LIBRARY

Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

“ DATE DUE DATE DUE DATE DUE

J
.L.

r"‘ ‘ iMTj ‘1 ‘
. 'LA‘L U L} ‘ ‘-

 

 

 

'

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L

 

MSU Is An Affirmative Action/Equal Opportunity Institution
twins-oi

 

 

 

 

A STRUCTURAL AND COMPUTATIONAL ANALYSIS OF MOLECULAR
MECHANISMS FOR MEMBRANE ADAPTATION TO EXTREME STRESS

By

Seunho Jung

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Biochemistry

1993

ABSTRACT

A STRUCTURAL AND COMPUTATIONAL ANALYSIS OF
MOLECULAR MECHANISMS FOR MEMBRANE
ADAPTATION TO EXTREME STRESS
By
Seunho Jung

One of the most complex systems in cells is the biological membrane. This system
displays a very sensitive, highly tuned adaptive mechanism to changes in environmental
parameters. This adaptive response includes changes in the composition of lipids as well
as changes in their structure. The lipid biosynthetic machinery is intimately coupled to the
dynamic state of the membrane.

A new adaptive system was discovered in Sarcina ventriculi and its molecular
mechanism was studied. In this response, a family of transmembrane lipid species is
formed in response to any perturbation which increases the dynamical state of the
membrane. This response, in effect, resulted in a transition from a bilayer to a bipolar
monolayer membrane system. A similar phenomenon was observed in Clostridium
thermohydrosulfuricum. It was determined that this transition is a spontaneous event
which does not require new protein synthesis and that the pairing partners is selected on a
strict probability basis. The possible effects of transmembrane species are capable of
significantly reducing the motional freedom of regular chain length species which are in
close proximity. They also lead to increased packing of the void which usually exists in the
area between the bilayer leaflets.

To my parents, Mr. Takban Jung and Mrs. Joongson Jun Jung
and my wife, Wonjae
and my son

Sungmo

ACKNOWLEDGMENTS

I thank Rawle I. Hollingsworth for being an excellent guidance and precious advice
during my graduate studies. His enthusiastic passion for the truth is a great role model to
me. I also thank my wife, Wonjae Rho Jung, for her tolerance, love and prayer. I also
thank Drs. Ester McGroarty, Zach Burton, Ned Jackson and William Wells for their kind
service as members of my guidance committee. I thank those persons who have helped me
in this endeavor. Thank to Drs. Gregory Zeikus and Susan Lowe for helping and allowing
me to use of their cell culture facilities; to Dr. YounSuk J ang for helping me about FAB-
MS experiments; to Dr. Long Le for the NMR training. I also thank our lab. colleagues,
Debbie, Chuck, Maria, Rob, Luc, Ben and Ying. Particularly, I thank K. Kim for his
kindness throughout the research. I also thank my friends, Jaeman Lee, JaeNeung Kim
and Dongkuy Choi for their encouragement and friendship.

I thank my parents for their love, support, encouragement and prayer during this
endeavor. Most importantly, I thank my Lord, Jesus Christ for His grace for every

moments that I have met since I came here.

iv

TABLE OF CONTENTS

PAGE
LIST OF TABLES ............................................................................. xii
LIST OF FIGURES .......................................................................... xiv
LIST OF ABBREVIATION ................................................................ xxi
CHAPTE I: INTRODUCTION ............................................................. 1
Membrane Dynamics ....................................................................... 2
Membrane Adaptation Mechanism of the Bacteria to Environmental Stress ......... 5
Physiology and Biochemistry of Anaerobic Eubacteria Sarcina ventriculi .......... 8
Membrane Lipid Structures in Eubacterial Thermophiles ............................ 10
Computational Studies on the Membrane Lipids ...................................... 11
Overview .................................. 12
References .................................................................................. 14

CHAPTER II: STUDY ON THE NEW FAMILY OF THE VERY LONG
CHAIN FATTY ACIDS IN THE MEMBRANE OF THE
SARCINA VENTRICULI IN RESPONSE TO DIFFERENT

FORMS OF ENVIRONMENTAL STRESS ......................... 18
INTRODUCTION ..................................................................... 19
MATERIALS AND METHODS ................................. . .................. 20

Organism and Culture Conditions ............................................. 20

Membrane Preparations ......................................................... 20

Total Fatty Acids Analysis ...................................................... 21

Extraction of Lipids .............................................................. 21

Analysis of Lipids ............................................................... 22

Isolation of a,tn-Dicarboxylic Acid Dimethyl Ester ......................... 23

Isotope Labeling ................................................................. 24

1H NMR and 13C NMR Spectroscopy ........................................ 24

TABLE OF CONTENTS (cont’d)

PAGE
Fourier Transform Infrared Spectroscopy .................................... 24
RESULTS AND DISCUSSIONS ................................................ 25
Two Dimensional TLC Analyses on the Lipids .............................. 25
Compositional Analyses of Isolated Lipids ................................... 25
GC/MS Analyses of Dicarboxylic Dimethyl Esters .......................... 30
Spectrosc0pic Analyses of Isolated Dicarboxylic Dimethyl Esters ......... 35
Induction of Dicarboxylic Acids in Response to Various Forms of
Environmental Stress ............................................................ 42
Effect of Alcohol and Thermal Stress on Lipid Composition ............... 43
CONCLUSIONS ...................................................................... 51
REFERENCES ........................................................................ 52

CHAPTER III: CHEMICAL PROOF FOR THE FORMATION OF VERY
LONG a,m-BIFUNCTIONAL ALKYL SPECIES IN THE

MEMBRANE OF SARCINA VENTRICULI BY TAIL-TO-
TAIL COUPLING OF EXISTING ALKYL CHAINS FROM

OPPOSITE SIDES OF THE BILAYER ............................... 55
INTRODUCTION .................................................................... 56
MATERIALS AND METHODS ................................................... 63

Organism and Culture Conditions ............................................. 63

Membrane Preparations ......................................................... 63

Total Fatty Acids Analysis . ...................................................... 63

TABLE OF CONTENTS (cont’d)

PAGE
Isolation of the onto-Dicarboxylic Acid Dimethyl Esters .................... 64
Isotope Labeling ................................................................. 65
1H NMR and 13C NMR Spectroscopy ........................................ 65
Fourier Transform Infrared Spectrosc0py .................................... 65
Optical Rotation Analysis (Polarimetry) ....................................... 65
Reductive Ozonolysis ........................................................... 66
RESULTS AND DISCUSSION ...................... g ........................... 67
GC Analyses of Fatty Acids of S. ventriculi ................................ 67
Mass Spectrometric Analysis of Very Long Bifunctional Fatty Acids ..... 67
NMR and FI'IR Analyses of Isolated Dicarboxylic Dimethyl Esters ...... 77
Reductive Ozonolysis and Mass Spectrometry ............................... 83
Molecular Mechanism for the Formation of Very Long Chain
Dicarboxylic Acids ............................................................... 92
The General Significance of the Coupling Mechanism ... .................. 99
CONCLUSIONS ..................................................................... 100
REFERENCES ....................................................................... 102

CHAPTER IV: A MATHEMATICAL MODEL EXPLAINING THE
MOLECULAR WEIGHTS AND DISTRIBUTION OF VERY
LONG CHAIN DICARBOXYLIC ACIDS FORMED
DURING THE ADAPTIVE RESPONSE OF SARCINA
VENTRICULI .............................................................. 103
INTRODUCTION ................................................................... 104

TABLE OF CONTENTS (cont’d)

PAGE
MODELS AND PROBABILITY CALCULATIONS ...................... 106
Model Construction ............................................................ 106
Definition of Parameters for Model .......................................... 106
Calculation of P(Fij) ........................................................... 106
Calculation of P(F1:Fj) [%] .................................................... 109
RESULTS ............................................................................. 1 11
Random Coupling Model (RCM) ............................................ 1 1 1
Comparison of ‘Predicted Data’ with ‘Observed Data’ for
Transmembrane Fatty Acids .................................................. 1 12
DISCUSSION ........................................................................ 117
CONCLUSIONS ..................................................................... 118
REFERENCES ................................. L ..................................... H9

CHAPTER V: A NEW FAMILY OF MONOGLUCOSYLDIACYL
GLYCERIDE DIACYL GLYCEROL LIPIDS CONTAINING
VERY LONG CHAIN BIFUNCTIONAL ACYL CHAINS IN

SARCINA VENTRICULI ............................................... 120
INTRODUCTION ................................................................... 121
MATERIALS AND METHODS ................................................. 122

Culture of Cells ................................................................. 122

Membrane Preparation ......................................................... 122

Extraction of Lipids ........ v. ................................................... 122

Analysis of Lipids .............................................................. 123

Head Group Analyses ......................................................... 123

TABLE OF CONTENTS (cont’d)

PAGE
Fatty Acids Analysis ........................................................... 124
FAB (Fast Atom Bombardment) Mass Spectrometry ...................... 125
1H NMR Spectroscopy ........................................................ 125
Fourier Transform Infrared Spectroscopy ................................... 126
RESULTS AND DISCUSSION ................................................ 127
Structural Characterization of Isolated Glycolipids ......................... 127
NMR Analyses of Glycolipids ............................................... 132
Head Group Analyses of Glycolipids ....................................... 132
FTIR Analysis of Glucolipids ................................................ 136
FAB (Fast Atom Bombardment) Mass Spectrometry Analysis
of the Glucolipids .............................................................. 144
Biochemical Significance of Glucolipids Containing Bifunctional
Fatty Acyl Chains .............................................................. 152
CONCLUSIONS .................................................................... 153
REFERENCES ....................................................................... 154

CHAPTER VI: A FAMILY OF VERY LONG CHAIN onto-DICARBOXYLIC
ACIDS IS A STRUCTURAL COMPONENT OF THE
MEMBRANE LIPIDS OF CLOSTRIDIUM

THERMOHYDROSULFURICUM .................................... 155
INTRODUCTION ................................................................... 156
METERIALS AND METHODS ................................................. 158

TABLE OF CONTENTS (cont’d)

PAGE
Bacterial Cultures and Membrane Isolation ................................. 15 8
Total Fatty Acid Analysis ...................................................... 158
Isotope Labeling ................................................................ 159
Isolation of (1,0)-13,16«Dimethyloctacosanedioate Dimethyl Ester ....... 159
DQF—COSY and DEPT Experiments ......................................... 159
Fourier Transform Infrared Spectroscopy ................................... 160
RESULTS AND DISCUSSION ................................................. 161
Total Fatty Acids Analyses of Cl. thermohydrosulfuricum ............... 161
Mass Spectrometric Analyses of C30-Dicarboxy1ic Dimethyl Ester ...... 161
NMR / FTIR Analyses of C30 - Dicarboxylic Dimethyl Ester ............ 168
Mass Spectrometric Analyses of C29, C31 and C32 Dicarboxylic
Dimethyl Esters ................................................................ 176
The General Significance of These onto-Dicarboxylic
Acyl Components .............................................................. 187
CONCLUSIONS ..................................................................... 192
REFERENCES ....................................................................... 193

CHAPTER VII: COMPUTATIONAL STUDIES ON THE EFFECT OF THE
TRANSMEMBRANE ALKYL CHAINS ON THE

STRUCTURE AND DYNAMICS OF MEMBRANE ............ 195
INTRODUCTION ................................................................... 196
MODELS AND METHODS OF SIMULATION ............................ 200

Models for Simulations ....................................................... 200

TABLE OF CONTENTS (cont’d)

PAGE

Force Fields for Energy Calculation ........................................ 200

Methods for Simulations ...................................................... 201
RESULTS AND DISCUSSION ................................................. 202

RMS Distance Fluctuation for Two Hydrocarbon Chain

Model during 10 ps MD Simulations ........................................ 202

RMS Distance and Angle (CD) Fluctuations for Two Lipid Model

during 10 ps MD Simulations ................................................. 202

Effect of the Transmembrane Alkyl Chains on the Structure and

Dynamics of Membrane ....................................................... 215
CONCLUSIONS .................................................................... 228
REFERENCES ....................................................................... 229

LIST OF TABLES

PAGE
CHAPTER II
Table 1. Analysis of 70 ev electron impact mass spectral fragments
of 0t,(u-15,16-dimethyl trocotanedioate dimethyl ester ............................... 34
Table 2. Effect of environmental stress on chain length and the degree
of saturation of the fatty acids in Sarcina ventriculi .................................. 48
CHAPTER 111
Table 1. Analysis of 70 ev electron impact mass spectral fragments
of proposed structure 2_ .................................................................. 71
Table 2. Optical rotation analyses of peak B, D and E ........................................ 93
CHAPTER IV
Table 1. Mathematical formular for the prediction of probability for the
formation of transmembrane fatty acids .............................................. 114

LIST OF TABLES (cont’d)

PAGE
CHAPTER V
Table 1. The calculated m/z values of ionic clusters based on the
compositional analyses ................................................................. 149
CHAPTER VI
Table 1. Analysis of electron impact mass spectral fragments of peak B ................. 169

LIST OF FIGURES

PAGE
CHAPTER I
Figure l. The characteristic frequencies of molecular motions of membrane
proteins nad lipids compared with the frequency ranges .............................. 3
Figure 2. Thermal regulation of fatty acid biosynthesis ........................................ 6
Figure 3. The proposed biochemical pathway for glucose metabolism
in S. ventriculi ............................................................................. 9
CHAPTER II
Figure 1. Two dimensional TLC analyses of the lipids extracted from
the cell of Sarcina ventriculi grown at pH 7 (A) versus pH 3 (B) ................... 27
Figure 2. Gas Chromatographic analyses of lipid components in the
membrane of S. ventricli grown at pH 7 (A) versus pH 3 (B) ...................... 29
Figure 3. Electron impact mass spectrum (70 ev) of cam-15,16-
dimethyltriacotanedioate dimethyl ester without (A)
and with isotope labeling (B) ............................................................ 32
Figure 4. Electron impact mass spectral fragmentation pattern of
onto-15,16-dimethyltriacotanedioate dimethyl ester .................................. 33
Figure 5. The 300 MHz 1H NMR spectrum of onto-15,16-
dimethyltriacotanedioate dimethyl ester ................................................ 37

Figure 6. The 125 Mhz 13C NMR spectrum of

xiv

LIST OF FIGURES (cont’d)

PAGE

a,co- 15, 16-dimethyltriacotanedioate dimethyl ester .................................. 39
Figure 7. The Fourier transform infrared spectrum of

0t,o)-15,16-dimethyltriacotanedioate dimethyl ester .................................. 41
Figure 8. Gas chromatographic analyses of lipid components

extracted from cells of S. ventriculi grown at pH 7.0

in the presence of various forms of environmental stress ............................ 46

CHAPTER III

Figure 1. Gas chromatographic analyses of the total methanolysate

of the lipids in the membrane of Sarcina ventriculi cell grwn at pH 7.0

at 37°C and then shifted to 45°C at late log phase for 3hrs ......................... 59
Figure 2. Rationalization of stereochemistry of the products formed by

tail-to—tail coupling of alkyl chains from opposite sides of the bilayer .............. 61
Figure 3. Electron impact mass spectrums of peak D ......................................... 69
Figure 4. Electron impact mass spectral fragmentation pattern

of peak D (assigned structure 2) ........................................................ 73
Figure 5. The electron impact mass spectrum of peak E(assigned structure 2) ........... 75
Figure 6. Electron impact mass spectral fragmentation pattern

of peak E (assigned structure 3,) ........................................................ 76
Figure 7. The 1H NMR spectrum of proposed structure 2 (peak D) ....................... 78
Figure 8. The 1H NMR spectrum of proposed structure 3 .................................. 79
Figure 9. The 13C NMR spectrum of proposed structure 2, ................................. 80

XV

LIST OF FIGURES (cont’d)

PAGE

Figure 10. The 13C NMR spectrum of proposed structure 3 ................................ 81
Figure 11. The Fourier Transform Infrared spectrum of proposed structure 2 ............ 82
Figure 12. Fragments expected for the reductive ozonolysis of

structure 2. (A) and 3, (B) ................................................................ 85
Figure 13. The electron impact mass spectrum of 4 .......................................... 87
Figure 14. The electron impact mass spectrum of 5_ .......................................... 90
Figure 15. The electron impact mass spectrum of 6 .......................................... 91
Figure 16. Mass spectral fragments of C321) 0t,0)-dicarboxylic very long

chain acyl species (9) in the case of propanol (0.15M) induction ................... 95
Figure 17. The electron impact mass spectrum of 1 .......................................... 98

CHAPTER IV

Figure 1. Gas chromatographic profile of methyl ester derivatives

of total fatty acids of Sarcina ventriculi at pH 3.0 and 37°C ....................... 105
Figure 2. Random coupling model for the synthesis

of transmembrane fatty acids .......................................................... 108
Figure 3. Prediction of the distribution of transmembrane fatty acids

based on random coupling model ..................................................... 116

CHAPTER V

Figure 1. Total ion chromatogram of Gas Chromatography/Mass

xvi

LIST OF FIGURES (cont’d)

PAGE

Spectrometry analysis for the esten'fied fatty acyl components

of one of the isolated glycolipids ...................................................... 129
Figure 2. Electron impact mass spectrum of peak k ......................................... 130
Figure 3. Mass spectral fragmentation pattern of peak k

(to-formyl-( 17,18-dimethyl)-cis-1 lhentriacotanemethyl ester) .................... l3 1
Figure 4. 1H NMR spectrum of the glycolipids containing

bifunctional acyl chains ................................................................. 134
Figure 5. Gas chromatographic profile of the alditol

acetates of hydrolysates obtained from 2 M TFA

(Trifluoroacetic acid) hydrolysis of the lipids ........................................ 135
Figure 6. Mass spectral analysis of peak B ................................................... 138
Figure 7. EI mass spectrum of peak A ........................................................ 139
Figure 8. 1H NMR spectrum of the hydrolysates

obtained from the TFA hydrolysis on glycolipids ................................... 141
Figure 9. Fourier Transform Infrared spectrum of the glucolipids ........................ 143
Figure 10. Positive FAB-Mass Spectrum of the glucolipids ............................... 145
Figure 11. Positive FAB-Mass spectrum of two ion .

clusters ([M+Na]+, [M+Na - l62]"') ................................................. 147
Figure 12. Basic model structure of the monoglucosyldiacylglyceride

diacyl glycerol in S. ventriculi ........................................................ 151

xvii

LIST OF FIGURES (cont’d)

PAGE
CHAPTER VI

Figure 1. Total ion chromatogram of Gas Chromatography/Mass

Spectrometry analysis for the esterified fatty acyl

components of the membrane of Cl. thermohydrosulfuricum ..................... 163
Figure 2. Electron impact mass spectrum (70 ev) of peak B

without (A) and with (B) isotope labeling ............................................ 165
Figure 3. Analysis of mass spectral fragmentations of peak B ............................. 167
Figure 4. 1H and 13C NMR spectrum of peak B ............................................ 171
Figure 5. The DQF-COSY spectrum (in the region between 0 to 2.6 ppm)

of peak B, in CDC13 at 500 MHz ...................................................... 173
Figure 6. The correlation analysis of DQF-COSY spectrum of peak B ................... 174
Figure 7. DEPT spectrum of peak B in CDC13 after making all peaks positive .......... 175
Figure 8. Fourier Transform Infrared spectrum of Peak B ................................. 177
Figme 9. Electron impact mass spectrum (70ev) of peak A

without (A) and with isotope labeling (B) ............................................ 179
Figure 10. Mass fragmentation pattern of peak A .......................... . ................ 181
Figure 11. Electron impact mass spectrum (70 ev) of peak C

without (A) and with isotope labeling (B) ............................................ 183
Figure 12. Mass fragmentation pattern of peak C ........................................... 184
Figure 13. Electron impact mass spectrum (70 ev) of peak D

without (A) and with isotope labeling (B) ............................................ 186
Figure 14. Mass fragmentation pattern of peak D ........................................... 188

Figure 15. The determined structures of a family of very

xviii

LIST OF FIGURES (cont’d)

PAGE
long chain on, (u-dicarboxylic dimethyl esters ........................................ 191
CHAPTER VII

Figure 1. Electron micrograph of S. ventriculi cells grown

at pH 7.0 (A) and at pH 3.0 (B) after freeze fracturing and coating ............... 198
Figure 2. Computer generated pictures of the trajectory

snapshots of two octadecane molecules for 10 ps

molecular dynamic simulations ........................................................ 204
Figure 3. RMS (Root Mean Square) distance fluctuation

based on the inital minimized structure as a reference coordinate .................. 205
Figure 4. Computer generated pictures of the snapshots of the

trajectory files obtained every 0.4 ps during the 10 ps MD simulations .......... 207
Figure 5. Simplified pictures of the motional spectrum of the lipids ...................... 208
Figure 6. Angle fluctuation between the bilayer form and

monolayer form during the MD simulations ......................................... 210
Figure 7. RMS (Root Mean Square) distance fluctuation

of each carbon atom in a sn-l chain (C18) of the typical bilayer

lipid as a function of MD simulation time ............................................ 211
Figure 8. RMS (Root Mean Square) distance fluctuation

of each carbon atom in a sn-2 chain (Cl6)-of the typical bilayer

lipid as a function of MD simulation time ............................................ 212

Figure 9. RMS distance fluctuation of the half of the

LIST OF FIGURES (cont’d)

PAGE

transmembrane monolayer (C36) acyl chain (sn-l). ................................. 213
Figure 10. RMS distance fluctuation of the other free chain

(Sn-2, C16) in the monolayer lipid form .............................................. 214
Figure 11. Fluctuation amplitude of the atomic distance of the lipids

during the MD simulations at 1 ps (A), 5 ps (B) and 10 ps (C) ................... 217
Figure 12. RMS distance fluctuation of the lipid based on

the initial lipid coordinates ( Ops) as the reference ................................... 218

Figure 13. Computer generated pictures of the reminirnized

structures after the annealed MD simulations (5 annealing cycles)

before (A) and after (B) the transtion from the bilayer to monolayer form ....... 220
Figure 14. Comparison of two structures before (right)

and after (left) the transition to monolayer in different models

( in ball stick model for (A), vector model for (B)) ................................. 222
Figure 15. Inter- and intradigitation of the acyl chains

shown in the transmembrane structures obtained

after the annealed dynamic simulations ............................................... 225
Figure 16. Electronic distribution of the lipids before and

after the transition to the monolayer form ............................................ 226
Figure 17. Progressive intra- and interdigitation of the acyl chains

in the transmembrane during simulated annealing calculations .................... 227

XX

ACP
AMU

BD

CL

DEPT
DQF-COSY
EI-MS

CV

FA
FAB-MS

GC
GC/MS

LIST OF ABBREVIATIONS

Acyl Carrier Protein

Atomic Mass Unit

Brownian Dynamics

Cardiolipin

Distortionless Enhanced by Polarization Transfer
Double Quantum Filtered-Correlation Spectroscopy
Electron Impact-Mass Spectrometry

electron volt

Fatty Acids

Fast Atom Bombardment-Mass Spectrometry
Fourier Transform Infrared

Gas Chromatography

Gas Chromatography/Mass Spectrometry

Monte Carlo

Molecular Dynamics

N itro Benzyl Alcohol

Neutral Lipids

Nuclear Magnetic Resonance
Phosphatidyl choline

Phosphatidyl ethanolamine

Phosphatidyl serine

RCM
RMS
TFA
TLC

LIST OF ABBREVIATIONS (cont’d)

Random Coupling Model
Root Mean Square
Trifluoroacetic acid

Thin Layer Chromatography

CHAPTER I

INTRODUCTION

Membrane Dynamics

All biological structures are dynamic, and the extent of the rate and motion are
important in considering biological function. This is true most certainly for membranes.
The fluid mosaic model (1, 2) has helped focus attention on the mobility of membrane
components by conceptualizing the membrane as a sea of lipid in which embedded proteins
are freely floating. The basic motivation of membrane dynamics is its relevance to
biological functions. Figure 1 shows the characteristic frequencies of molecular motions of
membrane proteins and lipids. Very wide ranges of motions are observed; from molecular
vibrations occurring in about 10'14 sec (0.01 ps) to transbilayer flip-flop of lipids (3, 4)
which can take many days to occur. The small frequencies of molecular vibrations
indicate that there are always vibrational motions of the lipid chains. Furthermore, this
vibrational motion would be very sensitive to subtle changes of external energy. Lipid
rotation (5) and trans/gauche isomerization (6) also show the small frequency ranges (10 ps
- 1 ns). Therefore, the relative impacts by environmental challenges can be explained by
this wide range of motional frequency. For real time motional dynamics, 10 ps can cover
the range from vibration to lipid rotation, including trans/gauche isomerization. There are
some other motions for membrane components. Isotropic rotation (7) implies equivalent
rotation in all directions with no favored rotational axis. This would be. expected for a
sphere rotating in a continuum and has been applied to the interpretation of the rotational
properties of small hydrophobic probes dissolved in a bilayer. Conical constraint (8)
describes the motion of amphipathic probes such as fatty acid derivatives. These can be
considered in simple models as rigid rods tethered at the membrane surfaces.

All the motions of membrane components work under three general forces. These
three long range forces (9) are the electrostatic forces, polarization (or induction) forces and
London-van der Waals dispersion forces. Electrostatic forces are due to the mutual

Coulomb attraction or repulsion of the net charges, or electric moments, carried by two

 

Lipid "flip—flop"

Protein rotation

 

 

Laternai "hopping" (lipids and proteins)

-— Lipid rotation (about long axis)

 

Trans/Gauche isomerization
---Cl~tz vibration Frequency (sec‘)

1 l l 1 l r l l
l l l l l

1015 10'0 105 102 10‘2 10'5

--

-—
fl
-

-

Figure l. The characteristic frequencies of molecules of molecular motions
of membrane proteins and lipids compared with the frequency ranges. The
characteristic times are obtained by taking the reciprocal of the indicated frequencies.
Boundariesareveryapproximate.

4
interacting molecules. Polarization forces arise from the charge separation on one
molecule by the charges or permanent electric moments of the other. Typical examples are
interactions between polar groups on one molecule and polarized groups on the other.
London-van der Waals forces arise between all molecules, even between neutral non-polar
molecules, and are due to the average interaction of an instantaneous electric moment on
one molecule, brought about by charge density fluctuations, and the moment it induces on
the other molecule. These forces have a purely quantum-mechanical nature. This
dispersion force is of paramount importance in holding the lipid chains together through
additive interactions.

For an ordinary liquid, the term “fluidity” is defined as the inverse of viscosity.
This is a well-defined and easily measured physical characteristic. Viscosity is essentially a
measure of the frictional resistance encountered when adjacent “layers” of fluid are moving
with different velocities. Viscosity can be measured by simply observing the velocity with
which a marble falls through the liquid. As applied to the membranes, the term “fluidity” is
usually thought of in a more qualitative sense. It is generally meant to represent a measure
of the resistance to various types of movements in the membrane. Generally, fluidity is
measured by observing the motion of spin probes or fluorescent probes incorporated in the
bilayer. Since the measurements are sensitive to both the rate of motion and any constraints
to that motion, information about the dynamics and molecular order gets intermixed (10).
Biological membranes are generally in the liquid crystalline phase, and it appears that the
maintenance of membrane fluidity is critical to their function. A decrease in the fluidity of
a biological membrane can cause a phase transition to a gel phase. The gel phase has
different dynamic and functional properties (11). It does not allow biological function.
The most dramatic evidence is from studies showing adaptations of various organisms to

environmental stress.

Membrane Adaptation Mechanism of the Bacteria to Environmental Stress

The biological membrane is one of the most vulnerable components of bacteria to
environmental stress. Many studies have been performed on the thermal and alcoholic
stress in relation to the membrane. As for thermal stress, most bacteria develop an
adaptive response system which involves fatty acid synthesis. E. coli, along with most (if
not all) other microorganisms, synthesizes lipids with a greater proportion of unsaturated
fatty acids when grown at low temperature (e. g. 25°C) rather than at high temperatures
(e.g. 42°C) (12). This regulatory system is believed to be designed to ameliorate the
effects of the temperature change on the physical state of all membrane lipids. In fact,
lowering the temperature deprives the system’s (membrane) energy. This low-energized
system requires the disordering components (unsaturated fatty acids) to maintain! the
motional functionality of the membrane. Membrane functionality depends on the motional
dynamics of the membrane components, particularly lipid components. In terms of that,
the proportion of disordered (unsaturated) lipids to ordered (saturated) lipids in cell
membranes plays a major role in membrane function. Increased incorporation of
unsaturated fatty acids into the lipids decreases the temperature at which transition from
ordered to disordered membrane lipids occurs, whereas increased incorporation of
saturated fatty acids has the opposite effect (13). The thermal regulatory system can thus
adapt the membrane lipids for optimal functioning at the new growth temperature.

Figure 2 shows the thermal regulation of fatty acid biosynthesis. According to the
in vitro studies on purified enzymes (14, 15), the decreased growth temperature alters the
activity of 3-Ketoacy1-ACP synthase II, whichin tum regulates the fatty acid composition
by producing more cis-vaccenoyl—ACP for incorporation into the lipids (16). Interestingly,
the thermal regulation of the membrane depends on the activity of this one enzyme and

occurs independently of new gene synthesis.

3' OH - DeconOyl - ACP

\

(Is - 3- DecenOyl- ACP trans - 2 - Decenoyl— ACP
{ +3Molonyl-ACP l + 3Molonyl-ACP
i i

 

 

 

 

 

 

 

 

Polmitoleoyl -ACP / Polmitoyl-ACP

c/s - Voccenoyl - ACP

 

 

 

 

/sn - Glycerol 3 - Phosphate

 

V
Phospho t idic ACIO

Phospholipid Synthesis

Phospholipid Biloyer.

Figure 2. Thermal regulation of fatty acid biosynthesis. 3-Ketoacyl-ACP
synthase II is primarily responsible for the temperature control of E. coli fatty acid
composition by being more active in the conversion of palmitoleate to cis-vaeoenate at

lower temperatures than at higher temperatures

As another regulation to thermal stress, fatty acid chain length can be regulated.
The 3-ketoacyl-ACP synthase I and II are the likely candidates for the site of chain length
regulation in that both enzymes catalyze the elongation reaction. Substrate specificity
studies in vitro indicates that one reason why membrane lipids are devoid of chains of more
than 18 carbons is in part the reduced activity of synthases I and II on C13 substrates
( 59,60). Furthermore, synthase II mutants are defective in the elongation of palrrritoleate
to cis-vaccenate (61). It indicates, therefore, that synthase 11 plays an important role in
determining the amount of C13 fatty acids in the membrane. Although these data indicated
that condensing enzymes play a significant role in determining the chain length,
physiological experiments indicate that the level of G3P (glycerol-3-phosphate)
acyltransferase activity is also important. When phospholipid biosynthesis is slowed or
arrested at the acyltransferase step, the fatty acids synthesized have abnormally long chains
compared with the normal distribution of the fatty acids synthesized in the presence of
G3P (glycerol-3-phosphate) acyltransferase activity (17). These data indicate that
competition between the rate of elongation and the rate of utilization of the acyl-ACPS by
the acyltransferase is a significant determinant of fatty acid chain length in E.coli.

Many studies (18, 19) have been performed focusing on ethanol effect on the
membrane. Alcohol is an amphipathic molecule that affects microbial processes at high
molar concentrations. These changes appear to be caused by colligative effects on the
aqueous medium and within the membrane rather than being mediated by specific
receptors. Colligative properties altered by alcohol include changes in dielectric properties,
replacement of water by alcohol as a hydrogen bonding partner, and a weakening of the
strength of hydrophobic bonds (20, 21). Alcohols decrease the effectiveness of the
hydrophobic core of the membrane as a barrier, increasing membrane permeability and

membrane leakage. Cells compensate for this effect in part by increasing the average acyl

8
chain length of membrane lipids, increasing the thickness of the hydrophobic core and
restoring this essential barrier function (22, 23).

A membrane adaptive response to thermal or alcoholic stress in E. coli is to change
the lipid composition by regulating the ratio of unsaturated to saturated fatty acids or by the
changing chain length by two or four carbons. Besides temperature or alcohol, there are
other forms of environmental stress such as pH, nutrient concentrations, osmotic pressure
and pressure. However, the relevance of these factors to the membrane architecture is not

known in detail.

Physiology and Biochemistry of Anaerobic Eubacteria Sarcina ventriculi

As a model system for membrane adaptive response to various forms of
environmental stress, anaerobic facultative acidophilic eubacteria, Sarcina ventriculi was
studied. Sarcina ventriculi was first observed in 1842 by Goodsir (24) in the contents of
human stomach. The organism has been cultivated from the garden soil (25, 26) and the
stomach contents (27), and it has also been enriched and isolated from sand (28), river mud
(29), and peat bog sediments (30). The prevalence of this organism in sedimentary
environments and acid or alkaline soils that have been stored for months to years (31)
suggests the presence of resistant structures or spores.

Sarcina ventriculi is an obligate anaerobe capable of growth from pH 8.0 to 2.0.
When grown at pH 3.0, the internal pH was 4.3, and this increased to 7.1 with an
environmental pH 7.0 (32). Distinct morphological changes in the ultrastructure of S.
ventriculi were observed when cells were grown in a medium of constant composition at
pH extremes of 3.0 and 8.0. (33). Figure 3 shows the proposed biochemical pathway for
glucose metabolism in S. ventriculi (34). The organism changed carbon and electron flow
from acetate, forrnate and ethanol production. at neutral pH, to predominantly ethanol
production at pH 3.0. Increased level of pyruvate dehydrogenase (relative to pyruvate
decarboxylase) and acetaldehyde dehydrogenase occurred when the cells were grown at

9

Influence at pH on Glucose Fermentation ot
Sarcina ventricu/i
GLUCOSE
ATP

ADP

GLYCERALDEHYDE-B-PHOSPHATE
ZADP 2 NAD

ZATP amen,
PEP

ADP
ATP

PYRUVATE

H!
as a... 23$»-<m.

 

pyruvate
«carbon 1
’0“ CoASH dehydrogenase
NAD NAOH;
ACETALDEHYDE oeet II I I ACE'WLPEoA
X
CoA WW
- NAG“:
olcohot ACEm mosmre
MAO acetate kinose

ATP

“30 fl KTABOLSQA NEUTRAL pH METABCIJSM.

Figure 3. The proposed biochemical pathway for glucose metabolism in S.

ventriculi.

10

neutral pH, indicating the predominance of carbon flux through the oxidative branch of the
pathway for pyruvate metabolism. When the organism was grown at acidic pH, there was
a significant increase in pyruvate decarboxylase levels and a decrease in acetaldehyde
dehydrogenase, causing flux through the non-oxidative branch of the pathway.

Of interest is that S. ventriculi predominantly produced ethanol when the organism
was grown at pH 3.0. External perturbation such as excess proton concentration (acidic
condition) was involved in increasing alcohol concentration inside the cell by changing
some metabolic pathway. S. ventriculi shows a wide dynamic life cycles since it survives

from pH 2.0 to pH 8.0.

Membrane Lipid Structures in Eubacterial Thermophiles

As another model for membrane adaptation study, anaerobic thermophilic
eubacteria, Clostridium themohydrosulfuricum was investigated. The optimum
temperature for growth of the organism is 67-69°C; the maximum temperature at which
growth occurs is 76-78°C; the minimum temperature is 42°C (35). Clostridium
thermohydrosulfuricum was isolated from Octopus Spring at Yellowstone National Park
(36). It ferrnents starch and a wide variety of hexose- or pentose-derived saccharides
including xylose and glucose into ethanol as the major reduced end product (37). In terms
of the fact that it produced ethanol and experienced constantly high thermal stress,
membrane adaptation mechanism was studied.

The apolar hydrocarbon chains, which are capable of hydrophobic interaction and
provide the core of the membrane, have been the most extensively studied components of
thermophile lipids. The majority of the hydrocarbon chains, usually pairs, are bound to
glycerol molecules through one of three classes of chemical bonds : acyl esters (fatty
acids), vinyl ethers (fatty aldehyde), or alkyl ether (alcohols). The derived glycerolipids in
turn are known to provide the typical bilayer assembly (38). Thermophiles tend to possess

1 1
longer hydrocarbon chains and methyl (iso— and anteiso-) branched chains (39, 40).
Unsaturated hydrocarbon chains, which significantly lower melting points, are rare (41).
In comparison to the number of reports dealing with the nature and function of the apolar
chains of thermophilic bacteria, little is really known about the structure or the function of
the complex lipids. The question on inherent stability of the membrane in therrnophiles
remains an enigma because methyl-branching as a characteristic of the apolar chains acts
as a “fluidizer” rather than “stabilizer” (42, 43, 44). A complete understanding of the
molecular basis for thermostable membranes requires structural studies on the new and

unique lipids.

Computational Studies on the Membrane Lipids

Model monolayers and bilayers of the chain molecules have been a popular subject
of computer simulations over the years. These simulations have been performed either by
Monte Carlo (MC)(45), molecular dynamics (MD)(46,47), or Brownian dynamic
(BD)(48,49) simulations, at various levels of detail, on ensembles of chain molecules of
varying lengths and packing densities. This has resulted in a growing understanding of the
structures of real, complex systems such as the biological membrane (50, 51). One of the
first detailed MD simulations of a smectic liquid crystal was reported by van der Ploeg and
Beredsen (52, 53) on a system consisting of multilayers of the three component decanoate-
decanol~water system. Recently Schurmann et al., (46) performed MD simulation on a
model bilayer of 48 chain molecules with fixed head groups. The effects of packing
density and temperature on the extent of spatial and temporal correlation in the ensemble
have been studied by analyzing the trajectories from the MD simulation. The structure and
dynamics of water between the bilayers was also studied by MD simulation (54). This
research showed that the thermal motion of thepolar head groups has no influence on the
orientational polarization of water but had a large influence on the dynamics of the water

interface. Large scale motional simulations in membrane lipids have been performed with

L, H
'h -'1

p013:
h r
”In:

hwy

OVEF

12

the Monte Carlo method (55) as models of lipid-protein interactions in bilayers (56) or for
the lipid-cholesterol interaction in a model membrane (47). In general the results of both
the molecular dynamics (MD) and the Monte Carlo (MC) simulations of systems consist of
chain order, RMS (Root Mean Square) distance parameters and “snapshots” of
instantaneous configurations of the systems (the MD simulation also give some dynamical
information over very small time scales). The general agreement between the MC and MD
calculations and experimental data indicates that the computer calculations have included
most of the relevant interactions for the study of chain packing issues.

The MD simulation of a lipid micelle was also performed by Salemme et a1. (57).
They used 85 LPE (IySOphosphatidylethanolamine) and 1591 water molecules in a 100 ps
real-time simulation. Throughout the comparison of the initial and equilibrated micelles,
substantial differences were proposed both in LPE hydrocarbon chain conformation and
polar head-group-solvent interactions. Their work could predict the polymorphic
pretransition from a spherical toward a cylindrical micelle structure with a molecular
viewpoint.

Computer simulations were also performed for a large class of phenomenon of
lipids such as fluctuating membranes with a simplified lipid model. Leibler et a1.
performed MD simulations with simplified membrane-like objects (58). They chose
anisotropic, multibody forces so as to mimic real interactions between amphiphilic
molecules. Their research showed the possibility of the successful MD simulations of a

very large system such as a phase transition or membrane fusion.

OVERVIEW

The biological membrane is the most vulnerable cellular component since it is the
structure most directly exposed to the challenges of environmental stress. External stress
always accompanies a change in the “external input energy to the organism”. Another

viewpoint is that, molecular adaptation is the converting process of “external input energy“

13

to “cellular output energy” such as the regulation of lipid biosynthesis. This energy flow
around the membrane causes changes in the motional dynamics of membrane components.
The bacterial membrane should maintain an optimal dynamic state of the lipids for its
proper function. Lipid components respond to external energy perturbations by changing
the motional dynamics of the membrane. Therefore, the structural studies on the lipid
components induced during the stress response should be performed from the standpoint of
motional dynamics.

The basic approach throughout this thesis work was to elucidate membrane adaptive
mechanisms of extremOphiles such as S. ventriculi or Cl. thermohydrosulfuricum from the
standpoint of membrane dynamics. New adaptive mechanisms of membranes to various
forms of environmental stress will be discussed in detail. In Chapter II, the general
membrane adaptive mechanism in S. ventriculi will be discussed. The detailed chemical
and mathematical proofs for this mechanism will be discussed in Chapters III and IV.
Studies on the head groups of transmembrane lipids of S. ventriculi cells will be discussed
in Chapter V. Studies on the generality of this phenomenon and descriptions of unusual
structures of the membrane components of Cl. thermohydrosulfuricum cells will be
presented in Chapter VI. Finally, the possible structural and physiological roles of
transmembrane chains will be examined by using the computational approaches such as

molecular mechanics and molecular dynamics (Chapter VII).

14

REFERENCES
1. Singer, SJ, and Nicolson, G. L. (1972) Science 175, 720-731
2. Singer, S. J. (1974) Ann. Rev. Bloch. 43, 805-833
3. Op den Kamp, J.A.F. (1981) In The Asymmetric Architecture of Membranes (J .B.
Finean and Michell, R.H. Eds.) pp83-126, Elsevier, New York
4. Op den Kamp, J.A.F. (1979) Ann. Rev. Bloch. 48, 47-71
5. Ameloot, M.H., Hendrickx, H., Herreman, W., Pottel, H.,Cauwelaert, V. F., and
Van Der Meer, W. (1984) Biophy. J. 46, 525-539
6. Govil, G., and Hosur, R. V. (1982) In Conformation of Biological Molecules: New
Results for NMR. Springer-Verlag, New York
7. Chapman, D., and Benga, G. (1984) In Biomembrane F luidity-Studies of Model and
Natural Biomembranes. Biological Membranes (D. Chapman, Ed.) Vol. 5, pp1-56
Academic Press, New York
8. Kinosita, K., Kawato, S., and Ikegami, A. (1977) Biophys. J. 20, 289-305
9. Salem, L. (1962) Can. J. Biochem. and Physiol. 40, 1287-1298
10. Shinitzky, M. (1984) In Membrane fluidity and Cellular Functions (M. Shintzky, Ed.)
vol. 1, ppl-Sl CRC Press, Boca Raton
11. Lee, A.G. (1983) In Lipid Phase Transitions and Mixtures. Membrane F luidity in
Biology. (R.C. Aloia, Ed.) vol. 2 pp43-88 Academic Press, New York
12. Marr, A.G., and Ingraham, J.L. (1962) J. Bacteriol. 84, 1260-1267
13. Cronan, J.E., Jr., and Gelmann, ER (1975) Bacterial. Rev. 39, 232-256
14. Garwin, J.L., Klages, A.L., and Cronan, J.E., Jr. (1980) J. Biol. Chem. 255,
3263-3265
15. Garwin, J.L., Klages, A.L., and Cronan, J.E., Jr. (1980) J. Biol. Chem. 255,
1 1949-1 1956
16.de Mendoza, D., and Cronan, J.E., Jr. (1983) Trends Biochem. Sci. 8, 49-52

15

17. Cronan, J.E., Jr., Welsberg, L.J., and Allen RC. (1975) J. Biol. Chem. 250,5835-
5840

18. Ingram, L. O. (1990) Crit. Rev. Biatechnol. 9, 305-319

19. Ingram, L. 0., and N. S. Vreeland. (1980) J. Bacteriol. 144, 481-488

20. Yaacobi, M., and Ben-Naim, A. ( 1974) J. Phys. Chem. 78, 175-182

21. Ben-Naim, A. (1980) In Hydrophobic Interactions, plenum Press, New York

22. Buttke, T. M., and Ingram, L. O. (1978) Biochemistry 17, 637 — 644

23. Ingram, L. O. (1980) J. Bacteriol. 149, 166-172

24. Goodsir, J. (1842) Edinb. Med. Surg. J. 57, 430-443

25. Beijerinck, M.W. (1905) Amsterdam 13, 608-614

26. Beijerinck, M.W. (1906) Arch. Neerl. Sci. Exact. Natur. Ser. 2, 11, 608 - 614

27. Beijerinck, M.W. (1911) Amsterdam 19, 1412-1415

28. Srnit, J. (1933) J. Pathol. Bacterial. 36, 455-468

29.Canale-Parola, E., and Wolfe, RS. (1960) J. Bacterial. 79, 857-859

30.Goodwin, S., and J. G. Zeikus. (1987) Appl. Environ. Microbiol. 53, 57-64

31. Knoll, H. (1965) Monatsber. Dtsch. Akad. Wiss. Berl. 7, 475-477

32. Goodwin, S., and J. G. Zeikus. (1987) J. Bacteriol. 169, 2150-2157

33. Lowe, S.E., Pankratz, H.S., and Zeikus, J. G. (1989) J. Bacteriol. 171,
3775-3781 .

34. Lowe, S.E., and Zeikus, J. G. (1991) Arch. Microbial. 155, 325-329

35. Lovitt. R. W, Longin, R., and Zeikus, J. G. (1984) Appl. Environ. Microbiol.
48, 171-177

36. Zeikus, J .G., Ben-Bassat, A., and Hegge, P. (1980) J. Bacteriol. 143, 432-440

37. Ng, T.K., Ben-Bassat, A., and Zeikus, LG. (1981) Appl. Environ. Microbiol. 41,
1337-1343

38. J. Isrelachvili (1978) In Light Transducing Membranes (D.W. Deamer, Ed.) Academic
Press, New York

16
39. Goldfine, H. (1982) Curr. Top. Membr. Transp. 17, 1-9
40. Kaneda, T. (1991) Microbiol. Rev. 55, 288-302
41. Langworhty, TA. (1982) In Methods Enzymol. 88, 396-416
42. Silvius, J. R., and McElhaney, R. N. (1979) Chem. Phys. Lipids 24, 287-296
43. Silvius, J. R., and McElhaney, R. N. (1980) Chem. Phys. Lipids 26, 67-77
44. Kannenberg, E., Blume, A., McElhaney, R. N. and Poralla, K. (1983) Biachim.
Biophys. Acta 733, 111-1 16
45. Quinn, P.J. , and Chapman, D. (1980) Crit. Rev. Biochem. 8, 1-12
46. Biswas, A., and Schurmann, BL. (1991) J. Chem. Phys. 95, 5377-5386
47. Scott, H.L., and Kalaskar, S. (1989) Biochemistry 28, 3687-3691
48. Pastor, R, Venable, R.M., and Karplus, M. (1988) J .Chem.Phys. 89,
11 12-1 127
49. Pastor, R, Venable, R.M., Karplus, M., and Szabo, A. (1988) J. Chem.
Phys. 89, 1128-1 140
50. Egberts, B., van Gunstem, W. F., and H.J.C. (1988) J. Chem. Phys.
89, 3718-3732
51. Watanabe, K., Kelin M. L. (1989) J. Phys. Chem. 93, 6987-6901
52. van der Ploeg, P., and Berendsen, H.J.C. (1982) J. Chem. Phys. 76,
3271-3279
53. van der Ploeg, P., and Berendsen, H.J.C.( 1983) Mol. Phys. 49, 233-
241
54. Raghavan, K., Reddy, M.R., and Berkowitz, M.L. (1992) Langmuir, 8,
233-240
55.Metropolis, N ., Rosenbluth, A., Rosenbluth, M., Teller,A., and Teller, E.
(1953) J. Chem. Phys. 21, 1087-1096
56. Scott, H.L. (1986) Biochemistry, 25, 6122-6126
57. Wendoloski, J.J., Kimatian, S.J., Schutt,C.E., and Salemme, FR. (1989) Science,

17

243, 636-638
58. Droufff,J-M., Maggs, A., and Leibler, S. (1991) Science, 254, 1353-1356

CHAPTER II

STUDY ON THE NEW FAMILY OF THE VERY LONG CHAIN
DICARBOXYLIC FATTY ACIDS IN THE MEMBRANE OF THE SARCINA
VENTRICULI IN RESPONSE TO DIFFERENT FORMS OF
ENVIRONMENTAL STRESS

18

19

INTRODUCTION

Bacterial membranes are extremely dynamic, complex systems which interface
directly with the environment and carry out many important functions. They provide
compartmentalization, transport, and acts as a matrix for cytoplasmic macromolecules and
membrane proteins.

Bacteria in their natural ecosystem may experience many changes in environmental
factors, including temperature, pH, solvent concentration, nutrient levels and oxygen
concentration. In this respect, the membrane, which is directly exposed to the
environment, is one of the most critical and vulnerable components of the cell and must
adapt to and survive these changes.

Sarcina ventriculi is a strictly anaerobic bacterium which can grow on sugars over a
wide pH range, from pH 2 to 8 (3,4). Detailed physiological studies have been done on
the influence of environmental pH range on growth, fermentation product formation and
the proton motive force in this organism (9, 25 , 39).

S. ventriculi also undergoes morphological adaptations in response to changes in
environmental pH. Regular tetrads are formed at low pH, whereas cells are irregular in
shape with higher numbers of cells within each packet at neutral pH, and spores are formed
at alkaline pH (24).

The purpose of this report is to examine the effect of perturbation of environmental
parameters (specifically of proton and solvent concentrations, and temperature) on
membrane composition and structure. We report that the lipid composition and fatty acid
structure dramatically change in response to these perturbations resulting in the synthesis of
a family of long chain a,a)-dicarboxylic fatty acids. These new lipids are believed to be

important for maintaining membrane integrity under the new conditions.

20

MATERIALS AND METHODS

Organism and Culture Conditions

All chemicals were reagent grade or better and were obtained from Sigma Chemical
Co., St. Louis, Mo., or Mallinckrodt, Inc., Paris, Ky. All gases were at least 99.9%
pure and were passed over copper-filled Vycor furnaces (Sargent Welch Scientific Co.,
Skokie, 111.) to remove oxygen. S. ventriculi JK was cultivated as described previously
(9). For growth under pH control, 4 liter Kimax jars (Baxter Scientific Products,
Romulus, Mich.) containing 3 liters of medium were used. The jars were equipped with a
pH probe, and the culture mixed by placing the jars on a magnetic stirrer. When the
organism was grown at pH 3.0, the initial pH did not change during the fermentation; at
pH 7.0 the pH was controlled by the addition of 5 M NaOH. Cultures were harvested as
outlined previously (24) under aerobic conditions.

For temperature shift experiments and growth in the presence of solvents, 750 ml
fermentation vessels (New Brunswick Sci., New Brunswick, NJ) containing 350 ml of
medium were used. The vessels were agitated at 200 rpm and pH maintained either at pH
3.0 or pH 7 .0. To determine the effect of solvents on membrane composition, vessels
containing medium with 0.25 M ethanol or 0.05 M butanol at pH 7.0, were inoculated with
a 5 % (v/v) inoculum of cells grown at pH 7.0 in the absence of solvent. . The cells were
harvested at midexponential phase, washed twice with distilled water and stored at -70°C
for further analysis.

Membrane Preparations

Cells were disrupted by passage through a French Pressure cell (American
Instruments Co., Inc., Silver Spring, Md.) at 20,000 lb/in2. Tire disrupted cells were
centrifuged at 20,000 x g to remove unbroken cells, and the supernatant was centrifuged at

110,000 x g to sediment the membranes, which were washed twice with distilled water.

21

Total Fatty Acids Analysis

Fatty acid analyses were performed on whole cells or isolated membrane fractions
by treatment with methanolic HCl using either of two procedures. Procedure (a) was
employed for whole cells and procedure (b) for isolated membranes. (a) cells (1-5 mg)
suspended with 0.3 ml chloroform and 1.5 ml 5% methanolic HCl solution, were sealed in
a teflon-lined screw-capped vial, and heated in a water bath or oven at 72°C for 24 hrs.
Chloroform (3 ml) was added every 8 hrs followed by mild sonication for 5 minutes.
After concentration to dryness under nitrogen gas, samples were partitioned between water
and chloroform and the aqueous layer washed several times with chloroform or hexane.
The combined solutions were filtered through glass wool. (b) Three(3) ml of chloroform
was added to 1 ml of membrane suspension followed by 15 ml 5% methanolic-HCI
solution. The flask was sealed and heated in an oven at 72°C for 12 hrs. Three (3) ml of
chloroform was added every 6 hrs followed by mild sonication for 5 minutes. The
mixture was then concentrated on the rotary evaporator to dryness and extracted with
chloroform. The combined organic fraction was redissolved in 1 ml of hexane. The fatty
acid methyl esters prepared by either procedure (a) or (b) were subjected to Gas
Chromatography analysis on a 25 M J &W Scientific DB1 capillary column using helium as
the carrier gas and a temperature program of 150°C initial temperature, 0.00 min hold time
and 3.0 deg/min rate, to a temperature of 200°C. A second ramp of 4.0 deg/min was then
immediately started until the final temperature of 300°C was obtained. This temperature
was held for 30 min. The relative proportion of lipid components were calculated from the
integrated peak areas. The fatty acid identification and molecular weight were determined
using GC/MS analysis using a Jeol JMS-AX505H spectrometer interfaced with a Hewlett-
Packard 5890A Gas Chromatograph.

Extraction of Lipids

22

Lipids were extracted from the isolated membrane or whole cells using procedures
(a) and (b) respectively. (a) To each 5-10 m1 membrane suspension 30 vol. of
chloroform/methanol (5:1,v/v) was added and then mixed to produce a single phase. The
mixture was shaken or stirred vigorously, with intermittent sonication (approximately 5
min every 30 min), for 2 hrs at 45°C. The combined extracts were taken to dryness in a
rotary evaporator. The residue was partitioned between 10 ml chloroform/methanol
(5:1,v/v) and 2.5 ml water. The lower phase was taken to dryness and redissolved in 1 ml
chloroform/methanol (9:1,v/v). (b) Cells of S. ventriculi from 50 liters of culture
medium were harvested by centrifugation at 10,000 x g for 10 min. Lipids from
approximately 50 g wet weight of cells were extracted at 45°C with 400 ml of a mixture of
chloroform/methanol/water (15: 3: 2, by vol.) for 2 hrs, followed by 200 ml of chloroform
/ methanol (5:1, v]v). Extraction was performed with intermittent sonication for 2 hrs.
After centrifugation at 20,000 x g the pellet (cell debris) was extracted again with the same
solvent systems. After centrifugation, the supernatant was taken to dryness in a rotary
evaporator, dissolved in 10 ml of chloroform/methanol (5:1,v/v) and then shaken with 2.5
ml water. The lower phase containing the lipids was taken to dryness and dissolved in 1

ml of chloroform.

Analysis of Lipids

As a preliminary step for comparing the total lipid profiles between cells of S.
ventriculi growing at pH 3.0 versus pH 7.0, the lipids were separated by 2-dimensional
TLC using chloroform] methanol] ammonia] water ( 3.3: 10:01: 0.05, by vol.) for the first
dimension and chloroform /methanol / water (7: 1.6:0.2, by vol.) for the second dimension.
Analyses were performed on silica-gel plates (Merck). Spots were made visible either by
spraying with 50% ethanolic-sulfuric acid and heating at 250°C to char the organic
components, or by spraying with a 0.1% solution of 2’,7,-dichlorofluorescein in aqueous

ethanol ( 1:1) and viewing under ultraviolet light (29). Standard phospholipids such as

23

phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylserine (PS) and
cardiolipin (CL) and neutral lipids (NL) were used as standards in addition to free fatty
acids (FA). Spraying agents (30) for the detection of components included ninhydrin for
PE or PS, dragendorff agent for PC, orcinol for glycolipid, and molybdenum blue for

phosphate.

Isolation of the onto-Dicarboxylic Acid Dimethyl Ester

The total lipids (100 mg) extracted from the cells as described before were
methanolysed with 5 % (w/v) HCl in methanol (5 ml) for 12 hrs at 72°C. Chloroform (1
ml) was added every 4 hrs followed by mild sonication for 5 minutes. The mixture was
then concentrated on the rotary evaporator to dryness and extracted with chloroform. The
combined organic fraction was redissolved in 1 ml of hexane. This fraction was applied to
a silica flash chromatography column and eluted with chloroform/hexane (1:1, by vol).
Fractions were assayed by Gas Chromatography. Fractions containing very long chain
fatty acid methyl esters were then concentrated and rechromatographed on 5% AgNO3
impregnated-preparative thin layer silica chromatography plates which were eluted with
petroleum ether] diethyl ether lacetone (10:1:0.5, by vol). Spots were made visible either
by spraying with 50% ethanolic-sulfuric acid and heating at 250°C to char the organic
components, or by spraying with a 0.1% solution of 2’,7,-dichlorofluorescein in aqueous
ethanol (1:1) and viewing under ultraviolet light (29). Bands were scraped from the plate
into a column fitted with a sintered disc and the material was eluted from the silca gel with
methanol and chloroform. Each fraction was concentrated by evaporation and redissolved
in choloroform for further analysis. Purity of each fraction was assayed by GC/MS. One
band containing the pure a,(u-15,16-dimethyltricotanedioate dimethyl ester was subjected

to NMR and FTIR analysis.

24

Isotope Labeling

Isotope labeling was used to aid in deducing the structures of esterified lipid
components using GC/MS. Methyl esters of fatty acids obtained by acid methanolysis
were further treated with 5% deuterated methanolic-HCl for 6 hrs at 72°C. Deuterated

methyl esters of fatty acids were extracted and analyzed as described for the natural

abundance samples.

1H NMR and 13C NMR Spectroscopy

Proton NMR spectra were recorded at 300 MHz on solutions in CDC13. Fourier
transform 13C NMR spectra were recorded at 125 MHz on solutions in CDC13. Chemical
shifts are quoted relative to the chloroform resonances taken at 7.24 ppm for proton and 77

ppm for 13C measurements, respectively.

Fourier Transform Infrared Spectroscopy

Spectra were obtained with a Nicolet model 710 FT-IR spectrometer on a 10%

(w/v) solution of dicarboxylic acid dimethyl ester in chloroform

25

RESULTS AND DISCUSSION

Two Dimensional TLC Analysis on Lipids

Experiments were initiated to examine lipid composition changes when cells of S.
ventriculi were grown at pH 7.0 versus 3.0. Two dimensional TLC analyses on the lipids
from intact cells or isolated membrane of cells are shown in Figure 1. Analysis of the
components for Specific lipids such as PE, PS and CL indicated that only traces of these are
present and the major components are glycolipids. There are markedly different
proportions of phospholipids and glycolipids in the cells grown at the two pH values.

In order to ensure efficient extraction of membrane lipids from the cells of S.
ventriculi grown at pH 3.0, a relatively non-polar solvent system was selected and used to

solubilize the strong hydrophobic lipid groups at elevated temperature of 45°C.

Compositional Analyses of Isolated Lipids

GC analyses of fatty acid methyl esters extracted from cells grown at pH 7.0 versus
3.0 are shown in Figure 2A and 2B respectively. Most of the fatty acids in the membrane
of S. ventriculi grown at pH 7.0 were between 14 and 18 carbon atoms long (Figure 2A)
but when cells were grown at pH 3.0, a family of novel lipid components appeared (Figure
2B). The relative proportion of the unusual and unique lipid components. was more than
50% by mass at pH 3.0, and only trace levels (< 7%) were present during growth at pH
7.0. Further analysis of these lipids from pH 3.0 cells using GC after methanolysis
revealed that peaks a, f, i, and j of Figure 1A contained unusual fatty acid groups. This is a
major difference between the cells grown at acid versus neutral pH. This indicated that the
unusual fatty acids are integral lipid components and do not occur in free uncombined

form.

Figure 1. Two dimensional TLC analyses of the lipids extracted from the
cell of Sarcina ventriculi grown at pH 7 (A) versus pH 3 (B). Equal amounts
(~ 0.8 mg) of total lipids of the membranes isolated from the cells of S. ventriculi grown at
different pH values were compared in 2D-TLC plates. A(a,b,c,j), B(a,b,k); phospholipids,
A(e,f,g,h,i), B(c,d,f,g,h,i,j); glycolipids. Spots .A(a,f,i,j) contained unusual fatty acids

with long retention times on GC/MS.

 

 

 

 

 

 

 

 

 

 

Figure l

figure

Membra
1355 Wit
trim}
'Q’bOxy
(13.-1 .Q
limos

4

tr) ' dice

Figure 2. Gas chromatographic analyses of lipid components in the
membrane of S. ventriculi grown at pH 7 (A) versus pH 3 (B). Total fatty
acids within the membrane were analde as fatty acid methyl ester derivatives after
methanolysis. 1. C14;o - carboxylic acid methyl ester, 2. C173 - fatty aldehyde, 3. C153
- carboxylic acid methyl ester, 4. C16:0 - carboxylic acid methyl ester, 5. unknown, 6.
C133 - carboxylic acid methyl ester, 7. C131) - carboxylic methyl ester, 8. C321) - (1,0)-
dicarboxylic dimethyl ester (mm-15,16-dimethyltricotanedioate dimethyl ester), 9. C343 -
(1,0) - dicarboxylic dimethyl ester , 10. (36:2 - a,co - dicarboxylic dimethyl ester .

UQCOQmOL LOQOQuQQ

Detector response

L

 

 

 

 

 

 

 

 

 

56
4
3 7
i
4 6
5
7
9
8 10

Retention time (effective boiling points)

Figure 2

GC’)

mOICC]
dClllcn
mass L‘-

1301ch

30

Figure 3A and 3B show the mass spectra of one of the unusual lipid components
which appeared in the membrane of S. ventriculi grown at high proton concentration (pH
3.0). The same lipid components were also present in cells grown in the presence of high

alcohol or higher temperatures at pH 7.0.

GC/MS Analyses of Dicarboxylic Dimethyl Esters

The electron impact mass spectrum in Figure 3A shows major ions at m/z 538, 506
and 475 including characteristic McLafferty fragment ions 74, 87. These corresponded to
the molecular ion of a C32-0t,(u—dicarboxylic dimethyl ester (M'l') with the sequential losses
of methanol (CH3OH) and a methoxy (CH3O) group respectively. The ion 464 mlz
represents the structure obtained by sequential losses of methanol (CH3OH) and ketene
(CHZCO) (Figure 4, Table l).

The mass spectrum also contained a series of ions the general structure of which
was CH30CO-(CH2)n, begining at m/z 73, charateristic of a saturated methyl ester. This
series of ions continued up to and included two prominent ions at m/z 269 and 297 (296 is
due to loss of one hydrogen from m]z the 297 fragment). The intense clusters of ions 28
mass units apart centered at m]z 269 (n=15) and mlz 297 (n=l7) strongly indicated the
presence of a vicinal dimethyl group (Figure 4). The ions at m/z 237 and 265 represent the
loss of methanol (CH3OH) from the mlz 269 and 297 ion fragments respectively. Two
other major group of ions at m/z 238 and 266 were assigned to acylium ions (m/z 238; OC-
(CH2)13-CH(CH3), m/z 266; OC-(CH2)13-CH(CH3)CH(CH3)) corresponding to loss of
methoxy groups from the ions of m/z 269 and 297 ions respectively.

Figure 3B shows the electron impact mass spectrum of the deuterium labeled
molecule obtained by rrrethanolysis with the D-4 methanol/HQ solution (to produce the tri-
deuterated methyl ester). The most striking change was the shift of the molecular ion by 6
mass units which confirmed methanol (CD3OH) and a methoxy group (CD30) from the
molecular ion (544 mlz) to give ions at m/z 509 and 510 were also observed in addition to

31

Figure 3. Electron impact mass spectrum (70ev) of a,w-lS,l6-dimethyltri-
cotanedioate dimethyl ester without (A) and with isotope labeling (B) (A)
El spectrum contained major ions at m/z 538, 506 and 475. These corresponded to
molecular ion (M"') with the sequential losses of methanol and methoxy group respectively.
(B) It shows the El Spectrum of deuterium labeled molecules obtained by deuteration with
D—4 methanolic-HCl solution. The presence of two carboxylic and two branching methyl

groups was confirmd by isotope labeling.

 

 

 

7‘"

05

«M

 

 

T

 

Relative Abundance

 

so.

 

a?

8 cl.

 

Relative Abundance

 

Figure 3

1
(Al
Al
dl
I
l
l
l
l
l
I
l
I
I
l
l

I
K ‘J 113 C)
ocn3 \ “
227 ‘1 \\ OCH3
---------- \

r--1

o I
-_ __2_ra_.<_.'_—3_‘.:_251_JC"3 i
I
I :
.32 -42
237 -<+————269 i454 g
I

-31 -32
475 :— 506*——r MT

. McLafferty Fragment
74

Figure 4. Electron impact mass spectral fragmentation pattern of 41,0)-
15,l6-dimethyltricotanedioate dimethyl ester. All numbers indicate the ionic
fragments (mlz). 31 mass units indicate the methoxy group (~CI-I3O), 32 for methanol
(CH3OH), 42 for ketene(CI-12CO). The ion at mlz 74 is a characteristic McLafferty

fragment of aliphatic ester groups.

Table 1. Analysis of 70 ev electron impact mass spectral fragments of (1,00-

15,16-dimethyltriacotanedioate dimethyl ester.

 

 

Structure of ionic fragment. mass(m/z) deuteriated mass(mlz)
ca,oco<cnz),,CHCH,CH,CH(CH2).,ococn, 538 544
cn,oc0(crrz),,cncrt,crr,crr(crr,).,cu=co 506 509
0C(CH,).,CHCH,CH,CH(CH,),2CH=CO 475 . 475

506 - l8 (H20) 488 491
CH3OCO(CH,).,CHCH3CH3CH(CH2),0CH=CH2 464 467
CH,oc0(cnz),,crrcrr,crr,crr 3” 3‘3
cnrocmcaaucarcnacmcua 297 30°
crr,oc0(crrz),,cncn, 269 272
camera)“ 2‘" 7’“
cotcrrgucncn, 238 238
wm0M1 74 77

Cl-I,OG)(CH¢). mat-13 73 + l4‘n 76 + l4‘n
cannery), : oar-43 55+l4*n 55+14‘n
CH,-(CH2). . as 3-12 57+l4‘n - 5mm:

cnfcn-(CHau :n=Z-ll 55+14tn 55mm

35

the deuterated McLafferty fragments (m/z 77). The ions at m/z 272 ((CD3OCO-
(CH2)13CH(CH3)) and 300 (CD3OCO-(CH2)13CH(CH3)CH(CH3)) corresponded to
deuterated m/z 269 (CH3OCO-(CH2)15) and mlz 297 (CH3OCO-(CH2)13 CH (CH3)
CH(CH3)) respectively. The ions at m]z 237 and 265 corresponded to loss of tri-
deuterated methanol (35 mass units) from the ions at m/z 272 and 300 respectively. Most of
the other unusual long chain fatty acid components gave spectra with similar properties
indicating that they had structures consistent with a,co—very long chain dicarboxylic acids.
Some of them contained one unsaturated double bond (C34;1-0t,co-dimethyldicarboxylic

acids) or two unsaturated double bonds (C36;2-mw-dimethyldicarboxylic acids).

Spectroscopic Analyses of Isolated Dicarboxylic Dimethyl Esters

To definitively prove the general structure of the family of dicarboxylic acids, one
of these was isolated by a combination of column chromatography and preparative T.L.C.
This component corresponded to 0t,(o-15,l6-dimethyltricotanedioate dimethyl ester.

Proton NMR analysis (Figure 5) also confirmed the presence of branched methyl
groups by the presence of a doublet at 0.74 ppm (J = 7.2 Hz) and the methoxy group by
the presence of a sharp singlet at 3.65 ppm. Resonances at 8 1.28 (broad singlet) were
assigned to methylene groups of a lipid chain. The multiplet at 8 1.58 was assigned to the
protons of the B-carbons of the molecule. Resonances at 5 2.28 (triplet) were assigned to
the methylene groups at to the carbonyl group (-CHz-CO-). Ester methoxy groups (CH3O-
) were assigned to singlet at 8 3.65. The 13C NMR spectrum (Figure 6) showed the ester
carbonyl carbon at 8 174.5, methoxy carbon at 8 51.4 and branched methyl carbons at 5
14.5. Moreover, all these resonances were present as singlets, indicating magnetic
equivalence due to molecular symmetry.

The Infrared spectrum (Figure 7) showed strong C-H stretching absorption in the
area of 2840-3000 cm'l. The bending vibrations of the C-H bonds in the methylene
groups occurred around 1465 cm'1 (scissoring) and 1350-1150 cm'1 (twisting and

Figure 5. The 300 MHz 1H NMR spectrum of anor-15,16-dimethyltricotan-
edioate dimethyl ester. Signals at 8 3.65 which were assigned to methyl groups of a
methoxycarbonyl function. The peak at 8 1.28 is due to the methylene groups in the
saturated hydrocarbon chains. Note also 3 characteristic signals, at 8 2.28 (t, J =13.2Hz)
representing the methylene group adjacent to the methoxy carbonyl function, at 8 1.58 due
to the protons of the B-carbons of this molecule and at 8 0.74 (d, J=7.2 Hz) corresponding
to the protons of a vicinal methyl group. The singlet at 8 7.24 is due to the chloroform.

Peaks marked with asterisks are due to contaminants from the silica layer.

 

 

 

 

Tflfi

g.

 

A

Figure 5

Figure 6. The 125 MHz l3C-NMR spectrum of a,w-15,16-dimethyltrico-

tanedioate dimethyl ester. The presence of a methyl ester and branching methyl group

are confirmed by signals at 8 51.4 and 14.5 respeCtively. The signal at 8 174.5 was

assigned to the carbonyl carbon of the ester group.

Figure 6

 

 

 

 

 

 

 

 

 

“4‘44‘11‘1‘1‘4444‘d“‘~“r‘“l4“"44‘Jdd‘14qdi‘“l‘l‘lfl“““.“‘ ‘r-r‘4“““4‘4“‘-““d‘r‘“““"““‘4‘1‘4-““‘4“4““‘

N00 .00 «mo .A0 .8 .00 00 Go no no 0 con

Figure 7. The Fourier transform infrared spectrum of a,w-lS,l6-dimethyl-
tricotane dioate dimethyl ester. It shows strong C-H stretching in the area of 2840-
3000 cm'l. The bending vibrations of the C-H bonds in the methylene group occurred
around 1465 cm'1 (scissoring) and 1350-1150 cm'1 (twisting and wagging). The
characteristic C=O absorption band of the aliphatic ester group appeared at 1735 cm‘l.

There was no evidence of unsaturation or of oxygen combined in hydroxy or simple ether

linkages.

XTranenlttence

I00

02.410 70.020 07.000

34.000

”0.4000 17.000

 

 

 

LN

 

3775.0 5409.2 5043.: £077.. £332.: 3040.3. iaeom 1234.0 -oea.n

41

 

 

 

 

 

 

 

 

l

4

Hovenuuber (cu-1)

Figure 7

42

wagging). The characteristic C=O absorption band of the aliphatic ester group appeared at
1735 cm'l. It also showed the stretching of the carbon-oxygen bond of the alkoxy group
(O=C-O-R) in the area of 1175-1250 cm‘l.

With all these spectroscopic data, the structure of this component was confirmed as

-a,m-15,16-dimethyltricotanedioate dimethyl ester.

Induction of Dicarboxylic Acids in Response to Various Forms of
Environmental Stress

The alteration in metabolic end product formation occurring as a function of pH
includes cessation in acid production at low pH and an increase in ethanol formation.
Either event may cause the organism to alter cellular components to adapt to these changes.
The cell may synthesize these unusual very long chain onto-dicarboxylic fatty acids in
response to the presence of acids during growth at pH 3.0, in order to prevent translocation
of protons across the cytoplasmic membrane by the protonated form of the acid and
dissipation of the proton motive force. In other bacteria the undissociated forms of these
acids are able to permeate through the cytoplasmic membrane and accumulate inside the cell
at large ApH values and decrease the internal pH (19).

The presence of solvents would increase fluidity of the membrane and could trigger
production of longer chain fatty acids to increase membrane integrity. As growth of S.
ventriculi at low pH is inhibited in the presence of elevated levels of metabolic acids, but
not solvent, experiments were conducted at neutral pH in the presence of high levels of
solvents to determine if any changes occurred in the lipid composition of the cell. S.
ventriculi grew at pH 7.0 in solvent concentrations as high as 0.6 M methanol, 0.5 M
ethanol, 0.3 M propanol and 0.10 M butanol. To examine the effect of solvents on the
fatty acid composition of S. ventriculi at neutral pH, the organism was grown in the
presence of 0.25 M ethanol or 0.05 M butanol (concentrations at which growth was

unaffected), and the fatty acid composition detemnined. If the ethanol levels produced in

43

the cells of S. ventriculi grown at pH 3 can trigger the synthesis of unusual long chain lipid
component then this mechanism would be supported by demonstrating that addition of
exogenous ethanol to cells grown at pH 7.0 has the same effect on membrane composition
as the addition of exogenous acid (i.e., growth at pH 3.0). Cells of S. ventriculi grown at
pH 7.0 in the presence of ethanol or butanol (Figure 8A and B) contained a similar family
of unusual lipid components as those observed in cells grown at pH 3.0. These findings
suggest that high levels of ethanol, butanol or protons (organic acids) can trigger the

synthesis of the very long chain fatty acids, and that the phenomenon is not specific.

Effect of Alcohol and Thermal Stress on Lipid Composition

There are numerous reports on the adaptations bacteria make in lipid composition
when grown in the presence of alcohols (1, 6, 14, 15, 16, 22). Membrane integrity is in
many cases the primary site of alcohol, in particular ethanol, damage although alcohol
clearly affects the properties of all biological macromolecules to some degree. Ethanol
adaptive changes have been divided into two groups; those that strengthen the hydrophobic
barrier and those that decrease the lipid sites available for passive leakage (17). The
former include the production of longer chain fatty acids and increasing the proportion of
nonpolar lipids in the membrane. This mechanism involves the extension of the average
chain length by a few carbon chains after the perturbation (18). The other group of
adaptive changes include a reduction of the lipid to protein ratio and a reduction of the lipid
pitches on the membrane surface sites available for passive leakage (7,15).

Our results indicate a much more dramatic mechanism because the average chain
length of the normal fatty acids remain the same after addition of alcohol and an entirely
new class of very long chain membrane lipids are synthesized. Since the effect of diverse
alcohols on membrane composition seems to be similar in S. ventriculi, it became apparent
that the synthesis of these unusual lipid components might be a general response to

environmental factors which might disrupt the membrane and reduce its fluidity. If the

44

fluidity of the membrane were the real controlling element for this adaptive response, heat
as another external perturbation would be expected to have a similar effect as alcohol or
protons. Temperature effects on the composition of the membrane have been thoroughly
studied in E. coli (16,27,28). The configuration of unsaturated carbons (cis- or trans) or
hydrocarbon chain length of fatty acids are known to be responsible for temperature
adaptive changes because configurations and increased chain length can influence the
packing density and hence fluidity. The optimum temperature for growth of S .ventriculi
is 30-37°C(5). Experiments were performed under pH 7.0 controlled conditions to
determine the effect of prolonged incubation (4hrs) at 45°C or 55°C on membrane
composition. Figure 8C shows the GC profile of membrane lipid components after
methanolysis of the membrane lipids of the cells of S. ventriculi grown at pH 7 after a
temperature shift from 37°C to 55°C. Incubation at elevated temperature resulted in
synthesis of the long chain fatty acids even though the organism became nonviable. Cells
of S. ventriculi were viable after a shift in temperature to 45°C, and synthesis of the onco-
dicarboxylic very long chain acids occurred at this temperature. For the reversibility of the
synthesis of these unusual membrane components, the temperature was shifted to 45°C
from 37°C at the mid-log phase of growing cells of S. ventriculi at neutral pH. A small
amount ranging from close to zero and rarely exceeding 7% of long chain lipids were
always found in cells of S. ventriculi grown at neutral pH, and probably represents the
basal level of these lipids. After shifting the temperature to 45°C, the amounts of long
chain lipid components dramatically increased and returning the temperature to 37°C
reduced their proportions. In order to demonstrate that decrease in very long chain fatty
acids was due to the decrease in temperature, rather than turnover or dilution before the
temperature was returned to 37°C, part of the culture was removed and the incubation at
45°C continued. The temperature of the remainder of the culture was brought down to

37°C and after 90 min the very long chain dicarboxylic acid content in the 45°C culture

Figure 8. Gas chromatographic analyses of lipid components extracted
from cells of S. ventriculi grown at pH 7.0 in the presence of various
forms of environmental stress. (A) 0.25 M ethanol, (B) 0.05 M butanol, (C)
elevated temperature (55°C), 1. C14;o - carboxylic methyl ester, 2. C173 - fatty
aldehyde, 3. C151) - carboxylic acid methyl ester, 4. unknown, 5. C133 - carboxylic acid
methyl ester, 6. C131) - carboxylic acid methyl ester, 7. C323) - (1,0) - dicarboxylic
dimethyl ester (a,a)-15,16-dimethyl tricotanedioate dimethyl ester), 8. (234,1 - a,co -
dicarboxylic dimethyl ester, 9. C353 - (1,0) - dicarboxylic dimethyl ester .

 

 

 

 

 

 

 

 

Retention time

 

 

 

 

 

 

 

Figure 8

47

was much higher than the culture at 37°C (data not shown) demonstrating that synthesis of
which fatty acids was reversible and dependent on temperature.

E .cali compensate for an increase in growth temperature or the presence of ethanol
by increasing the relative abundance of saturated fatty acids (14, 26, 37). The same
changes are also observed when Clostridium acetobutylicum is grown in the presence of
alcohol ( 1, 22). Our finding suggest that the ratio of saturated to unsaturated fatty acids is
not decreased by increasing temperature or solvent but rather that S. ventriculi overcomes
the increase in membrane fluidity by increasing the ratio of the very long chain fatty acids to
regular length fatty acids (Table 2). Such a response has also been observed in C.
acetabutylicum with increasing growth temperature (1, 22), but the increase in chain length
is not as dramatic as that observed in S. ventriculi, whereby the chain length almost
doubles. In Zymamonas mobilis longer chain fatty acids appear to be beneficial for the
high ethanol tolerance exhibited by the organism (6). In Z. mobilis the mechanism by
which ethanol tolerance is achieved appears to be a concerted shift in the relative amounts
of phospholipids, hopanoids and proteins in the cell envelope. The large ethanol-
dependent shifts in the hopanoid content suggest a major function of these sterol-like
substances for membrane stabilization (34).

There have been a few other reports of bacteria producing unusually long chain
fatty acids. Lactabacillus heterohiochii is a spoilage organism of Japanese wine, growing
in concentrations of alcohol greater than 20%, possessing short chain fatty acids and
considerable amounts of saturated and mono-unsaturated fatty acids ranging in length from
C20 to C30 (40, 41). Sulfate-reducing bacteria, Desulfatomaculum sp. and three strains of
D. ruminis contained or-hydroxy and onto-dicarboxylic acids with more than 20 carbons,
and in addition very long chain fatty acids up to C34, were identified (33). A Butyrivibrio
spp. (strain 82) occurring in the rumen has an absolute requirement for fatty acid, and a
new series of long chain dicarboxylic acids (C2035) were found to be major components of

the lipids (10,11,20). Another unusual C30 fatty acid is present in Clostridium

4.92m N. mama. on maaqoaamaa «2.0% o: 95”: Saws. 8:. =5 non—.8 on $3332.

.3333 5533:.

2. =5 83. mean 3

 

 

35. ”9333. ex So em So + em So + um flo em 90 um Po
9.9 8.8 3 mass. 3.9 8.8 K was? 3.9 3.9 9.9 3.9
$2.33 3..“ who who uoh wok. a: l
camESSa 3b 3..» .33 gob 3.0 2b
mamas AA 88 v 3 mm.» um.» um.» 3:0 ab
<3. Loam 255 Av 83 A .N fem firm 3.“ 3.0 u:
ea. 2 gm. N... we. we. 3.. 9.3 :3
wwao on. <%. 0.8 0.3 . 0.2 0.3 0.8 rou

 

a were?“ 858:8 cm Swansea? $8889 amaze comm... 5»: No 830: £083 2. <3. Loam are: 858 9.5 mm 8383 mp3.

more £2... aoanasoa 3. 90 030530: om Samara“ Raw 808 on on gamma a. 80: 8.533.

a csmmegoa 35. mowemmgsd 33. man.

0 S5. 83m 93: $3. woMEwomEB. 95. man

49

thermohydrosulfuricum and Cl. thermosulfurogenes, and this dicarboxylic acid accounts
for about 10 and 23% respectively of the total apolar chains of these organisms (21). Fatty
acids such as 27-hydroxyoctacosanoic acid bipolar long chain fatty acids are thought to be
key membrane components for controlling the fluidity of bacterial membranes in other
systems where oxygen diffusion has to be regulated or where the organism might have to
survive intracellularly within eucaryotic organisms (2,12,13).

Recently, homeoviscous adaptations in the archaebacterum, Methanococcus
jannashcii, have been reported in which changes in the proportions of diether, macrocyclic
ether and tetraether lipids of the membrane were found at different temperatures (36). The
relative amounts of tetraether lipid (C40) increased with increasing temperature relative to
the level of diether lipid (C20). This indicates that the transmembrane lipid components
function to modulate the reduced viscosity due to increased temperature thus, maintaining
the optimum fluidity.

Increased temperature, solvents, or proton levels disrupt membrane order and lead
to increased molecular motion, affecting the viscosity of the membrane. This may trigger
the synthesis of the very long hydrophobic lipid components which would reduce the
increased viscosity. There are several classical mechanisms which are known to be
activated in response to these changes. The first is the heat shock gene phenomenon (23)
whereby proteins are synthesized in response to various forms of stress including heat
(31), ethanol (32, 38, 42), puromycin (8), anaerobiosis (35) and pH (7). It is not clear to
what extent the phenomenon we see in S. ventriculi is related to heat shock or stress
proteins.

Om' results suggest that there is a common membrane response to various
environmental stresses in S. ventriculi. This response involves the synthesis of a new
family of very long cam-dicarboxylic acids which may control membrane fluidity. From
the structural data the two polar groups on the dicarboxylic groups are present at either end,

which suggests that this unusual lipid component probably spans the membrane, a position

50

which would be more energetically favorable.

51

CONCLUSIONS

Changes in the composition of membrane lipids in a strictly anaerobic, facultative
acidophilic eubacterium, Sarcina ventriculi were studied in response to various forms of
environmental stress. Changes in lipid composition and structure occurred in response to
changes in environmental pH. At neutral pH, the predominant membrane fatty acids
ranged in chain length from C14 to C13. However, when cells were grown at pH 3.0, a
family of unique very long chain fatty acids containing 32 to 36 carbon atoms was
synthesized and accounted for 50 % of the total membrane fatty acids. These acids were
identified as very long chain cum-dicarboxylic acids ranging in length from 28 to 36
carbons by electron impact mass spectrometry of methyl and (perdeuterio) methyl ester
derivatives. These methyl esters all bore a vicinal dimethyl group toward the center of the
chain. The assignment of the structures was confirmed by isolating one of the very long
chain unusual fatty acids as the ester form after methanolysis and performing further
analyses including 1H and 13C NMR (Nuclear Magnetic Resonance) spectroscopy and
Fourier Transform infrared spectroscopy. Coupling this information with the data from
Gas Chromatography/Mass Spectrometry (GC/MS) analysis, the exact structure was
confirmed as maHS,l6—dimethyltricotanedioate dimethyl ester.

Addition of alcohols, either metabolic (0.25 M ethanol) or nonmetabolic (0.05 M
butanol) to cells grown at pH 7.0, or thermal stress (growth temperature at pH 7.0 was
raised from 37°C to 45°C or 55°C) also resulted in the synthesis of these very long chain
fatty acids. Synthesis of these very long chain (mu-dicarboxylic acids was reversed by
reducing the temperature back to 37°C. S. ventriculi is also unusual in that the membrane
components are not the usual phospholipid components but appear to be predominantly

glycolipids.

1.

99°99?"

10.

11.
12.

13.

14.
15.
16.
17.

52

REFERENCES

Baer, S. H., H. P. Blaschek, and T. L. Smith. (1987) Appl. Environ. Microbial.
53, 2854-2861

Bhat, U.R., Mayer, H., Yokota, A., Hollingsworth, R. I., and Carlson, R. W.
(1991) J. Bacterial. 173 , 2155-2159

Canale-Parola, E. (1970) Bacterial. Rev., 34, 82-97

Canale-Parola, E. (1986) Genus Sarcina Goodsir 1842, 434AL, p. 1100-1103. In P.
H. A. Sneath, N. S. Mair, M. E. Shaarpe and J. G. Holt (ed.), Bergey’s manual
of systematic bacteriology, vol. 2. The Williams & Wilkins Co., Baltimore

Carey, V. C., and L. 0. Ingram. (1983) J. Bacterial. 154, 1291-1300

Dombek, K. M., and L. 0. Ingram. (1984) J. Bacterial. 157, 233-239

Foster, J. W., and H. K. Hall. ( 1990) J. Bacterial. 172, 771-778

Goff, S. A., and A. L. Goldberg. (1985) Cell 41, 587-595

Goodwin, 8., and J. G. Zeikus. (1987) J. Bacterial. 169, 2150-2157
Hazlewood, G. P., N. G. Clarke, and R. M. C. Dawson. (1980) Biochem. J.
191, 555-569

Hazelwood, G. P., and R. M. C. Dawson. ( 1979) J. Gen. Microbial. 112, 15-27
Hollingsworth, R. 1., and Carlson, R. W. (1989) J. Biol. Chem. 264, 9300-
9303

Hollingsworth, R. 1., and Lill-Elghanian, D. A. (1989) J. Biol. Chem. 264,
14039-14042

Ingram, L. O. (1976) J. Bacterial. 125, 670-648

Ingram, L. O. (1977) Can. J. Microbial. 23, 779-789

Ingram, L. O. (1982) J. Bacteriol. 149, 166-172

Ingram, L. O. (1990) Crit. Rev. Biatechnal. 9, 305-319

18. Ingram, L. 0., and N. S. Vreeland. (1980) J. Bacteriol. 144, 481-488

19.

20.

21.

22.

23.
24.

25.

26.

27.

28.

29.
30.

31.

32.

33.

34.

53

Kell., D. B., M. W. Peck, G. Rodger, and J. G. Morris. (1981) Biochem.
Biophys. Res. Comm. 99, 81-88

Klein, R. A., P. G. Hazlewood, P. Kemp, and R. M. C. Dawson. (1979)

Biochem. J. 183, 691-700

Langworthy, T. A., and J. L. Pond. (1986) Membranes and lipids of
thermophiles, p. 107-135. In T. D. Brock (ed.), Thermophiles, General,
Molecular, and Applied Microbiology. Wiley & Sons, New York.

LePage, C., F. Fayolle, M. Hermann, and J.-P. Vandecasteele. (1987) J. Gen.
Microbial. 133, 103-110

Lindquist, S. (1986) Ann. Rev. Biochem. 55, 1151-1191

Lowe, S. B., H. S. Pankratz, and J. G. Zeikus. (1989) J. Bacteriol. 171, 3775-
3781.

Lowe, S. B., and J. G. Zeikus. (1991) Arch. Microbial. 155, 325-329

Marr, A. G., and J. L. Ingraham. (1962) J. Bacteriol. 84, 1260-1267
deMendoza, D., and J. E. Cronan, Jr. ( 1983) Trends Biochem. Sci. 8, 49-52
deMendoza, D., A. K. Ulrich, and J. E. Cronan, Jr. (1983) J. Biol. Chem. 258,
2098-2101

Morris, J. L. (1962) Chem. & Ind. 1238-1240

Holme, J. D. and Peck, H. (1983) Analytical Chemistry pp 436-467 Longman,
Inc. New York .
Neidhardt, F. C., R. A. VanBogelen, and V. Vaughn. (1984) Ann. Rev. Genet.
18, 295-329

Plesset, J., C. Palm, and C. S. McLaughlin. ( 1982) Biophys. Res. Commun.
108, 1340-1345

Reznak, T., M. Yu. Sokolov, and I. Viden. (1990) FEMS Microbial. Ecol. 73,
231-238

Schmidt, A., S. Bringer-Meyer, K. Porolla, and H. Sahm. (1986) Appl. Microbial.

54

Biotechnol. 25, 32-36
35. Spector, M. R, Z. Aliabadi, T. Gonzalez, and J. W. Foster. (1986) J.
Bacterial. 173, 3907-3910.
36. Sprott, G. D., M. Meloche, and J. C. Richards. (1991) J. Bacterial. 173,
3907- 3910
37. Sullivan, K. H., Hegeman, G. D. and Cordes, E. H. (1979) J. Bacteriol.
138, 133-138
38. Terracciano, J. S., E. Rapaport, and E. R. Kashket. (1988) Appl. Environ.
Microbiol. 54, 1989-1995
39. Tilak, K. V. B. R. (1970) Sci. Cult. 36, 399-400
40. Uchida, K. (1974a) Biachim. Biophys. Acta 348, 86-93
41. Uchida, K. (1974b) BiachimBiaphys. Acta 369, 146-155
42. VanBogelen, R. A., P. M. Kelley, and F. C. Neidhardt. (1987) J. Bacteriol. 169,
26-32

BI

CHAPTER III

CHEMICAL PROOF FOR THE FORMATION OF VERY LONG (1,0)-
BIFUNCTIONAL ALKYLSPECIES IN THE MEMBRANE OF SARCINA
VENTRICULI BY TAIL-TO-TAIL COUPLING OF EXISTING ALKYL

CHAINS FROM OPPOSITE SIDES OF THE BILAYER

55

Cliff

men
tern]
hot
Com
imp
thcs

11nd;

firm
the 1
mor

Iran

10 11.
W011.

mac]

56

INTRODUCTION

The membrane is a universal structural feature which is responsible for signal and
energy transduction, compartmentalization of biological processes and maintenance of the
structural integrity of the cell in all living systems, from microorganisms through plants and
animals. Membranes play a central role in critical events such as DNA replication, cell
division, protein synthesis, electron transfer and photosynthesis. Despite this, due largely
to the preeminence of molecular genetics approaches, most of the current research on
cellular adaptation to environmental stress focuses on proteins and nucleic acids and not on
lipids which are the primary membrane components.

There is the growing concept that proteins are the structural entities which control
membrane fluidity at elevated temperatures rather than lipids and that changes in external
temperature is sensed directly by proteins with no stated involvement of lipids (2, 3, 4).
Protein changes are easier to demonstrate and understand genetically. Lipid changes are, in
contrast, much more difficult to demonstrate, more difficult to understand and close to
impossible to trace back to a specific gene or genes. Probably because of a combination of
these factors, the roles of lipids in ensuring cell survival during stress is often
underestimated and very poorly understood

Proteins are thought to be the universal entities which structurally preserve cell
function and integrity. This model is that proteins sense environmental changes and cause
the synthesis of new proteins which then alter membrane lipid structure (5, 6). A simpler,
more desirable system would be one in which changes in membrane dynamics are
transmitted through the lipids to trigger a constitutive, fluidity-regulated, dynamic
adaptation system. This system would act on the membrane lipids to return the membrane
to its original fluidity or dynamic state (at which the adaptation system is inactive). This
would constitute the ultimate in adaptation systems since the entire transcription/translation

machinery would not have to be set in motion. It is therefore, independent of the very

57

cellular events (the viability of which are in question during stress) which it is trying to
preserve. In the case of bacteria, a likely course would be for the organism to adopt a
membrane structure which is typical of bacteria whose normal habitat is similar to the new
environmental conditions. In the case of substantial temperature increases from normal
ambient temperatures, the new membrane structure might resemble that found in
thermophilic eubacteria (7, 8, 9) or archaebacteria (10, 11) in some important ways.
Bacteria with a very dynamic range of environmental tolerance are especially suited as
model systems for studying these adaptation phenomena since the chemistry of the
membrane can be studied as a function of the perturbed state and mutants which are tolerant
to the new conditions do not have to be obtained.

Sarcina ventriculi is a gram positive, strict anaerobic eubacterium which can adapt
to environmental pH values between 3 and 9 (12, 13, 14). The organism synthesizes very
long chain axe-dicarboxylic acids in response to lowering pH, the addition of organic
solvents or raising the temperature (1). The gas chromatography profile of the
methanolysate of the total cell mass after these perturbations (Figure 1) showed two
separate domains of fatty acid chain lengths. An early eluting one due to regular chain
length fatty acyl components and a later one due to components with lengths twice that of
the regular fatty acyl components. This latter domain contained the methyl ester derivatives
of the auto-dicarboxylic acids. The unusual chain length distribution and the. large variety of
very long chain dicarboxylic species formed in the membrane of S. ventriculi during the
response to various environmental perturbations suggest the possibility of an unusual, new
mechanism for their formation. The typical 2-carbon addition process involving the acetyl
Co-A cannot be used to explain the structures(a,(o-bifunctional) and chain length
distribution of the very long chain components of the membrane of this species. Here, we
report chemical evidence for the mechanism of formation of very long chain bifunctional
alkyl species by tail-to-tail addition. of hydrocarbon chain from opposite sides of the

bilayer. This would constitute an entirely new, very sophisticated mechanism for

Figure l
lipids in
and then
tits. isd
l-Cm-
19612"
icfil'
{Cm
4CBJ
B-Cm
CCKJ
0‘ C1931‘

Ecfifli

58

Figure 1. Gas Chromatographic analyses of the total methanolysate of the
lipids in the membrane of Sarcina ventriculi cells grown at pH 7.0 at 37°C
and then shift to 45°C at late log phase for 3hrs. The later eluting cluster of
peaks is due to very long chain ogre-bifunctional fatty acids. The major components are :

l. C17;1-fatty aldehyde (HCO(CH2)9CH=CH(CH2)4CH3 )

2. C16;o-carboxylic acid methyl ester (CH3OCO(CH2)14CH3)

3. Cum-carboxylic acid methyl ester (CH3OCO(CH2)9CH=CH (CH2)5CH3)

4. Clam-carboxylic acid methyl ester (CH30CO(CHz)15CH3)

A. C333 co-formylmethyl ester

B. C32;o-a,co-dicarboxylic acid dimethyl ester

C. C35:1 (e-formylmethyl ester

D. C34;1-0t,0)-dicarboxylic acid dimethyl ester

E. C35;2-a,a)-dicarboxylic acid dimethyl ester

GKCCQIGL

LGUUH-dnve

 

Detector respOnse

810R?

 

59

 

 

 

 

 

 

 

 

 

Retention time (effective boiling points)

Figure I

comb
intol'
speci:
spcci:
bk": .t
acid.

two is
the Ill
mcml
the 12
Unsat
glen;
3Cyl 5
an as:

the Cl

10 dc:
dun‘n
Chair,
Plasn
regill;

PTOpQ

60

membrane adaptation. The conclusions about the mechanism are drawn from the
satisfaction of certain structural and stereochemical features of the new bifunctional lipids
imposed by the nature of the mechanism.

It is clear from the structure of the fatty acid methyl ester (15,16-dimethyltriacotane-
1,30-dicarboxylic acid dimethyl ester) described in our earlier study (1) that the
combination of the regular fatty acyl chains to form the new bifunctional species could
involve coupling between the (0-1 positions of the chains of two hexadecanoyl (C16)
species leading to the vicinal dimethyl group. Hence coupling between two C15;o acyl
species should yield a 15 ,16-dimethyl C30 bifunctional molecule. A C30 dimethyl
bifunctional molecule can also be obtained by the similar coupling of a C13 and a C14 fatty
acid. However, the vicinal methyl groups will be located at the 13 and 14 positions. These
two isomeric acids can be distinguished by mass spectrometry since fragmentation between
the two methyl groups is dominant. The only unsaturated, regular fatty acid in S. ventriculi
membranes is cis-vaccenic acid (a C13 acid). In this molecule, the double bond occurs at
the 11 position. If the (0-1 carbon coupling theory is correct, the very long, bifunctional
unsaturated alkyl species should have a cis double bond at the 11 position and a methyl
group at the 17 position relative to the closest carbonyl carbon. In instances when the two
acyl species are identical, combination at the (0-1 positions by this mechanism should yield
an asymmetric bifunctional species provided that the two hydrogen atoms which are lost in
the coupling are removed from the same side of the intermediate, partially bonded species
(Figure 2). Another consequence of this proposed mechanism is that it should be possible
to detect co-formyl long chain acids among the very long chain bifunctional species formed
during the membrane adaptation process. These would be formed by coupling of the acyl
chains of phospholipid fatty ester groups with the alkenyl chains of plasmalogens.
Plasmalogens are known to be present since fatty aldehydes were detected among the
regular chain length components of the methanolysate (Figure 1) in comparatively minor

proportions. It should therefore be possible to rationalize the detailed structural features of

61

O OCH; Chiral Synthesis

cm H
CH3 H
l

on, ‘0

Cl8:l &Cl8:l

 
  

(Enzyme-mediated)

cn,

 

 

(336:2 (Optically inactive)

Figure 2. Rationalization of the stereochemistry of the products formed by

tail-to-tail coupling of alkyl chains from opposite sides of the bilayer.

62

any of the bifunctional lipid species by considering the Structural features of the regular
chain length species. Satisfaction of these requirements should constitute proof of the
transmembrane coupling mechanism. In this work we describe the isolation and
characterization of a variety of fatty acids from S. ventriculi cultured under conditions
which promote the synthesis of bifunctional, very long chain fatty acids. This allowed the

definitive chemical proof of the transmembrane lipid coupling mechanism.

63

MATERIALS AND METHODS

Organism and Culture Conditions

S. ventriculi JK was cultivated as described previously (1). For growth under pH
control, 4 liter Kimax jars (Baxter Scientific Products, Romulus, Mich.) containing 3 liters
of medium were used. The jars were equipped with a pH probe, and the culture mixed by
placing the jars on a magnetic stirrer. The organism was grown at pH 3.0. The cells
were harvested at midexponential phase, washed twice with distilled water and stored at

-70°C for further analysis.

Membrane Preparations

Cells were disrupted by passage through a French Pressure cell (American
Instruments Co., Inc., Silver Spring, Md.) at 20,000 lb/in2. The disrupted cells were
centrifuged at 20,000 x g to remove unbroken cells, and the supernatant was centrifuged at

110,000 x g to sediment the membranes, which were washed twice with distilled water.

Total Fatty Acids Analysis

Fatty acid analyses were performed on whole cells or isolated membrane fractions
by treatment with methanolic HCl using either of two procedures. Procedure (a) was
employed for whole cells and procedure (b) for isolated membranes. (a) cells (1-5 mg)
suspended with 0.3 ml chloroform and 1.5 ml 5% methanolic HCl solution, were sealed in
a teflon-lined screw-capped vial, and heated in a water bath or oven at 72°C for 24 hours.
Chloroform (3 ml) was added every 8 hours followed by mild sonication for 5 minutes.
After concentration to dryness under nitrogen gas, samples were partitioned between water
and chloroform and the aqueous layer washed several times with chloroform or hexane.

The combined solutions were filtered through glass wool. (b) 3 ml of chloroform was

64
added to 1 ml of membrane suspension followed by 15 ml 5% methanolic-HCI solution.

The flask was sealed and heated in an oven at 72°C for 12 hours. Three (3) ml of
chloroform was added every 6 hours followed by mild sonication for 5 minutes. The
mixture was then concentrated on the rotary evaporator to dryness and extracted with
chloroform. The combined organic fraction was redissolved in 1 ml of hexane. The fatty
acid methyl esters prepared by either procedure (a) or (b) were subjected to Gas
Chromatography (GC) analysis on a 25 M J&W Scientific DBl capillary column using
helium as the carrier gas and a temperature program of 150°C initial temperature, 0.00 min
hold time and 3.0 deg/min rate, to a temperature of 200°C. A second ramp of 4.0 deg/min
was then immediately started until the final temperature of 300°C was obtained. This
temperature was held for 30 min. The relative proportion of lipid components were
calculated from the integrated peak areas. The fatty acid identification and molecular weight
were determined using GC/MS analysis using a Jeol JMS-AXSOSH spectrometer interfaced
with a Hewlett-Packard 5890A Gas Chromatograph.

Isolation of the a,co-Dicarboxylic Acid Dimethyl Esters

The total lipids (100 mg) extracted from the cells as described above were
methanolysed with 5 % (w/v) HCl in methanol (5 ml) for 12 h at 72°C. Chloroform (1
ml) was added every 4 hours followed by mild sonication for 5 minutes. The mixture was
then concentrated on the rotary evaporator to dryness and extracted with chloroform. The
combined organic fraction was redissolved in 1 ml of hexane. This fraction was applied to
a silica flash chromatography column and eluted with chloroform/hexane (1:1, by vol).
Fractions were assayed by Gas Chromatography. Fractions containing very long chain
fatty acid methyl esters were concentrated and rechromatographed on 5% AgNO3
impregnated-preparative thin layer silica chromatography plates which were eluted with
petroleum ether/ diethyl ether lacetone (10:1:0.5, by vol). Spots were made visible either

by spraying with 50% ethanolic-sulfuric acid and heating at 250°C to char the organic

65

components, or by spraying with a 0.1% solution of 2’,7,-dichlorofluorescein in aqueous
ethanol (1:1) and viewing under ultraviolet light . Bands were scraped from the plate into a
column fitted with a sintered disc and the material was eluted from the silca gel with
methanol and chloroform. Each fraction was concentrated by evaporation and redissolved

in choloroform for further analysis. Purity of each fraction was assayed by GC/MS.

Isotope Labeling

Isotope labeling was used to aid in deducing the structures of esterified lipid
components using GC/MS. Methyl esters of fatty acids obtained by acid methanolysis
were further treated with 5% deuterated methanolic-HCl for 6 hours at 72°C. Deuterated

methyl esters of fatty acids were extracted and analyzed as described above.

1H NMR and 13C NMR Spectrosc0py
Proton NMR spectra were recorded at 500 MHz on solutions in CDC13. Fourier

transform 13C NMR spectra were recorded at 125 MHz on solutions in CDC13. Chemical
shifts are quoted relative to the chloroform resonances taken at 7.24 ppm for proton and 77

ppm for 13C measurements, respectively.

Fourier Transform Infrared Spectroscopy
Spectra were obtained with a Nicolet model 710 FT-IR spectrometer on a 10%

(w/v) solution of dicarboxylic acid dimethyl ester in chloroform.

Optical Rotation Analysis (Polarimetry)
Optical polarimeter (Perkin Almer) was used for measuring the optical activity with
the mercury (Hg) light at various wavelengths (578 nm, 546 nm and 436 nm).

Measurements were performed on 20mg samples dissolved in 1 ml chloroform solution.

66

Reductive Ozonolysis (15)

For ozonolysis of unsaturated lipids, a solution of 50 mg sample in 2 ml methylene
chloride was cooled to -70 oC. Ozone was passed through the solution at a rate of 60
mg/hr from a ozone generator while maintaining the temperature at -70°C. The extent of
the reaction was checked every 10 min. by TLC. Ozonides were decomposed to
aldehydes by reductive cleavage using Lindlar catalyst. These shorter chain aldehydes then
were isolated by preparative TLC or flash column chromatography (16). The products
were further analyzed by GC and GC/MS.

67

RESULTS AND DISCUSSION

GC Analyses of Fatty Acids of S.vcntriculi

Sarcina ventriculi were grown at 370C, pH 7.0, and then shift to 45°C at late log
phase for 3hrs. GC analyses of fatty acid methyl esters extracted from cells are shown in

Figure 1. The peaks from 1 to 4 are due to typical membrane fatty acyl components ranging

from 14 to 18 carbons. Major regular fatty acids are hexadecanoic acid (peak 2, C161,) and
cis-vaccenic acid ( peak 3, C1330”). The peaks from A to E, however, correspond to

very long bifunctional fatty acids components containing from 32 to 36 carbon atoms.

Mass Spectrometric Analyses of Very Long Bifunctional Fatty Acids
The mass spectrum of peak B indicated that it was the previously described (1)

or,m-diacid 1. The electron impact mass spectrum of Peak D (Figure 3A) contained major

ions at rn/z 564, 532 and 501.
0 CH H OCH
01 OWWMW 3
’ CH, 9H o
1

onto-(15,16-dimethyl)-triacotanedioate dimethyl ester

These corresponded to the molecular ion of a C34-a,m—dicarboxylic dimethyl ester (M+)
with the sequential losses of methanol (CHgOH) and a methoxy (CH3O) group,
respectively. The ions at m/z 514 was assigned to the structure obtained by sequential
losses of methanol (CH30H) and water (HzO) from the parent ion. The mass spectrum

also contained a series of ions, the general structure of which corresponded to CH2=CH-

68

Figure 3. Electron impact mass Spectrums of peak D. (A) Note the major ions
at m/z 564, 532 and 510. These corresponded to molecular ion (M+) with the sequential
losses of methanol and methoxy respectively. The cluster of ions at m/z 263, 296 and 324
are due to the vicinal methyl groups. (B) The electron impact mass spectrum of
deuterium labeled molecule obtained by deuteration with D-4 methanolic HCl solution. The
presence of two carboxyl groups was confirmed by the shift of 6 units of the M+ ion.

Other predicted mass increases due to deuterium were easily observed.

 

 

 

 

 

69

 

 

5'."

 

 

“1.1L.
. . a
o n
4
4
I (I
.. a
.\
a a. 4 t
C
2 (.{m
3 .t
o d W
. 2
W I
3 H
91 .
7 .1
.. m
8
a
6 -
S

 

10° r—f

A

Rel-(Ive abundance

 

llldtldtt . b l L It ll

1

l
6“
m

 

’ I

ll

‘
.1 11‘

v

a

'—

 

/I
I
Iain-414.1-

Q

4*

 

 

1 l 4 1 4 1“ I l 4

Relative abundance

Figure 3

7O
(CH2)n (n=2,3,4..), characteristic of an alkene. It also contained a series of ions, the
general structure of which corresponded to CH3OCO-(CH2)n, beginning at m/z 73 (n=1),
as well as the characteristic McLafferty fragment at m/z 74 (CH3OCOH=CH2) of a
saturated methyl ester. This series of ions beginning at m/z 73 continued up to and
included two prominent ions at m/z 263 and 296 suggesting the presence and location of
vicinal methyl groups. Scission between these methyl groups explained the primary
fragments at m/z 269 and 295. The general structure of the molecule was consistent with

the coupling between hexadecanoic and cis-l l-OCtadecanoic acid at their (1)-l positions to

give structure 2.

0 CH3 “3‘ H O
CH OA/VWVWWWK
3 CH3 H 00";
2

a,w-(17,18-dimethyl)-cis-ll-dotriacotaenedioate dimethyl ester

The ion at m/z 263 corresponded to the loss of methanol (CH30H) from the m/z 295
fragment. The ion at m/z 296 is due to a rearrangement resulting in elimination of a
hydrogen atom from m/z 297 fragment or capture of a hydrogen atom by the m/z 295
fragment. The ion at m/z 237 represented the losses of methanol (CH3OH) from the m/z
269. The fragments are tabulated and assigned in Table 1. Figure 313 shows the electron
impact mass spectrum of the deuterium labeled molecule obtained by methanolysis with D-
4 methanol/HCI solution to produce the trideutero methyl ester. The most striking change
was the upward shift of the molecular ion by 6 mass units which confirmed the presence of
two methoxy groups. The losses due to trideutero methanol (CD301!) and the sequential
loss of a trideutero methoxy group (CD30) from the molecular ion (m/z 570) give ions at
m/z 535 and 501. The deuterium labeled McLafferty fragment (m/z 77) was also observed.
Ions at m/z 272 and 299 corresponding to m/z 269 and m/z 296 in the unlabeled

71

Table 1. Analyses of 700v electron impact mass spectral fragments of

preposed structure 2.

 

 

Strucrurc of ionic fragments mass(m/z) deuteriated mass(m/z)
CH,OCO(CH,),CH=CI{(CHI).CHCH,CH,CH(CH,)‘,OCOCH, $64 $70
CH,OCO(CH,),CH=CH(CH2).CHCH,CH,CH(CH,_),,CH=CO 532 535
0C(CH,),CH=CH(CH,).CHCH,CH,CH(CH,)ucu=co so: 501

532 - l8 (H20) 5:4 517
CH,oc0(CH,),CH=CH(cr4,).CHCH,CH,CH(CH,).004:an 490 493
CH,0C0(CH,),CH=CH(CH,).CHCH,CH,CH - KB) 324 327
CH,0CO(CH,),CH=CH(CH,).CHCH, 297 300

297 - r (H) 296 299
C111:0C0(C|“Iz)tsCHCH: 259 272
oqcam=ca(crt,).cncn, 266 266
magnum, 237 237
CH,OC(OH)=CH, * 74 77
CH,OC0(CH,). ; 1:21-13 73 + 1011 . 76 + [4‘11
mad-[(011). : n=l~l3 55+l4'n 554-1011
Cay-(CHI). : as 342 5744011 574-100

mm :oaaa-u 55+l4‘n 55444.11

CS

‘7...)
’ 3

C0

C01

ab
mt

COL

Stab

E.

M

72
ester were also observed. The ion at m/z 263 corresponded to the loss of trideuterated

methanol (CD3OH) from the ions at m/z 298. Other major fragments are listed in Table 1.
Although the mass spectra indicated the exact positions of vicinal methyl groups and the
presence of one degree of unsaturation in the molecule, the location of the double bond
could not be deduced and was proven by chemical degradation. The mass spectral
fragmentation is illustrated in Figure 4.

Figure 5A shows the electron impact mass spectrum of peak E. This spectrum
contained major ions at m/z 590, 558 and 527. These corresponded to the molecular ion of
a bis-unsaturated C36—a,co-dicarboxylic dimethyl ester (M+) with the sequential losses of
methanol (CH3OH) and a methoxy (CH3O) group, respectively. This was consistent with

coupling of two cis—l l-octadecanoic acid chains to form structure 3.

O CH: 3“ H
WWW/W 0 C1 !3
CHJO '0.
CH, ‘H o

3
(mo-(17.l8-dimethyl)-cis-11,23-hentriacotadienedioate dimethyl ester

The prominent ion cluster at m/z 293 indicated fragmentation between the vicinal dimethyl
groups in the acyl chain followed by the elimination of hydrogen to produce two identical
stabilized. bis-unsaturated acyl chain fragments . The 28 mass unit difference between the
primary 293 and 321 chain cleavage fragments also indicated the presence and location of
vicinal dimethyl groups.

Figure SB shows the electron impact mass spectrum of the deuterium labeled peak
E molecule obtained by methanolysis with D-4 methanol/HG solution to produce the tri-
deuterated methyl ester. A 6 mass unit increase in the molecular ion (consistent with two

methyl ester functions) was observed as well as the sequential losses of uideutero methanol

73

p--------------------o

o x 0
cu, ,’ “on 241 41%le
x‘ .‘ ’1’- ------------------
/ 1
CH30 3265—. 31-29.-" "2% 5 : i ’
CH3 ‘1’ ’H‘\\ ‘90 42 H’)
-32 . l ‘ l
263 «— 295-<-l 323 i ' ~31 .32 i
--------------------- : "591:.52"? “‘0‘”

-23 ~32
245 <— 26} +—- 295

I---------------------

Figure 4. Electron impact mass spectral fragmentation pattern of peak D
(assigned structure 2). All numbers indicate mass fragments. 42 Mass units indicate
ltetene (01200), 32 indicate methanol (CH3OH), 31 a methoxy (ongoj group and 28 for
ethylene (CHz-=CH2). The ion at m/z 74 is the characteristic Mclafferty fragment of an

aliphatic methyl ester group.

74

Figure 5. The electron impact mass spectrum of peak E (assigned structure
3 ). (A) Major ions appear at m/z 564, 532 and 510. These correspond to the molecular
ion (M+) with the sequential losses of methanol and a methoxy group respectively. The
cluster of ions at m/z 265, 293 and 321 confurn the presence and location of the vicinal
methyl groups. (B) The electron impact mass spectrum of deuterium labeled peak E
obtained by deuteration with D-4 methanolic HCl solution. The presence of two carboxyl
groups was‘confirmed by isotOpe labeling. Mass increases due to deuterium were easily

observed. The most striking change is the increase of the molecular ion by 6 mass units.

75

 

SSO

 

 

 

 

 

 

 

 

 

 

 

e
v.
2
S
III.
7
S
m 4%
6
O 7
4
fl 4
x t
a 3. 9
O m 53 ,’
1 3
b
3
t a
4.1 8
Z 3 3
3 .
Z 1
o O
' Z.
0 Z
0
1 8
Z
I‘
o
7
JO.
3 1 a
9 1 I?
.8
”ll! tli.e
. 3ll 1
film
r .. l
OII’J 11111 (1‘11!th 0
0 t
m m m 4 n.

Rclntivc abundAncc

 

8

 

1 I ‘ I ‘

t Mam

 

ll};
Figure 5

 

 

 

 

 

Reintlvc abundance

76

 

_234-+—2—'- 265 +33— 267 x 3,” ‘~ 41 O:

"""""""""" , ~ 2.9§.—*.--_-2_9}.—*____2§_2.__ :

H2 tr V“. I

.‘__ 4__' ‘ -32 '

293 321 323 4' ~,_ 295 —> 263 E
------------------------- 42 -----------------

 

516 <
or .32
527 +——— 5534— M‘(590)

- ............................ J

Figure 6. Electron impact mass spectral fragmentation pattern of peak E
(assigned structure 3 ). All numbers indicate mass fragments (mlz). 112 mass units
indicate ketene (012(1)), 32 indicate methanol (CH30H), 31 a methoxy (CI-130) group
and 28 for ethylene (CHz=CHz).

77
(CD3OH) and a trideutero methoxy group (CD30). Figure 6 shows the mass spectral

fragmentation pattern of peak B.

NMR and FTIR Analyses of Isolated Dicarboxylic Dimethyl Esters

Proton NMR (Nuclear Magnetic Resonance) analysis of peaks D and E (Figure 7
and 8, respectively) confirmed the presence of the double bond and vicinal methyl groups
deduced from GC/MS data. In the spectrum of peak D (Figure 7), the methyl groups
resulted in a 6H doublet at 8 0.74 (J=7.20 Hz). A triplet at 8 5.32 (J=6.92 Hz) in both
spectra was assigned to the vinyl protons. In the case of peak D, this triplet integrated for

two protons. However, in the case of peak B it integrated for four protons. Resonances at

8 1.27 were assigned to the methylene groups of the lipid chain. The multiplet at 8 1.58
was assigned to the methylene protons [3 to the carbonyl group. Resonances at 8 2.28 (t,
J=7.69 Hz) were assigned to the methylene groups a to the carbonyl function. A 6H
singlet at 8 3.65 was assigned to ester methoxy group resonances . The 4H multiplets at 8
2.00 was assigned to the protons of the methylene groups a to the vinyl carbons. The
13C NMR spectrum of peak D (Figure 9) contained resonances ascribed to the ester
carbonyl carbons at 8 174.5, methoxy carbon at 8 51.4 and branched methyl carbons at 8
14.5. The vinylic carbons appeared at 8 131.5. In the 13C NMR spectrum of peak E
(Figure 10), an increase in intensity of the signal at 8 130.8 (vinylic carbons) was
observed This spectrum only contained 19 signals confirming the proposed symmetry of
the molecule. Information on the configuration of the double bonds as well as other
information confu'ming the specific chemical functional groups was obtained by Fourier
Transform Infrared Spectroscopy. The infrared spectrum of peak D (Figure 11) showed a
strong aliphatic C-H asymmetric stretching absorption at 2928 cm1 and symmetric
stretching at 2856 cm4. The characteristic alkene stretching (=C-H) was observed at
3020 cm‘l. A band due to the bending vibration of aliphatic C-H bonds in the methylene

groups appeared at 1464 cm'1 (scissoring) and one due to the twisting and wagging

78

 

 

 

 

Figure 7. The 1H NMR spectrum of pr0posed structure 2 (peak D). The
signal at 8 5.32 (t, J=6.92Hz) represented a methine proton on an unsaturated carbon atom.
The mutiplet at 8 2.00 was assigned to the protons of methylene groups adjacent to the
unsaturatedcarbons.1hepeakat8127isduetothemethylenegroupsofflrehydrocarbon
chains. The signal at 8 3.65 is due to the methyl group of the methoxycarbonyl functions.
Themultiplet at 81.58arisesfiomtheprotonsontbe carbons fitnthecarbonylfunction.
The doublet at 8 0.74 (J=7.20Hz) is due to the protons of the vicinal methyl groups. The

signal at 8 7.24 is due to chloroform.

79

 

 

Figure 8. The 1H NMR spectrum of pr0posed structure 3. The signal at 8 5 .32
(t. J=6.83Hz) is due to a methine proton on an unsaturated carbon atom. The mutiplet at 8
2.00 is assigned to the protons of the methylene groups adjacent of the unsaturated carbons.
The intense peak at 5 127 is due to the methylene groups ofthe hydrocarbon chains. The
signal at 8 3.65 result from methyl group of the methoxycarbonyl functions. The multiplet
at 81.58 fortheprotons ofthecarbons Btothemethylcarbonylfunetion. Thedoubletat8

0.70 (J=7.58Hz) arise from the protons of the vicinal methyl group.

1?

Fl:

80

 

 

 

 

 

 

 

 

 

Figure 9. The 13C NMR spectrum of proposed structure 2,. Signals at 8 131.5
correspond to olefinic carbons. Signals at 8 174.5 and at 8 51.4 confirmed the presence of
methyl ester group. The vicinal methyl group gives rise to the signal at 814.5.

71':

81

 

 

 

 

 

 

 

 

 

. . ~ ' "‘ ' ":’. . ‘-'.:“” l'7- IV. 0. 'u. ' - .l- .'-' a . t d ..O 'v ' - r
u. RAtLLtntJJul'wu'. s'bheua‘Uwvus-rrl’!‘HMil»Ill-int;ill.t'.'u..tt.hrw:-it~‘t allliv'nau'urh-w'dn'eil h'qu-lnlbuutsulr

I” I“ 8“ I ran .00 on ‘0 d I. C".

Figure 10. The ”C NMR spectrum of proposed structure 1. Signal: at 8 130.8
correspond to olefinic carbons. Signals at 8174.8 and at 8 51.5 confirmed the presence of
methyl ester group. The vicinal methyl group gives rise to the signal at 8 14.8.

82

103.9

825

 

611
*3-

 

VoTtrnsmiluncc

 

 

39.7

 

 

 

A‘

 

 

 

18.3

 

A A A l
v V ‘

 

3469 31;: 2327 2:95 2184 um lsol ll70 839 so:
What-l)

Figure 11. The Fourier Transform Infrared spectrum of proposed structure

2. The characteristic olefinic C-H stretching vibration appeared 3020 cm" and the out of

plane cis C—-C-H bending vibration at 667 cm'1 . Typical C=O stretching vibration at

1732 cm1 for the carbonyl group is also present.

83

deformation at 1315 cm'l. The characteristic C=O absorption bond of the aliphatic ester
group appeared at 1732 cm'1 and the ester alkoxy stretch at 1213 cm'l. The
stereochemisuy (cis- or trans-) of the unsaturated carbons was determined by the clear
presence of a strong peak at 667 cm-1 due to the C=C~H bending deformation for cis—
alkenes (17). The IR spectrum of peak E also indicated a cis- configuration for the two
unsaturated double bonds in the molecule. Note that the configuration of the double bonds
of the regular length C13;1 (11) acyl chains was cis from cis vaccenic acid (1 1-octadecenoic
acid). With regard to the mechanism of the formation of peak D or B, one of the most
important questions was the location of the unsaturation in the acyl chain. The regular
length monofunctional unsaturated acid contained the double bond at the 11 position of its
carbon chain. The determination of the exact position of unsaturation of the bifunctional
acyl chain was therefore very important since if it were located at any position other than

the 11 position (with respect to the closest end) our model would be invalidated.

Reductive Ozonolysis and Mass Spectrometry

The exact locations of the unsaturation in 2 and 3 were confirmed by ozonolysis
followed by reduction of the ozonide with zinc/acetic acid. This led to several fragments
which were assigned structures A , 5, and 6, (Figure 12) based on their mass spectra and
were consistent with the proposed locations of the double bonds. As expected, structure 2
produced two fragments on ozonolysis in contrast to three fragments (two of which were
identical) for structure 3. The mass spectrum of 4 (Figure 13A) contained major ions at
m/z 186, 183, 171, 74 and 87. These corresponded to the ions of M (m/z 214) - 28
(CH2=CH2), M - 31 (CH3O), M - 43 (CH2=CH=O) and the typical McLafferty fragment
(m/z 74) of an aliphatic methyl ester. The ion at m/z 139 was assigned to the sequential
losses of CH2=CH-O and a methoxy group from the molecular ion. Figure 13B shows

the EI mass spectrum of the deuterium labeled molecule with structure 4 obtained by

84

Figure 12. Fragments expected for the reductive ozonolysis of structure 2
(A) and 3 (B). Each product was analyzed by GC/MS with and without isotope
labeling.

C1130

033°

0 CH3 _“ 0
Wk 03 ZNACOH
CH30 - OCH3 .____*_ _____,__
C113 “u
.2 (MW 564)
W H “H O
‘H
1(M.W. 214) 5. (M.W. 382)
CH: H 03
5‘ anAcOH
- OCR; k 7"-
CK: "H o
3

1(M.W. 214)

85

 

Figure 12

86

Figure 13. The electron impact mass spectrum of 1. (A) Note the major
ions at 186, 183, 171 and 74. These correspond to loss of ethylene (CHz=Cl-I2) from the
molecular ion (M+=214), loss of methoxy group frem M+, loss of CHz=CH-O from M+
and the typical McLafferty fragment (at m/z 74). The ion at m/z 139 was assigned to the
sequential losses of CHz=CH—O and a methoxy group from the molecular ion. (B) The
electron impact mass spectrum of deuterium labeled :1 obtained by deuteration with D-4

methanolic HCl solution. Note the expected mass increases because of deuterium.

 

87

 

 

 

 

 

 

 

 

 

171

 

 

87

 

w

100‘

 

Rclgtive abundance

 

 

T

l-

,1.

1

.li‘

11

1‘1:

1

fi

‘

 

r50

 

 

 

Rclct‘vc Rhflfl‘ancc

 

1 1 i ‘ ‘

Figure 13

88

performing the methanolysis with D-4 methanol/HCI solution. The expected mass
increases due to the deuterium were easily observed .

The EI mass spectrum of 5 (Figure 14A) contained major ions at m/z 382, 364,
357 and 350. These were attributed to the molecular ion (M+) and the losses of water
(H20), ethene and methanol respectively. It also contained the typical ion series of the
general structure (CH3OCO-(CH2)n), beginning at m/z 73, for a saturated methyl ester.
There were three prominent ions at m/z 237, 269 and 297 (296 is due to the loss of one
hydrogen from m/z 297). The intense clusters of ions 28 mass units apart centered at m/z
269 And 297 indicated the presence of a vicinal dimethyl group. The ion at m/z 237
represented the loss of methanol (CH3OH) from the ion at m/z 269. Figure 14B shows
that the EI mass spectrum of the deuterium labeled molecule with structure 5. The
expected increases in mass units due to isotope labeling were easily observed. Figure 15
shows the electron impact mass spectrum of the dialdehyde (5,). It contained major ions at
m/z 208, 182 and 95. These were assigned to M+-18(H20), M-44(CIlz=CH-OH) and m/z
113-18(H20). The ion at m/z 113 was designated to the product formed by the
fragmentation between the two methyl branches. There was no difference in the EI mass
spectrum before and after the deuterium labeling, thus indicating the absence of a methyl
ester function. The dimethyl acetal which should have been formed would have been
hydrolysed back to the aldehyde during work up. .

These reductive ozonolysis studies provided structures consistent for peaks D and
E. They also provided conclusive proof for the proposed coupling mechanism since the
exact positions of unsaturation of the bifunctional acyl chains and the position of the vicinal
methyl groups were completely consistent with (1)—1 coupling of two monocarboxylic acid
acyl chains, one of which was cis-vaccenic acid. The combined spectroscopic and
chemical methods confirmed the proposed structures corresponding to peaks D and E as
onto-(17,]8-dimethyl)-cis-1l-dotriacotaenedioate dimethyl ester (1) and onto-(17,18-

dimethyl)-cis-11-23-hentriacotadienedioate dimethyl ester Q), respectively. Compounds

89

Figure 14. The electron impact mass spectrum of 5. (A) Note the major
ions at 382, 364, 339 and 269. These can be attributed to M+, M-18 (H20), M-28
(CH2=CH2), M-43(CH2=CH-O) and M-113 (HCO(CH2)5 CH (CH3)). (B) The
electron impact mass spectrum of deuterium labeled 5 obtained by deuteration with D-4

methanolic HCl solution. Mass increases because of deuterium are easily observed.

 

 

“ C“ ,u

 

 

 

90

 

 

 

 

 

 

 

Relative abundance

A

 

 

 

 

 

 

1 1 1 1 1 1 1 1

I 1 1 1 1 1 1 1

Relative abualaacc

Figure 14

91

 

 

 

 

 

 

 

we 93
~ 1
r; 1 H or, H
. am 0W0
'3 1 a" v“ H
v l
e ‘ é
a 68‘
b 6
,, ‘r‘
n 1
d .
A ‘0 s 1 3 1
n 1 a
c 1
. 23111 I
1 ._ Zr
’ ' 1; 3L fikv V f 1 V i .1 i ' V w v v 1 fi'fi 1
100 280 380 408 588 seam

Figure 15. The electron impact mass spectrum of a. Note the major ions at
m/z 208, 182 and 95. The molecular ion peak (M+) was not observed but M+ - 18 (1120)
giverisetotheion atria/2208. 1hestrongpeakatm/z95correspondstothelossofwaér

from the fragmentation products obtained from cleavage between the two methyl

branches.

92
1. 2 and 3, all displayed an optical rotations (Table 2). This precluded the possibility of

any of these bifunctional acyl chain species being racemates or meso—compounds.

Molecular Mechanism for the Formation of Very Long Chain Dicarboxylic
Acids

The presence of very long bifunctional lipids in the membrane of S. ventriculi could
have arisen by coupling the tails of fatty acids of membrane lipids from the same side of the
bilayer. This, incidently, was ruled out by freeze-fracture studies which demonstrated that
the leaflets could not be separated and that the very long chain fatty acids were
transmembrane (data not shown). Several pieces of information showed unequivocally that
the formation of these very long fatty acids occurred by the chemical linkage at the 00-1
positions of the tails of fatty acids from opposite halves of the bilayer. The structures and
gross relative proportions of individual very long chain fatty acids were closely correlated
with the structures and the relative proportions of normal chain fatty acids in the same
isolates. Since C151) and cis-C13;1(u) fatty acids were the major regular chain species, the
predominant very long chain fatty acid species were combinations of (316:0 plus C151),
C15;o plus cis-C13;1(11) and cis-C13;1(11) plus cis-C13;1(11). In this case the C133 species
was cis-l 1-octadecenoic acid (ctr—vaccenic acid) thus explaining the positions of the double
bonds of the unsaturated onto-dicarboxylic acids. When an odd-numbered carbon alcohol
was added to the growth medium, the regular chain fatty acids were all of odd carbon
number (mostly C17;o, cis-C17;1 and C15;o). This is consistent with propanol being
converted to propionyl CoA which was used to initiate fatty acid synthesis with a C-3 unit
leading to odd-numbered fatty acids. However, in this case, the new very long fatty acids
were all of even chain length and predominantly saturated C-3O (strictly speaking dimethyl
C-28), monounsaturated C-32 (dimethyl C-30)-and di-unsaturated C-34 (dimethyl C-32)
dicarboxylic acids. Further proof of the (0-1 coupling mechanism was obtained by the

structural comparison of the two different 032 (or vicinal methyl branched C-30) octo-

93

 

 

 

 

 

 

Species
Hg light (“m 1 Z a
[a 1578 0.51 5.44 5.76
[“15“ l 0.89 6.30 6.46
[(11436 5.49 11.39 11.49

 

 

 

 

Table 2. Optical rotation analyses of peak B, D and E. It indicates that all of
bifunctional molecules are chiral.

94

Figure 16. Mass spectral fragments of C32“) onto-dicarboxylic very long
chain acyl species (2) in the case of propanol (0.15M) induction. The
location of the vicinal methyl groups was determined by GC/MS analyses. Figure 17A
shows the EI mass spectrum. Figure 16B shows the EI mass spectrum of deuterium

labeled 2. The ions at m/z 258 and 286 corresponded to the trideutrated C151) fatty acid
methyl ester and trideuterated C171) fatty acid methyl ester.

95

___-_____._——-

r
. 255 JL—r' 223

{283 ——’—3—2—+—- 25l

C17:0

(L004 14, lS-dimethyl)-triucotanedioak dimethyl ester

 

 

 

 

 

 

‘D.
R $:F
C
I
a “r
t
i
v
c 1
“1
a
3
u 1
.. 5:
J a. b
C
a
c
c
z" 4 .
1
j
5' ‘U
m

 

 

 

 

 

 

8.
a ‘ 9“
f 1
a 'r
1
I
V
‘ 1
“1
a
O
U
C ,7 b
1 “ .
a ‘ 45‘.
. C .
c a l 8 ‘ ,, , i-
I ‘;: I‘ ’ .1
1W! l .’ . l r i ;.'1' t "t" ”3 . l a ' .
an an 5“ ‘fl
”0 39'

Figure 16

96

dicarboxylic fatty acid esters obtained from the two different induction conditions: propanol
induction and temperature induction. They are both vicinal methyl—branched C-3O 0t,00-
dicarboxylic acids. However, careful analyses of the mass Spectra clearly indicated that the
one formed on propanol induction (Figure 16) bore the vicinal methyl group in a position
consistent with the combination of C151) plus C171) to give structure 2. The mass
spectrum contained major ions at m/z 538, 506 and 475. These corresponded to the
molecular ions and sequential losses of methanol (CH3OH) and a methoxy (CH3O) group,
respectively. Important fragments giving information on the position of the vicinal methyl
groups appeared at m/z 255 and 283. These prominent peaks corresponded to the
fragmentation products formed by cleavage between the two methyl branches. In contrast,
the mass spectrum of the acid with the symmetrically placed vicinal methyl group
(consistent with a combination of C151) plus C15;o) obtained from acidic or high
temperature culture conditions showed the predominent alkyl chain cleavage ion m/z 269
(1). Another proof of the proposed mechanism came from the isolation of a very long
bifunctional species formed by the combination of a fatty acid and a fatty aldehyde. Using
GC/MS (Figure 17A), it was possible to demonstrate the presence of normal chain fatty
aldehydes such as 1 indicating the presence of plasmalogens. Corresponding to-
formylmethyl esters with structures such as 8 (structure 1 plus C15;o) were identified
(Figure 17B) by mass spectrometry.

The optical rotations of the dicarboxylic acids clearly showed them to be chiral.
This is consistent with a transition state which gives rise to the stereochemistry having two
centers of asymmetry. In an enzymatic process leading to coupling of the alkyl chains by
the proposed mechanism, the methyl groups would be expected to be directed away from
the face of the enzyme and would therefore appear on the same side of the Neuman
projection (Figure 2). A possible mechanism would be absuaction of a hydrogen radical
from each of the two Opposing fatty acids from the same side followed by coupling of the
radical species. The secondary radical is more stable than a primary radical hence coupling

97

Figure 17. The electron impact mass spectrums of Z. (A) Note the major ions
at m/z 252, 224 and 55+l4n (n=1, 2 ..). These corresponded to molecular ion (M+), M+ -

28 (C0) and a long chain alkene series respectively. (B) The electron impact mass
spectrums of 8 contained major ions at m/z 520, 492 and 460. These corresponded to

molecular ion (M+) with the sequential losses of carbon monoxide (CO) and methanol

(CH3OH) respectively.

98

 

 

 

 

 

 

 

 

 

 

 

 

_ r e
_
_ 7. .m
o m
.m
.Zi
a 11
w;
a
4m
1_
n.
will: _
r mlihi
T .1. 11...:
-1...
Tl l a
14.7. r c
m a m 14.11%- 1..

Relative Abundance

1.}..r3‘00 crittr.‘ {rt},

v v V

v

'N
“I“

v

459
1.

EA

3‘

f

 

 

f. 5.

a"?

2 i 11:”.
”31;!‘1‘l:. 1 9‘

‘

9:3 ‘

vvfi—erV

o
r
in

'70:!
M

4 “3.4.2411.

Relative Abundance

Figure 17

99
at the (0-1 instead of (1)-positions. The resulting molecule would have the (R, R) or (S, S )
stereochemistry and would be chiral even if the two chains are identical as is the case in
su'uctures 1 and 3. If coupling of two identical chains occurs on the same side of the
bilayer by the same mechanism, the resulting product would be mesa and have the (R, S)

or (S, R) stereochemistry.

The General Significance of the Coupling Mechanism

The unusual mechanism has very important implications for adaptive processes in
bacteria. The tail of the fatty acids of a lipid can be expected to have the highest freedom
of motion in the entire molecule since the other end is firmly attached to the head group. It
is, therefore, desirable for the organisms to develop methods or mechanisms to restrict this
motion especially under conditions (such as increased temperature or the presence of
organic solvents) which will tend to increase it. One very important aspect of the adaptive
response that we describe here is that it accomplishes this motional restriction by the most
direct mechanism one could imagine: the direct coupling of fatty acid tails across the bilayer
in a random, indiscriminate fashion. Another striking aspect of this adaptive response is
that it utilizes pre—existin g fatty acids rather than initiating the complete biosynthesis of the
new species in a de novo fashion. This makes the response much more effective since the

element of speed (which is of the utmost importance for survival) is ensured.

100

CONCLUSIONS

Uncoupling the maintenance system and survival mechanisms of biological
membranes from events which depend on them for their proper execution is a key step in
adaptation. Such a system was studied in a strictly anaerobic, facultative acid0philic
eubacterium Sarcina ventriculi. In a previous study, we demonstrated that this organism is
capable of adjusting to alterations in environmental conditions such as increase in
temperature, lowering of pH or addition of exogenous organic solvents by the synthesis of
unusual fatty acids (1). These fatty acids were a family of onto—dicarboxylic acids ranging
from 28 to 36 carbons long. One of these fatty acids, 15,16—dimethyltriacotane-l,30-
dicarboxylic acid was isolated as its dimethyl ester and characterized in that study. The
chain lengths and relative abundance of very long dicarboxylic acids found in S. ventriculi
suggest they may be formed after the perturbation by combinations of existing regular
monofunctional fatty acids and not de nova by direct 2-carbon addition of acetyl Co-A.
One possible explanation is that these new, very long chain, bifunctional species are
formed by combination of existing lipid species by tail-to-tail coupling of the acyl chains
across opposite sides of the bilayer. Definite chemical proof for this mechanism was
obtained by analyzing the structures and stereochemistry of the very long bifunctional
species in the light of the structures of the regular monofunctional species. The exact
structures of fatty acid components were detemrined by various spectroscopic and chemical
methods including GC (Gas Chromatography), GC/MS (Gas Chromatography/Mass
Spectrometry), 1H NMR and 13C N MR (Nuclear Magnetic Resonance) spectroscopy, FTIR
(Fourier Transform Infrared) spectroscopy, polarimetry and reductive ozonolysis. These
analyses were performed on the methylester derivatives which were isolated after acid-
catalyzed methanolysis by a variety of chromatographic techniques. The proposed
mechanism was supported by precise structural and stereochemical information such as the

position of substitution of the acyl chain by methyl groups, position and configuration of

101
double bonds, analysis of chirality and by the identification of predicted co-formyl acids
among the bifunctional fatty acid species. This mechanism allows the rationalization of the
structures and identities of the various bifunctional, very long alkyl species elaborated by

S. ventriculi in response to environmental stress.

102

REFERENCES
l. Jung, 8., Lowe E.S., Hollingsworth I. R., and Zeikus, J. G. (1993) J. Biol. Chem.
268, 2828-2835
2. Lindquist, S. (1986) Ann. Rev. Biochem. 55, 1151-1191
3. Plesset, J ., Palm, C., and McLaughlin, CS. (1982) Biochem, Biophys. Res.
Commun. 108, 1340-1345
4. VanBogelen, R. A., Kelley, P. M., and Neidhardt, F. C. (1987) J. Bacteriol. 169,
26-32
5. deMendoza, D., and Cronan, J. E. Jr. (1983) Trends Biochem. Sci. 8, 49-52
6. deMendoza, D., Ulrich, A. K., and Cronan J. E. Jr. (1983) J. Biol. Chem. 258,
2098-2101
7. Silvius, J. R., and McElhaney, R. N. (1979) Chem. Phys. Lipids 24, 287-296
8. Silvius, J. R., and McElhaney, R. N. (1980) Chem. Phys. Lipids 26, 67-77
9. Kannenberg, E., Blume, A., McElhaney, R N. and Poralla, K. (1983) Biachim.
Biophys. Acta 733, 11 1-116
10. Langworthy, T. A. (1982) Curr. Top. Membr. Transp. 17, 45-77
11. Langworthy, T. A. (1977) Biochim. Biophys. Acta 487, 37-50
12. Lowe, S. B., Panlqatz, H. S., and Zeikus, J. G. (1989) J. bacterial. 171, 3775-3781
13. Lowe, S. E., and Zeikus, J. G. (1991) Arch. Microbiol. 155, 325-329
14. Tilak, K. V. B. R. (1970) Sci. Cult 36, 399-400
15. Knowles, W.S., and Thompson, Q. E. (1960) J. Org. Chem. 25, 1031-1033
16. Still, W.C., Kahn, M., and Mitra, A. (1978) J. Org. Chem. 43, 2923-2927
17. Shreve, O. D., and Heether, M. R., (1950) Analytical Chemistry 22, 1261-1264

CHAPTER IV

A MATHEMATICAL MODEL EXPLAINING THE MOLECULAR
WEIGHTS AND DISTRIBUTION OF VERY LONG CHAIN
DICARBOXYLIC ACIDS FORMED DURING THE ADAPTIVE RESPONSE
OF SARCINA VENTRICULI

103

104

INTRODUCTION

The membrane is one of the most critical structures in all biological systems. In
bacteria, it is the central structure across which the electric potential gradient which drives
the cellular process is maintained. Bacteria therefore must develop very sophisticated
mechanisms for repair, maintenance and regenerating of this important structures (3, 4, 5,
6, 7, 8).

In a recent study (1, 2), we described an adaptive mechanism in a strict anaerobic,
facultative acid0philic eubacteria, Sarcina ventriculi in which a family of very long chain
fatty acid species is synthesized in response to environmental stress (Figure 1). The size
and the distribution of these fatty acid species (Figure 1) indicates that a very sophisticated
mechanism which does not require de novo fatty acid synthesis must be at work. The
molecular weights and distributions of these unusual bifunctional species could not be
explained by simple 2-carbon additions of acetyl-CoA.

One possible mechanism for the synthesis of these unusual species is the chemical
coupling of lipid acyl chains across opposite sides of the bilayer leaflet by an enzymatic
activity which is located at the interface between the leaflets. This activity would then select
the ends of two fatty acid chains (independently of head group) one from either leaflet and
chemically couple them to form new lipids containing axe-dicarboxylic transmembrane
fatty acyl species. If there is no asymmetry in the distribution of the fatty acyl species
between the two leaflets, it should be possible to prove or disprove this theory by a
mathematical model. Here we propose a model where the regular length fatty acid chains
are treated as units which are selected for combination to form transmembrane acyl species

by a coupling entity.

105

 

 

 

 

 

 

 

 

 

 

E
8.
3 l l Car-Cu
a: , L]
8
g; H
0
Q
l l 1
Mill -
Retention Time

Figure 1. Gas chromatographic profile of methyl ester derivatives of total

fatty acidsof Sarcina ventriculi at pH 3.0 and 37°C. The family of peaks
labeled cu - C35 dicarboxylic acids are not synthesized at pH 7.0 at 379C in the absence

of exogenous organic solvents.

106

MODELS AND PROBABILITY CALCULATIONS

Model Construction

The random coupling model is shown in Figure 2. It shows an enzymatic activity
randomly selecting and coupling acyl chains from lipid species on opposite sides of the
membrane bilayer to form transmembrane lipid species. The randomness is ensured by the
rapid rotation (about the lipid long-axis) and lateral diffusion of the lipid species. Acyl
chain (Fatty acids) in each leaflet is represented by Pi and the other Fj (i or j=1, 2, 3..n

where n = total number of fatty acids species).

Definition of Parameters for Model

F1, F2,....Fi, Fjv-"Fn := Fatty acid species ( total number of species is equal to n)

N1, N2, ....Ni, Nj,....Nn := Number of each fatty acid (ex. Fi for Ni )

Fij := Transmembrane fatty acid synthesized by the ordered coupling of the fatty
acid Pi and Fj

P(Fij) := The probability of the formation of the transmembrane fatty acid Fij-

Fi:Fj := Transmembrane fatty acid made by the two fatty acids Pi and Fj (Fij or Fji)

P( FitFj ):= The probability of the formation of the transmembrane fatty acid Fisz

P(Fi:Fj) [%] := The relative percentage of P(Fi:Fj)

Calculation of P(Fij)
P0711) = (N 1C1*N 1C1) / (ENiCr*2NiCr) = (N1*N1) / (IINi)2
P(F12) = (N1C1*N2C1) / (zNiC1*2:NrC1) = (N1*N2) / (END2

P(Fij) = (NiCr*NjC1) / (ZNiC1*2NiCr) = (Ni*Nj) / (END2

107

Figure 2. Random coupling model for the synthesis of transmembrane
fatty acids. (A) Conceptual model showing an enzymatic activity randomly selecting
and coupling acyl chains from lipid species on opposite sides of the membrane bilayer to
form transmembrane lipid species. The randomness is ensured by the rapid rotation
(about the lipid long-axis) and lateral diffusion of the lipid species. (B) Model in which
the lipid chains in one leaflet is represented by Pi and the other Fj (i or j=l, 2, 3, ..n
where n = total number of fatty acids species). Fisz represents the general
transmembrane species formed from the ith acyl chain Pi and jth acyl chain Fj. Let
N i=number of units of the ith fatty acid, N j=number of units of the jth fatty acid,
P(Fi:Fj)=probability of forming the transmembrane species Fisz or szFi then P(Fi:Fj)=
P(Fi)+P(Fj) where P(Fij)=P(FJ-i); probability of Fi being connected to F5.

(A)

 

108

(B)

Connecting Box

 

 

 

 

 

 

 

 

 

 

 

 

 

 

BoxG) BOX“)

Figure 2

 

109

On the other hand,
P(Fi:Fj) = 2P1 * P(Fij) = 2*P(Fij)
The number (N i) of each fatty acid species can be expressed in terms of relative ratio to the
first fatty acid species (Dividing by N1)
N1=N1*f1, N2=N1*f2, N3=N1*f3,..,Ni=N1*fi,NJ-=N1*fj,. Nn=Nl"‘fn
IJNi=N1+N2-r-...+Ni~r-....+Nn
=N1*f1+N1*f2+ N1*f3+...+ N1*fi+....+ N1"‘fn
=N1*(fl+f2+f3+...+fi+....+fn)
= N1 * 2 fi
Therefore,
P(Fij) = (NiC1*NjC1)/ (ZNiC1*2NiC1) = (N i*N j) / (2N02 = (fi* fj )/ (302

P(F{:Fj) = 2P1 * PO25) = 2*P(Fij)=2*(fi* f} )/ (£1.02

Calculation of P(Fi: Fj) [ %]
P(F1=Fr) [%] == 100* {P(F1=Fr)l/{ ( P(F1=Fr) + P(F1=F2) +.-+P(F1=Fn) )+
(P(F2:F2) + P(F2:F3) +....+ P(F2:Fn) ) + ( P(F3:F3) +
P(F3:F4) +....+ P(F3:Fn) ) + ( ) + ( P(Fn:Fn) }
= 100 * 1 (f1* f1 )/ (>392 } / { (P(F1=Fr) + P(F1=F2) +--+P(F1=Fn) )+
(P(Fz:F2) + P(F2:F3) +....+ P(F2:Fn) ) + ( P(F3:F3) + .
P(F3:F4) +....+ P(F3:Fn) ) + ( ) + ( P(Fn:Fn) }
Here, P(Fl:F1) + P(F1:F2) +....+ P(F1:Fn) = f1*( f1 + 2* (f2+ f3 +...+fn )) l ()3pr
P(F2:F2) + P(F2:F3) +....+ P(F2:Fn) = f2*( f2 + 2* (f3+ f4 +...+fu )) / (Efi)2
P(F3:F3) + P(F3:F4) +....+ P(F3:Fn)= f3*( f3 + 2* ( f4+ f5 +...+fn )) / (ZZfi)2

110
Therefore,
P<F1:F1) [%] = 100*r(f1*f1>/(2f,>2}/{(>:f,2 + 2*:(fi *(zzfpn/(Zfozr

= 100*{(r1arm/(mymm,2 + 22a?i *(ij))}/(Zfi)2}
= rooming/{m} + 2*f, *(ij))}

Similarly,

may) [%]:1oo*2*<fi*f,-)/{z<f,2 + 2*f, «213-m

Here, Zfi = Zi=1..n fi E(fi *(ij)) = 2i=1..n (fi * (Zj=i+1..nfj ))

111

RESULTS

Random Coupling Model (RCM)

Figure 2 shows the conceptual model for this mechanism. Acyl chains are selected
with a probability based on their molar proportions. The molar proportions are deduced
from gas chromatography data. The idea is that the gas chromatography profile of the very
long chain fatty acids can be predicted from that of regular-chain fatty acids. The
membrane model in Figure 2A in which the enzymatic coupling activity randomly selecting
lipid species is replaced by the connecting box in Figure 2B. Box(i) replaces one
membrane leaflet and box(j) the other. Fi and Fj are fatty acid chains which are selected to
form the new transmembrane fatty acid FiiFj. The probability of selecting any given fatty
acid species is proportional to the intensity of its peak in the gas chromatogram after
correcting for chain length differences by dividing the area of the peak by the molecular
weight of the peak to give molar proportions. This is accurate since the flame ionization
detector is a mass detector and the detector response for fatty acids of similar chain lengths
is the same. The probability, P(Fi:Fj), of forming the transmembrane fatty acid Fisz (or
szFi) is equal to 2*(Ni*Nj)/{(2Ni)*(ZNj)}. Here the number of units or moles of ith
fatty acid species (N i) is normalized to the number of units (moles) in the first eluting peak
in the gas chromatogram (N1) and is expressed as the real number (fi) times the number of
units of the lst fatty acid species (i.e. Ni=fi*N1, and f1=1). Then 2Ni=N1+ N2 + N3 +
= N1*f1 + N1*f2 +N1*f3 +...= N1*(f1 + f2+..) = N 1* (Zfi). Since the molecular
weight identities and relative molar proportions of the regular-chain length fatty acids are
known, and if we assume that they are distributed without bias between the leaflets of the
bilayer and combine randomly between them, we can then predict the molecular weight of

the transmembrane fatty acid species and the relative abundance (%). In other words,

P(Fi:Fj) becomes 2*(fi*fj)/[(2fi)*(2fj)}. The initial abundance by gas chromatography

112
of each regular fatty acid species (g1, g2, g3,... gn) is converted to f 1, f2, f3,...fn by
dividing initial abundance values by that of the first fatty acid species (f1=g1/g1,
f2=ggjg1,...fn=gn/g1). Each fatty acid species such as Fi, Fj or Fi:Fj in the case of new
transmembrane species, was identified by its molecular weight. These were obtained by
Gas Chromatography/Mass Spectrometry and there was no redundancy between weights of
regular chain fatty acids. Table 1 shows a combination grid in which all of the pair-wise

combinations of normal fatty acid species (Fn) to form transmembrane species (Fi:Fj) are

examined and their likelihood (probability) calculated

Comparison of ‘Predicted Data’ with ‘Observed Data’ for Transmembrane
Fatty Acids

In Figure 3A the results of Table l are presented in a graphical fashion. In Figure
3B the measured distribution and chain lengths (molecular weights) of the new fatty acid
species are compared with the theoretically calculated values on a mass axis. There is one
instance of redundancy in the molecular weights of transmembrane species; the
combination of two 16 carbon fatty acyl group gives a transmembrane species of the same
molecular weight as the combination of a 14 carbon and an 18 carbon acyl group. It is very
clear that the measured and calculated profiles are practically identical. This proves that the
new, very-long chain fatty acids are formed by the random combination of existing fatty
acid chains by the tail-to-tail coupling. The assertion that the coupling occurs between
fatty acid chains in opposing leaflets and not in the same leaflet is borne out by freeze-
fracture electron microscopy which does not show any instances of cleavage between the
bilayer leaflets in cells containing these new very-long chain dicarboxylic acids. The result
identifies this adaptive process which is characterized by a high degree of speed and

efficiency by a hitherto unreported mechanism. -

113

Table 1. Mathematical formula for the prediction of probability for the
formation of transmembrane fatty acids. (A) The relative abundance (%) of
predicted transmembrane fatty acid species (Fisz) from the initial ratios of the abundance of
regular-length fatty acid species (f1, f2...fn). P(Fi:Fj) (%) = lOO*P(Fi:Fj)/ {(2P(Fi:FJ-)}
where P(Fi:Fj)= {2*(fi*fj)}/ [(Zfi)*(2fj)], ZP(Fi:FJ-)= 22(fi2 + 2*fi*(2fj))= Emma? +
2*fi*(£j=i+1. n fj)) (B) The calculated relative abundance (%) of transmembrane fatty acid
species from the experimental initial ratios of the abundance of regular-length fatty acid
Species. 'Ihese transmembrane fatty acids were formed on changing the pH of the medium
from 7.0 to 3.0. Bolded numbers indicated the molecular weights of regular chain fatty
acids (ie. 242:C14 fatty acids, 298: C13 fatty acids). The numbers in parenthesis showed

the predicted molecular weights of transmembrane fatty acids.

(A)

114

 

I P(Fi:Fj)(%)

F.

F.

F

 

IF'

 

roo - «(1‘) taxi} - 2 -

t, - 0:5)»

100 - 211,") and} + 2
‘ 1:011)»

100 - 2-(1; r“) and" +2
4,111,)»

 

 

 

025m

[00°(li'15)l(£(f‘2¢2-f‘.°

roo- 2~(r‘-r.)I(r:(ti2-2
4301)»

 

Table l

 

 

roo- (quanta-2
4:01)»

 

 

 

115

Figure 3. Prediction of the distribution of transmembrane fatty acids based
on random coupling model. (A) Calculated distribution of transmembrane fatty acid
species from the grid in Table 1. Each value of fi was determined by analytical gas
chromatography. (B) Comparison of the distribution of the observed peaks in gas
chromatogram with that of the predicted ones by the random coupling model calculation.

 

116

(A)

 

25

 

N
O

 

-
M

 

—
O

 

Relative abundances (%)
l
I

M
“

—

 

 

 

 

 

 

 

 

 

11 ll 41 I

0 V f fi__ V t i V i V ‘
482 493 504 $15 $26 537 548 559 $70 581 592

 

 

 

Molecular Weight
(8)
l l l

i

r?

a:

'1 .4

ég 1 l cflmservexllmeakar

LL

 

 

 

 

 

 

 

simulated peaks ——->

 

Figure 3

117

DISCUSSION
In this work, a random coupling mechanism of establishing the unique distribution
and the relative abundance for a new family of transmembrane fatty acid species is
introduced. The accurate coincidence with the experimental results provides the conclusive
proof for the existence of this new synthetic pathway in S. ventriculi. Due to this
coupling mechanism, the empty spacing between the C14-C13 and C23-C35 region in the
GC profile in the distribution of fatty acids (Figure 1) can be explained and successfully

predicted as well as the structural uniqueness of onto-dicarboxylic acid species. The

generality of this mechanism in eubacteria is still unexplored.

118

CONCLUSIONS

A simple mathematical model is presented to explain a recent new discovery of an
unusual membrane adaptive response in Sarcina ventriculi . In this response, this organism
synthesizes very long chain 0t, co-dicarboxylic acids ranging from 28 to 36 carbon atoms in
length. The distribution of chain lengths of the new fatty acid species is not consistent
with de novo synthesis but suggests elaboration from the existing regular-chain fatty acids
by a coupling process. Here, it is demonstrated, using a mathematical model, that if the
molecular weights and relative abundance of regular chain fatty acids are known, then the
molecular weights and relative abundance of the new, very long chain dicarboxylic fatty
acid species can be predicted using a model based on the random, pairwise combination of
regular chain species. This combination takes place across the bilayer leaflet to form
transmembrane fatty acids. It is proposed that this coupling phenomenon is regulated by

the motional dynamics of the membrane.

119

REFERENCES

1. Jung, 8., Lowe, S. B., Hollingsworth, R.I., & Zeikus, G. J. (1993) J. Biol. Chem.
268, 2828.

2. Jung, 8., and Hollingsworth, RI. (1993) J. Biol. Chem. (submitted)
3. Andreoli, T. (1974) Ann N.Y. Acad. Sci. 235, 448.
4. Cronan, J. J. Biol. Chem. (1975) 240, 7074.
5. Hollingsworth, R & Carlson, R. W. (1989) J. Biol. Chem. 264, 9300.
6. deMendoza, D. & Cronan, J. (1983) Trends Biochem. Sci. 8, 49.
7. Sinensky, M. (1971) J. Bacterial. 106, 449.
8. Teuber, M & Bader, J. (1976) J. Arch. Microbiol. 109, 51.

CHAPTER V
A NEW FAMILY OF MONOGLUCOSYLGLYCERIDE DIACYL

GLYCEROL LIPIDS CONTAINING VERY LONG CHAIN
BIFUNCTIONAL ACYL CHAINS IN SARCINA VENTRICULI

120

121

INTRODUCTION

Membrane lipids are known to play various important roles in the physiological
functions of bacteria. These roles include the communication of external information,
compartmentalization between the cell and external medium and a matrix for localization
and proper functioning of membrane proteins. The membrane serves as a control center
(1, 2, 3) for theregulation of the motional dynamics of the membrane proteins. From the
standpoint of signal transduction (4, 5), the structural studies of membrane lipids and the
induction of the synthesis of unusual structures have their special importance. Recent
studies on the adaptive processes in Sarcina ventriculi shows that very long chain
bifunctional acyl components formed in membranes during stress can function in regulating
the membrane dynamics (6, 7). The mode of formation of these unusual lipid species in S.
ventriculi appears to be by the tail-to-tail condensation of alkyl chains across the two
leaflets of the membrane bilayer. Although the structures of the isolated fatty acyl species
support this idea, no information on the structures of the intact lipid species is available.
The focus of this study, therefore, was to isolate and characterize intact lipid species from
the membrane of S. ventriculi (11,12). Special emphasis was placed on the isolation and
characterization of membrane components which contained the unusual very long chain,
bifunctional fatty acyl species. Information on the structures of such lipid components is
critical for our understanding of the architecture of the membrane as well as our

appreciation for the possible modes of synthesis of these unusual but physiologically

important components.

122

MATERIALS AND METHODS

Culture of Cells
S. ventriculi was cultured at pH 3.0 in liquid medium as described before (6).

Membrane Preparations

Cells were disrupted by passage through a French Pressure cell (American
Instruments Co., Inc., Silver Spring, Md.) at 20,000 lb/in2. The disrupted cells were
centrifuged at 20,000 x g to remove unbroken cells, and the supernatant was centrifuged at
110,000 x g to sediment the membranes, which were washed twice with distilled water by

suspending and recenuifugation.

Extraction of Lipids

Lipids were extracted from the isolated membrane or whole cells using procedures
(a) or (b) respectively. (a) To each 5-10 ml membrane suspension was added 30 vol. of
chloroform/methanol (5:1,v/v). The suspension was then mixed to produce a single phase.
The mixture was shaken or stirred vigorously at 45°C, with intermittent sonication (for
approximately 5 min every 30 min), over a total of 2 hours and then taken to dryness on a
rotary evaporator. The residue was partitioned between 10 ml chloroform/methanol
(5:1,v/v) and 2.5 ml water. The lower organic phase was taken to dryness and
redissolved in 1 ml chloroform/methanol (9:1,v/v). (b) Cells of S. ventriculi from 50
liters of culture medium were harvested by centrifugation at 10,000 x g for 10 min. Lipids
from approximately 50 g wet weight of cells were extracted at 45°C with 400 ml of a
mixture of chloroform/methanol/water (15:3:2, by vol.) for 2 hours, followed by 200 ml
of chloroform/methanol (5:1,v/v). Extraction was performed with intermittent sonication

over 2 hours or described in procedure (a). After centrifugation at 20,000 x g, the pellet

123
(cell debris) was extracted again with the same solvent system. After centrifugation, the
supernatant was taken to dryness on a rotary evaporator, dissolved in 10 ml of
chloroform/methanol (5:1,v/v) and then shaken with 2.5 ml water. The lower phase

containing the lipids was taken to dryness and the residue dissolved in 1 ml of chloroform.

Analysis of Lipids

Membrane lipids were separated by 2-dimensional TLC using chloroform] methanol/
ammonia! water ( 3.3:1.0:0.l: 0.05, by vol.) for the first dimension and chloroform
[methanol / water (7:1.6:O.2, by vol.) for the second dimension. Analyses were
performed on silica— gel plates (Merck). Spots were made visible either by spraying with
50% ethanolic-sulfuric acid and heating at 250°C to char the organic components, or by
spraying with a 0.1% solution of 2’,7,-dichlorofluorescein in aqueous ethanol ( 1:1) and
viewing under ultraviolet light (8). Standard phospholipids such as phosphatidylcholine
(PC), phosphatidylethanolamine (PE), phosphatidylserine (PS) and cardiolipin (CL) and
neutral lipids (NL) were used as standards in addition to free fatty acids (FA). Spraying
agents (9) for the detection of components included ninhydrin for PE or PS, dragendorff
agent for PC, orcinol for glycolipid, and molybdenum blue for phosphate. Each lipid band
was scraped from the plate into a column fitted with a sintered disc and the material was
eluted from the silca gel with the mixtures of methanol and chloroform. Each fraction was
concentrated by evaporation and redissolved in chloroform or a choloroform [methanol
(5/1) mixture for further analysis. One of the orcinol-positive lipid bands (A(a), see
reference 1) containing the bifunctional acyl chains (by GC/MS analysis) was chosen for

further study.

Head Group Analyses
Isolated lipids were hydrolyzed with 2 M TFA (trifluoroacetic acid) at 120°C for 3
hours. The hydrolysate was concentrated to dryness, and then 1 ml of water was added and

124
the solution concentrated to dryness to remove trace amounts of TFA. The hydrolysate
was further analyzed by proton NMR (Nuclear magnetic Resonance) spectroscopy or by
conversion to alditol acetates for GC/MS analysis. For GC/MS analysis using alditol
acetates, the TFA hydrolysate was extracted with 2 volumes of 1 ml chloroform. The
chloroform extracts were discarded and the aqueous layer was concentrated to dryness
under niu'ogen. The aqueous residue was dissolved in methanol and reduced with sodium
borohydride for 1 hr. 3 M HCl was then added to decompose excess borohydride. The
solution was repeatedly concentrated to dryness several times from methanol. To the
residue was added 0.1 ml pyridine and 0.1 ml acetic anhydride. The solution was briefly
sonicated and left at room temperature for 16 hrs. The mixture was concentrated to dryness
and then 1 ml chloroform and 1 ml 3M HCl were added to the residue. The chloroform
layer was washed once with 1 ml 0.5 M NaCl, dried over anhydrous sodium sulfate, and
subjected to CC analysis on a DB225 capillary column with an initial temperature of
200°C, hold time 5.00 min., rate of 2.0 deg/min., final temperature of 230°C, final hold
time of 55 min., and total run time of 75 min. Retention times were compared with alditol
acetate derivatives of a number of alditol acetate standards. GC/MS analyses were then

performed on a JEOL JMA DASOOO mass spectrometer using a D3225 capillary column.

Fatty Acids Analysis

Fatty acid analyses were performed on the isolated lipids by treatment with
methanolic HCl using either of two procedures. Three (3) ml of chloroform was added to
1 ml of lipids suspension followed by 15 ml 5% methanolic-HCI solution. The flask was
sealed and heated in an oven at 72°C for 12 hours. 3 ml of chloroform was added every 6
hours followed by mild sonication for 5 minutes. The mixture was then concentrated on
the rotary evaporator to dryness and extracted-with chloroform. The combined organic
fraction was redissolved in 1 ml of hexane. The fatty acid methyl esters prepared by above
procedure were subjected to Gas Chromatography analysis on a 25 M J &W Scientific DB1

125
capillary column using helium as the carrier gas and a temperature program of 150°C initial
temperature, 0.00 min hold time and 3.0 deg/min rate, to a temperature of 200°C. A
second ramp of 4.0 deg/min was then immediately started until the final temperature of
300°C was obtained. This temperature was held for 30 min. The relative proportion of
lipid components were calculated from the integrated peak areas. The fatty acid
identification and molecular weight were determined using GC/MS analysis using a JEOL
JMS-AXSOSH spectrometer interfaced with a Hewlett-Packard 5 890A Gas

Chromatograph.

FAB (Fast Atom Bombardment) Mass Spectrometry

FAB-MS was performed on JEOL HXlOO (Peabody, MA) double-focusing mass
spectrometer (EB configuration) equipped with a high-field magnet operated in the positive
ion mode. Ions were produced by bombardment with a beam of Xe atoms (6kv). The
accelerating voltage was 10 kv, and the resolution was set at 1000 or 3000 according to the
mass range of interest. The samples were dissolved in chloroform solution. Generally , 1
- 1.5111 of sample was mixed with lul of NBA (Nitro-Benzyl Alcohol) on the FAB-MS
stainless steel probe tip. Calibration was performed using Ultramark (443 or 6121) or
(CsI)nI' cluster ions, depending on the mass range of interest. A JEOL DA-SOOO data
system was used for recording the spectra. The spectrum was scanned over 2 min from

m/z 0 -3000. Data presented were obtained in a single scan and found to reproducible.

1H NMR Spectroscopy
Proton NMR spectra were recorded at 300 MHz and 500MHz on solutions in
CDC13 or D20. Chemical shifts are quoted relative to the solvent resonances taken at

7.24 ppm for chloroform and at 4.65 ppm for water.

126

Fourier Transform Infrared Spectroscopy
Spectra were obtained with a Nicolet model 710 FT-IR spectrometer on a 10%

(w/v) solution of isolated lipids in chloroform.

127

RESULTS AND DISCUSSION

Structural Characterization of the Isolated Glycolipids

One of the orcinol-positive glycolipid fractions on 2—dimensional TLC was isolated
by preparative thin layer chromatography. The total ion chromatogram from the GC/MS
analysis of the fatty acid methyl esters obtained from this fraction is shown in Figure l.
The peaks from a to e are due to typical membrane fatty acyl components ranging from 14
to 18 carbons. The major regular fatty acid methyl esters are hexadecanoic acid methyl
ester (peak c, C16:0)- cis-vaccenic acid methyl ester ( peak d, C13;1(11)) and stearic acid
methyl ester (peak e, C13;o). The peaks from f to to, however, correspond to very long
bifunctional fatty acyl components containing from 28 to 36 carbon atoms. The major
bifunctional acyl species were identified as 0t,0)-15,16-dimethyltricotanedioic acid dimethyl
ester (peak j), a,a)-(17,l8-dimethyl)-cis-l l-dotriacotaenedioic acid dimethyl ester (peak I)
and 0t,(o-(l7,l8-dimethyl)-cis-11-23-hentriacotadienedioic acid dimethyl ester (peal m). A
detailed structural characterization of these components has been presented earlier (1, 2).
Another unusual, very long chain component (peak k) was identified on an to—formyl
methyl ester. Figure 2 shows the El (electron impact) mass spectrum (70ev) of this
molecule. Major ions appeared at m/z 534, 502 and 474. These correspond to the
molecular ion (M+) with the sequential losses of methanol (CH3OH) and ethylene
(CH2=CH2) or carbon monoxide, respectively. The ions at m/z 516 and 484 are due to
the loss of water from the ions at m/z 534 and 502 respectively. Primary fragmentation of
the alkyl chains between the vicinal methyl groups (Figure 3) give rise to the fragments at
m/z 295 and 239. The ions at m/z 264 and 221 correspond to the loss of methoxy group
(CH3O) from the ion at m/z 295 and to the loss of water from the ion at m/z 239,

respectively. The characteristic McLafferty fragment of aliphatic methyl ester appears at

128

Figure 1. Total ion chromatogram of Gas Chromatography/Mass Spectrom-
etry analysis for the esterified fatty acyl components of one of the isolated

glycolipids. The later eluting cluster of peaks is due to the membrane spanning (1,0)-

bifunctional fatty acids.

C1620 fatty aldehyde (OHC(CH2)14CH3)

C141) carboxylic acid methyl ester (OCH3CO(CHZ)12CH3)
C171) carboxylic acid methyl ester (OCH3CO(CI-12)15CH3)
C133 carboxylic acid methyl ester (OCH3CO(CI~12)9CH=CH(CH2)5CH3)
C131) carboxylic acid methyl ester (OCH3CO(CHz)15CI-I3)
Unknown

C313 0)-formylmethyl ester (M.W. = 490)

C33 m—formylmethyl ester (M.W. =520)

C303 onto-dicarboxylic dimethyl ester (M.W. = 508)

C321) (LCD-dicarboxylic dimethyl ester (M.W.= 538)

C343 (0—formylmethyl ester (M.W.: 534)

C343 onto-dicarboxylic dimethyl ester (M.W.: 564)

C353 axe-dicarboxylic dimethyl ester (M.W = 590)

flax 8055.46

announce-:0

 

129

2‘0 39 49 A 543 «mg . abund .

 

1 A A ‘AA—LJ‘J ‘

 

 

 

V v v i V v V ‘ fifi

 

1506 7 2636 Tasha 9600
SC an

Figure l

GODDQDCFD fl<—-r‘0—flm

130

 

166

 

 

 

 

 

 

 

 

 

 

 

 

 

5 3
. 97 502
‘55
80‘
5:34
‘ 09 ‘
‘ l.
63 266 .
l
* l
40- 2 2
. 1 7
l
20‘ l '11 z z ‘ 3 4 474
l l
; , l * . . 4
0 l ’ l .‘ fi , . pi -' - -. l I . l i l . AiLAr‘A‘ j‘g‘ Li ‘4 ‘ v V
100 280 300 400 508 608 700
"/2

Figure 2. Electron impact rrrass spectrum of peak R. It contained major ions at
m/z 534, 502 and 474. These correspond to the molecular ion (1%") with the sequential
hanesrfihnxkhanolflCIkfiIHDzuuledndene(KJHQ=(3H2)orcarbonrnonoxhha nuqxxxhnfly.

131

r ————————————————————— >-

- -28
323 ———>—31 292 —>—254

I
I
I
‘ a
: ~ , ............... «>-
I 12 5 -2 ‘32
, , ———> 293 ——>'261
I -
. ,/ L931 264
I

o 'C I 0
I I

, __
I, CH3
I
- -2
1944-13- 237+——| /
I

-28
217<——239 ,’ -18

I
I
I
I
I
I
..(. ———————————— I 516
E -18 I ‘
221 '
: -28 -32
l
I
I
I

4 74 +—<_ __5_02 g +— ____ _5_3_4(M.+)

-18
249*—] ,
~44 -1 484 18
a22<———256<———267 _,

Figure 3. Mass fragmentation pattern of peak k (to-formyl-(l7,18-dimethyl)-
cis-ll-hentriacotanemethyl ester). lzhydrogen, 2: hydrogen molecule (Hz), 18:
water (HzO), 28: ethylene (CH2=CH2), 31: methoxy (CH3O), 32: methanol (CH3OH),
43: CH2=CH-O, 44: CHz=CH-OI-I ‘

132
m/z 74. The other major fragments are shown in Figure 3. The structure of peak k was,
therefore, determined to be m-formyl-(17,18-dimethyl)-cis-1l-hentriacotanemethyl ester
(Figure 3). The existence of (o-formylmethyl ester after methanolysis suggested the
presence of the vinyl ether (-CH=CH-OC-) functional groups in the parent lipid. This
functionality defines the lipid class as a plasmalogen. The molar ratio of the regular length
(Cu-C20) fatty acids to very long chain bifunctional acyl chains was approximately 2:1.
The variety of fatty acid species indicates that, although a single discrete spot is obtained on
analysis of this lipid fraction, there is still considerable heterogeneity with respect to alkyl
chain length and even type. This heterogeneity is quite normal and an important aspect of

membrane structure.

NMR Analyses of Glycolipids

The Proton NMR spectrum of the glycolipid fraction is shown in Figure 4. The
intense resonance at 5 1.27 was assigned to the methylene groups of the lipid chain. The
multiplet at 5 1.58 were assigned to the methylene protons B to the carbonyl group. The
resonances at 5 2.28 (t, J=7.69 Hz) were assigned to the methylene groups a to the
carbonyl function. The characteristic 3H-doublet at 5 0.74 (J=7.20 Hz) confirmed the
presence of vicinal methyl groups of the 00-1 linked bifunctional acyl chains (6,7). A
triplet at 5 5.36 (J=6.92 Hz) was assigned to the vinyl protons of the unsaturated fatty
acids. The multiplet at 5 2.00 were assigned to the protons of the methylene groups a. to
the vinyl carbons. Multiple peaks in the range of 5 3.00 to 5 5.00 strongly indicated the
presence of glycerol and carbohydrate functions. This was confirmed by further structural

studies.

Head Group Analyses of the Glycolipids
Figure 5 shows the gas chromatographic profile of alditol acetates obtained from the
hydrolysate of the lipid fraction by treatment with 2M TFA. There were two major

133

Figure 4. 1H NMR spectrum of the glycolipids containing bifunctional acyl
chains. Signals at 0.74 ppm ((1, J=7.20Hz) and 1.27 ppm are characteristic of the methyl
and methylene groups of long chain acyl components. Resonance at 2.28 ppm (t,
J=7.69Hz) represented methylene groups a to the carbonyl function. The multiplet at 1.58
ppm was assigned to the protons of the B carbons of the molecule. The triplet at 5 5.36
(J=6.92 Hz) was assigned to the vinyl protons. The multiplet at 5 2.00 was assigned to the
protons of the methylene groups a to the vinyl carbons. Peaks in the range of 5 3.00 to
5 5 .00 strongly indicate the presence of glycerol and carbohydrate functions. The signal at

5 7.24 was assigned to the chloroform.

134

 

14d1ddd—‘dfldd—u‘l‘dd-dddd-1“d-Jddd-qddd‘ddfldqflddl.

uh uh 0.0 ..u 90 «.0 ab ~.u col

 

 

Figure 4

 

135

 

 

 

 

FL.“ L . 1.1L - ALL -- -

IVORY

Figure 5. Gas chromatographic profile of the alditol acetates of hydroly-
sates obtained from the 2 M TFA (Trifluoroacetic acid) hydrolysis of the

lipids. A; glycerol triacetates, B; glucitol hexaacetates

136

components present. Peak B(the later peak) had the same retention time of standard
glucitol hexaacetate. The presence of this alditol acetate was confirmed by mass
spectrometry. The electron impact mass spectrum (70ev) of peak B is shown in Figure
6A. Typically, alditol acetate do not give molecular ions, but (M-CH3C02)+ is found in
low abundance by elimination of an acetoxyl group, or by cleavage of the alditol chain, as
shown in Figure 6C. In this way, m/z 375 (M-CH3C02)+ and five other primary fragments
are formed from glucitol hexaacetates (Figure 6B). Figure 7 shows the EI mass spectrum
of peak A. It contains the major ions at m/z 159, 145 and 103. They corresponded to
(M— CH3C02)+, (M— CHzOCOCH3)+ and the loss of ketene (-CH2CO) from m/z 145,
respectively. The presence of these ions was indicative of the presence of glycerol
triacetate. These data indicated that glucose and glycerol are the components of the
headgroup. This fact was, further, confirmed by the analyses of the proton NMR
spectrum of hydrolysates obtained from TFA hydrolysis (Figure 8). The, oc-glucose
anomer was conf'umed by the characteristic coupling constant of the H-1 proton at 5 4.45
(d, J = 4.00Hz). Characteristic signals of glycerol appeared at 5 3.50 (dd J=4.2 Hz) and 5
3.41 (dd J=4.2 Hz). These data for the glycolipids strongly indicated that they are glucosyl
diglycerides that contain the bifunctional acyl chains. Further evidence on the ether linkage
between glucose and glycerol was obtained by the Fourier Transform Infrared

spectroscopic analysis.

FTIR Analysis of Glucolipids

The Infrared spectrum (Figure 9) showed a strong aliphatic C-H asymmetric
stretching absorption at 2918 cm-1 and symmetric stretching at 2856 cm'l. The
characteristic alkene stretching (=C-H) was observed at 3038 cm-1. Broad -OH stretching
band (3600 -3000 cm“) and hydroxyl stretching vibration at 3411 cm-1 were also present.
The C-O stretching vibration in the alcohol produced the strong band at 1166 cm'l. A
band due to the bending vibration of aliphatic C-H bonds in the methylene groups appeared

137

Figure 6. Mass spectral analysis of peak B. (A) Electron impact mass Spectrum
(70ev) of peak B. (B) Ionic fragmentation pattern of the glucitol hexaacetates. Note the
presence of ions at m/z 375 (MCI-13002)+ and masses corresponding to five other primary
fragments. (C) (M-CH3C02)+ is produced by elimination of an acetoxyl group, or by
cleavage of the alditol chain

OOQOQDCD’D

138

 

160‘
60':
40:

28‘

 

QL.-

 

 

 

 

 

 

 

 

 

 

 

 

 

1 s
187
145 ’
157 b
179 217 .
103 '
259 “
279 1
as ~1
59 if
#41. L1 iv#fi.f,ai. ' - r
we zoo 369 see sea
"/2
73 CHZOAc
---------- run-"as.- C '- +
1.. u——0Ac "gm
""""""""""" HOOAc

 

Figure 6

NODDQDCU’D 0(~«PD-—flm

139

 

 

 

 

 

 

 

 

 

 

106‘ 133 1 S
88" 86 73 CHzoAC
.------- """"""12§"
‘ H OAc
60- 73 145
j 5‘ CHzoAc 73
‘ 111.6
4o~ d
1*320v ‘
za
1 1T;
1 111+ - - g - - ‘ - - - g
180 209 see 406 533 sea

"/2

Figure 7. EI mass spectrum of peak A . It contains major ions at m/z 159, 145
and 103. These correspond to the (M- CH3C02)+ , (M- CI-IzOCOCII-I3)+ and loss of
ketene (~CHzCO) from m/z 145, respectively.

140

Figure 8. 1H NMR spectrum of the hydrolysates obtained from the TFA
hydrolysis on glycolipids. It shows the anomeric proton of glucose at 5 4.45 (d, J =
4.00Hz). Characteristic signals due to glycerol appear at 5 3.50 (dd J =4.2 Hz) and 5 3.41

(dd J :42 Hz). Chemical shifts are quoted relative to the water resonance at 4.65 ppm for

proton.

 

 

 

 

 

 

141

 

4 .

4.

4.

4

4.0

Figure 8

3.8

142

Figure 9. Fourier Transform Infrared spectrum of the glucolipids. Note the
presence of aliphatic C-H asymmetric stretching absorption at 2918 cm'land symmetric
stretching at 2856 cm“. The characteristic alkene stretching (=C-H) was observed at 3038
cm“. Broad -OH stretching band (3600-3000 cm“) and hydroxyl stretching vibration at
3411 cm-1 were shown. The characteristic C=O absorption bond of the aliphatic ester

group appeared at 1733 cm" and the ester alkoxy stretch at 1263 cm*‘. The strong
I asymmetrical C—O-C stretching band appeared at 1099 crrr‘.

XTronemxttance

51.070 71.422 91.775 112.13

30.717

 

143

 

 

 

 

 

4126.8

3733.1

1
I
3339.3

1
2945.6

5551.9 5:53.: 41734.4
Havonumbor (cm-1)

Figure 9

3370.7 573.9: 553.2:

C

144
at 1465cm'l (scissoring) and one due to the twisting and wagging deformation at 1340 cm-
1. The characteristic C=C absorption bond of the aliphatic ester group appeared at 1733
cm:1 and the ester alkoxy stretch at 1263 cm'l. The stereochemistry (cis- or trans-) of the
unsaturated carbons was determined by the clear presence of a strong peak at 660 cm'1 due
to the C=C-H bending deformation for cis- alkenes (10). The characteristic response of
ethers in the IR is associated with the stretching vibration of the C-O-C system. The
strong asymmetrical C-O-C stretching band appeared at 1099 cm'l. The C=C stretching
band of vinyl ethers appeared at 1623 cm'1 and 1636 cm‘1 as a doublet because of the

presence of the rotational isomers.

FAB(Fast Atom Bombardment) Mass Spectromety Analysis of the
Glucolipids

Positive FAB mass spectra of the glucolipids were obtained to determine the range
of molecular weight of the lipids and the number of attached glucose groups (Figure 10).
Due to the variety of fatty acyl components a cluster of molecular ions ([M+Na+H]+) was
observed. The spectra contained two clusters separated by 162 a.m.u. The lower mass
cluster was derived from the higher mass cluster by the loss of one glucosyl residue.
Figure 11 shows the two clusters at the high mass end of the spectrum. The numbers of
the ionic clusters ([M+Na+H]+) indicated various kinds of monoglucosyl lipids. Each
ionic cluster numbered in Figure 11 could be predicted from the general structure proposed
in Figure 12. This structural formula is based on the compositional data of the lipids, in
which the fatty acyl chains and head group components were separately quantitated and
analyzed by various spectrosc0pic methods. There is an approximate 2:1 molar ratio
between the regular length acyl chain and membrane spanning bifunctional acyl chains.
Based on the fatty acyl components present, the distribution and molecular weights of the
peaks in the clusters could be predicted using an unbiased, statistically weighted

combination algorithm. Table 1 shows the calculated m/z values of two ionic clusters.

145

 

 

 

 

 

 

 

 

1
R .
e I
‘ I
am
t I
i .
U 4
260
a 4
b .
U40
:1 .
d
a
n f
C
e
23»
“/2

 

Figure 10. Positive FAB-Mass Spectrum of the glucolipids. Note the cluster

of molecular ions [M+Na]+. The other lower mass cluster is due to the loss of one glucose

group ([M+Na - 162 ]+).

146

Figure 11. Positive FAB-Mass spectrum of two ion clusters ([M+Na]+,
[M+Na - 162 1+) separated by 162 mass units. It continued the presence of one
glucose molecule. Each numbered ionic cluster was predicted by the calculated m/z values

based on the compositional analyses. Table 1 showed the calculated ionic clusters from the

basic frame in Figure 12.

147

mosmazcc> (D<"""‘°’ID—(Djj

 

 

 

127.123-322.33‘ - .Nmmv

lhll 2.423.. 3 wow - de IIJ

 

 

 

 

 

Figure 11

 

 

 

148

Table l. The calculated m/z vaues of ionic clusters based on the
compositional analyses. MSFC; Membrane Spanning Fatty Acyl Components, Fy 1:
Fatty acyl component 1, Fy 2; Fatty acyl componet 2. Small alphabets (a, b... l, m)
indicated the peak numbers in the GC/MS analyses (Figure 3). The asterisked small
number in each column point out the peak number (1, 2, .., 9) of the ion fragments
observed in FAB-MS. Almost all of the observed ionic clusters were predicted with the
proposed lipid structures based on the structural model shown in Figure 12.

149

 

.2»

23.428; 23+

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9528+
35: 2:53 25.: 258 £58 2353 :58 258 £58
293 2.33 2.33 2933 2:33 2.33 233 £3.33 2:33
$3.28 :5: :5: :58 :2: :55”; :58 :88 :38
293 2.33 2.33 2933 2:33 .233 .233 ..2333 ._ 2:33
$3.23 :58 :58 :5: :58 :5: :88 :58 :5:
293 2.33 2.33 2933 2:33 .. 2.38 .. 2.33 a 2338 .. 2:33
$3.53 :58 :5: :5: :58 :88 :58 :58 :58
293 2.33 2.38 2938 2:38 2.38 ..238 a 2338 a 2:38
$5.58 :5: :58 :58 :58 :88 :58 :5: :58
293 2.98 2.98 2393 2:98 2.98 a 2.98 a 2398 a 2:98
$5.28 :5: :58 :58 :5: :88 :58 :5: :58
93 2.98 2.98 2393 2:98 a 2.98 42.98 .. 2393 .. 2:98
$5.58 :58 :58 :58 :58 :58 :5: :58 :58
2.3 2.98 2.98 2398 2:98 a 2.98 a 2.98 .. 2398 .. 2:98
$5.53 :58 :8: :5: :58 :58 :58 :58 :58
3.3 2.93 2.93 2993 2:93 .. 2.93 .. 2.93 2393 .. 2:93
$5.53 :58 :58 :5: :88 :5: :58 :58 :5:
293 2.98 2.98 3.93 2:93 a 2.98 .. 2.98 2398 .. 2:98
$5.58 :8: :5: :58 :5: :58 :58 :58 :58
5.8 2.98 2.98 3.98 2:98 .. 2.98 .. 2.98 23.98 .. 2:98
$5.58 :88 :58 :58 :E :58 :3: :58 25E].

 

 

 

 

 

 

Table 1

150

Figure 12. Basic model structure of the monoglucosyldiacylglyceride diacyl
glycerol in S. ventriculi. This frame was constructed based on the compositional
analyses of the glucolipids. Letters point out the peaks in Figure l. The numbers in
parenthesis indicated the molecular weights of fatty acyl components obtained from the
methanolysis of the glucolipids. MSFC; Membrane Spanning Fatty Acyl Components , Fy
l; a regular fatty acyl component

Fy 2; another. regular fatty acyl component

151

 

 

('3H2 - '. MSFC “(‘3'12
CH - Fy 1 FY 2 ‘CH

l l

CH2- Glucose OH-CH2

 

 

 

 

 

 

 

 

 

 

MSFC: 1(533) =>OCO-(CH2)13-(CH3)CH-CH(CH3)-(CH2)13-0CO
l(564) =>OCO-(CH2)13-(CH)3CH-CH(CH3)-(CH2)4-CH=CH(CH2),OCO
m(590)=>OCO(CH2)90H=CH(CH2)4-(CH3)CH-CH(CH3)-(CH2)4CH=CH(CH2).OCO
k(534)=>oc0(cuz)gcH=cH(cu2).-(CH3)ca-cmcug-wug,2cHzcuo

FY 1; FY 2

(b) 242 => OCO-(CH2)12-CH3
(C) 270 => OCO'(CH2)14'CH3

(d) 296 -_-> ocomuzoficu-_-cu(cui.)5-CH3

(e) 293 => 000(cH.‘,),.5.-cu3

Figure 12

152

Most of the members of the cluster of peaks observed were predicted with this calculation.
The asterisked small numbers in the columns in Table 1 points out the numbered peaks (1,
2,.., 9) observed in FAB-MS analyses (Figure 11). This combined with the structural data
presented, confirmed that this lipid fraction is composed of a new family of monoglucosyl

diglycerides containing bifunctional fatty acyl chains.

Biochemical Significance of Glucolipids Containing Bifunctional Fatty Acyl
Chains

One possible mechanism for the synthesis of these unusual lipid components
which was alluded to earlier (6, 7), is the random, pairwise, combination of fatty acyl
chains in a tail-to-tail manner across the interface between the leaflets of the bilayer. Based
on this scenario, we should expect to isolate intact membrane species of constant and
invariant head group structure which contain fatty acyl components that reflect the
distribution in the entire membrane. In this work, one such family of lipid species has been
described. These results are powerful evidence in favor of a general adaptive response in
S. ventriculi in which existing lipid species become chemically linked during certain stress

conditions to transform the bilayer membrane into a bipolar monolayer membrane.

153

CONCLUSIONS
Recent research on the fatty acyl chains in the membrane lipids in Sarcina ventriculi has
shown that unusually long chain, bifunctional fatty acyl components are the major
components of the total lipid. These studies did not yield any information on the complete
structures of the lipids species containing these fatty acids. In this study, the structures of
a new family of glucolipids containing bifunctional acyl chains are described. These
structures were determined using NMR (Nuclear Magnetic Resonance) spectroscopy,
GC(Gas Chromatography)/MS (Mass Spectrometry), FTIR (Fourier transform Infrared)
spectroscopy and FAB (Fast Atom Bombardment) mass spectrometric studies. One of the
major bifunctional acyl components of the (It—glucolipids was an m-formylmethyl ester
indicating the presence of a plasmalogen. The general structure of the lipid components is
one in which the two head groups are separated by a membrane-spanning acyl species.
One head group component is glycerol and the other is a glyceryl glucoside. Two regular
chain fatty acids, one on the glycerol moiety of each head group, are also present and meet

in the middle of the membrane, roughly equi-distant from each head group.

154

REFERENCES
1. Silvius, J. R., and McElhaney, R. N. (1979) Chem. Phys. Lipids 24, 287-296
2. Silvius, J. R., and McElhaney, R. N. (1980) Chem. Phys. Lipids 26, 67-77
3. Silvius, J. R., Mak, N., and McElhaney, R. N. (1980) in Membrane F luidity (Kates,
and Kuksis, A., Eds.) pp 213-222, Humana Press, Clifton, NJ .1.
4. Bazzi, M.D., and Nelsestuen, G.L. (1987) Biochemistry 26, 5002-5008
5. Bell, R.M. (1986) Cell 45, 631-632
6. Jung, 8., Lowe E.S., Hollingsworth, I. R ., and Zeikus, J. G. (1993) J. Biol. Chem.
2828—2835
7. Jung, 8., and Hollingsworth, LR. (1993) in preparation (JBC submitted)
8. Morris, J. L. (1962) Chem. & 1nd,. 1238-1240
9. Holme, J. D. and Peck, H. (1983) In Analytical Chemistry pp 436-467 Longman,
Inc. New York
10. Shreve, O. D., and Heether, M. R., (1950) Analytical Chemistry 22, 1261-1264
11. Canale-Parola, E. (1986) Genus Sarcina Goodsir 1842, 434’“, p. 1100-1103. In P.
A. Sneath, N. S. Mair, M. E. Shaarpe and J. G. Holt (ed.), Bergey’s manual
of systematic bacteriology, vol. 2. The Williams & Wilkins Co., Baltimore
12. Goodwin, S., and J. G. Zeikus. (1987) J. Bacteriol. 169, 2150—2157

CHAPTER VI
A FAMILY OF VERY LONG CHAIN a,co-DICARBOXYLIC ACIDS IS A

STRUCTURAL COMPONENT OF MEMBRANE LIPIDS OF
CLOS TRIDI UM THERM 0H Y DROS ULF URI C UM

155

156

INTRODUCTION

The optimum temperature for growth of the thermOphilic eubacterium, C lostridium
thermohydrosulfuricum, is 67-690C The maximum temperature at which growth occurs
is 76-780C and the minimum temperature is 42°C (1). The structural basis for the
stability of membranes of thermophilic bacteria such as Clostridium has been somewhat of
an enigma because a number of physical studies, employing model membranes or natural
membranes enriched in typical regular chain branched-chain fatty acids (the predominant
type found in these organisms), have indicated that they act as membrane ”fluidizers”
rather that “stabilizers” (2,3,4). Based on these studies, the relatively obscure process of
homeophasic adaptation (5), instead of homeoviscous adaptation (6, 7), was invoked in
order to explain the roles of these branched fatty acids in the thermophilic membrane. The
concept of homeOphasic adaptation in the membrane of thermophilic bacteria was proposed
mainly in an effort to explain how branched fatty acids components of regular length (C12-
C20) could function in maintaining the stability of these membranes at such high
temperatures. The process of homeophasic adaptation requires that the relative proportions
of the different lipid phases ( e.g. hexagonal, cubic etc.) always remain constant as the
temperature changes. The difficulty with this concept is that there is no clear physical
parameter one could use to follow these phase contributions. The possibility exists that
there is some other major structural feature of the membrane of thermophilic bacteria
which, when combined with the presence of the branched fatty acids, could easily explain
membrane stability. In a recent study (1, 2) we reported an interesting adaptive
mechanism in Sarcina ventriculi in which there was chemical cross-linking of the tails of
fatty acids from opposite leaflets of the membrane bilayer in response to thermal , pH or
solvent stress. This resulted in the formation of very long chain, transmembrane,
dicarboxylic acid species. The possibility that mechanisms such as this may be possible in

Cl. thermohydrosulfuricum was therefore explored. This seems reasonable in view of the

157
fact that S. ventriculi and Cl. thermohydrosulfuricum are not too distantly related. In this
study, we explore the possibility that the mechanism used by S. ventriculi to adapt to high

temperature might be employed by the thermophile C l. thermohydrosulfuricum during its

nOrmal growth.

158

MATERIALS AND METHODS

Bacterial Cultures and Membrane Isolation

Cl. thermohydrosulfuricum 39E was isolated from Ocopus Spring at Yellowstone
National Park (13) and is listed in the American Type Culture Collection (Rockville, Md.)
as ATCC 33223. The cells were grown on TYE medium which contained yeast extract,
trypticase, trace salts, and vitamins (TYE medium) with 0.5% of either xylose or glucose
as the fermentable carbohydrate in stringent anaerobic condition at 65°C (1). Bacterial

membranes were isolated as described before (8).

Total Fatty Acid Analysis

To 1 ml of a membrane suspension in water was added 3 ml of chloroform
followed by 15 ml of a 5% methanolic HCl solution. The suspension was heated at 72°C
for 12hrs and 3 ml portions of chloroform solution were added after every 6 hrs. with mild
sonication for 5 minutes after each addition. The mixture was then concentrated to dryness
on the rotary evaporator and the fatty acid methyl esters isolated by extraction several times
with chloroform. The fatty acid methyl esters prepared above were subjected to gas
chromatography analysis on a 25 m J&W Scientific DBl capillary column_using helium as
carrier gas. The temperature program started at 150°C with 0.00 min. hold time and a 3.0
deg/min. temperature ramp to a final temperature of 200°C with a hold time of 0.00 min.
There was another temperature ramp at a rate of 4.0 deg/min. to a final temperature of
300°C. The temperature was held at 300°C for 30 min. GC/MS analysis was performed
on a Jeol JMS-AXSOSH spectrometer interfaced with a Hewlett-Packard 5890A Gas
Chromatograph.

159
Isotope Labeling
Methyl esters of fatty acid obtained by acid methanolysis were deuterium labelled
by treatment with 5% D-4 methanolic-HCI solution for 6 hrs at 72°C. The deuterated

methyl esters of fatty acid were extracted and analyzed as described before.

Isolation of 0t,(o-l3,lG-Dimethyloctacosanedioate Dimethyl Ester

The mixture of fatty acid methyl ester was applied to preparative silver TLC (thin
layer chromatography) plates which were eluted with chloroform/hexane (l.5:1.0,v/v)
solutions. Bands were visualized by treatment with Iodine vapor. One of the fractions
containing the dicarboxylic acid dimethyl esters, determined by GC analysis of all the
bands, was subjected to flash column chromatography (14). The column was sequentially
eluted with 3 times the column void volume (Vc) of a chloroform/hexane (1.5:1,v/v)
mixture, 2 Vc of chloroform/hexane (4:1, v/v) and Vc of chloroform at a flow rate of 20
ml/min. Separated components were conveniently detected by spotting 5-10 ul of each
fraction on TLC plates followed by elution with chloroform and charring. Similar
fractions were pooled and concentrated on a rotary evaporator. Each product was

dissolved in hexane and then subjected to further analyses.

DQF—COSY and DEPT Experiments .

Proton N MR spectra were recorded at 500 MHz on solutions in CDC13, Correlation
data was obtained by 2D DQF-COSY (15) experiment. 13C NMR spectra were recorded at
125 MHz on solutions in CDC13. DEPT (Distortionless Enhanced by Polarization
Transfer) experiments were performed for establishing l3C chemical shift assignments.
Chemical shifts are quoted relative to the chloroform resonance taken as 7.24 ppm for

proton and 77 ppm for 13C measurements, respectively.

160

Fourier Transform Infrared Spectroscopy
Spectra were obtained with a Nicolet model 710 FT -IR spectrometer in a 10% (w/v)

solution of dicarboxylic acid dimethyl ester in chloroform.

161

RESULTS AND DISCUSSION

Total Fatty Acids Analyses of Cl. thermohydrosulfuricum

The total ion chromatogram from the gas chromatography/mass spectrometric
analysis of fatty acids methyl esters extracted from cells grown at 67°C is shown in Figure
l. The peaks from 1 to 6 are due to typical membrane fatty acyl components ranging from
14 to 17 carbons. Major regular fatty acids are iso-pentadecanoic acid (peak 3, iso-C15;o)
and iso-heptadecanoic acid (peak 5, iso-C17:0). The peaks from A to D, however,

correspond to unusual components of apparent lengths much greater than the usual lengths.

Mass Spectrometric Analyses of C30 - Dicarboxylic Dimethyl Ester

Peak B is the most abundant fatty acyl component (about 36 %) of the membrane of
Cl. thennohydrosulfuricum. The electron impact mass spectrum of Peak B is shown in
Figure 2. Figure 2A shows major ions at m/z 510, 478 and 446. These corresponded to
the molecular ion of a C3o-a,(o-dicarboxylic dimethyl ester (M+) with the sequential losses
of methanol (CH30H) and methoxy (C1130) groups, respectively. The predominant ion at
m/z 297 suggested dimethyl branching and represents the structure obtained by
fragmentation at a secondary carbon (Figure 3A). The ion m/z 265 was assigned to the
loss of methanol (CH3OH) from the ion m/z 297. The proposed positions for the
placement of the methyl groups was supported by the presence of the strong peak at m/z
255 corresponding to the loss of propene (CH2=CHCH3) from the ion m/z 297 by way of
a single bond inductive cleavage mechanism (Figure 3B) (17). The ion at m/z 199 was
similarly assigned to the loss of propene (CI-12=CHCH3) from the other secondary cationic
fragment at m/z 241. The characteristic McLafferty fragment ion at m/z 74
(CH3OC(OH)=CH2) was also detected. Figure 2B shows the electron impact mass

spectrum of the deuterium-labeled molecule obtained by methanolysis with the D-4

162

Figure 1. Total ion chromatogram of Gas Chromatography/Mass Spectrom-
etry analysis for the esterified fatty acyl components of the membrane of
Cl. thermohydrosulfuricum. The later eluting cluster of peaks is due to very long
chain cam-bifunctional fatty acids.

1. C153 iso-branched fatty aldehyde (OHC(CH2)12(CH3)CHCH3)

2. C141) iso-branched carboxylic acid methyl ester (OCH3CO(CH2)10(CH3)CHCH3)

3. C151) iso-branched carboxylic acid methyl ester (OCH3CO(CH2)11(CH3)CHCH3)

4. C151) iso-branched carboxylic acid methyl ester (OCH3OO(CH2)12(CH3)CHCH3)

5. C171) iso-branched carboxylic acid methyl ester (OCH3CO(CH2)13(CH3)CHCH3)

.6. unknowns

A. C29;o (mo—dicarboxylic dimethyl ester

B. C301) cam-dicarboxylic dimethyl ester

C. C311) cam-dicarboxylic dimethyl ester

D. C320 (LCD-dicarboxylic dimethyl ester

163

Retention Time

 

 

 

 

 

 

 

 

 

 

 

 

g 1 ea 1 14c 43 A 49 A A 6p A 89
b 3 8
U
n
d 80-
a
n
C . 5
c 60-
: 2 l
40-
l
20-
Via/L
é . --,----.-z-. ~.-~fi~'x1.a
1009 2080 3888 4008 5898 600056
on
Figure 1

164

Figure 2. Electron impact mass spectrum (70ev) of peak B without (A) and
with (B) isotope labeling. (A) Major ions appear at m/z 510, 478, and 447. These
corresponded to molecular ion (M+) with the sequential losses of methanol and a methoxy
group, respectively. The ion at m/z 255 is due to the single bond inductive cleavage of the
major fragment ion at m/z 297. (B) It shows the El spectrum of deuterium labeled
molecules obtained by deuteration with D-4 methanolic HCl solution. The presence of two

carboxylic and two branching methyl groups was confirmed by isotope labeling.

165

3

 

SIG

 

 

 

100

 

446

 

400

300

 

200

 

 

80
50.
40

8523a asuemm

 

 

400

 

 

 

 

7d! {I M

 

m w w w a 11H“

852.3% o>$£em

Figure 2

166

Figure 3. Analysis of mass spectral fragmentations of peak B. (A) Mass
fragmentation pattern of peak B. 18: water, 31: methoxy, 32: methanol, 42: pr0pene, 56:
butane (B) Single bond inductive cleavage fragmentation mechanism. It produced m/z 255

ion as a major product.

167

 

-42 -31
[r - > 255 '—>' 224
7“ ”trz""“"t;, """ r
-32 -42 \\ /’ 297 265 -——)._ 209
168 -<-—— 199 \ ,’
\ / r---;l- --------------

------------ ~‘ I I
‘ CH3/ 213———>-182 0

4:11-241 ‘ “
CH30Y\/\}l{/V\/\:1’M\/\/\/V\/u\
- - OCH
182 _(_3‘_213J” \\ /‘\_24-l--—_>.:‘---2-.19-—-—-— 3

 

: O ---------- CH3 ’V‘
' 223 ~4—— 255<—-—j \‘
' -32 -42
l
' + -32 -32
‘ M (510) —_>'478 ——+446
L _______________________
460
-18
3
(11+ singlebondinduetivecleavage
0 >-
CH3
III/7.297
CH3°W + .
( O ) + CHz=CHCH3
M255

Figure 3

168
methanol/HCI solution (to produce the trideuterated methyl ester). The most striking
change was the shift of molecular ion by 6 mass units confirming the presence of two
methoxy groups. Two trideuterio-methanol (CD3OH) losses from the molecular ion (516
m/z) to give ions at m/z 481 and 446 were also observed. The ions at m/z 258
(CD3OCO(CH2)11 CH(CH3)CH2) and 300 (CD3OCO(CH2)11CH(CH3)CH2CH2
CH(CH3)) corresponded to deuterated m/z 255 (CH3OCO(CI-I2)11CH(CH3)CH2) and 297
(CH3OCO(CH2)11CH (CH3)CH2CH2CH (CH3)), respectively. Figure 3A and Table 1

explains the origins of most other ionic fragmentation products.

NMR / FTIR Analyses of C30 - Dicarboxylic Dimethyl Ester

The 1H NMR spectrum (Figure 4A) of peak B contained signals at 0.83 and 1.25
ppm characteristic of the methyl and methylene groups, respectively of the long chain fatty
acyl components. Ester methoxy groups (CH3O-) were assigned to a singlet at 5 3.65.
The correlation of the peaks in the region between 0 and 2.6 ppm was investigated by 2D
N MR (DQF COSY) spectroscopy (Figure 5). Correlation analysis showed the
relationship between peaks such as peak A with D, peak D with A and B, peak E with F
and C and peak C with B and E (Figure 4). The doublet at 0.83 ppm (J=6.3Hz) was
assigned to the branched methyl groups of the fatty acyl chain. Resonance at 5 1.61 were
assigned to the protons of the B-carbons of the molecule. Resonance at 5 2.30 (t,
J=7.5Hz) was assigned to the methylene function a to the carbonyl group (-CHz-CO).
The 13C NMR spectrum (Figure 4B) showed the ester carbonyl carbon at 5 174.2,
methoxy carbon at 5 50.9, and branching methyl carbons at 5 19.2. Moreover, all these
resonances were present as singlets, indicating magnetic equivalence due to molecular
symmetry. In order to confirm the number of branched methyl carbons, DEPT
(Distortionless Enhancement by Polarization Transfer) experiments were performed (16).
Figure 7 shows the DEPT spectrum after making all signals positive. All 32 carbons (16
pairs) were assigned by this analysis. The presence of a signal for one methine carbon

169

Table 1. Analysis of electron impact mass spectral fragments of peak B

 

 

Structure of ionic fragments mass(m/z) deuteriated mass(m/z)
CH,OCO(CH,),.CH(CH,)CH2CH2(CH,)CH(CH2).,OCOCH, 510 $16
CH,0CO(CH,_),,CH(CH,)CH,CH,(CH,)CH(CH2).OCH=CO 478 481
co=CH(CH,),,CH(CH,)CH,CH,(CH,)CH(CH,).ocu=co 446 446

478 - ts (H20) 460 463
CH,0C0(CH,).,CH(CH,)CH,CH,CH(CH,) 297 300
CO=CH(CH,).0CH(CH3)CH,CH,CH(CH,) 265 268
CH,OCO(CH,)"CH(CH,)CH2 255 258
capootcngucmcrrp 2'“ 2“
OQCHAtCEKCHQGiz 22‘ 224
CH,000(CH,).. 213 216
Gama)" ‘32 I82

(madam, 74 . 77
019006112)- :n=l~13 73 4» Wt! 76 + 14m
(IO-CH(CI-lz). : nut-13 55+14‘n 55+l4‘n
Gig-(01,). : n- 3~12 $741011 57+14‘n

arm-(aim :uIZ-ll SS+14‘n 554-1911

 

170

Figure 4 . 1H and 13C NMR spectrum of peak B. (A) 1H NMR spectrum of
peak B It contained signals at 0.83 ppm ((1, J=8.3Hz) and 1.25 ppm characteristic of the

methyl and methylene groups of long chain acyl components. Resonance at 2.30 ppm (t,

=7.5Hz) and at 3.65 ppm (3) represented methylene groups of a to the carbonyl and ester
methoxy group, respectively. The multiplets at 1.61 ppm was assigned to the protons of
the [3 carbons of a molecule. The signal at 5 7.24 .was assigned to the chloroform. (B) The
125 MHz 13C NMR spectrum of peak B. The presence of a methyl ester and branching
methyl group are confirmed by signals at 50.9 and 19.2 ppm respectively. The signal at
174.2 ppm was assigned to the carbonyl carbon of the ester group.

171

 

 

 

 

 

 

9.0 6.0 7.0 6.0 5.0 4.0 3.0 2.0 ‘l .0 0.0 (pm
'1 ‘1‘ l l l
200 100 ISO 140 120 1“ 80 ‘0 40 20 "In

Figure 4

172

Figure 5 The DQF-COSY spectrum (in the region between 0 to 2.6 ppm)
of peak B, in CDCI3 at 500MHz. Correlation analysis showed the relationship
among the peaks such as peak A with D, peak D with A and B, peak E with F and C and
peak C with B and E (Figure 6). The doublet at 0.83 ppm (J =6.3Hz) was assigned to the

branched methyl groups of the fatty acyl chains and resonance at 5 1.61 as the protons of

the B-carbons of the molecule. Resonance at 5 2.30 (t, J=7.5Hz) was assigned to the

methylene groups Otto the carbonyl group (-CI-12-CO).

 

 

F2“;
(00mg

N I-l M h-h u t.- O
O m 03 A N O Q
nnluulunluuluulnuluuluulturluulnuluul r l‘nllluallnntnl

N
N

N
A

N
at
lunlnni

173

 

 

06) 0
"0'6 9

as

S 88

 

 

"WWWWWWWWWW

2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.9 0.8 0.6
F1 (ppm)

Figure 5

174

Peale IPcakB

 

 

 

 

 

l__l
PeakE MA PeakD PeakF

Figure 6. The correlation analysis of DQF-COSY spectrum of peak B.

175

 

 

 

 

 

 

CH3 carbons
CH2 carbons l ‘1
WW
CH carbons
all protmated carbons mu
v“ ”‘V‘w‘ *‘ru‘m‘—::‘::wr:eer‘rvi‘“ writinw‘ ‘_r::.:.:::‘..~ - L#_
WW“! ’Wfl—fiq’q'T-B fl'q-P'q—‘q" “0" 'H’ I‘fiq-P' -0 '1". 0' r "H 1 1 1 v W 1 j 'fifi' I j ‘fifi
so 55 so 45 no 35 30 25 zo 15 oo-

Figure 7. DEPT spectrum of peak B in CDC13 after making all peaks
positive. The presence of one methine carbon (CH) and two methyl carbons (CH3)
confirmed the 2 branched methyl groups and two methoxy groups in the structure.

176
(CH) confirmed the presence of the 2 branching methyl groups in the structure. Figure 8
shows a FI‘IR spectrum. It contains very strong signals for methylene asymmetrical
stretching at 2928 cm“1 and symmetrical stretching at 2856 cm'l. The asymmetric
stretching of methyl groups occur at 3017 cm‘l. The symmetric bending vibration of
methyl groups appear at 1378 cm“1 and asymmetric bending at 1458 cm'l. The absorption
band at 1378 cm'1 arises from the asymmetrical bending of the methyl C-H bonds. The
scissoring band of the methylene groups occur at 1463 cm‘l. The methylene twisting and
wagging vibrations occur at 1119 - 1378 cm'l. The characteristic C=O absorption band of
the aliphatic ester group appeared at 1733 cm'l. There was no evidence for alky chain

unsaturation in the molecular structure.

Mass Spectrometric Analyses of C29, C31 and C32 Dicarboxylic Dimethyl
Esters

Figure 9. shows the electron impact (70eV) mass spectrum of peak A. Major ions
appear at m/z 496, 464 and 432. These corresponded to the molecular ion of a C29-0t,a)-
dicarboxylic dimethyl ester (M+) with the sequential losses of two methanol (CH30H).
The major ion m/z 281 represents the structure obtained by the loss of one hydrogen
molecule on the secondary cations at m/z 283. The ion m/z 255 was assigned to the loss of
propene(CH2=CHCH3) from the ion at m/z 297 by way of a single bond inductive
cleavage mechanism (17). The ion at m/z 241 was similarly assigned to the loss of propene
(CH2=CHCH3) from the secondary cationic fragment at m/z 283. The characteristic
McLafferty fragment ion at m/z 74 (CI-I3OC(OH)=CH2) was also detected. Figure 9B
shows the electron impact mass spectrum of the deuterium-labeled molecule obtained by
methanolysis with D-4 methanol/HCI solution to produce the trideutrated methyl ester.
The most striking change was the shift of the molecular ion by 6 mass units, which
confirmed the presence of two methoxy groups in the molecule. The loss of two molecules

of trideuterio-methanol (CD3OH) from the molecular ion (502 m/z) to give ions at m/z 467

177

93.1

p

(9'0)

$0.8

 

 

manhunt.

 

 

 

 

S

 

A A
1 TV

4006 36! 7 3228 2840 2451 aocz 1674 1 205 09$ ' $07
mitt-her (an-t)

011.,
J

Figure 8. Fourier Transform Infrared spectrum of peak B. It shows strong
methylene stretching at 2856 cm‘1 (symmetrical), 2928 cm-1(asyrnmetrical). The bending
vibrations of the OH bonds in the methylene group occurred around 1463 cm"1
(scissoring) and 1119-1378 cm‘1 (twisting and wagging). The characteristic C=C
absorption band of the aliphatic ester group at 1733 cm'l. There was no evidence of

unsaturation or of oxygen combined in hydroxy or simple ether linkages.

178

Figure 9. Electron impact mass spectrum (70ev) of peak A without (A) and
with isot0pe labeling (B). (A) El spectrum contained major ions at m/z 496, 464 and
432. These corresponded to the molecular ion (M+) with the sequential losses of two
methanol molecules. The characteristic McLafferty fragment of aliphatic esters appears at
74 (m/z). (B) It shows the El Spectrum of deuterium labeled molecules obtained by

deuteration with D-4 methanolic—HG! solution. The presence of two carboxylic and two

branching groups was confirmed by isotope labeling.

179

 

 

 

 

 

 

 

[ «9

 

 

 

 

 

 

 

 

 

 

 

‘9 vii ll 3h Ii I”
z .w. 3 ”'Hh
z m _.. l
1 .H ll 9 l illllu

1 .11 Q

3 I lull

4

1 It. a. .’II
1 it” 1. ‘ til-
8 1 till... . will:
9 7 m a z
4 8 1 --1
..fl e! i I If;

S h lltll! B [I it
4 9 s 1...:
Z s | 'Ht
4 4. fqifi

1 Li all
0 . . . - - . -1100 1
w m m 0. a - - 1 -

852.5% 0302mm .

I
M m t w a 1 ‘ 1

.oofldvgfid ¢>mud~om

 

Figure 9

180

and 432 was also observed. Figure 10. shows most of the other fragments and their
origin. Figure 11A. shows the electron impact mass spectrum of peak C. Major ions
appeared at m/z 524, 492 and 461. These correspond to the molecular ion of a C31-0t,co-
dicarboxylicacid dimethyl ester (M’t) with the sequential losses of methanol (CH3OH) and
a methoxy (CH3O) group, respectively. The major ion at m/z 478 and 447 corresponded to
the sequential loss of methanol and methoxy groups, respectively from the second cationic
fragment at m/z 509. The ion at /mz 509 is due to the loss of a methyl group from the
molecular ion (M+). The major ion m/z 311 represents the structure obtained by
fragmentation of the alkyl chain to generate secondary cations . The ion m/z 269 was
assigned to the loss of propene (CH2=CHCH3) from the ion m/z 311 by way of single
bond inductive cleavage mechanism . The strong peak at m/z 255 corresponds to the loss
of propene (CH2=CHCH3) from the ion at m/z 297. The characteristic McLafferty
fragment ion at m/z 74 (CH3OC(OH)=CH2) was also detected. Figure 11B shows the
electron impact mass spectrum of the deuterium-labeled molecule obtained by methanolysis
with D-4 methanol/HG] solution (to produce the trideuterated methyl ester). The most
striking change was again the shift of the molecular ion by 6 mass units, confirming the
presence of two trideuterio-methyl groups. Loss of two molecules of CD3OH from the
molecular ion (530 m/z) to give ions at m/z 495 and 447 was also observed. Figure 12
shows the origin of the other fragmentation products.

The electron impact mass spectrum of peak D is shown in Figtne 13A. Major ions
appeared at m/z 538, 506 and 475. These correspond to the molecular ion of a C32-0t,a)-
dicarboxylic dimethyl ester (M‘) with the sequential losses of methanol (CH3OH) and a
methoxy (CH3O) group, respectively. The ions at m/z 297 and 325 arise from
fragmentation of the alkyl chain to form secondary cations. The ion at m/z 265 was
assigned to the loss of methanol (CH3OH) from the one at m/z 297. The ion at m/z 294
was assigned to the loss of a methoxy group from the secondary cation at m/z 325. The
strong peak at m/z 255 corresponded to the loss of propene (CH2=CHCH3) from the ion

181

 

 

-15
295 -H2
................ ., >281
-32 -H I l .42 -31
--(——311 g
278 4—310 : r2“ 210
296 4—11—1 ,--—L _________________ _
\ II '32 ~56
-32 -42 \ , , 283—->-251—).155
rss-<——t99 \ ,’ I
------------ 1-“ x ,’ : ’,-__.51.---_-__-_--__-
CH 0 210--<-l:11—-24l \ XX CH3/ ‘99 'L—‘F “53
3 " \ ’ OCH
\tl/WWW 3
' o Alix—2134’ CH ‘y’ 2123 .. —*' .... L 92-- o
' -31 ".- 3’ ‘ cl ,
: ___2_5)-<——— _ _____ 2 ctr-(i 2 7_,’ \ , 42 185 .—_>.32 153
_______ f, \ -
I
: 223+——255 .42 ‘5
' -32 g
' + '32 '32 I
t M (496) ———>-464 ——>-332
I. .......................
: 446
, -18
I
I
I
:- 74(McLafl’eryfiagment)

Figure 10. Mass fragmentation pattern of peak A. 15: methyl, 18: water, 31.
methoxy, 32: methanol. 42: propene, 56: butane, 74: McLafferty fragment

182

Figure 11.E|ectron impact mass spectrum (70ev) of peak C without (A) and
with isotope labeling (B). (A) El spectrum contained major ions at m/z 524, 492 and
461. These corresponded to the molecular ion (M+) with the sequential losses of methanol
and a methoxy group, respectively. The major ion at m/z 478 was due to the loss of a
methoxy group from the fragment at m/z 509 which was produced after the loss of methyl
group from the molecular ion (N14). Another major ion at m/z 446 was due to the loss of
methanol from the ionic fragment at m/z 478. The characteristic McLafferty fragment of
aliphatic ester appears at 74 (m/z). (B) It shows the El spectrum of deuterium labeled
molecules obtained by deuteration with D-4 methanolic-HCI solution. The presence of

two carboxylic and two branching groups was confirmed by isotope labeling.

183

 

 

 

 

 

 

 

 

 

 

 

 

 

“h
S 0 _
.2l 2 m
m a s
.4
6
4 -
4 .l
..,. m
4 4
4
1 1
S ltllrt - M
Z a.
4 2 it.
2
e
0
- Z
... t.
4 [H
1 M
1 MN
2 1 till”.
a. 1 l m
3 1.
9 8 ..
7 6 t
S
S
o . . . . .
m w w w &
mocmvczna m>$£wm
A

 

 

 

 

 

 

 

..1
6t
.5
4
1 l.
4 ..
K,»
94 P
4
c. l
I, it
r-l.
x. 4
I
u.
r l
7. a" -W
7” '1
S G I
S

 

 

100

m m m -w

8532: 253mm

Figure 11 .

184

 

-42 -31
l- , > 269 -—>- 238
7" -73-2'---'--. ------- '—
-32 \ x' 3” —*‘ 279 ——>—56 223
168 +— 199 \ 1’

\ I r -------------------

------------ -‘ I I -
31 \ ‘4 CH3” 227-ALV196

’ 241
CH3O\n/\/\}!{/W/\/\/\Wocm
182 AZISJII \\ /‘\_255 #224 _ O

0 —————————— CH3 ’8" ——————————
210424144112 71’ \ 213 ———->-l8l
--------------- f— x 42 ~31
223 -<-—- 255 \\
_32 '42

474
-I8

-31 -32
——->- 478 —-—>-446

 

I
I
I
I
I
I
I
I
I
I.
I
I
I
I
I
I
I
I
I

Figure 12. Mass fragmentation pattern of peak C. 15: methyl, 18: water, 31.
methoxy, 32:methanol, 42: propene, 56:butane, 74: McLafferty fragment

185

Figure 13. Electron impact mass spectrum of peak D without (A) and with
isotope labeling (B). (A) El spectrum contained major ions at m/z 538, 506 and 475.
These corresponded to molecular ion (M+) with the sequential losses of methanol and a
methoxy group respectively. The ions at m/z 255 and 283 are due to the single bond
inductive cleavage of the major fragment ions at m/z 297 and 325, respectively. (B) It
shows the BI spectrum of deuterium labeled molecules obtained by deuteration with D—4

methanolic-HCI solution. The presence of two. carboxylic and two branching methyl

groups was confirmed by is0t0pe labeling.

186

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A
L
8 4
3i
5
6 n
... .4
s m a m
‘ A,
7 4 inw
« h
3
4
8h
9
0 4 I 4 '1 , d I o
w. m w m a

macawcsna 9253mm

 

 

 

544

 

 

12

 

 

 

a: a- «at.»

4
aocavgna 9530mm

Figure 13

187
m/z 297 by way of a single bond inductive cleavage mechanism. The ion at m/z 283 was
similarly assigned to the loss of propene (CH2=CHCH3) from the secondary cationic
fragment at m/z 325. The characteristic McLafferty fragment ion at m/z 74
(CH3OC(OH)=CH2) was also evident. Figure 138 shows the electron impact mass
spectrum of the deuterium-labeled molecule obtained by methanolysis with the D-4
methanol/HCI solution to produce the trideuterated methyl ester. The shift of the molecular
ion by 6 mass units confirms the presence of two methoxy groups. Loss of two molecules
of trideuterio-methanol (CD3OH) from the molecular ion (544 m/z) to give ions at m/z 509
and 475 was also evident. The ions at m/z 258 (CD3OCO(CH2)11 CH(CH3)CH2) and 286
(CD3OCO(CH2)13CH(CH3)CH2) correspond to deuterated forms of m/z 255
(CH3OCO(CH2)11CH(CH3)CH2) and 283 (CH3OCO(CH2)13CH (CH3)CH2),
respectively. The other fragmentation products are shown in Figure 14. Figure 15 shows

the structures of the members of the new family of onto-dicarboxylic acids in CI.
thermohydrosulfuricwn. They are auto-13,16-dimethylheptacosanedioate dimethyl ester
(C29) a,to-13,16—dimethyloctacosane dimethyl ester (C30), a,co-13,16-dimethylnona
cosanedimethyl ester (C31) and 0t,tr)-13,16—dimethyltriacotane dimethyl ester (C32).

The General Significance of These a,to-Dicaboxylic Acyl Components

A recent study of the membrane of a strict anaerobic, facultative acidophilic
eubacterium, Sarcina ventriculi demonstrated that a.m~dicarboxylic fatty acyl components
are synthesized as a general response to the perturbation of the membrane structure by the

addition of exogenous organic solvents or increasing temperature (8). Interestingly, the

membrane of Clostridium thermohydrosulfuricum contains similar cue-dicarboxylic fatty
acyl components to those formed inS. ventriculi. The critical difference is the position of

internal methyl branches. It appears the synthetic mechanism of formation of (1,0)-

dicarboxylic fatty acyl components is similar to that found in S. ventriculi. The proposed

mechanism is by way of tail-to-tail (tn-1) coupling of the lipid bilayer across from the

188

 

-42 .32
[— - > 283 ——>- 251
I'" "'51 """"""""""" —
-32 -42 \ /I 325 ———*' 29“ —>-'42 252
168 -<—— 199 \ ,’
- ----------- ——\ \ ,’ /"""§,— """"""" 6""
- -31 24, \ A CH} I 241-'——->—210
CH3O\"/\/\}l{/\/\/;{:(WW\/\/\/\/u\
- \ - OCH
, o 132 .(_3_‘_2n_// CH \y/‘\_263__—->-il___2.3§______ 3
. "i3". """ 3’ x 1 ,
I 210-4——241-<—'5-§-2 7J1 \ 227 —_,_195
| ---------------- r- \\ 42 -32
I
I
I
I

Figure 14. Mass fragmentation pattern of peak D. 15: methyl, 18: water, 31.
methoxy, 32: methanol, 42: propene, 56: butane, 74: McLafferty fragment

 

 

 

 

 

011’

C811

aci

US

di.

189

opposite sides of the membrane. The difference in the position of internal methyl branches
can be easily rationalized if one considers the fact that there are predominantly iso-branched
regular fatty acids in this organism. Figure 15 shows the possible correlation between the
cue—dicarboxylic acids and the regular fatty acids. The most abundant regular chain fatty
acid in CI. thermohydrosulfuricum is the iso-pentadecanoic acid. This fatty acid could be
used as a general frame for the coupling step ((1)-coupling) to synthesize the (1,011—
dicarboxylic acids. The family of archaebacteria that can easily grow at extremely high
thermal condition (more than 100°C) provides an interesting analogy. These organisms
contain the diether lipid as well as tetraether lipid groups in their membranes The
tetraether lipids are thought to be made by way of the tail-to-tail coupling mechanism
(10,11,12). The fact that Cl. thermohydrosulfuricum or S. ventriculi are eubacteria
strongly supports the idea that the tail-to-tail coupling mechanism of the membrane lipids is
very general throughout the bacterial kingdom.

190

Figure 15. The determined structures of a family of very long chain a,m-
dicarboxylic dimethyl esters. A: cam-13,l6-dimethylheptacosanedioate dimethyl
ester (C29) B: a,m-13,16-dimethyloctacosane dimethyl ester (C30), C: (Leo-13,16-
dimethylnonacosane dimethyl ester (C31), D: mtg-13,16-dimethyltriacotane dimethyl ester
(C32) All assigned axe-dicarboxylic acids could be constructed by (1)-coupling of opposite

fatty acyl groups of the membrane.

 

CH

191

 

 

 

 

 

 

 

 

i30'C1520
I 1 CH3
0 CH . o
3 [ ’50’C14-o J
A
iSO'C1520
l I CH3 o
(350ka
OCH3
0 CH3
iso-C
[ 15:0 J
B
iSO'Crszo
I I CH3
CHBOWAAAWAAAAn/OCHB
O O
CH .
3 l “O'crsw J
C
“O‘clsfl
I l 3
CH3°\”/\/W\/\/fi/\fl/\/\/\/W\/E\
0CH3
° CH
3 I ISO-Crko J
D

Figure 15

 

 

 

 

 

 

char

 

therq
“er:
met
chr
65b
Sp:
SP
5?

192
CONCLUSIONS

A new family of cam-dicarboxylic, very long chain fatty acids was isolated and
characterized from the lipids of thermophilic anaerobic eubacterium, Closm'dium
thermohydrosulfuricum. After the isolation of the membrane, the fatty acyl components
were released by converting them to their methyl ester forms by acid-catalyzed
methanolysis. The esterified fatty acyl components were purified by a variety of
chromatographic techniques and analyzed by GC and GC/MS. One of the isolated,
esterified, a,a)-dicarboxylic, very long chain fatty acids was characterized by mass
spectrometry, 1H NMR and 13C NMR spectroscopy and Fourier transform infrared
spectroscopy. NMR experiments employed included double quantum filtered correlated
spectroscopy (DQF-COSY) to assign proton spin-spin coupling and polarization transfers
(DEPT) to measure the multiplicity of carbon signals split by protons. The other

components were identified by mass spectrometry. The new long chain fatty acid methyl

esters are a.m-l3,l6-dimethylheptacosanedioate dimethyl ester (C29), cum-13,16-
dimethyloctacosane dimethyl ester (C30), a,a)-13,16-dimethylnonacosanedimethyl ester
(C31) and a,co-l3,16-dimethyltriacotane dimethyl ester (C32). This family of mm-
dicarboxylic fatty acids accounts for about 40 % of fatty acyl components of the
membrane of Clostridium thennohydrosub‘uricum. Interestingly almost all (>90%) of the
very long chain, cam—dicarboxylic fatty acid was a,co-l3,l6-dimethyloctacosanedioic acid.
A careful study of the structures of the family of cam-dicarboxylic acids indicated that the
synthetic mechanism for their formation involves the tail-to-tail (0)) coupling of regular iso-

branched fatty acids across opposite sides of the membrane bilayer.

193
REFERENCES
l. Lovitt. R. W, Longin, R., and Zeikus, J. G. ( 1984) Appl. Environ. Microbial.
48, 171-177
2. Silvius, J. R., and McElhaney, R. N. (1979) Chem. Phys. Lipids 24, 287-296
3. Silvius, J. R., and McElhaney, R. N. (1980) Chem. Phys. Lipids 26, 67-77
4. Kannenberg, B., Blume, A., McElhaney, R. N. and Poralla, K. ( 1983) Biachim.
Biophys. Acta 733, 111-116
5. Silvius, J. R., Mak, N., and McElhaney, R. N. (1980) in Membrane Fluidily (Kates,
H, and Kuksis, A., Eds.) pp 213-222, Humana Press, Clifton, NJ.
6. Sinensky, M. (1974) Proc. Natl. Acad. Sci. USA 71, 522-525
7. Okuyama, H., Fukunaga, N ., and Sasaki, S. ( 1986) J. Gen. Appl. Microbial. 32,
473-482
8. Jung, 8., Lowe E.S., Hollingsworth I. R., and Zeikus, J. G. (1993) J. Biol. Chem.
268, 2828-2835
9. Chan, M., Himes, R. H., and Akagi, J. M. (1971) J. Bacteriol. 106, 876-881
10. Langworthy, T. A. (1982) Curr. Top. Membr. Transp. 17, 45-77
11. Langworthy, T. A. (1977) Biochim. Biophys. Acta 487, 37-50
12. Ross, H. N. M., Collins, M. D., Tindall, B. J., Grandt, W.D. (1981) J. Gen.
Microbiol. 123, 75-80 .
13. Zeikus, J. G., Ben-Bassat, A., and Hegge, P. W. (1980) J. Bacterial. 143, 432-440
14. Still, W.C., Kahn, M., and Mitia, A. (1978) J. Org. Chem. 43, 2923-2926
15. Rance, M., Sorensen, O. W., Bodenhausen, G., Wagner, (3., Ernst, R. R. and K.
Wuthrich ( 1983) Biochem. Biophys. Res. Commun. 113, 967-974
16. Derome, A. E. (1988) In Modern NMR Techniques for Chemistry Research
(Baldwin, J. E. ed.) pp 143-151, Pergamon Press, New York
17. McLafferty, F. W. (1980) In INTERPRETATION OF MASS SPECTRA (T urro, N.
J. ed.) ppl41-146, University Scirnce Books, California

194
18. Salem, L. (1962) Can. J. Biochem. and Physiol. 40, 1287-1298
19. Langworthy, T. L. (1985) in The Bacteria (Woese, CR, and Wolfe, R. S. eds.)
p459 , Academic Press, New York
20. De Rosa, M., Gambacorta, B., and Nicolaus, B. (1980) Phytochemistry 19, pp791-
795

CHAPTER VII
COMPUTATIONAL STUDIES ON THE EFFECT OF THE

TRANSMEMBRANE ALKYL CHAINS ON THE STRUCTURE AND
DYNAMICS OF MEMBRANE

195

 

 

mon

 

 

 

Phi“
W35

Silt

 

 

 

 

 

196

INTRODUCTION

Our research on the structural transition of membrane from the bilayer to a bipolar
monolayer form via (1)-1 or (o-linked acyl chains raises interesting questions about the
physiological role for this phenomenon (l, 2, 3). In Sarcina ventriculi this phenomenon
was responsible for the general adaptive response to the various forms of environmental
stress such as increasing the growth temperature, the addition of exogenous solvents and
lowering the pH (1).

Elevated temperature as well as the addition of the alcoholic solvents increased the
freedom of motion of the lipids. The reported adaptive response to this is to regulate the
fluidity of the membrane by changing the relative amounts of unsaturation of the acyl
chains or by increasing chain length of the acyl groups by 2 or 4 carbons (4, 5, 6). The
presence of very long bifunctional lipids in the membrane of S. ventriculi could have arisen
by coupling the tails of fatty acids of membrane lipids from the same side of the bilayer.
This, incidently, was ruled out by freeze-fracture studies which demonstrated that the
leaflets could not be separated and that therefore the very long chain fatty acids were
transmembrane (Figure 1).

Our important question which arises from this research is what is the effect of
transmembrane lipids on membrane fluidity and dynamics. Membrane fluidity studies are
usually performed with spin probes using electron paramagnetic resonance spectrosc0py
(7, 8). One critical disadvantage of this method is that probe molecules might change the
local fluidity of the system.

The theoretical and computational methods offer tremendous insight into molecular
processes and make real-time simulation possible. Molecular mechanics and molecular
dynamics simulation techniques have proven to be especially suited to the study of

biological macromolecules and their assemblies. This is very evident from the vast

197

Figure 1. Electron micrograph of S. ventriculi cells grown at pH 7.0 (A)
and at pH 3.0 (B) after freeze fracturing and coating. (A) Note the

concave/convex surfaces (arrows) characteristic of fracture between bilayer leaflets. (B)

Note that cleavage through the cytoplasm is the only form of cleavage observed.

gun..." . c u”
. ., .. a m. , .. .
. . l ..

. (£1 .
. y
b v... ..,!
. $5,541»

3
a
Jh

198

 

Figure 1

199

amount of work done on the proteins, nucleic acids and lipids during the last decades
(9,10,11,12). These techniques were used for studying the effects of transmembrane alkyl
species on membrane structure and dynamics. RMS (Root Mean Square) distance and
specific angle fluctuation were investigated for alkyl chain atoms and bonds respectively.
Three different systems were studied. Firstly, a molecular dynamics simulation
(10ps) of two free, opposed, unconnected hydrocarbon chains (bilayer form) was
performed. The simulation was repeated after coupling the ends of chains (linked
monolayer form). Secondly, 10 ps dynamic simulation was also performed for two
opposed, unconnected and two connected phospholipid molecules with fixed headgroups.
The third system studied was an array of phospholipids (phosphatidyl cholines, sn-l for
C131), sn-2 for C16;1) arranged in a bilayer membrane. Simulated annealing dynamics
(13, 14) was used for the prediction of molecular geometry of the acyl chains before and
after the transition of the bilayer to bipolar monolayer. In this case, 32 of the bilayer lipids
(50 %) were changed to monolayer form by coupling only one alkyl chain to the opposing
one cross the bilayer. Using results obtained from these studies, it was possible to

rationalize the effects of transmembrane lipids on the structure of the membrane.

200

MODELS AND METHODS OF SIMULATION

Models for Simulations

For the simulation of free hydrocarbon chain, two octadecane (C13) molecules were
used . For the second model, two typical phosphatidyl choline lipids with sn-l C13;o and
sn-2 (316:0 were used. For the membrane simulation, 64 phosphatidyl choline lipids with
sn-l cum and sn-2 C153 were used. For the transition to the monolayer form, 50% of

lipids were randomly connected across the bilayer at the (1)-1 positions.

Force Fields for Energy Calculation
All molecular mechanics and dynamics calculations were performed under the
Dreiding-II force filed (15). The forms of potential energy functions for bond stretching,

angle bending and torsions are given as

 

 

 

Ebs = lflkbs(r're)2 (a)
Eab = 1l2kab(e’6e)2 (b)
and Ema = 1/2ij(1 - COSIlijk (<1) - ¢e)2] (C)

where kbs and kab are the force constants for bond stretching and angle bending; re and 96
are the equilibrium bond length and bond angle. Eiju represents the torsional interaction
potential between two bonds ij and kl connected through a common bondjk. In eq. (c) (1)
is the dihedral angle (angle between ijk and jkl plane) ij is the rotational barrier. njk is
the multiplicity and (be is the equilibrium value of the dihedral angle. The van der Waals
interactions are represented through a 12-6 Lennard-Jones potential.

EU=D0(P°12' 2M) (d)

Where p is the radial distance scaled by van der Waals radius R0, p=RlRo and Do is the

 

depth of the potential at minimum, R0 The electrostatic interactions are calculated by using
Eq=3 32.0637 Qin/ERij (e)

 

201
Where Qi and Qj are charges in electron units, Rij is the distance in Angstroms. 8 is the
dielectric constant, and 332.0637 converts EQ to kcal/mol. Interactions are not calculated
between atoms bonded to each other (1, 2 interactions) or involved in angle terms (1, 3
interactions) since these are assumed to be contained in the bond and angle interaction. For
charge calculation the Gasteiger estimates (16) were used. In calculating the non-bond
interactions, a cut-off distance of 9.0 A was used along with a switching function.
Switching function smoothly reduces the interaction potential and forces to zero using on-
distance at 8.0 A and off-distance 8.5 A. The force field parameters are kept as simple as

possible.

Methods for Simulations

All molecular mechanics and dynamics calculations were performed using the
BIOGRAF 3.0 software package (Molecular Simulations, Inc.) on a Silicon Graphics Iris
4D super-minicomputer. Initial constructed models were relaxed with 500 steps of
conjugate gradient minimization methods (17) and relaxed with 500 steps of molecular
dynamics, and reminimized with 5000 steps of conjugate gradients methods. The global
minimum structure of the 64 lipids systems was obtained after the reminimization of the
structures obtained by 5 cycles of annealed dynamic simulation with the initial 300K to
600K temperature program from the initial minimized structure. A 1 fs (10'15 s) time step
was used in the integration for dynamic calculations. These minimized structures were
used as the initial coordinates for the molecular dynamics calculations for 10 ps in the
adiabatic conditions. The equations of motions in MD simulations were integrated with the
Verlet algorithm (18). For the dynamic simulations of 2 lipids (second model), a point in
the head group was fixed during the dynamics simulation. This is a reasonable assumption
because the motional dynamics (translational diffusion) of the head group relative to the
vibrational motions of the hydrocarbon chains is static over the relevant time scale (10 ps)

(19).

202

RESULTS AND DISCUSSION

RMS Distance Fluctuation for Two Hydrocarbon Chain Model during 10 ps
MD Simulations

Figure 2A shows the trajectory snapshots for every 0.5 ps at the dynamic
simulations for two C13 molecules over 10 ps. Each end of the free octadecane molecules
showed a high degree of freedom of molecular motion. Figure 2B shows the snapshots of
the trajectory files obtained every 0.5 ps from the end-to—end condensed form (one C35
hydrocarbon). As expected, coupling of the ends of the molecules led to considerable
restriction of the motion of the chain. As shown in Figure 2B, the relative displacements
of the alkyl chain atoms are severely restricted. Figure 3 shows the RMS (Root Mean
Square) distance fluctuation based on the molecular coordinate of the initial minimized
structure (0 ps). As shown in Figure 3, the tail-linked form (C35 chains) has smaller
values of RMS distance fluctuation than free form (C13 atoms). This relative rigidity of
the C36 alkyl chains immobilized at each end can, in principle, be transmitted to
neighboring chains via dispersion forces. This possibility was explored using a model

system with two phopholipid molecules.

RMS Distance and Angle (<b) Fluctuations for Two Lipid Model during 10
ps MD Simulations

Figure 4 shows the pictures of snap shots of the trajectory files obtained every 0.4
ps during the 10 ps MD simulation. The unconnected lipids displayed a much wider
spectrum of motion (Figure 4A) compared to the (0-1 linked form (Figure 4B). Figure 5
shows a simplified picture of the motional spectrum of the lipids, (Dm indicate the

maximum angle sunbtended by the two alkyl chains at the head group. It is a measure of

the extent to which the two alkyl chains are splayed Figure 5A is an amplitude of dim for

203

Figure 2. Computer generated pictures of the trajectory snapshots of two
octadecane molecules for the 10 ps molecular dynamic simulations. Figure
2A shows the trajectory snap shots for every 0.5 ps (pico second) dynamic simulations for
two C13 molecules. Each end of the free octadecane molecules showed high degree of
freedom of molecular motions. Figure ZB shows the snapshots of the trajectory files

obtained every 0.5 ps from the tail-linked (ta-l linked) form (one C35 hydrocarbons).

There were no fixed points during the simulation.

Figure 2

 

205

 

 

 

 

15
000
o
o o
12- 0° 0 o
o
A
O 0 0°
2 o ‘A
0
v9- A
o A
g A
a .‘
:1
«- ... O A
36 AAAA
'3 A
'E’
A
M 0 AA
3..
A
A freefomr
2 ‘ tail-linkedform
oollLI‘l‘l‘lnl‘ltllllllllllll
on ‘0
o 2 153 2’3 '3 32 3"?”2’333
Trme(ps)

Figure 3. RMS (Root Mean Square) distance fluctuation based on the

initial minimized structure as a reference coordinate.

206

Figure 4. Computer generated pictures of the snapshots of the trajectory
files obtained every 0.4 ps during the 10 ps MD simulations. One point in the
lipid head group was fixed during the MD simulations. (A) Bilayer form of the two lipids
(B) Monolayer form (One of the acyl chains linked at the (a-l positions).

Figure 4

 

 

 

 

 

 

 

 

 

 

 

 

Figure 5. Simplified pictures of the motional spectrum of the lipids. Qm

indicated the maximum angle subtended by the two alkyl chains at the head group. It is a
measure of the extent to which the alkyl chains are sprayed. (A) Bilayer form (B)

Monolayer form

 

209

the unconnected lipid chains and Figure SB describes the situation when one alkyl chain of
one lipid species is connected to an opposing alkyl chain of the other. Figure 6 shows the
angle fluctuation (CD) between the bilayer form and monolayer form. During the first 3 ps,
both systems showed similar angular motional dynamics. However, after the 4 picosecond
point, the angular fluctuation differences between them dramatically increased. The
simulation indicated that the monolayer form has very restricted angular motional dynamics
because of the linkage at the center. Early similarity of the angular fluctuations between
them was due to the intrinsic angular fluctuation around the connected centers. It also
indicated that transmembrane acyl chains could strongly restrict the isotropic rotations of
neighboring lipids in the membrane.

Figure 7 shows the RMS fluctuation of each carbon atom in one chain (sn-l; C18)
as a function of MD simulation time. It confirmed that as time went on, the degree of
motional freedom in one chain became greater towards the end of the chain. Hence, at the
end of MD simulation (10 ps) there was an almost a linear relationship between the chain
length (carbon number) and the magnitude of the degree of motional freedom. A similar
phenomenon was observed in the other chain (sh-2 C16) of the same lipids (Figure 8).
Figure 9 shows the RMS fluctuation pattern of the carbon atoms of the “transmembrane”
alkyl chains for two opposed lipids when two chains are connected to form a
transmembrane system. Because of symmetry, only half of the molecules is considered,
the half of the transmembrane (C36) acyl chains. The front portion of the chain (Cl-C4)
shows the big amplitude of the motional fluctuation relative to the center of the chain. The
magnitude of the amplitude became small as the number of carbon increases. Finally, after
10 ps, all of the first 18 carbon atom of the chain showed similar values of RMS distance
fluctuation. The other free chain (C16) displayed an interesting pattern of the RMS
distance fluctuation (Figure 10). It displayed an oscillatory pattern of small magnitude (30
% of that observed when both chains are unconnected) because of attraction to the relatively

immobile “transmembrane” chain. This result proves that transmembrane lipid chains can

210

 

 

 

 

125
0“" bilayer form

100 "‘
A ........°....._. monolayer(to-l linked) form
3?: 75 -
c:
o
'5
cu
B
o
:r
c: 50 ..
.53
co
5

fi —

o I I
o 9 10

 

Time (ps)

Figure 6. Angle fluctuation between the bilayer form and monolayer form
during the MD simulations. The angle 4) is determined by the trajectories of the
lipids referred to the initial position (0 ps). The monolayer form showed very restricted

angular motional fluctuations because of linked structme at the center.

211

 

 

 

 

 

35
—cr— Zps B
30— "—4“ 4ps E‘E’
""0"" 6 ,E’ A
25" ps I3 ’/
”A." 8ps ,8! f
E’ /
20- ""8"- lo‘” .B’ / o
E ’83, A”
a; 33' / ,0
g 15— m" /A’ :9 o
:3 I / O" /
c: ' o" 0
U) x51 /,A .x /
E 10" Q‘. ’tm 'o'o' /°
8¢-¢D‘-6'E 0'0 o
5-
0 I I I I I I I I I I I I I I I I I
fiNMV'fl‘OhOOOsO—‘NMV’V‘A‘OFOOO‘
H o-c v-1 c—4 v—I v—t v—t v-l c—e c—q
Carbonnumber(C18)

Figure 7. RMS (Root Mean Square) distance fluctuation of each carbon
atom in a sn-l chain (C18) of the typical bilayer lipid as a function of MD
simulation time. The degree of motional freedom in the chain became greater towards

the end of the chain.

212

 

 

 

 

35 ’8
—C}— ,E
30" '3’
““0“” 4ps B’
Auto
25‘ ""0““ 6ps ’m x/
l’ [A
~‘fi”" 8ps a A’
A ’I /
o3; ,0- -------ar e A, ,o
I: x ”D
.5 x93 /A/ 0
:3 ’3 A ”o
g 15.— I’ ',"o
a. ,8! 0,0
m , ,x
E 10 _ E- -3 “a 'oo'a ”W0
exfi’»? ..O-"d'
.v “0"
5" W
0 I I I I I I I I
.—t (la (In it In :0 IL «'30 as g S g :3 E :2 ‘3 2

Carbon number (C16)

Figure 8. RMS (Root Mean Square) distance fluctuation of each carbon
atom in a sn-2 chain (C16) of the typical bilayer lipid as a function of MD
simulation time. The degree of motional freedom in the chain became greater towards
the end of the chain.

213

 

 

 

 

Q
—C}-— 2ps .._...o........ dps
7_ it ---E--- lOps ----o---- 6p:
3. "'Q "“18"" 8ps
6" O .0.
Q \‘ . ’a‘\ E
v _ 32:/Ox . ,0 I la’ \ o--0
5 sm \\ ’m‘ I \ E o'o
E) Q XE!” .v. E in’ ‘\\ ”’8‘ l '
.6 - \‘e. \ ' .
3 4— °'°t_ \O \ E fie a"
0 '2. A“ s ' ‘ "
a ‘1 b.‘ A I ‘l
‘o s.\ I ‘
8 3 0‘0 ‘ Q ‘9} ‘ ‘ "
- .\ . '\‘
g. \*-... ‘0" " :9“ 't F
a e... - .v'° ' ‘t. /
m 2- \ ' b ‘ I
E ' It
i . °
1- l ‘ f
I o
I . -
0 I 1 I Ifi I i I I I I I I I I I I
H N M V V} \O l‘ 00 O\ O H N m V V) \O l‘ 00 O‘
H H v-4 q—I v-d 0-1 o-4 v-t v—I v-t
Carbonnumber(Cl8)

Figure 9. RMS distance fluctuation of the half of the transmembrane

monolayer (C36) acyl chain (sn-l). It shows some kinds of oscillation of the

degree of motional freedom.

214

 

 

 

 

 

12
10“

A

o< 8...

v

t:

O

'5

g

0 6‘

:3

I:

0

E5

.9 4'-

Q

U)

E

a:
2—
OIIIIIIIIIIIIIII
Humvmsohoeoto-«vamxot‘

Ho—{HI-(v—lq—(q—(H

Carbon number (C16)

Figure 10. RMS distance fluctuation of the other free chain .(sn-Z, C16) in
the monolayer lipid form. It also showed some oscillation pattern of the magnitude
of the fluctuation probably due to the motional restriction of the sn-1 transmembrane acyl
chain. Itconfirmedthat the geometrictransitiontothem—l linkedacylchainmodulatedthe

fluctuation of the other acyl chain of the same lipid.

215
modulate the fluctuation of other free acyl chains in close proximity. Average values of the
atomic fluctuation of the lipid chains (C 16) were also studied. Figure 11 shows how the
amplitude of the fluctuation of the atomic distance changed as time went on from 1 ps, 5 ps
to 10 ps. This fact strongly confirmed that the transition to the monolayer from the bilayer
makes an important contribution to the reduction of the amplitude of the motional
fluctuation. Finally, total RMS distance fluctuation of the lipid was investigated based on
the initial lipid coordinates as the reference in Figure 12. Early similarity was broken after
the 3 ps. These kinds of data eventually explained the molecular viewpoints of the

homeoviscous adaptation of the monolayer form lipid

Effect of the Transmembrane Alkyl Chains on the Structure and Dynamics
of Membrane.

In annealed dynamics, the energy of the structure is minimized gradually without
allowing it to be trapped in a conformation that has a lower energy that nearby
conformations but higher energy than more distant conformations (i.e. a local energy
minimum). The structure with the lowest potential energy from the annealing cycle is
minimized after the end of the cycle and then written to a trajectory file. Figure 13 shows
reminimized structures after the annealed MD simulations (5 annealing cycles) for model
membranes of the pure bilayer type (A) and 50 % bipolar monolayer type (B). The typical
bilayer phase of the membrane is shown in (A). Here, the width of the bilayer is around 38
A, the typically observed value. In contrast, the (1)-l linked systems of the bilayer is much
more compact and significantly more narrow (around 24 A). Figures 14A and B show
views of the two membrane model systems using different representations (ball-stick model
for A and vector model for B). Both structures indicate a clear contraction of the membrane
in going from the separated to linked forms. Approximately a 30 % contraction of the
membrane width was observed in going from the separated to cross-linked model

membranes.

216

Figure 11. Fluctuation amplitude of the atomic distance of the lipids
during the MD simulations at 1 ps (A), 5 ps (B) and 10 ps (C). Transition

to the monolayer from the bilayer made an important contribution for lowering the

amplitude of the motional fluctuation.

 

217

 

 

 

on

I2

lo—

I...

 

33 :2: 33¢ 85:5

 

 

 

 

 

on
D
of; .. 2
.0!
of.
IO/
0 ..2
0\M\.o t n
_ _ o
u m s 0
AM; £23302. 85:5
8

Monolayer format-1 linked form)

Bilayer form

 

 

ca

IS

IS

In

 

35

.
3

A «J cos-30:: 85:5

 

IS“
10"
5

Carbon number

Figure 11

 

218

 

 
   

 

 

20
+ bilayer form
16 "
.......o........ monolayer form (to-llinked form)
A
°$ 12-
t:
o
'::r
.5
a 8 ‘
C:
U)
2 MN
M
4 ..
0 I I I I I I I I I I I

 

0 l 2 3 4 5 6 7 8 9 10 11

Time (ps)

Figure 12. RMS distance fluctuation of the lipid based on the initial lipid

coordinates (0 ps) as the reference.

219

Figure 13. Computer generated pictures of the reminimized structures after
the annealed MD simulations (5 annealing cycles) before (A) and after (B)

the transition from the bilayer to monolayer form. Typical bilayer phase (Liquid

crystalline phase) of the membrane was shown in (A).

3
l
e
r
u
.m:
F

 

221

Figure 14. Comparison of the two structures before (right) and after (left)
the transition to monolayer in different models (in ball-stick model for (A),

vector model for (B)). It shows the clear contraction of the membrane after the

transition.

Figure 14

 

223

Figure 15A shows the energy-minimized structures of the two connected and two
unconnected lipids. The head-to-head distance of the two opposed lipids with the linked
acyl chains is significantly smaller than is the case when the lipids are unconnected. This
contraction serves to increase the chain-chain interaction (van der Waals interaction)
between the tails of the two unconnected lipid chain. There is, therefore, an increased
electronic density at the middle of the lipid bilayer (Figure 16). This would also lead to
increased interdigitation of the alkyl chains at the middle of the bilayer. Since the Keough’s
group proposed the interdigitation of the acyl chains of the lipids for membrane packing
(20), much research has been performed to prove that interdigitations of the natural
membranes could be a general and significant phenomenon (21, 22, 23). Interdigitation of
the acyl chains is due to the motional dynamics of the electronic distribution of the acyl
chains of the lipids. Figure ISB shows the portions of the monolayer lipids inside the
membrane obtained after the transition. Inter- and intradigitation of the acyl chains were
clearly shown. The total change of the membrane shape, i.e., contraction of the membrane
can be explained by the enhanced inter- and intradigitation of the acyl chains. Figure 17

shows the progressive intra- and interdigitation of the acyl chains of the membrane.

224

Figure 15. Inter- and intradigitation of the acyl chains shown in the
transmembrane structures obtained after the annealed dynamic simulations.
(A) A computer generated picture of the energy-minimized structures of a single lipid
before (left) and after (right) the transition. The lipid with the linked acyl chains made an
initial contraction of its size. (B) A computer generated picture of the portions of the
monolayer lipids inside the membrane. This transmembrane structure was obtained after the
annealed dynamic simulations. Intra- and interdigitation of the acyl chains were shown.

The total change of the membrane shape, i .e., contraction of the membrane was due to the

enhanced inter- and intradigitation of the acyl chains.

Figure 15

 

226

W

Electron density

distance

 

 

 

Figure 16. Electronic distributions of the lipids before and after the
transition to the monolayer form. The change of electronic distribution of the middle
of the lipids would be expected. This changed electronic distribution in the middle made a
serious effect on the electronic interactions such as van der Waals interaction (chain-chain

interactions).

227

 

 

llllllllllll t

 

 

 

 

 

 

 

I
Interdigitation
:2 of acyl chalns
‘ g
intradigitation
l , of the monolayer
l acyl chain

 

I W Wm

Figure 17. Progressive intra- and interdigitation of the acyl chains in the

 

 

 

transmembrane during the simulated annealing calculations. Contraction of the
membrane after the transition to monolayer could be explained by intra- and interdigitation

of the acyl chains.

228

C O N C L U S IO N S

The effects of transmembrane (to-1 linked hydrocarbon chains) on the molecular
motions of model systems from the free hydrocarbon chains to whole membrane model (64
lipids) were studied using molecular mechanics and molecular dynamics simulation
methods. RMS (Root Mean Square) distance fluctuation and angle parameters were used
for the indices for the description and characterization of the molecular motions. The
molecular motional parameters of the lipid in the typical bilayer form and monolayer form
(to-1 linked form) were compared during a 10 ps dynamic simulation. A dramatic decrease
in the degree of the motional freedom was observed on going from the bilayer to the bipolar
monolayer form. This change of the motional dynamics appears to induce the favorable
interdigitation of the tails of the chains thus enhancing van der Waals attraction.
Simulated annealing method was used for the general prediction of the membrane structures
before and after the transition from the bilayer to monolayer systems. After the transition to
the monolayer form, the interdigitation of the acyl chains of the membrane resulted in
compaction of the membrane structure. These findings have special significance in light
of the recent discovery of the structural transition from bilayer to bipolar monolayer in

Sarcina ventriculi.

229

REFERENCES
l. Jung, 8., Lowe E.S., Hollingsworth I. R ., and Zeikus, J. G. (1993) J. Biol. Chem.
268, 2828-2835
2. Jung, 8. and Hollingsworth I. R .(1993) in preparation (JBC submitted)
3. Jung, S. and Hollingsworth I. R .(1993) in preparation (JBC submitted)
4. Ingram, L. O. (1982) J. Bacteriol. 149, 166-172
5. Ingram, L. O. (1990) Crit. Rev. Biotechnal. 9, 305-319
6. Ingram, L. O., and N. S. Vreeland. (1980) J. Bacteriol. 144, 481-488
7. Schreir, S., Polnaszek, C. F., and Smith, I.C. ( 1978) Biachim. Biophys. Acta 515,
395- 436
8.Campbell, ID, and Dwek, R.A. (1984) In Biological Spectroscopy
Benjamin/Comming Publishing Co., Menlo Park, California
9. McCammon, J .A., and Harvey,S.C. (1987) In Dynamics of Proteins and Nucleic
Acids, Cambridge University Press, Cambridge
10. Brooks, C.L. IH, Karplus, M., and Pettitt, M. (1988) Adv. Chem. Phys. LXXI
11. Scott, H.L. (1984) Chem. Phys. Lett. 109. 570 - 577
12. Raghavan, K., Rami, R.M.and Berkowitz, M.L. (1992) Langmuir 8, 233 -240
13. Kirkpatrick, S., Gelatt, CD, and Vecchi, MP. (1983) Science 220, 671- 680
14.Bounds, D.G. (1987) Nature 329, 215 - 219
15. Mayo, S.L., Olafson, B.D., and Goddard, W.A. III (1990) J. Phys. Chem. 94,
8897- 8909
16. Gasteiger, J ., and Marsili, M (1980) Tetrahedron 36, 3219 - 3226
17. Fletcher, R., and Reeves, CM. (1964) Comput, J. 7, 149 -156
18. Verlet, L. Phys. Rev. 1967, 159, 98 -105
19. Gennis, RB. (1989) In Biamembrane (Cantor, C.R Eds.) ppl66-198, Springer-
Verlag, New York, NY

230
20. Keough, K. M.W. and Davis, P.J. (1979) Biochemistry 18, 1453-1459
21. Huang, C. and Mason, J.T. ( 1986) Biachim. Biophys. Acta 864, 423-470
22. Mattai, J., Sripada, PK. and Shipley, G.G. ( 1987) Biochemistry 26, 3287-3297
23. McGibbon, L., Cossins, A.R., Quinn, P.J., and Russell, NJ. (1985) Biochim.
Biophys. Acta 820, 115-121