MIAMI UNIVERSITY – THE GRADUATE SCHOOL

CERTIFICATE FOR APPROVING THE DISSERTATION

We hereby approve the Dissertation

Of

Elvis K. Tiburu

Candidate for the Degree:

Doctor of Philosophy

Dr. Gary A. Lorigan, Director

Dr. Christopher A. Makaroff, Reader

Dr. Robert E. Minto, Reader

Dr. Richard T. Taylor, Reader

Dr. David G. Pennock Graduate School Representative ABSTRACT

DEVELOPMENT OF NEW METHODS FOR THE ALIGNMENT OF LONGER CHAIN PHOSPHOLIPIDS IN BICELLES AND SOLID-STATE NMR STUDIES OF PHOSPHOLAMBAN

by Elvis K. Tiburu

Magnetically aligned phospholipid bilayers or bicelles are model systems that mimic biological membranes for magnetic resonance studies. A long chain phospholipid bilayer system that spontaneously aligns in a static magnetic field was characterized utilizing solid-state NMR spectroscopy. The oriented membrane system was composed of a mixture of the bilayer-forming phospholipid palmitoylstearoylphosphatidylcholine (PSPC) and the short-chain phospholipid dihexanoylphosphatidylcholine (DHPC) that breaks up the extended bilayers into bilayered micelles or bicelles that are highly hydrated. Traditionally, the shorter 14-carbon chain phospholipid dimyristoyl- phosphatidylcholine (DMPC) has been utilized as the bilayer-forming phospholipid in bicelle studies. The effect of cholesterol in bicelles containing chain perdeuterated

2 DMPC, a partially deuterated (a-[2,2,3,4,4,6- H6]) cholesterol, and stearic acid-d35 has been reported as a function of temperature using 2H solid-state NMR spectroscopy. The order parameters of the labeled probes were calculated and compared with values obtained from unoriented samples in the literature. In addition, 2H solid-state NMR spectroscopy was used to investigate the orientation and side chain dynamics of specific- labeled methyl groups of leucines in PLB in unoriented as well as in magnetically and mechanically aligned phospholipids bilayers. Phospholamban (PLB), is a 52 amino-acid protein that assembles into a pentamer in cardiac sarcoplasmic reticulum membranes. The protein is a key regulator of the calcium ATPase through an inhibitory association that can be reversed by phosphorylation. From the 2H NMR studies, the exhibited line shapes were characteristic of either methyl group reorientation about the Cg-Cd bond axis, 15 or by additional wobbling motion about the Ca-Cb and Cb-Cg bond axes. Using N NMR spectroscopy, the backbone dynamics of PLB as well as the orientation of PLB in the model membranes were also studied. A comparative study of PLB in shorter chain DMPC as well as in longer chain DOPC bilayers was conducted to determine the effect of hydrophobic length on peptide orientation. Taken together, the results are discussed in terms of the structure of PLB in phospholipid bilayers previously proposed on the basis of mutational and molecular modeling studies, thus, providing an understanding of the structure of PLB in model membranes. DEVELOPMENT OF NEW METHODS FOR THE ALIGNMENT OF LONGER CHAIN PHOSPHOLIPIDS IN BICELLES AND SOLID-STATE NMR STUDIES OF PHOSPHOLAMBAN

A DISSERTATION

Submitted to the Faculty of Miami University in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Chemistry and Biochemistry

By

Elvis K. Tiburu Miami University Oxford, Ohio 2004

Dissertation Director: Gary A. Lorigan Table of Contents

List of Figures vii List of Tables vi

Chapter 1: Development of Magnetically Aligned Phospholipid Bilayers in Mixtures of Palmitoylstearoylphosphatidylcholine and Dihexanoylphosphatidylcholine by Solid-State NMR Spectroscopy 1 1.0 Introduction 3 1.1 Phospholipid bilayers as model systems for protein studies 3 1.2 Materials 4 1.3 Methods 5 1.3.1 Sample preparation 5 1.3.2 NMR spectroscopy 6 1.4 Results 6 1.4.1 Magnetic alignment of phospholipid bilayers with 2H NMR spectroscopy 6 1.4.2 Magnetic alignment of phospholipid bilayers with 31P NMR spectroscopy 12 1.5 Discussion 12 1.6 References 18

Chapter 2: Solid-State 2H NMR Studies of the Effects of Cholesterol on Acyl Chain Dynamics of Magnetically Aligned Phospholipid Bilayers 21 2.0 Introduction 23 2.1 The effects of cholesterol and stearic acid on phospholipid bilayers 23 2.2 Materials 26 2.3 Methods 26 2.3.1 Sample preparation 26 2.3.2 NMR spectroscopy 27 2.3.3 Order parameter calculations 28

ii 2.4 Results and discussion 29 2.4.1 Cholesterol dynamics in phospholipid bilayers 29 2.4.2 Magnetic aligned DMPC/DHPC bicelles investigated by

stearic acid-d35 36

2.4.3 Calculating order parameters for stearic acid-d35 inserted into magnetically aligned phospholipid bilayers 39 2.5 Conclusion 41 2.6 References 44

Chapter 3: An Improved Synthetic and Purification Procedure for the Hydrophobic Segment of the Peptide Phospholamban 49 3.0 Introduction 51 3.1 The role of phospholamban in Ca2+ transport 51 3.2 Materials 54 3.3 Methods 55 3.3.1 Peptide synthesis 55 3.3.2 Peptide purification 56 3.4 Results and discussion 56 3.5 References 63

Chapter 4: Investigating Structural Changes in the Lipid Bilayer upon Insertion of the Transmembrane Domain of the Membrane-Bound Protein Phospholamban Utilizing 31P and 2H Solid-State NMR Spectroscopy 68 4.0 Introduction 70 4.1 Protein-lipid interaction 70 4.2 Materials 71 4.3 Methods 72 4.3.1 Synthesis and purification of PLB 72 4.3.2 NMR Sample preparation 72

iii 4.3.3 NMR spectroscopy 72 4.3.4 NMR data analysis 73 4.4 Results and discussion 74 4.4.1 31P NMR study of PLB incorporated into POPC bilayer 74 4.4.2 2H NMR study of PLB incorporated into POPC bilayer 80 4.5 Conclusion 84 4.6 References 86

Chapter 5: Investigating the Dynamic Properties of the Transmembrane Segment of Phospholamban Incorporated into Phospholipid Bilayers Utilizing 2H and 15N Solid-State NMR Spectroscopy 92 5.0 Introduction 94 5.1 Dynamics of PLB in phospholipid bilayers 94 5.2 Materials 98 5.3 Methods 99 5.3.1 Peptide synthesis 99 5.3.2 Peptide purification 99 5.3.3 Solid-state NMR sample preparation 100 5.3.4 2H solid-state NMR spectroscopy 100 5.3.5 15N soilid-state NMR spectroscopy 101 5.3.6 NMR data analysis 101 5.4 Results 102 5.4.1 2H side chain dynamics 102 5.4.2 15N backbone dynamics 105 5.4.3 Helical orientation of PLB with respect to the lipid bilayer 107 5.5 Discussion 109 5.5.1 Dynamics of PLB in phospholipid bilayers 109 5.5.2 Structural implications of the leucine side-chain motions 109 5.5.3 Structural model of TM-PLB in lipid bilayers 115 5.6 Conclusion 116 5.7 References 118

iv Chapter 6: 2H and 15N Solid-State NMR Spectroscopic Studies of the Helical Tilt of the Transmembrane Segment of Phospholamban Incorporated into Magnetically Aligned and Mechanically Aligned Phospholipid Bilayers 125 6.0 Introduction 127 6.1 Incorporation of PLB into lipids of different chain lengths 127 6.2 Materials 130 6.3 Methods 131 6.3.1 Peptide synthesis 131 6.3.2 Peptide purification 131 6.3.3 2H and 15N aligned solid-state NMR sample preparation 132 6.3.4 2H solid-state NMR spectroscopy 133 6.3.5 15N soilid-state NMR spectroscopy 133 6.3.6 2H tilt angle calculations 133 6.3.7 15N tilt angle calculations 135 6.4 Results 136 6.5 Discussion 141 6.6 References 150

v List of Tables

Table 5.1 CSA values measured from 15N-labeled PLB 114 Table 6.1 5N chemical shifts from 15N-labeled PLB 143

vi List of Figures

Figure 1.1 2H NMR of a 25% (W/W) q=3.5 DMPC/DHPC 7 Figure 1.2 2H NMR of a 25% (W/W) q=1.8 PSPC/DHPC 8 Figure 1.3 2H NMR of PSPC/DHPC as a function of temperature 9 Figure 1.4 2H NMR of a 25% (W/W) q=2.0 PSPC/DHPC/Yb3+ 10 Figure 1.5 31P NMR of a 25% (W/W) q=1.8 PSPC/DHPC 14

Figure 2.1 Structure of stearic acid-d35 29 Figure 2.2 2H NMR of deuterated cholesterol in DMPC/DHPC 30 Figure 2.3 2H NMR of DMPC/DHPC with cholesterol 32 Figure 2.4 2H NMR of a 25% (W/W) q=3.5 DMPC/DHPC/Yb3+ 33

1 Figure 2.5 Temperature-dependent SCD order parameter profile 35 2 3+ Figure 2.6 H NMR of stearic acid-d35 in DMPC/DHPC/Yb 37 2 3+ Figure 2.7 H NMR of stearic acid-d35 in DMPC/DHPC/ Chol/ Yb 38 1 3+ Figure 2.8 Temperature-dependent SCD with Yb 40 2 3+ O Figure 2.9 H NMR of stearic acid-d35 in DMPC/DHPC/Yb at 40 C 42

Figure 3.1 Two models of PLB 53 Figure 3.2 Fmoc deprotection of amino acids 58 Figure 3.3 Reverse-phase HPLC profile of PLB 59 Figure 3.4 MALDI-TOF MS of PLB 61

Figure 4.1 31P NMR of POPC bilayers as a function of temperature 75 Figure 4.2 31P NMR of POPC bilayers with 4 mol% PLB 76 Figure 4.3 31P NMR of POPC bilayers with 6 mol% PLB 78 Figure 4.4 2H NMR powder pattern spectra of PLB in POPC bilayers 81 Figure 4.5 2H NMR powder pattern spectra of PLB with temperature 82

1 Figure 4.6 Temperature-dependent SCD order parameter profile 84

vii Figure 5.1 The two models of PLB in lipid bilayers 97 Figure 5.2 The amino acid sequence of PLB 102 Figure 5.3 2H NMR powder pattern spectra of PLB in POPC 103 Figure 5.4 15N NMR powder pattern spectra of PLB in POPC 106 Figure 5.5 15N solid-state NMR of PLB in oriented DOPC bilayers 108

2 Figure 5.6 Simulations of CD3-LeuPLB NMR lineshapes from H spectra 111 Figure 5.7 A model of pentameric PLB in phospholipid bilayer 116

Figure 6.1 2H NMR spectra of a DMPC/DHPC/2H-labeled PLB bicelle Sample 138 Figure 6.2 2H NMR spectra of DOPC/DOPE/2H-labeled PLB samples 139 Figure 6.3 One-dimensional solid-state 15N NMR spectra of specifically 15N-labeled PLB in oriented DMPC/DHPC bicelles with q ratio 3.5 and DOPC/DOPE phospholipid bilayers 142 Figure 6.4 Contour plots representing the chemical shifts calculated for the rotation, r from 00 to 3600 and for tilt angle, t from 00 to 900 144 Figure 6.5 The two models showing the orientation of monomeric PLB incorporated into phospholipid bilayers 148

viii Dedication

This work is dedicated to Jesus Christ, whom I am privileged to know as Lord, Savior and Friend.

“Religion without science is blind. Science without religion is lame”. (Albert Einstein)

ix Acknowledgement

I am grateful to Professor Gary A. Lorigan who has broadened my perspective and deepened my appreciation for science. You are so wonderful and I am proud of you sir. The Professors who occupy Hughes Science building, here at Miami University have blessed me with their friendship and guidance. To all of you, thank you. But in particular I am grateful and proudly indebted to my parents, Nipontra and John Tiburu. Because of you mom and dad I had the opportunity to go to school. What I learned about science, I learnt first from you. May the Lord give you everlasting rest in Heaven. To Sandy, the Gilberts and my Church family, I am so thankful for your support. Your Prayers are deeply appreciated. To Mark, Sam, Essel, Ajaloo and Amila you are all part of the success story and I wish you all the best. The Lorigan group, past and present, Bill, Tom, Nusreen, Ethan, Paresh, Damu, Prem, Shadi, Lu, Dana, Marc, Johnson, Simon, Tia and the list goes on. I really enjoyed working with you all. You are the finest and friendliest people I ever worked with. Keep the lab vibrant because the best is yet to come. Last but not least, of course, are the members of my committee. I really appreciate your willingness to serve on this committee. May God Bless you abundantly.

x CHAPTER 1

Elvis K. Tiburu, Dana Moton, and Gary A. Lorigan, Development of Magnetically Aligned Phospholipid Bilayers in Mixtures of Palmitoylstearoylphosphatidylcholine and Dihexanoylphosphatidylcholine by Solid-State NMR Spectroscopy (2001), Biochim. Biophys. Acta , 512, 206-214

1 Abstract

This study reports the solid-state NMR spectroscopic characterization of a long chain phospholipid bilayer system which spontaneously aligns in a static magnetic field. Magnetically aligned phospholipid bilayers or bicelles are model systems which mimic biological membranes for magnetic resonance studies. The oriented membrane system is composed of a mixture of the bilayer forming phospholipid palmitoylstearoylphosphatidylcholine (PSPC) and the short chain phospholipid dihexanoylphosphatidylcholine (DHPC) that breaks up the extended bilayers into bilayered micelles or bicelles that are highly hydrated (approximately 75% aqueous). Traditionally, the shorter 14 carbon chain phospholipid dimyristoylphosphatidylcholine (DMPC) has been utilized as the bilayer forming phospholipid in bicelle studies. Alignment

(perpendicular) was observed with a PSPC/DHPC q ratio between 1.6 and 2.0 slightly above Tm at 50 oC with 2H and 31P NMR spectroscopy. Paramagnetic lanthanide ions (Yb3+) were added to flip the bilayer discs such that the bilayer normal was parallel with the static magnetic field. The approximate 1.8 (PSPC/DHPC) molar ratio yields a thicker membrane due to the differences in the chain lengths of the DMPC and PSPC phospholipids. The phosphate-to-phosphate thickness of magnetically aligned PSPC/DHPC phospholipid bilayers in the La phase may enhance the activity and/or incorporation of different types of integral membrane proteins for solid-state NMR spectroscopic studies.

2 1.0 Introduction 1.1 Phospholipid bilayers as model systems for protein studies.

Magnetically aligned phospholipid bilayers (bicelles) have been demonstrated to be useful models for studying the structural and dynamic properties of membrane systems and integral membrane proteins using solid-state nuclear magnetic resonance (NMR) spectroscopic techniques (1-8). The magnetic alignment of bicelles is due to the anisotropy of the overall magnetic susceptibility of the system. The negative sign of the diamagnetic susceptibility anisotropy tensor (Dc < 0) for phospholipid bilayers dictates that the bicelles align with their bilayer normal oriented perpendicular to the direction of the static magnetic field. The addition of paramagnetic lanthanide ions with a large positive Dc (Eu3+, Er3+, Tm3+, and Yb3+) can cause the bicelles to flip 90o such that the average bilayer normal is collinear with the direction of the static magnetic field. One advantage of flipping the phospholipid bicelles is that the spectral resolution is dramatically increased, because the spectral width is spread over a broader frequency range. Additionally, in a uniaxially aligned system, the highly anisotropic spectral data can yield the orientation and structure of different segments of the protein with respect to the magnetic field and the lipid bilayer (9-11). The development of dilute magnetically aligned aqueous liquid-crystalline media has dramatically increased the refinement of structural studies of globular proteins with high- resolution NMR spectroscopy (12-15). Generally, in an isotropic solution, internuclear dipolar couplings average to zero as a result of rotational molecular motion. By dissolving proteins in a dilute magnetically oriented aqueous liquid-crystalline medium (bicelle), a tunable degree of solute alignment with respect to the magnetic field can be created while retaining both the resolution and sensitivity of the regular isotopic NMR spectrum. In this system, the dipolar couplings no longer average to zero and can be accurately measured. This approach has been shown to significantly improve the accuracy of structures determined by solution NMR spectroscopy, and extend the size limit. The typical bicelle consists of long chain phospholipids such as 1, 2-dimyristoyl-sn- glycero-3-phosphocholine (DMPC) and a detergent such as 1,2-dihexanoyl-sn-glycero-3- phosphocholine (DHPC). The diameter of the bicelle increases as the molar ratio between the long-chain (n1) and short-chain phospholipid (n2), q=n1/n2 increases (3, 7, 16, 17). Phospholipids

3 that are present in a bilayer undergo various temperature-dependent phase transitions. Above the phase transition temperature (Tm), the phospholipids are in a relatively fluid liquid crystalline state. This transition can be of great significance for enzyme activity (18). The lamellar liquid- crystalline phase is also termed the La-phase (19-21). The transition from the gel phase to the La phase above the Tm is a requirement for magnetically aligned phospholipid bilayers and can be explained in terms of the strong forces in the polar head sheets (due to electrostatic interactions between headgroups) and the weak forces between the acyl chains. When a phospholipid bilayer system is heated above Tm , the van der Waals forces between the hydrocarbons become weak compared to the thermal motions (7, 19). Thus, the chains are transformed into a state of disorder with high degree of gauche conformation (7, 19). Since a bilayer arrangement of phospholipids is characterized by a distinct central hydrophobic region bounded by two polar interfacial regions, the expectation is that the thickness of the hydrophobic region will have influence on the structure and function of transmembrane proteins (22-24). Despite the success of using DMPC/DHPC bicelle systems to study the structural and dynamic properties of membrane proteins with solid-state NMR spectroscopy there are size limitations on the length and size of proteins to be studied (25). Natural biological lipids are dominated by chains with 16 to 18 carbons in length. Also, a-helical transmembrane proteins generally contain 25 residues, which may be too long to assemble into the standard DMPC/DHPC bicelle matrix. The objective of the current study is to develop a method to magnetically align longer chain phospholipids in a static magnetic field. Due to the longer acyl chains the lipid bicelles of such a system will extend the thickness of the lipid bilayer and has the potential to provide a better model for a biological membrane and possibly enhance activity for solid-state NMR investigations. In the present study, we have investigated the magnetic alignment of a mixture of 1- palmitoyl-2-stearoyl-sn-glycerol-3-phosphatidylcholine (PSPC) and DHPC at various phospholipid concentration (q) ratios and temperature. The results obtained from this new magnetically aligned phospholipid bilayer system are presented.

1.2. Materials

DMPC, PSPC, DHPC, DMPC-d54 and DPPC-d62 were purchased from Avanti Polar Lipids Inc. (Alabaster, AL). All phospholipids were dissolved in chloroform and stored at –20

4 OC prior to use. Ytterbium(III) chloride hexahydrate and HEPES (N-[2-hydroxyethyl]piperazine- N'-2-ethanesulfonic acid]) were obtained from Sigma/Aldrich. Deuterium-depleted water was obtained from Isotec Inc. (Miamisburg, OH).

1.3. Methods

1.3.1. Sample Preparation

The standard DMPC/DHPC bicelle sample, consisting of 25% (w/w) phospholipid to solution with a q = 3.54, was made in two separate 15 mL or 25 mL pear-shaped flasks. In one flask DMPC and DMPC-d54 were mixed together in chloroform at molar ratios of 0.1/0.005 respectively while in the second flask DHPC was added. The mixture in both flasks were rotovaped down to remove the chloroform solvent from the phospholipid mixture and both flasks were placed under high vacuum overnight. The following day a 100 mM HEPES buffer of pH 7.0 was prepared using deuterium- depleted water and the appropriate amount was added to the flask containing the DHPC. The flask was then vortexed briefly, sonicated for 5 min and vortexed again. The sample was sonicated with a FS30 (Fisher Scientific) bath sonicator with the heater turned off. Several freeze(77K)/thaw cycles(room temperature) made the dispersion more homogenous and also removed all of the air bubbles. The sample was then transferred into the flask containing the

DMPC and DMPC-d54, and vortexed and sonicated again. Similar procedures were also used to prepare the PSPC/DHPC/DPPC-d62 bicelle samples except that all the phospholipids were mixed together in one 25-mL pear-shaped flask. Typically, the total mass of the prepared bicelle sample was 300 mg. An aqueous solution of ytterbium (III) chloride hexahydrate was prepared fresh with deuterium-depleted water. The bicelle sample was placed into a 21-mm NMR flat- bottom tube with a 5-mm o.d. on ice via a Pasteur pipette. To monitor bicelle alignment with 2H NMR spectroscopy a 20:1 ratio of regular and chain-perdeuterated phospholipids was used in this study.

5 1.3.2. NMR Spectroscopy

All solid-state NMR experiments were carried out on a modified Bruker AVANCE 7.05 T narrow bore 300/54 magnet configured to conduct high-power solid-state NMR studies. The

1 2 resonance frequencies were 300.0 MHz for H, 46.1 MHz for H, and 121.5 MHz for 31P. The solid-state NMR spectra were gathered with a static double-tuned 5-mm round-coil solid-state NMR probe purchased from Doty Scientific Inc. 2H NMR spectra were recorded using a standard quadrupole-echo pulse sequence (3.0-ms 90O pulses, 45-ms interpulse delay, 5.12-ms acquisition time, and a 0.4-sec recycle delay). 31P NMR were recorded with proton decoupling under static conditions using a 11.7-ms excitation pulse. 31P chemical shifts were referenced to

85% H3PO4. The NMR data processing was carried out on a 300 MHz Power Macintosh G3 computer running Igor Pro 3.12 (Wavemetrics, Lake Oswego, OR) and MacNuts (Acorn NMR Inc., Livermore, CA).

1.4. Results 1.4.1. Magnetic alignment of phospholipid bilayers with 2H NMR spectroscopy

2H solid-state NMR spectra of a 25% w/w phospholipid bicelle of DMPC/DHPC to solution with a q =3.54 containing chain-perdeuterated DMPC-d54 were acquired over a temperature range from 25 OC to 60 OC. The results of the temperature study are displayed in Figure 1.1. At 25 OC, the DMPC/DHPC bicelle discs are not magnetically aligned. However, partial orientation is observed at 30 OC. At 35 OC and 40 OC, the spectra are characteristic of well-aligned liquid-crystalline bicelle discs. In this case, the bilayer normal is oriented perpendicular to the direction of the static magnetic field. At 45 OC, the DMPC/DHPC bicelle sample loses some of its orientational characteristics. Finally, between 50 OC and 60 OC powder characteristics are observed along with a strong isotropic component centered at 0 kHz. The standard DMPC/DHPC 2H bicelle NMR spectra serve as an excellent basis for comparing the magnetic alignment of different acyl chain length phospholipid bilayer discs. Similar 2H NMR studies are shown in Figure 1.2 for a 25% w/w PSPC/DHPC phospholipid bicelle sample with q=1.8 over a temperature range from 40 OC to 60 OC. At 40 OC and 45 OC, the spectra reveal only an isotropic component. The same isotropic component was

6 25 OC

30 OC

35 OC

O 40 C

45 OC

50 OC

55 OC

60 OC

60 40 20 0 -20 -40 -60 2 H (kHz) Figure 1.1 2H NMR spectra of a 25% (w/w) q= 3.54 DMPC/DHPC bicelle sample investigated as a function of temperature. The sample was doped with chain-perdeuterated DMPC to monitor bicelle alignment in a deuterium-depleted aqueous buffer. 2H NMR spectra were obtained with the quadrupole-echo pulse sequence incorporating 3.0-ms 90° pulses, a 45-ms interpulse delay, and a 0.4-s recycle delay. 2048 transients were acquired, and the free induction decay was processed with 200 Hz of line broadening. Each sample was allowed to equilibrate in the magnet for 30 minutes prior to signal acquisition.

7

40 OC

45 OC

50 OC

55 OC

60 OC

60 40 20 0 -20 -40 -60 2H (kHz)

Figure 1.2 2H NMR spectra of a 25% (w/w) q= 1.8 PSPC/DHPC bicelle sample investigated as a function of temperature. The sample was doped with chain- perdeuterated DPPC (DPPC-d62) to monitor bicelle alignment in a deuterium- depleted aqueous buffer. 2H NMR spectra were obtained with the quadrupole- echo pulse sequence incorporating 3.0-ms 90° pulses, a 45-ms interpulse delay, and a 0.4-s recycle delay. 2048 transients were acquired, and the free induction decay was processed with 200 Hz of line broadening. Each sample was allowed to equilibrate in the magnet for 30 minutes prior to signal acquisition.

8 q=1.4

q=1.5

q=1.6

q=1.8

q=2.0

q=2.25

q=2.5

q=2.75

q=3.0

q=10.0

60 40 20 0 -20 -40 -60 2 H (kHz)

Figure 1.3 2H NMR spectra of a 25% (w/w) PSPC/DHPC bicelle sample investigated at 50 oC as a function of the PSPC/DHPC molar q ratio. The sample was doped with chain-perdeuterated DPPC to monitor bicelle alignment in a deuterium-depleted aqueous buffer. 2H NMR spectra were obtained with the quadrupole-echo pulse sequence incorporating 3.0-ms 90° pulses, a 45-ms innerpulse delay, and a 0.4-s recycle delay. 2048 transients were acquired, and the free induction decay was processed with 200 Hz of line broadening. Each sample was allowed to equilibrate in the magnet for 30 minutes prior to signal acquisition.

9 (A) 0% Yb3+

(B) 5% Yb3+

60 40 20 0 -20 -40 -60 2 H (kHz)

Figure 1.4 2H NMR spectra of a 25% (w/w) q= 2.0 PSPC/DHPC bicelle sample investigated with and without the paramagnetic lanthanide ion Yb3+. (A) Bicelle spectrum contains 0% Yb3+ and is consistent with the bilayer normal being perpendicular to the magnetic field. (B) Bicelle spectrum contains 5% molar Yb3+ with respect to PSPC and is consistent with the bilayer normal being parallel to the direction of the static magnetic field. The sample was doped with chain- perdeuterated DPPC to monitor bicelle alignment in a deuterium-depleted aqueous buffer. 2H NMR spectra were obtained with the quadrupole-echo pulse sequence incorporating 3.0-ms 90°, a 45-ms interpulse delay, and a 0.4-s recycle delay. 2048 transients were acquired, and the free induction decay was processed with 200 Hz of line broadening. Each sample was allowed to equilibrate in the magnet for 30 minutes prior to signal acquisition.

10 observed at 25 OC, 30 OC, and 35 OC (data not shown). At 50 oC, the 2H NMR spectrum indicates that the PSPC/DHPC phospholipid bicelle discs are magnetically aligned such that the normal of the lipid bilayer is perpendicular to the static magnetic field. Above 50 OC, the spectra reveal a powder-like spectrum with a large isotropic component centered at 0 kHz. Unlike the DMPC/DHPC phospholipid discs, the temperature range at which the PSPC/DHPC bicelle discs aligns is higher and rather narrow. In order to investigate the orientational properties of PSPC/DHPC bicelle samples at various q ratios at 50 OC, 2H NMR spectra of samples with PSPC/DHPC q ratios between q =1.4 – 10 were obtained and the corresponding spectra are displayed in Figure 1.3. At q =1.4, a large isotropic peak centered at 0 kHz is observed. At q =1.5, a broader unoriented and isotropic component is observed. Between q =1.6-2.0, the powder and isotropic components disappear and orientational La-like phase characteristics are observed. In this case, the bilayer normal is aligned perpendicular to the static magnetic field. Although the individual deuterons along the acyl chain cannot be fully resolved, the spectra resemble the shape and breadth of previously published 2H NMR spectra of DMPC/DHPC bicelle samples (4, 16). As the q ratios increase from 2.25 – 10.0, orientational characteristics disappear and the spectra broaden out to reveal powder characteristics along with some isotropic components. As discussed previously, the addition of lanthanide ions such as Yb3+ or Tm3+ with a large positive Dc to DMPC/DHPC bicelle samples causes the phospholipid discs to flip 90o such that the bilayer normal is aligned parallel with the direction of the static magnetic field. Figure 1.4(A) shows a 2H NMR spectrum of a q=2.0 PSPC/DHPC bicelle sample at 50 oC prepared in the absence of lanthanide ions. In order to change the sign of the net magnetic susceptibility anisotropy tensor of the bicelles so that their normal is parallel to the applied magnetic field, 5% molar Yb3+ with respect to PSPC was added to the PSPC/DHPC bicelle sample and the results are shown in Figure 1.4(B). The quadrupolar splittings observed in the 2H NMR spectrum are spread out and much better resolved. The spectrum indicates that the PSPC/DHPC bicelle discs have flipped such that their bilayer normal is now parallel with the magnetic field. The 2H lineshape and breadth of the magnetically aligned lanthanide-doped PSPC/DHPC bicelle sample observed in Figure 1.4(B) are similar to previous DMPC/DHPC/lanthanide spectra in the literature (1, 4). The breadth is slightly larger for the PSPC bicelles in both the parallel and

11 perpendicular orientations when compared to the 2H spectra of standard DMPC/DHPC bicelle samples.

1.4.2 Magnetic alignment of phospholipid bilayers with 31P NMR spectroscopy

31P NMR spectroscopy is an excellent technique for studying magnetically aligned phospholipid bilayers in a magnetic field due to the presence of one phosphorus atom in the phospholipid headgroup (16, 26). 31P NMR resonance lineshapes and shifts contain information concerning headgroup conformation, membrane morphology, and the orientation of the lipid bilayer with respect to the magnetic field (27). The 1H-decoupled 31P NMR spectra of

O O PSPC/DHPC bicelles with q=1.8 are shown in Figure 1.5 from 40 C through 50 C. Below Tm, the two spectra at 40 OC and 45 OC reveal two closely spaced isotropic peaks at -0.55 ppm and - 0.75 ppm corresponding to a fluid-like membrane. These two spectra agree with the isotropic

2 O O O H NMR spectra at 40 C and 45 C shown in Figure 1.2. Above Tm at 50 C, two new peaks are clearly observed in the 31P spectrum. These two peaks are shifted upfield from the isotropic peaks. Based upon previous DMPC/DHPC bicelle 31P NMR studies, the broad peak centered at - 7.0 ppm corresponds to the phosphorus atoms of the long-chain PSPC phospholipids (16). The sharp peak observed at -3.0 ppm is due to the short-chain DHPC phospholipids located at the edges of the bicelle. Signal intensity integration of the two peaks indicates that the broad peak is twice as large as the narrow resonance transition. This agrees with the peak assignment due to the molar q ratio of 1.8/1 for the PSPC/DHPC bicelle sample.

1.5. Discussion

The results in this chapter provide important clues about the morphologies of longer chain magnetically aligned phospholipid bilayers (PSPC/DHPC) when compared to the traditional shorter chain DMPC/DHPC systems. Previous studies have indicated that the addition of detergents such as DHPC at the appropriate concentrations break up the extended DMPC phospholipid bilayers to form disc-shaped bicelles that magnetically align when placed in

12 a static magnetic field. In Figure 1.1, the DMPC/DHPC bicelle sample maintained La-phase aggregates in which DMPC forms a lipid bilayer while the shorter chain DHPC phospholipids cap the end of these lipid bilayers or assembles into the DMPC bilayers thereby stabilizing the structures. These aggregates align such that the bilayer normal is oriented perpendicular to the direction of the static magnetic field. The degree of orientation depends on several factors including q ratios (DMPC/DHPC), sample temperature, level of hydration, and concentration of lanthanide ions. Phospholipids with longer hydrocarbon chains were used in place of DMPC for several different reasons in this study. First of all, the application of DMPC/DHPC bicelles may not be optimal for solid-state NMR studies for all integral membrane proteins. The Sanders group has shown that activities vary for the integral membrane protein diacylglycerol kinase (DAGK) at different detergent:phospholipid ratios (25). They concluded that the phosphate-to-phosphate membrane thickness of lipid bicelles in the La phase affects the activity of proteins within the membrane (25). Thus, by investigating the magnetic alignment characteristics of longer chain phospholipid discs, which extend the thickness of the bicelle membrane, we hope to expand the bicelle technique to include additional integral membrane proteins for solid-state NMR studies. Also, larger membrane proteins that contain large hydrophobic regions may be more active or magnetically align better in a longer chain phospholipid disc such as PSPC/DHPC when compared to a standard 14 carbon acyl chain in the DMPC/DHPC bicelle matrix. Figures 1.1 and 1.2 indicate that the longer chain PSPC/DHPC bicelle discs magnetically align at a higher temperature when compared to the standard DMPC/DHPC phospholipid bilayer arrays. Characteristics of magnetic alignment are observed only at 50 OC, whereas the DMPC/DHPC phospholipid bilayer arrays align between 34 OC and 43 OC. DMPC/DHPC bicelle discs are known to magnetically align above the gel to liquid crystalline phase transition

O temperature of pure DMPC (Tm = 23 C) (7, 16, 23). Similarly, our results indicate that the O PSPC/DHPC bicelles align above the T m of pure PSPC (approximately 48 C) (23). Additionally, the range of temperatures at which the PSPC/DHPC bicelle discs align is very narrow when compared to the DMPC/DHPC system which could be due to instability of the

O system above or below the Tm value of 48 C (23).

13

40 OC

45 OC

50 OC

20 10 0 -10 -20 31P (ppm)

Figure 1.5 31P NMR spectra of a 25% (w/w) q= 1.8 PSPC/DHPC bicelle sample investigated as a function of temperature. 31P NMR spectra were obtained with proton decoupling utilizing a 11.7-ms 900 pulse and a 4-s recycle delay. 512 transients were acquired, and the free induction decay was processed with 25 Hz of line broadening. Each sample was allowed to equilibrate in the magnet for 30 minutes prior to signal acquisition.

14 For a DMPC/DHPC bicelle system the dimensions of the discs were calculated based on X-ray diffraction studies of the thickness of the DMPC hydrophobic chain and on a simple theoretical model of the mixed phospholipid bicelle (1). In this model, an expression

R = 0.5rp[p + (p2 + 8/q)1/2] (1) for the dimensions of the bicelle is calculated by analyzing a round bilayered center with radius R containing the length of the long chain phospholipids and a smaller rim portion containing the length of the short chain lipids with radius r, and p is the perimeter of the bicelle (1). For DMPC/DHPC bicelle system with a q ratio of 3.0, the diameter of the planar region has been estimated to be approximately 200 Å (1). This corresponds to a total bilayer thickness of approximately 40 Å for the DMPC/DHPC bicelle system (1, 8). For our study, the length of the hydrophobic bilayer of pure 16:0 and 18:0 phospholipids has been estimated by x-ray scattering techniques to be about 26 Å and 29.5 Å respectively in the La phase (24). In a similar fashion, we have used the same theoretical mixed phospholipid bicelle model (equation 1) and adjusted for the longer average length of the PSPC hydrophobic chain to estimate the dimensions of a PSPC/DHPC bicelle disc. Based upon this model, we estimate that the thickness for the lipid bilayer region of a PSPC/DHPC bicelle disc with a q ratio of 1.8 in the La phase should be approximately 46 Å and that the diameter of the planar section is approximately 160 Å (1). The 2H NMR spectrum in Figure 1.4(B) clearly demonstrates that the PSPC/DHPC bicelles are flipped in a similar fashion when Yb3+ ions are added to the bicelle matrix. Spectral resolution is increased because the ordering of the bilayer director has changed from Szz = -1/2 to

Szz = 1 causing the quadrupolar splittings to approximately double and spread out over a wider frequency range. For the well-resolved terminal methyl groups (CD3) the quadrupolar splitting is 3.8 kHz (perpendicular alignment) and 8.9 kHz (parallel alignment) for the PSPC/DHPC bicelle

o spectra at 50 C. The quadrupolar splittings of the plateau methylene (-CD2) deuterons located near the top of the phospholipid headgroups are 23 kHz (perpendicular alignment) and 54 kHz (parallel alignment). The quadrupolar splittings do not exactly double as expected from the change in the order parameter (1). For both the terminal methyl deuterons and the plateau methylene deuterons quadrupolar splitting, (DQ) increases by a factor of 2.3. This is expected because the addition of trivalent cations can increase the quadrupolar splittings up to 10%. For

15 comparison, DQ increases by a factor of between 2.1 and 2.2 for a DMPC/DHPC bicelle system (data not shown) (1). The individual terminal methyl deuterons are clearly resolved in Figures 1.4(A) and 1.4(B). The individual methylene deuterons are not fully resolved at the perpendicular orientation, but the resolution is significantly improved at the parallel orientation. 31P NMR spectroscopy can be used to study the different phases formed by phospholipid membranes because differences in the composition of the hydrocarbon chains can lead to

O O 31 dramatic changes in the chemical shifts and/or lineshape. Below Tm (40 C and 45 C), the P spectra of the PSPC/DHPC bicelle sample reveal two closely spaced isotropic peaks suggesting a

O 31 fluid-like phase with no orientational characteristics. Above Tm (50 C), the P NMR spectrum shown signs of magnetic alignment based upon comparisons with similar DMPC/DHPC bicelle spectra (16, 28). When DMPC/DHPC bicelle discs are magnetically aligned, the DMPC

2 31 phospholipid bilayer is in a liquid crystalline La-like phase (7, 16). Our H and P NMR spectra o suggest that the PSPC/DHPC bilayered discs are in a comparable La-like phase at 50 C. Thus, the PSPC/DHPC bilayers go directly from a fluid phase into a liquid crystalline phase around 50 oC. Similarly, temperature dependent 2H and 31P NMR studies of DMPC/DHPC phospholipid bilayers have indicated a highly ordered La-like phase above Tm, a mixed morphology consisting of an isotropic fluidic phase and a gel-like randomly dispersed phase slightly below Tm, and a completely fluid-like phase is observed below Tm(16). The 2H and 31P NMR spectra shown in Figures 1.4 and 1.5 reveal qualitative information on the overall degree of alignment of the phospholipid bilayers with respect to the direction of the static magnetic field. The broad peak at -7.0 ppm arises from 31P PSPC atoms and indicate that the phospholipid bilayer discs are not perfectly aligned with the magnetic field when compared to DMPC/DHPC 31P bicelle spectra (data not shown) (7, 16, 28). This agrees well with the lack of complete resolution of the methylene deuterons in the 2H NMR PSPC/DHPC bicelle spectra shown in Figure 1.4 (A). Although PSPC/DHPC phospholipid bilayers are not perfectly aligned they do provide an alternative membrane system for conducting solid-state NMR structural studies of integral membrane proteins. Depending upon the size, hydrophobicity, and/or activity of a particular membrane protein, a PSPC/DHPC bicelle system may be a good alternative to the standard DMPC/DHPC matrix. Our results indicate that longer chain phospholipids such as PSPC when mixed with detergents such as DHPC can be magnetically aligned and they can serve as potential model

16 systems for studying integral membrane proteins. Now that we have optimized magnetic alignment conditions (temperature and q ratio) for solid-state NMR studies, future investigations will be carried out on spin-labeled PSPC/DHPC bicelle systems containing 1-palmitoyl-2- stearoyl(n-DOXYL)-sn-glycerol-3-phosphatidylcholine with electron paramagnetic resonance (EPR) spectroscopy. The corresponding PSPC spin label is commercially available at positions 5, 7, 10, 12, and 16 (Avanti Polar Lipids, Inc., AL). The results of these studies will provide important dynamic information on the bicelle system and will be compared with parallel solid- state NMR results. Finally, the research presented in this chapter will enable us to carry out independent studies on the structural and dynamic properties of reconstituted integral membrane peptides inserted into magnetically aligned phospholipid bilayers utilizing both solid-state NMR spectroscopy and spin-label EPR spectroscopy (29, 30).

17 1.6. References

1. Prosser, R. S., Hunt, S. A., DiNatale, J. A., and Vold, R. R. (1996) Magnetically aligned membrane model systems with positive order parameter: Switching the sign of S-zz with paramagnetic ions, J. Am. Chem. Soc. 118, 269-270. 2. Prosser, R. S., Volkov, V. B., and Shiyanovskaya, I. V. (1998) Solid-state NMR studies of magnetically aligned phospholipid membranes: taming lanthanides for membrane protein studies, Biochemistry 76, 443-451. 3. Vold, R. R., and Prosser, R. S. (1996) Magnetically oriented phospholipid bilayered micelles for structural studies of polypeptides. Does the ideal bicelle exist?, J. Magn. Reson. B. 113, 267-271. 4. Howard, K. P., and Opella, S. J. (1996) High-resolution solid-state NMR spectra of integral membrane proteins reconstituted into magnetically oriented phospholipid bilayers, J. Magn. Reson. B. 112, 91-94. 5. Howard, K. P., and Prestegard, J. H. (1996) Conformation and dynamics of membrane- bound digalactosyldiacylglycerol, J. Am. Chem. Soc. 118, 3345-3353. 6. Losonczi, J. A., and Prestegard, J. H. (1998) Nuclear magnetic resonance characterization of the myristoylated, N-terminal fragment of ADP-ribosylation factor 1 in a magnetically oriented membrane array, Biochemistry 37, 706-716. 7. Sanders, C. R., Hare, B. J., Howard, K. P., and Prestegard, J. H. (1994) Magnetically- oriented phospholipid micelles as a tool for the study of membrane-associated molecules, Prog. NMR Spect. 26, 421-444. 8. Sanders, C. R., and Prosser, R. S. (1998) Bicelles: a model membrane system for all seasons?, Structure 6, 1227-1234. 9. Opella, S. J. (1997) NMR and membrane proteins, Nat. Struct. Biol. 4, 845-848. 10. Marassi, F. M., Ramamoorthy, A., and Opella, S. J. (1997) Complete resolution of the solid-state NMR spectrum of a uniformly N-15-labeled membrane protein in phospholipid bilayers, Proc. Natl. Acad. Sci. U. S. A., 8551-8556. 11. Marassi, F. M., and Opella, S. J. (1998) NMR structural studies of membrane proteins, Curr. Opin. Struct. Biol. 8, 640-648.

18 12. Tjandra, N., and Bax, A. (1997) Direct measurement of distances and angles in biomolecules by NMR in a dilute liquid crystalline medium, Science 278, 1697-1697. 13. Ottiger, M., Delaglio, F., Marquardt, J. L., Tjandra, N., and Bax, A. (1998) Measurement of dipolar couplings for methylene and methyl sites in weakly oriented macromolecules and their use in structure determination, J. Magn. Reson. 134, 365-369. 14. Ottiger, M., and Bax, A. (1998) Characterization of magnetically oriented phospholipid micelles for measurement of dipolar couplings in macromolecules, J. Biol. NMR 12, 361- 372. 15. Vold, R. R., Prosser, R. S., and Deese, A. J. (1997) Isotropic solutions of phospholipid bicelles: A new membrane mimetic for high-resolution NMR studies of polypeptides, J. Biol. NMR 9, 329-335. 16. Sanders, C. R., and Schwonek, J. P. (1992) Characterization of magnetically orientable bilayers in mixtures of dihexanoylphosphatidylcholine and dimyristoylphosph- atidylcholine by solid-state NMR, Biochemistry 31, 8898-8905. 17. Struppe, J., Whiles, J. A., and Vold, R. R. (2000) Acidic phospholipid bicelles: A versatile model membrane system, Biophys. J. 78, 281-289. 18. Deems, R. A. (2000) Interfacial enzyme kinetics at the phospholipid/water interface: Practical considerations, Anal. Biochem. 287, 1-16. 19. Larsson, K. (1994) Lipids-molecular organization, physical function and technical application, oily, glassglow.,5. 20. Tate, M. W., Shyamsunder, E., Gruner, S. M., and Damico, K. L. (1992) Kinetics of the lamellar inverse hexagonal phase-transition determined by time-resolved x-ray- diffraction, Biochemistry 31, 1081-1092. 21. Yamashita, Y., Kinoshita, K., and Yamazaki, M. (2000) Low concentration of DMSO stabilizes the bilayer gel phase rather than the interdigitated gel phase in dihexadecylphosphatidylcholine membrane, Biochimica et Biophysica Acta 1467, 395- 405. 22. Johannsson, A., Keightley, C. A., Smith, G. A., Richards, C. D., Hesketh, T. R., and Metcalfe, J. C. (1981) The effect of bilayer thickness and normal-alkanes on the activity of the (Ca2+/Mg2+)-dependent ATPase of sarcoplasmic-reticulum, J. Biol. Chem. 256, 1643-1650.

19 23. Haydon, D. A., and Hladky, S. G. (19972) Ion transport across thin lipid membranes: a critical discussion of mechanisms in selected systems., Quart. Rev. Biophys. 5, 187-282. 24. Lewis, B. A., and Engelman, D. M. (1983) Bacteriorhodopsin remains dispersed in fluid phospholipid-bilayers over a wide-range of bilayer thicknesses, J. Mol. Biol. 166, 203- 210. 25. Czerski, L., Sanders, C. R. (2000) Functionality of a membrane protein in bicelles, Anal. Biochem. 284, 327-333. 26. Picard, F., Paquet, M. J., Levesque, J., Belanger, A., and Auger, M. (1999) P-31 NMR first spectral moment study of the partial magnetic orientation of phospholipid membranes, Biophys. J. 77, 888-902. 27. Hemminga, M. A., and Cullis, P. R. (1982) P-31 NMR-Studies of oriented phospholipid multilayers, J. Magn. Reson. 47, 307-323. 28. King, V., Parker, M., Howard, K. P. (2000) Pegylation of magnetically oriented lipid bilayers, J. Magn. Reson 142, 177-182. 29. Garber, S. M., Lorigan, G. A., and Howard, K. P. (1999) Magnetically oriented phospholipid bilayers for spin label EPR studies, J. Am. Chem. Soc. 121, 32403241. 30. Mangels, M. L., Cardon, T. B., Harper, A. C., Howard, K. P., and Lorigan, G. A. (2000) Spectroscopic characterization of spin-labeled magnetically oriented phospholipid bilayers by EPR spectroscopy, J. Am. Chem. Soc. 122, 7052-7058.

20 CHAPTER 2

Elvis K. Tiburu, Paresh C. Dave, and Gary A. Lorigan, Solid-State 2H NMR Studies of the Effects of Cholesterol on Acyl Chain Dynamics of Magnetically Aligned Phospholipid Bilayers (2004), Magn. Reson. Chem., 32, 132-138

Paresh C. Dave, Elvis K. Tiburu, and Gary A. Lorigan, Calculating Order Parameter Profiles Utilizing Magnetically Aligned Phospholipid Bilayers for 2H Solid-State NMR Studies (2003), Solid State Nucl. Magn. Reson., 24, 137-149

21 Abstract We report the utilization of magnetically aligned phospholipid bilayers (bicelles) to study the effects of cholesterol in phospholipid bilayers for both chain perdeuterated DMPC, a partially

2 2 deuterated (a-[2,2,3,4,4,6- H6]) cholesterol, and stearic acid-d35 using H solid-state NMR spectroscopy. The quadrupolar splittings at 40 OC were 25.5 kHz and 37.7 kHz respectively for

2 2 the 2,4- Heq and 2,4- Hax deuterons of cholesterol when the bilayer normal of the discs were aligned perpendicular to the static magnetic field. The quadrupolar splittings were doubled when Yb3+ ions were added to flip the bicelles 900 such that the bilayer normal was collinear with the magnetic field. The results of our studies suggest that cholesterol is incorporated into the bicelle discs. For chain perdeuterated DMPC-d54, incorporated into DMPC/DHPC bicelles discs, the individual quadrupolar splittings of the methylene and methyl groups doubled going from the perpendicular alignment to the parallel alignment. Also, the presence of cholesterol increased

i the overall ordering of the acyl chains of the phospholipids. SCD calculations were extracted 2 directly from the H quadrupolar splittings of the chain perdeuterated DMPC and stearic acid-d35. i The order parameter, SCD calculations clearly indicated an overall degree of ordering of the acyl chains in the presence of cholesterol for both DMPC and stearic acid. We also noted disordering effects at higher temperatures. This study demonstrates the ease in which 2H order parameters can be calculated utilizing magnetically aligned phospholipid bilayers when compared to randomly dispersed membrane samples.

22 2.0 Introduction 2.1 The effects of cholesterol and stearic acid on phospholipid bilayers

The behavior and physical properties of phospholipid bilayers and biological membranes in the presence of cholesterol are of significant importance in biophysical and biological research (31-37). It has been possible to study the behavior of cholesterol directly in mechanically aligned membranes using 2H and 13C solid-state NMR spectroscopy (38). It has been shown using 2H-NMR data derived from specifically deuterated cholesterol incorporated into membranes that cholesterol is oriented perpendicular to the membrane surface (39). The rate of rotation as well as the internal motions of cholesterol have also been determined by 13C-NMR measurements revealing that the hydrophobic tail of cholesterol undergoes rapid rotational motion between different conformations (40). This is in contrast to the A-ring of cholesterol which maintains some flexibility. The orientation of cholesterol perpendicular to the bilayer normal (with the OH group pointing to the hydrophilic region of the lipids) has previously been confirmed by X-ray diffraction and neutron diffraction studies (41). The data from such studies indicate that cholesterol is located such that the polar hydroxyl group is in the immediate vicinity of the headgroup of the phospholipid acyl chain (40-42). Cholesterol is known to induce an overall increase in ordering for the phospholipids in the liquid-crystalline phase (La). This property is often referred to as the condensing effect of cholesterol which occurs at a temperature above the gel-to-liquid crystalline phase transition of the phospholipids and a disordering effect below the gel-to-liquid crystalline phase transition (40, 42-46). In a pure phospholipid membrane, the average orientation of the hydrocarbon chains defines the structure of the membrane whereas the fluctuations in the polar head groups impose different segmental motions that define the reorientational dynamics of the bilayer (47). These properties can be addressed if the segmental order parameters and the T1 and T2 relaxation times are measured using deuterated phospholipids. Binary mixtures of lipid-cholesterol composition have been investigated in detail using deuterium solid-state NMR spectroscopy (32, 39, 40, 47- 49). Also, several papers have discussed the application of molecular dynamic methods to study the lipid bilayer properties (50, 51). Pastor and co-workers have used a mean field approach to model the Brownian dynamics of the interior of a lipid bilayer (50). Also, Scott and co-workers

23 used the Monte Carlo approach to study the effects of cholesterol on the acyl chain dynamics (51). Although extensive studies of the effects of cholesterol on phospholipids have been carried out, most of these solid-state NMR studies have utilized unoriented or mechanically aligned membrane samples. The use of multilamellar unoriented vesicles in solid-state NMR spectroscopic studies give rise to broad spectra as well as poor signal-to-noise. Though vesicles are in a biologically relevant environment, the broad spectra from such systems may render them unsuitable for structural or dynamic studies. To obtain pertinent structural and/or dynamic information entails deconvolution or the so called de-Pakeing to resolve multiple splittings in the 2H NMR spectra (49, 52). However, the quality of de-Paked spectra will depend on the signal- to-noise ratio of the experimental powder spectra. Also, structural information can be obtained from mechanically aligned bilayers which may not be in a biological relevant environment when compared to magnetically aligned phospholipid bilayers (22-31). Recently, bicelles have been developed that mimic the properties of biological membranes (53-61). Bicellar solutions may consequently be used as a macro-ordered matrix that permits residual dipolar coupling to be used in the NMR structure determination of proteins. The use of lyotropic liquid crystals has been previously proposed for structural biology work, but phospholipid bicelles have proved particularly valuable for this purpose because enzymes have been shown to maintain activity in this environment (58). High resolution NMR techniques are now routinely employed to study the structure of complex macromolecules in solution (58, 60, 62-64). An alternative approach to structural studies of membrane macromolecules is the determination of the orientational and structural properties via solid-state NMR spectroscopy (64). Aligned bicelles allow the acquisition of high resolution solid-state NMR spectra of comparable quality to those obtained with samples that have been mechanically oriented on glass or polymer solids (54, 65). The use of bicelles to align membrane proteins is attractive for several reasons, but in particular because a wide variety of peptides and proteins can be easily reconstituted into bicelle discs while their biological activity is retained (58, 66). There has been considerable interest in the development of biomembrane-mimetic materials that can provide an ordered matrix in which biomolecules such as proteins and peptides can be spatially organized (58, 61, 66, 67). For example, polymer-grafted membrane liquid crystalline gels, consisting of a quaternary mixture of a phospholipid, a lipopolymer comprising

24 low-molecular-weight poly(ethylene) oxide (PEG) terminally grafted onto the phosphate head group of phospholipids, and a co-surfactant dispersed in water has been reported (53, 54). The magnetically induced alignment process has been studied by using 2H solid-state NMR spectroscopy. Moreover, our group has incorporated a small amount of PEG polymer into the DMPC/DHPC bicelle discs for EPR and NMR experiments (68, 69). The addition of the PEG polymer has increased the stability of the bicelle discs and is used routinely for low field EPR studies. Interestingly, various liquid crystalline materials have been developed in recent years for a wide variety of different technological applications (for e.g. liquid crystalline displays) (70). Liquid crystals consist of rod like shapes composed of Schiff’s base complexes, b- diketonates, or disc like shapes such as those prepared from phospholipid bicelles, porphyrins, and phthalocyanine complexes. They have orientational and positional dependence characteristics. These properties change with temperature and can be studied using 2H order parameter profiles. Very little information is known about mixtures that include long-chain and short-chain phospholipids and cholesterol in the presence of lanthanide ions for structural and dynamic studies. Several articles have discussed the interactions between cholesterol and phospholipid in bilayers, while few reports mention the effect of cholesterol on the bicelle system (45, 69, 71, 72). Our present report is exploring the possibility for easily calculating 2H order parameter profiles for different types of oriented systems. The initial objective of the present study was to

utilize the spontaneous magnetic alignment of a DMPC/DHPC/Chol-d6 bicelle system to describe the orientation of cholesterol in the bicelle disc and to evaluate the behavior of cholesterol with temperature in the presence and absence of lanthanide ions. The oriented deuterium spectrum does not require spectral deconvolution or the so called de-Pakeing, to resolve multiple splittings in the 2H NMR spectra (52, 73). The quadrupolar splittings provide a direct measure of the orientational and dynamic ordering of C-D bonds (SCD) embedded in anisotropic media. Also,

DMPC/DHPC/DMPC-d54 bicelle system in the presence of cholesterol will also be carried out and the effects of temperature on the ordering of the acyl chain evaluated from the quadrupolar splitting obtained from the aligned bicelle system. The results will be compared with results derived from molecular dynamic studies and solid-state 2H NMR studies of cholesterol in mechanically aligned and randomly dispersed phospholipid bilayers of DMPC (46, 50, 51, 74).

25 The second objective is to investigate the order parameter profiles obtained by inserting the isotopically labeled NMR probe stearic acid-d35 into magnetically aligned phospholipid bilayers. Previous studies have determined that deuterated fatty acids make excellent, non-perturbing NMR spin probes (75, 76). We incorporated this 2H-labeled fatty acid into a DMPC/DHPC bicelle system (q=3.5) and examined the results via solid-state NMR spectroscopy. Also, cholesterol has been added to the DMPC/DHPC/fatty acid bicelle system to examine the effect cholesterol has on the physical properties of the phospholipid bilayers using stearic acid-d35 as isotopic probe (37, 39, 46, 74, 77). We carried out temperature-dependent solid-state NMR experiments at two different cholesterol concentrations (0 mol% and 10 mol% with respect to DMPC). 2H NMR experiments were conducted in the presence of Yb3+. In the present work, we are focusing on three main points (1) The calculation of order parameters using quadrupolar splittings obtained directly from magnetically aligned 2H NMR spectra, (2) the effects of cholesterol and temperature on a DMPC/DHPC/ Yb3+/fatty acid bicelle system (3) the future application of this technique to non-biological systems.

2.2 Materials 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), 1,2-dihexanoyl-sn-glycero-3- phospho -choline (DHPC), deuterated 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC-d54) and stearic acid-d35 were purchased from Avanti Polar Lipids (Alabaster, AL). a-[2,2,3,4,4,6-

2 H6] cholesterol was purchased from Cambridge Isotope Laboratories (Andover, MA). All phospholipids were dissolved in chloroform and stored at –20 0C prior to use. Ytterbium(III) chloride hexahydrate and N-[2-hydroxyethyl]piperazine-N’-2-ethanesulfonic acid (HEPES) were obtained from Sigma-Aldrich. Deuterium-depleted water was obtained from Isotec (Miamisburg, OH). Cholesterol 95% pure was obtained from Alfa Aesar (Ward Hill, MA).

2.3. Methods 2.3.1 Sample Preparation The standard DMPC/DHPC bicelle samples, consisting of 25% (w/w) phospholipid to solution with a q = 3.5, were made in 25 mL pear-shaped flasks. The q ratio is defined as the

2 molar ratio of DMPC to DHPC. DMPC, DHPC and DMPC-d 54 (or a-[2,2,3,4,4,6- H6] cholesterol) were mixed together in chloroform at molar ratios of 0.1/0.03/0.01, respectively. 10

26 mol% nondeuterated cholesterol with respect to the DMPC long chain phospholipids was also added to the sample. The chloroform was removed using a rotavaporator and the sample placed under high vacuum overnight. For deuterated cholesterol studies 5% mol non-deuterated cholesterol and 5% mol of the deuterated cholesterol was used, bringing the total cholesterol concentration to 10% mol. The mixture in the flask was rotovaped down to remove chloroform and the sample was placed under high vacuum overnight. The stearic acid bicelle sample was also prepared similarly. The following day, a 100 mM HEPES buffer of pH 7.0 was prepared using deuterium- depleted water and the appropriate amount was added to the flask. The flask was then vortexed briefly, sonicated with a FS30 (Fisher Scientific) bath sonicator with the heater turned off. The sample was subjected to several freeze (77K)/thaw cycles at room temperature until the dispersion was homogenous and all the air bubbles were removed. Typically, the total mass of the prepared bicelle sample was 300 mg. An aqueous solution of YbCl3 (100 mM) was prepared fresh with deuterium-depleted water and 10 mL was added to the bicelle sample. The bicelle sample was placed in a NMR flat bottom tube with a 21-mm long and 5-mm o.d. on ice via a Pasteur pipette. To monitor bicelle alignment with 2H NMR spectroscopy, a 10:1 mole ratio of nondeuterated and chain deuterated phospholipids (DMPC-d54 ) and stearic acid-d35 were used in this study.

2.3.2 NMR Spectroscopy All solid-state NMR experiments were carried out on a modified Bruker AVANCE 7.05 T narrow bore 300/54 magnet configured to conduct high-power solid-state NMR studies. The solid-state NMR spectra were gathered with a static double-tuned 5 mm round-coil solid-state NMR probe purchased from Doty Scientific. 2H NMR spectra were recorded at 46.07 MHz using a standard quadrupole-echo pulse sequence (3.0-ms 900 pulses, 45-ms interpulse delay, 5.12-ms acquisition time, 0.4-s recycle delay, and a 150-kHz sweep width). Typically 20,000 scans were accumulated in the quadrature detection mode. An exponential line broadening of 300 Hz was applied to the free induction decay before Fourier transformation. We carried out the experiments in the temperature range from 25 to 65 0C in 5 0C increments.

27 2.3.3 Order Parameter Calculations Order parameters depend upon several averaging modes provided by intramolecular, intermolecular, and collective motions. The relative ease in deuterating molecules and using solid-state NMR spectroscopy has enabled C-D bond order parameter profiles (SCD), to be easily calculated. This C-D bond profile has been proposed by Seelig and co-workers as a measure of the thickness and fluidity of the hydrophobic phospholipid bilayer (78, 79). SCD describes local orientational or dynamic perturbations of the C-D bond vector from its standard state due to perturbations of DMPC phospholipid conformations or dynamics (80, 81). The segmental order parameters were calculated using the equations given by Prosser et al. (56). In our study, the

i i order parameter defined as SCD is analogous to Slp as previously defined by Prosser and co- workers (56).

i i i SCD = Dl / Dp (1)

i th where SCD is the order parameter for a deuteron attached to the i carbon of the phospholipid i th i acyl chain, Dl is the observed quadrupolar splitting for a deuteron attached to i carbon and Dp is the splitting that would be observed for a stationary deuteron in a C-D bond pointing along the direction of the external magnetic field. Order parameters were calculated assuming a quadrupole coupling constant of e2qQ/h = 168 kHz (56). In the case where the bicelle discs are aligned parallel to the static magnetic field (addition of Yb3+), equation (1) reduces to:

i i SCD = Dl / 252 (2)

i 2 when Dp is equal to 3/2 e qQ/h. Thus, the corresponding order parameters for the individual C-

D methylene groups and the terminal methyl groups of the acyl chains of DMPC-d54 and stearic acid-d35 incorporated into the aligned bicelle systems were directly evaluated from the quadrupolar splittings of the 2H NMR spectra. The 2H peaks in the NMR spectra were assigned based upon the dynamic properties of the individual CD2 and CD3 groups. The quadrupole splittings for the CD3 methyl groups at the end of the phospholipid acyl chain are the smallest 2 and closest to 0 kHz. In DMPC-d54, the next smallest splitting was assigned to the H attached to C-13 and so forth along the phospholipid acyl chain. The quadrupolar splittings for the

28 deuterons in the plateau region were estimated by integration of the last broad peak according to the literature (32, 82). The order parameters calculated for the CD3 quadrupolar splittings have been multiplied by three (39, 75). The carbon numbering system for the deuterated NMR probe stearic acid-d35 is given in Figure 2.1.

D2 D2 D2 D2 D2 D2 D2 D2 HOOC C C C C C C C C

C C C C C C C C CD3 D2 D2 D2 D2 D2 D2 D2 D2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

stearic acid-d35

2 Figure 2.1 Structure and numbering of deuterium labeled stearic acid-d35 used for this H NMR study.

2. 4 Results and Discussion 2.4.1. Cholesterol dynamics in phospholipid bilayers In addition to the behavior of cholesterol in a phospholipid membrane, we must also know the location of cholesterol in such a membrane. In this study a standard bicelle with a molar ratio of DMPC to DHPC equal to 3.5 (q = 3.5) was prepared for solid-state 2H NMR studies. We incorporated deuterated cholesterol (a-[2,2,3,4,4,6-2H]-cholesterol) into a bicelle sample and studied the orientation using 2H solid-state NMR spectroscopy as shown in Figure 2.2. Relative changes in the spectral width were observed from 25 OC to 65 OC. At 25 OC a single line resonance at around 0 kHz representing an isotropic peak was observed due to the random fast motion of the unoriented phospholipid bilayers (data not shown). 2H powder pattern spectra were also observed at 60 OC and were presumed to be caused by the overlap of different orientations of the phospholipids (data not shown). Figure 2.2 displays the 2H NMR spectra of labeled a-cholesterol incorporated into magnetically aligned phospholipid bilayers oriented perpendicular and parallel (Yb3+) with respect to the magnetic field at 40 OC. Figure 2.2(A) shows the perpendicular alignment where the inner doublets with the smallest quadrupolar splittings (1.9 kHz) represent the 6-deuterons due to their fast motion. The second smallest quadrupolar splittings (25.5 kHz) represent the 2,4-

29 H3C 2 CH3 [2,4- H]eq CH3 2 CH3 [2,4- H]ax CH3 D [6-2H] 1 2 2 D (A) 3- H 3 5 HO 4 6 D D D D

(B)

60 40 20 2 0 -20 -40 -60 H (kHz)

Figure 2.2 2H NMR spectra of deuterated-[2,2,3,4,4,6] cholesterol incorporated into magnetically aligned DMPC/DHPC phospholipids bilayers at 40 OC. (A) In the absence of lanthanide ion, Yb3+, the bicelles aligned with their bilayer normal perpendicular to the magnetic field tensor. (B) In the presence of lanthanide ion,Yb3+, the bicelles aligned with their bilayer normal parallel to the magnetic field tensor. The quadrupolar splitting in the parallel alignment is twice that of the perpendicular aligment. equatorial deuterons on the fused ring of cholesterol and the last set of doublets represent the quadrupolar splitting (37.7 kHz) of 2,4-axial deuterons on the fused ring of cholesterol as well

0 as the deuteron at the C3-position. Figure 2.2 (B) displays the parallel alignment at 40 C with similar results except that the quadrupolar splittings were slightly twice that of the perpendicular alignment due to the effect of Yb3+ lanthanide ions. The linewidths of the resonance lines in Figure 2.2 (B) are broader than those in Figure 2.2 (A). One explanation for this is that the lanthanide ion reduces the wobbling of the cholesterol in the phospholipid bilayer due to tighter

30 packing of the acyl chain next to the rigid sterol ring system of the cholesterol and thus, decreasing the correlation times but still causing ordering. These results demonstrate that cholesterol was successfully incorporated into the bicelle sample between the temperatures of 35 OC and 55 OC and that the cholesterol molecule is embedded inside the bicelle membrane. As depicted in Figure 2.2(A), the quadrupolar splitting at 40 OC was 25.5 kHz and 37.7 kHz

2 2 respectively for the 2,4- Heq and 2,4- Hax at the perpendicular alignment. This was comparable with Dufourc’s data for the deuterated cholesterol incorporated in DMPC bilayers (9). However, it is worthwhile to note that our values were lower because we used 10% mol cholesterol in the sample preparation instead of 30% mol with respect to DMPC (39). Also, the wobbling of the bicelle disks along the bicelle normal may cause the spectra to shift in, thus reducing the quadrupolar splitting. Recently, Trouard and coworkers investigated the effects of cholesterol on DMPC bilayers using 2H solid-state NMR studies (47). The quadrupolar splitting was doubled upon the addition of 1:1 mol cholesterol to the DMPC phospholipids. Our results support the ascertion that 10% cholesterol affects the ordering of the acyl chains on phospholipid bilayers. Figure 2.2(B) represents the parallel alignment of the bicelle discs in the presence of Yb3+ ions.

O 2 The quadrupolar splitting at 40 C was 53.6 kHz and 79.5 kHz respectively for 2,4- Heq and 2,4- 2 Hax at the parallel alignment. We can clearly see that the quadrupolar splitting for all the deuterons are approximately doubled. The doubling of the quadrupole splitting indicated that the OH group of cholesterol is directed towards the aqueous surface whereas the hydrophobic portion is directed into the hydrophobic core of the acyl chains of the bicelles. In order to probe the effects of cholesterol in the bicelle sample we used the standard

2 bicelle q ratio of 3.5 with DMPC-d54 as an isotopic probe for H solid-state NMR studies. Figure 2.3 (A) shows the alignment of the bicelle discs sample at 40 OC with the bilayer normal perpendicular to the static magnetic field due to the negative magnetic susceptibility anisotropy (Dc< 0) of the bicelle discs. The dotted spectra represent the aligned bicelle samples without cholesterol, whereas the solid spectral lines represent aligned spectra with 10 mol% cholesterol with respect to DMPC. The two sharp peaks close to the center represent the methyl deuterons whereas the outer broad peaks describe methylenes in the plateau region. The quadrupolar splitting of the broad outer peak was 19.6 kHz without cholesterol and 23.8 kHz in the presence of cholesterol. Upon addition of Yb3+ ions to the sample, the bicelle discs were

31 (A)

(B)

40 20 2 0 -20 -40 H (kHz)

2 Figure 2.3 H NMR spectra of DMPC-d54 incorporated into magnetically aligned DMPC/DHPC phospholipid bilayers at 40 OC (A) In the absence of the lanthanide ion Yb3+, the bicelles aligned with their bilayer normal perpendicular to the magnetic field tensor. The dotted (red) spectra represent the alignment in the absence of 10 mol% cholesterol and the solid line represents the alignment in the presence of 10 mol% cholesterol. (B) In the presence of the lanthanide ionYb3+, the bicelles aligned with their bilayer normal parallel to the magnetic field tensor. The dotted (red) spectra represent the alignment in the absence of 10 mol% cholesterol and the solid line represents the alignment in the presence of 10 mol% cholesterol. flipped such that the bilayer normal was parallel to the magnetic field as shown in Figure 2.3(B). The quadrupolar splitting of the outer broad peak was 41.7 kHz without cholesterol and 47.4 kHz in the presence of 10 mol% cholesterol with respect to DMPC. The quadrupolar splittings were increased by a factor of 2 in the parallel alignment when compared to the perpendicular aligned bicelles represented in Figure 2.3(A). Also, the spectra are shifted out at 40 OC in the presence of

32 cholesterol indicating an overall increase in ordering of the acyl chain in the liquid crystalline phase. This is because in the presence of cholesterol, there is restricted motion of the acyl chains causing the phospholipids in the bicelles in the liquid crystalline state to be more ordered and shifted outward when compared to the 2H spectra without cholesterol. Figure 2.4 shows the 2H solid-state NMR spectra of DMPC/DHPC bicelles containing

+3 chain deuterated DMPC-d54 in the presence of 5 mol% Yb ions with respect to DMPC as a function of temperature.

0 % Cholesterol 10 % Cholesterol

35 ºC

40 ºC

45 ºC

50 ºC

55 ºC

60 ºC

40 20 0 -20 -40 2 H (kHz)

2 Figure 2.4 H NMR spectra of DMPC-d54 incorporated into magnetically aligned DMPC/DHPC phospholipid bilayers in the presence of lanthanide ion Yb3+ as a function of temperature. (A) The dotted spectra represent the alignment in the absence of cholesterol. (B) The solid line represents the alignment in the presence of 10 mol% cholesterol with respect to DMPC.

The dotted spectra represent 0 mol% cholesterol in the bicelle sample and the solid spectra represent 10 mol% cholesterol in the bicelle sample. The spectra are all characteristic of Yb3+- doped aligned bicelles in the liquid crystalline phase. The effect of temperature was observed at the parallel alignment as shown in Figure 2.4. A general decrease in the degree of ordering with increasing temperature in the presence or absence of cholesterol was observed. At higher

33 temperatures, the amplitude of allowed motions of the acyl chain increases leading to a decrease in spectral width. Also, at higher temperatures, the hydrophobic and hydrophilic interactions in the presence of cholesterol becomes less significant and therefore cholesterol tends to disturb the parallel packing of the acyl chain. At 35 OC, the acyl chains are more ordered in the presence of cholesterol. It was observed that the difference in quadrupolar splitting becomes smaller with increasing temperature as revealed in Figure 2.4. This observation can be explained in terms of the tilt angle between the long molecular axis of the cholesterol molecule and the bilayer normal. This tilt angle increases with increasing temperature at the liquid crystalline state, thus, causing changes in the cholesterol concentration within the hydrophobic region of the membrane. Molecular dynamic studies also support this observation (50, 51). The changes in the chain order can better be characterized through the calculations of the

i i order parameter, SCD of the C-D bond of the acyl chain. SCD reflects the properties internal to the bicelle phospholipids, including disorder of individual bicelles normal relative to the sample director, motion of individual phospholipids molecules, and the orientation of particular C-D

i bonds relative to the molecular axis. The SCD was directly calculated from the quadrupolar splittings obtained from the parallel alignment spectra in Figure 2.4. We are presenting the

i absolute values for the SCD whether samples have positive or negative order with respect to the i magnetic field. SCD was plotted as a function of carbon number of the phospholipid acyl chain at different temperatures. The results are displayed in Figure 2.5 for the parallel alignment with Yb3+ ions and with 0 mol% cholesterol (Figure 2.5(A)) and 10 mol% cholesterol (Figure 2.5(B)).

i The error associated with the measurement of the quadrupolar splitting used for SCD calculation i is ±0.005. With such a small quadrupolar splitting, the error bar associated with each SCD measurement was so small that it did not reflect on Figure 2.5. Similar data were also extracted for the perpendicular alignment (data not shown). Figure 2.5 reveals that the order parameters for the C-D bond decreases in the acyl chains from the methylenes close to the headgroup of the phospholipids to the end of the chain towards CD3 at all temperatures studied. In Figure 2.5(A) the acyl chains are more ordered at 35 OC and less ordered at 45 OC. The quadrupolar splitting decreases with increasing temperature indicating a decrease in overall order. The same trend is seen in Figure 2.5(B) as evidence in the methylene groups close to the headgroup being more ordered in the presence of cholesterol.

34 (A) 0 60 C 0 0.25 55 C 0 50 C 0 0.20 45 C 0 40 C 0 0.15 35 C

0.10

CD 0.05

0.00 (B)

0.25 Order parameter, S

0.20

0.15

0.10

0.05

0.00

0 2 4 6 8 10 12 14

Carbon no.

1 Figure 2.5 Temperature-dependent S CD order parameter profiles for the acyl

chains of DMPC-d54 incorporated into magnetically aligned DMPC/DHPC phospholipid bilayers in the presence of Yb3+. (A) In the absence of cholesterol corresponding to the 2H solid-state NMR spectra in Fig. 4 (dotted line) and (B) in the presence of 10 mol% cholesterol with respect to DMPC corresponding to the spectra in Fig. 4 (solid line). The error associated with each data point measurement is negligible.

i The effect of cholesterol is more pronounced close to the headgroup with the characteristic SCD i at carbon position 2 being much higher than the SCD at carbon position 13 as shown in Figure 2.5. This observation indicates that cholesterol increases the order of the plateau region with less effect on the properties of the bilayer at the center and ends of the acyl chain. Increasing the motional order of the plateau region corresponds to a decrease in the trans-gauche isomerization in the plateau region. The increased degree of ordering at the plateau region suggests that the planar rigid steroid ring is in contact with the acyl chain preventing the lateral diffusion of the acyl chains. Similar effects were observed for the perpendicular alignmnent (data not shown).

35 While the order parameter data indicate the motional ordering of the acyl chain, deuterium NMR studies cannot distinguish the slight order parameter differences between sn-1 and sn-2 acyl chains of DMPC phospholipid.

2.4.2 Magnetic alignment of DMPC/DHPC bicelles investigated by stearic acid-d35 The bicelle system under investigation is composed of DMPC and DHPC phospholipids at a molar ratio of 3.5:1. Solid-state NMR studies have indicated that the addition of fatty acids in small concentrations to phospholipid bilayers has no marked effects on the bilayers, this is in contrast to the dramatic effects of other systems such as cholesterol and long chain alcohols (37,

75). Our goal is to calculate order parameters utilizing stearic acid-d35 as a NMR probe incorporated into a DMPC/DHPC bicelle system. We have carried out 2H NMR experiments on

3+ the stearic acid-d35 incorporated into bicelles in the presence of Yb . Stearic acid incorporates into bicelle discs such that the acyl chains of the DMPC phospholipids and the stearic acid are collinear with respect to each other. Thus, the molecular axis of stearic acid-d35 is parallel to the bilayer normal axis of the bicelle discs and the static magnetic field. Figure 2.6 represents the 2H

3+ NMR spectra for stearic acid-d35 incorporated into a DMPC/DHPC/Yb bilayer with 0 mol% cholesterol as a function of temperature. The spectra are all characteristic of well-aligned phospholipid bilayers. As the temperature increases, the resolution of the 2H peaks increases because the Yb3+-doped bicelle discs are in the magnetically aligned liquid crystalline smectic phase (56). The quadrupolar splittings are increased by a factor of 2 when compared to the splittings observed in the case where the bilayer normal is aligned perpendicular to the static magnetic field (data not shown).

36 0 35 C

0 40 C

0 45 C

0 50 C

0 55 C

0 60 C

40 20 0 -20 -40 2 H (kHz)

2 Figure 2.6 H NMR spectra of stearic acid-d35 embedded inside magnetically aligned DMPC/DHPC phospholipid bilayers in the presence of Yb3+ investigated as function of temperature. No cholesterol was added to the sample.

The increase is due to the parallel orientation of the phospholipid bilayers with respect to the magnetic field. In Figure 2.6, the sharp quadrupolar splittings closest to zero kHz represent the terminal CD3 groups, while the others represent the methylene (CD2) groups along the acyl 2 chain. In the case of stearic acid-d35, seventeen different peaks should be resolved in the H

NMR spectra (sixteen from CD2 and one from CD3), but only thirteen are resolved.

37 0 35 C

0 40 C

0 45 C

0 50 C

0 55 C

0 60 C

40 20 0 -20 -40 2 H (kHz)

2 Figure 2.7 H NMR spectra of stearic acid-d35 embedded inside magnetically aligned DMPC/DHPC phospholipid bilayers in the presence of Yb3+ and 10 mol% cholesterol with respect to DMPC investigated as a function of temperature.

Figure 2.7 illustrates the 2H quadrupole echo NMR spectra of DMPC/DHPC bicelles containing chain-deuterated stearic acid-d35 and 10 mol% cholesterol with respect to DMPC. The spectra (Figures 2.6 and 2.7) clearly indicate that by increasing the temperature the resolution of the 2H acyl chain peaks increases. The addition of cholesterol also has been shown to increase the phase transition temperature of the bicelle discs (37). Also, the deuterium quadrupolar splittings increase with the addition of cholesterol into the bicelle discs. This result suggests that in the

38 presence of cholesterol, the fluidity of the bicelle discs decreases and the acyl chains are more ordered in the liquid crystalline phase. An isotropic component is observed at 45 0C and above in the absence of cholesterol (Figure 2.6), while the isotropic peak is absent in the presence of 10 mol% cholesterol (Figure 2.7). On the NMR timescale, small molecules undergo fast isotropic tumbling due to the low viscosity of the medium and can give rise to an isotropic peak. It is difficult to pinpoint the exact origin of the isotropic component in the NMR spectra. However, the slight presence of mixed micelles or small discoidal objects could cause this.

2.4.3 Calculating order parameters for stearic acid-d35 inserted into magnetically aligned phospholipid bilayers Figure 2.8 depicts the orientational order parameter profiles for the fatty acid chain derived from the spectra in Figures 2.6 and 2.7 for (A) 0 mol% and (B) 10 mol% cholesterol with respect to DMPC as a function of temperature. SCD was calculated using equation 2 as a function of the carbon number i. The numbering scheme for stearic acid-d35 is shown in Figure 2.1. SCD reflects the internal dynamic properties of the stearic acid incorporated to the bicelle discs. The magnitude of the order parameters indicates that the bicelle discs are in the liquid crystalline phase. Figure 2.8 reveals that the order parameters for the C-D bond decrease in the acyl chains from the top of the fatty acid to the CD3 end. As generally observed for the acyl chains of the phospholipid bilayers, two regions can be distinguished: a plateau region from C1 to C7 with a maximal value of SCD = 0.23 and a region in which SCD gradually decreases from a value of approximately 0.17 to a final value of 0.03 for the terminal C17 methyl group. The values for SCD for stearic acid-d35 in a DMPC/DHPC bicelle system closely resemble previous studies of stearic acid-d35 incorporated into unoriented monounsaturated and polyunsaturated phosphocholine bilayer systems (32). The data indicates that there is more disorder and motion in the center and at the ends of the acyl chains when compared to the head group regions. Also, the order parameter profiles indicate that by increasing the temperature, the value of SCD decreases and the mobility of the acyl chains increases. The effect temperature has on the order parameters of stearic acid-d35 agrees well with the previous report for stearic acid-d35 incorporated into egg lecithin (76).

39 (A)

0.25 0 35 C

CD 0 0.20 40 C 0 45 C 0 0.15 50 C 0 55 C 0 0.10 60 C

Order Paremeter, S

0.05

0.00

(B)

0.25 CD 0.20

0.15

0.10 Order Parameter, S

0.05

0.00

0 2 4 6 8 10 12 14 16 18 20 Carbon no.

Figure 2.8 Temperature-dependent SCD order parameter profiles for the acyl

chains of stearic acid-d35 incorporated into magnetically aligned DMPC/DHPC phospholipid bilayers in the presence of Yb3+. (A) In the absence of cholesterol corresponding to the 2H solid-state NMR spectra in Figure 6 and (B) in the presence of 10 mol% cholesterol with respect to DMPC corresponding to the spectra in Figure 2.7.

X-ray neutron diffraction, and 2H NMR spectroscopic studies indicate that cholesterol embeds into phospholipid bilayers in such a way that its polar hydroxyl group is located in the aqueous phase and the hydrophobic steroid ring is oriented parallel to and buried in the hydrocarbon chains of the phospholipids (70). Thus, the acyl chains of stearic acid-d35 are parallel to the normal of the long axis of cholesterol. The effect of cholesterol in the bilayer is well reported in the literature (37). Based on the SCD profiles for 0 mol% and 10 mol% cholesterol with respect to

40 DMPC, acyl chain mobility of stearic acid-d35 decreases with the addition of 10 mol% cholesterol to the bilayer. Distal deuterons have a higher mobility when compared to the deuterons close to the head group as reflected in the order parameter profiles in Figure 2.8. It should be noted that the temperature variation has no significant effect on the order parameter profiles of the distal deuterons. Figures 2.8(A) and (B) suggest that the order parameters for the acyl chains decrease slowly from 35 0C to 45 0C in the absence of cholesterol when compared to the 10 mol% cholesterol sample. For simplicity, Figure 2.9(A) shows the 2H NMR spectra of stearic acid-d35 in the absence and in the presence of 10 mol% cholesterol with respect to DMPC at 40 0C. The spectra clearly indicate that the quadrupolar splittings for the C-D bonds increase upon addition of cholesterol to the phospholipid bilayers. Figure 2.9(B) displays the corresponding order parameter profiles for the 2H NMR spectra in Figure 2.9(A). The data indicates that the acyl chains of the deuterated fatty acid are well ordered in the presence of cholesterol in the liquid crystalline phase at 40 0C. The order parameter profiles of the last three carbons of the fatty acid with and without cholesterol are nearly identical. Thus, cholesterol does not significantly alter the dynamic properties of the CD2 and CD3 groups near the end of the acyl chain of the fatty acid. This is not surprising because cholesterol is located near the top of the phospholipid head groups and the cholesterol rings are in close proximity to positions 2-10 of the acyl chains of the DMPC phospholipids.

2.5 Conclusions. In this chapter magnetically aligned bicelle samples were used to demonstrate that cholesterol is located within the hydrophobic core of the bicelle membrane. The effects of cholesterol on the acyl chain ordering was also demonstrated by extracting quadrupolar splitting from the aligned spectra and also from the order parameter calculations using the quadrupolar splitting from the aligned spectra. The ordering effect of cholesterol on the acyl chain of the phospholipids was observed without actually affecting the spectral shape.

41 (A) 10% Chol. 0% Chol.

40 20 2 0 -20 -40 H (kHz)

0.25 (B) 10% Chol. 0% Chol.

0.20 CD

0.15

0.10 Order Parameter, S

0.05

0.00

0 5 10 15 20 Carbon no.

2 Figure 2.9 (A) H NMR spectra of stearic acid-d35 incorporated into magnetically aligned DMPC/DHPC phospholipid bilayers at 40 0C. The boldface line spectrum contains 10 mol% cholesterol with respect to DMPC and the dashed line spectrum contains no cholesterol. (B) Chain-order parameter profiles corresponding to the

stearic acid-d35 spectra represented in (A).

It was shown that the plateau region had the most ordering effect in the presence of cholesterol, indicating that the interaction of cholesterol with the phospholipids is mainly close to the headgroup. These results agree well with previously studied systems using both oriented and

42 unoriented samples for 2H solid-state NMR and molecular dynamic techniques. These results confirm that magnetically aligned phospholipid bilayers can serve as an excellent model membrane system. Also, we can extract quadrupolar splittings for SCD calculations much easier without using de-Pakeing algorithms. Quadrupolar splittings for individual deuterons were easily measured from the well

2 resolved H NMR spectra of stearic acid-d35 incorporated into magnetically aligned bilayers. Our results agree with previous studies that have utilized a more complicated de-Pakeing method for measuring the quadrupolar splittings of stearic acid-d35 inserted into unoriented phospholipid bilayers. The SCD order parameter profiles for stearic acid-d35 suggest that the mobility of the acyl chains increases with increasing temperature. Interestingly, the addition of cholesterol to the bicelle system decreases the fluidity of the membrane, except near the end of the acyl chains. This study indicates that non-perturbing NMR probes can be utilized for studying the structural and dynamic properties of membrane systems. Also, liquid crystalline molecules in a mesophase can be aligned in the presence of external magnetic field. Molecules in a crystalline lattice possess both orientational and positional order. The present investigation suggests that the properties of non-biological liquid crystalline materials can be more easily studied by using solid-state 2H NMR spectroscopy through the direct measurement of quadrupolar splittings of oriented systems.

43 2.6 References 1. Gabdoulline, R. R., G. Vanderkooi, and C. Zheng. (1996) Comarision of the structures of dimyristoylphosphatidylcholine in the presence and absence of cholesterol by molecular dynamics simulations, J. Phys. Chem. 100, 15942-15946. 2. Huster, D., K. Arnold, and K. Garwrisch. (1998) Influence of docosahexaenoic acid and cholesterol on lateral lipid organization, Biochemistry 37, 17299-17308. 3. Oldfied, E., Meadows, M., Rice, D., and Jacobs, R. (1978) Spectroscopic studies of specifically deuterium labelled membrane systems. Nuclear magnetic resonance investigation of the effects of cholesterol in model systems., Biochemistry 17, 2727-2740. 4. Brustowicz, M. R., Stillwell, W., and Wassal, S. R. (1999) Molecular organization of cholesterol in polyunsaturated phospholipid membranes: a solid state deuterium investigation, FEBS Lett. 451, 197. 5. Silvius, J. R., Giudice, D., and Lafleur, M. (1996) Cholesterol at different bilayer concentrations can promote or antagonize lateral segregation of phospholipids of differing acyl chain length, Biochemistry 35, 15198-15208. 6. Epand, R. M., Bach, D., Epand, R. F., Borochov, N., and Wachtel, E. (2001) A new high temperature transition of crystalline cholesterol in mixtures of phosphatidylserine, Biophys. J. 81, 1511-1520. 7. Ohvo-Rekila, H., B. Ramstedt, P. Leppimaki, and J. P. Slotte. (2002) Cholesterol interactions with phospholipids in membranes, Progress in Lipid Research 41, 66-97. 8. Taylor, M. G., Akiyama, T., and Smith, I. C. P. (1982) Direct observation of the properties of cholesterol in membranes by deuterium NMR, Chem. Phys. Lipids. 31, 359- 379. 9. Dufourc, E. J., E. J. Parish, S. Chitrakorn, and I. C. P. Smith. (1984) Structural and dynamical details of cholesterol-lipid interaction as revealed by deuterium NMR, Biochemistry 23, 6062-6071. 10. Weisz, K., G. Grobner, C. Mayer, J. Sother, and G. Kothe. (1992) Deuteron nuclear magnetic resonance study of the dynamic organization of phospholipid/cholesterol bilayer membranes: Molecular properties and viscoelastic behavior, Biochemistry 31, 1100-1112.

44 11. Worcester, D. L., and Franks, N. P. (1976) Structural analysis of hydrated egg lecithin and cholesterol bilayers, J. Mol. Biol. 100, 359-378. 12. Yeagle, P. L. (1985) Cholesterol and the cell membrane, Biochem. Biophys. Acta 822, 267-287. 13. Weisz, K., Grobner, G, Mayer, J. S., and Kothe, G. (1992) Biochemistry 31, 1100-1112. 14. Yeagle, P. L. (1985) Cholesterol and the cell membrane, Biochim. Biophys. Acta 822, 267-287. 15. Leonard, A., Escrive, C, Laguerre, M, Pebay-Peyroula, E, Neri, W, Pott, T, Katsaras, J, Dufourc, EJ. (2001) Location of cholesterol in DMPC membranes. A comparative study by neutron diffraction and molecular mechanics simulations, Langmuir 17, 2019-2030. 16. Leonard, A., C. Escrive, M. Laguerre, E. Pebay-Peyroula, W. Neri, T. Pott, J. Katsaras, and E. J. Dufourc. (2001) Location of cholesterol in DMPC membranes. A comparative study by neutron diffraction and molecular mechanics simulations, Langmuir 17, 2019- 2030. 17. Trouard, T. P., Nevzorov, A. A., Alam, T. M., Job, C., Zajicek, J., and Brown, M. F. (1999) Influence of cholesterol on dynamics of dimyristoylphosphocholine bilayers as studied by deuterium NMR relaxation, J. Chem. Phys 110, 8802-8818. 18. Davis, J. H., Jeffrey, K. R., Bloom, M., and Valic, M. I. (1976) Quadrupolar echo deuteron magnetic resonance spectroscopy in ordered hydrocarbon chains, Chem. Phys. Lett. 42. 19. Jacobs, R., and E. Oldfield. (1979) Deuterium nuclear magnetic resonance investigation of dimyristoyllecithin-dipalmitoyllecithin and dimyristoyllecithin-cholesterol mixtures, Biochemistry 18, 3280-3285. 20. Pastor, R. W., Verable, R. M., and Karplus, M. (1988) Brownian dynamics simulation of a lipid chain in a membrane bilayer, J. Chem. Phys 89, 1112-1127. 21. Scott, H. L. (1991) Lipid-cholesterol interactions. Monte Carlo simulations and theory, Biophys. J. 59, 445-455. 22. Bloom, M., J. H. Davis, and A. L. Mackay. (1981) Direct determination of the oriented sample NMR spectrum for systems with local axial symmetry, Chem. Phys. Lett. 80, 198- 201.

45 23. Firestone, M. A., P. Thiyagarajan, and D. M. Tiede. (1998) Structure and optical properties of a thermoresponsive polymer-grafted, lipid-based complex fluid, Langmuir 14, 4688-4698. 24. Firestone, M. A., D. M. Tiede, and S. Seifer. (2000) Magnetic field-induced ordering of a polymer-grafted biomembrane-mimetic hydrogel, J. Phys. Chem. B 104, 2433-2438. 25. Prosser, R. S., H. Bryant, R. G. Bryant, and R. R. Vold. (1999) Lanthanide chelates as bilayer alignment tools in NMR studies of membrane associated peptides, J. Magn. Reson. 141, 256-260. 26. Prosser, R. S., J. S. Hwang, and R. R. Vold. (1998) Magnetically aligned phospholipid bilayers with positive ordering: A new model membrane system, Biophys. J. 74, 2405- 2418. 27. Prosser, R. S., S. A. Hunt, J. A. DiNatale, and R. R. Vold. (1996) Magnetically aligned

membrane model systems with positive order parameter: Switching the sign of Szz with paramagnetic ions, J. Am. Chem. Soc. 269, 269-270. 28. Sanders, C. R., and G. C. Landis. (1995) Reconstitution of membrane proteins into lipid- rich bilayered mixed micelles for NMR studies, Biochemistry 43, 4030-4040. 29. Sanders, C. R., and J. P. Schwonek. (1992) Characterization of magnetically orientable bilayers in mixtures of dihexanoylphosphatidylcholine and dimyristoylphosphatidylcho- line by solid state NMR, Biochemistry 31, 8898-8905. 30. Vold, R. R., R. S. Prosser, and A. J. Deese. (1997) isotropic solutions of phospholipid bicelles: A new membrane mimetic for high-resolution NMR studies of polypeptides, J. Biomol. NMR 9, 329-335. 31. Tiburu, E. K., D. M. Moton, and G. A. Lorigan. (2001) Development of magnetically aligned phospholipid bilayers in mixtures of palmitoylstearoylphosphotidylcholine and dihexanoylphosphatidylcholine by solid state NMR spectroscopy, Biochim. Biophys. Acta 1512, 206-214. 32. Ottiger, M., and A. Bax. (1999) Bicelle-based liquid crystals for NMR-measurement of dipolar couplings at acidic and basic pH values, J. Biomol. NMR 13, 187-191. 33. Cavagnero, S., H. J. Dyson, and P. E. Wright. (1999) Improved low pH bicelle system for orienting macromolecules over a wide temperature range, J. Biomol. NMR 13, 387-391.

46 34. Tjandra, N., and A. D. Bax. (1997) direct measurement of distances and angels in biomolecules by NMR in a dilute liquid crystalline medium, Science 278, 1111-1114. 35. Bechinger, B., and S. J. Opella. (1991) Flat-coil probe for NMR spectroscopy of oriented membrane samples, J. Magn. Reson. 95, 585-588. 36. Prosser, R. S., and I. V. Shiyanovskaya. (2001) Lanthanide ion assisted magnetic alignment and model membranes of macromolecules, Conc. Magn. Reson. 13, 19-31. 37. Sanders, C. R., and J. H. Prestegard. (1990) Magnetically orientable phospholipid bilayers containing small amounts of a bile salt analogue, chapso, Biophys. J. 58, 447- 460. 38. Garber, S. M., G. A. Lorigan, and K. P. Howard. (1999) Magnetically oriented phospholipid bilayers for spin label EPR studies, J. Am. Chem. Soc. 121, 3240-3241. 39. King, V., M. Parker, and K. P. Howard. (2000) Pegylation of magnetically oriented lipid bilayers, J. Magn. Reson. 142, 177-182. 40. Luchette, P. A., T. N. Vetman, R. S. Prosser, R. E. W. Hancock, M. Nieh, C. J. Glinka, S. Kruger, and J. Katsaras. (2001) Morphology of fast-tumbling bicelles: A small angle neutron scattering and NMR study, Biochim. Biophys. Acta 1513, 83-94. 41. Smondyrev, A. M., Berkowitz, ML. (1999) Structure of dipalmitoylphosphatidyl cholinecholesterol bilayer at low and high cholesterol concentrations: Molecular dynamics simulation, Biophys. J. 77, 2075-2089. 42. Kessel, A., Ben-tal, N, May, S. (1989) Towards understanding the electrostatic free energy determinants of lateral domain formation in membranes, Biophys. J. 81, 643-658. 43. Sternin, E., M. Bloom, and A. L. Mackay. (1983) De-Pakeing of NMR spectra, J. Magn. Reson. 55, 274-282. 44. Kessel, A., N. Ben-tal, and S. May. (2001) Interactions of cholesterol with lipid bilayers: The preferred configuration and fluctuations, Biophys. J. 81, 643-658. 45. Stockson, G. W., C. F. Polnaszek, A. P. Tulloch, F. Hasan, and I. C. P. Hasan. (1976) Molecular motion and order in single bilayer vesicles and multilamellar dispersions of egg lecithin and lecithin-cholesterol mixtures. A deuterium nuclear magnetic resonance study of specifically labeld lipids, Biochemistry 15, 954-966.

47 46. Smith, I. C. P., G. W. Stockson, A. P. Tulloch, C. F. Polnaszek, and K. G. Johnson. (1977) Deuterium NMR and spin label ESR as probes of membrane organization, J. Coll. Inter. Sci. 58, 439-451. 47. Smondyrev, A. M., and M. L. Berkowitz. (1999) Structure of dipalmitoylphosphatidyl cholinecholesterol bilayer at low and high cholesterol concentrations: Molecular dynamics simulation, Biophys. J. 77, 2075-2089. 48. Seelig, A., and J. Seelig. (1974) The dynamic structure of fatty acyl chains in a phosph- atidylcholine bilayer measured by deuterium magnetic resonance, Biochemistry 13, 4839-4845. 49. Seelig, J., and W. Niederberger. (1974) Deuterium-labeled lipids as structural probes in liquid crystalline bilayers. A deuterium magnetic resonance, J. Am. Chem. Soc. 96, 2069- 2072. 50. Douliez, J. P., Leonard, A., and E. J. Dufourc. (1995) Restatement of order parameter in biomembranes: Calculation of C-C bond order parameters from C-D quadrupolar splittings, Biophys. J. 68, 1727-1739. 51. Otten, D., M. F. Brown, and K. Beyer. (2000) Softening of membrane bilayers by detergents elucidated by deuterium NMR spectroscopy, J. Phys. Chem. B 104, 12119- 12129. 52. Lafleur, M., Fine, B., Sternin, E., Cullis, P. R., and Bloom, M. (1989) Smoothed orientational order profile of lipid bilayers by deuterium nuclear magnetic resonance, Biophys. J. 56, 1037-1041.

48 CHAPTER 3

Elvis K. Tiburu, Paresh C. Dave, Jason F. Vanlerberghe, Thomas B. Cardon, Robert E. Minto, and Gary A. Lorigan, An Improved Synthetic and Purification Procedure for the Hydrophobic Segment of the Peptide Phospholamban (2003), Anal. Biochem., 318,146-151

49 ABSTRACT The 29-residue (transmembrane segment (Ala24-Leu52)) of the sarcoplasmic reticulum regulator protein phospholamban (24ARQNLQNLFINFCLILICLLLICIIVML52L) was synthesized and purified utilizing improved methods. Applying standard Fmoc strategies generally produce low yields of hydrophobic peptides including phospholamban. An optimized Fmoc procedure was designed to synthesize phospholamban (PLB) and comparative studies were carried out using both the standard and optimized Fmoc procedures. The UV trace indicates that the optimized Fmoc protocol was a better synthesis than the standard protocol. Purification of the crude peptide from the standard synthesis and optimized synthesis were carried out on silica-based and polymer-based C4 columns for comparison. The percent yield of the pure peptide was 12% and 22%, respectively, based upon loaded resin for the standard synthesis. The percent yield from the optimized synthesis was 18% for the silica-based column and 37% for the polymer column. The MALDI-TOF data from the polymer column indicated a high purity of the PLB peptide. The circular dichroism spectra of the purified peptide were recorded in 100% TFE and 2% SDS/phosphate buffers and the a-helical content was approximately 65% which agrees well with previously published results. These data demonstrate that using the optimized synthetic method and a polymer column for hydrophobic peptide purifications dramatically improves the yield of this widely-studied peptide.

50 Introduction

3.1 The role of Phospholamban in Ca2+ Transport Phospholamban (PLB) is a homopentameric transmembrane protein that regulates Ca2+- ATPase, which controls Ca2+ transport across the sarcoplasmic reticulum (SR), leading to muscle relaxation (1). Fuji and co-workers elucidated the complete primary structure of PLB by amino acid sequencing and found that the molecular mass of the PLB monomer is 6082 Da (2). Also they determined that PLB is a pentamer consisting of 5 identical subunits (2). Mutagenesis studies of the transmembrane domain section of PLB indicates that the monomeric form of PLB is a more effective inhibitor of Ca2+-ATPase than the pentameric form (3, 4). Thus, PLB inhibits SERCA2a, an isoform of Ca2+-ATPase, in its unphosphorylated (monomeric PLB) form whereas the phosphorylated (pentameric PLB) form dissociates from SERCA2a (5). Phosphorylation of PLB at Ser16 and Thr17 by both cAMP- and calcium/calmodulin-dependent protein kinases in response to b-adrenergic stimulation resulted in the formation of pentameric complexes (6). PLB is a 52 amino acid transmembrane protein and consists of three structural domains: residues 1-20 which comprise the hydrophilic cytoplasmic domain, residues 21-30 are a hinge segment, and residues 31-52 encompass the hydrophobic a-helical membrane-spanning region (7, 8). Solution NMR studies in organic solvents have shown that monomeric PLB has a disjointed structure, with the intervening domain as either a short flexible turn or in a b-turn type III conformation (9, 10). The a-helical transmembrane segment of PLB consists of 22 amino acid residues (31LFINFCLILICLLLICIIVMLL52). These residues are believed to span the membrane domain and anchor the protein into the membrane (11, 12). Engelman and co-workers utilized site- directed mutagenesis studies to define the interacting surfaces between the a-helices of phospholamban that are responsible for the formation of the pentamer (13). Subsequent studies using chimeric constructs expressed in E. coli explored a wide range of sequence alterations (14). SDS-PAGE assays were employed to determine the influence of substitution on pentamer formation (14). There is a repeating pattern of disruption when substitutions are made along the length of the PLB peptide suggesting that the helices may interact with each other. Mutations at specific residues such as Leu37, Ile40, Leu44 and Leu47 disrupted pentamer formation and are thought to be involved in isoleucine/leucine zipper formations. These helical structures permit

51 nonpolar side chains from one strand to fit into gaps in the surface of another strand, the so- called “knobs-into-holes” bonding arrangement (6). Furthermore, the inhibitory association of PLB with Ca2+-ATPase involves both the cytoplasmic and transmembrane domain of PLB (14- 16). Phosphorylation of PLB activates the Ca2+ pump of the cardiac sarcoplasmic reticulum (SR) and increases the Ca2+ uptake by a mechanism which is still unclear (6). Thus, the dynamic regulation of the protein-protein interactions is key to the understanding of the Ca2+ pump regulation by PLB. Direct measurement of protein dynamics and interactions using site-specific spectroscopic probes will be crucial to the elucidation of this molecular mechanism. Structural analysis through molecular dynamic techniques have been instrumental in understanding the motion and dynamics of PLB in the sarcoplasmic reticulum (17). The existence of a stable PLB pentamer in the SR and its small size makes it suitable for solid-state NMR spectroscopic studies. There is still disagreement on the structure of PLB embedded in lipid membranes even though studies have shown that PLB readily associates in lipid bilayers to form a homopentamer which has been shown to function as a Ca2+ channel (18, 19). Presently, there are two structural models that have been proposed through spectroscopic studies and molecular modeling techniques for pentameric PLB as shown in Figure 3.1 (20, 21). In one model, Figure 3.1(A), PLB is composed of two a-helices connected by a b-sheet with the cytosolic domain tilted in a range of 50-60O with respect to the bilayer normal (20, 21). Another model, Figure 3.1(B), has proposed a continuous a-helix of about 40 amino acid residues with a tilt angle of about 28O for PLB with respect to the bilayer normal (20, 21). Previous studies have shown that the transmembrane helices alone are sufficient to drive pentamer formation; thus, giving rise to interest in the residues that are involved in the structural organization of PLB (6, 22, 23). The isolation of large quantities of native PLB through molecular biology techniques has been hampered due to difficulties encountered in the bacterial overexpression of phospholamban cDNA (24, 25). Recently, PLB has been prepared by chemical synthesis using standard solid- phase peptide synthesis and purification in organic solvents (26-28). Synthetic PLB exhibited biochemical characteristics identical to native PLB (26-28). Thus, the increasing availability of synthetic PLB should facilitate future studies directed towards the biochemical and structural characterization of PLB. However, biophysical studies of transmembrane proteins require the synthesis and purification of large quantities of synthetic peptides, which is difficult due to their

52 high degree of hydrophobicity (29). Phospholamban is no exception to this group of proteins. Studies on phospholamban have been hindered due to its extremely hydrophobic nature and insolubility in most organic solvents (25, 29, 30).

Figure 3.1 The two models showing the orientation of monomeric PLB in phospholipid bilayers. In model (A), PLB has two a-helices separated by an unstructured region presume to be a b sheet structure. In model (B), PLB is a continuous a-helical protein.

Although there have been literature reports describing the purification of the transmembrane segment of phospholamban by reverse-phase high performance liquid chromatography (HPLC), the yields are typically low and harsh organic solvents, such as formic acid, are required for

53 solvation of PLB. The application of solvents containing high concentrations of formic acid cause chemical modifications to PLB due to formylation or oxidation of methionine (31). Furthermore, very few of these procedures actually utilize the wild-type sequence for the transmembrane domain of PLB with their methods (32-35). A straightforward reverse-phase HPLC strategy that allows the purification of the wild-type PLB TM peptide without any sequence modifications is therefore essential. We report here an improved solid-phase synthesis and purification methods that have been developed in our laboratory to increase the overall yield of the transmembrane portion of PLB. First, amino acid residues corresponding to the transmembrane segment of PLB (Ala24- Leu52) were synthesized using a modified FastMoc chemistry. This modification has improved the overall yield of the resin-bound transmembrane PLB segment in comparison to the standard FastMoc synthesis protocols. Secondly, we carried out comparative purification studies on silica- and polymer-based reverse-phase HPLC columns. Though these methods apply specifically to the purification of PLB, they are applicable to other hydrophobic transmembrane peptides as well. The purity of PLB was analyzed using MALDI-TOF and the conformational studies of the pure PLB peptide were carried out using circular dichroism spectroscopy to characterize its secondary structure.

3.2 Materials

Hexafluoro-2-propanol, formic acid, trifluoroacetic acid and 2,2,2-trifluoroethanol were purchased from Aldrich Chemical Co. (St. Louis, MO). Fmoc-amino acids, diisopropylethylamine, 1-hydroxybenzotriazole, O-benzotria-zo-1-yl-N,N,N,N-tetramethyluron- ium hexafluorophosphate, N-methylpyrrolidone, and piperidine were obtained from Applied Biosystems Inc. (Foster City, CA). Methyl tert-butyl ether was bought from Fisher Scientific (Fairlawn, NJ) and preloaded Leu-resin for the synthesis was obtained from Applied Biosystems Inc. (Foster City, CA). HPLC-grade acetonitrile and 2-propanol were obtained from Pharmco (Brookfield, CT) and were filtered through a 0.22-mm nylon membrane before use. Water was purified using a Nanopure reverse osmosis system (Millipore, Bedford, MA). Dodecylphosphocholine was used for circular dichroism studies and was obtained from Avanti Polar Lipids (Miamisburg, OH).

54 3.3 Methods

3.3.1 Peptide synthesis The polypeptide corresponding to the transmembrane segment of PLB (Ala24-Leu52) was synthesized on an Applied Biosystems Inc. (ABI) 433A solid-phase peptide synthesizer controlled by a G3 Macintosh computer with SynthAssistTM 2.0 software. The SynthAssistTM 2.0 software comes with several pre-programmed Fmoc chemistries developed by ABI. We used the standard FastMoc Chemistry developed by ABI and a customized version of the FastMoc chemistry developed by our lab to synthesize PLB. For both protocols, 128 mg of pre-loaded Leu-HMP resin was added to the reaction vessel to synthesize a theoretical yield of 0.1 mmol peptide. The standard FastMoc chemistry developed by ABI utilized 4 steps. The first step involved Fmoc deprotection (removal) of the growing peptide by reacting two equivalents of piperidine with the Fmoc-peptide-resin. One equivalent of piperidine acts as a general base to remove the base-labile Fmoc group from the N-terminus of the peptide, whereas a second equivalent of piperidine covalently binds to the Fmoc group and forms a fulvene-piperidine adduct. The fulvene-piperidine adduct has a UV absorbance at 301 nm. The degree of Fmoc removal is monitored with a specifically UV configured detector set at 301 nm. The program of the FastMoc chemistry contains feedback loops (UV feedback monitoring) to decipher whether or not additional Fmoc deprotection steps were necessary and whether Fmoc deprotection reaction times were to be extended from 2 minutes to 10 minutes (conditional extended deprotection). The second step was the activation of the free Fmoc-amino acid with a mixture of 0.45 M HOBt/0.45 M HBTU in DMF. The third step was the coupling of 0.1-mmol activated Fmoc-amino acid to the deprotected peptide-resin (single coupling) with conditional extended coupling. The fourth step was conditional “capping” (acetylation). If conditional extended deprotection was performed, then conditional capping with Ac2O/HOBt/DIEA was performed to acetylate the N-terminus of the remainder deprotected peptide chains after coupling. It took approximately 30 hours to synthesize PLB. The modified FastMoc chemistry (developed in our lab) was similar to the standard FastMoc chemistry (developed by ABI) except for the following modifications. Conditional extended deprotection and conditional extended coupling time were changed for residues Ile33 through Ala24. Also, double coupling (2 mmol of activated Fmoc-amino acid was coupled to

55 the deprotected peptide-resin) was performed for Cys46 through Asn34. The vortexing time for the coupling step was increased by 1 hour for Ile47 and 6 hours for Cys46 through Ala34. It took approximately 10 days to synthesize the transmembrane segment of PLB utilizing this procedure. The crude peptide yields prior to purification were calculated using the following expression: (mass cleaved peptide/theoretical mass PLB*100).

3.3.2 Peptide purification The lyophilized peptide was dissolved in HFIP/FA (4:1, 1 mg/mL) and centrifuged to eliminate insoluble particulates. The crude peptide was purified on an Amersham Pharmacia Biotech AKTA‚ Explorer 10S HPLC controlled by Unicorn version 3 system software. Two reverse-phase C4 analytical columns were acquired from Grace Vydac Inc. (Hesperia, Ca), a polymer-supported column (259 VHP5415, 5 mm, 300-Å pore size, 0.46 x 15 cm) and a silica- based column (214TP104, 10 mm, 300-Å pore size, 0.46 x 25 cm). Columns were equilibrated with 95% solvent A:5% solvent B (30). Solvent A consisted of H2O + 0.1% TFA and solvent B was 38% MeCN + 57% IPA + 5% H2O. A 200-mL (1 mg/mL) peptide sample was injected to the column and the gradient was ramped from 5% to 100% Solvent B at a flow rate of 1 mL/min. Peptide elution was achieved with a linear gradient to a final solvent composition of 93% solvent B (31) . The purified peptide was lyophilized and analyzed by MALDI-TOF mass spectroscopy using a matrix of 2,5-dihydroxybenzoic acid at the Campus Chemical Instrumentation Center (Ohio State University, Columbus, OH). To determine the concentration of a stock PLB solution in TFE that was used for the CD samples, the ninhydrin colorimetric assay was used (33, 36). CD spectra were gathered using a nitrogen-flushed Jasco J-810 spectropolarimeter (Japan Spectroscopic Co., Japan) controlled by Spectra Manager software version 1.15.00 (34, 35).

3.4 Results and Discussion While theoretically the standard FastMoc solid phase (developed by ABI) procedures mentioned in the method section was quite facile, there had been many complications arising from the synthesis of PLB due to the hydrophobicity of the transmembrane domain. Figure 3.2(A) shows the UV trace of the feedback UV monitoring of the PLB synthesis utilizing the standard FastMoc chemistry (developed by ABI). Conditional extended deprotection was

56 performed for amino acid residues Cys46-Leu39. The difficulty in this part of the synthesis was ascribed to the formation of secondary structures. This is a common problem when the peptide chain is between 8-20 amino acids long. Furthermore, difficulty in the synthesis of the peptide due to aggregation effects become more significant when the peptide length increases. It was observed that the intensity of the first peak of the Fmoc deprotection step for the amino acid residues Ala24-Ile38 decreased significantly relative to the amino acid residues lle47-Leu52. The average yield of crude PLB from this method was 47% based upon 128 mg of starting Leu- HMP resin which would give a theoretical yield of 0.1 mmol PLB peptide. The MALDI-TOF results after cleavage showed several additional peaks indicating peptide truncations (data not shown). We applied several modifications to the standard FastMoc chemistry (ABI) in order to improve the percent yield of crude PLB based upon the analysis of the PLB synthesis using the standard FastMoc chemistry (Figure 3.2(A)). The concentration of the activated amino acids for the difficult amino acids was doubled from 1 mM to 2 mM. The increase in the molar ratio of the activated amino acids to deprotected N-terminal peptides and the increase in reaction times raised the coupling efficiency as shown in Figure 3.2(B). Acetylation (capping) was applied to all amino acids after each coupling step. This eliminates the possibility of random coupling where the incoming amino acid can attach either to the complete growing peptide or to the truncated peptides. The average yield of crude PLB from this FastMoc modified synthesis was 85% based upon 128 mg of starting Leu-HMP resin which would give a theoretical yield of 0.1 mmol PLB peptide. The MALDI-TOF results after cleavage showed few impurity peaks indicating that the formation of truncated peptides was minimal (data not shown). Thus, comparing the synthetic UV trace of the product produced by the modified FastMoc synthesis with the standard FastMoc synthesis, it is apparent that the yield of crude PLB increased significantly through doubling the concentration of Fmoc-amino acids as well as using extended decoupling techniques and UV monitoring (modified FastMoc synthesis). The increase in yield of crude PLB from the standard FastMoc synthesis to the modified FastMoc synthesis was approximately two-fold (47% to 85%). A major problem in the purification of a hydrophobic peptide such as PLB was finding a suitable solvent to solubilize the peptide. Previous studies have indicated that neat formic and trifluoroacetic acids solubilize PLB (31, 37-40). The limitation of pure formic acid is that it

57 (A)

L L M V I I C I L L L C I L I L C F N I F L N Q L N Q R A 52 51 50 49 48 47 46 45 44 43 42 41 40 39 38 37 36 35 34 33 32 31 30 29 28 27 26 25 24

(B)

L L M V I I C-C I-I L-L L-L L-L C-C I-I L-L I-I L-L C-C F-F N-N I F L N Q L N Q R A 52 51 50 49 48 47 46 45 44 43 42 41 40 39 38 37 36 35 34 33 32 31 30 29 28 27 26 25 24

Figure 3.2 Fmoc deprotection of the amino acids corresponding to the transmembrane segment of PLB monitored by UV detection. (A) The Fmoc deprotection of the amino acids using the standard FastMoc chemistry (ABI). In this protocol all the amino acids are singly coupled. (B) The Fmoc deprotection of the amino acids using the modified FastMoc procedure. In this protocol residues 34 through 46 were double-coupled with additional deprotection times of 1 to 6 hours (Cys46 and Asn34-Leu45, respectively).

58 causes peptide formylation, whereas trifluoroacetic acid causes aggregation (31, 37-40). After analyzing the solubility behavior of PLB in several solvents, hexafluoro-2-propanol/formic acid at a 4:1 (v/v) ratio was found to be the best solvent for initially solubilizing the TM PLB peptide. The solubility was improved when formic acid was added to the peptide before the hexafluoro-2- propanol. The duration of contact between PLB and formic acid was minimized, reducing the extent of peptide modification.

(A) 61.5 55.2 53.5

Silica-based 54.3

Polymer-based 57.8 58.6 63.2

(B)

Silica-based 54.7 46.0

Polymer-based

30 40 50 60 70

Elution time (min) Figure 3.3 Reverse-phase HPLC profile for crude PLB from the standard and modified synthesis on silica-based and polymer-based C4 columns. (A) The standard FastMoc synthesis was purified on a silica-based analytical C4 column (dotted line) and polymer-based C4 column (solid line). (B) The modified FastMoc synthesis was purified on a silica-based C4 column (dotted line) and a polymer-based analytical C4 column (solid line). The numbers in bold represent the retention times for PLB.

59 First, the crude PLB peptide from the standard FastMoc synthesis was purified on a silica-based C4 reverse-phase and polymer-based C4 columns for comparison as shown in Figure 3.3(A). An optimized HPLC protocol using a linear gradient from 5% solvent B to 100% solvent B eluted the PLB peptide at 93% solvent B (31, 37, 38). The elution time per run was about 73 minutes. The HPLC results from the silica-based column of the standard synthesis showed peaks which were less intense and unresolved (Figure 3.3(A), dotted line). The fraction corresponding to the peak with retention time 61.5 min ( Figure 3.3(A), standard synthesis, dotted line) was PLB peptide and other impurity peaks based upon MALDI-TOF analysis (data not shown). The impurity peaks in the MALDI-TOF indicated that the truncated peptides were not all resolved by the silica-based column. Also, the peaks at retention times 53.5 min and 55.2 min were one leucine and two leucines short of PLB respectively based on MALDI-TOF analysis. These leucines were among the most difficult residues (Cys46-Asn34) to couple to the growing peptide on the resin (Figure 3.2(A)). The synthesis of PLB by the standard FastMoc chemistry resulted in low yields of pure PLB which may have accounted for the low intensity of the HPLC peaks in Figure 3.3(A) (dotted line). The percent yield of PLB based upon injection of a 1 mg/mL crude peptide sample was 12%. The purification of crude PLB peptide from the standard synthesis was repeated on a polymer-based C4 reverse-phase column for comparison with the results from the silica-based column purification (Figure 3.3(A), solid line). The PLB peaks were much more intense with the appearance of fewer impurity peaks indicating the impurity peaks may have different elution profiles in the polymer column. The peak at retention time 54.3 min was analyzed by MALDI- TOF and confirmed to be PLB with minor impurities (data not shown). The split observed in the chromatogram is probably caused by the overlapping of different aggregates of PLB (Figure 3.3(A), solid line). The percent yield of pure PLB was increased to 22% based upon injection of a 1 mg/mL peptide sample. Similarly, the purification of crude PLB peptide from the modified FastMoc synthesis was performed on the silica-based column and the polymer-based column as shown in Figure 3.3(B). The impurity peaks on the silica-based column were more intense suggesting the silica- based column had better elution profile for the truncated peptides (Figure 3.3(B), dotted line). The fractions corresponding to the two intense peaks at retention times 58.6 min and 57.8 min were collected and analyzed by MALDI-TOF and the results showed these peaks were one

60 leucine and a leucine and asparagine short respectively of PLB. The broad and less intense peak at retention time 63.2 min represents the PLB peptide as confirmed from the MALDI-TOF analysis results. The percent yield of pure PLB was 18% after purifying the peptide from the modified synthesis with the silica-based column.

Sequence of the transmembrane segment of PLB: ARQNLQNLFINFCLILICLLLICIIVMLL

PLB

1000

800

600 a.i.

400

PLB+23 200

0 2800 3000 3200 3400 3600 m/z

Figure 3.4: The MALDI-TOF MS results of the pure PLB from the polymer- based column using the modified synthesis. The molecular weight corresponding to the PLB peak is 3375. The second intense peak at position 3398 is the sodiated- PLB.

We repeated the purification of crude PLB from the modified synthesis on the polymer-based C4 column as shown in Figure 3.3(B) (solid line). There was a peak at retention time 46.0 min which was not characterized. The intense peak at retention time 54.7 min corresponds to pure PLB based upon MALDI-TOF analysis. Unlike the chromatograms from the silica-based column (Figure 3.3(B), dotted line), the PLB peak was better resolved and more intense in separations carried out with the polymer column. The percent yield of pure PLB from the polymer column was 37%. Figure 3.4 shows the corresponding MALDI-TOF data of pure PLB fraction collected from the polymer column (Figure 3.3(B), solid line). The intense peak at m/z

61 3375 represents the protonated transmembrane segment of PLB. The second less intense peak in the MALDI-TOF spectrum corresponds to PLB + Na+. From these results it is evident that the polymer column was more efficient in separating the peptide fragments than the silica-based column. The synthesis and purification of the hydrophobic transmembrane domain of PLB was improved dramatically utilizing the modified FastMoc synthetic procedure with UV monitoring and a polymer-based column. The increase in yield of PLB using the modified FastMoc synthesis was 50% (12% to 18%) on the silica-based C4 column. The increase in yield of PLB utilizing the modified FastMoc synthesis was 68% (22% to 37%) on the polymer-based column. Temperature has been shown to play an important role in peptide separation, and often causing an increase in both resolution and purity. We tested the effect of temperature on the separation of PLB for both columns and there were visible changes in the resolution of the peaks (data not shown). With the polymer column the intensity of the PLB peak decreased and became broad at higher temperatures (45 OC – 66 OC). This suggests the possibility of peptide aggregation at elevated temperatures. Similar observations were made with the silica-based column as well (data not shown). In order to determine the structural conformation of PLB in different solvents and detergents, CD spectroscopy was used. The CD spectra of 25-mM synthetic PLB in 100% TFE, 2% SDS and in DPC/SDS mixed micelles were obtained and the a-helical content determined as described previously (36). The a-helical content was found to be 65±3% at all temperatures (data not shown). The CD results agree with those quoted in the literature for full-length PLB (35) and verify the structural integrity of the PLB peptide synthesized by our improved methods. We have demonstrated the utility of a modified FastMoc synthetic method to dramatically improve the yield of the hydrophobic transmembrane segment of phospholamban utilizing UV monitoring. In addition, we have shown that polymer-based HPLC columns are better than the silica-based columns for purification and quantitation of the hydrophobic domain of phospholamban. Polymer columns are stable in strong acids or bases at ambient temperatures, and thus provide a broad choice of mobile phases where silica-based columns would be destroyed. These methods have enormous potential to increase the yield and quality of many synthetic hydrophobic peptides like PLB.

62 3.5 References 1. Li, H., Cocco, M. J., Steitz, T. A., and Engelman, D. M. (2001) Conversion of phospholamban into soluble pentameric helical bundle, biochemistry 40, 6636-6645. 2. Fuji, J., M. Kadoma, H. Tada, and F. Sakiyama. (1986) Characterization of structural unit of phospholamban by amino acid sequencing and electrophoretic analysis, Biochem. Biophys. Res. Commun. 138, 1044-1050. 3. Cornea, R., Autry, J., Chen, Z., and Jones, L. (2000) Re-examination of the role of the leucine/isoleucine zipper residues of phospholamban in inhibition of the Ca2+-pump of cardiac sarcoplasmic reticulum, J. Biol. Chem. 275, 41487-41494. 4. Kimura, Y., Kurzydlowski, K., Tada, M., and MacLennan, D. H. (1996) Phospholamban regulates the Ca2+-ATPase through intramembrane interactions, J. Biol. Chem. 249, 6166-6173. 5. Minamisawa, S., Hoshijima, M., Chu, G., Martone, M. E., Wang, Y., Ross, J., Jr., Kranias, E. G., Giles, W. R., and Chien, K. R. (1999) Chronic phospholamban- sarcoplasmic reticulum calcium ATPase interaction is the critical calcium cycling defect in dilated cardiomyopathy, Cell 99, 313-322. 6. Simmerman, H. K., Collins, J. H., Theibert, J. L., Wegender, A. D., and Jones, L. R. (1996) Sequence analysis of phospholamban. Identification of phosphorylation sites and two major structural domains, J. Biol. Chem. 261, 13333-13341. 7. Simmerman, H. K., Collins, J. H., Theibert, J. L., Wegener, A. D., and Jones. L. R. (1986) Sequence analysis of phospholamban: identification of phosphorylation sites and two major structural domains, J. Biol. Chem. 261, 13333-13341. 8. Mascioni, A., Karim, C., Zamoon, J., Thomas, D. D., and Veglia, G. (2002) Solid-state NMR and rigid body molecular dynamics to determine domain orientations of monomeric phospholamban, J. Am. Chem. Soc. 124, 9392-9393. 9. Pollesello, P., Annila, A., and Ovaska, M. (1999) Structure of the 1-36 amino-terminal fragment of human phospholamban by nuclear magnetic resonance and modeling of the phospholamban pentamer, Biophys. J., 1784-1795. 10. Lamberth, S., Schimidt, H., Muenchbach, M., Vorherr, T., Krebs, J., Carafoli, E., and Griesinger, C. (2000) NMR solution structure of phospholamban, Helv. Chim. Acta. 83, 2141-2152.

63 11. Arkin, I. T., Adams, P. D., MacKenzie, K. R., Lemmon, M. A., Brunger, A. T., and Engelman, D. M. (1994) Structural organization of the pentameric transmembrane alpha helix of phospholamban, a cardiac ion channel, EMBO. J. 13, 4757-4764. 12. Simmerman, H. K. B., Lovelace, D. E., and Jones, L. R. (1989) Secondary structure of detergent-solubilized phospholamban, a phosphorylatable, oligomeric protein of cardiac sarcoplasmic reticulum, Biochim. Biophys. Acta, 322-329. 13. Birmachu, W., and Thomas, D. D. (1990) Rotational dynamics of the Ca-ATPase in sarcoplasmic reticulum studied by time-resolved phosphorescence anisotropy, Biochemistry 29, 3904-3914. 14. Sasaki, T., Inui, M., Kimura, Y., Kuzuya, T., and Tada, M. (1992) Molecular mechanism of regulation of Ca2+ pump ATPase by phospholamban in cardiac sarcoplasmic- reticulum - effects of synthetic phospholamban peptides on Ca2+ pump ATPase, J. Biol. Chem. 267, 1674-1679. 15. Toyofuku, T., Kurzydlowski, K., Tada, M., and MacLennan, D. H. (1994) Amino-acids Glu(2) to Ile(18) in the cytoplasmic domain of phospholamban are essential for functional association with the Ca2+-ATPase of sarcoplasmic-reticulum, J. Biol. Chem. 269, 3088-3094. 16. Hughes, J., East, J. M., and Lee, A. G. (1994) The hydrophilic domain of phospholamban inhibits the Ca2+ transport step of the Ca2+-ATPase, Biochem. J. 303, 511-516. 17. Adams, P. D., Arkin, I. T., Engelman, D. M., and Brunger, A. T. (1995) Computational searching and mutagenesis suggest a structure for the pentameric transmembrane domain of phospholamban, Nat. Struct. Biol. 2, 154-162. 18. Ahmed, Z., Reid, D. G., Watts, A., and Middletown, D. A. (2000) A solid state NMR study of the phospholamban transmembrane domain: local structure and interactions with Ca+2-ATPase, Biochim. Biophys. Acta 1468, 187-198. 19. Hughes, E., and Middleton, D. A. (2003) Solid-state NMR reveals structural changes in phospholamban accompanying the functional regulation of Ca2+-ATPase, J. Biol. Chem. 278, 20835-20842. 20. Arkin, I. T., Rothman, M., Ludlam, C. F., Aimoto, S., Engelman, D. M., Rothschild, K. J., and Smith, S. O. (1995) Structural model of the phospholamban ion channel complex in phospholipid membranes, J. Mol. Biol. 248, 824-834.

64 21. Tatulian, S. A., Jones, L. R., Reddy, L. G., Stokes, D. L., and Tamm, L. K. (1995) Secondary structure and orientation of phospholamban reconstituted in supported bilayers from polarized attenuated total reflection FTIR spectroscopy, Biochemistry 34, 4448-4456. 22. Frank, S., Kammerer, R. A., Hellstern, S., Pegoraro, S., Stefeld, J., Lustig, A., Moroder, L., and Engel, J. (2003) Toward a high-resolution structure of phospholamban: Design of soluble transmembrane domain mutants., Biochemistry 39, 6825-6831. 23. Simmerman, H. K. B., Kobayashi, Y. M., Autry, J. M., and Jones, L. R. (1996) A leucine zipper stabilizes the pentameric membrane domain of phospholamban and forms a coiled-coil pore structure, J. Biol. Chem. 271, 5941-5946. 24. Fuji, J., Ueno, A., Kitano, K., Tanaka, S., Kadoma, M., and Tada, M. (1987) Complete complementary DNA-derived amino acid sequence of canine cardiac phospholamban, J. Clin. Invest. 79, 301-304. 25. Yao, Q., Bevan, J. L. Weaver, R.F., and Bigelow, D. J., (1996) Purification of porcine phospholamban expressed in Escherichia coli, Protein expression and purification 8, 463-468. 26. Collins, J. G., Kranias, E. G., Reeves, A. S., Bilezikjian, L. M., and Schwartz, A. (1981) Isolation of phospholamban and a second proteolipid component from canine cardiac sarcoplasmic reticulum, Biochem. Biophys. Res. Commun. 99, 796-803. 27. Otterson, K. M., Noble, R. L., Hoeprich, P.D., Jr., Shaw, K. T., and Ramage, R. (1993) UV monitoring of deprotection for FastMoc-SPPC on the model 433A: Feedback control to optimize peptide synthesis performance, Research News, Applied Biosystems, 1-10. 28. Mayer, E. J., McKenna, E., Garsky, V. M., Burke, C. J., Mach, H., Middaugh, C. R., Sardana, M., Smith, J. S., and Johnson R. G., Jr. (1996) Biochemical and biophysical comparison of native and chemically synthesized phospholamban and a monomeric phospholamban, J. Biol. Chem. 271, 1669-1677. 29. Hellstern, S., S. Pegoraro, C. B. Karim, A. Lustig, D.D. Thomas, L. Moroder, and J. Engel. (2001) Sarcolipin, the shorter homologue of phospholamban, forms oligomeric structures in detergent micelles and in liposomes, J. Biol. Chem. 276, 30845-30852.

65 30. Lew, S., E. London. (1997) Simple procedure for reversed-phase high performance liquid chromatographic purification of long hydrophobic peptides that form transmembrane helices, Anal. Biochem. 251, 113-116. 31. Torres, J., P.D. Adams, and I.T. Arkin. (2000) Use of a new label, 13C=18O, in the determination of a structural model of phospholamban in a lipid bilayer. spatial restraints resolve the ambiguity arising from interpretations of mutagenesis data, J. Mol. Biol. 300, 677-685. 32. Frank, S., R.A. Kammerer, S. Hellestern, S. Pegporaro, J. Stetefeld, A. Lustig, L. Moroder, and J. Engel. (2000) Toward a high-resolution structure of phospholamban: design of soluble transmembrane domain mutants, Biochemistry 39, 6825-6831. 33. Rosen, H. (1957) A modified ninhydrin colorimetric analysis for amino acids, Archiv. Biochem. Biophys. 67, 10-15. 34. Sreerama, N., R.W. Woody. (1993) A self-consistent method for the analysis of protein secondary structure from circular dichroism, Anal. Bichem. 209, 32-44. 35. Pelton, J. T., and McLean, L.R. (2000) Spectroscopic methods for analysis of protein secondary structure, Anal. Chem. 277, 167-176. 36. Minto, R. E., W.J.Jr. Gibbons, T.B. Cardon, and G.A. Lorigan. (2002) Synthesis and confomational studies of a transmembrane domain from a diverged microsomal D12- desaturase, Anal. Bichem. 308, 134-140. 37. Karim, C. B., J.D. Stamm, J. Karim, L.R. Jones, and D.D. Thomas. (1998) Cysteine reactivity and oligomeric structures of phospholamban and its mutants, Biochemistry 37, 12074-12081. 38. Ludlam, C. F. C., Arkin, I.T., Liu, X. , Rothman, M.S. , Rath, P., Aimoto, S.,. Smith, S.O, Engelman, D.M., and Rothchild, K.J. (1996) Fourier transform infrared spectroscopy and site-directed isotope labeling as a probe of local secondary structure in the transmembrane domain of phospholamban, Biophys. J. 70, 1728-1736. 39. Karim, C. B., Marquardt, C. G., Stamm, J. D., Berany, G. and Thomas. D.D. (2000) Synthetic null-cysteine phospholamban analogue and the corresponding transmembrane domain inhibit the Ca-ATPase, Biochemistry 39, 10892-10897.

66 40. Karim, C. B., Paterlini, M.G., Reddy, L.G., Hunter, G.W., Barany, G., and Thomas. D.D. (2001) Role of cysteine residues in structural stability and function of a transmembrane helix bundle, J. Biol. Chem. 276, 38814.

67 CHAPTER 4

Paresh C. Dave, Elvis K. Tiburu, Krishnan Damodaran, and Gary A. Lorigan, Investigating Structural Changes in the Lipid Bilayer upon Insertion of the Transmembrane Domain of the Membrane-Bound Protein Phospholamban Utilizing 31P and 2H Solid-State NMR Spectroscopy (2003), Biophys. J., 86, 1564-1573

68 Abstract

Phospholamban (PLB) is a 52-amino acid integral membrane protein that regulates the flow of Ca2+ ions in cardiac muscle cells. In the present study, the transmembrane domain of PLB (24-52) was incorporated into phospholipid bilayers prepared from 1-palmitoyl-2-oleoyl-sn- glycero-phosphocholine (POPC). Solid-state 31P and 2H NMR experiments were carried out to study the behavior of POPC bilayers in the presence of the hydrophobic peptide PLB at temperatures ranging from 30 ºC to º60 C. The PLB peptide concentration varied from 0 mol% to 6 mol% with respect to POPC. Solid-state 31P NMR spectroscopy is a valuable technique to study the different phases formed by phospholipid membranes. 31P NMR results suggest that the transmembrane protein phospholamban is incorporated successfully into the bilayer and the effects are observed in the lipid lamellar phase. Simulations of the 31P NMR spectra were carried out to reveal the formation of different vesicle sizes upon PLB insertion. The bilayer vesicles fragmented into smaller sizes by increasing the concentration of PLB with respect to POPC (only sn-1 chain was deuterated for the studies) phospholipid bilayers when the PLB peptide was inserted into the membrane.

69 4.0 Introduction

4.1 Protein-Lipid Interaction Methods for understanding interactions between proteins and lipids are essential for elucidating biological structure-function relationships. Peptide-lipid interactions can affect both protein and bilayer structure (1-3). Previous studies have suggested that these interactions can serve several regulatory roles such as controlling the membrane-association of lytic peptides, modulating membrane-protein activity, promoting peptide aggregation, segregation of proteins within the membrane, determining protein sorting during secretion recognition (4-7). Three typical models of biological membranes are planar lipid bilayers, vesicles (liposomes) and monolayers. According to the Singer-Nicholson model of a cell membrane, a lipid bilayer resembles a biomembrane closer than a monolayer (8). Liposomes or phospholipid dispersions are commonly used to study membrane structure upon peptide insertion (9, 10). In an aqueous solution, phospholipids self-assemble to form lipid bilayers rather than micelles. The reason is that phosphatidylcholine phospholipids have two acyl chains that are more or less parallel to one another. The overall shape of a phospholipid molecule is approximately rectangular, so these molecules are too bulky to fit in the interior of micelles. These well-defined synthetic membranes are used as model systems to mimic the properties of biological membranes. Interestingly, the lipid bilayer in biological membranes is generally in the liquid-crystalline phase where the axis of symmetry of the acyl chain motion is perpendicular to the plane formed by the polar head group. High-resolution NMR techniques are now routinely employed to study the structure of complex macromolecules in solution (11-14). An alternative approach to solution structural studies of membrane macromolecules is the determination of their structural and dynamic properties using solid-state NMR spectroscopy. Solid-state NMR spectroscopy has been widely used to study the structure and dynamics of peptides including those that associate with membranes or with biomineral surfaces (15-25). For membrane-bound peptides, solid-state NMR spectroscopy has the ability to probe lipid bilayers in the presence of a peptide, which can be poised in a biologically relevant liquid-crystalline state. Membrane proteins reconstituted into synthetic phospholipid bilayers, simulate the biological membrane better than detergent micelles

70 such as those formed from sodium dodecyl sulfate (SDS) (26-29). Solid-state NMR spectroscopy can probe the molecular structure and dynamics of membrane proteins directly without bulky extrinsic probes, thus minimizing perturbations of the native conformation of the proteins (22, 30). Determining the structure of PLB and its interactions with the lipid bilayer is central for understanding its regulatory role (31-33). The transmembrane segment of PLB (24-52) has been synthesized using solid-phase peptide synthesis and purified according to the modified method reported recently by our group (34). In the present study, the transmembrane domain of PLB (24- 52) was incorporated into phospholipid bilayers prepared from 1-palmitoyl-2-oleoyl-sn-3- glycero-phosphatidylcholine (POPC). Solid-state NMR spectroscopy has been used to monitor the interactions between the lipid bilayers and PLB, by exploiting the 31P nuclei as a natural spin reporter on the head group of POPC. Solid-state 31P NMR spectroscopy is a valuable technique to study the different phases formed by model phospholipid membranes (35, 36). The 31P NMR line shapes have distinct characteristics for different lipid phases such as the gel and liquid crystalline lamellar phases, the inverted hexagonal phase, and isotropic phases such as small vesicles or micelles (37). The low chain melting point of POPC makes it possible to examine PLB in membranes at a physiologically relevant temperature utilizing solid-state NMR spectroscopy. In the present paper, we are focusing on three main points: (1) lipid-peptide interactions of PLB incorporated into POPC bilayers utilizing 31P NMR spectroscopy, (2) perturbations of the phospholipid bilayers using POPC-d31 as a NMR probe and (3) the effects of various concentrations of PLB and temperature on the dynamic properties of the phospholipid bilayers.

4.2 Materials Fmoc-amino acids and other chemicals for peptide synthesis were purchased from

Applied Biosystems Inc. (Foster City, CA). POPC and POPC-d31 were purchased from Avanti Polar Lipids (Alabaster, AL). Prior to use, phospholipids were dissolved in chloroform and stored at –20 0C. Chloroform, hexafluoro-2-propanol, formic acid, and 2,2,2-trifluoroethanol (TFE) were purchased from Aldrich Chemical Co. (St. Louis, MO). HPLC-grade acetonitrile and 2-propanol were obtained from Pharmco (Brookfield, CT) and were filtered through a 0.22- mm nylon membrane before use. Water was purified using a Nanopure reverse osmosis system

71 (Millipore, Bedford, MA). N-[2-hydroxyethyl]piperazine-N’-2-ethane sulfonic acid (HEPES) and EDTA were also obtained from Sigma-Aldrich (St. Louis, MO).

4.3 Methods 4.3.1 Synthesis and Purification of PLB PLB was synthesized according to the recently published procedure (34). In brief, PLB was synthesized using modified Fmoc-based solid-phase methods with an ABI 433A peptide synthesizer (Applied Biosystems, Foster City, CA). The sequence of the synthesized transmembrane segment of PLB (24-52) is ARQNLQNLFINFCLILICLLLICIIVMLL. The crude peptide was purified on an Amersham Pharmacia Biotech AKTA explorer 10S HPLC controlled by Unicorn (version 3) system software. A reverse-phase C4 semi-preparative polymer supported column (259VHP82215) was acquired from Grace Vydac Inc. (Hesperia, CA). Columns were equilibrated with 95% solvent A and 5% solvent B. Solvent A consisted of

H2O and solvent B was 38% MeCN + 57% IPA and 5% H2O. Elution of the peptide was achieved with a linear gradient to a final solvent composition of 93% solvent B. The purified peptide was lyophilized and characterized by matrix-assisted laser desorption ionization time of flight (MALDI-TOF) mass spectrometry.

4.3.2 NMR Sample Preparation The POPC-rich bilayer samples, containing various mol% of peptide to phospholipid, were prepared following a slightly modified protocol given by Rigby and co-workers. (29).

POPC (76 mg) and PLB were dissolved in CHCl3 and TFE, respectively and added to a 12 x 75 mm test tube. The solvents were removed under a steady stream of N2 gas for approx. 15-20 min. The test tube was placed in a vacuum desicator overnight to remove any residual solvents. The peptide/lipid mixture was resuspended in 190 mL HEPES buffer (5 mM EDTA, 20 mM NaCl and 30 mM HEPES, pH 7.0) by heating in a water bath at +50 0C along with slight frequent sample agitation to avoid frothing the mixture. After all the phospholipids were fully dissolved, the sample was transferred to a NMR sample tube. POPC-d31 (4 mg) was added to the samples when conducting the 2H NMR experiments.

72 4.3.3 NMR Spectroscopy 31P NMR spectra were recorded on a Bruker Avance 500 MHz solid-state NMR spectrometer operating at 202.4 MHz using a Bruker double resonance 5-mm round coil static probe. The 31P NMR spectra were recorded with 1H decoupling using a 4-ms p/2 pulse for 31P and a 5-s recycle delay. For the 31P NMR spectra 1028 scans were taken and the free induction decay was processed using 100 Hz of line broadening. The spectral width was set to 150 ppm. 2H NMR spectra were recorded on the same NMR spectrometer operating at 76.77 MHz using the same 5-mm static round coil NMR probe. The quadrupolar echo pulse sequence was employed using quadrature detection with complete phase cycling of the pulse pairs (38). The 900 pulse length was 3 ms, the inter-pulse delay was 20 ms, the recycle delay was 0.4 s, the spectral width was set to 100 kHz. A total of 12 K transients were averaged for each spectrum and processed using 200 Hz line broadening. The sample was held at the desired temperature for 10 min prior to signal acquisition.

4.3.4 NMR data analysis Simulation of 31P NMR spectra was carried out using the software program called DMFIT (39). This program permits the fitting or modeling of experimental 1D and 2D spectra to a sum of lines or contributions characterized by their corresponding NMR parameters. The spectral fittings were conducted using a minimum number of species. Static chemical shift anisotropy spectral patterns were considered for all the species. Lorentzian broadening was used for all simulations.

2 Powder-type H NMR spectra of multilamellar dispersions of POPC-d31 were numerically deconvoluted (dePaked) using the algorithm of McCabe and Wassall (40, 41). The spectra were deconvoluted such that the bilayer normal was perpendicular with respect to the direction of the static magnetic field. The quadrupolar splittings were directely measured from the dePaked spectra and converted into order parameters according to (42, 43) the following expression:

i 2 i DnQ = 3/4 (e qQ/h) SCD

i th 2 where DnQ is the quadrupolar splitting for a deuteron attached to the i carbon, e qQ/h is the 2 i quadrupolar splitting constant (168 kHz for deuterons in C- H bonds), and SCD is the chain order

73 parameter for a deuteron attached to the ith carbon of the acyl chain of POPC. The 2H nuclei attached to the terminal methyl carbons were assigned carbon number 15. The remaining 2H assignments were made in decreasing order along the phospholipid acyl chain. Thus, the corresponding order parameters for the individual C-D methylene groups and the terminal methyl groups of the acyl chains were directly evaluated from the quadrupole splittings of the dePaked 2H NMR spectra. The 2H peaks in the NMR spectra were assigned based upon the dynamic properties of the individual CD3 and CD2 groups. The quadrupole splittings of the CD3 methyl groups at the end of the acyl chains are the smallest and closet to 0 kHz because they rotate at the fastest frequency. The next smallest splitting was assigned to the 2H attached to C- 14 and so forth along the acyl chain. The quadrupole splittings for the deuterons in the plateau region were estimated by integration of the last broad peak according to the literature (44). The order parameters calculated for the CD3 quadrupole splitting were multiplied by three according to the literature (45, 46).

4.4 Results and Discussion

4.4.1 31P NMR Study of PLB Incorporated into POPC Bilayer 31P NMR spectroscopic measurements were performed to determine the nature of the various lamellar and non-lamellar phase transitions exhibited by the transmembrane segment of PLB incorporated into POPC phospholipids bilayers. The static 31P NMR spectra of the POPC phospholipid bilayer with and without 1 mol% of the hydrophobic segment of PLB are shown in Figures 4.1(A) and 4.1(B), respectively. The 31P powder pattern NMR spectra were recorded at temperatures ranging from 30 0C to 60 0C. The motionally averaged powder pattern spectra are characteristic of phospholipid bilayers in the liquid crystalline phase (La) and are expected for POPC at a temperature well above its chain melting transition temperature of –3 0C (36). The spectra in the absence and in the presence of 1 mol% PLB showed a similar powder lineshape, both with h = 0 (axially symmetry). The spectra indicate that the lipid bilayers remain in the La phase even after addition of 1 mol% phospholamban with respect to POPC, and do not form isotropic or inverse hexagonal phases with high curvatures.

74 (B) (A)

0 30 C

0 45 C

0 60 C

40 20 0 -20 -40 40 20 0 -20 -40 31 31 P (ppm) P (ppm)

Figure 4.1 31P NMR spectra of POPC phospholipid bilayers investigated as function of temperature. (A) In the absence of PLB, spectra are shown for pure POPC bilayers (left column). (B) 31P spectra are shown for the POPC bilayer in the presence of 1 mol% PLB with respect to POPC (right column). The solid line spectra represent the experimental results and the dotted spectra represent best-fit simulated spectra corresponding to the experimental spectra.

In the La phase, a POPC bilayer has been determined to have a hydrophobic thickness of approximately 27 Å (47, 48). The entire thickness of the POPC phospholipid bilayer is approximately 50-54 Å (49). If the hydrophobic region of PLB (24-52) is 100% a-helical then it would be approximately 43 Å in length (50). This indicates that the POPC bilayer is a close match in thickness to the transmembrane segment of PLB and hydrophobic mismatch is not a serious problem under these conditions. The peptide is incorporated into the bilayer without distorting the lipid structure. Hydrophobic mismatch for model peptides KK(LA)15KK (45 Å hydrophobic length) have been studied in detail using 31P and 15N NMR spectroscopy (47). Those results clearly suggest that the POPC bilayer is a perfect match for hydrophobic peptides

75 with a 36 Å to 45 Å in length. A DOPC bilayer has been used previously for incorporation of AFA-PLB into mechanically oriented membrane system (32).

(B) (A)

0 30 C

0 45 C

0 60 C

40 20 0 -20 -40 40 20 0 -20 -40 31 31 P (ppm) P (ppm) Figure 4.2 31P NMR spectra of POPC phospholipid bilayers in the presence of 4 mol% PLB with respect to POPC investigated as function of temperature. (A) Experimental spectra (left column). (B) The dotted line spectra represent the simulated bilayer species having different vesicle sizes possessing different CSA widths. The solid line spectra (in right column) are the sum of the dotted line spectra and represent the best-fit simulated spectra corresponding to the experimental spectra.

The 31P NMR spectra of POPC samples containing no PLB and with 1 mol% PLB were found to possess a chemical shift anisotropy (CSA, in this paper CSA is sll - s^) width of 44 ppm and 41 ppm, respectively (26, 27, 36). A somewhat smaller (~ 3 ppm) 31P CSA width is detected for the membrane-bound sample in our study as seen from the line shape simulations. Similarly, in both cases by increasing the sample temperature, the CSA width decreases indicating that the molecular motions in the phospholipid bilayer increases with temperature. Higher

76 concentrations of PLB were also incorporated into POPC bilayers to probe the peptide-lipid interactions. Figures 4.2(A) and 4.2(B) show the experimental and simulated 31P NMR spectra of POPC bilayer samples containing 4 mol% PLB with respect to phospholipid, respectively. Similarly, Figures 4.3(A) and 4.3(B) represent the experimental and simulated 31P NMR spectra of POPC bilayer samples containing 6 mol% PLB with respect to phospholipid, respectively. The results clearly indicate that PLB interacts with the head groups of the POPC bilayer. This study was performed at temperatures ranging from 30 0C to 60 0C. The unoriented multilamellar liposome samples containing 0 mol% to 6 mol% peptide, were all found to produce 31P NMR

0 powder pattern spectra representing POPC in the La phase even at higher temperatures (60 C). The membrane remains in the lamellar phase upon binding of PLB even at the highest peptide concentration (6 mol%) studied. The overall CSA spectral width is about 3 ppm smaller when compared to the pure POPC membrane. This can be attributed to a faster rotation of the lipids when the membrane associated PLB peptide partially disrupts the hydrogen bonding network between the lipid head groups. At higher concentrations of PLB, the spectra represent super- positions of the different lamellar phases. Previously, the presence of two different anisotropic or lamellar phases were observed in the spectral simulations of 31P NMR spectra of cardiotoxin incorporated into DMPA phospholipid bilayers (51). Also, Strandberg and co-workers reported the presence of three different lamellar phases upon incorporation of the peptide KK(LA)8KK into DOPC bilayers utilizing 31P NMR spectroscopy (52). In order to fully understand the peptide-lipid interactions from 31P NMR, spectral simulations have been carried out as explained in the Materials and Method section (Figures 4.2(B) and 4.3(B)). It is important to note here that the 31P chemical shift of phosphodiester, such as is found in membrane lipids, depends upon the molecular motions and orientation of the group with respect to the magnetic field of the spectrometer (37).

77 (B) (A)

0 30 C

0 45 C

0 60 C

40 20 0 -20 -40 40 20 0 -20 -40 31 31 P (ppm) P (ppm)

Figure 4.3 31P NMR spectra of POPC phospholipid bilayers in the presence of 6 mol% PLB with respect to the phospholipid POPC bilayers investigated as function of temperature. (A) Experimental spectra (left column). (B) The dotted line spectra represent the simulated bilayer species having different vesicle sizes possessing different CSA width. The solid line spectra (in right column) are the sum of the dotted line spectra and represent the best-fit simulated spectra corresponding to the experimental spectra.

The orientation of the head group depends upon the chemistry, hydration, hydrogen bonding and charge interactions of each phospholipid head group. 31P NMR spectra are very sensitive to the rate of motions of lipids in the liquid crystalline phase, which usually have motional rates in the fast-limit region. Burnell and co-workers were able to simulate experimental spectra of dioleoyl phosphophatidyl choline (DOPC) vesicles of different sizes at various temperatures and viscosities of the medium (53). The general results from these studies indicate that the spectral line shape and CSA width are dependent upon the size of the vesicles. The CSA spectral width

78 decreases with decreasing vesicle sizes, and isotropic lines are observed for very small vesicles. The 31P NMR in Figures 4.2(A) and 4.3(A) at higher concentrations of PLB clearly indicates that the presence of an isotropic peak and a shallow high-field shoulder (near to s^). The spectral simulations reveal the presence of different species of POPC bilayers (Figures 4.2(B) and 4.3(B)). The interpretation of the 31P NMR spectra with the isotropic components is not certain (54). In some studies, they have been attributed to vesicular or micellar structures on or within the bilayers, possible associated with different phases (35, 55). Others have attributed them to the presence of smaller diameter vesicles induced upon protein binding (56, 57). Alternatively, the line shapes may arise from different 31P species, which undergoes molecular motion that yield isotropic chemical shift transitions. Despite the non-bilayer spectral component being broader from lipid-peptide complexes when compared to peptide-free bilayers, there is no

31 indication of a well-defined hexagonal HII phase. The spectral simulation of P NMR line shape clearly indicates the presence of two different lamellar phases and one isotropic phase. The presence of two lamellar phases having different CSA width values also indicates the presence of multi-lamellar vesicles (MLVs) of different sizes forming different micro domains. The presence of an isotropic component indicates the existence of micellelar components. Interestingly, spectral simulations suggest the presence of different species at higher PLB concentrations, when compared to 1 mol% PLB in which case only one species is present (Figure 4.1). The 31P NMR spectral lineshapes clearly indicate (Figure 4.1, 4.2 and 4.3) that the phospholipid bilayers are disrupted by increasing the concentration of PLB. The presence of one species having a similar large CSA width (41 ppm) was observed for the control and 1 mol% PLB sample and suggests the presence of large size MLVs. The smaller CSA width obtained for the other lamellar phases when 4 and 6 mol% PLB was added to the POPC bilayer suggests the presence of smaller size vesicles (53). The large size vesicles can be fragmented to form micro domains composed of small size vesicles. The diameter of the vesicles can be calculated from the CSA width for different MLVs species (53). At 1 mol % PLB with respect to POPC, larger MLVs with an approx. diameter of 25000 Å are formed. When 4 mol% PLB was embedded into the bilayers two different size vesicles were formed with an approx. diameter 20,000 and 1500 Å. Additionally, the isotropic components suggests the presence of very small vesicles having diameters smaller than 1000 Å. Analysis of the 6 mol% PLB/POPC data indicates that the extended bilayer fragmented and formed vesicles possessing diameters of approximately 10,000,

79 2500 Å, and one isotropic component with a diameter less than 1000 Å. The contributions of each component were calculated by integrating the area of the individual species and comparing it with the entire CSA width of the experimental spectrum. These results indicate that when 4 mol% PLB was incorporated into POPC contributions from component I (large CSA width), component II (small CSA width) and component III (isotropic species) is approx. 82%, 17% and 1%, respectively at 30 0C. In addition contributions from component I decrease from 82% to 67% and contributions from component II and component III increase from 17% to 26% and 1% to 7%, respectively, when the temperature was increased from 30 0C to 60 0C. Interestingly, the sample at 30 0C containing 6 mol% PLB/POPC was found to have contributions from component I, component II and component III of approx. 67%, 25% and 8%, respectively. Contributions from component I decrease from 67% to 39% and increase for component II from 25% to 44% and for component III from 8% to 17% when the temperature is increased from 30 0C to 60 0C. A comparison of Figures 4.2 and 4.3 indicate that the contribution from the anisotropic phases having smaller CSA widths increases as more PLB is incorporated into the phospholipid bilayers. One can see clearly from Figures 4.2 and 4.3 that upon increasing the temperature the contributions from the different species change. At higher temperatures, molecular motions increase and small vesicles rotate faster. Also, lateral diffusion influences the 31P NMR line shape of phospholipid bilayer (35, 37, 53). Another possibility is that by increasing the amount of PLB incorporated into the POPC bilayer, the lateral diffusion rate of the lipids at higher temperature increases resulting in a decrease in the CSA width of the 31P NMR spectra.

4.4.2 2H NMR Study of PLB Incorporated into POPC Bilayer

The effect of PLB on the order and dynamics of the acyl chains of the POPC bilayer have

2 been studied using POPC-d31 (deuterated palmitoyl acyl chain). The H NMR spectra of a 0 dispersion of POPC-d31 in the absence and in the presence of PLB at 35 C are shown in Figure 4.4. Several conclusions can be immediately drawn from the 2H NMR line shapes in Figure 4.4. First, the spectra are characteristic of axially symmetric motions of the phospholipids about the bilayer normal and the spectra consist mostly of overlapping doublet resonances that result from the different CD2 segments of the acyl chain.

80 Control

1 mol% PLB

2 mol% PLB

4 mol% PLB

6 mol% PLB

40 20 0 -20 -40 2 H (kHz)

Figure 4.4 2H NMR powder spectra of PLB at various concentrations

0 incorporated into POPC/POPC-d31 phospholipid bilayers at 35 C. 5.55 mol%

POPC-d31 was doped into the POPC sample. The concentration of PLB with respect to POPC is noted on the right side of each spectrum.

The central doublet corresponds to the terminal CD3 group. Secondly, the spectral width is a measure of the fluidity of the lipid bilayer. The range observed in the POPC/POPC-d31 is typical for acyl chains in a liquid-crystalline bilayer (58-60). In Figure 4.4, the spectral width of POPC- d31 marginally decreases by increasing the PLB concentrations when compared to the control spectrum of pure POPC/POPC-d31. The marginal decrease in the spectral width suggests that the presence of the PLB peptide disorders the acyl chains to some extent for all the different PLB/POPC concentrations. It also reveals that the POPC bilayer is still in the liquid-crystalline

31 La phase. This supports our P NMR results as discussed earlier. The spectral resolution deteriorates as the concentration of PLB increases, manifested by the disappearance of the sharp edges of the peaks. This suggests intermediate-timescale motions of the lipids induced by the

81 PLB peptide. The changes in the spectral resolution of the 2H NMR spectra confirm that PLB interacts with the acyl chains of the lipid bilayers. Interesting features of the 2H NMR spectra is the appearance of an isotropic peak in the presence of PLB and the intensity of which increases with the amount of PLB. This indicates that the large vesicles are fragmented into smaller size vesicles by increasing the amount of PLB in the POPC bilayers. These results agree with our 31P NMR data.

0 mol% PLB 4 mol% PLB

0 30 C

0 35 C

0 40 C

0 45 C

0 50 C

0 55 C

0 60 C

20 10 0 -10 -20 2 H (kHz)

Figure 4.5 Temperature-dependent 2H NMR spectra of PLB incorporated into

POPC/POPC-d31 phospholipid bilayers. The solid line spectra represent

POPC/POPC-d31 sample prepared in the absence of PLB, whereas the dotted line spectra represent samples prepared with 4 mol% PLB with respect to POPC. The temperature at which each spectrum was taken is noted on the right side of each spectrum.

2 Figure 4.5 shows the H NMR spectra of POPC-d31 samples in the presence (4 mol%, dotted line) and in the absence (pure POPC/POPC-d31 bilayers, solid line) of PLB obtained over a

82 temperature ranging from 30 0C to 60 0C. The shape of the 2H NMR spectra of the control

2 POPC/POPC-d31 is a typical H phospholipid bilayer lineshape (La phase), and looks very similar to those previously reported (48, 49, 60). As the temperature is raised, the spectra retain their overall line shape in both cases with and without PLB. The spectral features become narrower due to increased mobility by raising the temperature. The 2H NMR spectra with and without 4 mol% PLB did not undergo any substantial changes in the spectral breadth or quadrupolar splitting. Interestingly, the isotropic component observed by the addition of 4 mol% PLB indicates the formation of small vesicles due to the fragmentation of large vesicles throughout the temperature range from 30 0C to 60 0C. The intensity of the isotropic peak increases as the temperature increases, indicating an increase in the rapid tumbling of small vesicles and/or an increase in lateral diffusion of the phospholipids (61).

The smoothed segmental C-D bond order parameters (SCD) were calculated by dePakeing the powder spectra represented in Figure 4.5 for the POPC-d31 as a function of temperature

(Figure 4.6). The SCD order parameters depend upon several averaging modes provided by intramolecular, intermolecular, and collective motions. The segmental SCD order parameter describes local orientation or dynamic perturbations of the C-D bond vector from its standard state due to perturbations of the POPC phospholipid conformations or dynamics as a result of the addition of PLB. The magnitude of the order parameters (SCD ~ 0.20- 0.30) indicates that the phospholipid bilayers are in the liquid crystalline phase (44, 60). A characteristic profile of decreasing order parameters with increasing distance from the glycerol backbone was obtained both for the pure bilayer and for the PLB-bound bilayer. These values are comparable with the results previously determined using labeled POPC by NMR, and also recently calculated by using molecular dynamics simulations (49, 58). The data indicates that there is more disorder and motion in the center and at the end of the acyl chain when compared to the head group region in the presence of PLB. Interestingly, the order parameter profile obtained from the sample in the presence of PLB (4 mol%) closely resembles the order parameter profile of pure POPC bilayers. This indicates that the POPC bilayer acyl chains are not significantly perturbed by the addition of PLB at these concentrations. Also, the order parameter profiles indicate (for both cases) that by increasing the temperature, the SCD values decreases. This is due to a combination of the increase in the mobility of the acyl chains, rapid tumbling of vesicles, and possible increases in lateral diffusion of the POPC phospholipids.

83 0.35 (A) 0.30

0.25 CD

0.20 0 30 C 0 0.15 35 C 0 40 C

Order parameter, S 0.10 0 45 C 0 50 C 0.05 0 55 C 0 60 C 0.00 0.35 (B) 0.30

0.25 CD

0.20

0.15

Order parameter, S 0.10

0.05

0.00

0 2 4 6 8 10 12 14 16 Carbon no.

Figure 4.6 Temperature-dependent smoothed acyl chain (POPC-d31) orientational

order SCD profiles calculated from the dePaked spectra of Figure 4.5. (A) For pure POPC bilayer and (B) 4 mol% PLB embedded into the POPC bilayer.

4.5 Conclusion In the present study, the 31P NMR spectra indicated that the hydrophobic segment of phospholamban was incorporated successfully into the fully hydrated dispersed POPC phospholipid bilayers. The spectral simulations of 31P NMR spectra of samples containing higher concentrations of PLB indicate the presence of different sizes of vesicles. The large

84 phospholipid bilayer vesicles fragmented into smaller vesicles by increasing the concentration of PLB with respect to POPC. Interestingly, the contribution from smaller vesicles increased as the temperature increased. The data indicates that small vesicles are tumbling faster at higher temperatures. Another possibility is that by increasing the amount of PLB incorporated into the POPC bilayer, the lateral diffusion rate of the lipids increases at higher temperatures. In addition

31 2 to the P NMR studies, H NMR spectroscopic studies were carried out using POPC-d31 to understand the lipid-peptide interactions in the hydrophobic region of the phospholipid bilayers. The data suggests that smaller size vesicles are formed by the addition of 4 mol% PLB into the

31 bilayers and supports our P NMR spectral study. The segmental SCD order parameter indicates that there are no significant changes in the ordering of the acyl chains of the phospholipid bilayers. It suggests that the incorporation of PLB into the POPC bilayers did not significantly perturb the acyl chains of the phospholipids bilayers at these concentrations.

85 4.5 References 1. Bloom, M. (1991) Physical properties of the fluid lipid-bilayer component of cell membrane: a perspective, Q. Rev. Biophys. 24, 293-397. 2. Lemmon, M., and Engelman, D. (1994) Specificity and promiscuity in membrane helix interactions, FEBS Lett. 346, 17-20. 3. Watts, A. (1993) in Phospholipids handbook (Cevc, G., Ed.) pp 687-744, Marcel Dekker Inc., New York, USA. 4. Lemmon, M., and Engelman, D. (1994) Specificity and promiscuity in membrane helix interactions, Q. Rev. Biophys. 27, 157-218. 5. Arora, A., and Tamm, L. K. (2001) Biophysical approaches to membrane protein structure determination, Curr. Opin. Struct. Biol. 11, 540-547. 6. Watts, A. (1981) Protein-lipid interactions: do the spectroscopists now agree?, Nature 294, 512-513. 7. Dempsey, C. E., Ryba, N. J. P., and Watts, A. (1986) Evidence from deuterium nuclear magnetic resonance for the temperature dependent reversible self association of erythrocyte band-3 in dimyristoylphosphatidylcholine bilayers, Biochemistry 25, 2180- 2187. 8. Singer, S. L., and Nicolson, G. L. (1972) The fluid mosaic model of the structure of cell membranes, Science 175, 720-731. 9. Epand, R. M. (1998) Lipid polymorphism and protein-lipid interactions, Biochim. Biophys. Acta 1376, 353-368. 10. Liu, F., Lewis, R. N. A. H., Hodges, R. S., and McElhaney, R. N. (2001) A differential scanning calorimetric and 31P NMR spectroscopic study of the effect of transmembrane a-helical peptides on the lamellar-reversed hexagonal phase transition of phosphatidylethanolamine model membranes, Biochemistry 40, 760-768. 11. Brunner, E., Ogle, J., Wenzler, M., and Kalbitzer, H. R. (2000) Molecular alignment of proteins in bicellar solutions: Quantitative evaluation of effects induced in 2D COSY spectra, Biochem. Biophys. Res. Comm. 272, 694-698. 12. Cavagnero, S., Dyson, H. J., and Wright, P. E. (1999) Improved low pH bicelle system for orienting macromolecules over a wide temperature range, J. Biomol. NMR 13, 387-391.

86 13. Ottiger, M., and Bax, A. (1999) Bicelle-based liquid crystals for NMR-measurement of dipolar couplings at acidic and basic pH values, J. Biomol. NMR 13, 187-191. 14. Vold, R. R., Prosser, R. S., and Deese, A. J. (1997) Isotropic solutions of phospholipid bicelles: A new membrane mimetic for high-resolution NMR studies of polypeptides, J. Biomol. NMR 9, 329-335. 15. Long, J. R., Shaw, W. J., Stayton, P.S., and Drobny, G. P., (2001) Structure and dynamics of hydrated staherin on hydroxyapatite as determined by solid-state NMR, Biochemistry 40, 15451-15455. 16. Cross, T. A. (1997) Solid-state nuclear magnetic resonance characterization of gramicidin channel structure, Meth. Enzymol. 289, 672-696. 17. Nicholson, L. K., and Cross, T. A. (1989) Gramicidin cation channel-an experimental determination of the right handed helix sense and verification of beta-type hydrogen- bonding, Biochemistry 28, 9379-9385. 18. Long, J. R., Dindot, J. L., Zebroski, H., Kiihne, S., Clark, R. H., Campbell, A. A., Stayton, P. S., and Drobny, G. P. (1998) A peptide that inhibits hydroxyaptite growth is in an extended conformation on the crystal surface, Proc. Natl. Acad. Sci USA 95, 12083- 12087. 19. Shaw, W. J., Long, J. R., Dindot, J. L., Campbell, A. A., Stayton, P. S., and Drobny, G. P. (2000) Determination of statherin N-terminal peptide conformation on hydroxyapatite crystals, J. Am. Chem. Soc. 122, 1709-1716. 20. Watts, A. (1998) Solid-state NMR approaches for studying the interaction of peptides and proteins with membranes, Biochim. Biophys. Acta 1376, 297-318. 21. Marassi, F. M., Ma, C., Gratkowski, H., Straus, S. K., Strebel, K., Orblatt-Montal, M., Montal, M., and Opella, S. J. (1999) Correlation of the structural and functional domains in the membrane protein Vpu from HIV-1, Proc. Natl. Acad. Sci USA 96, 14336-14341. 22. Marassi, F. M., and Opella, S. J. (1998) NMR structural studies of membrane proteins, Curr. Opin. Struct. Biol. 8, 640-648. 23. Barre, P., Zschoring, O., Arnold, K., and Huster, D. (2003) Structural and dynamical changes of the bindin B18 peptide upon binding to lipid membranes, Biochemistry 42, 8377-8386.

87 24. Nakazawa, Y., and Asakura, T. (2003) Structure determination of a peptide model of the repeated helical domain in Samia cynthia ricini silk fibroin before spinning by a combination of advanced solid-state NMR methods, J. Am. Chem. Soc. 125, 7230-7237. 25. Marcotte, I., Dufourc, E. J., Ouellet, M., and Auger, M. (2003) Interaction of the neuropeptide Met-Enkephalin with zwitterionic and negatively charged bicelles as viewed by 31P and 2H solid-state NMR, Biophys. J. 85, 328-339. 26. Sharpe, S., Barber, K. R., Grant, C. W. M., Goodyear, D., and Morrow, M. R. (2002) Organization of model helical peptides in lipid bilayers: Insight into the behavior of single-span protein transmembrane domains, Biophys. J. 83, 345-358. 27. Sharpe, S., Barber, K. R., Grant, C. W. M., and Morrow, M. R. (2002) Evidence of a tendancy of self-association of the transmembrane domain of ErbB-2 in fluid phospholipid bilayers, Biochemistry 41, 2341-2352. 28. Morrow, M. R., and Grant, C. W. M. (2000) The EGF-receptor transmembrane domain: peptide-peptide interactions in fluid bilayer membranes, Biophys. J. 79, 2024-2032. 29. Rigby, A. C., Barber, K. R., Shaw, G. S., and Grant, C. W. M. (1996) Transmembrane region of the epidermal growth factor receptor: behavior and interactions via 2H NMR, Biochemistry 35, 12591-12601. 30. Fu, R., and Cross, T. A. (1999) Solid-state nuclear magnetic resonance investigation of protein and polypeptide structure, Annu. Rev. Biophys. Biomol. Struct. 28, 235-268. 31. Mascioni, A., Karim, C., Barany, G., Thomas, D. D., and Vegila, G. (2002) Structure and orientation of sacrolipin in lipid enviornments, Biochemistry 41, 475-482. 32. Mascioni, A., Karim, C., Zamoon, J., Thomas, D. D., and Vegila, G., (2002) Solid-state NMR and rigid body molecular dynamics to determine domain orientations of monomeric phospholamban, J. Am. Chem. Soc. 124, 9392-9393. 33. Ying, W., Irvine, S. E., Beekman, R. A., Siminovitch, D. J., and Smith, S. O. (2000) Deuterium NMR reveals helix packing interactions in phospholamban, J. Am. Chem. Soc. 122, 11125-11128. 34. Tiburu, E. K., Dave, P. C., Vanlerberghe, J. F., Cardon, T. B., Minto, R. E., and Lorigan, G. A. (2003) An improved synthetic and purification procedure for the hydrophobic segment of the transmembrane peptide phospholamban, Anal. Biochem. 318, 146-151.

88 35. Cullis, P. R., and de Kruijff, B. (1979) Lipid polymorphism and the functional roles of lipids in biological membranes, Biochim. Biophys. Acta 559, 399-420. 36. Seelig, J. (1978) 31P Nuclear magnetic resonance and the head group structure of phospholipids in membranes, Biochim. Biophys. Acta 515, 105-140. 37. Smith, I. C. P., and Ekiel, I. H. (1984) in Phosphorus-31 NMR (Gorenstein, D. G., Ed.) pp 447-475, Academic Press (Hughes Sci. Lab. Miami Univ.). 38. Davis, J. H., Jeffrey, K. R., Bloom, M., and Valic, M. I. (1976) Quadrupolar echo deuteron magnetic resonance spectroscopy in ordered hydrocarbon chains, Chem. Phys. Lett. 42, 390-394. 39. Massiot, D., Fayon, F., Capron, M., King, I., Le calve, S., Alonso, B., Durand, J. O., Bujoli, B., Gan, Z., and Hoatson, G. (2002) Modelling one- and two-dimensional solid- state NMR spectra, Mag. Res. Chem. 40, 70-76. 40. Alan McCabe, M., and Wassall, S. R. (1997) Rapid deconvolution of NMR powder spectra by weighted fast fourier transormation, Solid State NMR 10, 53-61. 41. Alan McCabe, M., and Wassall, S. R. (1995) Fast-fourier-transform depaking, J. Magn. Res. B 106, 80-82. 42. Huster, D., Yao, X., Jakes, K., and Hong, M. (2002) Confomational changes of colicin Ia channel-forming domain upon membrane binding: a solid-state NMR study, Biochim. Biophys. Acta 1561, 159-170. 43. Dave, P. C., Tiburu, E. K., Nusair, N. A., and Lorigan, G. A. (2003) Calculating order parameter profiles utilizing magnetically aligned phospholipid bilayers for 2H NMR studies, Solid State NMR 24, 328-339. 44. Huster, D., Arnold, K., and Garwrisch, K. (1998) Influence of docosahexaenoic acid and cholesterol on lateral lipid organization, Biochemistry 37, 17299-17308. 45. Stockson, G. W., Polnaszek, C. F., Tulloch, A. P., Hasan, F., and Hasan, I. C. P. (1976) Molecular motion and order in single bilayer vesicles and multilamellar dispersions of egg lecithin and lecithin-cholesterol moxtures. A deuterium nuclear magnetic resonance study of specifically labeld lipids, Biochemistry 15, 954-966. 46. Dufourc, E. J., Parish, E. J., Chitrakorn, S., and Smith, I. C. P. (1984) Structural and dynamical details of cholesterol-lipid interaction as revealed by deuterium NMR, Biochemistry 23, 6062-6071.

89 47. Harzer, U., and Bechinger, B. (2000) Alignment of lysine-anchored membrane peptides under conditions of hydrophobic mismatch: a CD, 15N and 31P solid-state NMR spectroscopy investigation, Biochemistry 39, 13106-13114. 48. Nezil, F., A., and Bloom, M. (1992) Combined influence of cholesterol and synthetic amphiphilic peptides upon bilayer thickness in model membranes, Biophys. J. 61, 1176- 1183. 49. Huber, T., Rajamoorthi, K., Kurze, V. F., Bayer, K., and Brown, M. F. (2002) Structure of docosahexaenoic acid-containing phospholipid bilayers as studied by 2H NMR and molecular dynamics simulations, J. Am. Chem. Soc. 124, 298-309. 50. Jones, D. T., Taylor, W. R., and Thornton, J. M. (1994) A model recognition approach to the prediction of all-helical membrane protein structure and topology, Biochemistry 33, 3038-3049. 51. Auger, M. (1997) Membrane structure and dynamics as viewed by solid-state NMR spectroscopy, Biophys. Chem. 68, 233-241. 52. Strandberg, E., Sparrman, T., and Lindblom, G. (2001) Phase diagrams of systems with cationic a-helical membrane-spanning model peptides and dioleoylphosphatidylcholine, Advances in Colloid and Interface Science 89-90, 239-261. 53. Burnell, E. E., Cullis, P. R., and De Kruijff, B. (1980) Effects of tumbling and lateral diffusion on phosphatidylcholine model membrane 31P NMR line shapes, Biochim. Biophys. Acta 1980, 63-69. 54. Pinheiro, T. J. T., Duer, M. J., and Watts, A. (1997) Phospholipid headgroup dynamics in

2 31 DOPC-d5-cytochrome c complexes as revealed by H and P NMR: the effects of a peripheral protein on collective lipid fluctuations, Solid State NMR 8, 55-64. 55. de Kruijff, B., and Cullis, P. R. (1980) Cytochrome-c specifically induces non-bilayer structures in cardiolipin-containing model membranes, Biochim. Biophys. Acta 902, 477- 490. 56. Pinheiro, T. J. T., and Watts, A. (1994) Lipid specificity in the interaction of cytochrome- c with anionic phospholipid-bilayers revealed by solid-state P-31 NMR, Biochemistry 33, 2451-2458.

90 57. Pinheiro, T. J. T., and Watts, A. (1994) Resolution of individual lipids in mixed phospholipid-membranes and specific lipid cytochrome-c interactions by magic-angle- spinning solid state P-31 NMR, Biochemistry 33, 2459-2467. 58. Seelig, A., and Seelig, J. (1974) The dynamic structure of fatty acyl chains in a phosphatidylcholine bilayer measured by deuterium magnetic resonance, Biochemistry 13, 4839-4845. 59. Seelig, J., and Niederberger, W. (1974) Deuterium-labeled lipids as structural probes in liquid crystalline bilayers. A deuterium magnetic resonance, J. Am. Chem. Soc. 96, 2069- 2072. 60. Lafleur, M., Fine, B., Sternin, E., Cullis, P. R., and Bloom, M. (1989) Smoothed orientational order profile of lipid bilayers by 2H -nuclear magnetic resonance, Biophys. J. 56, 1037-1041. 61. Davis, J. H. (1979) Deuterium magnetic resonance study of the gel and liquid crystalline phases of dipalmitoyl phosphatidylcholine, Biophys. J. 27, 339-358. 62. Sanders, C. R., and Schwonek, J. P. (1992) Characterization of magnetically orientable bilayers in mixtures of dihexanoylphosphatidylcholine and dimyristoylphosphatidyl- choline by solid state NMR, Biochemistry 31, 8898-8905.

91 CHAPTER 5

Elvis K. Tiburu, Ethan S. Karp, Paresh C. Dave, Krishnan Damodaran, and Gary A. Lorigan, Investigating the Dynamic Properties of the Transmembrane Segment of Phospholamban Incorporated into Phospholipid Bilayers Utilizing 2H and 15N Solid-State NMR Spectroscopy (2004), Biochemistry (accepted for publication)

92 Abstract 2H and 15N solid-state NMR spectroscopic techniques were used to investigate the membrane composition, orientation and side chain dynamics of the transmembrane segment of phospholamban (TM-PLB), a sarcoplasmic Ca2+ regulator protein. 2H NMR spectra of 2H- labeled leucine (deuterated at one terminal methyl group) incorporated at different sites (CD3-

Leu28, CD3-Leu39, and CD3-Leu51) along the TM-PLB peptide exhibited line shapes characteristic of either methyl group reorientation about the Cg-Cd bond axis, or by additional 2 librational motion about the Ca-Cb and Cb-Cg bond axes. The H NMR lineshapes of all CD3 labeled leucines are very similar below 0 OC, indicating that all the residues are located inside the lipid bilayer. At higher temperatures, all three labeled leucine residues undergo rapid

2 reorientation about the Ca-Cb, Cb-Cg and Cg-Cd bond axes as indicated by H lineshape simulations and reduced quadrupolar splittings. At all the temperatures studied, the 2H NMR spectra indicate that the Leu51 sidechain has less motion than Leu39 or Leu28 which is attributed to its incorporation in the pentameric PLB leucine zipper motif. The 15N powder spectra of Leu39 and Leu42 residues indicated no backbone motion, while Leu28 exhibited slight backbone motion. The chemical shift anisotropy tensor values for 15N-labeled Leu TM-

PLB were: s11 =50.5 ppm, s22 = 80.5 ppm, and s33 = 229 ppm within ± 3 ppm experimental error. The 15N chemical shift value from the mechanically aligned spectrum of 15N-labeled Leu39 PLB in DOPC/DOPE phospholipid bilayers was 220 ppm and is characteristic of a transmembrane peptide that is nearly parallel with the bilayer normal.

93 5.0 Introduction

5.1 Dynamics of PLB in phospholipid bilayers Phospholamban (PLB) is a homopentameric transmembrane protein that regulates Ca2+- ATPase, which controls Ca2+ transport across the sarcoplasmic reticulum (SR), leading to muscle relaxation (1). Fuji and co-workers elucidated the complete primary structure of PLB by amino acid sequencing and found that the molecular mass of the PLB monomer is 6082 Da. Also, they determined that PLB is a pentamer consisting of 5 identical subunits (2). Mutagenesis studies of the transmembrane domain section of PLB indicate that the monomeric form of PLB is a more effective inhibitor of Ca2+-ATPase than the pentameric form (3, 4). Thus, PLB inhibits SERCA2a, an isoform of Ca2+-ATPase, in its unphosphorylated (monomeric PLB) form whereas the phosphorylated pentameric PLB form dissociates from SERCA2a (5). Phosphorylation of PLB at Ser16 and Thr17 by both cAMP- and calcium/calmodulin-dependent protein kinases in response to b-adrenergic stimulation results in the formation of pentameric complexes (6). PLB is a 52 amino acid transmembrane protein and consists of three structural domains: residues 1-20 which comprise the hydrophilic cytoplasmic domain, residues 21-30 that create a hinge segment, and residues 31-52 that encompass the hydrophobic a-helical membrane-spanning region (7, 8). Solution NMR studies in organic solvents have shown that monomeric PLB has a disjointed structure, with the intervening domain as either a short flexible turn or a b-turn type III conformation (9, 10). The a-helical transmembrane segment of PLB consists of 22 amino acid residues (31LFINFCLILICLLLICIIVMLL52) and is believed to anchor the protein into the membrane (11, 12). Engelman and co-workers utilized site-directed mutagenesis studies to define the interacting surfaces between the a-helices of phospholamban that are responsible for the formation of the pentamer (13). Subsequent studies using chimeric constructs expressed in E. coli explored a wide range of sequence alterations (14). SDS-PAGE assays were employed to determine the influence of substitution on pentamer formation (14). There is a repeating pattern of disruption when substitutions are made along the length of the PLB peptide suggesting that the helices may interact with each other. Mutations at specific residues such as Leu37, Ile40, Leu44 and Leu47 disrupt pentamer formation and are thought to be involved in

94 isoleucine/leucine zipper formations (6, 11, 14, 15). The oligomeric state of phospholamban in lipid bilayers has also been determined using spin-labeled EPR and refined by molecular dynamic simulations. These studies suggest that these helical structures permit nonpolar side chains from one strand to fit into gaps in the surface of another strand, the so-called “knobs-into- holes” bonding arrangement (16). Furthermore, the inhibitory association of PLB with Ca2+-ATPase involves both the cytoplasmic and the transmembrane domain of PLB (17-19). Phosphorylation of PLB activates the Ca2+ pump of the cardiac sarcoplasmic reticulum (SR) and increases the Ca2+ uptake by a mechanism which is still unclear (20). Thus, the dynamic regulation of the protein-protein interactions is the key to understanding the Ca2+ pump regulation by PLB. Direct measurement of protein dynamics and interactions using site-specific spectroscopic probes will be crucial to the elucidation of this molecular mechanism. Structural analysis through molecular dynamic techniques have been instrumental in understanding the motion and dynamics of PLB in the sarcoplasmic reticulum (21). The existence of a stable PLB pentamer in the SR and its small size makes it suitable for solid-state NMR spectroscopic studies. There is still disagreement on the structure of PLB embedded in lipid membranes even though studies have shown that PLB readily associates in lipid bilayers to form a homopentamer that has been shown to function as a Ca2+ channel (22, 23). Presently, there are two structural models that have been proposed based upon spectroscopic studies and molecular modeling techniques for pentameric PLB as shown in Figure 5.1 (24, 25). In one model (Fig. 5.1(A)), PLB is composed of two a-helices connected by an unstructured/b-sheet region with the cytosolic domain tilted in a range of 50O - 60O with respect to the bilayer normal (25). Another model (Fig. 5.1(B)) has proposed a continuous a-helix of about 40 amino acid residues with a tilt angle of about 28O for PLB with respect to the bilayer normal (24). Previous studies have shown that the transmembrane helices alone are sufficient to drive pentamer formation; thus, giving rise to interest in the residues that are involved in the structural organization of PLB (6, 26). Deuterium NMR spectroscopy is well suited to study the structural and dynamic properties of membrane proteins in phospholipid bilayers (27-30). The technique has developed into an excellent probe for dynamic processes utilizing the corresponding quadrupolar splitting and lineshapes (31, 32). 2H NMR spectroscopy has been used to study the molecular dynamics of side chain residues in site-specific 2H labeled integral membrane proteins (33-36). The motions

95 of the methyl groups in aliphatic side chains reflect those of the backbone and methylene sites of these residues. The ability to selectively introduce a CD3 group at a specific residue in a protein through peptide synthesis, and to examine its corresponding quadrupolar splitting and lineshape associated with its motion through NMR spectroscopy is a powerful technique used by many research groups (15, 34, 37). Methyl group motions have been well characterized by 2H NMR studies of CD3-labeled sites of alanines and valines in other transmembrane peptides (38-41). The primary amino acid sequence of the transmembrane segment of PLB clearly shows a repeating pattern of isoleucine/leucine (isoleucine/leucine make up about 60% of the transmembrane segment) residues along the length of the peptide, suggesting a possible interacting role of these residues in stabilizing the pentameric structure inside the membrane. Very limited information is available on 2H solid-state NMR dynamic studies on long aliphatic side chains such as those of leucine for transmembrane proteins such as PLB (34, 36). This study will enhance our knowledge of long side chain dynamics by investigating the motional properties of various CD3 Leu residues of TM-PLB.

For leucine residues, the side chain can be isotopically labeled at the d- and/or e-CD3 sites, and the deuterium NMR powder pattern line shapes will be strongly influenced by the motions about the Cg-Cd bond axis and by additional librational motion about the Ca-Cb and Cb-Cg bond axes at various temperatures (34, 42). Furthermore, knowledge of the precise location of the residues within the bilayer is required for understanding the structural and dynamic organization of the peptide within the membrane. For solid-state 2H NMR spectroscopy the allowed transitions correspond to +1 ´ 0 and 0 ´ -1 and give rise to a quadrupolar splitting of the absorption line with separation Dv between peak maxima (assuming an axial symmetric electric field gradient tensor) of

Dv = (3/4)(e2qQ/h)(3cos2q-1) (1) where e2qQ/h is the quadrupole coupling constant and q defines the orientation of the principal axis of the electric field gradient tensor with respect to the laboratory coordinate system (35).

Generally, for CD3-Leu in a polycrystalline solid, all values of q are possible and the so-called “powder pattern” is obtained, having a quadrupolar splitting (Dv) of 127 kHz for solid aliphatic chains (35, 43). Three-fold methyl hops or rotations cause a reduction in the quadrupolar

96 Figure 5.1 The two models representing the structure of monomeric PLB in phospholipid bilayers are depicted above. In model (A), PLB has two a-helices separated by an unstructured region presumed to be a b-sheet structure. In model (B), PLB is a continuous a-helical protein.

splitting (Dv) from 127 kHz to 42.3 kHz (44). However, such spectra still retain an axially symmetric gradient tensor. Further reduction in the splitting followed by a certain degree of asymmetry in the spectra can be caused by additional motions; such as (i) rotational motions about the peptide long molecular axis, (ii) rotation about the Ca-Cb and Cb-Cg bond axes, (iii) librational motions associated with wobbling of methyl groups or (iv) backbone mobility (44, 45).

97 In addition, selective 13C-13C and 13C-15N interatomic distances along the backbone of wild type PLB have been measured via solid-state NMR spectroscopic techniques such as rotational echo double resonance (REDOR) and rotational resonance (23). The results suggest that the wild type PLB is pentameric with an a-helical structure. However, 15N solid-state NMR studies involving the mutant analogue of the AFA-PLB (where A36, F41, and A46 have replaced the three corresponding transmembrane cysteine residues) monomer in uniaxially aligned lipid bilayers have shown that the transmembrane and cytosolic domains are perpendicular to each other (8). The conflicting results may reflect structural differences between the pentameric and monomeric forms of PLB due to conformational changes. Thus, it is one objective of this study to use site- specific isotopically labeled TM-PLB peptides to probe the backbone and side chain motions of the transmembrane segment of the wild type PLB in phospholipid bilayers. In the present study, several TM-PLB peptides were synthesized with selective incorporation of 15N-labeled or 2H-labeled leucine at specific positions. We have examined side-

2 chain motions of [5,5,5- H3]Leu-labeled TM-PLB at residues Leu28, Leu39, and Leu51 separately incorporated into 1-palmitoyl-2-oleoyl-phosphocholine (POPC) bilayers in order to investigate the structural and dynamic properties of the leucine side chain using solid-state 2H NMR spectroscopy. Also, 15N-labeled TM-PLB at residues Leu28, Leu39, and Leu42 were studied using both 15N static and 15N CP-MAS NMR spectroscopy to investigate backbone mobility. Using the anisotropic 15N chemical shift as an orientational constraint, single-site 15N- labeled TM-PLB (at residue Leu39) was studied in 1,2-dioleoylphosphocholine (DOPC) bilayers oriented between thin glass plates. The resultant 15N chemical shift provided helical orientational information for TM-PLB with respect to the membrane.

5.2 Materials

Fmoc-amino acids and other chemicals for peptide synthesis were purchased from

15 Applied Biosystems Inc. (Foster City, CA). Fmoc-leucine-5,5,5-d3 and Fmoc-leucine- N derivatives were purchased from Isotec (Miamisburg, OH). 2H-depleted water was purchased from Isotec. Inc. (Miamisburg, OH). POPC and DOPC were purchased from Avanti Polar Lipids Inc. (Alabaster, AL). The phospholipids were shipped already dissolved in chloroform at a concentration of 20 mg/mL and stored at –20 OC. Chloroform, hexafluoro-2-propanol (HFIP),

98 formic acid and 2,2,2-trifluoroethanol (TFE) were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). HPLC-grade acetonitrile and isopropanol were obtained from Pharmco (Brockfield, CT) and were filtered through a 0.2-mm nylon membrane before use. Water was purified using a nanopure reverse osmosis system (Millipore, Bedford, MA). N-[2- hydroxyethyl]piperizine-N-2-ethane sulfonic acid (HEPES), trifluoroacetic acid (TFA), and EDTA were also obtained from Sigma-Aldrich (St. Louis, MO).

5.3 Methods 5.3.1 Peptide Synthesis Several 2H-labeled and 15N-labeled TM-PLB peptides were synthesized using predefined procedures (46). The polypeptides corresponding to the transmembrane segment of PLB Ala24- Leu52 (2H-Leu PLB and 15N-Leu PLB) were synthesized on an Applied Biosystems Inc. (ABI) 433A solid-phase peptide synthesizer controlled by a G3 Macintosh computer with SynthAssistTM 2.0 software as described previously (46).

5.3.2 Peptide Purification The synthesized 2H-labeled and 15N-labeled TM-PLB peptides were purified using the following procedure. The synthesized peptide was removed from the synthesizer and cleaved from the resin. The lyophilized peptide was dissolved in HFIP/FA (4:1, at a concentration of 5 mg/mL) and centrifuged to eliminate insoluble particulates. The crude peptide was purified on an Amersham Pharmacia Biotech AKTA‚ Explorer 10S HPLC controlled by UnicornTM version 3 system software. A polymer-supported column (259 VHP82215, 5 mm, 300-Å pore size, 2.2 x 15 cm) from Grace-Vydac was used to purify PLB. The column was equilibrated with 95% solvent A:5% solvent B. Solvent A consisted of H2O + 0.1% TFA and solvent B was 38%

MeCN + 57% IPA + 5% H2O (46). A 1-mL aliquot of the 5 mg/mL peptide sample was injected into the column and the gradient was ramped from 5% to 100% of solvent B at a flow rate of 10 mL/min. Peptide elution was achieved with a linear gradient to a final solvent composition of 93% solvent B. The purified peptide fraction was lyophilized and analyzed by MALDI-TOF mass spectroscopy using a matrix of 2,5-dihydroxybenzoic acid. The overall yield of PLB was about 37% (46).

99 5.3.3 Solid-State NMR Sample Preparation For the 2H NMR experiments, 76 mg of POPC dissolved in chloroform was mixed with 13.4 mg of appropriately 2H-labeled PLB (24-52) (4 mol% with respect to POPC) dissolved in a minimal amount of TFE. The sample was dried by passing a stream of nitrogen gas over the sample in the flask and then dried in a vacuum dessicator overnight. 190-mL HEPES buffer, prepared in 2H-depleted water, (5 mM EDTA, 20 mM NaCl, 30 mM HEPES, pH 7.0) was added to the peptide/lipid mixture in a 12 x 75 mm test tube and the peptide/lipid mixture was allowed to sit in a water bath at 50 OC for 30 minutes. The peptide-lipid mixture was occasionally agitated on a vortexer until a homogenous phase was established. The sample was loaded into a flat-bottom round glass tube of 5-mm o.d. for the NMR spectroscopy studies. After drying overnight in the dessicator the 15N NMR powder sample was transferred and packed into a 4-mm ZrO2 rotor. The hydrated sample was prepared by placing the rotor containing the dry sample in a relative humidity (~93%) chamber of saturated ammonium monophosphate. The sample was incubated for 6-12 h at a temperature of about 45 OC. The mechanically aligned 15N-labeled sample was prepared by cosolubilizing TM-PLB and DOPC/DOPE (4:1) in chloroform at a 1:200 mole ratio. The TM-PLB peptide was dissolved in a minimal amount of TFE prior to cosolubilizing with the lipids. The solution was spread onto 30 8.5 x 14 mm glass plates and allowed to air dry for 30 minutes before vacuum drying for another 24 h. Deuterium-depleted water was added onto the peptide/lipid mixture and the glass plates stacked on top of each other. The stacked glass plates were then placed in a humidity chamber consisting of saturated ammonium monophosphate at a relative humidity of about 93% at 42 OC for 12 h.

5.3.4 2H Solid-State NMR Spectroscopy All 2H solid-state NMR spectra were acquired on a Bruker Avance 500 MHz WB solid- state NMR spectrometer operating at a resonance frequency of 76.8 MHz for 2H using a double- resonance solid-state NMR probe equipped with a 5-mm solenoid coil. The quadrupolar echo pulse sequence was used with quadrature detection capabilities and complete phase cycling of the pulse pairs (47). A 3.0-ms 900 pulse, a sweep width of 100 kHz, a recycle delay time of 400 ms, and a 30-ms interpulse delay were used to accumulate 150,000 transients. Prior to Fourier

100 transformation, an exponential multiplication with 200-Hz line broadening was performed on the spectra. The spectra were acquired over a temperature range from –25 to 60 OC.

5.3.5 15N Solid-State NMR Spectroscopy Unoriented membrane protein samples were placed into a 4-mm sample rotor and inserted into in a Bruker triple resonance CP-MAS solid-state NMR probe (Bruker Avance 500 MHz WB solid-state NMR spectrometer) operating at 50.7 MHz for 15N. 15N solid-state NMR spectra were collected utilizing a standard cross polarization pulse sequence with 1H decoupling (48). The following pulse sequence parameters were used: 4.3-ms 1H 900 pulse, 500-ppm sweep width, 1.5-ms contact time and a 4-s recycle delay with 1H decoupling. Samples were spun at 5 kHz for the CP-MAS experiments. 40k scans were averaged for the static experiments and 4k scans were averaged for the CP-MAS experiments. The spectra were referenced to an external

15 standard of ( NH4)2SO4 (27 ppm). The experiments were recorded at temperatures ranging from -25 to 25 °C. The mechanically aligned 15N spectrum was collected using a Bruker double resonance flat-coil solid-state NMR probe operating at 25 °C. A standard cross-polarization pulse sequence was used with the following parameters: 4.5-ms 1H 90° pulse, 1.0-ms contact time, 600-ppm sweep width, and a 4-s recycle delay with 1H decoupling. 24 k scans were averaged and 300-Hz line broadening was used to process the data.

5.3.6 NMR Data Analysis Simulations of the 15N NMR spectra were carried out using the DMFIT software program that is capable of modeling one- and two- dimensional solid-state NMR spectra (49). The principal elements of the chemical shift tensors are represented according to the convention s33 ≥ s22 ≥ s11. Leucine deuterium lineshape simulations were performed using two independent axes defined with the MXQET software program (50). The first axis defined the tetrahedral three-site methyl group hopping in which the deuterons were tilted 75° with respect to the axis and with jumps of 120°. The second axis defined two-site methyl group jumping in which the methyl groups are tilted 75° and jump by 109.5°. The asymmetry parameter (h) was 0.05 for all of the 2H NMR simulations (34).

101 5.4 Results

The sequence corresponding to the peptide representing TM-PLB is displayed in Figure 5.2. The sequence represents part of the hinge segment (residues 24-30) and the entire transmembrane segment (residues 31-52) of PLB. The specific isotopically labeled sites (2H- labeled and 15N-labeled) are indicated in bold face.

28 39 42 51 ARQNLQNLFINFCLILICLLLICIIVMLL

Figure 5.2 The amino acid sequence corresponding to part of the hinge region and the transmembrane region of PLB is shown above. The amino acids in bold face represent the 2H-labeled or 15N-labeled sites used in this study.

The placement of the site-specific isotopically labeled residues was based upon the following conditions; Leu28 was chosen to be towards the N-terminus of the peptide and close to the cytosolic side of the membrane. Other residues were chosen so that they would be buried within the phospholipid bilayers (Leu39 and Leu42), and Leu51 was chosen because it is located towards the C-terminus and either close to the headgroup of the phospholipid bilayer or buried within the phospholipid hydrophobic core.

5.4.1 2H Side Chain Dynamics

2 Figure 5.3 shows the solid-state H NMR powder pattern spectra of specific labeled CD3-Leu28,

CD3-Leu39 and CD3-Leu51 TM-PLB samples incorporated into unoriented POPC bilayers. The spectra were recorded over the temperature range from –25 OC to 60 OC. In the gel phase (-25 OC), the deuterium lineshapes and quadrupolar splittings indicate that the only significant motion is methyl group rotation, since global motion is eliminated from the spectra by lowering the

102 temperature of the sample below the gel to liquid-crystalline phase transition of –3 OC (34). The quadrupolar splittings, Dv, for all three leucines are slightly different. The CD3-Leu28, CD3-

Leu39 and CD3-Leu51 spectra revealed quadrupolar splittings of about 30 kHz, 32 kHz and 36 kHz respectively. The 2H lineshapes of Leu28 and Leu39 are broad and rounded (bell-like shape), whereas the spectrum corresponding to Leu51 has a flatter top powder lineshape characteristic of more restricted motion. The marginal differences in the splitting and the lineshapes for all three leucines is attributed to the slightly different environment of these residues within the bilayer.

(A) Leu-28 (B) Leu-39 (C) Leu-51

0 0 0 60 C 60 C 60 C

0 0 0 45 C 45 C 45 C

0 0 25 C 0 25 C 25 C

0 0 0 0 C 0 C 0 C

0 -25 C 0 0 -25 C -25 C

40 20 0 -20 -40 40 20 0 -20 -40 40 20 0 -20 -40 2 2 2 H (kHz) H (kHz) H (kHz)

2 Figure 5.3 H NMR powder pattern spectra of L-leucine-5,5,5-d3 incorporated at specific sites of PLB and inserted into POPC phospholipid bilayers. 2H NMR

spectra are shown for (A) CD3-Leu28 PLB, (B) CD3-Leu39 PLB, and (C) CD3- Leu51 PLB incorporated into POPC phospholipid bilayers at lipid/peptide molar ratios of 25:1. The experiments were conducted at temperatures ranging from –25

103 to 60 OC. A total of 150,000 transients were accumulated for each spectrum and processed with a line broadening of 200 Hz.

At 0 OC, which is very close to the liquid-crystalline phase transition of the POPC phospholipids, additional leucine side chain motions are involved that decrease the quadrupolar

2 splitting of CD3-Leu28 and CD3-Leu39 significantly. The H peak centered at 0 kHz in the three spectra is due to residual 2H nuclei in the 2H-depleted water. Upon heating the sample to 25 OC, there is a dramatic change in the spectral width of the CD3-Leu28 spectrum characterized by a relatively narrow Pake doublet peak which exhibits a 2H lineshape that is characteristic of a h > 0 pattern (44). This is explained by an increase in side chain motions associated with two-fold jumps between predominant leucine side chain rotamers as well as three-fold methyl group reorientations (15). This causes a reduction in the spectral width with a quadrupolar splitting of about 6 kHz, a typical value for leucine buried within the hydrophobic region of a membrane at that temperature (29). Small changes in the quadrupolar splitting are observed at higher temperatures (45 OC and 60 OC), indicating faster motional averaging attributed to faster two-fold and three-fold jump rates. The shoulders at the edge of the powder are due to rapid methyl group reorientations that are faster than the NMR time scale (10-4 s). Our current model of TM-PLB suggests that Leu28 is located towards the N-terminus of the peptide, which is often a more mobile section of a transmembrane a-helix (38).

O At 25 C, CD3-Leu39 exhibits a spectral line shape with characteristics not typical of Pake doublets and this may be due to additional modes of motions (32). The narrow Lorentzian lineshapes are indicative of either fast side-chain reorientation both rapidly and isotropically or off-axis motion. At higher temperatures (45 OC and 60 OC), the spectral line shapes are maintained, though the breadth observed at 25 OC has merged with the isotropic peak. The disappearance of the breadth component at higher temperatures could also be due to global motion of the peptide along the long molecular axis and increased librational motions of the side chain resulting in an averaging of the electric field gradient tensors, and consequently giving a narrow powder pattern line shape (45).

O 2 At 25 C, H NMR powder spectra of CD3-Leu51 exhibit bell-shaped characteristics indicative of slower methyl group reorientations than the NMR time scale. The appearance of a small Pake doublet at this temperature could also indicate the presence of a more rapidly moving

104 population of PLB superimposed on a slower, broader component. Narrowing of the spectra is

O O further observed at higher temperatures (45 C and 60 C), but not as significantly as in the CD3-

Leu28 and CD3-Leu39 spectra. The unique features associated with these spectra could be explained by the condition that the Leu51 side chain is involved in the so-called “knobs-into- bolts” bonding arrangement in which the side chain of Leu51 from one a-helix is locked into the groove of a second a-helix (of the pentamer) so that the two a helices are fastened to each other. Smith and coworkers have also reported that Leu44 is buried within the core of pentameric phospholamban and also observed the general bell-shaped characteristics (34). This type of structural arrangement is believed to help stabilize the structure of the pentamer (6, 11, 15).

5.4.2 15N Backbone Dynamics The 15N chemical shift powder pattern from unoriented PLB protein samples in lipid bilayers can distinguish between mobile and rigid backbone sites on the 10-4 s timescale (36). Generally, it has been found that 15N spectral lineshapes of unoriented samples provide reliable evidence for the presence of mobile sites. To determine the degree of backbone mobility of TM- PLB in POPC bilayers, site-specific 15N-Leu TM-PLB peptides were incorporated into POPC phospholipid bilayers and the spectra were collected at -25 OC and 25 OC. 15N CP-MAS spectra were collected for each of the samples at a spinning speed of 5 kHz. Figure 5.4(D) shows the 15N CP-MAS spectrum of Leu28. All spectra showed single isotropic peaks near 120 ppm. The 15N static chemical shift powder pattern spectra of the hydrated samples of 15N-Leu TM-PLB (Leu28, Leu39, and Leu42) in POPC bilayers are displayed in Figure 5.4 at 25 OC. The smooth axially symmetric powder patterns (A, B & C) and the absence of any major isotropic components are an indication that the PLB peptide has limited backbone mobility in the POPC phospholipid bilayers. CP-MAS chemical shift values of about 129 ppm are characteristic of b-sheet structures, whereas chemical shift values lower than 122 ppm are characteristic of an a-helix (51). The CPMAS chemical shift value of 120 ppm corresponds to a leucine residue contained within an a-helical structure in a lipid bilayer, implying the parameters found in the powder pattern spectra are relevant to the secondary structure of PLB in POPC phospholipid bilayers. The lineshapes and intensities of the 15N NMR spectra exhibited small deviations from each other at each labeled site along the backbone, indicating the absence of motional averaging, and that the residues are all

105 (A) Leu28

(B) Leu39

(C)

Leu42

(D) CPMAS Leu28

300 200 100 0 15 N (ppm)

Figure 5.4 15N NMR powder pattern spectra of 15N-labeled PLB incorporated into POPC bilayers at a lipid to peptide ratio of 25:1. The static powder pattern spectra (A) Leu28, (B) Leu39, and (C) Leu42 were obtained after 200,000 transients were collected. The experimental spectra were collected at 25 OC after hydration in a relative humidity chamber (93%) for 12 h. The simulated solid- state NMR spectra (dotted line) are superimposed on the experimental spectra for comparison. Chemical shift anisotropy values, with errors within ± 3 ppm, were calculated from the simulated spectra and are shown in Table 1. (D) High power proton-decoupled CP-MAS spectrum of hydrated 15N-labeled Leu28 PLB inserted into POPC bilayers and spun at 5 kHz.

transmembrane. Also, the lineshapes for specific sites exhibited small changes as the temperature was increased from -25 OC (data not shown) to 25 OC. Thus, the experimental spectra in Figure 5.4 show 15N amide powder pattern spectra indicative of leucines in PLB that

106 have immobile peptide bonds located inside the membrane (36). This conclusion is further supported by the appearance of a single resonance peak and the absence of any other components in the CP-MAS spectra (one example is shown in Figure 5.4 (D)). Thus, highly constrained backbone sites of TM-PLB suggest that the side chain dynamics are not influenced by backbone fluctuations. The experimental spectra of Figure 5.3 were simulated and the results are displayed as dotted lines. The experimental spectra and the simulated spectra were superimposed and the chemical shift tensor elements extracted from the single site labeled PLB peptides. The chemical shift tensor elements are displayed in Table 1 for comparison. The 15N chemical shift tensor

15 values for N-Leu39 PLB were found to be s11 = 50.5 ppm, s22= 80.5 ppm, and s33 = 229 ppm ± 3 ppm. These values are comparable to other 15N tensoral components found in the literature (51). Similar parameters were also obtained from the 15N-labeled Leu42 sample (Table 1). The observation of chemical shift values close to the extremes of the chemical shift anisotropy and the absence of any isotropic components demonstrates that all of the labeled residues investigated are transmembrane and are motionally restricted by interactions with the lipid membrane (52). However, it is interesting to note that the CSA corresponding to Leu28 was about 170 ppm, which is less than the CSA corresponding to Leu39 (178 ppm) and Leu42 (179 ppm). This implies that slight backbone motion is centered around Leu28 when compared to the other residues.

5.4.2 Helical Orientation of PLB with Respect to the Lipid Bilayer

The one-dimensional 15N solid-state NMR spectrum of 15N-labeled Leu39 PLB was obtained in fully hydrated DOPC/DOPE bilayers mechanically aligned on glass plates. As expected for a well-oriented sample, the spectrum in Figure 5.5(A) consists of a relatively narrow single line with a resonance frequency within the span of the 15N chemical shifts anisotropy powder pattern.

107 (A)

s^

(B)

sII

300 250 200 150 100 50 0

15 N (ppm)

Figure 5.5 (A) One-dimensional solid-state 15N NMR spectrum of 15N-labeled Leu39 PLB inserted into DOPC/DOPE phospholipid bilayers mechanically oriented on glass plates. (B) For comparison, the solid-state 15N NMR spectrum of 15N-labeled Leu39 PLB inserted into unoriented POPC phospholipid bilayers.

15 15 The N NMR spectra were referenced to external NH4(SO4)2 at 27 ppm.

An amide N-H bond approximately parallel to the direction of the static magnetic field and the

15 membrane normal has an N resonance frequency near that of the principal tensor element s||. Conversely, an amide N-H bond perpendicular to the field and membrane normal has a 15N resonance frequency near that of the principal tensor element s^ of the chemical shift powder

108 pattern spectra. Comparison of the spectra in Figure 5.5 clearly shows the effect of sample orientation. The most striking feature of the aligned spectrum shown in Figure 5.5(A) is that the single line 15N resonance observed from the specific 15N-labeled Leu39 sample has a frequency near the s|| edge of the amide chemical shift powder pattern in Figure 5.5(B). Thus, the chemical shift value of 220 ppm indicates that the peptide plane is transmembrane and nearly parallel with the membrane normal and the static magnetic field in aligned DOPC/DOPE phospholipid bilayers. This is further evidence that the PLB peptide has been successfully incorporated into DOPC/DOPE bilayers.

5.5 Discussion

5.5.1 Dynamics of PLB in Phospholipid Bilayers Model membranes have been used in recent years to study the orientation and dynamics of membrane-bound proteins in both randomly dispersed and oriented systems (36, 53-57). Even though the study of such systems is challenging, especially when the lipids are in the liquid- crystalline phase characterized by a high degree of molecular disorder, much progress has been made in terms of understanding the physiological role and structure of integral membrane proteins (29). The current study summarizes our understanding of the dynamics of side chain leucines, the location, and the backbone mobility of the peptide representing the transmembrane segment of PLB in model membranes of POPC. The general shapes of the spectra are discussed in terms of pentameric TM-PLB incorporated into phospholipid bilayers. Several labs have shown that a PLB/lipid mole ratio as low as 1:100 will spontaneously assemble into a pentamer in DOPC phospholipid bilayers (16, 58, 59). However, such studies have also indicated the possibility of dynamic exchange between monomeric and pentameric structures of PLB (16, 58, 59).

5.5.2 Structural Implications of the Leucine Side-Chain Motions It is assumed in all of our discussions that the Leu side chain exists predominantly in only 2 of the 9 rotamer conformations separated by an angle of about 110O (60). Previous studies by Ying and coworkers (34) indicated that rotational, jumping, and librational motions about these bonds result in unique features of the deuterium lineshapes. Wittebort and coworkers have used

109 2H NMR spectroscopy to analyze lineshapes in anisotropic media (61). In order to better understand the dynamics of our 2H side chain motion, the spectra shown in Figure 5.3 were simulated and representative best fits of the data are displayed in Figure 5.6.

The lineshapes corresponding to CD3-Leu51 (Figure 5.6(A)) indicate methyl site hopping motions with jump rates between 104 and 106 Hz, and a superposition of a broad isotropic peak.

O The spectrum corresponding to CD3-Leu at –25 C is much broader than the other spectra collected at higher temperatures due to the slower side chain motions at low temperatures. The 2H quadrupolar splitting at –25 OC is consistent with previous studies that indicate that these residues are located inside the interior membrane, where the hydrocarbon chains are packed tightly in a crystalline lattice. Generally, for leucine residues in a crystalline lattice environment, the quadrupolar splitting is approximately 40 kHz (34, 44). Therefore, the values recorded from these studies, provide evidence for the penetration of all three labeled leucines within the hydrophobic region of the bilayer, though such values are slightly smaller than 40 kHz. Because Leu51 is believed to be involved in helix-helix interactions (13), the resulting quadrupolar splitting is greater than that of the corresponding CD3-Leu39 and CD3-Leu28 spectra which are more free to have other side chain motions such as jumping between the predominant rotamers. This additional side chain motion on the order of 102 Hz creates the more bell-shaped appearance of the Leu39 and Leu28 spectra at –25 °C as well as decreases the corresponding quadrupolar splittings by about 6 kHz (34). The broad isotropic peak in Figure 5.6(A) is attributed to residual water. At 0O C, the simulations of Leu28 and Leu39 in Figure 5.6(B) suggest that the 2H NMR spectra that contain three different components. The first component results from methyl group three-site jump rates of 1 x 107 Hz and two-site hopping motions with jump rates of 2 x 104 Hz, the second is a broader peak as in the lower temperature simulation, and the third is a sharp isotropic residual water peak. These two spectra are in stark contrast to the spectrum corresponding to Leu51 at 0 °C which looks similar to the –25 °C spectra of Leu28 and Leu39, but with a sharp residual water isotropic peak in the center. This indicates that at 0 °C Leu51 is undergoing much slower methyl group motion with three-site rates equal to 1 x 106 Hz and two site jumping rates on the order of 102 Hz. - At temperatures above 25 °C, the shape of the spectra of Leu39 (Figure 5.3(B)) is not easily simulated because of possible protein-lipid interactions. Molecular dynamic studies have

110 indicated that fluctuations of both the lipids and the protein are much stronger at the hydrophobic core center of the lipid bilayer causing the side chain to be very mobile in that area (62, 63). The 2H NMR spectra of Leu39 are more complicated than the terminal amino acids because of the very different environments and different types of motions (such as librational) that are involved. At temperatures above 25 °C, Leu28 side chains undergo rapid motions associated with methyl group rotations as well as two-site hoping motions at jump rates corresponding to 1 x 108 Hz and 6 x 108 Hz respectively as determined from the simulated spectrum which also include a sharp residual water isotropic component (Figure 5.6(D)).

(A)

3

(B)

2

(C)

1

(D)

0 40 20 0 -20 -40 3 2x10 H (kHz)

2 Figure 5.6 Simulations of CD3-Leu NMR lineshapes from the experimental H NMR spectra displayed in Figure 5.3. Parameter inputs used in the MXQET

111 program to create these model simulations are specified in the Materials and Methods and Discussion sections. Simulations correspond to models for (A) all Leu-PLB spectra at –25 °C, and Leu51 at 0 °C, (B) Leu28 and Leu39 at 0 °C and Leu51 at 25 °C, (C) Leu51 at 45 °C and 60 °C, and (D) Leu28 at 25 °C, 45 °C, and 60 °C.

2H NMR spectra of Leu51 at 25 °C, 45 °C, and 60 °C provide the most unique information about TM-PLB in phospholipid bilayers. It is readily apparent that the quadrupolar splittings in the spectra are broader than both Leu39 and Leu28 at the same temperatures. This is the first indication of slower Leu51 side chain motions, when compared to the other two residues. Also, the 2H Leu51 NMR spectra (Figure 5.3) are more complicated. At 25 °C, the Leu51 spectrum has one component with similar features as the spectra of Leu28 and Leu39 at a lower temperature of 0 °C, and a second component consisting of a Pake doublet with a splitting of about 8 kHz. We hypothesize that the unique side chain motions observed in the Leu51 spectrum at 25 °C result from the superposition of two different oligomerization states of TM- PLB. This agrees with previous studies that indicated the coexistence of both pentameric and monomeric populations of PLB (16, 58, 59). The fast moving less intense component (monomer) has rates comparable to Leu28 at 25 °C (Figure 5.6(D)), and a second slower dynamic component (pentamer) similar to that of Figure 5.6(B). Also, two components were not needed to simulate the Leu28 and Leu39 2H NMR spectra because they are not directly involved in the same Leu zipper formation as Leu51. At 45 °C, Leu51 can be simulated (Figure 5.6(C)) as a fast moving side chain with a strong isotropic peak in the center and three-site jump speeds of greater than 1 x 106 Hz and two-jump rates of greater than 1 x 106 Hz. The Pake doublet is very similar to that of Leu28 except for a slightly larger quadrupolar splitting. The larger quadrupolar splitting indicates more restricted side-chain motion. The second population of PLB in the monomeric form was taken under consideration in the 45 °C and 60°C spectral simulations by modeling a fast moving side chain nearly identical to Leu28, and a superimposition of a broader isotropic peak. The second component can be easily seen in the formation of the rounder bell shape 2H NMR spectrum of Leu51 at 60 °C. Our results from the simulations agree well with previous studies on the PLB pentameric channel (15). Ying and coworkers have targeted Leu42, Leu43 and Leu44 in previous studies to

112 establish the relative motion of the side chains and rotational reorientation of the phospholamban helices. The observed lineshapes at 25 °C are identical with the lineshapes from our 2H NMR studies. The conclusions drawn from our studies agree with the mutagenesis studies by Adams and coworkers. Previous work by MacLennan and coworkers has proposed that Leu51 of PLB is one of the residues lying on the interior face of the helix, stabilizing the pentamer (64). Thus, the slower Leu51 motion as simulated with the slower jump rates is probably due to restricted motion of this residue as a result of the sidechain orientation towards the helix interfaces that are involved in the leucine zipper motif. Recent studies by Asahi and coworkers have observed enhanced physical interaction between SERCA1a and the Leu28 mutant indicating that Leu28 is oriented towards the helix core (65). However, several studies have been carried out reporting contradictory results about the secondary structure of PLB in the membrane. One such model proposed that Leu28 lies in a position where it is a part of an antiparallel b-sheet region of the full length PLB (25). Arkin and coworkers have argued on the basis of FTIR studies that the full-length PLB is a-helical (24). The partial 15N backbone motional averaging of Leu28 (CSA of 170) indicates an a-helical peptide residue located towards the amino-terminus and buried within the bilayer. As shown in

Table 1, the magnitude of the difference between s11 and s33 corresponding to the breadth of the observed axially symmetric powder pattern are slightly different for all three residues within the limits of experimental error. The large breadth observed for Leu42 and Leu39 (CSA width of about 178 ppm) indicates that the carboxy-terminus residues of the peptide are relatively stable by virtue of backbone i Æ i + 4 intramolecular H-bonding (38). However, the 15N CSA width of 15N-labeled Leu28 is slightly reduced to 170 ppm, indicating the residue is weakly stabilized by backbone i Æ i + 4 intramolecular H-bonding when compared to Leu39 and Leu42 (38). It was also noted that at –25 OC, the 15N powder pattern spectrum of 15N-labeled Leu28 has a slightly broader CSA (180 ppm) powder pattern when compared to the spectrum at 25 OC (data not shown). The difference in CSA width at the two temperatures is further evidence of increased

15 mobility at the N-terminus of the peptide at higher temperatures in the La phase. The N CSA width of both 15N-labeled Leu42 and Leu39 were 179 ppm at the two temperatures (–25 OC and 25 OC). Chemical and molecular biological evidence supports the hypothesis that phospholamban exerts its action on Ca2+-ATPase through direct protein-protein interactions (21). The

113 Table 1 15N chemical shift tensor values of 15N-labeled PLB (Leu28, Leu39 and Leu42) inserted into POPC phospholipid bilayers. These values were extracted from the simulated solid-state NMR spectra in Figure 5.3.

CSA values measured from single site 15N-labeled Leu-PLB powder

15 Samples [in ppm relative to ( NH4)2SO4 solution referenced to 27 ppm]

Sites s11 s22 s33 siso CSA

Leu28PLB 55.0 82.5 225.0 121.0 170.0

Leu39PLB 50.5 80.5 229.0 120.0 178.0

Leu42PLB 52.0 84.8 230.0 121.0 179.0 hydrophobic side chains could fit into complementary pockets of the pump thereby accounting for hydrophobic and electrostatic interaction of the two proteins. Evidence of such interactions was shown with an I40L mutation on a zipper domain residue of I40LPLB, which was able to partially assemble into a pentamer yet decreased the apparent Ca2+ affinity of the pump as potently as a complete monomeric mutant substituted with Ala (66). Therefore, it may be appropriate to assume that any residue within the zipper domain should have a potent effect on the Ca2+ pump and any residue adjacent to the pump has no effect on the pump. A model pentameric structure of PLB was previously determined with a PDB file # 1N7L and will be used to discuss our results (67). As shown in Figure 5.7(A), the structure of the PLB pentamer is composed of a-helical monomers, associated together through the interaction of specific residues. From our data Leu51, which is shown in green, is among such residues involved in the helix-helix interaction inside the channel and is buried within the helix core, whereas Leu39 (shown in yellow) and Leu42 residues are pointing towards the lipid acyl chains and are not involved in helix-helix interactions (23). Therefore, the stabilization of the PLB pentamer is due to the side chains of leucine and isoleucine of specific residues of the monomers interlocking in

114 the crevices of opposite strands. The evidence of the type described above suggests that residues such as Leu51 within the zipper domain may be important regulatory components interacting with the pump, and a mutation of these residues will have potent effect on the activity of the pump.

5.5.3 Structural model of TM-PLB in lipid bilayers Activity of a number of membrane proteins including the Ca2+-ATPase has been shown to be dependent on bilayer thickness (62, 68). The activity of the pump is highest when the ATPase is reconstituted into bilayers of di(18:1)PC, whereas the activity is lower in bilayers of shorter fatty acyl chains (62, 69, 70). The activity varies little in the chain length range from C16 to C20, which is the range of fatty acyl chains found in the sarcoplasmic reticulum (SR) (71). The structural model displayed in Figure 5.7(B) for TM-PLB, indicates the location of the residues in POPC phospholipid bilayers. The hydrophobic thickness of POPC has been estimated to be about 27 Å. Assuming the number of amino acids per turn in an a-helix peptide is 3.6 residues, and the pitch is 5.4 Å, then the translation per residue along the a-helical peptide will be 1.5 Å. Thus, to be able to span the hydrophobic region of the membrane, 18 residues are required. That position Leu39 near the center of the hydrophobic core of the POPC phospholipid

2 bilayers. This agrees well with our H NMR data of increased CD3-Leu side chain motion caused by acyl chains in the center of the bilayer. The phospholipid headgroups have a thickness of about 5 Å, which places Leu28 and Leu51 within the hydrophobic region of the lipids and very close to the polar headgroup towards the N-terminal and the C-terminal regions respectively of the PLB peptide as depicted by the model in Figure 5.7 (B). This supports a transmembrane PLB segment that almost completely spans the bilayer and matches the bilayer thickness, thus having little to no tilt with respect to the bilayer.

115 (A)

(B)

Figure 5.7 Structural models of phospholamban in phospholipid bilayers showing the location of the leucine residues. (A) Pentamer model looking down the channel. Leu39, shown in yellow, is indicated as a surface residue and not involved in helix-helix interactions. Leu51, shown in green, is involved in stabilizing the pentamer and is considered to be oriented inside the pentamer. (B) PLB is displayed as a a-helical monomeric unit (for visual simplicity) showing the location of specific labeled residues within the bilayer.

5.6 Conclusion

In this work, we have probed the dynamics of site-specific labeled -CD3 groups of leucine in TM-PLB as a function of temperature. The quadrupolar splittings associated with the side

116 chain reorientations at lower temperatures were influenced by three-state jumps of the CD3-Leu group. Upon increasing the temperature, Ca-Cg and Cg-Cd rotations decreased the quadrupolar splittings. Leu51 spectra at all temperatures are slower moving than Leu28 and Leu39 because of its involvement in the leucine zipper motif. Site-specific 15N-labeled PLB NMR studies were used to characterize the backbone dynamics and our results indicated restricted motion for the transmembrane segment of PLB, but subtle backbone mobility centered around Leu28. Finally, the 220-ppm 15N chemical shift value for 15N-labeled Leu39 oriented in DOPC lipid bilayers indicates a helix that is nearly collinear with the bilayer normal. These results are displayed pictorially in the structural model of the monomer form of PLB in the membrane (Figure 5.7(B)).

117 5.7 References 1. Li, H., Cocco, M. J., Steitz, T. A., and Engelman, D. M. (2001) Conversion of phospholamban into soluble pentameric helical bundle, Biochemistry 40, 6636-6645. 2. Fuji, J., M. Kadoma, H. Tada, and F. Sakiyama. (1986) Characterization of structural unit of phospholamban by amino acid sequencing and electrophoretic analysis, Biochem. Biophys. Res. Commun. 138, 1044-1050. 3. Kimura, Y., Kurzydlowski, K., Tada, M., and MacLennan, D. H. (1996) Phospholamban regulates the Ca2+-ATPase through intramembrane interactions, J. Biol. Chem. 249, 6166-6173. 4. Cornea, R., Autry, J., Chen, Z., and Jones, L. (2000) Re-examination of the role of the leucine/isoleucine zipper residues of phospholamban in inhibition of the Ca2+-pump of cardiac sarcoplasmic reticulum, J. Biol. Chem. 275, 41487-41494. 5. Minamisawa, S., Hoshijima, M., Chu, G., Martone, M. E., Wang, Y., Ross, J., Jr.,, Kranias, E. G., Giles, W. R., and Chien, K. R. (1999) Chronic phospholamban- sarcoplasmic reticulum calcium ATPase interaction is the critical calcium cycling defect in dilated cardiomyopathy, Cell 99, 313-322. 6. Simmerman, H. K., Collins, J. H., Theibert, J. L., Wegender, A. D., and Jones, L. R. (1996) sequence analysis of phospholamban. Identification of phosphorylation sites and two major structural domains, J. Biol. Chem. 261, 13333-13341. 7. Simmerman, H. K., Collins, J. H. Theibert, J. L., Wegener, A. D., and Jones. L. R.(1986) Sequence analysis of phospholamban: identification of phosphorylation sites and two major structural domains, J. Biol. Chem. 261, 13333-13341. 8. Mascioni, A., Karim, C., Zamoon, J., Thomas, D. D., and Veglia, G. (2002) Solid-state NMR and rigid body molecular dynamics to determine domain orientations of monomeric phospholamban, J. Am. Chem. Soc. 124, 9392-9393. 9. Pollesello, P., Annila, A., and Ovaska, M. (1999) Structure of the 1-36 amino-terminal fragment of human phospholamban by nuclear magnetic resonance and modeling of the phospholamban pentamer, Biophys. J., 1784-1795.

118 10. Lamberth, S., Schimidt, H., Muenchbach, M., Vorherr, T., Krebs, J., Carafoli, E., and Griesinger, C. (2000) NMR solution structure of phospholamban, Helv. Chim. Acta. 83, 2141-2152. 11. Simmerman, H. K. B., Lovelace, D. E., and Jones, L. R. (1989) Secondary structure of detergent-solubilized phospholamban, a phosphorylatable, oligomeric protein of cardiac sarcoplasmic reticulum, Biochim. Biophys. Acta, 322-329. 12. Arkin, I. T., Adams, P. D., MacKenzie, K. R., Lemmon, M. A., Brunger, A. T., and Engelman, D. M. (1994) Structural organization of the pentameric transmembrane alpha helix of phospholamban, a cardiac ion channel, Embo. J. 13, 4757-4764. 13. Arkin, I. T., Adams, P. D., Aimoto, S., Engelman, D. M., and Smith, S. O. (1996) Structure of the transmembrane cysteine residues in phospholamban, J. Membr. Biol. 14. Birmachu, W., and Thomas, D. D. (1990) Rotational dynamics of the Ca-ATPase in sarcoplasmic reticulum studied by time-resolved phosphorescence anisotropy, Biochemistry 29, 3904-3914. 15. Smith, S. O., Kawakami, T., Liu, W., Ziliox, M., and Aimoto, S. (2001) Helical structure of phospholamban in membrane bilayers, J. Mol. Biol. 313, 1139-1148. 16. Cornea, R. L., Jones, L. R., Autry, J., and Thomas, D. D. (1997) Mutation and phosphorylation change the oligomeric structure of phospholamban in lipid bilayers, Biochemistry 36, 2960-2967. 17. Sasaki, T., Inui, M., Kimura, Y., Kuzuya, T., and Tada, M. (1992) Molecular mechanism of regulation of Ca2+ pump ATPase by phospholamban in cardiac sarcoplasmic- reticulum -effects of synthetic phospholamban peptides on Ca2+ pump ATPase, J. Biol. Chem. 267, 1674-1679. 18. Toyofuku, T., Kurzydlowski, K., Tada, M., and MacLennan, D. H. (1994) Amino-acids Glu(2) to Ile(18) in the cytoplasmic domain of phospholamban are essential for functional association with the Ca2+-ATPase of sarcoplasmic-reticulum, J. Biol. Chem. 269, 3088-3094. 19. Hughes, J., East, J. M., and Lee, A. G. (1994) The hydrophilic domain of phospholamban inhibits the Ca2+ transport step of the Ca2+-ATPase, Biochem. J. 303, 511-516.

119 20. Simmerman, H. K. B., Kobayashi, Y. M., Autry, J. M., and Jones, L. R. (1996) A leucine zipper stabilizes the pentameric membrane domain of phospholamban and forms a coiled-coil pore structure, J. Biol. Chem. 271, 5941-5946. 21. Adams, P. D., Arkin, I. T., Engelman, D. M., and Brunger, A. T. (1995) Computational searching and mutagenesis suggest a structure for the pentameric transmembrane domain of phospholamban, Nat. Struct. Biol. 2, 154-162. 22. Ahmed, Z., Reid, D. G., Watts, A., and Middleton, D. A. (2000) A solid-state NMR study of the phospholamban transmembrane domain: local structure and interactions with Ca2+-ATPase, Biochim. Biophys. Acta 1468, 187-198. 23. Hughes, E., and Middleton, D. A. (2003) Solid-state NMR reveals structural changes in phospholamban acompanying the functional regulation of Ca2+-ATPase, J. Biol. Chem. 278, 20835-20842. 24. Arkin, I. T., Rothman, M., Ludlam, C. F., Aimoto, S., Engelman, D. M., Rothschild, K. J., and Smith, S. O. (1995) Structural model of the phospholamban ion channel complex in phospholipid membranes, J. Mol. Biol. 248, 824-834. 25. Tatulian, S. A., Jones, L. R., Reddy, L. G., Stokes, D. L., and Tamm, L. K. (1995) secondary structure and orientation of phospholamban reconstituted in supported bilayers from polarized attenuated total reflection FTIR spectroscopy, Biochemistry 34, 4448-4456. 26. Frank, S., Kammerer, R. A., Hellstern, S., Pegoraro, S., Stefeld, J., Lustig, A., Moroder, L., and Engel, J. (2003) Toward a high-resolution structure of phospholamban: Design of soluble transmembrane domain mutants., Biochemistry 39, 6825-6831. 27. Prosser, R. S., Daleman, S. I., and Davis, J. H. (1994) The structure of an integral membrane peptide: A deuterium NMR study of gramicidin, Biophys. J. 66, 1415-1428. 28. Sharpe, S., Barber, K. R., and Grant, W. M. (2002) Evidence of a tendency to self- association of the transmembrane domain of ErbB-2 in fluid phospholipid bilayers, Biochemistry 41, 2341-2352. 29. Koenig, B. W., Ferreti, J. A., and Gawrisch, K. (1999) Site-specific deuterium order parameters and membrane-bound behavior of a peptide fragment from the intracellular domain of HIV-1 gp41, Biochemistry 38, 6327-6334.

120 30. Mack, J. W., Torchia, D. A., and Steinert, P. M. (1988) Solid-state NMR studies of the dynamics and structure of mouse keratin intermediate filaments, Biochemistry 27, 5418- 5426. 31. Beshah, K., Olejniczak, E. T., and Griffin, R. G. (1987) Deuteriun NMR study of methyl group dynamics in L-alanine, J. Chem. Phys. 86, 4730-4736. 32. Keniry, M. A., Kintanar, A., Smith, R. L., Gutowsky, H. S., and Oldfield, E. (1984) Nuclear magnetic resonance of amino acids and proteins. Deuterium nuclear magnetic resonance relaxation of deuteriomethyl-labeled amino acids in crystals and in Halobacterium halobium and E. coli cell membranes, Biochemistry 23, 288-298. 33. Batchelder, L. S., Sullivan, C. E., Jelinski, L. W., and Torchia, D. A. (1982) Characterization of leucine side-chain reorientation in collagen-fibrils by solid-state deuterium NMR, Proc. Natl. Acad. Sci. 79, 386-389. 34. Ying, W., Irvine, S. E., Beekman, R. A., Siminovitch, D. J., and Smith, S. O. (2000) deuterium NMR reveals helix packing interactions in phospholamban, J. Am. Chem. Soc. 122, 11125-11128. 35. Kinsey, R. A., Kintanar, A., and Oldfield, E. (1981) Dynamics of amino acid side chains in membrane proteins by high field solid state deuterium nuclear magnetic resonance spectroscopy, J. Biol. Chem 256, 9028-9036. 36. Leo, G. C., Colnago, L. A., Valentine, K. G., and Opella, S. J. (1987) Dynamics of fd coat protein in lipid bilayers, Biochemistry 26, 854-862. 37. Whiles, J. A., Brasseur, R., Glover, K. J., Melacini, G., Komives, E. A., and Vold, R. R. (2001) Orientation and effects of mastoparan X on phospholipid bicelles, Biophys. J. 80, 280-293. 38. Jones, D. H., Rigby, A. C., Barber, K. R., and Grant, W. M. (1997) Oligomerization of the EGF receptor transmembrane domain: A 2H NMR study in lipid bilayers, Biochemistry 36, 12616-12624. 39. Van der Wel, P. C. A., Strandberg, E., Killian, J. A., and Koeppe II, R. E. (2002) Geometry and intrinsic tilt of a tryptophan-anchored transmembrane alpha-helix determined by deuterium NMR, Biophys. J. 83, 1479-1488.

121 40. Killian, J. A., Taylor, M. J., and Koeppe II, R. E. (1992) Orientation of the valine-1 side chain of the gramicidin transmembrane channel and implications for channel functioning: a deuterium NMR study, Biochemistry 31, 11283-11290. 41. Lee, k.-C., Huo, S., and Cross, T. A. (1995) Lipid-peptide interface:valine conformation and dynamics in the gramicidin channel in a lipid bilayer, Biochemistry 34, 857-867. 42. Janin, J., and Wodak, S. (1978) Conformation of amino acid side-chain in proteins, J. Mol. Biol. 125, 357-386. 43. Keniry, M. A. (1989) Solid-state deuterium nuclear magnetic resonance spectroscopy of proteins, Meth. Enzymol. 176, 376-386. 44. Batchelder, L. S., Sullivan, C. E., Jelinski, L. W., and Torchia, D. A. (1982) Characterization of leucine side-chain reorientation in collagen-fibrils by solid-state 2H NMR, Proc. Natl. Acad. Sci. 79, 386-389. 45. Lee, K.-C., and Cross, T. A. (1994) Side-chain structure and dynamics at the lipid- protein interface: Val1 of the gramicidin A channel, Biophys. J. 66, 1380-1387. 46. Tiburu, E. K., Dave, P. C., Vanlerberghe, J. F., Cardon, T. B., Minto, R. E., and Lorigan, G. A. (2003) An improved synthetic and purification procedure for the hydrophobic segment of the transmembrane peptide phospholamban, Anal. Biochem. 318,146-151 47. Davis, J. H., Jeffrey, K. R., Bloom, M., and Valic, M. I. (1976) Quadrupolar echo deuterium magnetic resonance spectroscopy in ordered hydrocarbon chains, Chem. Phys. Lett. 42, 390-394. 48. Pines A., Gibby, M., and Waugh, J. S. (1973) Proton-enhanced NMR of dilute spins in solids, J. Chem. Phys. 59, 569-590. 49. Massiot, D., Fayon, F., Capron, M., King, I., Le calve, S., Alonso, B., Durand, J. O., Bujoli, B., Gan, Z., and Hoatson, G. (2002) Modeling one- and two-dimensional solid- state spectra, Magn. Reson. Chem. 40, 70-76. 50. Greenfield, M. S., Ronemus, A. D., Vold, R. L., Vold, R. R., Ellis, P. D., and Raidy, T. E. (1987) Deuterium quadrupole-echo NMR spectroscopy III. Practical aspects of lineshape calculations for multiaxis rotational processes, J. Magn. Reson. 72, 89-107. 51. Wildman, K. A. H., Lee, K.-D., and Ramamoorthy, A. (2003) Mechanism of lipid bilayer disruption by the human antimicrobial peptide, LL-37, Biochemistry 42, 6545-6558.

122 52. Hartzell, C. J., Whitfiled, M., Oas, T. G., and Drobny, G. P. (1987) Determination of the 15N and 13C chemical shift tensors of L-[13C]alanyl-L[15N]alanine from the dipole- coupled powder patterns, J. Am. Chem. Soc. 109, 5966-5969. 53. Ketchem, R. R., Hu, W., and Cross, T. A. (1993) High-resolution conformation of gramicidin-A in a lipid bilayer by solid-state NMR, Science 261, 1457-1460. 54. Keniry, M. A., Gutowsky, H. S., and Oldfield, E. (1984) Surface dynamics of the integral membrane-protein bacteriorhodopsin, Nature 307, 383-386. 55. Opella, S. J., Frey, M. H., and Cross, T. A. (1979) Three-dimensional structure of the membrane embedded M2 channel-lining segment from nicotinic acetylcholine receptor, J. Am. Chem. Soc. 101, 5856-5857. 56. Ulrich, A. S., Watts, A., Wallat, I., and Heyn, M. P. (1994) Distorted structure of retinal chromophore in bacteriorhodopsin resolved by deuterium NMR, Biochemistry 33, 5370- 5375. 57. Prosser, R. S., and Davis, J. H. (1994) Dynamics of an integral membrane peptide: a deuterium NMR relaxation study of gramicidin, Biophys. J. 66, 1429-1140. 58. Reddy, L. G., Jones, L. R., and Thomas, D. D. (1999) Depolymerization of phospholamban in the presence of calcium pump: A fluorescence energy transfer study, Biochemistry 38, 3954-3962. 59. Li, M., Reddy, L. G., Bennett, R., Silva, N. D., Jr.,, Jones, L. R., and Thomas, D. D. (1999) A fluorescence energy transfer method for analyzing protein oligomeric structure: Application to phospholamban, Biophys. J. 76, 2587-2599. 60. Lovell, S. C., Worde, J. M., Richardson, J. S., and Richardson, D. C. (2000) The penultimate rotamer library, Proteins 40, 389-408. 61. Wittebort, R. J., Olejniczak, E. T., and Griffin, R. G. (1987) Analysis of deuterium nuclear magnetic resonance line shapes in anisotropic media, J. Chem. Phys. 86, 5411- 5420. 62. Lee, A. G. (2003) Lipid-protein interactions in biological membranes: a structural perspective, Biochim. Biophys. Acta 1612, 1-40. 63. Berendsen, H. J. C., and Tieleman, D. P. (1998) Lipid-protein systems: history and methods, Bichim. Biophys. Acta 3, 1639-1650.

123 64. Johnson, R. G., and Kranias, E. (1998) Cardiac sarcoplasmic reticulum function and regulation of contractility, Ann. New York Acad. Sci. 853, 31-42. 65. Asahi, M., Green, N. M., Kurzydlowski, K., Tada, M., and MacLennan, D. H. (2001) Phospholamban domain IB forms an interaction site with the loop between transmembrane helices M6 and M7 of sarco(endo)plasmic reticulum Ca2+ ATPases, Proc. Natl. Acad. Sci. 98, 10061-10066. 66. Simmerman, H. K., Kobayashi, Y. M., Autry, J. M., and Jones, D. H. (1996) A leucine zipper stabilizes the pentameric membrane domain of phospholamban and forms a coiled-coil pore structure, J. Biol. Chem. 270, 5941-5946. 67. Zamoon, J., Macioni, A., Thomas, D. D., and Veglia, G. (2003) NMR solution structure and topological orientation of monomeric phospholamban in dodecylphosphocholine micelles, Biophys. J. 85, 2589. 68. Johannsson, A., Smith, G. A., and Metcalfe, J. C. (1981) Th effects of bilayer thickness on the activity of (Na+/K+)-ATPase, Biochim. Biophys. Acta 641, 416-421. 69. Johannsson, C. A., Keightley, G. A., Smith, C. D., Richards, T. R., Hesketh, J. C., and Metcalfe, J. C. (1981) The effects of bilayer thickness on the activity of (Na+/ K+)- ATPase, J. Biol. Chem. 256, 1643-1650. 70. Starling, J. P., East, J. M., and Lee, A. G. (1993) Separate effects of long chain phosphatidylcholine on dephosphorylation of the Ca2+-ATPase and on Ca2+ binding, Biochemistry 32, 1593-1600. 71. Gould, G. W., McWhirter, J. M., East, J. M., and Lee, A. G. (1987) Uptake of Ca2+ mediated by the (Ca2+/Mg2+)-ATPase in reconstituted vesicles, Biochem. J. 245, 751- 755.

124 CHAPTER 6

Elvis K. Tiburu, Krishnan Damodaran, Ethan S. Karp, and Gary A. Lorigan, 2H and 15N Solid-State NMR Spectroscopic Studies of the Helical Tilt of the Transmembrane Segment of Phospholamban Incorporated into Magnetically Aligned and Mechanically Aligned Phospholipid Bilayers (2004), J. Am. Chem. Soc. (submitted)

125 Abstract

The effect of phospholipid acyl chain length on the orientation of the transmembrane segment of the integral membrane protein phospholamban (TM-PLB) was investigated utilizing 2H and 15N solid-state NMR spectroscopic techniques. Using anisotropic 15N chemical shifts as orientational constraints, site-specific 15N-labeled PLB was incorporated into hydrated mechanically aligned 1,2-dioleoylphosphocholine/1,2-dioleoylphosphoethanolamine (DOPC/ DOPE) phospho-lipid bilayers and magnetically aligned 1,2-dimyristoylphosphocholine/1,2- dihexanoylphospho-choline (DMPC/DHPC) phospholipid bilayers (bicelles). The chemical shifts from the mechanically aligned spectra of site-specific 15N-labeled LeuPLB in DOPC/DOPE phospholipid bilayers range from 193 ppm for Leu28 to 220 ppm for Leu42. In the DMPC/DHPC bicelles, the chemical shifts range from 121 ppm for Leu28 to 220 ppm for Leu42. Assuming PLB is a-helical in the lipid environments, the helical tilt angles in DOPC/DOPE bilayers and DMPC/DHPC bicelles were determined to be 8 ± 3° and 26 ± 6° with respect to the bilayer normal, respectively. The results obtained from the 15N NMR analysis were compared with results from aligned 2H-labeled site-specific NMR spectra of the side chain orientation of Leu TM-PLB sites in two different lipid environments. A tilt angle of 11 ± 5° was determined in mechanically aligned DOPC/DOPE phospholipid bilayers and the corresponding tilt angle in magnetically aligned DMPC/DHPC bicelles was 30 ± 4°. The effects on protein- lipid interactions are discussed in terms of hydrophobic mismatch between the hydrophobic length of the peptide and the hydrophobic thickness of the phospholipid bilayers.

126 6.0 Introduction

6.1 Incorporation of PLB into phospholipid bilayers of different chain lengths

The incorporation of an intrinsic protein into a lipid membrane is driven by both hydrophobic effects associated with creating a lower free energy of the system and matching between the hydrophobic length of the integral membrane protein and the hydrophobic thickness of the lipid bilayer (hydrophobic mismatch) (1-3). Hydrophobic mismatch affects lateral segregation of proteins in membranes, the lipid melting transition, and protein activity (4-8). One key example of this phenomena and its role in membrane organization is that a variety of membranes between the Golgi and the plasma membrane have different thicknesses and associated proteins can be routed through the secretory pathway by increasing their hydrophobic thickness via mutagenesis (9, 10). In this way, proteins can only pass from one membrane to another more closely matching their new length (9, 10). Also hydrophobic mismatch affects the way in which the stability and the topology of transmembrane helices change as a function of their hydrophobic length (9). Such effects have been examined on the molecular level using synthetic hydrophobic peptides incorporated in different chain lengths of hydrated phospholipid membranes (11-16). There are various ways in which hydrophobic mismatch between a protein and a lipid changes the three-dimensional characteristics of the system. In order to obtain maximum stability, for example, the lipid bilayer around the protein can either lengthen or shrink in size to match the hydrophobic thickness of the protein (17). Alternatively, a prime way for transmembrane a-helices to compensate for the mismatch (if the transmembrane segment is too long to match the surrounding lipid bilayer), is to tilt to reduce its effective length across the bilayer (17). A number of different methods have been employed to determine whether the tilt angle of a transmembrane protein depends upon the extent of hydrophobic mismatch (18). Solid-state NMR spectroscopy is a powerful method for studying the structure and topology of polypeptides reconstituted into lipid bilayers (19-22). Solid-state NMR investigations indicate that many hydrophobic peptides assume transmembrane alignments (23-28). In a series of site-specific 15N- labeled hf20 analogue peptides in aligned bilayers of various phospholipid chain lengths, the 15N

127 chemical shifts were found to decrease with increasing extent of mismatch, indicating tilt angles ranging from 00 for long chain PC (22:1) to an unknown value of about 350 for hf20 in 10:0 PC (23). Also, Van der Wel and coworkers utilized 2H NMR spectroscopy to determine the tilt angle of a 19-amino acid WALP peptide in different bilayer-forming lipids (29). The results from such studies are consistent with a well defined helical tilt and spectral averaging about the bilayer normal. In a study conducted by Sharpe and coworkers, a detailed 2H NMR investigation

2 employing H-labeled Ala side-chains of Lys-flanked analogues (Ac-KK(LA)8LKK-Am and Ac- 0 0 KK(LA)11LKK-Am) were determined to have tilt angles of about 7 to 17 respectively in vesicles of (16:0/18:1)PC, indicating a mismatch-dependent tilt (30). Thus, NMR methods for investigating transmembrane helix orientation in a mismatch-dependent manner can complement optical techniques such as circular dichroism and ATR-FTIR spectroscopies (31-34). For example, optical techniques provide valuable information in establishing the approximate secondary structure composition of these proteins in a fast and convenient manner. These techniques allow the spectral changes to be followed in a time-dependent manner. However, when more specific information of bilayer-associated proteins is required the choice has been to use modern solid-state NMR spectroscopy in combination with isotopic labeling (34). The activity of a number of membrane proteins has been shown to be dependent on bilayer thickness and this includes Ca2+-ATPase. Ca2+-ATPase transports Ca2+ across the sarcoplasmic reticulum (SR), leading to muscle relaxation (35). The activity of Ca2+-ATPase is highest when the ATPase is reconstituted into bilayers of di(18:1)PC, whereas the activity is lower in bilayers with shorter acyl chains such as (14:0)PC (17, 36, 37). The marked effects of bilayer thickness on the ATPase activity indicates distinct changes in the conformation of the ATPase channel(38). Although several studies have been conducted to probe the effects of bilayer thickness on the function of ATPase, they are often very complex because many steps in the reaction sequences are involved in the function (38). One of the proteins involved in the Ca2+-ATPase reaction scheme is phospholamban (PLB). PLB is believed to be a homopentameric transmembrane protein that regulates the Ca2+-ATPase (39, 40). PLB inhibits an isoform of Ca2+-ATPase, SERCA2a, in its unphosphorylated (monomeric PLB) form whereas the phosphorylated (pentameric PLB) form dissociates from SERCA2a. PLB is a 52 amino acid transmembrane protein and consists of three structural domains: residues 1-20 which comprise the hydrophilic cytoplasmic domain, residues 21-30 create an unstructured or b-sheet hinge

128 segment (33,3-41), and residues 31-52 which encompass the hydrophobic a-helical membrane- spanning region. The secondary structure of PLB in phospholipid bilayers has already been characterized and the studies have indicated that PLB is predominantly a-helical in phospholipid bilayers (39, 41, 42). Presently, there are two structural models that have been proposed through spectroscopic studies and molecular modeling techniques for pentameric PLB (33, 43). In one model, PLB is composed of two a-helices connected by a b-sheet with the cytosolic domain tilted in a range of 50-60O with respect to the bilayer normal, and the transmembrane segment of PLB is almost collinear with the bilayer normal (43). Another model has proposed a near continuous a-helix with a tilt angle of about 28O for PLB with respect to the DMPC bilayer normal (33). The evidence of such tilt angles indicates that there is some degree of mismatch and by tilting away from the bilayer normal, the PLB peptide could reduce its effective length and compensates for a positive mismatch between the bilayer thickness and the a-helical protein length. 2H NMR spectroscopic studies on the dynamics and orientation of side chain residues in site-specific 2H labeled integral membrane proteins is currently being pursued by many labs (29, 44-46). The results can be used as a complementary technique to the conventional 15N solid-state NMR data to understand phenomena such as the structure-function relationship that underlies the transport of Ca2+ through the interaction between PLB and the Ca2+-ATPase (47). Methyl group

2 motions have been well characterized by H NMR studies of CD3-labeled sites of alanines and valines in other transmembrane peptides (29, 44-46). A practical advantage of using a CD3 Leu label as an NMR probe in addition to the conventional 15N solid-state NMR spectroscopic studies is that highly intense signals are observed due to the methyl group intrinsic mobility, and the presence of three chemically equivalent deuterons on each side chain (16). 2H and 15N solid-state NMR spectroscopy studies of samples mechanically aligned between glass plates is an established method used for membrane protein studies (26-28, 48-51). Although, highly oriented mechanically aligned systems such as gramicidin and coat proteins created between glass plates have been achieved, the use of glass plates is experimentally demanding (24, 26, 52-58). The NMR probe coil volume is always occupied with a stack of glass plates instead of sample and this results in lower sensitivity. Furthermore, sample heating can be a major limitation for mechanically aligned samples due to the high RF powers involved that may destroy the sample. Alternatively, magnetically aligned phospholipid bilayers (bicelles)

129 comprising DMPC/DHPC with a q ratio 3.5 have recently been developed and used for transmembrane protein studies (56, 59). The possibility of conducting structural studies of integral membrane proteins solubilized in bicelles is important for several reasons. First of all, unlike multilamellar systems, the bicellar assemblies are monodispersed and fully hydrated and more accurately depict a cell membrane (56). The second reason is that the bicelles spontaneously align in the magnetic field and the ability to conduct both solution and solid-state NMR studies on samples which are identical except for modest differences in lipid-detergent ratios (60, 61). Specialized flat-coil NMR probes are not needed for bicelle studies. Finally, the dimensions of the bicelle can be changed by varying the detergent of the phospholipids without changing the aggregate or overall morphology of the system (56). The objective of the present study is to determine the helical tilt angle of the peptide representing the transmembrane segment of PLB (TM-PLB) using both conventional 15N solid- state NMR and 2H NMR spectroscopic techniques. In order to probe the orientational properties of TM-PLB in phospholipid bilayers using these two techniques, the peptide was first incorporated into longer chain DOPC/DOPE phospholipid bilayers mechanically aligned between glass plates, and and then, secondly, into magnetically aligned shorter chain DMPC/DHPC phospholipid bilayers. Also, the shorter chain DMPC lipids in the bicelles will be used to compare protein-lipid interaction between DOPC and DMPC lipid environments, and the effect of hydrophobic thickness on the tilt angle. This is the first demonstration of studies involving TM-PLB in bicelles (C14 phospholipid) as well as in DOPC (C18 phospholipid) phospholipid bilayers to compare hydrophobic matching of TM-PLB in the two systems using both 2H and 15N solid-state NMR spectroscopy. This paper is unique because for the first time both 2H and 15N solid-state NMR spectroscopy has been used to monitor the degree of helical alignment for both bicelles and mechanical aligned bilayers with the same peptide. The resolution and sensitivity observed in the NMR spectra of the bicelles samples were directly compared with previous work obtained from DOPC/DOPE phospholipid bilayers (62).

6.2 Materials

Fmoc-amino acids and other chemicals for peptide synthesis were purchased from

15 Applied Biosystems Inc. (Foster City, CA). Fmoc-leucine-5,5,5-d3 and Fmoc-leucine- N

130 derivatives were purchased from Isotec, Inc. (Miamisburg, OH). 2H-depleted water was purchased from Isotec. Inc. (Miamisburg, OH). 1,2-dioleoylphosphocholine (DOPC), 1,2- dioleoylphosphoethanolamine (DOPE), 1,2-dimyristoylphosphocholine (DMPC), and dihexanoylphosphocholine (DHPC) were purchased from Avanti Polar Lipids Inc. (Alabaster, AL). The phospholipids were shipped already dissolved in chloroform at a concentration of 20 mg/mL and stored at –20 OC. Hexafluoro-2-propanol (HFIP), formic acid and 2,2,2- trifluoroethanol (TFE) were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). HPLC-grade acetonitrile and isopropanol were obtained from Pharmco (Brockfield, CT) and were filtered through a 0.22-mm nylon membrane before use. Water was purified using a nanopure reverse osmosis system (Millipore, Bedford, MA). N-[2-hydroxyethyl]piperizine-N-2- ethane sulfonic acid (HEPES), trifluoroacetic acid (TFA), and EDTA were obtained from Sigma- Aldrich (St. Louis, MO).

6.3 Methods

6.3.1 Peptide synthesis Several 2H-labeled and 15N-labeled TM-PLB peptides were synthesized using predefined procedures (63). The polypeptides corresponding to the transmembrane segment of PLB Ala24- Leu52 (2H-Leu PLB and 15N-Leu PLB) were synthesized on an Applied Biosystems Inc. (ABI) 433A solid-phase peptide synthesizer controlled by a G3 Macintosh computer with SynthAssistTM 2.0 software as described previously (63).

6.3.2 Peptide purification The 2H-labeled and 15N-labeled TM-PLB peptides were purified using the following procedure. The peptides were removed from the synthesizer and cleaved from the resin. The lyophilized peptides were dissolved in HFIP/FA (4:1, at a concentration of 5 mg/mL) and centrifuged to eliminate insoluble particulates. The crude peptide was purified on an Amersham Pharmacia Biotech AKTA‚ Explorer 10S HPLC controlled by UnicornTM version 3 system software. A polymer-supported column (259 VHP82215, 5 mm, 300 Å pore size, 2.2 x 15 cm) from Grace-Vydacc was used to purify PLB. The column was equilibrated with 95% solvent

A:5% solvent B. Solvent A consisted of H2O + 0.1% TFA and solvent B was 38% MeCN +

131 57% IPA + 5% H2O (63). A 1 mL aliquot of the (5 mg/mL) peptide sample was injected into the column and the gradient was ramped from 5% to 100% of solvent B at a flow rate of 10 mL/min. Peptide elution was achieved with a linear gradient to a final solvent composition of 93% solvent B. The purified peptide fraction was lyophilized and analyzed by MALDI-TOF mass spectroscopy using a matrix of 2,5-dihydroxybenzoic acid. The overall yield of PLB was about 37% (63).

6.3.3 2H and 15N aligned solid state NMR sample preparation The DMPC/DHPC bicelle samples, consisting of 25% (w/w) phospholipid to solution with a q = 3.5 was used. The q ratio is defined as the molar ratio of DMPC to DHPC. The samples were prepared in 2 pear-shaped flasks. In one flask, DHPC and TM-PLB (dissolved in a minimal amount of TFE) were mixed together in chloroform, and in the second flask DMPC was dissolved in chloroform. The mole ratio of DMPC/DHPC/PLB was 3.5/1/0.01 respectively. The chloroform was removed with rotary evaporator and the sample was placed under high vacuum over night. The following day 300 mL of 100 mM HEPES buffer solution (pH 7.0) was added to the flask using deuterium-depleted water. The flask was then vortexed briefly, sonicated with a FS30 (Fisher Scientific) bath sonicator with the heater turned off. The sample was subjected to several freeze (77 K)/thaw cycles at room temperature until the dispersion was homogenous and all the air bubbles were removed. Typically, the total mass of the prepared bicelle sample was 300 mg. For parallel-aligned bicelle samples, a 5-mL aqueous solution of Yb3+ (100 mM) was added to the bicelle sample using deuterium-depleted water. The bicelle sample was placed in an NMR flat bottom tube (5-mm o.d. by 21-mm height) on ice via a Pasteur pipette. The mechanically aligned samples were prepared by dissolving PLB in a minimal amount of TFE and the solution was mixed with DOPC/DOPE (4:1 mole ratio) dissolved in chloroform in a pear-shaped flask. Nitrogen gas was passed over the resulting mixture to reduce the volume of the chloroform to a third of the original volume. The sample was spread onto 25 glass plates of dimensions 8.5 mm x 14 mm and the glass plates were allowed to dry in a dessicator overnight. The peptide to lipid mole ratio was 1:100. Deuterium-depleted water was added onto the lipid-peptide mixture and the glass plates were stacked on top of each other. The stacked glass plates were then placed in a humidity chamber of ammonium monophosphate at a relative humidity of about 93% at 42 OC for 24 h.

132 6.3.4 2H solid-state NMR spectroscopy The magnetically aligned phospholipid bilayers (bicelles) solid-state NMR spectra were acquired on a Bruker Avance 500 WB solid-state NMR spectrometer operating at a resonance frequency of 76.8 MHz for 2H. A double-resonance round coil bruker solid-state NMR probe equipped with a 5-mm solenoid coil was used for the experiments. For the mechanically aligned samples, a double resonance flat coil solid-state NMR probe of dimension 8.5 mm x 14 mm was used operating at 25 OC and 40 OC. The quadrupolar echo pulse sequence was used with quadrature detection capabilities and complete phase cycling of the pulse pairs (64). A 3.0-ms 900 pulse, a sweep width of 100 kHz, a recycle delay time of 400 ms, and a 30-ms interpulse delay were used to accumulate 150,000 transients. Prior to Fourier transformation, an exponential multiplication of 200-Hz line broadening was performed on the spectra. The bicelle spectra were acquired at 40 OC. Nearly identical 2H NMR spectra were obtained at both 25 OC and 40 OC for the mechanically aligned TM-PLB samples.

6.3.5 15N solid-sate NMR spectroscopy For the magnetically aligned samples, a 5-mm round coil probe was used operating at a resonance frequency of 500 MHz for 1H and at 50.7 MHz for 15N. The following pulse sequence parameters were used: 4.5-ms 1H 90o pulse, 1.0-ms contact time, 600-ppm sweep width, and a 4-s recycle delay with 1H decoupling. 24k scans were averaged and 300-Hz line broadening was used to process the data. The bicelles samples were collected at 40 OC. The mechanically aligned 15N spectra were collected using a Bruker double resonance flat-coil solid-state NMR probe operating at 25 OC and 40 OC. The following pulse sequence parameters were used: 4.7-ms 1H 900, 500-ppm sweep width, 1.5-ms contact time and a 4-s recycle delay with 1H decoupling. Nearly identical 15N NMR spectra were obtained at both 25 OC and 40 OC for the mechanically aligned TM-PLB samples.

6.3.6 2H tilt angle calculations The helical tilt angles for both 2H and 15N NMR spectra were calculated using equations previously derived in the literature (16, 65-67) The observed quadrupolar splitting for a deuterated amino acid side chain in a peptide incorporated into aligned spectra (Dnq) can be expressed as (16, 65-67): e|| is the angle between the peptide helical axis and the side chain bond

133 vector. t is the tilt angle defined as the angle between the peptide helix axis and the bilayer normal (assume to be along the magnetic field direction) (16). d is the rotation angle and it depends on the three angles shown below (16).

2 2 Dnq = (3/4)K(3cos e|| (cos t - sin t cos d tan e||) - 1) (1)

Where Dnq is the quadrupolar splitting, K is the quadrupolar coupling constant, and S is the order parameter. K is defined as:

K = (e2qQ/h)S (2) e2qQ is the quadrupole interaction and, h is Planck’s constant.

d = r + e^ + J (3)

where r is the rotation of the helix when compared to a standard orientation. e^ is the angle of the bond vector with respect to a vector of the side chain to the peptide axis, and J is the angle between a reference residue and a labeled residue given by (n-1)*100 where n is the number of residues from the reference residue. The tilt angle of the protein in DOPC bilayers was calculated using a program written in C++. The parameter r was varied from 0 to 360° in increments of 0.1° and t was varied from 0 to

50° in increments of 0.1°. Values for e|| and e^ of each leucine residue were gathered assuming that the C-D bond vectors were averaged about the Cg-Cd bond, using the molecular modeling software Mol-Mol and phospholamban from the protein data bank (model 1FJK).(68) Leu28 was chosen as the arbitrary reference amino acid with r= 0°. The root mean squared deviation

0 (RMSD) was used to rank the possible values of r and t. e|| =70 , d =222° for a perfect a-helix, and e^ = 32°. For fast methyl group rotation, the quadrupolar coupling constant, was reduced by over third ((44, 53). Thus a quadrupolar coupling constant of 19 kHz was used for the calculation using the equation derived by Strandberg and coworkers (16).

134 6.3.7 15N tilt angle calculations For a regular helix such as TM-PLB with 3.6 residues per turn in an oriented sample, the z axis of the director frame is parallel to the magnetic field B0. The chemical shift at residue k is given by the following algorithm (58):

f(s11,k, s22,k, s22,k, r, t, k) = B(r, t, k-1)RsR’ B(r, t, k-1)’ (4) where the prime denotes the matrix transpose. The position of the helix axis relative to B0 (equivalent to the bilayer normal) is determined by the polar coordinates r and t. Where t is the tilt angle between the helix axis and the bilayer normal, and r is the helix axis rotation angle. s is the principal axis frame and is represented by:

(5) È 0 0 ˘ Ís 11,k ˙ s = 0 0 Í s 22,k ˙ Í 0 0 ˙ Î s 33,k˚

R is the matrix that transforms the coordinates from the principal axis frame, PAF(k) to the helical axis frame,† HAF(k-1) which is computed as,

È- 0.83 0.56 - 0.04˘ R Í 0.55 0.80 0.22˙ (6) = Í - ˙ ÎÍ- 0.09 - 0.21 - 0.97˚˙

B(r, t, k) is the vector that gives the coordinates of B0 in HAF(k) and has the following form:

B(r, t, k) = sint cos(r-(k-26)100°) sint sin(r-(k-26)100°) cost (7)

Using the two sets of 15N chemical shift values for DOPC and DMPC in Table 6.1, we determined the best fit for the orientation of the PLB peptide helix relative to the bilayer normal.

The magnetic field axis Bo is rotated over all r and t space and the chemical shifts were calculated for each experimentally characterized 15N site. The root-mean squared deviation (RMSD) between the observed and calculated chemical shifts is given by the following expression:

135 N (8) 1 calc Expt 2 RMSD = Â(s i - s i ) N i=1

6.4 Results

TM-PLB peptide (NH2-Ala-Arg-Gln-Asn-Leu28-Gln-Asn-Leu-Phe-Ile-Asn-Phe-Cys-Leu-

Ile-Leu39-Ile-Cys-Leu42-Leu-Leu-Ile-Cys-Ile-Ile-Val-Met-Leu51-Leu-COOH) was synthesized with selective 2H- and 15N-labels at positions Leu28, Leu39, Leu42 and Leu51. In the current investigation a total of six 2H and 15N specifically labeled peptides were incorporated into fully hydrated DMPC/DHPC magnetically aligned bicelles as well as in DOPC/DOPE phospholipid bilayers mechanically aligned on glass plates. The results of our 2H NMR studies of Leu site- specific 2H-labeled TM-PLB in DMPC/DHPC bicelles are shown in Figure 6.1. The DMPC/DHPC bicelle samples containing one of the 2H-labeled Leu28, Leu39 and Leu51 TM- PLB peptides were aligned in two different orientations with respect to the static magnetic field. The direction of alignment is dependent on the magnetic susceptibility anisotropy tensor (Dc) of the bicelle system. The bicelle will spontaneously align such that the bilayer normal is perpendicular to the magnetic field director because Dc < 0 (56, 65, 66, 69, 70). In the presence of lanthanide ions such as Yb3+(Dc > 0), the phospholipid bilayers are aligned such that the bilayer normal is parallel to the magnetic field director resulting in an increased spectral resolution. The 2H quadrupolar splittings of the perpendicular aligned spectra of DMPC/DHPC bicelle of Leu39 (Figure 6.1(C)) and Leu51 (Figure 6.1(E)) were 0.84 kHz and 0.92 kHz respectively. In the parallel aligned bicelle spectra, the quadrupolar splitting of Leu28 was about 0.52 kHz (Fig. 6.1(B)), the quadrupolar splittings of Leu39 (Figure 6.1 D)) was 1.72 kHz, and that of Leu51 (Figure 6.1 (F)) was 2.10 kHz. Thus, the quadrupolar splitting of the parallel aligned spectra were slightly more than double their perpendicular aligned quadrupolar splittings. The decreased quadrupolar splittings of Leu28 especially in the perpendicular alignment where the splitting disappears into the residual water isotropic peak is most likely due to an increased mobility of the Leu28 residue. The smaller quadrupolar splittings were also observed in randomly dispersed 2H-labeled Leu samples of unoriented TM-PLB incorporated into POPC bilayers (62). A phenomenon which was associated with the proximity of Leu28 being close to

136 the N-terminus where peptide mobility is faster (62). The resolved quadrupolar splittings clearly indicate alignment when compared to powder spectra (62). The error associated with the measurement of the quadrupolar splitting in each data point was estimated to be ± 0.1 kHz by measuring the variance in three different trials. To calculate the DOPC tilt angle using the deuterium quadrupolar splittings, equations derived by Strandberg and coworkers were used (16). The average calculated tilt angle of TM-PLB with respect to the bilayer normal based upon the quadrupolar splittings of all the labeled sites was 30 ± 4˚ in the DMPC/DHPC bicelle samples. Additionally, the helical tilt of TM-PLB in mechanically aligned DOPC/DOPE phospholipid bilayers was next investigated with 2H solid-state NMR spectroscopy. The choice of DOPC/DOPE for these studies is three-fold. First, the hydrophobic thickness of DOPC is 4 Å longer than DMPC (71) and will serve as a good model to compare the extent of hydrophobic matching of the TM-PLB peptide with the hydrophobic thickness of the lipid. Secondly, site- specific 15N NMR studies of aligned TM-PLB incorporated into DOPC/DOPE phospholipid bilayers has already been studied and will serve as a useful comparison with the bicelle data (62). Finally, DOPC/DOPE phospholipid bilayers have been shown to be an excellent model phospholipid system for reconstituting PLB (50). The results of the 2H NMR study of TM-PLB in DOPC/DOPE phospholipid bilayers on mechanically aligned glass plates are displayed in Figure 6.2. The quadrupolar splittings were 0.9, 2.2 and 2.5 kHz for Leu28, Leu39 and Leu51, respectively, as shown in Figure 6.2. The error associated with the measurement of the quadrupolar splitting in each data point was estimated to be ± 0.2 kHz determined from the spread of values in multiple trials. The 2H quadrupolar splittings of all the labeled sites in DOPC/DOPE phospholipid bilayers were slightly larger than those obtained from the parallel aligned DMPC/DHPC/Yb3+ bicelle samples. Because DMPC/DHPC/Yb3+ bicelles and DOPC/DOPE phospholipid bilayers are aligned in the same direction (parallel to the magnetic field), we should expect them to have similar quadrupolar splittings in similar chemical environments. However, the difference in splitting of the labeled sites in the two different phospholipid bilayers may be for several reasons; differences in temperature, differences in protein-lipid interaction, a decrease in wobbling of the TM-PLB peptide in DOPC phospholipid bilayers, a change in the tilt of the TM-PLB peptide, or a combination of all the different effects. Another striking difference between the spectra from the bicelle 2H NMR samples and that from the DOPC phospholipid bilayers is that the spectra from the later are broad. This could be due to

137 the level of hydration that is 30 % in the DOPC/DOPE samples when compared to the bicelle sample with 75 % hydration. Also, additional wobbling of the bicelle may yield sharper 2H linewidths.

Leu28 (n ^ Bo) (A)

(B) Leu28 (n || Bo)

Leu39 (n ^ Bo) (C)

Leu39 (n || Bo) (D)

Leu51 (n ^ Bo) (E)

(F) Leu51 (n || Bo)

8 6 4 2 0 -2 -4 -6 -8

2H (kHz)

Figure 6.1 2H NMR spectra of 2H-labeled site-specific 25% (w/w) q= 3.5 DMPC/DHPC/TM-PLB bicelle samples investigated with and without the

138 paramagnetic lanthanide ion Yb3+. (A) 2H-labeled Leu28PLB bicelle spectrum contains 0% Yb3+. The bilayer normal is perpendicular to the static magnetic field. (B) 2H-labeled Leu28PLB bicelle spectrum contains 5% molar Yb3+ (when compared to DMPC). The bilayer normal is parallel to the direction of the static magnetic field. (C) 2H-labeled Leu39PLB bicelle same as (A) and, (D) 2H-labeled Leu39PLB same as (B). (E) 2H-labeled Leu51PLB bicelle same as (A), and (F) 2H-labeled Leu51PLB bicelle same as and (B).

(A) Leu28

(B) Leu39

(C) Leu51

8 6 4 2 0 -2 -4 -6 -8

2 H (kHz)

Figure 6.2 2H NMR spectra of 2H-labeled site-specific DOPC/DOPE/TM-PLB mechanically aligned phospholipid bilayer samples. The spectra were collected at 25 OC. (A) 2H-labeled Leu28, (B) 2H-labeled Leu39 and (C) 2H-labeled Leu51.

First of all, temperature did not have any major effect on the 2H quadrupolar splittings shown in Fig 6.2. The same splittings were obtained at both 25 OC and 40 OC for the mechanically aligned samples. In order to compensate for the slight wobbling of the bicelle, an order parameter, S of 0.84 (calculated from the powder spectra) was used in the bicelle tilt angle

139 calculation of 30±4°(56). In the mechanically aligned DOPC/DOPE phospholipid bilayers TM- PLB yielded a tilt angle of 11±5°(56). Most of the error associated with the tilt angle is due to the inability to perfectly determine the S values for each labeled site. Estimates of its value were taken from randomly oriented 2H labeled PLB incorporated in POPC bilayers that have been previously studied.(62) K values from which S can be found were estimated for Leu28, Leu39, and Leu51 to be 8 kHz, 17 kHz, and 20 kHz respectively (62). To complement the results obtained from the 2H NMR studies, anisotropic 15N chemical shifts were observed for each of these oriented samples in magnetically aligned DMPC/DHPC/Yb3+ and mechanically aligned DOPC/DOPE phospholipid bilayers (Fig. 6.3). In oriented systems, an amide N-H bond approximately parallel to the direction of the applied magnetic field and the membrane normal has an 15N resonance frequency near that of the principal tensor element s33. Conversely, an amide N-H bond perpendicular to the magnetic field and the membrane normal has an 15N resonance frequency near that of the principal tensor

15 element s11 and s22 of the chemical shift powder pattern. Therefore, the N solid-state NMR spectra of site-specific 15N-labeled TM-PLB peptides reconstituted into oriented phospholipid bilayers will also provide a direct measurement of the helical tilt of the TM-PLB peptide with respect to the bilayer normal. Comparison of the 15N NMR TM-PLB spectra in Figure 6.3 clearly shows the effect of sample orientation when compared to the powder pattern spectra (Figure 6.3(G)). The corresponding 15N chemical shifts for 15N-labeled residues in DMPC/DHPC/Yb3+ bicelles range from 121 ppm for Leu28 (Figure 6.3(A)) to 220 ppm for Leu42 (Figure 6.3(E)), and from 193 ppm for Leu28 (Figure 6.3(B)) to 220 ppm for Leu42 (Figure 6.3(F)) in the hydrated DOPC/DOPE glass plate samples (see Table 6.1). The chemical shifts were all

15 referenced to a saturated solution of ( NH)SO4 (27 ppm). A chemical shift value of 220 ppm is indicative of a transmembrane a-helix. As shown in Figure 6.3 (A), in DMPC/DHPC phospholipid bicelles, TM-PLB (labeled at position Leu28) has a chemical shift at about 120 ppm which is close to an isotropic value. However, in Figure 6.3 (B), 15N-labeled Leu28 TM- PLB yields a chemical shift of about 190 ppm when reconstituted into oriented DOPC/DOPE phospholipid bilayers. This value is much closer to s33, a transmembrane orientation for TM- PLB. In both lipid environments Leu39 and Leu42 exhibited a transmembrane orientation as shown in Figures 6.3(C) through (F). The NMR spectra in Figure 6.3 also showed single line resonances with linewidths ranging from 6-15 ppm which supports the presence of a single

140 structural conformation of an aligned PLB peptide in a hydrated phospholipid bilayer (58). The signal-to-noise ratio is , however, poor due to resonances from stable structures. Previous studies have shown that increased tilt angles of peptides are often associated with decreased chemical shifts in aligned spectra (2, 23). In this study, going from the long-chain DOPC to the short-chain DMPC resulted in a decrease in chemical shifts suggesting an increase in the corresponding tilt angle. The static principal values of the tensor elements, s11, s22 and s33 were determined previously by simulating the 15N chemical shift powder pattern spectra of the unoriented samples of DOPC and DMPC for each labeled site (Table 6.1) (62). The aligned 15N chemical shift measurements as well as the chemical shift tensor values from the unoriented samples were used to accurately determine the helical tilt angle in the DOPC and DMPC bilayer systems. The method established by Kovacs and coworkers was employed for these calculations as outlined thoroughly in the introduction (58). Figure 6.4(A) shows the contour plots of the 15N angle calculations and show a unique minimum on the RMSD surface corresponding to t = 26 ± 2° and r = -250° in DMPC/DHPC bicelle samples. Additionally, the RMSD surface corresponding to the DOPC/DOPE phospholipid bilayer samples gave a unique minimum at a tilt angle, t = 8 ± 3° and peptide rotation, r = 50° asshown in (Figure 6.4 (B)). These values correspond to the most likely orientations of PLB in DMPC and DOPC phospholipid bilayers. As shown in Figure 6.4, there is only one minimum identified to predict the orientation of TM-PLB. Thus, if we assume the a- helical tilt angle of the TM-PLB peptide reflects the hydrophobic thickness of the phospholipid bilayer and the length of the peptide, then the tilt angle stated in this paper is reasonable. Clearly with such a large value for the associated error, we can conclude that the accurate determination of helix tilt is far from perfect. It is however certain from this studies that the helix tilt is very sensitive to the chemical shifts, and that we can determine the tilt angle with some degree of accuracy.

6.5 Discussion The unique 2H and 15N solid-state NMR spectroscopic data obtained from site-specific 2H- and 15N-labeled TM-PLB in oriented phospholipid bilayers allow us to compare for the first time the orientation of TM-PLB in two different model membrane systems. Striking differences were observed in the 2H NMR data form the DMPC phospholipid bicelles and the corresponding

141 DOPC phospholipid bilayers. The resonance peaks from the later were broad most likely because the level

(A) Leu28 bicelles

Leu28 glass plates (B)

Leu39 bicelles (C)

Leu39 glass plates (D)

Leu42 bicelles (E)

(F) Leu42 glass plates

s^

(G) s|| Leu42 powder

300 200 100 0 15 N (ppm)

142 Figure 6.3 One-dimensional solid-state 15N NMR spectra of site-specific 15N- labeled TM-PLB in oriented DMPC/DHPC phospholipid bilayers and DOPC/DOPE phospholipid bilayers. The phospholipid bicelles correspond to a q ratio of 3.5. The spectra displayed were collected at 40 OC for the bicelles and at 25 OC for the DOPC/DOPE phospholipid bilayers. (A) 15N-labeled Leu28 in DMPC/DHPC bicelles and (B) 15N-labeled Leu28 in DOPC/DOPE bilayers. (C) 15N-labeled Leu39 same as (A) and (D) 15N-labeled Leu39 same as (B). (E) 15N- labeled Leu42 same as (A) and (F) 15N-labeled Leu42 same as (B). The peptide/lipid mole ratio was 1:100. of hydration was much less (about 30%) in DOPC bilayers when compared to the fully hydrated DMPC phospholipid bicelles (about 75%). Theoretically, the peptide backbone could adapt to a positive mismatch by decreasing its pitch and forming for example a π-helix (72). However, this

Table 6.1 15N Chemical shift data from the single site 15N-labeled Leu-PLB

15 aligned Samples [in ppm relative to ( NH4)2SO4 solution referenced to 27 ppm].

Sites 15N(ppm) CSA (ppm) In DMPC or DOPC

DMPC/DHPC DOPC/DOPE s11 s22 s33

Leu28PLB 122 193 55 82 225

Leu39PLB 179 215 50 80 229

Leu42PLB 223 220 52 84 230

143 does not seem to be a favorable response because TM-PLB shows a stable a-helical conformation as evidenced from the 15N chemical shifts. Thus, we did not expect the backbone of TM-PLB to significantly change upon incorporation into the different phospholipid environments. It is also interesting to note that the significant change in the helical tilt orientation of TM-PLB upon reconstitution into the shorter-chain DMPC does not suggest bilayer fragmentation induced by the peptide. The presence of such sharp resonance peaks from the 2H DMPC/DHPC and 15N chemical shifts are evidence of intact bilayers. Because we have ruled out straining as a possible consequence of the change in orientation of PLB in the DMPC bilayer, the only other possibility is the hydrophobic thickness of the DMPC lipid bilayer. The bilayer thickness from carbonyl-to-carbonyl of DMPC and DOPC is about 23 Å and 27 Å respectively (73, 74). Theoretically, since DOPC is approximately 4 Å longer than DMPC, we should expect a change in the helix tilt of TM-PLB of about 20° assuming the helix tilt reflects the hydrophobic thickness and the length of the peptide (11, 58, 74). Both the 2H and 15N NMR results clearly demonstrate that this is the case in the two phospholipids bilayer environments.

40

30 (A) t 20

10

0 -100 -200 r

40

30 (B)

t 20

10

0 0 -100 -200 r 144 Figure 6.4 Contour plots representing the 15N chemical shifts calculated for the rotation, r from 00 to 3600 and for tilt angle, t from 00 to 900. The RMSD between the experimental and calculated chemical shifts was calculated for each r and t pair and plotted. In (A), the plot has a minimum at r = -2500 and t = 260 and corresponds to the DMPC/DHPC bicelle sample. In (B), the plot has a minimum at r = 500 and t = 80 corresponding to DOPC/DOPE phospholipid bilayers.

While the objective of this study is concerned with the helical tilt of TM-PLB in phospholipid bilayers, the data cannot be used to explain the extent of oligomerization of the TM-PLB peptide in the lipid environments of different chain lengths. However, evidence in the literature indicates that a peptide/lipid mole ratio of as low as 1:100 will result in the formation of pentameric PLB (62, 75, 76). Thus, there is indeed an association of the monomers of TM- PLB within the lipid bilayers at the concentration used for this study (1:100). Additional, evidence indicates that oligmerization can result in the overall shortening of a peptide which could inevitably facilitate matching of the peptide to the bilayer thickness. Based upon the present data shrinking did not change the overall tilt of TM-PLB in DMPC/DHPC phospholipid bilayers. Secondly, the pentameric TM-PLB peptide could not cause the formation of non-bilayer structures which would have resulted in broad linewidths of the 15N spectra as noted previously by Kovacs and coworkers for the Influenza A Virus (58). This is clear evidence that pentameric PLB did not shrink to the extent that it could fit into the DMPC hydrophobic environment, hence the reason for such a large tilt angle in the DMPC phospholipid bicelles. The present solid-state NMR spectroscopic data can be used to propose a structural model for TM-PLB as displayed in Figure 6.5. The model indicates that the lipid/peptide organization is similar to that of lamellar lipid/peptide complexes. In such a model, the effective hydrophobic length is the projection of the peptide length in the membrane, Leff =Lpcos t, where

Leff is the peptide length, t is the tilt angle, and Lp is the hydrophobic thickness of the lipid.(16) Using the calculated tilt angle of 8° for the TM-PLB peptide as well as the hydrophobic

145 thickness of the DOPC/DOPE phospholipid bilayer, which is estimated to be about 27 Å (71), the length of the TM-PLB peptide that is required to span the membrane is about 27 Å. Assuming the number of amino acids per turn in an a-helix is 3.6 residues, and the pitch is 5.4 Å, then the translation per residue along the a-helical peptide will be 1.5 Å. Thus, to be able to span the hydrophobic region of the DOPC/DOPE hydrophobic region, 18 amino acid residues are required, indicating a perfect match between the TM-PLB peptide and the DOPC/DOPE bilayers. Using the same concept for DMPC, with a DMPC/DHPC bicelle of thickness 23 Å and a length of TM-PLB of about 27 Å, will give an angle corresponding to 32°. For 18 amino acid residues to be buried within the hydrophobic core of DMPC/DHPC, the PLB peptide must tilt at an angle of 32° to compensate for the mismatch. The experimental tilt angle of about 26° from the 15N NMR and 30° from the 2H NMR within the limits of experimental error is similar to the estimated value of 32°. The helical tilt angle of PLB is currently under debate. Ludlam and coworkers used Fourier transform infrared spectroscopy and site-directed isotopic labeling to probe the local structure of the transmembrane segment of PLB in DMPC phospholipid bilayers. Their results showed an axial orientation of the a-helical PLB that is approximately 28° relative to the membrane normal of DMPC bilayers (77). Thus, the helical tilt angle of 26° obtained by solid- state NMR 15N spectroscopy and 30° by 2H spectroscopy in the present study are very similar to their results. Additionally, Mascioni and coworkers have conducted solid-state 15N NMR studies of the mutant full length AFA-PLB (where A36, F41, and A46 have replaced the three corresponding transmembrane cysteine residues) in DOPC/DOPE phospholipid bilayers (50). Although, this mutant forms monomeric PLB, its orientation with respect to the bilayer normal is very close to the present study’s 11° determined with 2H solid-state NMR and 8° determined with 15N solid-state NMR. The conventional 15N solid-state NMR spectra indicate that the PLB peptide is transmembrane and that the long molecular axis is oriented at two different tilt angles with respect to the bilayer normal in the two different lipid environments. That position Leu39 near the very center of the hydrophobic core of both DMPC/DHPC bicelles and DOPC/DOPE phospholipid bilayers. This agrees well with our 15N NMR data for both Leu39 and Leu42. The 15N chemical shift values corresponding to Leu39 and Leu42 indicate that the residues are transmembrane, and are buried within the hydrophobic core of the lipid. The only unique 15N

146 chemical shift in the two lipid environments is associated with Leu28. In the DMPC/DHPC bicelle sample, Leu28 has an 15N isotropic chemical shift of 120 ppm, indicating it is not transmembrane. However, in longer chain DOPC/DOPE phospholipid bilayers, Leu28 has a 15N chemical shift value of 193 indicating Leu28 is transmembrane. Theoretically, in DMPC/DHPC bicelles, Leu28 and Leu51 will be located in the lipid/water interface for the TM-PLB peptide that is oriented such that the peptide plane is collinear with the normal of the bilayer. However, with such a large tilt (26°) from the NMR data of TM-PLB incorporated into DMPC/DHPC bicelles, as well as the large 15N chemical shift corresponding to Leu51 (220 ppm), it is certain that Leu51 is transmembrane and buried within the hydrophobic core of the lipids. The 15N chemical shift of Leu28 (120 ppm) indicates that it is a surface residue. Previous studies on 15N backbone dynamics have indicated slight backbone motion for 15N labeled Leu28 when compared to Leu39 and Leu51 in POPC bilayers (62). As depicted in Figure 6.5(A), a structural model describes the orientation of TM-PLB in a DMPC bilayer. In this model, TM-PLB is tilted such that Leu28 is still exposed to the lipid/water interphase, whereas Leu39 and Leu42 are buried within the hydrophobic core. This is consistent with the 15N NMR and 2H NMR spectroscopic data. In DOPC/DOPE phospholipid bilayers, at a TM-PLB tilt angle of 8°, Leu28 is now transmembrane together with Leu39 and Leu42 as shown in Figure 6.5(B). Thus, clearly hydrophobic thickness is extremely important in peptide organization and function in model membranes. For instance if PLB is to be reconstituted in model membranes together with Ca2+- ATPase, the membrane thickness will determine the level of protein/lipid interaction. While this study is focusing on the type of mismatch in the two lipid environments, it will be interesting to test the activity of the Ca2+-ATPase under these lipid environments to ascertain the effect of phospholipid chain length on activity. Thus, clearly hydrophobic thickness is extremely important in peptide organization and function in model membranes. Previous studies have shown that the activity of the pump is dependent on the nature of the phospholipids surrounding the pump as well as on the thickness of the hydrophobic region of the phospholipid bilayer (17). In conclusion, we have been able to incorporate TM-PLB into DMPC/DHPC bicelles and DOPC/DOPE phospholipid bilayers to compare the extent of peptide tilting. The bicelle is highly hydrated and mimics a membrane environment, and therefore, could serve as a complementary technique for protein structural determination. The high degree of tilting in DMPC bicelles is a demonstration that the peptide representing the transmembrane segment of PLB exceeds the

147 hydrophobic thickness of DMPC. To compensate for the mismatch, the TM-PLB peptide is tilted in order to reduce the effective length for the positive hydrophobic match. However, the tilt was not sufficient in bringing Leu28 within the hydrophobic core of DMCP/DHPC bicelles as deduced from the 2H and 15N chemical shift data. Conversely, TM-PLB in DOPC/DOPE bilayers indicated that all the labeled residues are transmembrane indicating a perfect match within the hydrophobic core of DOPC/DOPE bilayers. We have also demonstrated that CD3 can serve as alternative probe to 15N for peptide tilt angle determination using solid state NMR due to its inherent sensitivity.

z r t

(A)

z r t (B)

Figure 6.5 Structural models describing the position of the specific residues and the orientation of the peptide representing TM-PLB in (A) DMPC/DHPC bicelles, and (B) DOPC/DOPE phospholipid bilayers. In each case t represents the tilt

148 angle, r is the rotation of the peptide along the long molecular axis, and z is the normal of the lipid bilayer.

149 6.6 References

1. Mouritsen, O. G., and Bloom, M. (1984) Mattress model of lipid-protein interactions in membranes, Biophys. J. 46, 141-153. 2. Morein, S., Koeppe, R. E., II,, Lindblom, G., Kruijff, B., and Killian, J. A. (2000) The effects of peptide/lipid hydrophobic mismatch on the phase behavior of model membranes mimicking the lipid composition in Escherichia coli membranes, Biophys. J. 78, 2475- 2485. 3. Hildebrand, P. W., Preissner, R., and Frommel, C. (2004) Structure features of transmembrane helices, FEBS Lett. 559, 145-151. 4. Froud, R. J., Earl, C. R. A., East, J. M., and Lee, A. G. (1986) Effects of lipid fatty acyl chain structure on the activity of the (Ca2+/Mg2+)-ATPase, Biochim. Biophys. Acta 860, 354-360. 5. Johannsson, A., Smith, G. A., and Metcalfe, J. C. (1981) The effects of bilayer thickness on the activity of (Na+/K+)-ATPase, Biochim. Biophys. Acta 641, 416-421. 6. Marsh, D. (1995) Lipid-protein interactions and heterogenous lipid distribution in membrane, Mol. Membr. Biol. 12, 59-64. 7. Piknova, B., Perochon, E., and Tocanne, J. F. (1993) Hydrophobic mismatch and long- range protein-lipid interactions in bacteriorhodopsin phosphatidylcholine vesicles, Eur. J. Biochem. 218, 385-396. 8. Lehtonen, J. Y. A., and Kinnunen, P. K. J. (1997) Evidence for phospholipid microdomain formation in liquid crystalline liposomes reconstituted with Escherichia coli lactose permease, Biophys. J. 72, 1247-1257. 9. Duque, D., Li, X., Katsov, K., and Schick, M. (2002) Molecular theory of hydrophobic mismatch between lipids and peptides, J. Chem. Phys. 116, 10478-10484. 10. Pelham, H. R., and Munro, S. (1993) Sorting of membrane proteins in the secretory pathway, Cell 75, 603-605. 11. de Planque, M. R. R., Greathouse, D. V., Koeppe, R. E., II,, Schaffer, H., Marsh, D., and Killian, J. A. (1998) Influence of lipid/peptide hydrophobic mismatch on the thickness of diacylphosphatidylcholine bilayers. A H-2 NMR and ESR study using designed transmembrane alpha-helical peptides and gramicidin A, Biochemistry 37, 9333-9345.

150 12. de Planque, M. R. R., Kruijtzer, J. A. W., Liskamp, R. M. J., Marsh, D., Greathouse, D. V., Koeppe, R. E., II,, de kruijff, B., and Killian, J. A. (1999) Different membrane anchoring positions of tryptophan and lysine in synthetic transmembrane alpha-helical peptides, J. Biol. Chem. 274, 20839-20846. 13. Morrow, M. R., Huschilt, J. C., and Davis, J. H. (1985) Simultaneous modelling of phase and calorimetric behavior in an amphiphilic peptide/phospholipid model membranes, Biochemistry 24, 5396-5406. 14. Zhang, Y.-P., Lewis, R. N. A. H., Hodges, R. S., and McElhaney, R. N. (1995) Peptide models of helical hydrophobic transmembrane segments of membrane-proteins .2. differential scanning calorimetric and FTIR spectroscopic studies of the interaction of AC-K-2-(LA)(12)-K-2-amide with phosphatidylcholine bilayers, Biochemistry 34, 2362- 2371. 15. Ying, W., Irvine, S. E., Beekman, R. A., Siminovitch, D. J., and Smith, S. O. (2000) Deuterium NMR reveals helix packing interactions in phospholamban, J. Am. Chem. Soc. 122, 11125-11128. 16. Strandberg, E., Ozdirekcan, S., Rijkers, D. T. S., van der Wel, P. C. A., Koeppe II, R. E., Liskamp, R. M. J., and Killian, J. A. (2004) Tilt angles of transmembrane model peptides in oriented and non-oriented lipid bilayers as determined by 2H solid-state NMR, Biophys. J. 86, 37093721. 17. Lee, A. G. (2003) Lipid-protein interactions in biological membranes: a structural perspective, Biochim. Biophys. Acta 1612, 1-40. 18. de Planque, M. R. R., and Killian, J. A. (2003) Protein-lipid interactions studied with designed transmembrane peptides: role of hydrophobic matching and interfacial anchoring, Mol. Membr. Biol. 20, 271-284. 19. McDowell, L. M., and Schaefer, J. (1996) High-resolution NMR of biological solids, Curr. Opin. Struct. Biol. 6, 624-629. 20. Nicholson, L. K., and Cross, T. A. (1989) Gramicidin cation channel: An experimental determination of the right-handed helix sense and verification of beta-type hydrogen bonding, Biochemistry. 28, 9379-9385.

151 21. Griffin, R. G. (1998) A variety of homo- and heteronuclear transfer sequences have emerged recently which operate by manipulating dipolar interactions, Nat. Struct. Biol., 508-512. 22. Bechinger, B. (1999) The structure, dynamics and orientation of antimicrobial peptides in membranes by solid-state NMR spectroscopy, Biochim. Biophys. Acta 1462, 157-183. 23. Harzer, U., and Bechinger, B. (2000) Alignment of lysine-anchored membrane peptide under conditions of hydrophobic mismatch: A CD, 15N and 31P solid-state NMR spectroscopy investigation, Biochemistry 39, 13106-13114. 24. Kovacs, F. A., and Cross, T. A. (1997) Transmembrane four-helix bundle of influenza A M2 protein channel: structural implications from helix tilt and orientation, Biophys. J. 73, 2511-2517. 25. Lazo, N. D., Hu, W., and Cross, T. A. (1995) Low-temperature solid-state 15N NMR characterization of polypeptide librations, J. Magn. Reson. 107, 43-51. 26. Marassi, F. M., and Opella, S. J. (2003) Simultaneous assignment and structure determination of a membrane protein from NMR orientational restraints, Protein Sci. 12, 403-411. 27. Montal, M., and Opella, S. J. (2002) The structure of the M2 channel-lining segment from the nicotinic acetylcholine receptor, Biochim. Biophys. Acta 1565, 287-293. 28. Nevzorov, A. A., Mesleh, M. F., and Opella, S. J. (2004) Structure determination of aligned samples of membrane proteins by NMR spectroscopy, Magn. Reson. Chem. 42, 162-171. 29. Van der Wel, P. C. A., Strandberg, E., Killian, J. A., and Koeppe II, R. E. (2002) Geometry and intrinsic tilt of a tryptophan-anchored transmembrane alpha-helix determined by deuterium NMR, Biophys. J. 83, 1479-1488. 30. Sharpe, S., Barber, K. R., Grant, C. W. M., Goodyear, D., and Morrow, M. R. (2002) Organization of model helical peptides in lipid bilayers: insight into the behavior of single-span protein transmembrane domains, Biophys. J. 83, 345-358. 31. Strandberg, E., Sparrman, T., and Lindblom, G. (201) Phase diagrams of systems with cationic a-helical membrane-spanning model peptides and dioleoylphosphatidylcholine, 89-90, 239-261., Adv. Coll. Interface Sci. 89-90, 239-261.

152 32. Arkin, I. T., Adams, P. D., MacKenzie, K. R., Lemmon, M. A., Brunger, A. T., and Engelman, D. M. (1994) Structural organization of the pentameric transmembrane alpha helix of phospholamban, a cardiac ion channel, EMBO. J. 13, 4757-4764. 33. Arkin, I. T., Rothman, M., Ludlam, C. F., Aimoto, S., Engelman, D. M., Rothschild, K. J., and Smith, S. O. (1995) Structural model of the phospholamban ion channel complex in phospholipid membranes, J. Mol. Biol. 248, 824-834. 34. Harzer, U., and Bechinger, B. (2000) Alignment of lysine-anchored membrane peptide under conditions of hydrophobic mismatch: A CD, 15N and 31P solid-state NMR spectroscopy investigation Biochemistry. 39, 13106-13114. 35. Li, H., Cocco, M. J., Steitz, T. A., and Engelman, D. M. (2001) Conversion of phospholamban into soluble pentameric helical bundle, Biochemistry 40, 6636-6645. 36. Johannsson, C. A., Keightley, G. A., Smith, C. D., Richards, T. R., Hesketh, J. C., and Metcalfe, J. C. (1981) The effects of bilayer thickness on the activity of (Na+/K+)- ATPase J. Biol. Chem. 256, 1643-1650. 37. Starling, J. P., East, J. M., and Lee, A. G. (1993) Separate effects of long chain phosphatidylcholine on dephosphorylation of the Ca2+-ATPase and on Ca2+ binding Biochemistry 32, 1593-1600. 38. Gould, G. W., McWhirter, J. M., East, J. M., and Lee, A. G. (1987) Uptake of Ca2+ mediated by the (Ca2+/Mg2+)-ATPase in reconstituted vesicles, Biochem. J. 245, 751- 755. 39. Simmerman, H. K. B., Lovelace, D. E., and Jones, L. R. (1989) Secondary structure of detergent-solubilized phospholamban, a phosphorylatable, oligomeric protein of cardiac sarcoplasmic reticulum, Biochim. Biophys. Acta, 322-329. 40. Simmerman, H. K., Collins, J. H., Theibert, J. L., Wegender, A. D., and Jones, L. R. (1996) Sequence analysis of phospholamban. Identification of phosphorylation sites and two major structural domains, J. Biol. Chem. 261, 13333-13341. 41. Arkin, I. T., Adams, P. D., Aimoto, S., Engelman, D. M., and Smith, S. O. (1996) Structure of the transmembrane cysteine residues in phospholamban, J. Membr. Biol. 42. Smith, S. O., Kawakami, T., Liu, W., Ziliox, M., and Aimoto, S. (2001) Helical structure of phospholamban in membrane bilayers, J. Mol. Biol. 313, 1139-1148.

153 43. Tatulian, S. A., Jones, L. R., Reddy, L. G., Stokes, D. L., and Tamm, L. K. (1995) Secondary structure and orientation of phospholamban reconstituted in supported bilayers from polarized attenuated total reflection FTIR spectroscopy, Biochemistry 34, 4448-4456. 44. Lee, k.-C., Huo, S., and Cross, T. A. (1995) Lipid-peptide interface:valine conformation and dynamics in the gramicidin channel in a lipid bilayer, Biochemistry 34, 857-867. 45. Killian, J. A., Taylor, M. J., and Koeppe II, R. E. (1992) Orientation of the valine-1 side chain of the gramicidin transmembrane channel and implications for channel functioning: a deuterium NMR study, Biochemistry 31, 11283-11290. 46. Jones, D. H., Rigby, A. C., Barber, K. R., and Grant, W. M. (1997) Oligomerization of t he EGF receptor transmembrane domain: A 2H NMR study in lipid bilayers, Biochemistry 36, 12616-12624. 47. Kinsey, R. A., Kintanar, A., and Oldfield, E. (1981) Dynamics of amino acid side chains in membrane proteins by high field solid state deuterium nuclear magnetic resonance spectroscopy, J. Biol. Chem. 256, 9028-9036. 48. Kochendoerfer, G. G., Jones, D. H., Lee, S., Oblatt-Montal, M., Opella, S. J., and Montal, M. (2004) Functional characterization and NMR spectroscopy on full-length Vpu from HIV-1 prepared by total chemical synthesis, J. Am. Chem. Soc. 126, 2439-2446. 49. Valentine, K. G., Mesleh, M. F., Opella, S. J., Ikura, M., and Ames, J. B. (2003) Structure, topology, and dynamics of myristoylated recoverin bound to phospholipid bilayers, Biochemistry 42, 6333-6340. 50. Mascioni, A., Karim, C., Zamoon, J., Thomas, D. D., and Veglia, G. (2002) Solid-state NMR and rigid body molecular dynamics to determine domain orientations of monomeric phospholamban, J. Am. Chem. Soc. 124, 9392-9393. 51. Nicholson, L. K., Asakura, T., Demura, M., and Cross, T. A. (1993) A method for studying the stucture of uniaxially aligned biopolymers using solid state 15N-NMR: Application to Bombyx mori silk fibroin fibers, Biopolymers 33, 847-861. 52. Ketchem, R. R., Hu, W., and Cross, T. A. (1993) High-resolution conformation of gramicidin-A in a lipid bilayer by solid-state NMR, Science 261, 1457-1460. 53. Lee, K.-C., and Cross, T. A. (1994) Side-chain structure and dynamics at the lipid- protein interface: Val1 of the gramicidin A channel, Biophys. J. 66, 1380-1387.

154 54. Marassi, F. M., and Opella, S. J. (2000) A solid-state NMR index of helical membrane protein structure and topology, J. Magn. Reson. 144, 150-155. 55. Massiot, D., Fayon, F., Capron, M., King, I., Le calve, S., Alonso, B., Durand, J. O., Bujoli, B., Gan, Z., and Hoatson, G. (2002) Modeling one- and two-dimensional solid- state spectra, Magn. Reson. Chem. 40, 70-76. 56. Howard, K. P., and Opella, S. J. (1996) High-resolution solid-state NMR spectra of integral membrane proteins reconstituted into magnetically oriented phospholipid bilayers, J. Magn. Reson. B 112, 91-94. 57. Hartzell, C. J., Whitfiled, M., Oas, T. G., and Drobny, G. P. (1987) Determination of the 15N and 13C chemical shift tensors of L-[13C]alanyl-L[15N]alanine from the dipole- coupled powder patterns, J. Am. Chem. Soc. 109, 5966-5969. 58. Kovacs, F. A., Denny, J. K., Song, Z., Quine, J. R., and Cross, T. A. (2000) Helix tilt of the M2 transmembrane peptide from influenza A virus:An intrinsic property, J. Mol. Biol. 295, 117-125. 59. Glover, K. J., Whiles, J. A., Wood, M. J., Melacini, G., Komives, E. A., and Vold, R. R. (2001) Conformational dimorphism and transmembrane orientation of prion protein residues 110-136 in bicelles, Biochemistry 40, 13137-13142. 60. Sanders, C. R., and Landis, G. C. (1995) Reconstitution of membrane proteins into lipid- rich bilayer mixed micelles for NMR studies, Biochemistry 34, 4030-4040. 61. Bloom, M., Evans, E., and Mouritsen, O. G. (1991) Physical properties of the fluid lipid- bilayer component of cell membrane: a perspective , Q. Rev. Biophys. 24, 293-397. 62. Tiburu, E. K., Karp, E., Dave, P. C., Damodaran, K., and Lorigan, G. A. (2004) Investigating the dynamic properties of the transmembrane segment of phospholamban incorporated into phospholipid bilayers utilizing 2H and 15N solid-state NMR spectroscopy†, Biochemistry (submitted). 63. Tiburu, E. K., Dave, P. C., Vanlerberghe, J. F., Cardon, T. B., Minto, R. E., and Lorigan, G. A. (2003) An improved synthetic and purification procedure for the hydrophobic segment of the transmembrane peptide phospholamban, Anal. Biochem., 146-151. 64. Davis, J. H., Jeffrey, K. R., Bloom, M., and Valic, M. I. (1976) Quadrupolar echo deuterium magnetic resonance spectroscopy in ordered hydrocarbon chains, Chem. Phys. Lett. 42, 390-394.

155 65. Whiles, J. A., Glover, K. J., Vold, R. R., and Komives, E. A. (2002) Methods for studying transmembrane peptides in bicelles: consequences of hydrophobic mismatch and peptide sequence, J. Magn. Reson. 158, 149-156. 66. Whiles, J. A., Brasseur, R., Glover, K. J., Melacini, G., Komives, E. A., and Vold, R. R. (2001) Orientation and effects of mastoparan X on phospholipid bicelles, Biophys. J. 80, 280-293. 67. Jones, D. H., Barber, K. R., VanDerLoo, E. W., and Grant, W. M. (1998) Epidermal growth factor receptor transmembrane domain: 2H NMR implications for orientation and motion in a bilayer environment, Biochemistry 37, 16780-16787. 68. Lamberth, S., Schimidt, H., Muenchbach, M., Vorherr, T., Krebs, J., Carafoli, E., and Griesinger, C. (2000) NMR solution structure of phospholamban, Helv. Chim. Acta. 83, 2141-2152. 69. Tiburu, E. K., Moton, D. M., and Lorigan, G. A. (2001) Biochim. Biophys. Acta 512, 206-214. 70. Mangels, M. L., Cardon, T. B., Harper, A. C., Howard, K. P., and Lorigan, G. A. ( 2000) Spectroscopic characterization of spin-labeled magnetically oriented phospholipid bilayers by EPR spectroscopy, J. Am. Chem. Soc. 122, 7052-7058. 71. Nagle, J. F., and Tristram-Nagle, S. (2000) Lipid bilayer structure, Curr. Opin. Struct. 10, 474-480. 72. Killian, J. A. (2003) Synthetic peptides as model for intrinsic membrane proteins, FEBS letters 555, 134-138. 73. Lewis, B. A., and Engelman, D. M. (1983) Lipid bilayer thickness varies linearly with acyl chain length in fluid phosphatidylcholine vesicles, J. Mol. Biol. 166, 211-217. 74. Weiner, M. C., and White, S. H. (1992) Structure of a fluid dioleoylphosphochline bilayer determined by joint refinement of x-ray and neutron diffraction data. III. Complete, Biophys. J. 61, 434- 447. 75. Li, M., Reddy, L. G., Bennett, R., Silva, N. D., Jr., Jones, L. R., and Thomas, D. D. (1999) A fluorescence energy transfer method for analyzing protein oligomeric structure: Application to phospholamban, Biophys. J. 76, 2587-2599.

156 76. Reddy, L. G., Jones, L. R., and Thomas, D. D. (1999) Depolymerization of phospholamban in the presence of calcium pump: A fluorescence energy transfer study, Biochemistry 38, 3954-3962. 77. Ludlam, C. F., Arkin, I. T., Liu, X., Rothman, M. S., Rath, P., Aimoto, S., Smith, S. O., Engelman, D. M., and Rothchild, K. J. (1996) Structural model of the phospholamban ion-channel complex in phospholipid-membranes, Biophys. J. 70, 1728-1736.

157