A‘\ fl. .
- nA ‘
'3, ;,‘.{‘,J"“ '1..",:'a~ ‘ ,,‘ 5“
- . ,n- . ‘ A .. . ‘ .
'..‘,“‘ - . 7n.- uwe " Xhfimx‘m-Jath
u . ' "M" ‘ w- r . ig‘yx‘sl «51m
’ r ,a. "flr. “”3.
(ch

as 4‘ Hr ‘ u. .
' .‘ I“, l.‘ km
(in -k—x um ‘i‘ -. 1..
'1‘}: H‘ .\|A‘.“
a? "I‘d“.larl‘
L; {on 'rmu \
‘\PI.‘."1M~'\IPMHG
H { . 15'“qu

. V“ ‘
f3?“

“Fr-w

"win
a ӣ33,;
u" Irv

.
14'

f‘. '
, fin, :- A ’ . E}. 41...} V . .n
o 5" §< -' «EM “A“ \ .

1*!
Mum: 1m."

'7’
MI! r - 'A«
Mr? ,3, ,

NET. y."
t ,, ,

. .,, ..

an“;

I

... ,
m, u...
I— ...

r q... ,
‘1..,«~

v ‘ .
r,»-‘ r! M
. h,” -u

m
....

- . '-«o-; $
. up ... u, . 4 -h
n w l~v’»’ m .; ,-.‘,, >. .. , y-ovca

. . . m . q I - .4...r.........
- ~ ,. .. ..,. v -.

1- _. '- . wer-

 

UNIVERSITY LIBRARIES

Illlllll‘lllllllll‘lllll Illlll ll I

3 1293 00913

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

Neisseria gonorrhoeae IgAl Protease:
Construction and Characterization of
Mutants and the Response to Iron Limitation

 

presented by

Russell James Shoberg

has been accepted towards fulfillment
of the requirements for

 

Ph.D. degreein Microbiology

Major professor

 

Date My /3 /?Q/
V a

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

LIBRARY
Mlchlgan State
Unlverslty

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before due due.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU lo An Affirmative ActlorVEquel Opportunity Institution
amine-m

 

 

 

NEISSERIA GONORRHOEAE IgAl PROTEASE:
CONSTRUCTION AND CHARACTERIZATION OF
MUTANTS AND THE RESPONSE TO IRON LIMITATION

BY

Russell James Shoberg

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Microbiology and Public Health

1991

ABSTRACT
NEISSERIA GONORRHOEAE IgAl PROTEASE:

CONSTRUCTION AND CHARACTERIZATION OF
MUTANTS AND THE RESPONSE TO IRON LIMITATION

By

Russell James Shoberg

The IgAl proteases produced by Neisseria gonorrhoeae are theorized
to be important virulence factors for the microbe. ' One of the
objectives of this research was to address the question as to what other
functions these enzymes might be involved in. The gene encoding IgAl
protease (iga2) was cloned from N. gonorrhoeae GCM 740 in Escherichia
coli. Following characterization, the iga2 locus was mutated at two
sites by site-specific deletion and disruption and Iga- variants of GCM
740 were constructed by transformation. These Iga- variants, GCM 740A4
and COM 740A2, were characterized with regard to potential Iga-related
phenotypes. Altered membrane protein profiles were consistently
observed on stained SDS-PAGE gels. Through analysis of variants with
restored igaZ alleles, an outer membrane protein of 56 kDa was observed
to be expressed in an Iga+-dependent fashion as well as a 34.5 kDa
cytoplasmic membrane protein which behaved similarly in GCM 740A4. A
second objective was examination of the IgAl protease susceptibility of
outer membrane proteins in the Iga- variants. Proteins susceptible to
IgAl protease hydrolysis were observed in the outer and cytoplasmic
membranes of both of the Iga- variants and the Iga+ parental strain.
IgAl protease-susceptible outer membrane proteins were demonstrated in

E. coli and Actinobacillus pleuropneumoniae as well. The identities and

functions of the susceptible proteins are undefined. A third objective
was the examination of the effects of ferric iron limitation on IgAl
protease activity. Growth curve experiments with GCM 740 in iron-
limited and iron-replete media demonstrated an iron-associated effect as
a two to six-fold increase in extracellular IgAl protease activity was
observed under iron-limited conditions. The mature enzyme was shown to
be unaffected by the concentration of ferric iron in the media.
Further, the increased protease activity was not an effect of the media
on the mature enzyme. A potential, well-conserved Fur repressorobinding
site was identified in the promoters of two iga2 genes. However, this
potential regulatory region was not regulated by Fur in a known E. coli
Fur- strain. Also, DNA hybridization analysis demonstrated the lack of

a fur homologue in the Neisseria gonorrhoeae genome.

This work is dedicated to:

Judi- my life partner and best friend
None of this could (or would) have happened
without all of your efforts

my parents-
for making learning fun

and
Dr. A. E. Duwe
(1920-1991)
for teaching me that science could be a be a way of life.

iv

ACKNOWLEDGEMENTS

There are so many of you- and so little space on this page!

I would like to express my gratitude to the members of my advisory
committee: Drs. Robert Brubaker, Susan. Conrad, Thomas Deits, Loren
Snyder, and Clarence Suelter.. A hearty thanks and congratulations to
my major professor, Dr. Martha Mulks, your first Ph.D. student has
finally made it out! This would not have been possible without the
assistance and counsel of all of you.

To the "Bru"-crews- past and present, and those in Esselman's lab-
Thanks for 'being there to talk to on those late nights and. more
importantly, for sharing equipment, reagents, and wisdom. You know who
you are.

To my fellow students and technicians- Thanks a million for the
ideas, critiques, and help with new procedures. Every one of you left
an imprint on my training. Again, you know who you are.

To Judi- for holding all of the dinners and for trusting me to
actually be working at weird hours. To Judi and Abby- for being there
to listen when things didn't work and for being there to listen when
they did.

Lastly, to my brothers Uracil and other Adenine- I guess it's time
to shut out the lights and head on home, the party's over.

TABLE OF CONTENTS

List of tables ...................................................... viii

List of figures ....................................................... ix

Introduction:
Gonorrhea .................................................. 1
Experimental model systems ................................. 5
Genetics ................................................... 7
Virulence functions ....................................... 12
IgAl proteases ............................................ 32
IgAl proteases of other species ........................... 50
References ................................................ 77

Chapter 1: Construction and Characterization of Isogenic IgAl
Protease-Deficient Mutants of Neisseria gonorrhoeae

Abstract ................................................ 102
Introduction ............................................ 103
Materials and Methods ................................... 110
Results ................................................. 126
Discussion .............................................. 160
Acknowledgements ........................................ 163
References .............................................. 164

Chapter 2: Proteolysis of Bacterial Membrane Proteins by Neisseria
gonorrhoeae Type 2 IgAl Protease

Abstract ................................................ 170
Introduction ............................................ 171
Materials and Methods ................................... 173
Results ................................................. 179
Discussion .............................................. 189
Acknowledgements ........................................ 193
References .............................................. 194

Chapter 3: Expression of the Neisseria gonorrhoeae Type 2 IgAl
Protease Gene Under Iron Limited Conditions

Abstract ................................................ 199
Introduction ............................................ 200
Materials and Methods ................................... 204
Results ................................................. 210
Discussion .............................................. 227
Acknowledgements ........................................ 234
References .............................................. 235

vi

Summary and Conclusions .............................................. 240

Appendix: Restriction Site Polymorphism in Genes Encoding Type
2 but not Type 1 Gonococcal IgAl Proteases .............. 243

vii

LIST OF TABLES

INTRODUCTION:
Table 1. Antibiotic resistances observed in N. gonorrhoeae

isolates .................................................... 30
Table 2. Permissive substrates for the N. gonorrhoeae type 2

IgAl protease ............................................... 37
Table 3. Molecules that are nonpermissive substrates for IgAl

proteases ................................................... 38
Table 4. Techniques used to analyze the IgAl proteases ................ 52
Table 5. Bacterial species demonstrated to produce IgAl

protease .................................................... 59
Table 6. Comparison of the better described IgAl proteases ............ 63
CHAPTER 1:
Table 1. Properties of plasmids constructed and used ................. 111
Table 2. Summary of pulse-chase labelling data ....................... 144
CHAPTER 3:
Table l. The effect of exogenous iron concentration on activity

of an enriched preparation of IgAl protease ................ 219
Table 2. The effect of relevant growth media on activity of a

purified IgAl protease preparation ......................... 220
APPENDIX:
Table 1. Characteristics of N. gonorrhoeae strains used in

this study ................................................. 246

viii

LIST OF FIGURES

INTRODUCTION:

Figure

Figure

Figure

Figure

l. The IgAl hinge region duplicated amino acid octamer

showing the N. gonorrhoeae types 1 and 2 IgAl

protease cleavage sites .................................... 34
2. Diagramatic representation of the N. gonorrhoeae

type 2 IgAl protease secretion process ..................... 44
3. Diagramatic representation of the proteins produced

during the processing of N.gonorrhoeae IgAl

proteases as proposed by Halter, et al ..................... 47
4. Alignment of the IgA1,IgA2m(1) and IgA2m(2) hinge

regions showing the defined and proposed sites of

cleavage of several IgAl protease producing bacteria ....... 61

CHAPTER 1:

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

1. Restriction site maps of recombinant plasmids
constructed in this study ................................. 114
2. Diagram depicting the steps involved in construction
of the various recombinant deletion-disruption
plasmids used in this study ............................... 117
3. Restriction site maps of the region around the iga2b
locus in strain GCM 740 and the A-iga2b loci of its
isogenic strains GCM 740A2 and GCM 740A4 .................. 128
4. Growth rates of N. gonorrhoeae GCM 740 and two isogenic
Aiga2 mutants in a defined broth medium at 37°C ........... 129
5.Coomassie blue-silver stained SDS-PAGE gel of isolated
outer membranes of GCM 740 and its two isogenic A-iga2
mutants, GCM 74OA4 and GCM 740A2 .......................... 131
6. Coomassie blue-silver stained SDS-PAGE gel of isolated
cytoplasmic membranes of GCM 740 and its two isogenic
A-iga2 mutants, GEMS74OA4 and GCM 740A2 ................... 134
7. Autoradiograph of I surface labeled N. gonorrhoeae
cells separated over a 7.5-12.5% SDS-PAGE ................. 136
8. Extracted outer membrane proteins from N. gonorrhoeae
GCM 740, GCM 74OA4, and GCM 740A2 grown under aerobic
and anaerobic conditions on GCB solid medium .............. 139
9. Growth curves of N. gonorrhoeae GCM 740, GCM 740A4,
and GCM 740A2 grown under iron-replete and iron-
limited conditions ........................................ 140
10. Coomassie blue-silver stained SDS-PAGE gels of
isolated outer membranes from N. gonorrhoeae GCM 740,
GCM 740A4, and GCM 740A2 separated over a 7.5-12.5%
linear polyacrylamide gradient ........................... 142

ix

CHAPTER 1 (cont'd).

Figure
Figure
Figure
Figure

Figure

Figure

Figure

CHAPTER 2:

Figure

Figure

Figure

11.

12.

13.

14.

15.

16.

17.

1.

2.

3.

Figure 4.

CHAPTER 3:

Figure

Figure

Figure

Figure

Figure

Figure

1.

Autoradiographs

of N. gonorrhoeae cells3pulse-chase
labeled during

exponential phase with
Autoradiographs of N. gonorrhoeae cellgspulse-chase

labeled during stationary phase with
Autoradiograph of IgAl protease assay cleavage

products ................................................. 152

Southern blot analysis of WT, A-iga2 mutant, and

revertant strains of N. gonorrhoeae GCM 740 .............. 153

Coomassie blue-silver stained 7.5-12.5% SDS-PAGE gel of
isolated outer membranes from WT, A-iga2 mutant, and

revertant strains of N. gonorrhoeae GCM 740 .............. 155

Coomassie blue-silver stained 7.5-12.5% SDS-PAGE gel of
isolated outer membranes from WT and A-iga2 variants

of N. gonorrhoeae GCM 740 ................................ 157

Coomassie blue-silver stained 7.5-12.5% SDS-PAGE gel of
isolated cytoplasmic membranes from WT, A-iga2 mutant,

and revertant strains of N. gonorrhoeae GCM 740 .......... 159

Restriction maps of Neisseria gonorrhoeae GCM 740 and

two isogenic gene disruption mutants ...................... 174

Silver stained SDS-PAGE gel of N. gonorrhoeae 0M3 fol-
lowing incubation with and without exogenous pure gono-

coccal type 2 IgAl protease ............................... 182

Silver stained SDS-PAGE gel of N. gonorrhoeae CMs fol-
lowing incubation with Iga2 buffer, heat-inactivated
type 2 gonococcal IgAl protease in Iga2 buffer, and

active IgAl protease in Iga2 buffer ....................... 185

Silver stained SDS-PAGE gel of E. coli DHl 0M5 fol-
lowing incubation with and without exogenous pure type

2 gonococcal IgAl protease ................................ 188

Restriction map of pRJ2, a pBR322-iga2b recombinant

construct encoding IgAl protease activity ................ 205
. The effects of 8-hydroxyquinolone-extracted defined

media on growth rate (A) and extracellular IgAl
protease activity (B) of Neisseria gonorrhoeae

GCM 740 ................................................... 211
. The effects of Desferal-induced iron-limitation on

growth rate (A) and IgAl protease expression (B) of

Neisseria gonorrhoeae GCM 740 ............................. 214
. Silver stained 7.5-12.5% linear gradient SDS-PAGE

analysis of outer membranes of N. gonorrhoeae GCM 740

grown under iron-limited and iron replete conditions ...... 217
. Nucleotide sequence alignments of parts of the IgAl

protease promoter region and the promoter region of the
gonococcal fbp gene with a reported consensus E. coli

Fur binding site .......................................... 221
. Growth (A) nd IgAl protease expression (B) of E. coli

BN402 (Fur ) plus pRJ2 and E. coli BN4020 (Fur-) plus

pRJ2 in fBHIt, an iron rich medium ........................ 223

S-methionine ..... 146

S-methionine ...... 149

CHAPTER 3 (cont'd).
Figure 7.Southern analysis of Neisseria gonorrhoeae and E.coli

genomic DNAs probed with an E. coli fur gene probe ........ 226
APPENDIX:
Figure 1. Restriction and Southern blot analysis of genomic DNA
purified from nine different strains of N. gonorrhoeae....248

Figure 2. Physical maps of gonococcal IgAl protease (iga) genes
and flanking sequences, as determined by restriction
enzyme analysis of cloned genes and by chromosomal
Southern blot analysis with the 4.2 Kb HindIII fragment
from a cloned iga-Zb gene used as a probe ................. 249

xi

INTRODUCTION

The nonmotile, gram negative bacterium, Neisseria gonorrhoeae (aka
the gonococcus) was first described in 1879 by Neisser and was later
demonstrated to be the causative agent of the sexually transmitted
disease gonorrhea. The microbe shows very little diversity in the
sources of energy that it can utilize. Glucose, pyruvate, and lactate
are the only carbohydrates that the organism can use as its sole source
of carbon, respiring primarily via the Entner-Doudoroff and pentose
phosphate pathways (181).

The organism is an obligate human pathogen and has never been
isolated from any other species. However, the chimpanzee has been used
infrequently as an experimental model system and has been shown to
manifest the clinical signs of an uncomplicated gonococcal infection
(145) and the ability to transmit the disease venereally (30).

Gonorrhea. N. gonorrhoeae infection results in an average of 600-
800,000 reported cases of gonorrhea in the United States every year
(264). The disease is most often thought of in the context of a
sexually transmitted disease affecting the genitourinary tracts of
infected persons. This type of infection, known as an uncomplicated
gonococcal infection (UGI), is the most frequent outcome of infection
(89). The UGI, if diagnosed and treated with an appropriate course of
antibiotic therapy is normally resolved without incident. However,
there are more serious ramifications of the disease which although

1

2
rarely fatal, are certainly dangerous.

These complications are a result of bacterial dissemination via
ascension of the reproductive tract or the vascular system. Ascension
of the reproductive tract can result in pelvic inflammatory disease
(PID) in females and a small array of secondary site infections in the
male (89,162). Salpingitis, an inflammation of the fallopian tubes, is
a serious complication which is currently thought to be a leading cause
of infection-produced infertility in the United States (162). A
disseminated gonococcal infection (DGI) can occur in both males and
females after the gonococcus gains access to the circulatory system and
spreads to remote regions of the body, considering the portal of entry
(178).

A UGI in males is characterized by a viscous exudate containing
numerous neutrophils which show intracellular gonococci (180). This
exudate is stimulated by the colonization, damage, and invasion of the
epithelial mucosa in the urethra (267). The invasion and resulting
damage to the mucosa triggers infiltration of neutrophils and plasma
leakage leading to the clinical signs of inflammation, pain, and
swelling (180). Approximately 75% of infected males exhibit these
clinical signs with the remainder considered asymptomatic carriers (89).
Microscopic examination of the exudate and demonstration of gram
negative, intracellular diplococci is considered to be the hallmark of
gonococcal infection of the male urethra (89).

Untreated urethritis in males may also result in infection of the
periurethral glands, prostate, seminal vesicles, and epididymis
(89,267). Although ‘not as potentially dangerous as a DGI, these

sequelae have the potential to cause a urethral stricture and result in

sterility in the male (89).

In the female, a UGI typically results in endocervicitis, a slightly
different outcome from that in the male. The vaginal mucosal
epithelium, unlike the epithelia of the cervix and uterus, is not
permissive for attachment and invasion by the gonococcus (267). While
there is an inflammatory reaction to the bacteria, the clinical signs
may be very slight (89). This is manifested by an alarmingly high
percentage of females who carry an asymptomatic infection (30-70%; 180)
and are only detected by diagnosis of sexual contacts, or through
secondary complications of the disease.

A UGI may also occur at body sites other than the urogenital tract.
Conjunctivitis, a rare occurrence in adults, can be acquired either by
neonates during passage of the birth canal (one reason for the
administration of topical antibiotics in the form of eye drops after
birth) or as a common type of laboratory acquired infection due to
formation of aerosols. Either (or both) rectal and pharyngeal
infections may be commonly encountered in persons who are simultaneously
diagnosed with a urogenital infection, with these secondary sites
commonly being asymptomatic (>90%; 89).

Pelvic inflammatory disease may arise from about 10-15% of UGI cases
and is considered to be due to bacterial ascent through the endometrial
cavity by either 1) migration of bacteria in phagocytes; 2) by adherence
to spermatazoa (25); or 3) by diffusion in the lumenal fluid (162). As
the name indicates, there is a strong inflammatory reaction to the
bacteria, resulting in a febrile state, intense pain, and nausea. This
sequel to the undiagnosed or untreated UGI requires hospitalization and

aggressive antibiotic therapy to prevent the spread of the organism

4
throughout the body in the form of a DGI.

Salpingitis is a, more precisely defined syndrome of PID which
results in a severe inflammatory response in the fallopian tubes. The
swelling and damage that occurs frequently causes permanent disruption
of the fallopian tube, with a final result of infertility or an
increased risk of ectopic pregnancy (162). As with PID, salpingitis
necessitates hospitalization and aggressive antibiotic therapy for
successful resolution and prevention of a D61. However, even if the
course of the infection is halted, there may still have been severe and
irreparable harm done to the reproductive tract.

Probably the most severe and certainly, the most difficult syndrome
to resolve, is the systemic DGI, occurring in ~1% of males and up to ~3%
of females (267). Once the gonococcus has gained entry to the vascular
system, it is rapidly transported to other sites of the body showing a
tropism for the axial limbs (178). This tropism is postulated to be due
to its optimal growth temperature of ~36°C or slightly below human core
temperature (183). The presence of these bacteria in circulation
results in a strong febrile response which can progress to shock if not
treated.

Most commonly, DGI shows a clinical sign, of erythematous skin
lesions having hemorrhagic and/or necrotic centers (178). These skin
lesions often contain culturable gonococci (~50-70% of cases; 180).
Also commonly observed is tenosynovitis, an inflammation of the tendon
sheaths, particularly of the hands and feet (178). Other clinical signs
include polyarthralgias or purulent arthritis, especially of the knees,
elbows, ankles, and hands. This interesting pattern of affected joints

also indicates the tropism of the gonococcus for slightly cooler regions

5

of the body. The affected joints frequently may contain viable bacteria
in the synovial fluid, as well as an increased number of neutrophils and
immune complexes (178). The clinical signs of a DGI can be mitigated by
parenteral antibiotic therapy with the infection being eradicated, and
the risk of permanent damage to the arthritic joints being minimal
(178). There is a hypothesis that some of the arthritic process is due
to the cell wall component, peptidoglycan, and this form of arthritis is
not directly resolvable by antibiotic therapy (68).

Other unusual, but observed, manifestations of a DGI include
endocarditis, hepatitis and perihepatitis, gonococcal meningitis,
pneumonia, and osteomyelitis (178).

Persons having a complement deficiency of C6, C7, or C8 have been
shown to be at a greater risk of acquiring a DGI (207). Characteristics
which have been correlated with the N. gonorrhoeae strains which cause
DGIs are expression of the PorA serotype (38), resistance to normal
human serum killing (243), high susceptibility to penicillin (179), the
arginine-hypoxanthine-uracil requiring auxotype (127), and production of
a type 1 IgAl protease (190).

Experimental model systems. A majority of information regarding the
pathogenesis of N. gonorrhoeae has come from descriptions of the courses
of disease in patients. The only whole animal models for the disease
which simulate the course of gonorrhea and allow for sexual transmission
of the disease are human volunteers and chimpanzees (145). Because of
the ethical and/or logistical restrictions in the ability to do
informative studies on 'virulence, a ‘number of experimental systems
(mainly heterologous) have been deve10ped for studying gonococcal

pathogenesis.

6

There are two main types of homologous systems, thosewhich use
either isolated human cells or tissues. A number of laboratories
utilize freshly prepared human cells to study the interactions of the
gonococcus with professional phagocytes in vitro (226). Similarly, a
number of laboratories have utilized transformed human cell culture
lines to study the separable events of gonococcal adherence and invasion
of human cells (260). Experimental systems have been developed using
HEC-l-B, HeLa, Chang conjunctival, and HEp-2 transformed cell lines
(251,260). These approaches have provided important information with
regard to the host-pathogen interactions at a cellular level, but both
types of approach suffer the lack of an immune system and the lack of
other microbial flora as would occur in a whole animal model.

An approach which takes the cellular system one step further, is the
human fallopian tube organ culture system (FTOC) using explants of
normal human tissue as a target for gonococcal adherence and invasion
(260). It was the use of this model which provided data implicating: l)
the nonciliated columnar epithelial cells of the mucosa as the target
for gonococcal adherence (100), and 2) the cytotoxic effects of
lipopolysaccharide and peptidoglycan for mucosal tissues (48). This
model does allow for some host immune function as the tissue presumably
contains plasma cells, T cells, and macrophages (46). However, the
system lacks important portions of the immune response such as the
ability to have infiltration of plasma components and the inflammatory
response which play a role in the progress of the disease. Experiments
performed using chimpanzee FTOC have demonstrated that this tissue is
also susceptible to adherence, cytotoxic damage, and invasion by N.

gonorrhoeae while those tissues of the baboon are not (163).

7

Animal model systems have been developed and used by various
laboratories to address particular phenomena of interest (4). Examples
include the mouse intraperitoneal and rabbit intraocular models
(176,177); the subcutaneous chamber system. applied in guinea. pigs,
hamsters, and mice to assess microbial ability to scavenge iron or the
roles of lipopolysaccharide in the pathogenic process (3); inoculation
of uterine strictures or transcervical inoculation in mice in an attempt
to simulate urogenital infection (282); the rat model of gonococcal
arthrogenesis (68); two different rabbit models used. to test the
effects of gonococcemia and the somnogenic properties of gonococcal
peptidoglycan fragments (238); and the chicken embryo system used to
assess relative 'virulence of' colony type 'variants or strains (32).
While these models have all been used to test particular facets of
virulence, all fall short of being a bona fide model system because N.
gonorrhoeae: 1) does not establish itself and grow for any length of
time, and 2) does not simulate the course of a gonococcal infection (1.
e. differs in the mode and site of bacterial entry and in the subsequent
events to cause the particular phenomenon being tested).

Genetics. N. gonorrhoeae possesses a circular chromosome, typical of
most bacteria. Estimates have placed its complexity at about 2.1-2.2
Mbp (55). The 51% G+C content is similar to that of E. coli (22). No
bacteriophages have been described that can infect N. gonorrhoeae (39).
Progress has been made in attempts to generate both physical and genetic
maps of the genome which will allow for evolutionary comparisons to be
made (288).

The gonococcus has been shown to possess up to three different types

of plasmids in a strain variable manner. Most strains contain either of

8

the two ~36 kbp conjugative plasmids (24.5 or 25.2 MDa varieties), which
have both been implicated in interstrain and interspecies transfer of
plasmid DNAs (73,233). The 24.5 MDa plasmid has been shown to mobilize
other plasmids to E. coli and other species of Neisseria, while not
mobilizing itself (231). In contrast, the 25.2 MDa plasmid can mobilize
both heterologous plasmids as well as itself to other strains and
species (233). The two plasmids show extensive DNA homology (>60%) and
it has been hypothesized that the 25.2 MDa plasmid arose from the 24.5
MDa plamsid via an insertion event (described below; 231). Prevalence of
conjugative plasmid carriage shows both geographic and temporal
variation (234).

About 96% of clinical strains also contain an ~4.2 kbp (2.6 MDa)
cryptic plasmid (234). Although the complete nucleotide sequence of the
cryptic plasmid has been identified (and therefore its open reading
frames), no defined functions have been ascribed to any of the ten
proposed gene products (133). Interestingly, this plasmid has never
been identified in proline-arginine-uracil requiring strains of the
gonococcus (57). However, neither the presence or absence of the
cryptic plasmid has been correlated with virulence in N. gonorrhoeae
(133).

The third type of plasmid demonstrated in the gonococcus is the R-
factor, with two examples. Gonococci can possess one of five different
fi-lactamase encoding plasmids, named for the geographic regions from
which they were first isolated. Examples are the 4.4 MDa Asian, the 4.0
HDa Nimes, the 3.2 MDa African, the 3.05 MDa Toronto, and the 2.9 MDa
Rio plasmids (58). The Rio and Toronto plasmids are thought to have

arisen from the Asian plasmid by deletion, and are not as common as

9
either of the parental Asian or African plasmids (58). Similarly, the
Nimes plasmid is thought to be an insertion derivative of the African
plasmid (82). These fl-lactamase plasmids can. be mobilized in an
interstrain and even in an interspecies fashion, by both of the
conjugative plasmids (73).

The second type of R-factor plasmid isolated from N. gonorrhoeae
specifies tetracycline resistance and is carried on the 25.2 MDa
conjugative 'plasmid (128). The tetM gene is considered. to ‘be of
streptococcal origin and it is presumed to have inserted into the 24.5
MDa plasmid by transposition (231). This marker has been shown to be
carried both independently of and in conjunction with a fi-lactamase
encoding plasmid (128).

Genetic exchange. With regard to gene transfer in N. gonorrhoeae,
both transformation and conjugation have been shown to occur (39). The
gonococcus is naturally competent for DNA uptake throughout its growth
cycle (12). Seemingly, the only criteria for this competence are the
production of pili, or certain pilin subunits (14); divalent cations;
and an energy source (12). Currently, it is not understood what the
role of pili are in the transformation process. A ten base pair
oligonucleotide uptake sequence (5'-GCCGTCTGAA-3') has been reported to
be necessary for high efficiency uptake of DNA by the gonococcus (79).
This sequence has been demonstrated to bind specifically to several DNA
binding proteins found in the outer membrane and on blebs (59) and is
frequently observed near the ends of: some N. gonorrhoeae genes
(79,88,159). These data suggest that the uptake sequence may play an
important role in the high frequency and specific uptake of gonococcal

DNAs via transformation.

10

After plasmid DNA is taken up by a.txansforming gonococcus, it is
randomly cleaved into small fragments by a nonspecific endonuclease, and
l) religated to form deleted. plasmids (11) or 2) recombined. with
homologous DNA (13). The latter situation is observed more frequently
and seems to require the RecA function (130). These random cleavage
events are thought not to be the result of a specific restriction
endonuclease (14). However, this mechanism of gene transfer has also
been shown to be dependent upon the donor's and the host's restriction-
modification systems (259).

Conjugation requires the presence of either the 24.5 or 25.2 MDa
conjugative plasmids, but does not require pili (65). Cell to cell
contact is required for conjugation and certain of the Opas have been
correlated with decreased conjugative efficiencies (10). Aside from
pilus independence, conjugation in the gonococcus also differs from that
in E. coli in that there is no F'-like system. In N. gonorrhoeae, a
conjugative event results in only the transfer of plasmids and not
chromosomal DNA (14). Conjugation allows for the transfer and faithful
maintenance of plasmids, which differs from transformation in the
gonococcus, apparently in conjugation's independence of the recipient's
restriction-modification system (259).

This relatively low degree of genomic variability suggests that N.
gonorrhoeae possesses a very efficient, or selective host restriction
system. In fact, gonococci have been shown to produce at least seven
different restriction-modification. systems (50). In four of these
systems, both the restriction endonuclease and its corresponding
methylase have been cloned (50). So far, all of these are isoschizomers

of enzymes characterized in other species.

11

Many gonococcal strains methylate the following sequence 5'-GATC-3'
(199), but a corresponding methylase-dependent restriction endonuclease
has not been described, suggestive of the E. coli dam methylase system.
However, Dam- gonococci are not hypermutable (as Dam- E. cell are; 199)
and there is some evidence of a gonococcal restriction enzyme that
cleaves the 5'-GHeATC-3' sequence in a strain which does not produce the
methylase (63). Therefore, the functional role of this putative N.
gonorrhoeae-Dam methylase is unclear at present.

Genetic regulation. N. gonorrhoeae is known for undergoing extensive
genetic regulation, exemplified by phase and antigenic variation of some
of its surface antigens. Initially, these were described as colonial
morphotype and color changes that occurred as the bacteria were serially
passaged in vitro, leading to the Tl-4 colony type/relative virulence
classification scheme described by Kellogg and coworkers (113).

Further characterization of these gross differences in colonial
morphotype defined the changes as correlating with variable expression
of pili (99) and/or outer membrane proteins referred to as opacity
proteins (Opas; 266). In summary, Tl colonial types are lightly
piliated, intermediate to opaque colonies (eg. expressing Opas) and have
been shown to be highly virulent in trial infection studies (21,113).
The T2 colony type is more heavily piliated and opaque, and is also
virulent (113). The T3 and T4 colonies are considered to be nonvirulent
and are both nonpiliated, with expression and nonexpression of Opas,
respectively (111,267). The regulation of each of these colonial
morphotype determinants will be discussed in greater detail in the
particular determinant's section to follow.

Other evidence of genetic regulation in N. gonorrhoeae comes from

12

the environmental stimulus-dependent expression or repression of
described outer membrane proteins. There are examples of outer membrane
proteins or structures which are induced by deprivation of iron, oxygen,
and cysteine (42) and heat shock (125). There are also examples of
outer membrane proteins which are repressed during oxygen limitation
(44). More detail on. each of these regulatory phenomena. will be
provided in the specific sections to follow.

Much of the genetic knowledge that exists regarding N. gonorrhoeae
is a result of studies done on particular virulence functions. Prime
examples include the phase 'variable phenomena. of ‘pilus and. opacity
protein expression. The study of gonococcal virulence functions has
unveiled different phenomena and has resulted in a wealth of information
relevant not only to the pathogenic processes involved in gonorrhea, but
to pathogenic processes and gene regulation in general.

Virulence functions. The following discussion of the virulence
functions of N. gonorrhoeae, will address 1) molecules or structures
involved in adherence to the host cell, examples being pili and Opas; 2)
important structures and antigens implicated in virulence other than as
adhesins, examples being Por, Rmp, Lip, lipopolysaccharide (LPS), and
peptidoglycan; and 3) functions or attributes involved in pathogenesis,
examples being serum resistance, antibiotic resistances, invasion of
host cells, and proteases.

Pili. Probably the single most studied component of the gonococcal
cell surface has been the pilus. Pili have been shown to be essential
for virulence, probably functioning as the primary bacterial adhesin for
the host cell (90). Nonpiliated gonococci have been shown to be less

virulent than piliated variants of the same strain (21,111), and have

13

been shown to be important mediators of bacterium-host cell interactions
in vitro as well (90,283). Additionally, pili are one of the entities
responsible for differentiation of gonococcal colonial morphotypes (99;
described in the preceding text). The pilus, or the expression of
pilin, is also correlated with the ability of N. gonorrhoeae to be
transformed, but the role of pili in the transformation process is
currently undefined (14). Due principally to its role in adherence to
the host epithelium, the pilus is considered to be an essential
virulence function.

The pilus is composed of polymerized 17-22 kDa pilin molecules and
some minor pilus associated accessory proteins (186,236). Pilin
production (and hence expression of pili) is a phase variable
phenomenon. Phase variable expression has been shown to be due to 1)
insertion of a promoterless pilin gene cassette from one of several of
the p118 (silent or nonexpressing) loci to the pilE (expression) locus
in an out-of-frame orientation (132) or 2) due to deletions in piIE, the
expression locus (8,247).

Pili are highly immunogenic in vivo (33) and this would suggest that
they might be an important vaccine candidate. However, the antigenic
type of pilin produced is variable, and this variation has been observed
to occur in vivo as well (271). A pilus preparation subunit vaccine was
tested in a trial and it was found that the vaccinees were not protected
against challenge with heterologous pilin producing gonococci, and were
only marginally protected against challenge with the homologous pilin
producing bacteria (278).

The pilin molecule is comprised of three domains based on

comparisons of interstrain antigenic homology: 1) the amino terminal C

14

region (conserved), the first ~53 amino acids which are highly
conserved; 2) the SV region (semivariable), from amino acid 54-114; and
3) the HV region (hypervariable), from amino acid 115-160 (90).
Antigenic variability results from the insertion of one of several
different pilin gene minicassettes from one of the many p118 loci, in a
faithful (i. e. in frame) orientation behind the pilE promoter (171).
Recombination is proposed to occur via both intragenomic reciprocal
recombination, and during stationary phase by transformation with
genomic DNA from lysed cells (171).

The pilE locus has also been reported to be regulated in trans by
the gene products of the pilA and pilB loci (274). These putative
regulatory loci are located ~3—5 kilobase pairs downstream of pilE and
have amino acid sequence homology with some members of the bacterial
two-component regulatory systems (273). PilA reportedly functions as a
transcriptional activator and pilB serves to down-regulate piIE
expression by an undefined mechanism (273). It is interesting that N.
gonorrhoeaei'has evolved such elegant regulatory' mechanisms for the
expression of pili. This would suggest that the ability to modulate
pilus expression is a vital function for the bacterium in the pathogenic
process.

There has been a report describing a 23 kDa gonococcal surface
protein which participated in pilus-independent, in vitro binding to
glycolipids extracted from human cervical epithelial cells (51). Pilus-
negative gonococci bound. to the glycolipids, but antibodies raised
against a pilus preparation immunoprecipitated. the 23 kDa. putative
adhesin, suggesting that it may be a part of the pilus structure (51).

This type of pilus-associated adhesin has been described in E. coli as

15
well as being proposed for N. gonorrhoeae (146).

Opas. A. major proteinaceous component of the gonococcal outer
membrane which has also been the subject of intense study is the set of
25-30 kDa proteins referred to as Opas (for opacity proteins; 94) and
previously called PIIs (268). Opas exhibit heat-modifiable behavior on
SDS-PAGE analysis, manifested. as an increase in apparent. molecular
weight after denaturation at 100°C relative to that of Opas denatured at
lower temperatures (161). The term opacity proteins refers to the
correlation between the color or opacity of an agar colony and the
expression of these surface exposed, phase variable proteins (266).
Antigenic variation of these highly immunogenic proteins has been shown
to occur, both in vivo and in vitro (246,262).

Some Opas have been shown to be involved in adherence to eukaryotic
epithelial cells (136) and to cohort gonococci (266). Some Opas have
also been reported to stimulate association of gonococci with human
neutrophils and therefore were once given the name "leukocyte
association proteins" (272). Opas can also be the target of serum
bactericidal antibody (16). For several or all of these reasons, Opas
are considered to be important virulence factors of the gonococcus.

The field of study on Opas grew rapidly due to the problems of phase
and antigenic variability. The gene family (thought to be twelve; 171)
coding for Opas are all complete genes which are constitutively
transcribed (261). Therefore, phase and antigenic variable expression
occurs at the level of translation by a umchanism apparently novel to
the gonococcus. There is a five nucleotide pair coding repeat (5'-
CTCTT-3'), present in the early translated region of the gene which,

depending upon the number of repeats present, gives rise to either an

l6

in-frame or out-of-frame gene (261). Therefore, it is the permissive
translation (1. e. production of a full-length Opa polypeptide) of opa
genes that is responsible for the "ON" phase and depending upon which
Opa gene(s) is/are "ON", dictates the antigenic type(s) of Opa expressed
on the surface of the cell. A gonococcus can express zero to several
different Opas simultaneously (246). Variation in the number of coding
repeats has been 1) suggested to occur by DNA slippage during
replication, 2) shown to be independent of recA, and 3) shown to occur
in E. coli suggesting that specialized gonococcal functions are not
necessary for the event to occur (171).

For. The predominant gonococcal outer membrane protein is the porin
protein (For, formerly called PI; 94,268). For has been implicated in
attachment to host cells in vitro (260) and is also the basis of a
serotyping scheme used in characterizing strains for epidemiological
research (275). A given For has an apparent molecular weight of ~3l-35
kDa by SDS-PAGE, which is not modified by either heat or reduction
(268).

There are two mutually exclusive serotypes, PorA and PorB, which
branch out further including 24 serovars of PorA and 32 serovars of PorB
(240). A given strain can express a Per of only one serovar, unlike the
case of Opas, and Por proteins are neither phase nor antigenically
variable within an isolate (41). Per serovar expression has been
associated with serum resistance, antibiotic resistances, nutritional
auxotype, IgAl protease type, and disseminated infections (106).

The Por protein is the primary constituent of the nonspecific,
hydrophilic pore in the gonococcal outer’ membrane, occurring in a

trimeric arrangement (160). There are no described porin proteins in

17

the gonococcus other than PorA and PorB, and a constitutively Por-
gonococcal variant has never been described (40,41). Pors have been
shown to interact with Rmp (198) and LPS (93), two other components of
the gonococcal cell surface, and peptidoglycan (106), a component of the
cell wall. PorA and PorB are oriented differently in the outer membrane
as demonstrated by the resistance of PorA to in situ cleavage by trypsin
and a-chymotrypsin while PorB is susceptible (7).

With regard to its function(s) in virulence, Por can be the target
for both serum bactericidal and opsonic antibodies (103,240). However,
it is less immunogenic in vivo than either pilin or Opas (28). Purified
Por has been demonstrated to insert into liposomes, and to fuse
liposomes with the plasma membranes of eukaryotic cells and hence may
play a role in attachment to and/or invasion of host cells (147,260).
This proposed function in invasion is also supported by data
demonstrating that anti-PorA and anti-PorB monoclonal antibodies block
gonococcal invasion in a monolayer cell culture model system (91).

Rmp. Another major component of the gonococcal outer membrane and
the hydrophilic pore is known as Rmp (aka PIII; 20,94). The acronym
Rmp, refers to its reduction modifiable behavior on SDS-PAGE (161).
When analyzed by SDS-PAGE, reduced-Rmp exhibits an increased apparent
molecular weight relative to its non-reduced form. Rmp is considered to
be structurally and antigenically conserved among gonococcal strains
(105,270), has been cloned in E. coli in a faithful manner (1. e. it is
expressed on the bacterial surface and is immunoreactive; 80), and has
also been purified to homogeneity (148).

In the hydrophilic pore structure, Rmp is closely, but

noncovalently, associated with For (20). Crosslinking experiments have

18
indicated that the pore exists with a 3Porlemp stoichiometry (198).
However, this arrangement is not a requisite for in vitra growth as
Rmp- variants have been constructed (20). At this time the function(s)
of Rmp is/are still undescribed.

The antigenic stability and surface exposure of Rmp once suggested
that it would make an attractive candidate for a gonococcal vaccine.
However, Rmp is only weakly immunogenic in viva, relative to pilin and
Opas (137). Also, Rmp is antigenically related to the E. cali OmpA
protein (a bacteriophage receptor; 18), and anti-OmpA antibodies have
been demonstrated to inhibit complement activation in bactericidal
assays using N. ganarrhoeae (228). By virtue of this property, it is
considered to be an important virulence factor for the gonococcus.

Laz and Lip. A minor constituent of the gonococcal outer membrane is
now referred to as Lip (for lipoprotein; 94). Lip was originally
defined as the H.8 antigen due to its reactivity with a particular
monoclonal antibody (37). However, it has since been shown that there
are at least two other conserved outer membrane proteins which react
with the H.8 antibody (eg. Laz; 80).

Both Laz and Lip are lipoproteins (263,292). Lip has been suggested
to be associated with both the outer membrane and peptidoglycan (36).
Laz has been proposed to function in electron transport but this has yet
to be conclusively shown (36,81). Laz is also observed in many of the
commensal Neisseria species and therefore is not considered. to be
important in virulence (36). Its mention here is due to the former
confusion with Lip, which has been suggested to be involved in virulence
(36).

Both of these proteins have been cloned in E. cali (15,81). Their

19

corresponding N. ganarrhaeae mutants have been constructed and it
appears that the functions of neither Laz nor Lip are required for
viability, at least in vitra (36). The primary structures of both
proteins show a repetitive sequence of AAEAP, found frequently in the N—
terminus of the 17 kDa Laz (81) and comprising the entire 6-7 kDa Lip
protein (108). One cloned Lip has a predicted molecular mass of 6.3
kDa, yet it migrates with an apparent SDS-PAGE molecular weight of ~20
kDa, the difference potentially being: 1) a function of the post-
translationally added lipid moiety, 2) the unusual amino acid
composition, or 3) its repetetive structure (36).

Lip is known to vary in molecular mass from strain to strain (from
18-30 kDa) but is invariant within a strain. Expression of Lip has been
demonstrated in all isolates of N. ganarrhaeae and N. meningitidis
examined thus far .(37). While Lip is immunogenic in patients with
gonorrhea and meningococcal infections (28), there is conflicting
evidence as to whether Lip is exposed on the surface of viable gonococci
(36). In humans, an anti-Lip response is considered to be nonprotective
upon challenge as indicated by patients having sequential gonococcal
infections and detectable anti-Lip titers (28), but there is also some
evidence that a monoclonal antibody directed against the H.8 epitope of
Lip can be bactericidal and opsonic in vitro (36). Therefore, the
importance of Lip as a virulence function is still unresolved. The
conservation of Lip among the pathogenic, and its exclusion from the
commensal Neisseriae suggests a role in virulence.

Lipopolysaccharide. Also known as endotoxin, and by some authors as
lipooligosaccharide (LOS; 94), LPS is an important component of the

gonococcal cell surface. Gonococcal LPS is chemically similar to the

20
rough LPS described in Salmonella, in that it is composed of a lipid A
moiety covalently bound to a core oligosaccharide, but lacking the
repetitive polysaccharide structures known as O-antigen (242,258). The
LPS of N. ganarrhoeae exhibits a low apparent molecular weight of 3.2-
7.1 kDa by SDS-PAGE (84,242).

In N. ganarrhaeae, LPS shows intrastrain antigenic variability, both
in vitra and in viva (155) as well as showing some antigenic similarity
to human erythrocyte surface antigens (153). LPS is highly immunogenic
in viva, and patient sera from both DGI and UGI cases show reactivity
(152). Some types of LPS expressed by N. ganarrhaeae have been
associated with stable serum resistance (227) and LPS is the target
molecule for modification in unstable serum resistance (204). LPS has
also been demonstrated as being a major cause of the cytotoxic effects
observed in the FTOC model system (83). Due to the cytotoxicity of
gonococcal LPS and its roles in serum resistance, LPS is considered to
be a very important virulence factor for N. ganarrhaeae. Serum
resistance will be discussed in further detail in a separate section.

Peptidoglycen. The major component of the gonococcal cell wall,
peptidoglycan, has been reported to possess a variety of biological
properties. Structurally, it has been reported to be associated with
the Por protein, possibly to stabilize the outer membrane (106). It is
immunogenic in vivo and reportedly activates the classical complement
pathway (208). It is also a potent cytotoxin in FTOC analyses (167).
Degradative fragments of peptidoglycan have demonstrated arthrogenic
potential in an in viva rat model system (68) and induction of slow-wave
sleep in a rabbit model system (238). These degradation fragments may

be liberated in vivo and may be partially responsible for the strong

21
inflammatory response associated with gonorrhea (237).

Iron acquisition. The ability of the host to withhold iron from
invading microbes is an important nonspecific bacteriostatic mechanism.
Therefore, the ability of a microbe to acquire iron from its host is an
important virulence factor as nearly all of the Fe3+ in the human body
is rendered unavailable for microbial assimilation due to host produced
iron chelators and/or intracellular location (287). In the vascular
system, the majority of extracellular iron is bound by transferrin, a
glycoprotein functioning in iron transport, while in the external
secretions, the glycoprotein lactoferrin is the major host iron binding
factor (165,287). Smaller amounts of extracellular iron in the
circulatory system are bound by hemopexin and haptoglobin, which bind
hemin and hemoglobin, respectively. Intracellular iron in the host is
stored as hemoglobin and ferritin, in erythrocytes and all other human
cells, respectively (287). Some of the functions of iron in both the
host and the invading microbe include serving: l)as a cofactor in
metalloenzymes involved in redox potential, 2) the electron transport
chain, 3) DNA synthesis, and 4) tRNA modification (138).

Microbes have evolved with two basic mechanisms of acquiring
chelated irmn. The first mechanism to be described in detail was the
microbe-produced, siderophore-mediated iron binding and uptake system
(197). The second mechanism, proposed and described more recently,
involves the direct binding of specific host iron binding molecules to a
microbial receptor and assimilation of the iron. without a soluble
microbial iron chelating intermediary (205,206). Typically, the first
mechanism is utilized by microbes which spend at least a portion of

their existences outside of a host and must compete for iron with other

22
free-living microbes, while the second mechanism seems to be utilized by
more specialized microbes which live dependent upon a suitable host
species. There are examples of bacteria which utilize both of these
mechanisms (eg. Shigella spp.; 205).

N. ganarrhaeae has been a controversial species in the debate over
siderophore production. Initially, it was reported that the gonococcus
produced its own siderophore, termed gonobactin, which stimulated
gonococcal growth in iron-poor media (293). More careful studies
indicated that the source of "gonobactin" was the medium itself, and
that while it was capable of stimulating growth, the siderophore was not
produced by N. ganorrhaeae (202,289).

Most strains of N. ganarrhaeae are able to assimilate host iron
bound by human transferrin (173), human lactoferrin (172), hemoglobin
and hemin (173), as well as the siderophore, aerobactin (289). Some
gonococci are also able to utilize the iron from a hemoglobin-
haptoglobin, complex but are unable to acquire iron from a ‘hemin-
hemopexin or a heme-albumin complex (64). The ability of N. ganarrhoeae
to use these iron sources makes sense when it is considered that the
gonococcus is strictly a human pathogen and not found free-living at any
time. The mechanism by which the gonococcus acquires siderophore-bound
iron is undefined at present. Probably this ability is exploited by
the microbe as an evolutionary vestige which gives it a competitive edge
in the mixed cultures that exist in viva. What is known about the
mechanisms involved in iron acquisition from host iron binding proteins
will be discussed in the following section.

Iron regulated proteins. There are a number of environmental stress-

induced outer membrane proteins that have been suggested to be involved

23
in the virulence of N. ganarrhaeae. With regard to iron-stress, at
least 13 different outer membrane Frps (Ferric Regulated Proteins; 94)
are inducible under iron-limited conditions (110,174,200,289). West and
Sparling have reported that aside from the 37 kDa protein, the number of
Frps produced by a given strain may vary depending upon the method used
to render the media iron-limited (289).

Of this set of 13 proteins, the only one with a described function
is the 37 kDa (Fbp, formerly Mirp; 94,174) which has been demonstrated
to bind Fe3+ in a specific, reversible manner (9). There are also
reported to be outer membrane receptors specific for both lactoferrin
and transferrin (140,164). One laboratory has proposed transferrin and
lactoferrin binding proteins of 100 and 102 kDa, respectively (49),
while another laboratory has proposed molecular weights of 37 kDa and 77
kDa for the gonococcal transferrin and lactoferrin receptors,
respectively (281). In N. meningitidis, a closely related species,
different apparent molecular weights have been reported for the
transferrin and lactoferrin receptors, 71 kDa and 105 kDa, respectively
(244,245). At present, the true identities of these proteins or
receptors remains to be resolved. At least some of these iron-
repressible outer membrane proteins are expressed in viva, as
convalescent patient serum shows immunoreactivity with them, including
Fbp (69).

Oxygen stress. The abilities to respond to both high oxygen tension
and/or anaerobic stress are also considered to be virulence functions
for bacteria. While N. ganarrhaeae was initially considered to be an
obligate aerobe, able to survive but not grow under anaerobic conditions

(112), it was later demonstrated that the gonococcus is capable of

24
anaerobic growth, if sufficient nitrite is available for its use as a
terminal electron acceptor (126). Further, in continuous culture
experiments, it has been reported that if N. ganorrhaeae is grown under
a lowered oxygen tension, the culture is enriched for the more virulent,
piliated organisms (109,151).

It has also been shown that N. ganarrhaeae is able to stimulate
increased consumption of molecular oxygen in the presence of lactate
(26). This ability is presumed to be of significant importance to N.
ganarrhaeae during its existence in the oxygen radical-rich environment
of pus since it has been reported that some gonococcal strains do not
produce superoxide dismutase activity (201). Gonococci constit-utively
produce both a catalase and a peroxidase which would also assist in
survival in pus (31).

Oxis and Anis. When N. ganarrhaeae is propagated under anaerobic
conditions, a set of oxygen-repressible outer membrane proteins is
expressed (Anis), and a set of oxygen-inducible outer membrane proteins
is repressed (Oxis; 44). The names Anis and Oxis supercede the previous
designations of Pans and Poxs, respectively (94). At least some of the
Anis are expressed in viva as patient sera from both uncomplicated
infections and from pelvic inflammatory disease show reactivity with
Anil (45). Although there are no defined functions for any of the Ani
or Oxi proteins they are proposed to be involved in virulence due to the
observations that 1) there is an antibody response raised against at
least one of them, and 2) they are regulated by relevant host
encountered conditions. The expression of other well defined outer

membrane antigens is not affected by growth under anaerobic conditions

(41+) .

25

Heat shock proteins. A third class of environmental stress inducible
proteins are the heat shock proteins produced by N. ganarrhaeae (125).
Little is known about the identities or function(s) of these proteins in
N. gonorrhaeae. However, since the heat shock response has been
suggested to play a role in recovery of microbes from stresses other
than heat, the induction of heat shock proteins may play an important
role in keeping the gonococcus viable under the rapidly changing,
hostile conditions in the host.

Blebs. N. ganarrhaeae has been shown to shed membraneaus vesicles in
vitro when grown on agar media or in liquid media (61,269). Blebs of
two different densities have been isolated and shown to originate from
either the outer or the cyoplasmic membrane (60). The denser bleb
fraction, BII, has been shown to contain the following outer membrane
antigens by silver stain and immunoanalyses: Por, Rmp, LPS, and Lip
while the less dense, BI fraction lacked all of these but Par (60,269).

Chromosomal and plasmid DNAs, as well as RNA, were observed to be
associated with both fractions BI and B11 and further, the 811 fraction
was demonstrated to contain plasmid DNA in a DNAase resistant state
(60). The BII blebs have been implicated in transfer of plasmids
between a donor and a recipient strain and are proposed to play a role
in intergonoccal genetic exchange (60).

Serum resistance. The ability to resist complement-mediated killing
in normal human sera is an important virulence characteristic of some
strains of N. ganarrhaeae. There are several means by which this can
occur: 1) the absence of bactericidal antibodies against surface exposed
antigenic determinants; 2) antigenic variation of surface exposed

molecules; 3) the presence of blocking antibodies directed against

26

surface exposed antigens which inhibit complement fixation at nearby
antibody-reactive, surface exposed determinants; and 4) modification of
surface exposed determinants rendering them non—immunogenic. N.
gonorrhoeae is able to escape complement mediated killing in normal
human serum by all of these mechanisms (227). However, serum resistant
strains are not entirely resistant to complement-mediated killing as
they are efficiently killed by immune sera from convalescing DGI
patients (102).

Most research on serum resistance/sensitivity has centered around
killing by the classical pathway as it has been demonstrated that this
pathway is more effective in killing the gonococcus in vitro (95). The
degree of serum sensitivity of a given gonococcal isolate has been shown
to be related to the contribution of the alternate pathway in formation
of the C3 convertase (98) and was related to the molecular form of LPS
produced (227). The alternate pathway' has been shown. to kill N.
ganarrhaeae with the same rate of efficiency as the classical pathway,
but the alternate pathway has a slower initiation period (56).

Serum resistance in N. ganarrhoeae occurs in both unstable and
stable forms. Stability is determined by the effects of in vitra
subculture and resultant serum resistant phenotype of these passaged
bacteria. ‘Unstable serum resistance is exemplified by recovery of a
serum resistant isolate from a host, which upon subculture, becomes
serum sensitive. This phenotype was first described by Ward, et al. who
demonstrated that a freshly isolated, serum resistant strain was not
killed by serum from other infected patients or immune rabbit sera, but
soon became sensitive to normal human serum upon subculture (284).

Unstable serum resistance can also be acquired in vitra by

27

propagation in a medium which contains ultrafiltrates from guinea pig
sera (230), human sera (157), or human secretions (158). This shift to
serum resistance involves modification of the LPS structure, evidenced
by: l) a shift in the LPS migration on SDS-PAGE (276), 2) altered
sensitivity to pyocins (290), and 3) a lack of immunoreactivity with
previously reactive antisera (52,276). The molecular alteration was
later shown to be due to sialidation of the LPS by a low molecular
weight component in the serum in secretions (204) and has also been
demonstrated to occur in viva, catalyzed by a gonococcal sialyl
transferase (154).

A recent paper described the binding of vitronectin (aka S protein,
a host serum glycoprotein) to the surface of N. ganarrhaeae (5). A
serum function of vitronectin is the inhibition of fluid phase, membrane
attack complex formation by preventing C9 polymerization (217). This
may be an important mechanism conferring unstable serum resistance to
gonococci, as well as a potential aid in invasion of host cells because
vitronectin has specific receptors on the surfaces of endothelial cells
(239). This report may lead to a better understanding of some of the
microbial processes that occur in establishment of a DGI.

Stable serum resistance is often the result of expression of
particular forms of LPS and Lip by the gonococcus (209). Expression of
the PorA serotype (92,97) and lack of Opa expression (97) have
previously been reported to be related to serum resistance. However, a
subsequent report indicated that there was no association between the
Por subclass expressed and a given strain’s serum resistance (209; this
study did not address the contribution of Opas).

In particular, the expression of the 2-1-L8 epitope (as defined by a

28

monoclonal antibody) on the LPS molecule, seems to correlate with stable
serum resistance (209,241). This epitope is rarely immunogenic in
humans, possibly due to inaccessibility in situ, helping to impart serum
resistance to its bearer (227). However, if a host does respond to the
2-l-L8 epitope, the antiserum is capable of killing serum resistant
strains (227). Exposure of the 2-l-L8 epitope has been shown to
increase as the length of the oligosaccharide chains decrease (62).

Another form of stable serum resistance, with regard to a given
specimen of normal serum, is the presence of blocking antibodies which
prevent complement fixation at other nearby determinants (227). An
example of this has been discussed in the previous section on Rmp.
Although the mechanism by which blocking antibodies protect the
bacterium is unknown, they may prevent deposition of complement at
adjacent sites, or result in complement deposition at a site where the
membrane attack complex is non-functional (101).

Invasion of host cells. The ability of a pathogen to enter a host
cell and isolate itself from the humoral immune system is often critical
to the survival of that pathogen and establishment of an infection. N.
gonarrhaeae has frequently been observed within the nonciliated columnar
cells of the urethral, cervical and fallopian tube epithelial mucosae
(260), but it was not until more recently that the events of invasion
have been studied and described.

The FTOC system demonstrated that gonococci were able to penetrate
the epithelial barrier by some undescribed mechanism (100). It was
shown that the bacteria bound to the nonciliated host cell's apical
surface and entered the cytoplasm in a membrane bound-vacuole (100).

These envacuolated gonococci were transported through the host cell's

29
cytoplasm and released in an unbound state via an orderly parting of the
host cell's basal membrane (260). Also, blebs are internalized by these
tissues and may be one mechanism for the cytoxicity of LPS and
peptidoglycan (48).

More recently, a human epithelium cell culture model (HEC-l-B
monolayers) has been described which allows the penetration of gonococci
into the host cell's cytoplasm (251). Cytosolic bacteria were observed
to be in an unbound state and multiplying until they lysed the host cell
(251).

Bacterial entry of these monolayers was suggested to occur in a
manner similar to phagocytosis by professional phagocytes, excepting for
the lack of bacterial destruction (260). Invasion requires viable
gonococci and involves microtubule function (229), suggesting that this
is an example of parasite-directed endocytosis. Par is involved in
entry (91), the role(s) of Opas is still unclear (53), and pili are not
required for invasion (260).

Antibiotic resistance. The ability to survive antibiotic therapy and
persist in a host is considered to be an important virulence factor. It
should be noted that wild type N. ganarrhoeae is resistant to low levels
of vancomycin, and this resistance has been exploited in selective media
developed for isolation of the gonococcus from clinical specimens (180).
Thus far, N. gonarrhaeae with several different antibiotic resistances
have been isolated either from patients or in the laboratory, and these
resistances are summarized in Table 1.

Antibiotic resistance phenotypes have been observed to be encoded on
the chromosome, and in some cases on plasmids (also noted in Table 1).

For the most part, resistances provided by chromosomal-borne mutations

30

Table l. Antibiotic resistances observed in N. ganarrhaeae isolates.

 

 

Antibiotic Lgcus Locationa Reference
ampicillin ampA,ampBb C 104
" ampC,ang C 104
chloramphenicol cap C 257
erythromycin ery C 149
fusidic acid fus C 256
nalidixic acid nald C 288
penicillin pem C 285
" penA C 257
" penB C 256,285
rifampin rif C 256
spectinomycin spc C 149
streptomycin str C 255
tetracycline tete C 256
" tem C 256,285
several mtrf C 149,257
" nsz‘g C 35
p-lactams bla P 232
tetracycline tetM P 182
a - C, chromosomal and P, plasmid-borne.
b - Stepwise additional resistance to ampA and ampB.
c - Previously called chl.
d - Additional resistance in combination with penA and mtr.
e - Additional resistance in combination with rat and mtr.
f - Results in penicillin, tetracycline, erythromycin, chlor-

amphenicol, rifampin, and fusidic acid resistances.
Results in similar resistances to mtr but has been mapped to
a different locus.

0:)
I

31

are low level resistance to a specific drug, with those provided by
plasmids being relatively higher. Nonspecific mutations have also been
characterized which give a broad range of low level antibiotic resist-
ances (eg. .mtr’ and. nsr), exerting their effects by' decreasing the
permeability of the gonococcal cell wall to various antibiotics, dyes,
and detergents (149,257). Some of these mutations are additive,
conferring increased resistance to a. drug as compared to a single
mutation (eg. penA, penB, mtr, and tem for penicillin; 256). Other
mutations must occur simultaneously to express the resistant phenotype
(eg. ampA and ampB; 104). All of the chromosomal antibiotic resistances
are heritable and can be transformed from a resistant donor strain into
a sensitive recipient strain, conferring resistance. Similarly, the
plasmid borne resistances can be mobilized by either the 24.5 MDa or
25.2 MDa conjugative plasmids to a sensitive recipient strain where they
confer antibiotic resistance, as discussed in the previous sections on
genetics and genetic exchange.

Proteases. Production of a few proteolytic enzymes has been
described for N. ganarrhaeae and a review article proposed some roles in
virulence for these enzymes (203). Most of the proteases are
nonspecific, and their relevant substrates in pathogenesis have not been
proposed or defined. Examples of nonspecific endopeptidases include: 1)
a trypsin-like enzyme that hydyrolyzes gelatin, azoalbumin, and p-tosyl-
L-arginine methyl ester HCl (TAME; 265), and 2) gonocosin, characterized
as being a neutral metalloprotease (pH optimum of 7.5-8.0, inhibited by
EDTA, and resistant to inhibitors of serine proteases and sulfhydryl
inhibitors; 43). Gonocosin is reported to cleave succinylated elastin

but was not tested against gelatin, azoalbumin, or TAME (43) so

32
comparisons to the trypsin-like enzyme can not be made. The ability to
cleave elastin may be an indication of gonocosin's potential role in
virulence, as elastin is a host protein component of connective tissues.

N. ganarrhoeae has also been described as producing two
exopeptidases, an aminopeptidase and a proline iminopeptidase (43). The
aminopeptidase was inhibited by EDTA; had an optimum pH of 7.5—8.0; and
was described as removing aminoterminal residues adjacent to proline, an
activity similar to the aminopeptidase-P of E. cali (43).
Aminopeptidase activities have been described in some of the
nonpathogenic neisseriae as well (286), so the contribution of these
enzymes to virulence is questionable.

The proline iminopeptidase is sensitive to both heavy metal and
iodoacetamide inactivation, and is activated by dithiothreitol,
suggesting that it is a thiol-activated protease (43). Its functional
activity involves the removal of aminoterminal proline residues. This
enzyme is considered to be similar to a proline iminopeptidase produced
by E. cali (43). It is more likely involved in the acquisition of amino
acids than in pathogenic processes per se.

IgAl proteases. The secreted fluids on mucosal surfaces have
secretory IgA (sIgA) as their ‘primary immunoglobulixi with. a lesser
amount of IgG and IgM (165,277). This pool of sIgA is approximately
equally comprised of both IgAl and IgA2 (54). This is in contrast with
serum, ‘where the predominant form. of IgA. is monomeric, and. ~80-9O
percent of serum IgA is of the IgAl subclass (54).

The sIgA antibodies are secreted by plasma cells in the lamina
propria beneath the mucosal epithelium, in the form of dimeric IgA, two

idiotypically identical IgA monomers (Mr of ~160,000) linked non-

33

covalently via their CH2 domains through the association of J chain (Mr
of 15,000; 2). This IgA dimer diffuses through the interstitial fluids,
and is bound by secretory component (SC; a glycoprotein of ~80,000), the
IgA receptor on the basolateral side of the epithelial cells (2,169).
The IgA dimer-SC complex is internalized by the epithelial cell and
extruded through the apical face of the cell into the lumen as an sIgA
molecule (2). This receptor-ligand interaction is unique in that the
receptor-ligand. complex is not dissociated intracellularly, and. the
receptor is not recycled for further use.

The IgAl proteases are the best studied of the proteolytic enzymes
produced by N. gonarrhaeae and are considered to be a virulence factor.
Nonpathogenic Neisseria spp. do not produce the enzyme (193). These
enzymes are classified as extracellular, metal-requiring endopeptidases,
which cleave human immunoglobulins of the Al subclass yielding
monovalent Faba and intact Fca fragments, while not cleaving those of
the A2 subclass (187,213). One of the characteristics held in common
among the set of bacteria which produce IgAl proteases is that they are
all considered to be human pathogens which reside on the mucosal
surfaces.

It has been demonstrated that a given species of bacteria can
produce one of a few different types of IgAl protease differentiated by
the site of IgAl cleavage. These different activities can be identified
by the migration of the IgAl cleavage products (Fda and Fca) visible
after SDS-PAGE (194). N. ganarrhaeae has been shown to produce two
types of cleavage site specificity, with a pure isolate producing only
one specificity (188). The type 1 protease cleaves the PS peptide bond

between residues 237-238 and the type 2 protease cleaves the PT peptide

34
bond between residues 235-236 (190,213). Figure l diagrams a portion of
the IgAl hinge region and shows the cleavage sites defined for the two
N. ganorrhaeae IgAl proteases. There has been no description of a
gonococcal isolate either switching specificity types or ceasing to
produce these enzymes upon repetitive subculture, and production has

been demonstrated by all four of the Kellogg colony types (213).

225 2?0 235~ 2:0
STPPTPSPSTPPTiSPSC

type 1
Type 2

Figure 1. The IgAl hinge region duplicated amino acid octamer showing

the N. gonorrhoeae types 1 and 2 IgAl protease cleavage sites (190,213).

It is intriguing that N. ganorrhoeae produces two different IgAl
proteases which cleave IgAl in the sequence 23(‘PI’TPSPSZI‘0. In either
situation, the products contain an amino-terminal T or S residue which
could be removed by the aminopeptidase to expose the adjacent proline,
which could be removed in turn by the proline iminopeptidase. In the

case of the type 2 IgAl protease, a second pair of amino acids could be

35

removed sequentially. A study of N. gonarrhaeae defined only 15.5% of
type 1 IgAl protease producers as proline-requiring auxotypes while
46.2% of the type 2 IgAl protease producers required proline (190; both
of these percentage groups include strains which require nutrients in
addition to proline). These data suggest a selective advantage for
proline-requiring gonococcal strains which also produce a type 2 IgAl
protease.

The type 1 IgAl protease has been shown to be associated with the
expression of only a few Por serovars, namely PorA-1 and PorA-2, with
very few examples of PorB-l and PorB-2 included (190). Additionally,
these type 1 IgAl protease producers were clustered into the arginine-
hypoxanthine -uracil - and prol ine - arginine -hypoxanthine - urac il - requiring
auxotypes with a few examples of proline auxotrophs (190). In contrast,
type 2 IgAl protease producers are found among all of the other 45
serovars tested except the PorA-2 group, and are of auxotypes other than
the arginine~hypoxanthine-uracil- and proline-arginine-hypoxanthine-
uracil-requiring groups (190).

As stated above, the IgAl subclass is susceptible to cleavage while
the IgA2 subclass is not. The molecular basis for this differentiation
is a 13 amino acid deletion in the hinge region of IgA2, between the CH1
and CH2 domains (144,170,279). The hinge region of IgAl contains a
duplicated amino acid octamer between residues 225 and 240
(NTPPTPSPS/TPPTPSPSC), where all but two of the defined cleavage sites
for IgAl proteases occur (71,184,187). Ironically, it is the unusual
repetitive structure of the hinge region that makes IgAl more resistant

to other commonly used proteinases (168). Also, the presence of

secretory component has been reported to provide sIgA with increased

 

36
resistance to proteolysis (165), but this is not the case with regard to
IgAl protease hydrolysis as both monomeric IgAl and sIgAl are cleaved
efficiently (76,150,226).V

There are very few examples of other substrates cleaved by the IgAl
proteases. The type 2 enzyme is the better studied of the two
gonococcal IgAl protease types with regard to alternative substrate
cleavage. While a large number of potentially relevant serum proteins
are not cleaved by these enzymes, it has been shown that there are a
small number of other peptides or proteins which are permissive
substrates. 'These include a 'pair of' synthetic IgAl 'hinge peptide
analogues (291), the type 2 IgAl protease proenzyme (218), and a set of
bacterial membrane proteins with undefined identities (252, and chapter
2 here). These permissive and nonpermissive substrates are listed in
Tables 2 and 3, respectively.

The type 2 IgAl protease of N. ganarrhaeae is considered to be a
serine protease, on the basis of boronyl peptide inhibitor studies and
by comparison of the predicted amino acid sequence with other known
serine proteases (6). A comparative study of the two purified enzymes
indicated that they have different sizes, the type 2 enzyme being 114
Kda which converts to 109 kDa, and the type 1 enzyme being a 112 kDa
form that could not be separated from two smaller proteins (34 and 31
kDa) without a complete loss of activity (253). The same study also
suggested that the two enzymes were very different on the basis of
stability and response to chemical inhibitors (253). The type 1
protease of N. gonorrhaeae was extremely sensitive to the metal
chelators tested» phenylmethylsulfonyl fluoride (a serine 'proteinase

inhibitor), and p-chloromercuriphenylsulfonic acid (a thiol-specific

37

Table 2. Permissive substrates for the N. ganarrhaeae type 2 IgAl

protease.

 

Substrate
IgAl

IgAl protease proenzyme

Cholera toxinzlgAl protease
fusion proteins

Synthetic peptides

Bacterial membrane proteins

W8

PP-TP

AP-SP
PP-SP
PP-AP

PP-AP
PP-TP

PP-SP
PP-AP

NDb

Refegence
213

336

124

291

252

 

a - Adjoining amino acids given in the one letter code.

b - ND, not determined.

38

Table 3. Molecules that are nonpermissive substrates for IgAl proteases.

 

Meleeele Seurcea Referenee
Albumin cow,egg,human 119,191
Azocasein 235
Azocoll 187
fl-microglobulin human 187
blood group substances A,B human 187
casein 213
collagen 213
gelatin 213
glycophorin 187
hemoglobin cow,human 213
IgA cow,dog,monkeys,mouse 107,187
orangutang,pig,rat,rabbit
IgA2 human 213
IgD 187
lgE human 213
IgG human,mouse,rabbit 187
IgM human 213
Insulin (5 chain) cow 213
J chain human 117
Light chain (n,A) human 117
Mucin human 187
Ovomucoid egg 187
Secretory component human 117
Synthetic IgAl hinge peptides 6,34,291
Transferrin 187

 

a - sources not listed in the table were not defined in the
reference.

39

inhibitor), while the type 2 protease was a minimum of ten times as
resistant to these same compounds (253). Kilian, et al. also reported
that the gonococcal IgAl proteases were not inhibited by three different
serine protease inhibitors (123). These latter sets of chemical data
pertaining to serine protease inhibitors are at odds with that presented
by Bachovchin, et a1. (6), but the specific inhibitors used. were
different so the results can not be compared critically.

Data indicating that the enzymes are produced in viva came from the
work of Blake, et a1., who demonstrated that active enzyme could be
recovered in vaginal wash fluids of patients infected with gonorrhea
(17). Another study indicated that some patients infected with
gonorrhea also respond with a serum antibody titer against IgAl
protease, detectable by immunoblot analysis (27,28). However,
additional data presented by the same group indicated that there was no
mucosal immune response to IgAl protease detectable in genital
secretions (27,28). This is in contrast to data involving persons
colonized or infected with Neisseria meningitidis, who responded to the
meningococcal IgAl protease with serum titers that inhibit the activity
of both meningococcal and gonococcal IgAl proteases (28).

This antibody inhibition data is in agreement with other data
comparing the antigenic relatedness of IgAl proteases from several
species. These results showed. that a gonococcal IgAl protease is
immunologically similar (1. e. is inhibited by heterologous antisera) to
an IgAl protease produced by one of two strains of N. meningitidis
(123).

These proteases can be inhibited, seemingly paradoxically, by both

human secretory IgA and human sera. Two reports by the same group have

4O

determined that serum from four of six patients acutely infected with
gonorrhea, inhibited the activity of a type 2 IgAl protease preparation
(76,77). The inhibitor was suggested. to be IgG on the basis of
chromatographic separations of the inhibitory sera (76,77). In the case
of inhibition by sIgA, it was determined that the inhibitory action was
via the Fab portion of sIgA, presumably functioning as an antibody with
idiotypic specificity for the IgAl protease molecule (77). These data,
while disparate with that of Brooks and Lammel (27,28), may be explained
by either: 1) these four patients had truly responded to the gonococcal
enzyme, or 2) these sera contained antibodies that were directed against
the meningococcal IgAl protease and happened to inhibit the gonococcal
enzyme.

The genes encoding both types of IgAl protease have been cloned in
Escherichia coli and characterized by several independent laboratories
(67,87,88,131,189,l96), and are discussed below in chronological order.
The first gene encoding the type 2 enzyme (iga2) from strain F62 was
cloned by Koomey, et a1., who demonstrated that it produced an ~140 kDa
extracellular enzyme in E. cali, and that the enzyme activity produced
in E. coli was only ~0.5-2.0 percent of that produced by the parent
gonococcal strain (131). The cloned iga2 gene was mutated in vitro and
transformed back into the iga2 locus of F62, with no apparent phenotypic
effects other than the inability to hydrolyze IgAl (131). Phenotypes
that were reported to be unaffected included: 1) the stained, total
protein profiles as analyzed by SDS-PAGE, 2) growth rate and viability
on complex solid media, and 3) colonial morphotype conversions (i. e.
from piliated to nonpiliated, and from transparent to opaque, and vice

versa; 131).

41

A subsequent report by Halter, et al. characterized the iga2 gene
from strain M811 as having a single copy in the genome, and being
expressed at comparable levels extracellularly, in both E. cali and the
gonococcus (87). They also reported that there were differences in the
restriction site maps of the F62 and MSll clones, the first published
observation of restriction site polymorphism among the IgAl protease
genes. This was also the first reported purification of the type 2
enzyme from 'both N. gonorrhoeae and. E. cali, indicating that the
gonococcal form of the protease was a 106/105 kDa doublet while that
produced in E. call was a single band of 105 kDa (87).

Mulks and Knapp reported the cloning of an iga2 gene from the strain
GCM 740, and confirmed that there was a single copy of the gene in the
genome (189). They also compared genomic restriction site maps of the
three strains from which the iga2 gene had been cloned with those of
known type 1 enzyme producers (igal). Their data indicated that there
is extensive restriction site polymorphism among the iga2 loci, and that
all three iga2 loci are different from the igal loci (189). The igal
loci however, showed no restriction site heterogeneity among the set
analyzed (189). This hypothesis was later expanded by Mulks, et al. to
include eight different genotypic restriction patterns for strains other
than type 1 producers, and a single genotypic pattern for all of the
type 1 producers (196).

The first report of a cloned igal gene from N. ganorrhaeae (strain
32819) came in 1985 when Fishman, et al. reported a recombinant phage
and plasmids that encoded the production of an extracellular, type 1
IgAl protease in E. cali (67). They also determined that this igal

clone from N. gonorrhaeae was not able to complement in trans, a lesion

42
in a weakly expressed iga clone from Haemaphilus influenzae (24) by
cloning the two genes into a single plasmid vector. However, their igal
plasmid subclones were toxic and unstable in E. coli and this is the
sole report from the group regarding the cloning of the type 1 IgAl
protease.

In 1987, Pohlner, et al. determined the complete nucleotide sequence
of the iga2 gene from M811 and presented a model for the secretion of
the type 2 IgAl protease in N. ganarrhaeae and E. cali (218). The gene
was characterized as 4,596 base pairs in length, encoding a 1,532 amino
acid prepraenzyme (169 kDa) with four domains: 1) a 3 kDa leader
peptide, 2) the 106 kDa structural IgAl protease, 3) a 12-15 kDa a-
domain with no defined function, and 4) a 45 kDa fi-domain necessary for
secretion of the protease (218). The leader domain is presumably
removed ‘by the cell's leader endopeptidase during the cytoplasmic
membrane translocation of the proenzyme. In the periplasmic space the
molecule folds into a conformation which allows the fl-domain to interact
with the outer membrane and form a pore structure through which the
proenzyme is extruded. On the extracellular surface, an autoproteolytic
event occurs to release the Iga-a intermediate, processed via further
autoproteolytic events to yield the mature IgAl protease and the a-
peptide(s) (218). There are two potential cleavage sites at the Iga-a
junction (a and b) and one at the a-fl junction (c) which are clipped
during these processing events, none of which are identical to the
sequence cleaved in IgAl (Figures 1 and 2; 213,218). Selection of the a
or b site is apparently nonpreferential as both sites are used. Figure
2 diagrams the events involved in secretion of the IgAl protease as

described by Pohlner, et al. (218).

43

Figure 2. Diagramatic representation of the N. gonorrhoeae type 2 IgAl
protease secretion process. Restriction endonuclease sites are
designated by the following abbreviations: H, HindIII; Hp, HpaI; and P,
PstI. Designated domains and features of the preproenzyme and proenzyme
are in the convention of Pohlner, et al. (218; L, leader peptide; Iga,
structural IgAl protease domain; a and fl domains, respectively; a, b, c,
the three potential internal proteolytic processing sites; and Cys, the

two cysteine residues proposed to be in the active site).

 

 

 

 

44

 
  

 

 

 
 

 

 

 

 

Figure 2
106 kDa W + D 3 kDa
Iga °‘
Autoproteolgsts b
° it] 15 kDa
109 kDa W + or
[go D 12 kDa
O!
Autopmteolgsls
ab
12: kDa W:
Iga 0‘
Autoproreolgsls {l B
Extracellular
0M pore formation by ,8
“I.” f
W
me kDa 19° l o, ,9
Pertplasmlc
space
CM
3 Idle 13-
\ Cytoplasm
Signal peptldase
”‘3‘" “.P f c
169 kDa 9' ‘9' “"29"”
L Iga ex ,8
T Translation
. t
H Hp Hp HpHp P P P H

Iga2 gene

45

The B-domain contains all of the requisite functions for export as
demonstrated by the surface expression of a.txenslational fusion gene
(cholera toxin B subunit-fl domain) in Salmonella typhimurium (124). The
exposed fusion protein was susceptible to trypsin and could be released
from the cell surface by treatment with exogenous IgAl protease (124).

The restriction site polymorphism question was later addressed by
Halter, et al., who reported. that there were several clusters of
nucleotide substitutions throughout the gene, in the structural protease
domain and in the a-fi junction region (88). They also reported the
first stable cloning of a gonococcal igal gene (88). They proposed a
scheme for classifying iga genes based on the presence or absence of a
83111 or PstI restriction site near the a-fl junction, which influences
the c autoproteolysis site hydrolyzed during secretion (88). Strains
exhibiting a 83111 site process the Iga proenzyme to yield a 33 kDa fi-
peptide (H2 strains) while strains with a PstI site process the Iga
proenzyme with a 45 kDa fl-peptide (H1 strains; 88). This differential
processing is diagrammed in Figure 3 as described by Halter, et a1.
(88). They also reported the presence of the gonococcal uptake sequence
(79) at the downstream end of four sequenced iga genes, and proposed
that the nucleotide sequences of iga genes had diverged over time due to
interstrain exchanges via transformation (88).

The IgAl protease genes of N. ganarrhaeae show significant DNA
homology with those of N. meningitidis, as reported by Koomey and Falkow
(>80%; 129). When these iga2 genes were compared to the IgAl protease
gene of H. influenzae, lower homology was observed (67-80%), and no
homology was detected between the cloned iga2 gene and the genome of

seven species of Neisseria and Branhamella catarrhalis, all considered

46

Figure 3. Diagramatic representation of the proteins produced during the
processing of N. ganorrhoeae IgAl proteases as proposed by Halter, et
al. (88). Restriction endonuclease sites are designated by the
following abbreviations: H, HindIII; Hp, Hpal; and P, PstI. The domains
and features of the preproenzyme and the proenzyme are given in the
convention of Pohlner, et al. (218; L, leader peptide; Iga, structural
IgAl protease domain; a and fl domains, respectively; a, b, c, the three
potential internal proteolytic processing sites; and Cys, the two

cysteine residues proposed to be in the active site).

 

 

47

Shite

2939...

m 050E

48
to be commensals (129) and previously shown not to produce IgAl
proteases (193).

The effects of IgAl protease on the immune system have been studied
and data from a variety of investigations suggests that IgAl protease
may play an important role in the virulence of the gonococcus. For
example, gonococcal IgAl protease activity causes deagglutination of
particles held in an IgAl bound-immune complex and lowers the
agglutinating titer of antibody preparations that include IgAl as an
agglutinin (215). These data indicate that IgAl is susceptible to IgAl
protease cleavage when free in solution or when bound to antigen, and
implies that the antibody activity of the IgAl molecules is destroyed.

Further data indicates that treatment of an sIgA preparation with
exogenous gonococcal IgAl protease reduces the ability of that sIgA
preparation to inhibit the adherence of gonococci to human cells (195).
These experiments were performed by incubating the sIgA preparation with
exogenous IgAl protease prior to exposing the sIgA preparation to the
target gonococcal cells, leaving the possibility open that if the sIgA
preparation were reacted with the target gonococci first, the results
may have been different. If the data presented by Plaut, et al. (215)
can be extrapolated, then the IgAl antibodies bound to the gonococci
would still have been susceptible to hydrolysis, and a decrease in the
inhibition of gonococcal adherence would have been observed.

One of the functions that has been proposed for IgA is a dampening
role in the inflammatory response (118). If degradation of IgAl
abrogates this dampening effect, there would be an increased
inflammatory response (118), similar to what is observed in cases of

gonococcal infection. Many of the clinical signs of gonorrhea are a

49
result of inflammation and perhaps efficient transmission of the
bacteria themselves is dependent upon a strong inflammatory response.
Also, there is some evidence that IgAl protease treatment of a
monomeric IgAl paraprotein causes a reduction in the ability of IgAl to
inhibit the movement of polymorphonuclear leukocytes, presumably an ECG-

mediated phenomenon (280). Other Fca -mediated functions proposed to be

1
inhibited by cleavage of IgAl include: 1) opsonization and phagocytosis
by macrophages and polymorphonuclear leukocytes, 2) monocyte-dependent
bactericidal activity, 3) antibody-dependent cellular cytotoxicity, and
4) sIgA enhancement of nonspecific bactericidal mechanisms (eg.
lactoferrin bactericidal activity) (118).

The monovalent Faba released after IgAl protease hydrolysis of IgAl
are reported to retain the ability to bind antigens (156). These Faba
fragments may obscure antigenic determinants susceptible to Fc-mediated
events involving other immunoglobulin classes (eg. complement activation
by IgG or IgM, and opsonophagocytosis by professional phagocytes), or by
coating the bacterium with host proteins, render the gonococcus less
immunogenic, also an advantage in establishing and. maintaining an
infection.

Pohlner, et al. have proposed that the CD.8 marker antigen on the
surface of cytotoxic T cells and a cytokine, granulocyte/macrophage
growth stimulating factor (GM-GSF) are potential substrates for the type
2 gonococcal IgAl protease (219). Cleavage of these two immunologically
relevant molecules could potentially alter the normal immune response
and change the outcome of an infection.

As a final point regarding the impact of the gonococcal type 2 IgAl

protease on virulence, Cooper, et al. compared a set of isogenically

50

constructed Iga+ and Iga- gonococcal strains (131) on the basis of their
abilities to adhere, cause cellular damage, and invade the human fallop-
ian tube organ culture model system. They observed that the Iga- mutant
strain is as virulent as the wild type parental strain for all three
parameters measured (47). The data suggested that IgAl protease has no
effect on the events that lead to initial mucosal colonization.
However, it must be kept in mind that the amount of sIgA secreted by the
tissue explant may not be equivalent to in situ production. These in
vitro experiments also lack the protective effects of mucus secretion
and flow, and can not stimulate inflammation which plays an important
role in the establishment of gonococcal infection.

In summary, the role(s) of IgAl protease in the virulence of N.
ganarrhaeae is unclear. There is some indirect evidence that it may
play a part in neutralizing the IgAl antibody-mediated effects of sIgA
and stimulation of the immune response. However, until the relative
virulence of isogenic Iga+ and Iga- strains can be examined in either a
human or chimpanzee experimental infection, the real contribution(s) of
these enzymes in virulence will remain speculative.

IgAl proteases of other species. A broader discussion of the IgAl
proteases will provide background information and serve to explain why
this set of enzymes has been of interest to so many researchers over the
years. There have been intense efforts from a variety of laboratories
and the information that has been generated merits some attention.

The first description of an enzyme which cleaved human IgA came in
1973 with the isolation of intact Fca fragments from the clarified feces
of six normal individuals (166). The isolation of the fragment was

noteworthy because up to that time, the relative proteinase resistance

51

of IgA had made it a frustrating molecule for structural studies.
Mehta, et al. demonstrated that the activity which generated the Foo
fragment was produced by an unidentified microorganism and observed in
the clarified supernatant from aerobic broth cultures of fecal material.
This activity was proposed to be metal-dependent based on its
sensitivity to 5 mM EDTA and was resistant to the effects of serine
protease inhibitors. Attempts to identify the specific microbe(s) that
produced the enzyme were unsuccessful. However, the authors noted that
an organism identified as an E. cali did make the enzyme transiently but
failed to produce it upon repeated subculture (166).

A subsequent report from the same group provided further information
about this enzyme activity but still no producing species was able to be
identified, which is the case today. Plaut, et al. reported that the
enzyme had a molecular weight of 40-60,000 based on its sedimentation in
sucrose density gradient ultracentrifugation and that it was inactivated
by heating at 55°C for thirty minutes (212). The enzyme was proposed to
be a metalloprotease due to its sensitivity to inhibition by EDTA and
the fact that activity could be restored by divalent cations (ZnH,
Co“, and MnH were better than CaH and MgH; 212). Due to the
inability to identify the source of this protease, efforts to find other
bacteria which produced the proteolytic activity intensified and many
IgAl protease producing bacteria have been identified.

IgAl protease methodologies. Many different methods of analyzing a
sample for the presence of IgAl protease have been developed through the
years. Some of these are in widespread use by the laboratories in the
field and others have predictably been passing fads. Table 4 lists some

of the methodologies developed and/or used in this research and gives a

52

Table 4. Techniques used to analyze the IgAl proteases.

 

s c e des tie a ve flag.
1. Those requiring complex formation.

A. Type 4 group A streptococci Fc receptors + 141
Uses radiolabeled IgAl as a sfibstrate
Protease activity -> lower number of cpm bound

B. Beckman Immunochemistry System nephelometry + 142
Assays degree of light scatter due to a-IgA:
IgAl complexes
Protease activity -> lower amount of scatter

C. Immunoelectrophoresis - 166
Uses a-Hc and a-Lc to detect cleavage products
after separation
IgAl protease activity -> altered precipitation
pattern

H-

D. Rocket immunoelectrophoresis 139
Uses a-Hc in the agar solid phase for
detection during separation
IgAl protease activity -> altered precipitation

with larger area in "cone of rocket"

E. ELISA + 223
Solid phase contains a-Lc:IgA complex
Detection with a-Hc after protease treatment
IgAl protease activity -> decreased binding of a-Hc

II. Those which physically separate IgAl cleavage products.

A. Cellulose acetate electrophoresis + 216
Uses either unlabeled or radiolabeled IgAl
Detect by staining or autoradiography, respectively
IgAl protease activity -> staining or radioactivity
in Fca and Faba bands

B. SDS-PAGE
1. Under reductive conditions + 19
Uses either unlabeled or radiolabeled IgAl
Detect by staining or autoradiography, respectively
IgAl protease activity -> staining or radioactivity
in Fca and Fda bands

2. Under native or nonreductive conditions - l
Detect by immunoblotting with a-Hc
IgAl protease activity -> reactivity in PC
and Faba bands a

53

Table 4 (cont'd).

 

e o e tio Quantitativea Ref,

C. Solid phase plate assays
l. Overlay bacterial colonies with immobilized -

radiolabeled IgAl in top agar

After incubation, overlay with a nitrocellulose
filter and allow for diffusion of Fab and

binding to the filter a

Detect by autoradiography

IgAl protease activity -> dark spot on X—ray film

2. Overlay colonies with radiolabeled sIgAl bound to -
nitrocellulose filter by a-secretory component
After incubation, detect by autoradiography
IgAl protease activity -> light spot on X-ray film

3. Overlay bacterial colonies with immobilized -
radiolabeled IgAl in top agar
After incubation, overlay with a nitrocellulose
filter having a-Lc to bind free Fab
Detect by autoradiography a
IgAl protease activity -> dark spot on X-ray film

D. Analytical ultracentrifugation -
Uses unlabeled IgAl or sIgAl
Detect by photography of separated bands
IgAl protease activity -> lower density cleavage
products (Fca and Faba)

E. High Pressure Liquid Chromatography +
Uses either unlabeled or radiolabeled IgAl
Separate on a gel filtration column
Detection by absorbance or determination of
radioactivity
IgAl protease activity -> lower molecular weight
cleavage products (Pea and Faba)

74

87

29

119

185

 

a - +, quantitative; i, semi-quantitative; -, qualitative only.

54
brief summary of the relevant points of each.

All have the following in common, the researcher must react a sample
containing human IgAl with the enzyme and analyze the results of that
reaction to identify the presence or absence of IgAl cleavage products.
The prototypic reaction was presented by Plaut, et al. and involves
mixing a culture supernatant, bacterial extract, whole bacteria, or some
processed version of these with a buffered reaction mix (pH range 7.0-
8.5) containing serum IgA, sIgA, purified IgAl, or purified sIgAl (214).
The reaction mix is incubated at 31-37°C until termination by addition
of EDTA, SDS and reducing agents, or heating.

The analysis procedure must entail some means of identifying the
cleavage fragments and the original IgAl molecule. Examples of
identification include immunoreactivity of' Hca and Faba or La and
physical separation based on size or charge of the cleaved fragments and
intact IgAl. For quantitative purposes, the IgAl protease unit has been
defined as that amount of activity cleaving 1 pg of human serum IgAl in
one minute at pH 7.0 and 37°C (214).

These techniques will be discussed in the order listed in Table 4.
With the exception of the first technique, the assays in set I are based
on the formation of an an immune complex using the IgAl molecule and its
cleavage fragments as antigens, and a specific antibody or anti-serum
preparation (eg. anti-Hca, a-Lc, or a-SC) as the antibody component. A
common disadvantage to most of these techniques is that their accuracy
is dependent upon the quality of the reagents used in detection.

The first assay exploits the class specific, Fca receptors expressed
on the surface of a particular subset of group A streptococci (type 4;

incidentalLy, an IgAl protease negative group; 141). This particular

55
assay technique is not widely used. The disadvantages of this procedure
include 1) the necessity of a laboratory to culture and work with the
virulent streptococci and 2) the necessity of a radioactive probe. The
advantage is that the assay is relatively rapid as compared to some of
the others.

The second assay described is not used by very many laboratories due
to the required instrumentation (142). The Beckman Immunochemistry
System is not a commonly held piece of laboratory equipment, except for
possibly in a hospital or other clinical laboratory. However, the assay
does have the advantages of not requiring a radioactive substrate and
being relatively rapid.

The third assay listed, immunoelectrophoresis or IEP, was the method
of choice in the early days of IgAl protease research (166), but its use
has declined due to the length of time required to obtain results (~24
hours) and the inability to quantitate enzyme activity. This technique
has also been replaced by methods which allow for more samples to be
processed in the same length of time (eg. ELISA and SDS-PAGE).

Rocket immunoelectrophoresis, a variation of the IEP, offers the
advantages of processing more samples in the same length of time and the
ability to quantitate a protease preparation (139). The chosen
definition of semiquantitative refers to the ability to compare the
activity of an experimental sample to the activity of a standard sample
but not to determine the activity of a sample de nova. This technique
is not widely used due to the lack of ability to quantitate unknown
samples and the inherently low degree of cleavage specificity that can
be assessed.

The final example of a technique based on formation of antigen-

56

antibody complexes is the ELISA, used by some laboratories in the IgAl
protease field (223). The principle is based on solid phase immune-
csqpture of the the IgAl molecule with a-Lc. Following treatment with
'Igfltl protease and rinsing, the secondary antibody (a-Hca) is added to
identify remaining uncleaved IgAl. The assay is quantitative in that by
using a known quantity of IgAl, the amount of Fca remaining and the
degree of hydrolysis can be determined. The ELISA technique offers the
ability to perform many several assays simultaneously, safely, and
relatively inexpensively.

The second set of techniques listed is based upon physical separa-
tion of the cleavage products from the intact IgAl molecule. Many of
‘these techniques are commonly used to assay IgAl proteases today, with
the exception of analytical ultracentrifugation.

Cellulose acetate electrophoretic separation is still used by some
Ilaboratories, but generally has been replaced by techniques allowing for
Inore samples to be processed in a given time frame (216). The technique
is sensitive and specific, can use either unlabeled or radiolabeled
Substrate, and is a rapid assay to perform and analyze. The chief
tirawback is that the apparatus limits the number of samples that can be
separated in a given run.

The SDS-PAGE-based assays are probably the method of choice in use
today by most laboratories studying IgAl proteases. The technique
tallows for’ multiple samples to 'be 'processed simultaneously, offers
consistency, is amenable to quantitative studies, and is relatively
rapid. While the technique is frequently performed with reducing
conditions and. radiolabeled. substrate to maximize sensitivity’ while

conserving substrate, it may also be performed with unlabeled substrate

57
for detection by staining (19).

A variation of the SDS-PAGE separation technique can be performed
with nonreducing conditions to preserve the tertiary structure of the
cleavage products for detection with immunoblot analysis (1). This
alternate technique is used less frequently but has the advantage of
demonstrating that the cleavage fragments were intact rather than
dissociated. The disadvantages of the immunoblot procedure are the
longer time necessary to obtain results, the inability to quantitate
enzyme activity, and the quality limitations of the detection reagents.

The solid-phase plate type assays are used. for screening large
numbers of isolates simultaneously and are most often used to select an
IgAl protease-producing isolate from a population of non-producers. All
of these methods are strictly qualitative. The first described method
utilizes radiolabelled IgAl immobilized via an a-Fca antibody (74). The
basis of the technique involves overlaying a bacteria covered surface
with a layer of immobilized IgAl. Secreted IgAl protease from
producing colonies will cleave the immobilized IgAl allowing the freed
Faba to diffuse so that it can ‘be recovered, usually employing a
nitrocellulose membrane (74). IgAl protease producing isolates are
identified by a location-corresponding radioactive signal.

Two variations have arisen from this method. One involves prior
immobilization of the labeled IgAl on the nitrocellulose membrane and
incubation on a bacteria covered surfacef After allowing for diffusion
of the liberated Faba, IgAl protease producers can be detected by a
decrease in the amount of radioactive signal (87). Another variation
incorporates a-Lc antibodies bound to the nitrocellulose membrane to

improve the efficiency of Faba recovery and differentiates IgAl

58
protease-positive isolates by increased radioactive signal (29).

Analytical ultracentrifugation has been used to verify the
composition of IgAl protease cleavage products but is rarely used today
due to the equipment, time, and expense involved in the procedure (119).

Finally, the chromatographic separation of intact and cleaved IgAl
has been used to identify and quantitate proteolytic activity (185).
The advantages of the technique are its speed and its consistency in
separation of the fragments. The disadvantage is that the HPLC
equipment required to perform the assay is not a common piece of
laboratory equipment due to the cost.

IgAl protease comparisons. A fairly extensive listing of bacterial
species have now been demon-strated to produce an IgAl protease and
these are listed in Table 5. There are many more examples of gram
negative than gram positive species, and a single mycoplasma. Although
the number of producing species has grown to include many previously
considered to be negative (187,210), all of the bacteria listed in the
table still reside on or in the mucosal epithelia. To have been
included in this listing, an organism must have met the following
criterion: IgAl is degraded and an intact Fca must have been
demonstrated. An organism may have also shown the capacity to degrade
IgG, as in the case of some of the Bacteraides spp. (70,114). However,
the ability to degrade multiple immunoglobulins was not shown to be a
function of a single enzyme, therefore including the possibility that a
second type of protease is being produced. Likewise, the unstable
nature of enzyme production shown for some of the enteric organisms
isolated from patients with urinary tract infections did not disqualify

those species from being included due to the original papers'

59

Table 5. Bacterial species demonstrated to produce IgAl protease.

 

Species

Acinetabacter calcaaceticus

var . ani tra tus

A. calcaaceticus var. lwaffi
Bacteroides asaccharalyticus

. buccae
. buccalis
. denticala
. laeschii
. melaninogenicus
. aralis
aris
. veraralis
Bifidabacterium sp.
Capnacytophaga gingivalis
C. achracea
C. sputigena
Clostridium ramasum
Escherichia coli
Gemella haemalysans
Haemaphilus aegyptius
H. influenzae

monumental-nu:

H. parahaemalyticus
H. parainfluenzae
Klebsiella pneumoniae
Moraxella spp.
Neisseria ganarrhoeae
N. meningitidis
Proteus mirablis

P. penneri

P. vulgaris
Pseudomonas aeruginasa
Serratia marcescens

Streptococcus mitis biovar l

S. aralis

S. pneumoniae

S. sanguis

Ureaplasma urealyticum

O
O

100
100
100
100

Site8 Refs,
UT 175
UT 175
0C 114
P0 70
DP 70
DP,PO 70
P0 70
SP 70,114
P0 70
P0 70
0C 70
IBD 72,248
NR 70
OC,PO 70
NR 70
IBD 72,248
UT 175
DP 123
Conjunctiva 119
NP,CSF,B,Ear 96,119
150,192
NP,OC 123,150,192
NP 123,150
UT 175
UT 175
UT 213, d
NP,CSF,B 213
F,UT,SP,Ear 175,249,250
F 250
F,UT,AB,Ear 175,250
UT 175
UT 175
DP 225
DP 225
NP,CSF,B,SP 119,150,192
DP 225
UT 107,115,235

 

m
I

Abbreviations are as follow: UT, urinary tract infection; 0C,

oral cavity; P0, periodontal infection; DP, dental plaque; SP,
sputum; IBD, inflammatory bowel disease; NR; not recorded; B,
blood; NP, nasopharynx; CSF, cerebrospinal fluid; F, feces;

and AB, abcess.
b - NR; not recorded.

Q0
II

Discrepancy between the references.
M. H. Mulks, unpublished observation.

60

descriptions of these enzymes as being transiently produced (166,175).

Some of the sites of cleavage have been identified by terminal amino
acid sequencing of the cleavage products (117,122,l90,l94,213,225).
Figure 4 shows an alignment of the hinge regions of IgAl and the
allotypes of IgA2 (170). The cleavage specific sites are also indicated
as they have been defined by peptide sequencing with the exceptions of
the Ureaplasma urealyticum (115) and Capnacytaphaga achracea (70)
proteases which have been suggested, but have not been proven, to cleave

the 2351,1236 2231,8224

and peptide bonds, respectivley. With four
exceptions, all characterized IgAl proteases cleave the IgAl molecule in
the duplicated octamer. The proteases that cleave IgAl beyond amino
acids 225 or 240 are produced by Clastridium ramasum (71), Bacteraides
melaninagenicus (184), and C. achracea (all shown in Figure 4) with
Proteus mirablis as the fourth. Based on SDS-PAGE analyses, the P.
mirablis IgAl cleavage site is suggested to be outside of the hinge
region but this has not been conclusively demonstrated (249). The enzyme
produced by C. ramasum is noteworthy in that it also is able to cleave
the IgA2m(1) allotype molecule as well as the IgAl isotype (72).

Some of the species which are able to produce more than one speci-
ficity of IgAl protease show relationships between the enzyme type
produced and other phenotypes. Examples are the previously stated rel-
ationships between protease type and For serovar and auxotype in N.
gonorrhaeae (189), and the relationship between protease type and
capsular serotypes in both N. meningitidis and H. influenzae
(96,192,191,194). In H. influenzae, similar clustering arrangements

were proposed on the basis of antigenic relatedness and serotype

(122,123) and between the restriction site polymorphism pattern of iga

61

IgA2m(2) CRV—‘ —————— — PPPPPCC
IgA2m(1) CP+V ———————— PPPPPCC

225 238

235 2‘10
|

I I
lgA1 CPVPSTPPTPSPSTPPTPSPSCC

Cl. runosmm
S. mitts
S. aralis
S. pneumoniae
S. sanguls

 

1 B. melanlnogenicus
C. ochracea

Figure 4. Alignment of the IgAl, IgA2m(l) and IgA2m(2) hinge regions
showing the defined and proposed sites of cleavage of several IgAl
protease producing bacteria. The alignment of the molecular hinge
regions is as presented by Mestecky and Russell (170) and the IgAl

protease cleavage sites are as reported previously (70,71,115,ll7,

122 , 190, 194,213 , 225) .

l

 

l

N. gonorrhoeae 1
N. meningmdls l

H. influenzae 2
N. gonorrhoeae 2
N. meningltldls 2
U. urealgflcum

H. aeggptius
H. influenzae l

62
loci and biotype (221). Interestingly, a previous study of H.
influenzae suggested that there was no relationship between the enzyme
specificity and the biotype (191).

Table 6 summarizes some of the information known about the chemistry
of these enzymes, including the class of protease that a given IgAl
protease belongs to and the molecular weight, where reported. Many of
the species listed in Table 5 are not included in Table 6 due to the
lack of information regarding the chemistry of the enzyme. Some of the
highlights of the data are the facts that: 1) three of the four classes
of proteases are represented, with aspartic proteases being the
exception, and 2) these enzymes show a wide range of molecular weights.
It appears that this set of proteases with similarly narrow ranges of
substrate specificity is actually a grouping of very diverse enzymes
which were clustered into a group based solely on that one
characteristic.

In addition to N. gonarrhaeae, genes encoding IgAl protease activity
have been cloned from N. meningitidis (129,222), and H. influenzae
(23,24,85,86,l29,220). One comparative DNA hybridization study of the
relationships between these iga genes indicates that the two neisserial
genes and the H. influenzae gene share significant homology, and
indicates that the commensal Neisseria and Haemophilus species lack any
significant homology to these cloned iga genes (129). A second study
confirmed the latter observation within the genus Haemaphilus (24).
Heteroduplex analysis of a N. ganarrhaeae—N. meningitidis iga clones-
hybrid demonstrated a large region of nonhomology in the 5' end of the
gene with very good homology over the 3'-most one-third (222). A

comparison of the N. ganarrhaeae and H. influenzae gene sequences showed

63

Table 6. Comparison of the better described IgAl proteases.

 

Species

Bacteraides melaninagenicus
Capnacytaphaga spp.
'Fecal”d
Haemaphilus influenzae (type 1)
Neisseria gonarrhaeae (type 1)
N. ganarrhaeae (type 2)
N. meningitidis
Proteus spp.
Streptococcus aralis
S. pneumoniae
S. sanguis

Ureaplasma urealyticum

Enzyme classa M;

Referenceg
T 62 70,184
M NRC 70
M 40-60 166,212
5 108 6,220
NR 112 253
s 105-109 6,218,253
M NR 213
M NR 249,250
NR 100 225
M NR 117
M 100-186 78,135,225
M NR 235

 

b
I

Molecular weight (103).

0'
I

O
I

Not reported.

Q.
I

M, metalloprotease; S, serine

protease; T, thiol protease.

The first described example of an IgAl protease.

64
good homology in the 3' end of the gene (1‘. e. the fi-domain) with
clusters of divergence scattered over the upstream sequences (220).
This observation coupled with conservation of some of the important
amino acid positions suggested to the authors that the H. influenzae
IgAl protease may be secreted in a fashion similar to the N. ganarrhaeae
type 2 enzyme (220).

The iga gene from Streptococcus sanguis has also recently been
cloned (75). Comparative analyses have shown that there was no
detectable sequence homology between the S. sanguis gene and the genome
of non-IgAl protease producing S. sanguis strains, S. pneumoniae (all
Iga+), S. mutans (all Iga'), Clastridium ramasum, Bacteraides
melaninogenicus, H. influenzae, and N. ganarrhaeae (75,78).

When the enzymes are compared on the basis of their abilities to be
inhibited by polyclonal anti—IgAl protease sera, extensive heterogeneity
is indicated. Studies by the Kilian laboratory have suggested that
there are more than 15 different antigenic types of IgAl protease
produced by H. influenzae alone (117,123). These studies also indicate
that: l) the enzyme produced by Streptococcus pneumoniae while cleaving
the same peptide bond in IgAl as the enzymes from S. sanguis and S.
aralis (formerly S. mitiar; 225), is antigenically distinct from that of
the latter two; 2) two different strains of N. meningitidis each produce
two antigenically different IgAl proteases, one of which is similar to a
gonococcal enzyme; and 3) other than the two neisserial enzymes, there
are no close immunologic relationships between the IgAl proteases
produced by H. influenzae, N. ganarrhaeae, N. meningitidis, and S.
pneumoniae (123). Comparison of the IgAl proteases produced by the

Bacteraides and Capnacyaphaga species indicated that each Bacteraides

65
enzyme is serologically distinct and that the C. achracea and C.
sputigena proteases are identical but unrelated to the enzyme produced
by C. gingivalis (70). The point of this discussion is to emphasize the
heterogeneity that exists among the set of enzymes that we have
arbitrarily clustered as "the IgAl proteases".

Specific IgAl proteases. Some of the IgAl proteases have been
studied extensively and will be discussed briefly in order to highlight
some of the important or interesting data which has been generated. The
best studied examples other than N. gonarrhaeae come from the bacteria
that are most frequently responsible for causing septic meningitis (N.
meningitidis, H. influenzae, and S. pneumaniae) and S. sanguis, one of
the bacteria responsible for the initiation of dental plaque.

The two types of N. meningitidis IgAl protease cleave IgAl at the
same peptide bonds as the gonococcal enzymes (see Figure 4) and are
produced in a mutually exclusive fashion with the distribution of type
1- versus type 2-producing strains, approximately equal (194). Enzyme
type specificity is correlated with some capsular serotypes as serotype
A strains produce only type 1 IgAl protease and serotypes X and Y only
produce a type 2 protease (194). Serotypes B, C, D, W—l35, 29-E, Z, and
the non-serotypable strains were shown to produce either type of IgAl
protease (194). While iga genes have been cloned from N. meningitidis,
little information is available on these genes (129,222).

H. influenzae has been studied intensively, if not aggressively,
with regard to the production. of IgAl proteases. In fact, three
different, independent laboratories published the identification of
enzyme production by the species (119,150,192). The species has also

been controversial with regard to specificity types. H. influenzae has

66

been proposed to produce enzymes with two to four different cleavage
site specificities, and enzymes of two different specificities have been
proposed to be produced simultaneously ‘by a clonal isolate of .H.
influenzae (23,96,119,122,123,191). This "double cutting" is associated
with nonserotypable strains (96,191). Currently, the consensus opinion
from the literature is that there are two different specificities of H.
influenzae enzyme, and that some strains produce both at the same time.
Evidence for these unusual strains having more than one copy of the gene
has been presented (221).

The type of enzyme specificity produced has been correlated with
serotype in H. influenzae as serotypes c and e make only the type 2
protease, serotypes b and f only produce the type 1 protease, and sero-
types a, d, and non-serotypables make either or both types of protease
(96,191,192). Similar correlations between antiserum inhibition profile
and serotype (122), biotype and restriction site pattern (221), and
serotype and restriction site pattern (23) have been presented.

The H. influenzae iga genes encoding for both types 1 and 2 have
been cloned by three different laboratories (23,24,85,86,129,220).
Comparison of the types 1 and 2 genes showed them to be very homologous
to each other with two major regions of divergence; one, near the 5' end
of the gene, contained a 370 bp section that was shown to be responsible
for the IgAl cleavage site specificity (86).

Like H. influenzae, three separate laboratories published the ident-
ification of an IgAl protease in Streptococcus pneumoniae
(119,150,191). However, since those reports, the amount of progress has
been less than that observed in H. influenzae and N. gonarrhaeae. Some

interesting results have been presented regarding the coproduction of

67
glycosidases by S. pneumoniae and their ability to alter the migration
of IgAl cleavage fragments on IEP and SDS-PAGE analysis (117).

Streptococcus sanguis was the first microbe identified as producing
an IgAl protease (211) and much of the early IgAl enzymology centered
around this enzyme (117,135,143,214). Originally, it was thought that
only some strains of the species produced IgAl protease (~70%; 116,210)
but recently, streptococcal taxonomy has been revised and it appears
that 100% of strains now identified as S. sanguis produce IgAl protease
while those strains which were Iga- are now classified as S. gordanii
(225).

The gene encoding IgAl protease activity in S. sanguis has been
cloned (75). Analysis suggests that it: 1) is a metallo-protease based
on conservation of a gram positive bacterial zinc-binding-site sequence,
2) lacks an apparent signal sequence, yet is secreted to the periplasmic
space in E. cali, and 3) has an unusual P-, G-, and E-rich stretch of 20
amino acids, repeated ten times (78). The classification as a
metalloprotease is substantiated by inhibitor studies and by site
specific mutagenesis of the proposed zinc binding site (78,214).

Two of the other oral streptococci also produce IgAl protease, S.
aralis (previously S. mitiar) and S. mitis biovar 1 (225). Due to the
reorganization of this taxonomic group, the organisms formerly classed
as S. mitiar have been redefined as S. aralis, all strains of which
produce IgAl protease, and the division of S. mitis into two biovars
places the IgAl protease producing strains into biovar l, 20% of which
produce the enzyme while none of the S. mitis biovar 2 strains do (225).
Of interest is the observation that all of the S. aralis strains tested

also degrade the carbohydrate moieties on IgAl as well as proteolyzing

68

it into Faba and Fca (225). A similar phenomenon was observed for the
two IgAl protease producing biovar 1 strains of S. mitis. The authors
reported that if the IgAl preparations are pre-treated with either
neuraminidase or an S. mitis glycosidase-enriched preparation, that IgAl
has increased susceptibility to these streptococcal IgAl proteases while
pretreatment with other commercially prepared endoglycosidases had no
effect on the susceptibility of the IgAl (225). It was reported that
the S. aralis IgAl protease is consistently nonimmunogenic in
experimental animals (225).

Other oral pathogens involved in periodontal disease, Bacteraides
spp. and Capnacytaphaga spp., produce IgAl proteases (70,114). Some
strains of these species also cleave or degrade IgG and/or IgA2 (114),
but other of the Bacteraides species (eg. B. gingivalis) have been
demonstrated to produce nonspecific proteases, so it is not out of the
question that other species of the genus may too. By definition, the
IgAl proteases cleave human IgAl and produce intact Faba and ECG
fragments, and when analyzed, some strains of the species of the
Bacteraides and Capnacytaphaga included in Table 5 were observed to have

at least an intact Fca remaining (114). Perhaps examination of the

l
taxonomy of these important oral pathogens will clear the picture as it
did for the oral streptococci. Also, similar to the observation of low
immunogenicity with the S. aralis protease, the IgAl proteases of the
Bacteraides and Capnacytaphaga are reported to be difficult to get
protease activity inhibiting titers against in experimental animals
(70).

Relatively little is known about the IgAl proteases produced by

Clastridium ramasum and an unidentified species of Bifidabacterium. The

69

report of enzyme production in Bifidabacterium (72) is the only one of
its kind and another report from some members of the same research group
indicated that they have not been able to find other strains of
Bifidabacterium which produce the protease (248). That same report also
suggested that the ability of C. ramasum to produce IgAl protease is
considered to be a rare event. What is interesting about the
clostridial enzyme is the fact that it not only cleaves IgAl but it also
is able to specifically hydrolyze IgA2 of the A2m(1) allotype at the
same peptide bond that it hydrolyzes IgAl (71; Figure 4).

Similarly, the IgAl protease expressed by Ureaplasma urealyticum is
not ‘very' well characterized. other than. the suggested. site of IgAl
cleavage (115) and the suggestion that it may be a metalloprotease
(235).

Of interest are the reports of IgAl protease production by members
of the Enterobacteriaceae, recalling that the original description of
this enzyme grouping was from a fecal specimen (166,175,249,250). One
research group described the unstable production of an IgAl protease by
eight different species of eight different genera isolated from patients
with urinary tract infections (175). Also, two reports by another group
indicates that several clinical isolates of Proteus species produce a
specific IgAl protease (249,250). Every P. mirablis and P. penneri
strain tested cleave IgAl outside of the hinge region as do one-half of
the P. vulgaris strains tested (250). These enzymes are sensitive to
inhibition by 5 mM EDTA which is also reminiscent of the fecal sample
enzyme described by Mehta, et a1. (166,249,250).

Roles in virulence. In addition to the studies with the gonococcal

IgAl protease and its effect(s) on virulence or immune system functions,

70
several similar studies have been reported for the IgAl proteases of the
other enzyme producing species, and some of these have been reviewed
previously (118, 120,134). Some of these data will be summarized for
the purpose of describing the proposed important role for the IgAl
proteases in the virulence of these organisms.

Several studies have described the production of the IgAl proteases
in vivo by' either isolating the enzyme activity or IgAl cleavage
products from patient specimens including saliva (134), dental plaque
(l), intestinal contents (166), cerebrospinal fluid of H. influenzae
meningitis patients (120), and nasopharyngeal secretions (120).
Bacteria with Fabal bound to their surfaces have been isolated from
fresh dental plaque samples (1). Bacteria (both Iga+ and Iga-) which
were incubated in vitra with whole saliva also bound Fabal’ suggesting
that even species that do not produce the IgAl protease may still
encounter benefits from their production in viva (1).

One of the proposed benefits that may occur via hydrolysis of IgAl
is the abrogation of the anti-adherence functions ascribed to IgAl,
tested by examining the adherence of oral streptococci to
hydroxyapatite, a component of dental enamel (121). One of the
conclusions presented was that sIgA has no inhibitory effect on the
adherence of IgAl protease producing bacteria while that the same sIgA
preparations are inhibitory for the adherence of non-protease producing
strains. That study also suggested that if the sIgA is directed
specifically against some undefined antigen(s), potentially an adhesin
molecule, production of IgAl protease is of no advantage for the
bacteria, and their adherence is inhibited (121).

A subsequent study suggested that IgAl protease pretreatment of the

71
sIgA, or of bacteria that have been incubated with sIgA, destroys the
adherence inhibition function of sIgA for these oral streptococci (224).
Examination of the bacteria exposed to treated sIgA indicated that they

had Faba bound to their surfaces and that it may have actually been of

1
benefit in attaching to the surface (224). A conclusion presented sug-
gested that the production of IgAl protease promoted bacterial adherence
to these surfaces.

It is interesting that the oral streptococci which are among the
first bacteria to colonize the surface of the tooth, produce one or more
enzymes which inactivate one of the primary antibacterial systems in the
secretions. It would also be of interest to examine this suggested
cooperative effect with the streptococcal glycosidases and some of the
other well characterized IgAl proteases to see if the effect is general
or specific to the streptococcal proteases.

Observations have been presented indicating that the IgAl protease
isolated from S. aralis is consistently nonimmunogenic in experimental
animals (123,225). These observations prompt the question as to whether
the proteins are truly nonimmunogenic or are the vaccinees tolerized to
them prior to the trial immunization? Presumably, the test animals have
been exposed to these oral streptococci their entire lives and due to
continual exposure, may have become tolerized to the antigen. There are
no reports of humans raising either a secretary or serum antibody titer
to the oral streptococcal IgAl proteases. Gilbert, et al. reported that
both serum and secretory IgA which were inhibitory for the type 2
gonococcal protease had. no effect on the S. sanguis enzyme (77).

Perhaps in humans these enzymes are not reacted to by either the serum

or mucosal immune systems which predisposes us to dental caries?

72

Experimental data from an in vitra virulence study of Iga+ and Iga-
H. influenzae mirrors that presented from similar experiments with Iga+
and Iga- strains of N. ganarrhaeae (66). In the nasopharyngeal organ
culture system, there is no difference in the abilities of encapsulated
Iga+ and Iga- H. influenzae to attach, damage, and invade the mucosal
epithelium. Likewise, the lack of production of a capsule has no
detectable influence in the tested virulence of these Iga+ and Iga-
strains in this in vitra system.

Paradoxically, it appears that the ability to produce IgAl proteases
may be of value to the oral streptococci while affording no obvious
benefit to pathogens of the nasopharynx or genitourinary tract. One
problem with comparing the two types of experimental systems is that in
the streptococcal adherence system, the time frames for experiments are
on the order of a few hours and the use of whole saliva which had been
clarified by centrifugation but had not been sterilized; and in the
organ culture systems, the experiments last for up to 72 hours and
incorporate conditions which are essentially sterile with regard to
normal microbial flora. Perhaps either the long time frames or the
near-xenobiotic conditions of the organ culture systems are influencing
the measured response(s), or perhaps in the streptococcal adherence
system, the suspected presence of other flora is influencing the
response(s).

A model has been presented for the invasion of some of the IgAl
protease producing bacteria, specifically for H. influenzae, but is
applicable to the other invasive species (120). It has been suggested
that some children who present with H. influenzae type b meningitis have

preexisting higher serum and nasopharyngeal titers against these

73

organisms than children who are carriers of these strains (120).
According to the model, IgAl protease-producing bacteria are able to
cleave IgAl and inactivate its proposed Fc-mediated functions, enabling
the bacteria to adhere to the mucosa and colonize a host. The host
responds to the bacteria with a secretary immune response (assumed to be
sIgA with some of these antibodies directed against, and inhibitory for
the IgAl protease), effectively neutralizing the pathogen and keeping
its virulence in check, resulting in carrier status.

However, if an immunized host is subsequently challenged with a
larger dose of the same strain, or with antigenically different strain,
the bacteria are able to cleave the sIgAl, coating themselves in Fab“1
and shielding themselves from a further immune response, facilitating
the adherence and invasion which results in systemic infection.

The key difference between the two outcomes is the presence of IgAl
protease-inhibitory sIgA in the carrier host and the absence of same in
the second host. This also lends additional significance to the data
from Kilian's group which pointed out the great antigenic diversity
observed among the H. influenzae IgAl proteases (117,122,123).

One final observation on the roles of IgAl protease in immune
(dys)function concerns an observation that suggested that children with
allergies might be colonized more heavily by IgAl protease producing
bacteria than those children which do not present with allergic
symptoms. Kilian, et a1. presented data from 97 children which shows
that greater than sixty percent of these children with allergies have
IgAl cleavage products in their nasopharyngeal secretions while only
twelve percent of the non-allergic children do (254). Both groups were

reported to produce similar levels of IgA in their secretions, so the

74

difference is not due to altered production or lack of IgA production by
either group (254). Some proposed reasons for this correlation include
the inability of the cleaved sIgAl in allergic children to exclude
allergens, or perhaps the proposed down-regulatory effect of sIgA is
abrogated due to the inactivated sIgAl, and that predisposes the
children to an unregulated immune response allowing for reactions to
innocuous allergens.

As evidence for functions of the IgAl proteases continues to accumu-
late and is reviewed, it becomes apparent that these enzymes may play
important roles in the pathogenic processes of several species of
bacteria as well as the regulation of the "normal" immune response. It
also points out that the roles of these enzymes are far from well
understood and that there likely are functions for these enzymes which
still need to be proposed and tested.

The amount of information that exists on the Neisseria gonarrhaeae
IgAl proteases is as large as that which exists for other protease-
producing species, but needs further study. The observation that the
enzymes are stably produced under the relatively "comfortable",
nutrient-rich, conditions used in vitra prompted examination into the
potential for previously undescribed IgAl protease "housekeeping"
functions and the potential for environmental regulation, which may
relate to either normal cellular function or to virulence. The previous
descriptions of an isogenic set of N. gonarrhaeae IgAl protease mutants
did not discuss some phenotypes such as the ability to respond to
environmental stresses and the protein profiles of isolated membranes
(47,131). With that in mind, the experiments and data presented in this

thesis are important steps in addressing these problems.

75

To address these problems, a pair of N. gonarrhaeae deletion-
disruption iga mutants were constructed by recombinant DNA technology.
No obvious phenotypic differences in viability were correlated with the
mutant allele. However, when isolated membrane protein profiles of the
wild type and mutant variants were compared, differences were consist-
ently observed. Experiments to correlate the production of the type 2
IgAl protease with membrane protein profiles suggested that isogenic
strains which fail to produce the protease also failed to insert a 56
kDa protein into their outer membranes and that one of two Aiga
deletion-disruption mutant strains also failed to insert a 34.5 kDa
protein into the cyoplasmic membrane. The identities of these IgAl—
protease associated proteins are unknown but several potential
identities have been ruled out.

Data are also presented which describe new substrates for the type 2
gonococcal IgAl protease which may later be shown to be important in
either normal cellular functions or in pathogenesis. The identities
and/or functions of these substrate membrane proteins are not presently
known. IgAl protease-susceptible outer membrane proteins were also
demonstrated in an unrelated strain of N. gonarrhoeae, Escherichia cali
and Actinabacillus pleuropneumaniae suggesting that the protease may
play a broader role in pathogenesis than previously suggested.

Also presented are data which indicate Fe3+ concentration-regulation
of the type 2 IgAl protease gene. These data indicated that the
observed increase in enzyme activity under low iron conditions was not
due to allosteric modulation of the IgAl protease. These data are
suggestive additional evidence for the hypothesis that the IgAl

proteases are a virulence factor in N. gonarrhaeae. While a putative

76

Escherichia cali Fur-binding regulatory sequence was identified in the
promoter of the 13.32 gene, experiments to test its function in a defined
E. coli genetic background suggested that the sequence was non-
functional. Further experiments to identify a proposed fur homologue in
N. ganorrhaeae were unsuccessful.

In conclusion, the data presented here give further evidence that
the IgAl protease is a virulence factor and suggest that it may play
other roles in the maintenance of the gonococcal cell during growth in

vi tra .

10.

11.

12.

77

REFERENCES

. Ahl, T., and J. Reinholdt. 1991. Detection of immunoglobulin Al

protease-induced Fab fragments on dental plaque bacteria. Infect.
Immun. 59:563-569.

. Ahnen, D. J., W. R. Brown, and T. M. Kloppel. 1985. Secretary

component: The polymeric immunoglobulin receptor. Gastroenterology
89:667-682.

. Arka, R. J. 1972. Neisseria ganarrhaeae: experimental infection of

laboratory animals. Science. 177:1200-1201.

. Arko, R. J. 1989. Animal models for pathogenic Neisseria species.

Clin. Microbiol. Rev. 2:856-859.

. Arko, R. J., C.-Y. Chen, W. O. Schalla, S. K. Sarafian, C. L.

Taylor, J. S. Knapp, and S. A. Horse. 1991. Binding of S protein by
Neisseria ganarrhoeae and potential role in invasion. J. Clin.
Microbiol. 29:70-75.

. Bachovkin, W. W., A. G. Plaut, G. R. Flentke, M. Lynch, and C. A

Kettner. 1990. Inhibition of IgAl proteinases from Neisseria
ganarrhaeae and Hemophilus influenzae by peptide prolyl boronic
acids. J. Biol. Chem. 265:3738-3743.

. Barrera, 0., and J. Swanson. 1984. Proteins IA and IB exhibit

different surface exposures and orientations in the outer membrane
of Neisseria ganarrhaeae. Infect. Immun. 44:565-568.

. Bergstrom, 8., K. Robbins, J. M. Koomey, and J. Swanson. 1986.

Piliation control mechanisms in Neisseria ganarrhaeae. Proc. Natl.
Acad. Sci. USA 83:3890-3894.

. Berish, S. A., T. A. Mietzner, L. W. Mayer, c. A. Genco, B. P.

Holloway, and S. A. Marse. 1990. Molecular cloning and charac-
terization of the structural gene for the major iron-regulated
protein expressed by Neisseria gonorrhoeae. J. Exp. Med. 171:1535-
1546.

Biswas, G. D., E. Y. Blackman, and P. F. Sparling. 1980. High-
frequency conjugal transfer of a gonococcal penicillinase plasmid.
J. Bacteriol. 143:1318-1324.

Biswas, G. D., K. L. Burnstein, and P. F. Sparling. 1986.
Linearization of donor DNA during plasmid transformation in
Neisseria gonarrhaeae. J. Bacteriol. 168:756-761.

Biswas, G. D., T. F. 80:, E. Blackman, and P. F. Sparling. 1977.
Factors affecting genetic transformation of Neisseria gonarrhaeae.
J. Bacteriol. 129:983-992.

l3.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

78

Biswas, G. D., and P. F. Sparling. 1981. Entry of double-stranded
deoxyribonucleic acid during transformation in Neisseria
gonarrhaeae. J. Bacteriol. 145:638-640.

Biswas, G. D., S. A. Thompson, and P. F. Sparling. 1989. Gene
transfer in Neisseria gonarrhaeae. Clin. Microbiol. Rev. 2:824-828.

Black, W. J., and J. G. Cannon. 1985. Cloning of the gene for the
common pathogenic Neisseria H.8 antigen from Neisseria gonarrhaeae.
Infect. Immun. 47:322-325.

Black, W. J., R. S. Schwalbe, I. Nachamkin, and J. G. Gannon. 1984.
Characterization of Neisseria gonarrhaeae protein 11 phase varia-
tion by use of monoclonal antibodies. Infect. Immun. 45:453-457.

Blake, H., R. K. Holmes, and J. Swanson. 1979. Studies on
gonococcus infection. XVII. IgA -c1eaving protease in vaginal
washings from women with gonorr ea. J. Infect. Dis. 139:89-92.

Blake, H. S., E. J. Lytton, M. E. Seiff, and E. C. Gotschlich.
1988. Studies on gonococal protein III, p. 85-97. In M. A. Horwitz
(ed.), Bacteria-host cell interaction. Alan R. Liss, Inc., New
York, NY.

Blake, H. S., and J. Swanson. 1978. Studies on gonococcus infect-
ion. XVI. Purification of Neisseria ganarrhoeae immunoglobulin A1
protease. Infect. Immun. 22:350-358.

Blake, M. S., L. M. Wetzler, E. c. Gotschlich. and P. A. Rice.
1989. Protein III: Structure, function, and genetics. Clin.
Microbiol. Rev. 2:860-863.

Boslego, J., J. Ciak, P. Hitchcock, J. Swanson, E. Tramont, J.
Sadoff, and J. M. Koomey. 1988. Neisseria gonorrhaeae human
infection study: clinical aspects, abstr. VC3, p. 107. 6th Int.
Pathog. Neisseria Conf., Pine Mountain, GA.

Bovre, K. 1984. Family VIII. Neisseriaceae Prevot 1933, 119, p.
288-290. In N. R. Krieg and J. G. Holt (ed.), Bergey's manual of
systematic bacteriology, vol. 1. The Williams and Wilkins, Co.,
Baltimore, MD.

Brisker, J., M. Mulks, E. R. Moron, A. G. Plaut, and A. Wright.
1985. Physical and genetic analysis of DNA regions encoding the
immunoglobulin A proteases of different specificities produced by
Haemophilus influenzae. Infect. Immun. 47:370-374.

Bricker, J., M. H. Mulks, A. G. Plaut, E. R. beon, and A. Wright.
1983. IgAl proteases of Haemophilus influenzae: cloning and char-
acterization in Escherichia cali K-l2. Proc. Natl. Acad. Sci. USA
80:2681-2685.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

79

Britigan, B. E., M. S. Cohen, and P. F. Sparling. 1985. Gonococcal
infection: a model of molecular pathogenesis. N. Eng. J. Med.
26:1683-1694.

Britigan, B. E., D. Klapper, T. Svendsen, and H. S. Cohen. 1988.
Phagocyte-derived lactate stimulates oxygen consumption by
Neisseria ganarrhaeae. An unrecognized aspect of the oxygen
metabolism of phagocytes. J. Clin. Invest. 81:318-324.

Brooks, G. F., and C. J. Lammel. 1988. Immune response to gonococ-
cal infection, abstr. MT42, p. 182. 6th Int. Pathog. Neisseria
Conf., Pine Mountain, GA.

Brooks, G. F., and C. J. Lammel. 1989. Humoral immune response to
gonococcal infection. Clin. Microbiol. Rev. 2:85-810.

Brown, T. A., and I. G. Leak. 1989. A solid-phase immunoassay for
detection of IgAl protease activity on agar plates. J. Immunol.
Methods 123:241-247.

Brown, W. J., T. T. Lucas, and U. S. G. Kuhn. 1972. Gonorrhoea in
the chimpanzee: Infection with laboratory-passed gonococci and by
natural transmission. Brit J. Ven. Dis. 48:177-178.

Bucca, H. A. 1943. The effects of germicides on the viability and
on the respiratory activity of the gonococcus. J. Bacteriol.
46:151.

Buchanan, T. H., and E. C. Gotschich. 1973. Studies on gonococcal
infection. III. Correlation of gonococcal colony morphology with
infectivity for the chick embryo. J. Exp. Med. 137:196-200.

Buchanan, T. E., J. Swanson, K. K. Holmes, S. J. Kraus, and E. C.
Gotschlich. 1973. Quantitative determination of antibody to
gonococcal pili. Changes in antibody levels with gonococcal
infection. J. Clin. Invest. 52:2896-2909.

Burton, J., S. G. Wood, M. Lynch, and A. G. Plant. 1988. substrate
analogue inhibitors of the IgAl proteinases from Neisseria
ganarrhaeae. J. Med. Chem. 31:1647-1651.

Bygdeman, S., H. Beckman, D. Danielsson, and M. Norgren. 1982.
Genetic linkage between serogroup specificity and antibiotic
resistance in Neisseria ganorrhaeae. Acta. Pathol. Microbiol.
Immunol. Scand. Sect. B. 90:243-250.

Cannon, J. G. 1989. Conserved lipoproteins of pathogenic Neisseria
species bearing the H.8 epitope: Lipid-modified azurin and H.8
outer membrane protein. Clin. Microbiol. Rev. 2:81-84.

Cannon, J. G., W. Black, I. Nachamkin, and P. Stewart. 1985.
Monoclonal antibody that recognizes an outer membrane antigen
common to pathogenic Neisseria species but not to most
nonpathogenic Neisseria species. Infect. Immun. 43:994-999.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

80

Cannon, J. G., T. H. Buchanan, and P. F. Sparling. 1983. Confirma-
tion of protein I serotype of Neisseria ganarrhoeae with ability to
cause disseminated infection. Infect. Immun. 40:816-819.

Cannon, J. G., and P. F. Sparling. 1984. The genetics of the
gonococcus. Annu. Rev. Microbiol. 38:111-133.

Carbonetti, N., V. Simnad, C. Elkins, and P. F. Sparling. 1990.
Construction of isogenic gonococci with variable porin structure:
effects on susceptibility to human serum and antibiotics. Mol.
Microbiol. 4:1009-1018.

Carbonetti, N. H., V. I. Simnad, H. S. Seifert, M. So, and P. F.
Sparling. 1988. Genetics of protein I of Neisseria gonarrhaeae:
Construction of hybrid proteins. Proc. Natl. Acad. Sci. USA
85:6841-6845.

Chen, C.-Y., C. A. Genco, J. P. Rock, and S. A. Horse. 1989.
Physiology and metabolism of Neisseria ganarrhaeae and Neisseria
meningitidis: implications for pathogenesis. Clin. Microbiol. Rev.
2:835-840.

Chen, K. C. S., and T. H. Buchanan. 1980. Hydrolases from Neisseria
gonarrhaeae. The study of gonocosin, an aminopeptidase-P, a proline
iminopeptidase, and an asparaginase. J. Biol. Chem. 255:1704-1710.

Clark, V. L., L. A. Campbell, D. A. Palermo, T. M. Evans, and K. W.
Klimpel. 1987. Induction and repression of outer membrane proteins
by anaerobic growth of Neisseria gonarrhaeae. Infect. Immun.

55:1359-1364.

Clark, V. L., J. S. Knapp, S. Thompson, and K. W. Klimpel. 1988.
Presence of antibodies to the major anaerobically induced
gonococcal outer membrane protein in sera from patients with
gonococcal infections. Microb. Pathog. 5:381-390.

Cooper, H. D., D. Dever, K. Tempel, E. J. Moticka, T. Hindman, and
D. S. Stephens. 1987. Characterization of lymphoid cells from ten
human fallopian tube mucosa. Adv. Exp. Med. Biol. 16:387-394.

Cooper, M. D., 2. A. McGee, H. H. Mulks, J. H. Koomey, and T. L.
Hindman. 1984. Attachment to and invasion of human fallopian tube
mucosa by an IgAl protease-deficient mutant of Neisseria
gonarrhaeae and its wild-type parent. J. Infect. Dis. 150:737-744.

Cooper, H. D., P. A. McCraw, and H. A. Kelly. 1986. Localization of
gonococcal lipopolysaccharide and its relationship to toxic damage
in human fallopian tube mucosa. Infect. Immun. 51:425-430.

Cornelissen, C. N., G. Biswas, J. Tsai, D. Paruchuri, and P. F.
Sparling. 1990. Genetics of pathogenic Neisseria iron scavenging
systems: Molecular cloning of transferrin receptor genes, abstr.,
p. 17. 7th Int. Pathog. Neisseria Conf., Berlin.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

81

Davies, J. K. 1989. DNA restriction and modification systems in
Neisseria ganarrhoeae. Clin. Microbiol. Rev. 2:878-882.

Deal, C. D., N. Stromberg, G. Nyberg, S. Normark, K.-A. Karlsson,
and M. So. 1988. Pilin independent binding of Neisseria ganarrhaeae
to immobilized glycolipids, p. 245-250. In J. T. Poolman, et al.
(ed.), Gonococci and meningococci. Kluwer Academic Publishers,
Dordrecht, the Netherlands.

de Hormaeche, R. D., H. Jessap, and K. Senior. 1988. Gonococcal
variants selected by growth in viva or in vitra have antigenically
different LPS. Microb. Pathog. 4:289-297.

Dekker, N. P., C. J. Lammel, R. E. Handrail, and G. P. Brooks.
1990. Opa (protein II) influences gonococcal organization in
colonies, surface appearance, size and attachment to human
fallopian tube tissues. Microb. Pathog. 9:19-31.

Delacroix, D. L., C. Dive, J. C. Rambaud, and J.-P. Vaerman. 1982.
IgA subclasses in various secretions and in serum. Immunology
47:383-385.

Dempsey, J. F., W. Litaker, and J. G. Cannon. 1988. Physical
mapping of the gonococcal chromosome using pulsed field
electrophoresis, abstr. MB28, p. 67. 6th Int. Pathog. Neisseria
Conf., Pine Mountain, GA.

Densen, P., C. HcRill, and S. C. Ross. 1988. The contribution of
the alternate and classical complement pathways to gonococcal
killing and C3 fixation, p. 639-697. In J. T. Poolman, et al.
(ed.), Gonococci and meningococci. Kluwer Academic Publishers,
Dordrecht, the Netherlands.

Dillon, J. R., and H. Pause. 1981. Relationship between plasmid
content and auxotype in Neisseria ganarrhoeae isolates. Infect.
Immun. 33:625-628.

Dillon, J. R., and K.-H. Yeung. 1989. fi-lactamse plasmids and
chromosomally mediated antibiotic resistance in pathogenic
Neisseria species. Clin. Microbiol. Rev. 2:8125-8133.

Dorward, D. W., and C. F. Caron. 1989. DNA-binding proteins in
cells and membrane blebs of Neisseria ganarrhaeae. J. Bacteriol.
171:4196-4201.

Dorward, D. W., c. F. Caron, and R. c. Judd. 1989. Export and
intercellular transfer of DNA via membrane blebs of Neisseria
gonarrhaeae. J. Bacteriol. 171:2499-2505.

Dorward, D. W., and R. C. Judd. 1988. The isolation and partial
characterization of naturally-evolved outer membrane blebs of
Neisseria gonarrhaeae, p. 349-356. In J. T. Poolman, et al. (ed.),
Gonococci and meningococci. Kluwer Academic Publishers, Dordrecht,
the Netherlands.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

82

Dudas, K. C., and M. A. Apicella. 1988. Selection and immuno-
chemical analysis of lipooligosaccharide mutants of Neisseria
ganarrhaeae. Infect. Immun. 56:499-504.

Duff, M. K., and J. K. Davies. 1988. Multiple restriction-
modification systems in Neisseria ganarrhaeae. Gene 74:227-228.

Dyer, D. W., E. P. West, and P. F. Sparling. 1987. Effects of serum
carrier proteins on the growth of pathogenic neisseriae with heme-
bound iron. Infect. Immun. 55:2171-2175.

Eisenstein, B. I., T. Sax, G. Biswas, E. Blackman, and P. F.
Sparling. 1977. Conjugal transfer of the gonococcal penicillinase
plasmid. Science 195:998-1000.

Farley, M. M., D. S. Stephens, M. H. Mulks, M. D. Cooper, J. V.
Bricker, S. S. Mirra, and A. Wright. 1986. Pathogenesis of IgAl
protease-producing and -nonproducing Haemophilus influenzae in
human nasopharyngeal organ cultures. J. Infect. Dis. 154:752-759.

Fishman, Y., J. Bricker, J. V. Gilbert, A. G. Plaut, and A. Wright.
1985. Cloning of the type 1 immunoglobulin Al protease from
Neisseria gonarrhoeae and secretion of the enzyme from Escherichia
cali, p. 164-168. In G. K. Schoolnik, et al. (ed.), The pathogenic
Neisseriae. Am. Soc. Microbiol., Washington, DC.

Fleming, T. J., D. E. Wallsmith, and R. S. Rosenthal. 1986.
Arthropathic properties of gonococcal peptidoglycan fragments:
implications for the pathogenesis of disseminated gonococcal
disease. Infect. Immun. 52:600-608.

Fohn, M. J., T. A. Mietzner, T. W. Hubbard, S. A. Morse, and E. W.
Hook, III. 1987. Human immunoglobulin G antibody response to the
major gonococcal iron-regulated protein. Infect. Immun. 55:3065-
3067.

Frandsen, E. V. G., J. Reinholdt, and M. Kilian. 1987. Enzymatic
and antigenic characterization of immunoglobulin Al proteases from
Bacteraides and Capnacytaphaga spp. Infect. Immun. 55:631-638.

Fujiyama, Y., M. Iwaki, K. Hodohara, S. Hosoda, and K. Kobayashi.
1986. The site of cleavage in human alpha chains of IgAl and
IgA2:A2m(l) allotype paraproteins by the clostridial IgA protease.
Mol. Immunol. 23:147-150.

Fujiyama, Y., K. Kobayashi. S. Sends, Y. Denna, T. Bamba, and S.
Hosoda. 1985. A novel IgA protease from Clastridium sp. capable of
cleaving IgAl and IgA2 A2m(1) allotype but not IgA2 A2m(2) allotype
paraproteins. J. Immunol. 134:573-576.

Genco, C. A., J. S. Knapp, and V. L. Clark. 1984. Conjugation of
plasmids of Neisseria ganarrhaeae to other Neisseria species:

potential reservoirs for the fl-lactamase plasmid. J. Infect. Dis.
150:397-401.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

83

Gilbert, J. V., and A. G. Plant. 1983. Detection of IgA protease
activity among multiple bacterial colonies. J. Immunol. Methods
57:247-251.

Gilbert, J. V., A. G. Plaut, Y. Fishman, and A. Wright. 1988.
Cloning of the gene encoding streptococcal immunoglobulin A

protease and its expression in Escherichia cali. Infect. Immun.
56:1961-1966.

Gilbert, J. V., A. G. Plaut, and B. Longmaid. 1983. Inhibition of
bacterial IgA proteases by human secretory IgA and serum. Ann. N.
Y. Acad. Sci. 409:625-634.

Gilbert, J. V., A. G. Plaut, B. Longmsid, and M. E. Lamm. 1983.
Inhibition of microbial IgA proteases by human secretory IgA and
serum. Mol. Immunol. 20:1039-1049.

Gilbert, J. V., A. G. Plaut, and A. Wright. 1991. Analysis of the
immunoglobulin A protease gene of Streptococcus sanguis. Infect.
Immun. 59:7-17.

Goodman, S. D., and J. J. Scocca. 1988. Identification and
arrangement of the DNA sequence recognized in specific trans-
formation of Neisseria ganarrhaeae. Proc. Natl. Acad. Sci. USA
85:6982-6986.

Gotschlich, E. C., M. 8. Blake, J. M. Koomey, M. Seiff, and A.
Derman. 1986. Cloning of the structural genes of three H8 antigens
and of protein III of Neisseria ganarrhaeae. J. Exp. Med. 164:868-
881.

Gotschlich, E. C., and M. E. Seiff. 1987. Identification and gene
structure of an azurin-like protein with a lipoprotein signal
peptide in Neisseria ganarrhaeae. FEMS Microbiol. Lett. 43:253-255.

Gourby, A., G. Bourg, and M. Remus. 1986. Previously undescribed
6.6-kilobase R plasmid in penicillinase-producing Neisseria
ganarrhaeae. Antimicrob. Agents Chemother. 29:1095-1097.

Gregg, C. R., M. A. Melly, and Z. A. McGee. 1980. Gonococcal lipo-
oligosaccharide: a toxin for human fallopian tube mucosa. Am. J.
Obstet. Gynecol. 138:981-984.

Griffiss, J. M., H. Schneider, R. E. Mandrell, G. A. Jarvis, J. J.
Kim, 8. Gibson, and M. A. Apicella. 1987. The immunochemistry of
neisserial LOS. Antonie van Leeuwenhoek J. Microbiol. 53:501-507.

Grundy, F. J., A. G. Plaut, and A. Wright. 1987. Haemophilus
influenzae immunoglobulin Al protease genes: cloning by plasmid
integration-excision, comparative analyses, and localization of
secretion determinants. J. Bacteriol. 169:4442-4450.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

84

Grundy, F. J., A. G. Plaut, and A. Wright. 1990. Localization of
the cleavage site specificity determinant of Haemophilus influenzae
immunoglobulin A1 protease genes. Infect. Immun. 58:320-331.

Halter, R., J. Pohlner, and T. F. Meyer. 1984. IgA protease of
Neisseria ganarrhaeae: isolation and characterization of the gene
and its extracellular product. EMBO J. 3:1595-1601.

Halter, R., J. Pohlner, and T. F. Meyer. 1989. Mosaic-like organ-
ization of IgA protease genes in Neisseria gonarrhaeae generated by
horizontal genetic exchange in viva. EMBO J. 8:2737-2744.

Handsfield, H. H. 1984. Gonorrhea and uncomplicated gonococcal
infection, p.205-220. In K. K. Holmes, et al. (ed.), Sexually
transmitted diseases. McGraw-Hill Book Co., New York, NY.

Heckels, J. E. 1989. Structure and function of pili of pathogenic
Neisseria species. Clin. Microbiol. Rev. 2:S66-S73.

Heckels, J. E., M. Virji, K. Zak, and J. N. Fletcher. 1987.
Immunobiology of gonococcal outer membrane protein I. Antonie van
Leeuwenhoek J. Microbiol. 53:461-464.

Hildebrandt, J. F., L. W. Mayer, S. P. Wang, and T. M. Buchanan.
1978. Neisseria ganarrhaeae acquire a new principal outer-membrane
protein when transformed to resistance to serum bactericidal
antibody. Infect. Immun. 20:267-272.

Hitchcock, P. J. 1984. Analysis of gonococcal lipopolysaccharide in
whole cell lysates by sodium dodecyl sulfate-polyacrylamide gel
electrophoresis: stable association of lipopolysaccharide with the
major outer membrane protein (protein I) of Neisseria ganarrhaeae.
Infect. Immun. 46:202-212.

Hitchcock, P. J. 1989. Unified nomenclature for pathogenic
Neisseria species. Clin. Microbiol. Rev. 2:S64-S65.

Ingwer, I., B. H. Peterson, and G. Brooks. 1978. Serum bactericidal
action and activation of the classic and alternate complement
pathways by Neisseria ganarrhaeae. J. Lab. Clin. Med. 92:211-220.

Insel, R. A., P. 2. Allen, and I. D. Berkowitz. 1982. Types and
frequency of Haemophilus influenzae IgAl proteases, p. 225-231. In
L. Weinstein and B. N. Fields (ed.), Seminars in infectious
diseases, vol. 4. Grune and Stratton, New York, NY.

James, J. F., E. Zurlinden, C. J. Lammel, and G. F. Brooks. 1982.
Relation of protein I and colony opacity to serum killing of
Neisseria ganarrhaeae. J. Infect. Dis. 145:37-43.

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

85

Jarvis, G. A., H. Schneider, M. K. Alberson, and J. M. Griffiss.
1988. Strain-specific direct binding of properdin accounts for
variable lysis of Neisseria gonorrhaeae, p. 717-722. In J. T.
Poolman, et al. (ed.), Gonococci and meningococci. Kluwer Academic
Publishers, Dordrecht, the Netherlands.

Jephcott, A. E., A. Reyn, A. Birch-Andersen. 1971. Neisseria
ganarrhaeae. III. Demonstration of presumed appendages to cells
from different colony types. Acta Pathol. Microbiol. Scand. Sect. B
79:437-439.

Johnson, A. P., D. Taylor-Robinson, and z. A. McGee. 1977. Species
specificity of attachment and damage to oviduct mucosa by Neisseria
ganarrhaeae. Infect. Immun. 18:833-839.

Joiner, K. A., R. Scales, K. A. Warren, M. M. Frank, and P. A.
Rice. 1985. Mechanism of action of blocking immunoglobulin G for
Neisseria ganarrhaeae. J. Clin. Invest. 76:1765-1772.

Joiner, K. A., K. A. Warren, G. Hammer, and M. M. Frank. 1985.
Bactericidal but not bactericidal C5b-9 is associated with
distinctive outer membrane proteins in Neisseria ganarrhaeae. J.
Immunol. 134:1920-1925.

Joiner, K. A., K. A. Warren, M. Tam, and M. M. Frank. 1985.
Monoclonal antibodies directed against gonococcal protein I vary in
bactericidal activity. J. Immunol. 134:3411-3419.

Jones, F., E. J. Cunningham, T. E. Shockley, and J. H. Jackson.
1985. Genetic analysis of spontaneous resistance to ampicillin in
Neisseria ganarrhaeae. Antimicrob. Agents Chemother. 28:21-27.
Judd, R. C. 1982. 125I-peptide mapping of protein 111 isolated from
four strains of Neisseria ganarrhaeae. Infect. Immun. 37:622-631.

Judd, R. C. 1989. Protein 1: Structure, function, and genetics.
Clin. Microbiol. Rev. 2:S4l-S48.

Kapatais-Zoumbos, K., D. K. F. Chandler, and M. F. Barile. 1985.
Survey of immunoglobulin A protease activity among selected species
of Ureaplasma and Mycaplasma: specificity for host immunoglobulin
A. Infect. Immun. 47:704-709.

Kawula, T. H., S. M. Spinola, D. G. Klapper, and J. G. Cannon.
1987. Localization of a conserved epitope and an azurin-like domain
in the H.8 protein of pathogenic Neisseria. Mol. Microbiol. 1:179-
185.

Keevil, C. W. 1988. Characteristics of gonococci grown in
continuous culture, abstr. MT28, p. 168. 6th Int. Pathog. Neisseria
Conf., Pine Mountain, GA.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

86

Keevil, C. W., D. B. Davies, B. J. Spillane, and E. Mahenthir-
alingham. 1988. Influence of iron-limited and excess continuous
culture on the virulence properties of Neisseria ganarrhaeae. In J.
T. Poolman, et al. (ed.), Gonococci and meningococci. Kluwer
Academic Publishers, Dordrecht, the Netherlands.

Kellogg, D. 8., I. R. Cohen, L. C. Norins, A. L. Schroeter, and G.
Reising. 1968. Neisseria ganarrhaeae. II. colonial variation and
pathogenicity during 35 months in vitro. J. Bacteriol. 96:596-605.

Kellogg. D. 8., J. A. Crawford, and C. 8. Callaway. 1983.
Cultivation of Neisseria ganarrhaeae under low oxygen conditions.
J. Clin. Microbiol. 18:178-184.

Kellogg, D. 8., Jr., W. L. Peacock, W. E. Deacon, L. Brown, and c.
I. Pirkle. 1963. Neisseria ganarrhaeae. I. Virulence genetically
linked to clonal variation. J. Bacteriol. 85:1274-1279.

Kilian, M. 1981. Degradation of immunoglobulins A1, A2, and G by
suspected principal periodontal pathogens. Infect. Immun. 34:757-
765.

Kilian, M., M. B. Brown, T. A. Brown, E. A. Freundt, and G. H.
Cassell. 1984. Immunoglobulin Al protease activity in strains of
Ureaplasma urealyticum. Acta. Path. Microbiol. Immunol. Scand.
Sect. B. 92:61-64.

Kilian, M., and K. Holmgren. 1981. Ecology and nature of immuno-
globulin Al protease-producing streptococci in the human oral
cavity and pharynx. Infect. Immun. 31:868-873.

Kilian, M., J. Mestecky, R. Kulhavy, M. Tomana, and W. T. Butler.
1980. IgAl proteases from Haemophilus influenzae, Streptococcus
pneumoniae, Neisseria ganarrhaeae, and Streptococcus sanguis: comp-
arative immunochemical studies. J. Immunol. 124:2596-2600.

Kilian, M., J. Mestecky, and M. W. Russell. 1988. Defense mechan-
isms involving Fc-dependent functions of immunoglobulin A and their
subversion by bacterial immunoglobulin proteases. Microbiol. Rev.

52:296-303.

Kilian, M., J. Mestecky, and R. E. Schrohenloher. 1979. Pathogenic
species of the genus Haemophilus and Streptococcus pneumoniae
produce immunoglobulin A1 protease. Infect. Immun. 26:143-149.

Kilian, M., and J. Reinholdt. 1987. A hypothetical model for the
development of invasive infection due to IgAl protease-producing
bacteria. Adv. Exp. Med. Biol. 2168:1261-1269.

Kilian, M., K. Roland, and J. Mestecky. 1981. Interference of
secretory immunoglobulin A with sorption of oral bacteria to
hydroxyapatite. Infect. Immun. 31:942-951.

122.

123.

124.

125.

126.

127.

128.

129.

130.

131.

132.

133.

87

Kilian, M., B. Thomson, T. E. Petersen, and H. Bleeg. 1983.
Molecular biology of Haemophilus influenzae IgAl proteases. Mol.
Immunol. 20:1051-1058.

Kilian, M., B. Thomsen, T. E. Petersen, and H. 8. Bleeg. 1983.
Occurrence and nature of bacterial IgA proteases. Ann. N. Y. Acad.
Sci. 409:612-623.

Klauser, T., J. Pohlner, and T. F. Meyer. 1990. Extracellular
transport of cholera toxin B subunit using Neisseria IgA protease
fi-domain: conformation-dependent outer membrane translocation. EMBO

J. 9:1991-1999.

Klimpel, K. W., and V. L. Clark. 1988. High resolution two-
dimensional analysis of the proteins of Neisseria ganarrhaeae
colony type 1 and 4, heat shocked and non-heat shocked, p. 393-399.
In J. T. Poolman, et al. (ed.), Gonococci and meningococci. Kluwer
Academic Publishers, Dordrecht, the Netherlands.

Knapp, J. 8., and V. L. Clark. 1984. Anaerobic growth of Neisseria
gonarrhaeae coupled to nitrite reduction. Infect. Immun. 46:176-
181.

Knapp, J. 8., and K. K. Holmes. 1975. Disseminated gonococcal
infections caused by Neisseria ganarrhaeae with unique nutritional
requirements. J. Infect. Dis. 132:204-208.

Knapp, J. 8., J. M. Zenilman, J. W. Biddle, G. H. Perkins, W. E.
DeWitt, M. L. Thomas, 8. R. Johnson, and 8. A. Morse. 1987.
Frequency and distribution in the United States of strains of
Neisseria gonorrhoeae with plasmid-meditated, high-level resistance
to tetracycline. J. Infect. Dis. 155:819-822.

Koomey, J. M., and S. Falkow. 1984. Nucleotide sequence homology
between the immunoglobulin Al protease genes of Neisseria
ganarrhaeae, Neisseria meningitidis, and Haemophilus influenzae.
Infect. Immun. 43:101-107.

Koomey, J. M., and S. Falkow. 1987. Cloning of the recA gene of
Neisseria gonarrhaeae and construction of gonococcal recA mutants.
J. Bacteriol. 169:790-795.

Koomey, J. M., R. E. Gill, and 8. Falkow. 1982. Genetic and
biochemical analysis of gonococcal IgAl protease: cloning in
Escherichia cali and construction of mutants of gonococci that fail
to produce the activity. Proc. Natl. Acad. Sci. USA 79:7881-7885.

Koomey, J. M., E. C. Gotschlich, K. Robbins, S. Bergstrom, and J.
Swanson. 1987. Effects of recA mutations on pilus antigenic and
phase transitions in Neisseria ganarrhaeae. Genetics 117:391-398.

Korch, G., P. Hagblom, H. Ohman, M. Goransson, and S. Normark.
1985. Cryptic plasmid of Neisseria gonarrhoeae: complete nucleotide
sequence and genetic organization. J. Bacteriol. 163:430-438.

134.

135.

136.

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

88

Kornfeld, 8. J., and A. G. Plant. 1981. Secretary immunity and the
bacterial IgA proteases. Rev. Infect. Dis. 3:521-543.

Labib, R. 8., N. J. Calvanico, and T. B. Tomasi, Jr. 1978. Studies
an extracellular proteases of Streptococcus sanguis: purification
and characterization of a human IgAl specific protease. Biochim.
Biophys. Acta. 526:547-559.

Lambden, P. R., J. E. Heckels, L. T. James, and P. J. Watts. 1979.
Variations in surface protein composition associated with virulence
properties in opacity types of Neisseria ganarrhaeae. J. Gen.

Microbiol. 114:305-312.

Lammel, C. J., R. L. Sweet, P. A. Rice, J. 8. Knapp, G. K.
Schoolnik, D. C. Heilbron, and G. F. Brooks. 1985. Antibody-antigen
specificity in the immune response to infection with Neisseria
ganarrhaeae. J. Infect. Dis. 152:990-1001.

Lankford, C. E. 1973. Bacterial assimilation of iron. CRC Crit.
Rev. Microbiol. 2:273-331.

Lassiter, M. 0., J. C. Kindle, L. C. Hobbs, and R. L Gregory. 1989.
Estimation of immunoglobulin protease activity by quantitative
rocket immunoelectrophoresis. J. Immunol. Methods 123:63-69.

Lee, B. G., and A. B. Schryvers. 1988. Specificity of the
lactoferrin and transferrin receptors in Neisseria ganarrhaeae.
Mol. Microbiol. 2:827-829.

Lindahl, L., C. Schalen, and P. Christensen. 1981. New method that
uses binding of immunoglobulin A to group A streptococcal immuno-
globulin A Fc receptors for demonstration of microbial immunoglob-
ulin A protease activity. J. Clin. Microbiol. 13:991-993.

Lindler, L. E., and F. J. Stutzenberger. 1983. Quantitative
screening of clinical isolates for immunoglobulin A protease
production. J. Clin. Microbiol. 18:365-371.

Lindler, L. E., and F. J. Stutzenberger. 1983. Inactivation and
stabilization of IgA protease from the human oral bacterium
Streptococcus sanguis. Arch. Oral Biol. 28:977-979.

Liu, Y.-8. V., T. L. K. Law, A. Infants, and F. W. Putnam. 1976.
Complete covalent structure of a human IgAl immunoglobulin. Science
193:1017-1020.

Lucas, C. T., F. Chandler, Jr., J. E. Martin, Jr., and J. D.
Schmale. 1971. Transfer of gonococcal urethritis from man to
chimpanzee. J. Am. Med. Assoc. 216:1612-1614.

Lund, B., F. Lindberg, and 8. Normark. 1988. Structure and
antigenic properties of the tip-located pilus proteins of
uropathogenic Escherichia coli. J. Bacteriol. 170:1887-1894.

147.

148.

149.

150.

151.

152.

153.

154.

155.

156.

157.

89

Lynch, E. C., M. 8. Blake, E. C. Gotschlich, and A. Mauro. 1984.
Studies on porins: spontaneously transferred from whole cells and
from proteins of Neisseria ganarrhaeae and Neisseria meningitidis.
Biophys. J. 45:104-107.

Lytton, E. J., and M. 8. Blake. 1986. Isolation and partial
characterization of the reduction-modifiable protein of Neisseria
ganarrhaeae. J. Exp. Med. 164:1749-1759.

Meier, T. W., L. Zubrzycki, and M. B. Coyle. 1975. Genetic anal-
yses of drug resistance in Neisseria ganarrhaeae: identification
and linkage relationships of loci controlling drug resistance.
Antimicrob. Agents Chemother. 7:676-681.

Male, C. J. 1979. Immunoglobulin Al protease production by
Haemophilus influenzae and Streptococcus pneumoniae. Infect. Immun.
26:254-261.

Manchee, R. J., C. N. Wiblin, and A. Robinson. 1980. Growth of
Neisseria gonarrhaeae in continuous culture. FEMS Microbiol. Lett.
7:115-118.

Mandrell, R., M. Apicella. J. Boslego, R. Chung, P. Rice, and J. M.
Griffiss. 1988. Human immune response to monoclonal antibody-
defined epitopes of Neisseria ganarrhaeae lipooligosaccharides, p.
569-574. In J. T. Poolman, et al. (ed.), Gonococci and meningo-
cocci. Kluwer Academic Publishers, Dordrecht, the Netherlands.

Mandrell, R. E., J. M. Griffiss, and B. Macher. 1988. Lipooligosac-
charides (L08) of Neisseria ganarrhaeae and Neisseria meningitidis
have components that are immunologically similar to precursors of
human blood group antigens. J. Exp. Med. 168:107-126.

Mandrell, R. E., A. J. Lease, J. V. Sugai, M. Shero, J. M.
Criffiss, J. A. Cole, N. J. Parsons, H. Smith, 8. A. Morse, and M.
A. Apicella. 1990. In vitro and in viva modification of Neisseria
ganarrhaeae lipooligosaccharide epitope structure by sialylation.
J. Exp. Med. 171:1649-1664.

Mandrell, R., H. Schneider, M. Apicella. W. Zollinger, P. A. Rice,
and J. M. Griffiss. 1986. Antigenic and physical diversity of
Neisseria ganarrhaeae lipooligosaccharides. Infect. Immun. 54:63-
69.

Manse, 8., and M. Kilian. 1986. Retained antigen-binding activity
of Fab fragments of human monoclonal immunoglobulin A1 (IgAl)
cleavea by IgAl protease. Infect. Immun. 52:171-174.

Martin, P. M. V., P. V. Patel, N. J. Parsons, and H. Smith. 1981.
Induction of phenotypically determined resistance of Neisseria

gonarrhoeae to human serum by factors in human serum. J. Gen.
Microbiol. 127:213-217.

158.

159.

160.

161.

162.

163.

164.

165.

166.

167.

168.

169.

170.

171.

90

Martin, P. M. V., P. V. Patel, N. J. Parsons, and H. Smith. 1982.
Induction in gonococci of phenotypic resistance to killing by human
serum by human genital secretions. Br. J. Vener. Dis. 58:363-365.

Martin, P. R., J. W. Cooperider, and M. H. Mulks. 1990. Sequence of
the argF gene encoding ornithine transcarbamoylase from Neisseria
ganarrhaeae. Gene 94:139-140.

Mauro, A., M. 8. Blake, and P. Labarca. 1988. Voltage gating of
conductance in lipid bilayers induced by porins from outer membrane
of Neisseria ganarrhaeae. Proc. Natl. Acad. Sci. USA 85:1071-1075.

McDade, R. L., Jr., and K. H. Johnston. 1980. Characterization of
serologically dominant outer membrane proteins of Neisseria
ganarrhaeae. J. Bacteriol. 141:1183-1191.

McGee, z. A. 1984. Gonococcal pelvic inflammatory disease, p. 220-
229. In K. K. Holmes, et al. (ed.), Sexually transmitted diseases.
McGraw-Hill Book Co., New York, NY.

McGee, z. A., C. R. Gregg. A. P. Johnson, 8. S. Kalter, and D.
Taylor-Robinson. 1990. The evolutionary watershed of susceptibility
to gonococcal infection. Microb. Pathog. 9:131-139.

McKenna, W. R., P. A. Mickelsen, P. F. Sparling, and D. W. Dyer.
1988. Iron uptake from lactoferrin and transferrin by Neisseria
gonorrhoeae. Infect. Immun. 56:785-791.

McNabb, P. C., and T. B. Tomasi. 1981. Host defense mechanisms at
mucosal surfaces. Annu. Rev. Microbiol. 35:477-496.

Mehta, S. K., A. G. Plaut, N. J. Calvanico, and T. B. Tomasi, Jr.
1973. Human immunoglobulin A: production of an Fc fragment by an
enteric microbial proteolytic enzyme. J. Immunol. 111:1274-1276.

Melly, M. A., Z. A. McGee, and R. 8. Rosenthal. 1984. Ability of
monomeric peptidoglycan fragments from Neisseria ganarrhaeae to
damage human fallopian tube mucosa. J. Infect. Dis. 149:378-386.

Mestecky, J., and M. Kilian. 1985. Immunoglobulin A (IgA). Methods
Enzymol. 116:37-75.

Mestecky, J., J. R. McGhee, 8. S. Crago, S. Jackson, M. Kilian, H.
Kiyono, J. L. Babb, and 8. M. Michalek. 1980. Molecular-cellular
interactions in the secretory IgA response. J. Reticuloendothel.
Soc. 28:458-608.

Mestecky, J., and M. W. Russell. 1986. IgA subclasses. Monogr.
Allergy 19:277-301.

Meyer, T. F., and J. P. M. van Putten. 1989. Genetic mechanisms and
biological implications of phase variation in pathogenic Neisser-
iae. Clin. Microbiol. Rev. 2:8139-8145.

172.

173.

174.

175.

176.

177.

178.

179.

180.

181.

182.

183.

184.

91

Mickelsen, P. A., E. Blackman, and P. F. Sparling. 1981. Ability of
Neisseria gonarrhaeae, Neisseria meningitidis, and commensal
Neisseria species to obtain iron from lactoferrin. Infect. Immun.
35:915-920.

Mickelsen, P. A., and P. F. Sparling. 1981. Ability of Neisseria
ganarrhaeae, Neisseria meningitidis, and commensal Neisseria
species to obtain iron from transferrin and iron compounds. Infect.
Immun. 33:555-564.

Mietzner, T. A., G. H. Luginbuhl, E. Sandstrom, and S. A. Morse.
1984. Identification of an iron-regulated 37,000 dalton protein in
the cell envelope of Neisseria ganarrhaeae. Infect. Immun. 45:410-
416.

Milazzo, F. H., and G. J. Delisle. 1984. Immunoglobulin A proteases
in gram-negative bacteria isolated from human urinary tract
infections. Infect. Immun. 43:11-13.

Miller, C. P. 1948. Experimental gonococcal infection of the
rabbit's eye. Am. J. Vener. Dis. 32:437-444.

Miller, C. P., and W. D. Hawk. 1939. Experimental infection of mice
with gonococcus. Arch. Pathol. 28:654.

Mills, J., and G. F. Brooks. 1984. Disseminated gonococcal
infection, p. 229-237. In K. K. Holmes, et al. (ed.), Sexually
transmitted diseases. McGraw-Hill Book Co., New York, NY.

Mbrello, J. A., 8. A. Lerner, and M. Bohnhoff. 1976. Character-
istics of atypical Neisseria ganarrhaeae from disseminated and
localized infections. Infect. Immun. 13:1510-1516.

Horse, 8. A. 1986. Neisseria and Branhamella, p. 278-287. In A. H.
Samiy, et al. (ed.), Infectious diseases and medical microbiology,
2nd. ed. W. B. Saunders Co., Philadelphia, PA.

Morse, 8. A., and B. H. Hebeler. 1978. Effect of pH on the growth
and glucose metabolism of Neisseria ganarrhaeae. Infect. Immun.
21:87-95.

Morse, 8. A., 8. R. Johnson, J. W. Biddle, and M. C. Roberts. 1986.
High-level tetracycline resistance in Neisseria ganarrhaeae is the
result of acquisition of a streptococcal tetM determinant.
Antimicrob. Agents Chemother. 30:664-670.

Morse, S. A., R. D. Miller, and B. H. Hebeler. 1977. Physiology and
metabolism of Neisseria gonarrhoeae, p. 213-253. In R. B. Roberts
(ed.), The gonococcus. John Wiley and Sons, New York, NY.

Mbrtensen, 8. B., and M. Kilian. 1984. Purification and character-
ization of an immunoglobulin Al protease from Bacteraides
melaninagenicus. Infect. Immun. 45:550-557.

185.

186.

187.

188.

189.

190.

191.

192.

193.

194.

195.

196.

197.

92

Mbrtensen, S. B., and M. Kilian. 1984. A rapid method for the
detection and quantitation of IgA protease activity by macrobore
gel-permeation chromatography. J. Chromatogr. 296:257-262.

Muir, L. L., R. A. Strugnell, and J. K. Davies. 1988. Proteins that
appear to be associated with pili in Neisseria ganarrhaeae. Infect.
Immun. 56:1743-1747.

Mulks, M. H. 1985. Microbial IgA proteases, p. 81-104. In I. A.
Holder (ed.), Bacterial enzymes and virulence. CRC Press, Inc.,
Boca Raton, FL.

Mulks, M. H., and J. S. Knapp. 1983. Distribution of IgAl protease
types in Neisseria gonarrhaeae, abstr. 12, p. 44. Program Abstr.
5th Int. Meet. Int. Soc. Sex. Transm. Dis. Res. 1983.

Mulks, M. H., and J. S. Knapp. 1985. Immunoglobulin A1 protease
types of Neisseria ganarrhaeae, p. 51-56. In G. K. Schoolnik, et
al. (ed.), The pathogenic Neisseriae. Am. Soc. Microbiol.,
Washington, DC.

Mulks, M. H., and J. S. Knapp. 1987. Immunoglobulin A1 protease
types of Neisseria ganarrhaeae and their relationship to auxotype
and serovar. Infect. Immun. 55:931-936.

Mulks, M. H., 8. J. Kornfeld, B. Frangione, and A. G. Plaut. 1982.
Relationship between the specificity of IgA proteases and serotypes
in Haemophilus influenzae. J. Infect. Dis. 146:266-274.

Mulks, M. H., 8. J. Kornfeld, and A. G. Plaut. 1980. Specific
proteolysis of human IgA by Streptococcus pneumoniae and
Haemophilus influenzae. J. Infect. Dis. 141:450-456.

Mulks, M. H., and A. G. Plaut. 1978. IgA protease production as a
characteristic distinguishing pathogenic from harmless Neisseri-
aceae. New Eng. J. Med. 299:973-976.

Mulks, M. H., A. G. Plaut, H. A. Feldman, and B. Frangione. 1980.
IgA proteases of two distinct specificities are released by
Neisseria meningitidis. J. Exp. Med. 152:1442-1447.

Mulks, M. H., A. G. Plaut, and M. Lamm. 1980. Gonococcal IgA
protease reduces inhibition of bacterial adherence by human
secretory IgA, p. 217-220. In S. Normark and D. Danielsson (ed.),
Genetics and immunobiology of pathogenic Neisseria. Proc. EMBO
Workshop, Hemavan, Sweden.

Mulks, M. H., D. A. Simpson, and R. J. Shoberg. 1987. Restriction
site polymorphism in genes encoding type 2 but not type 1 gonococ-
cal IgAl proteases. Antonie van Leeuwenhoek J. Microbiol. 53:471-
478.

Neilands, J. B. 1981. Microbial iron compounds. Annu. Rev. Biochem.
50:715-731.

198.

199.

200.

201.

202.

203.

204.

205.

206.

207.

208.

209.

210.

211.

93

Newhall, W. J., V, W. D. Sawyer, and R. A. Haak. 1980. Cross-
linking analysis of the outer membrane proteins of Neisseria
ganarrhaeae. Infect. Immun. 28:785-791.

Norlander, L., J. Davies, A. Norqvist, and S. Normark. 1981.
Deoxyribonucleic acid modifications and restriction endonuclease
production in Neisseria ganarrhaeae. J. Bacteriol. 145:788-795.

Norqvist, A., J. Davies, L. Norlander, and S. Normark. 1978. The
effect of iron starvation on the outer membrane protein composition
of Neisseria ganarrhaeae. FEMS Microbiol. Lett. 4:71-75.

Norrod, E. P., and S. A. Morse. 1979. Absence of superoxide
dismutase in some strains of Neisseria ganarrhaeae. Biochem.
Biophys. Res. Comm. 90:1287-1294.

Norrod, P., and R. P. Williams. 1978. Growth of Neisseria
ganarrhaeae in media deficient in iron without detection of
siderophores. Curr. Microbial. 1:281-284.

O'Reilly, T. M., and A. R. Bhatti. 1986. Proteases of the patho-
genic Neisseriae: possible role in infection. Microbios 45:113-129.

Parsons, N. J., P. V. Patel, E. L. Tan, J. R. C. Andrade, C. A.
Nairn, M. Goldner, J. A. Cole, and H. Smith. 1988. Cytidine 5'-
monophospho-N-acetyl neuraminic acid and a low molecular weight
factor from human blood cells induce lipopolysaccharide alteration
in gonococci when conferring resisance to killing by human serum.
Microb. Pathog. 5:303-309.

Payne, 8. M. 1989. Iron and virulence in Shigella. Mol. Microbiol.
3:1301-1306.

Payne, 8. M., and R. A. Finkelstein. 1977. Detection and differ-
entiation of iron-responsive avirulent mutants on Congo red agar.
Infect. Immun. 18:94-98.

Petersen, B. H., T. J. Lee, R. Snyderman, and C. F. Brooks. 1979.
Neisseria meningitidis and Neisseria ganarrhaeae bacteremia asso-
ciated with C6, C7, or C8 deficiency. Ann. Intern. Med. 90:917-920.

Petersen, B. H., and R. S. Rosenthal. 1982. Complement consumption
by gonococcal peptidoglycan. Infect. Immun. 35:442-448.

Pettit, R. K., J. C. Scuba, and R. C. Judd. 1990. Characterization
of fourteen strains of Neisseria ganarrhaeae: structural analyses
and serum reactivities. Mol. Microbiol. 4:1293-1301.

Plaut, A. G. 1983. The IgAl proteases of pathogenic bacteria. Annu.
Rev. Microbiol. 37:603-622.

Plaut, A. G., R. J. Genco, and T. B. Tomasi, Jr. 1974. Isolation of
an enzyme from Streptococcus sanguis which specifically cleaves
IgA. J. Immunol. 113:289-291.

212.

213.

214.

215.

216.

217.

218.

219.

220.

221.

222.

223.

94

Plaut, A. G., R. J. Genco, and T. B. Tomasi, Jr. 1974. Production
of an Fc fragment from human immunoglobulin A by an IgA-specific
bacterial protease. Adv. Exp. Med. Biol. 45:245-249.

Plaut, A. G., J. V. Gilbert, M. S. Artenstein, and J. D. Capra.
1975. Neisseria gonarrhoeae and Neisseria meningitidis: extra-
cellular enzyme cleaves human immunoglobulin A. Science 190:1103-

1105.

Plaut, A. G., J. V. Gilbert, and I. Heller. 1978. Assay and
properties of IgA protease of Streptococcus sanguis. Adv. Exp. Med.
Biol. 107:489-495.

Plaut, A. G., J. V. Gilbert, and R. Wistar, Jr. 1977. Loss of
antibody activity in human immunoglobulin A exposed to extracel-
lular immunoglobulin A proteases of Neisseria gonarrhoeae and
Streptococcus sanguis. Infect. Immun. 17:130-135.

Plaut, A. G., R. Wistar, Jr., and J. D. Capra. 1974. Differential
susceptibility of human IgA immunoglobulins to streptococcal IgA
protease. J. Clin. Invest. 54:1295-1300.

Podack, E. R., K. T. Preissner, and H. J. Muller-Eberhard. 1984.
Inhibition of C9 polymerization within the SCSb-9 complex of
complement by S-protein. Acta. Pathol. Microbiol. 284:92z89-96.

Pohlner, J., R. Halter, K. Beyreuther, and T. F. Meyer. 1987. Gene
structure and extracellular secretion of Neisseria ganarrhaeae IgA
protease. Nature (London) 325:458-462.

Pohlner, J., R. Halter, and T. F. Meyer. 1987. Neisseria
ganarrhaeae IgA protease. Secretion and implications for
pathogenesis. Antonie van Leeuwenhoek J. Microbiol. 53:479-484.

Poulsen, K., J. Brandt, J. P. Hjorth, H. C. Thfgersen, and M.
Kilian. 1989. Cloning and sequencing of the immunoglobulin A1
protease gene (iga) of Haemophilus influenzae serotype b. Infect.
Immun. 57:3097-3105.

Poulsen, K., J. P. Hjorth, and M. Kilian. 1988. Limited diversity
of the immunoglobulin A1 protease gene (iga) among Haemophilus
influenzae serotype b strains. Infect. Immun. 56:987-992.

Rahr, 8., R. Halter, H. Muller, J. Pohlner, and T. F. Meyer. 1985.
Genetic analysis of neisserial immunoglobulin A proteases, p. 157-
163. In G. K. Schoolnik, et al. (ed.), The pathogenic Neisseriae.
Am. Soc. Microbiol., Washington, DC.

Reinholdt, J., and M. Kilian. 1983. A sensitive enzyme-linked
immunosorbent assay for IgA protease activity. J. Immunol. Methods
63:367-376.

224.

225.

226.

227.

228.

229.

230.

231.

232.

233.

234.

235.

236.

237.

95

Reinholdt, J., and M. Kilian. 1987. Interference of IgA protease
with the effect of secretory IgA on adherence of oral streptococci
to saliva-coated hydroxyapatite. J. Dent. Res. 66:492-497.

Reinholdt, J., M. Tomana, S. B. Mortensen, and M. Kilian. 1990.
Molecular aspects of immunoglobulin A1 degradation by oral
streptococci. Infect. Immun. 58:1186-1194.

Rest, R. F., and W. M. Shafer. 1989. Interactions of Neisseria
gonorrhoeae with human neutrophils. Clin. Microbiol. Rev. 2:883-
891.

Rice, P. A. 1989. Molecular basis for serum reistance in Neisseria
ganarrhaeae. Clin. Microbiol. Rev. 2:8112-8117.

Rice, P. A., H. E. Vayo, M. R. Tam, and M. 8. Blake. 1986. Immuno-
globulin G antibodies directed against protein 111 block killing of
serum-resistant Neisseria gonorrhaeae by immune serum. J. Exp. Med.
164:1735-1748.

Richardson, W. P., and J. C. Sadoff. 1988. Induced engulfment of
Neisseria ganarrhaeae by tissue culture cells. Infect. Immun.
56:2512-2514.

Rittenberg, 8. C., C. W. Penn, N. J. Parsons, D. R. Veale, and J.
Smith. 1977. Phenotypic changes in the resistance of Neisseria
gonarrhoeae to killing by normal human serum. J. Gen. Microbiol.
103:69-75.

Roberts, M. C. 1989. Plasmids of Neisseria ganarrhaeae and other
Neisseria species. Clin. Microbiol. Rev. 2:818-823.

Roberts, M., L. P. Elwell, and S. Falkow. 1977. Molecular charact-
erization of two beta-lactamase specifying plasmids isolated from
Neisseria ganarrhaeae. J. Bacteriol. 131:557-563.

Roberts, M. C., and J. S. Knapp. 1988. Host range of the
conjugative 25.2-megadalton tetracycline resistance plasmid from
Neisseria ganarrhaeae. Antimicrob. Agents Chemother. 32:1430-1432.

Roberts, M., P. Plot, and S. Falkow. 1979. The ecology of
gonococcal plasmids. J. Gen. Microbiol. 114:491-494.

Robertson, J. A., M. E. Stemler, and G. W. Stemke. 1984. Immuno-
globulin A protease activity of Ureaplasma urealyticum. J. Clin.
Microbiol. 19:255-258.

Robertson, J. N., P. Vincent, and M. E. Ward. 1977. The preparation
and properties of gonococcal pili. J. Gen. Microbiol. 102:169-177.

Rosenthal. R. 8., and T. J. Fleming. 1988. Gonococcal peptidoglycan
(PG) and inflammation, abstr. MT4, p. 144. 6th Int. Pathog.
Neisseria Conf., Pine Mountain, GA.

238.

239.

240.

241.

242.

243.

244.

245.

246.

247.

248.

249.

250.

96

Rosenthal, R. 8., and J. M. Krueger. 1988. Promotion of sleep by
gonococcal peptidoglycan fragments. Structured requirements for the
somnogenic activity, p. 581-590. In J. T. Poolman, et al. (ed.),
Gonococci and meningococci. Kluwer Academic Publishers, Dordrecht,
the Netherlands.

Ruoslahti, E., and M. D. Pierschbacher. 1987. New perspectives in
cell adhesion: RGD and integrins. Science 238:491-497.

Sarafian, S. K., M. R. Tam, and S. A. Morse. 1983. Gonococcal
protein I-specific opsonic IgG in normal human serum. J. Infect.
Dis. 148:1025-1032.

Schneider, H., J. M. Griffiss. R. E. Mandrell, and G. A. Jarvis.
1985. Elaboration of a 3.6-kilodalton lipooligosaccharide, antibody
against which is absent from human sera, is associated with serum
resistance of Neisseria ganarrhaeae. Infect. Immun. 50:672-677.

Schneider, H., T. L. Hale, W. D. Zollinger, R. C. Seid, Jr., C. A.
Hammack, and J. M. Griffiss. 1984. Heterogeneity of molecular size
and antigenic expression within lipooligosaccharides of individual
strains of Neisseria ganarrhaeae and Neisseria meningitidis.
Infect. Immun. 45:544-549.

Schoolnik, G. K., T. M. Buchanan, and K. K. Holmes. 1976. Gonococci
causing disseminated gonococcal infection are resistant to the
bactericidal action of normal serum. J. Clin. Invest. 58:1163-1173.

Schryvers, A. B., and L. J. Morris. 1988. Identification and
characterization of the transferrin receptor from Neisseria
meningitidis. Mol. Microbiol. 2:281-288.

Schryvers, A. B., and L. J. Morris. 1988. Identification and
characterization of the lactoferrin binding protein of Neisseria
meningitidis. Infect. Immun. 56:1144-1149.

Schwalbe, R. 8., P. F. Sparling, and J. G. Cannon. 1985. Variation
of Neisseria ganarrhaeae protein II among isolates from an outbreak
caused by a single gonococcal strain. Infect. Immun. 49:250-252.

Segal, E., E. Billyard, M. So, 8. Stbrzbach, and T. F. Meyer. 1985.
Role of chromosomal rearrangements in N. gonorrhoeae pilus phase
variation. Cell 40:293-300.

Senda, 8., Y. Fujiyama, and S. Hosoda. 1987. IgA protease-producing
bacteria in patients with ulcerative colitis. Gut 28:917.

Senior, B. W., M. Albrechtesen, and M. A. Kerr. 1987. Proteus
mirablis strains of diverse type have IgA protease activity. J.
Med. Microbiol. 24:175-180.

Senior, B. W., M. Albrechtesen, and M. A. Kerr. 1988. A survey of
IgA protease production among clinical isolates of Prateeae. J.
Med. Microbiol. 25:27-31.

251.

252.

253.

254.

255.

256.

257.

258.

259.

260.

261.

262.

263.

264.

97

Shaw, J. H., and S. Falkow. 1988. Model for invasion of human
tissue culture cells by Neisseria ganarrhaeae. Infect. Immun.
56:1625-1632.

Shoberg, R. J., and M. H. Mulks. 1991. Proteolysis of bacterial
membrane proteins by Neisseria gonarrhoeae type 2 immunoglobulin Al
protease. Infect. Immun. in press.

Simpson, D. A., R. P. Hausinger, and M. H. Mulks. 1988. Purif-
ication, characterization, and comparison of the immunoglobulin Al
proteases of Neisseria ganarrhaeae. J. Bacteriol. 170:1866-1873.

Sorenson, C. H., and M. Kilian. 1984. Bacterium-induced cleavage of
IgA in nasopharyngeal secretions from atopic children. Acta.
Pathol. Microbiol. Immunol. Scand. Sect. C. 92:85-87.

Sparling, P. F. 1966 Genetic transformation of Neisseria gonor-
rhaeae to streptomycin resistance. J. Bacteriol. 124:1364-1371.

Sparling, P. F. 1977. Antibiotic resistance in the gonococcus, p.
111-136. In R. B. Roberts (ed.), The gonococcus. John Wiley and
Sons, Inc., New York, NY.

Sparling, P. F., F. A. Sarubbi, Jr., and E. Blackman. 1975.
Inheritance of low-level resistance to penicillin, tetracycline,
and chloramphenical in Neisseria ganarrhaeae. J. Bacteriol.

127:740-749.

Stead, A., J. S. Main, M. E. Ward, and P. J. Watt. 1975. Studies on
lipopolysaccharides isolated from strains of Neisseria ganarrhaeae.
J. Gen. Microbiol. 88:123-131.

Stein, D. C., S. Gregoire, and A. Piekarowicz. 1988. Restriction of
plasmid DNA during transformation but not conjugation in Neisseria
ganarrhaeae. Infect. Immun. 56:112-116.

Stephens, D. S. 1989. Gonococcal and meningococcal pathogenesis as
defined by human cell, cell culture, and organ culture assays.
Clin. Microbiol. Rev. 2:8104-8111.

Stern, A., M. Brown, P. Nickel, and T. F. Meyer. 1986. Opacity
genes in Neisseria gonarrhoeae: control of phase and antigenic
variation. Cell 47:61-71.

Stern, A., P. Nickel, T. F. Meyer, and M. So. 1984. Opacity
determinants of Neisseria ganarrhaeae: Gene expression and
chromosomal linkage to the gonococcal pilin gene. Cell 37:447-456.

Strittmatter, W., and P. J. Hitchcock. 1986. Isolation and
preliminary biochemical characterization of thet gonococcal H.8
antigen. J. Exp. Med. 164:2038-2048.

Summary of notifiable diseases. 1988. Morbid. Mortal. Weekly Rep.
37:3.

265.

266.

267.

268.

269.

270.

271.

272.

273.

274.

275.

276.

98

Swanson, J. 1977. Surface components affecting interactions between
Neisseria ganarrhaeae and eucaryotic cells. J. Infect. Dis.
136:8138-8143.

Swanson, J. 1978. Studies on gonococcus infection. XIV. Cell wall
protein differences among color/opacity colony variants of
Neisseria ganarrhaeae. Infect. Immun. 21:292-302.

Swanson, J. 1980. Adhesion and entry of bacteria into cells: a
model of the pathogenesis of gonorrhea, p. 17-39. In H. Smith, et
al. (ed.), The molecular basis of microbial pathogenicity.
Springer-Verlag KG, Berlin. 107.

Swanson, J., and J. Heckels. 1980. Proposal: nomenclature of
gonococcal outer membrane proteins, p. xxi-xxvi. In S. Normark and
D. Danielsson (ed.), Genetics and immunobiology of pathogenic
Neisseria. Proc. EMBO Workshop, Hemavan, Sweden.

Swanson, J., S. J. Kraus, and E. C. Gotschlich. 1971. Studies on
gonococcus infection. I. Pili and zones of adhesion: their relation
to gonococcal growth pattern. J. Exp. Med. 134:886-906.

Swanson, J., L. W. Mayer, and M. R. Tam. 1982. Antigenicity of
Neisseria ganarrhaeae outer membrane protein(s) III detected by
immunoprecipitation and western blot transfer with monoclonal
antibody. Infect. Immun. 38:668-672.

Swanson, J., K. Robbins, O. Barrera, D. Corwin, J. Boslego, J.
Ciak, M. Blake, and J. M. Koomey. 1987. Gonococcal pilin variants
in experimental gonorrheea. J. Exp. Med. 165:1344-1357.

Swanson, J., E. Sparks, D. Young, and C. King. 1975. Studies on
gonococcus infection. X. Pili and leukocyte association factor as
mediators of interactions between gonococci and eukaryotic cells in
vitro. Infect. Immun. 11:1352-1361.

Taha, M. K., B. Dupuy, W. Saurin, M. So, and C. Marchal. 1991.
Control of pilus expression in Neisseria ganarrhaeae as an original
system in the family of two-component regulators. Mol. Microbiol.
5:137-148.

Taha, M. K., M. So, E. Billyard, and C. Marchal. 1988. Pilin
expression in Neisseria ganarrhaeae is under both positive and
negative transcriptional control. EMBO J. 7:4367-4378.

Tam, M. R., T. M. Buchanan, E. G. Sandstrom, K. K. Holmes, J. 8.
Knapp, A. W. Siak, and R. C. Nowinski. 1982. Serological classi-
fication of Neisseria ganarrhaeae with monoclonal antibodies.
Infect. Immun. 36:1042-1053.

Tan, E. L., P. V. Patel, N. J. Parsons, P. M. V. Martin, and H.
Smith. 1986. Lipopolysaccharide alteration is associated with
induced resistance of Neisseria ganarrhaeae to killing by human
serum. J. Gen. Microbiol. 132:1407-1413.

277.

278.

279.

280.

281.

282.

283.

284.

285.

286.

287.

288.

289.

99

Tomasi, T. B., Jr., E. M. Tan, A. Solomon, and R. A. Prendergast.
1965. Characteristics of an immune system common to certain
external secretions. J. Exp. Med. 121:101-125.

Tramont, E. C. 1989. Gonococcal vaccines. Clin. Microbiol. Rev.
2:S74-S77.

Tsusukida, T., C. C. Wang, and F. W. Putman. 1979. Structure of the
A2m(1) allotype of human IgA-a recombinant molecule. Proc. Natl.
Acad. Sci. USA 76:1104-1108.

Van Epps, D. E., A. C. Plaut, G. M. Bernier, and R. C. Williams,
Jr. 1980. IgA paraprotein inhibition of human neutrophil chemo-
taxis. Reduced activity following treatment with IgA-specific
protease from Neisseria ganarrhaeae. Inflammation 4:137-144.

van Putten, J. P. M., Y. Pannekaek, J. F. L. Weel, and C. T. P.
Mopman. 1988. Neisserial proteins induced by environmental stress,
abstr. MT38, p. 178. 6th Int. Pathog. Neisseria Conf., Pine
Mountain, GA.

Vedros, N. A., and J. Kenyon. 1978. Mouse model for study of
gonococcal infection, p. 314-316. In G. F. Brooks, et al.

(ed.), Immunabiolagy of Neisseria ganarrhaeae. Am. Soc. Microbial.,
Washington, DC.

Virji, M., and J. E. Heckels. 1984. the role of common and type-
specific pilus antigenic domains in adhesion and virulence of gana-
cocci for human epithelial cells. J. Gen. Microbial. 130:1089-1095.

Ward, M. E., P. J. Watt, and A. A. Glynn. 1970. Gonococci in
urethral exudates possess a virulence factor last on subculture.
Nature (London) 227:382-384.

Warner, P. E., L. J. Zubrsycki, and M. Chila. 1980. Polygenes and
modifier genes for tetracycline and penicillin resistance in
Neisseria ganarrhaeae. J. Gen. Microbial. 117:103-110.

Watson, R. R., and S. Perrine. 1978. Aminopeptidase profiles from
cells of Neisseria ganarrhaeae, Neisseria sicca, Neisseria flava,
and Branhamella catarrhalis, p. 35-37. In G. P. Brooks, et al.
(ed.), Immunabialagy of Neisseria ganarrhaeae. Am. Soc. Microbial.,
Washington, DC.

Weinberg, E. D. 1978. Iran and infection. Microbial. Rev. 42:45-66.

West, S. E. M., and V. L. Clark. 1989. Genetic loci and linkage
associations in Neisseria ganarrhaeae and Neisseria meningitidis.
Clin. Microbiol. Rev. 2:SS92-SIO3.

West, S. E. H., and P. F. Sparling. 1985. Response of Neisseria
gonorrhoeae to iron limitation: alterations in expression of
membrane proteins without apparent siderophore production. Infect.
Immun. 47:388-394.

290.

291.

292.

293.

100

Winstanley, P. P., C. C. Blackwell, E. L. Tan, P. V. Patel, N. J.
Parsons, P. M. V. Martin, and H. Smith. 1984. Alteration of pyocin-
sensitivity pattern of Neisseria ganarrhaeae is associated with
induced resistance to killing by human serum. J. Gen. Microbiol.
130:1303-1306.

Wood, 8. G., and J. Burton. 1991. Synthetic peptide substrates for
the immunoglobulin Al protease from Neisseria ganarrhaeae (type 2).
Infect. Immun. 59:1818-1822.

WOods, J. P., J. P. Dempsey, T. H. Kawula, D. S. Barrit, and J. G.
Cannon. 1989. Characterization of the neisserial lipid-modified
azurin bearing the H.8 epitope. Mol. Microbiol. 3:583-591.

Yancey, R. J., and R. A. Pinkelstein. 1981. Sideraphare production
by pathogenic Neisseria spp. Infect. Immun. 32:600-608.

Chapter 1

CONSTRUCTION AND CHARACTERIZATION OF ISOGENIC IgAl

PROTEASE-DEFICIENT MUTANTS OF Neisseria ganarrhaeae

Russell J. Shoberg and Martha H. Mulks

Department of Microbiology and Public Health
Michigan State University

East Lansing, MI 48824 USA

101

102
ABSTRACT

This study was undertaken to determine any role(s) the type 2 IgAl
protease plays in the normal physiology of Neisseria gonorrhoeae. A set
of site-specific deletion-disruption mutants was constructed in the
cloned iga2 gene by standard recombinant DNA methodology and these
mutated alleles were introduced into the chromosome of strain GCM 740 by
transformation. The growth and membrane protein profiles of the Aiga
variants, GCM 740A4 and GCM 740A2, were compared to the parental strain
under a number of different standard in vitro growth conditions. The
wild type and the mutants grew at a similar rate in our standard defined
broth medium. However, the outer and cytoplasmic membrane protein
profiles of the mutants showed a number of differences detectable by
silver-staining of SDS-PAGE gels. Two of these differences were
demonstrated to be related to the Iga+ phenotype, as they "cured" when
the mutant allele was restored to a wild type allele; the p56 outer
membrane protein in bath A-iga mutant variants, and a 34.5 kDa
cytoplasmic membrane protein in GCM 740A4. Some of these unique and/or
missing outer membrane proteins were surface exposed. The mutants grew
and responded similarly to the wild type under anaerobic conditions and
under iron limited conditions, both of which have been shown to induce
specific alterations in the outer membrane protein profile. Also, the
mutants and, wild type were equally capable of posttranslationally

processing proteins and being transformed.

103
INTRODUCTION

The extracellular enzymes known as IgAl proteases are produced by a
set of bacteria which also share the characteristic of being human
pathogens capable of initiating infection on the mucosal epithelium
(20,34,43). The enzymes are produced by both gram negative and gram
positive bacteria in a variety of genera including Neisseria,
Haemophilus, Bacteraides, Capnocytophaga, Streptococcus, Clostridium,
and Ureaplasma. Thus far, only the pathogenic species of the genera
Neisseria and Haemophilus have been demonstrated to produce an IgAl
protease (34,37,43). Both N. ganarrhaeae and N. meningitidis have
been shown to be capable of expressing one of two described IgAl
proteases and in the case of H. influenzae, up to four different
enzymatic activities have been demonstrated (36,38). The type of enzyme
produced (e. g. gonococcal type 1 or type 2) is defined by the specific
peptide bond cleaved in the hinge region of IgAl and is determined by
apparent molecular weights of the resultant Pea and Fda observed on SDS-
PAGE (34).

The enzymes were initially defined as having a single substrate, the
human immunoglobulin Al, which was later shown to also include the IgAls
produced by the African gorilla and chimpanzee (34). The theorized
reason for this substrate exclusivity was proposed to be the presence of
a duplicated proline-rich amino acid octamer in the hinge region of IgAl
and absent in IgA2 as well as other sequenced immunoglobulins (34,43).

Since then, the substrate list for the Neisseria gonorrhoeae type 2
IgAl protease has been expanded to include the IgAl protease proenzyme
(45), some synthetic peptides with primary sequence homology to the IgAl

hinge region (62), and a set of gram negative bacterial outer and

104

cytoplasmic membrane proteins with undefined identities or function
(chapter 2 here and 50). For those first two mentioned substrates, the
specificity of the protease is at sites with reasonable sequence
homology to the IgAl hinge region. With regard to the final set of new
substrates, since the proteins are not defined any further than their
cellular compartment and apparent SDS-PAGE molecular weights, the amino
acid sequence of each cleavage site is not known.

As stated above, the enzyme has been demonstrated to be extracel-
lular in all species examined. A model for the secretion of the type 2
enzyme from N} ganarrhaeae has been proposed (45). According to this
model, the primary translational product (preproenzyme) is a 169
kilodalton (kDa) molecule which via the action of a leader endopeptidase
(in both E. coli and N. ganarrhaeae) loses an N-terminal leader peptide
in processing and is translacated to the periplasmic space as a 166 kDa
proenzyme. The C-terminus of this intermediate is referred to as the fi-
peptide or helper region (45 kDa) and is required for extracellular
secretion. Once the C-terminus of the preproenzyme is in the periplasm,
presumably the molecule exists in some conformation which favors
insertion of the fl-peptide into the outer membrane where it forms a pore
through which the rest of the enzyme is extruded. After the
preproenzyme is translacated through the outer membrane, an
autoproteolytic event releases the 121 kDa proenzyme from the surface.
Once the 121 kDa proenzyme is in its cell free form, it undergoes a
final autoproteolytic event to release either a 12 or a 15 kDa peptide
(both known as a) of unknown function. Both the 109 and 106 kDa forms
are presumed to be active against IgAl. It is not known whether the

autoproteolytic cleavages are cis or trans-mediated events. Conversion

105
from the 109 to 106 kDa forms has also been observed in vitro (51).

The gene coding for the N. ganarrhaeae type 2 IgAl protease has been
cloned by at least three different laboratories (15,21,35). A set of
isogenic Iga+ and Iga- variants of the gonococcal strain F62 has been
constructed previously and these variants were reported to differ from
the parental strain in only the ability to cleave IgAl (21). This
report also stated that there were no differences in total protein
profile, in growth on a complex agar medium, or in the ability to alter
piliation phenotype and colony color morphotype. A subsequent report
failed to demonstrate any differences in the mutant's ability to infect
and invade a human fallopian tube organ culture model system (10).
However, our laboratory subsequently reported that these two variants
possess different outer and cytoplasmic membrane protein profiles when
propagated in a defined in vitro broth medium (40).

The ability to produce an IgAl protease has been proposed to be a
virulence factor for several years (20,34,43). There is a wealth of
indirect evidence to support this hypothesis but there has yet to be a
direct comparison of a set of isogenic iga variants in an appropriate
model system (i. e. human or chimpanzee) to answer the question of its
contribution to virulence. Some of the indirect evidence includes: 1)
only the pathogenic species of a genus produce the enzyme; 2) only
pathogens of the human mucosal epithelia produce the enzyme; 3) the
enzyme has been demonstrated to be active in vivo by isolation of Fca1
(28) and active enzyme from patient specimens (34,43); 4) infected
patients exhibit an antibody response to the enzyme, and these
antibodies are inhibitory for enzyme activity (13); and 5) recently,

fresh samples taken from dental plaque have demonstrated the presence of

106
Fabal on the surface of oral streptococci, and after incubation of the
non-IgAl protease producing S. gordanii, in whole saliva, observed

cleaved IgAl Faba on its surface (1).

1

Some of the better characterized gonococcal virulence factors are
major proteinaceous components of the outer membrane. Most often, these
have been described from bacteria grown in either complex or defined in
vitro laboratory media under standard conditions of increased CO2
tension (5-10%) at 35-37’C. The outer membrane proteins relevant to
these studies are discussed below. There are also some outer membrane
components which have been described in detail which are induced or
repressed by relevant environmental stresses. These will be discussed
later in this section.

The major gonococcal outer membrane protein is the porin protein
(Par, formerly called PI; 16) which is the basis of a serotyping scheme
used in characterizing strains for epidemiological research (58). There
are two major Par serotypes, PorA and PorB. These proteins are neither
phase nor antigenically variable within an isolate. Par has an apparent
molecular weight of ~34-38 kDa by SDS-PAGE and is neither heat nor
reduction modifiable (57). There are no other described porin proteins
in the gonococcus. A role in virulence has been proposed as the Par
proteins are known to be targets of serum bactericidal and opsonic
antibodies (17,47).

Another component of the gonococcal pore is a 34 kDa protein known
as Rmp (formerly P111, 16). The acronym Rmp refers to its reduction
modifiable behavior on SDS-PAGE (26,57). Rmp is invariant with regard

to apparent molecular weight or antigenic type among all gonococcal

strains tested. Rmp is antigenically related to the E. coli porin

107
‘protein, OmpA (3). This relationship gives it a ‘passive role in
virulence as anti-OmpA antibodies have been demonstrated to function as
blocking antibody and prevent complement activation via the classical
pathway (46).

A third major protein component of the outer membrane which has been
the subject of intense study is the set of 24-30 kDa proteins referred
to as Opas (opacity, l6) and previously called PIIs. Opas exhibit heat-
modifiable behavior on SDS-PAGE, increasing in. apparent. molecular
weight after denaturation at 100°C relative to that of Opas denatured at
lower temperatures (26,57). The term opacity proteins refers to the
correlation between the color or opacity of a colony grown on agar and
the expression of these surface exposed, phase variable proteins (55).
Opas are also antigenically variable (48). Their functional role in
virulence has been suggested as an adhesin (23).

There are a number of environmental stress-induced outer membrane
proteins that are also implicated in virulence of the organism. With
regard to iron-stress, outer membrane proteins of 104 or 103; 97; 88 or
86; 80 or 79; 76, 74, or 73; 70; 45 or 41; 37 or 36; 29; 25; 23; 20.5;
and 19.5 or 19.0 kDa are inducible under iron-limited conditions
(l9,29,30,4l,60). Numbers with "or" between them refer to the same
protein described in different laboratories. Of this set of proteins,
the only one with a described function is that of 36 or 37 kDa (aka Fbp
and Mirp; 16) which has been demonstrated to bind Fe3+ in a specific,
reversible manner (32). There are reported to be specific outer
membrane receptors for lactoferrin and transferrin, two human iron
binding proteins, but no particular molecular weight has been reported

for either receptor in the gonococcus (4,24). These receptor proteins

108
would function in the essential role of iron acquisition from the host
and therefore would be vital for virulence. Presumably, at least some
of these iron-repressible outer membrane proteins are expressed in vivo,
as convalescent human serum shows immunoreactivity with them, including
Fbp (11).

When N. ganarrhaeae is propagated under anaerobic conditions, a set
of oxygen-repressible outer membrane proteins is expressed (Ani, 16;
previously Pan) and a set of oxygen-inducible outer membrane proteins is
repressed (Oxi, 16; previously Pox, 7). Anis 1-3 have been defined as
having molecular weights of 54, 46, and 31 kDa while Oxis 1-5 have
defined molecular weights of 62,27, 27, 22, and 16.5 kDa. At least one
of the Anis is expressed in vivo as patient sera from uncomplicated
infections and cases of pelvic inflammatory disease show reactivity with
Anil (8). Although there are no known functions for any of the Ani or
Oxi proteins, they are implicated in virulence due to the fact that
there is an antibody response raised against at least one of them.

Compared to the outer membrane, very little is known about the
identities of any of the proteins which are part of the cytoplasmic
membrane. In fact, other than a report on the specific density, only
one other paper describes the protein components (18,31).

Although a set of variants had been previously constructed and
described, we felt that there were more questions to be addressed than
those referred to in the previous report (21).

We decided to address the question as to what other functions the
IgAl protease might perform for the gonococcus in vitro by constructing
a set of isogenic Iga+ and Iga- variants with a defined background that

we could consider an IgAl protease-deficient phenotype. The gonococcus

109

expresses IgAl protease activity constitutively in vitro (and presumably
in vivo) and has never been demonstrated to lose the ability to produce
the enzyme in vitro, even after repeated subculture. Therefore, we
hypothesized that the IgAl protease had some required function in vitro.
We compared the relative abilities of the wild type (WT) and mutant
variants to grow in a standard defined medium for the gonococcus, to
posttranslationally process proteins, to be transformed, to shift colony
color and morphotype, and also compared the membrane protein profiles.

Due to our previous observation that IgAl protease-susceptible
proteins were present in the outer membrane of the gonococcus, and the
fact that the enzyme has been shown to function in the extracellular
milieu, we chose to concentrate on the role of the enzyme in outer
membrane protein profile responses in situ. Some of the other questions
that we attempted to answer involved the potential role of the IgAl
protease in maturation of the membrane protein profiles and in
modification of those profiles in response to relevant environmental
stimuli such as iron limitation and growth under anaerobic conditions.
These two environmental stress stimuli are particularly relevant given
the first niche that the gonococcus occupies in the human host. The
male 'urethra and. the female ‘vagina. would. have 'very low' levels of
available iron. due to the secretion. of lactoferrin. by the mucosal
epithelium (27). Due to the presence of normal anaerobic microbial
flora in the genitourinary tract and cervix, these microenvironments are
also considered to be anaerobic.

A second set of questions that we attempted to address was, would
the gonococcal protein secretion apparatus be confused or disrupted by

translation of truncated A-iga2 gene products? We attempted to

110
determine whether these truncated polypeptides would be detectable in
the membranes of the organism and if their synthesis interfered with the
normal secretion of other membrane proteins.
[Portions of this work have been presented at the 6th. International
Pathogenic Neisseria Conference (40) and at the 90th. Annual meeting of

the American Society for Microbiology (49).]

MATERIALS AND METHODS

Bacterial strains. The Neisseria ganarrhaeae strain GCM 740 is a
prototrophic, type 2 IgAl protease producing (genotype iga2b), serotype
PorIB2 clinical isolate which has been described previously (39). The
A-iga2 isogenic variants GCM 740A2 and GCM 740A4 were constructed in
this laboratory via deletion mutagenesis and insertion of a p-lactamase
gene cassette. Construction of these isogenic variants has been
presented previously (40). The revertant variants, GCM 740A2R-2, GCM
740A2R-40 and GCM 740A4R-83, were constructed in this laboratory by
transformation of the Aiga2 variants with the cloned iga2b gene.
Construction of the A-iga2 isogenic variants will be described in a
separate section of the Materials and Methods.

The Escherichia coli strains DHl and HB101 were used in the
manipulation of recombinant plasmids and have been described (25). The
E. coli strain Q358 has also been described previously and was used as a
host for the recombinant phage, A710 EMBL3. This recombinant
bacteriophage contained gonococcal insert DNA which encoded IgAl
protease activity in a lysate and has also been described (35). A
listing of plasmid vectors and recombinant iga2b plasmid constructs is

given in Table 1 and restriction maps of the various constructs are

111

Table 1. Properties of plasmids constructed and used.

 

 

Hamid ILL—ian e c s z “ Igélb L1! mm
pRH. 43 - R fidssUfly
pRJ2 9.4 + R This study
pRJ21 9.4 + S This study
pRB éfi + R nfissuMy
pRJ21Al 8.6 - S This study
pRJ21A2 8.3 - S This study
pRJ21A2Ar 10.7 - R This study
pRJ2lA3 7.1 - S This study
pRJ2lA4Ar 7.5 - R This study
pBR322 NAd NA R This laboratory
pBR322' NA NA S This study
pLES2 NA NA R V. L. Clark (54)
pNa16 10 NA 8 D. C. Stein (52)
a - Insert size in kilobase pairs
b - IgAl protease phenotype denoted as follows: +, Iga2+; -, Iga2-.

c - Beta-lactamase phenotype denoted as follows: R, resistant; S,

sensitive.

d NA, not applicable.

112
presented in Figure 1.
Media. N. ganarrhaeae was usually propagated on CC Base medium
(Difco Laboratories, Detroit, MI) containing Kellogg's supplement (61)

at 37°C in a 5-10% CO atmosphere. If it was necessary to propagate the

2
bacteria in broth, the bacteria were grown in the defined medium of
Morse and Bartenstein (33). Hereafter, these media will be referred to
as GCB and NEDA, respectively.

For selection and antibiotic sensitivity characterization of A-iga2
variants, gonococci were propagated on GCB containing either nalidixic
acid at 1.0 pg per ml or penicillin G at 0.025 pg per ml. All
antibiotics used in these studies were from Sigma Chemical Co. (St.
Louis, MO).

E. coli variants were propagated in LB broth or on LB agar plates
(25). When required for selective purposes, ampicillin or tetracyline
was added to the medium at 50-100 pg per ml and 12.5 pg per ml,
respectively. Propagation of A710 EMBL3 in E. coli Q358 was in LB
broth containing 0.2% maltose.

Subcloning of the iga2b gene. The recombinant bacteriophage A710
EMBL3 was identified by plaque hybridization (25) using a cloned iga
gene from Haemophilus influenzae (6) as a probe. A restriction site map
was determined and presented by Mulks and Knapp (35). DNA modification
enzymes and restriction endonucleases were purchased from Bethesda
Research Laboratories, Inc. (Gaithersburg, MD), Boehringer Mannheim
Biochemicals (Indianapolis, IN), or New England BioLabs, Inc. (Beverly,
MA) and were used as recommended by the suppliers.

Using this restriction site data, portions of the insert DNA from

A 710 EMBL3 were subcloned into pBR322 as a series of plasmids (pRJl,

113

Figure 1. Restriction site maps of recombinant plasmids constructed in

*
this study. Restriction sites are indicated as follows: A , Ach; A',

AvaI; A", AvaII; B', BglI; B", BglII; C, ClaI; H, HindIII; M, MluI; N,

NsiI; P, PstI; S, SmaI; and S', Sall. The direction of transcription is

as reported previously (45) and confirmed in our laboratory.

114

 

Figure 1
lgaZ
. :
H H' H'A'B' H' NA'PN PPH
W1 W

C A' M 8' MH H' H'A'B' H’ NA'PN PPH A'M B' A'SS'C

WWW
pR121

A' M 8‘ MH H' H'A'B' H'NA'PN PPH A'

pRJS L—I—l.—L—L—J—.LLL—.J—l—J—LL.——IJ—I_I

C A' M B‘ M H H' H'A'B’ H' N N PPH A'M B‘ A'SS'C

h—H—L—LJ—l—UJ—A-l———I-—U-l—U——l—-U-H

mm
C A' M B‘ M H H H'A’B' H NA‘P PH A'M B' A'SS'C
m -
W—-M
W
C A' M B‘ M H H H NA‘PN PPH A'M B‘ A'SS'C
h—d—l—l—H—J—--—-L—J—l—Ll——U—l—JJ——J—4LH
[9.21113
0 A' M B‘ M H H NM 8' A'SS‘C
MM" I—I—H—d—I— ------------- I—H——I—-H-H
I I I I l l I I I I I
kbp I I I I I I I I I I I

 

115
pRJ2, and pRJ3), in order to allow for the study of the IgAl protease
gene in E. coli on a multicopy vector. Transformation of E. coli HBlOl

was by the CaCl method (9). Restriction site maps are presented in

2
Figure 1 and pertinent information is summarized in Table 1.

Qualitative IgAl protease assays were performed as described by
Plaut, et al. (44). Briefly, experimental samples were incubated from
one to twelve hours at 37°C with four volumes of a reaction mix
comprised of 125I-labeled human IgAl in 50 mM Tris-HCl (pH 7.5), 10 mM
MgC12, 10 mM CaClz, and 0.5% bovine serum albumin (w/v). Reactions were
terminated by addition of four volumes of SDS-PAGE sample buffer (22).
Samples were boiled for five minutes and then resolved over 9.0% SDS-
PAGE (22). Reaction products (Hca, Pea, and Fda) were identified by
autoradiography at -70°C with an intensifying screen and X-Omat AR film
(Eastman Kodak Co. , Rochester, NY).

Construction of A-iga mutants in E. coli. The schema used to
construct the A-iga2b mutants are summarized in Figure 2. This first
involved the conversion of pBR322 (Ampr) to pBR322' (Amps) by frameshift
mutagenesis; this was accomplished by linearization of pBR322 with PstI,
blunt ending with the Klenow fragment of DNA polymerase plus dNTPs, and
religation with T4 DNA ligase. E. coli H8101 was transformed with the
ligation mix and propagated on LB containing tetracycline. Mutated
plasmids were identified by growth of the transformed strain on LB
containing tetracycline and failure to grow on LB containing ampicillin.
Site specificity of the mutation was confirmed by failure to be cleaved
by PstI, assayed visually by gel electrophoresis.

To allow for further selection based on ampicillin

resistance/sensitivity, pRJ21 was constructed by subcloning the 9.4 kbp

116

Figure 2. Diagram depicting the steps involved in construction of the

various recombinant deletion-disruption plasmids used in this study.

117

as. as

 

2862.. as. 2...
88.. (20 E. .v A. 734 . mum
it.” . eagle Bacon.
. .2: 28.82. £2. 8:. a
«31 ES. 3.238 a...
8:... S. ta :5. .8 a
51.5.3583: aegis
Sm... se. 2.38 a...
8:: a... ta 3...... .8 a
.3... so...
.888. 8.8) SB...
8.. to at... .8 ..

  

 

swoo- (zo E. a
2.252. .3...

.599 $533:

1&3 \
9: Wm. 15 Sou-<20 c» a

.48st .832. .2
3.. ad. 3:3 .oo 2

       

 

 

88. (zo E. a
Esp—Do: 39..

£325.32.

118
ClaI-bounded insert from pRJ2 (Iga2+) into the unique ClaI site of
pBR322’. Ligation mixes were used to transform, E. coli DHl and
recombinant constructs were identified by colony hybridization using the
4.3 kbp HindIII fragment of pRJ2 as a probe (25). This construct served
as the basis for all subsequent deletion mutants.

Deletion mutants were constructed by digestion of pRJ21 with either
HpaI, NsiI, or PstI followed by elution of the appropriately sized band
from agarose gels, religation, and transformation of E. coli DHl with
selection on LB containing tetracycline. DNA was prepared from the
transformants by an alkaline lysis miniprep procedure and analyzed for
loss of the appropriate restriction sites by endonuclease digestion and
agarose gel electrophoresis. Mutants constructed in this manner were
pRJ21A1 (deleted 0.8 kbp NsiI fragment), pRJ2lA2 (deleted 1.1 kbp PstI
fragment), and pRJ21A3 (deleted 2.3 kbp HpaI fragment). Figure 1 gives
a graphic representation of the restriction sites used to generate these
mutants.

In order to mark these deleted constructs with a phenotype which
could be used to select for eventual gonococcal transformants, a scheme
described by Koomey, et al. was utilized (21). This involved insertion
of a fi-lactamase gene (bla) cassette at the original sites of deletion
and. the schema are summarized. in Figure 2. 'The bla cassette was
prepared initially from pLESZ, an E. cali-N. ganarrhaeae shuttle vector
(54), by gel purification (25) of the 2.4 kbp HindIII fragment. The
HindIII termini were blunted by treatment with Klenow DNA polymerase
plus dNTPs, and this blunt ended fragment was ligated into the
similarily blunt ended site of the recipient A-iga2 construct.

A notable exception to this scheme was the construction of

119

pRJ21A4Ar. This mutant construct was generated by performing a partial
HindIII restriction digest on pRJ21, elution of the ~10 kbp band from an
agarose gel (25), and direct ligation of the 2.4 kbp HindIII bla gene
cassette, utilizing the cohesive ends of each fragment. E. coli DHl was
transformed with recombinant constructs selected on LB containing
ampicillin. Due to the presence of the complicating HindIII site in the
vector, and the presence of cohesive termini, this strategy was used
rather than the blunt end ligation scheme described above.

Transformation of N. ganarrhaeae GCM 740 to 13.32.. Initially,
several attempts were made to transform the gonococcal WT strain GCM 740
with the band purified, 10.7 kbp insert of pRJ21A2Ar by a standard
procedure (2). Prior to all transformations, a T1 colony morphotype was
selected and expanded on GCB to ensure competence. Competent gonococci
were resuspended in NEDA and one microgram of purified DNA was added.
After an incubation period at 37°C, the bacteria were plated onto
sterile membrane filters (Millipore Corp., Bedford, MA) preplaced on
warm GCB plates and the bacteria were incubated at 37°C under 5-10% CO2
for a 6 hour expression period. Following the expression period, the
membranes were aseptically transferred to a warm GCB plate containing
0.025 pg per m1 penicillin G and reincubated at 37°C for 48-72 hours.
At this time, penicillin resistant colonies were identified and selected
for further testing, including IgAl protease production (12,44),
penicillin resistance, and restriction mapping by Southern analysis
(25).

Due to difficulties with the timing of the expression period, the
consistency of penicillin G concentrations in the lots of media, and the

inability to recover transformants with the desired constellation of

120
phenotypes, an alternate selection procedure was employed to identify
putative transformants. As described above, competent gonococci were
selected and expanded on GCB, and were incubated at 37°C with the
transforming DNA (the 7.9 kbp ClaI insert of pRJ2lA4Ar or the 10.7 kbp
CIaI insert of pRJ21A2Ar). The expression period however, involved
spreading of the transformation mix directly onto the surface of a GCB

plate and a 6 hour incubation at 37°C under 5-10% CO Following this

2.
expression period, 0.1 ml of a 10 pg per m1 penicillin G solution was
spread over the central one-half of the plate. The plate was
reincubated as above, and the pencillin G was allowed to form a
diffusion gradient. At 48-96 hours after application of the antibiotic,
penicillin resistant colonies were observed growing in the zone between
the site of application and the periphery. These putative transformants
were selected and tested further for IgAl protease production (12,44)
and penicillin resistance, and were restriction mapped by Southern
analysis (25).

Preparation of bacterial membranes. Membranes were prepared from
bacteria grown to late exponential phase in NEDA by the sucrose-
lysozyme-EDTA spheroplasting method (42). Spheraplasts were disrupted
by ultrasonication and membranes were separated by isopycnic sucrose
density gradient ultracentrifugation (l8). Membrane protein
concentrations were determined by the Bradford dye binding assay (5)
using a commercial kit with bovine serum albumin as a standard (Bio-Rad
Laboratories, Richmond, CA).

Membrane protein profile analysis. Protein profiles were analyzed by

separation of 3 pg of membrane proteins per lane over 7.5-12.5% linear

polyacrylamide gradient SDS-PAGE (22) and a combination Coomassie blue-

121
silver staining procedure (14). Specific, described gonococcal outer
membrane proteins were identified by their relative abundance and
mobility, and in some cases by their heat- or reduction-modifiable
behavior (26,57). Apparent molecular weights were determined by
comparison to mobilities of a set of molecular weight standards (Bio-
Rad).

Assessment of surface exposed outer membrane proteins. To determine
whether any of the proteins which differed between the variants were
exposed on the surface of the bacterium, we radiolabeled the surface
exposed proteins using NalZSI (Amersham Corp., Arlington Heights, IL)
and lactoperoxidase (Sigma) by the procedure of Swanson (56). Briefly,
a 24 hour culture of gonococci were swabbed off a GCB plate and washed
once with phosphate buffered saline (PBS). Cells were resuspended in
PBS, 100 pg of lactoperoxidase in PBS and 100 pCi of Nalzsl were added,
followed by 40 p1 of 0.6% H202 (v/v). Additional H202 was added at
three minute intervals until fifteen minutes, when an excess volume of
PBS containing 5 mM cysteine was added to terminate the reaction. Cells
were pelleted by centrifugation and washed once in PBS plus cysteine
prior to final resuspension in PBS. Specific activity as cpm per ml was
determined by direct counting of aliquots in a GammaSSOOB gamma counter
(Beckman Instruments, Inc., Fullerton, CA). Analysis was performed by
separation on 7.5-12.5% SDS-PAGE gels, Coomassie blue staining, and
autoradiography of dried gels. Identification of described gonococcal
outer membrane proteins was accomplished by relative abundance and

mobilities, and/or heat- or reduction-modifiable behavior as described

above .

122
Growth of A-igaZ mutant gonococci under anaerobic conditions. The
methodology described by Clark, et al. was used to propagate this set of
variants on GCB in an anaerobic environment (7) . Gonococci were
inoculated on GCB with a 1.5 cm filter paper disc (Whatman, Inc.,
Clifton, NJ) placed at the center of the plate. The disk was inoculated

with 120 pl of 1 M NaNO and the plate was placed in a GasPak jar (BBL

2
Microbiology Systems, Cockeysville, MD). The jar was sealed and filled
with an anaerobic gas mixture containing 5% CO2 (BBL). Incubation was
at 37°C with growth assessed at 48 hours.

Outer membranes were prepared from 24 hour cultures of anaerobically
grown bacteria as well as from bacteria grown under standard aerobic
conditions by the sucrose-lysozyme-osmotic shock procedure of Clark, et
a1. followed by ultrasonication and extraction of outer membrane
proteins by Na n-lauroylsarcosine (7). Membrane protein concentrations
were determined and analysis of outer membrane proteins was performed as
described above.

Growth of A-igaz mutant gonococci in iron-limited media. Two
methods of rendering media iron-limited were utilized. One involved the
extraction and removal of ferric iron from NEDA by the chelator 8-
hydroxyquinoline (8-HQ; Sigma; 59) and the second involved chelation of
ferric iron in NEDA by the siderophore, deferrioxamine B mesylate
(Desferal; CIBA-GEIGY Corp., Summit, NJ). Gonococci have previously
been shown to be unable to assimilate iron chelated by Desferal (41) .

Ability to grow and growth rate under conditions of iron limitation
were assessed by growth in 8-HQ extracted NEDA. The ability to express

previously described iron-repressible outer membrane proteins was

determined by preparation of outer membranes as described above and

123
separation of the cytoplasmic and outer membranes by isopycnic sucrose
density gradient ultracentrifugation. Outer membrane protein
concentrations were determined, and analyses of membrane protein
profiles were performed as described above.

Pulse-chase analysis of wild type and A-igaZ mutant gonococcal
proteins. Gonococci were passaged twice in methionine-deficient NEDA
and radiolabeled in a 25 ml culture during either mid-exponential phase
or in stationary phase with 35S-Met (20 pCi per ml). Labelling was
accomplished with a one minute pulse followed by a chase period from 0-
120 minutes with excess unlabeled methionine (0.1 mM or 5000-fold
excess). At intervals, 1 m1 aliquots of bacteria were removed and
disrupted in an equal volume of cold 10% trichloroacetic acid (w/v).
The precipitate was recovered by 15 minutes of centrifugation at 10,000
xg in a microfuge (Brinkmann Instruments, Inc., Westbury, CT).
Insoluble pellets were washed twice with cold 5% trichloroacetic acid
and twice more with a cold ethanolzether (1:1) solution to remove the
trichloroacetic acid with excess solvent drained after each
centrifugation step. After the final centrifugation, excess solvent was
drained and evaporated overnight at 37°C. Dried pellets were
resolubilized in SDS-PAGE sample buffer (22) and if necessary,
neutralized with 1 N NaOH, boiled for five minutes and centrifuged for
two minutes to remove any insoluble debris. Incorporated radioactivity
was quantitated by liquid scintillation spectrophotometry in a Beckmann
LS counter using EcoLume aqueous scintillation cocktail (ICN).
Equivalent amounts of radiolabeled protein were loaded in each lane of a

7.5-12.5% SDS-PAGE gel as described above. Resolved proteins were

identified by Coomassie blue staining and autoradiography of dried gels

124
at room temperature.

Restoration of the A-iga2 loci to wild type. As N. ganarrhaeae is
transformable with a frequency of 0.2-1.0% (2,52), two different
procedures were employed in an attempt to identify the nonselectable, WT
transformants from the population of gonococci in a transformation mix.

Initially, competent Iga- gonococci (GCM 740A4 and GCM 740A2) were
transformed as per Atkinson and Mulks (2) with the gel purified 9.4 kbp
ClaI insert from pRJ2. Appropriate dilutions were plated directly onto
GCB to yield 100-200 colonies per plate. Isolated colonies were
transferred in duplicate to GCB and to GCB containing 0.025 pg per m1
penicillin. G in an attempt to identify the desired Penr to PenS
transformants.

A screening variation utilizing the IgAl protease plate assay
described by Gilbert and Plaut was also attempted (12). Following
transformation as described above, dilutions of cells were plated on GCB
and allowed to grow for 24 hours. At that time, a tap agar overlay
containing 125I-labeled IgAl (immobilized on polyacrylamide beads by
anti-human a chain; Bio-Rad) was applied. After a digestion period of 2
hours at 37°C, a nitrocellulose membrane (BA85; Schleicher 6: Schuell,
Inc., Keene, NH) was applied. Following a diffusion period of 20-30
minutes at 37°C, the membrane was removed, washed twice in PBS, and air
dried. The membrane was exposed to X-ray film overnight at -70°C with
an intensifying screen. By aligning the membrane, the autoradiograph,
and the original plate, putative IgAl protease producing colonies were
identified, picked through the overlay, and transferred to GCB plates
for expansion.

A second method of recovering the desired transformants was

125
attempted in order to increase the probability of screening gonococci
that had been transformed. Gonococci can be simultaneously transformed
with multiple pieces of DNA by a process termed congression (52). This
trait was exploited in order to reduce the number of colonies that
needed to be screened to isolate the desired restored phenotype. A
congression experiment using a limited amount of the non-linked
nalidixic acid resistance gene and an excess amount of the non-
selectable pRJ2 DNA, was performed, similar to that described by Stein
(52). Briefly, competent gonococci were resuspended in CC broth
containing 10 mM MgCl2 and 0.042% HCOS' (53) and a predetermined,
limiting amount of pNa16 DNA (the cloned Nalr gene) plus microgram
quantities of pRJ2 were added. Following a 30 minute transformation
period at 37°C, DNAaseI (Sigma) was added to 50 pg per ml, and the mix
was reincubated at 37°C for a 6 hour expression period, after which
dilutions were spread directly onto GCB plates containing 1.0 pg per ml
nalidixic acid. After 24 hours of incubation at 37°C under 5-10% C02,
Nalr colonies were identified and transferred in duplicate to GCB plates
for direct IgAl protease plate assays.

IgAl protease producing, Nalr isolates were confirmed by single
cloning of isolated colonies and subsequent testing of antibiotic
resistance/susceptibility as well as IgAl protease phenotype. Isolates
which were identified as being the desired revertants were restriction
mapped by Southern blot analysis with the 4.3 kbp HindIII iga2b gene
fragment and the 2.4 kbp HindIII bla gene cassette to verify that the
locus had been restored faithfully.

Transformants of the desired genotype were analyzed by comparison of

prepared outer and cytoplasmic membranes with those of the initial

126
parental type (GCM 740) and the intermediate A-iga2 mutants (GCM 740A2
and GCM 740A4). Membranes were prepared by procedures described above
involving separation of outer and cytoplasmic membranes by
ultracentrifugation over isopycnic sucrose density gradients.
Comparison of resolved membrane protein profiles was by analysis of

Coomassie blue-silver stained 7.5-12.5% SDS-PAGE.

RESULTS

Cloning of the iga2b gene and construction of A-iga mutants. Three
plasmid constructs were generated which contained insert DNA
corresponding to portions of the iga2b gene. Two of these produced
active IgAl protease (pRJ2 and pRJ3; inserts of 9.4 kbp and 6.6 kbp,
respectively; Figure 1) while the third, pRJl, contained ~90% of the
structural gene (insert of 4.3 kbp), presumably lacking the promoter
(45).

Four different deletion mutant constructs were generated. One
construct, pRJ2lA3, deleted 2.4 kbp in the 5' half of the gene,
presumably covering the active site (Figure 1; 45). Two of the
constructs, pRJ21A1 and pRJ21A2, each deleted ~17-25% (0.8 kbp and 1.1
kbp, respectively) of the gene covering the approximated junctions of
the structural protease domain, the a domain, and the 8 domain (Figure
l; 45). One further deletion was generated (pRJ21A4, 4.3 kbp deleted),
which removed ~90% of the structural protease domain, the entire a
domain and most of the 5 domain (Figure 1; 45). As predicted, all of
these constructs failed to produce active IgAl protease as assayed using
cell free supernatants and extracts of sonicated whole cells (data not

shown) .

127

Two of these deletion constructs were successfully interrupted with
the bla gene cassette and introduced via transformation into GCM 740.
These were pRJ21A2Ar and pRJ21A4Ar, with their recipient variants being
designated GCM 74OA2 and GCM 740A4, repectively (Figure 3). Disruption
of pRJ21A1 and pRJ21A3 with the bla cassette was not successful and
therefore, these constructs were not transformed into the gonococcus.

Characterisation of the A-iga mutants. When the growth rates of
these two mutant gonococcal variants were compared with that of the
parental WT, GCM 740, in a defined medium (Figure 4), no differences
were observed. During propagation on GCB, colony color and morphotype
changes were observed indicating that the A-iga mutants were capable of
undergoing both pilin and Opa phase variations (data not shown).

Numerous profile differences were observed when the isolated outer
and cytoplasmic membranes of the mutant variants were compared with
those of GCM 740. Figures 5 and 6 depict representative stained SDS-
PAGE gels of these comparisons for outer and cytoplasmic membranes,
respectively.

GCM 740 outer membranes contained unique bands of 130, 88, 48, and
25 kDa (Figure 5., lane 1) which were not present in either GCM 740A4 or
GCM 740A2 (Figure 5, lanes 2 and 3). GCM 740A4 was also observed to
lack a band at 94 kDa, present in both the WT and in the other mutant,
GCM 740A2. GCM 740A2 was observed to lack bands of 34, and 26 kDa that
were present in both GCM 740A4 and the WT, and to contain a unique band
of 28 kDa which was missing in the WT and in GCM 740A4. When the set of
isogenic Iga variants constructed by Koomey et al. was compared, there
were no differences observed in protein profile between the WT and the

mutant (data not shown;40).

128

 

k :

C A' ME MH H HA'B’ H’NA‘PN PPH NM 8’ A'SS’C

m I-I-I-I—I-H-IH—I-I-I-H—-II+—+I—-I—I+H

 

c A' M 8 MH H HA’B‘ H NA‘P PH NM 3' A'SS'C
G.,... I--I—I-I—I-I—I-I++—I+-H--+I—II-I-I+H
c A' M 3' MH H A'M s A'SS'C
Gwmm H-H_.H- ———————————— M
l l l I l _J
"m If i If i i I I I I I I
o 1 2 a ‘4 5 a '7 a 9 1a

Figure 3. Restriction site maps of the region around the iga2b locus in
strain GCM 740 and the A-iga2b loci of its isogenic strains GCM 740A2
and GCM 740A4. Sites are designated as follows: A*, Ach; A', AvaI; A",
AvaII; B', 8311; B", 33111; C, 01.31; H, HindIII; M, MluI; N, NsiI; P,

PstI; s, SmaI; and 5', Sell.

129

 

  

I I I I I I I I47 I I I
h NEDA A ‘
1 7. .1
o I :
N . .
L0 _ ..
Q ' -I
‘3 GCM 740 I
A GCM 7M3A4
a GCM 740.2
0.1;: -1
I 1
I- -I
l l I l J, I I l I

 

 

 

I I
0123456789101112

time (hours)

Figure 4. Growth rates of N. ganarrhaeae GCM 740 and two isogenic Aiga2

mutants in a defined broth medium (33) at 37°C.

130

Figure 5. Coomassie blue-silver stained SDS-PAGE gel of isolated outer
membranes of GCM 740 and its two isogenic A-iga2 mutants, GCM 740A4 and
GCM 740A2. The gel contains three micrograms of outer membrane proteins
per lane and is a 7.5-12.5% polyacrylamide linear gradient, stained as
per Gorg, et a1. (14). Lane 1, GCM 740; lane 2, GCM 740A4; and lane 3,
GCM 740A2. Missing protein bands are represented by a and unique bands
are represented by O . Positions of major gonococcal outer membrane
proteins and molecular weight standards are shown along the right margin

(kDa; Bio-Rad).

131

200
116

2

Figure 5

fnP
DLH no

1... ... .... . .2... .05.. ., no.1

.......::...... .0 a. . .pI...

. . .-....3 :‘E...ufl§$..'1..m;§w

132

The only protein which has been possible to define by standard
methods is the 26 kDa protein present in GCM 740 and GCM 740A4 which
showed heat modifiability and therefore belonged to the Opa class of
proteins (data not shown). All other differing pr6teins were
demonstrated not to be Por, Rmp, or Opas by standard procedures (data
not shown).

GCM 740 cytoplasmic membranes were observed to contain a unique
doublet band at 24 kDa and to be missing a band at 19 kDa (Figure 6,
lane 1). GCM 740A4 cytoplasmic membranes contained a unique band of 39
kDa which was missing in both the WT and GCM 740A2 (Figure 6, lane 2).
GCM 740A2 was missing bands of 49 and 27 kDa which the other variants
exhibited, and had unique bands of 110, 62, and 44 kDa. Similarly, when
the isogenic set of variants constructed by Koomey, et al. was compared,
the Iga' mutant was observed to lack a band of 65 kDa and to express a
novel band of 56 kDa (data not shown;40).

Surface exposure of the variant outer membrane proteins. As
demonstrated by the ability to be labeled with 125I by the
lactoperoxidase procedure of Swanson, some of the 'variant proteins
described above for GCM 740A2 were exposed on the cell surface (56). In
particular, the major protein of 28 kDa, unique to GCM 740A2 was labeled
(Figure 7, lane 3) while the missing band at 26 kDa (present and labeled
weakly in both GCM 740 and in GCM 740A4) was observed in lanes 1 and 2.
Also, a second protein band of 12.5 kDa was absent in GCM 740A2 that was
present and labeled in the other variants. By virtue of the apparent
SDS-PAGE molecular weight and heat-modifiable behavior, the 26 kDa band

present in GCM 740 and in GCM 740A4 was considered to be an Opa while

the unique 28 kDa protein observed in GCM 740A2 was not heat-modifiable

133

Figure 6. Coomassie blue-silver stained SDS-PAGE gel of isolated
cytoplasmic membranes of GCM 740 and its two isogenic A-iga2 mutants,
GCM 74OA4 and GCM 740A2. The gel contains three micrograms of
cytoplasmic membrane proteins per lane and is a 7.5-12.5% polyacrylamide
linear gradient, stained as per Gorg, et a1. (14). Lane 1, GCM 740;
lane 2, GCM 74OA4; and lane 3, GCM 740A2. Missing protein bands are
represented by a and unique bands are represented by 0. Positions of
the molecular weight standards are shown along the left margin in kDa

(Bio-Rad).

134

Figure 6

200

 

135

Figure 7. Autoradiograph of 1251 surface labeled N. ganarrhaeae cells

separated over a 7.5-12.5% SDS-PAGE. Cells were labeled with
lactoperoxidase as described previously (56). Lane 1, GCM 740; 2, GCM
740134; and 3, GCM 740A2. Unique proteins are represented by O and
missing proteins are represented by CL Positions of major gonococcal
outer membrane proteins are given for reference and molecular weight

standards are given along the left margin in kDa (Bio-Rad).

Figure 7

136

 

 

137
and therefore was not an Opa (data not shown; 26,57). None of the other
unique bands described in Figure 5 were observed to be surface exposed
by their ability to be labeled.

Growth and analysis of outer membrane protein profiles under
anaerobic conditions. All three variants of N. ganarrhaeae were capable
of growth on GCB solid medium supplemented with NaNO2 (data not shown).

The outer membrane protein profiles of the three variants showed
some intervariant differences when grown under anaerobic conditions.
These interstrain novel proteins expressed under anaerobiosis differed
from those observed when the variants were grown under aerobic (i. e.
normal) conditions. GCM 740 expressed a unique protein band of 31.5
kDa, GCM 740A4 expressed a unique band of 66 kDa, and GCM 740A2
expressed a unique band of 20.5 kDa (Figure 8; lanes 7-9).

The Aiga2 variants were demonstrated to be capable of expressing the
appropriate anaerobic stress proteins in their outer membranes as
demonstrated in Figure 8. All three variants expressed Anis 1-3 (lanes
1, 3, and 5) and repressed the synthesis of Oxis l-5 (lanes 2, 4, and
6), as defined by Clark, et a1. (7).

A number of other oxygen-repressible outer membrane proteins were
observed for the set of three variants including proteins of: 125, 94,
64, 38, 25, 22.5, and 16 kDa which were not defined by Clark, et a1.
(7). Similarly, two previously undescribed oxygen-inducible outer
membrane proteins of 92 and 45 kDa were expressed by all three variants.

Growth and analysis of outer membrane protein profiles under iron-
limited conditions. All three variants were capable of growth under
iron-limited conditions (Figure 9). All three variants grew to a higher

final optical density when propagated in iron-replete media as compared

138

Figure 8. Extracted outer membrane proteins from N. ganarrhaeae GCM 740,
GCM 740A4, and GCM 740A2 grown under aerobic and anaerobic conditions on
GCB solid. medium. IEach lane contains 3 pg, of ‘n-lauroylsarcosine-
insoluble proteins (as per Clark, et al., 7) separated by 7.5-12.5%
linear polyacrylamide gradient SDS-PAGE and stained with Coomassie blue-
silver. For lanes 1 through 6, odd numbered lanes contain membranes from
cells grown under anaerobic conditions and even numbered lanes contain
membranes from cells grown under aerobic conditions. Lanes 1 and 2, GCM
740; 3 and 4, GCM 740A4; 5 and 6, GCM 740132. Positions of oxygen-
repressible outer membrane proteins (Pans) are given to the left of each
pair and are represented by 0 while oxygen-induced outer membrane
proteins (Poxs) are represented by ‘I on the right edge of each pair.
Positions of previously described Pan and Pox (7) and major gonococcal
outer membrane proteins are given for reference points. Lanes 7 through
9 contain. n-lauroylsarcosine-insoluble proteins extracted from Icells
grown under anaerobic conditions (7, GCM 740; 8, GCM 740A4; and 9, GCM
740A2). Positions of unique outer membrane proteins are shown byIO to
the left of each lane. The positions of molecular weight standard
proteins are given in kDa along the right edges of the figure (Bio-

Rad).

 

 

139

Figure 8

7 8 9
200
116
97
Pox1
F’am’° 66
Pen2- ,
43
a -- C
Pans. . Par
Pox2 Rmp
POX3 s Q. 31

 

 

140

 

 

740 A 4——NEDA + Df
740 A 2—-NEDA '
740 A 2—-—NEDA + Df

001 L l l l I I
0 1 2 3 4 5 6 7

time (hours)

I I I I I I
1 _—- 0 1
P ‘ 3
r _
- /‘ .I
o ,_ _
N “ff/AMA
.1. '/
8 0.1 r _/ —:
I
/ 740—-NEDA I
- 740——NEDA + Df ~
- 740 A 4—-NEDA a

IUDDOO

 

 

 

Figure 9. Growth curves of N. ganarrhaeae GCM 740, GCM 740A4, and GCM
740A2 grown under iron-replete and iron-limited conditions. Bacteria

were grown in either NEDA plus 10 pM Fe3+ (added as Fe(NO -9H 0; iron-

3’3 2
replete; NEDA) or in the same medium plus 50 pM Desferal (iron-limited;

NEDA + Df). Both media were supplemented with 0.252 mM CaCl2 and 0.042%

NaHCO3 (w/v) and incubated at 37°C with shaking.

141

Figure 10. Coomassie blue-silver stained SDS-PAGE gels of isolated outer
membranes from N. ganarrhaeae GCM 740, GCM 740134, and GCM 740A2
separated over a 7.5-12.5% linear polyacrylamide gradient. Lanes 1, GCM
740 in iron-limited medium; 2, GCM 740 in iron-replete medium; 3, GCM
740A4 in iron-limited medium; 4, GCM 740134 in iron-limited medium; 5,
GCM 740A2 in iron-limited medium; and 6, GCM 740232 in iron-limited
medium. Iron-replete medium was NEDA plus 10 pM Fe3+, 0.252 mM CaC12,
and 0.042% NaHCO3 (w/v). Iron-limited medium was prepared by addition
of 50 pM Desferal to the same medium. Iron-repressible outer membrane
proteins are represented by O and positions of major gonococcal outer

membrane proteins are given for reference. Positions of molecular

weight standards are given along the right margins in kDa (Bio-Rad).

 

 

 

142

Figure 10

 

 

 

143
to iron-limited media.

All three of the variants were also able to express some of the
appropriate described iron-regulated outer membrane proteins as
demonstrated in Figure 10. In particular, the conserved Fbp protein (35
kDa here; 29) was expressed. Other previously described iron-
repressible outer membrane proteins that were observed included: 102,
85, 78, 73, 57, and 41 kDa.

Pulse-chase labeling analysis of WT and A-iga2 variants. To
determine whether differences existed in the ability to process proteins
between the WT and the mutant gonococci, we performed a series of pulse-
chase labeling experiments using Tran3SS-label (ICN Pharmaceuticals
Inc., Irvine, CA) in methionine-deficient NEDA. When N. ganarrhaeae
cells were pulsed during exponential phase, proteins were observed to
chase "out” (i. e. decrease in density over time) after addition of an
excess of cold carrier methionine while others chased "in". This
indicated that both of the mutant variants are capable of post-
translational processing with similar efficiency as the WT. These data
are summarized in Table 2 and shown in Figure 11.

In general, the set of variants were observed to process relatively
few proteins during exponential growth, as only 2 proteins were observed
to chase "out" for GCM 740 (Figure 11A), and one unique protein was
observed for each of the two mutant variants (Figure 118 and C). The
proteins observed in the WT were of 35.5 and 27 kDa. A unique protein
of 48 kDa was observed in GCM 740A4 and a unique protein of 113 kDa was
observed for GCM 740A2 (Figure 11B and C, respectively).

When these autoradiographs were analyzed with regard to the protein

bands which chased "in" (i. e. increased in density over time), we

144

Table 2. Summary of pulse-chase labelling data.

 

Bands processed during exponential phase8

 

 

 

 

Bands chasing "out" Bands chasing "in"
GCM Z49 GCM 240A4 GCM Z4042 GCM Z40 GCM Z4094 GCM Z4OA2
113 116 116
48 112
35.5 97 97 97
27 89
67
55
39
28
Bands processed during stationary phase8
Bands chasing "out" Bands chasing "in"
GCM Z49 GCM Z4QA4 GCM Z40A2 GCM Z40 GCM Z40A4 GCM Z40A2
35.5 215
162
143 143
140
120 120 120
116
109
107 107
104
95 95
89
78 78 78
69 69
66
61.5
47.5
46
33

 

Apparent molecular weights given in kDa.

145

Figure 11. Autoradiographs of N. ganarrhaeae cells pulse-chase labeled
during exponential phase with 35S-methionine. Macromolecules were
precipitated with trichloroacetic acid, resolubilized by boiling in SDS-
PAGE sample buffer, and separated by 7.5-12.5% SDS-PAGE. Panels are as
follow: A) GCM 740, B) GCM 740A4, and C) GCM 740A2. Lanes are samples
taken at intervals post-chase as follow: lane 1, 0 minutes; 2, 30
seconds; 3,1 minute; 4, 2 minutes; 5, 4 minutes; 6, 8 minutes; 7, 16
minutes; 8, 32 minutes; 9, 64 minutes, and 10, 128 minutes. Positions
of protein bands which chased "out" following the pulse of label are
given along the left margin as are some major gonococcal outer membrane
proteins. Positions of protein bands which were observed to chase "in"
after the labeling period as well as the molecular weight standards (in

kDa, Bio-Rad) are shown along the right margin.

 

 

 

146

.Immmmmmmuum1 Hummumuuua~1

'III1 ..lII .1 -
.1uu1mu111m1 111......1nu .1. .
wuummuuuawa... Isles-IL. 1......

'.I I

”(I “.'...‘.' I

.‘IVIII'I '9‘
0° IuIlJIJWIIWI

II II'II
11‘.

NOI .I'IlIlI-lltslI

0:1“

IIIIIIII".

Owanhnmvnaw

Figure 11 ‘

[1|

....... -80-
."-...|l1'1'

-.
II':.1U.HI
'

II.'|'| l"' IIuI
1“. I

Oranhom'nflw

'I.‘
, I:
.10,
1| 1
WM.- 1
1

OFGONODQMNF

147

observed a similar situation in that a low number of proteins were being
processed (Figure 11). Three different proteins were observed to chase
”in" for the WT with each of the mutants having similarly few bands, 3
for GCM 740A4 and 5 for GCM 740A2. A band of 97 kDa was observed to be
processed by all three variants. A band of 116 kDa was processed by
both of the mutants, but not by the WT. GCM 740 also processed unique
proteins of 67 and 28 kDa. GCM 740A4 uniquely processed an additional
protein of 55 kDa while GCM 740A2 uniquely processed proteins of 112,
89, and 39 kDa.

When the cells were labeled during stationary phase, virtually no
protein bands were observed to be chased "out" while a relatively larger
number chased "in". These data from stationary phase cells are also
summarized in Table 2 and are shown in Figure 12. With regard to the
number of proteins chased "out", only a single protein was observed to
be processed in GCM 740A2 (Figure 120; 35.5 kDa)

This low amount of processing was not observed however when the
variants were analyzed for proteins which chased "in", as the WT
processed 9, GCM 740A4 processed 11, and GCM 740A2 processsed 6
proteins. Two proteins of 120 and 78 kDa were processed by all three
variants. Three proteins of 107, 95, and 69 kDa were processed by both
of the mutants but not by the WT. A protein of 143 kDa was processed by
both the WT and GCM 740A4, but not by GCM 740A2. GCM 740 processed
unique proteins of 116, 104, 89, 61.5, 47.5, and 46 kDa. GCM 740A4
processed unique proteins of 162, 140, 109, 66, and 33 kDa while GCM
740A2 processed a single unique protein of 215 kDa.

The 89 kDa protein uniquely processed by GCM 740 during stationary

phase (Figure 12A) and the protein of 28 kDa uniquely processed during

148

Figure 12. Autoradiographs of N. ganarrhaeae cells pulse-chase labeled
during stationary phase with 35S-methionine. Macromolecules were
precipitated with trichloroacetic acid, resolubilized by boiling in SDS-
PAGE sample buffer, and separated over 7.5-12.5% SDS-PAGE. Panels are
as follow: A) GCM 740, B) GCM 740134, and C) GCM 740A2. Lanes are
samples taken at intervals post-chase as follow: lane 1, 0 minutes; 2,
30 seconds; 3,1 minute; 4, 2 minutes; 5, 4 minutes; 6, 8 minutes; 7, 16
minutes; 8, 32 minutes; 9, 64 minutes, and 10, 128 minutes. Positions
of protein bands which chased "out" following the pulse of label are
given along the left margin as are some major gonococcal outer membrane
proteins. Positions of protein bands which were observed to chase "in"

after the labeling period as well as the molecular weight standards (in

kDa, Bio-Rad) are shown along the right margin.

 

 

149

Figure 12

 

or a w .n n. n v a. N w

1..........1

1.5.551

or o 0 .n o .m v a N w

9:...
tom

IO
. ‘

V

‘

'1

V

IN. t‘l'llrlo 11
7.
1W.
V.

or a o h o m w a N _.

 

 

IIO= .

IIOON.

150
exponential phase (Figure 11A), may correspond to the unique GCM 740
outer membrane protein bands of 88 and 25 kDa shown in Figure 5 with
the differences in apparent molecular weight due to experimental
variation. However, since the pulse-Chase analysis utilized disrupted
whole cells, this is not considered proof of identity.

Restoration of the igaZb allele in the Aiga variants. In general,
the proposed scheme for identification of Iga' to Iga+ revertants
constructed by transformation with the 9.4 kbp CIaI fragment was found
to be unreliable. Attempts to identify the reverted iga allele by a
shift from Penr to PenS were difficult due to the degree of subjectivity
in differentiating Penr from Pens colonies. The use of a direct IgAl
protease plate assay (12) to screen the plated transformation mix for
Iga+ colonies was also considered unreliable in that even though
putative Iga+ transformants could be identified, recovery of viable
organisms from. beneath the top agar overlay was not consistently
successful. using these two screening methods, over 18,000 GCM 740A4
and 23,000 GCM 740A2 potential transformants were tested without
success. In general, the low transformation frequency expected also
made these attempts to identify a nonselectable transformant extremely
difficult.

Use of the congression experimental approach (52) was successful in
identifying Iga+ revertants for both GCM 740A4 and GCM 740A2. By first
selecting for transformants using Nalr as a selectable phenotype, two
GCM 740A2 revertants, GCM 740A2R-2 and GCM 740A2R-40, were isolated from
117 screened (1.71% congression) and one GCM 740A4 revertant, GCM
740A4R-83, was isolated from 708 screened (0.14% congression).

Secondary screening for the desired Iga+ revertants involved duplicate

151

plating on GCB and a direct IgAl protease plate assay to identify any
Iga+ colonies from the pool of Nalr transformants selected.

Characterization of the A-igaZb revertant variants. Use of a
standard qualitative IgAl protease assay (44) confirmed that these three
revertants were producing the specific IgAl protease (Figure 13).

Southern. analysis of chromosomal DNA from the set of 'variants
demonstrated that the restoration event had taken place faithfully.
HindIII digested DNA from the revertants showed an iga2b homologous
band of the appropriate size (4.3 kbp; Figure 14A). The revertants had
lost any DNA homology with the his gene cassette (Figure 14B). Finally,
the appropriate variants lacked homology with the 1.1 kbp PstI fragment
of an iga2b gene while the revertants regained homology (Figure 14C).
Southern analysis using the ~10 kbp PstI insert fragment of pNal6,
demonstrated that this probe hybridized to a different band (~6 kbp)
than that of the 4.3 kbp HindIII iga2b probe indicating that the two
loci were not physically linked (data not shown).

Isolated outer membranes were analyzed by protein profile (Figures
15 and 16). Some profile differences were observed consistently, such
as the novel bands of 102, 86, and 80 kDa and the missing band of 82 kDa
which correlated with Nalr. One protein band of 56 kDa was observed to
correlate with the Iga+ phenotype as it was not observed in either GCM
740A4 or GCM 740A2 (Figure 16). However, this protein was not noted as
being unique to the WT in Figure 5.

Isolated cytoplasmic membrane proteins were also analyzed by protein
profile also (Figure 17). One of the proteins which varied between
mutant and WT was observed to be "cured" by reversion of the A-iga2b

lesions. The 34.5 kDa band which is missing in GCM 740A4 and present in

152

12345678910

Hum

Fd» .' ' . .

Figure 13. Autoradiograph of IgAl protease assay cleavage products.
Assays were performed as described in Materials and methods. Lanes are
as follow: 1, GCM 740; 2, GCM 740Nalr; 3, can 7110114; 4, ccn 740A4Nalr;
5, GCM 740A4R-83; 6, GCM 740A2; 7, GCM 740A2Nalr; 8, GCM 740A2R-2; 9,
GCM 740A2R-40; and 10, negative control. Positions of intact heavy
chain (Han) and one of the cleavage products (Fda) are shown along the

left margin.

153

I 2 3 4 5 e 1 g 9 1 2 3 4 5 5 7 a 9 1 2 3 4 5 G 7 8 9
. ' _ U
—— —23.1—, —2:I.1 —. —2:I.1
O
— . ~91 — ‘ _. 9.. _ _..‘
— -6.7— - _ 6.7 — —6.7
_-- - “~—4.C-— -—4.4 —-- - ‘--—4.4
- a -
. -O C-
— ' ' -2~3— 1 —2.:I : :2:
__1 1 —2o —— _2o .
A a C

Figure 14. Southern blot analysis of WT, A-iga2 mutant, and revertant
strains of N. ganarrhaeae GCM 740. All DNAs were digested with HindIII
and separated over 0.65% agarose gels in TBE (25). Panels were probed
as follow: A) 4.3 kbp HindIII iga2b fragment, B) 2.4 kbp HindIII bla
cassette, and C) 1.1 kbp PstI fragment of iga2b gene. For each panel,
lanes are as follow: 1, GCM 740; 2, GCM 740Nalr; 3, GCM 740A4; 4, GCM
740A4Nalr; 5, GCM 740A4R-83; 6, GCM 740A2; 7, GCM 740A2Nalr; 8, GCM
740A2R-2; and 9, GCM 740A2R-40. Molecular weight standards are shown
along the right margin of each panel (A DNA digested with HindIII;

Bethesda Research Laboratories, Inc., Gaithersburg, MD).

154

Figure 15. Coomassie blue-silver stained 7.5-12.5% SDS-PAGE gel of
isolated outer membranes from WT, A-iga2 mutant, and revertant strains
of N. ganarrhaeae GCM 740. Three micrograms of outer membrane proteins
were loaded per lane as follow: 1, GCM 740; 2, GCM 740Nalr; 3, GCM 740;
4, GCM 740A4; s, GCM 740A4Nalr; 6, GCM 74OA4R-83; 7, GCM 740Nalr; 3, GCM
740; 9, GCM 740A2; 1o, GCM 740A2Nalr; 11, GCM 740A2R-2; 12, GCM 740A2R-
40; and 13, GCM 740A4Nalr. Unique protein bands are represented by 1.
and missing proteins are represented by a. Both are shown to the left
of the indicated band. The 56 kDa protein band which correlated with
the Aiga2 lesion in strain GCM 740-A4 is represented by +-.. Positions
of major gonococcal outer membrane proteins are given along the right
margin as are the molecular weight standards in kDa (Bio-Rad). The
position of p56 is noted and the positions of the set of three bands

which correlated with Nalr is represented by N3.

 

 

 

155

Figure 15

SN

.9

.V.Oa

up 3.

. .OF was

fl
0

SN

 

 

8N

 

 

156

Figure 16. Coomassie blue-silver stained 7.5-12.5% SDS-PAGE gel of
isolated. outer membranes from ‘WT and. A-iga2 mutant variants of N.
gonorrhoeae GCM 740. Three micrograms of outer membrane proteins were
loaded per lane as follow: 1, GCM 740; 2, GCM 740A4; and 3, GCM 740A2.
Unique protein bands are represented by O and missing proteins are
represented by 0. Both are shown to the left of the indicated band.
Positions of major gonococcal outer membrane proteins are given along
the right margin as are the molecular weight standards in kDa (Bio-Rad).

The position of p56 is noted.

 

 

were
hOAZ.
; are
band
aloné

MM

Figure 16

200

 

158

Figure 17. Coomassie blue—silver stained 7.5-12.5% SDS-PAGE gel of
isolated cytoplasmic membranes from WT, A-iga2 mutant, and revertant
strains of N. gonorrhoeae GCM 740. Three micrograms of cytoplasmic
membrane proteins were loaded per lane as follow: 1, GCM 740; 2, GCM
740Nalr; 3, GCM 740; 4, GCM 740A4; 5, GCM 740A4Na1r; 6, GCM 740A4R-83;
7, GCM 740Na1r; 8, GCM 740; 9, GCM 740232; 10, GCM 740A2Nalr; 11, GCM
740A2R-2; 12, GCM 740A2R-ao; and 13, GCM 740A4Nalr. Unique protein
bands are represented by 0’ and missing proteins are represented by 0.
Both are shown to the left of the indicated,'bandq Positions of
molecular weight standards are shown along the right margin in kDa (Bio-

Rad).

159

Figure 17

 

 

160
the WT, does show up in the revertant GCM 740A4R-83 (Figure 17; lanes 3-

6). Again, this band was not observed to be a variant band in Figure 6.

DISCUSSION

We have been able to successfully construct a pair of Neisseria
gonorrhoeae A-iga2 variants isogenic with GCM 740, a WT clinical
isolate. These two deletion-disruption mutants were capable of growing
at a similar rate in a standard defined medium for N. ganarrhaeae
(Figure 4) and were capable of posttranslationally processing proteins
during both exponential growth and stationary phases. Both of the
mutants were competent for transformation as demonstrated by recovery of
transformed variants with a faithfully restored iga2 locus (Figure 14).
They were also capable of phase variable pilus expression, evidenced by
the recovery of both P+ and P- colony variants (data not shown).
Likewise, the mutants were able to express Opas in a phase variable
manner as determined both grossly, by changes in colonial color on GCB
and molecularly, by assessment of heat modifiable behavior of outer
membrane proteins (data. not shown).

The mutants exhibited altered membrane protein profiles (Figures
5,6,15,16, and 17) as compared to. their WT parent. HOwever, except in
the cases of one outer membrane protein (p56, present only in the WT)
and a cytoplasmic membrane protein of 34.5 kDa missing in GCM 740A4 and
GCM 740A2 and present in GCM 740A4R-83), none of these other membrane
protein differences appeared to "cure" with restoration of the WT iga2
locus (Figures 15-17).

With regard to the differences observed in the outer membrane

protein profiles between WT and A-iga2 mutants, it was not possible to

161
identify the majority of these differing minor component proteins as
previously described outer membrane proteins or environmental stress
response outer membrane proteins. The sole exception to this was the 26
kDa protein missing in GCM 740A2 but present in both the WT and in GCM
740A4. In the two variants which exhibited this protein band, it was
identified as an Opa based on heat-modifiable behavior (data not shown).

We had initially hypothesized that the mutants would not be capable
of inserting the relevant environmental stress proteins into their outer
membranes based on the separate observations that 1) the membrane
protein profiles were different when propagated in a "normal" defined
medium and 2) that there are proteins present in the outer membrane that
were permissive substrates for IgAl protease in vitro (39,46). We
hypothesized that these differences were the result of some IgAl
protease trans-mediated effect. When this hypothesis was tested we
observed that the mutants were capable of growth under both iron-limited
(Figure 9) and anaerobic conditions. Furthermore, both of the mutants
were completely capable of expressing a full range of previously
described oxygen-repressible and iron—repressible outer membrane
proteins (Figures 8 and 10).

We tested the abilities of the mutants to process proteins
posttranslationally in an attempt to correlate these differences in
membrane protein profiles with the iga2 locus. Our results indicated
that the mutants had the ability to process proteins with similar
efficiency as their WT parent (Figures 11 and 12). Therefore, the
hypothesized pleiotrophic phenotype of the iga locus was not supported
by these data.

The 28 kDa protein which was observed to be unique to GCM 740132

162

(Figure 5), exhibited phase variable expression as it was not observed
in all membrane preparations from this variant (Figures 5 and 16).
This, considered with its apparent SDS-PAGE molecular weight and surface
exposure (Figure 7), was suggestive of an Opa. However, when
preparations which contained the 28 kDa protein were analyzed to define
heat modifiable proteins, the protein did not exhibit heat modifiable
behavior (data not shown). This apparent contradiction suggests that
the 28 kDa outer membrane protein is not an Opa but rather some
undescribed, phase variable outer membrane protein.

The fact that not all of the differences were observed on all gels
with a single preparation, and that all differing bands could not be
observed from preparation to preparation, indicated that the majority of
these detected differences were not directly related to the A-iga2
lesion. Rather, the fact that only one of the bands, p56, correlated
with the 13.32 phenotype suggests that the bands which differed from
preparation to preparation were likely due to the relative yields of a
particular protein in that particular preparation. Alternatively, there
may have been secondary mutations which had arisen either as a
compensatory result of the original A-iga2 lesion or as completely
unrelated lesions.

Most of the observed differences in cytoplasmic membrane protein
profiles were not observed to be directly related to the 1332 phenotype.
This was indicated by the failure of most of the profile differing bands
to be "cured" by restoration of the WT allele in the deletion mutant
variants. The sole exception was a 34.5 kDa protein band present in WT
but not in GCM 740A4 or GCM 740A2 and expressed in the revertant GCM

740A4R-83 (Figure 17).

163

In summary, most of the observed differences in the outer and
cytoplasmic membrane protein profiles that were presented in the Results
section were apparently not the result of the particular iga2 allele
present in the bacterium. However, the p56 outer membrane protein and
the 34.5 kDa cytoplasmic membrane protein did "cure" when the A-iga2
allele was restored to WT and therefore must be associated with the Iga
phenotype in some fashion. It is not known whether the genes might be
physically linked or if the production of IgAl protease influences these
other proteins in some posttranslational fashion. Overall, the observed
intervariant variability in the membrane protein profiles is not solely
accounted for by the Iga phenotype.

Finally, the phenotypic ability to produce an igal protease does not
affect the physiologic status of N. ganarrhaeae under the sets of
conditions tested in the laboratory. Two phenotypic differences were
observed to correlate with the A-iga2 allele: 1) the inability to
hydrolyze human IgAl and 2) the lack of expression of a 56 kDa outer

membrane protein and a 34.5 kDa cytoplasmic membrane protein.

ACKNOWLEDGEMENTS
I wish to thank Dr. V. L. Clark for providing me with pLESZ, Dr. D.
C. Stein for providing pNa16 as well as for helpful discussions relevant
to gonococcal transformation, and Dr. H. S. Seifert also offered many
salient suggestions on goncoccal transformation.
I would also like to thank H. F. Beitenmiller and Dr. A. K. Sample
for their many helpful suggestions regarding pulse-chase labeling

experiments.

10.

11.

164

REFERENCES

. Ahl, T., and J. Reinholdt. 1991. Detection of immunogolobulin A1

protease-induced Fab fragments on dental plaque bacteria. Infect.
Immun. 59:563-569.

. Atkinson.B. A., and H. H. Mulks. 1988. Linkage of genetic markers

in AHU strains of Neisseria gonorrhoeae, p. 223-230. In J. T.
Poolman, et a1,, (ed.), Gonococci and Meningococci. Kluwer Academic
Publishers, Dordrecht, The Netherlands.

. Blake, H. 8., E. J. Lytton, H. E. Seiff, and E. C. Gotschlich.

1988. Studies on gonococcal protein III, p. 85-97. In M. A. Horwitz
(ed.), Bacteria-host cell interactions. Alan R. Liss, Inc., New
York.

. Blanton, K. J., G. D. Biswas, J. Tsai, J. Adams, D. W. Dyer, S. M.

Davis, G. G. Koch, P. K. Sen, and P. F. Sparling. 1990. Genetic
evidence that Neisseria gonorrhoeae produces specific receptors for
transferrin and lactoferrin. J. Bacteriol. 172:5225-5235.

. Bradford, H. 1976. A rapid and sensitive method for the

quantitation of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal. Biochem. 72:248-254.

. Bricker, J., H. H. Mulks, A. G. Plaut, E. R. HOxon, and A. Wright.

1983. IgAl proteases of Haemophilus influenzae: cloning and
characterization in E. coli K-12. Proc. Natl. Acad. Sci, USA
80:2681-2685.

. Clark, V. L., L. A. Campbell, D. A. Palermo, T. M. Evans, and K. W.

Klimpell. 1987. Induction and repression of outer membrane proteins
by anaerobic growth of Neisseria gonorrhoeae. Infect. Immun.
55:1359-1364.

. Clark, V. L., J. S. Knapp. S. Thompson, and K. W. Klimpell. 1988.

Presence of antibodies to the major anaerobically induced
gonococcal outer membrane protein in sera from patients with
gonococcal infections. Microbial Pathogenesis. 5:381-390.

. Cohen, S. N., A. C. Y. Chang, and L. Hsu. 1973. Nonchromosomal

antibiotic resistance in bacteria: genetic transformation of
Escherichia coli by R-factor DNA. Proc. Natl. Acad. Sci. USA.
69:2110-2114.

Cooper, H. D., 2. A. McGee, H. H. Mulks, J. H. Koomey, and T. L.
Hindman. 1984. Attachment to and invasion of human fallopian tube
mucosa by a gonococcal IgAl protease deficient mutant and wild type
parent. J. Infect. Dis. 150:737-744.

Fohn, H. J., T. A. Hietzner, T. W. Hubbard, S. A. Horse, and E. W.
Book III. 1987. Human immunoglobulin G antibody response to the

major gonococcal iron-regulated protein. Infect. Immun. 55:3065-
3069.

12.

13.

14.

15.

l6.

l7.

l8.

19.

20.

21.

22.

23.

165

Gilbert, J. V., and A. G. Plaut. 1983. Detection of IgA protease
activity among multiple bacterial colonies. J. Immunol. Meth.
57:247-251.

Gilbert, J. V., A. G. Plaut, B. Langmaid, and H. E. Lamm. 1983.
Inhibition of microbial IgA protease by human secretary IgA and
serum. Mol. Immunol. 20:1039-1049.

Gorg, A., W. Pastel, J. Heser, H. W. Schivaragin, and W. M. Boesken.
1985. Horizontal SDS electrapharesis in ultrathin pore-gradient gels
for the analysis of urinary proteins. Science Tools (Sweden)
32:5-9.

Halter, R., J. Pohlner, and T. F. Meyer. 1984. IgA protease of
Neisseria ganarrhaeae: isolation and characterization of the gene
and its extracellular product. EMBO J. 3:1595-1601.

Hitchcock, P. J. 1989. Unified nomenclature for pathogenic Neisseria
species. Clin. Microbial. Rev. 2:864-865.

Hook, E. W., III, D. A. Olsen, and T. M. Buchanan. 1984. Analysis
of antigen specificity of the human serum immunoglobulin G immune
response to complicated gonococcal infection. Infect. Immun. 43:

706-709.

Johnston, K. H., and E. C. Gotschlich. 1974. Isolation and
characterization of the outer membrane of Neisseria ganarrhaeae. J.
Bacteriol. 119:250-257.

Keevil, C. W., D. B. Davies, B. J. Spillane, and E. Hahenthiral-
ingam. 1988. Influence of iron-limited and excess continuous culture
on the virulence properties of Neisseria gonorrhoeae. p. 727-729. In
J. T. Poolman, et al. (ed.), Gonococci and Meningococci. Kluwer
Academic Publishers, Dordrecht, The Netherlands.

Kilian, M., J. Mestecky, and M. W. Russell. 1988. Defense mechanisms
involving Fc-dependent functions of immunoglobulin A and their
subversion by bacterial immunoglobulin proteases. Microbiol. Rev.
52:296-303.

Koomey, J. H., R. E. Gill, and S. Falkow. 1982. Genetic and
biochemical analysis of gonococcal IgAl protease: cloning in
Escherichia coli and construction of mutants of gonococci that fail
to produce the activity. Proc. Natl. Acad. Sci. USA 79:7881-7885.

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

Lambden, P. R., J. E. Heckels, L. T. James, and P. J. watt. 1979.
Variations in surface protein composition associated with virulence

properties in opacity types of Neisseria ganarrhaeae. J Gen.
Microbial. 114:305-312.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

166

Lee, B. G., and A. B. Schryvers. 1988. Specificity of the
lactoferrin and transferrin receptors in Neisseria ganarrhaeae. Mol.
Microbiol. 2:827-829.

Haniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular
cloning, a laboratory manual. Cold Spring Harbor, New York.

HcDade, Jr., R. L., and K. H. Johnston. 1980. Characterization of
serologically dominant outer membrane proteins of Neisseria
gonorrhoeae. J. Bacterial. 141:1183-1191.

McNabb, P. C. and T. B. Tomasi. 1981. Host defense mechanisms at
mucosal surfaces. Ann. Rev. Microbiol. 35:477-496.

Mehta, S. K., A. G. Plaut, N. J. Calvanico, and T. B. Tomasi, Jr.
1973. Human immunoglobulin A: production of an Fc fragment by an
enteric microbial proteolytic enzyme. J. Immunol. 111:1274-1276.

Hietzner, T. A., G. H. Luginbuhl, E. Sandstrom, and S. A. Horse.
1984. Identification of an iron-regulated 37,000-dalton protein in
the cell envelope of Neisseria ganarrhaeae. Infect. Immun. 45:410-
416.

Mickelsen, P. A. and P. F. Sparling. 1980. Iron acquisition by the
Neisseria. p. 225-227. In D. Danielsson and S. Normark (ed.),
Genetics and Immunobiology of Pathogenic Neisseria. University of
Umea Press, Umea, Sweden.

Miller, M. A., C. S. Schirli, and J. Mills. 1980. Cytoplasmic
membrane proteins of spectinomycin-susceptible and -resistant
strains of Neisseria ganarrhaeae. J. Bacteriol. 143:198-204.

Horse, S. A., T. A. Hietzner, G. Bolan, A. LeFaou, and G.

Schoolnik. 1987. Characterization of the major iron-regulated
protein of Neisseria ganarrhaeae and Neisseria meningitidis. Antonie
van Leeuwenhoek J. Microbiol. 53:465-469.

Harse, S. A. and L. Bartenstein. 1980. Purine metabolism in
Neisseria gonorrhoeae: the requirement for hypoxanthine. Can. J.
Microbiol. 26:13-20.

Mulks, H. H. 1985. Microbial IgA Proteases, p. 81-104. In I. A.
Holder (ed.), Bacterial Enzymes and Virulence. CRC Press, Inc., Boca
Raton, FL.

Mulks, H. H., and J. S. Knapp. 1985. Immunoglobulin Al protease
types of Neisseria gonorrhoeae, p. 51-56. In G. Schoolnik, (ed.),
The Pathogenic Neisseriae. American Society for Microbiology,
Washington, DC.

Mulks, H. H., S. J. Kornfeld, B. Frangione, and A. G. Plaut. 1982.
Relationship between the specificity of IgA proteases and serotypes
in Haemophilus influenzae. J. Infect. Dis. 146:266-274.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

167

Hulks, M. H., and A. G. Plaut. 1978. IgA protease production as a
characteristic distinguishing pathogenic from harmless Neisseri-
aceae. N. Engl. J. Med. 299:973-976.

Hulks, H. H., A. C. Plaut, H. A. Feldman, and B. Frangione. 1980.
IgA proteases of two distinct specificities are released by
Neisseria meningitidis. J. Exp. Med. 152:1442-1447.

Mulks, H. H., D. A. Simpson, and R. J. Shoberg. 1987. Restriction
site polymorphism in genes encoding type 2 but not type 1 gonococcal
IgAl proteases. Antonie van Leeuwenhoek J. Microbiol. 53:471-478.

Hulks, H. H., D. A. Simpson, and R. J. Shoberg. 1988. Proteolysis of
gonococcal membrane proteins by IgAl proteases, abstr. P20, p. 20.
6th Intl. Pathogenic Neisseria Conf. Pine Mountain, GA. 1988.

Norqvist, A., J. Davies, L. Norlander, and S. Normark. 1978. The
effect of iron starvation on the outer membrane protein composition
of Neisseria gonorrhoeae. FEMS Microbiol. Letters. 4:71-75.

Osborn, H. J., and R. Hunson. 1974. Separation of inner
(cytoplasmic) and outer membranes of gram-negative bacteria. Meth.
Enzymol. 31:642-653.

Plaut, A. G. 1983. The IgAl proteases of pathogenic bacteria. Ann.
Rev. Microbiol. 37:603-622.

Plaut, A. G., J. V. Gilbert, and I. Heller. 1978. Assay and
properties of IgA protease of Streptococcus sanguis. Adv. Exp. Med.
Biol. 107:489-495.

Pohlner, J., R. Halter, K. Beyreuther, and T. F. Heyer. 1987. Gene
structure and extracellular secretion of Neisseria ganarrhaeae IgA
protease. Nature 325:458-462.

Rice, P. A. 1989. Molecular basis for serum resistance in Neisseria
ganarrhaeae. Clin. Microbiol. Rev. 28112-8117.

Sarafian, S. K., H. R. Tam, and S. A. Horse. 1983. Gonococcal
protein I-specific opsonic IgG in normal human serum. J. Infect.
Dis. 148:1025-1032.

Schwalbe, R. S., P. F. Sparling, and J. G. Cannon. 1985. Variation
of Neisseria gonorrhoeae protein II among isolates from an outbreak
caused by a single gonococcal strain. Infect. Immun. 49:250-252.

Shoberg, R.J., and H.H. Mulks. 1990. Effect of media iron
concentration on immunoglobulin Al protease activity in Neisseria
ganarrhaeae, abstr. B-l9, p.29. Abstr. 90th Annu. Meet. Am. Soc.
Microbial., 1990.

Shoberg, R.J., and H.H. Mulks. 1991. Proteolysis of bacterial
membrane proteins by Neisseria gonorrhoeae type 2 IgAl protease.
Infect. Immun. in press. IAI 186-91

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

168

Simpson, D. A., R. P. Hausinger, and H. H. Mulks. 1988.
Purification, characterization, and comparison of the immunoglobulin
Al proteases of Neisseria gonorrhoeae. J. Bacteriol. 170:1866-1873.

Stein, D. C. 1989. Introduction of cloned genes into Neisseria
gonorrhoeae. Clin. Microbiol. Rev. 2:8146-8149.

Stein, D. C., F. E. Young, F. C. Tenover, and V. L. Clark. 1983.
Characterization of a chimeric fl-lactamase plasmid of Neisseria
gonorrhoeae which can function in Escherichia coli. Mol. Gen. Genet.

189277-84.

Stein, D. C., L. E. Silver, V. L. Clark, and F. E. Young. 1983.
Construction and characterization of a new shuttle vector pLESZ,
capable of functioning in Escherichia coli and Neisseria
ganarrhaeae. Gene 25:241-247.

Swanson, J. 1982. Colony opacity and protein II composition of
gonococci. Infect. Immun. 37:359-368.

Swanson, J. 1978. Studies on gonococcus infection. XII. Colony
color and opacity variants of gonococci. Infect. Immun. 19:320-331.

Swanson, J., and J. Heckels. 1980. Proposal: nomenclature of
gonococcal outer membrane proteins, p. xxi-xxvi. In S. Normark and
D. Danielsson (ed.), Genetics and immunobiology of pathogenic
Neisseria. University of Umea Press, Umea, Sweden.

Tam, H. R., T. H. Buchanan, E. G. Sandstrom, K. K. Holmes, J. S.
Knapp, A. W. Siadak, and R. C. Nowinski. 1982. Serological
classification of Neisseria gonorrhoeae with monoclonal antibodies.
Infect. Immun. 36:1042-1053.

Waring, W. S. and C. H. Werkman. 1942. Growth of bacteria in an
iron-free media. Arch. Biochem. 1:303-310.

West, 8. E. H. and P. F. Sparling. 1985. Response of Neisseria
gonorrhoeae to iron limitation: Alterations in expression of
membrane proteins without apparent siderophore production. Infect.
Immun. 47:388-394.

White, L. A., and D. S. Kellogg, Jr. 1965. Neisseria gonorrhoeae
identification in direct smears by a fluorescent antibody-
counterstain method. Appl. Microbiol. 13.171-174.

WOod, S. G., and J. Burton. 1991. Synthetic peptide substrates for
the immunoglobulin Al protease from Neisseria gonorrhoeae (type 2).
Infect. Immun. 59:1818-1822.

Chapter 2

PROTEOLYSIS OF BACTERIAL MEMBRANE PROTEINS BY Neisseria gonorrhoeae

TYPE 2 IgAl PROTEASE

Russell J. Shoberg and Martha H. Mulks

Department of Microbiology and Public Health,
Michigan State University,

East Lansing, MI 48824 USA

Infection and Immunity. 1991. in press.

169

170
ABSTRACT

The IgAl proteases of Neisseria ganarrhaeae have been defined as
having human IgAl as their single permissive substrate. However, in
recent years there have been reports of other proteins which are
susceptible to the proteolytic activity of these enzymes. To examine
the possibility that gonococcal membrane proteins are potential
substrates for these enzymes, isolated outer and cytoplasmic membranes
of N. ganarrhaeae were treated in vitro with exogenous pure IgAl
protease. Analysis of silver stained SDS-PAGE gels of outer membranes
indicated that there were two outer membrane proteins of 78 and 68 kDa
which were cleaved by IgAl protease in vitro, in GCM 740 (a wild type
strain) and in two isogenic IgAl protease negative variants. Similar
results were observed with a second gonococcal strain, F62, and its
isogenic IgAl protease negative derivative. When GCM 740 cytoplasmic
membranes were treated with protease, three minor proteins of 24.5,
23.5, and 21.5 kDa were cleaved. In addition, when outer membranes of
E. coli DHl were treated with IgAl protease, several proteins were
hydrolyzed. While the identities of all of these proteolyzed proteins
are unknown, the data presented indicate that there are several proteins
found in the isolated membranes of gram-negative bacteria which are

permissive in vitro substrates for gonococcal IgAl protease.

171
INTRODUCTION

The enzymatic activity now labelled IgAl protease was first
described in 1973 by Mehta, et a1., when they demonstrated the presence
of a bacterial enzyme in human stool samples which produced intact Faba
and Fca fragments from human immunoglobulin A (19). Since that time,
several bacteria have been reported to produce enzymes with this
activity and these enzymes have been the subject of several review
articles (13,21,30). Bacteria which produce IgAl proteases are
clustered into a broad grouping consisting of human mucosal surface
pathogens. This group includes the gram-negative bacteria Neisseria
gonorrhoeae and N. meningitidis (31), Haemophilus influenzae
(14,17,25), Bacteraides melaninogenicus (11), B. asaccharalyticus
(11), and Gapnocytophaga sp. (11). There are also several gram positive
IgAl protease producers such as Streptococcus pneumoniae (14,17,25), S.
sanguis (32), S. mitis (35), S. aralis (12,35) and Clostridium
ramosum, (6) as well as one mycoplasma, Ureaplasma urealyticum (36).
Because many of the organisms that produce IgAl proteases are mucosal
pathogens of man and the IgAl proteases specifically cleave a major
component of the secretory immune system, the production of IgAl
protease has been postulated to be a virulence factor for these
organisms (13).

It: has been. reported repeatedly' that this group of' enzymes is
specific for a single substrate, the hinge region of human IgAl (21,30).
While a number of relevant human and animal proteins have been tested as
potential substrates, the only permissive substrates for these enzymes
were the IgAl molecules from man and higher primates. Each IgAl

protease cleaves a single peptide bond within the duplicated octet of

172
amino acids that is present in the hinge region of human IgAl and
missing in IgA2 due to a deletion (21,30). N. ganarrhaeae, as well as
N. meningitidis and H. influenzae, produce at least two types of IgAl
proteases as defined by the specific peptide bond cleaved (22,24,26).
Other of the IgAl protease producing species cleave within this same
hinge region but at different peptide bonds (21,30).

The enzymes from, N. ganarrhaeae have been particularly well
studied and the genes encoding them have been cloned by several
laboratories (8,15,23). The entire nucleotide sequence of an iga gene
has been reported as well as a proposed model for the secretion of the
gene product from the gonococcus (33). According to the model, an N-
terminal leader peptide is responsible for secretion of the entire 169
kDa iga2 gene product through the cytoplasmic membrane of the
gonococcus. Next, a 45 kDa carboxyl terminal helper region (5 protein)
mediates attachment to the outer membrane and forms a pore through which
the proenzyme is extruded. Once exposed on the extracellular side of
the outer membrane, the proenzyme releases itself from the outer
membrane via an autoproteolytic event. Finally, in the fluid phase,
IgAl protease makes a second clip in its own sequence to release either
a 12 or 15 kDa peptide (a) of unknown function (33). This report was
the first description of a substrate other than human IgAl for the
gonococcal enzyme. This same research group has subsequently proposed
that the gonococcal IgAl protease may be capable of cleaving two other
human proteins, the granulocyte macrophage growth stimulatory factor
(GH-GSF) and the CD.8 surface antigen of cytotoxic T lymphocytes (34).
These data raise the question that there might be other unrecognized

substrates for the IgAl proteases.

173

We have developed procedures for the purification of both the types
1 and 2 gonococcal IgAl proteases (39). The ability to produce pure
gonococcal type 2 IgAl protease and the construction of an isogenic set
of IgAl protease positive and negative variants of N. ganarrhaeae has
allowed us to examine the question of whether IgAl proteases may be
involved in the modification of the gonococcal cell surface via
proteolysis of gpnacoccal membrane proteins and whether these enzymes
may have more permissive substrates than had previously been described.
We have examined the susceptibility of isolated outer and cytoplasmic
membranes to IgAl protease in an in vitro assay system using both wild
type gonococcal strains and IgAl protease-deficient variants. We also
examined the effect of IgAl protease on the isolated outer membrane of
an E. cali laboratory strain. The data presented here demonstrate that
the type 2 gonococcal IgAl protease is capable of hydrolyzing several
proteins in the outer and cytoplasmic membranes of the gonococcus and
further indicate that this effect is neither strain nor species limited.

(Portions of this work were presented at the 6th International

Pathogenic Neisseria Conference in Pine Mountain, GA,in 1988 [28].)

MATERIALS AND METHODS
Bacterial strains and media. N. ganarrhaeae GCM 740 is a
protatrophic type 2 IgAl protease-producing strain of ParB2 serovar that
has been. described previously (23). The pair of Aiga2 deletion-
disruption mutants, 74OA4 and 740A2 were constructed in this laboratory
for use in other experiments and their construction will be described

below (Figure 1). In GCM 740A2, the deletion removes 1.1 kbp (24% of

174

 

 

 

 

 

 

W

C =
C A 5* ? 7 77'? 1‘ 8"
m7“ } % I I I II I I fl
0 A H B P 7H JA (I:
6074an I I I I Lunfl I 1
c A H H A JG

1
$740A4 1 IL I ------------------ 1 I j

on-

Figure 1. Restriction maps of Neisseria gonorrhoeae GCM 740 and two
isogenic gene disruption mutants. Deleted regions are represented by
dashed lines and both resultant constructs were negative for IgAl
protease activity (restriction site abbreviations: A, AvaI; B', 83111;

C, 0181; H, HindIII; P, PstI).

I

175

the iga gene), and in GCM 740A4, the deletion encompasses 4.3 kbp (93%
of the iga gene). The gonococcal strain F62 and its Iga2- isogen F62-
VDlll, (PorB7, proline requiring) have been described previously by
Koomey, et a1. (15). As well as differing in Par serovar and auxotype,
these two sets of strains have different iga2 genotypes as defined by
restriction site polymorphism (27). The E. coli strains DHl and HBlOl
were used in cloning of the iga2 locus from GCM 740 and in construction
of the deletion-disruption mutants. Actinobacillus pleuropneumoniae
ATCC 27088 is a serotype 1 type strain.

Gonococci were propagated on GC base solid medium (Difco
Laboratories, Detroit, MI) plus Kellogg's defined supplement (43, GCB)
and grown in the defined broth medium of Morse and Bartenstein (20).
The Aiga mutants were isolated after transformation on GCB containing a
diffusion gradient of penicillin G. Following isolation and
confirmation of phenotype, these mutants were propagated on GCB. A.
pleuropneumoniae was propagated on brain heart infusion agar and in
brain heart infusion broth, both containing 10 pg/ml NAD (Difco). E.
coli strains were propagated in LB broth or on LB agar solid medium (18)
containing either ampicillin (100 pg/ml) or tetracycline (12.5 ug/ml) as
necessary. Antibiotics and NAD were from Sigma Chemical Co., St. Louis,
MO.

Construction of Aiga mutants. Cloning of the iga2 gene from GCM 740
has been reported previously (23). Standard recombinant DNA methodology
(18) involving endonuclease restriction and fragment elution from
agarose gels was performed to construct deletion mutants (Figure l). A
2.4 kbp fi-lactamase gene cassette was removed from pLESZ, a N.

ganarrhaeae-E. cali shuttle vector (40) and ligated into these deleted

176

constructs in order to provide a marker which could be used to select
for gonococcal transformants. Resultant Iga-Ampr plasmids were
identified and amplified in .E. cali. Insert DNA. was removed. by
endonuclease digestion and gel purified prior to transformation of GCM
740.

Transformation of GCM 740 was performed as previously described (1)
with the following modifications. Fbllowing the transformation period
(60 minutes at 37°C), cells were plated on GCB for a 6 hour expression

period. at 37°C under 5% CO Following expression, 100 pl of a

2.
penicillin G solution (20 pg/ml) was spread over the inner two-thirds of
the plate and allowed to form a diffusion gradient. Penicillin
resistant colonies were isolated at 72 hours and subcultured on GCB plus
0.025 ug/ml penicillin G to verify the phenotypic stability. Penicillin
resistant colonies were assayed for IgAl protease phenotype by SDS-PAGE
analysis of 125I-labeled IgAl cleavage products (24). Southern blots of
endonuclease restricted genomic DNA also were probed with the 4.3 kbp
HindIII iga fragment and the 2.4 kbp fl-lactamase cassette fragment to
verify the genotype of the mutants (18,27).

Preparation of cytoplasmic and outer membranes. Bacteria were grown
to mid-exponential phase in appropriate broth media and harvested by low
speed centrifugation. Spheraplasts were prepared by the sucrose-
lysozyme-EDTA method (29) and were disrupted by ultrasonication.
Cytoplasmic (CM) and outer (OM) membrane fractions were separated by
ultracentrifugation over isopycnic sucrose density gradients (10).
Protein concentrations of each fraction were determined using the Bio-

Rad protein assay kit with bovine serum albumin as a standard (2, Bio-

Rad Laboratories, Richmond, CA).

177

IgAl protease purification and quantitation. IgAl protease was
purified from GCM 740 as previously described (39). Briefly, cell-free
supernatant from cells grown in defined medium. was harvested and
concentrated by ultrafiltration in a positive pressure filtration cell
(Amican Corp., Lexington, MA). Bulk protein was removed by batch anion
exchange treatment with DE52 at pH 7.5 (Whatman Inc., Clifton, NJ). Two
FPLC chromatographic steps were used to prepare the pure IgAl protease,
a Mono-P chromatofocussing column followed by a Superose 12 gel
filtration column (both Pharmacia, Inc., Piscataway, NJ). After both
column runs, fractions containing IgAl protease activity and silver
stainable IgAl protease were pooled, concentrated, and dialyzed into
Iga2 buffer [50 mM TrisHCl (pH 7.5) + 5 mM EDTA] by ultrafiltration in
a Centricon-30 (Amican Corp.). Determination of protease units,
defined as pg IgAl cleaved minute.1 m1-1, was by quantitative SDS-PAGE
analysis as described previously (32).

IgAl protease treatment of membrane preparations. Three protein
micrograms of each membrane preparation was mixed with one-tenth volume
of 10X IgAl protease assay buffer [0.5 M TrisHCl (pH 7.5), 0.1 M MgClz,
0.1 M CaClz] and from 0.17-4.0 units of IgAl protease per microgram of
membrane protein. In control reactions, either Iga2 buffer or heat-
inactivated IgAl protease was added instead of active enzyme.
Incubation was at 37°C overnight and the reaction was terminated by
addition of one-half volume of 2X sample buffer (16) with boiling for
five minutes. Reaction mixtures were separated by discontinuous system
SDS-PAGE (16) over a 7.5-12.5% linear polyacrylamide gradient followed
by dual Coomassie blue-silver stain (7). Wet gels were analyzed with

transillumination to detect the presence of bands which had changed in

178
staining intensity. A decrease in staining intensity was assumed to be
due to proteolysis and a resultant decrease in the concentration of
protein at that position since the buffer control lane had been treated
in an identical fashion save the presence of the protease. An increase
in the staining intensity of a protein band was assumed to be due to an
increase in the concentration of protein at that position.

For control experiments utilizing heat-inactivated IgAl protease,
the enzyme was inactivated by heating at 95°C for ten minutes. A
nonheated sample with an equivalent volume of the same preparation was
used as the active enzyme control.

Antiserum preparation and immunoblot analysis. A polyclonal
antiserum was prepared in New Zealand white rabbits by standard
procedures. Injections were subcutaneous using electrophoretically pure
IgAl protease (prepared as described above) in Freund's adjuvant (Sigma)
as the antigen. Antisera titers were determined by assaying dilutions
of the antiserum for inhibition of IgAl protease activity in an SDS-PAGE
assay (32). This antiserum preparation gives 50%-inhibition of IgAl
protease activity at a dilution of 1:595.

Western blot analysis was used in an attempt to determine whether
the low molecular weight bands that increased in staining intensity were
due to degradation of the exogenously supplied IgAl protease by
intrinsic membrane proteases. Following SDS-PAGE as described above,
proteins were electroblotted onto a nitrocellulose membrane (BA 85;
Schleicher & Schuell, Inc., Keene, NH) by the method of Towbin, et a1.
(42) for sixty minutes at 100 V, constant voltage. Filters were blocked
with 5% powdered skim milk in PBS-Tween 20 (0.05%) and probed overnight

with anti-IgAl protease antiserum. Washes were in PBS-Tween 20 and

179
12"SI-labelled Protein A was used to detect antigen-antibody complexes
with exposure to Kodak X-OMAT AR film (Eastman Kodak Co. , Rochester,
NY).

Human serum proteins. Human serum albumin (Cohn fraction V) and
human gamma globulins (purified from Cohn fractions 11 and III) were
purchased from Sigma. The human IgAl and IgAZ paraproteins (Mar and
Mapp) were purified from patient sera by standard procedures.

Digestion of human serum proteins. Aliquots of purified type 2 IgAl
protease from Neisseria ganarrhaeae GCM 740 (0.5 units) were incubated
overnight with 0.25 pg of IgAl (Mar), 0.25 pg of IgA2 (Mapp). 0.50 pg of
gamma globulins, or 0.25 pg of albumin in a reaction mix containing 50
mM Tris-H01 (pH 7.5), 10 mM MgC12, and 10 mM CaClz. Reactions were
terminated by addition of one-half volume of double strength SDS-PAGE
sample buffer (16) and boiled for five minutes. Reaction products were
resolved over either 9.0% or 7.5-12.5% SDS-PAGE gels, stained, and

analyzed with transillumination as described above for any degradation

which had occurred in the presence of the added IgAl protease.

RESULTS

Specificity of purified IgAl protease preparation. The purified IgAl
protease preparation was demonstrated to be specifically active against
a human IgAl (Mar) paraprotein and to lack proteolytic activity against:
1) a human IgA2 (Mapp) paraprotein, 2) a human gamma globulin
preparation, and 3) human serum albumin following overnight incubation
(data not shown).

Incubation of isolated outer membranes with exogenous IgAl protease.

Outer membranes of GCM 740 and its two isogenic IgAl protease negative

180
derivatives, 740A2 and 740A4, were incubated with electrophoretically
pure type 2 gonococcal IgAl protease prepared from GCM 740 to determine
whether this enzyme could proteolyze gonococcal OM proteins. Figure 2
shows a typical silver stained SDS-PAGE gel comparing the protein
composition of OHS incubated either with or without exogenous pure
protease.

Two proteins of 78 and 68 kDa were proteolyzed, i. e. decreased in
staining intensity, in all three of the variants in the COM 740 set
(Figure 2; lanes 4, 6, and 8). In contrast, an increase in staining
intensity was observed for bands of 61, 38, 24.5, and 14 kDa in all
three of these variants. In addition, a band of 88 kDa which was
present in the wild type (lanes 3 and 4) and missing in both mutants in
the absence of exogenous protease, was observed to increase in staining
intensity for bath of the mutant variants (lanes 6 and 8). Control
experiments where OM were incubated with either buffer alone (Figure 2,
lanes 3, 5, and 7) or with heat-inactivated IgAl protease (data not
shown) indicated that the proteolytic activity observed was heat-labile
and associated. with the addition of electrophoretically pure IgAl
protease. Immunoblot analysis with anti-IgAl protease antiserum
demonstrated that the lower molecular weight bands which appeared post-
IgAl protease treatment were not degradation products of the added IgAl
protease (data not shown).

Based on relative staining intensity, all of the proteins which were
affected by the addition of protease were minor components of the OM.
The identities of these protein bands are not known. However, it was
determined by SDS-PAGE analysis and with heat and reduction modification

experiments that they did not correspond to some of the previously

181

Figure 2. Silver stained SDS-PAGE gel of N. ganarrhaeae OMS following
incubation. with and without exogenous pure gonococcal type 2 IgAl
protease. Three micrograms of OM proteins were incubated for 14 hours
at 37°C with either Iga2 buffer or 12 protease units of IgAl protease in
Iga2 buffer. Following incubation, mixtures were boiled for five
minutes in sample buffer and separated over a linear 7.5-12.5% gradient
SDS-PAGE gel. Lanes: 1, 12 units of IgAl protease; 2, 12 units of IgAl
protease after 14 hours of incubation at 37°C; 3 and 4, GCM 740; 5 and
6, GCM 740A4; 7 and 8, GCM 740A2; 9 and 10, F62; 11 and 12, F62-VD111.
Odd numbered lanes contain OMs incubated with Iga2 buffer and even
numbered lanes contain OMs incubated with protease. Down-pointed
triangles indicate bands which decreased in intensity following
incubation with protease and up-pointed triangles indicate bands which
increased in intensity. Positions of molecular weight standards (Bio-
Rad) are shown along the left margin in kilodaltons as are the major

gonococcal OM proteins (Por and Rmp) and IgAl protease (a).

 

 

 

 

182

Figure 2

123456789101112

#‘k'k

a

* -7- —6- :fi*:
116
0

97 -2-;; ..

O"
-
no

' ...-

Ill U! f:
m

43

For
Rmp

 

 

 

 

 

183
described gonococcal OMPs or antigens; i. e. they were not For (PI), Rmp
(P111), or Hmps (PIIs) based an apparent molecular weights (9,41).

To determine whether prateolysis of gonococcal OML proteins was
strain specific, outer membranes of N. ganarrhaeae F62, a type 2a IgAl
protease producer (27), and its isogenic protease-negative mutant, F62-
VDlll, were incubated with purified type 2b protease from GCM 740.

When the F62 set was analyzed, the two variants had a single
protealyzed band of 48 kDa in common (Figure 2, lanes 10 and 12).
Protein bands of 24.5 and 14 kDa increased in staining intensity for
both of these variants and these corresponded to bands of the same
molecular weight also observed to increase in the GCM 740 set. The IgAl
protease deficient mutant, F62-VD111, exhibited a band of 58 kDa which
increased in staining intensity following treatment with protease
(Figure 2, lane 12). As stated above, the identities of these bands
were not determined but some identities were ruled out based an apparent
molecular weight and behavior.

Incubation of isolated cytoplasmic membranes with exogenous IgAl
protease. These experiments were performed in a similar fashion to
those with the OMs to determine whether the observed prateolysis was
limited to OM proteins. We found that there were also CM proteins which
altered in staining intensity following incubation with exogenous, pure
IgAl protease (Figure 3; lanes 4, 7, and 10). Control reactions
incubated. with 'buffer alone or ‘with heat-inactivated IgAl protease
demonstrated that this prateolysis was heat labile and dependent upon
the addition of active IgAl protease (Figure 3, lanes 3, 6, and 9).

When the GCM 740 isogenic set was analyzed (Figure 3, lanes 2-10),

three protein bands of 24.5, 23.5, and 21.5 kDa were protealyzed and two

184

Figure 3. Silver stained SDS-PAGE gel of N. ganarrhaeae CMs following
incubation with Iga2 buffer, heat-inactivated type 2 gonococcal IgAl
protease in Iga2 buffer, and active IgAl protease in Iga2 buffer. Three
micrograms of CM proteins were incubated for 13.5 hours at 37°C as
described below. Following incubation, mixtures were boiled for 5
minutes in sample buffer and separated over a linear 7.5-12.5% gradient
SDS-PAGE gel. Lanes: 1, 6 units of IgAl protease after 13.5 hours of
incubation at 37°C; 2, 3, and 4, GCM 740; 5, 6, and 7, GCM 740A4; 8, 9,
and 10, GCM 740A2. Lanes 2, 5, and 8 contain CMs plus buffer; 3, 6, and
9 contain CMs plus inactivated protease (6 protease unit equivalents);
and lanes 4, 7, and 10 contain CMs plus 6 units of protease (2 units per
microgram of CMP). Down-pointed triangles indicate bands which
decreased in intensity following incubation with protease and up-pointed
triangles indicate bands which increased in intensity. Positions of
molecular weight standards (Bio-Rad) are shown along the left margin in

kilodaltons as is IgAl protease (o).

 

Figure 3

 

200

' 116

66

43

31

22

I

185

2 3 4 ,5. 6 7 8,__9JO,_

‘

44

III“ I IIIIIII '

1 Hi‘ H

mm; nu u
m: 1
”I IIIIIIII

Ct
Ill
:37:
.
GI,
n
.il

17.1: “1111..

b

IIIIEHI Illl'.

TI? izlfili‘l'fl mm“.

alt-u um.
um: unit
tune,

I

'I"

“ “"-~ V -o 0"
—-— —- —-'
~V —~v -" "V
—"-V ——V-—-- —V--

 

186
protein bands of 57 and 20.5 kDa were observed to increase in intensity
in all three variants. The two mutants, GCM 74OA4 and 740132, also
shared a protein band of 121 kDa which was protealyzed (lanes 7 and 10).
Strain GCM 74OA4 exhibited two protein bands of 107 and 39 kDa which
were protealyzed (lane 7); these protein bands were not observed for
either GCM 740 or GCM 740A2.

When analyzed in a similar experiment, the F62 set of variants
showed a single CM protein band of 45 kDa in both the wild type and the
mutant which increased in staining intensity following protease
treatment (data not shown).

Incubation of E. cali outer membranes with exogenous IgAl protease.
To determine whether the observed prateolysis was restricted to gpno-
coccal OM proteins, OMs of two other gram negative bacteria, E. coli DHl
and Actinobacillus pleuropneumoniae, were incubated with purified type 2
gonococcal IgAl protease. Several E. coli DHl OM proteins changed in
staining intensity following incubation with protease (Figure 4).

E. coli DHl OM bands which were protealyzed included protein bands
of 177, 90, 43, 38.5, 27.5, 22.5, and 19.7 kDa. Two minor bands of 31
and 28.5 kDa increased in staining intensity. A. pleuropneumoniae OMs
exhibited a band of 37 kDa which was cleaved and two bands of 33 and 25
kDa that increased in staining intensity (data not shown).

While the identities of these protein bands are unknown, the fact
that some protein ‘bands are observed to be degraded ‘while others
increase in staining intensity supports the hypothesis that the type 2
gonococal IgAl protease has the ability to cleave OM proteins of gram

negative species in addition to the gonococcus.

187

Figure 4. Silver stained SDS-PAGE gel of E. coli DHl OMs following
incubation ‘with and. without exogenous pure type 2 gonococcal IgAl
protease. Three micrograms of OM proteins were incubated for 14 hours
at 37°C with either Iga2 buffer or IgAl protease (0.17 protease units
per microgram of OMS) in Iga2 buffer. Following incubation, mixtures
were boiled for 5 minutes in sample buffer and separated over a linear
7.5-12.5% gradient SDS-PAGE gel. Lane 1, OMs plus Iga2 buffer and lane
2, OMs plus protease. ‘Down-pointed triangles indicate 'bands ‘which
decreased in intensity following incubation with protease and up-pointed
triangles indicate bands which increased in intensity. Positions of
molecular weight standards (Bio-Rad) are shown along the left margin in

kilodaltons as is IgAl protease (o).

 

188

Figure 4

 

 

189
DISCUSSION

In this study, we have demonstrated that the type 2 IgAl protease of
Neisseria gonorrhoeae is capable of cleaving proteins found in both the
outer and cytoplasmic membranes of the gonococcus, and also of cleaving
outer membrane proteins of other gram-negative bacteria. This
protealytic activity' was associated. with a silver stainable,
electrophoretically' pure, heat labile preparation of IgAl protease.
Proteolysis of bacterial membrane proteins, or of any substrate of
bacterial origin other than the protease itself, by any IgAl protease,
has not been reported previously.

Our conclusion that the phenomenon observed, i.e., the disappearance
of some proteins and appearance of others in membrane preparations
incubated with IgAl protease, is due to specific prateolysis of membrane
proteins by this enzyme, is supported by several control experiments.
First, this phenomenon requires addition of active IgAl protease;
incubation of OM or CM preparations with heat-inactivated protease led
to no alteration of proteins. Second, appearance of new protein bands
was not due to autodegradation of the IgAl protease, as evidenced by the
integrity of the pure protease post-incubation overnight at 37°C (Figure
2, lanes 1 and 2). Third, appearance of new proteins was not due to
degradation of the added protease by any hydrolytic activity present in
the membrane preparations; immunoblot experiments demonstrated no bands
other than. the 109 kDa intact protease reactive with. monospecific
antiserum against IgAl protease. Fourth, the lack of non-specific,
contaminating protealytic activity in the purified IgAl protease
preparation was demonstrated by the lack of digestion of some relevant

human serum proteins, including human IgA2.

190

We constructed Iga- isogenic variants specifically to allow us to
examine membrane proteins produced in the absence of protease as
potential substrates. The ability to isolate and propagate these IgAl
protease deficient N. ganarrhaeae variants indicates that neither the
ability to produce IgAl protease nor the events of enzyme production and
secretion are required for in vitro growth of the gonococcus. Our Iga-
mutant variants grow equivalently to the WT in standard broth media and
are capable of growing anaerobically and under iron limitation, and of
expressing the appropriate previously defined stress response OM
proteins (4,5,37,38). However, we did observe some protein profile
differences between the WT and Iga- isogenic variants (Figure 2, lanes
3, 5, and 7) (28,38). We initially hypothesized that these differences
would be ”cured” by the addition of protease to the membrane
preparations. This did occur in the case of an 88 kDa 0M protein,
present in the WT GCM 740 but not in the mutants in the absence of
protease, and found in both mutants after incubation with active
protease. However, in general the proteins which differed between WT
and mutant variants were not apparent substrates for the protease.

Most of the altered proteins, whether in the outer or cytoplasmic
membrane, were found in all members of an isogenic set. This is
strongly supportive of specific prateolysis of unique common substrate
proteins rather than generalized degradation. The presence of IgAl
protease susceptible proteins in the OM and CM of WT gonococcal strains
(including both GCM 740 and F62) which produced active enzyme, as well
as in the Iga- mutants, may be due to the in vitra accessibility of both
sides of the OM and CM. In contrast, in an in viva situation the active

form of IgAl protease, as measured against human IgAl, is restricted to

191

the extracellular fluids (33). However, during the process of
secretion, the intact IgAl protease proenzyme (enzyme plus a and fl
peptides) does exist in the periplasmic space, and therefore has access
to the periplasmic sides of both the OM and CM. While it has been
reported that this intermediate form of the enzyme has no activity
against human IgAl, it is not known whether the proenzyme has
protealytic activity against any other substrate (33). It is therefore
possible that prateolysis of membrane proteins not exposed on the cell
surface could occur in vivo.

We do not know the specific identities of any of these membrane
proteins that are degraded by IgAl protease. However, we do know that
none of these are among the previously well described gonococcal major
outer membrane proteins; i.e., not Par, Rmp, or Hmp. We also have
examined the susceptibility of defined iron limitation and anaerobiosis
induced outer membrane proteins, and have found that these stress
induced proteins also are not substrates for IgAl protease hydrolysis
(38). The data does suggest that some of the substrate proteins may be
conserved gonococcal membrane proteins, since degradation products of
the same apparent molecular weight were observed in all five variants
examined (Figure 2).

Alteration of protein bands in both the GCM 740 and F62 sets of
strains by protease purified from GCM 740 indicated that the prateolysis
of membrane proteins is not strain specific (Figure 2). This
protealytic effect is also not restricted by species, as demonstrated in
Figure 4, which shows that OM proteins of E. coli DHl are susceptible
to cleavage by the exogenously added IgAl protease. Similar results

were observed when isolated OMs of a porcine pathogen, Actinobacillus

192
pleurapneumaniae, were treated with IgAl protease.

The fact that there are proteins in the OM of the gonococcus that
are permissive substrates for the IgAl protease prompts the question as
to whether the enzyme performs this function in viva. If membrane
proteins were processed in viva by IgAl protease, one would expect that
either there would be proteins found in the mutants (and not in the WT)
which would be degraded by added protease, or that there would be
proteins absent in the mutants (and found in the WT) that would appear
after incubation with added protease. The 88 kDa OM protein present in
WT GCM 740 and absent in the mutants, which appears post-IgAl protease
treatment of the mutants, is suggestive that this type of processing of
membrane proteins by IgAl protease does occur in viva.

In summary, the data presented here indicate that there are proteins
present in the outer and cytoplasmic membranes of N. ganarrhaeae that
are permissive in vitra substrates for gonococcal type 2 IgAl protease.
Proteolysis of these membrane proteins was dependent on the addition of
active IgAl protease and was neither strain nor species restricted.
This report defines an entire range of new potential substrates for the
IgAl proteases, whose role in the physiology of the organism or the

pathogenesis of gonococcal infection remains to be determined.

193

ACKNOWLEDGEMENTS
This work was supported by Public Health Service grant AI-21264 from
the National Institute for Allergy and Infectious Diseases to M. H. M.
and a Michigan State University College of Natural Sciences Continuing
Doctoral Fellowship to R. J. S.
The provision of a purified human IgA2 paraprotein (Mapp) by Dr. A.
G. Plaut and the N. ganarrhaeae strain F62 and its isogenic variant,

VDlll by Dr. M. D. Cooper are gratefully acknowledged.

10.

ll.

12.

194

REFERENCES

. Atkinson, B. A., and H. H. Mulks. 1988. Linkage of genetic markers

in AHU strains of Neisseria ganarrhaeae, p. 223-230. In J. T.
Poolman, et al., (ed.), Gonococci and Meningococci. Kluwer
Academic Publishers, Dordrecht, The Netherlands.

. Bradford, H. 1976. A rapid and sensitive method for the

quantitation of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal. Biochem. 72:248-254.

. Bricker, J., H. H. Mulks, A. G. Plaut, E. R. Hoxon, and A. Wright.

1983. IgAl proteases of Haemaphilus influenzae: cloning and
characterization in E. cali K-12. Proc. Natl. Acad. Sci, USA
80:2681-2685.

. Chen, C.-Y., C. A. Genco, J. P. Rock, and S. A. Horse. 1989.

Physiology and metabolism of Neisseria ganarrhaeae and Neisseria
meningitidis: implications for pathogenesis. Clin. Microbiol. Rev.
2:835-840.

. Clark, V. L., L. A. Campbell, D. A. Palermo, T. M. Evans, and K. W.

Klimpell. 1987. Induction and repression of outer membrane proteins
by anaerobic growth of Neisseria ganarrhaeae. Infect. Immun.
55:1359-1364.

. Fujiyama, Y., K. Kobayashi, S. Senda, Y. Benno, T. Bamba, and S.

Hosoda. 1985. A novel IgA protease from Clastridium sp. capable of
cleaving IgAl and IgA2m(l) allotype but not IgA2m(2) allotype
paraproteins. J. Immunol. 134:573-576.

. Gorg, A., W. Pastel, J. Weser, H. W. Schivaragin, and W. H. Boesken.

1985. Horizontal SDS electrophoresis in ultrathin pore-gradient gels
for the analysis of urinary proteins. Science Tools (Sweden) 32:5-9.

. Halter, R., J. Pohlner, and T. F. Meyer. 1984. IgA protease of

Neisseria ganarrhaeae: isolation and characterization of the gene
and its extra-cellular product. EMBO J. 3:1595-1601.

. Hitchcock, P. J. 1989. Unified nomenclature for pathogenic Neisseria

species. Clin. Microbiol. Rev. 2:S64-S65.

Johnston, K. H., and E. C. Gotschlich. 1974. Isolation and
characterization of the outer membrane of Neisseria ganarrhaeae. J.
Bacteriol. 119:250-257.

Kilian, H. 1981. Degradation of immunoglobulins Al, A2, and G by
suspected principal periodontal pathogens. Infect. Immun. 34:757-
765.

Kilian, H., and K. Holmgren. 1981. Ecology and nature of
immunoglobulin Al protease-producing streptococci in the human oral
cavity and pharynx. Infect. Immun. 31:868-873.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

195

Kilian, H., J. Mestecky, and H. W. Russell. 1988. Defense mechanisms
involving Fc-dependent functions of immunoglobulin A and their
subversion by bacterial immunoglobulin proteases. Microbiol. Rev.
52:296-303.

Kilian, H., J. Mestecky, and R. E. Schrohenloher. 1979. Pathogenic
species of the genus Haemaphilus and Streptococcus pneumoniae
produce immunoglobulin Al protease. Infect. Immun. 26:143-149.

Koomey, J. H., R. E. Gill, and S. Falkow. 1982. Genetic and
biochemical analysis of gonococcal IgAl protease: cloning in
Escherichia coli and construction of mutants of gonococci that fail
to produce the activity. Proc. Natl. Acad. Sci. USA 79:7881-7885.

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

Male, C. J. 1979. Immunoglobulin A1 protease production by
Haemaphilus influenzae and Streptococcus pneumoniae. Infect. Immun.
26:254-261.

Haniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular
cloning, a laboratory manual. Cold Spring Harbor, New Yark.

Mehta, S. K., A. G. Plaut, N. J. Galvanico, and T. B. Tomasi, Jr.
1973. Human immunoglobulin A: production of an Fc fragment by an
enteric microbial protealytic enzyme. J. Immunol. 111:1274-1276.

Horse, S. A., and L. Bartenstein. 1980. Purine metabolism in
Neisseria ganarrhaeae: the requirement for hypoxanthine. Can. J.

Microbiol. 26:13-20.

Mulks, H. H. 1985. Microbial IgA Proteases, p. 81-104. In I. A.
Holder (ed.), Bacterial Enzymes and Virulence. CRC Press, Inc., Boca
Raton, FL.

Hulks, H. H., and J. S. Knapp. 1983. Distribution of IgAl protease
types in Neisseria ganarrhaeae, abstr. 12, p. 44. Program and abstr.
5th. Intl. Meet. Intl. Soc. for Sexually Transmitted Dis. Res. 1983.

Mulks, H. H., and J. S. Knapp. 1985. Immunoglobulin Al protease
types of Neisseria ganarrhaeae, p. 51-56. In G. Schoolnik, (ed.).
The Pathogenic Neisseriae. American Society for Microbiology,
Washington, DC.

Mulks, H. H., S. J. Kornfeld, B. Frangione, and A. G. Plaut. 1982.
Relationship between the specificity of IgA proteases and serotypes
in Haemaphilus influenzae. J. Inf. Dis. 146:266-274.

Hulks, H. H., S. J. Kornfeld, and A. G. Plaut. 1980. Specific
prateolysis of human IgA by Streptococcus pneumoniae and Haemaphilus
influenzae. J. Inf. Dis. 141:450-456.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

196

Mulks, H. H., A. G. Plaut, H. A. Feldman, and B. Frangione. 1980.
IgA proteases of two distinct specificities are released by
Neisseria meningitidis. J. Exp. Med. 152:1442-1447.

Mulks, H. H., D. A. Simpson, and R. J. Shoberg. 1987. Restriction
site polymorphism in genes encoding type 2 but not type 1 gonococcal
IgAl proteases. Antonie van Leeuwenhoek J. Microbiol. 53:471-478.

Mulks, H. H., D. A. Simpson, and R. J. Shoberg. 1988. Proteolysis of
gonococcal membrane proteins by IgAl proteases, abstr. P20, p. 20.
6th Intl. Pathogenic Neisseria Conf. Pine Mountain, GA. 1988.

Osborn, H. J., and R. Hanson. 1974. Separation of inner
(cytoplasmic) and outer membranes of gram-negative bacteria. Meth.
Enzymol. 31:642-653.

Plaut, A. G. 1983. The IgAl proteases of pathogenic bacteria. Ann.
Rev. Microbiol. 37:603-622.

Plaut, A. G., J. V. Gilbert, H. S. Artenstein, and J. D. Capra.
1975. Neisseria ganarrhaeae and Neisseria meningitidis:
extracellular enzyme cleaves human immunoglobulin A. Science

190:1103-1105.

Plaut, A. G., J. V. Gilbert, and I. Heller. 1978. Assay and
properties of IgA protease of Streptococcus sanguis. Adv. Exp. Med.
Biol. 107:489-495.

Pohlner, J., R. Halter, K. Beyreuther, and T. F. Meyer. 1987. Gene
structure and extracellular secretion of Neisseria ganarrhaeae IgA
protease. Nature 325:458-462.

Pohlner, J., R. Halter, and T. F. Meyer. 1988. Neisseria ganarrhaeae
IgA protease: secretion and implications for pathogenesis, p. 427-
432. In J. T. Poolman, et al., (ed.), Gonococci and Meningococci.
Kluwer Academic Publishers, Dordrecht, The Netherlands.

Reinholdt, J., H. Tomana, S. B. Hortensen, and H. Kilian. 1990.
Molecular aspects of immunoglobulin A1 degradation by oral
streptococci. Infect. Immun. 58:1186-1194.

Robertson, J. A., H. E. Stemler, and G. W. Stemke. 1984.
Immunoglobulin A protease activity of Ureaplasma urealyticum. J.
Clin. Microbiol. 19:255-258.

Shoberg, R.J., and H.H. Mulks. 1990. Effect of media iron
concentration on immunoglobulin A1 protease activity in Neisseria
ganarrhaeae, abstr. B-19, p.29. Abstr. 90th Annu. Meet. Am. Soc.
Microbial., 1990.

Shoberg, R.J., and H. H. Mulks. Unpublished data.

39.

40.

41.

42.

43.

197

Simpson, D. A., R. P. Hausinger, and H. H. Hulks. 1988.
Purification, characterization, and comparison of the immunoglobulin
Al proteases of Neisseria ganarrhaeae. J. Bacteriol. 170:1866-1873.

Stein, D. G., L. E. Silver, V. L. Clark, and F. E. Young. 1983.
Construction and characterization of a new shuttle vector pLESZ,
capable of functioning in Escherichia cali and Neisseria
ganarrhaeae. Gene 25:241-247.

Swanson, J., and J. Heckels. 1980. Proposal: nomenclature of
gonococcal outer membrane proteins, p. xxi-xxvi. In S. Normark and
D. Danielsson (ed.), Genetics and immunobiology of pathogenic
Neisseria. Proc. EMBO Workshop, Hemavan, Sweden 1980.

Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic
transfer of proteins and nucleic acids from polyacrylamide gels to
nitrocellulose sheets: procedure and some applications. Proc. Natl.
Acad. Sci. USA 76:4350-4354.

White, L. A., and D. S. Kellogg, Jr. 1965. Neisseria ganarrhaeae
identification in direct smears by a fluorescent antibody-
counterstain method. Appl. Microbiol. 13.171-174.

Chapter 3

EXPRESSION OF THE Neisseria ganarrhaeae TYPE 2 IgAl PROTEASE

GENE UNDER IRON LIMITED CONDITIONS

Russell J. Shoberg and Martha H. Mulks

Department of Microbiology and Public Health
Michigan State University

East Lansing, MI 48824-1101

198

199

ABSTRACT

The effect of iron limitation on Neisseria ganarrhaeae and its
expression of IgAl protease activity was examined by growing bacteria
under conditions of iron limitation and iron excess. When bacteria were
grown in an iron limited medium, there was an initial burst of high IgAl
protease activity in the supernatant, which was more than double that
observed with cells grown in an iron replete medium and over five-fold
more than cells that were grown in a high iron medium. This effect was
consistently observed regardless of the chemical means of iron
limitation. The activity of the mature enzyme was not affected by
either the medium iron concentration or the type of medium used for
bacterial propagation. The existence of a well conserved nucleotide
sequence homology to an Escherichia cali operator fur binding site in
the promoter of the iga2 gene is presented. However, when the cloned N.
ganarrhaeae iga2b gene was transformed into a set of Escherichia cali
Fur+/- isogenic strains, fur-mediated regulation of the iga2b gene in E.
cali was not observed. Southern analysis failed to detect a chromosomal
fur homolog in N. ganarrhaeae when the cloned E. cali fur gene was used

as a probe. These apparent discrepancies are discussed further.

200
INTRODUCTION

The ability to produce the enzymes referred to as IgAl proteases is
a phenotype shared among a relatively small set of bacteria having a
number of other characteristics in common, such as being human pathogens
which initiate infections of the mucosal surfaces (37,45).
Representative IgAl protease producing bacteria include: Neisseria
ganarrhaeae (46), N. meningitidis (46), Haemaphilus influenzae
(25,29,38), Streptococcus pneumoniae (25,29,38), S. sanguis (47), S.
aralis (23,49), S. mitis (49), Clastridium ramasum (l6), Bacteraides
meIaninagenicus (23), B. asaccharalyticus (23), Capnacytaphaga spp.
(23), and Ureaplasma urealyticum (50).

Production of an IgAl protease has long been postulated to be a
virulence factor for these microbes and there is a body of indirect
evidence for this (see 24 for a review of the subject). However, since
many of these microbes are obligate human pathogens and lack a bana fide
animal model system (e. g. , N. ganarrhaeae) and since the "relevant"
substrate for IgAl protease is human IgAl, convincing experiments have
yet to be performed.

The sites at which N. ganarrhaeae may invade and inhabit a host
include the mucosal surfaces of the urogenital tract, rectum, pharynx,
and conjunctiva in an uncomplicated gonococcal infection; as well as the
bloodstream, synovial fluid, and cerebrospinal fluid during cases of
diseminated gonococcal infection. The glycoprotein lactoferrin is
secreted on the mucosal surfaces of the host where it functions as a
nonspecific, antimicrobial mechanism by reducing the availability of
free Fe3+ (31). Evidence that the urogenital tract is a low iron

environment for bacterial growth came from the work of Shand, et a1.,

201

who demonstrated that bacteria isolated directly from a patient with a
urinary tract infection were expressing iron repressible outer membrane
proteins (52). In the vascular system and interstitial fluids,
transferrin, an iron transport glycoprotein is secreted, where it also
functions as a nonspecific, antimicrobial mechanism. Therefore,
regardless of the site of colonization, the gonococcus resides in an
environment with little available iron, and must respond to that
metabolic stress. Most strains of gonococci are able to acquire and
utilize iron bound to the following human organic iron sources:
transferrin, lactoferrin, hemin, and hemoglobin (59). N. ganarrhaeae
does not produce siderophores (low molecular weight iron acquisition
molecules) but interestingly, has been demonstrated to assimilate iron
bound to the siderophore aerobactin (59).

A number of putative gonococcal virulence functions that are
regulated by the in vitra medium concentration of iron have been
reported. These include the Fbp (Ferric binding protein), a 37,000
dalton major outer membrane protein which has been demonstrated to bind
iron reversibly (4,34), and the separate outer membrane protein
receptors specific for human transferrin and lactoferrin (6,28). The
expression of these proteins has been demonstrated to be regulated in a
coordinate fashion by the limited availability of iron in the medium
(34,159). There are also a number of other iron-repressible outer
membrane proteins (1. e. induced by the limitation of iron in the
medium) that have been described for N. ganarrhaeae but most of these
have no described function as yet (22,43,59). The subset of iron
repressible outer membrane proteins expressed by a given strain has been

demonstrated to be dependent upon the mode of iron limitation (59).

202
Most of these have been demonstrated to be expressed in viva as well,
evidenced by the host's immune response to the protein (15a).

Several bacterial toxins are regulated by iron concentrations
including diphtheria toxin of Corynebacterium diphtheriae (57), exotoxin
A produced by Pseudamanas aeruginasa (5), the Shigella dysenteriae Shiga
toxin (13), shiga-like toxin produced by strains of Escherichia cali
(8), and the hemolysins of Vibria chalerae (56) and Listeria
manocytagenes (11). Also under iron regulation are several bacterial
siderophore iron acquisition systems (12) and a variety of bacterial
surface proteins (42).

There are two major described mechanisms by which bacteria regulate
gene expression with regard to iron-mediated expression. The first
mechanism, used by a variety of bacteria, is the Fur transcriptional
repression system as first described in Salmonella typhimurium (14) and
further characterized by Hantke (19,20). According to the model, the
repressor exists as a Fur homodimer in the cytosol. In the presence of
sufficient cytoplasmic ferrous iron, the Fur dimer complexes with Fe2+,
subsequently binding to an operator sequence (the "iron box") in the
promoter region of Fur regulated genes, therein preventing the
association of RNA polymerase with these genes and resulting in a lack
of gene expression (3). To date, Fur regulatory systems have been
proposed and described in such varied organisms as Yersinia pestis (55),
Vibria cholerae (8), and Carynebacterium diphtheriae (7). The second
described mechanism involves the cooperative transcriptional activation
of regulatory genes governing the synthesis and uptake of the Vibria

anguillarium siderophore, anguibactin, by the trans-acting factors Taf

and AngR (12 , 51) .

203

Since a number of N. ganarrhaeae genes are known to be regulated by
the availability of ferric -iron, and since a number of' microbial
virulence factors are also regulated by iron limitation, we felt that it
would be prudent to study the effects of low iron concentration on the
expression of IgAl protease. Since we had evidence that both wild type
and Iga- gonococci could respond to iron limitation and express the
appropriate, previously described iron regulable outer membrane proteins
(Chapter 1 of this thesis), we concluded that the ability to produce
IgAl protease was not required to respond to a low iron stress. Since
the IgAl proteases have been postulated and presented as virulence
factors (24,37,45), we wondered. whether they might be induced or
regulated in some fashion by the available iron concentration.

To examine the effect(s) of iron on the activity of the Neisseria
ganarrhaeae IgAl protease gene we studied it in comparison to the well
described fur regulatory system. We measured enzyme activity expressed
under controlled conditions of iron limitation and iron excess in the
gonococcus, and in a characterized E. cali Fur background to examine
the effect of the fur locus on expression of the cloned iga2 gene. We
also used the cloned E. cali fur gene as a molecular probe to search
for a gonococcal fur homologue.

We present evidence that the type 2 IgAl protease of N. ganarrhaeae
is influenced by the medium iron concentration and further, that this
effect is not due to modulation of the activity of the mature enzyme.
We were unable to demonstrate fur regulation of the cloned iga2 gene in
E. cali or the presence of a fur homologue on the gonococcal chromosome.
Potential pitfalls in the experimental design are discussed concerning

the lack of fur regulation in E. cali.

204
(Portions of this data were presented at the 90th. Annual Meeting of

the American Society for Microbiology in Anaheim, CA, in 1990 [53].)

MATERIALS AND METHODS

Bacterial strains and media. Neisseria ganarrhaeae GCM 740 is a
prototrophic, type 2 IgAl protease-producing (iga2b genotype) strain of
serovar PorB2 and N. ganarrhaeae NRL 30465 is an arginine-,
hypoxanthine-, and uracil-requiring (AHU-), type 1 IgAl protease-
producing (igal genotype) strain of PorAl serovar. These strains have
been described previously (39). The Escherichia coli strains BN402 and
BN4020 are fur isogens, being Fur+ and Fur-, respectively, and have been
described previously (2).

Gonococci were grown in NEDA, the chemically defined medium of Horse
and Bartenstein (36). The medium was made iron-limiting in one of two
manners, either extraction with the chelator 8-hydroxyquinolone (8-HQ;
Sigma Chemical Co., St. Louis, MO; 58) or by addition of the hydroxamate
siderophore, desferrioxamine B mesylate (Desferal; CIBA-GEIGY Corp.,
Summit, NJ), to 50 pM (43). To ensure that media induced the iron-
limitation response, gonococci were grown to mid-exponential phase and
outer membranes were isolated for analysis as described below.

E. cali strains were grown on LB agar medium (30) prior to
transformation (10) with the previously described type 2 IgAl protease-
encoding plasmid pRJ2 (Figure 1; 40). Following transformation, the
recombinant strains were selected and propagated on LB agar medium
containing 12.5 pg tetracycline per m1 (Sigma Chemical 00.). For growth
curves and IgAl protease assays, the E. cali strains were propagated in

fBHI containing 12.5 pg tetracycline per ml (fBHIt). We prepared fBHI

205

 

~
1’

C A' M 8‘ MH H H'A'B‘ H'NA'PN PPH A'M B’ A'SS‘C

Figure 1. Restriction map of pRJ2, a pBR322-iga2b recombinant construct
encoding IgAl protease activity. The direction of transcription is
indicated and restriction sites are designated as follows: A*, Ach; A' ,
AvaI; A”, Avail; B', BglI; B”, BglII; C, ClaI; H, HindIII; M, MluI; N,

NsiI; P, PstI; S, SmaI; and S', SalI.

206
by ultrafiltration of brain heart infusion broth medium (Difco
Laboratories, Detroit, MI) through a 10,000 molecular weight cutoff
membrane (type YMlO; Amicon Corp., Lexington, MA) under positive
nitrogen pressure. Media prepared in this fashion had been shown to be
iron-replete by its failure to induce the expression of iron-limitation
outer membrane proteins in N. ganarrhaeae, analyzed as described below.

Isolation of outer membranes and analysis. Gonococci were grown to
mid-exponential phase in the medium being tested. Cells were harvested
by low speed centrifugation and spheroplasts were prepared by sucrose-
1ysozyme-EDTA. treatment (44). Spheraplasts ‘were disrupted. 'by
ultrasonication and outer and cytoplasmic membranes were separated by
ultracentrifugation over isopycnic sucrose gradients (21). Membrane
protein concentrations were determined with a Bradford assay kit (Bio-
Rad Laboratories, Richmond, CA) using bovine serum albumin (Sigma
Chemical Co.) as a standard. Membrane proteins were resolved on 7.5-
12.5% linear gradient SDS-PAGE gels (27) and visualized using a
combination Coomassie blue-silver stain (18). Wet gels were analyzed
with transillumination and iron-repressible proteins were identified by
comparison of membranes from cells grown under iron replete conditions
with membranes from cells grown under iron-limited conditions.
Expression of the ~37,000 dalton Fbp protein was considered evidence of
iron-limitation by the medium (34,59).

IgAl protease assays. Quantitation of IgAl protease production was
determined by digestion of a 125I-labelled human IgAl paraprotein (Mar)
with a clarified supernatant sample prepared by centrifugation of an
aliquot of broth culture for two minutes in a microcentrifuge (Brinkmann

Instruments Inc., Westbury, NY). IgAl cleavage products were separated

207

by 9.0% SDS-PAGE as previously described (47). The amount of
radioactivity in the IgAl cleavage products (Fca and Fdh) was determined
using a Beckman Gamma5500B gamma counter (Beckman Instruments Inc. ,
Fullerton, CA). These values were used to calculate the percent
cleavage and in turn, the number of units of IgAl protease per digest
and the number of units per ml of culture supernatant. One IgAl
protease unit has been previously defined as one pg of IgAl cleaved per
minute at 37°C (47).

Assays examining the effect(s) of exogenous ferric iron on the
activity of ‘preformed IgAl protease were performed. by diluting an
enriched IgAl protease preparation (54) 1:400 in assay buffer [50 mM
Tris-H61 (pH 7.5), 10 mM CaClZ, 10 mM MgClz, 0.5% bovine serum albumin
(w/v)] and incubating with exogenous iron in the assay mix. Iran was

added in the form of Fe(N0 9H 0. Assays were performed in triplicate

3’3 2

at 37’C for fifteen minutes. The number of IgAl protease units per ml
was calculated for each replicate along 'with the mean enzymatic
activity per group. The mean number of units per ml for each treatment
group was compared to that of the control group (no Fe3+ added) by a
Student's t-test to compare sample means.

To examine the effect(s) of the media on the activity of preformed
IgAl protease, assays were performed by diluting an electrophoretically
pure preparation of IgAl protease (54) 1:100 and 1:200 in the medium
being tested. Media tested were 8-HQ extracted NEDA containing 252 pM

Ca2+ and 0.042% HCO - (w/v) and NEDA containing 50 pM Desferal, 10 pH

3
Fe3+, 252 pM Ca2+, and 0.042% HCO3' (w/v). Assay mixes were incubated

for fifteen minutes at 37°C in triplicate. The number of IgAl protease

units per ml was calculated and likewise, the mean protealytic activity

208
for each treatment group. Mean treatment group protease activities were

compared to that of a control group [NEDA containing 10 pM Fe3+, 250 pM

Ca2+, and 0.042 % H003- (w/v)] using a Student's t-test to compare
sample means.

Nucleotide sequencing and analysis. The reported promoter region of
the type 2b IgAl protease gene was subcloned into the bacteriophage
vector Ml3mpl8 by standard recombinant DNA techniques (30) and single
stranded DNA was isolated. Dideoxynucleotide sequencing reactions using
[a-3SS]dATP (New England Nuclear Corp., Boston, MA) were performed using
the Sequenase Version 2.0 kit as recommended by the supplier (United
States Biochemical, Cleveland, OH). Sequencing primers included the M13
universal primer supplied with the kit, and a 20-mer oligonucleotide
primer (GTTTGTTGACATCTGCAACG), selected and synthesized according to our
preliminary sequence data (Research. Genetics, Huntsville, AL).
Sequencing reaction products were resolved over 8% polyacrylamide
denaturing gels and identified by autoradiography.

Comparisons of the nucleotide sequences of the iga2b and iga2c (48)
promoter regions to each other, to an E. cali consensus Fur binding site
(41), and the fbp promoter region (4) were performed using the GenePro
version 4.20 software system (Riverside Scientific, Seattle, WA).

E. cali fur experiments. The E. cali strains BN402 (Fur+) and
BN4020 (Fur-) were transformed to the Iga2b+ phenotype (i. e. producing
IgAl protease) with pRJ2 by the CaCl2 procedure (10), selecting for
transformants on LB agar plus 12.5 pg tetracycline per ml. Fur
phenotypes were confirmed by the Arnow test for siderophore production

(1) and Iga2b phenotype was confirmed by IgAl protease assay (47).

The IgAl protease-producing fur isogens were propagated in iron-rich

209
fBHIt for two passages and transferred to fBHIt for analysis of growth
rate (OD6OO) to late exponential phase and IgAl protease production as
described above.

Southern blot analysis. To examine the possibility that the
gonococcal genome contained. a fur‘ homologue, we performed Southern
analysis by standard procedures (30). Chromosomal DNA was prepared from
N. ganarrhaeae strains GCM 740 and NRL 30465 and the E. coli fur
isogens, BN402 and BN4020, digested with the restriction endonuclease
ClaI as recommended by the supplier (Boehringer Mannheim Biochemicals,
Indianapolis, IN), and separated over a 0.7% agarose gel in Tris-boric
acid-EDTA. buffer (30). Transfer to a nitrocellulose sheet (BA85;
Schleicher & Schuell, Inc., Keene, NH) was by capillary action with 10X
SSC overnight at room temperature (30). Sheets were blocked and probed
in an aqueous hybridization solution at 6X SSC (0.9 M NaCl) overnight at
either 68°C (A; high stringency) or 42°C (B; low stringency). Twa
washes of thirty minutes each were done at room temperature in either:
(A) 2K SSC, 0.1% SDS; or in (B) 6X SSC, 0.1% SDS. Two subsequent
washes of thirty minutes each at 42°C were done in these respective
solutions for both A and B. The sheets were reblocked, probed with the
anti-hapten antibody, and developed as per the kit supplier's
instructions.

The E. cali fur gene probe was prepared by agarose gel
electrophoretic band elution of the 2.0 kilobase pair HindIII fragment
of pF3 (3,30). The probe was labeled by random priming using Klenow DNA
polymerase and a hapten-dUTP conjugate as per the Genius Nonradioactive

DNA Labeling and Detection Kit (Boehringer Mannheim Biochemicals).

210
RESULTS
Growth and IgAl protease production in iron limited media. When GCM
740 was propagated in 8-HQ extracted NEDA without additional Fe3+
(Figure 2A; C), it grew at a slower rate compared to growth in 8-HQ
extracted NEDA containing either 10 or 50 pM Fe3+ (Figure 2A; CF and
CPS, respectively). The iron limited gonococci were incapable of
growing to a final optical density as high as either of the iron
replete cultures. When the levels of IgAl protease per m1 of culture
supernatants were normalized with respect to the optical density and
compared (Figure 28), the iron limited culture (C) produced nearly twice
the amount of IgAl protease per ml per OD as that of the culture with 10

+ (CF) and over four times that of the culture with 50 pM Fe3+

pM Fe3
(CFS). This initial trend of increased IgAl protease activity was
observed up to mid-exponential phase («6 hours), where IgAl protease
activity in the iron replete culture (10 pM Fe3+; CF) surpassed the
activity of the iron limited culture, until these cultures reached
late-exponential phase (~9 hours). At late-exponential phase, the iron
limited culture (C) again exhibited more IgAl protease in the medium
than did either of the iron replete cultures (CF and CFS). Final
stationary phase IgAl protease production was 37 units per ml for the
iron limited culture, 24 units per ml for the culture containing 10 pM
Fe3+, and 12 units per ml for the culture containing 50 pM Fe3+.

Because these data were generated using a method of iron limitation
not previously described for N. ganarrhaeae, a similar experiment was

performed with GCM 740 in NEDA using Desferal to chelate the ferric iron

and render it nonassimilable to the gonococci (43). When optical

211

Figure 2A

 

IIIIIII
lllllll

     

O
5;; - // .
o
/ -
O
O
0.1, O, 5:
/ eCF :
: v (:F5 :

I
l

I
J

 

 

 

0.01 l l l l l l l l l l l l l l
012 3 4 5 6 7 8 9101112131415

time (hours)

Figure 2. The effects of 8-hydroxyquinolone-extracted defined media on
growth rate (A) and extracellular IgAl protease activity (B) of
Neisseria ganarrhaeae GCM 740. Media tested were as follows: 8-HQ
extracted NEDA (C), 8-HQ extracted NEDA plus 10 pM Fe(NO)3 (CF), and 8-
HQ extracted NEDA plus 50 pM Fe(NO)3 (CF5). All cultures were incubated.
with shaking at 37°C with 0.252 mM Ca2+ and 0.042% NaHCO3 (w/v) added“

IgAl protease assays were performed as previously described (47).

212

 

 

Figure 2B
‘I() I I If I I If I I I I I I I I
L. .J
I- / -
Q II- c.
O I T
L 30 —- -+
0) C) C _
0. T
r- -I
_ 0 CF
E v— V CPS ' . q
L.
a) 20 — —
CL _ a
O
(D L _
.t’ v
C I- e /\/v V ‘
3 I— . \ —I
N 10 “ ° : \V/ —
o O
9 . / 4
\V v -
c) I l l l l l l l I, 1 4L,

 

 

J l l
012 3 4 5 6 7 8 9101112131415

time (hours)

213

density data were analyzed (Figure 3A), the iron limited culture (Df)
grew at a greatly reduced rate and was unable to reach a maximal optical
density comparable to that of an iron replete culture (N). When the
data comparing units of IgAl protease per m1 of culture supernatant were
normalized with respect to the optical density (Figure 3B), the iron
limited culture (Df) exhibited at least three-fold more enzyme activity
initially than the iron replete culture (N). Though the enzyme activity
of the iron limited culture declined over time to a level approaching
that of the iron replete culture, the iron limited culture still
expressed more than twice the amount of IgAl protease activity per ml of
culture as that expressed by the iron replete culture.

These data suggested that some iron-mediated phenomenon was
occurring and resulting in increased amounts of IgAl protease activity
being expressed in the culture medium. We hypothesized that this could
be due to: 1) activation of preformed IgAl protease by the lowered Fe3+
concentrations, 2) destabilization of preformed IgAl protease by the
increased Fe3+ concentrations, or 3) a low iron-inductive effect on IgAl
protease synthesis.

To confirm that the media conditions tested were inducing a low iron
response, we prepared outer membranes of gonococci grown under our
assumed iron limited and iron replete conditions. A silver-stained SDS-
PAGE gel (shown in Figure 4) demonstrated that under the appropriate
medium conditions, previously described iron-repressible outer membrane
proteins were expressed (22,34,43,59). In particular, the ~37
kilodalton Fbp protein was only expressed in media containing Desferal

or in 8-HQ extracted media that was not supplemented with exogenous Fe3+

(lanes 2 and 4).

214

Figure 3A

 

AN

I IIIIII
llllllll

\\
{:1

0.1

OD 520
I IIIIII
I.
\D
\K

It
i.
....“A

I

 

0.01

llllI

 

l l 1 l l I l

0 1 2 3 4 5 6 7
time (hours)

on

Figure 3. The effects of Desferal-induced iron-limitation on growth
rate (A) and IgAl protease expression (B) of Neisseria ganarrhaeae GCM
740. Media tested were NEDA (N), and NEDA plus 50 pM Desferal (Df).
All cultures were incubated with shaking at 37°C with 10 ;di Fe(NO)3,
0.252 mM Ca2+ and 0.042% NaHCO3 (w/v) added. IgAl protease assays were

performed as previously described (47).

 

 

 

 

 

Figure 3B

#01010)
010010

ANNUbJ-b
UIOUIOU'IO

lgoZ units per ml per 00

10

5
O

215

 

 

._ \

 

0

0

Di

\C),,,/O\o\O

l A
4 5
time (hours)

I

6

/\A\ “A.,-K AfiAmA/Ax‘a
k I I\A/: #1
1 2 3

7

 

8

216

Figure 4. Silver stained 7.5-12.5% linear gradient SDS-PAGE analysis of
outer membranes of N. ganarrhaeae GCM 740 grown under iron-limited and
iron replete conditions. Bacteria were grown in broth media as follows:
lane 1, NEDA + 10 pM Fe(NO)3; lane 2, NEDA + 10 pM Fe(NO)3 + 50 pM
Desferal; lane 3, 8-HQ extracted NEDA + 10 pM Fe(NO)3; and lane 4, 8-HQ
extracted NEDA. All media contained 0.252 mM Ca2+ and 0.042% HCO3-
(w/v). Iron-repressible proteins are shown along the left margin (0)
as are some of the defined major gonococcal outer membrane proteins.

Positions of molecular weight standards are shown along the right margin

in kDa (Bio-Rad).

 

217

Figure 4

 

 

218

Effects of exogenous Fe3+ and various media on activity of
preformed IgAl protease. When an enriched preparation of IgAl protease
was incubated with exogenous Fe3+, there were no statistically
significant differences between enzyme activities in the presence of
exogenous iron (concentration range from 0.1-100 pM Fe3+) as compared to
a control reaction lacking exogenous Fe3+. Table l is a summary of the
data.

When a preparation of purified IgAl protease was diluted and
incubated in the various media tested, there was not a statistically
significant increase in the IgAl protease activity between the treated
samples and. the control samples (Table 2). IHowever, there was a
statistically significant decrease in enzyme activity when IgAl protease

in 8-HQ extracted NEDA (plus Ca2+ and HCO -) was compared to the control

3
IgAl protease samples in NEDA (plus Fe3+, Ca2+, and HCO3-).

These data taken together, indicated that the increase in
extracellular IgAl protease activity was not due to an activating or
destabilizing effect on preformed enzyme, caused by either treatment of
the medium or the medium iron concentration.

Examination of the promoter sequences of two iga2 genes and
comparison to the E. cali Fur binding site. The promoter nucleotide
sequence of the iga2b gene was identical to that of the previously
reported iga2c gene over the region of interest (48; Figure 5). These
two sequences exhibited high homology (14 of 19 bases or ~74%) to an E.
cali consensus iron box, the Fur binding site in the operator of genes
regulated by the iron-dependent Fur repressor system (41). When the

putative iga2 iron box was compared to the putative gonococcal fbp iron

box (4,35), less intraspecies homology was detected. The fbp iron box

219

Table l. The effect of exogenous iron concentration on activity of an

enriched preparation of IgAl protease.a

 

HH_£§;i I I] 1 -1 1-1 b > 0 g5c
100 14.13 i 0.20 -
10 14.56 i 0.79 -
1 13.92 i 0.60 -
0.1 14.72 i 0.57 -
Control 14.72 i 0.35 NDd

 

a - Enzyme preparation was diluted 1:400 in 50 mM Tris-H01 (pH 7.5),

10 mM CaC12, 10 mM M5012,

the stated concentration of Fe(NO3)3 9H20 and quantitative
125

assays using I-labeled IgAl were performed for 15 minutes at

0.5% bovine serum albumin (w/v) plus

37°C as previously described (47).
b - Mean i SD of three assays (x 103)

c i 2.776

' to(2) 0.05 '
d - Not determined

220

Table 2. The effect of relevant growth media on activity of a purified

IgAl protease preparation.a

 

Medium pg IgAl nin'1 1'1 b > 05°
8-HQ extracted NEDA 2.05 i 0.16 +
+ Ca2+, Hco3’
NEDA + 50 pM Desferal 3.23 i 0.30 -
+ Ca2+, Fe3+, Hco3'
NEDA control 3.00 r 0.40 NDd

+ Ca2+, Fe3+, H003'

 

a - Enzyme preparation was diluted 1:100 and 1:200 in the stated
medium and quantitative assays using 125I-labeled IgAl were
performed for 15 minutes at 37°C as previously described (47).

b - Mean i SD of six assays (x 103)

- ta(2) 0.05 - i 2.228

d - Not determined.

221

HOMOLOGY
_P_Z. cali consensus GATMTGAT$AT%AH%C
iga2b tAaAATGcaAAaCATTATC 14/ 1 9
fbp aATAATtATTthATTTat 13 / 19
E. coli consensus GA T TT
TAATGAT%ATEATTXEC

Figure 5. Nucleotide sequence alignments of parts of the IgAl protease
promoter region and the promoter region of the gonococcal fbp gene (4)
with a reported consensus E. cali Fur binding site (41). Conserved
nucleotides. are marked with dots and those shared among either

gonococcal sequence and the consensus are in upper case letters.

shared 13 of 19 bases with the E. cali consensus (~68% homology), while
the, iga2 and Vfbp iron boxes had only 8 of 19 bases in common (~42%
homology). These data suggested that the gonococcal iga2 gene may have
the potential to be regulated in a manner similar to the Fur repressor
system described in E. cali (3,20).

Response of a cloned iga2b gene in a known Fur background. The E.
cali Fur isogens were transformed with pRJ2, a recombinant pBR322
plasmid encoding the type 2 IgAl protease of GCM 740 (iga2b, Figure l).
Transformants were tested for both the Fur and Iga2 phenotypes by the
Arnow test (1) and qualitative IgAl protease assays (47) prior to the
growth curve experiment with quantitative IgAl protease assays. The
growth curve experiment was performed using bacteria grown in an iron
rich medium with the presumption that Fur repressed genes would be
derepressed in the Fur- strain (BN4020). Therefore, if the cloned iga2

gene was fur-regulable in E. cali, we predicted that we would see an

increase in IgAl protease activity with the Fur. mutant due to

222
derepression.

In fBHIt, BN402 containing pRJ2 (Fur+) did not produce catechol, an
intermediate in siderophore production, and did produce IgAl protease
(data not shown). The mutant BN4020 containing pRJ2 (Fur-) produced
catechol in the same medium and produced IgAl protease (data not shown).

The two strains grew with similar growth rates in fBHIt (Figure 6A).
The two strains also expressed nearly identical amounts of IgAl protease
at all points along the growth curve as followed (Figure 68). From
these data we concluded that the gonococcal iga2b gene is not regulated
by the E. cali fur system, as shown by the equivalent expression in the
E. cali Fur isogens grown in an iron rich medium.

Southern analysis of the Neisseria ganarrhaeae chromosome with the
E. cali fur gene. Figure 7 shows Southern analysis of chromosomal DNAs
from two strains of N. ganarrhaeae, GCM 740 and NRL 30465 and from E.
cali BN402 and BN4020, the Fur+ and Fur- isogens. Figure 7a shows
homologies detected under high stringency hybridization and wash
conditions. No homologies were detected between the E. cali fur probe
and the chromosome of either gonococcal strain. Figure 7b shows
homologies detected under low stringency hybridization and wash
conditions. Note the hybridization of the E. cali fur probe to the
molecular weight standards, A DNA digested with HindIII (lane 1); the
lack of hybridization to an empty lane (lane 2); the lack of apparent
specific hybridization to any region of the digested gonococcal DNAs
(lanes 3 and 4); and the still detectable specific, as well as
nonspecific hybridization to the digested E. cali DNAs (lanes 5 and 6).

These data suggested that there is no appreciable homology between the

223

 

 

 

 

Figure 6A
_ I 1 I I I 1 _
h— / —I
/g/
i 5‘ / “2
- // -
.— / _
I / -
o t /’ ‘
O - / _
LO ,1?
- // -i
Q //
O //
0 1 — 49 —
° : / o BN402 + pRJ2 :
e BN4020 +pRJ2 I
e _
O 01 I l I I l L
O 1 2 3 4 5 6 7
time (hours)
Figure 6. Growth (A) and IgAl protease expression (B) of E. cali BN402

(Fur+) plus pRJ2 and E. cali BN4020 (Fur-) plus pRJ2 in fBHIt, an iron

rich medium.
12.5 pg tetracycline per m1.

previously described (47).

Cultures were incubated with shaking at 37°C and contained

IgAl protease assays were performed as

Figure 6B

10

lgoZ units per ml

0.01

224

 

r Illlffll

 

 

  

    

l l Llllll

 

 

time (hours)

— O —

: 1

C 2

r- -i

b s

.— -J

F o BN402 + pRJ2 '

. BN4020 + pRJ2 -:

1

.1

‘- 1
l J I l I

O 1 3 4 5 6 7

225

Figure 7. Southern analysis of Neisseria gonorrhoeae and E. coli
genomic DNAs probed with an E. coli fur gene probe. Panel A was
hybridized and washed at high stringency. Panel B was hybridized and
washed at a low stringency (see Materials and Methods for conditions).
All DNAs were digested with ClaI unless specified otherwise. Lane 1” A
DNA digested with HindIII; lane 2, no DNA; lane 3, N. ganarrhaeae (X3!
740; lane 4, N. ganarrhaeae NRL 30465; lane 5, E. coli BN402; and larua
6, E. coli BN4020. Positions of molecular weight standards are given

along the margins in kilobase pairs (Bethesda Research Laboratories).

 

 

Figure 7

23d-

9Jh
615

QAh

2.3

   

 

226

23.13":

94-
6. 7. on. .

 

4,4- ----~ - -
is

2.3- -
2.0- , -

227
fur gene of' E. coli K-12 and the genome of either strain of .N.
ganarrhaeae tested.
DISCUSSION

We have demonstrated an iron-mediated effect which resulted in an
increase in the extracellular activity of the type 2 IgAl protease of
Neisseria ganarrhaeae. This increase in enzymatic activity when
gonococci were propagated in iron limited media correlated with the
expression of several described iron-repressible outer membrane proteins
(22,34,43,59). The increased IgAl protease activity was not due to an
effect of either high or low ferric iron concentrations, altering the
activity of preformed enzyme, and was not due to an effect of the medium
on the activity of preformed enzyme.

The promoter regions of both the iga2b and iga2c genes showed good
nucleotide sequence homology to a consensus E. coli Fur binding site
which suggested that the gonococcal IgAl protease gene might be
regulated by Fur in an iron-dependent manner. However, when we examined
the effect of iron limitation in a known Fur background, it indicated
that the E. coli fur system did not regulate the cloned N. ganarrhaeae
iga2b gene. It was also shown that there is no detectable DNA homology
between the cloned fur gene of E. coli K-12 and the genomes of two N.
ganarrhaeae strains, GCM 740 and NRL 30465.

It is not known at the present time how the medium iron
concentration influences or regulates the expression of the gonococcal
1382b gene, and whether the putative iron box in the promoter of the
iga2 genes is functional or anecdotal. Experiments to address these
problems are proposed in the following discussion.

We initially were interested in defining the gonococcal response to

228
iron limitation with regard to the expression of IgAl protease because
of its proposed function as a virulence factor for the organism
(24,37,45). We hypothesized that if IgAl protease were truly a
virulence factor, it might be upregulated in response to environmental
stimuli. One of the pertinent stimuli that we felt was important to
test was the limitation of available iron, as there is very little free
iron in or on the human body. Estimates have reported the free Fe3+
concentration as low as 10.18 M (15).

The first niche that N. ganarrhaeae would be expected to encounter
in a host is a mucosal epithelial surface, and any free iron which might
exist in these sites would be rapidly chelated by the host glycoprotein,
lactoferrin (31). Gonococci are able to acquire and utilize the iron
chelated by lactoferrin (32), but the events necessary for production of
the lactoferrin binding protein and expression on the surface of the
gonococcus are themselves triggered by a low iron response (6,28).
Therefore, the microbe is required to respond to the stress of iron
limitation prior to being able to express virulence functions which may
serve to alleviate this stress. There are a number of gonococcal outer
membrane proteins which are regulated. 'by the available iron
concentration, many of which have no defined function (22,43,59). Some
of these outer membrane proteins with described functions include the
Fbp (demonstrated to bind Fe3+, 35) and the undefined receptors for
lactoferrin and transferrin (6,28).

We observed an increase in the level of extracellular IgAl protease
when the bacteria were grown in iron limited media (Figures 28 and 38),

3+

regardless of the chemical reason for limitation of Fe . These same

media induced expression of Fbp (Figure 4) and other iron repressible

229

outer membrane proteins. The gonococci grown under iron limited
conditions produced from two to six-fold more enzyme than gonococci
grown in the same medium supplemented with 10 pH Fe3+ (Figures 28 and
38). For one set of media tested, gonococci grown to stationary phase
in a five-fold excess of Fe3+ demonstrated less than one-third the
amount of IgAl protease expressed by bacteria grown in an iron limited
culture (Figure 23; cultures C and CPS). From these data, we concluded
that the IgAl protease response was due to the limitation of iron.

Although the expression of IgAl protease activity is induced by only
two to six-fold when the cells are grown in iron limited media, we feel
that this response is of an acceptable magnitude to 'be considered.
genuine. There are a number of previously described iron regulated
functions which have inductive responses of this same order of
magnitude.

The E. coli outer membrane protein FhuA (aka TonA) shows an
induction of 2-3 fold under iron limited conditions in contrast to other
iron regulated functions such as Cir and FepA, 12 and 26 fold,
respectively (19). This moderate overexpression of FhuA. has been
observed by two different laboratories (19,42). A study of the kinetics
of the fhuA induction response demonstrated a rapid increase in protein
synthesis (as measured by 3H-leucine incorporation) which was followed
by reaching a maximal rate of synthesis, and then a steady decline over
time (26). This short term induction response is unusual in that only
one other iron regulated membrane protein with an undefined function
showed a similar response, while three different iron regulated proteins
(including FepA and Cir) showed a slower induction time but had a

continually increasing rate of synthesis (26).

230

Another example of a low induction ratio in E. coli is the manganese
containing metalloenzyme, superoxide dismutase (SodA). Inn qualitative
enzyme activity assays, there appeared to be no activity regulation with
regard to the tested strain's Fur phenotype. However, when the sodA
promoter was fused to the quantifiable lacZ reporter gene, a 3-4 fold
induction was observed (41).

Further examples of low inductive ratios have been demonstrated with
an E. coli ompA-Iacz transcriptional fusion construct, pSC27.1
(containing a synthetic E. coli consensus iron box), which shows
inductive ratios of 3.3 to 4.4-fold in two different E. coli host
strains (9), and about 2 fold in Yersinia pestis (55). Also, the cloned
diptheria toxin gene from Cbrynebacterium diphtheriae, shows a 4 to 5-
fold inductive ratio in E. coli (57).

We hypothesized that there could be two broad mechanisms that would
result in this outcome: 1) the change in enzymatic activity was due to
an effect of either increased or decreased Fe3+ concentration upon the
preformed IgAl protease enzyme or 2) that the decrease in available Fe3+
induced de novo synthesis of the iga2 gene product with a resultant
increase in enzyme activity.

We tested the former hypothesis by performing in vitro quantitative
enzyme assays using preformed or mature IgAl protease. Incubating the
enzyme in the presence of from 0.1-100 pM Fe3+, one order of magnitude
greater and two orders lower than the Fe3+ concentration in the iron
replete media tested, had no significant effect on its proteolytic
activity (Table 1). Similarly, when we incubated IgAl protease in the
iron limited media and compared its proteolytic activity to that when

incubated in iron-replete media, no significant increase in activity was

231
detected (Table 2). Rather, in the media rendered iron limited by 8-
hydroxyquinolone extraction, we observed a statistically significant
decrease in IgAl protease activity. From these data we concluded that
the effect of iron limitation on the activity of IgAl protease was not
due to activation or destabilization of mature enzyme by the media
concentration of Fe3+.

The observation that a sequence bearing good homology to a consensus
iron box is present in the promoter of the iga2 genes prompts the
question as to whether it is functional (Figure 5). Our data suggest
that this putative iron box is not functional, at least as far as the
gonococcal iga2 gene being fur-regulable in the E. coli strains. It was
determined by phenotypic assays that the strains were functioning in an
appropriate manner (1. e. BN4020, the Fur- strain, was expressing fur-
regulated functions sudh as catechol production in an iron rich media)
so the experiment yielded valid results with respect to the function of
the fur repressor system. However, the degree of gene expression is at
least four-fold lower in E. coli relative to the gonococcus (data not
shown) and in order to perform quantitative IgAl protease assays which
are statistically valid, one must incubate for long periods of time
during which the breakdown of the IgAl protease molecule itself may
affect the results obtained.

Another potential problem 'with the system 'used to address the
question of the fur-regulability of iga2 is the fact that the IgAl
protease molecule is a foreign protein and has no known homologue in E.
coli. That, coupled with the fact that the IgAl protease molecule is
secreted by a mechanism which also has no described homologue in E.

coli, may introduce unknown and uncontrolled variables into the system.

232

Addressing the question of cross functionality of different
regulatory systems, we also observed that there is not a well conserved
gonococcal fur' homology' at the DNA level, as detected. by Southern
hybridization analysis with the cloned E. coli fur gene (Figure 7).
Therefore, if N. ganarrhaeae has a fur-like repressor system for iron
regulation, it may be sufficiently divergent from the broadly conserved
enteric fur system that the two systems may not cofunction. Even among
closely related bacteria, there are reports of known Fur regulable genes
being weakly regulated in other systems. An example is the poorly
regulated cloned E. coli aerobactin operon in Shigella flexneri (2).

However, we do not feel that this is the case due to a recent report
which indicated that the purified Fur protein binds to the promoter
region of the gonococcal fbp gene, demonstrated in a gel retardation
assay (35). This data suggests that indeed there may be a Fur-like
mediated system in N. ganarrhaeae and that it may be similar enough to
the E. coli system to allow for cross-function.

In order to better address the problem, we propose that an iga2b-E.
c011 fusion protein be constructed in order to allow for the putative
iron regulated expression of an E. coli protein which may be
permissively regulated in a fur-dependent fashion. A suggested fusion
protein would be the promoterless phoA gene used in TnphoA mutagenesis
as described by Goldberg, et a1. (17). Briefly, the phoA coding
sequence is randomly inserted downstream from the promoter of the gene
under study via TnphoA mutagenesis. Resultant mutants can be identified
by plating on a differential, low-iron inductive medium also containing
S-bromo-h-chloro-B-indolylphosphate, a chromogenic substrate for

alkaline phosphatase (the phoA gene product; 17). Alkaline phosphatase

233

activity can be quantitated spectrophotometrically, which would allow
for assays to be done at points along the growth curve in a manner
similar to those done for IgAl protease. We propose that this sort of
assay using the putative regulatory region of the iga2b gene fused to an
E. coli reporter gene may produce more reliable data and allow the
question of whether the putative IgAl protease iron box is really a
functional entity or just an interesting looking stretch of DNA.

In summary, we have demonstrated that there was an increase in the
amount of extracellular IgAl protease activity expressed when Neisseria
gonorrhoeae was grown in an iron limited media. This increase in enzyme
activity was accompanied by the expression of previously described iron
stress outer membrane proteins and occurred under two different forms of
iron limitation. The increase in IgAl protease activity was not due to
an activation of the preformed enzyme caused by a shift to a lower Fe3+
concentration nor to a destabilization of preformed enzyme caused by an
increased Fe3+ concentration.

A potential Fur repressor binding site was identified in the
promoter of the iga2 gene, having 74% nucleotide sequence homology with
the E. coli consensus Fur binding site sequence. Functional assays
involving the cloned gonococcal type 2 IgAl protease gene in a known
Fur- E. coli background indicated that the iga2b gene is not under iron
regulation as controlled by the E. coli Fur repressor system. Also, DNA
hybridization analysis failed to detect specific homology between a
cloned E. coli fur gene and chromosomal DNA from two strains of N.
ganarrhaeae. The significance of the putative gonococcal iron box
located in the promoter of the iga2 gene and the apparent lack of fur

regulation are under further investigation.

234

ACKNOWLEDGEMENTS
The authors wish to thank Drs. R. R. Brubaker and R. D. Perry for
helpful discussions relevant to iron and virulence. They also
gratefully acknowledge Dr. J. B. Neilands and. M. Coy for ‘helpful
discussions regarding the E. coli fur system and for providing the E.
call for isogenic strains BN402 and BN4020 as well as pF3, the cloned

E. coli fur plasmid.

This work was supported by Public Health Service grant AI-21264 from
the National Institute for Allergy and Infectious Diseases to M. H. M.
and a Michigan State University College of Natural Sciences Continuing

Doctoral Fellowship to R. J. S.

10.

ll.

12.

235

REFERENCES

. Arnow, L. E. 1937. Colorimetric determination of the components of

3,h-dihydroxyphenylalanine-tyrosine mixtures. J. Biol. Chem.
118:531-541.

. Bagg, A., and J. B. Neilands. 1987. Molecular mechanism of

regulation of siderophore-mediated iron assimilation. Microbiol.
Rev. 51:509-518.

. Bagg, A., and J. B. Neilands. 1987. Ferric uptake regulation protein

acts as a repressor, employing iron(II) as a cofactor to bind the
operator of an iron transport operon in Escherichia cali.
Biochemistry. 26:5471-5477.

. Berish, S. A., T. A. Mietzner, L. W. Mayer, C. A. Genco, B. P.

Holloway, and S. A. Morse. 1990. Molecular cloning and
characterization of the structural gene for the major iron-
regulated protein expressed by Neisseria ganarrhaeae J. Exp. Med.
171:1535-1546.

. Bjorn, M. J., B. H. Iglewski, S. K. Ives, J. C. Sadoff. and M. L.

Vesil. 1978. Effect of iron on yields of exotoxin A in cultures of
Pseudomonas aeruginasa PA-103. Infect. Immun. 19:785-791.

. Blanton, K. J., G. D. Biswas, J. Tsai, J. Adams, D. W. Dyer, S. M.

Davis, G. G. Koch, P. K. Sen, and P. F. Sparling. 1990. Genetic
evidence that Neisseria ganarrhaeae produces specific receptors for
transferrin and lactoferrin. J. Bacteriol. 172:5225-5235.

. Boyd, J., M. N. Oso, and J. R. Murphy. 1990. Molecular cloning and

DNA sequence analysis of a diphtheria tax iron-dependent regulatory
element (dth) from Carynebacterium diphtheriae. Proc. Natl. Acad.
Sci. USA 87:5968-5972.

. Celderwood, S. B., and J. J. Mekalanos. 1987. Iron regulation of

Shiga-like toxin expression in Escherichia cali is mediated by the
fur locus. J. Bacteriol. 169:4759-4764.

. Calderwood, S. B., and J. J. Mekalanas. 1988. Confirmation of the

Fur operator site by insertion of a synthetic oligonucleotide into
an operon fusion plasmid. J. Bacterial. 170:1015-1017.

Cohen, S. N., A. C. Y. Chang, and L. Hsu. 1973. Nonchromosomal
antibiotic resistance in bacteria: genetic transformation of

Escherichia cali by R-factor DNA. Proc. Natl. Acad. Sci. USA.

69:2110-2114.

Covert, R. E., and B. G. Foster. 1981. The role of iron in the
production of hemolysin by Listeria monacytagenes. Curr. Microbial.
6:287-290.

Cross, J. H. 1989. Genetics and molecular biology of siderophore-
mediated iron transport in bacteria. Microbiol. Rev. 53:517-530.

l3.

14.

15.

15a.

l6.

17.

18.

19.

20.

21.

22.

23.

236

Dubos, R. J., and J. W. Geiger. 1946. Preparation and properties of
shiga toxin and toxoid. J. Exp. Med. 84:143-156.

Ernst, J. F., R. L. Bennett, and L. I. Rothfield. 1978. Constitutive
expression of the iron-enterochelin and ferrichrome uptake systems
in a mutant strain of Salmonella typhimurium. J. Bacteriol. 135:928-
934.

Finkelstein, R. A., C. V. Sciortino, and M. A. McIntosh. 1983. Role
of iron in microbe-hast interactions. Rev. Inf. Dis. 5:8759-8777.

Pohn, M. J., T. A. Mietzner, T. W. Hubbard, s. A. Morse, and E. W.
Hook, III. 1987. Human immunoglobulin G antibody response to the
major gonococcal iron-regulated protein. Infect. Immun. 55:3065-
3067.

Fujiyama, Y., K. Kobayashi. S. Senda, Y. Benno, T. Bamba, and S.
Hosada. 1985. A novel IgA protease from Clastridium sp. capable of
cleaving IgAl and IgA2m(1) allotype but not IgA2m(2) allotype
paraproteins. J. Immunol. 134:573-576.

Goldberg, M. B., V. J. DiRita, and S. B. Calderwood. 1990.
Identification of an iron-regulated virulence determinant in Vibria
chalerae, using TnphaA mutagenesis. Infect. Immun. 58:55-60.

Gorg, A., W. Pastel, J. Weser, H. W. Schivaragin, and W. M. Boesken.
1985. Horizontal SDS electrophoresis in ultrathin pare-gradient gels
for the analysis of urinary proteins. Science Tools (Sweden)
32:5-9.

Hantke, K. 1981. Regulation of ferric iron transport in Escherichia
cali K12: Isolation of a constitutive mutant. Mol. Gen. Genet. 182:
288-292.

Hantke, K. 1984. Cloning of the repressor protein gene of iron-
regulated systems in Escherichia cali K12. Mol. Gen. Genet. 197:337-
341.

Johnston, R. H., and E. C. Gotschlich. 1974. Isolation and
characterization of the outer membrane of Neisseria ganarrhaeae. J.
Bacteriol. 119:250-257.

Keevil, C. W., D. B. Davies, B. J. Spillane, and E. Mahenthiral-
ingam. 1988. Influence of iron-limited and excess continuous culture
on the virulence properties of Neisseria ganarrhaeae, p. 727-729. In
J. T. Poolman, et al. (ed.), Gonococci and Meningococci. Kluwer
Academic Publishers, Dordrecht, The Netherlands.

Kilian, M. 1981. Degradation of immunoglobulins Al, A2, and G by
suspected principal periodontal pathogens. Infect. Immun. 34:757-
765.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

237

Kilian, M., J. Mestecky, and M. W. Russell. 1988. Defense mechanisms
involving Fc-dependent functions of immunoglobulin A and their
subversion by bacterial immunoglobulin proteases. Microbiol. Rev.
52:296-303.

Kilian, M., J. Mestecky, and R. E. Schrohenloher. 1979. Pathogenic
species of the genus Haemophilus and Streptococcus pneumoniae
produce immunoglobulin Al protease. Infect. Immun. 26:143-149.

Klebba, P. E., M. A. McIntosh, and J. B. Neilands. 1982. Kinetics
of biosynthesis of iron-regulated membrane proteins in Escherichia
cali. J. Bacteriol. 149:880-888.

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

Lee, B. G., and A. B. Schryvers. 1988. Specificity of the
lactoferrin and transferrin receptors in Neisseria ganarrhaeae. Mol.
Microbiol. 2:827-829.

Male, C. J. 1979. Immunoglobulin Al protease production by
Haemophilus influenzae and Streptococcus pneumoniae. Infect. Immun.
26:254-261.

Maniatis, T., E. P. Fritsch, and J. Sambrook. 1982. Molecular
cloning, a laboratory manual. Cold Spring Harbor, New York.

McNabb, P. G., and T. B. Tomasi. 1981. Host defense mechanisms at
mucosal surfaces. Annu. Rev. Microbiol. 35:477-496.

Mickelsen, P. A., E. Blackman, and P. F. Sparling. 1982. Ability of
Neisseria ganarrhaeae, Neisseria meningitidis, and commensal

Neisseria species to obtain iron from lactoferrin. Infect. Immun.
35:915-920.

Mietzner, T. A., R. c. Barnes, Y. A. Jeanlouis, W. M. Shafer, and S.
A. Morse. 1986. Distribution of an antigenically related iron-
regulated protein among the Neisseria spp. Infect. Immun. 51:60-68.

Mietzner, T. A., G. H. Luginbuhl, E. Sandstrom, and S. A. Morse.
1984. Identification of an iron-regulated 37,000-dalton protein in
the cell envelope of Neisseria ganarrhaeae. Infect. Immun. 45:410-

416.

Horse, 8. A. 1990. Structure, function and regulation of the
gonococcal and meningococcal iron-binding protein (Fbp), abstr., p.
16. 7th Intl. Pathogenic Neisseria Conf. Berlin, GER. 1990.

Morse, S. A., and L. Bartenstein. 1980. Purine metabolism in
Neisseria ganarrhaeae: the requirement for hypoxanthine. Can. J.

Microbiol. 26:13-20.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

238

Mulks, M. H. 1985. Microbial IgA Proteases, p. 81-104. In I. A.
Holder (ed.), Bacterial Enzymes and Virulence. CRC Press, Inc., Boca
Raton, FL.

Mulks, M. H., S. J. Kornfeld, and A. G. Plaut. 1980. Specific
prateolysis of human IgA by Streptococcus pneumoniae and Haemophilus
influenzae. J. Inf. Dis. 141:450-456.

Mulks, M. H., D. A. Simpson, and R. J. Shoberg. 1987. Restriction
site polymorphism in genes encoding type 2 but not type 1 gonococcal
IgAl proteases. Antonie van Leeuwenhoek J. Microbiol. 53:471-478.

Mulks, M. H., D. A. Simpson, and R. J. Shoberg. 1988. Proteolysis of
gonococcal membrane proteins by IgAl proteases, abstr. P20, p. 20.
6th Intl. Pathogenic Neisseria Conf. Pine Mountain, GA. 1988.

Niederhoffer, E. G., C. M. Naranjo, K. L. Bradley, and J. A. Pee.
1990. Control of Escherichia cali superoxide dismutase (sadA and
sadB) genes by the ferric uptake regulation (fur) locus. J.
Bacteriol. 172:1930-1938.

Neilands, J. B. 1982. Microbial envelope proteins related to iron.
Annu. Rev. Microbiol. 36:285-309.

Norqvist, A., J. Davies, L. Norlander, and S. Normark. 1978. The
effect of iron starvation on the outer membrane protein composition
of Neisseria ganarrhaeae. FEMS Microbiol. Letters. 4:71-75.

Osborn, M. J., and R. Munson. 1974. Separation of inner
(cytoplasmic) and outer membranes of gram-negative bacteria. Meth.
Enzymol. 31:642-653.

Plaut, A. G. 1983. The IgAl proteases of pathogenic bacteria. Annu.
Rev. Microbiol. 37:603-622.

Plaut, A. G., J. V. Gilbert, M. S. Artenstein, and J. D. Capra.
1975. Neisseria ganarrhaeae and Neisseria meningitidis:

extracellular enzyme cleaves human immunoglobulin A. Science
190:1103-1105.

Plaut, A. G., J. V. Gilbert, and I. Heller. 1978. Assay and
properties of IgA protease of Streptococcus sanguis. Adv. Exp. Med.
Biol. 107:489-495.

Pohlner, J., R. Halter, K. Beyreuther, and T. F. Meyer. 1987. Gene
structure and extracellular secretion of Neisseria ganarrhaeae IgA
protease. Nature (London) 325:458-462.

Reinholdt, J., M. Tomana, S. B. Mortensen, and M. Kilian. 1990.
Molecular aspects of immunoglobulin A1 degradation by oral
streptococci. Infect. Immun. 58:1186-1194.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

239

Robertson, J. A., M. E. Stemler, and G. W. Stemke. 1984.
Immunoglobulin A protease activity of Ureaplasma urealyticum. J.
Clin. Microbiol. 19:255-258.

Salinas, P. G., M. E. Tomalsky, and J. H. Cross. 1989. Regulation of
the iron uptake system in Vibria anguillarium: evidence for a
cooperative effect between two transcriptional activators. Proc.
Natl. Acad. Sci. USA 86:3529-3533.

Shand, G. H., H. Anwar, J. Kadurugamuwa, M. R. W. Brown, S. H.
Silverman, and J. Melling. 1985. In vivo evidence that bacteria in
urinary tract infection grow under iron-restricted conditions.
Infect. Immun. 48:35-39.

Shoberg, R.J., and M.H. Mulks. 1990. Effect of media iron
concentration on immunoglobulin Al protease activity in Neisseria
ganarrhaeae, abstr. B-l9, p.29. Abstr. 90th Annu. Meet. Am. Soc.
Microbial., 1990.

Simpson, D. A., R. P. Hausinger, and M. H. Mulks. 1988.
Purification, characterization, and comparison of the immunoglobulin
Al proteases of Neisseria ganarrhaeae. J. Bacteriol. 170:1866-1873.

Staggs, T. M., and R. D. Perry. 1991. Identification and cloning of
a fur regulatory gene in Yersinia pestis. J. Bacteriol. 173:417-425.

Stoebner, J. A., and S. M. Payne. 1988. Iron-regulation of hemolysin
production and utilization of heme and hemoglobin by Vibria
chalerae. Infect. Immun. 56:2891-2895.

Tai, S-P. S., and R. K. Holmes. 1988. Iron regulation of the cloned
diphtheria toxin promoter in Escherichia cali. Infect. Immun.
56:2430-2436.

Waring, W. 8., and C. H. Werkman. 1942. Growth of bacteria in an
iron-free media. Arch. Biochem. 1:303-310.

West, S. E. H., and P. F. Sparling. 1985. Response of Neisseria
ganarrhaeae to iron limitation: Alterations in expression of
membrane proteins without apparent siderophore production. Infect.
Immun. 47:388-394.

SUMMARY AND CONCLUSIONS

One of the objectives of this research was to identify other
potential functions of the Neisseria ganarrhaeae type 2 IgAl protease.
Two different deletion-disruption mutant variants of the strain GCM 740
were constructed, GCM 74OA4 and GCM 740132, isogenic except for their
abilities to produce the IgAl protease.

After comparison of the Iga' variants with GCM 740, the parental
strain, only minor differences were observed between the two phenotypes.
The Iga status of a variant did not influence its ability to: l) grow in
a chemically defined medium, 2) grow under anaerobic conditions on a
complex agar medium, 3) grow under iron limited conditions in defined
broth media, 4) process polypeptides, either co- or post-
translationally, or 5) to be transformed.

Consistent differences in the outer (0M) and cytoplasmic membrane
(CM) protein profiles of the WT and mutants were observed. This
suggested a relationship between the variant's iga phenotype and the
membrane protein profiles. Differing proteins were undefined with
regard to identity or function, but based on relative abundance,
apparent SDS-PAGE molecular weight, and heat or reduction modifiable
behavior, several previously described OM proteins were ruled out.
Other than an OM protein of 26 kDa, which corresponded to an Opa, these
variant-specific proteins were identified not to be any of the other
previously well described gonococcal antigens (i. 9. Par, Rmp, or
environmental stress proteins). The A-iga mutants were also capable of

240

241
expressing appropriate, previously described OM stress response proteins
under anaerobic or iron limited conditions.

In an attempt to verify that some of these intervariant protein
profile differences were linked to the iga allele, the A-iga allele in
each mutant was restored by transformation. These Iga+ transformants
were demonstrated to produce kal protease and to have a restriction
site map identical to the parental strain. The Iga+ revertants
expressed a 56 kDa OM protein of unknown function, which exhibited Iga+-
dependent expression. One of the variants, GCM 740A4R-83 also exhibited
a similar Iga+-dependent 34.5 kDa cytoplasmic membrane protein, that was
not observed in either of two GCM 740A2 Iga+ revertants.

Another objective was to examine the susceptibility of the OM
proteins in the A-iga mutants to exogenous IgAl protease. The outer and
cytoplasmic membranes of these three variants were shown to contain
proteins which can be cleaved by exogenous, pure type 2 IgAl protease in
vitra. This was the first evidence of bacterially produced substrates
for these enzymes, save the autohydrolysis of the IgAl protease pro-
enzyme. The ability of the type 2 IgAl protease to alter some of the
membrane proteins in vitra, suggested that it might also perform this
function in viva. The ability of IgAl protease to cleave bacterial
membrane proteins was not limited to the particular enzyme-producing
strain or its variants (eg. GCM 740), but was also observed with another
N. ganarrhaeae strain, F62 and an A-iga variant; Escherichia cali DHl;
and a reference strain of Actinabacillus pleurapneumoniae.

When N. gonarrhaeae was propagated in an iron-limited medium, the
initially measured IgAl protease activity was from 2 to 6-fold as great

as that of the same strain propagated in similar media to which Fe3+ had

242

been added. This overall trend of increased protease activity was
maintained throughout mid-exponential phase and resumed in stationary
phase. Increased enzyme activity was not due to modulation of the
activity of preformed mature IgAl protease by either the ambient Fe3+
concentration or the chemical means of rendering the medium iron
limited, which suggests an increase in enzyme production.

Comparison of the DNA sequence of type 2 IgAl protease genes with an
E. cali consensus Fur binding site suggested the presence of an "iron
box", as the iga2 genes and the consensus sequence shared 74% nucleotide
identity. However, functional analyses examining the expression of the
cloned iga2 gene in characterized E. cali Fur host strains indicated
that the cloned iga2 gene was not Fur-regulated in E. cali. We were
also unable to detect a fur homologue in the N. ganarrhaeae genome by
DNA hybridization analysis. The mechanism of IgAl protease-regulation
by ferric iron and the presence of a putative Fur-binding site in the

promoter of the N. ganarrhaeae iga2 gene is presently unexplained.

APPENDIX

RESTRICTION SITE POLYMORPHISM IN GENES ENCODING TYPE 2 BUT NOT

TYPE 1 GONOCOCCAL IgAl PROTEASES

M. H. Mulks, D. A. Simpson, and R. J. Shoberg

Department of Microbiology and Public Health
Michigan State University

East Lansing, Michigan 48824 USA

Antonie van Leeuwenhoek J. Microbiol.

53:471-478. 1987.

This paper is appended to the dissertation because of the contribu-
tions of R. J. S. including: subcloning of the iga2b gene used as a mol-
ecular probe, experimental preparation of some of the blot panels shown
in Figure 1, interpretation of the data, and presentation of the data in
poster format at the 5th International Pathogenic Neisseria Conference

in Noordwijkerhout, the Netherlands (1986).

243

244

\

Antonie van Ltvnnenhoek 53: 47l 478 ( WK?)
t Martinus Nijltnll' Publishers. Dordru‘ht Printed in the Netherlands

Restriction site polymorphism in genes encoding type 2 but not type
I gonococcal IgAl proteases

M. H. MULKS. D. A. SIMPSON. 8L R. J. SHOBERG
Dept. nfMir-robiologr and Public Health. Michigan Stale Unimsiry. East farming. MI 48824. USA

Abstract. Neisseria ganarrhaeae produces two phenotypically distinct types of IgAl proteases. each
of which cleaves a specific peptide bond in the hinge region of the human lgAl havy chain. The
penesencoding IgAl protease from twenty-eight different strains of N. ganarrhaeae. including twelve
ahich produce type i enzyme. thirteen which produce type 2 enzyme. and three which are protease
negative. were analyzed. Nine restriction site patterns were found in the in genes. All twelve type
I strains showed identical restriction maps of the iga gene. which differed from all the iga-z variants.
The three protease negative strains each contained DNA homologous to the probe. While strain .
to strain variation in restriction maps of specific genes is not unique and has been reported in N.
gmorrhoeae previously. the existence of such restriction site polymorphism among 539-2 genes can-
u'asts strongly with the lack ofsuch variation among iga-l genes. The bass for this lack ofdiversity
among the iga-l genes is under further investigation.

Introduction

The immunogloban Al proteases are extracellular protealytic enzymes that
have been correlated with virulence in several species of bacteria which are path-
ogenic for humans (Mulks 1985; Plant 1983). Neisseria ganarrhaeae produces
two distinct phenotypic types of lgAl protease, each of which cleaves a different
specific bond in the hinge region of the human IgAl heavy chain. A given isolate
produces one but not both of the two phenotypic types of enzyme. The type
'of IgAl protease produced correlates with nutritional auxotype, outer mern-
brane Protein 1 serovar, and dam methylation in this organism. Strains which '
produce type I IgAl protease generally require arginine, hypoxanthine, and ura-
cil for growth, belong to the Protein IA-l or lA-2 serovars. and are dam methy-
lase positive (M ulks &. Knapp I985). Strains which produce type 2 enzyme are
extremely diverse in both auxotype and serotype, and may be either dam+ or
(1012- . Initial studies in this laboratory (Mulks & Knapp I985) indicated that
there is also polymorphism in the physical structure of the iga gene. as deter-
mined by restriction site maps. We report. here that there is extensive restriction -
site polymorphism in genes encoding type 2 but not type I IgAl proteases.

245

472 .II. II. Mulks et ul.
Materials and methods

Bum-rial strains. Isolates of Neisseria gmmrrltm'ae used in this study are listed
in Table I. All isolates with either NRL or CDC designations were received
front Dr Joan S. Knapp. Neisseria Reference Laboratory. Seattle. Washington.
Isolates were grown on GC base medium (Difco Laboratories) plus I% Kellogg‘s
supplement (Knapp & Holmes I975) or in defined broth medium (Morse 8:,
Bartenstein I980). and were stored at —70C in 2% tryptone + 20% glycerol.

IgAl protease asters. I gAl protease type was determined by analysis of IgAl
cleavage products on 9% polyacrylamide SDS gels (M ullts et al. I980).

Aaron-(ting. Each isolate was auxotyped on defined media (Knapp & Holmes
I975) and scored for requirements for arginine. hypoxanthine. uracil. proline,
and methionine. In some cases. isolates were also tested for requirements for
ornithine. serine. leueine, and isoleucine/valine. All isolates required cystine-
cysteine for growth.

Seton-ping. Serotyping of isolates into Protein I serovars by coagglutination
with monoclonal antibody reagents (Knapp et al. 1984) was performed by Dr
_J. S. Knapp.

DNA purification. Chromosomal DNA. from all gonococcal strains was pre-
pared (Bricker et al. I983) from cultures grown in defined broth medium (Morse
& Bartenstein I980).

Restriction analysis and southern hybridizations. Restriction enzymes (Bethesda
Research Laboratories, Inc., and New England BioLabs, Inc.) were used accord-
ing to the suppliers‘ recommendations. Electrophoresis of DNA was on agarose
submarine gels in $83 buffer. (Bricker et al. I983). Molecular weight standards
were phage l-Hindlll and I kilobase ladder fragments (Bethesda Research La- '
boratories, lnc.). DNA was transferred onto nitrocellulose paper by the South-
ern procedure and hybridized with Flu-labeled probe DNA at 42C in 50% for-
mamide (Maniatis et al. I982). The probe was the 4.2 Kb Hindlll fragment -
from a cloned igo-Zb gene (see Fig. 2). isolated from an agarose gel after electro-
phoresis and labelled with Pu-dCI‘P by nick translation. (M ullts 81. Knapp '
I985). . . .
’ Dom methylase determinations. Purified chromosomal DNA from each gono- 4'
coccal isolate was digested with the following restriction endonucleases: Mbol,
which only recognizes and cleaves the DNA sequence GATC- if the adenine re- '

'a- I” a. --
.

246

(ir'ttttlypir' variation in grtttrtr'ru'r'ttl [gal I proteases 473
sidue is not methylated. Dpnl. which only cleaves this sequence if the adenine

is methylated. and SunBA l, which cleaves regardless of methylation. If the DNA
. was cleaved by Dpttl and/or Sau3Al but not lllltol. the DNA was considered

Tultlr' I. Characteristics of N. mitts-or strains used in this study. .

 

Strain ‘ Auxotype' . Scrovar Dam IgAl Protase Source“

 

 

Methylan Activity Genotype

 

I. NRL 30465 AHU IAI + I iga-l Japan

2. NRL328I9 AHU . IAI + I iga-I Wash.

3. NRL905 AHU IAI + I iga-I ' .
4. NRL9396 AHU [Al + " I iga-l Copen.
5. CDC6 AHU (argC) IAI + . l iga-I Iowa

6. CDCI’. AHUIargOHV IAI + I iga-I ' Fla.

7. CDC209 AHUIVLeuMet lAI + ' I iga-I Ohio

8. cpcsr AHU (argOISer m + I igs-l Iowa

9. cacrzs AHU ' I82 + I iga-l Ohio

I0. CDC208 AHU(argO)IV. I82 + I iga-I Ohio

u. NRL31774 Pro tat + I 59.1 . Copen.

. 12. NRL33296 Pro IBI + I iga-l - Copen.

I3. F62 Pro 137 — 2 iga-za '

l4. MKB Pro I37 - 2 ip-2a ‘ _.
I5. CDC3 WT 1B3 - 2 iga-za Colo. -
16. CDCI79 Argo IAI - 2‘ iga-za Colo. "
l7. GCM 740 WT IBZ ' — 2 iga-Zb Mass.
l8. MS-II WT IB9 - 2 iga-Ze ’ '

I9. "-12 WT 139 - ' 2 iga-Zc r
20. CDC96 . Pro IA3 ' + 2 iga-Zd NJ.
21. coc253 Argo Met taz - 2 iga-Zd ~ Colo.
22. CDC34 . ProArg Metl..eu IBI - . 2 iga-Ie Iowa
23. CDCSO Pro IBI + 2 iga-ze Fla.
24.-com Argo I824 - , 2 iga-Zf Fla. -
25. CDC9 Pro 1A3 - 2 ian-2: . Mass.
26. NRL7904 WT 1A5 +- - iga-za ,- Copen. .
27. warms wr . IA5 +. . . -. , lea-2a _ Copen.
28. NRL3I778 ‘ Metf . . IBIS + - iga-Zh Copen. ..

 

'Auxotype: WT - prototrophic: Arg - requires arginine for growth: Argo -arginine requirement .

can be satisfied by ornithine. citrulline. or arginine; ArgC - arginine requirement can be satisfied

. by citrulline or arginine but not ornithine: AHU - requires arginine. hyponrttltine. and uracilz- .
. Pro - requires proline: Met - requires methionirte: Set - requires serine: and IV - requires

isoleucine and/or valine. -.-:

' _ |'Source: Copen. - Copenhagen. Denmark: others include Colorado. Florida. Iowa. Massachu- .

sens. New Jersey. Ohio. and Washington states.

‘NRL 905 was meived from J. S. Knapp: F62 was received from P. F. Sparling: MKB ts the

F62 derivative used by J. M. Koomey and was received from H. D. Cooper MS-Il and Its derivative

strain IF-2 were received from M. So.

247

474 M. II. .llttllt'sr'l al.

to be ntetltylated and the strain designated dam-r: if the DNA was cleaved by
.Illml and Sutt3A| but not Dpttl. the strain was designated tlum- (Kolodltin ct
al. 1982).

Results and discussion

IgAl protease genes from several strains of N. ganarrhaeae have been cloned
and carefully mapped with restriction enzymes (M ulks & Knapp I985: Fishman
et al. I 985: Rahr et al. 1985: Halter et al. 1984: Koomey et al. I982). Comparison
of these physical maps of the igo genes and their flanking sequences (Mulks
& Knapp 1985) demonstrated significant strain-to-strain variation in the
number and location of restriction sites within the iga gene. In order to evaluate
the extent and significance of these variations in the genes encoding gonococh
IgAl protease activity, we analyzed twenty-eight different isolates of Neisseria
ganarrhaeae (Table l) . Each isolate was assayed for IgAl protease type. auxo-
type. serotype. and production of dam methylase activity. Specific isolates to
be studied were selected as representative of the phenotypic diversity to be found
among protease type I and type 2 strains. In our collection of over I 100 gonococ-
cal strains, only the three which are included in this study have been found to
be protease negative.
. Chromosal DNA was purified from each isolate and digested with four restric-
tion enzymes: Hindlll, Avol, Bglll. and PstI. The restriction fragments were
separated by electrophoresis on agarose gels and the fragrrtents homologous
to the iga gene were identified by Southern hybridization, using as a probe the
4.2 Kb Hindlll fragment from our cloned igo-Zb gene (see Fig. 2). Representa-
tive Southern blots are shown in Fig. I. Using the data from these blots plus
restriction maps of cloned iga genes, nine distinct restriction maps of gonococcal
iga genes were identified (Fig. 2). Each of these varied slightly in pattern within
the region specifying IgAl protease activity. but all nine were essentially identical
in the structure of - the flanking sequences. Among 13 isolates producing type .
2 protease, 7 variations in the restriction map of the iga gene (which we have
' designated genotypes iga-Za to igo-2g) were found. No correlation could be.
made between specific iga-2 genotypes and auxotype. Protein I serovar, produc-
tion of dam methylase. or source of the isolates. In contrast. 12 isolates produc-
ing type 1 protease. regardless of the auxotype. serovar, or source of the isolate
(all were dom+). had identical restriction maps of the iga gene (iga-I), which
. differed from all 7 iga-2 genotypes. The three strains which do not produce de-
tectable levels of IgAl protease activity contained apparently intact iga genes
with either iga-Za restriction patterns or an eight variant (iga-Zh), suggesting-
. that the iga gene is present in these strains but is defective or that other genes
are necessary for full expression of protease activity.

(iwmtt'pic' animal»: in gaunt-mull Iga-l I proteases 475

248

Southern Hybridization of Genomic DNA

 

 

 

 

 

 

 

 

 

 

AvaI Hindm
2:.t- 23.1- -
a.s- -‘ 9.4-
s-
“ ‘ *_- ' - “-51
“‘ ‘-‘1 .-~.-0 an-
. ‘ - -
- - 2.3-t
2.3 - 2.1)"
2.0 ‘
n - Kb
t2a2b2c2d20212g2n izanacaazoztzgaa
Boll! PstI
.- -
t "' .Qur.‘ t" --—"‘"-
as. g -. ..-r' 0."
9.... 57°. “-
s.s- :. .. ..
4.4- —
in 2.3-
4' 2.0"l
2:3: .. .7 o
- to- -- -
Kb Kb

 

 

 

 

 

 

1 2l2b2c2d202f292h 12.202621120212021!

Fig. 1. Restriction and Southern blot analysis of genomic DNA purified from nine dilferent strains
of N. ganarrhaeae. DNA from each strain was digested with each of four restriction enzymes. includ- .
ing Am l, 3:! II. Hind III. and Pr: l. DNA fragernents were separated on 0.7% agarose gels. trans- .
fer-red onto nitrocellulose. and hybridized to the Pin-labelled 4.2 Kb Hind III fragment isolated
_t'romaeloned iga-Zbgene(see Fig. 2). Inall fourgels. lanesareddgnated awarding totheign'

mypeofmesmimikixa-l andiga-Za through 2h. SinsofDNA molecularweightnnrkers.
irt Khanindicatedonthelefiofeachautondiograph.

The significance of these variationsin gonococcal 530 genes to the physiology.
of the organism or the pathogenesis of infection is not clear. Differences in the .
genes encoding type I and 2 proteases were expected, despite the fact that these
genes are highly homologous at the DNA level, the type I and type 2 enzyme
proteins are biochemically and antigenically distinct (Simpson, Hausinger &
Mulks, J. Bacteriology, In press). However, while the genes encoding type I
enzymes are unique and distinct from those encoding type 2 enzymes. the varia-

249

476 M. II. Mulks” al.

 

 

 

 

 

 

 

 

 

:2 Wm Restriction Site 3 W et sum!
j. ; I. III? ..
w; I I HIE.
”3. t I III.
..... 3 I II I III .
I I I 1 III .
.. I I I I m .
.. I I I IL I III .
.. I I I I III-I.
.. I I I II IIiI .

"s : f; : r : r 1 r e

 

fig. 2. Physical maps of gonococcal IgAl protease (ige) genes and flanking sequences. asdetermined
by restriction enzyme analysis of cloned genes and by chromosomal Southern blot analysis with
the 4.2 Kb Hind III fragment from a cloned iga-Zb gene used as a probe. Nine distinct arrangements
of restriction sites were found. each varying slightly in pattern within the region specifying IgAl
protease activity but all being essentially identical in the structure of the flanking sequences. The
igestrueturalgenespamtheregion betweenthetwofiindlll sites. andistranseribed fromleft
to right from a promoter ~03 Kb upstream from the Hind III site. Thenumber of strains belonging
to each genotype is indicated on the right: further information concerning the phenotypes of each
strain is given in Table I.

tions in iga-2 genes are not correspondingly responsible for obvious differences
in the proteins they encode. Whether subtle differences in the type 2 enzyme
proteins exist has not been determined. .

. Strain-to-strain variation in DNA sequences of specific genes is not unique
to the IgAl protease genes of N. ganarrhaeae. Both genotypic and phenotypic .
polymorphism also occurs among the IgAl proteases of Homophilus influenza:
(Bricker et-al. I983. I985: Mulks et al. I982). In this species. however. the type
of protease produced and the genotype of the iga gene correlates with the capsu-
lar serotype of the organism. No similar correlation of iga genotypes to other
known phenotypes has been detected in N. ganarrhaeae. Other gonococcal genes
which demonstrate restriction site polymorphism include proA and proB, which.
specify two enzymes involved in the biosynthesis of proline (Stein et al. I984).
A variety of other enzymes and proteins produced by the gonococcus also show
considerable strain-to-strain antigenic or electrophoretic diversity. The multiple

250

Gtva_t'pit' variation in gmtm-m't'ul Iga-I I proteases 477

serovars of outer membrane protein 1 and the multiple isozymes used to establish
"electrophoretic types“ are examples which probably reflect significant differ-
ences in the DNA sequences of those genes. It is of particular interest that. while
extensive diversity in these genes is apparently the nornt for this species. AHU
strains of N. ganarrhaeae are highly conserved for protein l serovars (Knapp
et al. I985: Mulks & Knapp I985) and isozymc electrophoretic types (J. Musscr.
pers. comm.). as well as for I gAl protease genotype. The mechanism which me-
diates this apparent limitation of genetic variation in AH U strains of N. gonorr-
Imeae remains to be determined.

Acknowledgements

This research was supported by I’ublie Health Service grant AI-21264 from the
National Institute of Allergy and Infectious Diseases.

' References

Bricker. J., M. H. Mulks. A. G. Plaut. E. R. Moxonth. Wright (I983) IgAl proteasesofflaanoplti-
in: influensae: Cloning and characterization in E. cali K-I2. Proc. Natl. Acad. Sci. USA 80:
2681—2685

Bricker.J..M. I-I. Mulks.E. n Meson. A. G.Plaut&A. wmmussn Physical and madam
of DNA regions encoding the immunoglobulin A proteases of difl'erent specificities produced
by Hmophilus influenza. Infect. Immun. 47: 370-374

Fishman, V., J. Bricker. J. V. Gilbert. A. G. Plaut 8:. A. Wright (I985) Cloning of the type I IgAl
preteasefrom Neisseringonorritoeeeand secretion oftheenzymefrom E. roii. InzG. Schoolnik
(Ed) The Pathogenic Neisseriae (pp. l64-I68). American Society for Microbiology. Washington.
D. C.

Halter. R..J. PohlnerhT. Meyer( I984) IgA protease of Nemmagannrrhm notationand charac-
terization of the geneand its extracellular product. EMBOJ. 3: IS95-l60l

Knapp. J. S. a K. K. Holmes (I975) Disseminated gonme infection caused by Neiseriegm
rhoeaewith unique nutritionalrequirernents. J. Infect. Dis. l32: 204—208

Knapp. J. 8.. M. H. Mulks. I Lind. H. B. Short 8; V. L. Clark (I985) Evolution ofgonoeoecal

. populationsin Copenhagert. Denmark. l928-l979. In: G. Schoolnik (Ed)ThePathogenicNeisser-
iae(pp. 82—86). American Society for Microbiology. Washington. D. C.

Knapp. J. S.. M. R. Tarn. R. C. Nowinski. K. K. Holmes & E. G. Sandstrom (I984) Serological
classification of Neisseria ganarrhaeae using monoclonal antibodies directed against gonococcal
outer membrane protein I. J. Infect. Dis. ISO: “—48

Kolodkin. A. B., V. L. Clark. F. C. Tenover. & F. E. Young ( I982) High correlation of the presence
of methyladenine in Neisseria ganarrhaeae DNA with the AHU auxotype. Infect. Immun. 36:
586-590 ‘ ’- -

Koomey. J. M.. R. E. Gill 4: S. Falkow (I982) Genetic and bioehemial analysis of gonococcal
IgAl protease: Cloning in Escherichia cali and construction of mutants of gonococci that fail
to produce the activity. Proc. Natl. Acad. Sci. USA 79: 7881-7885

251

478 If. H. Mulkset al.

Maniatis. T.. E. Fritseh & J. Sumbrook (I982) Molecular Cloning. Cold Spring Harbor Laboratory,
Cold Spring Harbor. New York

Morse. S. A. 8: L. Bartenstein (I980) Purine metabolism in Neisseria ganarrhaeae: the requirement
for hypoxanthine. Can. J. Microbiol. 26: I3—20

. Mulks. M. H. I985. Microbial IgA proteases. In: I. A. Holder (Ed) Microbial Enzymesand Virulence

° (pp. 8I- I04). CRCPmBocaRatonJ’L

Mulks. M. H. &J. S. Knapp (I985) Immunoglobulin Al protase types ofNeineriagmorrlteeae.
In: G. Schoolnik (Ed) The Pathogenic Neisseriae (pp. 51—56). American Society for Microbiology,
Washington. D. C.

Mulks. M. H.. S. J. Kornfeld. B. Frangione & A. G. Plaut ( I 982) Relationship between the specificity
of IgA protases and aerotypa in Hmopltiim influensee. J. Infect. Dis. l46: 266—274 ‘

Mulks. M. H.. A. G. Plaut. H. A. Feldman & B. Frangione (I980) IgA protases of two distinct
specificities are relased by Neisseria mingitidtlr. J. Exp. Dis. 152: l442-I447

Plaut. A. G. (I983) The IgAl protases of pathogenic bacteria. Ann. Rev. Microbiol. 37: 603-622

Rahr. S. R. Halter, H. Muller.J. Pohlner&T. F. Meyer(l985)GeneticanaIysisofneisserial immun-
oglobulin A protases. In: G. Schoolnik (Ed) The Pathogenic Neisseriae (pp. l57-I63). American
Society for Microbiology, Washington, D. C.

Stein. D. C.. L. E. Silver,V. L. Clark.& F. E.Young(l984) Cloninggenesforproline biosynthesis
from Neisseria ganarrhaeae: Identification by interspeeific complernentation of Escherichia eoii
mutants. J. Bacteriol. I58: 696-7“)