LiBRARY
Michigan State
University

 

 

PLACE m RETURN Box to remove this checkout troni your record.
TO AVOID FINES Mum on or before data due.

[ DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

STRUCTURE-FUNCTION RELATIONSHIPS
IN RAT BRAIN HEXOKINASE:
A STUDY USING MONOCLONAL ANTIBODIES
By

Allen D. Smith

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Biochemistry

1991

k..—' 9/( 1

ABSTRACT
STRUCTURE-FUNCTION RELATIONSHIPS
IN RAT BRAIN HEXOKINASE:

A STUDY USING MONOCLONAL ANTIBODIES
BY
Allen D. Smith

Based on a combination of direct and competitive epitope mapping methods, 23
monoclonal antibodies against rat brain hexokinase could be divided into 9 distinct classes,
each recognizing epitopes located within a specific region of the molecular surface. Six of the
epitopic regions were located on the previously proposed structure for the enzyme.

Seven monoclonal antibodies were mapped to defined regions in the N-terminal half of
the enzyme. Based on the effect of these monoclonal antibodies on binding of hexokinase to
mitochondria, and their ability to recognize the mitochondrially bound enzyme, the Spatial
relationship between bound enzyme and the mitochondrial surface was deduced.

Several monoclonal antibodies of defined epitopic location were used to examine the
effects of ligands on the tryptic digestion pattern of hexokinase. Four tryptic sites, located in
the C-terminal domain of hexokinase, were identified that were markedly influenced by
binding of ligands. Results from these experiments suggest that binding of both the substrate
glucose to the catalytic site and the inhibitor glucose 6-phosphate to the allosteric regulatory
Site, induce conformational changes in the catalytic C-terminal domain of hexokinase similar
to the closing of the catalytic cleft observed when glucose binds to the homologous hexokinase
from yeast. Inhibitory hexose 6—phosphates have been shown to prevent binding of the
substrate ATP, and evidence presented here indicates that this results from reduced
accessibility of ATP to the catalytic site as a result of cleft closure.

The effect of ligand-induced conformational changes on the immunoreactivity of

monoclonal antibodies identified five regions in hexokinase that were affected by binding of

ligands. Antibodies binding to regions F and G, located on the N- and C-terminal domains of
hexokinase, respectively, exhibited reduced binding in the presence of inhibitory hexose 6-
phosphates and markedly increased the K; for the inhibitor 1,5-anhydroglucitol 6—phosphate,
thus, identifying these regions as being important for the regulation of enzyme activity. Two
antibodies binding to region H in the C-terminal domain virtually abolished enzyme activity.
Binding of these antibodies to hexokinase was significantly reduced in the presence of glucose
indicating that they recognize the "open" conformation of the enzyme and inhibit enzyme
activity by preventing the conformational changes necessary for catalysis. A similar effect on
immunoreactivity was observed with inhibitory hexose 6-phosphates, suggesting that these
ligands also induced the "closed" conformation of the enzyme, thus, providing additional
evidence that inhibitory hexose 6—phosphates may reduce access of the substrate ATP to the

catalytic Site.

Chapter 3 has been reprinted with permission from Archives of Biochemistry and
Biophysics, copyright Academic Press, Inc..

Smith, AD. and Wilson, LE. , Disposition of Mitochondrially Bound Hexokinase at the
membrane Surface, Deduced from Reactivity with Monoclonal Antibodies Recognizing
Epitopes of Defined Location, Arch. Biochem. Biophys., 287, 359-366, (1991).

Chapters 4 and 5 have been submitted for publication to Archives of Biochemistry
and Biophysics, Academic Press, Inc..

Smith, A.D., and Wilson, J .E., Effect of Ligand Binding on the Tryptic Digestion
Pattern of Rat Brain Hexokinase: Relationship of Ligand-Induced Conformational
Changes to Catalytic and Regulatory Functions, (1991).

Smith A.D. , and Wilson, LE. , Epitopic Regions Recognized by Monoclonal Antibodies
against Rat Brain Hexokinase: Association with Catalytic and Regulatory Function,
(1991).

TO MY PARENTS

ACKNOWLEDGEMENTS

First and foremost, I would like to thank my parents for all of their love and
support over the years. Without their help, none of this would have been possible.
Secondly, I would like to thank my major professor, John Wilson, for all of his
intellectual and financial support through the years. John provided an environment
that promoted independent drinking and research, but at the same time was always
available for advice when needed. I doubt if I will be fortunate enough to ever work
with a person of his caliber again. There are many friends and colleagues which also
deserve recognition, but space does not allow for everyone to be mentioned.
However, I would like to acknowledge several people who have been supportive
through the years. Steve and Karen Wresinski have been and will continue to be the
closest friends I have. Their support has helped me get through both professional and
personal hard times. Kari Tomashik is another very special friend who has always
had an open ear and heart; I love her dearly for all She has done for me. Steve Pittler
is another life long friend who has provided support in all aspects of my life.
Although we are miles apart, we both know that the next time we get together it will
seem like we were never apart; that is how it is with best friends. I feel fortunate in
that my list of special friends has not yet ended. One additional person, Mike Roach,
deserves mention as well. Mike has provided me with many hours of stimulating
intellectual discusssions as well as introducing me to a plethora of alternative music.
All of the people mentioned above have demonstrated to me the importance of love
and friendship in one’s life and I thank them all. Two lab mates, Annette Thelen and
Hector BeltrandelRio, were both colleagues and friends and I will miss them both. I
would also like to acknowledge Doug Wiesner. Although we have not known each
other all that long we have established a friendship that will surly last into the future.
Last, but by no means least, I would like to acknowledge Beta Borer. She has been
very supportive during the trying times I have experienced in finishing my Ph’D. She
helped me hang on to what little sanity I had left. Without her cooking and loving
care it would have been infinitely harder for me to have accomplished the task. She
holds a special place in my heart and words cannot convey the depth of my feelings
for her.

vi

TABLE OF CONTENTS

LIST OF TABLES ....................................... x
LIST OF FIGURES ....................................... xi
LIST OF ABBREVIATIONS ................................ xiii
CHAPTER 1
Literature Review ....................................... 1
Introduction ...................................... 2
Rat Brain Hexokinase ................................ 3
Mammalian Hexokinases .............................. 10
Yeast and Starfish Hexokinase ........................... 13
The Use of Mabs in the Study of Protein Structure .............. 14
REFERENCES .................................... 20
CHAPTER 2
Generation of Monoclonal Antibodies ........................... 26
Introduction ...................................... 27
MATERIALS AND METHODS ......................... 29
Materials .................................... 29
Affinity purification of rabbit anti-hexokinase
antibodies ............................... 30
Purification of rat brain hexokinase and measurement
of enzyme activity .......................... 31
Immunization protocols ............................ 31
In Vitro immunization ............................ 34
Production of hybridoma cell lines ..................... 34
Solution phase immunoassay ......................... 36
ELISAS ..................................... 37
Assessment of murine anti-hexokinase immune response ........ 39
RESULTS ....................................... 40
Comparison of adjuvant systems ...................... 40

Development of a modified ELISA that selectively
detects monoclonal antibodies recognizing

native antigen ............................. 44

Generation of anti-hexokinase Mabs .................... 48
DISCUSSION ..................................... 56
REFERENCES .................................... 60

vii

CHAPTER 3
Disposition of Mitochondrially-Bound Hexokinase at the
Membrane Surface, Deduced from Reactivity with Monoclonal Antibodies

Recognizing Epitopes of Defined Location ........................ 62
Introduction ...................................... 63
MATERIALS AND METHODS ......................... 65

Materials ................................... 65
Hexokinase .................................. 65
Monoclonal Antibodies ........................... 66
Mitochondria ................................. 67
Effect of Monoclonal Antibodies on Binding of Rat Brain
Hexokinase to Mitochondria ................... 67
Preparation of Nonbindable Hexokinase ................. 68
Effect of Modification of Sulthydryl Residues on
Immunoreactivity of Rat Brain Hexokinase .......... 68
Recognition of Mitochondrially-bound Hexokinase by
Monoclonal Antibodies ...................... 69
Computer Modelling ............................ 70
Electrophoretic and Immunoblotting Procedures ............ 70
Epitope Mapping ............................... 70
RESULTS ....................................... 75
Immunoprecipitation of Truncated Versions of Brain Hexokinase . 75
Further Definition of the Epitopes Recognized by These Mabs . . . 79
Effect of Mabs on Binding of Hexokinase to Mitochondria ..... 87
Recognition of Mitochondrially-Bound Hexokinase by Mabs . . . . 90
DISCUSSION ..................................... 93
REFERENCES .................................... 98
CHAPTER 4

Effect of Ligand Binding on the Tryptic Digestion Pattem of
Rat Brain Hexokinase: Relationship of Ligand-Induced

Conformational Changes to Catalytic and Regulatory Functions ........... 102
Introduction. ...................................... 103
MATERIALS AND METHODS ......................... 105

Materials ................................... 105

Hexokinase activity and protein determinations ............ 105

Proteolytic digestion of hexokinase in the presence of ligands . . . 105

RESULTS ....................................... 107
Ligands Affect the Tryptic Cleavage Pattern of Rat Brain

Hexokinase .............................. 107

Analysis of Tryptic Cleavage Patterns .................. 114

Effect of Hexoses on Tryptic Cleavage ................. 119

Effect of Hexose 6-Phosphates on Tryptic Cleavage ......... 119

viii

Generation of Discrete N- and C-Terminal Halves by Cleavage

at T3 .................................. 123

Effects of Hexose 6-Phosphates on Proteolysis of the C—
Tenninal Domain .......................... 128
Effect of NTPS and Their Mg+ + Chelates on Tryptic Cleavage . . 128
DISCUSSION ..................................... 134
REFERENCES .................................... 140

CHAPTER 5
Epitopic Regions Recognized by Monoclonal Antibodies
against Rat Brain Hexokinase:

Association with Catalytic and Regulatory Function .................. 143
Introduction ...................................... 144
MATERIALS AND METHODS ......................... 147

Materials ................................... 147
Monoclonal Antibodies ........................... 148
Hexokinase Activity and Inhibition Studies ............... 149
Protein Determinations ........................... 149
Immunoprecipitation of Hexokinase ................... 150

Preparation of isolated N- and C-terminal halves of hexokinase . . 150
In vitro synthesis of a truncated version of rat brain hexokinase . . 151

Epitope mapping using a competitive ELISA .............. 151
Effect of ligands on the interaction of Mabs with hexokinase,
determined by ELISA ....................... 152
Effect of Mabs on the tryptic cleavage pattern of hexokinase . . . . 154
Molecular modeling ............................. 154
RESULTS ....................................... 155
Epitope Mapping ............................... 155
Ligands affect the interaction of Mabs with hexokinase ....... 17 8
Ligands affect the interaction of Mabs with the isolated N-and
C-terminal domains of hexokinase ................ 182
Effect of Mabs on the kinetic parameters of hexokinase ....... 187
DISCUSSION ..................................... 190
REFERENCES .................................... 196
CHAPTER 6
Summary and Perspectives .................................. 200
SUMMARY AND PERSPECTIVES ....................... 201
REFERENCES .................................... 206

LIST OF TABLES

CHAPTER 1
I. Comparison of the various kinetic constants

for the mammalian hexokinase isozymes ..................... 12
CHAPTER 2
1. Summary of methods used to prepare microtiter

plates for ELISA ................................... 38
II. Comparison of hybridoma cell lines secreting anti-hexokinase

Mabs resulting from different immunization protocols ............ 41
III. Comparison of anti-hexokinase antibody titers ................. 45
CHAPTER 3
1. Location of Epitopes within the Sequence

of Rat Brain Hexokinase .............................. 78
CHAPTER 5
1. Summary of competitive ELISA epitope mapping results ........... 159
II. Effect of ligands on the interaction of Mabs with intact .

hexokinase ....................................... 179
II. (Continued) ...................................... 180
III. Comparison of the effect of ligands on the interaction of Mabs

with the isolated N- or C-tenninal domain or intact hexokinase ....... 183
III. (Continue) ..................................... 184
IV . Effect of Mabs on the kinetic parameters of intact brain

hexokinase and the isolated C-terminal catalytic domain ........... 188

LIST OF FIGURES

CHAPTER 1
1. Model of rat brain hexokinase ........................... 4
2. Proposed evolutionary relationship between the hexokinases ......... 15
CHAPTER 2
l. The effect of adjuvant on anti-hexokinase immune

response in mice ................................... 42
2. Analysis of the anti-hexokinase immune response of mice

immunized with the isolated C-terminal domain of hexokinase ....... 53
CHAPTER 3
1. Map of Plasmid HKpT7-l ............................. 72
2. Immunoprecipitation of truncated versions of hexokinase ........... 76
3. Reactivity of monoclonal antibodies with native and

chymotrypsin—treated hexokinase .......................... 80
4. Comparison of amino acid sequence of N- and C-terminal

halves of rat brain hexokinase ........................... 83
5. Effect of modification of sulfhydryl residues on irnmunoreactivity ..... 85
6. Effect of Mabs on binding of hexokinase to mitochondria .......... 88
7. Reactivity of Mabs with mitochondrially-bound hexokinase .......... 91
8. Schematic representation of the orientation of

hexokinase bound at the mitochondrial surface .................. 94
CHAPTER 4
1. Schematic representation of the tryptic cleavage pattern of

rat brain hexokinase ................................. 108
2. Effect of hexoses on tryptic digestion of hexokinase .............. 110
3. Effect of hexose-6—phosphates, NTPS and NTP-Mg2+ on

tryptic digestion of hexokinase ........................... 112
4. Immunoblotting analysis of tryptic fragments .................. 115

5. Effect of hexose 6-phosphates and analogs on the tryptic

digestion pattern of brain hexokinase ....................... 120
6. Detection of enzymatically-active tryptic cleavage products ......... 124
7. Time course for digestion of brain hexokinase in the presence

of hexose 6-phosphates ............................... 129
8. Effect of nucleoside triphosphates and their Mg+ + chelates

on the tryptic digestion pattern of brain hexokinase .............. 131
CHAPTER 5
1. Competitive ELISA ................................. 156
2. Distribution of epitopic regions on the molecular surface of

rat brain hexokinase ................................. 161
3. Reactivity of Mabs with isolated N- and C-terminal domains of

rat brain hexokinase ................................. 165
4. Venn diagram, depicting relationship between epitopes in regions

D and E, deduced from results of competitive ELISA ............. 167
5 . Effect of Mabs on the tryptic digestion pattern of hexokinase ........ 171
6. Immunoprecipitation of intact hexokinase and the isolated

C-terminal domain by Mabs ............................ 173
7. Immunoprecipitation of a truncated version of hexokinase by Mabs . 176

BCA

BSA
2deoxyGlc
2-deoxyGlc-6-P
DMEM

FBS

ELISA

Fm

Fru-6—P

Gal

Gal-6-P

Glc

Glc-6-P
GlcN
GlcN-6-P
GlcNac
Hepes

IgG

13M

Mab

Man
Man-6-P
NHS-LC-biotin
NTP

PMSF

SDS

STI

5-thioGlc
5-thioGlc-6-P
‘I'PCK

LIST OF ABBREVIATIONS

bicinchoninic acid

bovine serum albumin

2-deoxyglucose

2-deoxyglucose 6—phosphate

Dulbecco’s modified eagle medium

fetal bovine serum

enzyme—linked irnmunosorbant assay
fructose

fructose 6—phosphate

galactose

galactose 6-phosphate

glucose

glucose 6—phosphate

glucosamine

glucosamine 6-phosphate
N-acetylglucosamine
4-(2-hydroxyethyl)-1-piperasineethane-sulfonic acid
immunoglobulin G

immunogloban M

monoclonal antibody

mannose

mannose 6-phosphate
sulfosuccinimidly-6—(biotinamido) hexanoate
nucleoside triphosphate
phenylmethylsulfonyl fluoride

sodium dodecyl sulfate

soybean trypsin inhibitor

5 -thiog1ucose

5-thioglucose 6-phosphate
L-l-toslyamide-Z-phenylethyl chloromethyl ketone

xiii

CHAPTER 1

Literature Review

Introduction

Rat brain hexokinase catalyzes the initial step in glycolysis and plays a crucial
role in governing glucose metabolism in the brain. Understanding how this enzyme
functions and its role in regulation of glucose metabolism has been a major goal of
this laboratory. Until recently, very little was known about the structure of
hexoh'nase and its relationship to the function of this enzyme. To this end, several
different projects have been attempted, including cloning of the cDNA for the
enzyme. Another approach, and the major goal of the present work, was to generate
a library of monoclonal antibodies (Mabs) against rat brain hexokinase to use as
Probes in establishing structure-function relationships within hexokinase. The focus
Of this chapter is to provide background information on hexoldnase structure and
function at the onset of this project and other relevant developments since that time.
In addition a brief review of how MabS have been used in structure-function studies

of other proteins will also be presented.

W
Rat brain hexokinase (ATP: D-hexose 6—phosphotransferase, EC 2.7.1.1.),
consists of a single polypeptide chain with molecular weight approximately 100 kDa
(1). The complete amino acid sequence of the rat enzyme has been deduced from the
cloned cDNA (2). Extensive similarity in sequence of the N- and C-tenninal halves
of the enzyme, and between rat and yeast hexokinase (2), supports the view that the
mammalian enzyme is the evolutionary product of duplication and fusion of a gene
coding for an ancestral 50 kDa hexokinase related to the present-day yeast
hexokinase. The sequence similarity of mammalian and yeast hexokinase has been
presumed to indicate a corresponding conservation of secondary and tertiary structural
features, and the proposed structure for the mammalian enzyme (2) Shown in Fig.1 iS
based on the x-ray structure of yeast hexokinase (3-5).
Hexokinase catalyzes the conversion of glucose to glucose 6—phosphate using

ATP as phosphoryl donor. One of the products of the reaction, glucose 6—phosphate,
is a potent inhibitor of enzyme activity (1). The other product, ADP, also inhibits
hexokinase activity (6) but most of the cellular ADP is thought to be
illtmmitochondrial (7) and therefore ADP probably does not have a physiologically
Significant role in regulation of extramitochondrial enzymes such as hexokinase. The
nonchelated form of ATP, a competitive inhibitor of the magnesium chelate of ATP
(ATP-Mg“) (8), is found at Significant intracellular levels, and may inhibit
hexokinase activity in vivo (1). The cellular concentrations of other nucleotides are

well below that of ATP and are not thought to affect hexokinase activity (1).

Figure 1. Model of rat brain hexokinase. The model is nearly identical to the one
proposed by Schwab and Wilson (2), generated using the software package Mosaic,
courtesy of the UpJohn Co. . The model shown here was constructed using the
software package Insight II on a Silicon Graphics Personal his computer. In this

orientation, the N-terminal domain of hexokinase is on the left.

 

Two other ligands, glucose 1,6—bisphosphate and inorganic phosphate (Pi) also
play important roles in regulation of hexokinase activity. Glucose 1,6-bisphosphate
inhibits hexokinase, and like Glc-6-P, competes with ATP for binding to enzyme.
The reported Ki value for glucose 1,6—bisphosphate is 100 ”M (9) and well within the
range of concentration estimated to exist in brain tissue (10,11) indicating this
compound may play a role in regulation of hexokinase activity. Both Glc-6—P and
glucose 1,6-bisphosphate inhibition are antagonized by Pi (1). Antagonism suggests
mutually exclusive binding of these ligands to a common site on the N -tenninal half
of hexokinase (12). Inorganic phosphate itself has no effect on hexokinase activity
except at high, physiologically-irrelevant concentrations.

The nature of the Glc-6-P binding Site was a matter of controversy for a
number of years. Crane and Sols (13), proposed that a allosteric regulatory site for
Glc-6-P existed, based on the dissimilar specificity of the binding sites for hexoses
(14) and hexose 6-phosphates (13). Fromm and colleagues (15-17) argued that
inhibition by Glc-6-P was not allosteric and but resulted from binding of Glc-6-P at a
site that overlaps with the binding site for ATP. This results in direct competition
between the binding of the phosphate moiety of Glc-6—P and the y-phosphate of ATP
and places the binding site for Glc-6-P on the catalytic domain of hexokinase. Lazo
et al. (18) presented evidence for the existence of two discrete Glc-6aP binding sites
on hexokinase, one high affinity and one low affinity site, but technical difficulties
complicated interpretation of these experiments. In combination, the results from the

studies by Crane and Sols (13,14) and Lazo et. a1. (18) argue against the one site

model proposed by Fromm.

More recent work provides clear evidence in support of a discrete (from the
catalytic Site) allosteric Glc-6-P binding site. White and Wilson (19), using selective
proteolysis techniques, demonstrate that a N-terminal, 52 kDa fragment of rat brain
hexokinase is preferentially protected from denaturation and proteolysis by Glc-6-P
while the remaining 48 kDa C—terminal fragment, shown to contain the binding Sites
for Glc and ATP (20,21), and catalytic activity, (22) is not protected. A Kd of 10
M for Glc-6-P was estimated for the binding site on the 52 kDa fragment (19), in
excellent agreement with previous estimates for the high affinity Glc-6-P allosteric site
(1,23). Moreover, the specificity of the N-terminal site was shown to correlate with
inhibitory effectiveness of various hexose-6-phosphates or analogs. These results
provide direct evidence for the existence of a discrete regulatory domain containing a
binding site for the effector, Glc-6—P, located in the N-terminal half of hexokinase,
and distinct from the catalytic site located in the C-terminal domain.

Additional work by Hutny and Wilson (24), provided evidence that binding of
Glc—6-P to the allosteric site leads to conformational changes in both the N- and C-
terminal halves of hexokinase. At low concentrations of Glc-6-P (< 10 I‘M)
sulfhydryls present in the 52 kDa N—terminal domain of hexokinase were significantly
protected against reaction with 2-bromoacetamido-4-nitmphenol, a sulfhydryl specific
reagent. Reactivity of sulfhydryls present in the C-terminal domain were partially
protected at low concentrations of Glc-6-P and at higher concentrations of Glc-6-P,

250-1000 uM, reactivity was further reduced, suggesting that two sites for binding of

Glc-6-P exist, a high affinity Site on the N-terminal domain and a low affinity site on
the C-terminal domain. In a similar manner, low concentrations of Pi protected
sulfhydryls present in the N-terrninal half of hexokinase from modification, but
substantially higher concentrations were needed to protect the sulfhydryls located in
the C-terminal half of the enzyme. Together, these results provide further evidence
that Glc-6-P and Pi bind at a high affinity Site located on the N-tenninal half of
hexokinase and that Glc-6-P induces conformational changes in the C-terminal portion
of the enzyme at physiologically relevant concentrations of the ligand.

Hexokinase activity is also regulated by its interaction with the outer
mitochondrial membrane (25 ,26). Hexokinase exists in an equilibrium between a
mitochondrially bound and a soluble form of the enzyme (25). The distribution of the
enzyme between the two forms is influenced by the ligands Glc-6-P and Pi. Glc-6-P
shifts the equilibrium towards the soluble State and Pi increases mitochondrial
association by antagonizing the negative influence of Glc-6-P on binding (1). The N-
terminal residues of hexokinase are critical for mitochondrial binding (27) and are
inserted into the mitochondrial lipid bilayer (28). Binding is also dependent on
divalent cations (29). A hexokinase binding protein has been isolated from the outer
mitochondrial membrane (30) that subsequently was Shown to be the pore forming
protein, porin (31,32). Thus, bound hexokinase is in excellent position to have
preferential access to ATP generated intramitochondrially by oxidative

phosphorylation. Recent work by BeltrandelRio and Wilson (33) provide evidence in

support of this.

At least some portion of bound hexokinase exists as a tetramer (34). Based on
the estimated Size of a porin pore (35) (2.5 nm inner diameter and 5.0 nm outer
diameter) and the estimated size of rat brain hexokinase (2), it would be possible to
orient four hexokinase molecules radially around each pore such that the catalytically
active C-terminal domain is oriented toward the lumen of the pore. Such an
orientation would place hexokinase in position to have preferential access to ATP-
Mg2+ resulting from oxidative phosphorylation and would also place the N-terminal
domain in position to interact with the lipid bilayer (34).

Rat brain hexokinase can exist in a number of different ligand-induced
conformational states. Initially, the various conformational states were defined by the
ability of ligands to protect hexokinase against the inactivating agents glutaraldehyde,
heat, chymotrypsin, and DTNB (23,36,37). In general, preferred sugar substrates
(Glc and Man) and strong inhibitors (Glc-6-P and 1,5 -anhydroglucitol-6-P) are quite
effective at protecting hexokinase against inactivation by all four reagents while poor
substrates and weak inhibitors do not provide protection. With the exception of ATP-
Mg2+ , nucleoside triphosphates and their Mg2+ salts, provide varying degrees of
protection against chymotrypsin and heat but not against glutaraldehyde and DTNB
(34). Although ATP-Mg2+ is more tightly bound to hexokinase than free ATP (23),
Mg2+ reverses the protection seen with free ATP and Mg2+ alone did not protect
hexokinase from inactivation by any of the reagents used (23,37).

These results were supported by additional data on the effects of ligand-

induced conformational changes on the reactivity of specific sulfhydryl residues (24).

10

Again Glc, Glc-6-P, nucleoside triphosphates and ADP induce conformational changes
leading to protection of some or all of the sulfhydryls analyzed. The Mg2+ salts of
CTP, GTP and UTP are less effective at protecting certain sulfhydryls but continued
to protect others. This is in stark contrast to the results seen with ATP-Mg2+ and
ADP-Mg2+ where reactivity of most sulfhydryls was increased to levels much higher
than that seen in the absence of any ligands. ATP-Mg2+ destabilizes hexokinase to
the denaturant guanidine hydrochloride (12). These combined results suggest that
ligands such as Glc and Glc-6-P induce a closed conformational state, which protects
the enzyme from inactivation, whereas ATP-Mg2+ and ADP-Mg2+ induce
conformational changes that " open up" the structure and allow inactivation to occur at

accelerated rates.

m ' H xo ' S

There are at least four distinct isozymes of hexokinase in mammalian tissue
(38,39). There may exist some microheterogeneity in Type I and II isozymes but it is
clear in most cases that this is not genetically based (1). The four isozymes are
named A, B, C, and D (38), or alternatively, I, II, III, and IV (40), according to
differences in their isoelectric points. Type I hexokinase is the only isozyme found in
brain, and is present at Significant levels in most tissues except muscle (1). While all
four isozymes are found in liver, the Type IV isozyme, also known as glucokinase,

predominates (40). In addition to being the predominant form of hexokinase in

11

muscle, the Type II isozyme is also the major form present in other insulin sensitive
tissues such as the diaphragm, mammary gland, and adipose tissue (1). In contrast to
the other three isozymes, the Type III isozyme has not been found to be the major
form in any tissue studied (1).

All mammalian hexokinases are composed of a single polypeptide chain. The
molecular weights of isozymes I-III are approximately 100,000 Da and the Type IV
isozyme is approximately 50,000 Da. Isozymes I-III are inhibited by Glc-6-P, but the
Type IV isozyme is not and in this respect the Type IV isozyme is Similar to yeast
hexokinase (discussed below). Based on these considerations it has been proposed
that isozymes I-III may have arisen through gene duplication and fusion of an
ancestral form of hexokinase of molecular weight 50,000 (41-46). Comparison of the
sequences of these enzymes provided substantial evidence for this evolutionary
relationship. All four isozymes Share significant amounts of sequence identity with
each other (2,47-49) and isozymes I-HI each Share significant internal homology
between their N- and C-terminal halves.

The four isozymes exhibit significantly different kinetic parameters as
summarized in Table I (41). In addition, inhibition of type I hexokinase by Glc-6-P is
antagonized by Pi while inhibition of type II hexokinase is not relieved by Pi (1). The
effect of Pi on inhibition of type III hexokinase is not known. Both types I and II
isozymes reversibly bind to mitochondria and this has also been reported for the type
three isozyme (50,50a), but the type III isozyme lacks the N -terminal sequence Shown

to be critical for binding (27) that is conserved between isozymes I and II, and

12

Table I. Comparison of the various kinetic constants for the mammalian

hexokinase isozymes“.

 

Hexokinase
7Parwmeter - I II III Iv“
xi Glc 0.04b 0.13 0.02 4.5
In are 0.42 0.07 1.29 0.49
x1 Glc - 6 - P
V“ ATP 0.026 0.021 0.074 15

 

a-Table adapted from Ureta (41) , and references therein.
b-All apparent kinetic constant values are expressed in mM.

13

therefore, it is not clear whether binding of type III hexokinase is specific and

reversible.

Yeast and Starfish Hexokinase

Two isozymes of hexokinase are present in yeast and are referred to as PI and
P11 based on their order of elution from DEAE-cellulose by a decreasing pH gradient
(42). Alternatively the isozymes are designated A and B, respectively (51). The
sequence of both isozymes is known (52,53), and the two isozymes share 378
identical residues out of 485 total. Like the type IV mammalian isozyme, yeast
hexokinase has a molecular weight of 50,000, is insensitive to inhibition by Glc-6-P
(42), and does not reversibly bind to mitochondria. In addition, yeast hexokinase has
extensive sequence similarity to all four mammalian isozymes (2,4749). The x-ray
crystallographic structure of both isozymes, determined by Steitz and colleagues (3-5) ,
Shows that yeast hexokinase is composed of two "lobes" , with a deep cleft formed
between them. Glucose binds at the base of the cleft (3), and induces a
conformational change in the enzyme that results in cleft closure (3). More detailed
analysis of the conformational change associated with cleft closure indicates that a
majority of the structural changes occur in the small lobe and the hinge region
between the two lobes (3).

Also of interest is a hexokinase isolated from starfish. Like yeast hexokinase
and glucokinase, it has a molecular weight of approximately 50,000 Da but its amino

acid sequence is not known. Kinetically it resembles type I mammalian hexokinase

14

more closely than the other isozymes (54). The KTn for Glc and ATP are nearly
identical to the type I isozyme and both enzymes are strongly inhibited by Glc-6-P.
Thus starfish hexokinase may represent the modern version of an ancestral hexokinase
gene that coded for a 50,000 Da enzyme that had evolved a Glc-6-P regulatory site,
which subsequently underwent gene duplication and fusion to ultimately lead to the
family of modern day 100 kDa hexokinases. The sequence similarity between the
mammalian and yeast isozymes of hexokinase and the kinetic similarities between
starfish and type I mammalian hexokinase has led to the proposed evolutionary

scheme for the hexokinase family as shown in Figure 2 (22).

The Use of Mabs in the Study of Protein Stmcture.

The development of somatic hybridization techniques that allow isolation of
hybridoma cell lines producing monoclonal antibodies (55) has led to extensive use of
such antibodies in characterizing proteins at both the structural and functional levels.
Due to their specificity (usually for one site on the molecule), Mabs have been used
to map domains in structural proteins (56-61) or the structure of proteins associated
with membranes (62-64), to analyze structural changes in proteins upon ligand binding
(65—74), and to study the structure of enzyme active Sites (72-80).

Several examples of how Mabs have been used to study protein structure and function
will be examined in greater detail in the remaining portions of this chapter.

One elegant use of Mabs to map protein topography involved a study of the

structural protein Spectrin(56). Spectrin iS composed of two unique high molecular

15

Figure 2. Proposed evolutionary relationship between the hexokinases. The
catalytic site is represented by a filled circle, and the Glc—6-P binding regulatory Site
by a filled square. In this scheme, a 50 kDa ancestral hexokinase, insensitive to
inhibition by Glc-6-P, evolved in two directions, one of which led to the present day
yeast hexokinase (insensitive to inhibition by Glc—6-P). In the second evolutionary
pathway, a Glc-6-P binding Site evolved on the 50 kDa protein, resulting in an
enzyme with properties similar to those seen in the present-day starfish enzyme. The
ancestral gene coding for the 50 kDa Glc—6-P sensitive hexokinase then underwent
gene duplication and fusion giving rise to the family of 100 kDa mammalian

hexokinases (22).

16

A8 82;? e5 0%? soc 815$

30:30er co=oEEoz

— p
- ‘

 

 

OmOCEOXOI metOum I . I OmUCEOXOI uw00>

 

 

lliil _. O
4/ E
1.1.1:}... . e

/ \\

 

«emacfloxom 358:4

17

weight subunits, a and 6. Two dimensional peptide mapping indicated the a-subunit
contained five antigenically distinct and proteolytically resistant structural domains.
Nine Mabs were isolated which bound the a-subunit of spectrin. The specificity of
the antibodies for individual peptides of known origin confirmed the previously
established alignment. In no cases were antibodies reactive with one domain found to
cross-react with another domain, confirming the antigenic uniqueness of each of the
five domains. In addition, the identity of smaller proteolytic peptide fragments
resulting from further cleavage of the domains was established. The results made it
possible to construct a linear map of the primary structure of the protein to be
constructed. Analogous studies were also done with fodrin (57) and fibronectin (5 8).
Several investigators have isolated Mabs that modulate enzymatic activity and
used them to investigate the nature of the catalytic site in proteins (72-80). Cazaubon
et al. (72) generated Mabs against protein kinase C , a
monomeric protein composed of two functional domains. The catalytic domain can
function as a Ca2+- and phospholipid-independent protein kinase, while the regulatory
domain can function as a phospholipid-dependent phorbol-ester—binding protein.
Consistent with their epitope location, two Mabs that bind to the regulatory
domain inhibit cofactor-dependent activity. Further analysis indicated that antibody
blocked cofactor binding. Conversely, antibody did not bind to protein kinase C
preincubated with phosphatidylserine and phosphatidylserine/phorbol ester. Epitope
mapping results indicated that the antibodies bound to a Site on the molecule distinct

from the cofactor binding site and therefore, negative interactions observed between

18

these Mabs and the cofactors did not result from competition for the same Site. The
authors concluded that loss of irnmunoreactivity was the consequence of
conformational changes in the protein masking the epitopes. Thus, the Mabs
identified a region within protein kinase C that underwent structural rearrangement
upon binding of ligand that was critical for enzyme activity.

Another interesting example that illustrates the effect that Mab binding can
have on the kinetic parameters, comes from work with liver phenylalanine
hydroxylase (74). Binding of PH-l to phenylalanine hydroxylase is dependent on a
phenylalanine induced conformational change, and in the absence of ligand, binding
does not occur. Normally phenylalanine acts a positive effector of enzyme activity
but in the presence of Mab PH-l , the co-operativity is lost leading to reduced
maximal enzyme activity. The inhibition by PH-l is non-competitive with respect to
tetrahydropterin cofactor. The authors suggest that binding of PH-l occurs at or near
the regulatory or activator site for phenylalanine on the enzyme molecule and
provides evidence for the existence of discrete catalytic and regulatory Sites. Other
investigators working with phosphofructokinase/ fructose 2,6-bisphosphatase (79) and
asparagine synthetase (80) used Mabs to establish either enzyme bifunctionality, or the
topographical separation of catalytic sites, respectively.

Mabs have also been used to study ligand binding by enzymes and other
proteins. Five Mabs against tryptophan synthase (66) had their affrrrity modified by
the ligands pyridoxal 5 ’-phosphate and L-Serine. Presumably, the binding of specific

ligands brought about conformational changes which altered the binding site for the

l9

antibodies. Similarly, three Mabs against protein C, can bind only in the presence of
Ca2+, known to produce a Significant conformational change in this protein (67). In
addition, the three conforrnationally sensitive Mabs cross-react with two other vitamin
K dependent proteins, prothrombin and Factor x, in the presence of Ca2 + , indicating
that the proteins share, at least partially, a common metal ion-induced three-
dirnensional structure.

Using a similar line of reasoning, Dixit et a1. (68) screened a panel of Mabs
raised against thrombospondin for Mabs that would react with the protein only in the
absence of Ca2+ , thus identifying epitopes not present after a Ca2 +- induced
conformational change. The location of the epitopes within the primary and tertiary
structure was determined, thus establishing a link between structure and function
within the protein. From these, and other examples (65 ,69-72), it is evident that
Mabs are sensitive and precise probes of protein conformational States and can be

used to establish structure-function relationships within a complex globular protein.

l)
2)

3)

4)

5)

6)

8)

9)

10)
I 11)
12)
13)
14)

15)

REFERENCES
Wilson, J.E.(l984) in Regulation of Carbohydrate Metabolism (Beitner, R.,
Ed.), pp. 45-85, CRC Press, Boca Raton, FL.

Schwab, DA, and Wilson, J.E. (1989) Proc. Natl. Acad. Sci. USA 86, 2563-
2567.

Harrison, R. (1985) Ph.D. Thesis (Yale Univ., New Haven, CT).

Anderson, C.M., Stenkamp, R.E., McDonald, RC, and Steitz, T.A. (1978)
J. Mol. Biol. 123, 15-33.

Bennett, W.S., Jr., and Steitz, T.A. (1980) J. Mol. Biol. 140, 211-230.
Grossbard, L., and Schimke, RT. (1966) J. Biol. Chem. 241, 3546-3560.
Mandel, P. (1964) in Progress in Nucleic Acid Research and Molecular
Biology, Vol 3, (Davidson, J .N., and Cohn, W.E., Ed.), Acedemic Press,
New york.

Ning, J., Purich, D.L., and Fromm, HI. (1969) J. Biol. Chem. 244, 3840.

Rose, I.A., Warrns, J.V.B., and Kosow, DP. (1974) Arch. Biochem.
Biophys. 164, 729-735.

Bartlett, G.R. (1968) J. Biol. Chem. 156, 231-239.

Passonneau, J.V., and Lowry, OH. (1969) J. Biol. Chem. 244, 902-909.
White, T.K., and Wilson, J.E. (1990) Arch. Biochem. Biophys. 277, 26-34.
Crane, R.K., and Sols A. (1954) J. Biol. Chem. 210, 597-606.

Sols, A., Crane, R.K. (1954) J. Biol. Chem. 210, 581- 596.

Purich, D.L., Fromm. H.J., and Rudolph, RB. (1973) in Advances in
Enzymology (Meister, A., Ed.), Vol. 39, pp. 249-326.

20

l6)

17)
18)

19)

20)

21)

22)
23)
24)
25)
26)

27)

28)

29)

30)

31)

32)

21

Casazza, LR, and Fromm, HI. (1976) Arch. Biochem. Biophys. 177, 480-
487.

Solheim, LR, and Fromm, H]. (1981) Arch. Biochem. Biophys. 211, 92-99.

Lazo, P.A., Sols, A., and Wilson, J .E. (1980) J. Biol. Chem. 255, 7548-
7551.

White, T.K., and Wilson, J .E. (1987) Arch. Biochem. Biophys. 259, 402-411.

Schirch, D.M., and Wilson, J.E. (1987) Arch. Biochem. Biophys. 254, 385-
396.

Nemat-Gorgani, M., and Wilson, J .E. (1986) Arch. Biochem. Biophys. 251,
97-103.

White, T.K., and Wilson, J.E. (1989) Arch. Biochem. Biophys. 274, 375-393.
Wilson, J.E. (1979) Arch. Biochem. Biophys. 185, 88-99.

Hutny, 1., and Wilson, J.E. (1990) Arch. Biochem. Biophys. 283, 173-183.
Wilson, J.E. (1968) J. Biol. Chem. 243, 3640-3647.

Wilson, J.E. (1978) Trends Biochem. Sci. 3, 124-125.

Polakis, P.G., and Wilson, J .E. (1985) Arch. Biochem. Biophys. 236, 328-
337.

Xie, G., and Wilson, J.E. (1988) Arch. Biochem. Biophys. 267, 803-810.

Felgner, P.L., and Wilson, J.E. (1977) Arch. Biochem. Biophys. 182, 282-
294.

Felgner, P.L., Messer, J.L., and Wilson, J .E. (1979) J.Biol. Chem. 254, 4946-
4949.

Linden, M., Gellerfors, P., and Nelson, B.D. (1982) FEBS Lett. 141, 189-
192.

Fiek, C., Benz, R., Roos, N ., and Brdiczka, D. (1982) Biochim. Biophys.
Acta 688, 429-440.

33)

34)
35)
36)

37)
38)

39)

40)

41)

42)

43)

44)

45)

46)

47)

48)

49)

22

BeltrandelRio, H., and Wilson, J .E. (1991) Arch. Biochem. Biophys. 286,
183-194.

Xie, G., and Wilson, J.E. (1988) Arch. Biochem. Biophys. 276, 285-293.
Bens, R. (1985) CRC Crit. Rev. Biochem. 19,145-190.

Wilson, J.E. (1979) Arch. Biochem. Biophys. 196, 79-87.

Baijal, M. and Wilson, J.E. (1982) Arch. Biochem. Biophys. 218, 513-524.

Gonzalez, C., Ureta, T., Babul, J., Rabajille, E., and Niemeyer, H. (1967)
Biochemistry, 6, 460.

Katzen, H.M., Soderrnan, DD, and Nitowsky, HM. (1965) Biochem.
Biophys. Res. Comm. 19, 377.

Katzen, H.M., and Schimke, RT. (1965) Proc. Natl. Acad. Sci. USA., 54,
1218.

Ureta, T. (1982) Comp. Biochem. Physiol. 71B, 549-555.

Colowick, SP. (1973) in The Enzymes (Boyer, P.D., Ed.), 3rd ed., Vol. 9,
pp.1-48, Acedemic Press, New York.

Easterby, IS, and O’Brien, M.J. (1973) Eur. J. Biochem, 38, 201-211.

Rose, I.A., Warms, J.V.B.,and Kosow, DP. (1974) Arch. Biochem. Biophys.
164, 729-735.

Holroyde, M.J., and Trayer, LR, and Cornish-Bowden, A. (1976) FEBS Lett.
62, 213-219.

Gregoriou, M., Trayer, LR, and Conrish-Bowden, A. (1983) Eur. J.
Biochem. 134, 283—288.

Andreone, T.L., Printz, R.L., Pilkis, S.J., Magnuson, M.A., and Granner,
DR. (1989) J. Biol. Chem. 264, 363-369.

Schwab, D.A., and Wilson, J.E. (1991) Arch. Biochem. Biophys. 285, 365-
370.

Thelen, A.P., and Wilson J.E. (1991) Arch. Biochem. Biophys. in press.

50)

50a)

51)

52)

53)

54)
55)

56)

57)

58)

59)

60)

61)

62)

63)

64)

23

Parry, D.M., and Pedersen, PL. (1983) J. Biol. Chem. 258, 10904-10912.

Denis-Puxviel, C., Riesinger, I, Buhler, C., Brdiczka, D., and Murat, J.
(1987) Biochim. Biophys. Acta 902, 335-348.

Lazarus, N.R., Rarnel, A.H., Rustum, Y.M., and Barnard, EA. (1966)
Biochemistry 5, 4003-4016.

Kopetzki, E., Entian, K.-D., and Mecke, D. (1985) Gene(Amst.) 39, 95-102.

Stachelek, C., Stachelek, J., Swan, J ., Botstein, D., and Konigsberg, W.
(1986) Nucleic Acids Res. 14, 945- 963.

Mochizuke, Y., and Hori, SH. (1977) J.Biochem. 81, 1849-1855.
Kéhler, G., and Milstein, C. (1975) Nature 256, 495.

Yurchenco, P.D., Speicher, D.W., Morrow, J.S., Knowles, W.J., and
Marchesi, V.T. (1981) J.Biol. Chem. 257, 9102-9107.

Glenney, J.R., Jr., Glenney, P., and Weber, K. (1983) J. Mol. Biol. 167,
275-293.

Dziadek, M., Richter, H., Schachner, M., and Timpl, R. (1983) FEBS Letts.
155, 321-325.

Ishiguro, H., Higashiyama, S., Ohkubo, I., and Sasaki, M. (1987)
Biochemistry 26, 7021-7029.

Strickland, M.S., Thompson, N.E., and Burgess, RR. (1988) Biochemistry
27, 5755-5762.

Weldon, S.L., and Taylor, S.S. (1985) J. Biol. Chem. 260, 4203-4209.

Dunn, S.D., Tozer, R.G., Antczak, DE, and Heppel, LA. (1984) J. Biol.
01cm. 260, 10418-10425.

Ratman, M., Sargent, P.B., Sarin, V., Fox, L.J., Nguyen, D.L., Rivier, J.,
Criado, M., and Lindstrom, J. (1986) Biochemistry 25, 2621-2632.

Morgan, DO, and Roth, RA. (1986) Biochemistry 25, 1364-1371.

65)

66)

67)
68)

69)

70)

71)

72)

73)

74)

75)

76)

77)

78)

79)

24

Jennings, I. G., Russell, R.G.Mcr., Armarego, W.L.F., and Cotton, R.G.H.
(1986) Biochem. J. 235, 133-138.

Djavadi-Ohaniance, L., Friguet, B., and Goldberg, M.E. (1986) Biochemistry
25, 2502-2508.

Wakabayashi, Y.S., and Aoki, N. (1986) J. Biol. Chem. 261, 11097-11105.

Dixit, V.M., Galvin, N.J., O’Rourke, K.M., and Frazier, W.A. (1986) J.
Biol. Chem. 261, 1962-1968.

Friguet, B., Djavadi-Ohaniance, L., and Goldberg, M.E. (1986) Eur. J.
Biochem. 160, 593-597.

Koepsell, H., Korn, K., Raszeja-Specht, A., Bemotat-Daneilowski, S., and
Ollig, D. (1988) J. Biol. Chem. 263,

Ratnam, M., Xuehai, T., Prendergast, N.J., Smith, P.L., and Freisheirn, J .H.
(1988) Biochemistry 27, 4800-4804.

Cazaubon, S., Webster, C., Camoin, L., Strosberg, A.D., and Parker, P.
(1990) Eur. J. Biochem. 194, 799-804.

Cazaubon, S., Marais, R., Parker, R, and Strosberg, AD. (1989) Eur. J.
Biochem. 182, 401-406.

Pamiak, M.A., Jennings, 1.6., and Cotton, R.G.H. (1989) Biochem. J. 257,
383-388.

Accolla, R.S., Cina, R., Montesoro, E., Celada, F. (1981) Proc. Natl. Acad.
Sci. USA. 78, 2478-2482.

Djavadi-Ohaniance, L., Friguet, B., and Goldberg, M.E. (1984) Biochemistry
23, 97-104.

Lad, P.J., Schenk, DB, and Leffert, H.L. (1984) Arch. Biochem. Biophys.
589-595.

Dunn, SD, and Tozer, R.G. (1987) Arch. Biochem. Biophys. 253, 73-80.

Van Schaftingen, E., Coulie, P.G., Van Snick, J., and Hers, H.-G. (1986)
Eur. J. Biochem. 160, 367-373.

25

80) Pfeiffer,N.E., Mehlhaff, P.M., Wylie, DE, and Schuster, SM. (1987) 262,
11565-11570.

CHAPTER 2

Generation of Monoclonal Antibodies

26

Introduction

Numerous monoclonal antibodies (Mabs) against rat brain hexokinase have
previously been produced in the laboratory and ten of these have been partially
characterized by Finney et a1. (1) and Wilson and Smith (2). However, the majority
of Mabs generated were of the IgM class and did not react with native hexokinase. In
addition, those antibodies that were reactive with the native enzyme were directed
against epitopes in the N-terminal portion of the molecule. Since the goal of my
project was to establish structure-function relationships within rat brain hexokinase, it
was necessary to isolate additional Mabs reactive with native hexokinase. Therefore,
this project initially focused on methods to enhance the production of Mabs to native
hexokinase.

The large numbers of IgM antibodies produced were an indication that the
immunization protocol was inadequate, and led to a poor immune response (3). In
this early work, mice were first primed with hexokinase emulsified in complete
Freund’s adjuvant, then boosted with further injections emulsified in incomplete
Freund’s adjuvant. Initial tests of serum from mice immunized in this fashion
indicated that only a very poor anti-hexokinase response was being elicited. To
circumvent this problem, a number of alternative immunization protocols were tested
to try to enhance the immune response of mice to hexokinase as well as produce

27

28

Mabs against additional unique regions of the molecule. The results from these

experiments will be presented in this chapter.

MATERIALS AND METHODS

Materials

Tissue culture reagents were obtained from the following sources: Dulbecco’s
modified Eagle media (cat. no. 430—1200), normal horse serum, fetal bovine serum,
and gentamicin were all obtained from Gibco Laboratories (Grand Island, NY).
Falcon (Becton Dickinson and Company, Oxnard, CA) or Corning (Corning, NY)
tissue culture ware were used in all phases of hybridoma development. Nunc 96-well
microtiter plates (cat. no. 269620) for use in ELISAS were obtained from Thomas
Scientific (Swedesboro, NJ).

CD-l (Swiss outbred) mice were from the Charles River Co. (Wilmington,
MA). Sp2/O-Ag-14 mouse myeloma cells use in the cell fusion procedure were
kindly provided by Dr. William Smith (Department of Biochemistry, Michigan State
University). Freund’s adjuvant, complete and incomplete, were obtained from Difco
Laboratories (Detroit, MI). RIBI adjuvant (cat. no. R-700) was obtained from RIBI
ImmunoChem Research, Inc. (Hamilton, MT). 100 mesh stainless steel screen used
to generate single cell suspensions from spleen or popliteal lymph nodes was obtained
from Newark Wire Cloth Company (Newark, NJ).

Horseradish peroxidase coupled to goat anti-mouse Ig or IgG was obtained

29

30

from Cooper Biomedical (Malvem, PA), or BioRad Laboratories (Richmond, CA),
respectively. Rabbit anti-mouse immunoglobulin serum was purchased from Accurate
Chemical and Scientific Corp. (W estbury, NY). Colchicine, aminopterin,
hypoxantlrine, thymidine, adjuvant peptide (N -acetylmuramyl-L-alanyl-D-
isoglutarnine) and o-phenylene diamine were obtained from Sigma (St. Louis, MO).
Polyethylene glycol, mol. wt. 1450, for use in cell fusions was obtained from
J.T.Baker Inc. (Phillipsburg, NJ). The BCA Protein Assay kit was obtained from
Pierce (Rockford, IL). All other common biochemicals were obtained from
commercially available sources.

Denatured hexokinase was prepared as described by Finney et al. (1).
Polyclonal antiserum raised against rat brain hexokinase followed the general
procedure described in (4) with the substitution of RIBI adjuvant R-730 in place of
Freund’s adjuvant. S. aureus cells were obtained from The Enzyme Center (Malden,

MA) and prepared as described by Finney et al. (1).

. n . '-h x . . i

Affinity purified anti-hexokinase antibodies were obtained by affinity
chromatography on an Affi-Gel 10 (Bio-Rad Laboratories, Richmond, CA) column to
which purified rat brain hexokinase had been covalently linked. Rabbit serum (2 ml)
was applied to the column, the effluent collected and reapplied to the column. This
was repeated two additional times and then the column washed with 15 column

volumes of 50 mM Tris, pH 7.5, containing 120 mM NaCl, 0.5% NP-40, and 0.02%

31

NaN3. After washing with 10 column volumes of PBS (10 mM sodium phosphate,
pH 7.0, containing 154 mM NaCl) to remove the NP-40, the antibodies were eluted
off of the column with 50 mM Tris-glycine pH 2.5. One ml fractions were collected
into tubes containing 0.1 ml of l M Tris pH 9.0. Fractions containing protein were
combined, concentrated, and dialyzed against PBS. The final antibody concentration

was determined using an A280 m of 1.35 for a 1 mg/ml solution (5).

Purifigafig' n Qf 131 brain hexokinase and measurement of enzyme activity.

Rat brain hexokinase was purified as described previously (6), and enzyme

activity measured spectrophotometrically using a Glc-6-P dehydrogenase—coupled

assay (7)-

Several different immunization protocols were tested as described below.
When Freund’s adjuvant was used, it was prepared by mixing equal parts of antigen
solution and adjuvant, and then briefly (app. 30 sec) sonicated to produce an
emulsion. RIBI adjuvant came supplied as a lyophilized preparation that was
reconstituted with PBS and vigorous vortexing for several minutes. Mice injected
with RIBI adjuvant received 50 pg each of monophosphoryl lipid A and trehalose
dimycolate, the active ingredients in the adjuvant. In addition, some mice were also
injected with colchicine (1.0 mg/kg body weight) dissolved in PBS. Colchicine

injections were concurrent with hexokinase immunizations.

32

Protocol 1- Fifty micrograms of hexokinase (specific activity > 55 U/ mg) emulsified
in Freund’s complete was injected i.p. into mice. Three weeks later, mice were
boosted with 25 pg of hexokinase i.p. emulsified in Freund’s incomplete. Mice were
then rested at least two weeks. Three days prior to the fusion procedure mice were

boosted with 25 ug of hexokinase in Freund’s incomplete adjuvant.

Protocol 2- Protocol 2 was identical to protocol 1 except that RBI adjuvant was

substituted for Freund’s adjuvant.

Protocol 3- Mice were injected according to protocol 2 but also received a separate

injection of colchicine (1.0 mg/kg body weight) in PBS.

Protocol 4- Fifty micrograms of hexokinase in RBI adjuvant (50 pg in app. 0.2 m1
adjuvant) was injected into the hind footpads (0.1 ml/footpad) of mice. On day 13
after the initial injection, the procedure was repeated. On day sixteen the mice were
sacrificed, and lymphocytes isolated from the popliteal lymph nodes as described

under "Production of hybridoma cell lines".

Protocol 5- Hexokinase in 0.1 M potassium phosphate pH 7.0 containing 0.5 mM
EDTA, 0.1 M glucose and 10 mM thioglycerol (storage buffer) was digested with
trypsin at a trypsin:hexokinase ratio of 1:2 for two hours at room temperature. The

resulting digest was subjected to SDS-PAGE on a 6.5-20% gradient and the bands

33

subsequently visualized by soaking the gel in ice cold 0.2 M potassium chloride (8).
The 40 kDa band, representing the C-terrninal portion of hexokinase, was cut out of
the gel and the protein was electroeluted. After exhaustive dialysis in PBS , the
protein concentration was determined by the Pierce BCA Protein Assay using BSA as
a standard following manufacturers instructions. A sample of the isolated fragment
was subjected to SDS-PAGE to confirm purity. Mice were immunized with 20 pg of
the isolated fragment in RBI adj uvant. Twenty-three days later they were reinjected
with 25 pg of this fragment in RBI adjuvant and also injected with colchicine. After
resting the mice for one month, they were reinjected in an identical manner three days

prior to the cell fusion procedure.

Protocol 6- This procedure was adapted from the method used by Hockfield (9).
Neonatal mice were injected i.p. with 36 pg of hexokinase in storage buffer without
adjuvant every other day from birth through day 10. Then at six weeks of age, the
mice were reinjected with 4 pg of hexokinase in RBI adjuvant i.p. and 2 pg in each
footpad, 4 days apart, for five consecutive immunizations. One day after the last
injection the mice were sacrificed and cell fusions performed on lymphocytes isolated

from both the spleen and popliteal lymph nodes.

Protocol 7- The 48 kDa C-terminal domain of hexokinase was isolated using the
procedure of White and Wilson (10). Mice were initially injected with 30 pg of

hexokinase in RBI adjuvant. Two weeks later the mice were reinjected with 20 pg

34

of the C-terminal domain and after an additional two weeks, with 10 pg of the C-
terminal domain in RBI adjuvant with 1.0 mg/kg colchicine. After resting the mice
for three months, they were challenged with 15 pg of C-tenninal domain and 1.0
mg/kg colchicine in RBI adjuvant. Three days later the isolated spleen lymphocytes

were used in the cell fusion procedure.

The method of Boss (1 l) was employed for an in vitro immunization and will
be briefly described here. Single cell spleen lymphocyte preparations from two
unimmunized or two previously primed mice (protocol 2) were incubated with 2 pg of
hexokinase in DMEM, 50 pM 2-mercaptoethanol, 2 mM glutamine and 1 mM sodium
pyruvate for 1.5 hours. Fetal calf serum (20%) and adjuvant peptide (20
pg/ml,Sigma cat. no. A 9519) were then added, and cells cultured for four
days. Cells remaining after the incubation period were used for fusions to produce

hybridoma cell lines.

2 l . E! l . I 11 1' .

The cell fusion protocol is a modification of the procedure of Galfre et al.
(12,13). Spleens or lymph nodes from immunized mice were passed through a 100
mesh stainless steel screen to generate a single cell suspension. Cells were suspended
in DMEM containing 20 mM Hepes pH 7.6 (DMEM/Hepes) and centrifuged (7

minutes at 300 X g) rpm to pellet cells. The media was removed and 5 mls of 0.2%

35

NaCl added and the cells resuspended by pipetting up and down. After 30 seconds, 5
ml of 1.6% Nacl was added, followed after 30 seconds by 10 ml of DMEM/Hepes.
Red blood cell debris was removed, and the cell suspension centrifuged. After
removing the supernatant, the pellet was resuspended in 10 ml of DMEM/Hepes and
the number of lymphocytes present quantitated with a hemacytometer.

Sp2/o-Ag-14 mouse myeloma cells (app. 15-100 mm dishes in logarithmic
growth phase) were harvested, washed once with DMEM /Hepes, and the number of
myeloma cells quantified. Lymphocytes and Sp2 cells were mixed at a ratio of 2:1,
lymphocytes:Sp2 cells, in a silanized glass corex tube with cap and centrifuged. After
removal of the supernatant, 1 ml of fusing solution (7 gm of polyethylene glycol-
1000, previously autoclaved and stored at 37°C to keep liquid, to which 1 ml of
DMSO and DMEM/Hepes, to 20 ml, had been added) prepared just prior to use, was
added slowly and the pellet disrupted, but not resuspended, by tapping the tube
against the counter. This step was critical, and as many small clumps of cells as
possible need to be generated in the first 30 seconds for maximum yield of
hybridomas (personal observation).

At one and two minutes after addition of the fusing solution, 1 ml and 2 ml of
DMEM/Hepes were added, respectively. At four minutes after addition of fusing
solution, 4 ml of HT-Complete1 media was added, and at seven minutes, an additional
8 ml of media was added. The cells were centrifuged ten minutes after addition of

fusing solution, and resuspended in 150 ml of HT-Complete. Cells were then

187-Couple“, DMEM containing 10% FBS, 10% horse serun, 2 m4 glutamine, 50 ml of acre-109 media.

36

dispensed into 96 well tissue culture plates. Hybrids were selected by using a media
containing hypoxanthine, thymidine, and aminopterin (14).

Wells containing cells producing anti-hexokinase antibodies were identified by
using one of two different ELISA methodologies (described under EIJ_S_A_S), and were
cloned by the method of limiting dilution (15). Positive cell lines were propagated
both for antibody production and for storage under liquid N 2 (cells were frozen in calf

semm containing 10% DMSO).

MW

To determine whether Mabs were precipitating native or denatured hexokinase,
a solution phase immunoassay was employed. To 100 pl of spent culture medium,
11.1 pl of 1 M TriSCl-5 mM EDTA, pH 8.5, was added. After addition of 0.5 pg of
either native or l4C-labeled denatured hexokinase, the samples were incubated for 45
minutes at room temperature. S. aureus cells, 45 pl of a 20% wt/vol suspension that
had been pretreated with rabbit anti-mouse irnmunoglobulins (1), were added and the
incubation continued for an additional 45 minutes. After centrifugation to pellet the
cells, the supernatant was assayed for residual hexokinase activity in the case of the

native enzyme, or radioactivity when denatured enzyme was used.

37

ELISAS

Four different variants of the ELISA were used. These methods were identical
except for the manner in which the antigen was immobilized on the microtiter plates.
For convenient reference, these methods are summarized in Table I. In method A,
0.1 ml of an aqueous solution containing 1 pg of hexokinase was placed in the wells
and the plates allowed to air dry; this was the method used by Finney er al. (1) in the
initial studies on Mabs against rat brain hexokinase. In method B, 0.1 ml of a
solution containing 10 pg hexokinase per ml in 50 mM sodium borate, pH 8.5 , was
added to the wells and the plates incubated overnight at room temperature in a humid
chamber. In methods C and D, the wells were first coated with polyclonal anti-
hexokinase antibodies (rabbit) by incubation ovenright in a humid chamber with 0.1
ml per well of a solution containing 10 pg of the affinity purified antibodies per ml in
50 mM sodium borate pH 8.5. The plates were then thoroughly washed with PBS-
Tween (this solution, PBS containing 0.05% Tween-20, was used for all washes in
the ELISA procedures) and incubated 1.5 hours with a solution of 0.5 % gelatin in
PBS to block all remaining protein binding sites. For method C, the plates were then
incubated for 1.5 hours with a solution (0.1 ml per well) containing 10 pg purified
hexokinase per ml in PBS-Tween containing 0.2% BSA.

After treatment with one of the above methods, plates were thoroughly washed
with PBS-Ween. In methods A, B, and C, 0.1 ml of culture medium containing
various antibodies, or control medium was added directly to the wells. For method

D, the medium was preincubated for 1 hour at room temperature with 0.5 pg

38

Table 1. Summary of methods used to prepare microtiter plates for ELISA.

Method
A

B

Antigen (hexokinase) Immobilized By:
Air drying directly onto plastic surface
Adsorption onto plastic surface from 50 mM sodium borate; pH 8.5 .

Binding of antigen by polyclonal (rabbit) antibodies adsorbed to plastic
surface.

Binding of pre-fonned antigen-Mab complex by polyclonal (rabbit)
antibodies adsorbed to plastic surface.

39

hexokinase (reduced to 0.1 pg in later experiments) before addition to the wells.
After addition of medium, the plates were incubated for one hour at room
temperature, washed, and incubated for another hour with a 1:1000 dilution of
horseradish peroxidase-conjugated to goat anti-mouse immunoglobulin (0.1 ml per
well). After washing, adsorbed peroxidase activity was detected using 2.2 mM 0-
phenelene diamine 0.03% H202 in 50 mM sodium citrate, pH 4.0 (0.1 ml per well).
Color reactions were terminated by addition of 0.1 ml per well of 4 N 1-12SO4 to each

well and absorbance measured at 490 nm using a BioTek EL-307 plate reader.

Assessment of murine anti-hexokinase immune gspgnse.

To monitor the anti-hexokinase immune response of mice immunized by
various protocols, mice were tail-bled and the anti-hexokinase antibody titers
determined. Routinely, 20 pl of blood was diluted into PBS containing 1 mg per ml
of heparin sulfate and centrifuged to remove red blood cells. The diluted sera was
frozen at -20 °C until assayed via ELISA method B or D. Prior to analysis, the
samples were further diluted into PBS-Tween containing 0.2 % BSA. When using
method B for quantitative measurements, the titer of a mouse was defined as the
dilution of serum which gave an absorbance in the ELISA of 0.15 at 490 nm after 5

minutes incubation with substrate.

RESULTS

ngpag' son Qf adjuvan; systems.
Initial attempts at producing Mabs by fusing Spleen lymphocytes isolated from

CD-l nrice immunized with hexokinase in Freund’s adjuvant, to Sp2/O-Ag14 mouse
myeloma cells resulted in very few Mabs of the IgG class. Line 1 in Table II
summarizes the results of eight initial fusions that resulted in production of 34 lines
producing anti-hexokinase Mabs; 27 of the Mabs did not react with native hexokinase.
These results were attributed, in part, to a poor immune response of the mice to
hexokinase. In an attempt to increase the immune response, a alternative adjuvant
from RBI ImunoChem Research, Inc. (Hamilton, MT), was tested to determine if it
could elicit a stronger immune response to hexokinase in mice than did Freund’ s
adjuvant. In addition, the drug colchicine was co-administered with RBI adjuvant to
assess whether the immune response could be further stimulated by addition of a drug
previously shown to enhance the immune response of mice to antigens (16,17). The
results shown in Fig. 1 are the anti-hexokinase immune response of mice at 14 days
after the second immunization with hexokinase in either Freund’s (Protocol 1) or RBI
adjuvant (Protocols 2 and 3). As the results clearly indicate, immunization of mice

with hexokinase in RBI adjuvant, with or without colchicine, produced superior

41

Comparison of hybridoma cell lines secreting

anti-hexokinase Mabs resulting from different immunization

Table II.
protocols.
Immunization
Method
1) 1
2) 2
3) 3
4) 3
5) 4
6) a)in vitro-
unprimed
b)in vitro-
primed
7) 5
8a)b 6
b)c 6
9) 7

ND-not determined

ELISA(+'
Method A or B

34

26

28

107

28

ll

59

66

109

s)

# of
Ing

27

15

6

66

109

# of
IgGs

7
11
ND

588

6

“6

ELISA
Method D

7

3

58

62

aClass determined only on cell lines positive in ELISA
method D.
bLymphocytes used in fusion were derived from spleens.

cLymphocytes used in fusion were derived from popliteal
lymph nodes.

42

Figure l. The effect of adjuvant on anti-hexokinase immune response in mice.
Mice were immunized with hexokinase emulsified in complete Freund’s or in the
RBI adjuvant. A third group, immunized using RBI adjuvant, was also injected
with colchicine. After three weeks, mice were boosted and periodically tail-bled.

The mouse sera obtained were diluted 1:100 and tested for anti-hexokinase antibodies
using ELISA method B. Each bar represents the anti-hexokinase immune response of
an individual mouse 14 days after the second immunization. All assays were done in

triplicate and the standard deviation determined.

43

 

 

 

VVx““‘\\““x‘—‘—rvxwfi \\\\\\\\\
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

R\

 
   

      
  
  
   

         

   

+R\\\'\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

 

   

   

 

\\
\\

I

 

I
I

        

 

 

 

|

i
I I 1 I
"a O. to. 0.
1- 1- o o

uru 0517 eoueqrosqv

ne

RIBI+Co|chici

IBI

Freund's

44

immune responses to those of the group treated with Freund’s adjuvant.

No significant differences were seen between the RBI adjuvant and RBI
adjuvant plus colchicine treatments under the assay conditions used (1:100 dilution of
serum for all samples). However, further dilution of the serum obtained from the
RBI adjuvant and RBI adjuvant plus colchicine immunized mice revealed significant
enhancement of the anti-hexokinase immune response in mice treated with colchicine.
In Table III, the anti-hexokinase titers of mice in the two groups are compared at days
4, 7, 10, and 14 after the second immunization. Although a fair amount of variability
in response existed between animals, by day 10 the response of the mice to
hexokinase in the RBI-colchicine group was twice that of the group that received
only RBI adjuvant. Based on these results all subsequent immunization of mice was
done using hexokinase in RBI adjuvant and, in some cases, colchicine was also co-

administered.

stslgpmgnt of a mgflififi ELISA that sslgptivsly detects mgnmlgnal antibodies
W

As the fusion described in line 2 of Table 11 indicates, use of RBI adjuvant
alone led to a substantially larger number of anti-hexokinase Mabs being generated
than was seen with 8 previous fusions using mice immunized with Freund’s adjuvant
(Table 11, line 1). However, Mabs of the IgM class were still being generated and
even the Mabs of the IgG subclass were not necessarily recognizing native

hexokinase. We noted, as well as several other investigators (18-23), that the

45

douche 830 2:58 o>¢ a .85 2 .o me E: omv “a 859.48% ca 63» 3 Emcee:
EBB mo sesame 2: ma Beau—c was .48: BE. .39:on new £532 5 concave as m 352:
«Sam .3 358.42% :2: 863 ES: avenues 034.232-5944. .33 :8 863 some 05 nous—58::
958m 05 Bee 3 use .3 K .v 996 :0 .m 28 N $8803 8 mew—:88 ESE—Em 263 83.

 

 

 

“-

ens

050228

883848 Rosalia Sovfiewme 823a; tam

33334 Binge 8443884 8%85 8.1552
, H :1 cm H ._   04 an , . _ 4...”..48 18M

 

 

 

 

 

 

 

(883 3355 Reflexoaéea mo courses—cu .E 035.

46

standard ELISA (often used to screen for hybridoma cell lines secreting antibody of
interest), with antigen adsorbed to the wells, did not discriminate between Mabs
recognizing native versus denatured antigen. Thus, it was apparent that adsorption to
plastic resulted in partial denaturation of the protein. Since the project necessitated
that a large library of Mabs be generated against native hexokinase, and due to the
high costs associated with the production of Mabs, it was desirable to develop a
methodology that would quickly identify those cell lines secreting Mabs against native
hexokinase. A solution phase assay (1), which used rabbit anti-mouse IgG coated S.
aureus to immunoprecipitate Mab-hexokinase complexes, was available that
discriminated between Mabs recognizing native versus denatured hexokinase.
However, the assay was not suited for the large scale screenings associated with
hybridoma production and therefore, a capture ELISA procedure was devised
(methods C and D) in which affinity purified rabbit polyclonal anti-hexokinase
antibodies adsorbed onto 96-well microtiter plates were used to capture Mab-
hexokinase complexes. This method was then compared to the previously used
screening techniques (methods A and B).

The results of ELISA using the methods A-D are presented in Table IV, along
with the results of immunoprecipitation experiments with both native and denatured
enzyme. In accord with the observations of many previous investigators (18-23)
working with other antigens, all Mabs show substantial reactivity in the conventional
ELISA (methods A and B), but the reactivity shows no correlation with the relative

ability of the Mabs to recognize the native antigen in the immunoprecipitation assay.

‘47

 

 

.ncoduchELouov omen» toe co4pm4>mo engages» a cmoe on» ma osam> comma

 

 

 

 

 

muom mama oouOHmouo mononmowo mononsmuo zououmwwo mom
4“» Puzm mo.oumo,o ooqonoowo motowmmto ~0.o«wm.o mo—
mam— Famm oouonoouo Pouonppqo wouonzmuo moqowmmto 2m,
mpam Fume Fotou—owo wo.oa-wo zo.o«~o.o zotoumoto .m—
Pao warm Fotowoowo ooqoaoo.o zowo«0muo moqowwowo mm.
mam .«m: motowmoqo oo.o«oouo mowonaoto Foqow—mto <NF
mam FAQ» sououmoqo mouowcoto Nouowpowo moqowmowo «m
Naom MANm mowowzwto mo.o«mp.o Fowoazow— cowonpotp Fom
mumm 0H0. ootoaomuo mowoazzuo zpaoamouo zowoumoto N04
oamm mn—P zo.on>m.o mo.onomto No.0«mo.— zo.o«oo.o For
mfimo. sz— Fowouomwo ootonmm.o oouoaomuo cowoaom.p no,
oaoo_ sum. Po.o«—>.o mo.o«mmwo No.0«Nm.o mo.o«—m.o mm—
oaoo_ Pup» Fouo«>mwo o—woammqo Foqonm—u— mowowmow. 4m
zwmw mam mo.oummto mptoamm.o mo.o«moq— po.ouzo.— om
mama —«~> m—wonmwt— mo.o«m>to o—wonzmw— o—qo«=ow— @—
oaoh. :«m mo.o«om.o mo.o«mmwo po.oumm.o ~0.o«mw.o m4
owoow zap motowzmqo po.o«~oqo nowonop.— zpwo«>—.— 0m
zapm mam —owoumowo Fo.o«mmqo motoH—mq— no.0woo.— mm
meow mum mo.o«om.o zo.ouo>.o No.0«m~.F >—.o«m—.P m—
ommconxm: onmcaxoxox o vogue: o venue: m conga: < vogue: xoonquc<
0>aumz UOLSuwcoo HGCOHOOCOZ

 

 

umump4aaowtaocseeH m Nddn<v nuaanmm «qum

mnoduonauc< HmcoHoocoz
E $2303: 595 pmm nmesumcma new 2332 Lo 53343330:ng use <m3m .>H manna.

48

Moreover, with the single exception of Mab 12A, reactivity is Similar with either
method A or B, implying that comparable denaturation accompanies the direct
adsorption of antigen to the plastic surface regardless of whether it occurs during the
drying process or from mildly alkaline solution.

In marked contrast, methods C and D effectively distinguish Mabs recognizing
the native antigen from those that do not. This was particularly the case with method
D where the Mab is allowed to react with free antigen in solution prior to addition to
the microtiter plate. Thus, every Mab showing Significant reactivity in ELISA by
method D was also found to be at least moderately effective at immunoprecipitating
the native enzyme. Conversely, every Mab that reacted poorly in method D was
shown to react preferentially if not exclusively with the denatured enzyme. It should
also be noted that concurrent screening by method D and by either method A or B
immediately reveals the presence of Mabs that react selectively with the native antigen
(positive in both methods) or with the denatured enzyme (positive in ELISA method A
or B). ELISA method D alone or in conjunction with method B were used in all

subsequent post-fusion screenings.

With the development of the modified ELISA for post-fusion screening of
hybridomas cell lines for anti-hexokinase antibodies it was possible to rapidly assess
whether cell lines secreting Mabs reactive with native hexokinase had been generated.

In the fusion described on line 3 of Table II, a mouse was used that had been

49

immunized with hexokinase in RBI adjuvant, and also injected with colchicine at 1.0
mg/kg body weight. This mouse had previously been shown (Figure 1, Table III,
RBI-colchicine group) to have a high anti-hexokinase titer as measured using ELISA
method 2 (hexokinase adsorbed directly onto microtiter plate). However, while cell
lines secreting anti-hexokinase antibodies reactive in ELISA method B were identified,
no cell lines secreting Mabs reactive with the native enzyme were identified using
ELISA method D. Thus it became apparent that some mice could generate a strong
anti-hexokinase immune response, and not produce antibodies reactive with native
hexokinase. We thus attempted to determine, a priori, which mice produced
antibodies against the native enzyme by prescreening sera from mice before fusions
were canied out.

Since ELISA method D selectively detects antibodies reactive with native
hexokinase, sera from mice immunized to determine the efficacy of the different
adjuvants used (Freund’s, RBI and RBI-colchicine groups previously described),
were tested for the presence of antibodies reactive with native hexokinase via ELISA
method D. Although several mice in the RBI and RBI-colchicine groups Showed
reactivity in this assay, one mouse reacted better than the rest. This mouse was
boosted one additional time with hexokinase in RBI adjuvant and colchicine. Spleen
cells from this mouse were used for hybridoma production. The results of the fusion
are summarized in line 4 of Table II. All wells with hybridomas were screened by
ELISA methods B and D. One hundred and seven ELISA method B positive cell

lines were identified of which only 58 were positive in ELISA method D. Since we

50

were interested in antibodies reactive with the native enzyme, only cell lines reactive
in method D were cloned.

Subsequent analysis of the Mabs indicated that the antibodies produced by
these hybridomas reacted with about 5-7 unique epitopes, and that all of the Sites were
located on the N-terminal half of the protein. Although immunization with
hexokinase in RBI adj uvant, and treatment with colchicine, coupled with
prescreening of mouse sera for the desired type of response, had led to the isolation
of a significant number of cell lines secreting anti-native hexokinase antibodies, the
immunization protocol did not allow us to isolate hybridomas secreting antibodies
reactive with the C-terminal portion of the native enzyme. Therefore an attempt was
made to generate additional cell lines secreting MabS against the C-terminal portion of
the native enzyme using several new protocols.

Mice were immunized according to protocol 4 (hind footpads with hexokinase
in RBI adjuvant) , and their popliteal lymph nodes removed. The lymphocytes derived
from these nodes were used in a cell fusion. The results of one such fusion is
summarized in line 5 of Table II. In this experiment only ELISA method D was used
for screening and 28 positive cell lines were identified. Antibodies from these clones
were then characterized for their effect on kinetic parameters and their
immunogloban class. Five cell lines were then chosen for further propagation. The
antibodies produced by hybridomas from this fusion were shown to be directed
against two unique epitopic regions. One antibody, 4A6, binds to the C-tenninal

portion of hexokinase, inhibits hexokinase activity, and also blocks Glc-6-P inhibition

51

of enzyme activity. The characteristics of these antibodies will be discussed in
greater detail in subsequent chapters. Thus the route and protocol for immunization
appear to have Significant effects on both the magnitude, as well as, the type of
response elicited, and subsequently, the specificity of antibodies produced after
fusion.

Further attempts were made at producing Mabs against the C-tenninal portion
of the enzyme and to other regions of the N-terminal domain. Most of these, such as
the in vitro immunization, and immunization with denatured C-terminal 40 kDa
tryptic fragment, produced no Mabs reactive with the native enzyme (lines 6 and 7 of
Table II). However, one other method did produce MabS reactive with the C-terminal
portion of hexokinase.

Particular epitopes are often immunodominant in proteins and an animal will
respond preferentially, but not necessarily exclusively to these sites. The results
described above, suggest that the N -terminal domain contains epitopes that are
immunodominant. An attempt was then made to circumvent the normal immune
response. Mice were immunized at birth with hexokinase to induce tolerance against
the immunodominant epitopes on the N -terminal half of hexokinase and then rested.
Subsequent immunization of these young adult mice with hexokinase did not produce
a significant response to hexokinase. However, we conducted fusions using both
spleen and popliteal lymph node derived lymphocytes fiom these mice. The results
are summarized in lines 8a and b of Table 11. Use of spleen derived lymphocytes did

not result in the generation of any anti-hexokinase producing cell lines as measured by

52

ELISA method D, while use of popliteal lymph node derived lymphocytes did result
in six cell lines that secreted antibodies reactive with native hexokinase. Of these,
three have subsequently been Shown to bind to the C-terminal domain of hexokinase,
inhibit catalysis, and will discussed in greater detail in subsequent chapters.

More recently it has become possible to isolate a 48 kDa fragment of
hexokinase representing the C-terminal domain of the enzyme that retains full
enzymatic activity (10). Since the immunodominant N-tenninal portion of the enzyme
is removed in these preparations, this provided an ideal opportunity to selectively
generate Mabs against the C-terminal portion of the enzyme. Mice were injected as
described in protocol 8 and their anti-hexokinase immune response was monitored by
ELISA methods B and D. AS the results in Fig. 2, ELISA method 1 indicate, the
mice did not produce a strong anti-hexokinase response when compared to the
response using a similar immunization protocol and the intact enzyme (Fig. 1 and
Table III). To determine if the immune response was directed against the native
molecule, serum samples were tested for reactivity against the isolated 48 kDa
fragment or intact hexokinase using ELISA method D. The results in Fig. 2, ELISA
methods 2 and 3 , demonstrate that an immune response was directed against surface
determinants present on both the isolated 48 kDa fragment and intact hexokinase as
well. Based on these results the mice were immunized an additional time, rested,
reinjected, and lymphocytes isolated for hybridoma production. The results of this
fusion are summarized in line 9 of Table II. Although 109 cell lines were detected

secreting anti-hexokinase antibodies by ELISA method B and 62 via method D, all of

53

Figure 2. Analysis of the anti-hexokinase immune response of mice immunized
with the isolated C-terminal domain of hexokinase. Mice were immunized as
described under protocol 7 and all animals were tail-bled 9 days later. Their anti-
hexokinase immune response was determined using ELISA methods B and D with
intact hexokinase as the antigen. The immune response against the C-terminal domain
of hexokinase was assessed using ELISA method D with the isolated C-terminal
domain. Each bar represents the response of an individual mouse. Serum samples
were diluted 1:100, assayed in triplicate and standard deviations determined. 1-
ELISA method B (antigen adsorbed to plastic) using intact hexokinase; 2- ELISA
method D using intact hexokinase; 3- ELISA method D using the isolated C-terminal

domain.

 

 

‘ ‘ » _.

  

 

 

 

 

 

°°.
o

f

I
“l
O

0.0

«2 Sr.
0 , O

qu 0517 eoueqrosqv

ELISA Method

the antibodies were of the IgM class, and therefore of limited usefulness.

55

DISCUSSION

The development of somatic hybridization techniques that allow isolation of
hybridoma cell lines producing MabS (24), has led to extensive use of such antibodies
in characterization of proteins at both the structural and functional levels. Production
of Mabs is relatively expensive and a labor intensive undertaking, so strategies which
optimize the production of desired Mabs are useful. Since successful production of
hybridomas is dependent on the ability to induce a strong immune response against
the immunogen, our primary focus was on designing Strategies to enhance the immune
response.

Choice of mouse Strain can be another critical factor in hybridoma production.
Most strains of mice have limited immunological repertoires. Initially, Balb C mice
were used but this was changed to outbred Swiss CD-1 mice when it became apparent
that the Balb C strain was unable to respond to hexokinase.

As was shown in Figure 1, and Table III, choice of adjuvant can also play a
major role in determining the magnitude of an immune response. It is quite apparent
that the RBI adjuvant is superior to Freund’ s adjuvant in eliciting an anti-hexokinase
immune response in CD-1 mice. The major components of the adjuvant, trehalose
dimycolate and monophosphorylipid A, have been Shown to be nonspecific inducers

of immune responses in a number of different systems (25-27).

56

57

In addition, drugs such as colchicine (16,17) and cyclophosphamide (28) are
known to modulate the immune response in mice. We found that co-admirristration of
colchicine led to a two-fold increase in the immune response to hexokinase. Both of
the drugs are thought to produce their effect by inhibiting suppressor T cell function
(28). These cells normally limit the immune response against self antigens and help
in down regulating the normal immune response to foreign antigens. These drugs
help circumvent the normal regulatory components of the immune system and allow
strong immune responses from weak irnmunogens.

The route of administration of the immunogen can also significantly affect the
type of immune response seen. Although i.p. injections are most common, other
effective methods include subcutaneous injections, intravenous injections, and
injection into the hind footpads of mice. Numerous Mabs were generated against the
N-terminal portion of hexokinase by i.p injections and subsequent fusion of Spleen
lymphocytes but only one against the C-terminal half of hexokinase. It appears that
when hexokinase is administered i.p. , the immune response is directed against the N-
terminal portion of the molecule. One explanation for the observed result is the
existence of immunodominant epitopes on the N-terminal half of hexokinase.
However, removal of the N -tenninal portion of hexokinase did not transform the C-
terminal portion of hexokinase into a good immunogen as shown by the results in
Figure 2 and the fusion described in Table II,line 9. The basis for the lack of
immunogenicity observed with the C-terminal portion of hexokinase remains obscure.

Four Mabs that inhibit enzyme activity and react with the C-terminal half of

58

hexokinase were produced. All four resulted from fusions with popliteal lymph node
derived lymphocytes from mice injected in the hind footpads. Three of these Mabs
were obtained from mice that had been injected (9) with hexokinase at birth (protocol
6) in hopes of inducing tolerance against immunodominant epitopes. These mice had
also been injected i.p., and their spleen lymphocytes were used in a separate fusion
but did not produce any cell lines secreting antibodies against hexokinase. Thus, the
splenic immune response may have been negatively effected. It is not possible to say
if induction of tolerance altered the immune response but it is clear that use of this
alternative immunization route (hind footpads) , led to the production of Mabs against
the C-terminal half of hexokinase, all of which inhibit enzymatic activity.

It is also important to be able to easily identify those mice likely to produce
the desired response. If possible, serum should be pretested to determine if the nrice
are producing the desired antibodies. Use of an ELISA with antigen adsorbed onto
the microtiter plate provides a fast and easy way to qualitatively or quantitatively
analyze serum samples. However, this ELISA method does not differentiate between
those antibodies recognizing denatured versus native antigen or the class of the
antibodies being produced. Alternative ELISAS, such as ELISA method D, provide a
means to analyze antibody populations for specific characteristics, such as the ability
to react with native antigen. For example, the mouse used in the fusion described in
line 4 of Table I, was one of several mice producing a strong anti-hexokinase immune
response (RBI and RBI-colchicine groups in Figure 1 and Table l) as measured by

ELISA method B, but was chosen because its serum gave the highest response in

59

ELISA method D. If antibodies in western blots are desired then the serum Should be
screened for the presence of antibodies reactive on western blots. These same criteria
should also be used in post-fusion screenings to identify cell lines secreting antibodies
of interest.

It is also advisable to assay for antibody class, since IgG Mabs are generally
of higher affinity and more suitable for applications were the size or lower affinity of
IgM antibodies may be a problem. Since most secondary antibodies used in ELISAS
are not completely class specific, neither ELISA methods B or D give any indication
as to the class of antibody being produced in an immune response. Thus strong
reactivity in these assays does not guarantee that an IgG response is occurring.
ELISA method B can easily be converted to test for antibody class by treating
replicate wells incubated with mouse serum, with rabbit anti-class specific antisera
followed by horseradish peroxidase coupled to goat anti-rabbit IgG.

In conclusion, there is no one correct protocol for producing Mabs with the
desired characteristics. What is appropriate for one protein, may be inappropriate for
another. Until a better understanding of the basis of immunogenicity is achieved, any
effort to generate Mabs will be empirical. However judicious use of screening
procedures can increase the chances of obtaining desired cell lines, and in this

author’s opinion, should be the cornerstone of any effort to generate Mabs.

10.

11.

12.

13.

14.

REFERENCES
Finney, K.G., Messer, J.L., DeWitt, D.L., and Wilson, J.E. (1984) J. Biol.
Chem. 259, 8232-8237.
Wilson, J .E., and Smith, AD. (1985) J. Biol. Chem. 260, 12838-12843.

Harlow, E., and Lane,D. (1988) in Antibodies, A Laboratory Manual, pp. 46-
47, Cold Spring Harbor Laboratory, Cold Sring Harbor, NY.

Wilken, G.P., and Wilson, J.E. (1977) J.Neurochem. 29, 1039-1051.

Fasman, G.D., Ed. (1976) in Handbook of Biochemistry and Molecular
Biology, Vol. II, pp. 454, CRC Press, Cleveland, OH.

Polakis, P.G., and Wilson, J .E. (1984) Arch, Biochem. Biophys. 234, 341-
352.

Wilson, J.E. (1989) Prep. Biochem. 19, 13-21.
Magar, D.A., and Burgess, RR. (1980) Anal. Biochem. 109, 78—86.
Hockfield, S. (1987) Science 237, 67-70.

White, T.K., and Wilson, J.E. (1989) Arch. Biochem. Biophys. 274, 375-
393.

Boss, B. (1984) Brain Research 291, 193-196.

Galfre, G., Howe, S.C., Milstein, C., Butcher, G.W., and Howard, J .C.
(1977) Nature (Lond.) 266, 550-552.

DeWitt, D.L., Day, 1.8., Gauger, J.A., and Smith, W.L. (1982) in Methods
in Enzymology (Hire, C.H.W., and Timasheff, S.N., Eds.), Vol. 86, 299-240.

Littlefield, J. W. (1964) Science 145, 709-710.

15.

16.
17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

61

Harlow, E., and Lane,D. (1988) in Antibodies, A Laboratory Manual, pp.
222-223, Cold Spring Harbor Laboratory, Cold Sring Harbor, NY.

Shek, P.N., and Coons, AH. (1978) J. Exp. Med. 147, 1213-1227.

Shek, P.N., Waltenbaugh, C., Coons, AH. (1978) J. Exp. Med. 147, 1228-
1235.

Mierendorf,Jr., RC, and Dirnond, KL. (1983) Anal. Biochem. 135, 221.

Djavadi-Ohaniance, L., Friguet, B., and Goldberg, M.E. (1984) Biochemistry
23, 97.

Friguet, B., Djavadi-Ohaniance, L., and Goldberg, M.E. (1984) Mol.
Immunol. 21, 673.

Finney, K.G., Messer, J .L., DeWitt, D.L., Wilson, J .E. (1984) J. Biol.
Chem. 259, 8232-8237.

Dunn, S.D., Tozer, R.G., Antczak, DE, and Heppel, LA. (1985) J. Biol.
Chem. 260, 10418.

Vaidya, H.C., Dietzler, D.N., and Iadenson, J.H. (1985) Hybridoma 4, 271.
Kéhler, G., and Milstein, C. (1975) Nature 256, 495 .
Ribi, E. (1984) J. Biol. Response. Modifiers 3, 1-9.

Ribi, E. (1984) in Advances in Carriers and Adjuvants for Veterinary
Biologics, symposium, Ames Iowa, May 7-8, 1984.

Ribi, E., Cantrell, J., and Takayama, K. (1985) Clinical Immunology
Newsletter 6, 33-36.

Matthew, W.D., and Sandrock, A.W. (1987) J. Immunol. Methods 100, 73-
82.

CHAPTER 3

Disposition of Mitochondrially-Bound Hexokinase
at the Membrane Surface, Deduced from Reactivity with Monoclonal

Antibodies Recognizing Epitopes of Defined Location

62

Introduction

Rat brain hexokinase (ATP: D-hexose 6-phosphotransferase, EC 2.7.1.1)
consists of a single polypeptide chain with mol. wt. of approximately 100 kDa (1).
The complete amino acid sequence has been deduced from the cloned cDNA (2).
Extensive similarity in sequence of the N- and C-terminal halves of the enzyme, and
between these and yeast hexokinase (3), supports the view that the mammalian
enzyme is the evolutionary product of duplication and fusion of a gene coding for an
ancestral 50 kDa hexokinase related to the present-day yeast enzyme. The sequence
similarity to yeast hexokinase has been presumed to indicate a corresponding
conservation of secondary and tertiary structural features (4,5), leading to a proposed
structure for the mammalian enzyme (2) based on the x-ray structure of yeast
hexokinase (6).

Using selective proteolysis techniques (7-9), discrete N— and C-terminal halves
of rat brain hexokinase have been isolated with retention of functional properties
including catalytic activity and the ability to bind various ligands. Catalytic function
has been shown to be associated with the C-terminal half of the enzyme (8), while
regulatory function, including the allosteric binding site for the product inhibitor, Glc-
6-P, is associated with the N-terminal half (7,9).

Another function that has been associated with the N-tenninal half of

hexokinase (10,11) is reversible binding to the mitochondrial pore Structure, through

63

which metabolites such as ADP and ATP traverse the outer mitochondrial membrane
(12-14). This is believed to be of significance in regulation of hexokinase activity in
vivo, and to provide a topological basis for interactions between hexokinase and
intramitochondrial ATP-generating processes (15,16). Binding is critically dependent
on an intact hydrophobic sequence at the extreme N -terminus of hexokinase (11),
which is inserted into the hydrophobic core of the membrane when the enzyme is
bound (17). Other aspects of this interaction need to be defined before the molecular
basis for the metabolic relationship between hexokinase and intramitochondrial
processes can be understood.

In the present study, the location of epitopes recognized by several monoclonal
antibodies (Mabs) has been defined within the overall sequence of brain hexokinase,
and related to the proposed structure of the enzyme (2). Based on the effect of these
Mabs on binding of hexokinase to mitochondria, and their ability to recognize the
mitochondrially-bound enzyme, the spatial relationship between bound enzyme and the

mitochondrial surface has been deduced.

MATERIALS AND METHODS

Matsrials.

Enzymes used in plasmid construction and modification (restriction enzymes,
ligases, nucleases) were products of Bethesda Research Laboratories (Gaithersburg,
MD), Boehringer Mannheim (Indianapolis,IN) or New England Biolabs (Beverly,
MA). The pT7-l plasmid was obtained from Boehringer-Mannheim, while the
Sequenase DNA sequencing kit and T7 polymerase were from U.S. Biochemicals
(Cleveland, OH). L-[4,5-3H]-Leucine (Cat. No. TRK.683) and L-[35S]-methionine
(Cat. No. SJ .15 15), used for in vitro translation, were purchased from Amersham
(Arlington, IL). Horseradish peroxidase-conjugated goat anti-mouse IgG was obtained
from Bio-Rad Laboratories (Richmond, CA), rabbit anti-mouse immunoglobulin from
Accurate Chemical and Scientific Corp. (W estbury, NY), and agarose—coupled goat
anti-mouse IgG from Sigma Chemical Co. (St. Louis, MO). The BCA" Protein Assay
Reagent and bovine serum albumin standard were obtained from Pierce Chemical Co.
(Rockford, IL). Microtiter plates (Immuno Plate 11 F) used for ELISA were from

Nunc (Thousand Oaks, CA).

Harem.
Rat brain hexokinase was purified by affinity chromatography, as described
previously (18). Before use, the enzyme was chromatographed on a Sephadex G-25

65

66

(fine) column equilibrated with 50 mM HEPES, 0.5 mM EDTA, 10 mM

monotlrioglycerol, pH 7.5; this is subsequently referred to as HET buffer.
Hexokinase activity was determined by the coupled assay using Glc-6-P

dehydrogenase (18), and hexokinase concentrations determined from A280”, using a

molar extinction coefficient of 5.1 x 104 M'1 cm_l (1).

Monmlgnsl Antibodies.
Several of the Mabs (Mabs 2B, 18, 21, and 3A2) used in the present work

have been described previously (7,10). The others (MabS 4D4, 1C5, and 4C5) were
generated by similar methods; all are of the IgG class. Mabs were precipitated from
serum-containing culture media using ammonium sulfate (19), dialyzed exhaustively
against HET buffer, and stored at -20° C; Mab concentrations were determined with
a radial immunodiffusion assay kit from ICN Biochemicals (Costa Mesa, CA).
Alternatively, mm were produced in serum-free media containing Nutridoma
SP (Boehringer Mannheim) and isolated by affinity chromatography on Protein A-
agarose (Repligen, Cambridge, MA). Collected media was made to 1.5 M glycine and
3 M NaCl, and pH adjusted to 8.9 with NaOH. After adsorption to the Protein A-
agarose column and washing with 1.5 M glycine, 3M NaCl, pH 8.9, Mabs were
eluted with 0.1 M sodium citrate, pH 4.0 (Mabs of the IgG1 or Ing, subclasses) or
3.5 (Mabs of the IgGZb subclass); during elution, monitored by Azsonm, 1 ml fractions
were collected into tubes containing 100 pl 1 M TrisCl, pH 9.0. After exhaustive

dialysis against 0.05 M HEPES, 0.5 mM EDTA, pH 7.5, Mab concentration was

67

determined with the BCA’ Protein Assay Reagent and bovine serum albumin as
standard. The Mab solutions were diluted with an equal volume of glycerol before

storage at -20° C.

Mi h

Rat brain mitochondria were isolated by the procedure of Whittaker (20),
suspended in 0.32 M sucrose at 6-8 units of mitochondrially-bound hexokinase per
m1, and stored at -20° C. Rat liver mitochondria were prepared as described by
Bustamante et al. (21), suspended in 0.25 M sucrose at 20 mg protein per ml
(determined with BCA reagent), and stored at -20° C. Before use, mitochondria were

washed and resuspended in a buffer appropriate for the experiment (see below).

Effsg 9f Mgngglonal Antime s on Binding of Rat Brain ngokinase to
Mimhondn'a.

One pg of hexokinase was incubated with increasing amounts of Mab in 20
mM HEPES, 250 mM sucrose, pH 7.5, containing 3 mg bovine serum albumin per
ml (HSB buffer); total volume was 47 pl. After 30 min at room temperature, 3 pl of
100 mM MgC12 was added, followed by 50 pl of liver mitochondria in HSB buffer to
which 102 mM NaCl had been added. The samples were placed on ice for 30 min,
then centrifuged at 12,000 x g for 7 min. The supematants were removed and pellets

dispersed in HSB buffer containing 60 mM NaCl plus 0.5% (v/v) Triton X-100.

68

Hexokinase activity in both supematants and pellets was determined, with results

expressed as percent of the total activity bound to the pelleted mitochondria.

Mimi gf Nonbindable Hexokinase.
Hexokinase (approx. 0.5 mg/ml) in 0.1 M potassium phosphate, 0.1 M Glc,

0.5 mM EDTA. 0.01 M thioglycerol, pH 7.0, was treated with chymotrypsin
(hexokinase: chymotrypsin, 100: 1) for 1 hr at room temperature, and digestion
terminated by addition of 1 mM PMSF (22). This has previously been shown (1 l) to
result in selective proteolysis of a hydrophobic N -terminal segment critical for binding

of hexokinase to mitochondria.

' as f _ 01L .-. 'n f ulfh u 1 Re idue n Immun ‘1; 'vir f ' -. B '
Beam.

The enzyme was denatured with guanidine hydrochloride and sulfhydryl
groups modified by reaction with 2-bromoacetamido-4-nitrophenol (23), as previously
described (24). Alternatively, sulfhydryl groups of the denatured enzyme were
modified by reaction, under these same conditions, with 5 mM concentrations of
either iodoacetic acid or iodoacetamide. As control, hexokinase was denatured, but
incubated without addition of sulfhydryl modifying reagent. After incubation (4 hr,
37" C.), the enzyme was dialyzed exhaustively against 0.05 M TrisCl, 0.5 mM
EDTA, pH 8.0. The precipitated protein was redissolved by addition of 0.2 % SDS

(w/v), and quantitated with the BCA reagent.

69

Replicate aliquots, each containing 3 pg protein, were subjected to SDS gel
electrophoresis and irnmunoblotting (see below). Replicate blots were probed with
Mabs ZB, 1C5, and 4C5. Immunoreactivity was quantitated by scanning of the blots
using a Visage 110 (Biolrnage, Ann Arbor, MI) image analyzer in the reflective

mode, and expressed as a percentage of that seen with the control enzyme.

ngm’tign 9f Mitochondrially-bound nggkinsse by Monoclgnsl Ang'bodiss.

A fixed amount (0.5 pg) of Mab was incubated with increasing amounts of
hexokinase bound to rat brain mitochondria, suspended in HSB buffer containing 60
mM NaCl; total volume was 400 pl. After 1 hr on ice, the samples were centrifuged
to pellet the mitochondria, and residual Mab in the supernatant determined by ELISA.

An aliquot (50 p1) of the supernatant was diluted to 500 pl with HSB buffer
containing 60 mM NaCl. Aliquots (100 pl aliquots, in triplicate) were added to the
wells of a microtiter plate that had been previously coated with 10 pg/ m1 rabbit anti-
mouse immunoglobulin in 50 mM sodium borate, pH 8.5, followed by blocking with
5 % nonfat dry milk in the same buffer. After 1 hr at room temperature, the wells
were thoroughly washed with 50 mM TrisCl, 154 mM NaCl, 0.05% (v/v) Tween-20,
pH 7.5. A 1:1000 dilution of horseradish peroxidase-conjugated anti-mouse IgG in
HSB buffer containing 60 mM NaCl was added. Following a further 1 hr incubation
and washing as above, all wells received 100 pl of 50 mM sodium citrate, pH 4.0,
containing 0.4 mg/ml o-phenylene diamine and 0.03 % H202. After a suitable period,

color formation was terminated by addition of 100 pl 4 N H28 04. Absorbance at 490

Irv-l

70

nm was determined with a Bio-Tek EL-307 plate reader (Bio-Tek Instruments,
Burlington, VT). Results are given as mean i SD for each set of triplicate

determinations.

ngpmer Mmslling.
Location of epitopes within the proposed Structure for rat brain hexokinase (2)

was done in the Computational Chemistry Facility at The Upjohn Company,

Kalamazoo, MI, using the MOSAIC software package.

h ' Im n 10 ° res.

Unless indicated otherwise, SDS gel electrophoresis on 6-20 % linear gradient
acrylamide gels and immunoblotting were performed as described previously (25)
except for the following modifications: a) electroblotting was done using the carbonate
buffer system suggested by Dunn (26), b) 5 % nonfat dry milk was used in place of
gelatin as a blocking agent on nitrocellulose blots, and c) the blots were developed

with the tetrazolium method of Taketa and Hanada (27).

E . I I . .
Location of epit0pes within the overall sequence of rat brain hexokinase was

done using the approach of Lorenzo er al. (28). The strategy is to determine the

ability of Mabs to immunoprecipitate various truncated versions of the protein, with

deletion of the region containing the epitope reflected by loss of the ability to

71

immunoprecipitate.

An EcoR I-Pst I fragment of the previously described (2) cDNA clone, HKI
1.4-7, was isolated; this fragment includes 91 bp of untranslated 5’ sequence, the
entire coding sequence for rat Type I hexokinase, and 311 bp of 3’ untranslated
sequence. It was directionally cloned into the p'l‘7-1 plasmid such that the 5’ end of
the insert was located just downstream of the T7 RNA polymerase promoter. This
plasmid was designated as HKpT7-l, and a map indicating relevant restriction Sites is
shown in Fig. 1. Linearization of the plasmid with BamH I followed by transcription
with T7 polymerase provided a full length mRNA that coded for full length
hexokinase when translated in the rabbit reticulocyte lysate system (29). Plasmids
containing ordered deletions from the 3’ end of the coding sequence were prepared
according to Henikoff (30), using HKpT7-1 that had been digested with Pst I and an
additional restriction enzyme (Nco I, Sty I, or Bgl I) located near the region of
interest in the insert. Exact location of deletions was determined by DNA sequencing.

Truncated mRNAs, coding for truncated versions of hexokinase, were
generated by transcription of linearized plasmids with T7 polymerase (2 8). HKpT7-l
was linearized with a restriction enzyme cutting within the coding sequence (Fig. 1),
while plasmids containing ordered deletions were linearized by cutting at the Hind 111
Site located just 3’ of the insert. The mRNAs were translated in the rabbit reticulocyte
system, as previously described (29) except that KCl was replaced by 0.05 M
potassium acetate; either 3H-leucine or 35S-methionine was used as label. Translation

products were diluted 10-fold (to 300 pl) in 10 mM TrisCl,. 0.154 M NaCl, 2 mM

Fig. 1. Map of Plasmid HKpT7-l.

72

5920

73

a 5mm ooom
3 3m H com

4 3m
\

‘l
H aoom
{Sm coSaZomcob \ / S of:

 

omm4 ass
4 8m

.Illlv H 662 .
/ 83

        

  

4:41.52:

\ ommm
_
4

H 4mm

74

EDTA, 1% (v/v) NP-40 containing 3 pg of Mab. After incubation overnight at 4° C.,
40 pl of agarose-coupled goat anti—mouse IgG (1 :1 suspension in the dilution buffer)
was added, followed by further 2 hr incubation with gentle agitation at room
temperature. After centrifugation, the pellet was washed 5 times with the above

buffer, and once with 10 mM sodium phosphate, 0.155 M NaCl, pH 7.0.

Irnmunoplecipitated products were dissociated by addition of 100 p1 of the
SDS-containing sample loading buffer (25), and separated by SDS gel electrophoresis

on 15% acrylamide gels. The gel was fixed in 10% methanol-10%

acetic acid for 1 hr, lightly stained with Coomassie blue, then soaked in two changes

of water and in l M sodium salicylate (31), 30 min each time. The gel was dried and

put up for fluorography at -70° C.

Slabs

(from

smll

Hg.

RESULTS

Immungprecipitation of Truncated Versions of Brain Hexokinase.

Results of a typical experiment are shown in Fig. 2. In this case, the ability of
Mabs 1C5, 4C5, and 23 to immunoprecipitate truncated proteins containing the first
(from the N-terminus) 365, 385, or 401 amino acid residues was examined. In each
case, the largest component represents the full-length translation product, while
smaller species represent the result of aborted translation or degradation by proteases
shown to be present in the reticulocyte lysate (results not presented). As shown in
Fig. 2, all three Mabs were able to immunoprecipitate truncated versions of 385 or
401 residues, but could not immunoprecipitate the 365 residue species. Thus, the
epitopes recognized by these antibodies must include some portion of the segment
defined by residues 365-385. Using this method, the epitopes for seven Mabs were
located to segments of 20-50 residues, with results summarized in the second column
of Table I.

The approximate location of epitopes recognized by Mabs 2B, 18, and 21 had
previously been determined by two-dimensional peptide mapping techniques (10,25);
subsequent two dimensional peptide mapping studies (results not Shown) demonstrated
that Mabs 1C5 and 4C5 reacted with the same peptides previously shown (10) to be
reactive with Mab 2B.Location of the epitopes recognized by Mabs 3A2 and 4D4 in a
10 kDa fragment derived from the N—terrninus of the enzyme has also been previously

75

76

Fig. 2. Immunoprecipitation of truncated versions of hexokinase. RNAs coding
for truncated versions of hexokinase were translated in the rabbit reticulocyte lysate
system, and ability of Mabs to immunoprecipitate the translation products determined
(see Methods). Lanes 1, 6, and 11 Show translation products obtained with mRNAs
coding for truncated versions of hexokinase containing 401, 385, and 365 residues,
respectively (as indicated at top of figure); in each case, the largest component is the
full-length translation product, with smaller species resulting from aborted translation
or proteolysis by endogenous proteases. Lane 2 shows the absence of detectable
immunoprecipitation with a control Mab (against Sheep cyclo-oxygenase) which does
not react with hexokinase. Lanes 3-5 Show products immunoprecipitated with Mabs
1C5, 4C5, and 2B, respectively. Lanes 7-10 and Lanes 12-15 are analogous to Lanes
2-5 , and Show the results of immunoprecipitation with the 385 and 365 residue

products, respectively.

77

 

 

P

"r .

o
I

want; do". . _

 

5...: 39cm? use b.

   

 

      

use a... ., 43.4%. . .éu

78

manlmun moaniuam
moulded moflcfiumfl
ouiun umflkuam

and mosoauom

manimon uoflkuom
mouidmd MOSGflMOK

umtun uaflcfluOK

m Ml." MOH—Uflmflm

 

meowumueowwcoo um

no venom

 

:uo cowunufi ween cause

no venom

newummmm

onnsaxoxom sauna new no endeavom on» Rana“:

moe one .moa .mu
flu can «a

«fin

vav

 

nomouflmw uo nofiunooq

79

reported (7,32). The present results are consistent with this previous work, but permit

more refined location of the epitopes.

Fgghsr mfinitign 9f the Epitopss Recogng' ed by These Mabs.

Mab 4D4 was remarkable in its ability to immunoprecipitate truncated versions
of hexokinase as small as 35 residues (Table I). Further definition of the epitope was
obtained by proteolytic modification of this N—terminal region. Treatment of
hexokinase with chymotrypsin has been shown to remove up to nine of the
hydrophobic residues at the N-terminus (11). This treatment completely abolished the
ability of Mab 4D4 to recognize hexokinase, while having no effect on
immunoreactivity with Mabs 3A2 or 21 (Fig. 3). Thus, the first nine residues must
comprise at least some portion of the epitope recognized by Mab 4D4.

In preliminary experiments (results not shown), each of the Mabs used in the
present study was Shown (by both immunoblotting and ELISA techniques) to react
with the isolated N-terminal half of the molecule (7 ,9) but not with the C-terminal
half (8). It thus follows that regions of sequence identity between the two halves (2)
cannot contain the epitopes. The sequences of the N— and C-terminal halves of rat
brain hexokinase (2) are compared in Fig. 4, withregions containing the epitopes
recognized by Mabs other than Mab 4D4 denoted by dashed lines. Residues
highlighted in black are identical in the two halves. Based on this sequence

comparison, significant portions of the potential epitopic regions (dashed lines) must

80

Fig. 3. Reactivity of monoclonal antibodies with native and chymotrypsin-treated
hexokinase. Each panel Shows a replicate blot obtained after SDS-gel electrophoresis
of hexokinase before (Lane 1) and after (Lane 2) digestion with chymotrypsin, which
selectively removes a hydrophobic sequence at the extreme N-terminus of the enzyme
(11). Panel A, blot stained for protein. Panels B-D, irnmunoblots probed with Mabs

4D4, 3A2, and 21, respectively.

1

Awe} r,” e

81

 

.. -so.u- --.—-...._

 

82

be considered as highly unlikely to contain the epitopes. The epitope recognized by
Mab 3A2 was located to the region between residues 36 and 86 (Table I).
Examination of this region, and the corresponding region in the non-irnmunoreactive
C-terminal half, discloses extensive sequence identity in the segment comprised of
residues 61 to 86. Thus, it seems likely that the epitope for Mab 3A2 is located in the
preceding segment, residues 36-60. By similar reasoning, residues 204-209 can be
eliminated as a possible location for the epitopes recognized by Mabs 18 and 21; the
latter must, therefore, be within the segment defined by residues 191-203. Sequence
dissirnilarities in the segment represented by residues 365-385, which contains the
epitopes recognized by Mabs 1C5, 4C5, and ZB, are so extensive that none of this
region can be discounted based on these sequence comparisons.

These additional considerations permit assignment of the epitopes to 10-25
nE‘-Si<1ue segments within the anrino acid sequence (third column in Table I).

The segment defined by residues 365 -3 85 includes (at least portions of)
ePitopes recognized by Mabs 23, 1C5, 4C5. The N-tenninal half of this segment
contains two sulfhydryl groups, at positions 368 and 375. As shown in Fig. 5,
ClEBI‘ivatization of the sufhydryl groups with iodoacetate, iodoacetamide, or 2-
bl'Otnoacetamido-4-nitrophenol had a Significant effect on immunoreactivity with each

01' these antibodies. In each case, reactivity decreased progressively as the substituent
on the sulfhydryls changed from a small uncharged (iodoacetamide) to a small

negatively charged (iodoacetate) moiety and finally to a large aromatic (2-nitrophenol)

83

Fig. 4. Comparison of amino acid sequence of N- and C-terminal halves of rat
brain hexokinase. The sequence is that reported by Schwab and Wilson (2), with
alignment to illustrate the extensive similarity in sequence of the N - and C-terminal
halves. Dashed lines underlie regions which, in the N-terminal half, contain epitopes
recognized by the indicated Mabs, as determined by ability to immunoprecipitate
truncated versions of the enzyme (see text). Within these regions, residues which are
identical in both N- and C-terminal halves are highlighted in black. Residues 36-86,
epitope for Mab 3A2; residues 191-209, epitopes for Mabs 18 and 21; residues 365-

385, epitopes for Mabs 213, res, and 4c5.

N

02

02

02

02

OZ

()2

84

4O 50 60 70

30
i

I l i
1 HIMQLLAYYFTELKDDQVKKIDKYLYAMRLSDEIIIDIL‘Ifll-‘KJWSE nymRTifismLLm
476 HFRLSKQIEMEVKIQLRTEIIEIQ EKETBSKBWSFVEE

80

|
71 MSI-IMSSFRILRVQVNHEKNQN’VSHESEIYDTPENIVHGSGTQLFDHVADCLGDFMEKK
519 true EIGDF - once NFRVLLVKIRSGKKRTVEHHNKIYSIPLEIHQGTCDELFDHIVSCISDFLDYH
190 200 ' 210
I

 

\

I

I I
142 xrxnxxLvar-‘rrsPRCRQsxrnsanrmxarxascvscanvmwrfirfirqacoyamIch
s90 GIKGRRNRLGRTRSRRCHQTNLDCGILIsmxcrxamcecnovasuanfivfiafieerawvmrvc

213 THHTCGYDDQQCEVGLIIGTGTNACYMEELRHIDLVEGDEGRHCINTEWGAFGDDGSLEDIRTEFDRELDR
661 THHTCAYEEPTCEIGLIVGTGTNACYHEEHKNVEHVEGNQCQHCINHEWGAFGDNGCLDDIRTDFDKVVDE

284 GSLNPGKQLFEKHVSGHYMGELVRLILVKMBKEGLLFEGRITPELLTRGKFNTSDVSAIEKDKEGIQNAKE
732 YSLNSGKQRFEKHISGHYLGEIVRNILIDFTKKGFLFRGQISEPLKTRGIFETKFLSQIESDRLBLLQVRA

360 370 380 390

i I I
355 ILTRchapsovEcvs‘mnTerrIlrfiisijngvaarmarLNRLRDNKGT‘RSLRr'rvcvocsuxnnpqrsR
803 ILQQLGLNSTCDESIWKTVEGWAGMMWEKIRENRGLDHLNVTVGVDGTLYKLHPHFSR

.425 RFHKTLRRVVPDSDVRFLLSESGTGKGABMVTAVAIRLAEQHRQIEETLA
l874 IHHQTVKELSPKCTVSFLLSEDGSGKGAALITAVCVRL RG DPSIA

85

Fig. 5. Effect of modification of sulfhydryl residues on immunoreactivity. Rat
brain hexokinase was denatured and sulfhydryl groups modified by reaction with
iodoacetamide (IAm), iodoacetate (IA), or 2-bromoacetamido-4-nitrophenol (BNP).
Immunoreactivity with monoclonal antibodies ZB, 1C5 , and 4C5 was determined on
immunoblots, and expressed relative to reactivity seen with control enzyme
(denatured, but no modification of sulfhydryl groups). Values Shown are means 1: SD

for three determinations, each with a different immunoblot.

 

mumuummn
s\\\\\\\\\\\\\\\\\ E

llllllllllllllllllllll
s\\\\\\\\\\\\\\\\\\\\

llllllllllllllllllllllllllllllllllllllllllllll E
s\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ 5

''''''''''

E 405
I 105
E 23

 

 
 
 

 

IA

 

   

 

 

'IOHLNOO :IO %

87

ring; this pattern is consistent with progressive weakening of antibody-antigen
interactions by electrostatic and steric factors. More importantly, with each of these
reagents, the effects on immunoreactivity decreased in the order 4C5 > 1C5 > 2B.
A reasonable interpretation of these results would be that the epitopes for these Mabs
are ordered within this segment, with that for Mab 2B lying in the N -terminal half
(and thus highly susceptible to effects of modification of sulfhydryls in this region),
the epitope for Mab 4C5 located in the C-terminal half (and thus relatively resistant to

modification of sulfhydryls), and that for Mab 1C5 at an intermediate location.

fM nBindin fHexkin Mi hn'

The effect of these Mabs on the ability of brain hexokinase to bind to rat liver
mitochondria is shown in Fig. 6. Based on these results, the Mabs can be divided into
two groups. As previously reported (10,19), Mabs 2B, 18, and 21 did not prevent
binding, even though, at these same concentrations, they effectively
iIIllIrunoprecipitated the enzyme (results not shown, but see refs. 10 and 19). In
marked contrast, Mabs 3A2, 4D4, 1C5, and 4C5 virtually abolished the interaction

0f hexokinase with the mitochondria.

88

Fig. 6. Effect of Mabs on binding of hexokinase to mitochondria. Hexokinase was
incubated with the indicated amounts of Mab, and the ability of the enzyme to bind to
rat liver was subsequently determined. Results are expressed as a percent of the
amount bound in the absence of antibody. Antibodies used were: 23 ( A ), 21 ( O ),
3A2 ( El ), 4C5 ( O ), and 4D4 ( I ); results (not shown) with Mabs 18 and 1C5

were virtually identical to those with Mabs 21 and 4C5, respectively.

 

 

 

 

 

 

89

 

 

 

 

 

DlJPuoqoouw or punos >iH %

 

[Mob] ug

90

Recoggition Qf Mitochondrially-Bound Hexokinase by Mabs.

The ability of mitochondrially-bound hexokinase to react with these Mabs is
compared in Fig. 7. Mabs 18, 21, and 2B all readily interact with the soluble enzyme
(10,19). Mabs 18 and 21 retain this characteristic with the mitochondrially-bound
hexokinase, but reactivity with Mab 2B is markedly hindered by binding of
hexokinase to mitochondria. Mabs 4D4, 3A2, 1C5, and 4C5 have no Significant
reactivity with mitochondrially-bound hexokinase, though the results in Fig. 6 (and

immunoprecipitation results, not Shown) confirm that these Mabs readily interact with

the soluble form of the enzyme.

91

Fig. 7. Reactivity of Mabs with mitochondrially-bound hexokinase. A fixed
amount of the indicated Mab was incubated with increasing amounts of
mitochondrially-bound (rat brain mitochondria) hexokinase; after removal of the
mitochondria by centrifugation, remaining Mab in the supernatant was determined by
ELISA (see Methods). Results shown are with Mab 4D4 ( I ), 4C5 ( A ), 3A2 ( D ),
2B ( 0 ), and 18 ( O ); results obtained with Mabs 1C5 and 21 were virtually

identical to those shown for Mabs 4C5 and 18, respectively.

92

Amtcsv 68:20me 6766535

 

 

 

 

 

 

 

0% f0 MMO NO :0 0.0
_ — 9 n 0.0
10.00
mm M a .
t H. ll
[I ”H. H ma U
m m 'uIH-r I]. 0.00%

110110103 U; Bululowel qow %

93

DISCUSSION

The overall organization of brain hexokinase into two structurally similar but
functionally distinct halves is clear (2,7-11). The focus in the present situation is the
N -terminal half, which has been directly implicated in binding of the enzyme to
mitochondria (10,11). Based on the results above, the approximate location of the
epitopes recognized by seven Mabs has been defined within the amino acid sequence
of the N—terminal domain of rat brain hexokinase. A schematic representation of these
locations in the context of the proposed structure for this enzyme (2) is shown in Fig.
8, which serves as a useful reference when considering the following remarks.

Based on the extensive sequence similarity with yeast hexokinase, both N- and
C-terminal halves of brain hexokinase are reasonably presumed (4,5) to be similar to
the structure of the yeast enzyme, as determined by x-ray crystallographic studies (6).
That structure consists of two "lobes" (represented by different cross-hatching in Fig.
8). The epitopes recognized by Mabs 18 and 21 lie within the "small lobe" while the
epitopes for the other Mabs are located in the "large lobe" of the N-terminal domain.

All Mabs that block binding of the enzyme to mitochondria (Mabs 4D4, 3A2,
1C5: 4C5) recognize epitopes within the large lobe of the N—terminal domain; Mabs
18 and 21 , with epitopes in the small lobe, do not block binding. A reasonable

interpretation of these results would be that binding is mediated via the large lobe,

94

Fig. 8. Schematic representation of the orientation of hexokinase bound at the
mitochondrial surface. Only the N—terminal domain (involved in binding) is
represented. This domain is divided into " small" (horizontal cross-hatching) and
"large" (diagonal cross-hatching) lobe. It is the large lobe that is apposed to the
membrane when the enzyme is bound, with insertion of a hydrophobic N-terminal
sequence into the lipid core of the membrane (17). Locations of epitapes recognized

by the Mabs used in the present work are represented by blackened rectangles.

96

and that the epitopes recognized by Mabs 4D4, 3A2, 1C5, and 4C5 are located in
regions that come into close proximity with the mitochondrial surface when the
enzyme is bound, with prior binding of the Mab thereby precluding the enzyme-
membrane interaction. In the case of Mab 4D4, this is certainly the case since it
reacts with an epitope in the N—terminal segment that is actually inserted into the
hydrophobic core of the membrane in the course of binding (17). It also follows that
Mabs recognizing epitopes in regions that are in close proximity to the nritochondrial
surface when the enzyme is bound would be prevental from reacting with the
mitochondrially-bound enzyme. Mabs 4D4, 3A2, 1C5, and 4C5 fulfill this prediction
(Fig . 7).

Although Mab 2B recognizes an epitope in close proximity to that of Mabs
1C5 and 4C5, it is distinct from the latter Mabs in that it does not block binding of
the enzyme to mitochondria. This may appear to conflict with the above
interpretation, but a more detailed consideration of the proposed structure of
hexokinase (2) suggests an explanation that is both enlightening and consistent with
exPerimental results. As noted above, the segment defined by residues 365-385
includes, in order (from N-terrninus), the epitopes recognized by Mabs 2B, 1C5 , and
4C5 . Comparison with the structure of the homologous yeast hexokinase (6) indicates
that this region is dominated by an extended surface loop which leads into an a-helix.
The initial portion of the sequence, with which the epit0pe for Mab 2B is associated,
is located on one face of the large lobe. The latter portion of this segment is located

on another face of the large lobe, which also includes the epitopes for Mabs 4D4 and

97

3A2. Thus, location of the epitopes for Mabs 1C5 and 4C5 in the latter portion of this
segment would place epitopes for all Mabs that block binding - and which are also
incapable of recognizing the bound enzyme - on this common face of the large lobe.
Location of the epitope for Mab 2B in the initial portion of the turn, removed from
the face of the lobe that is apposed to the mitochondrial surface, would explain the
inability of Mab ZB to prevent binding while the proximity of this region to the
adjacent mitochondrial membrane would explain the hindered reactivity of Mab 28
with the mitochondrially-bound enzyme.

The above analysis is consistent with disposition of the enzyme on the
mitochondrial surface as depicted schematically in Fig. 8. Although not included in
this figure, the catalytically-active C-terminal domain (8) would extend from the small
lobe of this N -terminal domain (2). Orientation of the C-terminal domain into the
central pore structure would place the enzyme in excellent position for interaction

with intramitochondrial ATP-generating processes (15,16).

l.

2.

10.

11.

12.

13.

14.

REFERENCES

Chou, A.C., and Wilson, J.E. (1972) Arch. Biochem. Biophys. 151, 48-55.

Schwab, D.A., and Wilson, J .E. (1989) Proc. Natl. Acad. Sci. USA 86,
2563-2567.

Stachelek, C., Stachelek, J ., Swan, J ., Botstein, D., and Konigsberg, W.
(1986) Nucleic Acids Res. 14, 945-963.

Creighton, TE. (1983) Proteins, Freeman, New York.

Rossman, M.G., Liljas, A., Branden, CL, and Banaszak, LJ. (1975) in The
Enzymes (Boyer, P.D., Ed.), 3rd ed., Vol. 11, pp. 61-102, Academic Press,
New York.

Harrison, R. (1985) Crystallographic Refinement of Two Isozymes of Yeast
Hexokinase and Relationship of Structure to Function. Ph.D. , thesis, Yale
University, New Haven, CT.

White, T.K., and Wilson, J.E. (1987) Arch. Biochem. Biophys. 259, 402-411.
White, T.K., and Wilson, J.E. (1989) Arch. Biochem. Biophys. 274, 375-393.
White, T.K.,and Wilson, J .E. (1990) Arch. Biochem. Biophys. 277, 26-34.
Wilson, J.E., and Smith, AD. (1985) J. Biol. Chem. 260, 12838-12843.

Polakis, P.G., and Wilson, J.E. (1985) Arch. Biochem. Biophys. 236, 328-
337.

Felgner, P.L., Messer, J .L., and Wilson, J.E. (1979) J. Biol. Chem. 254,
4946-4949.

Linden, M., Gellerfors, P., and Nelson, B.D. (1982) FEBS Lett. 141, 189-
192.

Fiek, C., Benz, R., Roos, N., and Brdiczka, D. (1982) Biochim. Biophys.
Acta 688, 429-440.

98

15.

16.

17.

18.

19.

20.

21.

22.

23.
24.

25.

26.
27.

28.

29.

30.

31

99

Wilson, J.E. (1984) in Regulation of Carbohydrate Metabolism (Beitner, R. ,
Ed.), Vol.1, pp.45-85, CRC Press, Boca Raton, FL.

BeltrandelRio, H., and Wilson, J .E. (1991) Arch. Biochem. Biophys., in press.

Xie, G., and Wilson, J.E. (1988) Arch. Biochem. Biophys. 267, 803-810.
Wilson, J.E. (1989) Prep. Biochem. 19, 13-21.

Finney, K.G., Messer, J.L., De Witt, D.L., and Wilson, J.E. (1984) J. Biol.
Chem. 259, 8232-8237.

Whittaker, V.P. (1969) in Handbook of Neurochemistry (Lajtha, A., Ed.),
Vol. 2, pp. 327-364, Plenum Press, New York.

Bustamante, E, Soper, J.W., and Pedersen, PL. (1977) Anal. Biochem. 80,
401-408.

Alexander, J ., Wilson, 1.3., and Katz, R. (1963) J. Biol. Chem. 238, 741-
744.

Gardner, J.A., and Matthews, KS. (1987) Anal. Biochem. 167, 140-144.
Hutny, J., and Wilson, J.E. (1990) Arch. Biochem. Biophys. 283, 173—183.

Polakis, P.G., and Wilson, J.E. (1984) Arch. Biochem. Biophys. 234, 341-
352.

Dunn, SD. (1986) Anal. Biochem. 157, 144-153.
Taketa, K., and Hanada, T. (1986) J. Immunol. Meth. 95, 71-77.

Lorenzo, F., Jolivet, A., Loosfelt, H., Thu Vu Hai, M., Brailly, S., Perrot-
Applanat, M., and Milgrom, E. (1988) Eur. J. Biochem. 176, 53-60.

Pittler, S.J., Kozak. LR, and Wilson, J.E. (1985) Biochim. Biophys. Acta
843, 186-192.

Henikoff, S. (1984) Gene 28, 351-359.

Chamberlain, LP. (1979) Anal. Biochem. 98, 132-135.

100

32. White, T.K., Kim, J.Y., and Wilson, J.E. (1990) Arch. Biochem. Biophys.
276, 510-517.

101

ACKNOWLEDGEMENTS

We are grateful to The Upjohn Company for allowing us access to their
computer graphics facility, and to Dr. Gerry Maggiora and his colleagues for their
advice and assistance in use of the MOSAIC program. A monoclonal antibody against
sheep cyclo-oxygenase, used as a control in the immunoprecipitation experiments, was
the generous gift of Dr. Willian L. Smith, Dept. of Biochemistry, Michigan State
University.

CHAPTER 4
Effect of Ligand Binding on the Tryptic Digestion Pattern of
Rat Brain Hexokinase: Relationship of Ligand-Induced

Conformational Changes to Catalytic and Regulatory Functions

102

Introduction.

Three isozymes of hexokinase (ATP:D-hexose 6—phosphotransferase, EC
2- 7.1.1) are found in mammalian tissues. While differing in some other respects
(reviewed in ref. 1), the Type I, H, and III isozymes all have a molecular weight of
approximately 100,000 and are allosterically inhibited by the product, Glc-6-P. The
amino acid sequences for the Type I isozyme from rat (2), human (3), and mouse (4),
and for the Type II (5) and Type III (6) isozymes from rat have been deduced from
cloned cDNAs. There is extensive sequence similarity between the N— and C-terminal
halves of these molecules, and between these sequences and that of yeast hexokinase
(7,3), which has a molecular weight of about 50,000. This supports previous
sPeculation (9-14) that the mammalian isozymes evolved by duplication and fusion of
a gene coding for an ancestral 50 kDa hexokinase related to the yeast enzyme.

Following the precept that similarity in sequence implies similarity in structure
(15 , 16), a structure has been proposed (2) for the Type I isozyme based on the x-ray
ClZYStallogr'aphic structure of the yeast enzyme determined by Steitz and his colleagues
07-20). The enzyme is considered to have two domains, each of which resembles the
“ulcture of yeast hexokinase. Other studies have demonstrated that catalytic function
is associated with the C-terminal half of the molecule (21-23), while the allosteric
binding site for Glc-6-P is located in the N—terminal domain (24). Thus regulation by
Glc‘6~P requires propagation of a signal, initiated by binding of Glc-6-P to the N-

103

104

terminal domain, that results in loss of the ability to bind substrate ATP at the
catalytic site in the C-terminal domain (25,26). Extensive conformational changes
have also been shown to result from binding of other ligands to the enzyme (26—30).
Defining the nature of these conformational changes is fundamental to understanding
the structural basis for regulation of hexokinase, an enzyme generally accepted as
playing a major role in governing the rate of cerebral glucose metabolism.

In the present study, the effect of various ligands on the tryptic digestion
pattern of rat brain hexokinase has been investigated. The result has been to focus
attention on a particular region of the C-terminal domain that, judging from its altered
susceptibility to trypsin, is markedly influenced by the binding of several ligands.
Based on analogy with structural transitions shown to occur upon ligand binding to
the yeast enzyme (17), these results provide additional insight into the ligand-induced
conformational changes in this enzyme and the mechanism by which Glc-6-P

inhibition occurs.

MATERIALS AND METHODS

Materials.

Rat brain hexokinase was purified as previously described (31). 1,5-
Anhydroglucitol—6-P was synthesized according to Ferrari et al. (32). Yeast Glc-6-P
dehydrogenase was a product of Boehringer-Mannheim Biochemicals (Indianapolis,
IN). PMSF, TPCK-treated trypsin, and other biochemicals were obtained from

Sigma Chemical Co. (St. Louis, MO).

ex ' e 'vi and rotein determin tion .

Hexokinase activity was determined spectrophotometrically using the Glc-6-P
dehydrogenase-coupled assay (31). Hexokinase and trypsin concentrations were
determined from absorbance at 280 nm, on the basis of a molar extinction coefficient
of 5.1 X 10“M'lcm71 for hexokinase (33), and an A280 m of 1.43 for 1 mg/ml trypsin

solution (34).

' ' 'nfhxkin inhe n fli .
For reasons of stability, purified hexokinase is stored in 0.1 M potassium

phosphate, 0.1 M glucose, 0.5 mM EDTA and 10 mM thioglycerol, pH 7.0 (31).
Prior to use in the present experiments, the enzyme was chromatographed on a
column of Sephadex 625 (fine) equilibrated in 50 mM Hepes containing 0.5 mM

105

106

EDTA and 10 mM thioglycerol, pH 7.5 (HET buffer).

Ten ug of hexokinase in HET, with or without added ligand, was digested
with 2 pg of trypsin in a total volume of 50 it] for 50 minutes at room temperature.
Digestion was terminated by addition of PMSF to 2 mM (35). Where indicated,
ligands were present at the following concentrations: hexoses and analogs were 10
mM; hexose 6-phosphates and analogs were 1 mM except for the extremely potent
inhibitor, 5-thioGlc-6-P (36), which was used at 0.1 mM; NTPS were 10 mM, and
where the Mg’r + chelates were desired, MgClz was also added at a concentration of
12 mM.

Digests were analyzed by SDS-electrophoresis on 6-20 % linear acrylamide
gradient gels as previously described (23,24). Alternatively, analysis of digestion
products under nondenaturing conditions was done by electrophoresis on 2-20 % linear
acrylamide gradient gels using the Tris-borate-DTA system of Lambin and Pine
(37), followed by staining for hexokinase activity as described previously (38). To
determine the tryptic fragments present in catalytically active species, individual bands
were excised, soaked in SDS-sample buffer (0.063 M Tris, pH 6.8, 2 % (w/v) SDS,
4.5 % (v/v) 2-mercaptoethanol, 0.0005 % (w/v) bromophenol blue) for 15 minutes,
then placed in sample wells along with 100 pl SDS-sample buffer for analysis by SDS
gel electrophoresis as above.

Immunoblots were prepared (23,24) using monoclonal antibodies (Mabs) that
recognized epitopes at previously defined locations within the amino acid sequence

(39,40).

RESULTS

Limds Affect the Tmtic Cleavage Pattern of Rat Brain Hexokinase.

Previous studies (39) had demonstrated that, in the presence of relatively high
levels of Glc and Pi, limited digestion of hexokinase with trypsin resulted in
preferential cleavage at two sites, designated T1 and T2 (Fig. 1), giving rise to
cleavage products of 10 kDa, 50 kDa, and 40 kDa, listed in order of their position
from N— to C-terminal in the sequence. Partial cleavage products of 90 kDA and 60
kDa, corresponding to cleavage at only T1 or T2, respectively, were seen as
intermediates. The cleavage sites T1 and T2 are located at Lys 101 and Arg 551,
respectively (2). An alternative digestion pattern has also been detected (24) , with
trypsin cleaving at site T3 to give 52 kDa and 48 kDa fragments corresponding to the
N— and C-terminal halves of the molecule, respectively; further cleavage of the 52
kDa fragment at T1 gave rise to a 42 kDa fragment (Fig. 1). Site T3 was identified as
Arg 462 (2).

As shown in Figs. 2 and 3, these and other digestion products were seen after
limited proteolysis of rat brain hexokinase in the presence of various ligands. More
detailed analysis of these digestion patterns is provided below. For the present, it is
simply noted that these patterns differ in both the number of fragments detected and

in the relative distribution of protein among them.

107

108

Figure 1. Schematic representation of the tryptic cleavage pattern of rat brain
hexokinase. The amino acid sequence of the 100 kDa enzyme is represented linearly
at the top of the figure. The locations of tryptic cleavage sites, T1 through T5, within
the sequence are indicated, with specific residues identified at the bottom of the
figure; tentative identifications (see text) are shown in parenthesis. Various fragments
resulting from cleavage at specific sites are represented as linear segments, with
approximate molecular weights indicated above each. The location of epitopes (39,40)
recognized by monoclonal antibodies 4D4, 3A2, 21, and 5A are shown at the top of
the figure; proteolytic fragments containing these regions were identified by

immunoreactivity with the corresponding antibody.

109

see he .20 9<.~8 e358 93 :5 9s

"H-

esm nae 8e 94.
_ _ _ _

z: 9:

 

 

xov

Xmm
xom

 

xoo

 

Xem

 

v:.m

 

 

xmv

v.3
xmm

_Y__‘(_._V_ Y_

 

v.3

_V

 

 

<m

110

Figure 2. Effect of hexoses on tryptic digestion of hexokinase. Proteolysis
products were analyzed by SDS gel electrophoresis after tryptic digestion of
hexokinase in the absence or presence of ligands; the gel was stained with Coomassie
blue. Lane 1, no ligand; lane 2, Glc; lane 3, Man; lane 4, Gal; lane 5, Fru; lane 6,
GlcNAc; lane 7, 5-thioGlc; lane 8, 2-deoxyGlc; lane 9, 1,5 anhydroglucitol. The
sizes of the major tryptic fragments produced are indicated at the left. The unmarked
band between the 60 and 50 kDa fragments has an M, of approximately 57 kDa and
the band between the 50 and 40 kDa fragments has an M, of approximately 47 kDa.

The band marked T is trypsin.

111

        

 

.3'

«a ‘
w- . .w.

or

 

‘ 3 ‘ 5.“

n r

\vfle

« ._.r
.‘k‘ v. I
.‘b‘ 4
3‘...‘ a

,. .

-:.: ‘-
‘no' u
1"“ v
I

‘-
‘2

‘5'
3"”.

 

- 3:35-
‘-
‘Q
‘

 

112

Figure 3. Effect of hexose-6-phosphates, NTPS and NTP-Mg2+ on tryptic
digestion of hexokinase. Proteolysis products were analyzed by SDS gel
electrophoresis after tryptic digestion of hexokinase in the absence or presence of
ligands; the gel was stained with Coomassie blue. Lane 1, no ligand; lane 2, Glc-6-
P; 1ane 3, 1,5-anhydroglucitol-6-P; , lane 4, 5-thioGlc—6-P; lane 5, 2-deoxyGlc-6-P;
lane 6, Fru-6-P; lane 7, Man-6-P; lane 8, Gal-6-P; lane 9, GlcN-6—P; lane 10, ATP;
lane 11, CTP; lane 12, GTP; lane 13, UTP; lane 14, ATP-Mg“; lane 15, CTP-
Mg“; lane 16, GTP-Mg“; lane 17, UTP-Mg“. The sizes of the major tryptic

fragments are indicated as in Fig. 2; the band marked T is trypsin.

113

 

 

 

 

‘9

Ah

.aO'
.AOD

as

“a”.

 

 

114

Analysis of Tmtic Cleavage Patterns.

Interpretation of tryptic cleavage patterns was facilitated by immunoblotting,
using Mabs 4D4, 3A2, 21, and 5A, which have previously been shown (39,40) to
recognize epitopes at the positions indicated in Fig. 1. This was critical since some of
the cleavage fragments were very similar in molecular weight and hence in close
proximity on SDS gels, but these could readily be distinguished based on their
immunoreactivity. Representative results are shown in Fig. 4. The epitopes for Mabs
4D4 and 3A2 are located at residues 1-9 and 36-60, respectively (40). Thus, both are
present within the 10 kDa N-terminal fragment resulting from cleavage at T1, and
gave identical staining patterns on immunoblots; only the results with Mab 4D4 have

been included here.

Several of the cleavage products (Fig. 4) were as described previously (24,39).

For example, the 60 kDa fragment arising from cleavage at T2 (39) is readily
identified based on its molecular weight and its reactivity with Mabs 4D4 and 21;
further cleavage of this fragment at T1 yields a 50 kDa fragment that retains its
reactivity with Mab 21 but not with Mab 4D4 (see Fig. 1). Similarly, a 52 kDa
fl”figment resulting from cleavage at T3, representing the N-terminal half of the
molecule and reactive with both Mab 4D4 and Mab 21 , was seen - though in
relatively low amounts - in either the absence of ligands or in the presence of Glc

(Fig. 4, Lanes 1 and 2); further cleavage of this fragment at T1 led to production of a
42 kDa fragment reactive with Mab 21 but not with Mab 4D4 (see Fig. 1). The

complementary fragment from cleavage at T3, representing the C-terminal half of the

115

Figure 4. Immunoblotting analysis of tryptic fragments. Hexokinase was digested
in the absence (lane 1) or presence of Glc (lane 2), Glc-6-P (lane 3) or ATP-Mg2+
(lane 4). Replicate blots were probed with Mabs recognizing epitopes of defined
location within the amino acid sequence. The blot shown in the panel at the left was
stained for total protein with amido black. Right panels, irnmunoblots probed with
Mabs 4D4, 21, and 5A, which recognize epitopes located in the 10, 50, and 40 kDa
fragments, respectively, that result from cleavage of the enzyme at sites T1 and T2
(see Fig. 1). Sizes of the major tryptic fragments are indicated at the left; the band

marked T is trypsin.

116

8...... em... 1% ._.....

 

 

vaupcnuwcaupvauw

117

molecule, was a 48 kDa fragment reactive with Mab 5A; although certain ligands
stabilized this fragment (e. g., ATP-Mg++, shown in Fig. 4, Lane 4), further
cleavage at T2 leading to formation of the 40 kDa C-terminal fragment was commonly
observed. Based on the immunoblotting analysis, proteolytic fragments having
molecular weights of 10, 40, 42, 48, 50, 52, 60, and 90 kDa and resulting from
previously defined (24, 39) cleavage sites T1, T2, and T3 were identified (Fig. 1).
Novel proteolytic fragments were also identified in the present work,
corresponding to cleavage at additional sites whose expression was markedly sensitive
to the ligands present during digestion. One of these was a 57 kDa fragment (Fig. 4,
Lanes 1, 2, and 4) reactive with Mabs 4D4 and 21, and thus identified as originating
from the N-terminal half of the molecule after cleavage at a site designated T4;
further cleavage of the 57 kDa fragment at T1 produced a 47 kDa fragment reactive
With Mab 21 but not with Mab 4D4 (Fig. 4). From comparison of the sizes of the
relevant fragments (Fig. 1), it is evident that cleavage site T4 is somewhat beyond the
midpoint in the 89 residue sequence between sites T3 and T2. There is a single Arg
residue, Arg 517, located at this position in the sequence (2), representing the likely
Cleavage site. A complementary fragment with molecular weight of approximately 43
kDa, reactive with Mab 5A and representing the C-terminal portion of the sequence,
shOuld also result from cleavage at T4, but this was never observed, very likely as a
1‘Bsult of rapid cleavage at T2 to give the 40 kDa fragment.
Only in the presence of Glc-6-P and certain analogs (discussed below) was a

59 kDa cleavage product seen. Reactivity with Mab 21 but not with Mab 5A indicated

118

its source as the N-terminal region, while lack of reactivity with Mab 4D4 indicated
that the 10 kDa N-terminal segment had been removed by cleavage at T1; abundant
formation of the 90 kDa fragment demonstrated that cleavage at T1 was facile in the
presence of these ligands. Thus, the 59 kDa fragment was deduced to arise from
cleavage at T1 and a new site, designated T5 (Fig. 1). Based on relative sizes of the
50 kDa and 59 kDa fragments, site T5 was estimated to lie approximately 90 residues
beyond T2 in the sequence. A triplet of basic residues, Lys 642, Arg 643, and Arg
644, seems a likely site although cleavage at Arg 638 is also a possibility; there are
no other basic residues in this region of the sequence (2). A complementary C-
terminal fragment, with molecular weight of about 31 kDa and reactive with Mab 5A,
would also be expected to result from cleavage at T5; this was not seen, presumably
indicating great susceptibility of this fragment to further tryptic attack under the
Present conditions, with resulting formation of fragments undetected by the present
Procedures.

T5 may be identical with a tryptic cleavage site, located in this same region
and called ”Tz" , that was seen in previous experiments in which the C-terminal
domain had been covalently modified with a reactive analog of the substrate, Glc
(22) . In the previous work, a fragment reactive with Mab 5A and corresponding to

the C-terminal region was detected.

119

Effgt of Hexoses on Tryptic Cleavage.

Hexoses (or analogs) which serve as poor substrates (29,36) for brain
hexokinase (Gal, Fru, 5-thioGlc, GlcNAc, and 1,5-anhydroglucitol) had little effect
on the digestion pattern (Fig. 2). In contrast, hexoses that effectively serve as
substrates for this enzyme (Glu, Man, and 2-deoxyGlc) caused marked increase in the
amount of the 90 kDa fragment, indicative of a substantial decrease in the rate of
cleavage at sites lying in the central portion of this segment. Most notable was the
virtually complete abolition of cleavage at T4, as seen from the decrease in the
amount of the 57 kDa fragment and its 47 kDa derivative. These results are consistent
with previous work (28,29) which indicated that the ability of hexoses or their analogs
to serve as substrates for brain hexokinase was correlated with their ability to induce
Specific conformational changes prerequisite for catalysis. The present study
demonstrates that the region around site T4 is involved in this required conformational

Change.

Effect of Hexose 6-Phosphates on Tryptic glgvage.
All hexose 6-phosphates (or analogs) examined provided substantial protection

against tryptic attack in the region of T2-T5, with resulting accumulation of the 90
kDar fragment (Figs. 3 and 5). Although protection in this region was also seen with
Substrate hexoses (results above), the effect with hexose 6—phosphates was markedly
greater. This was readily evident from visual examination of the gels (Figs. 2 and 3),

but was also confirmed by densitometric analysis (results not shown). There was,

120

Figure 5. Effect of hexose 6-phosphates and analogs on the tryptic digestion
pattern of brain hexokinase. Digests were prepared as described in Methods. After
SDS gel electrophoresis, immunoblots were probed with Mabs 21 (Panel A) and 5A
(Panel B), which recognize epitopes in the N- and C-terminal domains, respectively
(see Fig. 1). Molecular weights of major species are indicated at the left. Ligands
present during the digestion were: lane 1, none; lane 2, Glc-6—P; lane 3, 1,5-
anhydroglucitol-G—P; lane 4, 5-thioGlc—6-P; lane 5, Man-6-P; lane 6, Gal-6-P; lane 7,

Fru-6-P; lane 8, 2-deoxyGlc-6-P; lane 9, GlcN-6—P.

121

A
1 2 34 5 6 73 9

’33:”---3-33

‘0’---“ u----

C7D.-

“""— - -—- “-‘;

‘° “-CI- —------

 

122

however, a notable difference in the ability of various hexose 6-phosphates to protect
againstcleavage in the Tz-Ts region, with Fru-6-P and potent inhibitors (29,36) such
as Glc-6-P, 1,5-anhydroglucitol-6—P, or 5-thioGlc-6-P being considerably more
effective than other hexose 6-phosphates examined.

The novel cleavage site, T5, was expressed in the presence of all hexose 6-
phosphates or analogs examined, as evidenced by detection of the 59 kDa fragment,
reactive with Mab 21 but not with Mab 4D4 (Fig. 4), in a closely spaced doublet (less
well resolved in Fig. 5) with the 60 kDa fragment resulting from cleavage at T2.

Another notable difference between the various hexose 6-phosphates was in
their effect on cleavage at T4 to give a 57 kDa fragment or its 47 kDa derivative.
Potent inhibitors (29,36) such as Glc-6-P, 1,5-anhydroglucitol-6-P, or 5-thioGlc-6-P
markedly decreased formation of these cleavage products while digestion in the
presence of poor inhibitors such as Man-6-P, Gal-6-P, 2-deoxyGlc-6—P, and GlcN-6-P
gave amounts comparable to digests done in the absence of added ligands (Fig. 5).
Again, these results are consistent with previous work (28,29) which indicated a
correlation between the ability of various hexose 6-phosphates to induce specific
conformational change in the molecule and their inhibitory effectiveness; it is evident
that this conformational change results in decreased susceptibility of T4 to trypsin.
The sole exception to this correlation is Fru-6-P, which is relatively effective at
protection against cleavage at T4 despite its limited potency as an inhibitor.

Protection against cleavage at sites T2 and T4, as well as induction of cleavage

at T5 - all sites located in the C-temrinal half of the molecule - supplements previous

123

evidence (30) for the extensive conformational effects induced by binding of Glc-6-P
at a site in the N-terminal half (24).

The occurrence of cleavage at T3 was evidenced by detection of the 42 kDa
fragment reactive with Mab 21 (Fig. 5A). The virtual absence of the complementary
48 kDa C-terminal fragment (Fig. 5B) indicated the occurrence of further cleavage at
T2 to give the 40 kDa species reactive with Mab 5A. Thus, after cleavage at T3,
protection by hexose 6-phosphates against cleavage at T2 had been diminished. This
was found to be attributable to disruption of interactions between the N- and C-

terminal domains by cleavage at T3.

Quotation of Discrete N- Ed Q—Terminal Halves by Cleavage at T3.
Detection of both the C-terminal 48 kDa species and the 42/52 kDa fragments

representing the N-terminal half (see Fig. 1) demonstrated the occurrence of cleavage
at T, (Fig. 5). It has previously been shown that the N- and C-terminal halves
resulting from cleavage at T3 can be isolated as discrete species with preservation of
functional properties such as catalytic activity and ligand binding (23 ,24) . The
conditions used in the previous work were, however, quite different from those used
here (specifically, the digests had been done in the presence of low concentrations of
the denaturant, guanidine hydrochloride), and thus it was pertinent to ask if cleavage
at T3 under the present conditions (absence of denaturant) resulted in generation of
independent fragments corresponding to N - and C-terminal domains.

Electrophoresis of tryptic digests under nondenaturing conditions revealed three

124

Figure 6. Detection of enzymatically-active tryptic cleavage products. Hexokinase,
digested in the presence of GlcNAc to stabilize the 48 kDa C-terminal fragment (23),
was analyzed by electrophoresis on nondenaturing acrylamide gradient gels followed
by staining for hexokinse activity, as described in Methods (lane 1); the undigested
enzyme is shown in lane 2. Three bands of hexokinase activity were observed (labeled
1-3 at left of figure) in the digested enzyme; tryptic fragments present in these species
determined after further analysis by SDS gel electrophoresis and immunoblotting (see

text).

125

 

 

 

 

 

126

catalytically active species (Fig. 6). Fragments present in Bands 1-3 were determined
by immunoblotting analysis using Mabs 21 and 5A, as described in Methods (results
not shown).

Band 1 had a mobility indistinguishable from undigested enzyme;
immunoblotting with Mabs 21 and 5 A demonstrated that Band 1 included some intact
100 kDa enzyme but was largely enzyme that had been cut at T1 (i.e., the 90 kDa
fragment was predominant). In contrast, although minor amounts of the 100 kDa and
90 kDa species were also detected in Band 2, the predominant components were the
40 kDa and 50 kDa fragments. Polakis and Wilson (39) had previously shown that the
fragments resulting from cleavage at T1 and T2 remain tightly associated under
nondenaturing conditions, and migrate in this gel system at a rate slightly faster than
the intact enzyme; the increased mobility is presumably due to either additional
charge resulting from cleavage of the peptide bond or to conformational changes
resulting from cleavage and affecting hydrodynamic properties. Thus, Band 2 includes
this previously described species.

Also detected in Band 2 was the 48 kDa fragment, without detectable amounts
of the complementary 52 kDa (or its derivative 42 kDa) fragment. These results thus
indicated that cleavage at T3 did result in dissociation of the N- and C-terminal
halves, with the catalytically-active (23) 48 kDa species migrating at the position of
Band 2, while the catalytically-inactive (23,24) N-terminal fragment migrated
elsewhere on the gel. It was confirmed that the 48 kDa C-terminal fragment, prepared

as described previously (23), did indeed migrate with the mobility of Band 2 (results

127

not shown).

Immunoblotting disclosed that Band 3 contained the 40 kDa C-terminal
fragment; no N—terminal fragment reacting with Mab 21 was detected. Coomassie
blue staining revealed that, in addition to the 40 kDa fragment, a smaller component
of about 8 kDa was also present. Thus, Band 3 represents the 48 kDa C-terminal
fragment which has been further cleaved at T2, with the resulting 8 kDa and 40 kDa
species remaining associated under nondenaturing conditions and migrating as a
catalytically-active species having a mobility slightly greater than that of the uncleaved
(at T2) 48 kDa fragment; the increased mobility of the cleaved species can again be
attributed to increased charge or modified hydrodynamic properties resulting from
cleavage. Thus, these results also indicate that cleavage at T3 resulted in a free
component representing the C-terminal half of the molecule.

It is evident from the above that fragments resulting from cleavage at internal
sites T1 and/ or T2 remain associated by noncovalent interactions. In contrast, cleavage
at T3 results in facile dissociation of resulting fragments. Both of these observations
are consistent with the proposed structure of the enzyme (2) since T1 and T2 occur at
surface loops, which could be severed without disrupting the extensive interactions
occurring in more central regions of the molecule. In contrast, T3 is located in the
segment linking the N- and C-terminal domains (2), and it is not difficult to envisage
that cleavage of this segment might permit separation of the domains, as has been

seen to result from similar cleavage with other proteins (41).

128

Eff f Hexose 6-Pho hates on Proteol is of the C-Termin Dom ' .

As discussed above, binding of hexose 6-phosphates to a site in the N-terminal
domain (24) provided substantial protection against tryptic cleavage in the C-terminal
domain, judging from the increased amount of the 90 kDa fragment seen in digests
prepared in the presence of these ligands. However, cleavage at T3 results in
dissociation of the N- and C—terminal halves (results above), revealing an additional
Glc-6-P binding site that is latent in the C-terminal half of the intact 100 kDa enzyme
(23). Results shown in Fig. 7 demonstrate that binding of Glc—6-P to the newly-
revealed (by cleavage at T3) site did decrease the rate of cleavage at T2, resulting in
transient appearance of the 48 kDa species during the digestion. In contrast, Man-6-P,
a poor inhibitor of both the intact enzyme (29) and the catalytically-active C-terminal
fragment (23) , provided little if any protection, resulting in rapid cleavage at T2 to
yield the 40 kDa fragment with no detectable accumulation of the 48 kDa

intermediate.

Effgj of N'I'Ps ad Their Mg+ + Qhelates on Tryptic Clgvage.

The effect of various NTPS on the tryptic digestion pattern are shown in Fig.
8. In the absence of Mg++ (lanes 2-5), ATP, CTP, GTP, and UTP were virtually
indistinguishable, with prominent increase in the 90 kDa fragment indicative of
substantially decreased cleavage in the T2-T5 region. The whole array of cleavage
products was seen, with no particular site being uniquely sensitive to protection by

these ligands.

129

Figure 7. Time course for digestion of brain hexokinase in the presence of hexose
6-phosphates. Hexokinase was digested with trypsin as described in Methods, except
that aliquots were removed and further cleavage halted by addition of PMSF at times
indicated, in minutes, at the top of the figure. Glc-6—P or Man-6-P were present
during the digestions shown in Panels A and B, respectively. After SDS gel
electrophoresis, immunoblots were prepared and probed with Mab 5 A, detecting

fragments originating from the C-terminal half of the enzyme (see Fig. 1).

130

.-“'lle ll!

'---' 'nl! ------‘ c8
' al."... 1' "IIII'zaOOp

segoumpcwm 0 833899...» 0
. m <

 

 

131

Figure 8. Effect of nucleoside triphosphates and their Mg+ + chelates on the
tryptic digestion pattern of brain hexokinase. Digests were prepared as described
in Methods. After SDS gel electrophoresis, immunoblots were probed with Mabs 21
(Panel A) and 5A (Panel B), which recognize epitopes in the N- and C-terminal
domains, respectively (see Fig. 1). Molecular weights of major species are indicated
at the left. Ligands present during the digestion were: lane 1, none; lane 2, ATP; lane
3, CTP; lane 4, GTP; lane 5, UTP; lane 6, ATP-Mg“: lane 7, CTP-Mg‘”; lane 8,

GT'P-Mg++; lane 9, UTP-Mg‘”.

132

011111? 47>-—

dm n>—--“- -—
si-U

(hillbsl'.

3mm! . B .. .,
1234567819
firmly;

fig”: 129:;

 

133

The presence of the 42 kDa and 5 2 kDa N—terminal species demonstrated that
appreciable cleavage at T3 had occurred but, as with hexose 6-phosphates discussed
above, the virtual absence of the complementary 48 kDa C-terminal fragment
indicated that rapid further cleavage to the 40 kDa species had occurred. Substantial
amounts of the 48 kDa fragment were seen when the corresponding Mg+ + chelates
were present, with ATP-Mg+ + being particularly effective at stabilizing this C-
terminal fragment. In contrast, the chelated ATP markedly destabilized the N -tenninal
region of the molecule, with the 42 kDa and 52 kDa fragments, complementary to the
48 kDa C-terminal fragment, being reduced to virtually undetectable levels; this was
not seen with the chelates of other NTPS. Thus, the discrete N - and C-terminal
halves, produced by cleavage at T3 and each having a nucleotide binding site (23,27) ,
differ greatly in the effect of ATP-Mg+ + on their stability against further tryptic
attack. This difference is also seen by examining the products resulting from cleavage
at T4, where the amounts of the N-terminally—derived 47 kDa and 57 kDa fragments
are markedly reduced in the presence of ATP-Mg+ + while there is no corresponding
decrease in the amount of the 40 kDa C-terminal fragment (Fig. 8). The results
shown in Fig. 8 confirm and extend previous work (23,26,27) which indicated that,
relative to chelates of other NTPS, ATP-Mg+ + had unique effects on the
conformation of this enzyme; it is likely that this reflects the specificity for ATP-

Mg+ + as substrate.

DISCUSSION

The extensive similarity in sequence between the N - and C-terminal halves of
brain hexokinase surely suggests that their structures will also exhibit overall
similarity. However, as noted previously (2), sequence differences such as substantial
insertions/ deletions make it evident that significant structural differences must exist.
These must underlie the observed functional differences between the domains, the C-
terminal region being catalytic (21-23) and the N-terminal domain serving a
regulatory (24) role.

Etch of the cleavage sites in the C-terminal domain has its counterpart in the
N-terminal domain. Using the sequence alignment of Schwab and Wilson (2): T1 (Lys
101) in the N-terminal domain conesponds to Arg 551 (T?) in the C-terminus; Arg 69
in the N-terminus conesponds to Arg 517 (TA) in the C-terminal half; Lys 194, Lys
195, and Arg 196 correspond to Lys 642, Lys 643, and Arg 644 (Ts) in the C-
terminal domain. The failure to detect cleavage products resulting from proteolysis at
the N —terminal counterparts to T4 and T5 implies their lack of structural equivalence
in the two domains. Except in the presence of ATP-Mg“ + , the N-terminal domain is
considerably more resistant to tryptic attack, implying an overall "tighter" structure
than that of the C-terminal domain.

A striking observation from the present work is that all of the effects of ligand
binding on tryptic cleavage were exerted either at the central site, T3, or at sites T2,

134

135

T4, or T5 in the C-terminal domain. This is true whether the ligand binds selectively
to the N-terminal domain as with the hexose 6-phosphates (24), to the C-terminal
domain as with the hexoses (22), or to sites available in both domains as with
nucleoside triphosphates (23,27,30). It is thus evident that a major component of the
conformational changes resulting from binding of various ligands must involve the C-
terminal region. This is perhaps not surprising in a functional sense since the ligands
used are either themselves substrates or regulatory effectors of the enzyme and hence
must have an effect on catalytic function associated with the C-terminal domain (23).
Nonetheless, it seems notable to us that a ligand such as Glc-6-P, which protects at
cleavage site T2 when bound to the isolated C-terminal domain, does not have any
significant effect on cleavage at the N -terminal counterpart, T1, when bound to the N-
terminal domain of the intact 100 kDa enzyme. It is evident that fusion of the
domains has resulted in appreciable modification of the conformational response to the
binding of Glc-6-P. Another example is the response to binding of Pi, which is a
relatively potent inhibitor of the isolated catalytic C-terminal domain (23) but binds to
the intact enzyme without effect on catalytic activity and, in fact, effectively serves as
an "activator" of the enzyme by antagonizing inhibition by Glc-6-P; the likely
physiological importance of this remarkable alteration of the response to P, has been
noted earlier (23).

As discussed previously (2), there are undoubtedly significant structural
differences between the N- and C-terminal halves of brain hexokinase, and between

these and the homologous enzyme of yeast. Nonetheless, the structure proposed (2) on

136

the basis of the x—ray structure of yeast hexokinase (1 1-14) has proven useful in
interpreting various experimental findings with the brain enzyme, e.g. , the model
predicts the location of T1 and T2 at surface loops, consistent with the facile tryptic
cleavage at these positions (in the absence of protective ligands). We believe that
analysis of the present results in terms of previous work with the yeast enzyme is also
useful.

The overall structure of yeast hexokinase is divided into two major regions,
generally referred to as the large and small " lobes " , divided by a deep cleft (17). The
substrate Glc is bound at a site located deep within the cleft, causing a conformational
change that results in closing of the cleft. All of the residues thought to be involved in
binding of Glc to the yeast enzyme (17) are conserved in the catalytic C-terminal
domain of mammalian hexokinases (2-6), and the recent site-directed mutagenesis
study of Arora et al. (42) has demonstrated their importance for catalysis. It is thus
reasonable to expect that binding of substrate hexoses to the catalytic domain of brain
hexokinase occurs in a similar manner, and with similar conformational response. X-
ray studies (17) have indicated that the structural changes associated with cleft closure
are largely restricted to the small lobe and hinge region between the lobes.

Although the binding site for ATP in the yeast enzyme has not been so well
defined, there is considerable evidence to support the view that it is associated with a
B-sheet structure within the small lobe (43). Based on comparison with characteristics
of ATP-binding sites in other enzymes, the corresponding region in the brain enzyme

had previously been proposed (44) as the ATP binding site; several residues

137

postulated to be important in the binding of ATP and catalytic function (44) have
since been shown to be highly conserved in all mammalian hexokinases (2-6). Thus, a
reasonable view of the catalytic events as they occur on yeast hexokinase and the
related catalytic C—terminal domain of mammalian hexokinases would be that binding
of Glc within the cleft induces a conformational change ( ” cleft closure") (17) that
brings ATP, bound to the small lobe, into proximity with the 6-hydroxy1 of the
hexose and with other residues likely to be of catalytic importance, e. g. , residues
corresponding to Asp 211 in yeast hexokinase (42, 44).

Cleavage site T4 is most likely at Arg 517 in brain hexokinase, corresponding
to Met 71 in the yeast enzyme (2). Met 71 is located in the hinge region connecting
the large and small lobes (l7) and this region obviously must be influenced by closing
of the cleft. Thus, the reduced cleavage at T4, seen in the presence of hexoses which
serve as good substrates for brain hexokinase but not with other hexoses, can be
attributed to a substrate-induced closing of the cleft in the C-terminal domain of brain
hexokinase. The correlation between the ability to serve as substrate and the ability to
induce specific conformational change has been noted previously (29); the present
work, viewed in the context of studies with the yeast enzyme (17), leads to a more
specific suggestion that the required conformational change is cleft closure in the C-
terminal domain.

T2 (Arg 551 in brain hexokinase, corresponding to Arg 103 in the yeast
enzyme) is predicted to be located near a surface turn connecting the first two B-

strands of the 5-stranded B-sheet involved in binding of ATP (42,43). As noted above,

138

cleft closure induced by binding of substrate hexose (17), results in movement of this
region into the cleft, bringing the bound nucleotide into proximity with the hexose and
catalytic residues. More restricted access to T2 would be an expected result of this
compaction of structure, consistent with the observed decreased cleavage at T2 seen in
the presence of hexoses (or analogs) inducing cleft closure.

Binding of Glc-6-P or other inhibitory hexose 6-phosphates (or, anomalously,
Fru-6-P) also results in reduced cleavage at T2 and T4, consistent with the view that
binding of the inhibitory ligand to the N-terminal site induces conformational changes
in the C-terminal domain that are similar to those induced by binding of substrate
hexoses. The theoretical analysis of Weber (45) makes it clear that dissimilar ligands
that induce similar conformational changes should exhibit mutual enhancement in their
binding, consistent with the observed synergistic binding of hexose and hexose 6-
phosphates (29,33).

The protection afforded to the C-terminal domain by binding of inhibitory
hexose 6-phosphates to the N-tenninal domain is even greater than that resulting from
direct binding of substrate hexoses to the C-terminal domain, suggesting that cleft
closure might be more complete in the presence of the hexose 6-phosphates.
Restricted access to the ATP binding site could account for the preclusion of ATP
binding (25 ,26,46) that underlies the inhibitory action of these compounds. The
protection by the weak inhibitor, Fru-6-P, is an obvious anomaly and suggests
additional factors may be involved. Previous work (29), using alternative methods for

detecting ligand-induced conformational change, has demonstrated that Fru—6-P is not

139

equivalent to inhibitory hexose 6-phosphates in its action on the enzyme.

In contrast to the substrate hexoses, hexose 6-phosphates also induce a
conformational response that results in expression of a new cleavage site, T5. This
site conesponds to the region of Lys 642, Arg 643, and Arg 644 in brain hexokinase,
with the equivalent being Ser 197, Lys 198, and Arg 199 in the yeast enzyme. These
residues are located (17) in a long a-helix that runs across the small lobe, adjacent to
the B-sheet involved in binding of substrate ATP. The increased cleavage at T5
implies a structural alteration in this ATP-binding region which may contribute to the
loss of ability to bind the nucleotide substrate (25,26,46).

The present lack of a detailed structure for brain hexokinase precludes a
detailed analysis of structural transitions that accompany binding of various ligands.
Moreover, the present work as well as previous studies have clearly demonstrated that
the conformational response to binding of physiologically relevant ligands is complex.
This multiplicity of functionally important conformations will certainly complicate
detailed analysis by x-ray or NMR methods, should they become feasible. Despite its
limitations, we feel that comparison with the structure and conformational transitions
of the homologous yeast hexokinase offers a useful approach toward understanding the

conformational changes that underlie regulation of the mammalian hexokinases.

10.

ll.

12.

13.

REFERENCES
Wilson,J.E. (1984) in Regulation of Carbohydrate Metabolism (Bietner,
R.,Ed.), Vol. 1, pp.45-85, CRC Press, Boca Raton, FL.

Schwab, D.A., and Wilson, J .E.(1989) Proc. Natl. Acad. Sci. USA 86, 2563-
2567.

Nishi, S., Seino, S., and Bell. 6.1. (1988) Biochem. Biophys. Res. Commun.
157, 937-943.

Arora, K.K., Fanciulli, M., and Pedersen, PL. (1990) J. Biol. Chem. 265,
6481-6488.

Thelen, A. P., and Wilson, J .E. (1991) Arch. Biochem. Biophys. 286, 645-
651.

Schwab, D.A., and Wilson, J .E. (1991) Arch. Biochem. Biophys. 285, 365-
370.

Kopetzki, E., Entian, K.-D., and Mecke, D. (1985) Gene 39, 95-102.

Stachelek, C., Stachelek, J., Swan, D., and Konigsberg, W. (1986) Nucleic
Acids Res. 14, 945-963.

Colowick, S.P. (1973) in The Enzymes (Boyer, P.D., Ed.), 3rd ed., Vol. 9,
pp.1-48, Acedemic Press, New York.

Easterby, 1.8., and O’Brien, M.J. (1973) Eur. J. Biochem, 38, 201-211.

Rose, I.A., Warms, J.V.B.,and Kosow, DP. (1974) Arch. Biochem. Biophys.
164, 729-735.

Holroyde, M.J., and Trayer, LR, and Cornish-Bowden, A. (1976) FEBS Lett.
62, 213-219.

Ureta, T. (1982) Comp. Biochem. Physiol. B: Comp. Biochem. 71, 549-555.

140

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

141

Gregoriou, M., Trayer, LR, and Cornish-Bowden, A. (1983) Eur. J.
Biochem. 134,283-288.

Rossmann, M.G., Liljas, A., Branden, C.I., and Banaszak, L.J. (1975) in The
Enzymes (Boyer, P.D., Ed.), 3rd ed., Vol. 11, pp. 61-102, Academic Press,
New York.

Creighton, TE. (1983) Proteins, pp. 252-262. Freeman Publications, New
York.

Harrison, R. (1985) Ph.D. Thesis, Yale University, New Haven, CT.

Anderson, C.M., Stenkamp, RE, and Steitz, T.A. (1978) J. Mol. Biol. 123,
15-33.

Anderson, C.M., Stenkamp, R.E., McDonald,R.C., and Steitz, T.A. (1978)
J. Mol. Biol. 123, 207-219.

Bennett, W.S., Jr., and Steitz, T.A. (1980) J. Mol. Biol. 140, 211-230.

Nemat-Gorgani, M., and Wilson, J.E. (1986) Arch. Biochem. Biophys. 251,
97-103.

Schirch, D.M., and Wilson,J.E. (1987) Arch. Biochem. Biophys. 254, 385-
396.

White, T.K., and Wilson, J.E. (1989) Arch. Biochem. Biophys. 274, 375-393.
White, T.K., and Wilson, J .E. (1987) Arch. Biochem. Biophys. 259, 402-411.

Ellison, W.R., Lueck, J.D., and Fromm, H]. (1974) Biochem. Biophys.
Res. Commun. 57, 1214-1220.

Baijal, M., and Wilson, J.E. (1982) Arch. Biochem. Biophys. 218, 513-524.
White, T.K., and Wilson, J.E. (1990) Arch. Biochem. Biophys. 277, 26-34.
Wilson, J.E. (1978) Arch. Biochem. Biophys. 185, 88-99.
Wilson, J.E. (1979) Arch. Biochem. Biophys. 196, 79-87.

Hutny,J., and Wilson, J.E. (1990) Arch. Biochem. Biophys. 173-183.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

142

Wilson, J.E. (1989) Prep. Biochem. 19, 13-21.

Ferrari,R.A., Mandelstam, P., and Crane, R.K. (1959) Arch. Biochem.
Biophys. 80, 372-377.

Chou, A.C., and Wilson, J.E. (1972) Arch. Biochem. Biophys. 151, 628-633.

Worthington, C.C., Ed. (1988) Worthington Manual, p.320, Worthington
Biochemicals Corp. , Freehold, NJ.

Alexander, J., Wilson, LB, and Katz, R. (1963) J. Biol. Chem. 238, 741-
744.

Wilson, J .E., and Chung, V. (1989) Arch. Biochem. Biophys. 269, 517—525.
Lambin, P., and Fine, J.M. (1979) Anal. Biochem. 98, 160-168.
Needels, D.L., and Wilson, J.E. (1983) J. Neurochem. 40, 1134-1143.

Polakis, P.G., and Wilson, J.E. (1984) Arch. Biochem. Biophys. 234, 341-
352.

Smith, A.D., and Wilson, J .E. (1991) Arch. Biochem. Biophys., in press.

Wilson, J.E. (1991) in Methods of Biochemical Analysis (Suelter, C.H., Ed.),
Vol. 35, pp. 207-250, Wiley-Interscience, New York.

Arora, K.K., Filbum, C.R., and Pedersen, PL. (1991) J. Biol. Chem. 266,
5359-5362.

Arora, K.K., Shenbagamurthi, P., Fanciulli, M., and Pedersen, PL. (1990) J.
Biol. Chem. 265, 5324-5328.

Schwab, D.A., and Wilson, J.E. (1988) J. Biol. Chem. 263, 3220-3224.
Weber, G. (1972) Biochemistry 11, 864-883.

Ellison, W.R., Lueck, J.D., and Fromm, H]. (1974)
Biochem. Biophys. Res. Commun. 57, 1214-1220.

CHAPTER 5
Epitopic Regions Recognized by Monoclonal Antibodies
against Rat Brain Hexokinase:

Association with Catalytic and Regulatory Function

143

Introduction

There are three isozymes of hexokinase (ATP:D-hexose 6-phosphotransferase,
EC 2.7.1.1) found in mammalian tissues, all with M, of approximately 100,000
(reviewed in ref. 1). Pronounced similarities between sequences of the N- and C-
terminal halves of the Type I (2-4), 11 (5), and III (6) isozymes isolated from various
sources, and between these and the sequence of the 50 kDa hexokinase of yeast (7,8),
support the suggestion (9-14) that the 100 kDa mammalian enzymes have evolved by
duplication and fusion of a gene coding for an ancestral hexokinase related to the
present-day yeast enzyme.

The Type I isozyme isolated from rat brain has been studied most thoroughly.
The extensive sequence similarities (2) make it reasonable to expect (15 , 16) that the
structures of the N - and C-terminal halves of the enzyme are similar to each other and
to that of yeast hexokinase, which has been determined by x-ray crystallography (17-
20); on this basis, a structure for the enzyme has been proposed (2). Although the N-
and C-terminal halves may be structurally similar, they are functionally distinct. The
C-terminal half serves the catalytic function (21—23), while regulation of the enzyme
results from binding of the product inhibitor, Glc-6-P, to an allosteric site in the N-
terminal half (24). The N-terminal half is also involved in binding of the enzyme to

144

145

the outer mitochondrial membrane (25). Under appropriate conditions, the N- and C-
tenninal halves of the enzyme can be cleaved by trypsin and isolated with retention of
functions such as catalysis and ligand binding (23,24,26). Studies with the N— and C-
tenninal fragments demonstrated that both possess binding sites for the substrates, Glc
and ATP, as well as for Glc-6—P. In the intact 100 kDa enzyme, the binding site for
Glc in the N-terminal half of the molecule and the binding site for Glc-6-P in the C-
terminal half are latent.

Binding of various ligands to rat brain hexokinase has been shown to induce
distinctive conformational changes which have been related to function (23,24,26-30).
For example, the ability of various hexoses or nucleoside triphosphates to induce
specific conformational changes has been correlated with their ability to serve as
substrates for the enzyme (23,27-29). A similar correlation between induced
conformational change and inhibitory effectiveness of various hexose 6-phosphates or
their analogs has been noted (23,24,26-29). Indeed, the functional organization of the
enzyme makes it evident that extensive conformational change must be involved in its
regulation, with the inhibitory signal resulting from binding of Glc-6-P to the N-
terminal half being propagated to the catalytic C-terminal half.

Monoclonal antibodies (Mabs) represent useful probes for establishing
structure-function correlations in proteins (25,31). In the present study, epitopes
recognized by a collection of 23 Mabs against rat brain hexokinase have been mapped
within the structure of the enzyme. By analysis of the effects of these Mabs on

catalytic and regulatory function and the effects of various ligands on conformational

146

epitopes recognized by the Mabs, additional insight into structure-function

relationships within brain hexokinase has been obtained.

MATERIALS AND METHODS

Materials.

Nunc 96-well microtiter plates (catalog number 269620 or 442404) for use in
ELISAS were obtained from Thomas Scientific (Swedesboro, NJ). Glc-6-P
dehydrogenase was a product of Boehringer-Mannheim Biochemicals (Indianapolis,
IN). Streptavidin-biotinylated horseradish peroxidase complex was purchased from
Amersham Corp. (Arlington Heights, IL). Horseradish peroxidase-conjugated goat
anti-rabbit IgG and Tween-20 were obtained from Biorad Laboratories (Richmond,
CA). Rabbit anti-mouse IgG, TPCK-treated trypsin, soy bean trypsin inhibitor-
conjugated agarose, PMSF, o-phenylenediamine, and other biochemicals were
purchased from Sigma Chemical Co. (St. Louis, MO). NHS-LC-biotin, BCA‘ Protein
Assay Reagent and bovine serum albumin standard were products of Pierce Chemical
Co. (Rockford, IL). S. aureus cells were obtained from The Enzyme Center (Malden,
MA) and prepared as described by Finney et al. (32). 1,5-Anhydroglucitol-6-
phosphate had been synthesized according to Ferrari et al. (33).

Hexokinase was prepared from rat brain by affinity chromatography (34).
Rabbit polyclonal antiserum against rat brain hexokinase was produced as described
previously (35), and antibodies against hexokinase purified by affinity chromatography
(36).

147

148

Monoclonfl Antimdies.

Several of the Mabs used in this study have been described previously
(25,32,37). Others were generated by similar procedures. Briefly, lymphocytes from
spleens or popliteal lymph nodes of Swiss white mice that had been immunized with
rat brain hexokinase were fused with Sp2/O-Ag l4 mouse myeloma cells, using
polyethylene glycol as fusative agent (38). Hybridomas secreting antibodies against rat
brain hexokinase were detected by ELISA methods (39) and cloned by limiting
dilution. Mabs were isolated from culture media, and concentrations determined, as
described previously (25).

Class and subclass were determined by ELISA using affinity purified subclass
specific antisera from ICN Biomedicals (Irvine, CA). Of the 23 Mabs used in the
present study, two (Mabs lDl and 2B6) were of the Inga subclass, three (Mabs 3A2,
13, and 1B2) were of the IgG2b subclass, and the others were IgGls.

All Mabs used in this work were shown to react with native hexokinase (39)
and hence recognize epitopes located on the surface of the folded structure.

For use in competitive ELISA experiments, Mabs were derivatized with NHS-
LC-biotin. Mabs purified by Protein A affinity chromatography (25) were dialyzed
against 10 mM sodium phosphate, 154 mM NaCl, pH 7.5. NHS-LC-biotin, dissolved
in dimethylsulfoxide, was added to give an NHS-LC-biotin: Mab molar ratio of
100:1. After 2 hr at room temperature, Mabs were dialyzed extensively against 50
mM Hepes, 0.5 mM .TA, pH 7.5. Alternatively, purified Mabs were dialyzed

against 50 mM Hepes, 0.5 mM EDTA, pH 7.5, biotinylated by addition of NHS-LC-

149

biotin, and used directly. Biotinylated Mab concentrations were determined with the

BCA‘ Protein Assay.

Hexokinase Activity and Inhibition Studies.
Hexokinase activity was measured spectrophotometrically using a Glc-6-P

dehydrogenase-coupled assay (34); one unit is defined as the amount of enzyme
catalyzing the formation of 1 pmole of G1c-6-P per minute.

Determination of KIn values for Glc and ATP was done in the same assay
system with the nonvaried substrate at saturating concentration (3.6 mM for Glc and
6.6 mM for ATP). Kinetic data were analyzed using the EZ-Fit program (40).
Inhibition by the Glc-6-P analog, 1,5-anhydroglucitol-6-P, was studied with 3.6 mM
Glc present; ATP was 0.55 mM when using intact hexokinase or 0.7 mM for
experiments with the isolated C-terminal fragment (23). The concentration of the

inhibitor was varied, and the K,- estimated from a Dixon plot.

in e ' ’on .
Hexokinase and trypsin concentrations were determined from the absorbance at
280 nm, using a molar extinction coefficient of 5.1 X 104 M'lcm'l for hexokinase
(41), and an A280 m of 1.43 for a 1 mg/ml trypsin solution (42). Mab concentrations

were determined as described previously (25).

150

Immunoprecipigtign of Hexokinase.
Immunoprecipitation was performed according to Finney et al. (32), with

rabbit anti-mouse IgG-coated S. aureus cells used to precipitate hexokinase-Mab

complexes.

Preparation of isolated N- and C-terrninal halves of hexokinase.
For reasons of stability, purified hexokinase is stored in 0.1 M potassium

phosphate, 0.1 M glucose, 0.5 mM EDTA and 10 mM thioglycerol, pH 7.0 (34).
Prior to use, the enzyme was chromatographed on a column of Sephadex G25 (fine)
equilibrated with 50 mM Hepes containing 0.5 mM EDTA and 10 mM thioglycerol,
pH 7.5 (HET buffer). The N- and C-terminal fragments of hexokinase were prepared
as described by White and Wilson (23), with the following modification. When
preparing the N -terminal fragment, PMSF was initially omitted and the samme
immediately dialyzed for 1 hour to reduce the concentration of guanidine
hydrochloride. The sample was then incubated with STI-agarose, using 7 pl of
packed gel per pg of trypsin, for 15 minutes and centrifuged to pellet the agarose.
PMSF was then added to 2 mM and the sample exhaustively dialyzed.

Under these conditions, the isolated C-terminal fragment retains full catalytic
activity and its concentration was estimated from activity measurements, using a
specific activity of 110 U/ mg (23). The concentration of the isolated N-terminal
fragment, which is not catalytically active, was determined using the Pierce BCA‘

Protein Assay with bovine serum albumin as standard and a modified protocol (43)

151

which avoids interference from endogenous thiols.

In vi Q synthesis of a truncated version of rat brain hexokinase.

A 453-residue segment was synthesized in vitro, and the ability of various
Mabs to immunoprecipitate the truncated product was assessed according to

previously described procedures (25).

Epitgm mapping using a comrgtitive ELISA.
The ability of pairs of Mabs to simultaneously bind hexokinase was assessed

using a modified ELISA (39). Hexokinase (0.32 pg) in HET buffer containing 3

mg/ ml BSA was first incubated for 30 min at room temperature with a 20-fold molar
excess of Mab, also in HET buffer; total volume was 400 pl. A second Mab (2-5 p1
in HET), which had been labeled with biotin, was then added at 5-fold molar excess,
and the incubation continued for an additional 30 minutes. Aliquots (0.1 ml, in
triplicate) were added to wells of a microtiter plate that had been coated by overnight
incubation (room temperature, humid atmosphere) with 5 pg/ ml affinity purified
rabbit polyclonal anti-hexokinase antibodies in 50 mM sodium borate, pH 8.5 ,
followed by blocking with 5 96 nonfat milk in the same buffer. After 1 hr at room
temperature, the wells were thoroughly washed with 50 mM TrisCl, pH 7 .5,
containing 154 mM NaCl and 0.05% (v/v) Tween 20 ('I'I'BS). A 1:1250 dilution of
streptavidin-biotinylated horseradish peroxidase complex was then added (100

pllwell). Following a 1 hr incubation and washing as above, all wells received 100

152

pl of 50 mM sodium citrate, pH 4.0, containing 0.4 mg/ml o-phenylenediamine and
0.03 % H202. After a suitable period, color formation was terminated by addition of
100 pl 4N HZSO4. Absorbance at 490 nm was determined with a Bio-Tek EL-307
plate reader (Bio-Tek Instruments,Burlington, VT). Results were compared with a

control to which only the biotinylated antibody was added.

Eff fli ds nth inte tin bewithhexkin de min b ELIA.
Microtiter plates were prepared by placing 0.1 ml of a solution containing 2
pg Mab per ml in 50 mM sodium borate, pH 8.5, into the wells, followed by
incubation overnight at room temperature in a humid chamber. One Mab, designated
4A6, was found to exhibit exceptional reactivity when intact enzyme was used as
antigen in ELISA, and the concentration of Mab 4A6 used in microtiter plate
preparation was reduced to 0.1 pg per ml in this case; the affinity of Mab 4A6 for the
isolated C-terminal fragment was significantly reduced (see below), and plates were
prepared in the usual fashion when reaction with the C-terminal fragment was to be
measured. After overnight incubation, plates were washed with TTBS and then
blocked by incubating for one hr with 200 pl per well of 5 % nonfat milk in 50 mM
sodium borate, pH 8.5. Wells, in triplicate, then received 0.1 ml of a solution
containing intact hexokinase or isolated N- or C-terminal fragment, dissolved in HET
containing 3 mg/ml BSA, 136 mM NaCl, and indicated ligands; when NTPS or
MgC12 were present, the NaCl was reduced to maintain equivalent ionic strength. To

ensure adequate color production in the ELISA, the concentration of antigen used

153

varied with different Mabs. The N-terminal fragment was used at 1.0-2.0 pg/ml, the
C-terminal fragment at 0.2-0.5 pg/ml and intact hexokinase at 0.1-1.0 pg/ml. Where
indicated, ligands were added at the following concentrations: Glc and Gal were 10
mM; hexose 6-phosphates or analogs were 1 mM, except for the extremely potent
inhibitor, 5-thioGlc-6-P (44), which was 0.1 mM. Nucleotides were 10 mM, and
MgClz was included to 12 mM when the magnesium chelates of the nucleotides were
desired. After a 45 min incubation, the wells were washed with TTBS and incubated
with 0.1 ml of affinity purified rabbit anti-hexokinase antibodies at 1.0 pg/ml in HET
containing 3 mg/ml BSA. After 30 minutes wells were washed with TTBS and
incubated with 0.1 ml of a 1:1000 dilution of horseradish peroxidase-conjugated goat
anti-rabbit antibodies. After a final 45 minute incubation and wash with TTBS,
adsorbed peroxidase activity was detected as above. Results were expressed relative
to color formation in the sample containing no added ligand. Standard deviation for
the three replicate samples was routinely less than 5 % of the mean.

For several of these Mabs, the effects of ligands on immunoreactivity, as
assessed by ELISA, were confirmed by demonstrating comparable effects on
immunoprecipitation. However, the latter method was obviously more cumbersome -
particularly if effects on the catalytically inactive N-terminal domain were being

assessed - and thus ELISA was routinely the method of choice.

154

Effect of Mabs on the tmtic cleavage pattern of hexokinase.
Three pg of hexokinase in HET buffer was incubated with a 3-fold molar

excess of Mab for 30 minutes at room temperature; with Mab 5A3, a 5-fold excess
was used and 10 mM Glc was included to enhance binding of the Mab (see below).
Trypsin (0.6 pg) was then added. After 50 minutes, PMSF was added to 2 mM.
The entire samples were then analyzed by SDS-polyacrylamide gel electrophoresis as

previously described (45 ).

Mglgular mxeling.
Molecular modeling was done on a Silicon Graphics Personal Iris Model

4D/35 using the BioSym software package Insight 11.

RESULTS

Epitopp Mapping.
Several of the Mabs used in the present study have previously been shown to

recognize continuous epitopes lying within defined segments of the sequence for rat
brain hexokinase (25). The epitope recognized by Mab 4D4 is located within residues
1-9, and that for Mab 3A2 within residues 36-60. Mabs l8 and 21 recognize epitopes
in the region of residues 191-203, while the epitopes for Mabs 2B, 1C5, and 4C5 are
located at residues 365-3 85. These established epitopic locations served as reference
points for additional epitope mapping studies.

Representative results of a competitive ELISA epitope mapping experiment are
shown in Fig. 1. In this case, the ability of various Mabs to compete with biotinylated
Mab 2B for binding to intact brain hexokinase was measured. Mab 2B, as expected,
effectively competes with its biotinylated analog for binding to the enzyme, resulting
in almost total abolition of color formation. Three other antibodies, Mabs 1C5, 4C5 ,
and 4C4, also markedly inhibit binding of Mab 2B, and this was also seen in
reciprocal experiments (results not shown) in which Mab 2B was shown to compete
effectively against the biotinylated forms of Mabs 1C5, 4C5, or 4C4. This mutually
competitive binding pattern strongly suggests that these epitopes are located in

relatively close proximity on the enzyme’s surface (31,46). In the case of Mabs 2B,

155

156

Figure 1. Competitive ELISA. The ability of biotinylated Mab 2B to compete with
other Mabs for binding to rat brain hexokinase was determined as described in
Methods. The competing Mab was: 1, Mab 2B (i.e. , the nonbiotinylated Mab 2B); 2,
Mab 3C; 3, Mab 13; 4, Mab 18; 5, Mab 20; 6, Mab 21; 7, Mab 1B2, 8, Mab lDl;
9, Mab 2C5;, 10, Mab 5C2; ll, Mab 3B1; 12, Mab 5A3; 13, Mab 3B2; 14, Mab
2B6; 15, Mab 4C4; 16, Mab 2A1; 17, Mab 4D4; 18, Mab 2A4; 19, Mab 4A6; 20,
Mab 1C5; 21, Mab 4C5; 22, Mab 3A2; 23, control (no Mab competing with

biotinylated Mab 2B). The means :|: SD for three replicate samples are shown.

157

on «N _.u on or or be or or or or «r F? are

a h

 

uru 0517 eoueqrosqv

158

1C5, and 4C5, this is known to be the case since the epitopes recognized by these
antibodies lie, at least partly, within the same 20-residue segment of the sequence
(25). None of the other antibodies had any substantial effect on the binding of
biotinylated Mab 2B (Fig. 1), clearly indicating that the epitopes recognized by these
antibodies are spatially discrete from that recognized by Mab 2B.

The results of similar competitive ELISA experiments with these Mabs are
summarized in Table 1. Based on these and other results (presented below), the Mabs
were organized into nine groups recognizing epitopes lying within distinct regions,
designated A-I, of the molecular surface. Two of these groups contained a single
Mab. In the other groups, with the exception of the Mabs recognizing epitopes in
Region B (discussed below), all of the Mabs within a single group showed mutually
competitive binding.

In some instances, the competitive or noncompetitive character of the results
depended on the order in which the Mabs were allowed to interact with hexokinase.
Similar behavior has been seen in previous studies, and attibuted to conformational
changes induced by binding of one Mab that influence the binding of the second in
nonreciprocal fashion (31,46). Whatever the cause, the lack of mutually competitive
binding clearly indicates that the epitopes are not spatially overlapping.

The molecular regions recognized by Mabs in Groups A-D were readily
defined since each of these groups contains at least one Mab whose epitope location
within the sequence had been established by previous work (25). This was further

confirmed by immunoblotting experiments in which reactivity with tryptic fragments

Table I. Summary of competitive ELISA epitope mapping results‘.

 

 

li

 

G

 

F

 

 

 

 

 

21111404 eel-2013112“ 13121 zalircsltlcs lalrorlzaslaazlm zeslaml'saalscz 4A6 113712114 132

C
N

 

159

ON ON c 7

N N ON CM CM c .c

N N N ON ON iii-cf
N N N ON (ON

'- - 4 ' ' .;,;.;._.;. ......

NNcN

\. "33:1

N
N

0’

 

N N N N N N jac

N CNCNN

N N N N N N cc

N N N N N N if Cc

NNNNNNNNNC

N CNCNN

.j. .............

N CNCNN

 

 

"404 0,
ac
' 20

2A1

3A2 N

401

286 N

l

3’4C4

to 5'- N to N 22.

 

andthaaccondlcttcrindicateathamrltwhenthaantibodyatthctopofthccolumnwutho

‘C, mutually competitive binding. N. naturally noncompetitive binding. In positions were there is a two latter designation, the first letter indicates the result obtained

When the antibody indicated at the 10h was the biotinylated antibody.

biotinylated antibody.

160

corresponding to defined segments of the protein (45) was determined. All of the
Mabs in Groups A-D are reactive on immunoblots, consistent with the view that the
epitopes recognized by these Mabs are continous in nature and do not require native
conformation for recognition. Immunoblotting patterns obtained with Mabs 4D4 (47),
3A2 (24,47), 2B, 3C, 18, 20, and 21 (37) have been published previously. In the
present work, these patterns were confirmed and identical patterns demonstrated with
other Mabs from within the same group, e. g. , the immunoreactivity pattern with Mab
2A1 was shown to be identical with that seen with Mab 4D4 (47), while Mabs 1C5
and 4C5 gave patterns identical to that seen with Mab 2B (37).

Based on the above analysis, epitopic Regions B-D could be represented on the
surface of the proposed structure (2) for the enzyme (Fig. 2). These regions are
represented in CPK renderings in Fig. 2. It will facilitate understanding of subsequent
comments conceming these and other regions if reference is made to this figure.

The antigen binding region of an IgG has been estimated to cover a circular
area approximately 35 A in diameter (48) on the surface of the antigen. Residues
including the epitopes recognized by Mabs 3A2 (Region B) and 18 and 21 (Region C)
have previously been identified (25). With residues at the midpoint in these segments
as a central point, surface regions lying within a radius of 17.5 A were defined. All
surface residues within this radius and in position to interact with an antibody binding
to this segment were included in the epitopic region; surface residues that might be
within the defined radius but which were on a different face of the molecule, and

hence unable to participate in contact with an antibody binding to the central epitopic

161

Figure 2. Distribution of epitopic regions on the molecular surface of rat brain
hexokinase. The proposed structure for the enzyme (2) is used as a basis for this
representation. In Panels A and C, the a-carbon backbone of the molecule is depicted
as a ribbon with the N -terminal domain on the left. Epitopic regions are depicted as
CPK surfaces in Panels B and D. In Panels C and D, the molecule has been rotated

180° around the y-axis with respect to the view in panels A and B.

162

 

D coamx

3 .5531

 

m :23:

u .830:

w :23: a :29:

h a coawm; h

 

 

 

 

 

 

 

163

segment, were excluded.

Region D was defined in essentially the same manner except that additional
information regarding the disposition of Mabs recognizing epitopes in this region (25)
was also taken into consideration. Briefly, it was known that the epitopes recognized
by Mabs 2B, 1C5, and 4C5 are located sequentially, in that order, proceeding from a
surface loop (far left in Fig. 2, Panel A) and into the initial portion of a long helix
running across the bottom (as depicted in Fig. 2, Panel A) of the large lobe in the N-
terminal domain. Hence, as represented in Panel B of Fig. 2, the epitope for Mab 2B
is located in the portion of Region D at the left side of the molecule; epitopes for
other Mabs binding to this region, Mabs 1C5 and 4C5, are located, in order, more
toward the bottom of the molecule as it is oriented in Fig. 2, Panel B.

Region A conesponds to the extreme N—terminal region of the molecule. This
region of the molecule is locally disordered and could not be defined by the x-ray
studies of Steitz and his coworkers (17-20); hence, epitopic Region A is not depicted
in Fig. 2, but corresponds to a disordered segment extending out from the N-terminal
helical segment (off to the left, as represented in Fig. 2). Regions B and D cover
most of one side of the "large lobe" of the N-terminal domain, while Region C is
associated with the ”small lobe” in this domain.

Mabs recognizing epitopes in other regions were not reactive on immunoblots,
and thus may be presumed to recognize conformational epitopes, dependent on
maintenance of at least some semblance of the native structure. This precluded

utilization of methods that had been successfully employed in identification of

164

continuous epitopes in Regions A-D.

That Mabs in Groups E and F recognized epitopes lying within the N-terminal
half of the molecule was shown by ELISA using the isolated N- or C-terminal halves
(23) as antigen (Fig. 3). All of these antibodies reacted selectively with the N -terminal
fragment, as did Mabs 1C5 (Region C) and 21 (Region D) whose epitopes were
known (see above) to lie in the N-terminal domain. As apparent from Fig. 2, there
remained extensive surface area in the N-tenninal domain that were not attributable to
Regions A-D, and the question became one of defining epitopic Regions E and F
within this available area.

The starting point for this was the competitive ELISA results shown in Table
I, and understanding of the logic involved will be aided by reference to the Venn
diagram in Fig. 4. All of the Mabs in Group E showed mutually competitive binding
with Mab 3B2, indicative of spatial overlap with the epitopic regions recognized Mab
3B2. Other pairs of Mabs within this group also showed mutually competitive binding
(e.g., Mabs 2B6 and 13, Mabs l3 and lDl, Mabs 4C4 and 2B6) while other pairings
did not show mutually exclusive binding (e. g., Mabs 4C4 and 13 or lDl). Thus, the
relative disposition of the epitopes recognized by the Mabs in Group E was as
depicted in Fig. 4. The location of this region within the available (after discounting
Regions B-D) surface on the N-terminal domain was deduced from the observation
that Mab 4C4 also showed mutually competitive binding with Mabs located in the
previously defined Region D (Table I and Fig. 4). Hence, epitopic Region B was

deduced to be adjacent to Region D, as shown in Figs. 2 and 4.

165

Figure 3. Reactivity of Mabs with isolated N- and C-terminal domains of rat
brain hexokinase. Reactivity of various Mabs with the isolated N-terminal (stippled
bars) or C-terminal (crosshatched bars) domains was determined by ELISA, as
described in Methods. In Panel A, the Mabs used were: 1, Mab 1C5; 2, Mab 21; 3,
Mab 2C5; 4, Mab 5C2; 5, Mab 3B1; 6, Mab 5A3; 7, Mab 13; 8, Mab lDl; 9, Mab
4C4; 10, Mab 2B6; 11, Mab 3B2. In Panel B, the Mabs used were: 1, Mab 3C; 2,
Mab l8; 3, Mab 2B6; 4, Mab 3B2; 5, Mab 1B2; 6, Mab 1E7; 7, Mab 2A4; 8, Mab
4A6. Means i SD for three replicate samples are shown; the absence of visible error

bars is due to extremely small SD values.

166

 

8

 

 

7

 

 

6

 

2345

1

 

  

V, . . .
Mt; «.

KWNJI.‘ fivM“! F ' ‘5‘Hr'
:- . «1.1,: r\

  

«'"Mc’W-wrwt'e'k‘ww -- , .
4 ~12} ~ ‘4’” ‘ a - .1 ;. ~. 1 ., t 4:}
." '. J . - . c' A V V4W“‘"" "' ' '

   

91011

8

7
ab

     

  
 

     

' . we:
‘15., K ,r. 1..

    

 

 

 

LUU 0517 aoueqrosqv

ab

167

Figure 4. Venn diagram, depicting relationship between epitopes in regions D and

E, deduced from results of competitive ELISA.

168

 

169

Although the boundaries on three sides of Region E were defined by the
previously scribed epitopic Regions B-D, the fourth side (adjacent to the still-to-be-
described region F) was not so delimited. However, examination of the proposed
structure suggested the location of this final boundary. Region B proceeds along the
surface of the large lobe into the hinge region between the large and small lobes.
There is a marked discontinuity in the surface at this point (Fig. 2, though more
evident when examined in stereo representations (17) or by computer graphics), and it
seemed reasonable to suspect that this might mark a boundary between Region E and
another discrete surface region containing epitopes for the Group F Mabs, none of
which showed mutually exclusive binding (and hence no overlap) with the Mabs of
Group E.

The above logic thus placed the epitopes for the Group F Mabs on the face of
the small lobe that was opposite to that identified with Region C. The total lack of
any mutually competitive binding between Mabs of Group F and those of other
groups was, of course, fully consistent with this placement, but it did not provide
direct evidence for it. This was, however, forthcoming from an alternative
experimental approach. Examination of the surface region tentatively designated as
region F disclosed that a previously identified (2,45) tryptic cleavage site, called T1,
was located centrally within this region. It could be expected that Mabs binding in
this region would obstruct access of trypsin to T1, with resulting decreased cleavage
at this site. Based on the previously defined (45 ) cleavage pattern, hindered cleavage

at T1 would result in decreased formation of a 90 kDa fragment (formed by cleavage

170

of the 100 kDa enzyme at T1, with concomitant production of a 10 kDa fragment
representing the extreme N-terminal region) and increased amounts of a 60 kDa
fragment whose further processing requires cleavage at T1. As shown in Fig. 5, this
prediction was fulfilled with Mabs of Group F. Digestion of hexokinase complexed
with these Mabs resulted in virtually undetectable levels of the 90 kDa fragment being
formed, with accumulation of the 60 kDa fragment to levels much greater than that
seen in digests of the uncomplexed enzyme.

Regions A-F accounted for virtually the entire surface area of the N-terminal
domain of hexokinase (Fig. 2). It thus seemed most unlikely that the Mabs associated
with Groups G-I, could also recognize epitopes lying in this domain. Accordingly, it
was anticipated that the epitopes recognized by Mabs in Groups G-I must lie in the C-
terminaldomain. Direct demonstration of this was provided by preferential reactivity
of these Mabs with the C-terminal domain in ELISA (Fig. 3), although this was less
striking with Mab 4A6 due to the relatively low affinity of this antibody for the
isolated C-terminal domain (see below). Further evidence for the location of these
epitopes was provided by immunoprecipitation of the catalytically-active isolated C-
tenninal domain (23) with Mabs in Groups G-I (Fig. 6). Mabs 1E7, 2A4, and 1B2
immunoprecipitated the intact 100 kDa enzyme with affinity similar to that for the C-
terrninal domain (results not shown). In contrast, Mab 4A6 precipitated the intact
enzyme much more effectively than it did the isolated C-terminal domain (Fig. 6).

Conversely, Mab 4A6 (Mabs 1B2, IE7, and 2A4 were not used in this

experiment) could not immunoprecipitate a truncated version representing the N-

171

Figure 5. Effect of Mabs on the tryptic digestion pattern of hexokinase.
Hexokinase, either alone (lane 10) or complexed with Mab (lanes 1-9) , was digested
with trypsin as described in Methods followed by electrophoresis on SDS acrylamide
gels. The intensely stained bands corresponding to heavy (H) and light (L) chains of
the Mabs are evident. The location of the intact 100 kDa enzyme and of the 90 kDa
and 60 kDa tryptic fragments of interest here (see text) are marked at the left. Other
tryptic cleavage products (not marked) are evident but are not relevant in the present
context. Hexokinase in Lanes 1-9 was digested after complexation with: Lane 1, Mab
2C5; Lane 2, Mab 3B1; Lane 3, Mab 5C2; Lane 4, Mab 5A3; Lane 5, Mab 18; Lane

6, Mab 4C4; Lane 7, Mab 2B6; Lane 8, Mab 1D1; Lane 9, Mab 1C5.

172

12 3450 139110-

- .
--—---—

   

——-—-_~

173

Figure 6. Immunoprecipitation of intact hexokinase and the isolated C-terminal
domain by Mabs. The isolated C-terminal domain (0.25 pg) was incubated with
increasing amounts of Mab 4A6 (O), Mab 1E7 (D), Mab 2A4 (I), or Mab 1B2 (A)
and resulting complexes immuunoprecipitated as described in Methods. Activity
remaining in the supernatant is expressed as a percent of control (activity present in
samples incubated with no Mab). Results of immunoprecipitation of intact hexokinase

(0.5 pg) with Mab 4A6 are also shown (O).

174

 

 

100 A

 

 

2.0

 

 

.0
CU
. E
O U)
-. I 3
F
I
O
o
l l I I o.
o o o o O
co ‘9 q. N

IorruOO 1° %

175

terminal half of the enzyme (25), although this was immunoprecipitated by Mabs
recogizing epitopes in the N-terminal domain (Fig. 7). It is worth noting that all of
the Mabs used in the experiment shown in Fig. 7 recognized conformational epitopes
(see above); their ability to immunoprecipitate this truncated protein, synthesized in
vitro, implies that the truncated version is able to fold to something resembling the
native structure. This is consistent with both the proposed organization of hexokinase
into relatively discrete N- and 3The analog, l,5-anhydroglucitol-6-phosphate, is used
since it is neither a substrate nor inhibitor of Glc-6-P dehydrogenase (33) and hence
the convenience of the Glc-6-P-coupled spectrophotometric assay can be retained. It
has been confirmed that this analog, like Glc-6-P, is a competitive inhibitor (vs. ATP-
Mg” +) inhibitor of brain hexokinase (44).C- terminal domains (2,47) as well as with
the concept that domains represent independent folding units of a protein (49).
Binding of Mab 4A6 was mutually competitive (Table I) with Mabs
recognizing epitopes in Region B of the N-terminal domain. Attributing this to spatial
proximity of the epitopes leads to positioning of Region G in the area of the C-
terminal domain shown in Fig. 2, i.e., on one face of the large lobe in the C-terminal
domain. We have no basis for assigning Regions H and I to other areas on the C-
terminal domain. However, their failure to exhibit mutually competitive binding with
each other or with Mab 4A6 (Table I) makes it evident that these epitopic regions are
discrete. Certainly, there is more than adequate surface area on the C-terminal domain

to accommodate this requirement.

176

Figure 7. Immunoprecipitation of a truncated version of hexokinase by Mabs. A
truncated version of the enzyme, containing residues 1-453, was synthesized in vitro
using [3H]-1eucine as radiolabel (25). The figure shows an autoradiogram obtained
after analysis by SDS gel electrophoresis. Total translation products are shown in
Lane 1. The location of the 453 residue fragment is indicated at the left; smaller
fragments are the result of aborted translation or degradation by endogenous proteases
in the translation system (25). Other lanes show products immunoprecipitated with
Mab 2B6 (Lane 2), Mab 1D1 (Lane 3), Mab 5C2 (lane 4), Mab 2C5 (Lane 5), Mab

4A6 (Lane 6), Mab 13 (Lane 7), or a control Mab that did not recognize hexokinase

(Lane 8).

177

1 2 3466 78

   

45300»

178

Ligands affect the interaction of Mabs with hexokinase. Marked conformational
changes have previously been shown to result from binding of various ligands to brain
hexokinase, and correlations with function (e.g. , ability to serve as substrate or
inhibitor) have been established (23,24,26-30). These were also detectable by altered
reactivity with Mabs; not unexpectedly, conformational epitopes (Regions E-I) were
particularly sensitive (Table II).

Epitopes in Region B were virtually unaffected by binding of hexoses or
hexose 6-phosphates (or analogs). In contrast, substantial effects on immunoreactivity
were seen in the presence of NTPs. This was markedly dependent on the base present
in the NTP. GTP had a particularly devastating effect - virtually abolishing
immunoreactivity with Mab 3B2- while ATP decreased immunoreactivity to a lesser
extent and CTP still less. With Mabs 2B6, 3B2, and 4C4, the Mg+ + chelates had
effects that were indistinguishable from those seen with the uncomplexed forms of
these N'I'Ps; with Mabs 13 and lDl, however, chelation tended to diminish the effects
of NTPS on immunoreactivity. Overall, the effects of NTP binding on epitopes in
Region B seem to be dominated by the nature of base present rather than chelation
status, suggesting that the conformational changes evoked in Region B are primarily
determined by interactions with the pyrimidine or purine moiety of the NTP.

Altered immunoreactivity patterns in response to binding of ligands were quite
distinctive with Mabs recognizing epitopes in Region F, particularly with Mab 5A3.
Immunoreactivity with Mab 5A3 was more than doubled in the presence of the

substrate hexose, Glc, but was not affected by nonsubstrates such as Gal. None of the

179

$5310.?» 380E“. 303.3

coast: 05 «coop—ac.— boéooho .555 «gunmhe o5 .Aflm V 31283 30— 33 .21030 Sun? a £5? count; 05 35m .uouauufiée
2.2 :08 8 .5: umflflge E :32? a £80 3 3053583 8.231.. 55 .auofitomxo he poses: 2F 62!, can»: o: .3 .822. n

9 @8835 33 aqua: no 8:82.— 05 E 253 an: me .503: 05 e:- fioo— u- eoeuoe a; one»: no 833- 05 3 1.53 .32 cc 2508. 2C.
6350.2 e5 235:2 coon: 3388—. (min 05 new»... e883. 83 83223.— 55 3.2 ca 330233 05 acou- 8 85a: co 3:5. 2F.

an“?
2“?
3me
on”:
efima
mem
EH3

nHm~
mung
h E.
N 2
Emma

nfima
uflwa

Emma
he—

«13..
mu 5
5H5.
«Hun
@—
m3”:
.3”:
aHmm
mummy
«LUS—
H2:
«H8—
euro—

nflaw
nu
3H3
m—«hm
2H3
9H3
mafia:
«2:8
353

my”?

3 S
Zflew

m3"? a“:

Hm» Sumac

b nu“: nun:
SHE.
wfimv .3 my":
5H8 gfiue— He»
can—"em _ a
3": H8

«Nunc» mfieem Hme
H5 ~Hne— HHS
mu”; 3”: afiee
human 2:

R

«Hue
a

8—
mafia
mung
$23
in:

1:5 0H3 manna
«Hue mung
in:

vfiwa

manna

3H3 wfina
my”:

3“;
mung
mung
manna
N

wfiaa
«Ha

 

(88305: 885 .23 332 mo .8322an 05 no ”one”: we Locum d 055.

180

 

nude

guns

3 “In

Suns

nflmv

3H3

nflwm

3% .NM

 

3H3.

mafia

unfiwm

aH—fi

nu":

gnaw:

an”:

«HR .__

 

a“?

sun—"me

5 firm

SH?

1:.»

as“:

”fine

 

5H3

aflnn

9%:

Eng

3":

can: _

3ch

a:

 

by":

aflum

=an

manna

nflhn

cance—

Ena. Era

 

5H5

6H5

Sum—

an":

nflhu

mum—o

aHNN

a«_n_

 

sauna

“263

2 «ma

u—Hmcu

can:

333

«H3

«H3

 

Emma

can:

”Ham

3H3—

16a

9 H 93

36a

an": M

 

«Hag

3H3

NEH—m

«Haw

nan"

wfimhm

nflnm

33. w.

 

3H3—

3H9

can—”mm

3H9.

nfinm

3 .fi wen

23333...

 

3H8

”Han

mHNN

S Hem

nun:

mnHFm

vfiwn

has

 

mHR

”Hum

aflvm

”H3“

Inne—

«H2:

anS

3:. .__

 

«an:

QHNN

an“:

9H3 sauna

3H3"

NHS—

33

 

 

 

 

 

 

 

5%. Sifl stun

_...§ E 53 ea 53." __h__

m .5an

 

 

 

 

 

 

 

.Goscucouv .= 2an

181

other Mabs of Group F showed altered immunoreactivity in the presence of added
hexose, substrate or not.

Potent inhibitors, such as Glc-6-P, 1,5-anhydroglucitol-6—P, or 5-thioGlc-6-P,
caused a rather spectacular increase in immunoreactivity with Mab 5A3 while poor
inhibitors, Man-6—P or Gal-6-P, were less effective. Conversely, these same
compounds decreased immunoreactivity with other Mabs in Group F, with markedly
greater effects again being seen with the potent inhibitors.

It is thus evident that epitopes located in Region F are highly sensitive to
ligand-induced conformational changes related to catalytic and regulatory functions.

NTPs substantially decreased immunoreactivity with Mabs of Group F; a
single exception was the lack of effect of CTP on reactivity with Mab 5A3. As in the
case of Region B, it was evident that the evoked conformational changes were
sensitive to the base moiety, with ATP, GTP, and CTP each giving unique
immunoreactivity patterns. However, in contrast to the results with Mabs of Group E,
immunoreactivity with Mabs of Group F was markedly affected by chelation status of
the phosphate sidechain, with chelation partially reversing the decrease in
immunoreactivity seen with the unchelated NTPs; Mab 5A3 was again an exception,
with relatively modest difference between effects of chelated and nonchelated forms.

Binding of hexoses had little if any effect on conformation of Regions G and I,
as judged from lack of change in immunoreactivity with Mabs binding in these
regions. In contrast, another region of the C—terminal domain, Region H, was affected

in a manner reflecting conformational changes that were correlated with substrate

182

activity; Glc greatly decreased reactivity with Mabs 2A4 and IE7, while the poor
substrate, Gal, had much less effect.

Inhibitory hexose 6-phosphates (or analogs) substantially altered the reactivity
with Mabs binding in all three regions of the C-terminal domain, with Regions G and
H being particularly affected. This was related to inhibitory function since poor
inhibitors such as Man-6-P or Gal-6-P had virtually no effect on immunoreactivity
with any of these Mabs. A single anomaly was the notable variation in the effect of 5-
thioGlc-6-P on reactivity with Mab 2A4; we have no explanation for this.

Immunoreactivity with Mabs recognizing epitopes in the C-terminal domain
was also influenced by binding of NTPS, with both increases and decreases in
immunoreactivity seen. The reactivity patterns were dependent on both the nature of

the base moiety as well as chelation status of the phosphate sidechain.

Li .. d . a he int .. in f ab wi h h ml; ac N- . - rmin . main f
hexokinase.

Previous work (23,26) had demonstrated the existence of binding sites for
hexoses, hexose 6—phosphates, and NTPs on each isolated domain, with Glc and Glc-
6-P sites, latent in the intact enzyme, appearing on the isolated N— and C-terminal
fragments, respectively. It was thus of interest to compare the conformational
response of these isolated domains with that of the intact enzyme, as judged from
immunoreactivity with the various Mabs. These are compared in Table III. Again not

unexpectedly, Mabs recognizing conformational epitopes (Regions E-I) were

183

685.58 85:81. 3 82— 32.33. 2- E.- .n as;

S 8521 806 an 3.3 05 8‘ e 08.38 .085 05 5.3 338M .5133 1555-0 .338. no dad—ow 19.58- 2 18:3. .92 .8330on .035 r
.5385?» accent—v 3933

acacia.» on. 858.52 3268?. Even. cow-Sow 05 A: n V 31398 32 an? .uufitoauo 03% a .523 count: 05 8.5 dew-:38“. an:

53 o. :8: £853.!— 3 E521 a .58 5 mucus—Eu 8.235 .23 iguana he r355: 2F 4— 03¢. a. $3168“. a. vomoéaxo 8. £33..

 

hung a“: 13.— Emma 3H3. 3H9. Sauna 3th ~13» :th mu”: a“; sauna mfina 3H8 wanna .5»:th

 

EHoS 2H: Enac— nfiam 3me m2"; ”133 vnfinou vflca 2H3 3:8 3": vfimc— my"; 5H2: my": . «nu—ads

 

9H2: 3H3 nfiaw 5H: 3H"? an”: afien— 3ch 9H?” :HS SHE“ 36a aanam—Hom 3H2: 3H2 ...w,«mE.A—._.<

 

 

 

 

 

«Huh can: oHu— «H: wfimv .1an :Hom 5H3. 5H3: n“: «Hma nflam nu": mfina wfinw 1n; . . APO
wfiwa efima nHNS nHma 5H3 c—Huw m—HG SH: 3H2: nun: nu”; cfiaa mun—3 3"; «H3 my"; . .3 . PHD
canon mfiww 1”; ”Lu: wfimm 3va 2H; wfimv 3H2: nflva mfiwa «H2: sauna wanna an: mung . .Ek
afimo— w—Haa by”; nfimm 9H2. afiva 3H3 5H8 3"; 3me 363 mung“ u“; wanna mung nfiwa T.A—ééG
EH2: cNHS— ”Luna “Haa eNHNh afima 3H? on“: «fig mun—a «Hma “Hg "#3 «#3 1.”? «Ha . mééaz

 

auflm: cNHmS «Hmo— wHNo— 3H3 fifinuu 2H3 «ufiva «1:3 3%: can: «Hwa «HR SH? «flea 9H3 .A—«WEUfififi,

 

wNHNS 2H3 uflva bfiwa 3H3 Int: 9me 3H: 3H? mufimw a“; mung «Hg 5H3 1"; vaa .mééaéafiq—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2f: :3: in: in: a“: 2;: :Hmm 2“: 2i: was «a: 3: «“833: 32 mg: . , .536
we: 32: $2: $2: a“; «is a“: 3; same :52 1:: is. 38— %: «ea $2: ._ .6
228 32: 33 came a“; .32 :«m Z3 2“:— nfls a“: "a": 12: $3 1: a“: ,. go
5.9 65,. an: 63: 82: 52: an. 52 am“: an: 52 a: 58 as 5:: 6:." a:
azunz~ez_nz_,nz~.nz._nz,_nz._

m. : a . o m < :25

 

 

 

 

 

 

 

baggage: 835 .5 Eafiou
28:58-0 co -Z @2203 on. 53» Egg we coroSBE 05 :0 mean»: .8 “onto 05 mo 533800 .E oEaH

184

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

vHNw fiHNw ”Hon cufiue wand: Sums 3H3 3H3 5H3. ”Hem mfiem afivm H+ 4:.0
vfiww SH: SH; 2H3 wflnufi aura vflnh gun—um: 3H3. nu": aflew VANS 2+"w2dub
3H3. 3H3 a“? rnfimv 3H3— Sfioa :th may”: oufiem «Haw nfluh "Hum ,,.,+aw2.m.—.<
Nufioh 5H3 vfiwe afinn mung 5H3. 5H5 can: Sfion 3H8 =an vam . «:6
3H3. 5H2. m“: 5H? wflwa mu”: wfine efioo— wflvn 5H9. SHE. 3H3 ,, as
m“: 5H5 5H8 5H5 SH?» aHE. anv 9H5 wHw— aHNN wfiwv nun—n 5:2
«H2: mee— ofimo— 3H2: =va 2H2: wHS— vfiaw— Sum» «H3 NH: ”Han ,. $9.15
553 fiva bung aunt: n—Hab 3H2: my”: n—meu 2H3 9H3 «H:- wku wmééuE
$62 «”32 efimm mauve» Nfifimv wfiav :uma wfimhm mufimn nflnn HHS name A.W.”..Uofifim
2H2: Euro: .2va vnfimm Efiwm 3H3. Q33 waflwom 3an SHem mfinm Efivm mééaéfim;
3H2: w—HnN— has wfiwn 5H3 SHen summon anflfim 31.2 vaN ufiwm 5H3. A620
EUR mun; mea ”H: mfiww wanna vaE :52 ono— 353 «in? a“? 30
$33 ”flu: 9.an ”Luna aHmN eHR «Hm: m~Hm- :3“:— «HS— vflwou mfia .. , 30
am”: am“: Avvv<n Evin A336 9.35., 3:36 52m 5:: 3:: Ramos AnvaN , 352
90 H no n Go ~ ,QZ a D2 u DZ —_

H. I U W, m conned

 

 

 

.Gonauqoov .E oEaH

185

particularly affected, but some notable differences between intact enzyme and isolated
domains were observed.

Binding of hexoses to the substrate site in the C—tenninal domain (22,23) of
the intact enzyme had only a modest effect on reactivity with Mabs binding to Region
B of the N -terminal domain, and this was not correlated with effectiveness as
substrate (results above). With the isolated N -terminal domain, both Glc and Gal had
little if any effect on reactivity with Mabs 236, 3B2, and 4C4. With Mab 13,
however, Glc had significantly greater effect than did the poor substrate, Gal. Thus,
the N-terminal Glc binding site, revealed by separation of the domains, now manifests
its specificity with conformational changes affecting the epitope for Mab 13.

Hexose 6-phosphates had little effect on reactivity of the intact enzyme with
Mabs recognizing epitopes in Region B. This remained the case with the isolated N -
terrninal domain, although the effects of hexose 6-phosphates on reactivity with Mabs
13 and 236 were somewhat greater than seen with intact hexokinase. Thus, it is
evident that the conformational epitopes of Region B are relatively insensitive to
binding of hexose 6ophosphates to either the intact enzyme or isolated N-terminal
domain.

As with the intact enzyme, conformational changes induced by binding of
NTPS to the isolated N—terminal domain and affecting Region B depended on the
nature of the base moiety, although chelation status of the phosphate sidechain was
also clearly a factor (e. g. , compare reactivity with Mab 13 in the presence of the free

and chelated forms of the NTPS).

186

The effect of ligand binding on immunoreactivity with Mabs 2C5 and 381,
recognizing epitopes in Region F, was rather similar in the intact enzyme and isolated
N-terminal domain. For example, in both cases, immunoreactivity was insensitive to
binding of hexoses, and was decreased by binding of inhibitory hexose 6-phosphates
(or analogs) but was much less affected by binding of poor inhibitors. Similarly, with
both Mabs, immunoreactivity was sensitive to the particular NTP used as ligand and
to the chelation status of the phosphate sidechain. However, in marked contrast to
results with the intact enzyme, reactivity of the isolated N-terminal domain with Mab
5A3 was not significantly affected by any of the hexoses or hexose 6-phosphates (or
analogs) used. Thus, the conformational changes induced by binding of hexoses or
hexose 6-phosphates that resulted in altered reactivity of Mab 5 A3 with the intact
enzyme are dependent on the presence of the C-tenninal domain. The sensitivity of
reactivity with Mab 5A3 to changes evoked by binding of NTPS or their Mg” +
chelates, seen with the intact enzyme, was preserved in the isolated N-terminal
domain.

In contrast to the lack of effect of hexoses on reactivity of the intact enzyme
with Mab 4A6 (Region G), reactivity of the isolated C-terminal domain was markedly
decreased by the binding of Glc while Gal had much less effect, again reflecting
substrate preference. Thus, relieved of constraints present in the intact enzyme,
functionally related conformational changes in Region G are manifested with the
isolated catalytic domain.

With hexose 6—phosphates, effects on reactivity with Mab 4A6 were

187

qualitatively similar for the intact enzyme and the isolated C-terminal domain. It
should be noted that, despite the similarity in conformational response in this region
of the molecule (as judged from effect on immunoreactivity), the mechanism by which
it is evoked must be quite distinct. With the isolated C-terminal domain, it results
from binding of the ligand directly to this domain (23) while in the intact enzyme, it
requires transmission of conformational information through the molecule from the N-
terminal binding site for Glc-6-P (24).

As with the intact enzyme, binding of NTPS or their chelated forms had
relatively modest effects on reactivity with Mab 4A6.

With the other Mabs recognizing epitopes in the C-tenninal domain (Regions
H and I), effects of ligand binding on immunoreactivity were rather similar between
intact enzyme and isolated C-terminal domain although the effects of NTPS, either
free or chelated with Mg+ + , were generally diminished with the isolated C-terrninal

domain.

Eff t f s n ’ ' te f h

The effect of these Mabs on kinetic properties of hexokinase was examined.
Specifically, effects on Kms for Glc or ATP-Mg”r + and on the K, for inhibition by the
Glc—6—P analog, 1,5-anhydroglucitol-6-P3 (44), as well as effects on Vm were
determined. None of the Mabs binding to Regions A-E affected any of these
parameters, nor did Mab 5A3, which bound in Region F, or Mab 132, which bound

in Region I of the C-terminal domain. The lack of effect of Mabs 2B, 3C, 13, 18, 20,

188

.nm: spas Umxmamfioo uoc ofihuco on» npfi3 comm 51> no unooumm»
.uouwnflncw mm oumnmmonmtmtaoufiozamoun>ACMIm.H mcflmb.

 

OCH
OCH

mm
OCH
OCH
OCH
OCH

0H «.H mm mfle no: 0
HH N.H Hm HOHuCOU

 

new mum answfiumutu

«em mm.H hma o<¢ an:
no mm.o we mom an:
mm Hm.o ow Hmm an:
mm mm.o me mom no:
0H mm.o we Houucoo

«Edging

‘34th

 

V... >

_ _ 22.31:...IH

 

I . I.

.Eane cubs—8 Rams—~86 338$ 05 23 82388:
:35 835 we £80883 ozone— ofi co 322 he Beam .3 Bank

189

and 21 on kinetic parameters of hexokinase had been previously noted (32,37).

In contrast, other Mabs binding in Region F (Mabs 2C5, 3B1, and 5C2) had
no substantial effect on Km or me values, but caused a 5 to 7-fold increase in K
(Table IV). The effect with Mab 4A6 (Region G) was even more dramatic, increasing
the Km values for both substrates by 2 to 3-fold and the K, for 1,5-anhydroglucitol-6-
P by over 30-fold, while decreasing V“m to 59 % of that for the uncomplexed (with
Mab) enzyme. The enzyme could be completely immunoprecipitated with Mab 4A6
(Fig. 6) , demonstrating that the residual activity seen in the presence of excess Mab
4A6 was not due to hexokinase that, for some reason, might have remained
uncomplexed with the Mab. In contrast to results with the intact enzyme, Mab 4A6
had no substantial effect on K"I or K values, nor on the me of the isolated C-
terminal domain (Table IV).

Complexing with Mabs binding to Region H, Mabs IE7 or 2A4, virtually
abolished (V m < 5 % of the free enzyme) catalytic activity of the intact enzyme and
reduced the V“m of the isolated C-terminal domain to approximately 20 % of that seen
with the uncomplexed domain. In the latter case, the effect of antibody binding was

shown to be on Vm, with no significant effect on K“I for either substrate.

DISCUSSION

We begin this discussion with some observations on the distribution and nature
of the epitapes recognized by Mabs generated against rat brain hexokinase. It will be
noted that the above study included many more Mabs recognizing epitopes in the N-
terminal domain than in the C-terminal half of the molecule. This is not due to any
selection on our part but rather reflects our experience in several different fusion
experiments: many more Mabs directed against the N -terminal region are generated.
We have noted this previously (37,50) and pointed out that such "clustering” of
epitopes has been seen with other proteins (e.g., 51-55). What would make the C-
terminal domain so nonimmunogenic compared to the N -terminal half?

The concept of "self-tolerance" (56,57) leads to the expectation that highly
conserved regions of the molecule would be less irnmunogenic. The 100 kDa
hexokinase is comprised of a C-terminal half with catalytic function (21-23) shared by
all three isozymes, and an N-terminal half serving a regulatory role (24) which differs
somewhat among the isozymes (l , and references therein). It might be anticipated that
the requirements for maintenance of catalytic function would be reflected by greater
conservation of structure in the C-terminal domain. Consistent with this are sequence
comparisons (5) which demonstrate greater identity between catalytic domains than
regulatory domains of the rat isozymes: 76 % identity between the C-terminal domains
of the Type I and Type II isozymes compared with 68 % identity in the N-terminal

190

191

domains, and 66 % identity between the C-terminal domains of the Type II and Type
III isozymes compared with 44% in the N—terminal domains. On the other hand, the
32 differences between the sequences of the Type I isozymes from rat (2) and mouse
(4) are rather evenly distributed between the domains, with 18 being found in the N-
terminal halves. Hence, there is no a priori reason to expect that immunization of
mice with the rat enzyme would preferentially induce antibodies against the N-
terminal domain. Moreover, if immunogenicity were determined by the degree of
sequence divergence, then it would be expected that the major immunogenic regions
on the otherwise conserved C-terminal domain would be at surface loops or
protrusions, where greater divergence in sequence can be tolerated without
detrimental effect on structure (58). Such regions are typically associated with
continuous epitopes (31,59), and it would therefore be predicted that the latter would
predominate on the C-terminal domain. This is not in accord with the observation that
all of the Mabs in Groups G-I recognize conformationally—sensitive epitopes. Thus,
we do not find ”self-tolerance" to be a particularly satisfying explanation for the
observed lack of immunogenicity seen with the C-terminal domain, which remains
perplexing.

Another intriguing observation is that all of the epitopes located in Regions A-
D appeared to be of the continuous type, as indicated by the ability of the
corresponding Mabs to react with denatured fragments of the protein on immunoblots
or with truncated versions of the protein synthesized in vitro (25). In contrast, all

Mabs identified as binding in epitopic Regions E-I recognized conformationally-

192

sensitive epitopes and were unreactive with denatured or truncated fragments of the
protein. There seems no obvious reason to expect such divisions, e. g., Regions B and
G are localized at more-or-less equivalent regions of the N- and C-terminal domains.
Our present inability to explain the irnmunodominance of the N -terminal domain, or
the dominance of continuous epitopes in certain regions and conformational epitopes
in others, makes it evident that there is much still to be learned about the relationship
between structure and immunogenic activity.

Ligand-induced conformational changes were not reflected by substantially
altered reactivity with Mabs binding in Regions A-D. It would, however, be a mistake
to interpret lack of effect on immunoreactivity as an indication that ligand binding had
little effect on structure. Dramatic changes in reactivity of sulfhydryl residues in these
regions have been shown to result from binding of various ligands (30), and attest to
the occurrence of significant structural changes. Similarly, binding of ligands such as
Glc-6-P or NTPs has been shown to protect the N-terminal domain against proteolysis
(23,26), indicative of ligand-induced conformational changes in this region. Thus, the
lack of effect of ligand binding on immunoreactivity with Mabs of Groups A-D is
more appropriately attributed to the continuous nature of the epitopes recognized by
these Mabs. As noted above, such epitopes are likely to be associated with
hydrophilic surface loops or protrusions (31,59). Loss of reactivity with structures of
this nature would imply loss of accessibility by movement of these hydrophilic into
more interior - and more hydrophobic - regions of the molecule, which seems very

unlikely.

193

In terms of providing information about effects of ligand binding on structure,
results obtained with the Mabs recognizing conformational epitopes in Regions E-I are
more informative. While the precise nature of the conformational changes obviously
cannot be determined from observed effects on immunoreactivity, these results do
direct attention to specific regions of the molecule being affected by binding of
ligands. Several of these have been noted in the course of presenting the results
above. For example, although Region B was relatively unaffected by binding of
hexose 6-phosphates, it was remarkably sensitive to N'I'Ps.

Effects of hexose 6-phosphates on immunoreactivity were particularly notable
with Mabs binding in Regions F and G, and were correlated with the relative
inhibitory effectiveness of the hexose 6-phosphate or analog. These regions are
located near the interface between the N- and C-terminal domains. 17w selective
sensitivity to binding of inhibitory hexose 6-phosphates or analogs identifies these
regions as being involved in transmission of a conformational signal fiom the
regulatory binding site in the N-terminal domain to the catalytic C—terminal domain.
Binding of Mabs to these regions may be expected to impede conformational changes
associated with the binding of inhibitors, with resulting effect on Ki (29). The
increased Ki resulting from complexation of hexokinase with Mabs binding to Regions
F and G (Table IV) is in accord with this expectation.

The effects of binding Mab 4A6 on kinetic properties of intact brain
hexokinase were not seen with the isolated C-terminal domain (Table IV). Moreover,

the affinity of Mab 4A6 for the isolated domain was appreciably less than that seen

194

with the intact enzyme (Fig. 6). And finally, although binding of Glc had little effect
on immunoreactivity of the intact enzyme with Mab 4A6, the reactivity was markedly
reduced by binding of Glc to the isolated domain (Table III). All of these results
indicate that the presence of the N-terminal domain has a substantial effect on the
conformation in Region G (epitopic region for Mab 4A6) and on its response to
binding of ligands. Conversely, it is also evident from the results with Mab 5A3
(Table III) that removal of the C-terminal domain has a remarkable effect on the
response of Region F to binding of Glc-6—P in the N-terminal domain. This mutual
sensitivity of Regions F and G to the other’s presence surely suggests close
interactions between the domains in these regions, and is thus consistent with the
proposition that the inhibitory signal from the N-terminal domain may be transmitted
through these regions with resultant effect at the C-terminal catalytic site.

Binding of Glc to yeast hexokinase results in closure of the cleft (17). This is
critical for catalysis and presumably results in proper alignment of the 6-hydroxyl
group of the hexose, the terminal phosphoryl group of the ATP, and catalytically
important residues of the enzyme. Catalytically important residues of the yeast
enzyme (17) are completely conserved in the mammalian hexokinases (2-6,36,60),
and site-directed mutagenesis studies by Arora et al. (61) confirm that these are
critical for catalytic activity. It is thus reasonable to expect great similarity in
mechanism, with cleft closure induced by binding of the hexose substrate also being
an integral part of catalysis by the mammalian enzymes. Immunoreactivity with Mabs

2A4 or IE7 is greatly decreased in the presence of Glc but not in the presence of

195

nonsubstrate hexoses such as Gal. These Mabs recognize a conformationally-sensitive
epitope in the C-terminal catalytic domain, and it is reasonable to attribute the
decreased immunoreactivity to structural rearrangements associated with cleft closure.
Selective recognition of the " open cleft" conformation, preventing closure of the cleft
required for progression of the catalytic cycle, would account for marked inhibition of
catalytic activity seen to result from binding of these Mabs.

Inhibitory hexose 6-phosphates (or analogs) also decrease reactivity with Mabs
2A4 and IE7 , thereby implying that binding of these inhibitors at the N—terminal site
in the intact enzyme (24) results in cleft closure in the C-terminal domain. Results of
limited proteolysis studies (A.D. Smith and J.E. Wilson, unpublished work) are also
consistent with this view. Both substrate hexoses and inhibitory hexose 6-phosphates
decrease tryptic cleavage at a site located in the hinge region linking the large and
small lobes of the C-terminal domain; cleft closure could reasonably be expected to
hinder accessibility to cleavage sites in the hinge region. These results are therefore in
accord with the view that inhibitory hexose 6-phosphates induce closure of the cleft in
the catalytic C-terminal domain. Resulting preclusion of access to the ATP binding
site in the catalytic domain may account for the mutually exclusive nature of ATP and

hexose 6-phosphate binding (26,27,62).

10.

11.

12.

13.

REFERENCES
Wilson,J.E. (1984) in Regulation of Carbohydrate Metabolism (Bietner,
R.,Ed.), Vol. I, pp.45-85, CRC Press, Boca Raton, FL.

Schwab, D.A., and Wilson, J .E.(1989) Proc. Natl. Acad. Sci. USA 86, 2563-
2567.

Nishi, S., Seino, S., and Bell. 6.1. (1988) Biochem. Biophys. Res. Commun.
157, 937-943.

Arora, K.K., Fanciulli, M., and Pedersen, PL. (1990) J. Biol. Chem. 265,
6481-6488.

Thelen, A. P., and Wilson, J.E. (1991) Arch. Biochem. Biophys. 286, 645-
651.

Schwab, D.A., and Wilson, J .E. (1991) Arch. Biochem. Biophys. 285, 365-
370.

Kopetzki, E., Entian, K.-D., and Mecke, D. (1985) Gene 39, 95-102.

Stachelek, C., Stachelek, 1., Swan, D., and Konigsberg, W. (1986) Nucleic
Acids Res. 14, 945-963.

Colowick, S.P. (1973) in The Erzymes (Boyer, P.D., Ed.), 3rd ed., Vol. 9,
pp.1-48, Acedemic Press, New York.

Easterby, IS, and O’Brien, M.J. (1973) Eur. J. Biochem, 38, 201-211.

Rose, I.A., Warms, J.V.B.,and Kosow, DP. (1974) Arch. Biochem. Biophys.
164, 729-735.

Holroyde, M.J. , and Trayer, LR, and Cornish-Bowden, A. (1976) FEBS Lett.
62, 213-219.

Ureta, T. (1982) Comp. Biochem. Physiol. B: Comp. Biochem. 71, $49-$55.

196

14.

15.

16.
17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

197

Gregoriou, M., Trayer, LR, and Cornish-Bowden, A. (1983) Eur. J.
Biochem. 134,283-288.

Creighton, TE. (1983) Proteins, pp. 252-262. Freeman Publications, New
York.

Chothia, C., and Lesk, A.M. (1986) EMBO J. 5, 823-826.
Harrison, R. (1985) Ph.D. Thesis, Yale University, New Haven, CT.

Anderson, C.M., Stenkamp, RE, and Steitz, T.A. (1978) J. Mol. Biol. 123,
15-33.

Anderson, C.M., Stenkamp, R.E., McDonald,R.C., and Steitz, T.A. (1978)
J. Mol. Biol. 123, 207—219.

Bennett, W.S., Jr., and Steitz, T.A. (1980) J. Mol. Biol. 140, 211-230.

Nemat-Gorgani, M., and Wilson, J.E. (1986) Arch. Biochem. Biophys. 251,
97-103.

Schirch, D.M., and Wilson,J.E. (1987) Arch. Biochem. Biophys. 254, 385-
396.

White, T.K., and Wilson, J.E. (1989) Arch. Biochem. Biophys. 274, 375-393.
White, T.K., and Wilson, J.E. (1987) Arch. Biochem. Biophys. 259, 402-411.
Smith, A.D., and Wilson, J.E. (1991) Arch. Biochem. Biophys. 287, 359-366.
White, T.K., and Wilson, J.E. (1990) Arch. Biochem. Biophys. 277, 26-34.
Baijal, M., and Wilson, J.E. (1982) Arch. Biochem. Biophys. 218, 513-524.
Wilson, J.E. (1978) Arch. Biochem. Biophys. 185, 88-99.

Wilson, J.E. (1979) Arch. Biochem. Biophys. 196, 79-87.

Hutny,J., and Wilson, J.E. (1990) Arch. Biochem. Biophys. 283, 173-183.
Wilson, J.E. (1991) in Protein Structure Determination. Methods of

Biochemical Analysis (Suelter, C.H., Ed.), Vol. 35, pp.207-250, Wiley-
Interscience, New York.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

48.

198

Finney, K.G., Messer, J.L., DeWitt, D.L., and Wilson, J.E. (1984) J.
Biol. Chem. 259, 8232-8237.

Ferrari,R.A., Mandelstam, P., and Crane, R.K. (1959) Arch. Biochem.
Biophys. 80, 372—377.

Wilson, J.E. (1989) Prep. Biochem. 19, 13-21.

Wilkin, G.P., and Wilson, J .E. (1977) J. Neurochem. 29, 1039-1051.
Schwab, D.A., and Wilson, J .E. (1988) J. Biol. Chem. 263, 3220-3224.
Wilson, J .13., and Smith, AD. (1985) J. Biol. Chem. 260, 12838-12843.

DeWitt, D.L., Day, 1.8., Gauger, J.A., and Smith, W.L. (1982) Meth.
Enzymol. 86, 229-240.

Smith, A.D., and Wilson, J.E. (1986) J. Immunol. Meth. 94, 31-35.
Perrella, F. W. (1988) Anal. Biochem. 174, 437-447.
Chou, A.C., and Wilson, J.E. (1972) Arch. Biochem. Biophys. 151, 628-633.

Worthington, C.C., Ed. (1988) Worthington Manuel, pp. 320, Worthington
Biochemicals Corp., Freehold, NJ.

Hill, H.D., and Straka, J.G. (1988) Anal. Biochem. 170, 203-208.

Wilson, J .E., and Chung, V. (1989) Arch. Biochem. Biophys. 269, 517-
525.43.

Polakis, P.G., and Wilson, J.E. (1984) Arch. Biochem. Biophys. 234, 341-
352.

Kordossi, A.A., and Tzartos, SJ. (1987) EMBO J. 6, 1605-1610.

White, T.K., Kim, J.Y., and Wilson, J.E. (1990) Arch. Biochem. Biophys.
276, 510-517.

Tzartos, S.J., Rand, D.E., Einarson, BL, and Lindstrom, J.M. (1981) J.
Biol. Chem. 256, 8635-8645.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

61.

62.

199

Jaenicke, R. (1991) Biochemistry 30, 3147-3161.

Wilson, J.E. (1987) in Chemical Modification of Enzymes (Eyzaguirre, J .,
Ed.), pp.17l-181, Ellis Horwood Ltd., Chichester, United Kingdom.

Crawford, G.D., Correa, L., and Salvaterra, PM. (1982) Proc. Natl. Acad.
Sci. USA 79, 7031-7035.

Lillenhoj, H.-S., Choe, B.-K., and Rose, NR. (1982) Proc. Natl. Acad. Sci.
USA 79, 5061-5065.

Vartio, T., Zardi, L., Balza, E., Towbin, H., and Vaheri, A. (1982) J.
Immunol. Meth. 55, 309-318.

Sams, C.F., Hemelt, V.B., Pinkerton, F.D., Schroepfer, G.J., Jr., and
Mathews, KS. (1985) J. Biol. Chem. 260, 1185-1190.

Weldon, S.L., and Taylor, S.S. (1985) J. Biol. Chem. 260, 4203-4209.
Berzofsky, J.A. (1985) Science 229, 932-940.
Jemmerson, R., and Patterson, Y. (1986) BioTechniques 4, 18-31.

Bowie, J.U., Reidhaar-Olsen, J .F., Lirn, W.A., and Sauer, RT. (1990)
Science 247, 1306-1310.

Barlow, D.J., Edwards, MS, and Thornton, J.M. (1986) Nature 322, 747-
748.

Schirch, D.M., and Wilson, J.E. (1987) Arch. Biochem. Biophys. 257, 1-12.

Arora, K.K., Filbum, C.R., and Pedersen, PL. (1991) J. Biol. Chem. 266,
5359-5362.

Ellison, W.R., Lueck, J.D., and Fromm, H]. (1974) Biochem. Biophys. Res.
Commun. 57, 1214-1220.

CHAPTER 6

Summary and Perspectives

200

SUNIMARY AND PERSPECTIVES

The results presented in this dissertation clearly indicate that Mabs can be used
as probes of structure-function relationships within hexokinase. Seven Mabs have had
their epitopes located to short segments of polypeptide chain. By correlating the
effects of these antibodies on the ability of hexokinase to bind to mitochondria, and to
recognize mitochondrially bound hexokinase, with the location of their epitopes, it
was possible to deduce the disposition of bound hexokinase with respect to the
mitochondrial surface (1).

Mabs also served as useful "tags" to identify the origin of tryptic fragments of
hexokinase generated in the presence or absence of ligands. Cleavage at several
tryptic sites were markedly sensitive to the ligands present during digestion, thus
identifying sites in hexokinase that undergo structural rearrangement in response to
ligand binding. Results from this work indicated that binding of hexose 6—phosphates
to the allosteric site located in the N-terminal domain (2) induced conformational
changes in the C-terminal domain that were similar to those produced by binding of
substrate hexoses. The ability to produce this change is correlated with the inhibitory
potency of the hexose 6-phosphates. Protection against cleavage at the C-terminal
sites, resulting from binding of substrate hexoses, can be attributed to substrate-

induced closing of a cleft in the C-terminal domain, as previously shown to occur

201

202

upon binding of substrate hexoses to the homologous yeast hexokinase (3). Thus,
inhibitory hexose 6-phosphates are inducing conformational changes similar to cleft
closure which may account for the inhibitory action of these compounds.

Additional studies identified other regions on the N- and C-terrninal domain
that undergo conformational changes in response to binding of substrates or inhibitory
hexose 6-phosphates (or analogs). Two regions, designated F and G in the model
shown in Chapter 5 , were markedly sensitive to conformational changes induced by
binding of inhibitory hexose-6-phosphates. In addition, four of five Mabs binding to
these regions increased the Ki for the Glc-6-P analog, 1,5 -anhydroglucitol-6-P and
thus identified these regions as being involved in the transmission of the
conformational signal from the regulatory N -terminal domain to the catalytic C-
terminal domain.

Mabs will continue to be useful tools for investigating the relationship
between structure and function in hexokinase. All of the work described in this
dissertation, has been done in vitro, but can Mabs be used to study hexokinase in
vivo? Collaborative work with Dr. Ron Lynch at the University of Arizona indicates
that Mabs are useful for studying hexokinase function in vivo as well. Dr. Lynch
microinj ected rhodamine labeled hexokinase into A7R5 cells, and found that it became
associated with mitochondria. Preincubation of rhodamine labeled hexokinase with
Mab 20, previously shown to block binding of hexokinase to mitochondria in vitro
(4), prevented hexokinase from becoming associated with mitochondria. Region B,

which contains the epitope for Mab 20, is located near the membrane surface in the

203

bound state in vitro (1). Inhibition of binding by Mab 20 in vi vo indicates the epitope
is in close proximity to the membrane surface in the bound state in vivo as well.

Dr. Lynch also analyzed the subcellular distribution of hexokinase relative to
mitochondria in paraformaldehyde-fixed astrocytes using irnmunocytochemistry and
quantitative three-dimensional confocal microscopy (5). In these experiments, Mabs
2B and 21, both of which recognize the bound and soluble forms of hexokinase (l),
were used to detect hexokinase. Under basal metabolic conditions, approximately
70 % of the cellular hexokinase was associated with mitochondria. However, if prior
to fixing, the cells were incubated with 2-deoxyglucose (an inhibitor Of glycolysis),
which leads to an increase in the intracellular Glc-6-P levels, the amount of bound
hexokinase was decreased by approximately 35 % . These results indicate that the
methodology can detect subcellular changes in the distribution of hexokinase between
bound and soluble forms. Continued collaborative efforts should yield additional
information concerning the effect of metabolites on the distribution between bound
and soluble forms of hexokinase and the role it plays in governing glucose
metabolism.

Work by Xie and Wilson (6) clearly established that at least some portion of
bound hexokinase exists as a tetramer. This suggests that two populations of
hexokinase may exist when bound to mitochondria. Two populations of porin have
been identified by subfractionation of mitochondria by treatment with digitonin (7, 8).
Dorbani et al. (7), found that 60-70 % of the hexokinase was not removed by digitonin

and 40 % of the porin remained associated with the membrane fraction. One of the

204

porin papulations binds most of the hexokinase and appears to be located at contact
points between the inner and outer membranes (7,9,10). Preliminary work indicates
that Mabs binding to regions E and F may prevent binding of hexokinase to a select
population of porin. In addition, Mabs l3 and 3A5 partially block and Mabs 3B1 and
5C2 totally block tetrarner formation. Thus antibodies binding to these regions may
provide additional insight into the distribution of hexokinase between different
populations of porin as well as identify regions of hexokinase that are in close
association when bound as a tetramer.

The cDNA for rat brain hexokinase (11) has been sequenced and mutagenesis
experiments are forthcoming. The sensitivity of particular Mabs to structural
perturbations in hexokinase provides a means of assessing structural changes brought
about by mutagenesis. As demonstrated in the body of the thesis, Mabs can serve as
monitors of conformational changes resulting from ligand binding. These antibodies
will be very useful in examining mutant forms of hexokinase for possible deviations
from the normal conformational changes that occur with ligand binding.

In addition to the cDNA coding for Type I hexokinase, the cDNAs for types
11, and III hexokinase have been isolated in our laboratory as well (12,13).
Pronounced similarities between sequences of the N- and C-terminal halves of the
Type I, II, and III isozymes, and the sequence of the 50 kDa hexokinase of yeast,
support a two domain model for these enzymes. Future work will involve generating
chimeric hexokinases in which the N- and C-terminal domains will be exchanged

between the various isozymes. This should provide a means of assessing the role of

205

each domain in regulation of the various isozymes. Again, the sensitivity of some of
the Mabs to structural perturbations in the enzyme will be useful in evaluating
structural changes that may result from the fusion of the N— or C-terminal domain of
Type I hexokinase with the C- or N-terminal domains of Types II and III isozymes,
respectively. In addition, Mabs that are sensitive to ligand induced conformational
changes, can be used to evaluate the effect of ligand binding on chimeric hexokinases.
In a similar manner, Mabs that alter the kinetic parameters of Type I hexokinase can

be tested for comparable effects with the chimeras.

1)

2)

3)
4)

5)

6)

8)

9)

10)

ll)

12)

13)

REFERENCES

Smith, A., and Wilson, J .E. (1991) Arch. Biochem. Biophys. 287, 359-
366.

White, T.K., and Wilson, J.E. (1987) Arch. Biochem. Biophys. 259, 402-
411.

Harrison, R. (1985) Ph.D. Thesis, Yale University, New Haven, CT.
Smith, A.D., and Wilson, J.E. (1985) J. Biol. Chem. 260, 12838-12843.

Lynch, R.M., Fogarty, K.E., and Fay, RF. (1991) J. Cell Biol. 112, 385-
395.

Xie, G., and Wilson J.E. (1990) Arch. Biochem. Biophys. 276, 285-293.

Dorbani, L., Jancsik, V., Linden, M.,Literrier, J.F.,Nelson, B.D., and
Rendon, A. (1987) Arch. Biochem. Biophys. 252, 188-196.

Jancsik, V., Linden, M., Dorbani, L., Rendon, A., and Nelson, B.D. (1988)
Arch. Biochem. Biophys. 264, 295-301.

Kottle, M., Adam, V., Riesinger, I., Bremm, G., Bosch, W. ,Brdiczka, D.,
Sandri, G., and Panfili, E. (1988) Biochim. Biophys. Acta 935, 87-102.

Adams, V., Bosch, W., Schlegel, J., Wallimann, T., and Brdiczka, D.
(1989)Biochim. Biophys. Acta 981, 213-225.

Schwab, D.A., Wilson, J.E. (1989) Proc. Natl. Acad. Sci. USA 86, 2563-
2567.

Thelen, A.P., Wilson, J.E. (1991) Arch. Biochem. Biophys. 286, 645-651.

Schwab, D.A., Wilson, J.E. (1991) Arch. Biochem. Biophys. 285, 365-370.

206