was acquired from

Oxchem Corporation (Los Angeles, CA, USA). Fmoc-Phe-OH, Fmoc-Gly-OH, Fmoc-

Ile-OH, Fmoc-Gln(Trt)-OH, Fmoc-Ser(tBu)-OH, and Fmoc-Val-OH were acquired from

Chem-Impex (Wood Dale, IL, USA). Fmoc-His(Trt)-OH, Fmoc-Arg(Pbf)-OH, Fmoc-

Thr(tBu)-OH, Fmoc-Nle-OH, and Hydroxybenzotriazole hydrate (HOBt•H2O) were obtained from AAPPTec (Louisville, KY, USA). 1,8-diazabicyclo[5.4.0]undec-7-ene

(DBU) from TCI America (Portland, OR, USA), piperidine from Sigma-Aldrich (St.

Louis, MO, USA), piperizine from Alfa Aesar (Ward Hill, MA, USA), O-(7-

Azabenzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium hexafluorophosphate (HATU) from Accela ChemBio Co., Ltd (Hong Kong, China), 2,4,6-trimethylpyridine (TMP) from Acros Organics (Geel, Belgium), and 1-Methyl-2-pyrrolidinone from Avantor

(Center Valley, PA, USA). The Fmoc-protected serine glycosides Fmoc--OGlc(OAc)4-

Ser-OH and Fmoc--OLac(OAc)7-Ser-OH were prepared in the Polt lab using previously published conditions.86

VIP and PACAP peptide and glycopeptide assembly was accomplished using a combination of manual and automated peptide methods.33 A general procedure follows:

The C-terminal amino acids were loaded onto Fmoc-Rink resin (Advanced ChemTech,

Louisville, KY, USA) at 0.1 mmol/g resin loading in 25 mL fritted syringes. Initially, the resin was swelled using DMF (~5 mL solvent per gram resin), agitating at RT for two minutes (x2). A solution of 2% DBU and 3% piperidine in DMF (v:v) was introduced and agitated for 5 minutes, refreshed, and agitated for an additional 10 minutes. The resin was then washed with DMF (x5), and NMP. In a separate vial, Fmoc--OGlc(OAc)4-Ser-

OH (0.12 mmol, 1.2 eq) was dissolved in 5 mL NMP, and HOBt•H2O (0.13 mmol, 1.3 eq) was added and allowed to mix for 5 minutes. Condensing agent DIC (0.26 mmol, 2.6 eq) was then added, and mixed for 5 minutes. This solution was added to the resin and

96 agitated for 10 minutes. Next, the syringe was placed in a microwave (Emerson 900W

Microwave – MW9338SB) set to power level 1 and irradiated for 10 minutes, stopping to shake the syringe every 90 seconds. The syringe was then agitated at RT for an additional

30 minutes. The resin was washed with NMP (x1), DMF (x5), and CH2Cl2 (x5), and dried in vacuo overnight.

Peptides and glycopeptides also were assembled on a Protein Technologies, Inc.

(Tucson, AZ, USA) Prelude® Peptide Synthesizer using the reaction scheme that follows.

Rink resin (100 mg) was placed into the fritted reaction vessels (RVs). Amino acids were dissolved in DMF at 250 mM concentration, HATU at 375 mM, and TMP at 3 M. The following steps were performed for coupling: DMF Top Wash (1.5 mL, 2 min mix and drain; x6), Deprotection (2% DBU/3% piperidine in DMF; 1.5 mL, 4 min mix and drain;

8 min mix and drain), DMF Top Wash (1.5 ml, 2 min mix and drain; x5), Amino Acid

Building Block (0.950 mL, 30 sec mix), Activator 1 (HATU, 0.650 mL, 30 sec mix),

Base (TMP, 0.300 mL, 35 min mix and drain), DMF Top Wash (1.5 mL, 2 min mix and drain; x2). After coupling aspartic acid D7, the deprotection solution was changed to 0.1

M HOBt•H2O/5% piperazine in DMF to minimize aspartimide formation.

Cleavage of peptides and glycopeptides from the resin was accomplished using a

TFA “cocktail” of F3CCOOH:Et3SiH:H2O:CH2Cl2:Ph-OCH3 (v:v:v:v:v,

9:0.3:0.2:1:0.05), agitating at RT for 2 hours. The resulting solution was expelled into a

15 mL centrifuge tube, evaporated under argon, precipitated in ice-cold Et2O, decanted, and rewashed with Et2O, then dissolved in H2O and lyophilized to afford fluffy, white solid.

Purification of the crude peptides and glycopeptides was accomplished by Reversed

Phase HPLC (RP-HPLC) with a preparative RP (C-18) Phenomenex (250 X 22 mm) column using a CH3CN-H2O gradient solvent system containing 0.1% F3CCOOH.

Fractions containing pure glycopeptide were combined and lyophilized. Homogeneity of the purified glycopeptides was confirmed by analytical RP-HPLC and high resolution mass spectrometry.

4. Secretin glycopeptide analysis and biological activity

Analysis of secretin family glycopeptides was accomplished in two ways: first, receptor activation was measured using fluorescence measurements; and second, by measuring their degradation in rodent serum samples.

4.1. Receptor activation analysis

Studies of receptor activation were performed in the laboratory of Dr. John

Streicher by Dr. Streicher and technician Katie Edwards at the University of New

England. This has biological significance because of the recognizance that both VIP and

PACAP promote calcium signaling in order to produce cell differentiation.198

4.1.1. Background

In order to quantify receptor activation, CHO or human embryonic kidney (HEK) cells are transfected to overexpress the human version of the PAC1R, VPAC1, or VPAC2 receptor. Activation is visualized using the Fluorescence Imaging Plate Reader (FLIPR) platform, which enables the measurement of calcium ion flux in real-time. This flux is measured through monitoring the conjugation of Ca2+ with an introduced fluorescent dye

(Calcium 6, Molecular Devices, USA). Briefly, one of the results of GPCR activation is

Ca2+ is release. The released Ca2+ subsequently conjugates to the dye and leads to a shift in the maximum fluorescence wavelength (λmax). As the dye’s extinction coefficient is known, the concentration of fluorescence-shifted dye can be accurately determined, thereby acting as a proxy for receptor activation. In our studies, receptor activation is normalized against commercially available PACAP27 or VIP28.

4.1.2. Results: PAC1R agonists and antagonists

For agonism experiments, cells expressing the PAC1R were incubated with each ligand, and the fluorescence output was measured and normalized to commercially available PACAP27 (100%) and vehicle (0%). N = 3 for each experiment, and mean and

SEM values were determined from curve fit.

Antagonism was measured using both variable and fixed concentration experiments. For variable concentration experiments, concentration curves of commercially available PACAP6-38 and synthesized PACAP6-27 were added to cells expressing the PAC1R, and after 2 minutes 5 nM PACAP27 was introduced. Response rates were determined in relation to 5 nM PACAP27 (max) and vehicle (min). Fixed concentration experiments introduced PACAP6-38 or PACAP6-27 to the cells for 2 minutes, followed by concentration curves of PACAP27. Response rates were normalized to max activation of the receptor by PACAP27 without antagonist (max) and with a vehicle (min).

Receptor activation results from a series of PACAP analogs expected to work as both agonists (Figure 4-1, Table 4-1) and antagonists (Figure 4-2, Table 4-2) were tested at the PAC1R expressed in CHO cells. For the PACAP analogs described, the nomenclature “PACAP27S-G” represents the substitution of Leu27 with a serine glucoside (see Section 3.2., Table 3-1, entry 6 for sequence of PACAP27, and entry 7 for

PACAP27S-G).

100

Figure 4-1: Agonist activity curves of PACAP derivatives at the PAC1R, measured by FLIPR. The native and glucosylated versions show good agonism, while the truncated forms show no activity. Courtesy of Katie Edwards.

PAC1R Agonist Data

Drug EC50 (nM) EMax (%)

PACAP27 0.95 ± 0.4 100 PACAP27S-G 5.7 ± 2.3 102 ± 2 PACAP6-27 NC PACAP6-27S-G NC PACAP6-27S-L NC PACAP6-27-Alt NC PACAP6-27SG-Alt NC PACAP6-27SL-Alt NC Table 4-1: Calculated agonist activity for PACAP derivatives at the PAC1R. NC = not converged, meaning no activity was observed.

101

Figure 4-2: Antagonist activity of PACAP derivatives at the PAC1R, measured by FLIPR. A. Variable concentration antagonism experiments with all proposed antagonists. B. Fixed concentration antagonism with PACAP6-38. C. Fixed concentration antagonism with PACAP6-27. Courtesy of Katie Edwards.

Based on the in vitro data presented above, it appears that this assay closely matches previously reported PACAP27 receptor activation values.101 This suggests that comparisons using data generated from this assay are valid as a starting point for PACAP analog agonist activity testing.

102

PAC1R Antagonist Data (Variable)

Drug IC50 (nM) IMax (%) PACAP6-38 >333 (79) PACAP6-27 NC PACAP6-27S-G NC PACAP6-27S-L NC PACAP6-27-Alt NC PACAP6-27SG-Alt NC PACAP6-27SL-Alt NC

PAC1R Antagonist Data (Fixed) Drug pA2 (nM) Schlid Slope PACAP6-38 200. ± 55 2.0 ± 0.1 PACAP6-27 NC Table 4-2: Calculated antagonist activity at the PAC1R using variable or fixed concentration experiments.

4.1.3. Results: VPAC1 and VPAC2 agonists

Agonist activity of VIP analogs at the VPAC1 and VPAC2 receptors was determined using the same methodology as the PAC1R experiments (see 4.1.2. Results:

PAC1R agonists and antagonists), using commercially acquired VIP28 as the standard.

Three VIP28 analogs were tested at both the VPAC1 (Figure 4-3, Table 4-3) and VPAC2

(Figure 4-4, Table 4-4) receptors, along with native VIP. In the analog cases, the modification (e.g. “-G”) indicates an amino acid appended to the C-terminus of the native

VIP28 sequence (G29, in that example). Their sequences can be found in Section 3.3.,

Table 3-2, entries 22-24, while VIP28 is in Section 3.2, Table 3-1, entry 12.

103

Figure 4-3: Agonist activity of VIP28 and derivatives at the VPAC1 receptor using FLIPR assay. Courtesy Katie Edwards.

VPAC1 Agonist Data

2 Drug EC50 (nM) EMax (%) R

VIP 24 52 0.97 VIP-G 15 50. 0.97 VIP-S-OGlc 7.5 58 0.96 VIP-S-OLac 19 57 0.97 Table 4-3: Calculated agonist activity of VIP28 and analogs at the VPAC1 receptor using FLIPR assay.

104

Figure 4-4: Agonist activity of VIP28 and derivatives at the VPAC2 receptor using FLIPR assay. Courtesy Katie Edwards.

VPAC2 Agonist Data

2 Drug EC50 (nM) EMax (nM) R

VIP <333.3 (47) 0.95 VIP-G <333.3 (29) 0.88 VIP-S-OGlc <333.3 (48) 0.94 VIP-S-OLac <333.3 (45) 0.94 Table 4-4: Calculated agonist activity of VIP28 and analogs at the VPAC2 receptor using FLIPR assay.

The experimentally determined agonist activity of VIP28 at the VPAC1 receptor is within an order of magnitude of literature data.101 However, multiple sources suggest that VIP28 should either have similar activity at the VPAC2 receptor,101 or that VIP activation of the VPAC2 receptor occurs in the sub-nanomolar regime,100,122 and in our assay we do not see evidence of this.

105

4.2. PACAP and VIP analog degradation & BBB penetration using MSn analysis

4.2.1. Background

In order to determine whether glycosylation of peptides improves both resistance to enzymatic degradation and BBB penetration, we collaborated with Dr. Michael

Heien’s lab at the University of Arizona. Dr. Nick Laude and Catherine Kramer have pioneered and expanded the utility of a method that has seen effective, if intermittent, use at measuring the presence of biologically relevant small molecules and peptides.199-201 In particular, their set-up allows “on-line” preservation and monitoring, where dialysate from the CNS is introduced into the LC-MSn (tandem MS) system using readily accessible, off-the-shelf components.

4.2.2. Collision-Induced Decay

In order to be able to track the degradation over time of a particular molecule, a unique identification tag needs to be generated for each compound. In this system, a fragmentation profile is generated by fragmenting a peptide at increasing energy through a process called collision-induced decay (CID), and recording which fragments are most prevalent via tandem MS (MS2, Figure 4-5). To further improve the selectivity of the process, the most prevalent fragment can be selectively isolated (by separating a specific mass/charge ratio from all other ions in the gas phase), and subsequently fragmented again (MS3).

106

107

Figure 4-6: CID fragmentation data for VIP28-S* (M17L) (23). (A) shows fragmentation occurs exclusively around the C-terminus. (B) shows the fragmentation states at +4 and +5 in terms of overall signal (left) and relative quantities (right). Courtesy Catherine Kramer.

108

109

4.2.3. Serum degradation

In vitro degradation and in vivo BBB penetration experiments were both performed using the same general procedure (Figure 4-8) by Dr. Nick Laude and

Catherine Kramer in the laboratory of Dr. Michael Heien at the University of Arizona. In serum degradation experiments, peptide is added to serum at 37 ⁰C, and at predetermined timepoints small aliquots of the mixture are removed. Acetic acid, to denature proteolytic enzymes, and an internal standard are then added to each sample. Samples are purified via a C18 ZipTip™, and injected onto an LCQ MS. In the MS, the parent ion is identified, isolated, and further fragmented (via MSn) to confirm the identity of the compound.

Impressively, this methodology can be applied to multiple analytes in a single serum sample concurrently, leading to a “shotgun” analysis approach.

Figure 4-8: Schematic representing the procedures to determine degradation rates (top) and BBB penetration (bottom) of peptides.

For BBB penetration studies, the freely moving rodent is fitted with a microdialysis tube that inserts into the region of the brain that is being probed.202

Dialysate from the brain is then mixed with AcOH and an internal standard in an on-line preservation step, then purified and analyzed via nano liquid chromatography-MS3

110 analysis. With endogenous neurotransmitters, including the peptides Leu- and Met- enkephalin, trace quantities could be detected in dialysate (limit of detection = 1.5 amol), suggesting this methodology could be applied advantageously to BBB penetration studies, in which we expect concentrations in the nM range.

Initial degradation studies using PACAP27 analogs were performed in a single experiment (“shotgun” approach, n = 3), in which each peptide was tested at approximately the same concentration (Figure 4-9, for full data see Figure 8-1 in

Appendix B. Supplementary Data).After fitting the concentrations over time, it was determined that the time to half maximal signal of the PACAP27 analogs were:

PACAP27 – 3.2 min; PACAP26-S* - 9.8 min; PACAP26-S** - 2.3 min.

4.3. Discussion

4.3.1. PAC1R, VPAC1, and VPAC2 receptor agonism and antagonism

PAC1R activity was tested on PACAP27 and PACAP26-S*, as well as a variety of PACAP6-27 analogs, using the FLIPR Ca2+ flux assay by Dr. John Streicher and Katie

Edwards at the University of New England. The PACAP6-27 analogs did not display any agonist activity at the receptor within the concentration range tested. We identified two encouraging agonist results from the PACAP27 analogs: first, our observed activity is close to reported PACAP27 binding and activity at the PAC1R;88,101 and second, that there is only a minimal decrease in activity with the addition of a carbohydrate. This suggests that the glycoside has is affecting the binding of the peptide to the receptor, as the modification is made to the C-terminal address of the peptide, while the N-terminus message sequence is unchanged. In addition, the PACAP analog tested at the PAC1R included the native Met17. Oxidation of the peptide to include methionine sulfoxide may

111 occur while in storage, transport, or during sample preparation, in turn decreasing the peptide’s biological activity.

Figure 4-9: Serum degradation of PACAP27 and two glycosylated analogs (A). Half-life of each is shown in B.

In contrast with the encouraging agonist results for PACAP27 and our glycosylated analog, we do not see substantial antagonism for any of the N-terminal truncated peptides tested. Literature shows that PACAP6-38 is an antagonist of PACAP at the PAC1R, and strongly suggests that the shorter PACAP6-27 also acts as an antagonist at that receptor.132 In our initial variable concentration experiment (Figure 4-2,

112

A), we do not see any antagonist activity for PACAP6-27 (reported IC50: 80 nM), and only slight activity for the extended PACAP6-38 (measured >333 nM, reported IC50: 80 nM). When comparing our results with that from Robberecht and coworkers, we recognized that their assays were run using a fixed concentration experiment, which provides higher sensitivity at the cost of using greater quantities of reagents. Variable concentration experiments only show relative potency. To validate our results we performed fixed concentration antagonist assays of PACAP6-38 and PACAP6-27 (Figure

4-2, B and C). In our hands we see a slight increase in antagonism for PACAP6-38 to 200 nM (via calculation of pA2, which is essentially the log of the binding constant, and is more pharmacologically relevant than IC50). PACAP6-27 still showed no antagonist activity.

We propose two reasons why our experiments show different results than those of

Robberecht and coworkers. First, their experiments were run using neuroblastoma cells.

These cells were isolated from living systems, which means that there are a variety of receptors present on their surfaces, and the potential exists for a number of off-target

PACAP binding sites. In contrast, our CHO cells only express PAC1R, meaning that our measurements are exclusively of our PACAP analogs at the desired receptor. Second, when analyzing the antagonism of PACAP6-38, we noted that the Schild slope of our experiment was 2. Generally, a Schild slope of 1 indicates competitive, reversible antagonism, and variation from that shows that one of those assumptions is not true.203

Therefore, we hypothesize that the behavior of PACAP6-38 may not be of a competitive antagonist, although further work is needed to validate this suggestion.

113

Similarly to the PAC1R agonist studies, testing VIP28 analogs for agonist activity at the VPAC1 receptor was broadly successful, with observed agonism with both the presence and absence of a carbohydrate at broadly the same concentration as suggested by binding data in the literature.101 However, unlike PACAP26-S*, we saw slightly improved activation of the VPAC1 receptor for the glycosylated peptides, as well as the

Gly-extended analog. This is likely due to one of three reasons. First, as previously mentioned, the oxidizable methionine residue was substituted out in the synthesized peptides, so that native VIP28 has similar receptor interactions regardless of whether an additional residue is present. Second, that appending the carbohydrate to the end of the peptide instead of substituting the final residue (as in the PACAP26-S* case) may conserve full address binding, whereas substitution could lead to the loss of an interaction between the ligand and receptor. Third, the extended VIP28-Xxx analogs were all prepared and purified in lab, and their purity has been confirmed (Figure 3-3). In contrast,

VIP28 was acquired from an outside source and may be lower purity, so that adding the same apparent molar quantities of synthesized and purchased VIP28 actually results in slightly different concentrations, and therefore affects the activation profiles. Dr.

Streicher acquired VIP28 from Tocris (catalog #1911), and while they claim high purity

(~96%), the sample purity was not confirmed in our lab.

Testing of the VIP28 analogs for agonism at the VPAC2 did not show significant activation. This directly contradicts reports that suggest that VIP28 activates the VPAC2 receptor at sub-nanomolar concentrations.100,122 We believe this is due to the observation that VPAC1 receptors can selectively couple to RAMP proteins, which induces a change in the concentration of calcium released.204,205 In contrast, VPAC2 receptors do not

114 appear to interact with RAMP proteins;205,206 as we are using a calcium flux assay, we may be seeing increased potency at the VPAC1 receptor due to the assay we are using.

We have not yet run these experiments using a different assay.

Finally, we have shown that VIP28 and PACAP27 both activate the VPAC1 receptor at similar concentrations as has been reported in literature (Appendix B. Supplementary

Data, Figure 8-2),101 although only a single experiment (n = 1) was performed and we have not yet tested further PACAP analogs at the VPAC1 receptor.

4.3.2. Peptide fragmentation and degradation

The CID fragmentation data for the VIP analogs shows how much more complicated larger peptides are to analyze using this particular approach. Each fragmentation of five amino acid peptide Leu-enkephalin202 shows a good signal to noise ratio, and the major fragment retains >80% of the max signal. This means that this analysis could have a very low limit of detection, as little signal is lost. However, it should be noted that this may be specific to the enkephalins. In contrast, the 29 residue

VIP analogs have many more locations for fragmentation, both throughout the peptide backbone and around the carbohydrates, and therefore much more of the signal decays.

However, it appears that glycosylation improves fragmentation, as breaking of the glycosidic bonds (either between the glycoside and the serine, or, in the case of the lactoside, between the glucose and galactose moieties) is the predominant fragmentation pathway. The major MS2 fragment in VIP28-S* and VIP28-S** retains 60-80% of the signal, while in the non-glycosylated VIP28 the major fragment retains about 10% of the signal.

115

One potential benefit of larger peptides is that a second fragmentation to MS3 may be unnecessary. With small molecules from crude samples, there exists the possibility that another molecule in the mixture may have the same parent ion. In contrast, because the secretins are much larger, between 3-5 kDa, and are more likely to exist in higher order charge states, their parent ions are expected to be unique, and therefore only a single tandem MS isolation is necessary to support their identities.

The degradation experiment is encouraging, although it provokes further questions. Initially it appears that glycosylation improves resistance to enzymatic degradation in the case of glucose, but not lactose, and that adding a lactose decreases the half-life of the peptide when compared with native PACAP27. We have initially attributed this result to the relationship between lactose and glucose; namely, that lactose is β-D-galactose-(14)-β-D-glucose. Therefore, cleavage of the galactose moiety from

PACAP26-S** would result in the presence of PACAP26-S* (Scheme 4-1). As the MS analysis only looks for specific ions, there is no way to distinguish between the introduced PACAP26-S* and the primary degradation product of PACAP26-S** in serum when utilizing the “shotgun” analysis approach without modifying one of the peptides. Due to this hypothesis, we propose that peptide lactosides act as pseudo- prodrugs for their related glucosides, as it appears that the cleavage of the Glc-Gal bond occurs much more quickly than cleavage of peptide backbones. We also expect that addition of a glucose moiety improves resistance to enzymatic degradation, but not to the extent that the values we calculated suggest.

One surprise from the degradation data was that the signal of each of the peptides stabilized at 30-50% of initial concentration, as we would not expect the enzymes to

116 cease functioning. The continued degradation of PACAP26-S* after the concentrations of the other two analogs stabilize around 20 minutes (Figure 4-9) suggests that viable proteolytic enzymes are present during the entire experiment. In addition, the limit of detection in the degradation experiments is typically around 1% of maximum signal, meaning that we should not be reaching that limit. Therefore, we hypothesize that amphipathic peptide aggregation, caused by the hydrophobic effect, effectively protects a portion of the peptide from enzymes. Literature supports this hypothesis, as Onyuksel and coworkers noted that VIP spontaneously aggregates into micelles at sub-micromolar concentrations.123 Conjugating a carbohydrate to a peptide increases its hydrophilic surface area at one end of the peptide, making it look more like a traditional self- aggregating amphiphile (such as sodium dodecylsulfate, or SDS) (Figure 4-10). We would therefore expect that self-aggregation occurs more easily with glycopeptides, which explains why a greater proportion of PACAP26-S** (and likely PACAP26-S*, although that curve does not stabilize over the course of 60 minutes) remains undegraded when compared with the native, non-glycosylated PACAP27.

Scheme 4-1: Proposed degradation of PACAP26-S** into PACAP26-S* via galactosidase action.

117

Figure 4-10: Representation of the hydrophilic (blue) and hydrophobic (red) areas of both VIP28-S* and SDS. Note that the red portion of the peptide adopts a helical conformation.

118

4.4. Materials and methods

4.4.1. Secretin receptor cell lines and cell culture

All cell experiments were performed in the laboratory of Dr. John Streicher at the

University of New England by Dr. Stricher and Katie Edwards. Custom DNA clones of the human PAC1, VPAC1, and VPAC2 genes with 3 hemagglutinin (HA) tags inserted 3’ to the signal peptide sequence (to avoid proteolytic loss) were obtained from

Genecopoeia (Rockville, MD, USA). A general procedure follows. The construct was electroporated into Chinese hamster ovary (CHO) cells and selected for with 500 µg/mL of G418. The resultant population was individually screened for high expressing clones, and one such clone selected for further analysis. The clonal cell line (PAC1-CHO) showed high receptor selectivity by Western blot and immunocytochemistry, and showed selective activation of signaling in response to PACAP1-27 (VIP1-28 for VPAC1 and

VPAC2 cell lines). The cell line expressing the specific receptor of interest was used in all subsequent molecular pharmacology experiments. Cells were maintained in

DMEM/F12 with 10% heat-inactivated FBS, 1X penicillin/streptomycin, and 500 µg/mL

G418, at 37 °C in 5% CO2 atmosphere.

4.4.2. Fluorescent Imaging Plate Reader (FLIPR) experiments

Experiments performed in the laboratory of Dr. John Streicher as above. All molecular pharmacology experiments were performed using a FLIPR Tetra from

Molecular Devices (Sunnyvale, CA, USA), using the manufacturer’s recommended settings and protocols and set to image calcium flux. One day prior to the experiment, cells from the appropriate cell line (for instance PAC1-CHO) were split into 384 well black-walled, clear bottom microplates, 10,000 cells per well. The cells were then

119 recovered overnight in grown medium as above. On the day of the experiment, growth medium was replaced with Calcium 6 dye (Molecular Devices) using the manufacturer- recommended buffer with 2.5 mM probenecid. The cells were incubated for 2 hours in the culture incubator, and removed during the final 15 minutes to allow equilibration to

RT. Compound as indicated below was added to the cells using a 384 tip block, with real- time monitoring of the cells before, during, and 15 minutes after compound addition. The resulting calcium flux was recorded, and the maximum response over the entire observation time calculated and reported as the mean ± SEM (4 wells per point).

In agonist mode experiments, compound was added in an 11 point concentration curve, with a vehicle control (buffer). The resulting response was normalized to the stimulation caused by native ligand (100%, PACAP1-27 or VIP1-28) and buffer (0%).

The response was analyzed using a 3 variable non-linear curve fit, and the EC50 (nM) and

EMax (%) were calculated and reported (Prism, GraphPad, La Jolla, CA, USA).

In antagonist mode experiments, a concentration curve (variable concentration mode) or fixed amount (fixed concentration mode) of antagonist was added to the cells, and allowed to equilibrate for 2 minutes. Then, either a 5 nM fixed concentration

(variable concentration mode) or an 11 point concentration curve (fixed concentration mode) of PACAP1-27 was added to the cells, and the max-min response recorded as above. For variable concentration mode experiments, the data was normalized to the stimulation caused by 5 nM PACAP1-27 (100%) and vehicle (0%), and analyzed as above, with the IC50 and IMax calculated and reported (Prism). For the fixed concentration mode experiments, normalization was to the maximum stimulation of PACAP1-27 with no antagonist present (100%) and vehicle (0%). The data was analyzed using a

120

203 207 Gaddum/Schild EC50 shift model, designed to determine competitive antagonism.

The results obtained were the pA2 (nM) and the Schild slope, which measures how closely the experimental data fits the operational model of competitive antagonism

(Prism). For all analyses, each independent experiment performed in quadripulicate is considered to be a sample size of 1. The pharmacology values are calculated separately for each experiment, then combined and reported as the mean ± SEM for the entire set of experiments.

4.4.3. Peptide degradation

Peptide degradation experiments were performed in the laboratory of Dr. Michael

L.Heien at the University of Arizona by Dr. Nicholas D. Laude and Catherine L. Kramer as previously reported.202 Peptides and glycopeptides were incubated in mouse serum at

37 °C. 10 µL aliquots were removed at various time points beween 1-60 minutes and spiked with 1 µL of a 10 µM solution of DADLE ([D-Ala2, D-Leu5]-enkephalin,

American Peptide Company, Sunnyvale, CA, USA) in 50% AcOH, and subjected to a

C18 ZipTip® desalting process. For desalting, the ZipTip® was equilibrated with three

10 µL volumes of CH3CN, followed by three 10 µL volumes of aqueous 0.1% trifluoroacetic acid (TFA). A 10 µL volume of sample was drawn up and dispensed 10 times to load peptides onto the C18 resin. The sample was then washed with two 10 µL volumes of 0.1% TFA and eluted in 10 µL of 60% CH3CN, 40% H2O with 0.1% TFA.

Once eluted from the ZipTip®, solutions were diluted to 100 µL in 50:50 CH3CN:H2O with 0.1% formic acid. Tandem mass spectrometry analysis (MSn) was performed to yield specific, quantitative signals proportional to the remaining peptide concentration at each time point. MSn analysis was performed using the following instrumentation.

121

Electrospray ionization of the samples was achieved using a Thermo Finnigan LCQ

(ThermoFisher Scientific, Waltham, MA, USA). Mass analysis was carried out in the LIT with radial ejection of ions for sensitive detection. Tandem MS was conducted with one or two isolation-and-fragmentation steps (MS2 or MS3) for all species, with specific analyte ions identified using unique peptide data previously in this chapter.

122

5. Design, synthesis, and analysis of angiotensin(1-7) glycopeptide

analogs

Analysis of the angiotensin derivatives was performed in collaboration with a number of labs at the University of Arizona, including those of Dr. Michael Heien in the

Department of Chemistry and Biochemistry, Dr. John Konhilas and Dr. Meredith Hay in the Department of Physiology, and Dr. Todd Vanderah in the Department of

Pharmacology.

5.1. Design rationale and synthesis

As discussed previously, Ang(1-7) is an attractive starting point for drug discovery and development due to its regulation of a wide range of biological systems.

Dr. Meredith Hay in the University of Arizona’s Department of Physiology asked if we could make a few glycosylated analogs of Ang(1-7). We initially proposed synthesizing the native peptide (34), an amidated version (35), and three peptides that successively increase the hydrophilicity of the native version by appending a serine, a serine glucoside, and a serine lactoside (36-38, Table 5-1). Because our peptide synthesizer has six reaction vessels, we proposed a version that replaced the Pro7 with a serine glucoside

(39).

123

# Peptide Sequence Terminus 34 Ang(1-7) DRVYI HP -OH

35 Ang(1-7)-NH2 DRVYI HP -NH2

36 Ang(1-7)-S DRVYI HPS -NH2

37 Ang(1-7)-S* DRVYI HPS* -NH2

38 Ang(1-7)-S** DRVYI HPS** -NH2

39 Ang(1-6)-S* DRVYI HS* -NH2 Table 5-1: Initial batch of Ang(1-7) analogs proposed and synthesized. S* = Ser(OGlc), S** = Ser(OLac).

Synthesis generally proceeded as expected, with good initial purity of the crude peptides (Figure 5-1). Purification was facile, with most of the peptides needing only a single injection onto a preparative HPLC to achieve >90% purity. However, while 1.5-3.0 mg of most of the peptides were acquired, only trace amounts of 36 were obtained due to an inseparable impurity, so that analog was not subjected to in vitro testing.

124

34 – Ang-(1-7)-OH 95%

35 – Ang-(1-7)-NH2 >99%

36 – Ang-(1-7)-S 76%

37 – Ang-(1-7)-S* 89%

38 – Ang-(1-7)-S** 75%

39 – Ang-(1-6)-S* >99%

Figure 5-1: Ang-(1-7) derivative crude (left) and pure (right) HPLC traces. Red arrows show desired peak in crude, and percentages represent approximate purity based on UV absorbance. Note that traces for 39 are from a separate experiment. S* = Ser(OGlc), S** = Ser(OLac).

125

5.2. In vitro and in vivo results – cognition model

5.2.1. NO production via Mas receptor activation

To first explore whether our modifications to Ang-(1-7) produced compounds with pharmaceutical value, an assay to measure the release of nitric oxide (NO) by activated Mas receptors using the fluorophore 4-amino-5-methylamino-2′,7′- difluorofluorescein diacetate (DAF-FM diacetate, Figure 5-2) was carried out (Figure

5-3). Briefly, DAF-FM was pre-incubated with human umbilical vein endothelial cells

(HUVEC), which naturally express the Mas receptor (Figure 5-4). The Ang-(1-7) derivatives were then added at 100 nM, incubated for five minutes, and the resulting fluorescence was measured and normalized to background. We also tested compound 39 in the presence of selective Mas receptor antagonist A779 [D-Ala7-Ang-(1-7)] to confirm that the observed NO flux is due to activation of the Mas receptor.208

Figure 5-2: Structure of DAF-FM diacetate.

126

Figure 5-3: (A) Relative fluorescence produced by the exposure of HUVEC cells to Ang- (1-7) and derivatives. (B) Concurrent introduction of selective Mas antagonist A779 with A5 (39) eliminates NO production, showing that 39 is acting at the Mas receptor. Also compared with NO donor SNAP (5 µM). Courtesy Dr. John Konhilas and Dr. Meredith Hay.

As expected, addition of Ang-(1-7) resulted in NO production. More surprising was the observation that the glycosylated analogs 38 (PN-A4) and 39 (PN-A5) displayed roughly a twofold increase in NO production over the native peptide. This activation was confirmed to take place at the Mas receptor by the coadministration of 39 with selective

Mas antagonist A779 (Figure 5-3), which completely halted production of NO. We also compared the NO response of NO donor SNAP209 at 5 µM to peptide 39 with encouraging results.

We attribute the increased activity of the glycosylated peptides to the fact that, unlike the activation assays in Chapter 4, HUVECs form an endothelial layer, or behave more like a living system than a classic in vitro cell line. Therefore, the increased hydrophilic character that the carbohydrates provide a glycopeptide amphiphile, and the subsequent opportunity for the membrane hopping mechanism to be active, may explain the observed increased receptor activation.

127

5.2.2. GPCR activation via cell impedence measurements

An alternative method to determine the activity of compounds is by using a Real-

Time Cell Analyzer (RTCA) for assays. This method, which uses electrodes to measure cell impedance (CI, or changes in the conductance of a layer of cells after perturbation by a ligand), can be used to observe the presence (or absence) of physiological changes. This system does not provide information about what change is happening.

The following experiments were performed in the laboratories of Dr. John

Konhilas and Dr. Meredith Hay at the University of Arizona in the Department of

Physiology. To determine dosage parameters, HUVEC cells were exposed to glycopeptide 39 at three concentrations; 1 pM, 10 pM, and 100 pM (Figure 5-4). We report that the 10 pM dosage produced a three-fold greater CI response over the 1 pM dosage. Coupled with the observation that the CI response between the 10 pM and 100 pM dosages is slight allows us to determine that the optimal dosage is near 10 pM, as saturation of receptors occurs above that concentration. The previously reported binding

152 of radiolabeled Ang-(1-7) to Mas receptors is 0.83 nM (via KD calculations), suggesting our model is yielding concentrations in the expected range of activity. We believe that the negative CI response is due to production of NO, which loosens the cells from the electrode-containing surface, although further studies are still needed to validate this hypothesis.

128

Figure 5-4: (A) HUVEC expression of the Mas receptor, shown by Western blot. (B) Administration of vehicle and three concentrations of 39 to HUVEC cells, measured by RTCA. The 10 pM dose shows 3-fold CI response compared with the 1 pM dose. (C) Graphic comparison of CI response of 1 pM and 10 pM administrations of 39 after 10 minute incubation. Courtesy Dr. John Konhilas and Dr. Meredith Hay.

5.2.3. CID and BBB penetration experiments

CID data was acquired by Catherine Kramer in the laboratory of Dr. Michael

Heien as previously described (4.2.2. Collision-Induced Decay). We tested Ang-(1-7) and two analogs (compounds 34, 35, and 39, respectively). The results from the fragmentation experiments follow.

129

130

131

Both compounds 34 and 39 have a single, major fragment that arises, meaning that much of the signal is retained after fragmentation. In turn, this should allow sensitive measurements of the peptides in serum, even at low concentrations.

To test BBB penetration, glycopeptide 39 was administered sub cutaneously (s.c.) at 1.0 mg/Kg (n = 4) and 10.0 mg/Kg (n = 4) to rats [for 7 days (1mg/kg) to chronic mice.

132 after 8th injection, sacrificed] in the laboratory of Dr. Todd Vanderah by Dr. Tally

Largent-Milnes. After 15 minutes, the rats were sacrificed, and both whole blood and cerebrospinal fluid (CSF) were removed from each animal, combined, stored at -80 ⁰C, and transferred to the lab of Dr. Michael Heien. For analysis, the samples were thawed, desalted using a ZipTip™, and loaded onto an LC-MS.

Based on our results, we notice that at both concentrations 39 is present in at least

10-fold greater concentrations in the CSF than in serum. This reveals two important characteristics of the glycopeptide: first, that 39 crosses the BBB in measurable

133 quantities; and second, that penetration of the BBB reduces the rate of clearance of the peptide (either via enzymatic hydrolysis, or by the liver) from the rat body. Taken together, these suggest that glycosylation of 39 is a modification that provides drug-like properties.

One inconsistency that may be present in the reported concentration values is that the calculations were performed using on-column measurements of concentration versus the internal standard. At this point we have not determined the percent recovery of sample from whole blood or CSF, but anticipate these values to be well under 100%, meaning we believe that the glycopeptide is present in higher concentrations than reported in both media. Additionally, it was observed that small quantities of blood were taken up with the CSF, thereby diluting the actual concentration of glycopeptide found in the CSF.

5.2.4. Memory and object recognition experiments

Finally, glycopeptide 39 was tested in the laboratories of Dr. John Konhilas and

Dr. Meredith Hay to determine if it can be used as a treatment for cognitive impairment following myocardial infarction (MI). To test this, a mouse model of cognitive impairment through MI-induced congestive heart failure (CHF) was used. The mice were separated into three groups, sham mice dosed with saline (n = 12), and those with induced myocardial infarction dosed with either saline (n = 6) or glycopeptide 39 (n =

11). Dosages began 8 weeks after MI surgery, administered daily. After 21 days of dosing, a Novel Object Recognition (NOR) test was performed (Figure 5-9A), and after

25 total days of dosing the mice were introduced to the Morris Water Maze (MWM)

(Figure 5-9B) (for a detailed explanation of each experiment, see Schroder, et al.,

134

2003210). These experiments measure, respectively, a non-spatial and spatial memory task and are standard for determining memory loss associated with damage to the hippocampus and perirhinal cortex.

The results of the CHF mice treated with 39 are broadly positive. First, the results of the NOR non-spatial task for CHF mice dosed with 39 were very similar to those of the sham mice, while CHF mice that were not given the compound showed a clear decrease in cognitive function. Second, after three days of training in the maze, CHF mice with 39 administration performed significantly better than the untreated CHF mice, and nearly as well as sham mice. Because administration of glycpeptide 39 was started 8 weeks after induced heart failure, loss of cognition and memory arising from CHF should have already occurred. Therefore, these experiments show that glycopeptide 39 can reverse the effects of cognitive impairment as measured by spatial and non-spatial memory tasks.

135

In addition, surgical removal of mice hearts from both sham and induced heart failure mice was performed (Figure 5-10). Generally, the results from this experiment show that glycopeptide 39 does not affect heart function, and therefore the entire benefit derived from CHF mice by its administration can be attributed to interactions in the brain.

Cancer on its own is rarely painful, but upon bone metastasis can become debilitatingly painful. This effect, known as cancer-induced bone pain (CIBP) is not well understood, but can be expressed in mouse models.211,212 Dr. Todd Vanderah and Dr.

Tally Largent-Milnes recognized that Mas receptors are overexpressed in some breast cancers, including those in murine cell line 66.1.213 Therefore, they tested whether Ang-

(1-7) had the ability to attenuate CIPB. Initial work (Figure 5-11) indicated that Ang-(1-

7) produced significant and prolonged antinociceptive effects as measured by changes in murine flinching and guarding after i.p. dosing. Administration of Ang-(1-7) following administration of selective Mas receptor antagonist A779208 eliminated this activity, confirming that antinociception was due to activation of the Mas receptor. Interestingly,

136

Ang-(1-7) also provides benefits to mechanical allodynia (measured using a Von Frey filament test).

Glycopeptide 39 was also tested for antinociception using the same assays (Figure

5-11). Similary to Ang-(1-7), 39 i.p. administration led to a maximum antinociceptive effect within 15 minutes, although unlike Ang-(1-7) the effect gradually decreased, fully disappearing after 2 hours. While the antinociceptive effects of 39 do not appear as persistent as Ang-(1-7) in the CIBP model, the glycopeptide still has the ability to fully attenuate both spontaneous and evoked bone pain after a single injection in a mouse model. At this point, the reduced effectiveness of 39 over time in this assay when compared with the sustained effects of native Ang-(1-7) are not understood.

137

Figure 5-11: The effect of Ang-(1-7) and 39 (AT5/ATn5) on cancer-induced bone pain based on measurements of (A) flinching and (B) guarding. (C) shows ability of peptides to treat mechanical allodynia. In general, Ang-(1-7) appears to have a longer duration of effect than 39. 5.4. The pseudo-proline motif

The apparent activity of our lead compound, 39, when compared with the native peptide was surprising (especially when considering the similar activities of 37 versus

34). In the case of the secretin peptides, as well as with opioid peptides tested in the group, we observe similar binding and activity for most glycosylated analogs as with their non-glycosylated parent peptides. Therefore, we suggest that the glycosidic linkage oxygen forms an internal hydrogen bond with the amide proton, leading to a 5-member

138 ring that forms a “pseudoproline” (Figure 5-12). While a 6-member ring to the C-terminal amide could also form, we believe that this would negatively affect receptor interactions, which our data do not suggest.

Pseudoprolines have been reported before, although typically these structures are composed of stable, covalent bonds.214-217 In 1987 Fasman and coworkers reported that in a cyclic peptide, NMR data shows the presence of a pseudoproline in which the ether- protected hydroxyl of a serine residue forms a hydrogen bond with the proton of its backbone amide, forming a β-turn motif.218,219 However, in 2001 Polt and coworkers observed that an internal serine glycoside (Ser(O-α-mannose)) maintained an α-helical configuration.220 Therefore, they hypothesized that the presence of a glycoside on a serine residue may reduce the propensity of that residue to adopt the previously described hydrogen bond-stabilized β-turn, thereby inhibiting its pseudoproline behavior.

We justify this discrepancy between the earlier observations from the Polt group and the currently proposed pseudoproline motif for two reasons. First, the Ser-O-α-Man residue was the tenth amino acid in a 17-residue, helical peptide. We would expect the

139 stability afforded by an α-helix to greatly outweigh that from a transient hydrogen bond and the steric disruption that an internal covalently-linked glycoside would provide. In contrast, glycopeptide 39 has a terminal glycoside, so hydrogen bond formation is less likely to induce unfavorable interactions between the carbohydrate and peptide backbone.

Second, the previous work used an α-linkage between the carbohydrate and peptide, while our recent work conjugated the sugar using a β-linkage. Based on modeling (Figure

5-13), it appears that β-glycosides have free rotation around the glycosidic bond while α- glycosides have significant steric barriers. Depending on the conformation of the glycopeptide at a membrane interface, a relatively fixed pseudoproline may also help the carbohydrate adopt a configuration that minimizes membrane interactions (Figure 5-14) while still allowing for helicity in the peptide backbone. This is illustrated using a model derived from NMR studies of modified Ang-(1-7) in the membrane-mimicking

221 environment d25-SDS and prepared using MOE® software and the AMBER-99 forcefield (Figure 5-15). We note that the glycopeptide is highly amphiphilic, with all of the lipophilic sidechains on one side of the helix, and the hydrophilic sidechains on the opposite face. In addition, the carbohydrate appears to be held far from the surface of the membrane, which may be due to the rigidity of the pseudoproline motif.

140

Figure 5-13: Proposed conformation of both an α- and β-serine glycoside pseudoproline (internal hydrogen bond is dashed). Note that the β carbohydrate has free, unhindered rotation around the indicated bond, while rotation of the α-glycoside is constrained by the peptide.

141

142

5.4. Materials and methods

5.4.1. Peptide synthesis

Unless otherwise noted, all solvents were obtained from EMD Chemicals

(Gibbstown, NJ), and used without further purification. Fmoc-protected amino acid

Fmoc-Tyr(tBu)-OH and condensing agent N,N’-diisopropylcarbodiimide (DIC) were acquired from Advanced ChemTech (Louisville, KY, USA). Fmoc-Asp(tBu)-OH was acquired from Oxchem Corporation (Los Angeles, CA, USA). Fmoc-Ile-OH, Fmoc-Pro-

143

OH, and Fmoc-Val-OH were acquired from Chem-Impex (Wood Dale, IL, USA). Fmoc-

His(Trt)-OH, Fmoc-Arg(Pbf)-OH, and Hydroxybenzotriazole hydrate (HOBt•H2O) were obtained from AAPPTec (Louisville, KY, USA). 1,8-diazabicyclo[5.4.0]undec-7-ene

(DBU) from TCI America (Portland, OR, USA), piperidine from Sigma-Aldrich (St.

Louis, MO, USA), piperizine from Alfa Aesar (Ward Hill, MA, USA), O-(7-

(Center Valley, PA, USA). The Fmoc-protected serine glycosides Fmoc--OGlc(OAc)4-

Ser-OH and Fmoc--OLac(OAc)7-Ser-OH were prepared in the Polt lab using previously published conditions.86

Angiotensin 1-7 and analogs assembly was accomplished using a combination of manual and automated peptide methods. A general procedure follows: The C-terminal amino acids were loaded onto Fmoc-Rink resin (Advanced ChemTech, Louisville, KY,

USA) at 0.83 mmol/g resin loading in 25 mL fritted syringes. Initially, the resin was swelled using DMF (~5 mL solvent per gram resin), agitating at RT for two minutes (x2).

A solution of 2% DBU and 3% piperidine in DMF (v:v) was introduced and agitated for

5 minutes, refreshed, and agitated for an additional 10 minutes. The resin was then washed with DMF (x5), and NMP. In a separate vial, Fmoc-β-OGlc(OAc)4-Ser-OH (0.60 mmol, 1.2 eq) was dissolved in 5 mL NMP, and HOBt•H2O (0.65 mmol, 1.3 eq) was added and allowed to mix for 5 minutes. Condensing agent DIC (1.30 mmol, 2.6 eq) was then added, and mixed for 5 minutes. This solution was added to the resin and agitated for 10 minutes. Next, the syringe was placed in a microwave (Emerson 900W Microwave

144

– MW9338SB) set to power level 1 and irradiated for 10 minutes, stopping to shake the syringe every 90 seconds. The syringe was then agitated at RT for an additional 30 minutes. The resin was washed with NMP (x1), DMF (x5), and CH2Cl2 (x5), and dried in vacuo overnight.

Peptides and glycopeptides were assembled on a Protein Technologies, Inc.

(Tucson, AZ, USA) Prelude® Peptide Synthesizer using the reaction scheme that follows. Rink resin (1 g) was placed into the fritted reaction vessels (RVs). Amino acids were dissolved in DMF at 250 mM concentration, HATU at 375 mM, and TMP at 3M.

The following steps were performed for coupling: DMF Top Wash (10 mL, 2 min mix and drain; x6), Deprotection (2% DBU/3% piperidine in DMF; 10 mL, 4 min mix and drain; 8 min mix and drain), DMF Top Wash (10 ml, 2 min mix and drain; x5), Amino

Acid Coupling Solutions were added manually (AA, HATU, TMP in DMF, 10 mL), then automation resumed for Mix and drain (30 min), DMF Top Wash (10 mL, 2 min mix and drain; x2).

Cleavage of peptides and glycopeptides from the resin was accomplished using a

TFA “cocktail” of F3CCOOH:Et3SiH:H2O:CH2Cl2:Ph-OCH3 (v:v:v:v:v,

9:0.3:0.2:1:0.05), agitating at RT for 2 hours. The resulting solution was expelled into a

15 mL centrifuge tube, evaporated under argon, precipitated in ice-cold Et2O, decanted, and rewashed with Et2O, then dissolved in H2O and lyophilized to afford fluffy, white solid.

Purification of the crude glycopeptides was accomplished by Reversed Phase

HPLC (RP-HPLC) with a preparative RP (C-18) Phenomenex (250 X 22 mm) column using a CH3CN-H2O gradient solvent system containing 0.1% F3CCOOH. Fractions

145 containing pure glycopeptide were combined and lyophilized. Homogeneity of the purified glycopeptides was confirmed by analytical RP-HPLC and high resolution mass spectrometry.

5.4.2. Mas receptor activation

In vitro Mas receptor activation was determined using a nitric oxide (NO) assay in the laboratories of Dr. Meredith Hay and Dr. John Konhilas at the University of Arizona, and was based on previously published methods.222 Commercially available human umbilical vein endothelial cells (HUVEC; ATCC, Manassas, VA, USA) were thawed and plated (25,000-30,000 cells per well) on a 96-well standard E-plate covered in gelatin, in media that contained modified Kaigan’s F-12, spiked with 10% FBS, 0.1 mg/mL heparin,

0.03 mg/mL endothelial cell growth supplement, 100 units/mL penicillin, and 100 g/mL streptomycin at 37 °C in a 5% CO2 atmosphere. Cells were cultured in T75 flasks, and data was collected within eight passages. Experiments were performed by preincubating cell cultures in Krebs solution for 30 minutes, followed by treatment with Ang-(1-7) analogs for an additional 15-30 minutes at 37 °C in a 5% CO2 atmosphere.

Fluorescence measurements were performed using the NO-specific dye DAF-FM diacetate. Cells treated with Ang-(1-7) derivatives were loaded with 5 M DAF-FM diacetate for 30 minutes at 37 °C. After loading, the cells were washed with media, then fluorescence intensity was detected using the BioTek Synergy 2 microplate reater

(BioTek Instruments, Inc., Winooski, VT, USA) with excitation at 485 nm and emission at 535 nm.

146

5.4.3. Cell impedence studies

In vitro impedance assays for real-time monitoring of physiological responses of

HUVECs to Ang-(1-7) derivatives were performed in the laboratories of Dr. Meredith

Hay and Dr. John Konhilas at the University of Arizona. Measurements were taken using the xCELLigence system Real-Time Cell Analyzer (RTCA; ACEA Biosciences, Inc.,

San Diego, CA, USA). HUVECs were treated with either media, 1 pM, 10 pM, or 100 pM glycopeptide 39. Relative impedance, expressed as the cell index (CI), was measured every 15 seconds for 36 hours to generate a continuous compound- and concentration- dependent CI readout. CI over time traces represent the average of 6 wells normalized to

CI at the time point when compounds was added.

5.4.4. Congestive heart failure model of memory loss in mice

In vivo congestive heart failure (CHF) models of memory loss and the effects of administration of Ang-(1-7) derivatives were performed in the laboratories of Dr.

Meredith Hay and Dr. John Konhilas at the University of Arizona.

Animal Selection

Animal Groups: A total of 33, male C57Bl/6J adult mice (Harlan, 8-10 weeks old) were used. Mice were randomly assigned to either the sham (n = 12) or CHF group (n =

21). Experimental groups were assigned as follows: sham + saline, CHF + saline, CHF +

39.

Animal Housing: The mice were housed together (2-4 per container) in a temperature-controlled cage rack based on their designated experimental group and maintained on a 12-hr light-dark cycle. Every mouse had access to food and water ad

147 libitum throughout the duration of the experiments. All experiments were performed using protocols that adhered to guidelines and approved by the Institutional Animal Care and Use Committee at the University of Arizona, and to 2012 NIH guidelines for care and use of laboratory animals.

Mouse Model of Myocardial Infarction (MI)

All mice prior to surgery were weighed and anesthetized. For the CHF mice, MI was induced by ligation of the left coronary artery (LCA) as was previously described.223

Under anesthesia (2.5% isoflurane in a mixture of air and O2) a thoracotomy was performed at the fourth left intercostal space and the LCA permanently ligated to induce a myocardial infarction (MI). Occlusion of the LCA was confirmed by observing blanching, a slight change in color of the anterior wall of the left ventricle downstream of the ligature. Sham mice underwent the same procedure with the exception of ligating the

LCA.

Implantation with angiotensin-(1-7) derivative 39

Mice were anesthetized with isoflurane and an osmotic pump (Alzet, model 1004,

Cupertino, CA, USA) was implanted in the right flank for subcutaneous infusions of 39

(Ang-(1-6)-S(OGlc)) or saline. Administration began 8 weeks post-MI surgery, when

CHF mice were treated with either daily injections of 39 (1 mg/kg/day) or saline.

148

Histology

Hearts were rapidly excised following euthanasia, washed, and placed in 10% neutral buffered formalin for 24 hours for fixation prior to histological evaluation. The fixed hearts were processed, embedded in paraffin, sectioned, and stained with hemotoxylin and eosin (H&E).

Novel Object Recognition (NOR)

Apparatus: The apparatus consisted of an evenly illuminated Plexiglas box (12 cm × 12 cm × 12 cm) placed on a table inside an isolated observation room. All walls of the apparatus were covered in black plastic, and the floor was grey with a grid that was used to ensure that the location of objects did not change between object familiarization and test phases. The mouse behavior and exploration of objects was recorded with a digital camera. The digital image from the camera was fed into a computer in the adjacent room. Two digital stopwatches were used to track the time the mouse spent interacting with the objects of the test. All data was downloaded to Excel files for analysis. Triplicate sets of distinctly different objects were used for the test.

Procedure: The novel object recognition task included 3 phases: habituation phase, familiarization phase, and test phase. For the habituation phase, on the first and second day, mice were brought to the observation room habituated to the empty box for

10 min per day. On the third day, each mouse had a “familiarization” trial with 2 identical objects followed by a predetermined delay period and then a “test” trial in which one object was identical to the one in the familiarization phase, and the other was novel. All stimuli were available in triplicate copies of each other so that no object needed to be

149 presented twice. Objects were made of glass, plastic or wood that varied in shape, color, and size. Therefore, different sets of objects were texturally and visually unique. Each mouse was placed into the box the same way for each phase, facing the center of the wall opposite to the objects. To preclude the existence of olfactory cues, the entire box and objects were always thoroughly cleaned with 70% ethanol after each trial and between mice. During the familiarization phase, mice were allowed to explore the 2 identical objects for 4 min and then returned to their home cages. After a 2 hour delay, the “test phase” commences. The mice were placed back to the same box, where one of the two identical objects presented in the familiarization phase was switched to a novel one and the mouse was allowed to explore these objects for another 4 min. Mouse “exploratory behavior” was defined as the animal directing its nose toward the object at a distance of

~2 cm or less.224 Any other behavior, such as resting against the object, or rearing on the object was not considered to be exploration. Exploration was scored by an observer blind to the mouse’s surgical group (CHF vs. sham). Finally, the positions of the objects in the test phases, and the objects used as novel or familiar, were counterbalanced between the 2 groups of mice.

Analysis: Discrimination ratios were calculated from the time spent exploring the novel object minus time spent exploring the familiar object during the test phase divided by the total exploration time. DRatio = (t novel – t familiar) / (t novel + t familiar). Data were analyzed from first 2 minutes of ‘test phase’. A positive score indicates more time spent with the novel object, a negative score indicates more time spent with the familiar object, and a zero score indicates a null preference.

150

All NOR data was examined using one-way analysis of variance, between subjects (ANOVA). Individual group differences were tested using the post hoc Tukey

HSD test. In comparisons between groups of different sample sizes, equal variance was tested using a modified Levene’s test. All statistical tests and p-values were calculated using MS Excel with Daniel’s XLtoolbox and alpha was set at the 0.05 level. Error bars represent SEM.

Morris Water Task: Testing Spatial Learning and Memory/Visual Test

Apparatus: The apparatus used was a large circular pool approximately 1.5 meters in diameter, containing water at 25 ̊C made opaque with addition of non-toxic white

Crayola paint. An escape platform was hidden just below the surface of the water. Visual, high contrast cues were placed on the walls of the test room. A digital camera connected to a computer in the adjacent room is suspended over the tank to record task progress.

Procedure: During the spatial version of the Morris water task, all animals were given 6 training trials per day over 4 consecutive days. During these trials, an escape platform was hidden below the surface of water. Mice were released from seven different start locations around the perimeter of the tank, and each animal performed two successive trials before the next mouse was tested.

The order of the release locations was pseudo-randomized for each mouse such that no mouse was released from the same location on two consecutive trials.

Immediately following the 24 spatial trials, the mice performed a probe trial in which the platform was removed and a mouse swam in the pool for 60 sec. Following the probe trial the animals were screened for visual ability where the escape platform was raised

151 above the surface of the water but the position of the platform changed between each trial. Performance on the swim task was analyzed with a commercial software application

(ANY-maze, Wood Dale, IL). Because different release locations and differences in swimming velocity produce variability in the latency to reach the escape platform, a corrected integrated path length (CIPL) was calculated to ensure comparability of mice performance across different release locations. The CIPL value measures the cumulative distance over time from the escape platform corrected by an animal’s swimming velocity, and is equivalent to the cumulative search error described by Gallagher and colleagues.225

Therefore, regardless of the release location, if the mouse mostly swims towards the escape platform the CIPL value will be low. In contrast, the more time a mouse spends swimming in directions away from the platform, the higher the CIPL value.

Analysis: The primary measures of spatial learning were path length and corrected integrated path length (CIPL). Analysis of Morris water task data was examined using a

Welch-T test. All statistical tests and p-values were calculated using MS Excel with

Daniel’s XLtoolbox and alpha was set at the 0.05 level. Error bars represent SEM.

Experimental Timeline

First, CHF was induced as described above. After 8 weeks, CHF mice received daily doses of either 39 (1 mg/kg/day) or saline. Following 21 days of treatment, animals were tested for object recognition using NOR testing, and compared with sham mice 4 weeks post-MI. The week after NOR testing, and approximately 25 days after beginning treatment, animals were tested for spatial memory using the Morris water maze task.

152

5.4.5. Blood-brain barrier penetration

Analysis of the persistence of glycosylated Ang-(1-7) derivative 39 and its ability to cross the BBB was performed in the laboratories of Dr. John Konhilas, Dr. Todd

Vanderah, and Dr. Michael Heien at the University of Arizona by Dr. John Konhilas,

Eleni Constantopoulos, Dr. Todd Vanderah, Dr. Tally M. Largent-Milnes, Catherine

Kramer, and Dr. Michael Heien.

10-week old male c57bl/6 mice were used. The left coronary artery was permanently ligated using an 8-0 non-absorbable polypropylene suture.

Echocardiography was used to confirm artery ligation. On days 1-7, animals were injected with glycopeptide 39 daily, subcutaneously at 1 mg/kg in 50 µL. On day 8, animals were injected with either 1 mg/kg or 10 mg/kg glycopeptide 39 (50 µL s.c.). 15 minutes after injection and under isoflurane anesthesia, the atlanto-occipital membrane was exposed, punctured with a 25 µL Hamilton syringe, and the CSF was pulled. Mice then underwent a terminal cardiac puncture (25 g needle) to harvest whole blood (200 –

500 µL). Samples were immediately preserved by adding acetic acid to 10% of total solution volume, and immediately stored in a -80 °C freezer until analysis.

When removed for analysis, 10 µL samples were removed from both the CSF and blood, and desalted using a C18 ZipTip®. For desalting, the ZipTip® was equilibrated with three 10 µL volumes of CH3CN, followed by three 10 µL volumes of aqueous 0.1% trifluoroacetic acid (TFA). A 10 µL volume of sample was drawn up and dispensed 10 times to load peptides onto the C18 resin. The sample was then washed with two 10 µL volumes of 0.1% TFA and eluted in 10 µL of 60% CH3CN, 40% H2O with 0.1% TFA.

Once eluted from the ZipTip®, solutions were diluted to 100 µL in 50:50 CH3CN:H2O

153 with 0.1% formic acid. Tandem mass spectrometry analysis (MSn) was performed to yield specific, quantitative signals proportional to the remaining peptide concentration in each sample. MSn analysis was performed using the following instrumentation.

Electrospray ionization of the samples was achieved using a Thermo Finnigan LCQ

(ThermoFisher Scientific, Waltham, MA, USA). Mass analysis was carried out in the LIT with radial ejection of ions for sensitive detection. Tandem MS was conducted with two isolation-and-fragmentation steps (MS3) for glycopeptide 39, with specific analyte ions identified using unique peptide data previously in this chapter. Quantification did not adjust for % recovery of peptide from whole blood or serum, and is an on-column concentration.

5.4.6. Cancer-induced bone pain model

The effects of treatment using Ang-(1-7) analog 39 on peripheral pain using a cancer- induced bone pain model was performed in the laboratory of Dr. Todd Vanderah at the

University of Arizona by Dr. Tally M. Largent-Milnes, Brittany L. Forte, and Lauren M.

Slosky.

Methods

Animals: All procedures were approved by the University of Arizona Animal

Care and Use Committee and conform to the Guidelines by the National Institutes of

Health and the International Association for the Study of Pain. Female BALB/cfC3H mice (Harlan, IN, USA) were 15 to 20 g prior to initiation of study (n = 5 animals per treatment group). Mice were maintained in a climate-controlled room on a 12-hour light/dark cycle and allowed food and water ad libitum. Animals were weighed on days 0

(day of surgery), 7, and 14. Clinical signs of morbidity were monitored and mice not

154 meeting inclusion parameters (e.g. paralysis, rapid weight loss of >20% in 1 week) were removed from the study.

Intramedullary implantation of 66.1 cells (11/4/2014, D0): Mice were anesthetized with ketamine:xylazine (80 mg :12 mg/kg, 10ml/kg injection volume;

Sigma-Aldrich, St. Louis, MO, USA). An arthrotomy was performed. The condyles of the right distal femoris were exposed and a hole was drilled to create a space for injection of 4 × 104 66.1 cells in 5 µL Opti-MEM or 5 µL Opti-MEM without cells in control animals within the intramedullary space of the mouse femoris. Injections were made with an injection cannula affixed via plastic tubing to a 10-µL Hamilton syringe (CI31,

Plastics One, Roanoke, VA, USA). Proper placement of the injector was confirmed through use of Faxitron X-ray imaging. Holes were sealed with bone cement.

Spontaneous pain: Flinching and guarding were observed for duration of 2 minutes during a resting state. Flinching was characterized by the lifting and rapid flexing of the right hind paw when not associated with walking or movement. Flinches were recorded on a five-channel counter. Guarding was characterized by the lifting the right hind limb into a fully retracted position under the torso. Time spent guarding over the duration of 2 minutes was recorded. Each were recorded on D7 at time = 0,15, 30, 60, 90, and 120 minutes after a single dose of drug was administrated to determine the efficacy of AT5 after acute administration as compared to native Ang-(1-7).

Tactile allodynia: The assessment of tactile allodynia consisted of measuring the withdrawal threshold of the paw ipsilateral to the site of tumor inoculation in response to probing with a series of calibrated von Frey filaments using the Chaplan up-down method with the experimenter blinded to treatment groups. The 50% paw withdrawal threshold

155 was determined by the nonparametric method of Dixon. On D7 at times = 0, 30, 60, 90, and 120 minutes after the bolus of vehicle/AT5 (i.p.), tactile thresholds were determined.

Drug treatment: The Mas receptor agonist AT5 diluted from 1mM to 0.8mM in

0.9% saline. On Day 7, mice received an intraperitoneal (i.p.) injection of either saline or

0.8µg/µL (200µL) per mouse for an average total dose of 800µg/kg. Native Ang-(1-7) was dosed at 1 µg/kg.

Statistics

Analysis of variance (ANOVA; post hoc: Neuman–Kuels) in FlashCalc (Dr. Michael H.

Ossipov, University of Arizona, Tucson, AZ, USA). Within group data were analyzed by non-parametric one-way analysis of variance (ANOVA; post hoc: Bonferroni) in FlashCalc

(Dr. Michael H. Ossipov, University of Arizona, Tucson, AZ, USA). Differences were considered to be significant if P≤ 0.05. All data were plotted in GraphPad Prism 6.

156

6. Summary and future directions

6.1. Synthesis of cysteine thioglycosides

When preparing glycopeptides, a number of groups have illustrated the variety of physiological and structural changes to peptides that can be achieved by modifying either the carbohydrate, the amino acid to which it is attached, or the location of the glycosylation along the peptide backbone.1,33,72,226-229 However, the vast majority of glycopeptides contain an O-, C-, or N-glycosylation. To our knowledge, only Kamerling and coworkers has successfully synthesized S-linked glycoside amino acids amenable to

230 incorporation into SPPS. Their process involved using excess Lewis acid (SnCl2), and achieved moderate yields (60-80%). Dr. Lajos Szabo in our lab developed a methodology for catalytic synthesis of S-linked glycosides (or thioglycosides).231 With Dillon

Hanrahan, I expanded the scope of the process to include the amino acid cysteine

(Scheme 6-1). This proceeded in similar or better yields to Kamerling using catalytic

Lewis acid.231 We would like to incorporate these into glycopeptides to determine how S- linked glycosides vary from O-linked in peptide systems. Unpublished work from our lab in collaboration with Dr. Jeanne Pemberton has observed a sizeable difference in solubility of O- and S-linked glycolipids in aqueous and methanolic conditions, and we anticipate that similar variations may be observed by changing the glycosidic linkage.

Therefore, incorporation of cysteine glycosides into a wide range of previously described

O-linked glycopeptides may result in interesting pharmacological variety, and potentially advances in drug development.

157

Scheme 6-1: Synthesis of four cysteine peracetate glycosides using catalytic InBr3. Glc = glucose, Mal = maltose, Lac = lactose, Cel = cellobiose.

158

6.2. SAA incorporation into selective MOR agonist DAMGO

Previous work in the Polt lab has shown that changing peptide amphipathicity by conjugating natural carbohydrates1 and phosphates166 onto the potent and selective MOR agonist DAMGO can be accomplished without disrupting receptor binding, and in some cases led to increased activity in vivo after peripheral administration. We have broadened the scope of modifications to include unnatural sugar-amino acid moieties without introducing negative drug-like properties. SAAs should allow more fine-tuning of amphipathicity beyond naturally occurring carbohydrates, as simple modifications to their synthetic formation (such as the introduction of halogens or stereocenter inversions) will affect their hydrophilicity. However, one drawback, at this point is that none of the modifications to the amphipathicity of DAMGO have been effectively quantified, and instead rely on calculations based on the estimated surface area and conformation of the glyco- or phosphopeptides. Therefore, it would be beneficial to apply known measures of carbohydrate hydrophilicity, such as an experimentally determined measure of g.u. (we have estimated g.u. values for the various DAMGO glycosides) using a dextran ladder, a standard measure for determine relative glycoside size in relation to glucose.232 The degradation in serum of these compounds compared to unmodified DAMGO and analogs with natural carbohydrates could prove instructive as well.

6.3. Glycosylation of the secretin family peptides VIP and PACAP

Development of drugs based on the endogenous secretin family neurotransmitter peptides VIP and PACAP is of interest due to the apparent ability of these neuropeptides to provide neuroprotective effects in models of PD and stroke.14,114,115,138,233

Modifications to VIP and PACAP by carbohydrate addition and substitution of the

159 oxidizable Met17 residue to hydrocarbon-based analogs Leu and Nle were performed.

Preliminary receptor activation studies based on calcium-flux assays show that glycosylated analogs of VIP and PACAP produce similar in vitro agonist activities to their non-glycosylated analogs at the VPAC1 and PAC1R receptors, respectively. As most assays of those receptors only measure binding instead of agonist activity,101 this represents an important advance for the analysis of this family of compounds. We have also observed39 that glycosylated PACAP penetrates the BBB in measurable quantities, while the non-glycosylated analog could not be detected. Taken together, this suggests that addition of a carbohydrate to a peptide provides additional physiological effects, rather than only affecting amphipathicity. Further experiments using radiolabeled or fluorescent peptides to visualize membrane crossing of glycopeptides may help to shed light on this aspect of glycosylation.

Whereas the agonism at VPAC1 and PAC1R performed as expected, changes to our methodology for analyzing antagonism and activity and the VPAC2 receptor are needed. For elucidating activity at the VPAC2, because VPAC2 does not interact with calcium-releasing RAMP proteins,205,206 we might simply not have any measurable change of calcium concentration. Therefore, switching to an assay that uses a different mechanism to quantify activation, such as Förster resonance energy transfer (FRET) or production of cyclic AMP may prove beneficial.113,234 Our current data suggests that if antagonism is occurring at the PAC1R, it is not competitive and reversible. Determining binding constants using radioligands and performing selective amino acid substitutions may help us elucidate what interactions occur between N-terminal truncated PACAP analogs and the PAC1R.

160

Synthesis of VIP and PACAP analogs is still hampered by the conversion of our desired peptide to branched or backbone-disrupted analogs due to aspartimide formation.

A recent paper discusses using a novel protecting group on the aspartic acid carboxylate

(Scheme 6-2) as a new approach to reducing aspartimide formation.235 This modification is very attractive when compared with the strategy of individually synthesizing aspartic acid dimers (Xxx-Asp) because it only requires preparation of a single synthetic target as opposed to the formation of a unique dimer for each Asp in the peptide sequence.

Scheme 6-2: Substitution of classic Asp protecting group tert-butyl to the newly reported 5-n-butyl-5-nonyl group235 is expected to help minimize Asi formation.

Peptide stapling63 as a method to stabilize peptide secondary structure and reduce susceptibility towards enzymatic digestion has been a popular modification in recent years. The formation of stapled analogs of VIP has been reported, and one of the analogs showed promise by improving the in vitro half-life of the peptide without negatively affecting receptor binding.122 To our knowledge, stapling PACAP, or combining glycosylation with stapling in any peptide has not been performed. I have begun synthesis of a PACAP glucoside analog with built-in stapling sites, and once completed and purified, I will determine whether the staple affects PAC1R activation and degradation.

We would also like to test the ability of the secretin family glycopeptides to form micelles. Work has shown that the native peptides can be covalently linked with

161 micelles,124,125 but to our knowledge the spontaneous self-aggregation of glycopeptide amphiphiles has not been measured. If successful, this could be used as a drug delivery vehicle comprised of the drug of interest, which would reduce introduction of exogenous compounds into the body during treatment.

6.4. Glycosylation of Ang-(1-7)

Angiotensin-(1-7) and the Mas receptor have been identified as providing regulation of vasoconstriction, but also affecting the processes that control memory and inflammation.161-164,236 We have shown that glycosylation of Ang-(1-7), and in particular the substitution of Ser(OGlc) for Pro7, improves activation of the Mas receptor in

HUVEC cells, and allows penetration of the BBB and increased in vivo persistence. In induced CHF mice, glycopeptide 39 reverses memory loss post-MI due exclusively to its effects in the brain. Glycopeptide 39 is the most potent of our first generation of Ang-(1-

7) derivatives. Future work will synthesize analogs with different amino acids (such as D-

Ser, hydroxyproline, and Thr) and carbohydrates (Gal, Lac, and Cel) in the 7-position.

Preliminary work also suggests that Ang-(1-7) and derivatives have therapeutic potential towards CIBP. Testing of a range of glycosylated derivatives may allow discovery of a new lead derivative.

6.5. Peptide and glycopeptide detection and degradation analysis by MSn

Collaboration with the Heien lab has yielded the development of an LC-MS-based microdialysis assay capable of detecting multiple peptides and glycopeptides in the sub- nanomolar range concurrently.202 Our work to date has shown that glycosylation improves the persistence of glycosylated PACAP in mouse serum when compared with native PACAP27, that the glycosylated PACAP crosses the BBB while PACAP27 does

162 not in measurable quantities, and that glycosylated Ang-(1-7) derivative 39 both crosses the BBB and can be found in higher concentrations in the NVU than in whole blood 15 minutes after injection into mice. However, some drawbacks still exist for this methodology. First, while the shotgun approach of characterizing multiple glycopeptides decreases necessary sample preparation and instrument time, because the glycosidic bond susceptible to enzymolysis, higher order carbohydrates can be degraded into lower ones

(for instance, lactose into glucose). To test this, we can introduce a selective galactosidase inhibitor237 into serum prior to adding the glycopeptides, and determine if the rate of degradation changes. Additionally, synthetic development of disaccharides linked with an atom other than oxygen, or introduction of a disaccharide such as cellobiose, should decrease enzymatic activity at that bond. We could also isotopically label one of the carbohydrates in the mixture to track degradation of that specific glycopeptide.

Another drawback is the use of the Orbitrap ion trap, which is expensive and therefore not widely distributed in academia and industry. An alternate approach, which is widely used to measure metabolites from living systems and can track and quantify multiple compounds in complex reaction and biological mixtures, is the multiple reaction monitoring assay on a triple-quadrupole MS.238-241 This method offers rapid, sensitive, and selective identification of a wide range of compounds. Therefore, adapting the current peptide identification methodology to a more common system, such as a triple- quadrupole MS, would expand the availability of the assay to a much wider group of research laboratories.

163

7. Appendix A. Calculation of the Connolly surface area of

glycopeptides

The original calculations of Connolly surface areas of the glycopeptides reported by Polt and coworkers for the initial DAMGO glycosides1 used a script within MOE that was provided by the manufacturer specifically for this calculation. In short, a user of

MOE assigned each atom of a molecule as either hydrophilic or lipophilic. In this case, the entire peptide was considered lipophilic, and the entire carbohydrate or reside at the

C-terminus was assigned as hydrophilic. This provided a rough approximation of the potential amphipathic character of each molecule. Importantly, this script was able to determine the surface area of the two assigned portions of each glycopeptide while using

MOE’s molecule-wide water-accessible surface area (ASA) (Figure 7-1).

164

However, the script that was provided was not transferred to subsequent versions of MOE, and was unavailable for the analysis of the DAMGO-SAA compounds. In order to be able to compare the two sets of data points, I set up MOE to assign the total surface area of the assigned lipophilic and hydrophilic separately (Figure 7-2). Unfortunately, this means that some slight overlap in surface area exists in the area between the two regions, which is not viewed in these calculations as containing space-occupying atoms.

In order to validate this approximation, I compared a couple calculated ratios between the hydrophilic and lipophilic surface areas, and between hydrophilic and total surface areas to unpublished results from the paper (Table 7-1).

165

LYM Method EMJ Method

DAMGO- Water/ Water/ Water/ Water/

Xxx/ Lipid Total Lipid Total e-W/T e-W/T Glycopeptide Ratio Ratio Ratio Ratio Lactoside 0.626 0.385 0.680 0.622 0.383 0.681 Glucoside 0.414 0.293 0.746 0.424 0.298 0.742 3 0.458 0.314 0.730

4 0.339 0.253 0.777 5 0.429 0.300 0.741

Encouragingly, the reported amphipathicity values1 (Table 7-1, LYM Method) are within a couple thousandths of those calculated using my method (Table 7-1, EMJ

Method).

166

8. Appendix B. Supplementary Data

Figure 8-1: Full normalized data for PACAP degradation studies. Experiments (x#) are presented in the same color, and each compound in the same shape. PACAP27 (6, ♦), PACAP26-S(OGlc) (7,■), PACAP26-S(OLac) (8,▲).

167

VPAC1 Agonist Data

2 Drug EC50 (nM) Emax R VIP28 55.1 80.0 0.98 PACAP27 44.4 64.8 0.87 Figure 8-2: VIP28 (12) and PACAP27 (6) calculated agonist activity at the VPAC1 receptor, measured by FLIPR. Similar activity is noted, although n = 1 for this experiment. Courtesy Katie Edwards.

168

NMR Data

1H – 7

169

13C – 7

170

1H – 8

171

13C – 8

172

1H – 9

173

13C – 9

174

1H – 10

175

13C – 10

176

1H – 11

177

13C – 11

178

1H – 1

179

13C – 1

180

1H watergate – 3

181

TOCSY – 3

182

COSY – 3

183

HSQC - 3

184

1H watergate - 4

185

TOCSY - 4

186

COSY - 4

187

HSQC - 4

188

9. Appendix C. Copyright Permissions

For: Figure 1-1

Dove Medical Press Ltd. LICENSE TERMS AND CONDITIONS Sep 22, 2015

This is a License Agreement between Evan Jones ("You") and Dove Medical Press Ltd. ("Dove Medical Press Ltd.") provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the terms and conditions provided by Dove Medical Press Ltd., and the payment terms and conditions. License Number 3714310413444 License date Sep 22, 2015 Licensed content publisher Dove Medical Press Ltd. Licensed content title Journal of Pain Research Licensed content date Jan 1, 2008 Type of Use Thesis/Dissertation Requestor type Academic institution Format Electronic Portion chart/graph/table/figure Number of 1 charts/graphs/tables/figures Title or numeric reference of Chapter 1, Figure 1. the portion(s) Title of the article or chapter N/A the portion is from Editor of portion(s) N/A Author of portion(s) N/A Volume of serial or N/A monograph. Page range of the portion 21 Publication date of portion October 2016 Rights for Main product Duration of use Life of current edition Creation of copies for the no disabled With minor editing privileges no For distribution to United States In the following language(s) Original language of publication With incidental promotional no use The lifetime unit quantity of Up to 499 new product Made available in the educational following markets The requesting Evan M Jones person/organization is: Order reference number None Author/Editor Evan M Jones The standard identifier 2015JonesEvanDissertation The proposed price 0.00 Title Design and Synthesis of Neurologically Active Glycopeptides for Neuroprotection and Antinociception Publisher The University of Arizona Expected publication date Oct 2016 Estimated size (pages) 200 Total (may include CCC user 0.00 USD fee)

189

For: Figure 1-2

ELSEVIER LICENSE TERMS AND CONDITIONS Sep 22, 2015

This is a License Agreement between Evan Jones ("You") and Elsevier ("Elsevier") provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the terms and conditions provided by Elsevier, and the payment terms and conditions. Supplier Elsevier Limited The Boulevard,Langford Lane Kidlington,Oxford,OX5 1GB,UK Registered Company 1982084 Number Customer name Evan Jones Customer address 1306 E University Blvd Tucson, AZ 85719 License number 3714320063269 License date Sep 22, 2015 Licensed content publisher Elsevier Licensed content Drug Discovery Today publication Licensed content title Peptide therapeutics: current status and future directions Licensed content author Keld Fosgerau,Torsten Hoffmann Licensed content date January 2015 Licensed content volume 20 number Licensed content issue 1 number Number of pages 7 Start Page 122 End Page 128 Type of Use reuse in a thesis/dissertation Portion figures/tables/illustrations Number of 1 figures/tables/illustrations Format electronic Are you the author of this No Elsevier article? Will you be translating? No Original figure numbers Figure 1. Title of your Design and Synthesis of Neurologically Active thesis/dissertation Glycopeptides for Neuroprotection and Antinociception Expected completion date Oct 2015 Estimated size (number of 200 pages) Elsevier VAT number GB 494 6272 12 Permissions price 0.00 USD VAT/Local Sales Tax 0.00 USD / 0.00 GBP Total 0.00 USD

190

For: Figure 1-9 Re: Figure Reproduction Query

Dear Dr. Jones,

You can use the figure for free.

Best Regards,

Xiu-Xia Song, Vice Director, Editorial Office

Baishideng Publishing Group Inc Telephone: +86-10-8538-1891 Fax: +86-10-8538-1893 E-mail:[email protected] http://www.wjgnet.com

2015-09-25

To whom it may concern, I would like to include a figure from one of your publications in my dissertation. This would be a non-commercial use, and will be fully cited and attributed in the dissertation. Do I need to complete any forms in order to receive permission to reproduce the following figure:

doi:10.4331/wjbc.v2.i6.146 Figure 1 Thank you, Evan Jones

2015-09-23

PERMISSIONS

Copyright ©1995-2014 Baishideng Publishing Group Inc. All rights reserved. Articles published by this open-access journal are distributed under the terms of the Creative Commons Attribution-Noncommercial License, which permits use, distribution, and reproduction in any medium, provided the original work is properly cited, the use is non commercial and is otherwise in compliance with the license.

191

For: Figure 1-10

NATURE PUBLISHING GROUP LICENSE TERMS AND CONDITIONS Sep 22, 2015

This is a License Agreement between Evan Jones ("You") and Nature Publishing Group ("Nature Publishing Group") provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the terms and conditions provided by Nature Publishing Group, and the payment terms and conditions. License Number 3714440558950 License date Sep 22, 2015 Licensed content publisher Nature Publishing Group Licensed content Nature Reviews Cardiology publication Licensed content title Angiotensin-converting enzyme 2 and angiotensin 1-7: novel therapeutic targets Licensed content author Fan Jiang, Jianmin Yang, Yongtao Zhang, Mei Dong, Shuangxi Wang, Qunye Zhang Licensed content date Apr 29, 2014 Volume number 11 Issue number 7 Type of Use reuse in a dissertation / thesis Requestor type academic/educational Format print and electronic Portion figures/tables/illustrations Number of 1 figures/tables/illustrations High-res required no Figures Protective effects of Ang-(1-7) on stroke.(Jiang et al., 2014) Author of this NPG article no Your reference number None Title of your thesis / Design and Synthesis of Neurologically Active Glycopeptides for dissertation Neuroprotection and Antinociception Expected completion date Oct 2015 Estimated size (number of 200 pages) Total 0.00 USD

192

Bibliography

1 Lowery, J. J. et al. Glycosylation Improves the Central Effects of DAMGO. Chemical Biology and Drug Design 69, 41-47 (2007). 2 Tsang, A. et al. Common Chronic Pain Conditions in Developed and Developing Countries: Gender and Age Differences and Comorbidity With Depression- Anxiety Disorders. Journal of Pain 9, 883-891, doi:10.1016/j.jpain.2008.05.005 (2008). 3 Pizzo, P. A. et al. Relieving Pain in America: A Blueprint for Transforming Prevention, Care, Education, and Research. (The National Academies Press, 2011). 4 Rao, P. P. & Mohamed, T. Current and emerging "at-site" pain medications: a review. Journal of pain research 4, 279-286, doi:10.2147/JPR.S11308 (2011). 5 Tracey, I. & Mantyh, P. W. The cerebral signature and its modulation for pain perception. Neuron 55, 377-391, doi:10.1016/j.neuron.2007.07.012 (2007). 6 Rowbotham, M. C. et al. Oral opioid therapy for chronic peripheral and central neuropathic pain. New England Journal of Medicine 348, 1223-1232, doi:10.1056/NEJMoa021420 (2003). 7 Cowan, W. M. & Kandel, E. R. Prospects for neurology and psychiatry. Jama- Journal of the American Medical Association 285, 594-600, doi:10.1001/jama.285.5.594 (2001). 8 Smith, S. C. Reducing the Global Burden of Ischemic Heart Disease and Stroke A Challenge for the Cardiovascular Community and the United Nations. Circulation 124, 278-279, doi:10.1161/CIRCULATIONAHA.111.040170 (2011). 9 Mattson, M. P. Apoptosis in neurodegenerative disorders. Nature Reviews Molecular Cell Biology 1, 120-129, doi:10.1038/35040009 (2000). 10 Bredesen, D. E., Rao, R. V. & Mehlen, P. Cell death in the nervous system. Nature 443, 796-802, doi:10.1038/nature05293 (2006). 11 Casson, R. J. et al. Translational neuroprotection research in glaucoma: a review of definitions and principles. Clinical and Experimental Ophthalmology 40, 350- 357, doi:10.1111/j.1442-9071.2011.02563.x (2012). 12 Andersen, J. K. Oxidative stress in neurodegeneration: cause or consequence? Nature medicine, S18-S25 (2004). 13 Zadori, D. et al. Mitochondrial disturbances, excitotoxicity, neuroinflammation and kynurenines: Novel therapeutic strategies for neurodegenerative disorders. Journal of the neurological sciences 322, 187-191, doi:10.1016/j.jns.2012.06.004 (2012). 14 Seidl, S. E. & Potashkin, J. A. The promise of neuroprotective agents in Parkinson's disease. Frontiers in neurology 2, 68-68, doi:10.3389/fneur.2011.00068 (2011).

193

15 Regenhardt, R. W., Bennion, D. M. & Sumners, C. Cerebroprotective action of angiotensin peptides in stroke. Clinical science 126, 195-205, doi:10.1042/CS20130324 (2014). 16 Hinkle, J. L. & Bowman, L. Neuroprotection for ischemic stroke. The Journal of neuroscience nursing : journal of the American Association of Neuroscience Nurses 35, 114-118 (2003). 17 Kaspar, A. A. & Reichert, J. M. Future directions for peptide therapeutics development. Drug discovery today 18, 807-817, doi:http://dx.doi.org/10.1016/j.drudis.2013.05.011 (2013). 18 Fosgerau, K. & Hoffmann, T. Peptide therapeutics: current status and future directions. Drug discovery today 20, 122-128, doi:http://dx.doi.org/10.1016/j.drudis.2014.10.003 (2015). 19 Lee, V. H. L. & Yamamoto, A. Penetration and enzymatic barriers to peptide and protein absorption. Adv. Drug Delivery Rev. 4, 171-207 (1990). 20 Bocci, V. Catabolism of therapeutic proteins and peptides with implications for drug delivery. Adv. Drug Delivery Rev. 4, 149-169 (1990). 21 Modi, N. B. Pharmacokinetics and pharmacodynamics of recombinant proteins and peptides. J. Controlled Release 29, 269-281 (1994). 22 Egleton, R. D. et al. Improved bioavailability to the brain of glycosylated Met- enkephalin analogs. Brain research 881, 37-46, doi:10.1016/S0006- 8993(00)02794-3 (2000). 23 Lipinski, C. A., Lombardo, F., Dominy, B. W. & Feeney, P. J. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Advanced Drug Delivery Reviews 46, 3-26, doi:10.1016/S0169-409X(00)00129-0 (2001). 24 Banks, W. A. & Kastin, A. J. Peptides and the Blood-Brain Barrier: Lipophilicity as a Predictor of Permeability. Brain Res. Bull. 15, 287-292 (1985). 25 Stevens, R. C. et al. The GPCR Network: a large-scale collaboration to determine human GPCR structure and function. Nature Reviews Drug Discovery 12, 25-34, doi:10.1038/nrd3859 (2013). 26 Pierce, K. L., Premont, R. T. & Lefkowitz, R. J. Seven-transmembrane receptors. Nature Reviews Molecular Cell Biology 3, 639-650, doi:10.1038/nrm908 (2002). 27 Stevens, R. C. et al. Vol. 2013. 28 Lord, J. A. H., Waterfield, A. A., Hughes, J. & Kosterlitz, H. W. Endogenous Opioid Peptides - Multiple Agonists and Receptors. Nature 267, 495-499, doi:10.1038/267495a0 (1977). 29 Dhawan, B. N. et al. International union of pharmacology .12. Classification of opioid receptors. Pharmacological reviews 48, 567-592 (1996). 30 Li, Y. et al. Opioid Glycopeptide Analgesics Derived from Endogenous Enkephalins and Endorphins. Future Science 4, 205-226 (2012). 31 Schwyzer, R. Molecular Mechanism of Opioid Receptor Selection. Biochemistry 25, 6335-6342, doi:10.1021/bi00368a075 (1986). 32 Dhanasekaran, M. et al. Glycopeptides related to beta-endorphin adopt helical amphipathic conformations in the presence of lipid bilayers. Journal of the American Chemical Society 127, 5435-5448, doi:10.1021/ja0432158 (2005).

194

33 Li, Y. et al. Can Amphipathic Helices Influence the CNS Antinociceptive Activity of Glycopeptides Related to beta-Endorphin? Journal of medicinal chemistry 57, 2237-2246, doi:10.1021/jm400879w (2014). 34 Egleton, R. D. et al. Biousian glycopeptides penetrate the blood-brain barrier. Tetrahedron: Asymmetry 16, 65-75 (2005). 35 Huang, W. J. et al. Structural insights into mu-opioid receptor activation. Nature 524, 315-+, doi:10.1038/nature14886 (2015). 36 Huang, W. M., A.; Venkatakrishnan, A.J.; Laeremans, T.; Feinberg, E.N.; Sanborn, A.L.; Kato, H.E.; Livingston, K.E.; Thorsen, T.S.; Kling, R.C.; Granier, S.; Gmeiner, P.; Husbands, S.M.; Traynor, J.R.; Weis, W.I.; Steyaert, J.; Dror, R.O.; Kobilka, B.K. Crystal structure of active mu-opioid receptor bound to the agonist BU72. doi:10.2210/pdb5c1m/pdb (2015). 37 Ballesteros, J. A., Shi, L. & Javitch, J. A. Structural mimicry in G protein-coupled receptors: Implications of the high-resolution structure of rhodopsin for structure- function analysis of rhodopsin-like receptors. Molecular pharmacology 60, 1-19 (2001). 38 Martin, B. et al. Class II G protein-coupled receptors and their ligands in neuronal function and protection. Neuromolecular Medicine 7, 3-36, doi:10.1385/NMM:7:1:3 (2005). 39 Anglin, B. Design and Synthesis of PACAP Based Glycopeptide Analogs; Effects of Glycosylation on Activity and Blood-Brain Barrier Penetration Dissertation/Thesis thesis, The University of Arizona, (2014). 40 Manglik, A. et al. 4DKL: Crystal structure of the mu-opioid receptor bound to a morphinan antagonist. Protein Data Bank, doi:http://dx.doi.org/10.2210/pdb4dkl/pdb (2012). 41 Manglik, A. et al. Crystal structure of the mu-opioid receptor bound to a morphinan antagonist. Nature 485, 321-326, doi:10.1038/nature10954 (2012). 42 Olejniczak, E. T. et al. 2JOD: Pac1-Rshort N-terminal EC domain Pacap(6-38) complex. Worldwide Protein Data Bank, doi:10.2210/pdb2jod/pdb (2007). 43 Sun, C. et al. Solution structure and mutational analysis of pituitary adenylate cyclase-activating polypeptide binding to the extracellular domain of PAC1-Rs. Proceedings of the National Academy of Sciences of the United States of America 104, 7875-7880, doi:10.1073/pnas.0611397104 (2007). 44 Gaudin, P., Couvineau, A., Maoret, J. J., Rouyerfessard, C. & Laburthe, M. Mutational Analysis of Cysteine Residues within the Extracellular Domains of the Human Vasoactive-Intestinal-Peptide (Vip) 1-Receptor Identifies 7 Mutants that are Defective in Vip Binding. Biochemical and biophysical research communications 211, 901-908, doi:10.1006/bbrc.1995.1897 (1995). 45 Tan, Y. V. et al. Peptide agonist docking in the N-terminal ectodomain of a class II G protein-coupled receptor, the VPAC1 receptor - Photoaffinity, NMR, and molecular modeling. Journal of Biological Chemistry 281, 12792-12798, doi:10.1074/jbc.M513305200 (2006). 46 Juarranz, M. G. et al. Vasoactive intestinal polypeptide VPAC(1) and VPAC(2) receptor chimeras identify domains responsible for the specificity of ligand binding and activation. European Journal of Biochemistry 265, 449-456, doi:10.1046/j.1432-1327.1999.00769.x (1999).

195

47 Tan, Y. V., Couvineau, A. & Laburthe, M. Diffuse pharmacophoric domains of vasoactive intestinal peptide (VIP) and further insights into the interaction of VIP with the N-terminal ectodomain of human VPAC1 receptor by photoaffinity labeling with [Bpa(6)]-VIP. Journal of Biological Chemistry 279, 38889-38894, doi:10.1074/jbc.M404460200 (2004). 48 Tan, Y. V., Couvineau, A., Van Rampelbergh, J. & Laburthe, M. Photoaffinity labeling demonstrates physical contact between vasoactive intestinal peptide and the N-terminal ectodomain of the human VPAC1 receptor. Journal of Biological Chemistry 278, 36531-36536, doi:10.1074/jbc.M304770200 (2003). 49 Gaudin, P., Couvineau, A., Rouyer-Fessard, C., Maoret, J. J. & Laburthe, M. The human vasoactive intestinal peptide pituitary adenylate cyclase activating peptide receptor 1 (VPAC1): Constitutive activation by mutations at threonine 343. Biochemical and biophysical research communications 254, 15-20, doi:10.1006/bbrc.1998.9883 (1999). 50 Siu, F. Y. et al. 4L6R: Structure of the class B human glucagon G protein coupled receptor. Worldwide Protein Data Bank, doi:10.2210/pdb4l6r/pdb (2013). 51 Siu, F. Y. et al. Structure of the human glucagon class B G-protein-coupled receptor. Nature 499, 444-+, doi:10.1038/nature12393 (2013). 52 Inooka, H. et al. Conformation of a peptide ligand bound to its G-protein coupled receptor. Nature structural biology 8, 161-165, doi:10.1038/84159 (2001). 53 Grace, C. R. R. et al. Structure of the N-terminal domain of a type B1 G protein- coupled receptor in complex with a peptide ligand. Proceedings of the National Academy of Sciences of the United States of America 104, 4858-4863, doi:10.1073/pnas.0700682104 (2007). 54 Gozes, I. et al. Mapping the active site in vasoactive intestinal peptide to a core of four amino acids: Neuroprotective drug design. Proceedings of the National Academy of Sciences of the United States of America 96, 4143-4148, doi:10.1073/pnas.96.7.4143 (1999). 55 Varshavsky, A. The N-end rule: Functions, mysteries, uses. Proceedings of the National Academy of Sciences of the United States of America 93, 12142-12149, doi:10.1073/pnas.93.22.12142 (1996). 56 Orskov, C., Wettergren, A. & Holst, J. J. Biological Effects and Metabolic Rates of Glucagon-Like Peptide-1 7-36 Amide and Glucagon-Like Peptide-1 7-37 in Healthy-Subjects are Indistinguishable. Diabetes 42, 658-661, doi:10.2337/diabetes.42.5.658 (1993). 57 Mentlein, R. Dipeptidyl-peptidase IV (CD26)-role in the inactivation of regulatory peptides. Regulatory peptides 85, 9-24, doi:10.1016/S0167- 0115(99)00089-0 (1999). 58 Taylor, A. Aminopeptidases - Structure and Function. Faseb Journal 7, 290-298 (1993). 59 Werle, M. & Bernkop-Schnurch, A. Strategies to improve plasma half life time of peptide and protein drugs. Amino acids 30, 351-367, doi:10.1007/s00726-005- 0289-3 (2006). 60 Drucker, D. J. & Nauck, M. A. The incretin system: glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase-4 inhibitors in type 2 diabetes. Lancet 368, 1696-1705, doi:10.1016/S0140-6736(06)69705-5 (2006).

196

61 Wu, S., Hopper, I., Skiba, M. & Krum, H. Dipeptidyl Peptidase-4 Inhibitors and Cardiovascular Outcomes: Meta-Analysis of Randomized Clinical Trials with 55,141 Participants. Cardiovascular Therapeutics 32, 147-158, doi:10.1111/1755- 5922.12075 (2014). 62 Roberts, M. J., Bentley, M. D. & Harris, J. M. Chemistry for peptide and protein PEGylation. Adv. Drug Delivery Rev. 64, 116-127 (2012). 63 Schafmeister, C. E., Po, J. & Verdine, G. L. An all-hydrocarbon cross-linking system for enhancing the helicity and metabolic stability of peptides. Journal of the American Chemical Society 122, 5891-5892, doi:10.1021/ja000563a (2000). 64 Jones, E. M. & Polt, R. CNS Active O-Linked Glycopeptides that Penetrate the BBB. Frontiers in Chemistry 3, doi:10.3389/fchem.2015.00040 (2015). 65 Osborne, R., Joel, S., Trew, D. & Slevin, M. Analgesic Activity of Morphine-6- Glucuronide. Lancet 331, 828-828 (1988). 66 Mulder, G. J. Pharmacological Effects of Drug Conjugates - is Morphine 6- Glucuronide an Exception. Trends in pharmacological sciences 13, 302-304, doi:10.1016/0165-6147(92)90095-N (1992). 67 Ueda, T. et al. Chemoenzymatic Synthesis of Glycosylated Glucagon-like Peptide 1: Effect of Glycosylation on Proteolytic Resistance and in Vivo Blood Glucose- Lowering Activity. Journal of the American Chemical Society 131, 6237-6245, doi:10.1021/ja900261g (2009). 68 Fisher, J. F. et al. PEPTIDE TO GLYCOPEPTIDE - GLYCOSYLATED OLIGOPEPTIDE RENIN INHIBITORS WITH ATTENUATED INVIVO CLEARANCE PROPERTIES. Journal of Medicinal Chemistry 34, 3140-3143, doi:10.1021/jm00114a026 (1991). 69 Negri, L. et al. Dermorphin and deltorphin glycosylated analogues: Synthesis and antinociceptive activity after systemic administration. Journal of medicinal chemistry 42, 400-404, doi:10.1021/jm9810699 (1999). 70 Negri, L., Lattanzi, R., Tabacco, F., Scolaro, B. & Rocchi, R. Glycodermorphins: opioid peptides with potent and prolonged analgesic activity and enhanced blood- brain barrier penetration. British journal of pharmacology 124, 1516-1522, doi:10.1038/sj.bjp.0701971 (1998). 71 Moradi, S. V., Mansfeld, F. M. & Toth, I. Synthesis and in vitro evaluation of glycosyl derivatives of luteinizing hormone-releasing hormone (LHRH). Bioorganic & Medicinal Chemistry 21, 4259-4265, doi:10.1016/j.bmc.2013.04.068 (2013). 72 Dangoor, D. et al. Novel glycosylated VIP analogs: synthesis, biological activity, and metabolic stability. Journal of Peptide Science 14, 321-328 (2008). 73 Filira, F., Biondi, L., Cavaggion, F., Scolaro, B. & Rocchi, R. SYNTHESIS OF O-GLYCOSYLATED TUFTSINS BY UTILIZING THREONINE DERIVATIVES CONTAINING AN UNPROTECTED MONOSACCHARIDE MOIETY. International Journal of Peptide and Protein Research 36, 86-96 (1990). 74 Gobbo, M. et al. Antimicrobial peptides: Synthesis and antibacterial activity of linear and cyclic drosocin and apidaecin 1b analogues. Journal of Medicinal Chemistry 45, 4494-4504, doi:10.1021/jm020861d (2002).

197

75 Gerwig, G. J., Hocking, H. G., Stoecklin, R., Kamerling, J. P. & Boelens, R. Glycosylation of Conotoxins. Marine Drugs 11, 623-642, doi:10.3390/md11030623 (2013). 76 Nomoto, M., Yamada, K., Haga, M. & Hayashi, M. Improvement of intestinal absorption of peptide drugs by glycosylation: Transport of tetrapeptide by the sodium ion-dependent D-glucose transporter. Journal of pharmaceutical sciences 87, 326-332, doi:10.1021/js970269p (1998). 77 Polt, R. et al. Glycopeptide Enkephalin Analogs Produce Analgesia in Mice - Evidence for Penetration of the Blood-Brain-Barrier. Proceedings of the National Academy of Sciences of the United States of America 91, 7114-7118, doi:10.1073/pnas.91.15.7114 (1994). 78 Poduslo, J. F. & Curran, G. L. Glycation increases the permeability of proteins across the blood-nerve and blood-brain barriers. Brain research.Molecular brain research 23, 157-162 (1994). 79 Dhanasekaran, M. & Polt, R. New Prospects for Glycopeptide Based Analgesia: Glycoside-Induced Penetration of the Blood-Brain Barrier. Curr. Drug Des. 2, 59- 73 (2005). 80 Otvos, L., Jr. et al. Development of a pharmacologically improved peptide agonist of the leptin receptor. Biochimica Et Biophysica Acta-Molecular Cell Research 1783, 1745-1754, doi:10.1016/j.bbamcr.2008.05.007 (2008). 81 Mabrouk, O. S., Falk, T., Sherman, S. J., Kennedy, R. T. & Polt, R. CNS penetration of the opioid glycopeptide MMP-2200: A microdialysis study. Neurosci. Lett. 531, 99-103 (2012). 82 Williams, S. A. et al. The effect of glycosylation on the uptake of an enkephalin analogue into the central nervous system. Biology and Physiology of the Blood- Brain Barrier: Transport, Cellular Interactions, and Brain Pathologies 46, 69-77 (1996). 83 Moradi, S. V., Varamini, P. & Toth, I. The Transport and Efflux of Glycosylated Luteinising Hormone-Releasing Hormone Analogues in Caco-2 Cell Model: Contributions of Glucose Transporters and Efflux Systems. Journal of Pharmaceutical Sciences 103, 3217-3224, doi:10.1002/jps.24120 (2014). 84 Maresh, G. A., Kastin, A. J., Brown, T. T., Zadina, J. E. & Banks, W. A. Peptide transport system-1 (PTS-1) for Tyr-MTF-1 and Met-enkephalin differs from the receptors for either. Brain Research 839, 336-340, doi:10.1016/s0006- 8993(99)01755-2 (1999). 85 Herve, F., Ghinea, N. & Scherrmann, J.-M. CNS Delivery Via Adsorptive Transcytosis. Aaps Journal 10, 455-472, doi:10.1208/s12248-008-9055-2 (2008). 86 Lefever, M. L. et al. Glycosylation of α-amino acids by sugar acetate donors with InBr3. Minimally competent Lewis acids. Carbohydr. Res. 351, 121-125 (2012). 87 Said, S. I. Vasoactive Intestinal Peptide Biologic Role in Health and Disease. Trends Endocrin Met 2, 107-112 (1991). 88 Harmar, A. J. et al. Pharmacology and functions of receptors for vasoactive intestinal peptide and pituitary adenylate cyclase-activating polypeptide: IUPHAR Review 1. British journal of pharmacology 166, 4-17, doi:10.1111/j.1476- 5381.2012.01871.x (2012).

198

89 Domschke, S., Domschke, W. & Bloom, S. R. Vasoactive intestinal peptide in man - pharmacokinetics, metabolic and circulatory effects. Gut 19, 1049-1053 (1978). 90 Dorsam, G. P., Benton, K., Failing, J. & Batra, S. Vasoactive intestinal peptide signaling axis in human leukemia. World journal of biological chemistry 2, 146- 160, doi:10.4331/wjbc.v2.i6.146 (2011). 91 Martinez, C. et al. Serum Levels of Vasoactive Intestinal Peptide as a Prognostic Marker in Early Arthritis. Plos One 9, e85248-e85248, doi:10.1371/journal.pone.0085248 (2014). 92 Moody, T. W. & Gozes, I. Vasoactive intestinal peptide receptors: A molecular target in breast and lung cancer. Current pharmaceutical design 13, 1099-1104, doi:10.2174/138161207780619000 (2007). 93 Said, S. I. & Faloona, G. R. Elevated plasma and tissue levels of vasoactive intestinal polypeptide in the watery-diarrhea syndrome due to pancreatic, bronchogenic and other tumors. N. Engl. J. Med. 293, 155-160 (1975). 94 Krejs, G. J. Vipoma syndrome. Am. J. Med. 82, 37-48 (1987). 95 Bloom, S. R., Polak, J. M. & Pearse, A. G. E. Vasoactive intestinal peptide and watery-diarrhoea syndrome. Lancet 302, 14-16 (1973). 96 Nussdorfer, G. G. & Malendowica, L. K. Role of VIP, PACAP, and Related Peptides in the Regulation of the Hypothalamo-Pituitary-Adrenal Axis. Peptides 19, 1443-1467 (1998). 97 Nussdorfer, G. G. Paracrine Control of Adrenal Corticol Function by Medullary Chromaffin Cells. Pharmacol. Rev. 48, 495-530 (1996). 98 Smalley, S. G. R., Barrow, P. A. & Foster, N. Immunomodulation of innate immune responses by vasoactive intestinal peptide (VIP): its therapeutic potential in inflammatory disease. Clinical and experimental immunology 157, 225-234, doi:10.1111/j.1365-2249.2009.03956.x (2009). 99 Laburthe, M., Couvineau, A. & Marie, J. C. VPAC receptors for VIP and PACAP. Receptors & channels 8, 137-153, doi:10.1080/10606820290005191 (2002). 100 Nicole, P. et al. Identification of key residues for interaction of vasoactive intestinal peptide with human VPAC(1) and VPAC(2) receptors and development of a highly selective VPAC(1) receptor agonist - Alanine scanning and molecular modeling of the peptide. Journal of Biological Chemistry 275, 24003-24012, doi:10.1074/jbc.M002325200 (2000). 101 Doan, N.-D. et al. Design and in vitro characterization of PAC1/VPAC1-selective agonists with potent neuroprotective effects. Biochemical pharmacology 81, 552- 561, doi:10.1016/j.bcp.2010.11.015 (2011). 102 Couvineau, A. et al. STRUCTURAL REQUIREMENTS FOR VIP INTERACTION WITH SPECIFIC RECEPTORS IN HUMAN AND RAT INTESTINAL MEMBRANES - EFFECT OF 9 PARTIAL SEQUENCES. Biochemical and Biophysical Research Communications 121, 493-498, doi:10.1016/0006-291x(84)90209-2 (1984). 103 Turner, J. T., Jones, S. B. & Bylund, D. B. A FRAGMENT OF VASOACTIVE- INTESTINAL-PEPTIDE, VIP(10-28), IS AN ANTAGONIST OF VIP IN THE

199

COLON-CARCINOMA CELL-LINE, HT29. Peptides 7, 849-854, doi:10.1016/0196-9781(86)90105-1 (1986). 104 Gozes, I. et al. An Antagonist to Vasoactive-Intestinal-Peptide Affects Cellular Functions in the Central-Nervous-System. Journal of Pharmacology and Experimental Therapeutics 257, 959-966 (1991). 105 Brenneman, D. E. & Eiden, L. E. Vasoactive intestinal peptide and electrical activity influence neuronal survival. Proc. Natl. Ada. Sci. USA 83, 1159-1162 (1986). 106 Rangon, C. M. et al. VPAC(2) receptors mediate vasoactive intestinal peptide- induced neuroprotection against neonatal excitotoxic brain lesions in mice. Journal of Pharmacology and Experimental Therapeutics 314, 745-752, doi:10.1124/jpet.105.086405 (2005). 107 Gozes, I. et al. From vasoactive intestinal peptide (VIP) through activity- dependent neuroprotective protein (ADNP) to NAP - A view of neuroprotection and cell division. Journal of Molecular Neuroscience 20, 315-322, doi:10.1385/jmn:20:3:315 (2003). 108 Gressens, P. et al. Vasoactive intestinal peptide prevents excitotoxic cell death in the murine developing brain. Journal of Clinical Investigation 100, 390-397, doi:10.1172/jci119545 (1997). 109 Waschek, J. A. VIP and PACAP: neuropeptide modulators of CNS inflammation, injury, and repair. British journal of pharmacology 169, 512-523, doi:10.1111/bph.12181 (2013). 110 Delgado, M., Munoz-Elias, E. J., Gomariz, R. P. & Ganea, D. Vasoactive intestinal peptide and pituitary adenylate cyclase-activating polypeptide enhance IL-10 production by murine macrophages: In vitro and in vivo studies. Journal of Immunology 162, 1707-1716 (1999). 111 Delgado, M., Pozo, D. & Ganea, D. The significance of vasoactive intestinal peptide in immunomodulation. Pharmacological Reviews 56, 249-290, doi:10.1124/pr.56.2.7 (2004). 112 Martinez, C. et al. Anti-inflammatory role in septic shock of pituitary adenylate cyclase-activating polypeptide receptor. Proceedings of the National Academy of Sciences of the United States of America 99, 1053-1058, doi:10.1073/pnas.012367999 (2002). 113 Persson, E., Voznesensky, O. S., Huang, Y. F. & Lerner, U. H. Increased expression of interleukin-6 by vasoactive intestinal peptide is associated with regulation of CREB, AP-1 and C/EBP, but not NF-kappa B, in mouse calvarial osteoblasts. Bone 37, 513-529, doi:10.1016/j.bone.2005.04.043 (2005). 114 Shioda, S. & Gozes, I. VIP and PACAP: Novel Approaches to Brain Functions and Neuroprotection. Current pharmaceutical design 17, 961-961 (2011). 115 Delgado, M. & Ganea, D. Neuroprotective effect of vasoactive intestinal peptide (VIP) in a mouse model of Parkinson's disease by blocking microglial activation. Faseb Journal 17, 944-+, doi:10.1096/fj.02-0799fje (2003). 116 White, C. M., Ji, S., Cai, H., Maudsley, S. & Martin, B. Therapeutic potential of vasoactive intestinal peptide and its receptors in neurological disorders. CNS Neurological Disorders - Drug Targets 9, 661-666 (2010).

200

117 Hamill, R. W. et al. Autonomic dysfunction and plasticity in micturition reflexes in human alpha-synuclein mice. Developmental Neurobiology 72, 918-936, doi:10.1002/dneu.20978 (2012). 118 Fujimori, N. et al. Vasoactive intestinal peptide reduces oxidative stress in pancreatic acinar cells through the inhibition of NADPH oxidase. Peptides 32, 2067-2076, doi:10.1016/j.peptides.2011.08.027 (2011). 119 Vacas, E. et al. Antioxidant activity of vasoactive intestinal peptide in HK2 human renal cells. Peptides 38, 275-281, doi:http://dx.doi.org.ezproxy1.library.arizona.edu/10.1016/j.peptides.2012.09.011 (2012). 120 Mentlein, R., Gallwitz, B. & Schmidt, W. E. DIPEPTIDYLPEPTIDASE-IV HYDROLYZES GASTRIC-INHIBITORY POLYPEPTIDE, GLUCAGON-LIKE PEPTIDE-1(7-36)AMIDE, PEPTIDE HISTIDINE METHIONINE AND IS RESPONSIBLE FOR THEIR DEGRADATION IN HUMAN SERUM. European Journal of Biochemistry 214, 829-835, doi:10.1111/j.1432- 1033.1993.tb17986.x (1993). 121 Umetsu, Y. et al. Structural difference of vasoactive intestinal peptide in two distinct membrane-mimicking environments. Biochimica et Biophysica Acta 1814, 724-730 (2011). 122 Giordanetto, F. et al. Stapled Vasoactive Intestinal Peptide (VIP) Derivatives Improve VPAC(2) Agonism and Glucose-Dependent Insulin Secretion. Acs Medicinal Chemistry Letters 4, 1163-1168, doi:10.1021/ml400257h (2013). 123 Onyuksel, H., Bodalia, B., Sethi, V., Dagar, S. & Rubinstein, I. Surface-active properties of vasoactive intestinal peptide. Peptides 21, 419-423 (2000). 124 Vukovic, L. et al. Structure and Dynamics of Highly PEG-ylated Sterically Stabilized Micelles in Aqueous Media. Journal of the American Chemical Society 133, 13481-13488 (2011). 125 Ashok, B., Rubinstein, I., Tsueshita, T. & Onyuksel, H. Effects of peptide molecular mass and PEG chain length on the vasoreactivity of VIP and PACAP1- 38 in pegylated phospholipid micelles. Peptides 25, 1253-1258 (2004). 126 Vaudry, D. et al. Pituitary Adenylate Cyclase-Activating Polypeptide and Its Receptors: 20 Years after the Discovery. Pharmacological reviews 61, 283-357, doi:10.1124/pr.109.001370 (2009). 127 Miyata, A. et al. Isolation of a Novel-38 Residue-Hypothalamic Polypeptide which Stimulates Adenylate-Cyclase in Pituitary-Cells. Biochemical and biophysical research communications 164, 567-574, doi:10.1016/0006- 291X(89)91757-9 (1989). 128 Miyata, A. et al. Isolation of a Neuropeptide Corresponding to the N-Terminal 27 Residues of the Pituitary Adenylate-Cyclase Activating Polypeptide with 38 Residues (Pacap38). Biochemical and biophysical research communications 170, 643-648, doi:10.1016/0006-291X(90)92140-U (1990). 129 Dejda, A. et al. Identification by photoaffinity labeling of the extracellular N- terminal domain of PAC1 receptor as the major binding site for PACAP. Biochimie 93, 669-677, doi:10.1016/j.biochi.2010.12.010 (2011). 130 Gourlet, P. et al. STRUCTURAL REQUIREMENTS FOR THE BINDING OF THE PITUITARY ADENYLATE-CYCLASE-ACTIVATING PEPTIDE TO

201

RECEPTORS AND ADENYLATE-CYCLASE ACTIVATION IN PANCREATIC AND NEURONAL MEMBRANES. European Journal of Biochemistry 195, 535-541, doi:10.1111/j.1432-1033.1991.tb15734.x (1991). 131 Ciccarelli, E. et al. Pharmacological Properties of 2 Recombinant Splice Variants of the Pacap Type-i Receptor, Transfected and Stably Expressed in Cho Cells. European Journal of Pharmacology-Molecular Pharmacology Section 288, 259- 267, doi:10.1016/0922-4106(95)90037-3 (1995). 132 Robberecht, P. et al. STRUCTURAL REQUIREMENTS FOR THE OCCUPANCY OF PITUITARY ADENYLATE-CYCLASE-ACTIVATING- PEPTIDE (PACAP) RECEPTORS AND ADENYLATE-CYCLASE ACTIVATION IN HUMAN NEUROBLASTOMA NB-OK-1 CELL- MEMBRANES - DISCOVERY OF PACAP(6-38) AS A POTENT ANTAGONIST. European Journal of Biochemistry 207, 239-246, doi:10.1111/j.1432-1033.1992.tb17043.x (1992). 133 Vandermeers, A. et al. ANTAGONISTIC PROPERTIES ARE SHIFTED BACK TO AGONISTIC PROPERTIES BY FURTHER N-TERMINAL SHORTENING OF PITUITARY ADENYLATE-CYCLASE-ACTIVATING PEPTIDES IN HUMAN NEUROBLASTOMA NB-OK-1 CELL-MEMBRANES. European Journal of Biochemistry 208, 815-819, doi:10.1111/j.1432-1033.1992.tb17252.x (1992). 134 Bourgault, S. et al. Strategies to Convert PACAP from a Hypophysiotropic Neurohormone Into a Neuroprotective Drug. Current pharmaceutical design 17, 1002-1024 (2011). 135 Ghatei, M. A. et al. DISTRIBUTION, MOLECULAR CHARACTERIZATION OF PITUITARY ADENYLATE CYCLASE-ACTIVATING POLYPEPTIDE AND ITS PRECURSOR ENCODING MESSENGER-RNA IN HUMAN AND RAT-TISSUES. Journal of Endocrinology 136, 159-166, doi:10.1677/joe.0.1360159 (1993). 136 Kim, S. J. et al. Parkin cleaves intracellular alpha-synuclein inclusions via the activation of calpain. Journal of Biological Chemistry 278, 41890-41899, doi:10.1074/jbc.M306017200 (2003). 137 Vaudry, D. et al. The neuroprotective effect of pituitary adenylate cyclase- activating polypeptide on cerebellar granule cells is mediated through inhibition of the CED3-related cysteine protease caspase-3/CPP32. Proceedings of the National Academy of Sciences of the United States of America 97, 13390-13395, doi:DOI 10.1073/pnas.97.24.13390 (2000). 138 Wang, G. et al. Neuroprotective effects of PACAP27 in mice model of Parkinson's disease involved in the modulation of K(ATP) subunits and D2 receptors in the striatum. Neuropeptides 42, 267-276, doi:10.1016/j.npep.2008.03.002 (2008). 139 Reglodi, D., Lubics, A., Szalontay, L. & Lengvari, I. Pituitary adenylate cyclase activating polypeptide protects dopaminergic neurons and improves behavioral deficits in a rat model of Parkinson's disease. Behav. Brain Res. 151, 303-312 (2004). 140 Wang, G., Qi, C., Fan, G. H., Zhou, H. Y. & Chen, S. D. PACAP protects neuronal differentiated PC12 cells against the neurotoxicity induced by a

202

mitochondrial complex I inhibitor, rotenone. Febs Letters 579, 4005-4011, doi:10.1016/j.febslet.2005.06.013 (2005). 141 Jansen-Olesen, I. et al. PACAP-38 but not VIP induces release of CGRP from trigeminal nucleus caudalis via a receptor distinct from the PAC(1) receptor. Neuropeptides 48, 53-64, doi:10.1016/j.npep.2014.01.004 (2014). 142 Onoue, S., Hanato, J., Kuriyama, K., Mizumoto, T. & Yamada, S. Development of PACAP38 Analogue with Improved Stability: Physicochemical and In Vitro/In Vivo Pharmacological Characterization. Journal of Molecular Neuroscience 43, 85-93, doi:10.1007/s12031-010-9415-0 (2011). 143 Zucker, I. H. & Zimmerman, M. C. The renin-angiotensin system in 2011: new avenues for translational research. Current Opinion in Pharmacology 11, 101- 104, doi:10.1016/j.coph.2011.02.006 (2011). 144 Santos, R. A. et al. Converting enzyme activity and angiotensin metabolism in the dog brainstem. Hypertension 11, I153-157 (1988). 145 Schiavone, M. T., Santos, R. A. S., Brosnihan, K. B., Khosla, M. C. & Ferrario, C. M. Release of Vasopressin from the Rat Hypothalamo Neurohypophyseal System by Angiotensin-(1-7) Heptapeptide. Proceedings of the National Academy of Sciences of the United States of America 85, 4095-4098, doi:10.1073/pnas.85.11.4095 (1988). 146 Keidar, S., Kaplan, M. & Gamliel-Lazarovich, A. ACE2 of the heart: From angiotensin I to angiotensin (1-7). Cardiovascular research 73, 463-469, doi:10.1016/j.cardiores.2006.09.006 (2007). 147 Jiang, F. et al. Angiotensin-converting enzyme 2 and angiotensin 1-7: novel therapeutic targets. Nature Reviews Cardiology 11, 413-426, doi:10.1038/nrcardio.2014.59 (2014). 148 Yamada, K., Iyer, S. N., Chappell, M. C., Ganten, D. & Ferrario, C. M. Converting enzyme determines plasma clearance of angiotensin-(1-7). Hypertension 32, 496-502 (1998). 149 Petty, W. J. et al. Phase I and Pharmacokinetic Study of Angiotensin-(1-7), an Endogenous Antiangiogenic Hormone. Clinical Cancer Research 15, 7398-7404, doi:10.1158/1078-0432.ccr-09-1957 (2009). 150 Shenoy, V. et al. Oral Delivery of Angiotensin-Converting Enzyme 2 and Angiotensin-(1-7) Bioencapsulated in Plant Cells Attenuates Pulmonary Hypertension. Hypertension 64, 1248-U1202, doi:10.1161/hypertensionaha.114.03871 (2014). 151 Shi, Y. et al. Angiotensin-(1-7) prevents systemic hypertension, attenuates oxidative stress and tubulointerstitial fibrosis, and normalizes renal angiotensin- converting enzyme 2 and Mas receptor expression in diabetic mice. Clinical Science 128, 649-663, doi:10.1042/cs20140329 (2015). 152 Santos, R. A. S. et al. Angiotensin-(1-7) is an endogenous ligand for the G protein-coupled receptor Mas. Proceedings of the National Academy of Sciences of the United States of America 100, 8258-8263, doi:10.1073/pnas.1432869100 (2003). 153 Chappell, M. C., Modrall, J. G., Diz, D. I. & Ferrario, C. M. Novel aspects of the renal renin-angiotensin system: Angiotensin-(1-7),ACE2 and blood pressure regulation. Kidney and Blood Pressure Regulation 143, 77-89 (2004).

203

154 Castor, M. G. M., Santos, R. A. S., Duarte, I. D. G. & Romero, T. R. L. Angiotensin-(1-7) through Mas receptor activation induces peripheral antinociception by interaction with adrenoreceptors. Peptides 69, 80-85, doi:10.1016/j.peptides.2015.04.011 (2015). 155 Vogels, R. L. C., Scheltens, P., Schroeder-Tanka, J. M. & Weinstein, H. C. Cognitive impairment in heart failure: A systematic review of the literature. European Journal of Heart Failure 9, 440-449, doi:10.1016/j.ejheart.2006.11.001 (2007). 156 Roach, G. W. et al. Adverse cerebral outcomes after coronary bypass surgery. New England Journal of Medicine 335, 1857-1863, doi:10.1056/nejm199612193352501 (1996). 157 Moller, J. T. et al. Long-term postoperative cognitive dysfunction in the elderly: ISPOCD1 study. Lancet 351, 857-861, doi:10.1016/s0140-6736(97)07382-0 (1998). 158 Stump, D. A., Rogers, A. T., Hammon, J. W. & Newman, S. P. Cerebral emboli and cognitive outcome after cardiac surgery. Journal of Cardiothoracic and Vascular Anesthesia 10, 113-119, doi:10.1016/s1053-0770(96)80186-8 (1996). 159 Chan, P. H. Reactive oxygen radicals in signaling and damage in the ischemic brain. Journal of Cerebral Blood Flow and Metabolism 21, 2-14, doi:10.1097/00004647-200101000-00002 (2001). 160 Reboldi, G. et al. Comparison between angiotensin-converting enzyme inhibitors and angiotensin receptor blockers on the risk of myocardial infarction, stroke and death: a meta-analysis. Journal of hypertension 26, 1282-1289, doi:10.1097/HJH.0b013e328306ebe2 (2008). 161 Mecca, A. P. et al. Cerebroprotection by angiotensin-(1-7) in endothelin-1- induced ischaemic stroke. Experimental physiology 96, 1084-1096, doi:10.1113/expphysiol.2011.058578 (2011). 162 Regenhardt, R. W. et al. Anti-inflammatory effects of angiotensin-(1-7) in ischemic stroke. Neuropharmacology 71, 154-163, doi:10.1016/j.neuropharm.2013.03.025 (2013). 163 Regenhardt, R. W. et al. Centrally administered angiotensin-(1-7) increases the survival of stroke-prone spontaneously hypertensive rats. Experimental physiology 99, 442-453, doi:10.1113/expphysiol.2013.075242 (2014). 164 Costa, A. C. O. et al. Participation of AT(1) and Mas receptors in the modulation of inflammatory pain. Peptides 61, 17-22, doi:10.1016/j.peptides.2014.08.010 (2014). 165 Zimmerman, M. C. Angiotensin II and angiotensin-1-7 redox signaling in the central nervous system. Current Opinion in Pharmacology 11, 138-143, doi:10.1016/j.coph.2011.01.001 (2011). 166 Yeomans, L. et al. Phosphorylation of Enkephalins: NMR and CD Studies in Aqueous and Membrane-Mimicking Environments. Chemical Biology and Drug Design 78, 749-756 (2011). 167 Yeomans-Maldonado, L. Glycopeptide and Phosphopeptide Analogs of DAMGO: A Study on the Role of Amphipathicity to Promote Blood Brain Barrier Penetration, The University of Arizona, (2009).

204

168 Chakraborty, T. K., Srinivasu, P., Tapadar, S. & Mohan, B. K. Sugar amino acids and related molecules: Some recent developments. J. Chem. Sci. 116, 187-207 (2004). 169 Gruner, S. A. W., Locardi, E., Lohof, E. & Kessler, H. Carbohydrate-Based Mimetics in Drug Design: Sugar Amino Acids and Carbohydrate Scaffolds. Chem. Rev. 102, 491-514 (2002). 170 Risseeuw, M., Overhand, M., Fleet, G. W. J. & Simone, M. I. A compendium of cyclic sugar amino acids and their carbocyclic and heterocyclic nitrogen analogues. Amino acids 45, 613-689, doi:10.1007/s00726-013-1521-1 (2013). 171 Risseeuw, M. D. P., Overhand, M., Fleet, G. W. J. & Simone, M. I. A compendium of sugar amino acids (SAA): scaffolds, peptide- and glyco- mimetics. Tetrahedron: Asymmetry 18, 2001-2010 (2007). 172 Grotenbreg, G. M. et al. A practical synthesis of gramicidin S and sugar amino acid containing analogues. Journal of Organic Chemistry 69, 7851-7859, doi:10.1021/jo0487449 (2004). 173 McDevitt, J. P. & Lansbury, P. T. Glycosamino acids: New building blocks for combinatorial synthesis. Journal of the American Chemical Society 118, 3818- 3828, doi:10.1021/ja9525622 (1996). 174 van Well, R. M. et al. Parallel synthesis of cyclic sugar amino acid/amino acid hybrid molecules. Tetrahedron letters 41, 9331-9335, doi:10.1016/S0040- 4039(00)01693-2 (2000). 175 Schmidt, S., Wilde, H., Hunger, J. & Sicker, D. Synthesis of furanoid sugar amino acids starting from fermentatively produced 2-oxo-D-gluconic acid. Zeitschrift Fur Naturforschung Section B-a Journal of Chemical Sciences 60, 1168-1174 (2005). 176 Kiso, M. & Hasegawa, A. Acetonation of some pentoses with 2,2- dimethoxypropane-N,N-dimethylformamide-p-toluenesulfonic acid. Carbohydr. Res. 52, 95-101 (1976). 177 Krishna, U. M., Deodhar, K. D. & Trivedi, G. K. Asymmetric oxidopyrylium- alkene [5+2] cycloaddition: a divergent approach for the synthesis of enantiopure oxabicyclo[5.4.0]undecanes. Tetrahedron 60, 4829-4836 (2004). 178 Ohrui, H. et al. C-Glycosyl Nucleosides .5. some Unexpected Observations on Relative Stabilities of Compounds Containing Fused 5-Membered Rings with Epimerizable Substituents. Journal of the American Chemical Society 97, 4602- 4613, doi:10.1021/ja00849a023 (1975). 179 van Well, R. M. et al. Synthesis and structural analysis of cyclic oligomers consisting of furanoid and pyranoid epsilon-sugar amino acids. European Journal of Organic Chemistry, 2303-2313, doi:10.1002/ejoc.200300031 (2003). 180 Chen, Y.-B. et al. A water-soluble ribosyl-based fluorescent sensor for Hg2+ and Cu2+ ions. Carbohydrate research 345, 956-959, doi:10.1016/j.carres.2010.02.003 (2010). 181 Wu, A.-T. et al. Synthesis of iminoalditol and N-alkyl iminoalditol derivatives of ribopyranosides. Carbohydrate research 343, 2887-2893, doi:10.1016/j.carres.2008.08.005 (2008).

205

182 Chir, J.-L. et al. Synthesis and biological evaluation of ether bridged bicyclic iminosugar derivatives. Carbohydrate research 344, 1639-1645, doi:10.1016/j.carres.2009.05.029 (2009). 183 Cubero, I. I., Lopez-Espinosa, M. T. P. & Kari, N. Synthesis of Optically-Active Chalcogran from L-Sorbose. Carbohydrate research 261, 231-242, doi:10.1016/0008-6215(94)84020-2 (1994). 184 O'Reilly, C., O'Brien, C. & Murphy, P. V. Synthesis of a constrained polyfunctional bicyclic iminocyclitol scaffold from L-sorbose via a tandem sequence including stereoselective intramolecular Huisgen cycloaddition. Tetrahedron letters 50, 4427-4429, doi:10.1016/j.tetlet.2009.05.060 (2009). 185 An, J.-N., Meng, X.-B., Yao, Y. & Li, Z.-J. A facile synthesis of 1,6- dideoxynojirimycin from L-sorbose. Carbohydrate research 341, 2200-2203, doi:10.1016/j.carres.2006.06.002 (2006). 186 Sorge, R. E. et al. Olfactory exposure to males, including men, causes stress and related analgesia in rodents. Nature Methods 11, 629-+, doi:10.1038/nmeth.2935 (2014). 187 Chesler, E. J., Wilson, S. G., Lariviere, W. R., Rodriguez-Zas, S. L. & Mogil, J. S. Identification and ranking of genetic and laboratory environment factors influencing a behavioral trait, thermal nociception, via computational analysis of a large data archive. Neuroscience and Biobehavioral Reviews 26, 907-923, doi:10.1016/s0149-7634(02)00103-3 (2002). 188 Brot, N., Weissbach, L., Werth, J. & Weissbach, H. ENZYMATIC REDUCTION OF PROTEIN-BOUND METHIONINE SULFOXIDE. Proceedings of the National Academy of Sciences of the United States of America-Biological Sciences 78, 2155-2158, doi:10.1073/pnas.78.4.2155 (1981). 189 Dangoor, D., Rubinraut, S., Fridkin, M. & Gozes, I. Novel analogs of VIP with multiple C-terminal domains. Peptides 28, 1622-1630, doi:10.1016/j.peptides.2007.03.014 (2007). 190 Gourlet, P. et al. Interaction of lipophilic VIP derivatives with recombinant VIP1/PACAP and VIP2/PACAP receptors. European Journal of Pharmacology 354, 105-111, doi:10.1016/s0014-2999(98)00435-x (1998). 191 Kashimoto, K. et al. Structure-activity relationship studies of PACAP-27 and VIP analogues. Annals of the New York Academy of Sciences; VIP, PACAP and related peptides 805, 505-510 (1996). 192 Gozes, I. et al. Neuroprotective strategy for Alzheimer disease: Intranasal administration of a fatty neuropeptide. Proc. Natl. Ada. Sci. USA 93, 427-432 (1996). 193 Subiros-Funosas, R., El-Faham, A. & Albericio, F. Aspartimide formation in peptide chemistry: occurrence, prevention strategies and the role of N- hydroxylamines. Tetrahedron 67, 8595-8606, doi:10.1016/j.tet.2011.08.046 (2011). 194 Haack, T. & Mutter, M. SERINE DERIVED OXAZOLIDINES AS SECONDARY STRUCTURE DISRUPTING, SOLUBILIZING BUILDING- BLOCKS IN PEPTIDE-SYNTHESIS. Tetrahedron Letters 33, 1589-1592, doi:10.1016/s0040-4039(00)91681-2 (1992).

206

195 Martinez, J. & Bodanszky, M. Side Reactions in Peptide-Synthesis .9. Suppression of the Formation of Aminosuccinyl Peptides with Additives. International journal of peptide and protein research 12, 277-283 (1978). 196 Michels, T., Doelling, R., Haberkorn, U. & Mier, W. Acid-Mediated Prevention of Aspartimide Formation in Solid Phase Peptide Synthesis. Organic Letters 14, 5218-5221, doi:10.1021/ol3007925 (2012). 197 Palasek, S. A., Cox, Z. J. & Collins, J. M. Limiting racemization and aspartimide formation in microwave-enhanced Fmoc solid phase peptide synthesis. Journal of Peptide Science 13, 143-148, doi:10.1002/psc.804 (2007). 198 Chafai, M. et al. Pituitary adenylate cyclase-activating polypeptide and vasoactive intestinal polypeptide promote the genesis of calcium currents in differentiating mouse embryonic stem cells. Neuroscience 199, 103-115, doi:http://dx.doi.org/10.1016/j.neuroscience.2011.09.062 (2011). 199 Mabrouk, O. S., Li, Q., Song, P. & Kennedy, R. T. Microdialysis and mass spectrometric monitoring of dopamine and enkephalins in the globus pallidus reveal reciprocal interactions that regulate movement. Journal of Neurochemistry 118, 24-33, doi:10.1111/j.1471-4159.2011.07293.x (2011). 200 Shen, H., Lada, M. W. & Kennedy, R. T. Monitoring of met-enkephalin in vivo with 5-min temporal resolution using microdialysis sampling and capillary liquid chromatography with electrochemical detection. Journal of Chromatography B 704, 43-52, doi:10.1016/s0378-4347(97)00436-2 (1997). 201 Emmett, M. R., Andren, P. E. & Caprioli, R. M. Specific molecular mass detection of endogenously released neuropeptides using in vivo microdialysis mass spectrometry. Journal of Neuroscience Methods 62, 141-147, doi:10.1016/0165-0270(95)00070-4 (1995). 202 Laude, N. D. M., Diana S.; Kramer, Catherine L.; Lemister, E.; Navratilova, E.; Porreca, F.; Heien, Michael L. Quantification of endogenous opioid peptide dynamics in the anterior cingulate cortex by online-preservation microdialysis. Scientific Reports (in revision) (2015). 203 Arunlakshana, O. & Schild, H. O. SOME QUANTITATIVE USES OF DRUG ANTAGONISTS. British Journal of Pharmacology and Chemotherapy 14, 48-58 (1959). 204 Dickson, L. & Finlayson, K. VPAC and PAC receptors: From ligands to function. Pharmacology & therapeutics 121, 294-316, doi:http://dx.doi.org/10.1016/j.pharmthera.2008.11.006 (2009). 205 Christopoulos, A. et al. Novel receptor partners and function of receptor activity- modifying proteins. Journal of Biological Chemistry 278, 3293-3297, doi:10.1074/jbc.C200629200 (2003). 206 Sexton, P. M. et al. Complexing receptor pharmacology - Modulation of family B G protein-coupled receptor function by RAMPs. Vip, Pacap, and Related Peptides: from Gene to Therapy 1070, 90-104, doi:10.1196/annals.1317.076 (2006). 207 Lazareno, S. & Birdsall, N. J. M. ESTIMATION OF COMPETITIVE ANTAGONIST AFFINITY FROM FUNCTIONAL INHIBITION CURVES USING THE GADDUM, SCHILD AND CHENG-PRUSOFF EQUATIONS. British Journal of Pharmacology 109, 1110-1119 (1993).

207

208 Santos, R. A. S. et al. CHARACTERIZATION OF A NEW ANGIOTENSIN ANTAGONIST SELECTIVE FOR ANGIOTENSIN-(1-7) - EVIDENCE THAT THE ACTIONS OF ANGIOTENSIN-(1-7) ARE MEDIATED BY SPECIFIC ANGIOTENSIN RECEPTORS. Brain Research Bulletin 35, 293-298, doi:10.1016/0361-9230(94)90104-x (1994). 209 Figueroa, S., Canadas, S., Arce, C., Oset-Gasque, M. J. & Gonzalez, M. P. SNAP, a NO donor, induces cortical neuron death by a mechanism in which the caspase pathway is implicated. Brain Research 1047, 168-176, doi:10.1016/j.brainres.2005.04.044 (2005). 210 Schroder, N., O'Dell, S. J. & Marshall, J. F. Neurotoxic methamphetamine regimen severely impairs recognition memory in rats. Synapse 49, 89-96, doi:10.1002/syn.10210 (2003). 211 Honore, P. et al. Murine models of inflammatory, neuropathic and cancer pain each generates a unique set of neurochemical changes in the spinal cord and sensory neurons. Neuroscience 98, 585-598, doi:10.1016/s0306-4522(00)00110-x (2000). 212 Slosky, L. M., Largent-Milnes, T. M. & Vanderah, T. W. Use of Animal Models in Understanding Cancer-induced Bone Pain. Cancer growth and metastasis 8, 47-62, doi:10.4137/cgm.s21215 (2015). 213 Passos-Silva, D. G., Brandan, E. & Souza Santos, R. A. Angiotensins as therapeutic targets beyond heart disease. Trends in Pharmacological Sciences 36, 310-320, doi:10.1016/j.tips.2015.03.001 (2015). 214 Keller, M. et al. Pseudoproline-containing analogues of morphiceptin and endomorphin-2: Evidence for a cis Tyr-Pro amide bond in the bioactive conformation. Journal of Medicinal Chemistry 44, 3896-3903, doi:10.1021/jm000332e (2001). 215 Liu, C. F. & Tam, J. P. PEPTIDE SEGMENT LIGATION STRATEGY WITHOUT USE OF PROTECTING GROUPS. Proceedings of the National Academy of Sciences of the United States of America 91, 6584-6588, doi:10.1073/pnas.91.14.6584 (1994). 216 Coin, I., Beyermann, M. & Bienert, M. Solid-phase peptide synthesis: from standard procedures to the synthesis of difficult sequences. Nature Protocols 2, 3247-3256, doi:10.1038/nprot.2007.454 (2007). 217 Hruby, V. J., Yamashiro., D. & Du Vigneaud, V. STRUCTURE OF ACETONE- OXYTOCIN WITH STUDIES ON REACTION OF ACETONE WITH VARIOUS PEPTIDES. Journal of the American Chemical Society 90, 7106-&, doi:10.1021/ja01027a040 (1968). 218 Hollosi, M., Kover, K. E., Holly, S. & Fasman, G. D. BETA-TURNS IN SERINE-CONTAINING LINEAR AND CYCLIC MODELS. Biopolymers 26, 1527-1553, doi:10.1002/bip.360260907 (1987). 219 Hollosi, M., Kover, K. E., Holly, S., Radics, L. & Fasman, G. D. BETA-TURNS IN BRIDGED PROLINE-CONTAINING CYCLIC PEPTIDE MODELS. Biopolymers 26, 1555-1572, doi:10.1002/bip.360260908 (1987). 220 Palian, M. M., Jacobsen, N. E. & Polt, R. O-linked glycopeptides retain helicity in water. Journal of Peptide Research 58, 180-189, doi:10.1034/j.1399- 3011.2001.00906.x (2001).

208

221 Patel, A. B., Srivastava, S. & Phadke, R. S. Interaction of Ile7 angiotensin III with dodecylphosphocholine micelles: NMR and molecular dynamics studies. Magnetic Resonance in Chemistry 37, 279-286, doi:10.1002/(sici)1097- 458x(199904)37:4<279::aid-mrc451>3.0.co;2-6 (1999). 222 Joshi, M. S. et al. Receptor-mediated activation of nitric oxide synthesis by arginine in endothelial cells. Proceedings of the National Academy of Sciences of the United States of America 104, 9982-9987, doi:10.1073/pnas.0506824104 (2007). 223 Gao, X. M., Dart, A. M., Dewar, E., Jennings, G. & Du, X. J. Serial echocardiographic assessment of left ventricular dimensions and function after myocardial infarction in mice. Cardiovascular Research 45, 330-338, doi:10.1016/s0008-6363(99)00274-6 (2000). 224 Ennaceur, A. & Delacour, J. A NEW ONE-TRIAL TEST FOR NEUROBIOLOGICAL STUDIES OF MEMORY IN RATS .1. BEHAVIORAL- DATA. Behavioural Brain Research 31, 47-59, doi:10.1016/0166- 4328(88)90157-x (1988). 225 Gallagher, M., Burwell, R. D. & Burchinal, M. SEVERITY OF SPATIAL- LEARNING IMPAIRMENT IN AGING - DEVELOPMENT OF A LEARNING INDEX FOR PERFORMANCE IN THE MORRIS WATER MAZE. Behavioral Neuroscience 107, 618-626, doi:10.1037/0735-7044.107.4.618 (1993). 226 Andersson, L., Stenhagen, G. & Baltzer, L. The site-selective glycosylation of a designed helix-loop-helix polypeptide motif. Journal of Organic Chemistry 63, 1366-1367, doi:10.1021/jo971999k (1998). 227 Ballet, S. et al. In Vitro Membrane Permeation Studies and in Vivo Antinociception of Glycosylated Dmt(1)-DALDA Analogues. Acs Medicinal Chemistry Letters 5, 352-357, doi:10.1021/ml4004765 (2014). 228 Craig, A. G. et al. Enzymatic glycosylation of contulakin-G, a glycopeptide isolated from Conus venom, with a mammalian ppGalNAc-transferase. Toxicon 39, 809-815, doi:10.1016/S0041-0101(00)00211-7 (2001). 229 Hang, H. C. & Bertozzi, C. R. The chemistry and biology of mucin-type O-linked glycosylation. Bioorganic & Medicinal Chemistry 13, 5021-5034, doi:10.1016/j.bmc.2005.04.085 (2005). 230 Maljaars, C. E. P. et al. Preparation of S- and N-linked glycosylated amino acid building blocks for solid-phase glycopeptide library synthesis. Journal of Carbohydrate Chemistry 24, 353-367, doi:10.1081/car-200066915 (2005). 231 Szabo, L. H., D.J.; Jones, E.M.; Pemberton, J.E.; Polt, R. Preparation of S- Glycoside Surfactants and Cysteine Thioglycosides. Carbohydr. Res. submitted (2015). 232 Guile, G. R., Rudd, P. M., Wing, D. R., Prime, S. B. & Dwek, R. A. A rapid high- resolution high-performance liquid chromatographic method for separating glycan mixtures and analyzing oligosaccharide profiles. Analytical Biochemistry 240, 210-226, doi:10.1006/abio.1996.0351 (1996). 233 Dejda, A., Sokolowska, P. & Nowak, J. Neuroprotective potential of three neuropeptides PACAP, VIP and PHI. Pharmacological Reports 57, 307-320 (2005).

209

234 Onoue, S. et al. Vasoactive intestinal peptide and pituitary adenylate cyclase- activating polypeptide attenuate the cigarette smoke extract-induced apoptotic death of rat alveolar L2 cells. European Journal of Biochemistry 271, 1757-1767, doi:10.1111/j.1432-1033.2004.04086.x (2004). 235 Behrendt, R., Huber, S., Marti, R. & White, P. New t-butyl based aspartate protecting groups preventing aspartimide formation in Fmoc SPPS. Journal of Peptide Science 21, 680-687, doi:10.1002/psc.2790 (2015). 236 Castro-Chaves, P., Cerqueira, R., Pintalhao, M. & Leite-Moreira, A. F. New pathways of the renin-angiotensin system: the role of ACE2 in cardiovascular pathophysiology and therapy. Expert Opinion on Therapeutic Targets 14, 485- 496, doi:10.1517/14728221003709784 (2010). 237 Park, A.-R. & Oh, D.-K. Effects of galactose and glucose on the hydrolysis reaction of a thermostable beta-galactosidase from Caldicellulosiruptor saccharolyticus. Applied Microbiology and Biotechnology 85, 1427-1435, doi:10.1007/s00253-009-2165-7 (2010). 238 Lange, V., Picotti, P., Domon, B. & Aebersold, R. Selected reaction monitoring for quantitative proteomics: a tutorial. Molecular Systems Biology 4, doi:10.1038/msb.2008.61 (2008). 239 Sabatine, M. S. et al. Metabolomic identification of novel biomarkers of myocardial ischemia. Circulation 112, 3868-3875, doi:10.1161/circulationaha.105.569137 (2005). 240 Stevens, J. F., Taylor, A. W. & Deinzer, M. L. Quantitative analysis of xanthohumol and related prenylflavonoids in hops and beer by liquid chromatography tandem mass spectrometry. Journal of Chromatography A 832, 97-107, doi:10.1016/s0021-9673(98)01001-2 (1999). 241 Kitteringham, N. R., Jenkins, R. E., Lane, C. S., Elliott, V. L. & Park, B. K. Multiple reaction monitoring for quantitative biomarker analysis in proteomics and metabolomics. Journal of Chromatography B-Analytical Technologies in the Biomedical and Life Sciences 877, 1229-1239, doi:10.1016/j.jchromb.2008.11.013 (2009).

210