v_—_

LIBRARY

Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

rJUL ; 2 4 2005

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

moo wmmu

_.

BUSH
POTENT]

EXAMINATION OF THE ROLE OF MEMBRANE DYNAMICS AND ORDER IN
POTENTIATING THE PHYSIOLOGICAL EFFECTS OF LIPOPOLYSACCHARIDES

By

Bernard Dale Hummel

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Biochemistry

1998

REMIX X
POTEWHT

Lipo
Gmm-nega:
in animals.
immune 5‘
septicemia

Panicularl

Sm,”

dim

ABSTRACT

EXAMINATION OF THE ROLE OF MEMBRANE DYNAMICS AND ORDER IN
POTENTIATING THE PHYSIOLOGICAL EFFECTS OF LIPOPOLYSACCHARIDES

By

Bernard Dale Hummel

Lipopolysaccharides (LPS) are complex glycolipids found on the surface of
Gram-negative bacteria and are capable of eliciting a wide array of biological responses
in animals. These responses are primarily due to the interaction of LPS with cells of the
immune system and are responsible for the deleterious effects of Gram-negative
septicemia. The studies presented here investigate the molecular mechanisms,
particularly at the membrane level, by which LPS activates immune cells.

A new understanding of the structural features of LPS required to elicit a
biological response has been obtained. Analysis of rhizobial LPS showed that they have
a drastically different lipid structure from enterobacterial LPS. These differences were in
the headgroup and fatty acyl structures as well as its charge functionalities. The
biological activities of rhizobial LPS were also distinct. They gave a strong response in
the Limulus amebocyte lysate and B-lymphocyte mitogen assays but only elicited a weak
response in the mouse toxicity, pyrogen and tumor necrosis factor production assays.
These results indicate that some physicochemical property of the lipid A portion of LPS,
and not the individual structural functionalities per se, is critical for its biological activity.

The effects of LPS on the properties of membranes and its role in the activation of
immune cells has been investigated. I demonstrate that enterobacterial LPS has a

dramatic effect on the order and dynamics of membranes. My results show that LPS

induces lateral

 

the acyl than
relationship be‘
biological am

Sl'nlhc’.
mechanism of
attii'izies of L
lipid lllhlbllor
mount of ll.
QRgaics ir
Perturbatiom

that can offs

induces lateral separation of model membrane lipids into distinct domains and reduces
the acyl chain mobility of the membrane. I also demonstrate that there is not a linear
relationship between membrane acyl chain mobility changes induced by LPS and its
biological activity.

Synthetic lipids that were designed to interact with lipid A were used to probe its
mechanism of action. The synthetic lipids were effective inhibitors of the biological
activities of LPS. The membrane dynamics and order changes induced by the synthetic
lipid inhibitors were also investigated. My results show that the inhibitors increase the
amount of fluid lipid species present in the membranes and disrupt the rigid LPS/lipid
aggregates in the membrane. These lipid inhibitors demonstrate that the membrane
perturbations induced by LPS are critical to its mechanism of action and that molecules

that can offset this perturbation can inhibit its ability to elicit a physiological response.

To my parents, Don and Joan Hummel

iv

1 mil“.
patience ii iii”.
from a more
findings. CVt‘

I {list
Mtflromy.
These mm
”1333ng .‘x

I Vi ,1;
Carol. Chucl
Jlt. Jim. Jim

“‘18- and Y-

ACKNOWLEDGEMENTS

I want to thank my thesis adviser, Dr. Rawle Hollingsworth for his guidance and
patience with me. I would also like to thank Rawle for teaching me how to see problems
from a more physical perspective and for teaching me to always question scientific
findings, even those considered to be ‘facts’.

I also want to thank my committee members, Dr. Fraker, Dr. Hausinger, Dr.
McGroarty, Dr. Strasburg, and Dr. Chaudry (who left MSU before I was completed).
These individuals truly lived up to their title as a ‘guidance committee’ and gave me
many good suggestions over the years.

I want to thank all of the past and present members of the Hollingsworth group:
Carol, Chuck, Debby, Du, Eunmi, Gabriela, Gang, Guangfei, Guijun, Hussen, Jeongrim,
Jie, Jim, Jimin, JJ, Jung, Kim, Lakshmi, Luc, Maria, Rob, Steve, Vladimir, Wayne, Yin,
Ying, and Yuanda. Their friendship made the Hollingsworth group a great place to work.

I am indebted to the mass spectrometry facility (Bev Chamberlin, Dr. Huang, and
Dr. Gage) and the NMR facility (Kermit Johnston and Dr. Long Le) for all of their help
with my experiments.

I also want to thank my parents, Don and Joan Hummel, and the rest of my family
for all of their love and support. I would also like to thank Mark, Laurie, Dani, Kati, Luc,
Odette, Sophie, Rob, Kerry and Alex for their friendship. They made Lansing a fun place

to live and a great place for my wife and I to start our family.

Finali}
They stuck b)
Without their

passibic.

Finally, I would like to thank my wife, Jean, and two sons, Zachary and Adam.
They stuck by my side at all times and put up with my unusual hours for many years.
Without their sacrifices, support, and unconditional love, this dissertation would not be
possible.
Thank you all

Ben

vi

list of Table
List of Figm
List of Abbi

Chapter 1
Iain
intro
Strut
Men
Men
Phys
Lipt
Siec
Con

Chapter 2
The
Lip
lntr
Mai
R65
Dix

Chapter 3
Li;
Mt
Int:

TABLE OF CONTENTS

List of Tables ................................................................................ ix
List of Figures ............................................................................... x
List of Abbreviations ....................................................................... xv
Chapter 1
Introduction to Biological Membranes and Lipopolysaccharides 1
Introduction ......................................................................... 2
Structure and Phase Behavior of Membrane Lipids ........................... S
Membrane Dynamics .............................................................. 10
Membrane Order .................................................................... l4
Physiological Relevance of Membrane Dynamics and Order ................. 18
Lipopolysaccharides ................................................................ 25
Mechanisms of Lipopolysaccharide Induced Biological Activities 32
Conclusion ........................................................................... 40
Chapter 2
The Relationship Between Structure and Biological Activities of
Lipopolysaccharides: Lessons from Rhizobial LPS ........................ 42
Introduction ......................................................................... 43
Materials and Methods ............................................................ 46
Results ............................................................................... 51
Discussion ........................................................................... 65
Chapter 3
Lipopolysaccharide Induced Changes in the Order and Dynamics of
Membranes ......................................................................... 72
Introduction .......................................................................... 73
Materials and Methods ............................................................. 76
Results and Discussion ............................................................. 79
Conclusions .......................................................................... 101
Chapter 4
Probing the Mechanism of Lipopolysaccharide Action with Synthetic
Lipids ................................................................................. 103
Introduction .......................................................................... 104

vii

Mute
Rest
C one

Chapter 5
Sum

Appendix
5an
Iain
Dish
Syn:
C ha
and
C on

List of Ref

Materials and Methods ............................................................. 108

Results and Discussion ............................................................. 111

Conclusions .......................................................................... 1 30
Chapter 5

Summary and Perspective ...................................................... 132
Appendix

Synthesis and Characterization of Five Synthetic Lipids ................. 145

Introduction .......................................................................... 146

Materials and Methods ............................................................ 146

Synthesis and Purification of EI—3, E14 and EI—S .............................. 147

Characterization of the Physical Properties of EI—l, EI—2, EI—3, EI-4

and EI—5 .............................................................................. 157

Conclusions .......................................................................... 168
List of References .......................................................................... 169

viii

Table 1.1: So

Table 2.1: S)‘
R. Infillii ANT

Table 2.2; ['S
R Irit‘olii AXl

Table 3.1: P4:

Table 3.2: T1
membranes.

Table 4.1: Per
01 the synthet

Table 4:; Bi
”Ethylene pn‘

LIST OF TABLES

Table 1.1: Some typical biological activities elicited by lipopolysaccharides. 27
Table 2.1: Systemic toxicity of S. typhimurium, R. meliloti 2011, and

R. trifolii ANU843 LPS. ..................................................................... 64
Table 2.2: USP Pyrogen assay of S. typhimurium, R. meliloti 2011, and

R. trifolii ANU 843 LPS. ..................................................................... 66
Table 3.1: Parameters of the deconvoluted wide-angle reflections. .................. 90

Table 3.2: T1 values for the methylene protons in PC and LPS-PC
membranes. ................................................................................... 99

Table 4.1: Percent change in the methylene proton signals upon addition
of the synthetic lipids. ....................................................................... 125

Table 4.2: Effects of the synthetic lipids on the T1 relaxation values for the
methylene proton signals in PC-LPS membranes. ...................................... 129

ix

LIST OF FIGURES

Figure 1.1: The fluid mosaic model of biological membranes in which the

proteins and lipids form a two-dimensional viscous fluid. The proteins

and lipids are free to move randomly in the plane of the membrane.

Adapted from reference 4. ................................................................. 4

Figure 1.2: Some common membrane lipids. ........................................... 6

Figure 1.3: Relationship between molecular geometry of lipids and the
phases they adopt. Taken from reference 7. ............................................ 9

Figure 1.4: A schematic of the phase behavior of
dipalmitoylphosphatidylcholine as a function of temperature. Modified
from reference 8. ............................................................................ 10

Figure 1.5: Rates of membrane motions and the techniques that are
sensitive to these motions. Taken from reference 9. ................................... 12

Figure 1.6: The coupling between membrane dynamics and order, the
function of the membrane, and the physiology of the cell. ............................ 24

Figure 1.7: Outer envelope of Gram-negative bacteria.
Lipopolysaccharides are located on the outer leaflet of the outer

membrane. .................................................................................... 26
Figure 1.8: Schematic diagram of the structure of lipopolysaccharides. ............ 30
Figure 1.9: Structure of Escherichia coli lipid A. ........................................ 31
Figure 1.10: Lipid A structure from Rhodobacter sphaeroides. ....................... 35

Figure 1.11: A schematic diagram of the two-step model proposed by
Jacobs and Price ( 121). The first step is the electrostatic interaction of
LPS with the cellular membrane. The second step is the insertion (fusion)
of the lipid A region of LPS into the membrane. The insertion of LPS into
the membrane is proposed to be the critical event leading to the activation
of the immune cell. The LPS aggregate is (shown as a monolayer) likely
to be a bilayer of LPS or LPS, lipids, and proteins depending on the source

and purity of the LPS used. ................................................................... 39
Figure 2.1: Structure of R. trifolii lipid A. From reference 132. ....................... 45

Figure 2.2: Purification of R. meliloti 2011 LPS by Sepharose 4B gel
filtration chromatography. ................................................................... 52

Figure 2.3: Purification of 358-1abelled R. meliloti 2011 LPS by gel
filtration chromatography. ................................................................... 53

Figure 2.4: Gas chromatograph of the methylated and peracetylated
carbohydrates from R. meliloti 2011 lipid A. The carbohydrates are
A. two isomers of glucose, B. heptose, and C. glucosamine. ........................... 56

Figure 2.5: Gas chromatograph of the fatty acid methyl esters from
R. meliloti 2011 lipid A. The peaks A-F correspond to 3-OH 14:0,
16:0, 3-OH 16:0, 18:0, 3-OH 18:0, and 27-OH 28:0, respectively. ................... 58

Figure 2.6: II-I-NMR spectrum of R. meliloti 2011 lipid A. Spectral
assignments are given in the text. ........................................................... 60

Figure 2.7: Tumor Necrosis Factor (TNF-a) assay of S. typhimurium
LPS, R. meliloti 2011 LPS, and R. trifolii ANU843 LPS. ............................... 63

Figure 2.8: Limulus Amebocyte Lysate (LAL) assay of S. typhimurium
LPS, R. meliloti 2011 LPS, and R. trifolii ANU843 LPS. Activity is
given as Endotoxin Units (EU) per ng LPS. ............................................... 67

Figure 3.1: Fluorescence anisotropy of DPH-PC in phophatidylcholine
membranes (0.05 mg/ml) as a function of LPS concentration. ......................... 81

Figure 3.2: X-ray diffraction pattern of phosphatidylcholine in D-PBS.
Arrows indicate low-angle reflections. .................................................... 85

Figure 3.3: X-ray diffraction pattern of phosphatidylcholine (67%) -
S. typhimurium LPS (33%) in D-PBS. Arrows indicate low-angle
reflections. ...................................................................................... 87

Figure 3.4: Wide-angle region of the x-ray diffraction patterns from
phosphatidylcholine (A) and 2:1 phosphatidylcholinezS. typhimurium

LPS (B). Solid lines are the deconvoluted signals and the sum of the

deconvoluted signals. ........................................................................ 89

Figure 3.5: ‘H-NMR spectra of s. typhimurium LPS (37°C) in D-PBS
with 0% (A), 25% (B), and 50% (C) of deuterated methanol added.
Note the large change in the methylene and methyl protons signals

xi

(0.8.1.5 pp:

Figure 3.6:
with 0‘} {A
h‘phimuriur

proton sign.

EnnBJ:

mflffir
approxinnut
non-hnearl

Egm4l:

figue42:
5. h'phimiu
idlidues;
5- Uphimiu

Emmi)

LPS hate C

Emmit
lelUIBS 0]
Exclusion.
hm ShOW 1

Figure 4.5;
LPS (11:! 1‘2

fiflmtg
m€mbriines
1:] 131. 5:]

(0.8-1.5 ppm) when methanol is added. ................................................... 92

Figure 3.6: lH-NMR spectra of phosphatidylcholine in D-PBS (37°C)

with 0% (A), 20% (B), 33% (C), 43% (D) and 50% (E) Salmonella

typhimurium LPS (by weight) added. Note the splitting of the methylene

proton signal (1.29 ppm) as LPS is added. ................................................ 94

Figure 3.7: lH-NMR T1 recovery plots for the methylene protons of PC

and PC-S. typhimurium LPS membranes at 37°C. The straight lines are

approximate fits to the early time points and are used to illustrate the

non-linear behavior of the data when LPS is added. ..................................... 98

Figure 4.1: Structures of the five synthetic lipids. ....................................... 107

Figure 4.2: Effect of the synthetic lipids on the activity of 0.1 ng/ml

S. typhimurium LPS in the Limulus amebocyte lysate (LAL) assay.

All values are reported as the percent activity relative to 0.1 ng/ml

S. typhimurium LPS. ......................................................................... 113

Figure 4.3: Tumor necrosis factor-0t (TNF-a) assay. Samples that contain
LPS have 0.1 ng/ml S. typhimurium LPS. ................................................ 1 14

Figure 4.4: Effects of S. typhimurium LPS and synthetic lipid-LPS

mixtures on RAW264.7 cell viability as judged by trypan blue

exclusion. The results are the average of duplicate counts and the error

bars show the range of the results. ........................................................ 116

Figure 4.5: Fluorescence anisotropy of DPH-PC in PC:S. typhimurium
LPS (1:1 ratio by weight) as a function of synthetic lipid concentration. ........... 1 18

Figure 4.6: lH-NMR spectra of PC:Salmonella typhimurium LPS
membranes (1 :1 ratio by weight) with a molar ratio of EI-1:LPS of 0 (A),
1:1 (B), 5:1 (C), and 10:1 (D). ............................................................ 120

Figure 4.7: ‘H-NMR spectra of PC:Salmonella typhimurium LPS
membranes (1:1 ratio by weight) with a molar ratio of EI-2:LPS of 0 (A),
1:1 (B), 5:1 (C), and 10:1 (D). ............................................................ 121

Figure 4.8: 'H-NMR spectra of PC:Salmonella typhimurium LPS
membranes (1:1 ratio by weight) with a molar ratio of EI-3:LPS of 0 (A),
1:1 (B), 5:1 (C), and 10:1 (D). ............................................................ 122

Figure 4.9: 'H-NMR spectra of PC:Salmonella typhimurium LPS

membranes (1:1 ratio by weight) with a molar ratio of EI—41LPS of 0 (A),
1:1 (B), 5:1 (C), and 10:1 (D). ............................................................ 123

xii

 

Emm4l
membrane
lJELS

Figure 4.1
hphinzuri
added. A
function c
thHA
[on enng 1
transition

figure A.

figmAa
mflH3l
meflmei

Figure A;
Spectrum

indicate St
in the text

Figure .-\.-‘

Fié’ure A:
SWCU'Um l
lDdlCate y
in The TCXI

figmAi

HflmAf
ipeermm l
lfldicate St
"1 the [fit

lipids.

FISUre Ar
‘ €31me

FlEUre A 1
A gSLTQm‘

Figure 4.10: 1H-NMR spectra of PC:Salmonella typhimurium LPS
membranes (1:1 ratio by weight) with a molar ratio of EI-5:LPS of 0 (A),
1:1 (B), 5:1 (C), and 10:1 (D). ............................................................ 124

Figure 4.11: Fluorescence anisotropy of tPnA in DPPC and DPPC-S.

typhimurium LPS (1:1 by weight) membranes with and without EI—2

added. A. Anisotropy of DPPC and DPPC-LPS membranes as a

function of EI-2 concentration at 37°C. B. Anisotropy vs. temperature.

In B., EI—2 is used at the highest concentration tested in A. Notice the

lowering of the transition temperature and the broadening of the phase

transition induced by EI-2. ................................................................. 127

Figure A.1: Synthetic scheme for EI-3. .................................................. 148

Figure A.2: Thin layer chromatography of EI-S (lane A), EI-4 (lane B),
and EI-3 (lane C). Solvent system was 10:1 chloroforrnzmethanol and
the plate was stained with phosphomolybdic acid (10%). .............................. 150

Figure A.3: Proton NMR spectrum (A) and positive-ion electrospray mass

spectrum (B) of 131-3. NMR spectrum assignments are shown. Asterisks

indicate solvent signals. Assignments of the mass spectrum are described

in the text. .................................................................................... 151

Figure A.4: Synthetic scheme for EI—4. .................................................. 152

Figure A.5: Proton NMR spectrum (A) and positive-ion electrospray mass

spectrum (B) of EI-4. NMR spectrum assignments are shown. Asterisks

indicate solvent signals. Assignments of the mass spectrum are described

in the text. .................................................................................... 153

Figure A.6: Synthetic scheme for EI-S. .................................................. 155
Figure A.7: Proton NMR spectrum (A) and positive-ion electrospray mass
spectrum (B) of EI—5. NMR spectrum assignments are shown. Asterisks

indicate solvent signals. Assignments of the mass spectrum are described
in the text. .................................................................................... 156

Figure A.8: Differential scanning calorimetry scans of the five synthetic
lipids. .......................................................................................... 158

Figure A.9: X-ray diffraction profile of EI-3. Reflections are labeled in
Angstroms. .................................................................................... 161

Figure A. 10: X-ray diffraction profile of EI-4. Reflections are labeled in
Angstroms. .................................................................................... 162

xiii

Figure A.11: X-ray diffraction profile of EI-S. Reflections are labeled in
Angstroms. .................................................................................... 163

Figure A. 12: X-ray diffraction profile of BL]. Reflections are labeled in
Angstroms. .................................................................................... 164

Figure A. 13: X-ray diffraction profile of EI-2. Reflections are labeled in
Angstroms. .................................................................................... 165

Figure A. 14: T1 values for the methylene protons of the five synthetic
lipids as a function of temperature. ....................................................... 166

xiv

CPM
CSF
DblEbl
DPBS

DPH-PC

DPPC
DSC
E. coil
El
ELISA
EL'

FRAP

LBp

CPM

CSF

DMEM

D-PBS

DPH-PC

DPPC

DSC

E. coli

EI

ELISA

EU

GC

GC-MS

LAL

LBP

List of Abbreviations

counts per minute

colony stimulating factor

Dulbecco’s modified eagles medium
Dulbecco’s phosphate buffered saline

2-(3-(diphenylhexatrienyl)propanoyl)— 1 -hexadecanoyl-sn-
glycero-3-phosphocholine

1 ,2—dipa1mitoyl-sn-glycero-3-phosphocholine
differential scanning calorimetry
Escherichia coli

endotoxin inhibitor

enzyme linked immunosorbent assay
endotoxin unit

fluorescence recovery after photobleaching
gas chromatography

gas chromatography-mass spectrometry
interferon

interleukin

3-deoxy-D—manno—octulosonic acid
lamellar liquid crystalline phase

lamellar gel phase

Limulus amebocyte lysate

lipopolysaccharide binding protein

XV

LL'V
NMR

P3

R. mez’h’uti
R. Irlltllii
S. nphtmnrttt-
Tl
TNT

TPnh

 

LPS

LUV

NMR

PB’

R. meliloti

R. trifolii

S. typhimurium
T1

TNF

TPnA

lipopolysaccharides

large unilamellar vesicles
nuclear magnetic resonance
ripple phase

Rhizobium meliloti
Rhizobium trifolii
Salmonella typhimurium
spin-lattice relaxation time
tumor necrosis factor

trans-parinaric acid

xvi

Chapter 1

Introduction to Biological Membranes and Lipopolysaccharides

 

 

 

 

 

 

 

 

lntmdu

in pan.
control"
step in

Ci’iltsf

inzplu
it is e
water.
hydro
rtgiur

is call

(:2;

(1’4
*1
(I:

Struct-

lipids
[Ultim-
Sun-3C1
[liken
dlscut,
Sillmm

[newbie

Introduction

Biological membranes are one of the critical components of cells. They function,
in part, to compartmentalize cells and allow the processes of the cell to be conducted in a
controlled environment. The isolation of prebiotic reactions may have been an essential
step in the evolution of life itself. Thus it is likely that membranes were features of the
earliest forms of life on earth (1).

Membranes have a common structural feature; they are all composed of
amphiphilic molecules (having both hydrophobic and hydrophilic moieties) called lipids.
It is energetically unfavorable for the hydrophobic regions of the lipids to interact with
water. In the presence of water, lipids aggregate in such a way as to sequester their
hydrophobic regions away from water. The unfavorable interaction of the hydrophobic
regions of lipids with water is the fundamental driving force for membrane formation and
is called the hydrophobic effect (2). Due to the hydrophobic effect, lipids are able to self
aggregate (without the aid of other molecules or covalent bonds) into highly defined
structures.

As early as 1925, it was postulated that membranes were composed of a bilayer of
lipids (3). This structure allows the hydrophobic regions of the lipids to be sequestered
towards the center of the bilayer and away from water. The hydrophilic regions are on the
surface of the membranes and are free to interact with water. There are many different
three-dimensional conformations, or phases, that the bilayer can form. These will be
discussed in more detail later in this chapter. However, it is the sheet-like, lamellar
structure that encapsulates cells and is, therefore, central to the studies of biological

membranes.

 

 

 

structure
that bio!
rennin

the plus

rhit th'
mentbt

E5545

thigh
heterc
limb

0T1 lllcl

COUnup,
Sm“Tlur

£70170”

lll llle {U

3
In 1972, Singer and Nicolson proposed the fluid mosaic model to explain the

structure, organization and motion of biological membranes (4). This model proposed
that biological membranes are a bilayer of lipid molecules that serve as a matrix for
membrane associated proteins (Figure 1.1). The lipids and proteins are free to diffuse in
the plane of the membrane. Thus, the membrane is viewed as a two dimensional viscous
fluid in which the lipids and proteins are randomly distributed. Although we now know
that this model is much to simple to explain the structure and function of biological
membranes, it has served as the foundation upon which all the modern membrane
research has been based.

Biological membranes are extremely complex systems. This is highlighted by the
number of different molecular species that may be present in a membrane. There are
many different classes of lipids (>10) which may be present (a lipid class is defined by its
headgroup and backbone structures). These lipids typically have two hydrocarbon chains
attached, of which there are over 20 different possibilities. On top of this lipid
heterogeneity, there are potentially hundreds of different proteins associated with the
membrane. This diverse set of molecules is part of an aggregate whose properties depend
on their intermolecular interactions and motions.

This chapter on biological membranes can only begin to scratch the surface of the
enormously diverse areas of research in membrane biology that has filled the pages of
countless journal articles and books. I begin this chapter with an overview of the
structure of lipids and the physical foundation of their organization. Next I discuss the
motion (dynamics) and order (structure and organization) of lipids and the role they have

in the function of the membrane. I end this chapter by discussing a class of lipids known

 

Figure 1.1: The fluid mosaic model of biological membranes in which the proteins and
lipids form a two-dimensional viscous fluid. The proteins and lipids are free to move
randomly in the plane of the membrane. Adapted from reference 4.

S
as lipopolysaccharides. These are unique and special lipids that are capable of eliciting a

very swift and severe immune response. It is the mechanism of their action that is at the

heart of this thesis.

Structure and Phase Behavior of Membrane Lipids

A tremendous diversity of lipid structures are found in biological membranes
including fatty acids, glycerophospholipids, sphingolipids, glycolipids, steroids, and some
vitamins. The structures of some of these lipids are shown in Figure 1.2. A more
complete review of membrane lipid structures can be found in reference 5. All of these
lipids have a common structural motif. They all have a non-polar, hydrophobic region
and a polar, hydrophilic region. The important differences lie in the relative proportions
of these two regions present, the charge state of the molecule, and the three dimensional
conformation which it adopts.

The class of lipids known as glycerophospholipids comprises the bulk of the
membrane lipids in eukaryotes. Their basic structural unit is a phosphoglycerol. To the
phosphate is attached one of several different polar molecules which generates the
hydrophilic headgroup of the lipid. Fatty acids are attached to the hydroxyl groups of the
glycerol to provide the hydrophobic region of the lipid. An example of a
glycerophospholipid is dipalmitoylphosphatidylcholine (see Figure 1.2). This lipid has a
choline group attached to the phosphate and two 16-carbon saturated fatty acids (palmitic
acids) attached to the glycerol.

The phase (aggregate structure) that a given set of lipids will form depends on the

intra- and inter-molecular forces of the lipid molecules. The hydrocarbon chain region of

How 0
O OH
3-hydroxytetradecanoic acid 0

How
0
palmitic acid

 

Cholesterol

= -OH phosphatidic acid

W(CH2)nW”O/RO ’OWN(CH3)3+ phosphatidylcholine
W(CH;,)nW"LO’{_O_(-3_0_R ,OwNH; phosphatidylethanolarnine
0'

general structure of a \ioz . .
diacylglycerophospholipid ’0 NH; p hosphatrdylserme

loflofl phosphatidylglycerol

Emit}! phosphatidylinositol

OH

 

O
OF-o’V 3 3
0'

dipalmitoylphosphatidylcholine

Figure 1.2: Some common membrane lipids.

7
the lipid experiences steric repulsive and van der Waals attractive forces. The polar

headgroup of lipids are influenced by steric repulsion, electrostatic interactions and
hydration forces (6). The molecular forces of both the hydrophobic and the hydrophilic
regions of the lipids are balanced to provide the minimum free energy of the lipid
aggregate. Thus, this delicate balance of forces dictates their aggregate properties and,
hence, their phase behavior.

The balance of molecular forces in the lipid aggregate results in each of the lipid
molecules having an effective (average) geometry. The packing constraints of this
geometry determine the overall 3-dimensional shape that the membrane will adopt (6).
This is illustrated in Figure 1.3. For instance, a lipid that has a cylindrical shape will tend
to pack together in planar sheets, thus giving rise to a lamellar phase. If, however, the
headgroups experience charge repulsion, their geometry may change and result in the
lipids adopting a different phase.

It can easily be seen that any factor that alters the delicate balance of
intramolecular and intermolecular forces can alter the phase state of the lipids. These
factors include temperature, pressure, hydration, pH, or ion concentration. The addition
of exogenous lipophilic molecules (such as proteins, antibiotics, organic solvents, or
lipids) will also perturb this balance of forces. If this perturbation causes the lipids to
favor a different aggregate structure, they will undergo a phase transition to the new,
lower energy phase. For example, below 35°C dipalmitoylphosphatidylcholine (DPPC) is
in the lamellar gel (Lg) phase (Figure 1.4). When this lipid is heated to 35°C, it
undergoes a transition to the ripple (ng) phase. Continued heating results in another

phase transition at 42°C. The thermal energy imparted into the system causes the lipids

Figure 1.3: Relationship between molecular geometry of lipids
and the phases they adopt. Taken from reference 7.

 

Micellar

 

Normal
Hexagonal

 

Lamellar

 

Inverted
Micellar

 

Inverted
Hexagonal

 

Figure 1.3

10
to undergo trans-gauche isomerizations around the carbon-carbon bonds of the acyl

chains and thus increase their motion. This is accompanied by the increased hydration of
the lipid headgroup and an increase in the overall area occupied by each lipid molecule.
The result is the lamellar liquid crystalline (Lu) phase. This is one example of lipid
polymorphism. Of course, the types of phases that a membrane will adopt and the
temperature (or other factors) which will cause a phase transition will be different for

each type of lipid studied.

WWW/M WWW/ll /. 111%”ch

WW4 35c WWW 42c it
WW WW. WM hhllllllltllii

Will/WW

Lamellar Gel Ripple Lamellar Liquid Crystalline

Figure 1.4: A schematic of the phase behavior of dipalmitoylphosphatidylcholine as a
function of temperature. Modified from reference 8.

Membrane Dynamics

The fluid mosaic model of biological membranes highlights one of the most
fundamental properties of membranes: their motion. Membranes are ordered two-
dimensional fluids that exhibit a range of different types and rates of motions while
maintaining a large degree of order. Therefore, they have the properties of both fluids

and crystals. Hence the general term, liquid crystals, is used to describe materials which

11
can form this unique state of matter. Membrane dynamics are one of the critical factors

that determine the physical state of the membrane. As will be discussed later in this
chapter, the dynamic state of a membrane also influences its function.

There are several types of motions (dynamics) that occur in membrane lipids.
These motions range from the very fast vibrations of the C-H bonds to the very slow
translation of lipids from one side of the bilayer to the other. Concomitantly, there are a
range of techniques which are sensitive to different time scales of motion and are used to
study membrane dynamics. The types and rates of motions found in membranes and the
time scales of the techniques used to study them are summarized in Figure 1.5.
Membrane dynamics are often described in terms of its 'fluidity'. Indeed, the membrane
literature in inundated with this term. Fluidity is simply the inverse of viscosity and
. viscosity is defined as the amount of resistance to motion in a medium. Fluidity and
viscosity are only valid when applied to isotropic substances and, therefore, are not
rigorously correct terms to describe anisotropic systems such as membranes. When used
to describe membranes, fluidity and viscosity are used in a more qualitative sense. In
reality, they describe a combination of motions. Therefore, they preclude the possibility
of providing a detailed picture of the molecular motions being measured and their use
should be avoided when such information is desired.

Lipid hydrocarbon chain motions constitute some of the fastest modes of motions
found in membranes and are also one of the most critical when trying to define its
dynamic state. They result from the trans-gauche isomerizations of the carbon-carbon
bonds in the acyl chains. Lipids that are in a rigid, solid-like phase (such as the LB: phase
of DPPC) have nearly all of their carbon-carbon bonds in an all-trans configuration. In

this configuration, the acyl chains are straight and can pack tightly together. Lipids that

12

. |-—-—-| Lipid “llip-tlop"
|-—-l Protein rotation
l—-—l Lateral “hopping" (lipids and proteins)
l——l Lipid rotation (about long axis)
l——l Trans/gauche isomerization
l—l -CH2- vibration
Frequency (sec ‘ ‘)

i i i ' i r 1
10“ 10‘° 106 to2 10‘2 10‘6
l——-l Raman/IR

|'-—-l ESR (nitroxide radicals)
4 Saturation transfer ESR

|-—-——-1 Fluorescence depolarization
' #1 Phosphorescence depolarization/triplet probes

I-—--l 2I-i- NMR (lineshape)

1 #4, NMR (relaxation)

 

 

Figure 1.5: Rates of membrane motions and the techniques that are sensitive to these
motions. Taken from reference 9.

13
are in a fluid-like phase (such as the La phase in DPPC) have an increased number of

gauche conformations in their acyl chains. The gauche conformations put 'kinks' in the
acyl chains and results in their inability to pack tightly together. Using infrared
spectroscopy, Casal and McElhaney have demonstrated that there is an average of 3.7
gauche conformers per acyl chain of DPPC at 5°C above the gel (P5) to lamellar-liquid
crystalline phase transition (10). The trans-gauche isomerizations of the carbon-carbon
bonds are constantly changing their position along the acyl chain.

A more detailed analysis of the lipid acyl chain motion has been performed by
Peterson and Chan (11). They have postulated that there are two major modes of motions
in the acyl chains of lipids. The first type of motion arises from large fluctuations of the
acyl chain as the result of single gauche conformers. A single gauche conformer would
produce a bend in the acyl chain and produce a large structural perturbation, especially if
it is located near the lipid headgroup. This large structural perturbation would be
unfavorable due to the steric interactions with neighboring lipids. Thus, this type of
motion is expected to be slow. Peterson and Chan (11) have suggested that this type of
acyl chain motion occurs at a rate of approximately 10’8 s. The unfavorable large
perturbation induced by a single gauche conformer can be partially overcome by a second
gauche conformer next to the first but in the opposite direction. Thus the acyl chain will
have neighboring gauche(+) and gauche(-) conformers resulting in a kink. A kink in the
acyl chain results in a much smaller structural perturbation than a single gauche
conformer and is therefore less energetically expensive. By successive trans-
gauche(+)/gauche(-)-trans isomerizations, the kink can translate up and down the acyl
chain. This is the second mode of acyl chain motion and is called 'kink diffusion'. The

rate of trans-gauche isomerizations that results in kink diffusion is approximately 10'10 s.

14
Membrane Order

The hydrophobic effect causes lipids and proteins to self-assemble into large
ordered aggregates (membranes). Order, therefore, is an inherent property of any
membrane. There are three basic types of membrane order: orientational order,
asymmetry, and lateral order. Orientational order is the degree to which the membrane
molecules adopt the same orientation relative to the bilayer normal. Asymmetry refers to
the amount of similarity between the two halves of the bilayer. The final type of order is
a measure of the lateral organization of the molecules in the plane of the membrane. I
will focus this discussion on this latter type of membrane order. An excellent discussion
of orientational order and asymmetry can be found in reference 12.

The fluid mosaic model proposed that the molecules of biological membranes are,
for the most part, free to move randomly in the plane of the membrane (4). Moreover,
this motional freedom results in an essentially homogeneous lateral distribution of
molecules in the membrane. We know now that this is not true. Membrane components
readily laterally separate into regions (domains) of distinct compositions, physical
properties and motions. Membrane domains usually refer to a lateral segregation of
lipids. They may or may not be accompanied by the lateral segregation of proteins.

The balance of entropy and intermolecular interaction energy (6) decides whether
or not lipids form domains. If a lipid has a particular structure that results in the strong
interaction with a neighboring lipid, then the two molecules will tend to remain together.
If this interaction is weak, however, then the entropy of mixing predominates and the
lipids will diffuse apart. This delicate balance of interaction energy and entropy results in
a dynamic equilibrium in which those lipids that have a stronger energy of interaction will

tend to associate together into distinct domains. Any perturbation in this equilibrium,

15
such as changes in temperature, pressure, pH, or the addition of lipophilic molecules, will

result in a change in the domain organization of the membrane. Depending on its
composition, each domain will have different structural, physicochernical and motional
properties.

The most definitive and elegant studies on membrane domains have been
performed with lipid monolayers (see ref. 13). In these experiments, synthetic lipids are
doped with fluorescent lipid analogs and allowed to assemble at the air-water interface.
The lipid monolayers are directly visualized by epifluorescence microscopy as a function
of surface pressure, temperature, or other factors. Peters and Beck (14) were the first to
visualize the formation of domains by this method during the 'fluid'-to-'crystalline' phase
transition. Studies by McConnell and co-workers as well as Miller and Mohwald have
greatly expanded these initial findings. They found that lipids, even those of a pure
molecular species, exhibit regular, periodic arrays of domains whose shape can be highly
complex and variable (15—17). These domains not only had distinctive shapes, but they
also had characteristic motional (15) and physical properties including their electrostatic
dipole moment (18, 19). Furthermore, these studies also demonstrated that the size and
shape of lipid domains are extremely sensitive to small amounts of added cholesterol (20,
21). Thus it is obvious that even small changes in lipid composition can lead to dramatic
long-range perturbations in membrane order.

Given the ability of membranes to form domains even with the very pure lipids
systems used in the monolayer studies, it is perfectly reasonable to expect that more
complex lipid systems will also be able to laterally segregate into domains. This is, in
fact, true for both biological membranes and liposomes composed of various lipid

mixtures. In the latter case, mixtures of lipids have been shown to exist in domains by

16
using differential scanning calorimetry (22, 23) and by measuring the steady state and

time resolved emissions from fluorescent probes (24-26). More direct evidence of
domain formation in phospholipid vesicles has been provided by Haverstick and Glaser
(27). This study visualized the formation of acidic phospholipid rich domains induced by
Ca2+ using epifluorescent microscopy and fluorescently labeled lipids.

Lipid domains have also been observed in cellular membranes. Some cells have
large, usually morphologically distinct, regions of their membranes that contain specific
lipid and protein compositions. This type of 'macrodomain' has been observed in
polarized epithelial cells (28), the sperm cell plasma membranes (29), inner- and outer-
membrane adhesion zones in gram-negative bacteria (30), and the thylakoid membranes
of chloroplasts (31).

Cellular membranes also have smaller membrane domains, sometimes referred to
as 'microdomains'. Due to the inherent structural complexity of cellular membranes, we
should expect that membrane microdomains are ubiquitous. However, because these
domains are relatively small and can not be isolated, studies on them have been difficult
and slow to evolve. We are now just beginning to understand and measure the properties
of these membrane domains. Lipids may be reorganized by their interaction with
membrane associated proteins. This has been observed for gramicidin and D—B-
hydroxybutyrate dehydrogenase (32), pyrularia thionin (33), and some ATPases (34, 35).
Of course, domains may also be created by the lateral segregation of the lipids
themselves.

A wide variety of methods have been employed to detect domains in biological
membranes (see ref. 36). Several of these studies provide definitive evidence for their

existence. An innovative use of fluorescence recovery after photobleaching (FRAP) has

17
been used by Yechiel and Edidin to detect domains in the plasma membrane of

fibroblasts (37). By varying the diameter of the laser beam, they were able to
demonstrate a large lateral heterogeneity in the diffusion rates of a fluorescent lipid. Yu
et al. (38) were also able to demonstrate the existence of domains in fibroblasts cell
membranes. These authors used a unique fluorescent probe, laurdan, and two-photon
scanning microscopy to demonstrate that there is lateral heterogeneity in the plasma and
nuclear membranes. The membrane of erythrocytes has also been shown to contain lipid
domains by fluorescence microscopy (27, 39, 40).

The reports discussed above demonstrate that all membranes, even those
composed of very pure lipid species, can laterally segregate into domains. These domains
will have different physical and functional properties. We must now focus not only on
detecting these domains, but also on developing an understanding of their physical
properties, what affects their formation and, most importantly, how they affect the

function of membranes.

So far, I have treated membrane dynamics and order as if they were separate
entities. However, it is important to realize that the motions and organization of
membranes are tightly coupled. If one type of membrane motion is altered, other modes
of motion as well as its organization may also be altered. These membrane properties are
intimately interconnected and can never be isolated. By measuring some of the modes of
dynamics or order, we are attempting to understand the membrane as a whole. Clearly,

the complexity of membranes makes this ultimate goal extremely difficult to achieve.

18
Physiological Relevance of Membrane Dynamics and Order

Traditionally, lipids have. been viewed as the inert matrix in which membrane
proteins are located. We now know that lipids play a vital role in the function of the
membrane. It is clear that, on top.of their structural role, lipids actively participate in
signal transduction events and are the substrates for some enzymes. However, the
dynamics and order of lipids also play an important role in regulating the functions of the
membrane.

The degree of membrane order and motion (often described as 'fluidity') has long
been thought to be critical to its function. Biological membranes generally have a
significant proportion of their lipids in the fluid-like liquid-crystalline phase and it was
proposed that normal membrane functions require this phase (41). The physiological
relevance of membrane dynamics and order is highlighted by studies on the adaptations of
organisms to environmental stress. Many organisms alter their membrane composition in
response to environmental perturbations (such as temperature, pH, organic solvents, etc.)
in order to maintain a constant dynamic state of the membrane (see ref. 42, 43). This
phenomenon has been called homeoviscous adaptation (44). Clearly, the maintenance of
a particular dynamic state of the membrane is critical to normal cellular functions.

The intimate interactions of membrane associated proteins with lipids suggest that
the functions of these proteins may be modulated by the physical state of the membrane.
A change in the 'fluidity' of the membrane has been correlated with a change in the
activity of several enzymes including succinate cytochrome c reductase, phospholipase
A2, various ATPases, adenylate cyclase, acyl CoA cholesterol acyl transferase, glucose-6-

phosphatase, and fatty acid desaturase (see ref. 45). Furthermore, many membrane-

19
associated enzymes display discontinuous Arrhenius plots, and the breaks in the

Arrhenius plots are often associated with the motional state of the membrane (45).

While the above mentioned examples lend credence to the idea that the dynamic
state of the membrane can regulate protein function, they do not examine the full scope of
lipid-protein interactions. Many of these reports use exogenous lipids, such as
cholesterol, to alter the motional state of the lipids. However, other physical properties of
the membrane may also be altered in these studies. In fact, cholesterol is known to affect
membrane properties in a very complex manner. It increases the order of fluid lipids (46)
as well as dramatically affecting the size and shape of lipid domains (20). As discussed
above, membrane dynamics and order are interconnected. The need to investigate other
aspects of the physical properties of membranes is emphasized by studies that argue that
'fluidity' per se may not be the critical factor regulating the activity of membrane bound
enzymes (47-49). These reports suggest that other physical properties of membranes may
be involved. Indeed, it has been demonstrated that phospholipase A2 (PLAz) is most
active at the onset of the bilayer-to-nonbilayer phase transition (50, 51). This indicates
that PLA2 activity is sensitive to the bilayer packing stress and defects. Senisterra and
Epand (52) observe a similar activity dependence with protein kinase C (PKC). They
observed that the activity of PKC is not simply regulated by membrane defects, but is also
closely associated with the hexagonal phase propensity of the membrane. They were
unable to determine the exact reason for this dependence. To this end, Tate et al. have
proposed that enzyme activity can be modulated by the build up of elastic strain in the
membrane as a result of non-lamellar forming lipids (53).

The lateral organization of lipids has also been demonstrated to be a critical

determinant of membrane function. Indirect evidence for this is provided by studies on

 

the effects
demonstra
the lateral
addition 0:
organizatio:
receptor act
shown that
reticulum m
More direct
demonstrate
hl‘droljze sc
Phase domai
into protein:
the activity C
mfimbrane (it
some CDmitri
favorable for
[sing
lateral difosi.
Compoblllon.
percolalion Iht
changes, the 3
number of “pr

Previously ISO

20
the effects of aging on myocyte and fibroblast cell membranes (54-57). These studies

demonstrate that the lipid compositional changes seen in aging cells results in a change in
the lateral mobility and organization of the membrane. These changes are reversed by the
addition of the depleted lipid species. Furthermore, they demonstrated that these
organizational changes which occur with aging resulted in a disruption of the muscarinic
receptor activity and ability to couple to G-proteins. Additionally, Asturias et al. have
shown that the structural and functional changes in the Ca2+-ATPase of sarcoplasmic
reticulum membranes is closely associated with the lateral organization of the lipids (58).
More direct evidence for the role of lipid domains in the membrane function has been
demonstrated with phospholipase A2. This enzyme has been shown to target and
hydrolyze solid-like lipid domains (59). Moreover, after a critical amount of the solid-
phase domains have been hydrolyzed, the enzyme inactivates itself by self-aggregating
into proteinacious domains in the membrane. In an elegant study by Dibble et a1. (60),
the activity of protein kinase C was found to be strongly dependent on the formation of
membrane domains with specific lipid compositions. They also demonstrated that, under
some conditions, the coexistence of domains with different lipid compositions was
favorable for PKC activity.

Using model membranes and FRAP, Ameida et a1. (61) have observed that the
lateral diffusion of lipids exhibits large and sudden changes with changes in the lipid
composition. They interpreted their results in terms of membrane domains and
percolation theory. As the lipid composition (or other parameter such as temperature)
changes, the solid-lipid domains (for example) grow as a result of an increase in the
number of lipids in that phase. At a critical point, the solid-lipid domains that were

previously isolated from one another now become connected. This is called the

 

percolatior

 

disconnecz:
change in

Muderhwa
dependent
preferential

and connec

 

 

 

 

regulation 0
membrane 1*.
121168 and }l(
segregation
functional 3.

Sinc.
it is ObViQu
membranes.
pmpertics O
the physical
See how dc
i“inactionS
membrane C
demonstrate

(65" This

Drowns ox

the lipid hEa

21
percolation threshold. At the percolation threshold, lipid domains that were previously

disconnected become continuous and visa-versa. This results in a sudden and dramatic
change in the lateral mobility of the lipids contained in these domains. Interestingly,
Muderhwa and Brockman have demonstrated that pancreatic lipase activity is strongly
dependent on lipid domain percolation (62). Because membrane proteins are
preferentially soluble in specific lipid domains (63), the organization, size, composition
and connectivity of membrane domains can provide a critical mechanism for the
regulation of membrane function. This has been demonstrated in a theoretical analysis of
membrane reactions in which it was found that domain size and connectivity can alter the
rates and yields of, enzymatic (or other) reactions in the membrane (64). Thus, the lateral
segregation of membranes into domains provides a mechanism for cells to create specific
functional areas within the membrane.

Since it is now well established that domains are found in biological membranes,
it is obvious that these domains are likely to be critical to the normal function of cell
membranes. Less clear is a molecular understanding of the physical and functional
properties of the domains. In the discussions above, I have given some insights into how
the physical properties of membranes can control the activity of enzymes. One can also
see how domains can create functional areas in the membrane and can control the
interactions of substrates and proteins in the membrane. The organization of the
membrane can, however, have additional roles in its function. For example, it has been
demonstrated that lipids can conduct protons laterally on the surface of the membrane
(65). This results in the efficient coupling of spatially separated proton pumps. The
protons move via an H-bond network that is critically dependent on the conformation of

the lipid headgroups. Domains of differing lipid compositions and conformations would

22
define the boundaries of the proton travel and therefore affect the proton motive force.

Membranes are also known to be ion storage devices (66). Ions, especially Ca2+, are
critical to normal cellular functions and signal transduction events. Interestingly, Ca2+ can
affect domain formation (27). Conversely, disruption of these Ca2+ rich domains will
release Ca2+. Thus, membranes are capable of absorbing and releasing Ca2+ in a spatially
defined way. Finally, membrane ‘defects’ found at the boundaries of domains may
induce special functions. I discussed above that membrane defects might be associated
with altered protein function (50-52). In addition to this, regions of membrane defects
cause an increase in the permeability of the membrane to water and small molecules (67,
68).

From the above discussions, it is easily seen that membrane dynamics and order
play a major role in the function of the membrane, which, in turn, can control the
physiology of the cell. Therefore, there is an intimate and dynamic coupling between the
physical state of the membrane, its function, and the function of the cell. This coupling is
summarized in Figure 1.6. Understanding the interplay between the dynamics and order
of the membrane and the physiology of the cell is essential to grasp an understanding of
membrane biology. While‘we know that this connection exists, we have only begun to
discover the full breadth of its potential.

In the sections above, I have discussed membrane structure, motions and
organizations. Furthermore, I reviewed how these membrane properties might influence
membrane and cellular function. In the remainder of this chapter, discussions will focus
on an unusual class of lipids (lipopolysaccharides) with an amazing ability to alter the
function of cells. The ability of lipopolysaccharides to alter cellular functions results in a

dramatic affect on the immune system of animals. As will be seen in the remainder of

 

23

.=8 2: .«o awe—033m 05 can .285an 05 mo canoes.“ 05
.520 28 88896 05582: concave wfiasoo 23. 6; 833m

24

ea semi

 

_ =~U ac «mix—mom Fame—3mg; _

 

my

33880::
5805-588?

 

moan—35m we SSE.

53:24 oazmem.

 

 

 

552:5 Eugene—z

 

 

 

A

v

 

32:89:00 .35

53:89.8 .8339 £5.53

H065 o

€388 38mg»: £288
382 £388 5.26 303
magic .

 

85.3.3.5 «53:52

 

25
this dissertation, understanding how lipopolysaccharides affect membrane dynamics and

order is important to understanding its mechanism of action.

Lipopolysaccharides

Lipopolysaccharides (LPS) are a unique class of glycolipids found on the surface
of Gram-negative bacteria (Figure 1.7). LPS are capable of eliciting a wide array of
biological responses when they interact with cellular systems of animals. A list of some
of these biological responses is given in Table 1.1. Some of these responses are
potentially beneficial, such as the ability to stimulate the immune system, to act as an
adjuvant and to induce tumor necrosis. Other effects, such as the induction of fever,
blood pressure depression, shock, and disseminated intravascular coagulation, can be
harmful. In fact, in severe cases, the biological effects elicited by LPS lead to the death
of the individual. A discussion of the mechanisms of these biological effects will be
given later in this chapter.

Lipopolysaccharides were first discovered by Richard Pfeiffer in 1892. He found
that Gram-negative bacteria produce toxins that are distinct from the more commonly
known exotoxins in that they were heat-stable and remained associated with the bacteria.
Due to these unique properties, he called them endotoxins (see ref. 69). Since this first
discovery over 100 years ago, a tremendous amount of effort went in to elucidating the
chemical structure of endotoxins. It was found that they consisted of both lipid and
polysaccharides. Thus a name reflecting its chemical composition, lipopolysaccharides,
was given. Today both endotoxins and lipopolysaccharides are used essentially
interchangeably although the latter term usually refers to the more highly purified form.

The majority of structural work on lipopolysaccharides has been accomplished

26

 

 

 

Figure 1.7: Outer envelope of Gram-negative bacteria. Lipopolysaccharides are located
on the outer leaflet of the outer membrane.

27
with enteric bacteria such as Escherichia coli or Salmonella. Therefore, discussions of

LPS structure imply that they are derived from these bacteria unless otherwise noted.
LPS are composed of two regions, a lipid region and a polysaccharide region (Figure 1.8).
The polysaccharide region is further divided into the O-specific polysaccharide and the
core oligosaccharide. The O—specific polysaccharide (O-antigen) is composed of
repeating oligosaccharide units (see ref. 70). An O-polysaccharide from wild type LPS
will have between 0—40 oligosaccharide repeats. The core oligosaccharide is linked to the
lipid moiety by an unusual sugar, 3-deoxy-D-manno-2-octulosonic acid (KDO). This
eight-carbon sugar is acid labile and facilitates the chemical cleavage of the lipid and
polysaccharide regions. The core oligosaccharide may also contain pentoses, hexoses,
heptoses, amino sugars and some substituents such as phosphate and

phosphorylethanolamine (see ref. 70, 71).

 

Table 1.1: Some typical biological activities elicited by lipopolysaccharides.

0 Lethal toxicity 0 Induction of prostaglandin synthesis

- Pyrogenicity 0 Induction of tumor necrosis factor

0 Bone Marrow Necrosis 0 Induction of interferon

0 Hypotension 0 Induction of interleukins

o Platelet aggregation - Induction of colony stimulating factor
0 Complement Activation 0 Mitogenic for B-lymphocytes

o Adjuvant Activity 0 Macrophage activation

0 Induction of Tumor Necrosis o Endothelial cell activation

0 Activation of clotting enzyme

cascade of the lysate of Limulus
amebocytes

 

28
The lipid region of LPS is called lipid A. The structure of lipid A, solved in 1983

(72), consists of a B-1,6 linked glucosarnine disaccharide headgroup that is

phosphorylated at the 1 (0t) and 4’ positions (Figure 1.9). Lipid A contains a total of six
fatty acids. Four 3-hydroxyl tetradecanoic fatty acids are attached directly to the
glucosamines by amide and ester linkages at the 2, 3, 2’ and 3’ positions. The 3-hydroxy
groups of the 2’ and 3’ fatty acids are esterified by a dodecanoic and a tetradecanoic fatty
acids, respectively. This structure was confirmed when it was produced by total organic
synthesis and was found to be indistinguishable from bacterial-derived lipid A (73).
Because this structure has the highest level of biological activity, it is considered to be the
‘typical’ or ‘model’ lipid A. However, some degree of heterogeneity is found in all lipid
A preparations. Additionally, lipid A derived from bacteria other than E. coli or
Salmonella may contain significantly different structures (see ref. 74).
Lipopolysaccharides (and lipid A) have not lent themselves easily to
physicochernical studies. This is partially due to the intrinsic heterogeneity of LPS
preparations. Further complications arise from differences in experimental conditions.
As with phospholipids, the physical properties of LPS are sensitive to environmental
conditions such as temperature, pressure, water content, pH, and ions. The negatively
charged groups in LPS make them extremely sensitive to these latter two parameters (75,
76). Cations such as Ca2+ and Mg2+ are especially critical to the physical properties of
LPS since they are capable of binding to the phosphates of lipid A and cross-bridging
them together to form a tight complex. These factors have made even the simple

observation of the chain melting transition produce tremendously variable results (7).

29

 

.moetmnooambomom=
Lo BBQ—Em 2: mo Emuwfle came—23m ”w; oSwE

 

30

3 2ng

e6< been ”I .ouafimonm n. .ocnasooamocoz ”O

 

 

l. coma—.8333 w

 

 

 

HIP—Bo l: J. comu:<.0 ._l

 

 

31

am
View
0
NH H 0
OH
I
NHO-E—O'
O
H

Figure 1.9: Structure of E. coli lipid A.

The lipid A moiety of LPS anchors it to the outer membrane of Gram-negative
bacteria where it helps to form a barrier to environmental stress (77). It is the physical
properties of the LPS aggregate, especially in the lipid A region, which provides it with
this ability. Labischinski et a1. (78) demonstrated that lipid A readily formed aggregates
with a remarkable degree of order. They postulated that this aggregate order imparts the
outer membrane with its ability to act as a permeability barrier. The ability to form highly
ordered aggregates is a result of the three-dimensional conformation of lipid A and its
ability to cross-link adjacent phosphates. Wang and Hollingsworth (79) have shown that
lipid A adopts a compact triangular wedge shape in which the acyl chains are rigid and
tightly packed together. This shape can readily pack into hexagonal arrays and explains

the crystal structures observed (80, 81). Additional molecular order is provided from the

 

 

cross-1m.

this resu
blechai

clinical
methan
ltspons
unrespo
follow
responsi
l’tnderer
These

maerOpl

e“dilthe
to LPS
imPOm
median
tumor

prOSlag
1mm U116
lOr [l1 6 I

SUPPOne

32
cross-linking of adjacent phosphates by divalent cations. Coughlin et a1. (75) found that

this results in a dramatic increase in the rigidity of the lipid A headgroups.

Mechanisms of Lipopolysaccharide Induced Biological Activities

The broad spectrum of biological activities produced by LPS (see Table 1) and its
clinical significance has produced a tremendous amount of interest in determining its
mechanism of action. Studies by Michaleck et a1. (82) demonstrated that many of the
responses to LPS are mediated by immune cells. They used a mouse strain that is
unresponsive to LPS (C3H/I-IeJ) and showed that LPS sensitivity could be restored
following strong irradiation by an adoptive transfer of bone marrow cells from LPS
responsive mice (C3H/HeN). They also found that LPS responsive mice could be
rendered LPS unresponsive by the adoptive transfer of C3H/HeJ bone marrow cells.
These results implicated bone marrow derived cells, such as lymphocytes and
macrophages, in mediating the biological affects induced by LPS.

It is now known that LPS affects (activates) many animal cells including platelets,
endothelial cells, B-lymphocytes and macrophages. The strong response of macrophages
to LPS has made it the center of much research and has implicated it as the most
important cell in mediating the response to LPS. Macrophages produce several soluble
mediators (cytokines) when stimulated by LPS. These include interleukin-1 (IL-1), lL-6,
tumor necrosis factor (TNF), interferon (IFN), colony stimulating factor (CSF),
prostaglandins and more (83). These mediators, in turn, modulate the response of other
immune cells in a complex manner. The immune response is believed to be responsible
for the physiological effects of LPS such as fever, hypotension, shock and death. This is

supported by studies that demonstrated that the soluble mediators released by

33
macrophages are capable of eliciting biological effects that are associated with LPS (85-

88).

Determining the biologically active moiety of LPS was of primary importance in
understanding its mechanism. A breakthrough came in 1967 when Kasai and Nowotny
discovered a Salmonella minnesota mutant that only synthesized a very short
polysaccharide region (89). This mutant LPS still maintained all the biological activity of
the wild type LPS. Thus, lipid A was suspected of being the active moiety. Irnoto et al.
(73) confirmed this when they synthesized lipid A and found that it contained nearly
equivalent biological activity as LPS. A study by Haeffner—Cavaillon et al. (90) has
modified this last assertion by demonstrating that the induction of IL-1 secretion was
more effective with the 3-deoxy-D-manno-2-octulosonic acid (KDO) residues on the lipid
A. While the proximal sugar residues are able to modulate its activity, the lipid A moiety
is generally recognized to contain most of the biological activities of LPS.

With the discovery of lipid A as the biologically active moiety of LPS came a
tremendous amount of interest in unraveling its structure-function relationship. This
topic has been the focus of many studies that utilized natural and synthetic lipid A, lipid
A analogues and partial structures. Results from these studies have been summarized in
several excellent reviews (91-94). Among all the structures tested, it is the lipid A
structure found in E. coli (Figure 1.9) that has optimal biological activity. Both the
headgroup and fatty acid structures are critical components of this activity. Removal of
either one of the phosphate groups results in a dramatic loss of activity. Loss of both
phosphates renders lipid A inactive. The biological activities of lipid A are also
dependent on the fatty acid structures. Addition or loss of one fatty acid, a change in the

type of fatty acids, or the rearrangement of their positions results in a decrease in activity.

34
The sensitivity of the fatty acid structures is, perhaps, most apparent with the lipid A from

Rhodobacter sphaeroides. This lipid A has the same headgroup structure as found in E.
coli and differs only in its fatty acid region (Figure 1.10) (95). However, this lipid A is
completely devoid of biological activity (95-99). A complete analysis of all the structure-
function studies led Reitschel et a1. (94) to conclude that the headgroup structures,
especially the phosphates, are essential for binding of lipid A to immune cells, while the
fatty acid region is responsible for the activation of the cells. Furthermore, they conclude
that “the binding of lipid A is a necessary but not sufficient event” for the activation of
immune cells, thereby implicating the hydrophobic interactions of the acyl chains as a
critical component in eliciting the signal which results in cellular activation.

The search for the molecular mechanism of lipid A-mediated activation of
immune cells has led to the proposal of specific membrane associated protein receptors
(for reviews, see ref. 100-102). This was fueled by the finding that Rhodobacter
sphaeroides lipid A is able to block the biological activities of enteric lipid A or LPS,
suggesting that it competes for a specific binding site (97-99). Lei and Morrison (103,
104) detected an 80 kDa membrane-associated protein on B-lymphocytes, T-lymphocytes
and macrophages by a photoactivatable LPS derivative. Hampton et al. (105) found that
several proteins from a macrophage-like cell line bound to lipid A, of which a 31 kDa and
a 95 kDa protein were predominant. Another report utilizing a macrophage-like cell line
identified a 55 kDa and a 65 kDa protein that bound to lipid A (106). Kirkland et al.
(107) found that 18 kDa, 25 kDa, and 28 kDa proteins from a B-lymphocyte cell line
bound to lipid A. A report by Wright and long (108) implicated CD11/CD18 in the
binding of LPS to macrophages. Finally, the LPS binding protein (LBP)/CD14 complex

was shown to bind to LPS (109).

35

OH
O
NH H 0
OH
I
H NH O—R—O'
O
H

Figure 1.10: Lipid A structure from Rhodobacter sphaeroides.

To date, the best characterized of the putative LPS receptors is the LBP/CD14
complex. Studies suggest a sequence of events that lead to cellular activation by
LBP/CD14 (see ref. 110). First, LBP, a soluble serum protein, binds to LPS. The LBP-
LPS complex then is recognized by CD14 and the LPS is transferred to it. Finally, the
LPS-CD14 complex elicits a signal transduction cascade. However, CD14 is not a
transmembrane protein and is only loosely associated with the membrane via a
phospholipid anchor (111). Therefore, another membrane protein, which interacts with
LPS or LPS-CD14, is thought to be involved. A recent report suggests that this may be
an 80 kDa protein (112). It is noteworthy, however, that LPS is capable of activating

macrophages in the absence of LBP or CD14 (109). This is supported by the finding that

36
94% of CD14 could be removed from macrophages by phorbol esters without blocking

their ability to respond to LPS (113). Another report showed that monocytes from CD14-
deficient patients were able to produce IL-1 in response to LPS (114). Therefore, the role
of CD14 in mediating an immune response to LPS is still undetermined. In fact, of all of
the putative LPS receptors identified, none of them have been definitively shown to be
responsible for the activation of immune cells.

While the role of LPS receptors remains unclear, it is clear that the lipid A region
of LPS interacts non-specifically with membranes. Because lipid A is an amphiphilic
molecule like any other lipid, it is obvious that it will be able to interact with and insert
into membranes. There have been numerous reports illustrating the non-specific
hydrophobic interaction of LPS with phospholipid, erythrocyte, platelet, lymphocyte, and
macrophage membranes (see ref. 115). A few of these reports provide more direct
support for the non-specific interaction of LPS with membranes. Schuster et al. (116)
showed that LPS perturbed cholesterol-lecithin bilayer membranes (black lipid
membranes). They introduced LPS into the aqueous phase next to the membrane and
observed a rapid change in the membrane electrical resistance and stability. Their results
suggest that the lipid A region of LPS inserted into the bilayer and caused a dramatic
change in its integrity. Experiments by Jackson et al. (117) support the role of lipid A in
binding to membranes. They attached spin labels to the carbohydrate groups of LPS and
monitored its interaction with macrophages by electron spin resonance. The rate of
motion of the labeled carbohydrates was not affected by the binding of LPS to
macrophage cells, thus showing that the polysaccharide region of LPS is not involved in
the binding to the macrophage membrane. Furthermore, this implicates lipid A as the

region that is responsible for the binding. Experiments by Carr and Morrison (118, 119)

37
used polymyxin B (a small polycationic peptide) to analyze the binding of LPS to

erythrocyte membranes. They demonstrated that the interaction of LPS with erythrocyte
membranes was time and temperature dependent and occurred in two discernable steps.
The first step is the ionic interaction of LPS with the membrane and can be prevented by
the preincubation with polymyxin B. The second step is temperature dependent and was
proposed by the authors to be due to the insertion of lipid A into the membrane. Jacobs
and co-workers obtained nearly identical results when they studied the interaction of LPS
with lymphocyte membranes (120). They also proposed that this interaction occurred in
two steps with the first step being ionic in nature and the second being the irreversible
insertion of lipid A into the membrane.

In 1987, Jacobs and Price (121) proposed a model to explain the molecular
mechanism of the interaction of LPS with immune cells which leads to its activation
(Figure 1.11). This model was formulated from an analysis of data from their lab as well
as published data. This ‘two-step’ model is similar to that discussed above (118-120) and
attempts to bring together the conflicting data on the role of LPS receptors and non-
specific membrane interactions. The first step is the electrostatic interaction of lipid A
headgroup and/or LPS polysaccharide region with the surface of the membrane. This is
supported by the structure-function studies discussed above which concluded that the
lipid A headgroup is involved in the binding to immune cells. The second step is the
rearrangement of the LPS so that the lipid A is inserted (fused) into the membrane
bilayer. This involves non-specific interactions of lipid A with the membrane
components. This step is temperature dependent due to the change in the physical

properties of the membrane and the LPS aggregates with temperature, thus affecting the

38

down mm..— 2: .8 >288 88 08:8 2: :o maze—Snow 8228a 88 .228— .mni
8 mm..— wo Home—E a on 8 38.: £232.88 a we 55an 2 288w? mm; 2;
.=8 28882 05 Co 82388 2: 9 2582 8o>o R226 05 on 8 eomomoa
2 e558o8 2t 98 mad he 8283.2 25. .o88n8u8 2t 28 mar— mo :2on
< 2&— ofi Lo 8225 c2888 2: 2 no; 8808 2S. 283808 88:8
2: :23 mm; Mo :28822 2:38.583 Eu 2 82m 35 2E. .CN: 025 28
302: E eomoaoa .2608 Q2292 2: Lo 88me 28838 < 2 _.2 0.82m

 

39

_ 2 ._ semi

85.32: "N 3am 82888:— ocfimcboofi— A 85m

       

O)

    

o
a
it.
28.5892 LI _ .

    

=oU «8: 0:55:52
0 as f

. .3 a (w assm

. o

. .
532a u. AU m

/ . or

a a

. a

. .. .

a... o

40
ability of lipid A to fuse with it. The authors propose that there may be proteins (or other

molecules) which facilitate the binding and insertion of LPS into membranes. These
molecules would have the properties characteristic of lipid A/LPS receptors.
Furthermore, they propose that the activation of immune cells by LPS is a consequence of
the insertion of lipid A into the membrane and not the binding to the surface of the
membrane. The authors suggest that a membrane perturbation resulting from the
insertion of lipid A may be the molecular signal which leads to cellular activation (122).
This notion is supported by the structure-function studies (discussed above) that
concluded that the acyl chain region of lipid A is responsible for the activation of immune
cells. To this end, Jacobs and co-workers have demonstrated that the insertion of lipid A

results in a decrease in the ‘fluidity’ of membranes (122-124).

Conclusion

To date, the two-step model proposed by Jacobs and Price (121) is the only
molecular model for the activation of immune cells by LPS. Its strength lies in its ability
to explain the seemingly disconnected data in the LPS literature. This model sparks new
questions in trying to understand the molecular events leading to the activation of
immune cells by LPS, and, more importantly, provides a framework to interpret
experimental results.

In this dissertation, I explore the molecular events of LPS-cell interactions,
specifically at the membrane level. The second chapter analyzes this interaction by
utilizing the unusual LPS structures from Rhizobium. Rhizobium are soil bacteria that are
not animal pathogens. However, they synthesize unusual LPS/lipid A structures and are

useful in dissecting the LPS-host interactions. In the third chapter, I focus on the effects

41
that LPS/lipid A has on membranes. Since membrane properties are a very important

factor in the function of membranes and cells, membrane perturbations elicited by LPS
may be the critical molecular signal leading towards the activation of immune cells. I test
this hypothesis and further our understandings of the non-specific LPS-membrane
interactions. In the fourth chapter, I explore the role of membrane perturbations in
mediating the biological effects of LPS. This is accomplished through the use of
synthetic lipid ‘probes’. Finally, the results obtained in this dissertation and their

significance are summarized in Chapter 5.

The]

Chapter 2

The Relationship Between Structure and Biological Activities of Lipopolysaccharides:

Lessons from Rhizobial LPS

42

43
Introduction

Lipopolysaccharides (LPS) elicit a wide spectrum of biological responses in
animals (see Table 1.1). It is now well established that most of the biological activities
associated with LPS are due to the lipid A moiety. Many of these activities are the result
of the interaction of lipid A with immune cells such as macrophages and B-lymphocytes.
This interaction results in the alteration of the physiological state of the immune cell
(usually to a more active state) and induces, for instance, the secretion of tumor necrosis
factor-0t (TNF-0t) from macrophages and the increased cell division in B-lymphocytes.
However, the molecular mechanism of the lipid A-host cell (or LPS-host cell) interaction
and the physical basis for the mechanism of the modulation of immune cell activity is
poorly understood.

Several researchers have proposed that the electrostatic interaction of lipid A with
the cell membrane is the first step in the association of lipid A with immune cells (119-
121). This interaction is thought to be mediated by the headgroup of lipid A, especially
the phosphate groups (94). It is not clear whether the electrostatic interaction between
lipid A and the immune cell is mediated by non-specific membrane interactions or is
mediated by a stereo-specific lipid A receptor. Many investigators have proposed the
existence of a receptor on the cell surface (see Chapter 1). However, no signal
transduction events resulting from these interactions have been uncovered. Therefore,
the role of a lipid A-specific receptor in the activation of immune cells remains to be
determined.

The physicochemical properties of lipid A have been implicated as important

determinants of biological activity. Using defined salt fornls of LPS, Galanos and co-

44
workers (125, 126) demonstrated that its biological activity is very dependent on its salt

form. Changes in the salt form of LPS are known to change its physical properties as
well as its supramolecular (aggregate) size and structure (76, 125-128). Brandenburg and
co-workers (7, 129) used Fourier transform infrared spectroscopy and X-ray diffraction to
analyze the supramolecular structure of LPS. They found a strong correlation between
the supramolecular conformation of LPS and its biological activity. These investigations
have led to the proposal of an ‘endotoxic conformation’ as a determinant of its biological
activity. The requirement of a particular supramolecular structure of LPS for the
maintenance of biological activity is appealing since it explains how subtle changes in the
lipid A headgroup or acyl chain structure, which might alter the supramolecular structure,
can alter biological activity. It also opens up the possibility that other unusual lipid A
structures can be biologically active provided that they have the proper physicochemical
properties.

Of course, knowledge of the structures of LPS required for biological activity is a
prerequisite to obtaining a detailed molecular understanding of the LPS-host cell
interaction and the subsequent events that lead to cellular activation. Many studies have
addressed this issue (see ref. 91-94). More recently, some very unusual lipid A structures
have been identified from species of Rhizobium. Rhizobial LPS have been found to be
distinct from enteric LPS in that they have a low phosphorous content (130) and contain
an unusual trans-membrane fatty acid, 27-hydroxyoctacosanoic acid (131). The structure
of R. trifolii lipid A was recently solved (Figure 2.1) (132). Its headgroup is comprised
of a galacturonic acid 1(a)-6 linked to glucosamine. It does not contain phosphates. A

lactic acid is glycosidically linked to the glucosamine. It also contains a total of five fatty

45

OH O OH
O
O
O
O O
O HO O O
o O OH
O NH
O O
HO H HE

OH
0

Figure 2.1: Structure of R. trifolii lipid A. From reference 132.

46
acids, four of which are attached directly to the headgroup carbohydrates at the 2, 3, 2’,

and 3’ positions, including the 27-hydroxyoctulosonic acid (27-OH 28:0). Interestingly,
this LPS still has some biological activity despite its unusual structure (133). One report
also found that the lipid A from R. meliloti has an unusual structure although it retains the
same headgroup as found in enteric bacteria (134). These authors report that it only
contains four fatty acids including a 27-OH 28:0. This LPS has been shown to be about
as biologically active as E. coli LPS (135). However, as will be discussed later in this
chapter, studies in our laboratory on LPS on a different strain of R. meliloti do not agree
with these reports.

The unusual lipid A structures found in species of Rhizobium provide an
opportunity to analyze the LPS-cell interaction event and gain an insight into the
molecular mechanisms by which LPS activates immune cells. In this chapter we describe
the structural analysis of R. meliloti 2011 LPS and its biological activities. These results
are compared to R. tnfolii LPS and Salmonella typhimurium LPS in an effort to
understand the structural features of lipid A that permit biological activity. In turn, these
results allow us to make important conclusions about the molecular mechanisms of the

LPS-cell interactions and the nature of the events that lead to cellular activation.

Materials and Methods

LPS Isolation and Punfication. Lipopolysaccharide from Salmonella
typhimurium (Sigma) was purified by gel filtration chromatography on a Sepharose 43
column using 0.05M ammonium forrnate, pH 5.5, as the buffer. The first peak eluting off

the column was positive for 3-deoxy-D-manno-octulosonic acid (136) and was pooled,

47
dialyzed against distilled H20 and lyophilized. Rhizobium meliloti 2011 was grown in

modified Bergensen’s medium (pH 6.9) and harvested in late exponential phase. Cells
were harvested by centrifugation at 8,000Xg for 15 min. The cell pellet was vigorously
stirred for 12 hrs with a waterzchloroform:methanolzbutanol (4:2:lzl) mixture at room
temperature. After centrifugation, the aqueous phase was removed and extracted
sequentially with chloroform and butanol. The aqueous phase (containing LPS) was
treated with a solution of DNase I, RNase A, and Mng to remove contaminating nucleic
acids. R. meliloti 2011 LPS was further purified by gel filtration chromatography on a
Sepharose 4B column as described above. Column fractions were monitored for total
carbohydrate content by the phenol-sulfuric acid assay (137), 3-deoxy-D-manno-2-
octulosonic acid (KDO) by the periodate/thiobarbituric acid assay (136), uronic acids by
the m-phenylphenol assay (138) and phosphorus by the method of Lowry et al. (139).
The LPS was located in the carbohydrate peak that contained large amounts of KDO and
no phosphate or uronic acids. LPS was dialyzed against distilled water, concentrated by
rotary evaporation, and lyophilized. Cells used for 35$ labeling studies were grown under
the same conditions with 35S-sulfate supplemented in the media. Column fractions
obtained from the extract of cells grown in 358 enriched media were further assayed by
scintillation counting. Rhizobium trifolii ANU843 was grown in modified Bergensen’s
medium (pH 6.9) and harvested at late exponential phase as described above. R. trifolii
ANU843 LPS was extracted by aqueous phenol as previously described (140). The
resulting LPS solution was dialyzed extensively against distilled H20, treated with DNase
I and RNase A, and purified by gel filtration chromatography as described above.

Galbraith Laboratories, Knoxville, TN, performed elemental analysis of LPS.

48
Lipid A Isolation. R. meliloti 2011 LPS was hydrolyzed with 5 ml of 1% acetic

acid at 100°C for 2 hrs, cooled to room temperature, and partitioned between water and
chloroform. The organic layer, which contained the lipid A, was dried under nitrogen.
Structural Analysis of Lipid A. R. meliloti 2011 lipid A was hydrolyzed with 1 ml
of 5% HCl in methanol for 30 hrs at 70°C. The reaction mixture was cooled to room
temperature and dried under a stream of nitrogen. The product was partitioned between
water and chloroform, and the organic phase containing fatty acid methyl esters was
concentrated to dryness under nitrogen. The fatty acid methyl esters were analyzed by
gas chromatography (GC) and gas chromatography/mass spectrometry (GC/MS). GC
was performed on a Hewlett Packard 5890 instrument equipped with a capillary DB1
column (J & W Scientific, Folsom, CA) using a temperature program of 150°C to 300°C
at 3°C/min, hold at 300°C for 20 min. GC/MS was performed on a Jeol JMS-AXSOSH
spectrometer in the electron impact (E1) mode at 70 eV. The methyl glycosides in the
aqueous phase were purified by loading the reaction mixture on a C18 cartridge (Sep-
Pak, Waters Associates, Inc., Milford, MA) and eluting with 4:1 waterzmethanol. The
purified methyl glycosides were peracetylated by a 1:1 mixture of pyridine2acetic
anhydride (0.2 ml) for 12 hrs at room temperature. After peracetylation, the reaction
mixture was acidified by the addition of 1 drop of 3M HCl, and partitioned between 0.5M
NaCl and chloroform. The peracetylated methyl glycosides were recovered in the
organic layer and concentrated under nitrogen. The sample was subjected to GC
(capillary DB225 column, from 180°C to 230°C at 2°C/min, hold at 230°C for 80 min)

and GC/MS analysis as described above.

49
Spectroscopy. Nuclear magnetic resonance (NMR) spectroscopy of R. meliloti

lipid A was performed using the Varian VXR-SOOS (500 MHz) spectrometer in CDCl3

with a small amount of CD30D added. Spectra were recorded at room temperature and
were referenced to the chloroform solvent signal at 7.24 ppm. 31P-NMR spectra were
obtained on a Varian VXR-300 spectrometer.

USP Pyrogen Assay. The ability of S. typhimurium, R. meliloti 2011, and R.
trifolii ANU843 lipopolysaccharides to induce fever in rabbits was determined according
to the standard USP pyrogen protocol at the Biological Test Center, Irvine, CA. S.
typhimurium LPS was administered at 0.05 ug/kg body weight. R. meliloti and R. trifolii
LPS was administered individually at 0.05 ug/kg and 1.0 ug/kg. Body temperature was
measured before and at 1, 2, and 3 hrs after injection and the maximum rise in
temperature was used. If one or more rabbits exhibited a temperature rise of 06°C or
greater or if the total temperature rise of the three rabbits exceeded 14°C, the test was
considered to be positive.

USP Systemic Toxicity. The toxicity of S. typhimurium, R. meliloti, and R. trifolii
LPS was determined in 5 mice. Tests were performed at the Biological Test Center,
Irvine, CA. LPS was administered intraperitoneally at 500 or 1000 [lg/animal. These
animals were observed at 0, 0.25, 4, 24, 48, and 72 hrs. Weights of the mice were
determined immediately before and after each trial. The sample is toxic if one of the
following occurred: two or more mice died, abnormal behavior in two or more mice, or
weight loss of more than 2 g in three or more mice.

Cell Lines. RAW264.7, a murine Abelson leukemia virus-transformed

lymphocyte with macrophage properties (141), was obtained from American Type

50
Culture Collection and grown in Dulbecco's modified Eagle's medium (DMEM)

supplemented with 10% fetal bovine serum and 100mg/L streptomycin sulfate (culture
medium). Subcultures were prepared by scraping and resuspending in fresh culture

medium. Cells were grown at 37°C in a humidified, 5% C02 atmosphere.

Tumor Necrosis Factor-a Assay. LPS samples were lyophilized, rehydrated in
Dulbecco’s PBS, sonicated, and vortexed thoroughly. RAW264.7 cells (2 x 105) were
placed in lml of culture medium in 24 well culture plates and allowed to grow for 15 hrs.
Samples (50 ul) were added to the cell culture wells and allowed to incubate with the

cells at 37°C, 5% C02. Cells were allowed to incubate with the samples for 8 hrs since a

time course experiment determined that the maximal TNF response occurred at this time.
After incubation, cell culture supematants were removed, centrifuged to remove cell
debris, and placed on ice. The TNF-or concentration in the culture supernatant was
determined using a mouse TNF-or ELISA kit (Genzyme Corp., Cambridge, MA). TNF—or
levels were quantitated with recombinant mouse TNF-0t according to the test kit
instructions. The results shown are the average of duplicate determinations and the error
bars show the range of the results.

Limulus amebocyte lysate (LAL) assay. The LAL assay was performed with the
Whittaker Bioproducts QCL-lOOO Quantitative Chromogenic LAL, Whittaker
Bioproducts, Inc. (W alkersville, MD). All assays were standardized to Endotoxin Units
(EU) using E. coli 0111:B4 LPS according to the assay instructions. Samples were
prepared in distilled water and then sonicated and vortexed thoroughly immediately prior

to being assayed. The concentration of 'LPS tested was 0.1 ng/ml. The assay was

5 1
performed following the test kit instructions. Results shown are the mean of at least three

determinations.

Results

Rhizobium meliloti 2011 Lipopolysaccharide isolation and characterization. LPS
from R. meliloti was extracted and purified by gel filtration chromatography (Figure 2.2).
LPS was contained in the first carbohydrate peak that eluted off the column. KDO, a
common LPS sugar, was detected only in this peak. This peak contained only trace
levels of uronic acids or phosphate, indicating that these are not major components of the
LPS. Also, no signal was detected by 3'P-NMR spectroscopy.
Interestingly, R. meliloti cells grown in 35S-sulfate enriched medium were found to
incorporate 35s into their LPS (Figure 2.3). The 35s is thought to be incorporated as
sulfate groups since it is readily removed by mild acid hydrolysis and the liberated sulfate
can be precipitated by Barium (Hollingsworth, unpublished results). Sulfate is a rare
component in LPS and has previously only been reported for Pseudomonas aeruginosa
LPS (142). A sulfur content of 0.5 % (w/w) for R. meliloti 2011 LPS was demonstrated
by elemental analysis. Given an approximate LPS molecular weight of 10000 g/mole,
this would be the equivalent of an average of 1.5 sulfate groups per LPS molecule. The
locations of the sulfate groups are difficult to determine because they are so labile.
However, chloroform extraction of bacterial cultures yielded free lipid A (without acid
cleavage) which was found to contain 358, thus demonstrating that at least some of the
sulfate groups are associated with the lipid A moiety (Hollingsworth, unpublished

results).

52

 

2.0

. —'— Phenol-Sulfuric Assay
. —°'— KDO Assay

1.5 '

1.0 ‘
* LPS

Absorbance

 

 

 

 

 

 

Fraction

Figure 2.2: Purification of R. meliloti 2011 LPS by Sepharose 4B gel filtration
chromatography.

Phenol-Sulfuric Assay (Abs. 490 nm)

53

 

 

 

 

 

 

0.3
—'— Phenol-SulfuricAssay
.—°'_ CPM
3s
S-LPS ..
*40005
O
‘ ”2000
\g‘“
0.0 --- “o
15 20 25 30 35 4o 45

Fraction

Figure 2.3: Purification of 35s-iabeueci R. meliloti 2011 LPS by gel filtration
chromatography.

 

 

 

chm
indit
iden
lipid
olig-
from

by

30
Ch
fat

01

54
R. meliloti 2011 lipid A was obtained by the mild acid hydrolysis of the LPS.

Analysis of the carbohydrates contained in the lipid A was accomplished by gas
chromatography (GC) and GC-mass spectrometry (GC-MS) (Figure 2.4). The results
indicate that glucosamine is the major component. Some glucose and heptose were also
identified. Although the possibility of glucose and heptose being an integral part of the
lipid A structure can not be ruled out, they are common core and 0-antigen
oligosaccharide components and are, therefore, believed to be contaminants remaining
from the acid hydrolysis of LPS. The fatty acids of R. meliloti lipid A were also analyzed
by GC and GC-MS (Figure 2.5). Three fatty acids were predominant: 3-
hydroxytetradecanoic acid (3-0H 14:0), 3-hydroxyoctadecanoic acid (3-0H 18:0), and
27-hydroxyoctulosonic acid (27-OH 28:0). They were present in an approximate molar
ratio of 3:1:1, respectively. Hexadecanoic acid (16:0), 3-hydroxyhexadecanoic acid (3—
0H 16:0) and octadecanoic acid (18:0) were present in approximate molar ratios of less
than 0.4.

The 1H-NMR spectrum of R. meliloti lipid A was acquired in a deuterated
chloroform/methanol solvent. This yielded a high resolution spectrum (Figure 2.6). The
very intense signal at 1.15 ppm and the triplet at 0.78 ppm are due to the methylene and
terminal methyl protons of the acyl chains, respectively. The group of signals between
2.1 ppm and 2.6 ppm arise from protons on the acyl chain or to the carbonyl carbon.
Some doublets of doublets (such as those at 2.29 ppm and 2.53 ppm) are resolved and are
characteristic of the protons coupled to the methine proton at the 3 position of hydroxy
fatty acids. The triplet at 2.19 ppm is most likely due to protons on the 2 position of 27-

OH 28:0 since this is the primary fatty acid present without a substituent on the third

55

.oagmooiw .U new .3890: .m .883» «o Eofiofi

o2 .< an 836285 2E .< 2% :8 .832: .z eoe 836285
noun—boomeom new cogs—>52: 65 mo nafiwofifiofio 30 ”ed ouswfi

ea 2:me

can

 

 

 

56

asuodsog Joiooiaq

 

57

ignoommoe .93 30.8 98 .92 IO-m .ouwm .92 30% .92
.9: sea 2 288:8 a: 938 2F .< ca: :8 coins .m
:8: $23 352: Boa 5.8 05 Mo casewoumEoEo $0 ”Wm EsmE

58

 

 

Time

 

Figure 2.5

 

 

-—_.l

osuodsou 10109193

59

4x8 2: E 53w 8d flcogwfimm
38on .< 29. :8 .eoEos .Eo 8.58% «22-x. 2 Beam

60

 

 

 

 

 

 

 

 

 

 

o." no"
PLrF— L»

h

3 2:3

o.n rim

~_r_Lth—

 

 

 

 

 

 

 

 

 

_

 

_

k

c.v

»

_

P

_

 

61
carbon. The complex multiplets between 1.2 ppm and 1.6 ppm are primarily due to

methylene protons in the acyl chains that are or to the hydroxy-substituted carbons or are
B to the carbonyl carbons. The intense signals at 3.28 ppm and 3.48 ppm are due to the
deuterated methanol and residual water in the sample, respectively. The remaining
signals between 3.0 ppm and 4.2 ppm are primarily from methine protons of the fatty
acid carbons that bear an unsubstituted hydroxyl group and from protons on the
carbohydrate headgroup. The group of signals between 4.5 ppm and 5.5 ppm are
primarily from anomeric protons, protons located on the carbohydrate carbons that are
acylated, and methine protons on the fatty acyl carbons that bear an acyloxyacyl
substituent. Interestingly, the spectrum contains a quartet at 4.85 ppm and a doublet at
1.07 ppm. These signals are identical to the signals found in R. trifolii lipid A and
assigned to a lactic acid group attached to the 1 position of the glucosamine residue
(132). The presence of these signals in R. meliloti lipid A indicates that it also has this
group.

Biological Activity of R. meliloti 2011 LPS. Four standard tests for biological
activity were carried out on the LPS of R. meliloti 2011 and the results compared to R.
trifolii LPS and Salmonella typhimurium LPS. The tests used to measure biological
activity were (1) the USP pyrogen test in rabbits, (2) the systemic toxicity test in mice,
(3) induction of tumor necrosis factor-0t release from a macrophage cell line, and (4) the
ability to activate the clotting enzyme cascade of the amebocyte lysate from Limulus
POIyphemus (LAL assay). With the exception of the tumor necrosis factor assay and
LAL assay, the biological activities for R. trifolii LPS and S. typhimurium LPS were

presented earlier (133) and are provided here for completeness.

62
The ability to induce the secretion of cytokines (immune mediators) from

macrophages is a critical measure of the biological activity of LPS. Tumor necrosis

factor-0t (TNF-or) is one of the primary mediators of the biological effects of LPS and is

the cytokine produced first by macrophages in response to LPS. In turn, TNF-0t
stimulates the secretion of other cytokines such as interleukin-1 (IL-1) (85). Together,
TNF-0L and IL-1 are thought to be the major mediators of the biological effects of LPS
including the induction of fever and toxicity (85-88). We tested the ability of rhizobial
and enteric LPS to induce the secretion of TNF-or from a macrophage-like cell line
(RAW264.7) known to be activated by LPS (141). LPS from R. trifolii and R. meliloti
were far less potent than that from S. typhimurium (Figure 2.7). At 1.0 ng/ml, R. trifolii
LPS and R. meliloti LPS induced 6% and 0% of TNF-0t compared to S. typhimurium
LPS, respectively. Both of the rhizobial LPS were able to induce the production of TNF-
a at higher concentrations.

Since it is generally thought that TNF-0t is the primary mediator of pyrogenicity
and toxicity induced by LPS, it would be predicted that rhizobial LPS would not be as
effective as enteric LPS in these tests. This is indeed what was found. The toxicity of the
LPS was tested in mice at 500 ug and 1000 pig LPS per animal (Table 2.1). The mice
were monitored for signs of toxicity and their weights were recorded before and after the
test as a measure of their health. The rhizobial LPS did show some signs of toxicity
including lethargy and moderate weight loss. However, they were clearly far less toxic
than S. typhimurium LPS. Interestingly, R. meliloti LPS appeared slightly more toxic
than R. trifolii LPS. This is the opposite of what we would predict from the TNF

induction assay. The ability of the LPS to induce fever in rabbits was tested (USP

63

 

 

 

 

5000
4000 "1 .
Salmonella typhimurium LPS .
A l
E
R:
3 3000 -
i?
g, Rhizobium trifolii LPS
Z
E-‘
2000 -
O
i . Rhizobium meliloti LPS
I 0
law I I l T l
0 2.5 5 7.5 10
LPS concentration (ng/ml)

12.5

Figure 2.7: Tumor Necrosis Factor (TNF-or) assay of S. typhimurium LPS, R. meliloti

2011 LPS, and R. tnfolii ANU 843 LPS.

Table 2.1: Systemic toxicity of S. typhimurium, R. meliloti 2011, and R. trifolii

 

 

ANU 843 LPS.
Weight Observations' Weight
Sample Dose Before (hrs) After
(pg) (g) 0 0.25 4 24 48 72 (g)
Salmonella 500 21 0 0 l D - - -
typhimurium 20 0 0 l D - - -
LPS 20 0 0 l l l l 14
19 0 0 l l l l 15
17 0 0 l l l l 14
1000 19 0 0 l D - - -
21 0 0 l D - - -
20 0 0 l D - - -
22 0 0 l D - -
l9 0 0 l D - - -
Rhizobium 500 22 l I l 0 0 0 27
meliloti 21 l l l 0 0 0 22
LPS 18 l l l 0 0 0 19
20 l l l 0 0 0 20
22 l l l 0 0 0 23
1000 19 l l l 0 l 0 17
23 l l l 0 l 0 22
23 l I l 0 l 0 21
22 l l l 0 l 0 1 8
21 | l I 0 0 0 20
Rhizobium 500 21 0 0 l 0 0 0 21
trifolii 21 0 0 l 0 0 0 23
LPS 21 0 0 l 0 0 0 23
22 0 0 l 0 0 0 24
20 0 0 l 0 0 0 22
1000 22 0 0 l l 0 0 21
20 0 0 l l 0 0 20
20 0 0 l l 0 0 20
1 8 0 0 l l 0 0 l9
l7 0 0 l l 0 0 l6

lObservations: 0 = no symptoms of toxicity, 1 = lethargy, and D = death

 

 

65
pyrogen assay). Again, the rhizobial LPS are less potent than S. typhimurium LPS in

inducing fever (Table 2.2). R. trifolii LPS produced a significantly higher fever than R.
meliloti LPS.

In 1964, Levin and Bang (143) reported that amebocytes obtained from the
horseshoe crab (Limulus polyphemus) were extremely sensitive to LPS. LPS activates a
clotting enzyme cascade causes the formation of gelatinous clot, or, if an artificial
substrate is used, into the development of a color. This was developed into an in vitro
assay to test the biological activity of LPS, called the Limulus amebocyte lysate (LAL)
assay, and is widely accepted as a measure of its biological activity. The results for the
LAL assay are shown in Figure 2.8. S. typhimurium LPS, R. meliloti LPS and R. trifolii
LPS had an LAL activity of 7.8 EU/ng, 13.3 EU/ng and 13.9 EU/ng, respectively.

Clearly, the rhizobial LPS are very active in this assay.

Discussion

The structural characterization of Rhizobium meliloti lipid A clearly demonstrates
that it has an unusual structure compared to E. coli or Salmonella lipid A. Although we
do not, at the present time, have its complete structure determined, we know that it is
substantially different than enteric lipid A. The headgroup is composed primarily of
glucosamine. This is consistent with a glucosamine disaccharide headgroup similar to
that found previously for R. meliloti (134) and enteric lipid A. However, instead of
phosphates, R. meliloti lipid A contains lactic acid and sulfate. The fatty acids present are
also significantly different. R. meliloti lipid A contains fatty acids ranging from 14-28

carbons long, which is much longer than found in enteric lipid A. From the molar ratios

66

Table 2.2: USP Pyrogen assay of S. typhimurium, R. meliloti 2011, and R. trifolii
ANU843 LPS.

 

Sample Dose Weight Max. Temp. Total Temp.
(Hg/kg) (kg) Rise (°C) Rise (°C)

Salmonella 0.05 2.63 1.18 2.89
typhimurium 3.26 1 . 13
2.74 0.58

Rhizobium 0.05 1.88 0.00 0.06
meliloti l .97 0.00
1 .86 0.06

1 .00 1 .92 0.00 0.65
l .69 0.28
l .77 0.37

Rhizobium 0.05 1.91 0.30 1 .56
trifolii 2.59 1.13
2.50 0.13

1.00 2.37 0.73 2.70
2. 15 1.19

2.24 0.78

67

 

15‘: 133 13.9

 

 

Rhizobium trifolii

Salmonella typhimurium
Rhizobium meliloti

Figure 2.8: Limulus Amebocyte Lysate (LAL) assay of S. typhimurium LPS, R. meliloti
2011 LPS, and R. trifolii ANU843 LPS. Activity is given as Endotoxin Units (EU) per
ng LPS.

of the fatty acids, R. meliloti lipid A appears to contain five fatty acids (three 3-0H 14:0,
one 3—0H 18:0, and one 27-0H 28:0). Overall, the structure of R. meliloti lipid A and R.
trifolii lipid A appear to be similar. They both contain a lactic acid, a disaccharide
headgroup, and similar length and number of fatty acids. The major difference lies in the
galacturonic acid and the lack of sulfate in the headgroup of R. trifolii lipid A.

The biological activities of the rhizobial LPS reflect these similarities. They both
have similar levels of biological activities. Both are lower in toxicity than S.
typhimurium LPS and are weaker in their ability to induce the production of TNF—0t and
induce fever. However, they still harbor these biological activities albeit at a lower level.

Interestingly, both rhizobial LPS were very powerful stimulators of the LAL assay.

68
Furthermore, Lill-Elghanian (133) reported that R. trifolii LPS is a potent immune

stimulant, being nearly as active as Salmonella LPS in stimulating the division of B-
lymphocytes (B-cell mitogen assay).

Many earlier structure-function studies on lipid A have attempted to utilize its
potentially beneficial biological activities (immune stimulation) while minimizing its
toxic effects. Some success has been found with monophosphoryl lipid A (MPLA) (144).
MPLA is very weakly toxic but is still active in the LAL assay and is a good B-
lymphocyte mitogen (93). LPS from R. meliloti 2011 and R. trifolii ANU843 also appear
to have these properties. They both have a low level of toxicity but are strong inducers of
LAL activity. Therefore, they may also have some therapeutic value similar to that found
for MPLA. Interestingly, Lill-Elghanian reported that, unlike MPLA, R. trifolii LPS is
not able to induce an antibody response from B-lymphocytes (133). The unusual
structures of Rhizobial LPS, along with their distinct biological activities, reaffirms the
notion that the biological activities elicited by LPS are not triggered by a single
mechanism. Furthermore, it questions the role of a universal LPS receptor (such as
LBP/CD14) in triggering the LPS-response of many different cell types including
macrophages and B-lymphocytes.

The unusual lipid A structures from R. meliloti and R. trifolii provides an insight
into the LPS-cell interaction. As mentioned above, this interaction is mediated by
electrostatic interactions. It has been postulated that the binding of lipid A to a
stereospecific receptor on the membrane surface induces a signal transduction cascade
that elicits a physiological response (see Chapter 1). However, R. meliloti and R. trifolii

LPS are able to elicit a physiological response even though they have completely

69
different charge functionalities than enteric lipid A. This clearly indicates that the LPS-

cell interaction is not very specific. The lack of specificity does not necessarily rule out
the possible involvement of membrane proteins. Indeed, the most promising of the
putative LPS receptors (CD14) has been found to be very non-specific (145, 146). CD14
readily binds to a variety of polyanionic substances and is, therefore, not a lipid A-
‘specific’ receptor. We feel that this lack of specificity severely questions a direct role of
lipid A receptors in the activation of immune cells. Perhaps, as suggested in the two-step
model (121), these receptors simply facilitate the incorporation of lipid A into the host
cell membrane. In fact, several recent reports have demonstrated that LBP and CD14
facilitate the incorporation of LPS into phospholipid bilayers (147-149).

As discussed in Chapter 1, the hydrophobic interactions of the acyl chains of lipid
A with the host cells is a critical component of eliciting a physiological response (94).
The biological activity of lipid A is extremely sensitive to its fatty acid structure. This is
apparent with the lipid A from Rhodobacter sphaeroides (Figure 1.10) that contains five
fatty acids but is completely devoid of biological activity (95-99). However, I
demonstrate here that the lipid A from Rhizobium contains the same number of fatty acids
but retains some biological activity. Obviously, the number of fatty acids present per se
is not the factor determining biological activity. The fatty acids must be influencing
some property of the lipid A that is able to modulate its activity. We can begin to
understand this by analyzing the differences between the fatty acyl structures found in
enteric, R. sphaeroides, and Rhizobium lipid A. It is known that enteric lipid A readily
forms highly ordered aggregates (78, 79). R. sphaeroides lipid A has only five short fatty

acids, one of which is unsaturated. This would be expected to significantly decrease the

70
order of the fatty acyl region of this lipid A as well as decrease its aggregate order. This

has recently been demonstrated by Bérubé (150). The lipid A from Rhizobium also has

only five fatty acids; however, they are all saturated fatty acids and are much longer than
found in enteric lipid A. This would be expected to increase the order of the fatty acyl
region. This is especially true for the trans-membrane 27-0H 28:0 fatty acid found in
Rhizobium. It has been demonstrated that trans-membrane fatty acids synthesized by
thermophilic bacteria help to stabilize their membrane (151). This same effect would be
expected for the 27-OH 28:0 fatty acid found in rhizobial lipid A. We believe that this
change in the physicochernical properties of lipid A (especially in the order of the fatty
acyl region) is responsible for the modulation of its biological activity.

As discussed in the introduction, the structure of lipid A from a different strain of
R. meliloti (10406) has been reported (134). Its structure contains a glucosamine
disaccharide backbone bearing two phosphates and four fatty acids. The fatty acids are
of the same length as the R. meliloti 2011 lipid A structure reported here, including a 27-
OH 28:0 fatty acid. Interestingly, LPS from R. meliloti 10406 was found to be very toxic
(135). The different biological activity elicited by this LPS, as well as the structural
studies, clearly shows that R. meliloti 10406 and R. meliloti 2011 lipid A have different
structures. This is surprising since the structure of lipid A is usually invariant among
closely related bacteria. However, because R. meliloti 10406 lipid A is proposed to have
only four fatty acids and is still toxic, the studies on it only serve to strengthen the
arguments made above.

In this study, I demonstrated that lipid A from Rhizobium have very unusual

structures but still retain biological activity. These results clearly demonstrate that the

71
LPS-host cell interaction is not highly specific. They also indicate that the

physicochernical properties of lipid A, especially in the fatty acyl region, is responsible
for the activation of immune cells. This is in agreement with the studies cited above that
found a correlation between the physicochernical properties of LPS and its biological
activity (7, 125-129). While it is difficult to determine the underlying physical
mechanism of this correlation, the importance of the order of the fatty acyl region of lipid

A suggests that it may involve a perturbation in the properties of the host cell membrane.

Chapter 3

Lipopolysaccharide Induced Changes in the Order and Dynamics of Membranes

72

73
Introduction

Lipopolysaccharides (LPS) are complex glycolipids located on the outer
membrane of Gram—negative bacteria. LPS are capable of eliciting a broad spectrum of
biological activities in animals including fever, hypotension, shock and death. These
effects are primarily due to the severe immune response induced by LPS.

The first interaction of LPS with the host (immune) cell occurs at the membrane.
Since LPS are amphiphilic molecules, it is understandable that insertion of its lipid
anchor (lipid A) into the host cell membrane should take place. Several early
investigations suggested this mode of interaction (152-156). Since then, many reports
have definitively demonstrated that the lipid A region of LPS inserts into membranes (see
Chapter 1 and ref. 116-120, 157). Moreover, it has been proposed that the activation of
immune cells by LPS requires the insertion of lipid A into the host cell membrane (117,
121, 15 8).

Although LPS have been the focus of intensive research for more than 20 years,
the signal that initiates an LPS-induced cellular response has remained elusive. This
signal could arise from perturbations in the host cell membrane properties that occur
when LPS inserts. This is reasonable since it is known that the physical properties of
membranes are sensitive regulators of their functions and, in turn, modulate cellular
functions (see Chapter 1). The ability of LPS to form highly ordered, negatively charged
aggregates (78, 79) suggests that it would be capable of inducing substantial
perturbations in the properties of membranes in which it was incorporated. The
requirement of the insertion of LPS into the host cell membrane to elicit a cellular

response is consistent with the involvement of such perturbations. Moreover, the

74
dependence of the activity of LPS on its structure and physicochemical properties

(detailed in Chapter 2) would also support a role of membrane perturbations. The
aggregate form of LPS and its degree of acyl chain order would be especially critical to
its ability to alter membrane properties and would explain why its fatty acid composition
is so critical to its ability to activate immune cells (94).

The above arguments indicate that membrane perturbations might be a critical
component in the activation of immune cells by LPS. However, relatively little is known
about how LPS causes such perturbations. The first indirect evidence that the
incorporation of LPS into phospholipids results in a change in its physical properties was
provided by Rothfield and Takeshita (159). They demonstrated that phospholipid-LPS
mixtures protected a transferase against thermal inactivation whereas neither LPS nor the
phospholipids were effective alone. Since then, several other studies have investigated
the effects of LPS on the dynamics and order of membranes. Lui et al. (160) showed that
LPS increased the transition temperature of phosphatidylethanolamine and phosphatidic
acid. This indicated that LPS increased the molecular order of these lipids. Their
observation that LPS decreased the width of the phase transition (thus increasing the
cooperativity) supports this conclusion. Rottem (161) found that LPS reduces the
mobility of the acyl chains in membranes by using spin-labeled phospholipids and
electron spin resonance. A series of studies by Jacobs and co-workers (121, 123, 124)
support this result. Using fluorescence depolarization of a lipid probe, they demonstrated
that LPS decrease the ‘fiuidity’ of phospholipid membranes. Interestingly, an NMR
spectroscopy study showed that LPS increased the acyl chain order of phospholipids in

the liquid crystalline phase but decreased the acyl chain order in the lamellar-gel phase

75
(162). Other investigators have measured the degree of acyl chain motion of the

membranes of intact macrophages (163, 164) and hepatocytes (165). These studies have
also noted a decrease in the acyl chain motion of the cellular membranes in response to
LPS. However, these changes were measured several hours after LPS was added.
Therefore, the possibility that the change in the physiological state of the cell were
responsible for these changes, and not LPS directly, could not be excluded. All of these
studies support the general conclusion that LPS is capable of reducing the acyl chain
motion and increasing the degree of order of lipids.

Other studies have analyzed the lateral order of LPS in phospholipids in an
attempt to understand the organization of the outer membrane of Gram-negative bacteria.
Fried and Rothfield (166) studied the changes in the equilibrium surface pressure of
phospholipid monolayers induced by LPS. Their results were consistent with a
homogeneous lipid-LPS mixture. A similar conclusion was reached in a 2H-NMR
spectroscopy study on LPS-phospholipid systems (162). However, reports by Takeuchi
and Nikaido (167, 168) contradict these findings. Using electron spin resonance and
spin-labeled phospholipids, they found that LPS aggregated together in stable domains in
the plane of the membrane. The conflicting results from these investigations leave the
issue of the lateral organization of phospholipid-LPS systems in question.

While the previous investigations on the effects of LPS on membrane properties
have led to the general conclusion that LPS decreases the ‘fluidity’ of membranes, they

provide only little molecular details of these changes. Additionally, all but one of these
studies (162) used lipid probes to investigate the membrane properties. These results are

limited to the region immediately around the probe and may not be representative of the

76
entire membrane. Furthermore, the effect of LPS on the lateral order of membranes

needs to be addressed. In this report, I investigate the effect that LPS has on
phosphatidylcholine model membranes. ,1 use a variety of biophysical techniques to
provide detailed information about how LPS affects the order and dynamics of
membranes. These include fluorescence depolarization, x-ray diffraction and lH-NMR
spectroscopy. Emphasis is placed on lH-NMR spectroscopy since this technique can
look at specific regions within the membrane and can be used to understand the
molecular environment around nuclei. NMR spectroscopy measures all the lipids in the
membrane and is not limited to the environment around a lipid probe. Furthermore, spin-
lattice relaxation times (T 1) of the nuclei are sensitive to their degree of motion and are
an excellent way to characterize the dynamic state of membranes. Through these
techniques, I demonstrate that LPS dramatically affects both the order and the dynamics

of membranes.

Materials and Methods

LPS Isolation and Purification. Lipopolysaccharide from Salmonella
typhimurium (Sigma) was purified by gel filtration on a Sepharose 4B column using
0.05M ammonium formate, pH 5.5, as the buffer. The first peak eluting off the column
was positive for 3-deoxy-D-manno-octulosonic acid (KDO) (136) and was pooled,
dialyzed against distilled H20 and lyophilized. Rhizobium meliloti 2011 was grown in
modified Bergensen’s medium (pH 6.9) and harvested in late exponential phase. Cells
were harvested by centrifugation at 8,000Xg for 15 min. The cell pellet was vigorously

stirred for 12 hrs with water:chloroform:methanol:butanol (4:2:lzl) mixture at room

77
temperature. After centrifugation, the aqueous phase was removed and extracted

sequentially with chloroform and butanol. The aqueous phase (containing LPS) was
treated with a solution of DNase I, RNase A, and MgCl2 to remove contaminating nucleic
acids. R. meliloti 2011 LPS was further purified by gel filtration chromatography on a
Sepharose 4B column as described above. Column fractions were monitored for total
carbohydrate content by the phenol-sulfuric acid assay (137) and 3-deoxy-D-manno-2-
octulosonic acid (KDO) by the periodate/thiobarbituric acid assay (136). The LPS was
located in the carbohydrate peak that contained large amounts of KDO. LPS was
dialyzed against distilled water, concentrated by rotary evaporation, and lyophilized.
Rhizobium trifolii ANU843 was grown in modified Bergensen’s medium (pH 6.9) and
harvested at late exponential phase as described above. R. trifolii ANU843 LPS was
extracted by aqueous phenol as previously described (140). The resulting LPS solution
was dialyzed extensively against distilled H20, treated with DNase I and RNase A, and
purified by gel filtration chromatography as described above.

Lipids and Reagents. 2-(3-(diphenylhexatrienyl)propanoyl)-1-hexadecanoyl-sn-
glycero-3-phosphocholine (DPH-PC) was obtained from Molecular Probes, Inc. (Eugene,
OR) and was dissolved in 100% ethanol at a concentration of 0.05 mg/ml. Phosphate
buffered saline (lOmM phosphate, 0.15M NaCl, pH 7.0) containing 0.9mM CaCl2,
2.7mM KCl and 0.5mM MgCl2 (Dulbecco’s PBS, D-PBS) was prepared. L—or-
phosphatidylcholine (PC), type XVI-E from fresh egg yolk, 99%, was purchased from
Sigma.

Fluorescence. Samples for fluorescence measurements were prepared by drying

0.125 mg phosphatidylcholine (PC) and DPH—PC in a vial. DPH-PC was used at a 500:1

78
PCzDPH-PC molar ratio. The lipids were hydrated in water and the appropriate amount

of LPS (in water) added. The samples were then sonicated for 5 minutes with a bath type
sonicator, frozen and lyophilized to dryness. Samples were rehydrated in 0.5 ml

Dubecco’s PBS (D-PBS), heated to 60°C and vortexed. The samples were then frozen,

reheated to 60°C and vortexed. This freeze-heat-vortex cycle was repeated several times.
Large unilamellar vesicles (LUVs) were prepared by passing the samples 19 times
through a liposome extruder equipped with a 100 nm filter (Avisten, Inc., Ottawa, ON).
The LUV preparation was diluted 5 fold with D-PBS to a final volume of 2 ml.
Fluorescence measurements were made on a SLM 4000 spectrofluorometer (Urbana, IL)
with detectors in a T-format and an excitation frequency of 381 nm. A Schott 418 kV
bandpass filter was used to remove scattered excitation light. The anisotropy of the
DPH-PC probe was calculated according to the following:
r = (I. - 1.0/(I. + 211)

where r is anisotropy, and I. and I i are the intensity of the parallel and perpendicular
fluorescence emission with the excitation light polarized in the parallel orientation. All
anisotropy values were corrected for scattered light by using an identical sample without
DPH—PC added. The temperature was controlled with a circulating water bath and a
jacketed cell. The temperature was measured using a Fluke digital thermometer and a

thermocouple placed directly in the sample. All measurements were made at 37.0 :1:

0.2°C.
Nuclear Magnetic Resonance Spectroscopy. Samples for NMR analysis were

prepared by adding PC (1 mg) in ethanol and drying the solution under a nitrogen stream.

The PC was hydrated in 0.6 ml D-PBS and LPS (in water) added. The sample was

79
sonicated thoroughly with a bath type sonicator, vortexed extensively, frozen and

lyophilized to dryness. The sample was taken up into 0.6 ml D20 and subjected to three
freeze-heat-vortexed cycles. The sample was then transferred to a NMR tube and sealed
under nitrogen. 1H-NMR spectra were recorded on a Varian 300 MHz spectrometer.
Spin-lattice (T1) relaxation times were measured by the inversion recovery technique

(169) using a delay-180°-t-90°-acquire pulse sequence where the time, t, was varied and

the delay was 2 4 seconds.

X-Ray Diffraction. X—ray diffraction patterns were recorded on a Rigaku rotaflex
rotating anode operating at 45 kV and 100 mA. The raw x-ray beam was collimated with
1/6° slits and the copper Km line (1.54 A) was selected by a monochromator. Samples for
x-ray analysis were prepared by drying PC, hydrating it into H20, and adding S.
typhimurium LPS. The PC-LPS mixture was sonicated several minutes, frozen, and
lyophilized to dryness. The lipids were rehydrated in D—PBS (2 60% D-PBS by weight)
and heated briefly to 60°C to facilitate hydration. Samples were placed into 0.7 mm glass
capillaries (Charles Supper Co.) and flame sealed. Capillaries were mounted in the x-ray
beam with a home-built capillary holder. Diffraction patterns were obtained in the
transmittance mode. All diffraction patterns were obtained at room temperature (22°C).

A one-dimensional scanning detector was used to measure the diffracted radiation.

Results and Discussion
Fluorescence Depolarization Analysis of PC-LPS Membranes. A series of
investigations by Jacobs and co-workers (121-124) used fluorescence depolarization to

examine the effects of LPS on membranes. They found that LPS caused a small decrease

80
in the ‘fluidity’ of the membranes. To date, these studies may be the most thorough and

comprehensive examination of the physical properties of phospholipid-LPS mixtures.
However, phospholipids in small unilamellar vesicles and diphenylhexatriene (DPH) as
the fluorescent lipid probe were employed in these studies. Small unilamellar vesicles
are severely strained due to their large curvature and may not exhibit the same properties
as non-strained membranes (170). Furthermore, DPH partitions into several regions
throughout the acyl chain region of lipid bilayers (171), thereby making the interpretation
of the results less clear.

I circumvented these issues by using large unilamellar vesicles and DPH that is
covalently linked to PC and is held in a defined position in the membrane. In addition,
we measured the effects of R. trifolii ANU843 LPS and R. meliloti 2011 LPS (as well as
S. typhimurium LPS) on phosphatidylcholine membranes. The unique structures and
biological activities of the LPS from R. trifolii and R. meliloti allow us to examine the
possible correlation between the biological activity and the change in acyl chain
dynamics induced by these molecules. The change in the anisotropy of the fluorescent
probe as a function of LPS concentration is shown in Figure 3.1. In this experiment, an
increase in the anisotropy of the fluorescent probe indicates an increase in the rigidity of
the membrane in the environment around the probe. Since the probe is anchored to PC
and is located in the fatty acid region of the bilayer, the motion sensed by it is primarily
due to the acyl chain mobility of the lipids. All LPS tested increased the anisotropy
indicating that they all decreased the mobility of the acyl chains of membranes in which
they are incorporated. R. meliloti LPS was the most effective at reducing the acyl chain

motion while R. trifolii and S. typhimurium LPS had almost equivalent effects. As seen

81

 

 

 

 

0.185
—D— Salmonella typhimurium LPS T
0-18‘ -0— Rhizobium meliloti LPS 1
——O-— Rhizobium trifolii LPS
1.
>. 0.175 J
§
0 T
a 0.17 "
<3
.I.
1- I
0.165 " 1- .
1. l
0.16'!
0.155 I l T I l I
0 0.01 0.02 0.03 0.04 0.05 0.06
Amount of LPS (mg/ml)

Figure 3.1: Fluorescence anisotropy of DPH-PC in phophatidylcholine membranes (0.05
mg/ml) as a function of LPS concentration.

82
from the results in Chapter 2, R. trifolii and R. meliloti LPS have a similar level and

spectrum of biological activities. S. typhimurium LPS, on the other hand, is much more
toxic and pyrogenic and is more effective at eliciting the production of TNF than the
rhizobial LPS. These results demonstrate that there is not a simple linear relationship
between the acyl chain mobility change induced by LPS (as judged by fluorescence
anisotropy) and its biological activity. This is similar to the results of Yeh and Price
(122).

The fluorescence depolarization measurements demonstrated the need for a more
careful and detailed analysis of the effects of LPS on membranes. To accomplish this,
we used x-ray powder diffraction and proton NMR spectroscopy.

X-Ray Difiraction Analysis of PC and PC-S. typhimurium LPS Membranes. X-
ray diffraction is a very powerful tool for studying the organization of lipids. In x-ray
diffraction, only regular repeating planes of atoms (lattices) diffract the radiation. The
angle of the reflection (diffracted x-rays) is governed by Bragg’s Law:

n)» = 2dsin0 ( 1)
where n is the order of the reflection, 7t is the wavelength of x-radiation, d is the spacing
of the repeating unit, and 0 is the angle of the incident and reflected radiation. Lipids
typically have two types of repeating structures. The first is due to the bilayer repeat
(large (1) and gives low-angle reflections (small 0). The second is due to the packing of
the acyl chains of the lipids (small d) and gives rise to wide-angle reflections (large 0).
The wide-angle region is useful in determining the order of the acyl chains. Acyl chains
which are rigid (all-trans conformation), such as found in gel phases, give sharp wide-

angle reflections. The other extreme is found in acyl chains that have many gauche

83
conformations and have a large degree of motional freedom (fluid-like phases). Acyl

chains in this state give broad wide-angle reflections.
The diffraction pattern of phosphatidylcholine (PC) in PBS at room temperature is

shown in Figure 3.2. Three equally spaced low-angle reflections (20 = 142°, 282°, and

436°) are indicative of a lamellar phase with a lamellar repeating distance (d) of 62.2 A.

The wide-angle (acyl chain) reflections of PC were broad and show that the acyl chains
were not highly ordered. Therefore, the PC membranes were in the fluid-like, lamellar
liquid crystalline phase. Addition of S. typhimurium LPS (33% by weight) did not alter

the phase of the PC membranes (Figure 3.3). Three discernible lamellar reflections were

observed with a similar lamellar repeating distance (64.0 A) as PC membranes. The PC-

LPS diffraction pattern also had broad wide-angle reflections thereby showing that the
membranes were in the lamellar liquid crystalline phase.

Upon closer inspection of the wide-angle region, it is obvious that the addition of
LPS resulted in significant changes in the acyl chain order of the PC membranes (Figure
3.4). The wide-angle region of both PC and PC-LPS consisted of two overlapping
reflections and indicated that the lipids had two acyl chain packing distances. The wide-
angle region was deconvoluted into its component reflections (using Lorentzian
functions) to aid in its analysis. The parameters of the deconvoluted reflections are given

in Table 3.1. Addition of S. typhimurium LPS to PC did not induce any significant

change to the 3.24 A reflection. However, LPS caused a dramatic decrease in the
intensity (area) of the 4.47 A acyl chain reflection. It also shifted the 4.47 A reflection to

a smaller repeating distance (4.30 A) thereby demonstrating that LPS decreased the acyl

84

.maouooce Ewcméfi 8865 @505
.mmEAH E oczonoacumnamonm mo 58qu 538.5% aarx ”Wm Emmi

85

3 2am

am
on mm om mm om m“ on m

FPL»F____P_L~PP»P~___._F_»_L—r_._»_._prr_r_._.ppbr—PL._pp._rwi—.P»_»_._._p_._

 

 

 

 

 

 

 

Kirsuoml [eufirs

86

.wGOSOOEDu

23-32 2865 :35 .349 a same was 553238 .m
I @th cam—ozoEEEH—amonm no E839 coconut? aarx ”mum 823m

87

3 Beam

2..
on mm om mm om ms 02 m

____L_____.—»P_—___—e_r—b___~—__pp_.___—____»~.____CPF_LPL~L_L_p_p_._r—L_~_p—

 

 

 

 

 

 

KJESUOWI leufiis

88

Figure 3.4: Wide-angle region of the x-ray diffraction patterns
from phosphatidylcholine (A) and 2:1 phosphatidylcholinezs.
typhimurium LPS (B). Solid lines are the deconvoluted signals
and the sum of the deconvoluted signals.

60

-20

60

40

Intensity

-20

 

 

I

 

 

 

10 20 30 40
20

Figure 3.4

90

Table 3.1: Parameters of the deconvoluted wide-angle reflections.

 

 

 

 

Reflection 1 Reflection 2
SampleI d (A) Peak Area d (A) Peak Area
Width2 Width2
PC 4.47:0.01 6.2:0.2 394:17 3.24:0.01 12.7:0.3 825:22
PC-LPS3 4.30:0.04 7.8:0.5 232:30 3.26:0.03 12.9w3 816:36

 

 

 

 

 

 

 

 

 

d=repeat distance

lsamples contain 2 60% Dulbecco’s PBS

2peak width is given in 20 degrees

3PC-LPS sample contains a 67% PC/33% S. typhimurium LPS by weight (before D-PBS is
added)

chain packing distance of the membrane. This shows that LPS is increasing the order of
the acyl chain. This correlates well with our fluorescence depolarization data as well as
previous reports (121, 123, 124, 160-165) which indicated an increase in the order of
membranes when LPS is added.

NMR Spectroscopy Analysis of S. typhimurium LPS. The NMR spectrum of pure
S. typhimurium LPS is shown in Figure 3.5. The very intense signal at 4.7 ppm is the H-
O-D resonance. The remaining signals between 3.5-5.5 ppm are primarily due to the
protons on the polysaccharide region of LPS. The methylene protons on the acyl chains
are observed at 1.0-1.4 ppm. The signal for the methyl protons of the acyl chains, which
is normally observed at 0.9 ppm for lipids, is absent. This is the result of severe line
broadening due to dipolar coupling and chemical shift anisotropy which occurs for slowly
moving nuclei (172) and demonstrates that the methyl groups are severely restricted in
their motion.

Of special interest are the signals arising from the methylene protons on the acyl

chains (1.0-1.4 ppm). Normally, for lipids in water, the methylene protons give rise to

one broad s
overlapping r
the methylen
and 1.34 ppm
more differe
spectrometer
local magnet
applied mag
chemical shi

described by

“here 011.

describe the
COSlneg Whi.
(.172), In .
averaged 0‘
(lipids) an (1
depends on
the two Sig}
at 1%“ t“
essentially
demonstral

l

91
one broad signal at approximately 1.3 ppm. These broad signals are due to the

overlapping resonances of most of the methylene protons in the acyl chain. Interestingly,
the methylene protons for S. typhimurium LPS are split into two broad envelopes at 1.22
and 1.34 ppm (Figure 3.5, A). These two signals are due to the protons existing in two or
more different orientations (domains) relative to the magnetic field of the NMR
spectrometer. The chemical shifts of nuclei in NMR spectroscopy are dependent on the
local magnetic field around the nuclei. The local magnetic field shields the nuclei to the

applied magnetic field to differing extents and gives rise to changes in the observed
.chemical shift. The local field is not necessarily uniform in all directions and can be
described by the chemical shift tensor, 0':

0' = 0111112 + 6221222 + (5333-332 (2)
where on, 0'22, and 0'33 are the principle values of the chemical shift tensor which
describe the magnitude in the three Cartesian coordinates and the A’s are the direction
cosines which describe the orientation of the molecule with respect to the applied field
(172). In ‘normal’ solution NMR, the molecules are moving fast and, therefore, 0 is
averaged over all orientations. In rigid poly-crystalline samples, such as liquid crystals
(lipids) and polymers, the molecular motion of the sample is slow and the chemical shift
depends on the orientation of the sample with respect to the magnetic field. Therefore,
the two signals arising from the methylene protons in S. typhimurium LPS must be due to
at least two different molecular orientations (domains) of the LPS. Interestingly,
essentially the same proton spectrum is obtained at any temperature tested (20-70°C),
demonstrating the high stability of the packing arrangement of LPS over a broad

temperature range (data not shown). The packing was perturbed, however, upon the

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IIIYTIfTTI7IETTTTITITTTIITfrItIIIIIIIIYlTerlTfijlfrTlllIIIi

5.5 5.0 4.5 6.0 3.5 3.0 2.5 2.0 1.5 1.0 m

Figure 3.5: 'H-NMR spectra of s. typhimurium LPS (37°C) in D-PBS with 0% (A), 25%
(B), and 50% (C) of deuterated methanol added. Note the large change in the methylene
and methyl protons signals (0.8-1.5 ppm) when methanol is added.

addition of
methanol (2
methanol in
methanol a
i indicating tl'.
NMR

order to fu1
membranes,
resultant Nh'
3.6, A) Shov
added ('Figur
emergence 0;
least two on:
pure LPS (F14.
COCXiStence (
COCXiSIed, am
CXpect the rat
at [he highesr
This indicated
is dUe to prm

i therefOTE. Clez

d .
Onlaln sullen

PrOIOn

93
addition of deuterated methanol to the sample (Figure 3.5, B-C). Addition of deuterated

methanol (37°C) completely altered the methylene signals. As the concentration of
methanol increased, the methylene signals began to merge into one broad resonance. The
methanol also dramatically increased the terminal methyl proton signal (0.9 ppm)
indicating that it decreased the rigidity of these protons.

NMR Spectroscopy and T1 Analysis of PC -S. typhimurium LPS Membranes. In
order to further examine the effects of S. typhimurium LPS on the properties of
membranes, we analyzed PC and PC-LPS vesicles by proton NMR spectroscopy. The
resultant NMR spectra are shown in Figure 3.6. The NMR spectrum of pure PC (Figure
3.6, A) showed only one methylene signal at 1.29 ppm. As S. typhimurium LPS was
added (Figure 3.6, B-E), we observe a shift of the methylene signal to 1.34 ppm and the
emergence of a second methylene signal at 1.22 ppm. This indicates the presence of at
least two different packing states in these lipid mixtures, analogous to the spectrum of
pure LPS (Figure 3.5, A). The two methylene signals can not be explained by the simple
coexistence of non-interacting LPS and PC membranes. If these two lipids simply
coexisted, and assuming that the PC methylene proton signals are at 1.34 ppm, we would
expect the ratio of the areas of the 1.34 ppm/1.22 ppm signals to be approximately 13.4/1
at the highest concentration of LPS (Figure 3.6, E). Instead we observe a ratio of 1.2/ 1.
This indicated that a large portion of the 1.22 ppm signal, as well as the 1.34 ppm signal,
is due to protons on the acyl chains of the PC. The splitting of the methylene signal,
therefore, clearly demonstrates the ability of LPS to induce a dramatic change in the
domain structure of PC membranes.

Proton NMR spin-lattice relaxation (T1) times of the methylene protons were

 

 

 

rfl[JIFUITTTTTITIllerleIllTT[PITUITTTIIITIIIIFFTIIIITITjjT‘I
6.0 5.5 5.0 0.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 pm

Figure 3.6: 1H-NMR spectra of phosphatidylcholine in D-PBS (37°C) with 0% (A), 20%
(B), 33% (C), 43% (D) and 50% (E) S. typhimurium LPS (by weight) added. Note the
splitting of the methylene proton signal (1.29 ppm) as LPS is added. ‘

95
analyzed in order to investigate the change in acyl chain dynamics induced by S.

typhimurium LPS. The NMR T1 experiments were performed using the inversion
recovery method (169). In these experiments, the magnetization is inverted (180° pulse)
and allowed to recover for variable lengths of time, t. The magnetization that recovers
along the z-axis (parallel to the magnetic field) is then rotated 90° into the x-y plane and
measured. The recovery of the magnetization to equilibrium is exponential and is
described by the equation:
M” - M. = 2Mmexp(-t/T 1) (3)

where M” and M. are the magnetization at equilibrium and at time = t, t is the delay time

and T1 is the spin-lattice relaxation time (172). For spin 1/2 nuclei, dipole-dipole
interaction is the primary mechanism for relaxation (i.e. the interaction of a proton with
another proton). T1 relaxation is dominated by the relaxation events occurring near the
same rate as the NMR spectrometer frequency. For lipids at 300 MHz (~10'9 8), this
corresponds to the trans-gauche isomerization of the carbon-carbon bonds in the acyl
chains. Kroon et al. (173) have demonstrated that the T1 relaxation of lipid acyl chains is
due to the intra-chain proton-proton interaction resulting from the trans-gauche
rearrangement. Hence, the T1 relaxation times are measuring the rate of trans-gauche
isomerization of the lipid acyl chains. For our experiments, the T1 relaxation is in the
fast correlation region where an increase in trans-gauche isomerization rate results in an
increase in T1.

Rearranging equation 3 and substituting the magnetization with the intensity of
the NMR signal (I) we obtain:

ln(I°° - L) =1n(21°°) + -t/T1. (4)

96

For a single exponential recovery, a plot of ln(I°° - L) vs. t will give a straight line with a
slope equal to -1/T 1. Plots of ln(I” - I.) vs. I for the methylene protons of PC and LPS-PC

membranes are shown in Figure 3.7. The one methylene signal observed for pure PC
membranes gives a straight line plot indicating that it is a single exponential recovery
with T1 = 0.54 seconds. None of the samples that contained LPS gave a straight line and
are, therefore, not adequately described by a single exponential recovery. The data for
these samples were fit to a double exponential equation:

I” - I. = Caexp(-t/T1;.) + Cbexp(-t/T lb)] (5)
where Ca and Cb are variable terms used to describe the 21” term in equation 4. All of the

inversion recovery plots for the methylene signals were fit very well by this equation
(correlation coefficients 2 0.999). The T1 values obtained by this analysis are given in
Table 3.2. The values of T1a for the 1.34 ppm peak did not change much from pure PC
membranes demonstrating that LPS had little influence on the trans-gauche isomerization
rate of these protons. At low concentrations of LPS, the T18, values for the 1.22 ppm
signal were very similar to that of the 1.34 ppm signal indicating that both of these
domains had a similar rate of trans-gauche isomerization in their acyl chains. At high
LPS concentrations, Tla for the 1.22 ppm signal began to decrease. Thus, the rate of
trans-gauche isomerization for the protons in this signal is decreased by high levels of
LPS. From the data it can be seen that the second exponential term (which gives le)
was highly variable. This is due in large part to the fact that this exponential term is most
strongly influenced by the data at long recovery times where small errors in the intensity
of the NMR signal will result in a large fluctuations in the T1 value. While we know that

the second exponential term is due to slowly relaxing protons (fast motion), the

97

Figure 3.7: lH-NMR T1 recovery plots for the methylene
protons of PC and PC-S. typhimurium LPS membranes at
37°C. The straight lines are approximate fits to the early
time points and are used to illustrate the non-linear behavior
of the data when LPS is added.

 

 

98

 

 

 

% LPS

0%

 

 

 

 

 

 

 

 

 

5432154321543215432

3-85.:

20%

 

 

$.65

33%

 

 

cases

43 %

 

 

9-3:...

50%

 

1

time (s)

time (s)

Figure 3.7

99

Table 3.2: T1 values for the methylene protons in PC and LPS-PC membranes.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S. typhimurium 1.34 ppm signal 1.22 ppm signal
LPS
(% PC by weight) T1a (s) T1b(s) T1a (s) Th, (5)
0% 054‘ na na na
20% 0.61 3.34 0.54 2.03
33% 0.56 1.47 0.64 -2.48
43% 0.58 9.43 0.49 1.71
50% 0.56 3.80 0.38 0.67
na=not applicable
11.29 ppm

variability of the le values precludes a quantitative examination of this term.

As discussed above, both the 1.34 ppm and 1.22 ppm signals are comprised of
methylene protons from LPS and PC. Therefore, it is impossible to determine
unequivocally whether or not each of the exponential recoveries arise from one of the
lipid species or from a mixed aggregate that has the same motional properties. It is clear,
however, that each of the two methylene signals observed for mixed LPS-PC membranes
contains protons in at least two different motional states.

To obtain a clear understanding of the effects that LPS has on membranes, it is
instructive to compare the fluorescence depolarization and NMR T1 measurements. An
excellent study by Petersen and Chan (11) showed that two types of motions could best
describe acyl chain dynamics. The first is the large motions of the acyl chains induced by
a single gauche conformation (chain reorientation) and occurs at approximately 10'7 s.
The second is from the coupled trans-gauche isomerizations that give rise to kinks in the

acyl chains. Successive trans-gauche(+)/gauche(-)-trans isomerizations result in the

100
diffusion of the kink along the acyl chain and occur at approximately 10"0 s. Since the

fluorescence measurements are sensitive to motions occurring at about 10’8 8, they would
be mainly measuring the chain reorientation motions. Likewise, the NMR T1
measurements (at 300 MHz) are sensitive to motions occurring at about 10'9-10'l0 s and
would be primarily measuring the trans-gauche isomerizations in the acyl chains that
give rise to kink diffusion. In the NMR results, only one of the methylene signals
showed a decrease in T1 in response to LPS and this decrease was relatively small. This
indicates that Salmonella LPS has a relatively small influence on the rate of kink
diffusion in the membrane. The fluorescence experiments, on the other hand,
demonstrated that LPS caused a significant decrease in the chain reorientation motions of
phosphatidylcholine membranes. Thus it appears that LPS alters membrane dynamics
primarily by decreasing the large acyl chain fluctuations. However, this does not account
for the sub-population of methylene protons observed in the NMR T1 measurements that
had an increased rate of acyl chain kink diffusion. We would expect that these regions of
increased motions would decrease the fluorescence anisotropy of the lipid probe,
assuming that the probe partitions into this area of the membrane. However, these
steady-state fluorescence measurements could not distinguish sub-populations of
different acyl chain motions because they are derived from an averaged signal.
Therefore, the different populations of motions measured by NMR T1, including the
increase in the rate of kink diffusion, will be averaged into the fluorescence anisotropy

results.

101

Conclusions

From the fluorescence depolarization, x-ray diffraction, proton NMR, and proton
NMR Tl relaxation measurements, we can see the large influence that LPS has on the
dynamics and order of PC membranes. X-ray diffraction measurements clearly showed
that LPS (at the concentration and temperature tested) did not alter the phase of PC
membranes but did decrease the acyl chain packing distance. The fluorescence
depolarization measurements demonstrate that LPS decreases the acyl chain mobility of
membranes in which it is incorporated. The NMR Tl measurements also showed a small
decrease in the acyl chain mobility induced by LPS. Interestingly, the fluorescence
depolarization measurements utilizing rhizobial LPS indicated that there is not a linear
relationship between acyl chain mobility and its ability to elicit a physiological response.
This suggests that the relationship between membrane physical property changes and
physiological responses elicited by LPS may be very complex and likely involves many
aspects of membrane dynamics and order.

Analysis of the NMR spectra from PC-LPS mixtures indicates that LPS causes a
large change in membrane order. LPS induced the lateral separation of lipids into distinct
domains. The two NMR signals observed for these domains had different responses to
LPS; the 1.22 ppm protons decreased in motion as the concentration of LPS increased
whereas the 1.34 ppm protons did not. Both of these signals, however, were not
consistent with a single exponential recovery (as seen for pure PC membranes) and
therefore are comprised of protons in at least two different motional states. This included
a fast motion component and suggests that LPS induced regions of membrane defects or

regions of different domains.

102
Our results have provided a much more detailed molecular understanding of the

membrane perturbations induced by LPS than was previously known. These results
definitively demonstrate that S. typhimurium LPS causes a dramatic change in both the
dynamics and order of membranes in a very complex manner. Furthermore, they lend
credence to the hypothesis that membrane perturbations are a critical component of the
physiological responses elicited by LPS. Our findings clearly demonstrate that
membrane perturbations are a possible, if not probable, mechanism by which LPS
activates immune cells.

Since LPS presumably trigger a cellular response through alterations in membrane
dynamics and order, it should be expected that other factors that affect membrane
properties could affect this response. Environmental conditions such as temperature,
ionic strength, pH and pressure will affect membrane properties. Lipophilic molecules
such as solvents and lipids will also induce membrane perturbations. If LPS act at the
membrane level, then we should expect that environmental conditions would alter its
ability to induce a cellular response. Moreover, the membrane perturbations elicited by
LPS can be thought of as a stress or shock to the cell, similar to other environmental
conditions such as temperature. Interestingly, there is a link between thermal shock in
macrophages and their ability to produce cytokines. Several investigators demonstrated
that heat shock decreases the production of tumor necrosis factor (TNF) from
macrophages (174-176). Cold shock, on the other hand, enhances the production of TNF
and interleukins from LPS stimulated macrophages (177, 178). These results indicate
that LPS effects and thermal effects may be related. This is consistent with a role of

membrane perturbations in the mechanism of LPS action.

Chapter 4

Probing the Mechanism of Lipopolysaccharide Action with Synthetic Lipids

103

104
Introduction

Lipopolysaccharides (LPS) are complex glycolipids located on the outer leaflet of
the outer membrane of Gram-negative bacteria. LPS are capable of eliciting a wide range
of biological activities when they interact with the immune system of animals and are
responsible for the deleterious effects of Gram-negative septicemia. These biological
activities include the induction of synthesis of tumor necrosis factor (TNF) and
interleukins by macrophages, the induction of the proliferation of B-cells, and the
activation of the clotting enzyme cascade of the lysate of amebocytes from the horseshoe
crab (see Chapter 1).

The critical structural elements of LPS that determine biological activity are
found in the lipid A region of the molecule. There is a very fine relationship between
lipid A structure and the various biological activities of LPS. The structures of the fatty
acids are critical elements of this relationship. For example, the lipid A head group might
be completely exposed, as in free lipid A, or covered with hundreds of carbohydrate
groups containing substituents such as phosphate groups (smooth form
lipopolysaccharide) without really influencing its biological activity (73, 89). However,
the insertion of double bonds into fatty acids and modification of fatty acid chain lengths
are sufficient to completely abolish endotoxin activity. The lipid A of Rhodobacter
sphaeroides is identical in structure to that of the Enterobacteriaceae except that some of
the fatty acyl chains are shorter and one is unsaturated (95), yet this lipopolysaccharide is
completely devoid of endotoxin activity (95—99).

In 1987, Jacobs and Price (121) proposed a ‘two-step’ model to explain the

mechanism by which lipid A interacts with and activates immune cells. This model

105
attempts to bring together the seemingly distant views of receptor mediated versus

membrane mediated events in the activation of immune cells by LPS. This model was
based, in large part, on studies that showed that LPS interacted with cells in two distinct
steps. First, lipid A (or LPS) binds to the membrane of the host cell. Binding is
proposed to involve an electrostatic interaction between lipid A and the cell surface.
These authors suggest that a membrane-associated protein (LPS receptor) may mediate
this interaction. Next, lipid A inserts (fuses) into the membrane of the immune cell. The
insertion of lipid A into the membrane results in the activation of the immune cell.

After LPS inserts into the host cell membrane, it induces a signal that causes the
cell to become more active and, for instance, secrete cytokines. While the origin of this
signal is still undefined, it may come from LPS-induced perturbations in the properties of
the host cell membrane. This contention is supported by our results (Chapter 3) which
clearly demonstrate that LPS are capable of dramatically altering the dynamics and order
of membranes. The ability of LPS to affect lipid dynamics and order would be dependent
on its molecular and aggregate structure. Manipulating LPS in such a way as to change
its molecular order or alter its ability to form highly ordered aggregates would alter its
effects on the properties of membranes. If membrane perturbations were involved in the
mechanism of LPS action, these changes should affect its biological activities. In fact,
the biological activities of LPS are highly dependent upon its structural and
physicochemical properties (7, 125, 126, 129). Furthermore, we should expect that
molecules that alter these properties of LPS or offset the membrane perturbations induced

by LPS would drastically alter the biological activities of LPS.

106
In 1994, our lab reported the synthesis of two positively charged lipids that were

designed to interact with the negatively charged LPS (179). These lipids were found to
block the ability of LPS to activate the clotting enzyme cascade of the horseshoe crab
(Limulus) amebocyte lysate (LAL assay) and were named endotoxin inhibitors (El-l and
EI-2). The mechanism by which EL] and EI-2 were able to block the activity of LPS
was not determined. However, because these lipids were designed to interact with LPS,
it was suggested that they disrupt the high degree of order of LPS aggregates and thus
alter the ability of LPS to induce membrane perturbations. Furthermore, the ability of E1-
1 and EI-2 to disrupt cells at high concentrations led Lill-Elghanian to suggest that these
lipids may act as membrane ‘fluidizers’ (133).

In this report, I am using RH and EI-2, as well as three additional lipids that were
synthesized (El-3, EI—4 and EI-S), to investigate the mechanism by which LPS activates
immune cells (Figure 4.1). These five synthetic lipids can be viewed as molecular probes
to manipulate the interaction between LPS and immune cells. By understanding how the
five synthetic lipids affect the biological activity of LPS and how they alter the LPS-host
cell interaction, we may gain an insight into the mechanism of LPS action. The
synthesis, purification, structural characterization, and physical properties of EI—3, EI—4
and EI-S are described later (Appendix I). Here I describe the effects that these five
lipids have on the biological activities of LPS (LAL assay and TNF-0t induction from
macrophages) and their ability to alter the dynamics and order of LPS-phospholipid

membranes.

107

mtmm

22a: 6:05.25 95 05 we mEBosbm

vtmm

”2. 2:5

mrmm

0
=2./.
2
o \ z\) o
:2 z:

mumm—

, \
212
\m N]
O .0

fl .5

\I
l\

O
O

>(

108
Materials and Methods

LPS, Lipids and Reagents. Lipopolysaccharide from Salmonella typhimurium
(Sigma) was purified by gel filtration chromatography on a Sepharose 4B column using
0.05M ammonium formate, pH 5.5, as the buffer. The first carbohydrate-containing peak
(137) eluting off the column was positive for 3-deoxy-D-manno-octulosonic acid (KDO)
(136) and was pooled, dialyzed against distilled H20 and lyophilized. EI-l and EI-2 were
prepared previously (179). 2-(3-(diphenylhexatrienyl)propanoyl)-l-hexadecanoyl-sn-
glycero-3-phosphocholine (DPH-PC) and trans-parinaric acid (tPnA) were obtained from
Molecular Probes, Inc. (Eugene, OR) and were dissolved in 100% ethanol at a
concentration of 0.05 mg/ml. Phosphate buffered saline (10mM phosphate, 0.15M NaCl,
pH 7.0) containing 0.9mM CaCl2, 2.7mM KCl and 0.5mM MgCl2 (Dulbecco’s PBS, D-
PBS) was prepared. bet-phosphatidylcholine (PC), type XVI-E from fresh egg yolk,
99%, was purchased from Sigma. Synthetic l,2-dipalmitoyl-sn-glycero-3-
phosphocholine (DPPC) was obtained from Avanti Polar Lipids, Inc., Alabaster,
Alabama.

Fluorescence. Samples for fluorescence measurements of PC-LPS membranes
were prepared by drying 0.125 mg phosphatidylcholine (PC), the synthetic lipids, and
DPH-PC in a vial. The synthetic lipids were added at the molar ratio of synthetic
lipid/LPS indicated, assuming a molecular weight of LPS of 15000 g/mole. DPH-PC was
used at a 500:1 PC2DPH-PC molar ratio. The lipids were hydrated in water and 0.125 mg
of LPS (in water) added. The samples for DPPC and DPPC-LPS membranes were
prepared in the same way except that DPPC was substituted for PC and tPnA was used as

the fluorescent membrane probe at a DPPC:tPnA molar ratio of 200:1. All of the samples

109
were sonicated for 5 minutes with a bath type sonicator, frozen and lyophilized to

dryness. Samples were rehydrated in 0.5 ml Dubecco’s PBS (D-PBS), heated to 60°C
and vortexed. The samples were then frozen, reheated to 60°C and vortexed. This
freeze-heat—vortex cycle was repeated several times. Large unilamellar vesicles (LUVs)
were prepared by passing the samples 19 times through a liposome extruder equipped
with a 100 nm filter (Avisten, Inc., Ottawa, ON). The LUV preparation was diluted 5
fold with D-PBS to a final volume of 2 ml. Fluorescence measurements were made on a
SLM 4000 spectrofluorometer (Urbana, IL) with detectors in a T—format and an
excitation frequency of 381 nm for DPH-PC or 320 nm for tPnA. A Schott 418 kV
bandpass filter was used to remove scattered excitation light. The anisotropy of the
DPH-PC probe was calculated according to the following:
r = (1' - I i)/(Ill + 21 J_)

where r is anisotropy, and I. and Il are the intensity of the parallel and perpendicular
fluorescence emission with the excitation light polarized in the parallel orientation. All
anisotropy values were corrected for scattered light by using an identical sample without
the fluorescent lipid probe added. The temperature was controlled with a circulating
water bath and a jacketed cell. The temperature was measured using a Fluke digital
thermometer and a thermocouple placed directly in the sample. Sample temperatures
were :1: 0.2°C from the reported value.

Nuclear Magnetic Resonance Spectroscopy. Samples for NMR analysis were
prepared by adding PC (1 mg) and the synthetic lipids and drying the solution under a
nitrogen stream. The lipids were hydrated in 0.6 ml D-PBS and LPS (in water) added.

The sample was sonicated thoroughly with a bath type sonicator, vortexed extensively,

110
frozen and lyophilized to dryness. The sample was taken up into 0.6 ml D20 and

subjected to three freeze-heat-vortex cycles. The sample was then transferred to an NMR
tube and sealed under nitrogen. lH-NMR spectra were recorded on a Varian 300 MHz
spectrometer. Spin-lattice (T1) relaxation times were measured by the inversion recovery

technique (169) using a delay-180°-t-90°-acquire pulse sequence where the time, t, was

varied and the delay was 2 4 seconds.

Cell Lines. RAW264.7, a murine Abelson leukemia virus-transformed
lymphocyte with macrophage properties (141), was obtained from American Type
Culture Collection. RAW264.7 was grown in Dulbecco's modified Eagle's medium
(DMEM) supplemented with 10% fetal bovine serum and 100 mg/L streptomycin sulfate
(culture medium). Subcultures were prepared by scraping and resuspending in fresh
culture medium. Cells were grown at 37°C in a humidified, 5% C02 atmosphere.

Tumor Necrosis Factor-a Assay. S. typhimurium LPS and the inhibitors (in
ethanol) were added to 1 ml of distilled water and sonicated. Samples were lyophilized,
rehydrated in Dulbecco’s PBS and then sonicated and vortexed thoroughly. RAW264.7
cells (2 x 105) were placed in 2 ml of culture medium in 24 well culture plates and
allowed to grow for 15 hrs. Samples (50 pl) were added to the cell culture wells and
allowed to incubate with the cells at 37°C, 5% C02. Cells were allowed to incubate with
the samples for 8 hrs since a time course experiment determined that the maximal TNF
response occurred at this time. After incubation, cell culture supematants were removed,
centrifuged to remove cell debris, and placed on ice. The TNF-0t concentration in the

culture supernatant was determined by using a mouse TNF-or ELISA kit (Genzyme

Corp., Cambridge, MA). TNF-0t levels were quantitated with recombinant mouse TNF-0t

111
according to the test kit instructions. The results shown are the average of duplicate

determinations and the error bars show the range of the results.

Limulus amebocyte lysate (LAL) assay. The LAL assay was performed with the
Whittaker Bioproducts QCL-1000 Quantitative Chromogenic LAL, Whittaker
Bioproducts, Inc. (Walkersville, MD). All assays were standardized to Endotoxin Units
(EU) using E. coli 0111:B4 LPS according to the assay instructions. Samples were
prepared by adding S. typhimurium LPS and the inhibitors (in ethanol) to distilled water.
The samples were sonicated and then lyophilized to dryness. The samples were
rehydrated in distilled water and sonicated and vortexed thoroughly immediately prior to
being assayed. The final concentration of S. typhimurium LPS was 0.1 ng/ml. Results
shown are the mean of at least three determinations and the error bars indicate the

standard deviation.

Results and Discussion

In Chapter 3, I demonstrated that LPS has a dramatic influence on the dynamics
and order of membranes. These results are consistent with the idea that LPS-induced
membrane perturbations are a critical component of its ability to elicit a severe immune
response in animals. However, little direct evidence exists to support the correlation
between membrane property changes induced by LPS and its ability to elicit a biological
response. The discovery that some positively charged synthetic lipids (EI—l and EI-2)
could block the biological response induced by LPS (179) suggested that these lipids
could be used to probe the mechanism of LPS action. Lill-Elghanian proposed that EI-l

and EI—2 might act as membrane ‘fluidizers’ (133). To examine this possibility further,

112
we synthesized three additional lipids that all have a different number of acyl chains

(Figure 4.1). By increasing the number of acyl chains in a lipid, we will increase the van
der Waals interactions within this lipid and thus, increase the molecular order of the lipid.
In turn, this will alter its effects on the properties of LPS and membranes.

The effects of the five synthetic lipids (EI-l through EI-5) on the biological
activities of S. typhimurium LPS were determined by their ability to reduce LPS
activation of the LAL assay and their ability to reduce LPS-dependent production of
tumor necrosis factor-or (TNF-0t) from a macrophage cell line. The LAL assay (Figure
4.2) showed conclusively that all of the synthetic lipids were able to reduce the activation
of the LAL cascade (using 0.1 ng/ml S. typhimurium LPS). EI—l and EI-2 were the most
potent inhibitors and completely blocked the LAL cascade at 100 ng/ml. EI—3, E14, and
EI-5 were less effective inhibitors and at 100 ng/ml gave 85%, 46%, and 50% inhibition,
respectively. EI—3, EI—4, and EI-5 were not able to elicit an LAL response by themselves.
EL] and EI-2 have been shown previously to be inactive in the LAL assay
(Hollingsworth and Lill-Elghanian, unpublished data). Using the same concentrations as
tested in the LAL assay, the synthetic lipids were also tested for their ability to inhibit the
production of TNF-or from RAW264.7 cells induced by 0.1 ng/ml S. typhimurium LPS
(Figure 4.3). RAW264.7 cells are a macrophage-like cell line known to be activated by
LPS (141). All of the synthetic lipids, with the exception of EI-4, were able to
completely block the induction of TNF-oz at the concentrations tested. The TNF-0t
induction assay also showed that EI—l and EI-2 did not alter the basal level of TNF-0t

produced. The decrease in TNF-0t production seen for the LPS treated cells was not due

to a nonspecific toxicity of the lipids since the basal level of TNF-or production did not

% Activity

é

llLLlLllelLULllellllljllllllllllLLLLLlllllllJll

\O
O

Nw-hUIQQOO
OOOOOOO

pa
0

O

113

 

 

EI-l (10 ng/ml) + LPS

EI-l (100 ng/ml) + LPS

131-2 (10 ng/ml) + LPS

 

EI-2 (100 ng/ml) + LPS

EI-3 (100 ng/ml) + LPS

EI-4 (100 ng/ml) + LPS

 

 

131-5 (100 ng/ml) + LPS

 

Figure 4.2: Effect of the synthetic lipids on the activity of 0.1 ng/ml S. typhimurium LPS
in the Limulus amebocyte lysate (LAL) assay. All values are reported as the percent
activity relative to 0.1 ng/ml S. typhimurium LPS.

114

 

 

2m + name 8: Em

 

F-

.. ms: + as»: 8: Em

 

. mi +35»: 8: 2m

 

I $3 + name 8: Sm

 

.. 2m + in: c: Sm

 

.. $5 + meme 8: Em

 

.. $3 + can: as Em

 

r GEE: fiov ma: gouge—mm

 

.. in: 8: Em

 

.. 2an o: Em

 

. can: 8 : Em

 

l.

.. can: o: Em

 

TITI.

. ammé 3.28

 

 

3500

duJiqiquW—iqdqq—qfl

mam

32

name somzm.

dq—dq.u—qqquqq
m m m
5 5
1 1

a

 

0

Figure 4.3: Tumor necrosis factor—0t (TNF-0t) assay. Samples that contain LPS have 0.1
ng/ml S typhimurium LPS

115
change upon the addition of EI-l and EI-2. Furthermore, cell viability (as judged by

trypan blue exclusion, Figure 4.4) and cell morphology was not altered after a 10 hour
incubation with the synthetic lipids.

One effect of LPS on the properties of membranes that I (Chapter 3) and others
(121-124) have observed is their ability to decrease the acyl chain motion (as measured
by fluorescence depolarization). If the synthetic lipids were simply acting to increase the
acyl chain motions of membrane (i.e. act as membrane ‘fluidizers’), then we should be
able to observe this by fluorescence depolarization. To test this I used a 1:1 PC:S.
typhimurium LPS mixture (by weight) and measured the effects of the synthetic lipids on
the anisotropy of a fluorescent lipid probe, DPH-PC (Figure 4.5). All five synthetic
lipids increased the anisotropy of the membranes (El-5 > EI—3 > EI—4 > EI-2 > EI-l).
This demonstrated that the synthetic lipids have the opposite effect than was expected.
They all decrease the acyl chain motion of PC-LPS membranes. My results clearly show
that these synthetic lipids do not block the biological activity of LPS by increasing the
rate of the large chain reorientation motions of membranes. This indicates that there is
not a simple relationship between acyl chain dynamics and biological activity. It is
noteworthy, however, that there did appear to be a rough correlation between the number
of acyl chains of the synthetic lipids and their ability to change the membrane acyl chain
dynamics. For example, the hexa-acyl lipid (El-5) is the most effective at reducing acyl
chain motions while the di-acyl lipids (El-1 and EI-2) are the least effective.

Proton NMR spectroscopy was used to further examine the effects of the five
synthetic lipids on the dynamics and order of PC-S. typhimurium LPS mixtures. The

resulting NMR spectra are shown in Figures 4.6-4.10. The NMR spectra show that all

116

00
O

 

Utmxl
COO

 

O
Lllllllljllfillllllllh]lllllllllllllIll

N

Number ofViable Cells (2 x 104)
0 ES o 3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

131-1 (100 ng/ml) + LPS
131-2 (100 ng/ml) + LPS
131-3 (100 ng/ml) + LPS
131-4 (100 ng/ml) + LPS
EI-S (100 ng/ml) + LPS

Salmonella LPS (0.1 ng/ml)

Figure 4.4: Effects of S. typhimurium LPS and synthetic lipid-LPS mixtures on
RAW264.7 cell viability as judged by trypan blue exclusion. The results are the average
of duplicate counts and the error bars show the range of the results.

117

 

dosabcooag Em: ocmfiim mo menace
a a some; 3 one w: mam Sargasso mam
E 0%me .«o Seaman ooaoomomoafi ”we 9.sz

118

no 2&5

6.353: one: use:

2 w c v

 

 

 

 

 

 

 

 

 

 

.(donosruv

119
the synthetic lipids caused an increase in the methyl and methylene proton signals. EI-l,

EI-2 and E15 resulted in a large increase in the NMR signals while both EI-3 and EI—4
gave only a modest increase. The two methylene proton signals (1.34 and 1.22 ppm)
were deconvoluted and the areas of the signals analyzed. The percent change in signal
area was calculated and the results are shown in Table 4.1. For the PC—LPS membranes
with EI-l, EI-2, and EI—S added, the percent increase of the two methylene signals is
much larger than would be predicted. For example, at a 10:1 EI-lzLPS molar ratio, 0.556
mg of EI-l was added to a mixture of 1.0 mg PC and 1.0 mg LPS. The addition of EI—l
will give approximately a 42% increase in the number of methylene protons in the
sample. However we observe an 81% increase in the total methylene proton signal
(102.7% in the 1.34 ppm signal and 42.4% in the 1.22 ppm signal, Table 4.1) in the NMR
spectrum (Figure 4.6, D).

To understand the reason for the unexpectedly large increase in the methylene proton
signals, we must first understand what the NMR experiment is measuring. In any
membrane system there are coexisting regions (domains) of solid-like and fluid-like
lipids. Solid-like domains may be especially prevalent in PC-LPS mixtures since we
know that LPS increase the order of membranes. However, the proton NMR signals from
slowly moving nuclei (such as those from the lipids in solid-like domains) would not be
observed due to the extreme signal broadening resulting from dipolar coupling and
chemical shift anisotropy. Therefore, we would not expect to see the methylene signals
for these rigid lipids. The large increase in the methylene proton signals observed upon
the addition of EI-l, EI-2 or EI-5 clearly shows that we are disrupting ('melting') some of

those regions and thereby increasing the number of protons observed. These synthetic

120

El- l/PC/LPS

 

 

 

 

 

 

 

 

 

 

I]IJIIIITIITUTIIFJIIWrIlllIllrIrTTllIIIllllIlIIIIIIIIIlTlll'

5.0 4.0 3.0 2.0 1.0 ppm

Figure 4.6: 1H-NMR spectra of PC:Salmonella typhimurium LPS membranes (1:1 ratio
by weight) with a molar ratio of EI—lzLPS of 0 (A), 1:1 (B), 5:1 (C), and 10:1 (D).

121

EI-2/PC/LPS

 

 

 

 

 

 

TIlilrIl—IlljlIIIIIIITI—TITIIIIITWTIIIIIIIIIIIIIIFITIFTIITUUII

5.0 4.0 3.0 2.0 1.0 ppm

Figure 4.7: 'H-NMR spectra of PC:Salmonella typhimurium LPS membranes (1:1 ratio
by weight) with a molar ratio of EI-2:LPS of 0 (A), 1:1 (B), 5:1 (C), and 10:1 (D).

122

 

 

 

 

l EI-3/PC/LPS
D
C
B
A
fi”T””l'”'l"'Tl""l""l"”l'"'ll'”l"'rr”"l”"l

5.0 4.0 3.0 2.0 1.0 pm

Figure 4.8: 1H-NMR spectra of PC:Salmonella typhimurium LPS membranes (1:1 ratio
by weight) with a molar ratio of EI-3:LPS of 0 (A), 1:1 (B), 5:1 (C), and 10:1 (D).

123

EI-4/PC/LPS

 

 

 

 

jilllllll'ITITIIIII]IIITIIIITITITIIIIIIIIIIIIIYTIIITIIII[Til

5.0 4.0 3.0 2.0 1.0 ppm

Figure 4.9: 1H-NMR spectra of PC:Salmonella typhimurium LPS membranes (1:1 ratio
by weight) with a molar ratio of EI-4:LPS of 0 (A), 1:1 (B), 5:1 (C), and 10:1 (D).

124

 

 

 

 

 

EI-5/PC/LPS
D
C
B
A
T‘l‘l‘“'IIT“I'W'I‘T‘W'”'l'”'l”"l"”l""l””l“”l

 

5.0 4.0 3.0 2.0 1.0 ppm

Figure 4.10: 'H-NMR spectra of PC:Salmonella typhimurium LPS membranes (1:1 ratio
by weight) with a molar ratio of EI—5:LPS of 0 (A), 1:1 (B), 5:1 (C), and 10:1 (D).

125

Table 4.1: Percent change in the methylene proton signals upon addition of the

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

synthetic lipids.
Molar Ratio Amount of synthetic % Change in Area % Change in Area
(LipidzLPS) lipid in samplg_mg) 1.34 ppm signal 1.22 ppm signal ‘
1:1 EI-l 0.056 21.0 -l6.6
5:1 EI-l 0.278 62.6 19.2
10:1 EI-l 0.556 102.7 42.4
1:1 EI-2 0.054 35.6 26.9
5:] EI-2 0.268 125.8 54.6
10:] EI-2 0.537 216.4 205.2
1:1 EI-3 0.059 18.1 48.3
5:1 EI-3 0.297 16.6 48.8
10:1 EI-3 0.593 27.5 20.4
1:1 EI-4 0.072 21.3 28.5
5:1 EI-4 0.358 16.2 55.8
10:1 EI-4 0.716 57.0 78.4
1:1 EI-5 0.103 5.8 30.6
5:1 EI-5 0.515 102.6 111.8
10:1 EI-S 1.031 300.9 264.9

 

 

 

 

 

 

lipids are, therefore, causing an increase in the amount of fluid lipids present in the

membrane.

The ability of the synthetic lipids to disrupt rigid lipid phases was also observed
by fluorescence depolarization measurements of DPPC membranes. Below 42°C, DPPC
exists in very highly ordered gel phases. At 42°C, DPPC undergoes a sharp melting
transition to the more fluid-like, lamellar liquid crystalline phase (see Figure 1.4). The
effects of EI-2 on DPPC and DPPC-S. typhimurium LPS mixtures were investigated
(Figure 4.11). EI-2 was able to disrupt DPPC and DPPC-LPS as measured by its ability
to decrease the fluorescence anisotropy of the membranes at 37°C (Figure 4.11, A).
Furthermore, EI-2 broadened the phase transition range and decreased the transition

temperature (Figure 4.11, B). This also indicated that EI—2 disrupted the rigid, gel phases

126

Figure 4.11: Fluorescence anisotropy of tPnA in DPPC
and DPPC-S. typhimurium LPS (1:1 by weight)
membranes with and without EI—2 added. A. Anisotropy
of DPPC and DPPC-LPS membranes as a function of El-
2 concentration at 37°C. B. Anisotropy vs. temperature.
In B., EI—2 is used at the highest concentration tested in A.
Notice the lowering of the transition temperature and the
broadening of the phase transition induced by EI-2.

 

 

 

0.32“
0.30‘

0.28 ‘

Anisotropy

0.26 ‘

0.24“

 

 

 

—D— DPPC
'—°— DPPC + Salmonella LPS

 

 

 

0.22

V V V I l V Vfi I r V I I V T V V I V I I V V U I r V V V ' V I v W

 

.9
u

Anisotropy

.9
N

 

 

 

 

 

1.? - \h
° DPPC .
° DPPC+S.t.LPS '_
'3 DPPC+S.t.LPS+EI-2 D .

 

 

 

 

 

0.1

r I V V I t T I T—l' ' I I I U I I I I I r T I I I V ' ‘fi—

20 25 30 35 40 45 50 55

Temperature (°C)

Figure 4.1 1

128
of DPPC.

The effects of the five synthetic lipids on the trans-gauche isomerization rate of
the acyl chains of PC-LPS membranes (37°C) were measured by T1 relaxation. As seen
previously for PC-LPS membranes (Chapter 3), the magnetization recovery plots for the
methylene protons of PC-LPS-synthetic lipid membranes were not single exponential
recoveries at all concentrations tested. The data were fit to a double exponential function
(Chapter 3, equation 5) and the T1 values obtained are shown in Table 4.2. The synthetic
lipids did not give any significant change in the T1a values for both the 1.34 ppm and
1.22 ppm signals. As in Chapter 3, the le values were not examined quantitatively due
to the large error associated with these values. The T1 results demonstrate that the
synthetic lipids do not have a large effect on the trans-gauche isomerization rate of the
acyl chains of the ‘fluid’ PC-LPS membranes that are observed by N MR spectroscopy.

A previous study by Lill-Elghanian (133) analyzed the effects of the interaction of
EI-l or EI-2 with lipopolysaccharides by determining whether these substances blocked
the binding of fluorescently—labeled S. typhimurium LPS to RAW264.7 cells. The binding
of the LPS (observed by fluorescence microscopy) to the RAW264.7 cells either did not
change or increased when EI-l or 131-2 was added to LPS. This clearly indicated that
these substances do not block the binding of the LPS into the cell membranes ruling out
the possibility that simple electrostatic blocking was taking place. As discussed above,
Jacobs and Price (121) suggested that the binding of LPS to the cell membrane might be
mediated by LPS receptors. Many putative LPS receptors have been reported (see
Chapter 1). These results clearly show that the endotoxin inhibitors are not competing

with these receptors since they do not inhibit binding of LPS to RAW264.7 cells. Yet,

Table 4.2: Effects of the synthetic lipids on the T1 relaxation values for the methylene

proton signals in PC-LPS membranes.

129

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.34 ppm
Molar El- 1 EI-3 EI—4 EI-5
Ratio
(lipidzLPS) T1, le T1, T1,, T1, T1,, T1, T1b T1, T1,,
0 0.36 1.0 0.56 6.4 0.56 6.4 0.56 6.4 0.56 6.4
l :1 0.38 1.0 0.46 2.1 0.52 1.6 0.56 6.0 0.56 9.1
5: l 0.50 1.7 0.50 2.8 0.48 2.9 0.48 2.2 0.49 2.0
10: 1 0.52 1.6 0.54 4.2 0.55 6.5 0.44 1.4 0.50 1.4
1.22 ppm
Molar EI-4
Ratio
(lipidzLPS) T1, T1, T1, T1, T1, T1, T1, T1, T1, T1,,
0 0.46 1 .4 0.59 6.5 0.59 6.5 0.59 6.5 0.59 6.5
1:1 0.48 1.7 0.45 1.8 0.40 1.0 0.56 -18.8 0.56 -9.3
5: 1 0.53 2.0 0.47 1.5 0.56 -4.1 0.52 2.1 0.54 1.5
10:1 0.52 2.0 0.58 -3.5 0.54 3.9 0.54 17.6 0.52 1.7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

they are able to block the biological activity of LPS as measured by the LAL and TNF
induction assays. This indicates that the mode of action of the inhibitors is not by
blocking the interaction of LPS with surface of the cells.

Together, these results are consistent with the endotoxin inhibitors functioning to
alter the ability of LPS to induce membrane perturbations. This could occur by either (1)
direct disruption of the LPS aggregate structure, thus altering the ability of LPS to induce
a membrane perturbation, or (2) change of the membrane properties that offset the
perturbations induced by LPS. In the former scenario, we would expect that a relatively
small quantity of the synthetic lipids is‘ needed to block activity. Intercalation of an
approximately equal molar ratio of the synthetic lipid into the LPS aggregate should alter

its aggregate structure and thus block activity. However, we observe that the biological

activity is blocked at a much higher molar ratio of the synthetic lipids. For example, at

130
10 ng/ml EL] and 0.1 ng/ml LPS, the molar ratio is approximately 1800:l EI—1:LPS

(assuming an average molecular weight of 15000 g/mole for LPS). The high level of the
synthetic lipids required indicates that the second scenario is the predominant mechanism

by which they are inhibiting the biological activity of LPS.

Conclusion

I definitively show that the five synthetic lipids (El-1 through EI—S) block the
ability of LPS to activate the clotting enzyme cascade of the Limulus amebocyte lysate
and to induce the secretion of TNF from macrophages. My results also demonstrate that
the synthetic lipids do not just simply cause an alteration in the acyl chain mobility of
membranes, they also induce a disruption of the domain structure (order) of the
membrane. I propose that the perturbation in the membrane dynamics and order elicited
by the synthetic lipids is the mechanism by which they inhibit the biological activity of
LPS. Other possible mechanisms of inhibition by the synthetic lipids have also been
explored. The synthetic lipids do not show any signs of toxicity for the RAW264.7 cells
at the concentrations used thus eliminating the possibility that these lipids simply block
the TNF-or response by killing the macrophage cells. The synthetic lipids could also
inhibit the biological activity of LPS by blocking the electrostatic interaction of LPS with
the macrophages. This possibility was explored for EI-l and EI-2 by measuring the
binding of fluorescent LPS to the cells (133). EL] and EI-2 did not decrease the amount
of LPS bound to the macrophages, thus demonstrating that the inhibitors are not blocking
the binding of LPS to the macrophages. These results clearly show that these

possibilities are not responsible for the inhibition of the biological activity by the

131
synthetic lipids. This is consistent with the hypothesis that EI-l — EI—5 function by

affecting the properties of the host cell membrane.

My results indicate that the change in the lipid domain structure induced by the
synthetic lipids is a major contributing component in the inhibition of a physiological
response by LPS. Furthermore, the five synthetic lipids reduce the biological activity of
LPS by altering the domain structure of the membranes, not simply by increasing the
lipid acyl chain motion as previously suggested (133). Furthermore, the large
concentration of the synthetic lipids required to suppress the biological activity of LPS
suggests that they function by offsetting the membrane perturbations induced by LPS.

From our results, it should be expected that many other unrelated molecules may
be able to suppress the biological activity of LPS if they are able to interact with the cell
in such a way as to counteract the effects of LPS on the properties of the membrane. In
fact, suppression of the physiological effects of LPS is observed for such diverse
lipophilic structures as Rhodobacter sphaeroides lipid A (which has a very unusual fatty
acid structure) (97-99), lipid X and lipid X derivatives (180-182), polymyxin B (183,
184), tetrahydrocannabinol (THC, the active ingredient of marijuana) ( 185), ethanol (186,
187), and dimethyl sulfoxide (188). Clearly, the link between the effects LPS has on the
dynamics and order of membranes and its ability to elicit a biological response is very

complicated and will not be explained by a simple model.

Chapter 5

Summary and Perspective

132

133
Since Pfeiffer first discovered lipopolysaccharides over 100 years ago, an

enormous amount of research has been devoted to understanding its properties. Perhaps
Berry (189) summarized it best when he wrote: “An important reason why endotoxin has
been the object of so much research is its diversity of action. It poses challenges to
chemists, physiologists, immunologists, microbiologists, pharmacologists,
endocrinologists, biochemists, anatornists or structural biologists, pathologists, internists,
surgeons, infectious disease specialists and others. A substance with an appeal as broad
as this is rare in science.” The interest in LPS continues unabated. Even as this
dissertation is being written, approximately 200 journal articles that study, use, or are
related to LPS research are being published every month.

Despite this intense multidisciplinary scrutiny, the molecular mechanism
underlying the LPS-induced activation of the immune systems of animals remains
elusive. The severe immune response elicited by LPS is, in large part, responsible for the
physiological effects characteristic of Gram-negative septicemia. Gram-negative
septicemia continues to be a major clinical problem and results in the death of thousands
of individuals each year. Understanding how LPS elicits such a severe immune response
is critical to developing effective therapies for this problem. The ever increasing number
of antibiotic resistant bacterial strains only adds urgency to this issue. Moreover,
understanding LPS’s potent immuno-stimulatory power may lead to therapies for
immune compromised individuals such as those undergoing chemotherapy.

I think that it is fair to say that anyone who has examined all the aspects of LPS
would agree that its mechanism of action is tremendously complex. Accordingly, the

LPS literature is inundated with seemingly unrelated or even contradictory results. To

134
date, the two step model proposed by Jacobs and Price (121) has been the only model

describing the molecular mechanism of the LPS-host cell interaction that is consistent
with the bulk of the known data on LPS (see Figure 1.11). This model divides the LPS-
host cell interaction into two distinct steps. The first step is the electrostatic interaction of
LPS with the surface of the immune cell membrane. The second step is the
rearrangement of the LPS so that the lipid A region inserts into the membrane bilayer.
They propose that there may be proteins (receptors) that facilitate the binding and
insertion of LPS into membranes. Furthermore, they propose that the activation of
immune cells by LPS is a consequence of the insertion of LPS into the membrane and not
the binding to the surface of the membrane.

Lets examine the aspects of this model in light of the data presented in this
dissertation.

Electrostatic Interaction of LPS with Membranes. For more than a decade, the
concept of an LPS-specific receptor has dominated the LPS literature (see Chapter 1). In
general, it is thought that there is a membrane associated protein that, when bound to
LPS, induces a signal transduction cascade to activate the immune cell. A true LPS
receptor would need to have the properties of being specific for lipid A (i.e. can accept
only limited structural diversity within lipid A) and, most importantly, be able to transmit
a functional signal across the membrane in response to binding of LPS.

The initial interaction of LPS with cell membranes is electrostatic in nature as
shown by the experiments of Carr and Morrison (118, 119) and Jacobs and co-workers
(120). Furthermore, there is little doubt that proteins may be involved in this interaction.

In fact, this is quite reasonable since proteins are such a major component of membranes.

135
However, the question that must be addressed is whether or not these interactions are

specific and result in a signal being transmitted across the membrane. Our results from
Chapter 2 demonstrate that the interaction of LPS with immune cells does not require a
highly conserved lipid A structure. We found that lipid A from species of Rhizobium had
several of the biological activities typical of enteric lipid A despite their highly unusual
structure. They not only have unique headgroups, which would be involved in the
electrostatic interaction with the membrane, but they also have very unusual fatty acid
structures. The ability of these LPS to elicit a biological response is especially amazing
when we consider the fact that the lipid A from Rhodobacter sphaeroides, which only
differs in its fatty acid structure (Figure 1.10), is completely devoid of biological activity.
This clearly demonstrates that a specific interaction of lipid A with the host cell
membrane is not a requirement for biological activity. In fact, the most promising of the
LPS ‘receptors’ found to date, CD14, has been shown to be very non-specific for LPS. It
binds to a wide range of polyanionic molecules (145, 146). Furthermore, CD14 is not a
transmembrane protein and therefore, can not transduce a signal across the membrane in
response to LPS. Therefore CD14 can not, in the true sense of the word, be called an
LPS receptor.

Unfortunately, the term receptor has been used very loosely in the LPS literature.
Of course, all molecules interact with each other to some extent. In the case of LPS, its
charge and amphiphilic nature make some of these interactions favorable. Those proteins
that have exhibited a favorable interaction with LPS are termed receptors even though
they may not be truly specific or functional receptors. In this loose definition of

receptors, almost any molecule, whether a protein or not, could potentially be called an

136
LPS receptor. The numerous LPS receptors identified so far may actually bind to LPS as

a result of serendipitous favorable molecular interactions and not because they were
designed to seek out LPS and transmit a signal across the membrane in response to it.

So what is the role of these LPS ‘receptors’? Jacobs and Price suggested that
some proteins on the surface of the membrane might act to facilitate the incorporation of
LPS into the membrane (121). This appears to be a likely scenario. In fact,
lipopolysaccharide binding protein and CD14 have been shown to facilitate the
incorporation of LPS into phospholipid bilayers (147-149). The incorporation of LPS
into membranes by these proteins may simply be a result of their ability to efficiently
bring LPS into close proximity to the membrane.

From the discussions above, we can see that the binding of LPS to the cell surface
is not the step that induces a cellular signal. Instead, the events that occur when LPS is
incorporated into the membrane are the essential points that must be addressed in order to
understand how LPS activates immune cells.

Effects of LPS on Membranes. In the two step model, Jacobs and Price proposed
that the activation of immune cells by LPS is a consequence of the insertion of LPS into
the membrane (121). They, as well as others, demonstrated that LPS is capable of
decreasing the ‘fluidity’ of membranes (121, 123, 124, 160, 161). However, they fall just
short of explaining how the insertion of LPS into the cell membrane results in the
activation of the cell.

Cellular membranes are extremely complex. They exist as an aggregate with
differing types and degrees of motions and order. The dynamics and order of membranes

have a tremendous influence on the function of the membrane (see Chapter 1). They can

137
control the activity of membrane-associated proteins, the flux of enzyme substrates and

protein-protein interactions. This, in turn, will affect the function of the cell. However,
the dynamics and order of membranes are not constant. Instead, they are sensitive to
environmental conditions such as temperature and pH. Furthermore, molecules that
interact with membranes will also affect its properties. Thus, these molecules will also
have the potential to alter the normal functions of the membrane and the cell.

The ability of LPS to incorporate into cellular membranes may result in a change
in the function of the membrane. If this was involved in the mechanism of LPS action,
then we should be able to observe changes in membrane properties induced by LPS. In
Chapter 3, we described the changes in membrane properties induced by LPS. We found
that LPS induced dramatic changes in both the dynamics and the order of phospholipids.
Furthermore, the results obtained from synthetic lipid probes (Chapter 4) indicate that
membrane perturbations are central to the physiological responses elicited by LPS.

My data also indicate that the relationship between the effects of LPS on
membrane properties and its ability to elicit a biological response is not simple. We show
that there is not a linear relationship between the biological activity of LPS and the
changes in membrane acyl chain motion. In hindsight, its seems unreasonable to expect a
simple relationship to exist. Membrane dynamics and order are extremely complex and
interrelated. There is a complex interplay between membrane dynamics and order and by
changing one parameter, all of the properties are affected. Moreover, our understanding
of the relationship between these changes in membrane properties and the function of the
membrane are still in its infancy. We can not yet fully predict the functional changes that

may be expected from perturbations in membrane properties.

138

Model for the Mechanism of LPS-Induced Immune Cell Activation. The two step

model proposed by Jacobs and Price remains the only molecular model to describe the

LPS-host cell interaction. My results are consistent with this model and also extends

some of the original points of this model:

1.

The first step is the electrostatic interaction of LPS (or lipid A) with the cell
membrane. This interaction is not highly specific and may involve a variety
of membrane molecules, including membrane proteins. This non-specific
interaction does not result in a cellular signal.

LPS intercalates into the cell membrane.

The insertion of LPS results in the perturbation of the membrane dynamics
and order. The extent and nature of the membrane perturbations is dependent
on the physicochemical properties of the individual LPS molecules and of its
aggregates.

The membrane perturbations induced by LPS results in a functional change in
the membrane and this, in turn, triggers the cellular response such as the

secretion of cytokines.

This last point requires further discussion. To the best of my knowledge, this is

the first time that membrane perturbations have been proposed as the decisive molecular

event that leads to the activation of immune cells by LPS. How might this occur? The

data presented in this dissertation clearly demonstrated that LPS affect both the dynamic

and organizational state of membranes. In Chapter 1, we discussed the physiological

significance of membrane dynamics and order. The activities of many membrane-

associated proteins are dependent upon the physical state of the membrane. Furthermore,

139
the degree of connectivity of domains can create functional areas in the membrane and

can control the flux of substrates and protein-protein interactions. Two particularly well
studied membrane proteins (with respect to membrane effects) are phospholipase A2
(PLA2) and protein kinase C (PKC). Both of these proteins are important components of
signal transduction pathways. PLA2 activity is affected by membrane ‘fluidity’ (45),
bilayer packing stress and defects (50, 51), and the formation of solid-like lipid domains
(59). Interestingly, LPS alters all of these membrane properties. I found that LPS
decreases lipid acyl chain motion (‘fluidity’), induces regions of fast motion (membrane
defects), and alters the domain organization of membranes. As would be predicted from
these effects, LPS is known to activate PLA2 (190). PKC activity is also dependent upon
domain formation (60) and has been shown to be closely associated with the hexagonal
phase propensity of the membrane (52). With respect to this last point, Brandenburg and
co-workers have found that the biological activity of LPS is correlated to its ability to
form inverted phases such as hexagonal phases (7, 129). Toxic LPS favored inverted
phases. As might be expected, LPS also activates PKC (191, 192). While the examples
of PLA2 and PKC illustrate how the perturbations induced by LPS can alter the
membrane function, they only begin to show the potential for these types of effects. As
our knowledge of the physiological significance of membrane dynamics and order
increases, so will our knowledge of how LPS induced membrane perturbations may
influence cellular functions.

The extended two step model described above has some important predictions.
First, this model does not necessitate the direct interaction of LPS with membrane

proteins in order to elicit a biological response. Second, any lipophilic molecule may

140
affect the biological activity of LPS if this molecule is capable of offsetting or enhancing

the membrane perturbations induced by LPS. Finally, the ability of LPS to elicit an
immune cell response may also be altered by environmental parameters that affect
membrane dynamics and order, such as temperature, pressure, ionic strength, and pH. It
should also be recognized that lipophilic molecules or environmental parameters might
be able to mimic some of the biological responses induced by LPS.

We already see evidence supporting these predictions. Many lipophilic molecules
are capable of reducing the biological activity of LPS. These include R. sphaeroides lipid
A (97-99), lipid X and lipid X derivatives (180-182), polymyxin B ( 183, 184),
tetrahydrocannabinol (THC) (185), ethanol (186, 187), and dimethyl sulfoxide (188).
While we do not know the exact mechanism by which they decrease the biological
activity of LPS, their ability to interact with the membrane is consistent with our model.
Temperature has also been shown to affect tumor necrosis factor (TNF) secretion from
macrophages. Increasing the temperature of macrophages, which would increase the
motion of the membrane, results in a decreased production of TNF-or in response to LPS
(174—176). The opposite effect is observed when the temperature is lowered. This
treatment would decrease the motion of the macrophage membrane (analogous to the
addition of LPS to membranes). Lowering the temperature increases the production of
TNF-or from LPS-stimulated macrophages ( 177, 178).

Other typical biological activities induced by LPS can be linked to the
involvement of membranes. These include the Limulus amebocyte lysate (LAL) assay,
thrombosis, and B-lymphocyte activation. The LAL assay utilizes the blood clotting

cascade of the horseshoe crab as a sensitive and specific measure of LPS toxicity. The

141
blood clotting cascade of the horseshoe crab is highly homologous to the mammalian

blood clotting mechanism and they are believed to be evolutionarily related (193). In
mammals, the enzyme complexes responsible for clotting depend on and assemble in
membranes (194). Calcium ions are essential for the proper folding of the membrane
binding domains of these proteins and are also involved in the cross-linking of the acidic
protein residues to the phosphates on the membrane lipids (195). Clearly, the membrane
is a critical component of the clotting cascade in mammals. Unfortunately, the role of the
membrane in the horseshoe crab clotting cascade has been largely ignored. There is,
however, strong evidence that at least some of the enzymes of this cascade are also
dependent on membrane interactions. Calcium ions are needed for the clotting cascade
analogous to the mammalian clotting cascade (196). Lui et al. have also shown that the
purified proclotting enzyme of this cascade requires detergent to be active (197),
indicating that this enzyme also requires an interaction with a membrane for activity. We
believe that the physical properties of the membranes with which the enzymes interact
will affect the activity of the enzymes of the clotting cascade. The ability of LPS to alter
the dynamics of these membranes will thus result in an alteration of the clotting cascade.
LPS also affects B-lymphocytes. LPS acts as a mitogenic signal for the induction
of B-lymphocyte proliferation and differentiation into antibody secreting plasma cells.
When LPS binds to B-lymphocytes, it induces a major lateral rearrangement of the
plasma membrane (capping) (121, 198). The LPS response of B-lymphocytes has
garnered a lot of attention since the discovery of a genetically altered mouse strain
(C3H/HeJ) whose B-lymphocytes were hyporesponsive to LPS (199). It was originally

believed that the reason for the low response to LPS was due to a missing or altered LPS

142
receptor. However, binding of LPS to C3H/HeJ B-lymphocytes was not altered and no

missing receptor has been identified (199, 200). One difference between the normal- and
hyporesponsive B-lymphocytes is their membrane composition. Chaby et al. (201) noted
a change in the ganglioside composition in C3H/HeJ B-lymphocytes. A later study by
Yohe, et al. (202), demonstrated that there is a change in the structure of the C3H/HeJ B-
lymphocyte membrane that alters the surface accessibility of the gangliosides. These
studies clearly show that LPS hyporesponsive B-lymphocytes have altered membrane
compositions. The decreased response of these cells is consistent with the idea that the
membrane structural changes alter the membrane perturbations induced by LPS and, thus,
affect the cell’s ability to respond to LPS.

The findings of this dissertation highlight the importance of understanding LPS-
membrane interactions. My research has advanced our understanding of the effects LPS
has on the properties of membranes. However, we are only just beginning to fully
understand LPS-membrane interactions. This area of LPS research needs more attention
in the future. Some experiments should focus on understanding the effects of other LPS
structures (such as Rhizobium or Rhodobacter LPS) on the properties of membranes. In
addition to the experiments described in this dissertation, I think that deuterium NMR
may be very powerful for this purpose. With this technique, specific lipid species can be
2H-labeled, thus allowing a definitive measure of the properties of individual lipid species
in a complex mixture. Furthermore, it may be possible to isolate 2H-labeled lipid A from
bacteria by using minimal media with a deuterated carbon source. Magic angle spinning
NMR spectroscopy should also be explored since it is able to resolve many of the

methylene signals in the lipid acyl chains (203). This would allow an even more detailed

143
look at the motions of the lipid acyl chains. My results showed that LPS induces domain

formations in membranes. This also needs to be investigated further. Fluorescence
microscopy on model membranes may provide useful information related to LPS induced
domains. This could be done with multilamellar, bilayer (black lipid membranes),
monolayer, or cellular (such as erythrocyte) membranes (see ref. 14-21, 27, 36, 39, 40).
Fluorescence recovery after photobleaching (FRAP) has also provided information on
lipid domains and could be used (see ref. 36, 37, 61). Furthermore, FRAP would
measure the change in the lateral diffusion of lipids in response to LPS. This would also
be a good parameter to assess the changes induced by LPS. Finally, I think that the effect
of LPS on different lipid species needs to be analyzed. One previous study addressed this
issue (160). Given the negative charge of lipid A and its high degree of order, we might
expect that it would have a strong preference to associate with negatively charged
phospholipids (due to ionic bridging) and saturated lipids (due. to more favorable van der
Waals interactions with the acyl chains). The preference of lipid A to associate with
specific lipid species will surely be an important component of its ability to elicit
membrane perturbations.

These experiments will allow future investigators to further our understanding of
LPS-membrane interactions. In turn, this will help us to understand how LPS elicits such
diverse and severe biological responses. In this quest, we must continually remember
that LPS is just a lipid and the key to understanding its effects lies in its ability to interact
with other lipids. After all, the main function of LPS is to serve as a barrier on the
surface of Gram-negative bacteria to protect the cell from environmental stress. It is

unlikely that LPS was designed to wreck havoc on the immune systems of animals. One

144

thing that is certain, however, is that the interaction of LPS with animals is extremely

complex and will capture the attention of many more researchers in the years to come.

APPENDIX

APPENDIX

Synthesis and Characterization of Five Synthetic Lipids

145

146
Introduction

In Chapter 4, I describe the results obtained from five synthetic lipids (El-1 to E1-
5) (see Figure 4.1). These lipids were found to be good inhibitors of the biological
activities of LPS. My results indicate that these lipids function by altering the physical
properties of membranes and offset the perturbations induced by LPS.

To understand how these lipids affect membrane properties, one must first
understand their physical properties. The synthesis of EI-l and EI-2 was described
previously (174). In this report, I describe the synthesis, purification and characterization
of EI-3, EI-4 and EI-S. I also characterize the physical properties of all five of the lipids
by differential scanning calorimetry (DSC), x-ray powder diffraction, and proton NMR

Tl relaxation measurements.

Materials and Methods

Mass Spectrometry. Electrospray ionization mass spectrometry was performed on
a Fisons Platform instrument using a 4:1 CHClgzmethanol mobile phase. Samples were
prepared in the same mobile phase with 1% formic acid added and positive ion spectra
were recorded. Spectra were processed by subtracting the baseline and smoothing the
data.

Nuclear Magnetic Resonance Spectroscopy. 1H-NMR spectra were recorded in
CDC13 on a Varian 500 MHz VXR spectrometer. EI-3 and EI—4 required the addition of
a couple of drops of deuterated methanol to dissolve the samples. The CDC13 solvent
peak (7.24 ppm) was used as the reference. Spin-lattice (T1) relaxation measurements

were performed on a Varian 300 MHz spectrometer using the inversion recovery

147
technique (169). An 8 s delay was used to allow for complete equilibration of the nuclei.

The temperature was allowed to equilibrate for at least 15 min before the measurements
were taken. All of the samples for T1 measurements were dissolved in the same stock
solution of 1:1 CDClngD3OD. This solution was used because it was the most polar
solvent system that would dissolve all of the lipids.

Difi‘erential Scanning Calorimetry (DSC). DSC measurements were made on a
Microcal-2 high sensitivity scanning calorimeter at 60°C/hr. Approximately 5 mg of
synthetic lipids were hydrated in water (1.5 ml) and multiple scans (up to 5) were
recorded from 5°C to 98°C.

X-Ray Diffraction. X-ray diffraction patterns were recorded on a Rigaku rotaflex
rotating anode operating at 45 kV and 100 mA. The raw x-ray beam was collimated with
1/6° slits and the copper K0, line (1.54 A) was selected by a monochromator. Dry
samples were deposited on a microscope slide in a thin film and the x-ray diffraction
pattern obtained in the reflection mode. Samples in solution were placed into 0.7 mm
glass capillaries (Charles Supper Co.) and mounted with a home-built capillary holder.
Diffraction patterns were obtained in the transmittance mode. All diffraction patterns
were obtained at room temperature. A one—dimensional scanning detector was used to

measure the reflected radiation.

Syntth and Purification of EI-3, E14 and EI-S

EI-3 was synthesized by the scheme shown in Figure A.1. Tris(2-
aminoethyl)amine and approximately 4 equivalents of fatty acid methyl esters (palmetic
acid, technical, methanolized in 5% HCl in methanol) were mixed in ethanol and refluxed

for 3 hours. The solvent was evaporated under reduced pressure and the remaining white

148

$2

Hzf

.m-_m m8 25:8 ocofiim :.< 0.5me

mmfi I\Iz£

5:2 z
mom

£2
nmoo

149

solid was washed several times with hexanes to remove excess fatty acid methyl esters.
The product was purified by silica gel flash chromatography in 10:1 CHCl3zmethanol.
Fractions were assayed by thin layer chromatography (TLC) on silica gel 60 in the same
solvent system. Fractions that contained only EI-3 were pooled. The resulting pure
product ran as a single spot on TLC (Figure A2). The 1H-NMR spectrum of the EI-3
shows the expected signals with trace levels of impurities detected. The assignments of
the proton signals are given in Figure A.3, A. The small resonance at 3.4 ppm is due to
deuterated methanol that was used along with CDC13 (7.24 ppm) to help completely
dissolve the sample. Positive-ion electrospray ionization mass spectrometry (El-MS)
shows that the pure EI-3 is composed of a mixture of saturated fatty acids ranging from
C14 to C18 (Figure A.3, B). The ion at 834.1 m/z corresponds to a protonated EI—3 that
contains one C14 and two C16 fatty acids. The ion at 946.3 m/z corresponds to the
protonated species containing three C18 fatty acids. From this spectra we deduced that
the primary fatty acids are C16 and C18 with lesser amounts of C14, C15, and C17
present. This lipid is judged to be 2 98% pure.

The EI-4 synthetic scheme is shown in Figure A.4. Methanolysis of
ethylenediaminetetraacetic acid (free acid) was performed in 5% HCl in methanol and
refluxed for 4 hrs. The solvent was dried and the methylated product was dissolved in
ethanol. Approximately 6 equivalents of l-tetradecylamine were added and refluxed for
3 hrs. Solvent was removed by evaporation under reduced pressure and the product was
washed several times with hexanes to remove excess l-tetradecylamine. EI-4 was
purified by silica gel flash chromatography with 100: 10:0.2

CHCl3:methanolzconcentrated ammonium hydroxide as the solvent. Fractions were

 

Figure A.2: Thin layer chromatography of EI-5 (lane A), EI-4 (lane B), and EI-3 (lane C).
Solvent system was 10:1 chloroform:methanol and the plate was stained with
phosphomolybdic acid (10%).

assayed by TLC in the same solvent system and the fractions that contained only EI-4
were pooled. TLC of the purified EI-4 gave nearly a single smeared spot with only trace
levels of faster migrating components detected (Figure A.2). lH-NMR spectroscopy of
EI-4 and the assignment of the proton signals are shown in Figure A5, A. Only very low
levels of impurities were detected. The signals at 3.35 ppm and 7.24 ppm are due to
deuterated methanol and CDC13, respectively. Positive ion EI-MS (Figure A.5, B) shows
the protonated molecular ion of EI-4 at 1073.9 m/z. The ion at 556.6 m/z corresponds to
the doubly charged calcium adduct. The small peak at 214.8 rn/z is due to protonated
tetradecylamine that may arise from the fragmentation of EI-4 or may be a small impurity
in the sample. The remaining cluster of ions from 437 m/z to 640 m/z could not be
assigned. These ions may be due to multiply charge lipid aggregates or impurities in the

sample. This lipid is judged to be .>_ 95% pure.

 

151

 

 

 

 

 

B
\
A
E\(fD
/\.N/
RNIW\/\A
F C B
e
1 F ED
1
C
G
U1.
,;. l I; fifij. i a" i'flr‘n-i
am
913.:
“1.1
946.3
nu
. .......thli .. a
soossocooosomwomssomssoimimnoo

 

350'“: 450

Figure A.3: Proton NMR spectrum (A) and positive-ion electrospray mass spectrum (B)
of EI-3. NMR spectrum assignments are shown. Asterisks indicate solvent signals.

Assignments of the mass spectrum are described in the text.

152

aim m8 080:8 ammofigm ”vac. Emmi

me}

15:8 .moom

 

1.5538882 £8 \/\/

z/Jfioonm 0.82256: .8 5le

 

at?“
Z/\/

153

 

 

 

 

 

 

 

 

 

 

/B
A
A :NHi)\NCH/Iz:\H ANHF
/ \l \A
E C B
*
F
E D
\’ C
’ Ll.
- ..
r 1 TIT 1 l‘jfTfi "IT'I'T
7 s 4 3 2 1 pp-
100 564.6
7.. 556.6 582.1
5505
‘ 2“" 271 6 4370 5404 1073.03,“
. ' ‘ 5‘33 ' 1046.0
0_ . l. 1 . m
2.. 3.. .. .. .1. 7.. an .., 1000 1100 1200 1300

Figure A.5: Proton NMR spectrum (A) and positive-ion electrospray mass spectrum (B)
of EI-4. NMR spectrum assignments are shown. Asterisks indicate solvent signals.
Assignments of the mass spectrum are described in the text.

154
EI—5 was synthesized according to the scheme shown in Figure A.6. Tris(2-

aminoethyl)amine was mixed with 4 equivalents of S-3-hydroxy-y-butyrolactone.
Nucleophilic attack of the primary amine on the carboxylic carbon opened the ring of the
lactone to yield the corresponding aminolysis product. The viscous mixture was
dissolved as completely as possible in 1:1 pyridinezacetonitrile and excess lauroyl
chloride added. Addition of excess lauroryl chloride helped dissolve the reaction mixture
to allow the reaction to proceed quickly. After one hour, the solvent was removed by
evaporation under reduced pressure. Analysis of the reaction mixture by TLC showed at
least eight components. Partial purification was accomplished by silica gel flash
chromatography with two different solvents. Lauric acid was first eluted off the column
with chloroform. The solvent was then changed to 98:2 chloroformzmethanol to elute E1-
5 that co-eluted with three other components. All of these components migrated very
closely together on TLC. Preparative TLC in 10:1 chloroformzmethanol was used for the
final purification. The band corresponding to EI-S was collected and product removed
with 85: 15 chloroformzmethanol. Purified EI-S migrated as a single major spot in TLC
with only trace levels of slower migrating components present (Figure A2). The lH-
NMR spectrum of EI—5 and the assignments of the proton signals are shown in Figure
A.7, A. Only trace amounts of impurities are detected. Positive-ion EI-MS spectrum
shows the protonated molecular ion (M-I-H")+ at 1546.8 m/z and the sodium adduct
(M+Na+)+at 1568.8 m/z (Figure A.7, B). Assignment of the other ions are as follows:

785.1 m/z, (M+H++Na+)2*; 795.9 m/z, (M+2Na+)2*; 811.9 m/z, (M+2K+)2+. The purity

of EI-5 is judged to be 2 98%.

155

.Em com cacao. omoeemm a... seam

misc...

EEESRE 653:3

 

‘

onto—no 3.8.52 :2

156

 

 

 

 

 

 

 

 

 

E E
\'\\"/G C B
RNA/NH 1 0%
. 1W \.
F 01
D
* 1)
E
. F C
j r H G
L 1 31 ilL
1 ; 1*”? r: t r;
100* 785.!
1 B
*4
861.0 1m
‘3‘... ... a. a. m ,.,,,,+ ..... e We ............... e ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

 

Figure A.7: Proton NMR spectrum (A) and positive-ion electrospray mass spectrum (B)
of EI-S. NMR spectrum assignments are shown. Asterisks indicate solvent signals.
Assignments of the mass spectrum are described in the text.

157
Characterization of the physical properties of EM, EI-2, EI-3, E14 and EI-S

In this report, I used x-ray diffraction, differential scanning calorimetry (DSC),
and proton NMR Tl relaxation measurements to characterize the properties of the five
lipids. This combination of techniques provides an extensive analysis of the dynamics
and order of these lipids.

I observed the physical properties of the five synthetic lipids. Both EI—l and EI-2
are waxy solids when dry and at room temperature. They readily hydrate when water is
added to form a suspension. EI-3 and EI-4 were crystalline at room temperature. These
lipids were very hydrophobic and were very resistant to forming a suspension in water.
EI-S is a waxy solid at room temperature. This lipid was also very hydrophobic and
resistant to forming a suspension in water. Addition of an ethanol solution of EI-3, EI—4
or EI-S to water was found to be the best way to obtain a fine aqueous suspension of the
lipids. Vigorous sonication with a probe type sonicator worked to a lesser extent.

One of the most useful aspects of DSC is the ability to determine the temperature
and characteristics of the lipid phase transitions. The width of the phase transition is
correlated with the cooperativity of the transition. Sharp transitions indicate a large
degree of intermolecular cooperation between the lipid molecules. The DSC scans from
5°C - 98°C of the five synthetic lipids are shown in Figure A.8. One of the lipids, EI-l,
showed no detectable phase transitions in this temperature range. EI—2 experienced a
fairly sharp phase transition at 74°C. Previous studies on this lipid indicate that the acyl
chains of this lipid were not well ordered at 37°C (see Chapter 4). This is corroborated
by x-ray diffraction data (shown below) which show that this lipid has fluid acyl chains at

room temperature. Therefore, the observed phase transition is not a chain melting

Cp (arb. units)

158

 

 

 

 

 

 

 

 

 

-7 Y"V'V'IV"'I'r'f'v'Uv‘Ui‘vfi'I'V"I’“ rji'U'V'IIT“

O 10 20 30 40 50 60 70 80 90 100

Temperature (°C)

Figure A.8: Differential scanning calorimetry scans of the five synthetic lipids.

159
transition. EI-3 and E14 had more complex DSC traces and exhibited multiple high

temperature phase transitions. EI—3 had a fairly broad phase transition at 82°C. The
increase in the baseline at the end of the DSC trace (90°C-98°C) indicates the beginning
of another phase transition at a higher temperature. EI-4 showed four distinct phase
transitions at 47°C, 86°C, 91°C, and 97°C. EI-S underwent a sharp phase transition at
27°C. These results indicate that these lipids, with the exception of EI-l, undergo
cooperative phase transitions.

X-ray diffraction is a very powerful tool for studying the organization of lipids.
In x-ray diffraction, only regular repeating planes of atoms (lattices) diffract the radiation.
The angle of the reflection (diffracted x-rays) is governed by Bragg’s Law:

n?» = 2dsin9 (1)
where n is the order of the reflection, A. is the wavelength of x-radiation, d is the spacing
of the repeating unit, and 6 is the angle of the incident and reflected radiation. Lipids
typically have two types of repeating structures. The first is due to the bilayer repeat
(large (1) and gives low-angle reflections (small 9). The second is due to the packing of
the acyl chains of the lipids (small (1) and gives rise to wide-angle reflections (large 6).
The wide-angle region is useful in determining the order of the acyl chains. Acyl chains
which are rigid (all-trans conformation), such as found in gel phases, give sharp wide-
angle reflections. The other extreme is found in acyl chains that have many gauche
conformations and have a large degree of motional freedom (fluid-like phases). Acyl
chains in this state give broad wide-angle reflections.

X-ray diffraction was used to study the molecular packing of the synthetic lipids.

EI-3 had a well ordered x-ray diffraction spectrum (Figure A9). The low angle region

160
contained five sharp reflections indicating a well ordered lattice. A sharp and intense

chain packing reflection (4.16 A) demonstrated that the acyl chains were rigid and very
well ordered. EI-4 also gave several low angle reflections indicating a well-ordered
lattice (Figure A.10). Three sharp reflections in the wide-angle region indicated three
different chain packing distances (4.61 A, 4.06 A, and 3.74 A). All of these reflections
were sharp which indicated well ordered, rigid acyl chains. EI—5 gave two sharp low
angle reflections (Figure All). The acyl chains gave three small, very sharp reflections
superimposed on a broad reflection. The three sharp reflections would indicate a rigid
acyl chain region while the broad reflection would indicate less ordered acyl chains
typical of a fluid lipid phase. This may be explained by a coexistence of at least two
phases, a phase with rigid acyl chains and a second with fluid acyl chains. These x-ray
results indicate that all three of these synthetic lipids are well ordered. Both EI—l and E1-
2 show similar x-ray diffraction patterns when in water (Figures A.12 and A.13). These
lipids had a very broad acyl chain packing reflection that is characteristic of lipid phases
with fluid acyl chains. This demonstrates that RH and EI—2 have a low degree of order
in their acyl chains.

The methylene proton T1 relaxation times for the five synthetic lipids were
measured as a function of temperature (Figure A.14). In these experiments, the
magnetization is inverted (180° pulse) and allowed to recover to equilibrium. The
recovery of the magnetization to equilibrium is exponential and is described by the
equation:

Mm — M, = 2Mmexp(-t/T 1) (2)

 

161

mm

.mEobmma/w E @232 8a mcouoocum .m-_m me 059:. cougar—o marx Hod Emmi

on
om mm om m“ on m

P—h—h—b—F—h—Phh—me—Pp—pub»—~_PhFhP—bmbt—hbhlemIr—h—h—hmh—b_b
<

.me_Ppp

\

3.:
8.3

3.3

owe—

 

» 3.3
2+

 

hysuoiul [eufirs

 

 

162

.mEobmwé E @232 2a maouoocom Jim .8 058m 888%? awry" A: .< 03me

on
mm om mm om mfi

on“

m

.
_—___rhb__..pbp—Fm_—_.»___L.pr—|F_PmrL.lrbr—~.~_pmp_____Lw».Fm.__..p_.-|P

 

i<|

cod

 

l<li

 

 

 

2.2

 

 

56v

Kitsuoiul [Ruins

 

 

 

 

163

ow

.maoawfi. a 332 2a 888%“ .Em co amen Sesame 52-x H:1“. 2%

mm om

_ .
hP—h—Pbfihr—hbflhphr—th—hPr—L—LP—p—hph—b—b—hopp—bhibipbb. . . l—ibWb—bfin—hmp—hn

 

 

‘I‘l‘lj

mm

 

nod

om
om

m...

 

 

mos

3

m

 

 

2.3

Kitsuoiul [eufits

 

 

 

164

.magmwé a 332 on 2.883. .75 co 28a 888% 5-x é .< aim

mm

ow mm om mm om ma 3.

rt—rFPp—P—h—p—_b——Pmbrp*FrWlP.—h—p—Fhrbbflh—nmh—thbh_P—hmFL-mn—

m

pb..+.bh__.m___._—.l

 

mmé

 

34m

firsumul [eufits

 

 

 

165

.mEobmw=< 5 @232 2a meoaoocom .NAm .8 0383 838.56 .35“ ”2 .< 235

mm

3. mm om mm om ma 3 m

hfL..____—.pp.rc_cm._.~___~._»—_....—»Fr_»»p_pw~__—._rr._pb»_

pm._____..___._br.

 

omé

 

3.9m \

5.:

 

 

 

Kitsuaiul [eufitg

166

 

 

 

T1 (3)

 

 

 

 

15 20 25 3O 35 4O 45 50
Temperature

Figure A. 14: T1 values for the methylene protons of the five synthetic lipids as a function
of temperature.

167
where M” and M, are the magnetization at equilibrium and at time = t, t is the delay time

and T1 is the spin-lattice relaxation time. For spin 1/2 nuclei, dipole-dipole interaction is
the primary mechanism for relaxation (i.e., the interaction of a proton with another
proton). Tl relaxation is dominated by the relaxation events occurring near the same rate
as the NMR spectrometer frequency. For lipids at 300 MHz (~10'9 5), this corresponds to
the trans-gauche isomerization of the carbon-carbon bonds in the acyl chains. Kroon et
al. (173) have demonstrated that the T1 relaxation of lipid acyl chains are due to the intra-
chain proton-proton interaction resulting from the trans-gauche rearrangement. Hence,
our T1 relaxation times are measuring the rate of trans-gauche isomerization of the lipid
acyl chains. For our experiments, the T1 relaxation is in the fast correlation region where
an increase in trans-gauche isomerization rate results in an increase in T1. As expected,
all of the lipids exhibited an increase in T1 (increase in the trans-gauche isomerization
rate) with an increase in temperature. EI-l, 131-2, and EI—S had the highest Tl values,
indicating that their acyl chains have the fastest rate of motion. EI—3 and EI-4 had
significantly lower T1 values indicating a lower rate of acyl chain motion in these lipids.
This correlates well to the x-ray diffraction data. X-ray diffraction showed that EI-l, E1-
2, and EI—5 had a large proportion of poorly ordered (‘fluid’) acyl chains. The x-ray
diffraction patterns from EI—3 and E14 exhibited much sharper acyl chain reflections.
This clearly demonstrates that these two lipids have a high degree of acyl chain order.
The DSC data is also consistent with this result. We know that EL] and EI-2 do not
undergo any phase transitions near or below room temperature and therefore, in
conjunction with the x-ray diffraction data, shows that they exists in a fluid-like phase at

least down to 5°C. EI—5 exhibits a sharp transition near room temperature (27°C) and

 

168
may be due to a chain ‘melting’ transition. EI-3 and EI—4, however, were much more

stable and only had phase transition well above room temperature. This suggests that
these lipids have a very rigid acyl chain region. It should be noted, however, that these
NMR Tl measurements might not be sensitive to very rigid acyl chain motions due to
dipolar coupling and chemical shift anisotropy (see Chapter 4). This could be a factor in
EI-5 since x-ray diffraction indicated that some of its acyl chains might be highly

ordered.

Conclusions

Our results clearly show that all five synthetic lipids form ordered systems typical
of membranes. With the exception of EI-l, all of the lipids undergo concerted thermal
transitions between 5°C and 98°C. According to x-ray diffraction, all of the lipids have
regular repeating lattices. The x-ray diffraction and NMR Tl experiments indicate that
EI—3 and EI-4 have well ordered, rigid acyl chains. The acyl chains of BL], 131-2, and El-
5 were less ordered and exhibited a higher rate of trans-gauche isomerizations than EI-3
or EI-4. However, it should be noted that EI-S had some x-ray reflections in the wide-
angle region that were very sharp. This indicates EI-5 may have a portion of its acyl

chains highly ordered.

LIST OF REFERENCES

10.

LIST OF REFERENCES

Deamer, D. W. (1986) Role of Amphiphilic Compounds in the Evolution of
Membrane Structure on the Early Earth. Orig. Life Evol. Biosph. 17, 3-25.

Tanford, C. ( 1980) The Hydrophobic Effect: Formation of Micelles and
Biological Membranes. John Wiley & Sons, Inc., New York, New York.

Gorter, E. and Grendel, F. (1925) On Biomolecular Layers of Lipid on the
Chromacytes of the Blood. J. Exp. Med. 41, 439-443.

Singer, S. J. and Nicolson, G. L. (1972) The Fluid Mosaic Model of the Structure
of Cell Membranes. Science 175, 720-731.

Hauser, H. and Poupart, G. (1992) Lipid Structure. in "The Structure of Biological
Membranes" (Yeagle, P., Ed.) pp 3-71. CRC Press, Boca Raton, Florida.

Israelachvili, J. N., Marcelja, S., Horn, R. G. (1980) Physical Principles of
Membrane Organization. Q. Rev. Biophys. 13, 121-200.

Seydel, U. and Brandenburg, K. (1992) Suprarnolecular Structure of
Lipopolysaccharides and Lipid A. in “Bacterial Endotoxic Lipopolysaccharides”
Vol. 1 (Morrison, D. C. and Ryan, J. L., Eds.) pp 225-250. CRC Press, Boca
Raton, Florida.

Lewis, R. N. A. H. and McElhaney, R. N. (1992) The Mesomorphic Phase
Behavior of Lipid Bilayers. in "The Structure of Biological Membranes" (Yeagle,
P., Ed.) pp 73-155. CRC Press, Boca Raton, Florida.

Gennis, R. B. ( 1989) Membrane Dynamics and Protein-Lipid Interactions. in
“Biomembranes: Molecular Structure and Function” pp 166-198. Springer-
Verlag, New York, New York.

Casal, H. L. and McElhaney, R. N. (1990) Quantitative Determination of
Hydrocarbon Chain Conformational Order in Bilayers of Saturated
Phosphatidylcholines of Various Chain Lengths by Fourier Transform Infrared
Spectroscopy. Biochemistry 29, 5423-5427.

169

/
C

13.

14.

15.

l6.

17.

18.

19.

 

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

170

Petersen, N. O. and Chan, S. I. ( 1977) More on the Motional State of Lipid
Bilayer Membranes: Interpretation of Order Parameters Obtained from Nuclear
Magnetic Resonance Experiments. Biochemistry 16, 2657-2667.

van der Meer, B. W. (1993) Fluidity, Dynamics and Order. in “Biomembranes:
Physical Aspects” Vol. 1 (Shinitzky, M., Ed.) pp 97-158. VCH Publishers, New
York, New York.

McConnell, H. M. (1991) Structures and Transitions in Lipid Monolayers at the
Air-Water Interface. Annu. Rev. Phys. Chem. 42, 171-195.

Peters, R. and Beck, K. ( 1983) Translational Diffusion in Phospholipid
Monolayers Measured by Fluorescence Microphotolysis. Proc. Natl. Acad. Sci.
USA 80, 7183-7187.

McConnell, H. M., Tamm, L. K., and Weis, R. M. (1984) Periodic Structures in
Lipid Monolayer Phase Transitions. Proc. Natl. Acad. Sci. USA 81, 3249-3253.

Weis, R. M. and McConnell, H. M. (1984)Two-Dimensiona1 Chiral Crystals of
Phospholipid. Nature 310, 4749.

Miller, A. and Mohwald, H. (1987) Diffusion Limited Growth of Crystalline
Domains in Phospholipid Monolayers. J. Chem. Phys. 86, 4258-4264.

Miller, A. and Mohwald, H. (1986) Collecting Two-Dimensional Phospholipid
Crystals in Inhomogenious Electric Fields. Europhys. Lett. 2, 67-74.

Klinger, J. F. and McConnell, H. M. (1993) Field-Gradient Electrophoresis of
Lipid Domains. J. Phys. Chem. 97, 2962-2966.

Weiss, R. M. and McConnell, H. M. ( 1985) Cholesterol Stabilizes the Crystal-
Liquid Interface in Phospholipid Monolayers. J. Phys. Chem. 89, 4453-4459.

Keller, D. J ., McConnell, H. M., and Moy, V. T. (1986) Theory of Superstructures
in Lipid Monolayer Phase Transitions. J. Phys. Chem. 90, 2311-2315.

Estep, T. N., Mountcastle, D. B., Barenholz, Y., Biltonen, R. L., and Thompson,
T. E. (1979) Thermal Behavior of Synthetic Sphingomyelin-Cholesterol
Dispersions. Biochemistry 18, 2112-2117.

Melchoir, D. L. ( 1986) Lipid Domains in Fluid Membranes: A Quick-Freeze
Differential Scanning Calorimetry Study. Science 234, 1577-1580.

Parasassi, T., Conti, F., Glaser, M., and Gratton, E. (1984) Detection of
Phospholipid Phase Separation. J. Biol. Chem. 259, 14011-14017.

t lie-I11 [CL N

29
30.
31.

 

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

171

Klausner, R. D., Kleinfeld, A. M., Hoover, R. L., and Kamovsky, M. J. (1980)
Lipid Domains in Membranes. J. Biol. Chem. 255, 1286-1295.

Parasassi, T., Ravagnan, G., Rusch, R. M., and Gratton, E. ( 1993) Modulation and
Dynamics of Phase Properties in Phospholipid Mixtures Detected by Laurdan
Fluorescence. Photochem. Photobiol. 57, 403-410.

Haverstick, D. M. and Glaser, M. (1937) Visualization of Ca2+-Induced
Phospholipid Domains. Proc. Natl. Acad. Sci. USA 84, 4475-4479.

Spiegel, S., Blumenthal, R., Fishman, P. H., and Handler, J. S. (1985)
Gangliosides Do Not Move From Apical to Basolateral Plasma Membrane in
Cultured Epithelial Cells. Biochim. Biophys. Acta 821, 310-318.

dc Curtis, 1., Fumagalli, G., and Borgese, N. (1986) Purification and
Characterization of Two Plasma Membrane Domains From Ejaculated Bull
Spermatozoa. J. Cell Biol. 102, 1813-1825.

Ishidate, K. Creeger, E. S., Zrike, J ., Deb, S., Glauner, B., MacAlister, T. J ., and
Rothfield, L. I. (1986) Isolation of Differential Membrane Domains from
Eschericia coli and Salmonella typhimurium, Including a Fraction Containing
Attachment Sites Between the Inner and Outer Membranes and the Murein
Skeleton and the Cell Envelope. J. Biol. Chem. 261, 428-443.

Murphy, D. J. (1986) The Molecular Organization of the Photosynthetic
Membranes of Higher Plants. Biochim. Biophys. Acta 864, 33-94.

Wang, S., Martin, J. C., Omann, G., and Glaser, M. (1988) Distribution of
Phospholipids Around Gramicidin and D-B-Hydoxybutyrate Dehydrogenase as
Measured by Resonance Energy Transfer. Biochemistry 27, 2033-2039.

Wang, F., Naisbitt, G. H., Vernon, L. P., and Glaser, M. (1993) Pyrularia Thionin
Binding to and the Role of Tryptophan-8 in the Enhancement of
Phosphatidylserine Domains in Erythrocyte Membranes. Biochemistry 32, 12283-
12289.

Warren, G. B., Toon, P. A., Birdsall, N. J ., Lee, A. G., and Metcalfe, J. C. (1974)
Reconstitution of a Calcium Pump Using Defined Membrane Components. Proc.
Natl. Acad. Sci. USA 71, 622-626.

Brotherus, J. R., Jost, P. C., Griffith, O. H., Keana, J. F., Hokin, L. E. (1980)
Charge Selectivity at the Lipid-Protein Interface of Membranous Na, K-ATPase.
Proc. Natl. Acad. Sci. USA 77, 272-276.

Tocanne, J. F. (1992) Detection of Lipid Domains in Biological Membranes.
Comments Mol. Cell. Biophys. 8, 53-72.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

48.

172

Yechiel, E. and Edidin, M. (1987) Micrometer-Scale Domains in Fibroblast
Plasma Membranes. J. Cell Biol. 105, 755-760.

Yu, W., So, P. T., French, T., and Gratton, E. (1996) Fluorescence Generalized
Polarization of Cell Membranes: A Two-Photon Scanning Microscopy Approach.
Biophys. J. 70, 626-636.

Rodgers, W. and Glaser, M. (1993) Distributions of Proteins and Lipids in the
Erythrocyte Membrane. Biochemistry 32, 12591-12598.

Rodgers, W. and Glaser, M. (1991) Characterizaiton of Lipid Domains in
Erythrocyte Membranes. Proc. Natl. Acad. Sci. USA 88, 1364-1368.

Luzzati, V. and Husson, F. ( 1962) The Structure of the Liquid Crystalline Phases
of Lipid-Water Systems. J. Cell. Biol. 12, 207-219.

Shinitzky, M. (1984) Membrane Fluidity and Cellular Functions. in "Physiology
of Membrane Fluidity" (Shinitzky, M., Ed.) Vol. 1, pp 1-51. CRC Press, Boca
Raton, Florida.

Cossins, A. R. and Sinensky, M. (1984) Adaptation of Membranes to
Temperature, Pressure, and Exogenous Lipids. in "Physiology of Membrane
Fluidity" (Shinitzy, M., Ed.) Vol. 2, pp 1-20. CRC Press, Boca Raton, Florida.

Sinensky, M. (1974) Homeoviscous Adaptation-A Homeostatic Process that
Regulates the Viscosity of Membrane Lipids in Escherichia coli. Proc. Natl.
Acad. Sci. USA 71, 522-525.

Lenaz, G. and Degli Esposti, M. (1993) Membrane Bound Enzymes. in
"Biomembranes: Structural and Functional Aspects" Vol. 2 (Shinitzky, M., Ed.)
pp 83-197, VCH Publishers Inc., New York, New York.

Stockton, B. W. and Smith, I. C. P. (1976) A Deuterium Nuclear Magnetic
Resonance Study of the Condensing Effect of Cholesterol on Egg
Phophatidylcholine Bilayer Membranes. Chem. Phys. Lipids 17, 251-263.

Lee, A. G. (1991) Lipids and Their Effects on Membrane Proteins: Evidence
Against a Role of Fluidity. Prog. Lipid Res. 30, 323-348.

Silvius, J. R., Mak, N., and McElhaney, R. N. (1980) in "Membrane Fluidity:
Biophysical Techniques and Cellular Regulation" (Kates, M. and Kuksis, A.,
Eds.) pp 213-222, Humana Press, Clifton, New Jersey.

 

49.

50.

51.

53.

54.

55.

56.

57.

58.

 

 

 

 

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

173

Zakim, D., Kavecansky, J ., and Scarlata, S. (1992) Are Membrane Enzymes
Regulated by the Viscosity of the Membrane Environment? Biochemistry 31,
11589-11594.

Sen, A., Isac, T. V., and Sek-Wen, H. (1991) Bilayer Packing Stress and Defects
in Mixed Dilinoleoylphosphatidylethanolamine and

Palmitoyloleoylphosphatidylcholine and Their Susceptibility to Phospholipase A2.
Biochemistry 30, 4516-4521.

Dawson, R. M. C., Irvine, R. F., Bray, J ., and Quinn, P. J. (1984) Long-Chain
Unsaturated Diacylglycerols Cause a Perturbation in the Structure of

Phospholipid Bilayers Rendering them Susceptible to Phospholipase Attack.
Biochem. Biophys. Res. Comm. 125, 836-842.

Senisterra, G. and Epand, R. M. (1993) Role of Membrane Defects in the
Regulation of the Activity of Protein Kinase C. Arch. Biochem. Biophys. 300,
378-383.

Tate, M. W., Eikenberry, E. R, Turner, D. C., Shyamsunder, E., and Gruner, S.
M. (1991) Nonbilayer Phases of Membrane Lipids. Chem. Phys. Lipids 57, 147-
164.

Yechiel, E., Barenholz, Y., and Henis, Y. I. (1985) Lateral Mobility and
Organization of Phospholipids and Proteins in Rat Myocyte Membranes. J. Biol.
Chem. 260, 9132-9136.

Yechiel, E., Henis, Y. 1., and Barenholz, Y. (1986) Aging of Rat Heart
Fibroblasts: Relationship Between Lipid Composition, Membrane Organization
and Biological Properties. Biochim. Biophys. Acta 859, 95-104.

Moscona-Amir, E., Henis, Y. 1., Yechiel, E., Barenholz, Y., and Sokolovsky, M.
(1986) Role of Lipids in Age-Related Changes in the Properties of Muscarinic
Receptors in Cultured Rat Heart Myocytes. Biochemistry 25, 8118-8124.

Moscona-Amir, E., Henis, Y. 1., and Sololovsky, M. (1989) Aging of Rat Heart
Myocytes Disrupts Muscarinic Receptor Coupling That Leads to Inhibition of
cAMP Accumulation and Alters the Pathway of Muscarinic-Stimulated
Phosphoinositide Hydrolysis. Biochemistry 28, 7130-7137.

Asturias, F. J ., Pascolini, D., and Blasie, J. K. (1990) Evidence that Lipid Lateral
Phase Separation Induces Functionally Significant Structural Changes in the Ca2+
ATPase of the Sarcoplasmic Reticulum. Biophys. J. 58, 205-217.

 

59.

61.

62.

63.

65.

66.

67.

68.

69.

7O.

 

59.

60.

61.

62.

63.

65.

66.

67.

68.

69.

70.

174

Grainger, D. W., Reichert, A., Ringsdorf, H., and Salesse, C. (1989) An Enzyme
Caught in Action: Direct Imaging of the Hydrolytic Function and Domain
Formation of Phospholipase A2 in phosphatidylcholine monolayers. F EBS Lett.
252, 7 3-82.

Dibble, A. R. G., Hinderliter, A. K., Sando, J. J., and Biltonen, R. L. (1996) Lipid
Lateral Heterogeneity in Phosphatidylcholine/Phosphatidylserine/Diacylglycerol
Vesicles and Its Influence on Protein Kinase C Activation. Biophys. J. 71, 1877-
1890.

Almeida, P. F. F., Vaz, W. L. C., Thompson, T. E. (1993) Percolation and
Diffusion in Three-Component Lipid Bilayers: Effect of Cholesterol on an
Equimolar Mixture of Two Phosphatidylcholines. Biophys. J. 64, 399-412.

Muderhwa, J. M. and Brockman, H. L. (1992) Lateral Lipid Distribution Is a
Major Regulator of Lipase Activity. J. Biol. Chem. 267, 24184-24192.

Jacobson, K., Sheets, E. D., and Simson, R. (1995) Revisiting the Fluid Mosaic
Model of Membranes. Science 268, 1441-1442.

Melo, E. C. C., Lourtie, I. M. G., Sankaram, M. 8., Thompson, T. E., and Vaz, W.
L. C. (1992) Effects of Domain Connection and Disconnection on the Yields of
In-Plane Bimolecular Reactions in Membranes. Biophys. J. 63, 1506-1512.

Teissié, J ., Prats, M., LeMassu, A., Stewart, L., and Kates, M. (1990) Lateral

Proton Conduction in Monolayers of Phospholipids from Extreme Halophiles.
Biochemistry 29, 59-65.

Watts, A. (1987) Nuclear Magnetic Resonance Methods to Characterize Lipid-
Protein Interactions at Membrane Surfaces. J. Bioenerg. Biomembr. 19, 625-653.

Marsh, D. Watts, A., and Knowles, P. F. (1976) Evidence for Phase Boundary
Lipid: Permeability of Tempo-choline into Dimyristoylphosphatidylcholine
Vesicles at the Phase Transition. Biochemistry 15, 3570-3578.

Carruthers, A. and Melchoir, D. L. (1983) Studies of the Relationship Between
Bilayer Water Permeability and Bilayer Physical State. Biochemistry 22, 5797-
5807.

Rietschel, E. T. and Brade, H. ( 1992) Bacterial Endotoxins. Scientific American,
August, 54-61.

Rietschel, E. T., Brade, L., Lindner, B., and Zéihringer, U. (1992) Biochemistry of

Lipopolysaccharides. in “Bacterial Endotoxic Lipopolysaccharides” Vol. 1
(Morrison, D. C. and Ryan, J. L., Eds.) pp 3-41. CRC Press, Boca Raton, Florida.

 

71.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

 

 

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

175

Holst, O. and Brade, H. (1992) Chemical Structure of the Core Region of
Lipopolysaccharides. in “Bacterial Endotoxic Lipopolysaccharides” Vol. 1
(Morrison, D. C. and Ryan, J. L., Eds.) pp 135-170. CRC Press, Boca Raton,
Florida.

Takayama, K., Qureshi, N ., and Mascagni, P. (1983) Complete Structure of Lipid
A Obtained from the Lipopolysaccharides of the Heptoseless Mutant of
Salmonella typhimurium. J. Biol. Chem. 258, 12801-12803.

Irnoto, M., Yoshimura, H., Sakaguchi, N., Kusumoto, S., and Shiba, T. (1985)
Total Synthesis of Eschericia coli Lipid A. Tetrahedron Lett. 26, 1545.

Takayama, K. and Qureshi, N. (1992) Chemical Strucutre of Lipid A. in
“Bacterial Endotoxic Lipopolysaccharides” Vol. 1 (Morrison, D. C. and Ryan, J.
L., Eds.) pp 43-65. CRC Press, Boca Raton, Florida.

Coughlin, R. T., Peterson, A. A., Haug, A., Pownall, H. J. and McGroarty, E. J.
(1985) A pH Titration Study on the Ionic Bridging Within Lipopolysaccharide
Aggregates. Biochim. Biophys. Acta 821, 404-412.

Coughlin, R. T., Haug, A., and McGroarty, E. J. (1983) Physical Properties of
Defined Lipopolysaccharide Salts. Biochemistry 22, 2007-2013.

Nikaido, H. and Vaara, M. (1985) Molecular Basis of Bacterial Outer Membrane
Permeability. Microbiol. Rev. 49, 1-32.

Labinschinski, H. Barnickel, G., Bradaczek, H., Naumann, D., Rietschel, E. T.,
and Giesbrecht, P. (1985) High State of Order of Isolated Bacterial
Lipopolysaccharide and Its Possible Contribution To the Permeation Barrier
Property of the Outer Membrane. J. Bacteriol. 162, 9-20.

Wang, Y. and Hollingsworth, R. I. (1996) An NMR spectroscopy and Molecular
Mechanics Study of the Molecular Basis for the Suprarnolecular Structure of
Lipopolysaccharides. Biochemistry 35, 5647-5654.

Kato, N., Ohta, M., Kido, N ., Ito, H., Arakawa, Y., and Naito, S. (1990) Electron
Microscopic Observation of Crystals of Escherichia coli K-12
Lipopolysaccharide. Microbiol. Immunol. 34, 665-674.

Kato, N. (1993) Crystallization and Electron Microscopy of Bacterial
Lipopolysaccharide (LPS). Micron 24, 91-114.

Michalek, S. M., Moore, R. N., McGhee, J. R., Rosenstreich, D. L., and
Mergenhagen, S. E. (1980) The Primary Role of Lymphoreticular Cells in the
Mediation of Host Responses to Bacterial Endotoxin. J. Inf. Dis. 141, 55-63.

 

 

83.

84.

85.

86.

87.

88.

89.

90.

91.

 

 

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

176

Morrison, D. C. and Ryan, J. L. (1987) Endotoxins and Disease Mechanisms.
Ann. Rev. Med. 38, 417-432.

Alving, C. R. (1993) Lipopolysaccharide, Lipid A, and Liposomes Containing
Lipid A as Immunologic Adjuvants. Immunabiol. 187, 430-446.

Dinarello, C. A., Cannon, J. G., Wolff, S. M., Bemheim, H. A., Beutler, B.,
Cerami, A., Figari, I. S., Palladino, Jr., M. A., O’Connor, J. V. (1986) Tumor
Necrosis Factor (Cachectin) is an Endogenous Pyrogen and Induces Production of
Interleukin-1. J. Exp. Med. 163, 1433-1450.

Galanos, C. and Freudenberg, M. A. (1990) Tumor Necrosis Factor Mediates
Endotoxin Shock: The Protective Effects of Antibodies and Cortisone. in “Tumor
Necrosis Factor: Structure, Mechanism of Action, Role in Disease Therapy”
(Bonavida, B. and Granger, G., Eds.) pp 187-193. Karger, Basel.

Murphy, P. A., Simon, P. L., and Willoughby, W. F. (1980) Endogenous
Pyrogens Made by Rabbit Peritoneal Exudate Cells are Identicle with LAF’s
Made by Rabbit Alveolar Macrophages. J. Immunol. 124, 2498-2501.

Okusawa, S., Gelfand, J. A., Ikejima, T., Connolly, R. J ., and Dimarello, C. A.
(1988) Interleukin-l Induces a Shock-Like State in Rabbits. Synergism With
Tumor Necrosis Factor and the Effect of Cyclooxygenase Inhibition. J. Clin.
Invest. 81, 1162-1172.

Kasai, N. and Nowotny, A. (1967) Endotoxic Glycolipid from a Heptoseless
Mutant of Salmonella minnesota. J. Bacteriol. 94, 1824-1836.

Haeffner-Cavaillon, N ., Caroff, M., and Cavaillon, J .-M. (1989) Interleukin-1
Induction by Lipopolysaccharide: Structural Requirement of the 3-deoxy-D-
manna-2-octulosonic acid (Kdo). Mol. Immunol. 26, 485-494.

Rietschel, E. T., Brade, H., Brade, L., Brandenburg, K., Schade, U., Seydel, U.,
Zéihringer, U., Galanos, C., Liideritz, O., Westphal, O., Labischinski, H.,
Kusumoto, S., and Shiba, T. (1987) Lipid A, the Endotoxic Center of Bacterial

Lipopolysaccharides: Relation of Chemical Structure to Biological Activity.
Prog. Clin. Biol. Res. 231, 25-53.

Rietschel, E. T., Brade, L., Holst, O., Kulshin, V. A., Lindner, B., Moran, A. P.,
Schade, U. F., Zéihringer, U., and Brade, H. (1990) Molecular Structure of

Bacterial Endotoxin in Relation to Bioactivity. in “Cellular and Molecular
Aspects of Endotoxin Reactions” Vol. 1 (Nowotny, A., Spitzer, E. J ., Ziegler, E.
J ., Eds.) pp 15-32. Elsevier Science Publishers B. V., Amsterdam.

 

 

 

 

93.1

94.

95.

96.

 

97.

98.

99.

100

103.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

177

Takada, H. and Kotani, S. (1992) Structure-Function Relationships of Lipid A. in
“Bacterial Endotoxic Lipopolysaccharides” Vol. 1 (Morrison, D. C. and Ryan, J.
L., Eds.) pp 107-134. CRC Press, Boca Raton, Florida.

Reitschel, E. T., Kirikae, T., Schade, F. U., Ulmer, A. J ., Holst, O., Brade, H.,
Schmidt, G., Mamat, U., Grimmecke, H.-D., Kusumoto, S., and Zéihringer, U.

(1993) The Chemical Structure of Bacterial Endotoxin in Relation to Bioactivity.
Immunobiol. 187, 169-190.

Strittmatter, W., Weckesser, J ., Salimath, P. V., and Galanos, C. (1983) Nontoxic
Lipopolysaccharide from Rhodapseudomonas sphaeroides ATCC 17023. J.
Bacteriol. 155, 153-158.

Qureshi, N., Takayama, K., Meyer, K. C., Kirkland, T. N ., Bush, C. A., Chen, L.,
Wang, R., and Cotter, R. J. (1991) Chemical Reduction of 3-Oxo and Unsaturated
Groups in Fatty Acids of Diphosphoryl Lipid A for the Lipopolysaccharide of
Rhodopseudamonas sphaeroides. J. Biol. Chem. 266, 6532-6538.

Kirkland, T. N., Qureshi, N ., and Takayama, K. (1991) Diphosphoryl Lipid A
Derived from Lipopolysaccharide (LPS) of Rhodopseudamanas sphaeroides
Inhibits Activation of 7OZ/3 Cells by LPS. Inf. Immun. 59, 131-136.

Takayama, K., Qureshi, N., Beutler, B., and Kirkland, T. N. (1989) Diphosphoryl
Lipid A from Rhodapseudomanas sphaeroides ATCC 17023 Blocks Induction of
Cachectin in Macrophages by Lipopolysaccharide. Infi Immun. 57, 1336-1338.

Qureshi, N., Takayama, N., and Kurtz, R. (1991) Diphosphoryl Lipid A Obtained
from the Nontoxic Lipopolysaccharide of Rhadopseudomonas sphaeroides Is an
Endotoxin Antagonist in Mice. Inf. Immun. 59, 441-444.

Raetz, C. R. H., Ulevitch, R. J ., Wright, S. D., Sibley, C. H., Ding, A., and
Nathan, C. F. (1991) Gram-Negative Endotoxin: An Extraordinary Lipid With
Profound Effects on Eukaryotic Signal Transduction. FASEB J. 5, 2652-2660.

Morrison, D. C. (1989) The Case for Specific Lipopolysaccharide Receptors
Expressed on Mammalian Cells. Microbial. Pathogen. 7, 389-398.

Lei, M.-G. and Chen, T.-Y. (1992) Cellular Membrane Receptors for
Lipopolysaccharide. in “Bacterial Endotoxic Lipopolysaccharides” Vol. 1
(Morrison, D. C. and Ryan, J. L., Eds.) pp 253-267. CRC Press, Boca Raton,
Florida.

Lei, M.-G. and Morrison, D. C. (1988) Specific Endotoxic Lipopolysaccharide-
Binding Proteins on Murine Spenocytes. 1. Detection of Lipopolysaccharide-
Binding Sites on Splenocytes and Spenocyte Subpopulations. J. Immunol. 141,
996-1005.

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

178

Lei, M.-G. and Morrison, D. C. (1988) Specific Endotoxic Lipopolysaccharide-
Binding Proteins on Murine Spenocytes. H. Membrane Localization and Binding
Characteristics. J. Immunol. 141, 1006-1011.

Hampton, R. Y., Golenbock, D. T., Raetz, R. H. (1988) Lipid A'Binding Sites in
Membranes of Macrophage Tumor Cells. J. Biol. Chem. 263, 14802-14807.

Kara-Kuge, S., Amano, F., Nishijima, M., and Akamatsu, Y. (1990) Isolation of a
Lipopolysaccharide (LPS)-Resistant Mutant, With Defective LPS Binding, of
Cultured Macrophage-Like Cells. J. Biol. Chem. 265, 6606-6610.

Kirkland, T. N., Virca, G. D., Kuus-Reichel, T., Multer, F. K., Kim, S. Y.,
Ulevitch, R. J ., and Tobias, P. S. (1990) Identification of Lipopolysaccharide-
Binding Proteins in 702/3 Cells by Photoaffinity Cross-Linking. J. Biol. Chem.
265, 9520—9525.

Wright, S. D. and Jong, M. T. C. (1986) Adhesion-Promoting Receptor on Human
Macrophages Recognize Eschericia coli by Binding to Lipopolysaccharide. J.
Exp. Med. 164, 1876-1888.

Wright, S. D., Ramos, R. A., Tobias, P. S., Ulevitch, R. J ., and Mathison, J. C.
( 1990) CD14, a Receptor for Complexes of Lipopolysaccharide (LPS) and LPS
Binding Protein. Science 249, 1431-1433.

Raetz, C. R. (1993) Bacterial Endotoxins: Extraordinary Lipids That Activate
Eucaryotic Signal Transduction. J. Bacterial. 175, 5745-5753.

Simmons, D. L., Tan, S., Tenen, D. G., Nicholson-Weller, A., and Seed, B.
(1989) Monocyte Antigen CD14 is a Phospholipid Anchored Membrane Protein.
Blood 73, 284-289.

Schletter, J., Brade, H., Brade, L., Kruger, C., Loppnow, H., Kusumoto, S.,
Rietschel, E. T., Flad, H.-D., and Ulmer, A., J. (1995) Binding of
Lipopolysaccharide (LPS) to an 80-Kilodalton Membrane Protein of Human Cells
is Mediated by Soluble CD14 and LPS-Binding Protein. Inf. Immun. 63, 2576-
2580.

Bazil, V. and Strominger, J. L. (1991) Shedding as a Mechanism of Down-
Modulation of CD14 on Stimulated Human Monocytes. J. Immunol. 147, 1567-
1574.

 

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

124.

125.

179

Cavaillon, J .-M., Munoz, C., Fitting, C., Couturier, C., and Haeffner-Cavaillon,
N. (1990) Signals Involved in Interleukin-1 Production Induced by Endotoxins. in
“Cellular and Molecular Aspects of Endotoxin Reactions” Vol. 1 (Nowotny, A.,
Spitzer, J. J ., and Ziegler, E. J ., Eds.) pp. 257-267. Elsevier Science Publishers B.
V., Amsterdam.

Morrison, D. C. (1985) Nonspecific Interactions of Bacterial Lipopolysaccharides
With Membranes and Membrane Components. in “Handbook of Endotoxin” Vol.
3 (Berry, L. J ., Ed.) pp 25-55. Elsevier Science Publishers B. V., Amsterdam.

Schuster, B. G., Palmer, R. F., and Aronson, R. S. (1970) The Effect of Endotoxin
on Thin Lipid Bilayer Membranes. J. Membr. Biol. 3, 67-72.

Jackson, S. K., James, P. E., Rowlands, C. C., Mile, B. (1992) Binding of
Endotoxin to Macrophages; Interactions of Spin-Labeled Saccharide Residues.
Biochim. Biophys. Acta 1135, 165-170.

Carr, Jr., C. and Morrison, D. C. (1984) Lipopolysaccharide Interaction with
Erythrocyte Membranes. Inf. Immun. 43, 600-606.

Carr, Jr., C. and Morrison, D. C. (1984) A Two Step Mechanism for the
Interaction of Re-Lipopolysaccharide with Erythrocyte Membranes. Rev. Inf. Dis.
6, 497-500.

Jacobs, D. M. (1984) Structural Features of Binding of Lipopolysaccharides to
Murine Lymphocytes. Rev. Inf. Dis. 6, 501-505.

Jacobs, D. M. and Price, R. M. ( 1987) A Model for Lipopolysaccharide-
Membrane Interaction. Adv. Exp. Med. Biol. 216A, 691-699.

Yeh, H. Y. and Jacobs, D. M. (1992) Characterization of Lipopolysaccharide
Fractions and Their Interactions with Cells and Model Membranes. J. Bacteriol.
174, 336-341.

Jacobs, D. M., Yeh, H., and Price, R. M. (1990) Fluorescent Detection of
Lipopolysaccharide Interactions With Model Membranes. Adv. Exp. Med. Biol.
256, 233-245.

Price, R. M. and Jacobs, D. M. (1986) Fluorescent Detection of
Lipopolysaccharide Interactions with Model Membranes. Biochim. Biophys. Acta.
859, 26-32.

Galanos, C. (1975) Physical State and Biological Activity of Lipopolysaccharides.
Toxicity and Irnmunogenicity of the Lipid A Component. Z. Immun.-Farsch. Bd.
149, 214-229.

 

 

 

 

a)

ci-l

 

 

 

12‘

131

13

13

13¢

126.

127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

180

Galanos, C. and Liideritz, O. (1984) Lipopolysaccharides: Properties of an

Amphipathic Molecule. in “Handbook of Endotoxin” Vol. 1 (Rietschel, E. T.,
Ed.) pp 46-58. Elsevier Science Publishers B. V., Amsterdam.

Brandenburg, K, Koch, M. H. J ., and Seydel, U. (1992) Phase diagram of Deep
Rough Mutant Lipopolysaccharide from Salmonella minnesota. J. Struct. Biol.
108, 93-106.

Kato, N., Ohta, M., Kido, N., Naito, S., and Kuno, T. (1984) Ultrastructure of
Klebsiella 03 Lipopolysaccharide Isolated from Culture Supernatant: Structure of
Various Uniform Salt Forms. Microbiol. Immunol. 28, 545-557.

Brandenburg, K., Mayer, H., Koch, M. H. J ., Weckesser, J ., Rietschel, E. T., and
Seydel, U. (1993) Influence of the Suprarnolecular Structure of Free Lipid A on
its Biological Activity. Eur. J. Biochem. 218, 555-563.

Vincent, J. M. (1982) Serology. in “Nitrogen Fixation: Rhizobium” Vol. 2
(Broughton, W., Ed.) pp 235-273. Clarendom Press, Oxford.

Bhat, U. R., Mayer, H., Yokota, A., Hollingsworth, R. 1., and Carlson, R. W.
(1991) Occurrence of Lipid A Variants with 27-Hydroxyoctacosanoic Acid in
Lipopolysaccharides from Members of the Family Rhizobiaceae. J. Bacteriol.
173, 2155-2159.

Ying Wang (1995) Major Revision of the Proposed Structure of the Lipid A
Region of the Lipopolysaccharides of Rhizobium leguminosarum Biovars. in
“Analysis of the Molecular Basis for the Suprarnolecular Structure of
Lipopolysaccharides” PhD dissertation. pp 86-138. Michigan State University,
Department of Chemistry.

Lill-Elghanian, D. A. (1992) Biological Responses Induced by Rhizobium trifolii
ANU 843 LPS. in “Dissection of Endotoxin Biological Activities Using Variant
Lipid A and Synthetic Antagonists” PhD dissertation. pp 79-140. Michigan State
University, Department of Biochemistry.

Urbanik-Sypniewska, T., Seydel, U., Greck, M., Weckesser, J ., and Mayer, H.
(1989) Chemical Studies on the Lipopolysaccharide of Rhizobium meliloti. Arch.
Microbiol. 152, 527-532.

Loppmow, H., Libby, P., Freudenberg, M., Krauss, J. H., Weckesser, J ., and
Mayer, H. (1990) Cytokine Induction by Lipopolysaccharide (LPS) Corresponds
to Lethal Toxicity and Is Inhibited by Nontoxic Rhodobacter capsulatus LPS. Inf.
Immun. 58, 3743-3750.

Weissbach, A. and Hurwitz, J. ( 1958) Formation of 2-Keto-3-Deoxyheptonic
Acid in Extracts of Escherichia coli B. J. Biol. Chem. 234, 705-709.

 

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

181

Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A., and Smith, F. (1956)
Colorimetric Method for Determination of Sugars and Related Substances. Anal.
Chem. 350-354.

Blumenkrantz, N. and Asboe-Hansen, G. (1973) New Method for Quantitative
Determination of Uronic Acids. Anal. Biochem. 54, 484-489.

Lowry, O. H., Roberts, N. R., Leiner, K. Y., Wu, M.-L., and Farr, L. (1954) The
Quantitative Histochemistry of Brain. 1. Chemical Methods. J. Biol. Chem. 207,
1-17.

Westphal, O. and J ann, K. (1965) Bacterial Lipopolysaccharides: Extraction with
Phenol-Water and Further Applications of the Procedure. in “Methods in
Carbohydrate Chemistry” Vol. 5 (Whistler, R. L., Ed.) pp 83-91. Academic Press,
New York, New York.

Raschke, W. C., Baird, 8., Ralph, P., and Nakoinz, I. (1978) Functional
Macrophage Cell Lines Transformed by Abelson Leukemia Virus. Cell 15, 261-
267.

Rivera, M. and McGroarty, E. J. (1989) Analysis of a Common-Antigen
Lipopolysaccharide from Pseudomonas aeruginosa. J. Bacteriol. 171, 2244—2248.

Levin, J. and Bang, F. (1964) The Role of Endotoxin in the Extracellular
Coagulation of Limulus Blood. Bull. Johns Hopkins Hosp. 115, 265-274.

Von Eschen, K. (1992) Monophosphoryl Lipid A and Irnmunotherapy. in
“Bacterial Endotoxic Lipopolysaccharides” Vol. 2 (Ryan, J. L. and Morrison, D.
C., Eds.) pp 411—428.

Pugin, J ., Heumann, D., Tomasz, A., Kravchenko, V., Akamatsu, Y., Nishijima,
M., Glauser, M. P., Tobias, P. S., and Ulevitch, R. J. (1994) CD14 is a Pattern
Recognition Receptor. Immunity 1, 509-516.

Delude, R. L., Savedra, Jr., R., Zhao, H., Thieringer, R., Yamamoto, S., Fenton,
M. J ., and Golenbock, D. T. (1995) CD14 Enhances Cellular Responses to
Endotoxin Without Imparting Ligand-Specific Recognition. Proc. Natl. Acad. Sci.
USA 92, 9288-9292.

Wurfel, M. M. and Wright, S. D. (1997) Lipopolysaccharide-Binding Protein and
Soluble CD14 Transfer Lipopolysaccharide to Phospholipid Bilayers: Preferential
Interaction with Particular Classes of Lipid. J. Immun. 158, 3925-3934.

 

148.

149.

150.

151.

152.

153.

154.

155.

156.

157.

182

Yu, 8., Hailman, E., and Wright, S. D. (1997) Lipopolysaccharide Binding
Protein and Soluble CD14 Catalyze Exchange of Phospholipids. J. Clin. Invest.
99, 315-324.

Schromm, A. B., Brandenburg, K., Rietschel, E. T., Flad, H. D., Carroll, S. F.,
and Seydel, U. (1996) Lipopolysaccharide-Binding Protein Mediates CD14-

Independent Intercalation of Lipopolysaccharide into Phospholipid Membranes.
FEBS Lett. 399, 267-271.

Bérubé, L. R. (1995) Endotoxic Lipopolysaccharide from Salmonella
typhimurium Modifies Lipid Alkyl Chain Dynamics in Model Membranes But
Non-Endotoxic Lipopolysaccharide from Rhodopseudamanas sphaeroides Does
Not. in “Studies on the Dynamics of Lipid Alkyl Chains in Biological Membranes
Subjected to Environmental Stress” PhD Dissertation, pp 115-144, Michigan
State University, Department of Biochemistry.

Bérubé, L. R. and Hollingsworth, R. 1. (1995) Membrane Lipid Alkyl Chain
Motional Dynamics is Conserved in Sarcina ventriculi Despite pH-Induced
Adaptative Structural Modifications Including Alkyl Chain Tail-to-Tail Coupling.
Biochemistry 34, 12005-12011.

Bara, J ., Lallier, R., Trude], M., Brailovsky, C., and Nigam, V. N. (1973)
Molecular Models on the Insertion of Salmonella minnesota R-forrn Glycolipid
into the Cell Membrane of Normal and Transformed Cells. Eur. J. Biochem. 35,
495-498.

Shands, J. W. (1973) Affinity of Endotoxin for Membranes. in “Bacterial
Lipopolysaccharides” (Kass, E. H. and Wolff, S. M., Eds.) pp 189-193. University
of Chicago Press, Chicago, Illinois.

Kabir, S. and Rosenstreich, D. L. (1977) Binding of Bacterial Endotoxin to
Murine Spleen Lymphocytes. Inf. Immun. 15, 156-164.

Symons, D. B. A. and Clarson, C. A. (1979) The Binding of Lipopolysaccharide
to the Lymphocyte Surface. Immunology 38, 503-508.

Morrison, D. C., Oades, Z. G., Dipietro, D. (1981) Endotoxin-Initiated Membrane
Changes in Rabbit Platelets. in “Pathophysiological Effects of Endotoxins at the
Cellular Level” (Majde, J. A. and Person, R. J ., Eds.) pp 47-64. Alan R. Liss, New
York, New York.

Rothfield, L. and Home, R. W. (1967) Reassociation of Purified
Lipopolysaccharide and Phospholipid of the Bacterial Cell Envelope: Electron
Microscopic and Monolayer Studies. J. Bacteriol. 93, 1705-1721.

 

 

 

,
1x

 

16

16

158.

159.

160.

161.

162.

163.

164.

165.

166.

167.

168.

169.

183

Morrison, D. C. and Rudbach, J. A. (1981) Endotoxin-Cell Membrane
Interactions Leading to Trans-Membrane Signaling. Contemp. T op. Mol.
Immunol. 8, 187-218.

Rothfield, L. and Takeshita, M. ( 1965) The Role of the Cell Envelope
Phospholipid in the Enzymatic Synthesis of Bacterial Lipopolysaccharide:

Binding of Transferase Enzymes to a Lipopolysaccharide:Lipid Complex.
Biochem. Biophys. Res. Comm. 20, 521-527.

Lui, M.-S., Onji, T., and Snelgrove, N. E. (1982) Changes in Phase Transition
Temperature of Phospholipids Induced by Endotoxin. Biochim. Biophys. Acta
710, 248-251.

Rottem, S. (1978) The Effect of Lipid A on the Fluidity and Permeability
Properties of Phospholipid Dispersions. FEBS Lett. 95, 121-124.

Mackay, A. L., Nichol, C. P., Weeks, G., and Davis, J. H. (1984) A Proton and
Deuterium Nuclear Magnetic Resonance Study of Orientational Order in Aqueous
Dispersions of Lipopolysaccharide and Lipopolysaccharide-
Dipalmitoylphosphatidylcholine Mixtures. Biochim. Biophys. Acta. 774, 18 1- 187.

Esser, A. F. and Russell, S. W. (1979) Membrane Perturbation of Macrophages
Stimulated by Bacterial Lipopolysaccharide. Biochem. Biophys. Res. Comm. 87,
532-540.

Larsen, N. E., Enelow, R. I., Simons, E. R., and Sullivan, R. (1985) Effect of
Bacterial Endotoxin on the Transmembrane Electrical Potential and Plasma
Membrane Fluidity of Human Monocytes. Biochim. Biophys. Acta 815, 1-8.

Salgia, R., Becker, J. H., and Sayeed, M. M. (1993) Altered Membrane Fluidity in
Rat Hepatocytes During Endotoxic Shock. Mol. Cell. Biochem. 121, 143-148.

Fried, V. A. and Rothfield, L. I. (1978) Interactions Between Lipopolysaccharide
and Phosphatidylethanolamine in Molecular Monolayers. Biochim. Biophys. Acta
514, 69-82.

Takeuchi, Y. and Nikaido, H. ( 1981) Persistence of Segregated Phospholipid
Domains in Phospholipid-Lipopolysaccharide Mixed Bilayers: Studies with Spin-
Labeled Phospholipids. Biochemistry 20, 523-529.

Takeuchi, Y. and N ikaido, H. (1984) Physical Interaction Between Lipid A and
Phospholipids: A Study with Spin-Labeled Phospholipids. Rev. Infi Dis. 6, 488-
492.

Vold, R. L., Waugh, J. S., Klein, M. P., and Phelps, D. E. (1968) J. Chem. Phys.
48, 3831-3832.

170.

171.

172.

173.

174.

175.

176.

177.

178.

179.

180.

184

Lewis, R. N. A. H. and McElhaney, R. N. (1992) The Mesomorphic Phase
Behavior of Lipid Bilayers. in “The Structure of Biological Membranes” (Yeagle,
P., Ed.) pp 73-155. CRC Press, Boca Raton, Florida.

Lentz, B. R. (1989) Membrane “fluidity” as Detected by Diphenylhexatriene
Probes. Chem. Phys. Lipids 50, l71-190.

Freeman, R. (1988) “A Handbook of Nuclear Magnetic Resonance” John Wiley
and Sons, Inc., New York, New York.

Kroon, P. A., Kainosho, M., and Chan, S. I. (1976) Proton Magnetic Resonance
Studies of Lipid Bilayer Membranes: Experimental Determination of Inter- and

Intramolecular Nuclear Relaxation Rates in Sonicated Phosphatidylcholine
Bilayer Vesicles. Biochim. Biophys. Acta 433, 282-293.

Fourqueray, B., Philippe, C., Amrani, A., Perez, J ., and Baud, L. (1992) Heat
Shock Prevents Lipopolysaccharide-Induced Tumor Necrosis Factor-0t Synthesis
by Rat Mononuclear Phagocytes. Eur. J. Immunol. 22, 2983-2987.

Tomasovic, S. P. and Klostergaard, J. (1991) Bacterial Endotoxin
Lipopolysaccharide Modulates Synthesis of the 70 kDa Heat Stress Protein
Family. Int. J. Hyperthermia 7, 643-651.

Ensor, J. E., Wiener, S. M., McCrea, K. A., Viscardi, R. M., Crawford, E. K., and
Hasday, J. D. (1994) Differential Effects of Hyperthermia on Macrophage

Interleukin-6 and Tumor Necrosis Factor-0t Expression. Am. J. Physiol. 266,
C967-C974.

Chancellor-Freeland, C., Zhu, G. F., Kage, R., Beller, D. 1., Leeman, S. E., and
Black, P. H. ( 1995) Substance P and Stress-Induced Changes in Macrophages.
Ann. N. Y. Acad. Sci. 771, 472-484.

Zhu, G. F., Chancellor-Freeland, C., Berrnan, A. S., Kage, R., Leeman, S. E.,
Beller, D. 1., and Black, P. H. (1996) Endogenous Substance P Mediates Cold
Water Stress-Induced Increase in Interleukin-6 Secretion from Peritoneal
Macrophages. J. Neuroscience 16, 3745-3752.

Kim, K.-I., Lill-Elghanian, D. A., and Hollingsworth, R. I. (1994) Design,
Synthesis and Biological Properties of Structurally Non-Related Endotoxin
Inhibitors. Bioorg. Med. Chem. Lett. 4, 1691-1696.

Danner, R. L., Van Dervort, A. L., Doefler, M. E., Stuetz, P., Parrillo, J. E. (1990)
Antiendotoxin Activity of Lipid A Analogues: Requirements of the Chemical
Structure. Phann. Res. 7, 260-263.

181.

182.

183.

184.

185.

186.

187.

188.

189.

190.

191.

185

Danner, R. L., Joiner, K. A., Parrillo, J. E. (1987) Inhibition of Endotoxin-
Induced Priming of Human Neutrophils by Lipid X and 3-Aza-Lipid X. J. Clin.
Invest. 80, 605-612.

Shiozaki, M., Deguchi, N., Macindoe, W. M., Arai, M., Miyazaki, H., Mochizuki,
T., Tatsuta, T., Ogawa, J ., Maeda, H., Kurakata, S. (1996) Carbohyd. Res. 283,
27-51.

Jacobs, D. M. and Morrison, D. C. (1977) Inhibition of the Mitogenic Response to
Lipopolysaccharide (LPS) in Mouse Spleen Cells by Polymyxin B. J. Immunol.
118, 21-27.

Cavaillon, J. M. and Haeffner-Cavaillon, N. (1986) Polymyxin-B Inhibition of
LPS-Induced Interleukin-1 Secretion by Human Monocytes is Dependent Upon
the LPS Origin. Mal. Immunol. 23, 965-969.

Friedman, H. (1990) Endotoxin Effects on Marijuana Induced Suppression of
Immune Responses. in “Cellular and Molecular Aspects of Endotoxin Reactions”
Vol. 1 (Nowotny, A., Spitzer, J. J ., Ziegler, E. J ., Eds.) pp 379-385. Elsevier
Science Publishers, New York, New York.

Bermudez, L. E., Wu, M., Martinelli, J ., Young, L. S. (1991) Ethanol Affects
Relaease of TNF and GM-CSF and Membrane Expression of TNF Receptors by
Human Macrophages. Lymphakine C ytokine Res. 10, 413-419.

Nair, M. P., Schwartz, S. A., Kronfol, Z. A., Hill, E. M., Sweet, A. M., Greden, J.
F. (1994) Suppression of Tumor Necrosis Factor Production by Alcohol in
Lipopolysaccharide-Stimulated Culture. Alcohol Clin. Exp. Res. 18, 602-607.

Liang, J. F. and Akaike, T. (1997) Inhibition of Lipopolysaccharide and Cytokine
Mixture-Mediated Hepatocyte Nitric Oxide Synthesis by Dimethyl Sulfoxide.
Biochem. Biophys. Res. Comm. 239, 517-521.

Berry, L. J. (1985) Introduction. in “Handbook of Endotoxin: Cellular Biology of
Endotoxin” Vol. 3 (Berry, L. J ., Ed.) pp xvii-xxi. Elsevier Science Publishers,
B.V., Amersterdam.

Mohri, M., Spriggs, D. R., and Kufe, D. (1990) Effects of Lipopolysaccharide on
Phospholipase A2 Activity and Tumor Necrosis Factor Expression in HL-60
Cells. J. Immunol. 144, 2678-2682.

McKenna, T. M., Fan, S. X., and Li, S. (1997) Lipopolysaccharide-Responsive
Protein Kinase C Isotypes in the Adult Rat Aorta. Shock 7, 269-273.

192.

193.

194.

195.

196.

197.

198.

199.

200.

201.

186

Shapira, L., Takashiba, S., Champagne, C., Amar, S., and Van Dyke, T. E. (1994)
Involvement of Protein Kinase C and Protein Tyrosine Kinase in
Lipopolysaccharide-Induced TNF-0t and IL-IB Production by Human Monocytes.
J. Immunol. 153, 1818-1824.

Levin, J. (1985) The History of the Development of the Limulus Amebocyte
Lysate Test. in “Bacterial Endotoxins: Structure, Biomedical Significance, and

Detection with the Limulus Amebocyte Lysate Test” (ten Cate, J. W., Biiller, H.
R., and Sturk, A., Eds.) pp 3-30. Alan R. Liss, Inc., New York, New York.

Mann, K. G. (1984) Membrane-Bound Enzyme Complexes in Blood Coagulation.
in “Progress in Hemostasis and Thrombosis”, Vol. 7, pp 1-23. Grune and Stratton.

Nelsestuen, G. L. (1984) Calcium Function in Blood Coagulation. in “Metal Ions
in Biological Systems”, Vol. 17 (Sigel, H., Ed.) pp 353-380. Marcel Dekker, Inc.,
New York.

Levin, J. (1979) The Reaction Between Bacterial Endotoxin and Amebocyte
Lysate. in “Progress in Clinical and Biological Research: Biomedical
Applications of the Horseshoe Crab (Limulidae)”, Vol. 29 (Cohen, E., Ed.) pp
131-146. Alan R. Liss, Inc., New York, NY.

Lui, T.-Y., Seid, R. C. Jr., Tai, J. Y., Liang, S.-M., Sakmar, T. P., Robbins, J. B.

( 1979) Studies on Limulus Lysate Coagulation System. in “Progress in Clinical
and Biological Research: Biomedical Applications of the Horseshoe Crab
(Limulidae)”, Vol. 29 (Cohen, E., Ed.) pp 147-158. Alan R. Liss, Inc., New York,
NY.

Bona, C., Juy, D., Truffa-Bachi, P., and Kaplan, G. J. (1976) Binding, Capping
and Intemalization of Lipopolysaccharide in Thymic and Nonthymic
Lymphocytes of the Mouse. Biological and Autoradiographical Studies. J.
Microbiol. Cell 25, 47-56.

Rosenstreich, D. L. (1985) Genetic Control of Endotoxin Response: C3H/HeJ
Mice. in “Handbook of Endotoxin” Vol. 3 (Berry, L. J ., Ed.) pp 82-122, Elsevier
Science Publishers, B.V., Amsterdam.

Haeffner-Cavaillon, N., Cavaillon, J .-M., and Szabo, L. (1985) Cellular Receptors

for Endotoxin. in “Handbook of Endotoxin” Vol. 3 (Berry, L. J ., Ed.) pp 1-24,
Elsevier Science Publishers, B.V., Amsterdam.

Chaby, R., Morelec, M.-J., Enserqueix, D., and Girard, R. (1986) Membrane
Glycolipid and Phospholipid Composition of Lipopolysaccharide—Responsive and
-Nonresponsive Murine B Lymphocytes. Infect. Immun. 52, 777-785.

202.

203.

187

Yohe, H. C., Berenson, C. S., Cuny, C. L., and Ryan, J. L. (1990) Altered B-
Lymphocyte Membrane Architecture Indicated by Ganglioside Accessibility in
C3H/HeJ Mice. Infect. Immun. 58, 2888-2894.

Davis, J. H., Jeffrey, K. R., and Bloom, M. (1978) Spin-Lattice Relaxation as a
Function of Chain Position in Perdeuterated Potassium Palmitate. J. Magn.

Resonance 29, 19 1-199.