Characterization of group B streptococcal surface antigens delta, epsilon, and glyceraldehyde-3-phosphate dehydrogenase

MISSING IMAGE

Material Information

Title:
Characterization of group B streptococcal surface antigens delta, epsilon, and glyceraldehyde-3-phosphate dehydrogenase
Physical Description:
xiv, 163 leaves : ill. (some col.) ; 29 cm.
Language:
English
Creator:
Seifert, Kyle Nikoli, 1975-
Publication Date:

Subjects

Subjects / Keywords:
Streptococcus -- genetics   ( mesh )
Streptococcus -- isolation & purification   ( mesh )
Streptococcus -- pathogenicity   ( mesh )
Antigens -- isolation & purification   ( mesh )
Antigens -- genetics   ( mesh )
Glyceraldehyde-3-Phosphate Dehydrogenases -- isolation & purification   ( mesh )
Glyceraldehyde-3-Phosphate Dehydrogenases -- genetics   ( mesh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2004.
Bibliography:
Includes bibliographical references (leaves 142-162).
Statement of Responsibility:
by Kyle Seifert.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 56512421
ocm56512421
System ID:
AA00011189:00001

Full Text











CHARACTERIZATION OF GROUP B STREPTOCOCCAL SURFACE ANTIGENS
DELTA, EPSILON, AND GLYCERALDEHYDE-3-PHOSPHATE
DEHYDROGENASE














By

KYLE SEIFERT


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2004
































Copyright 2004

by

Kyle Seifert





















To my mother and father.














ACKNOWLEDGMENTS

I would like to thank past and present members of the Brady/Bleiweis lab for

making my time in the lab both challenging and enjoyable. I have been fortunate to work

with a number of bright, dedicated, and fun individuals. I would like to thank Nikki for

being my "lab buddy" and more importantly, my friend for the last 5 years. She has been

an inspiration with her tireless work ethic. I would like to thank both Dr. Bleiweis and

Dr. McArthur for their guidance, wisdom, and constant support. They have been

instrumental in both my personal and professional development during this process. I

would especially like to thank Jeannine for allowing me to join the lab and continue her

work. She showed patience with me in allowing me to pursue my own experiments. She

also provided optimism when progress on the project slowed, being a role model for what

a researcher and mentor should be. I would also like to thank the other two members of

my committee, Dr. Paul Gulig and Dr. Nancy Denslow for offering helpful suggestions

and showing tremendous interest in my project.

I would like to thank my parents, Tom and Kris, for allowing me to be the person I

am today. They have offered their constant support during my education, even though it

came with personal losses to both of them (having their son 1500 miles away). I owe

them both a tremendous debt of gratitude for the sacrifices they have made for my

individual achievements and personal well-being.

Finally, I would like to thank the two women in my life who make the good times

better and the bad times bearable: my wife, Rochelle; and my daughter, Allison. I wish to








express my appreciation to Rochelle for believing in me enough to move far away from

home to an unknown place. She has made numerous sacrifices so that I could complete

this project. Both Rochelle and Allison provide a constant reminder as to what things are

really important in my life. I share this accomplishment with them.















TABLE OF CONTENTS
page

ACKNOW LEDGM ENTS .............................................................................................. iv

LIST OF TABLES........................................................................................................ viii

LIST OF FIGURES ..................................................................................................... ix

KEY TO ABBREVIATIONS ......................................................................................... xi

ABSTRACT..................................................................................................................... xiii

CHAPTER

1 INTRODUCTION .....................................................................................................1...

Group B Streptococcal Disease.......... ....................................................................... 1
Group and Type Specific Antigens of Group B Streptococci ...................................3...
Other Surface Antigens of Group B Streptococci..................................................5...
Pathogenesis of Group B Streptococci....................................................................9...
Avoidance of Immune Clearance ............................................................................ 13
Vaccine Development............................................................................................... 15
Specific Aims ......................................................................................................... 19

2 ISOLATION AND CHARACTERIZATION OF THE 8 AND c ANTIGENS.........24

Introduction......................................................................................................... 24
M materials and M ethods ............................................................................................ 26
Results ......................................................................................................................34
Discussion............................................................................................................ 40

3 EVALUATION OF GENETIC COMPONENTS INVOLVED IN 8 AND e
EXPRESSION ......................................................................................................... 61

Introduction........................................................................................................ 61
M materials and M ethods ............................................................................................ 63
Results .............................................................................................................. 70
Discussion.......................................................................................................... 76









4 THE ASSOCIATION OF 8 AND e EXPRESSION WITH PATHOGENESIS OF
SERO TY PE III G BS ............................................................................................... 93

Introduction.......................................................................................................... 93
M materials and M ethods ............................................................................................ 94
R esults............................................................................................................... 96
D discussion ........................................................................... ...............................97

5 CHARACTERIZATION OF GROUP B STREPTOCOCCAL
GLYCERALDEHYDE-3-PHOSPHATE DEHYDROGENASE: SURFACE
LOCALIZATION, ENZYMATIC ACTIVITY, AND PROTEIN-PROTEIN
IN TERA CTION S .................................................................................................. 102

Introduction ......................................................................................................... 102
M materials and M ethods ........................................................................................... 104
R esults..................................................................................... .. ......................111
D iscussion............................................................... ............................................ 114

6 SUMMARY AND CONCLUSIONS .................................................................. ..... 125

Isolation and Characterization of the 8 and s Antigens.....................126
Determination of Genetic Components Involved in 6 and E Expression................131
The Association of 6 and e Expression in the Pathogenesis of Serotype III GBS ... 136
Characterization of Group B Streptococcal Glyceraldehyde-3-Phosphate
Dehydrogenase: Surface Localization, Enzymatic Activity, and Protein-Protein
Interactions.............................. .................................................... ........... 138
C onclusions..................... .................................................... .............................139

LIST OF REFERENCES ................................................................................................... 142

BIOGRAPHICAL SKETCH ....... .............................................................................. 163















LIST OF TABLES


Table page

1-1 Expression of 8/c antigens on serotype III group B streptococci isolated from septic
vs. colonized sources............................................................................................. 22

1-2 In vitro opsonophagocytic killing of group B streptococci by human peripheral
blood leukocytes.................................................................................................... 23

3-1. Prim ers used in this study...................................................................................... 83

3-2. Epsilon expression by group B streptococci (GBS) of RDP types 1-4.................89

4-1. Results of passive protection assays.....................................................................101














LIST OF FIGURES


Figure page

1-1. Identification of s reactivity on serotype III, 8-positive group B streptococci by
radioim m unoassay ................................................................................................. 20

1-2. Evaluation of anti-8e, anti-5, and anti-s reagents with group B streptococcal (GBS)
w hole cells by ELISA ................................. .......................................................... 21

2-1. Detection of 8/e in sonic extracts of group B streptococcal (GBS) whole cells by
com petition ELISA ................................. .............................................................. 48

2-2. Anion-exchange chromatography of GBS sonic extracts.....................................49

2-3. Detection of 8 and s in anion-exchange column fractions by competition ELISA..50

2-4. Evaluation of proteinase K sensitivity of the 8 antigen on A909 serotypee Ic/apy5)
w hole cells by ELISA ................................. .......................................................... 51

2-5. Evaluation of sodium metaperiodate sensitivity of the 8 antigen on A909 serotypee
Ic/apy8) whole cells by ELISA.............................................................................52

2-6. Evaluation of anti-86 antiserum reactivity with material contained in anion-
exchange column fractions from DL700 serotypee III) and J48 serotypee III/8s) ...53

2-7. Evaluation of anti-86 antiserum reactivity with sodium-dodecyl sulfate (SDS)-
extracted material from representative &e-surface positive and &e-surface negative
G B S strains...............................................................................................................54

2-8. Visualization of E by Western immunoblot...........................................................55

2-9. Evaluation of proteinase K sensitivity of the e antigen on J48 serotypee III/86)
w hole cells by ELISA ........................................................................................... 56

2-10. Evaluation of sodium metaperiodate sensitivity of the & antigen on J48 serotypee
III/8s) whole cells by ELISA ................................................................................ 57

2-11. Evaluation of sodium metaperiodate sensitivity of the e antigen assessed by
W western im m unoblot............................................................................................. 58









2-12. Western immunoblot analysis of high-molecular weight material gel-purified from
sonic extracts of DL700 (lane 1) and J48 (lane 2) ................................................59

2-13. Western immunoblot analysis of SDS-extracted material from group B
streptococcal (GBS) whole cells. .......................................................................... 60

3-1. Western immunoblot analysis of recombinant GBS acyl-carrier protein .............84

3-2. Polymerase Chain Reaction (PCR) amplification of srr DNA from GBS strains...85

3-3. Detection of srr by Southern hybridization...........................................................86

3-4. Ribonucleic Acid (RNA) dot blot to detect srr mRNA............................................87

3-5. Comparison between J48 serotypee III/8s) and DL700 serotypee III) of ORFs
contained within an operon including the srr variant common to Ue-negative GBS
strain s. ...................................................................................................................... 88

3-6. Detection of a variant srr ORF associated with serotype III/e GBS by Southern
hybridization.................................................................... ..... ............. ............... 90

3-7. Western immunoblot analysis of recombinant Srr from J48 serotypee III/86) ........91

3-8. Comparisons of srr-containing loci representative of non-RDP III-3 and RDP II1-3
G B S .................................................................................................... .................. ..92

4-1. Comparison of LD90 values of serotype III 8s-positive and 8s-negative strains ...100

5-1. Protein stain and Western immunoblot analysis of GBS GAPDH ........................119

5-2. Detection of surface-localized GBS GAPDH by ELISA using an anti-Plr
m onoclonal antibody................................................................. ........................120

5-3. Evaluation of GAPDH enzymatic activity associated with GBS whole cells........121

5-4. Evaluation of GBS gapC copy number by Southern hybridization...................... 122

5-5. Western immunoblot analysis to detect binding of extracellular matrix and
cytoskeletal proteins to immobilized GBS GAPDH ............................. ..........123

5-6. ELISA to detect binding of GBS GAPDH to immobilized extracellular matrix and
cytoskeletal proteins............................................................................ ... .......... ......124












KEY TO ABBREVIATIONS

cpm counts per minute

Ci Curie

CPS capsular polysaccharide

CsC1 cesium chloride

DNA deoxyribonucleic acid

EDTA ethylenediamine-tetraacetate

ELISA enzyme linked-immunosorbent assay

g gravity

GAPDH glyceraldehyde-3-phosphate dehydrogenase

GAS group A streptococci

GBS group B streptococci

h hours

IgA immunoglobulin A

IgG immunoglobulin G

kDa kilodaltons

kb kilobase

min minutes

Mr relative molecular weight

NADH nicotinamide adenine dinucleotide, reduced form

nm nanometer

nt nucleotide









OPD o-phenylene-diamine

PAGE polyacrylamide gel electrophoresis

PBS 0.15 M phosphate buffered saline, pH 7.4

PCR polymerase chain reaction

s seconds

Tris tris(hydroxymethyl)aminomethane

TT tetanus toxoid

SDS sodium dodecyl sulfate















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

PARTIAL CHARACTERIZATION OF GROUP B STREPTOCOCCAL SURFACE
ANTIGENS DELTA, EPSILON, AND GLYCERALDEHYDE-3-PHOSPHATE
DEHYDROGENASE

By

Kyle Seifert

May, 2004

Chair: L. Jeannine Brady
Major Department: Oral Biology

Group B streptococci (GBS) are an important cause of disease in both neonates and

immunocompromised adults. Of the nine serotypes identified, serotype III GBS were

found in over 90% of isolates from cases of meningitis. Previous studies show, that in

addition to the group and serotype polysaccharides, proteins are also found at the cell

surface. Two novel antigenic reactivities, 8 and e, are co-expressed on a subset of

serotype III GBS, and are present more frequently on isolates from septic vs. colonized

sources. This study demonstrates preliminary characterization of the 8 and s antigens.

The 8 antigen is destroyed by proteolysis and is associated with one of the three known

GBS acyl-carrier proteins. The s antigen is a high molecular weight glycoprotein with

similarity to serine rich repeat (Srr) proteins such as GspB and Hsa of Streptococcus

gordonii and Fapl of Streptococcus parasanguis. Restriction digest pattern (RDP) typing

has identified a more pathogenic subset of serotype III GBS, designated RDP III-3. All









RDP III-3 (but not RDP III-1, 2, or 4) isolates express the s antigen. The srr gene from

RDP III-3, 8s-positive GBS differs from srr from all non-III-3 strains. Anti-E antiserum

binds to the recombinant Srr from a RDP III-3 strain, suggesting that the Srr protein is the

antigenic backbone of the E glycoprotein. The srr gene in both III-3 and non-III-3 strains

are contained within loci including accessory secA and secY homologues. The

organization of the srr-containing loci is different for RDP III-3 and non-III-3 GBS

strains. In a mouse model of neonatal sepsis, serotype III/86-positive strains demonstrated

LD9o values 2-4 logs lower than serotype III/8c-negative strains. Results of this study

indicate both antigenic and genetic differences that distinguish a more pathogenic subset

of serotype III GBS and may lead to improved diagnostic capabilities or improved

vaccine candidates in an effort to reduce the incidence of GBS disease.














CHAPTER 1
INTRODUCTION

Group B Streptococcal Disease

Lancefield group B streptococci (GBS) are a major cause of neonatal sepsis (Baker

and Edwards 1994) and are important pathogens in immunocompromised adults (Farley

2001). There are approximately 19,000 cases of invasive GBS disease per year in the

United States, resulting in 1,900 deaths (CDC 2002). There are two distinct syndromes of

neonatal GBS disease: early- and late-onset. Early-onset disease occurs within the first

week of life and is vertically transmitted from mother to infant. It is assumed that the

main route of infection is aspiration of GBS during parturition, resulting in subsequent

colonization of the respiratory epithelium (Baker and Edwards 1994). All known GBS

serotypes have been isolated from infants suffering from early-onset disease; however,

serotypes la, Ib, II, III, and V are more commonly isolated in the United States (Harrison

et al. 1998). Three manifestations of early-onset disease include sepsis with no known

focus of infection, meningitis, and pneumonia. Before the Centers for Disease Control

and Prevention (CDC) established guidelines for the treatment and prevention of GBS

infections in 1996 (CDC 1996), 80% of all GBS infections were early-onset (Baker and

Edwards 1994). Intrapartum prophylaxis has reduced the percentage of early-onset

disease to 50% of total GBS disease (CDC 1997).

Important risk factors for early-onset disease include maternal carriage of GBS

(Boyer et al. 1983), having a previous infant with GBS disease, having a heavily

colonized mother (Gerards et al. 1985, Lim et al. 1982), or having a mother with a low









level of antibodies to the capsular polysaccharide (Baker et al. 1977). There are also a

number of obstetric risk factors, which include premature rupture of membranes

(Schuchat et al. 1994), prolonged labor (Schuchat et al. 1994), intrapartum fever (Boyer

et al. 1983, Schuchat et al. 1994), and urinary tract infection during pregnancy (Liston et

al. 1979, Schuchat et al. 1994).

Late-onset disease occurs from 1 week to 3 months after birth. Serotype III GBS

were responsible for approximately 90% of all late-onset infections (Baker and Edwards

1994), with meningitis being a common clinical manifestation that results in a substantial

number of survivors suffering permanent neurological damage. The acquisition of GBS

resulting in late-onset disease is less well understood. It is speculated that the organism

can be acquired during passage through the birth canal similar to early-onset disease, as

50% of GBS isolated from late-onset infections are the same serotype as the colonized

mother (Anthony et al. 1979, Dillon et al. 1987). Nosocomial and community sources are

also believed to contribute to late-onset infections (Trager et al. 1996). Intrapartum

prophylaxis has been successful in reducing the incidence of early-onset disease, but the

incidence of late-onset disease has remained unchanged (Schrag et al. 2000).

Although invasive GBS disease occurs in healthy adults, most disease in non-

pregnant adults occurs in those with underlying medical conditions such as diabetes,

alcoholism, and cancer (Farley et al. 1993, Farley 2001, Schuchat 1998). Clinical

manifestations are numerous and varied, and the mode of transmission is unknown. Skin

and soft tissue infections are the most frequently reported syndromes (Farley 2001).

Infected foot ulcers are common in patients with diabetes. Breast cellulitus has also been

reported in instances of breast conservation therapy for stage I or II breast cancer (Mertz









et al. 1998). Up to 23% of invasive GBS disease in adults presents as a urinary tract

infection (Farley et al. 1993, Munoz et al. 1997, Schwartz et al. 1991, Trivalle et al. 1998,

Tyrrell et al. 1996, Verghese et al. 1986). Meningitis can occur, but is not common in

adults (Farley et al. 1993). Other manifestations also include pneumonia, endocarditis

(Bayer et al. 1976, Farley et al. 1991, Gallagher and Watanakunakom 1985, Lerner et al.

1977), bone and joint infections (Garcia-Lechuz et al. 1999), and pyogenic arthritis

(Nolla et al. 2003). The most common serotypes associated with adult disease are Ia, III,

and V, in approximately equal proportion with minor geographic variations (Tyrrell et al.

2000).

GBS are sensitive to treatment with penicillin G and ampicillin (Baker et al. 1981,

Bayer et al. 1976). Clindamycin and erythromycin are also used clinically, although

resistance to these antibiotics is increasing (Pearlman et al. 1998). Aminoglycosides have

little activity when used alone, but may provide a synergistic effect when combined with

ampicillin or penicillin G (Baker et al. 1981).

Group and Type Specific Antigens of Group B Streptococci

Lancefield (1933, 1934, 1938) devised a system of categorizing streptococci based

on hot-acid extracted surface polysaccharides, called group antigens. The group B

carbohydrate of GBS is comprised of four sugars: rhamnose, N-acetylglucosamine,

galactose, and glucitol (Carey et al. 1980, Heidelberger et al. 1967, Kane and Karakawa

1977, Michon et al. 1987, 1988, Pritchard et al. 1981,1984), in a complex, highly

branched arrangement (Michon et al. 1991). It has been known for some time that the

group B polysaccharide is in close proximity to the cell wall (Wagner et al. 1980) and it

was recently suggested that the group B antigen is covalently linked to N-acetylmuramic

acid residues of the disaccharide repeating unit of cell wall peptidoglycan (Deng et al.









2000). Expression of the group B carbohydrate is not regulated by growth rate (Ross et

al. 1999). In mice, antibodies against the group B carbohydrate are not protective against

lethal challenge of GBS (Lancefield et al. 1975). Antibodies to the group B carbohydrate

do not appear to contribute to protective immunity against GBS in humans either

(Anthony et al. 1985).

With few exceptions, all GBS strains are encapsulated. Four carbohydrates

(glucose, galactose, N-acetylglucosamine, and sialic acid) comprise the GBS

polysaccharide capsule with the exception of serotypes VI and VIII, which lack

N-acetylglucosamine (Kogan et al. 1996, von Hunolstein et al. 1993). GBS were

subdivided into serotypes based on the linkages of the sugars that comprise the capsule.

To date, nine serotypes have been identified: la, Ib, II, III, IV, V, VI, VII, and VIII; and

the structures of the capsular polysaccaharides of all have been determined (DiFabio et al.

1989, Hunolstein et al. 1993, Jennings et al. 1983a, 1983b, Kogan et al. 1994, 1995,

1996, Wessels et al. 1987, 1991).

Investigators have suggested that the capsular polysaccharide is linked to

peptidoglycan (DeCueninck et al. 1982, Yeung and Mattingly 1986). Analysis of digested

insoluble cell-wall fragments from lysates of a serotype III GBS suggests that serotype

polysaccharide is linked via a phosphodiester bond and an oligosaccharide linker to the

N-acetylglucosamine repeating unit of peptidoglycan, and its linkage to peptidoglycan is

independent of the group B carbohydrate (Deng et al. 2000).

The capsular polysaccharide appears to be a major factor associated with virulence

of GBS. Acapsular and asialo mutants are more susceptible to opsonophagocytosis in the

presence of complement and PMNs than their encapsulated parent strains (Marques et al.









1992, Rubens et al. 1987, Wessels et al. 1989). The serotype III capsular polysaccharide

is one of the three major capsular types associated with invasive neonatal infection and

the most common serotype isolated from cases of meningitis (Dillon et al. 1987, Wenger

et al. 1990). The importance of the terminal sialic acid residue of the serotype III

polysaccharide has been documented. An asialo GBS mutant was less virulent than its

fully encapsulated parent in opsonophagocytosis assays (Wessels et al. 1989). Sialic acid

was also shown to contribute to complement activation and generation of a specific

antibody response (Markham et al. 1982).

Other Surface Antigens of Group B Streptococci

In addition to the group B carbohydrate and serotype-specific polysaccharides,

numerous protein antigens can be associated with the GBS cell surface. Several surface

proteins from group B streptococci have been identified in an attempt to characterize the

complex nature of the cell wall and polysaccharide capsule. Wilkinson and Eagon (1971)

first described the c-protein in 1971 using a polyclonal rabbit antiserum made against

whole formalin-killed cells of CDC strain A909 serotypee la). Antibodies against the

group- and serotype-specific carbohydrates in the polyclonal rabbit antiserum were

eliminated by exhaustive adsorption with a c-protein-negative serotype la strain, and the

non-serotype Ia antigens that the antiserum bound were called the c-protein. It was later

recognized that c-protein typing serum actually recognizes at least four antigenic

moieties: a, 0, (Bevanger 1985, Johnson and Ferrerri 1984, Wilkinson and Eagon 1971)

y, and 8 (Brady et al. 1988, Chun 1991). The genes encoding the a and P proteins have

been cloned, (Cleat and Timms 1987, Michel et al. 1991), and the proteins characterized.









The a protein is a large, highly repetitive protein resistant to cleavage by trypsin

(Wilkinson and Eagon 1971). This protein contains nine long tandem repeats of 82 amino

acids each. These repeats constitute >75% of the total size of the protein, and the protein

itself has a typical wall-anchoring LPXTG motif (Michel et al. 1992). Deletion of tandem

repeats is associated with escape from protective immunity, increased pathogenicity, and

vertical transmission in mice (Gravekamp et al. 1998, Madoff et al. 1992, 1996). The 3

antigen has been characterized as a human IgA-binding protein (Russel-Jones et al.

1984). The y reactivity has been found to correspond to a variant amino-terminus of the a

protein (Madoff, personal communication). The nature of the 8 antigen has been partially

determined and is described below. In addition to antigens reactive with c-protein typing

antiserum, additional GBS protein antigens include protein Rib (Stalhammar-Carlemalm

et al. 1993), the R and X proteins (Ferrieri 1988), C5a peptidase (Beckmann et al. 2002,

Cleary et al. 1992), surface immunogenic protein (Sip) (Brodeur et al. 2000), laminin-

binding protein (Lmb) (Spellerberg et al. 1999), Fbs, and a Rib-like protein (Areschoug et

al. 1999).

The a, P, and y antigens are expressed primarily by serotype la, Ib, and II

organisms, while 8 and Rib have been reported to be expressed by most serotype III

strains (Brady et al. 1988, Chun, 1991, Stalhammar-Carlemalm et al. 1993). Because of

the higher association of 6 antigen expression with serotype III GBS isolates from septic

vs. colonizing sources (Brady et al. 1996), a polyclonal rabbit antiserum against J48

serotypee III/5e) whole cells was generated to facilitate further study of this marker. The

anti-J48 antiserum was rendered more specific for the previously described 8 antigen by

adsorption with GBS strain DL700 serotypee III) whole cells to remove antibodies









against the serotype III polysaccharide and group B carbohydrate, as well as any other

common surface determinants. The adsorbed polyclonal rabbit antiserum (Brady et al.

1988) was subsequently absorbed with GBS strain A909 serotypee Ic/apy8) to remove

antibodies against 8. The DL700- and DL700/A909-adsorbed reagents were then used to

evaluate reactivity of a panel of known 8-positive and 8-negative GBS strains by

radioimmunoassay (Figure 1-1). The DL700-adsorbed reagent reacted with A909

serotypee Ic/a3py8) and serotype III, 8-positive strains; whereas the DL700/A909-

adsorbed reagent antiserum was not reactive with strain A909 but retained reactivity with

serotype III, 8-positive strains. These results indicated that another immunological

reactivity absent from strain A909 serotypee Ic/apy8) was associated with the 8 antigen

on serotype III strains. The newly identified antigen was designated e. Thirty serotype

III/8 strains tested all expressed c. None expressed e without 8, and vice versa. Polyclonal

reagents reactive with both 8 and e antigens (anti-&e), or with each of antigen individually

(anti-8 or anti-e) can be generated by selective adsorption of polyclonal rabbit antisera as

described (Brady et al. 1988, 1996). For the work described below, a 8-reactive reagent

was generated by adsorption of anti-c protein typing antiserum with strain ss618

serotypee Ib/apy) to remove anti-a, P, and y antibodies. A 6g-reactive reagent was

generated by adsorption of an anti-J48 whole cell antiserum with strain DL700, and an e-

reactive reagent was generated by adsorption of the anti-J48 antiserum with strain DL700

and A909. The reactivity of these reagents with strains of GBS including those used for

immunizations and adsorptions is shown in Figure 1-2.

In addition to the polysaccharide capsule, serotype III GBS isolates can be further

classified into four distinct lineages based on HindIII and Sse83871 restriction digest









patterns (RDPs) of chromosomal. These have been designated RDP types III-1, 2, 3, and

4 (Takahashi et al. 1998). Most (91%) serotype III invasive neonatal isolates have been

reported to be RDP III-3, suggesting that these strains may represent a more virulent

lineage of serotype III GBS.

Other methods for subtyping GBS have also been reported and include multilocus

enzyme electrophoresis (MLEE) (Quentin et al. 1995), ribotyping (Chatellier et al. 1996),

randomly amplified polymorphic DNA (RAPD) (Chatellier et al. 1997), and most

recently by pulsed-field gel electrophoresis of SmaI-digested chromosomal DNA

(Rolland et al. 1999). None of these methods subdivide GBS strains into

pathogenic/nonpathogenic groups.

Pulsed-field gel electrophoresis of 10 mother/child isolate pairs (Melchers et al.

2003) supports the findings of other investigators (Chatellier et al. 1996, 1997, Quentin et

al. 1995, Rolland et al. 1999) that the heterogeneity within GBS isolates makes it difficult

to definitively identify virulent strains based on entire chromosomal DNA. Capsular

polysaccharide genes (Kong et al. 2002a), surface protein genes (Kong et al. 2002b), and

mobile genetic elements (Kong et al. 2003) have been used in combination to

characterize 224 GBS isolates. Such a combinatorial approach identified 56 different

genotypes and eight genetic clusters, again illustrating the heterogeneity of GBS isolates.

A multilocus sequence typing system (MLST) using seven housekeeping genes has also

been developed, and has identified a sequence type (ST-17) that correlates with neonatal

invasive disease, much like that of RDP III-3 strains (Jones et al. 2003).

The RDP III-3 data reported by Takahashi et al. (1998) showed that while 33% of

vaginal isolates were III-3, 91% of invasive neonatal isolates were III-3. These data were









reminiscent of an epidemiological study by Brady et al. (1996), in which 85/

co-expression was detected on 83.3% of serotype III GBS from septic sources, compared

to 58.8% of serotype III isolates from colonized sources (Table 1-1). Taken together,

these results suggested that the 8/, antigenic markers and RDP III-3 genetic marker might

identify the same highly pathogenic lineage of serotype III GBS.

Pathogenesis of Group B Streptococci

GBS are associated with asymptomatic colonization in healthy adults. However, in

susceptible individuals, GBS can be a devastating pathogen. Pathogenesis is

multi-factoral, and is generally considered to follow a progression that begins with

adhesion to host epithelial cells, invasion of tissue, direct injury to host tissue, and finally

induction of sepsis (Fischetti et al. 2000). In addition to affecting the host directly, a

major component of GBS pathogenesis is avoiding host immune mechanisms.

The initial step in GBS disease of neonates is asympomatic colonization of the

female urogenital tract. As expected, GBS bind efficiently to human vaginal cells

(Zawaneh et al. 1979). GBS can also adhere to placental membranes, alveolar epithelium

and endothelium, and pharyngeal mucosa, which could be relevant to vertical

transmission and induction of disease in a newborn infant (Fischetti et al. 2000). The

mechanism of adherence to host tissues has not been completely elucidated. However a

number of potential adhesins and GBS factors have been shown to affect GBS binding to

host cells (Tamura et al. 1994, Wibawan et al. 1992).

After initial adherence, GBS must penetrate and survive inside different host

tissues. GBS have been shown to migrate through freshly isolated chorioamniotic

membranes (Galask et al. 1984), as well as isolated chorion cells from human









cesarean-section placentas (Winram et al. 1998). Invasion of these cell types was

necessary to cause amniocentesis and exposure of the fetus to GBS. Interestingly,

although able to invade chorion cells, GBS were unable to invade amnion cells under a

number of assay conditions (Winram et al. 1998), suggesting that the amnion may

provide a formidable host barrier to infection.

In most instances, GBS must traverse the alveolar epithelium, the pulmonary

interstitium, and the pulmonary endothelium to gain access to the systemic circulation

and cause bacteremia and sepsis after aspiration of infected amniotic fluid. Studies using

a primate model of infection have shown GBS within alveolar epithelial cells, interstitial

fibroblasts, and capillary endothelial cells (Gibson et al. 1993, Rubens et al. 1991).

Tissue-culture studies have also shown invasion of human lung epithelial cells (Rubens et

al. 1992), as well as human umbilical vein and piglet endothelial cells (Gibson et al.

1993). Quantitative tissue-culture studies have shown that GBS isolates of serotypes la,

Ib, Ic, II, and III are capable of invading alveolar epithelial cells with a small degree of

strain-to-strain variation (Hulse et al. 1993, Rubens et al. 1992). Interestingly, the

capsular polysaccharide appears not to be necessary for invasion of respiratory epithelial

cells (Hulse et al. 1993) and attenuates invasion of umbilical vein endothelial cells

(Gibson et al. 1993) and alveolar epithelium (Tamura et al. 1994), suggesting that factors

other than the capsular polysaccharide are responsible for invasion of host tissue. In vivo

and in vitro studies have shown that GBS were able to invade and transverse the host

cells necessary to gain access to the systemic circulation of an infected host.

After gaining access to the circulatory system of the host, GBS must bind and

invade brain microvascular endothelial cells, which constitute the blood-brain barrier, in









order to cause meningitis. GBS have been shown to invade brain microvascular cells in

tissue culture (Nizet et al. 1997). Serotype III GBS were better at invading brain

microvascular cells than other serotypes, suggesting this as at least one reason serotype

III cause >90% of GBS meningitis cases. The same study suggested that GBS were able

to directly injure the blood-brain barrier, although the exact mechanism of such damage

is not fully understood.

Direct injury of host tissue by GBS appears to contribute to pathogenesis. One of

the clinical presentations of early-onset disease is severe pneumonia. GBS was shown to

cause extensive pulmonary damage in pneumonia (Vollman et al. 1976). A hemolysin of

GBS has cytopathic effects on pulmonary epithelial cells (Gibson et al. 1999, Nizet et al.

1996), suggesting that the hemolysin of GBS contributes to pneumonia. Degradation of

host-cell hyaluronic acid by hyaluronate lyase (Lin et al. 1994) may also play a role in

GBS pathogenesis, although the exact biological role is uncertain. GBS were able to

degrade collagen (Jackson et al. 1994), which may be important for penetration of the

chorioamnion. Lipoteichoic acid (LTA) has been shown to be cytotoxic for a variety of

human cell monolayers in tissue culture, including human embryonic brain cells and

human embryonic amnion cells (Miyazaki et al. 1988), although LTAs role in vivo is

uncertain. It has also been suggested that unidentified proteases are able to degrade

placental tissue (Schoonmaker et al. 1989).

If GBS are able to avoid host immune clearance mechanisms and establish

bacteremia, sepsis may be the consequence. Sepsis is characterized by extensive

hemodynamic alterations leading to a decreased cardiac output, acidosis, and finally,

multiple organ failure. Recent work using animal models of neonatal infection has









improved the understanding of how GBS activate host inflammatory mediators and

contribute to sepsis and circulatory shock. In animals, intravenous injection of GBS leads

to acute (<1 h) increases in pulmonary artery pressure and decreased arterial oxygenation

(Gibson et al. 1992). Pulmonary hypertension and hypoxemia persist through the late

phase of sepsis (2-4 h). A role of the inflammatory mediator TNF-a is suggested but not

established. In septic individuals, there is increased TNF-a in the bloodstream, and is

associated with edema, neutropenia, collapse of blood vessels, and eventually, multiple

organ failure. TNF-c can be detected in the blood, urine, and cerebral spinal fluid in

infants with invasive disease (Williams et al. 1993). However, anti-TNF-a antibodies do

not affect GBS mortality in mice or rats (Teti et al. 1992, 1993). Cell wall preparations of

GBS cause TNF-a release from human monocytes (Medvedev et al. 1998). The group B

carbohydrate and peptidoglycan have both been reported to be better than either LTA or

capsular polysaccharide at stimulating TNF-ca release from monocytes (Vallejo et al.

1996). Serotype III capsular polysaccharide was reported to have no effect on TNF-a

production by human mononuclear cells in vitro (Williams et al. 1993). Activation of

nitrous oxide by GBS hemolysin has been reported to contribute to sepsis and

hypotensive shock (Griffiths and Rhee 1992, Ring et al 1998). The hemolysin of GBS has

recently been identified within a cluster of at least 10 open reading frames in a single

operon (Spellerberg et al. 1999).

Antibodies and complement are both necessary for immune clearance of GBS.

Antibodies directed against the capsular polysaccharide, but not the group B

carbohydrate, have been shown to be protective in a neonatal mouse model of infection

(Lancefield 1934, 1938). Numerous studies have been conducted in an attempt to identify









antigens in addition to the capsular polysaccharide capable of eliciting opsonic

antibodies. Lancefield et al. (1975) showed protection by rabbit antiserum against

unidentified "c-protein" antigens independent of the polysaccharide capsule of GBS.

Madoff et al. (1991) showed that monoclonal antibodies recognizing the Ca antigen are

protective in mice. It has since been shown that antibodies to both the amino-terminal

domain and the repeat region of the a protein are protective in an animal model of GBS

infection and mediate opsonophagocytic killing in vitro (Kling et al. 1997). Antibodies

against the 13 antigen are also protective in mice (Fusco et al. 1997, Madoff et al. 1992,

Michel et al. 1991). Antibodies to protein Rib confer protective immunity (Stalhammar-

Carlemalm et al. 1993), as does antiserum recognizing an a-like protein present on

serotype V GBS (Lachenauer and Madoff 1996). These studies suggest that antibodies to

surface structures in addition to the capsular polysaccharide are important targets of

protective immunity.

Avoidance of Immune Clearance

GBS have many ways to resist opsonophagocytosis and survive inside

macrophages. For example, in serotype III GBS, resistance to opsonophagocytosis is

proportional to the sialic acid content (Edwards et al. 1982, Marques et al. 1992). RDP

III-3 strains have significantly more capsular sialic acid content than either III-1 or III-2,

suggesting that high sialic content may contribute to their increased virulence (Takahashi

et al. 1999). However, GBS were able to avoid killing after phagocytosis by monocyte-

derived macrophages, even in the presence of opsonic antibody and complement

activation (Marodi et al 2000). In addition, using signature-tagged mutagenesis (STM),

Jones et al. (2003a) identified a mutation in a gene encoding a penicillin-binding protein









(ponA) resulted in decreased virulence in a neonatal rat model of infection and decreased

resistance to phagocytosis, suggesting a role for ponA in the pathogenesis of GBS. Jones

et al. (2003b) identified avirulent mutant with a transposon insertion in rpoE, encoding

the delta subunit of RNA polymerase using STM. This mutation affected the ability of

GBS to inhibit the alternate pathway of complement activation, rendering it more

susceptible to phagocytic killing. Poyart et al. (2001) identified a Mn-cofactored

superoxide dismutase (SodA) as a virulence factor of GBS. SodA mutants were less

virulent in a mouse model of infection and more susceptible to bacterial killing by

macrophages than wild-type strains. As expected, sodA mutants were more susceptible to

oxidative stress generated by addition of hydrogen peroxide in culture medium. Survival

of oxidative stress following engulfment by professional phagocytes would represent

another GBS virulence mechanism involved in surviving the host immune response.

GBS have been shown to rapidly inactivate the complement component C5a by the

enzyme C5a-ase (Bohnsack et al. 1991, Hill et al. 1988). Inactivation of C5a reduces

PMN recruitment to sites of inflammation and stimulation of PMN phagocytosis, both

important factors in immune clearance of GBS (Bohnsack et al. 1997). The work by

many investigators suggests that like GBS pathogenesis itself, avoidance of immune

clearance is a multi-factoral system. Mutations in seemingly unrelated genes (such as

ponA, rpoE, and sodA) resulted in increased susceptibility to phagocytic killing in rat or

mouse models of infection (Jones et al. 2003a, 2003b, Poyart et al. 2001), illustrating that

surviving the host immune system contributes to GBS pathogenesis. Opsonophagocytosis

assays conducted by Brady (1996) suggested that 6/E-expressing serotype III GBS may

be more resistant to intracellular killing by PMNs than 8/e-negative strains. GBS strain









DL700 serotypee III) was effectively killed following incubation with anti-J48 serotypee

III/8 ) whole-cell polyclonal rabbit antiserum, while strain J48 serotypee III/8) itself was

not killed in the presence of the same antiserum (Table 1-2). GBS can also survive inside

of macrophages and interfere with the bactericidal function of macrophages.

Cornacchione et al. (1998) showed that a serotype III strain was able to survive inside

murine peritoneal macrophages for up to 48 h and impaired the protein kinase C-

dependent transduction pathway. A serotype III strain was also able to survive in a mouse

macrophage cell line, and survival was enhanced when entry was opsonin-independent

(Valentin-Weigand et al. 1996). Serotype III GBS were also shown to induce apoptosis in

murine macrophages (Fettucciari et al. 2000), possibly delaying or hindering the

development of a specific immune response.

Vaccine Development

Intrapartum prophylaxis has decreased the rate of early-onset, but not late-onset

GBS disease in neonates (Schrag et al. 2000). However, the most effective long-term

solution for GBS disease, both early- and late-onset, likely lies with the development of

an effective vaccine. An ideal vaccine would be protective against all nine serotypes of

GBS, but is currently unavailable. Initial work by Lancefield (1934, 1938) showed that

protection against GBS infection could be achieved in mice using polyclonal rabbit

antiserum made against capsular polysaccharide (CPS), but not group B carbohydrate.

The first experimental vaccine used in a phase I clinical trial was purified serotype

III CPS (Baker 1980). Healthy adults safely received the GBS CPS vaccine at a dose

range of 10 to 150 jg (Baker et al. 1978, Baker and Kasper 1985, Fischer et al. 1983,

Kasper et al. 1983). The safe use of the GBS CPS vaccine in healthy adults led to its









administration to pregnant women at a mean gestation of 31 weeks (Baker et al. 1988).

This vaccine was shown to be safe in pregnant women, and there was a direct correlation

between maternal and cord blood levels ofCPS-specific IgG (Baker et al. 1988).

Unfortunately, only 60% of vaccine recipients mounted a specific anti-serotype III GBS

CPS response. This result led to attempts to increase the immunogenicity of the CPS.

Coupling of a bacterial CPS to a carrier protein can engage T-cell help and improve

immunogenicity, as has been shown with Haemophilus influenza type b CPS (Englund et

al. 1993). When tested in animals, the first conjugate vaccines of GBS serotype III CPS

coupled to tetanus toxoid (TT) elicited opsonically active anti-CPS IgG at levels higher

than those achieved by immunization with CPS alone (Lagergard et al. 1990, Wessels et

al. 1990). The success of the serotype III CPS-TT conjugate vaccine led to development

and preclinical testing of conjugate vaccines with of nine GBS CPSs (Paoletti and Kasper

2002, Paoletti et al. 1992, 1994, Wessels et al. 1990, 1993, 1995). Work in both mice and

baboons has also shown that maternal vaccination with GBS glycoconjugate vaccines

induces high levels of functionally active IgG that crossed the placenta (Madoff et al.

1994, Paoletti et al. 1996, 2000). Maternal administration of a CPS-TT conjugate vaccine

resulted in increased in survival of pups challenged with a lethal dose of

serotype-matched GBS compared to pups born to mothers immunized with uncoupled

CPS or carrier protein only (Madoff et al. 1994, Paoletti et al. 2000).

This promising work in animal trials led to numerous clinical trials of GBS type III

conjugate vaccine, the first with purified type III CPS and monomeric TT (III-TT)

(Kasper et al. 1996). The vaccine was administered to nonpregnant women of

childbearing age at one of three dosage levels. Controls included age-matched women









immunized with either uncoupled CPS or saline. The vaccination was well tolerated,

serum levels of serotype III CPS-specific were dose-dependent, and elevated levels of

specific antibody persisted for 26 weeks. In mice and baboons, the administration of alum

adjuvant in combination with GBS conjugate vaccines improved the immunogenicity of

the vaccine (Guttormsen et al. 1998, Paoletti et al. 1996). Studies have subsequently been

conducted in healthy human adults with III-TT adsorbed to aluminum hydroxide gel, but

alum did not improve the immune response to the III-TT conjugate vaccine in humans

(Paoletti et al. 2001).

Phase 1 and 2 clinical trials have also been conducted with CPS conjugate vaccines

prepared with serotypes la (Baker et al. 1999), Ib (Baker et al. 1999), II (Baker et al.

2000), and V (Paoletti et al. 1998). To determine if a multivalent vaccine could protect

against multiple serotypes, a trial was conducted to evaluate the immune response against

two different GBS CPS conjugate vaccines administered in combination (Baker et al.

2003). When GBS serotype II-TT and serotype III-TT were administered together, the

immune response to each individual CPS was similar to that achieved using the

monovalent vaccines administered singly at the same dose. The bivalent vaccine was well

tolerated, and no immune interference was noted from combining the two antigens. More

work on the effectiveness of multivalent vaccines needs to be conducted. In the United

States, such a strategy would need to include the frequently isolated serotypes (Ia, Ib, II,

III, and V), while in Japan, serotypes VI and VIII would need to be included as well

(Lachenauer et al. 1999, Paoletti et al. 1999).

Since GBS is a neonatal pathogen, an effective vaccine would need to elicit

antibodies that cross the placenta. After showing that administration of a serotype III-TT









conjugate vaccine was safe and effective in healthy women (Kasper et al. 1996), pregnant

women were immunized with III-TT (Baker et al. 2001, 2003). The vaccine was safe and

well tolerated in these patients, with healthy babies born to each of the vaccinated

women. Type III CPS-specific IgG was obtained from maternal serum and cord blood

and indicated placental transfer of specific antibody. Importantly, the specific anti-III

CPS antisera obtained from 1- and 2-month-old infants were functionally active in

opsonophagocytosis assays.

In addition to CPS, several surface proteins have been explored as vaccine

candidates and as potential carrier proteins. Immunogenic GBS surface proteins in

various phases of development include the a antigen (Gravekamp et al. 1999, Madoff et

al. 1994), the P antigen (Gravekamp et al. 1999, Madoff et al. 1994), Rib (Larsson et al.

1996), Sip (Brodeur et al. 2000), and C5a peptidase (Cheng et al. 2001, 2002). Other

surface proteins include glutamine synthetase (Suvarov et al. 1997), a-enolase (Pancholi

and Fischetti 1998), and Hsp70 (Hammel et al. 1996). Hughes et al. (2002) have also

identified major surface proteins, including novel protein vaccine candidates using

alkaline phosphatase fusions (Hughes et al. 2003). Both the a and p proteins have been

tested for their effectiveness as a carrier protein in a CPS conjugate vaccine with

promising results (Gravekamp et al. 1999, Madoff et al. 1994). The a, P, and Rib proteins

are each associated with given serotypes. The a and p proteins are more commonly

expressed by serotypes la, Ib, and II, while Rib and the 6 and E antigens are more

commonly associated with serotype III GBS. Therefore, immunization with these

antigens would not confer protection against all serotypes. A desirable, but as yet elusive

alternative would be to identify an antigen capable of eliciting a more broadly protective









immune response. Although no GBS vaccine is yet available, research suggests that the

development of a multivalent GBS vaccine may reduce neonatal GBS disease.

Specific Aims

Initial epidemiology and opsonophagocytosis studies have suggested that the GBS

8 and s antigens are markers for a more pathogenic lineage of serotype III, and may

contribute to the virulence of this subset of serotype III GBS (Brady et al. 1996). The first

specific aim of this study was to isolate and characterize the antigens recognized by 6-

and e-specific antibodies. The second specific aim of this study was to determine genetic

components responsible for 8 and e expression. Lastly, the third specific aim of this study

was to evaluate the potential role of the 6 and E antigens in the pathogenesis of serotype

III GBS. In addition to the work characterizing the 8 and e antigens, a surface-localized

glyceraldehyde-3-phosphate dehydrogenase was identified and characterized.

















2 3 4


5 6 7


GBS strains
Figure 1-1. Identification of s reactivity on serotype III, 8-positive group B streptococci
by radioimmunoassay. GBS whole cells were incubated with anti-J48
serotypee III/6) antiserum adsorbed with DL700 serotypee III) (hatched bars)
or anti-J48 antiserum adsorbed with DL700 and A909 (solid bars). Results are
given as cpm/bacterial pellet. Column 1, A909 serotypee Ic/ap3y6), column 2,
J48 serotypee III/8), column 3, PEH serotypee III/8), column 4, DL797
serotypee III/), column 5, DL769 serotypee III/8), column 6, DL771
serotypee III/), column 7, DL686 serotypee III), column 8, DL775 serotypee
III), column 9, DL700 serotypee III).


25000


20000 -
15000 -
10000 -
5000 -
0


i




0


8 9


_I










0.45

0.4 -

0.35

0.3

I 0.25

0 0.2

0.15

0.1

0.05



A909 J48 DL700 ss617 ss618
serotypee serotypee serotypee III) serotypee Ia) serotypee
Ic/aPy6) IIHE/) Ib/apy)



Figure 1-2. Evaluation of anti-8&, anti-8, and anti-s reagents with group B streptococcal
(GBS) whole cells by ELISA. Solid black bars represent reactivity of anti-&8
antiserum. Hatched bars represent reactivity of anti-5 antiserum. Solid white
bars represent reactivity of anti-s antiserum.











Table 1-1. Expression of 8/e antigens on serotype III group B streptococci isolated from
septic vs. colonized sources


Source of isolates Number of Number of % 8/e positive
serotype III strains serotype III/8s serotype III strains
strains

Infant blood
Placenta 18 15 83.3
Chorioamniotic
fluid


Colonized infant
Colonized mother 34 20 58.8
with healthy
infant









Table 1-2. In vitro opsonophagocytic killing of group B streptococci by human peripheral
blood leukocytes.

GBS test strain Antiserum CFU-O' CFU-60'
None 0 0
J48 None 396 TNTC'
DL700 352 TNTC

J48 Unadsorbed 423 356
DL700 Anti-J48 whole cell 351 40

J48 Anti-8 specific 399 371
DL700 351 TNTC

Results are the mean of triplicate observations. Less than 5% variability was detected
between replicate plates. 'Too numerous to count (>750).














CHAPTER 2
ISOLATION AND CHARACTERIZATION OF THE 8 AND e ANTIGENS

Introduction

Serotypes la, Ib, II, III, and V are the most common serotypes isolated from

early-onset cases of GBS in the United States (Harrison et al. 1998). Serotype III GBS

are responsible for approximately 90% of all late-onset infections (Baker and Edwards

1994), with meningitis being a common clinical manifestation that results in a substantial

number of survivors suffering permanent neurological damage. Serotype III GBS are also

isolated from cases of adult GBS infection. Therefore, serotype III GBS are a highly

pathogenic serotype involved in all manifestations of GBS disease.

In addition to the group- and serotype-specific carbohydrates, a number of studies

have been conducted in an attempt to identify and characterize surface components that

may contribute to the virulence of serotype III GBS. One of the most well characterized

proteins associated with most serotype III GBS is protein Rib (Stalhammar-Carlemalm et

al. 1993). Rib is a highly repetitive protein with structural and sequence similarity to the

cc protein (Wastfelt et al. 1996). Antibodies to protein Rib confer protective immunity in

a neonatal mouse model of infection (Stalhammar-Carlemalm et al. 1993). Brady et al.

(1988) identified an immunologic reactivity different from Rib and associated with an

overlapping but distinct subset of serotype III GBS which was designated 8. Adsorbing a

polyclonal rabbit antiserum made against a 6-expressing serotype III strain (J48) with a

5-negative serotype III strain (DL700) rendered the reagent more highly specific for the 8









antigen present on GBS whole cells. When the 8-reactive antiserum was further adsorbed

with GBS strain A909 serotypee Ic/ap3y8) to remove antibodies against 8, an additional

reactivity remained with all serotype III strains that express the 8 antigen. These results

indicate that another immunological reactivity is associated with the 8 antigen on a

serotype III strains. This antigen is absent from strain A909. The newly identified antigen

was designated s.

Preliminary opsonophagocytosis assays and epidemiological data suggest that

expression of the 6 and e antigens is a marker for a more pathogenic subset of serotype III

GBS and may be involved in the pathogenicity of this subset (Brady et al. 1996).

Although 8 and were discovered based on their antigenic reactivity, the nature of the

antigens themselves was unknown. The experiments described in this chapter were

designed to isolate and begin to characterize the 6 and e antigens. Initial work was

conducted to extract the antigens from GBS whole cells. Anion-exchange

chromatography was used to separate 8 and e from other material in sonic extracts and a

competition ELISA was used to detect the fluid-phase antigens. To characterize

cell-associated 8 and e antigens, whole bacteria were treated with proteinase K or sodium

metaperiodate. Column chromatography fractions enriched in 8 and e were also analyzed

by Western immunoblot using appropriate antisera to identify reactive material. In an

attempt to generate a more e-specific polyclonal antiserum, e-containing material present

in a sonic extract and eluted from an SDS-polyacrylamide gel was used as an

immunogen. As a basis for comparison, corresponding material from an e-negative

serotype III strain was also used as an immunogen.









Materials and Methods

Bacterial Strains and Growth Conditions.

GBS strains used in this study were A909 serotypee Ic/apy8), J48 serotypee III/8 ),

J52 serotypee III/8e), PEH serotypee III/8s), COH1 serotypee III/81), DL1104 serotypee

III), DL700 serotypee III), and ATCC 12403 serotypee III). Strain A909 serotypee

Ic/acpy8) was obtained from Richard Facklam at the Centers for Disease Control and

Prevention (Atlanta, GA, USA). Strains J48 serotypee III/8), J52 serotypee III/66), and

PEH serotypee III/8s) were clinical isolates from Shands Hospital, University of Florida.

Strain COH1 serotypee III/8c) was a generous gift from Dr. Craig Rubens (Children's

Hospital Regional Medical Center, Seattle, WA, USA). Strains DL 1104 serotypee III)

and DL700 serotypee III) were clinical isolates kindly provided by Dr. Daniel Lim

(University of South Florida, Tampa, FL, USA. Strain ATCC12403 serotypee III) was

obtained from American Type Culture Collection (Manassas, VA). All streptococci were

grown to late exponential phase in Todd-Hewitt Broth (Becton Dickinson, Cockeysville

MD) at 370C without shaking.

Generation of Polyclonal Anti-8s, Anti-8, and Anti-s Antiserum

Polyclonal anti-86, anti-8, and anti-E rabbit antisera were made by immunizing

New Zealand white rabbits with formalin-killed whole cells, followed by adsorption with

appropriate GBS strains. To make an anti-8s reagent, a polyclonal rabbit antiserum was

generated against J48 serotypee III/8s) whole cells. Cells from a 200 mL stationary phase

culture were fixed by treatment with 10 mL of 4% formalin, 0.85% saline solution at 4C

overnight. The density of the immunogen was adjusted to an OD660 of 0.4. A rabbit was

injected intravenously with 0.5mL of the bacterial suspension 3 times/week for 1 week









and then with I mL 3 times/week for 3 weeks. The rabbit was rested for 3 weeks and then

injected with I mL 3 times/week for 2 weeks. The rabbit was exsanguinated by cardiac

puncture under anesthesia. Serum was stored at -200C. This polyclonal rabbit anti-J48

antiserum was made 8s-specific by adsorption with GBS strain DL700 serotypee III).

PBS-washed DL700 whole cells (from a 100 mL overnight culture) were rotated with an

aliquot (1 mL) of the antiserum end-over-end at 4C for 1 h to remove antibodies to the

serotype III and group B polysaccharides, as well as any other common surface antigens

such as protein Rib (Stalhammar-Carlemalm et al. 1993). This procedure was repeated

using fresh cells until reactivity against the adsorbing strain was eliminated. Appropriate

adsorption and specificity was indicated by reactivity with GBS strains J48 and A909, but

not DL700 as determined by whole cell ELISA. To make an e-specific polyclonal

antiserum, the anti-8e antiserum was further adsorbed with GBS strain A909 serotypee

Ic/aopy ), a serotype la strain that expresses 8, but not s. The adsorbed polyclonal rabbit

antiserum was determined to be e-specific if it was reactive with strain J48, but not

strains DL700 or A909 by whole cell ELISA. To make polyclonal anti-5 antiserum,

rabbits were immunized with GBS strain A909. Polyclonal rabbit anti-A909 antiserum

was made 8-specific by adsorbing the antiserum with strains ss617 serotypee Ia) and

ss618 serotypee Ib/oapy) to remove antibodies to all known antigens except 8. The

resulting reagent was determined to be 8-specific if it was reactive with strains A909 and

J48, but not strains ss617, ss618, or DL700. A representative ELISA assay showing

specificity of each of the adsorbed antiserum is shown in Figure 1-2.









Extraction of 8 and & Antigens

Extracts of GBS strains J48 serotypee III/8s) and DL700 (III) were made by

sonicating bacteria three times for 30 s in 20 mM Tris-HCl pH 8.0 with P-800 Potter's

glass beads (1:1:1, wet weight per volume per weight) on ice with a Sonic 300

Dismembrator (ARTEK Systems Corporation, Farmingdale, NY). The glass beads and

bacterial debris were removed from the sonicated mixture by centrifugation at 10 x g for

5 min. Wall membranes were removed in a Beckman L7-55 Ultracentrifuge at > 100,000

x g for 2 h. The non-pelleted material was collected and filtered through 0.2 mm

Acrodiscs (Gelman Sciences, Ann Arbor, MI).

Detection of 8 and s Antigens by Competition ELISA

GBS strain J48 serotypee III/56) whole cells were used to coat Costar High Binding

plates (Costar, Coming, NY). Sample wells were coated overnight at 40C in a moist

chamber with 100 [l of 0.1M carbonate-bicarbonate buffer (pH 9.6) containing 0.02%

sodium azide and approximately 107 streptococcal whole cells. The plates were blocked

with PBS containing 0.03% Tween-20 (PBS-Tw). Test wells were incubated with 100 pl

of bacterial sonic extract (serial 3-fold dilutions) and 100 [l of anti-8o antiserum (1:100

dilution in PBS) for 2 h at 37C. After washing, wells were incubated with horseradish

peroxidase (HRP)-labeled goat-anti-rabbit IgG at a 1:1000 dilution (ICN Biomedicals,

Inc, Aurora, OH) for 2 h at 37C. Plates were washed, and developed with o-

phenylenediamine in 0.1 M citric acid/0.2M sodium phosphate buffer in the presence of

0.03% hydrogen peroxide. OD450 was determined using a Model 550 microplate reader

(Bio-Rad, Hercules, CA). Inhibition was determined by comparing OD450 values of test

wells with OD450 values of control wells containing 100 p1 of buffer only instead of a









column fraction. Results are given as % inhibition. Percent Inhibition = 1 [ OD450

(average of wells containing inhibitor and antibody) (secondary antibody only

background)] / [OD450 (average of wells containing antibody and buffer only) -

(secondary antibody only background) ] x 100.

Anion-Exchange Chromatography

Bacterial sonic extracts were diluted 1:3 in 20 mM Tris-HCI, pH 8.0, and loaded

onto an UNO-Q1 (Bio-Rad, Hercules, CA) anion-exchange column. Unbound material

was removed from the column by washing with 2.5 column volumes of 20 mM Tris-HCI

pH 8.0. Bound molecules were eluted from the column using a stepwise gradient (5 mL

for each step) of 0-1 M NaCI of 0.05 M NaCl increments in 20 mM Tris-HCl pH 8.0.

Detection of 8 and c in Individual Anion-Exchange Column Fractions

To detect 8, GBS strain A909 whole cells were used to coat Costar High Binding

plates (Costar). For c detection, GBS strain J48 was used. The presence of the 6 and/or e

antigens was determined by competition ELISA as described above.

Proteinase K and Sodium Metaperiodate Sensitivity of Cell-Associated 8 Antigen

GBS strain A909 serotypee Ic/capy8) whole cells were used to coat Costar High

Binding plates (Costar) as described above. Test wells were incubated with 100 tl of

either PBS or PBS containing 0.05% proteinase K (Sigma Chemical Co.) for 2 h at room

temperature, or with 100 ptl of either 0.05 M sodium acetate buffer, pH 5.5 or acetate

buffer containing 0.05 M sodium metaperiodate (Sigma Chemical Co.) for 16 h at room

temperature. After treatment, the proteinase K or sodium metaperiodate was removed

from the wells by multiple washes with PBS-Tw. The plates were blocked with PBS-Tw,

and incubated with 100 pl of anti-86 polyclonal antiserum (1:100 dilution in PBS) for 2 h









at 37C, washed and developed as described above. The presence of anti-6 antibodies in

the anti-58 reagent was irrelevant in these experiments since A909 does not express this

antigen.

Enrichment of 8-Specific Antibodies by Immunoprecipitation

An immunoprecipitation method was used in an attempt to increase the

concentration of anti-8 antibodies in the polyclonal anti-6e rabbit sera. Whole cells from

replicate 10 mL overnight cultures of A909 serotypee Ic/capy8) were harvested by

centrifugation. Ten milliliters of 4% formalinized 0.85% saline was added to the cells,

mixed, and stored overnight at 4C. Cells were checked for lack of viability after 24 h by

streaking on Todd Hewitt Broth plates. Five-hundred microliters of anti-6s antiserum was

incubated with one aliquot of GBS strain A909 whole cells and rotated for 2 h at 4C.

Cells were pelleted, and the supernatant was transferred to another aliquot of A909 whole

cells. This procedure was repeated 10 times. Pelleted cells were incubated with 200 jl of

0.5 M acetic acid, 0.15 M NaCI to elute bound anti-6 antibodies. Eluted antibodies were

dialyzed against PBS and concentrated with a Sieze X Protein G Immunoprecipitation

Kit (Pierce Biotechnology, Rockford IL) according to the manufacturer's instructions.

Detection of the 8 Antigen by Western Immunoblot

Proteins contained in anion-exchange column fractions #14 from J48 and DL700

(Figure 2-2) were separated on duplicate 7.5% SDS-polyacrylamide gels, electroblotted

onto PVDF (NENTM Life Science Products, Boston, MA) or nitrocellulose membranes

(Protran, Scleicher & Schuell, Keene, NH) and blocked with PBS-Tw. The PVDF

membrane was incubated with AuroDye ForteTM (Amersham, Piscataway, NJ). The

nitrocellulose was incubated with antiserum enriched for anti-6 antibodies (1:500)









overnight at room temperature. After washing, the blot was incubated with HRP-labeled

goat-anti-rabbit IgG at a 1:1000 dilution (ICN Biomedicals, Inc) for 2 h at room

temperature, washed, and developed with ECLTM detection reagents (Amersham Life

Sciences, San Francisco, CA). An immunoreactive band (-150 kDa) present in the J48

sample was excised from the PVDF membrane and subjected to N-terminal sequencing

by the Interdisciplinary Center for Biotechnology Research (ICBR), University of

Florida, Gainesville, USA using automated Edman chemistry.

In addition to analysis of chromatography fractions, cells from overnight 10 mL

cultures of a panel of 6 surface-positive and 6 surface-negative GBS strains were

centrifuged, washed with 1 mL PBS, and resuspended in 200 pl IX SDS-sample buffer.

Resuspended cell suspensions were boiled for 10 min, and debris removed by

centrifugation. Proteins present in the supernatants were analyzed by Western

immunoblot using antiserum enriched for anti-8 antibodies as described above.

Immunoreactive bands present in samples from 8-positive (-25 kDa) and 5-negative

strains (-35 kDa) were excised from the PVDF membrane and subjected to N-terminal

sequencing by the ICBR.

Detection of the e Antigen by Western Immunoblot

Material contained in an s-containing anion-exchange column fraction #14 from

J48 (Figure 2-2) was analyzed by Western immunoblot as described above for detection

of the 8 antigen with the adsorbed anti-E antiserum (1:1000) as the primary detection

reagent. The corresponding column fraction from DL700 served as a negative control.









Proteinase K and Sodium Metaperiodate Sensitivity of Cell-Associated s Antigen

Experiments were performed as described above for evaluation of the 8 antigen

except that GBS strain J48 serotypee III/8s) was used instead of A909 and anti-s

antiserum (1:100) was used as the detection reagent.

Sodium Metaperiodate Sensitivity of Isolated e Antigen by Western Immunoblot

Material contained in an s-containing anion-exchange column fraction was

separated on 7.5% SDS-polyacrylamide gels and electroblotted onto duplicate

nitrocellulose membranes (Protran, Scleicher & Schuell). Before being blocked with

PBS-Tw, the membranes were treated with 0.05 M sodium acetate buffer pH 5.5 or

acetate buffer containing 0.05 M sodium metaperiodate (Sigma Chemical Co.) for 16 h at

room temperature. After blocking, the treated and untreated membranes were incubated

with anti-s polyclonal antiserum (1:1000) overnight at room temperature. After washing,

the blots were incubated with HRP-labeled goat-anti-rabbit IgG at a 1:1000 dilution (ICN

Biomedicals, Inc) for 2 h at room temperature, washed, and developed with ECLT

detection reagents (Amersham Life Sciences).

Gel Purification of e-Containing Material from an SDS-Polyacrylamide Gel:
Comparison of Comparable Material from an s-Negative Strain

Material from bacterial sonic extracts of GBS strains J48 serotypee III/5s) and

DL700 serotypee III) was separated on replicate 7.5% preparatory SDS-polyacrylamide

gels. A 2 mm gel-slice (above the 250-kDa protein standard) was excised from each gel.

The s antigen was shown previously by Western immunoblot to migrate at this apparent

molecular mass. Each gel slice was finely minced with a razor blade and placed in a 1.5

mL tube. Material was eluted from the minced gel slices with 500 pl. of 50 mM









ammonium bicarbonate pH 7.4, 1% SDS and rocking for 2 h at 370C. The sonic extracts

and gel-eluted materials were analyzed by Western immunoblot as described above with

anti-c antiserum, anti-8 antiserum, and anti-serotype III antiserum (Accurate Chem.,

Westbury, NY) as primary detection reagents. The gel-eluted materials from each strain

were assayed for the presence of group B carbohydrate antigen by latex agglutination

(Phadabact, Boule Diagnostics, Huddinge, Sweden). The gel-eluted materials were also

analyzed for total amino acid content by the ICBR, and for carbohydrate content by the

Complex Carbohydrate Research Center, University of Georgia, Athens USA, supported

in part by the Department of Energy-funded (DE-FG09-93ER-20097) Center for Plant

and Microbial Complex Carbohydrates.

Gel-purified materials from J48 and DL700 were used as immunogens to produce

polyclonal rabbit antisera (Lampire Biological Laboratories, Pipersville, PA). Rabbits

were injected with 100 ptl of each eluted sample mixed with 100 pl. of Incomplete

Freund's Adjuvant on days 1, 7, 14, and 28. On day 42, a 2 mL test bleed was removed.

Rabbits were injected again on days 56 and 84, and a termination bleed was obtained on

day 98.

Sonic extracts of representative of 8e-positive and 8s-negative GBS strains were

separated on duplicate 7.5% SDS-polyacrylamide gels, electroblotted onto duplicate

nitrocellulose membranes (Protran, Scleicher & Schuell), and blocked with PBS-Tw.

Each membrane was incubated with either polyclonal rabbit antiserum made against e-

containing gel-eluted material from J48 (1:1000) or comparable material from DL700

sonic extracts. After washing, blots were incubated with HRP-labeled goat-anti-rabbit









IgG at a 1:5000 dilution (ICN Biomedicals, Inc) for 2 h at room temperature, washed, and

developed with ECLTm detection reagents (Amersham Life Sciences).

Results

Extraction and Detection of the 8 and c Antigens

Sonic extracts of strain J48 serotypee III/6s) and DL700 were prepared and the

presence of the 6 and s antigens evaluated by competition ELISA. This approach was

chosen because the adsorbed antisera recognize the antigens under study specifically on

whole bacterial cells, but as polyclonal reagents they also recognize additional antigens

present in cellular extracts.

The J48 serotypee III/Ss) sonic extract inhibited binding of anti-8s antiserum to J48

serotypee III/8s) whole cells in a dose-dependent manner whereas an extract of DL700

serotypee III) did not (Figure 2-1). Similar results were observed when anti-8 or anti-s

antisera were used, confirming the presence of both antigens in the extract of J48 but not

DL700 (data not shown).

Enrichment for the 8 and e Antigens by Anion-Exchange Chromatography

Bacterial sonic extracts were diluted in 20 mM Tris-HCl, pH 8.0, and loaded onto

an UNO-Q1 (Bio-Rad, Hercules, CA) anion-exchange column. Comparisons of

representative anion-exchange column runs from J48 serotypee III/8s, Figure 2-2, Panel

A) and DL700 serotypee III, Figure 2-2, Panel B) are shown. While the 8 and s antigens

are detectable in the J48 but not the DL700 starting material, the elution profiles of these

two strains are remarkably similar.

The competition ELISA used to detect the presence of 8 and s in the J48 sonic

extract was also used to detect the antigens in individual anion-exchange column









fractions (Figure 2-3). Material eluted from the column with 0.2M NaCi (fractions 13-17)

from J48 inhibited anti-5& antiserum reactivity with A909 serotypee Ic/ap3y8) whole cells

(Panel A), while other column fractions from J48 and column fractions from DL700 did

not. Anti-s reactivity with J48 whole cells was also inhibited by fractions 13-17 from J48

(Panel B), indicating that both 8 and 6 co-eluted from the anion-exchange column in a

peak with 0.2 M NaCl. These fractions were subsequently demonstrated to also contain

the group B and serotype III polysaccharides (data not shown).

Proteinase K Sensitivity of the 8 Antigen on A909 Whole Cells.

To determine if the 8 antigen was sensitive to proteolysis, A909 serotypee Ic/apy8)

whole cells were treated with proteinase K or buffer only (Figure 2-4). Treatment of

A909 whole cells with proteinase K substantially reduced the binding of anti-8e

antiserum compared to treatment with buffer only. This suggests that the 8 antigen is

comprised of protein.

Sodium Metaperiodate Sensitivity of the 8 Antigen on A909 Whole Cells.

To further characterize the 8 antigen, A909 whole cells were also treated with

sodium metaperiodate or buffer only (Figure 2-5). Treatment of A909 whole cells with

sodium metaperiodate had no discernible effect on anti-6s binding suggesting that 8 does

not have a carbohydrate component.

Evaluation of a 8-Containing Anion-Exchange Fraction by Western Immunoblot

Material from an anion-exchange column fraction from strain J48 serotypee III/8e)

that contained the 8 antigen as determined by competition ELISA was separated by SDS-

PAGE and transferred to either nitrocellulose or PVDF. The nitrocellulose membrane

was incubated with antiserum enriched for anti-6 antibodies by immunoprecipitation with









strain A909 serotypee Ic/apy8) (Figure 2-6). An immunoreactive band migrating at -150

kDa was present in the column fraction from J48, but not the corresponding column

fraction from DL700. The -150-kDa J48-specific band present on a replicate PVDF

membrane stained for protein was subjected to N-terminal sequencing by the ICBR using

automated Edman chemistry. The sequence obtained was AVFEKVQEIIVEXLGKD.

BLAST analysis (Altschul et al. 1997) demonstrated homology (16/17 amino acids) with

the deduced amino acid sequence (amino acids 2-18 of 74 total) of gbs0332 (Glaser et al.

2002), an acyl-carrier protein of GBS.

Western Immunoblot of SDS-Extracted Material with Anti-8-Enriched Antiserum

SDS-extracted material from a panel of GBS strains was tested to identify

8-reactive bands present in material extracted by a method other than sonication. SDS-

extracted material from representative 6-positive and 8-negative GBS strains was

separated by SDS-PAGE and transferred to either nitrocellulose or PVDF. The

nitrocellulose membrane was incubated with the antiserum enriched for anti-8 antibodies

by immunoprecipitation with strain A909 serotypee Ic/apy5) (Figure 2-7). An

immunoreactive band migrating at -25 kDa (left arrow) was observed for strain A909

and the four serotype III/8s-positive strains tested. In contrast, an immunoreactive band

migrating at -35 kDa was observed for serotype III/86-negative strains (right arrow).

Both the -25-kDa and 35 kDa bands were excised from a replicate PVDF membrane

stained for protein, and subjected to N-terminal sequencing by the ICBR using automated

Edman chemistry. The amino acid sequence obtained from the 25-kDa band from

8-positive strains was ADVF(E/D)KVQ(E/G)IXDEXL(G/Q). The amino acid sequence

obtained from the 35 kDa band from 5-negative strains was









A(I/V)FE(D/K)VQ(A/E)(R/I)(E/X)V(L/E)RK. BLAST analysis (Altschul et al. 1997) of

both anti-8 immunoreactive bands demonstrated homology (amino acids 2-17 and 2-15,

respectively) with the deduced amino acid sequence of gbs0332 (Glaser et al. 2002), the

same acyl-carrier protein associated with the 150-kDa anti-8 reactive band observed upon

analysis of anion-exchange column fraction #14 of J48 sonic extract.

Visualization of s by Western Immunoblot

Material from a column fraction that contained E as determined by competition

ELISA was electrophoresed on a 7.5% SDS-polyacrylamide gel, transferred to a

nitrocellulose membrane, and incubated with anti-s antiserum to identify immunoreactive

bands present in the J48 serotypee III/8F) sample, but not the corresponding column

fraction from DL700 serotypee III) (Figure 2-8). A prominent band migrating at >250

kDa reactive with anti-s antiserum and present only in the column fraction from J48 was

putatively identified as s. This material was subsequently eluted from an SDS-gel slice

and was able to inhibit anti-s binding to J48 whole cells by competition ELISA,

confirming that the high molecular weight material contained the e antigen (data not

shown).

Proteinase K Sensitivity of the s Antigen on J48 Whole Cells.

To determine whether the e antigen was sensitive to proteolysis, J48 serotypee

III/56) whole cells were treated with proteinase K or buffer only (Figure 2-9). The ability

of anti-s antiserum to bind to proteinase K-treated but not buffer-treated J48 serotypee

III/8s) whole cells was substantially reduced, suggesting a protein component of the s

antigen.









Sodium Metaperiodate Sensitivity of the s antigen from J48

To determine if carbohydrate moieties contribute to the s antigen, J48 serotypee

III/8e) whole cells were treated with sodium metaperiodate or buffer only (Figure 2-10).

Treatment of J48 whole cells with sodium metaperiodate, but not buffer only, reduced

anti-s binding by over 50%, suggesting that in addition to protein, the c antigen is

comprised of a carbohydrate component as well.

Western immunoblot was also used to determine whether reactivity of e-containing

material was destroyed by treatment with sodium metaperiodate. Increasing amounts of

material contained in an e-containing anion-exchange column fraction from J48 were

separated on 7.5% SDS-polyacrylamide gels and electroblotted onto duplicate

nitrocellulose membranes. Membranes were incubated with sodium metaperiodate or

buffer only before incubation with anti-c antiserum (Figure 2-11). Treatment of the e

antigen immobilized on a nitrocellulose membrane with sodium metaperiodate eliminated

the ability of anti-c antibodies to bind to the >250-kDa immunoreactive band.

Gel-Purification e-Containing Material from an SDS-Polyacrylamide Gel:
Comparison of Comparable Material from an s-Negative Strain

Western immunoblot was used to localize anti-e reactive material from J48

serotypee III/8e) to >250 kDa on 7.5% SDS-polyacrylamide gels, although no protein

stained bands were detected in this region of a replicate gel. Assuming the antigen would

be present, e-containing material was extracted from an SDS-polyacrylamide gel of a

sonic extract of J48 using the 250-kDa protein standard as a guide. As a basis for

comparison, similar sized material was gel-purified from a sonic extract of DL700

serotypee III) as well. The extracted materials were analyzed by Western immunoblot









(Figure 2-12), incubating replicate membranes with anti-8, anti-6, and anti-serotype III

antisera. Anti-s antiserum bound only to the gel-purified material from J48. Anti-8 and

anti-serotype III antiserum did not bind to the gel-purified material from either strain.

Gel-purified material from both strains demonstrated reactivity with anti-group B

carbohydrate antibodies as determined by latex agglutination (Phadabact, Boule

Diagnostics, Huddinge, Sweden). The samples were not reactive with antibodies against

groups A, C, D, or G carbohydrates by latex agglutination (data not shown). The gel-

purified materials were subjected to total amino acid and carbohydrate composition

analysis and both contained the polysaccharides rhamnose, mannose, galactose, and

glucose. Rhamnose and galactose are known components of the group B streptococcal

carbohydrate antigen, while mannose and glucose are not. Amino acid analysis suggested

the presence of either a glycine- or serine-rich protein in each sample (data not shown).

Gel-purified e-containing material from J48 and corresponding material from

DL700 containing group B carbohydrate were used as immunogens to produce

polyclonal rabbit antisera. These antisera were used in Western immunoblot analysis of

SDS-boiling cell extracts from representative strains of GBS (Figure 2-13). The

polyclonal rabbit antiserum made against gel-purified material from J48 bound to the

>250-kDa band known to represent the P antigen (arrow). The anti-DL700 rabbit antisera

made against corresponding material was not reactive with this >250-kDa band and did

not recognize any unique bands in SDS-extracted material from serotype III/8s-negative

GBS. To remove antibodies against common determinants, the polyclonal antiserum

generated against gel-purified material from J48 was adsorbed with DL700 whole cells.

The adsorbed reagent retained reactivity with J48 whole cells, but binding to DL700









whole cells was drastically reduced indicating that the adsorbed antiserum contained

antibodies against the 6 antigen (data not shown). To determine if the polyclonal

antiserum made against the gel-purified material from DL700 recognized any antigens

unique to serotype III/8c-negative strains and not expressed by serotype III/68-positive

strains, this antiserum was adsorbed with J48 whole cells. The reactivity of the adsorbed

reagent was eliminated against both DL700 and J48 indicating the gel-purified material

from DL700 did not contain any additional antigens not also contained in the

corresponding sample from J48 (data not shown).

Discussion

To isolate and characterize the 8 and e antigens, multiple techniques were

employed to determine an efficient method of extraction, including treatment with

mutanolysin, boiling whole cells in SDS, sonication, and autoclaving whole cells (Rantz

and Randall 1955). Western immunoblot analysis using anti-8 and anti-& specific reagents

and SDS-polyacrylamide gels stained with Coomassie Brilliant Blue protein stain

suggested that sonication provided extraction of the desired antigens and the highest total

protein yield (data not shown). Therefore, sonication was chosen as a method for

generating starting material for subsequent isolation and purification techniques. A major

drawback to this method was that a number of irrelevant cytosolic proteins were

extracted as well. In spite of this drawback, sonication was used because relatively large

amounts of extract could be generated without substantially altering structures that may

have been modified by mutanolysin treatment or by autoclaving whole cells.

Although adsorbed polyclonal rabbit antisera detect 8 and c specifically on the GBS

cell surface, these reagents recognize additional antigens contained in cellular extracts.









Therefore, a competition ELISA was used to detect fluid-phase 8 and s antigens (Figure

2-1). The binding of anti-56 antiserum to J48 serotypee III/8) whole cells was inhibited

in a dose-dependent manner by a sonic extract of this strain but not by a sonic extract of

DL700 serotypee III). Similar results were observed with anti-5 and anti-s antisera (data

not shown), indicating that sonication of GBS whole cells was effective in extracting both

the 5 and s antigens.

Because a sonic extract of J48 serotypee III/6s) whole cells contained 8 and c this

material, in comparison with a sonic extract of DL700 serotypee III), was used as the

starting material for anion-exchange chromatography in an attempt to isolate and identify

these antigens (Figure 2-2). Elution of bound material with a linear gradient of increasing

NaCl concentrations resulted in 2 broad peaks. One small peak was observed spanning

NaCl concentrations of 0-0.25 M, while one large peak encompassed 0.25-2 M NaCI

(data not shown). Multiple smaller fractions corresponding to more well-defined peaks

were eluted with a stepwise gradient of 0-2 M NaCl. Therefore, this approach was used

for enrichment of the 8 and s antigens.

A competition ELISA was used to detect 6 and e contained in individual anion-

exchange column fractions (Figure 2-3). To detect 8, column fractions were tested for

inhibition of anti-86 antiserum binding to A909 serotypee Ic/apy8) whole cells. This

approach was chosen because the adsorbed anti-8 antiserum was of low titer and reserved

for other experiments. Although the anti-86 antiserum contains antibodies that recognize

s, A909 serotypee Ic/acpy6) does not express s, so these antibodies are irrelevant to the

assay. Inhibition of anti-s antiserum binding to J48 serotypee III/5s) was used to identify

fractions containing E. Both 8 and c eluted with 0.2 M NaCl (column fractions #13-17) in









anion-exchange column runs of J48 sonic extracts. In addition, material reactive with

antibodies against the serotype III and group B polysaccharides were detected in the same

fractions (data not shown), suggesting that multiple surface structures co-purify by this

technique. The presence of additional cellular components in these column fractions has

not been evaluated.

To determine if the 8 antigen was sensitive to protease treatment, A909 serotypee

Ic/apy6) whole cells were treated with proteinase K. Proteinase K has broad substrate

specificity and cleaves at the carboxyl side of hydrophobic, aliphatic, and aromatic amino

acids. Therefore, proteinase K should digest most surface proteins of GBS. Treatment of

A909 whole cells with proteinase K substantially reduced binding of the anti-8

antiserum (Figure 2-4), suggesting that the 8 antigen was proteinaceous.

Because group B and serotype III polysaccharide antigens were detected in column

fractions also containing 8, sensitivity of 6 to treatment with sodium metaperiodate was

evaluated (Figure 2-5). Sodium metaperiodate cleaves adjacent hydroxyl groups in

carbohydrate molecules. Anti-5s antiserum was able to bind to A909 serotypee III/8s)

whole cells after treatment with sodium metaperiodate, which suggests that carbohydrate

does not contribute to the 8 antigen. The possibility still exists that a carbohydrate moiety

unaffected by sodium metaperiodate is a component of the 6 antigen.

Because adsorbed anti-8 antiserum derived from c-protein antiserum made against

strain A909 (Ic/apy8) was very low titer, an attempt was made to increase the

concentration of 8-specific antibodies contained in anti-8 antiserum made against J48

serotypee III/86). To this end, the anti-86 antiserum was reacted with A909 (Ic/aP3y6)

whole cells, unbound antibodies washed away, and bound antibodies eluted from the









whole cells. This reagent was used to identify immunoreactive bands present in a 8-

containing anion-exchange column fraction from J48 serotypee III/8s), by comparison

with the comparable column fraction from DL700 serotypee III) (Figure 2-6). A 150-kDa

immunoreactive band was present in the column fraction from J48, but not from DL700

and was determined to have N-terminal sequence homology to a GBS acyl-carrier protein

(ACP). ACP is a small (-10-kDa), common enzyme known to carry substrates to six

separate enzymes for sequential processing during fatty acid biosynthesis (Wakil et al.

1983). In addition to its role in fatty acid biosynthesis, it has also been shown to be

involved in membrane-based oligosaccharide synthesis (Therisod et al. 1986, Tang et al.

1997) and lipid biosynthesis (Rock and Jackowski 1982) in E. coli. ACP is also involved

in the formation of capsular polysaccharide of rhizobia (Epple et al. 1998). It is unknown

why the GBS ACP was isolated from a band migrating at -150 kDa when ACP itself is

only predicted to be 10 kDa. The most likely explanation is that ACP is bound to another

molecule. This would be consistent with the recent findings of Gully et al. (2003), who

showed that ACP from E. coli could interact with proteins in addition to those necessary

for fatty acid biosynthesis. Furthermore, ACP formed a disulfide bond with IscS, a

protein involved in the synthesis of sulfur-containing molecules. Additional experiments

will need to be completed to determine if ACP is coupled to GBS proteins other than the

putative 8 antigen contained in the -150-kDa band observed by Western immunoblot.

In a further attempt to identify the GBS 8 antigen, SDS-extracted material from

representative 6-positive and 6-negative GBS strains were also analyzed by Western

immunoblot and incubated with immunoprecipitated, 6-enriched antiserum (Figure 2-7).

A -25-kDa immunoreactive band present in samples from all 6-surface positive GBS also









demonstrated N-terminal sequence homology to ACP. An immunoreactive band

migrating at -35 kDa present in samples from all 5-surface negative strains also

demonstrated N-terminal sequence homology to ACP. Demonstration of ACP migrating

at two different molecular weights depending on whether the GBS strain expressed 6 on

its surface was an interesting observation. However, these results are difficult to interpret

without further experimentation. Recombinant ACP was not reactive with anti-8

antibodies indicating that ACP itself does not represent 6. Because the anti-8 reagent

likely recognizes non-8 moieties present in cellular extracts, it is not possible to

determine whether reactivity of the anti-8 reagent with the -35-kDa band in 8-negative

strains represents a non-8 ACP-associated molecule or whether this band represents a 8

variant not exposed on the surface of this subset of strains.

Epsilon was identified by Western immunoblot of an anion-exchange column

fraction of J48 serotypee III/8s) as a high molecular weight antigen (>250-kDa) that was

not detectable by conventional protein stains (Figure 2-8). Binding of anti-E antiserum to

J48 was decreased by treatment with proteinase K (Figure 2-9) and sodium metaperiodate

(Figure 2-10), suggesting that both protein and carbohydrate components contribute to

the e antigen. Treatment of J48 whole cells with proteinase K completely eliminated

binding of anti-s antibodies. Although treatment with proteinase K would destroy the

protein backbone of a glycoprotein antigen, such treatment could also extract the antigen.

Treatment of J48 whole cells with sodium metaperiodate reduced antibody binding by

approximately 50% suggesting that c most likely represents a glycoprotein antigen.

Anti-s antiserum no longer bound to e-containing material after sodium

metaperiodate treatment of blotted antigens on nitrocellulose (Figure 2-11). This









complete ablation on anti-s reactivity contrasted with the results observed with whole

cells where antibody binding decreased by only 50% after treatment with sodium

metaperiodate. It is possible that the protein component recognized by anti-e on whole

cells was denatured and no longer recognized by anti-s antibodies on the Western

immunoblot.

The >250-kDa e-containing material from the J48 serotypee III/8e) sonic extract

and the corresponding material from DL700 serotypee III) were eluted from

SDS-polyacrylamide gels and analyzed by Western immunoblot (Figure 2-12). Eluted

material from J48, but not DL700, was recognized by anti-e, but not anti-6 or anti-

serotype III antisera. The material eluted from J48, but not DL700 was able to inhibit

anti-s antiserum binding to J48 whole cells in a competition ELISA supporting the

hypothesis that this high-molecular weight antigen is s (data not shown). Eluted materials

from both strains were reactive with a group B carbohydrate latex agglutination typing

reagent. Since e and anti-group B reactive material co-purify by anion-exchange

chromatography and are in the same >250-kDa eluted material, one possibility is that

these two entities are closely associated on the cell surface and sonication or SDS-

extraction do not separate the two. Another possibility is that both entities are of similar

molecular weight and co-migrate by SDS-PAGE. Carbohydrate analysis of the eluted

material from both strains demonstrated the presence of rhamnose, mannose, galactose,

and glucose in similar ratios for each strain (4:1:1:6, data not shown). The group B

carbohydrate consists of rhamnose, N-acetylglucosamine, galactose, and glucitol (Michon

et al. 1988). Therefore, in addition to having rhamnose and galactose in common with the

group B carbohydrate, the eluted material from each strain also contained mannose and









glucose. The absence of N-acetylglucosamine and glucitol from the gel-purified materials

of strains J48 and DL700 did not preclude reactivity with the group B specific typing

reagent. This suggests that immunologic detection of group specific determinants do not

depend on the presence of the complete native antigen.

Preliminary amino acid analysis of the e-containing gel-purified material from J48

and corresponding material from DL700 suggested the presence of a glycine-rich or

possibly serine-rich protein. Examination of the genome database of a serotype III strain

(Glaser et al. 2002) revealed no glycine-rich proteins. The genome did however, contain

an open reading frame (ORF) with a deduced amino acid sequence of a serine-rich

protein. Since serine residues are potential sites for either N- or O-linked carbohydrates,

a serine-rich protein could serve as a backbone for glycosylation and account for the

presence of sugars found in each sample.

In an attempt to generate a more specific and higher titer anti-s antiserum, eluted 8-

containing material from J48 serotypee III/68) was used to immunize rabbits. For

comparison, an antiserum was generated against the corresponding material from DL700

serotypee III) as well. Both polyclonal antisera were incubated with SDS-extracted

antigens from numerous GBS strains (Figure 2-13). The antiserum made against the J48

material bound to the >250-kDa antigen present in all 8-positive, but not 8-negative,

strains. The antiserum made against the DL700 material did not recognize any bands

unique to 8-negative strains by Western immunoblot. When antiserum made against J48

was adsorbed with DL700 whole cells, the remaining antiserum bound to J48, but not to

DL700 whole cells. However, when antiserum made against DL700 gel-purified material

was adsorbed with J48 whole cells, antibody binding was decreased against both strains









(data not shown). These results indicate that determinants are recognized by the anti-J48

reagent on 6-positive GBS, but the converse is not true. That is, the anti-DL700 reagent

does not recognize additional determinants present on 6s-negative, but not on 86-positive,

serotype III strains.

Initial characterization of the 8 and e antigens suggests that 5 is a protein and that it

is associated with one of the three know GBS ACPs. It is not yet known whether this

association is specific for the 8 antigen or a more general phenomenon involving an

interaction of the ACP with multiple GBS proteins. Initial work to isolate and

characterize the s antigen suggests it is a glycoprotein rich in glycine or serine and that it

may contain sugars that in part are known to contribute to the group B carbohydrate in

addition to sugars not reported to be related to the group B-specific antigen.







48



1 00 .... ... ....... ... ......... ......................................................... ................... ................. ......
100 -. -bU -------

90

80
70

.-60
50

40

30

20

10

0 0 0 e 0 G 0
1:3 1:9 1:27 1:81 1:243 1:729
Dilution of Clarified Supernatants



Figure 2-1. Detection of 6/s in sonic extracts of group B streptococcal (GBS) whole cells
by competition ELISA. Sonic extracts of J48 serotypee III/816s) (closed
diamonds and DL700 (III) (open circles) were used to inhibit binding of anti-
8c antiserum to J48 serotypee III/8s) whole cells. Results are given as %
inhibition of anti-86 antibody binding.






49


A.















B.













Figure 2-2. Anion-exchange chromatography ofGBS sonic extracts. A) Elution profile of
J48 serotypee ffIII/8s) sonic extract. B) Elution profile of DL700 (III) sonic
extract. OD280 is indicated on the y-axis and fraction number on the x-axis.
The blue lines represent relative levels of eluted protein. The red lines
indicate the concentration of NaCl.











A.

100 -
90
80
70
o 60
.5 50
S 40
30
20
10
0o





B.

100
90
80
70
S60 -
72 50
-E 40
30
20
10




Figure 2-3.


Co COO


5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Column Fraction
Detection of 8 and c in anion-exchange column fractions by competition
ELISA. A) Inhibition of anti-6e antiserum binding to A909 serotypee
Ic/apy6) whole cells B) Inhibition of anti- e antiserum binding to J48
serotypee III/5s) whole cells. Closed diamonds (*) represent inhibition by
J48 (III/8e) column fractions. Open circles (o) represent inhibition by
DL700 (III) column fractions. Results are given as % inhibition as compared
to inhibition in the presence of buffer only.


5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Column Fraction



























1:100 1:200 1:400 1:800 1:1600
Dilution of Anti-DeltalEpsilon


. Evaluation of proteinase K sensitivity of the 8 antigen on A909 serotypee
Ic/apy8) whole cells by ELISA. A909 whole cells were treated with either
proteinase K or buffer only and incubated with anti-8e antiserum. Closed
diamonds (*) represent binding of anti-8e antiserum to proteinase K-treated
A909 whole cells, while open circles (o) represent binding of anti-8s
antiserum to A909 whole cells treated with buffer only.


0.05 -


0 -


Figure 2-4











0.35--------

0.3

0.25

o 0.2

0 0.15 -.

0.1

0.05

0
1:100 1:200 1:400 1:800 1:1600
Dilution of Anti-Delta/Epsilon



Figure 2-5. Evaluation of sodium metaperiodate sensitivity of the 6 antigen on A909
serotypee Ic/apy8) whole cells by ELISA. A909 whole cells were treated with
either sodium metaperiodate or buffer only and incubated with anti-85
antiserum. Closed diamonds (*) represent binding of anti-8 antiserum to
sodium metaperiodate-treated A909 whole cells, while open circles (o)
represent binding of anti-8s antiserum to A909 whole cells treated with buffer
only.









1 2



-250

0-150

-100

-75



-50





Figure 2-6. Evaluation of anti-8e antiserum reactivity with material contained in anion-
exchange column fractions from DL700 serotypee III) and J48 serotypee
II/6s). Anion-exchange column fractions #14 from DL700 (Lane 1) and J48
(Lane 2) were separated on a 7.5% SDS-polyacrylamide gel and
electroblotted onto nitrocellulose, and incubated with anti-5s antiserum
enriched for anti-8 antibodies by immunoprecipitation with strain A909
serotypee Ic/apy5).









1 2 3 4 5 6 7 8


25 kDa 35kDa





Figure 2-7. Evaluation of anti-8s antiserum reactivity with sodium-dodecyl sulfate
(SDS)-extracted material from representative Us-surface positive and 5e-
surface negative GBS strains. Whole bacterial cells were boiled in SDS
sample buffer and extracted material was separated on a 7.5% SDS-
polyacrylamide gel, electroblotted onto nitrocellulose, and incubated with
anti-85 polyclonal rabbit antiserum enriched for anti-8 antibodies by
immunoprecipitation with strain A909 serotypee Ic/apy8). Lanes: 1, A909, 2,
COH1 serotypee III/85), 3, ATCC 12403 serotypee III), 4, PEH serotypee
III/56), 5, DL1104 serotypee III), 6, J52 serotypee III/85), 7, DL700 serotypee
III), 8, J48 serotypee III/8 ). Molecular mass markers in kDa from top are:
250, 150, 100, 75, 50, 37, and 25.










1 2


4-


Visualization of s by Western immunoblot. Materials contained within anion-
exchange column fractions #14 of DL700 serotypee III) (Lane 1) and J48
serotypee III/8c) (Lane 2) were separated on a 7.5% SDS-polyacrylamide
gel, electroblotted onto nitrocellulose, and incubated with anti-C antiserum.
A high molecular weight band (>250 kDa) present in the J48 serotypee
III/8S) column fraction was identified as the s antigen (arrow). Molecular
mass markers in kDa from top are: 250, 150, 100, 75, and 50.


Figure 2-8.














0 .1 4 .------.......-... ..... .... ... .. ..................... ..- .. .. .. .. .
0.12 G4


0.1

o 0,08
Oa

0 0.06

0.04

0.02

0
1:100 1:200 1:400 1:800 1:1600
Dilution of Anti-Epsilon

Figure 2-9. Evaluation of proteinase K sensitivity of the s antigen on J48 whole cells by
ELISA. Closed diamonds (*) represent binding of anti-E antiserum to
proteinase K-treated J48 serotypee III/56) whole cells, while open circles (o)
represent binding of anti-s antiserum to J48 whole cells treated with buffer
only.














0.3



0.25








0.05 --
o 0.2 N-------------.-------------------








0
1:100 1:200 1:400 1:800 1:1600
Dilution of Anti-Epsilon

Figure 2-10. Evaluation of sodium metaperiodate sensitivity of the e antigen on J48
serotypee III/8s) whole cells by ELISA. Closed diamonds (*) represent
binding of anti-s antiserum to sodium metaperiodate-treated J48 whole cells,
while open circles (o) represent binding of anti-s antiserum to J48 whole cells
treated with buffer only.










2 3 4


5


9 Amount of
6 7 8 ^Column Fraction
6 7 8


Buffer Only + Sodium Metaperiodate


Figure 2-11. Evaluation of sodium metaperiodate sensitivity of the e antigen assessed by
Western immunoblot. Anion-exchange column fraction #14 from GBS strain
J48 serotypee III/8s) containing 6 was separated on a 7.5% SDS-
polyacrylamide gel, electroblotted onto duplicate nitrocellulose membranes,
and treated with either buffer or sodium metaperiodate as indicated.
Membranes were incubated with anti-e antiserum. Molecular mass markers in
kDa from top are: 250, 150, 100, 75, and 50.


1











Anti- 8


1 2


1 2


1 2


1 2


Figure 2-12. Western immunoblot analysis of high-molecular weight material gel-
purified from sonic extracts of DL700 (lane 1) and J48 (lane 2). A) Starting
material incubated with anti-c antiserum. B) Gel-purified material incubated
with anti-c, anti-S, or anti-serotype III antisera. Molecular mass markers in
kDa from top are: 250, 150, 100, 75, and 50.


Anti-e


Anti-E


Anti-III













1 23 4 5 6 7 8


1 2 3


45 67


Figure 2-13. Western immunoblot analysis of SDS-extracted material from group B
streptococcal (GBS) whole cells. A) Reactivity with polyclonal rabbit
antiserum made against e-containing material from strain J48 serotypee
III/8&). B) Reactivity with polyclonal rabbit antiserum made against
corresponding material from DL700 serotypee III). Lanes: 1, A909 serotypee
Ic/ap3y,), 2, COH1 serotypee III/8e), 3, ATCC 12403 serotypee III), 4, PEH
serotypee lhI/fe), 5, DL1104 serotypee fI), 6, J52 serotypee IIfhI/), 7, DL700
serotypee III), 8, J48 serotypee I/668s). Arrow indicates binding of antiserum
from J48 to the >250-kDa e antigen. Molecular mass markers in kDa from top
are: 250, 150, 100, 75, and 50.


A.













CHAPTER 3
EVALUATION OF GENETIC COMPONENTS INVOLVED IN 8 AND F
EXPRESSION

Introduction

The 8 and s were initially characterized based on antibody binding. Therefore, the

genetic basis for these immunologic reactivities was of interest. Initial work to isolate and

characterize the 8 antigen by SDS-PAGE and Western immunoblot suggested that 8 was

associated with altered migration of one of the three acyl-carrier proteins (ACP) shown to

be encoded in the genome of GBS (Glaser et al. 2002). In E. coli, ACP is involved in

carrying substrates to six separate enzymes for sequential processing during fatty acid

biosynthesis (Wakil et al. 1983). It has also been shown to be involved in membrane-

derived oligosaccharide synthesis (Tang et al. 1997, Therisod et al. 1986) and lipid

biosynthesis (Rock and Jackowski 1982). ACP can also interact with proteins in addition

to those necessary for fatty acid biosynthesis (Gully et al. 2003). In an attempt to whether

ACP represented a component of the 8 antigen itself, the gene encoding the GBS ACP

associated with anti-8 antibody reactivity was cloned and expressed in E. coli. The

purified recombinant product was tested by Western immunoblot for reactivity with anti-

8 specific antiserum.

In addition to the division of serotype III GBS into two groups based on the

co-expression or absence of the 8 and s antigens, serotype III GBS are also categorized

into 4 distinct lineages based on HindIII and Sse83871 restriction digest patterns (RDPs)

of chromosomal DNA (Takahashi et al. 1998). The phenotypic consequences of these








genetic differences are not yet understood. In collaboration with Dr. John Bohnsack and

Dr. Elisabeth Adderson, the relationship between RDP type and 6 antigen expression was

explored.

As described in the preceding chapter, the e antigen was putatively identified as a

>250-kDa glycoprotein. Amino acid analysis of e-containing material suggested the

presence of a protein rich in either glycine or serine. A search of the available GBS

genomic databases (Glaser et al. 2002, Tettelin et al. 2002) revealed no ORFs encoding a

glycine-rich protein. However, an ORF encoding a serine-rich protein was identified. The

predicted amino acid sequence of this serine-rich protein was that of a -125-kDa protein

with a repetitive di-serine (XS) motif comprising approximately two-thirds of the protein.

Serine-rich repeat (Srr) proteins have recently been described in Streptococcus gordonii

(Bensing and Sullam 2002, Takahashi et al. 2002) and Streptococcus parasanguis (Wu et

al. 1998, 1999).

High molecular weight glycoproteins with a highly repetitive di-serine (XS) motif

backbone such as Fapl of S. parasanguis (Wu et al. 1999) and GspB (Bensing and

Sullam 2002) and Hsa (Takahashi et al. 2002) of S gordonii have been reported for

several streptococcal species and in the case of GspB are encoded within a locus

containing a dedicated accessory secretary apparatus (Bensing and Sullam 2002). These

glycoproteins have been characterized as adhesins. Fap from S. parasanguis binds to

saliva-coated hydroxyapatite (Wu et al. 1998, 1999), while S. gordonii GspB mediates

binding to human platelets (Bensing and Sullam 2002). A similar protein found in

another strain of S. gordonii, Hsa binds to sialic acid (Takahashi et al. 2002). Although

the members of the newly recognized Srr protein family share an XS-motif throughout









the majority of the protein backbone, binding to a given substrate is specific for each

individual protein within this family.

A possible relationship between e antigen expression and a GBS Srr protein was

explored. Experiments were conducted in an attempt to identify genetic components

involved in, or responsible for, e expression.

Materials and Methods

Bacterial Strains and Growth Conditions

Group B streptococcal (GBS) strains used in this study were A909 serotypee

Ic/apy8), ss617 serotypee Ia), ss618 serotypee Ib), 9B200 serotypee II/ay), J48 serotypee

III/8e), J52 serotypee III/8s), PEH serotypee III/85), COH1 serotypee III/8e), COH1-13

serotypee III/8s), DL1104 serotypee III), and DL700 (III). Strains A909, ss617, and ss618

were obtained from Richard Facklam at the Centers for Disease Control and Prevention

(Atlanta, GA, USA). Strains 9B200, J48, J52, and PEH were clinical isolates from

Shands Hospital, University of Florida. Strains COH1 and COH1-13 were generous gifts

from Dr. Craig Rubens (Children's Hospital Regional Medical Center, Seattle, WA,

USA). COH1-13 is an acapsular mutant of the parent strain COH1 (Rubens et al. 1987,

Wessels et al. 1989). Strains DL1104 and DL700 were clinical isolates kindly provided

by Dr. Daniel Lim (University of South Florida, Tampa, FL, USA. All streptococci were

grown to late exponential phase in Todd-Hewitt Broth (Becton Dickinson, Cockeysville

MD) at 370C without shaking.

Cloning and Expression of Recombinant Acyl Carrier Protein (ACP) from J48

Chromosomal DNA from J48 serotypee III/68) was isolated as described by

Nagano et al. (Nagano et al. 1991) and further purified on CsCI gradient. DNA encoding








acp was amplified by PCR with J48 DNA as template using the forward primer KS46

and the reverse primer KS47 (Table 3-1) as determined from the gbs0332 sequence of

NEM316 (Glaser et al. 2002). DNA was first denatured at 94C for 1 min. Thirty

amplification cycles were performed as follows: 1 min of denaturation at 940C, 2 min of

annealing at 600C, and 1 min of extension at 720C, followed by one cycle of 5 min at

720C. The PCR product was analyzed by gel electrophoresis in 0.7% agarose. The

amplified product was ligated to the pBAD-TOPO (Invitrogen, Carlsbad, CA)

expression vector and used to transform E. coli strain TOP 10 according to the

manufacturer's protocol. This plasmid was designated pKS90. The pBAD-TOPO vector

contains the araBAD promoter for tightly regulated expression by L-arabinose, a C-

terminal V5 epitope tag for detection of an induced protein, and a C-terminal

polyhistidine tag for purification. Correct orientation and reading frame with the C-

terminal V5 epitope was confirmed by DNA sequence analysis provided by the ICBR

using the universal primers supplied with the pBAD-TOPO kit (Invitrogen).

Purification of Recombinant Acyl Carrier Protein (ACP) and Western Immunoblot

Ten milliliters of mid-log phase E. coli cells (OD600 = 0.5) containing pKS90 was

induced with 10-fold decreasing concentrations of in L-arabinose (from 2 to 0.0002%) to

express recombinant ACP. Cells were grown for 4 h after induction. Induced cells were

harvested by centrifugation and resuspended in 1 mL lysis buffer (50m M NaH2PO4 pH

8.0, 0.3 M NaCl, 10 mM imidazole). One milligram of lysozyme (Sigma Chemical Co)

was added to each tube and placed on ice for 30 min. Cells were disrupted with a Mini

Beadbeater-8 (Biospec Products) 6 x 10 s (5 s rest). Residual intact cells and insoluble

material were pelleted by centrifugation for 10 min at 10,000 x g and the supernatant was








loaded onto a Ni-NTA Spin Kit column (Qiagen, Valencia CA), and centrifuged through

the Ni resin. Unbound material was removed from the column with two washes using

50m M NaH2PO4 pH 8.0, 0.3 M NaC1, 20 mM imidazole. Bound molecules were eluted

from the column with 50m M NaH2PO4 pH 8.0, 0.3 M NaCi, 250 mM imidazole. As

controls, uninduced cultures were treated identically. Material eluted from the column

was separated on 10-20% SDS-polyacrylamide gradient gels (Bio-Rad), electroblotted

onto nitrocellulose membranes (Protran, Schleicher & Schuell) and blocked with PBS-

Tw. One membrane was incubated with anti-V5 monoclonal antibody (Invitrogen) at a

1:5000 dilution overnight at room temperature. After washing, the blot was incubated

with HRP-labeled goat-anti-mouse immunoglobulin (Ig, H+L chain) at a 1:1000 dilution

(ICN Biomedicals, Inc) for 2 h at room temperature, washed, and developed with 8 mL of

4-chloro-1-naphthol solution (7 mL PBS, 1 mL 4-chloro-1-naphthol [Sigma Chemical

Company; 3 mg/mL in methanol], 8 tl of 30% hydrogen peroxide [Fisher Scientific]).

The other membrane was incubated with anti-8 polyclonal rabbit antiserum overnight at

room temperature. After washing, the blot was incubated with HRP-labeled goat-anti-

rabbit Ig at a 1:1000 dilution (ICN Biomedicals, Inc) for 2 h at room temperature,

washed, and developed with ECLTM detection reagents (Amersham Life Sciences).

Polymerase Chain Reaction (PCR) Amplification of srr from GBS Chromosomal
DNA

Chromosomal DNA from GBS was isolated as described by Nagano et al. (1991),

purified on CsCI gradients, and used as template for PCR amplification ofsrr. DNA

encoding srr was amplified using primers KS 13(F) and KS 14(R) (Table 3-1) as

determined from the srr sequence ofNEM316 (gbsl529, Glaser et al. 2002). DNA was

first denatured at 940C for 1 min. Thirty amplification cycles were performed as follows:








1 min of denaturation at 94C, 2 min of annealing at 600C, and 2 min of extension at

72C, followed by one cycle of 5 min at 720C. The PCR products were analyzed by gel

electrophoresis in 0.7% agarose.

Detection of srr by Southern Hybridization

DNA hybridization was performed by the method of Southern (Sambrook 1989).

Chromosomal DNA from all GBS strains was isolated as described by Nagano et al.

(1991) and further purified on CsCl gradients. Two micrograms of genomic DNA from

each of the GBS strains tested was digested with EcoRV (Promega, Madison, WI) at

370C overnight. The digested DNA was separated by electrophoresis through 0.7%

agarose gels and transferred to a nylon membrane (Boehringer Mannheim, Indianapolis,

IN) by a capillary blot procedure according to the manufacturer's instructions. The

membrane was probed with the DL700 serotypee III) srr ORF labeled during PCR

amplification with Digoxigenin-11-dUTP (DIG-dUTP, Boehringer Mannheim) at 42C

overnight. DNA encoding srr from DL700 was amplified using primers KS 13(F) and

KS 14(R) (Table 3-1) as determined from the srr sequence of NEM316 (gbs 1529, Glaser

et al. 2002). DNA was first denatured at 940C for 1 min. Thirty amplification cycles were

performed as follows: 1 min of denaturation at 940C, 2 min of annealing at 600C, and 2

min of extension at 720C, followed by one cycle of 5 min at 72C. The PCR product was

analyzed by gel electrophoresis in 0.7% agarose. The blot was developed according to the

manufacturer's instructions.

RNA Dot Blot to Detect srr mRNA

Total RNA was isolated from streptococci using the RNeasy Mini Kit (Qiagen,

Valencia CA) according to the manufacturer's instructions. Serial dilutions of total RNA

were transferred to a nylon membrane (Boehringer Mannheim) using a dot blot apparatus









(Schleicher & Schuell). The membrane was incubated with the DL700 serotypee III) srr

ORF labeled with DIG-dUTP (Boehringer Mannheim) during PCR amplification. The

blot was developed according to the manufacturer's instructions.

Polymerase Chain Reaction (PCR) of GBS Genes Within the secA2-Y2 locus of
Sequenced GBS Strain NEM316

Forward and reverse primers were designed to amplify individual ORFs contained

in the secA2-Y2 locus (Bensing and Sullam 2002) identified within the published

NEM316 serotypee III, RDP III-1) genome sequence (Glaser et al. 2002). These primers

are listed in Table 3-1. Chromosomal DNA from J48 serotypee III/85) and from DL700

serotypee III) was used as template. DNA was first denatured at 940C for 1 min. Thirty

amplification cycles were performed as follows: 1 min of denaturation at 94*C, 2 min of

annealing at 600C, and 2 min of extension at 72C, followed by one cycle of 5 min at

720C. The PCR products were analyzed by gel electrophoresis in 1% agarose.

Evaluation of c Expression by GBS Strains of RDP III Types 1-4

Costar High Binding plates (Costar) were used. Sample wells were coated

overnight at 40C in a moist chamber with 100 pl of approximately 107 streptococcal

whole cells in 0.1 M carbonate-bicarbonate buffer (pH 9.6) containing 0.02% sodium

azide. The plates were blocked with PBS-Tw. Test wells were incubated with 100 pl of

anti-c antiserum (1:100 dilution in PBS) for 2 h at 370C. After washing 3 times with 300

uil PBS-Tw, 100 pl of affinity-purified HRP-labeled goat-anti-rabbit secondary antibody

(1:1000 dilution in PBS) (ICN Biomedicals, Inc) was added to the wells for 2 h at 370C.

Plates were washed, and developed with o-phenylenediamine in 0.1M citric acid/0.2M

sodium phosphate buffer in the presence of 0.03% hydrogen peroxide. OD450 of test wells

was determined using a Model 550 microplate reader (Bio-Rad). Control wells included









incubation of streptococcal whole cells with secondary antibody alone. A strain was

designated positive for s expression if the OD450 value observed was more than twice that

of the negative control strain DL700 serotypee III, RDP III-2).

Detection of RDP 111-3 Variant srr by Southern Hybridization

DNA hybridization was performed by the method of Southern (Sambrook 1989).

Chromosomal DNA from all GBS strains was isolated as described by Nagano et al.

(1991) and further purified on CsCl gradients. Two micrograms of genomic DNA from

each of the GBS strains tested was digested with EcoRV (Promega, Madison, WI) at

37C overnight. The digested DNA was separated by electrophoresis through 0.7%

agarose gels and transferred to a nylon membrane (Boehringer Mannheim, Indianapolis,

IN) by a capillary blot procedure according to the manufacturer's instructions. DNA

encoding srr from GBS strain J48 serotypee III/8s) was amplified by PCR using primers

KS51(F) and KS53(R) based on the sequence of an srr gene contained within clone DY-3

derived from strain 874391 (RDP III-3) (Bohnsack et al. 2002). This srr was similar to,

but different from, that identified in the genome of the sequence strain NEM316 (Glaser

et al. 2002). DNA was first denatured at 940C for 1 min. Thirty amplification cycles were

performed as follows: 1 min of denaturation at 94C, 2 min of annealing at 600C, and 2

min of extension at 720C, followed by one cycle of 5 min at 720C. The PCR product was

analyzed by gel electrophoresis in 0.7% agarose. The blot was developed according to the

manufacturer's instructions.

Cloning of srr from J48

The srr gene from GBS strain J48 serotypee III/8&) was amplified by PCR as

described above. The amplified product was ligated to pBAD-TOPO (Invitrogen)








expression vector and used to transform E. coli strain TOP 10 according to the

manufacturer's protocol. The resulting plasmid was designated pKS60. Correct

orientation and reading frame with the C-terminal V5 epitope was confirmed by DNA

sequence analysis provided by the ICBR using the universal primers supplied with the

pBAD-TOPO kit (Invitrogen).

Reactivity of Anti-s Antiserum with Recombinant Srr from J48

Ten milliliters of mid-log phase E. coli cells (OD6oo = 0.5) harboring pKS60 was

induced to express recombinant Srr with 10-fold increases in L-arabinose (from 0.0002 to

2%). Cells were grown for 4 h after induction, harvested by centrifugation, and

resuspended in 1 mL of IX SDS-sample buffer. The cells were boiled for 5 min and

pelleted. Fifty microliters of the supernatant was separated on 7.5% SDS-polyacrylamide

gels, electroblotted onto nitrocellulose membranes (Protran, Scleicher & Schuell), and

blocked with PBS-Tw. One membrane was incubated with anti-V5 monoclonal antibody

(Invitrogen) at a 1:5000 dilution overnight at room temperature. After washing, the blot

was incubated with HRP-labeled goat-anti-mouse Ig (H+L chain) at a 1:1000 dilution

(ICN Biomedicals, Inc, Aurora, OH) for 2 h at room temperature, washed, and developed

with 8 mL of 4-chloro-l-naphthol solution (7 mL PBS, 1 mL 4-chloro-l-naphthol [Sigma

Chemical Company; 3 mg/mL in methanol], 8 tl of 30% hydrogen peroxide [Fisher

Scientific]). The other membrane was incubated with anti-s antiserum overnight at room

temperature. After washing, the blot was incubated with HRP-labeled goat-anti-rabbit Ig

at a 1:1000 dilution (ICN Biomedicals, Inc, Aurora, OH) for 2 h at room temperature,

washed, and developed with ECLTM detection reagents (Amersham Life Sciences, San

Francisco, CA).








Results

Western Immunoblot of Recombinant Acyl-Carrier Protein

Because anti-8 antiserum reacted with bands that demonstrated sequence homology

to a GBS acyl-carrier protein, the gene encoding the ACP was amplified by PCR, cloned,

expressed to determine if ACP represented a component of the 8 antigen. The

recombinant ACP protein expressed in E. coli was purified on nickel resin as described

above. Starting material, wash material, and eluted material from the Ni column from

induced and uninduced cultures were tested for reactivity with anti-V5 and anti-8

antiserum by Western immunoblot (Figure 3-1). The anti-V5 reagent (panel A)

recognized a band of-12 kDa that was present in the starting material of the uninduced

culture (lane 1) and was eluted from the nickel resin column (lane 3). This is the

predicted size of the recombinant ACP with V5 epitope and histidine tag. The band was

not present in the samples derived from the uninduced culture (lanes 4-6) and was

concluded to represent the V5 epitope and histidine-tagged recombinant ACP. A band

migrating at -60 kDa recognized by the anti-V5 antibody and present in both the induced

(lane 1) and uninduced (lane 4) starting material was not purified on nickel resin and was

considered to be an irrelevant anti-V5 cross-reactive moiety unrelated to ACP. The anti-8

antiserum (panel B) did not react with the -12 kDa recombinant ACP present in the

eluted sample from the induced culture (lane 3) indicating that ACP itself is not the 8

antigen. Reactivity of the polyclonal anti-8 antiserum with components present in the

starting materials of the induced (lane 1) as well as uninduced (lane 4) cultures represents

background reactivity of this reagent with E. coli proteins unrelated to expression of

recombinant ACP.









Polymerase Chain Reaction (PCR) Amplification of srr from GBS Chromosomal
DNA

Preliminary carbohydrate and amino acid analysis of s-containing material

suggested the presence of a protein component rich in either glycine or serine.

Examination of the GBS genomic sequence (Glaser et al. 2002) revealed no hypothetical

glycine-rich protein, but did identify one open reading frame (ORF) with a deduced

amino acid sequence rich in serine. PCR was used to attempt to amplify DNA

corresponding to this srr serinee rich repeat) gene from multiple GBS strains and

amplified products were separated on a 1% agarose gel (Figure 3-2). This srr gene was

only amplified from 6S-negative strains of GBS including 8s-negative serotype III strains

and serotype Ic, Ib, and II strains. Other serotypes were not tested.

Southern hybridization with srr from DL700

To determine if the lack of PCR amplification of srr observed for serotype

III/56-positive strains was a result of a lack of the gene itself or a consequence of

inadequate amplification using primers designed based on the NEM316 sequence (Glaser

et al. 2002), Southern analysis was performed (Figure 3-3). The srr gene was amplified

and labeled from the 6e-negative serotype III strain DL700 and used as a probe to detect

the presence of the srr gene in the same panel of GBS strains used for PCR amplification.

Again, no 86-positive strains demonstrated the presence of the srr gene; whereas,

chromosomal DNA from all 5s-negative strains, including serotype III and non-serotype

III GBS, hybridized with the srr probe derived from DL700.

Dot Blot to Detect srr mRNA

Because srr DNA could not be detected in chromosomal DNA from e-positive

GBS by PCR or Southern hybridization, an RNA dot blot procedure was used to









determine if srr might be located on an extra-chromosomal element in those

strains.(Figure 3-4). To evaluate the presence of srr mRNA, hybridization was conducted

using total RNA from J48 serotypee III/85) and DL700 serotypee III) immobilized on a

nylon membrane and incubated with DIG-labeled srr PCR product amplified from

DL700 chromosomal DNA as the probe (Figure 3-4). RNA from S. mutans strain MK4

was included as a negative control. No ORFs encoding serine-rich repeat proteins are

present in the S. mutans genomic sequence (Ajdic et al. 2002). The probe hybridized with

RNA from DL700, but not J48, suggesting that srr is not present on an extra-

chromosomal element in 8s-positive serotype III GBS. The collective results of PCR,

Southern hybridization, and RNA dot blot suggested that the srr ORF present in the

NEM316 database (Glaser et al. 2002) was only present in e-negative strains, but not e-

positive GBS strains. This information also suggested that serotype III NEM316 would

be a 6e-negative strain. A whole-cell ELISA assay confirmed this to be the case (data not

shown).

Polymerase Chain Reaction (PCR) Amplification of ORFs Located in the srr-
Containing Locus

As stated above, Se-positive serotype III GBS strains did not possess the srr ORF

present in the NEM316 genome (Glaser et al. 2002). The srr ORF is located in a putative

operon similar to that of a locus in S. gordonii named secA2-Y2 because of the presence

of additional secA and secY homologues dedicated to the export of GspB, a Srr

homologue (Bensing and Sullam 2002). Since 6s-positive serotype III GBS strains did

not possess the srr ORF, it was possible that other ORFs in this locus were missing as

well. Therefore, PCR amplification of other ORFs in the locus was conducted using J48

serotypee III/8s) and DL700 serotypee III) chromosomal DNA as template (Figure 3-5).









PCR products were visualized for all ORFs amplified from DL700 chromosomal DNA.

The only PCR products visualized using J48 chromosomal DNA were mutT and uvrB. In

the NEM316 genome (Glaser et al. 2002), these genes are approximately 24 kb apart,

suggesting that 8s-positive GBS lack this 24-kb segment of DNA.

Evaluation of c-Expression by RDP III Types 1-4

In addition to immunological classification of serotype III GBS into two groups

based on the presence or absence of the 6 and E antigens, serotype III GBS can be

subdivided into four distinct lineages based on HindIII and Sse83871 restriction digest

patterns (RDPs) of chromosomal DNA (Takahashi et al. 1998). In collaboration with Dr.

John Bohnsack (University of Utah Health Sciences Center, Salt Lake City, UT) and Dr.

Elisabeth Adderson (St. Jude Children's Research Hospital, Memphis TN),

representatives of RDP III types 1, 2, 3, and 4 were assayed for e expression by whole

cell ELISA (Table 3-2) to determine if there was a correlation between e-expression and

RDP type. A strain was determined to be positive for s expression if the ELISA OD450

value observed was more than twice that of the negative control strain DL700. The OD450

values observed for RDP III-3 strains varied from a high value of 0.45 to a low of 0.13.

The OD450 value observed for the negative control strain DL700 serotypee III, RDP 11I-2)

was 0.06. Using this criterion, no RDP 1, 2, or 4 strain was positive for e expression,

while all fifteen RDP III-3 strains tested were e-positive. Subsequent molecular typing by

Dr. Bohnsack's lab of three serotype III/86-positive strains (J48, J52, and PEH) and two

serotype III/65-negative strains (DL700, DL1104) from our laboratory collection revealed

that the former strains were RDP III-3 strains, while the latter strains were non-III-3 (data

not shown). The majority (91%) of serotype III invasive neonatal isolates have been








reported to be RDP III-3 (Takahashi et al. 1998). Brady et al. (1996) also found that

serotype III GBS isolate from septic sources appeared to exhibit a higher frequency of

expression of 85 (83.3%) compared to serotype III isolates from colonizing sources

(58.8), although the sample size was to small to ascertain statistical significance. Taken

together, these results indicate that 8/s expression by serotype III GBS is an antigenic

marker for the highly pathogenic RDP III-3 lineage.

Variant srr Associated with RDP III-3 8c-Positive GBS

Bohnsack et al. (2002) used genomic subtractive hybridization to identify DNA

sequences unique to RDP III-3 strains. One clone, designated DY-3, contained DNA

encoding a large putative protein rich in serine, with a repetitive di-serine (XS) motif for

approximately two-thirds of the protein. This putative protein appeared to represent a

variant of the large serine-rich protein found in the published genomic databases of a

serotype III and serotype V strain (Glaser et al. 2002, Tettelin et al. 2002). Both variants

of srr encode an XS motif for approximately two-thirds of the entire protein, however,

the DY-3 srr is slightly smaller (3.6 kb) than the NEM316 variant (4.0 kb). In addition to

primary nucleotide sequence differences, in the NEM316 variant, the amino acids that

alternate with serine in the XS motif are alanine, threonine, and methionine, while the

DY-3 variant has alternating amino acids glutamic acid, isoleucine, valine, and serine.

There are also significant amino acid differences between the N-terminal small XS repeat

(-80 amino acids) and the large C-terminal XS repeat (-800 amino acids).Therefore,

GBS strains were tested by Southern hybridization to determine the relationship between

8/s expression by serotype III GBS and the presence of the serine-rich repeat (srr) ORF

contained within the DY-3 sequence (Bohnsack et al. 2002) (Figure 3-6). The








DIG-labeled srr PCR product amplified from J48 serotypee III/83) chromosomal DNA

using primers based on the DY-3 sequence hybridized with all of the serotype III/8s-

positive strains tested, but not with any of the e-negative strains including A909 serotypee

Ic/apy5). This suggests that RDP 11I-3, e-positive strains have a variant srr gene, which is

similar to, but different from, the srr gene contained within the NEM316 genome (Glaser

et al. 2002).

Recognition of a Recombinant Srr by Anti-e Antiserum

Reactivity of anti-s antiserum with GBS whole cells is sensitive to both proteinase

K and sodium metaperiodate, suggesting that the c antigen is comprised of both

carbohydrate and protein. In order to determine if the protein encoded by srr contained in

clone DY-3 represents a protein component of the s antigen, the srr ORF amplified by

PCR from J48 serotypee III/8e) was cloned into pBAD-TOPO (Invitrogen) expression

vector and induced to express the Srr protein. Cell lysates were assayed by Western

immunoblot for the presence of the recombinant protein using an anti-V5 monoclonal

antibody (Figure 3-7, Panel A). A reactive band was detected at >250 kDa. The expressed

recombinant Srr was also reactive with anti-e antiserum by Western immunoblot (Figure

3-7, Panel B). The apparent molecular mass of >250 kDa is bigger than the size predicted

for the recombinant Srr (125 kDa); however, when the protein was analyzed under

strongly denaturing conditions in the presence of 9 M urea, migration of Srr was as

predicted (data not shown). This suggests that Srr may have migrated as a dimer when

urea was absent from the gel. The reactivity of recombinant Srr with anti-s antibodies

suggests that the Srr protein from J48 is a component of the s antigen. Reactivity of anti-s

antibodies with J48 whole cells was diminished but not completely destroyed by








treatment of whole cells with sodium metaperiodate. Treatment of whole cells with

proteinase K virtually eliminated anti-s reactivity (see Chapter 2). These results are

consistent with Srr comprising the protein backbone of a glycoprotein corresponding to

the 6 antigen.

Schematic Representations of Variant srr-Containing Loci

The srr-containing locus identified in the NEM316 genome sequence (Glaser et al.

2002) has both secA2 and secY2 homologues as well as six ORFs encoding putative

glycosyltransferases and several ORFs encoding hypothetical proteins of unknown

function (Figure 3-8, panel A). In this organization, the srr and secY2 genes are separated

by approximately 12-kb of DNA that includes the glycosyltransferases genes. In the

variant locus associated with serotype III/6s-positive RDP III-3 GBS (Panel B), the secY2

gene lies immediately adjacent (113 nt) and downstream of the srr gene. The presence of

glycosyltransferases genes or homologues of other ORFs contained in the non-RDP III-3

locus is unknown at present as the DY-3 clone does not include DNA downstream of

secY2. It is interesting that insertion sequences for the Tn 10 transposon are contained

within the DY-3 clone. The proximity of such insertion sequences to the srr-containing

variant locus associated with RDP III-3 strains suggests a genetic recombination event in

this lineage.

Discussion

Initial work to isolate and characterize the 8 antigen by SDS-PAGE and Western

immunoblot suggested that 8 is associated with one of the three known GBS acyl-carrier

proteins (ACP) identified in the NEM316 (Glaser et al. 2002) genomic sequence. One of

the 5-reactive bands demonstrating an N-terminal sequence corresponding to ACP and








associated with A909 (Ic/apy8) and serotype III/8s-positive GBS had an apparent

molecular mass of-25 kDa. Curiously, serotype III/8 -negative GBS lacked this band

but demonstrated an anti-8 reactive band migrating at -35 kDa, also with N-terminal

sequence homology to the same ACP. The ACP has a predicted size of-10 kDa;

therefore, these bands likely represent ACP completed with other moieties. It is unclear

why anti-5 antibodies react with the -35 kDa band from serotype III/8s-surface negative

strains. It is possible that a variant form of 8 is expressed by these strains but that 8

epitopes are not exposed on the cell surface. It is also possible that the ACP-associated

moiety detected in these strains is unrelated to 8. As stated previously, the anti-8

antiserum recognizes 8 specifically on whole cells, but reacts with numerous additional

antigens in cellular extracts.

ACP is involved in fatty acid (Wakil et al. 1983), oligosaccharide (Tang et al. 1997,

Therisod et al. 1886), and lipid biosynthesis (Rock and Jackowski 1982). ACP can also

form a disulfide bond with a protein not involved in fatty acid biosynthesis (Gully et al.

2003). To determine whether ACP was the 8 antigen itself, the gene encoding the GBS

ACP was cloned and the recombinant protein expressed in E. coli. Anti-8 antiserum did

not bind to the purified recombinant ACP (Figure 3-1). Follow-up studies beyond the

scope of this project will be necessary to clarify the relationship between the GBS 8

antigen and ACP.

The E antigen was found to correspond to a high molecular weight antigen (>250-

kDa) with both protein and carbohydrate components. Amino acid analysis of e-

containing material suggested the presence of a protein rich in either glycine or serine. By

searching GBS genomic databases (Glaser et al. 2002, Tettelin et al. 2002), no glycine-









rich proteins were identified, while an ORF with a deduced amino acid sequence of a

serine-rich protein was identified. The deduced amino acid sequence of this serine-rich

protein suggested an -125 kDa protein with a di-serine (XS) motif comprising

approximately two-thirds of the protein. PCR amplification of this ORF was obtained

using chromosomal DNA only from 8e-negative GBS strains as template (Figure 3-2).

These strains included multiple serotypes. PCR amplification was not obtained for

serotype III/8c-positive strains.

It was possible that the srr ORF from 5s-positive strains was not amplified due to a

sequence variation from the designed primers. Therefore, Southern blot was performed to

determine if srr was present in 5e-positive strains (Figure 3-3). Hybridization with the

PCR-amplified srr from DL700 serotypee III) was only detected for e-negative strains,

suggesting that the srr sequence did not exist in the chromosome of &-positive strains, or

that there were significant differences in the srr sequences.

To exclude the possibility that Srr was encoded on an extra-chromosomal element

in 8e-positive strains, an RNA dot blot experiment was performed (Figure 3-4). The only

hybridization observed with the labeled srr PCR product from DL700 was with mRNA

from that strain. Strain J48 did not demonstrate the presence of srr mRNA. This inverse

relationship between e-expression and the presence of the srr ORF suggested that the

sequenced NEM316 serotype III strain should be 5e-negative. A whole-cell ELISA

proved that this was indeed the case (data not shown).

The srr ORF from the NEM316 genome (Glaser et al. 2002) is located in a locus

with significant similarity to a locus in S. gordonii described by Bensing and Sullam

(2002). In this organism, the Srr protein (GspB) is responsible for binding to human








platelets. The gene encoding GspB is located upstream of secA and secYhomologues,

designated secA2 and secY2. The entire locus has been called the secA2-Y2 locus.

Because 6-positive GBS did not contain a srr ORF as determined by PCR, DNA

hybridization, and RNA dot blot, PCR amplification was used to determine if J48

serotypee III/8c) lacked other ORFs in this locus (Figure 3-5). Only two genes, mutT and

uvrB were amplified from the J48 chromosome, while nine additional ORFs within the

24-kb separating mutT and uvrB in the NEM316 genomic sequence were amplified from

the DL700 serotypee III) chromosome. This suggests that a number of genes present in

the srr-containing secA2-Y2 locus of e-negative GBS are lacking in 6-positive strains;

however, technical difficulties associated with PCR amplification cannot be excluded as

an explanation for these results.

The work by Takahashi et al. (1998) suggested RDP type III-3 represents a more

pathogenic lineage of serotype III GBS, and preliminary epidemiological data (Brady et

al. 1996) suggested that 8/e expression was a marker for a more pathogenic subset of

serotype III GBS. Therefore, RDP type (1-4) and the presence or absence of the e antigen

on the cell surface of serotype III strains were evaluated in parallel. All e-positive strains

were III-3, while all 6-negative strains were non-III-3. These results indicate that the

highly pathogenic RDP III-3 lineage can be identified immunologically on serotype III

GBS, namely by the detection of the co-expressed 8 and e antigens.

Bohnsack et al. (2002) used genomic subtractive hybridization to identify DNA

sequences unique to RDP III-3 strains. One clone, designated DY-3, contained DNA

encoding a large putative protein rich in serine, with a di-serine (XS) motif for

approximately two-thirds of the protein. This putative protein was similar, but not









identical, to the homologous large serine-rich protein found in the published genomic

databases of a serotype III and a serotype V strain (Glaser et al. 2002, Tettelin et al.

2002). Therefore, a panel of GBS strains was tested by Southern hybridization to

determine which, if any, contained the serine-rich repeat (srr) ORF identified within the

DY-3 clone (Bohnsack et al. 2002) (Figure 3-6). A correlation was observed between e

expression and the DY-3 variant of srr. All RDP III-3/8e-positive GBS strains tested

contained the DY-3 srr variant; whereas all e-negative GBS contained the NEM316 srr

variant (Figure 3-3). Although approximately two-thirds of each Srr is composed of a di-

serine (XS) motif, there was not any detectable cross-hybridization between the variant

genes. There are two major differences in these two variant Srr proteins. In the NEM316

variant the amino acids that alternate with serine in the XS motif are alanine, threonine,

and methionine, while the DY-3 variant has alternating amino acids glutamic acid,

isoleucine, valine, and serine. There are also significant amino acid differences between

the N-terminal small XS repeat (-80 amino acids) and the large C-terminal XS repeat

(-800 amino acids).

Based on SDS-PAGE, Western immunoblot, and composition analysis, e was

determined to likely represent a high molecular weight glycoprotein rich in serine. Since

e-positive GBS contained a variant srr with similarity to genes encoding high molecular

weight di-serine glycoproteins of S. gordonii (Bensing and Sullam 2002) and S.

parasanguis (Wu et al. 1998), DNA encoding Srr from the s-positive GBS strain J48

serotypee III/86) was cloned into an expression vector and the recombinant Srr protein

was expressed in E. coli. Anti-s antiserum bound to the induced recombinant Srr protein

(Figure 3-7). The predicted molecular mass of this Srr is -125 kDa. It is unknown why









the recombinant protein migrates at -250 kDa. In the presence of 9 M urea, the

recombinant Srr migrates close to its predicted size. It is possible that under standard

SDS-PAGE conditions the recombinant Srr protein migrates as a dimer or that the

molecule is not sufficiently denatured and its migration is aberrant. Results suggested that

the e antigen is comprised of both protein and carbohydrate components. It is probable

then, that only part of E antigenic reactivity is represented by recombinant Srr. Additional

studies will be necessary to identify possible carbohydrate components and linkages that

may contribute to s antigenic reactivity. Of interest is the relationship of the J48

e-containing material, and the corresponding serine-rich material from the serotype III

e-negative strain DL700, with the group B carbohydrate antigen. Two sugars, rhamnose

and galactose, known to be contained within the group B carbohydrate were detected by

carbohydrate analysis of the >250-kDa gel-purified material from sonic extracts of these

two strains. The materials were reactive with a group B-specific typing reagent

suggesting that a component of the group B carbohydrate is associated with the serine-

rich protein variants from both e-negative and e-positive strains.

A comparison of the NEM316 genome (Glaser et al. 2002) and the DY-3 sequence

(Bohnsack et al. 2002) show that although both DNA sequences contain secY2

homologues located downstream of their respective srr ORF, there is approximately

12-kb of DNA encoding putative glycosyltransferases between the two ORFs in non-III-3

strains. Only 113 nt separate srr and secY2 in the DY-3 sequence. In Staphylococcus

aureus and Staphylococcus epidermidis, secY2 lies immediately downstream of srr,

similar to the organization observed in clone DY-3, while NEM316 has an organization

more similar to that of other streptococci (Takamatsu et al. 2004). It is unknown at this








time what functions the glycosyltransferases may play in the non-III-3 strains or whether

homologues of some or all of the genes encoding these glycosyltransferases are present

elsewhere in RDP III/81-positive strains. In addition to differences in the protein

backbone of the two Srr variants, there may be differences in carbohydrate modifications

as well, with different sugars, different linkages, or both. Further work focusing on

potential differences in carbohydrate modification of RDP II1-3 and non-III-3 strains will

likely yield additional information regarding the immunologic basis for c reactivity as

well as elucidate differences in cell surface structures that may be associated with

differences in pathogenic potential.










Table 3-1. Primers used in this study.

Primer Name Sequence 5'-3' Gene Amplified Reference
(Forward/Reverse) Sequence'
KS46 (F) atggcagtatttgaaaaagtaca gbs0332 (acp) NEM316
KS47 (R) ttttactttttcttcaacataagc gbs0332 (acp) NEM316
KS13 (F) atgtcccaaaagacttttggc gbs 1529 (srr) NEM316
KS14 (R) cgtcccaaaagggttgcaccagtca gbs1529 (srr) NEM316
KS31 (F) atgtcacgcagtcaaaaagtta gbs 1514 (mutT) NEM316
KS32 (R) accatcttcaagtcgcttact gbsl 514 (mutT) NEM316
KS39 (F) tgatagttacactcaatcaagtt gbs 1516 NEM316
KS40 (R) ctgagtagagcttgaaacatg gbs 1516 NEM316
KS41 (F) aaccatggaattggttgggc gbs 1517 NEM316
KS42 (R) aaccatggaattggttgggc gbs 1517 NEM316
KS33 (F) atgacagcctttaatagtttattt gbs 1518 (secA2) NEM316
KS34 (R) aagatctccattctcattgaag gbs 1518 (secA2) NEM316
KS35 (F) cctggtgtcttgatttctgct gbsl522 (secY2) NEM316
KS36 (R) ttctcctcattttcaaatacaga gbs1522 (secY2) NEM316
KS24 (F) agctatcgcgttagcggca gbs 1526 NEM316
KS30 (R) gccttgaccatgataagttgt gbs1526 NEM316
KS26 (F) ggtgcagatttccaatatcgt gbsl527 NEM316
KS27 (R) tcatctaagaggacctacttc gbs 1527 NEM316
KS28 (F) atgcgagtacatattacaagtat gbs1528 NEM316
KS29 (R) atctataaagcatatgcacagc gbs 1528 NEM316
KS22 (F) gataccccctcactatcctt gbsl530 (rofA) NEM316
KS23 (R) gatattactgatttgagagggt gbsl530 (rofA) NEM316
KS20 (F) atagatagaaaagatactaaccg gbsl531 (uvrB) NEM316
KS21 (R) tgcagctagctcaaagtccaataat gbsl531 (uvrB) NEM316
KS51 (F) tcacgcaaagttcgagttaaaa srr DY-3
KS53 (R) cgcagatttagtagctcctaa srr DY-3
NEM316 refers to the published genome sequence by Glaser et al. (Glaser et al. 2002).
DY-3 refers to the RDP III-3 genomic clone described by Bohnsack et al.
(Bohnsack et al. 2002)











A.
1 2 3 4 5 6


1 2 3 4 5 6


S250
150
100
75
50
37

25

15
10


Figure 3-1. Western immunoblot analysis of recombinant GBS acyl-carrier protein. A)
Membrane incubated with anti-V5 monoclonal antibodies. B) Membrane
incubated with anti-8 antiserum. Lanes: 1, starting material from induced
culture, 2, wash from induced culture, 3, elution from induced culture, 4,
starting material from uninduced culture, 5, wash from uninduced culture, 6,
elution from uninduced culture.










1 2 3 4 5 6 7 8 9 10


Figure 3-2. Polymerase Chain Reaction (PCR) amplification of srr DNA from GBS
strains. DNA encoding the srr ORF was amplified using primers designed
according to the NEM316 genome database. Lanes: 1, A909 serotypee
Ic/apy8), 2, ss618 serotypee Ib/apy), 3, 9B200 serotypee II/ay), 4, J48
serotypee III/8e), 5, J52 serotypee III/8S), 6, PEH serotypee III/8c), 7, DL700
serotypee III), 8, DL1104 serotypee III), 9, COH1 serotypee III/8s), 10, COHI
serotypee III/8s, acapsular mutant of COH1).






















Figure 3-3. Detection of srr by Southern hybridization. Hybridization of DL700 serotypee
III) srr DIG-labeled PCR product with EcoRV-digested chromosomal DNA
from a panel of GBS strains. Lane 1, A909 serotypee Ic), 2, ss618 serotypee
Ib), 3, 9B200 serotypee II/ay), 4, J48 serotypee III/8~), 5, J52 serotypee III/8p),
6, PEH serotypee III/8c), 7, COH1 serotypee III/58), 8, COHI-13 serotypee
III/8c), 9, DL700 serotypee III), 10, DL1104 serotypee III), 11, pPC185
(negative control), 12, DL700 srr PCR product (positive control).