Characterization of glyceraldehyde-3-phosphate dehydrogenase from group A streptococci and analysis of its role as a pla...


Material Information

Characterization of glyceraldehyde-3-phosphate dehydrogenase from group A streptococci and analysis of its role as a plasmin receptor
Physical Description:
vi, 181 leaves : ill. ; 29 cm.
Winram, Scott Budd, 1964-
Publication Date:


Subjects / Keywords:
Glyceraldehyde-3-Phosphate Dehydrogenases -- genetics   ( mesh )
Glyceraldehyde-3-Phosphate Dehydrogenases -- chemistry   ( mesh )
Streptococcus pyogenes -- genetics   ( mesh )
Plasmin -- physiology   ( mesh )
Genes, Structural, Bacterial -- genetics   ( mesh )
Genes, Structural, Bacterial -- chemistry   ( mesh )
Mutagenesis, Site-Directed   ( mesh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1995.
Includes bibliographical references (leaves 171-180).
Statement of Responsibility:
by Scott Budd Winram.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 49349710
System ID:

Full Text







This work is dedicated to my dog, Booger, without whose

lessons in blind faith this work would not have been



I would like to thank Dr. Richard Lottenberg for his

guidance and friendship during my thesis work. I would also

like to thank my outside examiner Dr. Donald Leblanc and the

members of my committee for their help in guiding this

project, especially Dr. Paul Gulig who contributed

significantly to my graduate education. Most of all I would

like to acknowledge my parents, Geraldine Holt, and Kathleen

O'Donnel for their unwavering love and support.



ACKNOWLEDGEMENTS ..........................................iii

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

INTRODUCTION ................................................1
Group A Streptococci ................................... 1
The Plasminogen System ................................. 4
Plasmin(ogen) Receptors ................................ 6
Specific Background and Goals ..........................8

Introduction ............................................10
Materials and Methods .................................11
Results ......................... ....................... 21
Discussion ............... ............................. 36

Introduction ..................... .......................46
Materials and Methods ..................................47
Results ................................................. 56
Discussion .................................... ........ 76

OF GROUP A STREPTOCOCCAL STRAIN 64/14 IN plr.............89
Introduction ................................... ........ 89
Materials and Methods ................................... 91
Results .............................. ................. 106
Discussion .................................... ....... 143

EPILOGUE ........ ............. .................... ......... 166

LIST OF REFERENCES ................ ... ..... ..... ..... .......171

BIOGRAPHICAL SKETCH ....................................... 181

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



Scott Budd Winram

August 1995

Chairman: Richard Lottenberg, M.D.
Major Department: Molecular Genetics and Microbiology

Group A streptococcus is a human pathogen which has the

capacity to cause highly invasive soft tissue infections.

Streptococci can interact with the plasminogen system by

secreting the plasminogen activator streptokinase and

subsequently capturing the serine protease plasmin on the

cell surface. To assess the contribution of this surface

proteolytic activity to the invasive potential of the

organism, receptor molecules must be identified. Plr, a

candidate plasmin receptor protein from group A strain 64/14,

has previously been isolated and the gene encoding this

protein, plr, cloned in our laboratory. In this work, Plr,

has been identified as a functional glyceraldehyde-3-

phosphate dehydrogenase and characterized both biochemically

and genetically. Plr was the only GAPDH identified in strain

64/14 and plr appears to be an essential gene. Analysis of

expressed products of mutated plr revealed that a lysine in

the C-terminal position was necessary for wild-type levels of

plasmin binding. Additional residues of Plr also appear to

participate in the binding interaction with plasmin.

Plasmin-binding deficient Plr molecules with alterations of

the C-terminal lysine retained GAPDH enzymatic activity. The

genes harboring these mutations have been introduced

successfully into group A strain 64/14 at the plr locus.

Streptococcal strains expressing the mutant Plr molecules and

wild-type streptococci demonstrated equivalent plasmin

binding activity. Additional investigations indicated that

the majority of plasmin receptors on strain 64/14 are

proteins and that C-terminal lysines are required for at

least some of the binding activity. Therefore, strain 64/14

expresses plasmin binding structures other than Plr. The

role of plasmin receptors in streptococcal pathogenesis

remains to be determined.


Group A Streptococci

Group A streptococci are Gram-positive bacteria capable

of causing a variety of diseases in humans ranging from acute

pharyngitis to potentially lethal invasive soft tissue

infections and a toxic shock-like syndrome (Stevens, 1992).

Rheumatic fever and glomerulonephritis are two serious post-

infectious sequelae of group A streptococcal infections

(Bisno, 1991). Group A streptococci possess many surface and

secreted factors that may contribute to virulence, but few

have been definitively identified as participating in the in

vivo pathogenesis of streptococci.

Surface structures that have been identified include the

hyaluronic acid capsule, streptolysins 0 and S, M-protein and

M-related proteins such as the IgG binding proteins,

fibronectin binding proteins such as Protein F, collagen

binding proteins, C5a peptidase, and plasmin(ogen)

receptor(s). The capsule mimics host cell surfaces and

thereby aids the bacteria in avoiding opsonization and

phagocytosis. The importance of the capsule in the virulence

of group A streptococci was demonstrated when transposon

mutagenesis of the region encoding genes required for

capsule synthesis resulted in a 100-fold decrease of

virulence in mice (Wessels et al., 1991). In addition to

binding several host proteins including fibrinogen, the

surface M-protein has also been shown to inhibit opsonization

and phagocytosis of streptococci in vitro. Over 80

antigenically distinct types of M-protein have been

identified to date. Opsonic antibodies are usually made

against M-protein at a variable region in the exposed NH2-

terminus of the protein. Because of this, protective

antibodies against one serotype of group A streptococci still

leaves the host susceptible to infection by the remaining

serotypes (reviewed by Fischetti, 1989). The fibronectin

binding protein, Protein F, has been identified as an adhesin

of group A streptococci (Hanski and Caparon 1992). The

cloned gene product expressed in E. coli increased adherence

of the bacteria to fibronectin. Lipoteichoic acid on the

bacterial surface has also been reported to bind fibronectin

(Simpson and Beachey, 1983). Recently it has been

recognized that certain environmental stimuli can modulate

expression of streptococcal surface proteins. The genes

encoding M- and M-like proteins and the C5a peptidase reside

together in an operon which is regulated in response to

environmental CO2 tension (Okada et al. 1993). By growing

streptococci strains under high or low CO2 concentrations to

increase or repress M-protein expression, preferential

binding was demonstrated to skin keratinocytes or to

Langerhans cells. Investigations using isogenic strains

indicated that Protein F was the Langerhan cell adhesin and

suggested that the character of the bacterial surface could

be altered during various stages of invasive infections

(Okada et al., 1994).

Many group A streptococcal strains also secrete several

toxins such as SpeA, SpeB, and SpeC. SpeA and SpeC have been

shown to have toxic effects in animals and act as T-cell

mitogens in vitro (Lee and Schlievert, 1989; Drake and

Kotzin, 1992). Epidemiologic studies have indicated that the

strains causing the toxic shock-like syndrome often express

these toxins (Hauser et al., 1991). The active form of SpeB

has been reported to be a cysteine protease and is able to

cleave the IL-1 cytokine precursor as well as the urokinase

plasminogen activator receptor (Kapur et al., 1993; Wolf et

al., 1994). The role of these secreted factors and their

regulation in the pathogenesis of group A streptococci

remains to be elucidated.

Group A streptococci have the potential to interact

directly with the human plasminogen system. These bacteria

secrete streptokinase which is a potent activator of

plasminogen (Castellino, 1979). Plasmin can degrade a wide

range of protein substrates including fibrin and other

protein that comprise tissue planes, and group A streptococci

can bind enzymatically active plasmin on the cell surface

(Lottenberg et al., 1987). These interactions of group A

streptococci with the plasminogen system may contribute to

the ability of the bacteria to traverse the extracellular

matrix during invasive infections but the contribution of

these interactions has not been defined.

The Plasminoaen System

The zymogen plasminogen contains 791 amino acids in its

mature form. Plasminogen is found in blood, saliva, and in

extracellular spaces. Plasminogen is converted into the

active serine protease plasmin by either host or bacterial

plasminogen activators via cleavage of the peptide bond

between arginine561 and valine562 (reviewed by Vassalli et

al., 1991). The human plasminogen activators, tissue

plasminogen activator (tPA) and urokinase plasminogen

activator (uPA), possess proteolytic activity. In contrast,

the prokaryotic plasminogen activators streptokinase and

staphylokinase, secreted by certain strains of streptococci

and staphylococci, respectively, have been reported to

initially form a one to one stochiometric complex with

plasminogen and this complex then converts other molecules of

plasminogen to plasmin (Castellino, 1979). The plasminogen

molecule consists of a heavy chain and a light chain linked

together by three disulfide bonds. The light chain contains

an enzymatic active site with an amino acid sequence of a

serine protease. The heavy chain contains five well

characterized loops referred to as kringles, which serve as

lysine binding sites (LBS) for other proteins. There are two

classes of lysine binding sites. Kringle 1 has a high

affinity binding site for lysine whereas the low affinity

sites are localized to kringles 2 through 5. The native NH2-

terminal amino acid of secreted plasminogen is glutamic acid

(glu-plasminogen). Plasmin will cleave the peptide bond

between lysine77 and lysine78 of glu-plasmin(ogen) to form

lys-plasminogen or lys-plasmin (reviewed by Castillino,

1995). The conversion of glu-plasminogen to lys-plasminogen

results in a dramatic conformational change. The interaction

of lysine or epsilon-amino-caproic acid (EACA) with glu-

plasminogen will also produce this conformational change

(Violand et al., 1975). This "lys" conformation has been

reported to be required for high affinity binding

interactions through a C-terminal lysine residue of some

ligands such as fibrin fragments and alpha-2-antiplasmin

(Christensen, 1988; Sasaki et al., 1986).

Plasmin has trypsin-like specificity, with the ability

to hydrolyze peptide bonds after lysine and arginine

residues. This protease has been characterized most

thoroughly for its role in the dissolution of fibrin clots,

but has also been shown to degrade a variety of proteins,

including laminin and fibronectin which comprise the

extracellular matrix. Plasmin also activates latent

metalloproteases such as collagenase which further amplifies

plasmin's degradative effects (Vassalli et al., 1991). The

inhibitor alpha-2-antiplasmin binds plasmin via the high

affinity LBS on kringle I and the active site, and is the

primary physiological regulator of plasmin (Aoki, 1995).

Plasmin(ogen) Receptors

A wide range of eukaryotic cells express surface

receptors for plasmin(ogen) and plasminogen activators. Some

of the eukaryotic cells which have been identified as binding

plasmin(ogen) include platelets, monocytes, lymphocytes,

microglial neurons, human keratinocytes, colonic and breast

carcinoma cell lines, and monocytoid cell lines (Miles and

Plow, 1988; Nakajima et al., 1994; Burge et al., 1992; Burtin

et al., 1988; Miles et al., 1991). Several specific receptor

molecules identified on these cell types are GPIIb/IIIa on

platelets (Miles et al., 1986), the glycolytic enzyme alpha-

enolase on U937 monocytoid cells (Miles et al. 1991) and on

the plasma membrane of rat microglial cells (Nakajima et al.,

1994), and recently a protein homologous to alpha-enolase was

identified on the colonic cell line SW1116 (Lopez-Alemany et

al., 1994). Surface bound plasmin remains proteolytically

active and is protected from inhibition by alpha-2-

antiplasmin. Urokinase and tPA have also been shown to bind

to cells which enhances plasminogen activation and cell-bound

plasmin activity. Receptor molecules have been identified

for urokinase, the urokinase plasminogen-activator receptor

(uPAR), and for tPA, annexin II (Vassalli et al., 1985;

Hajjar et al., 1994). The cell-bound plasmin and plasminogen

activators are hypothesized to play a role in the migration

of cells through the extracellular matrix for physiological

processes such as embryogenesis as well as in pathological

events including the inflammatory response of the host and

metastatic spread of tumor cells.

In addition to the eukaryotic plasmin(ogen) receptors,

there is a wide range of both Gram-positive and Gram-negative

bacteria that have been reported to bind plasmin(ogen)

including group A, C, and G streptococci, Staphylococcus

aureus, Proteus mirabilis, Haemophilis influenza, Pseudomonas

aeruginosa, Neisseria meningitidis and Neisseria gonorrhea,

Escherichia coli, and Borrelia burgdorferi (Ullberg et al.,

1989; Ullberg et al., 1990; Ullberg 1992; Kuusela et al.,

1990; Fuchs et al., 1994). Similar to the findings for

eukaryotic cells, the surface-bound plasmin is not regulated

by alpha-2-antiplasmin. In order to determine the role of

bacteria-associated plasmin in the pathogenesis of infections

caused by these organisms, the plasmin(ogen) binding

components of the cell surface must first be identified.

Several of the prokaryotic surface receptors for

plasmin(ogen) which have been reported thus far have been the

flagella and fimbriae of E. coli (Lahteenmaki et al., 1993;

Parkkinen and Korhonen 1989), the OspA protein of B.

burgdorferi (Fuchs et al., 1994), and certain M-related

proteins of group A, C and G streptococci (Berge and

Sjobring, 1993; Nasar et al., 1994).

q- if if, Rarkarniinri ?nd ( na1,

Our laboratory was the first to demonstrate that group A

streptococci can bind active plasmin with a high affinity on

the bacterial surface (Lottenberg et al., 1987). We have

also shown that group A streptococci grown in human plasma

can activate plasminogen via streptokinase, and subsequently

capture plasmin activity on the bacterial surface (Lottenberg

et al., 1992b). Many group A streptococci, including strain

64/14, specifically bind plasmin or lys-plasminogen with a

reduced affinity for native glu-plasminogen (Broder et al.,


Analogous to the eukaryotic plasmin receptors, surface

bound proteolytic activity can no longer be regulated by

physiological plasmin inhibitors. This unregulated protease

bound to the streptococcal surface may facilitate the

movement of bacteria through normal host tissue barriers.

Recently, a candidate plasmin receptor protein was

isolated from mutanolysin extracts of group A strain 64/14 by

our laboratory (Broder et al., 1991). A gene encoding this

protein, plr, has been cloned, sequenced, and expressed in E.

coli (Lottenberg et al., 1992a). The predicted amino acid

sequence of Plr exhibits extensive homology to the glycolytic

enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

The goals of this study were to characterize Plr and to

determine its contribution to the plasmin binding phenotype

of group A streptococci. Although GAPDH molecules from a

wide variety of organisms had been extensively examined,

GAPDH from pyogenic streptococci had not been characterized.

Therefore this study addresses, on both a biochemical and a

genetic level, the relationship between Plr isolated from

mutanolysin extracts and the cytoplasmic GAPDH of group A

streptococcal strain 64/14. Information generated from these

studies was utilized in developing a strategy to generate

isogenic mutant strains of group A streptococci. Genetic

manipulation of recombinant plr in vitro yielded mutant Plr

molecules that were deficient in plasmin binding activity

relative to Plr, yet retained GAPDH enzymatic activity.

These mutations were introduced into the group A strain 64/14

and the resulting strains were compared to wild-type bacteria

for the ability to bind plasmin.


Plr, a Mr -41,000 protein, was previously identified as

a candidate plasmin receptor from group A streptococcal

strain 64/14 (Broder et al., 1989). The gene encoding this

plasmin binding protein, plr, was cloned and the deduced

amino acid sequence of Plr had extensive homology with the

glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase

(GAPDH) (Lottenberg et al., 1992a). The goal of this series

of experiments was to characterize Plr from group A

streptococci strain 64/14. Because of the homology of Plr to

GAPDH, we hypothesized that Plr could be a functional GAPDH

enzyme. Group A streptococci could contain multiple isozymes

of GAPDH potentially serving different functions, and may

harbor more than one gene encoding this protein(s). To

definitively assess the role of Plr as a plasmin receptor, a

genetic approach was required and therefore a more extensive

characterization of Plr was necessary. I have isolated

cytoplasmic GAPDH of streptococcal strain 64/14 and examined,

on both structural and functional levels, its relatedness to

strain 64/14 Plr obtained from mutanolysin extraction, and to

recombinant Plr expressed in E. coli.

Glycolytic enzymes and the genes encoding them have not

previously been examined in group A streptococci. Studies

presented in this chapter demonstrated that Plr isolated from

cell wall extracts by mutanolysin treatment was identical to

GAPDH found in the cytoplasm of strain 64/14. Furthermore,

this protein was encoded by a single open reading frame on

the streptococcal chromosome and appears to be essential for

viability. These studies have provided novel information

about this glycolytic enzyme of group A streptococci and also

a basis for the initiation of mutagenesis on plr studies

presented in Chapters 2 and 3.

Materials and Methods

Bacterial strains and growth conditions. Group A

streptococcal strain 64/14 is an M-untypable clinical isolate

that has been previously mouse-passaged 14 times (Reis et

al., 1984). Other Group A streptococcal strains used in this

study are clinical isolates described previously (Wang et

al., 1994). E. coli X6060 [F' (traD36 proAB lacIq

AlacZM15)::Tn5 (Kmr)/araD139 A(ara leu) 7697 AlacX74 AphoA20

galE galK recAl rpsE argE (Am) rpoB thi ] and E. coli X2602

(LE392) [F- (hsd R514 (r-, m+) supE44, supF58, lacYl, or

D(lacIZY)6, galK2, galT22, metB1, trpR55] were used for

transformation and gene expression.

Streptococci were grown as a standing culture to

stationary phase at 370 C in Todd-Hewitt broth containing

0.3% wt/vol yeast extract (THY) or in a chemically defined

medium (CDM) (Van DeRijn et al., 1980) containing 1% wt/vol

glucose as the carbon source. In some experiments,

streptococci grown to stationary phase were inoculated into

CDM containing 0.5% wt/vol succinate and 0.5% vol/vol

glycerol, 1% wt/vol acetate, or 1% wt/vol pyruvate in lieu of

glucose. Standing cultures were incubated for several days

at 370 C. E. coli X6060 and E. coli X2602 were grown as a

shaking culture to stationary phase at 370 C. Antibiotics

were added at concentrations of 10 Rg/ml of kanamycin, 30

gg/ml chloramphenicol, and 34 Rg/ml tetracycline where


DNA manipulations and constructions of plasmids.

Plasmids were constructed using DNA techniques performed by

standard methodology (Maniatis et al., 1989).

The plasmid pRL015 contains a 2.7 kb DNA fragment

isolated from group A strain 64/14, which includes plr and

its flanking chromosomal sequences (see insert of figure 1-

5), ligated into the EcoRI restriction site of the low-copy

cloning vector pYA2204 (Galan et al., 1988).

The plasmid pACYC184 is a medium copy E. coli cloning

vector with both chloramphenicol and tetracycline resistance

genes and the origin of replication from plasmid pl5A (Chang

and Cohen, 1978).

A 2.3 kb BamHI-HindIII fragment from the 2.7 kb insert

of pRL015 was subcloned into the BamHI-HindIII restriction

sites of pACYC184 to generate pRL024. The plasmid pRL024

encodes chloramphenicol resistance (cmr) and plr is

transcribed in the opposite direction of the cmr gene


The 2.3 kb BamHI-HindIII insert of pRL024 was excised,

blunt ended, and subcloned into PvuII digested pACYC184 to

generate pRL026. The putative promoter of plr transcribes in

the opposite direction to the tetr promoter of pRL026.

A 2.3 kb EcoRI fragment containing a n kanamycin

resistant (Kmr) gene cassette was blunt ended and subcloned

into the unique PvuII site of pRL026 located within the plr

open reading frame (ORF) at bp 420. Therefore any potential

transcription of plr was interrupted by the Q Kmr cassette.

The cassette was originally from the plasmid pHP452-Km, which

is a derivative of the streptomycin resistant (Smr)

interposon of plasmid pHP450 (Fellay et al., 1987). The Smr

gene had been replaced with the Kmr gene of Tn5 to yield

pHP45Q-Km. The Kmr gene of the Q cassette is flanked by

transcriptional and translational terminators and the gene

product is functional in both E. coli and streptococcal


Electroporation of DNA into group A strain 64/14. DNA

was electroporated into strain 64/14 following the method of

Simon and Ferretti (Simon and Ferretti, 1991) or a similar

protocol by M. Caparon, St. Louis, Mo. (personal

communication). A 5 ml starter culture of THY and 20 mM

glycine was grown as a standing culture at 370 C, ON. The

culture was then diluted with the same medium so that the

initial optical density (O.D.) was 0.06 to 0.08 when measured

at an absorbance of 600nm. The culture was incubated for 1 to

2 hrs at 370 C until the O.D. 600nm was approximately 0.20.

Bacteria were pelleted by centrifugation at 140 C, suspended

in 5 ml of the spent culture media, and then heat shocked

for 9 min at 430 C. Cells were washed two times in 15%

glycerol and suspended in a final volume of 1 ml with 15%

glycerol. Between one and five micrograms of DNA and 100 il

of cells were added to pre-chilled electroporation cuvettes,

0.5 cm slit width (Biorad). The cells were subjected to a

single pulse of 1.75 kV, 400 ohms. The cuvettes were

subsequently kept on ice for 45 min, the bacterial suspension

was diluted in 10 ml fresh THY broth, and then was incubated

at-370 C for 1 hr. Bacteria were pelleted by centrifugation,

suspended to 0.5 ml in THY broth and plated on THY agar

containing 500 ig/ml kanamycin.

Both circular and linear pRL027 DNA were used in

electroporation experiments. Linear DNA was prepared by

either digesting the plasmid at the unique Nco I restriction

site, or alternatively by amplifying a 3.3 kb DNA fragment

consisting of the plr ORF and the Q cassette of pRL027 by

PCR. The oligonucleotide primers RL22 5'-GTTAATACCAATAACTAC


were complementary to the 5' and 3' ends, respectively, of

the plr ORF and were used in the PCR.

Protein purification. Streptococcal Plr (sPlr) was

isolated from a mutanolysin extract of strain 64/14 prepared

by a modification of a previously described protocol (Broder

et al., 1991). Bacteria were pelleted by centrifugation,

washed 3 times with phosphate buffered saline (PBS), pH 7.4

and suspended in PBS containing 30% raffinose (to prevent

cell lysis), 1 mM PMSF, and 1 mM TLCK. Mutanolysin (Sigma

Chemical Company, St. Louis, Mo.) was added at a

concentration of 180 units per gram wet weight of bacterial

pellet to degrade the peptidoglycan cell wall. The

suspension was mixed gently for 2 h at 370 C to release cell

wall-associated components. Protoplasts were pelleted by

centrifugation, and the supernatant was filtered through a

0.2 gm filter. Ammonium sulfate was added to the filtrate to

60% saturation and stirred slowly overnight at 40 C.

Precipitated proteins were removed by centrifugation at

15,000 X g for 10 min at 40 C. The remaining supernatant was

dialyzed extensively in PBS and contained primarily sPlr.

Recombinant Plr (rPlr) was separated from the majority

of E. coli proteins as described previously (Lottenberg et

al., 1992a). Briefly, ammonium sulfate was added to a

soluble lysate of E. coli X6060 (pRL024) to 55% saturation,

and stirred gently overnight at 40 C. Precipitated proteins

were pelleted by centrifugation and rPlr remained

predominantly in the supernatant.

Streptococcal cytoplasmic GAPDH (sGAPDH) was purified

from strain 64/14 protoplasts generated by the mutanolysin

treatment described above. Protoplasts were washed three

times with 10 mM potassium phosphate, pH 6.8 and lysed by two

passages through a French pressure cell. Unlysed protoplasts

were removed by centrifugation at 8,000 X g for 10 minutes

and the resulting supernatant was subjected to

ultracentrifugation at 30,000 X g for 30 min to remove the

remaining insoluble material. sGAPDH was purified from the

soluble extract by NAD+ affinity chromatography (Comer et

al., 1975). Briefly, NAD+-agarose (Sigma) was hydrated in 10

mM potassium phosphate buffer, pH 6.8, washed extensively,

and loaded into a chromatography column. Protein extracts

were incubated in the column matrix for 1.5 h at room

temperature (RT) by end-over-end rotation. The matrix was

washed extensively with phosphate buffer to remove unbound

proteins. Bound proteins were eluted from the column by the

addition of 10 mM NAD+ (Sigma) to the phosphate buffer.

The 41-kDa protein(s), sPlr/sGAPDH, used for amino acid

sequencing and amino acid composition analysis was isolated

from a whole cell extract of strain 64/14. Bacteria were

incubated with mutanolysin, and the mixture passed through a

French pressure cell twice. Insoluble material was pelleted

by centrifugation. The supernatant was loaded onto a NAD+

affinity chromatography column (as described above) to

isolate sPlr/sGAPDH.

SDS-PAGE and protein blotting. Sodium dodecyl sulfate

polyacrylamide gel electrophoresis (SDS-PAGE) was used to

resolve proteins (Lammeli, 1970). Proteins were identified

by staining with Coomassie brilliant blue. Proteins resolved

on polyacrylamide gels were prepared for electrotransfer to

nitrocellulose membranes by equilibrating the gel in 25 mM

Tris-HCL, 0.2 M glycine, pH 8.0 with 20% vol/vol methanol.

Proteins were electrotransferred from gels to nitrocellulose

membranes using a Trans-Blot cell (Bio-Rad, Richmond, CA).

Membranes were soaked in NET-gel (50 mM Tris, 150 mM NaCl, 5

mM EDTA, 0.05% vol/vol Triton X-100, and 0.25% wt/vol

gelatin) prior to incubation with either primary antibody or


For the Western blots, polyclonal mouse anti-sPlr

antibody (Broder et al., 1991) was used as primary antibody;

followed with goat anti-mouse IgG (Organon Teknika Corp.,

Durham, NC) as secondary antibody which was detected with 1251

labeled protein G (Sigma).

Human glu-plasminogen (American Diagnostica, Inc.,

Greenwich, CT) was converted to plasmin using urokinase

(Sigma) as the plasminogen activator (Broder et al., 1991)

immediately prior to incubation with blots. Plasmin ligand

blots were performed by incubating approximately 50,000

counts per minute of labeled plasmin per ml NET-gel with

nitrocellulose membranes containing proteins of interest for

1 hr at RT. Membranes were then washed three times with NET-

gel and exposed to autoradiography film overnight at -700 C.

Radiolabeled proteins. Protein G and plasminogen were

radiolabeled with [125I]Na (Amersham, Arlington Heights, IL)

by a lactoperoxidase reaction using enzymobeads (Bio-Rad),

and labeled proteins were separated from free label by gel

filtration using a PD-10 column (Pharmacia Biotech Inc.,

Piscataway, NJ).

Peptide map analysis. Peptide maps were generated for

sPlr, sGAPDH, and rPlr with V8 protease following the method

of Cleveland (Cleveland et al., 1977). Purified proteins

were electrophoresed on a SDS-10% polyacrylamide gel.

Protein bands were visualized by staining with Coomassie

brilliant blue, and the 41-kDa proteins were excised from the

gel. Gel slices were incubated for 30 minutes in buffer

(0.125 M Tris-HCL, 0.1% wt/vol SDS, 1 mM EDTA, pH 6.8). Gel

slices were loaded into the wells of a SDS-polyacrylamide gel

(4% stacking gel and a 15% separating gel) and overlaid with

buffer containing 20% vol/vol glycerol and bromophenol blue.

Five micrograms (15 U/mg) of V8 protease (Calbiochem-

Novabiochem Corporation, San Diego, CA) in 10% vol/vol

glycerol were added to each well. Samples were concentrated

in the stacking gel by electrophoresis, and the digest was

allowed to proceed for 3 h at RT before resuming

electrophoresis. Following electrophoresis, proteins were

visualized by staining with Coomassie brilliant blue.

Amino acid sequencina. The NH2-terminal amino acid

sequence had previously been determined for sPlr (Lottenberg

et al., 1992b). The 41-kDa protein isolated from the whole

cell extract and rPlr were electrophoresed on SDS-

polyacrylamide gel and electrotransferred onto an Immobilon

PVDF membrane (Millipore Corp., Lakeland, FL) in the presence

of 10 mM MES, pH 6.0 in 20% methanol vol/vol. Proteins were

then subjected to microsequencing using automated Edman

chemistry at the Interdisciplinary Center for Biomedical

Research (ICBR) Protein Chemistry Core Laboratory at the

University of Florida, Gainesville.

Amino acid composition. The amino acid composition was

determined for sPlr/sGAPDH and rPlr. Proteins were

electrophoresed on SDS-polyacrylamide gel and then

electrotransferred onto Immobilon PVDF membranes. The 41-kDa

proteins were visualized with Coomassie brilliant blue and

excised from the membrane. Amino acid composition was

performed by the ICBR Protein Chemistry Core Laboratory,

University of Florida, Gainesville. Proteins were hydrolyzed

in 6N HCL with 1% phenol for 22 hr at 1100 C, and amino acid

composition was determined using the sodium buffer system

with a Beckman System 6300 high performance analyzer.

GAPDH enzymatic assays. Purified proteins were assayed

for GAPDH enzymatic activity following the protocol of

Ferdinand (Ferdinand, 1964). Purified proteins in 50 1L

volumes were added to 100 p1 of 20 mM DL-glyceraldehyde-3-

phosphate (DL-GAP), 100 il of 10 mM NAD+, and 750 il of

reaction buffer (40 mM triethanolamine, 50 mM Na2HP04, and 5

mM EDTA, pH 8.6). Negative control assays were performed as

above without the addition of DL-GAP. The reduction of NAD+

to NADH was monitored spectrophotometrically at an absorbance

of 340 nm, and absorbances were recorded at 20 sec intervals

for 4 min using a Beckman model DU-70 spectrophotometer

(Beckman Instruments, Inc., Fullerton, CA). Absorbances were

converted to micromoles NADH using the molar absorption

coefficient of 6.22 x 10-3 (Horecker and Kornberg, 1948), and

protein concentrations were determined using a bicinchoninic

acid protein assay (Pierce, Rockford, Ill.) to calculate

specific activities expressed as pM NADH min-1 mg-.

DNA hybridization studies. DNA hybridization was

performed by the method of Southern (Maniatis et al., 1989).

Chromosomal DNA was isolated by the method outlined by

Caparon and Scott (Caparon and Scott, 1991), and some

preparations were further purified on CsCl gradients. Five

micrograms of genomic DNA were used per restriction enzyme

digest. For strain 64/14, digested DNA was separated by

electrophoresis in duplicate 0.7% wt/vol agarose gels and

transferred to nylon membranes by a capillary blot procedure

outlined in the manufacturer's instructions (Gene Screen

Plus, Dupont Boston, MA). The DNA probe consisted of the 1-

kb plr open reading frame (ORF) which was amplified by PCR

using the plasmid pRL024 as DNA template. The probe was

labeled with [32p]dCTP (Amersham, Corp., Arlington Heights,

Ill.) using a random priming kit (United States Biochemical

Corp., Cleveland, Ohio). The probe was then incubated with

the membranes in the presence of 10% wt/vol dextran sulfate,

50% vol/vol formamide, and 0.5% wt/vol SDS for 18 h at either

RT (low stringency) or 550 C (high stringency). Membrane

washes were consistent with hybridization temperatures of RT

or 550 C. Reactive bands were visualized by

autoradiography. Southern blots of the 19 other

streptococcal isolates were performed by the same method but

with a hybridization temperature of 420 C and the washes were

performed at 650 C.


Immunochemical analysis and plasmin binding activity of

sPlr, sGAPDH, and rPlr. I first examined fractions of total

strain 64/14 proteins with anti-sPlr polyclonal antibody to

identify proteins immunologically related to Plr. The

antibody detected sPlr in the mutanolysin extract and an

immunologically reactive protein band also migrating at ~41

kDa in the cytoplasmic fraction as shown in figure 1-1.

There do not appear to be any other cross-reactive proteins

in either the soluble cytoplasmic lysates or the mutanolysin

extract. Additionally, the antibody did not detect any

k A B
125. 125.
88.--- 88-
65- 65-
56- 56-

38.5- 38.5-
33- 33-

12 3 123

Figure 1-1. Anti-sPlr antibody reactivity of strain 64/14
lysate fractions. Samples were electrophoresed on triplicate
reducing SDS-10% polyacrylamide gels. One gel was stained
with Coomassie brilliant clue to visualize proteins (A) and
the proteins on other gels were electrotransferred to a
nitrocellulose membrane, blocked, and reacted with mouse
anti-sPlr polyclonal antibody raised against mutanolysin
extracted sPlr and detected with secondary antibody and
[125I]protein G as describe, in the Materials and Methods (B).
Lanes 1 contain soluble -cyoplasmic lysate, lanes 2 contain
the insoluble fraction, and lanes 3 contain mutanolysin
extracted proteins.

cross-reactive proteins in the insoluble fraction containing


To determine if the anti-sPlr immunoreactive protein in

the cytoplasmic fraction was a streptococcal GAPDH (sGAPDH)

and to examine the relatedness of this protein to sPlr and

rPlr, these proteins were purified as described in Materials

and Methods and resolved on triplicate SDS-PAGE. One gel was

stained with Coomassie brilliant blue to visualize proteins

(figure 1-2A) and revealed that the predominant band migrated

at -41 kDa for each preparation. The proteins resolved on

the other two gels were electrotransferred to nitrocellulose

membranes, and probed with either 125I-labeled plasmin or

polyclonal anti-sPlr antibody. As shown in figure 1-2B, all

three 41-kDa proteins reacted with anti-sPlr antibody which

indicates cross-reactive epitope(s). Additionally, each of

the proteins bound radiolabeled plasmin (figure 1-2C)

demonstrating a functional similarity among them. The three

protein samples were then examined at the primary amino acid

level to further determine the extent of similarity among


Analysis of Protein Composition. To determine if

homologies among sPlr, sGAPDH, and rPlr extended to amino

acid residue position, peptide maps were generated by V8

protease. V8 protease cleaves peptide bonds on the

carboxylic side of aspartate and glutamate. Digested

proteins were separated by SDS-PAGE and visualized by

88 1

38 -

1 2 3 1 2 3 1 2 3

Figure 1-2. Anti-sPlr antibody reactivity and plasmin
binding ability of purified proteins. Samples were
electrophoresed on triplicate reducing SSZ-10 1 polyacrylamice
gels. One gel was stai.ed with Coomassie brilliant blue to
visualize proteins (A). The proteins resolved on the other
two gels were transferred to nitrocellulose membranes,
blocked, and reacted with either mouse anti-sPlr antibody
(B), or with [125]plasmin (C) Lanes 1 contain
streptococcal Plr (sPlr), lanes 2 contain streptococcal
cytoplasmic GAPDH (sGAPDH), and lanes 3 contain recombinant
Plr (rPlr).

staining with Coomassie brilliant blue. Digests of sPlr,

sGAPDH, and rPlr yielded identical size peptides (figure 1-3)

revealing conservation of acidic residues throughout the

sequence of the three proteins, as well as approximately the

same number of amino acids between them.

NH2-terminal amino acid sequencing of rPlr and

sGAPDH/sPlr (from a whole cell preparation) was performed to

establish if homologies extended to the conservation of amino

acid sequence. Analysis of rPlr and sGAPDH/sPlr revealed the

identical NH2-terminus as that of the previously determined

sequence of sPlr. Unambiguous sequence was obtained for all

samples. The amino acid sequence of sGAPDH/sPlr was as

follows: V V K V G I N G F G R I G R L A F

R R I. Valine was also the NH2-terminal amino acid for

sPlr whereas the recombinant protein contained an

approximately 50:50 mixture of protein with and without the

NH2-terminal methionine. Although the NH2-termini were

identical, it was possible that amino acid differences exist

elsewhere in the proteins.

In addition to the NH2-terminal identity of the three

samples, amino acid composition analysis of rPlr and the

sPlr/sGAPDH mixture revealed no significant differences in

overall protein composition between samples (Table 1). This

further indicated the relatedness of the proteins at the

amino acid level. However, amino acid composition analysis

38 -1
33.5- "5



?- V8

Figure 1-3. Peptide map analysis of purified proteins
generated by staphylococcal V-8 protease digestion. Purified
proteins were prepared as described in Materials and Methods
and digested with V-8 protease in the gel. Peptides were
separated on a reducing SDS-15% PAGE and identified with
Coomassie brilliant blue. The arrow indicates the V-8
protease. Lane 1 contains sPlr, lane 2 contains sGAPDH, and
lane 3 contains rPlr.

I ir*p:


Table 1. Amino acid composition comparison of sGAPDH/Plr,
rPlr, and SDH.

Number of amino acid residues
amino sGAPDH/ predicted
acid rPlr Plr a.a. seq. SDHa
Ala 25.3 29.6 33 38.1
Arg 19.8 27.2 13 15.5
Gly 35.5 38.7 33 36.6
His 8.0 7.8 7 7.2
Ile 19.4 16.6 21 22.4
Leu 24.5 24.4 22 23.4
Lys 19.7 19.8 20 21.4
Met 3.8 2.1 8 1.8
Phe 12.7 12.0 12 13.8
Ser 20.7 21.1 15 16.8
Val 25.3 24.1 35 36.5
Asn/Asp 49.0 39.2 42 43.3
Gln/Glu 30.5 32.6 27 29.9
Pro 5.5 8.9 10 13.6
Thr 27.7 22.8 26 27.0
Tyr 7.7 8.3 7 9.1

aPancholi and Fishetti, 1992.

may not be sensitive enough to detect small quantitative

differences for some amino acids.

These procedures detected no differences among the sPlr,

sGAPDH, rPlr indicating that they are structurally similar

proteins. The analysis was extended to functional properties

of GAPDH to further examine the -41 kDa proteins.

NAD binding and GAPDH enzymatic activity. NAD+ is

reduced to NADH during the oxidative phosphorylation of

glyceraldehyde-3-phosphate into 1,3-bisphosphoglycerate by

GAPDH. GAPDH contains a NAD+ binding domain in the NH2-

terminal half of the protein (Harris and Waters, 1976).

Streptococcal GAPDH (sGAPDH) was purified from a cytoplasmic

extract prepared from strain 64/14 by NAD+ affinity

chromatography (figure 1-4B). In this experiment, a -41 kDa

protein was the predominant protein in the NAD+ elution

fractions and was confirmed to be a GAPDH by exhibiting

glycolytic enzyme activity (see below). The identical

protocol used to isolate sGAPDH was also utilized to

demonstrate that sPlr was a NAD+ binding protein. A

mutanolysin extract of strain 64/14 was applied to the

affinity column and sPlr was the predominant protein eluted

with 10 mM NAD+ as shown figure 1-4A. Similarly, rPlr was

affinity purified from a soluble lysate of E. coli X6060

(pRL024) (data not shown). Therefore sPlr, sGAPDH, and rPlr

share the GAPDH characteristic of having functional NAD+

binding domains.


1 2 3 4 5 6 7 8 9

B 125


1 2 3 4 5 6 7 8 9

Figure 1-4. NAD' affinity chromatography purification of the
41-kDa proteins. Mutanolysin extracted proteins from strain
64/14 (A) or soluble cytoplasmic material from strain 64/14
(B) were applied to a NAD+-agarose affinity column and bound
proteins eluted from the column with 10 mM NAD+. Proteins
were resolved by SDS-10% polyacrylamide gel electrophoresis
and identified with Coomassie brilliant blue. Lanes 1
contain starting material applied to the column. Lanes 2 and
3 contain wash fractions, and lanes 4-8 contain elution peak
fractions by the addition of 10 mM NAD' to the wash buffer.

The three purified 41-kDa proteins were assayed for

glycolytic enzyme activity to verify that they were

functional GAPDH enzymes. Analysis of purified sPlr, sGAPDH,

and rPlr yielded specific activities of 62.6, 153.1, and

206.3 pM NADH min-1 mg-1, respectively. There was never any

detectable spontaneous conversion of NAD+ to NADH without the

addition of DL-GAP to the reaction mixtures. Therefore the

three purified proteins demonstrate functional GAPDH

enzymatic activity and have specific activities within the

range reported for other GAPDH molecules of prokaryotic

origin (Branlant et al., 1983). Based on both the

structural and functional data, it appears that sGAPDH and

sPlr are the same primary protein in strain 64/14 and that

this protein is identical to recombinant Plr.

DNA hybridization studies. Organisms which express only

a single GAPDH may possess more than one gap gene.

Therefore, DNA hybridization analysis by the method of

Southern was performed on strain 64/14 chromosomal DNA using

a plr probe to examine whether the genome contained a single

or multiple copies of plr. To also identify possible related

genes or psuedogenes both hybridization with the plr probe

and washes of the membrane were done under conditions of low

stringency at room temperature (RT), to allow for maximum

mismatch detection using this technique (figure 1-5A).

Conditions of high stringency at 550 C, were performed for

comparison (figure 1-5B). In lanes containing chromosomal


7,- .0z

4.0- ,b .4.0-
74.0- ..0
3.0o- 0
2.0- V
2.0- 1.6-


Hind ll Xmn I PVUII
PV I U11 EcoRV BanH
Pv Sall mn BamHI

0.2 kb

Figure 1-5. DNA hybridization analysis of strain 64/14
chromosomal DNA to determine gene copy number of plr. DNA
was digested with the restriction enzymes indicated in the
figure, electrophoresed on duplicate 0.7% agarose gels, and
transferred to nylon membranes. The membranes were then
reacted with [32p]dCTP labeled probe consisting of the p1r
ORF, washed, and subjected to autoradiography. One
autoradiograph shows the overnight hybridization and washes
performed at RT (A). The other autoradiograph shows the
hybridization and washes performed at 550C (B) The diagram
indicates relevant restriction enzyme sites located on the
2.7-kb DNA fragment cloned from strain 64/14 which harbors
plr and flanking regions.

DNA digested with restriction enzymes which cut once within

plr (EcoRV, PvuII, and XmnI), the probe hybridized with two

fragments at both RT and 550 C. The probe hybridized with a

single fragment in lanes containing DNA digested with

restriction enzymes which cut outside of plr (BamHI, EcoRI,

HindIII, and Sail). These results are consistent with a

single copy gene. The BamHI/SalI double digest yielded a

single 2 kb fragment that is too small to harbor more than

one copy of plr (see restriction map of figure 1-5).

Therefore, plr is a single copy gene in group A strain 64/14.

To confirm that a single copy of plr is typical for

group A streptococci, a series of nineteen other group A

streptococcal strains isolated from both throat and blood

cultures (Wang et al., 1994) was examined by DNA

hybridization. Chromosomal DNA was digested with BamHI,

Sail, or a BamHI/SalI double digest. The hybridization

pattern using the plr probe for all nineteen strains was

identical to that of strain 64/14 as shown for the four

representative strains SHS 7, SHS 9, SHS 17, and 230041 in

figure 1-6. The BamHI and Sail digests yielded a 10 kb and a

3.0 kb size fragment, respectively, and the BamHI/SalI double

digests yielded a single 2.2 kb fragment for all isolates.

These results indicated that all strains tested contained a

single copy of plr and suggested that this may be typical for

group A streptococci.

10.0 kb


3.0 -



BamH I Sal I BamH ISal I

Figure 1-6. DNA hybridization analysis to determine gene
copy number of plr in group A streptococcal clinical
isolates. Strain 64/14 and four representative strains of 19
strains tested, are shown in this figure. Chromosomal DNA
was digested with the restriction enzymes BamHI, SalI, or a
BamHI/SalI digest; electrophoresed on a 0.7% agarose gel; and
transferred to a nylon membrane. The membrane was then
reacted with a (32p]dCTP labeled probe consisting of the plr
ORF. Following overnight hybridization at 420C, the membrane
was washed, and reactive bands visualized by autoradiography.
Lanes 1 strain 64/14, lanes 2 SHS-7, lanes 3 SHS-9,
lanes 4 SHS-17, lanes 5 strain 230041.

Strateav for insertional inactivation of olr. The goal

of this experiment was to generate an isogenic mutant of

strain 64/14 in plr by insertionally inactivating the gene

with a DNA cassette containing an antibiotic resistance

marker (figure 1-7A). Linear pRL027 DNA, containing the plr

ORF interrupted by the 0 cassette, was electroporated into

strain 64/14, and bacteria were plated on either blood agar

or THY agar plates containing 500 gg/ml kanamycin. Circular

pRL027 was also electroporated in separate experiments as a

control to verify that strain 64/14 was transformable and

that pRL027 could integrate into the chromosome (figure 1-

7B). In contrast to successful transformation with circular

DNA, repeated attempts failed to produce a single kanamycin

resistant colony using the linear DNA. The plasmid contained

over 1 kb of homologous DNA both 5' and 3' to the 0 cassette

which should be of sufficient length to allow a crossover

event to occur. Furthermore hybridization analysis of

chromosomal DNA from an isolate transformed with circular

pRL027 using both plr and Q cassette gene probes indicated

that the plasmid had integrated directly downstream to the

plr gene (data not shown). These results suggested that the

plr gene may be essential for viability in strain 64/14.

Growth media requirements of strain 64/14. By providing

a carbon source which could potentially be utilized for the

generation of ATP by enzymes that function in the glycolytic


strain 64/14

strain 64/14

- S* On61Kmr cas
-- nran 6414 NA mIu~ s

- pACYC 84 ONA s-quels

Figure 1-7. Potential recombination events resulting from
electroporation of plasmid pRL027 into strain 64/14.
Electroporation of linear pRL027 would potentially yield a
double crossover event by homologous recombination, indicated
by the arrows, resulting in inactivation of plr (A),
Circular pRL027 integrated via a single crossover event
occurring downstream of plr on the chromosome,and therefore
retained a full-length copy of plr (E).

I il F,-- Jn.,l~X;;X l- ,,, ------- V ^ s _. _. _



pathway subsequent to the reaction involving GAPDH, it may be

possible to eliminate the need for this enzyme, and thus

provide an opportunity to successfully inactivate the plr

gene. Group A streptococci can be grown in culture using a

chemically defined medium (CDM) (Van Derijn et al., 1980)

containing 1% wt/vol glucose as the primary carbon source.

Strain 64/14 was inoculated into a modified CDM prepared with

0.5% wt/vol succinate and 0.5% vol/vol glycerol. There was no

visible growth after several days of incubation at 370 C.

The bacteria grew only in succinate/glycerol CDM to which

glucose was added. Attempts at growing the streptococci in

CDM containing either 1% wt/vol acetate or 1% wt/vol pyruvate

instead of glucose were also unsuccessful indicating that, of

the reagents used in these experiments, glucose is the only

carbon source which can be transported into and/or utilized

by the glycolytic pathway of strain 64/14. This result was

consistent with plr being an essential gene.


Our laboratory previously demonstrated that group A

streptococci bind plasmin with high affinity (Broeseker et

al., 1988). A plasmin binding protein, sPlr, has been

isolated from mutanolysin extracts of group A strain 64/14

(Broder et al., 1991). The gene encoding this candidate

plasmin receptor has been cloned, sequenced, and expressed in

Escherichia coli. The entire deduced and partially

experimentally determined amino acid sequence of sPlr showed

significant homology to glyceraldehyde-3-phosphate

dehydrogenases of both prokaryotic and eukaryotic origins

(Lottenberg et al., 1992a). GAPDH is an enzyme of the

glycolytic pathway that has been extensively characterized

for a variety of organisms. As expected, it is usually

localized to the cytoplasm or to specific organelles within a

cell, although there are reports of surface localization for

GAPDH in several organisms (Fernandes et al., 1992; Goudot-

Crozel et al., 1989).

In this study the relationship between sPlr and

cytoplasmic GAPDH from strain 64/14 was examined. GAPDH is a

tetrameric enzyme of the glycolytic pathway responsible for

the phosphorylation of glyceraldehyde-3-phosphate to generate

1,3-bisphosphoglycerate (Harris and Waters, 1976). NAD+

binds to GAPDH at a specific site and serves as an electron

acceptor for the substrate during the reaction. By using

NAD+ affinity chromatography, I was able to isolate

streptococcal GAPDH (sGAPDH) from a cytoplasmic extract of

strain 64/14. A single protein, migrating at -41 kDa on SDS-

PAGE was eluted from the affinity column with excess free

NAD+. This 41-kDa protein was compared with both Plr

isolated from mutanolysin extracts of group A strain 64/14

(sPlr) and purified recombinant Plr (rPlr) to evaluate the

relatedness of the proteins. Amino acid analysis of sGAPDH,

sPlr, and rPlr revealed identical NH2-terminal amino acid

sequences for the three proteins by Edman degradation. This

amino acid sequence was in complete agreement with the

predicted amino acid sequence of Plr. V8 protease digestion

of the three purified proteins yielded peptides of equivalent

size when identified by SDS-PAGE indicating conservation of

residues throughout the proteins. In addition to these amino

acid homologies among the proteins, polyclonal antibody

raised against sPlr also recognized both rPlr and the sGAPDH

on Western blot analysis. Furthermore, no cross-reacting

proteins were detected other than sGAPDH and sPlr in a strain

64/14 soluble lysate and mutanolysin extract, respectively.

No differences in primary protein structure of sPlr, sGAPDH,

and rPlr were detected. Although anti-Plr polyclonal

antibody recognizes only a single protein in group A

streptococcal lysates, GAPDHs from a single organism may

differ significantly in amino acid composition and contain

different antigenic epitopes. For example, Trichoderma

koningii expresses two GAPDH isozymes, GAPDH I and GAPDH II.

The glycolytic activity of GAPDH I is inhibited by koninigic

acid whereas GAPDH II is resistant. Analysis of the NH2-

terminal amino acid sequences revealed only 70% similarity to

each other, and antisera against GAPDH II are only weakly

cross reactive with GAPDH I (Sakai et al., 1990).

The streptococcal proteins were functionally alike as

well. sPlr, sGAPDH, and rPlr each bound to a NAD+ affinity

column, possessed GAPDH enzymatic activity, and demonstrated

plasmin binding activity using a ligand blot assay. Based on

these structural and functional identities, we conclude that

sPlr and sGAPDH are the same primary protein. Furthermore,

using the techniques in this study, Plr/GAPDH cannot be

differentiated from rPlr, which was expressed in E. coli from

recombinant plr.

Prokaryotic and eukaryotic organisms may possess single

or multiple GAPDH genes. In this report DNA hybridization

studies revealed that strain 64/14 harbors a single copy of

plr, a gene encoding a streptococcal GAPDH. Furthermore,

studies of 19 group A streptococcal clinical isolates yielded

hybridization patterns identical to those of strain 64/14

indicating that these strains contained a single copy of plr

as well. These results suggest that group A streptococci may

typically possess a single gene encoding Plr. However, the

possibility'remains of a second, highly divergent gap gene

which is present in the streptococcal chromosome but is not

detectable by DNA hybridization analysis using a plr probe.

This situation would be analogous to E. coli which has at

least two GAPDH genes (Alefounder and Perham, 1989). The gap

A appears to be similar to GAPDHs of eukaryotic origins while

the gap B has highest DNA homology with other prokaryotic

GAPDH genes. Only expression of gap A has been reported. In

contrast, Sacharomyces cerevisiae has three GAPDH genes, all

of which are expressed (Holland, 1983).

There are organisms which utilize one GAPDH for

cytoplasmic functions while another GAPDH is targeted for

specific organelles (Michels et al., 1991). In the protozoa

Trypanosoma brucei, one GAPDH isozyme is located in

glycosomes, specialized glycolytic organelles containing the

first seven enzymes of the glycolytic pathway. A second

GAPDH isozyme resides solely in the cytoplasm. The

glycosomal GAPDH contains amino acid substitutions throughout

the protein relative to the cytosolic GAPDH as well as

several additional amino acids at the C-terminus which are

thought to be involved in targeting the protein to the

glycosomes. The genes encoding these GAPDHs have been cloned

and the predicted amino acid sequences are 55% identical.

DNA hybridization studies indicated there are two genes in

tandem that could potentially express the glycosomal form,

whereas only a single gene was detected for the cytosolic

enzyme (Michels et al., 1986). Leishmania mexicana also has

two GAPDH isozymes, a glycosomal and a cytosolic form. The

gap gene arrangement of L. mexicana is similar to T. brucei,

and the predicted amino acid sequences of the two GAPDHs are

only 55% identical (Hannaert et al., 1992). However, in the

present study there was no experimental evidence for either a

second GAPDH product in strain 64/14 or more than one

homologous gene encoding Plr. The DNA sequence of plr does

not contain the putative Gram-positive membrane anchor motif

sequence nor does it reveal a protein-secretion signal

sequence (Lottenberg et al., 1992a). Additionally, Plr was

not detected by Western blot analysis in the insoluble

fraction of a strain 64/14 lysate (figure 1-1B, lane 2)

indicating that Plr may not be associated with the bacterial


GAPDHs and putative GAPDHs have been reported to be

localized on the surface of several organisms. The yeast

Kluyveromyces marxianus flocculates when subjected to heat

stress. The accumulation of a 37-kDa protein in the yeast

cell wall during flocculation has been described (Fernandes

et al., 1992). NH2-terminal amino acid sequencing was

determined for V-8 protease generated peptides of the cell

wall purified protein which revealed over 80% identity to one

of the Sacharomyces cerevisiae GAPDH molecules.

Interestingly, the cell wall-associated GAPDH homologue is

glycosylated, as demonstrated by its susceptibility to

cleavage by endoglycosidase-H, whereas the cytosolic form is

not. Although post-translational modifications occur less

frequently in prokaryotes, analysis for potential post-

translational modifications of Plr/GAPDH has not yet been

performed, but could yield clues regarding putative secondary

functions and/or localization of Plr/GAPDH.

GAPDH has also been reported to be localized on the

surface of Schistosoma mansoni. Indirect immunofluorescence

was used to identify GAPDH on whole S. mansoni and Western

blot analysis identified GAPDH in isolated tegument

preparations using antibody raised against the cytosolic

GAPDH (Goudot-Crozel et al., 1989). Furthermore, antisera

from patients who had severe S. mansoni infections contained

low titers of antibody directed against S. mansoni, whereas

those who were less susceptible to the blood fluke had higher

titers of antibody that reacted with the 37-kDa protein. In

addition to GAPDH, another glycolytic enzyme, triose

phosphate isomerase has been reported to be exposed on the S.

mansoni surface (Harn et al., 1988). The potential functions

of glycolytic enzymes on the surface of this organism have

not yet been defined and the in vivo significance remains to

be elucidated.

Recently Pancholi and Fishetti reported a surface GAPDH

molecule, SDH, from group A streptococci (Pancholi and

Fischetti, 1992). Using methodology similar to the isolation

of Plr, SDH was purified from cell wall extracts of group A

streptococci that were prepared using the muramidase lysin.

SDH was reported to bind to fibronectin, lysozyme, actin, and

myosin on a ligand blot assay; and to have ADP-ribosylation

activity in vitro (Pancholi and Fischetti, 1993). The NH2-

terminal amino acid sequence of SDH is identical to the NH2-

terminal amino acid sequence of Plr with the exception of an

alanine33 in SDH in contrast to an arginine33 in Plr.

In this report, NAD+ affinity chromatography was used to

purify a 41-kDa protein isolated from a whole cell

preparation of strain 64/14. The NAD+ purified protein, as

well as rPlr, were subjected to acid hydrolysis to discern

amino acid composition of the proteins. The differences in

amino acid composition between sPlr/sGAPDH and rPlr were no

greater than differences between the deduced amino acid

sequence of rPlr and the experimentally determined amino acid

composition of rPlr. Furthermore these differences are no

greater than those of the reported amino acid composition of

SDH compared to that of the sPlr/GAPDH and rPlr proteins (see

Table 1). The amino acid composition analysis of SDH

detected 1.8 methionine residues compared to 7 residues in

the Bacillus stearothermophilus GAPDH (Pancholi and

Fischetti, 1992). The authors speculated that a family of

structurally diverse GAPDH molecules may be expressed on the

surface of group A streptococci which may differ from the

GAPDH(s) utilized for glycolysis found in the cytoplasm. Our

amino acid composition analyses also detected similar low

numbers of methionine residues; however, the predicted amino

acid sequence of Plr indicates 7 methionines in the mature

protein which is consistent with the B. stearothermophilus

GAPDH. Four of these seven residues have been confirmed

experimentally in our laboratory by cyanogen bromide

fragmentation of Plr and subsequent NH2-terminal amino acid

sequencing of these fragments (Lottenberg et al., 1992a). It

appears that Plr and SDH are structurally similar proteins.

It is not uncommon for small amino acid changes in GAPDH to

occur among different strains of the same species. For

example, comparison of gap A nucleotide sequences of 13 E.

coli strains and 16 Salmonella strains indicated 0.1 % and

1.1% amino acid differences, respectively, that had occurred

among the strains. A larger number of the predicted amino

acid substitutions resided in the NH2-terminal portion of the

molecule comprising the NAD+ binding domain of the molecule

(Nelson et al., 1991).

To definitively assess the function of Plr as a plasmin

receptor, the gene would be inactivated and the ability of

the bacteria to capture surface bound plasmin would be

compared to the wild-type strain. However, attempts at

insertional inactivation of plr were unsuccessful using

linearized plasmid pRL027 DNA, which contained a functional

selectable marker inserted into the plr ORF. A double

crossover of this DNA into the streptococcal chromosome would

have replaced the wild-type copy of plr. The introduction of

circular pRL027 into strain 64/14, which would integrate via

a single crossover and thereby retain the wild-type copy of

the gene, did yield kanamycin resistant transformants

indicating that the construct could integrate into the

chromosome and that the kanamycin resistance marker was

functional in strain 64/14. Subsequent mutagenesis

experiments (see Chapter 3) demonstrated that linear DNA can

be successfully transformed into strain 64/14 and that the

regions of DNA homology flanking the 2 cassette (300 bp 5'

and 600 bp 3' to the cassette) were of sufficient length to

allow the recombination events to occur. The integration of

circular plasmid pRL027 into the chromosome eliminated the

possibility that the 0 cassette which is flanked by

transcription terminators was disrupting expression of

potentially essential genes downstream of plr. Thus, these

data suggested that plr may be an essential gene.


Taken together the data indicate that a candidate

surface plasmin receptor, Plr, is the same primary protein as

a GAPDH isolated from the streptococcal cytoplasm. The plr

gene which expresses glycolytically active Plr resides as a

single copy in the clinical group A strains tested in this

study. The inability to successfully grow strain 64/14 in

media containing carbon sources other than glucose suggests

that Plr/GAPDH may be required for growth. Furthermore, the

failure to successfully inactivate plr indicates that this

gap may be an essential gene in group A streptococci.


Studies presented in the previous chapter demonstrate

that Plr is a streptococcal glyceraldehyde-3-phosphate

dehydrogenase (GAPDH) enzyme encoded by a single gene on the

chromosome. Furthermore plr appears to be an essential gene

in strain 64/14. Therefore one approach to evaluate the role

of Plr as a plasmin receptor is to generate a non-plasmin

binding Plr mutant that retains GAPDH enzymatic activity and

then introduce that mutation(s) at the plr locus of strain

64/14. A series of genetic alterations of recombinant plr

were constructed and the expressed products were analyzed for

immunoreactivity with anti-Plr antibody, plasmin binding

ability, and GAPDH enzymatic activity. The importance of a

C-terminal lysine residue of Plr in the interaction with

plasmin is clearly demonstrated in this chapter. Mutant Plr

containing alterations of this C-terminal lysine retain GAPDH

activity indicating that the two activities are at least

partially separable. These mutant plr genes are therefore

potential candidates for gene replacement of plr in strain


Materials and Methods

Bacterial strains and growth conditions. The bacterial

strains used in these studies were E. coli X2602, E. coli

X6060 and E. coli DE3. The genotypes of E. coli X2602 and

E. coli %6060 are listed in the Materials and Methods of

Chapter 1. The genotype of E. coli DE3 is endAl, hsdRl7 (rk-

, mk+), supE44, thi-1, recAl, gyrA96, lac [F' proAB, laclq,

ZAM15::Tnl0 (tetr)} (Novagen, Inc., Madison, WI). Bacteria

harboring plasmids were grown as shaking overnight cultures

at 370 C in Luria both. Chloramphenicol, ampicillin,

tetracycline, and kanamycin were used at concentrations of

30, 50, 34, and 10 4g/ml, respectively, where appropriate.

Protein expression from pTrc99C and pMal-c2 derivatives were

induced by the adding a final concentration of 10 mM IPTG to

the Luria broth.

Construction of plasmids. The plasmids used in these

studies are described in Table 2-1. DNA manipulations used

in construction of plasmids was performed by standard

methodology (Maniatis et al., 1989).

A series of 3' end deletions was generated in plr using

exonuclease III by the protocol of Henikoff (Henikoff, 1987).

Exonuclease III will degrade DNA which has a free overhanging

5' end or a blunt end, but not an overhanging 3' end. The

plasmid pRL024 was digested with SphI and BamHI. Both

restriction sites lie downstream of the plr ORF.


0e 0. 0 0 f. 0.
H a a
u E a

W0)0 0) W0
,(0 u0) u 0 I- _o a
o- o O O H-H
'00 00( m 0 0 a2
00 00 00 0 0 42
0)- Oa- ) oX 00 0 0U 4
u)O OO O U 0 0 0 03
0 3 0 W3 0 0 "3 u O c 4-4
0 0 0 0O 0 O W(0 0
SX O 000 4. O U 4

4OL 40 W (00W 04-4 c I0
0W0 CW 00 O U CO W

-00 0r.0 4- 0 C 0 r- 0.v
S0 0 4 00v O4 u) 0 00
O- X u O -H 2O r- c m0
JH 0 4-H 4-j O C- 30 -,1 U

a0 a a)1 ua C: 44 u r.

*dO0.-HW 00 N -H 0(4 0.0-H O 0 O-) l.
5 4J 0) ( o W 44- 0 n O- -H n o
ua o- o c u4 o c e- a L u c ( f
Q) -r M M00 -1 M 0 0 0) 3 u0 0 0 *00 ( u
43 4 ( .rl U (0l r ( a -- 4 0 -H0 4--H
p -^ 0 -4. *r a 30 U 0) =- H 0 Hc
4 -- -( (W 4 a W0 ( 0NO 4 F
0a) 01 a) 0 0 0 ) ) 0) 4- 44 (a m401 0,- i mH
l n 4 o O-.( 4-)T "0 4 ( o -H c ca
)(N cq 4m N -W m 0 (0 > C -;
H r( 0E o-H4 -

O o) 0 c: oW -1 o o 0 ) 3 4 o a)

V 000- 4-.000 0-f 0o -H 0 C: 0 0 4-) 4-4
04 0 0- A c 4-) C04-) 0) (0) 0 0 M 0-i ,01
44 l 0) a ) 1 04 ( 0 1 cP u 0 4) 4a
0 t) mX 04 > C c IC
Sa ,- n CN w n 4- (I n 4-1 Qa O) E-| O N
H C)c- l C)O .1 mo-14 C 3 44u r .
4 0 -- 4-4 4-.-0 4-. i -0 4-) Z

H0 0W0 00W -HOO 04 2- W
o 4C a) 42 ac 2 C0 S0 00E- 4 o )
4J C 4 3 04') 3 C4 D U 0 0 F 0)

W WO WWW 4WW '0 4 U W 30
0 HO O O() CO O ( _r O O Q4 c 04-l 4

(4J 4J 0 4 00 (4 0 0 4W Wc W 040
Q0) 4-) 4 0--1 01 4-W Q) 4J 0 .'0 ) 44l (a c c C-

0u E 4( -0- ) 4j 'a u 4J cO WO) C WO ( 4 0 0) .)

0 O)O 0HO0 0)O 0 0(N .>, .04 C)>
t (N O; C Cd O'- rQ O CnD c 0)OC-) 0T 0M :O-(42 r0 u- 0-43 0 Q)
So o 0 0 oU o (- j-_ o o a
le U 4J U M 4-) U Uui E- El-1 0 El E J

4 oo oX o m A m U

i c n (3) M 0m
0 0 0 0 0 0
u 1 >- .2 .21 >

0. 0 0. 0 0. 0.

Q. 04 Q. 04 0.

Sa0 o o0
m 0 0 1--n, ,. 0
a, C 0) rN 0
-1 : 0
a- 0 .-H 0 0- .3 C C c m

S 0 0 0-4 0 O C0 0o
-H1 O E .0 a 0C 0 UO C o- 01
4H O >1 X w 1 0 0H100 -H Q0 0- CN
=$CL<) 40 '4 m t31 1) 1-44l 1 p, (1
4J X 1 0. O 0. M 4)0 0 E-3.-4 t)
>'OI-H 11 c o o 0 c 0 o > 103

m e. 0 00 a, ) 0 0. 13 w ,4-0
00 4 C 0 0 r I : ) c H o
11343 (c C w *H 0o UJ 41) C r-

E Ij o e 41-, 0 e B )00 00 ) 4-
T3 a z < u ( f () E E m) :3 4 (1) fu

) 0)0 Oa Oe a Os C 0 4- iB
O.. oi n- -- s
4 4 1) C m 0 a O 0 00 -30
0 r. 1 0-3 0o o n C-o- (. a) 3,3 .0 c

H4 44 e wI 41 4 O -l 4 J C W --
4 E .'0 -4 ?UJ .-' 10 0 3 0 (Z o CN ) m
4J C 4 C: C; 4 l4 O 0 4.J M

t -H ) O O .H -H a (13 -O1 C 0 0
00 E-4 U 0 0E- 3 w 1 B-0 0C 4404 0

-I ) 0 0 00 1 0 4 x0 -H-H 0
m ~l -H 11.0 N-4H NW 0 B*4SU0'C S C 7i
Q., 1 4 0 e4 4J 0
o o 41 a < 4 (a H M 4-q -H

04- A 0 00 0. 4-I4 -H 0. 0
0U 0 c e 4 4 a O
I) 11 3 0 14 'i T

4- C x o o o ( "0
E4 E C :x a0 00 '0o41 4-%4 Z a
0 O>,-H-H 340 00 00 a) 0 113 o H -H 1
NH -- -4 .,HO -H-H -H- H c 0 0r '( I. 4
0130 011 131 0 a,00 0001 a o-H
44 CC 4J co (a 4- C -P U
(0 -O 0( 0 H- U H 431 4-u4-l -H u
c4 04cC 0 0 c a) c 4-4 -H 4-3 (D 0 (0mo (
110 -H a 03 0 3 a)1 a -m
c-H c E C O 0 w 1 *) 43 HE-E > -H HO
s 04 0 o 4 oo -O 4o 3 a (0 04- 1 3 4-- o 1 0- c
a UL C C 3 O -,f E 0 & -1l 1 .-
4J0-r O] Q) O a) Q) 3 ca w EH -,r
C 4 .1 4) OC 01 1O O1 -0 to ) z 3:< <40 -
.q 4J4l 4 4 0L 11 -H C r_ r '0 4 3-11 3 3 4, 1 "0
Hn4 a u >4n -0 -0 o 3> .) n f-4 a
(Q 0 4J 44 X E E > 4 ) 4- <1) -
o -. H a0 43H n- C1)-4c Vc
) > a r- 10-- 11 0 1-0004-
( C- 4 4- H 4J 3 1 4- -H r- 0 3 0 .0 -H
3)wN aV mI m l m-4 e S 1 4- 3 0.
r- 4- .-H- M1 A HiN 13) 31 0 n j4J i 1D 1 o1 0 0
o tC a)o a 0 T m C Wc W-i

3 $4W 0 0 N 0
-.-i 0-OI 04 U C) Ci
-. 40 04 04 I I 0
0 UH C 7 0 .0
CN UM N O U ct 1

0 o a0 o o 1 ar
o 0 4 0 0 0 k0

o W o o o o o

E-iE u
( p3 0

TABLE 2-2. Oligonucleotide primers used for PCR mutagenesis
of plr.

DNA sequence of oligonuceotides used to
Plasmid amplify plr inserts










aThe forward primer is listed first. The plasmid pRL024 was
used as DNA template for PCR reactions.


Exonuclease III was added to buffer containing linearized

plasmid DNA. Aliquots were removed and the reactions

terminated at 30 sec intervals for 10 min. Plasmid ends were

blunt ended with S1 nuclease treatment followed by the

addition of Klenow fragment and dNTPs. The plasmids were

ligated and transformed into E. coli X6060. Protein products

expressed by transformants were examined by subjecting E.

coli lysates to electrophoresis on polyacrylamide gels, which

were stained with Coomassie brilliant blue to visualize the

proteins, and for immunoreactivity with anti-Plr polyclonal

antibody by Western blot analysis. Three plasmids with

apparent 3' end deletions in plr, pRL036, pRL037, and pRL038,

were chosen for further study. There was no stop codon

introduced before ligation of digested plasmids, and

therefore translation of plr mutations continued into vector

derived-codons until a termination codon was reached. The

predicted amino acid sequence derived from experimentally

determined DNA sequence of the plr 3' end regions indicated

that the expressed Plr mutations contained C-terminal amino

acid residues encoded by vector derived sequences. The

plasmids pRL036 and pRL038 have 228 and 114 base pairs (bp),

respectively, deleted from the 3' end of wild-type plr. The

resulting expressed proteins, Plr-36 and Plr-38, represent 76

and 38 amino acid deletions of the wild-type Plr C-terminus.

There are predicted to be 23 amino acids fused to the C-

terminal end of Plr-36 and Plr-38; the C-terminal five

residues consisting of asparagine, alanine, glycine, valine,

and alanine. The plasmid pRL037 has 4 bp deleted from the 3'

end of plr. Its product, Plr-37, has only the wild-type C-

terminal lysine residue deleted. The C-terminus of this

protein is fused to 17 vector-encoded amino acids which

terminate with the predicted sequence tyrosine, tryptophan,

alanine, alanine, and serine.

Site-directed and deletion mutagenesis by PCR was

performed by designing DNA primers containing restriction

enzyme sites for cloning and complementary sequence to plr

except for the desired mutation or altered translational

start site. Primer sequences are listed in Table 2-2. The

plasmid pRL024 was used as the DNA template for PCR.

Amplified fragments were subcloned into the vectors pTrc99C

(Aman et al., 1988) or pMal-c2 (New England Biolabs, Beverly,

MA) as indicated in Table 2-1. In some instances, due to

difficulty in direct cloning, the PCR products were first

ligated into the vector pT7Blue (Novagen, Inc.), and inserts

of pT7Blue plasmid were then excised and subcloned into

pTrc99C or pMal-c2. The vector pTrc99C contains the lacI

gene, which represses the trc promoter unless induced with

IPTG. A NcoI site (containing an ATG start site) at the 5'

end of the multiple cloning site allowed for the expression

of Plr mutants lacking additional vector-encoded amino acids.

The multiple cloning site of pMal-c2 lies 3' to the malE gene

resulting in expression of maltose binding protein (MBP)/Plr

fusions when plr inserts were cloned in-frame with malE. The


wild-type signal sequence of the MBP is deleted in pMal-c2 to

prevent secretion of expressed fusion proteins into the cell

periplasm. The 5' and 3' end junctions of vector and insert

were sequenced to verify the presence of the mutation(s).

A 726 bp internal deletion of plr was generated by

utilizing restriction enzyme sites within the ORF. Digestion

of pRL024 with BstYI yields three DNA fragments. The 0.6 kb

BstYI fragment containing the 3' end of plr and downstream

strain 64/14 DNA sequences was ligated to the 3.0 kb BstYI

DNA fragment of pRL024, which harbors putative promoter

sequences of plr, the 5' end of plr, and pACYC184 DNA

(including the tetr gene and the origin of replication). The

plasmid pSW029 contains the first 99 bp of plr ligated in-

frame to the 3' end 180 bp of plr. The expressed protein,

Plr-29, is a fusion protein consisting of the NH2-terminal 33

amino acids and the C-terminal 60 amino acids of Plr.

Solublization of inclusion bodies and recombinant

protein purification. When many of the plasmids containing

plr mutations were transformed into E. coli, the expressed

proteins formed insoluble inclusion bodies. To isolate the

recombinant proteins, bacteria were first grown in media

overnight, shaking at 370 C. Bacteria were lysed by passage

through a French pressure cell and the inclusion bodies were

concentrated by centrifuging the lysates at 8,000 X g for 10

min. Inclusion bodies were then isolated using a

modification of a previously published method (Leong et al.,

1991). Pellets containing inclusion bodies were suspended in

1% vol/vol Triton-X 100 and rotated for 30 min at RT.

Inclusion bodies were pelleted by centrifugation and washed

two times with dH20. The inclusion bodies were initially

treated with 6 M guanidine. Solublized proteins were then

either dialyzed against 0.5 M guanidine or subjected to an

initial dialysis against 1.5 M guanidine, and then followed

by dialysis against 0.5 M guanidine. The precipitate was

removed by centrifugation, and dialysis was continued against

150 mM KC1, 50 mM Tris-HCL, pH 8.0 followed by 50 mM NaC1, 50

mM Tris-HCL, pH 8.0. Dialyzed samples were filtered through

membranes with a pore size of 0.2 pm, and then stored at -70

C. Recombinant proteins Plr-36, Plr-38, Plr-28, Plr-46, Plr-

49, and Plr-62 were solublized by this method. The

solublization of Plr-29 and Plr-61 was not successful using

this protocol.

Maltose binding protein/Plr fusion proteins were soluble

when expressed in E. coli and were purified by amylose-

agarose affinity chromatography following the manufacturer's

instructions (New England Biolabs). Soluble lysates of E.

coli X6060 (pMal-c2), E. coli X6060 (pRL070), and E. coli

X6060 (pRL071) were applied to the affinity column and

incubated at RT for one hour. Non-specifically bound

proteins were removed by washing the column extensively with

wash buffer consisting of 10 mM phosphate, 0.5 M NaCI, 1 mM

sodium azide, 10 mM betamercaptoethanol, pH 7.0, followed by

elution of bound proteins by the addition of 10 mM maltose to

the wash buffer. Factor Xa cleavage of MBP fusion proteins

was also performed following the manufacturer's instructions

(New England Biolabs).

The soluble Plr mutants Plr-38, Plr-45, and Plr-42

retained the NADI binding activity demonstrated by Plr and

were purified by NAD+ affinity chromatography as described in

Chapter 1.

Gel electrophoresis and Western blot assays. The

labeling of proteins with [125I]Na, the preparation of

plasmin, SDS-PAGE, and Western blot analysis using anti-Plr

polyclonal antibody for immunoblots and [125I]plasmin for

ligand blot assays, were performed as described in Chapter 1.

GAPDH activity assays of cell lysates. E. coli

producing recombinant Plr proteins were grown overnight at

370 C as shaking cultures. Cultures were centrifuged and

bacterial pellets washed two times with 50 mM KP04, pH 6.0

and suspended to 2 ml per gram of pellet with buffer. Cells

were lysed by two passages through a French pressure cell.

Unlysed bacteria were removed by centrifugation at 5,000 X g

for 10 min at 40 C, and insoluble debris was cleared by

centrifugation at 33,000 X g, 30 min, 40 C. Lysates were

assayed for protein concentration and for GAPDH enzymatic

activity following the same procedures outlined for

determination of enzymatic activity for purified proteins in

the Materials and Methods of Chapter 1. Specific activities

are expressed as gM NADH min-1 mg-1 protein extract.


Analysis of the plasmin ligand blot. Standardized

procedures were used to assess immunoreactivity and plasmin

binding activity of recombinant proteins (see Materials and

Methods of Chapter 1). Highly purified commercial

streptokinase (Kabi), recombinant GAPDH of Bacillus

stearothermophilus, and recombinant Plr were examined by

immunoblot and plasmin ligand blot. The absence of anti-Plr

polyclonal antibody reactivity for streptococcal proteins

other than Plr was shown in figure 1-1 of Chapter 1. The

lack of cross-reactivity of the antibody with the

streptococcal plasminogen activator streptokinase was

previously demonstrated (Broder et al., 1991). As depicted

in figure 2-2, additional specificity of this antibody was

demonstrated by its failure to react with B.

stearothermophilus GAPDH, even though this protein has 50%

identity with Plr at the predicted amino acid sequence level.

The streptokinase bound both plasmin and glu-plasminogen. In

contrast, Plr bound plasmin in this assay but failed to bind

significant amounts of glu-plasminogen, indicating a higher

affinity for plasmin compared to glu-plasminogen.

Interestingly, strain 64/14 also has a much higher affinity

for plasmin compared to glu-plasminogen (Broder et al.,


207 .c T0fA 0T C .LA f .TCCr0C aCITCoBg.M


17 I AI T 0T 0 0I70L0GAUl Ir
81955 CgeCTJCTCM-AcG rc AJuua 1023 CrTlUI srTGvXTC s T v v T T D z CTV SgV

,z xI A x T~O YI

Figure 2-1. Nucleotide sequence of plr. The predicted amino
acid sequence of Plr is indicated in single-letter code below
the nucleotide sequence (Lottenberg et al., 1992a).

11116- 116-
6- 86- 6-
,36--/i- 8-- 86-
66- 66-66-
56- 56- 56-

40.5- 405- 40.5- 40.5-
36- 36- 36- 36-

1 2 3 1 2 3 1 2 3 1 2 3

Figure 2-2. Western blot analysis of purified proteins for
anti-sPlr antibody reactivity and plasmin binding activity.
Proteins were electrophoresed on quadruplicate SDS-
polyacrylamide gels. One gel was stained with Coomassie
brilliant blue to visualize proteins (A), while the proteins
on the other three gels were electrotransferred to
nitrocellulose and probed with either anti-Plr antibody (B),
[125I]plasmin (C), or [125I]glu-plasminogen (D). Lanes 1
contain streptokinase, lanes 2 contain Plr, lanes 3 contain
recombinant Bacillus stearothermophilus GAPDH.

1989). The B. stearothermophilus GAPDH, an enzyme highly

homologous to Plr, did not bind plasmin and indicated that

specific regions are required for binding which are unique to

Plr. Therefore, using this assay, binding interactions can

be distinguished between either glu-plasminogen or plasmin

with their respective ligands immobilized on nitrocellulose.

The characterization of mutant Plr molecules generated in

this study are summarized in Table 2-3.

In vitro analysis of recombinant Plr fusion mutations.

To examine the contribution of C-terminal amino acids of Plr

to its ability to bind plasmin, Plr-36, Plr-37, and Plr-38

were analyzed by ligand blot assay. Plr-37 was soluble when

expressed in E. coli and was separated from irrelevant

proteins using NAD+ affinity chromatography. Plr-36 and Plr-

38 formed insoluble inclusion bodies in E. coli and were

solublized after isolation of the inclusion bodies prior to

SDS-PAGE. All three C-terminal truncations of Plr reacted

with the anti-Plr polyclonal antibody indicating that at

least some of the antigenic epitope(s) of Plr lie within the

NH2-terminal 269 of the 335 native amino acids (figure 2-3B).

When probed with 125I-labeled plasmin, Plr-36, Plr-37, and

Plr-38 demonstrated reduced binding activity compared to Plr,

indicating that the C-terminus of Plr may play a role in

plasmin binding (Figure 2-3C). Because Plr-37 has only the

wild type C-terminal lysine deleted preceding the additional

Table 2-3. Summary of Pir mutations and the resulting
phenotypes examined in this study.

Solub. Inu.obrN

P,~..In DnO.~n proa' n.=b subh fdlpb s Indn==b
G wiyp.dn v 4 NA m

-Pk38 3 Gd-t. d -on -
PW38 76a 0 nn~daktedn


PR-4 CM tNA N. D.

P 135 UH.,*na. dlMA N.D. N
Pft-6 13 t+ NH4-*ft~L dWi.lion MA.0 HJ.. *
P-r 33 NJ-nn. dMIon N D.
PI442 fN4rtbm. on ND. + + ILA.
RP-2 242 .nmb aNdN U l a NJD. ND. *

MNIBP+Pnrd *i, + IL
Pi-71 MBP 116 .- NH -4.A dlMan + hLA

fa Pruminofb. 0 OliW.rdu. I ,uAdl*in. 0 No.rr..du.
C MUok i.Lr +pin. I Lt.Z.
b N.D.. ldo: NA.a nl ipl able

kD A B C


38.5-- 7 385-5
33- 33-

1 2 3 4 5 12 3 4 5 12 3 4 5

Figure 2-3. Western blot analysis of exonuclease III-
generated plr mutations. Samples were electrophoresed on
triplicate reducing SDS-10% polyacrylamide gels. One gel was
stained with Coomassie brilliant blue to visualize proteins
(A). Proteins resolved on the other two gels were
transferred to nitrocellulose membranes, which were blocked
and reacted with either anti-sPlr antibody (B), or with
[125I]plasmin (C) Lanes 1 and 5 contain Plr, lanes 2 contain
Plr-36, lanes 3 contain Plr-38, and lanes 4 contain Plr-37.

vector-encoded amino acids, this residue may be important in

mediating the binding of plasmin to Plr. None of the vector-

derived sequences fused to the 3' end of plr-36, plr-37, and

plr-38 coded for lysine residues and therefore the above

results are consistent with a C-terminal lysine contributing

to plasmin binding. The -37 kDa plasmin binding protein

contaminating the Plr-37 preparation may be E. coli GAPDH

which bound to the NAD+ column during Plr-37 purification.

To examine whether other regions of Plr in addition to

the C-terminus were involved in plasmin binding, a 129 amino

acid NH2-terminal truncation of Plr was constructed. The

full length plr ORF and the 5' end deletion DNA fragments

were subcloned individually into the pMal-c2 vector creating

in-frame fusions with the malE gene which encodes the MBP of

E. coli. Expression of these gene fusions yielded the

proteins Plr-70 and Plr-71, respectively. Purified fusion

proteins and a MBP/LacZa fusion were examined for reactivity

with anti-Plr antibody and plasmin binding activity (see

figure 2-4). The proteins Plr-70 and Plr-71 are shown

migrating close to their predicted -Mr of 84,000 and 65,000,

respectively, in figure 2-4A. The MBP has a -Mr of 42,000 and

the LacZa -Mr 10,000, therefore the MBP/LacZa fusion migrates

at -Mr of 52,000. The anti-Plr antibody reacts with both Plr-

70 and Plr-71 but not the MBP (figure 2-4B) indicating

correct expression of the constructs and that

- 88.

38.5- -

38.5-0 *

1 2 3 4



12 3 4

1 2 3 4

Figure 2-4. Western blot analysis of MBP/Plr fusion
proteins. Samples were electrophoresed on triplicate
reducing SDS-10% polyacrylamide gels. One gel was stained
with Coomassie brilliant blue to visualize proteins (A).
Proteins resolved on the other two gels were transferred to
nitrocellulose membranes, which were blocked and reacted with
either anti-sPlr antibody (B), or with [125I]plasmin in buffer
(C) Lanes 1 contain Plr, lanes 2 contain MBP/LacZa, lanes 3
contain the MBP/NH2-terminal Plr deletion fusion protein Plr-
70 and lanes 4 contain the MBP/P1r fusion protein Plr-71.


antigenic epitope(s) of Plr lie within the C-terminal 206

amino acids. The lower Mr reactive bands have been observed

on previous blots, albeit with less intensity, and may

represent partial degradation products of the fusion

proteins. As shown in figure 2-4C, the two Plr fusion

proteins as well as the MBP/LacZa fusion are deficient in

plasmin binding relative to Plr. These results suggest that

a NH2-terminal regionss, in addition to a C-terminal region

of Plr, is also required for wild-type levels of plasmin

binding. NH2-terminal residues may participate by direct

interaction, or indirectly by contributing to conformational

requirements necessary for optimal exposure of the C-terminal

lysine. The removal of the MBP from the Plr-70 recombinant

protein by Factor Xa would yield an additional eight amino

acids fused to the native Plr NH2-terminus. The Plr products

from Factor Xa cleavage of Plr-70 and Plr-71 were identified

by reactivity with anti-Plr antibody reactivity. Consistent

with the analysis of the intact fusion proteins, neither Plr

product bound [125I]plasmin (data not shown). To further

explore the contributions of both NH2-terminal and C-terminal

regions of Plr to the plasmin binding phenotype, it was

necessary to construct non-fusion mutations of Plr.

In vitro analysis of non-fusion Plr mutations. To

address whether the C-terminal and/or the penultimate lysyl

residues of Plr were necessary for plasmin binding (see Plr

sequence in figure 2-1), the plasmid pRL028, which has the

four 3' end codons of wild-type plr replaced with a

termination codon, was constructed. Recombinant Plr-28

expressed in E. coli was purified from inclusion bodies and

was solublized. However the preparation shown in figure 2-5

is an insoluble preparation. Plr-28 was compared to Plr for

plasmin binding ability. The proteins are shown in figure 2-

5A, and reactivity of Plr-28 with anti-Plr antibody was

verified as shown in figure 2-5B. Similar to Plr-37, Plr-28

had reduced plasmin binding compared to Plr, indicating that

the last four amino acids of Plr are necessary for wild type

levels of plasmin binding (figure 2-5C). It therefore

appeared likely that either or both of the C-terminal lysine

residues were responsible for the plasmin binding activity of


To assess the contribution of each of the lysine

residues, plr mutations were generated which substituted a

leucine in place of one or the other lysines. The C-terminal

lysine of Plr was substituted with a leucine in Plr-45,

whereas in Plr-46 the penultimate lysine was replaced with a

leucine and the C-terminal lysine was left intact. Plr-46

was purified and solublized as described for inclusion body

proteins, whereas Plr-45 was soluble and was purified from an

E. coli lysate by NAD+ affinity chromatography as was

performed for wild-type Plr (see Chapter 1). Additionally,

to examine whether the NH2-terminus contributed to plasmin

binding as inferred from the MBP/Plr fusion proteins or

alternatively if the terminal

65 -


1 2 1 2 1 2

Figure 2-5. Western blot analysis of the C-terminal mutant
Plr-28. Samples were electrophoresed on triplicate reducing
SDS-10% polyacrylamide gels. One gel was stained with
Coomassie brilliant blue to visualize proteins (A). Proteins
resolved on the other two gels were transferred to
nitrocellulose membranes, which were blocked and reacted with
either anti-sPlr antibody (B), or with [125I]plasmin (C)
Lanes 1 contain Plr, and lanes 2 contain Plr-28. The
plasmin binding protein migrating at a -Mr slightly faster
than Plr-28 is an E. coli contaminant protein.

lysine(s) was sufficient for wild type levels of binding, a

33 amino acid NH2-terminal deletion of Plr, Plr-49, was

constructed by PCR mutagenesis of plr. Plr-49 was purified

and solublized from inclusion bodies. As a control to show

that the solublization procedure had no effect on the ability

of the mutant proteins to bind plasmin, wild-type Plr was

also subjected to the solublization procedure. Plr, Plr-45,

Plr-46, and Plr-49 reacted with anti-Plr antibody verifying

correct expression of these Plr derivatives (figure 2-6B).

When incubated with 1251 plasmin, Plr-46 appeared to bind

approximately equivalent amounts of plasmin (figure 2-6C) as

wild-type Plr, suggesting that the penultimate lysyl residue

of Plr is not necessary for plasmin binding. However, the C-

terminal lysyl residue is required for plasmin binding as

evidenced by the observed reduction in plasmin binding by

Plr-45. Additionally, EACA completely inhibited plasmin from

binding to all of the proteins immobilized on the

nitrocellulose membrane, consistent with the hypothesis that

lysyl residues of Plr may interact with the lysine binding

sites of plasmin (figure 2-5D). In agreement with the

MBP/Plr fusion proteins, Plr-49 was also deficient in plasmin

binding compared to Plr, suggesting that there are amino

acids in addition to the C-terminal lysine which are

necessary for wild type levels of binding.

If the above hypothesis were correct, then a Plr

mutation which contains the NH2-terminal portion which is

kD A B C D

1 234 5 12345 12345 12345

Figure 2-6. Western blot analysis of NH2-terminal and C-
terminal mutations of Plr. Samples were electrophoresed on
quadruplicate reducing SDS-10% polyacrylamide gels. One gel
was stained with Coomassie brilliant blue to visualize
proteins (A). Proteins resolved on the other three gels were
transferred to nitrocellulose membranes, which were blocked
and reacted with either anti-sPIr antibody (B), with
(125I]plasmin in buffer (C) or with [125Ijplasmin in buffer
containing 20 mM ACA (D). Lane 1 contain Plr, lanes 2
contain Plr-46 which has the penultimate lysine residue
substituted with a leucine, lane 3 contain Plr-45 which has
thie C-terminal lysine relaced with a leucine, lanes 4
contain the 33 amino acid NH-terminal deletion Plr-49, and
lanes 5 contain Pi that has been subjected to the
asolulizatien protocol described in Material and Methods.
solublization protocol described in Material and Methods.

missing in Plr-49, fused to a C-terminal peptide of Plr

should retain its ability to bind plasmin. To test this

hypothesis, an in-frame internal deletion mutant of Plr, Plr-

29, was constructed using available restriction enzyme sites

in plr. Plr-29 contains the NH2-terminal 33 amino acids

fused to the C-terminal 60 amino acid of Plr. Plr-29 was

expressed as inclusion bodies in E. coli, and was not able to

be solublized prior to SDS-PAGE as was done with the other

mutant Plr proteins. This protein migrates at a -Mr of 10

kDa in the Coomassie brilliant blue stained gel of figure 2-

7A and is recognized by anti-Plr antibody (figure 2-7B). A

soluble preparation of Plr-28 is shown for comparison.

Analysis of plasmin binding (figure 2-7C) revealed that Plr-

29 does indeed bind plasmin. The plasmin binding ability of

Plr-29 supports the concept that a NH2-terminal region, in

addition to the C-terminal lysine residue, is required for

wild-type binding.

Interaction of the NH2-terminus of Plr with plasmin may

occur by direct binding of the ligand or possibly indirectly

by positioning the C-terminal lysine in a suitable

orientation to allow high levels of binding. To further

explore NH2-terminal contributions of Plr to plasmin binding,

several additional mutations were made. Shorter 5' end

deletions of plr were constructed and analyzed to more

precisely delineate the contributing regions) or residue(s).

The mutant Plr-61 has the NH2-terminal thirteen amino acids

of native Plr

kD B C
=65= 88.
56--. 88
38.5 38.5- 56-
33- 33- 38.5-

123 123
1 2 3

Figure 2-7. Anti-sPlr antibody reactivity and plasmin
binding ability of the internal deletion Plr mutant, Plr-29.
Samples were electrophoresed on triplicate reducing SDS-15%
polyacrylamide gels. One gel was stained with Coomassie
brilliant blue to visualize proteins (A). Proteins resolved
on the other two gels were transferred to nitrocellulose
membranes, blocked, and reacted with either anti-sPlr
antibody (B), or probed with [1251]plasmin. Lanes 1 contain
Plr, lanes 2 contain Plr-29, and lanes 3 contain Plr-28.

deleted whereas Plr-62 has only the first five NH2-terminal

amino acids removed. Additionally, to test the possibility

that the lysine4 residue in close proximity to the NH2-

terminus of Plr could contribute to plasmin binding, Plr-42

was constructed which contains a glycine4 substituted for the

wild-type lysine4. Plr-42 was expressed as a soluble protein

and was purified by NAD+ agarose affinity chromatography.

Plr-61 and Plr-62 were in inclusion bodies but only Plr-62

was successfully solublized. Protein preparations were

subjected to SDS-PAGE and stained with Coomassie brilliant

blue to visualize the proteins as well as transferred to

nitrocellulose to assay for plasmin binding activity (figure

2-8A). Both soluble and insoluble Plr-62 samples were

assayed for-plasmin binding activity. Plr-49, the 33 amino

acid deletion of Plr which has reduced plasmin binding, is

shown for comparison in figure 2-8B. Plr-42 retained wild-

type levels of plasmin binding, which indicated that the

lysine in close proximity to the NH2-terminus does not

participate in the binding activity as detected by this

assay. Soluble Plr-62 has greatly reduced binding activity

relative to Plr. This result would suggest that residues

within the first five amino acids of recombinant Plr are

required for plasmin binding in addition to the C-terminal

lysine. In contrast, insoluble Plr-61, which is the larger

of these two deletions, and insoluble Plr-62 have similar

levels of plasmin binding as wild-type Plr. These


123 4 5 6
kD B


1 2 3 4 5 6

kD B


Figure 2-8. Plasmin binding ability of Plr mutants with NH2-
terminal alterations. Samples were electrophoresed on
duplicate reducing SDS-15% polyacrylamide gels. One gel was
stained with Coomassie brilliant blue to visualize proteins
(A). Proteins resolved on the other gel was transferred to a
nitrocellulose membrane which was blocked, and reacted with
[125Iplasmin (B). Lanes 1 contain Plr, lanes 2 contain Plr-
42 which has the lys4 substituted with gly4, lanes 3 contain
Plr-49 which is a 33 amino acid NH2-terminal deletion of Plr,
lanes 4 contain a crude insoluble preparation of Plr-61 which
is a 13 amino acid NH2-terminal deletion derivative of Plr,
lanes 5 contain a similar preparation of Plr-62 which has
only the first five amino acids of wild-type Plr removed, and
lanes 6 contain soluble Plr-62.
Figure 2-8. Plasmin binding ability of Plr mutants with NH2-
lanes 6 contain soluble Pit-62.?

differences in plasmin binding between soluble and insoluble

preparations of NH2-terminal Plr mutations have not been

observed for soluble and insoluble preparations of C-terminal

Plr mutants (e.g. Plr-28 shown in figures 2-5 and 2-7).

These results raise the possibility that NH2-terminal

residues are required to position the C-terminal lysine in an

accessible orientation for interaction with plasmin and that

this position effect is lost in mutant Plr after being

denatured with guanidine.

Assessment of GAPDH enzymatic activity of Plr mutations.

To generate a viable strain 64/14 containing a mutated Plr

with reduced plasmin binding activity, GAPDH activity needs

to be retained. Therefore it was necessary to assess the

GAPDH activity of the mutant proteins. Lysates of E. coli

X6060 containing either the vector pACYC184, pRL024 harboring

wild-type plr, or pRL037 containing a four base pair deletion

of the 3' end of plr, were assayed for GAPDH activity in two

separate experiments. The enzymatic activities of E. coli

X6060(pACYC184), E. coli X6060(pRL024), and E. coli

X6060(pRL037) lysates were 11.7, 154.2, and 49.0 pM NADH min-1

mg extract-1 respectively, in one experiment and 17.5, 185.5,

and were 71.8 PM NADH min-1 mg extract -1 in a second

experiment. The average increase in specific activity of

lysates containing Plr over host background alone was almost

12-fold, whereas the Plr-37 preparations had an approximately

4-fold increase over background activity. Additionally,

purified Plr-45 which has the C-terminal lysine substituted

with a leucine and demonstrated reduced plasmin binding

similar to Plr-37, was glycolytically active as well,

although the specific activity was reduced relative to the

purified wild-type protein (data not shown). These

experiments demonstrated that Plr-37 and Plr-45, each

harboring a mutation of Plr that resulted in reduced plasmin

binding ability, still retained glycolytic activity, albeit

with an approximate minimum of 3-fold less activity than that

of Plr, and therefore indicated it was possible to at least

partially separate glycolytic enzyme activity from plasmin

binding activity.

All of the 3' plr mutations which extended farther than

the C-terminal lysine codon were expressed as insoluble

proteins in E. coli hosts regardless of whether transcription

occurred from the wild-type putative promoter contained

within the BamHI-HindIII streptococcal DNA fragment harboring

plr or when PCR-generated mutations were subcloned behind and

expressed from an inducible trc promoter. In contrast, wild-

type plr and mutations of the C-terminal lysine codon were

expressed as soluble, enzymatically active proteins.

Soluble cell lysates of E. coli X6060(pRL036) and E.

coli X6060(pRL037), which express the exonuclease III

generated C-terminal deletions of Plr, showed no increase in

GAPDH enzymatic activity relative to E. coli X6060(pACYC184),

however the majority of mutant Plr visible by Coomassie

brilliant blue staining was visible in the insoluble

fractions of the lysate preparations. Similarly, a soluble

lysate of E. coli X2602(pRL028), expressing the four amino

acid deletion mutant Plr-28 protein, yielded an equivalent

GAPDH activity of 10 gM NADH min-1 mg extract-1 as E. coli

X2602(pTrc99C) harboring the vector alone. Insolubility of

the recombinant proteins could account for the lack of

enzymatic activity due to improper folding and/or an absence

of the proteins of interest in the soluble fraction.

Therefore conclusions cannot be drawn about the role of the

deleted NH2- and C-terminal regions other than the C-terminal

lysine in the GAPDH activity of Plr.

Recombinant Plr was subjected to the solublization

procedure and, importantly, this procedure did not affect the

plasmin binding ability of Plr. However, the procedure did

abolish GAPDH enzymatic activity of the protein, suggesting

that Plr had not refolded properly. Therefore, definitive

conclusions pertaining to the potential enzymatic activity of

purified solublized mutant Plr proteins cannot be made.

The maltose binding protein fusion mutants, Plr-70 and

Plr-71, were also assayed for enzymatic activity. There was

no detectable activity for either protein, indicating the

importance of the native NH2-terminus for enzymatic activity

as well as plasmin binding activity. The absence of activity

may be due to improper folding of the putative nucleotide

binding domain in Plr-71 and the lack of this domain in Plr-

70 resulting in deficiencies in NAD+ binding, stable tetramer

formation, or effects on the substrate binding domain.

A C-terminal lysine residue of Plr was determined to be

essential for the plasmin binding activity of Plr in vitro.

Analysis of additional mutant Plr molecules indicates that a

C-terminal lysine was not sufficient for plasmin binding and

that other regions were necessary as well. Mutant Plr

molecules lacking or containing a leucine instead of the C-

terminal lysine were soluble and retained GAPDH enzymatic

activity. These studies have therefore generated gene

candidates for substitution of plr in vivo to generate

isogenic mutants of strain 64/14.


The plasmin binding protein Plr has been hypothesized to

be a plasmin receptor of group A streptococcal strain 64/14

(Broder et al., 1991). The plr gene was cloned previously

from a strain 64/14 chromosomal DNA library (Lottenberg et

al., 1992a). In this study, through genetic mutations of

recombinant plr, specific amino acid residues of Plr have

been examined for their contribution to plasmin binding

activity. The goals of this series of experiments were to

not only characterize the interactions between Plr and

plasmin, but to also generate a non-plasmin binding mutant of

Plr that retains glycolytic activity. A summary of the

mutations generated in this study are presented in Table 2-3.

The amino acid lysine can efficiently elute plasmin from

the surface of group A streptococci. Lysine and lysine

analogs also prevent the association of plasmin with the

bacterial surface (Broeseker et al., 1988). This inhibition

of binding occurs in a concentration-dependent manner that

suggests the involvement of the high-affinity lysine binding

site (LBS) of plasmin. Furthermore, the domain of

plasmin(ogen) facilitating the interaction is localized to

the heavy chain which contains the LBS's (Broder et al.,

1989). The deduced amino acid sequence of Plr reveals the

presence of lysine residues in the C-terminus (see figure 2-

1) that could potentially mediate the reversible binding of

plasmin observed for intact streptococci and purified Plr

(Lottenberg et al., 1992a). As demonstrated in figure 2-6,

the presence of EACA, a lysine analog, inhibited the binding

of plasmin to Plr, implicating the participation of either or

both of these lysyl residues in this interaction. The

results of analyses of C-terminal mutations of Plr supported

this hypothesis. The C-terminal deletion fusion proteins

Plr-36, Plr-37, and Plr-38 were all deficient in binding.

Plr-37 has only the C-terminal lysine deleted, however the

predicted vector encoded amino acids placed the penultimate

lysine residue 17 amino acids proximal to its wild-type

position and therefore prevented interpretation of its

contribution to plasmin binding. Plr-28, a mutant protein

with the four C-terminal wild-type amino acids deleted

including both the penultimate and C-terminal lysine, was

also deficient in plasmin binding and so more definitively

implicated the C-terminal lysines in the plasmin binding

interaction. To assess the role of each of these lysine

residues in plasmin binding, the single amino acid

substitution mutants which contain a leucine residue in place

of the lysine in either the C-terminal or the penultimate

positions, Plr-45 and Plr-46, respectively, were constructed.

The reduction of plasmin binding of Plr-45 compared to Plr

verified that the C-terminal lysine of Plr was necessary for

wild-type levels of binding. In contrast, there was no

difference in plasmin bound by Plr-46 relative to Plr

suggesting that the penultimate lysine residue may not

participate in the binding interaction.

A penultimate lysyl residue, and in some cases arginine,

is predicted from the DNA sequences to be present within the

C-terminal four amino acids of GAPDHs from a wide spectrum of

organisms including the non-plasmin binding GAPDH from

Bacillus stearothermophilus (Branlant et al., 1989). A basic

residue at this location could potentially contribute to the

conformation of the protein and therefore not be available

for interaction with plasmin. Consistent with this

hypothesis, all C-terminal deletions of Plr extending beyond

the penultimate lysine were expressed as insoluble proteins,

as was Plr-46 containing a substitution of this residue.

This is consistent with improper protein folding, although

other factors resulting from high expression of foreign

proteins in E. coli could also account for this effect.

Interactions of plasmin(ogen) with C-terminal lysyl

residues have previously been reported for fibrinogen

fragments, the physiological inhibitor alpha-2-antiplasmin,

and for eukaryotic plasmin(ogen) receptors. By treating

fragments of fibrinogen with carboxypeptidase B (CPB), which

specifically cleaves C-terminal lysine and arginine residues,

Christensen demonstrated that treated fragments would no

longer bind to a plasminogen-Sepharose column (Christensen,

1985). Untreated fragments bound to the column and could be

eluted by the addition of EACA, a lysine analog, thereby

implicating C-terminal lysines of the fragments in the

binding interaction.

The plasminogen activator urokinase, in solution or when

bound to a specific eukaryotic urokinase receptor, can

activate plasminogen to plasmin. However, enhanced

activation occurs when plasminogen has undergone dramatic

conformational change by binding either free lysine or a C-

terminal lysine of a peptide through the high affinity lysine

binding site (LBS) in kringle 1 (Violand et al., 1975). The

study by Pannell et al., which addressed plasminogen

activation by urokinase, revealed that pre-treatment of

fibrin fragments with CPB reduced the rate of plasminogen

activation by urokinase thereby supporting the role of a C-

terminal lysine in plasmin(ogen) binding interactions

(Pannell et al., 1988).

The major physiological inhibitor of plasmin, alpha-2-

antiplasmin (AP), binds to plasmin by a two step mechanism

(Wiman et al., 1979). The first step can be inhibited by

lysine, and it has been shown that the C-terminal lysine of

AP interacts with the high affinity LBS of plasmin. The

second step is a lysine-independent mechanism occurring at or

near the active site of plasmin. Sasaki et al. demonstrated

inhibition of the plasmin:AP complex with a peptide

constituting the C-terminal 25 amino acids of AP (Sasakai et

al., 1986). Additionally, they hypothesized that a

penultimate lysine residue of AP may also contribute to the

high affinity of AP for plasmin. This penultimate lysine

lies seventeen amino acids from the C-terminus of AP.

Sugiyama et al. demonstrated increased inhibition of the

plasmin:AP complex formation using trypsin generated C-

terminal peptides of AP that included this penultimate lysine

residue (Sugiyama et al., 1988). Hortin et al. performed

additional inhibition experiments using synthetic peptides

containing arginine residues in place of the lysines to study

the AP:plasmin interaction. Peptides containing either the

C-terminal or penultimate lysine residue substituted with

arginine resulted in a 9-fold and 5-fold reduction,

respectively, in the inhibition of AP:plasmin complex

formation compared to a peptide of the wild-type sequence

(Hortin et al., 1989). Two hypotheses offered by these

authors are that the internal lysines may be interacting with

the low-affinity LBSs of plasmin, or that residues including

the penultimate lysine contribute to a conformation of the

peptide which places the C-terminal lysine in a favorable

position for binding to plasmin.

Recently an endothelial cell membrane protein, annexin

II, was described that binds both tissue plasminogen

activator (tPA) and plasminogen at different sites (Hajjar,

1991 and 1994). A C-terminal lysine residue of the receptor

is thought to mediate the binding of plasminogen via the

kringle domains. Hajjar showed a significant reduction of

plasminogen binding to endothelial cells following CPB

treatment whereas tPA binding was unaffected by this

treatment (Hajjar, 1993).

In examining plasminogen binding to monocytoid U937

cells, Miles et al. used a series of synthesized peptides to

inhibit glu-plasminogen from binding to the cell surface

(Miles et al., 1991). The peptides contained lysyl residues

at various positions in order to assess the importance of

lysines in the binding interaction. They reported that

peptides containing a C-terminal lysine residue were

essentially as effective in the inhibition of plasminogen

binding to the cell surface as were peptides containing both

a C-terminal and an internal lysine. Peptides containing

only an internal lysine would not inhibit plasminogen

binding. Interestingly, peptides consisting of at least the

terminal nineteen amino acids of alpha-2-antiplasmin, which

includes its the penultimate lysine, were 4.5-fold more

effective at competing plasminogen off the cell surface than

other peptides which also contained a C-terminal lysine.

These data led the authors to speculate that perhaps

structural features of proteins or peptides were important

for plasminogen binding in addition to a C-terminal lysine

residue. To further emphasize the requirement of a C-

terminal lysine for surface plasminogen capture, U937 cells

were treated with CPB, and this treatment resulted in a 67%

loss of plasminogen binding relative to untreated cells.

Treatment of the purified receptor molecule, the glycolytic

enzyme alpha-enolase, with CPB resulted in a decrease in

plasminogen binding activity using a ligand blot assay

similar to the assay used in this study to characterize

plasmin binding to Plr. Additionally, a synthetic peptide of

the C-terminus of alpha-enolase (containing a C-terminal

lysine) was able to inhibit plasminogen binding to untreated

alpha-enolase (Miles et al., 1991).

Internal lysine residues have also been implicated in

mediating plasminogen binding directly in lieu of C-terminal

lysines. In fibrinolysis, tPA must be bound to fibrin to

efficiently activate native glu-plasminogen, which is also

fibrin bound. Glu-plasminogen is thought to be bound via a

low affinity LBS interacting with internal lysine residues of

intact fibrin in this tri-molecular complex (Nieuwenhuizen et

al., 1983). Consistent with this low affinity alternative

binding site, Pannell et al. demonstrated that removing C-

terminal lysyl residues of intact fibrin by CPB had no effect

on the ability of tPA to activate fibrin-bound plasminogen.

Removal of C-terminal lysines was confirmed by amino acid

analysis of treated and untreated supernatants (Pannell et

al., 1988).

The ligand blot assay used in these studies to assess

plasmin(ogen) binding revealed that Plr binds plasmin with a

greater avidity than glu-plasminogen. In contrast to

streptokinase, Plr did not bind glu-plasminogen. This lack

of binding suggested that the penultimate lysine of Plr may

not interact with the low affinity LBS's of plasmin(ogen) as

do penultimate lysine residues in proteins such as fibrin.

It is possible that the penultimate lysine of Plr does play a

minor role in plasmin binding but that the ligand blot assay

is not sensitive enough to reflect such potential differences

in binding affinity. However, the C-terminal lysine appears

to be a major determinant.

Many group A streptococcal isolates, including strain

64/14, specifically bind plasmin or lys-plasminogen to a

greater extent than for glu-plasminogen. Although there are

streptococcal strains which bind both plasmin and glu-

plasminogen, albeit with different affinities, these strains

contain certain M protein serotypes which are reported to

modulate the lower affinity binding to glu-plasminogen

(Kuusela et al., 1992b). Furthermore, only lys-plasmin(ogen)

and, to a lesser extent, isolated heavy chain of lys-plasmin

containing the kringle regions, could efficiently compete

bound plasmin from the cell surface of strain 64/14 which

does not express a plasminogen binding M protein (Broder et

al., 1991, M. D. P. Boyle, personal communication).

Specificity of ligand binding to the whole bacteria is

further demonstrated by the inefficient binding of proteins

which contain kringle regions homologous to those of

plasmin(ogen), such as urokinase or tPA (DesJardin et al.,


The reduced plasmin binding of NH2-terminal deletion

mutants of Plr, intact MBP-fusions with Plr as well as the

resulting Factor Xa cleavage products indicated that the

presence of a C-terminal lysine on a protein is not

sufficient for plasmin binding activity. Therefore, in

addition to a C-terminal lysine, NH2-terminal residues may be

required for wild-type levels of binding either by direct

interaction or aiding in the optimal presentation of the C-

terminal lysine.

There are plasmin(ogen) binding proteins which are

reported to mediate binding through residues other than

lysines. As noted above, Berge and Sjobring identified a 43-

kDa plasmin(ogen) binding protein from M-protein type 53

group A streptococci (Berge and Sjobring, 1993). The

molecule has homology to the M-like family of proteins which

include IgG receptor proteins. These proteins have a

conserved signal sequence for targeting to the cell membrane

and a cell wall spanning region. M-like proteins are

anchored to the cell membrane via a domain near the C-

terminus thereby precluding the C-terminus in a plasmin

binding role. However, plasminogen bound to the cell surface

of these particular streptococci can be eluted by EACA, and

the region binding the plasminogen molecule to the cell

surface harbors the high-affinity LBS. The in vitro analysis

of the putative 43-kDa receptor was performed on acid

extracted material from the bacteria, and therefore the

receptor may be a cleavage product of the native molecule.

Both the recombinant and the endogenous 43-kDa protein were

treated with CPB prior to analysis of plasminogen binding,

however the removal of any potential C-terminal lysines was

not experimentally confirmed. Although the authors concluded

that binding is not mediated via a C-terminal lysine, they

have not yet demonstrated that inactivation of the gene

encoding this protein results in a reduction of plasminogen

binding to the bacteria.

Additional support for the contribution of a NH2-

terminal regions) of Plr to plasmin binding was demonstrated

by the reduced binding activity of Plr-49, a non-fusion 33

amino acid NH2-terminal deletion of Plr. Furthermore, when

this same NH2-terminal region was fused to the C-terminal 60

amino acids of Plr, plasmin binding was restored.

To further delineate the contribution of the NH2-

terminus of Plr in plasmin binding, several additional

mutations were generated. The mutant Plr-62, lacking the

five NH2-terminal amino acids, had substantial loss of

plasmin binding activity when a purified soluble preparation

of this protein was assayed. In recombinant Plr this region

consists of (M} V V K V. The B. stearothermophilus GAPDH

does not bind plasmin, and the C-terminus of this protein

does not contain a terminal lysine; instead a lysine is

followed by a glycine and a terminal leucine. This GAPDH

does, however, have an identical NH2-terminus to that of Plr

indicating that this sequence of residues is not sufficient

for directly binding plasmin. Contributions from the NH2-

terminal methionine are unlikely since it is present in only

50% of the recombinant Plr and is not present on mature Plr

isolated from group A strain 64/14 (Chapter 1).

Additionally, Plr-42, which contains the lysine4 substituted

with glycine4, bound wild-type levels of plasmin, thereby

ruling out participation of this residue in the binding

interaction. In addition to a C-terminal lysine, this would

leave the three remaining valine residues of the NH2-terminus

as candidates for the necessary amino acids required for

wild-type levels of binding.

The crude, insoluble preparations of Plr-62 and Plr-61,

which is a thirteen amino acid NH2-terminal truncation, both

demonstrated equivalent amounts of plasmin binding relative

to wild-type Plr. These results were in contrast to the lack

of plasmin binding demonstrated by soluble Plr-62. However,

the C-terminal deletion constructs containing wild-type NH2-

termini are deficient in binding regardless of whether the

protein preparation has been solublized or not. Plr taken

through the solublization procedure retains plasmin binding

activity, although the protein does lose GAPDH enzymatic

activity. Furthermore, the insoluble mutant Plr-46 retains

its plasmin binding activity after solublization. These

results indicate that this procedure does not have any direct

effects on plasmin binding ability. Even though proteins are

denatured with SDS during SDS-PAGE, the SDS is removed

following electrophoresis and limited refolding of the

protein may occur during transfer of proteins from the

polyacrylamide gels to nitrocellulose membranes. It is

possible that NH2-terminal residues are required for

appropriate presentation of the C-terminal lysine for

effective plasmin binding. When the NH2-terminal region was

removed, this positioning effect may have been lost. However

in the crude preparations of insoluble NH2-terminal Plr

mutants, some of the molecules may have retained appropriate

presentation of the lysine due to altered conformation of the

protein in inclusion bodies. Appropriate presentation of the

C-terminal lysine may be lost when these proteins are exposed

to guanidine prior to SDS-PAGE, thus accounting for the

differences in binding activity between the soluble and

insoluble preparations of Plr-62.

A C-terminal lysine residue is necessary but not

sufficient for the plasmin binding phenotype of Plr. The

data suggest that additional residues are required to

position the lysine for optimal accessibility to plasmin.

The soluble Plr mutants Plr-37 and Plr-45 which have

reduced plasmin binding activity retained GAPDH enzymatic

activity. Therefore, the genes encoding these proteins are

candidates for replacement of wild-type plr in group A strain

64/14. Alternatively, an additional candidate is the gene


encoding the non-plasmin binding GAPDH from Bacillus

stearothermophilus. Successful introduction of these

alternate genes at the plr locus would enable the

contribution of Plr to the plasmin binding phenotype of

strain 64/14 to be assessed.


Group A streptococci are highly invasive organisms

(Stevens, 1992). The mechanisms) utilized by the bacteria

to penetrate through tissue barriers has not yet been

elucidated. It has been hypothesized that proteolytically

active plasmin bound to the bacterial cell surface plays a

role in this degradative process (Lottenberg et al., 1987).

A useful approach to test this hypothesis is to compare the

pathogenic potential of wild-type streptococci to mutant

streptococci lacking the ability to bind plasmin.

Ideally, in studying putative virulence factors of

bacteria, one would like to generate an isogenic strain in

which the expression of only the factor of interest is

eliminated. This strain can then be compared to the wild-

type strain for virulence in an animal model. The initial

goal of these studies was to insertionally inactivate plr in

strain 64/14 in order to evaluate the role of the putative

plasmin receptor, Plr, in the plasmin binding phenotype of

group A streptococci. These isogenic strains could then be

tested in a mouse model where organisms are inoculated into a

subcutaneous air bleb (Raeder and Boyle, 1993). However,

attempts at insertional inactivation of plr were

unsuccessful. The extensive homology of the predicted amino

acid sequence of Plr with GAPDHs suggested that this may be a

glycolytic enzyme (Lottenberg et al., 1992) and could

therefore potentially be an essential gene of group A

streptococci. Characterization of Plr revealed that it is a

functional GAPDH enzyme. Furthermore, group A streptococci

have only a single copy of plr. The failure of strain 64/14

to grow using carbon sources other than glucose was

consistent with the inability to insertionally inactivate plr

and suggested that plr was in fact an essential gene in group

A streptococci.

Two alternative approaches were applied to generate

isogenic derivatives of strain 64/14. One approach was a

gene replacement strategy whereby the gap gene from Bacillus

stearothermophilus, which encodes a GAPDH which does not bind

plasmin, would replace plr on the streptococcal chromosome.

In addition, in vitro analysis of mutant Plr proteins, as

presented in Chapter 2, was performed to identify domains or

residue(s) of Plr which were important for binding plasmin

and to determine if these domains were distinct from those

required for glycolytic function. These studies revealed

that a C-terminal lysine of Plr was required for wild-type

levels of plasmin binding, and some of these Plr mutations

retained GAPDH activity. In the second approach presented in

this Chapter, strategies were applied to successfully

introduce these mutations at the plr locus in strain 64/14.

The strains expressing mutated Plr molecules were compared to

wild-type strain 64/14 in vitro for the ability to capture

plasmin on the bacterial surface. Additional experiments are

presented herein which further examined plasmin binding

components of strain 64/14 and the cellular localization of


Materials and Methods

DNA manipulations and plasmids. DNA manipulations used

in construction of plasmids was described in the Materials

and Methods of Chapter 2. DNA hybridization studies were

performed as outlined in the Materials and Methods of Chapter

1. The plasmids used in these studies are summarized in

Table 3-1, and the construction of these plasmids is

described in detail below.

Construction of integration plasmids. Cointegrative

plasmids were constructed using the plasmids pRL024, which

harbors wild-type plr, and pRL037, a derivative of pRL024

containing a 3' end deletion of plr generated by exonuclease

III. The plasmids pRL024 and pRL037 were subjected to

identical DNA manipulations as shown schematically in figure

3-1. The first 660 bp of plr and all of the strain 64/14

upstream sequences were removed by digesting the plasmids

with BstEII and HindIII. The DNA overhangs were blunt ended

using Klenow fragment and dNTPs, and the linear DNA was

ligated. These plasmids were linearized by digesting at a

41 41)

< a

n a a)
W 4140 r4

3 c
4 01 ()

W 0.
C (N 1)
0.)4 43 >*
40. ro

44H (d

0 70 C

4-4 ,

'0 44 U)
44 It 1) _c
NC -

4- 0 0
ca a 1

0 J
H) e

(0 0J
wc m

Q) -H(
4M 0)

41) 4-

H) 4

a) co
r0 0

Ha) C

3 1
0 L( *H


C m
m -, )

4-4 H
4 a)r


H~ (4)- 0r




Q) 0 o 4J

a o C
(00 00

( w
( 1 o


) ,-r4 4
~0.. 4J

a .
4 c4
) a) 0

0 0
cD > a)

4 a
i4 I- 3


J3 4-44



Q 4

X: (N

) 4, 3 (
j00 0)

10 4 (414)

(N40 (

4 o H4)1)

u a 1 Hc
o-. 0
0) HU43
4: a)

0 41) 4J

4J 04-J M1)

> am Hi 4
) ) (4I 4H

u: 4 o E)

04 3.44

0 JCne

(0 0



4: 4
:4 -

S0 4
0J 41)

w > 41 )
5 0 WE^i

S4 -4

*0 .4:

4) SC: 3

(N arr
4,) I~I r


o< a
so c

3 4-1

a)) 04-

an 1-


4) 4) -.
4ro- (0


.4 41 4!


o 4:r

4J 4)


( 41
0 -H
u4 0
Sm H

c) (a


4o 2

4-1 CO
(0 0 4J

04 0

,c a

o o
0'0 0

s c

E- q

.0 0

0 o

0 4a)

l- '-
c ^
-H 0

3 >o
E-' I

u u



r_ C

o H 'H

(4 ( 0 4J

cn I4t 04-H 0

0) : )
(00 -3HI0

0U 5 4 o

M 2 ) ( 4JU)
C S 4J -)

(u -1 c a)

.) Ve-H
-H Q)'
a) H'H 0 m

c3 0 4 C

,C-H V' I
3 *H -r a-

*H o0 0o
E Ua) 0 ()H WO o

s4 44e- 0

( 0 4) t

a. s a o
C-' o (04 I


~CYC1U 1 rr~ll~

irl ~I
c~ ~c~
p~oac pRLO1T


prmrU pSWOI1

SIpi EaIKmtcuasU rino-pkrcodon
utramin 614 DNA "quenne -pACYC184 DNA sequences

Figure 3-1. Construction of coincegrative plasmids for plr
mutagenesis of strain 64/14. The plasmids pRL024 and pRL037
described in Chapters 1 and 2, respectively, were subjected
to identical manipulations. Plasmids were digested with
HindIII and BstEII to remove the 5' end one third of the plr
ORF and upstream streptococcal sequences. The Q cassette,
encoding a kanamycin resistance gene was ligated into the
unique Sal I site of these plasmids to yield pSW024 and
pSW037 containing wild-type 3' end plr and a 3' end mutation
of plr, respectively. See text for further details.
Abbreviations: H, HindIII; 3, Sail; Bs, BstEII; E, EcoRI;
B, BamHI; P, putative promoter elements; PBS, ribosomal
binding site; Kme, kanramyci resistance gene; cam,
chloramphenicol resisLtnce g-en; ret, tetracycline resistance
gene. The solid bo:< of plr-37 rpr s Ints predicted vector
encoded sequences.