Characterization of fimbriae of Actinomyces naeslundii N16 using monoclonal and polyclonal antibodies


Material Information

Characterization of fimbriae of Actinomyces naeslundii N16 using monoclonal and polyclonal antibodies
Physical Description:
vi, 137 leaves : ill. ; 28 cm.
Bragg, Sandra L., 1947-
Publication Date:


Subjects / Keywords:
Pili (Microbiology)   ( lcsh )
Actinomyces   ( lcsh )
Periodontal disease   ( lcsh )
Microbiology and Cell Science thesis Ph.D
Dissertations, Academic -- Microbiology and Cell Science -- UF
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1988.
Includes bibliographical references (leaves 127-136).
Statement of Responsibility:
by Sandra L. Bragg.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001156735
oclc - 20606446
notis - AFQ6872
sobekcm - AA00004792_00001
System ID:

Full Text








The author wishes to express her sincere gratitude to Dr. Arnold S.

Bleiweis, chairman of her supervisory committee, for his valuable

guidance and encouragement throughout this study. She would also like

to thank the other members of her committee for their advice, and

especially Dr. Paul A. Klein, for his advice and moral support.

This study was supported by the Centers for Disease Control,

Atlanta, GA. The author would like to thank the Division of Mycotic

Diseases, especially Dr. Libero Ajello, Dr. Leo Kaufman, and Dr. Errol

Reiss for their infinite patience and cooperation.

The electron microscopy would not have been possible without the

expertise of Dr. Greg Erdos and Mary Lane Martin. The assistance of Ray

Simons and Don Howard for their excellent scientific photography is

gratefully acknowledged.

Special thanks go to Marianna Wilson; without her support and

expert assistance, completion of this dissertation would not have been




ACKNOWLEDGMENTS ................................................. ii

ABSTRACT........................................................ v

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

MATERIALS AND METHODS ........................................... 8

Antigen Preparation ........................................... 8
Cultures................................................... ... 8
Culture Conditions ................................. ........ 9
Preparation of Crude Soluble Antigens....................... 9
Preparation of Fimbriae for Purification.................... 11
Sonication............ .................................... 11
French press shearing..................................... 12
Gel Filtration Chromatography.............................. 13
Treatments of Fimbriae by Physical/Chemical Means........... 14
Antibody Preparation........................................... 17
Immunization of Mice......................................... 17
Monoclonal Antibodies....................................... 17
Isotyping................................................... 18
Polyclonal Antibodies....................................... 18
DEAE Chromatography.......................................... 19
Protein A-Sepharose Chromatography.......................... 20
Radiolabeling. ............................................. 20
Immunoaffinity Chromatography.............................. 20
Dissociation Experiment......................... ............. .. 21
Assays. ................................. ....................... 23
Electron Microscopy .......................................... 23
Hemagglutination............................................. 24
Coaggregation .............................................. 25
Bacterial Agglutination.................. ......... .... .... 26
Radioimmunoassay.................................. ........ 26
Indirect Enzyme Immunoassay................................ 27
Enzyme Immunodot Assay..................................... 28
Immunodiffusion ............................................. 28
Laurell Rocket Immunoelectrophoresis....................... 29
Crossed Immunoelectrophoresis with Autoradiography.......... 29
Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis... 29
Immunoblot.................................................. 30


RESULTS AND DISCUSSION .......................................... 31

Monoclonal Antibodies........................................ 31
Evidence for the Presence of Both Types
of Fimbriae on A. naeslundii Serotype 3 Strains............. 35
Evidence for the Presence of Type 1 Fimbriae ................ 35
Evidence for the Presence of Type 2 Fimbriae................ 43
Antigen Preparation: Results of Preliminary Experiments....... 46
Identification of N16 Type 1 and Type 2 Fimbriae
in Crude Antigen Extracts by XIEP-A......................... 49
Purification of N16 Type 2 Fimbriae from a Crude Sonicate..... 51
Batch A. .................................................... 51
Batch B...................................................... 54
Purification of N16 Fimbriae from the
French Press Supernatant................................... 60
Effects of Various Physical and Chemical Treatments
on N16 Fimbriae............................................. 64
Effects of Temperature and Reduction on N16 Fimbriae.......... 77
Assessment of the Purity
of Fimbrial Samples by SDS-PAGE-Immunoblot................... 83
Antigenic Relatedness of Actinomyces Fimbriae ................. 90
Ouchterlony Analysis........................................ 90
XIEP-A....................................................... 92
Bacterial Agglutination.................................... 106
Immunoblot Analysis....................................... 115

SUMMARY AND CONCLUSIONS ......................................... 119

REFERENCES...................................................... 127

BIOGRAPHICAL SKETCH....................................... ...... 137

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



Sandra L. Bragg

August 1988

Chairman: Dr. Arnold S. Bleiweis
Major Department: Microbiology and Cell Science

Two populations of fimbriae, which differ both in antigenicity and

biological activity, have been identified on Actinomyces viscosus T14V

cells. Although A. naeslundii serotype 1 isolates possess only one of

these fimbrial populations (type 2 fimbriae), there was functional

evidence to suggest that A. naeslundii serotype 3 strain N16 had both

types of fimbriae. The purpose of this study was to characterize the

fimbriae of A. naeslundii N16 immunologically by using both monoclonal

and polyclonal antibodies.

Three monoclonal antibodies (MAbs) to N16 were produced; all three

bound to N16 fimbriae as determined by immunoelectron microscopy. In a

solid-phase radioimmunoassay MAb 3B5.A1 reacted with 100% of the

A. naeslundii serotype 3 isolates tested, but it did not react with any

heterologous isolates. Type 1 and type 2 fimbriae were detected in

Lancefield extracts of N16 cells by crossed immunoelectrophoresis (XIEP)
using rabbit antiserum against N16 whole cells. When 125I-MAb 3B5.A1

was also incorporated into the gel, autoradiography indicated that MAb

3B5.A1 was specific for type 2 fimbriae. The N16 type 2 fimbriae were

purified by gel filtration and immunoaffinity chromatography on a MAb

3B5.A1 column.

Polyclonal antisera specific for N16 type 1 or type 2 fimbriae were

produced by immunizing rabbits with fimbrial immunoprecipitins excised

from XIEP gels. These antisera were used to identify fimbrial bands on

immunoblots of SDS-PAGE resolved N16 fimbrial antigens. Although N16

fimbriae could not be completely dissociated, type 1 fimbrial subunits

at 65 kd, 57-60 kd, and a weaker doublet at 53-54 kd and type 2 subunits

at 63 kd and a doublet at 39-40 kd were identified in fimbrial extracts

reduced at 1000C.

Fimbriae-specific polyclonal and monoclonal antibodies were used in

various immunological assays to determine that (a) N16 type 1 fimbriae

are not related antigenically to type 2 fimbriae, (b) each type of

fimbriae has epitopes that are present on the corresponding fimbriae of

certain heterologous strains, and (c) MAb 3B5.A1 recognizes a

serotype-specific epitope residing on the type 2 fimbriae of

A. naeslundii serotype 3 strains.


Many structures found on the cell surfaces of microorganisms

exhibit properties that enable microorganisms to cause disease. These

properties include (a) adherence to host surfaces and other microbes,

(b) evasion of host defenses, and (c) destruction of host tissues. In

addition, surface molecules carry antigenic determinants that form the

basis of many serological and immunological identification

procedures. The fimbriae of certain species of Actinomyces are one

example of surface components that not only mediate adherence both to

host surfaces and to other microbes but also possess important

antigenic determinants.

Actinomyces are gram-positive, nonacidfast, nonsporeforming,

nonmotile bacteria that are highly variable in morphology, most

characteristically diphtheroidal, or filamentous and branched. They

ferment carbohydrate with production of acid but no gas, and except

for A. viscosus, all species are catalase-negative. Their cell walls

do not contain diaminopimelic acid, a characteristic that

differentiates them from Arachnia propionica, a morphologically

similar organism. Their natural habitat is the oral cavity of man and

other animals (96). All Actinomyces species, as well as Arachnia

propionica, are potential agents of actinomycosis (45).

Classical actinomycosis is a chronic granulomatous disease

characterized by the formation of abscesses and draining sinuses. It

is an endogenously acquired infection that can involve the soft tissue

and bone in any area of the body. Clinically, actinomycosis is

usually categorized as cervicofacial, thoracic, or abdominal. Of

these categories, cervicofacial actinomycosis is the most common, and

Actinomyces israelii is the most important etiological agent of

actinomycosis in humans. However, the extraoral infections of

classical actinomycosis are not as prevalent as intraoral periodontal

infections, and it has been suggested that the clinical classification

scheme be broadened to include a periodontal category (56).

During the last twenty-five years, most of the research efforts

dealing with Actinomyces species have been directed toward trying to

delineate the role that Actinomyces species play in the formation of

dental plaque and the development of periodontal disease. In the

early 1960s A. viscosus was found to be the etiologic agent of a

transmissible periodontal disease in Syrian hamsters (59,65). That

discovery initiated an explosion of research activity focused on the

pathogenicity of Actinomyces species in the oral cavity of man and

other animals because periodontal disease, unlike classical

actinomycosis, was and still is a major public health problem.

Virtually all humans and many animals accumulate dental plaque on

their teeth, and the formation of plaque may be followed by the

development of both caries and periodontal disease (49).

Periodontal disease is a collective term for a variety of chronic

inflammatory diseases of the structures that support the teeth.

Periodontal disease can range from mild inflammation of the marginal

gingiva (gingivitis) to severe forms of periodontitis, in which

extensive destruction of soft tissues and resorption of alveolar bone

can result in loss of the teeth. Although the microbial etiology and

pathology of the different types or stages of periodontal disease may

vary (98), the formation of dental plaque is the common first step in

the development of periodontal disease (49). However, successful

colonization of the tooth surface or other surfaces of the oral cavity

depends upon the ability of oral microbes to anchor themselves, either

directly or indirectly, to these surfaces to avoid being swept away by

the flow of saliva (47,104). Thus, an understanding of the mechanisms

by which microbes attach to surfaces in the oral cavity and to other

microbes is crucial to understanding the sequence of events leading to

periodontal disease.

Studies with experimental animals indicated that both A. viscosus

and A. naeslundii could form dental plaque and initiate the

pathological changes associated with periodontal disease

(55,58,59,60,97). However, most of the attention has been focused on

A. viscosus, particularly in studies on mechanisms of attachment to

oral surfaces and other bacteria (6,13,14,16,19,21,22,27,48,72,81,82,

92,106,107) and mechanisms for tissue destruction (1,8,9,10,39,40,51,


There were two factors that contributed to the placing of more

emphasis on A. viscosus than A. naeslundii as a subject for

periodontal research. One was the recognition that these two species

tended to be distributed differently in the human oral cavity in that

A. viscosus preferentially colonized the teeth, whereas A. naeslundii

was associated with mucosal epithelial surfaces (34,35). The second

was the existence of the T14V-T14AV model system for comparative

studies of the properties that enable T14V to be virulent (i.e., cause

periodontal disease) in experimental animals and T14AV to be avirulent

(6,51). Although it was soon determined that the virulence-

associated differences in the antigens of T14V and T14AV were

quantitative rather than qualitative (19,22,90), T14V became the focus

of intensive study because it was found to have two different

adherence-related properties. These were the ability to coaggregate

with Streptococcus sanguis via a lactose-inhibitable lectin (82) and

the ability to adsorb well to saliva-coated hydroxyapatite (24).

A. naeslundii, on the other hand, coaggregated streptococci (36) but

did not bind well to saliva-coated hydroxyapatite (24). Some of these

early studies established an association between the presence of

fibrils or fimbriae on the surface of A. viscosus cells and

adherence-related functions (19,22,82).

Fimbriae are proteinaceous surface appendages that are found on

many gram-negative and gram-positive bacteria and that often mediate

attachment to host surfaces. Fimbriae from five strains of

A. viscosus and two strains of A. naeslundii have been isolated and

characterized as to their amino acid compositions (21,76,77,106).

Polar uncharged and nonpolar amino acids together made up 62-74% of

the total amino acids, whereas basic amino acids accounted for only

10-17% of the total; aspartic and glutamic acid generally comprised

20-24% of the total amino acids (76). The minimum molecular weight of

a fimbrial subunit as calculated from amino acid data was determined

to be approximately 25 kd for T14V (type 1) fimbriae (106) or 30 kd

for A. viscosus WVU627 fimbriae (77). However, a method for

completely dissociating Actinomyces fimbriae into subunits has not

been discovered (12).

Over the last decade, the fimbriae of A. viscosus and

A. naeslundii have been the subject of intensive study. A great deal

of progress has been made in identifying the adherence-related

functions of the fimbriae and defining the molecular basis of the

interactions of certain Actinomyces cells with other microorganisms

and surfaces within the oral cavity. Nearly all of the studies have

focused on only two strains, A. viscosus serotype 2 strain T14V and

A. naeslundii serotype 1 strain WVU45, and spontaneously occurring

mutants derived from these two strains. Two populations of

Actinomyces fimbriae, which differ both in antigenicity and biological

activity, have been identified (20,84). Type 1 fimbriae mediate

adherence to saliva-coated hydroxyapatite in vitro and to the salivary

pellicle on teeth in vivo (23). Type 2 fimbriae via their

lactose-inhibitable lectin activity mediate adherence to other oral

bacteria and mammalian cells (11,12,15,44,68,80,93,103). The

functional and molecular properties of the fimbriae from these two

strains have been recently reviewed (84).

Both types of fimbriae are present on T14V and other serotype 2

isolates of A. viscosus, whereas WVU45 and other A. naeslundii

serotype 1 isolates possess only type 2 fimbriae (20). In 1974 Jordan

et al. (57) described a new serotype of A. naeslundii based on their

studies of strain N16 and 20 similar isolates. These strains had been

isolated from cervical plaque in a sampling of 59 Down's syndrome

patients who had moderate to severe periodontal disease (64). In

studies with gnotobiotic rats and hamsters, Jordan and co-workers have

shown that N16 can cause heavy plaque deposits and severe periodontal

pathology, including alveolar bone loss, root surface caries,and even

enamel caries in some animals (57,60). Strain N16 and isolates that

are biochemically and serologically identical to it have been

designated A. naeslundii serotype 3 (46).

There is both functional and immunological evidence to suggest

that A. naeslundii serotype 3 strain N16 has both types of fimbriae.

The N16 isolate exhibits lactose-reversible coaggregation with

Streptococcus sanguis 34 (18). It produces its own neuraminidase and

agglutinates human erythrocytes in a reaction that is reversed by

lactose (29). It possesses numerous surface fibrils or fimbriae and

adsorbs to saliva-treated hydroxyapatite (SHA) in vitro as avidly as

T14V (26). In fact, Clark et al. (25) demonstrated that on average

A. naeslundii serotype 3 adsorbed better to SHA than any other species

or serotype of Actinomyces tested. Strain N16 agglutinated with MAbs

against T14V type 2 fimbriae (14). It reacted with rabbit IgG against

T14V type 1 fimbriae in a dot enzyme immunoassay (28), and its

adsorption to SHA was partially inhibited by the antibody (23,28).

Although most of our knowledge of Actinomyces fimbriae has been

based on studies of A. viscosus serotype 2 T14V and A. naeslundii

serotype 1 WVU45, A. naeslundii serotype 3 was the Actinomyces species

most frequently isolated from subgingival dental plaque from sites

with moderate, severe, or juvenile periodontitis (87), and in another

study A. naeslundii serotype 3 was considered to be one of the most

likely etiological agents of gingivitis (86). Thus, A. naeslundii

serotype 3 strain N16 appears to be a logical choice for a study of

surface antigens of a periodontopathic actinomycete.


The goals of this study were to (a) produce monoclonal antibodies

against surface antigens of A. naeslundii serotype 3 strain N16, (b)

select at least one MAb that exhibited serotype specificity, and (c)

use the MAb as a tool to isolate and characterize the surface

component carrying the serotype-specific epitope.


Antigen Preparation


A. naeslundii M16 (WVU820), WVU1267, WVU1468, WVU1527, and

WVU1528 were obtained from M. A. Gerencser, West Virginia University.

W1629, W2273, and W2821 were obtained from the Centers for Disease

Control, Atlanta, GA. UF92 and UF524 were fresh clinical isolates

provided by J. E. Beem, University of Florida, Gainesville, FL. The

identity of each of these ten isolates was confirmed by biochemical

tests and by direct staining with serotype-specific fluorescent

antibody reagents; all ten isolates were designated A. naeslundii

serotype 3. Approximately 50 additional isolates, including

representatives of all the recognized serotypes of A. naeslundii,

A. viscosus, A. israelii, A. odontolyticus, Arachnia propionica,

Rothia dentocariosa, and Bacterionema matruchotii were obtained from

culture collections at the Centers for Disease Control for use as

heterologous organisms in evaluating the specificity of monoclonal

antibodies. Other strains used in this study were kindly provided by

P. E. Kolenbrander, National Institute of Dental Research, Bethesda,

MD, and by W. B. Clark and P. J. Crowley, University of Florida.

Stock cultures of all isolates used in this study were maintained by

lyophilization in skim milk and by storage at -200C in Trypticase

soy broth (TSB; BBL Microbiology Systems, Cockeysville, MD) containing

20% (v/v) glycerol.

Culture Conditions

Cells for all procedures were obtained by culturing the isolates

in TSB supplemented with 2.5 g of K2HPO4 per liter in flasks that

could be chemically sealed to provide the atmospheric environment

appropriate for each isolate. After inoculation of a TSB-containing

flask with an actively growing TSB culture (2% v/v), the flask was

plugged with a rubber stopper through which an open-ended screw-capped

tube filled with cotton had been inserted (7). Prior to tightening of

the screwcap, the cotton was saturated with equal volumes of 10%

Na2CO3 and 1 M KH2PO4 to generate an aerobic + CO2

environment or with 10% Na2CO3 and pyrogallol solution to generate

an anaerobic + CO2 environment (45). The cultures were incubated

without shaking at 37 C. Cells were harvested from cultures in mid

exponential to early stationary phase by centrifugation (10,000 x g,

10 min), washed twice, and stored in buffer at 40C. The buffers

used for washing and making cell suspensions varied depending on the

procedure for which the cells were intended; they are noted in each

specific method.

Preparation of Crude Soluble Antigens

Several different methods were used to obtain crude soluble

antigens from whole cells.

Ammonium sulfate fractionation of French pressure cell

supernatants. N16 cells harvested from 8.5 1 of 36 h TSB cultures

were suspended in 0.05 M phosphate buffer, pH 7.2, to form a thick

paste; then the cell suspension was passed once through a French

pressure cell (Aminco Model J4-3337, American Instrument Co., Silver

Springs, MD) at 10,000 psi to remove the fimbriae (106). The


pressed cell suspension was centrifuged (48,300 x g, 20 min) to pellet

intact cells and cell walls; then the crude supernatant was

ultracentrifuged at 160,000 x g for 24 h to pellet the fimbriae. The

pellet was partially resuspended in 0.1 M Tris-HCl, pH 7.5, by

sonicating for 1 min at full power with a Kontes microultrasonic cell

disruptor (Kontes, Vineland, NJ). After centrifugation at 23,700 x g

for 10 min, the fimbriae-containing supernatant was processed by

ammonium sulfate fractionation to obtain precipitates at 10, 20, 30,

40, and 50% (w/v) (NH4 )2SO4. The precipitates were collected by

centrifugation (30,900 x g, 15 min), dissolved in and dialyzed against

deionized water, and lyophilized. Lyophilized antigens were dissolved

in 0.5% Triton X-100 by adding 100 4l to 2 mg dry weight and

sonicating for 10 sec. The lyophilized material was highly resistant

to solubilization, so lyophilization was not used on subsequent

batches of fimbriae.

Continuous flow sonication. N16 cells (10% packed cells by

volume) in Tris-buffered saline (TBS; 0.025 M Tris-HCl, 0.15 M NaCI,

10-4 M CaCI2, 10-4 H MgC 2, and 0.02% NaN3, pH 7.6) were

passed at a flow rate of 5 ml/min three times through a continuous

flow sonifier cell disruptor (Heat Systems Ultrasonics, Inc.,

Farmingdale, NY) set at 100 W (14).

Virtis homogenization. A suspension of N16 cells in buffer was

agitated twice at 23,000 rpm for 5 min in a Virtis Model 45

homogenizer (The VirTis Company, Inc., Gardiner, NY) (77).

Extraction with 8 M lithium chloride. N16 cells were suspended

in 8 H LiCI, pH 7.0, (10% packed cells by volume) and stirred

continuously for 1 h at 25C (82).

Lancefield extraction. N16 cells were suspended in 0.04 N HC1 in

saline (5 ml per g wet weight of cells); the suspension was heated for

15 min in a boiling water bath, cooled, and neutralized (90).

For the last four methods described above, the cells and

supernatants were separated by centrifugation (10,000 x g, 10 min),

and the supernatants were concentrated by ultrafiltration (PM10

membrane; Amicon Corp., Danvers, MA). The protein content of each

sample was determined by the Bio-Rad Protein Assay, a Coomassie blue

dye binding assay (3) (Bio-Rad Laboratories, Richmond, CA). Total

carbohydrate was estimated by the phenol-sulfuric acid procedure (33).

Preparation of Fimbriae for Purification


Batch A. N16 cells were harvested from 12.5 1 of 36 h TSB

cultures, washed, and resuspended in TBS to a total volume of 240 ml

(10-20% v/v). Fimbriae were removed by sonicating 30 ml aliquots in

15 sec bursts at 100 W. After 0.5, 1, 2, 3, 4, 5, and 6 min of

sonication, cells and supernatants were separated by centrifugation

(10,000 x g, 10 min). The cells were resuspended in fresh buffer, and

a few drops of the suspension were set aside for subsequent

examination by hemagglutination (HA) and electron microscopy (EM)

before the next cycle of sonication was started. All the supernatants

were pooled and centrifuged (27,000 x g, 30 min) to remove

particulates. The crude 6 min sonicate was concentrated to 10 ml by

ultrafiltration (Amicon PM10 membrane) and by addition of dry Sephadex

G-25 prior to chromatography on Bio-Gel A-5m.

Batch B. N16 cells were harvested from 27.5 1 of 24-29 h TSB

cultures, washed, and resuspended in TBS to a total volume of 750 ml

(10-20% v/v). Fimbriae were removed from 720 ml of cell suspension by

sonicating 30 ml aliquots in 30 sec bursts at 100 W. After sonication

for a total of 2 min, the cell suspensions were centrifuged at 10,000

x g for 10 min to separate intact cells from supernatants. The

sonicated cells were resuspended in fresh buffer for examination by

HA. The supernatant or crude 2 min sonicate was further clarified by

centrifugation (27,000 x g, 30 min), then concentrated by

ultrafiltration (Amicon PM10 membrane) to 180 ml. The Amicon

concentrated sonicate was mixed with an equal volume of saturated

(NH) 2SO4 to give 50% saturation. After incubating at 4 C

overnight, the precipitate was collected by centrifugation (10,000 x g,

30 min), dissolved in water, and re-precipitated with (NH4)2SO4

at 35% final saturation. The precipitate from the 35% (NH4)2SO4

saturation was collected by centrifugation, resuspended in water, and

dialyzed against TBS. This was the starting material for gel

filtration chromatography.

French press shearing

N16 cells were harvested from 30 1 of 16-24 h TSB cultures,

washed once in TBS, and resuspended in TBS (5 ml of TBS per I of

culture harvested). The cell suspension was passed once through a

French pressure cell (Model J4-3337; American Instrument Company,

Silver Spring, MD) at 10,000 psi.

The amount of lysis caused by the French press was estimated to

be 1.2%, as determined by the following equation:

% lysis = (A260/A280 of untreated) (A260/A280 of treated) X 100
A260/A280 of untreated

Intact cells were removed from the pressed cell suspension by

centrifugation at 10,000 x g for 10 min. The French press supernatant

was clarified further by centrifugation at 48,300 x g for 20 min; the

clear supernatant was stored at -80 C until processed.

Gel Filtration Chromatography

Columns were packed with Bio-Gel A-5m (exclusion limit 5 x 106

daltons for globular proteins; Bio-Rad Laboratories). The void and

total volumes were determined with blue dextran 2000 and potassium

ferricyanide, respectively. Crude sonicates were applied to the

columns and eluted with TBS. A Marriotte flask was used to maintain a

constant operating pressure. For all chromatographic techniques,

effluents and eluates were monitored by A280 with a Uvicord II UV
analyzer and recorder, and fractions were collected with an Ultrorac

fraction collector (LKB Instruments Inc., Rockville, MD). The Batch A

sonicate was run on a 12 x 750 mm column packed to a bed volume of

77.5 ml; the flow rate was 8 ml/h, and 3 ml fractions were collected.

The Batch B sonicate, a 35% saturated (NH4)2SO4 fraction, was

applied in several aliquots (20 ml each at 3.9 mg protein per ml) to a

25 x 1000 mm column packed to a bed volume of 410 ml; the flow rate

was 25 ml/h, and 5 ml fractions were collected. Column fractions were

assayed for (a) total protein by the Bio-Rad Protein Assay with bovine

albumin as the standard, (b) total carbohydrate by the phenol-sulfuric

acid method with glucose as the standard, and (c) fimbriae reactive

with MAb 3B5.Al by an enzyme immunodot assay. Some fractions were

also examined by SDS-PAGE.

Treatments of Fimbriae by Physical/Chemical Means

Acetone precipitation. Five volumes of cold acetone were added

to the sample. After having been mixed thoroughly, the sample was

incubated at -200C for at least 10 min. Precipitates were collected

by centrifugation (10,000 x g, 5 min) and dried by evaporation under

vacuum. Precipitates were reconstituted in an appropriate buffer.

Magnesium chloride precipitation. Four parts of 1 M MgC12 were

mixed with one part sample to obtain a final concentration of 0.1 M

MgCl2. The sample was incubated at 4 C for 24 h, then examined

for evidence of precipitation.

Freon extraction. Equal volumes of sample and Freon 113

(trichlorotrifluoroethane) were mixed thoroughly by vortexing. After

centrifugation (2000 x g, 30 min), the upper aqueous phase was


Heat. Samples were placed in a waterbath at 37 C for 60 min,

at 65 C for 30 min, or at 100C for 5 min or 60 min.

Sonication. Samples in capped microfuge tubes were placed in an

ice slurry in a cup horn sonicator (Heat Systems-Ultrasonics, Inc.,

Plainview, NY) and sonicated continuously at full power for 1, 5, or

10 minutes.

Urea. An acetone precipitate of the sample was dissolved in a

volume of 8 M urea in 0.05 M Tris-HC1, pH 8.0, equal to the original

volume of sample; then it was incubated at 37 C for 1 h.

Guanidine hydrochloride. An acetone precipitate of the sample

was dissolved in saturated (8.6 M) guanidine HC1 at a final

concentration of 1 ml of guanidine HC1 per mg total protein. After

incubation at 37 C for 1 h, the sample was diluted with and dialyzed

against 10 mM EDTA, 0.05 M Tris-HCl, 0.15 M NaCI, 0.1% NaN3, pH 7.5.

Acid. One aliquot was treated by Lancefield extraction; it was

adjusted to 0.04 N HCI in saline and heated at 100 C for 15 min.

Two others were adjusted to 0.1 N HCI and heated at 1000C for 5 min

or at 37 C for 1 h. All were neutralized by addition of NaOH after


Base. The sample was adjusted to 0.1 N NaOH and incubated at

37 C for 1 h; then it was neutralized with HC1.

Periodate oxidation. The sample was adjusted to 0.1 M sodium

metaperiodate in 0.05 M acetate buffer, pH 4.5, and incubated in the

dark at 4 C for 24 h; then ethylene glycol was added to 0.3 M final

concentration to consume excess periodate.

Enzymatic digestions. Solutions of lysozyme (Cat. No. 36-324,

Miles Laboratories, Naperville, IL), mutanolysin (M-3765, Sigma

Chemical Co., St. Louis, MO), papain (Sigma P-4762), and Pronase (Cat.

No. 537088, Calbiochem, San Diego, CA) were prepared in 0.1 M sodium

phosphate, pH 6.2. Solutions of a-chymotrypsin, trypsin (Sigma

T-8253, Type III), Staphylococcus aureus V8 protease (Sigma P-8400,

Type XVII), and proteinase K (Sigma P-0390, Type XI) were prepared in

0.05 M Tris-HCl, pH 8.0. Acetone precipitates of crude N16 French

press supernatants were dissolved in 0.1 M phosphate-buffered saline,

pH 6.2, or 0.05 M Tris-buffered saline, pH 8.0, depending on the

enzyme to be used.

Antigen samples containing 25 ug of total protein were

incubated (37 C, 30 min) with serial dilutions of lysozyme ranging

from 0.1 yg to an upper limit of 7 Vg (175 units) of lysozyme per

sample. Likewise, samples were treated with mutanolysin in amounts

ranging from 0.3 Units to 17.5 Units. To ensure that the lysozyme and

mutanolysin were active, a suspension of A. naeslundii PK19 cells was

treated with each enzyme, and muramidase activity was detected as a

decrease in turbidity (A 600) of the cell suspension.
Samples that were to be treated with proteases were heated at

100 C for 2 min in the presence of 0.5% SDS prior to the addition of

enzyme solutions in order to make the proteins more susceptible to

digestion and to inactivate any endogenous proteases that might be

present in the crude antigen sample. Samples containing 500 pg of

total protein were mixed with 50 Vg of enzyme (chymotrypsin,

trypsin, papain, Pronase, proteinase K, or V8 protease) for a 10:1

ratio of protein:enzyme, and the appropriate buffer was added to bring

the final enzyme concentration to 100 ig per ml. The

protease-treated samples were incubated in a 37 C waterbath for 1 h.

At the end of the incubation period each sample was prepared

immediately for SDS-PAGE and stored at -800C until it could be


Aliquots of bovine serum albumin were treated exactly like the

N16 samples as a positive control for activity of the proteases and

for detection of protease contamination in the muramidases. Samples

that received buffer in place of enzyme served as untreated controls.

Samples containing enzyme only were also used as controls.

Treated samples were subjected to SDS-PAGE; the SDS-PAGE resolved

proteins were transferred to nitrocellulose. These blots were reacted

with either R10 rabbit IgG against N16 type 1 fimbriae or R2P rabbit

IgG against N16 type 2 fimbriae in an indirect EIA to visualize

immunoreactive fimbrial bands.

Antibody Production

Immunization of Mice

The cell suspensions for immunizations were formalin-killed whole

cells suspended in phosphate-buffered saline (PBS; 0.01 M phosphate,

0.85% NaCI, pH 7.4) containing 0.3% formalin. Cell aggregates were

dispersed by sonication, and the turbidity of each cell suspension was

adjusted to a No. 8 McFarland Standard. To obtain antigen-primed

spleen cells for fusions, 10-week old female BALB/c mice were

immunized with A. naeslundii N16 cells. Two mice that received 0.1 ml

intravenously on days 1, 8, and 16 were sacrificed on day 20 for

fusion of their spleen cells with X63Ag8.653 myeloma cells. Two

additional mice received 0.1 ml mixed with an equal volume of Freund's

complete adjuvant intraperitoneally on day 1 and 0.1 ml without

adjuvant intravenously on day 19 and were sacrificed on day 22 for a

fusion with P3X63Ag8 cells.

Monoclonal Antibodies

Hybridomas were produced, using the method of Simrell and Klein

(95), by fusing 108 spleen cells from N16-immunized mice with 107

myeloma cells in the presence of 50% polyethylene glycol-1000. For

one fusion myeloma cell line P3X63Ag8, which secretes immunoglobulin

molecules with gamma-1 heavy chains and kappa light chains (66), was

used. For another fusion, the Kearney myeloma cell line X63Ag8.653,

which does not express any immunoglobulin chains (63), was used.

Hybrid cells were selected by growth in Dulbecco's modified Eagle's

medium (DMEM) containing hypoxanthine, aminopterin, and thymidine (HAT

medium). Culture supernatants were screened for specific antibody by

solid-phase radioimmunoassay (RIA). Hybridomas secreting antibody to

N16 were cloned by limiting dilution by seeding a 96-well tissue

culture plate at a density of one-half cell per well; each well

contained 5 x 104 mouse peritoneal exudate cells as a feeder layer.

Three different clones were selected for further evaluation. They

were propagated in tissue culture and as ascites tumors in female

BALB/c mice primed by intraperitoneal injection of 0.5 ml pristane

(2,6,10,14-tetramethylpentadecane; Aldrich Chemical Company, Inc.,

Milwaukee, WI). Culture supernatants and sera or ascites fluids from

hybridoma-bearing mice were stored at -20 C. Hybridoma cells in

DMEM containing 30% fetal calf serum and 10% dimethylsulfoxide were

stored in liquid nitrogen.


The class and subclass of immunoglobulin secreted by each

hybridoma were determined initially by immunodiffusion in which

subclass-specific anti-heavy chain antisera (Meloy Laboratories, Inc.

Springfield, VA) were reacted with the hybridoma culture

supernatants. These results were confirmed subsequently by testing

hybridoma culture supernatants or ascites fluids with an enzyme

immunoassay isotyping kit (Mouse Typer, Bio-Rad Laboratories).

Polyclonal Antibodies

Polyclonal antisera monospecific for type 1 and type 2 fimbriae

of A. naeslundii N16 were produced in female New Zealand white rabbits

by immunization with immunoprecipitins cut from crossed

immunoelectrophoresis (XIEP) gels (71). Type 1 and type 2 fimbriae

from partially purified samples of N16 fimbriae were separated and

precipitated by XIEP versus rabbit antiserum R29 raised against N16

whole cells. Sections of gel containing the immunoprecipitate were

cut from several XIEP patterns. The appropriate segments (a minimum

total of 10 cm per immunogen) were pooled, washed exhaustively with

saline to remove excess soluble reactants, and solubilized in 1 ml of

6 M KI. One-fourth of the immunogen was mixed with an equal volume of

complete Freund's adjuvant and injected subcutaneously into multiple

sites. The remainder of the immunogen was emulsified with an equal

volume of incomplete Freund's adjuvant and administered in 3

subcutaneous injections at approximately weeks 3, 5, and 7. Three

rabbits received type 1 fimbrial immunoarcs, and three were immunized

with type 2 fimbrial immunoarcs. Antibody responses were monitored by

microtiter plate EIA, and additional injections of immunogen were

given as needed to boost or maintain antibody titers. The rabbits

were bled periodically from the central artery of the ear throughout

the schedule and were exsanguinated by cardiac puncture 3-9 months

after the first injection.

R10 against N16 type 1 fimbriae and R2P against N16 type 2

fimbriae were the two rabbit antisera used most often in this study.

J. O. Cisar, National Institute for Dental Research, kindly provided

the following samples of monospecific rabbit IgG: R59 against

A. viscosus T14V type 1 fimbriae, R55 against T14V type 2 fimbriae,

and R70 against A. naeslundii WVU45 type 2 fimbriae. Rabbit antiserum

raised against immunoarcs of Histoplasma capsulatum was kindly

provided by P. Standard, Division of Mycotic Diseases, Centers for

Disease Control, for use as a negative control.

DEAE Chromatography

IgG was purified from rabbit antisera or hybridoma culture

supernatants by chromatography on DEAE Bio-Gel A (Bio-Rad

Laboratories). Antisera and supernatants were precipitated with

(NH )2SO4 at 50% saturation. The precipitates were dissolved

and equilibrated with 0.01 M phosphate buffer, pH 7.8. Samples were

applied to DEAE Bio-Gel A columns, and IgG was eluted with

equilibrating buffer. Purified IgG was concentrated by

ultrafiltration and stored at 4C with 0.1% NaN3 as preservative.

Protein A-Sepharose Chromatography

Protein A-Sepharose (Pharmacia Fine Chemicals, Piscataway, NJ)

was packed in a column with flow adaptors. Hybridoma culture

supernatants or rabbit antisera were precipitated with ammonium

sulfate at 50% saturation. The precipitates were dissolved in water

and equilibrated against starting buffer. Samples were applied and

the unbound fraction eluted in 0.05 M Tris-HCl, 0.25 M NaCI, 0.1%

NaN3, pH 8. IgG was eluted with 7 H urea in 0.05 M Tris-HC1, pH 8.

Purified immunoglobulins in PBS and Na2 I (5 VCi per ig

protein) were placed in glass tubes coated with IODO-GEN

(1,3,4,6-tetrachloro-3a,6a-diphenylglycoluril; Pierce Chemical

Co., Rockford, IL). The samples were incubated for 10 min at 25 C,

and then removed from the reaction vessel to terminate the iodination.
125 125
125I-labeled Igs were separated from free Na I by gel filtration

on PD-10 columns (Pharmacia) equilibrated with PBS containing 1%
bovine serum albumin and 0.1% NaN3. All 1I-labeled samples

contained 90-100% TCA-precipitable counts per minute.

Immunoaffinity Chromatography

MAb 3B5.A1, purified from hybridoma culture supernatants by

protein A-Sepharose chromatography, was coupled to Affi-Gel 10

(Bio-Rad) and packed in a K16/20 column with flow adaptors

(Pharmacia); bed volume was 9.5 ml. Samples were applied by

continuously cycling them through the column in an ascending direction

at a flow rate of 10 ml/h. Unbound and bound fractions were eluted at

a flow rate of 40 ml/h in an ascending or descending direction,

respectively. The unbound fraction was eluted with 0.05 M Tris-HCl,

0.25 H NaCI, 0.1% NaN3, pH 8. The bound fraction was eluted with 7

M urea in 0.05 M Tris-HC1, pH 8.

Rabbit IgG, 250 mg in 0.1 M HEPES, pH 7.5, purified by DEAE and

protein A-Sepharose chromatography from antisera monospecific for N16

type 1 or type 2 fimbriae, was coupled to 50 ml of Affi-Gel 10 and

packed in a K26/40 column with flow adaptors. The coupling efficiency

for each gel was >90%. Initially samples were eluted as described for

the MAb immunoaffinity column; however, on the basis of results of a

dissociation experiment, 6 M NaSCN was substituted for 7 M urea in the

buffer for elution of the bound fimbriae.

Dissociation Experiment

For protein A-Sepharose chromatography and initial experiments

with immunoaffinity chromatography, 7 M urea was used as the

dissociating buffer for release of bound molecules from the ligand.

However, to experimentally determine the most effective dissociating

buffer to use with each immunoaffinity column, various dissociating

buffers were evaluated in a microtiter plate EIA for their efficacy in

releasing antibody bound to antigen.

The wells of a 96-well Immulon 2 plate (Dynatech Laboratories,

Inc., Alexandria, VA) were coated with a crude fimbriae-containing

sonicate of N16 cells. Aliquots of the antibody samples used to

prepare the three immunoaffinity columns, i.e., MAb 3B5.A1, rabbit IgG

anti-N16 type 1 fimbriae, and rabbit IgG anti-N16 type 2 fimbriae,

were the antibody samples tested in the dissociating experiment; the

optimal dilution of each antibody had been determined by indirect EIA

versus the fimbriae-coated plates. An indirect EIA was performed as

described below, except that prior to the addition of the enzyme

conjugate, each well received 200 ul of a dissociating or control

buffer for 1 h at 25 C. Buffers were evaluated in triplicate, and

the average A490 of the triplicate samples was used in the following

formula for measuring the efficacy of each buffer in dissociating the

antigen-antibody bonds:

% release of Ab 1 (A490 with dissociating buffer)
from fimbriae = (A490 with control buffer) X 100

The control buffer was the buffer normally used to elute the unbound

fraction during immunoaffinity chromatography, i.e., 0.05 M Tris-HC1,

0.25 M NaCI, 0.1% NaN3, pH 7.5 at 25C.

In order to rule out the possibility that the dissociating

buffers themselves were removing the fimbriae from the plate or

adversely affecting the fimbriae in some other manner rather than

merely dissociating the antibody, an indirect EIA was performed on a

fimbriae-coated plate that was first incubated with dissociating

buffers at maximum concentration overnight at ambient temperature. If

the dissociating buffers were releasing fimbriae from the plates or

irreversibly denaturing fimbrial epitopes, the A490 of wells that

were incubated with dissociating buffers would be lower than that of



Electron Microscopy

For electron microscopy (EM) of whole cells, the specimens were

prepared by the pseudoreplica technique as described by Martin et al.

(75), using 0.5% uranyl acetate for negative staining. For

immunoelectron microscopy (IEM), N16 cells were incubated with a 1:100

dilution of rabbit antiserum overnight at 4 C, washed with pH 7.2

PBS, and incubated (2 h, 25C) with a 1:10 dilution of

gold-conjugated goat anti-rabbit IgG (Auroprobe EM GAR G10, Janssen

Life Sciences Products, Piscataway, NJ). Then the gold-labeled cells

were prepared for EM and negatively stained as described above.

For thin section IEM colloidal gold (diameter, approximately 15

nm) was prepared by citrate reduction of chlorauric acid (43) and

coupled to goat anti-mouse IgG by the method of Horisberger and Rosset


N16 cells were grown in brain heart infusion broth (48 h,

37 C), then harvested and washed with pH 7.2 PBS by vacuum

filtration on Nucleopore filters (45 p pore size). The cells were

reacted with monoclonal antibody by floating the filters on hybridoma

culture supernatant (immune sample) or on myeloma cell line P3X63Ag8

culture supernatant (negative control) for 1 h at 25 C. The filters

were washed extensively with PBS, and then floated on gold-conjugated

goat anti-mouse IgG for 1 h at 25 C. Filters were washed in PBS,

fixed for 1 h in 2.5% glutaraldehyde in PBS with 0.5% tannic acid,

post-fixed in 1% Os4, dehydrated, embedded in Spurr's resin, and

sectioned. Thin sections were stained with uranyl acetate and lead

citrate and examined on a JEOL 100-CX electron microscope.


The assay described below, as performed with unstabilized

erythrocytes (RBC), is basically that of Costello et al. (29).

Preparation of neuraminidase-treated RBC. Human type 0 blood was

drawn in anticoagulant and diluted with four volumes of pH 7.2 PBS.

RBC were collected by centrifugation at 750 x g for 10 min and washed

twice more in the same manner. A 15% (v/v) suspension of RBC in

neuraminidase (30 pg per ml; Cat. No. N2876 type V neuraminidase

from Clostridium perfringens, Sigma Chemical Co.) in pH 5.0 PBS was

incubated in a 37 C waterbath for 2 h. Neuraminidase-treated RBC

(NTRBC) were washed three times and suspended in TBS containing 0.4%

BSA. NTRBC had to be used within a day or two, unless they were

stabilized by treatment with formaldehyde or glutaraldehyde.

Formaldehyde stabilization of NTRBC. After neuraminidase

treatment an aliquot of NTRBC was washed in pH 7.2 PBS to remove the

enzyme and acid buffer. Then a 10% (v/v) suspension of NTRBC in PBS,

pH 7.2, was mixed with an equal volume of 3.7% formaldehyde. The

suspension was incubated at 25 C with occasional stirring for 4-6 h,

then at 37 C with continuous stirring for 14-18 h. The

formaldehyde-stabilized NTRBC were washed 4X with 10 volumes of TBS

and 3X with TBS containing 0.4% BSA. They were stored at 4C in

TBS-BSA. These cells were evaluated versus freshly prepared and aged

preparations of NTRBC that had not been treated with formaldehyde to

determine whether formaldehyde stabilization prolonged the "life" of

the NTRBC.

Glutaraldehyde stabilization of NTRBC. A 1-2% (v/v) suspension

of RBC in cold 1% glutaraldehyde in pH 8.2 PBS was incubated in an ice

bath for 30 min with intermittent mixing (2). The glutaraldehyde-

treated RBC (G-RBC) were collected by centrifugation (750 x g, 10 min)

and washed once in saline. To block potentially reactive free

aldehyde groups, a 10% (v/v) suspension of G-RBC in 0.1 M glycine was

incubated at 25 C for 1 h. Then G-RBC were washed 5-10 times in

TBS. They were stored in TBS with 0.4% BSA at 4C. Glutaraldehyde

stabilization could be used either before or after neuraminidase


Test procedure. Starting with a bacterial cell suspension

equivalent to A = 2.0, serial two-fold dilutions in TBS with
0.4% BSA were made in a U-bottom microtiter plate, leaving 25 1l per

well. Then 25 ul of a 1% (v/v) suspension of NTRBC in TBS-BSA were

added to each well and mixed for 1 min. The reactions were read

immediately and after overnight incubation at ambient temperature. To

test for lactose-reversibility, 50 il of 0.04 M lactose (0.02 M

final concentration) were added. Alternatively, the effects of

various inhibitors were tested by adding 50 Vl of inhibitor to the

bacterial cells prior to the addition of NTRBC.


The protocol described below is based on the methods of Mclntire

et al. (82) and Cisar et al. (18).

Tube assay. Bacterial cell suspensions in TBS were adjusted to

A650 = 2.0-2.1 (1-cm cuvette, Beckman Model 25 spectrophotometer).

In a 10 x 75 mm test tube 0.2 ml of an actinomycete cell suspension

was mixed with an equal volume of a streptococcal cell suspension.

For controls, 0.2 ml of buffer was added in place of one of the cell

suspensions. The suspensions were mixed by vortexing, then incubated

overnight at ambient temperature. They were mixed again before the

reactions were graded on a scale of 0 to 4+ : 0 = no visible

aggregates; 1+ = small uniform aggregates in suspension; 2+ = definite

aggregates that did not settle immediately; 3+ = large aggregates that

settled rapidly, leaving some turbidity in the supernatant; 4+ = large

aggregates that settled immediately, leaving a clear supernatant.

Microtiter plate assay. A microtiter plate coaggregation assay

was developed to replace the standard tube assay. The procedure was

similar to that described for the tube assay, except that 25 pV of

each cell suspension were mixed in a U-bottom microtiter plate.

Bacterial Agglutination

In a U-bottom microtiter plate serial dilutions of the antibody

samples were made in TBS containing BSA (4 mg/ml). Normal sera and

diluent were used as negative controls. To 25 pl of antibody were

added 25 Vl of a suspension of bacterial cells in TBS adjusted to

A 650 = 1.0 (1 cm cuvette, Beckman Model 25 spectrophotometer), and
the plate was shaken for 1-2 min. Reactions were read after overnight

incubation at 25 C (14).


A solid-phase radioimmunoassay (RIA) was developed as a rapid,

sensitive screening assay for detection of monoclonal antibodies to

surface antigens of N16 and other isolates. Polyvinyl chloride

microtiter plates (U-bottom, Dynatech Laboratories, Inc.) were coated

with formalin-killed whole cells by placing 25 pI of a No. 4

McFarland cell suspension in each well for 1 h at 25C. Excess

antigen was rinsed out, and free sites were blocked by the addition of

a few drops of 10% agamma horse serum to each well. The antigen-

coated wells were incubated (1 h, 25 C) with 25 pl of undiluted

hybridoma culture supernatant. The wells were washed three times with

PBS containing 1% fetal calf serum and incubated (1 h, 25 C) with 25

pV of 1I-labeled rabbit anti-mouse IgG (heavy and light chain

specific) containing 30,000-50,000 counts per minute (cpm). After the

unbound radiolabeled second antibody had been washed away, individual

wells were cut out, and the bound radioactivity was measured in a
gamma counter. Reactions in which 1I-cpm were at least twice that

of the negative control were considered to be positive. Culture

supernatants from the myeloma lines used for fusion served as negative

controls. Murine antisera were used as positive controls if positive

culture supernatants or ascites fluids were not available.

Indirect Enzyme Immunoassay

Indirect enzyme immunoassays (EIA) were performed in Immulon 2

96-well flat-bottomed polystyrene microtiter plates (Dynatech

Laboratories, Inc.). Wells were coated with antigens by placing 50

pi of whole cells or soluble antigens optimally diluted in TBS in

each well and drying at 37 C. Antigen-coated plates were stored at

ambient temperature until used. Plates were washed once with blocking

buffer (PBS with 1% BSA), then 3 times with PBS with 0.05%

polyoxyethylene sorbitan monolaurate (Tween 20) (PBST). Antibodies

diluted in PBST were added to the wells (25 pI per well) and

incubated at 25 C for 1 h. Plates were washed 3 times with PBST.

Then 100 pl of horseradish peroxidase-labeled goat anti-rabbit or

mouse IgG optimally diluted in PBST were added and incubated at 25 C

for 1 h. After the excess conjugate had been removed by washing, 200

pU of substrate solution (0.1 mg of ortho-phenylenediamine and

2 ~1 of 3% H202 per ml of citrate-phosphate buffer, pH 5.0) were

added and incubated at 25 C for 15-30 min. Reactions were stopped

with 25 pl/well of 4 M H2SO4, and the absorbance at 490 nm was

read on a Dynatech MR 600 Microplate Reader (Dynatech Laboratories,

Inc.). Absorbance readings >0.2 above the negative control were

considered positive.

Enzyme Immunodot Assay

Antigens were affixed to strips of nitrocellulose by applying

2-1p drops at 5-nnm intervals and allowing them to dry at 25 C

(79). The antigen-coated strips were processed by enzyme immunoassay

essentially as described by Tsang et al. (99). The strips were washed

4 times for 5 min each time with PBS with 0.3% Tween 20 (PBSTW) with a

quick rinse in deionized water after each PBSTW wash. The

antigen-coated strips were immersed in antibodies diluted in PBSTW and

incubated with gentle agitation for 1 h at 250C. Then the strips

were washed as before to remove unbound antibodies. The strips were

covered with the appropriate peroxidase-conjugated anti-Ig and

incubated for 1 h at 250C. Unbound conjugate was removed by washing

as described above, followed by a final wash with PBS. The strips

were incubated with substrate (50 mg of 3,3'-diaminobenzidine and 10

pl of 30% H202 per 100 ml of PBS) for 10 min or until the spots

had developed the desired degree of intensity. The reaction was

stopped by rinsing the strips thoroughly with water.


Inmunodiffusion was performed in 1% agarose in PBS with wells

formed by use of a microimmunodiffusion template (89).

Laurell Rocket Immunoelectrophoresis

Laurell rocket immunoelectrophoresis (LRI) was performed as

described by Powell et al. (90) in 0.75% agarose in 0.043 M sodium

barbital buffer, pH 8.3, at a constant current of 8 mA per 50 mm x 75

mm slide. Gels were dried and stained with 0.5% Coomassie brilliant

blue R-250 in ethanol:glacial acetic acid:water (4.5:1.0:4.5).

Crossed Immunoelectrophoresis with Autoradiography

Glass slides (50 mm x 75 mm) were coated with 7 ml of 0.75%

agarose in barbital buffer (0.0375 M barbital, 2 mM calcium lactate,

0.05% NaN3, pH 8.6). After the antigen wells were cut and filled,

the gels were electrophoresed at 8 mA constant current for 1-3 h for

separation of antigens in the first dimension. Then the gel above the

antigen wells was replaced with 3 ml of agarose containing an

unlabeled rabbit antiserum raised against whole cells and an
125I-labeled Ig (50). After electrophoresis in the second dimension

overnight at 4 mA constant current, gels were washed, dried, and

exposed to Kodak XAR-2 film with a Dupont Cronex Lightning Plus

intensifying screen at -700C. Gels were stained with Coomassie

brilliant blue R-250 and photographed.

Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis

Socium dodecyl sulfate-polyacrylamide gel electrophoresis

(SDS-PAGE) was performed as described by Tsang et al. (102). Samples

were prepared to obtain final concentrations of 1% SDS, 0.9 M urea,

0.05% bromophenol blue, 8% glycerol, and 0.25 mg total protein per

ml. Some samples also received dithiothreitol (DTT) or

2-mercaptoethanol for reduction of disulfide bonds. Samples were

heated at 37 C for 30 min, at 65 C for 15 min, or at 1000C for

5 min. Gels consisting of a 5-20% linear gradient resolving gel and a

3% stacking gel 0.75 mm thick were cast using the Pharmacia GSC-2 177

mm x 200 mm vertical slab system (Pharmacia). The gels were

electrophoresed in the Pharmacia GE 2/4 LS electrophoresis chamber

using a discontinuous buffer system with 0.424 M Tris-HC1, pH 9.18, in

the lower reservoir and 0.04 M boric acid, 0.04 H Tris, 0.1% SDS in

the upper reservoir. Generally 0.25 pg of protein per mm width of

sample lane was applied. Either high molecular weight markers

(Bio-Rad) and low molecular weight markers (Pharmacia) or prestained

protein molecular weight standards (Bethesda Research Laboratories,

Gaithersburg, MD) were included in each gel (101). After

electrophoresis protein bands were visualized by the silver stain as

described by Tsang et al. (100), which is a combination of the methods

of Merril et al. (85) and Morrissey (88).


SDS-PAGE resolved proteins were transferred electrophoretically

to nitrocellulose (0.2 V, Schleicher and Schuell, Keene, NH); then

antigens were visualized by an indirect enzyme immunoassay, using

peroxidase conjugates (Bio-Rad Western blotting grade) and

diaminobenzidine/ H202 as the substrate. These methods have been

described in detail by Tsang et al. (99,102). For detecting unstained

molecular weight markers or other proteins on blots, nitrocellulose

strips were stained with India ink (52) or Aurodye forte (Janssen Life

Sciences Products, Piscataway, NJ).


Monoclonal Antibodies

Three different hybridomas secreting antibodies to A. naeslundii

serotype 3 strain N16 were produced. Hybridomas 3B5.A1 and 2A3.B3,

both obtained from the fusion with the non-Ig-producing Kearney

myeloma cell line, secreted IgG2a and IgG3 monoclonal antibodies,

respectively. The third hybridoma, 2B5.B6, which was a product of the

fusion with an IgG (k)-secreting myeloma cell line, secreted IgM

monoclonal antibodies along with the Ig chains of the myeloma parent.

Preliminary indications of differences in specificity and the

range of reactivity of each of the three monoclonal antibodies were

obtained by radioinmunoassay, using whole cells as antigens. When

tested against a panel of ten isolates of A. naeslundii serotype 3,

monoclonal antibodies 3B5.A1 and 2A3.B3 reacted with all 10 isolates,

whereas monoclonal antibody 2B5.B6 reacted with only 4 of the 10

isolates (Table 1). On the basis of these results, monoclonal

antibodies 3B5.A1 and 2A3.B3 appeared to have potential as serotyping

reagents, provided that they did not also react with heterologous

serotypes; monoclonal antibody 2B5.B6 had too narrow a range of

reactivity to be useful for serotyping, unless it could be used as an

epidemiological tool for fine typing below the serotype level. To

determine whether or not the monoclonal antibodies would cross-react

with other organisms, the panel of cells for the RIA was expanded to

include 51 heterologous isolates: 17 isolates of A. naeslundii that

TABLE 1. Summary of RIA results for isolates within the N16 serogroup

Monoclonal antibody
Isolate 3B5.A1 2A3.B3 2B5.B6

N16 +a + +
WVU1267 + + -
WVU1468 + + -
WVU1527 + + +
WVU1528 + + +
W1629 + + -
W2273 + + -
W2821 + + +
UF92 + + -
UF524 + + -

a + = 125I-_pm > 1251-cpm of negative control x 2;
- = 125I-cpm < 125I-cpm of negative control x 2 for the average
125I-cpm of samples run in duplicate. 125I-cpm of negative
controls ranged from 53 to 78, whereas 125I-cpm of + reactions
ranged from 1919 to 6283.

did not belong to the N16 serogroup serotypee 3), 13 isolates of

A. viscosus, 8 isolates of A. israelii, and 13 isolates of less

closely related species. When assayed using polyclonal antisera,

A. naeslundii and A. viscosus are closely related antigenically, and

antisera raised to whole cells of N16 could be expected to cross-react

with A. viscosus serotype 2 in particular. However, monoclonal

antibody 3B5.A1 did not react with any heterologous species or

serotypes (Table 2). On the other hand, monoclonal antibody 2A3.B3

cross-reacted with 7 of 17 non-serotype 3 A. naeslundii isolates, as

well as 7 of 11 A. viscosus serotype 2 isolates. However, the
I-cpm for all but one of these cross-reactions were less than 34%
of the 1I-cpm of the least reactive A. naeslundii serotype 3

isolate. Monoclonal antibody 2B5.B6 cross-reacted with 4 of 17

isolates representing other serotypes of A. naeslundii, and these

cross-reactions were as strong as the homologous reactions. None of

the three monoclonal antibodies reacted with isolates of A. israelii,

A. odontolyticus, A. pyogenes, A. propionica, R. dentocariosa,

P. avidum, or Bacteroides gingivalis. These results suggested that

monoclonal antibody 3B5.A1 recognizes a serotype-specific antigenic

determinant but that monoclonal antibodies 2A3.B3 and 2B5.B6 see

epitopes common to other species or serotypes.

The location of the epitopes to which the monoclonal antibodies

were binding was presumed to be the cell surface because whole cells

were used as immunogens and as antigens in immunoassays to measure

monoclonal antibody activity. The most immunogenic and biologically

significant surface components of A. naeslundii and A. viscosus

isolates are their fimbriae. Electron microscopy of A. naeslundii N16

TABLE 2. Summary of the specificities
of three monoclonal antibodies to N16 as determined by RIA

No. isolates Monoclonal antibody
Species/serotype tested 3B5.A1 2A3.B3 2B5.B6

Actinomyces naeslundii/3 10 10a 10 4
A. naeslundii/l, 2, 4 17 0 7 4
A. viscosus/1 2 0 0 0
A. viscosus/2 11 0 7 0
A. israelii/1 6 0 0 0
A. israelii/2 2 0 0 0
A. odontolyticus 3 0 0 0
A. pyogenes 1 0 0 0
Arachnia propionica 3 0 0 0
Rothia dentocariosa 2 0 0 0
Propionibacterium avidum 3 0 0 0
Bacteroides gingivalis 1 0 0 0

a Number of isolates that gave positive reactions, as defined in
Table 1. The range of 125I-cpm for positive reactions was as
follows: 2547-6283 for 3B5.A1 against homologous isolates; 1919-6025
for 2A3.B3 against homologous isolates; 2097-2628 for 2B5.B6 against
homologous and heterologous isolates. Except for 3207 1251-cpm
against A. naeslundii W1250, 2A3.B3 gave 161-649 125I-cpm against
heterologous serotypes of A. naeslundii and 180-328 125I-cpm against
A. viscosus serotype 2 isolates.

cells revealed that they possess abundant fimbriae (Figure 1). The

location and density of the epitopes recognized by each monoclonal

antibody, as determined by EM of N16 cells immunolabeled with colloidal

gold, are shown in Figure 2. All three monoclonal antibodies bound to

epitopes residing on the fimbriae, but based on the far greater number

of gold particles seen on cells immunolabeled with MAb 3B5.A1, the

3B5.A1 epitope appeared to be more abundant than those recognized by

MAbs 2A3.B3 and 2B5.B6.

All three of the MAbs against N16 fimbriae were used on occasion

in later experiments. However, most of this study focused on the use

of MAb 3B5.A1 because preliminary evaluations indicated that (a) it

reacted with all 10 A. naeslundii serotype 3 strains by RIA, (b) it

appeared to recognize a unique, serotype-specific epitope, and (c) on

the basis of the density of immunogold labeling, it appeared to

recognize an epitope that was more abundant than those of the other

two MAbs.

Evidence for the Presence of Both Types of Fimbriae
on A. naeslundii Serotype 3 Strains

Evidence for the Presence of Type 1 Fimbriae

Although this study did not include any functional assays to

demonstrate the presence of type 1 fimbriae on N16 or other

A. naeslundii serotype 3 isolates, Clark et al. (23,25,26,28) have

presented evidence that 3 of the 10 strains listed in Table 1 possess

type 1 fimbriae. In one study they showed by EM that N16 and UF92

cells had fimbriae; both strains adsorbed very well to saliva-coated

hydroxyapatite (SHA) relative to the adsorption of T14V, 100% for N16

and 84% for UF92; and their adherence to SHA could be blocked by

purified T14V fimbriae (26). In another study N16 and G1468 (WVU1468)

Electron micrograph of A. naeslundii N16 cells negatively
stained with 0.5% uranyl acetate. Note abundant fimbriae
on the cell surface. Bar, 1 um.

Figure 1.

Electron micrographs of thin sections of immunogold labeled
A. naeslundii N16 cells demonstrating the binding of the
monoclonal antibodies to fimbrial epitopes. The cells were
incubated with MAb-containing culture supernatants of
myeloma or hybridoma cell lines and then with goat anti-
mouse IgG conjugated to colloidal gold (diameter, 15 nm).
Bar, 0.25 mn.
A) P3X63Ag8, negative control; B) MAb 2B5.B6; arrows
indicate gold particles; C) MAb 2A3.B3; D) MAb 3B5.A1;
E) MAb 3B5.A1; F) MAb 3B5.A1

Figure 2.





a o.


*>, :I *"
..\ \

Figure 2. --continued


ll < '*

! 4


.b 6


*'I %


Figure 2. --continued


E p' '



s es

.A .
,. t 1,

... I .

-''R r *

^^^Sol^ sL&Ihi
~t F~y4

adsorbed to SHA better than T14V, despite the fact that each had a

lower hydrophobic index than T14V (25). N16 and WVU1468 reacted in an

immunodot EIA with rabbit IgG specific for the T14V type 1 fimbriae

(28), and rabbit IgG anti-T14V type 1 fimbriae was able to partially

inhibit adsorption of N16 cells to SHA and give 91% inhibition for

WVU1468 (23,28). The ability of N16, WVU1468, and UF92 to adsorb well

to SHA and the fact that T14V fimbriae or antibodies to T14V type 1

fimbriae can inhibit the adsorption suggests that these A. naeslundii

serotype 3 strains possess type 1 fimbriae. Inmunological assays

performed in this study (data to be presented later) showed that all

of the A. naeslundii serotype 3 strains listed in Table 1 have type 1


Evidence for the Presence of Type 2 Fimbriae

Hemagglutination. Many human strains of A. viscosus and

A. naeslundii can agglutinate human types A, B, 0, and AB erythrocytes

(29,37). With untreated RBC, hemagglutination occurs immediately at

4 C, but the reaction takes longer at room temperature or 37 C.

With neuraminidase-treated RBC, hemagglutination occurs immediately at

all three temperatures (29). Hemagglutination by A. viscosus and

A. naeslundii cells can be completely reversed by lactose. The

mechanism proposed for this reaction involves two steps: 1) the

unmasking of B-galactoside-containing receptors on the RBC when

terminal sialic acid residues are removed by neuraminidase released

from Actinomyces cells, and 2) the multivalent binding of these

receptors by multiple low affinity, lactose-reversible lectin sites on

the surface of the actinomycete (29,37). Since the lactose-reversible

lectin activity is a function of type 2 fimbriae (20), HA can be used

to determine whether or not Actinomyces cells have type 2 fimbriae.

Costello et al. (29) reported that N16 and WVU820 (N16) produced

neuraminidase and agglutinated human RBCs as described above. On the

other hand, Ellen et al. (37) reported that strain 820, which is

presumably WVU820 (N16), did not agglutinate non-sialidase-treated

horse, sheep, guinea pig, or human RBCs. Since Ellen et al. measured

HA after only a 5 min incubation, the inability of WVU820 to

agglutinate RBCs was probably a result of not allowing sufficient time

for the WVU820 neuraminidase to remove the terminal sugar blocking the


In this study the 10 A. naeslundii serotype 3 strains listed in

Table 1 were tested for HA activity. Preliminary experiments were

performed with unfixed NTRBC, but unfixed NTRBC were found to be so

unstable that they generally had to be used within 24 h of

preparation. Even then they often gave weak HA reactions or lysed

during assays. In an effort to improve their shelf-life, NTRBC were

fixed with formaldehyde or glutaraldehyde and compared to unfixed

NTRBC. HA reactions with fixed NTRBC were stronger than with unfixed,

and glutaraldehyde-fixed RBC were superior to formaldehyde-fixed

cells. Not only did glutaraldehyde-fixed NTRBC agglutinate better

than unfixed cells, but they also had a shelf-life at 4C that could

be measured in years instead of hours. The characteristics of the

agglutination reactions between N16 cells and human RBC (in terms of

the effects of temperature, incubation time, etc. as reported by

Costello et al.) were the same for glutaraldehyde-fixed RBC or NTRBC

as for unfixed RBC or NTRBC, except that HA with fixed cells was

stronger and, therefore, more difficult to reverse. All 10

A. naeslundii serotype 3 isolates agglutinated G-NTRBC, and all of the

HA reactions were reversed by lactose. Twelve sugars and EDTA were

tested for their ability to inhibit HA by N16 cells. Glucose,

fructose, mannose, sucrose, xylose and a-methyl galactoside

exhibited no inhibition of HA at concentrations of inhibitor < 25 mM.

Of the substances that did inhibit at concentrations < 25 mM, the

relative potency of the inhibitors was as follows: EDTA > lactose >

B-methyl galactoside > talose > fucose > galactose > N-acetyl

galactosamine. These results on the relative potency of inhibitors of

the N16 HA reaction are similar to the results of Ellen et al. (37)

for the inhibition of HA by A. naeslundii serotype 2 W752 cells and

also the results of Mclntire et al. (81,82) for inhibition of

coaggregation between T14V and Streptococcus sanguis strain 34 cells.

N16 cells also agglutinated sheep and guinea pig RBC, and HA was

completely inhibited with EDTA or lactose at a final concentration of

25 mM.

Coaggregation. Many human strains of A. viscosus and

A. naeslundii can agglutinate certain strains of oral streptococci

(18,19,36,69,81,82). These coaggregation reactions involve different

mechanisms for interaction, one of which is the binding of a

carbohydrate moiety on S. sanguis 34 cells by a lectin on Actinomyces

cells in a reaction that can be completely reversed by lactose

(18,69,81,82). Coaggregation, like hemagglutination, is a function of

type 2 fimbriae (20,92), so it also can be used to determine whether

or not Actinomyces cells have type 2 fimbriae.

Cisar et al. (18) showed that WVU820 (N16) and W1527 (WVU1527)

exhibited lactose-reversible coaggregation with S. sanguis 34 cells.

Their observations were confirmed in this study, using both the tube

and microtiter plate coaggregation assays (data not shown).

In summary, there was ample evidence from our preliminary studies

to suggest that N16 and other A. naeslundii serotype 3 strains had

both type 1 and type 2 fimbriae. The next step was to find a good way

to remove the fimbriae from the cells and obtain crude soluble

fimbrial extracts suitable for purification of fimbriae and for use as

antigens in immunological assays.

Antigen Preparation: Results of Preliminary Experiments

In the early stages of this research several different methods

for preparing soluble fimbrial extracts of N16 cells were tried.

First, the method of Wheeler and Clark (106) for purifying VA-1

fibrils (type 1 fimbriae) from A. viscosus T14V cells by ammonium

sulfate fractionation of French pressure cell supernatants was

attempted. Laurell rocket immunoelectrophoresis of the N16 fractions

prepared in this manner, i.e. 10, 20, 30, 40, and 50% (w/v) ammonium

sulfate saturated samples, revealed multiple antigens in each fraction

when antiserum to N16 whole cells was employed. When these fractions
were used to coat wells in the RIA, MAb 3B5.A1 gave 1I-cpm with

the 20% and 30% ammonium sulfate fractions that were 5-8 times higher

than those obtained with the 10%, 40%, or 50% fractions; thus the 20%

and 30% fractions appeared to be enriched for the antigen that 3B5.A1

recognizes. Each fraction was examined by transmission electron

microscopy for the presence of fimbriae, but the 20% fraction was the

only one in which fimbriae were readily observed.

Although the French pressure cell was an effective means for

removing fimbriae from cells with minimal cell lysis, access to this

equipment was not readily available. So, several other methods for

extracting fimbrial antigens were investigated: Lancefield extraction,

sonication, Virtis homogenization, and extraction with 8 M lithium

chloride. N16 cells, before and after being treated by these various

procedures, were tested for lectin activity by hemagglutination.

Since type 2 fimbriae of Actinomyces exhibit lectin activity, the

efficacy of each treatment of the cells in removing type 2 fimbriae

could be compared in this manner; presumably, type 1 fimbriae would be

removed also. Sonication and Lancefield extraction reduced the lectin

activity of N16 cells from 2+ to 0, whereas Virtis homogenization and

extraction with 8 M lithium chloride had little, if any, effect on the

lectin activity. Electron microscopy confirmed that sonication and

Lancefield extraction were much more effective than Virtis

homogenization or lithium chloride in removing fimbriae from N16

cells. The observation that treatment of N16 cells with 8 M LiCI had

no detectable effect on the fimbriae is contrary to the observation of

Mclntire et al. (82) that A. viscosus T14V cells treated in such a

manner lost their fimbriae and, consequently, their ability to

coaggregate with S. sanguis 34.

The total protein and total carbohydrate extracted per gram wet

weight of cells by these procedures are compared in Table 3.

Sonication released greater than 15 times more protein than either

Virtis homogenization or lithium chloride extraction, and the protein

to carbohydrate ratio was greater than 1.0 in each case. For

Lancefield extraction the protein to carbohydrate ratio was only

0.14. Lancefield extraction was eliminated from consideration as a

method for extracting fimbriae for subsequent purification because it

released 7 times more carbohydrate than protein and because of the

likelihood of protein denaturation resulting from acid hydrolysis.

However, it was the preferred method for preparing soluble antigens

for immunoelectrophoresis because (a) it was the most convenient way

to extract antigens from a large number of strains, and (b) it

invariably gave strong, symmetrical type 1 and type 2 fimbrial rockets

in LRI and XIEP.

TABLE 3. Total protein and total carbohydrate
released from N16 cells by several methods for extracting fimbriae

Method of extraction CHOa Proteinb

Continuous flow sonication 5.0 11.2
Virtis homogenization 0.5 0.6
8 H LiC1 0.3 0.4
Lancefield extraction 22.8 3.1

a mg total carbohydrate released per g wet weight of cells.
mg total protein released per g wet weight of cells.
c A different method was used to determine the wet weight of
cells subjected to Lancefield extraction, so the quantity of CHO and
protein extracted by that method cannot be compared directly to the
quantities extracted by the other methods.

Although French pressure cell shearing and continuous flow

sonication were my first and second choices, respectively, for

obtaining crude extracts for purification of N16 fimbriae, lack of

access to the appropriate equipment at the proper time ultimately

dictated that the initial crude fimbrial extracts be obtained by a

batch method of sonication. Later in the study, a French pressure

cell extract was also processed.

Identification of N16 Type 1 and Type 2 Fimbriae in
Crude Antigen Extracts by XIEP-A

Immunoelectron microscopy had been used to demonstrate that the

3B5.A1 epitope was located on N16 fimbriae, but whether those fimbriae

were type 1 or type 2 fimbriae had not yet been established. Analysis

of N16 crude sonicates by XIEP against rabbit antiserum to whole cells

revealed the presence of two major antigens that were thought to

correspond to type 1 and type 2 fimbriae (Figure 3, upper panels).

Cisar et al. (14) had shown that in XIEP Ag 1 (type 1 fimbriae) of

T14V migrated closer to the anode than Ag 2 (type 2 fimbriae).

Whether the type 1 and type 2 fimbriae of all Actinomyces strains

maintained the same relative electrophoretic mobilities was not

known. In fact, since crude sonicates of N16 cells contained other

antigens besides fimbrial antigens, and the antiserum used to

precipitate the antigens was not specific for fimbriae, the two N16

rockets did not necessarily have to be fimbrial antigens.

In order to establish the identity of the two N16 rockets and to

determine which type of fimbriae MAb 3B5.A1 recognized,
fimbriae-specific 125I-Ab was incorporated into the Ab-containing

portion of the gel along with the non-radiolabeled polyspecific
antiserum. When 125I-Ab specific for T14V type 1 fimbriae was used,

autoradiography of the gel demonstrated that the N16 rocket closer to

the anode corresponded to type 1 fimbriae (Figure 3A). When the gel
contained 1I-MAb 3B5.A1, autoradiography revealed that MAb 3B5.A1

bound to the other rocket (Figure 3B). Since (a) MAb 3B5.A1 binds to

fimbriae, (b) it did not bind to type 1 fimbriae, and (c) only two

types of Actinomyces fimbriae have been identified, by a process of

elimination it was determined that the antigen that migrated more

Identification of A. naeslundii N16 type 1 and type 2
fimbriae by XIEP-A. N16 cell surface antigens extracted by
sonication were separated by electrophoresis in the first
dimension with the anode to the left. Then they were
precipitated by electrophoresis (at a right angle to the
first dimension) into gel containing R29 rabbit antiserum
against N16 whole cells and 125I-labeled IgG specific for
fimbriae. Two rockets precipitated by R29 are seen in the
Coomassie-stained gels (upper panel); the corresponding
autoradiographs (lower panel) indicate that the rocket
closer to the anode represents type 1 fimbriae, whereas the
other rocket represents type 2 fimbriae. In this and all
subsequent figures of XIEP patterns, the anode is on the
left for electrophoresis in the first dimension and at the
top for electrophoresis in the second dimension. The
type 1 fimbrial rocket is always on the left and the type 2
fimbrial rocket on the right.
A) 125I-R59 specific for the type 1 fimbriae of
A. viscosus T14V binding to the type 1 fimbriae of N16;
B) 125I-MAb 3B5.A1 binding to the type 2 fimbriae of N16.

Figure 3.

slowly toward the anode was type 2 fimbriae and that MAb 3B5.A1 was

specific for type 2 fimbriae.

Immunological detection of type 2 fimbriae throughout various

purification procedures and purification of type 2 fimbriae by

immunoaffinity chromatography were made possible by the availability

of MAb 3B5.Al. Also, XIEP-A proved to be a useful technique for

monitoring antigen fractions for the presence of both types of

fimbriae, as well as non-fimbrial antigens.

Purification of N16 Type 2 Fimbriae From a Crude Sonicate

Batch A

Aliquots of the untreated control cells and the sonicated cells

collected at different time intervals were assayed for lectin activity

by HA to monitor the removal of type 2 fimbriae as measured by the

decline in HA activity. Cells sonicated for only 30 sec had the same

HA titer as the untreated control cells. With cells sonicated from 1

to 6 min total, the HA titers of the cells declined as the total

sonication time increased. Most of the decrease in lectin activity

had occurred after 2 min of sonication, but even after sonication for

a total of 6 min, the cells still retained weak HA-positive activity.

Electron microscopy on the cells sonicated for 6 min revealed that

nearly all the cells had lysed and the cell walls appeared to be

devoid of fimbriae.

When the crude 6 min sonicate was applied to the Bio-Gel A-5m

column and eluted with TBS, two major protein-rich peaks were eluted,

one just after the void volume and the other near the total volume

(Figure 4). When the column fractions were dotted on nitrocellulose

and reacted with MAb 3B5.A1 in an indirect EIA, fractions 11-24 gave

0 1
C)l 1-
00 0

F 0 .0

U 000

40 >r
S Ln

0 E4 M

*I 01 t

0.' W
0 0)
a) 0604
0 E4U

0 0O
-4 u4 .

0 0

WsT m

*^ a

*t-1 rr 0


0 L1-

t 0
a -i

'. .V *


~ cu


SL -



Cu a ao CD cu 0 0 LOD C/u
Cu C li cl C-, C-) ,' 0 0 0 0


3-4+ reactions, but the intensity of the immunodots decreased sharply

beyond fraction 24. Although fractions 18-24 also contained type 2

fimbriae, only fractions 11-17 (designated A-5m peak 1) were pooled,

because fractions beyond the first major protein peak contained

increasing levels of carbohydrate. The relatively high carbohydrate

content of all the fractions in this batch probably was a reflection

of the substantial amount of cell breakage that occurred when the

cells were sonicated. When A-5m peak 1 was examined by XIEP against

rabbit antiserum to N16 whole cells, both types of fimbriae and some

nonfimbrial antigens were detected. A-5m peak 1 from Batch A was not

purified further because another batch, less contaminated with

cytoplasmic components, was prepared.

Batch B

Two minutes of sonication caused a 16-fold decrease in the 2+ HA

endpoint relative to the untreated cell suspension, suggesting

significant removal of type 2 fimbriae from the cells. Electron

microscopy of negatively stained cells confirmed that sonication

removed some but not all fimbriae from the cells. When the crude

sonicate was precipitated with (NH ) 2SO at 50% saturation and

again at 35% saturation, the supernatant fractions after dialysis and

concentration were examined for the presence of fimbriae by XIEP. The

results indicated that some of both types of fimbriae were lost to the

supernatant fraction with each precipitation; however, no attempt was

made to quantitate the amounts of fimbriae or total protein not


The precipitate from 35% (NH4)2SO4 saturation of the crude

2 min sonicate contained 390 mg total protein. This sample was

chromatographed in 5 separate aliquots on a Bio-Gel A-5m column; a

representative elution profile is shown in Figure 5. This elution

profile is similar to that obtained with Batch A (Figure 4), except

that the Batch B fractions contained much lower levels of total

carbohydrate. The first of two major protein-rich peaks that were

eluted from the column contained most of the MAb 3B5.A1 reactive

(type 2) fimbriae as determined by immunodot EIA on the column

fractions. All of the fractions representing the first peak (peak 1)

from each run, i.e., fractions 14-32 in Figure 5, were pooled; the

A-5m peak 1 pool contained 159 mg total protein.

A 5 mg aliquot of the peak 1 pool was chromatographed on the MAb

3B5.A1 immunoaffinity column. Analysis of the column fractions by

XIEP-A indicated that the unbound fraction resembled the starting

material, which suggested overloading of the column, whereas the bound

fraction appeared to contain only type 2 fimbriae.

The unbound fraction from the first immunoaffinity run was

rechromatographed on the MAb column. This resulted in a reduction in

type 2 fimbriae relative to type 1 in the unbound fraction and

recovery of type 2 fimbriae in the bound fraction with only minor

contamination with type 1 fimbriae (Figure 6). Thus, MAb 3B5.A1,

which is specific for the type 2 fimbriae of N16, is an effective tool

for the immunochromatographic purification of this structural entity.

The fimbrial immunoprecipitin arcs that were used to immunize

rabbits for the production of polyclonal antibodies specific for N16

type 1 or type 2 fimbriae were obtained from these fractions, as shown

in Figure 7 (refer also to Materials and Methods).


A .4 U

0 w
o 4) -"

a .C
C I ,t
o *
S *.

60 0 0

| 441 4 -,4
.&) CO 4J

P *.4 A
I 0 >e

0 .0 4.

.40 ( 0
Scd 0 4 H

0 O4 41
S.0C U1
04 0 Cd 4

0 0 0
4) J) .-4 4)

0 00 x0

.4 co A >
4> 4. a) lu *..

0a 0. 4

S0 CO04)
0 0 0
C O- 0 Ia

0 z 0 ca

b)F )

0 0 3 Ln

S C-



cn c cD

S0 0 0C 0 0 0 0CoD

ra r ti
cC* ^''rl
C- TS '
as W a,

\i/^ r *-
$ ^^^
(D ^ ^ ^
R~l $ '^ Ll
^ ^ ^
g j0 ^-
S ^0^ -^
s ^*0' -

^ Zs^

\\< ~~ ~ ---_^s -v I
^ ^ -^ ^A" *-

"-* o o o o o o o o o



Results of immunoaffinity chromatography of A-5m peak 1 on
the MAb 3B5.Al-Affi-Gel 10 column as demonstrated by
XIEP-A. The antibody-containing portion of each gel
included R29 antiserum against N16 whole cells and
125I-labeled MAb 3B5.A1 specific for type 2 fimbriae.
The Coomassie-stained gels (upper panel) show all the
antigens precipitated by the polyspecific rabbit antiserum,
whereas their corresponding autoradiographs (lower panel)
identify the type 2 fimbriae.
A) A-5m peak 1, the starting material for immunoaffinity
chromatography, contained both types of fimbriae; B) The
unbound fraction also contained both types of fimbriae, but
there was a reduction in type 2 fimbriae relative to type 1;
C) The bound fraction contained mostly type 2 fimbriae with
a barely detectable quantity of type 1 fimbriae (arrow).



a i

Figure 6.






Source of immunogens for production of rabbit antisera
monospecific for N16 type 1 or type 2 fimbriae. Segments
of fimbrial immunoprecipitates were excised from the areas
indicated on the XIEP gels. R29 against N16 cells was the
precipitating antiserum in both gels.
A) Type 1 fimbrial immunoarcs for production of R10
B) Type 2 fimbrial immunoarcs for production of R2P

Figure 7.

When R10 antiserum against N16 type 1 fimbrial immunoprecitins or

R2P against N16 type 2 fimbriae were examined by immunoelectron

microscopy against N16 cells, both antisera reacted with epitopes

located on the fimbriae (Figure 8; R10 reaction not shown). These

antisera were used to make polyclonal rabbit IgG immunoaffinity

columns for the purification of both types of fimbriae from the N16

French press supernatant.

Purification of N16 Fimbriae from the French Press Supernatant

In addition to the purification of N16 type 2 fimbriae from a

crude sonicate by sequential chromatography first on Bio-Gel A-5m and

then on the MAb immunoaffinity column, an attempt was made to purify

N16 type 1 fimbriae and type 2 fimbriae directly from a crude French

press supernatant by immunoaffinity chromatography on rabbit IgG

(RIgG) anti-N16 type 1 fimbriae and rabbit IgG anti-N16 type 2

fimbriae columns, respectively. When a 90 ml aliquot (252 mg protein)

was chromatographed on the polyclonal anti-N16 type 2 fimbriae

immunoaffinity column, the unbound fraction contained both types of

fimbriae, and the bound fraction, desorbed with 7 M urea buffer,

contained type 2 fimbriae with perhaps a trace of type 1 fimbriae, as

determined by immunodot EIA. Since the unbound fraction still

contained plenty of type 2 fimbriae, approximately half of it was

rechromatographed and eluted as before. The quantity of bound protein

desorbed on each run was so low that it was virtually undetectable by

A280; also, the recovery from the second run appeared to be lower

than the first, suggesting that the bound antigen was not being

desorbed completely with 7 M urea.

.~. d


Electron micrograph of A. naeslundii N16 cells showing
indirect immunogold labeling of type 2 fimbriae. The cells
were incubated with R2P antiserum (rabbit anti-N16 type 2
fimbrial immunoarcs) and then with goat anti-rabbit IgG
conjugated to colloidal gold (diameter, 10 nm). The
gold-labeled cells were prepared for EM by the
pseudoreplica technique and were negatively stained with
0.5% uranyl acetate. Bar, 1 pm.


Figure 8.

* *


A dissociation experiment was undertaken to assess the efficacy

of various buffers in dissociating the antigen-antibody complexes

formed between N16 fimbriae and the anti-fimbrial antibodies used to

prepare the immunoaffinity columns; the results are summarized in

Table 4. Although 8 M urea buffer released 85% of Ag-bound MAb

3B5.A1, it released only 28.8% of rabbit IgG anti-N16 type 1 fimbriae

and 18.5% of rabbit IgG anti-N16 type 2 fimbriae. Thus, the 7 M urea

buffer normally used as the dissociating buffer for immunoaffinity

chromatography would be ineffective at desorbing bound fimbriae from

either of the polyclonal rabbit IgG anti-N16 fimbriae immunoaffinity

columns. The most effective dissociating buffer for the polyclonal

immunoaffinity columns was 6 M NaSCN, which gave 95.4% release of

rabbit IgG anti-N16 type 1 fimbriae and 91.0% release for rabbit IgG

anti-N16 type 2 fimbriae. There was no evidence that any of the

dissociating buffers irreversibly denatured fimbrial epitopes or

released fimbriae from the plate during these assays.

Based on the results of the dissociation experiment, the

desorbing buffer was changed to 6 M NaSCN, 0.05 M Tris-HCl, pH 7.5,

and the rabbit IgG anti-N16 type 2 fimbriae immunoaffinity column was

desorbed again to release the fimbriae that were not released by 7 M

urea. The bound fractions recovered from the 'type 2' immunoaffinity

column were pooled, concentrated, and dialyzed versus TBS by

ultrafiltration (Amicon YM10). Crystalline (NH4)2SO4 was added

to 100% saturation, and the precipitated sample was stored at 4 C

until further processing.

The unbound material recovered from the rabbit IgG anti-N16 type

2 fimbriae immunoaffinity column was chromatographed on the anti-N16

TABLE 4. Summary of the efficacy of various dissociation buffers in
disrupting binding between antibodies to N16 fimbriae and N16 fimbriae
coated on Immulon 2 microtiter plates

Antibody samples
Dissociation buffer RIgG a-i RIgG a-2 MAb 3B5.A1

NaSCN, 6 H 95.4a 91.0 85.6
NaSCN, 3 M 47.8 23.4 83.6
NaSCN, 1 M 16.9 6.4 78.0

GuHC1, 6 H 93.0 84.5 ND
GuHCl, 3 M 46.5 30.2 82.2
GuHC1, 1 M 17.1 8.6 76.6

Urea, 8 M 28.8 18.5 85.0
Urea, 4 M 5.3 3.9 73.9
Urea, 2 H 3.1 2.1 33.2
Urea, 1 M 12.1 0.0 7.7

Glycine-HCl, pH 2.5 81.7 64.9 86.7
Glycine-HC1/10% C2H602 b, pH 2.5 72.3 54.0 85.6
NH40H, pH 11.5 16.8 7.4 82.8
NH40H/10% C2H602, pH 11.5 26.8 10.8 73.6

aPercent release of antibody from fimbriae-coated plates.
bEthylene glycol.
Ethylene glycol.

type 1 fimbriae immunoaffinity column. Since the column did not bind

all the type 1 fimbriae on the first pass, the unbound fraction was

rechromatographed for a total of three runs. The bound fractions were

eluted with 6 H NaSCN, pooled, concentrated, and dialyzed versus

saline by ultrafiltration. Examination of the unbound and bound

fractions by immunodot EIA revealed that rabbit Ig was leaching from

the column during desorption of the bound fimbriae with 6 M NaSCN.

Crystalline (NH4)2SO4 was added to the unbound and bound

fractions, and they were stored as 75% and 100% saturated solutions at


Fimbrial samples from the French press supernatant, as well as

samples obtained from the Batch A and Batch B sonicates, were examined

by SDS-PAGE and immunoblot. However, before one can understand the

results of those experiments, an explanation of the nature of

Actinomyces fimbriae and the patterns they exhibit on immunoblots is


Effects of Various Physical and Chemical Treatments
on N16 Fimbriae

The fimbriae of E. coli and other gram-negative bacteria are

polymers of smaller subunits, and under the appropriate conditions,

they can be dissociated into their constituent monomers. The

following experiments were undertaken to (a) determine whether or not

N16 fimbriae had a subunit architecture or exhibited other properties

similar to those reported for the fimbriae of E. coli, (b) provide

additional evidence that N16 type 1 fimbriae differed from type 2

fimbriae, (c) see whether certain procedures caused changes in the

molecular weight or antigenicity of fimbrial bands.

Aliquots of the N16 crude French press sample were subjected to a

variety of physical and chemical treatments. The first was addition

of MgCl2. Unlike E. coli fimbriae (83), N16 fimbriae did not

precipitate with a final concentration of 0.1 H MgCl2.

Since the N16 crude French press supernatant contained both types

of fimbriae, as well as non-fimbrial components, the effects of the

other treatments on each type of fimbriae were demonstrated by

SDS-PAGE immunoblot analysis, using antibodies monospecific for type 1

or type 2 fimbriae. For some of these experiments a convenient method

for exchanging buffers and concentrating samples was needed. Acetone

precipitation appeared to be the ideal choice, but first it had to be

established that acetone precipitation would give total recovery of

fimbriae without adversely affecting their immunological reactivity.

When acetone precipitates of the crude fimbrial sample were compared

to the original sample, they gave identical patterns on immunoblots.

Acetone precipitation was thus employed whenever necessary to

accomplish the objectives cited above.

The effects of some of the physical and chemical treatments on

N16 type 1 and type 2 fimbriae are shown in Figures 9 and 10,

respectively. Exposure to Freon appeared to have no effect on N16

fimbriae; therefore, if extensive lysis were to occur during

sonication or French press shearing of cells to remove fimbriae,

lipids could be extracted from the crude fimbrial sample with Freon

before applying it to gel filtration or immunoaffinity columns.

Treatment with mutanolysin or lysozyme lysozymee not shown) did

not change the fimbrial patterns observed on immunoblots. This

suggested that the ladderlike series of bands >100 kd represented

Immunoblot analysis of the effects of various physical and
chemical treatments on N16 type 1 fimbriae. Aliquots of
the N16 antigen sample were treated as indicated, then
prepared for SDS-PAGE by heating at 100 C for 5 min without
reduction. Each lane was loaded with 1 pg of total
protein, based on the protein concentration prior to
treatment. Prestained molecular weight standards were
mixed with two of the samples; their positions on the blot
were marked with a ballpoint pen, and approximate molecular
weights are expressed in kilodaltons. The blot was
developed with RIO IgG (anti-N16 type 1 fimbriae) at 1 pg
per ml as the primary antibody.

Figure 9.


u u
-4 -4

o o
W ri
& .
.f -4 4

0a 0 *
o ld
N .4 .4 4 i.

n -4 -4 -4
c ) 0000
I 0) IA i-1 .. .i. .,
4 *r*0 00
M C 6 0 0 >ne
K O O OU 1 0
B U t CO il r
I 0 00 0 -
o .'4 .'4~

-I D



~ -- u -
L IJ~,~
~ I r~ r

- -

e U

4 4 J
K 0a
o *

0 4



* a

-4 -4
0 0

200 -

97 -

68 -

43 -

26 -

18 -

14 -


Figure 10. Immunoblot analysis of the effects of various physical and
chemical treatments on N16 type 2 fimbriae. This
experiment was identical to that described in Figure 9,
except that the blot was developed with R2P IgG (anti-N16
type 2 fimbriae) as the primary antibody, and the lane
order of the samples is reversed.

c C

r. I:
o o
0 o
0 0

4 r-4
a a
4 r-4
*4 *4

4 .14 .4 "4 r-4

0 0 a 0
0 0 in 0

r0 0n C
to 0 0 0
( ) 0 0 0

r E

'it Ui

V 0
4J -<

L aO

c x


- 0

0 0 0
a a
0 0 .


Iii i

- 200

- 97

- 68

- 43

6 a 00 s a

1001 a v s-g m

- 26

- 18
- 14

o i'" awl

-rM.Pi---- -MM-tI I

fimbriae of different lengths rather than a fimbrial protein attached

to different lengths of peptidoglycan.

Sonication generated some fragments that were not evident in the

untreated control. The longer the sample was sonicated, the more

fragments were generated and the greater the intensity of the

immunological reaction. However, the generation of different

molecular weight bands by sonication did not appear to be completely

random. The highest molecular weight material was broken down into

smaller fragments, but the breakdown products generally were the same

molecular weight as fragments already present in the untreated

control. The fimbriae were fairly resistant to breakage into

fragments smaller than 100 kd; after 10 min of continuous sonication,

most of the fimbrial bands were still greater than 100 kd, even though

the most immunodominant bands were in the 35-65 kd range.

N16 fimbriae were not affected by 8 M urea or saturated guanidine

hydrochloride, treatments that would dissociate some types of E. coli

fimbriae into subunit monomers (61,62).

Another method that has been reported to disaggregate E. coli

fimbriae, i.e. 0.1 N HC1 at 100C for 5 min (31), did not have a

similar effect on N16 fimbriae. Instead of causing dissociation into

monomers, hot acid caused loss of band resolution, probably as a

result of acid hydrolysis. This also occurred with Lancefield

extraction (0.04 N HC1 at 100C for 15 min). However, 0.1 N HC1 at

37C for 1 h had little, if any, effect on N16 fimbriae; and heating

at 100 C in the absence of a pH<2 did not cause loss of band

resolution. So, it was the combination of a very acidic pH and

100C that was destructive to the N16 fimbriae.

Base was even more destructive than acid, as judged by the total

loss of band resolution and substantially reduced antigenicity when

N16 fimbriae were incubated with 0.1 N NaOH at 370C for 1 h.

Sodium periodate oxidation also caused some loss of band

resolution but not to the extent seen with acid or base hydrolysis.

There was some loss of antigenicity with increasing temperature

and, as might be expected, heating at 100 C for 1 h was the most


N16 type 1 fimbriae could be distinguished from type 2 fimbriae

by their different patterns on immunoblots, i.e. the different

molecular weight distribution of the unreduced inmnunodominant bands.

Disregarding their intrinsically different patterns, the two fimbrial

types generally behaved similarly in response to the various

treatments. For example, neither was affected by high concentrations

of urea or guanidine HC1, but both were most adversely affected by

incubation with 0.1 N NaOH. The decrease in antigenicity of N16

fimbriae when incubated with NaOH or heated at 1000C for 60 min is

consistent with the findings of Masuda et al. (77) for A. viscosus

WVU627 fimbriae. Although they reported that WVU627 fimbriae were

also labile when incubated with 0.1 N HC1 at 37C for 60 min, a

similar effect was not observed with N16 fimbriae. However,

immunoblot analysis of SDS-PAGE-resolved proteins may not have been

the best way to examine the effects of the various physical and

chemical treatments because denaturation of the proteins by heating in

SDS-PAGE sample buffer may mask the true effects of the other

treatments. It might have been better to use an assay that could

quantitate the amount of immunoreactive fimbriae remaining after

treatment, perhaps by using LRI, XIEP, EIA, or radioimmuno-

precipitation. On the other hand, SDS-PAGE immunoblots were probably

the best way to see the effects of sonication.

A. naeslundii N16 fimbriae were not disaggregated by any of the

methods that would dissociate E. coli fimbriae, pili, or flagella

(31,61,62,70,83). These results are consistent with the observation

that A. viscosus T14V fimbriae cannot be completely dissociated either


Immunoblots were also used to assess the effects of various

proteases on the molecular weight distribution and immunological

reactivity of N16 type 1 and type 2 fimbrial bands. All the proteases

eliminated the immunoreactive bands >200 kd for both types of

fimbriae. Samples treated with papain, Pronase, or proteinase K

exhibited a total loss of immunoreactive type 1 and type 2 fimbrial

bands. However, since papain caused only limited digestion of the BSA

control as determined from the presence of multiple bands in the range

below 35 kd on a silver-stained SDS-PAGE gel, it is possible that

papain and perhaps the other enzymes destroyed the fimbrial epitopes

without completely digesting the fimbriae.

The effects of chymotrypsin, trypsin, and V8 protease on N16 type

1 and type 2 fimbriae are shown in Figure 11 and Figure 12,

respectively. Each of these enzymes caused limited digestion of both

types of fimbriae, but the resulting immunoblot patterns were

different for each type of fimbriae. Trypsin caused a total loss of

band resolution for type 1 fimbriae; except for two discrete bands at

<14 kd, trypsin-treated type 1 fimbriae gave an immunoreactive smear

from 40-200 kd on immunoblot. On the other hand, the immunoreactive

Figure 11. Immunoblot analysis of the effects of digestions with
different proteases on N16 type 1 fimbriae. The lanes
containing untreated or enzyme-treated N16 antigens were
loaded with 2 Vg of N16 protein, based on the
concentration of the N16 sample prior to treatment,
whereas the lanes containing enzyme only (lanes 3, 5, and
8 from left to right) were loaded with quantities of
enzyme equivalent to those in the enzyme-treated samples.
All samples were unreduced, except for the one marked R.
Prestained molecular weight standards were mixed with the
untreated control in lane 1. The blot was developed with
R10 IgG (anti-N16 type 1 fimbriae) at 2.5 ig per ml as
the primary antibody.

200 -

97 -

68 -
4 -

97- -I


9 3#
4) 0@
I 44
00I I
>* a* a a a
0 e S J 4)4)4)
.1 4) 4) 0 0 0

M emebb
g; | r
EE 3

26 -

18 -
14 -

Figure 12.

Inmmunoblot analysis of the effects of digestions with
different proteases on N16 type 2 fimbriae. This
experiment was identical to that described in Figure 11,
except that the blot was developed with R2P IgG (anti-N16
type 2 fimbriae) as the primary antibody, and the lane
order reads right to left.

4 0

00 41
40 4J


w0 u 0 0 + )
4). 4 4 I


'- I

I 0
0 C


- 200

- 97

- 68

- 43

- 26

- 18
- 14

type 2 fimbrial bands after trypsin treatment were mostly between the

14 and 43 kd markers, and for the most part, the bands were discrete

rather than unfocused. The effect of chymotrypsin on type 1 fimbriae

was very similar to that of trypsin, whereas for type 2 fimbriae, the

pattern obtained by digestion with chymotrypsin was quite different

from that obtained with trypsin. V8 protease-treated samples were

less immunoreactive than the trypsin- or chymotrypsin-treated samples,

even though V8 protease appeared to be less efficient than the other

two enzymes at digesting the higher molecular weight fragments.

Effects of Temperature and Reduction on N16 Fimbriae

On immunoblots the most inmunoreactive N16 type 1 and type 2

fimbrial bands undergo a shift in apparent molecular weight in

response to increases in the temperature for SDS-PAGE sample

preparation, as demonstrated in Figure 13. In unreduced samples

heated at 37 C for 30 min there are a series of type 1

immunoreactive fimbrial bands at or slightly below the 43 kd molecular

weight marker, whereas after heating at 100 C for 5 min, the

immunodominant bands are in the 57-65 kd range. Both sets of bands

are present in the sample heated at 650C for 15 min, so the

immunoblot pattern at this intermediate temperature shows the

"transitional state" of proteins switching from their lower molecular

weights at 37 C to higher molecular weights at 100 C. Reduction

with 0.01 M DTT (or 1% 2-mercaptoethanol) further simplifies the

pattern of bands so that with samples reduced at 100 C the primary

type 1 fimbrial bands on immunoblots are as follows: the uppermost

band has an apparent molecular weight of about 65 kd; there is a broad

area of immunoreactivity in the 57-60 kd range, which represents at

Figure 13.

Immunoblot analysis of the effects of temperature and
reduction on the type 1 and type 2 fimbriae of
A. naeslundii N16. Aliquots of the N16 antigen sample
were treated with SDS-PAGE sample buffer, with or without
0.1 M dithiothreitol (DTT), by heating at the temperature
and times indicated. Each lane of the SDS-PAGE gel was
loaded with 1 Vg of protein. Approximate molecular
weights are expressed in kilodaltons. The left half of
the blot was developed with R10 IgG at 5 ig/ml, and the
right half was developed with R2P IgG at 5 yg/ml.

I + + + 1+1+

oO n an in 4n in in in n oo
(M P4 r iP44 Mi I n

0000 n
S. u1 in
* W. h *
m lO


U u
0 0
0 0
1-1 1-I

0 0 U PS 9%

P4,4m (

- 200

- 97

- 68

- 43

- 26

- 18
- 14

Anti-Type 1

Anti-Type 2



least two bands, including a component at 57 kd that reacts with

normal RIgG; then there is a weaker doublet at about 53-54 kd.

In unreduced samples heated at 37 C, the immunoblot pattern for

type 2 fimbrial bands shows a series of bands near the 43 kd marker,

whereas unreduced samples heated at 1000C show a very reactive

doublet at about 62 and 63 kd and several closely spaced bands at

about 39-40 kd. As with the type 1 pattern, the 65 C type 2 pattern

showed the transition from the 37 C to the 1000C pattern. Type 2

fimbriae were also affected by reduction in that the prominent doublet

at 62-63 kd in the nonreduced sample is replaced by a single 63 kd

band, and a minor doublet is still evident at 39-40 kd. On type 2

blots there was a very weakly immunoreactive band at about 23 kd that

was present in reduced and nonreduced samples prepared at any

temperature. A similar band at 18 kd (not visible in Figure 13) was

sometimes visible on type 1 blots. These low molecular weight bands

might represent fimbrial subunits (E. coli fimbrial subunits are

approximately this size); however, they may just be products of


The observation that A. naeslundii N16 fimbrial proteins are heat

modifiable is consistent with the report of Yeung et al. (108) that

A. viscosus T14V type 1 fimbriae have a heat modifiable subunit with

an apparent molecular weight of about 50 kd at 37C or 65 kd at

100 C. Similarly, Yeung et al. reported that the fimbrial subunit of

A. naeslundii WVU45 type 2 fimbriae has an apparent molecular weight

of about 48 kd at 37 C or 60 kd at 100 C (109, M. K. Yeung,

personal communication).

It is likely that the observed shifts in molecular weights of

fimbrial proteins in response to heat are the result of changes in

conformation. Strong intramolecular interactions at the lower

temperature may prevent complete unfolding of the protein and

saturation with SDS, thereby altering the mobility in gels. The

results obtained with the cloned T14V type 1 fimbrial subunit (108)

tend to support such an explanation. Intermolecular interactions of a

noncovalent or covalent nature could also prevent complete

denaturation of fimbrial complexes and account for some of the

observed differences in the 37C and 100 C immunoblot patterns.

The observation that N16 fimbriae were susceptible to reducing

agents differs from all previous reports on fimbriae from other

strains of Actinomyces (76,77,108). However, Masuda et al. (76) were

unable to get many of their fimbrial samples to migrate into 5 or 7%

SDS-PAGE gels, and their single percentage gels would not have the

resolving power of the 5-20% gradient gels used in this study. Thus,

solubility problems and inadequate resolution of proteins with very

similar molecular weights might have prevented them from detecting the

effects of reduction on fimbrial bands. This is especially likely in

light of the finding of Yeung et al. (108) that the cloned T14V type 1

fimbrial subunit migrated as a single band in 10 or 12% gels but as a

doublet in 5-12% gradient gels.

If one were examining the effect of reduction on a purified

protein complex that migrated as a single band on SDS-PAGE gels, one

would expect that a reduction would cause a single higher molecular

weight band to dissociate into one or more lower molecular weight

bands, depending on whether the complex was composed of identical or

dissimilar polypeptide chains held together by disulfide bonds.

However, fimbrial immunoblot patterns are too complex to allow such a

simple interpretation. A comparison of the differences between the

patterns of type 2 fimbriae prepared with or without reduction at

1000C indicates that reduction caused changes in the visible bands

at virtually every molecular weight level. There are too many bands

present to speculate on whether any particular band was derived from

another, but the fact that the reduced samples appear to have fewer

and better focused bands than nonreduced samples suggests that the

reduction of intra- or intermolecular disulfide bonds allows complete

denaturation and SDS saturation of the proteins, thereby overcoming

conformational variations that affect the mobilities and apparent

molecular weights of fimbrial fragments.

Another possible explanation for the presence of fewer reactive

bands in reduced samples is that reduction may destroy the fimbrial

epitopes on some bands so that they are no longer visible on

immunoblots. This explanation would be more plausible if the blots

were developed with monoclonal antibody; but since they were developed

with polyclonal antibodies, it is not likely that destruction of

epitopes could explain these results.

In summary, since complete dissociation of N16 fimbriae was not

possible, immunoblot patterns exhibited a multitude of fimbrial

bands. The simplest pattern of bands was obtained with fimbrial

samples heated at 100 C in the presence of DTT or

2-mercaptoethanol. In addition to the ladderlike series of bands in

the high molecular weight range, there were several strongly

immunoreactive bands (referred to as fimbrial subunits) in the 35-65

kd range. A. viscosus T14V type 1 and type 2 fimbriae have been

reported to give immunoblot patterns (32,108) similar to those

obtained in this study with N16 fimbriae. The fimbrial subunit

patterns for unreduced and reduced A. naeslundii N16 type 1 and type 2

fimbriae are summarized in Figure 14.

Assessment of the Purity of Fimbrial Samples by SDS-PAGE-Inmunoblot

When aliquots of the type 1 and type 2 fimbriae purified from the

polyclonal antibody immunoaffinity columns were examined by SDS-PAGE

and immunoblot, it was apparent from the presence of immunoglobulin

heavy and light chains on the silver-stained SDS-PAGE gel (not shown)

that both samples were contaminated with RIgG that had leached off the

immunoaffinity columns. Consequently, immunoblot analysis of these

two fractions was complicated by the reaction of the peroxidase

conjugate with RIgG present in the fimbrial samples. The immunoblot

patterns of these two samples as well as those for the crude French

press supernatant, Batch A A-5m peak 1, Batch B 35% saturated

(NH4) 2SO4 precipitate of the crude sonicate, and two samples of

the MAb-purified type 2 fimbriae, are shown in Figure 15.

The fimbrial subunits for both type 1 and type 2 fimbriae appear

to be present in all the samples except the two MAb-purified

fractions, although the weak type 1 bands in the type 2 fimbriae

purified from the RIgG column may actually be a reaction of conjugate

with immunoglobulin heavy chains contaminating the preparation. The

65 kd type 1 fimbrial subunit is readily apparent only in the type 1

fimbriae recovered from the RIgG column, indicating that this fraction

was enriched for type 1 fimbriae. That band is also present, but not

as obvious, in the French press supernatant, sonicate, and A-5m peak 1

samples, all of which should contain both types of fimbriae. The 63

kd type 2 fimbrial subunit is present in the type 2 fimbrial samples,

Type 1 Type 2

Figure 14.

Line drawing summarizing the "fimbrial subunit" patterns
obtained on immunoblots for A. naeslundii N16 type 1 and
type 2 fimbriae for samples heated at 100 C for 5 min.
R = reduced with 0.01 M DTT; NR = unreduced. Approximate
molecular weights are expressed in kilodaltons.

- 68

- 43



Figure 15.

Immunoblot analysis of several preparations of N16
fimbriae at different stages of purification. Crude
French press (FP) supernatant was the starting material
for immunoaffinity purification of type 1 and type 2
fimbriae from the RIgG anti-type 1 and anti-type 2
columns, respectively. The middle lane contains the
Batch B 35% (NH4)2SO4 precipitate of the crude 2 min
sonicate. Also shown are A-5m peak 1 from Batch A and two
samples of type 2 fimbriae purified on the MAb 3B5.A1
column. All samples were reduced with DTT (0.01 M final
concentration). The blot was developed with a combination
of R10 IgG at 5 pg/ml and R2P IgG at 2.5 pg/ml, so
both types of fimbriae are evident.


0 0 .0 0t
P M *I-4 E- E

o- 18
S i osa 14

0 .r4 N 0 0.N N



a [ 43




but the 39-40 kd doublet is not apparent. In fact, the 39-40 kd

doublet was present only in the crude French press supernatant, so

there is some question as to whether or not it is a type 2 fimbrial


An attempt was made to remove the RIgG contaminating the

immunoaffinity purified type 1 and type 2 samples by adsorption with

protein A-Sepharose, but SDS-PAGE analysis of the samples after

protein A adsorption indicated that immunoglobulin chains were still

present in the fimbrial samples.

The protein A-adsorbed fimbrial samples and several of the other

samples were examined again by inmunoblot analysis as shown in Figure

16. Duplicate blots were developed with either R10 against N16 type 1

fimbriae or R2P against N16 type 2 fimbriae. Also included for

comparison to N16 fimbriae were samples of purified T14V type 1 and

type 2 fimbriae. The T14V fimbrial samples appeared to be pure since

type 2 fimbriae did not react with RIgG against N16 type 1 fimbriae or

vice versa. On the other hand, the N16 type 2 fimbriae recovered from

the MAb immunoaffinity column exhibited minor contamination with type

1 fimbriae, as determined previously by XIEP (Figure 6C). The subunit

pattern for T14V type 1 fimbriae on the immunoblot developed with RIgG

against N16 type 1 fimbriae differs from the N16 type 1 fimbrial

subunit pattern, whereas the type 2 patterns for each appear to be

more similar. Another interesting observation is that the 63 kd N16

type 2 subunit (in the French press supernatant reacted with anti-type

2) appears as a doublet. On all the other immunoblots it appeared to

be a single band, but this sample was electrophoresed longer than

usual, and better resolution was obtained.

Figure 16.

Immunoblot analysis of A. naeslundii N16 and A. viscosus
T14V fimbrial samples. All samples were reduced with DTT
at 100 C. The left half of the blot was developed with
R10 IgG anti-N16 type 1 fimbriae, and the right half was
developed with R2P IgG anti-N16 type 2 fimbriae.

Anti-Type 1

0 U
0 0
14 M

I >

a >

200 -

97 -

68 -


(0 0



Anti-Type 2

ao a

rl N


E6 04 64 H


26 -

18 -

14 -



I i
e I

, 4

In summary, the N16 fimbrial samples obtained by immunoaffinity

chromatography were enriched for one of the fimbrial types, but none

of the samples were completely pure. The MAb immunoaffinity column

appears to offer the most promise for purification of type 2 fimbriae,

but it might be necessary to incorporate a detergent into the starting

buffer to prevent nonspecific adsorption of type 1 fimbriae. As for

the RIgG columns, it was probably too much to expect that they could

produce pure fimbriae from totally crude starting material. However,

since they required such harsh conditions to elute bound fimbriae that

some RIgG also was stripped from the columns, it is unlikely that

these immunoaffinity columns would be as useful as MAb columns.

Antigenic Relatedness of Actinomyces Fimbriae

Ouchterlony Analysis

Like the type 1 and type 2 fimbriae of A. viscosus T14V (20), the

type 1 and type 2 fimbriae of A. naeslundii N16 were found to be

unrelated antigenically, as demonstrated by the reactions of

non-identity between type 1 and type 2 fimbrial precipitins in

immunodiffusion (Figure 17A). A reaction of partial identity between

the type 2 fimbrial bands precipitated by MAb 3B5.A1 and the

polyclonal reference antiserum (Figure 17B) provided additional

evidence that MAb 3B5.A1 is specific for an epitope on the type 2

fimbriae; however, the presence of a spur suggests that there are

other epitopes on N16 type 2 fimbriae that elicit the formation of Abs

in a polyclonal response to the antigen. MAbs 2A3.B3 and 2BS.B6 did

not precipitate either type of fimbriae, perhaps because (a) the

proportions of antigen and antibody were not at equivalence or (b) the

*2B5 C3B5


2A3 3B5


Figure 17. Immunodiffusion reactions of antibodies to N16 fimbriae.
The center wells contained N16 fimbriae partially purified
from a French press supernatant by precipitation at 20%
(w/v) (NH4)2SO4. Rabbit antiserum R29 against N16
whole cells precipitated both type 1 and type 2 fimbriae.
A) A reaction of non-identity between type 1 and type 2
fimbriae. Antisera were R10 against N16 type 1 fimbriae
and R2P against N16 type 2 fimbriae.
B) A reaction of partial identity between the type 2
fimbrial bands precipitated by MAb 3B5.A1 and the
polyclonal reference antiserum R29. MAb 2A3.B3 and
MAb 2B5.B6 did not precipitate the fimbriae.

density of fimbrial epitopes was too low, as suggested previously by

the sparse distribution of gold particles in immunoelectronmicroscopy.


The antigenic relatedness among fimbriae from different isolates

of A. naeslundii and A. viscosus was examined by XIEP-A. Unlabeled

homologous antisera raised against whole cell antigens were

incorporated into gels to precipitate fimbrial and sometimes non-
fimbrial antigens; in addition, an 1I-labeled antibody specific

for one type of fimbriae was incorporated into the gel so that

autoradiography would reveal which type of fimbriae reacted with the

monospecific antibody. Six fimbriae-specific antibodies were examined,

and the results of reactions with A. naeslundii antigens are

summarized in Table 5.

R10 rabbit IgG anti-N16 type 1 fimbriae reacted strongly with the

type 1 fimbriae of all 10 A. naeslundii serotype 3 isolates but not

with the type 2 fimbriae (Figure 18). It cross-reacted with the type

1 fimbriae of A. naeslundii serotype 2 W1544, but it did not react

with A. naeslundii serotype 1 W826 (WVU45), a serotype that has been

reported to lack type 1 fimbriae (17).

R2P rabbit IgG anti-N16 type 2 fimbriae reacted strongly with the

type 2 fimbriae of the 10 A. naeslundii serotype 3 isolates, but it

was essentially negative for type 1 fimbriae and the fimbriae of other

serotypes of A. naeslundii. Heterologous fimbriae were sometimes

visible on autoradiographs, but the intensity was so weak that it was

difficult to determine whether the reactions were the result of

nonspecific binding or very low levels of cross-reactivity; normal

rabbit IgG gave similar results.

CH4N N'4S Ii-- 0

r-H r- -" --4 -4 H- 1 0



4 :3

C 4

. O1














>1 04 1-

0 E-



u .0

04 E4

S 0
0 E-4




00 r-N o
%0 CM CN
S3 No 0 '
r-4 -4 C
QV V39 :3
> > > cV-4

U0 ,


C* %


u 0


H- H-4 -4 l H- r- r- r-4 1 0

00000 I OOO-O

S- -qW -q -- 00



0 0 0 O
0 A .-4 *.*4 Li0
U-4 LA Li 0 J3 S
.0 .0 .0 O -A
A ,A ,-4 4 4.4
*-4 .4 .* 4 4.
ad rd l CM
l44.l 44M CM (

) 4) 0) 0.

.oto4J 4J
> > >
-4-4 r-4-4 -4

.Hi .1-4 .- .Hi .,i .H H

4J 4J 4 4J 4J 4.

A -4 (n -A .4 -A.
60.0 60 60 0 60
-o (.d 0uo co c0

U 14 WK WA o

*q .q -q -q -q -T Iq <



-4 0
3 o
0 4)
4 a

0 z
o 0

0 0

.C 0

) 0

0 0

-4 .-
0 *40

o ,
60 ).


0 *
-1 01

0 -4

a o.
0 4.)
0 4

a 0.

o a

*4) C




*rl \0

0 44

. 0 *a
C 0 0

CO i-ice
O1 0 .-4 a

0 0
40) 0 4

. M 0
*A4 v- 4 C%
o-4 )OC

4 I4 0 0
(-4 0A-

cc 0 -1 f4-

0 00 00
*4 0 4
-C I 0
0.0 i- 0

L W 4 0 0
0d l I44 6

0 0 W0
%.0 0 -41

0 r-4 4 O

0 *4 d0 0 )

ul (-4 0 1-4
Q0 0 V
p ao ea

a! 00

" a


V Ca
0 0
-4 .

0 4
0 o

G 0

4 4
44 CO
0~ 0
0) 4c
0 0

c c

g4 .
as a
4 4

0 a

0. 0
4) (U

e a0

4) .ca
0 0a

-C 4) 0 44
W La 44
90 '-0
*1 0

0 .0

3 t'4.
*a 0 0
4 *I 3
0 l <0i
0 0.

* 4ja 4