'IIIIIIIIIIIIIIrtr

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘l/le‘llj‘fimflf MW?!

w - 1
; Li; '
; Michigan 31m”
5 University

‘b.

This is to certify that the

dissertation entitled

Domain Structure and Structure-Function Relationships
in Rat Brain Hexokinase

presented by

Tracy Keith White

has been accepted towards fulfillment
of the requirements for

Ph . D. degree in Biochemistry

 

 

9‘44 5’ $an

Major professor

Date September 22, 1989

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

A.

(“r
C.

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

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

A
.’ ‘3 t".
4a“ the =:

tr

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

__» I
l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

’ l
MSU Is An Affirmative Action/Equal Opportunity Institution
cmmma-m

DOMAIN STRUCTURE AND STRUCTURE-FUNCTION RELATIONSHIPS
IN RAT BRAIN HEXOKINASE
By

Tracy Keith White

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Biochemistry
1989

 

 

.0‘1
p
I-

. . q' u . v
I “2" '.. .
..‘-¢

 

 

. . _

A ‘->-'
o... . '
"1v“ 0.-

 

 

 

 

.. g _.
6.; . u
x... .u. .. -"‘

..~. “
~ ‘ ':‘;~
.v.. -.‘ ..

 

 

 

 

 

2' .-
w...‘ A. ....
v ‘u.':
4‘.
.
“‘ :-~‘ -..
\‘l ‘.. t...
'o
. ‘. ‘.
q
\ b l ‘I -
. s l. .-
s‘. ‘
C - -
-_ .. I'
...u J. :’~
“"
a.
h
...":A‘ '
v ‘
‘
:V‘
~~‘.
’& .
u ‘ -
‘Q
‘~ ~‘—:;.‘ _.
. ‘~“l‘—
‘
I
"s‘ , .
“‘--~ .‘
. Q'. ~

 

—‘..

,7' -3, t -.-

ABSTRACT

DOMAIN STRUCTURE AND STRUCTURE-FUNCTION RELATIONSHIPS
IN RAT BRAIN HEXOKINASE

BY

Tracy Keith White

After denaturation in 0.6 M guanidine hydrochloride, rat
brain hexokinase becomes highly susceptible to proteolysis by
trypsin. Glc-G-P and its analog, 1,5-anhydroglucitol-6-P,
selectively protect the N-terminal half of the molecule from
proteolysis. The glucose analog, N-acetylglucosamine,
selectively protects the C-terminal half of the molecule and
catalytic activity. These results demonstrate that the binding
site for the allosteric effector, Glc-6-P, lies in the N—terminal
half of the molecule and is distinct from the catalytic site. It
is demonstrated that the Glc-6-P binding site in the N-terminal
region has all the characteristics described for the allosteric
effector site on this enzyme.

Using the same general strategy, i.e., denaturation in
guanidine hydrochloride followed by limited proteolysis,
isolation of intact N- and C-terminal fragments of the enzyme was
performed, and extensive characterization of the C-terminal,
catalytically active fragment is described. It is demonstrated
that both N- and C-terminal fragments have binding sites for
hexoses, nucleotides, and the effector, Glc-6-P. Results are
discussed in terms of the proposed evolution of the enzyme, and a
model is presented which depicts the possible relationships

between ligand binding sites on brain hexokinase, and how their

n9

 

~

l."

...¢- 0

 

s
,‘uou-«I 0

“.nooo..’-.~
p a A
-~t-‘...‘--

hfinoao‘.'n.-
.1 u a at
" “M.“‘u-

 

 

 

interaction might lead to the observed regulatory properties of
the enzyme.

In addition, the enzyme was studied using differential
scanning calorimetry under a variety of conditions. Under all
conditions tested, the thermal denaturation could be described by
two independent denaturational events, assumed to correspond to
the denaturation of discrete structural domains. Correlation of
the denaturational transitions with specific regions of the
molecule was established by proteolytic analysis of enzyme in
which the domain corresponding to the first transition was

selectively denatured by a partial scan in the calorimeter.

To my parents

. .-
a o O
.- 0“ i U.-‘.

...—.-- V.
n
Hiv- .
0 U‘. -

 

 

   

   

ID'A'...’
vi... _-.
..

 

 

‘ Q
a q “A
Q'yi.‘ :‘GV
.
I. - ..’
~ .'.~
A 4
.‘c‘k . I...
.
n
4";
“‘I _

" "hoe. . .-
u.“'

-..

-"o s ~'~c A
V

 

ACKNOWLEDGMENT S

It is a pleasure to acknowledge the intellectual and
financial support provided by my major professor, Dr. John
Wilson. His dedication to his work has made him a role model
that has surely left a lasting impression. I would also like to
extend my appreciation to the faculty who served as my guidance
committee. Dr. William Deal, Dr. Robert Hausinger, Dr. Clarence
Suelter, and Dr. Alexander Tulinsky all provided encouragement or
advice at opportune times.

I would also like to thank the former and present members of
the Wilson lab whose terms have overlapped with mine. Doug,
Annette, Jan, Dave, Guochun, Al, and Hector have made the lab a
fun and interesting place to spend time. Also, special thanks to
my good.friends Catherine Wernette and Linda Gregory for their
friendship and support.

I am especially grateful to my husband, John, for his

encouragement, support, and pride in my accomplishments.

vi

 

‘0-

~.

TABLE OF CONTENTS

Page
List of Tables ............................................. xiii
List of Figures ............................................. xiv
List of Abbreviations ..................................... xviii
Chapter 1 Introduction ........................................ 1
Brain Hexokinase ................................. 2
Mammalian Hexokinase Isozymes .................... 5
Yeast Hexokinase ................................. 8
Structural Domains in Proteins ................... 9
Use of Denaturation to Define Domain
Structure in Proteins ........................... 12
References ......................................... 15
Chapter 2 Characterization of the Denaturant Induced
Inactivation of Rat Brain Hexokinase ............... 19
MATERIALS AND METHODS .............................. 22
Materials ....................................... 22
Purification of rat brain hexokinase ........... 22
Measurement of protein concentration
and enzyme activity ............................. 22
Inactivation of hexokinase ...................... 22
Preparation of N-succinimidyl(2,3-9)Hpro—
pionate labeled hexokinase ...................... 23
Treatment of N-succinimidyl(2,3-9H)pro-
pionate-labeled samples for counting ............ 23
Measurement of N-succinimidyl(2leH)pro-
pionate-labeled hexokinase adsorption
to glassware .................................... 24

vii

-n.

t"

0"

.O-~

Uvi

o.- e

a
he-..

---
'5" ‘1

VI‘T

O...-

I)

N

 

 

Chapter 3

RESULTS ............................................ 26

Inactivation of hexokinase in urea

and GuHCl ..... ...... ............................ 26
Effect of ligands on inactivation of

hexokinase by urea and GuHCl .................... 34
Transition curves for hexokinase in GuHCl ....... 37

Effect of ligands on the activity transition

of hexokinase in GuHCl .......................... 41
Adsorption of protein to glassware .............. 45
DISCUSSION ............ - ............................. 52
REFERENCES ......................................... 56
Location of the Glc-6-P Regulatory Site on a
Domain in the N-Terminus of the Enzyme ............. 58
MATERIALS AND METHODS .............................. 61
Materials ....................................... 61
Hexokinase activity and protein
determinations .................................. 61
Inactivation of hexokinase ...................... 61
Proteolysis of hexokinase ....................... 62
Electrophoretic procedures ...................... 62
Immunoblotting procedures ....................... 63
RESULTS ............................................ 65

Inactivation of hexokinase in urea and
GuHCl. ......................................... 65

Selective protection of N- and C-terminal
regions of hexokinase by Glc-6-P and the
substrate analog, N-acetylglucosamine .......... 68

Estimation of the dissociation constant

for binding of Glc—6-P to the N—terminal
domain. ........... . ............................ 76

viii

Chapter 4

Correlation between stabilization of the
N-terminal domain, inhibitory effectiveness,
and conformational change induced by binding

of hexose-G-phosphates ........................ 81

Synergism between Glc-6-P and hexose

binding sites ................................. 84
DISCUSSION ......................................... 88
REFERENCES ......................................... 91
FOOTNOTES .......................................... 93

Isolation and Characterization of the
Discrete N- and C-Terminal Halves
of Rat Brain Hexokinase ............................ 94

MATERIALS AND METHODS
Materials ..................................... 96

Hexokinase activity, inhibition studies,
and protein determination ..................... 96

Electrophoretic and immunoblotting
procedures .................................... 98

Molecular weight determination under
nondenaturing conditions ...................... 99

Preparation of the N- and C-terminal
halves of rat brain hexokinase ................ 99

Isolation and N-terminal sequencing
of the C fragment ............................ 101

Effectiveness of ligands at protecting
activity of the C fragment against
denaturation by GuHCl ........................ 102

Effectiveness of ligands at protecting N
and C fragments from proteolysis in the

presence of GuHCl ............................ 102

Elution of N and C fragments from Affi-

Gel Blue ..................................... 103
RESULTS ........................................... 105

Preparation of N and C fragments ............. 105

ix

‘Uu

I‘l
’1‘

Chapter 5

N-terminal sequence of the C fragment ........ 110

Inhibition of the C fragment by Glc-6-P
and the Glc-6-P analog, 1,5-AnG6P ............ 115

Inhibition of the C fragment by inorganic
phosphate and its analogs .................... 116

The ability of various ligands to protect
the C fragment against denaturation and
subsequent proteolysis ....................... 130

Elution of the C fragment from Affi-Gel
Blue....... .................................. 138

The ability of various ligands to protect
the N fragment against denaturation and
subsequent proteolysis ....................... 144

DISCUSSION ........................................ 148

Evolutionary relationships among the
hexokinases .................................. 148

Ligand binding sites on mammalian

hexokinases .................................. 153
ACKNOWLEDGEMENT ................................... 155
REFERENCES ........................................ 156
FOOTNOTES ......................................... 159
Anion Binding Sites in the N-Terminal Region
of Rat Brain Hexokinase ........................... 161
MATERIALS AND METHODS ............................. 163

Materials .................................... 163

Hexokinase activity and protein
determinations ............................... 163

Proteolytic digestion in the presence
of partially denaturing concentrations
of GuHCl ..................................... 164

RESULTS ........................................... 165

Effectiveness of nucleoside triphosphates
at protecting rat brain hexokinase
against inactivation in GuHCl ................ 165

X

5-.
II .
r~u

.9,
I

 

n-

‘1

Chapter 6

Effectiveness of nucleoside triphosphates

at protecting rat brain hexokinase

against partial denaturation in 0.6 M

GuHCl and subsequent tryptic digestion ....... 167

Phosphate and its analogs protect the

N-terminal half of the rat brain

hexokinase against denaturation with

GuHCl and subsequent proteolysis ............. 173

Pyrophosphate and tripolyphosphate

protect the N-terminal half of rat

brain hexokinase against denaturation

with GuHCl and subsequent proteolysis ........ 174

Anionic species that selectively protect

the N-terminal region of rat brain

hexokinase against denaturation and

subsequent proteolysis also reverse

inhibition by the Glc-6-P analog, l,5-AnG6P..179

DISCUSSION ........................................ 185
REFERENCES ........................................ 192
FOOTNOTES ......................................... 195

A Differential Scanning Calorimetric Study
of Rat Brain Hexokinase: Domain Structure

and Stability ................... . ................. 197
MATERIALS AND METHODS ............................. 199
Materials .................................... 199
Determination of hexokinase activity
and protein concentration .................... 199
DSC analysis ................................. 199
Proteolysis of thermally denatured
hexokinase ................................... 201
RESULTS ........................................... 205

DSC of rat brain hexokinase in high
ionic strength buffer ........................ 205

DSC of rat brain hexokinase in low
ionic strength buffer ........................ 212

xi

 

-0

'-

l.‘
D

(In

I

Proteolytic digestion of thermally

denatured hexokinase ......................... 216

DISCUSSION ........................................ 232
ACKNOWLEDGEMENT ................................... 2 3 5

REFERENCES ........................................ 2 3 7

FOOTNOTES ......................................... 239

Chapter 7 Summary and Perspectives .......................... 240
REFERENCES ........................................ 246

xii

.
.npu..
‘ up

P.5-1. .

 

s;
.V

.A
a. ,..
4

 

.-
s” I’ .
D

o
.v.

k -
.-
."~

0
§
U

 

(4)

q
,-

Cr:

C.

..:

 

 

 

Chapter

1.

Chapter

1.

Chapter

1.

Chapter

1.

Chapter

1.

Chapter

LIST OF TABLES

Page
Comparison of the various kinetic constants
for the mammalian hexokinase isozymes ............... 7
Concentration of denaturant required to
demonstrate 50% inactivation of hexokinase
in the presence of various ligands..... ............ 44
Inactivation of hexokinase in the presence
absence of 1 mM Glc—6-P and 10 mM GlcNAc ........... 70
Comparison of starfish hexokinase with the C
fragment of rat brain hexokinase .................. 150
Effect of nucleoside triphosphates on inactiva-
tion of rat brain hexokinase in 0.6 M GuHCl ....... 166
Calorimetric data for the denaturation of rat
brain hexokinase in high ionic strength
TET buffer, pH 8.5 ................................ 206
T;s for transitions in low ionic strength
buffer ............................................ 215

xiii

V

-u-

 

 

.
ouv

’.v

a
el-

 

~ 1v

Ibo

ChapterZ

1.

Chapter 3

1.

LIST OF FIGURES

Page
Inactivation of hexokinase in various
concentrations of urea ............................. 28
Inactivation of hexokinase in various
concentrations of GuHCl ............................ 30
Effect of temperature on the inactivation
of hexokinase in urea .............................. 33
Effect of ATP and ATP-Mg on the rate of
inactivation of hexokinase in 2.5 M urea ........... 36
Twenty-four hour GuHCl inactivation curve
for hexokinase ..................................... 39
Twenty-four hour GuHCl inactivation curve for
hexokinase in the presence of 10 mM glucose ........ 43
Adsorption of hexokinase to glassware .............. 49
GuHCl inactivation curve for hexokinase in
the presence of 2.0 mg/ml BSA ...................... 51
Inactivation of hexokinase in urea or guanidine
hydrochloride ........... . .......................... 67
Effects of guanidine hydrochloride, glucose-
6-P, and N-acetylglucosamine on digestion of
rat brain hexokinase with trypsin .................. 72
Schematic representation of tryptic cleavage
pattern in the presence and absence of
guanidine hydrochloride ............................ 75
Protection of the N-terminal domain by
increasing Glc-6-P concentrations .................. 78

xiv

1n
l'l
In

t"

'....:, I

 

I
..h

F‘!
ii.

V-'

low.

a
1
'l

V o

A...
up“.

‘1‘
C

9 i
Qt...
9r...
:--~

(.1;
O

'0

cu. :;:‘
~..t‘
V.
bQQe

“- PC:

h
.
‘A
...e
A
v."-«
V

_
I"h\.

‘-

<2haxpfl:er 4

1.

10..

11..

12.

Estimation of the K; for binding of Glc-6-P,
based on its protective effect against
proteolysis ........................................ 80

Comparison of various hexose-6-phosphates in
protecting the N-terminal domain against
proteolysis ........................................ 83

Synergistic interactions between Glc-6-P and
hexose binding sites ............................... 86

SDS-Polyacrylamide gel patterns of N and C
fragment preparations ............................. 107

Molecular sieve chromatography of the C
fragment preparation under nondenaturing
conditions ....................................... 111

Molecular weight of the catalytically active
C fragment, determined by molecular sieve
chromatography under nondenaturing conditions ..... 114

Inhibition of the C fragment by 1,5-AnG6P,
with ATP as varied substrate ........ . ............. 117

Inhibition of the C fragment by 1,5-AnG6P,
with glucose as varied substrate .................. 120

Inhibition of the C fragment by inorganic
phosphate, with ATP as varied substrate ........... 123

Inhibition of the C fragment by inorganic
phosphate, with glucose as varied substrate ....... 124

Inhibition of the C fragment by inorganic
phosphate and its analogs ......................... 127

Effectiveness of inorganic phosphate and its
analogs at reversing inhibition by 1,5-AnG6P ...... 129

Effectiveness of various ligands in protecting
the C fragment against denaturation in GuHCl ...... 132

Effectiveness of various ligands in protecting
the C fragment against proteolysis in 0.6 M
GuHCl.... ..... ... ............. . .................. 134

Effectiveness of inorganic phosphate and its

analogs at protecting the C fragment against
denaturation by GuHCl and subsequent proteolysis..137

XV

 

r
'
v

 

o“—
.

1"

" ... ‘

-.‘v‘

-" a
...-I;
-.DU'
.

.u |
... o
.
.OR‘
‘nn.
3,--.
...-riv-
.

‘ i

.F
-..‘3 .
.-
-.Oo.
' 0—
-..U.

7".
.d. "
hug',
on]

- i

l3.
14”

215.

1.6.

Chapters

1.

ChapterG

l.

Elution of the C fragment from Affi-Gel Blue ...... 141
Elution of the N fragment from Affi-Gel Blue ...... 143

Effectiveness of various ligands at protecting
the N fragment against denaturation with GuHCl
and subsequent proteolysis ........................ 146

Proposed evolutionary relationship between
the hexokinases ................................... 152

Effects of nucleoside triphosphates and their
Mg” chelates on proteolysis of rat brain
hexokinase in 0.6 M guanidine hydrochloride ....... 169

Immunoblot analysis of tryptic digestion
products .......................................... 171

Effect of P1 and analogs on proteolysis of
rat brain hexokinase in 0.6 M guanidine
hydrochloride.... ................................. 175

Effect of P1 and homologs on proteolysis of
rat brain hexokinase in 0.6 M guanidine
hydrochloride ..................................... 178

Effectiveness of P3 and various nucleotide-Mg”
complexes at reversing inhibition by the
Glc-6-P analog, 1,5-AnGGP ....... . ................. 180

Effectiveness of P; and its homologs at reversing
inhibition by the Glc-6—P analog, 1,5-AnG6P ....... 184

Schematic representation of ligand binding
sites on rat brain hexokinase ..................... 190

Location of tryptic cleavage sites and
epitopes for monoclonal antibodies ................ 204

Representative DSC scans of rat brain
hexokinase ........................................ 207

Deconvolution analysis of DSC results obtained
in high ionic strength TET buffer ................. 209

DSC of rat brain hexokinase in 10 mM Glc
and low ionic strength TET buffer ................. 214

xvi

o‘-

-. -~ V.

a“ . . ... .3

.5 C. a. no.
3“

DSC of rat brain hexokinase in 1 mM Glc-6-P
and high ionic strength TET buffer ................ 219

SDS acrylamide gel analysis of tryptic
digestion products of native and partially
denatured rat brain hexokinase .................... 221

Immunoblot analysis of products from tryptic
digestion of partially denatured rat
brain hexokinase .................................. 224

DSC of rat brain hexokinase in low ionic
strength TET buffer containing 5 mM ATP-Mg”
and 10 mM GlcNAc .................................. 228

SDS gel electrophoresis and immunoblot analysis

of products from tryptic digestion of
partially denatured rat brain hexokinase .......... 231

xvii

l , S—AnGGP
13518
EHSC:
EXIFUB
Frc-G-P
Ghal.
(311:
Gal-G-P
G1 cNAc
Glc-G-P
GhaEHZI
Hepes

HIKE

Man-6-9
PGE
PMSF
SDS
TG

TPCK

LI ST OF ABBREVIATIONS

l,5-anhydroglucitol 6-phosphate

bovine serum albumin

differential scanning calorimetry
5,5-dithiobis-(2-nitrobenzoate)

Fructose 6-phosphate

galactose

glucose

galactose 6-phosphate

N-acetylglucosamine

glucose 6-phosphate

guanidine hydrochloride
4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
10 mM Hepes, 0.5 mM EDTA, 10 mM TG

mannose

mannose 6-phosphate

0.1 M Potassium Phosphate, 10 mM Glc, 0.5 mM EDTA
phenylmethylsulfonyl fluoride

sodium dodecyl sulfate

monothioglycerol

L-1-tosylamide-2-phenylethyl chloromethyl ketone

xviii

Chapter 1

Introduction

The phosphorylation of glucose by hexokinase represents the
first step in a variety of metabolic pathways, and the enzyme
represents one of several important control points in glucose
metabolism (1) . A major goal of the present work has been to
characterize the domain structure of the rat brain enzyme, and
iirnanfloiguously assign various functions, i.e., activity and
rregnilation, to specific structural components of the protein. In
general, this was accomplished by using denaturation, induced by
organic compounds or heat, to selectively denature catalytic or
regulatory domains of the enzyme. The focus of the current
Chapter is to provide a brief introduction to what was understood
CanceI‘ning the structure and function of hexokinases at the time
this research was initiated. This section will also provide an
introduction to some additional relevant literature. Subsequent
c“hapters will give a more detailed description of the studies
perfiormed in an attempt to gain understanding of structure-

f - . . . . .
uncItion relationships in brain hexokinase.

 

 

.5 ij‘ "

..Aocfl

..., ‘- nooo'i’

   

 

Q .
i-
a
I - a
9 up. .6 aqlvvn.’
-1-—. by vI-U-v
o
I . ‘
a..." ,. .’~

u...

 

’ Quay.

‘II a... A
I' a

‘0‘ ‘oe.“

n

 

 

 

Va

 

 

I,'.
..‘M F.

 

l.‘ .b‘
't I
. A
‘o c‘ .
"~~e>. A
§"
a
P I
'A. “I

Brain Hexokinase

 

Rat brain hexokinase (ATP: D-hexose 6-phosphotransferase,

EC 2.7.1.1.), a monomer of approximately 100 kDa (2), catalyzes
the first step of glucose metabolism in the brain, converting
glucose to glucose-6-P using ATP as the phosphoryl donor. As
might be expected, the enzyme is subject to complex regulatory
controls, including inhibition by the reaction product, glucose-
6—P, and antagonism of this inhibition by P1 (1) .

The sensitivity of the brain enzyme to physiologically
relevant levels of Glc-6-P was first observed by Weil-Malherbe
and Bone (3), and the subsequent attempts of several laboratories
to characterize this inhibition has led to an involved debate
concerning the actual nature of the Glc-6-P binding site. The
scope of this debate is outlined briefly here. The presence of
an allosteric regulatory binding site for Glc-6-P was initially
proposed by Crane and $013 (4) upon the observation of dissimilar
specificities for hexose (5) and hexose-G-P (4) binding sites.
F01: example, mannose is nearly as effective as glucose in serving
as a substrate in the hexokinase catalyzed reaction, with both
compounds demonstrating Kns in the low micromolar range. On the
Other hand, it is not true that Man-6-P is as effective as Glc-6-
P in inhibiting the same reaction. They interpreted these
results as suggesting that the substrate binding site does not
Serve as the site of inhibition via Glc-6-P, and that a separate,

i°e- allosteric, site must exist.

On the other hand, Fromm and his colleagues (6,7) have

 

.
' or
...;R .
.... i -
': “..-
.0“.

.

-- :Av " a
." a
E...

 

. .-
pnp:0~‘fi .‘
"...-corn!

fl
no...” ~fi~
-....“Ub -—‘

 

.v'.‘ v.

A.
n.1,. . g

' v

‘I

.
“' V. vow

 

\I—
- 0 c, M Q
" ' ' to.

.
'F'v:h.,.' P
h‘..r‘_--.. ~

 

.
. O

J; ." “a...
-.. I.

I- _
Vt”.-.-

 

Ve
‘1'.
o .’
"‘-o .:~:. '
.‘u.“..'
... .
u
“u, :_.:.:~':
.:---‘v
at; ‘
"..., D .
'v' ‘s 1
I.
Q.
'. ‘~
\.I‘._‘ 3-.“
V‘,
‘

 

5:
.“ \
o..‘ ‘ . I
' a .
.... q ... ‘
I ~ -
I .
...
\T
A:
Vs

 

“
A;
y‘
V“.
‘ N
._H.‘.,
"'r “‘1.

3
maintained that the binding site for Glc-6-P overlaps the binding
site for the substrate ATP, resulting in direct inhibition at the
active site. Their model proposes that, after phosphoryl bond
formation, the hexose moiety of the hexose-6-P moves to a

distinct site, while the phosphate moiety remains anchored in the
y—phosphate position of the ATP binding site (7) .

Both of these models can account for the non-competitive
nature of Glc-6-P binding vs. Glc, the competitive binding of
Glc-G-P vs ATP (8), and can also be reconciled with the
observation by at least two laboratories (9,10) that glucose is
able to enhance the binding of Glc-6-P, providing evidence that
the two compounds can bind to the enzyme simultaneously.
However, a variety of binding experiments and kinetic analyses
have continued to argue both for and against the presence of an
allosteric Glc-6-P binding site (e.g., 6,7,9 and 11), and the
ruatter remains unresolved. The view adopted by our laboratory,
which appears to be the most consistent with the available
evidence, is that the enzyme contains an allosteric Glc-6-P
binding site, i.e., a site distinct from the active site,

In addition to the effects of Glc-6-P, the enzyme is
regulated by P1 through the ability of this ligand to antagonize
the binding of Glc-6-P (12), an effect which has been attributed
to the mutually exclusive binding of these two ligands (10). In

the absence of Glc-6-P, P1 has no effect on brain hexokinase

activity, other than an inhibitory effect which occurs at

. e F
...”... ”3"
--'"t:-~.. ‘

—. ‘

er's 7:3 -..

 

n I so OD V

no- on un- .

‘C~l.-.Oe. ‘

"av-.... .

c

m.- ‘:‘ ’7-

' \

‘" "‘0 to»..-
0 a

: ""°;von V;

horn-..-, ‘v

'“° C“; “‘6:

."“ Ivy... ‘

" FAQ. ~

I- 3-.
‘.'."'O‘ Ugo

....a A:

p...
n.‘.' V. ‘.'~‘
u...‘ .
0‘ I. ‘
......gh

1, .

Ag.
'

Lune. 1
Th _ . .
n. : o. .
\ ." ho. a.

v
'U... - ...

 

.
II
n‘. 9".“
... A-
1.. .ar
"e .
\4 'I
I ..‘e ‘Gc‘s‘ .
s..‘.
h
‘0
l
H . I ‘Q
.... ( ‘
(J, .
3'.
t._:“‘
1“.

 

“ e
.'.' s A
‘ I
.‘VA. A‘
's
o
n..
n
'H. A...
~‘- ‘ '
u“

 

4
unpfliysiologically high levels of P1 09 - 35 mM) and presumably
occurs via interaction with the ATP binding site (13) .

There has also been some evidence which suggests that the
reaction product, ADP, as well as other nucleotides may play a
role in the regulation of hexokinase (13-16) . The non-
cxanpetitive inhibition of hexokinase by ADP vs. both glucose and
ATP has prompted proposals that this compound binds to an
allosteric regulatory site on the enzyme (8) . However, there has
been some debate concerning the presence of an allosteric
nucleotide binding site (e.g., 13,15 and 17), and the actual
nature of nucleotide regulation of the enzyme has not been
resolved.

Another factor thought to influence hexokinase activity in
vivo is the ability of this enzyme to reversibly bind to the
outer mitochondrial membrane. The enzyme interacts with the
outer mitochondrial membrane via its association with porin
(18, 19), an interaction which is thought to give the enzyme
Preferential access to mitochondrially generated ATP (20,21) .
The enzyme can be solubilized from the membrane by Glc-6-P (22) ,
and the resulting solubilized enzyme demonstrates a reduced Kn
f°r ATP (23) . The fact that Glc-6-P is a potent inhibitor of
enzyme activity, and can induce solubilization of the enzyme from
mitochondria makes it clear that the role of this ligand in the
re‘3‘~llation of hexokinase is complex.

Rat brain hexokinase is known to exist in several

c°nf0rmations which can be induced by the binding of various

5
ligands (24-26) . For example, the ability of a variety of
hexoses to protect against various inactivating agents
(chymotrypsin, glutaraldehyde, heat, and DTNB) has been
correlated with the ability of those compounds to serve as a
substrate in the hexokinase catalyzed reaction and with their
ability to induce a specific conformational change in the enzyme.
Thus, carbohydrates such as glucose or mannose, which are good
substrates for hexokinase, induce a specific conformational
change and afford the enzyme a significant degree of protection
against inactivation by these agents. Those hexoses not able to
protect the enzyme, i.e., those which are not good substrates,
are presumed to be ineffective because of their inability to
induce the conformational change that has been proposed as

necessary for catalysis .

Emlian Hexokinase Isozymes
Hexokinase isolated from brain tissue is only one of four

isozymes which exist in mammalian tissues (27) . The isozymes are
designated most commonly by Roman numerals I-IV in order of
increased electrophoretic mobility toward the anode when analyzed
in a starch gel (28) . Alternatively, isozymes are designated as
A~D based on their order of elution from DEAE cellulose columns

(29). Brain hexokinase represents the type I (or type A) form of

the enzyme. The isozymes I-III are similar but vary

Significantly from isozyme IV in both kinetic and structural

Preperties. Isozymes I-III are all monomers with a molecular

u l‘
' ..uv ""'-‘
.
lll' “..s'
_‘,.I....' vow—A
c
u .. .
. f I ;’~ I:
it Add, W‘s-O --

 

 

 

o...._-‘
...
. 'Igv..:-"-

‘~~A-‘ '

 

 

;I
v.5“ .’ ~
“ .' ;.
Qt.‘.‘~‘.,
b.
V
D .
..6

 

«In ~
" :
n ‘An.
~-
Iva e
5
‘
.u
‘i ::e
.,e ,_
I may
‘
I“\
‘c-. a ‘
. .~ .
“5‘ A. .
U ‘ “A
s._
V
‘N
5‘ :'c
...‘:~,' I;
1
o‘ ‘En-
.

 

6
mass of approximately 100 kDa, and demonstrate potent inhibition
by the reaction product G1c-6-P. On the other hand, the type IV
isozyme, often referred to as glucokinase, is a monomer of only
50 kDa, and is not inhibited by Glc-6-P (30). In addition, the
1g for glucose for types I-III is significantly lower than that
demonstrated for glucokinase (28) . Comparison of the apparent
kinetic constants for the various isozymes of hexokinase is
presented in Table I.

In addition, although types I-III are all inhibited by
physiologically relevant levels of Glc-6-P, there are some
notable differences concerning the nature of this inhibition
between types I and II. For example, differences in the time
required for Glc-6-P inhibition have been reported by Kosow eg.
Ll. (31,32), and indicate that the type II enzyme demonstrates a
markedly slower response to the effects of this ligand.
Furthermore, G1c-6-P inhibition of the type II enzyme is not
relieved by P1, and P1 itself has been demonstrated to be a rather
POtent inhibitor of the type II enzyme from rat muscle (33) .

The ability of various isozymes to respond to the presence
ligands, e.g. Glc-6-P and P1, has been related to the type of
tissues in which they reside, and the primary means of obtaining
energy in those tissues (1) . For example, brain tissue, which
I‘Elies almost exclusively on blood-borne glucose as an energy

Substrate, may rely on the activation of the type I hexokinase
Provided by increased levels of P1 which occur during periods of
high energy demand (1) . On the other hand, muscle tissue, which

contains predominantly the type II isozyme, is not dependent on

o
‘- ivy-v.-
.

. .
a up-

'2’: ' " ‘3:

“can“ ‘lvn"

::.:.‘o:'
DOOM D'.

f ‘9-

‘3‘:

v .

I i A ’ G

- ...-:-:
n t
'5. 3.:

Table I. Comparison of the various kinetic constants for the
mammalian hexokinase isozymes‘.

 

Hexokinase

Parameter I II III IV

Apparent kinetic constants (mM)

K. glucose 0.04 0.13 0.02 4.5
K_ATP 0.42 0.70 1.29 0.49
K. glc-6-P

vs. ATP 0.026 0.021 0.074 15

 

‘Table adapted from Ureta (30), and references therein.

.uol 0"“ ~

...-‘ .3...

...":O'lll 5‘

”..OluouvUO U‘

... Allfi‘.‘.
fissu ll

‘- "~'II'-“

 

o.

g.

‘Q: .:' .
..‘Oo n:

"0-
\ '3
A
~'.l_.: ;
. a.

N a'.

I.

I
‘.”~..“
'wota‘

. ¢.§

 

 

 

 

8

blood born glucose as a source of energy, and a similar

activation of hexokinase during periods of high energy demand is

not essential .

Yeast Hexokinase

Yeast hexokinase is among the most extensively studied non-

manmalian hexokinases. There are two naturally occurring

isozymes of yeast hexokinase, referred to as PI and PII which can
be distinguished by their order of elution from a DEAE cellulose

column when eluted using a pH gradient (34) . Alternatively, the

isozymes are designated A and B, corresponding to the PI and PII

isozymes, respectively (35) . Both the A and B isozymes have

subunit molecular weights of approximately 50 kDa, and can exist

as dimers containing two identical (1 or two identical 8 subunits,

respectively (36) . The dimer can be dissociated and suffer no

activity loss under a variety of conditions (37), and the monomer

may be the primary reactive species present under most assay

conditions (38) . In contrast to the regulatory properties

observed in the mammalian enzyme, the yeast enzyme is not
inhibited by physiologically relevant levels of Glc-6-P (34) .

Extensive structural studies of the yeast enzyme have been
performed by Steitz and his colleagues using x-ray

Crystallography (39-43) . Crystallographic data indicate that the

Protein is composed of two lobes, with the cleft between the

lobes containing the binding site for glucose. Upon the binding

 

9

of glucose, the conformation of the enzyme is altered inducing a
closing of the cleft (39) . In the dimeric form, the enzyme
exhibits a non-symmetrical arrangement of subunits, with
different surfaces of each of the identical subunits interacting.

More recently, the amino acid sequences of the yeast
hexokinases have been obtained (44,45). Stachelek gt. a_l_. (44)
cloned BK A and BK B genes from a yeast genomic DNA library.
Comparison of the amino acid sequences deduced from this data

indicated that the two isozymes share 378 identical residues out

of 485 total residues.

Structural Domains in Proteins

In addition to the several types of secondary structure
known to exist in proteins, a higher level of structural
complexity appears to be common, or perhaps universal in larger
proteins (46-48) . Many large proteins are divided into
structural regions referred to as domains, defined by Blake and
Johnson (46) as 'a folded structure looking like a complete small
protein molecule". It is often the case that domains are formed
from a continuous segment of polypeptide chain and that a single,
relatively unstructured region of polypeptide links domains
(48, 49) .

It has been proposed that structural domains may act as
"folding units” during the synthesis of proteins (46,50) .
Wetlaufer (50) has presented evidence from numerous protein

folding experiments which have demonstrated the self-assembly of

..n' D
.
.....-

L...

..H."

 

 

no.
\ ’
h'oo.

..A

-.,'--

u
‘1’ c
o .

t...
a ‘ -
...

 

 

 

 

 

 

 

 

 

_ ... ...de

10
jJndividual domains in a variety of denatured proteins. In
addition, domains appear to play an important role in protein
function; it is often the case that the structural domains of a
Ixrotein can be associated with particular functional
characteristics, and the movement of domains in a protein
:redative to one another provides a range of motion which may be
ianortant in various protein activities (47).

Because domains are often connected by relatively
Iinstructured regions of polypeptide chain, connecting regions are
susceptible to proteolytic attack (49), and limited proteolysis
(:an often be used as an initial step in the isolation of the
ciomains of a protein. For example, Ch-tetrahydrofolate synthase
:isolated from rat liver is a multifunctional enzyme which
<3atalyzes three reactions in one carbon metabolism; these
:reactions will be referred to here as simply synthetase,
cyclohydrolase, and dehydrogenase activities. Experiments
employing differential scanning calorimetry have indicated that
the enzyme is composed.of two structural domains (51). UBing
limited chymotryptic digestion followed by HPLC molecular sieve
cmromatography, Villar gt. El- (51) were able to separate the two
‘dbmains of this multifunctional enzyme, and demonstrate that one
<domain was associated with synthetase activity, while the other
icontained both cyclohydrolase and dehydrogenase activities.
‘Furthermore, the isolated domains demonstrated activities with Kn

values which were the same as those observed in the intact

enzyme.

 

 

as.“
o

 

n..-
“on

 

 

 

-
‘ \AI
'v

 

 

ll.’
‘l,

 

 

 

11

Another example of a protein in which structural/functional
cionains are easily isolated using limited proteolysis is provided
11y rat liver sulfite oxidase (52-54). The enzyme contains two
nuolybdenums and two hemes per molecule, and consists of distinct
rmolybdenum and heme containing domains, each of which function
waith discrete activities in the intact enzyme (54). Limited
ciigestion with trypsin, followed by molecular sieve
«chromatography results in separation of these two domains in a
:Eunctional state. Comparison of circular dichroism spectra for
tzhe intact enzyme and isolated domains indicated that proteolysis
<>f the enzyme and separation of the domains does not result in
reignificant structural alterations in the domains (53).

It has been of obvious interest to isolate the functional
tinits of brain hexokinase in a similar manner and by so doing
provide unambiguous information concerning domain structure and
function. Unfortunately, structural/functional domains of rat
ibrain hexokinase have not been as easy to isolate and
(maracterize. In previous experiments performed by Polakis and
Vfilson (55), limited proteolysis of the enzyme resulted in the
generation of three stable fragments via two tryptic cleavage
sites in the enzyme; a 10 kDa N-terminal fragment, a 50 kDa
icentral fragment, and a 40 kDa C-terminal fragment. The N-
‘terminal 10 kDa fragment has been shown to be critical for the
interaction of hexokinase with the mitochondria (55,56), and
binding sites for Glc (57) and ATP (58) have been located in the

40 kDa C-terminal fragment. The 50 kDa fragment was postulated

., AA».

" ...».

 

. ...:
.0 ‘
”OI"

 

I... Al
(I-
~“"

 

 

‘-
I‘
u.
-
3‘“:
‘M..
t“
.
Q ‘ F
‘ on

'1';" .
v...”

 

 

 

 

 

 

 

t
1i 8'
\ e
V
...5
A“
.I
Q I.
I
o . “

 

12
to contain the allosteric Glc-6-P binding site (57) . Although
attempts were made to isolate the fragments obtained following
limited tryptic digestion, the fragments remained closely
associated and were not separable under native conditions. The
"nicked” 100 kDa enzyme maintained nearly 80% of its original

activity, and suffered no change in K. for glucose, or K1 for the

inhibitor, Glc-6-P (55).

Use of Denaturation to Define Domain Structure of Proteins.

As mentioned in the preceding section, it is thought that
domains may serve as "folding units" during the translation of
mRNA in protein synthesis (46,50) . To test this hypothesis,
experiments have been performed on a variety of proteins in which
the protein is denatured and renatured while observing various
properties which reflect native structure (e.g., activity, CD
spectra). The results, some of which will be discussed below,
clearly indicate that the domains of proteins often act as
independently unfolding/ folding units, which may be
differentially affected by the concentration of denaturant used,
the presence of ligands, or other factors (reviewed in 50) . It
is clear that folding modules correspond to domains in many
globular proteins, and, of the most interest in the present
context, the study of denaturant induced unfolding of a protein
has the potential to provide some information about domain

structure and interactions and contribute to an understanding of

Protein structure and function.

 

 

...-o-
.

one-J

.‘u-V

 

. -
-uc.

 

 

 

I n
o.-

'n. .

.‘I. .

...

:‘l
.I-

 

'on.

 

-
"v.

M
V.

- v
. 1

'n

v
v‘ I

 

-
‘ I
Ob.‘

 

 

 

 

 

‘-
\. ‘A
\‘t‘
I
.-
>\\.
'5

 

 

13

Urea and GuHCl are organic compounds which can induce
complete denaturation of most proteins at high enough
cxancentrations, and both have been used in the study of
denaturation in proteins (59) . For example, Goldberg and his
cxolleagues (60-62) have extensively characterized the GuHCl
induced unfolding of the B2 subunit of E. Q}; tryptophan
synthetase, and enzyme which is cleaved into two non-overlapping
fragments following limited digestion with trypsin (60). The
linfolding of the intact enzyme is a multi-state process, as
:indicated by observing both circular dichroism and fluorescence
spectra during denaturation (62) . Upon isolation of the
Iproteolytic fragments, and comparison of their unfolding
Ixroperties to those of the intact enzyme, Zetina and Goldberg
(62) were able to elucidate the order of unfolding of the domains
(If the enzyme, and demonstrate that the multi-state denaturation
(if the enzyme corresponds to the successive unfolding of the
iJndividual domains. Furthermore, they were able to demonstrate
the importance of inter-domain interaction in the overall

stability of the intact enzyme.

Examination of the unfolding of the a subunit of tryptophan
synthetase was examined by another laboratory (63), and similar
results were obtained. This subunit can also be proteolytically
divided into two fragments, and biphasic unfolding behavior was
observed during denaturation of the intact enzyme in GuHCl (64)

or urea (65) when circular dichroism was monitored. Comparison

 

14
of the unfolding properties of the isolated domains and the
jJitact enzyme provided information on the order of domain
cumfolding, and inter-domain interaction in the intact enzyme
(63).

In addition to denaturation induced by organic compounds,
ciifferential scanning calorimetry, or DSC, provides a means by
tdhich the thermal denaturation of proteins can be analyzed, and
.is an important tool for observing the independent denaturation
<of the domains of proteins. Furthermore, DSC analysis has the
potential to provide substantial information concerning the
<effects of ligands on domain stability and domain interaction
(66-68).

In general, the unfolding of individual domains are
iJndicated by individual endothermic transitions in a DSC
tJnermogram, and, when the denaturation of individual domains
<>verlap, software is available which can elucidate the number of
Chomains represented by a single endotherm (68). The addition of
ILigands has the potential to differentially affect one transition
in the thermogram, and provide additional information concerning
protein function. Cytochrome c peroxidase, lactoperoxidase, and
aspartate transcarbamoylase are just a few of the proteins whose

structure has been characterized using this powerful technique

(69-71).

 

1C).

11”

12.

13.

14.

15.

16.

REFERENCES

Wilson, J.E. (1984) i3 Regulation of Carbohydrate Metabolism
(R. Beitner, Ed.), pp. 45-85, CRC Press, Boca Raton, FL.

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

 

Weil-Malherbe, H., and Bone, A.D. (1951) Biochem. g. 42,
339-347.

Crane, R.K., and $013, A. (1954) g. Biol. Chem. 210, 597-
606.

 

 

5013, A., and Crane, R.K. (1954) g. Biol. Chem. 210, 581-
595.

 

Casazza, J.P., and Fromm, H.J. (1976) Arch. Biochem.
Biophygw 177, 480-487.

 

 

Solheim, L.P., and Fromm, H.J. (1981) Arch. Biochem.
Biophys. 211, 92-99.

 

 

Fromm, H.J. (1981) $3 The Regulation of Carbohydrate
Formation and Utilization in Mammals (C.M. Veneziale, Ed.),
University Park Press, Baltimore.

Lazo, P.A., $013, A., and Wilson, J.E. (1980) g. Biol. Chem.
255, 7548-7551.

Ellison, W.R., Lueck, J.D., and Fromm, H.J. (1975) g. Biol.
Chem. 250, 1864-1871.

Mehta, A., Jarori, G.K., and Kenkare, U.W. (1988) g. Biol.
Chem. 263, 15492-15497.

Ellison, W.R., Lueck, J.D., and Fromm, H.J. (1974) Biochem.
Biophys. Res. Commun. 51, 1214-1220.

Purich, D.L., and Fromm, H.J. (1971) g. Biol. Chem. 246,
3456-3463.

 

Ureta, T., Lazo, P.A., and $013, A. (1985) Arch. Biochem.
Biophys. 239, 315-319.

 

 

Ning, J., Purich, D.L., and Fromm, H.J. (1969) g. Biol.
Chem. 244, 3840-3846.

Grossbard, L., and Schimke, R.T. (1966) g. Biol. Chem. 241,
3546-3560.

 

15

 

 

 

 

 

 

1'7.

218.

219.

2C).

21.

22L

22L

124.

225.
126.

2'7.

2E3.

253.

30.

31.

32.

33.

34.

35.

16

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

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

 

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

Gots, R.E., Gorin, F.E., and Bessman, S.P. (1972) Biochem.
Biophys. Res. Commun. 42, 1249-1255.

Inui, M., and Ishibashi, S. (1979) g. Biochem. 85, 1151-
1156.

Rose, I.A., and Warms, J.V.B. (1967) g. Biol. Chem. 242,
1635-1645.

Wilson, J.E. (1968) g. Biol. Chem. 243, 3640-3647.

 

Wilson, J.E. (1973) Arch. Biochem. Biophys. 159, 543-544.

 

 

Wilson, J.E. (1978) Arch. Biochem. Biophys. 185, 88-99.

 

 

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

 

Katzen, H.M., Soderman, D.D., and Nitowsky, H.M. (1965)
Biochem. Biophys. Res. Commun. 12, 377-382.

 

0P0,
Commun. 38, 376.

Katzen, H.M., and Schimke (1965) Proc. Natl. Acad. Sci.

 

 

U.S.A. 5_4, 1218-1225.

Gonzalez, C., Ureta, T., Sanchez, R., and Niemeyer, H.
(1964) Biochem. Biophys. Res. Commun. 16, 347-352.

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

Kosow, D.P., Oski, P.A., Warms, J.V.B., and Rose, I.A.
(1973) Arch. Biochem. Biophys. 157, 114-124.

Kosow, D and Rose, I.A. (1972) Biochem. Biophys. Res.

 

Lueck, J.D., and Fromm, H.J. (1974) g. Biol. Chem. 249,
1341-1347.

 

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

Lazarus, N.R., Ramel, A.H., Rustum, Y.M., and Barnard, E.A.
(1966) Biochemistgypg, 4003.

 

£36.

137.

138.

39h

40.

41.

42L

443.

444.

‘15.

4(5.
473
48.

49.

50.

51.

52.

53.

54.

17

Purich, D.L., Fromm, H.J., and Rudolph, F.B. (1973) Adv.
Enzympl. 32, 249-326.

Schulze, I.T., and Colowick, S.P. (1969) 2. Biol. Chem. 244,
2306-2316.

Furman, T.C., and Neet, K.E. (1983) 2. Biol. Chem. 258,
4930-4936.

 

Bennett, W.S., Jr., and Steitz, T.A. (1977) Proc. Fed. Amer.
Soc. 352. Biol. 32, 667.

Anderson, C.M., Stenkamp, R.E., McDonald, R.C., and Steitz,
T.A. (1978) 2. Mol. Biol. 123, 207-219.

 

Anderson, C.M., Stenkamp, R.E., and Steitz, T.A. (1978) 2.
Mol. Biol. 123, 15-33.

 

Bennett, W.S., Jr., and Steitz, T.A. (1980) 2. Mol. Biol.
140, 211-230.

Steitz, T.A., Fletterick, R.J., Anderson, W.F., and
Anderson, C.M. (1976) 2. Mol. Biol. 104, 197-222.

Stachelek, C., Stachelek, J., Swan, J., Botstein, D., and
Konigsberg, W. (1986) Nucl. Acids Res. 22, 945-963.

Kopetzki, E., Entian, K.-D., and Mecke, D. (1985) Gene 32,
95-102.

Blake, C.C., and Johnson, L.N. (1984) TIBS 2, 147-151.
Chothia, C. (1984) Ann. Rev. Biochem. 33, 537-572.

Babul, J. (1987) Arch. Biol. Med. 352. 22, 333-341.

 

Schulz, G.E., and Schirmer, R.H. (1979) Principles 92
Protein Structure, p. 84, Springer—Verlag, New York/Berlin.

Wetlaufer, D.B. (1981) Adv. Prot. Chem. 32, 61-92.

Villar, E., Schuster, E., Peterson, D., and Schirch, V.
(1985) 2. Biol. Chem. 260, 2245-2252.

Johnson, J.L., and Rajagopalan, K.V. (1977) 2. Biol. Chem.
252, 2017-2052.

Southerland, W.M., and Winge, D.R. and Rajagopalan (1978) 2.
Biol. Chem. 253, 8747-8752.

 

Southerland, W.M., and Rajagopalan, K.V. (1978) 2. Biol.
Chem. 253, 8753-8758.

555.

£56.

557.

58.

59.

60.

61.

62.

62L

€54.

€55.

66L

67.

68.

69.

70.

71.

18

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

 

 

Wilson, J.E., and Smith, A.D. (1985) 2. Biol. Chem. 2 0,
12838-12843.

 

 

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.

 

 

Bibbard, L;S., and Tulinsky, A. (1978) Biochemistry, 11,
5460-5468.

 

Hogberg-Raibaud A., and Goldberg, M.E. (1977) Proc. Natl.
Acad. Sci. U.S.A. 12, 442-446.

Hogberg-Raibaud, A., and Goldberg, M.E. (1977) Biochemistry
12, 4014-4020.

 

Zetina, C.R., and Goldberg, M.E. (1980) 2. Mol. Biol. 137,
401-414.

Miles, E.W., Yutani, K, and Ogasahara, K. (1982)
Biochemistry 21, 2586-2592.

Yutani, K., Ogasahara, K., and Sugino, Y. (1980) 2. Mol.
Biol. 144, 455-465.

Crisanti, M., and Matthews, C.R. (1981) Biochemistry 22,
2700-2706.

 

Krishnan, K.S., and Brandts, J.E. (1978) 12 Methods in
Enzymology (airs, C.H.W., and Timasheff, S.N., Eds.), Vol.
49, pp. 3-14, Academic Press, New York.

Donovan, J.W. (1984) Trends 12 Biochemical Sciences 2, 340-
344.

 

Brandts, J.E., Hu, C.Q., Lin, L.-N., and Mas, M.T. (1989)
Biochemistry, in press.

Kresheck, G.C., and Erman, J.E. (1988) Biochemistry 21,
2490-2496.

 

Pfeil, W. and Ohlsson, P.I. (1986) Biochim. 22. Biophys.
Acta 872, 72-75.

Vickers, L.P., Donovan, J.W., and Schachman, R.K. (1978) 21
Biol. Chem. 253, 8493-8498.

 

:3
n-

. '6‘
“”e r
van-

0‘.

';‘-o :-
.an

‘
.v 6 A
:«3-9.

 

3391 to a

 

t:‘nnn.'.. 4 p
N» ‘

v». ..‘. u ‘I
.:-; ...... a,“
.... ....“ CV“!
u-

‘\

a. . .
V. ‘ “We“ .‘ A
"“U“Qb ‘

¢::.;II§ ~A' a?“
. uvouv =VJ“.‘

~. ‘an.

"5“! s:

...

 

‘u.. .

a

.' ‘ ‘ ‘

“"941 f: n
‘m

I

.

. . ‘

.H

a. _

"‘ (ti-7)
5‘: I
.2, ‘

CHAPTER 2

Characterization of the Denaturant Induced

Inactivation of Rat Brain Hexokinase

Since White (1) and Anfinsen and Haber (2) first
ciemonstrated that denatured and reduced ribonuclease could be
:renatured to a fully active protein, many proteins have
<1emonstrated an ability to spontaneously refold when removed from
cienaturing conditions. At this time, much additional evidence
lias accumulated to suggest that the primary structure of many
Eiroteins governs their folding, to yield their ultimate secondary
and tertiary structures.

The study of denaturation and renaturation in globular
‘proteins has provided some insight into the mechanisms involved
in protein folding, as well as demonstrating some of the
relationships between structure and function in these proteins

(3). It has been noted that many globular, single-chain proteins
can be divided into distinct structural regions referred to as
cbmains (4-7), and that domains may demonstrate differing

stabilities in denaturing conditions (3). As indicated in the

19

 

,..,...‘..~o c A“
“..i‘."-vta’

, .
. .- o a
nun : VAR :
u.u.-‘-‘-.Vfi|

.
«-:v:o‘a.~~ aw

novov b‘ol‘-oed
4

' Q a
u, n..-
V“ gang-

0.. ,

. p.9'-9 V;

"' v0:.l...v .‘
o

'~A-A , ‘

... ‘ I O I

. ""V be. -..

'7 Pan
..
. a.-.
A ‘*:
.I‘IAC ‘ '
... ~0- -r‘v-A
. " b a ~-.
. U
':‘ ~o.,' .
...Vea' ..e?
I
"g A...

2;. .
. n -
.‘b v--
U§a
c
‘o.. .
A“ o
\.b‘ ~ru.“‘
v-
" 11‘ :5
.,' ..i’
.
r: '
‘I Q i
’ ~35 29m-
».
In...
‘
‘w:.£ be u .
5 r
‘-
. o
. ‘ ‘-
“w
‘0.
w “
\ A V
‘ stirrup, ,
u 7n,
i"-
~ R

20
iJitroduction, the study of denaturant induced unfolding of a
nuolecule and/or its domains has the potential to provide some
iJiformation about domain interactions and contribute to an
understanding of protein structure and function.
Our laboratory has an interest in the protein hexokinase,

‘the enzyme responsible for the initial step in the metabolism of

glucose in the brain. As discussed previously, the enzyme is

jpmimarily regulated by allosteric inhibition from its product,

giucose-G-phosphate, and an antagonism of this effect by P” In

aaddition, hexokinase demonstrates an ability to reversibly bind
‘the outer mitochondrial membrane (10), an ability also regulated

13y levels of Glc-6-P and P1.

Rat brain hexokinase is a large, monomeric protein and

.1indted tryptic digestion of the protein yields three fragments

b4, - 10, 50, and 40 kDa (11). The N-terminal fragment, Rt - 10

leDa, has been shown to be associated with the ability of the
enzyme to bind the outer mitochondrial membrane (12) . The M1 =
40 kDA fragment has recently been shown to contain the nucleotide

and.carbohydrate binding sites, and therefore has been associated

with catalysis (13,14). At the time of initiating these

experiments, the function of the 50 kDa fragment had not been
established although it was presumed to contain the allosteric
binding site for Glc-6-P.

Clearly, rat brain hexokinase is a large protein with a
structural complexity that reflects its functional complexity.

The activity and control mechanisms that exist imply a complex

 

u Av
«rm-:3 ..
"*‘novanOV

a t
I. ... g. a r
L: .56 V. ‘6

... r. n .A
a, ..

on.

'|'1

. u

~ A! :I . fl

vow-outaolr'
o

  

'
no. 7 ..g.‘ '3‘
"He. ‘u‘v‘esbb

 

 

21
<:ommunications network between various components of the protein.
'The use of denaturants such as guanidine and urea appears to be
<3ne way to perturb some of these interactions and to analyze the
relationships between the various domains, thereby gaining a

loetter understanding of structure and function in hexokinase.

 

1.... a. Q

 

'Ov;-v~‘~v
vovoi ‘1‘-

~ p
“...-i5); O'Ap
...:ouvi Os vu-

' I
:ev-Iuet A: ~
...-OCH. V. II!

n

’ I
... . ’ —
° ‘ a -
>00 you“. in...
a

I e .
6"“.‘0... A“
u.

Ari

 

1 you--

\“1

 

 

 

 

‘) n
cg ~ A
I ..t
. .
p
"v. ~ ..
“\1 5 Vane
.-

 

n
M
\ o.
c-c‘vh‘3“,
u ‘ .
‘ nut-3n
s N.‘
'1" :-
q “z:

\.
\.‘:-.
\‘_~,“'
I‘lpr
‘ ‘jr
h‘n‘
0:. :f ~‘
.,A
4‘Mb
“I“.

MATERIALS AND METHOD S

Materials

Ultra-pure guanidine HCl and ultra-pure urea were both
<obtained from.Schwarz-Mann. N-succinimidyl(2,3-9H)propionate was
a product of Amersham Corp. Other biochemicals were obtained

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

Purification of rat brain hexokinase.
Rat brain hexokinase was purified as described previously
(11).

Measurement of protein concentration and enzyme activity.
Hexokinase concentration was determined from absorbance
readings at 280 nm using an extinction coefficient of 5.1x10‘czm'1

mole"1 (15). Hexokinase activity was determined
spectrophotometrically using a Glc-6-P dehydrogenase coupled
assay as previously described (16); one unit is defined as the

amount of enzyme catalyzing the formation of 1 umole of Glc-6-P

per minute.
Inactivation of hexokinase.

Prior to use in inactivation experiments, enzyme was
duomatographed on a Sephadex G25 (fine) column equilibrated with
EmmM HEPES, 0.5mM EDTA, 10mM TG, pH 7.5 (HET buffer). After
dupmatography, enzyme solutions were routinely checked for the
removal of glucose that was present in the storage buffer.

Aliquots of enzyme solution prepared in this manner were diluted

22

 

y .-
..a. “....

00‘ I.

" . ~‘--

...-t-~V:"’

. l
-”‘.‘-‘hl'

, o
:uo‘fl": a F
an..~.-c-~ \-
II .

.I'A conga. ~ A
u -
Km. .OA'ufl'-‘
O
4
‘1. - R.’
no- “I Will

 

;""v--A¢u q
u... o..s U

   

 

I ‘F A.-
-:s "ayu

Uvoav-

a
‘OAA
v-4-

 

()
"h
..n

 

-

h: o ‘ ‘

‘- 1‘1.
1‘:

 

\in
. n
‘s A
5 v‘ “A
\ H “
s -

 

23
into HET buffer and inactivation was initiated by the addition of
denaturant with thorough mixing. The HET buffer contained
appropriate concentrations of ligands when the effects of ligands
were investigated. Initial enzyme activity was approximately 1.0
unit/ml. Controls, consisting of identical samples containing no
denaturant, were incubated at 25° C during inactivation
experiments and demonstrated no significant loss in activity.
Unless otherwise indicated, all inactivations were carried out at
25° C. Stock solutions of 8M urea or 6M GuHCl were prepared

daily using HET buffer.

Preparation of N-succinimidyl(2,3-3H)propionate labeled
hexokinase.

Twenty-five microliters of labeling reagent,
N-succinimidyl(2,3-’H)propionate in toluene, with a specific
activity of 107 Ci/mmole, was pipetted into a test tube and the
toluene solvent was evaporated under a stream of nitrogen. At
this time, approximately 1 mg of hexokinase in 1 ml of 0.1M
potassium phosphate, lOmM Glc, 0.5mM EDTA, pH 7.5, (PGE buffer)
was added to the test tube and the sample was mixed thoroughly.
The sample was allowed to incubate at room temperature for two
hours at which time TG was added to a final concentration of 10
mM. Free label was removed by dialysis against a total of two

liters of PGE buffer that also contained 10 mM TG; complete
removal of free label was ensured by chromatography of the sample

before use. Labeled hexokinase used for the indicated

 

"—5'.s '

‘-
nov-
-..
...-v

.

1i
‘l

.
'. .. Ayfl’VA ‘
. 1"- -..-wot
-.

 

 

v-
oa-pga p"
0" U--.
o

cop's. ... .‘.

.....w ...

 

_l
. .
:' vs. ,.‘~
...“.--‘-Ald
0
~......‘ ..A'

-'-—.I .-v..

'.; .1 5‘ V“
o

u-. g. U‘

be

3:}:- was i:

1:.‘A‘ .A '1

vvy‘ 5U L‘

12:. 3a:;:
sr~

 

 

‘2»
,~ A“
\v‘s .l" A
.V“»i
.-
‘1
\

‘s:“§n a
s“ .
“~3 e»-
V.
:3.
“ id
~...‘ H‘s'
‘2‘ 1
:u.
‘u ...n‘
\
Nh“ =“
‘5 CI
A
e‘ \“
.n‘.o
I \‘ ‘
set“:
I.
I‘ Q.
‘55“.‘
“s

 

 

 

24
experiments had a specific activity of approximately 50

mCi/micromole (10,400 dpm/microgram) .

Treatment of N-succinimidyl(2,3-’H)propionate-labeled sanmles for

 

counting .

The presence of GuHCl in samples created a precipitation
problem in the scintillation fluid that was eliminated by using a
solubilizing procedure. 150-200 microliters of sample in the
:inactivation buffer was added directly to a scintillation vial.
(Dne ml of NCS tissue solubilizer (Amersham) was added, and the
sample was incubated at 50° C for one hour. Samples were allowed
1:0 cool to room temperature at which time 0.030 ml of glacial
aacetic acid and 5.0 ml of Safety-Solve scintillation fluid were
aadded. Samples were mixed well and counted in a Packard 300 CD

using an external standard and a quench curve to calculate dpm.

fieasurement of N-succinimidyl(2,3-3H)proLionate-labeled

 

gyexokinase adsorption to glassware.

Test tubes that were used to incubate labeled hexokinase in
'various concentrations of denaturant were assayed for adsorbed
protein. After removal of the sample, test tubes (10x50 mm
silanized borosilicate) were rinsed with double distilled water,

fiflJed with Safety-Solve scintillation fluid, and deposited in a
mfintillation vial containing 5 ml of scintillation fluid. After
‘dgorous mixing, the test tubes were counted in a Packard 300 CD

as described previously. After all tubes were counted they were

':.Anal‘ ‘7‘, ‘

~ ‘
..-.n;oafl ‘ R :
ucfi'I-‘vd OH —

.
u! Ofi 'I‘ Am ’
"I. UV by you .

I
a u ..
: :"v‘ a
fivwoivu- v- u
t

"70' a:

.
\
.‘oaooeb ..

r....' ‘.|I h
"‘“" .uld

. V I
5"” “ "a i
.'"" i ..A

n
I

I "“V‘c.1 A
..."Ofiiy.!’ ue5"

5:25; :0 the c,

.O

.
i?
'V

eq'fi v
" -we Dy :h

 

 

'u..‘ u
. - m
+3”- assoc

 

 

25
removed from the scintillation fluid, rinsed with 95% ethanol,
and filled with 0.350 ml of 1% SDS. Test tubes were then
incubated in a boiling water bath for 15 minutes and allowed to
cool to room temperature. At this time, 0.175 ml aliquots of the
detergent solution were removed and treated with NCS tissue
solubilizer as previously indicated. Test tubes were again
rinsed with 95% ethanol, filled with scintillation fluid,
deposited in 5 ml of scintillation fluid and counted as
previously described. The total dpm in 0.350 ml of 1% SDS was
equal to the fraction of counts that had been removed from the
test tube by the SDS solubilization procedure. From these data,

total dpm associated with the test tube could be calculated.

 

'---o~nv:. « en
a

..
... .I‘v-v

A!!! G?

yore” r-
‘Ul! I! ‘P
«mess S‘:nn

"'3! s.
”we'd-'u .

 

’ 1
t'“"'-;ro K
..‘I‘-.."“ ~ ‘

. .

lb

.-_;:s::a:es

 

 

"‘- ...

_“':'e a In

:36 332356 <

Id :9 ."_’"
h
N “ ..“‘

‘0

I‘m-o.
"N3“ cf 1

'Id. ..
‘.

h‘obuved fi:

I
§‘-‘
\ ..

I‘v‘ses ‘Jr‘

I
‘- Q"~‘§
h‘: ‘
5 A.
s SVlL
“-~
Q
§
\.: e ‘ ‘ ‘
V
I “:
.Q
\:‘s,'
I ' .
~‘ ..
.5. C
0:
... ‘
"3:3‘.‘
‘

 

 

RESULTS

Inuactivation of hexokinase in urea and GuHCl

Many proteins unfold and refold by a simple two-state
gxrocess such that the protein exists in either the native or
<1enatured state and intermediate conformations are not
significantly represented at any time (17). Hexokinase
«demonstrates inactivation kinetics in denaturants that may
indicate a more complex unfolding pathway for this enzyme. The
time course of inactivation at ZS’Ciin several concentrations of
tuea are indicated in Figure 1. For these experiments a stock
.solution of native hexokinase was diluted with 8M urea to the
.indicated final concentrations of denaturant. It is clear that
'the inactivation shows a biphasic character at lower
<:oncentrations of urea. The biphasic character of the
:inactivation becomes less pronounced as the concentration of urea
:increases until finally, at relatively high concentrations,
:inactivation appears monophasic.

The inactivation of hexokinase in GuHCl shows nearly
:identical kinetics (Figure 2). These experiments were carried
(out in the same manner as the inactivation experiments in urea--a
6M GuHCl solution was used to dilute stock enzyme to the
indicated final concentrations of denaturant. Again, biphasic
inactivation appears to approach monophasic inactivation as the

concentration of denaturant increases. The lower

26

27

Figure l. Inactivation of hexokinase in various
concentrations of urea. A hexokinase solution was diluted to the
indicated final concentrations of urea using an 8M stock
solution. All inactivations were performed at 25°<3. In the
absence of denaturant samples lost less than 2% original

activity.

Activity (96 original)

10

6(

.1:-
{—3

 

28

 

€598 é 33:3

 

 

6O

40

Time" (min)

20

 

Figure 1

 

 

 

29

Figure 2. Inactivation of hexokinase in various
concentrations of GuHCl. A hexokinase solution was diluted to
the indicated final concentrations of denaturant using stock 6M
GuHCl. All inactivations were performed at 25°<3. In the
absence of denaturant samples lost less than 2% original

activity.

Ac tivity (96 original)

40-

20+

 

 

Activity (% original)

 

30

 

 

 

 

 

 

20

40
Time (min)

60

Figure 2

 

 

"“3"'3' ‘ A'
,,_'.u..¢--V“

... on! v“ F
“: 15....3. \v-
I

an. n A
. V

 

nov a
V. ‘iA\“ r
O. oav‘-‘

' Q
': :'~na v0 ‘15
--.¢..I..! .U

‘::“.’ tire o.‘

"""I 90- 5‘

Q
.3 ~:"=" «v
'- “..-...

-5

;?:-u' 4

.4..- A.
n . ' .n V.

...

o I‘I--.

......_es 3:

3:?”6" o- ‘
‘ b

0o“;“_. a,

--,. .
~:: ." .‘. a.
55- “<

.
Y‘s

'. .9-

~- .29»
0.. 8“. -

5 l
N
chlcn
#-
\_ “
I ‘k- ’ ‘
A.~ Hr‘
."::‘E> ‘I‘
I ‘L ‘
‘1
5': . .
I.‘ i ‘. a
‘c. f‘fi
5.
.; l‘
“ 2' " .
b V
5‘3“ c
«‘9‘
53"
\.‘3‘
C., a b.
i-\.
. »::..

 

 

 

31
concentrations of GuHCl required to effect this change reflect
its general characterization as a much stronger denaturing agent
than urea.

It should be noted that, although a biphasic character in
relatively low concentrations of denaturant is a consistent
result, the time span that the first phase occupies appeared to
be variable in experiments performed on different days. For
example, in one experiment the "break" between phases might occur
ten minutes after initiation of inactivation, and in another
experiment take more or less time. It has not been possible to
clarify the nature of this inconsistency in the course of these
experiments.

In early experiments investigating the effects of urea on
inactivation it became apparent that temperature also has a
pronounced effect on the rate of inactivation. Figure 3 depicts
the effect of temperature on the rate of inactivation of
hexokinase in 3.5 M urea. Two separate inactivation time courses
at 26° C demonstrate the reproducibility of the data within an
experiment when samples are maintained at identical temperatures.
Ekmever, it is clear that when the temperature is altered the
rate of inactivation responds accordingly. In this case, a drop
Of‘P C significantly decreased the rate of inactivation. In
addition, a biphasic character becomes evident at the lower

temperature.

32

Figure 3. Effect of temperature on the inactivation of
hexokinase in urea. Hexokinase was inactivated in 3.5 M urea at
22° (0) and 26° C (A) . Two separate inactivations at 26° C
demonstrate reproducibility when constant temperature is
maintained. In the absence of denaturant no significant

inactivation occurred.

Activity (96 original)

 

40

20

 

33

 

 

6O

4O

20.

 

 

.\
est...
. .\
O
I K A..||..II.-.|A\‘..h
\O\O‘O A‘Ak‘k‘
‘A\ u.\oA\\J.A\A\|\
Lalo m m o m
0 8 6 4 2

cacato «a .>:>=o<

Time“ (min)

Figure. 3

u
....o I

. I A
' 'chnu‘
:.eiovu-

on. :‘pa‘afiu’ A
—- fic..-’ v

“"‘00:‘. fl.

.

...-nooovflotb ‘0
.

.
L':"S ‘FF un-
'j"'- Ohm-cu...

. I V.
J u}. ._ i

n .I:q.." .“A
" ’Wv-oi

 

u-..
u'":’v:. ~‘NA
l-".‘. ’OU'

-
l ‘0 .... '

N r-
” in utae ‘e..:

‘é':-‘ .' a;
No... .“ Vfita:
"v .
‘:.e. a'. ‘. II
Mc‘e‘-:
2'." G:‘QAQ
- uscv.~:r C
A
: "an 'R
a. ...
i Uh. C

O...
I O.
5" ‘v
.....a..-cn H
.
Q».
N“ v-.

Ss a‘es t
t.
.

h
o H
'cr‘ufib
5
"°“a"‘ C“
u
.; . .
I‘ ‘.J

Q!
I
... c
e e~fen
i‘m‘
\ iv
5 h
5..
i: ‘0:
s“. ‘0 .
.‘Jac‘
a ‘C
"E -.
“‘3Se, L.
‘c
|.:.i:.:.'.

 

 

 

 

34

Effect of ligands on inactivation of hexokinase by urea and
GuHCl.

Previous experiments in this laboratory have demonstrated
the ability of appropriate ligands to afford the enzyme a
significant degree of protection against various inactivating
agents including chymotrypsin, DTNB, heat, and glutaraldehyde
(18-20). It was of interest to study the effect of these ligands
on inactivation of the enzyme in denaturants. The effects of the
substrate glucose, or the effector analog, l,5-anhydrog1ucitol-6-
P, on the denaturation in urea and GuHCl is discussed in some
detail in Chapter 3, and for this reason will not be presented
here. Briefly, the results indicated that both the substrate and
the effector provided significant protection against inactivation
in urea, but only the hexose provided effective protection when
denaturation was performed in GuHCl. In addition, Figure 4
demonstrates the effects ATP and the ATP-Mg complex on the
inactivation of hexokinase in 2.5 M urea. For these experiments,
solutions of hexokinase in buffer containing appropriate
concentrations of ligands were diluted to a final concentration
cm 2.5 M urea or 0.6 M GuHCl using stock denaturant solutions.

The effect of these ligands on inactivation by urea is
Similar to those effects previously reported for other methods
cm’inactivation. Glucose, the primary substrate for brain
meokinase, shows a marked protective effect on the enzyme.

1,S-anhydroglucitol-6-P, an analog of the allosteric effector

35

Figure 4. Effect of ATP and ATP-Mg on the rate of
inactivation of hexokinase in 2.5 M urea. Inactivations were
performed as previously described. The final concentration of

ATP or ATP-Mg in the inactivation solution was 10 mM.

Activity _ (% original)‘

1'0

36

30
Time (min)

Figure 4

 

37

Glc-G-P, also provides significant protection. The ATP-Mg
complex, which is the nucleotide substrate for the hexokinase
molecule, did not provide protection and actually appears to
further destabilize the enzyme in the presence of urea. On the
other hand, ATP in the absence of Mg‘+ afforded significant
protection against inactivation in urea. The ability to affect
the rate of inactivation appears to be specific to appropriate
ligands--poor substrates or effectors such as galactose and
galactose-G-P do not change the kinetics of the inactivation of
hexokinase in urea.

The physical bases for urea and GuHCl denaturation are
presumed to be similar and, as already mentioned, initial
inactivation experiments show similar kinetic trends for
denaturation in these compounds. However, kinetic experiments
suggest that the effects of ligands on inactivation by GuHCl are
not similar to their effects in urea. Although glucose shows a
significant protective effect in GuHCl, it was clear that the
effector 1,5—anhydroglucitol-6-P does not provide any protection
against inactivation in this denaturant (data presented in

Chapter 3).

Transition curves for hexokinase in GuHCl

The hexokinase activity remaining after 24 hours of
incubation in various concentrations of GuHCl is shown in Figure
5. Activity levels at this time appear to be stable since

measurements after 48 hours of incubation show no appreciable

38

Figure 5. Twenty-four hour GuHCl inactivation curve for
hexokinase. Hexokinase was incubated in different GuHCl
concentrations at 25°C2in.silanized borosilicate test tubes.
After 24 hours, the activity remaining in samples was plotted vs.

GuHCl concentration.

 

Activity/(96 control)

80

60

401

20' _

0:1

 

0.2

39

0.3

0.4

0.5 06
[Gu~HCl] (M)

Figure 5

 

0.7

. v

I.

 
 

 

 

40
shift in the curve. Clearly, the major loss of activity takes
place at relatively low concentrations of denaturant—~a sharp
transition between active and inactive protein occurs between 0.3
M and 0.4 M GuHCl. For many proteins a stable transition such as
that depicted in Figure 5 would suggest a simple equilibrium
process for denaturation, i.e., the native form of the molecule
is in direct equilibrium with the denatured form (NTZD)(21). For
these proteins one is able to deduce the fraction of protein in
the native or denatured state at any point along the curve.
Furthermore, one should be able to move back up the transition
curve by removing denaturant and shifting the equilibrium toward
the native form of the protein.

However, it appears that such an equilibrium process is not
involved in the formation of the activity transition curve for
hexokinase in GuHCl. Initial attempts to remove denaturant,
either by dilution or dialysis, failed to show any recovery of
lost enzyme activity. This was true for samples taken from any
point on the transition curve, and included dilution and dialysis
into buffers that contained various additions, including
substrates and effectors of the protein (data not shown). It
appears that there is a significant irreversible loss of enzyme
activity with increasing concentrations of denaturant under these
conditions. Clearly, the presence of an irreversible step in the
denaturation of hexokinase is inconsistent with what appears to
be a stable activity transition, and it was not possible to come

up with a reasonable explanation for this anomaly during the

 

N.

T'-

 

 

 

 

41

course of these experiments.

Effect of ligands on the activity transition of hexokinase
in GuHCl.

The effects of various ligands on the 24 hour inactivation
transition curves were also investigated. In light of
preliminary inactivation experiments in GuHCl, it was not
surprising that the most pronounced effect on the activity
transition was observed when the enzyme was inactivated in the
presence of 10 mM glucose. A typical transition in the presence
of glucose is depicted in Figure 6. Clearly, glucose has a
significant effect on the denaturation of hexokinase; the
majority of the transition from active to inactive enzyme now
takes place between 0.6M and 0.7M GuHCl, rather than the typical
transition seen at approximately half that denaturant
concentration in the absence of added ligand.

The effects of other compounds on the transition are given
in Table l. The transition in the presence of any compound was
similar in overall shape to the transition curve seen with no
ligand present. When the curves differed they did so only with
respect to the concentration of denaturant required to induce the
transition. Thus, the effects of ligands can be described by
noting the changes in the concentration of GuHCl at which the
enzyme had irreversibly lost 50% of its activity (Id-

Mannose, also a good substrate for hexokinase,

significantly shifts the transition to higher concentrations of

42

Figure 6. Twenty-four hour GuHCl inactivation curve for
hexokinase in the presence of 10 mM glucose. Hexokinase was
incubated as described for Figure 5. In addition, all samples
contained 10 mM glucose. After 24 hours of incubation, samples

were assayed for hexokinase activity.

43

100

. _ h
- 0 . 0
8 6 4

2228 so 33:2

. _ .
0
2

0.5
[<3u~Hcfl (M)

0.1

 

0.3

Figure 6

44

Table I. Concentration of denaturant required to
demonstrate 50% inactivation of hexokinase in the presence of
various ligands.

 

 

Added Ligand‘ gab [GuHCl] at 50% inactivation, M°
None ‘ 6 0.34 i 0.03

Glucose 4 0.65 i 0.02

Mannose 1 0.66

Galactose 2 0.36 i 0.00'

Glc-6-P 4 0.40 i 0.04

1,5-AnG-6—P 2 0.36 i 0.02'

Gal-6-P 2 0.33 i 0.02'

ATP 1 0.39

ATP-Mg 1 0.27

 

‘Concentrations of hexoses were 10 mM; concentrations of hexose-
6-phosphates were 1 mM; concentrations of nucleotides and Mg++
were 10 mM.

bNumber of measurements.

°Where appropriate, data are recorded as an average : the S.D.
Samples marked with an asterisk are recorded as the average : one
half the range.

45
denaturant. It should also be noted that N-acetylglucosamine, a
compound that is bound but not phosphorylated by the enzyme (21),
induces a small shift in the transition. Galactose shows no
effect, suggesting that the effects seen with other carbohydrates
are due to protection induced by the specific binding of these
ligands, and the conformational change that they induce.

Results of kinetic studies indicate that the effectors
1,5AG-6-P and Glc-6-P should not affect the inactivation
transition. In fact, transitions in the presence of l,5AG-6-P
were indistinguishable from those seen with no ligand present,
and the presence of lmM Glc-6-P induced only a slight shift of
the transition to higher concentrations of GuHCl. Man-6-P also
appears to have no marked effect on the transition. The only
compound tested that shifts the transition to lower
concentrations of denaturant is the ATP-Mg complex, the
nucleotide substrate for hexokinase. ATP alone shows a moderate

level of protection.

Adsorption of protein to glassware

It is clear that irreversible inactivation plays a
significant role in the guanidine-induced denaturation of
hexokinase under these conditions. Since the concentrations of
protein that were used were relatively low (approximately 15-20
ug/ml), experiments were performed to assess whether or not
adsorption of protein to glassware might be the source of

irreversible loss of enzyme activity in these experiments.

46
In order to perform adsorption experiments, enzyme was
labeled with N-succinimidyl(2p3JH)propionate as described in the
methods section. This labeling reagent reacts primarily with the

epsilon amino groups of lysine as depicted below (24):

0 g 0
ii
-C -NH-CH-iCH2)‘- NH2 4 CHJCHI- C - NC) ————>
i
s=° .
NH
1
0 0
u u
_ C " NH'?H‘(CH1"‘ NH‘C " CH2 CH3 9 N:
(=0
NH 0

Hexokinase labeled with this reagent did not suffer any loss
in catalytic activity, and, as indicated below, the derivatized
enzyme behaved in a manner similar to the non-derivatized enzyme
in response to denaturation in GuHCl.

As in previous experiments, enzyme was allowed to incubate
at 25° C for 24 hours in various concentrations of Gui-1C1. In
addition, some samples also contained 2.0 mg/ml BSA. At the end
of the incubation period samples were assayed for enzyme
activity, and samples and test tubes were radioassayed for
hexokinase. Using the methods described, the recovery of total
counts in the samples was 3 90%. Protein adsorbed in the

presence or absence of BSA in various concentrations of

47
denaturant is indicated in Figure 7. Two observations are of
particular interest: 1) clearly, the adsorption of hexokinase to
glassware in the absence of added protein occurs to a significant
extent at all denaturant concentrations; 2) although
adsorption appears to be a problem at these protein
concentrations, the concentrations where adsorption is most
significant suggest that 70-80% of the total protein still
remains in solution. It can be concluded that, although
adsorption is occurring to a significant extent, it does not
appear to be responsible for the major irreversible loss of
enzyme activity.

This latter point is substantiated further by Figure 8.
Although the presence of 2 mg/ml BSA virtually eliminates.
hexokinase adsorption, it does not prevent the irreversible loss
of activity in these samples. In fact, on a % control basis, the
transition in the presence of BSA is indistinguishable from the

transition curve formed in the absence of added protein.

48

Figure 7. Adsorption of hexokinase to glassware.
N-succinimidyl(2p3JH)propionate labeled hexokinase was incubated
in various concentrations of denaturant as described in Figure 5.
Incubations were performed with no BSA present (A0 and with 2.0
mg/ml BSA present (A). After 24 hour incubations, test tubes
were radioassayed for adsorbed hexokinase. Data points shown are

the average of duplicates.

dpm, test tube (% total)

30

20

10

49

 

A t

\A.

p

 

A

/ /\
/

 

 

.A‘A‘A/A\A\
~ I l l e—A
QJ 03‘ 0'3 4:07

 

[bu—HOD (M)

FigUre 7

50

Figure 8. Inactivation curve for hexokinase in the presence
of 2.0 mg/ml BSA and various concentrations of GuHCl. Hexokinase
was incubated in various concentrations of denaturant as
indicated in Figure 5. Samples contained either no BSA (0) or
2.0 mg/ml BSA (A). After 24 hours, samples were assayed for

enzyme activity.

51

.0A

100

80

f —
0 . O
6 4

2828 is £28...

20

I.

0.3

0.1

@u‘vHCfl (M)

Figure 8

DISCUSSION

Hexokinase is a large protein that is both structurally and
functionally complex. In the experiments presented, the
denaturation of hexokinase in urea and GuHCl has been examined in
hopes of finding conditions whereby the independent denaturation
of the domains of hexokinase might be observed.

In light of previous experiments performed on multi-domain
proteins (reviewed in introduction), it could be predicted that
the urea or guanidine-induced denaturation of a protein as large
as hexokinase might be complex. The preliminary kinetic
experiments described here suggest that the inactivation of
hexokinase is something other than a simple two-state process.
If the transition of hexokinase from an active to an inactive
protein was an "all or nothing" process, with one unfolding step
governing the denaturation, the loss of activity should be
described by a simple monophasic curve (22). At relatively low
concentrations of denaturants, this is clearly not the case.
Kinetic studies are generally not able to characterize the nature
of the different steps that a protein follows during
denaturation. However, they are often able to rule out a simple
two-state mechanism for unfolding--this appears to be the case
with hexokinase.

The effects of ligands on the rate of inactivation in urea

52

53
are similar to those seen in previous experiments in this
laboratory (ll-13). In most inactivation experiments, glucose
provides substantial protection for the enzyme-—its binding to
the active site of the enzyme induces conformational changes that
stabilize those regions of the protein important to catalytic
function. The presence of l,5AG-6-P also markedly slows the rate
of inactivation of hexokinase in urea, but probably not through
direct interaction with the active site. In this case,
stabilization is presumed to be induced by a conformational
change that is communicated from one part of the molecule to
another. That is, the conformational change induced by the
binding of the effector l,5AG-6-P to the allosteric binding site
is communicated to that part of the molecule responsible for
catalytic function, resulting in its stabilization and a reduced
rate of inactivation. It seems apparent that the binding of the
nucleotide substrate, ATP-Mg, induces an unstable conformation
that is further destabilized in the presence of urea. Those
compounds which are neither good substrates nor good effectors,
i.e. galactose and galactose-6-P, do not appear to alter the
enzyme's conformation. This is reflected by their lack of
influence on the rate of inactivation. The effects of
principal ligands on the rate of inactivation in guanidine
are not identical to those seen in urea. Although glucose
induces significant protection, it is notable that the
allosteric effector 1,5AG-6-P does not appear to provide

protection against inactivation. These results were also

54
reflected in the effect of ligands on the 24 hour denaturation
curves of the enzyme. It is possible that guanidine is able to
interfere with interaction between the domains of the molecule.
Binding of the effector may stabilize that part of the molecule
associated with allosteric control, but the stabilizing
conformational change is not imparted to that part of the
molecule associated with catalytic function. In any case, it is
clear that the denaturation processes in urea and guanidine are
distinct.

Activity can be one of the most sensitive parameters
reflecting native structure in an enzyme (23). For this reason,
it is not surprising that the inactivation of hexokinase shows a
sharp inactivation transition at fairly low concentrations of
GuHCl. As previously mentioned, activity loss appears to reach
an equilibrium plateau after 24 hours, although activity loss
does not appear to be representative of an equilibrium process
for the unfolding of the molecule. That is, it is not possible
to restore activity by removing denaturant--there is clearly an
irreversible step involved on the unfolding pathway under these
conditions.

The nature of the irreversible denaturation of hexokinase
remains unclear. The results presented for experiments performed
with N-succinimidyl(2p3JF)propionate-labeled hexokinase suggest
that adsorption of protein is not the source of irreversible
inactivation that is demonstrated by the transition curves, and

the data in Figure 8 suggest that the difference in protein

55
concentration under the two different conditions (i BSA) does not
affect the inactivation transition.

It seems clear that the denaturant induced unfolding of
hexokinase is complex, apparently including intermediates as well
as irreversible processes. However, the data presented suggest
that, under appropriate conditions, it may be possible to use
these processes to gain a further understanding of domain

structure and function in hexokinase.

10.

ll.

12.

13.

14.

15.

16.

17.

18.

19.

REFERENCES

White, Frederick H. (1961) g. Biol. Chem. 236, 1353-1358.

 

 

Anfinsen, Christian B. and Edgar Haber (1961) g. Biol. Chem.
236, 1361-1363.

 

wetlaufer, Donald B. (1973) Proc. Nat. Acad. Sci. USA 19,

 

 

Rossman, Michael G. and Partick Argos (1981) Ann. Rev.
Biochem. 50, 497-532.

Blake, C.C., and Johnson, L.N. (1984) TIBS 2, 147-151.

Chothia, C. (1984) Ann. Rev. Biochem. 53, 537-572.

 

 

Wetlaufer, Donald B. (1981) Adv. Prot. Chem. 34, 61-92.

 

 

Wilson, J.E. (1968) g. Biol. Chem. 243, 3640-3647.

Tuttle, J.P. and J.E. Wilson (1968) Biochim. Biophys. Acta.
212, 185-188.

 

 

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

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

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

Nemat-Gorgani, Mohsen and J.E. Wilson (1986) Arch. Biochem.
Biophys, in press.

 

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

 

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

 

Polakis, Paul G. and J.E. Wilson (1982) Biochem. Biophys.
Res. Comm. 107, 937-943.

 

Creighton, Thomas E. (1983) Proteins p. 289. W.H. Freeman
Co., New York.

Wilson, J.E. (1973) Arch. Biochem. Biophys. 159, 543-544.

Wilson, J.E. (1978) Arch. Biochem. Biophys. 185, 88-99.

 

20.

20.

21.

22.
23.

24.

57

Ghelis, Charis and Jeannine Yon (1982) Protein Folding p.
468, Academic Press, New York.

 

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

 

 

Baijal, M, and Wilson, J.E. (1982) Arch. Biochem. Biophys.

 

 

218, 513-524.

Ikai, A. and Tanford, C. (1973) g. Mol. Biol. 13, 145-163.

 

Tsou, C.-L. (1986) TIBS 427-429.

Tang, Y. S. (1983) Journal of Labelled Compounds and
Radiopharmaceuticals 20, 277-284.

 

Chapter 3

Location of the Glucose-G-P Regulatory Site on

a Domain in the N-Terminus of the Enzyme

Hexokinase is the enzyme responsible for the initial step of
glucose metabolism in the brain, catalyzing the conversion of Glc
to Glc-6-P using ATP as phosphoryl donor.- It is generally
accepted that hexokinase represents a major regulatory influence
on the rate of cerebral glucose utilization, and that inhibition
of the enzyme by its product, Glc-6-P, is a major factor
governing hexokinase activity in vivo; in addition, reversible
binding of the enzyme to the outer mitochondrial membrane has
been suggested to play a regulatory role (1). A major goal in
our laboratory has been to establish the structural basis for the
catalytic, regulatory, and membrane-binding functions of brain
hexokinase.

Rat brain hexokinase (like other mammalian hexokinases)
consists of a single polypeptide chain having a molecular mass of

about 100 kDa (1). Under nondenaturing conditions, limited

58

tryptic digestion of the enzyme yields three major fragments with

molecular masses of 10, 40, and 50 kDa (2), as depicted below‘.

Partial cleavage at only T1 or T2 generates transient
intermediates of Mr 90K or 60K, respectively. The N-terminal 10
kDa fragment has been shown to be critical for the interaction of
hexokinase with mitochondria (4,5). In addition, it has been
demonstrated that the binding sites for both substrates, ATP (6)
and Glc (7), and thus catalytic function, are associated with the
40 kDa C-termdnal fragment. The central 50 kDa fragment has been
proposed (7) to be the location of the allosteric binding site
for the effector, Glc-6-P. The results of the present work
support this proposal.

Structural domains may be independently denatured in the
presence of agents such as guanidine hydrochloride (GuHCl) or
urea. For example, Zetina and Goldberg (8) have demonstrated
independent unfolding, induced by GuHCl, of the two domains
comprising the £L_subunit of E. coli tryptophan synthetase. As
is frequently the case with other proteins, binding of certain

ligands has been shown to stabilize brain hexokinase against a

59

60
variety of inactivating agents (9-12). We therefore reasoned
that these ligands might provide similar stabilization against
denaturants such as urea or GuHCl. In the present investigation,
ligands which bind at the (hexose) substrate or allosteric
regulatory sites of brain hexokinase have been shown to
selectively protect discrete domains against denaturation by
GuHCl. By use of peptide mapping techniques, in conjunction with
Western blots probed with monoclonal antibodies having epitopes
of defined location within the molecule (5), it is possible to
identify the domains stabilized by, and thus the binding sites

for, these ligands.

MATERIALS AND METHODS

Materials. Rat brain hexokinase was purified as previously
described (2). GuHCl and urea, both ultra pure, were obtained
from Schwarz/Mann Biotech (Cleveland, OH).
1,5-Anhydroglucitol-6-P was synthesized according to Ferrari gt
El. (13). PMSF, TPCK-treated trypsin, and other biochemicals

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

Hexokinase Activity and Protein Determinations. Hexokinase
activity was determined spectrophotometrically using a Glc-6-P
dehydrogenase-coupled assay as previously described (12); one
unit is defined as the amount of enzyme catalyzing the formation
of l/umole of G1c-6-P per minute. Hexokinase concentrations were
determined by the absorbance at 280 nm using a molar extinction
coefficient of 5.1x10‘M“cm"1 (14). Trypsin concentrations were
also determined from Am based on an absorbance of 1.43 for a 1

mg/ml solution (15).

Inactivation of Hexokinase. The purified enzyme is stored
in Glc-containing phosphate buffer for reasons of stability (14).
Prior to use, protective ligands were removed by chromatography
on a Sephadex G-25 (fine) column equilibrated with 0.05 M HEPES,
0.5 mM EDTA, 10 mM Thioglycerol, pH 7.5 (HET). Removal of Glc
was verified by enzymatic assay of the hexokinase fractions after

chromatography. Aliquots of enzyme were diluted into HET buffer,

61

62

equilibrated at 25°C» and inactivation initiated by the addition
of a stock denaturant solution (prepared in HET) with thorough
mixing. Aliquots were removed at the indicated times for assay
of residual activity; after dilution into the assay solution, the
concentration of denaturants was such that it had no effect on
activity. Unless stated otherwise, the final concentration of
enzyme in the inactivation mixture was 0.130 mg/ml. Where
indicated, ligands were also present. Control samples (no
denaturant) demonstrated no significant (<5%) loss of activity

during the period of the experiment.

Proteolysis of Hexokinase. Enzyme samples denatured with
GuHCl were subjected to proteolytic digestion as an assay for
structural perturbation. Regions of the molecule not protected
by ligands could be expected to have undergone significant
unfolding, resulting in increased susceptibility to proteolysis.
After 60 minutes of incubation at 25° C in HET containing 0.6 M
GuHCl, trypsin (5 mg/ml in 1 mM HCl) was added to a final
concentration of 0.026 mg/ml: control digests were prepared
identically except that no GuHCl was added to the HET. Digestion
was allowed to proceed at 25° C for 20 minutes, at which time
samples were moved to ice and PMSF added to a final concentration

of 1 mM to inhibit further digestion.

Electrophoretic Procedures. SDS-gel electrophoresis on 6-20%

linear gradient acrylamide gels was performed as previously

63
described (2). The presence of 0.6 M GuHCl in the samples
resulted in formation of a precipitate in the Laemmli denaturing
solution, which initially made samples unsuitable for loading and
running on a gel: this could be avoided if samples were first
dialyzed to remove GuHCl. Samples (usually < 0.5 ml total
volume) were dialyzed vs. HET buffer to which PMSF had been added
to a final concentration of 1 mM, using a microdialysis apparatus
(Bethesda Research Laboratories, Gaithersburg, MD; Model 1200
MA). Dialysis was allowed to proceed for 1.5 hours at a buffer
flow rate of 60 ml/hr. Where indicated, gels were scanned with a
Hoefer Model 65300 scanning densitometer (Hoefer Scientific

Instruments, San Francisco, CA).

Immunoblotting Procedures. Gels were electroblotted to
nitrocellulose as previously described (2). Blots were washed,
blocked, and incubated with monoclonal antibodies according to
the protocol outlined by Polakis and Wilson (2). Three
monoclonal antibodies, designated as 18, 5A, and 3A2 were used in
the experiments described here. Characteristics of antibodies 18
and 5A have been previously described (5); briefly, 18 is an IgGl
which recognizes an epitope located in the central 50 kDa domain,
while 5A is an IgM recognizing an epitope in the C-termdnal
catalytic domain. Antibody 3A2 has not been described in
previous publications from this laboratory. However, using
techniques employed in characterization of other antibodies (5),

3A2 has been shown to be an IgGl recognizing an epitope in the

64
N-terminal 10 kDa domain of rat brain hexokinase (A.D. Smith and
J.E. Wilson, unpublished work). After incubation with the
primary antibodies, blots received 2 ten minute washes with
Tris-buffered saline (20 mM Tris-Cl, 0.5 M NaCl, pH 7.5),
referred to as TBS, to which 0.05% Tween 20 had been added
(TTBS). Additional blocking was achieved by subsequent
incubation for 30 minutes in a 5.0% (v/v) solution of horse serum
in TBS, which was followed by a 10 minute wash in TTBS.
Immunoreactive species were detected using the Vectastain ABC
kit, following the protocol supplied by the vendor (vector
Laboratories, Burlingame, CA). Before initiation of the color
reaction, blots received two ten minute washes in TTBS, followed
by a rinse with TBS. The color reaction was initiated by
immersing the blots in a freshly prepared solution composed of 10
ml 4-chloro-1-naphthol (3.0 mg/ml in chilled methanol) and 50 ml

0.015% H202 in TBS.

RESULTS

Inactivation of Hexokinase in Urea and GuHCl. Initial
studies were directed at determining the effects of various
ligands on the inactivation of rat brain hexokinase in the
presence of urea or GuHCl. With both denaturants, inactivation
did not follow simple first order kinetics but rather appeared to
be (at least) biphasic. Ligands used included Glc and the
effector, l,5-anhydroglucitol-6-P, an analogzcof Glc-6-P which
also serves as an effective inhibitor of the enzyme (11). As
shown in Figure 1A, both ligands provided substantial protection
against inactivation by urea. With GuHCl as denaturant, however,
Glc provided substantial protection while l,5-anhydroglucitol-6-P
had no detectable effect on the rate of inactivation (Figure 18).
In addition to demonstrating distinct differences in the action
of these denaturing agents, these results provided the first
indication that it might be possible, using GuHCl as denaturant,
to selectively affect the catalytic and regulatory regions of the
molecule. Thus, although the allosteric effector,
l,5-anhydroglucitol-6-P, was evidently not protecting the
catalytic region against the denaturing effects of GuHCl, it was
conceivable that this ligand (or Glc-6-P itself) was binding to
and stabilizing a structural region associated with allosteric

regulation.

65

66

Figure l. Inactivation of rat brain hexokinase by urea or
guanidine hydrochloride. A, 2.5 M urea; B, 0.6 M GuHCl. Where
indicated, 10 mM Glc or 1 mM 1,5-anhydroglucitol-6-P were added.

In this experiment, hexokinase concentration in the denaturing

solution was 0.025 mg/ml.

Activity Remaining, %

Activity Remaining. 7.

67

 

 

 

 

 

 

 

 

1004
lwgggaRCA’OeOOQRAQ (90
BUD Q A0
80-4 U
D U
' U
60~ D U D D.D-
0 D DD
404
.l
D No Ligand
A Glucose A
20 o 1,5—Anhydroglucitol-6-P
0 10 20 30 40 50
Time, minutes
100
Jan 0 U U U U U D
J 8 8 0 Q
60- Q A
e o 2
6 _
6 8 A A
40" 0 o a
J
o No Ligand B
A 1.5-Anhydroglucitol-6-P
20 u Glucose f fi
0 10 20 30 4-0 50

Time, minutes

Figure 1

68

Selective Protection of N- and C-Terminal Regions of Hexokinase
by Glc-6-P and the Substrate Analog, N-Acetylglucosamine. To
assess the ability of various ligands to selectively protect
specific domains in the molecule, samples denatured in GuHCl for
60 min, with or without addition of ligands, were subjected to 20
min of tryptic digestion (see Methods). For the experiments
discussed here, the ligands used were the allosteric effector,
Glc-6-P, and GlcNAc, an analog of Glc which binds at the
catalytic site but does not induce the extensive conformational
changes that result from binding of Glc itself (11). The
intention in using GlcNAc was to attempt to restrict any
ligand-induced stabilizing effects to the vicinity of the hexose
binding site. As noted above, Glc-6-P had no effect on the loss
of activity seen in the presence of GuHCl, while GlcNAc offered
significant protection against this denaturant (Table I). These
effects would be consistent with stabilization of the catalytic
region by GlcNAc but not by Glc-6-P. The extent of inactivation
depended somewhat on protein concentration, with less
inactivation seen (at a given time) at higher protein
concentrations. Thus the inactivation observed in the experiment
shown in Figure 1B, in which hexokinase was 0.025 mg/ml, was
greater than that at comparable times of inactivation for the
experiment shown in Table I (and all data presented subsequently)
in which the hexokinase concentration was 0.13 mg/ml. More
detailed analysis of both the kinetics and concentration

dependence of the GuHCl-induced inactivation, and its

69

TABLE I. Inactivation of Hexokinase in 0.6 M GuHCl in the
Presence and Absence of 1 mM Glc-6-P or 10 mM GlcNAc.

 

Time of Incubation

25 mdnutes
45 minutes

Activity Remaining, % Control'

No Ligand Glc-6-P GlcNAc

76.4 2.8 76.0 5.0
64.6 4.2 69.2 8.0

l+l+

3.2 8 3.6
4.1 7 3.1

l+|+
l+l+

 

‘All values represent the average i S.D. for five measurements.

70
reversibility, are in progress, but these results are not germane
in the present context. It was immediately apparent that the
digestion pattern in the presence of GuHCl was different from the
pattern seen in the absence of this denaturant (Figure 2). The
decreased overall intensity of staining testified to the
extensive proteolysis that had occurred, as might be expected in
the presence of a denaturing agent such as GuHCl. However,
several discrete cleavage intermediates were clearly evident. Of
particular interest in the present context was the observation
that digestion in GuHCl generated two new major fragments having
molecular weights of approximately 52K and 48K (Figure 2). It
was also evident that GlcNAc and Glc-6-P were selectively
protecting the 48K and 52K species, respectively.

The origin of the cleavage products seen in the presence of
GuHCl was determined by immunoblotting (Figure 2), using
monoclonal antibodies recognizing epitopes whose location within
the primary structure of hexokinase had been previously defined
(5). The 52 kDa fragment was immunoreactive with monoclonal
antibodies 18 and 3A2, whose epitopes were known to lie in the
central 50 kDa and N-terminal 10 kDa domains, respectively.

Thus, the 52 kDa fragment represents the N-terminal half of the
molecule. The 48 kDa fragment was immunoreactive with monoclonal
antibody 5A, which recognizes an epitope in the 40 kDa C-terminal
domain: thus the 48 kDa fragment represents the C-terminal half
of the molecule.

Staining of the 10 kDa species by antibody 3A2, (Figure 2,

71

Figure 2. Effects of guanidine hydrochloride, glucose-G-P,
and N-acetylglucosamine on digestion of rat brain hexokinase with
trypsin. Left panel, blot stained for total protein with Amido
iblack. Right panels, immunoblots probed with monoclonal
antibodies 18, 5A, and 3A2, which recognize epitopes located in
the 50, 40, and 10 kDa domains of rat brain hexokinase,
respectively (see text). In each panel: Lane 1, control digest,
jprepared in the absence of GuHCl, giving cleavage fragments with
the sizes indicated at left of figure (unmarked band between the
10K and 40K fragments is trypsin); Lane 2, digested with trypsin
after 60 min denaturation with 0.6 M GuHCl (see Methods); Lane 3,
as in Lane 2, except 1 mM Glc-6-P was present during the
denaturation and subsequent incubation with trypsin; Lane 4, as
in Lane 2, except 10 mM N-acetylglucosamine was present during

the denaturation and subsequent incubation with trypsin.

72

N 23E

mxowc
m<m

.lll'

 

 

«an. vamp

flxovc
<m

 

 

 

flxomo

 

 

 

 

szHOLQ

 

 

 

.... A x OP

Av-oAv
Axon
Av- o0
Axes
Av. OOP

 

73
right panel, lane 1) is barely visible in this photograph but was
clearly evident on the original blot. This species was not
detected after digestion in the presence of GuHCl (lanes 2-4) and
presumably indicates rapid proteolysis of this fragment after its
cleavage from the hexokinase molecule. It might be noted here
that the faint doublet bracketing the 10 kDa species, seen on the
blot stained for protein (Figure 2, left panel), does not arise
from hexokinase but from trypsin (7). The cleavage patterns
seen in the presence and absence of GuHCl are compared in Figure
3. The 10, 40, and 50 kDa fragments resulting from cleavage at
15 and.T; in the absence of denaturant (2), are indicated at the
center of the figure. In the presence of GuHCl, cleavage at a
new site, 15, becomes predominant, generating 48 and 52 kDa
fragments as indicated. Though not evident in Figure 2, a 42 kDa
fragment, reactive with antibody 18 but not with antibodies 5A or
3A2 and formed by cleavage of the 52 kDa fragment at 15, was also
detectable on the original blots (see also Figs. 4-6, below).
I These results demonstrate that GlcNAc, a glucose analog,
protects enzyme activity in the presence of GuHCl and stabilizes
the 48 kDa C-terminal domain against proteolysis. This is
consistent with previous work (7) showing that the substrate
hexose binding site and catalytic function were associated with
this domain. These results also demonstrate that the allosteric
effector, Glc-6-P, does not protect catalytic function associated
with the C-terminal domain, nor does it protect this domain from

proteolysis under these conditions. On the other hand, Glc-6-P

74

Figure 3. Schematic representation of tryptic cleavage pattern
in the presence and absence of guanidine hydrochloride. In the
absence of GuHCl, cleavage occurs primarily at T1 and T2 to give
10, 40, and 50 kDa fragments (2). In the presence of GuHCl,
cleavage at 13 becomes predominant, giving 52 and 48 kDa
fragments originating from the N- and C-terminal regions,
respectively. The various fragments were identified by
immunoblotting techniques (Figure 2) using monoclonal antibodies
recognizing epitopes in the regions indicated at the top

of the figure.

75

m 059“.

 

 

 

 

 

 

 

«... m _p

4 < <
.. _ «ERAsemesters
<0 Tlflllludul.

 

76
does stabilize the N-terminal half of the molecule against
proteolytic attack. 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 the molecule and distinct from the catalytic site located

in the C-terminal domain.

Estimation of the Dissociation Constant for Binding of Glc-6-P
to the N-Terminal Domain. As shown in Figure 4, increasing
Glc-6-P concentrations provided increasing protection of the 42
and 52 kDa fragments, derived from the N-terminus of the molecule
by cleavage in the presence of GuHCl. Total protein in each lane
(excluding trypsin) was estimated by densitometric scanning of
the Coomassie Blue-stained gel and expressed as a percentage of
the total protein in the digest done in the absence of GuHCl
(Figure 4, left lane). This is plotted as a function of Glc-6-P
concentration in Figure 5. The protective effect of Glc-6-P is
evident, and the data appear to conform to a hyperbolic
relationship. Subtracting the residual protein seen in the
absence of added Glc-6-P and plotting in a double reciprocal
format gave a linear result (inset to Figure 5), consistent with
protection resulting from binding of a single Glc-6-P.
Extrapolation of this plot gave an estimated Kd of lo/uM, in
excellent agreement with previous estimates of the dissociation
constant for binding of Glc-6-P to the allosteric regulatory site

of brain hexokinase (1,11).

77

Figure 4. Protection of the N-terminal domain by increasing
Glc-6-P concentrations. Left lane, control digest, in the
absence of GuHCl, giving fragments with sizes as indicated at
left of figure. The other lanes received equal amounts of
hexokinase that had been denatured with 0.6 M GuHCl and
subsequently digested with trypsin (see Methods), with the
concentration (uM) of Glc-6-P indicated at top of figure being
present throughout these treatments. This SDS-gel was stained
with Coomassie Blue. Protection of the 42 and 52 kDa fragments
derived from the N-terminal region of the molecule (see Figure 3)

by increasing Glc-6-P is readily apparent.

78

O 510 2050100

 

 

 

 

Figure 4

79

Figure 5. Estimation of the K; for binding of Glc-6-P, based
on its protective effect against proteolysis. The gel shown in
Figure 4 was scanned densitometrically, and total protein
(excluding trypsin) expressed as a percentage of that seen in the
digest done in the absence of GuHCl (Figure 4, left lane). The
inset shows the data from the ascending portion of the curve,
plotted in double reciprocal format after correction for residual

protein seen in the absence of Glc-6-P.

80

m 9:9“.

 

 

 

 

 

 

 

2: $1728
0: cor om om on om on 0..4 on ON 9 0 oil
_ r _ P _ _ _ _ _ _ _ O
:4 Tinned:
8.0.8.5 . Sac.S.o.2..|_o~_....onf :m
w
I.
12.0 m. 9
d
w
m. 1
Into 0 ON
1mm
0 Ton

 

 

on

 

IOJlUOO % ‘ummd 10101

79

Figure 5. Estimation of the K; for binding of Glc-6-P, based
on its protective effect against proteolysis. The gel shown in
Figure 4 was scanned densitometrically, and total protein
(excluding trypsin) expressed as a percentage of that seen in the
digest done in the absence of GuHCl (Figure 4, left lane). The
inset shows the data from the ascending portion of the curve,
plotted in double reciprocal format after correction for residual

protein seen in the absence of Glc-6-P.

80

m 9:9”.

 

 

 

 

 

 

 

 

2: $18128
0: GOP om om ON. cm on 0+4 on ON OF 0 orl
_ _ L p L _ _ '— _ _ c O
:4 Tinned:
one 8.8.2.8.2..o.o._....o~_....on.u 1m
/
v-
I.
128 M 1m;
.6
m.
Into m. low
TmN
a ton
on

 

 

IOJlUOO % 'UielOJd Iolol

m-"9
it“... ‘—

l n
...—2‘. 7‘.
1ird00"-.‘.
#

0.49.;8 k,
..I“'v¢ U1

Y I -
.3 at.“c‘
“9 ““

l a
-V' F‘
Ao:‘:--"»'e.‘L '

V.“ .‘ne "r

. P ‘ .
”ran-1‘1“

' .
..‘PR‘ “'5‘.
I r‘IVSAUUQ‘

teen Show
:::;e:e “1
sci: less
:ela-tive a

anti-«“4 tc
U

A .
.exck‘ “

-..ase

t

A

an

w‘i' A

AQV
‘

..ati

Ltd; fi
«RVa:

et

'9‘
“\U . '
\LA
651

...-~anhyd:

f

Fs‘D
. ,..:saha‘
‘ b

‘J
. HQ
|‘V Gfi‘

i~U§YSi

81

Correlation between Stabilization of the N-Terminal

 

Domain,Inhibitory Effectiveness, and Conformational Change
Induced by Binding of Hexose 6-Phosphates. As shown in Figure 6,
the ability of various hexose 6-phosphates to protect the 52 kDa
fragment, containing the N-terminal domain, against proteolysis
in the presence of GuHCl is correlated with their ability to
serve as inhibitors of hexokinase (11). Thus Glc-6-P, a potent
inhibitor, provides substantial protection while
l,5-anhydroglucitol-6-P is somewhat less effective. Other hexose
6-phosphates examined (2-deoxyglucose-6-P, fructose-G-P,
galactose-G-P, and mannose-6-P) are all poor inhibitors of the
enzyme and afford little protection against proteolytic digestion
of the N-terminal domain of brain hexokinase. It has previously
been shown (16) that the latter four hexose 6-phosphates do
compete with Glc-6-P for binding to the allosteric site, though
much less effectively than does 1,5-anhydroglucitol-6-P. The
relative ability of these hexose 6-phosphates to compete for
binding to the regulatory site, and to serve as inhibitors of
hexokinase, has been correlated with their ability to induce
conformational change in the enzyme (11). The present results
indicate that one of the consequences of the marked
conformational change induced by binding of Glc-6-P and
1,5-anhydroglucitol, and much less effectively by other hexose
6-phosphates, is stabilization of the N-terminal domain against

proteolysis in the presence of GuHCl.

82

Figure 6. Comparison of various hexose 6-phosphates in
protecting the N-terminal domain against proteolysis. Lane 1,
control digest, prepared in the absence of GuHCl, giving
fragments of the size indicated at left. Lane 2, hexokinase was
denatured with 0.6 M GuHCl and subsequently digested with trypsin
in the absence of added ligands. Lanes 3-8, as in lane 2 except
that various hexose-6-phosphates (all at 1 mM) were present
during denaturation and digestion. Lane 3, Glc-6-P; Lane 4,
1,5-anhydroglucitol-6-P; Lane 5, 2-deoxyglucose-6-P: Lane 6,

fructose-G-P: Lane 7, galactose-G-P; Lane 8, mannose-G-P.

83

 

  

1 '2 3 4 5 e 7 a
100K-» ......“ .... ....

90K» ~— '
60K» “ 7......
50K»

 

40K»

 

 

Figure 6

 

in)

\n"

‘A‘ 5

Plot

11".
fl

84

Synergism between Glc-6-P and Hexose Binding Sites.
Synergistic interactions between the Glc-6-P and hexose binding
sites of brain hexokinase have previously been demonstrated
(11,17). The results presented in Figure 7 demonstrate that the
regulatory and catalytic sites are also able to interact in a
synergistic manner to protect the enzyme against denaturation by
GuHCl. Although neither fructose or Glc-6-P alone provided
appreciable protection against inactivation by GuHCl, in
combination these ligands exhibited a marked protective effect.
In contrast, galactose, whether in the presence or absence of
Glc-6-P, gave no protection against inactivation by GuHCl. These
results are in accord with previous results demonstrating intense
synergistic interactions between Glc-6-P and fructose, but not
between Glc-6-P and galactose, in protecting hexokinase against a
variety of inactivating agents (11). Similar synergistic
interactions between Glc-6-P and fructose, but not between
Glc-6-P and galactose, were evident from the effects of these
ligands in protecting against tryptic digestion in the presence
of GuHCl (Figure 7). Thus, fructose alone provided little or no
protection while, as noted above, Glc-6-P protected the
N-terminal 52 kDa fragment arising by cleavage at.Tg (and the 42
kDa fragment arising from additional cleavage at T3). In
combination, however, these ligands markedly reduced digestion by
trypsin, with most of the protein being found as the intact (100
kDa) molecule or the 90 kDa fragment resulting from cleavage of

the N-terminal 10 kDa domain (see above); in both of these

85

Figure 7. Synergistic interactions between Glc-6-P and
hexose binding sites. Hexokinase was incubated at 259 C either
in the absence (Lane 1) or presence of 0.6 M GuHCl (Lanes 2-7),
with or without addition of ligands, and activity determined
after 45 min. This is shown at the top of the figure, expressed
as a percentage of original activity. After 60 min of
incubation, trypsin was added, incubation continued for an
additional 20 min, and samples prepared for SDS-gel
electrophoresis. The gel was stained with Coomassie Blue. Lane
2, no ligand added; Lane 3, 1 mM Glc-6-P; Lane 4, 10 mM fructose;
Lane 5, 1 mM Glc-6-P plus 10 mM fructose; Lane 6, 10 mM

galactose; Lane 7, 1 mM Glc-6-P plus 10 mM galactose.

86

100 so 67 62 91 58 66
a 4' 5 a 7

100 K»
90 K”

60 K>
50K»

40K»

 

1OK>

 

Figure 7

87
species, the 40 and 50 kDa fragments, which include the binding
sites for hexose substrates and hexose-6-P effectors,
respectively, remain intact. No such protective effect was seen

with galactose.

The re

' e
32:461.,

:zain hex:

.
unvaaae a
0v:.\,:‘ 5.

l- and C-1
"Ms-35.

a...“
‘ o

tuckinase

:"l'n

DISCUSSION

The results presented here clearly indicate that the 50 kDa
fragment, which represents the bulk of the N-terminal half of the
brain hexokinase molecule, contains a high affinity site for
Glc-6-P with all of the characteristics of the allosteric
effector site on this enzyme. In contrast, previous work (6,7)
has demonstrated that catalytic function is associated with the
40 kDa fragment in the C-terminal half of the molecule. This
segregation of catalytic and regulatory functions into discrete
N- and C-terminal regions of the molecule is consistent with the
proposed (18-24) evolutionary relationship between mammalian
hexokinases and less complex hexokinases, such as the yeast
enzyme. The latter have a subunit molecular weight of 50,000
(half that of mammalian hexokinases), and do not exhibit the
allosteric regulation by Glc-6-P characteristic of mammalian
hexokinases. Several investigators have suggested that mammalian
hexokinases have evolved from an ancestral "yeast type"
hexokinase. Thus, it is proposed that duplication of a gene
coding for an ancestral hexokinase with molecular weight about
50,000 followed by gene fusion gave rise to a hexokinase
possessing two catalytic sites, one of which subsequently evolved
into the allosteric regulatory site possessed by present-day
mammalian hexokinases. There is precedent for such an
evolutionary process with the mammalian phosphofructokinases,

which exhibit sophisticated regulatory processes not seen with

88

the si

JR'F‘

wag...

L.,..h
.6-..

these

A

‘gcmpe

‘ -
“F‘s
'5‘A'A“

89
the simpler bacterial phosphofructokinases having molecular
weights approximately half that of the mammalian enzymes.
Poorman et al. (25) have demonstrated internal homology between
the N- and C-terminal halves of mammalian phosphofructokinase and
bacterial phosphofructokinase, as would be expected from such an
evolutionary relationship. Recently, Marcus and Ureta (26) and
Schirch and Wilson (27) have demonstrated homology in amino acid
sequences of the yeast and mammalian hexokinases, providing
further support for the postulated evolutionary relationship of
these enzymes.

Although many investigators have accepted the concept of a
discrete allosteric site for Glc-6-P (reviewed in ref. 1), as
first proposed by Crane and Sols (28), Fromm and colleagues
(29-31) have maintained that the high affinity site for Glc-6-P
is, in fact, the catalytic site, with resultant inhibition due to
competition between the 6-phosphate group of Glc-6-P and the
terminal phosphate of ATP for a common anion binding site on the
enzyme. The present work clearly makes this view untenable.
Glc-6-P does not protect catalytic activity, now known to reside
in the C-terminal domain (6,7), nor does it protect this domain
against proteolysis. It does, as discussed above, protect the
N-tenminal half of the molecule. It might be argued that this is
indirect, and results from binding of Glc-6-P to the C-terminal
domain with resulting conformational changes resulting in
protection of the N-terminal half of the molecule. Such an

explanation is inconsistent with results which demonstrate

(

fl.

90
protection of the N-terminal region under conditions in which the
C-terminal domain, and thus the catalytic and putative Glc-6-P

binding site, is abolished due to extensive proteolysis.

10.
11.

12.

13.

14.

15.

16.

17.

REFERENCES

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

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

Schulz, G.E., and Schirmer, R.H. (1979) Principles 2;
Protein Structure, p. 84, Springer-Verlag, New York/Berlin.

 

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

 

Wilson, J.E., and Smith, A.D. (1985) J. Biol. Chem. 260,
12838-12843.

 

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.

Zetina, C.R., and Goldberg, M.E. (1980) ‘J. Mol. Biol. 137,
401-414.

 

 

Wilson, J.E. (1973) Arch. Biochem. Biophys. 159, 543-549.
Wilson, J.E. (1978) Arch. Biochem. Biophys. 185, 88-99.
Wilson, J.E. (1979) Arch. Biochem. Biophys. 196, 79-87.

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

Ferrari, R.A., Mandelstam, P., and Crane, R.K. (1959)
Arch. Biochem. Biophys. 89, 372-377.

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

 

Decker, L.A., ed. (1977) Worthington Enzyme Manual, p.
221, Worthington Biochemical Corp., Freehold, NJ.

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

Ellison, W.R., Lueck, J.D., and Fromm, H.J. (1975) g.
Biol. Chem. 250, 1864-1871.

91

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

92

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

Easterby, J.S., and O'Brien, M.J. (1973) Eur. g. Biochem.
38, 201-211.

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

Holroyde, M.J., and Trayer, I.P. (1976) FEBS Lett. 62,
215-219.

 

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

Gregoriou, M., Trayer, I.P., and Cornish-Bowden, A. (1983)
Eur. g. Biochem. 134, 283-288.

Manning, T.A., and Wilson, J.E. (1984) Biochem. Biophys.
Res. Commun. 118, 90-96.

 

Poorman, R.A., Randolph, A., Kemp, R.G., and Heinrikson,
R.L. (1984) Nature 309, 467-469.

 

Marcus, F., and Ureta, T. (1986) Biochem. Biophys. Res.
Commun. 139, 714-719.

Schirch, D.M., and Wilson, J.E. (1987) Arch. Biochem.
Biophys., in press.

 

Crane, R.K., and Sols, A. (1954) J. Biol. Chem. 210,
597-606.

 

Purich, D.L., Fromm, H.J., and Rudolph, F.B. (1973) in
Advances in Enzymology (A. Meister, Ed.), Vol. 39, pp.
249-326, John Wiley, New York.

Casazza, J.P., and Fromm, H.J. (1976) Arch. Biochem.
Biophys. 177, 480-487.

 

Solheim, L.P., and Fromm, H.J. (1981) Arch. Biochem.
Biophys. 211, 92-99.

 

FOOTNOTES

1At the time that this manuscript was initially prepared, it was
believed that these three tryptic fragments represented the
structural domains of hexokinase. Comments to this effect exist
in the original manuscript. However, this is not the present
view, and our current understanding of the structure of the
enzyme is more accurately presented in Chapters 4-6. This
chapter has been modified to avoid inconsistencies with

subsequent chapters.

2The use of this analog instead of Glc-6-P itself was based on
the desire to retain the very convenient Glc-6-P
dehydrogenase-coupled assay to follow the kinetics of
inactivation. 1,5-Anhydroglucitol-6-P is not a substrate for
Glc-6-P dehydrogenase and does not interfere with this assay.
While it is possible to use this assay even in the presence of
Glc-6-P by allowing oxidation of endogenous Glc-6-P prior to
initiation of the hexokinase reaction by addition of ATP (14),
this is difficult when successive assays at closely
timed intervals are required, as in the inactivation experiments

depicted in Figure l.

93

Chapter 4

Isolation and Characterization of the Discrete N- and

C-Terminal Halves of Rat Brain Hexokinase.

Like other mammalian hexokinases, the Type I isozyme of
hexokinase (ATP:D-hexose 6-phosphotransferase; EC 2.7.1.1) from
rat brain consists of a single polypeptide chain having a
molecular weight of approximately 100 kDa, and exhibits marked
sensitivity to inhibition by the product, Glc-6-P: the latter is
generally accepted to be a major regulatory influence on the in
3339 activity, and thus on the rate of cerebral glucose
metabolism for which hexokinase represents the initial catalytic
step (reviewed in ref. 1). In contrast to the mammalian
hexokinases, yeast hexokinase has a molecular weight of about 50
kDa and is not inhibited by physiologically relevant levels of
the product Glc-6-P (2). Based on comparison of these properties
of the yeast and mammalian enzymes, several investigators (2-7)
have proposed that the present day mammalian hexokinases evolved

from simpler “yeast-like" hexokinases via a process of gene

94

95
duplication and fusion, with the allosteric regulatory site for
Glc-6-P evolving from what was originally a duplicate catalytic
site.

The complete amino acid sequence for rat brain hexokinase
has been deduced from the nucleotide sequence of cloned cDNA (8),
and extensive similarity in the sequence of the N- and C-terminal
halves of the brain enzyme and yeast hexokinase has been
demonstrated, as would be predicted if the mammalian enzyme were
the product of duplication and fusion of a gene coding for an
ancestral yeast-type hexokinase. Taking similarity in sequence to
be indicative of similarity in secondary and tertiary structure,
and based on these sequence comparisons and the structure of
yeast hexokinase determined by Steitz and his colleagues (9-12),
Schwab and Wilson (8) have proposed a structure for brain
hexokinase which is consistent with various properties of the
enzyme that may be directly related to structure.

Also consistent with the proposed evolutionary relationship
between yeast and mammalian hexokinases are previous studies
indicating that catalytic activity resides in the C-terminal half
of the rat brain enzyme (13-15), while the regulatory site
binding Glc-6-P is associated with the N-terminal half (16). In
the course of the latter work, we observed that allosteric
effectors and substrates could selectively stabilize the N- and
C-termdnal halves, respectively, against tryptic digestion in the
presence of partially denaturing concentrations of guanidine

hydrochloride (GuHCl). We have further developed this approach,

96

permitting isolation of the N- and C-terminal halves of the
enzyme in amounts adequate for more extensive characterization.

The focus of the present report is primarily on the C-
terminal half, which retains virtually full catalytic activity.
Based on the proposed evolutionary relationship to yeast
hexokinase, mentioned above, it would be predicted that the C-
terminal catalytic fragment would not show marked inhibition by
Glc-6-P, which has been associated with binding to a regulatory
site in the N-terminal half of the intact (100 kDa) molecule
(16). This was not the case. Failure to satisfy this prediction
represents the first indication that the evolutionary scenario
outlined above is not adequate to describe the relationship
between yeast and mammalian hexokinases. The results presented
here provide the basis for an alternative evolutionary scheme
that relates the 100 kDa mammalian hexokinases to the 50 kDa
hexokinases of yeast and other organisms. In addition, they
provide further insight into the structure of a mammalian

hexokinase, the Type I isozyme from rat brain.

-
fl.
'5

{:1 (n J

(A)

r3. 9 h

(J

in.

()

MATERIALS AND METHODS

Materials. Rat brain hexokinase was purified by affinity
chromatography on Affi-Gel Blue (Bio-Rad Laboratories,
Richmond,CA) (17). Ultra-pure GuHCl was obtained from
Schwarz/Mann Biotech (Cleveland, OH). 1,5-anhydroglucitol-6-
phosphate had been synthesized according to Ferrari gt 31. (18),
and was the same preparation used in previous studies from this
laboratory (16). Glc-6-P dehydrogenase (from yeast) was purchased
from Boehringer Mannheim (Indianapolis, IN). Thioglycerol, PMSF,
TPCK-treated trypsin, and other biochemicals were from.Sigma
Chemical Company (St. Louis, MO). Fractogel TSK HW-55(S) was

obtained from MCB Manufacturing Chemists (Gibbstown, NJ).

Hexokinase activity, inhibition studies, and protein
detenminations. Unless indicated otherwise, hexokinase activity
was determined spectrophotometrically using the Glc-6-P
dehydrogenase coupled assay as previously described (17); one
unit is defined as the amount of enzyme catalyzing the formation
of 1 umole of Glc-6-P per minute. Inhibition studies were done
under standard assay conditions, except for indicated variations
in substrate concentrations. When inhibition by Glc-6-P was
studied, ADP formation in the hexokinase reaction was coupled to
NADH oxidation gig the pyruvate kinase and lactate dehydrogenase
reactions (19). For all inhibition studies, NADP was prepared in

distilled water rather than the usual 0.1 M sodium phosphate, pH

97

98

7.0, and Glc-6-P dehydrogenase was extensively dialyzed against
0.02 M TrisCl, pH 7.5, to remove ammonium.sulfate present in the
commercial preparation. Inhibition patterns were analyzed using
the Ez-FIT program (20).

Hexokinase concentrations were determined by the absorbance
at 280 mm using a molar extinction coefficient of 5.1 x 10‘M’1
cmfl (21). Trypsin concentrations were also determined from A”,
based on an absorbance of 1.43 for a 1 mg/ml solution (22).

Since preparations of the C-terminal fragment contained
substantial amounts of trypsin, determination of the
concentration of C-terminal fragment (necessary to calculate the
specific activity of this catalytically active fragment) based on
in“ or direct chemical methods was not considered reliable. As an
alternative, the amount of C-terminal fragment was determined by
quantitative densitometry after SDS gel electrophoresis (16),
using a standard curve generated from known amounts of intact

enzyme electrophoresed in parallel with the C-terminal fragment

preparation.

Electrophoretic and immunoblotting procedures. SDS-gel
electrophoresis on 6-20% linear gradient acrylamide gels was
performed as described previously (16). Unless indicated
otherwise, gels were silver-stained (23); alternatively, protein
bands were stained with Coomassie Blue (24). Where indicated,
gels were scanned with a Model GS 3000 scanning densitometer

(Hoefer Scientific Intruments, San Francisco, CA).

99
Immunoblotting procedures were as in earlier work (16).
Three monoclonal antibodies, designated 18, 5A, and 3A2, were
used. These antibodies, and the location of their epitopes within
the overall sequence of rat brain hexokinase, have been described

previously (16).

Molecular weight determinations under nondenaturing
conditions. Samples and molecular weight standards were
chromatographed on an 81 x 2.6 cm column of TSK HW-55(S)
equilibrated with 0.05 M TrisCl, 0.15 M NaCl, 0.5 mM EDTA, 10 mM
thioglycerol, 10 mM glucose, pH 8.5. Chromatography was done at
4° C, and the flow rate was maintained at approximately 20 ml/hr

by a peristaltic pump.

Preparation of the N- and C-terminal halves of rat brain
hexokinase. To enhance stability, the purified enzyme is stored
in a Glc-containing phosphate buffer (17). Prior to use,
protective ligands were removed by chromatography on a Sephadex
G-25 (fine) column previously equilibrated with 0.05 M Hepes, 0.5
mM EDTA, 10 mM thioglycerol, pH 7.5 (HET buffer). Removal of
glucose was verified by enzymatic assay of the hexokinase
fractions.

The C-terminal half of the molecule, hereafter referred to
as the ”C fragment" was prepared in the following manner.
Approximately 300 ug of intact hexokinase in HET buffer was

incubated with 0.4 M GuHCl, lOmM GlcNac, 5 mM ATP, and 6 mM MgCl2

100
in a total volume of 2.0 ml. Temperature of the sample was
maintained at 25° C by means of a thermostatted water bath. After
1 hr of incubation with the denaturant, proteolysis was initiated
by the addition of a stock trypsin solution (2 mg/ml in 0.001 N
HCl) to give a final ratio of 0.2 mg trypsin per mg hexokinase.
After a further 20 min at 25° C, PMSF (200 mM in 95% ethanol) was
added to a final concentration of 1 mM. The sample, in 100-250
ul aliquots, was then dialyzed for one hour against HET buffer
containing 1 mM PMSF using a microdialysis apparatus (Bethesda
Research Laboratories, Gaithersburg, MD) through which buffer
flow was maintained at approximately 60 ml/hr.

In experiments where complete removal of protective ligands
was necessary, additional dialysis was done as follows. After the
initial dialysis, aliquots were recombined, an additional 2 mM
PMSF added, and the sample incubated at room temperature for 30
minutes. After a further addition of 2 mM PMSF (total cumulative
concentration added now 5 mM), 100 ul aliqots were returned to
the microdialysis apparatus and dialysis against HET buffer (flow
rate, 60 ml/hr) continued at 4° C for approximately 24 hours.

The N-terminal half of the molecule, hereafter referred to
as the "N fragment", was prepared by a method essentially the
same as used for the C fragment except that the concentration of
GuHCl was 0.85 M, and the only ligand present during denaturation
was 1 mM Glc-6-P. Also, PMSF was not added to the N fragment
preparation until completion of the initial one hour dialysis

step, which was performed gs. HET buffer in the absence of PMSF.

101
As in the preparation of the C fragment, the N fragment
preparation received two additions of 2 mM PMSF, with an
incubation period of 30 minutes between additions, before
overnight dialysis was performed 33. HET buffer. The stability
of the N fragment at room temperature could be improved if, after
overnight dialysis, the fragment was again treated with 2 mM PMSF
and incubated for 30 minutes at room temperature, followed by a
further addition of 2 mM PMSF and a second overnight dialysis

against HET buffer as described above.

Isolation and N-terminal sequencingjof the C fragment. The C
fragment was isolated after SDS acrylamide gel electrophoresis
using the method of Hunkapillar 32 31. (25), with several
modifications. Protein was visualized by immersion of the gel in
ice-cold 0.2 M KCl. The area containing the 48 kDa C fragment
was sliced from the gel and chopped into approximately 1 mm
cubes. The gel fragments were moved immediately to an
electroelution cell (ISCO, Lincoln, NE) fitted with a dialysis
membrane having a molecular weight cutoff of 12-14 kDa. The gel
fragments were covered with a soaking buffer (2.0% SDS, 0.4 M
ammonium bicarbonate) and overlaid with elution buffer (0.1% SDS,
0.05 M ammonium bicarbonate) to fill the cell. The cell was then
lowered into a buffer circulating electroelution apparatus
(C.B.S. Scientific Co., Del Mar, CA) which was filled with
elution buffer. Samples were soaked for four hours, and

subsequently electroeluted at 50 V overnight. The elution buffer

102
in the tank was replaced with dialysis buffer (0.02% SDS, 0.01%
ammonium bicarbonate), and elution-dialysis continued at
approximately 80 V for 24 hours. SDS gel electrophoresis of the
electroeluted sample confirmed homogeneity, and N-terminal
sequencing of the fragment was performed by the Macromolecular

Structure Facility, Michigan State University.

Effectiveness of ligands at protecting activity of the C
fragment against denaturation by GuHCl. Aliquots of the isolated
C fragment in HET buffer, with or without addition of specified
ligands, were equilibrated at 25°l3. Inactivation was initiated
by the addition of a stock 6.0 M GuHCl solution (prepared with
HET buffer) so that the final concentration in the sample was 0.6
M GuHCl. Aliquots were removed after thirty minutes for assay of
residual activity; after dilution into the assay solution, the
concentration of GuHCl was such that it had no effect on
activity. Control samples, incubated in the absence of GuHCl,

lost no significant activity (< 2%).

Effectiveness of ligands at protecting N and C fragments
from proteolysis in the presence of GuHCl. N or C fragments were
incubated for 60 min at ZS’in HET buffer containing 0.6 M GuHCl
and indicated concentrations of various ligands; samples were
generally 100-250 ul in size, with fragment concentrations of 15-
20 ug/ml. Trypsin was then added to give a final concentration of

3-4 ug/ml. Digestion was allowed to proceed for 20 minutes, at

103
which time PMSF was added to a final concentration of 2 mM.
Samples were then dialyzed for 30 minutes against HET buffer
containing 1 mM PMSF prior to SDS polyacrylamide gel
electrophoresis; as noted previously (16), failure to reduce the
GuHCl concentration prior to addition of the Laemmli sample
buffer resulted in formation of a precipitate which rendered the

sample unsuitable for electrophoresis.

Elution of N and C fragments from Affi-Gel Blue. A 50% (v/v)
slurry of Affi-Gel Blue in HET buffer containing 75 mM KCl was
prepared. Aliquots (150-250 ul) were placed in 1.5 ml microfuge
tubes, followed by addition of 100-200 ul of fragment preparation
(in HET buffer also containing 75 mM KCl); the inclusion of 75 mM
KCl was prompted by preliminary experiments which showed that A
trypsin did not adsorb to the Affi-Gel Blue under these
conditions. After brief mixing, the gel was sedimented by
centrifugation for a few seconds in a microfuge, and the
supernatant removed. The pellet was washed by resuspension in 0.5
ml HET buffer containing 75 mM KCl followed by brief
centrifugation. This was followed by a similar wash in a buffer
containing 0.05 M TrisCl, 0.5 mM EDTA, 10 mM thioglycerol, pH 9.0
(TET buffer); in experiments with the C fragment, the latter also
contained 20% (v/v) glycerol. The washed gel was then resuspended
in 150-200 ul of TET buffer (with glycerol in experiments with
the C fragment) containing the indicated concentration of any

added ligands. After centrifugation, the supernatant was analyzed

104
for eluted N or C fragment. For the N fragment,_this was done by
quantitative SDS gel electrophoresis, while for the C fragment,

both electrophoresis and activity measurements were used.

RESULTS

Preparation of N and C fragments. The same general strategy
was used in preparing both N- and C-terminal halves of the
molecule. It was based on previous work (16, 26) which delineated
sites highly susceptible to attack by trypsin. The tryptic

cleavage pattern is depicted schematically below.

 

404
1'1 7:
10x l 50K _ 1 40K
52K l 48K
. 1',

In the native enzyme, cleavage is largely restricted to two
sites, designated as T1 and T1, yielding fragments of
approximately 10, 50, and 40 kDa which correspond to the extreme
N-temminus, more central portion, and C-terminal region,
respectively (26); 90 kDa and 60 kDa fragments, formed by
cleavage at only T1 or T2. are seen as intermediates (Figure 1,
Lane D). Tryptic digests prepared under these conditions were
routinely included as standards on SDS gels, shown in other
figures below. Identification of tryptic fragments generated
under partially denaturing conditions (16), as in preparation of
the N and C fragments, was facilitated by immunoblotting using
antibodies 3A2, 18, and 5A, which recognize epitopes lying within
regions defined by the 10, 50, and 40 kDa tryptic fragments

produced from the native enzyme.

105

106

Perturbation of the native structure by low concentrations of
the denaturant, GuHCl, results in much more extensive
proteolysis, with metastable species of 52 and 48 kDa,
corresponding to N- and C-terminal halves of the molecule, being
formed as intermediates (16). The latter correspond to cleavage
at a new site, designated.T3 (see schematic representation
above), which is located near the point at which the internally
repeated sequences are fused (8). (Further comments regarding the
disposition of tryptic cleavage sites in the molecule and their
relation to the proposed structure for rat brain hexokinase can
be found in Schwab and Wilson(8).) Addition of ligands binding to
either the N- or C-terminal half of the intact enzyme results in
selective stabilization of the corresponding half (16), providing
the means to generate substantial quantities of the discrete N
and C terminal fragments, as done in the present work.

Digestion of the 100 kDa rat brain hexokinase under the
conditions described for generation of the N fragment (see
Methods), resulted in virtually complete loss of catalytic
activity (recovery of < 1.0% of the initial activity) and
produced 10, 42, and 52 kDa species (Figure 1, Lane 1). The 52
kDa fragment corresponds to cleavage at.15 to generate the N-
terminal half of the molecule as demonstrated by its reactivity
on immunoblots (results not shown, but see ref. 16) with
monoclonal antibodies 3A2 and 18. The 52 kDa fragment did 325
react with monoclonal antibody 5A, which binds to an epitope

within the C-terminal half of the molecule. The 10 and 42 kDa

107

Figure l. SDS-Polyacrylamide gel patterns of N and C fragment
preparations. Lane D, limited tryptic digest of native
hexokinase, prepared under the conditions of Polakis and Wilson
(26); digestion under these conditions has been well
characterized (see text), and generates fragments with molecular
weights indicated at the left of the figure. Samples in Lanes 1
and 2 were N fragment and C fragment preparations, respectively,
obtained by proteolysis of the partially denatured enzyme in the
presence of ligands offering selective protection of the N- or C-

terminal half of the molecule.

”(10"3

\u;‘.

too- ....
90’ - _

| n v
. p
“ p." "'4'
v , l
. . _’ ....
. _u ‘ .. .
- .‘-l. .
' ...-.,.-l . _
. ,7 . .
. n
' 0.! ..Z. '
n A" '
. \‘n
. 71'.
l ' I. .0
’ .-.R o_.' . s
l"~ .-
. - 7’ uh . .
.. v...‘_e'. \ . J. .
. _ "r.o.l .-l"... ' ‘1 .v
r - 4“ n-
. . . r ‘ , ~ - ‘ ~. \‘
a . q . ‘ 1“, .‘ “
D a . - p; '~ ‘1’: ‘ ~ ‘1'...

 

. \ u ' .' - ._, >-
. 0‘. n/. .4 _... C '
\ . ...,» ‘:‘”Y,("rpfl"\’:_.fi_
J- "“,\ 1' 1" . :v‘-”\'- t

‘ . - ,6 .
. _. a Wu (H _- _‘w .2125,
- ‘ A «,- 4 1 -.: . . .
. . . .. _ , 1:? .3“ .1 a.) ‘v
8‘“ .."-r v W .' .~~Q
‘ ~ ... ~ :f. dt-sait-«oqu
- " - 1. .7. , :‘(J .‘"t'/.».‘.'m'_~l5_13‘-.
, .- a." . 4.2-.4. 0“: 1' E? v o .“l‘g, _‘.
~' ‘ ' I‘; 1.501 'a} .’.'_%.:m
' I i V ‘ " ' l.
‘ I ' I}; “i; "(an IL", -'§' {‘2 ~ w .
I ... I .' 4‘ ‘ ‘- t 5.?

.0 . ,I ‘
" ’ ~ ' '- . o .
'I_ ah. pt 4: .l..~I L .

- ;~’.. ' ..." "" I “'1‘ . ’1“ '~‘-{.".'J'
- 7 " '.l n.’ ~‘ i. Wig}! -

'4
'u , ‘ . Iv 5 ... ‘ _‘.~
' ’3 -. _ -. --..:.. .- ..
a. so . ‘ ~45 I
\ . .c _ ..2 i.
‘ 5" A u- . - .
"e‘.'-:-~‘-V,‘ . .».l - ,.
l-c , .. I" " " . . .
. ‘ ‘D u-‘ .. ‘ 0 f’ l
w. . ‘ . >
. e
r ’ ,' 'v ' . - ..‘A‘
'1. a ‘ - .0

Figure 1

109
species arise by further cleavage of the 52 kDa fragment at Tfi
this was confirmed by immunoblotting which demonstrated that
neither fragment was reactive with antibody 5A while the 10 kDa
species reacted only with antibody 3A2 and the 42 kDa species
only with antibody 18.

The presence of both 10 and 42 kDa species despite extensive
dialysis that could be expected to remove freely diffusing
species of 10 kDa molecular weight, suggests that the 10 kDa
fragment remained associated with the 42 kDa fragment, i.e., as a
"nicked" 52 kDa species, prior to encountering the harsh
denaturing conditions of SDS gel electrophoresis. This is
consistent with the proposed structure (8) in which the
corresponding regions are engaged in extensive noncovalent
interactions.

Digestion of the enzyme under the conditions described for
preparation of the C fragment did n93 cause virtually complete
loss of activity. The average activity (1 SD) retained in 11
different preparations of the C fragment was 36 i 6 % of the
initial activity. When examined by SDS gel electrophoresis, a 48
kDa component was the major species present (Figure 1, Lane 2);
that this component corresponded to the C-terminal half of the
molecule, derived by cleavage at.T;, was confirmed by its
immunoreactivity with antibody 5A but lack of reactivity with'
antibodies 3A2 and 18. Also present were variable (from
preparation to preparation) but always lesser amounts of the 40

kDa species, reactive with antibody 5A but not with 3A2 or 18,

110
which arises by cleavage at T2 (26) .

To confirm that the residual activity in the C fragment
preparation could be attributed to the smaller species,
principally the 48 kDa fragment, the molecular weight of
catalytically active components was determined by molecular sieve
chromatography under nondenaturing conditions (Figure 2). It was
evident that the catalytically active component in the C fragment
preparation chromatographed as a distinctly smaller species than
the native enzyme, and examination of the catalytically active
fractions by SDS gel electrophoresis confirmed that this was
indeed the 48 kDa fragment (Figure 2). Based on comparison with
standards chromatographed on the same molecular sieve column, the
molecular weight of the catalytically active C fragment was
estimated to be 50 kDa (Figure 3), very similar to the 48 kDa
estimated by SDS gel electrophoresis.

The specific activities determined for two different
preparations of the C fragment were 100 and 120 u/mg protein.
This is approximately twice the specific activity, 60 u/mg
protein, reported for the unmodified hexokinase (17,21) and
indicates that km for the C fragment is virtually identical to
that of the intact enzyme.

N-terminal sequence of the C fragment. The N-terminal
sequence detemmined for the C fragment was identical to that at
positions 463-471 in the amino acid sequence deduced from the
cloned cDNA (8), namely, Leu-Ala-Glu-Gln-His-Arg-Gln-Ile-Glu.

This defines the T, cleavage site as being at Argm. Based on the

111

Figure 2. Molecular sieve chromatography of the C fragment
preparation under nondenaturing conditions. Panel A shows the
elution pattern after chromatography of a mixture of C fragment,
intact (100 kDa) hexokinase, and cytochrome c. Hexokinase
activity (0), Au," (0). Panel B shows components detected after
SDS polyacrylamide gel electrophoresis. Lane A, standard digest
of native hexokinase (26), containing fragments with molecular
weights indicated at left. Lane B, intact hexokinase, eluting in
Peak 1 of Panel A. Lane C, C fragment, eluting in Peak 2 of Panel

A. This gel was stained with Coomassie Blue.

112

N 959".

e
. . I. ‘on

,, tenant.

M-.‘8

 

loo

ab— x i

044/0 ‘MIAHOV

m_E .cozsm U_o oE3_o>

O in com com com OWN

0.
‘—"
1

O.
N
1

O.
“’P

 

 

INd

14.0

Tod

two

 

 

09

L110 093 ‘eouoqlosqv

113

Figure 3. Molecular weight of the catalytically active C
fragment, determined by molecular sieve chromatography under
nondenaturing conditions. This figure represents a composite of
several chromatographic runs, each with various combinations of
mol. wt. standards, with or without added C fragment. As can be
seen from the replicate points plotted for lactate dehydrogenase
and cytochrome c, elution volumes were quite reproducible. The
single point for the C fragment is the average for two runs which
gave elution volumes within 1 ml of the plotted value. The

line was determined by linear regression.

114

m 059i

com mmw 0mm m$m CAN mmm owm mmm 0mm .mwm owN 0mm onm

2E 50:25 .6 oE:_o>

 

o mEoEootAo

    

«contact 0 0583930

30:30on 5.032

085083on 3303 o

no

to;

In;

 

og>1 'NbleM lolnoelow 50‘]

115
deduced amino acid sequence (8), the actual molecular weights of
the N and C fragments are calculated to be 51,585 and 50,749,
respectively. To maintain consistency with previous publications
fromtthis laboratory (8,13-16,26), we will continue to refer to
the N and C fragments as having molecular weights of 52 and 48kDa
(the estimates from SDS gel electrophoresis), respectively,
though obviously this is not precisely correct.

Inhibition of the C fragment binlc-6-P and the Glc-6-P
analog, 1,5-anhydroglucitol-6-P. Two different laboratories have
shown that brain hexokinase possesses a single high affinity
inhibitory site for Glc-6-P and its analogs (27-29); that site
has been associated with the N-terminal half of the 100 kDa
intact enzyme (16). Since the N-terminal half of the molecule is
obviously not present in the C fragment, it would follow that
inhibition of the C fragment by Glc-6-P and its analogs would not
be expected. Hence it was surprising to find that the Glc-6-P
analog, l,5-AnG6P, was quite effective as an inhibitor of the C
fragment. As with the intact enzyme (30), inhibition by l,5-AnG6P
is linearly competitive gs. ATP (Figure 4); previous studies have
shown that inhibition of intact hexokinase by Glc-6-P itself is
also linearly competitive gs. ATP (28, 31, 32). The analog, 1,5-
AnGGP (which is not a substrate for Glc-6-P dehydrogenase), was
used in the studies reported here so that the convenience of the
Glc-6-P dehydrogenase-coupled assay could be retained. However,
more limited studies using the pyruvate kinase-lactate

dehydrogenase-coupled assay confirmed that Glc-6-P itself also

116

Figure 4. Inhibition of the C fragment by l,5-AnG6P, with ATP
as varied substrate. Concentration of glucose was 3.33 mM, and
l,5-AnG6P was 0 (O), 0.010 (V). 0.025 (D), or 0.050 (A) mM.
Inset: secondary plot of slope gs. 1,5-AnG6P concentration; the

apparent Kg, determined by the EZ Fit program (20), was 0.01 mM.

117

 

e 0.59....

.25 .32-&$\ i

CID

_F'

2: 5.32:.

003380—
_._._._._

.

 

 

 

 

 

I ‘ '
.. v- "! ‘
-...‘-

‘.
4

... vi. -: vi .~ g ‘F~ .‘ o a Flu . .

.... a .1 ... .m. s i a e. a .m .... ......
: .i . .6“ h: . c. a .

......" a. .5“ C . u ... .3 m . . 4m a.» 0. mu.

.5 v A a . .nJ a. .. . . a if. .01..» mla \\/I and. W4“ urn.“ «new.

118

inhibited the C fragment.

1,5-AnG6P was a linearly uncompetitive (gs. Glc) inhibitor
of the C fragment (Figure 5). Previous studies with the intact
rat (31) or bovine (32) brain enzymes have indicated that
inhibition by Glc-6-P was mixed noncompetitive gs. Glc, though
examination of the published data indicates that the double
reciprocal plots did not stray far from the parallel pattern
characteristic of uncompetitive inhibition, and in one study with
bovine brain hexokinase (33), inhibition by Glc-6-P was
explicitly stated to be uncompetitive vs. Glc. In short,
inhibition of the C fragment by Glc-6-P or its analog, l,5-AnG6P,
is similar to that seen with the intact enzyme - competitive gs.
ATP and uncompetitive, or nearly so, gs. Glc.

The relative effectiveness of various hexose 6-phosphates
at inhibiting the C fragment appears to be similar to that seen
with the intact enzyme (34), decreasing in the order: 1,5-AnG6P
>> Fru-6-P > 2-DeoxyGlc-6-P > Man-6-P - Gal-6-P. For example,
under assay conditions in which 0.1 mM 1,5-AnG6P inhibited
activity by 85%, 0.1 mM Fru-6-P inhibited only 24%, and 0.1 mM
Man-6-P less than 2%. As with the intact enzyme (30), inhibition
appears to require a dianionic group at the 6 position; Glc-6-
sulfate did not inhibit the C fragment.

Inhibition of the C fragment by inorganic phosphate and its
analogs. With the intact enzyme, relatively low concentrations
of Pilantagonize the inhibition by Glc-6-P, an effect that can be

directly attributed to the mutually competitive binding of these

119

Figure 5. Inhibition of the C fragment by 1,5-AnG6P, with
glucose as varied substrate. Concentration of ATP was 6.6 mM, and
1,5-AnG6P was 0 (O), 0.050 (v). 0.10 (D). or 0.20 (A) mM. Inset:
secondary plot of intercept gs. l,5-AnGGP concentration; the
apparent K; was 0.06 mM. A linear uncompetitive inhibition
pattern was again observed when this experiment was repeated with

ATP at 1.1 mM, with an apparent K1L of 0.02 mM.

120

 

 

 

m 28E
.58 .momoo:_3\—
on 04 on 0N OF 0
_ _ _ _ _ 0
®\\...
o\
>\
\>\ to?
\mm\
4
«\.3§25\10N
3N - JG n _ O
.. .
M
[.0—
w -0...

 

 

 

on

A/[

121
two ligands (28) . At much higher concentrations (K1: 35 mM), an
inhibition by P1 itself, competitive v_s. ATP, becomes evident
(28 ) . The response of the C fragment to P1 was markedly different

from that of the intact enzyme, with linear competitive

inhibition (gs. ATP) being seen even at low [P1] (Figure 6).

Linear noncompetititve inhibition was seen v_s. Glc (Figure 7) .

The inhibitory effectiveness of several analogs of P1 was

examined. Arsenate, sulfate, and P1 were similar in this respect

(Figure 8); the absence of detectable inhibition by acetate
indicates that this is not simply a nonspecific ionic strength
effect‘.

The similarity in response of the C fragment to P1,
an: senate, and sulfate prompted us to reexamine the effect of
these anions as antagonists of inhibition of the intact enzyme by
the Glc-6-P analog, l,5-AnG6P. It had been reported (19) that
arsenate and P1 were effective at reversing inhibition of rat
brain hexokinase by Glc-6-P while sulfate was not; in contrast,
Kosow 2E _a__l_. (35) found all three anions to be capable of

reversing inhibition of the Type I isozyme from tumor cells by

Glc-G-P. In the present study, all three anions - but not

acetate - were found to be similarly effective in reversing
inhibition by l,5-AnG6P (Figure 9); modest inhibition, perhaps
analogous to that previously described for P1 (28), was also seen
with all three anions. The antagonism by these anions was not

Peculiar to the inhibition by l,5-AnG6P because other

experiments, using the pyruvate kinase-lactate dehydrogenase

122

Figure 6. Inhibition of the C fragment by inorganic phosphate,
with ATP as varied substrate. Concentration of glucose was 3.33
mM, and P1 was 0 (O), 5 (v), 15 (D), and 25 (A) mM. Inset:

secondary plot of slope _v_s_. P1 concentration: the apparent K1 was

6 mM.

123

o 229....

..25 .Hmzigx F
n N m

— p _ -

0\Mvm
\

\

 

 

:2 cages...
% 8

 

 

 

 

124

Figure 7. Inhibition of the C fragment by inorganic phosphate,
with glucose as varied substrate. Concentration of ATP was 6.6
mM, and P1 was 0 (0), 15 (n), 30 (A), and 45 (v) mM. Inset:
secondary plot of intercept gs, Pg concentration;the apparent K1

was 60mM.

 

125

.4. 2:9“.

L2: .«oe x Homoo:_3\e
N F

_ _

~4-
‘—ro

 

   

o >
n
.\m§
o \ as .352...
m » ... . ._. . ._. . ...

e

 

 

 

 

women“

 

o
rm

.2 W
1:
row

126

Figure 8. Inhibition of the C fragment by inorganic phosphate
and its analogs. Sodium salts of the indicated anions were used.

ATP concentration was 1.1 mM and Glc 3.33 mM.

ActIVIty, % Original

127

 

 

 

 

100-
80-
60-
40-
H Acetate
' H Pi
H Sulfate
20 H A'rsenatel . I .
0 10 20 30

Concentration, mlvl

Figure 8

128

Figure 9. Effectiveness of inorganic phosphate and its analogs
at reversing inhibition by l,5-AnGGP. Activity was measured with
3.33 mM Glc, 0.66 mM ATP, 0.1 mM 1,5-AnG6P, and increasing
concentrations of the sodium salts of the indicated anions. In
the absence of added 1,5-AnG6P, the activity was 5.7 umol/min,

and there was no activation by added anions.

129

a 059i

3:: .cozobcoocoo

 

 

 

 

 

one . om. cm on o 0.0
8.0500 0
80:09.0 0
30:3 a
r 30:39.3 0 to;
O

 

 

 

ulLu/loum’ ‘Ail/xlpv

130
coupled assay employed in the previous study (19), confirmed that
similar effects were seen with Glc-6-P as inhibitor. The reason
for the previous (19) inability to detect the reversal of Glc-6-P
inhibition by sulfate is not known.

The ability of various ligands to protect the C fragment
against denaturation and subsequent proteolysis. As with the
intact enzyme, the isolated C fragment becomes highly susceptible
to proteolysis following denaturation in 0.6 M GuHCl. Various
ligands were examined for their effectiveness in protecting the C
fragment against such denaturation and proteolysis.

The ability of several hexoses and hexose 6-phosphates to
protect the activity of the C fragment against GuHCl
denaturation is compared in Figure 10, and the ability of these
same ligands to protect against subsequent (to incubation with
GuHCl) proteolysis is shown in Figure 11. These results make
several things apparent. First, the ability of hexoses to
protect the fragment against denaturation may be related to their
ability to serve as a substrate in the hexokinase reaction. Thus,
glucose and mannose, both good substrates for brain hexokinase
(34,36), provide substantial protection, while a poor substrate
such as fructose provides only marginal protection.
N-Acetylglucosamine, a glucose analog which binds at the active
site but is not phosphorylated (36) provides an intermediate
degree of protection. The ability of various hexoses to serve as

substrates has been correlated with their ability to induce

131

Figure 10. Effectiveness of various ligands in protecting the C
fragment against denaturation by GuHCl. The remaining activity
was determined after 30 min incubation of the C fragment with 0.6
M GuHCl and the indicated ligands (see Methods). Results are
expressed as a percentage of the initial activity. The unlabeled
bar indicates the results with no added ligand, while the other
bars show the results with various hexoses, GlcNAc (GLCN), or
hexose 6-phosphates. Hexoses and GlcNAc were added at 10 mM and
hexose 6-phosphates at 1 mM concentrations. Each bar represents

the meathSD from three independent experiments.

 

 

til\\\\\\\\\\\\\\\\\\\\\\\\\\\

 

 

'— \\\\\\\\
t\\\\\\\\\\\\\\\\\\\\
t \\\\\\\
\\\\\\\
FH\\\\\\\\K
i\\\\\\\\\\\\\\\\\\ \

 

 

T

 

 

 

T

 

 

 

T

 

 

 

 

 

 

 

+

 

l\\\\\\\\\\\\\\\\\\\\\\\\\\\\

 

 

 

l-€t\\\\\\

 

 

 

100-1

604

l
O

l
O O
(I) Vi' (\i

Iowf‘wo % ‘Mwmv

GGP M6P

Fi'iu F'iiu

GLC MAN GLCN FRU GBP M6P

Figure 10

133

Figure 11. Effectiveness of various ligands in protecting the C
fragment against proteolysis in 0.6 M GuHCl. This gel was
obtained in one of the three experiments mentioned in the legend
to Figure 10; similar results were seen in the other two
experiments. After 1 hr incubation in GuHCl, with or without
added ligand(s), trypsin was added to initiate proteolysis. After
20 min, PMSF was added and the samples prepared for SDS gel
electrophoresis (see Methods). Lane D, standard tryptic digest of
intact hexokinase (26), generating fragments with the mol. wt.
indicated at the left. Lane 1 is a control, and represents C
fragment incubated in the absence of either GuHCl or trypsin, but
otherwise treated as the experimental samples shown in Lanes 2-
10. The following listing indicates the ligand(s) added to
samples in Lanes 2-10, and the amount of C fragment, determined
by densitometric analysis and expressed as a percentage of the C
fragment in the control lane (Lane 1), is given in parenthesis:
2, none (34); 3, Glc (104); 4, Man (104); 5, GlcNAc (63); 6, Fru
(48); 7, G1c-6-P (36); 8, Man-6-P (34); 9, Glc-6-P plus Fru (73);
10, Man-6-P plus Fru (48). Hexoses and GlcNAc were added at 10 mM

and hexose 6-phosphates at 1 mM.

134

 

: 059.“.

|. .441. .I .l .... t .-

U
III ’I’ ...it I . - . .. ‘0”
A...

- ‘8

cl
6
a
22
x
h
2

Refinance“.

135
specific conformational changes in the enzyme, as reflected by
protection of activity against a variety of inactivating agents
(34). The present results suggest that this may also be the case
with the isolated C fragment, and that the effectiveness of
various hexoses is similar to that seen with the intact enzyme.

Secondly, although it is clear from inhibition studies that
Glc-6-P is able to bind to the C fragment, Glc-G-P along provides
no protection against denaturation and proteolysis. This is
consistent with results obtained with the intact enzyme (16) in
which, in the absence of other ligands, Glc-6-P selectively
protects the N-terminal half of the molecule.

Thirdly, synergistic effects between hexose and hexose 6-
phosphate binding sites on the intact enzyme have been described
previously (28, 34). It is evident that similar interactions can
occur between hexose and hexose 6-phosphate sites within the C-
terminal half of the molecule. Thus, although neither fructose
nor Glc-6-P provides substantial protection against GuHCl and
subsequent proteolysis, when present together, marked protection
is observed. Such interactions are not seen when Man-G-P replaces
Glc-G-P.

The results of a similar experiment to examine the ability
of P1.and its analogs to protect against inactivation by GuHCl
and subsequent proteolysis are presented in Figure 12. Acetate
had no significant effect, while P3 and sulfate provided
substantial protection. Arsenate destabilized the C fragment,

increasing both the activity loss during incubation with GuHCl

136

Figure 12. Effectiveness of inorganic phosphate and its analogs
at protecting the C fragment against denaturation by GuHCl and
subsequent proteolysis. The results represent the mean'iSD for
three independent experiments. The experimental protocol was
identical to that described in the legends to Figs. 10 and 11,
except for the ligands added. The abscissa shows the activity
remaining after a 30 min incubation with GuHCl (analogous to the
values shown in Figure 10), while the ordinate represents the
amount of C fragment remaining after further incubation with
GuHCl and proteolysis (determined by densitometric analysis of a
gel such as that shown in Fig.11), expressed as a percentage of
the C fragment in the control sample incubated in the absence of
GuHCl and with no trypsin added. Ligands added as sodium salts,
all at 20 mM: none (0): acetate (v): arsenate (0): phosphate (I);

sulfate (A).

137

 

 

 

 

iOO

 

 

60

_lonuog) % ‘ugeion

 

_O
00
r a

i E -"
-% : co
s
O
T ' l (\l

O O 0

fi' N

Activity, % Original

Figure 12

138

and the loss of protein during the subsequent period of
proteolysis.

Elution of the C fragment from Affi-Gel Blue. Cibacron Blue
F3GA has been shown to be a potent inhibitor of brain hexokinase,
competitive 33. ATP, and similar effects have been seen with a
number of nucleotide binding enzymes (37). This ability of
Cibacron Blue F3GA to serve as a nucleotide analog makes the
immobilized dye generally useful in the purification of these
enzymes by affinity chromatography. In fact, that is the
currently favored method for purification of rat brain hexokinase
(17), with the enzyme being adsorbed to Affi-Gel Blue followed by
elution with Glc-6-P, a competitive inhibitor of nucleotide
binding (38). i

In the present study, it has been observed that both the N
and C fragments could be adsorbed to Affi-Gel Blue, and eluted by
inhibitory hexose 6-phosphates such as Glc-G-P but not by
noninhibitory analogs such as Gal-6-P (see below). This
observation in itself is worthy of note, implying as it does the
existence of nucleotide binding sites on both N- and C-terminal
halves of the molecule, with binding gig these sites modulated in
a similar manner by specific hexose 6-phosphates. This is
obviously consistent with the view that ligand binding sites on
the two halves of the molecule may share features that reflect a
common evolutionary origin, as expected for a product of gene
duplication and fusion.

Since trypsin did not adsorb to the matrix under the

139
conditions used, this offered a potential approach to
purification of the fragments. However, the recovery of the
fragments was low despite a number of attempts to improve it, and
thus this did not prove useful as a preparative method.
Nevertheless, adsorption of the fragments to Affi-Gel Blue and
subsequent elution by various ligands did offer another approach
for gaining at least qualitative information about the effects of
various ligands on the nucleotide binding site.

Glc-6-P was moderately effective at eluting the C fragment
frdm Affi-Gel Blue, while Gal-6-P was not (Fig.13). Inclusion of
Glc in the elution medium.markedly enhanced elution by bgth
hexose 6-phosphates, though Glc itself was not effective as an
eluting agent; this again reflects the synergistic interactions
between hexose and hexose 6-phosphate binding sites (34),
referred to above. That Glc was indeed enhancing elution, and not
simply stabilizing activity during the washes preceding elution
(which would be reflected as an apparent increase in subsequently
eluted activity), was confirmed by the finding that virtually
identical amounts of activity were eluted by Glc-G-P in the
presence of Glc, regardless of whether the preceding washes were
done with or without Glc in the buffers.

The ability of various hexose 6-phosphates to elute the N
fragment from Affi-Gel Blue was compared, with results shown in
Fig.14. Effectiveness in eluting the N fragment generally
correlated with efficacy as an inhibitor of the intact enzyme,

with Glc-6-P and 1,5-AnGGP being quite effective while Gal-6—P

140

Figure 13. Elution of the C fragment from Affi-Gel Blue. The C
fragment was bound to Affi-Gel Blue and elution performed as
described in Methods. In one set of samples, 10 mM glucose was
included in the buffers throughout the procedure. Ligands added
to elution buffer: 1, none; 2, 1 mM Glc-G-P; 3, 1 mM Gal-G-P.
Results presented are from an experiment in which duplicate
samples were assayed, and data represent an average : one half

the range.

V

UIIILD \A

IULUI

141

 

 

[2:] No Glc

52 10 mM Glc

 

 

 

.. ZZZ

 

 

 

 

 

 

 

5

3
2
1-
O

Q: XV BE: _38

"Figure 13

142

Figure 14. Elution of the N fragment from Affi-Gel Blue. The N
fragment was bound to Affi-Gel Blue and elution performed as
described in Methods. Aliquots of eluted N fragment were then
analyzed by SDS gel electrophoresis, and the amount of eluted
fragment quantitated by densitometry. Ligands (all at 1 mM) added
to elution buffer are listed below, and the amount of N fragment
eluted, expressed as a percentage of that eluted with Glc-G-P, is
given in parenthesis: 1, none (insufficient elution to allow
reasonable quantitation); 2, Glc-6-P (100); 3, 1,5-AnG6P (125);
4, Gal-G-P (30); 5, Man-6-P (48); 6, Fru-6-P (94); 7, 2-
deoxyglucose-G—P (69). Lane D, standard tryptic digest of intact
hexokinase (26), which generates fragments with the mol.wt.

indicated at the left.

143

er163 D 1234567
100..
90"

50’ 1- —-—- ....

     

  

_ a..."
I a O
.‘ fl . ~oor~~
I
‘
1 I
'\ o /
‘ /
‘ y
\
s | .
. .1 .7 ...
'Yps'n" - -. 1 . ' ~
‘ u ..s . 01- .
a . ~ , ' ,r
; . ' . ‘ fi . ‘
v 1 ‘_l l l.’ LIL" (n . ..1 - ., ~ ... i
. . ._. ’ ' - . ~
- : r . .- :I- .1 ..L.-. .e. -, 3;},- ' #48,. ... . y _“ ..."
r . _ .. 1. ~ .,-. . ; v; t -
I ... I S ‘ ,, 3 ' . N‘. “ . .. \ . , o
\. ._‘ g .2' .‘. ‘. 1v . '-._‘I ~. W'u'u' '. 17'
7"“! ; ' ' ' ~.. . ‘ > (‘fi .. ’2'; {

Figure 14

there
posse
{See

that

144
and Man-6-P were much less so. The correlation was not absolute,
however, since both Fru-6-P and 2-deoxyglucose-6-P were more
effective as eluting agents than might have been anticipated
based on their ineffectiveness as inhibitors of the intact enzyme
(16,34). Thus, the specificity of the hexose 6-phosphate binding
site in the isolated N fragment is apparently somewhat different
from that seen for binding of hexose 6-phosphates to the N—
terminal half of the intact enzyme (16).

Inclusion of 10 mM Glc in the eluting medium had no effect
on elution of the N fragment by these hexose 6-phosphates. Since
there is evidence to indicate that the N fragment does, in fact,
possess sites for binding both hexoses and hexose 6-phosphates
(see below), the inability of Glc to affect this elution suggests
that synergistic interactions do not occur between these sites on
the isolated N fragment, in contrast to the situation with the
intact enzyme (16,34) or isolated C fragment.

The ability of various ligands to protect the N fragment
against denaturation and subsequent proteolysis. The ability of
various ligands to protect the N fragment against denaturation in
0.6 M GuHCl and subsequent proteolysis is compared in Figure 15.
In contrast to the rather remarkable stability of the C fragment,
appreciable loss of the N fragment was observed during incubation
at 232 even in the absence of added GuHCl or trypsin (compare
Lanes 1 and 2 in Figure 15). The cause of this loss remains
unclear. Attempts to reduce adsorptive losses (e.g., by

silanizing vessels, coating them with bovine serum albumin (39),

145

Figure 15. Effectiveness of various ligands at protecting the N
fragment against denaturation with GuHCl and subsequent
proteolysis. Lane D, standard tryptic digest of intact hexokinase
(26), containing fragments with the mol. wt. indicated at the
left. Lane 1 is a control, and represents a sample of N fragment
which was incubated with no added GuHCl or trypsin, but otherwise
treated as samples in Lanes 2-11. Lane 2 represents a similar
sample, incubated with GuHCl but no trypsin. Samples in Lanes 3-
11 were incubated with GuHCl and trypsin (see Methods) plus the
indicated ligand: 3, none: 4, Glc-6-P; 5, 1,5-AnG6P; 6, Man-6-P:
7, Gal-6-P: 8, Glc; 9, Man; 10, GlcNAc: 11, Gal. Concentrations
of hexoses (and GlcNAc) and hexose 6-phosphates were 10 mM and 1

mM, respectively.

146

m. 2:9...

 

Pwn¥ 010.5 0 m v.0 N p.

..
.
r . .
.lv (1 .

 

..O—.

152:...

ADV

A00
‘60

AOOP

TOP x .2

or it
albu:
a va:
Show}
agai1
EffEI
latti

were

enzy

147
or inclusion of low concentrations of Triton X-100 (40) or serum
albumin in all solutions) were without effect, as was addition of
a variety of protease inhibitors. Whatever the cause, the results
shown in Figure 15 indicate that ligands effective at protecting
against denaturation by GuHCl and proteolysis were similarly
effective at preventing this loss of unknown origin. Despite the
latter complication, the protective effects of various ligands
were clearly apparent (Figure 15).

As with protection of the N-terminal half of the intact
enzyme (16), the effectiveness of various hexose 6-phosphates was
related to their efficacy as inhibitors. Thus, Glc-6-P and 1,5-
AnGGP afforded substantial protection (Figure 15, Lanes 4 and 5)
while Man-6-P and Gal-6-P (Figure 15, Lanes 6 and 7) did not. It
should be noted that this ability to protect the N fragment in
the absence of added hexose is distinctly different from the
effect seen with the isolated C-terminal half of the molecule,
the C fragment, in which the presence of a hexose was required
for manifestation of a protective effect by 6-phosphate
derivatives (see results above).

The presence of a hexose binding site on the N fragment was
indicated by the marked protective effect of hexoses such as Glc
or Man (Figure 15, Lanes 8 and 9). In contrast, N-
acetylglucosamine and Gal (Figure 15, Lanes 10 and 11) did not
offer significant protection. It will be noted that these
protective effects again are correlated with the relative ability

of these hexoses to serve as substrates for the enzyme.

_vua

assc
the

hexc

DISCUSSION

Evolutionary relationships among the hexokinases. There has been
considerable evidence, summarized in the introduction to this
paper, to support the suggestion (2—7) that the 100 kDa Glc-6-P-
sensitive mammalian hexokinases evolved from an ancestral 50 kDa
yeast-type hexokinase by a process of gene duplication and
fusion. According to this suggestion, catalytic activity will be
associated with one half of the mammalian hexokinase molecule -
the half that has retained the function of the ancestral
hexokinase - while the other half will be associated with
regulatory function, with the originally duplicated catalytic
site having evolved into an allosteric site for binding Glc-6-P.
Implicit in this is the potential for isolating a catalytically
active “half molecule"; based on previous work (13-16), it was
anticipated that this would be the C—terminal half of brain
hexokinase. The present work has realized that potential and
confirmed that the C-terminal half of the molecule is indeed the
site of catalytic function in brain hexokinase.

In view of the extensive similarity in sequence between rat
brain and yeast hexokinases (8), there can be little doubt that
they are indeed homologous (41) enzymes. However, the present
results, which demonstrate that the catalytically active C-
temminal half of rat brain hexokinase is itself sensitive to
potent inhibition by Glc-6-P, suggest that the ancestral form

most proximately related to the mammalian hexokinases was not a

148

hexo)
by G:
silky

C21 C1111

POSS.
hexoi

dip;

149
Glc-6-P-insensitive yeast-type enzyme but rather a 50 kDa
hexokinase that had already acquired a sensitivity to inhibition
by Glc-6-P. Do such hexokinases exist? Indeed they do - in
silkworm (42) and marine organisms such as starfish, sea
cucumber, and sea urchin (43,44). In fact, various properties of
starfish hexokinase bear a striking similarity to those of the C
fragment derived from the rat brain enzyme (Table I). To our
knowledge, there is no sequence information currently available
for these latter hexokinases, but it seems very likely that they
will be found to share extensive similarities with the mammalian
and yeast enzymes. These considerations lead us to propose an
alternative evolutionary scheme (Figure 16) which maintains the
role of gene duplication and fusion, suggested earlier (2-7), but
explicitly recognizes the probable acquisition of the inhibitory
response to Glc-6-P prior to the gene duplication and fusion
event that gave rise to the mammalian enzymes.

If, in fact, sensitivity to Glc-6-P - and thus the
possibility for physiologically important feedback inhibition of
hexokinase activity - had been acquired before the gene
duplication and fusion event leading to the mammalian enzymes,
what possible selective advantage could there be in the latter?
We believe that at least one important factor is the remarkably
different response to P1.'The 50 kDa starfish enzyme, and its
presumed homolog, the C fragment of the rat enzyme, are both
quite sensitive to inhibition by P3, a metabolite that has long

been known to play a central role in regulation of energy

Mal. Wt
Spec. I
3. (nu-‘4)
for C
for A
Inhibit

Glc-6

'Inhibi

150

Table I. Comparison of Starfish Hexokinase with the C

Fragment of Rat Brain Hexokinase

 

 

Starfish Hexokinase‘ C Fragment

M01. Wt. 51 kDa 51 kDa
Spec. Act. (u/mg) 113 110
K. NM!)

for Glc 0.05 0.06

for ATP 0.14 0.88
Inhibitionb

Glc-6-P 2- ATP C, K1-0.055 mM -

l,5-AnG6P 1;. ATP - c, K1-0.01 mM

Glc-6-P 33. Man M, K,-0.09 mM -

1,5-AnG6P ls. Glc - U, K1-O.02 mM

91 E- ATP c, K1-4.7 mM c, K1-6 mM

P1 .‘E- Glc N, K1-6.5 mM N, K1-60 mM

 

' Unpublished results of Dr. John S. Easterby.
” Inhibition patterns: C, competitive; M, mixed noncompetitive;

N,noncompetitive: U, uncompetitive.

151

Figure 16. Proposed evolutionary relationship between the
hexokinases. The catalytic site is represented by a filled
circle, and the G1c—6-P-binding regulatory site by a filled

square. The open circle and open square represent sites which

are latent in the intact enzyme.

152

9 2:9".

omOSxOXOI Caz—DEED:

.-lionnr o.l.li

 

 

30:30on £535 Till?- L T .lu-lll.
iliOlIii

 

305x98: 3.5322

30:30on «new»

 

 

153
metabolism (45-47). The antagonism between inhibition by Glc-6-P
and its relief by Plihas figured prominently in discussions of
the regulation of hexokinase in mammalian systems (1,47). It is
apparent that such antagonism does not exist in the simpler 50
kDa hexokinases. Hence a major physiologically important
consequence of the duplication and fusion event was that it
converted the inhibitory response to phosphate into what is, in
effect, an activating response (in the sense that it counteracts
inhibition by Glc-6-P), conferring on the mammalian enzyme a
sophisticated mode of regulation not seen with the simpler
enzymes.

Ligand binding sites on mammalian hexokinases. In the
originally suggested evolutionary scenario (2-7), it was
postulated that the Glc-6-P-binding regulatory site arose from
what had been a catalytic site, presumably with concomdtant loss
of the ability to bind substrates, Glc and ATP. Hence, one might
expect to find a single site (on the catalytic half of the
molecule) for each of the substrates, while the regulatory half
had acquired a single binding site for Glc-6-P. This fits very
neatly with results indicating the binding of both ATP and Glc to
the C-terminal catalytic half (13-15), and Glc-6-P to the N-
terminal regulatory half (16) of the intact 100kDa enzyme. And
direct measurements of ligand binding have, in fact, indicated
that there is a single site for binding of Glc and Glc-6-P”*with
physiologically relevant affinity (27,28,48); as implied by the

synergistic interactions between these ligands (28,34), the sites

154
are discrete and both Glc and its phosphorylated derivative can
be bound simultaneously to the enzyme.

This simple (at least in principle) picture of the mammalian
hexokinase molecule must now be reevaluated. Obviously,
generation of the mammalian enzyme by fusion of two smaller
hexokinases, similar in properties to the present-day starfish
enzyme, creates the potential for duplication of all sites
present in the starfish enzyme. In other words, unless they have
been lost during the subsequent evolutionary process, a mammalian
hexokinase might possess Egg sites for each of the ligands
capable of binding to the 50 kDa precursor enzyme, i.e., two
sites for Glc, ATP, Glc-6—P, and P1.'The present study has
provided direct evidence for the existence of sites, similar in
their specificity, for hexoses and hexose 6-phosphates on both
the N- and C-terminal halves of rat brain hexokinase. How can
this be reconciled with previous work indicating a single site
for Glc and for Glc-6-P? It obviously implies that one of the
sites for each ligand is latent in the intact enzyme. Based on
previous work (14,15), the hexose site in the N-terminal half is
nonfunctional, as is the Glc-6-P binding site in the C-terminal
half (16). Freed from the constraints present in the intact

enzyme, both sites are revealed in the N and C fragments.

ACKNOWLEDGEMENT

We are very grateful to Dr. John S. Easterby, Department of
Biochemistry, University of Liverpool, United Kingdom, for
providing the results with the starfish hexokinase, shown in
Table I, and for granting permission to include them in this

manuscript.

155

10.

11.

12.

13.

14.

15.

16.

17.

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

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

Easterby, J.S. and O'Brien, M.J. (1973) Eur. g. Biochem. 38,
201-211.

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

 

Holroyde, M.J., and Trayer, I.P. (1976) FEBS Lett. 62, 213-
219.

 

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

Gregoriou, M., Trayer, I.P., and Cornish-Bowden, A. (1983)
Eur. g. Biochem. 134, 283-288.

 

 

Schwab, D.A., and Wilson, J.E. (1989) Proc. Natl. Acad. Sci.
USA, in press. -

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

 

 

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

 

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

Harrison, R. (1985) Crystallographic Refinement of Two
Isozymes 2; Yeast Hexokinase and Relationship of Structure to
Function. Ph.D. Thesis, Yale University, New Haven, CT.

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. 2 4, 385—396.

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

 

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

Wilson, J.E. (1989) Prep. Biochem., in press.

 

156

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

157

Ferrari, R.A., Mandelstam, P., and Crane, R.K. (1959) Arch.

Biochem. Biophys. 89, 372-377.

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

 

Perrella, F.W. (1988) Anal. Biochem. 174, 437-447.

 

 

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

Worthington, C.C., ed. (1988) Worthington Manual, p. 320,
Worthington Biochemical Corp., Freehold, N.J.

Wray, W., Boulikas, T., Wray, V.P., and Hancock, R. (1981)
Anal. Biochem. 118, 197-203.

Wyckoff, M., Rodbard, D., and Chrambach, A. (1977) Anal.
Biochem. 18, 459-482.

Hunkapillar, M.W., Lujan, E., Ostrander, F., and Hood, L.E.
(1983) in Methods in Enzymology (Hirs, C.H.W., and Timasheff,
S.N., Eds.), Vol. 91, 227-236, Academic Press, New York.

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

Chou, A.C., and Wilson,-J.E. (1974) Arch. Biochem. Biophys.
165, 628-633.

Ellison, W.R., Lueck, J.D., and Fromm, H.J. (1975) g. Biol.
Chem. 250, 1864-1871.

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

Wilson, J.E., and Chung, V. (1989) Arch. Biochem. Biophys.
269, 517-525.

Grossbard, L., and Schimke, R. T. (1966) g. Biol. Chem. 241,
3546-3560.

 

Copley, M., and Fromm, H.J. (1967) Biochemistry 6, 3503-3509.

Fromm, H.J., and Zewe, V. (1962) g. Biol. Chem. 237, 1661
-1667.

 

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

 

 

Kosow, D.P., Oski, P.A., Warms, J.V.B., and Rose, I.A. (1973)
Arch. Biochem. Biophys. 157, 114-124.

36.

37.
38.
39.
40.
11

41.

42.

43.

44.
45.
46.

47.

48.

158

Sols, A., and Crane, R.K. (1954) J. Biol. Chem. 210, 581-595.

Wilson, J.E. (1976) Biochem. Biophys. Res. Commun. 12, 816-
823.

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

Felgner, P.L., and Wilson, J.E. (1976) Anal. Biochem. 14,
631-635.

Suelter, C.H., and DeLuca, M. (1983) Anal. Biochem. 135, 112-
9.

 

Reeck, G.R., deHaen, C., Teller, D.C., Doolittle, R.F.,
Fitch, W.M., Dickerson, R.E., Chambon, P., McLachlan, A.D.,
Margoliash, E., Jukes, T.H., and Zuckerkandl, E. (1987) Cell
‘59, 667.

Yanagawa, H.-A. (1978) Insect Biochem. 8, 293-305.

Mochizuki, Y., and Hori, S.H. (1977) J. Biochem. (Tokyo) 81,
1849-1855.

Mochizuki, Y. (1981) Comp. Biochem. Physiol. 70B, 745-751.

Wu, R., and Racker, E. (1959) g. Biol. Chem. 234, 1029-1035.

 

Wu, R., and Racker, E. (1959) J. Biol. Chem. 234, 1036-1041.

 

Rose, I.A., Warms, J.V.B., and O'Connell, E.L. (1964)
Biochem. Biophys. Res. Commun. 15, 33-37.

Ellison, W.R., Lueck, J.D., and Fromm, H.J. (1974) Biochem.
Biophys. Res. Commun. 51, 1214-1220.

 

FOOTNOTES
1It was conceivable that added anions might compete with ATP for
Mg”, and thereby reduce availability of ATP Mg” required as a
substrate. However, this was not the source of the observed
inhibition since increasing MgCl2 concentration had no effect on

activity.

2A second binding site with very low affinity for Glc-6-P is

thought to represent binding at the substrate hexose site (29).

159

Chapter 5

Anion Binding Sites in the N-Terminal Region of

Rat Brain Hexokinase

Hexokinase (ATP: D-hexose 6-phosphotransferase, EC 2.7.1.1)
catalyzes the initial step in glucose metabolism, converting it
to glucose-G-P using ATP as the phosphoryl donor. This reaction
assumes special importance in brain, for which glucose normally
represents virtually the sole substrate supporting a highly
active energy metabolism (1). Regulation at the hexokinase step
is a major factor governing the overall rate of cerebral glucose
utilization (2,3) and, as might be expected, the enzyme is
subject to rather SOphisticated control mechanisms (reviewed in
4). A major goal of our laboratory has been to establish the
structural basis for the catalytic and regulatory properties of
brain hexokinase.

Potent inhibition of brain hexokinase activity by the
product, Glc-6-P, was first reported by Weil-Mahlerbe and Bone
(5). These observations were followed by the classic study of

Crane and Sols (6), which led to the view that the enzyme

160

161
possessed a discrete (from the catalytic site) regulatory site to
which the Glc-6-P could bind. A few years later, Monod EE.§$- (7)
presented a more comprehensive treatment of the concept of a
distinct ligand binding site uniquely associated with regulatory
function, the "allosteric“ site. Inhibition by Glc-6-P is
competitive gs. ATP-Mg” (8-10), and is thought to result from a
conformational change, induced by binding of Glc-6-P at the
allosteric site, that results in loss of the ability to bind the
nucleotide substrate (11). Inhibition by Glc-6-P is antagonized
by P,(10,12), an effect that may be attributed to the mutually
exclusive binding of these ligands (10). It is generally accepted
that inhibition of hexokinase by its product, Glc-6-P, and
antagonism of this effect by P“ represent major factors in the
regulation of hexokinase, and thereby the glycolytic rate, in
brain (4,13) as well as other mammalian tissues (14-16).

In contrast to the mammalian hexokinases, yeast hexokinase
is not inhibited by physiologically-relevant levels of Glc-6-P
(17). Another striking difference is that the mammalian enzymes
have M,s of about 100,000 (4), whereas yeast hexokinase has an Mr
of approximately 50,000. These observations led several
investigators (18-23) to suggest that the mammalian hexokinases
evolved by duplication and fusion of a gene coding for an
ancestral yeast-type hexokinase, with the allosteric Glc-6-P
binding site evolving from what had been a duplicated catalytic
site. This suggestion received direct support with the finding of

extensive similarity between the deduced (from cloned cDNA) amino

162
acid sequences of the N- and C-terminal halves of rat brain
hexokinase and yeast hexokinase (24). Furthermore, the finding
that catalytic function was associated with the C-terminal half
of the molecule (25-27) while the regulatory Glc-6-P binding site
was associated with the N-termdnal half (28) was also in accord
with what would be predicted from the proposed evolutionary
scheme.

The isolation of the C-terminal half of rat brain hexokinase
with full retention of catalytic activity and susceptibility to
inhibition by Glc-6-P forced a reevaluation of this evolutionary
scenario (29). As modified, present-day mammalian hexokinases did
indeed arise by duplication and fusion of a gene coding for an
ancestral hexokinase. However, the latter was not a yeast-type
enzyme as originally proposed (18-23), but more like the
hexokinases found in starfish and other marine organisms (30,
31), which have M, of about 50,000 El susceptibility to
inhibition by Glc-6-P, i.e., the latter characteristic evolved
prior to the gene duplication and fusion event. In the present
work, we demonstrate the presence of a vestigial nucleotide
binding site in the N-terminal half of rat brain hexokinase and
propose a relationship between this site and the effects of Glc-
6-P and P1 on activity of this enzyme. Integration of this and
previous studies with the evolutionary history of mammalian
hexokinases has permitted development of a model that represents
the relative disposition of ligand binding sites and how they
might interact to produce the sophisticated regulatory behavior

characteristic of brain hexokinase.

MATERIALS AND METHODS

Materials. Rat brain hexokinase was purified as previously
described (32). Ultrapure GuHCl was obtained from.Schwarz/Mann
Biotech (Cleveland, OH). 1,5-Anhydroglucitol-6-P was synthesized
according to Ferrari et al. (33). Yeast Glc-6-P dehydrogenase
was a product of Boehringer Mannheim Biochemicals (Indianapolis,
IN). PMSF, TPCK-treated trypsin, and other biochemicals were from

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

Hexokinase Activity and Protein Detenminations. Hexokinase
activity was determined spectrophotometrically using a Glc-6-P
dehydrogenase- coupled assay (32). Inhibition by the Glc-6-P
analog, l,5-anhydroglucitol-6-P, was studied under these same
assay conditions, except that the MgCl,iconcentration was 7.5 mM
and ATP, 0.66 mM. When reversal of this inhibition by
tripolyphosphate was studied (see below), the MgCl2 concentration
was increased to maintain a 1-5 mM excess over the concentration
of tripolyphosphate; under these conditions, Mg” concentrations
(required for formation of the ATP-Mg” used as substrate) were
not a limdting factor.

Hexokinase and trypsin concentrations were determined from
absorbance at 280 nm, based on a molar extinction coefficient of
5.1x10‘ M‘1 cm‘1 for hexokinase (34) and an absorbance of 1.43 for

a 1 mg/md trypsin solution (35).

163

164

Proteolytic Digestion in the Presence of Partially
Denaturing Concentrations of GuHCl. As in our previous studies
(28, 29), the effect of various ligands on the structure of rat
brain hexokinase was assessed by examining their ability to
protect specific regions of the molecule against tryptic
digestion in the presence of partially denaturing concentrations
of GuHCl. The procedures used have been described in detail
(28,29). Briefly, the enzyme (0.13 mg/ml) was incubated for 1 hr
at 25° C in HET buffer (0.05 M Hepes, 0.5 mM EDTA, 10 mM
monothioglycerol, pH 7.5) containing 0.6 M GuHCl, with various
ligands added where indicated. Trypsin (2mg/ml in 1 mM HCl) was
then added to a final concentration of 0.026 mg/ml, and
incubation continued for 20 min prior to analysis by SDS
polyacrylamide gel electrophoresis. Control digests were prepared
identically except that no GuHCl was present. Identification of
proteolysis products was confirmed by immunoblotting, using
monoclonal antibodies recognizing epitopes of defined location
within the overall amino acid sequence. The locations of
preferential tryptic cleavage sites within the molecule and of
the epitopes for the monoclonal antibodies have been given in
previous publications (28, 29) and are conveniently summarized in

Figure 3 of reference 28.

RESULTS

Effectiveness of nucleoside triphosphates at protecting rat
brain hexokinase against inactivation in GuHCl. As reported
previously (28), substantial loss of activity is seen during
incubation of rat brain hexokinase in the presence of 0.6 M
GuHCl. Marginal protection, at best, against such inactivation
is seen in the presence of added nucleoside triphosphates (Table
I). When Mg” is also added to produce the chelated forms,
inactivation is significantly increased, with the effect of ATP-
Mg“ being particularly remarkable and resulting in virtually
complete loss of activity within 30 min (Table I). Since
catalytic activity is associated with the C-termdnal half of the
molecule (25-27, 29), these results indicate a destabilization of
this region by the chelated forms of these nucleotides. The
distinctive effect (compared with chelates of other nucleoside
triphosphates) of ATP-Mg” on GuHCl-induced inactivation can be
correlated with the specificity for ATP-Mg” as a substrate (8);
the Mg’+ chelates of other nucleoside triphosphates do not serve
as effective substrates (8) and are extremely poor inhibitors
with Kg'values in the high mM range (36, 37). Thus, these results
would be consistent with the view that destabilization results
frmm binding of ATP-Mg” to the substrate nucleotide binding site
located in the C-terminal half of the molecule (25). Similar
results were seen in a previous study employing chymotryptic

digestion or heat denaturation as inactivating agents (11).

165

166

TABLE I. Effect of Nucleoside Triphosphates on Inactivation of
Rat Brain Hexokinase in 0.6 M GuHCl.

 

 

Nucleotide Added‘ Activity Remaining (% Controlfl_
None 68': 2
ATP 76 i 5
GTP 74 i 7
CTP 74 i 7
UTP 73 1 5
ATP-Mg 15‘: 3
GTP-Mg 50‘: 1'
CTP-Mg 63 i 3
UTP-Mg 48': 3'

 

'Nucleotide concentrations were 5 mM; MgCl®,*where added, was 7.5
mM. MgCl2 alone had no effect on inactivation.

bActivity remaining after 30 minutes of incubation in 0.6 M GuHCl
at 25°<:., expressed relative to control which contained no GuHCl
or added ligand and which lost no significant activity (< 5%)
under these incubation conditions. Values given are means‘: SD
from three separate experiments, except for values marked with
asterisk, which are means 3 half the range in two experiments.

167

Effectiveness of nucleoside triphosphates at protecting rat
brain hexokinase against partial denaturation in 0.6 M GuHCl and
subsequent tryptic digestion. In HET buffer alone, hexokinase is
subject to limited tryptic cleavage to give the species seen in
Figure 1 (Lane 1). The location of these fragments with respect
to the overall sequence of the enzyme is not germane in the
present context. Suffice it to say that, in conjunction with the
immunoblotting results shown in Figure 2, these cleavage
fragments can be shown to be identical to species characterized
in previous studies of the tryptic digestion pattern seen with
this enzyme (28, 29, 38). What is important in the present
context is that inclusion of 0.6 M GuHCl during the incubation
and subsequent treatment with trypsin results in extensive
proteolysis (Figure 1, Lane 2; note the markedly reduced protein
staining compared with Lane 1), with formation of metastable
species having molecular masses of 48 and 52 kDa (faint bands
seen in Lane 2 of Figure 1) which have been shown (29) to
correspond to the C- and N-terminal halves of the molecule,
respectively. This is confirmed in the present study by the
immunoblotting results in Figure 2, i.e., the 48 kDa fragment is
reactive with monoclonal antibody 5A which recognizes an epitope
in the C-terminal region but not with antibodies 18 or 3A2 which
react with epitopes lying in the N-terminal half of the molecule
while the opposite pattern of immunoreactivity is seen with the
52 kDa fragment.

Inclusion of nucleoside triphosphates during the incubation

168

Figure 1. Effects of nucleoside triphosphates and their Mg”
chelates on proteolysis of rat brain hexokinase in 0.6 M
guanidine hydrochloride. Tryptic cleavage products, generated as
described in Methods, were analyzed by SDS gel electrophoresis:
the Coomassie Blue stained gel is shown. Lane 1, control digest,
done in the absence of guanidine: products include fragments that
have been characterized previously (28, 29, 38), with molecular
weights indicated at left of figure. Other lanes received equal
amounts of hexokinase that had been denatured in 0.6 M GuHCl, and
subsequently digested with trypsin, either without (Lane 2) or
with addition of indicated ligands. Lanes 3-6 show effect of
addition of 5 mM ATP, GTP, CTP, and UTP, respectively. Lane 7, 5
mM ATP plus 5 mM ”;Cl,. Lane 8, 5 mM GTP plus 5 mM MgC1,.IResults
with CTP and CT? in the presence of MgCl2 (not shown) were
indistinguishable from those with CTP-Mg“ (Lane 8). Positions of
48 kDa and 52 kDa fragments, corresponding to C-terminal and N-
terminal halves of the molecule (29), respectively, and which
will be of particular interest in the present context are

indicated at the right of the figure.

169

Mr x 163

 

100—
so-
so-
50»
4o» -_ _1 .

 

10’

 

Figure 1

170

Figure 2. Immunoblot analysis of tryptic digestion products.
Replicate SDS gels were run and electroblotted to nitrocellulose.
The panel at the left is a blot stained for protein using amido
black. Other panels are replicate immunoblots, probed with
previously described (28, 29) monoclonal antibodies 3A2, 18, or
5A, which recognize epitopes in distinct regions of the molecule
represented by the 10, 40, and 50 kDa tryptic fragments,
respectively, that were characterized by Polakis and Wilson (38).
In each panel: Lane 1 is a control digest, done in the absence of
added GuHCl; Lane 2, digest in 0.6 M GuHCl with no added ligand:
Lane 3, digest in 0.6 M GuHCl plus 5 mM ATP. Positions of the 48
kDa and 52 kDa fragments, corresponding to the C-terminal and N-
terminal halves of the molecule (29), respectively, are shown at

right in each panel.

171

N 230E

 

 

 

 

vice xom v.9 320$
Ao-
r 152;:
.... in... Ii... I. ..- O?! .... (O?
i ..l v i «0‘ i a. .... tom
.... I 1. Too
\Iliiut 'III- 1"!!! '8" ‘OG
160..

 

 

 

 

 

 

 

 

131—L 2 . ludi- . . a .6222

172
with GuHCl and subsequent proteolysis results in selective
protection of the 52 kDa fragment against digestion, with
ATP,GTP, CTP, and UTP all being effective (Figure 1, Lanes 3-6);
that this is indeed the 52 kDa fragment corresponding to the N-
terminal half of the molecule (28,29) is confirmed by
immunoblotting results (Figure 2). Nucleoside triphosphates also
decreased the rate of cleavage of the intact enzyme, as indicated
by the increased intensity of the 100 kDa band in Lanes 3-6 of
Figure 1 compared to that in Lane 2.

Inclusion of Mg" had a dramatic effect on the protection
afforded by ATP; in contrast to protection of the N-terminal
region seen with the free nucleotide, binding of the Mg” chelate
resulted in destabilization to the point of inducing virtually
complete proteolysis (Figure 1, Lane 7). A much more moderate
decrease (relative to that seen with the unchelated form) in the
protection afforded by other nucleoside triphosphates was also
observed, with GTP shown as a representative (Figure 1, Lane 8)‘.

Unchelated nucleoside triphosphates had no detectable effect
on proteolysis of a 48 kDa fragment corresponding to the C-
terminal half of the molecule (28,29), with minor but comparable
amounts of this species being detected after denaturation and
proteolysis in either the absence (Figure 1, Lane 2) or presence
(Lanes 3-6) of these ligands. However, when Mg” was present, the
C-terminal fragment was reduced to barely detectable levels with
nucleoside triphosphates other than ATP (e.g., GTP as shown in

Figure 1, Lane 8), and was completely proteolyzed in the presence

173
of ATP-Mg” (Lane 7). These results are consistent with a
destabilizing effect of chelated nucleoside triphosphates on the
C-terminal region, with ATP-Mg” being most effective, as deduced

from effects on activity (Table I).

Phosphate and its analogs protect the N-termdnal half of rat
brain hexokinase against denaturation with GuHCl and subsequent
proteolysis. Figure 3 shows the results of a similar experiment
in which Piiand other anions were included during the incubation
with GuHCl and subsequent proteolysis. P1, sulfate and arsenate
all resulted in substantial protection of the 52 kDa N-terminal
fragment; the identity of this fragment was again confirmed by
immunoblotting results (not shown). Acetate did not provide
protection (Figure 3, Lane 4), with proteolysis proceeding to
equivalent extents in either the absence (Figure 3, Lane 2) or
presence (Figure 3, Lane 4) of this anion.

Stabilization of the intact (100 kDa) enzyme by P5 and
sulfate, and destabilization by arsenate, were also evident. This
may reflect a propagation of the effects from a ligand binding
site in the N-terminal half to the C-terminal domain in the
intact enzyme. Alternatively, the ligand may be interacting with
a second binding site in the C-terminal domain itself (see
below). Stabilization of the isolated C-terminal half by P3.and
sulfate, and destabilization by arsenate, were observed
previously (29). The observed results would be consistent with

similar effects in the intact enzyme.

174

Based on the previous work with the isolated C—terminal
domain (29), stabilization and hence accumulation of the 48 kDa
C-terminal fragment in the presence of'Pg or sulfate might have
been expected, but this was not seen (Figure 3). The reason
for this is not known, but it presumably reflects differences in
the mode of proteolytic attack on the isolated C-terminal domain
as compared with the intact enzyme in which interactions with the
N-termdnal region may influence proteolysis and effects of ligand

binding on that process.

Pyrophosphate and tripolyphosphate protect the N-terminal
half of rat brain hexokinase against denaturation with GuHCl and
subsequent proteolysis. The above results indicated no detectable

difference in the effectiveness with which various unchelated

 

nucleoside triphosphates protect the N-terminal region of
hexokinase against denaturation and proteolysis; thus, the purine
or pyrimidine moiety plays little if any role in eliciting the
protective effect. Together with the observation of a similar
protective effect seen with P3 and other anions, this focused
attention on the polyphosphate sidechain of the nucleoside
triphosphates. Accordingly, the ability of analogs of this region
to exert a similar protective effect was examined. As with the
nucleoside triphosphates (Figure 1) and with P1 and its analogs
(Figure 3), selective protection of a 52 kDa fragment was seen
when pyrophosphate or tripolyphosphate was present (Figure 4).

Identification as the N-terminal half of the molecule was again

175

Figure 3. Effect of Piiand analogs on proteolysis of rat brain
hexokinase in 0.6 M guanidine hydrochloride. This is a Coomassie
Blue stained SDS gel showing products of proteolysis in: Lane 1,
control digest containing no GuHCl; Lane 2, digest containing 0.6
M GuHCl only; Lanes 3-6, GuHCl plus 5 mM sulfate, acetate,

arsenate, and phosphate (all as sodium salts), respectively.

176

--'~‘"-

‘52
‘48

 

tryPSin"

 

 

 

‘ Figulre 3

177

Figure 4. Effect of P1.and homologs on proteolysis of rat brain
hexokinase in 0.6 M GuHCl. This is a composite showing Coomassie
Blue stained SDS gels on which proteolysis products are
displayed. Lanes labeled A are control digests done in the
absence of GuHCl, while lanes labeled B show the results of
proteolysis in GuHCl with no added ligands. Other lanes show
digest done in GuHCl with indicated additions: Lanes 1-3, In at
l, 5, and 15 mM, respectively; Lanes 4-8, pyrophosphate at 1,
2.5, 5, 7.5, and 10 mM, respectively; Lanes 9-13,
tripolyphosphate at l, 2.5, 5, 7.5, and 10 mM, respectively.
Position of 48 and 52 kDa fragments corresponding to C-terminal
and N-terminal halves of the molecule (29), respectively, are

shown at right.

178

O?!
va

v 9:9...

 

 

.i 0" .0

i

. 4".
I... ' ' ' III ' " al' Ill-

‘ ‘
i ‘ , furl...

_

4

...OP

152::

.... 1°?
M ...6m

 

8
“mm.

....o. x .2

179

confirmed by immunoblotting experiments (results not shown).

Anionic species that selectively protect the N-temminal region
of rat brain hexokinase against denaturation and subsequent
proteolysis also reverse inhibition by the Glc-6-P analog, l,5-
anhydroglucitol-G-P. As shown above, Pu and analogs such as
sulfate and arsenate protect the N-terminal half of the enzyme
against denaturation and proteolysis, while acetate is not
effective in this regard. This correlates with the relative
ability of these anions to reverse inhibition by the Glc-6-P
analog, l,5-anhydroglucitol-6-P (29). Results shown in Figs. 5
and 6 demonstrate that other ligands affording protection to the
N-terminal half of the molecule are also effective at reversing
this inhibition. .

Since it is the Mg” chelate of ATP that is the actual
substrate for hexokinase, with the unchelated form being a
competitive inhibitor (36,37), it obviously was necessary to
maintain an excess of Mg” during the inhibition studies to avoid
complications resulting from competition between ATP and added
anionic species for a limited pool of Mg”. Chelation of
nucleoside triphosphates decreases their protective effect
(results above), and it seems likely that ability of these
compounds to reverse inhibition by l,5-anhydroglucitol-6-P might
have been similarly affected. Nonetheless, even in the presence
of excess Mg”, a modest reversal of inhibition by 1,5-
anhydroglucitol-6-P could be demonstrated with relatively low

concentrations of nucleoside triphosphates (Figure 5)’. At higher

180

Figure 5. Effectiveness of P,.and various nucleotide-Mg“
complexes at reversing inhibition by the Glc-6-P analog, 1,5-
anhydroglucitol-6-P. Activity was measured with 0.03 mM l,5-
anhydroglucitol-6-P, increasing concentrations of P3 or the
indicated nucleotide, and with other conditions as given in
Methods. In the absence of inhibitor, the activity was
approximately 4.8/umole/min, and there was no activation by added

P1 or nucleotides.

181

ON

m 659“. ,

28. 605820800
0;
_

0.0

 

 

 

1111

i
'4')

ugw/elow’fl ‘KliAilOV

l
d.

 

 

182
concentrations, these compounds act as competitive inhibitors
(gs. ATP-Mg”) of the enzyme (36,37). P5 and its homologs,
pyrophosphate and tripolyphosphate, were similar in their
effectiveness at reversing inhibition by 1,5-anhydroglucitol-6-P
(Figure 6) . At high concentrations (K1 35 mM), P1 acts as a
competitive (gs. ATP-Mg“) inhibitor (10), as reflected by the
decreased activity seen at higher concentrations of this ligand
(Figure 6). Pyrophosphate and tripolyphosphate appear to be
similar in this regard, since decreased activity was also seen at

elevated levels of these ligands.

183

Figure 6. Effectiveness of P1.and its homologs at reversing
inhibition by the Glc-6-P analog, 1,5-anhydroglucitol-6-P.
Activity was determined with 0.1 mM 1,5-anhydrog1ucitol-6-P,
increasing concentrations of P3 or its homolog, and other
conditions as given in Methods. Pyrophosphate could not be used
at concentrations above 5 mM because of the limited solubility of

its magnesium salt.

184

on

o 2:9".

2E .cozobcoocoo

ON on . o

. p .

 

 

Bocdmocabodtu. <
oponamocdotxa n. ..
30:30.3 0

 

o/ n. ..m.n
c/e/T ,

 

 

mun

ugLu/sloun'l ‘KllAilOV

541‘

In

IT

M

ha
en
5-

C0.

DISCUSSION

There is now convincing evidence that the 100 kDa brain
hexokinase represents the evolutionary result of duplication and
fusion of a gene coding for an ancestral 50 kDa hexokinase
similar to the starfish enzyme in its properties (29). This may
be expected to result in a 100 kDa enzyme, each half of which
possesses a complete set of the ligand binding sites found on the
ancestral starfish-like hexokinase. Minimally, these include the
substrates, Glc and ATP-Mg”, as well as the inhibitors, Glc-6-P
and.P;. Thus, the 100 kDa brain enzyme would be predicted to
contain Egg sites for each of these ligands. This prediction may
seem to conflict with reports of a single binding site for Glc
(39, 40) located in the C-terminal half (26), a single high
affinity site for Glc-6-P (10, 39) located in the N-terminal half
(28), and a single high affinity site for P1L (10) of undefined
location. However, as we shall see, this is not the case.

Both starfish hexokinase and the isolated C-terminal half of
rat brain hexokinase, in which catalytic function resides, are
inhibited (competitively gs, ATP-Mg”) by relatively low (K; 5
mM) levels of Pi (29). The inhibition of the isolated C-terminal
half by P3 contrasts with the effect of this ligand on the intact
enzyme, for which it serves to antagonize the inhibition by Glc-
6-P (10, 12) and becomes inhibitory itself only at elevated
concentrations (10). As pointed out by Ellison 22.2i- (10), this

bimodal effect of Pg on the intact enzyme clearly indicates the

185

 

ac

.
.
S ..

re

6X

CO]

at!

boi

ha:

(25

COAT

thi

COD

mol

Prm

186

existence of Egg distinct binding sites for Pi, with quite
different affinity and function. With the range of P1
concentrations used in the binding experiments of Ellison g; 31.
(10), only the high affinity site was seen. The present study has
demonstrated that relatively low levels of P3 selectively protect
the N-terminal half of the brain hexokinase molecule, and
accordingly this region is deduced to contain the high affinity
site for P1..As previously demonstrated (28), the N-terminal
region also contains the high affinity allosteric site for Glc-6-
P, and thus a straightforward interpretation of the mutually
exclusive binding of P1 and Glc-6-P (10) would be that they
compete for a common anion binding site in the N-terminal region.

Results presented previously (25) and in the present work
are consistent with the existence of nucleotide binding sites in
both N- and C-terminal halves of the molecule. The C-terminal
half contains the site at which the substrate nucleotide binds
(25). This site exhibits a rather strict specificity for ATP, as
evidenced by the requirement for ATP as a phosphoryl donor (8),
the limdted effectiveness of other nucleoside triphosphates as
competitive inhibitors (36, 37), and greater destabilization of
this region by ATP-Mg++ compared with other nucleotides. In
contrast, binding to a site in the N-termdnal half of the
molecule is, at least in the absence of Mg”, quite nonspecific
and results in stabilization against denaturation and
proteolysis. This protective effect is diminished, and some

selectivity for ATP revealed, when the Mg" chelates are compared;

T3

()

0;

¢ 1.

m
t‘:
0"
in
he

5
54

51:
61

Ca:

187

the latter may reflect a basic similarity between the N- and C-
temminal sites, as would be expected from the gene duplication-
fusion schema. The absence of strict specificity in binding of
nucleotides to this N-terminal site obviously focuses attention
on the polyphosphate sidechain, common to all of these compounds,
as a major determinant in the binding and resultant stabilization
of the N-terminal region and reversal of inhibition by 1,5-
anhydroglucitol-G-P. This is reinforced by the demonstration
that binding of other polyanions such as P3,;pyrophosphate, and
tripolyphosphate also produce these same effects. A
straightforward interpretation of these observations would be
that all of these ligands interact at a common anion binding
site, for which Glc-6-P and P, also compete as discussed above.

The proposal of a second nucleotide binding site on brain
hexokinase is not unprecedented. In fact, the existence of a
nucleotide binding site, discrete from the site for binding of
the substrate nucleotide and adjacent to the allosteric site for
Glc-6-P, had previously been suggested based on analysis of the
inhibition patterns seen with a multisubstrate analog for
hexokinase (23). Fromm and his coworkers (36, 37) also postulated
the existence of an "allosteric" nucleotide binding site based on
kinetic results that they felt were incompatible with a single
substrate nucleotide binding site. However, this putative
allosteric site was said to exclude ATP as a ligand and hence
cannot be equated with the N-terminal site whose existence was

deduced from results of the present study, to which all

nucfi

IBS‘

sit.

see

int
anci
(pa;
the

inte

the i
the f

calcri

188
nucleoside triphosphates examined, including ATP, bind with
resulting protection against denaturation and proteolysis.

Since the present work indicates that the nucleotide binding
site in the N-terminal region is relatively nonspecific, it may
seem surprising that it was not detectably labeled by the
photoactivatable ATP analog, 8-azido-ATP, successfully used to
locate the substrate nucleotide binding site in the C-terminal
domain (25). We suggest that it was because of the lack of
specificity that this occurred. Since binding is dominated by
interactions gig the polyanionic sidechain, as discussed above,
and virtually independent of the nature of the base moiety
(particularly in the absence of Mg”, which was the situation in
the previous study (25)), it seems unlikely that intimate
interactions occur between the base and the protein. Yet it is
precisely such interactions that would be required to permit
covalent coupling with the photoactivated 8-azido group.

Intact brain hexokinase contains a single high affinity site
for Glc (39, 40) and for Glc-6-P (10, 39), located in the C-
terminal (26) and N-terminal (28) halves, respectively. However,
the isolated N-terminal and C-terminal halves have been shown to
contain binding sites for both Glc and Glc-6-P (29), and it was
concluded that a Glc binding site in the N-terminal half as well
as a Glc-6-P binding site in the C-terminal half are latent in
the intact enzyme, presumably the result of interactions between
the fused N- and C-terminal domains. Recent differential scanning

calorimetric studies (41) have provided further support for this

In

..I

{1‘

{f

189

view, with glucose shown to stabilize the N-terminal domain
against thermal denaturation. Moreover, the specificity of the N-
terminal hexose binding site, revealed in the isolated N-terminal
fragment, is similar to that of the C-terminal site at which the
substate hexose is bound, and the specificity of the C-terminal
hexose 6-phosphate binding site, revealed in the isolated C-
terminal half, is similar to that seen for binding of these
ligands to the allosteric site in the N-terminal region of the
intact enzyme (29). These results again suggest a basic
similarity between the N- and C-terminal sites for these
ligands, consistent with their generation from a common precursor
by gene duplication and fusion.

We propose a model for brain hexokinase (Figure 7) that
integrates the proposed evolutionary origin of this enzyme with

the analysis of ligand binding presented above.

 

190

Figure 7. Schematic representation of ligand binding sites on
rat brain hexokinase. Duplication and fusion of a gene coding for
a 50 kDa starfish-like enzyme gave rise to a 100 kDa enzyme
depicted schematically at the top center of the figure; following
the usual convention, the N-terminal region is depicted at the
left of the duplicated structure. The C-terminal domain contains
overt sites to which the substrates, Glc and ATP, are bound (25-
27). As previously proposed (11), occupation of the C-terminal
ATP site precludes binding of Glc-6-P at an allosteric site in
the N-terminal domain (28); conversely, binding of Glc-6-P at the
site in the N-terminal domain induces a conformational change
resulting in loss of the C-terminal nucleotide binding site
(lower right in figure). This antagonism between the binding of
substrate ATP and Glc-6-P is consistent with the competitive
nature of the inhibition by Glc-6-P (36, 37) as well as binding
studies (11, 40) indicating these to be mutually exclusive
ligands. The N-tenminal domain also contains a nucleotide binding
site (lower left in figure), much less specific than its sibling

in the C-terminal domain. The y-phosphate group (indicated in
bold lettering) of the nucleotide interacts with an anion binding
site for which P, (and other anionic ligands such as
pyrophosphate or tripolyphosphate) and Glc-6-P may also compete,
resulting in mutually exclusive binding of these ligands (10) and
antagonism.of inhibition by Glc-6-P by P, (10, 12) and its
homologs as well as by nucleoside triphosphates (present work).
The location of a nucleotide binding site in close proximity to a
site binding the hexose moiety of Glc-6-P had previously been
suggested based on kinetic studies with a multisubstrate analog
for hexokinase (23). Interaction between the domains in the
intact enzyme results in masking of a Glc-6-P binding site in the
C-terminal domain and a Glc binding site in the N-terminal domain
(crosshatched areas); isolation of the discrete N- and C-terminal
halves of the molecule removes this constraint, and the latent
binding sites are revealed (29). Removal of the N-terminal region
also facilitates access of Pg'to the anionic binding site with

which the y-phosphate of the substrate ATP normally interacts,
resulting in relatively potent inhibition (competitive gg. ATP-
Mg”, with K;* 5 mM) of the catalytically active C-terminal
fragment by P, (29). In the intact enzyme, access to this site is
limited, and inhibition (competitive gg, ATP-Mg", withIK,= 35
mM) is seen only at elevated levels of P, (10). Conformational
differences are suggested by altered profiles of the domains, but
obviously these are not meant to be taken as accurate reflections
of the actual structures. Conformational differences resulting
from binding of chelated and nonchelated forms of nucleoside
triphosphates are also neglected; although it is apparent from
the present work as well as previous studies (11) that these are
appreciable, they are not central to the concepts meant to be
depicted in this representation.

191

 

 

 

 

 

 

 

 

Figure 7

N)
O

\J

11

12

14.

15.

10.

11.

12.

13.

14.

15.

16.

REFERENCES

Sokoloff, L., Fitzgerald, G.G., and Kaufman, E.E. (1977) Ea
Nutrition and the Brain (Wurtman, R.J., and Wurtman, J.J.,
Eds.), Vol. 1, pp. 87-139, Raven Press, New York.

Lowry, O.H., Passonneau, J.V., Hassleberger, F.X. and
Schulz, D.W. (1964) g. Biol. Chem. 239, 18-30.

 

Sacktor, B., Wilson, J.E., and Tiekert, C.G. (1966) g. Biol.
Chem. 241, 5071-5075.

Wilson, J.E. (1984) 33 Regulation of Carbohydrate Metabolism
(Beitner, R., Ed.), Vol. I, pp. 45-85, CRC Press, Boca
Raton, FL.

Weil-Malherbe, H., and Bone, A.D. (1951) Biochemm g. 52,
339-347.

Crane, R.K., and Sols, A. (1954) g. Biol. Chem. 210, 597-
606.

 

Monod, J., Changeux, J.-P., and Jacob, F. (1963) g. Mol.
Biol. g, 306-329.

Grossbard, L., and Schimke, R.T. (1966) g. Biol. Chem. 241,
3546- 3560.

 

Copley, M., and Fromm, H.J. (1967) Biochemistry g, 3503-
3509.

Ellison, W. R., Lueck, J.D., and Fromm, H.J. (1975) g. Biol.
Chem. 250, 1864-1871.

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

Tiedemann, H., and Born, J. (1959) g, Naturforsch. 14b, 477-
478.

 

Fromm, H.J. (1981) £3 The Regulation of Carbohydrate
Formation and Utilization in Mammals (Veneziale, C.M., Ed.),
pp. 45-68, University Park Press, Baltimore.

Rose, I.A., Warms, J.V.B., and O'Connell, E.L. (1964)
Biochem. Biophys. Res. Commun. gg, 33-37.

Uyeda, K., and Racker, E. (1965) g. Biol. Chem. 240, 4689-
4693.

Rijksen, G., and Staal, G.E.J. (1985) i3 Regulation of
Carbohydrate Metabolism (Beitner, R., Ed.), Vol. I, pp.87-

192

18.

19.

21.

22.

23.

24.

26.

28.

29.

30.

32.

33.

34.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.
32.

33.

34.

193
103, CRC Press, Boca Raton, FL.

Colowick, S.P. (1973) $2 The Enzymes (Boyer, P.D., Ed.), 3rd
Ed., Vol. 9, pp. 1-48, Academic Press, New York.

Easterby, J. S., and O'Brien, M.J. (1973) Eur. 8. Biochem.
88, 201-211.

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

 

Holroyde, M.J., and Trayer, I.P. (1976) FEBS Lett. 88, 213-
219.

 

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

Gregoriou, M., Trayer, I.P., and Cornish-Bowden, A. (1983)
Eur. 8. Biochem. 134, 283-288.

Manning, T.A., and Wilson, J.E. (1984) Biochem, Biophys.
Res. Commun. 118, 90-96.

 

 

Schwab, D.A., and Wilson, J.E. (1989) Proc. Natl. Acad. Sci.
USA, 88, 2563-2567.

 

 

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.

 

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

 

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., in press.

 

Mochizuki, Y., and Hori, S.H. (1977) 8. Biochem. (Tokyo) 88,
1849-1855.

Mochizuki, Y. (1981) Comp. Biochem. Physiol. 703, 745-751.

 

 

Wilson, J.E. (1989) Prep. Biochem. 88, 13-21.

Ferrari, R.A., Mandelstam, P., and Crane, R.K. (1959) Arch.
Biochem. Biophys. 88, 372-377.

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

 

35.

36.

37.

38.

39.

40.

41.

194
Worthington, C.C., Ed. (1988) Worthington Manual, p. 320,
Worthington Biochemical Corp., Freehold, NJ.

Ning, J. Purich, D.L, and Fromm, H.J. (1969) 8. Biol. Chem.
244, 3840-3846.

 

Purich, D.L, and Fromm, H.J. (1971) 8. Biol. Chem. 246,
3456-3463.

 

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

Chou, A.C., and Wilson, J.E. (1974) Arch. Biochem. Biophys.
165, 628-633.

Ellison, W.R., Lueck, J.D., and Fromm, H.J. (1974) Biochem.
Biophys. Res. Commun. 81, 1214-1220.

White, T.K, and Wilson, J.E., submitted for publication.

‘The

to ti
the j
the k
in ti
while
mode:

unpuk

ZReve
abili
comp)

*0

M9.
inhih
Cont:

natuz

FOOTNOTES

1The ability of ATP-Mg” to produce distinct effects upon binding
to the N-terminal domain has been confirmed in experiments with
the isolated N-terminal half (29) of the molecule. Specifically,
the binding of several monoclonal antibodies recognizing epitopes
in this region was markedly enhanced in the presence of ATP-Mg“,
while chelated forms of other nucleoside triphosphates had only
moderate effects on antibody binding (A.D. Smith and J.E. Wilson,

unpublished results).

2Reversal of inhibition by ATP-Mg” was not studied since the
ability of this complex to serve as substrate obviously would
complicate interpretation of effects on rate. Assuming that ATP-
Mg”, like chelates of other nucleotides, can reverse the
inhibition, it seems possible that this interaction may
contribute to the observed (36, 37) competitive (g8. ATP-Mg”)

nature of the inhibition by Glc-6-P.

195

hexoki
rat br
mass 0
allost
in ref
0f onl
levels
Prompt
enzYme
a“C9811:
alJ‘OSti
"hat wé
demons

the u-

Chapter 6

A Differential Scanning Calorimetric Study of
Rat Brain Hexokinase:

Domain Structure and Stability

Like other mammalian hexokinases, the type I isozyme of
hexokinase (ATP: D-hexose 6-phosphotransferase; EC 2.7.1.1) from
rat brain consists of a single polypeptide chain with a molecular
mass of approximately 100 kDa, and is quite sensitive to
allosteric inhibition by the reaction product, Glc-6-P (reviewed
in ref. 1). In contrast, yeast hexokinase has a molecular mass
of only 50 kDa, and is not inhibited by physiologically relevant
levels of Glc-6-P (reviewed in ref. 2). These observations
prompted several investigators to propose that the mammalian
enzymes evolved by duplication and fusion of a gene coding for an
ancestral hexokinase similar to the yeast enzyme, with the
allosteric regulatory site of the mammalian enzyme evolving from
what was once a duplicated catalytic site (2-7). The
demonstration of extensive internal sequence similarity between

the N- and C-terminal halves of the Type I isozyme of mammalian

196

hexoi
10).

asso<
funct

molec

like

sensi
to in;
fUSlOl
This I
half c
activj
the N-
Proced
shown

halVes
PhCSPhT
measur,
for GM
it was
intact

i

 

197
hexokinase (8, 9), and between these and yeast hexokinase (8,
10), is consistent with such an evolutionary scheme, as is the
association of catalytic and regulatory (Glc-6-P binding)
functions with C- (ll-13) and N-terminal (14) halves of the
molecule, respectively.

Recent work (15) has led to a modification of the above
proposal, with the more proximate ancestor of mammalian
hexokinases suggested to be not a yeast-type hexokinase but more
like the hexokinase of starfish, with M., 50,000 Q8 marked
sensitivity to inhibition by Glc-6-P (16, 17), i.e., sensitivity
to inhibition by Glc-6-P evolved before the gene duplication and
fusion event giving rise to the 100 kDa mammalian hexokinases.
This modification was based on the isolation of the C-terminal
half of rat brain hexokinase with retention of full catalytic
activity 338 sensitivity to inhibition by Glc-6-P. Conversely,
the N-terminal half of the molecule was isolated by similar
procedures and, though devoid of detectable catalytic activity,
shown to bind both hexoses and hexose 6-phosphates. Thus, 8288
halves of the molecule possess binding sites for both
phosphorylated and unphosphorylated hexoses; since direct
:measurements of ligand binding have indicated one binding site
for Glc (18) and one for Glc-6-P (18, 19) per 100 kDa moleculefi
it was concluded that one site for each ligand was latent in the
intact enzyme, being revealed only when the two halves were
separated by proteolytic cleavage.

The proposed evolutionary origin of mammalian hexokinase, as

st
i.‘

:00

198
well as the apparent division of functional properties between
the two halves of the intact enzyme (ll-14), suggest that the 100
kDa enzyme is comprised of at least two structural domains.
Schwab and Wilson (9) have proposed such a structure based on the
amino acid sequence comparisons mentioned above and the structure
of the yeast enzyme, as determined by the x-ray crystallographic
studies of Steitz and coworkers (21-24).

Differential scanning calorimetry (DSC) can be an important
tool in the study of protein structure in that it introduces the
ability to recognize individual domains as they undergo thermal
denaturation, and the effect of ligands on the thermal
denaturation of domains has the potential to provide useful
information concerning domain interaction and function (25-27).
We initiated DSC experiments with the hope of generating
additional information about domain structure and function in rat
brain hexokinase. The results confirm that the enzyme can be
considered to be composed of quasi-discrete N- and C-terminal
halves, each of which thermally denatures as a single cooperative
unit, with (in the absence of added ligands) the N-terminal half

being inherently more stable.

MATERIALS AND METHODS

Materials. Rat brain hexokinase was purified by affinity
chromatography on Affi-Gel Blue (Bio-Rad Laboratories, Richmond,
CA) as previously described (28). Thioglycerol, PMSF, TPCK-
treated trypsin and other biochemicals were from Sigma Chemical

Company, (St. Louis, MO).

Determination of hexokinase activity and protein
concentration. Hexokinase activity was measured
spectrophotometrically by coupling to NADPH formation gig the
Glc-6-P dehydrogenase reaction (28). Hexokinase concentrations
were determined from the absorbance at 280 nm using a molar
extinction coefficient of 5.1 x 10‘ M“ cm‘1 (29) . Trypsin
concentrations were also determined from Am based on an

absorbance of 1.43 for a 1 mg/ml solution (30).

DSC analysis. Protein was prepared for DSC analysis by
exhaustive dialysis at 4°.against the buffer of choice. For the
experiments described here, two buffer systems were employed.
The first was comprised of 50 mM TrisCl, 0.5 mM EDTA, 10 mM
monothioglycerol, pH 8.5, and is referred to as high ionic
strength TET buffer. The second was identical to the first,
except that the concentration of Tris, was 5 mM (low ionic

Strength TET buffer).

199

under
solut

any P

and i

J

in thé

ultra:
The 1!
incorg
Instru

for at

 

200

Samples and reference buffers were deaerated with stirring
under vacuum for 5-10 minutes. Following deaeration, protein
solutions were centrifuged for 10 minutes at 27,000 x g to remove
any particulate matter. The supernatants were carefully removed
and protein concentration determined from Am prior to loading
in the calorimeter.

DSC experiments were conducted in a Microcal MC-2
ultrasensitive scanning calorimeter (Microcal Inc., Amherst, MA).
The instrument was modified to increase sensitivity by the
incorporation of a Model 150B microvolt ammeter (Keithley
Instruments, Cleveland, OH), and was interfaced with an IBM XT
for automatic data collection and storage. Scans were performed
at a rate of 903flhr. Since scan rate can affect DSC behavior
with some proteins (e.g., 31), preliminary experiments were done
which established that there was no effect of scan rate in the
range of 15-909/‘hr on the results obtained.

Data were analyzed using the DA-2 data analysis software
provided by Microcal. Deconvolution analysis was performed using
the Deconv subroutine, with the assumption that data could be
represented by independent two-state transitions with AHmmmm
- AH"...t l,m; excellent fit to the experimental data was obtained
using this analysis option, and the fit was not improved using

other analysis options available in the Deconv subroutine.

(‘V

'1“)

201

Proteolysis of thermally denatured hexokinase. Partially or

 

completely thermally denatured samples of hexokinase were
subjected to proteolytic digestion as an assay for structural
perturbation; domains that were irreversibly denatured during
calorimetric scans could be expected to demonstrate an increased
susceptibility to proteolysis. Samples that had been heated in
the MC-2 to the indicated temperatures were cooled in the cell to
approximately room temperature, removed from.the cell, and
subsequently kept on ice until proteolysis could be performed.
Proteolysis was done at room temperature by the addition of a
stock trypsin solution (2 mg/ml in 0.001 N HCl) to give a final
ratio of 0.2 mg trypsin per mg hexokinase. Digestion was allowed
to proceed for 20 minutes, at which time the sample was made 1 mM
in PMSF by the addition of an appropriate aliquot of stock
solution (200 mM in 95% ethanol). Within 3-4 minutes of the
addition of PMSF, an SDS-containing denaturing solution was added
to give a final concentration of 0.063 M TrisCl, pH 6.8, 2 %
(w/v) SDS, 4 % (v/v) glycerol, 4.5 % (v/v) 2-mercaptoethanol, and
0.0025 % bromphenol blue; the samples were immediately heated at
100° for 2 mdn prior to analysis by SDS gel electrophoresis and
immunoblotting procedures, as described previously (14).
Interpretation of these experiments was facilitated by
previous work from this laboratory (14, 15, 32, 33), with
pertinent results being summarized in Figure 1 for convenient
reference. Briefly, under nondenaturing conditions (32), trypsin

Cleaves hexokinase primarily at two sites, designated T1 and T,,

202
to give major species with molecular masses of 10, 40, and 50
kDa, and partial cleavage intermediates of 90 kDa (cleavage at T1
only) and 60 kDa (cleavage at I; only). When the enzyme is
partially denatured with guanidine hydrochloride (14, 15),
extensive proteolysis occurs but metastable species, with
molecular masses of 52 kDa and 48 kDa are observed. These
correspond to the N- and C-terminal halves, respectively, and
arise by cleavage at another site, T3; further cleavage of the 52
kDa fragment at.Tg gives rise to the 10 kDa N-terminal fragment
and a 42 kDa species. Identification of these proteolytic
fragments on immunoblots is made possible by the availability of
several monoclonal antibodies recognizing epitopes of defined
location within the molecule (33). The disposition of the various
proteolytic fragments within the overall sequence of the enzyme,
and the location of the epitopes recognized by monoclonal
antibodies 5A, 4D4, 3A2, and 21, are represented schematically in

Figure 1.

203

Figure 1. Location of tryptic cleavage sites and epitopes for
monoclonal antibodies. The overall sequence of rat brain
hexokinase is represented by the heavy line. T” 1;, and T3
represent sites at which tryptic cleavage occurs to give
fragments with the indicated molecular masses (in kDa). The bars
at the top of the figure delineate regions known to contain
epitopes recognized by monoclonal antibodies designated 4D4, 3A2,
21, and 5A; fragments including these regions are recognized by

their reactivity on immunoblots.

204

F 88E

v.3

 

q.

v_ow

«.—

v_om

gum

 

_—

J

20-

 

Tlllliu

Pa

25 i.
2:.

rerunr
|

of hex
analyz
has be
irreve
justif
inve s t
therma
repres

and en

In

RESULTS

DSC of rat brain hexokinase in high ionic strength buffer.
A representative thermogram for hexokinase in high ionic strength
TET buffer is shown in Figure 2 (Curve A). Thermal denaturation
under the described conditions was irreversible, and resulted in
precipitation of the protein between 85° and 95° C. Cooling and
rerunning the sample gave no discernible transition.

Despite the irreversible nature of the thermal denaturation
of hexokinase under these conditions, it was still possible to
analyze the results in terms of equilibrium thermodynamics, as
has been found to be the case with other proteins exhibiting
irreversible denaturation in DSC experiments (27, 34-37); the
justification for such treatment has been discussed by previous
investigators (34,36). Deconvolution analysis revealed that
thermal denaturation of rat brain hexokinase could be
represented by two independent transitions (Figure 3), with T33
and enthalpies given in Table I.

Similar results were obtained when Glc was included in the
buffer (Figure 2, Curve B), although the endotherm was shifted to
appreciably higher temperatures. Deconvolution again gave an
excellent fit with two independent two-state transitions, with
1L3 and enthalpies given in Table I. Both component transitions
were affected to a similar extent. In contrast, Glc-6-P had a

more marked effect on the second transition, resulting in a

205

206

Figure 2. Representative DSC scans of rat brain hexokinase. A-
C, high ionic strength TET buffer; D, low ionic strength TET
buffer. A, no ligand added; hexokinase, 0.64 mg/ml. B, 10 mM Glc;
hexokinase, 0.57 mg/ml. C, 1 mM Glc-6-P; hexokinase, 0.47 mg/ml.
D, no ligand added; hexokinase, 0.52 mg/ml. Distance between tick

marks on the y-axis represents 1 mcal/degree.

mca1/°K

207

 

 

 

 

 

 

W
i

l L 1 L

15 30 45 60 75

Temperature (0C)

Figure 2

 

 

 

208

Figure 3. Deconvolution analysis of DSC results obtained in
high ionic strength TET buffer. The scan shown in Curve A of
Figure 2 was subjected to deconvolution using the Microcal DA-2
program, as described in Methods. The solid curve shows the
actual experimental data. The dashed curve, labeled A + B,
represents the sum of the two transitions (A and B) obtained by
deconvolution. Distance betwen tick marks on the y-axis

represents 15 Kcal/ mole/ degree K.

Excess Heat Capacity

209

 

 

 

 

 

Temperature

(° C)

Figure 3

 

210

.mo:H¢> headsuco Havoc
wound momosacouwm :« puudoflpcw mucoawwonxo «0 Moses: now cm H omduo><a

.xe g .:0waduu:oocoo mimIOHc “28 ca .mcoHaduucuocoo o<ono use 0H0.

mu M «Na a.o M o.~a m I
«L + Hod m.o + o.mv a any ca + sum oazoflo

a m «as s.o H o.uo a i
mm + an“ s.o H ~.Ha H Ame so + new miououo

«H m sea e.o m s.em a I
s + ova o.c + «.mm a .ev as + new amoosuo

m m as” v.“ m a.om a i
on + so“ o.“ + m.sv H Ave an + sow ocoz
:Ihaqa aqua: ... um. ..H 33.3 game—la? :jd .333
.saaarucw asnausucm Hosea euev<

.m.m an .uouasm awe raucoupm oucoH emu: ca
OMGCHu—OXGZ Cddhm 06¢ ho “Oavdhflvdflva 03H .HOM “unfin— OMQHOEMQOHGU «H GHQ“?

211
larger difference in the T;s (Table I) and consequent improved
resolution of the two transitions in the thermogram (Figure 2,
Curve C). The substantial increase in stability of the enzyme in
the presence of Glc or Glc-6-P is consistent with previous
studies (38) demonstrating that binding of these ligands induces
a substantial conformational change in the enzyme with resulting
protection against a variety of inactivating agents, including
heat denaturation. This contrasts with the effects of GlcNAc,
which binds to the enzyme as evidenced by its action as a
competitive (gg. Glc) inhibitor (39), but which does not induce
conformational changes similar to those seen with Glc or Glc-6-P
(38). Addition of GlcNAc to the buffer had no observable effect
on the DSC profile of brain hexokinase (Table I).

It is evident that the standard deviations for the
enthalpies given in Table I are higher than might normally be
expected in DSC studies. Similar variability has been seen by
other investigators (e.g., 34, 40, 41) and attributed to
difficultly in accurately defining appropriate baselines for data
analysis or to subtle variation in different protein
preparations. Whatever the cause(s) of this problem, we do not
believe it undermines the essential conclusions of this study,
which, as will be seen, do not rest on a strictly quantitative
interpretation of the results. Nonetheless, it may be noted here
that the increased enthalpy seen for denaturation in the presence
of Glc is significant at the P < 0.01 level (42), indicating that

there is substantial enthalpy associated with the binding of Glc

212
to brain hexokinase; this is of interest because the previous DSC
study by Takahashi g8 gi. (43) led to the conclusion that binding
of Glc to yeast hexokinase was primarily entropy-driven and did
not involve a substantial enthalpy term. Thus, despite the
extensive similarity in sequence and, presumably, structure of
these two enzymes (9), there may be appreciable differences in

their interaction with this substrate.

DSC of rat brain hexokinase in low ionic strength buffer.
Thermal denaturation of rat brain hexokinase was also
investigated using low ionic strength TET buffer. This was
prompted by the results of Takahashi g8. gi. (43), which
indicated a marked effect of ionic strength on the DSC behavior
of yeast hexokinase“. Specifically, these workers found that at
high ionic strength, this enzyme displayed a single transition
while, at low ionic strength, the DSC thermogram.could be
resolved into two independent two-state transitions.
Qualitatively similar results have been obtained with the brain
enzyme, as the apparent single transition (actually resolvable
into two overlapping but independent transitions after
deconvolution -see above) seen at high ionic strength (Figure 2,
Curve A) is clearly split into two distinguishable transitions at
low ionic strength (Figure 2, Curve D). Sumilar results were seen
with low ionic strength buffer in the presence of Glc (Figure 4,
Curve A) or Glc-6-P (not shown), with these ligands again causing

a substantial increase in the T;s (Table II).

213

Figure 4. DSC of rat brain hexokinase in 10 mM Glc and low
ionic strength TET buffer. Hexokinase, 0.51 mg/ml. A, full scan
to 90° C; B, scan stopped at 54° C; C, rescan of sample in B after

cooling. Distance between tick marks on y-axis represents 1

meal/degree.

mcal/°K

214

 

 

 

l l l

(—

 

15

30 45 60 75
Temperature (°C)

kane4

 

215

Table II. T;s for Transitions in Low Ionic Strength Buffer

 

 

Added Ligand‘ 2° Transition
None 4 1
2
Glucose 3 1
2
G1c-6-P 3 1
2
ATP-Mg”, 2 1
plus GlcNac 2

'Glc and GlcNAc, 10 mM; Glc-6-P, 1 mM; ATP-Mg",

bNumber of experiments.

°Average SD.

49.
55.

54.
63.

52.
63.

47
54

Ht»

.8,
.4,

00 00
(”ab L04!)

0
(DU?

48.0
55.0

5 mM.

216

For reasons that remain obscure, we experienced considerable
difficulties with erratic baselines in DSC experiments at low
ionic strength (e.g, see Figure 4, Curve A). Due to the degree of
this problem, we are not including the results of deconvolution
analysis and resulting enthalpies. It may be stated, however,
that in several experiments in which a baseline could be
established with a reasonable degree of confidence, the data
could be satisfactorily deconvoluted into two transitions with
enthalpies similar to those seen at high ionic strengths.
Proteolytic digestion of thermally denatured hexokinase. The
observation of two independent transitions within the overall
endotherm for brain hexokinase is consistent with the concept,
developed in the introduction to this paper, that this enzyme may
be comprised of discrete catalytic and regulatory domains. We
wished to associate the transitions seen in DSC analysis with
specific structural regions of the molecule, and to elucidate the
order in which these regions unfolded during thermal
denaturation. To accomplish this, we took advantage of the
irreversible nature of the thermal denaturation of rat brain
hexokinase. It seemed possible that, by proceeding only through
the first transition in the DSC profile, we might be able to
obtain a species in which the more labile domain was unfolded
while the more stable domain was not. Unfolded proteins are
typically more susceptible to proteolysis, as has been found to

be the case with brain hexokinase partially denatured with

 

217
guanidine hydrochloride (14,15), and hence the thermally
denatured domain could be expected to be readily, and
selectively, proteolyzed. This formed the basis for the stategy
used in these experiments.

Curve A in Figure 5 is the same thermogram shown as Curve C
in Figure 2, and indicates the partial resolution of the two
transitions seen in high ionic strength in the presence of Glc-6-
P, as noted above. Curve B in Figure 5 shows the results of a
partial scan in which the heating was stopped at a point just
beyond the T; for the first transition. When this sample was
cooled and rescanned through the entire temperature range,
(Figure 5, Curve C), only the second transition was seen, with a
magnitude and'n,comparable to that seen for this transition in a
sample that had not been preheated (e.g, Figure 5, Curve A). This
result indicates two things: first, the absence of a detectable
first transition in Curve C confirms the essentially complete and
irreversible denaturation of the corresponding domain during the
partial scan; second, the appearance of the second transition
with virtually the same intensity and.T;.as seen with the sample
that had not been preheated, indicates that the domain
corresponding to the second transition had largely retained its
folded structure.

The question now became: what regions of the molecule
correspond to the labile and more stable domains? Figure 6 (Lane
2) shows the results after tryptic digestion of an unheated

sample of hexokinase in high ionic strength buffer with Glc-6-P

Present.

218

Figure 5. DSC of rat brain hexokinase in 1 mM Glc-6-P and high
ionic strength TET buffer. Hexokinase, 0.47 mg/ml. A, full scan
to 90° C; B, scan stopped at 54° C; C, rescan of sample in B after

cooling. Distance between tick marks on y-axis represents 1

meal/degree.

mca1/°K

219

 

 

l

l

l

 

15

30

45

Temperature

60
(°C)

Figure 5

75

 

 

220

Figure 6. SDS acrylamide gel analysis of tryptic digestion
products of native and partially denatured rat brain hexokinase.
Lane 1, digestion under the conditions of Polakis and Wilson
(32), which generates the fragments having molecular weights
indicated at the left of the figure.

Lane 2, tryptic digestion in high ionic strength TET buffer
containing 1 mM Glc-6-P. Lane 3, same enzyme sample as that used
in Lane 2, but digested after partial DSC scan to 54°<3 (i.e.,
through the first transition), as in Figure 5, Curve B. See

Figure 1 and text for additional information.

 

221

er10.3’ 1 2

100-
90*

60-
50*

4o-

trypsin»

10*

 

Figure 6

222
Only limited digestion occurs under these conditions, principally
at the T3 site to generate the 90 kDa intermediate and the 10 kDa
N-terminal fragment (see Methods); also seen are minor amounts of
the 40, 50, and 60 kDa species, formed in much greater abundance
under the digestion conditions of Polakis and Wilson (32) (Lane
1). In contrast, digestion of a hexokinase sample that had been
heated through the first transition (i.e., as in Figure 5, Curve
B) gave rise to 42 kDa and 52 kDa species as major products.
Immunoblots (Figure 7) demonstrated that the latter fragments
originated from the N-terminal half of the molecule, i.e., they
were reactive with monoclonal antibodies recognizing epitopes in
this region, but not with antibody 5A which
recognizes an epitope in the C-terminal half of the molecule.
Thus, the 42 kDa and.52 kDa species appear to be identical with
the N-terminal fragments generated by tryptic digestion of the
enzyme after partial denaturation with guanidine hydrochloride
(14,15), and result from cleavage of the enzyme at or near the
site designated'np Except for a small amount of residual intact
100 kDa enzyme and 90 kDa fragment, there was a virtual absence
of any species which included the C-terminal half of the
molecule. Therefore, it is the C-temminal half of the molecule
that has been denatured by passage through the first DSC
transition and thereby become susceptible to virtually complete
proteolysis during the subsequent digestion with trypsin. The

retention of tertiary structure in the N-temminal half of the

223

Figure 7. Immunoblot analysis of products from tryptic
digestion of partially denatured rat brain hexokinase. In each
panel, Lane 1 is a tryptic digest prepared under the conditions
of Polakis and Wilson (32), generating the fragments with
molecular weights indicated at the left of the figure, and Lane 2
is a tryptic digest, prepared after heating hexokinase in high
ionic strength TET buffer containing 1 mM Glc-6-P through the
first transition, as in Figure 5, Curve B. The panels at the
right show duplicate samples after electroblotting and probing
the blots with monoclonal antibodies recognizing epitopes in the
50 kDa (antibody 21), 40 kDa (antibody 5A), or 10 kDa tryptic
fragments (antibodies 4D4 and 3A2) found in the digest prepared
according to Polakis and Wilson (32). See Figure 1 and text for

additional information.

1

224

 

“a."
10 K

 

 

12_

40 K

 

 

1

 

50K

 

Protein

 

trypsin-

Figure 7

225
partially denatured molecule, as evidenced by its continued
resistance to tryptic digestion, confirms the identification of
this as the region of the molecule that unfolds during the second
transition.

As noted above, in low ionic strength buffer, the two
transitions become partially resolved in either the absence
(Figure 1, Curve D) or presence (Figure 4, Curve A) of Glc. As
also shown in Figure 4, after a partial scan through the first ’7'
transition (Curve B) followed by cooling of the sample, a rescan

(Curve C) shows the virtual absence of the first transition while

 

the second transition occurs with a magnitude and.T; comparable
to that seen in the unheated sample (Curve A). Similar results
were seen in the absence of added ligand. When samples that had
been partially denatured by heating through the first transition,
either in the absence or presence of Glc, were subjected to
proteolysis followed by SDS gel electrophoresis and
immunoblotting, the results (not shown) were virtually identical
to those shown in Figs. 6 and 7. Thus, under three conditions in
which the resolution of the two transitions was adequate to
pemmit selective denaturation (by a partial scan) of the domain
corresponding to the first transition, it is the N-terminal
domain that exhibits the greater stability.

Binding of ATP-Mg’+ destabilizes 8988 the N- and C-terminal
halves of the brain hexokinase molecule, as evidenced by marked
increase in the rate of proteolysis in the presence of guanidine
hydrochloride (T.K. White and J.E. Wilson, unpublished results).

However, with further addition of GlcNAc, a ligand that binds

226
poorly if at all to the N-terminal half, selective stabilization
of the C-terminal half is obtained (15). Similar effects have now
been seen in DSC experiments, with a resulting reversal of the
order in which N- and C-terminal domains are thermally denatured.
Figure 8, Curve A, shows the thermogram obtained in low

ionic strength buffer and with ATP-Mg” and GlcNAc present. Curve
B represents a partial scan of a duplicate sample; after cooling,
the scan through the entire temperature range was done, with ‘
results shown as Curve C. As in previous experiments employing

this same strategy (see above), only the second transition is

 

.seen with the sample subjected to the partial scan. However, the
Inagnitude of this transition is considerably reduced compared to
that seen with the sample that had not been subjected to prescan
(Curve A), and undoubtedly reflects a significant amount of
denaturation of the corresponding domain during the prescan. On
'the other hand, the T; for this transition is virtually identical
to that seen for the second transition with the untreated sample
(Curve A), indicating that the residual domain corresponding to
‘the second transition has retained its folded structure. Analysis
by'SDS gel electrophoresis and immunoblotting (Figure 9)
demonstrates that the domain corresponding to the second
transition is the C-terminal half of the molecule, i.e., 40 kDa
and 48 kDa fragments (corresponding to cleavage at.15 and T“
respectively) reactive with antibody 5A are present, but no

fragments reactive with antibodies recognizing epitopes in the

227

Figure 8. DSC of rat brain hexokinase in low ionic strength TET
buffer containing 5 mM ATP-Mg“ and 10 mM GlcNAc. Hexokinase, 0.46
mg/ml. A, full scan to 9U’CL B, partial scan through the first
transition (to 48°C». C, rescan of sample in B, after cooling.

Distance between tick marks on y-axis represents 1 mcal/degree.

mca1/°K

228

 

 

 

i l L 1 J

 

15 25 35 45 55 55

Temperature (°C)

Figure 8

 

229

N-terminal half of the molecule are seen on the immunoblots. The
complete degradation of the 40 kDa and 48 kDa C-terminal species
after scanning through 8988 transitions (Figure 9, lane 4)
confirms the association of the C-terminal domain with the second
transition. The complete proteolysis of the N-terminal half of
the molecule even in the unheated sample (Figure 9, Lane 2)
demonstrates the lability resulting from the presence of ATP-
Mg“, mentioned above.

The T;s for the transitions shown in Figure 8, Curve A, are
48° and 55°, corresponding to denaturation of the N- and C-
‘terndnal domains, respectively. Comparison with the values in
Table I, 51° and 48° for the N- and C-terminal domains,
respectively, shows that the relative values of these T.,s are in
taccord with the view that the N-terminal domain is destabilized
and.the C-terminal domain stabilized in the presence of ATP-Mg“

plus GlcNAc.

 

230

Figure 9. SDS gel electrophoresis and immunoblot analysis of
products from tryptic digestion of partially denatured rat brain
hexokinase. The left panel is an SDS gel stained with Coomassie
Blue. The other panels are immunoblots that have been probed with
monoclonal antibodies having a defined location within the
sequence of rat brain hexokinase (as in Figure 7). In each panel:
Lane 1 is a tryptic digest prepared under the conditions of
Polakis and Wilson (32), generating the fragments with molecular
weights shown at the left of the figure; Lane 2, hexokinase
digested in low ionic strength TET buffer containing 5 mM ATP-
Mg“ and 10 mM GlcNAc; Lane 3, identical to hexokinase sample in
Lane 2, but heated through first transition (to 48° C, as in
Figure 9, Curve B) prior to digestion; Lane 4, identical to
hexokinase in Lane 2, but heated through both transitions (to 90°
C, as in Figure 9, Curve A) prior to digestion. No
immunoreactivity was seen with any of the antibodies when blots

of Lane 4 were probed (not shown).

231

m 2:9“.

v_Ow

 

 

 

 

 

 

 

 

 

 

 

 

 

10..

152;:

10¢
1cm
100

10¢
100-

 

 

£22..
..Iiil.‘ -
in“. tr
.. w
a
.v n N _.

anX 52

DISCUSSION

These results are consistent with the view that brain
hexokinase is composed of two distinct domains, corresponding to
the N- and C-terminal halves of the molecule, and that the N-
terminal domain is intrinsically more stable.

Interactions between domains can be expected to influence
their stability. Brandts gp_gi. (27) have recently provided a
treatment of interdomain interactions as they pertain to analysis
of DSC results, and shown that it may be possible to estimate the
strength of these interactions, expressed as an “interface term",

.AG”, which effectively represents the free energy for
interaction between the domains. AG” can be expected to be
substantial for brain hexokinase since previous studies have
indicated the existence of strong interactions between the N- and
C-terminal regions (9, 14, 32) . Evaluation of AG” from DSC
experiments depends on the ability to selectively shift the lower
transition to higher T;s by addition of a stabilizing ligand
binding to the corresponding domain, or alternatively, to shift
the T. for the second transition to lower values by addition of a
ligand that selectively destabilizes the corresponding domain,
i.e., to produce “crossover“ in the transitions. Neither of these
alternatives appear to be viable in the present situation for
lreaaons discussed below. Nonetheless, we believe the results
c>btained in this study do provide additional insight into the

1hterdomain interactions as they exist in brain hexokinase.

232

 

233

Glc and Glc-6-P cause an increase in.T; for 8988
transitions, and do not produce the selective stabilization
required by the first approach of Brandts gp,gi. (27). This is in
itself quite interesting and merits specific comment. As made
evident by the results presented above, the C-terminal domain is
intrinsically less stable and corresponds to the first transition
seen in the absence of ligands. Denaturation of this domain
obviously would disrupt interactions with the N-terminal domain
as they exist in the native structure. Or, as expressed by
Brandts giDgi, (27), AGA3 goes to zero. Hence, the second
transition is expected to reflect the intrinsic stability of the
N-temminal domain, devoid of any interactions with the C-terminal
domain. In the intact enzyme, Glc binds selectively to the C-
termdnal domain (12,13). Thus, denaturation of this domain (by
passing through the first transition) should preclude binding of
Glc as well as abolish interdomain interactions (i.e.,.AG” goes
to zero) and hence it is to be expected that the second
transition would pg; be affected by the presence of Glc. This is
not the case. Since an indirect effect of Glc on the N-terminal
domain (gig interdomain interactions with the previously
denatured C-terminal domain) can be excluded, these results
.indicate that, when interactions with the C-terminal domain have
Leer: abolished, Glc can bind directly to the N-terminal domain.
inhese results are therefore in accord with the previous
c=<Dnclusion (15) that the N-terminal domain contains a latent Glc
t>inding site, that becomes manifest only when the C-terminal

donain has been effectively removed, by proteolysis in the

 

234
previous work (15) or by thermal denaturation in the present
study.

The effect of Glc-6-P on both transitions can be explained
in a more straightforward manner since this ligand interacts with
the N-terminal domain (14). Hence the increased stability of the
C-terminal domain (increased T. for the first transition) in the
presence of Glc-6-P can be considered to be an indirect effect,
mediated by intense interdomain interactions with the N-terminal
domain, to which the ligand is actually bound.

"Crossover" did occur in the transitions observed in the
presence of ATP-Mg“ plus GlcNAc, but this resulted from an effect
on 2222 domains, and hence this too lacks the selective effect
required for the quantitative analysis of interdomain
interactions as proposed by Brandts gp_gi. (27).

Based on comparison of the results in Tables I and II, the
increased resolution of the transitions seen in low ionic
strength buffer appears to be principally due to an increase in
the T; for the second transition (N-terminal domain). While it is
obviously difficult to attribute any specific significance to
this result, it does suggest that this region of the molecule may

be most susceptible to effects of ionic strength.

 

ACKNOWLEDGEMENT

We grateful to Dr. John F. Brandts for his interest in this
work, and for providing us with a copy of a manuscript (Brandts,

J.F., Hu, C.Q., Lin, L.-N., and Mas, M.T. (1989) Biochemistry, in

 

press) prior to its publication.

235

 

10.

11.

12.

13.

14.

15.

16.

17.

18.

REFERENCES

 

Wilson, J.E. (1984) ig Regulation of Carbohydrate Metabolism
(Beitner, R., Ed.), PP. 45-85, CRC Press, Boca Raton, FL.

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

Easterby, J.S. and O'Brien, M.J. (1973) Eur. 8. Biochem. 88,
201-211.

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

Holroyde, M.J., and Trayer, I.P. (1976) FEBS Lett. 88, 213- H1_”fi
219. '

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

.4‘

 

Gregoriou, M., Trayer, I.P., and Cornish-Bowden A. (1983) 7
Eur. 8. Biochem. 134, 283-288. J

 

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

 

Schwab, D.A., and Wilson, J.E. (1989) Proc. Natl. Acad. Sci.
USA 88, 2563-2567.

 

Schwab, D.A., and Wilson, J.E. (1988) 8. Biol. Chem. 263,
3220-3224.

 

 

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.

 

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

 

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., in press.

 

Mochizuki, Y., and Hori, S.H. (1977) 8. Biochem. (Tokyo) 8i,
1849-1855.

Mochizuki, Y. (1981) Comp. Biochem. Physiol. 708, 745-751.

Chou, A.C., and Wilson, J.E. (1974) Arch. Biochem, Biophys.
165, 628-633.

236

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

237

Ellison, W.R., Lueck, J.D., and Fromm, H.J. (1975) 8. Biol.
Chem. 250, 1864-1871.

Lazo, P.A., Sols, A., and Wilson, J.E. (1980) 8. Biol. Chem.

 

255, 7548-7551.

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

Anderson, C.M., Stenkamp, R.E., and Steitz, T.A. (1978) 8.
Mol. Biol. 123, 15-33.

 

Anderson, C.M., Stenkamp, R.E., McDonald, R.C., and Steitz,
T.A. (1978) 8. Mol. Biol. 123, 207-219.

 

 

Bennett, W.S., Jr., and Steitz, T.A. (1980) 8. Mol. Biol.
140, 211-230.

Krishnan, K.S., and Brandts, J.P. (1978) ig Methods in
Enzymology (Hirs, C.H.W., and Timasheff, S.N., Eds.), Vol.
49, pp. 3-14, Academic Press, New York.

Donovan, J.W. (1984) Trends ig Biochemical Sciences 8, 340-
344.

Brandts, J.F., Hu, C.Q., Lin, L.-N., and Mas, M.T. (1989)
Biochemistry, in press.

Wilson, J.E. (1989) Prgp. Biochem. 18, 13-21.

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

Worthington, C.C., ed. (1988) Worthington Manual, p. 320,
Worthington Biochemical Corp., Freehold, N.J.

Roe, J.A., Butler, A., Scholler, D.M., Valentine, J.S.,
Marky, L., and Breslauer, K.J. (1988) Biochemistry 81, 950-
958.

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

Wilson, J.E., and Smith, A.D. (1985) 8. Biol. Chem. 260,
12838-12843.

 

 

Edge, V., Allewell, N.M., and Sturtevant, J.M. (1985)
Biochemistry 88, 5899-5906.

Edge, V., Allewell, N.M., and Sturtevant, J.M. (1988)
Biochemistry 81, 8081-8087.

Manly, S.P., Matthews, K.S., and Sturtevant, J.M. (1985)
Biochemistry 81, 3842-3846.

 

Eh

‘ CM-

.41 i

37.

38.

39.

40.

41.

42.

43.

238

Hu, C.Q., and Sturtevant, J.M. (1987) Biochemistry 88, 178-
182.

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

Tuttle, J.P., and Wilson, J.E. (1970) Biochim. Biophys. Acta
212, 185-188.

Ghosaini, L.R., Abraham, M.B., and Sturtevant, J.M. (1988)
Biochemdstry 81, 5257-5261.

Kresheck, G.C., and Erman, J.E. (1988) Biochemistry 81,
2490-2496.

Snedecor, G.W., and Cochran, W.G. (1980) Statistical
Methods, Seventh Edition, Iowa State University Press, Ames,
IA.

Takahashi, K., Casey, J.L., and Sturtevant, J.M. (1981)
Biochemistry 88,4693-4697.

FOOTNOTES

1Although there is evidence for a second, low-affinity site for
G1c—6-P (20), this is thought to represent binding at the site

normally used for binding of Glc.

2We have also conducted some DSC experiments with yeast
hexokinase obtained from Sigma Chemical Co., the same source
used by Takahashi gg_gi. (43). We have found the effect of Glc
to be highly dependent on the particular preparation used, with
some showing pg detectable effect of Glc on the DSC profile
even though the specific activity was comparable to
preparations that did show the stabilizing effect reported by
Takahashi EE,§$- It is thus apparent that the effect of Glc on
DSC behavior is pp; directly linked with the ability to use Glc

as a substrate.

239

Chapter 7

Summary and Perspectives

Based on the size of the enzyme and its complexity of
function, it seemed clear at the initiation of these experiments
that brain hexokinase was likely to be composed of at least two
structural/functional domains. Initial experiments by Polakis
and Wilson (1) had demonstrated two sites highly susceptible to
tryptic cleavage, termed T, and T,, which generated three
fragments of 10, 50, and 40 kDa from N- to C-termdnus,
respectively. Because it is frequently the case that polypeptide
regions linking domains are highly susceptible to proteolysis
(2), it was initially proposed that these fragments corresponded
to the structural domains of hexokinase. However, it is clear
from the data presented that this view is not consistent with the
current evidence, and, based upon these structural studies and
cDNA sequencing evidence (3,4), we have proposed that the enzyme
is composed of two domains of approximately the same size, one
being regulatory in function while the other is involved with

catalysis.

240

 

241

At this time, it appears that tryptic cleavage of the enzyme
in the polypeptide region between domains, i.e., at the site we
have designated as 1;, occurs to a significant extent only under
denaturing conditions. It is noteworthy that this observation is
consistent with the structure of the enzyme as proposed by Schwab
and Wilson (4), which is based upon comparison of amino acid
sequence in the rat brain and yeast enzymes. In this model, the
T; cleavage site is located at the interface between the two
strongly interacting domains, and it is clear that separation of
the domains via partial denaturation would be one way in which
accessibility to trypsin could be achieved (4). It is noteworthy
that this is not the case for all mammalian brain hexokinases,
i.e., not all hexokinases require the presence of denaturants to
generate tryptic cleavage resulting in two fragments
corresponding to N- and C-terminal halves of the enzyme. For
example, brain hexokinases from cat, dog, pig, and sheep all
demonstrate significant cleavage at a site similar to T, under
non-denaturing conditions (5), and therefore behave in a manner
more consistent with the general trend which indicates an
increased susceptibility to cleavage between domains (2).

In addition to defining the domain structure of rat brain
hexokinase, the present work has also allowed us to reexamdne the
proposed evolution of the mammalian hexokinases. It was an
unexpected result that both domains should contain a binding site
for the effector Glc-6-P, and one.that clearly called for a

reevaluation of the previous proposal that the present day

242
mammalian enzymes have evolved from a yeast-like ancestor (6-11).
Our proposal that the ancestral enzyme was a 50 kDa enzyme which
ggg sensitive to Glc-6-P inhibition is clearly consistent with
the existence of present day marine organisma which share these
properties. The question still remains as to the physiological
significance, if any, of sites for both substrates and effectors
remaining in both domains. Since the "extra" sites do not appear
to be functional in the intact enzyme, one must presume that they
serve some other purpose in the enzyme, e.g., perhaps the
residues that are involved are critical for maintaining
structural integrity in the molecule.

In general, the use of organic denaturants to probe the
domain structure of this enzyme was productive. Although both
urea and GuHCl induced rapid inactivation of the enzyme, GuHCl
was by far the more useful denaturant in these studies,
apparently through its ability to disrupt inter-domain
interactions - a property which allowed us to demonstrate
selective protection of catalytic or regulatory domains by
appropriate ligands. The ability of GuHCl to similarly affect
inter-domain interactions has been reported previously by at
least two laboratories (12,13) who also noted an increase in the
independent behavior of domains in this denaturant. This
property of GuHCl was also instrumental in the development of
methods for the isolation of the catalytic and regulatory domains
in a native state.

Differential scanning calorimetry was also a very useful

243
technique in the study of the domain structure of hexokinase.
Results obtained using this instrument clearly supported the
conclusions previously obtained using the organic denaturants,
and provided further information concerning the relative
stabilities of the two domains.

Although some important information was gained concerning
the structure of hexokinase, it is possible to suggest several
areas of experimentation which have the potential to yield
further information concerning structure and function in this
enzyme. For example, our initial inactivation experiments
succeeded in defining conditions whereby the domains could be
induced to act independently. However, in the course of gaining
this information, some interesting questions were raised which
this research did not address. For example, the nature of the
multiphasic character of inactivation kinetics or the nature of
the irreversible denaturation of the enzyme are two areas in
which future experimentation might yield some valuable
information.

Additionally, in their review on multifunctional proteins,
Kirschner and Bisswanger (14), suggest that the reassociation of
separated fragments is one approach to the study of interdomain
interactions. Although brain hexokinase is not a classical
multifunctional enzyme, this approach clearly has the potential
to yield further information concerning interaction of the
domains in the intact enzyme. It would be of interest to

attempt to define conditions whereby the 52 kDa and 48 kDa

244
fragments interact as observed in the intact enzyme. The nature
of inter-domain interactions could then be more definitively
examined.

The bulk of experiments which have been described in this
thesis are concerned with the binding of the effector, Glc-6-P,
and how its binding affects catalytic activity in the C-terminal
domain. Results suggest that one advantage gained by the
presence of an independent regulatory domain is the additional
level of regulation introduced by the ability of P,‘to antagonize
Glc-6-P binding, and thereby "activate" hexokinase. However,
experiments performed by Polakis and Wilson (1) have made it
clear that the N-terminal domain is also involved in the
reversible interaction of brain hexokinase with the outer
mitochondrial membrane - an interaction which is also governed by
levels of Glc-6-P and.P; in the cell (15). It is apparent that
some valuable information might be gained by studying the
interaction of the isolated N-terminal domain with the outer
mitochondrial membrane, especially in the presence and absence of
selected ligands.

Experiments that were performed using the differential
scanning calorimeter also suggest the follow-up experiments in
which thermal denaturation of the isolated domains is observed.
However, the current purification procedure for the domains does
not yield enough of either fragment to make these experiments
practical. If methods could be developed which led to improved

yield of fragments, further DSC experimentation using the

245
isolated domains would be of obvious interest.

Although the experiments which have been described in this
thesis have shed substantial light on the complex structure-
function relationships that exist in this enzyme, it is apparent
that there is still a great deal to learn, concerning both the
structure and regulation of this enzyme, and the way this complex

regulation has evolved from more simple ancestral molecules.

10.

ll.

12.

13.

14.

15.

REFERENCES

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

Schulz, G.E., and Schirmer, R.H. (1979) Principles pi
Protein Structure, p. 84, Springer Verlag, New York/Berlin.

 

Schwab, D.A., and Wilson, J.E. (1988) 8, Biol. Chem, 263,
3220-3224.

 

Schwab, D.A., and Wilson, J.E. (1989) Proc. Natl. Acad. Sci.
USA 88, 2563-2567.

 

 

Ureta, T., Medina, C. and Preller, A. (1987) Arch. Biol.
Med. Exper. 88, 343-357.

 

 

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

Easterby, J.S. and O'Brien, M.J. (1973) Eur. 8. Biochem. 88,
201-211.

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

Holroyde, M.J., and Trayer, I.P. (1976) FEBS Lett. 88, 213-
219.

 

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

 

Gregoriou, M., Trayer, I.P., and Cornish-Bowden, A. (1983)
Eur. 8. Biochem. 134, 283-288.

Miles, E.W., Katsuhide, Y., and Ogasahara, K. (1982)
Biochemistry 81, 2586-2592.

Zetina, C.R., and Goldberg, M.E. (1980) 8. Mol. Biol. 137,
401-414.

 

Kirschner, K. and Bisswanger, H. (1976) Ann. Rev. Biochem.
88, 143-166.

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

246

MICHIGAN STATE UNIV. LIBRARIES
ii i ii WI ii ii iii "ll" Iii Ii iii Iii Iii“ Iii W H" Ii i ii
31293007880440