Conus Brunneus

ISOLATION AND CHARACTERIZATION OF NOVEL CONOPEPTIDES

FROM THE MARINE CONE SNAIL: CONUS BRUNNEUS

Fred C. Pflueger

A Dissertation Submitted to the Faculty of

The Charles E. Schmidt College of Science

in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

Florida Atlantic University

Boca Raton, Fl

December 2008

ACKNOWLEDGEMENTS

I would like to thank Dr. Frank Mari for his friendship, time, and for giving me

the opportunity to conduct research under his guidance. I would also like to thank Dr.

Gregg Fields, Dr. Gary Perry, and Dr. Andrew Terentis for their time, help, advice, and

efforts as members of my dissertation committee.

Most importantly I would like to thank Jacqueline, my wife, friend, and good mother to our children; Fred, my father, mentor, and the man who gave me the world; and Claudia, my mother for her unconditional love and support.

This dissertation is dedicated to my children, Alexandria and Frederick, for giving me the strength and determination to finish this work.

iii

ABSTRACT

Author: Fred C. Pflueger

Title: Isolation and Characterization of Novel Conopeptides from the

Marine Cone Snail: Conus brunneus

Institution: Florida Atlantic University

Dissertation Advisor: Dr. Frank Mari

Degree: Doctor of Philosophy

Year: 2008

Cone snails are predatory marine gastropods that use venom for means of

predation and defense. This venom is a complex mixture of conopeptides that selectivity

binds to ion channels and receptors, giving them a wide range of potential pharmaceutical

applications. Conus brunneus is a wide spread Eastern Pacific cone snail species that

preys upon worms (vermivorous). Vermivorous cone snails have developed very specific biochemical strategies for the immobilization of their prey and there venom has not been

extensively studied to date. The main objective of this dissertation is the characterization

of novel conopeptides isolated from Conus brunneus. Chapter 1 is an introduction and

background on cone snails and conopeptides. Chapter 2 details the isolation and

characterization of a novel P-superfamily conotoxin. Chapter 3 presents the 3D solution

structure of the novel P-superfamily conotoxin. Chapter 4 details the isolation and iv characterization of two novel M-superfamily conotoxins. Chapter 5 covers the use of

nano-NMR to characterize a novel P-superfamily conotoxin using nanomole quantities of

sample. Chapter 6 is a reprint of a paper published in the Journal of the American

Chemical Society in which we combined and implemented techniques developed in the

previous chapters to report the presence of D-γ-Hydroxyvaline in a polypeptide chain.

This dissertation contains the first reported work of a P-superfamily structure obtained

directly from the crude venom therefore accurately representing native post-translational

modifications. In this paper, crude cone snail venom was characterized by: high

performance liquid chromatography, nuclear magnetic resonance spectroscopy, nano-

nuclear magnetic resonance spectroscopy, mass spectrometry, amino acid analysis,

Edman degradation sequencing, and preliminary bioassays.

TABLE OF CONTENTS

LIST OF FIGURES………………………………………………………….…………..xi

LIST OF TABLES………………………………………………………………..…...…xv

CHAPTER 1 ...... 1

Background and Introduction ...... 1

Neurotoxins ...... 3

Conopeptides...... 3

Post-translational modifications...... 7

Anatomy of the Cone Snail ...... 7

Anatomy of the Venom Apparatus...... 9

Conus brunneus...... 11

Conus gladiator...... 19

Summary ...... 20

REFERENCES...... 21

CHAPTER 2 ...... 25

Conotoxin bru9a, a Novel P-Superfamily Conopeptide from the Venom of Conus brunneus. Isolation and Characterization...... 25

ABSTRACT...... 25

INTRODUCTION...... 25

MATERIALS AND METHODS...... 27

Specimen Collection...... 27

Venom Duct Removal ...... 28 vi Extraction of Crude Venom...... 29

Purification of Peptides ...... 31

Mass Spectrometry ...... 32

Reduction and Alkylation...... 33

Partial Reduction and Alkylation ...... 33

Promotion and Stabilization of b1 peptide using N-terminal derivatives...... 33

Amino Acid Analysis ...... 34

NMR Spectroscopy ...... 34

Peptide Sequencing...... 34

Bioassays ...... 35

Nomenclature...... 35

RESULTS AND DISCUSSION ...... 36

Purification of Peptides ...... 36

Mass Spectrometry ...... 43

Reduction and Alkylation...... 44

Partial Reduction and Alkylation ...... 48

Promotion and Stabilization of B1 peptide using N-terminal derivatives...... 49

Peptide Amino Acid Analysis ...... 53

Peptide Sequencing...... 54

Bioassays ...... 54

REFERENCES...... 56

CHAPTER 3 ...... 58

Three–Dimensional Solution Structure of Conotoxin bru9a: a Novel P-Superfamily Conotoxin from the Venom of Conus brunneus with an Unusual Cystine Knot Motif. .. 58

ABSTRACT...... 58

vii INTRODUCTION...... 58

MATERIALS AND METHODS...... 61

Preparation of bru9a ...... 61

NMR Sample Preparation...... 61

NMR Spectroscopy ...... 62

Structural Assignments and Restraints...... 62

Structure Calculations...... 64

RESULTS AND DISCUSSION ...... 64

Structural Assignments...... 64

Secondary Structure...... 78

Disulfide Connectivity Determination...... 79

Structural Calculations ...... 81

Stability Test...... 90

REFERENCES...... 93

CHAPTER 4 ...... 97

Isolation and Characterization of Two Novel Mini-M Conotoxins isolated from the Venom of Conus brunneus...... 97

ABSTRACT...... 97

INTRODUCTION...... 97

MATERIALS AND METHODS...... 99

Specimen Collection...... 99

Venom Duct Removal ...... 99

Extraction of Crude Venom...... 99

Purification of Peptides ...... 99

Mass Spectrometry ...... 99

viii Reduction and Alkylation...... 100

Peptide Sequencing...... 100

Nomenclature...... 100

RESULTS AND DISCUSSION ...... 100

Purification of Peptide, bruF070303 ...... 100

Mass Spectrometry ...... 101

Reduction and Alkylation...... 102

Peptide Sequencing...... 103

Purification of Peptide, bruH090702...... 104

Mass Spectrometry ...... 105

Reduction and Alkylation...... 106

Peptide Sequencing...... 106

REFERENCES...... 111

CHAPTER 5 ...... 112

Characterization of a Novel P-superfamily Conotoxin using Natural Abundance High Resolution-Magic Angle Spinning Nano-NMR ...... 112

ABSTRACT...... 112

INTRODUCTION...... 112

MATERIALS AND METHODS...... 114

Preparation of bru9a ...... 114

NMR Sample Preparation...... 114

NMR Spectroscopy ...... 115

Structural Assignments and Restraints...... 115

RESULTS AND DISCUSSION ...... 116

1H 1D experiments...... 116

ix Structural Assignments...... 117

Secondary Structure...... 128

REFERENCES...... 132

CHAPTER 6 ...... 135

Polypeptide Chains Containing D-γ-Hydroxyvaline ...... 135

ABSTRACT...... 135

INTRODUCTION...... 136

RESULTS AND DISCUSSION ...... 140

EXPERIMENTAL SECTION ...... 157

Peptide Isolation...... 157

Peptide Sequencing and Synthesis...... 157

Mass Spectrometry...... 158

NMR Spectroscopy...... 158

Molecular Modeling...... 159

Acknowledgment...... 159

REFERENCES...... 161

LIST OF FIGURES

Figure 1.1. Diagram of the exterior shell structure of a cone snail...... 8

Figure 1.2. Diagram of the soft exterior anatomy of a cone snail...... 9

Figure 1.3. Diagram of the cone snail venom apparatus...... 10

Figure 1.4. Diagram of the cone snail radular teeth based on feeding class...... 11

Figure 1.5. Photograph of Conus brunneus, the “Brown Cone”...... 12

Figure 1.6. Dissection of Conus brunneus...... 13

Figure 1.7. Venom Duct of Conus brunneus...... 14

Figure 1.8. Radular tooth of Conus brunneus...... 15

Figure 1.9. Common polychaete Caribbean worms,...... 16

Figure 1.10. Conus brunneus using the hook and pull feeding strategy...... 17

Figure 1.11. Photograph of Conus gladiator...... 19

Figure 2.1. Primary structures of the well characterized P-superfamily conotoxins:...... 26

Figure 2.2 Map of Costa Rica indicating Conus brunneus collection sites...... 28

Figure 2.3. Elution profile of 10 mg of crude Conus brunneus venom...... 37

Figure 2.4. Graph of the retention time vs MW distribution ...... 38

Figure 2.5. Elution profile of 20 mg of crude Conus brunneus venom...... 39

Figure 2.6. Elution profile of peak 6 from size-exclusion column ...... 41

Figure 2.7. Elution profile of peak 5 from semi-preparative column...... 42

Figure 2.8. MALDI-TOF MS spectrum of bruF060502...... 44

xi Figure 2.9. ESI-Ion trap MS spectrum of native bruF060502...... 45

Figure 2.10. ESI-Ion trap MS spectrum of reduced, but not alkylated bruF060502...... 46

Figure 2.11. ESI-Ion trap MS spectrum of bruF060502, reduced and alkylated

(iodoacetamide)...... 47

Figure 2.12. ESI-Ion trap MS spectrum of bruF060502, reduced and alkylated (4-

Vinylpyridine)...... 48

Figure 2.13. Elution Profile of Partially Reduced & Alkylated bruF060502 ...... 49

Figure 2.14. Standard nomenclature for MS/MS peptide backbone cleavage ions...... 50

Figure 2.15. ESI-Ion trap MS spectrum of the PTC derivative of bruF060502...... 51

Figure 2.16. ESI-Ion trap MS/MS spectrum of the PTC derivative of bruF060502...... 52

Figure 3.1. Primary structures of bru9a from Conus brunneus and the 5 known P-

Superfamily conotoxins ...... 60

Figure 3.2. Contour plot of a TOCSY spectrum...... 65

Figure 3.3. Contour plot of a 15N-HSQC spectrum...... 66

Figure 3.4. Contour plot of a 15N-HSQC-TOCSY spectrum...... 67

Figure 3.5. Contour plot of a 13C-HSQC spectrum...... 68

Figure 3.6. Expansion of the 13C-HSQC (Figure 3.5) spectrum...... 69

Figure 3.7. Expansion of the 13C-HSQC (Figure 3.5) spectrum...... 70

Figure 3.8. Contour plot of a DQF-COSY spectrum...... 73

Figure 3.9. Contour plot of a 13C-HSQC-TOCSY spectrum ...... 74

Figure 3.10. Contour plot of a TOCSY spectrum...... 75

Figure 3.11. Contour plot of a NOESY spectrum in 100 D2O...... 76

Figure 3.12. Contour plot of a NOESY spectrum in 90% H2O/10% D2O ...... 77

xii Figure 3.13. Graphical representation of the number of NOE restraints per residue ...... 78

Figure 3.14. Comparison of the Hα chemical shift values of bru9a to the random coil chemical shift values...... 79

Figure 3.15. Elution Profile of Partially Reduced & Alkylated bru9a...... 81

Figure 3.16. Backbone overlay of the 20 best structures of bru9a...... 84

Figure 3.17. Illustration of the bru9a backbone with the Cys disulfide bonds ...... 85

Figure 3.18. Structural comparison of the two P-superfamily conotoxins ...... 86

Figure 3.19. Charge distribution on the surface of bru9a ...... 87

Figure 3.20. RMSD per residue for bru9a ...... 88

Figure 3.21. Backbone representation of bru9a showing a high degree of variability in the

Arg side chains...... 89

Figure 3.22. 1D-NMR spectrum of bru9a...... 90

Figure 3.23. Primary structure of bru9a from Conus brunneus and the cyclotide

KAB15_OLDAF from the plant Oldenlandia affinis...... 91

Figure 4.1. Primary structures of the novel mini-M conotoxins: bru3a and bru3b, from

Conus brunneus...... 98

Figure 4.2. Elution profile of 20 mg of crude Conus brunneus venom...... 101

Figure 4.3. MALDI-TOF MS spectrum of bruF070303...... 102

Figure 4.4. MALDI-TOF MS spectrum of bruF070303, reduced and alkylated with iodoacetamide...... 103

Figure 4.5. Elution profile of 20 mg of crude Conus brunneus venom...... 104

Figure 4.6. MALDI-TOF MS spectrum of bruH090303...... 105

Figure 4.7. Sequence alignments of mini-M conotoxins...... 108

xiii Figure 4.8. Conus brunneus (A) and Conus regius (B) specimens...... 109

Figure 4.9. Three-dimensional structures of mr3a...... 110

Figure 5.1. 1D nano-NMR spectra of 180 nanomoles of bru9a...... 116

Figure 5.2. Contour plot of a nano-NMR TOCSY spectrum...... 117

Figure 5.3. Contour plot of a nano-NMR 15N-HSQC spectrum...... 118

Figure 5.4. Contour plot of a nano-NMR 13C-HSQC spectrum ...... 120

Figure 5.5. Expansion of the nano-NMR 13C-HSQC (Figure 3.5) spectrum...... 121

Figure 5.6. Expansion of the nano-NMR 13C-HSQC (Figure 3.5) spectrum...... 122

Figure 5.7. Contour plot of a DQF-COSY spectrum...... 125

Figure 5.8. Contour plot of a nano-NMR NOESY spectrum at 0oC...... 126

Figure 5.9. Contour plot of a nano-NMR NOESY spectrum at 25oC...... 127

Figure 5.10. Comparison of the Hα chemical shift values of bru9a to the random coil chemical shift values...... 128

Scheme 6.1...... 139

Figure 6.1. Conopeptide isolation from venom of (A) C. gladiator and (B) C. mus...... 142

Figure 6.2. NMR spectra of the -hydroxyconophans from C. gladiator...... 143

Figure 6.3. MALDI-MS/MS of the -hydroxyconophans from C. gladiator ...... 145

Figure 6.4. NMR spectra of the conophan from C. gladiator...... 146

Chart 6.1...... 147

Figure 6.5. MS/MS of the conophan from C. gladiator...... 149

Figure 6.6. Molecular model of -hydroxyconophan structural motif H3CC(O)-Ser-D-

Hyv-Trp-NH2...... 153

xiv

LIST OF TABLES

Table 2.1 Elution Times for Crude venom fractions ...... 40

Table 2.2 Elution Times for Peak 6 from size–exclusion column...... 41

Table 2.3 Elution Times for Peak 5 from semi-preparative column...... 43

Table 2.4 Amino Acid Average residue Mass (Da)...... 52

Table 2.5 Reported Amino Acid Composition for bru9a ...... 53

Table 3.1 1H chemical shift values of the bru9a conotoxin ...... 72

Table 3.2 Determination of Disulfide Bonding Pattern...... 80

Table 3.3 Structural input data parameters for XPLOR-NIH ...... 81

Table 3.4 Structural Statistics for bru9a ...... 83

Table 5.1 1H chemical shift values of the bru9a conotoxin ...... 124

CHAPTER 1

Background and Introduction

Cone snails have long been of interest to collectors because of their beautiful

shells and large number of variations. However, since the 1960’s, the venom of cone

snails has been looked at for their pharmaceutical properties (Endean and Izatt 1965).

Over the recent years there has been an explosion of research into the venom of cone snails because of their pharmaceutical potential based on the high selectivity binding of the venom components. These conopeptides are being investigated as possible therapeutics to treat a wide range of symptoms from chronic pain to diabetes (Lewis and

Garcia 2003). The venom components are also being looked at as probes in investigating ion channel functions (McIntosh, Olivera et al. 1999) and as structural scaffolds in the pharmaceutical design process .

There are over 700 species of the Conus genus (Röckel, Korn et al. 1995)

distributed in the tropical and sub-tropical areas of the Atlantic, Indian, and Pacific

Oceans (Filmer 2001). The Conus genus is considered one of the largest, most diverse,

and successful genus of any marine invertebrate (Kohn and Perron 1994) and has successfully undergone 55 million years of evolutionary refinement (Mari and Fields

2003). Within each species, their venom can have 50 – 300 unique peptides components.

To realize the potential magnitude the genus Conus venom has, this equates to a Conus

1 peptide library with a conservative estimate of greater than 100,000 distinct peptide components. Additionally, it has been shown that the Conus genus is diversifying at an unprecedented rate (Duda and Palumbi 1999). To add to the diversity of the venom, cone snails are characterized into three broad categories depending on their primary prey:

Vermivorous (worms hunters); Molluscivorous (mollusk hunters); and Piscivorous (fish hunters), each with their own uniquely designed cocktail a venom components

(Baldomero, Olivera et al. 1988).

Cone snails are predatory marine gastropods that use venom for means of predation and defense (Baldomero, Olivera et al. 1988). Their venom is a complex mixture of neurotoxins used to simultaneously target multiple physiological mechanisms within their prey (Terlau, Shon et al. 1996). The high affinity binding peptides in the venom are essential in targeting the nervous and muscular systems within their prey, rendering them immobile (Newcomb and Miljanich 2002). The use of their venom as a defense mechanism has circumstantially been confirmed by the inadvertent stinging of humans during handling. Human envenomations by the Conus geographus have proven to be lethal (McIntosh and Jones 2001). The lethal injections of the toxins in humans has caused the α-conotoxins to be listed as an agent Toxin in the Handbook of Chemical and

Biological Warfare Agents and described with symptoms of extreme flaccid paralysis with respiratory and circulatory failure (Ellison 2000). Cone snails have been described as “the most fantastic and indigenous chemical “war machines” which nature has developed” (Edstrom 1992).

2 Neurotoxins

Neurotoxins are toxins that generally interact with membrane proteins and ion

channel receptors. Peptide neurotoxins are commonly found in the venom of cone snails,

scorpions, spiders, and snakes. These venom toxins in nature are primarily used for

predation and defense and are being studied for their pharmaceutical potential. Cone snail

peptide neurotoxins represent one of the largest single sources of peptide neurotoxin

variations.

Conopeptides

Conopeptides is the wide-ranging term used to classify all peptides components

within the venom composition of the genus Conus. Conopeptides are generically

separated into two distinct groups (Terlau and Olivera 2004):

1. Disulfide poor, being absent of disulfide linkages or containing only one. This

group consist of Contulakins, Conantokins, Conofamides, Conopressins, and the

Contryphans;

2. Disulfide rich, containing two or more disulfide bonds referred to as Conotoxins.

This group consist of the A, I, M, O, P, S, T, V, and Y superfamilies.

Within each superfamily there are distinct frameworks based on the Cysteine structural motifs. Within in certain frameworks the conotoxins are further identified into families based on their broad molecular targets. These frameworks (detailed below) are used for identification nomenclature. The nomenclature adapted follows the standard conopeptide

naming conventions (McIntosh, Santos et al. 1999).

We conducted a current literature search of the conopeptide groups. Listed below

are the major groups within the disulfide poor group and the conotoxins superfamilies

3 along with their known or suspected molecular targets. This list may not be all inclusive,

and there is a strong possibility that new groups and superfamilies will be discovered.

Contulakin

The contulakins are linear polypeptides with no disulfide bridges and contain no cysteine residues. The contulakins target the Neurotensin receptor and when injected intracerebroventricular into mice caused motor-control-associated dysfunction (Terlau

and Olivera 2004) (Craig, Norberg et al. 1999).

Conantokin

The conantokins are linear polypeptides with no disulfide bridges and contain no

cysteine residues. The conantokins target the NMDA receptor and produce both

potentiation and inhibition of the NMDA receptor (Terlau and Olivera 2004)

(Ragnarsson, Yasuda et al. 2006).

Conorfamide

The conorfamides are linear polypeptides with no disulfide bridges and contain no

cysteine residues. The conorfamides are FMRFamide related peptides and may target the

RFamide receptor. A conorfamide caused hyperactivity in mice when intracranially

injected (Terlau and Olivera 2004) (Aguilar, Luna-Ramirez et al. 2008).

Conopressin

The conopressins contain a single disulfide bond and contain two cysteine

residues. The conopressins are part of the vasopressin-like peptide family and may target

the vasopressin receptor. A conopressin was shown to act as a selective antagonist at the

human V-1a receptor (Terlau and Olivera 2004) (Dutertre, Croker et al. 2008).

4 Contryphan

The contryphans contain a single disulfide bond and contain two cysteine residues. The target of the contryphans is the Calcium (Ca) channel (Terlau and Olivera

2004) (Sabareesh, Gowd et al. 2006).

A-superfamily

The A-superfamily consists of cysteine framework I (CC-C-C), framework II

(CCC-C-C-C), and framework IV (CC-C-C-C-C). Framework I and II target the nicotinic acetylcholine receptor (nACh) and are further categorized as an α family. Framework IV targets either the nACh categorized as an αA family or the Potassium (K) channel and are categorized as a A family (Terlau and Olivera 2004) (Santos, McIntosh et al. 2004).

I-superfamily

The I-superfamily consists of cysteine framework XI (C-C-CC-CC-C-C) and framework XII (C-C-C-C-CC-C-C). These frameworks target the K channel (Terlau and

Olivera 2004) (Brown, Begley et al. 2005) (Jimenez, Shetty et al. 2003).

M-superfamily

The M-superfamily consists of cysteine framework III (CC-C-C-CC) and framework XVI (C-C-CC). Framework III targets either the Sodium (Na) channel termed the μ family, the nACh termed the Ψ family, or the K channel termed M family (Terlau and Olivera 2004) (Pi, Liu et al. 2006) (Corpuz, Jacobsen et al. 2005)

O-superfamily

The O-superfamily consists of cysteine framework VI (C-C-CC-C-C), and framework VII (C-C-CC-C-C). Both Frameworks target the Na channel and are termed either the δ family or the μO family, the K channels termed the κ family, or the Calcium

5 (Ca) channel termed the ω family (Terlau and Olivera 2004) (McIntosh, Hasson et al.

1995)

P-superfamily

The P-superfamily consists of framework IX (C-C-C-C-C-C). The molecular

target for this framework is currently unknown, but when injected into the central

nervous system of mice, it causes uncontrollable spasms (Terlau and Olivera 2004)

(Lirazan, Hooper et al. 2000).

S-superfamily

The S-superfamily consists of framework VIII (C-C-C-CC-C-C-C). This framework inactivates the 5-HT3 receptor, a serotonin-gated ion channel and is termed the

σ family (Terlau and Olivera 2004) (England, Imperial et al. 1998).

T-superfamily

The T-superfamily consists of the framework V (CC-CC) and framework X (CC-

C-OC). Framework V’s biological activity is unknown. Framework X when injected

intra-cerebroventricularly into mice caused either seizures or flaccid paralysis and is

termed the χ family (Terlau and Olivera 2004) (Balaji, Ohtake et al. 2000) (Walker, Steel

et al. 1999).

V-superfamily

. The V-superfamily consists of framework XV (C-C-CC-C-C-C-C) (Peng, Liu et

al. 2008). The molecular target is unknown.

Y-superfamily

The Y-superfamily consist of framework XX (C-C-CC-C-CC-C) (Yuan, Liu et al.

2008). The molecular target is unknown.

6 Post-translational modifications

A common feature of conopeptides is they contain a high percentage of post- translational modifications. The most common feature is the formation of disulfide bonds, but they also commonly posses the non-standard amino acids, hydroxyproline, γ- carboxyglutamic acid, and bromotryptophan. Additionally, they undergo C-terminal amidation, hydroxylation, and glycosylation, among a host of other post-translational modifications (Marx, Daly et al. 2006). The conotoxins at times contain unique modifications, such as the presence of D-γ-Hydroxyvaline first discovered in our lab and reprinted in Chapter 6 (Pisarewicz, Mora et al. 2005). These post translational modifications help define the conotoxins and expand their molecular diversity. Although the evolutionary reasons for all post-translational modifications is unknown, these modifications appear to effect the structural stability and targeted activity of conotoxins

(Craik and Adams 2007).

Anatomy of the Cone Snail

Cone snail species are currently classified based on a detailed description of the shell (Figure 1.1).

Figure 1.1. Diagram of the exterior shell structure of a cone snail (Walls 1979)

Figure 1.2 shows the soft exterior anatomy of a cone snail. Generally visible on living species is the long muscular foot, the siphon, the eye stalk, and the rostrum which houses

the proboscis.

Figure 1.2. Diagram of the soft exterior anatomy of a cone snail (Walls 1979).

Anatomy of the Venom Apparatus

Cone snails belong to a specialized group of marine snails called the Conacea or

Toxoglossa (‘poison tongue’) along with turrids and terebras. These families share the use of the radula as a device for injecting poison through a hollow barb (Walls 1979).

The venom apparatus consist of four general parts. A large venom bulb, the venom duct, and the radular sac, and the radular teeth (figure 1.3).

Figure 1.3. Diagram of the cone snail venom apparatus (Olivera et al. 1988).

The venom bulb is responsible for forcing the toxins through the venom apparatus, but plays no role in the venom production. The venom duct synthesizes the venom components and houses the mature venom components. The venom duct is often subjected to cDNA analysis to create conopeptide precursor libraries. One of the limitations of cDNA libraries is they do not accurately describe the significant number of post-translational modifications that help define the conopeptides in their native state.

The radular sac houses the radular teeth which are used to deliver the venom (Walls

1979) via the proboscis. The hollow radular tooth attached to the proboscis is used like a harpoon to inject the venom into its prey and to tether its immobilized prey for engulfing

(hook and pull). Each radular tooth is used only once and is disposable, breaking away from the proboscis. Besides the diversity in venom composition, each cone uses their specialized radular teeth designed specifically for their preferred prey; worm, mollusk, or

10 fish. As can been seen in Figure 1.4, there is a stark difference between the teeth of the

different groups of cones based on prey.

Figure 1.4. Diagram of the cone snail radular teeth based on feeding class (Keen 1971).

Conus brunneus

Conus brunneus is a common cone species located in intertidal areas to moderately deep

water throughout the Eastern Pacific from the Gulf of California to Ecuador. Conus

brunneus was first named in 1828 by Wood and is often referred to as the “Brown Cone”

due to its color (Walls 1979). Figure 1.5 is a photograph of Conus brunneus collected off

the Pacific coast of Costa Rica and photographed at Florida Atlantic University.

Figure 1.5. Photograph of Conus brunneus, the “Brown Cone”.

12 Conus brunneus has a large identifiable venom duct that is easily seen upon dissection (Figure 1.6). This specimen was collected off the Pacific coast of Costa Rica and transported live to Florida Atlantic University.

Venom Duct

Figure 1.6. Dissection of Conus brunneus.

13 The venom duct of Conus brunneus is large and is generally off white to yellow

in color (Figure 1.7). The venom duct was isolated from the dissection of a specimen

(Figure 1.6) collected off the Pacific coast of Costa Rica

Figure 1.7. Venom Duct of Conus brunneus.

14 In addition to cone identification by the shell description, the radular tooth can be used to confirm the identity of a cone and its primary prey. A radular tooth of Conus

brunneus was isolated and photographed in-house on an inverted cell culture microscope

in a cell culture plate at 10x, 20x, and 40x magnification (Figure 1.8).

A B C

Figure 1.8. Radular tooth of Conus brunneus photographed on an inverted cell culture microscope with 10x (A), 20x (B), and 40x (C) magnification.

From Figure 1.8 it can be seen that the radular tooth resembles that of a harpoon.

This tooth contains two barbs and is used to inject the venom and to snare it prey. Conus brunneus uses a hook and pull feeding strategy. Conus brunneus is a vermivorous cone snail that preys primary on polychaete worms.

Figure 1.9 is a photograph of the common polychaete Caribbean bristle worm,

Hermodice carunculata, and the Caribbean red-tipped fire worm, Chloria viridis. Under aquarium conditions in the laboratory, Conus brunneus preyed on both species of worms.

A B

Figure 1.9. Common polychaete Caribbean bristle worm, Hermodice carunculata (A), and the Caribbean red-tipped fire worm, Chloria viridis (B).

Both species of worms were collected during a night dive of the intra-coastal waterway near the Blue Heron Bridge, Riviera Beach, FL during slack tide. Photographs were taken in-house at Florida Atlantic University.

A B

C D

Figure 1.10. Conus brunneus using the hook and pull feeding strategy. Proboscis extended (A), injection of venom (B), pulling of prey (C), initial engulfing (D).

Figure 1.10 illustrates Conus brunneus using the hook and pull feeding strategy.

As seen, the cone extends it proboscis until it finds its prey. Once the prey is identified, it injects its harpoon forcing the release of venom and snaring its prey. The paralyzing effects of the toxin are immediately visible on the worm. The cone then pulls the worm in

using its proboscis with the snared radular tooth embedded. The interesting observation is that the paralyzing effect is only seen locally and does affect the entire worm (whether this is a localized affect of the venom or a defense mechanism of the worm isolating itself is unknown). If the cone secures the worm, the worm will pinch off near the injection site

17 and continue on with no residual effects on the rest of the worm. Rarely was the cone observed engulfing an entire worm. The other interesting observation was that if the worm was able to free itself from the cone, the paralysis would wear off and the worm would regain movement near the injection site. On freed smaller worms, the worm would still pinch off the section of its body near the injection site. These observations provide a preliminary analysis that the venom components in Conus brunneus are reversible and do not permanently bind to the targeted receptors. This reversible action is a desirable property in the development or possible pharmaceuticals.

18 Conus gladiator

Conus gladiator is a common cone species located in shallow water from the Gulf of

California to Peru on the Western coast of Central America (Walls 1979). Conus gladiator was first identified in 1833 by Broderip. Figure 1.11 is a photograph of Conus brunneus collected off the Pacific coast of Costa Rica and photographed at Florida

Atlantic University. Conus gladiator has a small identifiable venom duct that is not easily seen upon dissection. The small venom duct is generally off white to yellow color.

Figure 1.11. Photograph of Conus gladiator .

19 Summary

This chapter provided a background of the cone snail and the venom components.

The potential application of conopeptides justifies the broad research interest in this topic. The conopeptides are typically 10-40 residues in length with a MW of less than

4000 Da. This is an ideal range for the analysis of peptides using specialized structural techniques to include, separation strategies, Edman sequencing, mass spectrometry, and natural abundance NMR. A description of Conus brunneus was provided as this cone’s venom was used primarily for isolation, characterization, and structural determination in my research (chapter 2-5). A description of Conus gladiator was provided as we identified a unique D-γ-Hydroxyvaline residue in one of its components (chapter 6).

Conopeptides have been characterized by classical techniques such as Edman degradation which yields the sequence of the peptide. Subsequent chemical synthesis of such sequences can yield enough material to carry out techniques such as the 3- dimensional structure determinations by NMR or X-ray methods. The advent of high throughput mass spectrometry based proteomic techniques promises to expand the current studies of conopeptides. The introduction of nano-NMR techniques, as discussed in this paper, should lead to access the 3-dimensional structural intricacies of conopeptides in their native state bypassing the need of synthetic material.

REFERENCES

Aguilar, M. B., K. S. Luna-Ramirez, et al. (2008). "Conorfamide-Sr2, a -

carboxyglutamate-containing FMRFamide-related peptide from the venom of

Conus spurius with activity in mice and mollusks." Peptides 29(2): 186-195.

Balaji, R. A., A. Ohtake, et al. (2000). "-conotoxins, a new family of conotoxins with

unique disulfide pattern and protein folding. Isolation and characterization from

the venom of Conus marmoreus." J Biol Chem 275(50): 39516-22.

Baldomero, M., B. M. Olivera, et al. (1988). Marine Snail Venoms. Handbook of Natural

Toxins. A. TU, Marcel Dekker, Inc.

Brown, M. A., G. S. Begley, et al. (2005). "Precursors of novel Gla-containing

conotoxins contain a carboxy-terminal recognition site that directs -

carboxylation." Biochemistry 44(25): 9150-9.

Corpuz, G. P., R. B. Jacobsen, et al. (2005). "Definition of the M-conotoxin superfamily:

characterization of novel peptides from molluscivorous Conus venoms."

Biochemistry 44(22): 8176-86.

Craig, A. G., T. Norberg, et al. (1999). "Contulakin-G, an O-glycosylated invertebrate

neurotensin." Journal of Biological Chemistry 274(20): 13752-13759.

Craik, D. J. and D. J. Adams (2007). "Chemical modification of conotoxins to improve

stability and activity." ACS Chem Biol 2(7): 457-68.

21 Duda, T. F., Jr. and S. R. Palumbi (1999). "Molecular genetics of ecological

diversification: duplication and rapid evolution of toxin genes of the venomous

gastropod Conus." Proc Natl Acad Sci U S A 96(12): 6820-3.

Dutertre, S., D. Croker, et al. (2008). "Conopressin-T from Conus tulipa reveals an

antagonist switch in vasopressin-like peptides." J Biol Chem 283(11): 7100-7108.

Edstrom, A. (1992). Venomous and Poisonous Animals. Malabar, Krieger Publishing

Company.

Ellison, D. H. (2000). Handbook of Chemical and Biological Warfare Agents, CRC Press

LLC.

Endean, R. and J. Izatt (1965). "Pharmacological study of the venom of the gastropod

Conus magus." Toxicon 3(2): 81-93.

England, L. J., J. Imperial, et al. (1998). "Inactivation of a serotonin-gated ion channel by

a polypeptide toxin from marine snails." Science 281(5376): 575-8.

Filmer, R. M. (2001). A Catalogue of Nomenclature and Taxonomy in the Living

Conidae 1758-1998. Leiden, Backhuys Publisher.

Jimenez, E. C., R. P. Shetty, et al. (2003). "Novel excitatory Conus peptides define a new

conotoxin superfamily." J Neurochem 85(3): 610-21.

Keen, A. M. (1971). Sea Shells of Tropical West America. Stanford, Stanford University

Press.

Kohn, A. J. and F. E. Perron (1994). Life history and biogeography : patterns in Conus

Oxford University Press.

Lewis, R. J. and M. L. Garcia (2003). "Therapeutic potential of venom peptides." Nat

Rev Drug Discov 2(10): 790-802.

22 Lirazan, M. B., D. Hooper, et al. (2000). "The spasmodic peptide defines a new

conotoxin superfamily." Biochemistry 39(7): 1583-1588.

Mari, F. and G. B. Fields (2003). "Conopeptides: Unique pharmacological agents that

challenge current peptide methodologies." Chimica Oggi 21: 43-48.

Marx, U. C., N. L. Daly, et al. (2006). "NMR of conotoxins: structural features and an

analysis of chemical shifts of post-translationally modified amino acids." Magn

Reson Chem 44 Spec No: S41-50.

McIntosh, J. M., A. Hasson, et al. (1995). "A new family of conotoxins that blocks

voltage-gated sodium channels." J Biol Chem 270(28): 16796-802.

McIntosh, J. M. and R. M. Jones (2001). "Cone venom--from accidental stings to

deliberate injection." Toxicon 39(10): 1447-51.

McIntosh, J. M., B. M. Olivera, et al. (1999). "Conus peptides as probes for ion

channels." Methods Enzymol 294: 605-24.

McIntosh, J. M., A. D. Santos, et al. (1999). "Conus peptides targeted to specific nicotinic

acetylcholine receptor subtypes." Annual Review of Biochemistry 68: 59-88.

Newcomb, R. and G. Miljanich (2002). Neurotoxins of cone snail venom. Handbook of

Neurotoxicology: 617-651.

Peng, C., L. Liu, et al. (2008). "Identification of a novel class of conotoxins defined as V-

conotoxins with a unique cysteine pattern and signal peptide sequence." Peptides

29(6): 985-991.

Pi, C., J. Liu, et al. (2006). "Diversity and evolution of conotoxins based on gene

expression profiling of Conus litteratus." Genomics 88(6): 809-19.

23 Pisarewicz, K., D. Mora, et al. (2005). "Polypeptide chains containing D--

hydroxyvaline." J Am Chem Soc 127(17): 6207-15.

Ragnarsson, L., T. Yasuda, et al. (2006). "NMDA receptor subunit-dependent modulation

by conantokin-G and Ala(7)-conantokin-G." J Neurochem 96(1): 283-291.

Röckel, D., W. Korn, et al. (1995). Manual of the living Conidae Hackenheim, Verlag

Christa Hemmen.

Sabareesh, V., K. H. Gowd, et al. (2006). "Characterization of contryphans from Conus

loroisii and Conus amadis that target calcium channels." Peptides 27(11): 2647-

2654.

Santos, A. D., J. M. McIntosh, et al. (2004). "The A-superfamily of conotoxins: structural

and functional divergence." J Biol Chem 279(17): 17596-606.

Terlau, H. and B. M. Olivera (2004). "Conus venoms: a rich source of novel ion channel-

targeted peptides." Physiol Rev 84(1): 41-68.

Terlau, H., K. J. Shon, et al. (1996). "Strategy for rapid immobilization of prey by a fish-

hunting marine snail." Nature 381(6578): 148-51.

Walker, C. S., D. Steel, et al. (1999). "The T-superfamily of conotoxins." J Biol Chem

274(43): 30664-71.

Walls, J. (1979). Cone Shells: A Synopsis of the Living Conidae, T.F.H. Publications.

Yuan, D. D., L. Liu, et al. (2008). "Isolation and cloning of a conotoxin with a novel

cysteine pattern from Conus caracteristicus." Peptides 29(9): 1521-5.

CHAPTER 2

Conotoxin bru9a, a Novel P-Superfamily Conopeptide from the Venom of Conus

brunneus. Isolation and Characterization.

ABSTRACT

A novel P-superfamily conotoxin has been isolated directly from the venom of the

vermivorous cone snail, Conus brunneus. The novel toxin, bru9a, is a 24-residue polypeptide, SCGGSCFGGCWOGCSCYARTCFRD, containing the post-translationally modified amino acid, 4-hydroxyproline (O) and three disulfide bonds. The P-superfamily is defined by a Cysteine motif of six non-adjacent Cysteine residues with disulfide bonds between I - IV, II – V, III – VI which is classified as a IX conotoxin framework. bru9a was fully characterized and its disulfide bonding was determined to be, Cys2 - Cys14,

Cys6 - Cys16, and Cys10 - Cys21 matching that of the P-superfamily.

INTRODUCTION

Cone snails are a group of marine mollusks that produces peptide neurotoxins in

their venom as a means of predation. These neurotoxins, named conopeptides, are

classified into families based on unique Cysteine motifs. The majority of the

25 conopeptide families have known physiological targets. Some families, including the P-

superfamily, have no identified molecular targets.

The P-superfamily has been termed the “Spasmodic Peptide” as it causes a loss

of motor control and seizure like symptoms when injected into the central nervous system

of mice. (Lirazan, Hooper et al. 2000). Currently, the P-superfamily is one of the least

studied classes of conopeptides with only two sequences having been fully studied. The

two, tx9a and gm9a, from Conus textile and Conus gloriamaris are essentially homologs,

differing by only 3 amino acids in sequence (Miles, Dy et al. 2002).

bru9a SCGGSCFGG--CWOGCSCYARTCFRD- tx9a GCNNSCQγHSDCγSHCICTFRGCGAVN(NH2) gm9a SCNNSCQSHSDCASHCICTFRGCGAVN

Figure 2.1. Primary structures of the well characterized P- superfamily conotoxins: bru9a from Conus brunneus, tx9a from Conus textile, and gm9a from Conus gloriamaris. The characteristic cysteine bonding framework, I - IV, II – V, III – VI is shown.

Figure 2.1 shows the sequence alignment of the novel conotoxin bru9a, tx9a and

gm9a. As seen bru9a has 3 less amino acids than tx9a and gm9a making bru9a the

shortest know P-superfamily peptide. Additionally, bru9a has little sequence homology with tx9a and gm9a which could be attributed to the differences in their primary prey.

Conus brunneus is a worm hunter (vermivorous) while Conus textile and Conus gloriamaris are mollusk hunters (molluscivorous). In this paper we discuss the isolation and characterization of the novel P-superfamily conotoxin, bru9a.

26 MATERIALS AND METHODS

Specimen Collection

Live specimens of Conus brunneus were collected by SCUBA and inter-tidal pool exploration off the Pacific Coast of Costa Rica (Figure 2.2). Conus brunneus tend to be located along the coast in relatively shallow water (< 20 meters) consisting of a rock-sand and/or reef environment. A shallow water dive was conducted along the Pacific coast of

Costa Rica and several cone snails were obtained from rocks, coral, and sand.

Additionally, several specimens were collected from flipping rocks in the inter-tidal pools during low tide. The specimens were maintained in a saltwater environment and transported alive to the laboratory at Florida Atlantic University. The specimens were then placed in saltwater aquariums and kept alive until the extraction of crude venom was conducted.

Figure 2.2 Map of Costa Rica indicating Conus brunneus collection sites (2008)

Venom Duct Removal

Depending on the condition of the shell, one of two methods was used to remove the internal body from the exterior shell. If the exterior shell was damaged or in poor condition a destructive method was used. To use the destructive method the shell was cracked with a vise, the shell pieces were removed, and the interior body was left entirely intact. Although this method destroyed the shell it caused no damage to the interior body or the venom duct.

28 If the shell was to remain intact we employed a non-destructive method. To use

the non-destructive method live specimens of Conus brunneus, were placed in the -80

freezer for approximately 5 minutes. This was done to slightly freeze the interior body

and make the internal body more rigid. A thin firm dissecting needle was inserted into the

pointed end of the shell opening parallel to the shell body from Anterior to Posterior

(chapter 1, figure 1.1). The dissecting needle was used to gently turn the interior body

counter-clockwise in the direction of the shell opening. This procedure was carefully

repeated forcing the majority of the interior body out of the shell opening. Using this

method, we had to be careful to not damage or pierce the venom duct. Using the non-

destructive shell method caused some tearing of the interior body as it is connected

internally connected to the exterior shell. Using this method was successful the majority

of the time for removing the venom duct intact with no damage.

Once the interior body was removed, the body was carefully dissected to reveal

the venom duct. Conus brunneus venom ducts were easily identified as they are quite

large and generally white in appearance. A photo of the body and venom duct of Conus

brunneus was shown in the previous chapter 1.

Depending on the extraction method employed, the complete duct was either

frozen at -80 oC and then lyophilized or it was submitted directly for venom extraction.

Extraction of Crude Venom

The method of crude venom extraction employed was dependent on if the venom

duct itself was to remain partially intact. The venom duct was often kept intact as it can be used to create cDNA libraries of precursor proteins.

29 We generally employed the method of preserving the duct; although, we used

both methods extensively in the lab.

To use the duct preservation method, the duct was immediately used after

removal. The duct was kept intact with the venom bulb attached. A scalpel was used to create a clean cut at the very tip of the duct opposite the end with the venom bulb. Using

a rounded dissecting needle the venom duct was rolled/milked forcing the crude venom

out of the duct into an Eppendorf tube. This would be analogous of squeezing toothpaste

out of its tube. Although the majority of the crude venom was removed this way there

was still traces of venom in the ducts. The ducts were then cut into short 10 – 15 mm

sections and placed in a small amount of a 0.1% trifluoroacetic acid (TFA) / H20. The

ducts in solution were physically mixed then centrifuged with the supernatant pooled with the previously extracted crude venom. This procedure was repeated 2 additional times using 0.1% TFA / 30% Acetonitrile (ACN) / H2O and 0.1% TFA / 60% ACN / H2O respectively. After the third washing the duct were relatively transparent and showed no sign of holding any crude venom. The ducts were immediately frozen at -80 oC and

stored. The crude venom pooled solution was frozen at -80 oC and then lyophilized for

the purification stage.

To use the duct destruction method, the lyophilized intact venom duct was ground

up and dissolved in a small quantity of 0.1% TFA / H20. The solution was centrifuged

and the supernatant was collected. The remaining pellet was washed two additional times

using 0.1% TFA / 30% ACN / H2O and 0.1% TFA / 60% ACN / H2O respectively. After

each washing the sample was centrifuged and the supernatant was collected and pooled with the other washings. The crude venom pooled solution was frozen at -80 oC and then

30 lyophilized for the purification stage. A typical size Conus brunneus (length 40mm, width 25 mm) would produce approximately 20 – 30 mg of crude venom after extraction and lyophilization.

Purification of Peptides

We used a multi-stage High Performance Liquid Chromatography (HPLC) process to purify the samples. Lyophilized samples were dissolved in the appropriate solution and filtered prior to injection on any HPLC system.

Size Exclusion HPLC was done by subjecting the crude venom to a size exclusion column (Pharmacia Superdex-30, 2.5 x 100 cm, built in-house). The crude venom was dissolved in 1 ml of 0.1 M NH4HCO3 prior to injection. The mobile phase used was 0.1

M NH4HCO3. The flow rate was set at 1.5 ml/min with the UV detector measuring wavelengths at 220nm and 280nm respectively. Collected fractions were frozen, lyophilized, and then stored.

Semi-Preparative Reverse-Phase HPLC was done by subjecting the lyophilized samples obtained from size exclusion HPLC to the semi-prep column (Vydac, 218TP510,

10 x 259 mm, 5m particle diameter, 300 Ǻ pore size) reverse-phase HPLC. The size exclusion venom fractions were dissolved in 1 ml of 0.1% TFA / H20 prior to injection.

The mobile phase used a linear gradient of: 0-100 min. 100% A to 100% B; (Buffer A:

0.1% TFA / H2O, Buffer B: 0.1% TFA / 60% ACN / H2O). The flow rate was set at 3.5 ml/min with the UV detector measuring wavelengths of  = 220 nm and 280 nm respectively. Collected fractions were frozen, lyophilized, and then stored.

Analytical Reverse-Phase HPLC was done by subjecting the lyophilized samples obtained from Semi-Preparative HPLC to an analytical column (Vydac, 238TP54, 4.6 x

31 250 mm, 5 m particle diameter, 300 Ǻ pore size) reverse-phase HPLC. The analytical

reverse-phase fractions were dissolved in 0.5 ml of 0.1% TFA / H20 prior to injection.

The mobile phase used a linear gradient of: 0-100 min. 100% A to 100% B; (Buffer A:

0.1% TFA / H2O, Buffer B: 0.1% TFA / 60% ACN / H2O). The flow rate was set at 1.0

ml/min with the UV detector measuring wavelengths at  = 220nm and 280nm

respectively. Collected fractions were frozen, lyophilized, and then stored for further

analysis.

Mass Spectrometry

Mass spectrometry was carried out using either a matrix assisted laser desorption ionization with time of flight mass analysis (MALDI-TOF) or an electrospray ionization

(ESI) coupled to an ion trap analyzer.

For MALDI-TOF molecular mass determinations a Voyager-DE STR (Applied

Biosystems) instrument was used. Samples were prepared using the sandwich method of

layering the sample dissolved in 0.1% TFA / 60% ACN / H2O solution between two

layers of α-cyano-4-hydroxycinnamic acid matrix on a MALDI sample plate. To prepare the sample 0.5 l of matrix were added, dried up, and then 0.3 l of sample were added, let dry, and then add 0.5 l of matrix (sandwich method). Spectra were acquired in linear and reflector mode. The MALDI-TOF mass spectrometer was calibrated using Applied

Biosystem’s external calibration standard.

For ion trap MS measurements, a Finnigan LCQ Deca MS Ion Trap equipped with an electrospray assembly which induced ionization was used. ACN or acetic acid solutions were used as solvents.

32 Reduction and Alkylation

Isolated samples that were deemed pure were subjected to reduction and alkylation to determine the number of disulfide bonds using standard protocols (Coligan,

Dunn et al. 1996). Samples were deemed pure by analysis of the RP-HPLC analytical data and the MALDI-TOF MS sample data. A 1 l aliquot of the sample in water was used for reduction and alkylation. To the 1 l sample we added 3 l of 0.1 M Tris-HCL buffer (pH 6.2) and 6 l of 20mM dithiothreitol (DTT) for reduction. The sample was

o incubated at 60 C for 30 minutes. For the alkylation we added 3 l of NH4OH and either

5 l of iodoacetamide or 4-vinyl pyridine and incubated in the dark for 60 minutes at room temperature. The sample was purified by analytical reverse-phase HPLC or using a

C18 Zip Tip (Millipore).

Partial Reduction and Alkylation

The partial reduction and alkylation was completed using an established method

(van den Hooven, van den Burg et al. 2001). The sample is dissolved in 30 l 0.1M

Citrate buffer at pH 3 then reduced with 12 l of 0.1M tris-(2-carboxyehtyl)phosphine

(TCEP) and incubated for 10-15 minutes at room temperature. After reduction, the sample is immediately alkylated with 30 l of 0.1M N-Ethylmaleimide (NEM) and incubated for 30 minutes at room temperature. The sample is then purified using the analytical reverse phase HPLC.

Promotion and Stabilization of b1 peptide using N-terminal derivatives.

The promotion and stabilization of b1 and yn-1 peptides was carried out using N-

terminal Phenylthiocarbamyl (PTC) derivatives as described by Summerfield et al. (Scott

G. Summerfield 1997). An aliquot of sample containing approximately 100 pM of 33 peptide was lyophilized in an Eppendorf tube using a roto-vac. 2 l of a solution (5%, w/v) of phenyl isothiocyanate in ethanol-water-pyridine (1:1:1. v/v/v) was added to the dry sample. The sample was briefly vortexed to mix and then incubated at 45 oC for 30 minutes. The solvent was then removed under a stream of nitrogen. The sample was purified by analytical reverse-phase HPLC or by using a C18 Zip Tip (Millipore).

Amino Acid Analysis

Amino Acid Analysis of bru9a was performed at the University of Florida,

Biotech Core facilities.

NMR Spectroscopy

NMR spectra were recorded on Varian Inova 900 MHz, Inova 600 MHZ, and

Inova 500MHz spectrometers at 5 oC and 25 oC. Chapter 3 provides an in depth study of the use of NMR spectroscopy in determining the solution structure of bru9a. See

Materials and Methods of Chapter 3 for more details. This study made use of the

National Magnetic Resonance Facility at Madison, which is supported by NIH grants

P41RR02301 (BRTP/ NCRR) and P41GM66326 (NIGMS). Additional equipment was purchased with funds from the University of Wisconsin, the NIH (RR02781, RR08438), the NSF (DMB-8415048, OIA-9977486, BIR-9214394), and the USDA.

Peptide Sequencing

Reduced and alkylated peptides were subjected to sequencing using automated

Edman degradation. Alkylated peptides were adsorbed onto Biobrene-treated glass fiber filters and sequenced by Edman degradation using an Applied Biosystems Procise model

491A Sequencer equipped with a micro gradient delivery system. A reference elution profile for the amino acids was determined using a standard mixture of PTH-amino acids.

34 Bioassays

Initial bioassays were carried out by Cognetix Inc., Salt Lake City, UT. Screening

for biological activity was done using a di-8-ANEPPS membrane potential dye in cortical

cell cultures which indicates hyper polarization or depolarization of the cell membrane.

The fractions were further characterized using SBFI sodium sensitive dyes in cortical

cells.

Nomenclature

In this chapter we initially adopted an internal nomenclature to describe the

isolated samples. We used the nomenclature of 3 lower case letters followed by 1 upper case letter and then 6 numbers, example bruF060502. The 1st three lower case letters designated the cone species, the upper case letter designated the batch, the 1st 2 numbers

designated the size exclusion HPLC fraction, the next 2 numbers designated the semi- preparative reverse-phase HPLC fraction, and the last 2 numbers designated the analytical reverse-phase HPLC fraction. The example, bruF060502, would represent the

sample isolated from Conus brunneus, the 6th batch processed (F). The 6th peak from size

exclusion HPLC, the 5th peak from semi-preparative reversed phase HPLC, and the 2nd peak from analytical reverse-phase HPLC.

Once the sequence is determined the sample is appropriately renamed based on its

Cysteine framework to conform to the naming convention of McIntosh, et al. (McIntosh,

Santos et al. 1999). bruF060502 was renamed to the bru9a based on its comparable cysteine framework that defines the P-superfamily (framework IX).

35 RESULTS AND DISCUSSION

Purification of Peptides

We initially attempted to purify samples of Conus brunneus using primarily

reverse-phase HPLC. 10mg of crude venom was dissolved in 1 ml of 0.1% TFA / H2O and injected into a semi-prep reverse-phase column using the methods described previously in materials and methods. A total of 72 fractions were collected. Figure 2.3 shows the elution profile of 10 mg of crude Conus brunneus venom injected directly into a semi-prep reversed phase HPLC column.

Figure 2.3. Elution profile of 10 mg of crude Conus brunneus venom injected directly into a semi-prep reversed phase HPLC column

It can be seen that the majority of the components are in the overlapping regions in the center of the spectrum. The 72 fractions collected were submitted to MALDI for

Molecular Weight (MW) determination. Each fraction had several different compounds identified by MW. We plotted a graph of the retention time versus MW distribution to get a global view of the compounds (Figure 2.4).

Figure 2.4. Graph of the retention time vs MW distribution of crude venom fraction of Conus brunneus separated using a semi-prep reversed phase HPLC column

The graph shows the majority of the components coming out heavily overlapped in the region of 25 to 60 minutes. We employed different gradients, solvents, and tempuratures to try and improve the seperation. None of these changes made any significant differences in the seperation. It was at this point we knew we would have to employ a muti-pronged approach to get samples that were pure and adequetly seperated.

We employed the size exclusion HPLC as the intial step in purifying bru9a. 20mg of crude venom was dissolved in 1 ml of NH4HCO3 and injected into a size exclusion

column using the methods described previously in materials and methods. Figure 2.5

shows the elution profile of a crude batch of venom from Conus brunneus recorded at

220nm and 280 nm respectively. The elution of all components took 230 minutes to

38 complete. Initial analysis was conducted on the most abundant and low molecular

fractions

Peak 6 A

B Peak 6

0 230 minutes

Figure 2.5. Elution profile of 20 mg of crude Conus brunneus venom injected directly into a size exclusion HPLC column. 220 nm spectrum (A) and 280 spectrum (B)

Table 2.1 Elution Times for Crude venom fractions on Size Exclusion Column

Fraction Elution Time (minutes)

1 57.50 – 72.00 2 72.00 – 85.00 3 91.00 – 103.50 4 103.50 – 124.75 5 124.75 – 146.00 6 146.00 – 165.00 7 165.00 – 192.00 8 192.00 – 215.00 9 215.00 – 230.00

As can be seen from figure 2.5 there is a large peak in the 280nm range for peak

6. Although not shown, there is an additional large peak in the 254nm range for peak 6.

From initial observations peak 6 would appear to have peptides containing more aromatic

residues than the other peaks based on its increased absorbance in the 254/280nm range.

Peak 6 from the size exclusion HPLC was lyophilized and dissolved in 1 ml of

0.1% TFA / H2O and injected into a semi-prep reverse-phase column using the methods described previously in materials and methods. Figure 2.6 shows shows the Elution profile of peak 6 isolated on the size exclusion HPLC recorded at 220nm and 280 nm.

40 A Peak 5

B Peak 5

0 minutes 60

Figure 2.6. Elution profile of peak 6 from size-exclusion column injected directly into a semi-preparative reversed phase HPLC column. Chromatogram detected at  = 220 nm (A) and

chromatogram detected  = 280 nm (B)

Table 2.2 Elution Times for Peak 6 from size–exclusion column injected on semi- preparative reverse-phase column Fraction Elution Time (minutes)

1 8.74 – 9.35 2 12.20 – 13.15 3 21.30 – 22.30 4 23.00 – 24.16 5 25.18 – 26.45 6 26.45 – 28.70 7 28.70 – 29.60

41 As can be seen in the RP-HPLC, peak 5 appears to be the major single component

isolated from peak 6 of the size exclusion HPLC. Again, this peak is the largest

component when detecting at  = 280 nm, indicating a high content of Trp and Tyr amino

acids for this fraction.

Peak 5 from the size exclusion HPLC was lyophilized and dissolved in 0.5 ml of

0.1% TFA / H2O and injected into an analytical reverse-phase column using the methods

previously described in materials and methods. Figure 2.7 shows shows the elution

profile of peak 5 isolated from the semi-preparative HPLC recorded at  = 220 nm.

Peak 2

0 minutes 60

Figure 2.7. Elution profile of peak 5 from semi-preparative column injected directly into an analytical reverse-phase HPLC column, 220 nm.

Table 2.3. Elution Times for Peak 5 from semi-preparative column injected on semi- preparative reverse-phase Column Fraction Elution Time (minutes)

1 20.55 – 21.10 2 21.10 – 21.60 3 21.60 – 21.90 4 45.50 – 46.00

As seen in Figure 2.7 we have isolated a relatively pure peptide, peak 2, which was internally named bruF060502.

Mass Spectrometry

A 0.3 l aliquot of the bruF060502 sample purified above was placed on a

MALDI plate using the sandwich method as described previously and subjected to

MALDI analysis. Figure 2.8 shows the MALDI spectrum of bruF060502.

Figure 2.8. MALDI-TOF MS spectrum of bruF060502.

The MALDI-TOF MS spectrum shows a clear peak with a MW of 2535.19 Da.

The average mass for bru9a calculated from its sequence, 2534.90 Da, which is in good agreement with the experimental data.

Reduction and Alkylation

bruF060502 was subjected to reduction and alkylation to positively identify the number of Cys residues and the number of disulfide bonds. Reduction and alkylation was completed as previously described in materials and methods. All samples were run on the

Finnigan LCQ Deca MS Ion Trap equipped with an electrospray ionization source. Prior to MW determination, the samples were purified using analytical HPLC to remove any contaminants from the reduction / alkylation process. 44 An aliquot of bruF060502 was subjected to:

1. Native bruF060502, no alkylation / reduction;

2. Alkylation with no reduction;

3. Reduction and alkylation using iodoacetamide

4. Reduction and alkylation using 4-vinyl pyridine.

For comparison the native peptide was run and the spectrum recorded (Figure 2.9)

1268.1 10 0 80 A 60

40 12 78 . 7 20

Relative Abundance 1222.1 12 9 7. 3 143.9222.9 367.1 479.0 639.1 774.9 863.3 967.7 116 2 . 0 14 14 . 0 14 70 . 7 1702.1 1790.2 1964.8 0 200 400 600 800 10 0 0 12 0 0 1400 1600 1800 2000 m/ z

1268.1 10 0 12 6 7. 6 1268.5 80 B 1267.2 1269.1 60 40 1267.1 1269.6 20 1264.9 1266.0 12 70 . 0 12 70 . 5 Relative Abundance 1264.1 12 71. 4 12 71. 9 1265.1 12 70 . 6 12 72 . 2 0 12 6 5 1266 1267 12 6 8 1269 12 70 12 71 12 72 12 73 12 74 m/ z Figure 2.9. ESI-Ion trap spectra of native bruF060502. MW (A) and isotopic distribution (B).

From the spectrum it can be seen that the 2+ charge state of bruF060502 at 1268.1 Da.

This is confirmed by the isotopic distribution differing by 0.5 Da confirming this is the 2+ charge state corresponding to the native MW of 2535.00 Da.

A bruF060502 sample was subjected to alkylation with no reduction. This test was performed to determine if any free Cys residues occurred in the native peptide. The sample was alkylated with iodoacetamide as previously described in materials and methods and its MW was determined (Figure 2.10). 45

845.4 10 0 1267.2 80 A 60

40 852.8 805.4 629.4 673.4 12 78 . 2 20 541.4 858.8

Relative Abundance 12 2 1. 0 1298.1 393.2 497.3 941.0 110 4 . 2 1174 . 2 1399.3 1488.6 16 12 . 4 1655.4 1837.7 1896.6 0 400 600 800 1000 1200 1400 1600 18 0 0 2000 m/ z

12 6 7. 9 10 0 1267.4 80 B 1266.91267.5 1268.3 60 1267.4 1267.9 1268.4 1268.8 1267.0 40 1268.9 1269.3 12 6 7. 0 1268.3 1268.8 1269.9 20 1266.9 1267.8 1268.7 1269.3 1269.8 Relative Abundance 12 6 6 . 3 12 6 6 . 8 1268.1 1269.4 0 1266.5 1267.0 1267.5 1268.0 1268.5 1269.0 1269.5 12 70 . 0 m/ z

Figure 2.10. ESI-Ion trap MS spectra of reduced, but not alkylated bruF060502. MW (A) and isotopic distribution (B).

From the spectrum you can see the 3+ charge state at 845.4 Da and the 2+ charge state at 1267.2 Da. This is confirmed by the isotopic distribution differing by 0.5 Da for the 2+ charge state. These weights and charge states correspond to the native MW of

2535.00 Da. There was no mass increase indicating the peptide had no Cys residues in

their free state.

A bruF060502 sample was subjected to reduction and alkylation with iodoacetamide. The reduction process breaks the disulfide bonds and leaves free Cys

residues. The alkylation agent iodoacetamide reacts with the free Cys residues forming S-

carboxamidomethyl cysteine. The mass difference between S-carboxamidomethyl Cys

and native free Cys is 57.02 Da. The sample was reduced and alkylated with

iodoacetamide as described previously and its MW was determined (Figure 2.11).

46 962.1 10 0

80 A 1441.9 60 968.9

40 1063.6 13 17. 9 585.4 629.4673.4 833.6 914.8 113 7. 2 20 365.1 12 18 . 7 13 9 7. 2 1460.8 Relative Abundance 497.3 157 2 . 8 1657.1 170 5. 1 18 6 7. 5 19 58 . 0 0 400 600 800 1000 1200 1400 1600 18 0 0 2000 m/ z

1442.0 10 0 1441.5

80 B 1442.4 1441.5 1442.0 1441.5 1442.5 60 1443.4 1441.0 1441.1 1441.2 1442.1 1442.9 40 1443.4 14 4 3 . 9 1441.6 1442.5 1440.9 1442.3 1443.9 20 1440.7 14 4 1. 9 1442.7 1443.5 Relative Abundance 1440.2 1440.4 1443.2 1443.6 0 1440.5 1441.0 1441.5 1442.0 1442.5 1443.0 1443.5 1444.0 m/ z Figure 2.11. ESI-Ion trap spectra MS of bruF060502, reduced and alkylated (iodoacetamide). MW (A) and isotopic distribution (B).

From the spectrum you can see the 3+ charge state at 962.1 Da and the 2+ charge state at 1441.9 Da. This is confirmed by the isotopic distribution differing by 0.5 Da for the 2+ charge state. These weights and charge states correspond to a MW weight of

2881.8 Da, indicating a mass increase of 347.6 Da over the native MW of 2535.00 Da.

With each Cys alkylation adding 57.02 Da, this indicates the presence of 6 Cys residues and 3 disulfide bonds.

A bruF060502 sample was subjected to reduction and alkylation with 4- vinylpyridine. The reduction process breaks the disulfide bonds and leaves free Cys residues. The alkylation agent 4-vinylpyridine reacts with the free Cys residues forming

S-pyridylethyl Cys. The mass difference between S-pyridylethyl Cys and native free Cys is 105.06 Da. The sample was reduced and alkylated with 4-vinylpyridine as described previously and its MW was determined (Figure 2.12).

x2 635.2 10 0 A 793.5 1057.5

50 365.3 10 2 2 . 7 15 8 5 . 7 529.7 805.4 10 58 . 5 717.4 767.2 987.3 905.7 1064.4 13 2 7. 3 1481.0 159 7. 3 453.8 497.4 12 17. 5 173 5. 3 1792.9 1881.2

Relative Abundance 19 8 9 . 5 0 400 600 800 1000 1200 1400 1600 18 0 0 2000 m/ z

1057.7 10 0

80 1057.7 B 10 57 . 4 10 58 . 0 60 10 57 . 3 1057.7 1058.3 1058.6 10 57. 1 10 58 . 3 40 1057.0 10 56 . 8 20 10 56 . 3 10 56 . 4 10 58 . 8 10 59 . 4 10 59 . 7 10 59 . 8

Relative Abundance 10 58 . 5 10 59 . 2 10 56 . 7 0 10 56 . 5 1057.0 1057.5 10 58 . 0 10 58 . 5 10 59 . 0 10 59 . 5 1060.0 m/ z Figure 2.12. ESI-Ion trap MS spectra of bruF060502, reduced and alkylated (4-Vinylpyridine). MW (A) and isotopic distribution (B).

From the spectrum you can see the 3+ charge state at 1057.5 Da and the 2+ charge state at 1585.7 Da. This is confirmed by the isotopic distribution differing by 0.3 Da for the 3+ charge state. These weights and charge states correspond to a MW weight of

3169.6 Da which is a mass increase of 635.4 Da over the native MW of 535.00 Da. With

each Cys alkylation adding 105.06 Da this indicates the presence of 6 Cys residues and 3

disulfide bonds.

Through the use of multiple reduction alkylation steps we were able to positively

identify that bruF060502 has 6 Cys residues forming 3 disulfide bonds. Additionally we

confirmed that bruF060502 had no free Cys residues in its native state.

Partial Reduction and Alkylation

We completed a chemical conformation of the disulfide bonding framework of

bruF060502 (Pellicier 2006). The partial reduction and alkylation of bruF060502 was

48 completed using the method previously described in materials and methods and purified

using analytical reverse-phase HPLC (Figure 2.13).

Contaminent Native 2 Cys 4 Cys 6 Cys

38 minutes 80

Figure 2.13. RP-HPLC elution Profile of Partially Reduced & Alkylated bruF060502 in the Analytical Column

The partial reduction and alkylation of bruF060502 was completed and the

fractions were sequenced. The results revealed a Cys bonding patter of, I - IV, II – V, III

– VI for bruF060502.

Promotion and Stabilization of B1 peptide using N-terminal derivatives.

We used the N-terminal phenylthiocarbamoyl (PTC) derivative of bruF060502 to

promote and stabilize the B1 and Yn-1 fragments during MS/MS of the

peptide. Figure 2.14 shows the standard nomenclature used for peptide backbone

cleavage ions.

Figure 2.14. Standard nomenclature for MS/MS peptide backbone cleavage ions

The purpose of this experiment is two-fold. First it was used to determine if the

N-terminus was blocked. If the N-terminus is blocked the peptide is not suitable for sequencing using the standard Edman method (Allen 1989). Using MS/MS to determine the blockage uses very little sample as this can be detected into the pM and fM range.

The second part of this process determines the amino acid residue for the b1 / N-terminus

if it is not blocked. The bruF060502 PTC derivative was prepared as described previously

in materials and methods. The sample was purified using analytical reverse-phase HPLC to remove contaminants. The mass difference between the native peptide and the PTC derivatized peptide is 135.2 Da. The peptide was subjected to MS followed by low energy collisionally activated dissociation MS/MS to promote peptide fragmentation.

Figure 2.15 shows the MS spectrum.

50 1335.6 100 A 80

40 897.7 585.4 629.4 1267.1 717.4 1345.5 20 497.4 879.1 903.5

Relative Abundance Relative 1209.5 928.8 1366.3 1467.5 1739.6 1832.6 1925.8 0 400 600 800 1000 1200 1400 1600 1800 2000 m/z [] 1335.4 100 B 80 1334.9 1334.4 1335.0 1335.8 60 1335.3 1334.4 1335.8 1336.3 1336.8 40 1334.8 1335.9 1336.7 1337.3 1334.4 1336.2 1337.8 20 1337.2 1337.7 Relative Abundance Relative 1336.4 0 1334.5 1335.0 1335.5 1336.0 1336.5 1337.0 1337.5 1338.0 m/z Figure 2.15. ESI-Ion trap spectra of the PTC derivative of bruF060502. MW (A) and isotopic distribution (B).

From the spectrum you can see the 2+ charge state at 1335.6 Da. This is confirmed by the isotopic distribution differing by 0.5 Da for the 2+ charge state. The

weight and charge states correspond to a MW weight of 2670.2 Da which is a mass

increase of 135.2 Da over the native MW of 2535.00 Da. This evidently shows the

formation of the bruF060502 PTC derivative. Figure 2.16 shows the collisionally activated MS/MS spectrum of the 1335.6 Da ion.

Figure 2.16. ESI-Ion trap MS/MS spectra of the PTC derivative of

bruF060502.

From the spectrum there are two peaks of interest. The first is the 1267.8 Da peak

which represents the 2+ charge state of the native bruF060502 peptide correlating to a native MW of 2534.6 Da. This indicates the Phenyl isothiocyanate is being fragmented from the native peptide leaving the native peptide intact. The second peak of extreme interest is the 1224.1 Da peak in the 2+ charge state correlating to a MW of 2447.2 Da

representing the Yn-1 fragment. The MW difference between the native peptide and the

Yn-1 fragment is 87.4 Da (b1 fragment). Table 2.4 shows the amino acid average residue

mass for the lower MW amino acids.

Table 2.4. Amino Acid Average residue Mass (Da)

Glycine 57.05 Alanine 71.08 Serine 87.08 Proline 97.12 Valine 99.13

From the table, the b1 fragment is identified as Ser. This was later confirmed by sequence analysis. The b1 PTC derivative was not directly observed because its MW was below the detection limits of the instrument.

Peptide Amino Acid Analysis

The University of Florida received a small aliquot of bruF060502 to conduct

Amino Acid analysis. The peptide was hydrolyzed using standard procedures of 6N HCL.

This process destroys the Cys residues, which was acceptable because the number of cysteine residues had been determined to be 6 through the reduction and alkylation / mass spectrometry process as detailed above. The University of Florida facility had difficulty interpreting the acquired amino acid spectrum. This is later attributed to the presence of the Hyp residue as determined by sequence analysis. Table 2.5 lists the results of the amino acid composition as reported by the University of Florida minus the Cys and Trp residues.

Table 2.5. Reported Amino Acid Composition for bru9a.

Amino Acid # of residues

ASX (D,N) 1 SER (S) 1 GLY (G) 6 ARG (R) 2 THR (T) 1 ALA (A) 2 TYR (Y) 1 PHE (F) 2

53 The reported composition was missing 2 Ser residues, had an extra Gly residue, had an extra Ala residue, and was missing the Hyp or Pro residue. The amino acid composition provided and the actual sequence determined later was substantially different. At this stage of the project, we were attempting to sequence the peptide using

NMR, mass spectrometry, and the amino acid composition in order to avoid the expense required for Edman degradation sequencing. Amino acid analysis proved to be unreliable for this purpose; therefore, traditional Edman degradation was used to sequence the peptide.

Peptide Sequencing

The bruF050602 peptide was sequenced using the method previously described in

materials and methods. From the data acquired the sequence was determined to be:

SCGGSCFGGCWOGCSCYARTCFRD.

Where O represents the non-standard amino acid 4-hydroxy-proline (Hyp).

The molecular weight of this peptide was determined by the program

ProteinProspector (Burlingame 2008) to be 2534.9 Da which is in excellent agreement

with experimentally derived MW of 2535.0 da.

Bioassays

We submitted 72 isolated fractions from the crude venom of Conus brunneus

(Figure 2.3). These fractions were isolated strictly using semi-preparative HPLC and as

seen previously contained numerous compounds per fraction. Out of the collected

fractions, we submitted half of the sample to Cognetix Inc. for screening of initial

bioactivity. The samples were first lyophilized before shipping.

54 The di-8-ANEPPS membrane potential assay identified fractions 13, 41, 43, and

54 as causing a hyperpolarization of the cell membranes. No fractions were identified as being depolarizing.

The SBFI sodium dye assay identified fractions 5, 7, 11, 13, 14, 20, 24, 25, 31,

34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 46, 48, 50, 54, 59, 61, 66, 70 as all being TTX insensitive sodium channel blockers. Bru9a was identified as a component in peak 31.

Although this is an initial screening for bioactivity, the venom of Conus brunneus shows promise in bioactivity and should be explored further.

In this chapter we discuss the isolation and characterization of a novel P- superfamily conotoxin, bru9a isolated from Conus brunneus. As illustrated in Figure 2.3, the crude venom composition is extremely complex and bru9a is one of the major components. bru9a is unique in that it is the only P-superfamily conotoxin to contain the post-translational modification 4-hydroxy-proline (Hyp). bru9a shares little sequence homology with other known P-superfamily sequences and is the shortest known P- superfamily at only 24 residues. bru9a is the first P-superfamily isolated and fully characterized from the worm hunting cone snails. Although the biological activity is currently unknown, this class of conotoxin exhibits unusual responses when injected into mice. The characterization of bru9a relied heavily on HPLC, mass spectrometry, and sequencing. The complete solution structure of bru9a by NMR is detailed in the next chapter.

REFERENCES

(2008). Map of Costa Rica, Central America, MSN encarta, World Atlas.

Allen, G. (1989). Sequencing of proteins and peptides. Amsterdam, Elsevier.

Burlingame, A. (2008). "ProteinProspector." from http://prospector.ucsf.edu.

Coligan, J. E., B. M. Dunn, et al. (1996). Current Protocols in Protein Science, John

Wiley & Sons, Inc.

Lirazan, M. B., D. Hooper, et al. (2000). "The spasmodic peptide defines a new

conotoxin superfamily." Biochemistry 39(7): 1583-1588.

McIntosh, J. M., A. D. Santos, et al. (1999). "Conus peptides targeted to specific nicotinic

acetylcholine receptor subtypes." Annual Review of Biochemistry 68: 59-88.

Miles, L. A., C. Y. Dy, et al. (2002). "Structure of a novel P-superfamily spasmodic

conotoxin reveals an inhibitory cystine knot motif." J Biol Chem 277(45): 43033-

43040.

Pellicier, J. (2006). Isolation and Characterization of Novel Conopeptides from Conus

brunneus. The Charles E. Schmidt College of Science. Boca Raton, Florida

Atlantic University. Masters.

Scott G. Summerfield, M. S. B. S. J. G. (1997). "Promotion and Stabilization of b1 ions

in Peptide Phenythiocarbamoyl Derivatives: Analogies with Condensed-phase

Chemistry." Journal of Mass Spectrometry 32(2): 225-231.

56 van den Hooven, H. W., H. A. van den Burg, et al. (2001). "Disulfide bond structure of

the AVR9 elicitor of the fungal tomato pathogen Cladosporium fulvum: evidence

for a cystine knot." Biochemistry 40(12): 3458-66.

CHAPTER 3

Three–Dimensional Solution Structure of Conotoxin bru9a: a Novel P-Superfamily

Conotoxin from the Venom of Conus brunneus with an Unusual Cystine Knot Motif.

ABSTRACT

The three-dimensional solution structure of conotoxin bru9a, a novel P- superfamily peptide from the venom of the vermivorous cone snail, Conus brunneus, has been determined using 2D homo-nuclear, and 2D hetero-nuclear NMR spectroscopy.

The 24-residue polypeptide, SCGGSCFGGCWOGCSCYARTCFRD, contains the post- translationally modified amino acid, 4-hydroxyproline (O). bru9a contains six non- adjacent cysteine residues with disulfide bonds between Cys2 and Cys14, Cys6 and Cys16, and Cys10 and Cys21. This cysteine framework is defined by the P-superfamily (IX

framework) and contains a classic inhibitory cysteine knot motif. bru9a is the first

structural determination of a P-superfamily conotoxin isolated directly from the venom.

INTRODUCTION

Cone snails are a group of marine mollusks that produces peptide neurotoxins in

their venom as a means of predation. These neurotoxins are either disulfide rich

58 (conotoxins) or non-disulfide rich. The conotoxins are classified into families based on their Cys frameworks. A common feature among conotoxins is the high percentage of

post-translational modifications (Buczek, Bulaj et al. 2005).

Conus brunneus is a venomous marine mollusk which preys upon worms

(vermivorous). Conopeptides isolated from the venom of vermivorous cone snails make

good initial candidates for drug discovery based on their immediate paralyzing effect, but

they do not appear to permanently bind to the selected receptors and are reversible as

seen in the feeding observations as described in chapter 1. The major component isolated

directly from the crude venom of Conus brunneus is bru9a, a 24-residue polypeptide,

SCGGSCFGGCWOGCSCYARTCFRD. The isolation and characterization of bru9a has

been detailed in the previous chapter 2. bru9a contains the post-translationally modified

amino acid, 4-hydroxyproline (O) and has six non-adjacent cysteine residues with

disulfide bonds between Cys2 and Cys14, Cys6 and Cys16, and Cys10 and Cys21. This

cysteine framework belongs to the P-superfamily (IX framework) of conotoxins and

contains a classic inhibitory cysteine knot motif. bru9a has been designated by the

conotoxin naming conventions adopted by MacIntosh et al (McIntosh, Santos et al.

1999).

A search of the literature identified 3 known P-superfamily conotoxins: tx9a from

Conus textile which defined the P-superfamily (Lirazan, Hooper et al. 2000), gm9a from

Conus gloriamaris which provides the only solution structure of a P-superfamily

conotoxin (Miles, Dy et al. 2002), and beTXIIb from Conus betulinus which was reported before the P-superfamily was defined (Chen, Fan et al. 1999). A search of the web server,

Conoserver, (Kaas, Westermann et al. 2008) identified two additional P-superfamily

59 sequences from Conus regius and Conus amadis which were not identified in the literature. To date these are the only 5 known conotoxin sequences of the P-superfamily.

Figure 3.1 shows the sequence of bru9a and the 5 known P-superfamily conotoxins displaying the Cysteine bonding framework, I - IV, II – V, III – VI of the P-superfamily

(framework IX).

bru9a SCGGSCFGG--CWOGCS-CYART--CFRD- Tx9a GCNNSCQγHSDCγSHCI-CTFRG--CGAVN(NH2) Gm9a SCNNSCQSHSDCASHCI-CTFRG--CGAVN BeTXIIb GCGGVCAYGESCPSSCNTCYSAQ--CTAQ Rg9.1 FCGQACSSVK-CPKKCF-CHPEEKVCYREMRTKERD Ca SCNNSCQQHSQCASHCV-CLLNK--CRTVN

Figure 3.1. Primary structures of bru9a from Conus brunneus and the 5 known P-Superfamily conotoxins. The characteristic cysteine bonding framework, I - IV, II – V, III – VI for the P-superfamily is shown.

The P-superfamily has been termed the “Spasmodic Conotoxin” because when tx9a is injected into the central nervous system of normal mice they then display a similar behavior to that of a well known mutant, the spasmodic mouse. “The distinctive symptomatology that these mouse mutants exhibit is a extreme hypersensitivity to sensory stimuli, causing a loss of motor control and seizure-like symptoms” (Lirazan,

Hooper et al. 2000). Although the symptomatology of tx9a has been described the molecular target of the P-superfamily remains unknown.

The solution structure of bru9a is the first structural determination of a P- superfamily conotoxin isolated directly from the crude venom, therefore containing all 60 intact de facto post-translational modifications. bru9a is the shortest primary structure known (24 residues) for the P-superfamily and has little sequence homology with the other members of the P-superfamily. As a characteristic of conotoxins there can variability within the superfamilies as to sequence and peptide length.

In addition to 1H-based homonuclear 2D-NMR techniques, non-traditional,

natural abundance, hetero-nuclear NMR experiments, 13C-HSQC, 13C-HSQC-TOCSY,

13C-HMBC, 15N-HSQC, and 15N-HSQC-TOCSY were applied to the structural analysis

of bru9a. This paper demonstrates the successful use of hetero-nuclear NMR experiments

for the rapid and unambiguous assignments of spin systems and sequence specific

assignments of a conotoxin directly isolated from its natural source.

MATERIALS AND METHODS

Preparation of bru9a

The isolation and characterization of bru9a has been detailed in Chapter 2.

NMR Sample Preparation

The NMR samples were prepared by dissolving lyophilized bru9a in 130 l of

90% H2O/10% D2O containing DSS as an internal standard. The sample was adjusted to

pH 3.6 using 0.1 M solutions of NaOH and HCl at room temperature. The sample was

placed in a 3mm Shigemi D2O matched tube (Shigemi Inc, Allison Park, PA) with a

sample height of 18 mm. bru9a is highly soluble in aqueous solution and had a final

concentration of ~ 0.3mM.

After all 90% H2O/10% D2O experiments had been acquired, the sample was

removed from the 3mm Shigemi tube and lyophilized. The sample was then dissolved in

61 180 l of 100% D2O. The sample was placed in a dry 3mm Shigemi D2O matched tube

with a sample height of 18 mm.

NMR Spectroscopy

NMR spectra were recorded on Varian Inova 900, Inova 600, and Inova 500

spectrometers at 5 oC and 25 oC. The experiments included 1D 1H, 1H-DQF-COSY, 1H-

TOCSY, 1H-NOESY, 13C-HSQC, 13C-HSQC-TOCSY, 13C-HMBC, 15N-HSQC, and 15N-

HSQC-TOCSY. Solvent suppression was obtained using pre-saturation, WATERGATE

(Piotto, Saudek et al. 1992) or DPFGSE (Hwang and Shaka 1995) pulse sequences.

Spectra were processed on a Sun Spark 5 and Sun Blade workstations using Varian

VNMR 6.1C software and converted and analyzed using the program, SPARKY

(Goddard T. D.) . This study made use of the National Magnetic Resonance Facility at

Madison, which is supported by NIH grants P41RR02301 (BRTP/ NCRR) and

P41GM66326 (NIGMS). Additional equipment was purchased with funds from the

University of Wisconsin, the NIH (RR02781, RR08438), the NSF (DMB-8415048, OIA-

9977486, BIR-9214394), and the USDA.

Structural Assignments and Restraints

Proton resonance assignments were made according to widely accepted standard procedures (Wüthrich 1986). Amino acid spin systems were identified from the

DPFGSE-TOCSY spectrum recorded at 5 oC on a Varian Inova 600 MHz spectrometer

and 15N-HSQC spectrum recorded at 25 oC on a Varian Inova 500 MHz spectrometer.

Sequential Amino acid spin systems were derived from the 200 ms DPFGSE-NOESY

spectrum recorded at 5 oC on a Varian Inova 900 MHz spectrometer.

62 Backbone dihedral angle φ restraints were derived from the DQF-COSY spectrum

o 3 recorded at 25 C on a Varian Inova 500 MHz spectrometer. The JHN-Hα coupling

constants were measured parallel to the F1 dimension and determined using the delta

o 3 function on VNMR 6.1C software. Φ angle restraints were set to -120 ± 40 for a JHN-Hα

o 3 ≥ 8 Hz and to -60 ± 30 for a JHN-Hα ≤ 5 Hz (Pardi, Billeter et al. 1984). Additional Φ

and Ψ dihedral angle restraints were determined from 1H, 13C, and 15N chemical shift data

obtained from multiple NMR experiments. Accurate chemical shifts combined with

database information provide good quantitative predications of Φ and Ψ dihedral angles with confidence scores (Cornilescu, Delaglio et al. 1999). The chemical shifts were placed into STAR 2.1 format and run through the web server for predicting protein

torsion angle restraints, PREDITOR (Berjanskii, Neal et al. 2006).

Hydrogen bonding restraints were derived from D2O exchange 1D, TOCSY, and

NOESY NMR experiments.

Distance restraints were derived from the 200 ms DPFGSE-NOESY spectrum

recorded at 5 oC on a Varian Inova 900 MHz spectrometer. Cross peak volumes were

determined by integrating the individual cross peaks within the SPARKY program and

classified as strong, medium, weak, or very weak, corresponding to upper bounds interproton distance limits of 2.5, 3.5, 5.0, and 6.0 Å, respectively. All lower bounds interproton distance limits were set to 1.8 Å (van der Waals radius). Pseudoatom corrections were applied to non-stereo specifically assigned methylene protons, methyl protons, and magnetically equivalent protons on the aromatic side chains of the amino acids Phe and Tyr (Wuthrich, Billeter et al. 1983).

63 Structure Calculations

bru9a structures were calculated with the program XPLOR-NIH (Schwieters,

Kuszewski et al. 2003) using the torsion angle dynamics protocols provided within

XPLOR-NIH (Stein, Rice et al. 1997). Structures produced with XPLOR-NIH were viewed and analyzed using the program CHIMERA (Pettersen, Goddard et al. 2004).

RESULTS AND DISCUSSION

Structural Assignments

The 1H spin system resonance assignments for bru9a were achieved using a

combination of TOCSY, 13C-HSQC, 13C-HSQC-TOCSY and 15N-HSQC experiments. 22

of the 24 spin systems were identified in the amide fingerprint region of the TOCSY

spectrum (Figure 3.2). Ser1 and Hyp12 were not observed as expected since the N- terminus (Ser1) exchanges rapidly and Hyp12 does not have an amide proton. These

amino acids were later identified using the NOESY Hα (i) to HN (i+1) assignment

“walk” and various hetero-nuclear spectra as detailed below.

Figure 3.2. Contour plot of a TOCSY spectrum acquired on a 600 MHz spectrometer at 5 oC. Spin systems are identified with the HN- Hα cross peak.

As seen in Figure 3.2 there is overlap in the regions of Arg19 and Arg23, Cys21 and

Trp11, and Asp24 and Gly8 recorded at 5 oC. At 25 oC there is overlap between Asp24 and

Phe7. To unambiguously assign these residues we implemented natural abundance 15N-

HSQC and 15N-HSQC-TOCSY experiments. The 15N-HSQC should identify the amide protons for the spectrum minus Ser1 and Hyp12 (Figure 3.3). 65

Figure 3.3. Contour plot of a 15N-HSQC spectrum acquired on a 500 MHz spectrometer at 25 oC. Spin systems are identified with the HN-N cross peak.

The 15N-HSQC clearly identified the amide protons of each spin system and

provided valuable insight to overlapping spin systems within the TOCSY fingerprint

region. The 15N-HSQC did not identify Ser1 and Hyp12 as expected; however, it

surprisingly did not identify Cys2 either. This spectrum did identify the amide Hε in Arg8 and Arg23 plus the amide Hε in Trp11. Gly residues are easily identified in the 15N

spectrum, as the 15N chemical shifts are considerably less than the other residues. The Tyr

residue is also easily identified because its 15N chemical shift is higher than most other

residues. A complete list of protein chemical shift statistics was obtained from the

Biological Magnetic Resonance Bank (BMRB 2008) and used as a reference.

To further demonstrate the use of hetero-nuclear experiments in natural abundance quantities of non-labeled peptides we acquired a 15N-HSQC-TOCSY

spectrum (Figure 3.4) realizing the relative insensitivity of the experiment.

Figure 3.4. Contour plot of a 15N-HSQC-TOCSY spectrum o acquired on a 500 MHz spectrometer at 25 C. Spin systems are identified with the HN-N cross peak and observable Hα-N cross

peaks are identified.

The 15N-HSQC-TOCSY revealed the HN to the Hα connectivity of approximately

42% of the residues including all 5 of the Gly residues Gly3, Gly4, Gly8, Gly9, Gly13, Ser5,

Ser15, Thr20, Cys14, and Asp24.

The sequence specific resonance assignments were determined using the

unambiguously assigned spin systems as described above and implementing the widely

excepted methods of Wuthrich (Wüthrich 1986). The methyl region near 0.95 ppm was

used to identify Thr20 and the methyl region near 1.20 ppm identified Ala18. Hyp12 was identified by its Hδ near 3.5 and 3.7 ppm and its Hβ near 2.01 ppm with no corresponding

HN (Marx, Daly et al. 2006). Ser1 was identified near 3.99 ppm and 3.92 ppm with no

67 corresponding HN in the amide region. To identify the 5 Gly residues a 13C-HSQC spectrum was acquired (Figure 3.5) to be used in conjunction with the other spectra.

Methyl (CH3) and methine (CH) peaks are positive (red) and methylene (CH2) peaks are negative (black).

Figure 3.5. Contour plot of a 13C-HSQC spectrum acquired on a o 500 MHz spectrometer at 25 C. CH3 / CH plots are positive (red) and CH2 plots are negative (black)

The spectrum clearly identifies the aromatic side chains, the methyl region, and the unique Gly region. From the spectrum we were able to unambiguously identify the 5

Gly residues (Fig 3.6)

Figure 3.6. Expansion of the 13C-HSQC (Figure 3.5) spectrum

highlighting the unique Gly region. The Hα – Cα have been identified for all 5 Gly in the primary structure.

Although this experiment was acquired to positively identify the Gly residues, of

more significance we were able to uniquely identify the crowded overlapping side chain

11 region of the aromatic residues. All Hs on the Trp aromatic rings were identified, Hε3 –

7.7 ppm, Hζ3 – 7.26 ppm, Hη2 – 7.1 ppm, Hζ2 – 7.5 ppm, Hδ – 7.24 ppm, and the Hε –

10.23 ppm was previously identified from its unique shift in the TOCSY. The Tyr17 magnetically equivalent aromatic Hs were identified as Hδ1 / Hδ2 – 6.99 ppm and Hε1 /

Hε2 – 6.75 ppm. The 2 Phe residue aromatic H were , Phe7 Hδ1 / Hδ2 – 7.2 ppm, Hε1 /

69 Hε2 – 7.35 ppm, Hζ – 7.28 ppm, and Phe22 Hδ1 / Hδ2 – 7.31 ppm, ε1 / ε2 – 7.43 ppm, Hζ

– 7.28 ppm (Figure 3.7). The 13C-HSQC was used in conjunction with the TOCSY to

identify the individual spin systems on the aromatic rings.

Figure 3.7. Expansion of the 13C-HSQC (Figure 3.5) spectrum highlighting the aromatic side chain region. All the aromatic Hs on Phe7, Trp11, Tyr17, and Phe22 have been identified.

The NOESY Hα (i) to HN (i+1) was used to sequentially identify the Ser1,Cys2,

Gly3, Gly4, Ser5, Cys6, Phe7, Gly8, Gly9, Cys10, Trp11 the sequential chain breaks here as

Hyp12 has no HN. From the C-terminus side of Hyp12 the sequential NOESY observed

was Hyp12, Gly13, Cys14, Ser15, Cys16, Tyr17, Ala18, Arg19, Thr20, Cys21, Phe22, Arg23, and

Asp24. In the HN to HN region of the NOESY spectrum HN (i-1) to HN (i) or HN (i+1) 70 to HN sequential residues were observed, Gly3 – Gly4, Cys6 – Phe7, Gly9 – Cys10, Gly13 –

Cys14, Tyr17 – Ala18, Ala18 – Arg19, Arg19 – Thr20, , and Arg23 – Asp24. The entire primary structure was assigned unique unambiguous resonances. Using the combined homonuclear and hetero-nuclear NMR experiments, a complete resonance assignment for bru9a at 25 oC was achieved (Table 3.1).

71 Table 3.1 1H chemical shift values of the bru9a conotoxin Residue HN(ppm) H (ppm) H (ppm) Others Ser1 4.13 3.97, 3.92 Cys2 8.79 4.78 2.76, 2.69 Gly3 7.41 3.51, 3.01 Gly4 7.99 4.23, 3.73 Ser5 7.92 4.67 3.99, 3.83 Cys6 8.5 4.3 3.19, 3.12 Phe7 8.23 4.34 3.11, 2.98 Hδ 7.22 Hε 7.36 Hζ 7.29 Gly8 8.22 3.87, 3.82 Gly9 7.75 4.47, 3.66 Cys10 8.87 4.82 2.86, 2.65 Trp11 8.33 4.83 3.37, 3.12 Hδ 7.23 Hε1 10.27 Hε3 7.66 Hζ2 7.52 Hζ3 7.26 Hη2 7.13 Hyp12 4.59 2.31, 2.01 Hδ 3.72, 3.48 Gly13 8.98 4.37, 3.78 Cys14 8.35 5.43 3.58, 2.72 Ser15 9.37 4.74 3.80, 3.70 Cys16 8.90 4.87 3.15, 2.98 Tyr17 9.55 4.61 2.97, 2.84 Hδ 6.95 Hε 6.76 Ala18 8.78 3.81 1.2 Arg19 8.50 3.80 2.15, 2.10 Hγ 1.53 Hδ 3.23, 3.19 Hε 7.25 Thr20 8.17 4.72 4.01 Hγ 0.97 Cys21 8.38 5.10 2.63, 2.49 Phe22 9.50 4.82 3.18, 2.99 Hδ 7.31 Hε 7.39 Hζ 7.28 Arg23 8.46 4.53 1.81 Hγ 1.74 Hδ 3.21 Hε 7.21 Asp24 8.24 4.48 2.74, 2.64

72 To determine dihedral torsion angles, a multi-pronged approach was employed using experimental 1D and DQF-COSY data along with dihedral database search

3 information. We were able to identify and measure 20 JNH-Hα from a DQF-COSY experiment (Figure 3.8).

Figure 3.8. Contour plot of a DQF-COSY spectrum acquired on a 500 MHz spectrometer at 25 oC. Spin systems are identified with the HN-N cross peak.

Additionally, we acquired to a 13C-HSQC-TOCSY to extract additional 13C chemical shift information (Figure 3.9). We were able to record a large percentage of the 13C

73 chemical shift information for bru9a. This was also used to confirm the sequential

resonance assignments obtained from the homonuclear experiment.

Figure 3.9. Contour plot of a 13C-HSQC-TOCSY spectrum acquired on a 500 MHz spectrometer at 25 oC.

This chemical shift information was compiled into STAR 2.1 format and submitted to the

PREDITOR (predicting protein torsion angle restraints) web server. PREDITOR was

used to generate a list of φ restraints with confidence deviations. The experimentally

derived restraints with deviations were compared to the database constraints with

deviations and found to be in good agreement. A total of 16 φ restraints were used, Cys2,

Ser5, Cys6, Phe7, Cys10, Trp11, Cys14, Ser15, Cys16, Tyr17, Ala18, Arg19, Thr20, Cys21, Phe22, 74 and Arg23. No φ restraints were used for the 5 Gly residues, the Hyp residue, the N- terminus Ser residue, or the C-terminus Asp residue.

To determine the hydrogen bonding characteristic of the peptide, bru9a was fully lyophilized and transferred to a 100% solution of D2O. NMR spectra were immediately recorded in the following order 1D, TOCSY, followed by a NOESY. From the 1D and

TOCSY, 13 amides did not exchange rapidly and were detectable after 1 hour (Figure

3.10).

Figure 3.10. Contour plot of a TOCSY spectrum acquired on a 900 o MHz spectrometer at 25 C / 100% D2O. Spin systems are identified with the HN-Hα cross peak.

From the NOESY spectrum (Figure 3.11), after several hours only 3 HN did not exchange and provided NOE peaks. 75

Figure 3.11. Contour plot of a NOESY spectrum acquired on a 900 o MHz spectrometer at 25 C in 100% D2O. Spin systems are identified with the HN-Hα cross peak.

The HN of 3 residues: Cys14, Ser15, and Phe22 were determined to be strongly

hydrogen bonded. After initial structural calculations were complete, the idealized

geometry for bru9a from strictly the NOE data was reviewed and it was determined Cys14

HN was H-bonded to Hyp12 carbonyl O, Ser15 HN was H-bonded to Phe22 carbonyl O,

and Phe22 HN was H-bonded to Thr20 carbonyl O. The software package CHIMERA was

employed to determine any additional H-bonding from the initial idealized geometry

determined from NOE data. The program identified one additional H-bond besides the 3

listed above. Thr20 HN had a inter-proton distance of 1.8 Å to Ala18 which is a strong

76 indication of H-bonding. Ala18 was one of the slowly exchanging HN determined from the TOCSY (Figure 3.10). Based on this information, the 4 H-bond constraints were added for structural refinement.

A total of 296 non-redundant H-H distance restraints were derived from the 200 ms DPFGSE-NOESY spectrum recorded at 5 oC (Figure 3.12).

Figure 3.12. Contour plot of a NOESY spectrum acquired on a 900 MHz spectrometer at 5 oC in 90% H O/10% D O. 2 2

Cross peak volumes were integrated and classified as strong, medium, weak, or very weak. Figure 3.13 illustrates a graphical representation of the number of restraints

77 per residue and divides them into intra-residue NOEs (red), sequential NOEs (black), and

medium/long range NOEs (blue).

Figure 3.13. Graphical representation of the number of NOE restraints per residue: intra-residue NOEs (red), sequential NOEs (black), and medium/long range NOEs (blue).

Even with the high percentage (>20%) of Gly residues the NMR NOESY spectrum showed a large number of NOEs indicating this was a highly structured peptide.

Secondary Structure

We did an initial assessment of the secondary structure using the Chemical Shift

Index (Wishart, Sykes et al. 1992). The Hα of bru9a were compared to the random coil chemical shift values for Hα (for residue 12 the random coil shift of Pro was used as the random coil shift for Hyp is unknown) (Wüthrich 1986). As seen from Figure 3.14 there are large deviations from the random coil values which is indicative that bru9a is a highly structured peptide.

Figure 3.14. Comparison of the Hα chemical shift values of bru9a to the random coil chemical shift values (Pro Hα random coil chemical shift value was used for residue 12 (Hyp).

From the Chemical Shift Index the structure does not contain any helical secondary structure. Residues 2 – 8 are classified as a multiple coil/turn region, residues 9

– 17 are classified as a β-strand region, residues 18 – 19 as a coil/turn region, and residues 20 – 23 as a β-strand region. Using the Chemical Shift Index as a preliminary tool, it supports bru9a as being a highly structured peptide.

Disulfide Connectivity Determination

To determine the disulfide connectivity of bru9a we used an analysis of the global

NOE data (Klaus, Broger et al. 1993) (Cooke, Carter et al. 1992). Local NOE data was not applicable because no Hα – Hβ or Hβ to Hβ between the Cys residues was observed.

We generated structures for the 15 possible disulfide bonding patterns of bru9a. Using

XPLOR-NIH we generated 25 structures using solely the global NOE data. A list of accepted structures matching the NOE data (no NOE violation exceeding 0.3 Å) for each 79 of the disulfide bonding patterns was completed. Table 3.2 shows the percentage of

accepted structures for each of the disulfide bonding patterns matching the global NOE data.

Table 3.2 Determination of Disulfide Bonding Pattern Disulfide Bonding Framework Accepted Structures based on NOE (%) 2-6, 10-14, 16-21 0 2-6, 10-16, 14-21 0 2-6, 10-21, 14-16 0 2-10, 6-14, 16-21 0 2-10, 6-16, 14-21 5 2-10, 6-21, 14-16 0 2-14, 6-10, 16-21 0 2-14, 6-16, 10-21 76 2-14, 6-23, 10-16 0 2-16, 6-10, 14-21 0 2-16, 6-14, 10-21 0 2-16, 6-21, 10-14 0 2-21, 6-10, 14-16 0 2-21, 6-14, 10-16 0 2-21, 6-16, 10-14 16

From Table 3.2 it can been seen that pairing 2 – 14, 6 – 16, and 10 – 21 produced

the greatest percentage of accepted structures based strictly on the global NOE data. The

disulfide connectivities determined above additionally match the disulfide pattern that defines the P-superfamily (framework IX) framework, I - IV, II – V, III – VI. Although this brute force global NOE pattern determination scheme is time intensive, it allows easy interpretation of the disulfide bonding framework based on global NOE data sets.

We proceeded to carry out the chemical determination of the cystine pairing of

bru9a (Pellicier 2006). The partial reduction and alkylation of bru9a was completed using

80 conventional methods (van den Hooven, van den Burg et al. 2001). Figure 3.15 shows the

elusion profile of the partially reduced and alkylated products of bru9a.

Native Contaminant 6 Cys 2 Cys 4 Cys

38 80

Figure 3.15. Elution profile of partially reduced and alkylated bru9a in the RP-HPLC analytical column.

The products of partial reduction and alkylation of bru9a containing 2 and 4

alkylated Cys were sequenced. The results confirmed that bru9a has framework IX

framework arrangement (I - IV, II – V, III – VI) typical of the P-superfamily of

conotoxins.

Structural Calculations

The input data for the structural calculations is listed in Table 3.3.

Table 3.3 Structural input data parameters for XPLOR-NIH

Total Inter-proton NOEs 296 Intra-residue NOEs 156 Sequential NOEs 78 Medium and long range NOEs 62 Disulfide Bonds 3 Φ – dihedral constraints 16 Hydrogen bonds 4

81 XPLOR-NIH was used to generate 250 initial structures using the modified protein.par and protein.top files. These files were modified to incorporate the Hyp residue which was not defined in the default parameters. The standard torsion angle dynamics script included with XPLOR-NIH was used. The 250 structures generated went through an additional refinement using the standard refine script. The 250 structures were then evaluated and 26 were determined to have no NOE violations > 0.17 Å and no dihedral angle violations > 5o. The 26 structures were visually inspected and a final set of 20 structures were selected. The energy and geometric statistics of the final structures are listed in Table 3.5. The experimental rmsd values were calculated using the program

MolMol (Koradi, Billeter et al. 1996) and are listed in Table 3.5. XPLOR-NIH was used to generate an average structure from the 20 final structures. The average structure was submitted to PROCHECK-NMR (Laskowski, Rullmannn et al. 1996) with the

Ramachandran statistics recorded in Table 3.4. No dihedral angles were in the disallowed range of the plot.

82 Table 3.4 Structural Statistics for bru9a

Constraint violations # of NOE distance violations > 0.17 Å 0 # of dihedral angle violations > 5o 0 XPLOR-NIH energies (kcal mol-1) Overall 74.68 ± 3.77 Bonds 3.67 ± 32 Angles 37.98 ± 2.17 Improper 12.70 ± 1.4 VDW 12.92 ± 1.14 NOE 0.36 ± 0.32 CDIH 7.05 ± .77 RMSD experimental (Å) Mean Global backbone 0.75 ± 0.17 Global heavy 1.41 ± 0.24 Global backbone (2-23) 0.58 ± 0.18 Global heavy (2-23) 1.28 ± 0.24 Ramachandran statistics from PROCHECK-NMR Most favored regions (%) 60 Additionally allowed regions (%) 26.7 Generously allowed regions (%) 13.3 Disallowed regions (%0 0

Figure 3.16 shows an overlay of the backbone atoms for the 20 structures of bru9a and a ribbon representation of the average structure.

83 N

C A B

Figure 3.16. Backbone overlay of the 20 best structures of bru9a (A) and the ribbon representation of the average structure (B).

As can be seen, the N-terminus and the C-terminus are poorly resolved. Additionally the

Phe7, Gly8, and Gly9 tight turn shows some variability. These regions influenced the overall rmsd of the structure, which can be considered fairly rigid in spite of the high content of Gly residues in bru9a. Figure 3.17 illustrates the bru9a backbone with the Cys disulfide bonds shown in yellow.

84 Cys-2

Cys-14 Cys-21

Cys-16 Cys-10

Cys-6

Figure 3.17. Illustration of the bru9a backbone with the Cys

disulfide bonds shown in yellow

The illustration clearly identifies bru9a as forming the classic inhibitory “cysteine knot” motif. It can be seen that the Cys10 – Cys21 disulfide bond passes through the Cys2

– Cys14 and Cys6 – Cys16 disulfide loops.

A comparison to the other structurally known P-superfamily conotoxin, gm9a,

was done (Figure 3.18). The gm9a structure was downloaded from the Protein Data Bank

(2008) with PDB ID 1ixt.

85 A B

Figure 3.18. Structural comparison of the two P-superfamily conotoxins. bru9a (A) and gm9a (B)

As can be seen both structures are similar in their folding, turns, and overall structure. Both form a classical inhibitory “cysteine knot” configuration. Although they are similar, there are differences in the structural conformations. These differences can be attributed to the sequence variation, loop lengths, and the different post-translational modifications present in the compound. bru9a is described as a multi-turn structure. bru9a has an initial turn at Cys2, Gly3, and Gly4, a reverse turn at residues Cys6, Phe7, Gly8, and

Gly9 which is allowed by the presence of Gly8 and Gly9, a turn at Trp11, Hyp12, Gly13, and Cys14, and another turn Tyr17, Ala18, Arg19, and Thr20. As illustrated the Gly residues play a crucial part in the solution structure as they are involved in 75% of the turns of bru9a.

Figure 3.19 shows the electrostatic grid potentials for bru9a calculated using the program GRASP (Nicholls, Sharp et al. 1991) and displayed using the program Chimera.

A B

Figure 3.19. Charge distribution on the surface of bru9a. A is a front view and B shows a 180o rotation. Electrostatic potentials are color coded positive (blue) and negative (red).

87 We calculated the rmsd per residue values for the backbone and heavy atoms of

bru9a (Figure 3.20)

Figure 3.20. RMSD per residue for bru9a: backbone (black) and heavy atoms (red).

From figure 3.20 it can be seen there is a large amount of conformational variability of the side chain atoms on Phe7, Arg19, and Arg23. This is in contrast to a very rigid

backbone for these same amino acids. Figure 3.21 is a graphical structural representation

of bru9a showing the high range of conformational variability in the Arg side chains.

88 Arg19 Arg23

Figure 3.21. Backbone representation of bru9a showing a high degree of conformational variability in the Arg side chains.

As seen in Figure 3.21, the final set of calculated structures shows the side chains

of Arg19 and Arg23 having numerous possible orientations and being free to move about

the solvent. In other conotoxins, such the well studied α-conotoxin GI, it has been shown

that the side chain of Arg are very flexible; they are free to move in and out of the solvent

and had several possible orientations (Guddat, Martin et al. 1996). Nevetheless, it has

been determined that residue Arg9 of α-conotoxin GI is responsible for its high

selectivity towards the γ-agonist site on the electric organ Acetylcholine receptor (Hann,

Pagan et al. 1997). Although the role of Arg in bru9a is unknown and the molecular target for the P-superfamily is unknown, it is interesting that bru9a has two Arg residues that are free to move about the solvent. Once a target for bru9a is found, the role these

Arg residues in bru9a should be investigated by mutagenesis.

89 Stability Test

The solution structure of bru9a confirms this is a highly structured and highly restrained peptide; therefore, we conducted a long-term stability test of the peptide in

solution. The peptide was stored in an aqueous, non-frozen solution in a refrigerated environment at 3 oC. A 1D NMR spectrum was initially recorded to determine the state of

the peptide, which matched previously recorded NMR spectrum. After 2 years another

1D NMR spectrum was recorded and compared to the initial spectrum. bru9a showed no

signs of degradation, deterioration, or conformational changes and appeared to remain as

structurally intact as the initial structure. Figure 3.22 shows the initial 1D spectrum of

bru9a and the 1D spectrum of bru9a recorded after year 2 of storage in an aqueous

solution.

Figure 3.22. 1D-NMR spectra of bru9a: Initial 1D (A) and 1D after two years of storage in solution (B).

90 This chapter describes the 3-dimensional solution structure of bru9a, the major

component isolated directly from the venom of Conus brunneus. Despite its small size,

bru9a is able to fold into a well defined, disulfide dependent, 3-dimensional structure.

This structure clearly forms the classic cysteine knot motif described primarily by toxic

and inhibitory polypeptides (Pallaghy, Nielsen et al. 1994).

The cysteine knot motif known as “knottins” consist of a “disulfide through

disulfide knot” and are found in the toxins of: cone snails, spiders, scorpions, and the horseshoe crab. Knottins are also found in peptides from plants and antimicrobial peptides (Chiche, Gelly et al. 2008). A BLAST alignment search of the knottin database revealed a similar sequence homology between bru9a and the cyclotide KAB15_OLDAF

(Figure 3.23) (Chiche, Gelly et al. 2008)

bru9a ---SCGGSCFGG-CW-OGCSCYARTCFRD KAB15_OLDAF GLPVCGESCFGGSCYTPGCSCTWPICTRD

Figure 3.23. Primary structure of bru9a from Conus brunneus and the cyclotide KAB15_OLDAF from the plant Oldenlandia affinis. The characteristic cysteine bonding framework, I - IV, II – V, III – VI is shown.

There is no known structure for KAB15_OLDAF to do a structural comparison.

Cyclotides are polypeptides with 28–37 amino acids and have a head-to-tail-cyclised backbone. They contain 6 Cys residues and 3 disulfide bonds and appear to be used in

91 multiple aspects of plant defense, although their actual functions are unknown (Plan,

Goransson et al. 2007).

Cyclotides are being considered as a stable peptide scaffolds suitable for protein

engineering (Aboye, Clark et al. 2008). However, cyclotides are difficult to produce by

recombinant DNA technologies. bru9a’s similarity to KAB15_OLDAF based on

sequence homology and the cysteine knot motif makes bru9a a suitable non-cyclic

template for protein engineering. From the sequences bru9a is the only P-superfamily to

end its sequence in with the RD amino acids, and based on sequence alignment is a

recurring sequence in the cyclotides. As seen in previous illustrations of bru9a, its N-

terminus and C-terminus are spatial close, slightly resembling that of a cyclic peptide,

although not connected. bru9a is at least 4 amino acids shorter than the cyclotides, possibly making it a tighter structure for scaffold engineering.

The P-superfamily conotoxin bru9a provides an excellent initial framework for the discovery of novel new peptide therapeutics or scaffold frameworks. Its biological

effects appear to be reversibly. bru9a is highly constrained and highly stable with a good

“shelf life” in solution. bru9a, at 24 residues, is the shortest known conotoxin to form the classic cysteine knot motif. The P-superfamily is the only family containing 6 non- adjacent Cys residues. This Cys motif potentially allows for the most variability of all the

6 Cys residue superfamilies based on sequence variations and particularly loop lengths.

These are all favorable qualities for potential drug development and scaffold engineering.

REFERENCES

(2008). "PDB." RCSB Protein Data Bank, from www.rcsb.org/pdb.

Aboye, T. L., R. J. Clark, et al. (2008). "Ultra-stable peptide scaffolds for protein

engineering - Synthesis and folding of the circular cystine knotted cyclotide

cycloviolacin O2." Chembiochem 9(1): 103-113.

Berjanskii, M. V., S. Neal, et al. (2006). "PREDITOR: a web server for predicting protein

torsion angle restraints." Nucleic Acids Research 34: W63-W69.

BMRB (2008). Biological Magnetic Resonance Bank, A Repository for Data from NMR

Spectroscopy on Proteins, Peptides, Nucleic Acids, and other Biomolecules,

University of Wisconsin.

Buczek, O., G. Bulaj, et al. (2005). "Conotoxins and the posttranslational modification of

secreted gene products." Cell Mol Life Sci 62(24): 3067-79.

Chen, J. S., C. X. Fan, et al. (1999). "Studies on conotoxins of Conus betulinus." J Nat

Toxins 8(3): 341-9.

Chiche, L., J.-C. Gelly, et al. (2008). "The Knottin Database." from knottin.cbs.cnrs.fr.

Cooke, R. M., B. G. Carter, et al. (1992). "The solution structure of echistatin: evidence

for disulfide bond rearrangement in homologous snake toxins." Protein Eng 5(6):

473-7.

93 Cornilescu, G., F. Delaglio, et al. (1999). "Protein backbone angle restraints from

searching a database for chemical shift and sequence homology." J Biomol NMR

13(3): 289-302.

Goddard T. D., K. D. G. Sparky 3, University of California, San Francisco.

Guddat, L. W., J. A. Martin, et al. (1996). "Three-dimensional structure of the -

conotoxin GI at 1.2 angstrom resolution." Biochemistry 35(35): 11329-11335.

Hann, R. M., O. R. Pagan, et al. (1997). "The 9-arginine residue of -conotoxin GI is

responsible for its selective high affinity for the agonist site on the electric

organ acetylcholine receptor." Biochemistry 36(29): 9051-9056.

Hwang, T. L. and A. J. Shaka (1995). "Water suppression that works. Excitation

sculpting using arbitary waveforms and pulsed field gradients." Journal of

Magnetic Resonance 112(A): 275-279.

Kaas, Q., J. C. Westermann, et al. (2008). "ConoServer, a database for conopeptide

sequences and structures." Bioinformatics 24(3): 445-6.

Klaus, W., C. Broger, et al. (1993). "Determination of the disulfide bonding pattern in

proteins by local and global analysis of nuclear magnetic resonance data.

Application to flavoridin." J Mol Biol 232(3): 897-906.

Koradi, R., M. Billeter, et al. (1996). "MOLMOL: a program for display and analysis of

macromolecular structures." J Mol Graph 14(1): 51-5, 29-32.

Laskowski, R. A., J. A. Rullmannn, et al. (1996). "AQUA and PROCHECK-NMR:

programs for checking the quality of protein structures solved by NMR." J

Biomol NMR 8(4): 477-86.

94 Lirazan, M. B., D. Hooper, et al. (2000). "The spasmodic peptide defines a new

conotoxin superfamily." Biochemistry 39(7): 1583-1588.

Marx, U. C., N. L. Daly, et al. (2006). "NMR of conotoxins: structural features and an

analysis of chemical shifts of post-translationally modified amino acids." Magn

Reson Chem 44 Spec No: S41-50.

McIntosh, J. M., A. D. Santos, et al. (1999). "Conus peptides targeted to specific nicotinic

acetylcholine receptor subtypes." Annual Review of Biochemistry 68: 59-88.

Miles, L. A., C. Y. Dy, et al. (2002). "Structure of a novel P-superfamily spasmodic

conotoxin reveals an inhibitory cystine knot motif." Journal of Biological

Chemistry 277(45): 43033-43040.

Nicholls, A., K. A. Sharp, et al. (1991). "Protein Folding and Association: Insights From

the Interfacial and Thermodynamic Properties of Hydrocarbons." Proteins: Struc.,

Func. and Genet 11: 281-296.

Pallaghy, P. K., K. J. Nielsen, et al. (1994). "A common structural motif incorporating a

cystine knot and a triple-stranded beta-sheet in toxic and inhibitory polypeptides."

Protein Sci 3(10): 1833-9.

Pardi, A., M. Billeter, et al. (1984). "Calibration of the angular dependence of the amide

proton-C alpha proton coupling constants, 3JHN, in a globular protein. Use of

3JHN for identification of helical secondary structure " Journal of Molecular

Biology 180(3): 741-751.

Pellicier, J. (2006). Isolation and Characterization of Novel Conopeptides from Conus

brunneus. The Charles E. Schmidt College of Science. Boca Raton, Florida

Atlantic University. Masters.

95 Pettersen, E. F., T. D. Goddard, et al. (2004). "UCSF Chimera - a visualization system for

exploratory research and analysis." J Comput Chem 25(13): 1605-12.

Piotto, M., V. Saudek, et al. (1992). "Gradient-tailored excitation for single-quantum

NMR spectroscopy of aqueous solutions." J Biomol NMR 2(6): 661-5.

Plan, M. R., U. Goransson, et al. (2007). "The cyclotide fingerprint in Oldenlandia

affinis: elucidation of chemically modified, linear and novel macrocyclic

peptides." Chembiochem 8(9): 1001-11.

Schwieters, C. D., J. J. Kuszewski, et al. (2003). "The XPLOR-NIH NMR molecular

structure determination package." Journal of Magnetic Resonance 160(1): 65-73.

Stein, E. G., L. M. Rice, et al. (1997). "Torsion-angle molecular dynamics as a new

efficient tool for NMR structure calculation." J Magn Reson 124(1): 154-64. van den Hooven, H. W., H. A. van den Burg, et al. (2001). "Disulfide bond structure of

the AVR9 elicitor of the fungal tomato pathogen Cladosporium fulvum: evidence

for a cystine knot." Biochemistry 40(12): 3458-66.

Wishart, D. S., B. D. Sykes, et al. (1992). "The chemical shift index: a fast and simple

method for the assignment of protein secondary structure through NMR

spectroscopy." Biochemistry 31(6): 1647-51.

Wüthrich, K. (1986). NMR of proteins and nucleic acids. New York, Wiley.

Wuthrich, K., M. Billeter, et al. (1983). "Pseudo-structures for the 20 common amino

acids for use in studies of protein conformations by measurements of

intramolecular proton-proton distance constraints with nuclear magnetic

resonance." Journal of Molecular Biology 169(4): 949-961.

CHAPTER 4

Isolation and Characterization of Two Novel Mini-M Conotoxins isolated from the

Venom of Conus brunneus.

ABSTRACT

Two novel conotoxins have been isolated directly from the venom of the

vermivorous cone snail, Conus brunneus. The novel toxins, bru3a and bru3b, are

members of the M-superfamily of conotoxins within the mini-M group. bru3a is a 18-

residue polypeptide, CCRWPRCNVYLCGOCCOQ, containing the post-translationally

modified amino acid, 4-hydroxyproline (O) and three disulfide bonds. bru3b is a 15-

residue peptide, CCQAYCSRYHCLPCC, containing three disulfide bonds. The M- superfamily is defined by several Cysteine-stabilized motif frameworks. bru3a and bru3b are classified as a type III framework of six Cysteine residues, CC-C-C-CC, belonging to the mini-M (m2) subclass.

INTRODUCTION

Cone snails are a group of marine mollusks that produces peptide neurotoxins in their venom as a means of predation. These neurotoxins, have evolved over 55 million years to form very specific ion channel targeted peptides (Terlau and Olivera 2004) named conopeptides. Conopeptides have a diverse base of therapeutic

97 applications in humans with several conopeptides reaching human clinical trials primarily for chronic pain (Olivera 2006). This chapter describes the isolation and characterization two novel conopeptides that belong to the M-superfamily.

Conopeptides are classified into families based on unique Cysteine motifs.

Within the M-superfamily, the type III framework of six Cysteine residues is defined by a

CC-C-C-CC arrangement. Framework III is broken down further into 4 branches based

on the number of amino acid residues between Cys4 and Cys5: m1, m2, m3, and m4.

Several pattern for cystine pairing have been determined for framework III. The

following trend has been observed so far for the different subclasses: m1 has a I – V, II –

IV, III – VI motif, m2 has a I - VI, II – IV, III – V motif, m3 is not yet determined, and m4 has I – IV, II – V, III – VI motif (Han, Wang et al. 2006). This is an unusual

observation within a superfamily. The m1-m3 branches have been referred to as the Mini-

M branch and the m4 branch referred to as the Maxi-M branch. The two novel conotoxins

described here are members of the Mini-M branch based on the two residues located

between Cys5 and Cys6.

bru3a CCRWPRCNVYLCGOCCOQ* bru3b CCQ-AYCSRYHCLPCC

Figure 4.1. Primary structures of the novel mini-M conotoxins: bru3a and bru3b, from Conus brunneus. * = amidated C-terminal

98 Figure 4.1 shows the sequence alignment of the 2 novel mini-M conotoxins, bru3a

and bru3b isolated from the crude venom of Conus brunneus. In this chapter we discuss

the isolation and characterization of the two novel mini-M conotoxins, bru3a and bru3b.

MATERIALS AND METHODS

Specimen Collection

Refer to materials and methods chapter 2.

Venom Duct Removal

Refer to materials and methods chapter 2.

Extraction of Crude Venom

Refer to materials and methods chapter 2.

Purification of Peptides

Refer to materials and methods chapter 2.

Mass Spectrometry

Mass spectrometry was carried out using a matrix assisted laser desorption

ionization time of flight (MALDI-TOF) instrument, Voyager-DE STR (Applied

Biosystems). Samples were prepared using the sandwich method of layering the sample

dissolved in 0.1% TFA / 60% ACN / H2O solution between two layers of α-cyano-4-

hydroxycinnamic acid matrix on a MALDI sample plate. To prepare the sample 0.5 l of matrix were added, dried up, and then 0.3 l of sample were added, let dry, and then add

0.5 l of matrix (sandwich method). Spectra were acquired in linear and reflector mode.

The MALDI-TOF mass spectrometer was calibrated using Applied Biosystem’s external calibration standard.

99 Reduction and Alkylation

Refer to materials and methods chapter 2.

Peptide Sequencing

Refer to materials and methods chapter 2.

Nomenclature

Refer to materials and methods chapter 2.

RESULTS AND DISCUSSION

Purification of Peptide, bruF070303

bruF070303 was purified using a multi-phase High Performance Liquid

Chromatography (HPLC) separation scheme. bruF crude venom was subjected to size exclusion HPLC following the protocol as previously described in materials and methods

(Figure 4.2)

100 Peak 7

0 minutes 230

Figure 4.2. Elution profile of 20 mg of crude Conus brunneus venom injected directly into a size exclusion HPLC column and detected at  = 220 nm.

Peak 7 was injected into a semi-preparative reverse-phase HPLC and the major peak was collected, peak 3. Peak 3 was injected into an analytical reverse-phase HPLC giving a well separated peak, peak 3 (data not shown). Per internal nomenclature, this peptide was initially named bruF070303.

Mass Spectrometry

A 0.3 l aliquot of the bruF070303 sample purified above was placed on a

MALDI plate using the sandwich method as described previously in materials and methods. The sample yielded the MALDI-TOF MS spectrum shown in Figure 4.3,

101

2126.9

Figure 4.3. MALDI-TOF MS spectra of bruF070303.

which shows a clear peak with a MW of 2126.9 Da. The average mass for bru9a

calculated from its sequence, 2128.00 Da, is in good agreement with the experimental

data.

Reduction and Alkylation

bruF070303 was subjected to reduction and alkylation to positively identify the

number of Cys residues and the number of disulfide bonds. Reduction and alkylation was

completed as previously described in materials and methods. Prior to MW determination,

the samples were purified using a pre-equilibrated C18 Zip Tip (Millipore) to remove any

contaminants from the reduction / alkylation process.

A bruF070303 sample was subjected to reduction and alkylation with iodoacetamide. The reduction process breaks the disulfide bonds and leaves free Cys

residues. The alkylation agent iodoacetamide reacts with the free Cys residues forming S-

102 carboxamidomethyl cysteine. The mass difference between S-carboxamidomethyl Cys and native free Cys is 57.02 Da. The sample was reduced and alkylated with iodoacetamide and its MW was determined (Figure 4.4).

2475.3

Figure 4.4. MALDI-TOF MS spectrum of bruF070303, reduced and alkylated with iodoacetamide.

From the MS spectrum, the reduce/alkylated product had a MW weight of 2475.3

Da, indicating a mass increase of 348.4 Da over the native MW of 2126.9 Da. With each

Cys alkylation adding 57.02 Da, this indicates the presence of 6 Cys residues and 3 disulfide bonds.

Peptide Sequencing

The bruF070303 peptide was sequenced using the method previously described in materials and methods. From the data acquired the sequence was determined to be:

CCRWPRCNVYLCGOCCOQ* (* = amidated C-terminal).

103 Where O represents the non-standard amino acid 4-hydroxy-proline (Hyp).

The molecular weight of this peptide was determined by the program ProteinProspector

(Burlingame 2008) to be 2534.9 Da which is in excellent agreement with experimentally

derived MW of 2535.0 da. The CCXCXCXCC framework is consistent with that of the

M-superfamily. bruF070303 was appropriately re-named bru3a to correspond to the

framework III conotoxin framework.

Purification of Peptide, bruH090702

bruH090702 was purified using a multi-phase High Performance Liquid

Chromatography (HPLC) separation scheme. bruH crude venom was subjected to size

exclusion HPLC following the protocol as previously described in materials and methods

(Figure 4.5)

Peak 9

0 minutes 230

Figure 4.5. Elution profile of 20 mg of crude Conus brunneus venom injected directly into a size exclusion HPLC column. 220 nm spectrum.

104 Peak 9 was injected into a semi-preparative reverse-phase HPLC and the major peak was collected, peak 7. Peak 2 was injected into an analytical reverse-phase HPLC giving a well separated peak, peak 2 (data not shown). Per internal nomenclature, this peptide was initially named bruH090702.

Mass Spectrometry

A 0.3 l aliquot of the bruH090702 sample purified above was placed on a

MALDI plate using the sandwich method as described previously in materials and methods. The sample yielded the MALDI-TOF MS spectrum shown in Figure 4.6,

1746.11

Figure 4.6. MALDI-TOF MS specta of bruH090303.

105 which shows a clear peak with a MW of 1746.11 Da. The average mass for bru9a calculated from its sequence, 1746.00 Da, is in good agreement with the experimental data.

Reduction and Alkylation

bruH090702 was subjected to reduction and alkylation to positively identify the number of Cys residues and the number of disulfide bonds. Reduction and alkylation was completed as previously described in materials and methods. Prior to MW determination, the samples were purified using a pre-equilibrated C18 Zip Tip (Millipore) to remove any contaminants from the reduction / alkylation process.

A bruH090702 sample was subjected to reduction and alkylation with iodoacetamide. The reduction process breaks the disulfide bonds and leaves free Cys residues. The alkylation agent iodoacetamide reacts with the free Cys residues forming S- carboxamidomethyl cysteine. The mass difference between S-carboxamidomethyl Cys and native free Cys is 57.02 Da. The sample was reduced and alkylated with iodoacetamide and its MW was attempted unsuccessfully.

Peptide Sequencing

The bruH090702 peptide was sequenced using the method previously described in materials and methods. From the data acquired the sequence was determined to be:

CCQAYCSRYHCLPCC.

Where O represents the non-standard amino acid 4-hydroxy-proline (Hyp).

The molecular weight of this peptide was determined by the program ProteinProspector

(Burlingame 2008) to be 1746.0 Da which is in excellent agreement with experimentally derived MW of 1746.11 da. The CCXCXCXCC framework is consistent with that of the

106 M-superfamily. bruH090702 was appropriately re-named bru3b to correspond to the framework III of conotoxins.

In this chapter we discuss the isolation and characterization of two novel M- superfamily conotoxins, bru3a and bru3b isolated from Conus brunneus. These conotoxins are members of the M-superfamily (framework 3) and are known as a Mini-M conotoxins based on the 1-3 residues between Cys4 and Cys5. These Mini-M conotoxins are likely to have a unique disulfide bonding motif of I - VI, II – IV, III – V, which is different from the other branches of the M-superfamily.

A common modifications found in the M-superfamily is the presence of hydroxyproline (Corpuz, Jacobsen et al. 2005). bru3a has 2 hydroxyproline (O) residues and bru3b does not contain a hydroxylated proline. A strictly conserved feature seen in the Mini-Ms is the presence of a Proline or Hydroxyproline immediately adjacent to Cys5

(-PCC- or –OCC-) (Wang, Jiang et al. 2008). bru3a has the –OCC feature and bru3b has the –PCC feature consistent with Mini-M branch.

A search of the Conoserver database (Kaas, Westermann et al. 2008) revealed several Mini-M conotoxins with similar sequence homology and/or alignment (figure

4.7) to the bru3a and bru3b.

107 BeTXIa -CCK-QSC--TTCMPCCW-- Reg12g -CCM-ALCSRYHCLPCC--- Reg12i -CCT-ALCSRYHCLPCC--- bru3b -CCQ-AYCSRYHCLPCC--- bru3a -CCRWPRCNVYLCGOCCOQ- QcIIIA -CCS-QDC--LVCIPCCPNX QcIIIB -CCS-RHC—WVCIPCCPNX Mr3a GCCGSFACRFGCVOCCV

Figure 4.7. Sequence alignments of mini-M conotoxins.

Two of these sequences were reported out of our lab from Conus regius (Franco,

Pisarewicz et al. 2006). From the sequence alignment there is significant sequence homology between bru3b, reg12g, and reg12i differing by only two amino acids.

Although this is interesting, this is not surprising as we have referred to Conus regius as the Atlantic Ocean equivalent of the Pacific Ocean cone snail, Conus brunneus. This comparison is based on observed similarities of the shells external features (typical size and shape), the cones native environment (shallow rocky areas), feeding habits (worm hunters), and similar venom apparatus (size and color of venom duct). Figure 4.8 shows a side by side comparison of two equivalently sized Conus brunneus and Conus regius specimens. Both species were collected and photographed in house at Florida Atlantic

University.

108

A B

Figure 4.8. Conus brunneus (A) and Conus regius (B) specimens.

Based on the 3D solution structure of mr3a isolated from Conus marmoreus, the

Mini-M branch forms a well defined globular peptide with a “triple turn” scaffold. This

Mini-Ms are extremely constrained by the nature of their short sequences and contain 3 disulfide bridges. When viewed from the front, the structure of mr3a has been referred to as a “plucked chicken”, figure 4.9 (McDougal and Poulter 2004). This is the only known structure of a m2 subclass conotoxin.

109 Backbone Front

Figure 4.9. Three-dimensional structures of mr3a. Shown are the backbone structure along with front view of the surface of the peptide. Blue regions are hydrophobic, and red regions are hydrophilic. (McDougal and Poulter 2004)

Although the molecular target of the Mini-M branch is unknown, mr3a causes hyperactivity, barrel rolling, circular motion, and convulsions when injected cranially into mice (Han, Wang et al. 2006) (Corpuz, Jacobsen et al. 2005)

The Mini-M branch conotoxins are smaller than the other M-superfamily branches and are found primarily in worm hunting and mollusk hunting snails (Wang,

Jiang et al. 2008). They contain an extremely high degree of sequence divergency, as it can be seen in Fig. 4.7. Sequence homology is only observed in related Conus species as

discussed above. Although their biological function is unknown, the Mini-Ms appears to

play an important role in the venom mollusk-hunting and worm-hunting cones.

110

REFERENCES

Burlingame, A. (2008). "ProteinProspector." from http://prospector.ucsf.edu.

Corpuz, G. P., R. B. Jacobsen, et al. (2005). "Definition of the M-conotoxin superfamily:

characterization of novel peptides from molluscivorous Conus venoms."

Biochemistry 44(22): 8176-86.

Franco, A., K. Pisarewicz, et al. (2006). "Hyperhydroxylation: a new strategy for

neuronal targeting by venomous marine molluscs." Prog Mol Subcell Biol 43: 83-

103.

Han, Y. H., Q. Wang, et al. (2006). "Characterization of novel M-superfamily conotoxins

with new disulfide linkage." Febs J 273(21): 4972-82.

Kaas, Q., J. C. Westermann, et al. (2008). "ConoServer, a database for conopeptide

sequences and structures." Bioinformatics 24(3): 445-6.

McDougal, O. M. and C. D. Poulter (2004). "Three-dimensional structure of the mini-M

conotoxin mr3a." Biochemistry 43(2): 425-9.

Olivera, B. M. (2006). "Conus peptides: biodiversity-based discovery and exogenomics."

J Biol Chem 281(42): 31173-7.

Terlau, H. and B. M. Olivera (2004). "Conus venoms: a rich source of novel ion channel-

targeted peptides." Physiol Rev 84(1): 41-68.

Wang, Q., H. Jiang, et al. (2008). "Two different groups of signal sequence in M-

superfamily conotoxins." Toxicon 51(5): 813-22.

111

CHAPTER 5

Characterization of a Novel P-superfamily Conotoxin using Natural Abundance

High Resolution-Magic Angle Spinning Nano-NMR

ABSTRACT

The novel P-superfamily conotoxin, bru9a, isolated from the venom of Conus brunneus has been analyzed using 1D, 2D homonuclear proton, and 2D heteronuclear

NMR methods using nanomole amounts of sample on a High Resolution Magic Angle

Spinning (HR-MAS) Nano-NMR probe. The novel conotoxin toxin, bru9a, is a 24- residue polypeptide, SCGGSCFGGCWOGCSCYARTCFRD, containing the post- translationally modified amino acid, 4-hydroxyproline (O) and stabilized by a network of three disulfide bonds. Because of the extensive post-translational modifications conotoxins undergo and the limited amount of sample obtainable, we explore the use of nano-NMR as an efficient, non-destructive characterization method.

INTRODUCTION

Cone snails are predatory marine gastropods that use a toxic venom for mean of predation and defense (Olivera et al. 1988). Their venom is a complex mixture of neurotoxins (conopeptides) used to simultaneously target multiple

112 physiological mechanisms within their prey (Terlau, Shon et al. 1996). These conopeptides are a rich source of peptides that are highly specific towards their molecular targets. Certain conopeptides are the most ion channel specific ligands identified

(McIntosh, Olivera et al. 1999). In addition to their exciting pharmaceutical potential, the

3D structure of conotoxins are being explored as templates to map out ion channel structural features (Hill, Alewood et al. 1997) and peptide toxins have indicated conformational changes in the target channels upon binding (Lange, Giller et al. 2006).

Conopeptides are relatively small (10 – 50 residues) sequences (Arias and

Blanton 2000) with low molecular weights. They have frequent post-translational modifications. Most notably, they undergo a complex array of disulfide bonding patterns making these peptides extremely stable and constrained. Furthermore, they often contain non-standard amino acids, C-terminal amidations, and hydroxylations (Marx, Daly et al.

2006). The properties of conopeptides make them well suited for NMR structural analysis.

Some conopeptide sequences have been synthesized from cDNA libraries obtained from cDNA analysis of the venom duct. Although this information is valuable it does not represent the conopeptides in their native state with the full repertoire of post- translational modification. Because of the unique nature of conopeptides, we find it crucial to characterize the conopeptides in their native state preserving all post- translational modifications and structural features that are the product of over 55 million years of evolutionary design (Mari and Fields 2003).

Nuclear magnetic resonance (NMR) has been extensively used in the past several decades to study peptide structures and is the primary method used for the structural

113 determination of conotoxins. Based on the comprehensive conopeptide database,

Conoserver (Kaas, Westermann et al. 2008), there are 110 reported 3D structures of

conopeptides with greater than 91% being structurally elucidated by NMR. Typically, the

native conopeptides isolated from the crude venom are only available in nanomole

quantities or less making them poor candidates for traditional structural NMR analysis.

The use of HR-MAS nano-NMR as the first step of characterization is ideally suited for

the structural studies of these native conopeptides. The advantage of using NMR as the

initial step in characterization is its non-destructive nature. From the NMR analysis you

can get an initial approximation of concentration. This helps in prioritizing further

characterization using destructive techniques (Mass Spectrometry, Edman sequencing,

amino acid analysis, etc.).

MATERIALS AND METHODS

Preparation of bru9a

The isolation and characterization of bru9a has been detailed in Chapter 2.

NMR Sample Preparation

The NMR sample was prepared by dissolving lyophilized bru9a in 40 μl of 90%

H2O/10% D2O containing TSP as an internal standard. The sample was adjusted to pH

3.6 using 0.01 M solutions of NaOH and HCl at room temperature. The sample was

placed in a Varian D2O matched 40 μl nano-probe tube (Varian Inc). A spinning rotor

was attached to the tube using acrylamide (super-glue) following protocols provided with the sample tube. The tube was initially spun in the probe, external to the magnetic to

114 ensure a spinning rpm of 2500 was attainable. bru9a is highly soluble in aqueous solution

and had a final concentration of 190 nanomoles.

NMR Spectroscopy

NMR spectra were acquired on a Varian Inova 500 spectrometer at 0 oC and 25

oC. The spectrometer was equipped with a nano-gHX probe. The experiments included

1D 1H, 1H-DQF-COSY, 1H-TOCSY, 1H-NOESY, 13C-HSQC, and 15N-HSQC. Solvent

suppression was obtained using pre-saturation, WATERGATE (Piotto, Saudek et al.

1992), and excitation sculpting (Callihan, West et al. 1996) pulse sequences. The sample

was spun at 2450 rpm to put the residual sidebands outside the spectral window of the

sample. Spectra were processed on a Sun Spark 5 and Sun Blade workstations using

Varian VNMR 6.1C software and converted and analyzed using the program, SPARKY

(Goddard , Kneller 2008) .

Structural Assignments and Restraints

Proton resonance assignments were made according to widely accepted standard

procedures (Wüthrich 1986) Amino acid spin systems were identified from the nano-

NMR TOCSY, 15N-HSQC, and 13C-HSQC spectrum recorded at 25 oC on a Varian

Inova 500 MHz spectrometer. Sequential Amino acid spin systems were derived from the

200 ms nano-NMR NOESY spectrum recorded at 0 oC on a Varian Inova 500 MHz

spectrometer.

We acquired a nano-NMR DQF-COSY spectrum recorded at 25 oC on a Varian

3 Inova 500 MHz spectrometer. The JHN-Hα coupling constants were able to be measured

parallel to the F1 dimension and determined using the delta function on VNMR 6.1C.

115 These measurements are instrumental in determining the backbone dihedral angle φ restraints for structural refinement.

No hydrogen bonding D2O experiments were conducted using the nano-probe.

RESULTS AND DISCUSSION

1H 1D experiments

Figure 5.1 shows the 1D nano-NMR spectra of 180 nanomoles of bru9a at 0 oC

and 25 oC.

Figure 5.1. 1D nano-NMR spectra of 180 nanomoles of bru9a at 0 oC (A) and 25 oC (B).

From the high quality 1D spectrum the Hε1 proton on the Trp residue is easily recognized

along with the methyl groups of the Thr and Ala residues.

116 Structural Assignments

The 1H spin system resonance assignments for bru9a were achieved using a

combination of TOCSY, 13C-HSQC, and 15N-HSQC experiments. 21 of the 24 spin

systems were identified in the amide fingerprint region of the TOCSY spectrum at 25 oC

(Figure 5.2). Ser1 and Hyp12 were not observed as expected since the N-terminus (Ser1)

exchanges rapidly and Hyp12 does not have an amide proton. The Cys2 residue was

additionally not observed at 23 oC, but was observed in the TOCSY at 0 oC. These amino

acids were later identified using the NOESY Hα (i) to HN (i+1) assignment “walk” and various hetero-nuclear spectra as detailed below.

Figure 5.2. Contour plot of a nano-NMR TOCSY spectrum o acquired on a 500 MHz spectrometer at 25 C. Spin systems are identified with the HN-Hα cross peak.

117

As seen in Figure 5.2 there is overlap in the regions of Cys6 / Arg19 / Arg23, Cys14

/ Cys21, Trp11 / Asp24, and Phe7 / Gly8 / Thr20 recorded at 25 oC. Although this is a relatively small peptide, almost 50% of the identifiable residues in the TOCSY fingerprint region have significant overlap. To assist in unambiguously assigning these residues we implemented natural abundance 15N-HSQC and 13C-HSQC- experiments.

The 15N-HSQC should identify the amide protons for the spectrum minus Ser1 and Hyp12

(Figure 5.3).

Figure 5.3. Contour plot of a nano-NMR 15N-HSQC spectrum acquired on a 500 MHz spectrometer at 25 oC. Spin systems are

identified with the HN-N cross peak.

The 15N-HSQC clearly identified over 90% of the amide protons of each spin system and provided valuable insight to overlapping spin systems within the TOCSY

118 fingerprint region. As seen in figure 5.3 the spinning side band from the rotor is observed

in the spectrum. The 15N-HSQC did not identify Ser1 and Hyp12 as expected but additionally did not identify Cys2 or Ser15. This spectrum did identify the amide Hε in

Arg8 and Arg23 plus the amide Hε in Trp11. Gly residues are easily identified in15N

spectrum as the 15N chemical shifts are considerably less than the other residues. The Tyr

residue is also easily identified because its 15N chemical shift is higher than most other

residues. A complete list of protein chemical shift statistics was obtained from the

Biological Magnetic Resonance Bank (BMRB 2008) and used as a reference. Even with a

concentration of only 190 nanomoles the signal to noise on an insensitive natural

abundance experiment was quite good. Although this is not a standard experiment in

traditional natural abundance NMR structural determinations we found it extremely

useful in positively identifying amino acid spin systems.

We acquired a 13C-HSQC to provide further identification of the spin systems

(figure 5.4)

119

Figure 5.4. Contour plot of a nano-NMR 13C-HSQC spectrum o acquired on a 500 MHz spectrometer at 25 C. CH3 / CH plots are positive (red) and CH2 plots are negative (black)

The 13C-HSQC produced a very clean spectrum containing a large amount of

cross peaks for an insensitive natural abundance experiment using only 180 nanomole of

sample, at the top and bottom of the spectrum the spinning side bands are present which are a classic trademark of nano-NMR spectra. There are several easily identifiable regions within the 13C-HSQC. Initially the methyl region is easily identified, but was also

easily identified in the TOCSY and 1D. The Gly region is unique in it contains a unique

carbon chemical shift in the 45 ppm region with the CH2 easily identifiable in the HSQC.

Figure 5.5 is an expansion of the Gly region of the 13C-HSQC. 120

Figure 5.5. Expansion of the nano-NMR 13C-HSQC (Figure 5.4) spectrum highlighting the unique Gly region. The Hα – Cα have been identified for all 5 Gly in the primary structure.

The next easily identifiable region is the aromatic side chain region. We consider this region of this 13C-HSQC to be significant because it allowed for the unambiguous assignment of the aromatic side chains (Figure 5.6)

121 .

Figure 5.6. Expansion of the nano-NMR 13C-HSQC (Figure 5.4) spectrum highlighting the aromatic side chain region. All the aromatic Hs on Phe7, Trp11, Tyr17, and Phe22 have been identified.

All H on the Trp11 aromatic rings were identified, Hε3, Hζ3, Hη2, Hζ2, Hδ, and

the Hε which was previously identified from its unique shift in the TOCSY and 1D. The

Tyr17 magnetically equivalent aromatic Hs were identified as Hδ1 / Hδ2 and Hε1 / Hε2.

The Phe residue aromatic Hs (including magnetically equivalent aromatic Hs), Hδ1 /

Hδ2, Hε1 / Hε2, and Hζ were uniquely identified for Phe7 and Phe22. The 13C-HSQC was used in conjunction with the TOCSY to identify the individual spin systems on the aromatic rings.

The NOESY Hα (i) to HN (i+1) at 0 oC was used to sequentially identify the

Ser1,Cys2, Gly3, Gly4, Ser5, Cys6, Phe7, Gly8, Gly9, Cys10, Trp11 the sequential chain

breaks here as Hyp12 has no HN. From the C-terminus side of Hyp12 the sequential

NOESY observed was Hyp12, Gly13, Cys14, Ser15, Cys16, Tyr17, Ala18, Arg19, Thr20, Cys21, 122 Phe22, Arg23, and Asp24. In the HN to HN region of the NOESY spectrum HN (i-1) to

HN (i) or HN (i+1) to HN sequential residues were observed, Gly3 – Gly4, Cys6 – Phe7,

Gly9 – Cys10, Gly13 – Cys14, Tyr17 – Ala18, Ala18 – Arg19, and Arg19 – Thr20. The entire

primary structure was assigned unique unambiguous resonances. Using the combined

homo-nuclear and hetero-nuclear NMR experiments a complete resonance assignment for

bru9a at 25 oC was achieved (Table 5.1).

123 Table 5.1 1H chemical shift values of the bru9a conotoxin Residue HN(ppm) H (ppm) H (ppm) Others Ser1 4.12 3.99, 3.95 Cys2 8.79 4.78 2.76, 2.69 Gly3 7.35 3.56, 3.13 Gly4 7.97 4.22, 3.74 Ser5 7.92 4.67 3.99, 3.83 Cys6 8.44 4.3 3.19, 3.12 Phe7 8.17 4.35 3.07, 2.98 Hδ 7.22 Hε 7.35 Hζ 7.29 Gly8 8.16 3.87, 3.82 Gly9 7.75 4.47, 3.66 Cys10 8.79 4.84 2.86, 2.65 Trp11 8.27 4.81 3.36, 3.09 Hδ 7.25 Hε1 10.27 Hε3 7.64 Hζ2 7.51 Hζ3 7.26 Hη2 7.13 Hyp12 4.59 2.31, 2.01 Hδ 3.72, 3.48 Gly13 8.91 4.37, 3.72 Cys14 8.35 5.41 3.56, 2.75 Ser15 9.40 4.72 3.80, 3.72 Cys16 8.85 4.87 3.15, 2.98 Tyr17 9.52 4.61 2.95, 2.81 Hδ 6.94 Hε 6.73 Ala18 8.70 3.78 1.19 Arg19 8.46 3.85 2.10 Hγ 1.53 Hδ 3.20 Hε 7.18 Thr20 8.15 4.72 4.00 Hγ 0.97 Cys21 8.36 5.09 2.64, 2.49 Phe22 9.48 4.83 3.16, 2.99 Hδ 7.31 Hε 7.37 Hζ 7.26 Arg23 8.46 4.55 1.82 Hγ 1.74 Hδ 3.20 Hε 7.16 Asp24 8.27 4.56 2.79, 2.70

124 This table is in good agreement with the bru9a 1H chemical shift table of the spectrum recorded at 600 MHz and 900 MHz as listed in Chapter 3.

To determine dihedral torsion angles we acquired a nano-NMR DQF-COSY. We

3 were able to identify and measure 20 JNH-Hα from a DQF-COSY experiment (Figure 5.7).

Figure 5.7. Contour plot of a DQF-COSY spectrum acquired on a 500 MHz spectrometer at 25 oC. Spin systems are identified with the HA-HN cross peak.

125 The following spin systems were identified from the DQF-COSY Cys2, Gly3, Gly4, Ser5,

Cys6, Phe7, Gly8, Gly9, Cys10, Gly13, Cys14, Ser15, Tyr17, Ala18, Arg19, Thr20, Cys21, Phe22,

Arg23, and Asp24.

A nano-NMR NOESY was acquired at 0 oC and 25 oC (Figure 5.8, Figure 5.9).

Figure 5.8. Contour plot of a nano-NMR NOESY spectrum acquired on a 500 MHz spectrometer at 0 oC.

126

Figure 5.9. Contour plot of a nano-NMR NOESY spectrum acquired on a 500 MHz spectrometer at 25 oC.

From the NOESY at 0 oC it can be seen that bru9a is a highly structured peptide. As a common characteristic of standard NMR, the NOE peaks increase as temperature decreases (Figure 3.9, 3.10). Because of the heavy overlap of bru9a on the 500 MHz spectrum the entire set of cross peaks was not determined. In chapter 3 the structure of bru9a was elucidated in detail. The NOESY spectrum in chapter 3 was obtained on a 900

MHz spectrometer which increased the spectral window and allowed for dispersion of the overlapped residues.

127 Secondary Structure

We did a comparison assessment of the secondary structure using the Chemical

Shift Index (Wishart, Sykes et al. 1992). Based on the close alignment of the Hα chemical shifts determined in chapter 3 and the Hα chemical shifts determined using nano-NMR in this chapter we expected similar results. The Hα of bru9a derived from nano-NMR was compared to the random coil chemical shift values for Hα (for residue 12 the random coil shift of Pro was used as the random coil shift for Hyp is unknown)

(Wüthrich 1986). As seen from Figure 5.10 there are large deviations from the random coil values which is indicative that bru9a is a highly structured peptide.

Figure 5.10. Comparison of the Hα chemical shift values of bru9a to the random coil chemical shift values (Pro Hα random coil chemical shift value was used for residue 12 (Hyp).

128 As mentioned the secondary structure from the Chemical Shift Index matches the

structure determined in chapter 3. The structure does not contain any helical domains as

is a multi coil/turn structure.

In this chapter discussed and proved the viability of nano-NMR as an analytical

method to characterize conotoxins of very low quantity. We have refined our techniques and methodologies to allow for the acquisition of 1D, 2D homonuclear, and 2D heteronuclear spectra of conopeptide samples at nanomole concentrations (Pflueger,

Franco et al. 2001) (Pflueger, Franco et al. 2003).

The most obvious feature of using nano-NMR is the initial gain in signal to noise

based entirely on an increase in concentration. Compared to a typical NMR sample volume (5 mm tube) of greater than 600 μl, by reducing the volume to 40 μl you

immediately gain a concentration of 15x. Even when employing a specialized 3 mm

matched Shigemi tube (Shigemi Inc, Allison Park, PA), it requires a minimum volume of

120 μl. By using low volume nano-NMR this still equates to a 3x increase in

concentration. The volume of the sample in the NMR nano-tube is also adjustable and we

have successfully recorded quality 1D spectrum of sample volumes as low as 5 μl.

The nano-NMR probe achieves higher sensitivity because the entire sample is

contained within the RF coil and spun at the magic angle, 54.7o (Keifer, Baltusis et al.

1996) (Delepierre, ProchnickaChalufour et al. 1997). By employing magic angle spinning

we realize a relative sensitivity 5x greater than a standard room temperature probe with a

spectral resolution 2x greater (Mari and Fields 2003). The increased resolution is

increased in part by eliminating the line shape distortions by spinning at the magic angle

(Barbara 1994). Extensive shimming is alleviated as the entire sample is located with the

129 RF coil saving valuable spectrometer time. Another advantage of the nano-NMR tube is greater sample security. The nano-NMR tubes are built to withstand high spinning rates

(>2500 rpm) thereby making the tubes more durable than a standard NMR tube. The nano-NMR tubes have caps or plugs which prevent evaporation and sample spillage, further protecting the integrity of the sample.

The use of nano-NMR does not come without its disadvantages. Currently, the cost of a single nano-NMR tube (rotor included) equipped for a 40 μl sample on a Varian gHX nano-probe is $ 392.60 (Wilmad-Labglass 2008). Changing of the sample is not a simplified process because the entire NMR-probe assembly must be removed from the magnet for each sample change. The use of a nano-probe ties up the instrument and prevents it from being utilized as a general use characterization technique. This is of extreme consideration in the academic environment as the NMR instruments are typically used cross-platform and NMR time is a valuable commodity. The use of high spinning rates has been observed to cause some spinning side band artifacts in the acquired spectra. Nano-NMR sample preparations require meticulous protocols based on the low sample volumes. On a typical 5 mm sample preparation a 5 μl loss of sample would be negligible, but in a nano sample, just a 5 μl loss of sample equates to losing over 12% of the sample. Additionally, this low volume presents challenges in pH adjustments and requires specialized equipment. To further complicate matters, to spin the sample there is a need for an extended supply of extremely dry compressed air or a pressurized inert gas.

The use of cryogenic NMR probe technology has rapidly gained popularity due to gains in signal to noise sensitivity. The use of these cryo-probes has not been particularly targeted for low concentration, natural abundance experiments, but has been well suited

130 for multi-nucleotide labeled 3D and 4D experiments with the major gains being savings in spectrometer time (Crouch, Llanos et al. 2001). From our experience, nano-NMR is the premier NMR method to structurally characterize high value, low concentration, and natural abundance conopeptides.

131

REFERENCES

Arias, H. R. and M. P. Blanton (2000). " -conotoxins." Int J Biochem Cell Biol 32(10):

1017-28.

Baldomero, M., B. M. Olivera, et al. (1988). Marine Snail Venoms. Handbook of Natural

Toxins. A. TU, Marcel Dekker, Inc.

Barbara, T. M. (1994). "Cylindrical demagnetization fields and microprobe design in

high-resolution NMR." Journal of Magnetic Resonance Series A 109(2): 265-269.

BMRB (2008). Biological Magnetic Resonance Bank, A Repository for Data from NMR

Spectroscopy on Proteins, Peptides, Nucleic Acids, and other Biomolecules.,

University of Wisconsin.

Callihan, D., J. West, et al. (1996). "Simple, distortion-free homonuclear spectra of

peptides and nucleic acids in water using excitation sculpting." J Magn Reson B

112(1): 82-5.

Crouch, R. C., W. Llanos, et al. (2001). "Applications of cryogenic NMR probe

technology to long-range H1-N15 2D NMR studies at natural abundance."

Magnetic Resonance in Chemistry 39(9): 555-558.

Delepierre, M., A. Prochnicka-Chalufour, et al. (1997). "A novel potassium channel

blocking toxin from the scorpion Pandinus imperator: A H-1 NMR analysis using

a nano-NMR probe." Biochemistry 36(9): 2649-2658.

Goddard T. D., Kneller D. G. (2008) Sparky 3, University of California, San Francisco.

132 Hill, J. M., P. F. Alewood, et al. (1997). "Solution structure of the sodium channel

antagonist conotoxin GS: a new molecular caliper for probing sodium channel

geometry." Structure 5(4): 571-83.

Kaas, Q., J. C. Westermann, et al. (2008). "ConoServer, a database for conopeptide

sequences and structures." Bioinformatics 24(3): 445-6.

Keifer, P. A., L. Baltusis, et al. (1996). "A comparison of NMR spectra obtained for

solid-phase-synthesis resins using conventional high-resolution, magic-angle-

spinning, and high-resolution magic-angle-spinning probes." Journal of Magnetic

Resonance Series A 119(1): 65-75.

Lange, A., K. Giller, et al. (2006). "Toxin-induced conformational changes in a

potassium channel revealed by solid-state NMR." Nature 440(7086): 959-62.

Mari, F. and G. B. Fields (2003). "Conopeptides: Unique pharmacological agents that

challenge current peptide methodologies." Chimica Oggi 21: 43-48.

Marx, U. C., N. L. Daly, et al. (2006). "NMR of conotoxins: structural features and an

analysis of chemical shifts of post-translationally modified amino acids." Magn

Reson Chem 44 Spec No: S41-50.

McIntosh, J. M., B. M. Olivera, et al. (1999). "Conus peptides as probes for ion

channels." Methods Enzymol 294: 605-24.

Pflueger, F. C., A. Franco, et al. (2003). "Paper presented at the Keystone Symposia,

Frontiers of NMR in Molecular Biology VII."

Pflueger, F. C., A. Franco, et al. (2001). Paper presented at the 42nd ENC.

Piotto, M., V. Saudek, et al. (1992). "Gradient-tailored excitation for single-quantum

NMR spectroscopy of aqueous solutions." J Biomol NMR 2(6): 661-5.

133 Terlau, H., K. J. Shon, et al. (1996). "Strategy for rapid immobilization of prey by a fish-

hunting marine snail." Nature 381(6578): 148-51.

Wilmad-Labglass. (2008). "Varian gHx NanoProbe tubes." Retrieved 10/12/2008, from

http://www.wilmad-labglass.com/group/2132.

Wishart, D. S., B. D. Sykes, et al. (1992). "The chemical shift index: a fast and simple

method for the assignment of protein secondary structure through NMR

spectroscopy." Biochemistry 31(6): 1647-51.

Wüthrich, K. (1986). NMR of proteins and nucleic acids. New York, Wiley.

134

CHAPTER 6

Reproduced with permission from the Journal of the American Chemical Society. 2005

Polypeptide Chains Containing D-γ-Hydroxyvaline

Katarzyna Pisarewicz, David Mora, Fred C. Pflueger, Gregg B. Fields, and Frank

Marí

Contribution from the Department of Chemistry & Biochemistry and Center of

Excellence in Biomedical & Marine Biotechnology, Florida Atlantic University, 777

Glades Road, Boca Raton, Florida 33431

ABSTRACT

Life has an unexplained and distinct L-homochirality. Proteins typically

incorporate only L-amino acids into their sequences. In the present study, D-Val and D-γ-

hydroxyvaline (D-Hyv; V*) have been found within ribosomally expressed polypeptide

chains. Four conopeptides were initially isolated, gld-V*/gld-V*' from the venom of

Conus gladiator and mus-V*/mus-V*' from the venom of Conus mus. Their complete sequences (gld-V*/gld-V*' = Ala-Hyp-Ala-Asn-Ser-D-Hyv-Trp-Ser and mus-V*/mus-

V*' = Ser-Hyp-Ala-Asn-Ser-D-Hyv-Trp-Ser) were determined by a combination of nano/pico-NMR and MS/MS methods. The amino acid triad that contains the γ- hydroxylated residue, Ser-D-Hyv-Trp, is a novel structural motif that is

135 stabilized by specific interactions between the D-amino acid and its neighboring L-

counterparts. These interactions inhibit lactonization, a peptide backbone scission process

that would normally be initiated by γ-hydroxylated residues. Conopeptides possessing the

Ser-D-Hyv-Trp motif have been termed γ-hydroxyconophans. We have also isolated

analogous conopeptides (gld-V and mus-V) containing D-Val instead of D-Hyv; these are

termed conophans. γ-Hydroxyconophans and conophans are particularly atypical because

(i) they are not constrained as most conopeptides, (ii) they are extremely short in length,

(iii) they have a high content of hydroxylated residues, and (iv) their sequences have no close match with other peptides in sequence databases. Their modifications appear to be part of a novel hyperhydroxylation mechanism found within the venom of cone snails that enhances neuronal targeting. The finding of D-Val and D-Hyv within this family of peptides suggests the existence of a corresponding D-stereospecific enzyme capable of

D-Val oxidation.

INTRODUCTION

Modification of L-amino acids within polypeptide chains often delineates protein function and is a ubiquitous biochemical process (Krishna and Wold 1998). Peptides found in the venom of predatory marine mollusks belonging to the genus Conus (cone snails) possess diverse modifications that modulate their activities. The Conus venom is a complex mixture of peptides (conopeptides) that elicit a wide range of neurophysiological responses (Jones and Bulaj 2000; Newcomb and Miljanich 2002;

Mari and Fields 2003; Terlau and Olivera 2004). Several conopeptides have been shown to be valuable therapeutic agents for the treatment of a variety of neurologically related conditions (David J. Adams 1999; Shen, Layer et al. 2000; McIntosh and Jones 2001; 136 Heading 2002). The venom apparatus of cone snails consists of a muscular bulb used to

propel the venom and is coupled to a long duct (where the venom is produced), which is

ultimately connected to a specialized radular tooth that serves as both a harpoon and

disposable hypodermic needle. The polypeptide components of the venom are produced

by epithelial cells cells (Marshall, Kelley et al. 2002) in the venom duct and are

expressed in the ribosome as protein precursors that subsequently undergo

posttranslational modifications and proteolytic cleavage to form the mature conopeptide

(Milne, Abbenante et al. 2003). Conopeptides contain multiple combinations of modified

amino acids, such as cystines, hydroxyproline, γ-carboxyglutamate, Br-Trp, D-Trp, D-

Leu, pyro-Glu, glycosylated Ser/Thr, and sulfated Tyr (Craig, Bandyopadhyay et al.

1999; Craig, Norberg et al. 1999; Craig, Park et al. 2001). These modifications provide stability and exquisite specificity toward neuronal targets (Myers, Cruz et al. 1993; Craig,

Bandyopadhyay et al. 1999; McIntosh, Olivera et al. 1999), aiding these marine snails in

prey capture.

Modification of polypeptide chains by epimerization of standard L-amino acids to

produce their D-counterparts is rarely observed (Mitchell and Smith 2003). D-Amino

acids were believed to be produced only by prokaryotes, with most D-amino acid

containing polypeptide chains being of nonribosomal origin and assembled by enzymatic

processes within unicellular organisms. A few examples of D-amino acids occurring

within bioactive peptides and proteins of multicellular organisms have been reported,

thus challenging our understanding of the homochirality of life (Yang, Zheng et al. 2003).

D-Amino acids are found in position 2 of frog skin opionoid peptides such as dermorfine

(Fujii 2002), in neuropeptides of land mollusks such as achatin (Fujii 2002), and even in

137 mammalians, as revealed in the 39-residue C-type natriuretic peptide from platypus venom (Torres, Menz et al. 2002). D-Phe is found in position 3 of the 72-residue

crustacean hyperglycemic hormone (Soyez, Toullec et al. 2000). D-Trp (and sometimes

D-Leu) is found as the second residue after the first Cys in contryphans (Jimenez, Olivera et al. 1996; Jacobsen, Jimenez et al. 1998; Jimenez, Watkins et al. 2001). The modifications of L- to D-amino acids are determinants of stability and potency. For example, D-Ser46 in the 48-residue funnel-web spider venom ω-agatoxin IVB provides more resistance to the major venom protease and is a more potent blocker of the P-type voltage-sensitive calcium channels than its L-Ser46 (ω-agatoxin IVC) counterpart (Heck,

Siok et al. 1994). More recently, a 46-residue conotoxin belonging to the I-gene

superfamily has been found to possess D-Phe in position 44, which has been determined

to be essential for the neuroexcitatory properties of the conopeptide (Buczek, Yoshikami

et al. 2005). In general, D-amino acids within polypeptide chains can provide unique

structural determinants that allow the stabilization of turns (Imperiali, Fisher et al. 1992)

or resistance to enzymatic breakdown (Heck, Siok et al. 1994).

γ-Hydroxylation of non-Pro amino acids is an even rarer process than

epimerization, since a hydroxyl group in the γ-position of any amino acid (except Pro)

could undergo nucleophilic attack at the contiguous peptide bond to form a stable five- membered ring lactone, (Scheme 6.1).

138

Scheme 6.1

Nonetheless, γ-hydroxyarginine has been found as part of the sequence of

polyphenolic proteins that form the adhesive plaques of marine mussel species (Papov,

Diamond et al. 1995). The presence of γ-hydroxyarginine provides trypsin resistance to

mussel glue proteins. γ-Hydroxylysine (γ-Hyk) has been reported within the sequence of

crytonomad algae biliproteins (Sidler, Kumpf et al. 1985); however, the role of γ-Hyk in protein function has not been defined. The oxidation of Leu to produce hydroxyleucine

(presumably in either the δ- or γ-position) has been described as an unusual posttranslational modification present in unstable forms of hemoglobin associated with patients afflicted with hemolytic anemia (Brennan, Shaw et al. 1992; Brennan, Shaw et al. 1993). The significance of this modification has not been established. γ-

Hydroxyproline (Hyp) is commonly found in collagen and is vital for collagen structural stability (Harding and Crabbe 1992; Perret, Merle et al. 2001). The unique cyclic nature of Pro impedes lactonization. Nonproteinogenous γ-hydroxylated amino acids have been found within enzymatically produced cyclic peptides (Shoji and Hinoo 1975; McGahren,

Morton et al. 1977; Terui, Nishikawa et al. 1990; Morita, Kayashita et al. 1997).

The present study has identified D-Val and D-γ-hydroxyvaline (D-Hyv; V*) within ribosomally expressed polypeptide chains. D-Hyv was found within the sequences 139 of four conopeptides from the venom of Conus gladiator (gld-V* and gld-V*') and Conus

mus (mus-V* and mus-V*'). We have also isolated analogous peptides that contain D-Val

(gld-V and mus-V). Hyv was first described as a novel amino acid isolated from plants

(Pollard, Sondheimer et al. 1958). Hyv is an unexpected modified residue in proteins and

peptides, as its hydroxyl group could readily cleave a peptide bond by intraresidue cyclization to form a lactone (Scheme 6.1). The stability of Hyv within conopeptides has

been explained by the D-configuration at the α-carbon in conjunction with specific

interactions with its surrounding L-amino acids. The doubly modified D-Hyv along with

its neighboring residues defines a novel structural motif that characterizes a new family of conopeptides termed γ-hydroxyconophans.

RESULTS AND DISCUSSION

We initially isolated three unusual conopeptides, gld-V*, gld-V*', and gld-V

(Figure 6.1A), from the venom of Conus gladiator (species code gld), a cone snail species that inhabits the tropical Eastern Pacific region and preys upon worms.

Concurrent with isolation of these conopeptides, related conopeptides mus-V*, mus-V*',

and mus-V were isolated from Conus mus (species code mus), a cone snail species

related to C. gladiator that inhabits the Western Atlantic region (Figure 6.1B). The gld

conopeptides were isolated in nanomolar quantities, whereas the mus conopeptides were isolated in picomolar quantities. Nano/pico-NMR techniques (Barbara 1994; Barbara and

Bronnimann 1999) allowed the acquisition of their spectra and revealed almost identical compositions for these octapeptides, including an unusual amino acid for gld-V* and gld-

V*' (Figure 6.2), whereas gld-V showed Val in its place (Figure 6.4). The mass spectra of

140 gld-V*/gld-V*' and gld-V gave molecular ions of 863.3 and 847.3 Da, respectively. gld-

V* and gld-V*' had the same covalent structures. The mus octapeptides revealed sequence information identical to their gld counterparts, except that their molecular weights were shifted by 16 Da. Combined Edman degradation sequencing, MS/MS

(Figure 6.3), and NMR analyses (Figure 6.2) revealed the structures (Chart 6.1) of these octa-conopeptides [hydroxylated amino acids are shown in blue, including the modified amino acids (O and V*)].

141

Figure 6.1. Conopeptide isolation from venom of (A) C. gladiator and (B) C. mus. The Conus venom was fractionated using (top and middle) SE- HPLC (Superdex-30, buffer = 0.1 M NH4HCO3). The elution SE-HPLC profiles are shown at λ= 220 (top) and 280 nm (middle), respectively. The arrows indicate the selected fractions of Trp-containing gld and mus peptides. The Trp-containing fractions were further separated using (bottom) RP-HPLC (Vydac C18, H2O/60% CH3CN linear gradient over 100 min with 0.1% TFA).

142

Figure 6.2. NMR spectra of the γ-hydroxyconophans from C. gladiator (A gld-V* and B gld-V*') and C. mus (C mus-V* and D mus-V*'). A and B show the 1D proton spectra along with its corresponding 2D-TOCSY spectrum of 35 nmol of gld-V* and 22 nmol of gld-V*', respectively. Part A was recorded at 25 C using a gHX HR-MAS probe. The NMR assignments of the γ-hydroxyvaline (i.e., for gld-V*: HN: d 7.99, 8 Hz; αH: 4.45, m; γCH2 m 3.15; βCH m 1.89, 6.9 Hz; γCH3: d 0.52, 7.1 Hz) correlated well with the reported values of the synthetic amino acid (Easton and Merrett 1997). C and D show the 1D proton spectra of picomolar quantities of mus-V* and mus-V*', respectively. Spectra in B, C, and D were recorded using 3 mm tubes (see Experimental Section).

143 NMR analysis, along with MS/MS spectra of gld-V*, gld-V*', mus-V*, and mus-

V*', revealed the presence of Hyv in these conopeptides, representing the first examples

of this amino acid found within polypeptide chains. A shielded doublet (~0.5 ppm) that

corresponds to a methyl group appears in all these spectra. The 2D-TOCSY spectra of the

gld-V*/gld-V*' pair indicates that this shielded doublet is part of a spin system that

corresponds to Hyv. The presence of resonances at δ= 0.52, 1.89, 3.15, 4.45, and 7.99 ppm defines a unique spin system that matches the reported values of Hyv (Figure 6.2).

This assignment is supported by the fragmentation pattern in the MS/MS spectra, where the mass difference between the b6 and the b5 fragment also corresponds to Hyv (Figure

6.3). Other internal fragments, such as OANSV* and NSV*, provide additional support

for the presence of Hyv in these peptides.

144

Figure 6.3. MALDI-MS/MS of the γ-hydroxyconophans from C. gladiator: (A) gld-V* and (B) gld-V*'; and C. mus: (C) mus-V* and (D) mus-V*'. These spectra were recorded using the AB Q-TOF instrument (see Experimental Section). Assignments of the b fragments and others were carried out using standard procedures (Siethoff, Lohaus et al. 1999).

145

Figure 6.4. NMR spectra of the conophan from C. gladiator (A and B gld-

V) and its corresponding synthetic peptides containing L-Val-L-Trp (C), L- Val-D-Trp (D), and D-Val-L-Trp (E). Part A shows the 1H NMR 2D-

TOCSY spectrum of 21 nmol of native gld-V isolated from the venom of C. gladiator. B, C, D, and E show the 0.2-4.5 ppm region (αH and side- chains) of the 1D proton spectra of native gld-V (B), synthetic gld-V incorporating D-Val-L-Trp (C), synthetic gld-V incorporating L-Val-D-

Trp (D), and synthetic gld-V using L-Val-L-Trp (E). A and B were recorded at 25 oC using a gHX HR-MAS probe. C, D, and E were recorded using 3 mm tubes in the gHCX probe. The arrows in B indicate resonances of interest that substantiate (by comparison with the synthetic analogue) the chirality assignment of Val-6 in gld-V.

146

Chart 6.1

The configuration of the α-carbon in Hyv was determined by detailed analysis of

the proton chemical shifts and the splitting patterns caused by proton-proton coupling in the 1H NMR spectra. The chemical shifts and splitting patterns of the βH protons within

the Ser-Val-Trp triad are quite sensitive to the different chiralities of their respective α-

carbons (Figure 6.4). Direct comparison of the NMR splitting patterns and chemical

shifts between the gld-V*, gld-V*', and gld-V indicates that the stereochemistry of the α- carbon at residue six (Hyv for gld-V* and gld-V*' and Val for gld-V) has been preserved

upon hydroxylation. This is particularly noticeable for the βH signal of Trp-7 in gld-V*, gld-V*', and gld-V (Figures 6.2 and 6.44), since their splitting patterns are most sensitive to the different stereochemistries of the different analogues of gld-V. Splitting patterns

for the synthetic analogues of gld-V with L-configurations in residue six are completely

147 unrelated to the one observed in the native conopeptides. By way of contrast, these

patterns are identical to the D-Val synthetic analogue and are the same in all native

conopeptides (gld-V*, gld-V*', and gld-V). The NMR findings are consistent with the

MS/MS spectra of gld-V and its synthetic analogues (Figure 6.5), as the fragmentation of the native conopeptide corresponds only to the fragmentation observed in the synthetic analogue with D-Val. Since the synthetic analogues of gld-V clearly established the D-

configuration of residue six in gld-V, the absolute configuration of Hyv in gld-V* and gld-V*' also corresponds to the D-amino acid.

148

Figure 6.5. MS/MS of the conophan from C. gladiator, gld-V: (A) native gld- V; (B) synthetic gld-V incorporating L-Val-D-Trp; (C) synthetic gld-V incorporating D-Val-L-Trp; and (D) synthetic gld-V incorporating L-Val-L- Trp. These spectra were recorded using the AB Q-Star XL Q-TOF instrument (see Experimental Section). The stereochemistry of the Val residue was determined by comparing the NMR (Figure 6.4) and MS/MS spectra shown here of the native gld-V with its synthetic peptide analogues. The influence of the chirality of Val-6 on the MS/MS fragmentation patterns of these peptides supports the NMR evidence shown in Figure 6.4.

The gld-V*/gld-V*' and mus-V*/mus-V*' pairs have the same covalent structures, respectively. However, their chromatographic behavior revealed differences in hydrophobicity in a temperature-independent fashion. In principle, these differences might be attributed to cis/trans isomerism of the peptide bond involving Hyp in residue 2, 149 as suggested by NMR evidence in other related conopeptides (Pallaghy, He et al. 2000).

Ultraviolet resonance Raman spectroscopy suggested that two conformational states

within conopeptides could be attributed to the differences of the χdihedral angles of the

Trp within their sequence (Jimenez, Watkins et al. 2001). However, in these cases

temperature dependency of the distribution of the conformers has been observed

(Jacobsen, Jimenez et al. 1998). Closer analysis of the MS/MS data of the gld and mus γ-

hydroxyconophans reveals that the fragmentation patterns within the pairs differ in the

intensity of the b6 fragment (Figure 6.3), which suggests structural differences within the

Hyv residue. This difference is more significant when utilizing an ESI-ion trap

instrument to record the MS/MS spectra (data not shown, see Experimental Section).

Furthermore, the largest chemical shift differences between the gld-V* and gld-V*' are within the resonances of α-CH and γ-CH2 proton of Hyv (Figure 6.2). Hyv bares a chiral center at the β-carbon; therefore, gld-V* and gld-V*' are likely to be diastereomers, epimeric at the β-carbon of Hyv, as reflected in the differences in the NMR chemical

shifts of the groups attached to the β-position of Hyv. There are slight differences

reported for the chemical shifts of the 2S,3S and 2S,3R diastereomers of free Hyv;

(Easton and Merrett 1997) however, these differences cannot be used to determine the absolute configuration of Hyv within these diastereomeric conopeptides.

The γ-hydroxylation of any amino acid (except for γ-Hyp) would be unexpected, as it introduces susceptibility to lactonization. This susceptibility would explain why the

Edman degradation analysis of gld-V* yielded its sequence only up to the residue preceding Hyv (data not shown). The basic conditions (pH ≥ 9 at 40-55 oC) required for

the coupling reaction of phenylisothiocyanate to the free N-terminal amino group of a

150 polypeptide chain, or the strongly acidic conditions necessary to produce the PTH-amino acid (Lottspeisch, Houthaeve et al. 1999), might catalyze the intraresidue cyclization

process leading to early termination of the complete peptide sequence. Intraresidue

cyclization is possible in proteins/peptides containing γ-hydroxyarginine (Papov,

Diamond et al. 1995) and γ-hydroxylysine (Sidler, Kumpf et al. 1985). Both amino acids

have been reported as part of the sequences of proteins. However, these are positively

charged amino acids and their hydroxyl groups are secondary alcohols. These two factors

diminish the nucleophilicity of the hydroxyl group and thus deter lactonization.

Furthermore, in the case of γ-hydroxyarginine in the mussel adhesive protein Mefp-3, it

has been suggested that γ-hydroxyarginine is capable of hydrogen bonding with Dopa,

the next residue in the protein sequence. The interresidue interactions of this dyad of

contiguously modified amino acids represent a stable structural motif that confers Mefp-3

with resistance to proteases and appears to be involved in the molecular interactions

necessary for adhesion (Papov, Diamond et al. 1995).

What is the basis for the stability of Hyv within gld and mus conopeptides? The

distinctive allowed Ramachandran space for D-amino acids could place contiguous

residues in close proximity (Mitchell and Smith 2003). This is confirmed by X-ray and

NMR analyses of polypeptide chains that contain D-amino acids, such as dermophins and

achatins, where all side-chains of the Xaai-1-D-Xaai-Xaai+1 motif are on the same face of

the polypeptide chain because of the central D-amino acid configuration Kamatani,

Minakata et al. 1990; Ishida, In et al. 1992; Slabicki, Potrzebowski et al. 2004). In the

case of the γ-hydroxyconophans gld-V*/gld-V*' and mus-V*/mus-V*', structural stability can be explained by specific interactions of the D-Hyv with it neighboring L-amino acids.

151 Trp provides (i) steric hindrance and (ii) a stabilizing van der Waals interaction between the D-Hyv methyl group and the Trp aromatic ring that impede cyclization. Evidence of the van der Waals interaction is observed in the chemical shift of the Hyv methyl group, which is shielded (δ = 0.52 ppm) in the same manner as methyl groups are shifted by aromatic residues within folded proteins (Storch, Grinstead et al. 1999). X-ray data on the

Tyr-D-Ala-Phe triad found in dermophin indicates the presence of a C-H···π interaction between the D-Ala methyl group and the Tyr aromatic ring; this is a defining feature of the Tyr-D-Ala-Phe structural topology. The same interaction appears to be present in the gld-V*/ gld-V*' and mus-V*/mus-V*' Ser-D-Hyv-Trp triad, as indicated by the NMR spectra described above and rationalized by molecular modeling (Figure 6.6). Stabilizing interactions within the Ser-D-Hyv-Trp triad can also be provided by hydrogen bonding between Ser and D-Hyv side-chains due to their proximity. Overall, the D-Hyv is held in a "locked" conformation by the C-H···π interaction with Trp on one side and an H-bond with Ser on the other side, explaining the inability of Hyv to undergo lactonization.

152

Figure 6.6. Molecular model of γ-hydroxyconophan structural motif H3CC(O)-Ser-D-Hyv-Trp-NH2. This model was constructed on the basis of the X-ray structure of the dermorphin Tyr-D-Ala-Phe "message sequence" and illustrates the proximity of the γ-methyl of D-Hyv to the Trp side-chain and the hydroxyl group of Hyv to the Ser side-chain, which are shown as being hydrogen bonded. This proximity enables specific interactions between the side-chains of all these residues (see text). Evidence of the D- Hyv-Trp side-chain interaction is provided by the strongly shielded chemical shift of the D-Hyv γ-methyl group (see Figures 6.2 and 6.4). Trp also provides steric impediment to the lactonization process and aids the stability of polypeptides chains that include this structural motif.

The amino acid triad that contains the γ-hydroxylated amino acid, Ser-D-γHyv-

Trp, is a novel structural motif that defines a new class of conopeptides termed γ-

hydroxyconophans. The corresponding analogous peptides that contain just D-Val, such

as gld-V, are termed conophans. The initial finding of γ-hydroxyconophans within the venom of Conus gladiator led to their isolation from C. mus. C. gladiator and C. mus are

153 closely related cone snail species (Duda and Palumbi 1999). These species evolved separately from a common ancestor as the Isthmus of Panama separated the Atlantic and

Pacific Oceans 3-3.5 million years ago. It is remarkable that these two Conus species, which nowadays inhabit different oceans, bear the same biochemical imprint within their venom through a novel set of posttranslational modifications. While this is an indication of their common ancestral origins, γ-hydroxyconophans, conophans, and peptides that contain related structural motifs are likely to be found in other Conus species. Here, the difference between the gld conophans and their mus counterparts is Ser in residue 1 in C. mus as opposed to Ala in C. gladiator. This appears to be part of the hyperhydroxylation strategy used by cone snails to optimize their venom efficacy.

It is noteworthy that the sequences of these conopeptides have an extremely high number of hydroxylated residues. Perhaps, just as in the case of collagen, hydroxylation is the preferred strategy used by the cone snails to increase hydrogen-bonding capabilities. However, in the Conus case, hydrogen bonding is directed toward increasing binding strength and selectivity toward their neuronal targets. Polyhydroxylation is a recurrent feature in natural products of marine and land origins (Danieli and Riva 1994;

Zeng, Su et al. 1999; Cui, Wang et al. 2000; Watson, Fleet et al. 2001). Polyhydroxylated compounds, ranging from Taxol to Dermostatin, rely on the hydrogen-bonding capabilities of their hydroxyl groups to interact with their molecular targets. In the case of most polyhydroxylated natural products, their complex molecular scaffolds are the product of an intricate multienzymatic biosynthetic pathway that incorporates diverse metabolites in manners unique to the organisms that produce the compounds. On the other hand, cone snails have relied on a universal mechanism for protein synthesis and

154 modification for the production of small molecule-like structural scaffolds that fulfill the

specialized task of targeting specific neuronal receptors. These versatile peptide

engineers have developed a wide range of protein processing schemes that allow them to

efficiently carry out multiple posttranslational modifications on polypeptide chains across

the many conopeptide families. While this process is reminiscent of the biosynthetic production of complex natural products, the cone snails are using a reduced set of

enzymes capable of modifying a wide range of ribosomally expressed polypeptide chains.

This strategy is aimed at enhancing the molecular complexity and diversity of the venom.

The need for the development of such biochemically diverse venom is likely to be an

evolutionary adaptation designed to compensate for the lack of mobility of cone snails

when compared to other marine predators. Within this scheme, the epimerization and

subsequent hydroxylation of Val provides further diversity to the venom by adding a new

protein scaffold that is so far unique to Conus. However, just as other posttranslational

modifications found in Conus venom were previously described in other organisms

(Olivera 2002; Mari and Fields 2003), it will not be surprising to find the γ-

hydroxyconophan scaffold in other organisms. In addition to the unprecedented presence

of D-Hyv in their sequence, these γ-hydroxyconophans are unusual because (i) they are

linear conopeptides and not constrained like the conotoxin and contryphan families, (ii)

they are extremely short in length, (iii) they have a high content of hydroxylated residues,

and (iv) their primary structure has no close match in the sequence databases.

The epimerization of Val by cone snails has produced the first example of D-Val

within a ribosomally expressed polypeptide chain. Most epimerizations found in small

linear peptides occur near the N-terminal and preferentially at the second position. The

155 D-Val in these conophans is at the third amino acid from the C-terminal, the same

relative position as in the larger disulfide-constrained ω-agatoxin (Heck, Siok et al. 1994)

and the r11a I-superfamiliy conotoxin (Buczek, Yoshikami et al. 2005), which have 48

and 46 residues, respectively. Apparently, the epimerization has a strong preference at

this position near the C-terminal regardless of the nature of the amino acid (D-Ser in ω-

agatoxin, D-Phe in the r11a conotoxin, or D-Val in conophans) or size and nature of the

expressed protein. In fact, it is likely that the two-base enzymatic mechanism proposed

for the epimerization of D-Ser in ω-agatoxin is in effect in all these cases, as the

epimerase in the funnel-web spider is also know to epimerize other amino acids, such as

Ala, Cys, and O-methylserine (Heck, Faraci et al. 1996). However, the substrate for this

epimerase has a recognition site Leu-Xaa-Phe-Ala, observed neither in the r11 conotoxin

nor in the conophans. Furthermore, the spider epimerase is capable of converting Xaa in

small peptides at several positions within the polypeptide chain (Heck, Faraci et al.

1996). Therefore, it is likely that different epimerases with distinct specificities are

operating in each of these cases.

The presence of D-Hyv in gld-V*/gld-V*', as opposed to D-Val in gld-V, suggests the existence of an enzyme capable of D-Val oxidation. This putative enzyme could be using gld-V, or its precursor protein, as a substrate to modify D-Val and

generate the D-Hyv form of the toxin. This process would be analogous to Glu γ- carboxylation of certain conopeptides, which require the action of a specific carboxylase on the precursor form of the peptide (Bandyopadhyay, Colledge et al. 1998; Walker,

Shetty et al. 2001; Bandyopadhyay, Garrett et al. 2002; Czerwiec, Begley et al. 2002) to

156 produce conantokins and related Gla-containing conopeptides. The isolation and

identification of a hydroxylase with D-amino acid specificity is under investigation.

EXPERIMENTAL SECTION

Peptide Isolation.

Specimens of Conus gladiator (species code gld) were collected from several

locations off the Pacific coast of Costa Rica. The venom was dissected from the venom

ducts, pooled, lyophilized (~50 mg from 47 snails), and initially fractionated using size

exclusion-HPLC on a Pharmacia Superdex-30 column (2.5 × 100 cm) with elution by 0.1

M NH4HCO3 buffer at a flow rate of 1.5 mL/min (Figure 6.1A). The column eluent was

monitored on a PDA detector (TSP SM-5100) at λ= 220 and 280 nm. The material in the

major peak (gld_07) in the λ= 280-detected chromatogram was further separated using an

RP-HPLC Vydac C18 column (10 × 250 mm, 5 μm, 300 Å) eluted with a linear gradient

of H2O/60% CH3CN over 100 min (Figure 6.1B). 0.1% TFA was used as ion-pairing

reagent. Three peptide fractions were separated (gld-V*, gld-V*', and gld-V) and

subsequently analyzed by MS and NMR. Similarly, specimens of Conus mus were

collected off the Florida Keys (Plantation, Monroe County); 12 mg of crude venom was

extracted and processed as described for C. gladiator.

Peptide Sequencing and Synthesis.

Sequencing was carried out by Edman degradation chemistry on an Applied

Biosystems (AB) Procise cLC and Procise 491A instruments. Peptide synthesis was

performed on an AB 433A peptide synthesizer using Fmoc chemistry (Fields, Lauer-

Fields et al. 2001). Peptide-resin cleavage utilized appropriate scavengers (King, Fields et

157 al. 1990; Fields and Fields 1993) to avoid Trp modification. Cleaved peptides were purified by RP-HPLC as described above.

Mass Spectrometry.

The MS/MS spectra of all conopeptides were obtained either on an AB Q-Star XL

Q-TOF spectrometer equipped with an oMALDI-2 or on a Micromass Q-TOF micro instrument equipped with a nanospray source. Additional MS experiments were carried out using a Finningan LCQ-Deca instrument. Samples (~1 pmol) analyzed using the Q-

TOF instrument were desalted using a C18 ZipTip and introduced with a nanospray ion source (Wilm and Mann 1996; Wilm, Shevchenko et al. 1996). Glu-fibrinogen, m/z =

785.85 doubly charged, was used as an internal standard. Approximately 10 pmol of sample was applied for analyses using the LCQ instrument. Samples were analyzed by flow injection using 30%ACN/0.1% acetic acid as a carrier.

NMR Spectroscopy.

NMR spectra were acquired on a Varian Inova 500 MHz spectrometer equipped with three rf channels, pulse field gradients, and waveform generators. Initially, 1D- and

2D-TOCSY spectra were recorded using a gHX HR-MAS probe (Barbara 1994; Barbara and Bronnimann 1999) for 1 nmol of gld-V* in 35 μL. Larger sample quantities (20-35 nmol of the gld peptides) were analyzed using 3 mm sample tubes in 130 μL of NMR solution in a 5 mm gHCX triple resonance probe. 1D spectra were acquired using 512 scans, whereas 2D spectra were acquired using 96 increments in t1 with 256 scans per increment in a phase-sensitive mode. 2D spectra were processed using linear predictions in t1 to 1024 points and transformed to final size of 2k × 2k. The 1D spectra of picomolar amounts of the mus conopeptides were acquired overnight using 3 mm sample tubes in

158 130 μL of NMR solution in a 5 mm gHCX triple resonance probe. All spectra were

o recorded at 25 and 0 C in an NMR solution that consisted of 90% H2O/10% D2O using

TSP as an internal standard. The pH for this solution was adjusted to 3.6 using 0.01 M

solutions of HCl and NaOH and a Phoenix micro-pH probe. Water suppression was achieved using Watergate (Piotto, Saudek et al. 1992) and Excitation Sculpting (Callihan,

West et al. 1996) for the 2D experiments and WET (Smallcombe, Patt et al. 1995) and presaturation for the 1D 1H spectra. The resonance assignments were carried out using standard biomolecular NMR procedures (Wüthrich 1986).

Molecular Modeling.

Molecular models were built based on the X-ray structure of the dermophin

"message sequence" Tyr-D-Ala-Phe (Slabicki, Potrzebowski et al. 2004). The extended conformation consistent with the NMR data was used for the initial model and optimized to self-consistency by the MMX force field as previously described (Mari, Lahti et al.

1992). The structural motif that characterizes the gld/mus γ-hydroxyconophan family of conopeptides [Ser-D-Hyv-Trp] was capped using an acetyl group at the N-terminus and an amide group at the C-terminus to simulate a protein-like environment.

Acknowledgment

We thank J. Lauer-Fields, Mare Cudic, and Jose Rivera-Ortiz for synthesis of peptides; A. Franco for collection of specimens; M. Crawford for peptide sequencing; and C. Byrdwell and K. Stone for mass spectrometry analysis. We thank the government of the Republic of Costa Rica for providing the permits to collect the necessary specimens. This work was supported by the Florida Sea Grant College Program (R/LR-

MB-18) and the NIH (GM 066004 to F.M., CA 77402 to G.B.F.). This work is

159 contribution P200503 from the Center of Excellence in Biomedical and Marine

Biotechnology.

160

REFERENCES

Bandyopadhyay, P. K., C. J. Colledge, et al. (1998). "Conantokin-G precursor and its role

in gamma-carboxylation by a vitamin K-dependent carboxylase from a Conus

snail." J Biol Chem 273(10): 5447-50.

Bandyopadhyay, P. K., J. E. Garrett, et al. (2002). "-Glutamyl carboxylation: An

extracellular posttranslational modification that antedates the divergence of

molluscs, arthropods, and chordates." Proc Natl Acad Sci U S A 99(3): 1264-9.

Barbara, T. M. (1994). "Cylindrical demagnetization fields and microprobe design in

high-resolution NMR." Journal of Magnetic Resonance Series A 109(2): 265-269.

Barbara, T. M. (1994). "Cylindrical Demagnetization Fields and Microprobe Design in

High-Resolution NMR." Journal of Magnetic Resonance, Series A 109(2): 265-

269.

Barbara, T. M. and C. E. Bronnimann (1999). "Target field design for magic angle

gradient coils." Journal of Magnetic Resonance 140(1): 285-288.

Barbara, T. M. and C. E. Bronnimann (1999). "Target field design for magic angle

gradient coils." J Magn Reson 140(1): 285-8.

Brennan, S. O., J. Shaw, et al. (1992). "Beta 141 Leu is not deleted in the unstable

haemoglobin Atlanta-Coventry but is replaced by a novel amino acid of mass 129

daltons." Br J Haematol 81(1): 99-103.

161 Brennan, S. O., J. G. Shaw, et al. (1993). "Posttranslational modification of beta 141 Leu

associated with the beta 75(E19)Leu-->Pro mutation in Hb Atlanta." Hemoglobin

17(1): 1-7.

Buczek, O., D. Yoshikami, et al. (2005). "Post-translational amino acid isomerization: a

functionally important D-amino acid in an excitatory peptide." J Biol Chem

280(6): 4247-53.

Callihan, D., J. West, et al. (1996). "Simple, distortion-free homonuclear spectra of

peptides and nucleic acids in water using excitation sculpting." Journal of

Magnetic Resonance Series B 112(1): 82-85.

Craig, A. G., P. Bandyopadhyay, et al. (1999). "Post-translationally modified

neuropeptides from Conus venoms." Eur J Biochem 264(2): 271-5.

Craig, A. G., T. Norberg, et al. (1999). "Contulakin-G, an O-glycosylated invertebrate

neurotensin." Journal of Biological Chemistry 274(20): 13752-13759.

Craig, A. G., M. Park, et al. (2001). "Enzymatic glycosylation of contulakin-G, a

glycopeptide isolated from Conus venom, with a mammalian ppGalNAc-

transferase." Toxicon 39(6): 809-15.

Cui, J., Y. Wang, et al. (2000). "Polyhydroxylated sterols with biological activities in

marine organism." Tianran Chanwu Yanjiu Yu Kaifa 12(6): 94-100.

Czerwiec, E., G. S. Begley, et al. (2002). "Expression and characterization of

recombinant vitamin K-dependent -glutamyl carboxylase from an invertebrate,

Conus textile." Eur J Biochem 269(24): 6162-72.

162 Danieli, B. and S. Riva (1994). "Enzyme-mediated regioselective acylation of

polyhydroxylated natural products." Pure and Applied Chemistry 66(10-11):

2215-2218.

David J. Adams, P. F. A. D. J. C. R. D. D. R. J. L. (1999). "Conotoxins and their

potential pharmaceutical applications." Drug Development Research 46(3-4):

219-234.

Duda, T. F., Jr. and S. R. Palumbi (1999). "Molecular genetics of ecological

diversification: duplication and rapid evolution of toxin genes of the venomous

gastropod Conus." Proc Natl Acad Sci U S A 96(12): 6820-3.

Easton, C. J. and M. C. Merrett (1997). "Stereoselective synthesis of (2S,3S)-[gamma]-

hydroxyvaline utilizing an asymmetric radical hydrogen bromide addition."

Tetrahedron 53(3): 1151-1156.

Fields, C. G. and G. B. Fields (1993). "Minimization of tryptopahn alkylation following

9-fluorenylmethoxycarbonyl solid-phase synthesis." Tetrahedron Letters 34(42):

6661-6664.

Fields, G. B., J. L. Lauer-Fields, et al. (2001). Synthetic Peptides: A User's Guide. New

York, W. H. Freeman & Co.

Fujii, N. (2002). "D-Amino Acids in Living Higher Organisms." Origins of Life and

Evolution of Biospheres 32(2): 103-127.

Harding, J. J. and M. J. C. Crabbe (1992). Post-translational modifications of proteins.

Boca Raton, CRC Press.

Heading, C. E. (2002). "Conus peptides and neuroprotection." Current Opinion in

Investigational Drugs (PharmaPress Ltd.) 3(6): 915-920.

163 Heck, S. D., W. S. Faraci, et al. (1996). "Posttranslational amino acid epimerization:

enzyme-catalyzed isomerization of amino acid residues in peptide chains." Proc

Natl Acad Sci U S A 93(9): 4036-9.

Heck, S. D., C. J. Siok, et al. (1994). "Functional consequences of posttranslational

isomerization of Ser46 in a calcium channel toxin." Science 266(5187): 1065-8.

Imperiali, B., S. L. Fisher, et al. (1992). "A conformational study of peptides with the

general structure Ac-L-Xaa-Pro-D-Xaa-L-Xaa-NH2: spectroscopic evidence for

a peptide with significant b-turn character in water and in dimethyl sulfoxide "

Journal of the American Chemical Society 114(9): 3182-3188.

Ishida, T., Y. In, et al. (1992). "Effect of the D-Phe2 residue on molecular conformation

of an endogenous neuropeptide achatin-I. Comparison of X-ray crystal structures

of achatin-I (H-Gly-D-Phe-Ala-Asp-OH) and achatin-II (H-Gly-Phe-Ala-Asp-

OH)." FEBS Lett 307(3): 253-6.

Jacobsen, R., E. C. Jimenez, et al. (1998). "The contryphans, a D-tryptophan-containing

family of Conus peptides: interconversion between conformers." J Pept Res 51(3):

173-9.

Jimenez, E. C., B. M. Olivera, et al. (1996). "Contryphan is a D-tryptophan-containing

Conus peptide." J Biol Chem 271(45): 28002-5.

Jimenez, E. C., M. Watkins, et al. (2001). "Contryphans from Conus textile venom

ducts." Toxicon 39(6): 803-8.

Jones, R. M. and G. Bulaj (2000). "Conotoxins - new vistas for peptide therapeutics."

Curr Pharm Des 6(12): 1249-85.

164 Kamatani, Y., H. Minakata, et al. (1990). "Molecular conformation of achatin-I, an

endogenous neuropeptide containing D-amino acid residue. X-ray crystal

structure of its neutral form." FEBS Lett 276(1-2): 95-7.

King, D. S., C. G. Fields, et al. (1990). "A A cleavage method which minimizes side

reactions following Fmoc solid phase peptide synthesis." International Journal of

Peptide and Protein Research 36(3): 255-266.

Krishna, R. G. and F. Wold (1998). Proteins: Design and Analysis. R. H. Angeletti. San

Diego, Academic Press: 121-206.

Lottspeisch, F., T. Houthaeve, et al. (1999). Microcharacterization of Proteins. R.

Kellner, F. Lottspeisch and H. E. Meyer. Weinheim, Wiley-VCH.

Mari, F. and G. B. Fields (2003). "Conopeptides: Unique pharmacological agents that

challenge current peptide methodologies." Chimica Oggi 21: 43-48.

Mari, F., P. M. Lahti, et al. (1992). " Molecular modeling of the Wittig reaction. 3. A

theoretical study of the Wittig olefination reaction: MNDO-PM3 treatment of the

Wittig half-reaction of unstabilized ylides with aldehydes." Journal of the

American Chemical Society 114(3): 813-821.

Marshall, J., W. P. Kelley, et al. (2002). "Anatomical correlates of venom production in

Conus californicus." Biol Bull 203(1): 27-41.

McGahren, W. J., G. O. Morton, et al. (1977). "Carbon-13 nuclear magnetic resonance

studies on a new antitubercular peptide antibiotic LL-BM547beta." J Org Chem

42(8): 1282-6.

McIntosh, J. M. and R. M. Jones (2001). "Cone venom - from accidental stings to

deliberate injection." Toxicon 39(10): 1447-51.

165 McIntosh, J. M., B. M. Olivera, et al. (1999). "Conus peptides as probes for ion

channels." Methods Enzymol 294: 605-24.

Milne, T. J., G. Abbenante, et al. (2003). "Isolation and characterization of a cone snail

protease with homology to CRISP proteins of the pathogenesis-related protein

superfamily." J Biol Chem 278(33): 31105-10.

Mitchell, J. B. and J. Smith (2003). "D-amino acid residues in peptides and proteins."

Proteins 50(4): 563-71.

Morita, H., T. Kayashita, et al. (1997). "Cyclic peptides from higher plants .33.

Delavayins A-C, three new cyclic peptides from Stellaria delavayi." Journal of

Natural Products 60(3): 212-215.

Myers, R. A., L. J. Cruz, et al. (1993). "Conus peptides as chemical probes for receptors

and ion channels." Chemical Reviews 93(5): 1923-36.

Newcomb, R. and G. Miljanich (2002). Handbook of Neurotoxicology. E. J. Massaro.

Totowa, NJ, Humana Press: 617-651.

Olivera, B. M. (2002). "CONUS VENOM PEPTIDES: Reflections from the Biology of

Clades and Species." Annual Review of Ecology and Systematics 33(1): 25-47.

Pallaghy, P. K., W. He, et al. (2000). "Structures of the contryphan family of cyclic

peptides. Role of electrostatic interactions in cis-trans isomerism." Biochemistry

39(42): 12845-52.

Papov, V. V., T. V. Diamond, et al. (1995). "Hydroxyarginine-containing polyphenolic

proteins in the adhesive plaques of the marine mussel Mytilus edulis." J Biol

Chem 270(34): 20183-92.

166 Perret, S., C. Merle, et al. (2001). "Unhydroxylated triple helical collagen I produced in

transgenic plants provides new clues on the role of hydroxyproline in collagen

folding and fibril formation." J Biol Chem 276(47): 43693-8.

Piotto, M., V. Saudek, et al. (1992). "Gradient-tailored excitation for single-quantum

NMR spectroscopy of aqueous solutions." J Biomol NMR 2(6): 661-5.

Pollard, J. K., E. Sondheimer, et al. (1958). "Hydroxyvaline in Kalanchoe

daigremontiana." Nature 182(4646): 1356-8.

Shen, G. S., R. T. Layer, et al. (2000). "Conopeptides: From deadly venoms to novel

therapeutics." Drug Discov Today 5(3): 98-106.

Shoji, J. and H. Hinoo (1975). "Chemical characterization of new antibiotics, cerexins A

and B. (Studies on antibiotics from the genus Bacillus. II)." J Antibiot (Tokyo)

28(1): 60-3.

Sidler, W., B. Kumpf, et al. (1985). " Structural studies on cryptomonad biliprotein

subunits. Two different a-subunits in Chroomonas phycocyanin-645 and

Cryptomonas phycoerythrin-545." Biological Chemistry Hoppe-Seyler 366(3):

233-244.

Siethoff, C., C. Lohaus, et al. (1999). Microcharacterization of Proteins. R. Kellner, F.

Lottspeisch and H. E. Meyer. Weinheim, Wiley-VCH.

Slabicki, M. M., M. J. Potrzebowski, et al. (2004). "X-ray and nuclear magnetic

resonance (NMR) studies of signalizing the tripeptide sequence (Tyr-D-Ala-Phe)

of dermorphin and deltorphins I and II. Comparative analysis in the liquid and

solid phases." Journal of Physical Chemistry B 108(14): 4535-4545.

167 Smallcombe, S. H., S. L. Patt, et al. (1995). "WET solvent suppression and its

applications to LC NMR and high-resolution NMR spectroscopy." Journal of

Magnetic Resonance Series A 117(2): 295-303.

Soyez, D., J. Y. Toullec, et al. (2000). "L to D amino acid isomerization in a peptide

hormone is a late post-translational event occurring in specialized neurosecretory

cells." J Biol Chem 275(48): 37870-5.

Storch, E. M., J. S. Grinstead, et al. (1999). "Engineering out motion: a surface disulfide

bond alters the mobility of tryptophan 22 in cytochrome b5 as probed by time-

resolved fluorescence and 1H NMR experiments." Biochemistry 38(16): 5065-75.

Terlau, H. and B. M. Olivera (2004). "Conus Venoms: A Rich Source of Novel Ion

Channel-Targeted Peptides." Physiol. Rev. 84(1): 41-68.

Terui, Y., J. Nishikawa, et al. (1990). "Structures of cepafungins I, II and III." J Antibiot

(Tokyo) 43(7): 788-95.

Torres, A. M., I. Menz, et al. (2002). "D-Amino acid residue in the C-type natriuretic

peptide from the venom of the mammal, Ornithorhynchus anatinus, the Australian

platypus." FEBS Lett 524(1-3): 172-6.

Walker, C. S., R. P. Shetty, et al. (2001). "On a potential global role for vitamin K-

dependent gamma-carboxylation in animal systems. Evidence for a gamma-

glutamyl carboxylase in Drosophila." J Biol Chem 276(11): 7769-74.

Watson, A. A., G. W. Fleet, et al. (2001). "Polyhydroxylated alkaloids -- natural

occurrence and therapeutic applications." Phytochemistry 56(3): 265-95.

Wilm, M. and M. Mann (1996). "Analytical properties of the nanoelectrospray ion

source." Anal Chem 68(1): 1-8.

168 Wilm, M., A. Shevchenko, et al. (1996). "Femtomole sequencing of proteins from

polyacrylamide gels by nano-electrospray mass spectrometry." Nature 379(6564):

466-9.

Wüthrich, K. (1986). NMR of proteins and nucleic acids. New York, Wiley.

Yang, H., G. Zheng, et al. (2003). "D-Amino acids and D-Tyr-tRNA(Tyr) deacylase:

stereospecificity of the translation machine revisited." FEBS Lett 552(2-3): 95-8.

Zeng, L. M., J. Y. Su, et al. (1999). "Search for new compounds and biologically active

substances from Chinese marine organisms." Pure and Applied Chemistry 71(6):

1147-1151.

169