SOLID-STATE NUCLEAR MAGNETIC RESONANCE STUDIES OF THE STRUCTURES,
MEMBRANE LOCATIONS, CHOLESTEROL CONTACT, AND MEMBRANE MOTIONS
OF MEMBRANE-ASSOCIATED HIV FUSION PEPTIDE (HFP)
By
Lihui Jia

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Chemistry—Doctor of Philosophy
2017

ABSTRACT
SOLID-STATE NUCLEAR MAGNETIC RESONANCE STUDIES OF THE STRUCTURES,
MEMBRANE LOCATIONS, CHOLESTEROL CONTACT, AND MEMBRANE MOTIONS
OF MEMBRANE-ASSOCIATED HIV FUSION PEPTIDE (HFP)
By
Lihui Jia
Membrane fusion is the key step during HIV viral entry to cells, and the process is catalyzed by
HIV membrane fusion protein gp41. HFP is the ~25-residue N-terminal domain of gp41 and is
required for membrane fusion with significant decreases in fusion activity with point mutations.
Both viral and host cell membrane contain ~30mol % cholesterol (CHOL), and HFP induced
fusion is faster in membrane with CHOL. However, how HFP interacts with membrane lipids
and CHOL is unknown. In this thesis, we used the newly developed

13

C-2H Rotational Echo

Double Resonance (REDOR) solid-state NMR method to study the membrane location of HFP
in chemically-native membrane environment.
HFP is 13CO labeled at specific residue, and the membrane is deuterated at specific regions of the
membrane using selective regions deuterated phosphatidylcholine (PC) and CHOL. We study
HFP wild type, HFP_V2E and L9R mutants because these two mutants eliminate and decrease
fusion respectively. HFP is predominantly β sheet structure in bilayer membrane for both HFP
wild type and HFP_V2E mutant, HFP_L9R has a different structure and is likely helical. Both
HFP and HFP_V2E mutant have major deeply-inserted membrane location contacting membrane
center and minor shallowly-inserted membrane location contacting half way of one membrane
leaflet. The HFP_V2E mutant has bigger fraction of molecules with shallower membrane
location, which is consistent with the strong correlation between membrane location insertion
depth and the peptide fusogenicity. HFP_L9R mutant has majorly deeply inserted into membrane.

By comparing the HFP- PC and HFP- CHOL contact, there is preferential contact between HFP
and CHOL vs PC at several residues including G5, G10 and G16. The free energy difference for
contacting PC vs CHOL is ~ 0.57(5) kcal.mol-1 for T= 300K. HFP- CHOL contact geometry is
successfully modeled by Swiss Dock and YASARA energy minimization with two strands
antiparallel HFP (1→16/16→1 registry). There are two energetically favorable binding models
between HFP and CHOL, from docking, energy minimization and consistency with REDOR
results. The contact models reveal tilted and curved-up tail orientation of Chol_d7. Fusion may
be catalyzed by matching the curvature of lipids contacting HFPs with the membrane curvature
during the fusion intermediates like the stalk.
Membrane motion perturbation by HFP is studied by static deuterium NMR from deuterium
powder pattern spectrum, order parameter profile and T2 relaxation time. The DMPC-d54
spectrum becomes ~10% narrower in membrane without CHOL with 4% HFP and in membrane
with 33% CHOL with 1% HFP. Accordingly, the order parameter of lipid acyl chain becomes ~
1-10% disordered by HFP. However, the spectrum becomes 20% broader in membrane with 33%
CHOL with 4% HFP, and the order parameter of lipid acyl chain becomes ~ 20- 30% ordered by
HFP. With HFP at 37 °C, DMPC-d54 T2 decreases ~ 70 %, and the CHOL T2 decreases ~ 30%.
T2 reduction is probably associated with increased membrane curvature induced by HFP. With
greater membrane curvature, the C-D bond will experience more orientation diversity relative to
the external magnetic field. Thus, the quadrupolar field will have greater change, leading to
faster relaxation and shorter T2.
Gp41_V2E mutant eliminates cell-cell fusion. Our CD spectroscopy studies show that the
FPHM_V2E mutant is helical and the melting temperature is above 90 °C in 10mM Tris buffer +
0.2 % SDS at pH 7.4. Protein is trimer and induces no lipid mixing in PC: CHOL= 2:1 vesicles.

Copyright by
LIHUI JIA
2017

Dedicated to Xinliang Ding and Joshua J. Ding

v

ACKNOWLEDGEMENTS

First of all, I would like to thank my PhD advisor, Dr. David P. Weliky for his tremendous
guidance, patience and encouragement through my graduate life. Dr. Weliky has been very nice
and always supportive when I have questions and concerns. He always encourages me to ask
questions and teaches me to solve problems and become a scientist. I am very grateful for his
support and help.
I would like to acknowledge my committee members, Prof. John L. McCracken, Prof. A. Daniel
Jones and Prof. Benjamin G. Levine for their time, generous help and guidance through my PhD
life, which is very valuable for my research and development to a scientist. Prof. A. Daniel Jones
has also provided valuable suggestions to the peptide and protein purification and
characterization in my research. He also gives me helpful suggestions for my interview
preparations and future career development. I really appreciate that.
I would like to acknowledge MSU Mass Spectrometry Facility, Max. T. Rogers NMR Facility
for their assistance to my research. I thank Dr. Daniel Holmes and Dr. Li Xie for trouble
shooting the NMR instruments and willingness to devote time and effort to help.
I am also very grateful for the help and support from all the Dr. Weliky group members, Dr.
Charles Gabrys, Dr. Li Xie, Dr. Ujjayini Ghosh, Dr. Punsisi Ratnayake, Dr. Koyeli banerjee,
Shuang Liang, Robert Wolfe and Ahinsa Ranaweera. They have been very helpful in my
research, and very friendly in many other aspects.
At the end, I would like to thank all my friends for their support and best wishes. I thank my
family for their encouragement and unconditional love.

vi

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................................... x
LIST OF FIGURES ..................................................................................................................... xiii
KEY TO ABBREVIATIONS ................................................................................................... xxvii
Chapter 1 - Introduction .................................................................................................................. 1
1.1 NMR Introduction ........................................................................................................... 1
1.1.1 Zeeman interaction................................................................................................. 2
1.1.2 Radio frequency (RF) B1 field interaction ............................................................. 4
1.1.3 Rotating frame ....................................................................................................... 6
1.1.4 Chemical shift interaction ...................................................................................... 6
1.1.5 J coupling interaction ............................................................................................. 9
1.1.6 Dipolar coupling interaction ................................................................................ 10
1.1.7 Quadrupolar coupling interaction ........................................................................ 12
1.1.8 Magic angle spinning (MAS)............................................................................... 17
1.1.9 Rotational echo double resonance (REDOR) ...................................................... 20
1.1.10 Quadrupolar echo (QUECHO) .......................................................................... 27
1.2 HIV Introduction ............................................................................................................. 28
1.2.1 HIV virus and infection ....................................................................................... 28
1.2.2 HIV gp41 ............................................................................................................. 31
1.2.3 HIV fusion peptide (HFP) .................................................................................... 35
REFERENCES ............................................................................................................................. 38
Chapter 2 - Materials and methods ............................................................................................... 45
2.1 Materials ......................................................................................................................... 45
2.2 Peptide sequences, preparation and purification ............................................................. 45
2.3 Peptide associated membrane sample preparation for MAS and static solid state NMR 46
2.4 Solid state NMR.............................................................................................................. 48
2.4.1 MAS solid state NMR spectroscopy .................................................................... 48
2.4.2 Static solid state NMR spectroscopy ................................................................... 50
REFERENCES ............................................................................................................................. 52
Chapter 3 - Structure and Membrane Location Studies of HIV Fusion Peptide (HFP) and KALP
Peptide........................................................................................................................................... 55
3.1 Introduction ..................................................................................................................... 55
3.2 Results ............................................................................................................................. 59
3.2.1 Fitting of the 13C-2H REDOR data....................................................................... 59
3.2.2 Effect of sample preparation methods on 13C–2H REDOR △S/S0 ..................... 63
3.2.3 Effect of temperature on experimental △S/S0..................................................... 66
3.2.4 Effect of membrane charge on experimental △S/S0 ........................................... 67
3.2.5 HFP location in membrane without CHOL studied by 13C–2H REDOR............. 71

vii

3.2.6 HFP location in membrane with CHOL studied by 13C–2H REDOR.................. 76
3.2.6.1 HFP location in DPPC: DPPG: CHOL membrane studied by 13C–2H
REDOR ................................................................................................................. 76
3.2.6.2 HFP location in POPC: POPG: CHOL membrane studied by 13C–2H
REDOR ................................................................................................................. 82
3.2.6.3 HFP location in DOPC: DOPG: CHOL membrane studied by 13C–2H
REDOR ................................................................................................................. 86
3.2.7 HFPV2E 13C–2H REDOR results in membrane without CHOL ......................... 89
3.2.8 HFPL9R results in membrane without CHOL..................................................... 94
3.2.9 KALP results in membrane without CHOL....................................................... 101
REFERENCES ........................................................................................................................... 109
Chapter 4 - Preferential Contacts of HFP with CHOL vs PC lipid ............................................ 114
4.1 Introduction ................................................................................................................... 114
4.2 Results ........................................................................................................................... 116
4.2.1 Experimental 13C–2H REDOR results for HFP_G5, G10 and G16c ................. 116
4.2.2 Free energy of preferential contact of HFP with CHOL vs PC ......................... 124
4.2.3 CHOL binding to two strands antiparallel HFPs predicted by Swiss Dock and
two sets of REDOR experimental distance constraints .............................................. 130
4.2.4 Favorable CHOL binding geometry on the concave surface of an HFP–antiβ
sheet ............................................................................................................................ 139
REFERENCES ........................................................................................................................... 144
Chapter 5 - HFP Effect on Membrane Motion by 2H-NMR Studies .......................................... 150
5.1 Introduction ................................................................................................................. 150
5.2 Experimental conditions ............................................................................................... 152
5.3 Results ........................................................................................................................... 153
5.3.1 Solid echo or quecho experimental results for DMPC/DMPG membrane without
and with CHOL/ HFP ................................................................................................. 153
5.3.1.1 2H - NMR spectra features for DMPC-d54 and Chol_d6 ....................... 153
5.3.1.2 DMPC-d54 Segmental order parameters at 37 °C .................................. 159
5.3.1.3 Transverse relaxation studies of DMPC-d54 and Chol_d6 .................... 168
5.3.2 CHOL motion in POPC/POPG/CHOL membrane without and with HFP..... 184
5.3.2.1 Transverse relaxation studies of Chol_d6 ............................................... 184
5.3.2.2 Transverse relaxation studies of Chol_d7 ............................................... 190
5.3.2.3 Spin lattice relaxation studies of Chol_d7 ............................................... 196
5.4 Discussion ..................................................................................................................... 204
5.4.1 HFP disrupts lipid acyl chain packing in membrane both without and with CHOL
..................................................................................................................................... 205
5.4.2 Transverse relaxation studies of PC and CHOL ................................................ 209
5.4.3 Spin lattice relaxation studies of Chol_d7 ......................................................... 211
REFERENCES ........................................................................................................................... 212
Chapter 6 - Summary and Future work....................................................................................... 215
6.1 Summary of 13C-2H REDOR as a probe for membrane location study and future work of
HFP ..................................................................................................................................... 215
6.2 Summary of membrane motions perturbed by HFP and future work ........................... 217
viii

REFERENCES ........................................................................................................................... 219
APPENDICES ............................................................................................................................ 221
APPENDIX A Preparation and characterization of FPHM_V2E mutant of gp41 ectodomain
............................................................................................................................................. 222
APPENDIX B NMR file locations ..................................................................................... 234
APPENDIX C Data for fitting ............................................................................................ 245
REFERENCES ........................................................................................................................... 265

ix

LIST OF TABLES

Table 3.1 Best-fit exponential buildup parameters for HFP_G5c in membrane a………………71
Table 3.2 Best-fit exponential buildup parameters for HFP_L12c in membrane a……………...75
Table 3.3 Best-fit exponential buildup parameters for HFP_G5c and G16c in membrane with
Chol_d7 and Chol_d6 a…………..................................................................................................81
Table 3.4 Best-fit exponential buildup parameters for HFP_G10c and G16c in POPC: POPG
membrane with Chol_d7 and Chol_d6 a........................................................................................86
Table 3.5 Best-fit exponential buildup parameters for HFP_V2E_G5c in membrane without
CHOLa...........................................................................................................................................93
Table 3.6 Best-fit exponential buildup parameters for HFP_L9R in membrane without CHOL
a
....................................................................................................................................................100
Table 3.7 Best-fit exponential buildup parameters for KALP in membrane without CHOL
a
....................................................................................................................................................107
Table 4.1 Best-fit exponential buildup parameters for HFP in membrane with CHOL a...........123
Table 4.2 Alternative SIMPSON simulation results a.................................................................124
Table 4.3 C values for different peptide and membrane labeling a.............................................127
Table 4.4 Fractional probabilities of 13CO making contact to PC or CHOL with two molecules
contact model a.............................................................................................................................128
Table 4.5 △GPC-Chol values (the energy difference for peptide binding to PC vs CHOL) for
different samples..........................................................................................................................129
Table 4.6 Dockings of CHOL with HFP–antiβ resulting from different protocols....................133
Table 4.7 Energies of CHOL dockings meeting at least two experimental distance constraints
within 5Å a...................................................................................................................................135
Table 4.8 CHOL dockings meeting at least one experimental distance constraints within 5 Å on
the concave surface of an HFP-antiβ sheet..................................................................................140
Table 5.1 The HFP effects on DMPC-d54 best-fit 2H T2 (s) in membrane without and with
CHOL by quecho experiments. The fitting errors are in parenthesis.........................................183

x

Table 5.2 HFP effects on best-fit 2H T2 (s) of Chol_d6 and DMPC-d54 in membrane with 33%
CHOL studied by quecho experiment at 37 °C and pH 7.4. The fitting errors are in parenthesis.
The ratio is HFP: phospholipids mole ratio. NA means not applied. .........................................184
Table 5.3 HFP effects on best-fit Chol_d6 2H T2 (s) of in POPC: POPG membrane with 33%
Chol_d6 studied by quecho experiment at 37 °C and pH 7.4. The fitting errors are in parenthesis.
We got the best fit T2 by fitting the echo intensity. The ratio is HFP: phospholipids mole
ratio..............................................................................................................................................188
Table 5.4 Best-fit Chol_d7 2H T2 (s) values at different temperatures without and with HFP,
uncertainties are in parenthesis. T2 was fitted with ln (CD3 peak intensity) vs 2......................196
Table 5.5 Best-fit Chol_d7 CD3 T1 (ms) values at different temperatures, uncertainties are in
parenthesis. T1 was fitted with the CD3 peak intensity vs 1.......................................................204
Table 5.6 Peak splitting for CD3 and CD2 of DMPC-d54 powder pattern spectrum at
37 °C............................................................................................................................................207
Table A1 Mean residue molar elipticity of FPHM_V2E at 222nm (θ222) (deg.cm2.dmol-1.residue1
) at different temperatures……………………………………………………………………...229
Table C1 HFP_G10 and HFP_G16 dephasing △S/S0 in POPC: POPG: Chol_d7/d6, the data is
processed with 100Hz Gaussian line broadening and integration width of 3ppm for G10, and
1ppm for G16…………………………………………………………………………………...246
Table C2 HFP_G5 dephasing △S/S0 in DOPC: DOPG: Chol_d7/d6 (8:2:5), the data is processed
with 100Hz Gaussian line broadening and 2ppm integration width……………………………246
Table C3 HFPL9R_G5 and G10 dephasing △S/S0 in DPPC: DPPG (4:1), the data is processed
with 100Hz Gaussian line broadening and 3ppm integration width……………………………247
Table C4 KALP_A5, A7, A17 and A19 dephasing △S/S0 in DPPC: DPPG (4:1), the data is
processed with 100Hz Gaussian line broadening and 3ppm integration width………………...248
Table C5 DMPC-d54 order parameter (SCD) in d54: DMPG (4:1) and d54: DMPG: CHOL (8:2:5)
without and with HFP, and the order parameter percentage change by HFP. ΔSCD= SCD Lipid – SCD
(Lipid+ HFP)
………………………………………………………………………………………...250
Table C6 DMPC-d54 T2 fitting data for echo, CD3 and CD2 for both without and with HFP at
different temperature. The membrane is d54: DMPG= 4:1 mol ratio……………………….…251
Table C7 DMPC-d54 T2 fitting data for echo, CD3 and CD2 for both without and with HFP at
different temperature. The membrane is d54: DMPG: CHOL= 8:2:5………….……………...255
Table C8 Chol_d6 T2 fitting data for echo intensity for both without and with HFP at 37 °C in
membrane of DMPC: DMPG: Chol_d6= 8:2:5 and POPC: POPG: Chol_d6= 8:2:5………….259

xi

Table C9 Chol_d7 T2 fitting data with ln(CD3 peak intensity) for both without and with HFP at
different temperatures. Peak intensity is short as PI……………………………………………260
Table C10 Chol_d7 T1 fitting data (CD3 peak intensity) for both without and with HFP at
different temperatures. Peak intensity is short as PI…………………………………………....262

xii

LIST OF FIGURES

Figure 1.1 The two spin states m= Âą1/2 of a nucleus with I= 1/2 in the static magnetic field B0.
The corresponding magnetic moments Οι and Οβ make precession about B0 in the z direction
with Larmor frequency ω0 = γB0. ....................................................................................................3
Figure 1.2 Vector representation of M precession with 90X (a) and 180X (b) pulse.[4] …..........5
Figure 1.3 Shielding tensor (red) relative to the B0 field in PAF with principal axes values of σxx,
σyy and σzz (a), where  is the angle between the z-axis of PAF and B0, and  is the angle
between the x-axis of PAF and the projection of B0 on the x-y plane of PAF. PAF associated
Euler angles, ι, , and  with respect to B0 field (b). [7] ................................................................8
Figure 1.4 (a) CSA powder pattern of 13CO with the three principal chemical shift values of xx
= 247 ppm, yy = 176 ppm and zz = 99 ppm. (b) The PAF of 13CO in peptide or protein
backbone with zPAF perpendicular to the peptide bone C-CO-N plane, yPAF alone the CO bond
direction, and xPAF perpendicular to the CO bond in the C -CO- N plane.....................................9
Figure 1.5 Dipolar coupling between nuclear spin I and S with inter-nuclear distance (r) and
azimuthal angle ().  is the angle between the inter-nuclear vector and the magnetic field B0
which is alone z axis......................................................................................................................11
Figure 1.6 Prolate (a) and Oblate (b) charge distribution of quadrupolar nucleus and the
corresponding quadrupole moment shown in (c) and (d) respectively. Prolate moment (c) is
positive and Oblate moment (d) is negative. [2, 18]......................................................................13
Figure 1.7 The orientation dependence of static 2H spectra. Discrete lines are observed for the
allowed two transitions (m = +1  m = 0, and m = 0  m = -1) with (a)  = 0, (b)  = 54.7
and (c)  = 90 where  is the angle between the C – 2H bond and B0 field. (d) The quadrupolar
powder pattern for the allowed two transitions for all possible s in static sample.[2, 18]...........17
Figure 1.8 Magic angle spinning (MAS) for 13C- 2H inter-nuclear vector, the angle Îą is the angle
between the external magnetic field B0 and sample- spinning axis. When the angle Îą is fixed at
54.7°, the sample spinning is called magic angle spinning.[4] θ and β is the angle between 13C2
H distance vector and B0 and spinning axis respectively. ...........................................................18
Figure 1.9 The effect of different spinning speed on the observed spectrum. For this example,
the isotropic chemical shift is set at 0 Hz, the CSA is 5 kHz, and the asymmetry is 0. When the
spinning speed is slow, there are spinning sidebands spaced at spinning frequency; when the
spinning speed is fast enough to overcome the CAS, only isotropic chemical shift is observed
with high intensity.[1] And the figure is from reference 1…........................................................19
Figure 1.10 Pulse sequence for 13C - 15N REDOR …...................................................................20

xiii

Figure 1.11 Diagram of heteronuclear dipolar coupling evolution over rotor period for S 0
experiment in REDOR. The + and – sign represent positive and negative dipolar coupling
respectively. MAS represents the dipolar coupling spatial dependence over each rotor period; C
spin represents the observing spin operator and π pulse changes the sign of dipolar coupling; S0
represents the overall effects from MAS and C spin π pulses on dipolar coupling over each rotor
period. As we can see, the dipolar coupling for S0 is averaged to zero over each rotor period.....25
Figure 1.12 Diagram of heteronuclear dipolar coupling evolution over rotor period for S1
experiment in REDOR. The + and – sign represent positive and negative dipolar coupling
respectively. MAS represents the dipolar coupling spatial dependence over each rotor period; C
spin and N spin represent the observing and dephasing spin operator respectively and π pulse
changes the sign of dipolar coupling. S1 represents the overall effects from MAS, C spin
(detecting) and N (dephasing) spin π pulses on dipolar coupling over each rotor period. As we
can see, the dipolar coupling for S1 is re-introduced and is nonzero over each rotor period…....26
Figure 1.13 “Quecho” pulse sequence. Theoretically,  = 1.......................................................27
Figure 1.14 Electron microscopy (right panel) and the relevant model (left panel) of HIV- host
cell viral entry process. (a) HIV virus binding to host cell, (b) HIV and host cell membrane
hemifusion, (c) large viral pore formation and (d) HIV viral components released into host cell.
In the model, the spikes represent HIV viral membrane protein, black triangle is the viral RNA
and black dots are other components including proteins.[30] ......................................................30
Figure 1.15 HIV interacts with T cell surface protein CD4 and chemokine receptor CXCR4.
HIV gp120 protein interacts with T cell CD4 and CXCR4 sequentially and moves away, then
gp41 get exposed to interact with cell membrane.[38]………………..........................................31
Figure 1.16 Schematic representation (A) and partial sequences (B) of HIV-1 gp41 protein.
Colored boxes show the functional regions of gp41. Form N terminal to C terminal, FP is fusion
peptide region, NHR and CHR is N- heptad repeat and C-heptad repeat respectively, Loop is the
loop region, MPER is the membrane proximal external region, TM is the transmembrane domain,
and Endo is the endodomain or cytoplasmic domain. The ectodomain without FP is HM protein,
and the full ectodomain including FP of gp41 is FP-HM. FP containing protein could induce
more fusion.[39, 40] The amino acid sequence of HM and FP-HM (B) is color coded according
to the different domains.[39] A minimal six residues loop SGGRGG replaces the native loop and
does not affect the SHB assembly.[41,42] …………………………............................................32
Figure 1.17 Three major states of gp41 during fusion: (A) pre- fusion native state where gp41 is
trimeric and non-covalently associated with gp120, (B) extended pre-hairpin state where gp41
has conformational change and interacts with host cell membrane, (C) post fusion hairpin state
(six helical bundle states).[41,43]. Figures in (B) and (C) do not show gp120 to focus on the
change of gp41…………………………………………….……………………………………..33
Figure 1.18 X- ray crystallography of soluble gp140 (PDB: 4NCO), a near-native gp160 without
MPER and transmembrane domain of gp41 with N-terminal 4 residues of MPER and without
transmembrane and cytoplasmic domains of gp41: (a) trimer of gp140, gp120 is in yellow,

xiv

orange, and red; gp41 is in green. (b) Close view of dimer of gp120 and gp41 in the trimer of
gp140.[46] ………………….........................................................................................................34
Figure 1.19 Crystal structure of HIV gp41 ectodomain composed of FPPR-CHR-NHR-MPER
(Gly531 to Leu581 in blue and Met628 to Tyr681 in green; gp41 531-681) (PDB: 2X7R). The gp41
does not include the N-terminal fusion peptide (gp41 512-530). Residues are numbered
according to their positions in gp160 complex).[54] ....................................................................35
Figure 2.1 13C – 2H REDOR pulse sequence. Each sequence starts with a CP from 1H to the
observed 13C nucleus to enhance the intensity of 13C signal followed by a dephasing and
acquisition period. TPPM 1H decoupling was applied during the dephasing and the acquisition
periods. ..........................................................................................................................................50
Figure 2.2. “Quecho” pulse sequence. .........................................................................................51
Figure 3.1 (a) DPPC bilayer regular membrane (left) without chemically modified lipid and
interdigitated membrane (right) composed of C16- 19F DPPC with a 1H→ 19F substitution at C16,
(b) chemical structure of C16- 19F DPPC lipid.[15, 17] ...............................................................56
Figure 3.2 (A) 2H patterns and structures of deuterated DPPC lipids and CHOL and (B)
approximate 2H’s and 31P’s (P) membrane locations in the membrane without protein. DPPC
lipids are deuterated at different regions of the acyl chain. The lipid 2H and 31P locations are for
the membrane gel-phase without CHOL and CHOL 2H locations are for the liquid-ordered phase
with CHOL. The same color-coding is used in other figures through this thesis. ........................58
Figure 3.3 Experimental 13CO-2H (△S/S0)exp (red squares with error bars) for sample HFP_G5C
in PC_d10 membrane. The (△S/S0)exp are for the major lab G5 peak with  sheet structure and
error bars are calculated from spectral noise. Fitted (△S/S0) are displayed for three different
fitting approaches. The blue crosses (Approach I) and green stars (Approach II) are based on
models of two- (P1 and P2) and three-populations (P1, P2, and P3) of HFP_G5C molecules,
respectively. The (△S/S0) for each population is calculated with the quantum mechanics-based
SIMPSON program using a model of isolated 13CO-2H spin-pairs with a single dipolar coupling
(d). For the two population model, the best-fit parameter values for P1 are d = 53 Hz and
fractional population A = 0.71. The corresponding P2 = 1 – A = 0.29 with d = 0 Hz. For the threepopulation model, the best-fit values are d1 = 90 Hz, A1 = 0.27, d2 = 25 Hz, and A2 = 0.50 with P3
= 1 – A1 – A2 = 0.23 and d3 = 0 Hz. The black line (Approach III) is the best-fit to the
exponential buildup function A × (1 – e – ) with A = 0.63 and  = 44 Hz…………....................62
Figure 3.4 13C-detect / 2H-dephase REDOR S0 (black) and S1 (red) experimental spectra of
membrane - associated HFP G5c at  = 40ms by different sample preparation methods, (a)
organic cosolubilization method (organic) and (b) aqueous vesicle binding method (aqueous).
Each spectrum is the sum of ~ 40000 scans and processed with 150 Hz Gaussian line broadening
and polynomial baseline correction. The observed chemical shifts for G5 are 171ppm from both
organic and aqueous methods, and this is consistent with major  sheet structure of HFP in
membrane. The similar structure supports that HFP achieves thermodynamic equilibrium
structure when it’s associated with membrane. The membrane is composed of 40μmol
DPPC_d10 and 10Îźmol DTPG lipids. Sample prepared by organic methods contains ~ 2Îźmol

xv

and sample prepared by aqueous method contains ~ 1.3Îźmol HFP. The cooling N2 gas
temperature is ~ - 50 °C with corresponding sample temperature of ~ - 30 °C……....................64
Figure 3.5 The dephasing buildups of △S/S0 vs dephasing time () for different NMR sample
preparation methods: organic cosolubilization (closed square) and aqueous vesicle binding (open
square). The intensity for S0 and S1 were obtained by integration over a 3ppm width centered at
the highest peak intensity. The similar dephasing buildups support thermodynamic equilibrium
membrane location of HFP when it’s associated with membrane. ...............................................65
Figure 3.6 The spectra (top panel) and dephasing buildups (bottom panel) of △S/S0 vs dephasing
time () for sample investigated at ~ - 30 °C and ~ - 0 °C: the spectrum is for  = 40ms and the
black is S0 and red is S1. The data is processed with 150Hz Gaussian line broadening. ..............67
Figure 3.7 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra of
HFP_G5c in (a) neutral membrane and (b) negatively charged membrane at  = 40ms. The data
is processed with 100Hz Gaussian line broadening and polynomial baseline correction. The 171
ppm chemical shift indicates that HFP has predominant  sheet structure at the labeled Gly5
residue site.....................................................................................................................................69
Figure 3.8 The dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing time
() for HFP_G5c in (top panel) neutral membrane and (bottom panel) negatively charged
membrane. The neutral membrane is composed of PC lipid, and the negative membrane is PC:
PG lipids of 4:1mol ratio. The dephasing buildups are for the major peak at 171 ppm with
integration window of 3ppm for both S0 and S1. The colored line is the best fit exponential
buildup curve fitted by A × (1 – e – ) for samples with significant dephasing buildups. ............70
Figure 3.9 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra of
HFP_L12c in PC: PG =4:1 membrane. The data is processed with 20Hz Gaussian line
broadening and polynomial baseline correction…………………………………………............74
Figure 3.10 The dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing
time () for HFP_G12c in membrane composed of PC: PG = 4:1 ratio. The △S/S0 data is for the
major peak of  sheet conformation and the S0 and S1 is integration with 3ppm integration
window. The colored line is the best fit exponential buildup curve fitted by A × (1 – e – ).........74
Figure 3.11 HFP membrane location model in anionic membrane without CHOL: (a) major
deeply inserted membrane location and (b) minor shallowly inserted membrane location. The
membrane 2H positions represent the location without protein. ...................................................76
Figure 3.12 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
(top panel) and the dephasing buildups (bottom panel) of △S/S0 (colored squares) with error bars
vs dephasing time () for HFP_G5c in membranes with Chol_d7 and Chol_d6. The S0 and S1
spectra displayed are for  =40ms. The data is processed with 100Hz Gaussian line broadening
and polynomial baseline correction. The dephasing buildup is for the major  peak. S0 and S1 is
integration through 3ppm integration window. The colored line is the best fit exponential buildup
curve fitted by A × (1 – e – ) ........................................................................................................79

xvi

Figure 3.13 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra (a)
and the dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing time () for
HFP_G16c in membranes with Chol_d7 and Chol_d6 for 171ppm peak (b) and 174ppm peak (c).
The S0 and S1 spectra displayed are for  =40ms. The data is processed with 100Hz Gaussian line
broadening and polynomial baseline correction. The dephasing buildup is for the major  peak
with 171ppm chemical shift. S0 and S1 is integration through 1ppm integration window. The
colored line is the best fit exponential buildup curve fitted by A × (1 – e – )...............................80
Figure 3.14 Semi-quantitative HFP membrane location model with major population deeply
inserted and minor population shallowly inserted to the membrane hydrophobic core. The
membrane 2H positions represent the location without protein. ...................................................81
Figure 3.15 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
for HFP_G10c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. POPC: POPG: CHOL = 8:2:5 and 8:2:2.5 mole ratios. ............84
Figure 3.16 The dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing
time () for HFP_G10c in membrane composed of POPC: POPG: CHOL = 8:2:5 and 8:2:2.5
ratios. The △S/S0 data is for the major peak of  sheet conformation and the S0 and S1 is
integration with 3 ppm integration window. The colored line is the best fit exponential buildup
curve fitted by A × (1 – e – ). .......................................................................................................84
Figure 3.17 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
for HFP_G16c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial
baseline
correction.
POPC:
POPG:
CHOL
=
8:2:5. .............................................................................................................................................85
Figure 3.18 The dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing
time () for HFP_G16c in membrane composed of POPC: POPG: CHOL = 8:2:5 ratio. The
△S/S0 data is for the major peak of  sheet conformation and the S0 and S1 is integration with 1
ppm integration window. The colored line is the best fit exponential buildup curve fitted by A ×
(1 – e – ). ......................................................................................................................................85
Figure 3.19 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
for HFP_G5c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. DOPC: DOPG: CHOL = 8:2:5. Chol_d7 sample contains ~
1.5Îźmol peptides, and Chol_d6 sample contains ~ 1.3Îźmol peptides. .........................................88
Figure 3.20 The dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing
time () for HFP_G5c in membrane composed of DOPC: DOPG: CHOL = 8:2:5 ratio. The
△S/S0 data is for the major peak of  sheet conformation and the S0 and S1 is integration with 2
ppm integration window. ..............................................................................................................88
Figure 3.21 Schematic pictures of the hydrophobic patch (yellow shaded area) formed by the Nterminal most hydrophobic 12 residues of HFP based on amino hydrophobicity. (a) 20→1/1→20
registry for HFP_V2E mutant. (b) 16→1/1→16 registry for HFP wild type. ..............................91

xvii

Figure 3.22 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
(top panel) and the dephasing buildups (bottom panel) of △S/S0 (colored triangles) with error
bars vs dephasing time () for HFP_V2E_G5c. The spectra displayed is for  =40ms. The data is
processed with 100Hz Gaussian line broadening and polynomial baseline correction. The
quantitative dephasing buildup is for the major  peak with 171ppm chemical shift. S0 and S1 is
integration through 3ppm integration window. The colored line is the best fit exponential buildup
curve fitted by A × (1 – e – ). .......................................................................................................92
Figure 3.23 Semi-quantitative HFP_V2E membrane location model with major population
deeply inserted and minor population shallowly inserted to the membrane hydrophobic core.
There is more minor population compared to HFP wildtype with shallowly inserted membrane
location. How the lipids and CHOL are displaced by HFP molecules is not known and neither
the orientation the neighbor lipids and CHOL of HFP..................................................................94
Figure 3.24 Membrane location model of HFP_L9R mutant consistent with our REDOR
experimental data. A short helix is shown from G5 to G10 to reflect the helical conformation for
G5 and G10 residues. And the secondary structure of other residues in HFP_L9R mutant is not
determined and shown as line. The R9 sidechain is likely pointing out to the direction of the
membrane surface. However, the arginine side chain length is ~ 7.5 Å, which is shorter than the
hydrophobic thickness of half membrane leaflet. HFP_L9R mutant probably induces local
membrane thinning, and similar membrane curvature relative to membrane fusion intermediate.
This might help explain the fusogenicity of L9R mutant. ............................................................97
Figure 3.25 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra (a)
and the dephasing buildups (b) of △S/S0 (colored squares) with error bars vs dephasing time ()
for HFP_L9R_G5c. The spectra displayed is for  =40ms. The data is processed with 100Hz
Gaussian line broadening and polynomial baseline correction. The quantitative dephasing
buildup is for the major peak with 176 ppm chemical shift. S0 and S1 is integration through 3ppm
integration window. The colored line is the best fit exponential buildup curve fitted by A × (1 – e
–
). Fitting in (b) in done with A≤1, the χ2 for d10 fitting is 5. Fitting in (c) is done with no
limitation on A, and the χ2 for d10 fitting is 2.5. ..........................................................................98
Figure 3.26 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra (a)
and the dephasing buildups (b) of △S/S0 (colored squares) with error bars vs dephasing time ()
for HFP_L9R_G10c. The spectra displayed is for  =40ms. The data is processed with 100Hz
Gaussian line broadening and polynomial baseline correction. The quantitative dephasing
buildup is for the major peak with 176 ppm chemical shift. S0 and S1 is integration through 3ppm
integration window. The colored line is the best fit exponential buildup curve fitted by A × (1 – e
–
). Fitting in (b) in done with A≤1, the χ2 for d10 fitting is 45. Fitting in (c) is done with no
limitation on A, and the χ2 for d10 fitting is 10…….....................................................................99
Figure 3.27 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
for KALP_A5c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. PC: PG =4:1 mole ratio. ..........................................................103

xviii

Figure 3.28 The KALP_A5c dephasing buildup of △S/S0 (colored squares) with error bars vs
dephasing time. S0 and S1 is integration through 3ppm integration window. The colored line is
the best fit exponential buildup curve fitted by A × (1 – e – ). PC: PG =4:1 mole ratio. ...........103
Figure 3.29 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
for KALP_A7c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. PC: PG =4:1 mole ratio. ..........................................................104
Figure 3.30 The KALP_A7c dephasing buildup of △S/S0 (colored squares) with error bars vs
dephasing time. S0 and S1 is integration through 3ppm integration window. The colored line is
the best fit exponential buildup curve fitted by A × (1 – e – ). PC: PG =4:1 mole ratio. ...........104
Figure 3.31 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
for KALP_A17c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. PC: PG =4:1 mole ratio. ………………………………..........105
Figure 3.32 The KALP_A17c dephasing buildups of △S/S0 (colored squares) with error bars vs
dephasing time. S0 and S1 is integration through 3ppm integration window. The colored line is
the best fit exponential buildup curve fitted by A × (1 – e – ). PC: PG =4:1 mole ratio. ...........105
Figure 3.33 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
for KALP_A19c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. PC: PG =4:1 mole ratio. ……………………………..............106
Figure 3.34 The KALP_A19c dephasing buildup of △S/S0 (colored squares) with error bars vs
dephasing time. S0 and S1 is integration through 3ppm integration window. PC: PG =4:1 mole
ratio. ……………………………................................................................................................106
Figure 3.35 The membrane location model for KALP peptide with one representative lysine
sidechain near N- and C- terminal. (a) Major populations of A5 and A7 have close contact to
PC_d10 located near the membrane center. (b) Major populations of A17 and A19 have close
contact to PC_d10 located near the membrane center, but with A19 further away to PC_d10
compared to A17. (c) Significant populations of A5, A7, A17 and A19 make close contact of
PC_d8. (d) Amino acid sequence for KALP peptide. Snorkeling effect of terminal lysine
sidechains help extend the hydrophobic length of the peptide by pointing out to the aqueous
surface near the phosphate group. The molecular population of (a): (b): (c) ≈ 3:3:2, because the
dephasing for PC_d10: PC_d8 is ~ 3:2 for all A5, A7, A17 and A19 residues at τ=48ms. There
isn’t substantial dephasing buildup for PC_d4 might because the snorkeling effect of lysing
sidechains might make lysine sidechains displace PC_d4 and thus enlarge the inter-nuclear
distance between labeled 13CO and PC_d4 and is beyond the ~ 8Å detection limit of 13C-2H
REDOR. ......................................................................................................................................108
Figure 4.1 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
(top panel) of membrane - associated HFP_G5c at  = 40ms and quantitative △S/S0 buildups for
the major  peak in membrane with different 2H labeling. The colored line is best fitted curve by
equation of A × (1 – e – ). ..........................................................................................................118

xix

Figure 4.2 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
(top panel) of membrane – associated HFP_G5c at  = 40ms and quantitative △S/S0 buildups for
the major  peak in membrane with different 2H labeling. The colored line is best fitted curve by
equation of A × (1 – e – ). ..........................................................................................................119
Figure 4.3 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
(top panel) of membrane – associated HFP_G10c at  = 40ms and quantitative △S/S0 buildups
for the major  peak in membrane with different 2H labeling. The colored line is best fitted curve
by equation of A × (1 – e – ). .....................................................................................................120
Figure 4.4 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
(top panel) of membrane – associated HFP_G16c at  = 40ms and quantitative △S/S0 buildups
for the major  peak in membrane with different 2H labeling. ...................................................121
Figure 4.5 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra
(top panel) of membrane – associated HFP_G16c at  = 40ms and quantitative △S/S0 buildups
for the major  peak at 171 ppm in membrane with different 2H labeling. The colored line is best
fitted curve by equation of A × (1 – e – ). ..................................................................................122
Figure 4.6 Eight favorable dockings of CHOL (colored tubes) that meet two REDOR 4–5 Å
distance constraints: the YASARA energy minimized predicted HFP-antiβ and HFP-antiβmin
structures (drawn as lines, with residues labeled) in complex with the corresponding CHOL
binding mode are colored red for the least energetically favorable group (-8535 to -8841 kJ¡mol1
), green for the intermediate energy group (-8841 to -9147 kJ.mol-1) and purple, for the most
favorable group (-9147 to -9454 kJ.mol-1). Note the high occupancy of seven favorable CHOL
dockings spanning Ala1–Gly5 (position 1, lower left), with just one docking occupying position
2 (top center), as shown by thick purple tubes for the two most favorable dockings. CHOL
interactions with HFP residues are listed earlier in this chapter. ................................................136
Figure 4.7 The most favorable CHOL binding mode relative to HFP-antiβ: (a) CHOL binding
mode (purple tubes) is shown relative to the two HFP strands (sticks). (b) Details of HFP-antiβ
side chain interactions with CHOL protons monitored by REDOR (shown in white and labelled).
The binding mode is the same as shown for location 1 above, while rotated by roughly 180°
about the horizontal and vertical axes to enable viewing from above. (c) Same CHOL binding
mode shown above (position 1, purple), plus its symmetry mate at the opposite end of HFP-antiβ
(position 1’, magenta). ................................................................................................................137
Figure 4.8 The HFP–antiβ structural model in complex with (a) CHOL in its dominant favorable
position (position 1, purple tubes, as shown in Figure above) and (b) in the alternative binding
mode (position 2, shown here in green tubes). The two strands of HFP-antiβ (dark blue) with
cholesterol bound are shown in the context of a multi-stranded beta barrel structure (light blue
strands, from PDB entry 2iww) to explore the extent to which the two CHOL positions are
consistent with formation of a larger (more than two–stranded) antiparallel  sheet structure by
HFPs. The first cholesterol position (a) and its symmetry mate (not shown) would be compatible
with formation of a larger sheet structure by HFPs if the last four residues, GAAG, of every
second peptide in the sheet did not pair with the adjacent beta strand and shifted out of the way.

xx

The second CHOL site (b) would clearly block the addition of a beta strand (shown by
interpenetration between CHOL and the light–blue beta strand) and thus would preclude
formation of a larger sheet by HFPs............................................................................................138
Figure 4.9 Favorable dockings of CHOL (colored tubes) on the concave surface of HFP–antiβ:
Only the purple and green ones have within 5 Å distance between CHOL and HFP at one of the
sites measured by REDOR. There are eight favorable dockings predicted by Swiss Dock with
CHOL located on the concave surface of HFP-antiβ without blocking sheet extension. ...........140
Figure 4.10 The most favorable CHOL binding mode (purple tubes, location 3) for CHOL
located on the concave surface of HFP-antiβ, (a) the geometry of the most favorable CHOL
binding to the two stranded HFPs, (b) details of HFP-antiβ side chain interactions with CHOL
protons monitored by REDOR (shown in white and labelled)…………………………............141
Figure 4.11 Superimposition of this purple colored favorable binding (location 3) on PDB 2iww,
the purple colored CHOL is from the most favorable binding location 3. It’s still possible to have
this CHOL binding geometry if the neighbor strands accommodate the terminal CH3 protons of
CHOL during bigger sheet formation. …………………………………………………............142
Figure 5.1 Chemical structures of deuterated PC and CHOL used for 2H NMR study: (a) DMPCd54, (b) Chol_d6 and (c) Chol_d7…………...…………………………………………............152
Figure 5.2 2H - NMR spectra of DMPC-d54 taken at different temperature in membrane of
DMPC: DMPG= 40: 10μmol at pH 7.4.……………………………………………………......155
Figure 5.3 2H - NMR spectra of DMPC-d54 taken at different temperature in membrane of
DMPC: DMPG= 40: 10μmol with ~ 2μmol HFP at pH 7.4.………………….……………......156
Figure 5.4 2H - NMR spectra of DMPC-d54 taken at different temperature in membrane of
DMPC: DMPG: CHOL = 40: 10: 25μmol at pH 7.4……………………………………….......157
Figure 5.5 2H - NMR spectra of DMPC-d54 taken at different temperature in membrane of
DMPC: DMPG: CHOL = 40: 10: 25μmol with ~ 2μmol HFP at pH 7.4. …………………......158
Figure 5.6 2H - NMR spectra of DMPC-d54 without and with HFP at (a) 21C and (b) 37C in
membrane without CHOL at pH 7.4. Pure lipids are DMPC-d54: DMPG (40: 10Îźmol). HFP:
lipids ratio is 1:25. ………….......................................................................................................160
Figure 5.7 2H - NMR spectra of DMPC-d54 without and with HFP at (a) 21C and (b) 37C in
membrane containing 33% CHOL at pH 7.4. Membrane is DMPC-d54: DMPG: CHOL (40: 10:
25Îźmol). HFP: lipids ratio is 1:25. .............................................................................................161
Figure 5.8 2H - NMR spectra of DMPC-d54 without (top) and with different peptide to lipids
mole ratios, 1: 100 (middle) and 1: 25 (bottom) at 37C and pH 7.4. Pure lipids are DMPC-d54:
DMPG: CHOL with 40:10:25Îźmol. ...........................................................................................162

xxi

Figure 5.9 Chol_d6 2H-NMR spectra, without (top) and with HFP (bottom) at 37C and pH 7.4,
pure lipids are DMPC: DMPG: Chol_d7 with 40:10:25Îźmol. HFP: lipids mole ratio is 1:50, and
lipids do not include CHOL. .......................................................................................................163
Figure 5.10 DMPC-d54 de-Paked spectra, without (top) and with HFP (bottom) at 37C and pH
7.4. The membrane is DMPC-d54: DMPG with 40:10Îźmol. HFP: lipids ratio is 1:25. ............164
Figure 5.11 DMPC-d54 de-Paked spectra, without (top) and with HFP (bottom) at 37C and pH
7.4. The membrane is DMPC-d54: DMPG: CHOL with 40:10:25Îźmol. HFP: lipids ratio is
1:100. ..........................................................................................................................................165
Figure 5.12 DMPC-d54 de-Paked spectra, without (top) and with HFP (bottom) at 37C and pH
7.4. The membrane is DMPC-d54: DMPG: CHOL with 40:10:25Îźmol. HFP: lipids ratio is
1:25. ............................................................................................................................................166
Figure 5.13 HFP effects on the DMPC-d54 order parameters profile in membrane with and
without CHOL at 37C and pH 7.4. HFP decreases the order parameters along the acyl chain of
the lipid in membrane of DMPC-d54 (d54) and DMPG (pg) with 1:25 peptide to lipids ratio, and
membrane with additional 33% CHOL (+Chol) with 1:100 peptide to lipids ratio. However, HFP
increases the order parameters along the acyl chain of the lipid with 1:25 peptide to lipids ratio in
the membrane containing 33% CHOL.........................................................................................167
Figure 5.14 HFP effects on the order parameter change of DMPC-d54 along the lipid acyl chain
compared to pure membrane with and without CHOL at 37 C and pH 7.4. ΔSCD= SCD Lipid – SCD
(Lipid+ HFP)
. Positive value indicates decreased order parameter and negative value indicates
increased order parameter compared to pure membrane without HFP. The plot color and shape
are consistent with previous figure. ............................................................................................168
Figure 5.15 Representative stacked FID plots DMPC-d54 in membrane of DMPC-d54: DMPG
without HFP at 21C (top) and 37C (bottom) in static. The 2H FIDs were obtained by varying 
and 1. For each  and 1, the number of scans was 5000 (top) and 1000 (bottom),
respectively. ................................................................................................................................170
Figure 5.16 Representative stacked spectrum plots DMPC-d54 in membrane of DMPC-d54:
DMPG without HFP at 21C (top) and 37C (bottom) at pH 7.4 in static. The 2H spectra were
obtained by varying  and 1. For each  and 1, the number of scans was 5000 (top) and 1000
(bottom), respectively. All spectra were processed with 200 Hz line broadening, data shift = 11. ................................................................................................................................................171
Figure 5.17 Representative stacked spectrum plots DMPC-d54 in membrane of DMPC-d54:
DMPG with 2µmol HFP at 21C (top) and 37C (bottom) and pH 7.4 in static. The 2H spectra
were obtained by varying  and 1. The top spectra were arrayed to  = 750 and 1 = 731 ¾s. The
bottom spectra were arrayed to  = 1360 ¾s and  = 1341 ¾s. For each  and 1, the number of
scans was 5000 (top) and 1000 (bottom), respectively. All spectra were processed with DC offset
correction, data shift = -11. We processed the bottom spectra additionally with polynomial
baseline correction of the order 5. ...............................................................................................172

xxii

Figure 5.18 Representative stacked spectrum plots DMPC-d54 in membrane of DMPC-d54:
DMPG: CHOL (8:2:5) without HFP at 21C (top) and 37C (bottom) and pH 7.4 in static. The
2
H spectra were obtained by varying  and 1. The top spectra were arrayed to  = 1340 and 1 =
1321 ¾s. The bottom spectra were arrayed to  = 1320 ¾s and  = 1301 ¾s. For each  and 1, the
number of scans was 1000. All spectra were processed with DC offset correction, data shift = 11. ....... ........................................................................................................................................173
Figure 5.19 Representative stacked spectrum plots DMPC-d54 in membrane of DMPC-d54:
DMPG: CHOL (8:2:5) with 2µmol HFP at 21C (top) and 37C (bottom) and pH 7.4 in static.
The 2H spectra were obtained by varying  and 1. The top spectra were arrayed to  = 1340 and
1 = 1321 ¾s. The bottom spectra were arrayed to  = 1000 ¾s and  = 981¾s. For each  and 1,
the number of scans was 1000. All spectra were processed with DC offset correction, data shift =
-11. ..............................................................................................................................................174
Figure 5.20 Representative stacked spectrum plots DMPC-d54 in membrane of DMPC-d54:
DMPG: CHOL (8:2:5) without (top) and with HFP (HFP: phospholipids= 1:100) (bottom) at
37C (bottom) and pH 7.4 in static. The 2H spectra were obtained by varying  and 1. The top
spectra were arrayed to  = 1320 and 1 = 1301 ¾s. The bottom spectra were arrayed to  = 1160
¾s and  = 1141¾s. For each  and 1, the number of scans was 1000. All spectra were processed
with DC offset correction, data shift = -11. We processed the bottom spectra with additional 200
Hz Gaussian line broadening and polynomial baseline correction of the order
5....................................................................................................................................................175
Figure 5.21 Stacked spectrum plots Chol_d6 in membrane of DMPC: DMPG: Chol_d6 (8:2:5)
without (top) and with (bottom) 1µmol HFP at 37C and pH 7.4 in static. The 2H spectra were
obtained by varying  and 1. The spectra were arrayed to  = 340 and 1 = 321 ¾s. For each 
and 1, the number of scans was 10000. .....................................................................................176
Figure 5.22 Quecho experimental (red squares) and best fit (red lines) plots of tip intensity of
echo FID vs 2 for DMPC-d54 in membrane of DMPC-d54: DMPG without HFP under static
conditions at different temperatures and pH 7.4. The data are fitted with 𝐼(2𝜏) = 𝐼(0) × exp( −
2𝜏/𝑇2 ) + 𝐴, where A is the fitting offset. ...................................................................................177
Figure 5.23 Quecho experimental (red squares) and best fit (red lines) plots of tip intensity of
DMPC-d54 echo FID vs 2 in membrane of DMPC-d54: DMPG with HFP under static
conditions at different temperatures and pH 7.4. The HFP: lipids mole ratio is 1:25. The data are
fitted by equation 𝐼(2𝜏) = 𝐼(0) × exp( − 2𝜏/𝑇2 ) + 𝐴, where A is the fitting offset. There are
some fitting deviations relative to experimental data because there are fast decay components
(CD2) and slow decay components (CD3). .................................................................................178
Figure 5.24 Quecho experimental (red squares) and best fit (red lines) plots of tip intensity of
echo FID vs 2 for DMPC-d54 in membrane of DMPC-d54: DMPG: CHOL (8:2:5 mole ratio)
without HFP under static conditions at different temperatures and pH 7.4. The data are fitted by
equation 𝐼(2𝜏) = 𝐼(0) × exp( − 2𝜏/𝑇2 ) + 𝐴, where A is the fitting offset. There are some fitting
deviations relative to experimental data because there are fast decay components (CD2) and slow
decay components (CD3). ...........................................................................................................179

xxiii

Figure 5.25 Quecho experimental (red squares) and best fit (red lines) plots of tip intensity of
DMPC-d54 echo FID vs 2 in membrane of DMPC-d54: DMPG: CHOL (8:2:5 mole ratio) with
HFP under static conditions at different temperatures and pH 7.4. The HFP: phospholipids mole
ratio is 1:25. The data are fitted by equation 𝐼(2𝜏) = 𝐼(0) × exp( − 2𝜏/𝑇2 ) + 𝐴, where A is the
fitting offset. ................................................................................................................................180
Figure 5.26 Quecho experimental (red squares) and best fit (red lines) plots of DMPC-d54 echo
tip intensity of FID, CD3 and CD2 peak intensity vs 2 in membrane of DMPC-d54: DMPG:
CHOL (8:2:5 mole ratio) with HFP under static conditions at 37 °C and pH 7.4. The HFP:
phospholipids mole ratio is 1:100. The data are fitted by equation 𝐼(2𝜏) = 𝐼(0) × exp( −
2𝜏/𝑇2 ) + 𝐴, where A is the fitting offset. ...................................................................................181
Figure 5.27 Quecho experimental (red squares) and best fit (red lines) plots of Chol_d6 FID
echo tip intensity vs 2 in membrane of DMPC: DMPG: Chol_d6 (8:2:5 mole ratio) with and
without HFP under static conditions at 37 °C and pH 7.4. The HFP: phospholipids mole ratio is
1:50. The data are fitted by equation 𝐼(2𝜏) = 𝐼(0) × exp( − 2𝜏/𝑇2 ) + 𝐴, where A is the fitting
offset. ..........................................................................................................................................182
Figure 5.28 Stacked spectrum plots Chol_d6 in membrane of POPC: POPG: Chol_d6 (8:2:5)
without (top) and with (bottom) HFP at 37C and pH 7.4 in static. The 2H spectra were obtained
by varying  and 1. The top spectra were arrayed to  = 330 and 1 = 311 ¾s, and the bottom
spectra were arrayed to  = 300 and 1 = 281 ¾s. For each  and 1, the number of scans was
4000. ............................................................................................................................................186
Figure 5.29 Quecho experimental (red squares) and best fit (red lines) plots of Chol_d6 FID
echo tip intensity vs 2 in membrane of POPC: POPG: Chol_d6 (8:2:5 mole ratio) with and
without HFP under static conditions at 37 °C and pH 7.4. The HFP: phospholipids mole ratio is
1:50. The data are fitted by equation 𝐼(2𝜏) = 𝐼(0) × exp( − 2𝜏/𝑇2 ) + 𝐴, where A is the fitting
offset. ..........................................................................................................................................187
Figure 5.30 Stacked spectrum plots Chol_d6 in membrane of POPC: POPG: Chol_d6 (8:2:5)
with HFP at different temperatures and pH 7.4 in static. The 2H spectra were obtained by varying
 and 1. We processed the spectra with data shift, DC offset, 500 Hz Gaussian line broadening,
and baseline correction of the order 5. The number of scans was typically 10000. It is 35000 for
25 °C and 80000 for 0 °C. ...........................................................................................................189
Figure 5.31 Stacked spectrum plots Chol_d7 in membrane of POPC: POPG: Chol_d7 (8:2:5)
without and with HFP at -50C and pH 7.4 in static. HFP to phospholipids ratio is 1:50. The 2H
spectra were obtained by varying  and 1. For each  and 1, the number of scans was 2000 (top)
and 800 (bottom). We processed the data with -10 data shift pts, 2000 Hz Gaussian line
broadening and polynomial baseline correction of the order 5. ..................................................191
Figure 5.32 Stacked spectrum plots Chol_d7 in membrane of POPC: POPG: Chol_d7 (8:2:5)
without and with HFP at 37C and pH 7.4 in static. HFP to phospholipids ratio is 1:50. We
acquired the 2H spectra by varying  and 1. We arrayed  and 1 to 1000 and 975 ¾s (top), 1540
and 1521 ¾s (bottom). For each  and 1, the number of scans was 800. We processed the data

xxiv

with -11 data shift pts, 500 Hz Gaussian line broadening and polynomial baseline correction of
the order 5. ..................................................................................................................................192
Figure 5.33 Fitting plots of Chol_d7 ln (CD3 peak intensity) from membrane of POPC: POPG:
Chol_d7 (8:2:5) without HFP at different temperatures and pH 7.4. We acquired the data in static
condition. ....................................................................................................................................193
Figure 5.34 Fitting plots of Chol_d7 ln (CD3 peak intensity) from membrane of POPC: POPG:
Chol_d7 (8:2:5) with HFP at different temperatures and pH 7.4 in static. .................................194
Figure 5.35 Fitting plots of Chol_d7 ln (CD3 peak intensity) from membrane of POPC: POPG:
Chol_d7 (8:2:5) without HFP (a) and with HFP (b) at 37 °C and pH 7.4 in static. ....................195
Figure 5.36 “t1D_ir” pulse sequence for T1 relaxation study. ...................................................198
Figure 5.37 “t1D_ir” experimental (red squares) of Chol_d7 from membrane of POPC: POPG:
Chol_d7 (8:2:5) without HFP at 37 °C and pH 7.4 in static. The number of scans is 3000 for each
1. (a) 2H FID for 1 = 0.1 and 150.1ms, (b) Chol_d7 2H spectra for 1 = 0.1 through 150.1ms.
We did not show spectra for 1 = 180.1 through 510.1ms for view simplicity. All Spectra are
processed with 500 Hz Gaussian line broadening, -7 data shift points, and polynomial baseline
correction of order 5. ...................................................................................................................199
Figure 5.38 “t1D_ir” experimental (red squares) of Chol_d7 from membrane of POPC: POPG:
Chol_d7 (8:2:5) with HFP at 37 °C and pH 7.4 in static. The HFP to phospholipids mole ratio is
1:50. The number of scans is 3000 for each 1. (a) 2H FID for 1 = 0.5 and 120.5ms, (b) Chol_d7
2
H spectra for 1 = 0.5 through 120.5ms. We did not show spectra for 1 = 140.5 through
340.5ms for view simplicity. We processed the spectra with 500 Hz Gaussian line broadening,
data shift of -12, and polynomial baseline correction of order 5. ...............................................200
Figure 5.39 “t1D_ir” experimental (red squares) and best fit (red line) of Chol_d7 CD3 peak
intensity vs 1 from membrane of POPC: POPG: Chol_d7 (8:2:5) without and with HFP at 37 °C
and pH 7.4 in static. HFP: phospholipids mole ratio is 1:50.......................................................201
Figure 5.40 “t1D_ir” experimental (red squares) and best fit (red line) of Chol_d7 CD3 peak
intensity vs 1 from membrane of POPC: POPG: Chol_d7 (8:2:5) without HFP at different
temperatures and pH 7.4 in static.................................................................................................202
Figure 5.41 “t1D_ir” experimental (red squares) and best fit (red line) of Chol_d7 CD3 peak
intensity vs 1 from membrane of POPC: POPG: Chol_d7 (8:2:5) with HFP at different
temperatures and pH 7.4 in static. HFP: phospholipids mole ratio is 1:50. ................................203
Figure A1 FPHM_V2E amino acid and DNA sequences. The C-terminal
GGGGGGLEHHHHHH residues are non-native tag. SGGRGG is the engineered loop…...…224
Figure A2 SDS-PAGE of FPHM_V2E mutant. PBS wash is the supernatant of the inclusion
bodies in PBS buffer at pH 7.4. Filter through and wash 1-4 is in solubilization buffer, and
Elution 1 is in solubilization buffer + 250mM imidazole………………………………………226

xxv

Figure A3 Proteomics results of the SDS-PAGE band corresponding to FPHM_V2E protein.
Green color shaded M means there is detection of digested short peptides including oxidation
(+16). Green color shaded Q and N means there is detection of digested short peptides including
deamidated Q and N (+1) respectively. Green color shaded E means there is detection of digested
short peptides including dehydrated E (-18)……………………………………………………226
Figure A4 CD spectroscopy of FPHM_V2E (top panel) and melting temperature (bottom panel)
in 10mM tris(hydroxymethyl)aminomethane (Tris-HCl) + 0.2% SDS at pH 7.4. The melting
temperature plot is based on the mean residue molar elipticity at 222nm……………………...230
Figure A5 Gel filtration chromatograph of FPHM_V2E in 10mM Tris-HCl + 0.2% SDS +
150mM NaCl at pH 7.4. The highest peak is eluted at 14.23mL and corresponds to molecular
weight of 99kDa………………………………………………………………………………...231
Figure A6 Lipid mixing assay of FPHM_V2E in POPC: CHOL= 2:1 vesicles at pH 7.4. Protein:
Lipids = 1:300 mole ratio……………………………………………………………………….233

xxvi

KEY TO ABBREVIATIONS

AIDS: acquired immunodeficiency syndrome
CD: circular dichroism
CHR: C-terminal heptad repeat region
CHOL: cholesterol
Chol-d6: cholesterol-2,2,3,4,4,6-d6
Chol-d7: cholesterol-25,26,26,26,27,27,27-d7
CP: Cross polarization
CSA: chemical shift anisotropy
DMPC: 1,2-dipalmitoyl-sn-glycerol-3-phosphocholine
DMPC-d54: 1,2-dimyristoyl-d54-sn-glycero-3-phosphocholine
DMPG: 1,2-dipalmitoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (sodium salt)
DOPC: 1,2-dioleoyl-sn-glycero-3-phosphocholine
DOPG: 1,2-dioleoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (sodium salt)
DPC: dodecylphosphocholine
DPPC: 1,2-dipalmitoyl-sn-glycero-3-phosphocholine
DPPG: 1,2-dipalmitoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (sodium salt)
DTPG: dipalmitoylphosphatidylglycerol
Endo: endo-domain
FACTS: fast analytical continuum treatment of solvation
FID: free induction decay
FMOC: 9-Fluorenylmethoxycarbonyl

xxvii

FP: fusion peptide
FPPR: fusion peptide proximal region
FT: Fourier Transform
FWHM: full width at half maximum
HEPES: 2-(4-(2-hydroxyethyl)piperazin-1-yl) ethanesulfonic acid
HFP: HIV fusion peptide
HIV: Human immunodeficiency virus
HPLC: high pressure liquid chromatography
MALDI-TOF: matrix-assisted laser desorption ionization- time of flight
MAS: Magic angle spinning
MES: 2-(N-morpholino) ethanesulfonic acid
MPER: membrane-proximal external region
N-NBD-PE: N-(7-nitro-2, 1, 3-benzoxadiazol-4-yl) (ammonium salt)
dipalmitoylphosphatidylethanolamine
N-Rh-PE: N-(lissamine rhodamine B sulfonyl) (ammonium salt)
dipalmitoylphosphatidylethanolamine
NHR: N-terminal heptad repeat region
na: natural abundance
NMR: nuclear magnetic resonance
PC: phosphatidylcholine
PC-d4: 1,2-(dipalmitoyl-2,2,2,2-d4)-sn-glycero-3-phosphocholine
PC-d8: 1,2-(dipalmitoyl-7,7,7,7,8,8,8,8-d8)-sn-glycero-3-phosphocholine
PC-d10: 1,2-(dipalmitoyl-15,15,15,15,16,16,16,16,16,16-d10)-sn-glycero-3-phosphocholine

xxviii

PG: phosphatidylglycerol
PHI: pre-hairpin intermediate
POPC: 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine
POPG: 1-palmitoyl-2-oleoyl-sn-glycero-3- phospho-(1'-rac-glycerol) (sodium salt)
QCC: quadrupolar coupling constant
QUECHO: quadrupolar echo
REDOR: Rotational echo double resonance
RMSD: root mean square deviation
SDS: sodium dodecyl sulfate
SHB: six helical bundles
SPPS: solid phase peptide synthesis
ssNMR: solid state nuclear magnetic resonance
TFA: trifluoroacetic acid
TM: transmembrane domain
TPPM: tow-pulse phase modulated
Tris: (hydroxymethyl)aminomethane

xxix

Chapter 1 - Introduction
1.1 NMR Introduction
NMR spectroscopy bases on the properties of nuclear spin, and studies the nuclei with spin
quantum number I ≠ 0. The nucleus involves in multiple types of electric and magnetic field
interactions and each interaction type has its corresponding Hamiltonians.[1] The total
Hamiltonian expression is:
̂t = H
̂ ext + H
̂ int = (H
̂Z + H
̂ RF ) + (H
̂ CS + H
̂J + H
̂D + H
̂ Q)
H

1.1

̂ ext is the Hamiltonian for the interaction between the nuclear spin and the external
Where H
̂ Z and H
̂ RF , which corresponds to the static magnetic field B0 and the
magnetic field, including H
radio frequency (r.f.) field B1 respectively;
̂ int is the Hamiltonian for the interaction between the nuclear spin and the intrinsic field within
H
the sample including:
̂ CS is chemical shift interaction between the nuclear spin and the chemical shift field (electronic
H
shielding field) induced by B0;
̂ J is J coupling interaction or spin-spin coupling within one molecule;
H
̂ D is the direct dipolar coupling interaction between two nuclear spins;
H
̂ Q is quadrupole interaction between the nucleus (I ≥1) quadrupole moment and surrounding
H
electric filed gradient.
In the dissertation a vector is displayed in bold letters, the quantum mechanical operators are
shown with “^” above the letters and the vector-operators are shown in bold letters with “^”
above them.

1

1.1.1 Zeeman interaction
The interaction between the nuclear spin and the external static magnetic field (𝐁𝟎 ) is Zeeman
interaction and the Hamiltonian is:
̂ Z = −𝛍
H
̂ ∙ 𝐁𝟎

1.2

Where 𝛍
̂ is the nuclear magnetic moment operator, and it is related to nuclear spin operator 𝐈̂ as
𝛍
̂ = γћ𝐈̂

1.3

Where γ is the gyromagnetic ratio; and ћ is the reduced Plank’s constant. Then,
̂ Z = −𝛍
H
̂ ∙ 𝐁𝟎 = −γћ𝐈̂ ∙ 𝐁𝟎 = −γћ(Îx 𝐢 + Îy 𝐣 + Îz 𝐤) ∙ B0 ∙ 𝐤 = −γћÎz B0

1.4

Where i, j and k is the x, y and z direction’s unit vector respectively.
For a spin I nucleus, it has 2I+1 energy states in B0 field and each state has Eigen function as Ń°Im
or written as |I, m>. The m is the magnetic spin quantum number, and has the value of -I, ̂ Z and Îz have the same sets of eigenfunctions because H
̂ Z is proportional
I+1, .…., I-1, and I. H
to Îz . The energy value of eigenstate |I, m> is EI,m and
̂ Z |I, m > = EI,m |I, m > = −γћÎz B0 |I, m > = −γћB0 𝑚|I, m >
H

1.5

So,
EI,m = −γћB0 𝑚

1.6
1

1

For a spin 1/2 nucleus, I = 1/2 and m= 1/2. Then there are two possible eigenstates, |2 , + 2 >, and
1

1

1

1

|2 , − 2 >, which are also called α and β states respectively. E1,1 = − 2 𝛾ћ𝐵0, and E1,−1 = 2 𝛾ћ𝐵0.
22

2

2

The energy difference between the two states is ΔE. ∆E = ћ𝛾𝐵0 , and the corresponding
frequency is ω0 = γB0. This frequency is the rotation frequency of the net magnetization (M)
about the B0 field, and is Larmor frequency (Figure 1.1). The rotation of M comes from the

2

torque (T) derived from the net spin angular momentum (J) in the magnetic field. The relation is
derived as following: [1]
𝐌 = ∑ 𝛍𝐢 = ∑ γћ𝐈𝐢 = γ𝐉
i

𝐓=𝐌×𝑩=
So

d𝐌
dt

1.7

i

d𝐉 1 d𝐌
=
dt Îł dt

1.8

= γ𝐌 × 𝑩

1.9

Equation 1.9 predicts that M rotates about B0 with frequency ω0 = γB0.

Figure 1.1 The two spin states m= Âą1/2 of a nucleus with I= 1/2 in the static magnetic field B0.
The corresponding magnetic moments Οι and Οβ make precession about B0 in the z direction
with Larmor frequency ω0 = γB0.
Compared to internal magnetic fields such as dipolar field, the external magnetic field is million
times bigger, and the Zeeman interaction is a much stronger interaction. The local field Bloc is
divided into secular parts (Blocz) along B0 direction that commute with B0 and non-secular part
(Blocp) perpendicular to B0 direction that does not commute with B0. The secular part of the
Hamiltonian commutes with Zeeman Hamiltonian and will affect the energy level correction to
the first order, while the non-secular part does not commute with Zeeman Hamiltonian and is
3

thousand times smaller, thus it is not considered. This is Zeeman truncation or secular
approximation. [2, 3]
1.1.2 Radio frequency (RF) B1 field interaction
In NMR experiments, we apply radiofrequency (RF) pulse that produce time dependent B1 field
to the spin system. B1 field is oscillating magnetic field. There are two components, the resonant
B1res and the non-resonant B1non-res part.
1
𝐁𝟏𝐫𝐞𝐬 = B1 (cos(ωt)𝐱 − sin(ωt) 𝐲)
2
𝐁𝟏𝐧𝐨𝐧−𝐫𝐞𝐬 =

1.10

1
B (cos(ωt)𝐱 + sin(ωt) 𝐲)
2 1

1.11

Where x and y are the x and y direction’s unit vector respectively.
Only the 𝐁𝟏𝐫𝐞𝐬 part affects the spin states because 𝐁𝟏𝐫𝐞𝐬 part rotates about B0 clockwise in xy plane,
the same direction as the spin does while 𝐁𝟏𝐧𝐨𝐧−𝐫𝐞𝐬 rotates about B0 counter clockwise.
The B1 field produced by a 90X pulse is:
𝐁𝟏 = B1 (cos ωt) 𝐱 = B1 (cos 2πνt) 𝐱

1.12

Where ν = the frequency of the 90X pulse.
With B1 field, the spin will experience a torque 𝐓 = 𝐌 × 𝑩𝟏 , and rotates about B1 field with
frequency of ÎłB1, which is Rabi frequency, and the precession is Rabi precession, determined by
the cross product or right hand rule. If M is in z direction, and B1 is in x direction, then from 𝐓 =
𝐌 × 𝑩𝟏 = 𝒛 × 𝐱 = 𝐲, the M will process toward y direction. The right hand rule would be
placing the four fingers of the right hand along M (z- axis), followed by curving the four fingers

4

toward B1 direction (x- axis), and the direction the thumb is pointing to is the direction that the
M is rotating to (which is y- axis direction in this example).
The flip angle rf of M generated by B1 field is θ = ω1 τp = γB1 τp , where τp is the duration of
the pulse. If M is originally along z- axis, for a 90X pulse, = 90°, and M rotates 90° about x
axis to the y- axis; for a 180X pulse, = 180°, and M rotates 180° about x- axis to the z- axis.
The Hamiltonian for X pulse in rotating frame (see rotating frame section) is
̂ RF = −γB1 Îx
H

1.13

̂ RF is time independent in rotating frame and B1 appears static.
H

Figure 1.2 Vector representation of M precession with 90X (a) and 180X (b) pulse.[4]

5

1.1.3 Rotating frame
In laboratory frame, when placed in static magnetic field B0, M rotates about B0 with rapid
frequency ÎłB0 about hundreds of MHz. RF-pulses introduce B1 field and M rotates about B1 field
with a much slower precession frequency γB1. B1 rotates about B0 with frequency ωRF.[4, 5] The
motion of M is complicated in laboratory frame for NMR experiments and it is not obvious to
analyze the effect of B1 field on the motion of M. It is very helpful to consider the motion of M
in rotating frame. The frequency of the rotating frame is ωRotF. In rotating frame, the x-y plane
rotates about B0 with ωRotF set as the transmitter frequency. When transmitter frequency is set at
B1 precession frequency ωRF, B1 will appear static in the rotating frame.
In the rotating frame, the apparent precession frequency of M would be (ω0 - ωRotF), where ω0 is
the Larmor precession frequency and ω0 = γB0. Then the resonance offset Ω = ω0 - ωRotF, and the
offset field is Bo.f. = Ω/γ = (ω0 - ωRotF)/γ according to the relation of ω= γB.[6] Bo.f. is along z-axis
and is the reduced magnetic field in the rotating frame. When we set transmitter frequency ω
close to ω0 and ωRotF the same as ω, Bo.f would be close to zero, then the B1 field would be
dominant relative to Bo.f. M would majorly rotate about B1 field in the rotating frame. See
example in Figure 1.2.
1.1.4 Chemical shift interaction
Chemical shift interaction arises from electron behavior in magnetic field B0. There are electrons
around the nucleus and within the chemical bond in a molecule. When place the molecule in
external magnetic field B0, the electron motions react to B0 and move under the Lorentz force
from B0. This type of electron movement produces a secondary magnetic field that contributes to
the total field that the spin feels and thus affects the resonance frequency. Shielding interaction

6

(or chemical shift interaction) is the interaction of the nucleus with the secondary field that the
electrons produced. Shielding field decreases the magnitude of B0 experienced by a nucleus in
most cases. Chemical shift is the frequency shift the shielding interaction causes in the spectrum.
The shielding field can be in any direction because of various molecular orientations relative to
B0. However, only the B0 direction component is relevant according to secular approximation.
The Hamiltonian for this chemical shift interaction is,
̂ CS = γ𝐈̂. 𝛔. 𝐁𝟎
H

1.14

When B0 is in z direction,
̂ CS = −γћB0 𝛔. 𝐈̂
H

1.15

Where 𝐈̂ is the spin operator; σ is the shielding tensor. The electron distribution around the
nucleus interested is generally not spherically symmetric, and the size of the shielding is
dependent on the molecular orientation relative to B0. The shielding tensor σ is associated with
the principal axis frame (PAF), which has three axes of xPAF, yPAF and zPAF with corresponding
values of σxx, σyy and σzz (see Figure 1.3 a). The three values are also the three principal chemical
shifts in the chemical shift anisotropy (CSA) powder pattern. The orientation of PAF is
dependent on the electron cloud orientation relative to the B0 and is certain for a certain molecule.
The chemical shift is:
σ = σxx cos 2 ι + σyy cos2 β + σzz cos2 γ

1.16

Where ι, β and γ are the Euler angles and corresponds to the angles between B0 and the three
PAF axes (see figure 1.3 b).

7

(a)

(b)

Figure 1.3 Shielding tensor (red) relative to the B0 field in PAF with principal axes values of σxx,
σyy and σzz (a), where  is the angle between the z-axis of PAF and B0, and  is the angle
between the x-axis of PAF and the projection of B0 on the x-y plane of PAF. PAF associated
Euler angles, ι, , and  with respect to B0 field (b). [7]
For liquid which has rapid molecular tumbling motion and solids under magical angle spinning
(MAS) (see MAS section), the chemical shift orientation dependence (CAS) is averaged out and
isotropic chemical shift is observed. [7, 8].
σiso =

1
(σ + σyy + σzz )
3 xx

1.17

Without rapid molecular tumbling or MAS for a peptide or protein, a powder pattern is observed
(Figure 1.7 c). The total Hamiltonian for chemical shift contains isotropic and anisotropic and is:
1
̂ CS = σiso γћB0 + δCS ћ(3 cos2 θ − 1 − ηCS sin2 θ cos( 2ϕ))
H
2

1.18

Where the first part is for isotropic chemical shift and the second part is for the anisotropic
chemical shift; σiso is the isotropic chemical shift, 𝛿𝐶𝑆 is the reduced anisotropy, 𝜂𝐶𝑆 is the
asymmetry parameter, θ and ϕ (see Figure 1.3 a) are the polar angles of B0 in PAF. 𝛿𝐶𝑆 =

8

−γB0 (σ𝑧𝑧 − σ𝑖𝑠𝑜 ) and ηCS = −

σyy −σxx
σzz −σiso

.[9] Figure 1.4a displays a powder pattern of

13

CO

nucleus. A powder pattern contains all the possible orientations of the molecule and the intensity
reflects the population of the molecule with that orientation. σzz is the most shielded chemical
shift principal component, and σxx is the least shielded chemical shift principal component.
Figure 1.4 b shows the PAF for

13

CO. zPAF with principal chemical shift value of σzz is

perpendicular to the Cι-CO-N peptide bond plane, yPAF with principal chemical shift value of σyy
is along the CO bond direction, and xPAF with principal chemical shift value of σxx is
perpendicular to CO bond in the peptide bond plane. [8]

(b)

(a)

Figure 1.4 (a) CSA powder pattern of 13CO with the three principal chemical shift values of xx
= 247 ppm, yy = 176 ppm and zz = 99 ppm. (b) The PAF of

13

CO in peptide or protein

backbone with zPAF perpendicular to the peptide bone C-CO-N plane, yPAF alone the CO bond
direction, and xPAF perpendicular to the CO bond in the C -CO- N plane.
1.1.5 J coupling interaction
J coupling is also scaler coupling. It is the indirect dipole- dipole coupling between two nuclear
spins connected through chemical bonds. The coupling depends on the interaction between the

9

nuclear spins and the bonding electron spins.[10] The Hamiltonian for J coupling between spin j
and k is
̂ J = 2π𝐉jk 𝐈̂𝐣 ∙ 𝐈̂𝐤
H

1.19

Where Jjk is the J-coupling tensor, 𝐈̂𝐣 and 𝐈̂𝐤 are the nuclear spin operator for the jth and kth nuclei.
From the Hamiltonian, the J-coupling is only dependent on the molecular structure and
independent on magnetic field.[11] Therefore, J coupling remains a constant with differing
magnetic field. J coupling tensor becomes a number in isotropic liquids. In liquid state NMR, the
line width is usually narrow with a few hertz. Therefore, J coupling is an important interaction
because J coupling magnitude is about 10 Hz for a three bond 1H-1H J coupling of 1H-C-C-1H,
and about 140 Hz for 1H-C J coupling of C-H bond.[12] However, in solid state NMR, the line
width is generally about a few hundred hertz, and J- coupling will be within the broad linewidth.
[13]
1.1.6 Dipolar coupling interaction
Dipolar coupling is the direct magnetic dipole- dipole interaction through space. It arises from
the interaction of one nuclear spin with the magnetic field generated at its site from another
nuclear spin nearby. The strength of the interaction depends on the inter-nuclear distance r and
the angle  between the inter-nuclear distance vector and the magnetic field B0 along the z-axis
(shown in Figure 1.5).
The secular Hamiltonian for homonuclear dipolar coupling is

̂ DII = −
H

Îź 0 Îł2 1
ћ 3 (3 cos 2 θ − 1)(3Î1z Î2z − 𝐈̂𝟏 ∙ 𝐈̂𝟐 )
4π r12
2

10

1.20

Where I1 and I2 are two different spin of the same nucleus type, Âľ0 is the permeability of free
space, 𝐈̂𝟏 and 𝐈̂𝟐 are the vector operators of spin 1 and 2,  and r are the angles shown in Figure
1.5. The Hamiltonian for heteronuclear dipolar coupling for spin I and S of different nucleus type
is:
̂ DIS = −
H

Îź 0 ÎłI ÎłS 1
ћ 3 (3 cos 2 θ − 1)(2Îz Ŝz )
4π rIS
2
Îź

For dipolar coupling, 4π0 ћ

ÎłI ÎłS
r3IS

1.21

is dipolar coupling constant in unit of rad/s, and heteronuclear

dipolar coupling constant in unit of Hz is
𝑑=

Îź0 ÎłI ÎłS
1
μ0 ℎγI γS
ћ 3 ×
=
3
4π rIS
2𝜋 16π3 rIS

1.22

For 13C-31P dipolar coupling, d=12250/r3, for 13C-15N dipolar coupling, d=3066/r3, while for 13C2

H dipolar coupling, d=4662/r3, where d is in unit of Hz and r is in unit of angstroms.[14-16]

Figure 1.5 Dipolar coupling between nuclear spin I and S with inter-nuclear distance (r) and
azimuthal angle ().  is the angle between the inter-nuclear vector and the magnetic field B0
which is alone z axis.

11

̂ Z is the Zeeman Hamiltonian for homonuclear dipolar
𝐈̂𝟏 ∙ 𝐈̂𝟐 = Î1𝑥 Î2𝑥 + Î1𝑦 Î2𝑦 + Î1𝑧 Î2𝑧 . H
̂ Z = −γћB0 (Î1z + Î2z ), and following calculation
coupling. For homonuclear dipolar coupling, H
̂ Z commutes with 𝐈̂𝟏 ∙ 𝐈̂𝟐 .
can be done to see whether H
̂ Z , 𝐈̂𝟏 ∙ 𝐈̂𝟐 ] = −γћB0 [(Î1z + Î2z ), 𝐈̂𝟏 ∙ 𝐈̂𝟐 ]
[H
= −γЋb0 [(Î1z + Î2z ), ( Î1𝑥 Î2𝑥 + Î1𝑦 Î2𝑦 + Î1𝑧 Î2𝑧 )] = 0
̂ Z = −ћB0 (γI Îz + γS Îz ). The
The Zeeman Hamiltonian for heteronuclear dipolar coupling is H
[γI Îz + γS Îz , 𝐈̂ ∙ 𝐒̂] ≠ 0 because the two different γ present.[2, 17]
̂ Z for homonuclear dipolar coupling, but does not commute with
Therefore, 𝐈̂ ∙ 𝐒̂ commutes with H
̂ Z for heteronuclear coupling. According to secular approximation or Zeeman truncation, the
H
Hamiltonian is truncated more for heteronuclear dipolar coupling than homonuclear dipolar
coupling.
1.1.7 Quadrupolar coupling interaction
Quadrupolar coupling interaction exists in quadrupolar nuclei with spin quantum number I > 1/2
due to non-spherical charge distribution around the nucleus. Figure 1.6 shows the nucleus charge
distribution and the corresponding quadrupole moment.

12

Figure 1.6 Prolate (a) and Oblate (b) charge distribution of quadrupolar nucleus and the
corresponding quadrupole moment shown in (c) and (d) respectively. Prolate moment (c) is
positive and Oblate moment (d) is negative. [2, 18]
The quadrupolar nuclei has electric quadrupole moment, which interacts with the electric field
gradient produced by the distribution of other nuclei and the electrons near the nucleus at the
nucleus site, and this is known as quadrupolar coupling.[18] The coupling strength depends on
the magnitude of the quadrupole moment and the electric field gradient strength. The electric
quadrupole moment is eQ, where e is the proton charge and Q is the quadrupole moment specific
to nucleus type. Quadrupolar interaction is a relatively stronger interaction (~170 kHz for
aliphatic C- 2H) and affects the energy levels of the nuclear spin like other magnetic interactions
previously discussed. The secular quadrupolar interaction Hamiltonian is
̂D =
H

eQeq
1
1
(3 cos 2 θ − 1 − ηQ sin2 θ cos(2∅)) (3Îz2 − I(I + 1))
2I(2I − 1)ћ 2
2

1.22

Where e = the proton charge, Q = the quadrupole moment specific to nucleus type, q is
associated with electric field gradient tensor, I is the nucleus spin quantum number, θ and ϕ are

13

the polar angles of magnetic field B0 in PAF, ΡQ is the asymmetry parameter of the electric field
tensor, and Îz is the z-component of the spin operator.
eQeq
ћ

𝜒=

= the quadrupolar coupling constant (QCC), in unit of rad/s. QCC in Hz is
eQeq 1
eQeq
=
ћ 2π
h

1.23

The term (3 cos2 θ − 1 − ηQ sin2 θ cos(2∅)) is from the orientation dependence of the electric
field gradient tensor.
For 2H nucleus, spin quantum number I =1, and χ ≈ 170 kHz in aliphatic C-D bond[19]. Because
deuterated lipids and cholesterol have been widely used to study membrane structure and
dynamics, we are going to analyze the effect of orientation on the observed resonance frequency
of 2H. Since ΡQ is ~ 0 for aliphatic C-D bond due to the approximate uniaxiality of electron
density in the σ bond, and Îz2 |𝐼, 𝑚 >= 𝑚2 |𝐼, 𝑚 >, the 2H quadrupolar energy is written as,
EQ =

eQeq
1
1
(3 cos 2 θ − 1 − ηQ sin2 θ cos(2∅)) (3m2 − I(I + 1))
2I(2I − 1)ћ 2
2
π

= 4 χћ(3 cos2 θ − 1)(3m2 − 2)

1.24

Where m= -1, 0 and +1 for 2H. The allowed transitions are m = ±1 in NMR spectroscopy.
Therefore, there are two allowed transitions, m = +1  m = 0, and m = 0  m = -1. Next, we
are going to discuss orientation dependence of EQ by analyzing a few typical θ value, and see
how it affects the 2H transition frequency and the resulting 2H spectrum.
π

π

Example (1): When  = 0, EQ = 4 χћ(3 cos 2 θ − 1)(3m2 − 2) = 2 χћ(3m2 − 2). For the three
π

π

possible values of m: +1, 0, and -1, the corresponding EQ would be 2 χћ, −πχћ and 2 χћ. Besides
the quadrupolar interaction, there is also Zeeman interaction, which is the strongest interaction in

14

NMR. We assume the Zeeman energies for m = +1, 0, and -1 are – Ez, 0 and + Ez. For transition
π

m = +1 →m = 0, ∆E = E+1→0 = Ez − πχћ − 2 χћ = Ez −

3π
2

χћ. From ∆E = hν, the transition

3

π

frequency ν+1→0 = νz − 4 χ. Similarly, for transition m = 0 →m = -1, ∆E = E0→-1 = Ez + 2 χћ −
(− πχћ) = Ez +

3π
2

3

χћ , thus the transition frequency ν0→−1 = νz + 4 χ . νz is the Larmor

frequency of 2H in absence of quadrupolar interaction and νz = Ez / h. When the spectrometer
transmitter frequency is set at the 2H Larmor frequency νz, there will be two discrete signals
3

observed in the 2H spectrum, one at − 4 χ corresponding to transition of m = +1 →m = 0, and the
3

other one at 4 χ corresponding to transition of m = 0 →m = -1. The spectrum frequency axis is in
unit of Hz. This spectrum for this specific angle is shown in Figure 1.7 a.
π

π

Example (2): When  = 54.7, EQ = 4 χћ(3 cos 2 θ − 1)(3m2 − 2) = 2 χћ(3m2 − 2) = 0
because (3 cos 2 θ − 1) = 0. EQ = 0 regardless of m value. ∆E = E0→-1 = Ez for transition m = +1
→m = 0 and ∆E = E0→-1 = Ez for transition m = 0 →m = -1. Therefore, ν+1→0 = νz and ν0→−1 =
νz . When the transmitter frequency is set at the 2H Larmor frequency νz, both transitions will
give signal at the same frequency and appears at 0 Hz in the spectrum. Figure 1.7 b shows the
spectrum for this specific angle.
π

π

Example (3): When  = 90， EQ = 4 χћ(3 cos2 θ − 1)(3m2 − 2) = − 4 χћ(3m2 − 2). For the
π

π

π

three possible values of m: +1, 0, and -1, the corresponding EQ would be − 4 χћ, 2 χћ and − 4 χћ.
π

π

For transition m = +1 →m = 0, ∆E = E+1→0 = Ez + 2 χћ + 4 χћ = Ez +

3π
4

χћ. From ∆E = hν, the

3

transition frequency ν+1→0 = νz + 8 χ, where νz = Ez /h. Similarly, for transition m = 0 →m = π

π

1, ∆E = E0→-1 = Ez − 4 χћ − ( 2 χћ) = Ez −

3π
4

χћ, thus the transition frequency ν0→−1 = νz −

15

3
8

3

χ. In this example, there will be two discrete signals observed in the 2H spectrum, one at χ
8

3

corresponding to transition of m = +1 →m = 0, and the other one at − 8 χ corresponding to
transition of m = 0 →m = -1. Figure 1.7 c shows the spectrum for this specific angle.
For a static sample that  could have all possible values, the observed 2H spectrum would have a
powder pattern with doublet (Figure 1.7 d). The doublet corresponds to the two possible
transitions for the three spin states m=1, 0 and -1. The transitions are m = +1 →m = 0 and m = 0
3

→m = -1 for = 90o. The splitting of the two horns in the spectrum is 4 χ, which would be 127
kHz for aliphatic C-D bond.

16

Figure 1.7 The orientation dependence of static 2H spectra. Discrete lines are observed for the
allowed two transitions (m = +1  m = 0, and m = 0  m = -1) with (a)  = 0, (b)  = 54.7
and (c)  = 90 where  is the angle between the C – 2H bond and B0 field. (d) The quadrupolar
powder pattern for the allowed two transitions for all possible s in static sample.[2, 18]

1.1.8 Magic angle spinning (MAS)
17

Magic angle spinning (MAS) is a widely used technique to achieve high-resolution spectrum in
solid state NMR.[1, 20, 21] In liquid sate NMR, the observed high-resolution spectrum with
narrow peaks is because of rapid molecular tumbling in solution. The rapid motion in liquids can
average out the orientation dependence of CSA and dipolar coupling. In solid state NMR, the
rapid molecular reorientation is absent, so the peaks are generally broad due to anisotropic
interactions like CSA and dipolar coupling. In order to get high resolution spectrum in solids,
MAS was invented.[20] The angle between the rotor spinning axis and the external magnetic
field B0 equals magic angle 54.7° (Figure 1.8). In MAS, the sample-containing rotor spins at
speed ranging from a few to tens of kHz. Rapid MAS could average out chemical shift
anisotropy and only leave isotropic chemical shift observed. Spinning side bands shows up with
slow spinning speed in addition to the isotropic chemical shift (Figure 1.9).

Figure 1.8 Magic angle spinning (MAS) for 13C- 2H inter-nuclear vector, the angle Îą is the angle
between the external magnetic field B0 and sample- spinning axis. When the angle Îą is fixed at
54.7°, the sample spinning is called magic angle spinning.[18] θ and β is the angle between 13C2

H distance vector and B0 and spinning axis respectively.

18

Figure 1.9 The effect of different spinning speed on the observed spectrum. For this example,
the isotropic chemical shift is set at 0 Hz, the CSA is 5 kHz, and the asymmetry is 0. When the
spinning speed is slow, there are spinning sidebands spaced at spinning frequency; when the
spinning speed is fast enough to overcome the CAS, only isotropic chemical shift is observed
with high intensity.[1] And the figure is from reference 1.
Another important effect of MAS is to average out dipolar coupling which also broadens
spectrum line width. Take 13C-2H dipolar coupling for example, the 13C-2H distance geometry is
shown in Figure 1.8, ι is 54.7° (magic angle) and is the angle between the rotor spinning axis
and B0. θ and β is the angle between

13

C- 2H distance vector and B0 and spinning axis

19

respectively. The θ varied with time when the sample spins about the spinning axis. Over each
rotor period, the average of (3cos2θ–1) becomes 0.[1]
< 3cos2 θ(t) − 1 > =

1
(3cos 2 α − 1)(3cos2 β − 1) = 0
2

1.25

Where θ, ι, and β are defined in Figure 1.8.
1.1.9 Rotational echo double resonance (REDOR)
Rotational echo double resonance (REDOR) solid state NMR has been widely used to study the
inter-nuclear distance by recovering the heteronuclear dipolar coupling under MAS. Dipolar
coupling has a distance dependence of 1/r3. From REDOR, we can get both dipolar coupling and
inter-nuclear distance r. REDOR was developed by Terry Gullion and Jacob Schaefer and
originally illustrated with

13

C- and

15

N labeled alanine.[22] Typical REDOR pulse sequence is

shown in Figure 1.10 with example spins of 13C (detect) and 15N (dephasing).

Figure 1.10 Pulse sequence for 13C - 15N REDOR.

20

REDOR experiments require a three- channel spectrometer and a triple resonance probe for
peptide and protein studies. For

13

C -15N REDOR, the three channels are 1H,

13

C and

15

N. 90°

pulse is initially applied to 1H nuclei to generate a transverse magnetization by rotating the 1H
magnetization from B0 direction to the x-y plane; cross polarization (CP) is applied to both 1H
and

13

C channels to generate

13

C transverse magnetization by transferring 1H transverse

magnetization to 13C nucleus.
Cross polarization (CP) is a commonly used method to study rare spins like

13

C with low

abundance and very long relaxation times, which causes poor signal/noise ratio and requires long
gaps between scans respectively. A decent spectrum with good signal signal/noise ratio would
take long time due to thousands of scans needed in solid state NMR. CP could solve the
problems by transferring the magnetization from the nearby network of abundant spins like 1H to
the rare spins like

13

C. The process is mediated by 1H -13C dipolar coupling and can be

understood through doubly rotating frame.[1] In one rotating frame, the 1H B1 field is static;
while in the other rotating frame, the 13C B1 field is static. In both rotating frames, it assumes B1
field is the only magnetic field considering no resonance offset present. The 1H and 13C contact
pulses during CP must meet the Hartmann-Hahn matching condition:
Îł1H B1 ( 1H) = Îł13C B1 ( 13C)

1.26

Where B1(1H) and B1(13C) is the magnitude of 1H B1 field and 13C B1 field respectively. Equation
1.26 applies when there is no resonance offset. However, in real samples, there is resonance
offset Bo.f.. Then the matching conditions become
Îł1H Beff ( 1H) = Îł13C Beff ( 13C)

1.27

Where Beff is the magnitude of Beff and Beff = B1 + Bo.f. Beff = √(𝐁𝟏 + 𝐁𝐨.𝐟. )2.

21

The transition energy gaps for the two hetero-nuclei are equal in the doubly rotating frame. The
energy required for a

13

C α→β transition is from the energy released by a nearby 1H β→α

transition. There is energy redistribution between 1H and

13

C spins, but the total energy of the

spin system is conserved.
The Hartmann-Hahn matching condition described in equation 1.26 and 1.27 is for static sample.
However, in REDOR, MAS is used and MAS with CP complicates the Hartmann-Hahn
matching condition because MAS affects dipolar coupling by creating time dependent
orientation dependence <3 cos2θ(t) - 1>. The strength of the dipolar coupling depends on the
orientation between inter-nuclear distance vector and the B0 field as well as the magnitude of Îł.
For the 1H –13C dipolar coupling, the largest 13CO – 1H dipolar coupling is ~ 4 kHz in a peptide
because the closest distance between 1H and a labeled 13CO is ~ 2Å, which is between 13CO and
1

H in the peptide bond. In REDOR with 10 kHz MAS, the largest 4 kHz

13

CO – 1H dipolar

coupling is supposed to be averaged out. However, the 1H – 1H dipolar coupling is not averaged
out with 10 kHz MAS. The 1H – 1H dipolar coupling is much stronger due to 4 times bigger γ
compared to 13CO, and is typically 10-50 kHz.[23] The 1Hs are dipolar coupled as a network in
peptide, and there are rapid spin states exchange (α↔β transition) between 1Hs through the
homonuclear dipolar coupling. The spin states exchange rate is roughly equal to the 1H – 1H
homonuclear dipolar coupling. Therefore, the 1Hs will change its spin state over each rotor
period with a rate comparable or even faster than the 10 kHz MAS speed. Then the

13

CO – 1H

dipolar coupling is not averaged to zero over each rotor period because the heteronuclear dipolar
coupling is disrupted by the fast 1H spin states exchange. This is also the reason that
CP can be achieved through the

13

13

CO – 1H

CO – 1H heteronuclear dipolar coupling under 10 kHz MAS.

The Hartmann-Hahn matching condition under MAS is

22

γ1H Beff ( 1H) = γ13C Beff ( 13C) + nωR

1.28

Where n= 0, Âą1, Âą2 and represents the nth spinning sideband in the

13

C spectrum, and ωR is the

MAS frequency.[24] From equation 1.28, γ1H Beff ( 1H) ≠ γ13C Beff ( 13C), and the total energy of
the spins are not conserved.
MAS disrupt the

13

CO – 1H dipolar coupling by averaging out the orientation dependence.

Besides, there is distribution of resonance-offset frequency due to different molecular orientation,
chemical environment and dipolar couplings. Therefore, a ramp field is applied to 13C during CP
to increase the CP transfer efficiency. In REDOR, the maximum CP

13

C signal is achieved by

optimizing the field strength and ramp of 13C pulses.
Following CP, rotor synchronized π pulses are applied to
period except the last rotor period, and to

15

13

C channel at the end of each rotor

N channel at the middle of each rotor period. Two

types of signals are detected, the control signal S0 and the reduced signal S1. To obtain S0 signal,
13

C π pulses are applied and no 15N π pulses. To get S1, 15N π pulses are applied in the middle of

each rotor period along with 13C π pulses applied at the end of each rotor period. To understand
how MAS and π pulses affect dipolar coupling over each rotor period, we can look at the 13CO –
15

N heteronuclear dipolar coupling Hamiltonian:

̂ DCN = − μ0 ћ γC3γN 1 (3 cos 2 θ − 1)(2Îz Ŝz )
H
4π
r
2

1.29

CN

Where Îz and Ŝz are the z-component of spin operator for 13C and 15N respectively.
The 13CO – 15N dipolar coupling depends on spatial part, and 13C and 15N spin parts. MAS affect
the spatial part, and π pulses affect the spin parts. MAS average out the

13

CO –

coupling over each rotor period. π pulses change the sign of dipolar coupling.

15

13

N dipolar

C π pulses

change the sign of magnetic moment of the observed spins (13C), while 15N π pulses change the

23

sign of magnetic moment of the dephasing spins (15N) and thus alter the sign of the dipolar field
experienced by the observed spin (13C).
In S0,

13

C π pulses can refocus the isotropic chemical shift and average out chemical shift

anisotropy by MAS. Full

13

C signal is observed, and the dipolar coupling is averaged to zero

(Figure 1.11). In S1 experiment, the dipolar coupling is reintroduced, and results in reduced

13

C

signal S1 (Figure 1.12). The reintroduced dipolar coupling causes phase accumulation of the
magnetization over each rotor period. S1/ S0= cos (ϕ). 𝜙 =
1

𝜋

𝑁𝑐 𝑇𝑟 𝑑
𝜋

√2 sin 2𝛽 sin 𝛼 ;

𝑆1
𝑆0

=

2𝜋

∫ ∫ cos ∅ sin 𝛽 𝑑𝛼𝑑𝛽, where Nc is number of rotor periods, Tr is rotor period, d is dipolar
2𝜋 0 0
coupling in rad/s, and figure 1.8 shows the angle of θ and β. The observed S1 signal is reduced
because of all the possible θ and β values in the sample. The dephasing is given by the equation
ΔS/S0 = (S0 - S1)/S0 and is dependent on a dimensionless parameter λ = d×τ, where d is the
dipolar coupling, τ is the dephasing time.

13

C –

15

N dipolar coupling d =

Îź0 hÎł13C Îł15N
16π3 r3IS

=

3066
r3CN

,

where d is in unit of Hz and internuclear distance r is in unit of Å. d can be obtained by
simulating the dephasing buildup with SIMPSON program.[25] Once d is known, the internuclear distance can be calculated.

24

Figure 1.11 Diagram of heteronuclear dipolar coupling evolution over rotor period for S 0
experiment in REDOR. The + and – sign represent positive and negative dipolar coupling
respectively. MAS represents the dipolar coupling spatial dependence over each rotor period; C
spin represents the observing spin operator and π pulse changes the sign of dipolar coupling; S0
represents the overall effects from MAS and C spin π pulses on dipolar coupling over each rotor
period. As we can see, the dipolar coupling for S0 is averaged to zero over each rotor period.

25

Figure 1.12 Diagram of heteronuclear dipolar coupling evolution over rotor period for S1
experiment in REDOR. The + and – sign represent positive and negative dipolar coupling
respectively. MAS represents the dipolar coupling spatial dependence over each rotor period; C
spin and N spin represent the observing and dephasing spin operator respectively and π pulse
changes the sign of dipolar coupling. S1 represents the overall effects from MAS, C spin
(detecting) and N (dephasing) spin π pulses on dipolar coupling over each rotor period. As we
can see, the dipolar coupling for S1 is re-introduced and is nonzero over each rotor period.

26

1.1.10 Quadrupolar echo (QUECHO)
In 2H NMR experiment, the spectrum lines are broad due to quadrupolar coupling. Broad lines
generally have rapid decaying FID. Pulse ringing-down is much stronger signal relative to the
weak sample signal, thus prevents the measurement of sample signal until a short time (dead
time) after pulse.[1] For solid 2H NMR experiment with broad NMR resonance frequencies, the
proportion of signal loss during the dead time is significant and is overcome by using quecho
pulse sequence (shown in Figure 1.13) which refocuses the time evolution of spins.

Figure 1.13 “Quecho” pulse sequence. Theoretically,  = 1.
In the sequence, the first π/2 pulse is used to generate transverse magnetization, a sencond π/2
pulse is used to refocus the time evolution of spins and quadrupolar coupling. The two π/2 pulse
must be out of phase.  is the time interval between the first and the second π/2 pulse. 1 is the
time interval after the second π/2 pulse and before acq. When the time interval 1 = , the echo
appears with its maximum intensity. The can be understood through time evolution of density
operator 𝜌̂(𝑡).
The first π/2 pulse generates M along y axis, density operator at time 0 is 𝜌̂(0) = 𝐼̂𝑦 . The time
evolution of density operator is[17]
̂ (𝑡)𝜌̂(𝑡)𝑈
̂ −1 (𝑡)
𝜌̂(𝑡) = 𝑈

1.30

̂ (𝑡) is the time evolution operator.
Where 𝑈

27

𝜋

𝜋

̂ (2𝜏) = 𝑒 −𝑖𝐻̂𝑄 𝜏 𝑒 𝑖 2 𝐼̂𝑦 𝑒 𝑖𝐻̂𝑄 𝜏 = 𝑒 −𝑖𝜔𝜃 (3𝐼̂𝑧2 −𝑰̂2 )𝜏 𝑒 𝑖 2 𝐼̂𝑦 𝑒 𝑖𝜔𝜃 (3𝐼̂𝑧2 −𝑰̂2)𝜏
𝑈
2

= 𝑒 −𝑖𝜔𝜃 (3𝐼̂𝑧 −𝑰̂
2

= 𝑒 −𝑖𝜔𝜃 (3𝐼̂𝑧 −𝑰̂
2

2 )𝜏

2 )𝜏

𝜋

̂

2

𝑒 𝑖 2 𝐼𝑦 𝑒 𝑖𝜔𝜃 (3𝐼̂𝑧 −𝑰̂
2

𝑒 −𝑖𝜔𝜃 (3𝐼̂𝑥 −𝑰̂

2 )𝜏

2 )𝜏

𝜋

𝜋

̂

𝜋

̂

𝑒 −𝑖 2 𝐼𝑦 𝑒 𝑖 2 𝐼𝑦

̂

𝑒 𝑖 2 𝐼𝑦

1.31

𝜋
̂2 )𝜏 𝑖 𝐼̂𝑦

= 𝑒 𝑖𝜔𝜃 (3𝐼̂𝑦 −𝑰

𝜋

𝑒

2

𝜋

̂
̂
by using 1̂ = 𝑒 −𝑖 2 𝐼𝑦 𝑒 𝑖 2 𝐼𝑦 and 𝑰̂2 = 𝐼̂𝑥2 + 𝐼̂𝑦2 + 𝐼̂𝑧2 .

̂ (2𝜏)𝐼̂𝑦 𝑈
̂ −1 (2𝜏)
Then 𝜌̂(2𝜏) = 𝑈
2

𝜋
𝜋
̂2 )𝜏 𝑖 𝐼̂𝑦 ̂ −𝑖 𝐼̂𝑦 −𝑖𝜔𝜃 (3𝐼̂𝑦2 −𝑰̂2 )𝜏
𝑒 2 𝐼𝑦 𝑒 2 𝑒

= 𝑒 𝑖𝜔𝜃 (3𝐼̂𝑦 −𝑰
2

1.32

̂2 )𝜏 ̂ −𝑖𝜔𝜃 (3𝐼̂𝑦2 −𝑰̂2 )𝜏
𝐼𝑦 𝑒

= 𝑒 𝑖𝜔𝜃 (3𝐼̂𝑦 −𝑰
= 𝐼̂𝑦

by using [3𝐼̂𝑦2 − 𝑰̂2 , 𝐼̂𝑦 ] = 0.
𝜋

̂

𝜋

̂

By replacing the second (π/2)y pulse propogator 𝑒 𝑖 2 𝐼𝑦 with (π/2)-y pulse propogator 𝑒 −𝑖 2 𝐼𝑦 , same
results will be obtained, which means the solid echo is independent of the sign of the second
pulse.
Therefore, the time evolution results of 𝜌̂(2𝜏) = 𝐼̂𝑦 is the same as 𝜌̂(0) = 𝐼̂𝑦 which is generated
by the first (π/2)x pulse in the “quecho” pulse sequence. So the spin states and the signal at time
2τ is the same as at time 0. 2τ = τ +τ1.
1.2 HIV Introduction
1.2.1 HIV virus and infection
Human immunodeficiency virus (HIV) is a kind of retrovirus that causes disease of acquired
immunodeficiency syndrome (AIDS). The size of the mature HIV virus is 110 – 128 nm.[26]

28

Globally in 2015, an estimated 36.7 million people were living with HIV, ~ 1.1 million died
because of AIDS, and the newly infected people were ~ 2.1 million. Each year from 2000 to
2015, there were ~ 28 to 40 million people living with HIV with an increasing trend, 2 to 1
million people died of AIDS related diseases with a decreasing trend, and 3 to 2 million people
got newly infected with HIV with a decreasing trend. From the start of the epidemic, there have
been ~ 78 million people infected with HIV and ~ 35 million people died from AIDS related
illness.[27] Because increasing number of people living with HIV have access to the HIV
antiretroviral therapy, the number of death is decreasing, and the newly infection number is
bigger than the death number, the total number of people living with HIV has an increasing trend.
Even though the number of AIDS death has declined due to HIV antiretroviral treatments, there
is still no vaccine available against HIV. The typical cost for HIV treatment is ~ $ 25,000
/year.[28, 29]
HIV virus enters host cell through two different pathways. One is through direct membrane
fusion between the viral and host cell membrane; the other one is through endocytosis.[30-32]
There were electron microscopy studies of the HIV viral entry process in early 1990s. There
were major steps for the direct membrane fusion pathway. First, the HIV virus binds to host cell
membrane, then the outer leaflet of HIV and host cell membranes merge, followed by a
membrane pore formation, and the entry of the HIV contents into the host cell during infection.
Figure 1.14 shows the direct membrane fusion process and relevant fusion model. This process is
pH independent and the viral entry process happens within 1-3 mins at 37 °C. The endocytic
pathway was also observed by electron microscopy and the viral entry process happens a few
mins later.[30]

29

Figure 1.14 Electron microscopy (right panel) and the relevant model (left panel) of HIV- host
cell viral entry process. (a) HIV virus binding to host cell, (b) HIV and host cell membrane
hemifusion, (c) large viral pore formation and (d) HIV viral components released into host cell.
In the model, the spikes represent HIV viral membrane protein, black triangle is the viral RNA
and black dots are other components including proteins.[30]

30

Membrane fusion between the HIV virus and the target host cell membrane is the initial process
of AIDS infection. The fusion is mediated by the viral glycoproteins gp160, which is a dimer
composed of gp120 and gp41 via non-covalent interaction.[33] gp120 is the receptor binding
protein and gp41 is the transmembrane fusion protein.[34] gp120 and gp41 dimers are assembled
as trimers and there are ~ 14 trimers on each virion by cryoelectron microscopy.[35, 36] gp120
interacts with cell surface protein CD4 and chemokine receptor CXCR4 or CCR5 sequentially.
Then gp120 moves away and gp41 is exposed to interact with host cell membrane.[32, 37]

Figure 1.15 HIV interacts with T cell surface protein CD4 and chemokine receptor CXCR4.
HIV gp120 protein interacts with T cell CD4 and CXCR4 sequentially and moves away, then
gp41 get exposed to interact with cell membrane.[38]
1.2.2 HIV gp41
Transmembrane protein gp41 catalyzes the membrane joining or fusion between the viral and
host cell membrane. A general feature of HIV gp41 is shown in figure 1.16 A. From N- to Cterminal, it is the fusion peptide (FP), N-heptad repeat (NHR), loop linker, C-heptad repeat
(CHR), the transmembrane domain (TM) and the endo-domain (Endo) or cytoplasmic domain.

31

Figure 1.16 Schematic representation (A) and partial sequences (B) of HIV-1 gp41 protein.
Colored boxes show the functional regions of gp41. Form N terminal to C terminal, FP is fusion
peptide region, NHR and CHR is N- heptad repeat and C-heptad repeat respectively, Loop is the
loop region, MPER is the membrane proximal external region, TM is the transmembrane domain,
and Endo is the endodomain or cytoplasmic domain. The ectodomain without FP is HM protein,
and the full ectodomain including FP of gp41 is FP-HM. FP containing protein could induce
more fusion.[39, 40] The amino acid sequence of HM and FP-HM (B) is color coded according
to the different domains.[39] A minimal six residues loop SGGRGG replaces the native loop and
does not affect the SHB assembly.[41, 42]
During fusion, fusion protein gp41 undergoes three major states, which are pre- fusion native
state, extended pre- hairpin state, and post fusion hairpin state (six helical bundle or SHB) (figure
1.17).[43-45] In native state before fusion, gp41 exists as trimers covered by three gp120s. Xray crystallography reveals the native trimer state of soluble trimer gp140; a near-native gp160
with N-terminal four residues of MPER and without transmembrane and cytoplasmic domains of
gp41.The electron density for FP part is weak and diffuse and is likely lack of regular secondary

32

structure.[46] gp41 interacts with the cell membrane with FP part and exists as extended prehairpin state after gp120 moves away due to receptor binding. The structure of gp41 fusion
intermediates is not clear. There is evidence to support the existence of pre-hairpin folding by
inhibited membrane fusion with treatment of CHR analog short peptide such as C34 and T20
designed to bind to NHR and prevent CHR and NHR from folding to SHB state.[47-50] The
completion of SHB formation is essential for membrane fusion pore enlargement and inhibition
of SHB formation could cause low temperature (4 °C)- arrested fusion pore quickly and
irreversibly close.[51] There is high- resolution crystal structure for the final hairpin or SHB
state that contains the CHR and NHR part.[52-54]

Figure 1.17 Three major states of gp41 during fusion: (A) pre- fusion native state where gp41 is
trimeric and non-covalently associated with gp120, (B) extended pre-hairpin state where gp41
has conformational change and interacts with host cell membrane, (C) post fusion hairpin state
(six helical bundle states).[41, 43] Figures in (B) and (C) do not show gp120 to focus on the
change of gp41.

33

(a)

(b)

Figure 1.18 X- ray crystallography of soluble gp140 (PDB: 4NCO), a near-native gp160 without
MPER and transmembrane domain of gp41 with N-terminal 4 residues of MPER and without
transmembrane and cytoplasmic domains of gp41: (a) trimer of gp140, gp120 is in yellow,
orange, and red; gp41 is in green. (b) Close view of dimer of gp120 and gp41 in the trimer of
gp140.[46]
In SHB, the NHR helices form the interior parallel trimer through hydrophobic interaction and
the CHR helices pack in to the grooves on the surface of the interior trimer in an antiparallel
orientation to the NHR helices. Figure 1.19 shows a crystal structure of gp41 ectodomain
including MPER and part of FP region.

34

Figure 1.19 Crystal structure of HIV gp41 ectodomain composed of FPPR-CHR-NHR-MPER
(Gly531 to Leu581 in blue and Met628 to Tyr681 in green; gp41 531-681) (PDB: 2X7R). The gp41
does not include the N-terminal fusion peptide (gp41 512-530). Residues are numbered
according to their positions in gp160 complex).[54]
1.2.3 HIV fusion peptide (HFP)
HIV fusion peptide is the N terminal ~ 20 residues. Figure 1.16 displays the amino acid sequence
in red. Mutations in FP region of gp41 eliminate/decrease the membrane fusion activity as well
as infectivity compared to wild type protein.[55, 56] There is more membrane fusion induced by
gp41 ectodomain including the FP region compared to gp41 ectodomain without FP region.[39,
40, 43, 57] It is very important to study HFP to help understand the HIV fusion mechanism. HFP
is reasonable substitution as model peptides for gp41 because HFP itself can cause rapid fusion
and leakage of lipid vesicles.[57-60] There are HFP structure studies in both detergent micelles
and lipid membranes. HFP is majorly helical in dodecylphosphocholine (DPC) or sodium
dodecyl sulfate (SDS) micelles studied by liquid state NMR and CD spectroscopy.[61-63] HFP
is majorly helical from I4 to L12 in DPC micelles and from I4 to A15 in SDS micelles. In
membrane lipid bilayers, there is significant helical, or β sheet structures by CD spectroscopy

35

depending on the membrane lipid compositions or peptide to lipid ratio.[64, 65] Infrared
spectroscopy shows significant helical, or β sheet or a mixture of both structures depending on
membrane lipid composition and peptide to lipid ratio.[65, 66] Solid-state NMR chemical shift
measurements and 2D 13C- 13C correlation spectroscopy give continuous β stand conformation in
the first 16 residues in model membrane that reflects approximate lipid head-group and CHOL
composition of host cell of HIV-1 virus.[67-69] β sheet structure of HFP is probably the
biological relevant structure. There is FP structure dependence on CHOL. HFP has major β sheet
structure in membrane with CHOL, and has a mixture of β sheet and ι helical structure in
membrane without CHOL.[56, 70, 71] There is a significant population of antiparallel β stand
oligomer/aggregate states by both close proximity studies via 2D

13

C-

13

C correlation

experiments and distance measurements by 13C-15N REDOR.[68] A minor fraction of parallel β
strand structure, which is at most 0.15, and a major antiparallel is confirmed by quantitative
analysis of

13

C-15N REDOR data of HFP constructs with selected labeling at specific

residues.[72] A major antiparallel structure of FP is also supported for the strand conformation
when it is in the SHB state (FP-Hairpin) by SSNMR.[73]
Studies have shown that FP membrane location correlates with fusion. There is a strong
correlation between HFP insertion depth and fusion activity by REDOR SSNMR and lipid
mixing assays.[14, 55, 74] In SSNMR experiments, a specific residue of HFP backbone is
labeled.
since

31

13

13

CO

C-31P REDOR is extensively used to measure the distance from membrane surface

P comes from the lipid head-group.

13

C-

19

F REDOR is widely used to study the

proximity to the membrane center or the middle of one leaflet, while there is 1H→19F
substitution at 16-C or 5-C respectively. It turns out that HFP V2E mutant is located at the
membrane surface, while HFP monomer inserts into a single membrane leaflet and HFP trimer

36

inserts more deeply into the membrane center. Lipid mixing data shows that the membrane
fusion rate of HFP trimer is 15-40 times higher than that of HFP monomer.[74] Thus, there is
strong correlation between HFP membrane insertion depth and fusogenicity. Although
fluorinated lipid is widely used to study the peptide location in the membrane hydrophobic core,
using fluorinated lipid has the potential to change membrane bilayer structure.[75] A recent
developed

13

C- 2H REDOR by Dr. WELIKY has been used to study peptide location in native

membrane lipid bilayer because 2H substitution of 1H in the lipid is chemically equivalent and
will not change the structure of membrane.[2, 16, 70]
Besides FP membrane location, studies have shown that Cholesterol (CHOL) also correlates with
HFP fusogenicity. Depletion of cellular CHOL reduces HIV-1 binding to cells and inhibits HIV
virus induced cell-cell fusion. HFP induced model membrane fusion studies indicate that there is
faster fusion in membranes that contain CHOL and more fusion when there is more CHOL.[60,
74, 76, 77]
Our current work focuses on structure, membrane location and membrane dynamics study of
HFP- the N- terminal 23 residues of gp41 to help understand HIV membrane fusion mechanism.
We study the peptide structure and membrane location by peptide

13

C – membrane 2H REDOR

SSNMR with deuterated phospholipid and CHOL. To understand the role of CHOL in fusion, we
compare the peptide contact to phospholipid vs CHOL. We study deuterated phospholipid and
CHOL 2H relaxation times to help understand membrane perturbation by HFP and role of CHOL
in this membrane perturbation by static 2H – NMR method.

37

REFERENCES

38

REFERENCES

1.
Duer, M.J., Solid-State NMR Spectroscopy Principles and Applications. 2002: Blackwell
Science.
2.
Xie, L., Solid State Nuclear Magnetic Rosonance Studies of Structures and Membrane
Locations of Peptides, in Chemistry. 2014, Michigan State University.
3.
Ghosh, U., Solid State NMR Studies of Structure and Dynamics of Membrane Associated
Influenza Fusion Peptides, in Chemistry. 2016, Michigan State University.
4.
Thomas C. Farrar, E.D.B., Pulse and Fourier transform NMR Introduction to theory and
methods. 1971: Academic Press.
5.
Rabi, II, N.F. Ramsey, and J. Schwinger, USE OF ROTATING COORDINATES IN
MAGNETIC RESONANCE PROBLEMS. Reviews of Modern Physics, 1954. 26(2): p. 167-171.
6.

Keeler, J., Understanding NMR spectroscopy. 2005, Wiley.

7.
Cisnetti, F., et al., Determination of chemical shift anisotropy tensors of carbonyl nuclei
in proteins through cross-correlated relaxation in NMR. Chemphyschem, 2004. 5(6): p. 807-814.
8.
Loth, K., P. Pelupessy, and G. Bodenhausen, Chemical shift anisotropy tensors of
carbonyl, nitrogen, and amide proton nuclei in proteins through cross-correlated relaxation in
NMR spectroscopy. Journal of the American Chemical Society, 2005. 127(16): p. 6062-6068.
9.
Baldus, M.W.a.M., CHAPTER 1 Introduction to biological solid-state NMR, in Advances
in Biological Solid-State NMR: Proteins and Membrane-Active Peptides, A.N. Frances
Separovic, Editor. 2014, Royal Society of Chemistry. p. P001-608.
10.
Ramsey, N.F. and E.M. Purcell, INTERACTIONS BETWEEN NUCLEAR SPINS IN
MOLECULES. Physical Review, 1952. 85(1): p. 143-144.
11.
Hahn, E.L. and D.E. Maxwell, SPIN ECHO MEASUREMENTS OF NUCLEAR SPIN
COUPLING IN MOLECULES. Physical Review, 1952. 88(5): p. 1070-1084.
12.

Coupling constants. Available from: http://sopnmr.ucsd.edu/coupling.htm.

13.
Nethercott, M.J., Sample Preparation and 2D Solid State Nuclear Magnetic Resonance
Studies of the FP-Hairpin Construct of the HIV gp41 protein, in Chemistry. 2012, Michigan
State University.
14.
Qiang, W., Y. Sun, and D.P. Weliky, A strong correlation between fusogenicity and
membrane insertion depth of the HIV fusion peptide. Proceedings of the National Academy of
Sciences of the United States of America, 2009. 106(36): p. 15314-15319.

39

15.
Ghosh, U., et al., Closed and Semiclosed Interhelical Structures in Membrane vs Closed
and Open Structures in Detergent for the Influenza Virus Hemagglutinin Fusion Peptide and
Correlation of Hydrophobic Surface Area with Fusion Catalysis. Journal of the American
Chemical Society, 2015. 137(24): p. 7548-7551.
16.
Xie, L., et al., Residue-specific membrane location of peptides and proteins using
specifically and extensively deuterated lipids and C-13-H-2 rotational-echo double-resonance
solid-state NMR. Journal of Biomolecular Nmr, 2013. 55(1): p. 11-17.
17.
Spiess, K.S.-R.H.W., Multidimensional solid-state NMR and polymers. 1994:
ACADEMIC PRESS INC.
18.
Harris, R.K., Nuclear Magnetic Resonance Spectroscopy. Longman Scientific &
Technical. 1986, England.
19.
Seelig, J., DEUTERIUM MAGNETIC-RESONANCE - THEORY AND APPLICATION
TO LIPID-MEMBRANES. Quarterly Reviews of Biophysics, 1977. 10(3): p. 353-418.
20.
Andrew, E.R., A. Bradbury, and R.G. Eades, NUCLEAR MAGNETIC RESONANCE
SPECTRA FROM A CRYSTAL ROTATED AT HIGH SPEED. Nature, 1958. 182(4650): p. 16591659.
21.
Tycko, R., Biomolecular solid state NMR: Advances in structural methodology and
applications to peptide and protein fibrils. Annual Review of Physical Chemistry, 2001. 52: p.
575-606.
22.
Gullion, T. and J. Schaefer, ROTATIONAL-ECHO DOUBLE-RESONANCE NMR.
Journal of Magnetic Resonance, 1989. 81(1): p. 196-200.
23.
Yannoni, C.S., HIGH-RESOLUTION NMR IN SOLIDS - THE CPMAS EXPERIMENT.
Accounts of Chemical Research, 1982. 15(7): p. 201-208.
24.
Meier, B.H., CROSS POLARIZATION UNDER FAST MAGIC ANGLE SPINNING THERMODYNAMICAL CONSIDERATIONS. Chemical Physics Letters, 1992. 188(3-4): p. 201207.
25.
Bak, M., J.T. Rasmussen, and N.C. Nielsen, SIMPSON: A general simulation program
for solid-state NMR spectroscopy. Journal of Magnetic Resonance, 2000. 147(2): p. 296-330.
26.
Gentile, M., et al., DETERMINATION OF THE SIZE OF HIV USING ADENOVIRUS
TYPE-2 AS AN INTERNAL LENGTH MARKER. Journal of Virological Methods, 1994. 48(1): p.
43-52.
27.

http://www.unaids.org/sites/default/files/media_asset/UNAIDS_FactSheet_en.pdf. 2016.

28.

cdc.gov.

29.

https://aidsinfo.nih.gov/guidelines.

40

30.
Grewe, C., A. Beck, and H.R. Gelderblom, HIV - EARLY VIRUS-CELL INTERACTIONS.
Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology, 1990. 3(10): p.
965-974.
31.
Miyauchi, K., et al., HIV Enters Cells via Endocytosis and Dynamin-Dependent Fusion
with Endosomes. Cell, 2009. 137(3): p. 433-444.
32.
White, J.M., et al., Structures and mechanisms of viral membrane fusion proteins:
Multiple variations on a common theme. Critical Reviews in Biochemistry and Molecular
Biology, 2008. 43(3): p. 189-219.
33.

Chan, D.C. and P.S. Kim, HIV entry and its inhibition. Cell, 1998. 93(5): p. 681-684.

34.
Rizzuto, C.D., et al., A conserved HIV gp120 glycoprotein structure involved in
chemokine receptor binding. Science, 1998. 280(5371): p. 1949-1953.
35.
Zhu, P., et al., Distribution and three-dimensional structure of AIDS virus envelope
spikes. Nature, 2006. 441(7095): p. 847-852.
36.
Wyatt, R. and J. Sodroski, The HIV-1 envelope glycoproteins: Fusogens, antigens, and
immunogens. Science, 1998. 280(5371): p. 1884-1888.
37.
Liu, J., et al., Molecular architecture of native HIV-1 gp120 trimers. Nature, 2008.
455(7209): p. 109-U76.
38.
Balter, M., AIDS Research: New Hope in HIV Disease. Science, 1996. 274(5295): p.
1988.
39.
Banerjee, K. and D.P. Weliky, Folded Monomers and Hexamers of the Ectodomain of the
HIV gp41 Membrane Fusion Protein: Potential Roles in Fusion and Synergy Between the Fusion
Peptide, Hairpin, and Membrane-Proximal External Region. Biochemistry, 2014. 53(46): p.
7184-7198.
40.
Ratnayake, P.U., et al., pH-dependent vesicle fusion induced by the ectodomain of the
human immunodeficiency virus membrane fusion protein gp41: Two kinetically distinct
processes and fully-membrane-associated gp41 with predominant beta sheet fusion peptide
conformation. Biochimica Et Biophysica Acta-Biomembranes, 2015. 1848(1): p. 289-298.
41.
Sackett, K., et al., Hairpin Folding of HIV gp41 Abrogates Lipid Mixing Function at
Physiologic pH and Inhibits Lipid Mixing by Exposed gp41 Constructs. Biochemistry, 2009.
48(12): p. 2714-2722.
42.
Tan, K.M., et al., Atomic structure of a thermostable subdomain of HIV-1 gp41.
Proceedings of the National Academy of Sciences of the United States of America, 1997. 94(23):
p. 12303-12308.
43.
Sackett, K., A. TerBush, and D.P. Weliky, HIV gp41 six-helix bundle constructs induce
rapid vesicle fusion at pH 3.5 and little fusion at pH 7.0: understanding pH dependence of

41

protein aggregation, membrane binding, and electrostatics, and implications for HIV-host cell
fusion. European Biophysics Journal with Biophysics Letters, 2011. 40(4): p. 489-502.
44.
Melikyan, G.B., Common principles and intermediates of viral protein-mediated fusion:
the HIV-1 paradigm. Retrovirology, 2008. 5: p. 13.
45.
Tamm, L.K., J. Lee, and B.Y. Liang, Capturing Glimpses of an Elusive HIV Gp41
Prehairpin Fusion Intermediate. Structure, 2014. 22(9): p. 1225-1226.
46.
Julien, J.P., et al., Crystal Structure of a Soluble Cleaved HIV-1 Envelope Trimer.
Science, 2013. 342(6165): p. 1477-1483.
47.
Kilby, J.M., et al., Potent suppression of HIV-1 replication in humans by T-20, a peptide
inhibitor of gp41-mediated virus entry. Nature Medicine, 1998. 4(11): p. 1302-1307.
48.
Kliger, Y., et al., Mode of action of an antiviral peptide from HIV-1 - Inhibition at a postlipid mixing stage. Journal of Biological Chemistry, 2001. 276(2): p. 1391-1397.
49.
Weissenhorn, W., et al., Structural basis for membrane fusion by enveloped viruses.
Molecular Membrane Biology, 1999. 16(1): p. 3-9.
50.
Furuta, R.A., et al., Capture of an early fusion-active conformation of HIV-1 gp41.
Nature Structural Biology, 1998. 5(4): p. 276-279.
51.
Markosyan, R.M., F.S. Cohen, and G.B. Melikyan, HIV-1 envelope proteins complete
their folding into six-helix bundles immediately after fusion pore formation. Molecular Biology
of the Cell, 2003. 14(3): p. 926-938.
52.
Weissenhorn, W., et al., Atomic structure of the ectodomain from HIV-1 gp41. Nature,
1997. 387(6631): p. 426-430.
53. Chan, D.C., et al., Core structure of gp41 from the HIV envelope glycoprotein. Cell, 1997.
89(2): p. 263-273.
54.
Buzon, V., et al., Crystal Structure of HIV-1 gp41 Including Both Fusion Peptide and
Membrane Proximal External Regions. Plos Pathogens, 2010. 6(5): p. 7.
55.
Freed, E.O., et al., A MUTATION IN THE HUMAN-IMMUNODEFICIENCY-VIRUS
TYPE-1 TRANSMEMBRANE GLYCOPROTEIN-GP41 DOMINANTLY INTERFERES WITH
FUSION AND INFECTIVITY. Proceedings of the National Academy of Sciences of the United
States of America, 1992. 89(1): p. 70-74.
56.
Gabrys, C.M., et al., Solid-State Nuclear Magnetic Resonance Measurements of HIV
Fusion Peptide (CO)-C-13 to Lipid P-31 Proximities Support Similar Partially Inserted
Membrane Locations of the alpha Helical and beta Sheet Peptide Structures. Journal of Physical
Chemistry A, 2013. 117(39): p. 9848-9859.

42

57.
Cheng, S.F., et al., The fusion peptide domain is the primary membrane-inserted region
and enhances membrane interaction of the ectodomain of HIV-1 gp41. Molecular Membrane
Biology, 2010. 27(1): p. 31-44.
58.
Durell, S.R., et al., What studies of fusion peptides tell us about viral envelope
glycoprotein-mediated membrane fusion. Molecular Membrane Biology, 1997. 14(3): p. 97-112.
59.
Epand, R.M., Fusion peptides and the mechanism of viral fusion. Biochimica Et
Biophysica Acta-Biomembranes, 2003. 1614(1): p. 116-121.
60.
Lai, A.L., et al., Fusion Activity of HIV gp41 Fusion Domain Is Related to Its Secondary
Structure and Depth of Membrane Insertion in a Cholesterol-Dependent Fashion. Journal of
Molecular Biology, 2012. 418(1-2): p. 3-15.
61.
Li, Y.L. and L.K. Tamm, Structure and plasticity of the human immunodeficiency virus
gp41 fusion domain in lipid micelles and bilayers. Biophysical Journal, 2007. 93(3): p. 876-885.
62.
Jaroniec, C.P., et al., Structure and dynamics of micelle-associated human
immunodeficiency virus gp41 fusion domain. Biochemistry, 2005. 44(49): p. 16167-16180.
63.
Gabrys, C.M. and D.P. Weliky, Chemical shift assignment and structural plasticity of a
HIV fusion peptide derivative in dodecylphosphocholine micelles. Biochimica Et Biophysica
Acta-Biomembranes, 2007. 1768(12): p. 3225-3234.
64.
Chang, D.K., S.F. Cheng, and W.J. Chien, The amino-terminal fusion domain peptide of
human immunodeficiency virus type 1 gp41 inserts into the sodium dodecyl sulfate micelle
primarily as helix with a conserved glycine at the micelle-water interface. Journal of Virology,
1997. 71(9): p. 6593-6602.
65.
Rafalski, M., J.D. Lear, and W.F. Degrado, PHOSPHOLIPID INTERACTIONS OF
SYNTHETIC PEPTIDES REPRESENTING THE N-TERMINUS OF HIV GP41. Biochemistry,
1990. 29(34): p. 7917-7922.
66.
Castano, S. and B. Desbat, Structure and orientation study of fusion peptide FP23 of
gp41 from HIV-1 alone or inserted into various lipid membrane models (mono-, bi- and multibilayers) by FT-IR spectroscopies and Brewster angle microscopy. Biochimica Et Biophysica
Acta-Biomembranes, 2005. 1715(2): p. 81-95.
67.
Yang, J., et al., Oligomeric beta-structure of the membrane-bound HIV-1 fusion peptide
formed from soluble monomers. Biophysical Journal, 2004. 87(3): p. 1951-1963.
68.
Qiang, W., M.L. Bodner, and D.P. Weliky, Solid-state NMR Spectroscopy of human
immunodeficiency virus fusion peptides associated with host-cell-like membranes: 2D
correlation spectra and distance measurements support a fully extended conformation and
models for specific antiparallel strand registries. Journal of the American Chemical Society,
2008. 130(16): p. 5459-5471.

43

69.
Aloia, R.C., H.R. Tian, and F.C. Jensen, LIPID-COMPOSITION AND FLUIDITY OF
THE HUMAN-IMMUNODEFICIENCY-VIRUS ENVELOPE AND HOST-CELL PLASMAMEMBRANES. Proceedings of the National Academy of Sciences of the United States of
America, 1993. 90(11): p. 5181-5185.
70.
Jia, L.H., et al., REDOR solid-state NMR as a probe of the membrane locations of
membrane-associated peptides and proteins. Journal of Magnetic Resonance, 2015. 253: p. 154165.
71.
Yang, J., C.M. Gabrys, and D.P. Weliky, Solid-state nuclear magnetic resonance
evidence for an extended beta strand conformation of the membrane-bound HIV-1 fusion peptide.
Biochemistry, 2001. 40(27): p. 8126-8137.
72.
Schmick, S.D. and D.P. Weliky, Major Antiparallel and Minor Parallel beta Sheet
Populations Detected in the Membrane-Associated Human Immunodeficiency Virus Fusion
Peptide. Biochemistry, 2010. 49(50): p. 10623-10635.
73.
Kelly Sackett, M.J.N., Zhaoxiong Zheng and David P. Weliky, Solid-State NMR
Spectroscopy of the HIV gp41 Membrane Fusion Protein Supports Intermolecular Antiparallel β
Sheet Fusion Peptide Structure in the Final Six- Helix Bundle State. Journal of molecular
biology, 2014. 426(5): p. 1077-1094.
74.
Rong Yang, M.P., ‡ Francis J. Castellino,‡ and David P. Weliky*,†, A Trimeric HIV-1
Fusion Peptide Construct Which Does Not Self-Associate in Aqueous Solution and Which Has
15-Fold Higher Membrane Fusion Rate. Journal of American Chemical Society, 2004. 126(45):
p. 14722-14723.
75.
Hirsh, D.J., et al., A new monofluorinated phosphatidylcholine forms interdigitated
bilayers. Biophysical Journal, 1998. 75(4): p. 1858-1868.
76.
Liao, Z.H., et al., Lipid rafts and HIV pathogenesis: Host membrane cholesterol is
required for infection by HIV type 1. Aids Research and Human Retroviruses, 2001. 17(11): p.
1009-1019.
77.
Yang, S.T., et al., HIV gp41-mediated membrane fusion occurs at edges of cholesterolrich lipid domains. Nature Chemical Biology, 2015. 11(6): p. 424-+.

44

Chapter 2 - Materials and methods
2.1 Materials
Protected amino acids and Wang resins were purchased from Peptides International (Louisville,
KY), Sigma-Aldrich (St. Louis, MO) and Dupont, and lipids were purchased from Avanti Polar
Lipids (Alabaster, Al). The phosphatidylcholine (PC) lipid was typically 1,2-dipalmitoyl-snglycero-3-phosphocholine (DPPC) and the phosphatidylglycerol (PG) lipid was typically 1,2dipalmitoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (sodium salt) (DPPG). Some other lipids
were also used, including 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC), 1palmitoyl-2-oleoyl-sn-glycero-3- phospho-(1'-rac-glycerol) (sodium salt) (POPG), 1,2-dioleoylsn-glycero-3-phosphocholine

(DOPC),

1,2-dioleoyl-sn-glycero-3-phospho-(1'-rac-glycerol)

(sodium salt) (DOPG), 1,2-dipalmitoyl-sn-glycerol-3-phosphocholine (DMPC) and 1,2dipalmitoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (sodium salt) (DMPG). 1,2-(dipalmitoyl2,2,2,2-d4)-sn-glycero-3-phosphocholine

(PC-d4),

1,2-(dipalmitoyl-7,7,7,7,8,8,8,8-d8)-sn-

glycero-3-phosphocholine (PC-d8) and 1,2-(dipalmitoyl-15,15,15,15,16,16,16,16,16,16-d10)-snglycero-3-phosphocholine (PC-d10) lipids were custom-synthesized by Avanti (Alabaster, Al)
using deuterated palmitic acids obtained from CDN Isotopes. Protected labeled amino acids were
obtained from Cambridge Isotopes (Andover, MA) or Sigma-Aldrich (St. Louis, MO). Other
reagents including cholesterol-2,2,3,4,4,6-d6 (Chol-d6) and cholesterol-25,26,26,26,27,27,27-d7
(Chol-d7) were typically obtained from Sigma-Aldrich (St. Louis, MO).
2.2 Peptide sequences, preparation and purification
HFP: AVGIGALFLGFLGAAGSTMGARSWKKKKKKG;
HFP_V2E: AEGIGALFLGFLGAAGSTMGARSWKKKKKKG;
HFP_L9R: AVGIGALFRGFLGAAGSTMGARSWKKKKKKG;
45

KALP: Acetyl-GKKLALALALALALALALALKKA-NH2.
The underlined residues of HFPs are N-terminal fusion peptide (FP) residues of fusion protein
gp41 subunit of the HIV virus, LAV1a strain, without (HFP) and with point mutation (HFP_V2E
and HFP_L9R).[1, 2] The HFP peptides have a non-native W to permit FP quantitation by 280
nm absorbance, and a non-native C-terminal tag to increase the aqueous solubility of the peptides
which helps purification and further NMR sample preparation. Manual solid-phase peptide
synthesis (SPPS) was done using Fmoc chemistry for all the peptide sequences.[3, 4] Wang resin
with attached glycine and alanine were used for HFP sequences and KALP, respectively. HFP
was 13CO labeled at G5, G10, L12 and G16. HFP_V2E was 13CO labeled at G5, and HFP_L9R
was 13CO labeled at G5 and G10. KALP was 13CO labeled at A5, A7, A17 and A19. There was
only one residue 13CO labeled in each labeling sequence. The synthesized peptides were purified
using reversed phase HPLC with a preparative C4 column using water - acetonitrile gradient
containing 0.1% TFA. TFA is the ion- pair reagent and helps maintain the acidic pH (~ 2) of the
eluting solution and also neutralizes the carboxylate group of the peptide. The HPLC purification
program for each peptide is in Appendix. Peptide content in the fractions was analyzed by matrix
assisted laser desorption/ionization (MALDI)- TOF mass spectrometry using the intensity of the
peptide mass peak relative to the sum of mass peak intensities. The purest fraction contained 9095% peptide and was used to make solid-state NMR (SSNMR) samples. For a 200 mole scale
synthesis, this fraction contained ~ 12 mole HFP.
2.3 Peptide associated membrane sample preparation for MAS and static solid state NMR
Samples for NMR experiments were prepared by the same procedure to better analyze the data in
terms of dependence on sample composition. Each sample contained ~1Îźmole peptide, 40Îźmole
PC, 10Îźmole PG, and either 25, 12.5, 6.25, or 0Îźmole CHOL, which respectively correspond to
46

cholesterol lipid fraction fChol = 0.33, 0.20, 0.11, and 0. NMR samples were typically prepared by
organic cosolubilization method as following: phospholipids (with and without CHOL) were
dissolved in chloroform and solvent was then removed with nitrogen gas and overnight vacuum
pumping.[5, 6] The dry lipid film and lyophilized peptide were dissolved in 2,2,2trifuoroethanol:1,1,1,3,3,3-hexafluoroisopropanol:chloroform with 2:2:3 volume ratio, and
solvent was then removed with nitrogen gas and overnight vacuum pumping. The peptide +lipid
film was suspended in 3 mL aqueous buffer containing 5mM HEPES and 10mM MES at pH 7.4
with 0.01 % NaN3 preservative. The suspension was subjected to ten freeze/thaw cycles and each
cycle includes rapid freezing in liquid nitrogen followed by thawing in hot water at ~ 60 °C,
then with addition of 20 mL more buffer, and ultracentrifugation at 270000g for four hours, with
free peptide soluble in the supernatant. There was typically 0.85 fraction membrane-bound FP, as
compared to 0.5 fraction for samples without anionic lipid, which evidences both hydrophobic
and electrostatic contributions to binding. The centrifugation pellet was harvested, lyophilized,
and packed in a 4 mm diameter magic angle spinning (MAS) rotor which contained 10ÎźL buffer.
An additional 10ÎźL buffer was added after packing.
Some FP samples were prepared by initial FP binding to unilamellar vesicles in aqueous solution
which is also named as aqueous vesicle binding method as following: Lipids were dissolved in
chloroform and the solvent was removed by nitrogen gas followed by vacuum pumping
overnight. The lipid film was suspended in 2mL aqueous buffer at pH 7.4 followed by 10
freeze/thaw cycles. Large unilamellar vesicles were prepared by extrusion through a
polycarbonate filter with 100 nm diameter pores. The extrusion was repeated 20-25 times. There
is probably 10 % lipid loss during extrusion process, but the loss of lipid is minimized by
extruding ~ 1 mL buffer through the apparatus after collecting the vesicle suspension. The buffer

47

extrusion mixture is also harvested and the lipid loss is not considered. ~ 20 mL 0.1mM peptide
stock solution in pH 7.4 buffer was added dropwise to the extruded vesicles while maintaining
the pH 7.4. The feeding peptide: lipid mole ratio is 1:25. The lipid and peptide solution was
vortexed overnight and ultra-centrifuged at 270000 g for four hours. The quantity of membrane
bound peptide is considered the difference between the total and unbound quantities, which are
obtained by measuring A280 with Îľ280 = 5700 cm-1 M-1. The membrane bound peptide: lipid mole
ratio is 1:50 in the pellet. The membrane bound peptide fraction relative to total feed peptide is
typically 0.5, and this is probably because there is kinetic barrier for peptide incorporation into
the membrane. The typical fraction of membrane-bound FP is 0.85 in organic cosolubilization
method where there is no kinetically trapped peptide because peptide and lipids were premixed.
The pellet was lyophilized overnight and packed in 4 mm MAS rotor. Typically, ~ 10 L of
buffer (pH 7.4) was added to the rotor before and after packing the sample to rehydrate the
peptide bound membrane pellet.
The packed NMR sample was rehydrated overnight at room temperature before NMR
experiments.
2.4 Solid state NMR
2.4.1 MAS solid state NMR spectroscopy
Experiments were done with a 9.4 T Agilent Infinity Plus spectrometer using a MAS probe tuned
to 1H, 13C, and 2H frequencies. The sample was typically cooled with nitrogen gas at –50 °C with
corresponding sample temperature of ~ –30 °C. The REDOR pulse sequence was in time: (1) 1H
π/2 pulse; (2) 1H-13C cross polarization (CP); (3) dephasing period of duration (); and (4)

13

C

detection. S0 and S1 REDOR data were acquired alternately and differed in the pulses applied
during the dephasing period.[5] For both acquisitions, there was a 13C π pulse at the end of each

48

rotor cycle except the final cycle, and for S1, there was also a 2H π pulse at the midpoint of each
cycle. Typical parameters included: (1)

13

C transmitter at 160 ppm and 2H transmitter at the

center of the powder pattern; (2) 10 kHz MAS frequency and 1.5ms CP contact time; (3) 50 kHz
1

H π /2 pulse and CP fields; (3) 55-68 kHz 13C CP ramp; (4) 60 kHz 13C π pulses and 100 kHz

2

H π pulses with XY-8 phase cycling applied to all π pulses; and ~70 kHz two-pulse phase-

modulated (TPPM) 1H decoupling during dephasing and acquisition.[7, 8] Pulses were calibrated
using a lyophilized peptide that included a single 13CO- 2H spin pair with rCD = 5.0 Å.[9] Typical
recycle delays were 1 s ( = 2, 8, 16ms), 1.5 s ( = 24, 32ms), and 2 s ( = 40 and 48ms). The
typical numbers of summed S0 or S1 scans were ~ 4000, 7000, 12000, 22000, 32000, 40000, and
50000 for = 2, 8, 16, 24, 32, 40, and 48ms, respectively. 13C shifts were externally referenced to
the adamantane methylene peak at 40.5 ppm which allows direct comparison to liquid-state
NMR databases.[10] Data processing included 100 Hz Gaussian line broadening and baseline
correction. S0 and S1 are typically integrals over 3 ppm intervals, with △S/S0 = (S0 – S1)/S0 and 
△S/S0

= ((S0/S0)2 + (S1/S1)2)½  (S1/S0), where S0 and S1 are the standard deviations of 10

spectral noise regions with 3 ppm integration width.[11]

49

Figure 2.1
observed

13

13

C – 2H REDOR pulse sequence. Each sequence starts with a CP from 1H to the

C nucleus to enhance the intensity of

13

C signal followed by a dephasing and

acquisition period. TPPM 1H decoupling was applied during the dephasing and the acquisition
periods.
NMR

parameters

were

optimized

using

I4

peptide

with

sequence

of

Acetyl-

AEAAAKEAAAKEAAAKA-NH2 with C-terminal amidation and N-terminal Acetylation.[9]
The I4 peptide was synthesized by solid phase peptide synthesis (SPPS) and A8 CÎą- 2H and A9
13

CO labeled.[5] Solid state NMR studies have shown that lyophilized I4 peptide has majorly Îą

helical conformation. The distance between A8 Cα- 2H and A9 13CO labeled nuclei is 5.0 Å with
a corresponding 13C- 2H dipolar coupling of 37 Hz.
2.4.2 Static solid state NMR spectroscopy
The overall membrane (lipid/CHOL) structure and motions with and without HFP were
evaluated using static 2H NMR spectroscopy typically with quadrupolar echo (quecho) pulse

50

sequence.[12] The experiments were done on a 9.4 T Agilent Infinity Plus spectrometer with a
MAS triple resonance probe tuned to 2H frequency. The 2H frequency was 61. 2023333 MHz,
and the 2H /2 pulses were calibrated using D2O (99%). The quecho pulse sequence is, (/2)x - 
- (/2)y - 1 - detect (Figure 2.2), and this sequence is used to minimize the effects of pulse ringdown.[13, 14] The first /2 pulse is the excitation pulse and the second /2 pulse is the
refocusing pulse. The phase of the first /2 pulse is x and the phase of the second /2 pulse is
alternated between y and –y. The recycle delay is 1s. 2H spectra were acquired for a fixed  and
1 value at different temperatures. To obtain the 2H T2, decay of the acquired signals was
measured for different  and 1 with synchronous increment of  and 1. Typical static solid state
NMR parameters include 2.2Οs 2H /2 pulse, dwell time = 2Οs,  = 40Οs and 1 = 21Οs. The
quecho 2H FID data was typically processed with dc offset because uncorrected dc offset could
result a spike at the middle of the spectrum.[15] Experimentally, 1 is shorter than , so we also
need to do data shift to move the maximum echo signal at t = 0 before Fourier Transform (FT)
was performed, typically with -11 data shifts and 200 Hz Gaussian line broadening.

Figure 2.2. “Quecho” pulse sequence.

51

REFERENCES

52

REFERENCES

1.

Qiang, W. and D.P. Weliky, HIV Fusion Peptide and Its Cross-Linked Oligomers:
Efficient Syntheses, Significance of the Trimer in Fusion Activity, Correlation of beta
Strand Conformation with Membrane Cholesterol, and Proximity to Lipid Headgroups.
Biochemistry, 2009. 48(2): p. 289-301.

2.

Freed, E.O., et al., A MUTATION IN THE HUMAN-IMMUNODEFICIENCY-VIRUS
TYPE-1 TRANSMEMBRANE GLYCOPROTEIN-GP41 DOMINANTLY INTERFERES
WITH FUSION AND INFECTIVITY. Proceedings of the National Academy of Sciences
of the United States of America, 1992. 89(1): p. 70-74.

3.

Fields, G.B. and R.L. Noble, SOLID-PHASE PEPTIDE-SYNTHESIS UTILIZING 9FLUORENYLMETHOXYCARBONYL AMINO-ACIDS. International Journal of Peptide
and Protein Research, 1990. 35(3): p. 161-214.

4.

Merrifield, R.B., SOLID PHASE PEPTIDE SYNTHESIS .1. SYNTHESIS OF A
TETRAPEPTIDE. Journal of the American Chemical Society, 1963. 85(14): p. 2149-&.

5.

Xie, L., et al., Residue-specific membrane location of peptides and proteins using
specifically and extensively deuterated lipids and C-13-H-2 rotational-echo doubleresonance solid-state NMR. Journal of Biomolecular Nmr, 2013. 55(1): p. 11-17.

6.

Jia, L.H., et al., REDOR solid-state NMR as a probe of the membrane locations of
membrane-associated peptides and proteins. Journal of Magnetic Resonance, 2015. 253:
p. 154-165.

7.

Bennett, A.E., et al., HETERONUCLEAR DECOUPLING IN ROTATING SOLIDS.
Journal of Chemical Physics, 1995. 103(16): p. 6951-6958.

8.

Terry Gullion, D.B.B., Mark S Conradi, New, compensated Carr-Purcell sequences.
Journal of Magnetic Resonance, 1990. 89(3): p. 479-484.

9.

Long, H.W. and R. Tycko, Biopolymer conformational distributions from solid-state
NMR: alpha-helix and 3(10)-helix contents of a helical peptide. Journal of the American
Chemical Society, 1998. 120(28): p. 7039-7048.

10.

Morcombe, C.R. and K.W. Zilm, Chemical shift referencing in MAS solid state NMR.
Journal of Magnetic Resonance, 2003. 162(2): p. 479-486.

11.

Zheng, Z., et al., Conformational flexibility and strand arrangements of the membraneassociated HIV fusion peptide trimer probed by solid-state NMR spectroscopy.
Biochemistry, 2006. 45(43): p. 12960-12975.

53

12.

Davis, J.H., THE DESCRIPTION OF MEMBRANE LIPID CONFORMATION, ORDER
AND DYNAMICS BY H-2-NMR. Biochimica Et Biophysica Acta, 1983. 737(1): p. 117171.

13.

Duer, M.J., Solid-State NMR Spectroscopy Principles and Applications. 2002: Blackwell
Science.

14.

Spiess, K.S.-R.H.W., Multidimensional solid-state NMR and polymers. 1994:
ACADEMIC PRESS INC.

15.

Christian Schorn, B.T., NMR Spectroscopy: Data Acquisition. 2004, WILEY-VCH
Verlag GmbH & Co. KGaA.

54

Chapter 3 - Structure and Membrane Location Studies of HIV Fusion Peptide
(HFP) and KALP Peptide
3.1 Introduction
The cell membrane is the stable lamellar bilayer structure composed of lipids and is the physical
barrier to molecular diffusion. Aside from lipids, the membrane also contains many types of
proteins. And the membrane mass is about equal between lipids and protein for plasma
membranes of most animal cells.[1] The protein membrane locations and contacts between
protein specific residues and membrane lipid specific regions are important for their functions
and can also give insight to protein/membrane biophysical interaction.[2-4] There have been
high-resolution protein crystal structures that are obtained generally in non-lamellar media like
detergent micelles, detergent-rich bicelles, or lipidic cubic phase. These high resolution
structures sometimes provide the protein location information in the non-lamellar phase but
typically not in the bilayer phase which is the most relevant model of the cell membrane.
Structure of HFP studied in detergent micelle shows majorly helical conformation, while
structure of HFP studied in membrane without CHOL show both helical and  sheet secondary
structure, and studies in membrane containing ~ 33% CHOL show predominant  strand
structure.[5-10]
The membrane locations of HFP have been studied by REDOR NMR method.[11-14] The
peptide location relative to the membrane surface can be studied by

13

C-31P REDOR. The

peptide location in the membrane hydrocarbon core has been extensively studied by
incorporating fluorinated lipids. DPPC (C5and C16. DPPC (C5-

19

F) and DPPC (C16-

19

F) were fluorinated at C5

19

F) is used to investigate the location of the peptide relative to the half

way of one membrane leaflet, and DPPC (C16- 19F) to study the location of the peptide relative

55

to the membrane center. However, fluorinated DPPC is chemically modified lipid and there is
potential perturbation to the membrane structure. Studies have shown that 100% DPPC (C16- 19F)
would form interdigitated membrane.[15] To maintain membrane bilayer and achieve maximum
REDOR dephasing, generally only 10% fluorinated lipids were incorporated, but these dilute and
randomly distributed nuclei would give rise to the inaccuracy of inter-nuclear distance
correlation with membrane insertion depth.[16]

Figure 3.1 (a) DPPC bilayer regular membrane (left) without chemically modified lipid and
interdigitated membrane (right) composed of C16- 19F DPPC with a 1H→ 19F substitution at C16,
(b) chemical structure of C16- 19F DPPC lipid.[15, 17]
A recently developed method in our group is to use

13

C-2H REDOR method to study peptide

membrane location in the membrane hydrophobic core using deuterated lipids or CHOL.[11, 13,
14, 18, 19] This method is advantageous because, 2H is chemically equivalent to 1H, and would
not cause perturbation to the membrane. Besides, it would have a continuous band of 2H labeling

56

in the membrane, which could more accurately reflect the membrane location according to the
closest

peptide

-membrane

contact

(see

figure

3.2B).

Non-chemically

modified

phosphatidylcholine (PC) and CHOL are available with a wide variety of 2H labeling patterns
located in different bilayer regions (Figure 3.2A).
Previous studies have shown that HFP forms a small intermolecular antiparallel  sheet with a
distribution of antiparallel registries.[9, 20] Studies of HFP_F8c and HFP_G5c by

13

C- 2H

REDOR support major deeply inserted and minor shallowly inserted membrane locations for
HFP.[18] In this thesis, the membrane locations of HFP ware studied in membrane both without
and with CHOL. In membrane without CHOL, HFP was

13

CO labeled at either G5 or L12. In

membrane with CHOL, HFP was 13CO labeled at residue G5 and G16. To help understand HFP
induced membrane fusion mechanism, fusogenic L9R and non-fusogenic V2E mutants were also
studied in membrane without CHOL.[21] This method has also been applied to study the
membrane locations of the model transmembrane KALP peptide.

57

Figure 3.2 (A) 2H patterns and structures of deuterated DPPC lipids and CHOL and (B)
approximate 2H’s and

31

P’s (P) membrane locations in the membrane without protein. DPPC

lipids are deuterated at different regions of the acyl chain. The lipid 2H and 31P locations are for
the membrane gel-phase without CHOL and CHOL 2H locations are for the liquid-ordered phase
with CHOL. The same color-coding is used in other figures through this thesis.
The peptides were chemically synthesized with specific

13

CO labeling by FMOC solid phase

peptide synthesis (SPPS).[22] The method was described in Chapter 2, where the NMR sample
preparation method and NMR parameters were also discussed.

58

3.2 Results
3.2.1 Fitting of the 13C-2H REDOR data
Dephasing buildup plot is obtained from dephasing (△S/S0) at different dephasing time ().[23]
The S0 and S1 peak intensities are also denoted S0 and S1, and are obtained typically from 3ppm
integration windows of major peak. The dephasing (△S/S0) is calculated as
(S0 − S1 )
∆S
=
S0
S0

3.1

The S0 and S1 error is based on spectral noise, which is the standard deviation of 10 spectral
noise regions with 3ppm integration window.[24]
σS 2
σS 2 S1
σ∆S/S0 = √( 0 ) + ( 1 ) ×
S0
S1
S0

3.2

The dephasing is directly related to the dipolar coupling between the

13

C (detecting) and 2H

(dephasing) nuclei. The 13C – 2H inter-nuclear dipolar coupling (d) is dependent on the 13C – 2H
inter-nuclear distance (r), and
d(Hz) =

4642
r(Å)

3.3

3

The buildup of experimental dephasing (△S/S0)exp as a function of  provides the experimental
basis for evaluating the protein labeled (lab) 13CO - to - lipid 2H proximity and the labeled internuclear distance (r) values. The uncertainty of (△S/S0)exp is based on spectral noise and
calculated according to equation 3.2. The sample analyzed is 0.5Îźmol HFP G5c in 50Îźmol
PC_d10 membrane with peptide to lipid ratio of 1:100.
The data are fitted with three approaches that are denoted І, П and Ш. And best fit parameters
are based on the minimum χ2 value. Method І and П are simulated by quantum mechanics- based
SIMPSON program with a model of isolated

13

59

CO – 2H spin-pairs that has a single dipolar

coupling (d).[25] І is obtained with two populations (P’s) of peptides (HFP G5c molecules), and
one population (P) (single

13

CO – 2H spin-pair) has a nonzero dipolar coupling (d) and

contributes to the experimental dephasing buildup while the rest population (1-P) has a dipolar
coupling of zero and does not contribute to the experimental dephasing buildup. П is obtained
with three populations (P’s) of peptides, and two populations (two different 13CO – 2H spin-pairs)
have nonzero dipolar couplings and contribute to the experimental dephasing, while the rest
population has a dipolar coupling of zero and does not contribute to the experimental dephasing.
For approach І and П, each population has a single 13CO – 2H spin-pairs and a single d, and the
(△S/S0) is quantum mechanically calculated by SIMPSON program. For Ш, the (△S/S0)exp is
fitted to a single exponential buildup equation A  (1 – e

–  

) with A and  as the fitting

parameters. A is assigned as the approximate fraction of peptide (HFP molecule) with d ≈ 3γ/2.
The 3/2 ratio is based on approximately equal time spent in the three 2H m states (m= -1, 0, +1
states) during τ because of the 2H T1 relaxation. Experimental dephasing time  typically goes up
to 48ms, and there are m=0  m=1 2H transitions during the dephasing period because the 2H
T1  50ms.[17, 26] There isn’t buildup for a lab 13CO during the m = 0 times of 2H nuclei. The
stochastic variability of the m = 0 times due to 2H T1 relaxation among the sample 13CO’s is not
straightforwardly incorporated into quantum mechanical calculation of the buildup in SIMPSON
program. We approximate that each 2H is in the m = 0 state for 1/3 of the dephasing period so
that the observed buildup rate   2d/3. (1 – A) is the fraction of peptide (HFP molecule) with d 
0.
The data are fitted poorly by method I in part because the (S/S0)exp buildup has exponential
shape whereas the calculated buildup by SIMPSON program gives sigmoidal shape. Method II
gives better fitting and applied three populations of HFP G5
60

13

CO with four fitting parameters:

fractional populations A1 and A2 with couplings d1 and d2, respectively. The A3 = 1 – A1 – A2 and
d3 = 0. Method III also has good fitting with a single exponential buildup from equation: A  (1
– e –   ). Method III is consistent with a model of two populations. Population P1 has fraction A
with

13

CO-2H proximity of d  3/2 and population P2 has fraction 1 – A and d  0. We

approximate that the dipolar coupling is dominated by the closest 2H and the 13CO-2H distance r
is calculated from the d of P1 using equation 3.3.
There are several reasons for choosing method III rather than II for general fitting of sample
buildups. (1) The 2III is lowest for most samples fitting, and it is generally close to the number
of degrees of fitting  5 and this is statistically reasonable. The lowest 2III is achieved with two
fitting parameters instead of four fitting parameters in method II which also gives good fitting
buildups. This difference is especially significant because there are only seven data points to be
fitted. We choose method III also because it is simpler and probably more biophysically
plausible to have two instead of three peptide membrane locations. Additionally, stochastic
processes such as the non-radiative m=0  m=1 transitions commonly have exponential
dependence as a function of time.[26]

61

Figure 3.3 Experimental

13

CO-2H (△S/S0)exp (red squares with error bars) for sample HFP_G5C

in PC_d10 membrane. The (△S/S0)exp are for the major lab G5 peak with  sheet structure and
error bars are calculated from spectral noise. Fitted (△S/S0) are displayed for three different
fitting approaches. The blue crosses (Approach I) and green stars (Approach II) are based on
models of two- (P1 and P2) and three-populations (P1, P2, and P3) of HFP_G5C molecules,
respectively. The (△S/S0) for each population is calculated with the quantum mechanics-based
SIMPSON program using a model of isolated 13CO-2H spin-pairs with a single dipolar coupling
(d). For the two population model, the best-fit parameter values for P1 are d = 53 Hz and
fractional population A = 0.71. The corresponding P2 = 1 – A = 0.29 with d = 0 Hz. For the threepopulation model, the best-fit values are d1 = 90 Hz, A1 = 0.27, d2 = 25 Hz, and A2 = 0.50 with P3
= 1 – A1 – A2 = 0.23 and d3 = 0 Hz. The black line (Approach III) is the best-fit to the
exponential buildup function A × (1 – e – ) with A = 0.63 and  = 44 Hz.

62

3.2.2 Effect of sample preparation methods on 13C–2H REDOR △S/S0
To evaluate whether HFP achieves thermodynamic structure and membrane location, two
different sample preparation methods were compared, which is organic cosolubilization method
and aqueous vesicle binding method and the two different sample preparation methods have been
described in detail in Chapter 2. Most samples were prepared by organic cosolubilization method.
Vesicle binding method is more like peptide incorporation during viral fusion. In organic sample
preparation method, the lipids and peptides are well mixed in organic solvent mixtures, so there
will be no kinetic trapped structure and membrane location. In these samples, the HFP is

13

CO

labeled at Gly5 residue. The membrane is composed of PC_d10: DTPG with 4:1mol ratio. And
the peptide to lipid ratio is 1:25. Both samples have similar spectrum (Figure 3.3) with a major
CO peak centered at 171ppm chemical shift which suggests HFP form majorly  sheet

13

structure in membrane. The higher chemical shift 175ppm peak is from lipid carbonyl natural
abundance and peptide carbonyl natural abundance of residues other than glycine and
methionine.[27]
The dephasing buildups (Figure 3.4) are also similar with (S/S0)exp  0.7 for large . The results
support that HFP achieves thermodynamic equilibrium structure and membrane location.

63

(a)

Figure 3.4

(b)

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (red) experimental spectra of

membrane - associated HFP G5c at  = 40ms by different sample preparation methods, (a)
organic cosolubilization method (organic) and (b) aqueous vesicle binding method (aqueous).
Each spectrum is the sum of ~ 40000 scans and processed with 150 Hz Gaussian line broadening
and polynomial baseline correction. The observed chemical shifts for G5 are 171ppm from both
organic and aqueous methods, and this is consistent with major  sheet structure of HFP in
membrane. The similar structure supports that HFP achieves thermodynamic equilibrium
structure when it’s associated with membrane. The membrane is composed of 40μmol
DPPC_d10 and 10Îźmol DTPG lipids. Sample prepared by organic methods contains ~ 2Îźmol
and sample prepared by aqueous method contains ~ 1.3Îźmol HFP. The cooling N2 gas
temperature is ~ - 50 °C with corresponding sample temperature of ~ - 30 °C.

64

Figure 3.5 The dephasing buildups of △S/S0 vs dephasing time () for different NMR sample
preparation methods: organic cosolubilization (closed square) and aqueous vesicle binding (open
square). The intensity for S0 and S1 were obtained by integration over a 3ppm width centered at
the highest peak intensity. The similar dephasing buildups support thermodynamic equilibrium
membrane location of HFP when it’s associated with membrane.

65

3.2.3 Effect of temperature on experimental △S/S0
The effect of temperature on the experimental dephasing buildup is compared with sample
temperature at ~ - 30 °C and ~ - 0 °C with corresponding cooling gas temperature of - 50 °C and
~ - 20 °C. The sample analyzed is HFP_G5c in PC_d10: DTPG (4:1) membrane. And the sample
is prepared by organic cosolubilization sample preparation method. The same sample is
investigated at different temperature. The spectra acquired at the two different temperatures are
similar with a major peak at 171 ppm chemical shift corresponding to  sheet secondary
structure. The same  sheet structure at higher temperature supports that HFP structure doesn’t
change significantly when temperature varies and probably represents the most relevant structure
at physiological temperature. The dephasing buildup is smaller at higher temperature and the
(△S/S0)exp0 °C/ (△S/S0) exp –30 °C) ≈ 0.7 for a given . At higher temperature the signal per scan S0
is decreased when  increases. For =2ms, (S0)exp 0 °C/ (S0) exp –30 °C) ≈ 1.0, and it’s only 0.13 for 
= 48ms. This suggests that the 1H →13C CP is temperature independent and the T2 of peptide
13

CO is shorter at higher temperature. The T2 is shorter probably because the peptide motion

increases at higher temperature. And the increased peptide and lipids motions likely cause
motional averaging of the

13

CO- 2H dipolar coupling and thus a reduced (△S/S0)exp is

observed.[18] In order to obtain the biggest experimental (△S/S0), all samples are investigated at
~ - 30 °C.

66

Figure 3.6 The spectra (top panel) and dephasing buildups (bottom panel) of △S/S0 vs dephasing
time () for sample investigated at ~ - 30 °C and ~ - 0 °C: the spectrum is for  = 40ms and the
black is S0 and red is S1. The data is processed with 150Hz Gaussian line broadening.
3.2.4 Effect of membrane charge on experimental △S/S0
The membrane charge effect on HFP structure and membrane location is investigated with HFP
G5c in neutral membrane with 50Îźmol PC_d10, PC_d8 and PC_d4 with 2H labeling. The
negatively charged membrane has the same lipid labeling but with addition of anionic PG lipid

67

and the PC: PG mole ratio is 4:1. The negative charged lipid is added because most human cell
membranes including the HIV host cell include 0.1~ 0.2mol fraction of anionic lipids.[28] The
S0 spectra of HFP G5c samples are similar with predominant 171ppm chemical shift peak
regardless of the membrane charge. The 171ppm chemical shift corresponds to  sheet structure
of the lab G5

13

CO’s. The dephasing buildups also show similar trends. There are significant

buildups for HFP G5 13CO in membranes labeled with PC_d10 and PC_d8 lipids, but negligible
buildups in membranes with labeling in PC_d4 lipid. The chemical shift and dephasing buildups
features support that HFP  sheet is inserted into the membrane hydrophobic core. Exponential
fitting of the significant dephasing buildups give 4~5 Å closest inter-nuclear distances, which
indicates Van der Waals contact between HFP G5 13CO and the lipid 2Hs in PC_d8 and PC_d10.
The NMR sample is generally prepared with organic cosolubilization method if without special
instructions. The peptide incorporation efficiency is generally greater for sample in negatively
charged membrane. This greater bound peptide fraction is probably due to the electrostatic
attraction between the positively charged solubility tag in HFP and the negatively charged lipids
head-group in the membrane. Therefore, extra peptides were added to compensate the binding
efficiency difference when preparing samples with neutral membrane. Since the NMR sample is
the centrifuged pellet containing the membrane and the bound HFP, the unbound HFP was in the
supernatant and separated from and not contained in the NMR sample. Because of better binding
of HFP in anionic membrane, it gives better prediction of the quantity of HFP incorporated in
membrane. Thus, most samples in this thesis are prepared in membrane with negatively charged
lipids and the PC: PG = 4:1 mole ratio.

68

(a)

(b)

Figure 3.7

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra of

HFP_G5c in (a) neutral membrane and (b) negatively charged membrane at  = 40ms. The data
is processed with 100Hz Gaussian line broadening and polynomial baseline correction. The 171
ppm chemical shift indicates that HFP has predominant  sheet structure at the labeled Gly5
residue site.

69

Figure 3.8 The dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing time
() for HFP_G5c in (top panel) neutral membrane and (bottom panel) negatively charged
membrane. The neutral membrane is composed of PC lipid, and the negative membrane is PC:
PG lipids of 4:1mol ratio. The dephasing buildups are for the major peak at 171 ppm with
integration window of 3ppm for both S0 and S1. The colored line is the best fit exponential
buildup curve fitted by A × (1 – e – ) for samples with significant dephasing buildups.

70

Table 3.1 Best-fit exponential buildup parameters for HFP_G5c in membrane a

a

Membrane

A

 (Hz)

r (Å)

PC_d10

0.63(4)

44(5)

4.1(2)

PC_d8

0.60(3)

34(3)

4.5(1)

PC_d10:PG

0.89(2)

36(2)

4.4(1)

PC_d8:PG

0.44(6)

27(6)

4.9(3)

A and  are fitting parameters for experimental dephasing buildup from fitting equation of A ×

(1 – e – ), and r is calculated as d= 3/2 = 4642/r3.
3.2.5 HFP location in membrane without CHOL studied by 13C–2H REDOR
To study the membrane locations of HFP, we used deuterated PC lipids with addition of a small
fraction of anionic PG lipids and PC: PG = 4:1 mole ratio. The negative charged lipid is added
because the HIV host cell includes 0.1~ 0.2mol fraction of anionic lipids and also because better
peptide binding to membrane.[28] The HFP is

13

CO labeled at G5 and L12 residues. Former

student Dr. Li XIE from our group has studied the membrane contact of HFP with F8

13

CO

labeled in neutral membrane with deuterated PC lipid and showed that HFP_F8 has  sheet
structure and is inserted in the membrane hydrophobic core.[26] Additionally, for HFP_F8
studied in DMPC_d54 with per-deuterated lipid acyl chain, the dephasing goes up to ~ 1 rapidly,

71

which supports that all HFP molecules are deeply inserted into the membrane hydrocarbon core.
C – 15N REDOR studies have shown that HFP forms majorly  antiparallel sheet registry when

13

associated with ~ 30mol % CHOL. In membrane without CHOL, each HFP 13CO labeled sample
is studied with PC_d10, PC_d8 and PC_d4 2H labeling in the membrane to fully understand the
13

CO contacts with different regions of the membrane. For HFP_G5c in PC: PG (4:1) membrane,

the spectrum is displayed in Figure 3.7 (b), the dephasing buildup is displayed in Figure 3.8, and
the fitting results of significant buildups for PC_d10 and PC_d8 labeling is shown in Table 3.1.
The major peak for G5 has 171 ppm chemical shift, and the major peak for L12 has 175 ppm
chemical shift, these results support major  sheet structure at both G5 and L12 residue. The data
for sample containing HFP_L12c is processed with 20Hz Gaussian line broadening and
polynomial baseline correction. L12 and G5 have similar dephasing buildup features with
biggest dephasing with PC_d10, significant dephasing with PC_d8 and smallest dephasing with
PC_d4. Similar features are also observed for HFP_F8 samples.[26] Fitting of PC_d10 and
PC_d8 experimental dephasing buildups all give 4 ~ 5 Å inter-nuclear distances, which supports
Van der Waals contacts between HFP_G5 and L12 13CO and 2H of membrane lipid PC_d10 and
PC_d8. At = 48ms, the (△S/S0)d10/(△S/S0)d8 is ≈ 7:3 and Ad10: Ad8 ≈ 2:1 supports that there
are multiple membrane locations of HFP in membrane. And the major population is deeply
inserted into the membrane in contact with d10 2Hs, and the minor population is shallowly
inserted into the membrane and contacts d8 2Hs.
The interior of the HFP  sheet is likely located within the hydrophobic hydrocarbon core in the
membrane, because there is lower free energy due to hydrophobic effect from many of the
nonpolar sidechains of HFP amino acids. On the contrary, the terminal part of HFP is probably
located in near the membrane head group region rather than the membrane hydrocarbon region.

72

There are registry distributions of antiparallel HFP registry in membrane. The HFP terminal
residues have incomplete inter–residue hydrogen bonds with the neighbor strands. The free
energy is lowered by forming additional hydrogen bonds with water which has higher content
near the membrane head group region instead of the hydrocarbon core. The interior residues can
form nearly complete inter–residue hydrogen bonds in membrane hydrocarbon core. All HFP
antiparallel  sheet registries have G5 and L12 in the registry interior and G16 near the registry
terminal.[18, 29] There is probably small number (~ 10) of HFP molecules in the antiparallel 
sheet registries and is consistent with gp41 oligomerization including establishment of a dimer of
trimers of gp41ectodomain.[30]
It's not clear how the major deeply inserted and minor shallowly inserted membrane location is
advantageous to HFP induced membrane fusion. The majorly deep inserted HFP may reduce the
fusion activation energy to the membrane fusion intermediate by perturbing the local membrane.
The multiple membrane location of HFP may be correlated with HFP antiparallel  sheet registry
distributions and hydrophobicity in membrane. How the lipids are displaced by HFP molecule is
not known and neither the orientation of the contact lipids relative to the rest bulk lipid. Our data
is most consistent with insertion of the antiparallel  sheet registries in a single membrane leaflet,
but we can’t rule out a transmembrane model.

73

Figure 3.9

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra of

HFP_L12c in PC: PG =4:1 membrane. The data is processed with 20Hz Gaussian line
broadening and polynomial baseline correction.

Figure 3.10 The dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing
time () for HFP_G12c in membrane composed of PC: PG = 4:1 ratio. The △S/S0 data is for the
major peak of  sheet conformation and the S0 and S1 is integration with 3ppm integration
window. The colored line is the best fit exponential buildup curve fitted by A × (1 – e – ).

74

Table 3.2 Best-fit exponential buildup parameters for HFP_L12c in membrane a

a

Membrane

A

 (Hz)

r (Å)

PC_d8

0.50 (8)

25 (5)

5.0 (4)

PC_d10

0.88 (5)

31 (3)

4.6 (2)

A and  are fitting parameters for experimental dephasing buildup from fitting equation of A ×

(1 – e – ), and r is calculated as d= 3/2 = 4642/r3.

75

(a) Major population

(b) Minor population

Figure 3.11 HFP membrane location model in anionic membrane without CHOL: (a) major
deeply inserted membrane location and (b) minor shallowly inserted membrane location. The
membrane 2H positions represent the location without protein.
3.2.6 HFP location in membrane with CHOL studied by 13C–2H REDOR
3.2.6.1 HFP location in DPPC: DPPG: CHOL membrane studied by 13C–2H REDOR
HIV host cell membrane contains ~ 30mol % CHOL, so it’s important to study HFP membrane
location in membrane containing CHOL.[31] The HFP is

13

CO labeled at G5 and G16 residue.

One advantage of Gly-13CO labeling is that HFP has majorly  sheet structure in membrane. The

76

171ppm major peak of HFP Gly-13CO is separated from 175ppm lipid carbonyl natural
abundance peak. Previous 13C– 2H REDOR data shows that G5 is majorly deeply inserted into
membrane center contacting PC_d10 2Hs in membrane without CHOL, and

13

C–

31

P REDOR

data suggests that G5 is far away (at least 10 Å) from phosphorous head group. 13C–31P REDOR
data also suggests that G16 is close to phosphorous head group because A14 and A15 is ~ 5 Å
away from the phosphorous group.[16, 32] Therefore, G5 and G16 are reasonable candidates to
study HFP contacts to Chol_d7 and Chol_d6 with 2Hs located near the center and edge of the
membrane, respectively. Chol_d7 and Chol_d6 have 2Hs deuterated at the methyl and hydroxyl
regions, respectively. The membrane studied is PC: PG: CHOL (8:2:5) and the CHOL amount is
25Îźmol.
In the spectrum, both G5 and G16 have the major peak at 171ppm, which corresponds to sheet
secondary structure of HFP in membrane containing 33mol % CHOL. G16 also has a slightly
smaller peak at 174 ppm compared to the 171 ppm major peak. The 174 ppm peak likely
corresponds to coil conformation according to the glycine chemical shift distributions in
proteins.[27] Coil formation means lack of regular secondary structure (helix and β sheet), and
the dihedral angles could be any of the angles sterically allowed while different dihedral angles
give different chemical shifts. Those G16 residues form hydrogen bonds with water near the
membrane head group.[18, 27] A single major peak at 171 ppm for G5 suggests that almost all
HFP antiparallel registries include G5 residue and G5 forms complete inter-residue hydrogen
bonds.
The dephasing is calculated for major 171ppm peak corresponding to  sheet HFP structure. The
dephasing buildup trends are strikingly different for Chol_d7 and Chol_d6 labeling. G5 has large
and small dephasing buildup with Chol_d7 and Chol_d6, respectively. For G16, the trend is

77

opposite, with large and small dephasing buildup with Chol_6 and Chol_d7, respectively. Since
Chol_d7 have 2Hs located in the membrane center, the G5 dephasing buildup supports that most
HFP G5 residue is deeply inserted into the membrane hydrophobic core near the membrane
center in CHOL containing membrane. And this result is consistent with the G5 and L12 major
deeply insertion membrane location model discussed earlier in this chapter. The G16 dephasing
buildup evidences G16 is located near membrane surface, and this result is consistent with the
13

C–31P REDOR data. Exponential fitting of the significant buildups give ~ 4 Å

nuclear distances for G5

13

CO – Chol_d7 2H and G16

13

13

CO–2H inter-

CO – Chol_d6 2H, which support Van

der Waals contact between HFP and CHOL.
HFP–CHOL contact results support that there is multiple membrane locations of HFP in
membrane containing CHOL because (△S/S0)Chol_d7 /(△S/S0)Chol_d6 ≈ 3:1 at  = 40ms. The 3:1
ratio suggests that G5 has major population deeply inserted to the membrane center contacting
Chol_d7 2Hs and minor population shallowly inserted to the membrane surface contacting
Chol_d6 2Hs. The HFP membrane location model is similar in membrane containing 33mol %
and membrane without CHOL.
The CHOL contact studies of HFP with Chol_d7 and Chol_d6 labeling scheme strongly suggests
that this

13

C–2H REDOR methodology is effective to study HFP specific residue contact to

specific regions of CHOL. And this method is promising to be applied to study protein–CHOL
van der Waals contact with other proteins.

78

Figure 3.12

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

(top panel) and the dephasing buildups (bottom panel) of △S/S0 (colored squares) with error bars
vs dephasing time () for HFP_G5c in membranes with Chol_d7 and Chol_d6. The S0 and S1
spectra displayed are for  =40ms. The data is processed with 100Hz Gaussian line broadening
and polynomial baseline correction. The dephasing buildup is for the major  peak. S0 and S1 is
integration through 3ppm integration window. The colored line is the best fit exponential buildup
curve fitted by A × (1 – e – )
(a)

79

(b)

(c)

Figure 3.13 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra (a)
and the dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing time () for
HFP_G16c in membranes with Chol_d7 and Chol_d6 for 171ppm peak (b) and 174ppm peak (c).
The S0 and S1 spectra displayed are for  =40ms. The data is processed with 100Hz Gaussian line
broadening and polynomial baseline correction. The dephasing buildup is for the major  peak
with 171ppm chemical shift. S0 and S1 is integration through 1ppm integration window. The
colored line is the best fit exponential buildup curve fitted by A × (1 – e – ).

80

Table 3.3 Best-fit exponential buildup parameters for HFP_G5c and G16c in membrane with
Chol_d7 and Chol_d6 a

a

Peptide

Membrane

A

 (Hz)

r (Å)

HFP_G5c

PC:PG:Chol_d7

0.76 (3)

47 (3)

4.0 (1)

HFP_G16c

PC:PG:Chol_d6

0.67 (5)

64 (10)

3.6 (2)

A and  are fitting parameters for experimental dephasing buildup from fitting equation of A ×

(1 – e – ), and r is calculated as d= 3/2 = 4642/r3. The membrane composition is PC: PG:
CHOL (8:2:5) with ~ 1Îźmol peptide. The peptide to lipids (not including CHOL) ratio is 1:50.

Figure 3.14 Semi-quantitative HFP membrane location model with major population deeply
inserted and minor population shallowly inserted to the membrane hydrophobic core. The
membrane 2H positions represent the location without protein.

81

3.2.6.2 HFP location in POPC: POPG: CHOL membrane studied by 13C–2H REDOR
13

C–2H REDOR method is evaluated in POPC: POPG: CHOL membrane which is typically used

to study the fusion activity of fusion proteins such as HIV gp41 and influenza
hemagglutinin.[33-36] So it’s important to study the membrane locations in this membrane
composition to help correlate the membrane location function relationship. POPC and POPG
lipids have lower melting temperature, which is -2 °C compared to 41°C of DPPC and DPPG
lipids. The membrane will be in gel phase below the lipid melting temperature and liquid
disordered phase above the lipid melting temperature. Using POPC lipid, because it represents
the most abundant lipid head-group and common lipid acyl chains in HIV host cell
membrane.[28] In this study, HFP is 13CO labeled at G10 and G16. The membrane composition
studied with both G10 and G16 labeling is POPC: POPG: CHOL with 8:2:5 mole ratios. The
membrane is labeled at CHOL with Chol_d7 and Chol_d6. Less CHOL content membrane with
POPC: POPG: CHOL of 8:2:2.5 is compared with HFP_G10c and Chol_d7 labeling. The data is
processed with Gaussian 100Hz line broadening and polynomial baseline correction. S0 and S1
are integrated with 3ppm integration window. Substantial dephasing buildups with exponential
buildup trends are fitted with equation A × (1 – e – ).
All G10 samples have predominant peak with ~ 171ppm chemical shift, which is consistent with
 sheet structure in both 33mol % and 20mol % CHOL content membranes. G16 samples have
predominant peak with ~ 171ppm chemical shift, which is consistent with  sheet structure. G16
samples also have a significant peak at ~ 174ppm, which is consistent with coil conformation
and corresponding to the G16s that are not included in the peptide antiparallel registries and have
hydrogen bonding with water molecules near the membrane surface.[29] These spectra features
are similar to those in DPPC: DPPG: CHOL membrane, which suggests the HFP peptide

82

antiparallel  sheet with registration distribution structure is robust and likely the most biological
relevant structure. In membrane with POPC: POPG: CHOL= 8:2:5, the quantitative dephasing
trend is opposite for G10 and G16 samples. For G10, the dephasing is much bigger with Chol_d7
than Chol_d6, while the dephasing is bigger with Chol_d6 than Chol_d7 for G16. The
experimental dephasing results are consistent with G10 is inserted in the membrane hydrocarbon
core and G16 has a shallower location near the membrane surface. Compared to samples in
DPPC: DPPG: CHOL membrane, the experimental dephasing at longer dephasing time is
generally smaller and the buildup rate is also slower in POPC: POPG: CHOL membrane. This
might due to shallower membrane locations of HFP in the hydrophobic core relative to the
membrane center compared to membrane location in DPPC membrane. The contrary dephasing
buildup trends with Chol_d7 and Chol_d6 labeling for both G10 and G16 supports that this 13C–
2

H REDOR method can be used to differentiate residue specific membrane locations. 20mol %

Chol_d7 labeling for G10c gives about half the dephasing than 33mol % Chol_d7 labeling. This
is probably due to spin dilutions with less amount of Chol_d7. More residue specific labeling
will be necessary for a complete membrane location model of HFP in this POPC: POPG: CHOL
membrane system.

83

Figure 3.15

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

for HFP_G10c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. POPC: POPG: CHOL = 8:2:5 and 8:2:2.5 mole ratios.

Figure 3.16 The dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing
time () for HFP_G10c in membrane composed of POPC: POPG: CHOL = 8:2:5 and 8:2:2.5
ratios. The △S/S0 data is for the major peak of  sheet conformation and the S0 and S1 is
integration with 3 ppm integration window. The colored line is the best fit exponential buildup
curve fitted by A × (1 – e – ).

84

Figure 3.17

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

for HFP_G16c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. POPC: POPG: CHOL = 8:2:5.

Figure 3.18 The dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing
time () for HFP_G16c in membrane composed of POPC: POPG: CHOL = 8:2:5 ratio. The
△S/S0 data is for the major peak of  sheet conformation and the S0 and S1 is integration with 1
ppm integration window. The colored line is the best fit exponential buildup curve fitted by A ×
(1 – e – ).

85

Table 3.4 Best-fit exponential buildup parameters for HFP_G10c and G16c in POPC: POPG
membrane with Chol_d7 and Chol_d6 a

Peptide

A

 (Hz)

r (Å)

0.79 (10)

20 (3)

5.4 (3)

1.02 (61)

12 (9)

6.4 (1.6)

0.60 (38)

11 (8)

6.6 (17)

Membrane

PC:PG:Chol_d7
HFP_G10c
(8:2:5)
PC:PG:Chol_d6
HFP_G16c
(8:2:5)
PC:PG:Chol_d7
HFP_G10c
(8:2:2.5)

a

A and  are fitting parameters for experimental dephasing buildup from fitting equation of A ×

(1 – e – ), and r is calculated as d= 3/2 = 4642/r3. Each sample typically contains ~ 1μmol
peptide. The peptide to lipids (not including CHOL) ratio is typically 1:50.
3.2.6.3 HFP location in DOPC: DOPG: CHOL membrane studied by 13C–2H REDOR
The 13C–2H REDOR NMR method is also employed to study the membrane locations of HFP in
DOPC: DOPG: CHOL membrane. DOPC and DOPG lipids have much lower melting
temperature, which is -17 °C and -18 °C, respectively, compared to 41°C of DPPC and DPPG
lipids. The sample temperature during NMR data acquisition is ~ -30 °C. The lipids and peptides
would have much more motion compared to the POPC and DPPC membranes. The peptide is
13

CO labeled at G5. The NMR sample is prepared by aqueous vesicle binding method which is

described in chapter 2. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. The spectra feature for Chol_d7 and Chol_d6 are consistent with
86

the membrane compositions studied earlier. Major peak has 171ppm chemical shift, which is
corresponding to  sheet structure. For HFP_G5c in DOPC and DPPC containing membrane, the
full width at half maximum (FWHM) is 2.4 ppm and 2.7 ppm at τ = 2ms; 2.2 ppm and 2.5 ppm
respectively at τ = 40ms. The (S/N/scan) DPPC: (S/N/scan)DOPC ≈ 1 and 5 for S0 at τ = 2ms and τ =
40ms respectively. Compared to DPPC membrane, the FWHM is a little bit narrower and
S/N/scan is much worse in DOPC membrane. This is probably because increased membrane and
peptides motion, which leads to shorter T2 for HFP. S0 and S1 are integrated with 2 ppm
integration window because the peak is narrower, and the dephasing results are superimposable
to those obtained with 3 ppm integration window for HFP_G5c with Chol_d7 labeling. The
quantitative dephasing is calculated for major  peak. HFP_G5c has much greater dephasing
with Chol_d7 than Chol_d6, which supports that major HFP_G5c is inserted into the membrane
hydrocarbon core and contacting Chol_d7 2Hs. The membrane location features are consistent
with the results studied earlier with other model membranes. However, the dephasing buildups
are not showing exponential trends. The buildup rate is much slower; the (△S/S0)16ms is 13 % in
DOPC membrane, but 40% in DPPC membrane, and the (△S/S0)40ms is 46% in DOPC membrane,
but 76% in DPPC membrane. This is likely due to increased motional averaging of the coupling.
The initial data shows that the

13

C–2H REDOR NMR method is also able to differentiate

membrane core and membrane surface contacts of peptide. More residue specific labeling is
needed to obtain a complete membrane location model of HFP in the DOPC: CHOL membrane
system. This method should also be able to be applied to study residue specific contacts of other
peptides or proteins in this model membrane system.

87

Figure 3.19

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

for HFP_G5c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. DOPC: DOPG: CHOL = 8:2:5. Chol_d7 sample contains ~
1.5Îźmol peptides, and Chol_d6 sample contains ~ 1.3Îźmol peptides.

Figure 3.20 The dephasing buildups of △S/S0 (colored squares) with error bars vs dephasing
time () for HFP_G5c in membrane composed of DOPC: DOPG: CHOL = 8:2:5 ratio. The
△S/S0 data is for the major peak of  sheet conformation and the S0 and S1 is integration with 2
ppm integration window.

88

The membrane location of HFP in 33mol % CHOL with DPPC, POPC, and DOPC membrane is
compared by analyzing the average dephasing of 8- 40ms dephasing. The ratio of dephasing with
Chol_d7 to Chol_d6 is calculated for G5 labeling, and the ratio of dephasing with Chol_d6 to
Chol_d7 is calculated for G16. The average dephasing ratio is 2.9 for G5c in both DPPC and
POPC membrane, which supports similar membrane locations for G5 in both model membrane
studied. For Gl6c, the average dephasing of Chol_d6 relative to Chol_d7 is 3.7 and 2.0 in DPPC
and POPC membrane respectively, which support major fraction of G16 is located near the
membrane surface in both model membranes studied.
3.2.7 HFPV2E 13C–2H REDOR results in membrane without CHOL
It's interesting and significant to study V2E mutant because studies have shown that this mutant
could significant decrease fusion.[16, 21] A single residue V2E mutation in the whole HIV gp41
protein eliminates fusion, and even a mixture of small fraction of the V2E mutant protein
significantly decreases fusion. These results suggest that gp41 acts as oligomers during catalysis
of fusion. It has been proposed that wild type and V2E mutant HFP peptides have different
secondary structure and membrane locations.[8, 22, 37, 38] In membrane without CHOL,
HFP_V2E mutant has greater population of helical structure than HFP wild type.[8]

13

C–31P

REDOR data supports that HFP_V2E has much shallower membrane location then HFP
wildtype with greater contact to the membrane surface phosphorous head group.[16] In this
thesis, we are studying the membrane contacts with specific regions of the membrane by 13C–2H
REDOR method, the advantage of which has been discussed earlier.
The HFP_V2E peptide is

13

CO labeled at G5 residue. The peptide is chemically synthesized by

FMOC SPPS method, and purified by reverse phase HPLC. The peptide purity was checked by

89

MALDI–TOF mass spectrometry and the purity is ≥ 90 %. The membrane is PC: PG =4:1 mole
ratio and 2H labeled with PC_d10/d8/d4.
Compared to the spectra of wild type G5, the V2E mutant G5 has a greater 175ppm chemical
shift peak. This is probably due to higher fraction of helical conformation at G5 site compared to
wild type, which is consistent with the previous study results. There are substantial dephasing
buildups for samples with both PC_d10 and PC_d8 labeling, and smallest dephasing buildup for
PC_d4 labeling. The substantial buildups are well fitted to exponential equation A × (1 – e – ).
Fitting of both PC_d10 and PC_d8 sample give 4 ~ 5 Å

13

C–2H inter-nuclear distance. These

results support HFP_V2E G5 13CO has Van der Waals contact with both PC_d10 and PC_d8 2Hs,
which also support major deeply inserted and minor shallowly inserted membrane locations of
G5 residue. And the multiple membrane location feature of HFP_V2E is similar to G5 wild type.
However, the PC_d8 dephasing buildup is much bigger for the V2E mutant than the wild type,
which suggests bigger fraction of G5 in the V2E mutant has shallower membrane location than
wild type G5. The bigger fraction of shallower membrane location for V2E molecules positively
correlates with its less fusogenicity. There is antiparallel registry distribution difference between
V2E and wild type  sheet molecules.[29] There is major longer antiparallel  sheet registry for
HFP_V2E mutant than wildtype. For V2E mutant, ~ 0.44 fraction of antiparallel  sheet registry
is with residues 20→1/1→20. For wild type, ~ 0.30 fraction of antiparallel  sheet registry is
with residues 16→1/1→16 and 17→1/1→17.[29] The hydrophobic patch is formed by the most
hydrophobic first 12 residues of HFP N-terminal based on amino acid hydrophobicity. The nonfusogenicity of V2E is associated with the hydrophobic patch size formed from its registry. For
20→1/1→20 registry, the hydrophobic patch is formed from residue 9 to residue 12. However,
for 16→1/1→16 registry, the hydrophobic patch is formed from residue 5 to residue 12 (shown

90

in Figure 3.21). Thus, the hydrophobic patch is shorter with longer 20→1/1→20 antiparallel 
sheet registry compared to 16→1/1→16 registry. The bigger fraction of shallower membrane
location for V2E is probably due to the shorter hydrophobic patch which causes less perturbation
of the membrane. A semi quantitative membrane location model is proposed, but a more
complete model will need membrane contact studies of some other residue labeling.
(a)

(b)

Figure 3.21 Schematic pictures of the hydrophobic patch (yellow shaded area) formed by the Nterminal most hydrophobic 12 residues of HFP based on amino hydrophobicity. (a) 20→1/1→20
registry for HFP_V2E mutant. (b) 16→1/1→16 registry for HFP wild type.

91

Figure 3.22

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

(top panel) and the dephasing buildups (bottom panel) of △S/S0 (colored triangles) with error
bars vs dephasing time () for HFP_V2E_G5c. The spectra displayed is for  =40ms. The data is
processed with 100Hz Gaussian line broadening and polynomial baseline correction. The
quantitative dephasing buildup is for the major  peak with 171ppm chemical shift. S0 and S1 is
integration through 3ppm integration window. The colored line is the best fit exponential buildup
curve fitted by A × (1 – e – ).

92

Table 3.5 Best-fit exponential buildup parameters for HFP_V2E_G5c in membrane without
CHOL a

a

Membrane

A

 (Hz)

r (Å)

PC_d8:PG

0.70 (3)

35 (3)

4.5 (1)

PC_d10:PG

0.82 (4)

51 (4)

4.0 (1)

A and  are fitting parameters for experimental dephasing buildup from fitting equation of A ×

(1 – e – ), and r is calculated as d= 3/2 = 4642/r3. The membrane composition is PC: PG (4:1)
with ~ 1Îźmol peptide. The peptide to lipids ratio is 1:50.

93

Figure 3.23 Semi-quantitative HFP_V2E membrane location model with major population
deeply inserted and minor population shallowly inserted to the membrane hydrophobic core.
There is more minor population compared to HFP wildtype with shallowly inserted membrane
location. How the lipids and CHOL are displaced by HFP molecules is not known and neither
the orientation the neighbor lipids and CHOL of HFP.
3.2.8 HFPL9R results in membrane without CHOL
L9R mutant is another mutant that is worth study because L9R mutant could also significantly
reduce membrane fusion.[21] The fusion reduction effect is much less for L9R than V2E mutant.
In order to help understand HIV membrane fusion mechanism, we study the structure and
membrane contacts of HFP_L9R in lipid membrane.
The HFP_L9R peptide is chemically synthesized by FMOC SPPS method and purified by
reverse phase HPLC. The peptide purity is checked by MALDI-TOF mass spectrometry, and the
purity is ≥ 90 %. Each sample contains ~ 1.5μmol peptide. And the membrane is PC: PG =4:1.
The spectra have a major 13CO peak at 176ppm for both HFP_L9R G5 and G10 samples, which
is not corresponding to  sheet structure.[27] Then 176ppm chemical shift might indicate ι
helical structure because 13CO chemical shift for glycine in Îą helical protein is 175.51 (1.23)ppm,

94

where 175.51 ppm is the average value and 1.23ppm is the standard deviation.[27] G5 sample
has biggest dephasing with PC_d10 labeling, significant dephasing with PC_d4 labeling and
negligible dephasing for PC_d8. These results suggest that there are multiple locations of L9R
G5 reside, with most population deeply inserted into the membrane hydrocarbon core and
contacting PC_d10 2Hs, and minor population located near the membrane surface contacting
PC_d4 2Hs, and almost no contact to PC_d8 2Hs. At  = 48ms, (△S/S0)d10/ (△S/S0)d4 ≈ 2:1. G10
sample has similar buildup trends with G5, biggest dephasing obtained with PC_d10 labeling,
some dephasing with PC_d4 and negligible dephasing buildup with PC_d8 labeling. It’s just the
buildup extents for both PC_d10 and PC_d4 are different. For G10 sample, there is rapid
dephasing buildup and goes up to 90% at  = 48ms, which supports most G10 13CO are making
contacts with PC_d10 2Hs, and most L9R peptide molecules are deeply inserted into the
membrane center. There are ~ 25% molecules making contact with PC_d4 2Hs near the
membrane surface based on the  = 48ms dephasing ratio, (△S/S0)d10/(△S/S0)d4 ≈ 3:1. Substantial
buildups with PC_d10 labeling are fitted by A × (1 – e – ) with two options. One is with the
limitation of A≤ 1, because A is the molecular fraction, the maximum of which is 1. The other
one is without limitation of A. The best fitted 4 ~ 5 Å distance supports Van der Waals contact
between HFP_L9R molecule and PC_d10 at both G5 and G10 13CO site.
A more quantitative fitting of the PC_d4 buildups might need lipid natural abundance correction
because the lipid 13CO group is directly bonded to the carbon that has d4 2Hs attached. So there
will be a small fraction of rapid buildup from lipid

13

CO group due to 13C–2H dipolar coupling.

The lipid 13CO–2H distance is ~ 2.1 Å according to the distance between Cα–1H and 13CO within
the same residue in  strand structure. The corresponding d would be ~ 500Hz and  ≈ 333 Hz.
And the resulting dephasing would increase rapidly to ~1 at  = 16ms according to fitting

95

equation of A × (1 – e – ) with A= 1. So the dephasing from lipid natural abundance (na) at
longer dephasing time would depend on the fraction which is A value from lipid na 13CO. For a
typical sample containing 1.5Îźmol peptide in 50Îźmol lipids,
Lip−na

50μmol×1.1%

fLip−na = Lip−na+pep−na+lab = 50μmol×1.1%+30×1.5μmol×1.1%+0.99×1.5μmol = 0.23 , where Lip-na
is the amount of lipid natural abundance
abundance
Lip-na

13

13

CO, pep–na is the amount of peptide natural

CO, and lab is the amount of labeled peptide

13

CO. The corresponding (△S/S0)2ms

is ~ 0.10. For both G5 and G10 labeling, the (△S/S0)2ms is ~ 0.10, which is mostly from

the Lip-na dephasing. For G5 sample with PC_d4 labeling, (△S/S0)48msexp is ~ 0.45, then the
corrected (△S/S0)48mscor ~ 0.22. (△S/S0)d10/ (△S/S0)d4cor ≈ 7:2, which still supports major deeply
inserted and minor shallowly inserted membrane location for G5. For G10 sample with PC_d4
labeling, (△S/S0)48msexp is ~ 0.32, then the corrected (△S/S0)48mscor ~ 0.10, which supports almost
no dephasing buildup from the peptide with PC_d4. And the (△S/S0)d10/ (△S/S0)d4cor ≈ 9:1. So,
there is likely single deeply inserted membrane location of HFP_L9R at G10 residue.
A membrane location model consistent with our experimental results is that G10 is located near
the membrane center, and G5 is also close to the membrane center but relatively shallower
compared to G10 because (△S/S0)G5/ (△S/S0)G10 ≈ 0.7: 0.9. The complete membrane location
model would need more dephasing data from other residue labeling as well as the confirmation
of the secondary structure of the HFP_L9R mutant.

96

Figure 3.24 Membrane location model of HFP_L9R mutant consistent with our REDOR
experimental data. A short helix is shown from G5 to G10 to reflect the helical conformation for
G5 and G10 residues. And the secondary structure of other residues in HFP_L9R mutant is not
determined and shown as line. The R9 sidechain is likely pointing out to the direction of the
membrane surface. However, the arginine side chain length is ~ 7.5 Å, which is shorter than the
hydrophobic thickness of half membrane leaflet. HFP_L9R mutant probably induces local
membrane thinning, and similar membrane curvature relative to membrane fusion intermediate.
This might help explain the fusogenicity of L9R mutant.

97

(a)

(b)

(c)

Figure 3.25 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra (a)
and the dephasing buildups (b) of △S/S0 (colored squares) with error bars vs dephasing time ()
for HFP_L9R_G5c. The spectra displayed is for  =40ms. The data is processed with 100Hz
Gaussian line broadening and polynomial baseline correction. The quantitative dephasing
buildup is for the major peak with 176 ppm chemical shift. S0 and S1 is integration through 3ppm
integration window. The colored line is the best fit exponential buildup curve fitted by A × (1 – e
–

). Fitting in (b) in done with A≤1, the χ2 for d10 fitting is 5. Fitting in (c) is done with no

limitation on A, and the χ2 for d10 fitting is 2.5.

98

(a)

(b)

(c)

Figure 3.26 13C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra (a)
and the dephasing buildups (b) of △S/S0 (colored squares) with error bars vs dephasing time ()
for HFP_L9R_G10c. The spectra displayed is for  =40ms. The data is processed with 100Hz
Gaussian line broadening and polynomial baseline correction. The quantitative dephasing
buildup is for the major peak with 176 ppm chemical shift. S0 and S1 is integration through 3ppm
integration window. The colored line is the best fit exponential buildup curve fitted by A × (1 – e
–

). Fitting in (b) in done with A≤1, the χ2 for d10 fitting is 45. Fitting in (c) is done with no

limitation on A, and the χ2 for d10 fitting is 10.

99

Table 3.6 Best-fit exponential buildup parameters for HFP_L9R in membrane without CHOL a

a

Peptide

Membrane

A

 (Hz)

r (Å)

G5

PC_d10:PG

1.0 (3) b

21(7)

5.3 (6)

G5

PC_d10:PG

1.7 (7) c

11(5)

6.5(10)

G10

PC_d10:PG

1.0 (2) b

37 (11)

4.4 (4)

G10

PC_d10:PG

1.5 (2) c

30 (7)

5.4 (4)

A and  are fitting parameters for experimental dephasing buildup from fitting equation of A ×

(1 – e – ) with limitation of A≤ 1, and r is calculated as d= 3/2 = 4642/r3. The membrane
composition is PC: PG (4:1) with ~ 1.5Îźmol peptide. The peptide to lipids ratio is 3:100. The
uncertainty is in the parenthesis, and a sample 1.0 (3) means an error of Âą0.3.
b

data is fitted with limitation of A≤ 1.

c

data is fitted with no limitation on A .

To summarize, our study supports that L9R mutant is non β sheet structure at G5 and G10
residues. Membrane contacts between the peptide backbone and specific regions of membrane
support multiple membrane locations for G5 with major deeply inserted population contacting
the membrane center and a minor shallowly inserted population contacting the membrane surface

100

lipid carbonyl region. Different from G5, G10 likely has a single membrane location which is
deeply inserted into the membrane center contacting the lipid tail.
3.2.9 KALP results in membrane without CHOL
KALP peptide is a transmembrane α helical peptide in membrane.[39, 40] The 13C–2H REDOR
method is applied to study the residue specific membrane locations of helical transmembrane
peptide. The KALP peptide is chemically synthesized by Fmoc SPPS method and

13

CO labeled

at A5, A7, A17 and A19. The crude peptides are purified by reverse phase HPLC with a C4
column. The purification program is different than any HFP purification program and is shown
in appendix. To make NMR sample, 9.7 mg peptides are weighed and dissolved in 3mL 2,2,2trifuoroethanol, 3mL 1,1,1,3,3,3-hexafluoroisopropanol, and 4.5 mL chloroform. The mixture is
then divided into 3 portions, and each portion is added to PC_d10: PG, PC_d8: PG and PC_d4:
PG lipid films, respectively. So each membrane is feed 1.4Îźmol KALP peptides. The peptide
lipids mixture is mixed by sonication, and solvent is removed by a steam of N2 (g), followed by
vacuum overnight to remove any residual solvents. The dry film is then hydrated in 3 mL buffer
with pH 7.4, homogenized by 10 freeze– thaw cycles. Extra 20 mL buffer is added. The pellet is
harvested after centrifugation at ~ 270000g. Any unbound KALP peptide is in the supernatant.
All KALP samples with Ala 13CO labeling have dominant peak with ~ 179 ppm chemical shift,
which is consistent with Îą helical conformation. Dephasing of the four Ala labeling show similar
dephasing buildup trends, with dephasing (△S/S0) d10>d8>d4. There are dephasing buildups for d10
and d8, and no dephasing buildups for d4 labeled samples. At  = 48ms, (△S/S0) d10:d8 ≈ 4:3, 2:1,
2:1 and 3:2 ratio for A5, A7, A17 and A19, respectively. The bigger dephasing with d10 than d8
labeling supports that there are major contacts with d10 then with d8. The substantial
experimental dephasing buildups show exponential trends for A5, A7 and A17, and are fitting

101

with equation A × (1 – e – ). It's not exponential buildup trend for A17 data, which need to be
fitted by other fitting methods to find out the inter-nuclear distance, which will be used to
calculate the membrane location of the labeled

13

CO nucleus. The similar buildup extent at

longer dephasing time with much slower buildup rate might suggest a single longer inter-nuclear
distance.
To summarize, our studies show helical secondary structure of KALP peptide at A5, A7, A17
and A19 residue. Membrane contact study between the peptide backbone and specific regions of
membrane show similar membrane contacts at A5, A7 and A17 residue, with major contact with
membrane center PC_d10 and minor contact with PC_d8. A19 has similar dephasing value at
longer dephasing time (τ = 48ms) for PC_d10, d8 and d4 labeling, but with a much slower
buildup, which suggests a longer

13

C-2H inter-nuclear distance, and thus A19 is much further

away from membrane center. Figure 3.35 shows a membrane location model of KALP peptide
consistent with our experiment results. The hydrophobic length of LA residues is 25.5Å, which
is shorter than the ~ 31Å of the DPPC membrane hydrophobic thickness.[39, 40] The lysine
sidechains are pointing out to interact with the aqueous phase near the phosphate group, and the
hydrophobic length of the peptide is extended by lysine sidechains. This is consistent with the
proposed snorkeling effect of charged residues like lysine in transmembrane peptides.[41] The
different membrane locations are probably due to different snorkeling geometries of the lysine.
According to the dephasing at τ=48ms, the molecular population for making close contact with
PC_d10 relative to PC_d8 is ~ 3:2 for all A5, A7, A17 and A19 residues.

102

Figure 3.27

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

for KALP_A5c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. PC: PG =4:1 mole ratio.

Figure 3.28 The KALP_A5c dephasing buildup of △S/S0 (colored squares) with error bars vs
dephasing time. S0 and S1 is integration through 3ppm integration window. The colored line is
the best fit exponential buildup curve fitted by A × (1 – e – ). PC: PG =4:1 mole ratio.

103

Figure 3.29

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

for KALP_A7c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. PC: PG =4:1 mole ratio.

Figure 3.30 The KALP_A7c dephasing buildup of △S/S0 (colored squares) with error bars vs
dephasing time. S0 and S1 is integration through 3ppm integration window. The colored line is
the best fit exponential buildup curve fitted by A × (1 – e – ). PC: PG =4:1 mole ratio.

104

Figure 3.31

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

for KALP_A17c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. PC: PG =4:1 mole ratio.

Figure 3.32 The KALP_A17c dephasing buildups of △S/S0 (colored squares) with error bars vs
dephasing time. S0 and S1 is integration through 3ppm integration window. The colored line is
the best fit exponential buildup curve fitted by A × (1 – e – ). PC: PG =4:1 mole ratio.

105

Figure 3.33

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

for KALP_A19c at  = 40ms. The data is processed with 100 Hz Gaussian line broadening and
polynomial baseline correction. PC: PG =4:1 mole ratio.

Figure 3.34 The KALP_A19c dephasing buildup of △S/S0 (colored squares) with error bars vs
dephasing time. S0 and S1 is integration through 3ppm integration window. PC: PG =4:1 mole
ratio.

106

Table 3.7 Best-fit exponential buildup parameters for KALP in membrane without CHOL a

a

Peptide

Membrane

A

 (Hz)

r (Å)

A5

PC_d8:PG

0.47 (21)

10 (5)

6.8 (1.2)

A5

PC_d10:PG

0.47 (22)

14 (8)

6.0 (1.2)

A7

PC_d8:PG

0.44 (26)

10 (7)

6.8 (1.6)

A7

PC_d10:PG

0.84 (73)

10 (10)

6.8 (2.3)

A17

PC_d8:PG

0.22 (9)

19 (10)

5.5 (1.0)

A17

PC_d10:PG

0.72 (40)

10 (7)

6.8 (1.5)

A and  are fitting parameters for experimental dephasing buildup from fitting equation of A ×

(1 – e – ), and r is calculated as d= 3/2 = 4642/r3. The membrane composition is PC: PG (4:1)
with ~ 1Îźmol peptide.

107

(d) KALP sequence: Acetyl-GKKLALALALALALALALALKKA-NH2
Figure 3.35 The membrane location model for KALP peptide with one representative lysine
sidechain near N- and C- terminal. (a) Major populations of A5 and A7 have close contact to
PC_d10 located near the membrane center. (b) Major populations of A17 and A19 have close
contact to PC_d10 located near the membrane center, but with A19 further away to PC_d10
compared to A17. (c) Significant populations of A5, A7, A17 and A19 make close contact of
PC_d8. (d) Amino acid sequence for KALP peptide. Snorkeling effect of terminal lysine
sidechains help extend the hydrophobic length of the peptide by pointing out to the aqueous
surface near the phosphate group. The molecular population of (a): (b): (c) ≈ 3:3:2, because the
dephasing for PC_d10: PC_d8 is ~ 3:2 for all A5, A7, A17 and A19 residues at τ=48ms. There
isn’t substantial dephasing buildup for PC_d4 might because the snorkeling effect of lysing
sidechains might make lysine sidechains displace PC_d4 and thus enlarge the inter-nuclear
distance between labeled

13

CO and PC_d4 and is beyond the ~ 8Å detection limit of

REDOR.

108

13

C-2H

REFERENCES

109

REFERENCES

1.

Dupuy, A.D. and D.M. Engelman, Protein area occupancy at the center of the red blood
cell membrane. Proceedings of the National Academy of Sciences of the United States of
America, 2008. 105(8): p. 2848-2852.

2.

Michalek, M., C. Aisenbrey, and B. Bechinger, Investigation of membrane penetration
depth and interactions of the amino-terminal domain of huntingtin: refined analysis by
tryptophan fluorescence measurement. European Biophysics Journal with Biophysics
Letters, 2014. 43(8-9): p. 347-360.

3.

Lin, P.H., et al., Membrane attachment and structure models of lipid storage droplet
protein 1. Biochimica Et Biophysica Acta-Biomembranes, 2014. 1838(3): p. 874-881.

4.

Su, Y., R. Mani, and M. Hong, Asymmetric insertion of membrane proteins in lipid
bilayers by solid-state NMR paramagnetic relaxation enhancement: A cell-penetrating
peptide example. Journal of the American Chemical Society, 2008. 130(27): p. 88568864.

5.

Gabrys, C.M. and D.P. Weliky, Chemical shift assignment and structural plasticity of a
HIV fusion peptide derivative in dodecylphosphocholine micelles. Biochimica Et
Biophysica Acta-Biomembranes, 2007. 1768(12): p. 3225-3234.

6.

Jaroniec, C.P., et al., Structure and dynamics of micelle-associated human
immunodeficiency virus gp41 fusion domain. Biochemistry, 2005. 44(49): p. 1616716180.

7.

Li, Y.L. and L.K. Tamm, Structure and plasticity of the human immunodeficiency virus
gp41 fusion domain in lipid micelles and bilayers. Biophysical Journal, 2007. 93(3): p.
876-885.

8.

Gabrys, C.M., et al., Solid-State Nuclear Magnetic Resonance Measurements of HIV
Fusion Peptide (CO)-C-13 to Lipid P-31 Proximities Support Similar Partially Inserted
Membrane Locations of the alpha Helical and beta Sheet Peptide Structures. Journal of
Physical Chemistry A, 2013. 117(39): p. 9848-9859.

9.

Qiang, W., M.L. Bodner, and D.P. Weliky, Solid-state NMR Spectroscopy of human
immunodeficiency virus fusion peptides associated with host-cell-like membranes: 2D
correlation spectra and distance measurements support a fully extended conformation
and models for specific antiparallel strand registries. Journal of the American Chemical
Society, 2008. 130(16): p. 5459-5471.

10.

Yang, J., C.M. Gabrys, and D.P. Weliky, Solid-state nuclear magnetic resonance
evidence for an extended beta strand conformation of the membrane-bound HIV-1 fusion
peptide. Biochemistry, 2001. 40(27): p. 8126-8137.

110

11.

Toke, O., et al., Secondary structure and lipid contact of a peptide antibiotic in
phospholipid Bilayers by REDOR. Biophysical Journal, 2004. 87(1): p. 662-674.

12.

Xie, L., et al., Residue-specific membrane location of peptides and proteins using
specifically and extensively deuterated lipids and C-13-H-2 rotational-echo doubleresonance solid-state NMR. Journal of Biomolecular Nmr, 2013. 55(1): p. 11-17.

13.

Schmidt, A., R.A. McKay, and J. Schaefer, INTERNUCLEAR DISTANCE
MEASUREMENT BETWEEN DEUTERIUM (I = 1) AND A SPIN-1/2 NUCLEUS IN
ROTATING SOLIDS. Journal of Magnetic Resonance, 1992. 96(3): p. 644-650.

14.

Gullion, T., R. Kishore, and T. Asakura, Determining dihedral angles and local structure
in silk peptide by C-13-H-2 REDOR. Journal of the American Chemical Society, 2003.
125(25): p. 7510-7511.

15.

Hirsh, D.J., et al., A new monofluorinated phosphatidylcholine forms interdigitated
bilayers. Biophysical Journal, 1998. 75(4): p. 1858-1868.

16.

Qiang, W., Y. Sun, and D.P. Weliky, A strong correlation between fusogenicity and
membrane insertion depth of the HIV fusion peptide. Proceedings of the National
Academy of Sciences of the United States of America, 2009. 106(36): p. 15314-15319.

17.

Xie, L., Solid State Nuclear Magnetic Rosonance Studies of Structures and Membrane
Locations of Peptides, in Chemistry. 2014, Michigan State University.

18.

Jia, L.H., et al., REDOR solid-state NMR as a probe of the membrane locations of
membrane-associated peptides and proteins. Journal of Magnetic Resonance, 2015. 253:
p. 154-165.

19.

Sack, I., et al., Solid-state NMR determination of peptide torsion angles: Applications of
H-2-dephased REDOR. Journal of the American Chemical Society, 2000. 122(49): p.
12263-12269.

20.

Schmick, S.D. and D.P. Weliky, Major Antiparallel and Minor Parallel beta Sheet
Populations Detected in the Membrane-Associated Human Immunodeficiency Virus
Fusion Peptide. Biochemistry, 2010. 49(50): p. 10623-10635.

21.

Freed, E.O., et al., A MUTATION IN THE HUMAN-IMMUNODEFICIENCY-VIRUS
TYPE-1 TRANSMEMBRANE GLYCOPROTEIN-GP41 DOMINANTLY INTERFERES
WITH FUSION AND INFECTIVITY. Proceedings of the National Academy of Sciences
of the United States of America, 1992. 89(1): p. 70-74.

22.

Qiang, W. and D.P. Weliky, HIV Fusion Peptide and Its Cross-Linked Oligomers:
Efficient Syntheses, Significance of the Trimer in Fusion Activity, Correlation of beta
Strand Conformation with Membrane Cholesterol, and Proximity to Lipid Headgroups.
Biochemistry, 2009. 48(2): p. 289-301.

111

23.

Gullion, T., Introduction to rotational-echo, double-resonance NMR. Concepts in
Magnetic Resonance, 1998. 10(5): p. 277-289.

24.

Zheng, Z., et al., Conformational flexibility and strand arrangements of the membraneassociated HIV fusion peptide trimer probed by solid-state NMR spectroscopy.
Biochemistry, 2006. 45(43): p. 12960-12975.

25.

Bak, M., J.T. Rasmussen, and N.C. Nielsen, SIMPSON: A general simulation program
for solid-state NMR spectroscopy. Journal of Magnetic Resonance, 2000. 147(2): p. 296330.

26.

Xie, L., et al., Multiple Locations of Peptides in the Hydrocarbon Core of Gel-Phase
Membranes Revealed by Peptide C-13 to Lipid H-2 Rotational-Echo Double-Resonance
Solid-State Nuclear Magnetic Resonance. Biochemistry, 2015. 54(3): p. 677-684.

27.

Zhang, H.Y., S. Neal, and D.S. Wishart, RefDB: A database of uniformly referenced
protein chemical shifts. Journal of Biomolecular Nmr, 2003. 25(3): p. 173-195.

28.

Brugger, B., et al., The HIV lipidome: A raft with an unusual composition. Proceedings of
the National Academy of Sciences of the United States of America, 2006. 103(8): p.
2641-2646.

29.

Schimick, S., High resolution tertiary structure of the membrane-associated HIV fusion
peptides by solid state nuclear magnetic resonance, in Chemistry. 2012, Michigan State
University.

30.

Banerjee, K. and D.P. Weliky, Folded Monomers and Hexamers of the Ectodomain of the
HIV gp41 Membrane Fusion Protein: Potential Roles in Fusion and Synergy Between the
Fusion Peptide, Hairpin, and Membrane-Proximal External Region. Biochemistry, 2014.
53(46): p. 7184-7198.

31.

Lorizate, M., et al., Comparative lipidomics analysis of HIV-1 particles and their
producer cell membrane in different cell lines. Cellular Microbiology, 2013. 15(2): p.
292-304.

32.

Qiang, W., J. Yang, and D.P. Weliky, Solid-state nuclear magnetic resonance
measurements of HIV fusion peptide to lipid distances reveal the intimate contact of beta
strand peptide with membranes and the proximity of the Ala-14-Gly-16 region with lipid
headgroups. Biochemistry, 2007. 46(17): p. 4997-5008.

33.

Ratnayake, P.U., et al., pH-dependent vesicle fusion induced by the ectodomain of the
human immunodeficiency virus membrane fusion protein gp41: Two kinetically distinct
processes and fully-membrane-associated gp41 with predominant beta sheet fusion
peptide conformation. Biochimica Et Biophysica Acta-Biomembranes, 2015. 1848(1): p.
289-298.

34.

Ghosh, U., et al., Closed and Semiclosed Interhelical Structures in Membrane vs Closed
and Open Structures in Detergent for the Influenza Virus Hemagglutinin Fusion Peptide

112

and Correlation of Hydrophobic Surface Area with Fusion Catalysis. Journal of the
American Chemical Society, 2015. 137(24): p. 7548-7551.
35.

Lai, A.L., et al., Fusion Activity of HIV gp41 Fusion Domain Is Related to Its Secondary
Structure and Depth of Membrane Insertion in a Cholesterol-Dependent Fashion. Journal
of Molecular Biology, 2012. 418(1-2): p. 3-15.

36.

Ratnayake, P.U., et al., Full-length trimeric influenza virus hemagglutinin II membrane
fusion protein and shorter constructs lacking the fusion peptide or transmembrane
domain: Hyperthermostability of the full-length protein and the soluble ectodomain and
fusion peptide make significant contributions to fusion of membrane vesicles. Protein
Expression and Purification, 2016. 117: p. 6-16.

37.

Morris, K.F., X.F. Gao, and T.C. Wong, The interactions of the HIV gp41 fusion peptides
with zwitterionic membrane mimics determined by NMR spectroscopy. Biochimica Et
Biophysica Acta-Biomembranes, 2004. 1667(1): p. 67-81.

38.

Pereira, F.B., et al., Permeabilization and fusion of uncharged lipid vesicles induced by
the HIV-1 fusion peptide adopting an extended conformation: Dose and sequence effects.
Biophysical Journal, 1997. 73(4): p. 1977-1986.

39.

de Planque, M.R.R., et al., Sensitivity of single membrane-spanning alpha-helical
peptides to hydrophobic mismatch with a lipid bilayer: Effects on backbone structure,
orientation, and extent of membrane incorporation. Biochemistry, 2001. 40(16): p. 50005010.

40.

de Planque, M.R.R., et al., Different membrane anchoring positions of tryptophan and
lysine in synthetic transmembrane alpha-helical peptides. Journal of Biological
Chemistry, 1999. 274(30): p. 20839-20846.

41.

Strandberg, E., et al., Lipid dependence of membrane anchoring properties and
snorkeling behavior of aromatic and charged residues in transmembrane peptides.
Biochemistry, 2002. 41(23): p. 7190-7198.

113

Chapter 4 - Preferential Contacts of HFP with CHOL vs PC lipid
4.1 Introduction
CHOL is an important component in membranes.[1] HIV is an enveloped virus, and CHOL
constitutes about 30mol % in both HIV host cell plasma membrane and HIV virus membrane.[2]
Studies have shown that depletion of cellular CHOL reduces HIV-1 binding to cells and inhibits
HIV virus induced cell-cell fusion.[3] The significance of CHOL on HIV makes it important to
study how CHOL interfere with HIV induced fusion. HIV infection is initiated via viral gp120
binding to host cell CD4 receptor and co-receptor sequentially, which leads to conformational
change of gp120 and get gp41 exposed to interact the host cell membrane and catalyze
membrane fusion. Protein gp41 undergoes extended pre-hairpin intermediate (PHI) and folds
into final hairpin or six helical bundle state (SHB) (see figure 1.16 in chapter 1).[4-8] The HIV
gp41 N-terminal fusion peptide region (HFP) itself can induce both lipid mixing and contents
leakage of vesicles.[9-13] Both HFP_V2E and gp41_V2E mutant abrogate membrane
fusion.[14-16] Thus HFP has been widely studied as a model protein to understand its structure
and function in HIV-host cell membrane fusion. HFP induced model membrane fusion studies
indicate that there is faster fusion in membranes that contain CHOL and more fusion when there
is more CHOL.[12, 17-19] The vesicle fusion rate is faster for three different oligomer states of
HFP, monomer, dimer and trimer in membrane with CHOL than in membrane without
CHOL.[18] Vesicle fusion studies suggest that membranes with a coexistence of liquid ordered
and liquid disordered domains undergo more fusion, and fluorescence microcopy studies
suggests that HFP binds to membranes and promotes membrane fusion at the interface between
CHOL rich liquid ordered domains and liquid disordered domains.[17] Although all these studies
have illustrated that CHOL is important to HFP induced efficient membrane fusion, there is little

114

information about how HFP contact the nearby lipid and the affinity between HFP and the
membrane CHOL.
To study contact and affinity between HFP and membrane CHOL, 13C–2H rotational echo double
resonance (REDOR) solid-state NMR is employed to measure the dipolar couplings (d’s)
between
distance

13

CO–HFP and 2H–CHOL in this thesis. The d depends on the
(r)

as

d(Hz)=4642/r(Å)3.[20-22]

HFP

sequence

in

13

C–2H inter-nuclear

gp41

studied

is

AVGIGALFLGFLGAAGSTMGARS. The model membrane is 1, 2-dipalmitoyl-sn-glycero-3phosphocholine (PC): 1, 2-dipalmitoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (sodium salt) (PG)
in 4:1mol ratio typically with 33% CHOL. This composition reflects features of HIV host cell
membrane including significant fraction of PC lipid, anionic lipid fraction of ~ 0.15, and CHOL
fraction of ~ 0.3.[2] To study the HFP-CHOL contact, we also studied membrane with 20 %
CHOL and 11 % CHOL.
HFP forms majorly  sheet conformation in membrane with ~ 30mol % CHOL.[23] Therefore,
one advantage of Gly-13CO labeling in the  sheet is a Gly-13CO chemical shift (~171 ppm)
different from lipid and natural abundance chemical shifts of other residues (~175 ppm).[24, 25]
There are distributions of antiparallel registry length of the first 16 residues of HFP in membrane
with a total fraction of ~ 30% of antiparallel 16→1/1→16 and 17→1/1→17 registries.[23, 26]
The 13CO labeled at G16 (G16c) spectrum also has a higher chemical shift (~174 ppm)  sheet
peak that might due to HFP molecules with shorter antiparallel  sheet registries that do not
include G16 residue.[27] The membrane is 2H labeled either at PC lipids or CHOL displayed in
chapter 3. In membrane without protein, PC_d10 and Chol_d7 deuterons are located near the
membrane center, PC_d4 and Chol_d6 deuterons are located near the membrane surface, while

115

PC_d8 deuterons are located in the middle of PC lipid acyl chain. The overall preferential
contact with CHOL vs PC is through comparing experimental dephasing data with CHOL
labeling and PC lipid labeling. All the REDOR data was acquired with sample temperature at
about -30 °C to minimize sample motion which would reduce signal intensity and dephasing by
increasing T2 relaxation rate and reducing dipolar coupling, respectively.[27, 28]
4.2 Results
4.2.1 Experimental 13C–2H REDOR results for HFP_G5, G10 and G16c
The contact between HFP and membrane components is evaluated with same membrane
composition but different 2H membrane locations, shown in chapter 2. In Membrane PC: PG:
CHOL (8:2:5) which contains ~33mol% CHOL, G5c has biggest dephasing with Chol-d7,
significant dephasing with PC_d10 and relatively smaller dephasing with PC_d8, which suggests
closer contact between G5c and Chol_d7 deuterons than PC_d10 and PC_d8 deuterons (figure
4.1). Since the deuteron ratio is 7:16 in sample with Chol_d7: sample with PC_d10 and the
deuteron membrane location is similar at least in membrane without protein, the bigger
dephasing in membrane with Chol_d7 than PC_d10 evidences preferential contact with Chol_d7
compared to PC_d10. The greater dephasing in membrane with PC_d10 and Chol_d7 than
PC_d8 suggests most HFP G5

13

CO is deeply inserted into the membrane hydrophobic core,

assuming that the membrane structure locally around peptide is similar to the structure in the
absence of peptide. The contact between G5c and membrane is also assessed with smaller CHOL
content. In membrane with PC: PG: CHOL (8:2:2.5) which contains ~ 20% CHOL, comparable
dephasing is obtained with either Chol_d7 or PC_d10 deuteron labeling where the deuteron ratio
of Chol_d7: PC_d10 is only 7: 32 (figure 4.2). The CHOL amount is further reduced to ~ 11mol %

116

with membrane PC: PG: CHOL (8:2:1.25), and significant dephasing is still observed (figure
4.2), which further supports considerable preferential contact between G5c and Chol_d7.
Similar dephasing trend is also observed at G10

13

CO (figure 4.3). In membrane with PC: PG:

CHOL (8:2:2.5) which contains ~ 20% CHOL, comparable dephasing is observed with either
Chol_d7 or PC_d10 deuteron labeling where the deuteron ratio is only 7:32. The dephasing
buildups are also similar to G5c (figure 4.2), which suggests similar contact with Chol_d7
between G10 and G5 13CO.
Previous studies show that G16 in HFP is majorly located near the phosphorous head group
region. Our G16

13

CO REDOR dephasing plots all show minimal dephasing in membrane PC:

PG: CHOL (8:2:5) with deuteron labeling in PC_d10, d8, d4 and Chol_d7 (figure 4.4), which
indicates that most G16 13CO don't have close contact with membrane hydrophobic core and is
thus not inserted into the membrane interior. However, G16 13CO shows significant dephasing in
the same membrane with Chol_d6 deuterons labeling which is located near the membrane
surface (figure 4.5).[27, 29, 30] Smaller amount of CHOL (20mol %) containing membrane is
also tested (figure 4.5), and the dephasing is only reduced by ~ 20%, which supports preferential
contact between G16c and Chol_d6. If there is no preferential contact between G16c and
Chol_d6 compared to PC, there would be less or equal dephasing with Chol_d6 labeling than
with PC_d4 labeling because Chol_d6: PC_d4 = 5:8 and the corresponding deuterons ratio is
30:32.
The substantial dephasing buildups are semi-quantitatively fitted to exponential buildup equation
A × (1 – e – ). A is the fraction of molecule with d  3/2 (fully explained in chapter 3).[31] The
exponential fitting approach is chosen majorly due to T1 relaxation effects of 2H (shown in (table

117

4.1)). The data is also alternatively fitted with isolated

13

CO–2H spin pairs with a single dipolar

coupling (d) by SIMPSON (table 4.2). The data is well fitted by fitting approach П (discussed in
chapter 3) with 3 populations, two fractions with different d’s and the rest population with d=0.
Experimental parameters are incorporated into the SIMPSON simulation, including the MAS
frequency, radio frequency fields of the pulses,
2

13

C chemical shift offset and anisotropy and the

H quadrupole anisotropy. But the SIMPSON program does not take relaxation into account.[32]

Figure 4.1

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

(top panel) of membrane - associated HFP_G5c at  = 40ms and quantitative △S/S0 buildups for
the major  peak in membrane with different 2H labeling. The colored line is best fitted curve by
equation of A × (1 – e – ).

118

Figure 4.2

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

(top panel) of membrane – associated HFP_G5c at  = 40ms and quantitative △S/S0 buildups for
the major  peak in membrane with different 2H labeling. The colored line is best fitted curve by
equation of A × (1 – e – ).

119

Figure 4.3

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

(top panel) of membrane – associated HFP_G10c at  = 40ms and quantitative △S/S0 buildups
for the major  peak in membrane with different 2H labeling. The colored line is best fitted curve
by equation of A × (1 – e – ).

120

Figure 4.4

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

(top panel) of membrane – associated HFP_G16c at  = 40ms and quantitative △S/S0 buildups
for the major  peak in membrane with different 2H labeling.

121

Figure 4.5

13

C-detect / 2H-dephase REDOR S0 (black) and S1 (colored) experimental spectra

(top panel) of membrane – associated HFP_G16c at  = 40ms and quantitative △S/S0 buildups
for the major  peak at 171 ppm in membrane with different 2H labeling. The colored line is best
fitted curve by equation of A × (1 – e – ).

122

Table 4.1 Best-fit exponential buildup parameters for HFP in membrane with CHOL a

a

Peptide

PC lipid label

fChol

A

 (Hz)

r (Å)

G5c

PC_d8

0.33

0.46 (25)

16 (11)

5.8 (1.4)

G5c

PC_d10

0.33

0.77 (11)

20 (4)

5.4 (4)

G5c

Chol_d7

0.33

0.83 (5)

40 (5)

4.3 (2)

G5c

PC_d10

0.20

0.81 (3)

35 (2)

4.5 (1)

G5c

Chol_d7

0.20

0.75 (4)

42 (5)

4.2 (2)

G5c

Chol_d7

0.11

0.68 (9)

29 (7)

4.7 (4)

G10c

PC_d10

0.20

0.80 (4)

34 (3)

4.5 (1)

G10c

Chol_d7

0.20

0.77 (2)

41 (2)

4.2 (1)

G16c

Chol_d6

0.33

0.67 (5)

64 (10)

3.6 (2)

G16c

Chol_d6

0.20

0.46 (3)

61 (10)

3.7 (2)

A and  are fitting parameters for experimental dephasing buildup from fitting equation of A ×

(1 – e – ), and r is calculated as d= 3/2 = 4642/r3. The peak analyzed is the major  peak. The
sample typically contains ~ 1Îźmol peptide with peptide: phospholipid ratio of 1:50. Phospholipid
is DPPC: DPPG with 4:1 mole ratio. fChol is the CHOL amount relative to the total amount of
lipids including CHOL.

123

Table 4.2 Alternative SIMPSON simulation results a

a

Peptide

PC lipid label

fChol

f1

d1/Hz

r1/ Å

f2

d2 /Hz

r2/ Å

G5c

PC_d8

0.33

0.24

54

4.4

0.43

19

6.3

G5c

PC_d10

0.33

0.07

80

3.9

0.24

24

5.8

G5c

Chol_d7

0.33

0.44

81

3.9

0.56

22

6.0

G5c

PC_d10

0.20

0.34

85

3.8

0.58

22

6.0

G5c

Chol_d7

0.20

0.39

81

3.9

0.53

23

5.9

G5c

Chol_d7

0.11

0.23

66

4.1

0.47

23

5.9

G10c

PC_d10

0.20

0.33

83

3.8

0.58

22

6.0

G10c

Chol_d7

0.20

0.37

83

3.8

0.57

23

5.9

G16c

Chol_d6

0.33

0.40

77

3.9

0.48

33

5.2

G16c

Chol_d6

0.11

0.36

78

3.9

0.29

21

6.0

distance r was calculated from r= (4642/d)1/3. The membrane is DPPC: DPPG with 4:1 mole

ratio. fChol is the CHOL amount relative to the total amount of lipids. The best fitting results are
based on minimum χ2 achieved.
4.2.2 Free energy of preferential contact of HFP with CHOL vs PC
The membrane in our studies contains ~7–30mol % cholesterol, and there are liquid-ordered and
solid-ordered phase separations at room temperature according to the DPPC/CHOL binary phase

124

diagram.[33] The CHOL: PC ratio is 0.30:0.70 and 0.07:0.93 in liquid-ordered and solid-ordered
phases, respectively. There would be experimental dephasing as long as HFP

13

CO nucleus

makes contact with at least one labeled molecule. The average dephasing value (C) of 8-48ms or
8-40ms (when 48ms data isn’t measured) is a more reliable metric of peptide 13CO-membrane 2H
contact. The error of C is calculated from equation 𝜎(𝐶) =

1

2
∑𝑖=𝑛
𝑖=1 𝜎 ∆𝑆 and n=5 and 6 for

𝑛√

(

)

𝑆0 𝑖

average dephasing of 8–40ms and 8–48ms, respectively (table 4.3). G5c and G10c show 90% of
maximum C (in sample of PC_d10: PG with 4:1 mole ratio) in both PC_d10: PG: CHOL and PC:
PG: Chol_d7 membranes with 8:2:2.5 ratios. We consider a model that the nucleus contacts two
molecules. Because d depends on 1/r3, the dipolar coupling is dominated by the closest coupled
spin pair. Therefore, dephasing for contact of 13CO with two 2H labeled molecules is considered
equivalent to dephasing with one labeled molecule because there are many 2H nuclei in a labeled
molecule. The fractional probabilities (P) for HFP contacting a nearby lipid which result in
dephasing in our samples are statistically calculated (table 4.4). The C’s depend strongly on PC:
CHOL ratios in our samples which evidence peptide binding to both gel and liquid-ordered
phases because there is significant dephasing buildup for HFP in pure gel phase membrane
(fChol= 0%), pure liquid- ordered phase membrane (fChol= 33%) and a mixture of both phases
membrane (fChol= 20%). The C values also support preferential binding of HFP to CHOL vs PC.
For example, CChol_d7:CPC_d10 ≈ 1.7 and PChol: PPC ≈ 0.7 for PC: PG: CHOL (8:2:5) and
CChol_d7:CPC_d10 ≈ 1.0 and PChol: PPC ≈ 0.4 for PC: PG: CHOL (8:2:2.5). To evaluate the
preferential contact of HFP to CHOL, we consider that CPC= PPC × B and CChol = PChol × W × B
where B is a proportionality constant and W is the preference for peptide binding to CHOL vs
PC. W is evaluated from comparison of C and P values for different samples. For the same
peptide labeling and membrane composition, W = (CChol × PPC) / (CPC × PChol). A model for W is

125

W = exp(-△GChol_PC/RT) where △GChol_PC is the free energy difference between peptide contact
with CHOL vs PC. Then GPC-Chol= RTln(S/S0Chol-d7/S/S0PC-d10PPC/PChol) for the same
peptide labeling and membrane composition, and S/S0 is the average dephasing of 8-48ms
dephasing times, which is also denoted as C values. For a typical W=2.5 and T=300 K, △G ≈
0.57(5) kcal.mol-1 for all 6 independent samples, and the number in the parenthesis is the
standard deviation, which is 0.05 kcal/mol (table 4.5).

126

Table 4.3 C values for different peptide and membrane labeling a

a

Peptide

PC lipid label

fChol

C value

G5c

PC_d8

0.33

0.157 (8)

G5c

PC_d10

0.33

0.304 (5)

G5c

PC_d10

0.33

0.272 (6)b

G5c

Chol_d7

0.33

0.522 (5)

G5c

PC_d10

0.20

0.471 (5)

G5c

PC_d10

0.20

0.433 (5)b

G5c

Chol_d7

0.20

0.479 (6)

G5c

Chol_d7

0.11

0.350 (4)

G10c

PC_d10

0.20

0.460 (3)

G10c

Chol_d7

0.20

0.487 (4)

G16c

Chol_d6

0.33

0.477 (16)b

G16c

Chol_d6

0.20

0.329 (8)b

The C values are typically the average dephasing S/S0 of 8-48ms dephasing times, except

for G16c samples, it’s 8-40ms dephasing times. The phospholipid composition is PC: PG (4:1),
and fChol the fraction of CHOL of total lipids. Example C value of 0.157 (8) means C value is
0.157 with error of Âą 0.008.
b

The C values is the average dephasing S/S0 of 8-40ms dephasing times.

127

Table 4.4 Fractional probabilities of 13CO making contact to PC or CHOL with two molecules
contact model a

Membrane

PPC

PChol

PC: PG: CHOL (8:2:5)

0.782

0.556

PC: PG: CHOL (8:2:2.5)

0.870

0.360

0.917

0.210

PC: PG: CHOL
(8:2:1.25)

a

The fractional probabilities are based on a model that the 13CO nucleus contacts two molecules.

PPC=fPC2+2×fPC(fChol+fPG)= fPC2+2×fPC(1-fPC)= fPC(2-fPC), similarly, PChol= fChol(2-fChol). fPC and
fChol are the mole fraction of PC and CHOL in the relevant membrane compositions, respectively.

128

Table 4.5 △GPC-Chol values (the energy difference for peptide binding to PC vs CHOL) for
different samples

a

Peptide label

PC: PG: CHOL

Membrane labels

△GPC-CHOL/kcalmol-1

G5c

8:2:5

PC_d10 or Chol_d7

0.53

G5c

8:2:2.5

PC_d10 or Chol_d7

0.54

G5c

8:2:1.25

PC_d10 or Chol_d7

0.65a

G10c

8:2:2.5

PC_d10 or Chol_d7

0.56

G5c, G16c

8:2:5

PC_d10, Chol_d6

0.62b

G5c, G16c

8:2:2.5

PC_d10, Chol_d6

0.53b

Calculation is done with CPC_d10 =0.517 which is obtained from sample G5c in PC_d10:PG

(4:1).
b

Calculation is done by comparison of sample between G5c with PC_d10 labeling and G16c

with Chol_d6 labeling and based on

13

CO contact with one lipid model, and GPC-Chol =

RTln(S/S0Chol-d6/S/S0PC-d10fPC/fChol), and S/S0 is the average dephasing of 8-40ms
dephasing times.

129

4.2.3 CHOL binding to two strands antiparallel HFPs predicted by Swiss Dock and two
sets of REDOR experimental distance constraints
Prediction of the most favorable sites of interaction between CHOL and the antiparallel HFP
motif

(HFP-antiβ)

is

performed

by

using

the

Swiss

Dock

web

service

(http://www.swissdock.ch/docking) with default settings.[34-36] The HFP–anti motif is built by
sidechain mutation to the 16–residue HFP sequence in PyMOL (The PyMOL Molecular
Graphics System, Version 1.2r2, SchrĂśdinger, LLC) from two adjacent long strands in the betabarrel outer membrane protein G (OmpG) (Protein Data Bank (PDB) entry 2iww).[37, 38] The
OmpG sequences of

82

DFSFGLTGGFRNYGY97H and

106

TANMQRWKIAPDWDV121K are

both mutated to N-terminal HFP 16–residue sequence AVGIGALFLGFLGAAG. The
antiparallel  strand backbone structures of OmpG sequences are maintained. Any van der Waals
overlaps between side chains are alleviated by using alternative favorable side–chain orientations
from Dunbrack’s backbone-dependent rotamer library, as implemented in PyMOL.[39, 40] Both
the energy–minimized (HFP-antimin) and non–energy–minimized (HFP-anti) states of HFP–anti
are used as input to Swiss Dock, with energy minimization performed with default parameters by
the YASARA energy minimization server (http://www.yasara.org/minimizationserver.htm).[41]
Developed by crystallographers working with computational chemists, this energy minimization
tool maintains good stereochemistry in the protein and ligand, as well as improving interactions
between the two molecules.
CHOL ligand structures for docking are obtained from ZINC database entry 3875383
(http://zinc.docking.org) and from a CHOL molecule bound to the crystal structure of human 2adrenergic receptor, a membrane bound G protein-coupled receptor, PDB 2rh1.[42-45] In both
cases, the stereochemistry of CHOL is confirmed to be correct. A panel of all 17 low–energy

130

(favorable) conformations of the 2rh1 CHOL molecule is obtained by using Omega2 software
version 2.3.2 (OpenEye Scientific Software, Inc., Santa Fe, NM; http://www.eyesopen.com).[46,
47] For clarity in the following data tables, the CHOL structure from the ZINC database is
referred as CHOL, and the panel of all favorable CHOL conformers from the PDB entry 2rh1
cholesterol structure is labeled Chol17.
For predicting favored interactions between a ligand such as CHOL and the HFP protein, Swiss
Dock samples alternative ligand binding modes relative to the protein surface (e.g., HFP-antiβ)
by dihedral sampling, filters out redundant or poorly interacting orientations of the ligand, and
then sorts the dockings by their Simple-Fitness values (the CHARMM22 energy). The most
favorable orientations are then minimized by using the CHARMM force field, and the exact
CHARMM22 energy is calculated. This includes the bonded and non-bonded (electrostatic and
van der Waals interaction) energy of protein and ligand, and the non-bonded interaction energy
between the protein and the ligand. Finally, Swiss Dock spatially clusters the energy–minimized
dockings by RMSD, using a distance cutoff of 2Å. Within a cluster, each binding mode is ranked
according to its effective energy, which also includes the fast analytical continuum treatment of
solvation (FACTS) energy, such that the rank #0 docking in each cluster represents the lowest
energy configuration in that cluster.[34, 36, 48]
For identification of dockings that are consistent with the REDOR–identified interaction between
CHOL and HFPs, the coordinates of the lowest energy CHOL docking from each spatial cluster
are saved in a separate PDB-formatted file, and the peptide backbone carbonyl carbon to CHOL
1

H distances are analyzed in PyMOL. Table 4.6 summarizes the results. Those meeting a 4–5 Å

distance between the CHOL isopropyl 1Hs and the Gly5 or Gly10 13CO, or between one or more
of the Chol-2,2,3,4,4,6 protons and Gly16 or Ala1

131

13

CO, are considered to match the

13

CO–2H

REDOR experimental data and are sent to YASARA for energy minimization. After energy
minimization, CHOL– HFP–antiβ atomic coordinates are saved as PDB files following deletion
of the shell of water molecules. The final energy values of the docking conformers that meet at
least 2 REDOR distance constraints within 4–5 Å are found to lie within a favorable range of 8535 kJ.mol-1 to -9454 kJ.mol-1 (more favorable), though these energy values should be
considered approximate rather than absolute.
A summary of the docking modes that resulted is given in table 4.7. More binding orientations
relative to HFP-antiβ resulted when experimental distance constraints up to 4–6 Å are considered,
but these additional binding modes are distributed all around HFP–antiβ without showing
preferential binding in any region(s) and are deemed nonspecific. Further analysis focused on
those docking modes that meet the more stringent criteria of matching at least two experimental
distance constraints within 4–5 Å.

132

Table 4.6 Dockings of CHOL with HFP–antiβ resulting from different protocols
Number of docking
Number of docking
Protein
model

Number of

modes matching at

Swiss Dock

least 2 experimental

CHOL

modes matching at least

model

1 experimental distance
docking modes

distance constrains
constrains within 4–6 Å
within 4–5 Å

HFP-anti

CHOL

12

1

7

HFP-anti

Chol17

14

2

4

HFP-antimin

CHOL

6

1

3

HFP-antimin

Chol17

14

4

7

Docked CHOL molecules that match at least two REDOR 4–5 Å distance constraints to HFP–
antiβ (HFP-antiβ and HFP-antiβmin) are analyzed together in PyMOL to define the most
favorable sites of interaction (figure 4.6). The minimized energies of each docking are listed in
table 4.7, as well as the side chains in HFP–antiβ participating in each interaction of the lowest
energy docking from each cluster. There are two distinct interaction footprints on HFP–antiβ,
one around Ala1–Gly5 showing the highest catchment, including seven dockings (position 1),
and a second around Ala6–Gly10 containing only one docking (position 2). These positions
reflect two different ways of two distance constraints being satisfied. In the first docking position,

133

either Chol–2,2,3,4,4,6 protons around the hydroxyl group contact the Ala1 carbonyl carbon and
the CHOL isopropyl tail contacts the Gly5 carbonyl carbon in the same strand of HFP, or the
CHOL isopropyl tail contacts the Gly5 carbonyl carbon in one strand and the Gly10 carbonyl
carbon in the other strand of HFP. In the second docking position, the CHOL isopropyl tail
contacts the Gly5 carbonyl carbon in one strand and the Gly10 carbonyl carbon in the other
strand of HFP. Here we label the residues in one strand of HFP with sentence case (e.g., Gly) and
the residues in the second strand with capitals (e.g., GLY) to distinguish them. ILE4, GLY5 and
LEU7 sidechains are in frequent contact with cholesterol across all eight dockings.
The most energetically favorable docking (-9453.8 kJ.mol-1) is selected as the representative
from the seven dockings overlapping in the first (Ala1–Gly5) position (bold purple in position 1,
figure 4.6). In this docking, cholesterol makes sidechain contacts with residues VAL2, ILE4, and
LEU7 in one chain and Gly10 and Leu12 in the other (figure 4.7-b). Since the sequence of the
two HFP strands is the same, there is also a symmetry–related position in which cholesterol
interacts with the same residues in the other strand. The original and symmetry–related positions
of this docking are also shown (figure 4.7-c), and can be occupied concurrently. Furthermore,
considering the possibility that multiple HFPs together form a larger antiparallel beta sheet
imbedded in the membrane, cholesterol binding in this position and its symmetry–related
position would not interfere with the growth of a beta sheet (figure 4.8-a). However, the second
position for cholesterol binding, in the Ala6–Gly10 region (Chol location shown in green, figure
4.8-b) blocks the main–chain hydrogen bonding edge of the strand and thus would inhibit beta
sheet expansion.

134

Table 4.7 Energies of CHOL dockings meeting at least two experimental distance constraints
within 5Å a
Protein

CHOL

Energy /

Satisfied REDOR

model

model

(kJ.mol-1)

distance constraints a

CHOL

-9247.0

Side chains within 4 Å of CHOL

GLY5–Chol_d7,
HFP-anti

Leu7,Leu9,Gly10,Phe11,Leu12
Gly10–Chol_d7
GLY5–Chol_d7,

Gly10,Leu12,VAL2,ILE4,GLY5

Gly10–Chol_d7

,LEU7

-9453.8
HFP-anti

Chol17
GLY5–Chol_d7,
-9209.5

NA
ALA1–Chol_d6

HFP-

GLY5–Chol_d7,
CHOL

-8750.2

VAL2,GLY3,ILE4,GLY5,LEU7
ALA1–Chol_d6

antimin

GLY5–Chol_d7,
-9014.3

Leu12,ILE4,GLY5,LEU7
Gly10–Chol_d7
GLY5–Chol_d7,

-8964.7
HFP-

NA
Gly10–Chol_d7

Chol17
antimin

GLY5–Chol_d7,
-8535.7

NA
ALA1–Chol_d6
GLY5–Chol_d7,

-8816.6

NA
ALA1–Chol_d6

a

Chol_d7 refers to cholesterol deuterated at isopropyl 1H, and Chol_d6 refers to Chol deuterated

at 2,2,3,4,4,6 – 1H. HFP-antiβ and HFP-antiβmin side chains within 4 Å contact distance of CHOL
are listed for the most energetically favorable member of each docking cluster from Swiss Dock.

135

Figure 4.6 Eight favorable dockings of CHOL (colored tubes) that meet two REDOR 4–5 Å
distance constraints: the YASARA energy minimized predicted HFP-antiβ and HFP-antiβmin
structures (drawn as lines, with residues labeled) in complex with the corresponding CHOL
binding mode are colored red for the least energetically favorable group (-8535 to -8841 kJ¡mol1

), green for the intermediate energy group (-8841 to -9147 kJ.mol-1) and purple, for the most

favorable group (-9147 to -9454 kJ.mol-1). Note the high occupancy of seven favorable CHOL
dockings spanning Ala1–Gly5 (position 1, lower left), with just one docking occupying position
2 (top center), as shown by thick purple tubes for the two most favorable dockings. CHOL
interactions with HFP residues are listed earlier in this chapter.

136

(a)

(b)

(c)

Figure 4.7 The most favorable CHOL binding mode relative to HFP-antiβ: (a) CHOL binding
mode (purple tubes) is shown relative to the two HFP strands (sticks). (b) Details of HFP-antiβ
side chain interactions with CHOL protons monitored by REDOR (shown in white and labelled).
The binding mode is the same as shown for location 1 above, while rotated by roughly 180°
about the horizontal and vertical axes to enable viewing from above. (c) Same CHOL binding
mode shown above (position 1, purple), plus its symmetry mate at the opposite end of HFP-antiβ
(position 1’, magenta).
137

(a)

(b)

Figure 4.8 The HFP–antiβ structural model in complex with (a) CHOL in its dominant favorable
position (position 1, purple tubes, as shown in Figure above) and (b) in the alternative binding
mode (position 2, shown here in green tubes). The two strands of HFP-antiβ (dark blue) with
cholesterol bound are shown in the context of a multi-stranded beta barrel structure (light blue
strands, from PDB entry 2iww) to explore the extent to which the two CHOL positions are
consistent with formation of a larger (more than two–stranded) antiparallel  sheet structure by
HFPs. The first cholesterol position (a) and its symmetry mate (not shown) would be compatible
with formation of a larger sheet structure by HFPs if the last four residues, GAAG, of every
second peptide in the sheet did not pair with the adjacent beta strand and shifted out of the way.
The second CHOL site (b) would clearly block the addition of a beta strand (shown by
interpenetration between CHOL and the light–blue beta strand) and thus would preclude
formation of a larger sheet by HFPs.

138

4.2.4 Favorable CHOL binding geometry on the concave surface of an HFP–antiβ sheet
It’s also possible that CHOL is located on top of HFP-antiβ with extended sheet formation. All
binding positions on top of HFP–antiβ without blocking sheet extension predicted from Swiss
Dock are screened out. Most of the binding has CHOL located near the center residues of the
HFP strands. The distances between CHOL and HFPs at the sites measured by REDOR are
analyzed. Only the green with energy of -9080.4 kJ.mol-1 (energy within intermediate energy
group (-8841 to -9147 kJ.mol-1)) and purple with energy of -9228.5 kJ.mol-1 (energy within
most favorable group (-9147 to -9454 kJ.mol-1) ones have within 5Å distance between CHOL
and HFP at one of the sites measured by REDOR (table 4.8, figure 4.9). It’s between CHOL
isopropyl 1Hs and G5/G10 for green/purple colored binding respectively. Among the populated
CHOL binding positions, the purple one is chosen as the representative of the most energy
favorable binding position for CHOL located on the concave surface of HFP-antiβ. In this
docking, CHOL makes sidechain contacts with Leu7 and Leu9 in one strand, and PHE8, GLY10
and LEU12 in the other (figure 4.10-b). This binding mode might interfere with neighboring
sheet extension as seen when superimposing this binding geometry on PDB 2iww (figure 4.11).
However, it only needs space to accommodate the terminal CH3 protons of CHOL. It’s still
possible to have this CHOL binding geometry if the neighbor strands accommodate the terminal
CH3 protons of cholesterol during bigger sheet formation.

139

Table 4.8 CHOL dockings meeting at least one experimental distance constraints within 5 Å on
the concave surface of an HFP-antiβ sheet
Protein

CHOL

Energy

Satisfied REDOR

Sidechains within

Docking

model

model

(kJ.mol-1)

distance constraints

4Å of CHOL

location

HFP-antimin

CHOL

-9080.4

Gly5–Chol_d7

NA

4

HFP-antimin

Chol17

-9228.5

GLY10–Chol_d7

Leu7, Leu9, PHE8,
3
GLY10, LEU12

Figure 4.9 Favorable dockings of CHOL (colored tubes) on the concave surface of HFP–antiβ:
Only the purple and green ones have within 5 Å distance between CHOL and HFP at one of the
sites measured by REDOR. There are eight favorable dockings predicted by Swiss Dock with
CHOL located on the concave surface of HFP-antiβ without blocking sheet extension.

140

(a)

(b)

Figure 4.10 The most favorable CHOL binding mode (purple tubes, location 3) for CHOL
located on the concave surface of HFP-antiβ, (a) the geometry of the most favorable CHOL
binding to the two stranded HFPs, (b) details of HFP-antiβ side chain interactions with CHOL
protons monitored by REDOR (shown in white and labelled).

141

Figure 4.11 Superimposition of this purple colored favorable binding (location 3) on PDB 2iww,
the purple colored CHOL is from the most favorable binding location 3. It’s still possible to have
this CHOL binding geometry if the neighbor strands accommodate the terminal CH3 protons of
CHOL during bigger sheet formation.
In summary,

13

C-2H REDOR SSNMR reveals preferential contact between HFP and CHOL vs

PC lipid. Energy favorable contact between two strands antiparallel HFP and CHOL is
successfully modeled by Swiss Dock, energy minimization, and filtered by REDOR
experimental results. There are two energetic favorable models of close contact between HFP
and CHOL (Figure 4.7a and Figure 4.10 a). The models are consistent with REDOR
experimental results of G5/Chol_d7 and/or G10/Chol_d7 results. The contact models reveal
tilted and curved-up tail orientation of Chol_d7. Fusion may be catalyzed by matching the
curvature of lipids contacting HFPs with the membrane curvature during the fusion intermediates
like the stalk.[49] The antiparallel HFPs make lipid tail closer to the membrane surface and could
reduce activation energy of joining of the outer leaflets of viral and host cell membrane.[50] The
HFP preferential contact to CHOL vs PC lipid may increase fusion through two features of
CHOL vs PC: one is that CHOL has greater intrinsic curvature, and the other one is that CHOL

142

is a shorter molecule and the tail is closer to the membrane surface. The study of HFP contact to
CHOL vs PC lipid supports that this

13

C-2H REDOR SSNMR method can also be applied to

study other proteins – CHOL contact with residue specific labeling.

143

REFERENCES

144

REFERENCES

1.

Chapman, D., PHASE-TRANSITIONS AND FLUIDITY CHARACTERISTICS OF LIPIDS
AND CELL-MEMBRANES. Quarterly Reviews of Biophysics, 1975. 8(2): p. 185-235.

2.

Lorizate, M., et al., Comparative lipidomics analysis of HIV-1 particles and their
producer cell membrane in different cell lines. Cellular Microbiology, 2013. 15(2): p.
292-304.

3.

Liao, Z.H., et al., Lipid rafts and HIV pathogenesis: Host membrane cholesterol is
required for infection by HIV type 1. Aids Research and Human Retroviruses, 2001.
17(11): p. 1009-1019.

4.

White, J.M., et al., Structures and mechanisms of viral membrane fusion proteins:
Multiple variations on a common theme. Critical Reviews in Biochemistry and Molecular
Biology, 2008. 43(3): p. 189-219.

5.

Sackett, K., et al., Hairpin Folding of HIV gp41 Abrogates Lipid Mixing Function at
Physiologic pH and Inhibits Lipid Mixing by Exposed gp41 Constructs. Biochemistry,
2009. 48(12): p. 2714-2722.

6.

Kilby, J.M., et al., Potent suppression of HIV-1 replication in humans by T-20, a peptide
inhibitor of gp41-mediated virus entry. Nature Medicine, 1998. 4(11): p. 1302-1307.

7.

Furuta, R.A., et al., Capture of an early fusion-active conformation of HIV-1 gp41.
Nature Structural Biology, 1998. 5(4): p. 276-279.

8.

Chan, D.C. and P.S. Kim, HIV entry and its inhibition. Cell, 1998. 93(5): p. 681-684.

9.

Cheng, S.F., et al., The fusion peptide domain is the primary membrane-inserted region
and enhances membrane interaction of the ectodomain of HIV-1 gp41. Molecular
Membrane Biology, 2010. 27(1): p. 31-44.

10.

Durell, S.R., et al., What studies of fusion peptides tell us about viral envelope
glycoprotein-mediated membrane fusion. Molecular Membrane Biology, 1997. 14(3): p.
97-112.

11.

Epand, R.M., Fusion peptides and the mechanism of viral fusion. Biochimica Et
Biophysica Acta-Biomembranes, 2003. 1614(1): p. 116-121.

12.

Lai, A.L., et al., Fusion Activity of HIV gp41 Fusion Domain Is Related to Its Secondary
Structure and Depth of Membrane Insertion in a Cholesterol-Dependent Fashion. Journal
of Molecular Biology, 2012. 418(1-2): p. 3-15.

145

13.

Yang, J., C.M. Gabrys, and D.P. Weliky, Solid-state nuclear magnetic resonance
evidence for an extended beta strand conformation of the membrane-bound HIV-1 fusion
peptide. Biochemistry, 2001. 40(27): p. 8126-8137.

14.

Gabrys, C.M., et al., Solid-State Nuclear Magnetic Resonance Measurements of HIV
Fusion Peptide (CO)-C-13 to Lipid P-31 Proximities Support Similar Partially Inserted
Membrane Locations of the alpha Helical and beta Sheet Peptide Structures. Journal of
Physical Chemistry A, 2013. 117(39): p. 9848-9859.

15.

Freed, E.O., et al., A MUTATION IN THE HUMAN-IMMUNODEFICIENCY-VIRUS
TYPE-1 TRANSMEMBRANE GLYCOPROTEIN-GP41 DOMINANTLY INTERFERES
WITH FUSION AND INFECTIVITY. Proceedings of the National Academy of Sciences
of the United States of America, 1992. 89(1): p. 70-74.

16.

Delahunty, M.D., et al., Mutational analysis of the fusion peptide of the human
immunodeficiency virus type 1: Identification of critical glycine residues. Virology, 1996.
218(1): p. 94-102.

17.

Yang, S.T., et al., HIV gp41-mediated membrane fusion occurs at edges of cholesterolrich lipid domains. Nature Chemical Biology, 2015. 11(6): p. 424-+.

18.

Yang, J., et al., Oligomeric beta-structure of the membrane-bound HIV-1 fusion peptide
formed from soluble monomers. Biophysical Journal, 2004. 87(3): p. 1951-1963.

19.

Qiang, W. and D.P. Weliky, HIV Fusion Peptide and Its Cross-Linked Oligomers:
Efficient Syntheses, Significance of the Trimer in Fusion Activity, Correlation of beta
Strand Conformation with Membrane Cholesterol, and Proximity to Lipid Headgroups.
Biochemistry, 2009. 48(2): p. 289-301.

20.

Gullion, T., Introduction to rotational-echo, double-resonance NMR. Concepts in
Magnetic Resonance, 1998. 10(5): p. 277-289.

21.

Gullion, T. and J. Schaefer, ROTATIONAL-ECHO DOUBLE-RESONANCE NMR.
Journal of Magnetic Resonance, 1989. 81(1): p. 196-200.

22.

Schmidt, A., R.A. McKay, and J. Schaefer, INTERNUCLEAR DISTANCE
MEASUREMENT BETWEEN DEUTERIUM (I = 1) AND A SPIN-1/2 NUCLEUS IN
ROTATING SOLIDS. Journal of Magnetic Resonance, 1992. 96(3): p. 644-650.

23.

Schmick, S.D. and D.P. Weliky, Major Antiparallel and Minor Parallel beta Sheet
Populations Detected in the Membrane-Associated Human Immunodeficiency Virus
Fusion Peptide. Biochemistry, 2010. 49(50): p. 10623-10635.

24.

Qiang, W., M.L. Bodner, and D.P. Weliky, Solid-state NMR Spectroscopy of human
immunodeficiency virus fusion peptides associated with host-cell-like membranes: 2D
correlation spectra and distance measurements support a fully extended conformation
and models for specific antiparallel strand registries. Journal of the American Chemical
Society, 2008. 130(16): p. 5459-5471.

146

25.

Zhang, H.Y., S. Neal, and D.S. Wishart, RefDB: A database of uniformly referenced
protein chemical shifts. Journal of Biomolecular Nmr, 2003. 25(3): p. 173-195.

26.

Schimick, S., High resolution tertiary structure of the membrane-associated HIV fusion
peptides by solid state nuclear magnetic resonance, in Chemistry. 2012, Michigan State
University.

27.

Jia, L.H., et al., REDOR solid-state NMR as a probe of the membrane locations of
membrane-associated peptides and proteins. Journal of Magnetic Resonance, 2015. 253:
p. 154-165.

28.

Ghosh, U., et al., Closed and Semiclosed Interhelical Structures in Membrane vs Closed
and Open Structures in Detergent for the Influenza Virus Hemagglutinin Fusion Peptide
and Correlation of Hydrophobic Surface Area with Fusion Catalysis. Journal of the
American Chemical Society, 2015. 137(24): p. 7548-7551.

29.

Qiang, W., Y. Sun, and D.P. Weliky, A strong correlation between fusogenicity and
membrane insertion depth of the HIV fusion peptide. Proceedings of the National
Academy of Sciences of the United States of America, 2009. 106(36): p. 15314-15319.

30.

Qiang, W., J. Yang, and D.P. Weliky, Solid-state nuclear magnetic resonance
measurements of HIV fusion peptide to lipid distances reveal the intimate contact of beta
strand peptide with membranes and the proximity of the Ala-14-Gly-16 region with lipid
headgroups. Biochemistry, 2007. 46(17): p. 4997-5008.

31.

Xie, L., et al., Multiple Locations of Peptides in the Hydrocarbon Core of Gel-Phase
Membranes Revealed by Peptide C-13 to Lipid H-2 Rotational-Echo Double-Resonance
Solid-State Nuclear Magnetic Resonance. Biochemistry, 2015. 54(3): p. 677-684.

32.

Bak, M., J.T. Rasmussen, and N.C. Nielsen, SIMPSON: A general simulation program
for solid-state NMR spectroscopy. Journal of Magnetic Resonance, 2000. 147(2): p. 296330.

33.

Sankaram, M.B. and T.E. Thompson, CHOLESTEROL-INDUCED FLUID-PHASE
IMMISCIBILITY IN MEMBRANES. Proceedings of the National Academy of Sciences of
the United States of America, 1991. 88(19): p. 8686-8690.

34.

Grosdidier, A., V. Zoete, and O. Michielin, Fast Docking Using the CHARMM Force
Field with EADock DSS. Journal of Computational Chemistry, 2011. 32(10): p. 21492159.

35.

Grosdidier, A., V. Zoete, and O. Michielin, SwissDock, a protein-small molecule docking
web service based on EADock DSS. Nucleic Acids Research, 2011. 39: p. W270-W277.

36.

Grosdidier, A., V. Zoete, and O. Michielin, EADock: Docking of small molecules into
protein active sites with a multiobjective evolutionary optimization. Proteins-Structure
Function and Bioinformatics, 2007. 67(4): p. 1010-1025.

147

37.

Yildiz, O., et al., Structure of the monomeric outer-membrane porin OmpG in the open
and closed conformation. Embo Journal, 2006. 25(15): p. 3702-3713.

38.

Berman, H.M., et al., The Protein Data Bank. Nucleic Acids Research, 2000. 28(1): p.
235-242.

39.

Dunbrack, R.L. and M. Karplus, BACKBONE-DEPENDENT ROTAMER LIBRARY FOR
PROTEINS - APPLICATION TO SIDE-CHAIN PREDICTION. Journal of Molecular
Biology, 1993. 230(2): p. 543-574.

40.

Bower, M.J., F.E. Cohen, and R.L. Dunbrack, Prediction of protein side-chain rotamers
from a backbone-dependent rotamer library: A new homology modeling tool. Journal of
Molecular Biology, 1997. 267(5): p. 1268-1282.

41.

Krieger, E., et al., Improving physical realism, stereochemistry, and side-chain accuracy
in homology modeling: Four approaches that performed well in CASP8. ProteinsStructure Function and Bioinformatics, 2009. 77: p. 114-122.

42.

Irwin, J.J., et al., ZINC: A Free Tool to Discover Chemistry for Biology. Journal of
Chemical Information and Modeling, 2012. 52(7): p. 1757-1768.

43.

Irwin, J.J. and B.K. Shoichet, ZINC - A free database of commercially available
compounds for virtual screening. Journal of Chemical Information and Modeling, 2005.
45(1): p. 177-182.

44.

Hanson, M.A., et al., A specific cholesterol binding site is established by the 2.8
angstrom structure of the human beta(2)-adrenergic receptor. Structure, 2008. 16(6): p.
897-905.

45.

Cherezov, V., et al., High-resolution crystal structure of an engineered human beta(2)adrenergic G protein-coupled receptor. Science, 2007. 318(5854): p. 1258-1265.

46.

Hawkins, P.C.D. and A. Nicholls, Conformer Generation with OMEGA: Learning from
the Data Set and the Analysis of Failures. Journal of Chemical Information and Modeling,
2012. 52(11): p. 2919-2936.

47.

Hawkins, P.C.D., et al., Conformer Generation with OMEGA: Algorithm and Validation
Using High Quality Structures from the Protein Databank and Cambridge Structural
Database. Journal of Chemical Information and Modeling, 2010. 50(4): p. 572-584.

48.

Haberthur, U. and A. Caflisch, FACTS: Fast analytical continuum treatment of solvation.
Journal of Computational Chemistry, 2008. 29(5): p. 701-715.

49.

Tristram-Nagle, S., et al., HIV Fusion Peptide Penetrates, Disorders, and Softens T-Cell
Membrane Mimics. Journal of Molecular Biology, 2010. 402(1): p. 139-153.

148

50.

Larsson, P. and P.M. Kasson, Lipid Tail Protrusion in Simulations Predicts Fusogenic
Activity of Influenza Fusion Peptide Mutants and Conformational Models. Plos
Computational Biology, 2013. 9(3): p. 9.

149

Chapter 5 - HFP Effect on Membrane Motion by 2H-NMR Studies
5.1 Introduction
2

H - NMR is a useful tool to study the structure and motions of lipids as well as membrane

protein-lipids interactions in biological membranes.[1, 2] 2H and 1H are chemically equivalent
and

2

H substitution of

1

H is not perturbing the native lipids and protein membrane

environment.[3-5] Thus, it is advantageous to study membrane structure and motion changes by
2

H - NMR. We analyze the membrane motion through quadrupolar splitting (ΔνQ) and deuteron

relaxation (T2 and T1) studies.
Deuteron is a spin 1 nucleus, and there are two allowed transitions: ms= 1→0 and ms= 0→ -1,
which corresponds to the Pake doublet resonance in the spectrum. ΔνQ is the peak splitting
between the Pake doublets corresponding to a certain C-D bond orientation relative to B0. There
is discussion of the orientation dependence of the quadrupolar energy in chapter 1 quadrupolar
coupling interaction section. The 2H spectrum can provide information about the C-D bond
vector fluctuations by deuterium order parameter (SCD).
3

∆𝜈𝑄 = 4 × (

𝑒 2 𝑞𝑄
ℎ

3

) × 𝑆𝐶𝐷 = 4 𝜒𝑄 𝑆𝐶𝐷

5.1

χQ is the static quadrupolar coupling constant. For aliphatic C-D bond, χQ is ~ 170 kHz.[6, 7] The
1

order parameter is SCD. 𝑆𝐶𝐷 = 2 < 3 cos2 𝜃 − 1 >, and θ is the angle between the C-D bond and
the lipid principal reorientation axis which is the membrane normal. It is hard to get the ΔνQ for
each individual deuteron along the per-deuterated lipid acyl chain because individual peaks are
broad and not well resolved in the static powder pattern spectrum (Figure 5.2) for un-oriented
lipid vesicles. De-Pake-ing is one method to convert the un-oriented spectrum into 0°- oriented
spectrum with well-resolved peaks from the FID.[8] To calculate SCD for 2Hs attached at

150

different carbon positons along the acyl chain, we use the peak splitting of the Pake doublet from
the de-Paked spectrum. The more motion, the less ordered of C-D bond. Similarly, the less
motion, the more ordered of C-D bond. Bigger ΔνQ gives bigger SCD. There is more motion
toward the acyl chain terminus, and SCD for lipid decreases along the carbon positon toward the
membrane center. Perturbation of lipid motion can be analyzed by comparison of S CD in
membrane without and with HFP or/and CHOL.
Besides quadrupolar splitting and segmental order parameter, deuteron relaxation times are also
sensitive to molecular motions. Relaxation is the process to return to the thermal equilibrium. In
NMR, after pulses are applied, the nuclei relax to the Boltzmann equilibrium with rates of 1/T2
and 1/T1.
𝑡

𝑡

2

2

𝑀𝑥 (𝑡) = 𝑀𝑥 (0)𝑒𝑥𝑝 (− 𝑇 ) ; 𝑀𝑦 (𝑡) = 𝑀𝑦 (0)𝑒𝑥𝑝 (− 𝑇 )
𝑡

𝑀𝑧 (𝑡) − 𝑀𝑧 (0) = (𝑀𝑧 (0) − 𝑀0 )𝑒𝑥𝑝 (− 𝑇 )

5.2
5.3

1

T2 is the transverse or spin-spin relaxation time and T1 is the longitudinal or spin lattice
relaxation time. T2 is sensitive to slow motions with τC >> ω0-1, and T1 is sensitive to fast
motions with τC << ω0-1, where τC is the molecular correlation time.[9, 10]
In this study, we are using 2H NMR method to investigate the effect of HFP on membrane
structure and motion in different model membranes. Membrane without and with CHOL is used
to study the CHOL catalysis on HFP induced membrane fusion from membrane motion view.
We use DMPC-d54 to probe the PC motion including segmental order parameter SCD and T2
relaxation rates in membrane both without and with CHOL. It is interesting to study the
membrane with CHOL because studies have shown that presence of CHOL could catalyze HFP

151

induced fusion and previous studied in chapter 4 in this thesis support that HFP has preferential
contact with CHOL vs PC.[11, 12] In membrane with CHOL, we investigate CHOL motion by
T2 and T1 relaxation studies. In DMPC/DMPG/CHOL membrane, we investigated both DMPC
and CHOL motion. Besides, we also studied CHOL motion in POPC/POPG/CHOL membrane,
which is the most commonly use membrane composition for viral protein induced vesicle fusion
assay.[13]

Figure 5.1 Chemical structures of deuterated PC and CHOL used for 2H NMR study: (a) DMPCd54, (b) Chol_d6 and (c) Chol_d7.
5.2 Experimental conditions
Samples are prepared by oganic cosolubilization method, see chapter 2 for details. The
percentage of CHOL is mole percentage, and HFP percentage is mole of HFP relative to the total
moles of PC and PG. DMPC-d54 and Chol_d6 2H NMR spectrum at variable temperatures are
acquired by solid echo (quecho) experiment. Chol_d6 and Chol_d7 are used in another

152

membrane copositon. The chemical structure of the deuterated PC and CHOL are displayed. The
pulse sequence of quecho experiment is shown in Chaper 2. For spectra at different temperatures,
the sample is equlibrated for ~30 minutes with rotor inside the probe at each target temperature.
The data is acquired typically with  = 40 Οs and 1 = 21 Οs to gain best signal intensity and
minimize pulse ringdown interference. For DMPC-d54 and Chol_d6/d7 2H T2 studies, the data
are acquired for different  and 1 with same constant delay increments. For Chol_d7 T1
relaxation studies, the data are acquired for each 1 with fixed increments for one set of 1 array.
90° and 180° pulses are optimized using D2O with the transmiter frequency set at 61.2023333
MHz. The data is acquired at Varian 9.4 tesla NMR spectromete. Typical parameters are 3.0 Îźs
90° pulse, 6.3 Οs 180° pulse, 0.9 rf amplification and 1s pulse delay. For DMPC-d54 studies, 2.1
Οs 90° pulse is used with 0.4 rf amplification. The rabi frequency for 90° 2H pulse is calculated
from equation:

𝛾𝐵1
2𝜋

1

= 4 × 𝑝𝑤90𝑋 , and is 83 kHz and 119 kHz for 3 μs and 2.1 μs 90° pulse,

respectively.
5.3 Results
5.3.1 Solid echo or quecho experimental results for DMPC/DMPG membrane without
and with CHOL/ HFP
5.3.1.1 2H - NMR spectra features for DMPC-d54 and Chol_d6
In order to help understand membrane motions perturbed by HFP in membrane without and with
CHOL, 2H spectra of DMPC-d54 are taken in membrane without and with CHOL or/and HFP at
different temperatures (Figures 5.2- 5.5). We add ~ 20 % DMPG to reflect the negative charged
membrane composition in HIV host cells, and assist HFP membrane binding through
electrostatic interaction because the peptide is positively charged. Pure DMPC-d54 has a phase
transition temperature of ~ 23 °C.
153

In membrane without CHOL and HFP, Figure 5.2 shows a clear phase transition of DMPC-54
between 21 °C and 25 °C as evidenced from the spectrum shape change. The spectrum at 21 °C
is broad because the experimental temperature is below the phase transition temperature of the
PC and the PC is in gel phase with less C- 2H bond motion. At 21 °C with addition of 4% HFP,
the spectrum becomes significantly narrower and the peaks are similarly resolved compared to
the spectra taken at higher temperatures. This indicates that HFP increases PC motion below the
transition temperature. For the spectrum taken from 25 °C to 37 °C, the spectra are narrower and
well resolved because the temperature is above the phase transition temperature and the
membrane is in liquid disordered phase. In presence of HFP, the overall spectra become a little
bit narrower, and the individual peaks become broader and less resolved.
In membrane with ~ 33% CHOL and without HFP, figure 5.3 does not show a phase transition of
the membrane because the membrane is in liquid ordered phase with the presence of 33% CHOL
at the temperature range studied. With addition of 4% HFP, all spectra become broader at
different temperatures that support that HFP increase membrane order. In addition of 1% HFP,
DMPC-d54 2H spectrum becomes narrower with individual peaks broader. For Chol_d6, the
spectrum is similar without and with 2% HFP.

154

Figure 5.2 2H - NMR spectra of DMPC-d54 taken at different temperature in membrane of
DMPC: DMPG= 40: 10Îźmol at pH 7.4.

155

Figure 5.3 2H - NMR spectra of DMPC-d54 taken at different temperature in membrane of
DMPC: DMPG= 40: 10Îźmol with ~ 2Îźmol HFP at pH 7.4.

156

Figure 5.4 2H - NMR spectra of DMPC-d54 taken at different temperature in membrane of
DMPC: DMPG: CHOL = 40: 10: 25Îźmol at pH 7.4.

157

Figure 5.5 2H - NMR spectra of DMPC-d54 taken at different temperature in membrane of
DMPC: DMPG: CHOL = 40: 10: 25Îźmol with ~ 2Îźmol HFP at pH 7.4.

158

5.3.1.2 DMPC-d54 Segmental order parameters at 37 °C
We investigated the effect of HFP on segmental order parameter of DMPC-d54 in membrane
without and with 33% CHOL at physiological temperature 37 °C and physiological pH 7.4.
Compared to membrane without HFP, the overall 2H spectrum of DMPC-d54 becomes narrower
with addition of 4% HFP in membrane without CHOL. There is similar spectrum feature with
addition of 1% HFP in membrane with CHOL. The narrowing of the spectrum indicates
disordering of the lipid acyl chain. On the contrary, the spectrum becomes significantly broader
with addition of 4% HFP in membrane with CHOL, which suggests ordering of the lipid acyl
chain in this condition.
However, the DMPC-d54 2H spectrum results from all the 2Hs in the molecule. In order to get
the quadrupolar splitting for each 2H along the acyl chain, the spectrum need to be de-convoluted
(de-Paked) to get better-resolved individual peaks. Figures 5.6-5.8 show representative overall
2

H powder pattern spectra of DMPC-d54, and figures 5.10- 5.12 show de-Paked spectra of

DMPC-d54 without and with HFP. From the de-Paked spectrum, order parameter for each 2H
3

along the acyl chain is calculated as 𝑆𝐶𝐷 = ∆𝑣𝑄 /(4 𝜒𝑄 ) , where 𝜒𝑄 =

𝑒 2 𝑞𝑄
ℎ

, and is the

Qradrupolar coupling constant in Hz. For aliphatic C-2H bond, χQ is 170 kHz. Figure 5.14 shows
the percentage change of 2H order parameter relative to the pure membrane. In presence of 4%
HFP, the 2H order parameter decreases ~ 1-10% along the acyl chain in membrane without
CHOL, and increases about 20% to 30% along the acyl chain in membrane with 33% CHOL. In
presence of 1% HFP, the 2H order parameter decreases ~ 2 - 7% along the acyl chain in
membrane with 33% CHOL. There is greater effect toward the terminal of the acyl chain. Order
parameter change toward the terminal 2H indicates perturbation of HFP down to the membrane
center, which is consistent with REDOR results.

159

(a)

(b)

Figure 5.6 2H - NMR spectra of DMPC-d54 without and with HFP at (a) 21C and (b) 37C in
membrane without CHOL at pH 7.4. Pure lipids are DMPC-d54: DMPG (40: 10Îźmol). HFP:
lipids ratio is 1:25.

160

(a)

(b)

Figure 5.7 2H - NMR spectra of DMPC-d54 without and with HFP at (a) 21C and (b) 37C in
membrane containing 33% CHOL at pH 7.4. Membrane is DMPC-d54: DMPG: CHOL (40: 10:
25Îźmol). HFP: lipids ratio is 1:25.

161

Figure 5.8 2H - NMR spectra of DMPC-d54 without (top) and with different peptide to lipids
mole ratios, 1: 100 (middle) and 1: 25 (bottom) at 37C and pH 7.4. Pure lipids are DMPC-d54:
DMPG: CHOL with 40:10:25Îźmol.

162

Figure 5.9 Chol_d6 2H-NMR spectra, without (top) and with HFP (bottom) at 37C and pH 7.4,
pure lipids are DMPC: DMPG: Chol_d7 with 40:10:25Îźmol. HFP: lipids mole ratio is 1:50, and
lipids do not include CHOL.

163

Figure 5.10 DMPC-d54 de-Paked spectra, without (top) and with HFP (bottom) at 37C and pH
7.4. The membrane is DMPC-d54: DMPG with 40:10Îźmol. HFP: lipids ratio is 1:25.

164

Figure 5.11 DMPC-d54 de-Paked spectra, without (top) and with HFP (bottom) at 37C and pH
7.4. The membrane is DMPC-d54: DMPG: CHOL with 40:10:25Îźmol. HFP: lipids ratio is 1:100.

165

Figure 5.12 DMPC-d54 de-Paked spectra, without (top) and with HFP (bottom) at 37C and pH
7.4. The membrane is DMPC-d54: DMPG: CHOL with 40:10:25Îźmol. HFP: lipids ratio is 1:25.

166

Figure 5.13 HFP effects on the DMPC-d54 order parameters profile in membrane with and
without CHOL at 37C and pH 7.4. HFP decreases the order parameters along the acyl chain of
the lipid in membrane of DMPC-d54 (d54) and DMPG (pg) with 1:25 peptide to lipids ratio, and
membrane with additional 33% CHOL (+Chol) with 1:100 peptide to lipids ratio. However, HFP
increases the order parameters along the acyl chain of the lipid with 1:25 peptide to lipids ratio in
the membrane containing 33% CHOL.

167

Figure 5.14 HFP effects on the order parameter change of DMPC-d54 along the lipid acyl chain
compared to pure membrane with and without CHOL at 37 C and pH 7.4. ΔSCD= SCD Lipid – SCD
(Lipid+ HFP)

. Positive value indicates decreased order parameter and negative value indicates

increased order parameter compared to pure membrane without HFP. The plot color and shape
are consistent with previous figure.
5.3.1.3 Transverse relaxation studies of DMPC-d54 and Chol_d6
Samples are prepared by oganic cosolubilization method, see chapter 2 for details. To evaluate
the effect of HFP on membrane motion and the role of CHOL, we study the transverse relaxation
(T2 relaxation) by quecho experiments. For each sample,  and 1 were arrayed with same delay
increments. We hydrated the sample with pH 7.4 buffer and study PC motion with DMPC-d54 in
membrane both without and with CHOL. We used Chol_d6 for CHOL motion in membrane
containing 33% CHOL. We study DMPC-d54 at different temperatures and Chol_d6 at 37 °C.

168

Figure 5.15 and Figure 5.21 shows representative stack plots of FID of DMPC-d54 and Chol_d6,
respectively. Figure 5.16 to 5.20 show representative stack plots of processed spectra of DMPCd54 for CD3 and outer CD2 peak intensity. Both FID intensity and processed peak intensity show
exponential decay. We fit the echo intensity of FID to get the overall T2 that has contributions
from all the 2Hs for both DMPC-d54 and Chol-d6. CD3 and CD2 peak intensity from processed
spectra are used to get CD3 and CD2 2H T2. Below phase transition temperature, CD2 peak
intensity is weak and decays very fast, so we did not analyze the T 2 at 21 °C. CD3 peak intensity
has contribution from CD3 and CD2 deuterons, the T2 analyzed for CD3 would also contains part
effect from CD2. Since CD2 T2 is much shorter than CD3, inclusion of CD2 peak intensity would
make the measured CD3 T2 smaller than the actual value.
We fit the data by equation:
2τ

I(2τ) = I(0) × exp (− T ) + A

5.1

2

I(2) is the experimental measured echo or peak intensity, and 2 =  + 1 + data shift points ×
dwell time. I(0) and T2 are fitting parameters, and A is the fitting offset. Figures 5.22 to 5.25
display T2 fittings from echo intensity for DMPC-d54 in different sample compositions at
various temperatures. Figure 5.26 shows representative T2 fitting from echo intensity, CD3 peak
intensity and CD2 peak intensity. Figure 5.27 displays representative fitting plots for membrane
without and with HFP for Chol_d6 at 37 °C. The best-fit T2s from different samples at various
temperatures are in Table 5.1. Adding HFP will generally decrease PC T2, except for sample of
HFP in membrane composed of DMPC-d54 and DMPG at 21 °C. We compared the T2s in
membrane with 33% CHOL, shown in Table 5.2. Adding HFP causes T2 reduction for both
Chol_d6 and DMPC-D54. There is much bigger reduction for PC than Chol_d6. The T2

169

reduction is a little smaller with 1% HFP than 4% HFP when fitting with echo intensity, but
similar for CD3 and CD2.

Figure 5.15 Representative stacked FID plots DMPC-d54 in membrane of DMPC-d54: DMPG
without HFP at 21C (top) and 37C (bottom) in static. The 2H FIDs were obtained by varying 
and 1. For each  and 1, the number of scans was 5000 (top) and 1000 (bottom), respectively.

170

Figure 5.16 Representative stacked spectrum plots DMPC-d54 in membrane of DMPC-d54:
DMPG without HFP at 21C (top) and 37C (bottom) at pH 7.4 in static. The 2H spectra were
obtained by varying  and 1. For each  and 1, the number of scans was 5000 (top) and 1000
(bottom), respectively. All spectra were processed with 200 Hz line broadening, data shift = -11.

171

Figure 5.17 Representative stacked spectrum plots DMPC-d54 in membrane of DMPC-d54:
DMPG with 2µmol HFP at 21C (top) and 37C (bottom) and pH 7.4 in static. The 2H spectra
were obtained by varying  and 1. The top spectra were arrayed to  = 750 and 1 = 731 ¾s. The
bottom spectra were arrayed to  = 1360 ¾s and  = 1341 ¾s. For each  and 1, the number of
scans was 5000 (top) and 1000 (bottom), respectively. All spectra were processed with DC offset
correction, data shift = -11. We processed the bottom spectra additionally with polynomial
baseline correction of the order 5.

172

Figure 5.18 Representative stacked spectrum plots DMPC-d54 in membrane of DMPC-d54:
DMPG: CHOL (8:2:5) without HFP at 21C (top) and 37C (bottom) and pH 7.4 in static. The
2

H spectra were obtained by varying  and 1. The top spectra were arrayed to  = 1340 and 1 =

1321 ¾s. The bottom spectra were arrayed to  = 1320 ¾s and  = 1301 ¾s. For each  and 1, the
number of scans was 1000. All spectra were processed with DC offset correction, data shift = -11.

173

Figure 5.19 Representative stacked spectrum plots DMPC-d54 in membrane of DMPC-d54:
DMPG: CHOL (8:2:5) with 2µmol HFP at 21C (top) and 37C (bottom) and pH 7.4 in static.
The 2H spectra were obtained by varying  and 1. The top spectra were arrayed to  = 1340 and
1 = 1321 ¾s. The bottom spectra were arrayed to  = 1000 ¾s and  = 981¾s. For each  and 1,
the number of scans was 1000. All spectra were processed with DC offset correction, data shift =
-11.

174

Figure 5.20 Representative stacked spectrum plots DMPC-d54 in membrane of DMPC-d54:
DMPG: CHOL (8:2:5) without (top) and with HFP (HFP: phospholipids= 1:100) (bottom) at
37C (bottom) and pH 7.4 in static. The 2H spectra were obtained by varying  and 1. The top
spectra were arrayed to  = 1320 and 1 = 1301 ¾s. The bottom spectra were arrayed to  = 1160
¾s and  = 1141¾s. For each  and 1, the number of scans was 1000. All spectra were processed
with DC offset correction, data shift = -11. We processed the bottom spectra with additional 200
Hz Gaussian line broadening and polynomial baseline correction of the order 5.

175

Figure 5.21 Stacked spectrum plots Chol_d6 in membrane of DMPC: DMPG: Chol_d6 (8:2:5)
without (top) and with (bottom) 1µmol HFP at 37C and pH 7.4 in static. The 2H spectra were
obtained by varying  and 1. The spectra were arrayed to  = 340 and 1 = 321 ¾s. For each 
and 1, the number of scans was 10000.

176

Figure 5.22 Quecho experimental (red squares) and best fit (red lines) plots of tip intensity of
echo FID vs 2 for DMPC-d54 in membrane of DMPC-d54: DMPG without HFP under static
conditions at different temperatures and pH 7.4. The data are fitted with 𝐼(2𝜏) = 𝐼(0) × exp( −
2𝜏/𝑇2 ) + 𝐴, where A is the fitting offset.

177

Figure 5.23 Quecho experimental (red squares) and best fit (red lines) plots of tip intensity of
DMPC-d54 echo FID vs 2 in membrane of DMPC-d54: DMPG with HFP under static
conditions at different temperatures and pH 7.4. The HFP: lipids mole ratio is 1:25. The data are
fitted by equation 𝐼(2𝜏) = 𝐼(0) × exp( − 2𝜏/𝑇2 ) + 𝐴, where A is the fitting offset. There are
some fitting deviations relative to experimental data because there are fast decay components
(CD2) and slow decay components (CD3).

178

Figure 5.24 Quecho experimental (red squares) and best fit (red lines) plots of tip intensity of
echo FID vs 2 for DMPC-d54 in membrane of DMPC-d54: DMPG: CHOL (8:2:5 mole ratio)
without HFP under static conditions at different temperatures and pH 7.4. The data are fitted by
equation 𝐼(2𝜏) = 𝐼(0) × exp( − 2𝜏/𝑇2 ) + 𝐴, where A is the fitting offset. There are some fitting
deviations relative to experimental data because there are fast decay components (CD2) and slow
decay components (CD3).

179

Figure 5.25 Quecho experimental (red squares) and best fit (red lines) plots of tip intensity of
DMPC-d54 echo FID vs 2 in membrane of DMPC-d54: DMPG: CHOL (8:2:5 mole ratio) with
HFP under static conditions at different temperatures and pH 7.4. The HFP: phospholipids mole
ratio is 1:25. The data are fitted by equation 𝐼(2𝜏) = 𝐼(0) × exp( − 2𝜏/𝑇2 ) + 𝐴, where A is the
fitting offset.

180

Figure 5.26 Quecho experimental (red squares) and best fit (red lines) plots of DMPC-d54 echo
tip intensity of FID, CD3 and CD2 peak intensity vs 2 in membrane of DMPC-d54: DMPG:
CHOL (8:2:5 mole ratio) with HFP under static conditions at 37 °C and pH 7.4. The HFP:
phospholipids mole ratio is 1:100. The data are fitted by equation 𝐼(2𝜏) = 𝐼(0) × exp( −
2𝜏/𝑇2 ) + 𝐴, where A is the fitting offset.

181

Figure 5.27 Quecho experimental (red squares) and best fit (red lines) plots of Chol_d6 FID
echo tip intensity vs 2 in membrane of DMPC: DMPG: Chol_d6 (8:2:5 mole ratio) with and
without HFP under static conditions at 37 °C and pH 7.4. The HFP: phospholipids mole ratio is
1:50. The data are fitted by equation 𝐼(2𝜏) = 𝐼(0) × exp( − 2𝜏/𝑇2 ) + 𝐴, where A is the fitting
offset.

182

Table 5.1 The HFP effects on DMPC-d54 best-fit 2H T2 (s) in membrane without and with
CHOL by quecho experiments. The fitting errors are in parenthesis.

Temperature

d54pg

25

30

37

d54pgChol

(1:25)

(C)

21

HFP/d54pg

HFP/d54pgChol
(1:25)

120 (6)

235 (11)

726 (14)

289 (11)

echo

158 (7)

541 (50)

1255 (53)

710 (24)

CD3

NA

247 (5)

519 (16)

265 (6)

CD2

1170 (61)

301 (16)

1153 (56)

273 (9)

echo

1333 (108)

735 (46)

2060 (147)

663 (36)

CD3

1000 (25)

304 (4)

674 (58)

259 (2)

CD2

1206 (61)

327 (18)

1292 (112)

298 (7)

echo

1465 (119)

685 (42)

1935 (242)

726 (31)

CD3

991 (29)

322 (7)

807 (57)

241 (5)

CD2

1140 (36)

302 (26)

1818 (265)

264 (13)

echo

1206 (114)

715 (63)

2767 (454)

678 (53)

CD3

901 (28)

329 (16)

1140 (144)

222 (10)

CD2

183

Table 5.2 HFP effects on best-fit 2H T2 (s) of Chol_d6 and DMPC-d54 in membrane with 33%
CHOL studied by quecho experiment at 37 °C and pH 7.4. The fitting errors are in parenthesis.
The ratio is HFP: phospholipids mole ratio. NA means not applied.

T2/ Îźs

dmpcpgChol_d6

HFP/

d54pgChol

HFP/

HFP/

dmpcpgChol_d6

d54pgChol

d54pgChol

(1:50)

(1:100)

(1:25)

echo

358 (13)

230 (7)

1818 (265)

345 (13)

264 (13)

CD3

NA

NA

2767 (454)

671 (48)

678 (53)

CD2

NA

NA

1140 (144)

242 (9)

222 (10)

5.3.2 CHOL motion in POPC/POPG/CHOL membrane without and with HFP
It is interesting to study CHOL motion in POPC/POPG/CHOL membrane system with 33%
CHOL because this is a common membrane composition used to study the fusion activity of
fusion peptides and proteins of virus such as HIV and Influenza.[14-17] Studies have shown that
presence of CHOL can catalyze HFP induced vesicle fusion.[11, 12] We study HFP perturbation
on CHOL motion by T2 and T1 relaxation. We analyzed T2 relaxation for both Chol_d6 and
Chol_d7, T1 relaxation for Chol_d7. In membrane with HFP, the HFP: phospholipids mole ratio
is ~ 1:50.
5.3.2.1 Transverse relaxation studies of Chol_d6

184

We studies Chol_d6 T2 relaxation at 37 °C and pH 7.4. Figure 5.28 shows the stacked plots of
FID without and with HFP. We got T2 by fitting the echo intensity with equation 𝐼(2𝜏) = 𝐼(0) ×
exp( − 2𝜏/𝑇2 ) + 𝐴, shown in Figure 5.29 and Table 5.3.
Adding HFP reduces T2 of Chol_d6 fitting from echo intensity. The reduction extent is similar
compared to DMPC/DMPG/CHOL membrane system. We investigated the spectrum features at
-5°C, 0 °C, 25 °C, 37 °C and 45 °C, shown in figure 5.30. The spectrum becomes broader when
the temperature decreases, and this is consistent with less motional averaging of quadrupolar
couplings at lower temperature. However, the spectrum is not significantly different at and
higher than 0 °C, because the phase transition temperature of POPC and POPG is -2 °C.
However, the spectrum at -5 °C has much worse signal/noise ratio compared to spectrum taken
at 37 °C with similar number of scans. This is due to less motion at the temperature below the
phase transition temperature of the membrane.

185

Figure 5.28 Stacked spectrum plots Chol_d6 in membrane of POPC: POPG: Chol_d6 (8:2:5)
without (top) and with (bottom) HFP at 37C and pH 7.4 in static. The 2H spectra were obtained
by varying  and 1. The top spectra were arrayed to  = 330 and 1 = 311 ¾s, and the bottom
spectra were arrayed to  = 300 and 1 = 281 ¾s. For each  and 1, the number of scans was
4000. (File location: mb4b…./102416, Chold6popcpg_withHFP_102716 on mb4b )

186

Figure 5.29 Quecho experimental (red squares) and best fit (red lines) plots of Chol_d6 FID
echo tip intensity vs 2 in membrane of POPC: POPG: Chol_d6 (8:2:5 mole ratio) with and
without HFP under static conditions at 37 °C and pH 7.4. The HFP: phospholipids mole ratio is
1:50. The data are fitted by equation 𝐼(2𝜏) = 𝐼(0) × exp( − 2𝜏/𝑇2 ) + 𝐴, where A is the fitting
offset. (File location: mb4b…./102416, Chold6popcpg_withHFP_102716 on mb4b )

187

Table 5.3 HFP effects on best-fit Chol_d6 2H T2 (s) of in POPC: POPG membrane with 33%
Chol_d6 studied by quecho experiment at 37 °C and pH 7.4. The fitting errors are in parenthesis.
We got the best fit T2 by fitting the echo intensity. The ratio is HFP: phospholipids mole ratio.

T2/Îźs

No HFP

with HFP: phospholipids (1:50)

echo

312 (28)

224 (6)

188

Figure 5.30 Stacked spectrum plots Chol_d6 in membrane of POPC: POPG: Chol_d6 (8:2:5)
with HFP at different temperatures and pH 7.4 in static. The 2H spectra were obtained by varying
 and 1. We processed the spectra with data shift, DC offset, 500 Hz Gaussian line broadening,
and baseline correction of the order 5. The number of scans was typically 10000. It is 35000 for
25 °C and 80000 for 0 °C.

189

5.3.2.2 Transverse relaxation studies of Chol_d7
To evaluate the effect of HFP on the CHOL hydrocarbon tail motion, we investigated the
transverse relaxation (T2) at different temperatures: - 50°C, 5°C, 25°C, 37°C and 45°C. We fit
2𝜏

the peak intensity of CD3 (I2) to get best fit T2. The fitting equation is ln(𝐼2𝜏 ) = ln 𝐼0 − 𝑇 ,
2

where lnI0 and T2 are the fitting parameters and 2 =  + 1+ data shift points × dwell time.
Figure 5.31 and 5.32 display stack plots of processed spectra without and with HFP at - 50°C and
37 °C, respectively. Figures 5.33 to 5.35 show the fitting plots at different temperatures without
and with HFP. Table 5.4 displays the best-fit T2 values. The T2 values at - 50°C are much smaller
in membrane both without and with HFP compared to the other higher temperatures studied.
This is due to less motion at the temperature below the PC phase transition temperature (-2 °C).
At - 50°C, presence of HFP increases T2, which suggests that presence of HFP increases the
CHOL tail motion. However, at and above 5 °C, there is similar T2 at different temperatures for
the same sample, and similar T2 without and with HFP, which suggests that HFP does not affect
the CHOL tail motion at the temperature above the POPC phase transition temperature.

190

180 Îźs
161Îźs

180 Îźs
161Îźs

Figure 5.31 Stacked spectrum plots Chol_d7 in membrane of POPC: POPG: Chol_d7 (8:2:5)
without and with HFP at -50C and pH 7.4 in static. HFP to phospholipids ratio is 1:50. The 2H
spectra were obtained by varying  and 1. For each  and 1, the number of scans was 2000 (top)
and 800 (bottom). We processed the data with -10 data shift pts, 2000 Hz Gaussian line
broadening and polynomial baseline correction of the order 5.

191

Figure 5.32 Stacked spectrum plots Chol_d7 in membrane of POPC: POPG: Chol_d7 (8:2:5)
without and with HFP at 37C and pH 7.4 in static. HFP to phospholipids ratio is 1:50. We
acquired the 2H spectra by varying  and 1. We arrayed  and 1 to 1000 and 975 ¾s (top), 1540
and 1521 ¾s (bottom). For each  and 1, the number of scans was 800. We processed the data
with -11 data shift pts, 500 Hz Gaussian line broadening and polynomial baseline correction of
the order 5.

192

Figure 5.33 Fitting plots of Chol_d7 ln (CD3 peak intensity) from membrane of POPC: POPG:
Chol_d7 (8:2:5) without HFP at different temperatures and pH 7.4. We acquired the data in static
condition.

193

Figure 5.34 Fitting plots of Chol_d7 ln (CD3 peak intensity) from membrane of POPC: POPG:
Chol_d7 (8:2:5) with HFP at different temperatures and pH 7.4 in static.

194

(a) no HFP

(b) with HFP

Figure 5.35 Fitting plots of Chol_d7 ln (CD3 peak intensity) from membrane of POPC: POPG:
Chol_d7 (8:2:5) without HFP (a) and with HFP (b) at 37 °C and pH 7.4 in static.

195

Table 5.4 Best-fit Chol_d7 2H T2 (s) values at different temperatures without and with HFP,
uncertainties are in parenthesis. T2 was fitted with ln (CD3 peak intensity) vs 2.
Temperature
T2

T2 (with HFP)

-50

93(2)

157(4)

5

1043(25)

1040(26)

25

1000(17)

990(30)

37

952(22)

943(26)

45

926(28)

1016 (30)

(C)

5.3.2.3 Spin lattice relaxation studies of Chol_d7
To evaluate the effect of HFP on the CHOL hydrocarbon tail motion, we investigated the spin
lattice relaxation (T1) for Chol_d7 at different temperatures: - 50°C, 5°C, 25°C, 37°C and 45°C.
The sample is POPC/POPG/Chol_d7 without and with HFP. Chol_d7 composition is 33%, and
HFP: phospholipids ratio is ~ 1:50. The Chol_d7 T1 is studied in static and pH 7.4 by t1D_ir
pulsed sequence, which is inversion-recovery followed by quecho. The pulse sequence is π ‒ 1‒
(π/2)x ‒ 2 ‒ (π/2)y ‒ 3 ‒ acquisition, shown in Figure 5.36. The data is acquired by arraying 1
and 2 and 3 are fixed at 50 ¾s. For each 1 within the array, the number of acquisition is the
196

same and is typically 3000 scans. We processed the data with Gaussian line broadening, data
shift and polynomial baseline correction. Figure 5.37 and 5.38 display representative FID and
stacked plots of processed spectra for Chol_d7 without and with HFP, respectively. Chol_d7
spectrum has two pairs of horns (Pake doublet), one pair has bigger intensity with ~ 1.6 kHz
peak splitting corresponding to the six CD3 deuterons, and the other pair has much smaller
intensity with ~ 26 kHz peak splitting corresponding to the CD deuteron present in Chol_d7.
Chol_d7 T1 is typically from the CD3 peak intensity fitting. At - 50°C in pure membrane, the
sample temperature is ~ - 30°C, the peaks are broader, and the CD3 peak two horns are not
resolved due to less motional averaging of quadrupolar anisotropy at lower temperature.
Therefore, the integrated peak intensity is used, and the integration is with 300-ppm integration
width centered at the peak center. The peak intensity is fitted vs 1 by the equation:
𝜏1

𝐼(𝜏1) = 𝐼0 + ∆𝐼 × (1 − exp(− 𝑇 ))

5.2

1

Where I(1) is the experimental peak intensity, I0, ∆I and T1 are fitting parameters. I0 is I (1=0),
∆I = I (1→∞) – I (1=0). Figures 5.39 to 5.41 show the best-fit plots. Table 5.4 show best-fit T1
values. At and above 5 °C, presence of HFP does not affect Chol_d7 T1. T1 is slightly shorter
with HFP at - 50°C.

197

Figure 5.36 “t1D_ir” pulse sequence for T1 relaxation study.

198

(a)

(b)

Figure 5.37 “t1D_ir” experimental (red squares) of Chol_d7 from membrane of POPC: POPG:
Chol_d7 (8:2:5) without HFP at 37 °C and pH 7.4 in static. The number of scans is 3000 for each
1. (a) 2H FID for 1 = 0.1 and 150.1ms, (b) Chol_d7 2H spectra for 1 = 0.1 through 150.1ms.
We did not show spectra for 1 = 180.1 through 510.1ms for view simplicity. All Spectra are
processed with 500 Hz Gaussian line broadening, -7 data shift points, and polynomial baseline
correction of order 5.

199

(a)

(b)

Figure 5.38 “t1D_ir” experimental (red squares) of Chol_d7 from membrane of POPC: POPG:
Chol_d7 (8:2:5) with HFP at 37 °C and pH 7.4 in static. The HFP to phospholipids mole ratio is
1:50. The number of scans is 3000 for each 1. (a) 2H FID for 1 = 0.5 and 120.5ms, (b) Chol_d7
2

H spectra for 1 = 0.5 through 120.5ms. We did not show spectra for 1 = 140.5 through

340.5ms for view simplicity. We processed the spectra with 500 Hz Gaussian line broadening,
data shift of -12, and polynomial baseline correction of order 5.

200

Figure 5.39 “t1D_ir” experimental (red squares) and best fit (red line) of Chol_d7 CD3 peak
intensity vs 1 from membrane of POPC: POPG: Chol_d7 (8:2:5) without and with HFP at 37 °C
and pH 7.4 in static. HFP: phospholipids mole ratio is 1:50.

201

Figure 5.40 “t1D_ir” experimental (red squares) and best fit (red line) of Chol_d7 CD3 peak
intensity vs 1 from membrane of POPC: POPG: Chol_d7 (8:2:5) without HFP at different
temperatures and pH 7.4 in static.

202

Figure 5.41 “t1D_ir” experimental (red squares) and best fit (red line) of Chol_d7 CD3 peak
intensity vs 1 from membrane of POPC: POPG: Chol_d7 (8:2:5) with HFP at different
temperatures and pH 7.4 in static. HFP: phospholipids mole ratio is 1:50.

203

Table 5.5 Best-fit Chol_d7 CD3 T1 (ms) values at different temperatures, uncertainties are in
parenthesis. T1 was fitted with the CD3 peak intensity vs 1.
Temperature
T1/ms

T1 (with HFP)/ms

-50

18.4(7)

15.8(2)

5

62(1)

62(1)

25

97(1)

94(1)

37

122 (1)

128(3)

45

143(1)

137(1)

(C)

5.4 Discussion
The static quecho experiments studied in this chapter is majorly to investigate the membrane
motions including PC and CHOL perturbed by HFP. Our earlier REDOR results have shown that
HFP has preferential contact to CHOL vs PC when associated with membrane at several residue
sites along the peptide chain. Specifically, that is G5 and G10 residues inserted deeply into the
membrane center has preferential contact to Chol_d7 terminal deuterons than PC_d10 terminal
deuterons near the membrane center, and G16 has preferential contact to Chol_d6 deuterons than
PC_d4 deuterons near the membrane surface. However, there is little information about how

204

CHOL catalyze HFP induced membrane fusion. To help understand this, we investigate how
HFP perturbs membrane motion in membrane both without and with CHOL in this study by 2H NMR.
5.4.1 HFP disrupts lipid acyl chain packing in membrane both without and with CHOL
To study the CHOL effect on HFP-membrane interaction, we used DMPC-d54 with perdeuterated acyl chain. We added anionic lipid to reflect the negatively charged membrane of
HIV host cell. To investigate the effect of HFP on membrane motion, we studied membrane both
without and with CHOL. In membrane with CHOL, 33% CHOL represents the typical CHOL
composition in HIV host cell membrane. For all the 2H spectra at different temperatures and
different membrane composition, the overall spectrum shape is similar without and with HFP,
which suggests that the anionic membrane remains lamellar membrane phase in addition of HFP
regardless of CHOL. Instead of powder pattern spectrum, a narrow (~ 200Hz FWHM) isotropic
deuteron peak is likely corresponding to formation of isotropic, non-lamellar lipid phase.[18, 19]
In membrane without CHOL, the spectrum individual peaks become broader and less resolved in
presence of HFP, and there is ~ 0.75 kHz line broadening due to shorter T2 with 4% HFP. In
membrane without CHOL, at 21 C, the broad spectrum indicates that the membrane is in gel
phase. However, the static spectrum becomes much narrower with sharp peaks similar to the
spectrum at higher temperature when the membrane is in liquid disordered phase. These results
indicate that HFP lowers the DMPC-d54 phase transition temperature in anionic membrane
without CHOL because the phase transition temperature of DMPC-d54 is 23 C. Similar effect
has been observed for Influenza peptide at pH 5.0.[20] In membrane with 33% CHOL, the
spectrum feature at 37 C is similar with earlier studied of HFP in LM3 membrane at 35 C by
Jun Yang and Charles GABRYS. Adding HFP makes individual peaks broader and less resolved,

205

and there is ~ 0.90 kHz line broadening due to shorter T2 with 4% HFP. However, the peak
splitting is ~ 10% smaller for CD3 peak, and ~ 3% smaller for the CD2 peak in presence of 1%
HFP, while the peak splitting is ~ 15% bigger for CD3, and ~ 20% bigger for CD2 in presence of
4% HFP. In membrane without CHOL, the peak splitting is ~ 10% smaller for CD3 peak and
almost the same for CD2 peak in presence of 4% HFP. The results with 4% HFP in the
membrane without and with CHOL suggest a role of CHOL in membrane fusion. CHOL restricts
fast motion of PC because the spectrum is broader with CHOL. HFP slightly increases fast
motion of lipids in membrane without CHOL, but restrict the fast motions of lipids with CHOL.

206

Table 5.6 Peak splitting for CD3 and CD2 of DMPC-d54 powder pattern spectrum at 37 °C
CD3 peak splitting

CD2 peak splitting

Membrane

HFP amount
/kHz

/kHz

3.23

25.68

0%

2.89 (10 % ↓)

25.66 (0 %)

4%

5.71

37.29

0%

5.07 (10% ↓)

36.16 (3% ↓)

1%

6.58 (15% ↑)

44.70 (20% ↑)

4%

d54pg

d54pgCHOL

De-Pake-ing the data gives better-resolved spectra. Figures 5.10 to 5.12 show the de-Paked
spectra. The peaks with the smallest and biggest Q comes from the terminal CD3 deuterons
near the membrane center and CD2 deuterons near the membrane surface, respectively. The
largest Q peak corresponds to deuterons from C2 to C5. Other well-resolved peaks are from
deuterons attached from C6 to C13, with smaller Q coming from carbon positions closer to the
membrane center. There are similar trends of Q for middle deuterons.
In membrane without CHOL, HFP decreases acyl chain order of lipid by ~ 1-10% with 4% HFP.
However, HFP increases acyl chain order of lipid by ~ 20-30% in membrane with CHOL with 4%

207

HFP (figure 5.14). The opposite effect of HFP on membrane lipid acyl chain ordering is because
CHOL can help regulate the membrane ordering.[21] Adding HFP perturbs the local membrane
packing, CHOL is a smaller molecule relative to phospholipids, and able to pack well to the void
space between HFP and nearby lipid molecule, and thus restricts the motion of lipids and
increase acyl chain order parameter of lipid. From these results, CHOL probably catalyze
membrane fusion by stabilizing the high-energy membrane intermediate states and thus lowering
the energy barrier to achieve the membrane fusion intermediate states. While in membrane
without CHOL, HFP perturbs membrane packing, the less than 10% of lipid acyl chain
disordering is probably due to the local lipids nearby HFP. From the order parameter profile of
lipid acyl chain, HFP affects order parameter down to the membrane center independent of
CHOL, which is consistent with the earlier major deeply inserted HFP membrane location model
in membrane both without and with CHOL. The deeply inserted membrane location is
responsible for fusion.
In membrane with 33% CHOL, HFP perturbs DMPC-d54 acyl chain order dependent on HFP
concentrations. Specifically, with 1% and 4% HFP, lipid acyl chain order decreases and
increases, respectively. Figure 5.13 and figure 5.14 shows the order parameter and order
parameter change induced by HFP. In presence of 1% HFP, HFP increases membrane disorder
along the whole acyl chain, and with greater disordering extent toward the membrane center. In
presence of 4% HFP, HFP increases membrane order also along the whole lipid acyl chain, and
with bigger ordering extent from middle of acyl chain toward the membrane center. Both 1% and
4% HFP affect lipid acyl chain packing down to the lipid tail at the membrane center.

208

5.4.2 Transverse relaxation studies of PC and CHOL
To study the effect of HFP on lipid transverse relaxation (T2), we used DMPC-d54 in anionic
membrane both without and with 33% CHOL. Except in membrane without HFP at 21 °C,
adding HFP generally decreases PC T2 ~ 68 % independent of CHOL fitted from echo, CD3 and
CD2 intensity. In membrane without HFP at 21°C, adding HFP increased PC T2 fitted from echo,
CD3 and CD2 intensity while CD2 T2 is not determined in membrane without HFP because the
intensity is too weak and decays too fast due to broad CD2 spectrum in gel phase. T2 at and
above 25 °C is similar for the same sample. We also compared the PC T2 with different
concentrations of HFP. For 1% and 4% HFP, T2 fitted form echo intensity is ~ 23% smaller or
the T2 relaxation rate is ~ 23% faster with 4% HFP relative to with 1% HFP. However, T2 fitted
from CD3 and CD2 peak intensity is similar. These results suggest that the T2 effect of HFP on
CD3 and CD2 effect is the same with smaller and larger quantities of HFP, while larger quantities
of HFP has bigger T2 decrease on other CD2s along the acyl chain other than the terminal CD2s
(C2-C5). To investigate the effect of HFP on CHOL transverse relaxation (T2), where both
Chol_d6 and Chol_d7 where the deuterons are located near the membrane surface and membrane
center, respectively. In membrane with 33% CHOL at 37 °C, the Chol_d6 T2 is similar in DMPC
and POPC membranes fitted from echo intensity. In DMPC membrane, T2 is 358 (13) and 230 (7)
for without and with HFP respectively. In POPC membrane, T2 is 312 (28) and 224 (6) for
without and with HFP respectively. The T2 decreases or the T2 relaxation rate increases ~ 35%
and ~ 28% in DMPC and POPC membrane, respectively. We studied the T2 of Chol_d7 with
POPC in anionic membrane with 33% CHOL, which is the most common membrane
composition for vesicle fusion study of viral fusion peptides and proteins. At - 50°C, the T2
increases in presence of 2% HFP. At temperature above and including 5 °C, the T 2 is similar

209

without and with HFP, and is similar for the same sample regardless of temperature increasing.
These results suggest that the effect of HFP on Chol_d7 motion change is negligible on the time
scale of ~ 10-4 s.
T2 for DMPC-d54 and Chol_d6 decrease probably because the membrane curvature changes by
presence of HFP when it interacts with the membrane. With greater membrane curvature, the CD bond will experience more orientation diversity relative to the external magnetic field. Thus,
the quadrupolar field will have greater change, leading to faster relaxation and shorter T 2. T2
increases for DMPC-d54 in gel phase anionic membrane without CHOL and Chol_d7 in POPC
anionic membrane in presence of HFP, which is because HFP increases nearby C-D bond axial
rotation and the increased motion leads to the observed longer T2. However, the fast axial CD3
bond rotation and longitudinal diffusion induced by increased temperature overcomes the effect
induced by HFP, so the T2 is similar without and with HFP.
Interestingly, the T2 for DMPC-d54 decreases independent of CHOL, which suggests that HFP
induce membrane curvature independent of CHOL and helps explain HFP induce fusion in
membrane both without and with HFP.[11] The T2 decreases or the T2 relaxation rate increases ~
75% by 4% HFP in membrane without CHOL, and the T2 decreases or the T2 relaxation rate
increases ~ 85% by 4% HFP in membrane with CHOL. The greater reduction in T2 in membrane
with CHOL might due to greater membrane curvature induced by HFP in presence of CHOL.
The greater membrane curvature facilitates faster transition into the membrane fusion
intermediate states because of the smaller energy gap between the curved membrane and the
fusion intermediate states and leading to increased membrane fusion rate.

210

5.4.3 Spin lattice relaxation studies of Chol_d7
To investigate the spin lattice relaxation (T1 relaxation) for Chol_d7, we studied OPC/POPG
membrane with 33% CHOL at various temperatures. We analyzed the CD3 deuterons. At - 30°C,
T1 decreases (or T1 relaxation rate) slightly (~ 8%) in presence of ~ 2% HFP. This result
indicates that HFP restricts T1 relaxation of Chol_d7 CD3 deuterons. This is probably because
the interaction between HFP and Chol_d7 restricts the fast rotational and translational motions of
the CD3 that interacts with HFP. This is consistent with the preferential contact between HFP and
CHOL vs PC observed at - 30°C. At and above 5 °C, T1 is similar without and with HFP. T1 is
longer for the same sample at higher temperature due to increased motion.

211

REFERENCES

212

REFERENCES

1.

Seelig, J. and P.M. Macdonald, PHOSPHOLIPIDS AND PROTEINS IN BIOLOGICALMEMBRANES - H-2 NMR AS A METHOD TO STUDY STRUCTURE, DYNAMICS, AND
INTERACTIONS. Accounts of Chemical Research, 1987. 20(6): p. 221-228.

2.

Salnikov, E.S., A.J. Mason, and B. Bechinger, Membrane order perturbation in the
presence of antimicrobial peptides by H-2 solid-state NMR spectroscopy. Biochimie,
2009. 91(6): p. 734-743.

3.

Xie, L., et al., Multiple Locations of Peptides in the Hydrocarbon Core of Gel-Phase
Membranes Revealed by Peptide C-13 to Lipid H-2 Rotational-Echo Double-Resonance
Solid-State Nuclear Magnetic Resonance. Biochemistry, 2015. 54(3): p. 677-684.

4.

Jia, L.H., et al., REDOR solid-state NMR as a probe of the membrane locations of
membrane-associated peptides and proteins. Journal of Magnetic Resonance, 2015. 253:
p. 154-165.

5.

Xie, L., et al., Residue-specific membrane location of peptides and proteins using
specifically and extensively deuterated lipids and C-13-H-2 rotational-echo doubleresonance solid-state NMR. Journal of Biomolecular Nmr, 2013. 55(1): p. 11-17.

6.

Brown, M.F. and G.D. Williams, MEMBRANE NMR - A DYNAMIC RESEARCH AREA.
Journal of Biochemical and Biophysical Methods, 1985. 11(2-3): p. 71-81.

7.

Seelig, J. and W. Niederberger, DEUTERIUM-LABELED LIPIDS AS STRUCTURAL
PROBES IN LIQUID-CRYSTALLINE BILAYERS - DEUTERIUM MAGNETICRESONANCE STUDY. Journal of the American Chemical Society, 1974. 96(7): p. 20692072.

8.

Sani, M.A., et al., A practical implementation of de-Pake-ing via weighted Fourier
transformation. Peerj, 2013. 1: p. 8.

9.

Bloom, M. and E. Sternin, TRANSVERSE NUCLEAR-SPIN RELAXATION IN
PHOSPHOLIPID-BILAYER MEMBRANES. Biochemistry, 1987. 26(8): p. 2101-2105.

10.

Harris, R.K., Nuclear Magnetic Resonance Spectroscopy. Longman Scientific &
Technical. 1986, England.

11.

Rong Yang, M.P., ‡ Francis J. Castellino,‡ and David P. Weliky*,†, A Trimeric HIV-1
Fusion Peptide Construct Which Does Not Self-Associate in Aqueous Solution and Which
Has 15-Fold Higher Membrane Fusion Rate. Journal of American Chemical Society,
2004. 126(45): p. 14722-14723.

213

12.

Yang, S.T., et al., HIV gp41-mediated membrane fusion occurs at edges of cholesterolrich lipid domains. Nature Chemical Biology, 2015. 11(6): p. 424-+.

13.

Gabrys, C.M., et al., Solid-State Nuclear Magnetic Resonance Measurements of HIV
Fusion Peptide (CO)-C-13 to Lipid P-31 Proximities Support Similar Partially Inserted
Membrane Locations of the alpha Helical and beta Sheet Peptide Structures. Journal of
Physical Chemistry A, 2013. 117(39): p. 9848-9859.

14.

Ratnayake, P.U., et al., Full-length trimeric influenza virus hemagglutinin II membrane
fusion protein and shorter constructs lacking the fusion peptide or transmembrane
domain: Hyperthermostability of the full-length protein and the soluble ectodomain and
fusion peptide make significant contributions to fusion of membrane vesicles. Protein
Expression and Purification, 2016. 117: p. 6-16.

15.

Ratnayake, P.U., et al., pH-dependent vesicle fusion induced by the ectodomain of the
human immunodeficiency virus membrane fusion protein gp41: Two kinetically distinct
processes and fully-membrane-associated gp41 with predominant beta sheet fusion
peptide conformation. Biochimica Et Biophysica Acta-Biomembranes, 2015. 1848(1): p.
289-298.

16.

Sackett, K., et al., Hairpin Folding of HIV gp41 Abrogates Lipid Mixing Function at
Physiologic pH and Inhibits Lipid Mixing by Exposed gp41 Constructs. Biochemistry,
2009. 48(12): p. 2714-2722.

17.

Lai, A.L., et al., Fusion Activity of HIV gp41 Fusion Domain Is Related to Its Secondary
Structure and Depth of Membrane Insertion in a Cholesterol-Dependent Fashion. Journal
of Molecular Biology, 2012. 418(1-2): p. 3-15.

18.

Gabrys, C.M., et al., Nuclear magnetic resonance evidence for retention of a lamellar
membrane phase with curvature in the presence of large quantities of the HIV fusion
peptide. Biochimica Et Biophysica Acta-Biomembranes, 2010. 1798(2): p. 194-201.

19.

Yang, J., et al., Solid state NMR measurements of conformation and conformational
distributions in the membrane-bound HIV-1 fusion peptide. Journal of Molecular
Graphics & Modelling, 2001. 19(1): p. 129-135.

20.

Ghosh, U., Solid State NMR Studies of Structure and Dynamics of Membrane Associated
Influenza Fusion Peptides, in Chemistry. 2016, Michigan State University.

21.

Yang, S.T., et al., The role of cholesterol in membrane fusion. Chemistry and Physics of
Lipids, 2016. 199: p. 136-143.

214

Chapter 6 - Summary and Future work
Over the last ~ 4 years I have been mainly working on two projects:
(1) Using

13

C-2H REDOR NMR method as a probe to study the membrane locations of HFP

constructs including wildtype, HFP_V2E mutant and HFP_L9R mutant, and transmembrane
peptide KALP.
(2) Using 2H-static NMR method to study the membrane dynamics of lipid bilayer affected by
the presence of HFP.
6.1 Summary of 13C-2H REDOR as a probe for membrane location study and future work
of HFP
To study the membrane locations of HFP constructs and KALP in the membrane hydrophobic
core, the peptide backbone is

13

CO labeled at specific residue and membrane is selectively

deuterated at different regions of the lipid acyl chain or CHOL (labeling scheme see figure 2. ).
We study the peptide membrane contacts by 13C- 2H REDOR.[1-4]
The experimental REDOR dephasing supports multiple membrane locations for both HFP
wildtype and HFP_V2E mutant in DPPC: DPPG (4:1) membrane. In membrane containing ~ 33%
deuterated CHOL, there is similar multiple membrane locations for HFP wildtype in DOPC:
DOPG and POPC: POPG membranes compared to DPPC: DPPG membranes. These results for
HFP wildtype suggest that the multiple membrane locations distribution is robust and probably
regardless of the membrane composition. The multiple membrane locations of KALP are
because of the hydrophobic mismatch between the KALP hydrophobic length and the membrane
and the resultant snorkeling effect of lysine sidechains to the membrane head-group region.

215

As for the HFP wildtype membrane location study, an interesting discovery is that HFP has
preferential contact to neighboring CHOL than lipid acyl chain. In DPPC: DPPG: CHOL (8:2:5)
with HFP_G5c, there is ~2 times dephasing with Chol_d7 labeling than with PC_d10 (deuterated
DPPC) labeling, where the deuterons are both located near the membrane center without peptides.
The dephasing is comparable with Chol_d7 labeling and PC_d10 labeling when the membrane
components ratio is DPPC: DPPG: CHOL= 8:2:2.5. HFP_G10c give similar dephasing
compared to HFP_G5c in membrane DPPC: DPPG: CHOL= 8:2:2.5. HFP_G16c also has
preferential contact to Chol_d6 than lipid acyl chains near the membrane surface. The
preferential contact of HFP to CHOL vs PC might shed some light on understanding the
increased fusion in membrane with CHOL vs without CHOL. CHOL increase fusion via two
features of CHOL vs PC: (1) CHOL has greater intrinsic curvature which is more favorable in
membrane intermediate state during fusion; (2) CHOL tail is closer to the membrane surface
because CHOL is shorter.[5]
For future work of HFP, it will be interesting to study the residue specific membrane locations of
HFP in the membrane hydrophobic core using larger gp41 constructs including the HFP region.
The larger gp41 construct can be FP-Hairpin (FP-NHR-Loop-CHR) or FP-HM (FP-NHR-LoopCHR-MPER). We can study the membrane location in membrane with and without CHOL. In
membrane with CHOL, we can compare the dephasing with PC_d10 and Chol_d7, or PC_d4
with Chol_d6 to see whether there is preferential contact of FP in the larger gp41 constructs to
CHOL vs PC. In membrane containing ~ 33% CHOL, greater or similar dephasing with
deuterated CHOL than deuterated PC would support preferential contact to CHOL, while about
half or less dephasing with deuterated CHOL compared to PC would be consistent with no
preferential contact to CHOL.

216

6.2 Summary of membrane motions perturbed by HFP and future work
My second project is to investigate the changes in membrane structure and motions in presence
of HFP. We use per-deuterated PC (DMPC-d54) and deuterated CHOL (Chol_d6 and Chol_d7)
to investigate the change by evaluating the spectrum and T2 and T1 relaxation times by 2H- NMR.
We observed powder pattern of DMPC-d54 even in presence of 4% HFP suggests that the
membrane maintains the lamellar bilayer structure. The spectrum changes are CHOL and peptide
quantity dependent.
For changes of DMPC-d54, addition of HFP broadens the fine peaks in DMPC-d54 spectrum. In
membrane without CHOL, the CD3 and CD2 peak splitting is narrower with HFP relative to
without HFP in membrane without CHOL and the acyl chain order parameter decreases less than
5% along the acyl chain with 4% HFP relative to without HFP. However, in membrane with ~ 33%
CHOL, the peak splitting is significantly broader and the acyl chain order parameter increases
about 20% to 30% with HFP relative to without HFP. Therefore, HFP increases the motion on
the time scale of ~ 10-5s of membrane without CHOL, and decreases the motion on the time scale
of ~ 10-5s of membrane with CHOL. HFP generally decreases DMPC-d54 T2 ~ 68 %
independent of CHOL above the DMPC phase transition temperature. With 1% and 4% HFP, T2
fitted form echo intensity is ~ 23% smaller with 4% HFP relative to with 1% HFP. For changes
of Chol_d6, in membrane with 33% CHOL at 37 °C, the Chol_d6 T2 is similar in DMPC and
POPC membranes. The T2 decreases ~ 35% and ~ 28% in DMPC and POPC membrane,
respectively. These results suggest that presence of HFP restricts the motion of Chol_d6 and
there is little effect of membrane composition on Chol_d6 motion while the membrane is likely
in liquid ordered phase both in POPC/CHOL membrane and in DMPC/CHOL membrane at
37 °C.[6, 7] For changes of Chol_d7, for temperature above and including 5 °C, the T2 is similar

217

without and with HFP, and is similar for the same sample regardless of temperature increasing.
The decreased T2 of DMPC-d54 and Chol_d6 is probably due to increased membrane curvature
in presence of HFP. And the similar T2 for Chol_d7 indicates that the effect of HFP on Chol_d7
motion is negligible. Chol_d6 has the deuteron on the rigid ring system, and would have slow
motions, while Chol_d7 have deuterons in the isopropyl tail which has fast axial rotation of the
C-D bond.
For future work, it will be interesting to study the structure and motion changes in membrane
induced by HFP_V2E and HFP_L9R mutants to correlate the membrane change induced by HFP
constructs and fusogenicity. It’s worthwhile to study the membrane both without and with CHOL
to understand the role of CHOL in fusion. It’s also interesting to study the membrane change
induced by FP using larger gp41 constructs. The larger gp41 constructs can be FP-Hairpin (FPNHR-Loop-CHR) and FP-HM (FP-NHR-Loop-CHR-MPER) and the corresponding nonfusogenicity V2E and fusogenicity L9R mutants.

218

REFERENCES

219

REFERENCES

1.

Gullion, T., Introduction to rotational-echo, double-resonance NMR. Concepts in
Magnetic Resonance, 1998. 10(5): p. 277-289.

2.

Xie, L., et al., Multiple Locations of Peptides in the Hydrocarbon Core of Gel-Phase
Membranes Revealed by Peptide C-13 to Lipid H-2 Rotational-Echo Double-Resonance
Solid-State Nuclear Magnetic Resonance. Biochemistry, 2015. 54(3): p. 677-684.

3.

Jia, L.H., et al., REDOR solid-state NMR as a probe of the membrane locations of
membrane-associated peptides and proteins. Journal of Magnetic Resonance, 2015. 253:
p. 154-165.

4.

Xie, L., et al., Residue-specific membrane location of peptides and proteins using
specifically and extensively deuterated lipids and C-13-H-2 rotational-echo doubleresonance solid-state NMR. Journal of Biomolecular Nmr, 2013. 55(1): p. 11-17.

5.

Zimmerberg, J. and M.M. Kozlov, How proteins produce cellular membrane curvature.
Nature Reviews Molecular Cell Biology, 2006. 7(1): p. 9-19.

6.

Almeida, P.F.F., W.L.C. Vaz, and T.E. Thompson, LATERAL DIFFUSION IN THE
LIQUID-PHASES OF DIMYRISTOYLPHOSPHATIDYLCHOLINE CHOLESTEROL
LIPID BILAYERS - A FREE-VOLUME ANALYSIS. Biochemistry, 1992. 31(29): p. 67396747.

7.

Thewalt, J.L. and M. Bloom, PHOSPHATIDYLCHOLINE - CHOLESTEROL PHASEDIAGRAMS. Biophysical Journal, 1992. 63(4): p. 1176-1181.

220

APPENDICES

221

APPENDIX A
Preparation and characterization of FPHM_V2E mutant of gp41 ectodomain

222

A.1 Preparation of FPHM_V2E mutant
Figure A.1 shows the FPHM_V2E mutant amino acid and DNA sequence. Studies have shown
that V2E mutant in gp41 eliminates cell-cell fusion.[1] We study the structure and oligomeric
state of FPHM_V2E mutant to help understand gp41 catalyzed fusion. FPHM amino acid
sequence is shown in figure 1.16, and is from residue 512 to 683 according to g160 sequence
numbering for the HXB2 laboratory strain of HIV, and the region from residue 582 to 627 in
gp41 is replaced with SGGRGG loop which does not affect the SHB assembly.[2-4] Crystal
structure shows flexible (crystallography disordered) structure from residue 601 to 613 and
helical structure from residue 582 to 600 and 614 to 627.[5] PHM_V2E mutant is engineered
through point mutation of FPHM by polymerase chain reaction (PCR) using FPHM template
DNA and designed primers including the point mutation. FPHM template plasmid is extracted
from overnight grown cells. Primers are designed including the V2E point mutation. The forward
primer sequence is 5’ - CAT ATG GCC GAG GGT ATC GGT G - 3’. The reverse primer
sequence is 5’ – CAC CGA TAC CCT CGG CCA TAT G - 3’. After PCR, DNP1 enzyme is
added to the final PCR product to cleave mother template DNA (FPHM DNA). We run the PCR
product on agarose gel and compare with DNA ladder to check successfulness of the mutation.
The successful PCR product with the point mutation is taken up by BL21 (DE3) competent cells.
Then we incubate the cells on agar plate overnight in the incubator at 37 °C. Colony on the agar
plate is picked to grow overnight in 100mL LB at 37 °C. Glycerol stocks were made with 1mL
cell cultures and 50% glycerol and stored at -80°C.[6] The plasmid is extracted from overnight
grown cells using plasmid extraction kit (Promega) and sent for DNA sequencing to check the
successfulness of the V2E point mutation.

223

To express the protein, glycerol stocks of the cells containing the FPHM_V2E mutant plasmid is
incubated overnight in 50mL LB with 50mg/L kanamycin antibiotic at 37 °C with stirring at 180
rpm. Then the culture is added to 1L fresh LB in baffled flask. Expression of the protein is
induced by adding isopropyl β–D-1-thiogalactopyranoside (IPTG) to a final concentration of
2mM after the cell culture OD600 ≈ 0.8. After 5hrs protein expression, the cells are harvested by
centrifugation at 9000g for 10mins at 4 °C. The cell pellets are stored at -20 °C until use.
(a) FPHM_V2E amino acid sequence:
AEGIGALFLGFLGAAGSTMGARSMTLTVQARQLLSGIVQQQNNLLRAIEAQQHLLQLTV
WGIKQLQARILSGGRGGWMEWDREINNYTSLIHSLIEESQNQQEKNEQELLELDKWASL
WNWFNITNWLWYIKGGGGGGLEHHHHHH
(b) FPHM_V2E DNA sequence:
GCCGAGGGTATCGGTGCTCTGTTCCTGGGTTTCCTGGGTGCTGCTGGTTCGACGATG
GGTGCCCGCTCAATGACGCTGACGGTCCAAGCACGTCAGCTGCTGAGCGGCATTGT
GCAGCAACAGAACAATCTGCTGCGCGCGATCGAAGCCCAACAGCATCTGCTGCAGC
TGACCGTTTGGGGTATTAAACAACTGCAGGCTCGTATCCTGAGCGGCGGTCGCGGCG
GTTGGATGGAATGGGATCGTGAAATTAACAATTATACGAGCCTGATTCACTCTCTGA
TCGAAGAAAGTCAAAACCAACAGGAGAAAAACGAACAGGAACTGCTGGAACTGGA
CAAATGGGCCTCCCTGTGGAACTGGTTTAACATTACGAACTGGCTGTGGTACATCAA
AGGCGGCGGTGGCGGTGGTCTCGANCACCACCACCACCACCAC
Figure

A1

FPHM_V2E

amino

acid

and

DNA

sequences.

The

C-terminal

GGGGGGLEHHHHHH residues are non-native tag. SGGRGG is the engineered loop.

224

To get purified FPHM_V2E protein, ~ 4g cell pellet is tip sonicated in ~ 40mL pH 7.4 PBS
buffer (10mM Na2HPO4 + 2mM KH2PO4 + 3mM KCl + 140mM NaCl). Sonication is done with
60% amplitude for 0.6s and followed by 0.4s rest for 1min. And there are 5 rounds × 1min
sonication separated by 1min rest. Then the inclusion bodies which are rich in the FPHM_V2E
proteins are separated from the rest part of the cells by centrifugation at 48000g for 20mins at
4 °C. The inclusion bodies are subject for another 2 rounds of sonication in PBS buffer followed
by centrifugation. Next, the inclusion body pellets are solubilized in 8M Urea + 0.8% Nlauroylsarcosine (SRC) + 0.5% SDS PBS buffer (solubilization buffer) by tip sonication.
Sonication is done with 70% amplitude for 0.8s and followed by 0.2s rest for 1min. And there
are 5 rounds × 1min sonication separated by 1min rest. There is no visible solid after the mixture
is centrifuged at 48000g for 20 mins at 4 °C which supports complete solubilization.
The protein is purified affinity chromatography. First, 1mL Co2+ affinity resin is added to the ~
40mL solutions containing the FPHM_V2E protein. Then the mixture is agitated for 2hrs at
ambient temperature to allow protein binding to the resin. Next, the protein bound resin is
separated from solution through fritted column and followed by 1mL × 4 rounds solubilization
buffer wash. Then the protein is eluted with solubilization buffer + 250mM imidazole (elution
buffer). The protein is quantitated with A280, and the purification yield is ~ 0.7 mg from 1L
culture. Figure A.2 shows the SDS-PAGE of purification products. The elution shows pure band
around 15kDa and is consistent with FPHM_V2E molecular weight. This band is submitted to
MSU Proteomics Facility center to identify the FPHM_V2E by trypsin digestion and analysis of
the resultant peptides. There is 88% amino acid coverage which identifies FPHM_V2E.

225

Figure A2 SDS-PAGE of FPHM_V2E mutant. PBS wash is the supernatant of the inclusion
bodies in PBS buffer at pH 7.4. Filter through and wash 1-4 is in solubilization buffer, and
Elution 1 is in solubilization buffer + 250mM imidazole.

Figure A3 Proteomics results of the SDS-PAGE band corresponding to FPHM_V2E protein.
Green color shaded M means there is detection of digested short peptides including oxidation
(+16). Green color shaded Q and N means there is detection of digested short peptides including
deamidated Q and N (+1) respectively. Green color shaded E means there is detection of digested
short peptides including dehydrated E (-18).
A.2 Structure and oligomeric state of FPHM_V2E mutant
Before structural characterization of the protein, the protein eluent is refolded by adding equal
volume of ice-cold buffer (10mM Tris-HCl, 0.17% decylmaltoside (DM), 2mM EDTA, and 1M

226

L-arginine at pH 8.0).[6] The mixture is agitated overnight at 4 °C followed by ~ 4 days dialysis
against CD-buffer (10mM Tris-HCl + 0.2% SDS at pH 7.4) with 4 times buffer change. We use
CD spectroscopy to characterize the secondary structure of the protein. Typical parameters
include: protein concentration of ~ 14 ÎźM, 1mm path-length and 260-185 nm spectral range with
1nm band width and 1s response time. Each spectrum is accumulation of 3 scans. Each scan is
baseline subtracted, and baseline is absorbance from CD-buffer only. The CD spectrum shows
two minima at 208nm and 222nm, which is consistent with overall Îą- helical secondary structure.
According to the mean residue molar elipticity at 222 nm (θ222), the protein is 45% helical
according to the maximum θ222value of -33,000 for ι helix.[7] There are 146 residues totally in
the FPHM_V2E sequence, and helicity of NHR (residue 535- 581) and CHR (residue 628-662)
accounts for 56% helicity. HFP_V2E has partial helical structure from some residues like L9,
L12 and A14 in membrane bilayer.[8] The 45% helicity is probably from NHR and CHR, and
the decreased helicity is likely due to SDS denaturation of the protein. To get the melting
temperature, the CD spectrum is taken from 25 °C to 90 °C with 5 °C increment. The mean
residue molar elipticity at 222 nm is plotted. The protein is still helical at 90 °C and the melting
temperature is above 90 °C.
To get the oligomeric state of FPHM_V2E, protein solution is analyzed by gel filtration
chromatography. The protein eluent is dialyzed against gel-filtration-buffer (10mM Tris-HCl +
0.2% SDS + 150mM NaCl at pH 7.4) for ~ 4days with 4 times buffer change after refolding. The
instrument is Pathfinder 20 instrument (Bio-Rad) with Superdex 200 increase column (General
Electric). The column is equilibrated with gel-filtration-buffer before running protein sample.
The protein concentration is ~0.9 mg/mL protein concentration. Eluent flow rate is 0.3mL/min.
Detection is A280nm. There is a single peak and peak is eluted at 14.23mL which corresponds to

227

molecular weight of 99kDa and FPHM_V2E is likely in trimeric state with molecular weight of
50kDa, and the rest 49kDa corresponds to SDS detergent molecules. SDS aggregation number is
~62, which corresponds to 17.8kDa mass per micelle. The 49kDa of SDS detergent molecules is
~ 3 micelles size. Then in the trimer peak, the FPHM_V2E protein: SDS micelle is ~1:1 mole
ratio.

228

Table A1 Mean residue molar elipticity of FPHM_V2E at 222nm (θ222) (deg.cm2.dmol-1.residue1

) at different temperatures
Temperature/ °C

θ222/ deg.cm2.dmol-1.residue-1

25

-15079

30

-15008

35

-14691

40

-13928

45

-13632

50

-13443

55

-13316

60

-13016

65

-12803

70

-12407

75

-12164

80

-11804

85

-11598

90

-11337

229

Figure A4 CD spectroscopy of FPHM_V2E (top panel) and melting temperature (bottom panel)
in 10mM tris(hydroxymethyl)aminomethane (Tris-HCl) + 0.2% SDS at pH 7.4. The melting
temperature plot is based on the mean residue molar elipticity at 222nm.

230

Figure A5 Gel filtration chromatograph of FPHM_V2E in 10mM Tris-HCl + 0.2% SDS +
150mM NaCl at pH 7.4. The highest peak is eluted at 14.23mL and corresponds to molecular
weight of 99kDa.
A.3 Lipid mixing assay of FPHM_V2E mutant in POPC: CHOL (2:1) vesicle at pH 7.4
Lipid mixing assay of FPHM_V2E is done with POPC: CHOL (2:1) ratio. Lipids are dissolved
in chloroform. Chloroform is removed by N2 gas followed by vacuum pumping. The lipid film is
dissolved in 1mL pH 7.4 buffer (5mM HEPES + 10mM MES), and homogenized by 10 freezethaw cycles. The mixture is then extruded through membrane film with 100nm pore size ~ 20
times. There are non-labeled lipids and fluorescent labeled lipids. The fluorescent labeled lipid
also contained 2mol % fluorescent lipid N-(7-nitro-2, 1, 3-benzoxadiazol-4-yl) (ammonium salt)
dipalmitoylphosphatidylethanolamine (N-NBD-PE) and 2mol % quenching lipid N-(lissamine
rhodamine B sulfonyl) (ammonium salt) dipalmitoylphosphatidylethanolamine (N-Rh-PE).
Labeled and unlabeled lipids are mixed in 1: 9 ratios to achieve total phospholipid (POPC)

231

concentration ≈ 150μM. The vesicle solution is stirred at 37 °C. Protein to lipids ratio is 1:300.
The fluorescence is measured with 467nm excitation and 530nm detection. The baseline
fluorescence is F0, protein stock solution (40ÎźM in CD-buffer (10mM Tris-HCl + 0.5% SDS at
pH 7.4)) is added at time 0, and there is increased fluorescence ΔF(t) = F(t)-F0 because the
distance between fluorescent and quenching lipids is longer in the fused vesicle (labeled +
unlabeled) than in the original labeled vesicle. The asymptotic fluorescence is usually reached by
~ 600s. Then 12μL 10 % Triton X-100 is added to achieve the maximum fluorescence (ΔFmax). %
lipid mixing =ΔF(t)/ ΔFmax × 100. There is ~ 0.5% lipid mixing with protein: lipids ratio of
1:300. However, FPHM wildtype induces ~ 4% lipid mixing under same conditions from Shuang
Liang’s work in our group. Our result supports that FPHM_V2E mutant of gp41 is nonfusogenic.
For future work, the protein structure characterization and lipid mixing assay need to be repeated
to check consistency. Comparing with FPHM wild type is necessary to gain insight on structure
and fusion correlation. For lipid mixing assay, it is interesting to study the negatively charged
vesicles, and study fusion at low pH to compare the electrostatic and hydrophobicity contribution
to FPHM induced fusion.

232

Figure A6 Lipid mixing assay of FPHM_V2E in POPC: CHOL= 2:1 vesicles at pH 7.4. Protein:
Lipids = 1:300 mole ratio.

233

APPENDIX B
NMR file locations

234

Figure 3.3
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_041213
Figure 3.4
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_053113 (organic)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_051713 (aqueous)
Figure 3.5
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_053113 (organic)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_051713 (aqueous)
Figure 3.6
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_053113 (- 30 °C)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_071613 (0 °C)
Figure 3.7 and Figure 3.8
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_041213 (PC_d10)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_042513 (PC_d8)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_050813 (PC_d4)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_053113 (PC_d10: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_090313 (PC_d8: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_040814 (PC_d4: PG)
Figure 3.9 and Figure 3.10
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_112713 (PC_d10: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_121713 (PC_d8: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_122613 (PC_d4: PG)
Figure 3.12

235

/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_051214 (PC: PG: Chol_d7)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_052614 (PC: PG: Chol_d6)
Figure 3.13
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_061614 (PC: PG: Chol_d7)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_060214 (PC: PG: Chol_d6)
Figure 3.15 and Figure 3.16
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/ G10c_050216 (Chol_d7 5)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/G10c_051416 (Chol_d6 5)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/ G10c _101415 (Chol_d7 2.5)
Figure 3.17 and Figure 3.18
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/ G16c_052616 (Chol_d6)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/G16c_060316 (Chol_d7)
Figure 3.19 and Figure 3.20
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/ HFP_070714 (Chol_d7)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/G16c_071714 (Chol_d6)
Figure 3.22
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/V2E_031214 (PC_d10: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/V2E_031914 (PC_d8: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/V2E_032714 (PC_d4: PG)
Figure 3.25
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_012714 (PC_d10: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_011414 (PC_d8: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_010214 (PC_d4: PG)

236

Figure 3.26
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_022214 (PC_d10: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_021214 (PC_d8: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_030314 (PC_d4: PG)
Figure 3.27 and Figure 3.28
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A5c_033115 (PC_d10: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A5c_040815 (PC_d8: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A5c_041515 (PC_d4: PG)
Figure 3.29 and Figure 3.30
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A7c_031315 (PC_d10: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A7c_021315 (PC_d8: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A7c_022015 (PC_d4: PG)
Figure 3.31 and Figure 3.32
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A17c_082815 (PC_d10: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A17c_092315 (PC_d8: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A17c_090915 (PC_d4: PG)
Figure 3.33 and Figure 3.34
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A19c_073015 (PC_d10: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A19c_081815 (PC_d8: PG)
/export/home/hapi0/mb4c/data/Lihui/13C2H/KALP/A19c_081015 (PC_d4: PG)
Figure 4.1
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_042614 (PC_d10: PG: CHOL)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_081114 (PC_d8: PG: CHOL)

237

/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_051214 (PC: PG: Chol_d7)
Figure 4.2
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_090814 (PC_d10)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_091614 (Chol_d7 2.5)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_112014 (Chol_d7 1.25)
Figure 4.3
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_010815 (PC_d10)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_121914 (Chol_d7)
Figure 4.4
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_072814 (PC_d10)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_081814 (PC_d8)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_080414 (PC_d4)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_061614 (Chol_d7)
Figure 4.5
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_060214 (Chol_d6 5)
/export/home/hapi0/mb4c/data/Lihui/13C2H/HFP/HFP_120614 (Chol_d6 2.5)
Figure 5.2
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/21C_one_ 021517
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/25C_one _021317
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/30C_one _021417
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/d54dmpg_one_113016 (37 °C)
Figure 5.3
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/21C_one_withHFP_021617

238

/export/home/hapi0/mb4c/data/Lihui/DMPCPG/25C_one_withHFP_021817
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/30C_one_withHFP_022017
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/d54dmpg_one_withHFP_120116 (37 °C)
Figure 5.4
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/21C_one_d54pgchol_ 020217
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/25C_one_d54pgchol_ 020117
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/30C_one_d54pgchol_ 013117
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/d54_noHFP_091516_one (37 °C)
Figure 5.5
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/21C_one_020617
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/25C_one_020817
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/30C_one_020917
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/d54pgchol_one_withHFP_1227
16 (37°C)
Figure 5.6
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/21C_one_ 021517
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/21C_one_withHFP_021617
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/d54dmpg_one_113016 (37 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/d54dmpg_one_withHFP_120116 (37 °C)
Figure 5.7
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/21C_one_d54pgchol_ 020217
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/21C_one_020617
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/d54_noHFP_091516_one (37 °C)

239

/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/d54pgchol_one_withHFP_1227
16 (37°C)
Figure 5.8
/export/home/khafre0/mb4b/data/Lihui/d54 _noHFP_one_101716 (pure lipids)
/export/home/khafre0/mb4b/data/Lihui/d54 _withHFP_one_101516 (1:100)
/export/home/khafre0/mb4b/data/Lihui/2umolHFP/d54 _withHFP_one_102416 (1:25)
Figure 5.9
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/Chold6_noHFP_090216 (pure lipids)
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/Chold6_withHFP_090816 (lipids+HFP)
Figure 5.15
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/21C_T2_2_ 021417 (21 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/d54dmpg_113016 (37 °C)
Figure 5.16
Same as figure 5.15
Figure 5.17
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/21C_T2_withHFP_ 021617 (21 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/d54dmpg_withHFP_120116 (37 °C)
Figure 5.18
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/21C_T2_d54pgchol_ 020317 (21 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/37C_T2_2_ 021017 (37 °C)
Figure 5.19
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/21C_T2_ 020617 (21 °C)

240

/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/d54pgchol_withHFP_122716
(37 °C)
Figure 5.20
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/37C_T2_2_021017 (without HFP)
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/d54_wihHFP_090716 (HFP:PL= 1:100)
Figure 5.21
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/Chold6_090216
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/Chold6_withHFP_090816
Figure 5.22
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/21C_T2_2_021417 (21 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/25C_T2_021217 (25 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/30C_T2_021317 (30 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/d54dmpg_113016 (37 °C)
Figure 5.23
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/21C_T2_withHFP_021617 (21 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/25C_T2_withHFP_021817 (25 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/30C_T2_withHFP_021917 (30 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPG/d54dmpg_withHFP_120116 (37 °C)
Figure 5.24
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/21C_T2_d54pgchol_020317 (21 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/25C_T2_d54pgchol_020217 (25 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/30C_T2_2_ 020917 (30 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/37C_T2_2_ 021017 (37 °C)

241

Figure 5.25
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/21C_T2_020617 (21 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/25C_T2_020717 (25 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/30C_T2_020817 (30 °C)
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/2umolHFP/d54pgchol_withHFP_122716
(37 °C)
Figure 5.26
/export/home/hapi0/mb4c/data/Lihui/DMPCPGChol/d54_wihHFP_090716
Figure 5.27
Same as figure 5.21
Figure 5.28 and Figure 5.29
/export/home/khafre0/mb4b/data/Lihui/Chold6popcpg_102416 (no HFP)
/export/home/khafre0/mb4b/data/Lihui/Chold6popcpg__withHFP_102716 (with HFP)
Figure 5.30
/export/home/hapi0/mb4c/data/Lihui/Chold6_070516/Chold6_45C_071216 (45 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold6_070516/Chold6_37C_071216 (37 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold6_070516/Chold6_25C_070516 (25 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold6_070516/Chold6_0C_070616 (0 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold6_070516/Chold6_-5C_070816 (-5 °C)
Figure 5.31
/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_-50C_T2array2_080316

(no

HFP)
/export/home/hapi0/mb4c/data/Lihui/Chold7_062716/Chold7_-50C_T2array_071816 (with HFP)

242

Figure 5.32
/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_37C_T2_080816 (no HFP)
/export/home/hapi0/mb4c/data/Lihui/ Chold7_062716/Chold7_37C_T2array_072816 (with HFP)
Figure 5.33
/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_-50C_T2array2_080316
50 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_5C_T2_080316 (5 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_25C_T2_080316 (25 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_45C_T2_080916 (45 °C)
Figure 5.34
/export/home/hapi0/mb4c/data/Lihui/Chold7_062716/Chold7_-50C_T2array_071816 (-50 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold7_062716/Chold7_5C_T2array_072216 (5 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold7_062716/Chold7_25C_T2array_072116_1 (25 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold7_062716/Chold7_45C_T2array_072916 (45 °C)
Figure 5.35
Same as figure 5.32
Figure 5.37
/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_37C_T1_080816
Figure 5.38
/export/home/hapi0/mb4c/data/Lihui/Chold7_062716/Chold7_37C_T1array_072816
Figure 5.39
Same as figure 5.37 and figure 5.38
Figure 5.40

243

(-

/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_-50C_T1array_080216 (-50 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_5C_T1_081016 (5 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_25C_T1_080516 (25 °C)
/export/home/hapi0/mb4c/data/Lihui/Chold7_m_080216/Chold7_45C_T1_080916 (45 °C)
Figure 5.41
/export/home/hapi0/mb4c/data/Lihui/Chold7_062716/Chold7_-50C_T1array_071816
/export/home/hapi0/mb4c/data/Lihui/Chold7_062716/Chold7_5C_T1array_072216
/export/home/hapi0/mb4c/data/Lihui/Chold7_062716/Chold7_25C_T1array_072116
/export/home/hapi0/mb4c/data/Lihui/Chold7_062716/Chold7_45C_T1array_072916

244

APPENDIX C
Data for fitting

245

Table C1 HFP_G10 and HFP_G16 dephasing △S/S0 in POPC: POPG: Chol_d7/d6, the data is
processed with 100Hz Gaussian line broadening and integration width of 3ppm for G10, and
1ppm for G16.
HFP_G10 in

HFP_G10 in

HFP_G10 in

HFP_G16 in

HFP_G16 in

POPC:PG:Chol_d

POPC:PG:Chol_d

POPC:PG:Chol_

POPC:PG:Chol_

POPC:PG:Chol_

7 (8:2:5)

7 (8:2:2.5)

d6 (8:2:5)

d6 (8:2:5)

d7 (8:2:5)

ΔS/So

Dephasi

error

ng

of

time/ms

ΔS/So

2

0.0079

0.0168

ΔS/So

ΔS/So

error of

ΔS/So

error

ΔS/So

error of

ΔS/So

ΔS/So

of

error of
ΔS/So

ΔS/So
-

0.0146

-

0.0047

0.0135

0.0311

0.0252

0.0086

0.0169

0.0059

8

0.1168

0.0087

0.0612

0.0122

0.0412

0.0173

0.1033

0.0269

0.0335

0.0291

16

0.2188

0.0109

0.0921

0.0161

0.0478

0.0123

0.1642

0.0273

0.0424

0.0201

24

0.3007

0.0177

0.1302

0.0133

0.0770

0.0210

0.2010

0.0381

0.1026

0.0267

32

0.3659

0.0269

0.1948

0.0288

0.1148

0.0179

0.3335

0.0178

0.1964

0.0344

40

0.4349

0.0232

0.1980

0.0216

0.0915

0.0320

0.3767

0.0142

0.2267

0.0338

48

0.4931

0.0552

0.2513

0.0266

0.1713

0.0304

Table C2 HFP_G5 dephasing △S/S0 in DOPC: DOPG: Chol_d7/d6 (8:2:5), the data is processed
with 100Hz Gaussian line broadening and 2ppm integration width.
Chol_d7

Chol_d6

Dephasing time/ms

ΔS/So

error of ΔS/So

ΔS/So

error of ΔS/So

2

0.0000

0.0170

0.0030

0.0137

8

0.0773

0.0182

0.0155

0.0189

16

0.1271

0.0203

0.0357

0.0121

24

0.2373

0.0297

0.0569

0.0474

32

0.2988

0.0330

0.1047

0.0375

40

0.4630

0.0622

0.2088

0.0558

246

Table C3 HFPL9R_G5 and G10 dephasing △S/S0 in DPPC: DPPG (4:1), the data is processed
with 100Hz Gaussian line broadening and 3ppm integration width.

Dephasing

HFPL9R_G5 with

HFPL9R_G5 with

HFPL9R_G5 with

PC_d10

PC_d8

PC_d4

ΔS/So

error of

ΔS/So

error of ΔS/So

ΔS/So

ΔS/So

time/ms

error of
ΔS/So

2

0.0275

0.0075

-0.0041

0.0091

0.0970

0.0117

8

0.1528

0.0183

0.0058

0.0199

0.1628

0.0128

16

0.2804

0.0213

0.1259

0.0260

0.1958

0.0119

24

0.4246

0.0339

0.0485

0.0314

0.2497

0.0232

32

0.5086

0.0392

0.1393

0.0530

0.3566

0.0283

40

0.5901

0.0481

0.0980

0.0769

0.3859

0.0363

48

0.7171

0.1157

0.2282

0.1044

0.4305

0.0727

Dephasing

HFPL9R_G10 with

HFPL9R_G10 with

HFPL9R_G10 with

PC_d10

PC_d8

PC_d4

ΔS/So

error of

ΔS/So

error of ΔS/So

ΔS/So

ΔS/So

time/ms

error of
ΔS/So

2

0.0397

0.0127

-0.0056

0.0113

0.1024

0.0098

8

0.2081

0.0135

0.0269

0.0189

0.1156

0.0097

16

0.4095

0.0184

0.0906

0.0223

0.1689

0.0117

24

0.6029

0.0184

0.0852

0.0261

0.2041

0.0126

32

0.7384

0.0112

0.1121

0.0321

0.2249

0.0240

40

0.8231

0.0216

0.1589

0.0251

0.2378

0.0325

48

0.8946

0.0404

0.1826

0.0499

0.3213

0.0514

247

Table C4 KALP_A5, A7, A17 and A19 dephasing △S/S0 in DPPC: DPPG (4:1), the data is
processed with 100Hz Gaussian line broadening and 3ppm integration width.
A5 with PC_d10
Dephasing

ΔS/So

error of

A5 with PC_d8
ΔS/So

error of ΔS/So

A5 with PC_d4
ΔS/So

ΔS/So

time/ms

error of
ΔS/So

2

0.0048

0.0099

0.0084

0.0068

0.0265

0.0058

8

0.0388

0.0070

0.0344

0.0064

0.0170

0.0089

16

0.0995

0.0116

0.0694

0.0093

0.0234

0.0115

24

0.1493

0.0094

0.1104

0.0108

0.0559

0.0116

32

0.1871

0.0126

0.1150

0.0138

0.0581

0.0209

40

0.1892

0.0092

0.1619

0.0144

0.0487

0.0131

48

0.2402

0.0136

0.1776

0.0171

0.0224

0.0188

A7 with PC_d10
Dephasing

ΔS/So

error of

A7 with PC_d8
ΔS/So

error of ΔS/So

A7 with PC_d4
ΔS/So

ΔS/So

time/ms

error of
ΔS/So

2

-0.0134

0.0058

0.0041

0.0113

0.0067

0.0114

8

0.0570

0.0072

0.0284

0.0083

0.0114

0.0098

16

0.1321

0.0061

0.0576

0.0117

0.0050

0.0086

24

0.1825

0.0073

0.0953

0.0099

0.0367

0.0107

32

0.2106

0.0080

0.1307

0.0106

0.0284

0.0110

40

0.2841

0.0100

0.1348

0.0117

0.0224

0.0158

48

0.3077

0.0149

0.1626

0.0154

0.0397

0.0148

248

Table C4 (cont’d)
A17 with PC_d10
Dephasing

ΔS/So

error of

A17 with PC_d8
ΔS/So

error of ΔS/So

A17 with PC_d4
ΔS/So

ΔS/So

time/ms

error of
ΔS/So

2

-0.0075

0.0154

-0.0018

0.0112

0.0181

0.0103

8

0.0547

0.0119

0.0329

0.0085

0.0171

0.0119

16

0.0990

0.0118

0.0658

0.0084

0.0487

0.0115

24

0.1650

0.0111

0.0681

0.0125

0.0482

0.0095

32

0.1987

0.0130

0.0974

0.0106

0.0414

0.0146

40

0.2207

0.0174

0.1114

0.0136

0.0585

0.0196

48

0.2818

0.0203

0.1453

0.0178

0.0681

0.0131

A19 with PC_d10
Dephasing

ΔS/So

error of

A19 with PC_d8
ΔS/So

error of ΔS/So

A19 with PC_d4
ΔS/So

ΔS/So

time/ms

error of
ΔS/So

2

0.0134

0.0121

0.0240

0.0249

0.0137

0.0200

8

0.0408

0.0204

0.0318

0.0179

0.0367

0.0217

16

0.0655

0.0175

0.0333

0.0187

0.0254

0.0248

24

0.0707

0.0270

0.0683

0.0146

0.0852

0.0225

32

0.1366

0.0356

0.0923

0.0244

0.0247

0.0418

40

0.1846

0.0420

0.1790

0.0463

0.0457

0.0293

48

0.2753

0.0380

0.1879

0.0265

0.0463

0.0399

249

Table C5 DMPC-d54 order parameter (SCD) in d54: DMPG (4:1) and d54: DMPG: CHOL (8:2:5)
without and with HFP, and the order parameter percentage change by HFP.
ΔSCD= SCD Lipid – SCD (Lipid+ HFP).
C

HFP +

+CHO

+Chol and

+ Chol

HFP

and HFP

(1:100)

(1:25)

ΔSCD (

ΔSCD (+

ΔSCD (

Chol and HFP

0.028

+ Chol
and HFP
(1:25))
-0.196

2

0.219

0.216

0.341

0.332

0.408

d54pg+
HFP
(1:25))
0.014

3

0.219

0.216

0.341

0.332

0.408

0.014

0.028

-0.196

4

0.219

0.216

0.341

0.332

0.408

0.014

0.028

-0.196

5

0.219

0.216

0.341

0.332

0.408

0.014

0.028

-0.196

6

0.204

0.198

0.317

0.332

0.408

0.029

-0.046

-0.286

7

0.186

0.175

0.297

0.301

0.368

0.059

-0.014

-0.240

8

0.172

0.161

0.261

0.257

0.326

0.064

0.016

-0.248

9

0.151

0.143

0.252

0.257

0.326

0.053

-0.019

-0.292

10

0.142

0.134

0.22

0.210

0.274

0.056

0.044

-0.247

11

0.129

0.12

0.201

0.195

0.254

0.070

0.028

-0.262

12

0.113

0.105

0.16

0.156

0.202

0.071

0.027

-0.264

13

0.094

0.088

0.127

0.124

0.064

0.021

14

0.028

0.025

0.047

0.044

0.107

0.070

position

d54pg

d54pg
(1:25)

L

250

0.055

(1:100))

-0.160

Table C6 DMPC-d54 T2 fitting data for echo, CD3 and CD2 for both without and with HFP at
different temperature. The membrane is d54: DMPG= 4:1 mol ratio.
21°C

time/
Îźs

echo intensity
(HFP: PL=
1:25)

CD3 peak
intensity
(HFP: PL=
1:25)

CD2 peak
intensity
(HFP: PL=
1:25)

691988.312

101

115802128

2135285.5

1128808.5

406958.406

201

71780360

1612765.875

736819.812

267603.156

303

49382440

1391027

506400.656

179660.125

401

36058856

1245675.875

345419.438

136698.203

501

27391492

1072072.375

252804.359

100800.125

601

21214256

892813.25

169144.359

76095.414

701

17113696

816511.875

127413.523

62065.859

801

13937489

745940.625

94968.844

47807

901

11411520

646773.188

72931.93

34504.895

1001

9378537

542748.812

57735.293

27966.346

1101

7740526

481930.781

39720.156

1201

6470418

438686.75

30024.465

1301

5560921

378197.5

22200.453

1401

4489041

315689.594

1501

3629461

272308.406

CD3 peak
time/Îźs

echo intensity
intensity

101

63057788

181

29523848

263

16641136

341

10358777

421

7093353

501

5218229

581

3733626

661

2830889

741

2301231

821

1702988

903

1537200

251

Table C6 (cont’d)
25°C
echo

CD2 peak
intensity

time/
Îźs

echo
intensity
(HFP: PL=
1:25)

CD3 peak
intensity
(HFP: PL=
1:25)

CD2 peak
intensity
(HFP: PL= 1:25)

2026801.2
5

1144891

103

105781664

1996271.87
5

1073387.875

1702830

942213.5

261

57158708

776697.438

423

35837136

621514.75

583

24722928

1185154.5

513685.281

743

17524884

790177.125

150071.406

1019752.5

408583.469

903

12994199

673512.312

97908.492

860716.5

340108.438

1063

9936582

565712.375

61229.551

245394.234

1223

7366124

440080.594

41009.828

201877.359

1383

5627005

362828.469

32989.188

142585.859

1543

4321859

301345.438

16505.637

102665.383

1703

3189982

224314.828

39720.156

79838.648

1861

2436088

181833.453

30024.465

56270.645

2023

1753249

148932.031

22200.453

2183

1220028

107132.828

2343

898252

79102.477

2503

683723

67685.086

CD3 peak

time/Îźs
intensity

intensity

10263878
123
4
303

89085456

483

76725280

663

64093508

843

53143176

1023

43038292

1203

34194696

1383

26996276

1563

20587472

1743

15670842

1923

11683965

2103

8540030

2283

6174512

2463

4473311

2643

3339245

2823

2453834

1607489.2
5
1303193.1
25

739432.68
8
599586.81
2
496478.71
9
396061.5
337245.56
2
289892.15
6
264372.31
2
251591.45
3
245778.14
1

252

1413117.62
5
1199985.37
5
1004988.31
2

636620.812
378250.75
234601.484

Table C6 (cont’d)
30°C
echo
intensity

101

echo
intensity
(HFP: PL=
1:25)
98796512

CD3 peak
intensity
(HFP: PL=
1:25)
2101625.75

261

54628436

1458327

CD2 peak
intensity
(HFP: PL=
1:25)
1009767.06
2
611281.125

421

35466788

1185919.5

378962.469

581

25012956

743

18347956

1006188.37 245063.312
5
804359.188 164715.953

901

13766222

649429.188 115702.234

1061

10600092

551011.062

79574.047

1221

8097601

458949.5

57071.07

1381

6122391

349215.094

36059.785

1541

4697748

286665

26214.693

1703

3529706

229169.281

22468.465

1863

2652671

176373.281

2023

1982592

134109.656

2181

1433475

106230.781

CD3 peak

time/Îźs
intensity

CD2 peak
intensity

time/Îź
s

1184524.
625
993302.8
75
820749.2
5
665734.2
5
542333.3
75
456520.1
56
342164.9
69
299301.4
38
220743.6
25
158240.8
28
139159.8
59
97379.00
8
78159.41
4
62806.78
1

123

97945168

303

85633272

483

74481904

663

62519828

843

52011936

1023

43123512

1203

34664904

2168760.7
5
1804549.7
5
1676268.2
5
1434212.6
25
1200631.7
5
1099090.3
75
904017.75

1383

27490000

740479.25

1563

21481714

1743

16368111

1923

12453107

2103

9271540

2283

6856360

2463

4939892

639008.62
5
515123.34
4
429882.12
5
355275.12
5
298497.09
4
277561.37
5

2643

3653884

2345

1034288

80250.523

2823

2611025

2501

620372

63816.102

3003

1986086

2661

320609

46936.75

253

Table C6 (cont’d)
37°C
echo
intensity

83

echo
intensity
(HFP: PL=
1:25)
23368092

CD3 peak
intensity
(HFP: PL=
1:25)
509551.438

CD2 peak
intensity
(HFP: PL=
1:25)
265969.75

303

10879426

321844.75

133845.438

523

6756645

254588.141

77485

743

4737914

213737.406

50404.617

963

3319797

153440.141

31857.168

1183

2391862

115981.594

23216.367

1403

1703089

96738.625

15364.001

1623

1180113

69017.914

9958.122

1843

914771

47562.926

7405.187

CD3 peak

time/Îźs
intensity

CD2 peak
intensity

time/Îź
s

83

23420632

323

18562896

563

15225284

803

11926090

1043

9223867

1283

7104173

1523

5176301

1763

3844378

2003

2718262

2243

1909120

614511.68
8
454312.15
6
405409.43
8
340249.28
1
256865.29
7
236136.96
9
164442.90
6
130532.31
2
108489.31
2
73970.656

2483

1256475

67963.414

2723

890600

57693.52

290392.1
88
218417.3
44
169984.7
5
131170.9
84
100246.2
81
76389.07
8
51373.67
6
38250.26
6
27651.39
5
26960.74
6
20812.71
9
12761.95

2963

634516

52292.93

8685.154

2063

38422.871

2283

25491.84

2503

16364.961

2723

14591.285

254

Table C7 DMPC-d54 T2 fitting data for echo, CD3 and CD2 for both without and with HFP at
different temperature. The membrane is d54: DMPG: CHOL= 8:2:5.
21°C

Îźs

intensity

121

6301407

72370.742

41591.008

121

4860291

CD3 peak
intensity
(HFP: PL=
1:25)
56084.289

283

5166467

65204.785

30029.061

281

2628209

43005.797

18491.996

443

4146610

55714.715

20775.986

441

1563425

33106.707

10292.763

603

3271153

48981.867

14351.056

601

1030302

27663.021

5372.837

763

2519588

42008.625

9594.907

761

718111

23221.93

3298.103

923

1955328

35551.242

6124.931

923

515240

18150.461

1781.744

1083

1526812

31399.668

3877.846

1083

377137

15149.034

1243

1181775

27149.332

1243

294901

12619.528

1403

912202

21955.949

1403

201307

10034.387

1563

685291

18529.879

1563

164014

8021.642

1723

528695

15252.553

1723

117155

6082.786

1883

386965

12173.619

1883

77524

4938.036

2043

297120

9968.1

2043

60258

3905.18

2203

195505

7988.121

2203

44170

3327.426

2363

141331

7011.406

2363

42345

2660.875

2523

87149

6028.201

2523

30900

2019.662

2683

50449

5629.337

2681

18685

1762.267

time/

echo

CD3 peak

CD2 peak
intensity

intensity

time/Îź
s

echo intensity
(HFP: PL= 1:25)

255

CD2 peak
intensity
(HFP: PL=
1:25)
33658.496

Table C7 (cont’d)
25°C
echo

123

echo
intensity
(HFP: PL=
1:25)
19930866

CD3 peak
intensity
(HFP: PL=
1:25)
231827.156

35725.578

243

12540461

189542.797

84053.414

26925.637

363

8251461

153984.156

53628.488

19613.178

483

5775383

132002.016

33073.926

13957.897

603

4217832

115099.695

21174.137

9352.472

723

3192564

97583.891

13608.686

7540.693

843

2539507

87238.836

8325.301

5119.122

963

1964522

74099.109

5900.491

3069.832

1083

1614543

65692.82

3342.039

1203

1270232

54134.52

2740.498

1323

1047440

48528.965

1922.136

1443

797007

38951.594

1563

676338

34940.828

1683

542335

28156.535

1803

408628

24471.822

CD3 peak

time/Îźs
intensity

intensity

83

6347105

243

5532157

403

4780424

563

4073641

723

3370140

883

2779965

1043

2231754

1203

1778664

1363

1414178

1523

1084809

80792.64
8
73942.88
3
66615.60
2
60484.26
6
52796.31
2
46467.38
3
40227.84
8
34618.05
5
29852.29
5
24698.77

1683

829748

1843

632355

2003

455647

2163

327381

2323

226087

20795.04
3
16854.39
5
14082.30
1
11327.17
3
9006.731

CD2 peak
intensity

time/
Îźs

44365.863

256

CD2 peak
intensity
(HFP: PL= 1:25)

133459.281

Table C7 (cont’d)
30°C
echo
intensity

85

echo
intensity
(HFP: PL=
1:25)
100128624

CD3 peak
intensity
(HFP: PL=
1:25)
1097857.25

CD2 peak
intensity
(HFP: PL=
1:25)
678205.188

203

69532712

897913.438

427571.5

323

44843536

751978.562 247767.672

443

30772500

645994.062 158576.359

563

22570992

565815.438

94699.414

683

16819664

481651.812

62608.941

803

13050568

428669.906

41452.617

923

10396601

361066.094

28641.406

1041

8261454

310768.125

16430.283

1163

6420406

269916.812

8570.573

CD3 peak

time/Îźs
intensity

364004.90
6
342987.31
2
307423.25

CD2 peak
intensity

time/Îź
s

216588.0
31
174189.4
06
136185.6
09
100793.0
7
71622.75
8
49457.90
6
35640.70
3
22334.29
9
15393.72
1
9897.642

123

28706428

303

26174888

483

22768650

663

19179180

843

15775435

1023

12660307

1203

9940572

1383

7662337

1563

5786863

1743

4220267

204002.78
1
180911.46
9
146586.03
1
121667.30
5
97285.578

1923

3012119

76122.398

1283

5276130

222875.125

2103

2140265

62034.441

1403

4230648

183808.906

2283

1424155

48343.043

1523

3593334

159582.594

2463

907426

39500.703

1643

2832819

144842.5

2643

545976

36462.152

1763

2171168

118713.008

2823

317050

32676.654

1885

1880407

100648.93

273475.96
9
237977.25

257

Table C7 (cont’d)
37°C
echo
intensity

83

echo
intensity
(HFP: PL=
1:25)
13901748

CD3 peak
intensity
(HFP: PL=
1:25)
220394.219

CD2 peak
intensity
(HFP: PL=
1:25)
123708.883

203

8110308

161628.469

71819.789

323

5592441

155021.781

43283.867

443

3849416

121720.812

29284.15

563

3013031

112230.75

20009.635

72356.75

683

2273741

91241.977

11001.198

52255.43
8
37383.78
1
25970.53
7
17748.80
1
10881.54
6

803

1839586

75849.367

8655.914

923

1452028

68953.648

5539.538

1043

1233011

57797.215

1163

964646

52754.32

1283

821070

42714.387

CD3 peak

time/Îźs
intensity

CD2 peak
intensity

time/Îź
s

234696.2
66
203342.5
62
166953.3
91
128196.7
5
98458.82

123

27683546

303

26203104

483

23483512

663

20314432

843

17124336

1023

14080315

1203

11290925

1383

8922532

1563

6848299

1743

5189815

1923

3785671

413753.31
2
392187.21
9
354733.34
4
323332.18
8
281276.62
5
247663.79
7
214235.54
7
181751.64
1
149601.31
2
124471.78
1
96539.695

2103

2695042

78749.055

1403

656531

33184.688

2283

1791103

59661.094

1523

577684

32935.875

2461

1191651

51195.496

1643

427749

26385.656

2641

720066

42218.711

1763

395406

22643.736

1883

355069

20892.908

2003

288613

18788.369

258

Table C8 Chol_d6 T2 fitting data for echo intensity for both without and with HFP at 37 °C in
membrane of DMPC: DMPG: Chol_d6= 8:2:5 and POPC: POPG: Chol_d6= 8:2:5.
DMPCPGChol_d6

POPCPGChol_d6

time/Îźs

Echo intensity

time/Îźs

Echo intensity

133

18017108

Echo intensity
with HFP
17682278

63

17700144

Echo intensity
with HFP
58113512

183

15503495

14659419

93

15126436

51097308

233

13510561

11705473

123

13886196

44113644

283

11729365

9425289

153

12606192

38606532

333

10417644

7852361

183

12187828

34709568

383

9214424

6516831

213

11201064

30500508

433

8024377

5504747

243

9895180

27207600

483

6882644

4302516

273

8902792

24344344

533

6071274

3618283

303

8903644

21966172

583

5240645

2983362

333

7972644

19520700

633

4628060

2715867

363

6719492

18031672

683

4085062

2306738

393

6363652

15968128

423

6591000

14244916

453

6237668

12931472

483

5077704

11633468

513

4801908

10613544

543

4200040

8870172

573

4286332

8297860

603

4244240

6966928

633

3593688

663

2967528

259

Table C9 Chol_d7 T2 fitting data with ln(CD3 peak intensity) for both without and with HFP at
different temperatures. Peak intensity is short as PI.
-50°C
tim

5°C

ln(CD3

e/Îźs

PI)

75

11.290

95

10.989

115

10.695

135

10.538

155

10.371

175

10.135

195

10.003

215

9.653

235

9.509

255

9.409

275

9.059

295

8.770

25°C
With

ln(CD3

With

With
HFP

time/
Îźs

119

10.899

77

11.959

87

12.631

79

12.016

81

12.680

179

10.714

197

11.852

287

12.443

199

11.886

281

12.549

239

10.379

317

11.739

487

12.254

319

11.788

481

12.381

299

9.986

437

11.654

687

12.052

439

11.659

681

12.224

359

9.692

557

11.584

887

11.815

559

11.572

881

12.033

419

9.355

677

11.500

1087

11.530

679

11.481

1081

11.809

479

8.782

797

11.401

1287

11.277

799

11.345

1281

11.557

539

8.475

917

11.238

1487

11.075

919

11.239

1481

11.289

1037

11.049

1039

11.085

1681

11.067

1157

10.958

1159

10.982

1881

10.903

1277

10.781

1279

10.832

2081

10.799

1397

10.679

1399

10.670

1517

10.565

1519

10.582

1637

10.532

1639

10.449

1757

10.390

1877

10.312

PI)

time/
Îźs

ln(CD3

tim
e/Îźs

260

HFP

time/
Îźs

PI)

time/
Îźs

HFP

Table C9 (cont’d)
37°C
time/

45°C

ln(CD3

With

ln(CD3

Îźs

PI)

time/
Îźs

With
HFP

time/
Îźs

time/
Îźs

79

12.555

83

12.694

79

12.545

83

12.673

199

12.480

283

12.564

199

12.465

263

12.581

319

12.373

483

12.411

319

12.374

443

12.454

439

12.268

683

12.226

439

12.270

623

12.307

559

12.175

883

12.030

559

12.163

803

12.146

679

12.036

1083

11.803

679

12.039

983

11.982

799

11.920

1283

11.534

799

11.917

1163

11.788

919

11.810

1483

11.250

919

11.801

1343

11.597

1039

11.658

1683

11.060

1039

11.646

1523

11.362

1159

11.514

1883

10.851

1159

11.534

1703

11.183

1279

11.396

2083

10.723

1279

11.373

1883

10.963

1399

11.231

1399

11.186

2063

10.797

1519

11.088

1519

11.058

1639

10.948

1639

10.844

PI)

261

HFP

Table C10 Chol_d7 T1 fitting data (CD3 peak intensity) at different temperatures for both
membranes without and with HFP. Peak intensity is short as PI.
-50°C

5°C

CD3 PI
tau/ms

(300ppm
integration)

CD3

With HFP
(300ppm
integration)

tau/ms

PI

With
tau/ms

HFP

1

-1512.9

-3312.4

0.1

-434881.6

0.5

-945151.7

21

1116.0

3315.2

20.1

-169887.3

20.5

-369262.8

41

1888.2

5154.4

40.1

22743.7

40.5

46525.4

61

2184.2

5646.2

60.1

156616.8

60.5

345054.3

81

2307.9

5827.7

80.1

253943.0

80.5

560756.6

101

2340.9

5864.5

100.1

327334.4

100.5

718595.9

121

2347.8

5925.1

120.1

381858.5

120.5

817169.9

141

2329.7

5921.9

140.1

419204.4

140.5

913808.7

161

2358.0

5892.5

160.1

445182.2

160.5

960305.4

181

2337.8

5913.6

180.1

468484.2

180.5

1012877.5

201

2416.7

5892.3

200.1

484320.9

200.5

1039870.8

221

2480.8

5884.3

220.1

491316.6

220.5

1064254.9

241

2481.0

5924.8

240.1

500078.8

240.5

1072906.4

261

2434.8

5984.2

260.1

504772.7

260.5

1096844.9

281

2322.9

5917.3

280.1

509381.3

280.5

1111076.1

301

2311.4

5836.3

300.1

517757.5

300.5

1112866.4

262

Table C10 (cont’d)

25 °C

tau/ms

CD3 PI

tau/ms

With HFP

0.1

-442409.8

1

-893296.7

20.1

-255557.3

21

-507085.4

40.1

-109295.0

41

-219779.4

60.1

14820.2

61

19034.4

80.1

107549.8

81

204182.4

100.1

185117.4

101

354759.8

120.1

254591.7

121

481844.9

140.1

302620.8

141

583910.9

160.1

343963.7

161

664930.5

180.1

381466.1

181

732717.7

200.1

412364.4

201

786209.6

220.1

434873.4

221

832482.4

240.1

455804.0

241

869810.3

260.1

466611.2

261

899894.5

280.1

480669.6

281

921989.6

300.1

495197.5

301

943300.2

263

Table C10 (cont’d)
37°C

45°C

ln(CD3
time/Îźs
PI)

With

ln(CD3
time/Îźs

With
HFP

time/Îźs

PI)

time/Îźs

HFP

0.1

-932827.8

0.5

921270.1

0.1

-932013.2

1

30.1

-492471.4

20.5

30.1

-546767.1

31

60.1

-167003.8

40.5

60.1

-255563.2

61

-247503.8

90.1

90662.7

60.5

90.1

-23192.2

91

27236.7

120.1

289546.3

80.5

681174.2
394185.3
150763.0
57369.1

1075404.
5
-590307.9

120.1

174972.9

121

245609.6

150.1

450398.2

100.5

229344.6

150.1

334353.8

151

425383.2

180.1

570572.6

120.5

379687.8

180.1

458165.5

181

566971.8

210.1

665113.3

140.5

502600.3

210.1

564324.1

211

684175.6

240.1

742972.4

160.5

608695.4

240.1

646277.6

241

779078.9

270.1

802038.3

180.5

699460.5

270.1

715041.6

271

856705.8

300.1

848178.4

200.5

775730.8

300.1

771683.9

301

919107.6

330.1

887415.9

220.5

840829.8

330.1

818654.8

331

967508.3

360.1

916573.1

240.5

895715.1

360.1

854716.3

361

390.1

939465.1

260.5

942620.4

390.1

884128.0

391

420.1

959147.1

280.5

984891.4

420.1

907751.1

421

450.1

971403.8

300.5

450.1

926919.6

451

480.1

981659.6

320.5

480.1

946317.3

481

510.1

994294.1

340.5

1014429.
5
1044093.
1
1069516.
8

1013582.
9
1041437.
9
1069570.
9
1091981.
4
1108735.
9
1126726.
9

511

264

REFERENCES

265

REFERENCES

1.

Freed, E.O., et al., A MUTATION IN THE HUMAN-IMMUNODEFICIENCY-VIRUS
TYPE-1 TRANSMEMBRANE GLYCOPROTEIN-GP41 DOMINANTLY INTERFERES
WITH FUSION AND INFECTIVITY. Proceedings of the National Academy of Sciences
of the United States of America, 1992. 89(1): p. 70-74.

2.

Banerjee, K. and D.P. Weliky, Folded Monomers and Hexamers of the Ectodomain of the
HIV gp41 Membrane Fusion Protein: Potential Roles in Fusion and Synergy Between the
Fusion Peptide, Hairpin, and Membrane-Proximal External Region. Biochemistry, 2014.
53(46): p. 7184-7198.

3.

Sackett, K., et al., Hairpin Folding of HIV gp41 Abrogates Lipid Mixing Function at
Physiologic pH and Inhibits Lipid Mixing by Exposed gp41 Constructs. Biochemistry,
2009. 48(12): p. 2714-2722.

4.

Tan, K.M., et al., Atomic structure of a thermostable subdomain of HIV-1 gp41.
Proceedings of the National Academy of Sciences of the United States of America, 1997.
94(23): p. 12303-12308.

5.

Yang, Z.N., et al., The crystal structure of the SIV gp41 ectodomain at 1.47 angstrom
resolution. Journal of Structural Biology, 1999. 126(2): p. 131-144.

6.

Ratnayake, P.U., et al., Full-length trimeric influenza virus hemagglutinin II membrane
fusion protein and shorter constructs lacking the fusion peptide or transmembrane
domain: Hyperthermostability of the full-length protein and the soluble ectodomain and
fusion peptide make significant contributions to fusion of membrane vesicles. Protein
Expression and Purification, 2016. 117: p. 6-16.

7.

Greenfield, N., CITATION CLASSIC - COMPUTED CIRCULAR-DICHROISM
SPECTRA FOR THE EVALUATION OF PROTEIN CONFORMATION. Current
Contents/Life Sciences, 1982(26): p. 28-28.

8.

Gabrys, C.M., et al., Solid-State Nuclear Magnetic Resonance Measurements of HIV
Fusion Peptide (CO)-C-13 to Lipid P-31 Proximities Support Similar Partially Inserted
Membrane Locations of the alpha Helical and beta Sheet Peptide Structures. Journal of
Physical Chemistry A, 2013. 117(39): p. 9848-9859.

266