‘ ulna 115?.
l? 4*;“fi1mtfifi
" “ " "’ ”it‘fififl'éif

r ' I: ‘ ‘ . i' - . 5' [4325' .i v
_ u :‘Wfi 1' :f ; .' " 3 .. m!" {3131:51.

E“. _‘ N; _ . _ ..§.\.'~‘.-¢nl';..~:’”"
' ”1" " 't.‘ -:-- -.1.
mm . w «w
’ w 4 .'
' '. .
‘A

L5

J

4 . , a"?

.1 .
w w: . .‘. .
. I. ,

.

- ....__. .. up.-.
V “'<~..:""~
'6‘ .f harm-u:

'"P'n _.- -
. 43...")

 

'v.v\~4 ,,.: .
, ._
44?.“ _

 

 

w
‘5‘:'
I

.qJJIY:
r. 'A .. I

“-7 4—.

' mac?”
‘ J .z..

m
.1.

u

.35"
( ~
~97

A~v-vo

1:";
. #9

4m.
. ’1

L17":

‘ r ,_
‘ o' ? iv
‘ L-‘u‘ffi !,

 

.

 

-‘——.——d—-.
; ‘ 77:29.4

3‘55:

#:935-

: 119 L

"I; ‘L ‘73.;
74.11'1113-17- T4,
.‘Hfifinflfi ‘

 

THESTS

Z)

IllllllllllllllllllllltlllllllIlllllllllllllllllllllll

3 1293 016826

This is to certify that the
dissertation entitled

GENERAL PRINCIPLES OF MEMBRANE FUNCTIONS
IN ADAPTATION AND ADAPTABILITY OF BACTERIA
TO EXTREME ENVIRONMENTS

presented by

Carol Ann Mindock

has been accepted towards fulfillment
of the requirements for

PI”! . D . degree in mm

\A l/i 2W

Melli- profest r

 

Date O‘ILU fir

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

LIBRARY

M‘Chlgan 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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1193 WpGS-QM

 

GENERAL PRINCIPLES OF MEMBRANE FUNCTIONS IN ADAPTATION
AND ADAPTABILITY OF BACTERIA TO EXTREME ENVIRONMENTS

By

Carol Ann Mindock

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Biochemistry

1998

ABSTRACT

GENERAL PRINCIPLES OF MEMBRANE FUNCTIONS IN ADAPTATION
AND ADAPT ABILITY OF BACTERIA TO EXTREME ENVIRONMENTS

By

Carol Ann Mindock

Based on the hypothesis that membranes function to propagate the available thermal
energy spectrum to the cell interior, using an Arctic permafrost bacterium we further confirm
that there is a regular (mathematical) relationship between the available thermal energy
spectrum and the membrane acyl chain length distributions. These acyl chain length
distribution curves are Gaussian, in which the mean and variance are determined by
temperature, indicating that at least for Gram-positive organisms the acyl chain length
selection process is random. The relationship between the variances of the distribution
curves and their respective growth temperatures are shown to be linear over the broad
permissible growth range of the organism with deviation from linearity at the extremes of
its growth temperature limit. We show that this relationship is general for Gram-positive
organisms using data from the literature. This suggests that knowing the temperature, one
can predict the fatty acid composition over a temperature range for a particular bacterium.

We also show that effects of chemical perturbants, such as solvents and antibiotics,
alter the Gaussian acyl chain distribution curves in the same manner as increasing or
decreasing temperatures depending on whether they have a fluidizing or rigidifying effect on
the membrane, indicating that a temperature could be used to allow the bacterium to grow

in the presence of normally unfavorable solvent concentrations.

Since heat is the driving force for all cellular chemical processes, it was of interest
to determine whether temperature affects the balance of metabolic processes in the cell and
how this relates to membrane structure. NMR and l4C-labeling experiments show that the
Arctic permafrost bacterium utilizes glucose much more slowly at 4°C than at 24°C. This
bacterium makes primarily branched chain fatty acids, however, an increase in straight chains
(in particular n-Cm) is seen as the growth temperature increases from 4°C to 30°C. This is
consistent with the increased utilization of glucose at higher temperatures.

Here we show that Arctic permafrost strain 45-3 does not respond to near freezing
conditions by synthesizing so called compatible metabolites. Rather the adaptations seen are
changes in cell surface compositions, namely alterations in membrane acyl chains, the change
to nonpolar headgroups, and a reduction in the amount of polysaccharide capsule.

At the other extreme of temperature, we further characterize the ability of Sarcina
ventriculi to form transmembrane fatty acids in response to any fluidizing effect on its
membrane. We show that exogenously added fatty acids can be joined tail—to-tail at their (o-l
positions to fatty acids in S. ventriculi membrane preparations in the presence of a low
molecular weight ketone. This coupling is independent of the acyl chain headgroup as
shown by the coupling of exogenously added n-heptane and hexadecylpyridinium chloride.
The process is suppressed if cells are cultured in deuterium oxide indicating that the trigger
is a kinetic factor related to motional dynamics. These results and results from NMR
spectroscopy indicate that the process is regulated by motional dynamics and occurs via a
radical mechanism in which a hydrogen atom is extracted from each acyl chain and the two
acyl radicals join to form a new carbon-carbon bond. Further work is being done to isolate

possible enzymes involved in this mechanism.

Dedicated to my parents for all of their love and support.

iv

ACKNOWLEDGIVIENTS

In the many years it has taken me to accomplish this goal, I have had a great deal of
support and encouragement from a lot of wonderful people whom I wish to thank.

First and foremost, I wish to thank my mentor and friend Dr. Rawle I. Hollingsworth.
I feel privileged to have learned science and research under his guidance. He has taught me
to think beyond the current views of science by applying fundamental concepts of math,
physics, and chemistry to living systems. His continual support and encouragement as well
as his friendship have been instrumental in helping me reach this goal.

I also want to thank the members of my committee: Dr. William Wells, Dr. Estelle
McGroarty, Dr. J. Gregory Zeikus, and Dr. Melvin Yokoyama for their helpful advice and
suggestions regarding this dissertation project.

I want to thank past and present members of the Hollingsworth group for all of their
help, advice, support, and for making the lab such a fun place to work. I have alsoenjoyed
both the scientific and cultural discussions we have had.

I would like to thank Beverly Chamberlin in the Mass Spectrometry Facility for her
help in running the 505 GC/MS. I would also like to thank Dr. Long Le and Dr. Kermit
Johnson in the NMR Facility for their assistance in setting up numerous NMR experiments.

In my years here I have been fortunate to have met some very great friends. I would
especially like to thank Carla, Brian, and Casey Swan; Scott Herr; Kim and Greg Katch; and

V

Upal and Tracy Sengupta. Thank you so much for all of the fun times. Because of you I
have many special memories of my years here.

Most of all I wish to express my sincerest love, thanks, and appreciation to my
family: my parents, Jean and Jim; and Jim, Brenda, Joe, Katurah, Daniel, Stephanie, and
Jennifer; and my grandparents, J can and Del Hahn. My accomplishments are a result of your

continual unconditional love, support, and encouragement.

vi

TABLE OF CONTENTS

LIST OF TABLES ..................................................... ix
LIST OF FIGURES ..................................................... x
LIST OF ABBREVIATIONS ........................................... xiv
CHAPTER I
OVERVIEW OF GENERAL BACTERIAL ADAPT ATIVE MECHANISMS . . . . 1
Membrane Structure and Organization ................................. 3
General Theories on the Roles of Membranes in Adaption .................. 9
Molecular mechanisms of adaptation ................................. 15
Membrane alterations ........................................ 15
Protein stability and function .................................. 24
Osmotic balance ............................................ 28
Special issues in subzero temperature adaptation ........................ 3O
Salt/metabolite concentrations and oxygen availability .............. 31
Prevention of freezing of intracellular water ...................... 32
Maintenance energy ......................................... 32
Carbon source and heat generation ............................. 34
The universality of adaptive responses to various environmental
perturbations .................................................... 35
References ...................................................... 37
CHAPTER II
DEVELOPMENT OF A THEORY ON THE SIGNIFICANCE
OF FATTY ACID HETEROGENEITY ................................... 45
Abstract ........................................................ 46
Introduction ..................................................... 47
Materials and Methods ............................................. 54
Results and Discussion ............................................ 57
References ...................................................... 86

vii

CHAPTER III
THE INTERRELATION BETWEEN TEMPERATURE-INDUCED SHIFTS IN

CARBON METABOLISM AND MEMBRANE LIPID CHEMISTRY .......... 89
Abstract ........................................................ 90
Introduction ..................................................... 9 1
Materials and Methods ............................................. 94
Results and Discussion ............................................ 95
References ..................................................... 106

CHAPTER IV

THE ADAPTATION OF ARCTIC BACTERIA TO LOW PRE-FREEZING
TEMPERATURES INVOLVES MEMBRANE ALTERATIONS BUT NOT

INCREASED PRODUCTION OF OSMOLYTES .......................... 108
Abstract ....................................................... 109
Introduction .................................................... 109
Materials and Methods ............................................ 111
Results ........................................................ 1 17
Discussion ..................................................... 142
References ..................................................... 149

CHAPTER V

CHARACTERIZATION OF A MECHANISM OF ADAPTATION OF A pH AND

SOLVENT RESISTANT BACTERIUM Sarcina ventriculi .................. 150
Abstract ....................................................... 151
Introduction .................................................... 152
Materials and Methods ............................................ 156
Results and Discussion ........................................... 159
References ..................................................... 1 80

CHAPTER VI

SUMMARY AND PERSPECTIVES ..................................... 182
References ..................................................... 188

APPENDIX A

PREPARATION OF THE RADIOLABELED PHOTOAFFINITY PROBE FOR

IDENTIFYING PUTATIVE CROSS-LIN KIN G ACTIVITY ................ 189
Preliminary Results .............................................. 193
References ..................................................... 195

viii

LIST OF TABLES

Table 3.1: Percent of fatty acid methyl ester derivatives in the total methanolysate of
the cells grown at 4°C, 14°C, 24°C, and 30°C in order of elution. . . . . 96

Table 4.1: Percent of fatty acid methyl ester derivatives in the total methanolysate of
the cells grown at 4°C and 24°C in order of elution. .............. 141

ix

Figure 1.1:

Figure 1.2:
Figure 1.3:
Figure 1.4:
Figure 1.5:

Figure 2.1:

Figure 2.2:

Figure 2.3:

Figure 2.4:

Figure 2.5:

Figure 2.6:

LIST OF FIGURES

Representation of the cell envelopes of (A) Gram-positive bacteria and (B)

Gram-negative bacteria. ....................................... 5
Structures of some common classes of membrane lipids. ............. 6
Lipid polymorphism in the liquid crystalline state. ................. 13
Structures of novel lipids from bacteria. ......................... 21
Simplified pictures of the motional spectrum of lipids ............... 23

Diagram depicting the transfer of the thermal energy spectrum from the
environment via longitudinal accordian modes of membrane acyl chains to
the cell interior. ............................................ 49

Gas chromatographs of fatty acid methyl esters obtained from Arctic strain
45-3 cultured at (A) 4°C and (B) 24°C. ......................... 58

Plot of probability distributions at 4°C, 14°C, 24°C, and 30°C for Arctic
strain 45-3 ................................................ 60

Plots of probability distributions at several growth temperatures for (A)
Listeria monocytogenes (B) Bacillus amyloliquefaciens and (C) Bacillus
subtilis. ................................................... 62

Plot of probability distributions at 60°C, 70°C, and 80°C for Bacillus
caldolyticus. ............................................... 65

Plots of the variances of the acyl chain distribution curves versus growth
temperature (in degrees Kelvin) and the equations for the curves for (A)
Arctic strain 45-3, (B) L. monocytogenes, (C) B. amyloliquefaciens, and
(D) B. subtilis .............................................. 66

Figure 2.7:

Figure 2.8:

Figure 2.9:

Figure 2.10:

Figure 2.11:

Figure 2.12:

Figure 2.13:

Figure 3.1:

Figure 3.2:

Figure 3.3:

Figure 3.4:

Figure 3.5:

Figure 3.6:

Gas chromatograms of fatty acid methyl esters obtained from Arctic strain
45-3 cultured in the presence of two solvents. (A) 24°C control, (B) 0.1
M cyclohexane, and (C) 0.1M 1,4-cyclohexadiene ................. 69

Computer generated pictures showing the packing interactions between a
straight chain hydrocarbon and cyclohexane and an w-cyclohexyl acyl
chain from two different views (A) and (B). ...................... 72

Structures of the chemical perturbants used studies with Bacillus subtilis
PY79. .................................................... 75

Gas chromatogram of fatty acid methyl esters of Bacillus subtilis PY79
from a 37°C culture. ........................................ 76

The effects of cis- and trans-decalins on the plot of the probability
distributions of Bacillus subtilis PY79. .......................... 77

Effects of three antibiotics on the plot of the probability distributions of
Bacillus subtilis PY79 ........................................ 79

A three dimensional graph showing hypothetical growth rate curves for
conditions of high and low temperatures and solvent (fluidizing type)
concentrations ............................................. 8 1

Hypothetical energy profiles for two different metabolic pathways are
shown, a higher energy requiring profile in (A) and a lower energy
requiring profile in (B). ...................................... 92

1H NMR of TSB media before and after growth of Arctic strain 45-3. . . 98

TLC separation of l4C-glucose, l4C-valine, and l4C-leucine for one time
point and its corresponding analog profile indicating the relative intensities

of the three spots. .......................................... 100
The extent of utilization of 14C—glucose by Arctic strain 45-3 at 4°C and

24°C. ................................................... 101
Total protein profiles for Arctic strain 45-3 ...................... 103

Schematic representation of the relationship of the balance between two
types of cellular metabolism, the potential thermal or enzymatic control of
the metabolic processes, and the pools of available acyl-CoA primers for
fatty acid biosynthesis. ...................................... 104

xi

Figure 4.1:

Figure 4.2:

Figure 4.3:

Figure 4.4:

Figure 4.5:

Figure 4.6:

Figure 4.7:

Figure 4.8:

Figure 4.9:

Figure 4.10:

Figure 4.11:

Figure 4.12:

Figure 4.13:

Figure 4.14:

Figure 5.1:

Siberian permafrost bacterial sampling region in the tundra zone of north-
eastern Russia, Kolyma-Indigirka Lowland (152—162°E and
68-72°N). ................................................ 112

Electron micrographs (at the same magnification) of Arctic strain
45-3. .................................................... 1 18

Fluorescent micrographs using RIT C optics (at the same magnification) of
Artic strain 45-3 stained with 0.1% acradine orange attached to silanized
coverslips. ............................................... l 19

1H N MR spectra of the cytoplasmic metabolites. ................. 121

Two dimensional DQF-COSY spectrum of the cytoplasmic metabolites
from cells cultured at 4°C. ................................... 123

1H N MR spectra of some common osmolytes. (A) glucose (B) glycerol
(C) mannitol (D) sorbitol. ................................... 125

Gas chromatograph profiles of the capsular polysaccharides from the
bacteria cultured at (A) 4°C and (B) 24°C. ...................... 127

1H N MR spectra of the total lipids extracted from the bacteria grown at
(A) 4°C and (B) 24°C ....................................... 131

1H NMR spectra of the lipid headgroups from the bacteria grown at (A)
4°C and (B) 24°C. ......................................... 133

1H NMR spectrum of the methanol insoluble component from cells
cultured at 4°C. ........................................... 136

Positive ion electrospray mass spectrum of the triacylglycerol from cells
grown at 4°C. ............................................. 137

Gas chromatographs of fatty acid methyl esters obtained from bacteria
cultured at (A) 4°C and (B) 24°C .............................. 139

SDS-PAGE gel of total proteins of lysates from cells cultured at 4°C and
24°C. ................................................... 143

Model showing the solvation of the region between two bilayer
membranes. .............................................. 148

Conceptual model for the coupling of alkyl chains of the lipid components
by an integral membrane catalytic system. ...................... 154

xii

Figure 5.2:

Figure 5.3:

Figure 5.4:

Figure 5.5:

Figure 5.6:

Figure 5.7:

Figure 5.8:

Figure 5.9:

Figure A.1:

Figure A2:

The electron impact mass spectrum of one of the chimeric bifunctional
fatty acid dimethyl esters formed by tail-to—tail coupling between an
exogenously added fatty acid (heptadecanoic acid) and a hexadecanoic
acid residue from an S. ventriculi lipid component. ............... 161

The electron impact mass spectrum of a 23 carbon fatty acid methyl ester
resulting from coupling of n-heptane to a hexadecanoic acid residue from a
S. ventriculi lipid component at the w-l position ................. 163

1H NMR spectra of (A) total fatty acid methyl esters from S. ventriculi
cultured with hexadecylpyridinium chloride, and (B) isolated fatty acid
methyl esters containing hexadecylpyridinim coupled to native S.
ventriculi membrane lipids. .................................. 165

Bacterial (proposed) and synthetic mechanisms of the reduction and
subsequent oxidation of the pyridine ring of hexadecylpyridinium
chloride. ................................................. 168

Gas chromatogram of fatty acid methyl esters derivatized from total lipid
extracts of S. ventriculi in culture at pH 9.7 ..................... 169

Gas chromatographs of fatty acid methyl esters from S. ventriculi cells
cultured in (A) H20 and (B) D20. ............................. 172

Results from radical trap/scavenger experiments indicating the percent of
tail-to-tail coupling of S. ventriculi membrane lipids normalized to the
50°C controls. ............................................ 175
1H NMR spectra showing the effect of cyclohexane and 1,4-
cyclohexadiene on the lipid packing arrangement of dipalmitoyl
phosphatidylcholine. ....................................... 177
Synthesis of the l4C-labeled photoaffinity probe. ................. 194

Phosphoimage of the nylon membrane containing proteins from several G-
75 column fractions ........................................ 194

xiii

ACP
DGDG
DPG
DPPC
DQF—COSY
DSC
DTA
ESR
FAME
GC
GC-MS
1H NMR
LAM
MeOH
MGDG
PA

PC

PE

LIST OF ABBREVIATIONS

Acyl carrier protein

Diglucosyldiglyceride

Diphosphatidylglycerol (cardiolipin)
DL—a-phosphatidylcholine, dipalmitoyl

Double quantum filtered-correlated spectroscopy
Differential scanning calorimetry

Differential thermal analysis

Electron spin resonance spectroscopy

Fatty acid methyl esters

Gas chromatography

Gas chromatography-mass spectrometry

Proton nuclear magnetic resonance spectroscopy
Longitudinal accordian modes

Methanol

Monoglucosyldiglyceride

Phosphatidic acid

Phosphatidylcholine

Phosphatidylethanolamine

xiv

PG Phosphatidyl glycerol

PI Phosphatidylinositol

PS Phophatidylserine

RITC Rhodamine iso-thiocyanate

SDS-PAGE Sodium dodecylsulfate-polyacrylamide gel electrophoresis
TAG Triacylglycerol

TLC Thin layer chromatography

TOCSY Total correlated spectroscopy

TSB Trypticase soy broth

XV

CHAPTER I

OVERVIEW OF GENERAL BACTERIAL ADAPTATIVE MECHANISMS

Bacteria, as a class of organisms, occupy almost all the environmental niches known
on earth including many that are man-made, from extremes of temperature, pH, pressure,
salinity, and solvent concentrations. A key to adaptation in the face of such vastly different
circumstances is the maintenance of a functioning biochemical machinery and structural
integrity. What are the most important physical properties bacteria must maintain to survive?
Are there common adaptative mechanisms the various species have developed to survive
more than one type of environmental condition? Or are there different mechanisms used

even among species adapted to the same types of environments?

One characteristic bacteria must maintain is membrane integrity. However, it is not
enough just to keep the membrane intact to ensure survival to extreme environments or to
variations in those niches. It must also function under the new conditions. As single-celled
organisms, the membrane in bacteria is in constant contact with its environment, and as such
may serve as the primary sensor to environmental fluctuations. The membrane may also
serve to transmit the available thermal energy (determined by temperature) into the cell via
certain vibrations of the acyl chains [1]. Membrane fluidity is another important
consideration. It is commonly known that bacteria alter their lipid and fatty acid composition
as a result of various environmental perturbations. One theory for the purpose of these
alterations is to maintain the same degree of membrane fluidity, termed "homeoviscous
adaptation" [2]. Others argue that maintaining a certain amount of the necessary bilayer type

phase is more important, termed "homeophasic adaptation" [3].

Other characteristics are also important to cell survival such as maintaining
protein/enzyme stability and function; and altering the balance of metabolic pathways to
utilize available nutrients and generate needed energy. These factors also play a role in
determining membrane structure and composition. This literature review will focus on
theories of membrane functions in adaptation and adaptability of bacteria to extreme
environments as well as common molecular mechanisms of adaptation used by various types
of extremophiles. Since much of the work in this dissertation was done using a Siberian
permafrost bacterium, a section of this review focuses on special issues in subzero

temperature adaptation.

Membrane Structure and Organization

All cells require at least one membrane to contain the cytoplasm and all cellular
functions. Procaryotes, the simplest of all organisms, contain membranes only in their cell
envelope. The peripheral membrane(s) is not only a boundary for the cell, but also serves
as a semipermeable barrier to the outer environment as well as a location for many cell
processes to take place. While the composition of membranes is complex and varies greatly
between organisms, the basic structure is still considered to be the fluid mosaic system
described by Singer and Nicolson [4]. This model depicts membranes as containing integral

and peripheral proteins embedded in a fluid-like matrix consisting of a bilayer of lipids.

In addition to the cell envelope membrane(s), procaryotes usually also contain a rigid

cell wall composed of the macromolecule peptidoglycan. In Gram-positive bacteria this cell

wall lies outside the cytoplasmic membrane. In Gram-negative bacteria the cell wall lies
between the inner (cytoplasmic) and outer membrane (Figure 1.1). Peptidoglycan is a
polymer composed of glycan strands cross—linked by peptides. The glycan strands are usually
made of alternating units of N-acetylglucosamine and N-acetylmuramic acid which are B- l ,4-
linked. The short peptide chain is linked by its amino terminus to the carboxyl group of
muramic acid and contains L and D amino acids. These peptide chains are cross-linked by
interpeptide bridges enabling the formation of a rigid macromolecule that surrounds the cell.
While there is a high degree of variability in the peptidoglycan structures among various
procaryotes especially in the peptide structures, there are relatively few cases of phenotypic
alterations when a bacterium is cultured under different growth conditions. When structural

alterations do occur, it is usually when certain nutrients are limiting factors for growth [5].

The lipid composition of membranes, however, is highly variable not only among cell
types but also due to numerous environmental perturbations. This diversity is due to
variations in both the polar head groups and the fatty acyl chains. The most common
membrane lipids are shown in Figure 1.2. The membranes of eucaryotes primarily consist
of glycerol-based phospholipids, including phosphatidylcholine (PC),
phosphatidylethanolamine (PE), phosphatidylinositol (PI), phosphatidylserine (PS), and
cardiolipin. Another major class of eucaryotic membrane lipids are the sphingosine-based
lipids, including sphingomyelin and glycosphingolipids. Glycolipids, both glycerol- and
sphingosine-based, often function as cell-surface associated antigens and recognition factors

in eucaryotes. Eucaryotic membranes, especially mammalian plasma membranes, contain

 

 

oo‘ncfi-‘—--
/

b/ 5

j E

 

 

} cauwm

 

 

 

Figure 1.1: Representation of the cell envelopes of (A) Gram-positive bacteria and (B)
Gram-negative bacteria. (adapted from ref. 94)

O I 0"
,0! l
W0 CH,—O-:-—OH
o 0 ‘0-Gs-Cth-NOM.
‘0—CNn-M“NH-
-O—CH.—fH—NH.
00?
OH
—o-cn.-i:u—cu.ou
-o‘ p
0 OH
m.
’ OH
W\
,3? +
o a4.—o—r--o-a1.-ai.-m
o“ w-oa-as-m.
O
“\leY x?“ 0—
We ‘9‘!“0‘7‘0
o O
0‘ 7H.
0’ o\m_°_z_o
Ono’gag—o—E—Oxfix
0

Wm OH
3,-
M1 ‘ag-O-P-Ox+/
W I N-\-
O 0

phosphatidic acid
phosphatidylcholine
phosphatldylethanolamlne
phoephatldylserlne

phoopmudytgiyoeml

phoephatldyflnooltol

Figure 1.2: Structures of some common classes of membrane lipids.

cholesterol as well. While the eubacteria, cyanobacteria, and the archaea are typically
distinguishable by their lipid composition, a common feature among them (with few
exceptions) is that they all lack sterols. Among the eubacterial class phosphatidylglycerol
(PG), PE, and cardiolipin are the major membrane lipids. In addition, Gram-positive bacteria
also contain mono- and diglycosyldiacylglycerols. Cyanobacteria (which are photosynthetic)
typically contain PG and glycolipids. The archaea, which seem to lie phylogenetically
between procaryotes and eucaryotes, contain a special class of lipids in which the usual ester-
linked fatty acids of the above mentioned lipids are replaced with ether linked phytanyl alkyl
chains. The most common types are analogues of PG, phosphatidylglycerol phosphate
(PGP), and phosphatidylglycerol sulphate (PGS). Another degree of membrane lipid
diversity lies in the fatty acyl and alkyl structures. These structures vary in length,
unsaturation, and branching. Some have carbocyclic rings. The possible combinations
between lipid headgroups and hydrocarbon tails easily allow for hundreds of different lipid
molecules to make up the membrane lipid matrix. The lipid species mentioned above are
just the most common lipids found in membranes. Many organisms make other unusual

lipids, especially those organisms that live in extreme environments.

Another feature of biological membranes that may serve an important function in
adaptation and adaptability is the high degree of lateral organization. Experiments using
several techniques show that membrane lipids (and proteins) partition into many phases or
domains. These domains are either gel or liquid—crystalline in nature [6]. The driving forces

for phase formation are not well understood, though some possibilities are the minimization

of free energy and geometric packing constraints. Free energy minimization would be
achieved by organizational features that minimize the void spaces, the interaction of acyl
chains with the aqueous phase, and the gaucheztrans ratio of the carbon-carbon bonds in the
acyl chains for the lipid molecules in a domain. Free energy minimization would also be
achieved by maximizing the interchain interactions and by specific molecular associations

that give rise to phases of different symmetry [7].

The molecular geometric packing of lipids may also be a driving force for phase
separation. The size of the head group and degree of unsaturation of the acyl chains leads
to three dimensional lipid structures that are cylindrical, conical, or inverted conical shaped
[8]. In addition the 2-fatty acid chain of phospholipids initially extends perpendicularly to
the glycerol backbone, but it bends back at the second carbon atom to become parallel to the
first acyl chain. This nonequivalence of the acyl chains leads to tilting of some
phospholipids [9]. As a result phospholipids with different tilts and/or different acyl chain
lengths are more prone to separate into phases such that packing defects are plated out to the

domain boundaries [7].

The enormous membrane lipid diversity and the organization of membrane lipids
into different domains are two characteristics of membranes that enable procaryotes (and
simple eucaryotes) to survive in constantly changing environments. As will be discussed
further, bacteria alter their membrane lipid composition in response to numerous

perturbations. The sensory mechanism of these perturbations, though, may be triggered by

disruptions in the membrane phases.

General Theories on the Roles of Membranes in Adaption

Biological membranes exist in a phase transition between a gel state, in which the
lipid hydrocarbon tails are primarily in an all-trans configuration, and a liquid—crystalline
state, in which the hydrocarbon tails contain a number of gauche conformers. The motional
dynamics (fluidity) of these lipids is critical to the numerous functions carried out by the
membrane. A large proportion of membrane lipids needs to be in the liquid-crystalline
phase to maintain transport and certain enzymatic functions. Studies with Escherichia coli
and Acholeplasma laidlawii indicate that growth of these organisms is severely impaired
when more than 50% of their membrane lipids are in the gel phase [10]. Perturbations that
alter membrane fluidity often result in a rapid alteration of lipid acyl chains and/or the
modification of the lipid head groups. Typically shorter and more unsaturated acyl chains
are incorporated when perturbations cause the membrane to become more gel-like; while
longer and more saturated acyl chains are made in response to fluidizing effects. Other
modifications include acyl chain branching and cyclization. All of these modifications affect

membrane fluidity.

One theory for the alteration of membrane lipids is to maintain the same degree of
fluidity regardless of growth temperature (and possibly due to other variations in the
environment). This theory, termed homeoviscous adaptation, was first described for E. coli

using electron spin resonance spectroscopy (ESR) [2]. ESR was also used to demonstrate

10

homeoviscous adaption in two other organisms, A. laidlawii [1 l] and Bacillus
stearothermophilus [12]. All three organisms are able to alter their membrane lipid
composition in response to temperature changes to maintain a virtually constant membrane
fluidity. Other organisms also conserve membrane fluidity, though it does not always seem
to be a complete compensation over the growth temperature range for particular organisms.
Cossins and Sinensky proposed a coefficient of homeoviscous efficacy to quantify the
compensatory response. The efficacy was calculated as the ratio of the change in membrane
order to the difference in growth conditions (temperature, pH, etc.). An efficacy value of 1.0
indicates complete compensation of membrane fluidity, while values between 0 and 1.0
indicate partial fluidity compensation for the different conditions. The efficacy for
procaryote membranes varies between 0.2 and 1.0, while the efficacy for all eucaryotes is less
than 0.5 [13]. The large variation in efficacy values may be due in part to the techniques

and/or probes used.

Membrane fluidity is related to the properties of its composite lipids. The lipid
bilayer, however, is not a typical liquid and as such does not behave like a bulk fluid. For
typical liquids, like water, fluidity is defined as the reciprocal of viscosity. Membranes have
different characteristics which also must be taken into consideration when trying to
determine fluidity. For instance membranes are microscopic fluids with stringent boundary
conditions and are essentially two-dimensional. Also, while typical fluids have distinct
phases with regards to temperature (solid, liquid, or gas), membranes have numerous phases

that can exist simultaneously (solid and liquid-crystalline).

11

Due to these special characteristics of bilayers, "membrane fluidity" is not very well
defined and thus is hard to quantify. Typically membrane fluidity refers to the disorder of
the acyl chains (structure) and/or their ease of movement (dynamics). Therefore,
quantization of membrane fluidity encompasses several parameters such as diffusion
(translational and rotational), order, packing, and permeability. These parameters are related,
though how is not clearly understood [14]. As a result different techniques and different
spectroscopic probes may measure different membrane parameters--yielding widely varied
results in the measurement of membrane fluidity. Differential scanning calorimetry (DSC)
and differential thermal analysis (DTA) are nonperturbing methods used to measure
biomembrane phase transitions, but are not capable of direct measurement of fluidity within
a given phase state. ESR and fluorescence spectroscopy are often used to directly measure
lipid phase states and fluidity, but the probes used may partition into certain
nonrepresentative domains as well as perturb the membranes. Nuclear magnetic resonance
(NMR) techniques, such as Tl relaxation measurements, are capable of accurately monitoring
phase transitions and providing orientations and rates of motion of the lipid molecules [10].
NMR techniques are likely to provide the most accurate measurements of membrane lipid
fluidity, though currently they are not widely used. Despite the different techniques used,
it is clear that most organisms do alter their membrane compositions to maintain a constant

degree of membrane fluidity.

Another theory has been proposed for membrane lipid composition alterations based

on the highly variable homeoviscous efficacy coefficients. This theory, termed homeophasic

l2

adaption [15], suggests that it is more important for organisms to maintain a particular ratio
of bilayer to non-bilayer forming lipids. A large fraction of non-bilayer forming phases
would decrease the stability of the bilayer. The three dimensional structure of lipid
molecules depends on the size and shape of its headgroups and acyl chains. Lipids that have
a cylindrical shape (like PC, PS, and PG) would favor the typical bilayer phase. PE
containing unsaturated acyl chains, and cardiolipin or PA complexed with calcium form
conical molecular structures, which favor inverted hexagonal or cubic phases (Figure 1.3).
Lysophospholipids form inverted cones and would favor micelle type structures [8, 16].
Non-bilayer structures have been implicated in such physiological functions as adhesion,

joining, and fusion [16, 17].

In support of this theory Wieslander and coworkers proposed that the regulation of
lipid composition in A. laidlawii is governed by molecular shape. The dominate lipids in this
organism are monoglucosyldiglyceride (MGDG) and diglucosyldiglyceride (DGDG), which
have wedge and rodlike molecular shapes, respectively. The organism regulates the level of
these two lipids to maintain a balance of lamellar and hexagonal phases [18-20]. While this
theory is interesting, evidence supporting the formation of non-bilayer phases primarily
comes from experiments with single lipids, or a mixture of only a few types of lipids.
Membranes are extremely complex mixtures, and the fact that simple (few components) lipid
mixtures assume non-bilayer phases in water does not mean they will form the same

structures in a complex mixture.

13

 

 

 

 

 

 

 

 

 

Figure 1.3: Lipid polymorphism in the liquid crystalline state. (A) Lamellar phase (L,);
(B) Hexagonal inverted phase (Hn); (C) Hexagonal phase (HI); and (D) Cubic phase.
(adapted from ref. 95)

l4

Membrane acyl chain heterogeneity also serves another function in the adaptability
of organisms to their niches. As membranes completely surround cells, they are capable of
transferring the available thermal energy (which are propagated by collisions) to the interior
of cells. The heat or infrared quanta from the environment (of the magnitude "kT") is
transferred by molecular vibrations. These quanta are passed via water in biological systems
by collisions with other molecules. The acyl chains of membrane lipids serve as oscillators
to focus these quanta to the cell interior. The quanta are transferred by specific acyl chain
vibrations called longitudinal accordian modes (LAMS). By changing the lengths of the acyl
chains, the membrane is tailored to propagate the available thermal energy spectrum.
Evidence that membranes transform the available energy is shown in experiments in which
the probability distributions of the main acyl chain length series are plotted from organisms
at different growth temperatures and in the presence of D20. Probabilities of the energies
per chain are plotted. The graphs indicate a Gaussian distribution of the acyl chain lengths
and their energies, and there is a direct correlation between the variance (02) of the Gaussian
distributions and kT. As temperature decreases, kT decreases as well. The distribution of
available energies becomes more narrow as indicated by an increase in the height of the
Gaussian curve. Therefore, low temperature effects (including D20 effects) indicate cells
have a more limited set of available energies which can be utilized. This study indicates that
organisms are able to "tune" their membranes to the value of kT by adjusting the acyl chain

distributions of the membrane lipids [1].

15

Molecular mechanisms of adaptation

Procaryotes and the archaea, unlike most eucaryotes, have the ability to adapt to many
environmental extremes. Procaryotes and the archaea colonize most of the harshest
environments on earth, from extremes in temperature (hot and cold), pH (acidic and
alkaline), salinity, osmotic pressure, and they survive in the presence of many detrimental
chemical substances. To survive in all of these environmental conditions organisms must
maintain certain functions, including membrane structural and functional integrity, osmotic
balance, protein stability and function, and the ability to carry out numerous metabolic
processes. In adapting to the environment (and changes within) these organisms have certain
molecular mechanisms which they have altered (or can alter) to maintain critical cellular
functions. These alterations include changes in membrane and other cell surface structures,
osmolyte changes, protein stability and expression, and the ability to balance metabolic

processes.

Membrane alterations

Bacteria can alter their membrane structure in many ways. One way is to alter the
lipid composition, though it is not always affected by environmental changes--the most
commonly studied of which is temperature. In one study psycrophiles, and not psycrotrophs,
were found to alter their lipid composition in response to temperature [21]. In another study,
four psychrophilic Vibrio species were found to have thermosensitive conversion of PS to
PE or of PG to DPG, though the direction of the temperaure change that induced these

conversions was not the same in the four organisms [22]. On the other hand, several

16

psychrotropic organisms like Pseudomonas spp. [23] and Micrococcus cryophilus [24] do
not alter there phospholipid composition in response to temperature changes. It has been
observed for several organisms, including Tetrahymena [25], Mycobacterium smegmatis
[26], and several temperature-sensitive mutants of E. coli [27, 28], that the ratio of
cardiolipin to phosphatidylglycerol increases with temperature. The increase in cardiolipin
effectively cross-links the lipid headgroups to counteract increased membrane fluidity with
the increase in temperature. Similarly, Sarcina ventriculi also cross-links its lipid
headgroups in response to numerous perturbants that increase membrane fluidity [29]. One
strain of Bacillus stearothermophilus increases its phosphoglycolipid content when the
growth temperature is raised from 45 to 71°C [30] This result is in agreement with the trend
towards a general increase in the amount of glycolipids with increasing temperature observed
in organisms ranging from psychrophiles to moderate and extreme thermophiles [3]. In
Tetrahymena the alterations in lipid headgroups is much slower, and is suggested to be a
secondary adaptive mechanism to fine tune membrane dynamics after altering acyl chain

composition [31].

While bacteria may or may not alter their lipid headgroup composition, most
eubacteria and cyanobacteria alter their fatty acid composition in response to perturbations
in their environments. Eubacteria can alter the membrane fatty acid chain length,
unsaturation, branching, and cyclization. Cyanobacteria can only alter the chain length and
unsaturation of their acyl chains [3]. Typically changes to the membrane acyl chains occur

rapidly in response to perturbations. Shorter, more unsaturated or branched acyl chains occur

17

when bacteria are exposed to cold, high pressures, and substances that reduce the membrane
fluidity. Fatty acid cyclization may also decrease on lowering growth temperatures in Gram-
negative bacteria. This is seen in such organisms as E. coli [32, 33], P. fluorescens [34, 35],
P. halosaccharolytica [36, 37], Proteus mirabilis [38], and Serratia marcescens [39]. In
response to temperature changes the amount of branched chain fatty acids and/or the relative
proportion of anteiso versus iso-branched fatty acids is altered in Gram-positive bacteria.
Low temperatures result in an increase in the lower melting anteiso-fatty acids and
decreasing amounts of higher melting, odd-carbon, iso-fatty acids [40]. The reduction of
cyclopropyl fatty acids seen in Gram-negative bacteria as temperatures drops is probably due
to the synthetic mechanism for forming these fatty acids. The formation of these fatty acid
involves methylation by S-adenosylmethionine (SAM). The synthesis of SAM requires a lot
of energy (ATP). As will be discussed in more detail, the available thermal energy which
drives cellular processes is much less at low temperatures causing the formation of SAM to
be thermodynamically unfavorable, such that less SAM would be available at lower
temperatures. The increase in branched fatty acids in Gram-positive organisms also is most
likely due to thermodynamic alterations in the balance of cellular metabolic pathways. A
shift in metabolism at low temperatures to favor amino acid metabolism over carbohydrate
metabolism would increase the branched acyl-CoA primers available for fatty acid synthesis

causing an increase in the amount of branched lipid acyl chains.

In aquatic environments organisms must contend with hydrostatic pressure. On

average hydrostatic pressures below the ocean continental shelf are 200 atmospheres and

18

reach as high as 1100 atmospheres at the deepest part of the ocean. In addition temperatures
below 1000 m are usually less than 4°C. Hydrostatic pressure compresses lipid bilayer
causing an ordering effect on the acyl chains increasing the gel state of membranes. Hence
pressure acts similarly to low temperatures in causing a decrease in membrane fluidity and
as a result similar alterations in membrane acyl chains are observed [41]. The barophilic
marine bacterium CNPT3, most likely in the Vibrio genus, increases it's proportion of
unsaturated acyl chains particulary Cm and to a lesser extent C ,8, I at the expense of C 14,0,
Cm, and Cl 4. 1- While the amount of Cm decreases with increasing pressure this bacterium

still contains a larger percent of CW than most marine vibrios [42].

In addition to temperature and pressure, the chemical composition of the aqueous
environment surrounding cells influences cellular structures and functions. In regards to cell
membranes, a major portion of the lipids are anionic resulting in a negative surface potential
which is stabilized by cations in the aqueous phase. As a result the salinity and pH of the
surrounding aqueous phase effects the membrane surface potential and ionization of the
anionic phospholipids [41]. It has been observed that an increase in salinity results in similar
acyl chain changes cause by low temperatures and increased pressures. Namely in Gram-
negative bacteria cyclopropyl and unsaturated fatty acids increase while in Gram-positive
bacteria an increase in branched chain fatty acids is observed. The increase in cyclopropyl
fatty acids of Gram-negative bacteria must be interpreted with caution since in some strains
the major change is from unsaturated to cyclopropyl acyl chains, which would not alter the

fluidity greatly [43]. In the halotolerant (meaning no specific salt requirements for growth)

19

bacterium Staphylococcus epidermidis a large increase in anteiso-15:0 fatty acids occurs in
response to salt concentrations of 25%, which is again a similar response to low temperature
[44]. The increase in salinity would affect the polar headgroups of lipids, and cellular
membranes in general have an overall negative charge. In fact, an increase in the anionic
membrane lipid content relative to zwitterionic lipid has been observed with increased
salinity. In Gram-negative bacteria the major change is an increase in PG over PE, and
sometimes an increase in DPG as well. In Gram-positive bacteria, which usually have a
more complex lipid composition, an increase in phosphoglycolipids, PG, and/or DPG is seen
with increased salinity [43]. These changes in headgroups, however, are similar to the
fluidizing effects caused by increased temperatures. Some aspects of these trends in acyl
chain versus headgroup alterations are confusing and may arise from differences between the
systems studied and the ranges of salt concentrations used in the studies. For instance the
cations of the salts at low to medium concentrations may stabilize the negative charges of the
lipid headgroups through electrostatic interactions causing an ordering effect on the acyl
chains and rigidifying the membrane. As a result more unsaturated, cyclopropyl, and/or
branched acyl chains would offset this rigidifying effect. However, at high salt
concentrations the abundance of ions would have a chaotropic effect on the membrane. To
counteract this fluidizing effect increases in DPG (cardiolipin) would cross-link the lipid

headgroups and stabilize the membrane.

Longer, saturated and straight acyl chains occur in membranes when bacteria are

exposed to high temperatures, extremes of pH, and substances that increase membrane

20

fluidity. In addition to these general acyl chain structural changes, many bacteria have
special types of membrane lipids that allow these organisms to thrive in harsh environments
or adapt to habitats that undergo many extremes. These include, in particular, thermophilic
bacteria or bacteria that can adapt to high temperatures. Some of these specialized lipids are
shown in Figure 1.4. These lipids include the tetraether and diether lipids of the archaea.
The alkyl ether linkage is resistant to cleavage over the wide pH range encountered by these
organisms. All three classes of the archaea (extreme halophiles, methanogens, and
thermoacidophiles) have membranes composed of lipids derived from one basic core
structure, diphytanylglycerol ether. In addition, some archaea lipids contain one to four
diphytanylglycerol ether cyclopentane rings [45]. The membranes of these organisms are
very tolerant to extreme conditions. Archaea inhabit climates with temperatures above
100°C as well as niches with 36% salt content, pHs above 11 and less than 1, and pressures
as high as 1100 atmospheres in the deepest part of the ocean. [46]. The extreme halophilic
archaea contain only diether derivatives. This may, however, be a reflection of their typical
growth temperature rather than their tolerance for salt, meaning they may not be adapted for
extremely high growth temperatures which might require tetraether lipids. But as
temperature increases the percent of tetraether lipids increases to the point where the
therrnoacidophiles contain only tetraether derivatives. Similarly, Sarcina ventriculi forms
onto-fatty acids and cross-links its headgroups forming a bipolar monolayer, which is
reminiscent of the tetraether lipids in archaea. S. ventriculi synthesizes these lipids in the
presence of high temperatures, acidic and alkaline pHs, and in the presence of certain

fluidizing solvents [29, 47]. The anaerobic thermophilic bacterium T hermoanaerobacter

21

  

C3308
A
: o H
33 ' o
Err . r
H
B
EY-OH
c.a.c0— n. wen-carcmucn.
Cm
H ‘OCWCHDEE‘EETCHflu—COO-TH n H ‘03
HgC(CHq)n—CH-CH H: H [Eco—9
' c

OH

OH
OH OH

 

Figure 1.4: Structures of novel lipids from bacteria. (A) diether lipid and (B) tetraether
lipid from Archaea; (C) a diabolic acid-containing phospholipid from Butyrivibrio S2; and
(D) a tetrahydroxybacteriohopane lipid from Acetobacter.

22

ethanolicus 39E [48] and the hypertherrnophile Thermatoga maritima [49] also synthesizes
a,w-dicarboxylic fatty acids in response to high temperatures. Other specialized lipids

include hopanes, and co-cyclohexyl fatty acids.

Hopanes are the bacterial equivalent of cholesterol [50]. These structures have been
found in diverse bacteria including Acetobacter [51], cyanobacteria, obligate methylotrophs,
and in many Gram-negative and Gram—positive chemoheterotrophs [52]. w-Cyclohexyl fatty
acids are the predominant fatty acids in the thermoacidophiles isolated from hot springs all
over the world. Most have been initially classified as strains of Bacillus acidocaldarius. But
(o-cyclohexyl fatty acids are also major components in some mesophilic bacteria such as
Curtobacterium pusillum [53, 54]. Several strains of therrnoacidophilic bacteria isolated
from Japanese hot springs included anywhere from 74 to 93% of these fatty acids, and
biosynthesis of these lipids increases with increasing glucose in the medium [53]. This is
necessary because acids and solvents are made from glucose as a primary fermentation
activity. To protect its membrane, the levels of these (o—cyclohexyl fatty acids must increase
as glucose concentration increases and fermentation activity increases as a consequence. The
region of greatest hydrocarbon chain mobility in a biomembrane is the area between the two
leaflets (Figure 1.5) [55]. By incorportating an (o-cyclohexyl group in this zone, the
organism can reduce the extent of chain terminus rotation and wagging. This affords a very
high degree of motional restriction and hence control of dynamics. The flat (all trans-fused)

hopane molecule does the same, but along the entire length of the hydrocarbon chain.

23

 

 

 

 

 

 

 

 

 

 

Figure 1.5 : Simplified pictures of the motional spectrum of lipids. (Pm indicates the
maximum angle subtended by the two acyl chains at the headgroup. It is a measure of the
extent to which the acyl chains are sprayed. (A) Monolayer form, and (B) Bilayer form.

24

Protein stability and function

Cells must also maintain protein stability and function. The studies to date primarily
focus on protein stability at high temperatures, mainly from thermophilic organisms. While
the exact mechanisms that confer high temperature tolerance are not known, some
suggestions are fewer amino acids that are vulnerable to heat and the formation of "salt-
bridges," which consists of ionic bonds through the interior preventing denaturation.
Enzymes from thermophilic barophiles, isolated from deep sea vents, often consist of a more
dense hydrophobic interior as well as containing salt bridges. Increased pressure stabilizes
van der Waals interactions and is thought to destabilize salt bridges. In general enzymes
isolated from bacteria adapted to high temperatures are also extremely resistant to high
pressure [56]. Proteins from extreme halophilic and even some moderate halophiles have
been shown to contain a high content of acidic amino acids and are also relatively deficient
in nonpolar amino acids compared to nonhalophiles [43]. The tendency to increase the
amount of acidic arrrino acids relative to basic amino acids may not be general. Increased
proportions of polar amino acids will, in general, help in keeping the proteins hydrated in this
highly polar environment. Increases in N aCl and KCl increases the extent of interactions of
hydrophobic groups again by a "salting out effect." Proteins with less nonpolar amino acids
will maintain a more flexible conformation and a higher degree of hydration with increased

salinities.

In addition to protein stability, certain enzyme activities and metabolic processes

must be maintained. Since membrane alterations, especially changes in acyl chain structures,

25

occur in response to environmental perturbations, some aspect(s) of lipid biosynthesis must
be regulated by the effects of the environment on the cell. Temperature regulated changes
in membrane fatty acid composition have been studied the most. Two mechanisms for
making unsaturated lipids are known. These pathways are 'aerobic' or 'anaerobic' and can
usually be distinguished by the position of the double bond in the acyl chain. The aerobic
pathway requires molecular oxygen and results in palmitoleate (16:1A9cis) and oleate
(18: 1A9cis) as the primary unsaturated products. The anaerobic pathway yields palmitoleate
and cis-vaccenate (18:1A11cis) as the major unsaturated products [57]. All anaerobic and
facultative anaerobic bacteria use the anaerobic pathway of fatty acid biosynthesis. In E.
coli, which uses the anaerobic pathway, the ratio of saturated to unsaturated acyl chains is
thought to be determined by the thermal regulation of B-ketoacyl-ACP synthase. This
enzyme exists in two forms in E. coli, type I and H, and catalyzes the two carbon elongation
reaction of fatty acid biosynthesis. Type H is more efficient in elongating palmitoleate than
palmitate (16:0) and is more temperature labile. Thus at low temperatures palmitoleate is
elongated to cis-vaccenate and less palmitate is made. In this case the effect of longer chains
is outweighed by the effect of increased unsaturation. In this organism the thermal regulation
of fatty acid composition does not require de novo protein synthesis, only changes in activity

of existing enzymes occurs [58].

The aerobic pathway for producing unsaturated acyl chains involves desaturation of
saturated fatty acids. In this case the fatty acid synthetase yields saturated fatty acids only.

The desaturases, which are membrane-bound, are believed to work on membrane

26

phospholipids directly. Temperature then could regulate these enzymes by inducing
membrane phase changes to alter the enzymatic activity, or by induction of new protein
synthesis [58]. In Bacillus strains the desaturase enzyme is temperature labile resulting in
faster turnover rates with increased temperatures. Drops in temperature then cause an
increase in the amount of desaturase present in the membrane (slower turnover rate). Lower

temperatures also are known to induce desaturase activity in these organisms [59].

As noted earlier, changes in membrane acyl chain length occur with alteration of
temperature. It is proposed that an elongase enzyme that is thermally regulated is responsible
for this. Hence, Micrococcus cryophilus alters the ratio of oleic and palmitoleic acids, the
only fatty acids in its membrane, in response to temperature supposedly through an elongase
enzyme that interconverts Cl6 and C18 fatty acids by way of a C14 intermediate [58]. Russell
and coworkers suspect an essential component of this process such as acyl carrier protein
(ACP) turns over rapidly because inhibitors of protein synthesis only partially inhibit this
interconversion [58]. An alternative idea is that the acyl chain length selection process is
random, and that the most viable cells at a given temperature have a particular acyl chain
distribution. Dramatic shifts in temperature select for the bacterial populations with the best
suited acyl chain length distributions or the ones that can alter their distributions to function

at the new temperature.

Other alterations of fatty acid structures are seen in response to temperature including

changes in the cyclization, branching, and cross-linking of fatty acids, though the

27

mechanisms of thermal regulation of these syntheses has not yet been shown. The synthesis
of iso- and anteiso-branched fatty acids requires de novo fatty acid synthesis using primers
derived from amino acids. Temperature regulation of the transaminases that convert amino
acids to their a-keto acids has not been demonstrated as of 1984 [58], and a subsequent
literature search to the present has found no further evidence that this has been demonstrated.
As will be discussed later, the thermal regulation of the synthesis of the branched fatty acids
relates to the available pools of acyl-CoA primers as well as to the available thermal energy.
It has been shown that the tail-to—tail cross-linking of S. ventriculi membrane acyl chains is
temperature dependent as well. The regulation occurs presumably through activation of a
constitutively expressed enzyme, since inhibitors of protein synthesis do not affect this
activity [29]. Similarly, the Archaeon Methanococcusjannashcii has been shown to increase
the proportions of tetraether lipids (C40) over the level of diether lipids (C20) with increasing
temperature [60], indicating some aspect of the formation of tetraethers is also thermally

regulated.

In addition to regulating enzymes involved in controlling fatty acid composition many
organisms produce 'heat shock' proteins in response to detrimental stress factors. Though
they are called heat shock proteins, they are synthesized in response to extreme changes in
temperature (hot or cold), acid, ethanol, and other physical and chemical perturbations.
These proteins are constitutively expressed at very low levels, but with a sudden large
temperature increase (or extreme changes due to other perturbants) the levels of these

proteins rapidly increase. As the cells acclimate to the new environmental condition, the heat

28

shock protein levels decrease slowly, usually over a couple of hours, back to basal levels.
If cells, however, are slowly acclimated to a higher temperature (or other condition), the
levels of heat shock proteins do not increase [61]. The major heat shock proteins (Hsp) are
believed to function as molecular chaperones to aid in protein folding and prevent protein
aggregation of unfolded proteins [62]. Some of these same proteins induced by heat are also
induced by other stresses such as ethanol and inhibitors of oxidative phosphorylation [63].
E. coli, Bacillus cereus WSBC 10201, and Pseudomonasfragi all produce a new series of
proteins in response to cold shock, one in particular is CspA [64-66]. Some of these cold
shock proteins are thought to function in control of transcription and translation, namely to
prevent initiation of protein synthesis [67]. E. coli and Salmonella typhimurium are two
examples of bacteria that produce acid shock proteins [68, 69]. While the exact functions
of most shock proteins are not known, they may simply reflect the shifts in metabolic

pathways that must occur in response to changing environmental conditions.

Osmotic balance

Changes in concentrations of solutes in the aqueous environment surrounding
organisms determines the flow of water into and out of the cell. Cell membranes are readily
permeable to water but not to many solutes. As a result, these osmotic changes alter cell
volume and/or pressure, and changes in cell volume lead to fluctuations in intracellular solute
concentrations as well. While the cellular signals for sensing osmotic changes are not really
known at this time [70], several physicochemical changes that might be used by Gram-

negative bacteria have been suggested. These theories include baroreceptors for detecting

29

hydrostatic pressure changes in the interior bulk liquid phase of the cell, which have been
viewed as unlikely [71]; turgor pressure sensors in the inner membrane which respond to
compression against the peptidoglycan, and wall stretch receptors which respond to
tangential stress within the peptidoglycan layer [72]; stretch receptors for detecting changes
in membrane surface area [73]; and chemoreceptors for detecting internal and external solute

concentration or water activity [74].

It is thought that a common adaptive response for many organisms is to alter
intracellular concentrations of solutes that do not produce adverse effects on cellular
macromolecules [71, 75, 76]. These solutes are referred to as osmolytes or compatible
solutes. Compatible solutes include K+ ions; amino acids such as glutamate, glutamine,
proline, and glycine betaine (an N-methyl substituted amino acid); trehalose and
mannosucrose, which are disaccharides; and others such as monosaccharides and glycerol.
Osmolytes in high concentrations are seemingly not too toxic, but they do affect some
biochemical processes. It has been proposed that increased intracellular concentrations of
osmolytes has overcome temperature—sensitive mutations [76] as well as increased the upper
temperature growth limit [77, 78], and conferred cryotolerance [79] on many bacteria.
Results from these studies should be taken with caution, though, because the media were
altered and analysis on the membrane chemistry and the levels of protein expression were
not canied out. Bacteria that are thought to produce compatible solutes include Escherichia
coli [80-82], several marine methanogens [83], Rhizobium meliloti [84], and many others

[70]. It should be noted, however, that an increase in compatible solutes may not indicate

30

osmotic adaptation. The increase (or decrease) of certain metabolites may be a consequence
of the new conditions to which the cell has adapted. Alterations in metabolic pathways are
to be expected as environmental conditions change. So an increase in metabolites may not
be the cause of adaptation to changes in osmotic pressure. For instance, if the increase in a
particular osmolyte is still low compared to the external osmolality there is little significance
in this event with regards to adapting to changes in osmotic pressure. For example, the
glutamine pool in E. coli remains proportional to the external osmolality, but is 10% less
than the intracellular K+ ion concentration [80]. In halophilic archaea the glycine betaine
concentration is about 100 fold less than the major compatible solutes (K+ and Cl’ ions) [85].
So in these instances the presence of the osmolytes may serve a different function and/or be
a consequence of shifts in metabolism [70]. Again, it is thought that bacteria respond to the
reverse situation, hypoosmotic shock (water influx), by decreasing osmolyte concentrations
via excretion, dilution by growth, catabolism, and polymerization. As will be discussed in
more detail in chapter 4, osmolytes are also thought to aide in cryoprotection of
microorganisms, but this is not the only means by which bacteria can survive subzero
temperatures. In the next section special issues relevant to survival in subzero climates is

reviewed.

Special issues in subzero temperature adaptation
A great deal of research has focused on the adaptation of organisms to extreme heat.
One reason for this is the potential industrial use of enzymes and organisms that can

withstand such high temperatures. However from the aspect of adaptation of organisms to

31

extreme environments, the majority of the earth's biosphere is subjected to subzero
temperatures permanently or seasonally. Water can be supercooled to the extent that it can
still be a liquid down to -40°C and below. As such, organisms deal with stresses from
supercooled water as well from freezing. In supercooled water the degree of ionization is
greatly decreased, affecting its ability to serve as conjugate bases or acids in reactions. Also
the viscosity of the water increases and thus diffusion rates decrease [86]. In light of these
and other changes that occur as water is chilled or frozen, the organisms that inhabit these
climates have special circumstances which must be dealt with in order acclimate to subzero
temperatures. These issues include responding to changes in salt/metabolite concentrations
and oxygen availability; prevention of freezing of intracellular water; and provision of

cellular maintenance energy, carbon sources and heat generation.

Salt/metabolite concentrations and oxygen availability

As water freezes it generally separates out as pure water (ice) causing an increase in
salt and metabolite concentrations in the remaining liquid phase. Bacteria in the liquid phase
would need to adjust to these increases. However, bacteria that may be trapped within the
ice see the other extreme: a sharp decrease in available salts both organic and inorganic.

Diffusion of these salts and metabolites for use by the cells is now extremely limited.

Oxygen on this planet is generated by photosynthesis, which obviously requires light.
In most subzero climates, oxygen is primarily generated by algae but only on the surface of

snow. For microorganisms buried underneath these frozen habitats, oxygen becomes

32

limiting very quickly because it cannot diffuse well through the ice. The bacteria must then
develop efficient anaerobic mechanisms for generating energy using other inorganic terminal
electron acceptors that are available. These electron acceptors may be nitrate, carbon

dioxide, sulphate, thiosulphate, or sulfite [87].

Prevention of freezing of intracellular water

Freezing of intracellular water causes several problems, the major one is cell rupture
due to the expansion of frozen water. Another problem is an imbalance in electrolyte
concentration. As ice forms it forces out the solutes creating areas with high electrolyte
concentrations causing osmotic changes that result in the efflux of water. It is also possible
for freezing to result in eutectic mixtures in which some of the solutes remain within the ice,
preventing their availability for use by the cell. Cells must also contend with decreased

diffusion of metabolites as water becomes supercooled and/or freezes.

Maintenance energy

All cells expend a certain amount of energy to maintain organization at the molecular
level. This includes the spatial orientation of membranes, proteins, nucleic acids, etc. As
is well known, there exists a natural tendency towards randomness and disorder (increase in
entropy) in the universe. Therefore maintaining the necessary levels of organization as well

as carrying out the biosynthesis of macromolecules, requires energy.

More than two decades ago calculations were made by several researchers to estimate

33

the amount of ATP needed for maintenance energy by bacteria. John Pirt determined a
maintenance coefficient for aerobic and anaerobic continuous cultures of Aerobacter cloacae
[88]. This is an index that tells the amount of glucose needed to maintain a gram of cells.
This coefficient was defined as the slope of the line of plots of the reciprocal of the yield
value versus the reciprocal of growth rate. The values for the aerobic and anaerobic cultures
were 0.09 and 0.47 grams of glucose per gram dry weight bacterium, respectively. The
higher coefficient for anaerobic cultures is expected due to less efficient energy yields from
anaerobic metabolism. Using the coefficient for the anaerobic culture, the amount of ATP
needed to maintain one gram dry weight bacterium was one-twentieth (~0.75 mmole) the
ATP necessary to produce one gram dry weight of this bacterium. It should be noted,
though, that this value far exceeds the free ATP pool (~10 pmoles per gram dry weight
bacterium) usually available in a given microorganism [89]. About ten years later Tempest
and colleagues [90], by modifying Pirt's derivation, determined another linear relationship.
The slope of their plot of energy source/cells/hour versus growth rate is also defined as a
maintenance coefficient, but the error in this slope is primarily in the intercept. In the Pirt
plots, the error is in the slope. While these two derivations for determining maintenance
energy are often used, other calculations for determining maintenance yield have been done

[91].

While there is no one way to calculate maintenance energy, the concept is still very
important. The idea of maintenance energy was first suggested when it was observed that

low levels of energy sources were not sufficient for subculturing bacteria, indicating that

34

cells may need some energy just to maintain viability [92]. At low temperatures cells require
much more energy to produce one gram of cells (also partly due to decreased oxygen
availability), but the total amount of energy put into the system (as measured by carbon
source used) is even greater than that calculated to produce one gram of cells [89]. So at low
temperatures, the amount of energy needed to maintain viable cells is also much greater.
This can be explained by the driving force of cellular chemical processes--heat (thermal
energy). Since cell maintenance requires turnover of macromolecules, to drive the chemical
processes involved in turnover requires a certain amount of heat, which makes these
processes more difficult at low temperatures. Heat expenditure is also necessary to sustain
membrane dynamics at low temperatures. Low temperatures cause lipids to solidify,
preventing the diffusion and lateral motion of substrates and enzymes used in the many
chemical reactions that take place within the membrane. In addition maintaining the
electromotive force across the membrane is entropically unfavorable. So a certain amount
of energy is used to preserve membrane asymmetry. Enzymes, the biological
macromolecular catalysts, also require thermal energy to sustain the optimal vibrational
modes at which they function. Thus overall, with the decrease in available thermal energy
at colder temperatures, cells must burn vast amounts of chemical fuels just to maintain an

equivalent cell mass, much less allow for cell growth.

Carbon source and heat generation
This increased need for heat requires cells to be more efficient in generating heat and

in using energy resources. Typically cells use three classes of substances for deriving

35

chemical energy: carbohydrates, amino acids, and fatty acids. Carbohydrates are the most
oxidized of these compounds and hence generate the least amount of heat per gram than
amino acids and fatty acids. A substantial loss in the net generation of energy from burning
carbohydrates comes from the priming of them for metabolic degradation. This process
requires a large investment of energy through the generation of phosphorylated species.
Amino acids are an oxidation state lower than carbohydrates and fatty acids are the among
the most reduced substrates for microorganisms. Both would provide more energy than
carbohydrates and would be used with greater efficiency at subzero temperatures, as is
indicated by their respective heats of combustion. The heat of combustion for glucose is
669.94 kg-cal/g. While for leucine (855.6 kg-cal/g) and two fatty acids, palmitate (2,384.76
kg-cal/g) and oleate (2,657.4 kg-cal/g), the heats of combustion are higher, especially for the
fatty acids [93]. Further evidence for the preferred use of fatty acids as energy at subzero
temperatures is the observation of lipid droplets in many Arctic and other psycrophilic
organisms (Danielle Prévost, Agriculture and Agri-Food, Canada; personal communication).

These lipid droplets consist of triacylglycerides, which are the storage form of fatty acids.

The universality of adaptive responses to various environmental perturbations

From the evidence documented above one can see that numerous environmental
perturbations cause the same types of molecular mechanisms of adaptation. Therefore these
stresses may induce the same sets of sensory mechanisms in a given organism. The
peripheral membrane of procaryotes is the ideal candidate for sensing all of these variations

within an environment. In surrounding the cell it has constant contact with its surroundings,

36

it transmits the available thermal energy to the cell interior, and its organization and
adaptability allow for rapid alterations in its structure to compensate for any physical and
chemical perturbations it senses. The primary sensory mechanism could be simply
perturbations in the membrane dynamics sensed by disturbances of the gel and liquid-
crystalline phases. As discussed above both physical and chemical stresses (including
temperature, pressure, pH, salinity, and chemical substances) all result in membrane
alterations which are thought to maintain a certain degree of fluidity. Cold temperatures,
increased pressure, increased salinity (depending on the type and concentration of cation),
and certain solvents all rigidify the membrane by inducing closer packing of the lipid acyl
chains and/or by stabilizing electrostatic lipid headgroup interactions. Thus all these
perturbations result in membranes with greater unsaturated acyl chains and shorter chain
lengths, which disrupts this induced tighter packing. Other factors that fluidize the
membrane (high temperature, pH and salinity extremes, and some solvents) by disrupting
acyl chain packing and/or destabilizing electrostatic interactions of the lipid headgroups,
result in membranes with longer, more saturated acyl chains which increases the acyl chain

interactions causing a reduction in the acyl chain motion.

The membrane, by virtue of its component lipid acyl chains, is tuned to the available
thermal energy and transmits it to the cell interior. This available energy is what determines
which metabolic reactions can be carried out, and the cell's metabolic processes are adjusted
accordingly. The products of the metabolic process, however, determine the available pool

of acyl-CoA primers which can be used for fatty acid synthesis. So there is a circular

37

relationship between available thermal energy, metabolism, and membrane structure. A

change in any one of these parameters by any type of perturbation affects the other two.

Psychrophilic and psychrotropic bacteria offer a unique opportunity to establish the
relationship between energy, metabolism, and fatty acid structures. As discussed above, the
available thermal energy is much less as temperature drops. Thus the balance between
carbohydrate and amino acid metabolism is affected, which is also reflected in the membrane
lipid acyl chain structures. In this work, principles of adaptation and adaptability of
organisms to extreme environments are developed using an Arctic permafrost bacterium.

These principles are then applied to data regarding other bacteria to test their generality.

REFERENCES

l. Hollingsworth, R. I. (1995) Membrane lipid alkyl chain heterogeneity is a molecular
transform of the available thermal energy spectrum. J. Phys. Chem, 99: 3406-3410.

2. Sinensky, M. (1974) Homeoviscous adaptation--a homeostatic process that regulates the
viscosity of membrane lipids in Escherichia coli. Proc. Natl. Acad. Sci. USA, 71: 522-525.

3. Russell, N. J., and N. Fukunaga. (1990) A comparison of thermal adaptation of
membrane lipids in psychrophilic and thermophilic bacteria. FEMS Microbiol. Revs, 75:
171-182.

4. Singer, S. J. and G. L. N icolson. (1972) The fluid mosaic model of the structure of cell
membranes. Science, 175: 720-731.

5. Schleifer, K. H., and O. Kandler. (1972) Peptidoglycan types of bacterial cell walls and
their taxonomic implications. Bacteriol. Rev., 36: 407-477.

6. Karnovsky, M. J ., A. M. Kleinfeld, R. L. Hoover, E. A. Dawidowicz, D. E. McIntyre, E.
A. Salzman, and R. D. Klausner. (1982) Lipid domains in membranes. Ann. N. Y. Acad.

38

Sci., 401: 61-75.

7. Jain, M. K. (1983) Nonrandom lateral organization in bilayers and biomembranes. In
Membrane Fluidity in Biology, ed. R. C. Aloia, 1: 1-37.

8. Cullis, P. R., and B. DeKruijff. ( 1979) Lipid polymorphism and the functional roles of
lipids in biological membranes. Biochim. Biophys. Acta, 559: 399-420.

9. Lee, A. G. (1983) Lipid phase transitions and mixtures. In Membrane Fluidity in
Biology, ed. R. C. Aloia, 2: 43-88.

10. McElhaney, R. N. (1984) The relationship between membrane lipid fluidity and phase
state and the ability of bacteria and myc0plasmas to grow and survive at various
temperatures. Biomembranes, 12: 249-278.

11. Huang, L., S. K. Lorch, G. G. Smith, and A. Haug. (1974) Control of membranes lipid
fluidity in Acholeplasma laidlawii. FEBS Letters, 43: 1-5.

12. Esser, A. F., and K. A. Souza. (1974) Correlation between thermal death and membrane
fluidity in Bacillus stearothermophilus. Proc. Natl. Acad. Sci. USA, 71: 4111-4115.

13. Cossins, A. R., and M. Sinensky. (1984) Adaptation of membranes to temperature,
pressure, and exogenous lipids. In CRC Physiology of Membrane F luidity Vol. [1, ed. M.
Shinitzky, CRC Press, Boca Raton, FL. pp. 1-20.

14. van der Meer, W. (1984) Physical aspects of membrane fluidity. In CRC Physiology
of Membrane F luidity Vol. 1, ed. M. Shinitzky, CRC Press, Boca Raton, FL. pp. 53-71.

15. McElhaney, R. N. (1984) The structure and function of the Acholeplasma laidlawii
plasma membrane. Biochim. Biophys. Acta, 779: 1-42.

16. Rilfors, L., G. Lindblom, A. Wieslander, and A. Christiansson. (1984) Lipid bilayer
stability in biological membranes. Biomembranes, 12: 205-245.

17. Cullis, P. R., M. J. Hope, and C. P. S. Tilcock. (1986) Lipid polymorphism and the roles
of lipids in membranes. Chemistry and Physics of Lipids, 40: 127- 144.

18. Wieslander, A., A. Christiansson, L. Rilfors, and G. Lindblom. (1980) Lipid bilayer
stability in membranes. Regulation of lipid composition in Acholeplasma laidlawii as
governed by molecular shape. Biochemistry, 19: 3650-3655.

19. Wieslander, A., L. Rilfors, L. B.-A. Johansson, and G. Lindblom. (1981) Reversed
cubic phase with membrane glucolipids from Acholeplasma laidlawii. 'H, 2H, and diffusion
nulear magnetic resonance measurements. Biochemistry, 20: 730-735.

39

20. Wieslander, A., L. Rilfors, and G. Lindblom. (1986) Metabolic changes of membrane
lipid composition in Acholeplasma laidlawii by hydrocarbons, alcohols, and detergents:
Arguments for effects on lipid packing. Biochemistry, 25: 7511-7517.

21. Herbert, R. A., and M. Bhakoo. (1979) Microbial growth at low temperatures. In Cold
Tolerant Microbes in Spoilages and the Environment, eds. A. D. Russell and R. Fuller.
Academic Press, London. pp. 1-16.

22. Bhakoo, M., and R. A. Herbert. (1979) The effects of temperature on the fatty acid and
phopholipid composition of four obligately psychrophilic Vibrio spp. Arch. Microbiol., 121:
121-127.

23. Bhakoo, M, and R. A. Herbert. (1980) Fatty acid and phospholipid composition of five
psychrotrophic Pseudomonas spp. grown at different temperatures. Arch. Microbiol., 126:
51-55.

24. Russell, N. J. (1974) The lipid composition of the psychrophilic bacterium Micrococcus
cryophilus. J. Gen. Microbiol., 80: 217-225.

25. Yamauchi, T., K. Ohki, H. Maruyama, and Y. Nozawa. (1981) Thermal adaptation of
Tetrahymena membranes with special reference to mitochondria. Role of cardiolipin in
fluidity of mitochondrial membranes. Biochim. Biophys. Acta, 649: 385-392.

26. Taneja, R., and G. K. Khuller. (1981) Phospholipid metabolism in Mycobacterium
smegmatis ATCC 607 grown at 37°C and 27°C. Arch. Microbiol., 129: 81-84.

27. Hawrot, E., and E. P. Kennedy. (1978) Phospholipid composition and membrane
function in phosphatidylserine decarboxylase mutants of Escherichia coli. J. Biol. Chem,
253: 8213-8220.

28. Raetz, C. R. (1976) Phosphatidylserine synthetase mutants of Escherichia coli. Genetic
mapping and membrane phospholipid composition. J. Biol. Chem, 251: 3242-3249.

29. Lee, J ., S. Jung, S. Lowe, J. G. Zeikus, and R. I. Hollingsworth. (1995) A dynamically
regulated transformation of the bilayer membrane of bacteria to a cross-linked 2-dimensional
sheet during adaptation to unfavorable environmental pressures. Submitted to J. Amer.
Chem. Soc.

30. J urado, A. S., M. S. Costa, and V. M. C. Madeira. (1988) Therrnotropic behavior of the
polar lipids of B. stearothermophilus. Abstr. FEMS Int. Symp. on The Microbiology of
Extreme Environments and its Biotechnological Potential. Troia, Portugal.

31. Thompson, G. A., and Y. Nozawa. (1984) The regulation of membrane fluidity in
Tetrahymena. Biomembranes, 12: 397-432.

40

32. Marr, A. G., and J. L. Ingraham. (1962) Effect of temperature on the composition of
fatty acids in Escherichia coli. J. Bacteriol., 84: 1260.

33. Kito, M., S. Aibara, M. Kato, and T. Hata. (1972) Differences in fatty acid composition
among phosphatidylethanolamine, phosphatidyl glycerol, and cardiolipin of Escherichia coli.
Biochim. Biophys. Acta, 260: 475-478.

34. Cullen, J ., M. C. Phillips, and G. G. Shipley. ( 1971) The effects of temperature on the
composition and physical properties of the lipids of Pseudomonasfluorescens. Biochem. J.,
125: 733-742.

35. Gill, C. O. (1975) Effect of growth temperature in the lipids of Pseudomonas
fluorescens. J. Gen. Microbiol., 89: 293-298.

36. Hyono, A., S. Kuriyama, H. Hara, 1. Yano, and M. Masui. (1980) ESR studies on the
membrane properties of moderately halophilic bacterium. 1. Physical properties of lipid
bilayers in whole cells. J. Biochem., 88: 1267-1274.

37. Hara, H., A. Hyono, S. Kuriyama, I. Yano, and M. Masui. (1980) ESR studies on the
membrane properties of a moderately halophilic bacterium. II. Effect of extreme growth
conditions on liposome properties. J. Biochem., 88: 1275-1282.

38. Rottem, S., O. Markowitz, and S. Razin. (1978) Thermal regulation of the fatty acid
composition of lipopolysaccharides and phopholipids of Proteus mirabilis. Eur. J. Biochem. ,
85: 445-450.

39. Bishop, D. G., and J. L. Still. (1963) Fatty acid metabolism in Serratia marcescens: the
effect of temperature on fatty acid composition. J. Lipid Res., 4: 87.

40. Suutari, M., and S. Laakso. (1994) Microbial fatty acids and thermal adaptation. Crit.
Rev. Microbiol., 20: 285-328.

41. Hazel, J. R., and E. E. Williams. (1990) The role of alterations in membrane lipid
composition in enabling physiological adaptation of organisms to their physical environment.
Prog. Lipid Res, 29: 167-227.

42. DeLong, E. F., and A. A. Yayanos. (1985) Adaptation of the membrane lipids of a deep-
sea bacterium to changes in hydrostatic pressure. Science, 228: 1101-1103.

43. Russell, N. J. (1989) Adaptive modifications in membranes of halotolerant and
halophilic microorganisms. J. Bioenerg. Biomem, 21: 93-1 13.

44. Komaratat, P., and M. Kates. (1975) The lipid composition of a halotolerant species of
Staphylococcus epidermidis. Biochim. Biophys. Acta, 398: 464-484.

41

45. Kates, M. (1992) Archaebacterial lipids: structure, biosynthesis and function. Biochem.
Soc. Symp., 58: 51-72.

46. Pool, R. (1990) Pushing the envelope of life. Science, 247: 158-160.

47. Jung, S., S. E. Lowe, R. I. Hollingsworth, and J. G. Zeikus. (1993) Sarcina ventriculi
synthesizes very long chain dicarboxylic acids in response to different forms of
environmental stress. J. Biol. Chem, 268: 2828-2835.

48. Jung, 8., J. G. Zeikus, and R. I. Hollingsworth. (1994) A new family of very long chain
moo-dicarboxylic acids is a major structural fatty acyl component of the membrane lipids of
Thermoanaerobacter ethanolicus 39E. J. Lipid Res., 35: 1057-1065.

49. Carballeira, N. M., M. Reyes, A. Sostre, H. Huang, M. F. J. M. Verhagen, and M. W.
W. Adams. (1997) Unusual fatty acid compositions of the hyperthermophilic Archaeon
Pyrococcus furiosus and the bacterium Thermotoga maritima. J. Bacteriol., 179: 2766-

2768.
50. Ourisson, G. (1987) Bigger and better hopanoids. Nature, 326: 126-127.

51. Forster, H., K. Biemann, W. G. Haigh, N. H. Tattrie, and J. R. Colvin. (1973) The
structure of novel C35 pentacyclic terpenes from Acetobacter xylinum. Biochem. J., 135:
133-143.

52. Rohmer, M., P. Bouvier—Nave, and G. Ourisson. (1984) Distribution of hopanoid
triterpenes in prokaryotes. J. Gen. Microbiol., 130: 1137-1150.

53. Oshima, M., and T. Ariga. (1975) w-Cyclohexyl fatty acids in acidophilic thermophilic
bacteria. J. Biol. Chem, 250: 6963-6968.

54. Mantsch, H. H., C. Madec, R. N. A. H. Lewis, and R. N. McElhaney. (1989) An
infrared spectroscopic study of the therrnotropic phase behavior of phosphatidylcholines

containing w-cyclohexyl fatty acid chains. Biochim. Biophys. Acta, 980: 42-49.

55. Yeagle, P. (1992) The dynamics of membrane lipids. In The Structure of Biological
Membranes, P. Yeagle, ed., CRC Press, Inc., Boca Raton, FL. pp.157-174.

56. Flam, F. ( 1994) The chemistry of life at the margins. Science, 265: 471-472.

57. Kawaguchi, A., and Y. Seyama. (1984) Thermal control of fatty acid synthetases in
bacteria. Biomembranes, 12: 279-301.

58. Russell, N. J. (1984) Mechanisms of thermal adaptation in bacteria: blueprints for
survival. TIBS, pp. 108-112.

42

59. Fulco, A. J ., and D. K. Fujii. (1980) In Membrane F luidity. Biophysical Techniques and
Cellular Regulation, M. Kates and A. Kuksis, eds, Humana Press, pp. 77—98.

60. Sprott, G. D., M. Meloche, and J. C. Richards. (1991) Proportions of diether,
macrocyclic diether, and tetraether lipids in Methanococcus jannaschii grown at different
temperatures. J. Bacteriol., 173: 3907-3910.

61. Yamamori, T., and T. Yura. (1982) Genetic control of heat-shock protein synthesis and
its bearing on growth and thermal resistance in Escherichia coli K-12. Proc. Natl. Acad. Sci.
USA, 79: 860—864.

62. Craig, E. A., B. D. Gambill, and R. J. Nelson. (1993) Heat shock proteins: molecular
chaperones of protein biogenesis. Microbiol. Revs, 57: 402-414.

63. Craig, E. A. (1985) The heat shock response. CRC Crit. Rev. Biochem., 18: 239-280.

64. Goldstein, J ., N. S. Pollitt, and M. Inouye. ( 1990) Major cold shock protein of
Escherichia coli. Proc. Natl. Acad. Sci. USA, 87: 283-287.

65. Mayr, B., T. Kaplan, S. Lechner, and S. Scherer. (1996) Identification and purification
of a family of dimeric major cold shock protein homologs from the psychrotrophic Bacillus
cereus WSBC 10201. J. Bacteriol., 178: 2916-2925.

66. Hebraud, M., E. Dubois, P. Potier, and J. Labadie. (1994) Effect of growth temperatures
on the protein levels in a psychrotrophic bacterium, Pseudomonas fragi. J. Bacteriol., 176:
4017 -4024.

67. Russell, N. J. (1990) Cold adaptation of microorganisms. Phil. Trans. R. Soc. Lond. B,
326: 595-611.

68. Goodson, M., and R. J. Rowbury. (1989) Habitation to normally lethal acidity by prior
growth of Escherichia coli at a sublethal acid pH value. Lett. Appli. Microbiol., 8: 77-79.

69. Foster, J. W. ( 1991) Salmonella acid shock proteins are required for the adaptive acid
tolerance response. J. Bacteriol., 173: 6896-6902.

70. Csonka, L. N., and A. D. Hanson. (1991) Prokaryotic osmoregulation: genetics and
physiology. Annu. Rev. Microbiol., 45: 569-606.

71. Cram, W. J. ( 1976) Negative feedback regulation of transport in cells. The maintenance
of turgor, volume, and nutrient supply. In Encyclopedia of Plant Physiology, New Series,
Vol. 2A, ed. U. Luttge, M. G. Pitman. Berlin/New York: Springer. pp. 284-316.

72. Frey-Wyssling, A. (1957) Macromolecules in Cell Structure. Harvard University Press,

43

Cambridge, pp. 1 12.

73. Sant'Ambrogio, G. (1987) Nervous receptors of the tracheobronchial tree. Annu. Rev.
Physiol, 49: 611-27.

74. Macnab, R. M. (1987) Motility and chemotaxis. In Escherichia coli and Salmonella
typhimurium: Cellular and Molecular Biology. F. C. Neidhardt, J. L. Ingraham, K. B. Low,
B. Magasanik, M. Schaechter, and H. E. Umbarger, eds., American Society of Microbiology,
Washington, D. C., pp. 732-759.

75. Brown, A. D. (1990) Microbial Water Stress Physiology: Principles and Perspectives.
Chichester: Wiley.

76. Csonka, L. N. (1989) Physiological and genetic responses of bacteria to osmotic stress.
Microbiol. Rev., 53: 121—147.

77. Tesone, S., A. Hughes, and A. Hurst. (1981) Salt extends the upper temperature limit
for growth of food-poisoning bacteria. Can. J. Microbiol., 27: 970-972.

78. Troller, J. A. (1986) Water relations of foodbome bacterial pathogens—-an updated
review. J. Food Prot., 49: 656-670.

79. K0, R., L. Tombras Smith, and G. M. Smith. (1994) Glycine betaine confers enhanced
osmotolerance and cryotolerance on Listeria monocytogenes. J. Bacteriol., 176: 426-431.

80. McLaggan, D., T. M. Logan, D. G. Lynn, and W. Epstein. (1990) Involvement of y-
glutamyl peptides in osmoadaptation of Escherichia coli. J. Bacteriol., 172: 3631-3636.

81. Boos, W., U. Ehmann, H. Fork], W. Klein, M. Rimmele, and P. Postma. (1990)
Trehalose transport and metabolism in Escherichia coli. J. Bacteriol., 172: 3450-3461.

82. Cayley, S., M. T. Record Jr., and B. A. Lewis. (1989) Accumulation of 3-(N-
morpholino)-propanesulfonate by osmotically stressed Escherichia coli K-12. J. Bacteriol.,
171: 3597-3602.

83. Robertson, D. E., M. F. Roberts, N. Belay, K. O. Stetter, and D. R. Boone. (1990)
Occurance of B-glutamate, a novel osmolyte in marine methanogenic bacteria. Appl.
Environ. Microbiol. , 56: 1504-1508.

84. Botsford, J. L., and T. A. Lewis. (1990) Osmoregulation in Rhizobium meliloti:
production of glutamic acid in response to osmotic stress. Appl. Environ. Microbiol., 56:

488-494.

85. Nicolaus, B., V. Lanzotti, A. Trincone, M. De Rosa, and A. Gambacorta. (1989)

44

Glycine betaine and polar lipid composition in halophilic archaebacteria in response to
growth in different salt concentrations. FEMS Microbiol. Lett., 59: 157-160.

86. Franks, F., S. F. Mathias, and R. H. M. Hatley. (1990) Water, temperature and life.
Phil. Trans. R. Soc. Lond. B, 326: 517-533.

87. Rose, A. H. (1976) Chemical Microbiology: An Introduction to Microbial Physiology,
3rd Ed. Plenum Press, New York. pp. 181-243.

88. Pirt, S. J. (1965) The maintenance energy of bacteria in growing cultures. Proc. R. Soc.
London Ser. B, 163: 224-231.

89. Rose, A. H. (1976) Chemical Microbiology: An Introduction to Microbial Physiology,
3rd Ed. Plenum Press, New York. pp. 245-325.

90. Neijssel, O. M., and D. W. Tempest. (1976) Bioenergetic aspects of aerobic growth of
K lebsiella aerogenes NCTC 418 organisms growing in chemostat culture. Arch. Microbiol. ,
107: 215-221.

91. Russell, J. B., and G. M. Cook. (1995) Energetics of bacterial growth: Balance of
anabolic and catabolic reactions. Microbiol. Rev., 59: 48-62.

92. Buchanan, R. E., and E. I. Fulmer. (1928) Physiology and biochemistry of bacteria. The
Williams & Wilkins Co., Baltimore.

93. Weast, R. C., ed. (1985) Heat of Combustion. In CRC Handbook of Chemistry and
Physics, 66th ed. CRC Press, Boca Raton, FL. pp. D272-277.

94. Ratledge, C., and S. G. Wilkinson. (1988) Microbial Lipids, Vol. 1, Academic Press,
New York, pp. 1-22.

95. Gruner, S. M. (1992) In The Structures of Biological Membranes, P. Yeagle, ed., CRC
Press, Inc., Boca Raton, FL, pp. 211-250.

CHAPTER H

DEVELOPMENT OF A THEORY ON THE SIGNIFICANCE
OF FATTY ACID HETEROGENEITY

45

46

ABSTRACT

The theory that the chain length distribution of the fatty acyl groups of bacteria is a
Gaussian function the mean and variance of which depends on temperature was tested by
examining this distribution in a permafrost bacterium that is capable of growth from
temperatures close to freezing up to 30°C. The variance of the curve was a linear function
of absolute temperature over most of the range with a second order deviation towards the
upper end. This supports the idea that the chain length selection is a random process in
which the distribution reflects the external temperature, or more correctly momentum
distribution of molecules in the extracellular matrix, and is characterized by a smaller
variance at lower temperatures and a higher one at higher temperatures. Fatty acid
composition data from the literature for three other Gram-positive organisms were also used
to support the theory. Having established that there is a consistent relationship between
temperature and acyl chain distribution, then one can predict the fatty acid composition of
a given organism over a temperature range. By adding a variety of chemical perturbants to
Arctic strain 45-3, Bacillus subtilis PY79, and Sarcina ventriculi J K, we also show that there
is a temperature equivalent for a given solvent perturbation in a particular organism. This
is consistent with earlier isotope studies that indicated the mobility of the hydrocarbon chains
is actually the critical dynamic factor that is influenced by temperature. The membrane acyl
chain length alterations induced by chemical substances, such as solvents and antibiotics
could, therefore, be explained by the structures of the perturbants and how the structures

could influence the acyl chain packing. These results indicate that by knowing whether the

47

perturbant would fluidize or rigidify the membrane, the temperature tolerance of an organism
can be increased (or decreased) by adding exogenous substances such as very long or short
chain or fused-ring hydrocarbons to counteract the temperature effect. For extreme
perturbations, the mean of the distribution curve shifted abruptly to form a new series of
distribution curves about the new mean. For elevated temperatures, this is consistent with

the increased value of kT.

INTRODUCTION

"Why not more quantitative data?

Lipid data on the same strain, grown and analyzed
under the same conditions in two different laboratories, or
even in the same laboratory, may vary appreciably from a
quantitative point of view. Thus, the most important facts
to consider are the types of lipids that occur in a given
organism and which of these represent the major types.
Actual percentages will not be of much use. "

M. P. Lechevalier [1]

While this quote was made in 1982, it still seems to reflect our understanding of
membrane acyl chain heterogeneity today. Researchers have shown that many organisms
alter their membrane acyl chain lengths, unsaturation, branching, and cyclization in response
to fluctuations in temperature. The percentages are graphed or listed in many combinations,
but they still seem to just end up describing an overall trend of changes with no real
relationship between the magnitude of the temperature change and the percentages

determined. Here we show that there is, in fact, a regular (mathematical) relationship

48

between the acyl chain distribution of an organism and its growth temperature.

The reason for the many types of acyl chain alterations that have been documented
in response to temperature (and other perturbations) is thought to be an organism's response
to maintain a necessary degree of membrane fluidity. This phenomenon is termed
"homeoviscous adaption" [2]. Key questions that remain are: How do cells sense
temperature? and, What is the connection between this observed membrane heterogeneity

and temperature?

All cellular metabolic pathways are series of chemical reactions, and these reactions
all have activation energy barriers which must be overcome. The driving force to overcome
these barriers is heat (infrared quanta). Temperature dictates the available heat that can be
used for driving cellular processes, including the synthesis of chemical energy (ATP). One
theory is that the membrane lipid acyl chains are able to channel this energy (kT) to the cell
interior via special vibrational modes called longitudinal accordian modes (LAMS) [3]. This
concept is depicted in Figure 2.1. The sun emits a continuous spectrum of electromagnetic
(EM) radiation, which determines the thermal energy spectrum on earth. Though the infrared
part of the spectrum is of interest here because it is absorbed as heat in matter. The most
abundant quanta at a given temperature (T) has the magnitude kT. This thermal energy
spectrum is then transferred to the matter (water, gas, and other molecules) surrounding cells
and is stored in their rotational, vibrational, and translational modes. In biological systems

water plays an important role in transferring this energy through collisions with other

49

.8285;
82878838 :8 803:8 838—08 .3 Emma 8:25 :8 85

9 8:888: 85 fl $88 $5. .3326 888 86:88 3:63:82
8:8 888 R8 885:88 85 :0 8:2: 8:28.53 :8on

E 88?. :8 88582: 85 8 8:888: E $88 $5 82853 852
$838 :8 80658 8388:: km 8:8 2: w:_:::o::m $2822:
82w n 88w £2829: 2.3 u 883 mm 83$ 8:2: 2: .6 8:9:
3:8: 8:283:23 :8 A93 8:285; .30: 8:288: 85 E 888
a $88 :5. 8:88:88 n H :8 .3888 m.::w:§:om n « .588
:0 88:8: u > 3:388 8885 n a 88:3 ..5 u >: E 8288:: 85
mm: 888 8588: 28:: 85 833 :0 32m: 886:: 6:8:me0588
£88 :3 2:. £8.82: :8 :5 3 :52? .8: 25.552:

.3 88E 531.88: BEE—zwfiz :2, 385:838 2: 88m 55.58%
.388 12:85 85 a: 8.85:: 2: $538: Eauwfin— "fiN «Sumn—

vibrational

U)
C
(U
I_..
4...!
+
:9
>
+
4.:
O
L—

em radiation

50

    

\‘i‘; ,\
3'7“”?
‘?\/h’J

'3:
i 2

Figure 2.1

51

molecules and solvent-lattice vibrations. The membrane, due to its location around the cell,
has the opportunity to control the thermal energy spectrum that enters the cell. This is
because collisions with molecules surrounding the cell and solvent-lattice vibrations result
in energy being transferred to special vibrational modes of the membrane acyl chains called
longitudinal accordian modes (LAMS), which is the extension and compression of the acyl
chains. The net effect is such that the energy quanta are converted into highly tailored,
specific, oriented waves on passing through the membrane and into the cell. This energy
spectrum is then transmitted throughout the cell interior again via molecular collisions and
solvent-lattice vibrations. However, the energy that can be stored and propagated by a given
chain length is a function of the chain length, with longer chains storing higher energy

quanta. Therefore an acyl chain length distribution is equivalent to an energy distribution.

In a rigorous mathematical analysis of the relative energies of these LAM modes,
Hollingsworth showed that both the LAM energy distribution in the membrane and its
corresponding acyl chain length distribution were normally distributed, as described by the

Gaussian density function:

f(x) = { 1/[O(21t)"2] }e'l’2[(P-X)2/02

where f(x) is the probability, 0 is the standard deviation, p is the mean, and x is the acyl chain
length. He also showed that the most prominent energy stored in the most abundant acyl

chain length of the membrane was equal to kT at the growth temperature of the organism.

52

Thus it was proposed that altering of membrane acyl chain lengths allows organisms to

transmit the maximal available thermal energy spectrum [3].

It was also shown in this paper that not only is the composition of membrane fatty
acid chains normally distributed when ranked by chain length, but there is a relationship
between the variance (02) of this distribution curve and temperature as well. Namely, as
temperature decreases the distribution curve becomes taller and narrower (smaller 02). As
temperature increases the curve becomes shorter and broader (larger 02) [3]. The molecular
speed or momentum distribution of the molecules surrounding the cell, which is one aspect
of the kinetic energy (translational energy mode) distribution of the molecules, also has a
Gaussian type function described by the Maxwell-Boltzmann equation. This equilibrium
distribution function {fo(p)} describing the momentum (p) of a dilute gas at temperature T

has the form:

fo(P) : Ae-(p-pfi)2/2ka
where m is the mass of a gas particle, k is the Boltzmann constant, po is the mean momentum,
and A is a constant involving the total number of particles (11), m, k, and T. Because the
distribution function that describes molecular speed or momentum at a particular temperature
has the same form, a relationship between temperature in the Maxwellian distribution
function and 02 in the Gaussian density function is invited. This suggests several important
questions, the main one of which is: Is there a linear relationship between 02 and
temperature over the adaptation range of an organism? If temperature is sensed by

alterations of the phase structure of membranes, then membrane-active chemical agents that

53

disturb the packing and fluidize or melt certain regions should have the effect of increasing
temperature. On the other hand, those that cause a reduction or freezing of acyl chain motion
should have the effect of reducing temperature. These effects are not limited to organic
(hydrocarbon) soluble molecules. Hence antibiotics that interfere with the lipid headgroup
packing thus disturbing their long range order should have the same fluidizing effects as
increasing temperature. Therefore, there should be a temperature equivalence for a variety

of membrane-active chemical perturbants.

Modulation of phase equilibria is another aspect of membrane structure which may
enable cells to sense temperature variations. Membrane components are not uniformly
dispersed, rather they are organized into phases or domains which may differ in composition,
morphology, or both. The chemical make-up of the constituents (including lipid headgroups,
acyl chains, and membrane proteins) dictate the types and structures of the phases. These
domains exist in a very complex equilibrium between each other and with the bulk of the
membrane. The phase equilibria are extremely temperature sensitive. A small temperature
change can result in dramatic alterations in phase structures. Since the phase boundary is a
discontinuity, the specific heat should change abruptly across it and small changes in
temperature should be amplified. A change in temperature would affect gel to liquid-
crystalline and other phase transitions thus altering membrane fluidity. If the new dynamical
state does not adequately capture the thermal spectrum, it does so eventually. This is because
an essential characteristic of adaptation is that the molecular systems that capture the

available heat energy at the maximum of its distribution are the ones that are most viable.

54

One can, therefore, equate a change in temperature with a change in some dynamical

property of the membrane and vice-versa.

In this study we use a Siberian permafrost bacterium grown at four different
temperatures to determine the real dependence between 02 and temperature. We also
demonstrate the generality of this relationship using data from the literature regarding three
other Gram-positive organisms, Bacillus subtilis, Listeria monocytogenes. and Bacillus
amyloliquefaciens. We also investigate the effects of several solvents and antibiotics on
Arctic strain 45-3, Bacillus subtilis PY79, and Sarcina ventriculi JK to establish a
temperature equivalence for the effect of chemical perturbations on membrane acyl chain

distributions.

MATERIALS AND METHODS

Arctic strain 45-3 culture conditions
Arctic strain 45-3 was cultured in 100 ml cultures of trypticase soy broth at 4°C,

14°C, 24°C, and 30°C with shaking at moderate speeds.

Effects of cyclohexane and 1,4-cyclohexadiene on Arctic strain 45-3
Cultures (10 ml) of Arctic strain 45-3 were grown as described above except that 0.5
M and 1.0 M concentrations of cyclohexane (54 ul and 108 pl, respectively) and 1,4-

cyclohexadiene (47 pl and 95 ul, respectively) were added to individual cultures.

55

Cyclohexane was obtained from Sigma, and 1,4-cyclohexadiene was obtained from Aldrich.

Effects of cis- and trans-decalins on Bacillus subtilis PY79

Cultures (10 m1) of B. subtilis PY79 were grown to mid-exponential phase (LB
media: 10 g/l tryptone, 5 g/l yeast extract, and 10 g/l NaCl; pH 7.0) at 37°C with shaking
at 270 rpm. At this point 0.2 M (0.31 ml) cis—decalin and 0.2 M (0.32 ml) trans-decalin
(both obtained from Sigma) were added to individual cultures, and then incubated at 37°C
for another six hours. Cells were harvested by centrifugation. Fatty acid methyl ester

analyses were done as described below.

Effects of antibiotics on Bacillus subtilis PY79

A 2.5 1 culture of B. subtilis PY79 was grown, as described above, to mid-exponential
phase, and then split into 5, 500 ml cultures. Then 5 ml of 10 mg/ml stocks of
chloramphenicol (Sigma), streptomycin (Boehringer Mannheim), and tetracycline (Sigma)
were added to three of the cultures. One was a 37°C control. The last was an ethanol control
(5 ml added), since chloramphenicol was dissolved in ethanol. The cells were harvested by
centrifugation. The total lipids were extracted two times in 41114 CHCl3zMeOHszO by
stirring for 36 hours total. The organic layers were removed and evaporated to dryness.

Fatty acid methyl ester analyses were done as described below.

Fatty acid methyl ester analysis

About 1-2 mg of cells from cultures grown under the various conditions were

56

methanolysed in 5% HCl in methanol at 75°C for 36 hours. The suspensions were
evaporated to dryness under a stream of nitrogen, and 0.5 ml HZO, 1.5 ml hexane, and 1.5
ml chloroform were added. After vigorous shaking, the mixture was centrifuged and the
organic layer was removed and concentrated to dryness. The resulting fatty acid methyl
esters were analyzed by gas chromatography (GC) and GC-MS using a 30 m DB1 column.
The temperature program was a ramp from 160°C to 320°C at 3°/min. The final
temperature of 320°C was held for 5 min. The relative proportion of the fatty acid
components were calculated from the integrated peaks. The fatty acid methyl esters were
identified by GC-MS analysis using a Jeol JMS—AXSOSH spectrometer interfaced with a

Hewlett-Packard 5890A gas chromatograph.

Variance (02) calculations of acyl chain length distribution curves
Variances (02) of the acyl chain length distribution curves, which are normally

distributed were calculated using the Gaussian density function:

fix) = {1/[O(2n:)”2]}e'1’21<u-x)2/02

where f(x) is the probability, 0 is the standard deviation, u is the mean, and x is the acyl chain
length. By setting the value of x equal to u (the condition at the mean), the equation

simplifies to:

f(x) = l/o(21t)”2

57

which when solved for 0 becomes

0 = l/J’(x}l21t]“2

The variance is the standard deviation squared. The variances of each curve were then

plotted versus temperature.

RESULTS AND DISCUSSION

The gas chromatograms for Arctic strain 45-3 at 4°C and 24°C are shown in Figure
2.2A and B, respectively. The main series of acyl chains synthesized is the odd one, with the
anteiso-branched fatty acids predominating. In Figure 2.3, the odd series acyl chain length
probability distributions (inclusive of iso, anteiso, alkenyl, and straight chains of the same
length) are plotted for each growth temperature. These distributions had the expected
Gaussian profile with the curves becoming taller and narrower with decreasing temperatures.
The Gaussian character was easily verified by sampling the function at points equidistant
from and on opposite sides of the mean, and verifying that it had the same value. These
distributions are in agreement with those first shown by Hollingsworth [3], and support his
idea that the chain length selection process is random, and the mean and variance of these

distributions are determined by temperature.

To further support the generality of this relationship, fatty acid composition data for

58

Figure 2.2: Gas chromatographs of fatty acid methyl esters
obtained from Arctic strain 45-3 cultured at (A) 4°C and (B)
24°C. The main acyl chain components were (1) Cum (2) iso-
C14 (3) n--Cl4 (4) C15,, (5) anteisa-C15 (6) is0--C16 (7) n--Cl6 and (8)
anteiso-C”.

 

 

 

59

 

 

5
4
6
7
23 I?
5
67 8
234 J

 

 

 

Retention time ——-+

Figure 2.2

6O

 

 

 

 

 

 

 

 

 

0.9 4- -o- 40

0.3 .. +140

0.7 fl "*"24C
3‘ 0.6 ~- """300
:‘g 05+
E 0.4 T-

0.3 ‘1-

0.2 4*

0.1 4% K.‘

o 1 : : 4‘"- ~-_:-.-3H
12 13 14 15 16 17 18 19 20

Acyl Chain Length

Figure 2.3: Plot of probability distributions at 4°C, 14°C, 24°C, and 30°C for Arctic
strain 45-3.

61

three other Gram-positive organisms were taken from the literature and their probability
distributions plotted for the many series of acyl chains synthesized. The branched (anteiso
and iso) odd chain length series was plotted for Listeria monocytogenes (Figure 2.4A) [4],
and for Bacillus amyloliquefaciens (Figure 2.48) [5]. For Bacillus subtilis the anteiso, odd-
chain length series was plotted (Figure 2.4C) [6]. All three of these organisms have Gaussian
probability distributions in which the curves become shorter and broader with increasing
temperature. The peak of the distribution curve for B. subtilis at 35°C lies between the
curved for 20°C and 26°C, which does not follow the trend. This is the only discrepancy
found so far, and since this is from the literature it is unknown whether this is true or due to
experimental error. These results indicate, at least for Gram-positive bacteria, that the
random acyl chain length selection process is a general rule for these organisms as shown by
the Gaussian distributions, of which the mean and variance of the curves are determined by

temperature.

For L. monocytogenes and B. amyloliquefaciens, the means of the probability curves
also shift toward the right as temperature increases (Figure 2.4A and B, respectively),
showing that indeed longer chain lengths become more dominant. This suggests that bacteria
have "basins of stability," in which they can alter their acyl chain length distributions over
a certain range dictated by kT. At extremes of temperature (hot or cold) there are limits to
which the organism can adjust around a particular acyl chain length distribution. This would
be shown be the Gaussian profiles such that at high temperatures the curve would become

very broad (flatten out) while at low temperatures the curve would become very tall and

62

Figure 2.4: Plots of probability distributions at several
growth temperatures for (A) Listeria monocytogenes (B)
Bacillus amyloliquefaciens and (C) Bacillus subtilis.

Probability

63

 

IAALAAAJLLL‘A‘AA‘A

Probability
O
m

p
A
1

02‘

 

 

3..-
vvrrvwvrvv'v'vvvv'v—wvrrvvw—vvv

 

 

+4C
-¢-1OC
--.--37C

 

 

 

 

vvvvvvv

12 13

F)
(D

14 15 16 17

Acyl Chain Length

18

19

v v

20

 

C’
1:

AlAAAALLAAA

C’
V

 

 

 

 

 

 

 

 

rvvvvvv—v—r'vvrvtvvvv'vvvv'vvvvirvvrvvvv'

13

14 15 16 17

AcyIChalnLnngth

18

19

20

 

0.7 1

06~

05 ~

04 <

013*

02‘

 

 

 

 

T V V

15 16 17
Acyl Chaln Length

Figure 2.4

20

64

narrow. So as the temperature (kT) approaches the limits of this stability range, those
bacteria which can reestablish their acyl chain length distributions about a new mean (acyl
chain length) will survive. At higher temperatures this new mean would be shifted to longer
acyl chain lengths, and at lower temperatures the mean would be at shorter acyl chain
lengths. Consistent with this idea, the acyl chain length probability distributions for the
extreme thermophilic Gram-positive bacterium, Bacillus caldolyticus, is shown in Figure 2.5
[7]. As predicted, as the temperature increases the means of the distribution curves shift
from 15.1 to 17, indicating that at higher temperatures this bacterium is able to establish a

new distribution of acyl chains in which the mean is now reset to a longer fatty acid chain.

Since there is a definite pattern of height and width changes of the Gaussian
distribution curves with respect to temperature, we determined the variances (02), the
distribution about the mean, of these curves and plotted them versus the growth
temperatures. These graphs and the equations for the curves are shown in Figure 2.6 for
Arctic strain 45-3 (A), L. monocytogenes (B), B. amloliquefaciens (C), and B. subtilis (D).
Over the broad permissible growth range for Arctic strain 45-3 (Figure 2.6A), the
relationship between 02 and temperature is linear. The other three organisms also have
regions of linearity as well, particularly L. monocytogenes (Figure 2.6B) and B.
amyloliquefaciens (Figure 2.6C). However, towards the extremes of growth temperatures
for the organisms, the data for the variance plots have more pronounced curves. Thus, the
functions that best fit the data are all second order. In any event, these results show then that

for a given organism there is a smooth relationship between the variance of the acyl chain

65

0.9

 

 

-+—6OC

0.8 1
-I- 70 C

0.7 ~ --t--80 C

 

 

 

0.6 d

.0
U1
1

Probability
.0
b

0.3 1

0.2 1

0.1 -1

 

 

 

 

12 13 14 15 16 17 18 19 20
Acyl Chain Length

Figure 2.5: Plot of probability distributions at 60°C, 70°C, and 80°C for Bacillus
caldolyticus. Note the shift in the mean as the temperature increases.

66

Figure 2.6: Plots of the variances of the acyl chain
distribution curves versus growth temperature (in
degrees Kelvin) and the equations for the curves for
(A) Arctic strain 45-3, (B) L. monocytogenes, (C) B.
amyloliquefaciens, and (D) B. subtilis. Note that the
35 °C data point for B. subtilis was left out of the graph
for curve fitting purposes due to possible errors in the
data for that growth temperature (see text).

Variance

Variance

67

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

A B
0.23 0 47 ‘
0 22 0.42 : ,
0.21 0 37 ‘
o
o
02 S 032‘
'1:
a
0.19 > 0 27‘
0,18 0.22 ‘
017 017 ......., HY. ..
277 282 287 292 297 302 307 277 282 287 292 297 302 307 312
Growth Temperaure (Kelvin) Growth Temperature (Kelvin)
y = (7.70 x 105)x2 + (4.29 x 102)x + 6.15 y = (~2.86 x 105)x2 + (2.45 x 107)x - 4.42
C
04
0 38
0.36
0.34 3
0.32 .5
0 5
' >
0.28
0
0,24
0 22 0
293 295 297 299 301 303 307 309 311 288 293 298 303 308 313
Growth Temperature (Kelvin) Growth Temperature (Kelvin)
y = (-3.38 x10*‘)x2 + (2.12 x10‘)x - 33.1 y = (-6.11 x1t'r5)x2 + (3.81 x 102)x - 5.59

Figure 2.6

68

length distribution and the growth temperature. Therefore knowing the temperature, one can

predict the fatty acid composition over a temperature range for a particular bacterium.

The effects of two solvents, cyclohexane and 1,4-cyclohexadiene, on the membrane
acyl chain distribution of Arctic strain 45-3 was also determined. These two solvents were
chosen because they interact with the hydrocarbon chains and because L. Be’rube’ in our lab
demonstrated that cyclohexane had a rigidifying effect on S. ventriculi membranes while 1,4-
cyclohexadiene fluidized them. These solvent effects were confirmed by further analysis
using T1 NMR experiments (explained in detail in Chapter 5), in which these solvents mixed
with DPPC (DL-a-phosphatidlycholine, dipalmitoyl) vesicles indicated the terminal methyl
and methylene groups of the acyl chains had greater motional freedom in the presence of 1,4-

cyclohexadiene, but were more restricted in the presence of cyclohexane.

However, in the case of Artic strain 45-3 both cyclohexane and 1,4-cyclohexadiene
acted as fluidizers. This is best shown in the GC chromatograms (Figure 2.7 A, B, and C)
in which there is an increase in the amounts of longer acyl chains (namely anteisa-C17 and
n-Cm) in the presence of both cyclohexane and 1,4-cyclohexadiene as compared to the 24°C
control. This increase in saturated and longer chains fatty acids shows the solvents act like
high temperatures and fluidize the membrane. The difference in effects on the Arctic strain
45-3 and S. ventriculi bacteria can be explained by the structures of both the solvents and the
main membrane acyl chain structures. As shown in Figure 1.5, the greatest degree of acyl

chain motion is at the terminus of the chain in the middle of the bilayer. The T1 NMR

69

Figure 2.7: Gas chromatograms of fatty acid methyl
esters obtained from Arctic strain 45-3 cultured in
the presence of two solvents. (A) 24°C control, (B)
0.1 M cyclohexane, and (C) 0.1M 1,4-cyclohexadiene.
The main acyl chain components were (1) Cm (2) iso-
C14 (3) n-Cl4 (4) C15,l (5) am‘eiso-C15 (6) iso-Cl6 (7) n-C,6
and (8) anteiso-Cn.

7O

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.7

71

experiments discussed above, also show that the primary effect of these solvents takes place
at these methyl termini of the fatty acids. In Figure 2.8 the packing interactions of
cyclohexane and an (o-cyclohexyl acyl chain with a straight hydrocarbon chain are shown.
The chair conformations of both these cyclic hydrocarbons enable them to pack closely in
a nesting fashion with the methyl terminus of the straight hydrocarbon chain, effectively
reducing the motional freedom of the chain resulting in a rigidifying effect. 1,4-
Cyclohexadiene, which is a more planar molecule is not capable of this nesting arrangement
with the straight acyl chains and as a result fluidizes the membrane. This is what is observed
with S. ventriculi, which has primarily straight acyl chains. In the Arctic bacterium,
however, the main membrane fatty acids are anteiso- and iso-branched, and as such neither
cyclohexane nor 1,4-cyclohexadiene pack well with the branched acyl chain termini, and
both therefore would increase the random motion at the methyl termini. 1,4-Cyclohexadiene
was particularly toxic to the Arctic bacterium, such that the cells showed limited growth after
its addition. This is probably due to the more planar conformation of 1,4-cyclohexadiene
which disturbs the packing of the chain termini more, or to its greater diffusibility within the
membrane because of the absence of stable packing arrangements along the hydrocarbon

chain.

Further substance perturbation studies were done using cis- and trans-decalin and
three antibiotics (Figure 2.9) on B. subtilis PY79. The GC chromatogram for B. subtilis
PY79 at 37°C is shown in Figure 2.10. The primary membrane acyl chains are the branched,

odd chain length series. In Figure 2.11 the effects of cis- and trans—decalins on B. subtilis

72

Figure 2.8: Computer generated pictures showing the
packing interactions between a straight chain
hydrocarbon and cyclohexane and an w-cyclohexyl acyl
chain from two different views (A) and (B). Note the
trans conformations of the rings allows close packing in a
nesting fashion with the methyl terminus of the hydrocarbon
chain.

73

 

Figure 2.8

74

 

Figure 2.8 (cont.)

75

 

 

Figure 2.9: Structures of the chemical perturbants used studies with Bacillus subtilis
PY79. (A) cis-decalin, (B) trans-decalin, (C) chloramphenicol, (D) tetracycline, and (E)
streptomycin.

76

 

 

 

 

 

H

Figure 2.10: Gas chromatogram of fatty acid methyl esters of Bacillus subtilis PY79
from a 37 °C culture. The main acyl chain components were (1) iso-Cl4 (2) n-Cl4 (3) iso-Cls
(4) anteiso-C15 (5) iso-C16 (6) n-Cl6 (7) iso-Cl7 (8) anteiso-Cl7 (9) n-Cls.

77

 

0.80

 

--o--37c

  
  
  

0.70 1
- I- Cis-decalin

0.60 1

 

 

+ Trans-decalin

 

0.50 1

0.40 1

Probability

0.30 1

0.20 1

0.10 1

 

 

 

0.00

   

I

12 13 14 15 16 17 18 19 20
Acyl Chain Length

Figure 2.11: The effects of cis- and trans-decalins on the plot of the probability
distributions of Bacillus subtilis PY79.

78

acyl chains distributions is shown, in which cis-decalin has no perturbing effect on the
membrane while trans—decalin acts as a rigidifier. Again this can be explained by interaction
of the perturbants with the type of membrane acyl chains in this organism. cis-Decalin has
an L—shaped structure and as such the fluidizing effect of one ring is offset by the nesting
capability of the second ring, and thus would not disrupt the membrane packing anymore
than the branched acyl chains already cause. The two rings of trans-decalin, on the other
hand, are both chair conformers and can nest between the hydrocarbon chains along their
length causing an overall decrease in acyl chain motion. The effects of the antibiotics are
also related to their structures. Chloramphenicol and tetracycline are largely hydrophobic
and interact with the membrane acyl chains. Both of these antibiotics have planar ring
systems and would not pack well with the membrane acyl chains. As expected these two
antibiotics fluidize the B. subtilis membranes (Figure 2.12). In fact the large ring system of
tetracycline fluidizes the membrane to such an extent that the mean of the curve is shifted
towards longer chains lengths (similar to that seen by extremely high temperatures).
Streptomycin, though, is a polar molecule consisting of sugar derivatives. As a result it most
likely interacts with membrane headgroups such that the amidino groups, which are
positively charged, interact with the negative phospholipid charges and stabilize the
membrane. This headgroup stabilization would result in closer packing of the acyl chains,
rigidifying the membrane. Again this predicted effect is demonstrated with B. subtilis, as
seen by the taller, more narrow distribution curve for streptomycin as compared to the 37 °C

control (Figure 2.12).

79

 

 

  
  
  
  
  

 

 

 

 

 

 

0.9
"OHS? C Control
0.8 1
- I- EtOH Control
0.7 1 - 1&- - Chloramphenicol
0.6 q +Streplornyern
z. - ai- - Tetracyclin
E 0.5 1
.D
3
2 0.4 1
n.
0.3 1
0.2 1
0.1 1
O T T T l I If T
12 13 14 15 16 17 18 19 20

Acyl Chain Length

Figure 2.12: Effects of three antibiotics on the plot of the probability distributions of
Bacillus subtilis PY79.

80

These results indicate that the effect of chemical substances on acyl chain length
distributions can be explained based on their structures and the membrane acyl chain
components of a given organism. As a result there is a temperature equivalent shift in
membrane acyl chain length compositions based on whether the perturbant acts as a fluidizer
or rigidifier of the particular membrane. The ability of organisms to adapt to environmental
perturbations by alteration of their acyl chain distributions would be reflected in their growth
rate under these new conditions. These results indicate that there might be a relationship
between temperature, solvent concentration, and growth rate. In Figure 2.13 hypothetical
growth rate curves for conditions of high and low temperatures and solvent concentrations.
Such a three dimensional graph of these factors for a given bacterium would allow one to
determine the solvent concentration needed to offset the temperature shift to allow the
organism to grow outside its temperature range. A temperature offset could also be

determined to allow a bacterium to grow in the presence of higher solvent concentrations.

In living systems something must fuel the molecular motion of the cellular processes,
since these processes are not self-sustaining. This driving force is the heat (kinetic energy),
determined by temperature, available to the organism. Therefore, there has to be a match
between the thermal energy spectrum and the cellular chemical properties, and this match is
characteristic of a given organism. Organisms capture the available kinetic energy by
collisions with molecules whose momentum distribution is a Gaussian type function, the
mean and variance of which is determined by temperature according to the Maxwell-

Boltzmann distribution law. The collisions between the surrounding solvent and gas

81

 

 

&

¥

 

temperature

 

 

 

 

0 [solventL

 

“3.
O
2.
61
n
‘32:

 

Figure 2.13: A three dimensional graph showing hypothetical growth rate curves for
conditions of high and low temperatures and solvent (fluidizing type) concentrations.

82

molecules and the molecules in the boundary structures of the cells form the basis of the heat
exchange process. To effectively capture this kinetic energy distribution, the surface of the
cell must have an array of molecules or some molecular system that can absorb the energy
given up in these exchanges into some vibrational mode because the cellular chemistry is
driven by these vibrations. So then the best adapted organism should have some energy
harvesting system that traps this energy that is passed on in these collisions in vibrational
modes with the same mean energy as in the collisions and over a distribution that is narrowly
defined around the mean energy. Since the energy distribution of the collisions to start with
is Gaussian as well, it makes sense that the vibrational modes are also Gaussian, with the
most abundant acyl chain length in the membrane trapping the most abundant energy quanta

stored in the matter surrounding the cell.

The energy trapped by the membrane and propagated to the cell interior is used to
drive the cellular chemical processes. Not only does this energy determine the rate of
diffusion of molecules (and thus the collision rate), it also dictates various motions of
proteins that determines the activity of these proteins, and as a result the types of cellular
processes that can be carried out. Two types of large-scale motions observed in proteins are
hinge bending motions and single strand motions, both of which may produce large relative
displacements of atoms within proteins and are usually involved the biological activity of the
protein. As will be discussed the frequencies, by relation to the time scale of the motions,

of these motions are also of the magnitude kT.

83

In 1995, Hollingsworth showed that the frequency of the most abundant membrane
acyl chain length corresponded to 215.4 cm’1 (in the spectroscopic unit of wavenumbers) or
6 psec", which corresponded to kT at the growth temperature of the organism [3]. This
energy (as well as the rest of the energy distribution stored in the acyl chains) is then
transferred to the cell interior, mostly to the water molecules. Recently terahertz (Thz) laser
vibration-rotation-tunneling spectroscopy has been used to characterize the frequencies of
the water solvent cage. The first calculations were for trimers of water molecules, and they
found that the characteristic frequency (in wavenumbers) for one aspect of water vibrations
was 40-50 cm’l [8]. For water tetramers this frequency increased to 2.04 Thz (67.9 cm") [9].
This frequency continued to increase with the size of the water cage to 2.4 Thz (81.2 cm-l)
for pentamers [10] and 2.491 Thz (83.03 cm-l) for hexamers [1 1]. So one would expect that
as this water cage approached infinity, this water vibrational frequency would get larger and
eventually settle to a frequency probably in the range of 100-200 cm‘1 (since the rate of
increase in frequency between the pentamer and hexamer is smaller), which would
correspond to the characteristic energies transmitted by the membrane acyl chains through
their LAM modes. This is also only one aspect of water cage vibrations, and there are
certainly many others that would contribute to the total energy that is transmitted by the
solvent molecules. This energy stored in the solvent molecules in the cell interior is now
redistributed among the other low frequency modes of proteins and other molecules (such

as DNA and RNA) for use in carrying out many of the cell processes.

Many enzymes and binding proteins consist of several globular domains connected

84

by relatively flexible polypeptide chains [12-14]. The binding sites for substrates and ligands
are typically located between these domains, and thus the time scale and relative motion of
the domains would play an important role in the binding and activity of the enzymes
(proteins). These large scale motions would involve collective motions with the concerted
displacement of many atoms. Normal mode calculations done on the small protein bovine
pancreatic trypsin inhibitor (BPTI) showed that some of these large scale motions were low
frequency modes on the order of 3.6-6 ps'l (100—200 cm"), which typically involved motion
of several neighboring residues. More global motions have frequencies less than 3.6 ps‘1
(100 cm") [15]. Adiabatic mapping calculations for hinge bending motions, one type of
global motion, were first done for lysozyme, in which its active site is located in a cleft
between two globular domains [16]. These calculations, and further improved calculations
that take into account solvent damping effects, indicate that the hinge bending motion is
greater than 10 ps (less than 0.1 ps"). In fact with solvent damping effects, the hinge motion
is expected to be Brownian in nature with a typical fluctuation opening and closing the cleft
by 0.1 nm and will persist for about 20 ps [17]. In regards to other types of proteins,
fluorescence depolarization techniques were used to determine the time scale of hinge
bending motions of immunoglobulins is in the range of 10-50 ns [18, 19]. Also, indirect
evidence provided by X-ray crystal structures indicate that tRNA molecules may even have
hinge bending motions [20]. Initial investigations of this movement for tRNAPh" have been
done [21, 22], and molecular dynamics simulations show a periodic motion of the bending
between the two arms of the tRNAPhc molecule which is 24 ps [23]. It is expected that

solvent damping effects would increase this time scale to the characteristic nanosecond range

85

[12]. It has also been shown that several proteins have biologically important strand folding
and unfolding motions, which often occur in peptide segments near binding or active sites.
Upon binding of the substrate or ligand these segments assume ordered conformations in
which they cover or otherwise interact with the bound molecule [12-14]. For example, the
active site of trypsin, which consists of several polypeptide loops, is highly ordered; but in
its inactive precursor, trypsinogen, these loops are disordered. The apparent characteristic
time scale of this loop motion is about 11 ns [24], which is in agreement with studies done
on the growth kinetics of alpha helices [25]. The time scale of these global motions of
proteins and other macromolecules are limited by the viscosity of the solvent, which is water
in biological systems. Thus solvent damping impedes these global motions, and hence their
rates of motion are diffusion limited and are in a very low frequency range (less than 100

cm").

Based on all of this evidence, though some of it is indirect from various types of
simulations, one can perceive that the energy transmitted by the membrane acyl chains is
disseminated among the many vibrational modes of the solvent cage (and other molecules)
in the cell interior. These numerous solvent-lattice vibrational modes store and propagate
many frequencies of energy, in the low frequency range, to the proteins and other large
molecules in the cell, which are then stored in these molecules in their various vibrational

modes, thus allowing them to carry out the myriad cellular processes.

In this chapter, focus is on the acyl chain length heterogeneity of membranes and their

86

function in transmitting the available thermal energy spectrum. However, membrane lipid
acyl chain heterogeneity also includes differences in unsaturation, branching, and the
presence of carbocyclic rings. It has been shown that functionalization of an acyl chain by
single alkenyl, methyl, and cyclopropyl groups does not cause a significant difference of the
acyl chain LAM frequencies [26]. These functionalizations, though, do affect the packing
arrangement of the membrane acyl chains. Thus these alterations in acyl chain structure most
likely affect other aspects of membrane motional dynamics as is predicted by the
homeoviscous adaption theory. As a result these changes may play important roles in
maintaining other aspects of membrane function, such as their selective permeability
functions or control of certain enzymatic processes by altering the membrane domain

compositions.

REFERENCES

l. Lechevalier, M. P. ( 1982) Lipids in bacterial taxonomy. In CRC Handbook of
Microbiology, 2nd ed., Volume IV, eds. A. I. Laskin and H. A. Lechevalier. CRC Press, Inc.,
Boca Raton, FL, p. 436.

2. Sinensky, M. ( 1974) Homeoviscous adaptation--a homeostatic process that regulates the
viscosity of membrane lipids in Escherichia coli. Proc. Nat. Acad. Sci USA, 71: 522-525.

3. Hollingsworth, R. I. (1995) Membrane lipid alkyl chain heterogeneity is a molecular
transform of the available thermal energy spectrum. J. Phys. Chem, 99: 3406-3410.

4. Tadayon, R. A., and K. K. Carroll. (1971) Effect of growth conditions on the fatty acid
composition of Listeria monocytogenes and comparision with the fatty acids of
Erysipelothrix and Corynebacterium. Lipids, 6: 820-825.

5. Paton, J. C., E. J. McMurchie, B. K. May, and W. H. Elliott. (1978) Effect of growth
temperature on membrane fatty acid composition and susceptibility to cold shock of Bacillus

87

amyloliquefaciens. J. Bacteriol., 135: 754-759.

6. Suutari, M., and S. Laakso. (1992) Unsaturated and branched chain-fatty acids in
temperature adaptation of Bacillus subtilis and Bacillus megaterium. Biochim. Biophys.
Acta, 1126: 119-124.

7. Weerkamp, A., and W. Heinen. (1972) Effect of temperature on the fatty acid
composition of the extreme thermophiles, Bacillus caldolyticus and Bacillus caldotenax. J.
Bacteriol., 109: 443-446.

8. Viant, M. R., J. D. Cruzan, D. D. Lucas, M. G. Brown, K. Liu, and R. J. Saykally. (1997)
Pseudorotation in water trimer isotopomers using terahertz laser spectroscopy. J. Phys.
Chem, 101: 9022-9041.

9. Cruzan, J. D., M. R. Viant, M. G. Brown, and R. J. Saykally. (1997) Terahertz laser
vibration-rotation tunneling spectroscopy of the water tetrarner. J. Phys. Chem, 101: 9022-
9031.

10. Liu, K., M. G. Brown, J. D. Cruzan, and R. J. Saykally. (1997) Terahertz laser
spectroscopy of the water pentamer: structure and hydrogen bond rearrangement dynamics.
J. Phys. Chem, 101: 9011—9021.

11. Liu, K., M. G. Brown, and R. J. Saykally. (1997) Terahertz laser vibration-rotation
tunneling spectroscopy and dipole moment of a cage form of the water hexamer. J. Phys.
Chem, 101: 8995-9010.

12. McCammon, J. A., and S. C. Harvey. (1987) Global structural changes. In Dynamics
of Proteins and Nucleic Acids, Cambridge University Press, Cambridge. pp. 137-150.

13. Bennett, W. S., and R. Huber. (1984) Structural and functional aspects of domain
motions in proteins. CRC Crit. Rev. Biochem., 15: 291-384.

14. Petsko, G. A., and D. Ringe. (1984) Fluctuations in protein structure from X-ray
diffraction. Ann. Rev. Biophys. Bioeng., 13: 331-371.

15. G6, N., T. Noguti, and T. Nishikawa. (1983) Dynamics of a small globular protein, in
terms of low-frequency vibrational modes. Proc. Natl. Acad. Sci. USA, 80: 3696-3700.

16. McCammon, J. A., B. R. Gelin, and M. Karplus. (1976) The hinge-bending mode in
lysozyme. Nature, 262: 325-326.

17. McCammon, J. A., and P. G. Wolynes. (1977) Nonsteady hydrodynamics of biopolymer
motions. J. Chem. Phys, 66: 1452-1456.

88

18. Yguerabide, J ., H. F. Epstein, and L. Stryer. (1970) Segmental flexibility in an antibody
molecule. J. Mol. Biol, 51: 573-590.

19. Holowka, D. A., R. E. Cathou. (1976) Conformation of immunoglobulin M. 2.
Nanosecond fluorescence depolarization analysis of segmental flexibility in anti-e-l-

dimethylamino-S-naphthalenesulfonyl-L-lysine anti-immunoglobulin from horse, pig, and
shark. Biochemistry, 15: 3379-3390.

20. Westhof, E., P. Dumas, and D. Moras. (1985) Crystallographic refinement of yeast
aspartic acid transfer RNA. J. Mol. Biol, 184: 119-145.

21. Harvey, S. C., and J. A. McCammon. (1981) Intramolecular flexibility in phenylalanine
transfer RNA. Nature, 294: 286-287.

22. Tung, C-S., S. C. Harvey, and J. A. McCammon. (1984) Large-amplitude bending
motions in phenylalanine transfer RNA. Biopolymers, 23: 2173-2193.

23. Harvey, S. C., M. Prabhakaran, and J. A. McCammon. (1985) Molecular-dynarnics
simulation of phenylalanine transfer RNA. 1. Methods and general results. Biopolymers, 24:
1169-1188.

24. Butz, T., A. Lerf, and R. Huber. (1982) Intramolecular reorientational motion in
trypsinogen studied by perturbed angular correlation of 199"“Hg labels. Phys. Rev. Lett., 48:
890-893.

25. McCammon, J. A., S. H. Northrup, M. Karplus, and R. M. Levy. (1980) Helix-coil
transitions in a simple polypeptide model. Biopolymers, 19: 2033-2045.

26. Hsu, S. L., G. W. Ford, and S. Krimm. (1977) Longitudinal acoustic mode in polymers.
II. Solution of the general composite perturbed elastic rod. J. Polym. Sci., Polym. Phys. Ed.,
15: 1769-1778.

CHAPTER III

THE INTERRELATION BETWEEN TEMPERATURE-INDUCED SHIFTS IN
CARBON METABOLISM AND MEMBRANE LIPID CHEMISTRY

89

90

ABSTRACT

The available thermal energy, determined by temperature, drives the chemical
processes carried out in living organisms. The ability to utilize a particular substrate to
derive chemical energy (typically ATP) is dependent upon the transition state (activation)
energies of all of the steps in the metabolic pathway for that substrate. If the available
thermal energy is not enough to overcome the activation energy for one or more steps in the
metabolic process, the pathway requires significant energy investment or priming, and/or the
energy yield is too low, then the organism might not be able to utilize the given substrate
effectively. This is true for the more highly oxidized substrates like carbohydrates. As a
result, the balance of metabolism might conceivably shift to utilization of more reduced
substrates, like amino acids and fatty acids. The utilization of which might require lower
activation energies and certainly requires less of a chemical energy investment and can yield
much more chemical energy per gram. The notion that more energy is required to support
cellular upkeep at lower temperatures is well established and embodied in the concept of
maintenance energy. Based on this premise we investigated the ability of Arctic strain 45-3,
which thrives in extremely cold environments, to utilize glucose at 4°C and 24°C. One
special advantage of such a strain is that it allows a substantial spread of temperature. Most
organisms that are available in culture do not grow well below 15°C. Our results, using
NMR spectroscopy and radioactively labeled glucose, indicate that this strain utilizes glucose
much more slowly at 4°C than at 24°C. We also show that enzymatic control of metabolic
processes, such as carbohydrate and amino acid metabolism, must be done by regulation of

constitutively expressed enzymes because the total cellular protein profiles at the two

91

temperatures are essentially identical. The fact that this strain can use glucose more readily
at 24°C is in agreement with our observation that more straight membrane acyl chains are
produced at 24°C than at 4°C. These results support the idea that the available thermal
energy regulates the balance of chemical processes carried out in the cell, and this balance
determines the metabolic intermediate profile which is reflected in the membrane acyl chain

structure.

INTRODUCTION

As temperature fluctuates the amount of heat (infrared quanta) available for use by
cells changes in turn. This energy (kT), which is propagated into the cell interior by the
longitudinal accordian modes (LAMS) of the membrane acyl chains, provides the necessary
energy for various metabolic pathways [1]. In Figure 3.1 hypothetical energy profiles for two
different metabolic pathways are shown. At high temperatures (Tl) there is most likely
enough energy to carry out either metabolic pathway. At lower temperatures (T2), though,
the available energy may not be enough to get over the activation barriers of the pathway
indicated by profile A, so the cell has no choice but to use the metabolic pathway indicated
by profile B. In general the more oxidized substrates, such as carbohydrates, require more
chemical energy input in their metabolism. More reduced substrates such as amino acids and

fatty acids require less energy input.

The structure of the cell membrane reflects the type of metabolism carried out by the

92

 

 

Energy

 

 

 

 

 

 

Reaction progress

Figure 3.1: Hypothetical energy profiles for two different metabolic pathways are
shown, a higher energy requiring profile in (A) and a lower energy requiring profile
in (B). Note that at high temperatures (T,) there is most likely enough energy to carry out
either metabolic pathway. At lower temperatures (T2), though, the available energy may only
be enough to use the metabolic pathway indicated by profile B.

93

cell. Evidence of this is suggested by the literature documenting the greater occurrence of
branching in membrane fatty acids at low temperatures [2, 3]. The structures of these fatty
acids are determined by the acyl-CoA primers. For the branched fatty acids the primers
result from metabolism of the amino acids, leucine, valine, and isoleucine [4]. The tendency
to synthesize branched fatty acids at lower temperatures therefore could be interpreted to
mean that amino acid metabolism is more favorable than carbohydrate metabolism at low

temperatures.

Changes in metabolism can be accompanied by changes in either the activity or the
levels of synthesis in the enzymes involved. The controls of metabolic pathways could be
accomplished by de novo protein synthesis, or induction or suppression of constitutively
expressed enzymes. Changes in metabolism with temperature can also occur in the absence
of both these two control mechanisms. This is because at higher temperatures, there may be
enough activation energy to make some pathways active that are not active at low
temperatures. Here we studied the ability of an Arctic permafrost bacterium to utilize
glucose over its growth temperature range as well as monitor its protein profiles at two
temperatures. This bacterium, Arctic strain 45-3, utilizes glucose much more slowly when
grown at 4°C than at 24°C. The protein profiles for this bacterial strain at 4°C and 24°C are
essentially the same, indicating that the control of different metabolic pathways likely occurs

through thermal or constitutive enzyme control.

94

MATERIALS AND METHODS

Culture conditions of Arctic strain 45-3
Siberian strain 45-3 was cultured in trypticase soy broth (TSB) at 4°C and 24°C with

shaking at moderate speeds.

Analysis of trypticase soy broth before and after bacterial growth

Samples of TSB media before and after Arctic strain 45-3 was cultured at 4°C and
24°C (both cultures were grown to about the same optical density) were taken and
concentrated to dryness. The samples were then resuspended in 0.6 ml of DZO. 1H NMR
spectroscopy was done on each sample. Two dimensional DQF-COSY (double quantum
filtered-correlated spectroscopy) and TOCSY (total correlated spectroscopy) spectra were

also done for the unused TSB media sample.

Rate of utilization of carbon sources by Arctic strain 45-3

Five milliliter aliquots of TSB media were made, and 0.2 [Ci of l4C-glucose, l4C-
leucine, and l4C-valine (each) were added to the media. Then the media was inoculated with
Arctic strain 45-3, and cultures were grown at 4°C and 24°C. Samples (100 pl) were taken
before the cells were added and then every 4 hours for at least 36 hours (longer for the 4°C
cells). The cells were centrifuged out and then 20 ul of the media were spotted on TLC
plates. The metabolites on the TLC plates were separated using a solvent system of 7:3:1 1-

propanolzammonium hydroxideszO. The plates were then exposed to a phosphoimager.

95

The intensities of spots were determined using the integration routines available in the

IrnageQuantTM package.

Total protein profiles of Arctic strain 45-3 grown at 4°C and 24°C

Cultures of Arctic strain 45-3 at 4°C and 24°C were grown as described above. Total
cell lysates were prepared from 4°C and 24°C cultures by adding two volumes of 10 mM
Tris-HCl, pH 8.0 and 3 mg/ml lysozyme and incubating for one hour on ice. One volume
of 2X SDS-PAGE sample buffer solution [0.125 M Tris-HCL (pH 6.8), 0.04% sodium
dodecylsulphate (SDS), 10% glycerol, 5% B-mercaptoethanol, 0.0025% bromophenol blue]
was then added. The samples were mixed well, boiled for 5 minutes, and centrifuged briefly
to remove any precipitate. The lysates from 4°C and 24°C cultures were separated using
Laemmli’s discontinuous buffer system [5] on 12% polyacrylamide slab gels containing
0.32% bisacrylarnide, 0.375 M Tris-HCl (pH 8.8), and 0.1% SDS. The stacking gel
contained 4% polyacrylamide, 0.1% bisacrylamide, 0.125 M Tris-HCl (pH 6.8), and 0.1%
SDS. Electrophoresis was carried out at 80 V for about 3 hours. Gels were stained in a
solution of 0.1% Coomassie blue in 40% methanol/10% acetic acid, and destained in a

solution of 40% methanol/10% acetic acid.
RESULTS AND DISCUSSION

Arctic strain 45-3 has predominantly branched chain fatty acids indicating the use of

amino acid derived acyl-CoA primers for fatty acid synthesis. While amino acid metabolism

96

predominates in generating these primers throughout the growth temperature range, the
amount of C16 straight chain fatty acids increases by almost a factor of six in going from 4°C
to 30°C (Table 3.1), which is consistent with an increased utilization of carbohydrates. Also
at 24°C and 30°C, C18 straight chains (not shown) are detected. Based on the reasons that
the

Table 3.1: Percent of fatty acid methyl ester derivatives in the total methanolysate of

the cells grown at 4°C, 14°C, 24°C, and 30°C in order of elution. The numbered
components correspond to the numbered GC peaks in Figure 4.12.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Peak No. Structure % in 4°C % in 14°C % in 24°C % in 30°C

Cells Cells Cells Cells

1 CW 1.30 1.41 0.00 0.00

2 iso-CI4 2.65 3.18 2.24 1.08

3 n-Cl4 1.55 3.02 2.78 3.74

4 C15,, 17.58 9.09 0.87 0.28

5 anteiso-C15 62.67 64.53 68.82 60.87

6 iso-Cl6 8.72 10.22 10.38 6.31

7 n-C16 3.11 5.43 8.61 17.69

8 anteiso-C ,7 2.42 3.10 6.31 10.02

 

 

pathway requires significant energy investment or priming or the energy yield is insufficient
when the available thermal energy is too low, it appeared possible that carbohydrate
metabolism might be less effective at lower temperatures than other forms of metabolism.
We investigated the ability of Arctic strain 45-3, which survives in extremely cold climates,

to utilize glucose and amino acids at 4°C and 24°C. 1H NMR of TSB media before (Figure

97

3.2A) and after growth of Arctic strain 45-3 at 4°C (Figure 3.2B) and 24°C (Figure 3.2C)
are shown. The signals for the oc- and B-anomeric protons of glucose at 5.18 and 4.58 ppm,
respectively, have almost completely disappeared from the media of the 24°C culture, while

the relative decrease in these signal intensities for the 4°C culture is much smaller.

To further investigate this bacterium's ability to metabolize glucose, l4C-labeled
glucose, leucine, and valine were added to the TSB media for 4°C and 24°C cultures. Figure
3.3 shows the TLC separation of l4C-glucose, l4C-valine, and l4C-leucine for one time point
and its corresponding analog profile indicating the relative intensities of the three spots.
Figure 3.4 shows the extent of utilization of l4C-glucose by the bacterium at 4°C and 24°C.
The time points at 24 hours and 67 hours for the 24°C and 4°C, respectively, are when the
two cultures are at about the same optical density. The results from this radioactive label
experiment confirm the NMR results, which indicate that Arctic strain 45-3 utilizes glucose
at a greatly reduced rate at 4°C, while at 24°C the glucose becomes limiting very quickly.
Since the trypticase soy broth medium is largely peptides and amino acids (17 g/l pancreatic
digest of casein; 3 g/l papaic digest of soybean meal; and 2.5 g/l glucose), the percent
reduction of l4C-valine and l4C-leucine was small. These results, however, support our
hypothesis that the ability to use more oxidized substrates at low temperatures is impeded.
Thus the bacteria shift the balance of their metabolic pathways to use more reduced

substrates, amino acids for instance.

To determine whether this shift in metabolism is reflected in variations of this strain's

98

Figure 3.2: ‘H NMR of TSB media before and after growth
of Arctic strain 453. (A) Unused TSB media, (B) TSB
media from 4°C culture, and (C) TSB media from 24°C
culture. Most of the glucose signals are in the 3.2-3.6 ppm
range, while some of the free amino acid and polypeptide
signals are in the 1.0-1.5 ppm range. However, note the almost
complete disappearance of the (l) a-anomeric [5. 18 ppm] and
(2) B-anomeric [4.58 ppm] glucose signals in the 24°C culture,
as compared to the unused media and the 4°C culture.

99

 

 

 

1 2

l l
1 2
1 2

l 1‘ 1

Figure 3.2

 

 

Geranium LINE 1

 

 

glucose valine leucine

Figure 3.3: TLC separation of l‘C-glucose, l4C-valine, and 1“C-leucine for one time
point and its corresponding analog profile indicating the relative intensities of the three
spots.

101

50

 

 

251

Percent

201

151

 

101 ...--24c

 

 

+40

 

 

 

 

0 1o 20 30 4o 50 60 7o 80
Hours

Figure 3.4: The extent of utilization of l4C-glucose by Arctic strain 45-3 at 4°C and
24°C. The time points at 24 hours and 67 hours for the 24°C and 4°C cultures, respectively,
are when the optical densities for the two cultures are about equivalent.

102

enzyme production, total protein profiles from 4°C and 24°C cultures were determined.
These profiles are shown in Figure 3.5. The profiles for the two temperatures are essentially
the same, showing no appearance or disappearance of different proteins at the two
temperatures. The differences seen in some band intensities is not consistent upon repeated
culturing of the organism at the two temperatures. In many gels of the total cellular proteins
from subsequent cultures of this organism at the two temperatures, no significant differences
in band intensities was seen. This indicates that regulation of the metabolic pathways in this
organism must occur largely by induction or suppression of constitutive enzyme activity, or
by thermal regulation in which higher temperatures may provide enough activation energy

to make some pathways active that are not active at lower temperatures.

As has been discussed previously, there is a relationship between the types of
metabolism carried out by an organism and its membrane acyl chain structure. This
relationship is represented in Figure 3.6. The balance between carbohydrate and amino acid
metabolism, for instance, can be regulated by de novo enzyme synthesis, or by activation or
suppression of constitutively expressed enzymes, or by thermal regulation as described
above. The balance between these two types of metabolism determines the available acyl-
CoA primers for fatty acid synthesis. An increase in amino acid metabolism, in particular
the levels of leucine, isoleucine, and valine, would increase the pool of branched acyl-CoA
primers. In contrast, an increase in carbohydrate metabolism would increase the pool of

acetyl-CoA leading to straight acyl chains.

103

199 kDa m l

 

 

15 kDa

 

Figure 3.5: Total protein profiles for Arctic strain 45-3. Lane (1) low molecular weight
standards, (2) 4°C, and (3) 24°C.

104

 

Carbohydrate ‘ , Amino acid
metabolism metabolism
(more oxidized substrates) (more reduced substrates)
(higher activation barriers?) (lower activation barriers?)
(high initial energy investment) (low initial energy investment?)

\ /

enzyme suppression or de novo synthesis
versus thermal regulation of constitutive enzyme activity

1
1

Available acyl-CoA primer pool for
fatty acid synthesis

 

Figure 3.6: Schematic representation of the relationship of the balance between two
types of cellular metabolism, the potential thermal or enzymatic control of the
metabolic processes, and the pools of available acyl-CoA primers for fatty acid
biosynthesis.

105

The balance between these two metabolic pathways is controlled by the available
thermal energy. Since carbohydrates are more oxidized substrates, they generate less energy
per gram (669.94 kg-cal/g for glucose) than amino acids (855.6 kg-cal/g for leucine) [6].
Carbohydrates also require more energy to prime (in phosphorylation steps) and is oxygen
dependent. In low temperature climates not only is the available thermal energy reduced, but
the oxygen availability is also much less (due to decreased diffusion of oxygen through cold
or frozen soils). So as temperature drops the highly oxygen dependent and energy costly
metabolic pathways, i. e. carbohydrate metabolism, shut down in favor of utilization of the
more reduced substrates, amino acids and fatty acids. This is also consistent with the fact
that maintenance energy is higher at lower temperatures [7-9]. Therefore organisms that use
energy sources that yield more energy per gram would be more successful at low
temperatures as well. Therefore the balance of metabolic processes, which is determined by
the available thermal energy, is reflected by the structure of the membrane acyl chains. At
higher temperatures more straight chains are produced, while at lower temperatures more

branched chains are made.

Since the available thermal energy determines the balance of cellular metabolic
processes, the outcome of competing cellular reactions is determined by kinetics (the
activation energy barriers of the steps in the pathways) and not thermodynamics (the
difference in free energies of the starting materials versus the products). In other words, if
two chemical processes in the cell are competing, then the short-term outcome of the

competition is determined by the activation energy barriers the two processes face. The

106

ability to overcome these barriers is determined by the available thermal energy spectrum
that has been transferred to the cell interior. One reaction may have a lower activation
energy barrier, but the resultant product has a higher free energy content. So these products
are kinetic products, and they accumulate the fastest because of the lower activation barriers
to their formation and are not determined by the final free energies of these products. If the
cellular systems were closed systems and were in equilibrium, these higher energy kinetic
products, if there is some reversibility in the processes, would revert to the lower energy
thermodynamic products. However, cellular metabolic pathways are an intricate web of
reactions, and the kinetic products are often entrained in subsequent reactions such that the
entire process is determined by kinetics and not by thermodynamics (free energy differences).
Also in living systems, energy is constantly being fed in as heat, light, and/or nutrients, and
energy is constantly being given off in completely unpredictable patterns in swimming,
crawling, or walking. Thus an equilibrium state of cellular processes, as would be required
for thermodynamic control of chemical processes, is never reached. Consequently, the nature

of living systems is determined by kinetics, not thermodynamics.

REFERENCES

1. Hollingsworth, R. I. (1995) Membrane lipid alkyl chain heterogeneity is a molecular
transform of the available thermal energy spectrum. J. Phys. Chem, 99: 3406-3410.

2. Russell, N. J., and N. Fukunaga. (1990) A comparison of thermal adaptation of
membrane lipids in psychrophilic and thermophilic bacteria. FEMS Microbiol. Rev., 75:
171-182.

3. Suutari, M., and S. Laakso. (1994) Microbial fatty acids and thermal adaptation. CRC

107

Crit. Rev. Microbiol., 20: 285-328.

4. Kaneda, T. (1991) 180- and anteiso-fatty acids in bacteria: biosynthesis, function, and
taxonomic significance. Microbiol. Rev., 55: 288-302.

5. Laemrnli, U. K. (1970) Cleavage of structural proteins during the assembly of the head
of bacteriophage T4. Nature, 227: 680-685.

6. Weast, R. C., ed. (1985) Heat of combustion. In CRC Handbook of Chemistry and
Physics, 66th ed. CRC Press, Inc., Boca Raton, FL. pp. D272-277.

7. Buchanan, R. E., and E. I. Fulmer. (1928) Physiology and biochemistry of bacteria. The
Williams & Wilkins Co., Baltimore.

8. Rose, A. H. (1976) Chemical Microbiology: An Introduction to Microbial Physiology,
3rd Ed. Plenum Press, New York. pp. 245-325.

9. Russell, J. B., and G. M. Cook. ( 1995) Energetics of bacterial growth: Balance of
anabolic and catabolic reactions. Microbiol. Rev., 59: 48-62.

CHAPT ER IV
THE ADAPTATION OF ARCTIC BACTERIA TO LOW PRE-FREEZIN G

TEMPERATURES INVOLVES MEMBRANE ALTERATIONS BUT NOT
INCREASED PRODUCTION OF OSMOLYTES

108

109

ABSTRACT

An Arctic bacterial strain isolated from the Siberian permafrost grown at 4°C
produced undetectable levels of osmolytes such as glycerol, sugars, other sugar alcohols and
glycine betaine in the cytoplasm. The relative proportions in the total carbohydrate and
amino acid metabolic profiles between cells cultured at 4°C and 24°C using NMR
spectroscopy showed no significant changes. There was a net reduction by a factor of 4.6 in
the total amount of amino acids at 4°C compared to 24°C. In the cells cultured at 4°C,
major changes in membrane chemistry and cell surface chemistry were observed compared
to those grown at 24°C. There was a change in composition and a dramatic reduction in the
amount of polysaccharide capsule formed when cells were grown at 4°C compared to 24°C.
Because of this, cells from the lower temperature culture were hardly wettable. This would
negate the effect of freezing water on the membrane surface. Cells grown at the lower
temperature were smaller in volume by a factor of ~14. These results indicate that the major
mechanism for low temperature adaptation in this organism is alteration of the surface and

membrane structure.

INTRODUCTION

The processes bacteria use to allow them to survive and function at close to freezing
temperatures and below is a mystery that has fueled much interest and spawned several
theories. It is generally believed that there are two major physical or physico-chemical threats

that the bacterium needs to respond to in these temperature regimes. The first is ice formation

110

within the cell which might lead to cell lysis because of the volume increase on expansion
of water as ice is formed. The second is the increased salinity outside the cell as ice
formation leads to the separating out of pure water (as ice) and a corresponding increase in
salt concentration in the liquid state. It is thought that a general mechanism to counteract
both of these phenomena would be an increase of solute concentration inside of the bacterial
cell. A reasonable response then would be for the organism to increase the intracellular
amounts of certain metabolites especially glycine betaine, glycerol, mannitol and sorbitol
both as cryo-protectants and as osmolytes (1-3). There are certain to be other mechanisms
at work to protect against freezing since as temperature decreases the solubility of most
organic molecules and salts decreases. The decrease is rather dramatic in the case of
phosphates and carbonates. This solubility trend is in the opposite direction to that required
for freeze protection by increasing osmolyte concentration. There are other aspects of
bacterial cell structure that can contribute to this adaptation. These include membrane
structure, capsular polysaccharide structure and amount and cell size. For instance, the
membrane can be protected from freezing water by increasing its hydrophobicity. This can
be accomplished by changing the structure and reducing the amount of the hydrophilic
capsule. A change in structure might lead to a microenvironment that reduces the tendency
of water associated with it to freeze. Making the actual surface more hydrophobic by
removing charged groups from the lipids will reduce the interaction between the actual

membrane surface and external water.

Psychrophilic bacteria (organisms that are adapted to these near-freezing and below

111

conditions) are very good systems in which to examine the mechanisms of bacterial
adaptation to these extremes. In this study, we used a strain of bacteria that was obtained
from the permafrost sediment core of the age 60- 100 thousand years deposited in the Siberian
tundra zone on Kolyma-Indigirka Lowland (152-162°E and 68-72°N), see Figure 4.1. The
permafrost sample was received by rotary drilling without any solution or chemical reagents.
The temperature of the extracted core was no higher than -7°C. The methods of sampling,
storage, transportation and contamination control were described previously (4). Using these
methods, it was shown that the micro flora in the samples were there in situ and were not the
result of contamination. The organism studied was isolated from a sample at a core depth
of 4.0 meters and was named 45-3. The bacterium was a gram positive and according to its
morphological properties and 16S rRNA sequence (Jim Tiedje, Center for Microbial
Ecology, Michigan State University, Michigan; personal communication) was an
Arthrobacter sp. The bacterial morphology existed as rods and V-forms in young cultures
and coccus type in old cultures. The strain has a temperature range for growth of 4 to 30°C

and an optimal growth temperature of 23°C.

MATERIALS AND METHODS

Culture conditions
Siberian strain 45-3 was cultured in trypticase soy broth at 4°C and 24°C with

shaking at moderate speeds.

112

.Q 28 Soc 038mm 88 a So: @2202

£3 m-mv Edam Alembic ES mouefiém:
ecu—Bed axnmwmcEéEm—ov— 5325— 5893
-53: «o 0:8 «.655 2: E :Ewou wEEEam
3.58.3 «mamas—Lon car—2.5 “:0 9.53

113

3. 23E

0

RiVer

[ya

0
fl

2%
a

 

8985 f

 

-
A 582-3 .

 

114

Hydrophobicity analysis

Cells from 4°C and 24°C cultures were diluted to the same optical density. Glass
coverslips coated with chlorotrimethylsilane were dipped into the cultures for several
minutes and then rinsed thoroughly in a stream of fresh media. The coverslips were placed
on slides and the cells were then stained with several drops of 0.1% acridine orange for five
minutes. The coverslips were again washed thoroughly using fresh media. The cells were
viewed using rhodamine iso-thiocyanate (RITC) optics with excitation and primary barrier

filters at 529 and 550 nm respectively.

Total lipid extraction

Cells from the 4°C and 24°C cultures were stirred with 60 ml H20, 130 ml CHC13,
and 15 ml MeOH at 30°C for 12 hours. The layers were separated by centrifugation. The
aqueous layers were extracted once more with 30 ml CHCl3 for 15 min., and the organic
layers were then combined. The aqueous layers were again centrifuged and the clear
solutions were kept. The organic layers were analyzed by NMR spectroscopy and by thin
layer chromatography using 10:4:2:2:1 chlorofonn/acetone/methanol/acetic acid/water. TLC
analyses were performed on silica gel plates. The spots were visualized by spraying either
with orcinol or 10% phosphomolybdic acid in ethanol and heating at 120°C to visualize the

organic components.

Polysaccharide isolation
The aqueous layers from the total lipid extracts were concentrated to 4 ml and 12 ml

of ethanol was added. The polysaccharide which precipitated was recovered with a glass rod

115

and dissolved in 10 ml with water, and 10 mg MgClz, and 10 units each of RNase A and
DNase were added. The solutions were kept at room temperature for 2 hours. They were
then dialyzed for 4-5 hours, changing the water twice during that time, lyophilized and the
resulting solids were weighed. These polymers were analyzed by NMR spectroscopy and
by gas chromatography-mass spectrometry (GC-MS) after converting them to alditol acetate
derivatives [5]. The remaining aqueous alcoholic solution was centrifuged to remove any
precipitated solids. The supematants which contained free amino acids, carbohydrates, and
the like were removed and analyzed by NMR spectroscopy and a sample subjected to amino
acid analysis on a Waters Millipore system which included a WISP 7103 autosampler, a

Waters 510 pump, and a Waters 440 absorbance detector.

Fatty acid methyl ester analysis

Total cells from cultures grown at 4°C and 24°C were methanolysed in 2% HCl in
methanol at 75°C for 30 hours. The suspensions were blown to dryness, and 1 ml H20 and
2 ml hexane were added. After vigorous shaking, the mixture was centrifuged and the
hexane layer was removed and concentrated to dryness. The resulting fatty acid methyl
esters were analyzed by gas chromatography (GC) and GC-MS using a 30 m DB1 column.
The temperature program was a ramp from 160°C to 320°C at 3°/min. The final

temperature of 320°C was held for 20 min.

Lipid headgroup analysis
Approximately 5 mg of total lipids (isolated as described above) from cells cultured

at 4°C and 24°C were dried, and then 0.5 ml of 2 M trifluroacetic acid was added. The

116

lipids were hydrolyzed at 120°C for 1.5 hrs. The lipid headgroups were obtained by adding
0.5 ml of H20 and extracting twice with 2 volumes of CHC13. The chloroform layers were
removed and combined. The aqueous layers, containing the lipid headgroups, were dried and

analyzed by NMR spectroscopy.

Protein analysis by SDS-PAGE

Total cell lysates were prepared from 4°C and 24°C cultures by adding two volumes
of 10 mM Tris-HCl, pH 8.0 and 3 mg/ml lysozyme and incubating for one hour on ice. One
volume of 2X SDS—PAGE sample buffer solution [0.125 M Tris-HCL (pH 6.8), 0.04%
sodium dodecylsulphate (SDS), 10% glycerol, 5% B-mercaptoethanol, 0.0025%
bromophenol blue] was then added. The samples were mixed well, boiled for 5 minutes, and
centrifuged briefly to remove any precipitate. The lysates from 4°C and 24°C cultures were
separated using Laemmli’s discontinuous buffer system [6] on 12% polyacrylamide slab gels
containing 0.32% bisacrylarnide, 0.375 M Tris-HCl (pH 8.8), and 0.1% SDS. The stacking
gel contained 4% polyacrylamide, 0.1% bisacrylarnide, 0.125 M Tris-HCI (pH 6.8), and 0.1%
SDS. Electrophoresis was carried out at 80 V for about 3 hours. Gels were stained in a
solution of 0.1% Coomassie blue in 40% methanol/10% acetic acid, and destained in a

solution of 40% methanol/10% acetic acid.

Electron microscopy
Arctic strain 45-3 cells from 4°C and 24°C cultures were negatively stained with

1.5% potassium phosphotungstate and viewed by transmission electron microscopy.

117

RESULTS

The morphologies of the cells cultured at the two temperatures were very different.
The cells that were cultured at 4°C were waxy and oily. These cells would not form a
suspension even after prolonged agitation. In contrast, the cells cultured at 24°C were fluffy
and readily suspendable. This indicated that there was a significant layer of polysaccharide
material on their surface and that the 4°C cells did not have much of a polysaccharide
capsule. This was confirmed by the weight of capsular polysaccharide material obtained from
the cells. Even though the packed volume of the 4°C cells was over twice that of the 24°C
cells, the yield of polysaccharide in the latter cells was 2.4 times greater than from the 4°C
cells. The cells grown at 4°C were also much more susceptible to osmotic shock in a
hypotonic medium compared to those grown at 24°C. Hence these cells were completely
lysed on storing at 10°C in distilled water in less than 4 hours whereas the cells cultured at

24°C were still intact after 2 days under the same conditions.

Transmission electron microscopy revealed that there was a dramatic difference in
cell sizes for bacteria grown at the two temperatures. Those grown at the higher temperature
had twice the length and more than twice the diameter of the cells grown at the lower
temperature. Since bacteria were rods at both temperatures, this represented a difference in
volume of a factor of ~14 (Figure 4.2). Electron microscopy also indicated that the 4°C cells
had only a microcapsule and had a very non-polar membrane surface since these took up the
ionic stain only with difficulty. In agreement with these results, the hydrophobicity

experiment showed that more cells from the 4°C culture (Figure 4.3A) attached to the

118

 

 

Figure 4.2: Electron micrographs (at the same magnification) of Arctic strain 45-3. Bar
= 0.66 pm. (A) Cells cultured at 4°C. (B) Cells cultured at 24°C. Note that the 24°C are
about two times the width and length of the cells grown at 4°C.

_119

 

Figure 4.3: Fluorescent micrographs using RITC optics (at the same
magnification) of Artic strain 45-3 stained with 0.1 % acradine orange
attached to silanized coverslips. (A) Cells cultured at 4°C. (B) Cells
cultured at 24°C.

120

silanized coverslips. than from the 24°C culture (Figure 4.3B), which indicated the 24°C

cells have a more polar surface and were unable to attach to the coverslips.

Proton N MR spectroscopy was used to provide a comprehensive determination of the
relative amounts of the various metabolic intermediates elaborated by strain 45-3 at 4°C and
24°C. Analyses were carried out on the supernatant of the ethanol precipitation of the
aqueous layer which would contain free amino acids, sugars, sugar alcohols, and nucleotides.
Proton NMR spectra of these fractions from the 4°C (Figure 4.4A) and 24°C cells (Figure
4.48) indicated that the major components were amino acids. Two dimensional DQF-COSY
(Figure 4.5) and TOCSY (data not shown) experiments showed that these were glutamate,
valine, glycine, methionine, threonine, and a compound tentatively identified as 3-
aminobutyric acid. These results were in agreement with amino acid analyses. From the
NMR spectra, the relative proportions of the amino acids in each culture did not change
significantly between cells cultured at the two temperatures, which was again confirmed by
the amino acid analyses in which the relative amounts varied by about 1-4%. Signals for
common osmolytes such as free sugars and sugar alcohols such as glycerol, sorbitol and
mannitol, which have proton shifts in the 3.0 - 4.5 ppm region were noticeably absent or at
most present in only trace amounts (Figure 4.6). Signals for the methyl group of glycine

betaine (~3.2 ppm), another common osmolyte were also absent.

Gas chromatography mass spectrometry analysis of the capsular polysaccharides
obtained from the bacteria grown at 4°C and 24°C indicated that they contained fucose,

ribose, mannose, galactose, glucose, and glucosamine (Figures 4.7A and B). There were

121

Figure 4.4: 1H NMR spectra of the cytoplasmic metabolites. (A) Cells
cultured at 4°C. (B) Cells cultured at 24°C. The peaks are labeled with the
one letter amino acid codes. The major components were glutamate (E),
valine (V), glycine (G), methionine (M), threonine (T), and a compound
tentatively identified as 3-aminobutyric acid (see also figure 5.5). Note that
there are no significant differences in the relative pr0portions of amino acid
components.

122

 

 

 

 

 

 

T Y Y I 7 1* f v r T r r v y I r r Y I Y r v v I v r v v I v I v r I v I v I I v 1*! '.-_' \

 

 

 

 

 

Figure 4.4

123

.880 Boa 255 5:2 one 05 55 @223
8a 33m $80 2:. .009 “a con—5.3 £8
59: mom—onus... flame—98.3 2: «e 55.58%
VmOUd—On— Enema—35:. 95.“. "me 953%

124

m... 23E

 

v a 3“). n
v ’ TV. .If i! .p .

v =2- .9.

V was: . 8.. ..m. .
as ....

v a u. 1 a;

V I"

 

 

 

 

 

 

fabuwh. u. .n ...... “aile....3. 1&3... 31......5... u..?.«.........»t. 1.9.3... 18.

 

.57 .I .o’ar’ uéofibsn ..o\\\....

 

 

‘r r-Y—r—r‘r-r—i‘v Fr"! 1 r1‘1”'1'r"1"1 r r
In

In.
.ro.

1m.

I I
O

in

O

O

 

125

Figure 4.6: 1H NMR spectra of some
common osmolytes. (A) glucose (B) glycerol
(C) mannitol (D) sorbitol.

126

 

 

 

 

V I y—r f+1 Y Y Y

—4
—1
4

O 7 6 5 C

_" {fi- 1 Y fT v Y V r 1' v v Y 7 r v Y 1‘1 1

 

 

 

Y

4*,“ B
l

f' T I V r T r I V Y Y Y I V fl r I Y Y TfT ‘

l
5 3 2 pp.

... .- __ u L ll A....IL . u ._ 4].-..
—w . -w— - v—fi 7 -—~ ._..__' “V7 v v —~——— v—vw v.- v
‘ v v f 1 l v v v—r 1 v Y T—Y
5 I 3

 

 

 

 

I V V ‘7 Y I ff V Tfi ' Vfi'fi I V Y I ' 1' V f Y 1 I ‘

2 1 PP

Figure 4.6

127

Figure 4.7: Gas chromatograph profiles of
the capsular polysaccharides from the
bacteria cultured at (A) 4°C and (B) 24°C.
The peaks indicated are ( 1) fucose, (2) ribose,
(3) mannose, (4) galactose, (5) glucose, and (6)
glucosamine. (Note: The peak seen at 58 min.
was due to a transient shift in the baseline.)

2.5E6
2.8[8
1.5EB
.BEB
5.0ES

Abundance

1.8E8
.BES
.BES
.BES
.QES

Abundance
m A 07 CD

 

128

 

 

 

3
A
4
2
l
,1J.i ,. 1. ,. .L,- . ..
10 20 3G 49 SB 68
Time(min.)
3
B
2
5
l 4
J M- r .. .
10 20 30 4B 50 80 7a
Time(min.)

Figure 4.7

k...’

129

quantitative differences in the relative proportions between the cells from the two different
cultures. Mannose was the major component in both cases. Ribose was the next major
component in the case of the cells grown at higher temperature whereas galactose and
glucose were the next major components in the cells grown at lower temperature. Another
major difference was that the polymer from the 4°C cells contained a greater amount of

glucosamine.

Thin layer chromatography of the total lipids from the 4°C and 24°C cells showed
three major components namely a fast moving component which was tentatively identified
as triacylglycerol, diacylglycerol and free fatty acids. The latter two had retention times
consistent with those observed for standards run under the same conditions. The levels of
triacylglycerol were much higher in the cells grown at 4°C and appeared as a methanol
insoluble oil that was soluble in chloroform or hexane. In the case of the 4°C cells,
triacylglycerols accounted for more than half the total lipids. Diacylglycerol was the next
predominant component, while only traces of phosphoglycerol were present. Other minor
components were observed in the lipid preparations from the 24°C cells but the quantities
were too small to allow characterization. Triacylglycerol probably serves as an energy source
and not as a membrane component. Its presence, though, precluded a meaningful estimate
of the total amount of membrane lipids obtained from the 4°C cells. Triacylglycerol is only
sparingly soluble in methanol, however, and this was used as a way of separating it from the
total lipids for NMR analyses. The recovery of lipids (excluding triacylglycerol) from the

4°C cells was ~6 mg per gram (wet weight) and about half this amount for the 24°C cells.

130

The 'H NMR spectra of the total lipids from the 4°C and 24°C cells are shown in
Figure 4.8A and 4.8B, respectively. The two dimensional DQF-COSY and TOCSY
experiments (not shown) indicated that the signals between 5.30 and 5.43 ppm were due to
vinyl protons because of correlations with those at ~2.05 ppm. These were much more
abundant in the lipids from the 4°C culture than from the 24°C culture. Connectivities
between signals assignable to diacylglycerol and what later proved to be triacylglycerol were
also evident. In the spectrum of the 24°C lipids, extra signals due to the glycerol group of

phosphatidyl glycerol at 3.71 and 4.12 ppm were also observed.

In agreement with the TLC analyses and the 1H NMR spectra of the total lipids, the
1H NMR spectra of the isolated lipid headgroups of the 4°C cells (after precipitating
triacyl glycerol) showed definitively that the only lipid headgroup was glycerol (Figure 4.9A).
The two doublet of doublets at 3.47 ppm and 3.57 ppm were assigned to the C1 and C3
protons of glycerol. The multiplet at 3.7 ppm was assigned to the proton on carbon 2 of
glycerol. The lipid headgroup components of the 24°C cells were glycerol and
phosphoglycerol (Figure 4.9B). The chemical shifts of the glycerol signals were the same
as mentioned above. Two doublet of doublets at 3.65 ppm and 3.76 ppm were assigned to
the Cl and C3 protons of phosphoglycerol. The multiplet at 3.84 ppm was assigned to the
proton on carbon 2 of phosphoglycerol. From these data the proportion of phosphoglycerol
was determined to be 40% by integration of the peaks. This method is much more reliable

than charring the TLC plates and integrating densitometer traces.

1.
m

 

131

Figure 4.8: 1H NMR spectra of the total
lipids extracted from the bacteria grown at
(A) 4°C and (B) 24°C. Note the much higher
relative proportion of vinyl protons relative to
the head group protons in the 4°C cells. TAG =
triacylglycerol, DAG = diacylglycerol, PG =
phosphatidylglycerol.

132

 

 

 

 

[Eu—0| n—UII

 

 

 

I

 

ppm

 

 

 

lfilmulomol

 

 

 

O<n

 

2.135.111!

10181 \.

 

 

Figure 4.8

133

Figure 4.9: 1H NMR spectra of the lipid headgroups from the bacteria
grown at (A) 4°C and (B) 24°C. The only lipid headgroup in the 4°C cells
was glycerol. The two doublet of doublets at 3.47 ppm and 3.57 ppm were
assigned to the C1 and C3 protons of glycerol. The multiplet at 3.7 ppm was
assigned to the proton on carbon 2 of glycerol. The lipid headgroups of the
24°C cells were glycerol and phosphoglycerol. The chemical shifts of glycerol
were the same as mentioned above. The two doublet of doublets at 3.65 ppm
and 3.76 ppm were assigned to the Cl and C3 protons of phosphoglycerol. The
multiplet at 3.84 ppm was assigned to the proton on carbon 2 of
phosphoglycerol.

134

 

 

 

 

 

Wr'fi’YTTIYyVYVVIv:ITIV—fr‘rf‘rvvrrivaT‘YvVjvvrijTYervrvlrvvf‘T‘l—TT’TTTTTTTWTV—T—m
4.2 4.0 3.: 3.6 3.: 3.2 3.0 2.0 ppm

 

 

 

 

 

‘r—T'T'WTY'T—T‘fi-Y’rvv[rrvvtvfirfTv'vrI—IIIII—rfivfifwrv11rvvIIYTr111vIVIVIYIIIIrrjvv—trlvrvryvvverYY—r—l
4.2 4.0 3.8 3.6 3.4 3.2 3.0 2.. pp

Figure 4.9

135

The total lipid sample from the 4°C cultured cells contained a component that was
insoluble in pure methanol. With the aid of 'H NMR spectroscopy using homonuclear
decoupling (Figure 4.10) this component (in CDC13) was identified as triacylglycerol. Three
sets of signals in the downfield region could be assigned to a glycerol moiety using the
coupling information from the double resonance experiments. A multiplet at 5.25 ppm was
assigned to the H2 of glycerol. The C1 and C3 protons appeared at 4.1 ppm and 4.3 ppm.
The identity of this peak was confirmed by mass spectrometry using electrospray ionization
(Figure 4.11). The peak at m/z 605 is due to the protonated form of a triacylglycerol species
containing two C12 and one C10 fatty acid with a total of 3 double bonds between them.
Other homologous species separated by 14 or 28 mass units i.e. 1 or 2 methylene groups (e.g.

m/z 591 and 619) were also present.

The gas chromatograms of the fatty acid methyl esters of the membrane lipids from
the 4 and 24°C cells are shown in Figures 4.12A and 4.123. GC-MS analysis of these
methyl esters indicated the major lipid in both cell types was a anteiso-Cls lipid (peak 5).
However, the 4°C grown cells contained a large amount of a C”.l lipid (peak 4), and in
general had shorter chain lengths and a greater degree of unsaturated and branched chains
(Table 4.1). The cells cultured at 24°C contained a much smaller amount of the C 15.1 lipid
(peak 4), while the relative proportions of iso-Cl6 , n-C16 and anteiso-Cl7 (peaks 6, 7, and 8

respectively) increased over the 4°C cells.

136

 

 

 

 

 

 

 

Figure 4.10: 'H N MR spectrum of the methanol insoluble component from
cells cultured at 4°C. The two insets show double resonance experiments carried
out at 4.3 ppm and 5.25 ppm, which indicate the component was triacylglycerol.
The multiplet at 5.25 ppm was assigned to the H2 of glycerol. The C1 and C3
protons appeared at 4.1 ppm and 4.3 ppm.

137

.2805

cm? 803 83 93 Km NE. .wé SE: 3.2:

mm .0 E .3 3.838 360% msowoBEo:
850 .805 5953 £289 05:2. m .0 .38 a
5.3 Box .33. EU 25 US. 20 9.3 @8868
$6on 352:3»an a mo 85.. BESSE on.
o. 26 m. moo ES .m .33 on... .00.. an :39.»
239 E9... 3.392353... 2: we 55.58%
8...: b.3858? :3 @332..— uzé 0.53..“

138

@086
+mm snow

a. .2.

anu m..... u 0.30

2.. as”...

 

MVON. mvewm
0.9 B m. an
\
38;
w.moo

 

 

 

I
M _ i : _
wdm. tam NNNN
. NdaN
. 5. 33 23
v.93. \
«Non n63”
Qua“ Ndww
flown
mdnn

 

”.8.

.CN

NNN...

 

loo _.

139

Figure 4.12: Gas chromatographs of fatty
acid methyl esters obtained from bacteria
cultured at (A) 4°C and (B) 24°C. The main
acyl chain components were (1) C14.l (2) iso-Cl4
(3) n-Cl4 (4) C”.l (5) anteiso-C15 (6) iso—C16 (7)
n—Cl6 and (8) anteiso-Cn.

 

 

140

 

678

J -

Retention time ———.

234

 

 

Figure 4.12

141

Table 4.1: Percent of fatty acid methyl ester derivatives in the total methanolysate of
the cells grown at 4°C and 24°C in order of elution. The numbered components
correspond to the numbered GC peaks in Figure 4.12.

 

 

 

 

 

 

 

 

 

 

 

 

 

Peak No. Structure % in 4°C Cells % in 24°C Cells

1 C 14:1 1.75 0

2 iso-C14 3.55 1.89

3 n-Cl4 2.38 1.78

4 Cm 17.82 0.87

5 anteiso-Cl5 58.95 67.04

6 iso-Cl6 7.9 10.39

7 n-C 16 4.7 8.72

8 anteiso-C17 2.96 9.31

 

 

142

Total lysates from cells cultured at 4°C and 24°C were run on 12% SDS-PAGE gels (Figure
4.13). There were no significant changes in the protein profiles, either in new proteins
produced or suppressed, from the two cell cultures. Also there were no significant changes
in the amounts of the various proteins produced at both temperatures, as was confirmed by
running subsequent gels of total cellular proteins from other cultures grown at these two

temperatures.

DISCUSSION

As the growth temperature of bacteria is reduced two possible scenarios can occur
as the freezing point is reached. In the first scenario, pure water should separate out as ice
leading to a sudden increase in the extracellular salt concentration. In the second scenario,
if freezing does not occur, the ionic strength of the medium should actually decrease since
the dielectric constant increases. In addition, some salts might actually fall out of solution
because solubility is generally lower in cold water. This would also result in a decrease in
ionic strength. The ionic strength of the external medium therefore decreases progressively
as the freezing point is reached from above and then it suddenly increases at freezing. There
is no way that this can be anticipated by the bacteria and so it is not surprising that if
compatible solutes are synthesized to cope with the sudden increase in ionic strength on
freezing, none are synthesized even a few degrees above freezing. Hence, in this study, the
relative proportions of the components in the amino acid pools did not vary significantly
between cells cultured at 4°C and 24°C and the total mass of metabolites per unit volume

of packed cells decreased by a factor of 4.6 in going from 24 to 4°C. It is important to note

143

199 kDa W .

 

15 kDa

 

Figure 4.13: SDS-PAGE gel of total proteins of lysates from cells cultured at 4°C and
24°C. Lane (1) Low molecular weight standards (2) 4°C lysate (3) 24°C lysate.

144

here that the 4°C cells were much smaller than the 24°C cells and more of the former are
therefore present per unit volume of wet cells. This is also borne out by the fact that the total
lipids recovered per unit gram of the 4°C cells was much higher than the quantity recovered
from the 24°C. In other words, there are many more cells per unit volume or wet weight
present in the 4°C pellets compared to the 24°C pellets but the total quantity (as determined
by NMR spectroscopy) is much less. Electron microscopy definitively showed that there was
no breakage (thus leakage) of the cells at 4°C thus accounting for the lower amounts of

metabolites.

In addition to these findings (and consistent with them) there was no evidence for the
presence of such common compatible solutes as glycine betaine, glycerol, glucose, sorbitol,
and mannitol in the 4°C cells. This was definitively determined by amino acid analysis and
by NMR spectroscopy. Signals for these metabolites were completely absent from the NMR
spectra of the 4°C cells. These results indicated that, at least in this case, the induction of
catalytic systems (enzymes) involved in the synthesis of metabolites is not triggered by low
temperature, when within a few degrees of freezing. This was also clear from SDS—PAGE
where there were no new protein bands and no increase in the relative amounts of bands

found in 24°C cultures was observed for the 4°cells.

The large decrease in the intracellular metabolite and general ion concentration of the
organism as the temperature neared freezing had a predictable impact on the chemistry of the
bacterial membrane. The low solute concentration means that high concentrations of ions are

not available to act as counter-ions for the membrane lipids. As a result, the synthesis of

145

charged lipid components was almost totally suppressed. Only diacylglycerols and
triacylglycerols (which were likely contaminants of the membrane preparation) were found
in the membrane of the 4°C cells but the 24°C cells contained both phosphatidyl glycerol
and diacylglycerol. Smaller amounts of triacylglycerol were also found. The degree of
unsaturation was higher and the average length shorter in the fatty acid chains of 4°C cells
compared to 24°C as was expected if the average motional dynamics (fluidity) of the
membrane lipids was to be conserved. Because no charged lipids are present in the
membranes of the 4°C cells, there is no need for counter ions to be localized there to
stabilize the lipid headgroups. So these counter ions would now be in the actual free
circulating ion concentration in the bulk solution. Thus the cell now has a significantly
higher free circulating ion concentration without a real increase (by metabolite synthesis) in

ion concentration. This is another beneficial strategy for adaptation.

Another major change between the bacteria cultured at the two temperatures occurred
in the amount and composition of the capsular polysaccharides. The mass of the capsular
polysaccharide coating in the 24°C cells was reduced by a factor of 4.8 per unit packed cell
volume in the 4°C cells. A shift in the relative proportions of the carbohydrates making up
this capsule was also observed. This decrease in capsular polysaccharide is a very reasonable
observation. As the temperature decreases, the motional dynamics and diffusivity of
molecules slow as well. The presence of a thick capsular polysaccharide coating would
impede water flux through the cell and prevent nutrients from entering or metabolic waste
from leaving. This reduction in capsule size is another critical adaptative mechanism that

is probably as important as osmotic balance.

.

g
i.

 

146

From the above discussion, it can be concluded that the conditions of high salinity
in the extracellular medium that might accompany freezing cannot be anticipated by the
cellular mechanisms that respond to osmolality even fractions of a degree before freezing.
In fact, the indications of an incipient increase in salinity on freezing are quite to the contrary.
The imbalance in osmolality that accompanies the freezing event is therefore a sudden and
catastrophic one. The proposed shift in metabolite production of the organism in response
to this freezing event must therefore occur after the fact. Most of the adaptative events prior
to freezing revolve around maintaining the functioning of the cell membrane and coping with
the low ionic strength and diffusivity of the medium. Perhaps these conditions can help
counter a sudden increase in external electrolyte concentration since as we have stated
earlier, the absence of charged groups on the membrane lipids increases the availability of
ions to act as osmolytes. Such an adaptative mechanism to negate the effects of freezing of
the extracellular matrix is more in keeping with the ion and metabolite concentrations of the
cell just prior to freezing. The ordering of water at the membrane surface should also
increase as temperature drops providing further freeze protection. The membrane, therefore,
plays a central role in the physiology and physico-chemistry of cold adaptation. According
to SDS-PAGE, there was no dramatic increase in new protein synthesis on going to 4°C
from 24°C and so anti-freeze proteins, which would need to be produced in large quantities,

cannot be involved.

One limitation of the cryoprotection by colligative property model is that it treats the
environment inside of the cell and in its immediate vicinity as if the water there behaved as

“bulk” water and its properties could be controlled strictly by colligative effects. It is possible

147

that the tendency for water to freeze can be reduced by interaction with the membrane
surface. In fact, the ordering of water at regular hydrophilic surfaces is a well recognized
phenomenon (7). The fact that salt in aqueous solutions is excluded from an ordered surface
is utilized in desalination systems. In such systems, water is passed through membranes and
only pure water flows through because the water close to the pores in the membrane is
ordered and excludes salts. This is a very general effect. If it were an adaptative mechanism
to low temperature in bacteria, there would be no need for an increase in osmolyte
concentration at low temperatures. In fact, such an increase might be deleterious and a
decrease might be required instead. This allows us to propose a model for freeze resistance
which is consistent with our data. We propose that a good strategy would be for the cell to
increase the relative amount of (membrane) bound water by reducing the cell size and change
the membrane surface chemistry so that ions are not required on its surface. The water thus
arranged will not freeze and the small amount of bound water in the center of the cell will

contain all of the electrolytes and will not freeze because of colligative effects (Figure 4.14).

148

 

- Cell membrane

m Highly ordered hydntion sheath

 

Lees ordered hydration sheath

E cytoplasmic solutes

Figure 4.14: Model showing the solvation of the region
between two bilayer membranes. Note the ordered water
(checkered pattern) structure close to the membrane surfaces. In
this region ice formation cannot take place. Note that the solutes
(solid particles) are restricted to the disordered region where their
effective concentrations are now so high that ice formation does
not take place in this region either. The width of the ordered
solvation layers should increase as temperature decreases.

149

ACKNOWLEDGMENTS

We thank Dr. Vera Soina for isolating the Siberian permafrost strain 45-3, and Dr.
Yin Tang (from our laboratory) for doing the transmission electron microscopy.

This work was supported in part by the Center for Microbial Ecology, a National
Science Foundation Science and Technology Center (NSF grant# BIR 912-0006) and by

the National Science Foundation (Grant IBN 9507189 to RIH).

REFERENCES

l. Franks, F., S. F. Mathias, and R. H. M. Hatley. (1990) Water, temperature and life.
Phil. Trans. R. Soc. Lond. B, 326: 517-533.

2. Gould, G. W., and J. C. Measures. (1977) Water relations in single cells. Phil. Trans.
R. Soc. Lond. B, 278: 151-164.

3. K0, R., L. T. Smith, and G. M. Smith. (1994) Glycine betaine confers enhanced
osmotolerance and cryotolerance on Listeria monocytogenes. J. Bacteriol.,l76: 426-431.

4. Khlebnikova, G. M., D. A. Gilichinsky, D. G. Fedorov-Davidov, and E. A.
Vorobyeva. (1990) Quantitative evaluation of microorganisms in permafrost deposits and
buried soils. Microbiologiya, 59(1): 148-155 [in Russian].

5. Hollingsworth, R. I., M. Abe, J. E. Sherwood, and F. B. Dazzo. (1984)
Bacteriophage-induced acidic heteropolysaccharide lyases that convert the acidic

heteropolysaccharides of Rhizobium trifolii into oligosaccharide units. J. Bacteriol., 160:
5 10-5 16.

6. Laemmli, U. K. ( 1970) Cleavage of structural proteins during the assembly of the
head of bacteriophage T4. Nature, 227: 680-685.

7. Schultz, R. D., and S. K. Asunmaa. (1970) Ordered water and the ultrastructure of the
cellular plasma membrane. In Recent Progress in Surface Science Vol. 3. J. F. Danielli,
A. C. Riddiford, and M. D. Rosenberg, editors. Academic Press, New York and London.
pp. 291-33 1.

CHAPTER V

CHARACTERIZATION OF A MECHANISM OF ADAPTATION OF A pH AND
SOLVENT RESISTANT BACTERIUM Sarcina ventriculi

150

151

ABSTRACT

The formation of carbon-carbon bonds remains one of the most challenging and
important endeavors of synthetic organic chemistry. There are no chemical or biological
systems in which this has been realized at centers that are not activated by some functionality
such as a carbonyl group. Here we report on such a reaction that occurs at isolated,
completely unactivated carbon centers. Free foreign fatty acids were joined tail-to-tail at their
w-l positions to fatty acids in the membrane lipids of Sarcina ventriculi simply by adding
a membrane preparation to the fatty acid suspension in the presence of a low molecular
weight ketone. In similar experiments, n-heptane and hexadecylpyridinium chloride were
linked at the 2-position to the w-l position of S. ventriculi membrane lipids by simply adding
the hydrocarbons to growing cultures of the organism. The coupling mechanism is
temperature dependent, is triggered both by high and low pH values and is inhibited by thiols
but not by 1,4 cyclohexadiene although, interestingly, it is suppressed by the addition of
cyclohexane. The process is suppressed if cells are cultured in deuterium oxide indicating
that the trigger is a kinetic factor related to motional dynamics. These results and results from
NMR spectroscopy indicate that the process is regulated by membrane dynamics and occurs
via a radical mechanism in which a hydrogen atom is extracted from each alkyl chain and the
two alkyl radicals join to form a new carbon carbon bond. Experiments using the radical trap
2-nitroso-2-methyl propane (to which the process is insensitive) and the fact that coupling
is stereospecific indicated that the radicals are caged species that never escape. Radicals that

escape would be susceptible to combination or trapping by external agents and would

152

undergo rapid planarization and loss of chirality. The stereospecific formation of carbon-
carbon bonds between unactivated carbon centers by this system has very high potential for

use in synthetic organic chemistry.

INTRODUCTION

Because of its very unfavorable energetics, the formation of carbon-carbon bonds
between unactivated carbon centers by the removal of hydrogen atoms from the two bonding
sites and direct connection of the two centers is a process for which there are no known
biological nor many laboratory precedents. In the case of alkyl chains, carrying this out with
any regio- or stereospecificity is a daunting task because energetics would be expected to be
the same for any methylene group along the chain. The stability of the radical formed by
removal of a hydrogen atom from the unique terminal methyl groups will be less than that
of one formed from a methylene group and so the specific coupling of the chain termini
through the methyl groups is actually less likely. Despite the high degree of difficulty of this
coupling process it might, apparently, be a very common event in the membranes of some
bacteria. One such organism is the anaerobic gram positive bacterium Sarcina ventriculi. In
a series of earlier studies [1-3] we demonstrated that this organism adapted to changes in
environmental conditions such as temperature and pH by forming new lipids species
containing onto-very long chain dicarboxylic acids of up to 36 carbons in length. Over 80%
of the total membrane acyl chains can be these very long bifunctional molecules. Both the

relative proportions of these unusual fatty acid species and their structures suggested they

153

might be formed by tail to tail coupling of existing regular chain fatty acids probably via the
radical species formed by simple removal of a hydrogen atom from the 00-1 positions of two
opposing acyl chains (Figure 5.1). A vicinal dimethyl function is therefore introduced at the
site of linkage. Even if the two fatty acid species are identical, the isolated fatty acids (or
dimethyl ester) must have the R, R or S, S configuration at the two new stereocenters and

will be optically active. This has been observed [2].

There are several necessary elements which have to be understood if this reaction
could ever be exploited. Firstly, coupling two alkyl chains in this fashion with the liberation
of two molecules of hydrogen is an energetically uphill task. Fortunately, even bacteria have
to conform to the laws of thermodynamics so there is a chance that this activity could be
understood and harnessed. To drive such a process, the chain joining would have to be
coupled to a reduction event and so this process would require an oxido-reducing system.
One would not necessarily need to know what the intermediate redox system is, but it is
important to know the nature of the final molecular species accepting the two hydrogen
atoms. By adding this terminal hydrogen acceptor, the intermediate species that accepts the
two hydrogens from the coupling can be regenerated thus allowing the process to be a

catalytic one.

Another element of the mechanism which has to be defined is the trigger or
controlling factor regulating the activity. It seems clear that the coupling process is activated

by any stimulus or condition causing an increase in fluidity or motion in the membrane lipid

154

 

LIPID
MEDIA

MEMBRANE

 

 

Figure 5.1: Conceptual model for the coupling of alkyl chains of the lipid components
by an integral membrane catalytic system. Note that if the two hydrogen atoms are
removed from the same side, the final product will have the methyl groups on the same side.

155

ensemble. Chemical studies show that the amount of chain coupling in this bacterial species
grown at different temperatures increases with the growth temperature from essentially zero
at low temperature (~ 30°C) to greater than 80% at temperatures of 45°C. Studies using
dynamic NMR spectroscopy [4] show quite clearly that even when this high degree of
coupling occurs in response to depression of pH, the fluidity and dynamics of the alkyl
chains do not change when measured at the temperature and pH at which the bacteria are
grown. The coupling activity is therefore activated and also back-regulated by molecular
dynamics always conserving the motional freedom of the membrane lipids within a tight
range. This dynamic regulation is an important feature of the membranes of living systems

and has been termed "homeoviscous adaptation" [5].

Since this coupling reaction seems to be a general response to a perturbation of
membrane dynamics, a broad spectrum of perturbations should (and do) trigger it. One
parameter that should affect the expression of this catalytic activity is pH. This catalytic
activity should be buried in the inner core of the membrane but it should, however, be
sensitive to pH since the state of ionization of the membrane lipid head groups also helps
determine the closeness of their packing [6]. The headgroup packing will, in turn, influence
the order and dynamic state of the lipid alkyl chains. It should therefore be possible to use

pH (at either extreme) to activate and control the reaction.

The possibility of forming these transmembrane fatty acids by the joining of existing

chains suggests some very interesting and exciting possibilities. The most obvious one is

156

that if the enzyme system could be isolated, or harnessed directly using membrane
preparations or whole cells, it could be utilized in synthetic organic chemistry for the
stereospecific formation of carbon-carbon bonds between unactivated centers under very
mild conditions. If the coupling process is independent of the nature of the group at the
distal ends of the hydrocarbon chains, then an entire family of cam-bifunctional molecules
can be made. This would have important implications for the polymer industry. The
coupling activity could be used to stabilize liposomes by joining together alkyl chains at the
interface between the layers of lipids or it could be used for anchoring monolayers to
hydrocarbon substrates such as those formed by treating noble metal layers with alkane
thiols. Here we describe how this coupling activity can be controlled and regulated and

demonstrate that the potential uses cited above are well within the realms of possibility.

MATERIALS AND METHODS

General methods

Cells were grown under strict anaerobic conditions in liquid media as described
earlier [1]. All cell transfers were carried out in a glove box under an oxygen free
atmosphere provided by argon which was passed over heated copper to remove traces of
oxygen. Cells were harvested by centrifugation at 8000 rpm in a Sorvall centrifuge equipped
with a GSA rotor. For experiments that required pH control of growth conditions, bacteria
were grown in a chemostat equipped with a pH probe. The chemostat was also fitted with

an inlet for adding sodium hydroxide or hydrochloric acid solution via a peristaltic pump

157

under pH feedback. Cells were cultured at 37°C at pH 3, 7 or 9.7. The pH conditions were
maintained by adding 5 M sodium hydroxide for growth at pH 7 or 9.7 or hydrochloric acid
for growth at pH 3. For growth of cells in 99% deuterium oxide, cells were preconditioned
by culturing them in increasing concentrations of deuterium oxide until they were adapted
to growth at 99%. The growth temperature was 37°C. In these experiments, the pD or pH
of the growth medium was allowed to drop to ~4, conditions that normally trigger the
formation of very long or,(.)- bifunctional fatty acids. It should be noted that the measured

pD is typically about 0.4 units lower than the corresponding pH value [7].

Cell-free coupling involving free, exogenous fatty acids

For the cell-free coupling of free fatty acids to membrane lipids, cells (200 ml) were
cultured anaerobically in the absence of exogenous fatty acid at 37°C and a pH of 7. They
were then harvested in an inert atmosphere (argon) by centrifugation and resuspended in 2
ml of oxygen free growth medium. They were lysed in a French pressure cell under
anaerobic conditions (by setting up the pressure cell in a glove box, and collecting the lysate
in an anaerobic vial) and hexadecanoic acid (10 mg) and ethyl ketone (10 ml) were then
added. The mixture, under anaerobic conditions, was incubated at 45°C for 3 hours and then
the fatty acids were released from the lipids by methanolysis with 5% HCl in methanol at
75 °C for 24 hours. The mixture was concentrated to dryness under a stream of nitrogen and
the fatty acid methyl esters were extracted by partitioning between chloroform and water.
The chloroform extract was subjected to GC-MS analysis using a J eol JMS-AX505H

spectrometer interfaced with a Hewlett-Packard 5890A gas chromatograph.

158

Coupling of heptane chains to membrane lipid fatty acid termini

N-heptane (0.29 ml) was added to a growing 10 ml culture of Sarcina ventriculi cells
incubated at 37°C in roller bottles to effect mixing. After 2 hours, the cells were harvested
by centrifugation and the methyl ester derivatives of the membrane lipid fatty acids were

prepared and analyzed by GC-MS.

Coupling of hexadecylpyridinium chloride to membrane lipid fatty acid termini
Hexadecylpyridinium chloride (50 mg in 800 pl sterile water) was added to each of
four exponentially growing cultures (500 ml) of Sarcina ventriculi at 37°C. After 24 hours,
the cells were harvested by centrifugation. Membrane fatty acid methyl esters (FAMEs)
were made and isolated by refluxing for 2 days in 4:1 5% HCl in MeOH: 1,2-dichloroethane
at 74°C, then extracting twice by phase separtation in 50 ml H202100 ml CHCl32100 ml
hexane. The organic layers were removed and evaporated to dryness. 1H NMR was done
on the total FAMEs. A family of methyl esters that were visible by UV detection was
isolated by preparative thin layer chromatography (TLC) and characterized by various

spectroscopic methods.

Effects of chemical agents on transmembrane fatty acid synthesis

Cells (10 ml) were cultured at 37°C as described above except that tert-butyl thiol
(0.23 ml), n-butane thiol (0.21 ml), cyclohexane (0.22 ml), 1,4-cyclohexadiene (0.19 ml) or
2-nitroso-2-methyl propane (34.8 mg) was added. In the case of 2-nitroso-2-methyl propane,

the solid was predissolved in 1.5 ml of ethanol to a concentration of 23.2 mg/ml. The

159

reagents were added to the cultures in early log phase to the final concentrations indicated,
then the temperature was shifted to 50°C, and the cells were harvested in early stationary

phase (about 2 hours later). The pH was not regulated.

NMR spectroscopy analysis of the effect of cyclohexane and 1,4-cyclohexadiene on
membrane lipid alkyl chain dynamics

Five, 20 mg samples of dipalmitoyl phosphatidylcholine (DPPC) were lyophilized
for 12 hours. Each sample was resuspended in 0.6 ml of a 10 mM phosphate buffer (pH 7.0)
containing the following components at the final concentrations indicated: NaCl, 0.1 mM;
KCl, 2.5 mM; MgC12, 0.5 mM; and CaClz, 1.0 mM. The DPPC samples were then frozen
in dry ice/acetone bath, heated to 55°C for 10 min., and vortexed. These three annealing
steps were repeated five times. Cyclohexane (5 ,al) was added to the first sample and 10 al
to the second. The same amounts of 1,4-cyclohexadiene were added to the third and fourth.
The fifth was kept as a control. The samples were vortexed and from each 50 a] was
withdrawn and added to 0.6 ml of 5:1 DZOzD4-MeOH, mixed and transferred to NMR tubes.
1H NMR measurements were performed on a Varian VXR-BOOS spectrometer operating at

300 MHz. The temperature was held at 60°C, above the phase transition for DPPC.

RESULTS AND DISCUSSION

Since the very long bifunctional fatty acid species in Sarcina ventriculi are formed

by random, indiscriminate (but enzymatic) tail-to-tail joining of existing fatty acids, it should

160

be possible to demonstrate that foreign, exogenously-added fatty acids could be taken up
intact into membrane vesicles of Sarcina and could be incorporated (intact) into the
transmembrane lipid species. Addition of heptadecanoic acid (or any fatty acid which is not
found in membranes of Sarcina) to a growing culture of that organism at 37°C, followed by
cell lysis and a temperature shift to 45°C resulted in the formation of chimeric fatty acid
species with structures one half of which was formed from the foreign fatty acid and the
other from a native species (Figure 5.2). In this experiment, methyl ethylketone was added
as a final hydrogen acceptor (with the formation of 2-butanol, catalysed by native oxidases)
to ensure that all of the potential cofactors were always oxidized and able to accept the two
hydrogen atoms generated in the coupling process. The reaction did not take place to any
measurable extent if the ketone was left out (data not shown) indicating that the hydrogens
were funneled (presumably via a species such as ferredoxin) to a flavin cofactor and finally
to the ketone. The addition of n-heptane (which was added to a growing culture in mid-log
phase) to the (o-l position of the membrane fatty acyl chains was also realized. This was
evident from the GC-MS analyses in which a peak for a component with a mass spectrum
(Figure 5.3) corresponding to a 23 carbon fatty acid was observed, indicating the n-heptane
was added to a 16 carbon membrane lipid. The predominant alkyl chains observed in
Sarcina are 16:0, 18:0, and 18:1 at 37°C. However, carbon chains longer than 23 (as in
addition to the 18 carbon long chains) were not observed in the n—heptane experiment,
possibly because they adversely affect the membrane motional dynamics. Longer carbon
chains would probably extend too far into the opposing membrane leaflet or would have to

fold back on themselves, in either case increasing the fluidity of the membrane. To further

161

Figure 5.2: The electron impact mass spectrum of one
of the chimeric bifunctional fatty acid dimethyl esters
formed by tail-to-tail coupling between an exogenously
added fatty acid (heptadecanoic acid) and a
hexadecanoic acid residue from an S. ventriculi lipid
component. The fragmentation pattern is indicated above
the spectrum. Detailed discussions of the fragmentation
modes for these molecules have been presented earlier [1,

2].

crr,o

QOSIQSCFD fl<~op—ap

r—--?------------

  

162

 

.31 42
269 —v- 31-—+195
i'"'"""""", """"
: -293.-3—i-.>.3§5.-.‘+> 223
\ 4 o
. x - j ocn,
........... I ---.§3¢h .( s‘ ‘7‘ 42 H
221<—°14— 24! ’ : l .n I 42 :
I - rue—ma— s 2
231 «AZ—2194i 311' : """"°'""""~j-MJ
ooooooooooooooooooooo J I m .
I
I
l

87. 101. 115. 129. 143. ..

 

100

60‘

4a.

 

 

 

 

 

 

 

I
t ' 1 23
11 Z : s 4
l i" I. ‘ . .' I . .
200 300 400 500 600

Figure 5.2

 

163

 

 

 

 

 

 

 

19a 5
R
8 . O M
1 .
: BB. MeO n "3 1
i or
V I
e 66‘
9 j 74
'3 . are
n 43. 97
d 38
a b
n ‘ 2 7 ’
: 281 97 143
j 1 1 1 3171 1-4 2 9 27 ’
J 3 8 7 4 7 .
so me 150 290 259 see 350 493 450

M/Z

Figure 5.3: The electron impact mass spectrum of a 23 carbon fatty acid methyl ester
resulting from coupling of n-heptane to a hexadecanoic acid residue from a S. ventriculi
lipid component at the 00-1 position. The molecular weight is 368. The inset shows some
of the major fragmentation patterns.

164

show that tail-to-tail coupling of exogenous lipids occurs without regard to its headgroup,
the addition of hexadecylpyridinium chloride was also demonstrated. The 1H NMR spectrum
of the total fatty acid methyl esters (Figure 5.4A), which included free hexadecylpyridinium
chloride, regular chain fatty acids and transmembrane fatty acids, indicated that some of the
hexadecylpyridinium compound had been converted to a reduced pyridine system since there
were new signals in the region of 6.9 - 7.2 ppm. The total fatty acid methyl esters were
separated by thin layer chromatography. A family of transmembrane fatty acids (visible by
UV detection), now separated from the free hexadecylpyridinium chloride and regular chain
fatty acid methyl esters, gave a mixture in which the 1H NMR spectrum indicated that the
hexadecylpyridinium was coupled to native membrane lipids (indicated by the complex
splitting of the methyl signals at about 0.9 ppm), but the aromatic ring had been transformed
to a 1-alkyl-4-pyridone (Figure 5.4B). This was indicated by the disappearance of signals
for the pyridine nucleus and the appearance of a narrow doublet at 7.72 ppm (2-position
protons) and a multiplet at 7.54 ppm (3-position protons). The signals between 6.9 and 7.2
ppm were hardly visible in the spectra of the total extract prior to TLC indicating that they
probably arose by oxidation of a dihydropyridine during handling. This is not surprising
since the chimeric fatty acids were formed under reducing and anaerobic conditions. The
infrared spectrum contained strong resonances for the C-H stretch at 2956 cm", 2928 cm",
and 2859 cm". The peak for the ester carbonyl stretch was observed at 1725 cm“1 (m), while
the C=C stretch was observed at 1600 cm'I (w). Electrospray mass spectrometry also
supported the presence of chimeric fatty acid methyl esters containing the 4-pyridone

nucleus. To test the theory that the bacteria reduced the pyridine ring, with subsequent

165

Figure 5.4: 1H NMR spectra of (A) total fatty acid methyl
esters from S. ventriculi cultured with hexadecylpyridinium
chloride, and (B) isolated fatty acid methyl esters
containing hexadecylpyridinim coupled to native S.
ventriculi membrane lipids. The signals at 8.1 ppm, 8.45
ppm, and 9.25 ppm in (A) are due to the pyridine nucleus of
free hexadecylpyridinium chloride. Note the new signals in (A)
in the region of 6.9-7.2 ppm which indicated that some of the
hexadecylpyridinium had been converted to a reduced pyridine
system. In (B) note the essential disappearance of signals for
the reduced pyridine nucleus between 6.9 and 7.2 ppm, and the
appearance of the narrow doublet at 7.72 ppm (2-position
protons) and a multiplet at 7.54 ppm (3-position protons)
indicate that during handling the aromatic ring was oxidized to
a 1-alkyl-4-pyridone.

166

 

 

 

 

 

 

 

 

 

 

.rvrr vaIVV’yIv vvlvrwytyyvutrvvvy 7‘vlvavTvvyvlvvvarvryTvvrvl 7‘—

u 10 9 O 1 6 s C 3 2 1 0 fi-

 

 

 

 

Figure 5.4

167

oxidation to form the pyridone (probably during handling) (Figure 5.5), we reduced
hexadecylpyridinium chloride using sodium borohydride to form a dihydropyridine, and then
attempted to oxidize it to a pyridone using ceric ammonium nitrate (Figure 5.5). The
reduction step was very successful but oxidation yielded a variety of products and 1H NMR
spectroscopy indicated that only trace amounts of product corresponded to the desired

material.

In a prior study, it was demonstrated that Sarcina ventriculi carries out tail-to-tail
coupling of its membrane lipids in response to agents lowering the pH [1]. It might be
argued that the coupling response might have some ionic component which is acid catalyzed.
An explanation that appears to be more consistent with the profile of this response, however,
is simply that lowering of pH should suppress the ionization of the phosphate groups in the
membrane lipids thus reducing their electrostatic cross-linking with calcium ions thus
decreasing membrane stability. Such an argument would require that we also see coupling
at high pH, since then the phosphate groups would be completely ionized. The resulting
electrostatic repulsion should also destabilize the membrane thus triggering the process.
Observing coupling at high pH would also rule out any acid catalysis. The gas
chromatography profile of fatty acid methyl esters from Sarcina ventriculi cells cultured at
pH 9.7 is shown in Figure 5.6. A family of peaks with very long retention times is seen in
the pH 9.7 chromatograms. The inset shows a mass spectrum of peak 3, which corresponds
to a C34,,oz,w-dicarboxylic acid dimethyl ester formed at high pH. This was the case for all

of the other late eluting peaks in the profile. They were all readily identifiable as very long

168

fly an o

{J NAD(P)H [O]
—" —* l l
\’V
N N N
BH4'
H H O
Ceric ammo-
nium nitrate several Other
I I —> + products
N N
( Q15 (<) 15

Figure 5.5: Bacterial (proposed) and synthetic mechanisms of the reduction and
subsequent oxidation of the pyridine ring of hexadecylpyridinium chloride.

169

£2.82 0882808 86250

no 88: 920 m .08 3:0 a w8880 .3 808.5. 8 £023
.m 80m .8 88.00% 888 .0mm8_ 80800—0 8. £508 .02:
0:... 88.00 35086 025.88.08.38 3.0 A3 .8050
35086 208 0.3888063... .30 Amv .8080 3508:.
E08 0.388.806-38 SJ .8 .8080 3508388.

-3 3.0 A: .8050 3508 20.. 02.08880 3:0

5. 80501508 20... 02.98280 3:0 ADV 80.0.0 3508
28 28.88 3.0 a. .882. be. So 2. a.
8:08.88 8.88 05. .5. In 8 08:50 8 $30.58..
.m. .8 $8.58 Ba: 88. 89a 3882...... 8080
.288 E8 .38.. .8 888388980 80 5% 08m?—

170

8 2:5

 

.Al as: :2820:

 

we
.4!
N...‘
"1
"“‘1
J

 

 

 

 

 

 

 

 

 

 

U
N\:
00... 0mm corn r - aw? Gun SN 00.—
. 8.5 “.13.: j.
a n
v N u
v , 6.0.
m 0 u
m N n N h N . C
v m 0
.av n
0
A. ._
. c
. .uw u
8 . >
. no .
:8 L0 f a
C
{007%}; km.- ..8 _
u
«.8 mm 0 m.
Nflm .Oou

 

 

10109100

 

esuodso;

 

171

chain bifunctional species observed when these bacteria are cultured at low pH, high
temperature or in the presence of organic solvents again demonstrating the universality of

the response.

To further test the hypothesis that tail—to-tail coupling is triggered by changes in the
physical state of the membrane, Sarcina ventriculi was cultured in D20. The stronger
hydrogen bonds in D20 should slow down the lipid molecular motions, which would be
equivalent to a reduction in temperature. The proportion of transmembrane lipids should
differ between cells cultured in H20 versus those cultured in D20. In these experiments, the
pH was allowed to drop freely as a result of acetic acid production by the bacteria [8], usually
reaching a minimal pH of 4.0. A reduction in pH is known to promote the formation of long
chain lipids leading to a high proportion of transmembrane lipids. If the cells are cultured
in D20, however, the reduced motional dynamics in the hydrogen bonding network in the
lipid headgroup and in the various solvent cages should stabilize the membrane thus off-
setting the fluidizing effects of the low pH. This is exactly what was observed (Figure 5.7).
A high proportion of transmembrane fatty acids is seen in the cells grown in H20. In
contrast, a very small proportion is observed in the cells cultured in D20. These results are
consistent with the idea that substituting D20 for H20 is equivalent to a reduction in
temperature causing an increase in membrane viscosity. The cells respond by maintaining
a low level of long chain lipids. A similar kinetic effect of D20 on lipid composition has
recently been reported [9]. In this study Bacillus subtilis was grown in D20 and the

proportion of lipids with alkyl chain length of 16 carbons increased relative to those of 18

172

Figure 5.7: Gas chromatographs of fatty
acid methyl esters from S. ventriculi cells
cultured in (A) H20 and (B) D20. Note the
high proportions of late eluting peaks in A
compared to B.

173

 

 

 

 

Retention time

 

 

 

 

 

 

Retention time

Figure 5.7

 

 

174

or 20 carbons.

The proposed mechanism for tail-to-tail coupling of membrane lipids is that
hydrogens are removed from the w-l positions of two opposing alkyl chains forming radicals
which quickly combine (Figure 5.1). Since these bifunctional fatty acid species are chiral,
they should not be readily accessible to radical traps or scavengers. Radicals planarize
rapidly, much faster than the timescale for diffusion from the cage in which they are formed
[10-12]. This notwithstanding, it was of interest to assess the effects of various radical traps
and scavengers on this system. Such experiments were important from another perspective.
It was clear from previous experiments that oxygen had an adverse effect on the coupling
activity. This would indicate either a radical process or the presence of an oxygen labile
species such an iron-sulfur complex. Reagents which would allow us to test these
possibilities were added to Sarcina cultures growing at 37°C, then the temperature was
raised to 50°C to trigger the coupling reaction. The use of a radical trap is a very direct way
of testing whether any accessible radical species are generated in the coupling reaction. This
was done using 2-nitroso-2-methylpropane [13]. In these experiments, transmembrane cross-
linking was not inhibited but, in fact, increased to levels (19%) much higher than observed
in the 50°C control (Figure 5.8). Thiols too react very rapidly with radicals and would also
be good probes to see if, during the coupling process, any radicals which might be generated
are accessible. Two thiols, t-butyl thiol and n-butane thiol, were used and contrary to the
nitroso methylpropane results, both inhibited coupling (Figure 5 .8). Iron-sulfur species are

very sensitive to thiols [14-17] and this inhibition could be indicative of the presence of such

175

30-
l

% Coupling Relative to 50 c Control

 

 

-40 _-

2-NMP

t butyl tthI
n-butane thiol

1 ,4-cyclohexadiene
cyclohexane

Figure 5.8: Results from radical trap/scavenger experiments indicating the percent of
tail-to-tail coupling of S. ventriculi membrane lipids normalized to the 50°C controls.
2-Nitroso—2-methylpropane and 1,4-cyclohexadiene enhanced coupling, while the two thiols
and cyclohexane inhibited the process.

176

a center and not the presence of an accessible radical center. To further clarify this point,
1,4-cyclohexadiene was used as a radical scavenger. This is a radical-sensitive diene that
should have no effect on iron—sulfur centers. A similar result to the one seen with 2-nitroso-
2-methylpropane was obtained. The levels of cross linking actually rose (Figure 5.8). It
appeared that 2-nitroso-2-methylpropane and 1,4-cyclohexadiene were not reacting with
radicals but were further fluidizing membrane. 1,4-Cyclohexadiene is flattened and will not
stack well between the membrane lipid chains. Cyclohexane, on the other hand, has a chair
conformation and will stack well between the lipid chains. Since it would reduce the alkyl
chain mobility, it could be expected to reduce the level of coupling considerably. This is
exactly what was observed (Figure 5.8). The reduction of alkyl chain mobility by

cyclohexane has also been characterized by fluorescence spectroscopy [18].

The relative effects of 1,4-cyclohexadiene and cyclohexane on hydrocarbon chain
dynamics was investigated using NMR spectroscopy. The test system used was dipalmitoyl
phosphatidylcholine. This is a comparatively rigid molecular system and the signals for the
methyl and methylene groups at ~ 0.7 and 1.2 ppm are broad with very low intensities
because of very short spin spin and spin lattice relaxation times. In the NMR (Figure 5.9)
cyclohexane (spectra B and C) does not increase the fluidity of the lipids (rigidifies) as
compared to the control (spectrum A). The methylene and terminal methyl peaks at 1.19
ppm and 0.72 ppm are still barely visible in all three spectra indicating no increase in T 1 or
T2. 1,4-Cyclohexadiene, however, shows a marked affect on the lipids. Both the methylene

and terminal methyl groups show increased mobility as seen by the increase in intensity of

177

Figure 5.9: 1H NMR spectra showing the effect of
cyclohexane and 1,4-cyclohexadiene on the lipid packing
arrangement of dipalmitoyl phosphatidylcholine. (A)
dipalmitoyl phosphatidylcholine control, (B) 5g]
cyclohexane, (C) lOul cyclohexane, (D) 5 ul 1,4-
cyclohexadiene, an (E) 10 ul 1,4-cyclohexadiene. Signals
for the methylene protons (1.19 ppm) and methyl groups
(0.72 ppm) are indicated.

178

 

 

 

 

 

 

 

L L (E L L

 

 

Figure 5.9

179

the peaks at 1.19 ppm for the methylene (seen in both spectra D and E) and 0.72 ppm for the

methyl groups (which is more prominent in spectrum B), respectively.

All of the results obtained here, therefore, indicate the putative hydrocarbon radicals
are caged and therefore inaccessible to any external agents. It is clear that thiols are
inhibiting cross-linking via another mechanism and not by radical scavenging. One
possibility is that they are extruding iron-sulfur centers from the necessary cofactors such as
ferredoxin [17], thus breaking the electron transfer chain and stopping the recycling of the
hydrogen acceptor. The involvement of iron-sulfur centers is also consistent with the oxygen

lability of the coupling process.

The hydrocarbon chain coupling phenomenon demonstrated has no known precedent.
The coupling conditions are mild and the regio- and stereospecificity is extremely high. It
is independent of the nature of the functional groups on the distal ends of the carbon chains.
Other systems containing this activity will, for sure, be identified and knowing how to
optimize and control it will have important implications for synthetic organic chemistry. The
stereoselective formation of carbon-carbon bonds between unactivated centers is viewed as

the last frontier in carbon-carbon bond formation.

ACKNOWLEDGMENTS

I would like to acknowledge Seunho Jung, a past member of our laboratory, who did

180

the cell-free coupling of the heptadecanoic acid to S. ventriculi membrane fatty acids. I
would also like to acknowledge Luc Bérubé’s, another past member of our laboratory,
contributions to this work as well. He did the alkaline pH and D20 experiments. He also did
the experiments in which S. ventriculi was cultured in the presence of cyclohexane, 1,4-

cyclohexadiene, and tert-butyl thiol.

REFERENCES

1. Jung, S., S. E. Lowe, R. I. Hollingsworth, and J. G. Zeikus. (1993) Sarcina ventriculi
synthesizes very long chain dicarboxylic acids in response to different forms of
environmental stress. J. Biol. Chem., 268: 2828-2834.

2. Jung, 8., and R. I. Hollingsworth. (1994) Structures and stereochemistry of the very long
cam-bifunctional alkyl species in the membrane of Sarcina ventriculi indicate that they are
formed by tail-to-tail coupling of normal fatty acids. J. Lipid Res, 35: 1932-1945.

3. Jung, 8., and R. I. Hollingsworth. ( 1995) A mathematical model explaining the molecular
weights and distribution of very long chain dicarboxylic acids formed during the adaptive
response of Sarcina ventriculi. J. Theor. Biol, 172: 121-126.

4. Bérubé, L. R., and R. I. Hollingsworth. (1995) Membrane lipid alkyl chain motional
dynamics is conserved in Sarcina ventriculi despite pH-induced adaptive structural
modifications including alkyl chain tail to tail coupling. Biochemistry, 34: 12005-12011.

5. Sinensky, M. (1974) Homeoviscous adaptation--a homeostatic process that regulates the
viscosity of membrane lipids in Escherichia coli. Proc. Nat. Acad. Sci. USA, 71: 522-525.

6. Eibl, H. (1983) The effect of the proton and of monovalent cations on membrane fluidity.
In Membrane F luidity in Biology, ed. R. C. Aloia, 2: 217-236.

7. Glasoe, P. K., and F. A. Long. (1960) Use of glass electrodes to measure acidities in
deuterium oxide. J. Phys. Chem, 64: 188-190.

8. Canale-Parola, E. (1970) Biology of the sugar-fermenting Sarcinae. Bacteriol. Rev., 34:
82-97.

181

9. Hollingsworth, R. I. (1995) Membrane lipid alkyl chain heterogeneity is a molecular
transform of the available thermal energy spectrum. J. Phys. Chem, 99: 3406-3410.

10. Doering, W. v. E., M. Farber, M. Sprecher, K. B. Wiberg. (1952) On the
decarbonmonoxylation reaction. J. Amer. Chem. Soc., 74: 3000-3001.

11. Greene, F. D., and N. N. Lowry. (1967) 9-Decalyl free radicals. Decomposition of cis-
and trans-9-decalylcarbinyl hypochlorite. J. Org. Chem, 32: 875-882.

12. Walborsky, H. M., and P. C. Collins. (1976) Cyclopropanes. XXXVIII. Effect of 1-
substituents on the stereochemical stability of the cyclopropyl radical. J. Org. Chem, 41:
940-946.

13. Thie’ry, C. J., JP. L. Agnel, C. M. Fre’javille, and J. J. Raffi. ( 1983) Electron spin
resonance spin-trapping analysis of y-induced radicals in polycrystalline oc-D-glucose. J.
Phys. Chem, 87: 4485—4488.

14. Bobrik, M. A., L. Que, Jr., and R. H. Holm. (1974) Synthetic analogs of the active sites
of iron-sulfur proteins. IV. Ligand substitution reactions of the tetranuclear clusters
[Fe4S4(SR)4]2' J. Amer. Chem. Soc., 96: 285-287.

15. Que, L., Jr., J. R. Anglin, M. A. Bobrik, A. Davison, and R. H. Holm. (1974a) Synthetic
analogs of the active sites of iron-sulfur proteins. D(. Formation and some electronic and
reactivity properties of Fe4S4 glycyl-L-cysteinylglycyl oligopeptide complexes obtained by
ligand substitution reactions. J. Amer. Chem. Soc., 96: 6042-6048.

16. Que, L., Jr., M. A. Bobrik, J. A. Ibers, and R. H. Holm. (1974b) Synthetic analogs of
the active sites of iron-sulfur proteins. VII. Ligand substitution reactions of the tetranuclear
clusters [Fe4S4(SR)4]2' and the structure of [(CH3)4N}2 [Fq S4 (SQ l-g )]. J. Amer. Chem. Soc.,
96: 4168-4178.

17. Que, L., Jr., R. H. Holm, and L. E. Mortenson. (1975) Extrusion of Fe2S2* and Fe4S4*
Cores from the active sites of ferredoxin proteins. J. Amer. Chem. Soc., 97: 463-464.

18. Sikkema, J., J. A. M. de Bont, and B. Poolman. (1994) Interactions of cyclic
hydrocarbons with biological membranes. J. Biol. Chem, 269: 8022-8028.

CHAPTER VI

SUMNIARY AND PERSPECTIVES

182

183

The ability of procaryotes to survive in, or adapt to, many extreme environments is
impressive. The fact that they can not only withstand such extremes of temperatures, pH,
pressure, salinity, and high concentrations of detrimental chemical agents, but thrive under
these conditions is extraordinary. One of the most commonly observed cellular change,
however, in response to all of the different environmental perturbations is alteration of
membrane acyl (or alkyl) chain structures. As the cell envelope membrane(s) is in constant
contact with the environment, it makes sense that rapid alterations in membrane structure are

necessary as conditions change.

We have shown in this work that the extent of membrane acyl chain length
heterogeneity observed response to environmental perturbations is significant. The selection
of the acyl chain length distribution is a random process, in which the mean and variance of
the distribution curve depends on the available thermal energy spectrum. We know that
many plants and animals also make similar changes in their membrane acyl chains in
response to temperature as well, for instance fish adapt to cold temperatures by a similar
spectrum of membrane chemistry changes. This, then, seems to be a general response to
temperature changes. Thus aspects of the relationship between temperature and acyl chain
length distributions are likely to be universal. From an ecology and paleontology
perspective, the relationship between fatty acid chain length heterogeneity of a given
organism and temperature could be used basically as a temperature stamp. The fatty acid
compositions of bacteria isolated from different soil depths (or from other organisms like

fish) that have been preserved in some fashion (whether or not they are still viable), like the

184

Siberian permafrost bacteria, could give indications of global climate changes or catastrophic

events that have occurred in a particular region as the earth has evolved.

The fact that there is also a temperature equivalent for solvent and antibiotics effects
is another important finding. The possible relationship between temperature, solvent
concentration, and growth rate could be used to extend the industrial utility of
microorganisms. Using a three-dimensional graph like the one shown in chapter two, the
external operation temperature for microorganisms could be extended in both directions, by
adding the appropriate concentration of a fluidizing or rigidifying solvent. It may also be
possible to overcome drug resistance of some bacteria by determining the appropriate
combination of temperature and drugs to kill the bacteria. It has been shown, in fact, that
several bacterial species show temperature dependent sensitivity to certain antibiotics.
Several Pseudomonas strains were more resistant to gentamicin at 20°C and 30°C, but
became more susceptible at 37°C and 42°C [1]. In another study, all Yersinia kristensenii
strains were more resistant to B-lactam antibiotics at 22°C than at 37°C [2]. The reason for
the increased resistance to antibiotics at lower temperatures is probably due to the increased
rigidity of the membranes such that antibiotics are less effective in intercalating into the
membrane and so there is less uptake of the drug. Recent evidence for this was shown in
a study in which strains of Stenotrophomonas maltophilia that were more resistant to
aminoglycoside antibiotics at lower temperatures all had more rigid membranes than strains
that were susceptible at those temperatures as shown by ESR spectroscopy [3]. It may also

be feasible to kill cancer cells using the right concentrations of chemotherapy drugs and

185

localized heating of the tumor with lasers or probes.

Many researchers use the profile of fatty acid compositions of microorganisms as an
indication of the species, but these organisms are typically cultured under different conditions
in the laboratory than found in nature. However, we have shown that the balance between
metabolic pathways, for instance carbohydrate versus amino acid metabolism, could have
a significant impact on the membrane acyl chain structures. This balance is determined by
the available heat (kinetic energy) needed to carry out chemical processes. At lower
temperatures the amount of maintenance energy increases dramatically and cells need to use
carbon sources that require less energy to catabolize and yet yield more energy for
maintaining the necessary cellular functions. These shifts in metabolism directly affect the
types of fatty acid chains produced by changing the available acyl-CoA primer pool for fatty
acid synthesis. This finding also supports the randomness of membrane acyl structures, since
the cell has to use the primers available, a random selection process would favor whatever
is the most prominent primer in the pool. Therefore it would be more appropriate to
demonstrate that organisms produce the same spectrum of fatty acids under different
conditions. This is preferable to basing the relatedness of organisms on similar fatty acid
profiles when the organisms were most likely cultured under two very different conditions.
Also since it is expected that organisms that are adapted to subzero climates would most
likely use triacylglycerol stores for their primary energy source at low temperatures, as is
shown by large amounts of triacylglycerol found in Arctic strain 45-3 at 4°C, they should

have lipases which are very active at low temperatures. Isolation of these enzymes would

 

186

have many uses in industry. Work in our laboratory is currently being done to characterize

the activity of lipases in Arctic strain 45—3.

The findings that Arctic strain 45-3 does not produce compatible metabolites in
response to cold temperatures indicates that it is not clear that there is link between extreme
salinity and freezing. Also the current models on the ability of compatible solutes to confer
cryotolerance as well are certainly not general. As such more serious physico-chernical
models and theories for explaining the interactions between water and biological surfaces
need to be developed. The model we propose, in which cell surface alterations are sufficient
to provide cryoprotection, should also be expanded and subjected to a more physico-

chemical analysis in more systems.

The advantages of carbon-carbon bond formation at unactivated centers by S.
ventriculi has many potential synthetic organic chemistry and industrial applications. The
coupling of fatty acids in this manner could be used to make biodegradable plastics, liquid-
crystalline devices, and allow the grafting of materials onto polymers to help in the
biocompatibility of artificial implants. While the entire functioning enzymatic system will
probably not be able to be isolatable from the cells, since its regulation in the membrane is
likely very complex, characterizing aspects of the process would aid in understanding ways
of optimizing the system to effect the coupling of the desired lipid species. Initial work done

in isolating potential enzymes involved in this coupling process is presented in Appendix A.

187

A lot of work in studying adaptative responses of microorganisms to environmental
perturbations seems to result in "catalogs" of types of responses, the levels of the responses,
and in which organism they occur. Since these fine details of the adaptations seem to vary
so much in bacteria, little (if any) general conclusions are drawn from these observations.
In this work, we have hopefully provided new insights and theories as to the fundamental
principles involved in adapting to fluctuations in the environment. Bacteria, and even higher
organisms, have certain cellular functions which must be maintained regardless of the
environment. The manner in which organisms adapt to perturbations vary based on the types
of cellular processes which they are able to carry out. However, all of these chemical
processes rely on the available thermal energy (and thus kinetic energy) spectrum. There is
a circular relationship, though, between this available energy, the balance of metabolic
processes, and membrane structure. But, by "tuning" the membranes through continual
alterations in membrane structures to propagate the maximum of this available energy, an
optimal balance of metabolic pathways can be achieved to maintain necessary cellular
functions. The process by which organisms adapt to other environmental perturbations, other
than temperature, results from similar perturbations in the relationship of kinetic energy,
metabolism, and membrane structure. These changes are again detected by disturbances in
membrane phase equilibria, similar to temperature effects, which is why similar adaptative
responses are observed. So organisms typically have a general set of changes they can make
in response to many stress conditions, rather than a specific mechanism of adaptation for a
given type of environmental change. Thus the idea of specific receptors, channels, and

pumps for individual antibiotics or other chemical species is highly unlikely. Hopefully the

 

188

work presented here will elicit new views, ideas, and models as to the generality of
adaptation mechanisms to environmental perturbations and how they are tied to the types of
cellular processes than can be carried out by the cell. Also the application of the principles
we have developed for the adaptation and adaptability of bacteria to extreme environments,
along with the use of certain enzymatic functions of extreme adapted organisms has a great

deal of potential use in research and industry.

REFERENCES

1. Papapetropoulou, M., G. Rodopoulou, E. Giannoulaki, and P. Stergiopoulos. (1994)
Effect of temperature on antimicrobial susceptibilities of Pseudomonas species isolated from
drinking water. J. Chemother., 6: 404-407.

2. Bejar, V., C. Calvo, and A. Ramos-Cortnenzana. (1986) In vitro susceptibility of Yersinia
kristensenii strains to B-lactam antibiotics. Ann. Inst. Pasteur. Microbiol., 137A: 169-177.

3. Rahmati—Bahram, A., J. T. Magee, and S. K. Jackson. (1995) Growth temperature-
dependent variation of cell envelope lipids and antibiotic susceptibility in Stenotrophomonas

(Xanthomonas) maltophilia. J. Antimicrob. Chemother., 36: 317-326.

 

APPENDIX A

PREPARATION OF THE RADIOLABELED PHOTOAFFINITY PROBE FOR
IDENTIFYING PUTATIVE CROSS-LINKING ACTIVITY

189

190

We have begun work to try to isolate the enzymes involved in the tail-to-tail fatty acid
cross-linking mechanism in Sarcina ventriculi characterized in chapter 5. Below is described
the synthesis of a membrane l4C-labeled photoaffinity probe, which was then incorporated
into growing cultures of Sarcina ventriculi. Preliminary results of this experiment are then

shown.

Synthesis of a 1“C-labeled photoaffinity probe

A l4C-labeled photoaffinity probe was synthesized as outline in Figure A. l. (1) The
probe was made by combining 50 mg of the C,5-iodofatty acid with 40 pl (40 uCi) ”C-
phenol and 20 mg cold phenol (final concentration of 1 uCi/mg phenol), plus 50 mg
anhydrous potassium carbonate and 0.5 ml acetone. The reaction was stirred overnight. The
reaction was stopped by adding 1 ml of water and glacial acetic acid until effervesence
ceased. The product was extracted with chloroform and dried under a stream of nitrogen.
The product was dissolved in 2 ml acetic acid and the solution cooled in ice while stirring.
One milliliter of concentrated sulfuric acid was then added followed by the slow addition of
100 u] fuming nitric acid. The reaction became very dark orange/brown in color. After
about 10 minutes it was diluted with about 10 g of crushed ice, and then 15 ml chloroform
was added. The mixture was stirred, and the chloroform layer was then removed, washed
with cold water, and evaporated to dryness. The nitrated product was then hydrogenated to
form the amino derivative. This was done by taking up the product in 50 ml of absolute
ethanol and adding a few drops of water and acetic acid. Ten percent of palladium on carbon

(~20 mg) was added while passing nitrogen over the reaction. Hydrogen was supplied by

 

L

191

attaching two hydrogen filled balloons to the system. The reaction mixture was allowed to
stir for over 8 hours until the yellow color had disappeared. The palladium on carbon was
removed by vacuum filtration through methanol washed celite. The product mixture was
evaporated to dryness. The product was dissolved in 10 ml of 2 M sulfuric acid and 10 ml
of absolute ethanol and cooled in a salt/ice water bath. Sodium nitrite (0.2 g) was slowly
added. The reaction then continued to stir for 2 hours more. From this step on, the reactions
were protected from light. Sodium azide (0.2 g) was slowly added, and then the reaction was
stirred for one hour at room temperature, followed by a half hour at 60°C. The product
mixture was evaporated only long enough to remove the ethanol, leaving the product in about
5 ml of water. The l4C-photoaffinity probe, containing the azide groups, were extracted three
times with ethyl acetate and evaporated to dryness. The product was taken up in 5 ml of
ethyl acetate and 50 ul was counted on the scintillation counter. The product was then dried
again and stored at -20°C. The specific activity of the photoaffinity probe was about 55

mCi/mole.

Growth of Sarcina ventriculi with the l"C-photoaffinity probe

The ”C-photoaffinity probe was dissolved in ethanol and half of it was split between
two exponentially growing cultures of Sarcina ventriculi (50 ml) at 37°C. The final ethanol
concentration per culture was 0.2 M, and the cultures were wrapped in aluminum foil.
Continued 37°C incubation for 6 hours. Cells were protected from light at all times. Cells
were harvested by centrifugation, and washed with Dulbecco's phosphate buffered saline (10

mM phosphate, 0.15 M NaCl, pH 7.0) containing 0.9 mM CaClz, 2.7 mM KCl, and 0.5 mM

192

MgClZ. Cells were lysed by passage through a French pressure cell (American Instruments
Co., Inc., Silver Spring, MD) and by freeze/thawing several times. The probe was then
cross-linked membrane components by exposing the lysate (in a quartz cuvette) to a high

intensity UV lamp for two minutes.

Isolation of potential protein involved in the cross-linking mechanism
Half of the l4C-lysate (1 ml) was separated by using a Sephadex G-75 column. The
amount of radioactivity for the first 100 fractions was determined. Fractions containing the

most radioactivity were lyophilized.

SDS-PAGE gels and Western transfers

The radioactive Sephadex G-75 fractions were lyophilized and separated on duplicate
12% SDS-PAGE gels using Laemmli’s discontinuous buffer system [1] on 12%
polyacrylamide slab mini-gels containing 0.32% bisacrylamide, 0.375 M Tris-HCl (pH 8.8),
and 0.1% SDS. The stacking gel contained 4% polyacrylamide, O. 1% bisacrylamide, 0.125
M Tris-HCl (pH 6.8), and 0.1% SDS. Electrophoresis was carried out at 200 V for about 1
hour. One gel was stained in a solution ofO. 1% Coomassie blue in 40% methanol/10% acetic
acid, and destained in a solution of 40% methanol/10% acetic acid. The other gel was
Western transferred to ProblottTM (Applied Biosystems) nylon membrane, which is suitable
for protein sequencing, following manufacturer instructions: The nylon membrane was
briefly wetted with methanol and then soaked in electroblotting buffer (10 mM 3-

[cyclohexylamino]-l-propanesulfonic acid {CAPS}, pH 11 in 10% methanol). The gel was

193

soaked in electroblotting buffer for 5 minutes. The transblotting sandwich was assembled
according to Bio-Rad manufacturer's instructions, and the transfer was carried out at 100V
on ice for 1.5 hours. The nylon membrane was Coomassie stained with 0.1% Coomassie
Blue R-250 in 40% methanol/1% acetic acid for less than one minute and then destained
with 50% methanol and rinsed well with deionized water. The membrane was air-dried and

then exposed to a phosphoimager.

PRELINIINARY RESULTS

Currently, work is being done to isolate components of the enzymatic system for tail-
to-tail coupling of membrane fatty acids. A l4C-labeled fatty acid photoaffinity probe was
synthesized (Figure A. 1) such that the reactive azide group would be positioned in the middle
of the membrane bilayer. After growing S. ventriculi in the presence of this probe, the probe
was cross-linked using UV light to membrane components. Since the cross-linking occurs
in the center of the membrane, it is reasonable to expect that enzymes involved in this
process must span the central portion of the membrane. Phosphoimages of Western blots
containing proteins from the Sephadex column fractions containing the highest amounts of
radioactivity show that only three of the void volumn fractions contain a detectable labeled
protein (Figure A2). The other fractions may also contain labeled proteins, but either not
enough protein was loaded onto the gel, or there is too little radioactivity to detect. In these
three void volumn fractions, a single 15 kDa protein is detected. Since the fractionation

range of Sephadex G-75 is 3 - 80 kDa, this protein was likely a large aggregate as it passed

194

OH OH OH 0H
0
on
0 _
0H
+ 12:3; H2/Pd Name/Ho NaNa
l4C-Iabelled.
and unlabelled
NO; NH: m‘c1 :4,

Figure A.1: Synthesis of the l“C-labeled photoaffinity probe.

 

Figure A.2: Phosphoimage of the nylon membrane containing proteins from several G-
75 column fractions. Only lanes containing proteins from column fractions 21, 22, and 23
had a radioactively labeled protein detectable by phophoimaging. The other lanes probably
did not contain enough protein and/or enough radioactivity to detect.

195

through the column. This would be expected as integral membrane proteins are usually
hydrophobic and thus they would tend to aggregate. This protein has been submitted for

sequencing, though no results have been obtained to date.

REFERENCES

1. Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head
of bacteriophage T4. Nature, 227 : 680—685.

"‘illlillllllllllES