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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 a/CIRC/Dnoouepss-pds

 

THE ANTIMICROBIAL ACTIVITY OF GANGLIOSIDES ISOLATED
FROM LACTIC ACID BACTERIA, CULTURED DAIRY FOODS,
AND BOVINE MILK FAT
By

Carollyn Hope Boykins

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Animal Science

2001

ABSTRACT
THE ANTIMICROBIAL ACTIVITY OF GANGLIOSIDES ISOLATED
FROM LACTIC ACID BACTERIA, CULTURED DAIRY FOODS,
AND BOVINE MILK FAT
By

Carollyn Hope Boykins

There is substantial evidence suggesting that lactic acid bacteria (LAB) have an
inhibitory effect on pathogenic bacteria in the intestinal tract. LAB and LAB used for
manufacturing cultured dairy foods have been shown to modify the metabolic activities of
the colonic microflora, possess antimicrobial activity, and neutralize Escherichia coli (E.
coli) enterotoxin. Although various mechanisms have been proposed for these effects,
the involvement of LAB gangliosides contained in these products was the focus of this
research.

Mixed gangliosides were extracted and purified from sweet and cultured
buttermilk, LAB, and yogurt. These products were targeted because it is known that they
contain gangliosides. Using an agar overlay plate technique, mixed gangliosides
recovered from these products demonstrated bacteriostatic effects, as evidenced by the
retardation of E. coli growth around wells containing varying amounts of gangliosides.
Mixed gangliosides have also been demonstrated to inhibit mannose-resistant pilus-
mediated adhesion of E. coli. The hemagglutination-inhibition assay results suggest that
gangliosides may limit the adhesion of enterotoxigenic E. coli to the intestinal mucosa
and thus prevent diarrhea. The application of this research to Helicobacter pylori (a

bacterium strongly associated with gastric ulcers) was demonstrated by inhibition of

pilus-mediated adhesion of this organism as determined by the hemagglutination-

inhibition assay.

DEDICATION

To my mother, Mildred Boykins, my first teacher: Your patience and capacity for
love I can only hope to emulate. The knowledge you have endowed me with will span
my lifetime. You have made the education of others your life’s work, and I am honored
and fortunate to be counted among them.

To my father, James O. Boykins: By your example, you have impressed upon me
the values, perseverance, and attitude needed not only to succeed, but truly to experience
life. Through you, I have come to understand that the merit of one’s life lies not in one’s
accomplishments, but in his or her positive effects on others and the community as a
whole.

To my future husband, Michael A. Winrow: You have helped me unlock the
mysteries of deoxyribonucleic acid, as well as the secrets of my heart. To my family and
friends who have nurtured, encouraged, and believed in the dream; to the many educators
who have provided me with the foundation needed to achieve the dream; but most of all

to my parents, for making me a dreamer.

iv

ACKNOWLEDGMENTS

I would like to express my gratitude to Dr. Melvin Yokoyama, my major
professor, for training me in various microbial techniques, and for his guidance, support,
encouragement, and friendship, all of which contributed to my completion of the
doctorate degree.

I would like to thank Dr. Julius Jackson for rescuing me from the world Of work
and encouraging me to further my education at Michigan State University and, in the
process, introducing me to my future husband. My debt to him will likely never be
repaid.

I also would like to thank Dr. Sheila Grimes for her assistance, expertise,
friendship, and positive attitude in the face of slow, hectic, and sometimes impossible
odds.

I also wish to acknowledge the remaining members of my graduate committee,
Dr. Steven Bursian and Dr. John Kaneene, for their advice and assistance.

My sincere appreciation is expressed to Dr. Kathleen Moore, whose expertise in
chemical analysis allowed for the extraction of gangliosides, which was so crucial to this
research, and to Dr. Anna Ettinger, whose preliminary work in this area provided
fundamental data upon which much of our early research was based. They both were
invaluable resources for information, consultation, and advice.

I would also like to express my gratitude to Darigold, Inc. (Mary Jens) for
providing the powdered buttermilk and bacterial starter cultures for this research, to

Michigan State University’s Dairy Plant (Dr. John Partridge and Siggaard Torben) for

graciously supplying yogurt for ganglioside extraction, and Chris Hansen’s Biosystems
for supplying bacterial starter cultures.

1 must acknowledge the help of Dr. James Jay, who believed that I could
accomplish the task at hand. Without him, I might have given up long ago. He has
personally assisted in adding to the ranks of those holding postgraduate degrees in
science, and society’s debt to him will likely never be repaid. My appreciation also is
extended to Dr. Soegiarto for his endless assistance with the trials and tribulations of
computers and their programs, Maggie Hoffman and Ralph Common for their graphics
expertise, Dr. Robert Holland and Dr. Raul Hollingsworth for providing bacterial

samples, and Dr. Sue Hengemuehle for providing everything that she could.

vi

TABLE OF CONTENTS

LIST OF TABLES ...................................................... ix

LIST OF FIGURES ...................................................... x
CHAPTER

1. LITERATURE REVIEW .................................... 01

The Use of Lactic Acid Bacteria as a Probiotic .................... O3

Adhesion of E. coli ......................................... 07

Hemagglutination ........................................... IO

Gangliosides ............................................... 1 1

Antimicrobial Properties of Gangliosides ........................ l6

Antibacterial Treatment of Gastric Ulcers ........................ 17

II. GANGLIOSIDE EXTRACTION AND PURIFICATION FROM

III.

LACTIC ACID BACTERIA, BOVINE MILK FAT, AND

CULTURED DAIRY FOODS ................................. 19
Abstract .................................................. 19
Introduction ............................................... 2 1
Materials and Methods ....................................... 22
Results ................................................... 28
Discussion ................................................ 31

BACTERIOSTATIC EFFECTS OF GANGLIOSIDES DETER-
MINED BY A DOUBLE AGAR OVERLAY ASSAY

TECHNIQUE .............................................. 40
Abstract .................................................. 40
Introduction ............................................... 41
Materials and Methods ....................................... 42
Results ................................................... 46
Discussion ................................................ 62

vii

IV. INHIBITION OF ADHESION K99+ ESCHERICHIA COLI (E. COLI)
AND COLONIZATION FACTOR ANTIGENS-1+ E. COLI BY
PURIFIED MIXED GANGLIOSIDES EXTRACTED FROM
LACTIC ACID BACTERIA, BOVINE MILK FAT, AND

CULTURED DAIRY FOODS ................................. 67
Abstract .................................................. 67
Introduction ............................................... 68
Materials and Methods ....................................... 69
Results ................................................... 74
Discussion ............................................... 107
V. INHIBITION OF ADHESION OF HELICOBA CT ER PYLORI

BY MIXED GANGLIOSIDES IN LACTIC ACID BACTERIA,
BOVINE MILK FAT, AND CULTURED DAIRY FOODS ........ 112
Abstract ................................................. 112
Introduction .............................................. 1 13
Materials and Methods ...................................... 114
Results .................................................. 119
Discussion ............................................... 122
VI. SUMMARY .............................................. 130
APPENDIX ........................................................... 33
BIBLIOGRAPHY ..................................................... 139

viii

LIST OF TABLES

Table Page
A.1 Inventory of Ganglioside Preparations by Protocol ....................... 33
A2 Resorcinol Assay Quantitation of Crude Mixed Gangliosides .............. 36

A3 Qualitative Thin-Layer Chromatography Analysis of Crude Mixed

Gangliosides ..................................................... 36
A4 Crude Ganglioside Preparation Results ................................ 37
A5 Purified Ganglioside Preparation Results .............................. 37
A6 Inventory of Purified Gangliosides and Dairy Products ................... 38

ix

Figure

1.1

1.2

3.1

3.2

3.3

3.4

3.5

3.6

3.7

3.8

3.9

3.10

3.11

LIST OF FIGURES

Page
The Structures of Important Ganglioside Precursors and
Monosialoganglioside (GMl) ....................................... l3
Schematic Drawing of Ganglioside Biosynthesis ........................ 14
Photograph of Zones of Inhibition Caused by Mixed
Gangliosides Using a Double AgarOverlay Technique .................... 47
Formation of Inhibition Zones With Mixed Bovine
Brain Gangliosides in a lawn of K88+ Escherichia Cali .................. 48
F ormation of Inhibition Zones With Mixed Cultured
Buttermilk Gangliosides in a lawn of K88+ Escherichia Cali .............. 49
Formation of Inhibition Zones With Mixed Sweet
Buttermilk Gangliosides in a lawn of K88+ Escherichia Cali .............. 50
Formation of Inhibition Zones With Mixed Lactic Acid
Bacteria Gangliosides in a lawn of K88+ Escherichia Cali ................ 51
Formation of Inhibition Zones With MixedYogurt
Gangliosides in a lawn of K88+ Escherichia Cali ....................... 52
Formation of Inhibition Zones With Mixed Cultured
Buttermilk Gangliosides in a lawn of K99+ Escherichia cali ............... 53
Formation of Inhibition Zones With Mixed Bovine
Brain Gangliosides in a lawn of K99+ Escherichia Cali .................. 54
Formation of Inhibition Zones With Mixed Sweet
Buttermilk Gangliosides in a lawn of K99+ Escherichia Cali .............. 55
Formation of Inhibition Zones With MixedYogurt
Gangliosides in a lawn of K99+ Escherichia Cali ....................... 56
Formation of Inhibition Zones With Mixed Lactic Acid
Bacteria Gangliosides in a lawn of K99’ Escherichia Cali ................ 57

3.12

3.13

3.14

4.1

4.2

4.3

4.4

4.5

4.6

4.7

4.8

4.9

Formation of Inhibition Zones With MixedYogurt
Gangliosides in a lawn of Escherichia Cali 0157:H7 ..................... 58

Formation of Inhibition Zones With Mixed Sweet
Buttermilk Gangliosides in a lawn of Escherichia Cali 0157:H7 ............ 59

Bacteriostatic Effects of Mixed Gangliosides Using a Double Agar
Overlay Technique ................................................ 61

Diagrammatic Scheme of the Hemagglutination-Inhibition Assay Used
to Determine the Beneficial Antibinding Effect of Gangliosides in
Cultured Dairy Foods .............................................. 75

Photograph of Hemagglutination Assay Chambers Showing
Erythrocyte Binding to Enterotoxigenic E. cali .......................... 76

Inhibition of Hemagglutination of Sheep Erythrocytes With
Escherichia cali K99+47 by Competitive Binding With
Sweet Buttermilk Gangliosides ...................................... 77

Inhibition of Hemagglutination of Sheep Erythrocytes With
Escherichia cali K99+47 by Competitive Binding With Cultured
Buttermilk Gangliosides ........................................... 78

Inhibition of Hemagglutination of Sheep Erythrocytes With
Escherichia cali K99+47 by Competitive Binding With .
Yogurt Gangliosides .............................................. 79

Inhibition of Hemagglutination of Sheep Erythrocytes With
Escherichia cali K99+47 by Competitive Binding With
Lactic Acid Bacterial Gangliosides ................................... 80

Inhibition of Hemagglutination of Sheep Erythrocytes With
Escherichia cali K99+47 by Competitive Binding With
Bovine Brain Gangliosides ......................................... 81

Inhibition of Hemagglutination of Sheep Erythrocytes With

Escherichia cali K99+47 by Competitive Binding With
Monosialoganglioside ............................................. 82

Inhibition of Hemagglutination of Equine Erythrocytes With

Escherichia cali K99+47 by Competitive Binding With
Sweet Buttermilk Gangliosides ...................................... 83

xi

4.10

4.11

4.12

4.13

4.14

4.15

4.16

4.17

4.18

4.19

4.20

Inhibition of Hemagglutination of Equine Erythrocytes With
Escherichia cali K99+47 by Competitive Binding With
Cultured Buttermilk Gangliosides .................................... 84

Inhibition of Hemagglutination of Equine Erythrocytes With
Escherichia cali K99+47 by Competitive Binding With
Yogurt Gangliosides .............................................. 85

Inhibition of Hemagglutination of Equine Erythrocytes With
Escherichia cali K99’47 by Competitive Binding With
Lactic Acid Bacterial Gangliosides ................................... 86

Inhibition of Hemagglutination of Equine Erythrocytes With
Escherichia cali K99+47 by Competitive Binding With
Bovine Brain Gangliosides ......................................... 87

Inhibition of Hemagglutination of Equine Erythrocytes With
Escherichia cali K99+47 by Competitive Binding With
Monosialoganglioside ............................................. 88

Inhibition of Hemagglutination of Equine Erythrocytes With
Escherichia cali K823“ by Competitive Binding With
Sweet Buttermilk Gangliosides ...................................... 93

Inhibition of Hemagglutination of Equine Erythrocytes With
Escherichia cali K88+ by Competitive Binding With
Cultured Buttermilk Gangliosides .................................... 94

Inhibition of Hemagglutination of Equine Erythrocytes With
Escherichia cali K88+ by Competitive Binding With
Yogurt Gangliosides .............................................. 95

Inhibition of Hemagglutination of Equine Erythrocytes With
Escherichia cali Kss” by Competitive Binding With
Bovine Brain Gangliosides ......................................... 96

Inhibition of Hemagglutination of Chicken Erythrocytes With
Colony Forming Antigens by Competitive Binding With
Sweet Buttermilk Gangliosides ...................................... 98

Inhibition of Hemagglutination of Chicken Erythrocytes With

Colony Forming Antigens by Competitive Binding With
Cultured Buttermilk Gangliosides .................................... 99

xii

4.21 Inhibition of Hemagglutination of Chicken Erythrocytes With
Colony Forming Antigens by Competitive Binding With
Yogurt Gangliosides ............................................. 100

4.22 Inhibition of Hemagglutination of Chicken Erythrocytes With
Colony Forming Antigens by Competitive Binding With
Lactic Acid Bacterial Gangliosides .................................. 101

4.23 Inhibition of Hemagglutination of Chicken Erythrocytes With
Colony Forming Antigens by Competitive Binding With
Sweet Buttermilk Gangliosides ..................................... 102

4.24 Inhibition of Hemagglutination of Bovine Erythrocytes With
Colony Forming Antigens by Competitive Binding With
Cultured Buttermilk Gangliosides ................................... 103

4.25 Inhibition of Hemagglutination of Bovine Erythrocytes With
Colony Forming Antigens by Competitive Binding With
Yogurt Gangliosides ............................................. 104

4.26 Inhibition of Hemagglutination of Bovine Erythrocytes With
Colony Forming Antigens by Competitive Binding With
Lactic Acid Bacterial Gangliosides .................................. 105

5.1 Inhibition of Hemagglutination of Chicken Erythrocytes With
Helicabacter Pylari by Competitive Binding With Lactic
Acid Bacterial Gangliosides ....................................... 123

5.2 Inhibition of Hemagglutination of Chicken Erythrocytes With
Helicabacter Pylari by Competitive Binding With Cultured
Buttermilk Gangliosides .......................................... 124

5.3 Inhibition of Hemagglutination of Chicken Erythrocytes With
Helicabacter Pylari by Competitive Binding With Yogurt
Gangliosides .................................................... 125

5.4 Inhibition of Hemagglutination of Chicken Erythrocytes With

Helicabacter Pylari by Competitive Binding With Evaporated
Milk Gangliosides ............................................... 126

xiii

CHAPTER I

LITERATURE REVIEW

For many years, scientists have identified the microbial world as an arch enemy of
continuing healthful existence. For this reason, most researchers have emphasized the
harmful effects of the microbial environment. Investigations dealing with the protective
and regulatory role of normal microflora, however, recently have begun to increase.

Using “good” bacteria to fight or replace “bad” bacteria is far from a new concept,
with research on this issue dating back more than a century. Metchnikoff’s (107)
research with cholera in the early 19003 usually is recognized as the earliest attempt at
probiotic use. In his review, Florey (43) discussed work by Cantani in 1885, in which the
replacement of Bacterium tuberculosis with a “harmless organism” was claimed. In
1916, Nissle used a bacterial strain called “Bacti. coli,” which was made commercially
available as “multiflor” for treatment of various intestinal problems.

Staphylococcus aureus strain 502A was used in the early 19605 to prevent
intestinal colonization by S. aureus phage type 80/81, a virulent organism known to cause
high rates of morbidity and mortality among neonates (8, 155). This experiment involved
implanting the 502A strain in neonates before the more virulent S. aureus 80/81 strain
could become established. Use of the 502A strain became fairly widespread. However,

its use declined, partly because of decreasing need and also because the administration of

large doses of this strain was thought to be responsible for lesion development, which
could lead to the disease itself (29).

More recently, other attempts have been made to block S. aureus infections in
neonates; one such effort was Ehrenkrantz’s implantation of diphtheroids and
S. epidermidis (34). No harmful effects were produced, but the purpose also was not
achieved. In 1973, Ludinova implanted a benign type 083 Escherichia coli in the gut of
neonates to lessen intestinal colonization by bacteria and decrease the incidence of
diarrhea (98). More recently, Rueda and colleagues (139) added gangliosides to an
adapted milk formula and were successful in modifying the fecal flora of preterm infants.
Tuomola and Salminen (174) demonstrated adhesive properties of probiotics and dairy
lactobacillus strains. Implantation of bacterial strains is also being used for the
prevention of dental caries (83, 89, 90) and urinary-tract infections (137, 166).

In 1989, Fuller (45) defined a probiotic as “a live microbial feed supplement
which beneficially affects the host animal by improving its inteStinal microbial balance.”
This definition is somewhat limited, though, and a broader definition that does not limit a
probiotic to feed supplements or to animals would likely be more applicable today. The
scientific basis for the development of probiotics stems from the knowledge that the gut
flora is involved in protecting the host animal against colonization of the intestinal tract
by enteroinvasive or opportunistic nonindigenous organisms.

Although much of the knowledge that predates Metchnikoffian theories is not
well established, belief in the health-promoting properties of lactic acid bacteria (LAB)
was well ingrained in Eastern European and Russian traditions before Metchnikoff’s

publications (156, 159). Healthy animals and humans have normal microflora that are

present on internal and external surfaces. The skin, upper respiratory tract, lower
urogenital tract, oral cavity, and gastrointestinal tract all represent microbial habitats, each
of which has complex and different environmental conditions. In neonates, these
microbial habitats receive microbes directly from the mother and also are contaminated
during and directly following birth by the surrounding environment (15, 28, 143, 146).

Colonization of the body surfaces may be temporary or permanent, with mature
microflora being dependent on the specific animal species, the individual animal, and the
inhabiting environment (64, 146, 179). The well-established indigenous microflora of the
body surfaces are quite stable. Through competitive exclusion (44, 162, 182) or
colonization resistance (23, 24, 179), the colonization of the surfaces by nonindigenous
microorganisms is inhibited (146, 162).

The importance of indigenous microflora received renewed interest when
intestinal disturbances in mice were produced using oral antibiotics (179). Protective
effects of normal gut microflora have been demonstrated for Salmonella typhimurium
(97, 161), Escherichia coli (35,), Campylobacter (I35, 162), and Yersenia enterocolitica
(163). Several factors can influence the prevention of early colonization by
microorganisms and the development of normal microflora. The lack of breast feeding is
one factor that is thought to contribute to a deficiency of normal microflora development

in infants (17, 28, 139).

I] II [I '!°1E . Erli'
There is considerable interest in understanding how the microflora of the

intestines can be manipulated, particularly by dietary factors (30). Among food products

targeted as possible microflora manipulators, fermented dairy products have received
considerable attention. LAB are an integral component of many dairy products. The term
LAB refers to gram-positive, anaerobic microaerophilic or aero-tolerant, catalase-
negative rods or cocci that produce lactic acid as an important product from the
fermentation of sugars. Substantial evidence suggests that LAB have an inhibitory effect
on pathogenic bacteria in the intestinal tract (18, 35, 49, 50, 51, 59, 62, 90, 103, 109,
117, 130, 185, 189). Several theories have been proposed to explain this bacterial
antagonism. The most widely accepted theory is that of competitive exclusion, which
holds that LAB control the growth and colonization of Opportunistic enteric pathogens by
effectively competing for nutrients, lowering the luminal pH through the production of
lactic acid and the possible production of antibiotic-like substances. This theory is the
basis for the use of LAB as a biological prophylaxis for enteric diseases (23, 162, 182).
The long-standing belief that cultured dairy foods containing active LAB are beneficial in
the prevention and treatment of enteric diseases was first proposed by Metchnikoff(107)
in 1907, investigated further by Vincent et al. (178), and more recently studied by Wood
(185).

Various explanations have been proposed for the beneficial effects demonstrated
by cultured dairy foods. The involvement of LAB gangliosides, however, has not been
considered. LAB such as members of the genera Lactobacillus and Streptococcus are
used in the commercial production of fermented dairy products. Gangliosides constitute
about 1% to 2% of the total lipid content of buttermilk (65). The possible involvement of
LAB in the modification or production of milk-fat gangliosides also has been suggested

by the identification of a unique ganglioside 9-0-acetyI-GD3 in bovine buttermilk (14).

Goldin et al. (49, 50) conducted research on the oral administration of strains of
Lactobacillus acidophilus, which decreased the concentrations of some fecal enzymes.
The enzymes that are decreased (B-glucuronidase, azoreductase, and nitroreductase) have
the ability to transform procarcinogens to carcinogens in the colon (58, 68). The
beneficial effects suggested by these studies have encouraged further investigation by
others concerning the effects of commercial fermented products containing lactic acid
bacteria 100, 104).

The belief that breast feeding provides increased resistance to microbial
infections, thereby lessening digestive problems, has been supported in several studies
(17, 34, 50, 57, 69, 73, 111) and presents a good case for probiotic use. The neonatal
tract is initially germ free, but it is progressively colonized with bacteria from the mother
as well as external sources (15, 17, 34). The microflora of the gastric contents of a 5- to
IO-minute-old infant were found to already reflect the cervical flora of its mother (15,
28).

It is still a mystery how the neonatal gut selects a limited number of microbial
species from the enormous number to which it is exposed. The complex microflora that
develop are largely responsible for protection against many infectious diseases and play a
major role in resisting further colonization (23, 51). These protective microflora can be
modified by various factors, such as diet (49, 139), environmental changes (185), and
antibiotic treatment (130, 179), which result in a disturbance of the normal gut flora. This
disturbance frequently enables pathogens to multiply and is a major cause of diarrhea.
Infants ingesting milk-formula substitutes or milk from other species do not have the

same degree of microbial resistance as breast-fed infants (17, 85). In breast-fed

newborns, colonization of the intestine by Lactobacillus bifidus occurs quickly. In non-
breast-fed infants, other commensal microorganisms inhabit the intestine and become
established (15, 34). The magnitude of lactobacilli in breast-fed infants can be attributed
to the high concentration of a Specific growth-promoting factor for L. bifidus present in
human milk (69, 85). The high concentration of lactose in this growth promoter is
believed to ensure that, even after the food has passed into the large intestine, there is
sufficient lactose to produce acidic conditions due to lactose fermenters and to encourage
lactobacilli growth (23, 49, 50). Specifically, L. bifidus helps convert lactose into lactic
acid and acetic acid. The increased concentration of these acids in the intestinal tract
lowers the pH and interferes with the growth of many enteropathogenic organisms (23,
49, 103).

The growth-promoting factor for L. bifidus is a nitrogen—containing
polysaccharide with high lactose content, low protein, low bulk, and low buffering
capacity (69). This growth promoter is found in bovine milk in only negligible amounts
(69, 79, 85). In addition to the lactobacilli growth-promoting factor, other defense factors
found in human milk include an anti-staphylococcal agent, imrnunoglobulins,
complement components, lysozyme, peroxidase, lactoferrin, macrophages, lymphocytes,
and lipids (57, 69).

Since Metchnikoff (104) conducted his studies at the turn of the century, the belief
has persisted that LAB and cultured dairy foods produced by LAB hold the key to
prolonged life and improved health. The ability to significantly increase the numbers of
lactobacilli in feces within a few days by ingesting low-fat milk has been proven (18, 49,

189).

More recently, researchers have suggested that the use of fermented milk could
have therapeutic effects on malnourished and diarrhea] patients(2, 49, 116, 150, 156).
Examples are the resistance to Shigella sonnei, which was enhanced by feeding fermented
milk for 8 days (117), and the increased resistance to Salmonella typhimurium in subjects
fed fermented milk (97, 130). About 40% to 70% of children with Salmonella and
Shigella dysentery recovered when acidophilus milk was administered for a short time.
Continued long-term administration of acidophilus milk resulted in 100% recovery (97,
130).

Lactic acid bacteria therapy also has been reported to have positive effects in the
prevention and treatment of travelers’ diarrhea (103, 106, 156). This condition is
normally attributed to enterotoxigenic strains of Escherichia cali (E. cali). These E. coli
cause diarrhea through their ability to colonize the small intestine and produce
enterotoxin (64, 119, 127, 143). The mechanisms by which LAB may prevent or lessen
the effects of diarrhea (e.g., the production of anti-E. coli [2, 27, 109], the synthesis of
lactate with ultimate reduction in intestinal pH [23, 49], detoxification of endotoxin
[130], prevention of toxic amine synthesis [154], and adhesion to the gut wall preventing
colonization [2, 18, 130]) have been suggested in a number of reports. Treatment of
bacterial infections by competitive inhibition of bacterium-receptor interactions has been

proposed and successfully demonstrated ( 97, 103, 125, 130, 139, 150).

9 ll . E E i'

Enteric diseases, including enteropathogenic E. coli (EPEC), enterohemorrhagic
E. coli (EHEC), and enterotoxigenic E. cali (ETEC), all cause diarrhea, with ETEC being
the most common cause of diarrhea in young farm animals (64). Attachment of
enterotoxigenic bacteria to the susceptible host tissue is the initial step in the disease
process (119, 159). Most strains of E. coli are normal inhabitants of the gastrointestinal
tract. However, the expression of virulence factors, or mechanisms of action, that
enhance an organism’s ability to cause a variety of intestinal infections and neonatal
diarrhea in animals and humans has been well documented (5, 31, 34, 39, 40, 111, 119).
Pathogenic bacteria can cause diarrhea by (a) producing toxins that increase secretion by
the crypt cells of the small intestine; (b) invading the intestinal mucosa, causing
inflammatory reactions, which in turn increase secretion by crypt cells; or (c) destroying
absorptive epithelial cells, which causes malabsorption (119, 143, 159).

Two important factors that provide ETEC with means of colonization are the
expression of pili, which allow for the attachment of bacteria to the intestinal wall, and
the production of enterotoxins that influence intestinal secretions (39). Pili are finger-like
projections protruding from the bacterial surface; they are also designated as fimbriae,
surface adhesions, and colonization-factor antigens. Pili enable bacteria to adhere to the
small intestinal epithelium (19, 39, 47, 56, 123). The pili are filamentous proteins that
are located on the surface of the bacteria. Normally, peristalsis and the flow of intestinal
contents, coupled with the continuous surge of mucous over the surface of the constantly

moving intestinal villi, keep noncommensal bacterial populations small (5, 119, 143).

The ability of bacteria to adhere to the mucosal surface via pili would improve the
capacity for intestinal colonization by the organism.

Different types of pili are synthesized by various strains of E. coli, which allow
the bacteria to bind to the intestinal tract of several species. Antigens found on E. coli
that cause diarrhea in pigs are K88, K99, F41 , and 987P; K99 and F41 antigens cause
diarrhea in calves; K99 antigens are present on E. cali that cause diarrhea in lambs; and
the colonization-factor antigens, CF A I, CFA II, and CFA IV, are partially responsible for
diarrhea in humans (64, 110, 119). In general, one E. coli strain expresses one type of
antigen (64, 119), although there are some exceptions (149). Both glycoprotein (160,
165) and glycolipid (171) moieties have been implicated as receptors for K88 fimbriae in
the small intestines of pigs.

The mucosa of the intestine is the site of both absorption and secretion. The crypt
portion contains undifferentiated epithelial cells, which are the major sites for secretions
(5). These cells later proliferate and migrate out to the villi, thus losing their proliferative
and some secretory abilities. Crypt cells then become the sites of absorption and
digestion in the small intestine (111, 119). Absorption occurs more than secretion in
normal intestinal function. Intestinal secretion is regulated by cyclic adenosine
monophosphate (cAMP), which is in turn regulated by specific hormones (111, 119, 143).
Enterotoxin produced by E. coli is believed to act similarly to hormones, signaling the
increase of mucosal CAMP. The increase in cAMP causes the hypersecretion of the crypt
cells; this leads to increased secretion and reduced absorption, resulting in diarrhea (l 19).

The specificity of bacterial pili and the presence of receptors for these pili

producing clinical manifestations have been best evidenced for E. coli bacteria possessing

the K88 fimbrial adhesion, which causes diarrhea in neonatal pigs (36, 72, 149). The
presence of two plasmids, one encoding the genes for the expression of K88 fimbriae and
the other encoding the genes causing toxin production, has been found to correlate
directly with the incidence of diarrhea in neonatal pigs (146, 149). It has been further
demonstrated that the adhesion of bacteria to the intestinal brush borders is facilitated by
the K88 pili (36). Possession of the receptor for K88 pili has been proven to confer
sensitivity, whereas the lack of this specific receptor has been discovered in pigs that are
resistant to the K88-mediated diarrhea (12, 153, 160). In addition, K88-specific receptors
have been identified in the mucous layer that covers epithelial cells of the small intestines
of neonatal pigs (11, 12). The K99 pili possess an affinity for neuraminic acid derivatives
(84). Recent research has indicated that a receptor—active glycolipid
(monosialoganglioside II3 Nech-Lac) is present in the intestines of pigs (124). This

same receptor also has been identified as a receptor in horse erythrocytes (158).

W

The binding or adhesive properties of E. cali were first examined by Guyot (56),
who found that some strains had the ability to agglutinate erythrocytes (56). In 1955,
Duguid (31) was the first to link the attachment property of E. coli (agglutination) to
fimbrial projections on the surface. Hemagglutination has proven to be an important way
to detect and classify pili, as well as to indicate their presence. Distinction of pili as
mannose-sensitive (MS), which are inhibited by mannose and designated as Type I pili, or
mannose-resistant (MR), which are not affected by mannose and designated as Type II

pili, represents one form of classification (40, 126). ETEC possesses mannose-resistant

10

pili. The K88 antigen will hemagglutinate guinea pig and chicken erythrocytes (31, 32).
The 987P antigen exhibits only weak MR-hemagglutination with chicken erythrocytes
(39). The K99 antigen shows MR-hemagglutination with horse and sheep erythrocytes
(94, 158). The F41 antigen demonstrates MR-hemagglutination with sheep erythrocytes
(47, 92). The CF A I antigens possess MR-hemagglutination with human, bovine, and
chicken erythrocytes, whereas the CFA II antigen has MR—hemagglutination with chicken
and bovine erythrocytes (38, 39, 42).

The relationship between hemagglutination and adhesion was first established by
Evans et al. (37). Observations indicated that strains of E. coli from the intestines and
urinary tracts of diseased individuals showed increased hemagglutination with a variety of
cell types, as compared to normal intestinal microflora. This association has been
established for MS-hemagglutination of guinea pig erythrocytes with a variety of cell
types (123). Determining the specific membrane receptor molecules on erythrocytes and
epithelial cells is the most important factor in establishing the relationship between
hemagglutination and adhesion.

High concentrations of N-acetyl neuraminic acid have been found to inhibit
hemagglutination. Lindahl et al. (93) demonstrated that disialogangliosides (Type II,
Sigma) and highly purified GMZ gangliosides caused specific inhibition of
hemagglutination and binding of strains to isolated human intestinal epithelial cells. In
addition, researchers have found that terminal sialic acid residues inhibited K99Jr E. cali

and CFA II E. cali (93, 94, 99, 116, 120, 183).

11

fiangllgsidss

Glycolipids consist of carbohydrate and lipid moieties in covalent linkage. Two
important classes are cerebrosides and gangliosides, with the term “glycosphingolipid”
being a more accurate description of both groups. Gangliosides are naturally occurring
derivatives of cerebrosides containing a more complex carbohydrate moiety that includes
one or more N-acetyl neuraminic acid, also called sialic acid, units in the oligosaccharide
chain that is attached to ceramide (75, 88). ( Figure 1.1.)

Gangliosides are normal components of the plasma membrane and are abundant
in the nervous system. The ganglioside’s hydrophilic oligosaccharide chain is oriented
toward the extracellular environment, protruding on the membrane surface (88, 173). The
ganglioside pattern is well defined, but it varies greatly between the same tissues of
different species and also between species (75, 76). Sialic acid is the general term for the
series of compounds derived from neuraminic acid (5-amino-3, 5-dideoxy-D-glycero-
galactononulosonic acid). The major types of sialic acid detected in gangliosides are
N-acetylneuraminic acid and N-glycolylneuraminic acid (88, 173). The other
carbohydrates present in gangliosides are glucose, galactose, and either N-
acetylgalactosamine or N-acetylglucosamine (88).

The biosynthesis of gangliosides (Figure 12) occurs in the Golgi apparatus and
involves the sequential addition of saccharide units to ceramide, with each reaction owing
its catalyzation to glycosyltransferase. The strict specificity of many glycosyltransferases
holds the key to the formation of different gangliosides (75, 148, 164). Lactosyl-
ceramide is the precursor for all the gangliosides of the ganglio-series. Early insertion of

sialic acid into lactosylceramide leads to the ganglioside chain carrying a single sialic acid

12

 

 

 

 

CH2 OH O CHZOH CH3
ll
+H3N--C--H R—CfirC-H CH3-—N---CH3
| ¢H2
HO~C~H Acyl HO'C-H Acyl %H2 CDP
CoA CoA C-H CoA CoA Choline CMP
C-H , fi-H - O-P-O \J ,
H—L C CH--.CH..CH2
l l CH-NH
(CH2)12 (CH2)12 CH'é'O
(trim.
SPHINGOSINE CERAMIDE SPHINGOMYELIN
Galactog N-Acetylgalactosamine
Sialic acid Galactose
Glucose
Galactose Galactose
0H 3) Activated Sugars OH
CH— gI-I- H2 g. CH-CH—CHZ
CH— H CH—CH
CH-¢ =0 CH—C =0
(CHz)i2 (Qthz
H3 CH3
GALACTOCEREBROSIDE GM, GANGLIOSIDE

Figure 1.1: The structures of important ganglioside precursors.

13

 
  

       

Glycomotocns
... G""Qliosidn 5109mm

       
    
         
   
 

O

        
    

O : .0
. s
0.0 u

HlnNAc _ v.

09 Nova“— 5 "thesis

~

   
 

Direet ‘55
\m t) ”W" ’h

 

kcyc .
any Acid /
3!!an

NUCLEUS

 

Figure 1.2: Schematic drawing of ganglioside biosynthesis. From G. Tettamanti et al.,
1988.

14

residue linked to an inner galactose at the 3 position. The designation of ganglioside
monosialic acid 3 (GM3) is given to this specific ganglioside. Designations of GD3 and
GT3 are given to ganglioside disialic acid 3 and ganglioside trisialic acid 3, respectively.

An important factor leading to the testing of glycolipids as receptors for bacterial
adherence is the amount of diversity within this group (7, 75, 88, 121). Carbohydrate
molecules of varying complexities are normal features of cell surfaces throughout nature
and are termed collectively the glycocalyx (7). The glycocalyx encompasses
glycoproteins and glycolipids (which include gangliosides), as well as the mucosal
coating on epithelial surfaces, including the intestinal tract. These carbohydrate moieties
are usually in the form of oligosaccharide chains. The sugar sequences comprising these
chains show a high degree of complexity and variability, genetically determined, which
provides them with broad variability in their biological functions. These functions
include immunological specificity, cell-to-cell adhesions, and receptor-site capabilities
(75, 76, 87, 113, 115, 164). This diversity increases the possibility of these compounds
being specific receptors in many systems (164).

Binding has yet to be attempted using specific gangliosides as competitive
inhibitors of bacterial pili. Since the intestinal receptor sites for bacterial pili have been
demonstrated to be glycolipid in nature, the attempt to use gangliosides as a competitive
binding site is a relatively new approach (21, 44, 46, 54, 99, 150).

Cell-surface gangliosides have been shown to function as receptors for a variety of
toxins, lymphocytes, and biological mediators (4, 164). Gangliosides have also been
associated with tumors and used as differentiation markers (20, 22, 48, 61, 66). There is

substantial evidence to support the role of gangliosides in the regulation of immune

15

functions. The documented immune-regulative properties of gangliosides include

(a) inhibition of normal human lymphocytes’ response to mitogens and antigens (52, 86,
118, 132); (b) inhibition of cell proliferation, which is interleukin-2 dependent (105);

(c) inhibition Of cytotoxic functions (118, 169); and ((1) binding of CD4 on the
T-lymphocytes (78, 113, 122). These applications appear to have relevance in the area of

active immunotherapy with vaccines (96, 114, 115, 168).

Research has indicated that substances found in milk, especially colostrum, are
effective in immunologically protecting young animals as well as humans from diseases
(17, 50, 51, 57, 69, 125, 150). Milk, however, contains factors other than
immunoglobulins that also possess antimicrobial properties. Non-immunoglobulin
fractions identified in milk that have bacteriostatic or bacteriocidal effects include
chemotactic factors, lacto-ferrin, lysozyme, lactoperoxidase, interferon, and bifidous
factors (2, 73, 109).

Recently, lipid components of milk have been under investigation due to the
discovery of antimicrobial activity in this fraction. The antimicrobial activity resisted
boiling and digestion with trypsin but could be rendered ineffective by periodate
treatment (2, 73). This demonstrates that the activity is likely related to carbohydrate
residues, which may act as cell-surface receptors. Gangliosides naturally occur in milk
fat, as part of the milk fat globular membrane, and exhibit antimicrobial and enterotoxin-

inhibiting capabilities (2, 63, 79, 85, 87, 99).

16

Studies have shown that cell-surface gangliosides serve as cell-recognition sites
for the attachment of viruses, bacteria, and bacterial toxins (2, 13, 44, 54, 55, 76, 77, 78).
This attachment is attributed to the sialic acids and Oligosialyl structures in glycolipids,
which are associated with the microorganism’s cell wall (5, 146). Researchers now are
suggesting that the binding ability of gangliosides influences the growth, infectivity, and
pathogenicity of some microorganisms (44, 52, 60, 61, 68, 80, 82, 86). Lee and his
colleagues (89) found indirect evidence that milk oligosaccharides can bind to
Pseudomanas aeruginosa and inhibit the attaching of this organism to mucous. It has
also been found that the exotoxin of Vibrio cholerae can be neutralized by the binding of
gangliosides (63, 177) and that the binding of heat-liable enterotoxins from porcine and

human enterotoxigenic E. coli can be inhibited by gangliosides (150).

, .1 '11 E3 .11]

The acidic pH of the human stomach has caused scientists to conclude that
bacterial grth is not a problem in the stomach. In 1983, Marshall and Warren (101)
cultured a bacterium from a human gastric mucosa and showed an association between its
presence and gastric inflammation. This organism was named Helicobacter pylori
(H. pylori); it infects populations worldwide (9, 10). There is a strong association
between infection with H. pylori and gastric ulcers, with 70% of gastric ulcer patients
testing positive for H. pylori infection (131, 135, 175). If patients suffering from ulcers
induced by non-steroidal anti-inflammatory drugs are excluded, the association between
H. pylori and gastric ulcers increases to 96% (9, 10, 53, 131). Superficial gastritis can

result from the ingestion of H. pylori. If H. pylori is removed from the system, the

17

gastritis is alleviated (9, 53, 131). H. pylori is found almost exclusively in the stomach,
where it is closely associated with gastric epithelial cells or the mucosal layer. Critical to
the bacteria’s survival in the stomach is the production of urease. The production and
subsequent release of urease facilitate the production of ammonia, which functions as a
base and raises the pH of the immediate environment. H. pylori can therefore survive
well in the presence of urea (102).

Adhesion is an important prerequisite for many bacteria that colonize the mucosal
surfaces and may be important for the induction of inflammation by H. pylori (9, 74). H.
pylori expresses sialic-acid-specific hemagglutinins and also lipid-binding adhesions (95,
157). The H. pylori hemagglutinin also has been shown to have a high affinity for the N-
acetylneuraminyl-lactose (2-3 linkage) that predominates in bovine colostrum (95, 108,

157).

18

CHAPTER II

GANGLIOSIDE EXTRACTION AND PURIFICATION FROM
LACTIC ACID BACTERIA, CULTURED DAIRY FOODS,
AND BOVINE MILK FAT
Abstract

Gangliosides are a diverse group of naturally occurring complex sugar-containing
glycosphingolipids, which have sialic acid (N-acetylneuraminic acid) as an essential
component, as well as a hydrophobic, lipid component (ceramide). Brain, visceral organs
(especially the liver), erythrocytes, and other tissues contain a large number of different
gangliosides (75, 121, 148). Gangliosides of different types naturally occur in the milk
fat of all mammalian species, as part of the milk fat globular membrane, which is derived
from the plasma membrane of epithelial cells in the mammary gland (79, 170). The
average concentrations of gangliosides in the cow mammary gland and milk are 88.4 and
5.6 nmol/g wt, respectively, with nearly 90% of the milk gangliosides contained in the
milk fat globular membrane (79). Although the literature on this subject is incomplete, it
seems that the ganglioside compositions of milk fat from different mammalian species
can vary markedly (79, 85).

The ability to extract large amounts of crude mixed gangliosides from powdered
buttermilk and yogurt preparations was essential to Obtaining purified ganglioside

samples for further study. Several procedures were examined. The initial protocol

19

(Protocol 1) was based on F olch-like extraction as modified by Svennerholm and
Friedman (168). Sweet and cultured buttermilk were extracted at Oakland University,
while simultaneous extractions were performed on lactic acid bacteria (LAB) at Michigan
State University (MSU) laboratories. Identical procedures were used for extractions at
both laboratories. Subsequent attempts to remove non-ganglioside contaminants were
performed at Dr. Kathleen Moore’s laboratory at Oakland University and led to the use of
other protocols. Protocol 2 used an acetone pre-extraction step to remove natural lipids;
the extracted buttermilk powder was then processed as in Protocol 1. In Protocol 3,
buttermilk powder was resuspended in water, dialyzed, and lyophilized before extraction
by means of Protocol 1. Protocol 4 also involved pre-extraction dialysis and
lyophilization; the resuspended powder was then extracted with a series of
chloroform/methanol solutions and subsequently partitioned (138). Ultimately, an
extraction and Diethylaminoethyl (DEAE)-Sephadex A-25 (Sigma Chemical)
chromatography procedure as described by Bonafede et al. (14) yielded what were termed
purified preparations. These purified mixed ganglioside preparations, when quantitated
by the resorcinol assay, produced yields significantly lower than the crude preparations.
The yields from purified preparations produced a true purple/blue color (indicative of
gangliosides when sprayed with resorcinol) as opposed to the murky purple/brown color
obtained with crude preparations. The subsequent mixed ganglioside extractions were all

performed utilizing the purified preparation technique.

20

Introduction

There is substantial evidence that LAB have an inhibitory effect on pathogenic
bacteria in the intestinal tract (94, 103, 109, 178). Several theories have been proposed
for this bacterial antagonism; the most widely accepted theory is that of competitive
exclusion. This theory holds that LAB control the growth and colonization of
opportunistic enteric pathogens by effectively competing for nutrients, lowering the
luminal pH by producing lactic acid, and possibly producing antibiotic-like substances.
When an imbalance occurs in this arrangement, pathogens gain an advantage, ultimately
resulting in a disease state. This rationale is the basis for the controversial use of LAB as
a biological prophylaxis for enteric diseases in humans (45, 103).

Since the initial hypothesis of Metchnikoff (107), it has been believed that
cultured dairy foods containing LAB are beneficial in the prevention and treatment of
enteric diseases. LAB used for manufacturing cultured dairy foods have been shown to
modify the metabolic activities of the colonic microflora (103), possess antimicrobial
activity (178), and neutralize E. coli enterotoxin (109). Although various mechanisms
have been proposed for these beneficial effects, the involvement of LAB gangliosides has
not been considered as a mechanism.

Sialic acids have been detected in the cell membrane of Lactobacillus and
Streptococcus spp. (4, 144). Also, 1% to 2% of the total lipid content of buttermilk is
ganglioside, consisting of neuraminosyl lactosyl ceramide (20%) and dineuraminosyl
lactosyl ceramide (50%) (65). An unusual ganglioside, 9-0-acetyl-GD3, has been found
in bovine buttermilk (14), which suggests a possible involvement of LAB in modifying

milk fat gangliosides or producing other gangliosides not present in the original milk fat.

21

Mixed gangliosides were extracted from cultured dairy foods (cultured buttermilk,
yogurt), sweet buttermilk, LAB, and evaporated milk using a modification of the method
of Bonafede et al. (14). The extraction process, which was labor intensive and had not
previously been developed specifically for cultured dairy foods (CDF), bovine milk fat
(BMF), or LAB, required the application of existing protocols to the unique and
specialized challenges of ganglioside extractions from these products. In addition to the
extraction process, mixed gangliosides were purified with a DEAE- Sephadex A-25
(Sigma Chemical) column, producing samples with fewer contaminants. Although LAB
ganglioside extraction initially was undertaken at MSU, the major task of ganglioside
extraction from dairy products (cultured buttermilk, sweet buttermilk, yogurt, evaporated
milk) and LAB occurred at Dr. Kathleen Moore’s laboratory in the Department of

Chemistry at Oakland University.

Materials and Methods

3 l. . 1 5

Two starter cultures were used. One (designated 8803) was supplied by Darigold,
Seattle, WA, and used in their cultured buttermilk product. The second (designated R-39)
was supplied by Chris Hansen’s laboratory, Milwaukee, WI, and used in producing the
yogurt supplied by the MSU Dairy Plant. The starter cultures contained a mix of different
bacterial strains, including Streptococcus lactis, Streptococcus lactis subspecies
diacetylactis, Streptococcus cremoris, and Lactobacillus citrovarum. The specific
amounts of each bacteria strain are a trade secret and therefore, are not available for

reporting.

22

Cultured buttermilk was prepared from skim milk treated with the starter culture,
and sweet buttermilk was derived from the liquid remaining after butter was processed.
Both butterrnilks were obtained in the powdered form. Yogurt is also produced with a
starter culture, usually containing Streptococcus thermophilus and Lactobacillus
bulgaricus. Mixed ganglioside extractions and purifications were performed on LAB
obtained from a buttermilk starter culture (Darigold, Seattle, WA), sweet and cultured
powdered buttermilk (Darigold, Seattle, WA), and plain cultured yogurt (MSU Dairy
Plant, East Lansing, MI) using a modification of the method of Bonafede et al. (14).

The difficulty in growing large batches of LAB in order to extract gangliosides
proved to be a major challenge. Various media were used to facilitate growth of LAB
from a starter culture. Media tested included de Man-Rogosa-Sharpe (MRS) broth (Difco
Laboratories, Detroit MI) , 1/2 MRS with 1% lactose (Sigma Chemical), MacConkey’s
medium (Difco), and a powered milk-based medium. The most successful growth
occurred using the 1/2 MRS broth fortified with 1% lactose. Because of the small yields
of gangliosides produced through this labor-intensive method, the previously described

starter cultures were ultimately used as ganglioside sources.

1"131'ilEX ‘E 1
The initial extractions were based on a modification of the method of
Svennerholm and Friedman (168). This procedure depends on the ability of gangliosides
to partition from the chloroform-methanol phase into which they are extracted, into an
upper aqueous phase. This partitioning should ensure that most of the other undesired

lipids remain in the chloroform phase.

23

One hundred grams of dried dairy product were dissolved in 318 ml of deionized
water. This solution was stirred for 8 min using a magnetic stirrer bar. Then 455 ml of
methanol (MeOH) and 228 ml of chloroform (CHCl3) were added, and stirring continued
for 30 min. The solution was poured into a separatory funnel and allowed to stand for 4 h
until two distinct phases were observed. The upper phase was set aside for later use. To
the lower phase, MeOH at 1/2 the remaining volume (approximately 250 ml) and 0.01 M
KCl in water at 1/3 the volume of the lower phase (approximately 173 ml) were added
and lefi to stand overnight. The upper phase was then combined with the previous upper
phase, rotary evaporated, and ammonium sulfate added until saturated (approximately 60
to 70 g), and stirred 15 min. The sample was centrifuged for 20 min at 10,000 rpm and
the liquid poured into dialysis tubing. The sample was dialyzed against running tap water
for 4 to 6 h and refiigerated at 4°C overnight. The tap water was replaced with deionized
water at 4°C for 24 h, the water changed, and kept at 4°C an additional 24 h. The dilute
preparation was removed and lyophilized to produce a concentrated crude preparation.
The lyophilized sample was resuspended in deionized water and used for thin-layer
chromatographic (TLC) analyses. These preparations were subjected to TLC analyses;
sialic acid standards and ganglioside standards (purified mixed brain gangliosides and
Disialoganglioside 3 (GD3) were routinely run with samples.

In an attempt to obtain better yields and cleaner preparations, four different
protocols were employed in the isolation of crude ganglioside mixes. Protocol 1 involved
a Folch-like extraction as modified by Svennerholm and Friedman (168); this procedure
was described above. Protocol 2 used an acetone pre-extraction step to remove neutral

lipids (170); the extracted buttermilk powder was then processed as in Protocol 1. In

24

Protocol 3, the buttermilk powder was resuspended in water and dialyzed and lyophilized
before extraction by means of Protocol 1. The dialysis step has been shown to remove
more than 50% of the powder solids (138). Protocol 4 also involved pre-extraction
dialysis and lyophilization; the resuspended powder was then extracted with a series of

chloroform/methanol solutions and partitioned as described by Ren et al. (138).

19.513 1"1E 'El

Cultured buttermilk was prepared from skim milk treated with a starter culture
(designated 8803 and supplied by Darigold, Seattle, WA) , and sweet buttermilk was
derived from the liquid remaining after butter was processed. Both buttermilks were
obtained in the powdered form. Yogurt is also produced with a starter culture,
(designated R-39 and supplied by Chris Hansen’s laboratory, Milwaukee, WI.). The
yogurt starter culture usually containing Streptococcus thermophilus and Lactobacillus
bulgaricus. Mixed ganglioside extractions and purifications were performed on LAB
obtained from a buttermilk starter culture (Darigold, Seattle, WA), sweet and cultured
powdered buttermilk (Darigold, Seattle, WA), and plain cultured yogurt (MSU Dairy
Plant, East Lansing, MI) using a modification of the method of Bonafede et al. (14).

Specifically, 100 g of dried dairy product (powdered sweet buttermilk, powdered
cultured buttermilk, or lyophilized yogurt) were suspended in 1,000 ml of chloroform:
methanol (1:1) and allowed to stir overnight. After solid material was removed by
Buchner filtration, the extract was subjected to rotary evaporation for removal of
solvents. The dried extract was resuspended in 50 ml of water and dialyzed against

deionized water for 2 days in a cold box. The dialyzed sample was lyophilized and

25

resuspended in chloroformzmethanolzwater (C :M:W; 30:60:8) for anion-exchange
chromatography. The sample was applied to a 1.5 cm x 30 cm column of DEAE-
Sephadex A-25 (Sigma Chemical), which had been preequilibrated with the C:M:W
solvent. The column was then washed with 175 ml of the C:M:W solvent and 75 ml of
methanol. Gangliosides were eluded from the column with 150 ml of 0.4 M ammonium
acetate in methanol. Ganglioside-containing fractions were pooled, and methanol was
removed by rotary evaporation. The dried ganglioside fraction was resuspended in
deionized water and dialyzed against deionized water for 2 days to remove ammonium
acetate. The dialyzed ganglioside fraction was lyophilized and resuspended in minimal
deionized water for quantitation, TLC analyses, and hemagglutination studies. The

ganglioside fractions were stored at -32°C and thawed to room temperature prior to use.

i l' . . n

LAB gangliosides were extracted in a manner similar to purified gangliosides
except that the LAB cells were ruptured by glass beads in a chilled BeadBeator chamber
(Fisher Scientific, Atlanta, GA) and cell membranes were recovered after centrifugation
of homogenate. The membrane pellet was then extracted with a 20-fold volume of
chloroformzmethanol (1:1) and processed as described above.

Starter cultures used for the production of yogurt were obtained from Chris
Hansen’s laboratory (Milwaukee, WI). The starter culture (R-39) contained multiple
strains Of Streptococcus thermophilus and Lactobacillus bulgaricus. TO produce large
quantities of LAB for the isolation of the bacterial-cell-wall gangliosides, 0.2 ml of starter

culture was inoculated per liter of MRS Broth (BBL, Becton Dickinson Microbiology

26

Systems, Cockeysville, MD) and incubated at 39.5 °C for 72 h. The resultant culture
yielded approximately 7 g of packed wet cells per liter of medium when harvested by
centrifugation at 7,650 x g for 20 min. Approximately 1,400 g of wet bacterial cells were
initially propagated and harvested for ganglioside isolation using this method.

The starter culture (8803) for buttermilk, which was obtained from Darigold,
Seattle, WA, was also cultured. The bacterial strains used in this starter culture were
proprietary; therefore, the specific bacteria strains used were not identified. The
bacterial-cell-wall gangliosides were isolated and purified at Oakland University in the
laboratory of Dr. Kathleen Moore.

With respect to the isolation of gangliosides from LAB, the typical preparation
involved motor-driven homogenization of 60 g of bacterial cells (wet weight) with
100 ml of sucrose phosphate buffer, pH 7.4. The mixture was then transferred into a
chilled BeadBeator chamber (Fisher Scientific, Atlanta, GA). The bacteria cells were
ruptured by glass beads for three 1-min intervals; each l-min process was followed by a
10-min cool-down. The process was done in a cold box.

After treatment was complete, the beads were allowed to settle before decanting
of the homogenate. The beads were washed twice with 50 ml of buffer; the washes were
then combined with the homogenate. The homogenized bacteria mix was centrifuged for
20 min at revolutions per minute = 25,000. This procedure pelleted the bacterial
membranes, which contained most of the gangliosides. The supernatant was lyophilized
and then extracted for its minimal ganglioside content. The bacterial—membrane pellets
were weighed and gradually homogenized with chloroformzmethanol (1:1 v/v, 20 ml of

solvent per gram of pellet). After overnight extraction, the process continued as

27

described earlier, according to the method of Bonafede et a1. (14). Eventually,
lyophilized samples were extracted by Dr. Moore’s laboratory using the method

previously described.

i n ' ' 11

Thin layer chromatography (TLC) was performed using silica gel 60 TLC
precoated plates (glass plates 10 x 10cm Merck). The plates were activated at 100° C and
ganglioside preparations and ganglioside and N-acetylneuraminic acid (NANA) standards
were spotted on silica gel G plates and placed in a chamber containing chloroform:
methanol:0.25% CaCLz solution (30: 18:4). Plates were developed in the chambers for
twice the time it took the solvent to reach the top of the plate. Plates were then removed,
dried, sprayed with resorcinol-HCL for visualization, and heated at 110°C for 15 min.
Sialic-acid-containing material gave blue spots; organic material lacking sialic acid
charred brown. Densitometry of TLC plates was performed with a Shimadzu Dual-
Wavelength TLC-Scanner CS930. The chromatograms were obtained by single
wavelength zig-zag scanning of each lane at 620 nm (beam dimension: 0.4 X 0.4mm) in
the reflection mode (168). The purified preparations yielded blue bands of varying
patterns. All types of preparations had a band that migrated with standard GD3; LAB
preparations had only one additional band, whereas the dairy products had six to eight
separate ganglioside bands. The yields resulting from this process were small, but the

preparations were highly purified and devoid of extraneous contamination.

Results

28

Two basic methods were used to characterize the crude ganglioside mixes. Rough
quantitation was obtained by the resorcinol/HCI spectrophotometric assay method (168).
This assay quantitates sialic acid content of the preparations, but larger quantities of
ganglioside may be artifactually suggested due to enhancing interference by other
partially reactive contaminants in our crude preparations. Appendix A.1 (Appendix p.
33-3 5) contains an inventory of crude mixed ganglioside preparations by protocol (1
through 4); included are the volume extracted, the ganglioside concentration, and the total
ganglioside present on a micromole (umol) basis. The quantitative assessment of our
buttermilk and yogurt preparations are summarized in A2 (Appendix p. 36).

A summary of the present research with crude preparations is given in A3
(Appendix p. 36). Aliquots of purified gangliosides were reacted with resorcinol-HCL,
and absorption at 620 nm was measured. Sample absorption was compared to that of
NANA standards to determine sialic acid content. Mean yields on a nmol/ g dry weight
basis were as follows: sweet buttermilk = 11.6 i 8.1, cultured buttermilk = 7.3 :t 5.0,
yogurt = 13.2 it: 5.6, yogurt starter culture LAB mix (Chris Hansen R-39) = 16.7 i 1.5,
and buttermilk starter LAB mix (Darigold 8803) = 65.0 :1: 13.7. The crude mixed
ganglioside preparations were evaluated qualitatively by examining thin-layer
chromatograms. The developed plates were sprayed with resorcinol/HCI reagent and
heated; sialic-acid-containing bands stained a distinctive purple/blue color, whereas non-
sialic-acid contaminants gave nondistinctive brown spots. The brown material was
frequently at the origin and occasionally migrated into the plates. In terms of qualitative
observation of TLC plates in the bacterial preparations, 3 band that correlates with

disialoganglioside—3 was predominant, with a fainter, higher mobility unknown band. In

29

the yogurt preparations, up to eight bands, including disialoganglioside-3, were visible; in
terms of color, the bands were moderate to strong in intensity. For the cultured
buttermilk preparations, up to Six bands, including disialoganglioside-3, were visible;
again, color intensity was moderate. With the preparations from sweet buttermilk, up to
six bands, including disialoganglioside-3, were visible; color intensity was moderate to
strong. The TLC observations for the different protocols are summarized in A4
(Appendix p. 37).

Protocol 4 gave fairly consistent quantitative yields, as well as good quality
ganglioside bands. In addition, the mean value for sweet buttermilk (36.5 i 13.9 umol/
100 g powder) was close to the expected yield from sweet buttermilk powder (48 umol/
100 g powder) (65). From comparison with standards, the primary component of crude
ganglioside mixes from both the sweet and cultured buttermilk appears to be GD3.

The average total sialic acid present in umol for different dairy products and LAB
from purified preparations A.5 (Appendix p. 38) was (a) cultured buttermilk = 1.4 umol i
3.3, (b) sweet buttermilk = 2.6 umol d: 6.1, (c) yogurt = 2.18 umol :l: 3.25, and (d) LAB =
0.426 umol i 0.44.

Yogurt from the MSU Dairy Plant was subjected to extraction by means of
Protocols 1 and 4. Briefly, Protocol 1 was based on Folch-like extraction as modified by
Svennerholm and Friedman (168). Protocol 2 used an acetone preextraction step; the
extracted buttermilk powder was then processed as in Protocol 1. In Protocol 3, the
buttermilk powder was resuspended in water, dialyzed, and lyophilized before extraction
by means of Protocol 1. Protocol 4 also involved preextraction dialysis and

lyophilization; the resuspended powder was then extracted with a series of chloroform/

30

methanol solutions and subsequently partitioned (138). Protocol 1 extractions resulted in
extremely high resorcinol quantitation values; these same preparations showed no
evidence of ganglioside bands afier TLC. Because of these unsatisfactory results, we
switched to Protocol 4. Preparations performed using this protocol provided purer
quantitation results, and TLC analysis usually resulted in an N-acetylneuraminic acid
(NANA) band and some ganglioside bands (not GD3). The preextraction dialysis was
helpful in removing much of the unwanted yogurt solids; however, these preparations
remained viscous.

Of the four protocols performed, Protocol 4 provided the best results in terms of
purity, but still contained contaminants. Ultimately, an extraction and DEAE-Sephadex
A-25 (Sigma Chemical) chromatography procedure as described by Bonafede et al. (14)
yielded what were termed purified preparations. The ptu'ified preparations provided a
low-yield, highly purified sample. The subsequent mixed ganglioside extractions were all

performed using the purified preparation technique.

D' .
The isolation of large amounts of crude mixed gangliosides from powdered
buttermilk, yogurt, and LAB used to produce cultured dairy products was the initial goal
of this part of the present research project, which was to be accomplished by means of a
Folch-like extraction (168). This procedure (Protocol 1) was used for the LAB
extractions at MSU and for the sweet and cultured buttermilk and yogurt extractions at

the Oakland University laboratory. The difficulty in growing large batches of LAB in

31

order to extract gangliosides proved to be a major challenge. The most successful growth
occurred using a 1/2 minimal MRS medium fortified with 1% lactose.

In general, crude preparations are high-yield, low-purity processes that are useful
for estimating ganglioside content in dairy products and bacteria. Several different
procedures were tried, resulting in a range of values. Purified preparations are high-
purity, low-yield processes that are useful for in vitra anti-microbial studies. To eliminate
interference by nonganglioside product components, a large percentage of gangliosides
are lost in the process. The results of the purified preparation process are listed in A6
(Appendix p. 38-39). The purified preparation procedures achieved the goal of
decreasing the amount of contaminants present in samples, indicated by TLC results and a
decrease in mixed gangliosides (actually a reduction in ganglioside contaminants). The
lower yield of crude gangliosides from cultured buttermilk extracted by Protocol 4, Table
A (Appendix p. 34-36) is reasonable because skim milk contains less than 8% of the
normal milk gangliosides (65). The ganglioside content of our cultured buttermilk
powder was probably derived primarily from the bacterial components of the cultured
dairy product, due to the lack of fat content in skim milk. The dairy products, whether
cultured or not, appeared to have a more diverse group of ganglioside components than
did the LAB. When the purified mixed ganglioside preparations were quantitated by the
resorcinol assay, the yields were significantly lower than those calculated for the crude
preparations (sweet buttermilk = 1.62 3: 0.34 umol/ 100 g powder; cultured buttermilk =
0.82 d: 0.89 umol/ 100 g powder), but the color reaction was a true purple/blue, not a
murky mauve. Likewise, TLC analysis showed smaller but purer amounts; therefore, it

was decided to process all of our mixed ganglioside preparations by this method.

32

A. 1: Inventory of ganglioside preparations, by various protocols.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Concentration Total Gangliosides
Preparation No. Volume (ml)
(mM)a Present ([1)
PROTOCOL 1: CULTURED BUTTERMILK
4 6.1 11.10 67.71
5 24.0 4.75 114.00
6 10.5 2.90 30.45
7 14.0 2.70 37.80
8 12.6 1.99 25.07
9 17.5 1.58 27.65
10 15.5 1.99 30.84
12 15.5 3.13 48.15
13 13.0 2.50 60.00
16 13.0 2.80 36.40
17 7.7 0.83 6.39
18 12.0 3.13 37.20
27 6.7 6.50 42.90
28 9.5 11.20 616.00
47 12.5 6.00 75.00
49 9.5 5.70 54.15
50 9.3 8.00 74.40
PROTOCOL 1: SWEET BUTTERMILK
19 10.0 3.30 33.00

 

33

 

A. 1: Continued.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Concentration Total Gangliosides
Preparation No. Volume (ml)
(mM)a Present (u)
20 5.0 2.20 11.00
21 6.3 1.75 11.03
22 6.3 1.53 9.64
25 7.0 0.63 4.41
PROTOCOL 1: YOGURT
3 20.0 24.30 486.00
8 21.0 9.70 203.70
9 21.0 10.70 224.70
PROTOCOL 2: CULTURED BUTTERMILK
11 7.1 0.99 6.99
14 16.0 2.00 32.00
15 30.0 1.70 51.00
PROTOCOL 2: SWEET BUTTERMILK
24 11.0 2.70 29.70
26 7.0 2.90 20.30
PROTOCOL 3: CULTURED BUTTERMILK
29 44.0 11.20 616.00
37 10.0 1.52 15.20
PROTOCOL 3: SWEET BUTTERMILK
32 34.0 5.20 176.80
38 22.0 3.88 85.36
PROTOCOL 4: CULTURED BUTTERMILK
31 6.0 0.40 2.40

 

 

 

 

34

 

A. 1: Continued.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Concentration Total Gangliosides
Preparation No. Volume (ml)
(mM)a Present (11)
35 12.0 1.50 18.00
39 10.0 2.38 23.80
45 5.2 1.50 7.80
46 5.2 0.00 0.00
PROTOCOL 4: SWEET BUTTERMILK
30 7.0 3.00 21.00
36 9.0 0.90 8.10
40 10.0 2.70 27.00
41 7.3 5.40 39.42
42 8.2 5.70 46.74
43 8.5 5.50 46.75
44 5.7 7.60 43.32
51 5.2 7.85 40.56
52 8.5 5.60 47.60
PROTOCOL 4: YOGURT
1 6.0 2.70 16.20
2 6.0 3.20 19.20
4 8.5 2.8 23.80
5 19.0 6.25 56.25
6 9.0 1.50 18.75
a Micromolar

35

 

A.2: Quantitation of crude mixed gangliosides by resorcinol assay.

 

 

 

 

 

 

 

 

 

 

Ganglioside Content (umol sialic acid/ 100 6 powder)
Protocol
Cultured Buttermilk Yogurt Sweet Buttermilk
1 46.4 1 25.70 33.3 d: 34.8 304.8 :1: 24
2 30.0 d: 22.13 25.0 i 06.5 -—
3 25.0il4.42 71.9i51.1 —
4 16.5 i 8.13 35.6:t 13.9 32.0:t27

 

A.3:Qualitative thin-layer chromatography analyses of crude mixed gangliosides

from sweet and cultured buttermilk using four different protocols.

 

 

 

 

 

 

 

 

 

Mean Relative Prominence of Components
Protocol NANAa GD3b Gangliosides Contaminant
s
l—Cultured buttermilk + — + ++++++
Sweet buttermilk +++ +++++ ++++ ++++++
2—Cultured buttermilk + _. _ +++++++
Sweet buttermilk — +++—r _ ++++
3—Cultured buttermilk ++++ ++++ ++++ ++++
Sweet buttermilk ++++ ++++ _ ++++
4—Cultured buttermilk +++ +++ +++ ++++
Sweet buttermilk ++++ ++++ ++++ ++++

 

 

 

 

a N-acetylneuraminic acid

b Disialoganglioside-3

36

 

A.4: Summary of the ganglioside content of crude preparations from dairy products.

 

 

 

 

 

 

 

Ganglioside Content Product Concentration
Dairy Product
(nmol/g dry weight) (umglfl)
Sweet buttermilk 250-719 21-61
Cultured buttermilk 165-464 21-58
Yogurt 800 114
Whole milk — — 5.6

 

 

A.5: Summary of the ganglioside content of purified preparations from dairy

products and LAB.

 

Ganglioside Source

Ganglioside Contenta (nmol/g dry weight)

 

 

 

 

LAB
R-39 26.7 i 8.5
8803 65.0 i 13.7
Dairy products
Sweet buttermilk 11.6 i 8.1
Cultured buttermilk 7.3 i 5.0
Yogurt 13.2 i 5.6

 

alAll concentrations are based on sialic acid content as determined by resorcinol
assay. Ganglioside content (nmol/g dry weight) was calculated using an average
ganglioside molecular weight (1,500 g/mol), approximately that of GD3.

37

 

 

A.6: Inventory of purified ganglioside preparations from dairy products.

 

Preparation No.

 

Concentration (mM) Volume (ml)

 

 

Total Present (umol)

 

 

 

 

 

 

 

 

 

 

 

SWEET BUTTERMILK
1 0.126 34.0 4.300
3 0.126 15.0 1.887
4 0.107 19.0 2.033
5a 0.046 19.0 0.874
6 0.580 7.0 4.060
7 0.186 4.9 0.910
8 0.046 6.5 0.299
9b 0.080 10.5 0.840

10 0.104 48.4 5.030

11 0.760 10.0 0.760

12 0.315 21.7 6.836

13 0.127 21.1 2.668

14 0.306 11.4 3.488

18 0.334 15.0 5.010

19 0.465 9.0 4.185

24 0.264 5.0 1.320

CULTURED BUTTERMILK
1 0.023 18.0 0.414
2 0.074 20.0 1.480
3 0.030 7.0 0.212
4 0.154 8.0 1.232
5 0.232 8.0 1.859
6 0.050 3.1 0.155
7 0.034 10.1 0.343
8 0.136 14.8 2.013
9 0.086 6.2 0.533

10 0.136 19.5 2.652

11 0.136 19.7 2.679

12 0.408 8.5 3.468

13 0.082 7.5 0.615

17 0.360 6.2 2.230

18 0.940 5.8 0.545

19 0.316 10.0 3.160

 

38

 

A.6: Continued.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Preparation No. Concentration (mM) Volume (ml) Total Present (umol)
YOGURT
1 0.250 24.0 6.000
2 0.180 20.0 3 .600
3 0.000 0.0 0.000
4 0.064 16.0 1.024
5 0.015 3.2 0.048
6 0.176 22.3 3.925
7 0.256 8.8 2.258
8 0.406 8.0 3.248
12 0.483 5.4 2.608
18 0.260 4.7
LACTIC ACID BACTERIA
B-PP-S (R-39) 0.032 7.6 0.243
B-PP-6 (R-39) 0.048 6.1 0.293
B-PP-7 (R-39) 0.052 12.0 0.624
B-PP-ll (8803) 0.080 6.8 0.540
EVAPORATED MILK (12 02/350 ml can)
EVAP-l 0.372 5.0 1.860
EVAP-2 0.384 6.0 2.304
EVAP-3 0.336 9.0 3.024
EVAP-4 0.440 8.0 3.520
SIMILAC
SIMI-l 0.056 6.0 0.336
SOY MILK
SOY-1 0.200c 7.8 1.560
SOY-2 0.060c 12.0 0.720
aOld samples.
b(.15 M).

cTLC plates produced spots that were orange in color, indicating that they probably
were not gangliosides.

39

CHAPTER III

BACTERIOSTATIC EFFECTS OF GANGLIOSIDES AS DETERMINED
BY A DOUBLE AGAR OVERLAY ASSAY TECHNIQUE

£42m

Using a growth-inhibiting assay, the antimicrobial activity of mixed gangliosides
extracted from sweet buttermilk and cultured buttermilk, yogurt, and lactic acid bacteria
(LAB) against K88* and K99+ Escherichia coli (E. coli) was determined. Screening
small concentrations of extracted mixed gangliosides for growth inhibition of enteric
pathogens using conventional microbiological methods is tedious and often produces
conflicting results (2, 14, 123). A semi-quantitative plate assay was developed to
determine the antimicrobial activity of mixed gangliosides. A double agar overlay
technique, in which the top layer was a plated lawn of the test pathogen, was the basis for
this assay.

Mixed gangliosides were placed into shallow wells that were punched into the
pathogen lawn. Zones of clearing were evident around the wells and corresponded to
growth inhibition of the pathogenic bacteria. The zones were transitional, but they
persisted from 2 to 16 h after incubation. In addition, samples taken from wells and
incubated in a minimal medium for up to 16 h produced no bacterial growth.

Mixed gangliosides extracted from sweet buttermilk and cultured buttermilk,

yogurt, and LAB were effective in inhibiting the growth of K99+ E. cali (strain

40

431,0101:K30), K88+ E. coli (strain 19K974,0141:K85:K88), and E. cali 0157:H7 in the
range of 0.09 mM to 0.51 mM. Mixed gangliosides extracted from lactic acid bacteria
were the most effective in concentrations ranging from .052 mM to .028 mM in
producing zones of inhibition. Standard bovine brain gangliosides (BGG) (Sigma
Chemical, St. Louis) were the least effective, with a concentration of 8.4 mM being

required to produce inhibition zones.

Intmduction

Throughout the world, enterotoxigenic Escherichia coli (ETEC) is an important
cause of neonatal diarrhea in pigs, calves, and lambs, as well as in humans (34, 64, 119,
143). Two virulence determinants are necessary for ETEC to cause diarrhea. The first is
enterotoxin production, and the second is pilus production. Pili are hairlike structures
present on the outer cell membrane of E. coli, which are required for attachment to occur.

Hemagglutination was first recognized as an adhesive property Of E. cali in the
early 1900s. In 1955, it was demonstrated that the presence of pili on the E. cali surface
was correlated to this adhesive property (5, 33). Antigenically distinct pili present on
separate strains of E. coli are capable of infecting neonatal calves K99, F41 (19, 92), pigs
K99, K88, 987P, F41 (84, 94, 153), lambs K99 (47), and humans (CFA-I and CFA-II)
(37, 38, 39, 40). Pili allow E. coli to attach to specific receptors on the mucosal surface
of the small intestine, which is the initial step in the disease process (56, 64, 119).
Present knowledge about the chemical nature of the receptors indicates that they are
composed of glycolipids and/or glycoproteins (119). Previous research has indicated that

gangliosides, naturally occurring complex sugar-containing glycosphingolipids, which

41

have Sialic acid as an essential component, serve as recognition sites for the attachment of
viruses, bacteria, and bacterial toxins (75, 76, 80). This attachment is attributed to the
presence of sialic acids and oligosialyl structures in glycolipids that are part of the
microbial cell wall (76, 144). There is growing evidence that, because of this binding
ability, gangliosides can influence the growth, infectivity, and pathogenicity of
microorganisms. The neutralization of Vibrio cholera pathogenicity (63, 177) and the
inhibition of binding of Pseudomonas aeruginasa to mucous (80) are two examples of the
antimicrobial activity of gangliosides. Various glycoproteins and glycolipids may be able
to competitively inhibit the binding of piliated E. cali to brush borders, epithelial cells, or
erythrocytes (99, 116, 120, 183). The N-acetylneuramic acid portion appears to play an
important role in the inhibition of binding. Previous studies (133) have demonstrated that
mixed gangliosides at low levels exhibit bacteriostatic activity against K88+ E. coli and
K99+ E. coli. in vitra. A sensitive plate assay to determine the bacteriostatic activity of

mixed gangliosides was developed using a double agar overlay technique.

M ri l M h
l r v rl h i
To test the grth inhibition of enteric pathogens by mixed gangliosides, an assay
involving a double agar overlay was developed. MacConkey’S minimal medium (Difco
Laboratories, Detroit, MI) (1.5%) was used for the base layer, and de Man-Rogosa-Sharpe
(MRS) minimal medium (Difco Laboratories, Detroit, MI) (0.9%) , was used for the top
layer or lawn (containing the pathogen). Ten to 15 ml of MRS base agar, previously

prepared, were heated to boiling, allowed to cool, and poured into a sterile petri dish. Ten

42

to 15 ml of top agar stored in test tubes were heated to boiling and then cooled to 45°C (in
order to avoid the microbial killing reported at higher temperatures of only 50°C [164]).
The top agar was inoculated with an enteric pathogen either K88+ Escherichia coli (strain
80-022), K99+ Escherichia cali-45 (strain 455133), or K99+ Escherichia cali-4 7 (strain
472304P) provided by the lab of Dr. Robert Holland, Department of Large Animal
Clinical Science, Michigan State University or E. cali 0157:H7 provided by the lab of Dr.
Julius Jackson, Department of Microbiology, Michigan State University. All tests were
performed with 24-h cultures of bacteria (108/m1), grown in Minca broth with IsoVitaleX
(BBL/Becton Dickinson, Cockeysville, MD). The cultures were incubated aerobically at
37°C in a shaking water bath prior to use. The pathogens and top agar were then gently
titrated to encourage even distribution, and poured over the base agar. After the agar
congealed, wells (8 mm in diameter and approximately 4mm in depth) were punched
(using a brass cork corer) into only the top agar. Seventy-five mM of various mixed
gangliosides were added to wells. The plates with innoculated mixed gangliosides were
incubated at 34°C with grth inhibition zones measured at designated intervals.

Zones of clearing, or growth-inhibition zones, were measured and designated
qualitatively as follows: clear—no apparent growth, cloudy—a small amount of growth,
opaque—some growth but much less than outside of the zone, and growth—no apparent
inhibition. designations of clear to opaque. Zones of clearing were measured from the
edge of the well to the outer edge of the zone, in millimeters (mm). Growth within the
zone was also designated as clear (no apparent growth), cloudy (a little growth, but
printed text could be read when placed under the petri dish), opaque (some growth, but

printed text could not be read when placed under the petri dish), and growth (no apparent

43

growth inhibition). The zone sizes and designations of clarity were averages (a minimum
of three replications at each concentration), with the clarity being reproducible (no
averaging of clear and opaque to produce cloudy was done). Attempts were made to
reduce the zones of clearing by1/2 and 1/4, usingl/2 and 1/4 the amount of samples that
produced the original inhibition zones. Negative controls used were autoclaved
gangliosides, autoclaved distilled water, and empty wells.

Gangliosides used for inhibition were bovine brain gangliosides (Sigma
Chemical); two starter cultures one designated 8803 from a cultured buttermilk source
(Darigold, Seattle, WA) and the second designated R-39 (Chris Hansen’s laboratory,
Milwaukee, WI), which is used in the production of yogurt supplied to the Michigan State
University Dairy Plant; and lactic acid bacteria gangliosides obtained from a buttermilk
starter culture (Darigold, Seattle, WA). Sweet and cultured buttermilk, yogurt and lactic
acid bacteria gangliosides were extracted in the lab of Dr. Kathleen Moore, Department
of Chemistry at Oakland University, Michigan. The ganglioside extraction protocol used
was based on a F olch-like extraction as modified by Svennerholm and Friedman (168).
Modifications to this procedure included: an acetone pre-extraction step to remove
natural lipids; dialysis and lyophilization; the resuspension of powdered milk or yogurt in
water which was then extracted with a series of chloroform/methanol solutions and
subsequently partitioned (75, 168). Ultimately, an extraction and diethylaminoethyl
(DEAE)-Sephadex A-25 (Sigma Chemical, St. Louis, MO) chromatography procedure as
described by Bonafede et al. (14) yielded what were termed purified preparations. Thin
Layer Chromatography (TLC) was performed using silica gel 60 TLC precoated plates

(glass plates 10 x 10cm, Merck) for ganglioide quantitation. The plates were activated at

44

100° C and ganglioside preparations and ganglioside and N-acetylneuraminic acid
(NANA) standards were spotted on silica gel G plates and placed in a chamber containing
chloroform: methanol:0.25% CaCL2 solution (30:18:4). Plates were developed in the
chambers for twice the time it took the solvent to reach the top of the plate. Plates were
then removed, dried, sprayed with resorcinol-HCL for visualization, and heated at 110°C
for 15 min. Sialic-acid-containing material gave blue spots; organic material lacking
sialic acid charred brown. Densitometry of TLC plates was performed with a Shimadzu
Dual-Wavelength TLC-Scanner CS930. The chromatograms were obtained by single
wavelength zig-zag scanning of each lane at 620 nm (beam dimension: 0.4 X 0.4mm) in
the reflection mode (88, 168). The purified preparations yielded blue bands of varying
patterns. These purified mixed ganglioside preparations, were used for the grth

inhibition studies.

B ' - i

The concentration of viable cells is sometimes roughly estimated by a method
designated as the Most Probable Numbers (MPN) (30). This procedure is based on a
mathematical calculation of multiple cultures that fail to Show grth in a series of
dilution tubes containing an appropriate medium. As a modification of the MPN method,
ten ul of mixed gangliosides were recovered from sample wells where zones of clearing
were evident. Three replicates are the minimum performed for bacterial-growth studies.
Ten-microliter aliquots were recovered at 2, 4, 6, and 8 hour intervals. Wells from which
samples were obtained were not used for zones-of-clearing measurements because the

reduced amounts in the wells may have decreased the inhibition zones. The 10 pl

45

aliquots were then inoculated into 5-ml tubes of 1/2 de Man—Rogosa-Sharpe (MRS) broth
(Difco Laboratories, Detroit MI) with 1% lactose (Sigma Chemical, St. Louis MO) and
incubated at 37°C for the intervals indicated. Using a spectrophotometer, optical density

readings were taken at 2-, 4-, 8-, 12-, and 16-h intervals to test for any growth inhibition.

Results

A sensitive plate assay to determine the bacteriostatic activity of mixed
gangliosides was developed using a double agar overlay technique. A photograph of this
assay is depicted in Figure 111.1. The results of the double agar overlay technique with
enteric pathogens are presented in Figures 111.2 through 111.13. The ability of mixed
gangliosides to inhibit growth varied with the ganglioside concentration, the ganglioside
source, and the specific pathogen being tested. Purified mixed gangliosides were the only
samples tested experimentally. Early crude ganglioside extractions were used with this
technique, but the purified gangliosides produced more consistent reproducible results
due to the purification process. Purified mixed gangliosides from sweet buttermilk,
cultured buttermilk, yogurt, and lactic acid bacteria with ganglioside concentrations in the
range of 0.018mM to 19.4mM were effective in inhibiting the growth of K99+ E. cali,
K88+ E. cali, and E. coli 0157:H7.

Zones of clearing were transitory, with zones fading from clear to cloudy, turning
opaque, and gradually disappearing as bacterial growth covered the area. Depending on
the grth inhibition, some zones were still apparent after 14 to 18 h of incubation.
Zones appeared most evident after 4 to 6 h incubation, with the photograph in Figure 111.]

being taken at 5 h of incubation. As ganglioside concentration increased, the zone’s

46

 

Figure 111.1: Bacteriostatic effects of mixed gangliosides from sweet buttermilk, lactic
acid bacteria, and bovine brain gangliosides on K88+ E. coli using a double agar overlay
technique.a

aWell 1 contains bovine brain gangliosides (19 mM), Well 2 contains sterile water, Well 3 contains
lactic acid bacteria mixed gangliosides(0.32 mM), and Well 4 contains sweet butterrnilkmixed gangliosides

(0.58 mM).

bFive-hour-growth K88+ E. coli bacteria was used.

47

   

1—77 ill"777*7#—~' fiffiffii—g"
1 ' 7 fl 1 ‘ . . ,
A l i l ‘ l i ‘ ‘7 V . H I if 7 '
E l l l 2. ~ l
g 1.5 .. 1 ‘ ‘ l i ~t ~74 wax
u l , 1 l l} l l 7“ ‘fivee , 7 l
g i, i ’ l 7 . I t I i If T77 ' i 71°
I“ 1 l ‘ ‘ l ‘ , . i
z i l . .3» ‘ J ‘ ’_ l
a ‘1 ‘1 i l r W , G T" ""7 i
as -‘ i ‘ l ,. i T
o®‘sr l
. ‘ .. ‘1'
0.186 _ , l
. t
z
MIXED BOVINE BRAIN
GAPBLIOSIE CONCBITRATION
(mM) . ” TlME(hrs)t
19.4 /

 

 

Figure 111.2: Formation of inhibition zones around wells containing varying
concentrations of mixed bovine brain gangliosides. Each well was produced in a lawn of
K88+ Escherichia coli and contained 1 ml of ganglioside extract per well. Ganglioside
quantitation required that preparations and standards be spotted on silica gel G plates
(Merck), placed in a chamber containing chloroform: methanol:0.25% CaCLz solution
(30:18:4); developed; removed, dried, sprayed with resorcinol-HCL for visualization, and
heated at 110°C for 15 min. Zones of inhibition were measured at 2, 4, 6, 8, 12 and 14
hour intervals and values are means with a minimum of three replications at each
concentration.

48

ZONE SIZE [mm]

MIXED CULTURED BUTTERMLK

GANGLDSDE CONCENTRATDN

 

 

 

 

Figure 111.3: Formation of inhibition zones around wells containing varying
concentrations of mixed cultured butter milk gangliosides. Ganglioside quantitation
required that preparations and standards be spotted on silica gel G plates (Merck); placed
in a chamber containing chloroform: methanol:0.25% CaCL2 solution (30: 18:4);
developed; removed; dried; sprayed with resorcinol-HCL for visualization; and heated at
110°C for 15 min. Each well was produced in a lawn of K88+ Escherichia coli and
contained 1 ml of ganglioside extract per well. Zones of inhibition were measured at 2, 4,
6, 8, 12 and 14 hour intervals and values are means with a minimum of three replications
at each concentration.

49

 

ZONE SIZE (mm)

0.5 ',

     
  

./

-:,/ 6
l s
/

MIXED SWEI' BUTTHWIILK
GANGLIOSIDE CONCBJTRATION
(mM)

Figure 111.4: Formation of inhibition zones around wells containing varying
concentrations of mixed sweet buttermilk gangliosides. Ganglioside quantitation required
that preparations and standards be spotted on silica gel G plates (Merck); placed in a
chamber containing chloroform: methanol:0.25% CaCLz solution (30:18:4); developed;
removed; dried; sprayed with resorcinol-HCL for visualization; and heated at 110°C for
15 min. Each well was produced in a lawn of K88+ Escherichia coli and contained 1 ml
of ganglioside extract per well. Zones of inhibition were measured at 2, 4, 6, 8, l2 and 14
hour intervals and values are means with a minimum of three replications at each
concentration.

50

 

TlME(hrs)

 

   

1.5 1.“
LU i ‘7 77
'2' \‘t
0) A 1 4',
I.” E it /
6 v 1. -.
N 0.5 1 /
o - /
/
2
, , .7]
MXED LACTK: ACID BACTERIA ' \ 7 M N
GANGLIOSIDE CONCB‘ITRATION 12 HRS

(mM)

 

Figure 111.5: Formation of inhibition zones around wells containing varying
concentrations of mixed lactic acid bacterial gangliosides. Ganglioside quantitation
required that preparations and standards be spotted on silica gel G plates (Merck); placed
in a chamber containing chloroform: methanol:0.25% CaCLz solution (30: 18:4);
developed; removed; dried; sprayed with resorcinol-HCL for visualization; and heated at
110°C for 15 min. Each well was produced in a lawn of K88+ Escherichia cali and
contained 1 ml of ganglioside extract per well. Zones of inhibition were measured at 2, 4,
6, 8, 12 and 14 hour intervals and values are means with a minimum of three replications
at each concentration.

51

 

 

ZONE 3125 (mm)

 

Figure 111.6: Formation of inhibition zones around wells containing varying
concentrations of mixed yogurt gangliosides. Ganglioside quantitation required that
preparations and standards be spotted on silica gel G plates (Merck); placed in a chamber
containing chloroform: methanol:0.25% CaCLz solution (30: 18:4); developed; removed;
dried; sprayed with resorcinol-HCL for visualization; and heated at 110°C for 15 min.
Each well was produced in a lawn of K88+ Escherichia coli and contained 1 ml of
ganglioside extract per well. Zones of inhibition were measured at 2, 4, 6, 8, 12 and 14
hour intervals and values are means with a minimum of three replications at each
concentration.

52

ZONE SIZE
(mm)

 

BUTTWK GANGLIOSDE
CONCENTRATION
(mM)

 

 

 

Figure 111.7: Formation of inhibition zones around wells containing varying
concentrations of mixed cultured buttermilk gangliosides. Ganglioside quantitation
required that preparations and standards be spotted on silica gel G plates (Merck); placed
in a chamber containing chloroform: methanol:0.25% CaCL2 solution (30: 18:4);
developed; removed; dried; sprayed with resorcinol-HCL for visualization; and heated at
110°C for 15 min. Each well was produced in a lawn of K99+ Escherichia coli and
contained 1 m1 of ganglioside extract per well. Zones of inhibition were measured at 2, 4,
6, 8, 12 and 14 hour intervals and values are means with a minimum of three replications
at each concentration.

53

 

ZONE SIZE (mm)

   

TIME (hrs)

 

MIXED BOVINEBRAIN 1“” '5 ‘ _,
GANGLIOSIDE CONCBII'RATION '
(mM)

 

 

 

Figure 111.8: Formation of inhibition zones around wells containing varying
concentrations of mixed bovine brain gangliosides. Ganglioside quantitation required that
preparations and standards be spotted on silica gel G plates (Merck); placed in a chamber
containing chloroform: methanol:0.25% CaCL2 solution (30: 1 8:4); developed; removed;
dried; sprayed with resorcinol-HCL for visualization; and heated at 110°C for 15 min.
Each well was produced in a lawn of K99+ Escherichia coli and contained 1 ml of
ganglioside extract per well. Zones of inhibition were measured at 2, 4, 6, 8, 12 and 14
hour intervals and values are means with a minimum of three replications at each

concentration.

54

ZONE srzra (mm)
/

 

M IXH) SWEI' BUTTERMILK
GANGLIOSIE CONCBJTRATION

0.58 ' ' / u
, /
(mM) ‘ _

 

 

 

Figure 111.9: Formation of inhibition zones around wells containing varying
concentrations of mixed sweet buttermilk gangliosides. Ganglioside quantitation required
that preparations and standards be spotted on silica gel G plates (Merck); placed in a
chamber containing chloroform: methanol:0.25% CaCL2 solution (30:18:4); developed;
removed; dried; sprayed with resorcinol-HCL for visualization; and heated at 110°C for
15 min. Each well was produced in a lawn of K99+ Escherichia coli and contained 1 m1
of ganglioside extract per well. Zones of inhibition were measured at 2, 4, 6, 8, 12 and 14

hour intervals and values are means with a minimum of three replications at each
concentration.

55

/

_,,.- r
/' t
i

ZONE SIZE (mm)

 

   

Mixar YOGURT
GANGLIOSIDE CONCENTRATION
(mM)

 

 

Figure 111.10: Formation of inhibition zones around wells containing varying
concentrations of mixed yogurt gangliosides. Ganglioside quantitation required that
preparations and standards be spotted on silica gel G plates (Merck); placed in a chamber
containing chloroform: methanol:0.25% CaCL2 solution (30:18:4); developed; removed;
dried; sprayed with resorcinol-HCL for visualization; and heated at 110°C for 15 min.
Each well was produced in a lawn of K99+ Escherichia cali and contained 1 ml of
ganglioside extract per well. Zones of inhibition were measured at 2, 4, 6, 8, 12 and 14

hour intervals and values are means with a minimum of three replications at each
concentration.

56

 

ZONE SIZE
(mm)

 

 

 

 

Figure 111.11: Formation of inhibition zones around wells containing varying
concentrations of lactic acid bacterial gangliosides. Ganglioside quantitation required that
preparations and standards be spotted on silica gel G plates (Merck); placed in a chamber
containing chloroform: methanol:0.25% CaCLz solution (30: 18:4); developed; removed;
dried; sprayed with resorcinol-HCL for visualization; and heated at 110°C for 15 min.
Each well was produced in a lawn of K99+ Escherichia cali and contained 1 ml of
ganglioside extract per well. Zones of inhibition were measured at 2, 4, 6, 8, 12 and 14
hour intervals and values are means with a minimum of three replications at each
concentration.

57

 

  
        

 
  

2 \“7’” ll
1‘ t\
E 1.5 i" l
g \\ I‘.
111 _ .
5 _ .
m ‘ q ,
ru 1 i
z .
o \ ,
0.5 ‘1
\ .......
l
\
o »
ASKS“
\i I at“ ,_ , W3 fiX‘“
. . “v \Z\:E\\\R§\N\
0.125 V ‘ ~-.::<: \ \‘§;\?\‘Q3§.\- . ..\_ ‘3.
MIXEYOGIRT
GANGLIOSIDE
CONCBITRATION 035 \\ .
(mM) “‘x/

 

Figure 111.12: Formation of inhibition zones around wells containing varying
concentrations of mixed yogurt gangliosides. Ganglioside quantitation required that
preparations and standards be spotted on silica gel G plates (Merck); placed in a chamber
containing chloroform: methanol:0.25% CaCL2 solution (30: 1 8:4); developed; removed;
dried; sprayed with resorcinol-HCL for visualization; and heated at 110°C for 15 min.
Each well was produced in a lawn of Escherichia coli 0157:H7 and contained 1 m1 of
ganglioside extract per well. Zones of inhibition were measured at 2, 4, 6, 8, 12 and 14
hour intervals and values are means with a minimum of three replications at each
concentration.

58

 

\
A: t
l
l

l

 

 

 

zone SIZE (m)
-l

   
   

Iv , ; a
. : ._\ .
Q J V ,
l . Q . .. ‘ I ‘ 7‘ ‘~
.. p \g
I ‘I (9 “-
1‘ \ ‘\
1 r x
v i /
a
t
.
U

12

MIXED swear ammrtx “\
GANGLIOSIDE
concemnA'non
(mM)

 

 

 

Figure 111.13: Formation of inhibition zones around wells containing varying
concentrations of mixed sweet buttermilk gangliosides. Ganglioside quantitation required
that preparations and standards be spotted on silica gel G plates (Merck); placed in a
chamber containing chloroform: methanol:0.25% CaCL2 solution (30:18:4); developed;
removed; dried; sprayed with resorcinol-HCL for visualization; and heated at 110°C for
15 min. Each well was produced in a lawn of Escherichia cali 0157:H7 and contained 1
ml of ganglioside extract per well. Zones of inhibition were measured at 2, 4, 6, 8, 12 and
14 hour intervals and values are means with a minimum of three replications at each
concentration.

59

 

appearance remained clearer for a longer time. Bovine brain gangliosides at a
concentration of 19.4 mM produced a clear zone, which remained clear for 6 h, gradually
progressing to growth by 12 h (Figure 111.2). However, bovine brain gangliosides at .46
mM ganglioside concentration produced no apparent zone of inhibition (Figure 111.2).
Concentrations of LAB gangliosides as low as 0.28 mM produced zones that remained
clear for up to 14 h (Figures 111.5 and 111.11). Sweet buttermilk ganglioside
concentrations as minute as 0.8 mM produced some zone effect (cloudy) (Figure 111.13),
with clear zones being achieved at sweet buttermilk ganglioside concentrations of .58
mM (Figure 111.4). Cultured buttermilk at ganglioside concentrations of .037 mM also
produced observable results (cloudy zones), which remained cloudy for up to 4 h (Figures
111.6 and 111.10) persisted for 6 h in some cases (Figure 111.7). Yogurt samples with
ganglioside concentrations as low as .09 mM produced zones that were cloudy in
appearance and The clear zones appeared more frequently and persisted longer on lawns
of K99+ E. cali (Figures 111.7 through 111.11) as compared to K88)r E. coli (Figures 111.2
through 111.6). This was true for all ganglioside sources. When testing similar
ganglioside concentrations, there did not appear to be any discernible difference between
lawns of E. coli 0157:H7 (Figure 111.12 and 111.13) and those of either K99+ E. coli
(Figures 111.7 through 111.11) or K88+ E. coli (Figures 111.2 through 111.6). If plates were
removed from the incubator and placed at 5°C, zones remained for all samples
indefinitely, (plates were refrigerated for one week) because the bacterial growth was

apparently retarded by the low temperature.

60

1.6

1.4
1.2 ‘
l
E I
§ 1i r g
g i Fo—control
3‘ I —I—
'a 0.8 l I LAB '
5 l |—Lr—SBM l
2 l +3993
.3 0.6 1
Q
o
l

 

Time (hours)

Fig. 111.14: Growth inhibition of K99+ Escherichia cali (E. Coli). Ten microliter samples
of mixed bovine brain gangliosides (BBG), mixed lactic acid bacterial gangliosides
(LAB), mixed sweet buttermilk gangliosides (SWB) and sterile water (control) were
extracted from double agar overlay wells which exhibited zones of inhibition, and placed
in 5ml of K99+ E. coli. K99+ Escherichia cali growth occurred for two hours (h) prior to
ganglioside addition. Growth inhibition was measured by optical density readings
conducted at 2h intervals for 16 h at performed at 660 nanometers.

61

In order to determine if a bacteriostatic effect could be detected through growth
turbidity, gangliosides which produced zones of inhibition were extracted and inoculated
into samples of K99+ E. coli. Ten-microliter samples of gangliosides were extracted from
specific wells that produced zones of clearing at 2-, 4-, and 6-h intervals. The results of
ganglioside incubation are illustrated in Figure 111.14. Each value represents an average
of three replications. Bacterial grth was determined by turbidity readings using a
spectr0photometer at 660 nm. Comparing readings at initial incubation to 16 h after
incubation, all samples extracted from the wells after 8h produced bacterial growth,
consistent with 24-h grth samples (Figure 111.14). Mixed gangliosides, from lactic acid
bacteria, sweet buttermilk and bovine brain gangliosides retarded growth for 5h with slow
growth apparent for 6h with lactic acid bacteria and sweet buttermilk (Figure 111.14).
Growth retardation was still evident for LAB at 7h (Figure 111.14). Negative controls

produced bacterial grth when sampled at 2-, 4-, and 6-h intervals (Figure 111.14).

Discussion
The ability of lactobacillus strains to interfere with the growth of intestinal
pathogens has been previously investigated (23, 27). The mode of this inhibition has been
suggested to be attributed to lactic acid production (90), and more recently an effect
termed bacterial coaggregation (22). The research which produced the previously
mentioned results, utilized procedures which were labor intensive and often cost
prohibitive. The ability to screen low concentrations of mixed gangliosides for growth

inhibition of enteric pathogens using non-conventional microbial methods could prove to

62

be a valuable research tool. Developing a reliable procedure for screening low
concentrations of mixed gangliosides for growth inhibition did prove to be a challenge.
Approaches that proved to be unsuccessful included streaking ganglioside solutions over
inoculated plates, soaking filter-paper disks with gangliosides and placing them on
pathogen-inoculated plates, inoculating actively growing cultures with gangliosides, and
incubating pathogens with ganglioside preparations before plating. The double agar
overlay technique produced results that confirmed the applicability of the method for
rapidly screening small quantities of gangliosides and determining their growth-
inhibiting capabilities. Mixed bovine brain gangliosides required higher concentrations
to produce similar zones of clearing (Figure 111.2 and 111.8) when compared to
gangliosides from lactic acid bacteria (Figure 111.5 and 111.11), sweet buttermilk (Figure
111.4 and 111.9), yogurt (Figure 111.6 and 111.10) and cultured buttermilk (Figure 111.3 and
111.7). The lowest concentration of bovine brain gangliosides to prodUce a detectable
zone of inhibition was 8.4 mM (Figure 111.2). In contrast, a concentration of only 0.46
mM sweet buttermilk was effective in inhibiting growth. Sweet buttermilk was more
effective than cultured buttermilk at lower concentrations (Figures 111.3 and 111.4).
Gangliosides extracted from lactic acid bacteria were the most effective, with
concentrations ranging from 0.52 to 0.28 mM producing clear zones that persisted for up
to 8 h (Figure 111.5). The negative controls used were autoclaved samples of gangliosides,
sterile water, and empty wells. No zones of inhibition were observed surrounding any of
the negative control wells, indicating that the gangliosides were solely responsible for the
effect. Recovering sufficient quantities of gangliosides to accurately measure

their effects was a continuing challenge during this research project. In the initial studies,

63

ganglioside source and concentration were compared for inhibition ability and for
quantification with regard to ganglioside concentrations. Dose-responsive studies were
later performed that included dilution factors of 1/2 and 1/4 in an attempt to determine
whether growth inhibitions could be reduced by these amounts. The ability to curtail
bovine brain ganglioside (BBG) inhibition by approximately 1/2 was demonstrated by
reducing BBG content from 19.4 mM to 8.4 mM. Other ganglioside sources, however,
did not demonstrate this effect. In addition, BBG required higher concentrations to
produce similar zones of clearing when compared to gangliosides from lactic acid
bacteria and sweet and cultured buttermilks. The lowest concentration of BBG to
produce a zone of inhibition was 8.4 mM. A concentration of 0.46 mM sweet buttermilk
was effective in producing inhibition. Sweet buttermilk was more effective than cultured
buttermilk at lower concentrations. This effect could be attributed to ganglioside
composition, which has been demonstrated to play an important role with respect to
growth inhibition. Sweet buttermilk may contain specific cell membrane receptors for
K99+ E. coli. Gangliosides extracted from lactic acid bacteria were the most effective
when taken from wells at 2- and 4-h intervals. Sweet buttermilk, cultured buttermilk, and
yogurt mixed gangliosides produced varying results. In general, as ganglioside
concentration was increased by 1/2 or 1/4, the zones of clearing were increased by these
same increments. It was demonstrated, however, that there was a concentration effect for
some gangliosides. In general, the higher the ganglioside concentration of the BBG,
sweet buttermilk, cultured buttermilk, and yogurt sample, the larger, more persistent, and
clearer the zones appeared. BBG, however, was the least effective inhibitor. Since BBG

was likely the purest of the gangliosides because it was commercially prepared by Sigma

64

Chemical; the possibility did exist that other factors were present in the ganglioside
extractions and contributed to the bacteriostatic effects noted.

It is important to reemphasize that ganglioside composition pays an important role
with respect to growth inhibition and that certain gangliosides have been identified as the
specific cell membrane receptors, (14, 85, 87). The differences in inhibition rates can
possibly be attributed to ganglioside sample composition. Additionally, not all
gangliosides have demonstrated growth inhibition capabilities. Sample viscosity could be
an extra factor in determining the size and clarity of zones produced. Some samples,
especially yogurt, were more viscous than others due to their initial density prior to
extraction. The more vicious samples were apparently unable to diffuse into the
pathogen~containing lawn at a rate that could significantly retard bacterial growth. This
may explain why, even at high Concentrations, yogurt gangliosides produced only cloudy
zones. Although zones of clearing are transitory, the fact that bacterial growth was not
observed at intervals as great as 24 h for some mixed gangliosides does suggest the
possibility that the effects may have been bacteriostatic in nature. Although the
determination of zones as clear, cloudy, or opaque did not always allow for the most
accurate description, it did make comparisons between zones more quantifiable.

In general, the growth of bacterial samples in the presence of gangliosides taken
from wells mimicked the ability of these mixed gangliosides to form zones. The larger
and clearer the zones, the more reduced was the grth of bacteria taken from the wells,
which seemed to indicate a bacteriostatic effect. The gangliosides from LAB were
effective in eliciting no bacterial growth in minimal medium at 2- and 4-h intervals.

Whereas grth ultimately did occur in all other samples, it was greatly retarded in many

65

of them. The negative controls produced bacterial growth similar to the normal 24-h
growth of bacteria.

Although the specific mode of action of growth inhibition by gangliosides is
unknown, trisialoganglioside has been shown to inhibit the tyrosine kinase of epidermal
growth factor receptor, inhibiting cell growth and proliferation (171) . In viva the
pathogenicity of entertoxigenic E. coli in infectious diarrhea is due in large part to the
action of enterotoxins (28, 111 ). It is also known that adhesion of bacteria to the host
plays an important role in bacterial virulence (39). Adhesion of bacteria is an early
process in the development of an infection which involves binding of bacterial pili to
intestinal glycolipid receptors (111). Bacteria bind to the intestinal mucosa to begin the
colonization process and elicit disease. Any substances that affect the adhesion of
entertoxigenic E. coli to the intestinal mucosa cells will reduce the pathogenicity of
microorganisms(39, 111). The opportunity to use this newly developed technique in
determining the growth inhibition of small quantities of mixed gangliosides and other

materials is a realistic and achievable goal.

66

CHAPTER IV

INHIBITION OF ADHESION OF K99” Escherichia coli (E. cali)
AND COLONIZATION FACTOR ANTIGENS-1+ (CFA-1+ ) BY
PURIFIED MIXED GANGLIOSIDES EXTRACTED FROM
LACTIC ACID BACTERIA, BOVINE MILK FAT,

AND CULTURED DAIRY FOODS

Abitmszt

A lipid component of milk fat, referred to as gangliosides, which is also present in
cultured dairy food products and lactic acid bacteria (LAB) used to produce these dairy
food products, possesses antimicrobial activity. Mixed gangliosides recovered and
purified from LAB and cultured dairy products were effective in inhibiting the pilus-
mediated adhesion of K99+ Escherichia coli (E. coli) to sheep and horse erythrocytes, and
colonization factor antigen-I (CFA-I+ ) E. cali to chicken and cow erythrocytes.
Ganglioside inhibition of K99+ E. coli and CFA-I+-mediated adhesion was assessed using
a mannose-resistant hemagglutination assay. Bacteria were preincubated with mixed
gangliosides before erythrocyte addition, and nonagglutinated erythrocytes were counted
electronically with an electron particle counter (Coulter Counter). Using horse and sheep
erythrocytes with K99+ E. coli, hemagglutination was inhibited by 19% to 100% in
ganglioside-treated samples (concentrations of 0.018 mM to 0.074 mM). Incubating
CFA-I+ with cow and chicken erythrocytes and mixed gangliosides ranging from 0.026
mM to 0.176 mM produced hemagglutination inhibition ranging from 41% to 81%.

These hemagglutination-inhibition results suggest that mixed gangliosides extracted from

67

cultured dairy foods and LAB could play an important role in limiting the adhesion of
enterotoxigenic E. cali (K99+ and CFA-1+) to the intestinal mucosa, thus preventing or

lessening the effects of diarrhea.

Intmductinrr

Throughout the world, enterotoxigenic Escherichia coli (ETEC) is an important
cause of neonatal diarrhea in pigs, calves, and lambs, as well as in humans (34, 64, 119,
143). Two virulence determinants are necessary for ETEC to cause diarrhea. The first is
enterotoxin production, and the second is pilus production. Pili are hairlike structures
present on the outer cell membrane of E. coli.

Hemagglutination was first recognized as an adhesive property of E. coli in the
early 19003. In 1955, it was demonstrated that the presence of pili on the E. cali surface
was correlated to this adhesive property (33). Antigenically distinct pili present on
separate strains of E. coli are capable of infecting neonatal calves K99, F41 (19, 92), pigs
K99, K88, 987P, F41 (84, 94, 153), lambs K99 (47, 92), and humans (CFA-I and CFA-II)
(37, 38, 39, 40). Pili allow E. coli to attach to specific receptors on the mucosal surface
of the small intestine, which is the initial step in the disease process (56, 64, 119).
Present knowledge about the chemical nature of the receptors indicates that they are
composed of glycolipids and/or glycoproteins. Various glycoproteins and glycolipids
may be able to competitively inhibit the binding of piliated E. coli to brush borders,
epithelial cells, or erythrocytes (99, 116, 120, 183). The N-acetylneuraminic acid portion
appears to play an important role in the inhibiting bacterial adhesion to epithelia thereby

preventing bacterial infection at the stage of bacterial colonization (116, 120, 183).

68

Hemagglutination inhibition of K99+ ETEC has been demonstrated with a
glycoconjugate resembling the monosialoganglioside-Z (GMZ) (93). The carbohydrate
moiety of the sheep erythrocyte receptor also appears to resemble ganglioside GMZ. The
hemagglutination-inhibition assay mimics ETEC pili binding to the intestinal mucosa and

therefore, could serve as a good test system for evaluating gangliosides.

S E 3 I' . I
Mixed ganglioside extractions and purifications were performed in the laboratory
of Dr. Kathleen Moore, Department of Chemistry, Oakland University. The purifications
were conducted on lactic acid bacteria (LAB) obtained from a buttermilk starter culture
(Darigold, Seattle, WA), sweet and cultured powdered buttermilk (Darigold, Seattle,
WA), and plain cultured yogurt (MSU Dairy Plant, East Lansing, MI), by a modification
of the method of Bonafede et al. (14). Specifically, 100 g of dried dairy product
(powdered sweet buttermilk, powdered cultured buttermilk, or lyophilized yogurt) was
suspended in 1,000 ml of chloroform:methanol (1:1 v/v) and allowed to stir overnight.
After removal of solid material by Buchner filtration, the extract was subjected to
rotary evaporation to remove solvents. The dried extract was resuspended in 50 ml of
water and dialyzed against deionized water for 2 days in a cold box. The dialyzed sample
was lyophilized and resuspended in chloroform:methanol:water (C :M:W; 30:60:8 v/v) for
anion-exchange chromatography. The sample was applied to a 1.5-cm x 30- cm column

of DEAD Sephadex A-25 (Sigma Chemical) that was preequilibrated with the C:M:W

69

solvent. The column was then washed with 175 ml of the C :M:W solvent and 75 ml of
methanol. Gangliosides were eluded from the column with 150 ml of 0.4 M ammonium
acetate in methanol. Ganglioside-containing fractions were pooled, and methanol was
removed by rotary evaporation. The dried ganglioside fraction was resuspended in
deionized water and dialyzed against deionized water for 2 days to remove ammonium
acetate.

The dialyzed ganglioside fraction was lyophilized and resuspended in minimal
deionized water for quantitation, TLC analysis, and hemagglutination studies. LAB
gangliosides were extracted in a similar manner except that the LAB cells were ruptured
by glass beads in a chilled Bead Beator and cell membranes were recovered after
centrifugation of homogenate. The membrane pellet was then extracted with a 20-fold
volume of chloroform:methanol (1:1v/v) and processed as described above.

Starter cultures used for the production of yogurt were obtained from Chris
Hansen’s laboratory (Milwaukee, WI). The starter culture (R—39) contained multiple
strains of Streptococcus thermophilus and Lactobacillus bulgaricus. To produce large
quantities of LAB for the isolation of bacterial cell wall gangliosides, 0.2 ml of starter
culture was inoculated per liter of MRS Broth (BBL, Becton Dickinson Microbiology
Systems, Cockeysville, MD) and incubated at 39.5 °C for 72 h. The resultant culture
yielded approximately 7 g of packed wet cells per liter of medium when harvested by
centrifugation at 7,650 rpm for 20 min. Approximately 1,400 g of wet bacterial cells
were initially propagated and harvested for ganglioside isolation using this method. The

starter culture (8803) for buttermilk, which was obtained from Darigold, Seattle, WA,

70

was also cultivated. The bacterial strains used in this starter culture were proprietary;

therefore, specific bacterial identification was not possible.

Thin layer chromatography (TLC) was performed using silica gel 60 TLC
precoated plates (glass plates 10 x 10cm Merck). The plates were activated at 1000 C and
ganglioside preparations and ganglioside and N-acetyl neuraminic acid (NANA)
standards were spotted on silica gel G plates and placed in a chamber containing
chloroform: methanol:0.25% CaCL2 solution (30:18:4). Plates were developed in the
chambers for twice the time it took the solvent to reach the top of the plate. Plates were
then removed, dried, sprayed with resorcinol-HCL for visualization, and heated at 110°C
for 15 min. Sialic-acid—containing material gave blue spots; organic material lacking
sialic acid charred brown. Densitometry of TLC plates was performed with a Shimadzu
Dual-Wavelength TLC-Scanner CS930. The chromatograms were obtained by single
wavelength zig-zag scanning of each lane at 620 nm (beam dimension: 0.4 X 0.4mm) in
the reflection mode (168). The purified preparations yielded blue bands of varying
patterns. All types of preparations had a band that migrated with standard GD3; LAB
preparations had only one additional band, whereas the dairy products had six to eight
separate ganglioside bands. The yields resulting from this process were small, but the

preparations were highly purified and devoid of extraneous contamination.

MW

Erythrocytes were obtained from chickens, cows, horses and sheep located at the

71

Michigan State University, East Lansing, MI farm facilities. Twenty five milliliters of
whole blood were drawn from chickens, cows, horses and sheep weekly. Freshly drawn
blood was immediately aliquoted to prelabeled chilled test tubes containing an equal
volume of Alsever’s solution (Sigma Chemical, St. Louis, MO). Test tubes were inverted
four times for mixing and placed in an ice bath during transportation. Erythrocytes used

for procedures were never older than one week and were stored at 4°C .

S P n n r w n i i ns

Strain 80-5022 K88+ Escherichia cali , K99+-45 Escherichia cali (strain 455133),
K99+ -4 7 Escherichia coli (strain 472304P) was provided by the lab of Dr. Robert
Holland, Department of Large Animal Clinical Science, Michigan State University or
colony forming antigens-1+ Escherichia cali (CFA- 1*) from the lab of Dr. Raul
Hollingsworth, Department of Biochemistry, Michigan State University were utilized.
The negative control for K99+ E. coli hemagglutination assays was selected to be K88+ E.
coli. Sheep and horse erythrocytes have receptors for K99+ E. cali pili but do not possess
receptors for K88+ E. cali pili. All tests were performed with 24-h cultures of bacteria
(108/m1), K99"-45 E. coli, K99+-47 E. cali, OI' CFA-1" E. coli grown in Minca broth with
IsoVitaleX (BBL/Becton Dickinson, Cockeysville, MD). The cultures were incubated

aerobically at 37°C in a shaking water bath prior to use.

m l ' i -
The crude mixed gangliosides recovered from sweet and cultured buttermilk, yogurt, and

lactic acid bacteria (LAB) were tested for their ability to inhibit the pilus-mediated

72

adhesion of CF A-I+ E. cali to cow and chicken erythrocytes, and the K99+ E. coli
adhesion to horse and sheep erythrocytes. Red blood cells mixed with bovine brain
gangliosides (BBG) were used as a positive control. The use of hemagglutination (HA)
in the presence or absence of p-mannose and its analogs provides a basic distinction
between mannose-sensitive (MS) and mannose-resistant (MR) adhesions and is often
used as one form of classification. Enterotoxigenic E. cali possesses MR pili, which
means they are not inhibited by mannose, and it was therefore included as part of the
phosphate buffer.

Pili presence was tested by slide agglutination using CFA-1+, K99+, or K88+
antiserum. One hundred microliters of bacteria (108/ml) were mixed with 200 pm of
mixed gangliosides; both had been washed in 0.14 M of phosphate-buffered saline (pH
6.8) containing 0.1 M of alpha-methylmannoside (PBSM), and incubated at 4°C for 30
min. Then 100 111 of erythrocytes were added and the mixture was incubated for an
additional 60 min at 4°C. Clumping of erythrocytes or hemagglutination was counted in
20 m1 of Isoton (BBL/Becton Dickinson, Cockeysville, MD) solution, with an electronic
particle counter, Coulter Counter, (Coultronics, Margency, France) tube orifice, 100 pl.
All statistical calculations were performed using the Microsoft Excel 2000 statistical
package. The percentage of hemagglutination inhibition was calculated using the

following standardized formula (116):

RBCf in inhibition test - RBCf in 100% MRHA 100
x

 

RBCf 0% MRHA - RBCf in 100 MRHA

where:
RBC = Red blood cells

73

MRHA = Mannose resistant hemagglutination
RBCf in inhibition test = 100 pl RBC, 200 111 ganglioside, 100 pl bacteria
RBCf in 100% MRHA = 100 [.1] RBC, 200 111 buffer, 100 pl bacteria

RBCf in 0% MRHA = 100 )4] RBC, 300 pl buffer

Results

The hemagglutination-inhibition assay is depicted diagrammatically in Figure
IV. 1. A photograph of the hemagglutination-inhibition assay is shown in Figure 1V.2.
Images in this dissertation are presented in color. Purified mixed gangliosides extracted
from LAB and dairy foods (sweet and cultured buttermilk and yogurt) will inhibit the
pilus-mediated hemagglutination of enterotoxigenic E. cali to erythrocytes from different
livestock species. The percentage of hemagglutination inhibition was calculated using a
standardized formula (116). Hemagglutination-inhibition results representing the means
of percentage inhibitions at specific concentrations utilizing K99+-47 E. coli with sheep
and horse erythrocytes are shown in Figures IV.3 through IV. 14. As shown in these
figures, mixed gangliosides from sweet buttermilk, cultured buttermilk, yogurt and lactic
acid bacteria were effective in inhibiting K99+-47 E. coli binding to sheep and horse
erythrocytes at concentrations ranging from 0.0009 mM toO.38 mM . Hemagglutination
was decreased between 0.0% and 100% using these mixed gangliosides. Bovine brain
gangliosides (Figures IV.7 and 13) and monosialoganglioside (Figures IV.8 and 14)
produced inhibitions ranging from 0.0% to 100% with concentration ranging from
0.0005mM tol9.4mM.

As shown in Fig. IV.3 through IV.8, the use of sheep erythrocytes with extracted

74

l    

Red Blood Cell Piliated
bacteria

A B C

 

Figure IV. 1: Diagrammatic scheme of the hemagglutination—inhibition assay used to
determine the beneficial antibinding effect of gangliosides in cultured dairy foods. (3)
Red blood cells in buffer showing no binding effect; (b) hemagglutination of red blood
cells with piliated bacteria, exhibiting binding; (c) Mixed gangliosides binding to piliated
bacteria preventing red blood cells from binding to the piliated bacteria.

75

 

Fig. IV.2: Photograph of hemagglutination assay chambers, showing erythrocyte binding
to enterotoxigenic E. coli. (a) Horse erythrocytes in buffer showing no binding effect; (b)
hemagglutination of K99+ E. coli with horse erythrocytes, exhibiting binding; (c) horse
erythrocytes, K99+ E. cali, with mixed gangliosides from cultured dairy foods, showing
no binding; ((1) horse erythrocytes with K88+ E. cali (negative control), showing no
binding.

76

 

 

 

 

 

 

 

 

 

 

 

110 »— #~ —-——-— ——————————-—-——ewree——w————v_——J—e
100 e A ,4 4* ,___ ~ , M_3§__ L_______WL_.FI,_I
that
904 ._ _. __ ._ _ _ _ *._,..H_.______ _._._ _L__. __ _.*
3 core ewe-e e e e ~‘L , e—I—eee—v—
c E 0153 0221
o
E roPEe- ’9- — e g *—~_———m—— ipiww
.o
E
E 601— _____3 --__ — _— — # ~ ~— i i _ 7 _.—~
g 0.037
g a
g 5° 0025 T L* W V“*‘ I
gae-_ +2 e*__
or 0.316
E It
soe—H Wc,f‘_’,___
g .041 0.068
| 2° T”Tios'r— #0..“ — — m *7 *7 ‘7 — *7 —
10ii0:00fi—---—— —- + +-—-
0633mm, Each ML , I.-- __,__. _.
'0063 .05 .10 .15 .20 .25 .30 .35
Ganglioside Concentrations (mM)

 

 

 

Figure IV.3: Inhibition of sheep erythrocytes with K99+ 47 Escherichia cali by
competitive binding with sweet butter milk gangliosides. Bacteria (108/m1) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
point.

77

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1101 eeeeeeee e-_*2___z___hcflc

1001-7- 77ee_.3*_L_LL_ -LL
0055: 0098 0196
eoi—e ' fl_fiey-__-______z___L_fi
1' 110.083 '

$3 80 —— —~-* —~~- 1' i‘~—v~———J—v

c 3.053 1

.3, 10 e- ~e~uo063.e_—_.¢,___e-#__-flf-z

§ - 0.334

E

26° 0.042 T ”1"“"T*’T'**7

2 0.026

a

g so (its 1 wnfidt—w MFLLIMEE

‘5

a 40 410519— ._‘ L__ i_ *_ Jr”

3 1:00063 _

o 1

I ao-rLue *# g- 0.153L _-_7 _____ __c
20 _.__._ v 7_ _ _ _ _ _ _ _ -_

0.0052
1 101* e * We 7 A L_ ———————

oi! — +._ __._ _ __. _E .__ __ _____-._ ___

.05 .10 .15 .20 .25 .30 .35 .40
1 Ganglioside Concentration (mM)

 

 

Figure IV.4: Inhibition of sheep erythrocytes with K99+ 47 Escherichia coli by
competitive binding with cultured butter milk gangliosides. Bacteria (108/ml) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
point.

78

 

80 —~ —- —-~ - 7* .L -33 —I ;—~—;-—fu——————
0.125

70 _1:_. "___,V i _“3 W _ #. * 7W _— ___
1

 

I 0.045

 

 

 

I .00625 1'0-125
L 0931

501- -3___ e

 

 

 

" i0.0125'0.032
4O — —— - -

Hemagglutination Inhibition (%)

 

30 3 O-Q‘E

0.00125
20 ______,-____-3.. .-.____-.L____L - ——- 4
- 0000910.022

1 _ .05 I 7.10 35%» ".20 U .25
’ Ganglioside Concentration (m M)

 

 

 

 

Figure IV.5: Inhibition of sheep erythrocytes with K99+ 47 Escherichia cali by
competitive binding with yogurt gangliosides. Bacteria (108/ml) were preincubated with
gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

79

 

 

 

 

130333333 3333333333333333 ——
120!— 33333_3_3333333333333
110 T~ 3 3 333 # *——*~ ;_ ;__33_3_333 3 - *—— ———
100*» 3 3 3 3 77777777 33
a“? 90.
V +3.3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
§ .
E80 33 333333333 .333333333DD7433
E
-7033333333313 3333
8
Ewan3333 333333 .333300563330054333
"' 0.037
E sou-ta 3 3 3 ~ 3 33 _333333 33__33_3_333
a
gut
£ I
SDI—##-
20‘3333333333I.33333
0.052
10] *7 "W -3 "WW—I ’ A” ”"0348” "“ i i #H‘
0.026 .
0 t..3 __ .,. _3 333912130032 3 , _-_._r..-_ _ _ ._....-_3 ____,
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Ganglioside Concentration (Mm)

 

 

 

Figure IV.6: Inhibition of sheep erythrocytes with K99+ 47 Escherichia coli by
competitive binding with lactic acid bacteria gangliosides. Bacteria (lOS/ml) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
point.

80

 

Hemagglutination Inhibition ('lo)
8
7

l
l
l
l
l
l
l
l
l
l
l
l

 

 

20 #- ,_ feud, few-17f *A fliJigif v v. A .4, ~—

1 10 ~ a
1 . or9~9°11._, F“:
l
l
l

 

300372 3 4 5 6 7 a 9101112131415161718192021

Ganglioside Concentration (m M)

 

 

Figure IV.7: Inhibition of sheep erythrocytes with K99+ 47 Escherichia coli by
competitive binding with bovine brain ganglioside. Bacteria (lOs/ml) were preincubated
with gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

81

 

 

100L , Lv‘ 777*
0.00145
3: sol e 7 — —— ~ ~ — —
s . 1
E 80 “~- - —- 0.00072 -— — — — _- 7 7 - - -- 7 - * -——-
g 0.003
'5 70 l— ,7. —- ~- -—»— __77 ~77 7, —~ 7 -~ -»-—-——- —- 4
.5 60 .e I 7 7 - - 7 7 7
.5 0.0003{
.5
a ,.7,7, 7 777 7777 777777
.2 0.0005
o:
o:
a _ 7 _ _ 7 7 7 7 7 7 77 7 7
E
o
I 30 T07 7, 7 , ,,,,,, ,7_.. -7. 7.. ___ 7 7___ ,7- 7 ,7. 77
20 L __ AA, 7 7, 7, 7 7. _900194 __ 77 7 7 7 7 __
1o 7 7, .7 . 7 77 7 7- 7 _--__ 7 77777 -7- .7
i o ‘i" "7 l' ” i _771 i ”Tfifi "' "T"" ”—7 i '—— —'l
i 0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035
‘ Ganglioside Concentration (Mm)
7 J

 

Figure IV.8: Inhibition of sheep erythrocytes with K99+ 47 Escherichia coli by
competitive binding with monosialoganglioside. Bacteria (108/ml) were preincubated
with gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

82

 

 

 

 

 

Hemagglutination Inhibition (%)

 

 

 

 

l - L
0 +— 7.77 r_ 7 77 .-7 --7 ___ - . . 77- 7. 7

 

i

i ‘0 TOQ‘Y' # *‘I F"

' .04 .06 .08 .20 0.334 .40 .50
i Ganglioside Concentration (mM)

 

 

Figure IV.9: Inhibition of equine erythrocytes with K99+ 47 Escherichia coli by
competitive binding with sweet butter milk gangliosides. Bacteria (log/ml) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
point.

83

 

 

 

 

 

 

 

 

 

 

 

‘110‘777777777777777 7777777
1 ! ll
. 10°?" F f if F ' ’701-53 01387 7 if f f f f *
eoi ————— 7_7777777777777_77
.3 277777777 7777 7777
‘=’ so! $0.102 I I
.2 ‘ 0.068 0315
S 7017 7 7 7' 7. 1W1“ 7 7 7 7 0.22177 7 7 77
-= 1
EsoiAAA— 7777777777777
g | 1r0037
‘5 50177777 7 7 7 7 A 7_77_7__ 7 7 7 7 7
.E
‘5
a 4077l0943 7 7 A 7__77 77 A 7 77
O
N
§30T—— __7_7 7 A A 7 77777 7 A 7 _77
I _ 10.034
20 _ .7..-_7_ 7_ 7 77 7 7 7 7 7 7 7 7
I I0.021
10.77-77777 A A 77 _ A A A 7
1 0.041
1.0.0017
ol._77...__...7 77.— -j . 1b. , - ,.———.-.—,
.10 .15 .20 0-194 .30 .35 .40
. Ganglioside Concentration (mM)
{ .
Figure IV. 10: Inhibition of equine erythrocytes with K99+ 47 Escherichia coli by

competitive binding with cultured butter milk gangliosides. Bacteria (lOs/ml) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each

point.

84

 

 

 

 

‘ 0000;“
807TA7— — _77777777777777777
11L ? 0.69 #0128
0‘64- 7025.77 7 7 7 ,7 7 7 7 7 7 7 7.._

 

Hemagglutination Inhibition (%)

 

 

 

l
l
l .05 .10 i .13" “72'0"— 7' .2'5 .30
i Ganglioside Concentration (mM)

 

 

Figure IV.11: Inhibition of equine erythrocytes with K99+ 47 Escherichia cali by
competitive binding with yogurt gangliosides. Bacteria (108/m1) were preincubated with
gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

85

 

 

 

 

 

 

 

 

l 100.“ 4 4 47777” 77777777777
.1-----_777

I! .—

| 0704
780AAAAAA—AAAAAAA . 7777717
.3 0.074
C
0 .7777777777 7777..._77‘7777
9370' 8 0.064
a h
£60.777777777777.70.05& 77777 7
g l
EsoiAAAA7A:AA0.037777777777
§ 1
g: 40- —— —— —— —— —— ___ _— ___ ___ __ —_ __ ___ —_ __ ___ __ _7
(0
§ -
$3077777777_7777777777777

.-
I

~ 20 A 7_ A -7- A-- 0086”“ - .7 —AA— A 77- A
1 1°i““e*“”“ ** ***0‘*r’“ 058§C7_"‘__ “—“"r*“‘“*'"“ “"“’""
l i 0.04 0.06

0177 7 7777

 

 

 

.01 _T02 i .08 .04 .05 .08 T6? .08
Ganglioside Concentration (mM)

 

 

Figure IV.12: Inhibition of equine erythrocytes with K99+ 47 Escherichia coli by
competitive binding with lactic acid bacteria gangliosides. Bacteria (108/m1) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
point.

86

 

 

 

 

 

i
‘100A -77777777777777--7-
790AAA A A AAA AA.-AA AAA
°\°
=80777777777-777777777_
.9
3.:
9701777777777777777777
'5:
'5 60A 7 A 7 A 7 7 A AA 7 A-19;4—A AA
.9
.5507 77777777777777
.5
33.407 -
3
~m-4AAA 7777777777
E
w
:20- AAA
10‘» 777 7 7777
0.9

 

 

Gangioude Concertrafion (n'M)

 

 

Figure IV.13: Inhibition of equine erythrocytes with K99+ 47 Escherichia coli by
competitive binding with bovine brain gangliosides. Bacteria (108/ml) were preincubated
with gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

87

 

 

 

 

 

 

 

Ganglioside Concentration (mM)

1 10017 A A A 7-7 7- 7 A A A 77 7777-7777 -7 7
90. A 77777777777-7---7_
80 .7 7 7- 7- 7.. 7.7 77 777_77__7 7 7 7 7- 7_7
3:":
g 70A 7 A A A A A 7 7-- 777777 7 7
E
E 60177777 7777-7_7_ A A 7 A 7 7_—
.5
E 50% 77777777777 7 777
3
§40A A A AA _7__7__7 -7 -7 7 -77
E
O l
_I 304 A _ 7 7 77 77
.1 7 7 7 7 7 7
I
‘ l 0.98
10‘ A A _77 — 7
3 [0 I0097
1 hW‘FOJ 0.4 0'.6 SWEA- IIT fl
1
l

 

 

Figure IV. 14: Inhibition of equine erythrocytes with K99+ 47 Escherichia coli by
competitive binding with monosialoganglioside-3. Bacteria (lOS/ml) were preincubated
with gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

88

mixed gangliosides produced the following ranges of hemagglutination—inhibition
percentages, standard deviations and standard error of the mean. Sweet buttermilk
ganglioside concentrations (Fig. IV.3) in the range of 0.0017mM to 0.316mM produced
inhibitions ranging from 2% to 98.5%; standard deviations (Sd) ranging from 2.08
standard deviation units (SdU) to 33.9SdU and standard error of the mean (SE) ranging
from 1.49SEm units (SEmU) to 23.99SEmU. Additionally, when concentrations of
0.0017mM and 0.316mM sweet buttermilk gangliosides were compared, the percent
inhibitions were determined to be significant at a confidence level of 0.01 utilizing the t-
Test. Cultured buttermilk ganglioside concentrations (Fig. IV.4) in the range of 0.0052
mM to 0.334 mM produced inhibitions ranging from 19.6% to 100%; standard
deviations(Sd) ranging from 1.73 SdU to 49.49SdU and standard error of the mean (SE)
ranging from 0.99 SemU to 34.99 SemU. When concentrations of 0.0052mM and 0.33mM
cultured buttermilk gangliosides were compared, the percent inhibitions were determined
to be significant at a confidence level of 0.05 utilizing the t-Test. Yogurt ganglioside
concentrations (Fig. IV.5) in the range of 0.0009 mM to 0.202 mM produced inhibitions
ranging from 0% to 96%; Sd ranging from 2.12SdU to 50.2SdU and SEm ranging from
1.49SEmU to 35.35SemU. Additionally, when concentrations of 0.0009mM and 0.02mM
gangliosides were compared, the percent inhibitions were determined to be significant at
a confidence level of 0.05 utilizing the t—Test. Lactic acid bacteria ganglioside
concentrations (Fig. IV.6) in the range of 0.024 mM to 0.074mM produced inhibitions
ranging from 0% to 88%; Sd ranging from 0.7OSdU to 53.74SdU and SEm ranging from
0.49SemU to 37.47SemU. When concentrations of 0.024mM and 0.074mM gangliosides

were compared, the percent inhibitions were determined to be significant at a confidence

89

level of 0.01 utilizing the t-Test. Bovine brain gangliosides concentrations (Fig. IV.7) in
the range of 0.0097mM tol9.4mM produced inhibitions ranging from 0% to 63.5%; Sd
ranging from 0.0SdU to 26.87SdU and SEm ranging from 0.49SemU to 18.99SemU.
Additionally, when concentrations of 0.0097mM and 19.4mM gangliosides were
compared, the percent inhibitions were determined to be significant at a confidence level
of 0.01 utilizing the t-Test. It is important to note that bovine brain ganglioside
concentrations required to produce effective inhibitions was increased by a factor of 10,
as compared to the mixed gangliosides extracted from sweet and cultured butter milk and
lactic acid bacteria. This observation held true for the different gangliosides used, the
various pathogens inhibited, and the types of erythrocytes tested. An additional assay,
using monosialoganglioside-3 (GM3) (Sigma Chemical, St Louis, MO), which has been
reported to be the K99+ receptor site (84), was also conducted (Fig. IV.8).
Monosialoganglioside-3 (GM3) concentrations ranging from 0.0005mM to 0.003 mM
produced inhibitions ranging from 27% to 100%; Sd ranging from 0.7OSdU to 18.3 8SdU
and SEm ranging from 0.508EmU to 12.99SemU. Further, when monosialoganglioside-3
(GM3) (Fig. IV.8) and bovine brain gangliosides (BBG) (Fig. IV .7) were compared at the
concentration of 0.001mM utilizing sheep erythrocytes, monosialoganglioside-3 (GM3)
demonstrated a greater ability to inhibit agglutination. These findings were consistent
using either horse or sheep erythrocytes.

Depicted in Figures IV.9 through IV.13 are the results of equine erythrocytes
binding with K99+ E. coli and incubated with extracted mixed gangliosides, producing

the following ranges of hemagglutination-inhibition percentages, standard deviations and

90

standard errors of the mean. Sweet buttermilk ganglioside concentrations (Fig. IV.9) in
the range of 0.0052 mM to 0.38 mM produced inhibitions ranging from 0% to 100%; Sd
ranging from 0.7OSdU to 36.76SdU and SEm ranging from 0.49SemU to 25.99SEmU.
Additionally, when concentrations of 0.0052mM and 0.380mM gangliosides were
compared, the differences between percent inhibitions were determined to be significant
at a confidence level of 0.05 utilizing the t-Test. Cultured butter milk ganglioside
concentrations (Fig. IV.10) in the range of 0.0017 mM to 0.316 mM produced inhibitions
ranging from 4% to 100%; Sd ranging from 0.08dU to 32.52SdU and SElm ranging from
0.0SEmU to 22.99SemU. When concentrations of 0.0017mM and 0.316mM gangliosides
were compared, the differences between percent inhibitions were determined to be
significant at a confidence level of 0.01 utilizing the t-Test. Yogurt ganglioside
concentrations (Fig. IV.11) in the range of 0.0055mM to 0.252mM produced inhibitions
ranging from 4% to 100%; Sd ranging from 2.12SdU to 36.59SdU and SEm ranging from
1.49SEmU to 25.87SemU. Additionally, when concentrations of 0.005 SmM and 0.202mM
yogurt gangliosides were compared, the percent inhibition of the larger concentration was
determined to be significantly greater at a confidence level of 0.01 utilizing the t-Test.
Lactic acid bacteria ganglioside concentrations (Fig. IV. 12) in the range of 0.04 mM to
0.074 mM produced inhibitions ranging from 0% to 86%; Sd ranging from 6.36SdU to
40.3OSdU and SEm ranging from 4.49SEmU to 28.49SemU. Additionally, when
concentrations of 0.026mM and 0.074mM lactic acid bacteria gangliosides were
compared, the differences between percent inhibitions were determined to be significant
at a confidence level of 0.01 utilizing the t-Test. Bovine brain ganglioside concentrations

(Fig. IV.13) in the range of 0.04mM to 19.4mM produced inhibitions ranging from 0% to

91

60%; Sd ranging from 2.82SdU to 29.14SdU and SEm ranging from 1.99SEmU to
20.60SemU. Additionally, when concentrations of 0.04mM and 19.4mM gangliosides
were compared, the percent inhibition of the larger concentration was determined not to
be significantly greater at a confidence level of 0.05 utilizing the t-Test.The
hemagglutination assay using GM3 (Sigma Chemical, St. Louis, MO) (Fig. IV. 14) in the
range of 0.48 mM to 0.98 mM produced inhibitions ranging from 2% to 17.5%; Sd
ranging from 2.8ZSdU to 24.74SdU and SEm ranging from 1.99SEmU to 17.49SemU.
Additionally, when concentrations of 0.48mM and 0.98mM GM3 (Sigma Chemical, St.
Louis, M0) were compared, the percent inhibition of the larger concentration was
determined not to be significantly greater at a confidence level of 0.05 utilizing the t-Test.
Testing the percent inhibitions of 0. 10mM bovine brain ganglioside (Fig. IV.13) with
0.98mM monosialoganglioside-3 (Fig. 4.14) it was determined that they were not
significantly different at a confidence level of 0.05 utilizing the t-Test. The
hemagglutination inhibition percentage of lactic acid bacteria ganglioside at a
concentration of 0.074mM was evaluated using sheep erythrocytes and equine
erythrocytes and it was determined that they were not significantly different at a
confidence level of 0.05 utilizing the t-Test.

Figures IV .15 through IV. 1 8 demonstrate the inability of K88+ E. coli to inhibit
hemagglutination with equine erythrocytes due to the absence of a K88+ E. coli receptor
on the surface of equine erythrocytes. Fig. IV.15 illustrates the use of sweet buttermilk
gangliosides, Fig. IV.16 depicts the use of cultured buttermilk gangliosides, Fig. IV. 17
illustrates the use of yogurt gangliosides, and the use of bovine brain ganglioside is

illustrated in Fig. IV. 1 8. The mean percent inhibition for K88+ E. coli with sweet

92

 

 

N
O

 

s a
l
l
l
l
l

Hemagglutination Inhibition (%)
8

 

 

30 L. L L L L LL L L LLL L __ _L_ L _ L
20 l — f f f f f f i f —
10 f _- ptisf ‘——‘ “ 0 38
0.098 029 o 334
0 i ‘ — ' ‘—Y——— "r' J ' "—' ' ’ ‘r————'" I *.—‘ —Y' r r I r —'I —1
0 0.05 1 0.15 0.2 0.25 I03 0.35 0.4

.10 L- .L. . _ .- L . . - L L . LL
.20 i_ LLL- LL ____ ______ __ - LL LL _-

 

Ganglioside Concentration (mM)

 

 

Figure IV.15: Inhibition of equine erythrocytes with K88+ Escherichia coli by competitive
binding with sweet butter milk gangliosides. Bacteria ( 108/ml) were preincubated with
gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

93

 

 

 

; 130.‘—»J»—»»*»»»»»»—~»f»»1»*»»»
120il »» fffffff »J——-—»~‘»»»»»»——
11olr»mrvg——w»»»»w——

A100l—»—» — ~»*+»»J—**——

é l
=00~——~»»—v»#»» ffffff »»i»»f*—

.2

“

Baa—f” »‘+*f+frivf~f» ——

§

— 70L L LLL L L LL L L L LL L L L L
C

.9.

saw—Jivvi*;I——»fii‘iAflf~—

.5

H

aso-v—LAfi»fi#—~—————————fl———a
a:
3’40LLL LLLLLLL_LLLLL
E l
o

130‘4AfiivAAwrva»—
20l——e ———»»»»»»»»»—
ml L L L L .1... L L L L
0.158 ' 0221 I0.316

1 01» .—.—» L L.IL—_i.——' — L L» LL

l 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

L Ganglioside Concentration (mM)

 

Figure IV.16: Inhibition of equine erythrocytes with K88+ Escherichia coli by competitive
binding with cultured butter milk gangliosides. Bacteria (log/ml) were preincubated with
gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

94

 

Hemagglutination inhibition (%)
8
l
l
l
l
l
l
l
l
l
l
l

 

 

5 l, h h h, JP ___ _-. ._ -______
1 0.126 0302 Io.252 [0.34

0.05 0.1 E 0.15 0.2 0.25 0.3 5.35 0.4

-5 ~ —— 0.125

 

 

 

 

 

Ganglioside Concentration (mM)

 

 

Figure IV.17: Inhibition of equine erythrocytes with K88+ Escherichia coli by competitive
binding with yogurt gangliosides. Bacteria (108/ml) were preincubated with gangliosides
for 30 min. Erythrocytes were added and incubated an additional hour. Percent inhibitions
were calculated using a standard equation based on Coulter Counter measurements. All
values are means with a minimum of two replications at each concentration. Standard
deviation is represented by the ascending lines at each point.

95

60-

457

35 7 7 77
30 7 7 7 7

25- 7 7 77 7 77 7 777 7 7 7 7 7 7
20 1 7 7 7 7 7 7

15T 77 77 7 7 77 77 777 77 7 777 7 777

Hemagglutination inhibition (%)

1o } 7 7 7 7 7 77 7

1.3' 7 '
{0.1 _5__ ,_ 10 -15 ,, 20 25

.10. 7 77 7 , -4- M--- ,4 h
Ganglioside Concentration (mM)

 

 

Figure IV.18: Inhibition of equine erythrocytes with K88” Escherichia coli by competitive
binding with bovine brain gangliosides. Bacteria (lOg/ml) were preincubated with
gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

96

buttermilk ganglioside, cultured buttermilk gangliosides yogurt gangliosides and bovine
brain ganglioside was -1%. As demonstrated in Fig. IV.19 through IV.26, mixed
gangliosides from sweet buttermilk, cultured butter milk, and lactic acid bacteria were
effective in inhibiting the pilus-mediated agglutination of colonization factor antigen-I
(CFA-I) to chicken and bovine erythrocytes at concentrations ranging from 0.016mM to
0.176 mM. Hemagglutination was inhibited from 32% to 98.5% using these mixed
gangliosides. Depicted in Figures IV. 19 through IV.22 are the results of CFA+-I and
chicken erythrocytes which produced the following ranges of hemagglutination-inhibition
percentages, standard deviations and standard error of the mean. Sweet buttermilk
ganglioside concentrations (Fig. IV. 19) in the range of 0.066 mM to 0.136mM produced
inhibitions ranging from 44% to 98%; Sd ranging from 3.53 SdU to 14.46SdU and SEm
ranging from 2.49SEmU to 8.3SSemU. Additionally, when concentrations of 0.066mM and
0.136mM sweet buttermilk gangliosides were compared the differences between percent
inhibitions were determined to be significant at a confidence level of 0.01 utilizing the t-
Test. Cultured buttermilk ganglioside concentrations (Fig. IV.20) in the range of
0.079mM to 0.18mM produced inhibitions ranging from 46% to 49%; Sd ranging from
4.24SdU to 6.36SdU and Sem ranging from 2.94SEmU to 4.49SEmU. Comparing
concentrations of 0.079mM to 0.18mM cultured buttermilk gangliosides, the differences
between percent inhibitions were determined to be significant at a confidence level of
0.01 utilizing the t-Test. Mixed yogurt ganglioside concentrations (Fig. IV.21) in the
range of 0.016mM to 0.176 mM produced inhibitions ranging from 49% to 81%; Sd

ranging from 0.7OSdU to 5.74SdU and SEm ranging from 0.0SEmU to 2.87SemU.

97

 

90777777777777 7777 771,77

80 _ zzzz_ z_,,-,ofi3%z,
0.103

7077———————————————————————

 

Hemagglutination Inhibition (%)

0 ,,,_ _, ”T ___ ___fi ___ _-_,_____r_.__._ —7-- _. , .
.02 .04 .06 .08 .10 .12 .14 .16

l Ganglioside Concentration (m M)

l

 

 

Figure IV. 19: Inhibition of chicken erythrocytes with colonization factor antigen-I (CFA-
I) by competitive binding with sweet butter milk gangliosides. CFA-I (lOs/ml) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
point.

98

l

l 90.
l

1 80
I .

7077777

1
l
l
u
l
l
l
l
l 7 .
i
60: 777777777777777 --.__sz.zr -_L , -_fl
l
l
l
I
l
l
i
1

50?

 

40.

Hemagglutination Inhibition (%)

303—

.085 "1105"” T12? “ARM-.135” 11735“ ’C 5205”

l

l

'.

l 20

l

l

l

; Ganglioside Concentration (m M)

 

 

Figure IV.20: Inhibition of chicken erythrocytes with colonization factor antigen-1(CFA-
I) by competitive binding with cultured butter milk gangliosides. CF A-I (108/ml) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
pomt.

99

 

 

 

 

 

 

90 777777777777777777
30_,v;_-#__7*I_A
g? i 0.176
5 70 z A z z “.7 __ I + A .0130 E E f
i 0120 0.126
gml7777777777777
é |0016
£50777—[7—7777 77777
a 0.065
U: «ml 7 7 7 77 7 7 7 777 7 7 7 7
a
E
9
I 30._-_-H -. __.. ._ #
20 77 7777 7 77 7 7 7 77 7 777
10 77 77 77 77 777 7 777 77 7
l o .7 7 7777 . .7777 77 7 7
1 .05 .10 .15 .20

‘ Ganglioside Concentration (m M)

 

 

Figure IV.21: Inhibition of chicken erythrocytes with colonization factor antigen-I (CFA-
I) by competitive binding with yogurt gangliosides. CF A-I (108/ml) were preincubated
with gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

100

 

Hemagglutination Inhibition (%)

.09 .15

—_* —> Y_" “-1.— _" ___'V " ‘ #‘V'v’

. 20 . 25
Ganglioside Concentration (mM)

 

Figure IV.22: Inhibition of chicken erythrocytes with colonization factor antigen-I (CF A-
I) by competitive binding with lactic acid bacteria gangliosides. CFA-I (108/m1) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent'inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
point.

101

 

110 7 7 77 77777 7 7 7 7 7777777 7 7 777-77-- .7 ,7 #7”

1m,7 7 777 77777 7 7 777 7 7 A z,z-,z f,z __

W77~77 7*7 7 7 *7 7 777 7 7 77 777777 7 7 777

70777777777 7 777777 7777777777777777J7777777

 

 

 

 

Hemagglutination inhibition (%

 

 

 

 

30.7 77 7 7 7 77 7 7 i7 7 A ,
0.066

20 l 7 7 7 7 7 7 7 7 7 7 7 7 7
10 7 7 7 7 7 7 — 7 7 7 7 7

| 0 + 7 77.777 .7 7 77 7777 -77 7, 7.7.77 ,

.02 .04 .06 .08 .10 .12 .14
Ganglioside Concentration (Mn)
\ J

 

Figure IV.23: Inhibition of bovine erythrocytes with colonization factor antigen-I (CFA-I)
by competitive binding with sweet butter milk gangliosides. CFA-I (108/m1) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
point.

102

 

 

i
1004 77777777777 7 77777777777 7 77 7 7 77
907:7777
Agol___777______ 7_______7_7
ES I
:5: 701L7777777—7—
3:
9
£604"7 7777777777777777
g ‘E
Ewlt ,, ,zhhhfzonaoz 7-771-
8
73 40 j ,_..-__D.OLQ d 1, A- 1 1,-_______-___..,1 .1, 1, ,7 1. 1m_
II
E l
0:
130(7—
20I._.._____7-._-._ __ ________.___ ,_____ _fi__
10777 77 7 77 7 7 77 7777777 77 7 777
0 l '7 ’7 i'l ”7 ’i’ T #7, 'i'l C“ ' " 7 " *‘” m" 'V' ‘ ‘i i I 7" i _ fi
.10 .12 .14 .16 .18 .20
Ganglioside Concentration (Mm)
7 J

 

Figure IV.24: Inhibition of bovine erythrocytes with colonization factor antigen-I (CF A-I)
by competitive binding with cultured butter milk gangliosides. CFA-I (108/ml) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
point.

103

 

 

Hemagglutination Inhibition (%

 

 

.02 C C .0604 C” iio‘i "‘ .12? 31‘s“
Ganglioside Concentration (rrM)

 

Figure IV.25: Inhibition of bovine erythrocytes with colonization factor antigen-I (CF A-I)
by competitive binding with yogurt gangliosides. CFA-I (108/ml) were preincubated with
gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.

104

 

Hemagglutination Inhibition (%

’ .020” ’ :0‘30 “.035 Wm was

Ganglioside Concentration (mM)

 

 

Figure IV.26: Inhibition of bovine erythrocytes with colonization factor antigen-I (CFA-I)
by competitive binding with lactic acid bacteria gangliosides. CFA-I (lOs/ml) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each

point.

105

 

Additionally, when concentrations of 0.016mM and 0.176mM yogurt gangliosides were
compared, the differences between inhibition capabilities were determined to be
significant at a confidence level of 0.01 utilizing the t-Test. Lactic acid bacteria mixed
ganglioside concentrations (Fig. IV.22) in the range of 0.04 mM to 0.026mM produced
inhibitions ranging from 34.3% to 42.7%; Sd ranging from 3.46SdU to 16.07SdU and
SEm ranging from 1.99SEmU to 9.27SemU. When concentrations of 0.026mM and
0.04mM lactic acid bacteria gangliosides were compared, the differences between percent
inhibitions were determined to be significant at a confidence level of 0.05 utilizing the t-
Test.

Depicted in Figures IV.23 through IV.26 are the results of CF A-I+ and bovine
erythrocytes which produced the following ranges of hemagglutination-inhibition
percentages, standard deviations and standard error of the mean. Sweet buttermilk mixed
ganglioside concentrations (Fig. IV.23) in the range of 0.066 mM to 0.136mM produced
inhibitions ranging from 32% to 89%; Sd ranging from 0.7OSdU to 20.46SdU and SEm
ranging from 0.49SEmU to 1.99SemU. Comparing concentrations of 0.066mM and
0.136mM sweet butter milk gangliosides, the differences between percent inhibitions
were determined to be significant at a confidence level of 0.01 utilizing the t-Test.
Cultured buttermilk mixed ganglioside concentrations (Fig. IV.24) in the range of
0.079mM to 0.180mM produced inhibitions ranging from 46% to 56%; Sd ranging from
1.41SdU to 2.82SdU and SEm ranging from 0.99SEmU to 1.99SemU. Additionally, when
concentrations of 0.079mM and 0.180mM mixed cultured butter milk gangliosides were
compared, the differences between percent inhibitions were determined to be significant

at a confidence level of 0.01 utilizing the t-Test. Mixed yogurt ganglioside concentrations

106

(Fig. IV.25) in the range of 0.061 mM to 0.176mM produced inhibitions ranging from
50% to 80.5%; Sd ranging from 1.9ZSdU to 25.45SdU and SEm ranging from 1.35SEmU
to 25.99SemU . When concentrations of 0.061mM and 0.176mM mixed yogurt
gangliosides were compared, the differences between percent inhibitions were determined
to be significant at a confidence level of 0.05 utilizing the t-Test. Mixed lactic acid
bacteria ganglioside concentrations (Fig. IV.26) in the range of 0.026 mM to 0.052mM
produced inhibitions ranging from 34% to 43%; Sd ranging from 4.24SdU to 19.345SdU
and SEm ranging from 2.728EmU to 11.31SemU. Additionally, when concentrations of
0.026mM and 0.052mM mixed yogurt gangliosides were compared, the percent inhibition
of the larger concentration was determined not to be significantly greater at a confidence
level of 0.05 utilizing the t-Test.

Hemagglutination assays were initially performed using two different strains of
K99+ E. coli. one designated K99+-45 E. coli and the other K99+-47 E. coli. Limited
assays using K99+-45 E. coli were conducted. When it was determined that the abilities
of both strains to inhibit agglutination were comparable, the decision was made to
concentrate exclusively on the K99+-47 E. coli strain, conserving our limited gangliosides
resources for other comparisons. All the results reported are based on assays using K99+-

47 E. coli.

Discussion
Enteric diseases, including enteropathogenic E. coli (EPEC), enterohemorrhagic
E. coli (EHEC), and enterotoxigenic E. coli (ETEC), all cause diarrhea, with ETEC being

the most common cause of diarrhea in young farm animals (119, 143). Attachment of

107

enterotoxigenic bacteria to the susceptible host tissue is the initial step in the disease
process (119, 159). Two important factors that provide ETEC with means of colonization
are the expression of pili, which allow for the attachment of bacteria to the intestinal wall,
and the production of enterotoxins that influence intestinal secretions (119). The
hemagglutination-inhibition assay has the ability to mimic ETEC pili binding to the
intestinal mucosa and therefore is regarded as a good in vitro test system. Currently, the
treatment for ETEC centers on rehydration and antibiotic treatment of infected species
(12, 14, 27). The extensive use of antibiotics has lead to the development of antibiotic
resistant strains of E. coli (27, 45). Interest in providing alternatives to the standard
treatments for ETEC is an important basis for this research. Carbohydrate moieties on the
cell surface of sheep, horse, cow, chicken, and human erythrocytes resemble various
gangliosides. These chemical structures have some characteristics similar to the receptors
used by pathogenic agents to adhere to the intestinal epithelium. The assay used is based
on research previously done by Mouricout et al. (116).

The results generated by this research demonstrate that purified mixed
gangliosides from sweet and cultured buttermilk, yogurt, and LAB will inhibit the binding
of the livestock pathogen K99+ E. coli and the human pathogen E. coli CFA-I+ to varying
degrees. The mechanism by which binding of the bacterial cell to the red blood cell
occurs has been reported to be similar, if not identical, to the binding of the bacterial cell
to the intestinal mucosa (116, 124, 158). Therefore, it is reasonable to infer that the
adhesion of the enterotoxigenic E. coli to the intestinal mucosa, which is a prerequisite
for colonization, could be inhibited by the consumption of dairy products that contain

gangliosides.

108

Fujiwara et al. (44) demonstrated that enterobacteria have a high affinity for a
specific neutral glycolipid and that this binding can be inhibited through the use of
ganglioside-containing bifidiobacteria strains. Other researchers have indicated that
gangliosides extracted from LAB also have this bacterial-interference effect (139). The
mixed gangliosides contained in cultured dairy foods are contributed by both the bacteria
and the milk fat.

Determining why a 10-fold increase in ganglioside amount is needed when using
BBG to inhibit agglutination, as compared to mixed gangliosides from LAB and dairy
products, has several possible explanations. It is conceivable that the preparation of
mixed gangliosides contains proteins that may interfere with bacterial binding. It is also
reasonable that, among the mixed gangliosides, one or several possess a higher affinity
for the tested bacterial strains as compared to the affinity of BBG. When comparisons
have been made between the inhibition of hemagglutination by gangliosides, the specific
composition of the ganglioside is extremely important. Enterotoxins of Vibrio cholerae
will bind to monosialoganglioside-l with a higher affinity than other tested gangliosides
(63). Monosialoganglioside-3 has been reported to be the K99+ receptor by Kyogashima
et a1 (84). The findings of this research demonstrated that mixed ganglioside preparations
from sweet and cultured buttermilk, lactic acid bacteria and plain cultured yogurt were
more effective as inhibitors of K99+ E. coli hemagglutination than GM]. However, it is
important to note that when monosialoganglioside-3 (GM3) (Fig. IV.8) and bovine brain
gangliosides (Fig. IV.7) were compared at a concentration of 0.001mM utilizing sheep

erythrocytes, the data indicate that GM3 does exhibit the ability to inhibit agglutination

109

more severely than BBG. These findings were consistent using either horse or sheep
erythrocytes.

There appears to be no perceptible differences between the inhibiting capacities of
sheep erythrocytes and horse erythrocytes with K99+-47 E. coli. Sweet buttermilk,
cultured buttermilk or yogurt also did not exhibit any major differences as far as their
ability to inhibit hemagglutination. It is important to note that the data suggest
differences between the inhibitory abilities of commercial gangliosides, (GM3 and bovine
brain gangliosides) and those of gangliosides extracted from dairy products. The ability
of mixed gangliosides isolated from dairy products to inhibit agglutination at lower
concentrations, when compared to commercially prepared gangliosides, was
demonstrated with both horse and sheep erythrocytes.

As discussed earlier, the possibility of the mixed gangliosides from cultured dairy
food origins containing gangliosides with a greater affinity for the K99+ pili is a
reasonable explanation for this phenomenon. Hemagglutination-inhibition assays using
CFA-I+ were more consistent at specific concentrations and range variations extremely
small as compared to tests with K99+ E. coli. CFA-I+ hemagglutination assays also
produced similar inhibitions at constant concentrations when comparing cow percentage
inhibitions to chicken percentage inhibitions. BBG gangliosides were required in
amounts more than 10 times that of extracted mixed gangliosides, which proved to be the
case in agglutination tests with K99+ E. coli. CFA-I+ assays utilized only purified
gangliosides, which may account for the increase in repeatability.

E. coli K88: the negative control, produced very low or no hemagglutination

inhibition (see Fig. IV.15 through IV.18). The negative percentage inhibitions would

110

indicate a loss of erythrocytes. The data do not suggest that this was the case. In
actuality, the sampling of nonagglutinated erythrocytes was not an exact measurement,
but rather an estimate. Precautions were taken: Each assay was run in duplicate, each
chamber was sampled twice with an average used as the erythrocyte count, and only one
person performed the Coulter Counter sampling. The initial erythrocyte counts may have
been somewhat low and, when used in the equation, produced negative numbers for the
percentage inhibition of K88+ E. coli. It is reasonable to conclude that K88+ did not
impede the agglutination of erythrocytes from horse and sheep, and the addition of
gangliosides caused no disruption in this process. More tests were performed with K88+
E. coli hemagglutination; the percentage inhibitions included in Fig. IV.15 through IV. 18
represent an average of three tests at each concentration and also coincide with K99+ E.
coli concentrations, allowing for more accurate comparisons of the data.

These results underscore and support previous reseach (81, 100) as to the possible
benefits of regular consumption of cultured dairy foods in preventing enteric diseases. It
was calculated that a person would need to consume three to four 8-ounce containers of
yogurt per day to obtain concentrations similar to those shown to be effective in this

study.

111

CHAPTER V

INHIBITION OF ADHESION OF HELICOBA C T ER PYLORI BY MIXED
GANGLIOSIDES IN LACTIC ACID BACTERIA, BOVINE
MILK FAT, AND CULTURED DAIRY FOODS

Abstract

Gangliosides are a diverse group of naturally occurring, complex sugar-containing
glycosphingolipids, which are found in the milk fat of all mammalian species as part of
the milk fat globular membrane (2, 14, 51, 79). Gangliosides are also localized on the
outer plasma membrane, and studies have shown that cell-surface gangliosides serve as
recognition sites for the attachment of viruses, bacteria, and bacterial toxins (126).
Recently, the involvement of Helicobacter pylori (H. pylori) in gastric diseases has been
researched (9, 10, 53, 101, 102). Studies have indicated that H. pylori binds to
gangliosides, and that the adhesion of the bacteria to glycosphingolipid receptors is an
initial step in the pathogenesis and propagation of disease. Mixed gangliosides recovered
and purified from cultured dairy products and lactic acid bacteria (LAB) were found to be
effective in inhibiting the pilus-mediated adhesion of H. pylori to chicken erythrocytes.
Ganglioside inhibition was assessed using a mannose-resistant hemagglutination assay.
H. pylori was preincubated with gangliosides before their exposure to erythrocytes, and
nonagglutinated erythrocytes were counted electronically with a Coulter Counter.

Ganglioside concentrations in the cultured dairy products and LAB ranging from 0 .126

112

mM to 30.0 mM resulted in a decrease in hemagglutination ranging between 0% and 79%
when H. pylori was agglutinated with chicken erythrocytes. On the basis of these results,
mixed gangliosides may possibily limit the adhesion of H. pylori to the gastric epithelial

cells, thereby decreasing the organism’s contribution to gastric disease.

Introduction

Peptic-ulcer disease is a chronic inflammatory condition affecting the stomach
and duodenum; it is believed to cause complications to at least 10% of the US.
population at some time during their lives (9, 10, 101, 175). The pathogenesis of this
disorder was first believed to be related to stress and diet. Digestive secretions, such as
gastric acids, were later targeted as the cause. The role of the bacterium identified as
Helicobacter pylori (H. pylori) in gastric diseases recently has been under investigation
(9, 10, 101).

This bacterium is found almost exclusively in the stomach and duodenum, where
it is closely associated with the gastric epithelial cells and/or the mucous layer (9).
Evidence that H. pylori is involved in the etiology of duodenal and gastric inflammation
and ulcers, and that it is also a pathogen causing human gastritis, is compelling (9, 10,
101, 175).

The specific adherence of H. pylori to gastric epithelial cells may be critical to
maintenance of infection and subsequent pathogenicity. When sialic acid receptors were
removed from erythrocytes, the inhibition of the hemagglutination factor of H. pylori was
achieved (95). This same research paper concluded that the receptor specificity of the H.

pylori hemagglutin is in the same class as the K99+ fimbrial adhesion of enterotoxigenic

113

E. coli (95). Other studies have indicated that H. pylori binds to gangliosides and that the
adhesion of the bacteria to glycosphingolipid receptors is an initial step in the
pathogenesis and propagation of disease (135, 142, 143). These findings suggested that
H. pylori might be a good candidate to test for inhibition with extracted mixed
gangliosides from cultured dairy foods and lactic acid bacteria (LAB), as assessed by the

hemagglutination assay.

W

The purified mixed gangliosides recovered from cultured buttermilk, yogurt, lactic
acid bacteria (LAB) and evaporated milk (Carnation) were tested for their ability to
inhibit the pilus-mediated adhesion of H. pylori to chicken erythrocytes. Red blood cells
mixed with bovine brain gangliosides (Sigma Chemical, St. Louis, MO) (BBG), were
used as a positive control.
Scumfflangljgsidfi

Mixed ganglioside extractions and purifications were performed in the laboratory
of Dr. Kathleen Moore, Department of Chemistry, Oakland University. The purifications
were conducted on lactic acid bacteria (LAB) obtained from a buttermilk starter culture
(Darigold, Seattle, WA), cultured powdered buttermilk (Darigold, Seattle, WA), and plain
cultured yogurt (MSU Dairy Plant, East Lansing, MI), by a modification of the method of
Bonafede et a1. (14). Specifically, 100 g of dried dairy product (powdered sweet
buttermilk, powdered cultured buttermilk, or lyophilized yogurt) was suspended in 1,000
mi of chloroform:methanol (1 : 1) and allowed to stir overnight. After removal of solid

material by Buchner filtration, the extract was subjected to rotary evaporation to remove

114

solvents. The dried extract was resuspended in 50 m1 of water and dialyzed against
deionized water for 2 days in a cold box. The dialyzed sample was lyophilized and
resuspended in chloroform:methanol:water (C:M:W; 30:60:8) for anion—exchange
chromatography. The sample was applied to a 1.5-cm x 30- cm column of DEAE
Sephadex A-25 (Sigma Chemical) that was preequilibrated with the C:M:W solvent. The
column was then washed with 175 ml of the C:M:W solvent and 75 m1 of methanol.
Gangliosides were eluded from the column with 150 ml of 0.4 M ammonium acetate in
methanol. Ganglioside-containing fractions were pooled, and methanol was removed by
rotary evaporation. The dried ganglioside fraction was resuspended in deionized water
and dialyzed against deionized water for 2 days to remove ammonium acetate. The
dialyzed ganglioside fraction was lyophilized and resuspended in minimal deionized
water for quantitation, TLC analysis, and hemagglutination studies.

Starter cultures used for the production of yogurt were obtained from Chris
Hansen’s laboratory (Milwaukee, WI). The starter culture (R-39) contained multiple
strains of Streptococcus thermophilus and Lactobacillus bulgaricus. To produce large
quantities of LAB for the isolation of bacterial cell wall gangliosides, 0.2 m1 of starter
culture was inoculated per liter of MRS Broth (BBL, Becton Dickinson Microbiology
Systems, Cockeysville, MD) and incubated at 395°C for 72 h. The resultant culture
yielded approximately 7 g of packed wet cells per liter of medium when harvested by
centrifugation at 7,650 for 20 min. Approximately 1,400 g of wet bacterial cells were
initially propagated and harvested for ganglioside isolation using this method. The starter

culture (8803) for buttermilk, which was obtained from Darigold, Seattle, WA, was also

115

grown up. The bacterial strains used in this starter culture were proprietary; therefore,
specific bacterial identification was not possible.

Thin layer chromatography (TLC) was performed using silica gel 60 TLC
precoated plates (glass plates 10 x 100m Merck). The plates were activated at 100° C and
ganglioside preparations and ganglioside and N-acetylneuraminic acid (NANA) standards
were spotted on silica gel G plates and placed in a chamber containing chloroform:
methanol:0.25% CaCLz solution (30:18:4). Plates were developed in the chambers for
twice the time it took the solvent to reach the top of the plate. Plates were then removed,
dried, sprayed with resorcinol-HCL for visualization, and heated at 110°C for 15 min.
Sialic-acid-containing material gave blue spots; organic material lacking sialic acid
charred brown. Densitometry of TLC plates was performed with a Shimadzu Dual-
Wavelength TLC-Scanner CS930. The chromatograms were obtained by single
wavelength zig-zag scanning of each lane at 620 nm (beam dimension: 0.4 X 0.4mm) in
the reflection mode (168). The purified preparations yielded blue bands of varying
patterns. All types of preparations had a band that migrated with standard GD3; LAB
preparations had only one additional band, whereas the dairy products had six to eight
separate ganglioside bands. The yields resulting from this process were small, but the
preparations were highly purified and devoid of extraneous contamination.
W

Erythrocytes were obtained from chickens located at the Michigan State
University, (East Lansing, MI) poultry farm facility. Twenty five milliliters of whole

blood was drawn from chickens weekly. Freshly drawn blood was immediately aliquoted

116

to prelabeled chilled test tubes containing an equal volume of Alsever’s Solution (Sigma
Chemical, St. Louis, MO). Test tubes were inverted four times for mixing and placed in
an ice bath during transportation. Erythrocytes used for procedures were never older than
one week and remained stored at 4°C .

Smeafflaflrogflmdfimmhficndifim

Helicobacter pylori (strain ATCC 43504, NCTC 11637) cultures were provided
by the lab of Dr. Raul Hollingsworth in the Department of Biochemistry, Michigan State
University (MSU). H. pylori was grown in Minca broth with IsoVitaleX (BBL/Becton
Dickinson, Cockeysville, MD). The cultures were incubated aerobically at 37°C in a
shaking water bath prior to use.

The bacterial cell wall gangliosides were isolated and purified at Oakland
University in the laboratory of Dr. Kathleen Moore. With respect to the isolation of
ganglioside from LAB, the typical preparation involves motor-driven homogenization of
60 g of bacterial cells (wet weight) with 100 ml of sucrose phosphate buffer, pH 7.4. The
mixture was then transferred into a chilled BeadBeator chamber (Fisher Scientific,
Atlanta, GA). The bacteria cells were ruptured by glass beads for three 1-min intervals;
each l-rnin process was followed by a lO-min cool-down. The process was done in a
cold box. After treatment was complete, the beads were allowed to settle before
decanting of the homogenate. The beads were washed twice with 50 ml of buffer; the
washes were then combined with the homogenate.

The homogenized bacterial mix was centrifuged for 20 min at 25,000 revolutions

per minute (RPM). This procedure pelleted the bacterial membranes, which contained

117

most of the gangliosides. The supernatant was lyophilized and then extracted for its
minimal ganglioside content. The bacterial-membrane pellets were weighed and
gradually homogenized with chloroform:methanol (1 :1 v/v, 20 ml of solvent per gram of
pellet). Afier overnight extraction, the process continued as previously described,
according to the method of Bonafede et al. (14). Eventually, lyophilized samples were

extracted by Dr. Moore’s laboratory using the method described earlier.

- ' i i
The crude mixed gangliosides recovered from sweet and cultured buttermilk, yogurt,
lactic acid bacteria (LAB) and evaporated milk were tested for their ability to inhibit the
pilus-mediated adhesion of H. pylori to chicken erythrocytes. Chicken erythrocytes
mixed with bovine brain gangliosides (BBG) were used as a positive control. The use of
hemagglutination (HA) in the presence or absence of p-mannose and its analogs provides
a basic distinction between mannose-sensitive (MS) and mannose-resistant (MR)
adhesions and is often used as one form of classification. Enterotoxigenic E. coli
possesses MR pili, which means they are not inhibited by mannose, and it was therefore
included as part of the phosphate buffer.

One hundred microliters of H. pylori (108/ml) were mixed with 200 pm of mixed
gangliosides; both gangliosides and H. pylori were previously washed in 0.14 M of
phosphate-buffered saline (pH 6.8) containing 0.1 M of alpha-methylmannoside (PBSM),
and incubated at 4°C for 30 min. Then 100 pl of chicken erythrocytes were added and
the mixture was incubated for an additional 60 min at 4°C. Clumping of erythrocytes or

hemagglutination was counted in 20 ml of Isoton (Coultronics, Margency, France) with

118

an electronic particle counter (Coulter Counter, Coultronics, Margency, France, tube
orifice 100 pl). All statistical calculations were performed using the Microsoft Excel
2000 statistical package. The percentage of hemagglutination inhibition was calculated by

the following standardized formula (116):

RBCf in inhibition test - RBCf in 100% MRHA x 100

 

RBCf 0% MRHA 7 RBCf in 100 MRHA

where:

RBC = Red blood cells

MRHA = Mannose resistant hemagglutination

RBCf in inhibition test = 100 pl RBC, 200 pl ganglioside, 100 pl bacteria
RBCf in 100% MRHA = 100 pl RBC, 200 pl buffer, 100 pl bacteria

RBCf in 0% MRHA = 100 pl RBC, 300 pl buffer

limits
The hemagglutination-inhibition assay has the ability to mimic H. pylori pili binding to
the gastric epithelial cells and therefore serves as a good in vitro test system. The
hemagglutination-inhibition assay is depicted diagrammatically in Figure IV.1 and Figure
IV.2 is a photograph of the hemagglutination-inhibition assay. The percentage of
hemagglutination inhibition was calculated using a standardized formula (116). Purified
mixed gangliosides extracted from LAB and cultured buttermilk, yogurt and evaporated
milk will inhibit the pilus-mediated hemagglutination of H. pylori to chicken

erythrocytes.

119

Hemagglutination-inhibition results representing the means of percentage
inhibitions at specific concentrations utilizing H. pylori with chicken erythrocytes are
shown in Figures V.1 through V.4. As shown in these figures, mixed gangliosides from
cultured buttermilk, yogurt, lactic acid bacteria and evaporated milk were effective in
inhibiting H. pylori binding to chicken erythrocytes at concentrations ranging from 0.061
mM to 0.79 mM. Hemagglutination was decreased between 1% and 79% using these
mixed gangliosides. Each mean inhibition represents a minimum of two replications at
this concentration.

As shown in Fig. V.l the use of chicken erythrocytes with extracted mixed
gangliosides from lactic acid bacteria produced the following hemagglutination
inhibition percentages, standard deviations and standard errors of the mean. Lactic acid
bacteria ganglioside concentration of 0.079mM produced 0% hemagglutination inhibition
(HI); a standard deviation (Sd) of 0.0 and a standard error of the mean (Sem) of 0.0.
Lactic acid bacteria ganglioside concentration of 0. 103mM produced 46% HI; 11.31 Sd
and 7.99 Sem. Lactic acid bacteria ganglioside concentration of 0.132mM produced 48%
HI; 16.97 Sd and 11.99 Sem. Lactic acid bacteria ganglioside concentration of 0.136mM
produced 54% HI; 9.89 Sd and 6.99 Sem. Additionally, when concentrations of 0.079mM
and 0.136mM lactic acid bacteria gangliosides were compared, the percent inhibition of
the larger concentration was determined to be significantly greater at a confidence level of
0.01 utilizing the t-Test.

As shown in Fig. V.2 the use of chicken erythrocytes with extracted mixed
gangliosides from cultured buttermilk produced the following hemagglutination

inhibition percentages, standard deviations and standard errors of the mean. Cultured

120

 

 

so 77 77 7 7 777 7 7 7 7 77— 7 77 7 7 .777. .77 77 77 7

 

3o 7 7- 777 7.77 7 77 7_7 7 777—77.— 77777 7. ___—1

 

 

Hemagglutination Inhibition (%)

20 7 v 7 777 777 7 7— 7—7 7 --—~ -7——

 

‘0 F!" 7 ”-7177 ' 'v T079 # _ " ”7"—— _——‘ “C

o l;- ' I I __ ""———Y“' - __——“ _"' ' r I '
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16

Lactic Acid Bacteria Ganglioside Concentration (mM)

 

 

 

 

 

Figure V. 1: Inhibition of chicken erythrocytes with Helicobacter pylori by competitive
binding with lactic acid bacteria gangliosides. H. pylori (108/ml) were preincubated with
gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.
Ganglioside quantitation required that preparations and standards be spotted on silica gel
G plates (Merck), placed in a chamber containing chloroform: methanol:0.25% CaCL2
solution (30: 18:4); developed; removed, dried, sprayed with resorcinol—HCL for
visualization.

121

 

Hemagglutination Inhibition (%)
00
O

 

 

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
Cultured Buttermilk Gangliosides Concentration (mM)

 

 

Figure V.2: Inhibition of chicken erythrocytes with Helicobacter pylori by competitive
binding with cultured buttermilk gangliosides. H. pylori (108/ml) were preincubated with
gangliosides for 30 min. Erythrocytes were added and incubated an additional hour.
Percent inhibitions were calculated using a standard equation based on Coulter Counter
measurements. All values are means with a minimum of two replications at each
concentration. Standard deviation is represented by the ascending lines at each point.
Ganglioside quantitation required that preparations and standards be spotted on silica gel
G plates (Merck), placed in a chamber containing chloroform: methanol:0.25% CaCL2
solution (30:18:4); developed; removed, dried, sprayed with resorcinol-HCL for
visualization,

122

 

 

 

Yogurt Ganglioside Concentration (mM)

0
N
.5 ..
O)
on
_a.
O

Hemagglutination Inhibition (%)
8
l
l
l
l
l
l
I
l
l
l
l
l
l
l
l
1
l

 

Figure V.3: Inhibition of chicken erythrocytes with Helicobacter pylori by competitive
binding with yogurt gangliosides. H. pylori (108/ml) were preincubated with gangliosid

68

for 30 min. Erythrocytes were added and incubated an additional hour. Percent inhibitions

were calculated using a standard equation based on Coulter Counter measurements. All
values are means with a minimum of two replications at each concentration. Standard
deviation is represented by the ascending lines at each point. Ganglioside quantitation

required that preparations and standards be spotted on silica gel G plates (Merck), placed

in a chamber containing chloroform: methanol:0.25% CaCLz solution (30:18:4);
developed; removed, dried, sprayed with resorcinol-HCL for visualization.

123

 

 

 

( 100.77777777777777777777 777777777777777
I ..I
|
I
80] * ***’*’****‘ i 0.372 05,;4’“
? .
2:on ............ W 7“,---
s I I
g I 0.192
2 60 .7 7 7 7777 777 777 777777777777777 7 777
_‘E : 0:86
c .
o 50 7777777777777777 7777—77777—
: I
“ I
.E
L5 40 7 7 7 7 7 7 7 7 7 7 7777777777 77777—
O)
O)
N
e 30 7- 7 7 7777777777777777777
0
::
20177—777777 777777 7777777777
107 I 77777777777
I
I 0.096
0 F4" #_ ' _T ‘ ‘"‘ "‘.’"_" ;"‘ ’_T“_ _ " ”T—m "”— —‘T_ __ ‘_T" F “" F" '_._‘I"_“ —‘—‘1
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45

Evaporated Milk Ganglioside Concentration (mM)

 

 

 

Figure V.4: Inhibition of chicken erythrocytes with Helicobacter pylori by competitive
binding with mixed gangliosides from evaporated milk. H. pylori (108/ml) were
preincubated with gangliosides for 30 min. Erythrocytes were added and incubated an
additional hour. Percent inhibitions were calculated using a standard equation based on
Coulter Counter measurements. All values are means with a minimum of two replications
at each concentration. Standard deviation is represented by the ascending lines at each
point. Ganglioside quantitation required that preparations and standards be spotted on
silica gel G plates (Merck), placed in a chamber containing chloroform: methanol:0.25%
CaCLz solution (30: 18:4); developed; removed, dried, sprayed with resorcinol-HCL for
visualization.

124

buttermilk ganglioside concentration of 0.079mM produced 3% HI; 2.82 Sd and 1.99
Sem. Cultured buttermilk ganglioside concentration of 0.094mM produced 17% HI; 12.72
Sd and 8.99 Sem. Cultured buttermilk ganglioside concentration of O. 1 80mM produced
49% HI; 5.65 Sd and 3.99 Sem. Additionally, when concentrations of 0.079mM and
0.18mM cultured buttermilk gangliosides were compared, the percent inhibition of the
larger concentration was determined to be significantly greater at a confidence level of
0.01 utilizing the t-Test.

As shown in Fig. v.3 the use of chicken erythrocytes with extracted mixed
gangliosides from yogurt produced the following hemagglutination inhibition
percentages, standard deviations and standard errors of the mean. The yogurt ganglioside
concentration of 0.061mM produced 1% HI; 0.0 Sd and 0.0 Sem. The yogurt ganglioside
concentration of 0.065mM produced 10% HI; 8.48 Sd and 5.99 Sem.The yogurt
ganglioside concentration of 0.126mM produced 58% HI; 4.24 Sd and 2.99 Sem. The
yogurt ganglioside concentration of O. 130mM produced 54% HI; 5.65 Sd and 3.99 Sem.
Additionally, when concentrations of 0.065mM and 0.130mM yogurt gangliosides were
compared, the percent inhibition of the larger concentration was determined to be
significantly greater at a confidence level of 0.01 utilizing the t-Test.

As shown in Fig. V.4 the use of chicken erythrocytes with extracted mixed
gangliosides from evaporated milk produced the following hemagglutination inhibition
percentages, standard deviations and standard errors of the mean. The evaporated milk
ganglioside concentration of 0.096mM produced 7% HI; 5.65 Sd and 3.99 Sem. The
evaporated milk ganglioside concentration of 0.186mM produced 57% HI; 1.41 Sd and

0.99 Sem. The evaporated milk ganglioside concentration of 0.192mM produced 63% H1;

125

7.07 Sd and 4.99 Sem. The evaporated milk ganglioside concentration of 0.372mM
produced 79% HI; 14.14 Sd and 9.99 Sem. The evaporated milk ganglioside concentration
of 0.384mM produced 77% HI; 0.0 Sd and 0.0 Sem. Additionally, when concentrations of
0.096mM and 0.384mM evaporated milk gangliosides were compared, the percent
inhibition of the larger concentration was determined to be significantly greater at a

confidence level of 0.01 utilizing the t-Test.

Discussion

The correlation of H. pylori infections with peptic ulcer disease has been reported
(9, 10, 53, 101, 131, 135, 175). Marshall and others (9, 10, 101, 102, 131) demonstrated
that H. pylori is responsible for at least 70% of ulcers. Stress, and other causes, such as
prescription drug use may play a role, but these are not usually the direct cause of ulcers
(9, 10, 131). Additionally, it is important to note that the presence of H. pylori does not
indicate peptic ulcer formation will occur, but it is unlikely to detect the conditions of
peptic ulcer disease or gastritis without this bacterium’s presence. Previously, ulcers were
treated with a combination of diet alterations and stress reductions. These regimes usually
provided only limited relief. With the indication of H. pylori as an important causative
agent in the etiology of gastric ulcers, antibiotic therapy is currently the course of choice.
These antibiotics have proven to be extremely effective in the treatment and cure of
gastric ulcers.

Adhesion of bacteria in general and H. pylori specifically, to host epithelial
surfaces is usually the first step in the colonization process (119, 123, 131, 146, 175).

This binding can also cause to induce the expression of virulence factors (146). The

126

binding of H. pylori to erythrocytes has been reported to be via sialoprotein containing
the N-acetylneuraminyl-lactose (2-3) linkage in the same class as the K99 fimbrial
adhesion of enterotoxigenic Escherichia coli (ETEC) (3, 95, 108, 157). H. pylori isolates
have also been shown to bind strongly with rabbit, guinea pig, chicken, horse, cow, sheep,
and monkey erythrocytes (95, 157). These research findings were the basis for the
decision to test whether H. pylori would bind with the mixed gangliosides from cultured
dairy foods in a hemagglutination-inhibition assay.

The specific course of H. pylori-induced illnesses are complex. Motility appears
to be directly related to the ability to penetrate the epithelial mucus layer. Attachment of
the bacterium appears to prevent or severely impair mucus production, allowing this area
to be exposed to digestive gastric enzymes. The indication that adherence of H. pylori to
gastric epithelial cells may be critical to maintenance of infection and ultimate
pathogenicity also lends support to the approach of limiting H. pylori adherence. It is
relevant to indicate that H. pylori was demonstrated to have a high binding affinity for the
form of N-acetylneuraminyl-lactose which predominates in bovine colostrum (108). The
ability of extracted mixed gangliosides from dairy foods to inhibit the adhesion of
chicken erythrocytes to H. pylori pili indicates that gangliosides extracted from cultured
buttermilk, yogurt, LAB, and evaporated milk may provide a useful therapy for
combating gastric diseases through the reduction of H. pylori gastric epithelial cell
adherence.

The substantial evidence which suggests that LAB have an inhibitory effect on
pathogenic bacteria in the intestinal tract (18, 32, 35, 45, 49, 50, 51, 59, 62, 63, 90, 103,

109, 117, 185, 189) is generally attributed to the theory of competitive exclusion. This

127

theory maintains that LAB control the growth and colonization of opportunistic enteric
pathogens by effectively competing for nutrients, lowering the luminal pH through the
production of lactic acid and the possible production of antibiotic-like substances. H.
pylori possesses characteristic enzyme activity with urease being of special interest.
Urease activity facilitates the production of ammonia from gastric juice urea. The
ammonia functions as a base and raises the pH of the gastric environment, which allows
for the survival of H. pylori in the acidic environment of the stomach. As described
earlier, LAB control the growth and colonization of opportunistic enteric pathogens by
lowering the luminal pH through the production of lactic acid. In addition to
preventing/reducing epithelial binding, LAB can also adversely affect the gastric pH,
providing a less stable environment for H. pylori.

Although these studies were limited in scope—only two agglutination tests at each
concentration were performed—the goal of these experiments was to assess whether
cultured dairy foods and LAB could inhibit the hemagglutination of RBC by H. pylori. In
this respect, the studies were successful. Statistical comparisons were made but their
relevance may be over or under stated due to the of the lack of repetitive experiments.
There appeared to be no difference between the inhibition capabilities of lactic acid
bacteria, yogurt or evaporated milk with respect to similar concentrations. There were
significant differences when the smallest concentrations of specific gangliosides were
compared to the highest concentrations of these same gangliosides. The results therefore
indicate, that as mixed ganglioside concentrations increase, the resultant percentage

inhibition also increases.

128

Present knowledge about the chemical nature of the receptors indicates that they
are composed of glycolipids and/or glycoproteins. Various glycoproteins and glycolipids
may be able to inhibit the binding of piliated E. coli to brush borders, epithelial cells, or
erythrocytes (116, 120, 183). The N-acetylneuraminic acid portion appears to play an
important role in the inhibition of binding. The hemagglutination—inhibition assay
mimics ETEC pili binding to the intestinal mucosa and therefore serves as a good test
system.

The results of this research demonstrate that mixed gangliosides from cultured
buttermilk, yogurt, LAB, and evaporated milk will inhibit the binding of the human
pathogen H. pylori to chicken erythrocytes. Therefore, it is reasonable to infer that the
adhesion of H. pylori to gastric epithelial cells could be inhibited by the consumption of
dairy products that contain gangliosides. Additionally, it has been reported that H. pylori
is rarely seen in children. Although, the literature does not indicate an explanation for this
phenomenon, it is important to note that children are proliferate consumers of dairy
products. These results underscore the benefits of regular consumption of cultured dairy
foods in preventing enteric diseases. It was calculated that a person would need to
consume three to four 8-ounce containers of yogurt per day to obtain concentrations

similar to those shown to be effective in this study.

129

CHAPTER VI

SUMMARY

The basic premise of this research is that mixed gangliosides extracted from sweet
buttermilk, cultured buttermilk, yogurt and lactic acid bacteria possess antimicrobial
activity and enterotoxin inhibiting capabilities that will inhibit the pilus-mediated
adhesion of bacteria and that these gangliosides also possess bacteriostatic and
bacteriocidal capabilities. There is substantial evidence that lactic acid bacteria (LAB)
have an inhibitory effect on pathogenic bacteria in the intestinal tract (49, 50, 62, 103,
109, 178). Several theories have been proposed for this bacterial antagonism; the most
widely accepted theory is that of competitive exclusion. This theory holds that LAB
control the growth and colonization of opportunistic enteric pathogens by effectively
competing for nutrients, lowering the luminal pH by producing lactic acid, and possibly
producing antibiotic-like substances. When an imbalance occurs in this arrangement,
pathogens gain an advantage, ultimately resulting in a disease state. The antimicrobial
activity beneficially modifies the microflora and microbial activity of the intestinal tract
serving as a prophylaxis against enteric diseases.

Gangliosides are a diverse group of naturally occurring, sugar-containing
glycosphingolipids having sialic acid (N-acetylneuramic acid) as an essential component

and a hydrophobic lipid component (ceramide) (88). Studies have shown that cell-surface

130

gangliosides serve as recognition sites for the attachment of viruses, bacteria, and
bacterial toxins (71, 75, 76, 87, 88, 99). This attachment is attributed to the sialic acids
and oligosialyl structures in glycoproteins, glycolipids, and the complex polysaccharides
associated with the cell walls of microorganisms.

Gangliosides of different types naturally occur in the milk fat of all mammalian
species as part of the milk fat globular membrane, which is derived from the plasma
membrane of epithelial cells in the mammary gland (2, 14, 51, 79, 170). The
gangliosides in milk will inhibit the enterotoxins of V. cholera (63) and E. coli (12), with
human milk gangliosides showing a considerably higher (5-10x) inhibitory effect than
bovine or formula milk (45, 51, 57). The difference in activity has been attributed to the
ganglioside composition in human and bovine milk (85). It is also important to note that
changes in ganglioside composition occur in human and bovine milk throughout the
lactation cycle, which may affect the inhibitory ability of these gangliosides (170).

Throughout the world, ETEC is an important cause of neonatal diarrhea in pigs,
calves, and lambs, as well as in humans (64, 119, 143). E. coli and salmonella are the
most common causes of diarrhea in children in developing nations (15, 28, 119). These
illnesses are still a major cause of infant and childhood morbidity and mortality. Two
virulence determinants are necessary for ETEC to cause diarrhea: enterotoxin production
and pilus production. Pili allow E. coli to attach to specific receptors on the mucosal
surface of the small intestine (31, 32, 38, 42, 158).

The working hypothesis of this project was that mixed gangliosides extracted

from dairy products will inhibit the pilus-mediated adhesion of bacteria and that these

131

gangliosides also possess bacteriostatic and bacteriocidal capabilities. The specific
research objectives are as follows:

1. To extract gangliosides from sweet buttermilk, cultured buttermilk, yogurt and
lactic acid bacteria.

2. To test extracted mixed ganglioside fractions from sweet buttermilk, cultured
buttermilk, yogurt and lactic acid bacteria for enterotoxin inhibiting
capabilities.

3. To screen extracted mixed ganglioside fractions from sweet buttermilk,
cultured buttermilk, yogurt and lactic acid bacteria for antimicrobial activity
using human and animal enterotoxigenic E. Coli strains.

One of the most important impediments to ganglioside research is the inability to
extract large amounts of crude mixed gangliosides for analysis as well as the ultimate
purity level (lack of contaminants) of these gangliosides. In order to achieve the first
objective, the development of precise procedures, based on ganglioside extraction from
tissue (168) and bovine buttermilk (l4), and adapted to sweet and cultured buttermilk,
yogurt, evaporated milk and LAB from starter cultures, had to be accomplished. Mixed
gangliosides were extracted from cultured dairy foods (cultured buttermilk, yogurt), sweet
buttermilk, LAB, and evaporated milk using a modification of the method of Bonafede et
al. (14). The extraction process, which was labor intensive and had not previously been
developed specifically for cultured dairy foods (CDF), bovine milk fat (BMF), or LAB,
required the application of existing protocols to the unique and specialized challenges of
ganglioside extractions from these products. In addition to the extraction process, mixed

gangliosides were purified with a DEAE- Sephadex A-25 (Sigma Chemical) column,

132

producing samples with fewer contaminants. Although LAB ganglioside extraction
initially was undertaken at MSU, the major task of ganglioside extraction from dairy
products (cultured buttermilk, sweet buttermilk, yogurt, evaporated milk) and LAB was
undertaken at Dr. Kathleen Moore’s laboratory in the Department of Chemistry at
Oakland University. Thin layer chromatography (TLC) was performed using silica gel 60
TLC precoated plates (glass plates 10 x 10cm Merck), in order to quantify gangliosides.
The plates were activated; ganglioside preparations and ganglioside and N-
acetylneuraminic acid standards were spotted on silica gel G plates which were then
developed. Plates were removed, dried, sprayed with resorcinol-HCL for visualization,
and heated at 110°C for 15 min. Sialic-acid-containing material gave blue spots; organic
material lacking sialic acid charred brown. Densitometry of TLC plates was performed
with a Shimadzu Dual-Wavelength TLC-Scanner CS930. The chromatograms were
obtained by single wavelength zig-zag scanning of each lane at 620 nm (beam dimension:
0.4 X 0.4mm) in the reflection mode (168). The purified preparations yielded blue bands
of varying patterns. The qualitative evaluation of several protocols by examining TLC
plates provided information as to ganglioside composition of the samples (which were
predominantly GD3), and the ability of these protocols to provide gangliosides with the
least contamination. The yields resulting from this process were small, but the
preparations were highly purified and devoid of extraneous contamination. The success of
these extractions has proven and will continue to be an important part of future analysis
in the area of milk-product ganglioside research. The use of purified preparations also
provided more consistent results with the hemagglutination assays and the double agar

overlay technique.

133

The extraction of LAB gangliosides indicated that substantially more gangliosides
are present in these bacteria than in dairy products (sweet or cultured buttermilk, or
yogurt). Of the dairy products tested, yogurt had the highest ganglioside content. The
calculation of ganglioside content on a nmol/g dry-weight basis, using purified
preparations, produced the following results: cultured buttermilk 7.3 i 5.0 nmol/ g, sweet
buttermilk 11.6 :t 8.1 nmol/g, yogurt 13.2 :1: 5.6 nmol/g, yogurt starter culture 26.7 i 8.5
nmol/g, and buttermilk starter culture 65.0 :I: 13.7 nmol/g. The use of cultured plain
yogurt whose commercial inoculant was also used to grow LAB provided a unique
situation in attempting to determine the proportion of the contribution LAB may make to
the final ganglioside concentration. This was the premise for using the starter culture.
No definitive answers were obtained as to the relationship between the presence of
gangliosides in the starter culture to the amount of ganglioside present in the final
product, (cultured dairy foods).

Testing gangliosides isolated from actual milk products will likely provide
information not previously available, and is important in respect to both objectives 2 and
3. The discovery and subsequent isolation and characterization of ganglioside 9-O-acetyl-
GD3 (14) from bovine buttermilk is an important indication of what further research on
actual products could produce. The isolation of this unique ganglioside suggests that
LAB may be involved in the modification of milk fat gangliosides or the production of
other gangliosides not present in the original milk fat.

Rueda et al. (139) determined the effect on preterm infants of supplementing an
adapted milk formula with gangliosides at a total concentration similar to that found in

human milk. In this study it was concluded that ganglioside supplementation did promote

134

bifidobacteria while suppressing E. coli growth. Although the exact mechanism by which
dietary gangliosides reduce fecal levels of E. coli has yet to be determined, interaction
between specific E. coli strains was indicated. In addition, Drago and colleagues (27)
demonstrated growth inhibition of E. coli and S. enteritidis using specific strains of
Lactobacillus isolated from feces. Both of these studies support the fundamental
hypothesis of the present study regarding the feasibility of gangliosides as probiotics. Our
present knowledge is not sufficient to answer some of these important questions, but
essential strides were made in ganglioside research.

Our hemagglutination—inhibition studies clearly demonstrated that gangliosides in
buttermilk, yogurt, LAB, and evaporated milk inhibited the pilus-mediated adhesion of
E. coli, K99: CFA-1+, and H. pylori to erythrocytes. ETEC via pili (hairlike structures on
their surface) have the ability to bind to the villi of the intestine and various erythrocytes.
The receptors for bacteria on erythrocytes and the intestinal receptors were found to be
similar. The basic premise of this study was that if bacteria were to bind to exogenous
gangliosides, fewer pili would be available for adherence to the intestine, or to the in vitro
test group of erythrocytes.

Hemagglutination inhibition was determined by comparing the amount of
erythrocyte present in the supemate when gangliosides and bacteria were present in the
assay to the erythrocyte count when bacteria were included alone. The increase of
erythrocytes in the supernatant in the presence of gangliosides led to the assumption that
gangliosides bound to the bacteria and were not available to bind to the erythrocytes,

leaving them free in the supemate. It is therefore reasonable to conclude that the

135

adhesion of enterotoxigenic E. coli to the intestinal mucosa could be inhibited by the
consumption of dairy products containing gangliosides.

Interpretation of the data complied, indicates that there is an important difference
between the effective inhibition of gangliosides extracted from dairy products and
gangliosides obtained from commercial sources (BBG, GM3). The receptor for K99+ E.
coli has been reported to be GM3 (84), and this specific ganglioside was more effective
than the ganglioside combination present in BBG. The actual quantity of gangliosides
needing to be consumed to fall within the effective range was calculated to be that
contained in three to four 8-ounce yogurt containers. The negative control, K88+ E. coli,
did not agglutinate with the test erythrocytes and did not interfere with the binding that
was observed to occur with gangliosides.

In an attempt to extend the research findings to another relevant enteric disease,
mixed gangliosides from dairy products were tested for their ability to inhibit the
adhesion of H. pylori to chicken erythrocytes using the hemagglutination assay. Recently,
there has been considerable research into the involvement of H. pylori in gastric diseases
(9, 10). This organism is found almost exclusively in the stomach and duodenum, and the
evidence is strong that this organism is significantly involved in the etiology of duodenal
and gastric inflammation and ulcers (10). Epidemiological data indicate that millions of
people in the United States are chronically infected with H. pylori.

The ability of mixed gangliosides extracted from dairy products to inhibit the
adhesion of chicken erythrocytes to H. pylori indicates that the hemagglutination assay
could be useful in research involving this organism. The adherence of H. pylori to gastric

epithelial cells may be critical to the maintenance of infection and ultimate pathogenicity.

136

H. pylori was found to contain a factor that facilitated agglutination with all types of
human and many types of animal erythrocytes (53, 74). Determining whether individuals
suffering from the effects of H. pylori respond positively to ganglioside therapy requires
further research. It is reasonable, however, to assume that many of the findings involving
gangliosides and enteric disease would hold true for H. pylori.

Screening small concentrations of extracted mixed gangliosides for grth
inhibition of enteric pathogens using conventional microbiological methods is tedious
and often produces conflicting results (2, 14, 99). A semi-quantitative plate assay was
developed to determine the antimicrobial activity of mixed gangliosides. This double
agar overlay technique, in which the top layer was a plated lawn of the test pathogen, is
the basis for this assay. The double agar overlay technique has the potential to become an
important tool for the rapid screening of small quantities of gangliosides. This is
extremely important in light of the fact that commercial gangliosides are cost prohibitive
and laboratory extraction processes are labor intensive. Using a growth-inhibiting assay,
the antimicrobial activity of mixed gangliosides extracted from sweet buttermilk and
cultured buttermilk, yogurt, and lactic acid bacteria (LAB) against K88+ and K99+
Escherichia coli (E. coli) was determined and met the criteria of objective three.

Mixed gangliosides were placed into shallow wells that were punched into the
pathogen lawn. Zones of clearing were evident around the wells and corresponded to
growth inhibition of the pathogenic bacteria. The zones were transitional, but they
persisted from 2 to 16 h after incubation. In addition, samples taken from wells and

incubated in a minimal medium for up to 16 h produced no bacterial growth.

137

Mixed gangliosides extracted from sweet buttermilk and cultured buttermilk,
yogurt, and LAB were effective in inhibiting the growth of K99)r E. coli (strain
431,01012K3O), K88+ E. coli (strain 19K974,0141:K85:K88), and E. coli 0157:H7 in the
range of 0.09 mM to 0.51 mM. Mixed gangliosides extracted from lactic acid bacteria
were the most effective in concentrations ranging from .052 mM to .028 mM in
producing zones of inhibition. Standard bovine brain gangliosides (BGG) (Sigma
Chemical, St. Louis) were the least effective, with a concentration of 8.4 mM being
required to produce inhibition zones.

This technique involved placing gangliosides in shallow wells punched into pathogenic
bacterial lawns. Zones of inhibition produced around the wells were subsequently
measured and evaluated according to specific criteria. The results of using the double
agar overlay technique indicated that milk product gangliosides were effective in
inhibiting the growth of K99+ E. coli, K88+ E. coli, and E. coli 0157:H7. The effects
produced by the gangliosides were bacteriostatic in nature because the zones were no
longer observed after 16 h.

Cultured dairy products have long been regarded as beneficial in treating enteric
diseases such as diarrhea. The results from this research will contribute significantly to
understanding the beneficial effects of consuming dairy products. In addition, this
research has the potential to increase the understanding of ganglioside interactions in the

prevention of enteric diseases in humans, especially infants, as well as in young livestock.

138

BIBLIOGRAPHY

139

10.

11.

BIBLIOGRAPHY

Anderson, M. J ., Whitehead, J. S. and Kim, Y. S. (1980) Interaction of
Escherichia coli K88 antigen with porcine intestinal brush border membranes.
Infect. Immun. 29, 897-901

Ashkenazi, S. (1994) A review of the effect of human milk fractions on the

adherence of diarrheogenic escherichia coli to the gut in an animal model. Isr. J.
Med. Sci. 30, 335-338

Awad-Masalmeh, M., Moon, H. W., Runnels, P. L. and Schneider, R. A. (1982)
Pilus production, haemagglutination and adhesion by porcine strains of
enterotoxigenic Escherichia coli lacking K88, K99 and 987P antigens. Infect.
Immun. 35, 305-313

Baker, C. J. and Kasper, D. L. (1976) Identification of sialic acid in
polysaccharide antigen of group B Streptococcus. Infect. Immun. 13, 284-288

Beachey, E. H. (1981) Bacterial adherence: adhesion-receptor interactions
mediating the attachment of bacteria to mucosal surfaces. J. Infect. Dis. 143, 325-
345

Bellamy, J. E., Latshaw, W. K. and Nielsen, N. O. (1973) The vascular
architecture of the porcine small intestine. Can. J. Comp. Med. 37, 56-62

Bennett, H. S. (1963) Morphological aspects of extracellular polysaccharides.
.I. Histochem. Cytochem. 11, 14

Blair, E. B. and Tull, A. H. (1969) Multiple infections among newborns resulting
from colonization with Staphylococcus aureus 502A. Am. J. Clin. Pathol. 52, 42-
49

Blaser, M. J. (1990) Helicobacter pylori and the pathogenesis of gastroduodenal
inflammation. J. Infect. Dis. 161, 626-633

Blaser, M. J. (1992) Hypothesis on the pathogenesis and natural history of
Helicobacter pylori induced inflammation. Gasteroenter. 102, 720-727

Blomberg, L. and Conway, P. L. (1991) Influence of growth phase and carbon

source on the relative synthesis rate of K88 fimbria and the adhesive capacity of
Escherichia coli K88. Microb. Pathog. 11, 143-147

140

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Blomberg, L. and Conway, P. L. (1989) An in vitro study of colonization
resistance to Escherichia coli strain Bd 1107/7508 (K88) in relation to indigenous

squamous gastric colonization in piglets of varying ages. Microb. Ecd. Health
Dis. 2, 285-291

Blomberg, L., Krivan, H., Cohen, P. S. and Conway, P. L. (1993) Piglet ileal
mucus contains protein and glycolipid (Glactosylceramide) receptors specific for
Escherichia coli K88 fimbriae. Inft. and 1mm. 61, 2526-2531

Bonafede, D. M., Macala, L. J ., Constantin-Paton, M. and Yu, R. K. (1989)
Isolation and characterization of ganglioside 9-0-acetyl-GD3 from bovine
buttermilk. Lipids 24, 680-684

Brook, 1., Barett, C., Brinkman, C., Martin W. and Finegold, S. (1979) Aerobic
and anaerobic bacterial flora of the maternal cervix and newborn gastric fluid
conjunctive: a prospective study. Pediatrics 63, 451-455

Bryan, R., Kube, D., Perez, A., Davis, P. and Prince, A. (1998) Overproduction
of the CF TR R domain leads to increased levels of asialoGMl and increased

Pseudomonas aeruginosa binding by epithelial cells. Am. J. Respir. Cell Mol.
Biol. 19, 269-277

Bullen, C. L. and Willis, A. T. (1963) Resistance of the breast fed infant to
gastroenteritis. Br. Med. J. 3, 338

Burnett, G. S. and Hanna, J. (1963) Effect of dietary calcium lactate and lactic
acid on faecal Escherichia coli counts in pigs. Nature (London) 197, 815

Burrows, M. R., Sellwood, R. and Gibbons, R. A. (1976) Haemagglutinating and
adhesive properties associated with the K99 antigen of bovine strains of
Escherichia coli. J. Gen. Micro. 96, 269-275

Cahan, L. D., Irie, R. F. Singh, R., Cassidenti, R. and Paulson, J. C. (1982)
Identification of human neuroectoderman tumor antigen (ORA-l-2) as ganglioside
GDZ. Proc. Natl. Acad. Sci. 4, 381-383

Charych, D., Cheng, Q., Reichert, A., Kuziemko, G., Stroh, M., Nagy, J. 0.,
Spevak, W. and Stevens, R. C. (1996) A “litmus test” for molecular recognition
using artificial membranes. Chem. Biol. 3, 113-120

Collins, B. E., Kiso, M., Hasegawa, A., Tropak, M. B., Roder, J. C., Crocker,
P. R. and Schnaar, R. L. (1997) Binding specificities of the sialoadhesin family of
I-type lectins. Sialic acid linkage and substructure requirements for binding of

myelin-associated glycoprotein, Schwann cell myelin protein, and sialoadhesin. J.
Biol. Chem. 272, 16889-16895

141

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

Conway, P. L. (1989) Lactobacilli: fact and fiction. In The Regulation and
Protective Role of the Normal Microfiora (Grubbe, R., Midvedt, T., and Norin, E.,
eds.) pp. 263-281, Macmillan, New York

Dean-Nystrom, E. A. and Samuel, J. E. (1994) Age-related resistance to 987P
fimbria-mediated colonization correlates with specific glycolipid receptors in
intestinal mucus in swine. Infect. Immun. 62, 4789—4794

Deans, A. G., Ching, Y. C., Williams, R. G. and Harden, L. G. (1972) Test for
Escherichia coli enterotoxin using infant mice: application in a study of diarrhea
in children in Honolulu. J. Infect. Dis. 125, 407-411

DeGraff, F. K. and Rhoorda, I. (1982) Production, purification and
characterization of the fimbrial adhesive antigen F41 isolated from the calf entero-
pathogenic Escherichia coli strain B41M. Infect. Immun. 36, 751-758

Drago, L., Gismondo, M. R., Lombardi, A., de Haen, C. and Gozzini, L. (1997)
Inhibition of in vitro growth of enteropathogens by new Lactobacillus isolates of
human intestinal origin. FEMS Microbiol. Lett. 153, 455-463

Drasar, B. S. and Hill, M. J. (1974) Human Intestinal Flora. Academic Press,
London

Drutz, D. J., VanWay, M. H., Schaffner, W. and Kownig, M. G. (1966) Bacterial
interference in the therapy of recurrent staphylococcal infections: multiple

abscesses due to implantation of 502A strain of staphylococcus. N. Engl. J. Med.
275,1161-1165

Ducluzeau, R. (1989) Role of experimental microbial ecology in
gastroenterology. In Micro. Ecol. and Intest. Infect. (Berezin, B., ed.) p. 7,
Springer-Verlag, Paris

Duguid, J. P., Clegg, S. and Wilson, M. I. (1979) The fimbrial and non-fimbrial
haemagglutinins of Escherichia coli. J. Med. Micro. 12, 213-227

Duguid, J. P. and Old, D. C. (1980) Adhesive properties of Enterobacteriaceae.
In Bacterial Adhesion Receptors and Recognition (Beachy, E. T., ed.) ser. B, vol.
6, Hapman and Hall, London

Duguid, J. P., Smith, I. W., Dempster, G. and Edmunds, P. N. (1955) Non-
flagellar filamentous appendages (fimbriae) and haemagglutinating activity in
Bacterium coli. J. Path. and Bact. 70, 335-348

Ehrenkrantz, N. J. (1970) Bacterial colonization of newborn infants and
subsequent acquisition of hospital bacteria. J. Pediatr. 76, 839-847

142

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

Ellis, S. and Spratt, J. S. (1970) Lactobacillus overgrowth diet as an aid in
controlling Escherichia coli septicemia and endotoxemia in cancer patients: a case
report. J. Am. Geriatr. Soc. 18, 410-415

Erickson, A. K., Willgohs, J. A., McFarland, S. Y., Benfield, D. A. and Francis
D. H. (1992) Identification of two porcine brush border glyCOproteins that bind
the K88 adhesion of Escherichia coli and correlation of these glycoproteins with
the adhesive phenotype. Iinfect. Immun. 60, 983-988

Evans, D. G., Evans, D. J ., Jr. and Tjoa, W. (1977) Haemagglutination of human
group A erythrocytes by enterotoxigenic Escherichia coli isolated from adults
with diarrhea: correlation with colonization factor. Infect. Immun. 18, 330-337

Evans, D. G. and Evans, D. J., Jr. (1978) New surface-association heat-labile
colonization factor (CFA/II) produced by enterotoxigenic Escherichia coli of
serogroups 06 and 08. Infect. Immun. 21, 638-647

Evans, D. J., Jr., Evans, D. G. and Dupont, H. L. (1979) Haemagglutination
patterns of enterotoxigenic and enteropathogenic Escherichia coli determined
with human, bovine, chicken and guinea-pig erythrocytes in the presence and
absence of mannose. Infect. Immun. 23, 336-346

Evans, E. J ., Jr., Evans, D. G., Young, L. S. and Pitt, J. (1980) Haemagglutina-
tion typing of Escherichia coli: definition of seven haemagglutination types. J.
Clin. Micro. 12, 235-242

Evans, D. G., Evans, D. J., Jr., Moulds, J. J. and Graham, D. Y. (1988) N-
acetylneuraminyllactose-binding fibrillar hemagglutination of Campylobacter
pylori: a putative colonization factor antigen. Infect. Immun. 56, 2896-2906

Faris, A., Lindahl, M. and Wadstrom, T. (1980) sz-like glycoconjugate as
possible erythrocyte receptor for the CFA/I and K99 haemagglutinins of
enterotoxigenic Escherichia coli. FEMS Microbiol. Lett. 7, 625-629

Florey, H. W. (1946) The use of microorganisms for therapeutic purposes. Yale
J. Biol. Med. 19, 101-117

Fujiwara, S., Hashiba, H., Hirota, T. and Forstner, J. F. (1997) Proteinaceous
factor(s) in culture supernatant fluids of bifidobacteria which prevents the binding

of enterotoxigenic Escherichia coli to gangliotetraoslyceramide. Appl. Enviro.
Micro. 63, 506-512

Fuller, R. (1989) Probiotics in man and animals. J. App. Bact. 66, 365-378

143

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

Galustian, C., Childs, R. A., Yuen, C. T., Hasegawa, A., Kiso, M., Lubineau, A.,
Shaw, G. and F eizi, S. (197 7) Valency dependent patterns of binding of human L-
selection toward sialyl and sulfated oligosaccharides of Le(a) and Le(x) types:
relevance to anti-adhesion therapeutics. Biochem. 36, 5260-5266

Gaastra, W. and deGraff, E. K. (1982) Host-specific fimbrial adhesions of
noninvasive enterotoxigenic intestinal Escherichia coli strains. Micro. Rev. 46,
129-161

Giron, J. A., Ho, A. S. Y. and Schoolnik, G. K. (1991) An inducible bundle-
forming pilus of enteropathogenic Escherichia coli. Sci. 254, 710-713

Goldin, B. R., Swenson, L., Dwyer, J ., Seeton, M. and Gorbach, S. L. (1980)
Effect of diet and Lactobacillus acidophilus supplements on human fecal bacterial
enzymes. J. Nutri. Clin. 64, 255-260

Goldin, B. R. and Gorbach, S. L. (1984) The effect of milk and Lactobacilli
feeding on human intestinal bacteria enzyme activity. Am. J. Clin. Nutr.

Goldman, A. S. and Smith, C. W. (1973) Host resistance factors in human milk.
.1. Pediatr. 82, 1082-1090

Gonwa, T. A., Westrick, M. A. and Macher, B. A. (1984) Inhibition of mitogen
and antigen-induced lymphocyte activation by human leukemia cell gangliosides.
Cancer Res. 44, 3467-3470

Grahm, D. Y., Lew, G. M. and Klain, P. D. (1992) The effect of treatment of
Helicobacter pylori infection on the long-term recurrence of gastric or duodenal
ulcer: a randomized, controlled study. Ann. Intern. Med. 116, 705-708

Grange, P. A., Erickson, A. K., Anderson, T. J. and Francis, D. H. (1998)
Characterization of the carbohydrate moiety of intestinal mucin-type

sialoglycoprotein receptors for the K88ac fimbrial adhesion of Escherichia coli.
Infect. Immun. 66, 1613-1621

Grange, P. A. and Mouricout, M. (1996) Susceptibility of infant mice to F5 (K99)
E. coli infection: differences in glycosyltransferase activities in intestinal mucosa

of inbred CBA and DBA/2 strains. J. Glycoconj. 13, 45-52

Guyot, G. (1980) Bacterial haemagglutination. J. Bact., Parasit, Infect. and
Hygiene. 47, 640-653

Gyorgy, P. (1971) The uniqueness of human milk. Biochemical aspects. Amer. J.
Clin. Nutr. 27, 970

144

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

Hakomori, S. and Kahhagi, R. (1983) Glycosphingolipids as tumor associated
and differentiation markers. J. Nat. Cane. Inst. 71, 231-251

Herrick, J. B. (1972) Therapeutic nutrition using Lactobacillus species. Vet.
Med. and Sm. Anim. Clin. 67, 1249

Hevermann, D., Roggentin, P., Kleineidam and Schauer, R. (1991) Purification
and characterization of sialidases from Clostridium chauvoei NC08596. Glyco. J.
8, 9.5-101

Higashi, H., Hirabayashi, Y., Fukui, Y., Naiki, M., Matsumoto, M., Vada, S. and
Kato, S. (1985) Characterization of N-glycolyl-neuramic acid-containing

gangliosides as tumor-associated Hanganutziu-Deicher antigen in human colon
cancer. Cancer Res. 45, 2795-3802

Hill, I. R., Kenworthy, R. and Porter, P. (1970) The effect of dietary Lactobacilli
on in-vitro catabolic activities of the small intestine microflora of newly weaned
pigs. J. Med. Micro. 3, 593-605

Holmgren, J., Svennerholm, A. M. and Lindahl., M. (1983) Receptor-like
glycocompounds in human milk that inhibit classical and El Tor Vibrio cholerae
cell adherence (hemagglutination). Infect. Immun. 39, 147-154

Holland, R. E. (1990) Some infectious causes of diarrhea in young farm animals.
Clinical Micro. Reviews 23, 345-375

Huang, R. T. C. (1973) Isolation and characterization of the gangliosides of
buttermilk. Biochem. Biophys. Acta. 306, 82-84

Irie, R. F., Irie, K. and Morton, D. L. (1976) A membrane antigen common to
human cancer and fetal brain tissues. Cancer Res. 36, 3510-3517

Irie, R. F. and Morton, D. L. (1986) Regression of cutaneous metastatic
melanoma by intralesional injection with human monoclonal antibody to

ganglioside GDZ. Proc. Natl. Acad. Sci. 83, 8694-8698

Irie, R. F., Chandler, P. J. and Morton, D. L. (1989) Melanoma gangliosides and
human monoclonal antibody. In Human Tumor Antigens and Specific Tumor
Therapy (Metzgar, R. S. and Mitchel, M. S., eds.) pp. 115-126, Alan R. Liss, New
York

Jelliffe, D. B. and Jelliffe, E. F. P. (1971) The uniqueness of human milk. Am. J.
Clin. Nutr. 24, 968 '

145

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

Johnson, D. E. and Calia, F. M. (1979) The effect of Lactinex on rabbit ileal loop
reactions induced by enterotoxigenic Escherichia coli. Curr. Microbiol. 2, 207-
210

Jolivet-Reynaud, C., Launay, J. M. and Alouf, J. E. (1988) Damaging effect of
Clostridium perfringens delta toxin on blood platelets and their relevance to
ganglioside GMZ. Arch. Biochem. Biophys. 262, 59-66

Jones, G. W. and Rutter, J. M. (1974) The association of K88 antigen with
haemagglutinating activity in porcine strain of Escherichia coli. J. Gen. Micro.
84, 135-144

Kabara, J. J. (1980) Lipids as host resistance factors of human milk. Nutr.
Reviews 2, 65-73

Kamisago, S., Iwamori, M., Tai, T., Mitamura, K., Yazaki, Y. and Sugano, K.
(1996) Role of sulfatides in adhesion of Helicobacter pylori to gastric cancer
cells. Infect. Immun. 64, 624-628

Kanfer, J. N. and Hakamori, S. (1983) Handbook of Lipid Research—Sphingolipid
Biochemistry, Plenum, New York

Karlsson, K. A. (1982) Glycosphingolipids and surface membranes. Biological
Membranes, vol. 4, p. 7, Academic Press, New York

Karlsson, K. A., Teneberg, S., Angstrom, J ., Kjellberg, A., Hirst, T. R., Berstrom,
J. and Miller-Podraza, S. (1996) Unexpected carbohydrate cross-binding by
Escherichia coli heat-labile enterotoxin. Recognition of human and rabbit target

cell glycoconjugates in comparison with cholera toxin. Bioorg. Med. Chem. 4,
1919-1928

Kawaguchi, T., Nakakoma, H., Kagimoto, T., Shirono, K., Horikawa, K., Midake,
M., Iwamori, M., Nagai, V. and Takatsuke, K. (1989) Characteristic mode of
action of gangliosides in selective modulation of CD4 on human T lymphocytes.
Biochem. Biophys. Res. Common. 158, 1050-1059

Keenan, T. W. (1974) Composition and synthesis of gangliosides in mammary
gland and milk of the bovine. Biochem. Biophys. Acta. 337, 255-270

Krivan, H. C., Ginsburg, V. and Roberts, D. D. (1988) Pseudomonas aeruginosa
and Pseudomonas cepacia isolated from cystic fibrosis patients bind specifically
to gangliotetraosylceramide (asialo GMl) and gangliotriaosylceramide (asialo

GMZ). Arch. Biochem. Biophys. 260, 493-496

146

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

Kolars, J. C., Lewit, M. D., Aouji, M. and Savoiang, D. A. (1984) Yogurt: an
autodigesting source of lactose. New Eng. J. Med. 310, 1-3

Kozaki, S., Kamata, Y., Watarai, S., Nishiki, T. and Mochida, S. (1998)
Ganglioside thb as a complementary receptor component for Clostridium
botulinum neurotoxins. Microb. Pathog. 25, 91-99

Kurasz, A. B., Hanzer, J. M., Bazer, L. and Savoldi, E. (1986) In vitro studies of
growth and competition between S. salivarius Tore-R and mutans streptococci. J.
Dent. Res. 65, 1149-1153

Kyogashima, M., Ginsburg, V. and Krivan, H. (1989) E. coli K99+ binds to N-
glycolylsialoparagloboside and N-glycolylGM3 found in piglet small intestine.
Arch. Biochem. Biophys. 270, 291-297

Laegreid, A., Lolsto-Otnaess, A. B. and Fulgelsang, J. (1986) Human and bovine
milk: comparisons of ganglioside composition and enterotoxin inhibitory activity.
Pediatric Res. 20, 416-421

Ladish, S., Ulsh, L., Gillard, B. and Wong, C. (1984) Modulation of the immune
response by gangliosides: inhibition of adherent monocyte accessory function in
vitro. J. Clin. Invest. 74, 2074-2081

Ladish, S., Becker, H. and Ulsh, L. (1991) Carbohydrate structure and
immunosuppression by human gangliosides. Proc. Am. Assoc. Cancer Res. 32,
226

Ledeen, R. W. and Yu, R. K. (1982) Gangliosides: structure, isolation and
analysis. In Complex Carbohydrates (part D) (Ginsburg, V., ed.). In Methods in
Enzymology (Colowick, W. and Kaplan, N., eds.) pp. 139-191, Academic Press,
New York

Lee, K. K., Yu, L., Macdonald, D. L., Paranchych, W., Hodges, R. W. and Irvin,
R. T. (1996) Anti-adhesion antibodies that recognize a receptor-binding motif
(adhesintope) inhibit pilus/fimbrial-mediated adherence of Pseudomonas

aeruginosa and Candida albicans to asialo-GMl receptors and human buccal
epithelial cell surface receptors. Can. J. Micro. 42, 479-486

Lidbeck, A., Gustafsson, J. A. and Nord, C. E. (1987) Impact of Lactobacillus
acidophilus supplements on the human oropharyngeal and intestinal microflora.
Scand. J. Infect. Dis. 49, 64-72

Lidbeck, A., Edlund, D., Gustafsson, J. A., Kager, L. and Nord, C. E. (1989)

Impact of Lactobacillus acidophilus administration on the intestinal microflora
after enoxacin treatment. Microecol. Them. 18, 97-102

147

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

Lindahl, M. (1989) Binding F41 and K99 fimbriae of enterotoxigenic Escherichia
coli to glycoproteins from bovine and porcine colostrum. Microbiol.
Immunol. 33, 373-379

Lindahl, M., Paris, A. and Wadstrom, T. (1984) A GMz-like receptor for CFA/I
and K99. In Attachment of Organisms to Gut Mucosa (Boedecker, E. C., ed.) vol.
II, pp. 190-195, CRC Press, Boca Raton, FL

Lindahl, M., Brossmer, R. and Wadstrdm, T. (1987) Carbohydrate receptor
specificity of K99 fimbriae of enterotoxigenic Escherichia coli. Glycon. J. 4, 51-
58

Lingwood, C. A., Huesca, M. and Kuksis, A. (1992) The glycerolipic receptor for
Helicobacter pylori (and exoenzymes) is phosphatidylethanol-amine. Infect.
Immun. 60, 2470-2474

Livingston, P. O. (1993) Approaches to augmenting the IgG antibody response to
melanoma ganglioside vaccines. Annals of the N. Y. Academy of Sciences 690,
204-213

Lloyd, A. B., Cumming, R. B. and Kent, R. D. (1977) Prevention of Salmonella
typhimurium infection in poultry by pretreatment of chicken and poults with
intestinal extracts. Austr. Vet. J. 53, 82-87

Ludinova, R., Jouja, V. and Wagner, V. (1973) Serum immunoglobulins and
copro-antibody formation in infants after artificial intestinal colonization with
Escherichia coli 083 and oral lysozyme administration. Pediatr. Res. 7, 659-669

MacKenzier, C. R., Hirama, T., Lee, K. K., Altman, E. and Young, N. M. (1997)
Quantitative analysis of bacterial toxin affinity and specificity for glycolipid
receptors by surface plasmon resonance. .1. Biol. Chem. 272, 5533-5538

Mann, G. V. (1977) A factor in yogurt which lowers cholesterolemia in man.
Atherosclerosis 26, 335-340

Marshall, B. J. and Warren, J. R. (1984) Unidentified curved bacilli in the
stomach of patients with gastritis and peptic ulceration. Lancet 1, 1311-1314

Marshall, B. J., Barrett, L. J ., Prakash, C., McGallum, R. W. and Guerrant, R. L.

(1990) Urea protects Helicabacter (campylobacter) pylori from the bactericidal
effect of acid. Gastroenterology 99, 697-702

148

103.

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

Marteau, P., Pochart, P., Flourie, B., Pellier, P., Santos, L., Desjeux, J. F. and
Rambaud, J. C. (1990) Effect of chronic ingestion of a fermented dairy product
containing Lactobacillus acidophilus and Bifidobacterium bifidum on metabolic
activities of the colonic flora in humans. Amer. J. Clin. Nutr. 52, 685-688

McNamara, D. J ., Alison, E. L. and Sabb, J. E. (1987) Effect of yogurt intake on
plasma lipid and lipoprotein levels in normolipidemic males. Atherosclerosis 79,
167-171

Merritt, W. D., Bailey, M. and Pluznik, K. H. (1984) Inhibition of interleukin-2
dependent cytotoxic T-lymphocyte grth by gangliosides. Cell. Immunol. 89, 1-
10

Merson, M. H., Morris, G. K. and Sack, D. A. (1976) Travelers diarrhea in
Mexico. New Eng. J. Med. 294, 1299-1301

Metchnikoff, E. (1907) The Prolongation of Life (Mitchell, P. C., trans),
Heinemann, London

Miller, P. H., Bergstrom, J ., Milh, M. A. and Karlsson, K. A. (1997) Recognition
of glycoconjugates by Helicobacter pylori. Comparison of two sialiacid-
dependent specificities based on haemagglutination and binding to human
erythrocyte glycoconjugates. Glycoconj. J. 14, 467-471

Mitchell, 1. D. G. and Kenworthy, R. (1976) Investigations on a metabolite from
Lactobacillus bulgaricus which neutralizes the effect of enterotoxin from
Escherichia coli pathogenic for pigs. J. Appl. Bacterol. 41, 163-174

Moon, H. W., Nagy, B., Isaacson, R. E. and Orskov, I. (1977) Occurrence of K99
antigen on Escherichia coli isolated from pigs and colonization of pig ileum by
K99+ enterotoxigenic E. coli from calves and pigs. Infect. Immun. 15, 614-620

Moon, H. W. (1978) Mechanisms in the pathogenesis of diarrhea: a review.
J. Amer. Vet. Med. Assoc. 172, 443-448

Morris, J. A., Thorns, C. J., Wells, G. A. H., Scott, A. C. and Sopjka, W. J. (1983)
The production of F41 fimbriae by piglet strains of enterotoxigenic Escherichia
coli that lack K88, K99 and 987P fimbriae. J. Gen. Microbiol. 129, 2753-2759

Morrison, W. J., Offner, H. and Vandenbark, A. A. (1989) Specific ganglioside

binding to receptor sites on T lymphocytes that couple to ganglioside-induced
decrease of CD4 expression. Life Sci. 45, 1219-1226

149

 

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

124.

Morton, D. L., Holmes, E. C., Eilber, F . R. and Ramming, K. P. (1982) Adjuvant
immunotherapy of malignant melanoma: results of a randomized trial in patients
with lymph node metastasis. In Immunotherapy of Human Cancer (Rosenberg,
S. A., ed.) pp. 245-253, Excerpta Medica, New York

Morton, D. L. and Ravindranatah, M. H. (1993) Active specific immunotherapy
with vaccines. In Cancer Medicine 3rd ed. (Holland, J. F., ed.) pp. 913-926, Lea
and F ebiger, Philadelphia

Mouricout, M., Petit, J. M., Carias, J. R. and Julien, R. (1990) Glycoprotein
glycans that inhibit the adhesion of Escherichia coli mediated by K99 fimbriae:
treatment of experimental colibacillosis. Infect. Immun. 58, 98-106

Nader de Macias, M., Apella, M., Romero, N., Gonzales, S. and Oliver, G. (1992)
Inhibition of Shigella sonnei by Lactobacillus casei and Lactobacillus
acidophilus. J. Applied Bact. 73, 407-411

Nahar, T. 0., Williams, N. A. and Hirst, T. R. (1996) Cross-linking of cell
surface ganglioside GMl induces the selective apoptosis of mature CD8+ T
lymphocytes. Int. Immunol. 8, 731-736

Natro, J. P. and Kaper, J. B. (1998) Diarrheogenic Escherichia coli. Clin. Micro.
Rev. 11, 142-201

Nesser, J. R., Koelreutter, B. and Wuersch, P. (1986) Oligosaccharide-type
glycopeptides inhibiting adhesion of Escherichia coli strain mediated by type I
pili: preparation of potent inhibitors from plant glycoproteins. Infect. Imm. 52,
428-436

Nohara-Uchida, K. and Ohashi, M. (1987) Isolation and characterization of major
gangliosides from liver. J. Biochem. 102, 923-932

Offner, H., Thieme, T. and Vandenbark, A. A. (1987) Gangliosides induce
selective modulation of CD4 from helper T lymphocytes. J. Immunol. 139, 3295-
3305

Old, D. C. (1985) Haemagglutination methods in the study of Escherichia. The
Virulence of Escherichia: Reviews and Methods (Sussman, M., ed.) pp. 287-313,
Academic Press, New York

Ono, F ., Abe, K., Nakazawa, M. and Naiki, M. (1989) Ganglioside epitope
recognized by K99 fimbriae from enterotoxigenic Escherichia coli. Infect. and
Immun. 57, 907-911

150

 

n. i“

125.

126.

127.

128.

129.

130.

131.

132.

133.

134.

135.

Otnaess, A. B. and Svennerholm, A. M. (1982) Nonimmunoglobulin fraction of
human milk protects rabbits against enterotoxin-induced intestinal fluid secretion.
Infect. Immun. 35, 738-740

Parry, S. H. and Porter, P. (1978) Immunological aspects of cell membrane
adhesion demonstrated by porcine enteropathogenic Escherichia coli.
Immunol. 34, 41-49

Parry, S. H. and Rooke, D. M. (1985) Adhesions and colonization factors of
Escherichia coli. In The Virulence of Escherichia coli (Sussman, M., ed.) pp. 79-
155, Academic Press, London

Payne, D., O’Reilly, S. and Williamson, D. (1993) The K88 fimbrial adhesion of
enterotoxigenic Escherichia coli binds to Bl-linked galactosyl residues in
glycosphingolipids. Infect. Immun. 61, 3673-3677

Payne, S. M. and Finkelstein, R. A. (1977) Detection and differentiation of iron-
responsive avirulent mutants on Congo red agar. Infect. Immun. 18, 94-98

Perdigén, S. M., Nader de Macias, M. E., Alvarez, S., Oliver, G. and Pesce de
Ruiz Holgado, A. (1990) Prevention of gastrointestinal infection using
immunobiological methods with milk fermented with Lactobacillus casei and L.

acidophilus. J. Dairy Research 57, 255-264

Peterson, W. L. (1991) Helicobacter pylori and peptic ulcer disease. New Eng. J.
Med. 324, 1043-1048

Prokazova, N. V., Dyatloritskaya, E. V. and Bergelson, L. D. (1988) Sialylated
lactosylceramides. Possible inducers of non-specific immunosuppression and
atherosclerotic lesions. Eur. J. Biochem. 171, 1-6

Radecki, S. V., Miller, E. R. and Yokoyama, M. T. (1989) Antimicrobial effects
of bovine brain gangliosides on enterotoxigenic Escherichia coli K88 and K99
isolated from swine. Animal Science Swine Report, Michigan State University,
East Lansing

Rafiee, P., Leffler, H., Byrd, J. C., Cassels, F. J ., Boedeker, E. C. and Kim, Y. S.
(1991). A sialoglycoprotein complex linked to the microvillus cytoskeleton acts

as a receptor for pilus (AF /R1) mediated adhesion of enteropathogenic
Escherichia coli (RDEC-l) in rabbit small intestine. J. Cell Bio. 115, 1021-1029

Rauws, E. J., Langenberg, W., Houthoff, H. J ., Zanen, H. C. and Tytgat, G. N.
(1988) Campylobacter pyloridis-associated chronic active antral gastritis: a
prospective study in prevalence and the effect of antibacterial and antiulcer
treatment. Infect. Immun. 74, 33-40

151

136.

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

Reddy, G. V. and Shahani, K. M. (1971) Isolation of an antibiotic from
Lactobacillus bulgaricus. J. Dairy Sci. 54, 748

Reid, G., Chan, R. C., Bruce, A. W. and Costerton, J. W. (1985) Prevention of
urinary tract infection in rats with an indigenous Lactobacillus casei strain. Infect.
Immun. 49, 320-324

Ren, S., Araga, T., Scarsdale, J ., Zhang, Y., Slominski, A., Livingston, P., Ritter,
G., Kushi, Y. and Yu, R. K. (1992) Characterization of a hamster melanoma-

associated ganglioside antigen as 7-O-acetylated disialoganglioside GD3. J. Biol.
Chem. 267, 12632-12638

Reuda, R., Sabatel, J. L., Maldonado, J ., Molina-Font, J. A. and Gil, A. (1998)
Addition of gangliosides to an adapted milk formula modifies levels of fecal
Escherichia cali in preterm newborn infants. J. Pediatr. 133, 90-94

Robins-Browne, R. M. and Levine, M. M. (1981) The fate of ingested lactobacilli
in the proximal small intestine. Amer. J. Clin. Nutr. 34, 514-519

Rudin, A., McConnell, M. M. and Svennerholm, A. (1994) Monocolonial
antibodies against enterotoxigenic colonization factor antigen I (CF A/I) that cross-
react immunologically with heterologous CF As. Infect. Immun. 62, 4339-4346

Sack, D. A., Lastovica, A. J ., Chang, S. H. and Pazzaglia, G. (1998) Microtiter
assay for detecting Campylobacter spp. and Helicabacter pylori with surface
gangliosides which bind cholera toxin. J. Clin. Microbiol. 36, 2043-2045

Sack, K. B. (1975) Human diarrheal disease caused by enterotoxigenic
Escherichia coli. Ann. Rev. Microbial. 29, 333

Sakellaris, G., Kolisis, F. N. and Evangelopoulis, A. E. (1988) Presence of sialic
acids in Lactobacillus plantarum. Biochem. Biophys. Res. Comm. 115, 1126-
1132

Savage, D. C. (1977) Microbial ecology of the gastrointestinal tract. Ann. Rev.
Microbial. 31, 107-133

Savage, D. C. (1983) Mechanisms by which indigenous microorganism colonize
gastrointestinal epithelial surfaces. Progress in F 00d and Nutrition Science 7, 65-

68
Schaedler, R. W. and Dubos, R. J. (1962) The fecal flora of various strains of

mice. Its bearing on their susceptibility to endotoxin. J. Experimental Medicine
15, 1149-1160

152

148.

149.

150.

151.

152.

153.

154.

155.

156.

157.

158.

159.

Schauer, R. (ed.) (1982) Sialic Acids, Springer-Verlag, New York

Schneider, R. A. and To, S. C. (1982) Enterotoxigenic Escherichia coli strains
that express K88 and 987P pilus antigens. Infect. Immun. 36, 417-418

Schroten, H., Hanisch, F. G., Plogmann, R., Hacker, J ., Uhlenbruck, G., Nobis-
Bosch, R. and Wahn, V. (1992) Inhibition of S-fimbriated Escherichia cali to
buccal epithelial cells by human milk fat globule membrane components: a novel
aspect of the protective functions of mucins in the nonimmunoglobin fraction.

Infect. Immun. 60, 2893-2899

Schwartz, D. J ., Kimberg, D. V. and Sheerin, H. E. (1974) Vasoactive intestinal
peptide stimulation of adenylate cyclase and active electrolyte secretion in
intestinal mucosa. J. Clin. Invest. 54, 536-544

Schwertmann, A., Rudloff, S. and Kunz, C. (1996) Potential ligands for cell
adhesion molecules in human milk. Ann. Nutr. Metab. 48, 252-262

Sellwood, R. R., Gibbons, A., Jones, G. W. and Rutter, J. M. (1975) Adhesion of
enteropathogenic Escherichia cali to pig intestinal brush borders; the existence of
two pig phenotypes. J. Med. Microbiol. 8, 405-411

Shida, K., Takamizawa, K. and Osawa, T. (1996) Inhibition of Escherichia cali
heat-liable enterotoxin by an amino carbonyl product, lactose-alpha-lactalbumin.
Biasci. Biotechnol. Biochem. 60, 1492-1494

Shinefield, H. R., Ribble, J. C., Boris, M. and Eichenwald, H. F. (1963) Bacterial
interference: Its effect on nursery acquired infections with Staphylococcus aureus
1, Ii, III, IV. Am. J. Dis. Child. 105, 646-682

Shore, E. G., Dean, A. G., Holik, K. J. and Davis, B. R. (1974) Enterotoxin
producing Escherichia cali and diarrheal disease in adult travelers: a prospective
study. J. Infect. Dis. 129, 577-582

Simon, P. M., Goode, P. L., Mobasseri, A. and Zopf, D. (1997) Inhibition of
Helicobacter pylori binding to gastrointestinal epithelial cells by sialic acid-
containing oligosaccharides. Infect. Immun. 65, 750-757

Smit, H. W., Gaastra, J. P., Kamerling, J. F ., Gliegenthart, G. and deGraaf, F. K.
(1984) Isolation and structural characterization of the equine erythrocytes

receptor for enterotoxigenic Escherichia cali K99 fimbrial adhesion. Infect.
Immun. 46, 578-584

Smith, H. W. and Halls, S. (1967) Studies on Escherichia cali enterotoxins. J.
Pathol. Bacterial. 93, 531-543

153

160.

161.

162.

163.

164.

165.

166.

167.

168.

169.

170.

Smith, H. W. and Lingood, M. A. (1971) Observations on the pathogenic
properties of K88, Hly and Ent plasmids of Escherichia with particular reference
to porcine diarrhoea. J. Med. Microbial.

Snoeyenbos, G. H., Weinack, O. M. and Soerjade, A. (1978) Protecting chicks
and poults from salmonellae by oral administration or normal gut microflora.
Avian Dis. 22, 273-278

Soerjade, A. S., Snoeyenbos, G. H. and Weinack, O. M. (1982) Intestinal
colonization and competitive exclusion of Campylobacter fetus subsp. jejuni in
young chicks. Avian Dis. 26, 520-524

Soerjade-Liem, A. S., Snoeyenbos, G. H. and Weinack, O. M. (1984)
Establishment and competitive exclusion of Yersinia enterocolitica in the gut of
monoxenic and holoxenic chicks. Avian Dis. 28, 256.

Sonnino, S. D., Azquotti, V., Chigomo, M., Pitto, C., Montecucco, G., Schivo, G.,
Kirschner, S. and Tettamanti, G. (1988) Recent advances in the chemistry and the
technology of gangliosides. Indian Journal of Biochem. and Biophys. 25, 144-149

Staley, T. E. and Wilson, I. B. (1983) Soluble pig intestinal cell membrane
components with affinities for E. cali K88+ antigen. Molec. and Cellular
Biochemistry 52, 177-189

Stapleton, A. E., Stroud, M. R., Hakamori, S. I. and Stamm, W. E. (1998) The
globoseries glycosphingolipid sialosyl galactosyl globoside is found in urinary
tract tissues and is a preferred binding receptor in vitro for uropathogenic

Escherichia coli expressing pap-encoded adhesions. Infection and Immun. 66,

3856-3861

Svanborg, Edén, C. and Leffler, H. (1979) Glycosphingolipids of human urinary
tract epithelial cells as possible receptors for adhering Escherichia cali bacteria.
Symposium on Host Parasite Interactions. Scand. J. Infect. Dis. 24, 144

Svennerholm, L. and Friedman, P. (1980) A procedure for the quantitative
isolation of brain gangliosides. Biochem. Biophys. Acta. 617, 97-109

Tai, T., Cahan, L. D., Paulson, J. C. and lrie, R. F. (1983) Ganglioside GMZ as a
human tumor antigen (OFA-Il). Proc. Natl. Acad. Sci. 80, 5392-5396

Takamizawa, K., Iwamori, M., Mutai, M. and Nagai, Y. (1986) Selective changes

in gangliosides of human milk during lactation: a molecular indicator for the
period of lactation. Biochem. Biophys. Acta. 879, 73-77

154

171.

172.

173.

174.

175.

176.

177.

178.

179.

180.

181.

182.

Teneberg, S., Willemsen, P., deGraaf, F. K. and Karlsson, K. A. (1990) Receptor—
active glycolipids of epithelial cells of the small intestine of young and adult pigs
to susceptibility to infection with Escherichia coli. FEBS Letters 263, 10-14

Teneberg, S., Willemsen, P., deGraaf, F. K. and Karlsson, K. A. (1993) Calf
small intestine receptors for K99 fimbriated enterotoxigenic Escherichia coli.
FEMS Micro. Lett. 109, 107-112

Tettamanti, G., Ghidoni, R. and Trinchera, M. (1988) Recent advances in
ganglioside metabolism. Indian J. Biochem. Biophys. 25, 106-111

Tuomola, E. M. and Salminen, S. J. (1998) Adhesion of some probiotic and dairy
Lactobacillus strains to Caco-2 cell cultures. Int. J. Food Microbial. 41, 45-51

Tytgat, G. J. and Rauws, E. J. (1990) Campylobacter pylori and its role in peptic
ulcer disease. Gastraentera. Clin. North Am. 19, 183-196

Umesaki, Y., Okada, Y., Matsumoto, S., Imaoka, A. and Steoyama, S. (1995)
Segmented filamentous bacteria are indigenous intestinal bacteria that activate
intraepithelial lymphocytes and induce MHC class II molecules and fucosyl asialo

GMl glycolipids on the small intestinal epithelial cells in the ex-germ-free mouse.
Micro. Immun. 39, 555-562

VanHeyningen, W. E., Carpenter, C. J ., Pierce, N. F. and Greenough, W. B. 111
(1971) Deactivation of cholera toxin by ganglioside. J. Infect. Dis. 124, 415-418

Vincent, J. R., Veomett, R. C. and Riley, R. F. (1959) Antimicrobial activity
associated with Lactobacillus acidophilus. J. Bacterial. 78, 477-484

Waaij, D. F., Berghuis-DeVries, J. M. and Lekkerkerk-Wees, J. E. C. (1971)
Colonization resistance of the digestive tract in conventional and antibiotic-treated
mice. J. Hygiene 69, 405-411

Waaij, D. V. (1984) The immunoregulation of the intestinal flora: experimental
investigations on the development and the composition of the microflora in
normal and thymusless mice. Microecol. and Ther. 14, 63-74

Weigandt, H. (1985) In New Comprehensive Biochemistry (Neuberger, A. and
VanDeenen, L. L. M., eds.) vol. 10, pp. 199-260, Elsevier, Amsterdam

Weinack, O. M., Snoeyenbos, G. H., Smyser, C. F. and Soerjadi, A. S. (1981)

Competitive exclusion of intestinal colonization of Escherichia coli in chicks.
Avian Dis. 25, 696-705

155

183.

184.

185.

186.

187.

188.

189.

Wenneras, C.,. Nesser, J. and Svennerholm, A. (1995) Binding of the fibrillar
CS3 adhesion of enterotoxigenic Escherichia cali prevented by GalNABl-Gal-
containing glycoconjugates. Infect. Immun. 63, 640-646

Willemsen, P. T. and Graaf, F. K. (1992) Age and serotype dependent binding of
K88 fimbriae to porcine intestinal receptors. Micro. Pathol. 12, 367-375

Wood, B. J. (ed.) (1992) The Lactic Acid Bacteria in Health and Disease.
Elsevier Applied Science, New York

Yamamoto, K., Miwa, T., Taniguchi, H., Shimamura, K., Tanka, T. and Kumagai,
H. (1996) Binding specificity of Lactobacillus to glycolipids. Biochem. Biophys.
Res. Commun. 228, 148-152

Yogeeswaran, G. (1983) Cell surface glycolipids and glycoproteins in malignant
transformation. Adv. Cancer Res. 38, 289-350

Yuyama, Y., Yoshimatsu, K., Ono, E., Saito, M. and Naiki, M. Postnatal change
of pig intestinal ganglioside bound by Escherichia coli with K99 fimbrae. J.
Biochem. 113, 488-492

Zychowicz, C., Kowalczyk, S. and Cieplinska, T. (1975) Results of

administration of Lactobacillus acidophilus culture (acidophilus milk) in an
endemic focus of dysentery. Pediatria Palska 50, 429-435

156