v

35%
I .7 ‘

"u

,u, . .
‘n v

ta.‘r4'r_.‘.
'11 '.

v

g
II .I 7 .3».

. n
o.‘ ..;r'~'
r M

"‘ 7%

. _ .\
. ‘~.
1 .5

 

KY

0

‘ '._ 1.371,
Wag???“
m «a! ‘,,

.fi?§' ‘ I
"V“: 3 ?’ :4: . _
\‘a‘f- Ir! J ‘3‘“. ‘
1 .:.p;m, 4’va .3:- ‘V N N
,‘3‘ “i“ > .. . a
k; I >

5

O
5+»Efiéf.zi:éfi::>,<$:v¥}?—2>g,g
Iff‘l’J’J firrljy‘z‘F. |

.. I:
my).

33x

< ”a, ‘
~fI'I’V3'
' guru:
.14.‘4‘&'
/- - In. «.
tan-<- A
' , ’J {614?
(3‘ ’~ 7555'»

71? - ‘ :‘r ' .{
v-"r.-JI 1‘ n ‘r‘
, . A . .. _.
{fig-{£55, f}, “75$;
r; w (969% m a -
3'1~‘e‘-"J«”" fic-Tm‘x;£;-1=.~:;§ cc: .,
f" "’ , .
“i ’:v .4535.
I 'o
‘ I,
- k: 7" Q
‘1'.’é$f"’_o:’

Afr

' v: .

'\ O "
"qr,

. l
1 I ._‘
fitin-f
.IlevJ' ‘ m
.- 5","uf

(1‘ ‘
If, ";;J'En‘
. , 1-,

£1. ‘ ' -
. ' w ._ 3'."

7‘ (KI-"t. :7
MW‘zéJ» ,

3“ “3%:

J7. 5"? a
. JV _

n": 1*
1;}

m.- ,

. ’1
If}, - .
41.0, 2.35;:
”'7’... .7"; o“
a? ,1, .3‘ v"
"ff- 5"
‘J' r. .
41"? 1

7%“
.-,r_ '.' ,"

JI' v.
W

' 7}
.r,-',-‘.,r.lf?‘ :9
f 11'; 3,9211%,- ” I I;

”?W 7rd?

llllllllllllllllllllUlHlIll!llllilllllllllllllllllllllllllll

3 1293 00786 8148

LIBRARY
Mlchlgan State
University

 

r-—-wm..-.'

 

 

This is to certify that the

thesis entitled

CHARACTERIZATION OF THE PHYSICAL AND CHEMICAL STRUCTURE
OF PSEUDOMONAS AERUGINOSA LIPOPOLYSACCHARIDE

presented by

Mildred Rivera

has been accepted towards fulfillment
of the requirements for

Masters of Science degree in Biochemistry

 

 

Date Z/JP/YX/

0.7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

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

 

 

NDJLZJLZDUI
1Q1501

fil=—_l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An Affirmative ActiorVEquel'Opportunity lnditution
cmmmod

CHARACTERIZATION OF THE PHYSICAL AND CHEMICAL STRUCTURE

OF PSEUDOMONAS AERUGINOSA LIPOPOLYSACCHARIDE

 

BY

Mildred Rivera

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

MASTER OF SCIENCE

Department of Biochemistry

1988

ABSTRACT

CHARACTERIZATION OF THE PHYSICAL AND CHEMICAL STRUCTURE

OF PSEUDOMONAS AERUGINOSA LIPOPOLYSACCHARIDE

BY

Mildred Rivera

Pseudomonas aeruginosa is the dominant pathogen in respiratory
infections of patients with cystic fibrosis. During antibiotic
treatment of acute infections, this bacterium can alter its surface
structure so as to increase resistance to antibiotics and host defense
mechanism. Lipopolysaccharide (LPS) is a major component of the outer
membrane of gram-negative bacteria and its structure is critical in
forming a penetration barrier to amphipathic and hydrophobic compounds.
The structure of the LPS molecule is very heterogeneous and complex.
The first aim of this project was to determine the size heterogeneity of
LPS isolates from several 2, aeruginosa strains. The results from gel
filtration chromatography, sodium dodecyl sulfate-polyacrylamide gels
(SDS-PAGE). and Western blot revealed that LPS isolates from PA01
strains contain two distinct set of bands suggesting that PAOl strains
synthesize more than one type of chemically and antigenically distinct
molecule. Also, the results showed that the fraction of core
oligosaccharides carrying an O-specific polymer is less than 8%.

The second part of the project was to correlate antibiotic-
lipopolysaccharide interactions with antibiotic efficacy. The affinity

of cationic antibiotics, like polymyxin B and gentamicin, for LPS was

characterized. The alterations induced in the LPS aggregate structure
by polycations was measured. The magnesium (MgLPS) and calcium (CaLPS)
salts of the LPS from an antibiotic-sensitive, a revertant, and parent
strain of g. aeruginosa were formed. Relative affinities of various
polycationic antibiotics and polyamines for purified LPS was measured
using a cationic spin probe, CAT12. It was found that polycationic
antibiotics bind to and rigidify MgLPS and CaLPS complexes from E,
aeruginosa to different extents. The chemical basis for differences in
LPS-cation interactions among the three isolates may result from
substoichiometric modification of the core-lipid A. Furthermore, the
CaLPS complexes appeared to have a lower affinity for cationic
antibiotics.

Finally, to assess which chemical groups within the LPS molecule
are important in LPS-antibiotic interaction, the LPS from B-lactam and
aminoglycoside-resistant E. aeruginosa strains were analyzed by both gel
filtration and SDS-PAGE. Cation interaction with these LPS isolates and
with LPS from a rough mutant which contained lower amounts of phosphate
were studied. The results showed that, for the B-lactam-resistant
mutants, the LPS had an altered size distribution of the
O-polysaccharide chain lengths. LPS from the aminoglycoside resistant
mutant completely lacked O-polysaccharide chains. Differences in
binding of cationic compounds to the various LPS isolates was measured
with a cationic spin probe CAT12. Analysis of the partitioning of this
probe onto LPS isolates indicated that the length of the O-antigen and
the amount of phosphate in the core-lipid A region affected LPS

aggregate structure and polycation binding.

This thesis is dedicated to my Lord Jesus Christ

and my daughter Tanya Y. Collazo.

iv

ACKNOWLEDGEMENTS

I wish to express my gratitude to my mentor, Dr. Estelle J.
McGroarty, for her guidance, support, and friendship. Also, I am
grateful to Dr. Alfred Haug who provided critical insight into my
research and direct collaboration in the project.

I am indepted to my lab partners, Arnie and Warren, for sharing
with me the knowledge and misery in the world of scientific research.
Thanks are also due to Barb who taught me and helped me with the ESR
experiments. Special thanks to Sharon who gave me a lot of support in
the dark moments of chaos. I am very grateful to the members of my
committee for their valuable comments and suggestions during my
dissertation.

I would like to acknowledge the Cystic Fibrosis Foundation for
providing funds which supported this research.

I am grateful to Jan Wood for her patience and excellent
secretarial skills. Above all, I would like to acknowledge a special
group in the department of Biochemistry, my soul sisters Carmen and
Maria, for being a source of spiritual strength and intellectual
stimulus. To them and my other good friends Carol, Marco, Mark, Joan
and the people from Dr. Fraker's and Dr. Aust's lab: Thanks for the
moral support and friendship. I would also like to extend my gratitude
to my good friends in the Chemistry Department, especially the Puerto
Rican crowd, Aracelys, Tony, Fernando, Juan, Gladys, Aileen; and the
half P.R.C., Jeff and Dr. Harrison, for their support and friendship.

Finally, my deepest thanks is for my dear daughter Tanya and my
parents whose love and patience were my inspiration in the realization

of my goals.

TABLE OF CONTENTS

List of Figures . . . . . . . . . . . . . . . . . .
List of Tables . . . . . . . . . . . . . . . .

List of Abbreviations . . . . . . . . . . . . . . .

Chapter I: Literature Review . . . . . . . . . .
References . . . . . . . . . . . . . . . . . .

Chapter II: Heterogeneity of Lipopolysaccharides from
Pseudomonas aeruginosa: Analysis of Lipopolysaccharide

 

Chain Length . . . . . . . . . . . . . . . . .

Abstract . . . . . . . . . . . . . . .
Introduction . . . . . . . . . . . . . . . .
Materials and Methods . . . . . . . . . . . . .
Bacterial Strains . . . . . . . . . . . . .
Growth Media . . . . . . . . . . . . . . . .
Isolation of LPS . . . . . . . . . . .
SDS-PAGE . . . . . . . . . . . . . . . . . .
Column Chromatography . . . . . . . . . . .
Western Blots . . . . . . . . . . . . . . .
Assays . . . . . . . . . . . . . . . . . .
Results . . . . . . . . . . . . . . . . . . .
Discussion . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . .

Chapter III: Binding of Polycationic Antibiotics to
Lipopolysaccharide from an Antibiotic-sensitive
Pseudomonas aeruginosa . . . . . . . . .

 

Abstract . . . . . . . . . . . . . . . . . . .
Introduction . . . . . . . . . . . . . . . . .
Materials and Methods . . . . . . . . . . . . .
Bacterial Strains and Growth Media . . . . .
Isolation of LPS . . . . . . . . . . . . . .
Chemical Analysis. . . . . . . . . . . . . .
SDS-PAGE and Western Blots . . . . . . . . .
Partitioning of Spin Probe . . . . . . . . .
Chemicals. . . . . . . . . . . . . . . . . .
Results . . . . . . . . . . . . . . . . . . . .
Discussion . . . . . . . . . . . . . . . . . .
References. . . . . . . . . . . . . . . . . .

vi

Page

viii

xi

32

33
35
37
37
37
37

38
39
39
NO
63
69

Page
Chapter IV: Cation Interactions with Lipopolysaccharide from
Pseudomonas aeruginosa PAO Strains Altered in O-antigen
Chain Lengths . . . . . . . . . . . . . . . . . . . . . . . . 119

 

Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Introduction . . . . . . . . . . . . . . . . . . . . . . . . 122
Materials and Methods . . . . . . . . . . . . . . . . . . . . 125
Bacterial Strains and Growth Media . . . . . . . . . . . . 125
Isolation of LPS . . . . . . . . . . . . . . . . . . . . . 125
SDS-PAGE and Western Blots . . . . . . . . . . . . . . . . 126
Column Chromatography. . . . . . . . . . . . . . . . . . . 127
Chemicals Analysis . . . . . . . . . . . . . . . . . . . 127
Partitioning of Spin Probe . . . . . . . . . . . . . . . . 127
Chemicals. . . . . . . . . . . . . . . . . . . . . . . . 128
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . 1AA
References. . . . . . . . . . . . . . . . . . . . . . . . . . 151

Chapter V: Summary . . . . . . . . . . . . . . . . . . . . . . 159

Appendix A: Electron Spin Resonance Theory . . . . . . . . . . . 161

Appendix B: Publications . . . . . . . . . . . . . . . . . . . . 1?“

vii

LIST OF FIGURES

Figure Page

CHAPTER I

1 Schematic diagram of the cell envelope of gram-negative
bacteria 0 O I 0 O O O 0 O O 0 O O O O O O O O O O O O O O O 3

2 Proposed structure for the lipopolysaccharide molecule
Of £0 aeruginosa PAO1 O O O O O 0 O O O O O O O O O O O O O O 6

3 Proposed structure of LPS from the heptoseless mutants of
Escherichia coli D21f2 and D31mu . . . . . . . . . . . . . . 9

 

A Proposed structure for the lipid A of Pseudmonas

aeruginosa . . . . . . . . . . . . . . . . . . . . . . . . . 11

5 Schematic diagram of the self-promoted uptake pathway
in Pseudomonas aeruginosa . . . . . . . . . . . . . . . . . . 17

 

CHAPTER II

1 Silver-stained SDS-polyacrylamide gel and Western blots of
LPS from P. aeruginosa strains 1715, 1716, 503, and PA21 . . A2

2 Fractionation of LPS from P. aeruginosa strain 503 on
Sephadex G-200 . . . . . . . . . . . . . . . . . . . . . . . A6

3 Fractionation of LPS from P. aeruginosa strain 1715 on
sephadex 0-200 0 O O O I O O O O O I I O O O O O O O O O O O “8

A Fractionation of LPS from P. aeruginosa strain Z61 on
sephadex 6-200 0 O O O O O O O I O O O O 0 O O O O O 0 O O O 50

5 Silver-stained SDS-PAGE (11% acrylamide) and Western blots
of LPS fractions from P. aeruginosa reacted with monoclonal
anti-503 LPS antibody . . . . . . . . . . . . . . . . . . . . 57

6 Dot blots of pooled column fractions from P, aeruginosa

1715 LPS . . . . . . . . . . . . . . . . . . . . . 61

CHAPTER III

1 Silver-stained SDS-PAGE and Western blots of LPS from
P. aeruginosa FAQ 1670, FAQ 1716, and FAQ 1715 . . . . . . . 88

2 Scatchard plots of CAT12 binding to magnesium (o) and calcium
(0) salts of LPS from P. aeruginosa strain FAQ 1670 . . . . . 93

viii

Figure

3

The partitioning parameters ii measured as a function of added
(A) MgClz, (B) CaClz, (C) spermine, (D) gentamicin C, and
(E) polymyxin B . . . . . . . . . . . . . . . . . . . . . . .

Partitioning of the spin probe CAT12 upon addition of
cations to (A) MgLPS and (B) CaLPS from P. aeruginosa
strain FAQ 1715 . . . . . . . . . . . . . . . . . . . . . . .

Head group mobility of LPS from P. aeruginosa strain FAQ 1670

(0); FAQ 1716 (o); and PAO 1715 (A) was measured by the

hyperfine splitting parameter (ZTy, Gauss) of bound

CATIZ 0 O O 0 O O O O O O O O O 0 O O O O O O 0
CHAPTER IV

Silver-stained SDS-PAGE and Western blots of LPS from

P. aeruginosa strains PAO 503; PCC 118; PCC A5; PAO 503-18;

and H23" . . . . . . . . . . . . . . . . . . . . . . . . . .

Fractionation of LPS from P. aeruginosa strains (A) PAO 503,
(B) PCC 118, and (C) PCC A5 on Sephadex G- 200 . . . . . .

The partitioning paramter Vi measured as a function of

added (A) polymyxin B, (B) gentamicin C, (C) spermine, and
(D) caClz O O O O I O O O O O O O O O O O O O O 0 O O O O O O
The partitioning parameter Vi measured as a function of

added polymyxin B . . . . . . . . . . . . . . . . . . . . . .

APPENDIX

The spinning electron is characterized by a dipole moment
which will align itself with the line of force of a
Strong magnetic field 0 O O O 0 O O 0 O O O O O O I O O O O O

Cationic spin probe A-(dodecyldimethylammonio)-1-oxy-2,2,6,6-
tetramethylpiperidine bromide (CAT12) . . . . . . . . . . . .

Stimulated nitroxide spectra . . . . . . . . . . . . . . .

Rigid limit nitroxide spectrum (broken lines), illustrating the

the measurement of the parameter A'zz (-A'z) and
A22 (3A2) O O O O O O O I O O O O O I O O O O O O O I O I O I

ix

Page

96

99

101

131

13“

1A0

1A2

163

166
168

172

LIST OF TABLES

 

Table Page
CHAPTER I
1 Structure of O-repeating Units in Pseudomonas aeruginosa
Lipopolysaccharide . . . . . . . . . . . . . . . . . . . . . 1“
CHAPTER II

1 Chemical Analysis of P. aeruginosa B-Band LPS Fractions . . . 52
2 Composition of the Pooled Column Fractions of LPS from
P. aeruginosa Strains 1715 and PAZl . . . . . . . . . . . . . 55
CHAPTER III

1 Elemental Composition of Purified P. aeruginosa LPS
ISOlates O O O O O O D 0 O O O O 0 I O O O O O I O O O I O 0 89

2 Scatchard Analysis of CAT12 Binding to LPS Isolates of
£0 aeruginos I 0 O O O I O O O O O O O 0 O O O 0 O O O O O 91
CHAPTER IV

1 Percent KDO Recovery in Peaks 1, 2, 2a, and 3 from P.
aeruginosa LPS Fractionated on Sephadex G-200 . . . . . . . . 135

2 Elemental Composition of Purified P. aeruginosa MgLPS
Isolates . . . . . . . . . . . . . . . . . . . . . . . . . . 137

CaLPS

CAT12

CF

EDTA

ESR

FucNAc
GalNAc
GalNacUA
Glc
Glc(NAc)2UA
Gul(NAc)2UA
HEPES
Hex(NAc)20A
KDO

LPS

ManImUA
Man(NAc)2UA
MgLPS
2-OAtha
QuiNAc

Rha

SDS-PAGE

LIST OF ABBREVIATIONS

calcium salt of the LPS isolate

A-dodecyldimethylammonium-1—oxyl-2,2,6,6-
tetramethylpiperidine bromide

cystic fibrosis

ethylenediaminetetraacetate

electron spin resonance

N-acetylfucosamine

N-acetylgalactosamine

N-acetylgalactosaminuronic acid

glucose

2,3-diacetamido-2,3-dideoxyglucuronic acid

2,3-diacetamido-Z,3-dideoxyguluronic acid

N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid

2,3-diacetamido-2,3-dideoxy(hex)uronic acid

2-keto-3-deoxyoctulosonic acid

lipopolysaccharide

2-imidazolinomannuronic acid

2,3-diacetamido-2,3-dideoxy-D-mannuronic acid

magnesium salt of the LPS isolate

2-O-acetylrhamnose

N-acetquuinovosamine

rhamnose

sodium dodecyl sulfate-polyacrylamide gel

electrophoresis

xi

CHAPTER I: LITERATURE REVIEW

Cystic fibrosis (CF) is an inherited disease of children,
adolescents, and young adults. Most patients with CF develop lung

disease and the dominant pathogen in pulmonary infections is Pseudomonas

 

aeruginosa (2,A,18,2A). With antibiotic therapy P. aeruginosa replaces

Staphlococcus aureus as the main infectious agent (11). These lung

 

infections are often difficult to treat with drugs that are currently
available. However, antimicrobial chemotherapy has shown to be of
dramatic value (18). With prolonged therapy, bacterial isolates from
the sputum may become mucoid (A,11,18,23) or resistant to
aminoglycosides (A,5) or to B-lactams (25,68) and may become altered in
the level of O-polysaccharide bound to the lipopolysaccharide (LPS) in
the outer membrane (5.21). Low level aminoglycoside resistance that can
develop is reported to be a result of decreased antibiotic permeability
due to specific changes in LPS composition (5). Thus the structure of
the outer membrane surface of P. aeruginosa appears critical for
antibiotic permeation of the infecting cell during treatment of CF
patients. In this introduction, a description of the architecture and
components of the P. aeruginosa outer membrane will be given based on
the knowledge on the outer membrane of this and other gram-negative
bacteria, especially that of enteric bacteria.

The gram-negative cell envelope consists of two membranes separated
by a layer of peptidoglycan, which contributes to the mechanical
rigidity, and a cellular compartment called the periplasm (Fig. 1). The
innermost, cytoplasmic membrane is generally a phospholipid bilayer

embeded with a wide variety of polypeptides. The major functions of the

Figure 1.

Schematic diagram of the cell envelope of gram-negative bacteria.
The outer membrane structure is modeled from data obtained on

Pseudomonas aeruginosa. The labeled proteins are porins (P),

 

lipoproteins (L, which in P. aeruginosa are noncovalently

 

associated with the peptidoglycan but in enteric organisms are
partly covalently attached), and periplasmic substrate binding
proteins (B). Significant features of this diagram stressed in the
content of this chapter are (1) LPS as the major cell surface
lipidic molecule, (2) cross-bridging of adjacent LPS molecules by
Mg2+, (3) the presence of hydrophylic channels of defined exclusion
limit formed by outer membrane porins, and (A) the lack of porin
channels and presence of lipid bilayer (as Opposed to LPS; lipid

bilayer) in the inner (cytoplasmic) membrane.

A A A
SAAB BAA

Rim RKTA%A%I:T

cytoplasmic membrane proteins include energy generation, active and
facilitated transport of nutrients, export of toxic byproducts, and
enzymatic synthesis and translocation of cell envelope components
(20,62).

The periplasm contains a variety of enzymes, some of which function
as scavengers or processing enzymes for conversion of nontransportable
metabolites to substrates which can be transported (20,27). Little is
known about the barrier function of the periplasm nor the peptidoglycan,
although the peptidoglycan serves as a primary shape and osmotic
stability determinant of the cell (20).

The outer membrane is an unusual biological membrane in that its
outer monolayer contains lipopolysaccharide (LPS) as its major lipidic
molecule, while the inner leaflet contains phospholipids, with
phosphatidylethanolamine the predominant species (20,39,A8,A9). The
protein composition is unusual in that it is comprised of only several
"major" proteins (20,A7,A9). In the enterics, as well as P. aeruginosa,
this membrane contains several major proteins of approximately 30 to 36
kDa molecular weight; one class of these proteins, the porins, forms
trimers within the membrane which function as hydrophilic pores (A7,A9).

The outer membrane is very important for gram-negative bacteria,
making them resistant to host defense factors such as lysozyme, B-lysin
and various leukocyte proteins, which are very toxic to gram-positive
bacteria (12.51.59). For enteric gram-negative bacteria, the outer
membrane is a very effective barrier, giving protection to cells from
the detergent action of bile salts and from degradation by digestive
enzymes (A8). At the same time the outer membrane of a number of

gram-negative bacteria, including Pseudomonas, acts as a strong

 

Figure 2. Proposed structure for the lipopolysaccharide molecule of P.

aeruginosa PAO1 (see ref. 30,3A,37).

 

T

O-ontigen

A

 

 

 

 

 

 

CO re

—r®<9
.-

\

m?

KDO

 

 

e® ®9

Lipid A

l

 

rough

smooth

 

permeability barrier to many antibiotics that are effective against
other bacteria (20,A7-A9,70). Another important function of the outer
membrane is to make the cell surface strongly hydrophylic, important in
evading phagocytosis, in some complement resistance, and in the capacity
to avoid a specific immune attack due to alterations in the surface
antigen constitution (AO,A6,A9). These surface functions as well as the
construction of a highly impermeable membrane layer involve LPS, a
characteristic component of the outer membrane.

The biochemical structure of LPS from enterobacteria, Pseudomonas

 

and other gram-negative bacterias, can be divided into three parts: the
lipid A component, the polysaccharide core, and the O-antigenic side

chain (Fig. 2). Like most LPS's, the lipid A from Enterobactericeae

 

consists of a A-phosphoglucosaminyl-(1-+6)-glucosamine-1-phosphate
backbone, to which several fatty acid chains are attached via ester and
amide linkages (A7.58,67,73). In this structure (Fig. 3), the fatty
acid residues attached directly to the disaccharide backbone are all
3-OH-tetradecanoic acid, a component that occurs in the LPS of most
gram-negative bacteria (6,A7,A9,73). Interestingly, in P. aeruginosa
there is little of this fatty acid, which is replaced mainly by 2-OH and
3-OH-dodecanoic acids and some 3-OH-decanoic acid (Fig. A) (13,A7,56).
A certain variability is also noted with regard to the presence and
nature of polar phosphate head groups if enterobacterial lipid A
structures are compared (60). In many cases the substitution on the
phosphoryl residues are not stoichiometric (60). Thus, despite
differences of detail, the backbone of lipid A from P. aeruginosa is

similar to that of enterobacteria. Whether the phosphate residues in

Figure 3. Proposed structure of LPS from the heptoseless mutants of

Escherichia coli D21f2 and D31mu (taken from ref. 67).

 

R: H. lauroyl. WWW”-
Of 0-3-hydgoxyrnynst0yl

   

 

10

Figure U. Proposed structure for the lipid A of Pseudomonas aeruginosa.

 

11

IO

IO

0

IO

00-40 96/0 o «:0 __
o

12

the 1' and A positions of the glucosamine disaccharide serve as points
of attachment for other residues is not yet known (13,73).

Covalently attached to lipid A is the rough oligosaccharide core
containing in its proximal portion an unusual sugar, 2-keto-3-
deoxyoctonate (KDO), as well as a variety of more distal heptose and
hexose residues (Fig. 2) (38,39). In addition to these sugars, the
known characteristic components of core oligosaccharides from g.
aeruginosa are D-glucose, L-rhamnose, D-galactosamine, and L-alanine

("7,61,73). In 3. aeruginosa the phosphate content of the isolated core

 

oligosaccharide is much higher than that of Enterobactericeae (1N,N7,73).
Analyses of whole LPS suggest that there are 10 or more phosphate
residues per chain of core oligosaccharide (“7.73) compared with

Salmonella minnesota and Escherichia coli LPS, which reportedly contain

 

 

M-7 phosphate moieties (13,57). Also, the core region of enterobacteria
as well as E. aeruginosa exhibits a high degree of structural
heterogeneity (38,“7,73). For instance, in enterobacteria some of the
inner core heptose, phosphate, ethanolamine, and possibly
N-acetylglucosamine and galactose may not be present in molar amounts
(38). It seems likely that the inner core region of LPS from g.
aeruginosa basically follows the enterobacterial pattern but less is
known of the nature of its substoichiometric substitutions ("7).

In smooth strains, O-polysaccharide is attached to the core region
of LPS (26,52). As indicated in Fig. 2, the O-specific chains of LPS
are made up of repeating units of identical oligosaccharides (73). The
O-side chains from g. aeruginosa are rich in amino sugars, some of which
are unique among natural products (29-31,3u,u7). Neutral sugars

confirmed as integral components of many O-polymers are L-rhamnose and

13

D-glucose (73). Table 1 summarizes the available information on the
known components of the O-specific fractions representing different
serotypes or immuno-types. Strain PA01 of g. aeruginosa, which has been
used most frequently in genetic and outer membrane studies, was recently
shown to have an O-antigen composition identical to the Lanyi 3a,d type
(35,”7). The characterization of E. aeruginosa O-specific
polysaccharides has been complicated in some cases by chemical
heterogeneity of the polysaccharide chains (7,8,33,75). It has been
reported that the polymeric material can be resolved into amino
sugar-rich and neutral sugar-rich fractions (33.63.75.78). The
biological significanceof such fractions is unclear, but a possible
explanation is that E. aeruginosa strains produces multiple types of
molecules with chemically distinct polysaccharide chains.

As described earlier, LPS contains a variety of ionic groups, with
acidic phosphate and carboxyl moieties concentrated within the core and
lipid A head group region (38,39,u9). As a consequence, the LPS carries
a net negative charge at physiological pH resulting in a strong negative
charge on the surface of gram-negative cells (20). LPS is apparently
associated with a variety of cations including Mg2+ and Ca2+ (10,u9),
which may form ionic bridges between phosphate groups on neighboring LPS
molecules stabilizing the outer membrane structure (36,u9). Recent 31?
NMR studies (67) suggest the participation of both the carboxyl group of
KDO and phosphate groups in the lipid A in the coordinate bind of metal
ions. Treatment of gram negative cells with ethylenediaminetetraacetate
(EDTA) removes, by chelation, divalent cations and consequently disrupts
the outer membrane (36,“7). Therefore, the combination of negative

charge and divalent cation cross-bridging of LPS provides the

14

 

 

 

Table 1. Structure of O-repeating units in Pseudomonas aeruginosa
lipopolysaccharide

Lanyi IATS

type type O-repeating unit

2a,b 1O u)L-GalNAcUA(a1-3)D-QuiNAc(a1-3)L-2-OAtha(a1-3)L-2-OAtha(1-

2a,c U)L-GalNAcUA(a1-3)D-QuiNAc(a1-3)L-Rha(a1-3)L-Rha(a1-

3a,b 16 u)D-ManImUA(81-U)D-Man(NAc)2UA(81-3)D-FucNAc(81-

3(a),c 2 N)D-ManImUA(81-“)L-Gul(NAc)2UA(al~3)D-FucNA(B1—

3a,d 5 U)D-ManImUA(81-N)D-Man(NAc)2UA(81-3)D-FucNAc(a1-

3a,d,e N)D-ManImUA(81-u)L-Gul(NAc)2UA(a1-3)D-FucNAc(a1-

3(a),d,f M)D-ManImUA(81—?),Hex(NAc)2UA,D-FucNAc

6 1 u)D-GalNAc(a1-H)D-Glc(NAc)2UA(B1-3)D-FucNAc(a1-3)D-QuiNAc(a1-

7ab(ac) 11 3)L-FucNAC(a1~3)D-FUCNAC(81-2)D-GlC(B1-

 

Nonstandard abbreviations: QuiN, quinovosamine; FucN, fucosamine;
(NAc)ZUA, 2,3-diacetamido-2,3-dideoxy(hex)uronic acid; and ManImUA,
2-imidazolinomannuronic acid. IATS type refers to the serotype
according to the international antigen typing scheme commercially
marketed by Difco (taken from ref. “7).

15

gram-negative cell surface with a tight barrier, important for the
cells' resistance to hydrophobic antibiotics, bile salts, detergents,
protease, lipases, and lysozymes (36,“9).

The outer membrane, as mentioned before, may be considered a
hydrophobic bilayer interspersed with hydrophilic channels and
surrounded by a hydrophilic polysaccharide net (15,"9). An antibiotic
molecule, or any other molecule not using a specific outer membrane
transport system, can only reach its target site by one of three
possible pathways: the hydrophilic, the hydrophobic, and the
self-promoted pathway (20). The hydrophilic pathway is provided by the
aqueous pore. These pores allow the passage of small hydrophilic
compounds (e.g. B-lactams) of sizes smaller than 600 daltons for E. coli

and below 9,000 daltons for Pseudomonas (20,"9). The second pathway,

 

probably utilized by hydrophobic antibiotics, is through the hydrophobic
domain of the bilayer and is essentially inoperative in gram negative
bacteria. Thus, hydrophobic antibiotics generally are ineffective
against gram-negative bacteria, except with deep rough mutants in which
LPS is deficient in core structure (“5), or with certain antibiotic
supersensitive mutants (9). Finally, the self-promoted uptake pathway
has been postulated to be involved in the transport of polycationic
antibiotics, like polymyxin and aminoglycosides, across the outer
membrane of g. aeruginosa (20,u9). Compounds utilizing this pathway
generally are polycationic and are thought to displace divalent metal
cations destroying the ionic bridges between phosphate groups on
neighboring LPS molecules destabilizing the outer membrane structure
(Fig. 5) (20,u9). Thus, such compounds increase outer membrane

permeability to proteins and hydrophobic compounds (20,22,23,u9,66).

16

Figure 5. Schematic diagram of the self-promoted uptake pathway in

Pseudomonas aeruginosa. The labeled proteins are porins (P)

 

'and lipoproteins (L).

17

R Membrane
i 6 mm

Peptido-

/77//'/ [con
, 1/7/17—7% xi/fi 77/77 gy

Hydrophilic
LPS Pore ‘M ‘
J M9
{ fl“ Outer

<—- Periplosm

18

These cationic antibiotics bind to and perturb the packing arrangement
of isolated LPS (3.5“), suggesting that their site of interaction on the
outer membrane is with LPS. In addition, polymyxin B and gentamicin
induce blebbing and other structural perturbations of bacterial outer
membranes which may serve to make the outer membrane more permeable to
the antibiotic (32,42,6u,66).

Further evidence that disorganization of the outer membrane is
critical in the action of polymyxin was obtained from the study of

polymyxin-resistant (pmr A) mutants of Salmonella typhimurium (“1,71,72).

 

These mutants demonstrate low-level resistance to polymyxin, are
impermeable to lysozyme and deoxycholate upon treatment with polymyxin
(71), and are more resistant to Tris-EDTA and polycations such as
polylysine and protamine (69). Compared to the parental isolate, the
LPS of the pm: A mutants was shown to contain elevated amounts of
M-aminoarabinose and ethanolamine, making the LPS less negatively
charged (72), and thus less able to interact with polycations (53).
Similarly, the E. aeruginosa mutant H181 was cross resistant to the
polycationic antibiotics and aminoglycosides, and to the divalent cation
chelator EDTA due to an altered outer membrane (an). In wild type 3.
aeruginosa cells, a statistically significant relationship was observed
between a pseudoaffinity constant for outer membrane permeabilizing
ability and the MIC for eight different aminoglycosides (37).

Changes in the outer membrane LPS, phospholipid, and/or protein
composition have been associated with resistance due to a decrease in
antibiotic permeation (15,17,22,h9,53). Alterations in the LPS of
gram-negative bacteria could potentially result in changes in

sensitivity to antibiotics, including cell wall synthesis inhibitors

19

(50). It has been shown that some mutants of gram negative bacteria
resistant to low levels of polycationic (16,72) and B-lactam (16,17)
antibiotics resulting from low rates of penetration across the outer
membrane, have altered LPS structure. The types of alterations in the
LPS structure include changes in the length of the 0-antigen chain
(19.55) and a decrease in the overall negative charge of the core-lipid
A region (“3,65,72,76). It has been proposed that an alteration in LPS
O-antigen length may influence the opened state of the outer membrane
pores (1.3“). Yamada and Mizushima (77), using X-ray diffraction,
demonstrated that both the lipid A moiety and the polysaccharide moiety
of LPS interact with the 0-8 porin trimer of E. coli. Furthermore, they
showed that the core-oligosaccharide moiety is involved in stabilizing
the proper porin lattice conformation (77).

In the past several years knowledge of the structure and function
of the g. aeruginosa outer membrane has expanded greatly. This outer
membrane shows unusual functional properties, i.e., very low
permeability toward lipophilic solutes and high permeability toward
hydrophilic solutes. These functional attributes can be correlated with
the structure of LPS and porins, and the precise molecular organization
of these components. As mentioned, the LPS molecules are very tightly
packed because of the divalent cation bridging of neighboring molecules
containing the unusually large number of phosphate residues. It has
been proposed that the ability of cationic antibiotics to penetrate the
outer membrane may be related to their ability to bind LPS and alter the
packing of the LPS molecules. To test this hypothesis we persued the

following objectives:

20

1. To analyze the size heterogeneity of LPS isolates from several
3. aeruginosa strains altered in the rate of antibiotic penetration of
the outer membrane.

2. To form the magnesium (MgLPS) and calcium (CaLPS) salts of the
LPS isolates, and to measure the relative affinities of various
polycationic antibiotics and polyamines to these purified LPS salts from
E. aeruginosa using a cationic spin probe.

3. To determine if the antibiotic-LPS interaction affect the
structure organization of the LPS aggregate.

u. To determine the structural changes in the LPS molecule which

correlate with resistance toward the polycationic antibiotics.

REFERENCES

Angus, B.L., A.M. Carey, D.A. Caron, A.M.B. Kropinski, and R.E.w.
Hancock. 1982. Outer membrane permeability in Pseudmonas
aeruginosa: comparison of a wild-type with an antibiotic
supersusceptible mutant. Antimicrob. Agents Chemother. 21:299-309.
Anwar, H., M.R.w. Brown, A. Day, and P.H. Weller. 198”. Outer

membrane antigens of mucoid Pseudomonas aeruginosa isolated

 

directly from the sputum of a cystic fibrosis patient. FEMS
Microbiol. Lett. 2u:235-239.

Bader, J., and M. Teuber. 1973. Action of polymyxin B on
bacterial membranes. I. Binding to the O-antigenic lipopoly-

saccharide of Salmonella typhimurium. Z. Naturforsch. Teil C

 

28:“22-“30.
Blessings, J., B. Maybury, and N. Lewiston. 1981. Antimicrobial

susceptibilities of Pseudomonas from sputum of patients: 1975-1980.

 

In Current Problems and New Trends in Cystic Fibrosis. Monographs

 

 

in Pediatrics, Vol. 1N, pp. 115-119. S. Karger. Basal.

Bryan, L.B., K. O'Hara, and S. Wong. 198R. Lipopolysaccharide
changes in impermeability-type aminoglycoside resistance in
Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 26:250-255.
Bryn, K., and E.T. Rietschel. 1978. L-2-Hydroxytetradecanoic acid
as a constituent of Salmonella lipopolysaccharides (lipid A). Eur.
J. Biochem. 86:311-315.

Chester, I.R., and P.M. Meadow. 1975. Heterogeneity of the

lipopolysaccharide from Pseudomonas aeruginosa. Eur. J. Biochem.

 

58:273-282.

21

10.

11.

12.

13.

1”.

15.

22

Chester, I.R., and P.M. Meadow. 1973. The relationship between the
O-antigenic lipopolysaccharides and serological specificity in

strains of Pseudomonas aeruginosa of different O-serotypes. J.

 

Gen. Microbiol. 78:305-318.
ChOpra, I., and P. Ball. 1982. In A.H. Rose and J.C. Morris

(eds.), Advances in Microbial Physiology pp. 183-2M1. Academic

 

Press, London.
Coughlin, R.T., S. Tosanger, and E.J. McGroarty. 1983.
Quantitation of metal cations bound to membranes and extracted

lipopolysaccharide of Escherichia coli. Biochemistry 22:2002-2006.

 

DiSant'Agnese, P.A., and P.B. Davis. 1976. Research in cystic
fibrosis (part 1-3). Res. Cystic Fibrosis 295:597-602.

Donaldson, D.M., R.R. Roberts, H.S. Larsen, and J.C. Tew. 197R.

~Interrelationship between serum beta-lysin, lysozyme, and the

antibody-complement system in killing Escherichia coli. Infect.

 

Immun. 10:657-666.
Drewry, D.T., J.A. Lomax, G.W. Gray, and S.D. Wilkinson. 1973.
Studies of lipid A fractions from lipopolysaccharides of

Pseudomonas aeruginosa and Pseudomonas alcaligenes. Biochem. J.

 

133:563-572.

Droge, W., E. Ruschmann, 0. Laderitz, and 0. Westphal. 1968.
Biochemical studies on lipopolysacchrides of Salmonella R mutants.
u. Phosphate groups linked to heptose units and their absence in
some R lipopolysacchrides. Eur. J. Biochem. h:13U-138.

Godfrey, A.J., and L.E. Bryan. 198A. Intrinsic resistance and

whole cell factors contributing to antibiotic resistance, pp.

16.

17.

18.

19.

20.

21.

22.

23

113-1u5. In L.E. Bryan (ed.), Antimicrobial Drug Resistance.

 

Academic Press, Inc., Orlando, FL.

Godfrey, A.J., and L.E. Bryan. 198M. Resistance of Pseudomonas

 

aeruginosa to new B-lactamase-resistant B-lactams. Antimicrob.
Agents Chemother. 26:“85-u88.

Godfrey, A.J., L. Hatlelid, and L.E. Bryan. 198M. Correlation
between lipopolysaccharide structure and permeability resistance in

B-lactam- resistant Pseudomonas aeruginosa. Antimicrob. Agents

 

Chemother. 26:181-186.

Gordts, B. 198M. Pseudomonas in cystic fibrosis. Anton.

 

vanLeerwen. 50:292-293.
Grossman, N., M.A. Schmetz, J. Foulds, B.N. Klima, V. Jimenez, L.L.
Leive, and K.A. Joiner. 1987. Lipopolysaccharide size and

distribution determine serum resistance in Salmonella montevideo.

 

J. Bacteriol. 169:856-863.

Hancock, R.R.W. 198M. Alterations in outer membrane permeability.
Ann. Rev. Microbiol. 38:237-264.

Hancock, R.E.W., L.M. Mutharia, L. Chan, R.P. Darveau, D.P. Speert,

and 0.8. Pier. 1983. Pseudomonas aeruginosa isolates from

 

patients with cystic fibrosis: A class of serum sensitivive,
non-typable strains deficient in lipopolysaccharide O-side chains.
Infect. Immun. "2:170-177.

Hancock, R.E.W., V.J. Raffle, and T.I. Nicas. 1981. Involvement
of the outer membrane in gentamicin and streptomycin uptake and

killing in Pseudomonas aeruginosa. Antimicrob. Agents Chemother.

 

19:777-785.

23.

2M.

25.

26.

27.

28.

29.

30.

24

Hancock, R.E.W., and P.G.W. Wong. 198“. Compounds which increase

the permeability of the Pseudomonas aeruginosa outer membrane.

 

Antimicrob. Agents Chemother. 26:“8-52.

Hoiby, N. 197”. Pseudomonas aeruginosa infection in cystic

 

fibrosis. Acta. Path. Microb. Scand. Sec. B. 82:551-558.
Horby, N.A., Heilesen, and N.E. Moller. 1981. Development of

Pseudomonas aeruginosa strains resistant to Carbenicillin,

 

Azlocillin, Piperacillin and Tobramycin in cystic fibrosis patients.

In Current Problems and New Trends in Cystic Fibrosis. Monographs

 

 

in Pediatrics, Vol. 1”, pp. 103-107. S. Karger. Basal.

Jann, K., and B. Jann. 198M. Structure and biosynthesis of
O-antigen, pp. 138-186. In E.Th. Rietschel (ed.), Handbook of
Endotoxin, 191. 1: Chemistry of Endotoxin. Elsevier Science
Publishers B.V.

Kennedy, E.P. 1982. Osmotic regulation and the biosynthesis of

membrane-derived oligosaccharides in Escherichia coli. Proc. Natl.

 

Acad. Sci. USA 79:1092-1095.

Kilbourn, J.P. 198“. Composition of sputum from patients with
cystic fibrosis. Current Mibrobiol. 11:19-22.

Knirel, Y.A., B.V. Vinogradov, A.S. Shashkov, B.A. Dmitriev, N.K.
Kochetkov, E.S. Stanislavsky, and M. Mashilova. 1987. Somatic

antigens of Pseudomonas aeruginosa. The structure of the

 

O-specific polysaccharide chain of the lipopolysaccharide from P.
aeruginosa 013 (Lanyi). Eur. J. Biochem. 163:627-637.

Knirel, Y.A., B.V. Vinogradov, A.A. Shashkov, B.A. Dmitriev, N.K.
Kochetkov, E.S. Stanislavsky, and G.M. Mashilova. 1983. Somatic

antigens of Pseudomonas aeruginosa. The structure of O-specific

 

31.

32.

33.

3“.

35.

36.

37.

25

polysaccharide chains of P. aeruginosa O:3(a),c and O:3a,d,e
lipopolysaccharides. Eur. J. Biochem. 13H:289-297.

Knirel, Y.A., B.V. Vinogradov, A.A. Shashkov, B.A. Dmitriev, N.K.
Kochetkov, E.S. Stanislavsky, and G.M. Mashilova. 1982. Somatic

antigens of Pseudomonas aeruginosa. The structure of O-specific

 

polysaccharide chains of P. aeruginosa 0:3a,b and 0:3a,d
lipopolysaccharides. Eur. J. Biochem. 128:81-90.

Koike, M., K. Iida, and T. Matsuo. 1969. Electron micrOSCOpic
studies of mode of action of polymyxin. J. Bacteriol. 97:uu8-u52.
Koval, S.F., and P.M. Meadow. 1977. The isolation and
characterization of lipopolysaccharide-defective mutants of

Pseudomonas aeruginosa PAC1. J. Gen. Microbiol. 98:387-398.

 

Kropinski, A.M., C.V. Chan, and F.H. Milazzo. 1979. The

extraction and analysis of lipopolysaccharides from Pseudomonas

 

aeruginosa strain PAO, and three rough mutants. Can. J. Microbiol.
25:390-398.
Kuzio, J. and A.M.B. Kropinski. 1983. O-antigen conversion of

Pseudomonas aeruginosa by bacteriophage D3. J. Bacteriol.

 

155:203-212.

Leive, L. 197A. The barrier function of the gram-negative
envelope. Ann. N.Y. Acad. Sci. 235:109-129.

Loh, B., C. Grant, and R.E.W. Hancock. 1984. Use of the
fluorescent probe 1-N-phenylnaphthylamine to study the interactions
of aminoglycoside antibiotics with the outer membrane of

Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 26:5“6-551.

 

38.

39.

HO.

“1.

N2.

“3.

un.

26

Lflderitz, 0., M.A. Freudenberg, C. Galanos, V. Lehmann, E.Th.
Rietschel, and D.H. Shaw. 1982. Lipopolysaccharides of
gram-negative bacteria. Curr. Topics Membs. Transp. 17:79-150.
Lugtenberg, B., and L. vanAlphen. 1983. Molecular architecture
and functioning of the outer membrane of Escherichia coli and other
gram-negative bacteria. Biochim. Biophys. Acta 737:51-115.

Makela, P.H., D.J. Bradley, H. Brandis, M.M. Frank, H. Hahn,

W. Henkel, K. Jann, S.A. Marse, J.B. Robbins, L. Rosenstreich, H.
Smith, K. Timmis, A. Tomaz, M.J. Turner and D.C. Wiley. 1980.
Evasion of host defense group report, pp. 17N-197. In H. Smith,

J.J. Skehel, and M.J. Turner (eds.), The Molecular Basis of

 

Microbial Pathogenity. Dahlem Konferenzen 1980. Verlag Chemie

 

GmbH, Weinheim, Federal Republic of Germany.
Makela, P.H., M. Sarvas, S. Calcagno, and K. Lounatmaa. 1978.

Isolation and characterization of polymyxin-resistant mutants of

Salmonella. FEMS Microbiol. Lett. 3:323-326.

Martin, N.L., and T.J. Beveridge. 1986. Gentamicin interaction

with Pseudomonas aeruginosa cell envelope. Antimicrob. Agents

 

Chemother. 29:1079-1087.
Moore, R.A., and R.E.W. Hancock. 1986. Involvement of outer

membrane of Pseudomonas cepacia in aminoglycoside and polymyxin

 

resistance. Antimicrob. Agents Chemother. 30:923-926.
Nicas, T.I., and R.E.W. Hancock. 1980. Outer membrane protein H1

of Pseudomonas aeruginosa: Involvement in adaptive and mutational

 

resistance to ethylenediaminetetraacetate, polymyxin B and

gentamicin. J. Bacteriol. 1H3:872-878.

#5.

A6.

“7.

A8.

“9.

50.

51.

52.

53.

27

Nikaido, H. 1976. Outer membranes of Salmonella typhimurium.

 

Transmembrane diffusion of some hydrophobic substances. Biochem.
Biophys. Acta “33:118-132.
Nikaido, H. 1970. Lipopolysaccharide in the taxonomy of

Enterobacteriaceae. Int. J. Syst. Bacteriol. 20:383-AO6.

 

Nikaido, H., and R.E.W. Hancock. 1986. Outer membrane

permeability of Pseudomonas aeruginosa. In The Bacteria, Vol. X,

 

 

pp. 1A5-193. Academic Press.

Nikaido, H., and T. Nakae. 1979. The outer membrane of
gram-negative bacteria. Adv. Microb. Physiol. 20:163-250.

Nikaido, H., and M. Vaara . 1985. Molecular basis of bacterial
outer membrane permeability. Microbiol. Rev. ”9:1-32.

Parr, T.R., Jr., and L.E. Bryan. 198A. Nonenzymatic resistance to
B-lactam antibiotics and resistance to other cell wall synthesis

inhibitors. pp. 81-111. In. L.E. Bryan (ed.), Antimicrobiol Drug

 

Resistance, Academic Press, Inc., Orlando, FL.

Patterson-Delafield, J., R.J. Martinez, and R.I. Lehrer. 1980.
Microbicidal cationic proteins in rabbit alveolar macrophages: a
potential host defense mechanism. Infect. Immun. 30:180-192.
Perry, M.B., L. MacLean, and D.W. Griffith. 1986. Structure of
the O-chain polysaccharide of the phenol-phase soluble

lipopolysaccharide of Escherichia coli O:157:H7. Biochem. Cell

 

Biol. 6A:21-28.
Peterson, A.A., S.W. Fesik, and E.J. McGroarty. 1987. Decreased
binding of antibiotics to lipopolysaccharides from

polymyxin-resistant strains of Escherichia coli and Salmonella

 

typhimurium. Antimicrob. Agents Chemother. 31:230-237.

 

5“.

55.

56.

57.

58.

59.

60.

28

Peterson, A.A., R.E.W. Hancock, and E.J. McGroarty. 1985. Binding
of polycationic antibiotics and polyamines to lipopolysaccharides

of Pseudomonas aeruginosa. J. Bacteriol. 16A:1256-1261.

 

Peterson, A.A., A. Haug, and E.J. McGroarty. 1986. Physical
properties of short- and long-O-antigen-containing fractions of

lipopolysaccharide from Escherichia coli. O111:BA. J. Bacteriol.

 

165:116-122.

Pier, G.B., R.B. Markham, and D. Eardley. 1981. Correlation of the
biological responses of C3H/Hej mice to endotoxin with the chemical
and structural properties of the lipopolysaccharides from

Pseudomonas aeruginosa and Escherichia coli. J. Immunol.

 

 

127:18H-191.
Prehm, P., S. Stirm, K. Jann, and H.G. Boman. 1976. Cell wall

lipopolysaccharides of ampicillin-resistant mutants of Escherichia

 

coli K-12. Eur. J. Biochem. 66:369-377.
Qureshi, N., K. Takayama, D. Heller, and C. Fenselau. 1983.
Position of ester groups in the lipid A backbone of

lipopolysaccharides obtained from Salmonella typhimurium. J. Biol.

 

Chem. 258:129u7-12951.

Rest, R.F., M.H. Cooney, and J.K. Spitznagel. 1977.
Susceptibility of lipopolysaccharide mutants to the bactericidal
action of human neutrophil lysozomal fractions. Infect. Immun.
16:1A5-151.

Rietschel, E.Th., H.W. Wollenweber, H. Brade, U. ZAhringe, B.
Lindner, U. Seydel, H. Bradaczek, G. Barnickel, H. Labischinski,
and P. Giesbrecht. 198A. Structure and conformation of the lipid

A component of lipopolysaccharides. pp. 187-220. In E.Th.

61.

62.

63.

6A.

65.

66.

67.

29

Rietschel (ed.), Handbook of Endotoxin, Vol. 1: Chemistry of
Endotoxin. Elsevier Science Publishers B.V.
Rowe, P.S.N., and P.M. Meadow. 1983. Structure of the core

oligosaccharide from the lipopolysaccharide of Pseudomonas

 

aeruginosa PAC 1R and its defective mutants. Eur. J. Biochemm.
132:329-327.

Saier, M.H. 1979. In The Bacteria, Vol. VII, pp. 168-227. New
York Academic.

Sawada, S., T. Kawamura, Y. Masuho, and K. Tomibe. 1985. A new

polysaccharide antigen of strains of Pseudomonas aeruginosa

 

detected with a monoclonal antibody. J. Infect. Dis.
152:1290-1299.

Schindler, P.R.G., and M. Teuber. 1975. Action of polymyxin B on
bacterial membranes: morphological changes in the cytoplasm and in

the outer membrane of Salmonella typhimurium and Escherichia coli B.

 

 

Antimicrob. Agents Chemother. 8:95-1OA.
Sidorczyk, 2., U. ZEhringer, and E.Th. Rietschel. 1983.
Chemical structure of the lipid A component of the

lipopolysaccharide from Proteus mirabilis Re-mutant. Eur. J.

 

Biochem. 137:15-22.

Storm, D.R., K.S. Rosenthal, and P.B. Swanson. 1977. Polymyxin
and related peptide antibiotics. Ann. Rev. Biochem. A6z723-763.
Strain, S.M., S.W. Fesik, and I.M. Armitage. 1983. Structure and
metal-binding properties of lipopolysaccharides from heptoseless

mutants of Escherichia coli studied by C-13 and P-31 nuclear

 

magnetic resonance. J. Biol. Chem. 258:13A66-13U77.

68.

69.

70.

71.

72.

730

7”.

75.

3O

Szaff, M., and N. Horby. 1981. Antibiotic treatment of

Staphylococcus aerus infection in cystic fibrosis. In Current

 

Problems and New Trends in Cystic Fibrosis. Monographs in

 

 

Pediatrics, Vol. 1A, pp. 108-11“. S. Karger. Basal.

Vaara, M. 1981. Increased outer membrane resistance to
ethylenediaminetetraacetate and cations in novel lipid A mutants.
J. Bacteriol. 1A8:A26-A3A.

Vaara, M., and H. Nikaido. 1983. Molecular organization of

bacterial outer membrane. pp. 1-A5. In E.Th. Rietschel (ed.),

Handbook of Endotoxin, Vol. I: Chemistry of Endotoxins. Elsevier

Science Publishers B.V.
Vaara, M., and T. Vaara. 1981. Outer membrane permeability
barrier disruption by polymyxin in polymyxin-susceptible and

-resistant Salmonella typhimurium. Antimicrob. Agents Chemother.

 

19:578-583.

Vaara, M., T. Vaara, M. Jensen, 1. Helander, M. Nurminen, E.Th.
Rietschel, and P.H. Makela. 1981. Characterization of the
lipopolysaccharide from the polymyxin-resistant 2mg A mutants of

Salmonella typhimurium. FEBS Lett. 129:1A5-1A9.

 

Wilkinson, S.G. 1983. Composition and structure of

lipopolysaccharides from Pseudomonas aeruginosa. Rev. Infect. Dis.

 

5:89A1-S9N9.

Wilkinson, S.G. 1970. Cell walls of Pseudomonas species sensitive

 

to ethylenediaminetetraacetic acid. J. Bacteriol. 1OA:1035-1OUA.
Wilkinson, S.G., and L. Galbraith. 1975. Studies of

lipopolysaccharides from Pseudomonas aeruginosa. Eur. J. Biochem.

 

52:331-3H3.

76.

77.

78.

31

Wilkinson, S.G., L. Galbraith, and L.G. Lightfoot. 1973. Cell

walls, lipids, and lipopolysaccharides of Pseudomonas species.

 

Eur. J. Biochem. 33:158-17A.

Yamada, H., and S. Mizushima. 1980. Interaction between major
outer membrane protein (0-8) and lipopolyaccharide in Esherichia
coli K-12. Eur. J. Biochem. 103:209-218.

Yokota, S., S. Kaya, S. Sawada, T. Kawamura, Y. Araki, and E. Ito.
1987. Characterization of a polysaccharide component of

lipopolysaccharide from Pseudomonas aeruginosa IID 1008 (ATCC

 

2758A) as D-rhamnan. Eur. J. Biochem. 167:203-209.

CHAPTER II

HETEROGENEITY OF LIPOPCLYSACCHARIDES FROM PSEUDOMONAS

 

AERUGINOSA: ANALYSIS OF LIPOPOLYSACCHARIDE CHAIN LENGTH

32

ABSTRACT

Lipopolysaccharide (LPS) from smooth strains of Pseudomonas

 

aeruginosa 503, PAZ1, PAO1715, PAO1716, and 261 was fractionated by gel
filtration chromatography. Lipopolysaccharide samples from the first
four strains, all PA01 derivatives, separated into three major size
populations, while strain Z61, a Pae K799/WT mutant strain, separated
into two size populations. When column fractions were applied to sodium
dodecyl sulfate-polyacrylamide gels in their order of elution, molecules
of decreasing size were resolved, and the ladder of molecules with
different length O-antigens formed a diagonal across the gel. The LPS
from the PA01 derivatives contained two distinct sets of bands
distinguished on the gels as two sets of diagonals. The set of bands
with the faster mobility, the B bands, was found in column fractions
comprising the three major amino sugar containing peaks. In the sample
from 503, a fourth minor peak which contained B bands was resolved. The
slower moving set of bands, the A bands, were recovered in a minor peak.
LPS from strain 261 contained only one set of bands, with the higher
molecular weight molecules eluting from the column in a volume similar
to that of the B bands of the PA01 strains. Analysis of the fractions
of LPS from all strains indicated that less than 8% of the LPS molecules
had a long attached O-antigen. Analysis of the peak that contained
mainly A bands indicated a lack of reactive amino sugar and phosphate,
although heptose and 2-keto-3-deoxy-octulosonic acid were detected.
Reaction of isolated fractions with monoclonal antibody which is
specific for the PA01 O-antigen side chain, indicated that only the B

bands from the PA01 strains were antigenically reactive. The bands from

33

34

strain Z61 showed no reactivity. The data suggest that the A and B
bands from the PA01 strains are antigenically distinct. We propose that
PA01 strains synthesize two types of molecules that are antigenically

different.

INTRODUCTION

LipOpolysaccharide (LPS), a major component of the outer membrane
of gram-negative bacteria, is important in the structure (3A,38) and
function (3A,37) of this membrane. Structural microheterogeneity has
been demonstrated in several regions of LPS molecules from the

Enterobacteriaceae (3,15,23,38,39,A8,52) and Pseudomonas aeruginosa

 

 

(3A,58). Of the several methods used to separate the subclasses of LPS
from individual strains, sodium dodecyl sulfate-polyacrylamide gel
electrophoresis (SDS-PAGE) (21,25,A6,A8) and gel filtration (10.27.31,
33.35,A8) are the best. Either of these two methods by themselves,
however, may be insufficient to completely characterize the high and low
molecular weight fractions of LPS. Peterson and McGroarty (A8)
demonstrated that the SDS-PAGE of the column fractions of samples from

Salmonella typhimurium. Salmonella minnesota, and Escherichia coli was

 

 

 

instrumental in characterizing the various sized fractions. Analysis of
the isolated fractions allowed for the estimation of the average number
of O-antigen repeat units per LPS from each of the size fractions.
Compositional analysis of LPS from P. aeruginosa has indicated that
the LPS molecules are structurally similar to enterobacterial LPS
molecules, but possess several distinctive features (3A,58). The most
outstanding differences include the unusually high phosphate content
(3A,59). the presence of L-alanine in the core (3A,58), and the high
levels of amino sugars and uronic acids in the O-side chain (9,3A). The
characterization of P. aeruginosa O-specific polysaccharides has been

complicated in some cases by chemical heterogeneity of the

35

36

polysaccharide chains (9,10,31,59). In some instances, the polymeric
material has been resolved into amino sugar-rich and neutral sugar-rich
fractions (31,59). The biological significance of such fractions is
unclear, but a possible explanation is that P. aeruginosa strains
produce multiple types of molecules with chemically distinct
polysaccharide chains.

In this study, we have analyzed the size heterogeneity of LPS
isolates from several P. aeruginosa strains by both gel filtration and
SDS-PAGE. These studies have revealed that LPS isolates from PA01
strains contain two distinct sets of bands, suggesting that PA01 strains
are capable of synthesizing more than one type of chemically and
antigenically distinct molecule. We also present evidence that the
percentage of core oligosaccharides carrying the O-specific polymer is

less than 8%.

MATERIALS AND METHODS

Bacterial Strains

 

P. aeruginosa strain 261 was a mutant derived from strain Pae
K799/wt selected for antibiotic supersusceptibility; and PAZ1 met-28,
tip-6, _y§A-12, his-A, 113-226, and absA was a PA0222 derivative into
which the absA mutant gene from Z61 encoding antibiotic sensitivity has
been conjugated (2). Strains PAO1716 age—136, leg-8, rig-1 (revertant)
and PAO1715 age-136, 322'8' [if-1, tglA-12 (an aminoglycoside super-
sensitive mutant) were described previously (A1). Strain PA0503 met-
9011 is a methionine auxotroph of P. aeruginosa PA01. Strains PAZ1,

1715, 1716, and 503 were 0-5 serotype. Escherichia coli strains D21 and

 

D21f2 are derived from strain K-12 and were characterized as a Ra and Re

chemotype, respectively (A).

Growth Media

 

Strains PAO1716, PAO1715, 261 and PAZ1 were grown at 37°C to
mid-logrithmic phase in a 100 l fermentor containing 80 A of protease
peptone no. 2 medium from a 1 l overnight culture grown in the same
medium. Strain 503 was grown as previously described (5).

E. coli strains D21 and D21f2, grown as described by Coughlin gt

21° (11), were harvested in late log phase.

Isolation of LPS

 

LPS from P. aeruginosa strain 261 and PA01 derivatives 1715, 1716,
PAZ1, and 503 were isolated by the method of Darveau and Hancock (13),

followed by two extractions in chloroformzmethanol (1:1 v/v) resulting

37

38

in recovery of approximately 80% of the total LPS. The LPS from E. coli
strains 021 and D21f2 was isolated using the hot aqueous phenol (57) and

the chloroform-petroleum ether (17) extraction procedures, respectively.

SDS-PAGE

Sodium dodecyl sulfate-polyacrylamide (SDS-PAGE) gels were prepared
and run using the buffer system of Laemmli (36). Unless otherwise
noted, separating gels were formed with 15% acrylamide, 0.1% SDS, with a
7.5% acrylamide stacking gel. Samples were mixed 1:1 with sample buffer
(containing A% SDS) and applied to the gel. Electrophoresis was
performed with a constant current of 15 mA per gel until the tracking
dye entered the separating gel and then at 30 mA per gel until the
tracking dye reached the bottom of the gel. LPS bands were detected by

the silver staining method of Dubray and Bezard (16).

Column Chromatography

 

Samples were fractionated with a Sephadex G-200 (Pharmacia Fine
Chemicals) column (6A cm by 25 mm) at room temperature using the column
buffer system of Peterson and McGroarty (A8). Approximately 30 mg of
LPS was applied to the column, and 5 ml fractions were collected at a
flow rate of 8 ml per h. To remove detergent and buffer, pooled
fractions were extensively dialyzed (12,000 to 1A,OOO molecular weight
cutoff membranes) against column buffer without deoxycholate at 37°C and
then against distilled water at A°C. The dialyzed fractions were
lyophilized and resuspended to a concentration of 10 mg/ml in water.

All fractionations were done at least twice.

39

Western Blots

 

Western blots of SDS-polyacrylamide gels were prepared as
previously described (6,55). The gels were electrotransferred with a
model TE Transphor Electrophoresis apparatus (Hoefer Scientific
Instruments) at a constant current of 150 mA for 18 h using the
electrode buffer described by Otten and co-workers (A5) unless otherwise
noted. The nitrocellulose blots were visualized as described previously
(A5) with monoclonal anti-503 antibody (20, titer - 1:100,000) diluted
1:10,000 in blocking solution. In addition, dot blots were performed by
applying known quantities of LPS isolates directly on nitrocellulose.
The blots were washed and visualized using horse radish peroxidase
conjugated goat anti-mouse IgG antibody (Sigma Chemical Co.) as
described above or using a silver stain protocol identical to that

described by Dubray and Bezard (16).

Assays

Assays for amino sugars (18), heptose (60), and 2-keto-3-
deoxyoctulosonic acid (KDO) (29) were performed as described previously
except where noted. Phosphate analysis were performed using either a
colorimetric assay (1) or by inductively plasma emission spectroscopy
(11). Protein concentrations were estimated with the Pierce BCA protein
assay (Pierce Chemical 00., Rockford, IL) using bovine serum albumin as

a standard.

RESULTS

Silver staining of LPS from strains of P. aeruginosa separated by
SDS-PAGE revealed a progressive ladder-like pattern of bands up the gel

(Fig. 1A). For the Enterobacteriaceae and P. aeruginosa, these bands

 

have been reported to represent LPS molecules containing increasing
lengths of O-antigen (5,21,A6). The intensity of staining indicated
three to four regions of bands representing as many as four populations
of molecules (band sets 1, 2, 2a and 3, Fig. 1) differing in O-antigen
length. The electrophoretic pattern of LPS from strain 503 showed a set
of bands (set 2a) which were closely spaced and slower migrating bands
immediately above the 2a set which had greater spacing (Fig. 1A, lane 2).
In the banding pattern of LPS from PA01 derivatives 1715, 1716, and PA21
(Fig. 1A, lanes 1, 3, and 5, respectively), we observed irregularities
in the spacing and intensities of bands up the gel. In contrast, LPS
from strain 261 appeared to have a regular spacing and intensity in the
banding pattern (Fig. 1A, lane A). The average length of the highest
molecular weight LPS of strains 261 and PAZ1 seemed shorter than that of
the other PA01 derivatives, a phenomenon observed previously and
ascribed to the absA (antibiotic supersusceptibility) mutant locus (2).
When low amounts of LPS were applied to the gel, only the fastest
migrating bands were stained, and there was no difference in migration
pattern of LPS of strains 1715, 503, 1716, and PAZ1 (Fig. 1A, lanes 8,
9, 10, and 12, respectively). On the other hand, the low molecular
weight bands from strain Z61 migrated faster than that of the other P.

aeruginosa strains (Fig. 1A, lane 11) as previously observed (35), due

 

to an apparent truncation in the rough core of the short chain LPS

40

Figure 1.

41

(A) Silver-stained SDS-polyacrylamide gel of LPS from P.

aeruginosa strains 1715 (lanes 1 and 8), 503 (lanes 2 and 9).

 

1716 (lanes 3 and 10), Z61 (lanes A and 11), PA21 (lanes 5
and 12), and from E. coli strain D21 (Ra, lanes 6 and 13) and
D21f2 (Re, lanes 7 and 1A). Samples of either 5 pg (lanes 1
to 7) or 0.1 ug (lanes 8 to 1A) were applied to a 15%
acrylamide gel which had been polymerized overnight with a
butanol overlay. Arrows indicate the four intensively

stained regions of the P. aeruginosa samples: band sets 1,

 

2, 2a and 3. (B) Western blots of LPS from P. aeruginosa

 

strains 1715 (lane 15), 1716 (lane 16), 503 (lane 17), and
PAZ1 (lane 18) reacted with monoclonal anti-503 antibody.
Samples of 2.5 pg were applied to a 12% acrylamide gel which
had been polymerized overnight with a butanol overlay. The

gel was blotted as described in Materials and Methods.

 

N

2
4 .

, . _ Am
3 7
‘AF

9.29m. ¢_m_N_:O_m w hwmvm N_
m

<

43

molecules in this strain. LPS from P. coli D21 (Ra chemotype) and D21f2
(Re chemotype) was used to compare and characterize the electrophoretic
mobilities of the short chain populations.

Using antibodies specific to the PA01 O-antigen, Western blots of
the LPS separated with SDS-PAGE were analyzed to help clarify the
irregularities in the banding pattern. The blots of LPS isolated from

P. aeruginosa strains 1715, 1716, 503 and PA21 revealed a ladder pattern

 

of molecules that consisted of doublet bands (Fig. 18). Furthermore,
the level of one of the bands in the doublet was in lower amounts in the
isolates from strains 1715, 1716, and PAZ1 than from strain 503.
Presumably, this reflects a difference in substoichiometric modification
within the core-lipid A region of the molecules. Interestingly, spacing
and intensity of the ladder pattern seen in the Western blot was much
more regular than that of the silver stained gel. This suggested that
the stained sample may contain bands superimposed on the ladder pattern
of the main antigen.

If the irregular silver-stained banding pattern was a result of
heterogeneity in the LPS samples, this heterogeneity could have been due
to contamination of the culture or to true heterogeneity of the sample.
The possibility that the cultures were contaminated is very low since
after growth, all cultures were streaked out onto protease peptone no. 2
agar plates to observe characteristic colonial morphologies and
pigmentation and, in the cases of Z61, PA21 and PAO1715, tested for
characteristic antibiotic supersusceptibilities. Furthermore, the
irregular banding patterns were seen in samples from PA01 derivatives
(Fig. 1A) isolated in two different laboratories and from several

independently isolated LPS samples.

44

To further characterize the heterogeneity of the LPS isolates from
the PA01 strains, the samples were separated on a Sephadex G200 column.
The elution profile showed three major amino sugar-containing peaks for
strains 503 (Fig. 2), 1715 (Fig. 3), 1716 and PAZ1 (results not shown).
In contrast, the elution profile of the LPS sample from the Z61 strain
showed only two major peaks (Fig. A). Both gel permeation
chromatography and SDS-PAGE separate molecules on the basis of size;
therefore, a diagonal banding pattern should be expected across SDS-PAGE
gels of column fractions when applied in the order of elution. SDS-PAGE
of the column fractions of samples from each of the PA01 derivatives
studied revealed two distinct ladder patterns of apparently different
sizes, the A bands (later eluting ladder) and the B bands (earlier
eluting ladder) (see Figs. 2 and 3 for elution profiles of strain 503
and 1715, respectively). In contrast, the LPS isolate from strain Z61
showed only one ladder set (Fig. A). Peterson and McGroarty (A8)
reported that the SDS-PAGE profiles of fractionated LPS from Salmonella
species separated in this type of column as molecules of decreasing size.
The presence of two distinct ladder patterns suggests that either stable
aggregates were present (A8) or there were two types of molecules with
different charges; electrophoresis separates molecules on the basis of
both size and charge. To eliminate the possibility that A bands were
aggregates of the B bands stabilized by hydrogen bonding between
O-polymers, SDS-PAGE of the column fractions of samples from strain 503
was run in the presence of A M urea (final concentration). The same two
sets of bands were observed in AM urea-SDS gels (results not shown). In
addition, two-dimensional electrophoresis of column fractions of LPS

from strains 503 and 1716, which contained approximately equal amounts

Figure 2.

45

Fractionation of LPS from P. aeruginosa strain 503 on

 

Sephadex G-200. Fractions were analyzed for KDO (o) and
amino sugar (A). Silver-stained SDS-polyacrylamide gels of
column fractions are aligned under their appropriate fraction
number. A represents the slow-moving set of bands and B the

faster-moving set.

46

 

0.6

0.5

0.3 -

0.2 -

0.l -

A54e ' °
0
A
l

 

 

 

 

 

2
A
A
20
‘ ‘ x X
‘ A
A A A
AA. - A ‘1- "
20 40
3’ Fraction Number
1 ”I“... -‘- A’ v.
Olflofir.'

‘*

 

0.6

Figure 3.

47

Fractionation of LPS from P. aeruginosa strain 1715 on

 

Sephadex G-200. Fractions were analyzed for KDO (o) and
amino sugar (A). Silver-stained SDS-polyacrylamide gels of
column fractions are aligned under their appropriate fraction
number. A represents the slow-moving set of bands and B the

faster-moving set.

 

0.6 r—

 

05 -

A548 H
.0
A
l

}

 

0.3 -

D

0.2 .—

D,

 

 

 

 

" -\
2o

 

 

20 40 60

B F rociion Number
. H
"Ila..." A; “cup

"Z_r:* 029 V

 

‘ ..‘_‘__A_L_A._Ag‘ 0..-. .

49

Figure A. Fractionation of LPS from P. aeruginosa strain 261 on

 

Sephadex G-200. Fractions were analyzed for KDO (O) and
amino sugar (A). Silver-stained SDS-polyacrylamide gels of
column fractions are aligned under their appropriate fraction

number.

A54e' ’

0.6

0.5

0.4

0.3

0.2

0.1

50

 

 

 

 

 

 

Viv

 

Fraction

Number

in...

0.3

0.2

01

.OEQ V

51

of A and B bands, showed that both the A and B bands ran with the same
mobilities in the second dimension. Since the gels in both dimensions
consisted of standard SDS gels and the gel strip from the first
dimension was heated before running the second dimension, the results
indicate that the bands do not interconvert (results not shown). This
again indicates that one set of bands is not an aggregate or different
conformational state of the other. Amino sugar analysis of our column
fractions indicated that the fractions containing mainly A bands (peak
X, Fig. 3) lacked reactive amino sugars, while the other fractions
(peaks 1, 2, 2a, and 3; Figs. 2 and 3) showed reactivity.

The individual fractions from the elution profile of each of the
LPS samples separated in a Sephadex G-200 column were also analyzed for
KDO and phosphate content, and the relative molar ratios of amino sugar
to K00 and amino sugar to phosphate were determined for the 3 or A major
peaks. Also, the fractions corresponding to each of the major peaks in
the elution profile of strains 1715 and PA21 were pooled, dialyzed,
lyophilized and resuspended in water to a final concentration of 10
mg/ml and were analyzed as above. The data shown in Table 1 for the
pooled fractions and data from the individual fractions (not shown)
indicated that, using phosphate as an indicator of molar amounts of the
LPS, the short chain peak 3 sample represented 92 to 97% of the total
LPS molecules; peak 2a, 0.A to 2%; peak 2, 1 to 3%; and the very long
chain fraction, peak 1, 3 to A%. The fifth population of molecules, the
A bands, were resolved as a separate population on the column (peak X)
in the samples from 1715 (Fig. 3), 1716, and PAZ1, whereas A bands
overlapped with the 2a peak of the main ladder set in the 503

fractionation (Fig. 2). As stated above, strain 261 showed only two

52

Table 1. Chemical Analysis of P. aeruginosa B-Band LPS Fractions

 

 

s P Amino Amino P/KDO

Peak No. Recovereda Sugar/KDOb Sugar/Pc (mole/mole)
Strain Z61d

2 2.8 5.5 9.3 N.D.e

3 97.0 1.0 1 0 5.5
Strain PAZ1f

1 2.3 2u.8 30.8 u.8

2 3.2 19.5 18 8 6.1

3 9A.3 1 O .O 5.9
Strain 1715f

1 3.A 29.3 31.A 5.A

2 1.A 17.0 18.2 5.A

3 95.1 1.0 1.0 6.0

 

éPercentage of total amounts in each of the peaks.

bRelative molar ratio; K00 and amino sugar levels were normalized to a
value of 1.0 for peak 3 samples.

CRelative molar ratios; phosphate levels were determined for the
individual fractions of the 261 sample by using the colorimetric assay
and for the pooled fractions from samples of strains 1715 and PAZ1 by
inductively coupled plasma emission spectroscopy, as described in
Materials and Methods. Levels were normalized to a value of 1.0 for
peak 3 samples.

dIndividual fractions from the elution profile of the LPS samples
separated on a Sephadex G-200 column, as described in Material and
Methods, were analyzed and the amounts in all the fractions in each
peak were added together.

eN.D., not determined.
fPooled fractions corresponding to each of the peaks in the elution

profile of LPS samples separated on a Sephadex G-200 column were
dialyzed, lyophilized, and resuspended in distilled water for analysis.

53

different LPS size populations, the short chain fraction, peak 3 (97% of
total), and the long chain fraction, peak 2 (2.8% of total) (Fig. A).

The O-antigen of P. aeruginosa is reportedly rich in amino sugars
(3A,35,AO,58), although the reported sequence of the 05 serogroup
(equivalent to Lanyi type 3a,d) indicates that most of the amino groups
are acetylated (30). The phosphate and KDO assays measure residues in
the core region of LPS (35), whereas the assay for amino sugars detects
residues both in the core-lipid A region and O-antigen repeat region.
The data reported in Table 1 show that the molar ratios of PzKDO for the
three major B band-containing peaks of the pooled fractions are similar
for strains PA21 and 1715. Also, this ratio for peak 3, calculated from
the individual fractions for all strains, appeared to be relatively
constant from strain-to-strain and was determined to be between 5.5 and
6.6 (mole/mole, data not shown). Since the level of phosphate or KDO
can be used as a measure of the relative molar amount of LPS in each
peak, comparison of the ratios of amino sugar to KDO and/or amino sugar
to phosphate in the three major peaks of the pooled fractions should
reflect the O-antigen length. For the individual fractions a
discrepancy existed in these ratios where the values are lower than for
the pooled fractions. This may be due to the elution buffer present in
the fractions before dialysis. Since the pooled fractions were dialyzed
to remove the detergent and salts and resuspended in H20 to a known
concentration the pooled-fraction data were more reliable. Furthermore,
the low amount of sample in the individual fractions decreased the
sensitivity of the assays.

The chemical composition of the dialyzed A band samples from

strains 1715 and PA21 was analyzed, and the results were compared with

54

those of the three major amino sugar-containing peaks (Table 2). The
results indicated that peak X (A bands) comprised 10-15% of the total
LPS sample by weight. Furthermore, under the normal conditions of the
assay, it appeared to have low levels of reactive KDO, amino sugars and
phosphate. When the pooled samples were hydrolyzed for a longer period
of time (20 min) and at a higher acid concentrations (0.5 N HZSOu), the
K00 levels detected in the B band fractions were similar to previous
assays, but the KDO levels in the A band fraction increased 10-fold.
Thus, the A band material appeared to contain KDO residues which are
much less reactive than those in the B band isolates. Analysis of the
isolates from strain 1715 indicated that the A band material had levels
of heptose inbetween that found in peaks 2 and 3 (Table 2). Analysis
for protein in the pooled fractions indicated less than 1.5% (W/W) in
all fractions. From the SDS-PAGE of the separated fractions (e.g. Fig.
3) and from the chemical characterization of the pooled samples, it was
observed that the A bands of strains 1715 and PAZ1 were contaminated
with only minor amounts of B band material. The A band material is
probably a glycolipid since the isolate, suspended in the absence of
detergent at 10 mg/ml, was cloudy and not completely water soluble, and
since it contained heptose and KDO.

To better compare the migration patterns of equivalent LPS
populations from different strains, comparable column fractions of LPS
containing both A and B bands from the different strains were run
together on SDS-PAGE (Fig. 5). The A bands from all four PA01 strains
had very similar spacings (lanes 2-5), and the spacings of the B bands
of strain PAZ1, 1716, and 1715 (Fig. 5, lanes 2, A and 5, respectively)

appeared very similar. In contrast, the B bands of strain 503 (Fig. 5,

55

.om:#ELoumo go: ..a.zw

.mflnmofifiaam yo: 742.2.H

.m>L:o ucmccmpm ecu com com: mm: omouqocoHsmionoLmozamram

.Acoapmgucoocoo Hmcfiuv HE\mE P am maaemm m Lo pmcu o» copommgoo mm: 5: 0mm pm :ofiuqcomo<v

.mcocuo: ocm mHmHLoumz
:3 cmnwLowmo mm .xaoowocuomam cowmm_eo memqu mafiasoo >Hm>fiuosccfi >2 umcfiscmumu ego: mao>ma oumcamocmo

.HOOOSOLQ ccmccmum on» mcfims nmammmm mm: oaxn

.mfimzamcm Lou

flaxwe op no codpmgucoocoo m o» Loom: uoHHdumdu :fi coocoamsmog one: moaasmm one .uocmfimz ocm vowfiafinaoza
coca ucm .Lmumz ooHHHomwo zo umzoafiou oumaozomxoou usocufiz Louuzo casaoo umcfimwm commamfiu .umHooa
omo: .mma ho maqsmm copmcoduomgmc: cm cow: mcon .x xmmq ocm .m xmma .m xmma .P xmoa ho mcouuommm seafloom

 

 

 

 

 

 

 

mHaemm was
.o.z .a.z ~.m .o.z comp m.o.z NOF .<.z m.<.z omumcodpomcpcz
mm m.o 0.0 a s m m m— _P x
msm m.— m._ oco_ o:>_ New mom mm mm m
:m ®.m m.m oo_ on? Fm om mp op m
mm :.m m.: on? om. Pm mm 2F mm _
m_>_ PN<m mpspi Pm<a mps_ PN<m mps_ Pm<m mph.
mag me\moHOE: oHQEmm Hs\me\cmm< mag waxmoaoec mag me\moHoec cmcm>ooom Hmuoe u m.oz xwom
mmmouqo: cmcmmsm ocme< omumzamocm coax mm ucmfioz

—N<m vcm mph. mcfimcum mmocfimsaom .M song was go mcofipomcm cesHoo UmHoom ocp no cofiafimoaeoo .m mama?

Figure 5.

56

Silver-stained SDS-PAGE (11% acrylamide) (lanes 1 to 5) and
Western blots (lanes 6 to 10) of LPS fractions from P.

aeruginosa reacted with monoclonal anti-503 LPS antibody.

 

Lanes: 1 and 6, LPS from strain 261 (fraction 36); 2 and 7,
strain PAZ1 (fraction AA); 3 and 8, strain 503 (fraction A0);
A and 9, strain 1716 (fraction A2); 5 and 10, strain 1715
(fraction A8). Samples of 100 pl were applied to the SDS-
polyacrylamide gel, which had been polymerized overnight with
a butanol overlay. The gel was blotted as described in
Materials and Methods. Arrows indicate the A and B bands of

the respective fractions.

57

 

58

lane 3) appeared much more closely spaced compared to the same bands of
the other strains due to an increased amount of the second band in the B
band set of doublets (see also Fig. 1B, lane 17). Fractions from peak 3
of LPS from strains Z61, 503, PAZ1, 1715, and 1716 were run on an 18%
SDS-acrylamide gel. No differences were seen in band mobilities of the
short chain isolates except for that from strain Z61 which migrated
faster (results not shown), corroborating the results with
unfractionated LPS (Fig. 1A, lanes 8-12).

To determine the antigenic reactivity of the A and B bands, the LPS
fractions containing approximately equal amounts of A and B bands were
subjected to Western immunoblotting. Figure 5 (lanes 6-10) shows a
Western blot of these LPS fractions from strains PAZ1, 503, 1716, and
1715 and a fraction of Z61 whose LPS had an electrophoretic mobility
similar to that of the A bands of the other four samples. Reaction with
monoclonal anti-503 LPS antibody (specific for the O-antigen side chain)
indicated that only the B bands were antigenically reactive and not the
A bands. There was a weak antigenic reactivity of closely spaced bands
above the region corresponding to A bands which presumably was due to
reaction with aggregated B band-type LPS (Fig. 5, lanes 7-10). Peterson
and McGroarty (A8) have shown that the high-molecular weight fractions
of LPS from E. coli can migrate as multimers. As expected, the bands
from strain Z61 also showed no reactivity with the antibody since it is
known that the LPS from strain 261 is antigenically distinct from that
of PA01 derivatives (2).

To show that the lack of immunoreactivity of the A bands was not
due to lack of transfer or recovery of samples on the nitrocellulose,

Western blots were performed under different conditions. We used 150

59

and A00 mAmp and also transferred for 2A h. The gels were stained for
LPS with and without transblotting to determine the level of
electrotransfer of the different bands. The results indicated that, at
the lower voltage, A bands did not transfer as well as B bands,
consistent with their mobility on the gel. However, at the higher
transblotting voltage where A band-type of molecules were removed from
the gel no immunoreactivity was detected on the nitrocellulose (data not
shown).

We also performed a dot blot immunoassay using the pooled column
fractions to characterize differences in the antigenic reactivity of the
two types of molecules. The four pooled fractions of LPS from strain
1716 (peaks 1, 2, 3 and X) were dialyzed, lyophilized and resuspended in
water. A 10 pg portion of each of the fractions was spotted along the
top of a nitrocellulose strip. Subsequent rows were spotted with the
same volume of sample serially diluted 10-fold each row. The spots were
dried and visualized with the anti-503 monoclonal antibody as described
for the Western blots. The results (Fig. 6A) indicated that the very
long chain and long chain populations of B bands had similar reactivity,
and as little as 1 ng could be detected. In contrast, the short chain
population, containing no or only one O-repeat unit per molecule, showed
no reactivity. Interestingly, the A band sample showed weak reactivity
when 10 pg and 1 pg were applied, but no reactivity with lower amounts.
The reactivity seen with 10 pg of the A bands was similar to that seen
with 10 ng of long chain and very long chain fractions, suggesting that
the A band fraction was contaminated with < 1% (w/w) of B band-type
molecules. That the A band fraction remained bound to the nitrocellulose

was shown by staining the dot blot with the silver stain used for the

Figure 6.

6O

Dot blots of pooled column fractions from P. aeruginosa 1715

 

LPS. Peaks 1, 2, 3 and X were separated on Sephadex G-200,
dialyzed, and lyophilized as described in Materials and
Methods and suspended in distilled water to a final
concentration of 10 mg/ml. In the top row, 10 pg of each
sample was applied on the nitrocellulose strip, and
subsequent rows were spotted with an equal volume of 10-fold
serially diluted samples. The sample spots were dried, and
the nitrocellulose was either reacted with monoclonal
anti-503 antibody and developed as described in Materials and
Methods (A) or washed with 10 mM Tris, 150 mM NaCl (pH 7.0)
and reacted with the silver stain as indicated in the

Material and Methods for polyacrylamide gels (B).

61

 

'- '01“;
s—Ing

fi-OJpg

<—|Ong

<—|ng

62

polyacrylamide gels. We found the intensity of staining indicated the
approximate molar amount of material applied (Fig. 6B). The levels
detected were not changed if the blots were washed extensively with
buffer or the blocking solutions used in the Western blots prior to

staining (data not shown).

DISCUSSION

The LPS isolates of P. aeruginosa strains 503, PAZ1, 1715, 1716,
and 261 were separated by SDS-PAGE into as many as four major size
populations. The ladder-like banding patterns represent molecules with
increasing numbers of O-antigen repeat units (21,25,A6). Other
investigators have reported similar heterogeneity in the LPS from
strains 503 and Z61, as well as other P. aeruginosa smooth strains
(5,25,3A,35). An O-antigen ladder pattern has recently been described
for several smooth strains of P. aeruginosa using silver staining and
immunoblotting techniques (5“). Our Western blot of unfractionated LPS
(Fig. 1B) revealed a ladder-like banding pattern with regular spacing.
The bands were resolved as doublets (Fig. 1B, lanes 15-18) suggesting
substoichiometric modification in the core-lipid A similar to that seen

with Salmonella LPS (A3,56). The irregular spacing that we observed in

 

the SDS-PAGE silver-stained ladder pattern of LPS from strains 503,
PAZ1, 1715, and 1716 (Fig. 1A) suggested the possibility that PA01
derivatives may be producing LPS with more than one type of O-polymers.

Peterson and McGroarty (A8), using strains of Enterobacteriaceae,

 

have shown that LPS molecules of different sizes can be partially
separated with Sephadex G-200 in the presence of deoxycholate and EDTA.
Two or three major populations of LPS could be resolved as detected by
sugar analysis. These populations represent sets of molecules with
O-antigens of different lengths which are made in high amounts. Other
investigators have also demonstrated, using gel permeation
chromatography in combination with other methods, that the LPS from the

Enterobacteriaceae and other gram-negative bacteria could be resolved

64

into at least two main populations of LPS differing in the length of
their O-polysaccharide chain (10.27.31.33.35). In the results presented
here, we found that the LPS from strains 503 and 1715 (Figs. 2 and 3,
respectively), as well as strains 1716 and PA21, separated into three
major populations (peaks 1, 2, and 3) and two minor populations (peaks
2a and X). Interestingly, the short chain population comprised more
than 90% of the total sample on a molar basis (Table 1). In addition,
there were greater amounts of the very long chain population from the
PA01 strains compared to the long and intermediate chain populations.
These results agree with those of Wilkinson (58) and Hancock 3; 21'
(2A), where they estimated the mole percent of S-form LPS to be between

0.2-1A%. On the other hand, the Enterobacteriaceae show a distribution

 

of AA-60% of the LPS molecules in the low molecular weight population
and 30-50% in the high molecular weight fractions (37,A8). It has been

demonstrated that S. typhimurium synthesizes LPS molecules with over 80

 

O-antigen repeating units and this population constitutes about 6% of
the total LPS sample (A8). Since the hydrophilic O-polysaccharides
extend from bacterial surface into the aqueous environment, the observed
heterogeneity of O-chain lengths suggests that the surface topography of
the gram-negative bacteria is irregular and that accessibility of the
lipid A head group of the LPS could vary in different regions on the
surface and on different bacterial species. It has been shown that the
presence of O-antigen-containing LPS influences various cell surface
phenomena, including antibiotic binding to LPS (A9). antibiotic
susceptibility (2,5,19), LPS aggregate structure (A9), bacteriOphage
recognition (26,28), immunochemical characterization (9.10.50).

virulence (12,51). protection against the bactericidal action of serum

65

(22,A2), polyclonal B cell activation and macrophage cytotoxicity (AA).
The low level of LPS on P. aeruginosa that contains a long O-polymer,
however, may be sufficient to form a uniform cover over the cell since
the surface is inaccessible to rough core specific monoclonal antibodies
(53).

A striking feature of PA01 strains from P. aeruginosa is the
presence of the A bands which constitute a significant amount of the
isolated LPS (Table 2). This set of bands was observed as a slow moving
diagonal banding pattern across the SDS-PAGE of the column fractions
(peak X, Figs. 2 and 3). LPS from strain 261 contained only a single
diagonal banding pattern corresponding to the bands in peaks 2 and 3,
(Fig. A). Although SDS-PAGE separates LPS molecules according to size
(27,A6,A8). we propose that the anomalous migration in SDS-PAGE of the A
bands is due to a difference in the charge in the core-lipid A region of
the molecules. The lack of phosphate substituents in the A band sample
would make the molecules much less negatively charged than the B band
fractions which are high in phosphate groups. To further explore this
possibility, we compared the mobility of the A and B bands of the PA01
strains on an 11% SDS-PAGE (Fig. 5). We observed a difference in the
spacing between the A and B bands, and that the B bands consisted of
sets of doublet bands in which one of the two differ in their staining
intensity (Figs. 18 and 5). while the A bands seemed to lack this
doublet pattern. This doublet probably represents substoichiometric
modification in the core or lipid A of the B band components. It has
been reported for different gram-negative bacteria that there is
microheterogeneity in the structure of lipid A (phosphate levels and

types and numbers of fatty acids) (3.52.58), in the substituents in the

66

core (15.23,37-39.58). and in modifications of the O-antigen side chain
(37.38.58). The heterogeneity of LPS molecules presumably depends, in
part, on the strain and on growth conditions (1A.37).

We have presented several pieces of evidence which indicate that
the molecules represented by the A and B bands are chemically distinct.
Only the B bands from the PA01 strains reacted with anti-O-antigen
antibodies in Western blots (Fig. 5). Also. the A bands lacked reactive
amino sugars detected in the B band fractions. Since the pooled
fractions corresponding to A bands (peak X) contained very low levels of
reactive KDO unless hydrolyzed with high concentrations of acid, and no
phosphate or amino sugar was detected (Table 2). there is the
possibility that these A bands represent another type of molecule
different from LPS, that is, an O-repeat attached to a molecule which is
not lipid A. In the past 10 years the chemical structure of the lipid

A's from gram-negative bacteria other than the Enterobacteriaceae have

 

been studied, and the existence of "unusual" lipid A's has been noted

(A0). For instance, the lipid A from Pseudomonas paucimobilis contains

 

a number of sugars, in addition to glucosamine, in the "bound lipid"
fraction, and phosphate as well as KDO appears to be lacking (A0).
Also, species of Thermus have been reported to make LPS that lacks
detectable heptose, KDO, glucosamine, and phosphorus (A0). However,
negative reactivity in the thiobarbiturate assay may not reflect the
lack of KDO residues. Recently, Parr and Bryan (A7) demonstrated that
more rigorous hydrolysis conditions were required to release KDO from

Haemophilus influenza LPS compared to other LPS species. Subsequently,

 

Carof £3 21' (8) demonstrated that after treatment with aqueous

hydrofluoric acid. the presence of KDO could readily be demonstrated in

67

LPS of Bortedellas, Bacteroides, Aeromonas. and Vibrios which had been

 

 

reported to be KDO-deficient. Thus, we hydrolyzed our A band sample
with a higher acid concentration for longer times and observed a 10-fold
increase in the KDO level. This suggests that the A bands are resistant
to hydrolysis due to substitutions of the KDO units in position A and 5.
or 5 and 7 (7). Other evidence that suggests that the A bands represent
LPS molecules is the ability to silver stain the A band molecules;
Kropinski 22 21° (32) and Lam (personal communication) reported that the

silver staining reaction for P. aeruginosa LPS occurs in the lipid A

 

rather than the core sugars.

Analysis of the Western and dot blots (Fig. 5 and 6) yielded
several additional interesting observations. The pooled fractions
containing the higher molecular weight LPS (peaks 1 and 2) from the PA01
derivatives reacted with the antibody (Fig. 6), indicating that the B

bands comprise the serotype-specific LPS. However. P. aeruginosa

 

synthesizes a significant amount of A band type molecules (Table 2) with
presumably a different antigenicity. It has been shown that S.

paratyphi B and S. typhimurium can synthesize a T1 polysaccharide and an

 

O-polysaccharide attached to the same care (38,A2), and that the

synthesis of the two molecules is independent (38). Other P. aeruginosa

 

strains may also have the ability to synthesize more than one type of
LPS with different O-antigenic side chains (9,31,58,59). The high
molecular weight polysaccharide released from LPS of these strains has
been resolved into amino sugar-rich and neutral sugar-rich fractions
when separated by gel permeation chromatography (31.58.59). And

finally, Carof 33.31. (7) have shown that Bortedella pertussis produces

 

two lipopolysaccharides. one of which does not give a positive reaction

68

for KDO under normal thiobarbituric assay conditions. This is very
similar to our findings for the A and B band-type LPS of the PA01
derivatives.

The effect of the presence of a unique, A band-type of LPS on the
physical interactions within the outer membrane, as well as the
immunological reaction with the cell surface. may be important. It has
been proposed that bacteria synthesizing O-side chains of an unusual
structure might escape the immune system of the host, which might have
difficulties in producing effective antibody molecules against these
O-side chains (A2,51).

In summary, our data suggest that PA01 strains from P. aeruginosa

 

are capable of synthesizing more than one type of LPS-like molecule
differing in their antigenic reactivities. Although the LPS isolated
from the different strains used in this study was shown to be
heterogeneous on SDS-PAGE, this method by itself does not have the power
to predict the presence of more than one type of LPS-like molecule with
a different O-polymer. However, by combining SDS-PAGE with gel
permeation chromatography, Western blots, and sugar analysis, we have
been able to distinguish chemically distinct subclasses of molecules
from individual strains. We have also demonstrated that the percentage

of core oligosaccharides carrying the O-specific polymer is less than 8%.

REFERENCES

Ames, B.N., and D.T. Dubin. 1960. The role of polyamines in the
neutralization of bacteriophage deoxyribonucleic acid. J. Biol.
Chem. 235:769-775.

Angus. B.L.. J.A.M. Fyfe, and R.E.W. Hancock. 1987. Mapping and
characterization of two mutations to antibiotic supersusceptibility

in Pseudomonas aeruginosa. J. Gen. Microbiol. (in press).

 

Baltzer, L.H., and I. Mattsby-Baltzer. 1986. Heterogeneity of
lipid A: structural determination by 13c and 31? NMR of lipid A

fractions from lipopolysaccharide of Escherichia coli O111.

 

Biochemistry 25:3570-3575.

Boman. H.G.. and D.A. Manner. 1975. Characterization of

‘lipopolysaccharides from Escherichia coli K-12 mutants. J.

 

Bacterial. 121:A55-A6A.
Bryan, L.E.. K. O'Hara, and S. Wang. 198A. Lipopolysaccharide
changes in impermeability-type aminoglycoside resistance in

Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 26:250-255.

 

Burnette. W.N. 1981. "Western blotting": electrophoretic
transfer of proteins from sodium dodecyl sulfate-polyacrylamide
gels to unmodified nitrocellulose and radiographic detection with
antibody and radioiodinated Protein A. Anal. Biochem. 112:195-203.
Carof, M., S. Lebbar. and L. Szabo. 1987. Detection of
3-deoxy-2-octulosonic acid in thiobarbiturate-negative endotoxins.

Carbohyd. Res. 161:CA-C7.

69

10.

11.

12.

13.

1A.

15.

16.

7O

Carof, M., S. Lebbar. and L. Szabo. 1987. Do endotoxins devoid of
3-deoxy-D-manno-2-octulosonic acid exist? Biochem. Biophys. Res.
Commun. 1A3:8A5-8A7.

Chester, I.R., and P.M. Meadow. 1973. The relationship between
the O-antigenic lipopolysaccharides and serological specificity in

strains of Pseudomonas aeruginosa of different O-serotypes. J.

 

Gen. Microbiol. 78:305-318.
Chester, I.R., and P.M. Meadow. 1975. Heterogeneity of the

lipopolysaccharide from Pseudomonas aeruginosa. Eur. J. Biochem.

 

58:273-282.
Coughlin, R.T., S. Tonsager, and E.J. McGroarty. 1983.
Quantitation of metal cations bound to membranes and extracted

lipopolysaccharide of Escherichia coli. Biochemistry 22:2002-2007.

 

Cryz. Jr., S.J., T.L. Pitt, E. Ffirer, and R. Germanier. 198A.

Role of lipopolysaccharide in virulence of Pseudomonas aeruginosa.

 

Infect. Immun. AA:508-513.
Darveau. R.P., and R.E.W. Hancock. 1983. Procedure for isolation
of bacterial lipopolysaccharides from both smooth and rough

Pseudomonas aeruginosa and Salmonella typhimurium strains. J.

 

 

Bacterial. 155:831-838.
Day, D.P., and M.L. Marceau-Day. 1982. Lipopolysaccharide

variability in Pseudomonas aeruginosa. Curr. Microbial. 7:93-98.

 

Drdge, W., V. Lehmann, O. LUderitz, and O. Westphal. 1970.
Structural investigations on the 2-keto-3-deoxyoctonate region of
lipopolysaccharides. Eur. J. Biochem. 1A:175-18A.

Dubray, G.. and G. Bezard. 1982. A highly sensitive periodic

acid-silver stain for 1,2-diol groups of glycoproteins and

17.

18.

19.

20.

21.

22.

23.

71

polysaccharides in polyacrylamide gels. Anal. Biochem.
119:325-329.

Galanos, C.. O. Lflderitz and O. Westphal. 1969. A new method

for the extraction of R lipopolysaccharides. Eur. J. Biochem.
9:2A5’2A9.

Gatt, R., and E.R. Bermad. 1966. A rapid procedure for the
estimation of amino sugars on a microscale. Anal. Biochem.
15:167-171.

Godfrey, A.J., L. Hatlelid, and L.E. Bryan. 198A. Correlation
between lipopolysaccharide structure and permeability resistance in

B-lactam-resistant Pseudomonas aeruginosa. Antimicrob. Agents

 

Chemother. 26:181-186.
Godfrey, A.J., M.S. Shahrabadi. and L.E. Bryan. 1986.
Distribution of porin and lipopolysaccharide antigens on a

Pseudomonas aeruginosa permeability mutant. Antimicrob. Agents

 

Chemother. 30:802-805.
Goldman, R.C., and L. Leive. 1980. Heterogeneity of

antigenic-side chain length in lipopolysaccharide from Escherichia

 

coli 0111 and Salmonella typhimurium LT2. Eur. J. Biochem.

 

107:1A5-153.
Grossman, N., M. A. Schmetz, J. Foulds, E.N. Klima, V. Jiminez,
L.L. Leive, and K.A. Joiner. 1987. Lipopolysaccharide size and

distribution determine serum resistance in Salmonella montevideo.

 

J. Bacteriol. 169:856-863.
HAmmerling, 0.. V. Lehmann, and O. LUderitz. 1973.

Structural studies on the heptose region of Salmonella

 

lipopolysaccharides. Eur. J. Biochem. 38:A53-A58.

2A.

25.

26.

27.

28.

29.

30.

72

Hancock, R.E.W., L.M. Mutharia, L. Chan, R.P. Darveau. D.P. Speert,

and 0.8. Pier. 1983. Pseudomonas aeruginosa isolated from

 

patients with cystic fibrosis: A class of serum-sensitive.
nontypable strains deficient in lipopolysaccharide O-side chains.
Infect. Immun. A2:170-177.

Hitchcock, P.J.. and T.M. Brown. 1983. Morphological

heterogeneity among Salmonella lipopolysaccharide chemotypes in

 

silver-stained polyacrylamide gels. J. Bacteriol. 15A:269-277.

Ikeda. K., and F. Egami. 1973. Lipopolysaccharide of Pseudomonas

 

aeruginosa with special reference to pyocin R receptor activity.

 

J. Gen. Appl. Microbiol. 19:115-128.

Jann, B., K. Reske, and K. Jann. 1975. Heterogeneity of
lipopolysaccharides. Analysis of polysaccharide chain lengths by
sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Eur. J.
Biochem. 60:239-2A6.

Jarrell, K., and A.M. Kropinski. 1977. The chemical composition

of the lipopolysaccharide from Pseudomonas aeruginosa strain PAO

 

and a spontaneously derived rough mutant. Microbios. 19:103-116.
Karkhanis. Y.D.. J.Y. Zeltner, J.J. Jackson, and D.J. Carlo. 1978.
A new and improved microassay to determine 2-keto-3-deoxyoctonate
in lipopolysaccharide of gram-negative bacteria. Anal. Biochem.
85:595-601.

Knirel, Y.A., B.V. Vinogradov, A.S. Shashov, B.A. Dmitriev, N.K.
Kochetkov, E.S. Stanislavsky, and G.M. Mashilova. 1982. Somatic

antigens of Pseudomonas aeruginosa. The structure of O-specific

 

polysaccharide chains of P. aeruginosa O:3a.b and O:3a,d

 

lipopolysaccharide. Eur. J. Biochem. 128:81-90.

31.

32.

33.

3A.

35.

36.

37.

73

Koval. S.F., and P.M. Meadow. 1977. The isolation and
characterization of lipopolysaccharide-defective mutants of

Pseudomonas aeruginosa PAC 1. J. Gen. Microbiol. 98:387-398.

 

Kropinski, A.M., D. Berry, and E.P. Greenberg. 1986. The basis of
silver staining of bacterial lipopolysaccharides in polyacrylamide
gels. Cur. Microbial. 13:29-31.

Kropinski, A.M., L.G. Chan, and F.H. Milazzo. 1979. The

extraction and analysis of lipopolysaccharides from Pseudomonas

 

aeruginosa strain PAO, and three rough mutants. Can. J. Microbiol.

 

25:390-398.
Kropinski, A.M.. B. Jewell, J. Kuzio, F. Milazzo. and D. Berry.

1985. Structure and functions of Pseudomonas aeruginosa

 

lipopolysaccharide, p. 58-73. In D.P. Speert and R.E.W. Hancock

(eds.), Pseudomonas aeruginosa: New Therapeutic Approaches from

 

 

Basic Research. (Antibiotics and Chemotherapy; Vol. 36). Karger,

 

Basel.
Kropinski, A.M., J. Kuzio, B.L. Angus, and R.E.W. Hancock. 1982.
Chemical and chromatographic analysis of lipopolysaccharide from an

antibiotic-supersusceptible mutant of Pseudomonas aeruginosa.

 

Antimicrob. Agents Chemother. 21:310-319.

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

Lflderitz, 0., M.A. Freudenberg, C. Galanos, V. Lehmann, E.Th.

Rietschel. D.H. Shaw. 1982. Lipopolysaccharides of gram-negative

bacteria. Curr. Top. Membr. Transp. 17:79-151.

 

38.

39.

A0.

A1.

A2.

A3.

1111.

A5.

74

Makela, P.H., and B.A.D. Stacker. 198A. Genetics of
lipopolysaccharide, p. 59-137. In E.T. Rietschel (ed.), Handbook

23 Endotoxin, Vol. 1: Chemistry 9: Endotoxin. Elsevier Science

 

Publishers B.V.. Amsterdam.
Mayer, H., A.M.C. Rapin, G. Schmidt, and G. Boman. 1976.
Immunochemical studies on lipopolysaccharides from wild-type and

mutants of Escherichia coli K-12. Eur. J. Biochem. 66:357-368.

 

Mayer, H., and J. Weckesser. 198A. "Unusual" lipid A's:
structures, taxonomical relevance and potential value for endotoxin
research, p. 221-2A7. In E.T. Rietschel (ed.), Handbook 9:

Endotoxin, Vol. 1: Chemistry 9: Endotoxin. Elsevier Science

 

Publishers, B.V.. Amsterdam.

Mills, B.J., and B.W. Holloway. 1976. Mutants of Pseudomonas

 

aeruginosa that show specific hypersensitivity to aminoglycosides.

 

Antimicrob. Agents Chemother. 10:A11-A16.

Nikaido, H. 1970. Lipopolysaccharide in the taxonomy of
enterobacteriaceae. Int. J. Syst. Bacteriol. 20:383-AO6.

Nikaido, H., and M. Vaara. 1985. Molecular basis of bacterial
outer membrane permeability. Microbiol. Rev. A9:1-32.

Ohta, M., J. Rothmann. E. Kovats. P.H. Pham, and A. Nowotny. 1985.
Biological activities of lipopolysaccharide fractionated by
preparative acrylamide gel electrophoresis. Microbiol. Immunol.
29:1-12.

Otten, S.. S. Iyer, W. Johnson, and R. Montgomery. 1986.

Serospecific antigens of Legionella pneumophila. J. Bacteriol.

 

167:893’90A.

A6.

’47.

A8.

A9.

50.

51.

52.

53.

75

Palva. E.T., and P.H. Makela. 1980. Lipopolysaccharide

heterogeneity in Salmonella typhymurium analyzed by sodium dodecyl

 

sulfate/polyacrylamide gel electrophoresis. Eur. J. Biochem.
107:137-1A3.
Parr, T.R., and L.E. Bryan. 198A. Lipopolysaccharide composition

of three strains of Haemophilus influenzae. Can. J. Microbiol.

 

20:131-132.
Peterson, A.A.. and E.J. McGroarty. 1985. High-molecular-weight

components in lipopolysaccharides of Salmonella typhimurium,

 

Salmonella minnesota, and Escherichia coli. J. Bacteriol.

 

 

162:738-7A5.
Peterson, A.A., A. Haug, and E.J. McGroarty. 1986. Physical

properties of short- and long-O-antigen-containing fractions of

”lipopolysaccharides from Escherichia coli 0111:BA. J. Bacteriol.

 

165:116-122.
Pier, 0.8., M. Pollack, and M. Cohen. 198A. Immunochemical
characterization of high-molecular-weight polysaccharide from

Fisher immunotype 3 Pseudomonas aeruginosa. Infect. Immun.

 

A5:309-313.

Roantree, R.J. 1967. Salmonella O antigens and virulence. Ann.

 

Rev. Microbiol. 21:AA3-A66.
Rosner, M.R., J. Tang, I. Barzilay. and H.G. Khorana. 1979.

Structure of the lipopolysaccharide from an Escherichia coli

 

heptose-less mutant. I. Chemical degradations and identification
of products. J. Biol. Chem. 25A:5906-5917.
Sadoff, J.C.. D.C. Wright, S. Futrovsky, H. Sidberry, H. Collins.

and B. Kaufmann. 1985. Characterization of mouse monoclonal

5A.

55.

56.

57.

58.

59.

60.

76

antibodies directed against Pseudomonas aeruginosa

 

lipopolysaccharides. Antibiot. Chemother. 36:13A-1A6.

Sidberry, H., B. Kaufman, D.C. Wright, and J. Sadoff. 1985.
Immunoenzymatic analysis by monoclonal antibodies of bacterial
lipopolysaccharides after transfer to nitrocellulose. J. Immunol.
Methods 76:299-305.

Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic
transfer of proteins from polyacrylamide gels to nitrocellulose
sheets: procedure and some applications. Proc. Natl. Acad. Sci.
USA 76:A350-A35A.

Vaara, M., T. Vaara. M. Jensen, I. Helander, M. Nurminen, E.Th.
Rietschel, and P.H. Makela. 1981. Characterization of the
lipopolysaccharide from the polymyxin-resistant pmrA mutants of

Salmonella typhimurium. FEBS Lett. 129:1A5-1A9.

 

Westphal. 0., O. LUderitz, and F. Bister. 1952. Uber die
extraktion von bakterien mit phenol/wasser. Z. Naturforsch Teil
B7:1A8-155.

Wilkinson, S.G. 1983. Composition and structure of

lipopolysaccharides from Pseudomonas aeruginosa. Rev. Infect. Dis.

 

5:89A1-S9A9.
Wilkinson. 8.0., and L. Galbraith. 1975. Studies of

lipopolysaccharides from Pseudomonas aeruginosa. Eur. J. Biochem.

 

52:331-3A3.
Wright, 8.0.. and P.A. Rebers. 1972. Procedure for determining
heptose and hexose in lipopolysaccharides. Modification of the

cysteine-sulfuric acid method. Anal. Biochem. A9:307-319.

CHAPTER III

BINDING OF POLYCATIONIC ANTIBIOTICS TO LIPOPOLYSACCHARIDE

FROM AN ANTIBIOTIC-SENSITIVE STRAIN OF PSEUDOMONAS AERUGINOSA

 

77

ABSTRACT

The interactions of cationic antibiotics with lipopolysaccharides
(LPS) from PA01 strains 1715 (aminoglycoside supersensitive mutant).

1670 (parent). and 1716 (revertant) of Pseudomonas aeruginosa were

 

characterized using a cationic electron spin resonance probe. The
levels of phosphate in the three LPS isolates were similar, as were the
levels of bound divalent metal cations. Electrophoretic separation of
the three isolates showed similar LPS banding patterns. Analysis of the
Western blots of LPS isolates with anti-O-antigen specific monoclonal
antibody indicated that the ladder pattern was comprised of double bands
and the level of one of the bands in the doublet was in much lower
amounts in the isolate from the mutant than from the parent or revertant.
Titration of the Mg salt of the three samples with polymyxin B,
gentamicin C, spermine, MgClg. and C8C12 indicated that the 1715 isolate
had a much higher affinity for all the polycations tested compared to
the other two isolates. The order of affinity for all samples was

polymyxin >> gentamicin = spermine > Ca2+ > Mg2+. Also, the Ca salts
all had lower affinities for the polycations than the analogous Mg salts.
The affinity of the Ca salt of LPS from 1715 derivative for polycations
was lowered almost to that of the parent and revertant isolates. For
bath salts, the antibiotics induced rigidification of all of the LPS
samples but for the 1715 sample rigidification occurred at lower
antibiotics concentrations for the Mg salt compared to the Ca salt. Our

results suggest that the differences seen in LPS-cation interactions

among the three isolates studied may result from differences in the

78

79

level of substituents in the core-lipid A. Furthermore, the CaLPS

complexes appear to have a lower affinity for antibiotic binding.

INTRODUCTION

Lipopolysaccharide (LPS) molecules are one of the major components
found in the outer monolayer of the outer membrane of gram-negative
bacteria. An important function of the outer membrane is to protect
cells from amphipathic and hydrophobic compounds (27.35). LPS contains
a variety of ionic groups, with acidic phosphates and carboxyl moieties
concentrated within the core and lipid A head group region (26.27.35).

Enterobacterial LPS contains about four phosphate residues in the inner

 

core region and an additional two or three phosphate residues in lipid A

(35); in contrast, LPS from Pseudomonas aeruginosa may contain 10 or

 

more phosphate residues per molecule (62,63). As a consequence. the LPS
carries a net negative charge, at physiological pH values, resulting in
a strong negative charge on the surface of gram-negative cells (13).

LPS is associated with a variety of cations, including Mg2+ and Ca2+
(A,10,35). which form cross-bridges between phosphate groups on
neighboring LPS molecules stabilizing the outer membrane structure
(23.35).

It has been proposed that in some bacteria the ability of cationic
antibiotics to penetrate the outer membrane is related to their ability
to bind LPS, destroying the LPS-LPS cross-bridging and destabilizing the
outer membrane (13.35). Cationic antibiotics, such as polymyxin B and
gentamicin, have been reported to increase the permeability of the outer
membrane to lysozyme and hydrophobic compound (1A,15.53). In addition,
both classes of antibiotics induce blebbing of bacterial outer membranes
and produce small transient holes which makes the outer membrane more

permeable to the antibiotic (17.18,29.A6,A8). The initial action of

80

81

these antibiotics may be to disrupt outer membrane structure, allowing
themselves and other compounds to enter the cell and inhibit specific
metabolic processes (13,35).

Further evidence that disorganization of the outer membrane is
critical in the action of polymyxin was obtained from the study of

polymyxin-resistant (pmrA) mutants of S. typhimurium (28.59.60). These

 

mutants have a low-level resistance to polymyxin, were impermeable to
lysozyme and deoxycholate upon treatment with polymyxin (59). and were
more resistant to Tris-EDTA and polycations such as polylysine and
protamine (57). Compared to the parental isolate, the LPS of the pmgA
mutants was shown to contain elevated amounts of A-aminoarabinose and
ethanolamine, making the LPS less negatively charged (60).

The molecular arrangement of adjacent LPS, and thus LPS packing,
depends in part on the presence of ions. Different packing arrangements
may influence the barrier properties of this outer layer in the intact
outer membrane. Several studies have shown that cations associated with
LPS determine, in part, the type of aggregation of LPS (10,23). There
are many studies which indicate that the cations associated with the
cell surface influence the interaction of polycationic antibiotics on
intact cells (15.2A,33,53). In addition, the sensitivity of deep rough
mutants to hydrophobic inhibitors is reported to be decreased by Mg2+
(51). The bactericidal effect of normal serum on enteric bacteria is
also decreased by Mg2+ (32). Thus, the outer membrane appears to
contain cation binding sites involved in stabilizing the permeability
barrier and these sites are presumably on the LPS (1.5A).

In this study, we have measured the relative affinities of various

polycationic antibiotics and polyamines for purified LPS from P.

.. __

 

82

aeruginosa using a cationic spin probe. We have found that polycationic
antibiotics bind to and differentially rigidify MgLPS and CaLPS
complexes from P. aeruginosa strains PAO 1715 (aminoglycoside
supersensitive mutant), PAO 1716 (revertant). and FAQ 1670 (parent).
LPS-cation interactions may vary due to differences in substoichiometric
modification of the core-lipid A. No other chemical differences was
noted in the LPS isolates. Furthermore, the CaLPS complexes appear to

have a lower affinity for cationic antibiotics.

MATERIALS AND METHODS

Bacterial Strains and Growth Media

 

P. aeruginosa strains PAO 1716 (revertant) and PAO 1715 (an

 

aminoglycoside supersensitive mutant) are derived from strain PAO 1670
(parent) as described previously (30). All the three strains were grown
in mid-logarithmic phase in a 100 2 fermentor containing 80 A of
protease peptone No. 2 medium after inocculation with 1 i of an

overnight culture grown in the same medium.

Isolation of LPS

 

LPS from P. aeruginosa PA01 strains 1715. 1716, and 1670 were

 

isolated by the method of Darveau and Hancock (5). followed by two
extractions in chloroform:methanol (1:1 v/v) resulting in recovery of
approximately 80% of the total LPS. The LPS isolates were dialyzed
extensively against a buffer containing 0.2 M NaCl. 10 mM Tris. 1 mM
EDTA, and 0.01% NaN3, pH 8.0, at 37°C followed by distilled water. The
magnesium (MgLPS) and calcium (CaLPS) salts of the LPS isolates were
formed by dialysis against 10 mM MgC12 or CaClg, respectively, followed
by distilled water. All the LPS isolates were dialyzed simultaneously
to decrease the variation of ion content between preparations. Samples

were lyophylized and stored at -20°C.

Chemical Analysis

 

Inductively coupled plasma emission spectroscopy of wet ashed LPS
samples was used to quantitate phosphorus and metal ion content as

described previously (A). Levels of 2-keto-3-deoxyoctulosonic acid

84

(KDO) were quantitated by the thiobarbituric acid assay (8). Protein
content on LPS was determined by the Pierce BCA protein assay (Pierce

Chemical Co., Rockford, IL) using bovine serum albumin as a standard.

SDS-PAGE and Western blots

 

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis
(SDS-PAGE) was carried out as described previously (A3). using the
buffer system of Laemmli (22) and the silver staining procedure of
Dubray and Bezard (9). The separating gel was formed with 15%
acrylamide and 0.1% SDS, with a 7.5% acrylamide stacking gel.

Western blots on SDS-polyacrylamide gels (12% acrylamide) were
prepared as previously described (56). The gels were electrotransferred
with a model TE Transphor Electrophoresis apparatus (Hoefer Scientific
Instruments) at a constant current of 150 mA for 18 h using the
electrode buffer described by Otten and co-workers (36). The
nitrocellulose blots were visualized as described previously (36) with
monoclonal anti-503 antibody (11, titer ~1:100,000) diluted 1:10.000 in

blocking solution.

Partitioning of spin probe

 

Electron spin resonance (ESR) spectroscopy was carried out with a
Varian X-band spectrometer (model E-112). Sample temperature was
measured with a thermocouple placed within the cuvette. Titrations of
magnesium or calcium LPS suspended at 10 mg/ml were performed by
measuring the spectral parameters of the spin probe A-dodecyldimethyl-
ammonium-1-oxyl-2,2,6,6-tetramethylpiperidine bromide (CAT12; 18:1 molar

ratio, LPS:CAT12) after successive additions of cations to sample at

85

37°C. All the cations and antibiotics as well as the LPS were dissolved
in 50 mM KOH-HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid;
pH 7.0).

Upon the successive addition of cations to samples of LPS, the
spectra were analyzed for 2T”. the hyperfine splitting parameter whose
value is an indicator of probe mobility, and for 9;, the partitioning of
probe between aqueous (F) and LPS-bound (B) environments, calculated as
described by Coughlin fig El“ (1).

Scatchard plots of binding of CAT12 to the LPS isolates from P.
aeruginosa strains 1715. 1716, and 1670 were determined by suspending
either Mg- or CaLPS complexes at 5 pM in 50 mM KOH-HEPES (pH 7.2). The
LPS samples were mixed with different concentrations of CAT12 and the
concentrations of bound (B) and free (F) probe were determined from the

resultant ESR spectra, measured at 25°C.

Chemicals

Polymyxin B sulfate, gentamicin C sulfate, and spermine were
purchased from Sigma Chemical Co., St. Louis, MO.; the monoclonal
anti-503 antibody was a gift of L.E. Bryan; and CAT12 was synthesized as
previously described (1). All other chemicals were of reagent grade or

better.

RESULTS

Silver staining of LPS from PA01 strains 1715 (mutant), 1716

(revertant), and 1670 (parent) of P. aeruginosa separated by SDS-PAGE

 

revealed three major sets of bands (Fig. 1A, lanes 1-3) indicating as
many as three populations of molecules differing in O-antigen length
(12,20,21,38). At low concentrations, there was no difference in the
migration pattern of the short chain LPS of the three strains (Fig. 1A,
lanes A-6). Analysis of Western blot of LPS isolates with
anti-O-antigen specific monoclonal antibodies revealed the ladder
pattern of O-antigen-containing molecules with increasing length of
O-antigens (Fig. 1B). This blot indicates that the ladder consists of
doublet bands. and the level of one of the doublet bands is in much
lower amounts in the isolate from the mutant, 1715, than from the
parent, 1670. Presumably. this reflects a difference in
substoichiometric modification within the core-lipid A region of the
molecules (28,60).

The metal and phosphate content of the magnesium and calcium salts
from the three LPS isolates were determined by inductively coupled
plasma emission spectroscopy (ICP). There were approximately eleven
phosphates bound per LPS molecule for all three isolates (Table 1). and
no significant difference was observed in the phosphorus content of
magnesium and calcium salts (at P < 0.05). All of the samples studied
were in the magnesium or calcium salt forms. There was no detectable
protein associated with the LPS salts. The thiobarbituric acid assay
showed two reactive KDO residues per molecule of LPS for the three

strains.

86

Figure 1.

87

(A) Silver-stained SDS-PAGE of LPS from P. aeruginosa
strains PAO 1670 (lanes 1 and A); PAO 1716 (lanes 2 and 5);
and PAO 1715 (lanes 3 and 6). Samples of either 5 pg (lanes
1-3) or 0.10 pg (lanes A-6) were applied to a 15%
polyacrylamide gel which had polymerized overnight with a
butanol overlay. Arrowheads indicate the three regions of
intensively staining. (B) Western blots of LPS from P.

aeruginosa strains PAO 1715 (lane 7). FAQ 1716 (lane 8), and

 

PAO 1670 (lane 9) reacted with monoclonal anti-503-antibody.

123456

A
”I."

' ”No.

 

88

89

Table 1. Elemental Composition of Purified P. aeruginosa LPS Isolatesa

 

Strain Ca/Pb MS/Pb +/Pb’° Ca/Lpsb'd MS/Lpsb'd P/LPSb’d

 

MgLPS
1670 < 0.01 0.76:0.01 1.55:0.0A N.D.g 10.20:0.13 13.A:0.1
1716 < 0.01 0.75:0.01 1.52:0.02 N.D. 9.30:0.56 12.A:0.6

1715 < 0.01 0.72:0.01 1.A8:0.01 N.D. 8.39:2.1A 11.6:2.7

CaLPS
1670 0.78:0.0A < 0.01 1.60:0.09 7.6A:1.51 N.D.e 9.811.5
1716 0.81:0.0A < 0.01 1.69:0.0A 7.8A:O.71 N.D. 9.8:1.2
1715 0.77:0.0A < 0.01 1.57:0.09 7.66:0.01 N.D. 10.0:O.A

 

aThe metal and phosphate content of the magnesium and calcium salts from
the three LPS isolates were determined by inductively coupled plasma
emission spectroscopy (ICP). There was less than 0.01 metal per
phosphate (mole/mole) of Na, K, Fe, Al, and Zn ions associated with the
samples. Values are expressed as the Mean i S.D. Comparisons of the
values were made using the randomized block ANOVA statistical analysis
at P<0.05 (52).

bValues are expressed in mol/mol.
cTotal metal cation charges bound per phosphate.
dAssumes an average molecular mass of 9,000.

eN.D.. not detected at levels greater than 0.01 metal per phosphate.

90

A cationic electron spin resonance probe, CAT12, was used to
measure cation binding to the Mg2+ and Ca2+ salt of LPS from the parent
(1670), revertant (1716), and mutant (1715) strains. Scatchard analysis
of CAT12 binding was used to obtain a KD, the apparent dissociation
constant. and N, the number of binding site for the probe. Also the
Hill constant, up. was calculated to give an index of the cooperativity
of probe binding. Results from the Scatchard analysis indicated that
for the MgLPS isolates there were approximately 3.5 probe binding sites
per LPS molecule, one high affinity site and approximately 2.5 low
affinity sites (Table 2). The apparent KB of probe for the high
affinity site in the MgLPS of the mutant 1715 was higher than for that
of the parent strain 1670, indicating that the mutant isolate had a
lower affinity for binding probe to this site, indicating that the probe
can more readily be displaced by other cations. The apparent KD's of
the probe for the low-affinity sites and the total number of sites were
similar for the MgLPS of all three strains. At low ligand
concentrations, a concave-down shape of the Scatchard plot was observed
(Fig. 2). This, along with the Hill coefficient (Table 2), indicated
positive copperativity binding of CAT12 to MgLPS aggregates. Positive
cooperativity has also been observed for CAT12 binding to the MgLPS
complexes of S. 2913 (A0) and to the sodium salt of lipid A (McGroarty
and Chessen. unpublished data). Schindler and Osborn (A7). using
fluorometric titrations, reported KD values of 6 and 15 pM for Ca2+ and

Mg2+ binding to dansylated LPS from S. typhimurium. The differences in

 

the reported Kb's for binding of CAT12 (A0) and of divalent metal

cations (A7) to Enterobacterial LPS is likely the result of differences

 

in the physico-chemical prOperties of the cations, the type of label,

91

Table 2. Scatchard Analysis of CAT12 Binding to LPS Isolates of
P. aeruginosa3

 

 

 

 

 

High Affinity Site Low Affinity Site

Strain KD(pM) Nb 0H0 KD(pM) Nb 0H0
MgLPS

1670 0.11:0.03 1.0:o.1 2.6 2.2:o.5 2.5:o.3 1.0

1716 0.32:0.10 1.3:o.2d 1.5d 2.2:o.2 2.2:o.1 0.9

1715 o.uu:o.09d 1.A:O.1d 1.3 2.AiO.A 1.9:o.3 1.1
CaLPS

1670 0.29:0.02 1.0:o.1 1.ud 2.5:o.9 1.8:0.2 1.0

1715 0.28:0.02 1.0:0.1 1.1d 3.0i0.6 1.9:o.1 1.0

 

a5 pM samples of the calcium and magnesium salts of the LPS isolates
were suspended in 50 mM KOH-HEPES (pH 7.2) at 25°C and mixed with
different concentrations of CAT12. The levels of bound and free probe
were determined from the resultant ESR spectra. Results are averages
of three samples.

bN represents the number of probe-binding sites per LPS molecule.
Cap is the Hill coefficient (50).
dValues differ significantly at a P=0.05 from the value of the 1670

MgLPS sample. Analysis was determined using the Randomized ANOVA
statistical analysis (52).

Figure 2.

92

Scatchard plots of CAT12 binding to magnesium (o) and calcium

(0) salts of LPS from P. aergginosa strain PAO 1670. Samples

 

were prepared as described in Materials and Methods. Bound
probe is presented as moles of bound CAT12 relative to moles
of LPS. Error bars indicate the standard deviation for the

triplicate measurements made on the samples.

93

 

 

 

 

 

94

and in the experimental conditions used to determine the binding
affinities.

The probe binding affinities for the CaLPS and MgLPS complexes were
different (Fig. 2). Relative to MgLPS. strain 1670 exhibited a 2.6-fold
higher KD for the high affinity site in the CaLPS with no change in the
number of binding sites (Table 2). Surprisingly. the CaLPS isolate from
the mutant strain (1715) showed an approximately 2-fold decrease in its
relative KD for the high-affinity site indicating that the probe was
displaced less readily from this salt. On the other hand, the CAT12
exhibited a slightly higher KD for the low-affinity site in the parent
and mutant isolates but the parental sample had almost one less binding
site. Furthermore, at low CAT12 concentrations positive copperativity
was decreased significantly for the CaLPS (Fig. 2, and Table 2). At low
probe concentrations used in the antibiotic titration experiments
described below, most of the probe was initially bound to the
high-affinity site.

Titrations of LPS samples containing CAT12 with various cations
displaced different amounts of the probe. reflected in the partitioning
parameter 9;. Generally, addition of highly charged cations (polymyxin
B and gentamicin) displaced CAT12 at lower concentrations than did
cations with lower charge (e.g.. Ca2+ and Mg2+). Displacement of the
spin probe from the MgLPS complex of the mutant strain (1715) by any
cation occurred at much lower levels of added cation per LPS than for
the parent (1670) or revertant (1716) strains isolates (Fig. 3. left
panel). This difference in binding to the MgLPS complexes from the

various strains of P. aeruginosa is probably related to differences in

 

the substoichiometric modification in the core-lipid A region (35,60).

Figure 3.

95

The partitioning parmaters Vi measured as a function of added
(A) M8C12, (B) CaClg. (C) spermine, (D) gentamicin C, and (E)
polymyxin B. Partitioning of the spin probe CAT12 was

determined upon addition of cations to MgLPS (closed symbols)

and CaLPS (open symbols) from P. aeruginosa strains PAO 1670

 

(parent, 0 o); PAO 1715 (antibiotic supersensitive mutant, A
A); and FAQ 1716 (revertant a I). The amount of added cation

is plotted as the number of cation per LPS (mol/mol).

96

 

 

 

 

 

 

 

 

 

 

A m” I I I
075 - .1.
h Q50 . ..
025 - O .. J
CaLPS ' '
*4. .1. fig
CaLPS ' '
1.5 " "'
,A
" I0 " '0'
05 ./ m.
D r‘wtps I v Ctps I I
1.3 '- ~-
A
K
h 1.0 1- .. ~-
C
I
as- f ..
E u” I l “B r I
(A
‘5 1- . 41-
5 AA
" 1.0 b / di-
05 . 1.
01 no 10 to 13
CANON/LPS “TM/LPS

 

97

LPS from PA01 derivatives 1716 and 1670 exhibited similar cation binding
affinities. The binding affinities of the compounds for MgLPS of all
three strains was as follows: polymyxin B >> gentamicin C 5 spermine >>
Ca2+ > Mg2+ (Fig. A, left panel).

Samples of the calcium salts of the 1715 and 1670 LPS isolates were
also titrated with the various polycations (Fig. 3, right panel). The
competitive displacement curves showed a similar pattern of cation
affinities as the MgLPS complexes (Fig. A, right panel). However. the
9; values for CaLPS calculated for both strains were significantly
decreased when compared to the MgLPS isolates at the same level of added
cation; the mutant strain was the most affected by change in the salt
form (Fig. 3). The decrease in the cation affinity may suggest that a
rearrangement in the LPS aggregate structure occurred when changing the
counter ion from Mg2+ to Ca2+.

To further explore this possibility, the relative fluidity of the
bound CAT12 probe was measured by calculating ZTy as a function of
added cation concentrations (Fig. 5). The hyperfine splitting parameter
22¢ is related to the rotational mobility of the spin label and
therefore reports the local motion within the LPS head groups. High
values of 2T” reflects low motion (rigidification). It can be seen
from the data that: (1) upon titration of MgLPS with polycations, a
structural change in the 1715 sample occurred at lower cation
concentrations than for strain 1670 (Fig. 5. left panel), and (ii) in
the absence of added cations, the CaLPS aggregate had a more rigid
structure. than the MgLPS aggregate (Fig. 5. right panel). The
antibiotics polymyxin B and gentamicin C rigidified the MgLPS of strains

1715 to a greater extent than that of the parent strain (Fig. 5D and 5E,

Figure A.

98

Partitioning of the spin probe CAT12 upon addition of cations
to (A) MgLPS and (B) CaLPS from P. aeruginosa strain PAO 1715.
The partition function Vi was measured upon the addition of
polymyxin B (o). gentamicin C (A), spermine (a), CaC12 (A).
and MgClg (o). The amount of added cation is plotted as the

number of cation per LPS (mol/mol).

99

 

..mdd \ 99:8

 

 

Wad \ H2920“

 

on 0.. no on 0.. no
_ _ d ‘11" J
1 m‘h
I\
x4
\. .
\n
4
\. .
4
.1. o
41‘ I
-\
o
o
o
o
i J
958 as memos. .5
p — — P _ _

 

0.0

0..

m.—

 

 

Figure 5.

100

Head group mobility of LPS from P. aeruginosa strain PAO 1670

 

(a); FAQ 1716 (D); and PAO 1715 (A) was measured by the
hyperfine splitting parameter (2Ty, gauss) of bound CAT12.
This parameter was measured as a function of increasing
concentrations of (A) MgClg, (B) CaClg. (C) spermine, (D)
gentamicin. C, and (E) polymyxin B added to the magnesium

(closed symbols) and calcium (open symbols) salts of LPS.

101

 

 

 

 

 

58» T
5 ”P5 fl ' ' cm Y I '
62r 1’
A A A
/7 A _ < \ ‘
2.". A
M‘Vflk-rg ! 0
so 0‘ o .
Ir/ ‘3“ A

 

 

 

 

 

 

 

 

 

 

AK A
1W\
4+ \ \O O
O O A
w r J»
CaLPS ' t .
r
A
4
A
<5
53 r '1)
E wsj ' ' CaLPS ' ' '
up. A
A
1
1i-
“ r A
0.3 I0 30 03 '0 30

CANON/LPS CHM/LPS

 

102

left panel). Of these two antibiotics, polymyxin B rigidified the LPS
aggregate to a greater extent. In contrast, spermine, added at high
concentrations, tended to fluidize the MgLPS complexes (Fig. 5C, left
panel). Addition of Ca2+ and Mg2+ had only slight effects on the motion
within the MgLPS aggregates rigidifying the 1715 sample at intermediate
or high concentrations (Fig. 5A and SB, left panel). Similar changes in
head group motion upon cation addition have been observed previously
(A1).

Different structural changes were observed upon addition of
polycations to CaLPS (Fig. 5, right panel). At low concentrations of
cations, differences in motion were seen between the MgLPS from the
parent and mutant strains isolates. In contrast, for the Ca2+ salts,
similar values of 2T” were measured over the complete titrations of
both strains. The cation-induced changes in structure was more

pronounced in the MgLPS preparation than in the CaLPS.

DISCUSSION

The LPS isolates of P. aeruginosa strains 1670, 1716, and 1715 were

 

Shown to contain as many as three major size populations (Fig. 1A). The
Western blot of the LPS (Fig. 1B) revealed a ladder-like banding pattern
with regular spacing. In addition, the bands were resolved as doublets
which presumably represents substoichiometric modifications in the

core-lipid A similar to the doublets seen in the LPS from Salmonella and

 

S. coli (35,38.AO,60); such variable substitutions have been observed in

the 31P-NMR analysis of P. aeruginosa lipid A phosphates (M. Bateley.

 

and R.E.W. Hancock, unpublished). Elemental analysis indicated that
there were no significant differences in the level of phosphate and
metal cations for the three LPS isolates (Table 1). We propose that the
susceptibility of the cells to polycationic compounds may result from
differences in the substoichiometric modification of the core-lipid A
region of the LPS which affects the cation interactions at this Site but
not the overall negative charge.

In order to characterize the differences in cation affinity and
number of binding sites. we first analyzed the binding of the cationic
spin probe, CAT12, by Scatchard analysis. For all three strains. the
Scatchard plots were nonlinear (Fig. 2). The results indicated that
there were two types of binding sites on LPS for which CAT12 and
presumably other cations compete. We also analyzed the binding data by
calculating the Hill constant (an; Table 2); this constant is an index
of the cooperativity of binding. Values of 0H larger than one indicated
positive cooperativity, whereas aH=1 indicates no cooperativity (50). A

Hill coefficient with a value less than one or a curvilinear Scatchard

103

104

plot may indicate either negatively cooperativity interactions or
heterogeneity of receptor sites (50). Thus. the high affinity site in
all three LPS isolates showed positive cooperativity of CAT12 binding
whereas the low affinity site indicated no copperativity (Table 2). The
MgLPS isolate from the parent strain, 1670, showed a significantly
higher level of cooperativity in probe binding and a lower apparent KB.
This may reflect a difference in the ability of CAT12 to intercalate
between the head groups of the three MgLPS isolates. The cooperativity
of binding for all the isolates may be due to a CAT12-induced alteration
in LPS aggregate packing at low levels of CAT12 association. The
initial barrier to probe penetration at low probe concentrations may
result from strong polar and ionic interactions within the core-lipid A,
between O-polymers and between core sugars (A5).

The CaLPS complexes were different from the MgLPS isolates in their
interactions with the spin probe. The CAT12 probe had a lower KD for
the mutant CaLPS isolate than for the MgLPS. In addition, the parent
strain's isolate lost approximately half of a binding site when the Mg2+
was substituted with Ca2+. At low concentration of CAT12 there was a
significantly lower amount of probe bound to CaLPS complex than to the
MgLPS (see Fig. 2). and the positive cooperativity was decreased
dramatically for the CaLPS of the parent strain (on; Table 2).
Differences in the partitioning of the spin probe for the MgLPS and
CaLPS complexes may be explained based on the physico-chemical

properties of the Mg2+ and Ca2+ ions. The unhydrated ionic radius for

2
Mg + and 032* are 0.65 A and 0.99 A. respectively (16). This size is a
limiting factor in determining the coordination number of a cation while

it also determines the extent of hydration and solubility of its salts

105

(16). For cations of similar charge. the hydration energy increases
with decreasing cation radius. Mg2+ forms a six-coordinated complex of
regular structure. On the other hand, Ca2+, using the g orbitals in its
valence shell, can often form higher coordination complexes (7 or 8)
with irregular geometry (16,6A). The flexibility of Ca2+ in its binding
requirements means that it can bind more strongly to the irregular sites
offered by biological ligands and can be more effective in cross-linking
ligands (16,6A). LPS binds divalent cations strongly (A7). and the
nature of counter ions has a pronounced influence on the physical
structure of LPS aggregates (3.10). Strain and co-workers (5A) suggest
that both the carbonyl group of K00 and a phosphate group on lipid A are
involved in coordinating a single metal ion. We propose that the Ca2+
ion rearranges the anionic groups in the core-lipid A region into a more
compact or "close" structure whereas the Mg2+ ion forms a more "open"
type of structure. This is consistent with studies done on cations
interacting with phospholipid vesicles (39.AA). In these studies, it
was found that addition of Ca2+ to phosphatidylserine vesicles induced
the formation of an "anhydrous" complex of closely opposed membranes
with highly ordered crystalline acyl chains while the Mg2+ complexes
were more hydrated and had no acyl chain crystallization. Thus, one
would expect the Ca2+ ion to bind more tightly to LPS and not be as
readily displaced by a competitive ligand as Mg2+.

The ability to displace the CAT12 probe from the LPS isolates
depended on the cations used. The interaction of polycationic
antibiotics with LPS depends on net charge, size and conformation as
well as on the presence of hydrophobic sites on the antibiotic molecule

(37,A1.53.58). Since the LPS of P. aeruginosa has an unusually high

 

106

phosphate content (62,63) and two to three KDO residues (19.63), there
are numerous arrangements of cation-binding sites. some of which may
have high affinity for certain cations. Competition between cations
could be for the same site (31.33.A7.53). or for overlapping sites
(31.A1). As a given cation binds to LPS, the LPS aggregate may alter
its conformation to Optimize ionic interactions. and thus alter other
cation binding sites. This induced change in conformation is suggested
by the cooperative binding of polycations to LPS from S. 2911 and P.

aeruginosa (2A,31,A1,A2) and the physical properties of different LPS

 

salt forms (A,10). Thus, partitioning of the spin probe CAT12 upon

cation addition may depend on the affinity of the added cation, the

preferrential site to which the cation binds, and the ability of the
cation to induce a structural alteration upon binding.

CAT12 was readily displaced from the MgLPS complex by polymyxin B
(Fig. 3). Polymyxin B as well as the other polycations displaced CAT12
more readily from LPS of strain 1715 than that of strains 1670 and 1716,
in accordance with the Scatchard analysis. This difference in polymyxin
B binding to the MgLPS complexes might be a result of differences in
structure caused by the alteration in substoichiometric modification
observed in the Western blot of the LPS (Fig. 18). Vaara g; 21' (60)
and Peterson SE El: (A0) have demonstrated that decreases in polymyxin B

binding to Enterobacterial LPS was a consequence of substitution of the

 

core-lipid A phosphates. The nature of the substoichiometric
modifications in our LPS isolates has yet to be elucidated.

It is not surprising that polymyxin B has a high affinity for LPS
since this antibiotic has been reported to bind and to perturb the

packing arrangement of membranes and acidic phospholipids (31,37.A9.53).

107

It has been proposed that the binding of polymyxin B and related
polypeptide antibiotics is through electrostatic and hydrophobic
interactions (31.A9.53).

The polycations, gentamicin and spermine, displace much less spin
probe from MgLPS complexes when compared to polymyxin B (Fig. A).
Since, at neutral pH gentamicin and spermine have a net charge of +A.5
and +A, respectively, the nature of the binding for these molecules to
LPS is mainly by electrostatic interaction with the LPS core
(7,37,A1,58). Vaara and Vaara (58) suggest that at least five basic
charges are sufficient to increase outer membrane permeability if the
charges are in a favorable conformation. Thus, the size and shape of
the LPS molecule plays a significant role in outer membrane disruption.

Addition of the same concentration of competitive cations to CaLPS
and MgLPS showed less CAT12 displacement from the CaLPS than from the
MgLPS (Fig. A). Again this is compatible with the Scatchard data
obtained from the mutant strain and is consistent with the model
proposed for the two salt forms. The preference of Ca2+ over Mg2+, when
titrating either the Ca- or MgLPS complexes with divalent cations (Fig.
3). implies that the Ca2+ ion tends to complex more tightly to LPS than
Mg2+ (6A). This is due to the ease with which calcium accomodates a
variety of stereochemistries and a large number of coordinating groups
(6A). It has been shown that the packing of the LPS aggregates depend
on the salt form (3,10). Perhaps what we are observing. in the Ca2+
salts, are LPS aggregates different from the MgLPS. In the intact outer

2+ may have
membrane, regions with localized concentrations of Mg2+ or Ca

different LPS packing. These regions may differ in their ability to

form a permeability barrier. Newton (33) and others (6,51,61,65) have

108

reported that Ca2+ is protective against the action of antibiotics in

Enterobacteriaceae as well as in Pseudomonas.

 

 

The increased rigidity of LPS aggregates upon addition of cationic
antibiotics indicates an antibiotic-induced alteration of the LPS
packing (Fig. 5). Peterson §£_gl. (AO,A1,A2) observed similar
polycation induced structural alterations in the MgLPS complexes from

Enterobacteriaceae and Pseudomonas. Electron microscopy studies

 

 

(25,29,A8.58) have shown that polymyxin B and other polycationic
compounds alter the cell envelope, inducing blebs, rodlike projections,
or holes. It has been proposed that these antibiotics alter the
structure of the outer membrane and provide a pathway for their own
uptake (13.35). The alteration in LPS structure seen in the pure LPS
may reflect the changes that occur in the intact membrane.
Interestingly. the antibiotic-induced alteration in the LPS structure
reported here was more pronounced in the MgLPS complexes than the CaLPS
complexes supporting the idea that Ca2+ stabilizes and protects the
aggregates against binding of antibiotics. It has been shown that metal
cations can stabilize the outer membrane and decrease its permeability
(33.3A.51.65).

The polyamine. spermine, which is naturally present in the outer
membrane, also binds tightly to LPS yet does not immobilize the LPS head
groups (Fig. 3 and 5). Peterson 35 21° (A1,A2) observed that spermine
increases the motion of LPS head groups. Spermine has been shown to
stabilize lysozome-induced S. 9911 spheroplast against lysis in water
(55). Also, Vaara and Vaara (58) demonstrated that spermine was totally
inactive both in increasing outer membrane permeability and as an

antibacterial agent.

109

In summary, we have found that polycationic antibiotics bind to and

rigidify purified LPS isolates of P. aeruginosa. The nature of the

 

antibiotic binding to LPS is more than an ionic attraction between LPS
and the polycation. The affinity of a specific cation for LPS also
depends on the number and arrangement of the charges in the antibiotic
molecule as well as the presence of hydrophobic groups. The decreased
LPS motion, upon binding of antibiotic, further supports a
cation-induced alteration of the aggregate packing arrangement which
perturbs the membrane integrity and increases outer membrane
permeability (AO,A1).

The differences in antibiotic susceptibility observed among the
three strains used in this study may result from substoichiometric
modifications of the core-lipid A region. This would suggest that
changes other than the overall negative charge can affect cation binding.
The changes in LPS molecule in the antibiotic sensitive strain are
thought to result from an alteration in the level of substoichiometric
modification in the core-lipid A region which affects cation affinities
perhaps by affecting the conformation of the aggregates. Finally. these
data also suggest that the LPS aggregate structure and perhaps the
structure of the intact outer membrane may depend in part on the

divalent cations present and their physico-chemical properties.

REFERENCES

Coughlin, R.T., C.R. Caldwell. A. Haug, and E.J. McGroarty. 1981.
A cationic electron spin resonance probe used to analyze cation
interactions with lipopolysaccharides. Biochem. Biophys. Res.
Commun. 100:1137-11A2.

Coughlin, R.T., A. Haug, and E.J. McGroarty. 1983. Electron spin
resonance probing of lipopolysaccharide domains in the outer

membrane of Escherichia coli. Biochem. Biophys. Acta. 729:161-166.

 

Coughlin, R.T., A. Haug, and E.J. McGroarty. 1983. Physical
properties of defined lipopolysaccharide salts. Biochemistry
22:2007-2013.

Coughlin, R.T., S. Tosanger, and E.J. McGroarty. 1983.
Quantitation of metal cations bound to membranes and extracted

lipopolysaccharide of Escherichia coli. Biochemistry 22:2002-2006.

 

Darveau, R.P.. and R.E.W. Hancock. 1983. Procedure for isolation
of bacterial lipopolysaccharides from both smooth and rough

Pseudomonas aeruginosa and Salmonella typhimurium strains. J.

 

 

Bacterial. 155:831-838.
Davis, S.D., and A. Iannetta. 1972. Influence of serum and
calcium on the bactericidal activity of gentamicin and

carbenicillin on Pseudomonas aeruginosa. Appl. Microbiol.

 

23:775-779.

Day, D.F. 1980. Gentamicin-lipopolysaccharide interaction in
Pseudomona aeruginosa. Curr. Microbial. A:277-281.

Droge. W., V. Lehmann, O. LUderitz. and O. Westphal. 1970.
Structural investigations on the 2-keto-3-deoxyoctonate region of

lipopoly- saccharides. Eur. J. Biochem. 1A:175-18A.

110

10.

11.

12.

13.

1A.

15.

16.

111

Dubray, G., and G. Bezard. 1982. A highly sensitive periodic
acid-silver stain for 1,2-diol groups of glycoproteins and
polysaccharides in polyacrylamide gels. Anal. Biochem.
119:325-329.

Galanos, C.. and O. Lfideritz. 1975. Electrodialysis of
lipopolysaccharides and their conversion to uniform salt forms.
Eur. J. Biochem. 5A:603-610.

Godfrey, A.J., M.S. Shahrabadi. and L.E. Bryan. 1986.
Distribution of porin and lipopolysaccharide antigens on a

Pseudomonas aeruginosa permeability mutant. Antimicrob. Agents

 

Chemother. 30:802-805.
Goldman, R.C., and L. Leive. 1980. Heterogeneity of

antigenic-side chain length in lipopolysaccharide from Escherichia

 

coli 0111 and Salmonella typhimurium LT2. Eur. J. Biochem.

 

107:1A5-153.

Hancock, R.E.W. 198A. Alterations in outer membrane permeability.
Ann. Rev. Microbiol. 38:237-26A.

Hancock. R.E.W., V.J. Raffle, and T.I. Nicas. 1981. Involvement
of the outer membrane in gentamicin and streptomycin uptake and

killing in Pseudomonas aeruginosa. Antimicrob. Agents Chemother.

 

19:777-785.
Hancock. R.E.W., and P.G.W Wong. 198A. Compounds which increase

the permeability of the Pseudomonas aeruginosa outer membrane.

 

Antimicrob. Agents Chemother. 26:A8-52.
Hughes. M.N. 1981. The alkali metal and alkaline earth metal

cations in biology, p. 256-295. In M.N. Hughes (ed.), The

17.

18.

19.

20.

21.

22.

23.

112

Inorganic Chemistryng Biological Processes (Second Edition), John

 

 

Wiley and Sons, Ltd.

Iida, K., and M. Koike. 197A. Cell wall alterations of gram-
negative bacteria by aminoglycoside antibiotics. Antimicrob.
Agents Chemother. 5:95-97.

Koike. M., K. Iida, and T. Matsuo. 1969. Electron microscopic
studies on mode of action of polymyxin. J. Bacteriol. 97:AA8-A52.
Kropinski, A.M.. L.C. Chan. and F.H. Milazzo. 1979. The

extraction and analysis of lipopolysaccharides from Pseudomonas

 

aeruginosa strain PAO, and three rough mutants. Can. J. Microbiol.

 

25:390-398.
Kropinski, A.M., B. Jewell, J. Kuzio, F. Milazzo. and D. Berry.

1985. Structure and functions of Pseudomonas aeruginosa

 

lipopolysaccharide, p. 58-73. In D.P. Speert and R.E.W. Hancock

(eds.), Pseudomonas aeruginosa: New Therapeutic Approaches from

 

 

Basic Research. (Antibiotics and Chemotherapy; Vol. 36). Karger,

 

Basel.
Kropinski, A.M., J. Kuzio, B.L. Angus, and R.E.W. Hancock. 1982.
Chemical and chromatographic analysis of lipopolysaccharide from an

antibiotic- supersusceptible mutant of Pseudomonas aeruginosa.

 

Antimicrob. Agents Chemother. 21:310-319.

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

Leive, L. 197A. The barrier function of the gram-negative

envelope. Ann. N.Y. Acad. Sci. 235:109-129.

2A.

25.

26.

27.

28.

29.

30.

31.

113

Loh, B., C. Grant, and R.E.W. Hancock. 198A. Use of the
fluorescent probe 1-N-phenylnaphtylamine to study the interactions
of aminoglycoside antibiotics with the outer membrane of

Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 26:5A6-551.

 

Lounatmaa. K., P.H. Makela, and M. Sarvas. 1976. Effect of
polymyxin on the ultrastructure of the outer membrane of wild-type

and polymyxin-resistant strains of Salmonella. J. Bacteriol.

 

127:1AOO-1AO7.

Lflderitz. 0., M.A. Freudenberg, C. Galanos, V. Lehmann, E.Th.
Rietschel. and D.H. Shaw. 1982. Lipopolysaccharides of
gram-negative bacteria. Curr. Topics Membs. Transp. 17:79-150.
Lfigtenberg, B., and L. Van Alphen. 1983. Molecular

architecture and functioning of the outer membrane of Escherichia

 

29;; and other gram- negative bacteria. Biochem. Biophys. Acta.
737:51-115.

Makela, P.H., M. Sarvas. S. Calcagno, and K. Lounatmaa. 1978.
Isolation and characterization of polymyxin-resistant mutants of

Salmonella. FEMS Microbiol. Lett. 3:323-326.

 

Martin, N.L., and T.J. Beveridge. 1986. Gentamicin interaction

with Pseudomonas aeruginosa cell envelope. Antimicrob. Agents

 

Chemother. 29:1079-1087.

Mills, B.J., and B.W. Holloway. 1976. Mutants of Pseudomonas

 

aeruginosa that show specific hypersensitivity to aminoglycosides.

 

Antimicrob. Agents Chemother. 10:A11-A16.
Moore, R.A., N.C. Bates. and R.E.W. Hancock. 1986. Interaction of

polycationic antibiotics with Pseudomonas aeruginosa

 

32.

33.

3A.

35.

36.

37.

38.

39.

114

lipopolysaccharide and lipid A studied by using dansyl-polymyxin.
Antimicrob. Agents Chemother. 29:A96-500.

Muschel, L.H., and J.B. Jackson. 1966. Reversal of the
bactericidal reaction of serum by magnesium ion. J. Bacteriol.
91:1399-1A02.

Newton, B.A. 195A. Site of action of polymyxin on Pseudomonas

 

aeruginosa: antagonism by cations. J. Gen. Microbiol. 10:A91-A99.
Nikaido, H., P. Bavoil, and Y. Hirota. 1977. Outer membranes of
gram-negative bacteria. XV. Transmembrane diffusion rates in

lipoprotein-deficient mutants of Escherichia coli. J. Bacteriol.

 

132:1OA5-1OA7.

Nikaido, H., and M. Vaara. 1985. Molecular basis of bacterial
outer membrane permeability. Microbiol. Rev. A9:1-32.

Otten, S., S. Iyer, W. Johnson, and R. Montgomery. 1986.

Serospecific antigens of Legionella pneumophila. J. Bacteriol.

 

167:893'9OA.

Pache, W., D. Chapman, and R. Hillaby. 1972. Interaction of
antibiotics with membranes: polymyxin B and gramicidin 8.
Biochim. Biophys. Acta. 255:358-36A.

Palva. E.T., and P.H. Makelfi. 1980. Lipopolysaccharide

heterogeneity in Salmonella typhimurium analyzed by sodium dodecyl

 

sulfate/polyacrylamide gel electrophoesis. Eur. J. Biochem.
107:137-1A3.

Papahadjopoulos, D., A. Portis, and W. Pangborn. 1978. Calcium
induced lipid phase transitions and membrane fusion. Ann. N.Y.

Acad. Sci. 308:50-66.

A0.

“1.

A2.

A3.

uu.

A5.

A6.

A7.

115

Peterson, A.A., S.W. Fesik, and E.J. McGroarty. 1987. Decreased
binding of antibiotics to lipopolysaccharides from

polymyxin-resistant strains of Escherichia coli and Salmonella

 

 

typhimurium. Antimicrob. Agents Chemother. 31:230-237.

 

Peterson, A.A., R.E.W. Hancock, and E.J. McGroarty. 1985. Binding
of polycationic antibiotics and polyamines to lipopolysaccharides

of Pseudmonoas aeruginosa. J. Bacteriol. 16A:1256-1261.

 

Peterson, A.A.. A. Haug, and E.J. McGroarty. 1986. Physical
properties of short- and long-O-antigen-containing fraction of

lipopolysaccharide from Escherichia coli 0111:BA. J. Bacteriol.

 

165:116-122.
Peterson, A.A., and E.J. McGroarty. 1985. High-molecular-weight

components in lipopolysaccharides of Salmonella typhimurium,

 

"Salmonella minnesota, and Escherichia coli. J. Bacteriol.

 

 

162:738-7A5.

Portis, A., C. Newton, W. Pangborn. and D. Papahadjopoulos. 1979.
Studies on the mechanism of membrane fusion: evidence for an
intermembrane Ca2+-phospholipid complex, synergism with Mg2+, and
inhibition by spectrin. Biochemistry 18:780-790.

Rees, D.A. 1976. Stereochemistry and binding behavior of
carbohydrate chains. MTP (Med. Tech. Publ. Co.) Int. Rev. Sci.
Ser. One Physiol. 5:1-A2.

Rosenthal, K.S., P.B. Swanson. and D.R. Storm. 1976. Disruption

of Escherichia coli outer membrane by EM A9. A new membrane-active

 

peptide antibiotic. Biochemistry 15:5783-5792.
Schindler, M., and M.J. Osborn. 1979. Interaction of divalent
cations and polymyxin B with lipopolysaccharide. Biochemistry

18:AA25-AA30.

A8.

A9.

50.

51.

52.

53.

5A.

55.

116

Schindler, P.R.G., and M. Teuber. 1975. Action of polymyxin B on
bacterial membranes: morphological changes in the cytoplasm and in

the outer membrane of Salmonella typhimurium and Escherichia coli B.

 

 

Antimicrob. Agents Chemother. 8:95-1OA.

Sixl, F., and R.J. Galla. 1982. Calorimetric investigation of
polymyxin binding to phosphatidic acid bilayers. Biochim. Biophys.
Acta. 693:A66—A78.

Smith B.L.. R.L. Hill, I.R. Lehman. R.J. Lefkowitz, P. Handler, and

A. White (eds). In Principles 9: Biochemistry: General Asppcts

 

 

(7th ed.), Chapt. 1A, pp. 289-315. McGraw-Hill, Inc., New York,
1983.

Stan-Latter, H., M. Gupta, and K.E. Sanderson. 1979. The
influence of cations on the permeability of the outer membrane of

Salmonella typhimurium and other gram-negative bacteria. Can. J.

 

Microbial. 25:A75-A85.
Steel, R.C.D., and J.H. Torrie. 1980. Analysis of variance I:
the one-way classification, pp. 137-171. In C. Napier and J.W.

Maisel (eds.), Principles and Procedures pg Statistics: A

 

Biometrical Approach. McGraw-Hill, Inc. New York.

 

Storm, D.R., K.S. Rosenthal, and P.B. Swanson. 1977. Polymyxin
and related peptide antibiotics. Ann. Rev. Biochem. A6:723-763.
Strain, S.M., S.W. Fesik, and I.M. Armitage. 1983. Structure and
metal-binding properties of lipopolysaccharides from heptoseless

mutants of Escherichia coli studied by 13C and 31P nuclear magnetic

 

resonance. J. Biol. Chem. 258:13A66-13A77.
Tabor, C.W. 1962. Stabilization of protoplasts and spheroplasts

by spermine and other polyamines. J. Bacteriol. 83:1101-1111.

56.

57.

58.

59.

60.

61.

62.

63.

117

Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic
transfer of proteins from polyacrylamide gels to nitrocellulose
sheets: procedure and some applications. Proc. Natl. Acad. Sci.
USA 76 A350—A35A.

Vaara, M. 1981. Increased outer membrane resistance to
ethylenediaminetetraacetate and cations in novel lipid A mutants.
J. Bacteriol. 1A8:A26-A3A.

Vaara, M., and T. Vaara. 1983. Polycations as outer
membrane-disorganizing agents. Antimicrob. Agents Chemother.
2A:11A-122.

Vaara, M., and T. Vaara. 1981. Outer membrane permeability
barrier disruption by polymyxin in polymyxin-susceptible and

-resistant Salmonella typhimurium. Antimicrob. Agents Chemother.

 

19:578-583.

Vaara, M., T. Vaara, M. Jensen, I. Helander, M. Nurminen, E. Th.
Rietschel, and P.H. Makela. 1981. Characterization of the
lipopolysaccharide from the polymyxin-resistant pmrA mutants of

Salmonella typhimurium. FEBS Lett. 129:1“5-1A9.

 

Vaara, M., and P. Viljanen. 1985. Binding of polymyxin B
nonapeptide to gram-negative bacteria. Antimicrob. Agents
Chemother. 27:5A8-55A.

Wilkinson, S.G. 1983. Composition and structure of

lipopolysaccharides from Pseudomonas aeruginosa. Rev. Infect. Dis.

 

5:89A1-89A9.
Wilkinson, S.G., and L. Galbraith. 1975. Studies of

lipopolysaccharides from Pseudomonas aeruginosa. Eur. J. Biochem.

 

52:331*3A3.

6A.

650

118

Williams, R.J.P. 1972. A dynamic view of biological membranes.
Physiol. Chem. & Physics A:A27-A39.
Zimelis, V.M., and G.G. Jackson. 1973. Activity of aminoglycoside

antibiotics against Pseudomonas aeruginosa: specificity and site

 

of calcium and magnesium antagonism. J. Infect. Dis. 127:663-669.

CHAPTER IV

CATION INTERACTIONS WITH LIPOPOLYSACCHARIDE FROM PSEUDOMONAS

 

AERUGINOSA PAO STRAINS ALTERED IN O-ANTIGEN CHAIN LENGTHS

 

119

ABSTRACT

The lipopolysaccharide (LPS) isolates from an isogenic set of

antibiotic-resistant mutants of Pseudomonas aeruginosa, altered in

 

O-antigen chain length, were chemically and chromatographically analyzed
to determine heterogeneity in LPS size and phosphate content. Using gel
filtration the LPS samples from strain 503 and PCC A5 separated into
four major LPS populations: peak 1, 2. 2a. and 3. The amount of
highest molecular weight material, peak 1, was decreased in the LPS
isolate of strain PCC A5, but there was an increase in the amount of
peak 2a material. In contrast, the LPS sample from PCC 118 showed only
three major peaks: 2, 2a, and 3, and the intermediate length peak 2a
material was in higher amounts compared to that of strains 503 and PCC
A5. The levels of phosphate in the three LPS isolates were Similar.
These LPS isolates were compared to LPS sample from rough PAO strains
503-18, an aminoglycoside-resistant transductant of PAO 503, and H23A, a
rough mutant of PAO 307. The phosphate content from the 503-18 LPS
isolate was slightly lower than its parent, while the phosphate content
of the LPS from H23A was half that of the other strains. To determine
whether the decrease in the O-antigen length or change in phosphate
content seen in these mutant isolates affected LPS aggregate structure
or polycation binding, the magnesium salts of these LPS samples were
analyzed by electron spin resonance probing. Differences in head-group
motion was observed which appeared to be related to O-antigen length.
Using a cationic electron spin resonance probe, analysis of polymyxin B
binding indicated that the LPS isolate from the rough strain 503-18 had

a higher affinity toward polymyxin B and displaced more CAT12 than did

120

121

the LPS samples from the three smooth strains 503, PCC A5, and PC 118.
In contrast, gentamicin C and spermine each displaced similar amounts of
probe from all the four samples. Titration of the MgLPS from strains
503, 503-18, and H23A with polymyxin B revealed that the alterations in
phosphate levels appeared to affect polymyxin binding. The results
suggest that the presence of the long O-antigen chain and loss of
phosphate both affect LPS aggregate structure and the affinity of
polycations for LPS, but such alterations are not directly correlated
with the decreased permeation of antibiotics across the intact outer

membrane.

INTRODUCTION

The outer membrane of gram-negative bacteria acts as a major
barrier to the uptake of antibiotics (A,1A). A major component of the
outer membrane, the lipopolysaccharide (LPS), appears important in
forming this barrier (33.38). LPS is comprised of two regions: a
hydrophobic lipid A region, which generally contains five to seven fatty
acids linked to diglucosamine. and a hydrophilic polysaccharide portion
(oligosaccharide care i O-antigen polysaccharide) covalently bound to

the lipid A region (33.38). In Pseudomonad species, the O-antigen

 

contains high levels of amino sugars which are usually N-acetylated
(26,56), and the core-lipid A head-group region is highly phosphorylated
(56). Thus, LPS carries a net negative charge, resulting in the high
negative surface charge on gram-negative cells (A7). In the outer
membrane. the LPS associates very tightly with proteins and with
adjacent LPS by divalent cation cross-bridging (30,35). Therefore, the
combination of negative charge and divalent cation cross-bridging of LPS
provides the gram-negative cell surface with a tight barrier important
for the cell's resistance to hydrophobic antibiotics, bile salts,
detergents, protease, lipases, and lysozymes (29.38).

The outer membrane may be considered a hydrophobic bilayer
interspersed with hydrophilic channels and surrounded by a hydrophilic
polysaccharide net (1A,38). Molecules not using a specific outer
membrane transport system, reach their target sites only by one of three
possible outer membrane pathways: the hydrophilic. the hydrophobic, and
the self-promoted pathway (21). The hydrophilic pathway is provided by

aqueous pores that allow the passage of small hydrophilic compounds

122

 

123

(21,38). The second pathway, probably utilized by hydrophobic
antibiotics, is through the hydrophobic domain of the bilayer and is
essentially inoperative in gram negative bacteria. Thus, hydrophobic
antibiotics generally are ineffective against gram-negative bacteria,
except with deep rough mutants in which the LPS is deficient (36) or
with certain antibiotic-supersensitive mutants (A). Finally, a
self-promoted uptake pathway has been postulated for the transport of
polycationic antibiotics, like polymyxin and aminoglycosides, across the
outer membrane of P. aeruginosa (21,22,38). Compounds utilizing this
pathway are thought to displace divalent metal cations destroying the
ionic cross-bridging of LPS and destabilizing outer membrane structure
(21). Thus, such compounds increase outer membrane permeability to
proteins and hydrophobic compounds (21,38). These cationic antibiotics
bind to and perturb the packing arrangement of isolated LPS (2,A1).
suggesting that their site of interaction on the outer membrane is with
LPS. It has been Shown that some mutants of gram negative bacteria,
resistant to low levels of polycationic (AO,53) and B-lactam (15,16)
antibiotics due to low rates of penetration across the outer membrane,
have altered LPS structure. Thus it was proposed that such changes may
affect LPS interactions and thus outer membrane permeability. Although
it is thought that the affinity of cations for LPS depends on the
charges in the core-lipid A region, Peterson 33 21' (A2) have shown that
the length of the O-antigen can affect LPS aggregate structure and
polycation binding.

In this study, we analyzed the aggregate structure of LPS from
strains of P. aeruginosa with altered LPS O-antigen lengths and

phosphate content from strains which have decreased permeability of

 

124

either B-lactams or aminoglycosides across the outer membrane. Cation
interaction with these LPS isolates and with LPS from a rough mutant
which contained lower amounts of phosphates indicated that the length of
the O-antigen and the amount of phosphate in the core-lipid A region
both affected LPS aggregate structure and polycation binding but such
LPS alterations were not always correlated with alterations in

polycation permeability of the outer membrane on the intact cell.

MATERIALS AND METHODS

Bacterial Strains and Growth Media

 

Strain PAO 503 met-9011 is a methionine auxotroph of E. aeruginosa

 

PA01. Strain PAO 503-18, a transductant of PAO 503, was selected for
aminoglycoside resistance (3). PCC 118 and PCC A5 are isogenic B-lactam
resistant mutant and partial revertant strains constructed by mutation
of 503 with ethane methane-sulfonate (15). Confirmation of methionine
auxotropy of these stains was carried out using minimal medium with or
without methionine supplementation (55). Differences in B-lactam
resistance between PCC 118 and PCC A5 strains are partly accounted for
by differences in expression of B-lactamase. PCC A5 does not
constitutively express B-lactamase but continues to express a low-level
of B-lactam resistance and a decreased B-lactam permeability as does PCC
118 (15). H23A is a rough mutant of PAO 307 Egg C-Su, gal-2, FP'
(AK-37,25).

A single colony of the Pseudomonas 503 strains was inoculated into

 

100 ml Mueller Hinton broth and grown using a shaking incubator at 35°C
for approximately 8 h. At this time, 20 ml was inoculated into 2 2 of
Mueller Hinton broth and grown overnight at 35°C in a shaking incubator.
An aliquot of the overnight growth was removed, diluted, and inoculated
in trypticase soya agar for single colony growth to confirm methionine

auxotropy.

Isolation of LPS

 

LPS from g. aeruginosa PA01 strains 503, 503-18, PCC A5, and PCC

 

118 was isolated by the method of Darveau and Hancock (9), followed by

125

126

two extractions in chloroform:methanol (1:1 v/v), resulting in recovery
of approximately 80% of the total LPS. The LPS isolates were dialyzed
extensively against a buffer containing 0.2 M NaCl, 10 mM Tris, 1 mM
EDTA, and 0.01% NaN3, pH 8.0, at 37°C, followed by distilled water. The
magnesium (MgLPS) salts of the LPS isolates were formed by dialysis of
the sample against 10 mM MgClg followed by distilled water. All of the
LPS isolates were dialyzed simultaneously to decrease the variation of
ion content between preparations. Samples were lyophylized and stored
at -20°C. LPS from strain H23A, a gift from R.E.W. Hancock, was

isolated as described previously (“5).

SDS-PAGE and Western Blots

 

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis
(SDS-PAGE) was carried out as described previously (A3), using the
buffer system of Laemmli (28) and the silver-staining procedure of
Dubray and Bezard (11). The separating gel was formed with 15%
acrylamide and 0.1% SDS, with a 7.5% acrylamide stacking gel.

Western blots of SDS-polyacrylamide gels (12% acrylamide) were
prepared as previously described (51). The gels were electrotransferred
with a model TE Transphor Electrophoresis apparatus (Hoefer Scientific
Instruments) at a constant current of 150 mA for 18 h using the
electrode buffer described by Otten gt 21' (39). The nitrocellulose
blots were visualized as described previously (39) with monoclonal

anti-503 antibody (17, titer ~ 1:100,000) diluted in blocking solution.

127

Column Chromatography

 

Samples were fractionated with a Sephadex G200 (Pharmacia Fine
Chemicals) column (6“ cm by 25 mm) at room temperature using the column
buffer system of Peterson and McGroarty (“3). Approximately 30 to A0 mg
of LPS was applied to the column, and 5 ml fractions were collected at a

flow rate of 8 ml per h. All fractionations were done at least twice.

Chemical Analysis

 

Inductively coupled plasma emission spectroscopy of LPS samples was
used to quantitate phosphorus and metal ion content as described
previously (8). Levels of 2-keto-3-deoxyoctulosonic acid (KDO) were
determined by the thiobarbituric acid assay (10). The assay for amino

sugars was performed as described by Catt and Bermad (12).

Partitioning of Spin Probe

 

Electron spin resonance (ESR) spectroscopy was carried out with a
Varian X-band spectrometer (model E-112). Sample temperature was
measured with a thermocouple placed within the cuvette. Titrations of
MgLPS suspended at 10 mg/ml were performed at 37°C by measuring the
spectral parameters of the spin probe A-dodecyldimethylammonium-1-oxyl-
2,2,6,6-tetramethylpiperidine bromide (CAT12; 18:1 molar ratio,
LPS:CAT12) after successive additions of cations to the sample. All of
the cations and antibiotics, as well as the LPS, were dissolved in 50 mM
KOH-HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid; pH 7.0).

Upon the successive addition of cations to samples of LPS, the

spectra were analyzed for 2T”, the hyperfine splitting parameter,

128

whose value is an indication of probe mobility, and for vi, the
partitioning of probe between aqueous (F) and LPS-bound (B)
environments, calculated as described by Coughlin gt §l° (5).
Comparisons of the 2T” values were made using the randomized ANOVA

statistical analysis at P = 0.05 (A8).

Chemicals

Polymyxin B sulfate, gentamicin C sulfate, and spermine were
purchased from Sigma Chemical Co., St. Louis, MO; and CAT12 was
synthesized as previously described (5). All other chemicals were of

reagent grade or better.

RESULTS

In order to analyze the importance of O-antigen chain length on the
structure of LPS, the size heterogeneity of the different isolates was
characterized. Silver staining of LPS from PAO strains 503, PCC A5, and

PCC 118 of P. aeruginosa separated by SDS-PAGE revealed a progressive

 

ladder-like pattern of bands representing LPS molecules containing
increasing lengths of O-antigen (19,23,A5) (Fig. 1A, lanes 1-3). The
intensity of staining indicated three to four regions of bands
representing as many as four populations of molecules differing in
O-antigen length. The average length of the highest molecular weight
LPS (set 1) of strain PCC A5 and PCC 118 appeared shorter than that of
strain 503 (Fig. 1A, lanes 3, 2, and 1, respectively), as previously
observed (15). On the other hand, the LPS isolate from the rough
strains, 503—18 and H23A, lacked bands sets 1, 2, and 2a (Fig. 1A, lane
A and 5). When low amounts of the different LPS isolates were applied
to the gel, only the fastest migrating bands were stained, and there was
no difference in the migration pattern of LPS from strain 503, PCC 118,
and PCC A5 (Fig. 1A, lanes 6, 7, and 8, respectively). In contrast, the
low molecular weight band from strain 503-18 and H23A migrated faster

than that of the other P. aeruginosa strains (Fig. 1A, lanes 9 and 10)

 

indicating an altered, presumably deficient, core region (3). The
Western blots of LPS isolates using anti-O-antigen specific monoclonal
antibody revealed a regular ladder pattern of 0-antigen-containing LPS
from strains 503, PCC A5, and PCC 118 (Fig. 1B, lanes 11, 13, and 1A;
respectively). Furthermore, the ladder pattern consisted of doublet

bands detected in approximately equal amounts; it has been suggested

129

Figure 1.

130

(A) Silver-stained SDS-PAGE of LPS from P. aeruginosa strain
PAO 503 (lanes 1 and 6); PCC 118 (lanes 2 and 7); PCC A5
(lanes 3 and 8); FAQ 503-18 (lanes A and 9); and H23A (lanes
5 and 10). Samples of either 5 ug (lanes 1-3), 10 ug (lanes
A and 5), 0.1 pg (lanes 6-8), or 0.25 ug (lanes 9 and 10)
were applied to a 15% polyacrylamide gel which had
polymerized overnight with a butanol overlay. Arrows
indicate the four intensively stained regions of the P.

aeruginosa samples: band peaks 1, 2, 23, and 3. (B) Western

 

blots of LPS from P. aeruginosa strains PAO 503 (lane 11, PAC

 

503-18 (lane 12), PCC A5 (lane 13), and PCC 118 (lane 1A)
reacted with monoclonal anti-503 antibody. Samples of 2.5 ug
were applied to a 12% acrylamide gel which had been
polymerized overnight with a butanol overlay. The gel was

blotted as described in Materials and Methods.

t.
1
3
l

3 Q N— =
m

 

132

that these doublets represent substoichiometric modification within the
core-lipid A region of the molecules (A5). Interestingly, the LPS
isolate from strain 503-18 reacted very weakly with the anti-O-antigen
monoclonal antibody suggesting that the mutation is leaky (Fig. 1B, lane
12).

To better quantitate the size heterogeneity of the LPS isolates
from strains 503, PCC A5, and PCC 118, the samples were separated on a
Sephadex G200 column. The elution profiles showed four major amino
sugar-containing peaks for strains 503 and FCC A5 (Fig. 2, A and C,
respectively), and three major peaks for strain PCC 118 (Fig. 2B). The
relative amount of material in peak 1 was decreased in the LPS isolate
of strain PCC A5 and was in very low amounts in the LPS isolate from PCC
118 (Table 1, Fig. 2, C and B, respectively). The level of material in
peak 2 was also lower in the isolates of both PCC 118 and PCC A5
compared to that from the 503 LPS isolate (Table 1, Fig. 2A), and peak
2a was significantly increased in the PCC 118 and PCC A5 isolates (Table
1). These observations are consistent with the decrease in the high
molecular weight bands noted in the SDS gels (Fig. 1A). Even though the
data in Table 1 show differences in the percentage of amounts in peak 1
and peak 2a between PCC A5 and PCC 118, the overall average molecular
weight is similar for these two LPS isolates and significantly different
from that of PA0503.

It has been shown that gel permeation chromatography and SDS-PAGE
separate different sized molecules (A3,A5). Therefore, a diagonal
banding pattern should be expected across SDS-PAGE gels of column
fractions when applied in the order of elution. As reported before for

several P. aeruginosa PAO strains (A5), SDS-PAGE of the column fractions

 

Figure 2.

133

Fractionation of LPS from P. aeruginosa strains (A) PAO 503,

 

(B) PCC 118, and (C) PCC A5 on Sephadex 0200. Fractions were
analyzed for KDO (o) and amino sugar (0). Silver-stained
SDS-polyacrylamide gel of column fractions from strain PCC A5
is aligned under the appropriate fractions. (A) represents

the slow moving set of bands and (B) the faster moving set.

131+

 

>350 mecca olo

4 3

_. O.|

 

 

2 I.
O O O
— — 4 q

.106

~0.4

_02
0

 

 

 

 

 

 

$33 0.. Cox

 

Fraction Number

 

135

Table 1. Percent KDO Recovery in Peaks 1, 2, 2a, and 3 from
P. aeruginosa LPS Fractionated on Sephadex G200.a

 

 

 

 

Peak No. 503 PCC-A5 PCC-118
1 11.2 1.8 0.2
2 5.1 0.8 1.1
2a 3.1 7.1 13.3
3 87.5 86.3 8A.0
Ave. Mol. Wt.b 9000 6275 5901

 

aPercentage of total molar amounts in the four peaks.

bAverage molecular weights for the LPS isolate of each strain was
determined by assuming an approximate molecular weight of 70,000 for
peak 1; 35,000 for peak 2; 12,000 for peak 2a; and A,500 for peak 3.

136

of samples from strain 503, PCC A5, and PCC 118 revealed the two
distinct ladder patterns, the A bands (later eluting ladder) and the B
bands (earlier eluting ladder) which have been shown to occur in many
PAO strains (A5) (see Fig. 2 for the SDS-PAGE elution profile of strain
PCC A5).

To quantitate the phosphate content and to define the salt forms of
the LPS isolates, the metal and phosphate contents of the four PAO
isolates were determined by inductively coupled plasma emission
spectroscopy (ICP, Table 2). Variability in the P/LPS ratio observed in
LPS isolates is likely due to differences in growth conditions and/or
the LPS isolation procedure. Repeated analyses of samples from a given
strain have shown that a difference of 2 in the P/LPS ratio among the
LPS isolates is significant (P=0.05, unpublished). 0n the other hand,
the Mg/P ratio was less variable and remained fairly constant from
sample to sample. No difference was detected in the levels of phosphate
in the LPS from 503, PCC A5, and PCC 118, but strain 503-18 had
approximately 1.5 fewer phosphates (Table 2). The LPS from the rough
mutant strain H23A contained A.5 phosphate per LPS and 0.9 magnesium
ions per phosphate (data not shown). The LPS from strains 503-18, PCC
A5, and PCC 118 had significantly lower levels of magnesium ions per
phosphate than did LPS from the parental strain (PAO 503). These
results indicate differences in Mg2+ binding to the LPS of the
transductant (PAO 503-18) and mutant strains (PCC A5 and PCC 118) as
compared with the LPS of the parental strain.

To measure polycation binding to the LPS isolates, an ESR probe,
CAT12, was used. Titration of LPS samples, containing CAT12, with

various cations displaced different amounts of the probe as reflected in

137

Table 2. Elemental Composition of Purified P. aeruginosa MgLPS Isolatesa

 

 

 

Strain Mg/Pb +/Pb:0 Mg/LPSb’d P/LPSb'd
503 0.87 1.87 7.31 8.u
503—18 0.7u 1.62 5.07 6.8
PCC-A5 0.70 1.50 6.01 8.6
PCC-118 0.67 1.A2 5.69 8.5

 

8The metal and phosphate content of the magnesium salts from the four
LPS isolates were determined by inductively coupled plasma emission
spectroscopy (ICP). There was less than 0.01 metal per phosphate
(mol/mol) of Ca, Fe, Al, and Zn ions associated with the samples.

bValues are expressed in mol/mol.
CTotal metal cation charges bound per phosphate.

dAssumes an average molecular weight reported in Table 1, and of A500
for the LPS isolate of strain 503~18.

 

138

the partitioning parameter W1. We have shown that the standard
deviation of W1 for titrations of independent isolates is i 0.0A and is
approximately constant throughout the titration (Rocque, W.J., S.W.
Fesik, A. Haug and E.J. McGroarty, in press). The addition of polymyxin
B displaced high levels of CAT12 from the LPS sample, as indicated by
the large increase in W1; these results suggest that this antibiotic has
a high affinity for LPS compared to gentamicin C, spermine, and Ca2+
(Fig. 3). The LPS isolate from strain 503-18 was shown to have a higher
affinity and level of binding of polymyxin B than did the LPS samples
from strain 503, PCC A5, and PCC 118 (Fig. 3A), even though the LPS from
strain 503-18 had approximately 1.5 less phosphates. The data suggest
that the presence of outer core and long O-antigen decreases polymyxin B
binding to the LPS core region. Gentamicin C and spermine each
displaced probe in a similar manner for all samples (Fig. 3, B and C).
Interestingly, the LPS isolate from strain 503 showed a decrease
displacement of CAT12 by Ca2+ compared to the other isolates which may
reflect Ca2+ interacting with the long O-antigen chains in addition to
the CAT12 binding site.

To explore the possibility that phosphate levels in the core-lipid
A affect polymyxin binding, LPS was isolated from a rough PAO mutant of

P. aeruginosa which had low phosphate levels (A.5 phosphate per LPS).

 

Titration of polymyxin B onto the MgLPS complexes from strain 503,
503-18, and H23A revealed that decreased phosphate levels on the H23A
isolate correlated with decreased polymyxin binding (Fig. A). However,
the 503—18 isolate, as previously stated, had a higher affinity than the

503 isolate, presumably due to the lack of O-antigen chain and outer

Figure 3.

139

The partitioning parameter Vi measured as a function of added
(A) polymyxin B, (B) gentamicin C, (C) spermine, and (D)
CaClZ. Partitioning of the spin probe CAT12 was determined

upon addition of cations to MgLPS from P. aeruginosa strains

 

PAO 503 (o), PAO 503-18 (0), PCC A5 (A), and PCC 118 (o).
The amount of added cation is reported as cation per LPS

(mol/mol).

140

 

0.8..

0.4 (-

H
p

l

l

 

.0-
.1
A

L6,

12,.

04- A

 

A l l
I

 

081.

0.4 ..

 

 

P

 

 

“x

__$_9___¢

03 L0
CHM/LPS

Figure A.

141

The partitioning parameter Vi measured as a function of added
polymyxin B. Partitioning of the spin probe CAT12 was
determined upon addition of polymyxin B to MgLPS from P.

aeruginosa strains PAO 503 (0), FAQ 503-18 (A), and H23A (o).

 

The amount of added polymyxin B is reported as polymyxin B

per LPS (mol/mol).

142

O.

8.5 5:528;
on

0..

 

 

_

 

 

 

v.0

md

S,

N._

0..

143

core components. Such outer core residues may be important in
preventing polymyxin binding.

The aggregate structure of the isolates was measured at 37°C by
calculating the hyperfine splitting parameter, 2T” of the bound CAT12.
This parameter is related to the rotational mobility of the spin label
and, therefore, reports the local motion within the LPS head groups.
High values of 2Tfl reflects low motion (rigidification). A
significant difference (P=0.05) was observed in head-group motion of
MgLPS complexes between isolates of 503, and of 503-18, and PCC 118
suggesting differences in LPS aggregate packing. The average 2T¢ (in
gauss) for the MgLPS isolated from each of the strains were: 60.A i 0.6
for strain 503. 59.A i 0.6 for H23A, 59.2 1 0.7 for PCC A5, 58.2 1 0.6
for PCC 118, and 57.6 i 1.1 for 503-18. The data suggest that the MgLPS
aggregate from strain 503 was significantly more rigid than the MgLPS
aggregates from the resistant strains PCC 118 and 503-18. A similar
trend was also noted comparing samples from strains 503 and PCC A5.

This elevated rigidity presumably reflects the increased levels of Mg2+

bound per phosphate in the 503 LPS sample (see Table 2).

DISCUSSION

The LPS isolates of P. aeruginosa strains 503, PCC A5, and PCC 118,

 

shown to be heterogeneous when separated by SDS-PAGE (Fig. 1A),
contained different levels of the high-molecular weight components (Fig.
2, Table 1). Godfrey and Bryan (15) have reported a lower level of the
O-antigen-containing molecules in the LPS isolates from strains PCC A5
and PCC 118 compared to that from PAO 503. In the results presented
here, the LPS molecules of different sizes were partially separated on
Sephadex 0200 in the presence of deoxycholate and EDTA (A3,A5). and the
levels of the different size populations were quantitated. Three or
four major populations of LPS were resolved as detected by sugar
analysis representing three or four populations of molecules with
O-antigens of different lengths. We found that the LPS from strains PCC
118 and PCC A5 had decreased amounts in peak 1 material (very long
O-containing chain LPS molecules) and increased relative amounts of peak
2a material (intermediate 0-antigen-length LPS) compared to the LPS
isolates from strain 503 (Fig. 2; Table 1). These differences in the
level of molecules with different O-antigen lengths are reflected in
their average molecular weights (Table 1). The results show that the
LPS from PCC 118 and PCC A5 is composed mainly of LPS molecules with
intermediate and short chain O-antigen. Interestingly, all three of the
smooth strains appeared to contain an antigenically distinct A band
population, an unusual type of LPS lacking phosphate (A5). However,
this A band material, which has been shown to constitute approximately
10-15% of the LPS on a weight basis (A5), did not appear to vary in

amounts among the isolates from strain 503, PCC 118 and PCC A5. The

144

145

data presented here suggest that there is a variation in the length of
LPS molecules bearing the serotype specific O-antigen and this variation
might be important in the barrier function of the outer membrane, since
the mutant strains have a decreased permeability to B-lactams.

The LPS isolates from two other PAO mutants from P. aeruginosa

 

completely lacking an 0-antigen chain were also studied (strains 503-18
and H23A). Their electrophoretic mobilities on SDS-PAGE were greater
than any of the bands from the smooth strains suggesting that the LPS of
these rough mutants also lacked outer core sugars (Fig. 1A). Bryan and
coworkers (3) reported that the LPS from strain PAO 503-18 contained
less core component sugars (glucose, rhamnose, and glucosamine) and an
increase in the inner core components (KDO, alanine, and galactosamine)
compared to wild-type isolates. It is expected that any decrease in the
overall negative charge of the core-lipid A region would alter
polycation binding. Since the 503-18 strain is resistant to the
polycationic aminoglycosides, the loss of outer core may in the intact
outer membrane, result in decreased polycation binding and uptake by the
self-promoted pathway. As shown in Table 2, the three smooth LPS
isolates from strains 503, PCC A5, and PCC 118 had similar levels of
phosphate per LPS, whereas the LPS isolate from strain 503-18 had 1.5
fewer phosphate. The LPS from the H23A strain had dramatically lower
levels of phosphate concomitant with a lower affinity for polymyxin.

In previous studies we have shown that the ability of polycationic
antibiotics to displace CAT12 from LPS was correlated with the ability
of the antibiotic to increase outer membrane permeability (AO,A1). In
this study, we found that polymyxin B displaced less CAT12 from the

MgLPS complexes of the three smooth strains, PAO 503, PCC A5, and PCC

146

118, than from the MgLPS complex of the rough strain PAO 503—18 (Fig.
3A). Although the 503-18 LPS isolate had a slight decrease in phosphate
groups compared to the smooth LPS isolates (Table 2), the lower level of
polymyxin B binding to the smooth strain isolates suggests that the
presence of the O-antigen or sugars in the outer core prevent polymyxin
binding. This decreased affinity may result from interactions between
0-polysaccharides forming a network mesh which excludes large molecules
(AA). It has been shown that LPS is a highly ordered, almost
crystalline structure, and such tight interactions presumably are
critical in preventing outer membrane penetration (27,37). Polymyxin B
is a rather large amphipathic molecule with a molecular weight of
approximately 1200 (A9). Thus, the hydrophilic O-antigen side chain
attached to the LPS core may preclude polymyxin penetration of this
layer decreasing the amount that can reach the surface of the outer
membrane. Peterson 32 31. (A2) have demonstrated that binding of
polymyxin B to long-chain fractions and unseparated LPS of E. coli was
less than to the short-chain fraction, suggesting that the long
O-antigen masks anionic sites on LPS. Other investigators have
demonstrated that the O-polymer on LPS sterically hinders access of
large protein molecules to the outer membrane (18,20,2A,A6). Such
results indicate that the presence of a long O-antigen may be critical
in forming a barrier against large compounds.

In contrast, the smaller hydrophilic molecules, gentamicin C and
spermine (A60 and 202 molecular weight, respectively) had similar

affinities toward the smooth and the rough LPS isolates of P. aeruginosa

 

including the isolate from the aminoglycoside-resistant strain (Fig. 3,

B and C). Perhaps such molecules diffuse freely through the

147

O-polysaccharide matrix. Interestingly, when the LPS aggregates were
titrated with Ca2+, the isolate from the parent strain (PAO 503) showed
a lower affinity for Ca2+ than the other isolates (Fig. 30). Since the
O-antigen, containing potential Ca2+ binding sites as mannuronic acid
derivatives, is longer on the isolate of strain 503 compared to that of
strains PCC A5 and PCC 118 (3,15). the parental isolate may have more
binding sites for Ca2+ (AA). The parental isolate may bind Ca2+ to the
O-polymer at low concentrations and to the core-lipid A only at higher
concentrations. The MgLPS sample from strain 503 had a slightly higher
amount of Mg2+ bound per phosphate, possibly reflecting the small amount
of metal ion bound to the very long carbohydrate chain (AA). Magnesium
binding to O-polymers of E. coli has been detected previously (A2).

It was shown that the large decrease in phosphates in the
core-lipid A region of the H23A isolate resulted in a decrease in
polymyxin binding (Fig. A). Others have reported that LPS containing
lower negative charges had lower affinities for polycations
(3A,A0,50,52,53,5A,56). However, we found that polycation binding was
affected not only by the number of phosphate moieties, but also by the
presence and length of O-antigen side chain. Furthermore, we found that
head-group motion was different among the LPS isolates suggesting
differences in LPS aggregate packing. The motion within LPS aggregates
is known to depend upon ionic composition and pH (6,7). We found that
the MgLPS aggregate from strain 503-18 was more fluid compared to the
other MgLPS complexes. This increased motion could partially account
for the greater binding and penetration of polymyxin into the 503-18
isolate despite the small decrease in phosphate content. 0n the other

hand, the relatively high rigidity of the LPS from the H23A strain

148

combined with its significantly reduced phosphate content might explain
the lower polymyxin, penetration and binding. The MgLPS aggregate from
strain PCC 118 was intermediate in its head group motion when compared
to strains 503 and 503-18. The differences observed in the head group
motion of these aggregates may result from differences in the length of
the attached O-antigen. Peterson gt gt. (A2) reported that above 25°C
the head group motion of unseparated LPS from a smooth strain E. ggtt
was less than that of the short-chain fraction. Such differences in
aggregate packing may affect polycation association. Lounatmaa and
co-workers (31,32) reported that the effects of polymyxin on cell
surface morphology were dependent on the aggregate structure of LPS.
Gilleland and Farley (13) proposed that adaptative polymyxin resistance

which can occur in P. aeruginosa cultures grown in the presence of

 

polymyxin occurs by altering the aggregate nature of the outer membrane
which results in polymyxin exclusion.

Although LPS is important in determining outer membrane
permeability, outer membrane proteins can also play an important role in
the permeability barrier. Nikaido (36) proposed that hydrophilic
antibiotics diffuse through aqueous pores produced by the porin, whereas
hydrophobic antibiotics utilize a hydrophobic pathway, dissolving into

the hydrocarbon phase of the outer membrane. The E. aeruginosa strains

 

PCC 118 is resistant to B-lactam antibiotics while strain PCC A5 is a
partial revertant of this phenotype (15). No significant differences
were found in the outer membrane protein pattern of these two strains
but the permeability of both mutant cells to B-lactams was reduced
compared to the susceptible parent (15). Godfrey gt gt. (16) suggested

that changes in the composition of the LPS is correlated with decreased

149

permeability and increased resistance. Our data show that PCC A5 and
PCC 118 LPS have an altered SDS-PAGE pattern and a marked difference in
the size heterogeneity compared to that from strain 503 (Fig. 1 and 2;
Table 1). The LPS isolates from strains PCC 118 and PCC A5 showed
higher amounts of the peak 2a fractions, and this difference may be
important in outer membrane structure. In addition, we found that the
LPS phosphate content was not altered in the B-lactam resistance strain.
Thus, alteration in LPS O-antigen length may influence the opened state
of the outer membrane pores. Angus gt gt. (1) and Kropinski and
coworkers (26) in their analysis of the outer membrane permeability and

composition of an antibiotic supersusceptible mutant from g. aeruginosa

 

reported an increase in outer membrane permeability concomitant with an
alteration in LPS core, perhaps affecting outer membrane protein
structure. Yamada and Mizushima (57), using X-ray diffraction,
demonstrated that both the lipid A moiety and the polysaccharide moiety
of LPS interact with the 0-8 porin trimer of g. ggtt. Furthermore, they
showed that the core-oligosaccharide moiety is involved in stabilizing
the proper porin lattice conformation (57). Whether alterations in LPS
O-antigen length associated with aminoglycoside resistance can be
related to changes with cation-LPS interactions is questionable. As has
been noted, the isolate from the aminoglycoside resistant mutant,
503-18, has an altered affinity for polymyxin B, but not for gentamicin.
Thus, the aminoglycoside resistance of the 503-18 strain (3) is not
likely to be due to changes in direct interactions of the modified LPS
with aminoglycosides.

In summary, we have shown that the presence of a long O-antigen

chain and diminished LPS phosphate levels can decrease the binding of

150

amphipathic polycations to the aggregate. We also observed that such
changes in LPS structure can affect LPS head group motion or aggregate
packing. However such alterations in LPS structure and LPS-cation

interaction are not always correlated with changes in outer membrane

permeability.

LITERATURE CITED

Angus, B.L., A.M. Carey, D.A. Caron, A.M.B. Kropinski, and R.E.W.

Hancock. 1982. Outer membrane permeability in Pseudomonas

 

aeruginosa: comparison of a wild-type with an antibiotic

 

supersusceptible mutant. Antimicrob. Agents Chemother. 21:299-309.
Bader, J., and M. Teuber. 1973. Action of polymyxin B on
bacterial membranes. I. Binding to the O-antigenic

lipopolysaccharide of Salmonella typhimurium. Z. Naturforsch. Teil

 

C 28:A22-A30.
Bryan, L.E., K. O'Hara, and S. Wong. 198A. Lipopolysaccharide
changes in impermeability-type aminoglycoside resistance in

Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 26:250-255.

 

Chopra, 1., and P. Ball. 1982. In A.H. Rose and J.G. Morris

(eds.), Advances tg Microbial Physiology, pp. 183-2A1. Academic

 

Press, London.

Coughlin, R.T., C.R. Caldwell, A. Haug, and E.J. McGroarty. 1981.
A cationic electron spin resonance probe used to analyze cation
interactions with lipopolysaccharides. Biochem. Biophys. Res.
Commun. 100:1137-11A2.

Coughlin, R.T., A. Haug, and E.J. McGroarty. 1983. Physical
properties of defined lipopolysaccharide salts. Biochemistry
22:2007-2013.

Coughlin, R.T., A.A. Peterson, A. Haug, H.J. Pownall, and E.J.
McGroarty. 1985. A pH titration study on the ionic bridging
within lipopolysaccharide aggregates. Biochem. Biophys. Acta

821:AOA-A12.

151

10.

11.

12.

13.

111.

15.

152

Coughlin, R.T., S. Tosanger, and E.J. McGroarty. 1983.
Quantitation of metal cations bound to membranes and extracted

lipopolysaccharide of Escherichia coli. Biochemistry 22:2002-2006.

 

Darveau, R.P., and R.E.W. Hancock. 1983. Procedure for isolation
of bacterial lipopolysaccharides from both smooth and rough

Pseudomonas aeruginosa and Salmonella typhimurium strains. J.

 

 

Bacteriol. 155:831-838.

Droge, W., V. Lehmann, O. Lfideritz, and O. Westphal. 1970.
Structural investigations on the 2-keto-3-deoxy-octonate region of
lipopolysaccharides. Eur. J. Biochem. 1A:175-18A.

Dubray, G., and G. Bezard. 1982. A highly sensitive periodic
acid-silver stain for 1,2-diol groups of glycoproteins and
polysaccharides in polyacrylamide gels. Anal. Biochem.
119:325-329.

Gatt, R., and E.R. Bermad. 1966. A rapid procedure for the
estimation of amino sugars on a microscale. Anal. Biochem.
15:167-171.

Gilleland, H.E., Jr., and L.E. Farley. 1982. Adaptive resistance

to polymyxin in Pseudomonas aeruginosa due to an outer membrane

 

impermeability mechanism. Can. J. Microbiol. 28:830-8A0.
Godfrey, A.J., and L.E. Bryan. 198A. Intrinsic resistance and
whole cell factors contributing to antibiotic resistance, pp.

113-1A5. In L.E. Bryan (ed.) Antimicrobial Drug Resistance.

 

 

Academic Press, Inc., Orlando, FL.

Godfrey, A.J., and L.E. Bryan. 198A. Resistance of Pseudomonas

 

aeruginosa to new B-lactamase-resistant B-lactams. Antimicrob.

 

Agents Chemother. 26:A85-A88.

16.

17.

18.

19.

20.

21.

22.

153

Godfrey, A.J., L. Hatlelid, and L.E. Bryan. 198A. Correlation
between lipopolysaccharide structure and permeability resistance in

B-lactam-resistant Pseudomonas aeruginosa. Antimicrob. Agents

 

Chemother. 26:181-186.
Godfrey, A.J., M.S. Shahrabadi, and L.E. Bryan. 1986.
Distribution of porin and lipopolysaccharide antigens on a

Pseudomonas aeruginosa permeability mutant. Antimicrob. Agents

 

Chemother. 30:802-805.
Goldman, R.C., K. Joiner, and L. Leive. 198A. Serum-resistant

mutants of Escherichia coli 0111 contain increased

 

lipopolysaccharide, lack an O-antigen-containing capsule, and cover
more of their lipid A core with O-antigen. J. Bacteriol.
159:877-882.

Goldman, R.C., and L. Leive. 1980. Heterogeneity of

antigenic-side chain length in lipopolysaccharide from Escherichia

 

coli 0111 and Salmonella typhimurium LT2. Eur. J. Biochem.

 

107:1A5-153.
Grossman, N., M.A. Schmetz, J. Foulds, E.N. Klima, V. Jimenez, L.L.
Leive, and K.A. Joiner. 1987. Lipopolysaccharide size and

distribution determine serum resistance in Salmonella montevideo.

 

J. Bacteriol. 169:856-863.

Hancock, R.E.W. 198A. Alterations in outer membrane permeability.
Ann. Rev. Microbiol.. 38:237-26A.

Hancock, R.E.W., V.J. Raffle, and T.I. Nicas. 1981. Involvement
of the outer membrane in gentamicin and streptomycin uptake and
killing in Pseudomonas aeruginosa. Antimicrob. Agents Chemother.

19:777-785.

7-

23.

2A.

25.

26.

27.

28.

29.

154

Hitchcock, P.J.. and T.M. Brown. 1983. Morphological

heterogeneity among Salmonella lipopolysaccharide chemotypes in

 

silver-stained polyacrylamide gels. J. Bacteriol. 15A:269-277.
Joiner, K.A., N. Grossman, M. Schemtz, and L. Leive. 1986. C3
binds preferentially to long-chain lipopolysaccharide during

alternative pathway activation by Salmonella montevideo. J.

 

Immunol. 136:710-715.
Kropinski, A.M., L.C. Chan and F.H. Milazzo. 1979. The extraction

and analysis of lipopolysaccharides from Pseudomonas aergginosa

 

strain PAO, and three rough mutants. Can. J. Microbiol.
25:390-398.

Kropinski, A.M., J. Kuzio, B.L. Angus, and R.E.W. Hancock. 1982.
Chemical and chromatographic analysis of lipopolysaccharide from an

antibiotic-supersusceptible mutant of Pseudomonas aeruginosa.

 

Antimicrob. Agents Chemother. 21:310-319.

Labischinski, H., G. Barnickel, H. Bradaczek, D. Naumann, E.T.
Reitschel and P. Giesbrecht. 1985. High state of order of
isolated bacterial lipopolysaccharide and its possible contribution
to the permeation barrier property of the outer membrane. J.
Bacteriol. 162:9-20.

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

Leive, L. 197A. The barrier function of the gram-negative

envelope. Ann. N.Y. Acad. Sci. 235:109-127.

30.

31.

32.

33.

3A.

35.

36.

155

Leive, L., V.K. Shovlin, and S.E. Mergenhagen. 1968. Physical,
chemical, and immunological properties of LPS released from E. ggtt
by ethylenediaminetetraacetate. J. Biol. Chem. 2A3:638A-6391.
Lounatmaa, K., P.H. Makela, and M. Sarvas. 1976. Effect of
polymyxin on the ultrastructure of the outer membrane of wild-type

and polymyxin-resistant strains of Salmonella. J. Bacteriol.

 

127:1A00-1A07.

Lounatmaa, K., M. Sarvas, and P.H. Makela. 1975.
Ultrastructural effects of polymyxin on the outer membrane of
resistant mutants of Salmonella. J. Ultrastruct. Res. 50:391.
LUderitz, 0., M.A. Freudenberg, C. Galanos, V. Lehmann, E. Th.
Rietschel, and D.H. Shaw. 1982. Lipopolysaccharides of
gram-negative bacteria, pp. 79-151. In S. Razin and S. Rottem

(eds.), Current Topics tg Membranes and Transport, Vol. 17:

 

 

Membrane Lipids of Prokaryotes. Academic Press, New York.
Moore, R.A., and R.E.W. Hancock. 1986. Involvement of outer

membrane of Pseudomonas cepacia in aminoglycoside and polymyxin

 

resistance. Antimicrob. Agents Chemother. 30:923-926.
Mutoh, N., M. Furukawa, and S. Mizushima. 1978. Role of

lipopolysaccharide and outer membrane protein of Escherichia coli

 

K-12 in the receptor activity for bacteriophage TA. J. Bacteriol.
136:693-699.

Nikaido, H. 1976. Outer membranes of Salmonella typhimurium.

 

Transmembrane diffusion of some hydrophobic substances. Biochem.

Biophys. Acta A33:118-132.

37.

38.

39.

A0.

A1.

A2.

A3.

1111.

156

Niikaido, H., Y. Takeuchi, S. Ohnishi and T. Nakae. 1977. Outer

membrane of Salmonella typhimurium. Electron spin resonance

 

studies, Biochim. Biophys. Acta A65:152-16A.

Nikaido, H., and M. Vaara. 1985. Molecular basis of bacterial
outer membrane permeability. Microbiol. Rev. A9:1-32.

Otten, S., S. Iyer, W. Johnson, and R. Montgomery. 1986.

Serospecific antigens of Legionella pneumophila. J. Bacteriol.

 

167:893-9OA.
Peterson, A.A., S.W. Fesik, and E.J. McGroarty. 1987. Decreased
binding of antibiotics to lipopolysaccharides from

polymyxin-resistant strains of Escherichia coli and Salmonella

 

 

typhimurium. Antimicrob. Agents Chemother. 31:230-237.

 

Peterson, A.A., R.E.W Hancock, and E.J. McGroarty. 1985. Binding
of polycationic antibiotics and polyamines to lipopolysaccharides

of Pseudomonas aeruginosa. J. Bacteriol. 16A:1256-1261.

 

Peterson, A.A., A. Haug, and E.J. McGroarty. 1986. Physical
properties of short- and long-O-antigen-containing fractions of

lipopolysaccharide from Escherichia coli 0111:BA. J. Bacteriol.

 

165:116-122.
Peterson, A.A., and E.J. McGroarty. 1985. High-molecular-weight

components in lipopolysaccharides of Salmonella typhimurium,

 

Salmonella minnesota, and Escherichia coli. J. Bacteriol.

 

 

162:738-7A5.
Rees, D.A. 1976. Stereochemistry and binding behavior of
carbohydrate chains. MTP (Med. Tech. Publ. Co.) Int. Rev. Sci.

Ser. One Physiol. 5:1-A2.

“5.

A6.

A7.

A8.

A9.

50.

51.

52.

157

Rivera, M., L.E. Bryan, R.E.W. Hancock, and E.J. McGroarty. 1987.

Heterogeneity of lipopolysaccharide from Pseudomonas aeruginosa.

 

Analysis of lipopolysaccharide chain length by gel filtration and
SDS-PAGE. J. Bacteriol., (in press.).

Sanderson, K.E., T. MacAlister, and J.W. Costerton. 197A.
Permeability of lipopolysaccharide-deficient (rough) mutants of

Salmonella typhimurium to antibiotics, lysozyme, and other agents.

 

Can. J. Microbiol. 20:1135-11A5.
Sherbert, G.V. and M.S. Lakshmi. 1973. Characterization of

Escherichia coli cell surface by isoelectric equilibrium analysis.

 

Biochim. Biophys. Acta 298:50-58.

Steel, R.C.D., and J.H. Torrie. 1980. Analysis of variance I: the

one-way classification. pp. 137-171. In Princtptes and Procedures

 

gt Statistics (2nd Ed.) McGraw-Hill, Inc., New York.

Storm, D.R., R.S. Rosenthal, and P.E. Swanson. 1977. Polymyxin
and related peptide antibiotics. Ann. Rev. Biochem. A6:723-763.
Teuber. M. 1969. Susceptibility to polymyxin B of pencillin

G-induced Proteus mirabilis L forms and spheroplasts. J.

 

Bacteriol. 98:3A7-350.

Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic
transfer of proteins from polyacrylamide gels to nitrocellulose
sheets: procedure and some applications. Proc. Natl. Acad. Sci.
USA 76:A350-A35A.

Vaara, M. 1981. Increased outer membrane resistance to

ethylenediamine- tetraacetate and cations in novel lipid A mutants.

J. Bacteriol. 1A8:A26-A3A.

 

 

 

53.

5A.

55.

56.

57.

158

Vaara, M., T. Vaara, M. Jensen, I. Helander, M. Nurminen, E. Th.
Rietschel, and P.H. Makela. 1981. Characterization of the
lipopolysaccharide from the polymyxin-resistant gmr A mutants of

Salmonella typhimurium. FEBS Lett. 129:1A5-1A9.

 

Vaara, M., T. Vaara, and M. Sarvas. 1979. Decreased binding of

polymyxin by polymyxin-resistant mutants of Salmonella typhimurium.

 

J. Bacteriol. 139:66A-667.

Vogel, H.J., and Bonner, D.M. 1956. Acetyl ornithinase of E.
ggtt: Partial purification and some properties. J. Biol. Chem.
218:97-106.

Wilkinson, S.G., L. Galbraith, and L.G. Lightfoot. 1973. Cell
walls, lipids, and lipopolysaccharides of Pseudomonas species.
Eur. J. Biochem. 33:158-17A.

Yamada, H., and S. Mizushima. 1980. Interaction between major

outer membrane protein (0-8) and lipopolysaccharide in Escherichia

 

coli K12. Eur. J. Biochem. 103:209-218.

 

CHAPTER V: SUMMARY

The results of this project indicate that PA01 strains from

Pseudomonas aeruginosa are capable of synthesizing more than one type of

 

LPS-like molecule differing in their antigenic reactivities. Although
the LPS isolate from the different strains used in this study was shown
to be heterogeneous on SDS-PAGE, this method by itself does not have the
power to predict the presence of more than one type of LPS-like molecule
with a different 0 polymer. However, by combining SDS-PAGE with gel
permeation chromatography, Western blots, and sugar analysis, it was
possible to distinguish chemically distinct subclasses of molecules from
individual strains. It was also demonstrated that the fraction of core
oligosaccharides carrying the O-specific polymer is less than 8%.

The affinity of cationic antibiotics for E. aeruginosa LPS was

 

characterized using the magnesium (MgLPS) and calcium (CaLPS) salts from
an antibiotic sensitive, a revertant, and parent strain. As has been
reported, I found that polycationic antibiotics bind to and rigidify

purified LPS isolates of E. aeruginosa. The nature of the antibiotic

 

binding to LPS is probably not solely an ionic attraction between LPS
and the polycation. The affinity of specific cations for LPS also
appears to depend on the number of arrangement of the charges in the
antibiotic molecule as well as the presence of hydrophobic groups. The
decreased LPS motion, upon antibiotic binding indicated a cation-induced
alteration of the aggregate packing arrangement which may also occur in
the intact membrane perturbing membrane integrity and increasing outer

membrane permeability.

159

160

The differences in antibiotic susceptibility of the three strains
mentioned above may result from differences that we detected in the
substoichiometric modifications of the LPS, presumably in the core-lipid
A region. This difference in the substoichiometric modification may
affect cation affinities by modifying the conformation of the aggregates.
The data also suggest that the LPS aggregate structure and perhaps the
structure of the intact outer membrane may depend in part on the
divalent cations present and their physico-chemical properties.

Finally, to assess the effect of O-antigen length on polycation

 

interaction with LPS, LPS from a B-lactam and from an aminoglycoside

resistant E. aeruginosa strain both which contained shortened O-antigens

 

as well as the LPS from a rough mutant strain which contained lower
amounts of phosphate were studied. It was shown that the presence of
any length O-antigen chain or the loss of phosphate levels can decrease
the binding of certain polycations to LPS aggregates. It was also
observed that such changes in LPS structure can affect LPS head group
motion, i.e., aggregate packing. The data presented in this study
suggest that such alterations in LPS structure can specifically modulate

outer membrane permeability perhaps by several different mechanism.

APPENDIX A

Electron spin resonance probing techniques have been applied to the
study of biological membranes. The introduction of a stable free
radical ("spin label") enables one to use electron spin resonance (ESR)
spectroscopy to study specific environments within the membrane. The
ESR technique is sufficiently sensitive to detect signals from free-
radicals concentrations as low as 10'5 to 10'7 M.

In quantum mechanical notation electrons have a pr0perty known as
"spin", which can be conceptualized as a rapid rotation of the electron
about a specific axis. Because of the negative charge of the electron,
this spin will be associated with a magnetic moment, a vector quantity
having finite magnitude and direction, termed dipole moment. This
dipole moment will tend to allign itself with the lines of force of a
strong applied magnetic field (Fig. 1A). Application of electromagnetic
radiation of exactly the correct wavelength to an electron in a strong
magnetic field can induce the spining electron to "flip" the orientation
of its dipole moment with respect to the lines of force of the field
either from a parallel to an antiparallel direction or vice-versa (Fig.
1B) (5). The wavelength that induces the flip is called the resonance
frequency. Transitions in one direction absorb energy from the
electromagnetic radiation; the converse transition releases energy.
Since at equilibrium more electrons will be in low energy (parallel)
orientation than in the high energy (antiparallel) orientation, as the
resonance frequency is reached, the net observed effect will be a small

but measurable absorption of energy from the electromagnetic field (Fig.

161

162

Figure 1. The spinning electron is characterized by a dipole moment
which will align itself with the lines of force of a strong
magnetic field (A). If this electron is eXposed to electro-
magnetic radiation of the correct resonance frequency,
"flipping" of the electron dipole moment with respect to the
magnetic field is induced (B). When a varying frequency of
electromagnetic radiation is directed at a free electron in a
constant magnetic field, a peak of energy absorption
reflecting this flipping occurs at this reasonance frequency
(C). Plotting the differential of the absorption spectrum
produces a more easily interpretable signal (taken from

ref. 1).

163

fl-‘t [B
1. ’ t u
11 > 111‘

mognenc he":

11»
with electron ‘ (”’3’
I [I 1’

 

 

 

energy cosorbec “ 1

1 l ‘
microwave energy or / reversal of orientation
resononce frequency of magnetic moment

<1:
\ f
C :0 .
C V [A
/p\/) ; energy obsorpncn
J 151 Cenvotwe of energy cbscrphon

\f—

( — resonance freQuency

 

 

energy absorption

 

micr0w0ve {reduency (constant'tiem)
magnetuc field intensity (constant (reduency)

164

1C). In practice, changes in observed resonance frequency can be
related to physical factors in the environment of the electron.
Theoretically, the resonance absorption of a free electron in a
vacuum would produce a single peak in an ESR experiment. However, an
electron in a molecule is exposed to magnetic fields in addition to that
produce by the experimenter, the most significant of these fields is the
magnetic moment of the nucleus. Interaction of the nuclear magnetic
moment with the applied field shifts the electron resonance peak either
above or below that calculated for a free electron. Since the value of
the nuclear magnetic moment is constrained to certain discrete values by
quantum mechanical considerations, a finite number of discrete resonance
peaks will be seen for the electrons in a given population of radicals,
and these peaks will correspond to the number of permissible states of
the magnetic moment of their respective nuclei. For example, the free
radical most commonly used for studies of biological membranes is the
nitroxide radical (Fig. 2); in our case CAT12 (2). This is often used
in the form of an oxazolidine ring (3.A). This radical is unusually
stable due to the steric protection conferred by the methyl groups on
the carbon atoms adjoining the nitrogen. The unpaired electron spends
most of its time in a 2pm orbital oriented about the nitrogen atom along
an axis perpendicular to the plane of the carbon-nitrogen-oxygen (C-N-O)
surface (Fig. 3). A variable portion of its time is spent in a
similarly oriented orbital about the oxygen nucleus. The ESR absorption
spectrum is split into three peaks by interaction with the three
permissible states of the magnetic moment of the nitrogen nucleus and
the resonance process (Fig. 3A-C). These electron-nucleus interactions

are termed hyperfine interactions, and the resultant spectral

 

 

165

Figure 2. Cationic spin probe A-(dodecyldimethylammonio)-1-oxy-2,2,6,6-

tetramethylpiperidine bromide (CAT12).

166

CATIZ

167

Figure 3. Simulated nitroxide spectra. (A)-(C) Single crystal spectra
with the magnetic field oriented along each of the principal
axes, x, y, z. (D) Powder spectrum from a randomly oriented
collection of rigidly immobilized nitroxides. (E) Isotropic
spectrum from rapid and randomly tumbling nitroxides. (Taken

from ref. 5).

168

 

H11:

 

(o)

H"!

 

(b)

H"!

)

C

(

Rig'd 91055

((1)

50m 1 ion

(1:)

 

l
'_i

169

complications are termed hyperfine spectral splittings. Largest
hyperfine splitting is seen when the magnetic field is oriented along
the same axis as the electronic orbital (Fig. 3C). Much smaller
splittings are seen when the field is in the C-N-O plane (Fig. 3B-C).
The degree of spectral splitting produced by hyperfine interactions is
dependent on the disposition of other electrons within the molecule
containing the unpaired electron. Ring structures containing this
nitroxide radical can be used free or can be incorporated into various
biological molecules as probes of their local environment.

Once a stable free radical has been incorporated into the structure
of a larger probe molecule, new ESR spectral characteristics may become
apparent. The orbital containing the unpaired electron will have a
definite orientation within the molecular framework of the spin-label
probe. The magnitude of the nuclear hyperfine splitting will be
dependent on the orientation of this orbital and hence of the molecule
itself with respect to the applied magnetic field, giving rise to
positional dependency of the ESR spectrum. Spectra obtained from probe
molecules having a known uniform orientation with a host crystal lattice
will change dramatically when the orientation of the crystal and its
immobilized probe molecules are altered with respect to the applied
magnetic field (Fig. 3).

For a probe molecule in free solution, all potential orientations
with respect to the magnetic field are possible. If the probe molecule

is free to move very rapidly with respect to the time required for

electronic transition (rotation occurring within ~10"9 390-). mOIGCUIar
positional effects will be averaged, and a single well-resolved spectrum

reflecting these average effects will be seen (Fig. 3E). If the probe

170

molecules are immobilized in random orientation with respect to the
magnetic field and are not free to rotate within that time-frame
(rotation slower than 10’7 sec.), the resulting spectrum will be the sum
of all possible spectra, each with its specific possitional effect, and
a much more diffuse and less well defined experimental spectrum will be
obtained (Fig. 30). As probe motion in an isotropic solvent is varied
between complete freedom and complete immobility, a continuous family of
curves will be obtained (Fig. A). Careful analysis of spectral line
shapes in experimental systems can be used to draw inferences concerning
the freedom of movement of the probe in its microenvironment with
respect to the electronic transition times involved (10'7 to 10'9 sec.).
This spectral anisotropy is the fundamental basis of the usefulness
of the spin-label method. The description of anisotropic systems
usually requires six independent parameters conveniently ordered in a
symmetric array called tensor. When the spin-probe molecular system
assumes axial symmetry about the 3 molecular axis the calculations of
these parameters are simplified. For these cases, the principal values
of the hyperfine tensor constant A (or T) and g-tensor, g, values are:
A” = A22; A1 = Axx = Ayy, similarly g” = gzz; 81 . gxx = gyy.
Approximate values of Ay, AL’ fig, and 81 are obtained from
experimental spectra as shown in Fig. A. These parameteres, particularly
A” and A1’ are highly sensitive to changes in the motional
environment of the spin-label. For further discussion on this t0pic, see

Marsh (5) and Schreier gt gt. (6).

 

171

Figure A. Rigid limit nitroxide spectrum (broken lines), illustrating
the measurement of the parameters A'zz (=A'z) and A22 (=Az).

(Taken from ref. 5).

172

 

 

 

 

 

 

 

 

 

10 GAUSS

REFERENCES

Barchi, R.L. 1980. Physical probes of biological membranes in
studies of the muscular dystrophies. Muscle & Nerve 3: 82-97.
Coughlin, R.T., C.R. Caldwell, A. Haug, and E.J. McGroarty. 1981.
A cationic electron spin resonance probe used to analyze cation
interactions with lipopolysaccharides. Biochem. Biophys. Res.
Commun. 100: 1137-11A2.

Gaffney, B.J. 1976. The chemistry of spin labels, pp. 18A-232.

In L.J. Berline (ed.), Spin Labelligg Theogy and Application, New

 

York, Academic Press.

Keana, J.F., S.B. Keana, and P. Beetham. 1967. A new versatile
ketone spin label. J. Am. Chem. Soc. 89: 3055-3056.

Marsh, D. 1981. Electron spin resonance: spin labels, pp. 51-1A2.

In E. Grell (ed.), Molecular Biology, Biochemstry and Biophysics.

 

Membrane Spectroscopy. Springer-Verlag, Germany.

 

Schreier, S., C. Polnaszek, and I. Smith. 1978. Spin labels in
membranes. Problems in practice. Biochem. Biophys. Acta 515:

375-A36.

173

APPENDIX B

Publications:

"Phospholipases A2 as Probes of Phospholipid Organization in Cell
Surfaces." C.A. Medina, E.R. Resto, M. Rivera, and R.W. Morales,
1983. Egg. Etgg. fig, 320.

"Neutron Scattering Analysis of Bacterial Lipopolysaccharide Phase
Structure. Changes at High pH." J.B. Hayter, M. Rivera, and E.J.

McGroarty. 1987. g. Biol. Chem. 262, 5100-5105.

 

"Binding of Polycationic Antibiotics to Lipopolysaccharides of

Pseudomonas aeruginosa." M. Rivera and E.J. McGroarty, 1988. In

 

Antibiotic Inhibition gt Bacterial Cell Surface Assembly and

 

 

Function. P. Actor, L. Daneo-Moore, M. Higgins, M.R. Salton, and

G. Shockman (eds.), ASM, Washington.

"Heterogeneity of Lipopolysaccharide from Pseudomonas aergunosa:

 

Analysis of Lipopolysaccharide Chain Length." M. Rivera, L.E.
Bryan, R.E.W. Hancock, and E.J. McGroarty, 1988. g. Bacteriol.
112.

"Enhanced Binding of Polycationic Antibiotics to Lipopolysaccharide
from an Aminoglycoside Supersusceptivle, tgtg-mutant Strain of

Pseudomonas aeruginosa." M. Rivera, R.E.W. Hancock, J.G. Sawyer,

 

A. Haug, and E.J. McGroarty, 1988. Antimicrob. Agents Chemother.

 

 

(in press).

"Cation Interactions with Lipopolysaccharide from Pseudomonas

 

aeruginosa PAO Strain with Altered O-antigen Chain Length." M.

 

174

175

Rivera, L.E. Bryan, A. Haug, and E.J. McGroarty. 1988.
(submitted).
"Differential Interactions of Calcium and Magnesium Salts of

Lipopolysaccharide from Pseudomonas aeruginosa with Polycations."

 

M. Rivera, A. Haug, R.E.W. Hancock, and E.J. McGroarty, 1988. (in
preparation).
"Partial Chemical and Antigenic Characterization of Polyrhamnose

Containing Lipopolysaccharide from Pseudomonas aeruginosa PAO

 

Strains." M. Rivera, W. Rocque, and E.J. McGroarty, 1988. (in

preparation).

7 Abstracts published in various Conference Proceedings
(co-authors: R.E.W. Hancock, B.W. Morales, L.E. Bryan, W. Rocque,

and E.J. McGroarty).

MICHIGAN STATE UNIV. LIBRQRIES
1|1111111111111111111111111111111111111"11111111
31293007868148