a...
r...
t

L “.u‘rhWWJ...

,
aimmrh.

. }
maﬁﬁw
uVlr .- c ‘

2... 1.”. “aw.
4.1 a... t... x}
5):. .

5:

. 2’7

. $2

. .5: 3:3.
an 5‘

Fit...

. 3, :
31.9..» A...
.I. . 21.03 hr:

.{ I: .

 

 

 

 

 

 

..i.::.:; «at ,

an“? 3 £3...

:. sign? . :3...
ecu“. 1”?

’(u V

 

. a . 1 1.2 x
,. .,.J..l.v..\~. , . , ‘
émﬁgaﬁg

I. < v .. a

 

 

 

mjm .LlBRARY I
2006 Michigan State
University

 

 

This is to certify that the
dissertation entitled

SYNTHESIS, ENHANCED FUSOGENICITY, AND SOLID
STATE NMR MEASUREMENTS OF OLIGOMERIC HIV-1
FUSION PEPTIDE CONSTRUCTS

presented by

RONG YANG

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Chemistg

 

 

ﬁwﬂﬁ ﬂ [deﬁant

Major Professor’s’Signature

W 274%;

Date

MSU is an Affirmative Action/Equal Opportunity Institution

_ - —<.-.-.-o-n-I-I-I-o-I-o-o-o-I-I-a-o-c-o-I-o-I-c-o-0-0-0-1-»-.-

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/55 cﬁﬁwateDmJndd-p. 1 5‘

 

SYNTHESIS, ENHANCED FUSOGENICITY, AND SOLID STATE NMR
MEASUREMENTS OF OLIGOMERIC HIV-1 FUSION PEPTIDE CONSTRUCTS

By

Rong Yang

A DISSERTATION
Submitted to
Michigan State University
In partial fulﬁllment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry

2005

ABSTRACT

SYNTHESIS, ENHANCED FUSOGENICIT Y, AND SOLID STATE NMR
MEASUREMENTS OF OLIGOMERIC HIV-l FUSION PEPTIDE CONSTRUCT S

By

Rong Yang

The HIV-l fusion peptide (FF) is a ~20 residue sequence at the N -terminus of the
viral protein gp41 and plays a key role in catalyzing fusion between viral and target cell
membranes. Peptides composed of the FP sequence serve as useful model systems for
understanding some aspects of fusion catalysis. The gp41 protein is trimeric and during
viral/cell fusion, it is thought that three FPS interact with target cell membranes with their
C-termini in close proximity. In an effort to mimic this oligomeric topology in a peptide
model system, an FP construct (FPtr) was synthesized with three FP strands chemically
bonded at their C-termini through lysine sidechains. The yield is ~1umol for a 10pmol
synthesis. Analytical ultracentrifugation demonstrated that FPtr does not self-associate in
aqueous solution and therefore models the expected FP t0pology of gp41. To assess the
effect of the oligomeric strand topology on fusion, comparative functional fusion assays
were carried out using FPtr, FPdm (a cross-linked FP dimer construct), and FPmn (FP
monomer). The fusion rates (k) varied signiﬁcantly among the constructs with
ku>M>kmn and kt, ~ 40km. Thus, there is a strong correlation of the fusion rate with C-
terminal cross-linking and with the number of strands in the construct.

We applied solid state NMR techniques to characterize the membrane-inserted
structure of FPtr, FPdm, and FPmn. The N-terminal secondary structures of these

peptides are determined by rotational-echo double-resonance (REDOR) chemical shift

measurements. Our NMR results suggest that HIV-l fusion peptides may adopt parallel B
strand arrangement in the virus’ host cell membrane, which is consistent with the existing
high-resolution structure of gp41 soluble ectodomain. We proposed that the parallel
strand arrangement could be part of the reason for the enhanced fusion rate in FPtr and
FPdm because of (1) greater membrane perturbation due to placement of the apolar N-
terminal regions of PP strands on one end of the oligomer; and (2) greater local free
energy released when multiple FP strands bind to the membrane in close proximity. For
fusion peptide constructs inserted into a membrane with different lipid headgroup and
cholesterol composition, we observed helical structure in the vicinity of Phe-8, which
suggest that secondary structure of fusion peptide is lipid-dependent.

In addition, deuterium NMR is used to probe the speciﬁc effect of fusion peptide
on lipid motion and structure. Our ﬁndings are consistent with a model in which the

addition of fusion peptide promotes the formation of curvature in the lipid membranes.

Dedicated to my husband, Guangming Wang.

iv

ACKNOWLEDGMENTS

The guidance and sponsorship from my advisor, Dr. David P. Weliky, is essential
for this work. His intelligence and experience have helped me through many difﬁculties
in research. He is very supportive of independent thinking and always encourages me to
try out new ideas. His passion for truth is the most precious quality of a scientist. I would
also like to thank my committee members, Drs. John L. McCracken, Merling Bruening,
Timothy Watson, and John Allison for their support throughout my graduate study.

I am grateful of the generosity of Drs. James Geiger, Honggao Yan, Babak
Borhan, David L. DeWitt, and Francis J. Castellino, who have allowed me to use their
instruments. I also want to thank Drs. Mary Prorok and Qiuyun Dai who have kindly
assisted me in the analytical ultracentrifugation experiments. I am particularly thankful
for the help from Dr. Joseph Lecom who rescued our peptide synthesizer numerous
times. I appreciate the help from Dr. Alicia Pastor at the MSU Center for Advanced
Microscopy. Assistance from the MSU Mass Spectrometry Facility and the Max T.
Rogers NMR Facility are also appreciated.

I want to thank my family, particularly my husband, Guangming, for their love
and support. My husband came from China to join me, for this reason he gave up his own
business and had to start over from learning English and attending school. I cherished the
sacriﬁce and effort be made for the sake of love. After my daughter Shimiao was born,
Guangming helped me take care of her for half a year. I also want to thank my parents,
who have been taking care of Shimiao in Beijing for two years. Their selfless care for

Shimiao provided me a worry-free time to focus on research. I feel guilty with my

daughter since I was not able to accompany her in her early ages. Her leaving pained me
very much. However, it also motivated me to work harder in order to complete my degree
in order to reunite with her sooner. My mom has brought Shimiao to the US a week ago.
I am so happy to see them again.

Finally, I would like to thank every former and current member in Dr. Weliky’s
research team: Dr. Charles Gabrys, Bhagyashree Khunte, Paul Parkanzky, Christian
Canlas, Chris Wasniewski, Michele Bonder, Zhaoxiong Zheng, Jason Criscione, Vamshi
Cotla, and Qiang Wei. Their collaboration and friendship are very important for me to

complete this work.

vi

TABLE OF CONTENTS

 

 

 

 

Page
LIST OF TABLES ................................................................................. x
LIST OF FIGURES ................................................................................ xi
LIST OF SYMBOLS AND ABBREVIATIONS ............................................. xvii
CHAPTER I. INTRODUCTION ................................................................ 1
BACKGROUND ................................................................... 2
REFERENCE ................................................................... 14
CHAPTER II. DESIGN AND SYNTHESIS OF OLIGOMERIC HIV-l FUSION
PEPTIDE CONSTRUCT S ..................................................... l7
BACKGROUND .............................................................. 18
MATERIALS AND METHODS ............................................ 21
RESULTS AND DISCUSSION ............................................. 26
FPMN .................................................................... 26
FPDM ..................................................................... 26
FPT R (ONE STRAND OUT OF REGISTER) ...................... 26
FPTR (IN REGISTER) ................................................. 32 .
REFERENCE .................................................................. 34
CHAPTER IH. VERIFICATION OF THE SYNTHESIZEI) FUSION PEPTIDE
CONSTRUCTS’ SOLUTION AGGREGATION STATE ............... 36
BACKGROUND .............................................................. 37
MATERIALS AND METHODS ............................................. 38
RESULTS AND DISCUSSION ............................................. 40
CONCLUSION ................................................................ 41
REFERENCE .................... . ............................................. 45
CHAPTER IV. COMPARISION OF THE FUSOGENICITY BETWEﬂ

 

 

DIFFERENT FUSION PEPTIDE OL_IGOME_RS BY FLUORESCENCE

ASSAY .......................................................................... 46
BACKGROUND .............................................................. 47
MATERIALS AND METHODS ........................................... 49
RESULTS ....................................................................... 53

FUSION OF LM3 VESICLES ........................................ 53

vii

DISCUSSION .................................................................. 58

 

REFERENCE ................................................................... 63

CHAPTER V. KINETIC ANALYSIS OF FUSION ASSAY ............................. 65
BACKGROUND .............................................................. 66

MATERIALS AND METHODS ........................................... 67

RESULTS ....................................................................... 72

RATE OF FUSION .................................................... 72
TEMPERATURE-DEPENDENCE OF FUSION RATE ......... 74

DISCUSSION .................................................................. 76

REFERENCE .................................................................. 82

CHAPTER VI. ELECTRON NIICROSCOPY .............................................. 83
BACKGROUND .............................................................. 84

MATERIALS AND METHODS ............................................ 84

RESULTS ........................................................................ 86

DISCUSSION .................................................................. 87

REFERENCE .................................................................. 97

CHAPTER VII. REDOR CHEMICAL SHIFT MEASUREMENTS TO PROBE THE

SECONDARY STRUCTURE OF MEMBRANE-BOUND FUSION

 

PEPTIDE CON STRUCTS .................................................. 98
BACKGROUND ............................................................. 99
METHODS .................................................................... 102
RESULTS ...................................................................... 105
DISUSSION ................................................................. 108
REFERENCE ................................................................ l 12

CHAPTER VIII. DEUTERIUM SOLID STATE NMR EVIDENCE FOR MEMBRANE

BILAYER CURVATURE INDUCED BY THE HIV --1 FUSION

 

PEPTIDE ...................................................................... l 14
BACKGROUND ............................................................ 1 15
MATERIALS AND METHODS .......................................... 123
RESULTS AND DISCUSSION ........................................... 127
CONCLUSION ............................................................... 134
REFERENCE ................................................................ 136
CHAPTER IX. BINDING ASSAY ........................................................... 138
BACKGROUND ............................................................ 139
METHODS ................................................................... 140
RESULTS ..................................................................... 143
DISCUSSION ............................................................... 148

viii

REFERENCE ................................................................ 150

CHAPTER X. SUMMARY AND WORK IN PROGRESS ............................. 151
REFERENCE ............................................................... 158

ix

LIST OF TABLES

 

 

 

Page
CHAPTER IV
Table l. Fusogenicity Ratios as a Function of Peptide Strandzlipid ........................ 57
CHAPTER V
Table 2. Fitting Results for FP-induced Vesicle Fusion According to Equation 1 ......... 73
Table 3. Kinetic Parameters of FP-induced Vesicle Fusion .................................. 76
CHAPTER VIII
Table 4. Lipid T1 and T2 Relaxation Times for Various Peptidezlipid Ratios ............ 133

LIST OF FIGURES

Page
CHAPTER I

Figure 1. HIV Infection: (Left) Model of infection process. (Right) Freeze fracture
electron microscopy of a virion (a) binding to, (b-d) fusing with, and infecting host cell

[7] ...................................................................................................... 3
Figure 2: Model of HIV Infection. FP = fusion peptide region. Time Sequence: left
panel, right panel ...................................................................................... 4
Figure 3. Energy barrier (E) for viral/host cell membrane fusion .............................. 6

Figure 4 [10]. A model for the interaction with membranes of gp41 and other viral fusion
subunits, during membrane fusion. Left, before fusion, viral glycoproteins projects their
receptor-binding domains (spheres) towards the cellular membrane. Brackets, a
conformational change extends the N-terminal fusion peptide (F) towards the cellular
membrane. Center, after the outer layer of the fusion conformation has assembled, the N-
terminal fusion peptides (F) and the C-terminal transmembrane anchors (A) lie near each
other at a site of close apposition of the prefusion membranes. Flexible links between the
central rod and the F and A segments allow variable orientations of the rod with respect
to the two membranes. A second trimer, shown in dotted lines, indicates how such
trimers might aggregate at their hydrophobic ends at initial sites of fusion. Right, after
membrane fusion, both the fusion peptides (F) and transmembrane anchors (A) are
shown in the same membrane, as suggested previously [25-27] .............................. 8

Figure 5 [29]. Comparison of the soluble structure of SIV e-gp41 (ectodomain of gp41)
with the X-ray structures of truncated versions of HIV-1 e-gp41. The residue numbering
corresponds to that of SIV gp41. The X-ray structures were taken from Weissenhom et
al. [10], Chan et a1. [30], and Tan et al. [31] ..................................................... 10
CHAPTER 11

Figure 6 [4]. Ribbon diagram of SIV e-gp41 (soluble ectodomain of gp41) viewed from
the side with respect to the viral membrane ..................................................... 18

Figure 7. (a) The Cysteine cross-linking reaction [16]. (b) Formation of FPdm by cross-
linking ................................................................................................ 23

Figure 8. Cysteine cross-linking reaction to form FPtr (with one strand out of register)..24

Figure 9. FPtr synthesis scheme (by method 2) ................................................. 24

xi

Figure 10. HPLC chromatograms of FPmn, FPdm, and FPtr (one strand out of register).
The gradient was between two solvents A and B, 90/10 water/acetonitrile and 10/90
water/acetonitrile, respectively. (a) 280-nm detected chromatogram of FPCK3W. (b) 280-
nm detected chromatogram of cross-linking of FPCK3W. The main product is the cross-
linked FPdm dimer. (c) 220—nm detected and (d) 280-nm detected chromograms of cross-
linking of FPCCK3W and FPCK3 at 1:4 mol ratio. In panel c, the largest peak that goes
off-scale corresponds to cross-linked FPdm, and the following peak corresponds to cross-
linked FPtr trimer. In panel (I, the FPtr and FPdm peaks have comparable intensities
because at 280 nm, all FPtr is detected whereas only FPdm with at least one FPCCK3W
strand is detected. For panel c-d, the beginning and end of the chromatograms correspond
to 66 and 82 % acetonitrile, respectively ......................................................... 28

Figure 11. MALDI spectrum for (FPCK3W)2 dimer. The peak of 5594.1 corresponds to
FPdm, which has a theoretical molecular mass of 5594.6 g/mol. The peak of 2801.2
arises from the double protonated species and FPCK3W monomer formed by breaking of
disulﬁde bond in FPdm ............................................................................. 29

Figure 12. MALDI spectrum for FPtr (cross-linked product of FPCCK3W and FPCK3 in
a 1:4 mol ratio). Peak A (8104.7) corresponds to FPtr, which has a theoretical molecular
mass of 8121.7 g/mol. Due to the fact that the instrument was neither internally calibrated
nor recently externally calibrated, there was a disagreement of ~ 1 mass unit. The other
peaks arise from a variety of sources as indicated in Figure 13 ............................. 30

Figure 13. Possible reactions in the cross-linking of FPCCK3W and FPCK3, resulting in
the desired product A and undesired products B, C, and D. The theoretical molecular
masses of the products are listed on the right. B1 correspnds to the FPCKKK monomer
due to breaking of disulﬁde bond under laser and the double prononated species (BHz):.

C1 correspnds to the FPCCKKKW monomer and the double protonated species (CH2)2.

D1 corresponds to the FPCKKK monomer, D2 corresponds to the double prononated
species (DH2)2+, and D3 corresponds to the FPCKKKW monomer ........................ 31

Figure 14. MALDI spectrum and 280-nm detected chromatogram for FPtr (in-register).
The peak labeled “3” in the HPLC corresponds to FPtr with a theoretical molecular mass
of 9274.0 g/mol ..................................................................................... 33

CHAPTER III

Figure 15. Amino acid sequences of FP constructs in the analytical ultracentrifugation
experiment ........................................................................................... 39

Figure 16. Sedimentation equilibrium experiments of 104.3 uM FPmn. The main panel
shows the absorbance at 280 nm as a function of the centrifugal radius after reaching the
equilibrium in 20 hours at 45,000 rpm. The best ﬁt to the model for a single species was
obtained with a molecular weight of ~3026 (shown as a solid line through the
experimental points). The upper panel shows the residuals between the data and the ﬁt.42

xii

Figure 17. Sedimentation equilibrium experiments of 41.6 M FPdm. The main panel
shows the absorbance at 280 nm as a function of the centrifugal radius after reaching the
equilibrium in 20 hours at 32,000 rpm. The best ﬁt to the model for a single species was
obtained with a molecular weight of ~5729 (shown as a solid line through the
experimental points). The upper panel shows the residuals between the data and the ﬁt.43

Figure 18. Sedimentation equilibrium experiments of 27.8 uM FPtr. The main panel
shows the absorbance at 280 nm as a function of the centrifugal radius after reaching the
equilibrium in 20 hours at 32,000 rpm. The best ﬁt to the model for a single species was
obtained with a molecular weight of ~10863 (shown as a solid line through the
experimental points). The upper panel shows the residuals between the data and the ﬁt.44

Figure 19. Comparison of sedimentation equilibrium experimental ﬁtting curves of

 

FPmn, FPdm, and FPtr .............................................................................. 44
CHAPTER IV
Figure 20. Resonance energy transfer (RET) assay ............................................. 47

Figure 21. Lipid mixing assays. (a) Data for (i) FPmn (FPK3), (ii) FPmn (FPCK3), (iii)
FPdm (cross-linked product of FPCK3), and (iv) FPtr (cross-linked product of FPCCK3
and FPCK3). For lines i — iv, the peptide strand/lipid mol ratio was 0.010, and peptide
quantitation was made using the BCA assay. (b) Data for lines i — iii, FPmn (FPK3W)
and lines iv — vi, FPdm (cross-linked product of FPCK3W). In lines ii and v, the ratio was
0.010; and in lines iii and vi, the ratio was 0.020. For panel b, peptide quantitation was
made by AAA. For all of the runs in panel a, the same sets of liposomes were used, and
for all of the runs in panel b, the same sets of liposomes were used ......................... 55

Figure 22. Fusogenicity Mf vs peptide strand/lipid mol ratio. Open circles represent data
for FPmn (FPK3W), closed squares represent data for FPdm (obtained from cross-linking
FPCK3W), and crosses represent data for FPtr (obtained from cross-linking FPCCK3W
and FPCK3). All of the data were obtained using the same sets of liposomes. The
uncertainty in each Mf value is i l % ............................................................ 56

CHAPTER V

Figure 23. Amino acid sequences of FF constructs in the kinetic analysis of fusion
reaction ................................................................................................ 68

Figure 24. Fluorescence data and ﬁtting for FPdm-induced PC/PG lipid mixing at 37 °C.
The step-like features at longer times are due to the digitization of the data ............... 70

Figure 25. Arrhenius plot for FPtr-induced LM3 fusion (fast component) .................. 71

xiii

Figure 26. Stopped-ﬂow ﬂuorescence data for FP-induced lipid mixing in (a) LM3 and
(b) PC/PG vesicles at 37 °C and [total lipid] = 150 M, [FPmn] = 1.5 pM, [FPdm] = 0.75

 

M, and [FPtr] = 0.50 uM .......................................................................... 73
Figure 27. Arrhenius plots for the fast component of FP-induced lipid mixing in (a) LM3
and (b) PC/PG ....................................................................................... 75
Figure 28. Arrhenius plots for the slow component of FP-induced lipid mixing in (a)
LM3 and (b) PC/PG ................................................................................. 75
Figure 29. Collision model for vesicle fusion ................................................... 77
CHAPTER VI

Figure 30. 0.15 mM LM3 (Magnification = 40,000) ....................................... 89
Figure 31. 1 mM LM3 (Magniﬁcation = 27,000) ........................................... 89
Figure 32. 0.15 mM LM3 + 0.0015 mM FPmn (Magniﬁcation = 27,000) ................ 90
Figure 33. 1 mM LM3 + 0.01 mM FPmn (Magniﬁcation = 27,000) ....................... 90
Figure 34. 0.15 mM LM3 + 0.00075 mM FPdm (Magniﬁcation = 50,000) ............... 91

Figure 35. 1 mM 1M3 + 0.005 mM FPdm (Magniﬁcation = 50,000) ..................... 91

Figure 36. 0.15 mM LM3 + 0.0005 mM FPtr (Magniﬁcation = 50,000) .................. 92
Figure 37. 1 mM LM3 + 0.0033 mM FPdm (Magniﬁcation = 14,000) .................... 92
Figure 38. 0.15 mM PC/PG (Magnification = 14,000) ....................................... 93
Figure 39. 1 mM PC/PG (Magniﬁcation = 14,000) ........................................... 93
Figure 40. 0.15 mM PC/PG + 0.0015 mM FPmn (Magniﬁcation = 40,000) .............. 94
Figure 41. 1 mM PC/PG + 0.01 mM FPmn (Magniﬁcation = 40,000) ..................... 94
Figure 42. 0.15 mM PC/PG + 0.00075 mM FPdm (Magniﬁcation = 14,000) ............ 95
Figure 43. 1 mM PC/PG + 0.005 mM FPdm (Magnification = 40,000) ................... 95
Figure 44. 0.15 mM PC/PG + 0.0005 mM FPtr (Magniﬁcation = 40,000) ................ 96
Figure 45. 1 mM PC/PG + 0.0033 rnM FPtr (Magnification = 14,000) .................... 96
CHAPTER VII

 

xiv

Figure 46. Amino acid sequence of FPmn-, FPdm-, and FPtr-F8CL9N. Each peptide
strand has Cl3 carbonyl carbon label at Phe-8 and N15 amide nitrogen label at Leu-9...102

Figure 47. REDOR subtraction. (So), (Si), and (So-81) represents the full spectrum,
reduced spectrum, and difference spectrum, respectively .................................... 103

Figure 48. REDOR-ﬁltered So - 81 difference spectra for (a) FPmn, (b) FPdm, (c) FPtr
associated with LM3, (d) FPtr associated with PC/PG, and (e) FPmn associated with DPC
micelle. All the peptides have 13C carbonyl carbon label at Phe-8 and 15N amide nitrogen
label at Len-9. Sample (a) was made using 0.4 mol FPKKKW and 40 mole LM3
unilamellar vesicles. Sample (b) was made using 0.2 umol (FPCKKKW); and 40 umole
LM3 unilamellar vesicles. Sample (c) was made using 0.05 pmol FPtr and 15 mole
LM3 unilamellar vesicles. Sample ((1) has 0.05 mol FPtr and 15 umole PC/PG
unilamellar vesicles. Sample (e) contained 0.5 umol of FPK3W and 0.56 mmol of DPC
dissolved in 250 [AL 5 mM HEPES buffer. The displayed spectra are dominated by the
Phe-8 carbonyl signals ............................................................................ 107

Figure 49. Peptide oligomerization and membrane insertion models for (a) parallel strand
arrangement and (b) antiparallel strand arrangement. The arrows indicate the peptide
direction from N-terminus to C-terminus while the numbers label the indices of
individual residues ................................................................................. l 10

Figure 50. Membrane insertion topology according to the “splayed helix pyramid”
model ................................................................................................. 111

CHAPTER VIII

Figure 51 [2]. Left: Diagrammatic sketch of phospholipid arrangements. The bilayer is
the arrangement of phospholipids commonly found in biological and model membranes.
Right: The hexagonal phase (Hn) ................................................................ 116

Figure 52 [3]. Diagrammatic sketch of positive curvature strain (left) and micelle

Figure 53 [10, 11]. Scheme for membrane fusion and inverted phase formation. (A)
Represents two justaposed bilayers which can join to form the stalk, hemifusion
intermediate (B), followed by a transmembrane contact (TMC) intermediate (C), which
will lead to fusion pore formation during membrane fusion ................................. 118

Figure 54. The experimental Pake doublet 2H-NMR powder spectrum for DPPC-dz
deuterated in the til-position of the phosphocholine (PC) polar head group. It is obtained
as the Fourier transform of the quadrupolar echo signal [15] ................................ 119

Figure 55. Model for the bilayer lipid membrane in the external magnetic ﬁeld B0. 11 is
the bilayer local membrane normal. Each carbon position are numbered as 1, 2, 3, .....120

XV

Figure 56. 2-Dimyristoyl-dS4-sn-Glycero-3-Phosphocholine (DMPC-d54) .............. 121

Figure 57. An array of lipid mixture 2H-NMR spectra with varied d2 from 0 to 0.4
second ............................................................................................... 126

Figure 58. 2H spectra of unoriented samples. All spectra were taken at 35°C. The 2H
spectra of a LM3-DMPCdac sample is presented in (a). Similar spectra are presented for
FP23:LM3-DMPCdac at (b) 1:33 and (c) 1:10 peptide lipid ratios and for melittinzLMB-
dac at (d) 1:33 and (e) 1:10 peptide:lipid ratios. The (a) and (c)-(e) samples were made by
method (2) and (b) sample was made by method (1). Each (a)-(e) spectrum was processed
with 150 Hz line broadening and is the sum of 4000-6000 scans ........................... 129

Figure 59. 2H relaxation data. (3) Measurements of outer and inner intensities. low, is
measured between the horns of the outer Pake pattern and the spectrum baseline. line, is
measured between the horns of the inner Pake pattern and an approximate baseline from
the outer Pake pattern. (b) T1 inversion recovery data of the outer Pake pattern. The pure
LM3 (circles) and 1:10 FP23:LM3 (crosses) intensities are plotted as a function of (11.
There is little difference between LM3, 1:10 FP23:LM3 and 1:33 FP23:LM3 (data not
shown). (0) T2 data of the outer Pake pattern. The pure LM3 (circles), 1:33 FP23:LM3
(triangles) and 1:10 FP23:LM3 (crosses) data are plotted as a function of 21. In addition,
the best-ﬁts are shown for the LM3 and 1:10 FP23:LM3 samples. Addition of the peptide

 

causes a 10 - 15% reduction in T2 ............................................................... 132
CHAPTER IX
Figure 60. Amino acid sequences of FP constructs in the binding assessments ........... 142

Figure 61. Titration of FPmn, FPdm, and FPtr with IM3/NBD-PE LUV (98:2 mol %) as
monitored by ﬂuorescence. Fluorescence emission were measured for LUV at 150 M in
5 mM pH 7 HEPES buffer with increasing amounts of added peptide, and the
ﬂuorescence intensities in the range of 526 - 534 nm were integrated and plotted as a
function of fusion peptide strand concentration. Fluorescence data were obtained in two
consecutive experiments. The average values of the two measurements were used in the

plot. x, FPmn; A, FPdm; O, FPtr ............................................................... 146

Figure 62 (a), (b). Titration of FPmn, FPdm, and FPtr with PCPG/NBD-PE LUV (98:2
mol %) as monitored by ﬂuorescence obtained under similar condition in two sets of
experiment. Fluorescence emission were measured for LUV at 150 M in 5 mM pH 7
HEPES buffer with increasing amounts of added peptide, and the ﬂuorescence intensities
in the range of 526 - 534 nm were integrated and plotted as a function of fusion peptide

strand concentration. X, FPmn; A, FPdm; O , FPtr ........................................... 147

xvi

AAA

CD

DMAP

ESR

FMOC

FPdm

FPtr

FRET

HEPES

HPLC

LM

LUV

MALDI

MAS

N—NBD-PE

N-Rh-PE

LIST OF SYNHBOLS AND ABBREVIATIONS

Amino Acid Analysis
Circular Dichroism

4-Dimethylaminopyridine

Electron Spin Resonance

Free Induction Decay

9-Fluorenylmethoxycarbonyl

Fusion Peptide;

Fusion Peptide Dimer

Fusion Peptide Monomer

Fusion Peptide Trimer

Fluorescence Resonance Energy Transfer
N-2-hydroxyethylpiperazine-N’ -2-ethanesulfonic acid
Human Immunodeﬁciency Virus

High Performance Liquid Chromatography

Lipid Mixture

Large Unilamellar Vesicles

Matrix-Assisted Laser Desorption Ionization

Magic Angle Spinning

Nuclear Magnetic Resonance
N-(7-nitro-2,1,3-benzoxadiazol-4-yl)-phosphatidylethanolamine

N—(lissamine rhodamine B sulfonyl)-phosphatidylethanolamine

xvii

PI

POPC

POPE

POPS

REDOR

RFDR

SDS

TFA

TEM

TPPM

Phosphatidylinositol
1-Palmitoyl-2-oleoyl—sn-glycero-3phosphocholine
1-Palmitoy]-2-oleoyl-sn-glycero-3-phosphoethanolamine
1-Palmitoyl-2-oleoyl-sn-glycero-3-[phospho-L-serine]
Rotational Echo Double Resonance

Radio Frequency-dri ven Dipolar Coupling

Sodium Dodecyl Sulfate

Triﬂuoroacetic Acid

Transmission Electron Microscopy

Two Pulse Phase Modulation

xviii

CHAPTER I

INTRODUCTION

BACKGROUND

Membrane fusion is an essential process in life. It underlies many physiological
functions including egg fertilization, synaptic transmission, insulin secretion, and
intracellular protein transport. It is also a key step in the infectious process of enveloped
animal viruses. All these functions require transport of materials across the membranes of
cells, transport vesicles, organelles, and viruses.

Infection of mammalian cells requires the delivery of the viral gene to host cells,
which is made possible when the viral membrane fuses with its target cell membrane,
forming a fusion pore, which allows the viral gene to enter its host (cf. Figure 1). Fusion
between viral and target cell membranes is mediated by viral envelope proteins, which
undergoes conformational changes that bring the two membranes into apposition. It has
been hypothesized that the free energy released by the protein’s conformational changes
is transferred into membrane fusion activation energy and facilitates the merging of two
bilayer membranes [1].

Among the many viral infectious diseases, AIDS has been a serious threat to
human health in the past three decades. Its infection and spread are dependent on the
delivery of the HIV-1 viral RNA to two types of white blood cells: the T4 lymphocyte
cell and the macrophage cell [2].

HIV-1 viral/host cell fusion is mediated by the viral envelope glycoproteins
gp120 and gp41. In particular, the N-terminal region of gp41 (cf. Figure 2), termed the
fusion peptide (FP), plays an important role in initiating fusion. Synthetic fusion peptide
analogs have been reported to induce vesicle fusion and red blood cell fusion in the

absence of the

 

 

 

 

 

 

 

 

Figure 1 [7]. HIV Infection: (Left) Model of infection process. (Right) Freeze fracture
electron micrography of a virion (a) binding to, (b-d) fusing with, and infecting host cell.

 

 

Figure 2 [20]. Model of HIV Infection. FP = fusion peptide region. Time Sequence: left
panel, right panel.
rest of gp41 [3-5]. The site-directed mutation/fusion activity relationships are similar for
viral/cell fusion and fusion peptide-induced vesicle fusion [6-9]. Information about the
HIV-1 fusion peptide-induced membrane fusion is important for understanding HIV
infection and for the design of therapeutic and prophylactic strategies. In addition, the
fusion protein of many other viruses including inﬂuenza and moloney murine leukemia
viruses have the similar N-terminal fusion peptide region [10-12], so understanding the
HIV—1 fusion peptide should be generally useful for understanding the fusion peptide
from these viruses. Furthermore, the mechanism of viral/host cell fusion induced by
fusion proteins resembles that of intracellular vesicle fusion mediated by the SNARE
proteins [13], so investigating the function of viral fusion proteins may provide additional
insight into the understanding of cellular transport processes.

The overall goal of our research is to understand some aspects of the fusion

peptide-induced viral/target cell fusion. And my specific aim is to study how the

oligomerization state and secondary structure of HIV -1 fusion peptide affect its ability to
induce membrane fusion. My approach has been to design and synthesize oligomeric
HIV-1 fusion peptide constructs, compare their activities of inducing artiﬁcial bilayer
membrane fusion, and use solid-state NMR to characterize the membrane-insertion
topology of these peptides as well as the structural and dynamic properties of the bilayer

membrane lipids under the inﬂuence of inserted fusion peptide.

1. Structural Biology of Viral Envelope Proteins

A biomembrane consists of a ﬂuid phospholipid bilayer intercalated with proteins,
carbohydrates, and their complexes. Membrane fusion requires the viruses’ and host
cells’ lipid bilayer to merge into one united whole membrane. There are four steps in
viral/host cell membrane fusion [1]: (1) Viral/host cell binding; (2) formation of small
fusion pores through which electrolytes can pass; (3) mixing of viral and host cell lipids;
(4) formation of a large fusion pore through which larger molecules can pass and creation
of a single virus/host cell moiety. In Figure 1[14], step (1), (3), and (4) are illustrated in a
series of freeze fracture/electron rrricrographs which follow the time evolution of HIV
virion infection of a host cell. The fusion process is controlled on a temporal and

molecular basis such that the host cell remains intact and can serve to the virus.

 

 

 

 

 

 

 

 

Figure 3. Energy barrier (E) for viral/host cell membrane fusion.

Looking at the fusion process (cf. Figure 1), one would wonder how it start with
two membranes and end up with one membrane. The energy of the starting and ending
point are approximately the same, whereas there is an energy barrier to cross if the lipids
from the two bilayers are going to mix together (cf. Figure 3). It is generally believed that
the triggered conformational changes of the viral envelope glycoproteins gplZO and gp41
provide the free energy for the two membranes to merge [15, 16].

On the surface of the HIV-1 virion, gplZO and gp41 form non-covalently bond
complexes, with three sets of gplZO-gp41 complexes associating as trimer [17]. During
the initial step of HIV infection, the gp41-gp120 complex associates with the CD4
receptor and the chemokine receptor (CXCR-4) proteins of target human T or
macrophage cells [18, 19]. This interaction is followed by a series of undetermined
conformational changes that expose the N-terminus of gp41, termed the fusion peptide,

which then inserts into the target cell membrane and catalyzes fusion (cf. Figure 2) [10,

21]. This highly conserved ~ 20 amino acid region is named the “fusion peptide”
because mutations or deletions in this region greatly disrupt viral/host cell membrane
fusion and infection [6-8]. Additionally, radioactive labeling has demonstrated that the
fusion peptide is the only region of the inﬂuenza viral fusion protein which inserts deeply
into the membrane during fusion [22].

A commonly accepted model [10] predicted that during the conformational
change, receptor binding to the gp120 envelope protein moves the gp120 away from gp41
(cf. Figure 4). The fusion peptide of gp41 is exposed and inserted into the target cell
membrane. The target cell membrane is drawn towards the host cell membrane by a
putative gp41 conformational change from an undetermined structure to the observed
coiled-coil gp41 structure. The close proximity of the viral and host cell membranes at a
cluster of envelope protein trimers is then believed to lead to viral/host cell membrane
fusion. Recent models suggest that there are at least six trimers at the fusion site [23].
This number is in approximate agreement with a recent model of inﬂuenza fusion, in

which a minimum of eight HAS were predicted at the fusion site [24].

ml minmnmmmmtﬁ WWW
mmmmmmmai a?

   

          

 

 

 

 

 

 

 

 

 

I“ III". ”I gigimm r—mi ti mist
uttgltttttw “£18133!”
.- maintaining»
8% )0 . -~ mattiiainitar fig
,8 A 8‘ ‘~ ﬁgﬁ -b 4:“; “33
133313 in -,. it 3‘"
I 3‘ I 3333383 milligram “5’”
' Milli?

“aim inn éiiii gar ill”??? [[33ng

Viral membruw

 

Figure 4 [10]. A model for the interaction with membranes of gp41 and other viral fusion
subunits, during membrane fusion. Left, before fusion, viral glycoproteins project their
receptor-binding domains (spheres) towards the cellular membrane. Brackets, a
conformational change extends the N-terminal fusion peptide (F) towards the cellular
membrane. Center, after the outer layer of the fusion conformation has assembled, the N-
terrninal fusion peptides (F) and the C-terminal transmembrane anchors (A) lie near each
other at a site of close apposition of the prefusion membranes. Flexible links between the
central rod and the F and A segments allow variable orientations of the rod with respect
to the two membranes. A second trimer, shown in dotted lines, indicates how such
trimers might aggregate at their hydrophobic ends at initial sites of fusion. Right, after
membrane fusion, both the fusion peptides (F) and transmembrane anchors (A) are
shown in the same membrane, as suggested previously [25-27].

 

This model is supported by recent high-resolution structural data [10, 28, 29].
Figure 5 [29] displays the ribbon diagram of SIV (simian immunodeﬁciency virus) and
HIV -1 gp41 soluble ectodomain determined by solution NMR and X-ray crystallography.
The SIV gp41 is a good model system because its amino acid sequence is highly similar
to that of HIV-1 [29]. These structures are believed to correspond to the protein
conformations after fusion has occurred and perhaps during fusion as well. The
ectodomain of an individual gp41 molecule folds back on itself and the molecules
associate as very stable coiled coil trimers [10, 28]. The inferred proximity of the gp41
transmembrane and fusion peptide domains suggests that gp41 may catalyze membrane
fusion by bringing the virus and host cell close together [10, 29].

In each of these structures, the proteins are trimeric and the three N—termini
(corresponding to about residue 30 in the whole envelope protein) are in close proximity
at the end of an in-register helical coiled—coil. Due to the hydrophobicity of the fusion
peptide region, it was not included in these soluble constructs. However, these structures
end just several residues to the C-terminus of the fusion peptide and they imply that the C-
termini of three fusion peptide strands are in close proximity. Therefore it has been
hypothesized that at least three fusion peptides insert into the target cell membrane during

fusion with their C-termini close to each other.

   

HIV e—gp41 HIV e-gp41
(34-69/112-145) (34-67/112-139)

sw e-gp41 HIV e—gp41
(27-149) (29-76/112-149)

Figure 5 [29]. Comparison of the soluble structure of SIV e-gp41 (ectodomain of gp41)
with the X-ray structures of truncated versions of HIV -1 e-gp41. The residue numbering
corresponds to that of SIV gp41. The X-ray structures were taken from Weissenhom et
al. [10], Chan et al. [30], and Tan et al. [31].

The gp41 trimeric coiled-coil motif was also observed in the structure of the low
pH inﬂuenza haemagglutinin and many other viruses [11, 32-34]. This shared motif,
together with evidence from mutational studies and findings of multiple trimers/high
local fusion peptide concentration at the fusion site[l, 7, 10, 23, 24, 29], suggest a similar
membrane fusion mechanism for all of these viruses in which fusion peptide trimerization
may be a structural requirement for viral/target cell fusion.

There have been various experiments to investigate fusion peptide trimerization

and its effect on the membrane fusion. Recently, enhanced fusion has been observed with

inﬂuenza fusion protein constructs which have a similar trimer motif to gp41 and likely
contain fusion peptides in the biologically relevant trimeric topology [35, 36]. In
addition, solid-state NMR measurements suggested an oligomeric [3 sheet structure for
the HIV-1 fusion peptides associated with phospholipid membrane whose lipid
headgroup and cholesterol composition resembles that of the host cell of HIV-1 virus [37,
38]. Jointly, these experiments support the hypothesis that an oligomeric fusion peptide
topology may plan an important role in fusion.

Therefore, we thought that it would be useful to study the fusion peptide domain
of HIV-1 gp41 as a trimer with the C-termini of three strands chemically cross-linked.
We also thought that it would be interesting to compare the fusogenicity of the cross-
linked trimer and the previously studied non-cross-linked monomeric fusion peptide.

In our research, we chose to study the free fusion peptide (H-
AVGIGALFLGFLGAAGSTMGARS-NHZ) derived from the fusogenic region of the
HIV-1 envelope glycoprotein gp41. The free fusion peptide is a good model to study
viral/host cell fusion because it causes fusion of lipid vesicles and erythrocytes and
because its site-directed mutagenesis/fusion activity relationships are comparable to those
of the fusion peptide domain in the intact protein [6-9]. We designed C-terminal cross-
linked fusion peptide trimers, which are very likely the lowest order oligomerization state
of gp41 during fusion [29]. We also synthesized monomeric and dimeric fusion peptides
and compared these peptides’ capacity of mediating fusion in an intervesicular lipid
mixing assay and examined their structures by solid-state NMR. My studies showed that
the trimeric peptide has a 1540 fold enhanced fusion rate relative'to the monomer and

also has a reduced activation energy of vesicle fusion. These data suggest that

11

trimerization of the fusion peptide is important in viral/cell fusion. Our solid state NMR
measurements provide further insights into the fusion-active topology of HIV-1 fusion

peptide.

2. Solid state NMR measurements of the structure of oligomeric membrane-bound
HI V-I fusion peptide (F P) and the FP-induced membrane bilayer curvature

Solid state nuclear magnetic resonance (NMR) spectroscopy is a novel approach
to determine atomic-level structure and dynamics in biological systems. It is particularly
applicable to systems which are difficult to characterize with X-ray or solution NMR
including membrane, aggregated, and partially ordered proteins. There are two principal
advantages of solid state NMR over the more established techniques of X—ray
crystallography and solution NMR spectroscopy: (1) crystals are not required and (2)
large (>30,000 molecular weight) systems can be studied [39, 40]. Recently, systems
including the B-amyloid ﬁbrils implicated in Alzheimer’s disease [41, 42], the E. coli
serine receptor [43], and a HIV-1 peptide/neutralizing antibody complex [44] have been
studied by solid state NMR.

The techniques I used include: (1) Rotational-echo double-resonance (REDOR)
measurements [45] to probe the secondary structure of membrane associated fusion
peptide oligomers; and (2) Deuterium-NMR relaxation measurement to characterize the
dynamic properties of membrane lipids. In collaboration with Zhaoxiong Zheng and
Michele Bodner, our REDOR experiments revealed predominantly B strand conformation
for FPS associated with LM3 (a lipid mixture containing POPC, POPE, POPS,

sphingomyelin, PI, and cholesterol in a mol ratio of 10:5:2:2:1:10, which resembles the

12

lipid headgroup and cholesterol composition in the host cell of HIV -1 [46]) and helical
conformation for FPs associated with PC/PG (a lipid mixture containing POPC and
POPG in a 4:1 mol ratio). My deuterium-NMR results indicated that the fusion peptide

promotes more curvature in the membrane system.

13

REFERENCE

10.

11.
12.

13.

14.

15.

16.
17.

Bentz, J. (2000) Biophys. J. 78, 886-900.
Haseltine, W. A. and Wong-Stall, F. (1988) in The Science of AIDS (Piel, J ., Eds.)
pp 13-25, W. H. Freeman and Company, New York.

Durell, S. R., Martin, I., Ruysschaert, J. M., Shai, Y., and Blumenthal, R. (1997)
Mol. Membr. Biol. 14, 97-112.
Rafalski, M., Lear, J. D., and DeGrado, W. F. (1990) Biochemistry 29, 7917-22.

Slepushkin, V. A., Melikyan, G. B., Sidorova, M. S., Chumakov, V. M., Andreev,
S. M., Manulyan, R. A., and Karamov, E. V. (1990) Biochem. Biophys. Res.
Commun. 172, 952-7.

Freed, E. O., Myers, D. J ., and Risser, R. (1990) Proc. Natl. Acad. Sci. U. S. A.
87, 4650-4.

Freed, E. O., Delwart, E. L., Buchschacher, G. L., Jr., and Panganiban, A. T.
(1992) Proc. Natl. Acad. Sci. U. S. A. 89, 70-4.

Schaal, H., Klein, M., Gehrrnann, R, Adams, 0., and Scheid, A. (1995) J. Virol.
69, 3308-14.

Delahunty, M. D., Rhee, I., Freed, E. O., and Bonifacino, J. S. (1996) Virology
218, 94-102.

Weissenhom, W., Dessen, A., Harrison, S. C., Skehel, J. J., and Wiley, D. C.
(1997) Nature 387, 426-30.

Fass, D., Harrison, S. C., and Kim, P. S. (1996) Nat. Struct. Biol. 3, 465—9.
Hernandez, L. R., Hoffman, T. R., Wolfsberg, T. G., and White, J. M. (1996)
Annu. Rev. Cell Dev. Biol. 12, 965-974.

Weber, T., Zemelman, B. V., McNew, J. A., Westermann, B., Gmachl, M.,
Parlati, F., Sollner, T. H., and Rothman, J. E. (1998) Cell 92, 759-772.

Grewe, C., Beck, A., and Gelderblom, H. R. (1990) J. Acquir. Immune. Deﬁc.
Syndr. 3, 965-74.

Bentz, J. (2000) Biophys. J. 78.

Rothman, J. E., and Wieland, F. T. (1996) Science 272, 227-34.

Kwong, P. D., Wyatt, R., Sattentau, Q. J ., Sodroski, J ., and Hendrickson, W. A.
(2000) J. Virol. 74, 1961-1972.

14

 

18.

19.

20.

21.

22.

23.

24.
25.

26.

27.

28.
29.

30.
31.

32.

33.

34.

Blumenthal, R., and Dimitrov, D. S. (1997) in Handbook of Physiology, Section
14: Cell Physiology (Hoffman, J. F., and Jamieson, J. C., Eds.) pp 563-603,
Oxford, New York.

Lapham, C. K., Ouyang, J., Chandrasekar, N. Y., Dimitrov, D. S., and Golding,
H. ( 1996) Science 274, 602-605.

Balter, M. (1996) Science 274, 1988-1991.

Wyatt, R., and Sodroski, J. (1998) Science 280, 1884-8.

Durrer, P., Galli, C., Hoenke, S., Corti, C., Gluck, R., Vorherr, T., and Brunner, J.
(1996) J. Biol. Chem. 271, 13417-21.

Munoz-Barroso, I., Durell, S., Sakaguchi, K., Appella, E., and Blumenthal, R.
(1998) J. Cell. Biol. 140, 315-23.

Bentz, J. (2000) Biophys. J. 78, 227-45.

Bullough, P. A., Hughson, F. M., Skehel, J. J ., and Wiley, D. C. (1994) Nature
371, 37-43.

Wharton, S. A., Calder, L. J ., Ruigrok, R. W. H., Skehel, J. J ., Steinhauer, D. A.,
and Wiley, D. C. (1995) EMBO J. 14, 240-246.

Skehel, J. J ., Bizebard, T., Bullough, P. A., Hughson, F. M., Knossow, M.,
Steinhauer, D. A., Wharton, S. A., and Wiley, D. C. (1995) Cold Spring Harbor
Symposia on Quantitative Biology 60, 573-5 80.

Lu, M., Blacklow, S. C., and Kim, P. S. (1995) Nat. Struct. Biol. 2, 1075-82.
Caffrey, M., Cai, M., Kaufman, J., Stahl, S. J., Wingﬁeld, P. T., Covell, D. G.,
Gronenbom, A. M., and Clore, G. M. (1998) EMBO J. 17, 4572-84.

Chan, D. C., Fass, D., Berger, J. M., and Kim, P. S. (1997) Cell 89, 263-73.

Tan, K., Liu, J ., Wang, J ., Shen, S., and Lu, M. (1997) Proc. Natl. Acad. Sci. U. S.
A. 94, 12303-8.

Kobe, B., Center, R. J., Kemp, B. E., and Poumbourios, P. (1999) Proc. Natl.
Acad. Sci. U. S. A. 96, 4319-24.

Baker, K. A., Dutch, R. E., Lamb, R. A., and Jardetzky, T. S. (1999) Mol. Cell 3,
309-19.

Weissenhom, W., Carfi, A., Lee, K. H., Skehel, J. J., and Wiley, D. C. (1998)
Mol. Cell 2, 605-16.

15

35.

36.

37.

38.

39.

40.
41.

42.

43.

45.
46.

Epand, R. F., Macosko, J. C., Russell, C. J ., Shin, Y. K., and Epand, R. M. (1999)
J. Mol. Biol. 286, 489-503.

Lau, W. L., Ege, D. S., Lear, J. D., Hammer, D. A., and DeGrado, W. F. (2004)
Biophys. J. 86, 272-284.

Yang, J ., Gabrys, C. M., and Weliky, D. P. (2001) Biochemistry 40, 8126-37.
Yang, J ., and Weliky, D. P. (2003) Biochemistry 42, 11879-11890.

Grifﬁn, R. G. (1998) Nat. Struct. Biol. NMR Suppl. 508-12.

Marassi, F. M., and Opella, S. J. (1998) Curr. Opin. Struct. Biol. 8, 640-8.

Lansbury, P. T., Jr., Costa, P. R., Griffiths, J. M., Simon, E. J., Auger, M.,
Halverson, K. J ., Kocisko, D. A., Hendsch, Z. S., Ashburn, T. T., Spencer, R. G.,
and et al. (1995) Nat. Struct. Biol. 2, 990-8.

Benzinger, T. L., Gregory, D. M., Burkoth, T. S., Miller-Auer, H., Lynn, D. G.,
Botto, R. E., and Meredith, S. C. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 13407-
12.

Wang, J ., Balazs, Y. S., and Thompson, L. K. (1997) Biochemistry 36, 1699-703.
Weliky, D. P., Bennett, A. E., Zvi, A., Anglister, J ., Steinbach, P. J., and Tycko,
R. (1999) Nat. Struct. Biol. 6, 141-5.

Gullion, T., and Schaefer, J. (1989) J. Magn. Reson. 81, 196-200.

Aloia, R. C., Tian, H., and Jensen, F. C. (1993) Proc. Natl. Acad. Sci. U. S. A. 90,
5181-5.

16

 

CHAPTER H

DESIGN AND SYNTHESIS OF OLIGOMERIC HIV-1 FUSION PEPTIDE

CONSTRUCTS

l7

BACKGROUND

In HIV—1 envelope protein gp41, the N-terminal fusion peptide is a key motif in
initiating membrane fusion. This hydrophobic region is believed to penetrate into the
target cell membrane and initiate fusion. To function properly, the fusion peptide may
need to insert into the target membrane with a precise structure. Specifically, formation
of a correctly assembled fusion peptide oligomer has been postulated to be a requirement
in membrane fusion catalysis [1, 2]. In order to understand how the membrane insertion
topology contributes to the perturbations of the lipid bilayer that are necessary for two
membranes to fuse, the relevant oligomerization state of FP must be known.

The structure of the truncated HIV-l gp41 soluble ectodomain has been

determined by X-ray crystallography and solution NMR [3, 4] (cf. Figure 6).

 

Figure 6 [4]. Ribbon diagram of SIV e-gp41 (soluble ectodomain of gp41) viewed from
the side with respect to the viral membrane.

These structures end ~ 10 residues C-terminal of the fusion peptide. Due to the bulky,
nonpolar residues in the fusion peptide sequence, the peptide fragments that contain it
would have signiﬁcantly increased hydrophobicity. Such peptides tend to aggregate and
are difﬁcult to crystallize. Also, when the fusion peptide is bound with bilayer membrane,
it is nearly immobilized in the lipids and therefore does not have the fast molecular
tumbling property required by solution NMR for line-narrowing. As a result, neither X-
ray crystallography nor solution NMR techniques could resolve the structure of
membrane bound fusion peptide.

Although the fusion peptide region is not included in the established trimeric
structures of the gp41 ectodomain (cf. Figure 6), its likely topology can be imagined,
since the C-terrnini of the fusion peptides are just a few amino acids away from the N-
termini of the gp41 soluble ectodomain. In the trimeric coiled-coil structure of the gp41
soluble ectodomain, the close proximity of the N-termini of the three molecules implies
that at least three fusion peptides insert into the target cell membrane with their C-termini
near each other.

In addition to evidence from envelope protein trimerization, experiments and
modeling studies have shown that the fusion site contains multiple trimers and a
corresponding high fusion peptide concentration [3-7], which indicate the importance of
fusion peptide oligomerization. An important piece of evidence is that the functional
disruptive V2E mutation in the gp41 fusion peptide is trans-dominant, i. e. cells
expressing 10% mutant protein and 90% wild—type protein exhibit only 40% of the fusion

activity of cells with 100% wild-type protein [8]. These data are consistent with a model

19

in which the mutant peptide disrupts the correct assembly of a functionally essential
fusion peptide oligomer [1, 2].

Recently, enhanced fusion has been observed with inﬂuenza fusion protein
constructs which have a similar trimer motif to gp41 and likely contain fusion peptides in
the biologically relevant trimeric topology [9] [10]. In addition, solid state NMR
measurements suggest an oligomeric [3 strand structure for the HIV -1 fusion peptide [11,
12]. Jointly, these experiments support the proposition that an oligomeric topology of the
fusion peptide may plan an important role in fusion.

Synthetic peptides are good models for investigating some aspects of the full
protein-mediated fusion process. By comparing the effects of similar point mutations in
isolated fusion peptides to those studied in intact viruses, many investigators have found
direct correlations between the properties of the isolated fusion peptides and the intact
fusion proteins [13]. For example, the wild-type HIV-1 fusion peptide was compared
with a V2E mutant in which the Val-2 was replaced by Glu-2 [1]. This single substitution
in gp41 abolished the fusion activity of the virus, and substantially reduced activity when
coexpressed with excess wild type protein [8, 14]. Similarly, while the synthetic wild-
type fusion peptide was very fusogenic, the V2E mutant was totally non-fusogenic. In
addition, a mixture of the mutant and wild-type peptide was also non-fusogenic.

The advantage of working with small synthetic peptides is that a relatively large
quantity of these peptides can be acquired. Also, in spectroscopy there will be no
background from the remaining larger portions of the protein.

In most fusion peptide studies to date, the peptides were synthesized as

monomers. Aiming at mimicking the biologically relevant trimeric topology of the HIV-1

20

fusion peptide in its membrane inserted form, we designed and synthesized a fusion
peptide analogue composed of three HIV-1 fusion peptide strands cross-linked at their C-
termini. We characterized the synthetic peptides’ conformation, aggregation, and
interaction with membranes.

Our synthetic approach has greatly benefited from incorporating 3 to 6 C-terminal
lysine residues to increase the fusion peptide analogues’ solubility in aqueous solution.
This approach was originally invented by Tamm et al. [15]. The additional positively
charged C-terminal lysines significantly increase the hydrophilicity of the fusion peptide.
In addition, the lysines minimize self-association of the peptides, which is essential for
interpreting their oligomerization-fusogenicity relationship (cf. chapter 111).

The trimeric fusion peptide synthetic approach is described in this chapter. In
order to delineate the correlation between oligomerization state of fusion peptide and its
fusion catalyzing activity, monomeric and dimeric fusion peptide analogues are also
synthesized. The solution oligomerization state, fusogenicity and solid state NMR

structural measurements on these model peptides will be addressed in Chapters III - VIII.

MATERIALS AND METHODS

Materials. Rink amide resin was purchased from Advanced Chemtech
(Louisville, KY). Fmoc-B-Ala-Wang resin, N-OL-Fmoc-N-e-t-butoxycarbonyl-L—lysine
(Fmoc-Lys(Boc)), and other 9-ﬂuorenylmethoxycarbonyl (FMOC)-amino acids were
obtained from Peptides International (Louisville, KY). N-oc-Fluorenylmethoxycarbonyl-
N-8-4-methyltrityl-L-lysine (Fmoc-Lys(Mtt)) was purchased from Calbiochem-

Novabiochem (La Jolla, CA). All other reagents were analytical grade.

21

 

Fusion Peptide Monomer (FPmn). Monomeric fusion peptides are denoted as
FPmn. Several different constructs were synthesized: FP
(AVGIGALFLGFLGAAGSTMGARS), FPW (AVGIGALFLGFLGAAGSTMGARSW)
FPK3 (AVGIGALFLGFLGAAGSTMGARSKKK), FPCK3
(AVGIGALFLGFLGAAGSTMGARSCKKK), FPCCK3
(AVGIGALFLGFLGAAGSTMGARSCCKKK), FPK3W
(AVGIGALFLGFLGAAGSTMGARSKKKW), FPK6W
(AVGIGALFLGFLGAAGSTMGARSKKKKKKW), FPCK3W
(AVGIGALFLGFLGAAGSTMGARSCKKKW), FPCK6W
(AVGIGALFLGFLGAAGSTMGARSCKKKKKKW), and FPCCK3W
(AVGIGALFLGFLGAAGSTMGARSCCKKKW). These peptides were synthesized as
their C-terminal amides using a peptide synthesizer (ABI 431A, Foster City, CA)
equipped for FMOC chemistry. The sequences all contain the 23 N-terminal residues of
the LAVI, strain of HIV-1 gp41 and were sometimes followed by cysteine(s) for cross-
linking and/or three to six lysines for enhanced solubility and/or a tryptophan as a uv/vis
chromophore [15]. Peptides were cleaved from the resin in a three hour reaction using a
mixture of TFA:HzO:phenol:thioanisolezethanedithiol in a 33:2:2:2:1 volume ratio.
Peptides were subsequently puriﬁed by reversed-phase HPLC using a preparative C13
column (Vydac, Hesperia, CA) and a water/acetonitrile gradient containing 0.1% TFA.
Mass spectroscopy was used to verify peptide purity. Isotopically labeled forms of the
peptides FP23K3W and FP23CK3W were also synthesized with a 13C carbonyl label at
Phe-8 and a 15N label at Leu-9.

Fusion Peptide Dimer (FPdm). Figure 7 (a) displays the cross-linking reaction

22

schemes. Peptides at ~ 5 mM concentration were cross-linked in 10 mM pH 8.2 DMAP
buffer which was open to the atmosphere. Cross-linking was completed within one day.
FPdm dimers were formed in a solution containing monocysteine peptide. Cross-linked
FPdm were purified by reversed-phase HPLC with higher mass peptides eluting at higher
acetonitrile concentrations. Cross-linked peptide masses were checked using SDS gels

and mass spectrometry.

(a)

C00“ ([300—
HgN—(l’J—H Hgﬁ-C—H
Cysteine Cysteine
2H “5 2H
000’ C00 '

Hair—C—H H3N— I ——H

CH2-“l‘5 .‘i—CHZ
Cystine

(b)

FPCKKKKKKW
FPCKKKKKKW + FPCKKKKKKW ——> |
pkg, FPCKKKKKKW
FPdm

Figure 7. (a) The Cysteine cross-linking reaction [16]. (b) Formation of FPdm by
Cysteine cross-linking.

23

Fusion Peptide Trimer ( F Ptr)
Method 1. By Cysteine Cross-linking (with one strand out of register). As
displayed in Figure 8, the ﬁrst version of FPtr was formed in a solution containing

dicysteine peptide and monocysteine peptide in a 1: 4 ratio.

1311 = 3,2 FPCKKK
FPCCKKK + 2 FPCKKK —> FPCCKKK

FPCKKK

FPtr

Figure 8. Cysteine cross-linking reaction to form FPtr (with one strand out of register).

 

 

 

Fmoc-Lys (Mtt)
Fmoc-B-Ala-Resin o. 5 mmol F moc-Lys-B-Ala-Resi n 1% TFA Fmoc-Lys-B-Al a-Fiesin
I ———-» l
0.02 mmol Mtt NH2
Fmoc-Lys-B-Ala-Resin Fmoc-Lys-B-AIa-Resin Fmoc-Lys-B-AIa-Resin
Fmoc-Lys (Mtt) I o . . . .
0.5 mmol FmOC-LIYS 1 /° TFA Fmoc-Lys 32$?ngth FmOO'I-Ys
—— |
Mil NH2 NH2
0.02 mmol 0.01 mmol
Fmoc-L s-B-Ala-Res'n
Fmoc-Lys (80c) y ' Standard Fmoc peptide AVGIGALFLGFLGAAGSTWMGARSKKKKKKA
0 5 mmol Fm oc-Lys synthesus followed by |
' l 80% TFA cleavage AVGIGALFLGFLGAAGSTWMGARSKKKKKK
Fmoc-Lil’s AVGIGALFLGFLGAAGSTWMGARSKKKKKK
Boc FPtr

Figure 9. FPtr synthesis scheme (in—register trimer).

Method 2. By Lysine Scaﬂold (all strand in register). Figure 9 displays the
scheme for creating the trimer scaffold with coupling through amino groups on lysine
sidechains [17, 18]. Each Fmoc-Lys(Mtt) was added in a two hour coupling step using a

peptide synthesizer and the standard Fmoc chemistry for the instrument. After each

24

addition, the resin was taken out of the synthesizer reaction vessel and the Mtt protecting
group on the lysine sidechain was removed by gentle mixing in 4 mL of a 1:5:94 mixture
of trifluoroacetic acid (TFA):triisopropylsilane (TIS):dichloromethane (DCM). The
mixing in the 1% TFA solution was done for two minutes and was followed by removal
of the solution by filtration and a resin wash with DCM. The 1% TFA
reaction/filtration/DCM wash cycle was repeated six times. The clear TFA solution
became yellow when added to the resin and this yellow color was less intense with each
subsequent cycle.

After addition of each Fmoc-Lys(Mtt) and removal of its Mtt group, the trimer
scaffold was completed by addition of Fmoc-Lys(Boc) on the peptide synthesizer with
standard Fmoc chemistry and two hour coupling time. FPtr synthesis was then continued
on the synthesizer using standard chemistry and included addition of non-native lysines
to improve aqueous solubility and tryptophans as 280 nm chromophores for peptide
quantitation. Each amino acid was coupled for four hours and coupling was followed by
acetylation of free NH; groups to terminate any unreacted strands. The isotopically
labeled amino acids 1-13C Fmoc-Phe and 15N Fmoc-Leu were incorporated into the
peptide at Phe-8 and Leu-9, respectively. After completion of the synthesis, peptides
were cleaved from the resin in a three hour reaction using a mixture of
TFA2H202phenol:thioanisole:ethanedithiol in a 33:2:2:2:1 ratio. Peptides were
subsequently puriﬁed ~ by reversed-phase HPLC using a C13 column and a
waterzacetonitrile gradient which varied from 80:20 to 20:80 ratios. FPtr mass was

checked using mass spectrometry and analytical ultracentrifugation.

25

RESULTS AND DISCUSSION

FPmn. Figure 10 (a) displays the 280-nm detected chromatogram of FPCK3W.

FPdm. Figure 10 (b) displays the 280-nm detected reversed-phase chromatogram
of the products formed by cross-linking of FPCK3W. The main peak corresponds to the
FPdm dimer. Peak identiﬁcation was made by mass spectrometry (cf. Figure 10) with
supporting evidence that under the same chromatographic conditions, the FPCK3W
eluted at lower acetonitrile concentration than FPdm.

FPtr (one strand out of register). Figure 10 (0) displays the 220-nm detected
chromatogram of the cross-linking products formed from the reaction of FPCCK3W and
FPCK3 in a 1:4 ratio. The largest peak corresponds to the FPdm dimer, and the last
peak corresponds to FPtr trimer. In terms of increasing hydrophobicity, the molecules can
be ordered FPmn, FPdm, and FPtr, and this also matches the ordering of the acetonitrile
concentration at elution. Additional evidence for the trimer assignment was by detection
of the chromatogram at 280 nm, as displayed in Figure 10 (d). In this chromatogram, the
FPdm and FPtr peaks have approximately equal intensity. The peak intensities in panels
(c) and (d) are qualitatively consistent with 220-nm detection of all FPtr and FPdm and
280-nm detection of all FPtr and only FPdm containing at least one FPCCK3W strand.

The identiﬁcation of FPtr was also made by mass spectrometry (cf. Figure 12).
Figure 13 illustrates the assignment of multiple peaks in Figure 12. There are several
possible reactions in the cross—linking of FPCCK3W and FPCK3, resulting in the desired
trimer A (FPtr) and undesired dimers B, C, and D. Peak B1, D1 correspond to FPCK3,
which was generated by the laser power breaking the disulﬁde bond in B and D. The

double protonated species (3H2)2+ also contribute to Peak 81. Peak C1 correspnds to the

26

 

FPCCKKKW monomer and the double protonated species (CH2)2+. DZ arises from the
double prononated species (DH2)2+. D3 corresponds to the FPCKKKW monomer,
resulting from breaking of the disulﬁde bond in D. The signal intensity of peak A is
relatively small comparing to the other two major peaks. This is possibly due to signal
suppression, which is a common phenomenon in MALDI. The experimental mass of the
various products differs from their theoretical mass by 01-02 %.

As evidenced in Figure 10 (b), FPdm was the major product formed in cross-
linking of either pure FPCK3W or FPCK3. Using 6 mg of monomer starting material, ~ 3
mg of FPdm could be obtained after HPLC puriﬁcation. FPtr was not the major product
of cross-linking (cf. Figure 10 (c)), and after HPLC puriﬁcation and repuriﬁcation, ~ 0.2
mg of FPtr was obtained from a cross-linking reaction that began with 15 mg of

monomer starting material.

27

(a) FPmn

 

 

 

66 76 86 °/oB
(b) FPdm
l 1 1 1 1 l 1 1 1 1_l
68 78 88 %B
FPdm
(C)
W
l 1 1 1 1 l 1 1 J
66 76 82 %B
(d) M
l 1 1 1 1 l 1 1 l
66 76 82 %B

28

 

Figure 10. HPLC chromatograms of FPmn, FPdm, and FPtr (one strand out of register).
The gradient was between two solvents A and B, 90/10 water/acetonitrile and 10/90
water/acetonitrile, respectively. (a) 280-nm detected chromatogram of FPCK3W. (b) 280-
nm detected chromatogram of cross-linking of FPCK3W. The main product is the cross-
linked FPdm dimer. (c) 220-nm detected and (d) 280-nm detected chromograms of cross-
linking of FPCCK3W and FPCK3 at 1:4 mol ratio. In panel c, the largest peak that goes
off-scale corresponds to cross—linked FPdm, and the following peak corresponds to cross-
linked FPtr trimer. In panel d, the FPtr and FPdm peaks have comparable intensities
because at 280 nm, all FPtr is detected whereas only FPdm with at least one FPCCK3W
strand is detected. For panel c-d, the beginning and end of the chromatograms correspond
to 66 and 82 % acetonitrile, respectively.

5594. 1

2801.2

Counts
01
O
O
‘1’

 

 

 

 

 

' j I l ‘ l ‘ I I l
0 2000 4000 6000 8000 10000
Mass (m/z)

Figure 11. MALDI spectrum for (FPCK3W)2 dimer. The peak of 5594.1 corresponds to
FPdm, which has a theoretical molecular mass of 5594.6 g/mol. The peak of 2801.2
arises from the double protonated species and FPCK3W monomer formed by breaking of
disulﬁde bond in FPdm.

29

 
 
 
   
 

 

 

 

 

 

8000~
70004
6000—
{g — 31.01
8 5000— 2612.0 3
CD 1 5214s;
4000a
3000-
2000— 2753.0 02 o A
‘ 2897.9 C1. 03
1000~ 5501.7 8104.7
1
O ‘ l ' l ' l ' l ‘ l
0 2000 4000 6000 8000 10000

Mass (m/z)

Figure 12. MALDI spectrum for FPtr (cross-linked product of FPCCK3W and FPCK3 in
a 1:4 mol ratio). Peak A (8104.7) corresponds to FPtr, which has a theoretical molecular
mass of 8121.7 g/mol. Due to the fact that the instrument was neither internally calibrated
nor recently externally calibrated, there was a disagreement of ~ 17 mass unit. The other
peaks arise from a variety of sources as indicated in Figure 13.

3O

PH = 8.2
FPCCKKKW + 2 FPCKKK -—>

FPCKKK

I
FPCCKKKW

F PCKKK
A

FPCKKK

|
FPCKKK
B

F PCIDCKKKW

FPCCKKKW
C

F PCKKK

I
FPCCKKKW
0

M. W. (g/mol)

8121.7

5222.2 (2612.1)
3 B1

5801.0 (2901.5)

C C1

5511.6

D

(2612.1, 2756.8, 2901.5)
D1 D2 D3

Figure 13. Possible reactions in the cross-linking of FPCCK3W and FPCK3, resulting in
the desired product A and undesired products B, C, and D. The theoretical molecular
masses of the products are listed on the right. 81 correspnds to the FPCKKK monomer
due to breaking of disulﬁde bond under laser and the double prononated species (BH2)2+.
C1 correspnds to the FPCCKKKW monomer and the double protonated species (CH2)2+.
D1 corresponds to the FPCKKK monomer, D2 corresponds to the double prononated
species (DH2)2+, and D3 corresponds to the FPCKKKW monomer.

31

 

FPtr (in register). Using a waterzacetoniuile gradient varying from 80:20 to 20:80
ratios in HPLC purification, FPtr eluted at ~30:70 ratio and was identified by mass
spectroscopy (cf. Figure 14). An additional peak with +128 mass units was also observed
in the FPtr mass spectrum. We believe that this peak was due to peptides which had an
additional lysine on one strand. It has recently been reported that the free e-NHZ group of
a Lys residue can catalyze the removal of the Lys Fmoc group, and in the FPtr synthesis,
this premature Fmoc removal could allow coupling of an additional Lys on the ﬁrst or
second strand [19]. Premature removal of the Fmoc group on the ﬁrst Lys might also lead
to formation of a peptide tetramer but there was at most a minor (<15%) peak in the
tetramer region of the FPtr mass spectrum. In future FPtr syntheses, attempt will be made
to minimize the premature removal of the Fmoc group using the straightforward Mtt

deprotection and coupling protocol from the Albericio et al.[l9].

32

 

10000— ,v.;~i .655
80009
6000~

4000-

Ckants

 

2000‘

 

 

 

j

q

I ' I ‘ I ' I ' I ' I
6000 8000 10000 12000 14000 16000
Mass (m/z)

' 1 ' I
0 2000 4000

Figure 14. MALDI spectrum and 280-nm detected chromatogram for FPtr (in-register).
The peak labeled “3” in the HPLC corresponds to FPtr with a theoretical molecular mass

of 9274.0 g/mol.

33

REFERENCE

1.

10.

ll.

12.

13.

14.

15.

16.

Kliger, Y., Aharoni, A., Rapaport, D., Jones, R, Blumenthal, R., and Shai, Y.
(1997) J. Biol. Chem. 272, 13496-505.

Pritsker, M., Rucker, J., Hoffman, T. L., Doms, R. W., and Shai, Y. (1999)
Biochemistry 38, 11359-71.

Weissenhom, W., Dessen, A., Harrison, S. C., Skehel, J. J., and Wiley, D. C.
(1997) Nature 387, 426-30.

Caffrey, M., Cai, M., Kaufman, J., Stahl, S. J., Wingﬁeld, P. T., Covell, D. G.,
Gronenbom, A. M., and Clore, G. M. (1998) EMBO J. I 7, 4572-84.

Munoz-Barroso, I., Durell, S., Sakaguchi, K., Appella, E., and Blumenthal, R.
( 1998) J. Cell. Biol. 140, 315-23.

Bentz, J. (2000) Biophys. J. 78, 227-45.
Bentz, J. (2000) Biophys. J. 78, 886-900.

Freed, E. 0., Delwart, E. L., Buchschacher, G. L., Jr., and Panganiban, A. T.
(1992) Proc. Natl. Acad. Sci. U. S. A. 89, 70-4.

Epand, R. F., Macosko, J. C., Russell, C. J ., Shin, Y. K., and Epand, R. M. (1999)
J. Mol. Biol. 286, 489-503.

Lau, W. L., Ege, D. S., Lear, J. D., Hammer, D. A., and DeGrado, W. F. (2004)
Biophys. J. 86, 272-284.

Yang, J ., Gabrys, C. M., and Weliky, D. P. (2001) Biochemistry 40, 8126-37.
Yang, J ., and Weliky, D. P. (2003) Biochemistry 42, 11879-11890.

Durell, S. R., Martin, I., Ruysschaert, J. M., Shai, Y., and Blumenthal, R. (1997)
Mol. Membr. Biol. 14, 97-112.

Freed, E. 0., Myers, D. J ., and Risser, R. (1990) Proc. Natl. Acad. Sci. U. S. A.
87, 4650-4.

Han, X., and Tamm, L. K. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 13097-
13102.

Lehninger, A. L., Nelson, D. L., and Cox, M. M. (1997) Principles of
Biochemistry, 2nd Edition ed., Worth Publishers, New York, NY.

34

17.

18.

19.

Aletras, A., Barlos, K., Gatos, D., Koutsogianni, S., and Mamos, P. (1995) Int. J.
Pept. Protein Res. 45, 488-496.

In (1999) Novabiochem Catalog and Peptide Synthesis Handbook pp S 32-34,
Calbiochem-Novabiochem Corp.

Farrera-Sinfreu, J., Royo, M., and Albericio, F. (2002) Tetrahedron Lett. 43,
7813-7815.

35

CHAPTER III

VERIFICATION OF THE SYNTHESIZED FUSION PEPTIDE CONSTRUCTS’

SOLUTION AGGREGATION STATE

36

BACKGROUND

There are a significant number of bulky, nonpolar residues in the HIV-1 fusion
peptide sequence. If synthetic fusion peptides are dissolved in aqueous solution, they tend
to self-associate and form aggregates. In particular, our synthetic peptide analogues
containing two or three fusion peptide strands would aggregated more than non-cross-
linked monomeric peptides do.

In fusion assays, FPs bind to membranes from aqueous solution and the
correlation of fusogenicity with numbers of strands could be confounded by prior self-
association of FPS in aqueous solution. For example, analytical ultracentrifugation
demonstrated that 75 [AM FPK3W is monomeric in 5 mM pH 7 HEPES buffer. A mixture
of monomer and higher order oligomerization states were observed for FPdm (cross-
linked product of FPCK3W in a 1:1 ratio) and FPtr (cross-linked product of FPCCK3W
and FPCK3 in a 1:4 ratio) at concentrations of 40 and 20 11M, respectively [1]. Thus, the
possibility exists that some part of the enhanced fusogenicity of FPdm and FPtr is related
to their oligomeric state in aqueous solution. On the other hand, studies of different FPmn
peptide constructs that were either monomeric or oligomeric in aqueous solution did not
show a clear correlation between Mf and oligomeric state [2]. In addition, for the
influenza fusion peptide, there does not appear to be a correlation between the peptide
oligomeric state in aqueous solution and its ﬁnal oligomerization state in the membrane,
which suggests that oligomers can break up and/or (re)form at the membrane interface
[3].

One goal of the present work is to determine the correlation between the different

numbers of strands in FPmn, FPdm, and FPtr with their respective fusion catalyzing

37

activities. Therefore, it is crucial to obtain FPmn, FPdm, and FPtr constructs which do not
self-associate in aqueous solution. In order to speciﬁcally address the effect of
oligomerization, we have synthesized cross-linked fusion peptides with longer (6) C-
terminal lysine sequences. Analytical ultracentrifugation of inﬂuenza fusion peptides
showed that additional lysines could change the peptide from an oligomeric to a
monomeric state in solution [3]. Here, we observed similar effect for the cross-linked
HIV-l fusion peptides.

Analytical ultracentrafugation experiments were performed on synthetic fusion
peptide constructs in order to assess their self-association. Our results indicated that
FPmn, FPdm, and FPtr do not self-associate at FP concentrations comparable to those in

stock solutions of the functional assays.

MATERIALS AND METHODS

Peptides. The amino acid sequences of FPmn, FPdm, and FPtr tested by analytical
ultra-centrifugation are displayed in Figure 15. Their synthetic protocols were described
in Chapter H. The peptides were dissolved in 5 mM pH 7 HEPES buffer. Their
concentrations were ~100 M for FPmn, 40 11M for FPdm, and 25 1.1M for FPtr. These
concentrations were chosen so that they were comparable to those in stock solutions for

the functional assays. The 280-nm absorbance of these solutions were ~ 0.5 — 0.6.

38

 

FPmn AVGIGALFLGFLGAAGSTMGARSKKKKKKW

FPdm AVGIGALFLGFLGAAGSTMGARSCKKKKKKW

AVGIGALFLGFLGAAGSTMGARSCKKKKKKW

F Ptr AVGIGALFLGFLGAAGSTWMGARSKKKKKKA

AVGlGALFLGFLGAAGSTWMGARSKKKKKK
AVGIGALFLGFLGAAGSTWMGARSKKKKKK

Figure 15. Amino acid sequences of FP constructs in the analytical ultracentrifugation
experiment.

Sedimentation Equilibrium. Sedimentation equilibrium experiments at room
temperature were performed on an analytical ultracentrifuge (Beckman XL-I, Palo Alto,
CA) using an An-60 Ti rotor. The instrument was operated in absorbance mode at 280
nm. Samples were loaded into six-channel epon charcoal-ﬁlled centerpieces equipped
with quartz windows and were equilibrated at rotor speeds of 32000, 45000, or 52000
rpm. Data were ﬁtted using the analysis software supplied by Beckman to a single molar
mass (M) using:

(A/Ao) =exp1M0-vpxr2—roz)(62/2RT>1 (1)
where A and A0 are the experimental absorbencies at radius r and reference radius r0,
respectively, v is the partial specific volume of FPmn, FPdm, or FPtr, p is the buffer
density, (0 is the angular velocity, R is the ideal gas constant, and T is the temperature [4].
This equation assumes no baseline offset and a single value of M, i.e. a single self-
association state for all peptides in solution. The values of v for each peptide were
calculated from the mass average of the partial speciﬁc volumes of the individual amino

acids in FPmn, FPdm, and FPtr [5]. The value of p was set to 1.0 g/ml.

39

 

RESULTS AND DISCUSSION

Figure 16 displays the results of the sedimentation equilibrium experiment at 20
°C for a sample made with 100 11M FPK6W peptide in 5 mM pH 7 buffer. The rotor
speed was 45000 rpm. The top panel shows that the differences between the experimental
and ﬁtted absorbance were small (< 0.02) and random as a function of I, which indicates
that a single-species model is reasonable. The bottom panel displays 280 nm absorbance
(proportional to peptide concentration) as a function of centrifugal radius. With a partial
speciﬁc volume of 0.7703 ml/g and a solvent density of 1.0 g/rnl, the optimal ﬁt the data
set, determined by a non-linear least square method, was a molecular weight of ~ 3000,.
which is close to the monomeric FPK6W mass of 3080 g/mol. Data were also acquired at
52000 rpm and could be ﬁt well to equation (1) with values of M between 2560 and 2810
g/mol. Thus these results indicate that FPK6W at ~ 100 11M concentration is
predominantly monomeric in 5 mM pH 7 HEPES buffer.

Figure 17 and 18 display the sedimentation equilibrium data and ﬁttings for 40
11M FPdm, and 25 11M FPtr, respectively. The rotor speeds were 32000 rpm for both
samples. Using a single molar mass model according to equation (1), and with specific
volume of 0.7657 for FPdm and 0.7700 for FPtr, the ﬁtted molecular mass was ~ 5700
g/mol for FPdm, and ~ 11000 g/mol for FPtr, which are close to the actual non-associated
peptide masses of 6364 and 9274 for FPdm and FPtr, respectively. For both ﬁts, the
residual differences between the calculated and ﬁtted absorbances were randomly
distributed around zero with typical magnitudes less than 0.02. For FPdm, data were also
obtained at rotor speed of 45000 and 52000 rpm with the result of 4800 and 4500 g/mol,

respectively. For FPtr, result of 11600 and 10400 g/mol were respectively obtained at

40

rotor speeds of 32000 and 45000 rpm. These results demonstrate that FPdm and FPtr are
predominantly monomeric under conditions similar to that of the lipid mixing assay.

Figure 19 compares the results of the sedimentation equilibrium experimental curves at
20°C for the sample made with 104.3 uM FPmn, 41.6 uM FPdm, and 27.8 uM FPtr. The
fitted masses of the three peptides are 3000, 5700, and 11000 g/mol, respectively, and are

close to the actual non-associated peptide masses of 3080, 6364, and 9274 g/mol.

CONCLUSION

In conclusion, the analytical ultra-centrifugation results suggest that for fusion
peptide concentrations close to the stock peptide concentration used in ﬂuorescence assay
and solid state NMR measurements, our synthesized fusion peptide constructs do not
aggregate in aqueous solution. This information is essential for the later interpretation of

the impact of fusion peptide oligomerization on membrane fusion catalysis.

41

 

 

 

 

 

 

 

 

 

0.0 , . , . , . 1
6.9 7.0 7.1 7.2
Radius (cm)

Figure 16. Sedimentation equilibrium experiments of 100 uM FPmn. The main panel
shows the absorbance at 280 nm as a function of the centrifugal radius after reaching the
equilibrium in 20 hours at 45,000 rpm. The best fit to the model for a single species was
obtained with a molecular weight of ~3026 (shown as a solid line through the
experimental points). The upper panel shows the residuals between the data and the ﬁt.

42

 

 

 

 

 

 

 

 

 

0.0 . . . T . , . 1
6.8 6.9 7.0 7.1 7.2
Radius (cm)

Figure 17. Sedimentation equilibrium experiments of 40 M FPdm. The main panel
shows the absorbance at 280 nm as a function of the centrifugal radius after reaching the
equilibrium in 20 hours at 32,000 rpm. The best ﬁt to the model for a single species was
obtained with a molecular weight of ~5729 (shown as a solid line through the
experimental points). The upper panel shows the residuals between the data and the ﬁt.

43

 

 

 

 

 

 

 

 

6.9 7.0 7.1 7.2
Radius (cm)

Figure 18. Sedimentation equilibrium experiments of 25 uM FPtr. The main panel shows
the absorbance at 280 nm as a function of the centrifugal radius after reaching the
equilibrium in 20 hours at 32,000 rpm. The best ﬁt to the model for a single species was
obtained with a molecular weight of ~10863 (shown as a solid line through the
experimental points). The upper panel shows the residuals between the data and the fit.

 

1.0

Absorbance at 280 nm
0
S"

 

 

 

 

6.8 6.9 7.0 7.1 7.2
Radius (cm)

Figure 19. Comparison of sedimentation equilibrium experimental ﬁtting curves of
FPmn, FPdm, and FPtr.

 

REFERENCE

1.

2.

Yang, R., Yang, J ., and Weliky, D. P. (2003) Biochemistry 42, 3527-3535.

Yang, J., Prorok, M., Castellino, F. J., and Weliky, D. P. (2004) Biophys. J.
87, 1951-1963.

Parkanzky, P. D., Prorok, M., Castellino, F. J., and Weliky, D. P., unpublished
experiments.

Cantor, C. R., and Shimmel, P. R. (1980) Biophysical Chemistry, W. H.
Freeman and Company, New York.

Laue, T. M., Shah, B. D., Ridgeway, T. M., and Pelletier, S. L. (1992) in
Analytical Ultracentrlfugation in Biochemistry and Polymer Science (Harding, S.
E., Rowe, A. J., and Horton, J. C., Eds), Royal Society of Chemistry, Cambridge,
UK.

45

CHAPTER IV

COMPARISON OF THE FUSOGENICITY BETWEEN DIFFERENT FUSION

PEPTIDE OLIGOMERS BY FLUORESCENCE ASSAY

46

BACKGROUND

In this chapter, a resonance energy transfer (RET) assay is employed to compare
the fusogenicity among different fusion peptide oligomeric constructs. The RET assay
was designed by Struck et al. (1981) [1] to measure lipid mixing during membrane
fusion. It relies on the resonance energy transfer between N-(7-nitr0-2,1,3,-
benaoxadiazol-4-yl)phosphatidylethanolamine (N-NBD-PE), the energy donor, and N-
(lissamine Rhodamine B sulfonyl)phosphatidylethanolamine (N-Rh-PE), the energy
acceptor (cf. Figure 20). In the assay, if fusion occurs, the labeled lipids are diluted in the
unlabeled membranes. Since RET depends significantly on the distance between the

donor and acceptor [2], the fusion-associated dilution leads to an increase in ﬂuorescence.

.9” “"9.

gr... 8::

“-3- 55 9““:-
“333,213: Peptide § 19%:

+ = a

‘0“!!! .5. 2:-

~"3 - 3 m

.r x
“5'8: 93‘

= Fluorescence Donor

a = Fluorescence Acceptor

Figure 20. Resonance energy transfer (RET) assay.

47

Because of its sensitivity and the ability to continuously monitor the events of
fusion from the very beginning stage [3], RET assay is well suited to probe fusogenic
peptides and proteins. This approach has been commonly used in the study of
phospholipid bilayer fusion promoted by viral fusion proteins and synthetic analogues [4,
5].

Recently, fusion assays on inﬂuenza fusion protein constructs have demonstrated
that oligomerization of fusion peptide considerably enhances lipid mixing. Epand et al.
has studied a 127-residue FHA2 protein construct including both the fusion peptide
domain as well as the ~ 100-residues C-terrrrinal of the peptide [4-10]. Glutaraldehyde
cross-linking showed that this construct is predominantly trimeric in either 2% n-octyl B-
D-glucoside or 0.5% Triton X-100 detergent. When a solution of 100 uM FHA2 in 0.1%
Triton was added to a lipid vesicle solution, the induced lipid mixing was signiﬁcantly
higher than that induced by the original inﬂuenza fusion peptide. In another experiment,
DeGrado et al. [11] synthesized a peptide that incorporates the wild-type inﬂuenza
fusogenic peptide sequence (ch31) followed by a VaLd peptide sequence that self-
assembles into a coiled-coil trimer at neutral pH and aggregates at acidic pH [12]. At
acidic pH, the fusogenicity of ch31 is higher than that of the ﬂu peptide by itself.
Collectively, these results are consistent with the hypothesis that oligomerization of
fusogenic peptides promotes membrane fusion.

In this chapter, we compared the fusogenicity of the three synthetic HIV -1 fusion
peptide constructs: FPmn, FPdm, and FPtr in an intervesicle lipid mixing assay. At
peptide strand/lipid mol ratios between 0.0050 and 0.010, the ﬁnal extent of lipid mixing

for the dimer and trimer was 2-3 times greater than for the monomer. These data suggest

48

that the higher local concentration of peptide strand in the cross-linked peptides enhances
fusogenicity and that oligomerization of the fusion peptide in gp41 may enhance the rate
of viral/target cell membrane fusion. This effect is in addition to the role of gp41
trimerization in stabilizing the gp41 coiled-coil structure which may aid fusion catalysis
by bringing the viral and target cell membrane into apposition [13, 14]. The fusion
enhancement effect of C-terminal cross-linking observed in the present study suggests
that a higher local peptide concentration is an additional factor for gp41’s fusion
catalyzing activity.

It should be noted that the cysteine cross-linked FPtr employed in this chapter has
one strand out of register. In the next Chapter, the fusion assay of an in-register FPtr is
performed on a stopped-ﬂow spectroﬂuorimeter and demonstrates a dramatically '

enhanced fusion rate relative to the monomeric and dimeric fusion peptides.

MATERIALS AND METHODS

Materials. 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 1-
palmitoyl-Z-oleoyl-sn-glycero-3-phosphoethanolamine (POPE), l-palmitoyl-2-oleoyl-sn-
glycero-3-[phospho-L-serine] (POPS), phosphatidylinositol (PI), sphingomyelin, N-(7-
nitro-2,1,3-benzoxadiazol-4-yl)-phosphatidylethanolamine (N-NBD-PE), N-(lissamine
Rhodamine B sulfonyl)-phosphatidylethanolamine (N-Rh-PE) were purchased from
Avanti Polar Lipids, Inc. (Alabaster, AL). The Micro BCATM protein assay was obtained
from Pierce (Rockford, IL). N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid
(HEPES) and Triton X-100 were obtained from Sigma. All other reagents were analytical

grade.

49

Peptides. Monomer peptides are denoted as FPmn. Several different constructs
were synthesized according to the procedure described in Chapter II: FP
(AVGIGALFLGFLGAAGSTMGARS), FPW (AVGIGALFLGFLGAAGSTMGARSW)
FPK3 (AVGIGALFLGFLGAAGSTMGARSKKK), FPCK3
(AVGIGALFLGFLGAAGSTMGARSCKKK), FPCCK3
(AVGIGALFLGFLGAAGSTMGARSCCKKK), FPK3W
(AVGIGALFLGFLGAAGSTMGARSKKKW), FPCK3W
(AVGIGALFLGFLGAAGSTMGARSCKKKW), and FPCCK3W
(AVGIGALFLGFLGAAGSTMGARSCCKKKW). FPdm ((FPCK3)2 or (FPCK3W)2) are
the cross-linking product of monocysteine fusion peptide. FPtr are the product from
cross-linking monocystein and dicystein fusion peptide at 4 to 1 mol ratio. The cross-
linking schemes are described in Chapter II. The FPtr used here has one FP strand out of
register with the other two.

Lipid Preparation. A “LM-3” lipid/cholesterol mixture was used which
approximately reﬂects the lipid and cholesterol content of the HIV -1 virus and its target
T-cells [15, 16]. The LM-3 mixture had POPC, POPE, POPS, sphingomyelin, PI and
cholesterol in a 10:5:2:2:l:10 mole ratio. Lipid and cholesterol powders were dissolved
together in chloroform. The chloroform was removed under a stream of nitrogen
followed by overnight vacuum pumping. Lipid dispersions were then formed with
addition of pH 7 buffer containing 5 mM N-2-hydroxyethylpiperazine-N'-2-
ethanesulfonic acid (HEPES) and 0.01% NaN3 preservative. After homogenization of the
dispersion with ten freeze-thaw cycles, LUVs were prepared by extrusion through a

polycarbonate ﬁlter with 100 nm diameter pores [17].

50

Peptide Concentration Quantitation. Peptide concentrations were quantitated in
three different ways: (1) BCA assay; (2) 280 nm absorbance; and (3) quantitative amino
acid analysis (AAA). The calibration of the BCA assay was done by comparison with
weights of FF. Reproducibility of the BCA assay is i 10%. Quantitation from 280 nm
absorbance was made using an extinction coefficient of 6000 M'lcm‘1 for FPmn
(FPK3W), 12200 M"em-l for FPdm (made from cross-linking FPCK3W), and 7200 M‘
1cm" for FPtr (made from cross-linking FPCCK3W and FPCK3). The reproducibility of
an absorbance measurement was i 0.010 a.u. or typically i 2%. When 280 nm
absorbance gave FPmn, FPdm, and FPtr concentrations of 100 11M, 50 11M, and 25 11M,
respectively, AAA typically gave concentrations of ~ 90, 40, and 20 1.1M, respectively.
Because AAA was considered to be the most accurate technique for absolute
quantitation, concentrations determined by 280 nm absorbance for FPmn, FPdm, and FPtr
solutions were multiplied by 0.9, 0.8, and 0.8, respectively. As will be presented in
Figure 21 and table 1, similar fusion results were obtained when peptide was quantiﬁed
by AAA and by BCA assay.

Lipid Mixing Assay for Membrane Fusion. The resonance energy transfer assay of
Struck, et al. was used to monitor membrane fusion [1]. Two types of 100 nm diameter
LM-3 LUV were prepared. One set contained two mol % of the ﬂuorescent lipid N-NBD-
PE and two mol % of the quenching lipid N-Rh-PE while the other set only contained
unlabeled lipids. Fluorescently labeled and unlabeled vesicles were mixed in a 1:9 ratio.
Following addition of peptide, lipid mixing between labeled and unlabeled vesicles
caused dilution of the labeled lipids with a resulting increase of ﬂuorescence.

Fluorescence was recorded using 4 nm bandwidth on an Instruments S. A. Fluoromax-Z

51

(Edison, NJ) spectroﬂuorimeter operating at excitation and emission wavelengths of

465 nm and 530 nm, respectively. A siliconized glass cuvette was used with continuous
stirring in a thermostated cuvette holder. Measurements were carried out at 37 °C with

2 ml of 150 11M LUV in 5 mM pH 7 HEPES buffer. Peptide solution was added to the
liposome solution to achieve the desired peptide:lipid mol ratio and the change in
ﬂuorescence of the sample was monitored following this addition. The initial residual
ﬂuorescence intensity, F0, referenced zero lipid mixing. After addition of peptide, the
ﬂuorescence F(t) was monitored as a function of time (t). The maximum ﬂuorescence
intensity, me, was obtained following addition of 20 111 of 10% Triton X-100. Percent
lipid mixing at time t is denoted as M(t) and was calculated using:

M(t) = [(F(t)—Fo)/(Fmax-Fo)l X 100 (1)

When the peptide or Triton solution is added to the liposome solution, there are
two competing effects on ﬂuorescence: (1) increase due to lipid mixing; and (2) decrease
due to larger solution volume and corresponding lower ﬂuorophore concentration. The
largest added peptide solution volume was 60 11.1 which increased the liposome solution
volume by 3%. Experimentally, a 3% decrease in ﬂuorescence was also observed when
60 ul of pure buffer was added to the liposome solution. In calculating M(t), F(t) and Fmam
values were adjusted to take into account the small volume change which occurs upon
addition of peptide and detergent, respectively. With sufﬁcient time, M(t) reached a
constant value that is denoted as Mf, the ﬁnal extent of lipid mixing. Mf is used as a
measure of peptide fusogenicity.

In order to make meaningful comparisons between the lipid mixing abilities of

different peptides, the following protocols were deve10ped: (1) the same sets of unlabeled

52

and labeled liposomes were used for all of the peptides in each fusion assay; (2) peptide
concentrations were adjusted so that the same volume of each peptide was added for a
given peptide strand/lipid; and (3) for each peptide and peptide strand/lipid, two runs
were made. The Mf values for the two runs were usually within 2% of each other. The
stock peptide solution concentrations varied between 20 and 900 M, and Mf appeared to
be concentration-independent within this range. For most runs, peptide concentrations

were ~90 11M for FPmn, ~40 11M for FPdm, and ~ 20 11M for FPtr.

RESULTS

Fusion of LM3 Vesicles. Figure 21 (a) displays one example of intervesicle lipid
mixing by FPmn, FPdm, and FPtr at peptide strand:lipid = 0.010. In this case, FPmn was
either FPK3 or FPCK3, FPdm was formed from cross-linking of FPCK3 and FPtr was
formed from cross-linking of FPCCKKK and FPCKKK in a 1:4 mol ratio. Peptide
concentrations were determined by the BCA assay. Because FPdm and FPtr have
respectively twice and three times as many strands per molecule as FPmn, the
peptide:lipid mol ratios for FPmn, FPdm, and FPtr are 0.010, 0.0050, and 0.0033,
respectively. In these data, the Mf values for FPdm and FPtr are about two times greater
than the Mf for FPmn. Figure 21 (b) displays more lipid mixing data for FPmn (FPK3W)
and FPdm (made from cross-linking FPCK3W). In this case, peptide quantitation was
made from AAA. At peptide strand:lipid = 0.0050 and 0.010, FPdm has 2 - 3 times the
fusogenicity of FPmn, while at 0.020 ratio, the enhancement factor is about 1.3.

The ﬁnal extent of lipid mixing, Mf, was used as a general measure of peptide

fusogenicity. Figure 22 displays a plot of Mf as a function of peptide:lipid ratio for assays

53

using FPmn, FPdm, and FPtr with the same sets of liposomes. For these data, FPmn was
FPK3W, FPdm was obtained from cross-linking FPCK3W and FPtr was obtained from
cross-linking FPCCK3W and FPCK3 in a 1:4 mol ratio. The fusogenicity of FPdm and
FPtr are comparable over the experimental range of peptide strand/lipid mol ratios. At
lower (S 0.010) peptide strand/lipid ratios, FPdm and FPtr are signiﬁcantly more
fusogenic than FPmn while at ratios closer to 0.020, the fusogenicities of the three

peptides are much more comparable.

54

 

 

 

 

 

 

 

 

 

C)
c
:3 (m)
2
'o 20 '4 (iv)
'5.
'3
E (ii)
9 10.
a:
‘L (i)
0 1 1 1
0 500 1000
Time (s)
(b)
40 ‘ PM (vi)
a o.-
C (111)
IS 30 “
2
I9
'3 20-
E (V)
9
93 10 - -
0. (IV)
“ :- rf—‘Jr mLﬁ~w‘*~,.L:*rA_~vt-e-;L (ii)
5" - JAL'A';"":"“’r‘-"v‘=-“‘—;var‘:ve.:;-:: (i)
0 I T I
0 50° 1000
Time (s)

Figure 21. Lipid mixing assays. (a) Data for (i) FPmn (FPK3), (ii) FPmn (FPCK3), (iii)
FPdm (cross-linked product of FPCK3), and (iv) FPtr (cross-linked product of FPCCK3
and FPCK3). For lines i — iv, the peptide strand/lipid mol ratio was 0.010, and peptide
quantitation was made using the BCA assay. (b) Data for lines i — iii, FPmn (FPK3W)
and lines iv — vi, FPdm (cross-linked product of FPCK3W). In lines i and iv, the peptide
strand/lipid mol ratio was 0.0050; in lines ii and v, the ratio was 0.010; and in lines iii and
vi, the ratio was 0.020. For panel b, peptide quantitation was made by AAA. For all of the
runs in panel a, the same sets of liposomes were used, and for all of the runs in panel b,

the same sets of liposomes were used.

55

 

 

60

 

 

 

 

o
I
E 40— X
a)
E
G)
9;
.._ I
2 20‘ x
x' 0
o
0

l I l
0 0.005 0.010 0.015
Peptide StrandzLipid

Figure 22. Fusogenicity Mf vs. peptide strand/lipid mol ratio. Open circles represent data
for FPmn (FPK3W), closed squares represent data for FPdm (obtained from cross-linking
FPCK3W), and crosses represent data for FPtr (obtained from cross-linking FPCCK3W
and FPCK3). All of the data were obtained using the same sets of liposomes. The
uncertainty in each Mr value is i 1 %.

56

Table 1. Fusogenicity Ratios as a Function of Peptide Strandzlipid.

 

 

 

Peptiiiosltgﬁiupid Fusogenicity Ratio“
FPdm/FPmn FPtr/FPmn
0.0050 2.6 (0.5) 2.2 (0.3)
0.010 2.1 (0.3) 1.9 (0.4)
0.020 1.3 (0.1) n.d.b

 

 

 

 

 

" This is calculated from the ratios of Mf values. The average fusogenicity ratio is
followed by its standard deviation in parentheses.

b not deterrrrined.

57

When comparing runs using different sets of liposomes, there can be considerable
variation in the Mr values, even for the same peptide and peptide strand/lipid. A more
meaningful measure is the ratio of the Mr value of FPdm or FPtr to the Mf values of
FPmn. For runs from the same sets of liposomes, these fusogenicity ratios were
calculated at peptide strand/lipid mol ratios of 0.0050, 0.010, and 0.020. In some runs, the
data had been obtained at peptide strand/lipid different than these three standard values so
linear interpolation was used to calculate the fusogenicity ratios at the standard values.
There was general consistency in the fusogenicity ratios among different liposome
batches as evidenced in Table l, which displays the average values and standard
deviations of the fusogenicity ratios. For example, the FPdm/FPmn fusogenicity ratio at
peptide strand/lipid = 0.010 represents the average of measurements taken with six
different sets of liposomes. In addition, for a given peptide strand/lipid, the fusogenicity
ratio is independent of the presence or absence of the C-terrninal tryptophan in FPmn,
FPdm, or FPtr and is also independent of the presence or absence of the cysteine in FPmn

(cf. Figure 21 (a)).

DISCUSSION

The topology of insertion of the HIV-1 fusion peptide is strongly suggested from
high-resolution structures of the soluble portion of the gp41 ectodomain [13, 14, 18-20].
In this topology, the C-terrrrini of three fusion peptides are near one another and there is
the possibility that the fusion peptides insert into the membrane in close proximity. In the
present study, we have synthesized fusion peptides which are cross-linked at their C-

termini with a topology close to that of the full protein. Relative to the monomer peptide,

58

these peptides enhance fusogenicity, at least to the lipid mixing stage of membrane
fusion, which suggests that this topology may be a signiﬁcant structural factor in
catalyzing viral/target cell fusion. In the present chapter, we consider the ﬁnal extent of
lipid mixing as a measure of fusogenicity, but in the next chapter, we will carefully
investigate the rates for buildup of lipid mixing. Preliminary analysis shows that these
rates are generally larger for cross-linked peptides than for FPmn, and with further
analysis and temperature-dependent studies, in Chapter V we attempted to quantify the
effect of topology on the activation energy of membrane fusion.

The enhanced fusogenicities of cross-linked fusion peptide are consistent with the
hypothesis that lipid mixing induced by fusion peptides requires some critical local
concentration of peptide strands associated with the membrane. For FPdm and FPtr, the
local concentration is elevated by cross-linking so that relative to FPmn, the cross-linked
peptides allow greater lipid mixing at lower peptide strand/lipid. At higher peptide
strand/lipid, FPmn, FPdm and FPtr would all have surpassed the critical concentration
and would have more similar fusion activities, which is consistent with experimental
observation (cf. Figure 22 and Table 1). The proposed correlation between local peptide
concentration and fusogenicity has some similarity to an earlier hypothesis that fusion
requires a critical number of fusion peptides per vesicle [21]. However, this latter
hypothesis predicts that at constant peptide/lipid, fusion will be greater for larger vesicles,
and we observed that for FPmn, Mf is independent of LM3 vesicle size [22].

In addition to the higher local peptide concentration achieved with cross-linking,
the overall concentration of cross-linked peptide in LM3 could be higher than that of

FPmn because of higher lipid affinity of the cross-linked peptide. However, there is some

59

experimental evidence that points against different affinity of different peptides. We will
address this issue in Chapter IX.

In a more detailed molecular structural picture, there are at least two models that
correlate cross-linking with enhanced fusogenicity. In one model, the main effect is
greater peptide/lipid interactions that lead to larger membrane disruption. For example, a
splayed helix model for oligomers has non-interacting helices that insert obliquely into
the membrane [4]. In a trimer, each helix forms an edge of a trigonal pyramidal structure.
In a second model, the topology correlates with a particular structural arrangement of
interacting peptide strands, perhaps a parallel [3 sheet structure. For FPmn, solid state
NMR measurements have demonstrated that in LM3 at peptide/lipid 2 0.010, the N-
terminal and central regions of the peptides adopt a nonhelical B strand structure [16].
Additional solid state NMR REDOR studies have shown that there are both parallel and
antiparallel arrangements of strands in LM-associated FPmn [23]. However, the C-
terminal cross-linking topology of FPdm and FPtr suggests a parallel arrangement of
strands. Furthermore, it is possible that the parallel alignment is more fusogenic than the
antiparallel alignment. For parallel arrangement, the apolar N-terminal regions of two or
more peptide strands could insert into the hydrophobic interior of the membrane, which
could be more disruptive to the membrane than a single peptide strand. This particular
insertion topology is less likely with the antiparallel arrangement in which the polar and
apolar ends of strands would be on the same side of the oligomer. Thus, FPmn, which has
a mix of both parallel and antiparallel alignments, would be less fusogenic than FPdm or
FPtr. The actual strand arrangement of the membrane-associated cross-linked peptides

will be investigated by solid state NMR (cf. Chapter VI) and other experimental methods.

60

In addition, our REDOR measurement demonstrated that LM3-associated FPmn, FPdm,
and FPtr have similar chemical shifts of 171-173 ppm for the Phe-8 carbonyl carbon,
which is consistent with predominantly B strand conformation for LM3-associated FPs. It
is noted that in the solid state NMR experiments, the peptide structure is observed after
vesicle fusion has occurred. It is also possible that fusion requires a transient structure
that is different from the NMR structure that observed at the end state of fusion.

At peptide strand/lipid mol ratios between 0.004 and 0.015, FPdm and FPtr have
approximately the same fusogenicity (cf. Figure 21, 22 and Table 1). This may be a
general observation that the dimeric and trimeric topologies have approximately the same
effect on lipid mixing. However, it is also noted that FPdm has an in-register strand
alignment, whereas FPtr has one strand out of register with the other two strands (cf.
Chapter II: Figure 7 and 8). In gp41, the three strands are likely in-register, and it is
possible that this arrangement is a structural requirement for enhanced fusogenicity. In a
further experiment, we ﬁnd that vesicle fusion rate induced by an in-register FPtr is
signiﬁcantly faster than the FPdm (cf. Chapter V: Figure 26 and Table 2.). However, the
ﬁnal ﬂuorescence for FPdm and the in-register FPtr are about the same. It appears that
the third strand affects the rate but not the final amount of fusion.

Our results suggest that oligomeric fusion peptide topology may have also
contributed to the enhanced fusogencity observed for the 127-residue FHA2 domain of
inﬂuenza envelope protein [4-6] and for the construct containing residue 1-70 of gp41
[24]. In these systems, there is likely additional fusogenic enhancement because of the
presence of other regions of the fusion protein. For the gp41 system, the overall

fusogenicity of the 1-70 construct was about 10 times greater than that of the 1-23

61

construct, and this is a larger enhancement than we observed from cross-linking the 1-23
construct. In addition, although the 127-residue FHA2 induces cell-cell hemifusion, a 90-
residue construct containing the fusion peptide and ~ 70 C-terminal residues does not
induce cell-cell hemifusion [5]. One possible explanation for this observation is that the
37 extra residues in FHA2 contain a kinked loop region that is required for pH-dependent
association of FHA2 trimers and for cell-cell hemifusion [4, 6, 9, 10].

In this study as well as lipid mixing studies from other groups, it is possible that
the oligomerization of peptides in aqueous solution prior to interaction with the target
membrane impacts the rate and extent of lipid mixing. For our work, analytical
ultracentrifugation demonstrated that FPmn (FPK3W construct) is predominantly a
monomer in the assay buffer, whereas FPdm and FPtr are mixtures of monomeric and
oligomeric cross-linked peptides. Thus, the possibility exists that some part of the
enhanced fusogenicity of FPdm and FPtr is related to their oligomeric state in aqueous
solution. To address this question, we made FP constructs with more C-terminal lysines
and assessed their self-association by analytical ultracentrifugation. Our data revealed
that FPmn, FPdm, and FPtr with 6 C-terminal lysines are predominantly monomeric in
the buffer solution used in ﬂuorescence assay. The fusogenicities of these peptides with
definite solution oligomeric state will be investigated in the next chapter.

In summary, we have shown a correlation between C-terminal cross-linking and
enhanced fusogenicity of HIV-1 fusion peptides. The topology achieved through cross-
linking is similar to the fusion peptide topology thought to exist in the fusogenic form of

gp41. Thus, the topology may play a role in enhancing membrane fusion rates.

62

REFERENCE

1.

2.

10.

11.

12.

13.

14.

15.

16.

Struck, D. K., Hoekstra, D., and Pagano, R. E. (1981) Biochemistry 20, 4093-9.
Fung, B. K.-K., and Stryer, L. (1978) Biochemistry 17, 5241-5248.

Duzgunes, N., and Bentz, J. (1988) in Spectroscopic Membrane Probes (Loew, L.
M., Ed.) pp 117-159, CRC Press, Inc., Boca Raton, FL.

LeDuc, D. L., Shin, Y. K., Epand, R. F., and Epand, R. M. (2000) Biochemistry
39, 2733-9.

Leikina, E., LeDuc, D. L., Macosko, J. C., Epand, R., Shin, Y. K., and
Chemomordik, L. V. (2001) Biochemistry 40, 8378-86.

Epand, R. F., Macosko, J. C., Russell, C. J., Shin, Y. K., and Epand, R. M. (1999)
J. Mol. Biol. 286, 489-503.

Kim, C. H., Macosko, J. C., Yu, Y. Q, and Shin, Y. K. (1996) Biochemistry 35,
5359-65.

Macosko, J. C., Kim, C. H., and Shin, Y. K. (1997) J. Mol. Biol. 267, 1139-48.
Kim, C. H., Macosko, J. C., and Shin, Y. K. ( 1998) Biochemistry 37, 137-44.

Epand, R. F., Yip, C. M., Chemomordik, L. V., LeDuc, D. L., Shin, Y. K., and
Epand, R. M. (2001) Biochim. Biophys. Acta 1513, 167-175.

Lau, W. L., Ege, D. S., Lear, J. D., Hammer, D. A., and DeGrado, W. F. (2004)
Biophys. J. 86, 272-284.

Ogihara, N. L., Weiss, M. S., DeGrado, W. F., and Eisenberg, D. (1997) Protein
Sci. 6, 80-88.

Weissenhom, W., Dessen, A., Harrison, S. C., Skehel, J. J., and Wiley, D. C.
(1997) Nature 387, 426-30.

Caffrey, M., Cai, M., Kaufman, J., Stahl, S. J., Wingfield, P. T., Covell, D. G.,
Gronenbom, A. M., and Clore, G. M. (1998) EMBO J. 17, 4572-84.

Aloia, R. C., Tian, H., and Jensen, F. C. (1993) Proc. Natl. Acad. Sci. U. S. A. 90,
5181-5.

Yang, J., Gabrys, C. M., and Weliky, D. P. (2001) Biochemistry 40, 8126-37.

63

17.

18.

19.

20.

21.

22.

23.

24.

Hope, M. J., Bally, M. B., Webb, G., and Cullis, P. R. (1985) Biochim. Biophys.
Acta 812, 55-65.

Chan, D. C., Fass, D., Berger, J. M., and Kim, P. S. (1997) Cell 89, 263-73.

Tan, K., Liu, J., Wang, J., Shen, S., and Lu, M. (1997) Proc. Natl. Acad. Sci. U. S.
A. 94, 12303-8.

Yang, Z. N., Mueser, T. C., Kaufman, J., Stahl, S. J., Wingﬁeld, P. T., and Hyde,
C. C. (1999) J. Struct. Biol. 126, 131-44.

Nieva, J. L., Nir, S., and Wilschut, J. ( 1998) J. Lipos. Res. 8, 165-182.
Yang, J ., and Weliky, D. P., unpublished experiments.
Yang, J., and Weliky, D. P. (2003) Biochemistry 42, 11879-11890.

Sackett, K., and Shai, Y. (2002) Biochemistry 41, 4678-85.

64

CHAPTER V

KINETIC ANALYSIS OF FUSION ASSAY

65

BACKGROUND

In Chapter IV, the final extent of lipid mixing (Mf) was used as a measure of
fusogenicity. The Mf values associated with FPmn-, FPdm-, and FPtr-induced fusion
were compared to demonstrate that oligomerized HIV-1 fusion peptides could induce
more vesicle fusion than fusion peptides.

Mf values are representative of how much fusion has occurred. We are not only
interested in the amount of fusion, but also interested in how fast the fusion has occurred,
since an increased fusion rate is likely the most important fusion peptide effect on
viral/target cell fusion,

Preliminary analysis showed that fusion rates are generally larger for cross-linked
peptides than for FPmn. In this chapter, we analyzed the kinetics of stopped-ﬂow
ﬂuorescence data to determine the variation of fusion rate with fusion peptide constructs.
We also investigated the temperature dependence of fusion rates and derived the
activation energy, pre-exponential factor, and change of entropy for LM3 and PC/PG
vesicle fusion induced by FPmn, FPdm, and FPtr. A physical model is presented to
explain the observed trends of these variables.

We noted that the FPtr used in Chapter IV induced a similar amount of fusion as
FPdm (cf. Figure 21, 22 and Table 1 in Chapter IV). We speculated that this might be due
to the fact that FPdm has an in-register strand alignment, whereas FPtr has one strand out
of register with respect to the other two strands (cf. Figure 7 and 8, Chapter II).
According to the high-resolution structure of HIV-1 gp41 [1, 2], the most likely fusion
peptide topology would be an in-register trimer, and it is possible that this arrangement is

a structural requirement for enhanced fusogenicity. In order to study the most

66

biologically relevant system, we improved our synthetic strategy and created an FPtr with
an in-register strand arrangement. Our kinetic analyses in this chapter were made on the
new version of FPtr. With the in-register trimer, we found that the rate of vesicle fusion
induced by FPtr was substantially higher than the rate of vesicle fusion induced by FPdm
and FPmn.

Also, we increased the length of the C-terminal lysine sequence from three to six
lysines in order to reduce peptide aggregation. Analytical ultracentrifugation has shown
that the FPmn, FPdm, and FPtr studied in this chapter do not self-associate in an aqueous
solution at concentrations similar to those of the stock solutions used for fusion assays

(cf. Chapter III).

MATERIALS AND METHODS

Peptides. Monomeric, dimeric, and trimeric HIV -1 fusion peptides were
synthesized as described in Chapter H. In particular, FPtr is made as an in-register trimer
by method 2, which is different from the out-of-register trimer used in Chapter IV. All
three peptide constructs included 6 C-terminal lysines (cf. Chapter III). The amino acid

sequences of FPmn, Pde, and FPtr are displayed in Figure 23.

67

FPmn AVGIGALFLG F LGAAGSTMGARSKKKKKKW

F P dm AVGIGALFLGFLGAAGSTMGARSCKKKKKKW

AVGIGALFLGFLGAAGSTMGARSCKKKKKKW

FPtr AVGIGALFLGFLGAAGSTWMGARSKKKKKKA

AVGIGALFLGFLGAAGSTWMGARSKKKKKF
AVGIGALFLGFLGAAGSTWMGARSKKKKKK

Figure 23. Amino acid sequences of PP constructs in the kinetic analysis of fusion
reaction.

Preparation of large unilamellar vesicles (LUVs). Lipids and cholesterol were
purchased from Avanti (Alabaster, AL). LUVS were prepared with one of the two
compositions: (1) "LM3" — 1-palmitoyl-2-oleoyl-sn-glycero-3phosphocholine (POPC),
l-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE), l—palmitoyl-2-oleoyl-
sn-glycero-3-[phospho-L-serine] (POPS), sphingomyelin, phosphatidylinositol (PI) and
cholesterol in a 10:5:2:2:1:10 mol ratio. (2) “PC/PG” — 1-palmitoyl-2-oleoyl-sn-glycero-
3-phosphocholine (POPC) and 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-rac-(1-
glycerol)] (POPG) in a 4:1 mol ratio. LM3 reﬂects the approximate lipid headgroup and
cholesterol composition of host cells of the HIV-1 virus [3] and PC/PG has been a
common composition used in studies of FPS. In addition, solid state NMR structural
measurements are consistent with predominant B strand conformation for LM3-associated
FPS and with predominant helical conformation for PC/PG-associated FPS (cf. Chapter
VII). For either composition, ten percent of the LUVS were prepared with an additional 2
mol% of the ﬂuorescent lipid N-(7-nitro-2,1,3-benzoxadiazol-4-yl)-
phosphatidylethanolarrrine (N-NBD-PE) and 2 mol% of the quenching lipid N-(lissamine

Rhodamine B sulfonyl)-phosphatidylethanolamine (N-Rh-PE).

68

Large unilamellar vesicles (LUVS) of 100 nm diameter were prepared by
extrusion [4] using a procedure described in Chapter IV.

Peptide Concentration Quantitation. Peptide concentrations were quantitated by
280-nm absorbance. We used 6000 M‘lcm‘l, 12000 M‘lcm", and 18000 M'lcm’1 as the
extinction coefﬁcient for FPmn, FPdm, and FPtr, respectively. The reproducibility of an
absorbance measurement was i 0.010 a.u. or typically i 2%.

Stopped-ﬂow Fluorescence Assay and Analysis. In this assay, ﬂuorescence was
recorded on a stopped-ﬂow ﬂuorimeter (Applied Photophysics SX.18MV-R, Surrey, UK)
using excitation and emission wavelengths of 465 and 530 nm, respectively. In the
instrument, one syringe contained a mixture of labeled and unlabeled LUVS at 1:9 ratio
and 300 pM total lipid concentration, and the other syringe contained FPmn, FPdm, or
FPtr at 3, 2, or 1 11M concentration, respectively, as determined from 280 nm absorbance.
The buffer in each syringe solution was 5 mM HEPES pH 7 with 0.01% NaN3. Time zero
in the assay was set by the ~5 ms mixing of the two solutions and ﬂuorescence was then
measured every second for ~1000 s. The temperature of the system was set to a value

speciﬁed between 25 and 40 °C.

69

80-

A 0')
O O
1 1

Fluorescence (a.u.)
N
‘1’

 

 

 

I I I I I I l
o 100 200 300 400 soo 600
Time (s)

Figure 24. Fluorescence data and fitting for FPdm-induced PC/PG lipid mixing at 37 °C.
The step-like features at longer times are due to the digitization of the data.

As displayed in Figure 24, ﬂuorescence data F(t) in arbitrary units were fitted
with equation 1:

F(t) = F0 + F1(l — e'kl') + F2(l — e“) (1)
where F0, F1, F2, k], and k2 are fitting parameters. F0 represents the ﬂuorescence intensity
prior to mixing the LUV and FF solutions. The ﬂuorescence increase after mixing was
modeled as the sum of a fast buildup with overall ﬂuorescence change F1 and rate
constant k], and a Slow buildup with overall ﬂuorescence change F2 and rate constant k2.
Fitting was much poorer with a single buildup model. In a single data set, the best-fit
values of F1 and F2 are generally comparable, and the best-fit k1 ~ 10 k2. The fast
component likely represents the lipid mixing induced by initial interaction of FPS with
membranes and the associated k1 rate constants are listed as the rate constants k in Table
1. The origin of the Slow component in ﬂuorescence buildup is not yet understood.

Temperature-dependence of The Fusion Rate. In an Arrhenius model, k=Ae'E“/RT
and the variations in rate constants would have contributions from differences in
activation energy EE, and differences in pre-exponential factor A [5]. To quantify these

contributions, assays and kinetic analyses were carried out over a temperature range of

70

25-40 °C. As an example, Figure 25 plots the natural log of FPtr-induced fast fusion rate

vs. temperature. According to the Arrhenius equation, a linear relationship exists between

lnk and 1sz
lnk=lnA-E,,/RT (2)
A=(e2k3T/h)eA*S°/R (3)

In the Arrhenius model, the activation energy (E) represents the energy barrier
for lipid mixing. The pre-exponential factor (A) is related to the frequency of collision of
reactants and probability of favorably oriented collision. Their values can be calculated
from the slope and y-intersection of the Arrhenius plot, respectively. A1=S° is the change
of entropy from the initial state to the transition state of the fusion reaction, and its value

could be calculated from the A value using equation 3 [5].

-0.7-
l

-0.8- '\

. \-
-0.9~
-1.0-

-1.1-

-1 .2- I \
-1.3- \

-1.4- \-

Ink

 

 

‘i I 1 ﬁ f r ‘

' I ' I ' I ' I ' I ' 1
3.18 3.20 3.22 3.24 3.26 3.28 3.30 3.32 3.34 3.36
(1/T) x 1000

Figure 25. Arrhenius plot for FPtr-induced LM3 fusion (fast component).

71

RESULTS

Rate of Fusion. Figure 26 displays the stopped-ﬂow ﬂuorescence data for FP-
induced (a) LM3 and (b) PC/PG lipid mixing. In the assays, [FPmn]=2[FPdm]=3[FPtr]
and the strand concentration was constant among the different constructs. The long-time
changes in ﬂuorescence for both LM3 and PC/PG are ordered Aan<Ade<AFm which
suggests that the oligomeric constructs induce more vesicle fusion than FPmn. Because
an increased fusion rate is an important FP effect in viral/target cell fusion, the kinetics of
the stopped-ﬂow ﬂuorescence data were analyzed to determine the variation of fusion
rate with FP construct. In a single data set, the initial increase in ﬂuorescence could be
modeled by a dominant fast exponential buildup, while at longer time, there was an
additional contribution from a slower buildup. The data were ﬁtted with equation 1 and
the fitting results for FP-induced LM3 and PC/PG vesicle fusion at 37°C are listed in
Table 2. The fast component likely represents the lipid mixing induced by initial
interaction of FPS with membranes and the values of its rate are presented as k1. The
values of slow ﬂuorescence build up rates are presented as k2. For both PC/PG and LM3
vesicles, ku>kdm>knm with k1.,,~40k1.mn for LM3 and k1.t,~15k1.,,,n for PC/PG. Thus, there
is very significant correlation of the fusion rate with both the oligomeric topology

enforced by C-terminal cross-linking and with the number of PP strands in the construct.

72

 

 

(a) LM3

100‘ - A FPtr

80- FPdm
3’
9; 60' FPmn
”<5 40-

204

0

 

 

0 200 400 600 800 1000
T1me(s)

(b) PC/PG

 

 

100‘ - FPtr
80‘ FPdm
so-

40_ FPmn
20-
o-

 

 

0 200 400 600 800 1000
Time (s)

Figure 26. Stopped-ﬂow ﬂuorescence data for FP-induced lipid mixing in (a) LM3 and
(b) PC/PG vesicles at 37 °C and [total lipid] = 150 11M, [FPmn] = 1.5 uM, [FPdm] = 0.75

M, and [FPtr] = 0.50 11M.

Table 2. Fitting Results for FP-induced Vesicle Fusion According to Equation 1.

 

 

 

 

 

 

 

 

 

 

 

LM3 PC/PG
FPmn FPdm FPtr FPmn FPdm FPtr
k, (10'3 s“) 11 76 430 13 63 190
k2 (10'3 s") 3 11 40 3 7 11
F1 (10'3 a.u.) 98 90 123 29 80 145
F2 (10'3 a.u.) 60 119 51 92 116 89
F0 (10‘3 a.u.) -612 -604 -612 422 -464 457

 

 

 

 

 

The listed k and F values are at 37°C. Each k value has :20% uncertainty and each F

value has 110% uncertainty.

73

Temperature-dependence of Fusion Rate. We examined the temperature
dependence of reaction rate by obtaining the k values at 25°C, 28°C, 31°C, 35°C, 37°C,
and 40°C. Figure 27 and 28 display k1 and k2 Arrhenius plots and linear ﬁts for some of
the plots. Arrhenius plots of the data yielded values of E21 and lnA for FP-induced lipid
mixing in LM3 and PC/PG. As presented in Table 3, the values of E, and lnA are similar
for FPtr-induced fusion of either LM3 or PC/PG LUVS, and in PC/PG, E;,.t,<Ez..dn.,,Ea.mn
and Au<Adm,Amn. Thus, the changes in E2, and A appear to have competing effects on the
magnitude of k. The entropy of activation A81 at 37°C was approximately calculated with
the transition-state theory expression AS*=R[ln(Ah/kBT)—2] where R, h, and k3 are the
standard physical constants [5]. It is not yet clear why ASI<0 for all constructs and why
ASui<ASdm*,ASmn*. The PC/PG k2 Arrhenius plots were also linear with trends Ea-" < Ea-
dm <Ea."m and lnAtr < lnAdm < lnAmn, which are similar to the trends observed for the
PC/PG k1 Arrhenius plots. We do not yet understand the non-linearity of the Arrhenius

plots for FPmn and FPdm-induced fusion of LM3 LUVS.

74

(a) LM3 (b) PC/PG

FPtr
-1 -‘ w .2 d N

 

 

 

 

-2 .
, A A FPdm -3 - FPdm
J.- A A A 5,:
E 3 - A E
-4 .,
-4 - FPmn Mn
0 o 1
o o o o
‘5 r v 1* V 1 v 1 f r '5 r V I ' r ' I ' I
3.20 3.24 3.28 3.32 3.36 3.20 3.24 3.28 3.32 3.36
(1mx1000K (1fl')x1000K

Figure 27. Arrhenius plots for the fast component of FP-induced lipid mixing in (a) LM3
and (b) PC/PG.

 

 

 

 

(a) LM3 (b) PC/PG
-3.0 - m V
\k -4.5 . vax
'3-5 .. V\\\v~\‘\k\‘v \\ \v‘x‘N-VN
-5.0 ' \\
-4.0 1 \A\\
\
4.5 '1 AK -5'5 4 A\
N N
_ '5.0 'i 1N _ ’60 _‘ \
-5.5 1 \\°
-6.5 ~
-6.01 0 ° ° 0 o o
I I I ‘r I ’7.0 I I I I I
3.20 3.24 3.28 3.32 3.36 3.20 3.24 3.28 3.32 3.36
(1/T)x1000K (1/T)x1000K

Figure 28. Arrhenius plots for the slow component of FP-induced lipid mixing in (a)
LM3 and (b) PC/PG.

75

Table 3. Kinetic Parameters of FP-induced Vesicle Fusion.

 

 

 

 

 

 

LM3 PC/PG
FPmn FPdm FPtr FPmn FPdm FPtr
k (10‘3 s“) 11 76 430 13 63 190
Ea(kJ/mol) n.d. n.d. 33(2) 54(7) 48(3) 27(1)
lnA n.d. n.d. 12(1) 17(3) 16(1) 9(1)
ASIU/mol-K) n.d. n.d. -160 -120 -130 -190

 

 

 

 

 

 

 

 

 

The listed k and AS1 values are at 37°C. Each k value has 3:20% uncertainty and the
ﬁtting uncertainties in Ea and lnA are given in parentheses. Because of signiﬁcant
deviations from linearity, the Arrhenius plots for FPmn and FPdm in LM3 are not
presented.
DISCUSSION

In summary, enforcement of the trimeric biological FP strand topology by cross-
linking reduces E21 and increases the fusion rate by a factor of 15-40. The effect is
observed both for LM3 and for PC/PG fusion in which the membrane-associated FPS
have dominant B strand and helical conformations, respectively (cf. Chapter VII). For
FPdm or FPtr in either conformation, one reason for the fusogenic enhancement may be
placement of the apolar N-terrrrinal regions of strands on one end of the oligomer and
placement of the more polar C-terminal regions of strands on the other end of the
oligomer. FPtr likely has the largest apolar volume which may correlate with the greatest
membrane disruption and fusion rate for this construct. Fusion may also be enhanced by
the larger localized free energy released upon membrane binding of multiple FP strands
in FPdm and FPtr. We note that enhanced fusion has also been observed with inﬂuenza

protein constructs which likely contain FPS in the biologically relevant trimeric topology

[6, 7].

76

For both PC/PG and LM3 vesicles, ku>kdm>kmm In addition, Ea-" (LM3) ~ Ea-"
(PC/PG), lnAtr (LM3) ~ lnA,I (PC/PG). In PC/PG vesicle fusion, Em < 133.6111, Ea.Inn & A,
< Adm , Am. These trends suggest that E, and A play competing roles in affecting k
values.

We set up a simpliﬁed model to interpret our data. In this model, vesicle fusion
occurs in a two-step process:

Step1: P + V —* PV rate constanth

Step2: PV + PV —» A rate constant kp

 

Figure 29. Collision model for vesicle fusion.
where P stands for the fusion peptide, V stands for the unfused vesicle, PV stands for the
unfused vesicle with bound peptide, and A stands for the fused vesicle product.

We consider the maximum rates of the two reactions which are determined by
diffusion-limited collision rates [8].

leVlldt= kB[P][V1 Sknp'lellVl (4)

d[A]/dt = kp[PV]2 s kDPV'PV[PV]2 (5)
In general, the diffusion-limited rate constant kg for collision between particles A and B
can be described by [8]:

kg = 216,111,, + r3)2/ 311mm,, (6)

77

where k3 is Boltzmann’s constant, T is the temperature of reaction, 11(T) is the
temperature-dependent viscosity of the solution, and rA and n; are the radii of the two
particles. In our model, the radii of V and PV are ~ 500 A which are determined by our
sample preparation methods. In equation 4, we assign particle A as V and particle B as P,
while in equation 5, particles A and B are both PV. For equation 4, r3 is given by the
effective hydrodynamic radius (rH) of the fusion peptide [9]:

r11 = [3VMr/(47t1‘fx)]”3 + rw (7)
where V is the speciﬁc volume of the fusion peptide, M, is the molecular mass of the
peptide, N A is Avogadro constant, and rw is the thickness of a hydration layer. Assuming
the speciﬁc volume V = 0.73 cm3/g and a hydration layer rw = 2.4 A [9], the radius of
FPtr is ~ 16 A.

Steinfeld et al. gives a typical kD value of 6 x 109 M'ls'1 for collisions at 300 K
with rA and r3 equal to 2 A [8]. Using this number and our values for rA and r3, kDP'V is ~
50 x 109 M'ls'l, and k1)P V") v is ~ 6 x 109 M'ls'l. In the fusion assay, at FPtrzlipid mole
ratio of 1:300, the initial concentrations for different species are: [P]o = 5 x 10'7 M, and
[V]o = 1.5 x 10'9 M (using an estimated 100,000 lipid molecules per vesicle). In a model
in which step 1 occurs much more rapidly than step 2, the initial rates d[PV]o/dt 2

kB[P]o[V](), and d[A]o/dt z kF [Pv102z kF [V]02. Using k3: kDP'V and k1: = kDPV’PV, we
calculate (d[PV]0/dt)/(d[A]o/dt]), which is predicted by our model to be >> 1.

(dlpvlddt)/(d[A]0/dt]) = kBIPIOIVIO/ kethrﬁ
= kDP'ViPioivw 1<e"V"’V[\/1o2
= kDP'VlPlo/ kDW'WIVio

= 2.5 x 103 S'1/9 s;1 z 300 (8)

78

The result supports our model, and suggests that the second step is the rate-limiting step
in the fusion process.

The fastest experimentally observed fusion rate is ~ 0.4 s'1 and is for FPtr-induced
LM3 fusion (cf. Table 2). This is slower than the calculated PV-PV collision rate (~ 9 s")
and suggests that ~ 5% of the PV-PV collisions lead to fusion.

This two-step model provides some insights of the fusion process. Under our
experimental conditions, the predicted P-V collision rate is significantly greater than the
PV-PV collisions rate, and suggests that peptides ﬁrst bind to vesicles and then vesicles
fuse. In addition, the diffusion-limited rate of the rate-limiting step is 9 S", which is much
faster than the fastest experimental fusion rate (0.4 S") and suggests that relatively few
PV-PV collisions lead to fusion. A possible explanation is that the surface of the unfused
vesicles are not fully covered with fusion peptides, and that the intervesicle contacts in a
large number of collisions do not occur at FP binding sites. This idea is supported by the
following calculations of the vesicle surface area and the FPtr surface area on the vesicle.
The surface area of an LM3 vesicle with radius 500 A is:

Av = 41112 = 4n(500 AF 2 3 x 106 A2
If we consider FPtr as a sphere when it binds with the vesicle, the cross-sectional area of
the sphere can be calculated using rH ~ 16 A for FPtr,

AF = TH‘HZ z 800 A2
At FPtrzlipid mole ratio of 1:300 and 100,000 lipid molecules per vesicle, there would be
~100,000/300 z 300 FPtr bound with one vesicle. The fractional surface area of FPtr on
an LM3 vesicle would be:

300 x 800 112/ 3 x 106 A2 z 0.08

79

The result suggests that only 8% of the vesicle surface is covered with FPtr molecules.
These data support a model in which only PV-PV collisions at FP binding sites would
lead to fusion. This is consistent with the small value of experimentally observed fusion
rate (~ 0.4 8") relative to the PV-PV collision rate (~ 9 8").

Besides the fractional surface area, the membrane-insertion topology of the
different fusion peptide constructs may also affect the fusion rate. rH for FPmn is ~ 12 A
and m for FPdm is ~ 15 A. At a 1:100 FP strand to lipid mole ratio, their corresponding
fractional surface coverage are ~ 0.15 and 0.12, respectively. If the peptide surface
coverage was the only factor in determining the fusion rate, we would expect faster
fusion rate for FPmn- and FPdm-induced vesicle fusion comparing to FPtr-induced
vesicle fusion. However, the experimental results of ktr>kdm>kmn with k1-" ~ 15-40 km“,
suggests that FPtr may adopt a conformation which facilitates membrane fusion more
efﬁciently than FPmn and FPdm. The membrane-insertion topology of FPtr could be in-
register parallel strand arrangement or “splayed helix pyramid” [10], which may be
correlated with the greatest membrane disruption (cf. Chapter VII).

Another interesting topic is the dependence of fusion rate on fusion peptide and
vesicle concentration. We observed an approximately linear relationship between [FPdm]
and FPdm-induced LM3 fusion rate. We also observed an approximately quadratic
dependence of FPmn-induced LM3 fusion rate on [FPmn] (cf. Chapter IV, Figure 22).
We hypothesize that the actual fusion rate constant kg: is determined by the product of kg
and a fusion factor f,

kl: = ka f (9)

where f < 1.

80

If we consider that fusion peptides first bind to the surface of vesicles at a fast
rate, then increasing fusion peptide concentration would result in increasing the surface
density of fusion peptides, but not appreciably change kDPV’PV. The observed increase in
fusion rate with higher fusion peptide concentration would be a result of increasing f, and
is likely related to the greater membrane disruption associated with higher fusion peptide
concentration. The preliminary data on the orders of reaction of PP were obtained using a
regular ﬂuorimeter with an acquisition delay of ~ 1 s, and it would be interesting to
investigate the [FP] dependence of fusion rate for FPmn-_, FPdm-, and FPtr-induced
fusion using a stopped-ﬂow ﬂuorimeter, which has a much shorter acquisition delay (~ 5
ms). Because ﬂuorescence intensity builds up rapidly after mixing FP with vesicles, the
st0pped-ﬂow ﬂuorimeter would provide the short-time ﬂuorescence data which are
essential for more accurate kinetic analysis.

Equation 5 predicts that the fusion rate would have a quadratic dependence of
[PV]. This hypothesis can be tested by further experiments in which the variation of
fusion rates with lipid concentrations are examined. The model predicts that there would
be a quadratic relationship between fusion rate and [lipid], because of the linear

correlation between [lipid] and [PV].

81

REFERENCE

1.

10.

Caffrey, M., Cai, M., Kaufman, J., Stahl, S. J., Wingﬁeld, P. T., Covell, D. G.,
Gronenbom, A. M., and Clore, G. M. (1998) EMBO J. I 7, 4572-84.

Weissenhom, W., Dessen, A., Harrison, 8. C., Skehel, J. J., and Wiley, D. C.
(1997) Nature 387, 426-30.

Aloia, R. C., Tian, H., and Jensen, F. C. (1993) Proc. Natl. Acad. Sci. U. S. A.
90, 5181-5.

Hope, M. J., Bally, M. B., Webb, G., and Cullis, P. R. (1985) Biochim.
Biophys. Acta 812, 55-65.

McQuarrie, D. A., and Simon, J. D. (1997) Physical Chemistry: A Molecular
Approach, University Science Books, Sausalito, CA.

Epand, R. F., Macosko, J. C., Russell, C. J., Shin, Y. K., and Epand, R. M.
(1999) J. Mol. Biol. 286, 489-503.

Lau, W. L., Ege, D. S., Lear, J. D., Hammer, D. A., and DeGrado, W. F.
(2004) Biophys. J. 86, 272-284.

Steinfeld, J. I., Francisco, J. S., and Hase, W. L. (1999) Chemical Kinetics and
Dynamics, Prentice-Hall, Upper Saddle River, NJ.

Cavanaugh, J., Fairbrother, W. J., Palmer, A. G., and Skelton, N. J. (1996)
Protein NMR Spectroscopy Principles and Practice, Academic Press, San
Diego.

LeDuc, D. L., Shin, Y. K., Epand, R. F., and Epand, R. M. (2000)
Biochemistry 39, 2733-9.

82

CHAPTER VI

ELECTRON MICROSCOPY

83

BACKGROUND

Besides the ﬂuorescence assay, we applied transmission electron microscopy
(TEM) to assess the effect of the synthetic HIV-l fusion peptides on membrane fusion.

TEM relies on a focused beam of electrons to examine objects on a very ﬁne scale
and produces magnified images of up to > 100,000 times the original object Size.
Therefore it is suitable for visualizing small particles including viruses, proteins, and
liposomes. In the study of viral membrane fusion, TEM has been employed to
characterize the structural features of fusion intermediates [1,2] and the
aggregation/fibrilization of fusion peptides interacting with membranes [3]. In addition,
TEM has provided direct evidence that a synthetic construct of the HIV-1 fusion peptide
induces vesicle fusion [4]. Despite the broad application of TEM in studying viral
membrane fusion, there has been no TEM data to date that directly compares the
fusogenicity of monomeric HIV -1 fusion peptide and its oligomeric analogs.

In the present study, FPmn, FPdm and FPtr were added to LM3 and PC/PG
vesicles. Negative stain TEM images of the FP-vesicle samples and pure vesicle samples
were obtained. Our results evidenced that synthetic HIV -1 fusion peptides induce vesicle
fusion. This agrees with Nieva et al.’s report [4], where TEM showed that a membrane-
incorporated HIV-1 fusion pepide induces fusion of vesicles. In particular, our TEM data
demonstrated much larger aggregates of fused vesicles in the samples containing FPtr and
FPdm compared to samples containing FPmn. Therefore, TEM provided direct evidence

for the enhanced fusogenicity of oligomeric fusion peptides.

MATERIAL AND METHODS

84

Sample Preparation. LUVS of LM3 and PC/PG were prepared in a Similar
procedure as described in Chapter IV and V [5], except that no ﬂuorescent lipids were
added. FPmn, FPdm, and FPtr were similar to what have been used in Chapter V.
Peptides were dissolved in 5 mM pH 7 HEPES buffer and the concentrations were ~ 100,
50 and 30 M for FPmn, FPdm, and FPtr, respectively. Peptide and LUV were mixed and
gently vortexed ~ 15 minutes prior to negative staining and ~ 30 minutes prior to being
viewed from the electron microscope. Two types of sample were prepared using 1 mM
and 150 M lipid concentration. In all the samples, the peptide strand-to-lipid mol ratio
was 1:100.

Negative Staining. Negativly charged dyes containing heavy atoms commonly
derived from molybdenum, uranium, or tungsten are added to TEM samples in order to
increase the contrast between the objects and their background. This approach is called
negative staining, which is a simple technique for routine examination of structures. They
are commonly used to visualize the edges of protein complexes, macromolecules and
cells in suspension. The stain we used in the present work was 1% uranyl acetate.

During staining, 8 small drop of the peptide/vesicle mixture was deposited on the
carbon coated grid, allowed to settle for approximately one minute, blotted dry if
necessary, and then covered with a small drop of the stain ( 1% uranyl acetate). After a
few seconds, this drop is also blotted dry, and the sample is ready for viewing.

Transmission Electron Microscopy (TEM). Samples were observed using a JEOL
(Japan Electron Optics Laboratories) 100CXII microscope at an accelerating voltage of

100 kv. TEM images were obtained at 5,000 — 100,000 magniﬁcation.

85

RESULTS

Figures 30, 31 display the TEM images of LM3 vesicles at lipid concentration
0.15 mM and 1 mM, respectively. No fusion peptides were added. The average vesicle
size is ~ 100 nm in the 0.15 mM LM3 sample (cf. Figure 30) and is ~ 400 nm in the 1
mM sample (cf. Figure 31). The origin of the different vesicle sizes in the two samples is
not known. In both samples all vesicles are unilamellar, and the majority of vesicles are
unfused.

Figures 32, 34, and 36 display the TEM images of 0.15 mM LM3 in the prescence
of FPmn, FPdm and FPtr, respectively. The FP strand to lipid molar ratios were 1: 100. In
the FPmn-LM3 sample (cf. Figure 32), fused vesicles with diameter of ~ 500 nm co-exist
with unfused vesicles of ~ 100 nm. The number of fused vesicles is signiﬁcantly less than
that of unfused ones. In the FPdm-LM3 sample (cf. Figure 34), most vesicles are fused.
The average size of aggregations is ~ 500 nm. The image of the FPtr-LM3 sample is
dominated by vesicle aggregations bigger than 500 nm (cf. Figure 36). A comparison of
these images clearly indicates that as the oligomerization state of PP increases, the size
and number of aggregated/fused vesicles increase. FPS at higher oligomerization state are
appreciably more fusogenic than FPS in the lower oligomerization state.

Figures 33, 35, and 37 display the TEM images of 1 mM LM3 with FPmn, FPdm,
and FPtr, respectively. Similar trend of fusogenicity as observed in the 0.15 mM FP-LM3
samples are also found in these 1 mM LM3 samples. Increasing fusogenicity is associated
with higher oligomerization States. In the FPmn-LM3 sample (of. Figure 33), the majority
of vesicles are unfused. Their sizes range from ~ 100 nm to ~ 300 nm. In the FPdm-LM3

sample (cf. Figure 35), signiﬁcant amounts of vesicles are fused. The aggregates of fused

86

vesicles are approximately 200-500 nm in size. In FPtr-LM3 sample (Figure 37), the
vesicle aggregates are 2 2 um. Since the vesicles are more concentrated in these samples,
it is easier to ﬁnd fused vesicles in the TEM images, and the aggregates seem to be bigger
than those in the 0.15 mM samples.

Figures 38 and 39 display the TEM images of PC/PG vesicles at lipid
concentration 0.15 mM and 1 mM, respectively. No fusion peptides were added. The
majority of vesicles are unfused with a few exceptions. The vesicle sizes range from ~
100 nm to 300 nm.

Figures 40, 42, and 44 display the TEM images of 0.15 mM PC/PG in the
presence of FPmn, FPdm, and FPtr, respectively. Figures 41, 43, and 45 displayed the
TEM images of 1 mM PC/PG with FPmn, FPdm, and FPtr, respectively. Similar to the
trend observed in the FP-LM3 samples, larger aggregates of fused vesicles are present in

the samples containing higher oligomerization states of FPS.

DISCUSSION

TEM data for LM3 and PC/PG vesicles with addition of FPmn, FPdm, and FPtr
demonstrate that higher oligomerization states of FPS are associated with greater fusion
catalyzing activity. Speciﬁcally, FPtr is the most fusogenic construct among the three
peptides. This observation provides direct evidence for the enhanced fusogenicty of
trimeric fusion peptide.

Our TEM data were obtained under the same peptide/lipid molar ratio as in the
ﬂuorescence assay described in Chapter IV and V. In particular, the TEM images for

LM3 and PC/PG at lipid concentration 0.15 mM were obtained using the same vesicle

87

and peptide solutions as used in the kinetic analysis described in Chpater V. As the
ﬂuorescence assay established that FPtr possesses the greatest fusogenicity, a similar
ﬁnding using TEM further conﬁrms the point. We could conclude that fusion peptide
trimerization plays an important role in membrane fusion.

Using electron microscopy, at least two groups have observed that fusion peptides
form large fibrillar aggregates in neutral pH PBS or saline buffered D20 where fusion of
HIV virion occurs [3,4]. However, under our fusion peptide concentration and buffer
composition, no fibrils or aggregated fusion peptides were identiﬁed. This may be
because our data were obtained at a much lower peptide concentration than the other
researchers, i.e. at 0.0015 mM or 0.01 mM fusion peptide concentration, while Nieva [4]
used 0.1 mM fusion peptide and Karamov [3] used 0.5 mM. Karamov has suggested that
fusogencitiy of fusion peptides are correlated with their potential to form long ﬁlaments
in aqueous solution. On the contrary, our results do not support the direct correlation

between fusogencity and the HIV-1 fusion peptide’s ability to form fibrillar aggregates.

88

 

Figure 30. 0.15 mM LM3 (Magniﬁcation = 40,000) 5°° "m

 

Figure 31. 1 mM LM3 (Magniﬁcation = 27,000)

89

 

Figure 33. 1 mM 1M3 + 0.01 mM FPmn (Magniﬁcation = 27,000)

90

 

/ 500 nm

Figure 34. 0.15 mM LM3 + 0.00075 mM FPdm (Magnification = 50,000)

 

500 nm

Figure 35. 1 mM LM3 + 0.005 mM FPdm (Magniﬁcation = 50,000)

91

500 nm

 

 

Figure 37. 1 mM LM3 + 0.0033 mM FPdm (Magniﬁcation = 14,000)

92

 

Figure 39. 1 mM PC/PG (Magnification = 14,000)

93

 

 

500 nm
Figure 41. 1 mM PC/PG + 0.01 mM FPmn (Magniﬁcation = 40,000)

94

 

Figure 43. 1 mM PC/PG + 0.005 mM FPdm (Magniﬁcation = 40,000)

95

500 nm

 

 

Figure 45. 1 mM PC/PG + 0.0033 mM FPtr (Magniﬁcation = 14,000)

96

REFERENCE

1. Kanaseki, T., Kawasaki, K., Murata, M., Ikeuchi, Y., and Ohnishi, S-i. (1997) J.
Cell Biol. 137, 1041-56.

2. Stegmann, T., White, J. M., Helenius, A. (1990) EMBO J. 9, 4231-41.

3. Slepushkin, V. A., Andreev, S. M., Sidorova, M. V., Melikyan, G. B., Grigoriev,
V. B., Chumakov, V. M., Grinfeldt, A. E., Manukyan, R. A., and Karamov, E. V.
(1992) AIDS Res. Hum. Retroviruses. 8, 9-18.

4. Pereira, F. B., Goni, F. M., Muga, A., and Nieva, J. L. (1997) Biophys. J. 73,
1977-86.

5. Hope, M. J., Bally, M. B., Webb, G., and Cullis, P. R. (1985) Biochim. Biophys.
Acta 812, 55-65.

97

CHAPTER VII

REDOR CHEMICAL SHIFT MEASUREMENTS TO PROBE THE SECONDARY

STRUCTURE OF MEMBRANE-BOUND HIV -1 FUSION PEPTIDE CONSTRUCTS

98

BACKGROUND

The three dimensional structure of membrane-bound fusion peptide is essential
for understanding its function in fusion catalysis. The low solubility of fusion peptide
complicates the preparation of quality X-ray crystals. Therefore, X-ray crystallography is
not an appropriate technique to probe fusion peptide structure. Solution NMR is not very
suitable, either. Although the structure of fusion peptide solublized in detergent could be
determined by solution NMR [1, 2], it is difficult to study the fusion peptide virtually
immobilized in lipid membranes using solution NMR because of the lack of rapid
molecular tumbling.

Solid state NMR has some advantages relative to the conventional techniques in
detecting atomic-resolution structure of proteins in membrane environments. Unlike X-
ray crystallography, solid state NMR does not require crystals; and unlike solution NMR,
macromolecules do not have to truncate [3] [4].

Solid state NMR employs magic angle spinning (MAS) as a major approach to
examine the structure of membrane proteins embedded in unoriented bilayer samples.
MAS is deﬁned as mechanically rotating the sample at an axis tilted at 54.7° relative to an
external magnetic ﬁeld. In these experiments, structural information can be obtained by
measuring the dipolar interaction between coupled pairs of rare-spin spin-l/2 nuclei,
because the dipolar coupling strength is inversely proportional to the cube of the
internuclear distance. Solid-state NMR techniques for heteronuclear dipolar interaction
measurements have been well developed. Examples include dipolar exchange-assisted
recoupling (DEAR) [5], transfer-echo double-resonance (TEDOR) [6], and rotational-

echo double-resonance (REDOR) [7] [8] techniques.

99

As a simple and versatile heterodipolar recoupling technique, REDOR filters the
MAS signals of selectively labeled nuclei from the background Signal of natural
abundance isotopes. In the present work, REDOR was employed to perform chemical
shift measurements which probe the secondary structure of the N-terrninal region of our
synthesized fusion peptide constructs. Based on the experimental correlation between
secondary structure and NMR chemical shift [9], we investigated the local secondary
structure in the vicinity of Phe-8 using its l3C carbonyl carbon chemical shift.

The REDOR measurements were made on the consensus 23-residue HIV -1 fusion
peptide sequence, AVGIGALFLGFLGAAGSTMGARS, and its cross-linked analogs. It
has been demonstrated that the N-terrninal and central region of the fusion peptide inserts
into the target membrane during fusion [10] [11] [12]. Our central interest is to
investigate the secondary and tertiary structure of these fusion peptide analogues in their
membrane-bound states.

In the literature, there have been conﬂicting experimental data regarding the
secondary structure of membrane-bound HIV -1 fusion peptide from CD, infrared,
solution NMR, ﬂuorescence, and ESR experiments [1, 13-23]. Evidence for both helical
and [3 structures exists. Different conformations may arise from the differences in peptide
sequence, lipid composition, sample preparation, or hydration level.

In our group, Yang et al. have reported a predominantly [3 strand structure of the
fusion peptide N-terrrrinal and central region when it is associated with membranes whose
lipid headgroup and cholesterol composition is close to that of the host cell of the HIV -1
virus [24]. Moreover, REDOR experiments [12] probing the 13C-ISN proximity between

neighboring FP strands strongly support a structural model in which oligomeric B strands

100

are held together by inter-peptide hydrogen bonding. This experiment revealed equal
population of fusion peptide in parallel and antiparallel strand arrangements. In the
parallel alignment, the oligomeric strands are approximately in-register. Also, C-terminal
fraying is detected in the parallel arrangement which is consistent with a reasonable
biophysical model in which the apolar N-terrninal and central regions of the oligomer
insert into the membrane while the polar C-terminal region is outside the membrane (cf.
Figure 49 (a)).

The gp41 soluble ectodomain structure consists of three gp41 subunits associating
as an in-register helical coiled-coil, which ends just a few residues C-terminal of the
fusion peptide. It has been postulated that at least three fusion peptides insert into the
target membrane during fusion [25, 26]. In these contexts, it is attractive that a fusion
peptide population was indeed observed with an approximately in-register parallel strand
arrangement. This arrangement seems a natural extension of the soluble ectodomain
oligomeric structure, and this arrangement also has a distinct apolar region of the fusion
peptide oligomer which could penetrate into the membrane and catalyze fusion [12].

In the present study, we mimicked the trimeric topology of the HIV-1 fusion
peptide and used solid-state NMR REDOR chemical shift measurements to probe the N-
terrninal secondary structure of individual strands in FPmn, FPdm, and FPtr embedded in
LM3 and PC/PG bilayer membranes. Our result is consistent with a model in which the
interacting fusion peptide strands adopt a B strand structure in LM3 and a helical structure
in PC/PG. We hypothesized that the enhanced fusogencity of oligomeric fusion peptides

in LM3 may be partly due to their preference for parallel strand arrangement.

101

METHODS

Solid State NMR Sample Preparation.

(1) Fusion Peptide in LUVs. Samples were typically prepared using 5 mM
HEPES buffer (pH 7) which also contained 0.01% NaN3. Between 0.05 and 0.4 umol
FPmn-F8CL9N, FPdm-F8CL9N, or FPtr-F8CL9N (cf. Figure 46) were mixed with LUVS
containg 15 ~ 40 umol LM3 or PC/PG lipids in a total volume of ~ 4 ml or 36 ml, and the
mixture were kept at room temperature overnight to ensure maximum peptide/lipid
binding. The peptide/lipid mixtures were then ultracentrifuged at 100,000g for 4 to 5
hours so that the membrane and associated peptide pelleted. The peptide/lipid pellet
formed after ultracentrifugation was transferred by spatula to a 4 mm or 6 mm diameter

magic angle spinning (MAS) NMR rotor.

FPmn AVGIGALFLGFLGAAGSTMGARSKKKW
AVGIGALFLGFLGAAGSTMGARSCKKKW
FPd m
AVGIGALFLGFLGAAGSTMGARSCKKKW
FPtr AVGIGALFLGFLGAAGSTWMGARSKKKKKKA

AVGIGALFLGFLGAAGSTWMGARSKKKKKT
AVGIGALFLGFLGAAGSTWMGARSKKKKKK

Figure 46. Amino acid sequences of FPmn-, FPdm-, and FPtr-F8CL9N. Each peptide
strand has a 13 C carbonyl carbon label at Phe-8 and an ‘5 N amide nitrogen label at Leu-9.

(2) FPmn in DPC. 0.5 umol of FPK3W was dissolved in 250 HL 5 mM pH 7

HEPES buffer. DPC (dodecylphosphocholine) powder of ~ 50 umol was added to the
peptide solution. The sample was gently vortexed was for 1 hour. Subsequently the

sample was frozen in liquid nitrogen and stored in a — 4 °C freezer.

102

Solid State NMR Spectroscopy. The rotational-echo double-resonance (REDOR)
ﬁltering technique relies on the heteronuclear dipolar interation between bonded Spin 1/2
nuclei pairs. In our labeling approach, the peptide had a 13C label at the Phe-8 carbonyl
carbon and an 15N label at the directly bonded Leu-9 amide nitrogen. In REDOR
experiments, we took two l3C spectra, one with N15 1t pulses, and one without 15N 1:

pulses (of. Figure 47).

 

So—Sl

 

 

Figure 47. REDOR subtraction. S0, S1, and So-81 represents the full spectrum, reduced
spectrum, and difference Spectrum, respectively.

In the ﬁrst spectrum (So), the 13C signal arises from both the labeled carbonyl carbon and
natural abundance 13C in the protein and lipids. When taking the second spectrum (S1),
we apply two 15N 1t pulses per MAS rotor period to the N channel. Because of the dipolar
coupling between the carbonyl carbon of Phe-8 and its directly bonded amide nitrogen of
Len-9, the 13C Signal from Phe-8 carbonyl carbon is attenuated [27], while the 13C signal

from the protein backbone and lipids will remain unchanged. If we subtract the $1 from

103

the So, the difference spectrum (So — 81) will be the signal from the Phe-8 carbonyl 13C.
In this way, the background from natural abundance 13C is ﬁltered out. The chemical shift
of the difference spectrum can then be used to assess the distribution of local secondary
structures in the vicinity of Phe-8.

There exists an experimental correlation between peptide secondary structure and
NMR chemical shift [9]: For peptide or protein carbonyl carbons, helical secondary
Structure correlates with downfield (higher ppm) chemical shifts while extended [3
structure correlates with upﬁeld (lower ppm) chemical shifts. This correlation is the basis
of our interpretation of the NMR spectra.

Experiments were done on a 9.4 T spectrometer (Varian Infinity Plus) using a
triple resonance MAS probe. Spacers were placed in a 4 mm or 6 mm diameter rotor so
that the sample was in the central 2/3 volume of the coil length, and the 13C and 15N RF
ﬁelds in the NMR probe circuit were reduced by at most 10% from their maximum
values in the rotor center. The NMR detection channel was tuned to 13C at 100.8 MHz,
the decoupling channel was tuned to 1H at 400.8 MHz, and the third channel was tuned to
15N at 40.6 MHz. Experiments were canied out using a MAS frequency of 8000 Hz and
the spinning frequency was stabilized to i 2 Hz. A single spectrum was the average of
80,000 — 120,000 scans.

NMR spectra were taken using a REDOR filter of the 13C—‘SN dipolar interaction
so that the Phe-8 carbonyl was the only signal observed in the l3C-detected REDOR
difference spectrum [7, 28]. Between 1 and 2 ms of cross-polarization at 50 kHz was
followed by a l-ms REDOR dephasing period and then direct l3C detection. A single 50

kHz 13C refocusing 1: pulse was placed at the center of the dephasing time and 1H TPPM

104

decoupling of 70 kHz (6 mm rotor) or 100 kHz (4 mm rotor) was applied during both
dephasing and detection [29]. The 13C transmitter frequency was set to 155 ppm, and the
lsN frequency was near the isotropic peptide amide resonance (115 ppm). For the S1
acquisition, the dephasing time contained a 40 kHz '5 N 1: pulse at the middle and end of
each rotor period, while the So acquisition did not contain these pulses. XY-8 phase
cycling was used for the 15N pulses [30, 31]. During the dephasing period, pulses were
not actively synchronized to the rotor phase. To obtain optimal compensation of Bo, B),
and spinning frequency drifts, So and 81 FID values were acquired alternately. The
recycle delay was 1 8.

Chemical shifts were referenced to the methylene carbon resonance of
adamantane (38.2 ppm). At room temperature and at 0 °C, l3C NIVIR signals were
attenuated, presumably because of Slow motion. Hence, most measurements were made
at —50 °C where the samples are rigid and the 1H T, is < l S. With sufﬁcient signal
averaging time, spectra can also be obtained at room temperature and are similar to those

observed at -50 °C except that some of the lines are narrower.

RESULTS

Figure 48 (a-c) are REDOR difference spectra of FPmn-, FPdm-, and FPtr-
F8CL9N associated with LM3 bilayer membrane. Figure 48 ((1) displays a REDOR
difference spectrum of FPtr-F8CL9N associated with PC/PG bilayer membrane. Figure
48 (e) displays a REDOR difference spectrum of FPKKK associated with DPC micelle.
For samples (a) — (e), the FP strand-to-lipid mol ratio were all approximately 1:100. The

(a) - (e) Spectra were each derived from the difference between the So and the 81 free-

105

 

 

 

induction decays (FID) and included 50 Hz line broadening. Because of the specific
isotopic labeling and use of the REDOR difference, natural abundance signals are ﬁltered
out and only the Signal from the labeled Phe-8 is detected in each spectrum.

For (a) — (c), the spectra are very similar and consist of a Single line centered at
171 ppm with 2.3 ~ 2.5 ppm linewidth (full-width at half-maximum (FWHM)). This
strong similarity suggests that the three membrane-bound peptides have similar structures
in the vicinity of Phe-8. According to the chemical shift correlation table [9], a carbonyl
carbon chemical shift of 171 ppm corresponds to B strand structure. These results are
consistent with previous experiment [12] where LM3-associated structure in the vicinity
of Phe-8 for FP23 and FPKKK has been demonstrated to adopt [3 strand conformation. In
Figure 48 (b), the FPdm spectrum Shows partial resolution of three components, which is
not apparent in the displayed FPmn and FPtr spectra. However, we have also observed
these components in spectra of some other FPmn samples and hence conclude that they
are not uniquely related to FPdm.

In Figure 48 (d), the REDOR spectrum of FPtr associated with PC/PG bilayer
membrane in a 1: 100 peptide strand-to-lipid ratio yields a sharp line with a chemical Shift
peak at 176 ppm, which corresponds to helical structure in the vicinity of Phe-8.

In Figure 48 (e), the REDOR spectrum of FPKKK (F8CL9N) in frozen detergent
DPC yields a sharp line with a chemical Shift peak at 176 ppm, corresponding to helical
structure. This observation is consistent with the results from solution NMR which show

that FPmn is predominantly helical in detergent [1, 2].

106

(a)

FPmn in LM3
171 ppm

(b)

FPdm in LM3
171 ppm

(C)

FPtr in LM3
171 ppm

(d)

FPtr in PC/PG
176 ppm

(9)

FPmn in DPC
176 ppm

 

190 180 via 150 ppm

Figure 48. REDOR-ﬁltered So — S. difference spectra for (a) FPmn, (b) FPdm, (c) FPtr
associated with LM3, (d) FPtr associated with PC/PG, and (e) FPmn associated with DPC
micelle. All the peptides have 13C carbonyl carbon label at Phe-8 and 15N amide nitrogen
label at Leu-9. Sample (8) contained 0.4 umol FPKKKW and LUVS made from 40 umole
LM3 lipids. Sample (b) contained 0.2 umol (FPCKKKW)2 and LUVS made from 40
mole LM3 lipids. Sample (c) contained 0.05 umol FPtr and LUVS made from 15 umole
LM3 lipids. Sample ((1) has 0.05 umol FPtr and LUVS made from 15 umole PC/PG
lipids. Sample (6) contained 0.5 mol of FPK3W and 50 umol of DPC dissolved in 250
HL 5 mM HEPES buffer. The displayed spectra are dominated by the Phe-8 carbonyl
signals.

107

DISCUSSION

We observed ~ 171 ppm l3C chemical shift for FPmn-, FPdm-, and FPtr-F8CL9N
associated with LM3 bilayer membrane and ~ 176 ppm 13C chemical shift for FPtr-
F8CL9N associated with PC/PG. The result for LM3-associated FP constructs is
consistent with a model in which the N—termini of multiple fusion peptide strands insert
into an LM3 membrane in a B strand conformation, and the result for PC/PG membranes
is consistent with a model in which the N -termini of multiple fusion peptide strands insert
into a PC/PG membrane in a helical conformation. The different secondary structures for
the FP constructs embedded in different lipid bilayers indicated that the peptides’
conformation is dependent on the membrane’s lipid and cholesterol composition.

We examined the correlation between the secondary structure of the oligomeric
fusion peptide constructs and their enhanced fusogenicity. We first considered the LM3-
associated FPS that adopt B strand conformation. Additional solid-state NMR REDOR
studies have shown that there are both parallel and antiparallel arrangements of strands
for LM3-associated FPmn [12]. For FPdm and FPtr, their C-terrninal cross-linking
topology suggests parallel strand arrangement, since the C-terminal chemical bonds
would likely enforce a parallel alignment for the multiple peptide strands. Furthermore, it
is possible that the parallel alignment is more fusogenic than the antiparallel alignment.
For the parallel alignment, the apolar N-terminal regions of two or more peptide strands
could insert into the hydrophobic interior of the membrane, which could be more
disruptive to the membrane than a single peptide strand (cf. Figure 49 (a)). This particular
insertion topology is less likely with the antiparallel arrangement in which the polar and

apolar ends of strands would be on the same side of the oligomer (cf. Figure 49 (b)).

108

Thus, FPmn, which has a rrrix of both parallel and antiparallel alignments, would be less
fusogenic than FPdm and FPtr. The actual strand arrangement of the membrane-
associated cross-linked peptides will be investigated by solid State NMR inter-strand

distance measurements and other experimental methods.

109

 

 

HIV-1 FP apolar polar
f A \ K_H
AVGI GALFLGFLGAAGSTMGARS

 

(a)

Parallel Arrangement

 

A16G ‘
A
5GALFLGFLG
5GALFLGFLG
AA
15G ‘
(b)
Antiparallel Arrangement
F 5GALFLGFLGAAWG
16(3AAGLFGLFLA5G ﬂ

Figure 49. Peptide oligomerization and membrane insertion models for (a) parallel strand
arrangement and (b) antiparallel Strand arrangement. The arrows indicate the peptide
direction from N—terminus to C-terminus while the numbers label the indices of
individual residues.

110

i
i iii???

Figure 50. Membrane insertion topology according to the “splayed helix pyramid” model.

In PC/PG where the fusion peptide adopts helical conformation, substantial
fusogenic enhancement was also observed for oligomerized fusion peptide constructs. In
order to explain this phenomenon, I suggest the “splayed helix pyramid” model
developed from inﬂuenza FHA2 fusion peptide trimerization by LeDuc et al. [32]. In this
model, three N—terminal fusion peptides form a trimer, with each strand being a helix and
inserting into the target cell membrane at an angle of ~ 65° from the bilayer normal (cf.
Figure 50). The insertion topology resembles the shape of a pyramid, with the C-termini
of the three helices in close proximity. I think that FPtr could adopt the “splayed helix
pyramid” structure in PC/PG. The insertion of three FP strands into the bilayer would
cause more membrane disruption than the insertion of one or two FP strands, and would
correlate with higher fusogenicity.

It should be noted that in the solid-state NMR experiments, the peptide structure
is observed after vesicle fusion has occurred. It is also possible that fusion requires a
transient Structure that is different from the NMR structure observed at the end state of

fusion.

11]

REFERENCE

1.

2.

10.

11.

12.

l3.

14.

15.

16.

17.

Chang, D. K., Cheng, S. F., and Chien, W. J. (1997) J. Virol. 71, 6593-602.

Chang, D. K., Cheng, S. F., and Trivedi, V. D. (1999) J. Biol. Chem. 274, 5299-
309.

Grifﬁn, R. G. (1998) Nat. Struct. Biol. NMR Suppl., 508-12.
Marassi, F. M., and Opella, S. J. (1998) Curr. Opin. Struct. Biol. 8, 640-8.

Sachleben, J. R., Frydman, V., and Frydman, L. (1996) J. Am. Chem. Soc. 118,
9786-9787.

Hing, A. W., Vega, 8., and Schaefer, J. (1992) J. Magn. Reson. 96, 205-209.
Gullion, T., and Schaefer, J. (1989) J. Magn. Reson. 81, 196-200.

Gullion, T., and Schaefer, J. (1989) in Advances in Magnetic Resonance (Warren,
W. 8., Ed.) pp 57-83, Academic Press, New York.

Wishart, D. S., Sykes, B. D., and Richards, F. M. (1991) J. Mol. Biol. 222, 311-
33.

Durrer, P., Galli, C., Hoenke, S., Corti, C., Gluck, R., Vorherr, T., and Brunner, J.
(1996) J. Biol. Chem. 271, 13417-21.

Yang, J., Parkanzky, P. D., Khunte, B. A., Canlas, C. G., Yang, R., Gabrys, C.
M., and Weliky, D. P. (2001) J. Mol. Graph. Model. 19, 129-35.

Yang, J ., and Weliky, D. P. (2003) Biochemistry 42, 11879-11890.

Kliger, Y., Aharoni, A., Rapaport, D., Jones, R, Blumenthal, R., and Shai, Y.
(1997) J. Biol. Chem. 272, 13496-505.

Slepushkin, V. A., Andreev, S. M., Sidorova, M. V., Melikyan, G. B., Grigoriev,
V. B., Chumakov, V. M., Grinfeldt, A. E., Manukyan, R. A., and Karamov, E. V.
(1992) AIDS Res. Hum. Retroviruses 8, 9-18.

Gordon, L. M., Curtain, C. C., Zhong, Y. C., Kirkpatrick, A., Mobley, P. W., and
Waring, A. J. (1992) Biochim. Biophys. Acta 1139, 257-74.

Chang, D. K., Chien, W. J., and Cheng, S. F. (1997) Eur. J. Biochem. 247, 896-
905.

Rafalski, M., Lear, J. D., and DeGrado, W. F. (1990) Biochemistry 29, 7917-22.

112

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

Pereira, F. B., Goni, F. M., Muga, A., and Nieva, J. L. (1997) Biophys. J. 73,
1977-86.

Mobley, P. W., Waring, A. J., Sherman, M. A., and Gordon, L. M. (1999)
Biochim. Biophys. Acta 1418, 1-18.

Peisajovich, S. G., Epand, R. F., Pritsker, M., Shai, Y., and Epand, R. M. (2000)
Biochemistry 39, 1826-33.

Martin, I., Defrise-Quertain, F., Decroly, E., Vandenbranden, M., Brasseur, R.,
and Ruysschaert, J. M. (1993) Biochim. Biophys. Acta 1145, 124-33.

Martin, I., Schaal, H., Scheid, A., and Ruysschaert, J. M. (1996) J. Virol. 70, 298-
304.

Taylor, S. E., Desbat, B., Blaudez, D., Jacobi, 8., Chi, L. F., Fuchs, H., and
Schwarz, G. (2000) Biophys. Chem. 87, 63-72.

Yang, J., Gabrys, C. M., and Weliky, D. P. (2001) Biochemistry 40, 8126-37.

Caffrey, M., Cai, M., Kaufman, J., Stahl, S. J., Wingfield, P. T., Covell, D. G.,
Gronenbom, A. M., and Clore, G. M. ( 1998) EMBO J. 17, 4572-84.

Weissenhom, W., Dessen, A., Harrison, S. C., Skehel, J. J., and Wiley, D. C.
(1997) Nature 387, 426-30.

Gullion, T. (1998) Cone. Magn. Reson. 10, 277-289.

Yang, J., Parkanzky, P. D., Bodner, M. 'L., Duskin, C. A., and Weliky, D. P.
(2002) J. Magn. Reson. 159, 101-110.

Bennett, A. E., Rienstra, C. M., Auger, M., Lakshmi, K. V., and Griffin, R. G.
(1995) J. Chem. Phys. 103, 6951-6958.

Gullion, T., Baker, D. B., and Conradi, M. S. (1990) J. Magn. Reson. 89, 479-
484.

Gullion, T., and Schaefer, J. (1991) J. Magn. Reson. 92, 439-442.

LeDuc, D. L., Shin, Y. K., Epand, R. F., and Epand, R. M. (2000) Biochemistry
39, 2733-9.

113

 

CHAPTER VIII

DEUTERIUM SOLID STATE NMR EVIDENCE FOR MEMBRANE BILAYER

CURVATURE INDUCED BY THE HIV-l FUSION PEPTIDE

114

BACKGROUND

Many experiments have suggested that the rate and efﬁciency of membrane
fusion is modulated by the curvature properties of these membranes [1]. In the present
work, we investigated the inﬂuence of a synthetic HIV-1 fusion peptide analog on the
phospholipid membrane morphology. We used 2H-NMR spectroscopy to probe the
interaction of these peptides with model bilayer membranes similar in composition to
those of host cells of the virus. 2H-NMR spectra of the LM3 bilayer membrane system in
the presence and absence of HIV fusion peptide were obtained to probe the Speciﬁc effect
of fusion peptide on lipid motion and structure, with the goal of relating observed
changes to the fusion process. Our results suggest that the addition of fusion peptide

promotes the formation of curvature in the lipid membranes.

1. Membrane Morphology

The curvature of biological and model membranes are determined by the bilayer
arrangement of the phospholipid. This bilayer contains two opposing monolayers (cf.
Figure 51, bilayer). In a membrane bilayer comprised of two identical monolayers, any
curvature would lead to structural instability. Hence the bilayer has no tendency to curve.
Different from this kind of bilayer curvature, each monolayer has an intrinsic curvature.
This intrinsic curvature can be defined as the shape of the monolayer when it had attained
its most stable structure, i.e. when the head groups were hydrated with water and the
terminal methyl groups were surrounded with a nonpolar solvent [2, 3]. If both
monolayers adopted the intrinsic curvature, they would bend in similar direction

(assuming that the two monolayer have similar chemical composition), leading to

115

separation of the two monolayers. (cf. intrinsic curvature in Figure 49). Thus, the intrinsic

curvature would cause instability in the bilayer structure [2].

m1 W115:
1111111 gang

BILAYER
INTRINSIC

§Wyé CURVATURE
3 C:
ﬂﬁﬁ

HEXAGONAL PHASE

 

Figure 51 [2]. (Left) Diagrammatic sketch of phospholipid arrangements. The bilayer is
the arrangement of phospholipids commonly found in biological and model membranes.
(Right) The hexagonal phase (Hn).

Generally there are two types of intrinsic monolayer curvature strain: negative
curvature Strain and positive curvature strain (cf. Figure 51, 52). When the membrane
lipids have smaller cross-sectional area at the headgroup and larger cross-sectional area at
the acyl chain, the membrane tends to have negative curvature Strain. On the other hand,
positive curvature strain occurs when the membrane lipids have large, well-hydrated, or
repelling head groups and a smaller cross section at the acyl chain.

Hydrophobic substances such as hydrocarbons, diacyglycerols [2], and some viral
fusion peptides can promote negative curvature strain. When the negative monolayer
curvature rises to certain extent, the bilayer will convert to the hexagonal (Hn) phase.
The left panel of Figure 51 displays the cross-section of a single hexagonal phase

cylinder. This cylinder has a hydrophobic outlayer and is surrounded by other identical

ll6

 

cylinders in a hexagonal packing array with the structure extending perpendicular to the
plane of the page (cf. Figure 51 right panel).

Positive curvature strain could be enhanced by detergents. When positive
monolayer curvature strain becomes too large, the bilayer membrane forms micelles (cf.

Figure 52).

,1; .-

_“‘ .- x1. ‘ - g . Z
‘ _/“H:1-’Ir" L.."\“*-.
.r .- 1: ‘:
I'd/d‘ I C“ ‘ 1‘"
r.» 2 -,: ‘-
* _1 1 .
.5] ,—
:4 v:
:‘I I")
‘. {a
,- .4 :g" "f. 6.: '»— w
7 A I. > , ’ ", ‘ L<T “ > 1 ,
“/ILI'J. .3112. 51:?- Lr‘lxl) .-' i (
» -. ,- v ,. 1.; , c- -.

Positive Curvature Strain Micelle

Figure 52. [3] Diagrammatic Sketch of positive curvature strain (left) and micelle (right).

It has been suggested that viral fusion peptides promote membrane fusion by
inducing negative curvature strain in the membranes [4]. Studies on inﬂuenza fusion
peptide [5, 6], SIV fusion peptide [7, 8] and HIV fusion peptide [9] have demonstrated
correlation between fusogenicity of these peptides and their ability to promote the
formation of inverted hexagonal phase (H11), which has a high degree of negative
curvature [3].

These observations are consistent with the “stalk/TMC” fusion model [10, 11]. In
this model, membrane fusion begins with the “stalk” intermediate, in which the
contacting (cis) monolayers acquire negative curvature (cf. Figure 53). It has been
proposed that fusion peptide may promote the formation of “Stalk” intermediate by

inducing negative curvature [1]. From “stalk”, the membrane will proceed through a

117

series of morphological rearrangements that will lead to fusion pore formation.

 

 

 

 

 

 

g
1.6.

 

 

stolk

 

 

Figure 53 [10, 11]. Scheme for membrane fusion and inverted phase formation. (A) Two
juxtaposed bilayers join to form the stalk, hemifusion intermediate (B), followed by a
transmembrane contact (TMC) intermediate (C), which will lead to fusion pore formation
during membrane fusion.

In our work, we studied the inﬂuence of synthetic HIV-l fusion peptides on
model bilayer membranes whose lipid headgroup and cholesterol composition are similar

to that of the host cell of HIV-1 virus. The membrane structure and dynamics of the

membrane lipids are probed by 2H-NMR.

2. 2H-NMR for Fusion Peptide Effect on Membrane Structure & Dynamics
2.1. 2I1! Pake Doublet Powder Spectrum

Local structural information on ﬂuid bilayer membranes could be obtained from
the 2H-NMR Pake powder pattern spectra [12] [13], which are usually acquired using a
“Solid Echo” NMR pulse sequence which minimizes distortions in spectrum [14]. Figure

54 displays a Pake doublet powder spectrum:

118

05
l 1»
Bo {- 2w(0))..)

 

 

0-0 °

 

 

F I I l l I 1 r —1

-10.0 -7.5 -5.0 -2.5 0 2.5 5.0 7.5 10.0

Figure 54. The experimental Pake doublet 2H-NMR powder Spectrum for DPPC-d2
deuterated in the (ii-position of the phosphocholine (PC) polar head group. It is obtained
as the Fourier transform of the quadrupolar echo signal [15].

119

 

Figure 55. Model for the bilayer lipid membrane in the external magnetic field B0. 11 is
the bilayer local membrane normal. Each carbon position is numbered as 1, 2, 3,

For ﬂuid membranes, the 2H—NMR spectrum of a deuteron on a C-ZH bond of the
membrane lipid is a doublet with a splitting 203(0) [15], where 0 (cf. Figure 55) is the
angle between the external magnetic field and the bilayer normal. On the bilayer, the
phospholipid molecules rotate rapidly about the surface normal, such that:

(0(9) = In, SCD(3c0820-1)/2| (1)
where 8Q (125 kHz) is V2 times the largest (0:00) NMR splitting in a sample containing
rigid C-ZH, and SCD is the orientational order parameter defined by:

SCD = < ( 3005213 ‘1)/ 2 > fast motions (2)
< > stands for the average over fast motion, and B is the angle between the C-2H bond and

the bilayer normal. SCD varies between 0 and 1. These extrema correspond to rapid

120

 

isotropic motion and no motion, respectively. In Figure 54, the separation of peaks
corresponding to 0 = 90° in the powder spectrum yields the magnitude of the order
parameter I SCD I . According to equation 1:

SCI): 0)(90°) / 62.5 kHz (3)

The shape of the Pake doublet spectrum indicates the ﬂuidity of the lipid
membrane [15]. A smaller splitting in the Pake powder pattern (13cc I) is a sign of
more conformational freedom in the bilayer membranes. The observation of a Pake
doublet spectrum with I SCD I < 0.5 for a methylene deuteron on an acyl chain in a lipid
bilayer is associated with the type of rapid axially symmetric motion characteristic of a
ﬂuid bilayer [15]. These motions include: (1) Transitions between different molecular
conformations (often referred to as ‘tranS-gauche isomerizaton’ in the case of acyl
chains); and (2) rotations of the molecules around the bilayer normal.

In our experiment, the Pake powder spectra of side-chain 2H-labeled phospholipid
(DMPC-d54 cf. Figure 56) bilayer membranes are obtained in the presence and absence
of HIV -1 fusion peptide. The 2H quadrupolar splittings in the Pake patterns are compared

to study the fusion peptide effect on different regions of the lipid bilayer

O§C Dz Dz Dz Dz Dz Dz_

/ .
O H\\\0 2222222

11 ,
>NA/ °‘f‘°\)\/ “CW

/+ .0 0% 02 132 132 02 Dz 02

Figure 56. 2-Dimyristoyl-d54-sn-Glycero-3-Phosphocholine (DMPC-d54).

121

2.2. Relaxation-time Measurements

In NMR spectroscopy, the classical motion of the nuclear moment it in an applied
magnetic field Bo can be understood as a rotation of the direction of u in a cone with its
axis along Bo. This is called Larmor precession [16], whose angular velocity is — 118., (y,
magnetogyric ratio, is a constant for a given nucleus). The corresponding Larmor
frequency 000, is

(1)0 = Iy/Zrt I B0 (4)

Any system with a magnetic moment will give Larmor precession. For an NMR
sample containing many identical molecules, each having one magnetic nucleus, the total
magnetic moment M of the sample is the vector sum of the individual nuclear moments
in:

M=Zu (”

Therefore, the total magnetic moment M will be the resultant of the Larmor
precession of each nuclear magnetic moment.

In addition to Larmor precession, the total magnetic moment M is also affected by
the internal realignment of individual nuclear spins with the magnetic ﬁeld Bo. When the
sample is placed in B0, its magnetization changes from 0 to M0, the thermal equilibrium
value. The process for the magnetic moment to approach thermal equilibrium is referred
to as relaxation. In Bloch equations that describe motion of the magnetic moment [17],
the components of M decay to M0 exponentially, with the components of M parallel (M2)
and perpendicular (Mx, My) to M0 approaching their equilibrium values Mo and 0 with
different time constants T1 and T2.

Mz(t)-Mz(0)=[Mz(0)-Mo]eXp(-t/T 1) (6)

122

Mr(t) = Mx(0)exP(-t/T2), My(t) = My(0)eXp(-t/T 2) (7)

T1 and T2 are relaxation times [17]. T1 describes the energy ﬂow between the
nuclear spin system and the other degrees of freedom of the system know as ‘lattice’, and
is therefore named the spin-lattice relaxation time. It is also named the longitudinal
relaxation time. T2 describes the relaxation of MK and My to 0. T2 relaxation is caused by
the direct interaction between the spins of different nuclei without energy transferred to
the lattice, and is thus named the spin-spin relaxation time, or transverse relaxation time.

The T1 relaxation processes are sensitive to relatively fast dynamical molecular
motions characteristic of correlation times 1:1 5 (00 ‘1 [18], where (00 may lie between a

few MHz and about 500 MHz, depending on the magnetic field and nucleus being
studied [15]. The fast motions that contribute to T1 include transitions between different

molecular conformations and rotations of the molecules around the bilayer normal [15].
T2 is very sensitive to slow motions with correlation time 132 >> 000 ‘1 [18]. The slow

motions that contribute to T2 in are interpreted as associated with molecular diffusion
along curved membrane surfaces [18].

Relaxation-time measurements provide information on thermally driven
molecular and membrane motions. In our experiments, 2H spin-lattice relaxation (T1) and
spin-spin relaxation times (T2) are determined at different peptide:lipid ratios to

investigate the fusion peptide effect on the dynamic properties of membrane lipids.

MATERIALS AND METHODS
Materials. Dimyristoyl-sn-glycero-3-phosphocholine (DMPC), l-palmitoyl-2-

oleoyl-sn-glycero-3-phosphocholine (POPC), 1-palmitoyl-2-oleoyl-sn-glycero—3-

123

phosphoethanolamine (POPE), 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-L-serine]
(POPS), phosphatidylinositol (PI), sphingomyelin, and cholesterol were purchased from
Avanti Polar Lipids, Inc. (Alabaster, AL). The Micro BCATM protein assay was obtained
from Pierce (Rockford, IL). N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid
(HEPES) was obtained from Sigma. All other reagents were analytical grade.

Peptides. FP23 peptides corresponding to the 23 N-terrrrinal residues
(AVGIGALFLGFLGAAGSTMGARS) of the LAVla strain of HIV-1 gp41 were
synthesized as their C-terrninal amides using a peptide synthesizer (ABI 431A, Foster
City, CA) equipped for FMOC chemistry. A similar FP23KKK peptide was synthesized
by combining the FP23 sequence with three C-terrninal lysines. A 26 amino acid
polypeptide melittin (GIGAVLKVLTTGLPALISWIKRKRQQ) is also synthesized.
Melittin represents about 50 % of bee venom and its major function is to cause
dissolution or destruction of natural and artiﬁcial lipid membranes [19].

Lipid Preparation. Samples were prepared using a lipid/cholesterol mixture which
reﬂects the approximate lipid and cholesterol content of host cells of the HIV-1 virus
[20]. The lipid mixture, denoted “LM3-DMPCdac” had DMPC, POPE, POPS,
sphingomyelin, PI and cholesterol in a 10:5:2:2:1210 mole ratio. DMPC was
perdeuterated along the acyl sidechains. Another mixture, denoted “LM3” had POPC,
POPE, POPS, sphingomyelin, PI and cholesterol in a lO:5:2:2:1:10 mole ratio was also
prepared. Lipid and cholesterol powders were dissolved together in chloroform. The
chloroform was removed under a stream of nitrogen followed by overnight vacuum
pumping. Lipid dispersions were formed by addition of water or buffer containing 0.01%

NaN3 followed by homogenization with freeze-thaw cycles. Large unilamellar vesicles

124

(LUVS) of 100 nm diameter were prepared by extrusion [21]. In this approach, lipid
dispersions were extruded ~ 30 times through two stacked 0.1 pm polycarbonate ﬁlters
(Avestin, Inc., Ottawa, ON, Canada).

Static NMR Sample Preparation. Samples were prepared in two different ways.
Method (1): Peptide and lipid (either dispersion or 100 nm diameter LUV) were mixed in
36 mL of 5 mM HEPES buffer (pH 7.0) which also contained 0.03% NaN3. The mixtures
were kept at room temperature overnight and then ultracentrifuged at 130,000 x g for
four hours at 15 °C to pellet the peptide-lipid complex. Using the BCA assay, it was
shown that the peptide binds quantitatively to lipid under these conditions and unbound
peptide does not pellet [22]. The peptide/lipid pellet formed after ultracentrifugation was
transferred by spatula to a NMR tube. Method (2): Peptide and lipid were mixed in 0.5 -
1 mL of aqueous solution and kept at room temperature overnight. Peptide-lipid complex
was then transferred to an NMR tube. The peptide binding to lipid was nearly
quantitative under these conditions [22].

Static NMR Experiment. Experiments were done on a 9.4 T spectrometer (Varian
VXR or Varian Inﬁnity Plus, Palo Alto, CA) using a Varian Single resonance probe. The
NMR detection channel was tuned to 2H. The RF ﬁeld was about 40 kHz. The
quadrupolar echo sequence (rt/2) - 131 - (1r/2)- t2 - acquire [23] was used to minimize
effects from probe ring-down. The phase of the ﬁrst 1t/2 pulse was x and the phase of the

second rr/2 pulse alternated between y and —y. Each 11/2 pulse was 6.7 us. I] was 50 us,

and 12 was 40 us. The recycle delay was 1 s.

125

 

 

T1 Measurement. To determine the 2H T1, a (rt)x — d1 inversion-recovery module
[23] was inserted before the echo sequence, and the variation of the acquired signal was
measured as a function of d]. In the pulse sequence:

1! — d1 — (rt/2) — tl — (1t/2)— t2 — acquire
a it pulse inverts the magnetization from M20 to —M,o. This is followed by spin-lattice
relaxation during d1, when the magnetization tends towards its equilibrium value M20
according to equation 6. The NMR signal is acquired after the quadrupolar echo
sequence. The signal intensity is plotted as a function of recovery time (11. Figure 57
displays an array of lipid mixture 2H-NMR spectra with varied (11 from 0 to 0.4 8. At
short recovery time (11 the Signal will appear negative. At long times, full recovery is

obtained. At some intermediate time the signal will be zero.

 

   

l; i 1*

TT+~LH «IL-H L ikauLuvauuLuél-nwtvuih JLJLJL

1

Figure 57. An array of lipid mixture 2H-NMR spectra with varied d2 from 0 to 0.4 s.

 

1;);‘
l
l

126

 

 

T1 was determined by fitting the inversion-recovery data to the equation:

I(d.) =1i + AI(1 — e‘dl’“) (8)

where I(d1) is the measured echo intensity, and I, (the initial echo intensity), Al (the
difference between the equilibrium and initial echo intensities), and T1 (the longitudinal
relaxation time) are ﬁtting parameters.

T2 Measurement. Taking the natural log of both sides of equation 7 yields:

lnMy=lnMyo — t /T 2 (9)
Using the solid echo pulse sequence: (rt/2)—tl —(1t/2) —‘t2 - acquire, the decay of the
acquired signal was measured as a function of synchronous increment of 1:1 and 1.2. T2 can
be determined by plotting the natural log of peak height (lnI) as a function of the total
echo time, 21:, where ‘L' is the time from the end of the second m2 pulse to the echo
formation. The data were ﬁt with the equation:

ln[I(2‘t)] = ln[I(0)] — Z‘t/T 2 (10)
where 101) is the measured echo intensity, and 1(0) (the initial signal intensity) and T2

(the transverse relaxation time) are fitting parameters.

RESULTS AND DISCUSSION

Static Lipid NMR Spectra. Figure 58 displays 2H NMR spectra of various static
samples containing lipid or lipid plus peptide. All spectra were taken at 35 °C. In (a — e),
the samples contained LM3-DMPCdac dispersions with FP23zlipid mole ratios of (a) 0,
(b) 1:33, and (c) 1:10, and melittinzlipid mole ratios of (d) 1:33 and (e) 1:10. The (a), (c —
e) samples were made by method (2) and the (b) sample was made by method (1). A

sample with a composition identical to the (b) sample but made by method (2) gave a

127

 

 

 

spectrum similar to the displayed (b) spectrum (data not shown). Thus, the spectra appear
to be approximately independent of whether sample preparation was done by method (1)
or (2). In addition, the 31P NMR spectra of the (a - c) samples are similar to those of LM3
samples, indicating that lipid phase is independent of whether the choline lipid is DMPC
or POPC [24].

Each 2H spectrum in (a — e) is a superposition of individual Pake powder patterns
of the 2H along the DMPC acyl chain. Each methylene position experiences a different
amount of motion and thus has a different quadrupolar splitting. The motion increases
and the quadrupolar splitting decreases as one moves from the phospholipid headgroup to
the tail. As displayed in (a — c), there were minor changes in the 2H NMR spectra with
addition of fusion peptide. Relative to a pure LM3-DMPCdac sample, there appear to be
some small increases in quadrupolar splittings with bound fusion peptide which are in the
range of l — 3% for the 1:33 sample and 2 - 5% for the 1:10 sample. This is consistent
with slightly less chain motion in the presence of peptide. In addition, the features that
define the sharp “horns” of individual Pake patterns are broader in the presence of

peptide, which is consistent with a more heterogeneous distribution of lipid motion.

128

 

‘ l K I‘D.
1 /
\ / ‘a‘
r “iv Ii I.

 

 

 

 

 

 

HIIIHHEEIIH
50 0 -50kHz
(b) ‘
J1;
lilllllllTlllll
50 0 -50kHZ
(C) #
ill
H
r} I“
H. J
lat/'r‘
l
I l
lTll:ll]lll.[ll
50 0 ~50kHz
(d) (9)
Irfrrilrrrirr TIIWTIIITWIIY
50 0 -50 kHz 50 O -50 kHz

Figure 58. 2H spectra of unoriented samples. All spectra were taken at 35°C. The 2H
spectra of an LM3-DMPCdac sample is presented in (a). Similar spectra are presented for

129

 

 

 

FP23:LM3-DMPCdac at (b) 1:33 and (c) 1:10 peptide lipid ratios and for melittinzLM3-
DMPCdac at (d) 1:33 and (e) 1:10 peptide:lipid ratios. The (a) and (c)-(e) samples were
made by method (2) and (b) sample was made by method (1). Each (a)-(e) spectrum was
processed with 150 Hz line broadening and is the sum of 4000-6000 scans.

In large contrast, melittin induces a narrow isotropic 2H feature which has ~ 200
Hz full-width at half-maximum WIN) linewidth and which is shifted about 300 Hz
upﬁeld of the 2H20 resonance. This feature is likely associated with formation of micelles
which undergo fast isotropic reorientation.

The (a — e) spectra indicate that the HIV -1 fusion peptide and melittin have very
different effects on the same lipid distribution, which reﬂects the different biological
function of the two peptides. The role of fusion peptide is to induce viral/target cell
membrane fusion. It is essential for the viral and target cell membrane to remain intact
during fusion pore formation so that the cell could maintain its biological function after
fusion. Therefore, the fusion peptide should not disrupt the equilibrium structure of the
lipid membrane. Our NMR results indicate that in the presence of fusion peptide the
membranes remain as a bilayer with slightly less chain motion and a more heterogeneous
distribution of lipid motion. On the other hand, one of the major biological functions of
melittin is to cause cell lysis (dissolution or destruction of cell membrane)[l9], which is
consistent with the observed micellization of lipids in the presence of melittin. Other
investigators have also observed this effect of melittin in different lipid systems [19].

2H T, and T2 measurements. Measurements were made on LM3-DMPCdac
dispersions containing no fusion peptide, and for dispersions containing 1:33, and 1:10

FP23zlipid ratios. The 1:33 sample was made by method (2) and the pure dispersion and

1:10 samples were made by method (1). For each of our spectra, we measured the

130

 

 

 

intensities of “outer” and “inner” features, as is visually shown in the Figure 59 (a). The
intensity of the outer feature is mostly due to the 2H close to the headgroups while the
inner feature is mostly due to the 2H in the terminal methyl group. The outer feature
intensity was measured relative to the spectrum baseline while the inner feature intensity
was measured relative to the approximate baseline intensity of the non—methyl deuteron
transitions.

Figure 59 (b) displays the outer feature inversion-recovery data for the three
samples. The best-ﬁt relaxation times for these and other data sets are presented in Table
4. Within our ﬁtting uncertainties, there is no difference between the outer feature T1
values of the different samples. There are also not signiﬁcant differences between the
inner feature T1 values of the different samples. The relative uncertainties of the inner
feature T‘ values are greater than those of the outer feature values, in part because of the
difficulty of determining the inner feature baseline.

Figure 59 (c) displays plots of the outer feature data for the three different
samples and the best-ﬁt T2 values are presented in Table 4. There appears to be a 10 —
15% reduction in the outer feature T2 with addition of fusion peptide. There is probably
not a signiﬁcant difference between the outer feature T2 values of the 1:10 and 1:33
samples. The inner feature intensity data are qualitatively the same for the three samples.
However, for these data, a plot of ln[I(21:)] vs. 2': is non-linear, probably because of
inaccuracy in baseline intensity subtraction. Thus, it is not reasonable to fit the inner

feature data to equation 10.

131

 

 

 

 

 

 

 

 

 

I f l Iinner
I like. } l. f" l
I ‘1’ val .........
l !
louter l! I
u f I
.J/f ”v
,‘jkaWm—mAJaJ/ﬂ \“MfM/«WV'KJ‘W
I ‘ T r l
80 0 -80 kHz
(b)
6 ‘ r r F T I «
A x X 6 a U Q Q 9 Q Q
m o
a. 3 — 25 ‘
.o
5’ a
92‘
2 0 ‘ 9 _
£2
E
-3 —
T l l I
0 0 1 0.2 O 3
Time (s)
(c)
‘ ‘5 \' l l l
"'c.
E 1.0 - ~32“ ‘
5 ‘gl‘ﬂ
£5 0.5~ ‘Yf'o ‘
s; 31.50..
g 0.0“ \%~.‘_V°«; ‘
c ‘31. '9.
Q) x.
E x. ° .
I: '0.5-‘ V \x ‘0“ _
E '\
it,
-‘l.0 ' I ‘ I ‘ T I
O 200 400 600
Time (us)

Figure 59. 2H relaxation data. (a) Measurements of outer and inner intensities. Ioum is
measured between the horns of the outer Pake pattern and the spectrum baseline. Iinner is
measured between the horns of the inner Pake pattern and an approximate baseline from

132

 

the outer Pake pattern. (b) T1 inversion recovery data of the outer Pake pattern. The pure
LM3 (circles) and 1:10 FP23:LM3 (crosses) intensities are plotted as a function of d1.
There is little difference between LM3, 1:10 FP23:LM3 and 1:33 FP23:LM3 (data not
shown). (c) T2 data of the outer Pake pattern. The pure LM3 (circles), 1:33 FPZ3zLM3
(triangles) and 1:10 FPZ3zLM3 (crosses) data are plotted as a function of 21. In addition,
the best-ﬁts are shown for the LM3 and 1:10 FP23:LM3 samples. Addition of the peptide
causes a 10 — 15% reduction in T2.

Table 4. Lipid T1 and T2 Relaxation Times for Various Peptidezlipid Ratios

 

 

 

 

Nucleus Peptide: Lipid Mixture T1 (s) # T2 (us) #
Lipid Ratio*
Outer 0.053 (0.001) 327 (10)
0 LM3-DMPCdac
Inner 0.34 (0.01) 794 (53)
Outer 0.050 (0.002) 268 (5)
2H 1:33 LM3-DMPCdac

 

Inner 0.32 (0.06) 794 (84)

 

Outer 0.051 (0.002) 294 (10)
l: 10 LM3 -DMPCdac

 

Inner 0.31 (0.06) 709 (50)

 

 

 

 

 

 

 

 

* All fusion peptide in these samples was FP23KKK.

# Relaxation time measurements reported as value (error).

133

CONCLUSION

We applied deuterium NMR to hydrated lipid membranes with bound HIV-1
fusion peptide. The lipid headgroup composition reﬂected the composition of the host
cells of the virus. Phosphocholine lipid deuterated along its acyl chains was incorporated
into the membranes. Our 2H-NMR spectra are consistent with a predominant bilayer
structure in the presence of bound fusion peptide and this differs from the work of other
investigators who observed a signiﬁcant component of isotropic phase [25]. This
difference is likely due to the different lipid compositions used in the various studies.

The broadening of the “horns” of the individual Pake patterns in our 2H spectra
are consistent with a more heterogeneous distribution of lipid motion. The increase in
quadrupolar splittings is consistent with a slight decrease in the motion of the acyl chains,
possibly due to the insertion of the peptide into the membrane.

We noticed no significant change in the 2H T1 relaxation times with the addition
of fusion peptide, however the 2H and T2 times decreased with the addition of fusion
peptide. T1 is sensitive to fast molecular motions including transitions between different
molecular conformations and rotation around the bilayer normal. Our data for T1 suggest
that insertion of fusion peptide into the bilayer membrane does not signiﬁcantly affect
these molecular motions.

We interpret the decrease in T2 as a result of more curvature in the bilayer
membrane induced by insertion of fusion peptide. T2 is strongly inﬂuenced by diffusion
of phospholipid molecules along the membrane surface [18]. At our experimental
temperature, the phospholipids could diffuse along the membrane surface. During

diffusion, the 2H spin on the C-ZH bond in the phospholipid acyl chain will experience

134

 

some ﬂuctuation in the quadrupolar field due to variation of the C-ZH bond angle with
respect to the external magnetic field Bo (cf. Figure 55). Fluctuations in the quadrupolar
ﬁeld are essential for relaxation [26]. If the phospholipid was moving along a planar
surface, the angle between the C-ZH bond and the external field Bo would remain
approximately the same during diffusion. As a result, there would be little ﬂuctuation in
the quadrupolar ﬁeld experienced by the 2H spin, and therefore, slower relaxation. When
the phospholipid moves along a curved surface, the angle between the C-ZH bond with
the external field Bo will change, causing greater ﬂuctuation in the quadrupolar ﬁeld
experienced by the 2H spin. Thus, there is a correlation between the curvature in the
membrane and ﬁeld ﬂuctuation, which might then explain the faster relaxation and
shorter T2 in the presence of fusion peptide.

Taken together, these lipid NMR results support a model in which the peptide acts
as a fusion catalyst by inducing membrane curvature and reducing the energies of high-

energy, highly curved fusion intermediates.

135

 

REFERENCE

1.

2.

10.

ll.

12.

l3.

14.

15.

16.

17.

18.

Epand, R. M. E., R. F. (2000) Biopolymers 55, 358-363.

Epand, R. M., Shai, Y., and Segrest, J. P. (1995) Biopolymers (Pept. Sci.) 37, 319-
338.

Epand, R. M. (1998) Biochim. Biophys. Acta 1376, 353-368.
Epand, R. M. (1997) Biochem. Soc. Trans. 25 ( 3 ), 1073-1079.

Epand, R. M. a. E., R. F. (1994) Biochem. Biophys. Res. Commun. 202, 1420-
1425.

Colotto, A. E., R. M. (1997) Biochemistry 36, 7644-7651.

Epand, R. F., Martin, I., Ruysschaert, J.-M. and Epand, R. M. (1994) Biochem.
Biophys. Res. Commun. 205, 1938-1943.

Colotto, A., Martin, I., Ruysschaert, J.-M., Sen, A., Hui, S. W. and Epand, RM.
(1996) Biochemistry 35, 980-989.

Peisajovich, S. G., Epand, R. F., Pritsker, M., Shai, Y., and Epand, R. M. (2000)
Biochemistry 39, 1826-33.

Siegel, D. P. (1993) Biophys. J. 65, 2124-2140.
Siegel, D. P. a. E., R. M. (1997) Biophys. J. 73, 3089-3111.
Davis, J. H. (1983) Biochim. Biophys. Acta 737, 117-171.

Schmidt-Rohr, K., and Spiess, H. W. (1994) Multidimensional Solid-state NMR
And Polymers, Academic Press Limited, San Diego.

Davis, J. H., Jeffrey, K. R., Bloom, M., Valic, M. I., and I-Iiggs, T. P. (1976)
Chem. Phys. Lett. 42, 390-394.

Bloom, M., Evans, E., and Mouritsen, O. (1991) Q. Rev. Biophys. 24, 293-397.

Harris, R. K. (1983) Nuclear Magnetic Resonance Spectroscopy, a
Physicalchemical View, Pitman Books Limited. London, UK. Section 1-6.

Harris, R. K. (1983) Nuclear Magnetic Resonance Spectroscopy, a
Physicochemical View, Pitman Books Limited, London, UK. Section 3-2.

Bloom, M., and Stemin, E. (1987) Biochemistry 26, 2101-2105.

136

19.

20.

21.

22.

23.

24.

25.

26.

Dufourc, E. J ., Smith, I. C., and Dufourcq, J. (1986) Biochemistry 25, 6448-55.

Aloia, R. C., Tian, H., and Jensen, F. C. (1993) Proc. Natl. Acad. Sci. U. S. A. 90,
5181-5.

Hope, M. J., Bally, M. B., Webb, G., and Cullis, P. R. (1985) Biochim. Biophys.
Acta 812, 55-65.

Yang, J ., Gabrys, C. M., and Weliky, D. P. (2001) Biochemistry 40, 8126-37.

Harris, R. K. (1983) Nuclear Magnetic Resonance Spectroscopy, a
Physicochemical View, Pitman Books Limited, London, UK. Section 3-12.

Wasniewski, C., Canlas, C. G., and Weliky, D. P. Unpublished experiments.

Schanck, A., Peuvot, J., and Brasseur, R. (1998) Biochem. Biophys. Res.
Commun. 250, 12-14.

Harris, R. K. (1983) Nuclear Magnetic Resonance Spectroscopy, a
Physicochemical View, Pitman Books Limited, London, UK. Section 3-15.

137

CHAPTER IX

BINDING ASSAY

138

BACKGROUND

For fusion assay and solid-state NMR structural measurements, it is important to
examine the binding strength of peptide to lipid. In fusion assays, our aim was to assess
the correlation between numbers of strand in each peptide constructs and fusogenicity. In
order to make a fair comparison, we need to determine whether all the peptide constructs
bind quantitatively, or at least comparably to membranes. In solid-state NMR
measurements, the NMR signals are observed for fusion peptide in all types of
environment, i.e. membrane-bound or free. Since we are particularly interested in the
structure of membrane-bound peptide, the contribution from the free peptide must be
eliminated. Also, when we used 2H NMR to probe the change of membrane morphology
as a function of fusion peptide to lipid ratio, we need to verify quantitative binding of the
fusion peptide with lipid membranes. Otherwise, the putative peptide to lipid ratio would
not reﬂect their actual values.

For solid-state NMR samples, Yang et al. [1] demonstrated that there is strong
binding between FPmn and LM3 bilayer membrane. Yang et al. estimated the percentage
of binding by comparing the initial peptide concentration to that found in the supernatant
after addition of lipid and subsequent centrifugation. The BCA assay was used to
determine peptide concentration. During preparation of FPmn (FF or FPKKK) NMR
samples at 10 uM peptide and 1 mM lipid, BCA data showed little peptide in the
centrifugation supernatant [2]. As a control experiment, BCA assay measurement showed
that in the absence of LM3, FPmn did not pellet under centrifugation. Taken together,

these observations suggest quantitative association of FPmn with LM3.

139

For FPdm, BCA assay measurements also showed that it did not pellet under the
ultracentrifugation conditions used to make the solid-state NMR samples in the absence
of LM3 [3].

We did not directly test the binding affinity of FPdm and FPtr to LM3 and PC/PG
membranes. However, there was approximately the same NMR signal per scan for the
FPmn, FPdm, and FPtr sample (cf. Figure 48, Chapter VII), which is consistent with
comparable (quantitative) membrane binding for the three peptide constructs.

For samples made under the conditions of the fusion assay (1.5 uM FPmn or 0.75
M FPdm and 150 uM lipid), peptide binding was not accurately measured because of
large background signals in the BCA assay from HEPES and uncentrifuged lipid. In this
chapter, we describe the attempts to assess the binding strength of FPmn, FPdm, and FPtr

to LM3 and PC/PG under fusion assay conditions.

METHODS

Binding Assessments Based on Centricon Separation of Free Peptide from
Membrane-bound Peptide. In this experiment, we measured the binding of 1.5 uM FPmn
(FPGCK6W) solution to 150 uM LM3, which are similar to the concentrations used in
the fusion assays described in Chapter IV and V.

Solutions containing both PP and LM3 vesicles and reference solutions
containing either LM3 vesicles or FP were prepared. We denoted the mixture as FP+L,
and the two references as L and FF, respectively. These solutionswere centrifuged on a
Sorvall Ultracentrifuge (Newtown, CT) with a SW25.1 rotor (100,000 x g, 300 minutes).

Supematants were collected and lyophilized. The lyophilized supernatant for FP+L, L,

140

and FF were re-dissolved in 1.2 ml H20, and their 280-nm absorbance were measured.
Next, the 1.2 m1 solutions of FP+L, L, and FP were each placed in the top compartments
of Centricons YM-lO (Millipore, MA) and were centrifuged on a Sorvall RC-SB
centrifuge (Newton, CT) with a GSA rotor (14,000 x g, 300 — 420 minutes). Because the
Centricon YM-10 is equipped with a membrane ﬁlter between the top and bottom
compartments and is designed to sieve any particles greater than 10,000 Dalton, FPmn
bound with LM3 vesicles or LM3 vesicles will not pass through the ﬁlter, since the mass
of an LM3 vesicle (~ 100,000 lipid molecules per vesicle) is ~ 42,000,000 Dalton
whereas unbound FPmn (M.W.~ 3000 g/mol) will pass through the ﬁlter. Therefore,
separation of the membrane-bound peptide from unbound peptide can be achieved.

The 280-nm absorbance of the Centricon ﬁltrate from FP+L, L, and FF were
recorded and denoted as A(FP+L), Am, and Amp), respectively. The percentage of unbound
peptide is calculated with the equation:

Percent of Unbound FPmn = [A(pp+L) — A(L)]/ A(Fp) (1)
In this calculation, the percentage of unbound peptide is determined by dividing the
difference of 280-nm absorbance of FP+L and L ﬁltrate by the absorbance of pure FP
ﬁltrate. The reason to subtract the absorbance of L from FP+L is to correct for the
background resulting from the lipids which has passed through the Centricon ﬁlter.

Binding Assessments Based on Fluorescence Energy Transfer. FPmn, FPdm, and
FPtr all contain a Trp residue at the C-terminus of each fusion peptide strand. Their
amino acid sequences are displayed in Figure 60. Fluorescence resonance energy transfer
between the tryptophan (donor) and ﬂuorescently labeled phospholipid (acceptor) was

used to characterize the binding of the peptides with lipid vesicles. Binding of the

141

F Pmn AVGIGALFLGFLGAAGSTMGARSKKKKKKW

AVGIGALFLGFLGAAGSTMGARSCKKKKKKW
F Pdm l

AVGIGALFLGFLGAAGSTMGARSCKKKKKKW

FPtr AVGIGALFLGFLGAAGSTMGARSWKKKKKKA

AVGIGALFLGFLGAAGSTMGARSWKKKKKK
AVGIGALFLGFLGAAGSTMGARSWKKKKKK

Figure 60. Amino acid sequences of FF constructs in the binding assessments.

peptides to the NED-labeled vesicles puts the donor and acceptor in proximity such that
excitation of the T rp in the peptide leads to an increase in the emission of the NBD-lipid.

In this experiment, LM3 and PC/PG vesicles were prepared. The LM-3 mixture
had POPC, POPE, POPS, sphingomyelin, PI and cholesterol in a 10:5:2:2:1:10 mole
ratio. The PC/PG mixture had POPC and POPG in a 4:1 mole ratio. In both LM3 and
PC/PG lipid mixture, 2 mol % of N-NBD-PE lipids were added. Lipid and cholesterol
powders were dissolved together in chloroform. The chloroform was removed under a
stream of nitrogen followed by overnight vacuum pumping. Lipid dispersions were then
formed with addition of 5 mM pH 7 HEPES buffer. After homogenization of the
dispersion with ten freeze-thaw cycles, LUVS were prepared by extrusion through a
polycarbonate ﬁlter with 100 nm diameter pores [4]

Fluorescence was recorded using 4 nm bandwidth using an Instruments S. A.
Fluoromax-2 (Edison, NJ) spectroﬂuorimeter operating at excitation and emission
wavelengths of 290 nm and 530 nm, respectively. A siliconized glass cuvette was used
with continuous stirring in a thermostated cuvette holder. Measurements were carried out

at 37 °C with ~ 2 ml of 150 M LUV in H2O. Peptide solution (FP strand concentration =

142

100 uM) was titrated into the liposome solution and the change in ﬂuorescence of the

sample was monitored during the titration.

RESULTS
Binding Assessments Based on C entricon Separation of Membrane-bound Peptide
from Unbound Peptide. FPmn (FPGCK6W) was mixed with LM3 vesicles in 12 ml H2O.

The concentration of FPmn and LM3 were 1.5 and 150 “M, respectively. Two reference

solutions were prepared: 1.5 uM FPmn solution and 150 M LM3 solution. The sample
and references were denoted as FP+L, FP, and L, respectively. For simplicity, all three
solutions were made in H2O instead of 5 mM pH 7 HEPES buffer. The ultracentrifuged
supernatant of FP+L, FP, and L were lyophilized and re-dissolved in 1.2 ml H2O. The
280-nm absorbance for the re-dissolved FP+L, L, and FF were 1.075, 1.060, and 0.101,
respectively. The large absorbance of FP+L and L was due to light scattering from
unpelleted lipids. Each of three solutions was placed in the t0p compartment of Centricon
YM-10 and was centrifuged under a spinning force of 14,000 x g for 4 — 6 hours until
most of the top solution went through the Centricon membrane into the bottom
compartment of the Centricon. The 280-nm absorbance of the ﬁltrates of FP+L, L, and
FP, were 0.061, 0.052, and 0.053, respectively. Therefore we calculated the percentage of
unbound peptide according equation 1:

Percent of Unbound FPmn = (0.061 — 0.052)/0.053 = 17 %

The experiment was repeated and the 280-nm absorbances of the filtrate of FP+L,
L, and PP were 0.048, 0.049, and 0.045, respectively. The percentage of unbound FPmn

= (0.048 - 0.049)/0.045 ~ 0 %. These results are consistent with approximately

143

quantitative binding of FPmn with LM3 bilayer membrane under the fusion assay
conditions.

Since the result could be confounded by the possibility that the lipids that have
already adhered to the Centricon ﬁlter may block unbound peptide, we did two control
experiments. First, we re-centrifuged the FP ﬁltrate (280-nm absorbance = 0.053) through
the Centricon which had been used for centrifuging FP+L under similar spinning
condition. The 280-nm absorbance of the re-centrifuged ﬁltrate was 0.032. Comparing to
the initial absorbance of 0.053, the result indicated that ~ 60 % (0032/0053) of free
fusion peptide could pass through a Centricon contaminated with LM3 lipids. The other
control experiment we did was to ﬁlter an FP solution with a 280-nm absorbance of 0.124
through a clean Centricon ﬁlter under similar centrifuging condition. The absorbance of
the ﬁltrate was 0.076, which means only ~ 60 % (0076/0124) of PP passed through the
clean ﬁlter as well. Therefore, we concluded that LM3 lipids adhered to the Centricon
ﬁlter would not change the ﬁlter’s capability to trap unbound FP compared to a clean
Centricon ﬁlter. In either case ~ 40% of free FP will be hold by the ﬁlter.

This approach was only successful for assessing membrane binding strength of
FPmn. It was not suitable for measuring lipid afﬁnity of FPdm and FPtr. Due to the
oligomeric FPs’ larger size and possibly higher afﬁnity with the Centricon membrane,
they could not pass through the centricon ﬁlter in the absence of lipids. As a result,
separation of free oligomeric fusion peptide from membrane-bound peptide was not
successful.

Binding Assessments Based on Fluorescence Energy Transfer. Titration of FPmn,

FPdm, and FPtr to LM3 and PC/PG vesicles containing 2 mol % N—NBD-PE as

144

ﬂuorescent acceptor resulted in an increase in the NBD fluorescence emission [5]. The
fluorescence increase is interpreted as the result of energy transfer from the tryptophan
(donor) on the peptides to the NBD (acceptor) on the lipid. Following each titration, the
NBD ﬂuorescence increased and then leveled off after 200 - 1500 seconds (depending on
the different peptide used). This was recorded and the average ﬂuorescence intensity in
the platform was calculated. For each peptide, the average value of stabilized
ﬂuorescence intensity after each titration was plotted as a function of FP strand
concentration. The proﬁles of FPmn, FPdm, and FPtr associated with LM3 and PC/PG
are displayed in Figure 61 and 62, respectively. The ﬂuorescence proﬁles for FPmn,
FPdm, and FPtr are nearly identical in Figure 61, which indicates similar LM3 binding
afﬁnity for the three peptide constructs. In Figure 62, the patterns of ﬂuorescence
increase are also very similar for FPmn, FPdm, and FPtr, with the ﬂuorescence of FPdm
somewhat higher than that of FPmn and FPtr, suggesting the binding affinity of FPdm to
PC/PG may be a little higher than the other two peptides. The ﬂuorescence curve for FPtr
is slightly higher than that of FPmn in Figure 62 (a). However, the FPtr proﬁle is almost
identical to that of FPmn in Figure 62 (b). Therefore we could not conclude that the

binding strength of FPtr to PC/PG is higher than that of FPmn to PC/PG.

145

 

 

OW
GD

Fluorescence (a.u.)

"MM???”

6

0
n

 

 

I I I I ﬂ

l I I I
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
-6
PP strand conc. (10 M)

s5
a:

Figure 61. Titration of FPmn, FPdm, and FPtr with LM3/NBD-PE LUV (98:2 mol %) as
monitored by ﬂuorescence. Fluorescence emission were measured for LUV at 150 M in
5 mM pH 7 HEPES buffer with increasing amounts of added peptide, and the
ﬂuorescence intensities in the range of 526 — 534 nm were integrated and plotted as a
function of fusion peptide strand concentration. Fluorescence data were obtained in two
consecutive experiments. The average values of the two measurements were used in the

plot. x, FPmn; A, FPdm; 0, FPtr.

146

 

 

a
( ) 40000- A
35000
A
A 'l
3' 30000-
31 l A 0
25000-4 O
.. A x
20000- 0
‘ A x
II 0
15000- 8
‘ x
10000— Q
I I I I I I T
0.0 0.5 1.0 1.5 2.0 2.5 3.0
-6
PP strand conc. (10 M)
(D) 45000-
A
40000-
A
35000«
=2
59, 30000- A o
5 25000- A 6
§ 20000- A 6
II G
150004 8
10000 a

 

 

I I I I

0.0 0.5 1'0 1'5 2'0 2.5 3.0
FP strand conc. (10'6 M)

Figure 62. Titration of FPmn, FPdm, and FPtr with PCPG/NBD-PE LUV (98:2 mol %)
as monitored by ﬂuorescence obtained under similar condition in two sets of experiment
(a, b). Fluorescence emission was measured for LUV at 150 M in 5 mM pH 7 HEPES
buffer with increasing amounts of added peptide, and the ﬂuorescence intensities in the
range of 526 - 534 nm were integrated and plotted as a function of fusion peptide strand

concentration. X, FPmn; A, FPdm; 0, FPtr.

147

DISCUSSION

With Centricon separation of free peptide from membrane-bound ones, we have
determined that the binding of FPmn (FPGCK6W) to LM3 is 83 — 100% under fusion
assay conditions. Based on ﬂuorescence resonance energy transfer from Trp residue in
the membrane-bound peptide to NED-labeled lipids, we determined that the binding
strength of FPmn, FPdm, and FPtr to LM3 bilayer membrane are approximately equal.
The PC/PG lipid afﬁnity of FPdm seemed to be slightly higher than that of FPmn and
FPtr.

It is possible that the overall concentration of cross-linked fusion peptide in LM3
could be higher than that of FPmn because of higher lipid afﬁnity of the cross-linked
peptide, and this higher afﬁnity could be an additional reason for the enhanced
fusogenicty of the oligomeric fusion peptides, in addition to the higher local peptide
concentration achieved with cross-linking. Our observations in this chapter point against
different afﬁnities of the different peptides to LM3. For PC/PG, FPdm has a little higher
afﬁnity to PC/PG lipids than FPmn or FPtr. However, the fusion activities are ordered
FPtr > FPdm > FPmn so there is not a direct conclusion between binding affinity and
fusogenicity. Interestingly, the afﬁnity of the peptides for PC/PG and LM3 appear to be
about the same, and may be due to the comparable hydrophobicity and charges of PC/PG
and LM3 bilayers.

In related studies, other investigators have studied the lipid binding and
fusogenicities of peptides that contained between 16 and 70 of the N-terminal residues of
gp41 [6, 7]. Longer peptides had higher fusogenicities, but at 0.1 uM peptide

concentration there were negligible differences in the lipid afﬁnities of the different

148

peptides. These data are similar to the different fusogenicities and comparable binding

that we observed for FPmn, FPdm, and FPtr.

149

REFERENCE

1.

2.

Yang, J., Gabrys, C. M., and Weliky, D. P. (2001) Biochemistry 40, 8126-37.

Yang, J., Prorok, M., Castellino, F. J ., and Weliky, D. P. (2004) Biophys. J. 87,
1951-1963.

Yang, R., Yang, J., and Weliky, D. P. (2003) Biochemistry 42, 3527-3535.

Hope, M. J., Bally, M. B., Webb, G., and Cullis, P. R. (1985) Biochim. Biophys.
Acta 812, 55-65.

Lau, W. L., Ege, D. S., Lear, J. D., Hammer, D. A., and DeGrado, W. F. (2004)
Biophys. J. 86, 272-284.

Peisajovich, S. G., Epand, R. F., Pritsker, M., Shai, Y., and Epand, R. M. (2000)
Biochemistry 39, 1826-33.

Sackett, K., and Shai, Y. (2002) Biochemistry 41, 4678-85.

150

CHAPTER X

SUMMARY AND WORK IN PROGRESS

151

The overall goal of this research is to understand some aspects of the HIV-1
fusion peptide-induced viral/target cell membrane fusion. My speciﬁc aim is to study the
effect of fusion peptide oligomerization on its fusion catalyzing activity. The motivation
of my project is to test the hypothesis that during fusion at least three fusion peptides
insert into the target cell membrane with their C-termini in close proximity and that
fusion peptide trimerization may be a structural requirement for HIV-1 viral/target cell
fusion. Evidences for the signiﬁcance of trimerization include high-resolution structure of
trimerized viral envelope proteins [1, 2], ﬁndings of multiple trimers and corresponding
high local fusion peptide concentration at the fusion site [3-5], and mutational studies [6].
Despite the importance of fusion peptide oligomerization, there is not a high-resolution
structure to date for the fusion peptide domain in its membrane-inserted form because (1)
its hydrophobicity impairs the high protein solubility required by X-ray crystallography
and (2) the fast molecular tumbling required by solution NMR could not be satisﬁed due
to immobilization of the fusion peptide by membrane lipids. Therefore we thought it
would be interesting to study the membrane-insertion structure of HIV-1 fusion peptide
as a trimer with three fusion peptide strands chemically cross-linked at their C-termini.
We also thought it would be interesting to compare the fusogenicity of the trimeric fusion
peptide with previously studied monomeric fusion peptide. These goals were achieved
through peptide design/synthesis, analytical ultracentrifugation, fusion assay, electron
microscopy, and solid-state NMR techniques.

In our approach, a series of monomeric and oligomeric peptides have been
synthesized and are derived from the region of the HIV envelope glycoprotein that

mediates membrane fusion during infection. Three major products are denoted as FPmn,

152

FPdm, and FPtr. Each contains the original fusion peptide sequence plus 3-6 extra lysines
at the C-terminus to increase solubility and inhibit self-association of the peptide
constructs. FPmn was synthesized with an automatic peptide synthesizer. FPdm was
formed by cross-linking two monomeric fusion peptides through C-terminal cystine di-
sulﬁde bond. An in-register trimer, FPtr, was created by cross-linking three monomeric
fusion peptide through C-terminal lysine side-chains. Because fusion peptides contain
many bulky, nonpolar amino acid residues, they tend to self-associate and form
aggregates in aqueous solution. Since our aim was to study the correlation between
numbers of strand in each construct with their relative fusogenicity, it is crucial to know
that these peptide constructs do not aggregate in aqueous solution prior to interaction
with the membrane. Analytical ultracentrifugation results demonstrated that at the peptide
concentrations similar to those of the stock solution of fusion assays, the three peptide
constructs’ experimental ﬁtted masses were close to their theoretical monomer masses.
This indicated that there was no appreciable higher order aggregation in the assay buffer.

Fusogenicity of FPmn, FPdm, and FPtr were compared in a fusion assay in which
the increase in NBD ﬂuorescence due to lipid mixing induced by the fusion peptide
constructs was monitored as a function of time. In the two vesicle systems (LM3 and
PC/PG) that we have used, the long time changes of ﬂuorescence were ordered AF" >
Ade > Aan, which indicated that oligomeric fusion peptides induced more vesicle
fusion.

Because an increased fusion rate is likely the most important fusion peptide effect
on viral/target cell fusion, the kinetics of the ﬂuorescence data were analyzed to

determine the variation of fusion rate with different constructs. In a single data set, the

153

initial ﬂuorescence was modeled by a dominant, fast exponential build-up representing
the lipid mixing induced by the initial interaction of fusion peptide with membranes. At
longer times, there was an additional contribution from a slower build-up whose origin
has not yet been understood. For both LM3 and PC/PG fusion, the fast ﬂuorescence
build-up rates were ordered ku>kdm>kmm with ku~40kmn for LM3 and kn~15kmu for
PC/PG. These ﬁndings suggest that the cross-linked peptides induce both a greater ﬁnal
extent and a more rapid rate of fusion than their non-cross-linked analogs.

We also studied the temperature dependence of fusion rate. Fluorescence data
were obtained at ﬁve temperatures between 25 ~ 40°C and Ink was plotted over 1/1‘
according to the Arrhenius equation. From the Arrhenius plot, activation energies, Ba, and
pre-exponential factors, A were calculated for LM3 and PC/PG vesicle fusion induced by
FPmn, FPdm, and FPtr. The trends in the data were: (1) E3." (LM3) ~ Ea-" (PC/PG), lnA"
(LM3) ~ lnAtr (PC/PG); and (2) for PC/PG vesicle fusion, Ea-" < Ea-dm, Eann & At, < Adm
Am. We did not obtain the E1, and A values for LM3 fusion induced by FPmn and FPdm
since the Arrhenius plots for these two cases were not linear.

In addition to the fusion assay, images from transmission electron microscope
also provided evidence for a strong correlation between C-terminal cross-linking and
enhanced fusogenicity. In order to establish a structure-function relationship for the
oligomeric fusion peptides, we used solid-state NMR REDOR chemical shift
measurements to determine the secondary structure of FPmn, FPdm, and FPtr in their
membrane-inserted forms. Previous solid-state NMR results in this group have
demonstrated that in LM3 at peptide/lipid ratio 2 0.010, the N-terminal and central

regions of the peptide adopt a nonhelical B strand structure [7]. In Figure 48 (a-c) of

154

Chapter VII, the 1-D solid-state NMR spectra for the Phe-8 carbonyl are very similar for
LM3-associated FPmn, FPdm, and FPtr, which suggests that FPdm and FPtr also forms a
B strand at this residue. Additional solid-state NMR REDOR studies have shown that
there is ~ 50% fusion peptide population of LM3-associated FPmn which adopts an
approximately in-register parallel strand arrangement [8]. In the contexts that cross-
linked fusion peptides are signiﬁcantly more fusogenic than their non-cross-linked
analogs, this ﬁnding is interesting because the in-register parallel strand arrangement is
likely to be a reasonable extension of the known ectodomain oligomeric structure and this
arrangement also has a distinct apolar region of the fusion peptide oligomer which could
insert into the membrane and catalyze fusion [8]. The other 50 % of FPmn associated
with LM3 was found to adopt antiparallel strand arrangement. It does not seem clear to us
how the oligomer could insert into the membrane with this type of strand arrangement.
Therefore, LM3-associated FPdm and FPtr with their in-register parallel B strand
conformation enforced by their C-terminal cross-linking would be more fusogenic than
FPmn, which has a mix of both parallel and antiparallel alignment.

For PC/PG-associated fusion peptides, Figure 48 (d) indicates that PC/PG-
associated FPtr adopts helical structure. The “splayed helix pyramid” model may explain
the enhanced fusogenicity for the helical-structured fusion peptide trimer [9].

In addition to studying the secondary structure of membrane-bound fusion peptide
by solid-state NMR chemical shift measurement, we applied 2H solid-state NMR to probe
the fusion peptide effect on 2H labled membrane lipids. Our results are consistent with a
model in which the addition of fusion peptide promotes the formation of curvature in the

lipid membranes.

155

In summary, the present research demonstrated a strong correlation between C-
terminal cross-linking and enhanced fusogenicity of the HIV -1 fusion peptide. The fusion
enhancement effect of trimeric HIV-1 fusion peptides may be due to the following
reasons:

(1) Placement of apolar N-terrninal regions on one end of the oligomer may
facilitate membrane insertation.

(2) Larger apolar volume relative to FPmn and the consequent greater membrane
disruption and fusion rate.

(3) Larger free energy released upon membrane binding of multiple FP strands.

Our observations using various experimental techniques are consistent with the
hypothesis that trimerization of fusion peptide may be a structural requirement for
viral/target cell membrane fusion. The topology achieved through C-terminal cross-
linking should be similar to the fusion peptide topology thought to exist in the fusogenic
form of HIV-1 gp41.

Because it is possible to obtain FPtr in ~ 1 umol quantities and because it has
negligible self-association in aqueous solution, it should be possible to study its structural
and motional properties in aqueous, detergent, and membrane environments with a
variety of biophysical methods. These studies should provide further insight into its
enhanced fusion rate. The synthetic approach should also be applicable to fusion peptides
from other viruses and perhaps to other membrane-associated peptides derived from
proteins of known oligomeric stoichiometry.

There are many remaining questions including the actual strand arrangement of

FPdm and FPtr associated with LM3 membranes. When we tried to correlate the peptide

156

constructs’ membrane-bound structure with their enhanced fusogenicity, we hypothesized
that both FPdm and FPtr would prefer parallel strand arrangement in 1M3. This
assumption is based on the C-terminal cross-linking topology of FPdm and FPtr, in which
the covalent bond would be likely to enforce a parallel arrangement for the multiple
peptide strands. However, we have not yet obtained experimental evidence for this
assumption. Work in progress in this group involves solid-state NMR measurements of
the inter-strand distance for FPdm and FPtr. One possible approach would be to form the
FPdm by cross-linking equal amount of two types of fusion peptide monomer: one has
[13C] carbonyl label at Phe 8, and the other has [”N] amide label at Phe 8. If the two
peptides strands adopt an in-register parallel B arrangement in FPdm, the hydrogen
bonding between the two strands can be detected using 13C - 15N distance measurement.
For FPtr, a radio frequency-driven dipolar recoupling (RFDR) experiment [10] will be
employed to probe the inter-strand distance based on homonuclear dipolar coupling

between 13 C labeled carbonyl carbon on each strand.

157

REFERENCE

1.

10.

Weissenhom, W., Dessen, A., Harrison, S. C., Skehel, J. J., and Wiley, D. C.
(1997) Nature 387, 426-30.

Caffrey, M., Cai, M., Kaufman, J., Stahl, S. J., Wingﬁeld, P. T., Covell, D. G.,
Gronenbom, A. M., and Clore, G. M. (1998) EMBO J. I 7, 4572-84.

Munoz-Barroso, I., Durell, S., Sakaguchi, K., Appella, E., and Blumenthal, R.
(1998) J. Cell Biol. 140, 315-23.

Bentz, J. (2000) Biophys. J. 78, 227-45.
Bentz, J. (2000) Biophys. J. 78, 886-900.

Freed, E. O., Delwart, E. L., Buchschacher, G. L., Jr., and Panganiban, A. T.
(1992) Proc. Natl. Acad. Sci. U. S. A. 89, 70-4.

Yang, J ., Gabrys, C. M., and Weliky, D. P. (2001) Biochemistry 40, 8126-37.
Yang, J., and Weliky, D. P. (2003) Biochemistry 42, 11879-11890.

LeDuc, D. L., Shin, Y. K., Epand, R. F., and Epand, R. M. (2000) Biochemistry
39, 2733-9.

Bennett, A. E., 0k, J. H., Griffin, R. G., and Vega, 8. (1992) J. Chem. Phys. 96,
8624-8627.

158

 

  

    

    

L RRRRR

llllll

5075

N TEUN

llllllllllllllllllllllllljlll

30