I'n
)
'2

:7 *Q L .
. -. ' (.lr‘”. , 1'8}! "w
, .- . J 3 a). .o.:.
‘ N‘.'("'.'.v'" .) 9‘)
‘ ' ' ’5'. .4

v r

J p' ‘. hl| n
“H ’_..I.|.k.1u, )3.
. 1' '

[311;
3:1»;

Lit

. ‘

v
I‘-
\r
.,'- ‘ .,.IA .-
Ot'. Two. I“
g .:;»;f rt: -
I

,. A ,_ .
F .
' 1.93.3.“
‘ ‘ ‘- 'v u '
I "(P

o
y}?! r:- -

0‘-
w-

A‘
w -b>‘f

4:1
:1

\-

 

L
.4
't 5.

v.‘ r

"‘
. . - N.“
“‘:"'|‘A'o'y
; ‘. .
.d-w»
V .u.‘
. .-

., °'
-. fiz'f'P-‘tiaitti‘. '
‘f . “14:5: .,
x: ' "

,\

mg i L?

.,.:_' ‘
i.-- ‘ _

. ‘,‘.<.

 

ITY LIBRARIES

Wiiiiii‘j‘jiiigw“

 

Iliiii‘iiiiiiiiiiifif

; i i‘ W
(A 3 1293 01682

 

 

This is to certify that the

dissertation entitled

CHARACTERIZATION OF THE CATALYTIC MECHANISM
OF THIOLTRANSFERASE TOWARD THE NON-DISULFIDE
SUBSTRATES OF DEHYDROASCORBIC ACID AND ALLOXAN:
THE POTENTIAL TO PROTECT AND THE POTENTIAL TO

M
HAR presented by

MICHAEL P . WASHBURN

has been accepted towards fulfillment
of the requirements for

Ph. D. degreein Biochemistry

 

Major professor

 

Daté 1, /7%

MSU is an Affirmative Action/Equal Opportunity Institution 0-1277!

 

LIBRARY

MlchIgan State
University

 

 

 

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

 

MTE DUE DATE DUE MTE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/98 WMpGS-ou

 

CHARACTERIZATION OF THE CATALYTIC MECHANISM OF
THIOLTRANSFERASE TOWARD THE NON-DISULFIDE SUBSTRATES OF
DEHYDROASCORBIC ACID AND ALLOXAN: THE POTENTIAL TO PROTECT
AND THE POTENTIAL TO HARM

by

Michael P. Washbum

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Biochemistry and Institute for Environmental Toxicology

1998

ABSTRACT
CHARACTERIZATION OF THE CATALYTIC MECHANISM OF
THIOLTRANSFERASE TOWARD THE NON-DISULFIDE SUBSTRATES OF
DEHYDROASCORBIC ACID AND ALLOXAN: THE POTENTIAL TO PROTECT
AND THE POTENTIAL TO HARM
by
Michael P. Washbum

Unable to synthesize the antioxidant ascorbic acid (AA), humans, other primates, and
guinea pigs rely on dietary intake and AA regeneration from either its one or two electron
oxidation products semidehydroascorbate and dehydroascorbate (DHA) to maintain cellular
levels of this essential vitamin. A rapidly growing class of enzymes catalyze NADPH-
dependent or glutathione (GSH)-dependent DHA reduction. Based on the comparison of
their catalytic efficiencies, the GSH-dependent DHA reductases, recombinant pig liver
thioltransferase (RPLTT) and a 32kDa enzyme isolated from human erythrocytes, are the
most robust DHA reductases. Several DHA reductases including protein disulfide
isomerase, the thioredoxin/thioredoxin reductase system, and thioltransferase belong to a
class of enzymes known as thiolzdisulfide oxidoreductases, which contain -CXXC- at the
active site. The thiolzdisulfide exchange activities of these enzymes are well understood, but
little is known about their DHA reductase activities.

The catalytic mechanism of the GSH-dependent DHA reductase activity by
thioltransferase was characterized. Kinetic comparisons demonstrated RPLTT and the C258
mutant protein to have equivalent catalytic efficiencies, in excess of those of the R26V and

K27Q variants with DHA as a substrate. Iodoacetamide (IAM) inactivation studies

confirmed the role of cysteine in the reaction mechanism and demonstrated the ability of

DHA to bind to reduced enzyme and protect RPLTT and the R26V and K27Q variants but
not the C258 enzyme from IAM inactivation. As demonstrated by isoelectric focusing (IEF)
and electrochemical analysis, reduced RPLTT and the R26V and K27Q variants but not the
C258 enzyme could reduce DHA to AA in the absence of GSH and form oxidized enzyme.
The inability of the C258 enzyme to chemically reduce DHA to AA in the absence of GSH
suggested the possibility of an alternate route of thioltransferase catalysis where a postulated
thiohemiketal intermediate between DHA and the enzyme would be cleaved by GSH,
releasing AA and forming and enzyme/GSH mixed disulfide.

The diabetogenic agent alloxan has a similar structure to DHA, and RPLTT was
shown to be a GSH-dependent alloxan reductase. Based on catalytic efficiencies,
thioltransferase was a more robust alloxan reductase than DHA reductase. Essential to
alloxan’s toxicity, after the reduction of alloxan to dialuric acid, dialuric acid oxidizes back
to alloxan generating reactive oxygen species. RPLTT catalysis of the formation of dialuric
acid resulted in increases in oxygen consumption and production of superoxide and
hydrogen peroxide in the alloxan/GSH system, suggesting a role for thioltransferase in
alloxan’s toxicity. Like DHA, reduced RPLTT and the K27Q variant but not the C258
enzyme could chemically reduce alloxan forming oxidized enzyme, suggesting similar
catalytic mechanisms for the alloxan and DHA reductase activities of RPLTT. Interestingly,
IEF of C258 incubated with alloxan resulted in reduced enzyme and a more acidic species

which may be the formation of a thiohemiketal intermediate between alloxan and C258.

To the most important people of them all: my family

iv

ACKNOWLEDGMENTS

Who to thank, who to thank, who to thank ..... Well probably more people than ever
will get thanked, but here goes. Thanks to Dr. Tom Diets, Dr. Larry Fischer, Dr. Shelagh
Fergusson-Miller, Dr. Robert Hausinger, and Dr. Rawle Hollingsworth for their patience and
guidance with someone like myself. In addition, Dr. John Mieyal of the Department of
Pharmacology at Case Western Reserve University was like an additional committee
member. With his expertise in thioltransferase, Dr. Mieyal provided valuable insight
regarding my poster presentations at national meetings. Dr. John Wilson was also a great
assistance in getting me going on isoelectric focusing studies. Dr. Laurence Florens and Dr.
Claire Vielle provided much needed critical reviews of my writing and they demonstrated
that the French have a much better command of English than Americans...I guess it’s because
they live closer to England and have such a colorful history of interaction. Without the
support staff in this building we would never accomplish anything. As a result the Carols,
Vicki, Mary, Pappan, Tony, and Joyce, and all the other support staff deserve a great deal of
credit.

All of the past and present members of the Wells lab taught me a great deal about the
process of science with a group of people from nations all over the globe. Dian Peng, Chun-
Zhi, Guo Ping, Che-Hun, Leslie, Beta, Vicki, Lori, Rajashree, Van, Melissa, Helen, Katia,
Anita, Aaron, and Erin...what a crew. They were essential to my development as a scientist.
Many people in the department and out of the department greatly contributed to my well
being as person and my mental health. The list is to long to include in this thesis, and all of

these people know who they are.

I also must thank Dr. Larry Fishcer and the Institute for Environmental Toxicology
for accepting me into the program and placing me on the NIEHS training grant. Also, the
last year of my thesis would not have been possible without the generous financial support
of Dr. Hugh Riordan and the Center for the Improvement of Human Functioning
International, Inc.

Finally, my mentor. Dr. Wells is one of the nicest, most forgiving, and patient
persons I have ever met. In addition, Bill is an excellent, creative researcher. In spite of the
nature of science, Bill maintained his humor, spirit, and unrequited passion for research. I
learned much more from Bill than the process of asking questions, experimental design,
formulating hypotheses, and carrying out a project to completion. Bill brings his humanity
to science which is something science needs much more of. I can only hope that my career

can be as long and as productive as Bill’s, but I will always try to be a humane scientist.

vi

TABLE OF CONTENTS

LIST OF TABLES ...................................................... xi
LIST OF F IGURES .................................................... xii
LIST OF ABBREVIATIONS ............................................. xv
CHAPTER I
INTRODUCTION ....................................................... 1
I. Roles of ascorbic acid in biology. ................................... l
A. Role of ascorbic acid as a ferric iron reductant. .................. 2
B. Antioxidant functions of ascorbic acid. ........................ 3
l. Ascorbic acid protection against lipid peroxidation ......... 6
2. Antioxidant function of ascorbic acid in the plasma ......... 7
3. Role of ascorbic acid in immune system function. .......... 8
4. Antioxidant function of ascorbic acid in glutathione deficient
animals ....................................... 11
II. Ascorbic acid synthesis and regeneration. ........................... 12
A. Semidehydroascorbate reduction. ........................... 12
B. Dehydroascorbate reduction ................................ l3
1. NADPH-dependent dehydroascorbate reduction. .......... l3
2. GSH-dependent dehydroascorbate reduction. ............. l4
3. Chemical vs. enzymatic GSH-dependent dehydroascorbate
reduction ..................................... 14
III. Thiolzdisulfide oxidoreductases ................................... 16
A. DsbA, DsbB, and protein disulfide isomerase. .................. 16
B. Thioredoxin. ............................................ 17
C. Thioltransferase (glutaredoxin). ............................. l9
1. Purification, cloning, and sequences of thioltransferases. . . . . 22
IV. Structures of thiolzdisulfide oxidoreductases ......................... 22
A. 3D structure of thioredoxin ................................ 25
B. 3D structures of DsbA and protein disulfide isomerase ........... 25
C. 3D structures of thioltransferases. ........................... 27
l. Bacteriophage T4 thioltransferase. ..................... 27
2. E. coli thioltransferase. .............................. 27
3. Pig liver thioltransferase. ............................ 29

vii

4. GSH binding to thioltransferase ........................ 32
5. Solution structure of the mixed disulfide between E. coli

thioltransferase and glutathione. ................... 32
V. Function and mechanism of thioltransferase .......................... 34
A. Redox potentials and pKa of active site cysteines thioltransferase. . . 34
B. Protein deglutathionylation activity of thioltransferase. ......... 35
C. The catalytic mechanism of thioltransferase. ................... 37
D. The dehydroascorbate reductase activity of thioltransferase. ...... 41
B. Other potential non-disulfide substrates for thioltransferase. ....... 44
1. Alloxan toxicity to the 2-cells of the pancreas. ............ 47
VI Objectives of thesis. ............................................ 49
References ...................................................... 50
CHAPTER II
KINETIC CHARACTERIZATION AND COMPARISON OF MAMMALIAN
DEHYDROASCORBATE REDUCTASES ............................ 62
Abstract ........................................................ 63
Introduction ..................................................... 64
Materials and Methods ............................................. 65
Enzyme Purification ......................................... 66
Recombinant pig liver thioltransferase .................... 66
Bovine liver protein disulfide isomerase ................... 67
DHA reductase assay and kinetic characterization ................. 68
Results ......................................................... 69
DHA reductase activity of bovine liver PDI, RPLTT, and human
erythrocyte 32kDa DHA reductase ....................... 69
Discussion ...................................................... 69
Comparison of DHA reductase activities ......................... 76
Chemical vs. enzymatic DHA recycling ......................... 79
References ...................................................... 82
CHAPTER III

CATALYTIC MECHANISM OF THE DEHYDROASCORBATE REDUCTASE

ACTIVITY OF THIOLTRANSFERASE .................................... 86
Abstract ........................................................ 87
Introduction ..................................................... 88
Materials and Methods ............................................. 89

Expression and purification of wild type and mutant RPLTTs ........ 9O
DHA reductase assay and enzyme kinetics ....................... 91
Preparation of reduced thioltransferases ......................... 91

Iodoacetamide inactivation of thioltransferases DHA reductase activity 92
Reduction of DHA to AA by thioltransferases in the absence of GSH . . 92
Isoelectric focusing ......................................... 93

viii

Results and Discussion ............................................ 94
Kinetic characterization of wild type and mutant RPLTT DHA reductase

activity ............................................. 94
Proposed mechanism of RPLTT’s DHA reductase activity .......... 99
Iodoacetamide inactivation of the DHA reductase activity of wild type and
mutant RPLTT ...................................... 101
Ability of wild type and mutant RPLTTs to chemically reduce DHA
to AA ............................................. 109
Isoelectric focusing analysis of RPLTT and the C258, R26V, and K27Q
variants ............................................ 112
Role of R26 and K27 in RPLTT catalysis ....................... 120
Conclusion ..................................................... 120
Support of experiments for proposed mechanism of the DHA reductase
activity of thioltransferase ............................. 121
References ..................................................... 122

CHAPTER IV

GLUTATHIONE DEPENDENT REDUCTION OF ALLOXAN TO DIALURIC ACID
CATALYZED BY THIOLTRANSFERASE : A POSSIBLE ROLE FOR

THIOLTRANSFERASE IN ALLOXAN TOXICITY ......................... 125
Abstract ....................................................... 126
Introduction .................................................... 127
Materials and Methods . . . . . . . . . I .................................. 128

Kinetic characterization of RPLTT as an alloxan reductase ......... 130
Spectrophotometric study of uncatalyzed and RPLTT catalyzed
reaction ............................................ 130
Oxygen consumption during the GSH-dependent reduction of alloxan 131
Isoelectric focusing ........................................ 132
Results ........................................................ 132
Spectrophotometric study of the effect of RPLTT on the GSH reduction
of alloxan .......................................... 132

Kinetic characterization of thioltransferase as an alloxan reductase . . . 133
Oxygen consumption in the RPLTT catalyzed and uncatalyzed GSH

reduction of alloxan .................................. 139
Catalase and SOD inhibition of oxygen consumption in the RPLTT

catalyzed and uncatalyzed GSH reduction of alloxan ........ 139
Determination of enzyme reaction products by isoelectric focusing . . . 143 _

Discussion ..................................................... 145

Kinetic comparison of RPLTT substrates ....................... 145
Effect of RPLTT on the formation of compound 305 .............. 145
Effect of thioltransferase on dialuric acid oxidation and the production of

H202 andO2 " ....................................... 146
Proposed reaction scheme for alloxan-dialuric acid cycle ........... 147

Proposed catalytic mechanism of alloxan reductase activity of RPLTT 149

ix

References ..................................................... 15 1

CHAPTER V

SUMMARY AND PERSPECTIVES ...................................... 153
DHA reductases ................................................. 154
Catalytic mechanism of the DHA reductase activity of thioltransferase ...... 156
Alloxan reductase activity of thioltransferase .......................... 158
References ..................................................... 161

LIST OF TABLES

CHAPTER II

Table I - Kinetic constants for bovine liver PDI, RPLTT, and human erythrocyte

32kDa DHA reductase ............................................. 75
Table II- Kinetic constants for DHA reductases ......................... 77
CHAPTER 111

Table l - Wild type and mutant RPLTTs DHA reductase kinetic parameters . . . 98

CHAPTER IV

Table l - RPLTT kinetic parameters ................................. 138

xi

LIST OF FIGURES
CHAPTER I
Figure l - Structures of ascorbic acid and its oxidation products ............. 5
Figure 2 - Redox potentials of chemicals and thiol:disulfide oxidoreductases . . 18
Figure 3 - The proposed catalytic mechanism of thioredoxin ............... 20
Figure 4 - Amino acid sequence alignment of thioltransferases ............. 23

Figure 5 - The active site sequences of the thiol:disulfide oxidoreductases DsbA,
DsbB, Dst, PDI, and thioredoxin .............................. 24

Figure 6 - Kabsch-Sander representations of E. coli thioredoxin, the N-terminus of
protein disulfide isomerase, and E. coli DsbA. .................... 26

Figure 7 - Kabsch-Sander representations of T4, pig liver, and E. coli
thioltransferase. ............................................ 28

Figure 8 - Kabsch-Sander representations of E. coli thioredoxin and pig liver
thioltransferase. ............................................ 30

Figure 9 - Crystal structure of pig liver thioltransferase. ................... 31

Figure 10- Proposed catalytic mechanism of thioltransferase by Yang
and Wells ................................................. 42

Figure ll-Mechanism of thioltransferase proposed by Gravina and Mieyal. . . . 43

Figure lZ-Mechanism of thioltransferase’s DHA reductase activity proposed by
Yang and Wells. ............................................ 45

Figure l3-Alternate mechanism for the DHA reductase activity of thioltransferase
proposed by Mieyal et a1. ..................................... 46

Figure 14- Structural comparison of alloxan to dehydroascorbic acid and dialuric
acid to L-ascorbic acid. ...................................... 48

CHAPTER II

xii

Figure 1 - Kinetic analysis of the DHA reductase activity of bovine liver PDI at

pH 6.9 .................................................... 70
Figure 1 - Kinetic analysis of the DHA reductase activity of RPLTT

at pH 6.9 .................................................. 71
Figure l - Kinetic analysis of the DHA reductase activity of RPLTT

at pH 7.2 .................................................. 72
Figure 4 - Kinetic analysis of the DHA reductase acitivity human erythrocyte

32kDa DHA reductase at pH 6.9. .............................. 73
Figure 5 - Kinetic analysis of the DHA reductase acitivity human

erythrocyte32kDa DHA reductase at pH 7.2. ..................... 74

CHAPTER 111

Figure 1 - Kinetic analysis of the DHA reductase acitivity of C258. ......... 95
Figure 2 - Kinetic analysis of the DHA reductase acitivity of R26V. ......... 96
Figure 3 - Kinetic analysis of the DHA reductase acitivity of K27Q. ......... 97

Figure 4 - Proposed mechanism of GSH-dependent DHA reduction catalyzed by
RPLTT. ................................................. 100

Figure 5 - IAM inactivation of reduced RPLTT with and without preincubation
with substrates. ........................................... 105

Figure 6 - IAM inactivation of reduced C258 with and without preincubation with
substrates. ............................................... 106

Figure 7 - IAM inactivation of reduced R26V with and without preincubation
with substrates ............................................. 107

Figure 8 - IAM inactivation of reduced K27Q with and without preincubation
with substrates. ........................................... 108

Figure 9 - HPLC analysis of AA production by DHA incubation with reduced
RPLTT, C258, R26V, and K27Q. ............................ 111

Figure 10 - Isoelectric focusing analysis of incubation 03 reduced RPLTT with
various substrates. ......................................... 116

xiii

Figure 11 - Isoelectric focusing analysis of incubation 03 reduced C258 with
various substrates. ......................................... 117

Figure 12 - Isoelectric focusing analysis of incubation os reduced R26V with
various substrates. ......................................... 118

Figure 13 - Isoelectric focusing analysis of incubation 05 reduced K27Q with
various substrates. ......................................... 119
CHAPTER IV

Figure l - Structural comparison of alloxan to dehydroascorbic acid and dialuric
acid to L-ascorbic acid. ..................................... 129

Figure 2 - Broad range spectral Spectrophotometric time course study of the effect
of RPLTT on the alloxan and GSH reaction ...................... 134

Figure 3 - Time course study of the effect of RPLTT on the alloxan and GSH
reaction at 305 nm .......................................... 135

Figure 4 - pH profile of the GSH-dependent alloxan reductase activity of
RPLTT. ................................................. 136

Figure 5 - Kinetic analysis of the GSH-dependent alloxan reductase activity of
RPLTT. ................................................. 137

Figure 6 - Oxygen uptake in the uncatalyzed and RPLTT catalyzed alloxan +
GSH reaction .............................................. 141

Figure 7 - Catalase and SOD inhibition of oxygen uptake in the uncatalyzed and

RPLTT catalyzed alloxan + GSH reaction. ...................... 142
Figure 8 - Isoelectric focusing of incubation of reduced RPLTT, C258, and

K270. ................................................... 144
Figure 9 - Interactions between oxygen and dialuric acid. ................ 148

xiv

LIST OF ABBREVIATIONS

AI - amplitude

Abs-absorbance

Alx- alloxan

Alez-dialuric acid

'Ale-alloxan radical

AA - ascorbic acid

BSA-bovine serum albumin

BSO - buthionine sulfoximine

C258 - C258 mutant of RPLTT

DHA - dehydroascorbic acid or dehydroascorbate
DTT- dithiothreitol

E‘” — standard reduction potential

er - glutaredoxin

GSH - glutathione

3-Glu-Cys-Gly - glutathione

GSSG - glutathione disulfide

HED - hydroxyethyldisulfide

HZOZ-hydrogen peroxide

HOCI - hypochlorous acid

IAA - iodoacetic acid

IAM - iodoacetamide

IEF - isoelectric focusing

IPTG - isopropylthiogalactoside

K27Q - K27Q mutant of RPLTT

PDI - protein disulfide isomerase

RPLTT - wild type recombinant pig liver thioltransferase
R26V- R26V mutant of RPLTT

semiDHA - semidehydroascorbic acid or semidehydroascorbate
’OH- hydroxyl radical

02'"-superoxide

PMA - phorbol myristate acetate

PMSF - phenylmethylsulfonyl flouride

RBC - red blood cell

SOD - superoxide dismutase

TOH- K-tocopherol, vitamin E

TO ' - K—tocopheroxyl radical

TR - thioredoxin reductase

Vp-16-O ' - etoposide

32kDHAR - a 32 kDa GSH-dependent DHA reductase

XV

INTRODUCTION

The nemesis of long distance sea explorations from the 1400S to the whaling industry
of the 1800s, scurvy may have been first described in the Thebes Ebers papyrus of
approximately 1500 BC. (I). Scurvy is a condition initially marked by lethargy, malaise,
and weakness, followed by skin lesions on the lower leg, easy bruisability, bleeding gums,
and poor wound healing (2). The first scientific attempt to devise a cure for scurvy at sea
was carried out by James Lind in 1746 on the H.M.S. Salisbury. Dr. Lind described the
alleviation of the scorbutic condition by lemons and oranges on this particular voyage in “
A Treatise of the Scurvy” published in 1753. In spite of Dr. Lind’s discovery, scurvy
continued to plague long distance sea exploration until the 18003 and was a cause of
thousands of deaths during the American Revolutionary War and the American Civil War
when approximately 3000 Union soldiers imprisoned in Andersonville, Georgia, died from
scurvy. (I)

In 1895, Theobald Smith became the first person to induce scurvy in guinea pigs by
feeding them a diet of cats and bran, and be prevented scurvy in guinea pigs by adding
clover, grass, or cabbage to their diets (I). The development of the guinea pig as an animal
model led to the discovery and synthesis of ascorbic acid (vitamin C, AA) by Charles Glen
King and Albert Imre Szent-Gybrgyi in 1932 (1). Since the discovery of AA and the
development of animal models to investigate AA deficiency, multiple roles of AA in biology

have been characterized.

I. Roles of Ascorbic Acid in Biology

I. A. Role of Ascorbic Acid as a Ferric Iron Reductant.

In the AA deficient state of scurvy, the medical problems of bleeding gums and poor
wound healing are most likely the result of a defect in collagen biosynthesis (3). Prolyl-4-
hydroxylase and lysyl hydroxylase are involved in the post-translational modification of
collagen, respectively forming hydroxyproline and hydroxylysine (3). Hydroxyproline is
essential for the stabilization of the collagen triple helix (4), while hydroxylysine is necessary
for the formation of collagen intermolecular cross-links (5). Both enzymes are 1(-
ketoglutarate dependent dioxygenases which undergo the following generalized reaction (6):

(1362*)

1) 02 + K-ketoglutarate + specific substrate >
(Reductant)

 

succinate + C02 + hydroxylated substrate
Upon the oxidation of Fe“ to Fe3‘, the enzyme is no longer functional and ferric ion must
be reduced to ferrous ion (6). AA is the best known iron reductant for both prolyl-4-
hydroxylase and lysyl hydroxylase, and AA’s importance in vivo in the maintenance of these
enzyme activities is demonstrated by the condition scurvy (3, 6).

AA’s role in collagen biosynthesis surpasses its function as a cofactor for prolyl-4-
hydroxylase and lysyl hydroxylase. Incubation of AA with normal human skin fibroblasts
stimulated collagen synthesis independent of hydroxylation (7), and AA incubation with
PAT cells from chicken embryos resulted in a three fold rate of procollagen gene
transcription (8). This evidence suggests that AA may act as a signal to induce collagen
production.

AA also functions as a ferric ion electron donor for the enzymes prostaglandin

synthases I and II (9), the K-ketoglutarate dependent dioxygenases 3-butyrobetaine

dehydrogenase and 6—N-trimethyl- l -lysine hydroxylase in the camitine biosynthetic pathway
(6), and the monooxygenases peptidylglycine K-amidating monooxygenase and dopamine
2-hydroxylase (6). Dopamine 2-hydroxylase catalyzes the formation of norepinephrine
fiom dopamine (10):

2) Dopamine + 02 + 2 AA ——> Norepinephrine + 2 Semidehydroascorbic
acid

As an example of AA’s importance in vivo in the maintenance of these enzyme’s activities,
norepinephrine biosynthesis is enhanced by incubating cultured bovine chromaffin cells with
AA (10). Finally, AA stimulated the chlorinating activity of myeloperoxidase (11) by
reducing the inactive form, compound II, of myeloperoxidase back to active form of the
enzyme (12).

AA has been demonstrated to be essential for insulin release from the islets of
scorbutic guinea pigs (13). Treatment of normal and scorbutic guinea pig islets with
membrane depolarizing additions of KCl and the secretagogue D-glyceraldehyde
demonstrated the site of AA action to fall between the triose-phosphate level of glycolysis
and the increase in ATP levels leading to insulin release (14). AA may stimulate insulin
release by functioning as a ferric ion reductant for mitochondrial glycerol-3-phosphate
dehydrogenase. Addition of AA stimulated mitochondrial glycerol-3-phosphate
dehydrogenase from scorbutic guinea pig brain, liver, and skeletal muscle several fold, an
effect inhibited by iron chelators (15). Thus, the role AA plays in insulin release may also

be mediated by AA’s ability to reduced ferric ion to ferrous ion.

1. B. Antioxidant Functions of Ascorbic Acid.

Another major function of AA in addition to that of a ferric ion reductant, is as an
antioxidant. An antioxidant can be defined as “ any substance that, when compared to those
of an oxidizable substrate, significantly delays or prevents oxidation of that substrate, “(16).
Antioxidants provide an important defense against oxidative stress. Reactive oxygen species
such as superoxide (02"), hydroperoxyl radical (HOZ'), hydrogen peroxide (H202), and the
hydroxyl radical ('OH) can irreversibly modify and damage proteins, lipids, and nucleic acids
(reviewed in I 7). Free radical damage has been proposed to play a role in ageing,
atherosclerosis, carcinogenesis, and diabetes. AA is considered a primary defense because
it is able to directly interact with 02", which is the first one-electron reduction product of
oxygen (reviewed in 1 7). Other primary antioxidant defense systems include the vitamins
E and A, glutathione (GSH), uric acid, and antioxidant scavenging enzymes including
catalase, glutathione peroxidase, and superoxide dismutase (reviewed in 1 7).

AA readily oxidizes by consecutive one electron oxidations to semidehydroascorbate
(semiDHA) and dehydroascorbate (DHA) as shown in Figure 1 (18). The reaction of AA
to semiDHA has an E°’ of -282 mV at pH 7.0 and semiDHA to DHA an E°’ of 174 mV at pH
7.0 (18). AA is an excellent small molecule chain-breaking antioxidant because semiDHA
is neither a strongly oxidizing nor a strongly reducing radical, semiDHA reacts poorly with
superoxide, and only relatively small amounts of ascorbic acid need to be present to serve
as an effective antioxidant (18, 19). In solution, AA will primarily be in the form of the
ascorbate monoanion with a pKl = 4.1 (19). The ascorbate monoanion, (AA-) scavenged
'OH with a rate constant (k) = 1.28 x 10'0 M" s‘| at pH 7 .0, the quinone radical
hydroxyacetophenone- O' with a k = 1.7 x 109 M'1 s" at pH 9.5, and the phenoxy radical of

the chemotherapeutic agent etoposide (Vp- 1 6-0') with a k = 3.5 x 107 M" s‘l at pH 8.5 (19).

 

H+ H OH
0 J, 0
H — -
HT HO O
L-Ascorbic Acid H+ L-ASCOYbi¢.A¢id
Monoanron
e-
CHZOH
H OH e
O
H
O O.

Semidehydroascorbic Acid

e' CHZOH
H OH O
O
H
O O

Dehydroascorbic Acid

Figure 1: Structures of L-ascorbic acid, semidehydroascorbic acid, and
dehydroascorbic acid.

Under more physiologically relevant conditions, AA reduced Vp-16-O' in rat hepatocyte and
retinal cell homogenates (20). In vitro AA effectively scavenged 02" (21) because of the
combination of the following reactions:

3) 02" + 2H+ 9H202 ( E°’ = 940 mV)

4) AA -) semiDHA' + H+ (E°’ = -282 mV)
combining into

5) 02°” + 2H+ + AA 9H202 + semiDHA' + H+
The sum of the two reactions has an E°’ = 658 mV making it a spontaneous reaction (18).
The importance of AA as an antioxidant in physiology is demonstrated by protection by AA
against lipid peroxidation (22, 23), providing antioxidant capacity for the blood (24, 25), and

protecting against ischemia reperfusion injury (26, 27).

I. B. I. Ascorbic Acid Protection Against Lipid Peroxidation.

AA protected against lipid peroxidation in a variety of systems. Chakrabatty et a1.
(22) found that marginal AA deficiency in guinea pigs resulted in lipid peroxidation in the
cardiac tissue, and this damage was independent of decreases of other reactive oxygen
species scavengers such as superoxide dismutase, catalase, and glutathione peroxidase. In
a different system, AA protected against lipid peroxidation in human plasma induced by gas
phase oxidants of cigarette smoke (23). In contrast, in the presence of free metal ions like
iron, AA was prooxidative and induced lipid oxidation (28). It is generally believed that AA
does not directly protect against lipid peroxidation but carries out this function because of
its ability to regenerate vitamin E (K—tocopherol, TOH) from the tocopheroxyl radical (TO')

(18). Vitamin E is an important chain-breaking antioxidant in the lipid bilayer, and when it

canies out its function it is oxidized by one electron to TO’ (18). AA has been demonstrated
to regenerate TOH from T0“ in vitro (29, 30). The regeneration of TOH by AA from T0’
is a thermodynamically favorable reaction (18):

6) TO' + H+ > TOH (E°’ = 500 mV)

 

7) AA > semiDHA' + H+ (on = -282 mV)

 

Combining into the following reaction

 

8) TO' + AA >TOH + semiDHA' (E°’ = 218 mV)
As a result, for example, the protection against lipid peroxidation of rat liver microsomes by

AA was dependent on vitamin E (31).

I. B. 2. Antioxidant Function of Ascorbic Acid in the Plasma.

Extracellular fluids of the human body like blood plasma contain very low levels of
glutathione (GSH), superoxide dismutase, glutathione peroxidase, and catalase activities
which are all important intracellular antioxidant defenses (32). Substantial evidence exists
that AA is a pivotal plasma antioxidant. AA was the primary defense against lipid
peroxidation when human plasma was treated either with a water-soluble radical initiator or
by activated polymorphonuclear leukocytes (24, 25). AA is an effective scavenger of Oz"
(21) and hypochlorous acid (HOCl, 33, 34), which are both released by activated
polymorphonuclear leukocytes (2, 35). Human plasma devoid of AA, but no other
endogenous antioxidants like vitamin E or uric acid, was unable to prevent initiation of lipid
peroxidation (25). These results suggest that AA is an important antioxidant in the plasma.

In addition to being an important antioxidant in the plasma, AA protects against

ischernia/reperfusion injury. Sciamanna and Lee (26) demonstrated a dose dependent

 

protection of AA on ischemia/reperfusion injury in the rat forebrain. In this study, AA
provided partial protection even if given afier the onset of reperfusion, but maximal
protection occurred when AA was administered prior to the onset of ischemia (26). In
addition, AA protected against lipid peroxidation in rat liver when AA was administered
prior to ischemia/reperfusion (2 7). Recall that when AA functions as an antioxidant it is first
oxidized to the relatively stable radical semiDHA. SemiDHA has been identified as a
biological marker of ischemia/reperfusion injury in isolated rat hearts (36), and as a

biological marker of oxidative stress in human open heart surgery patients (37).

I. B. 3. Role of Ascorbic Acid in Immune System Function.

As a result of its antioxidant capacity, AA also plays an important role in the immune
system. Cells of the immune system such as neutrophils and macrophages accumulate at a
site of inflammation to engulf and destroy microbes in a process known as phagocytosis (2).
The process of microbial killing is predominantly carried out by oxygen dependent
mechanisms. The phagocytic vacuole contains NADPH oxidase which produces 02" also
resulting in the production of H202 by the spontaneous dismutation of 02‘". In addition,
myeloperoxidase generates HOCl from 02" and Cl". Both 02" and HOCl are important in
microbial killing. This process is known as the respiratory burst. The contents of the
phagocytic vacuole can also be released into the extracellular space producing a highly
oxidative environment around the immune cells. (2, 35).

Existing in a potential highly oxidative environment, lymphocytes and erythrocytes
must be able to protect themselves against both endogenous and exogenous oxidative stresses

and therefore require antioxidant defenses. In erythrocytes, using nitroxides as a model of

persistent free radicals, Melhom concluded that AA was the only reducing agent for
nitroxides in erythrocytes (38), a result confirmed by Zhang and Fung (39). Using the
membrane-impermeable oxidant ferricyanide, May et al. (40) found that AA in erythrocytes
was capable of protecting against ferricyanide damage with the subsequent formation of
DHA, which is reduced back to AA to provide ongoing antioxidant protection.

Erythrocytes have long been known to have impressive DHA uptake capacity and
GSH-dependent DHA reduction activity (41-43). Melhom (38) and May et al. (40) both
found the DHA reductive capacity of erythrocytes to be essential for the erythrocyte’s ability
to defend themselves against oxidative stress. When erythrocytes were depleted of AA and
incubated with DHA, their ability to reduce fenicyanide and nitroxides was restored (38, 40).
Furthermore, AA depleted, DHA treated cells have greater nitroxide reductive capacity than
buffered solutions of DHA plus GSH, suggesting the important role of a DHA reductase in
erythrocyte AA regeneration (38, 39).

Neutrophils also require robust antioxidant capacity to defend themselves against
endogenously produced oxidants. As discussed earlier, activated neutrophils undergo the
“respiratory burst” whereby they produce reactive oxygen species like 02", H202, and
HOCl. Human neutrophils stimulated with opsonized zyrnosan or phorbol myristate acetate
(PMA) oxidized 30-40% of their intracellular AA to DHA (44). Activated neutrophils, but
not quiescent neutrophils, consumed extracellular AA in a superoxide and HOCl dependent
process (45). Human neutrophils, incubated with a physiological plasma concentration of
AA (200 aM) and stimulated with PMA, N-forrnyl-methionyl-leucyl-phenylalanine, or NaF,
accumulated AA intracellularly above the levels of unstimulated neutrophils from 2 mM to

as high as 10 mM AA (46). Unstimulated neutrophils incubated with 200 uM AA in the

presence of the 02'“ generating system xanthine/xanthine oxidase increased their intracellular
AA concentration from 2 mM to 7 mM (46). Activated chronic granulomatous disease
neutrophils, which cannot produce 02'”, did not accumulate AA (46). These results
suggested that neutrophils preferentially take up DHA and reduce it to AA, and Welch et al.
(4 7) demonstrated that DHA transport in human neutrophils was at least 10-fold faster than
AA transport, and DHA and AA were taken up by separate mechanisms. Furthermore, once
DHA was transported into the neutrophil, it was immediately reduced to AA (47).
Recently, AA recycling in human neutrophils was induced by incubation of
neutrophils with Gram-positive and Gram-negative bacteria and the fungal pathogen
Candida albicans (48). Recycling was dependent on extracellular AA oxidation to DHA
because incubation of neutrophils, bacteria, and AA with superoxide dismutase and catalase
completely inhibited AA recycling (48). In addition, when neutrophils unable to make
oxidants from chronic granulomatous disease patients were incubated with bacteria, no AA
increase in neutrophils was seen (48). However, neutrophils from chronic granulomatous
disease patients incubated with DHA accumulated AA intracellularly. Wang et al. (48)
suggested that the pronounced ability of activated human neutrophils to transport DHA,
reduce it to AA and accumulate intracellular AA may be an integral component of the
antimicrobial function of neutrophils. In conclusion, both erythrocytes and neutrophils
accumulate AA intracellularly by transporting DHA across the plasma membrane and rapidly
reducing DHA to AA. This process provides antioxidant protection to these cells against
endogenously and exogenously produced oxidative stress and may be important in the body’s

ability to fight bacterial infection.

10

I. B. 4. Antioxidant Function of Ascorbic Acid in Glutathione Deficient Animals.

Perhaps the best in vivo evidence for AA as an essential antioxidant comes from
studies in glutathione deficient animals (GSH). GSH deficiency is induced in experimental
animals by administering buthionine sulfoximine (880) which is a transition state inhibitor
of 3-glutamylcysteine synthetase, an enzyme in the GSH synthesis pathway (49-50). GSH
is also an important antioxidant because it is a substrate of glutathione peroxidases which
scavenge H202 and lipid peroxides (51). In adult mice, severe GSH deficiency induced by
BSO led to damage in skeletal muscle, lung, and mucosal degeneration in the jejunum and
colon (51). In newborn rats, severe GSH deficiency induced by BSO led to 83% mortality
after seven days. In these animals, there were fewer mitochondria in brain, liver, lung , and
kidney, evidence of mitochondrial swelling and degeneration in all tissues except the heart,
and cellular damage in the lung, liver and brain (52). Martensson et al. (52) concluded that
BSO treated newborn rats, which do not synthesize AA, are excellent models for
endogenously produced oxidative stress. When AA was given to these animals along with
BSO, mortality after seven days decreased to l 1% and none of the tissue damage was seen
(52). In another study using adult mice, severe GSH deficiency induced by BSO led to
swelling and destruction of lung type 2 cell learnellar bodies and severe mitochondrial
degeneration, decreased phosphatidylcholine in the lung, decreased bronchioalveolar lining
fluid, and in the skeletal muscle, severe mitochondrial swelling and vacuolization, and
mitochondrial membrane disintegration ( 53). Concomitant addition of AA with BSO
abrogated all of these effects, and caused an increase in GSH levels in the lung, lung
mitochondria, muscle, and muscle mitochondria, but not to their original control levels (53).

Thus, AA had a sparing effect on GSH levels as part of its antioxidant function. These

11

studies by Meister and colleagues clearly demonstrated the interrelationship of GSH and AA

and the function of AA in vivo as an antioxidant.

II Ascorbic Acid Synthesis and Regeneration.

When AA carries out its function as an antioxidant or as a ferric ion reductant, it
undergoes two successive one electron oxidations to semiDHA and DHA (Figure 1). In
order to maintain sufficient tissue AA levels, AA must be synthesized , absorbed from the
diet, or its oxidation products recycled. Most species of animals are able to synthesize AA
fiom glucose (54). Humans, other primates, and guinea pigs cannot synthesize AA
endogenously because they lack the enzyme L-gulono-3-lactone oxidase which is required
for the conversion of L-gulono-3-1actone to 2-oxo-L-gulono-3-lactone, a tautomer of L-
ascorbic acid that spontaneously transforms into AA (54). Therefore, in species like humans
and guinea pigs, dietary vitamin C and recycling semiDHA and DHA are vital for the

maintenance of AA levels.

11. A. Semidehydroascorbate Regeneration.

SemiDHA is a relatively stable radical, neither strongly oxidizing nor strongly
reducing (18). SemiDHA may either be reduced back to AA by an NADH-semiDHA
reductase (55-5 7) or it rapidly disproportionates to AA and DHA with a rate constant (k) =
1.4 x 10" M" s" at pH 7.4 (19). NADH-semiDHA reductase has been identified in the
microsomal fraction of pig heart (55) and rat liver (56). In addition, a soluble NADH-

semiDHA reductase has been identified in human lens (57). However, NADH-semiDHA

12

reductase activity from any of these sources has not yet been purified and kinetically

characterized.

II. B. Dehydroascorbic Acid Regeneration.

DHA is potentially toxic to cells and tissues as highlighted by its in vitro toxicity to
pancreatic islets (58-59) and erythrocytes (60). As discussed previously, DHA is taken up
preferentially by cells like human erythrocytes (61), neutrophils (47), and fibroblasts (47)
and rapidly reduced to AA. May et al. (61 ) determined that DHA reduction by human
erythrocytes was GSH-dependent and did not involve semiDHA. In addition, Melhom (38)
and Zhang and F ung (39) concluded that the DHA reductive capacity of erythrocytes could
not be accounted for by GSH alone, suggesting the importance of a DHA reductase activity
in erythrocyte AA recycling. Therefore, recycling of DHA is important especially in species

like humans whose only source of AA is from the diet.

11. B. 1. NADPH-dependent Dehydroascorbic Acid Regeneration.

Evidence exists that both NADPH-dependent and GSH-dependent DHA reduction
are important in viva. DHA can be reduced to AA in an NADPH dependent manner by both
3Khydroxysteroid dehydrogenase (62) or the selenoprotein thioredoxin reductase (63). Rats
fed a selenium deficient diet had an 88% reduction in thioredoxin reductase activity and a
33% reduction in liver ascorbate levels providing evidence for selenium dependent enzymes,
like thioredoxin reductase, playing a role in DHA recycling (63). However, 67% of control
levels of liver ascorbate remained in selenium deficient rats, suggesting a role for GSH-

dependent DHA reductases, whose activities were not affected by selenium deficiency (63).

13

[1. B. 2. GSH-dependent Dehydroascorbic Acid Regeneration.

Unable to initiate synthesis of AA, when newborn rats were made GSH deficient by
administering the 3-glutamylcysteine synthetase inhibitor buthionine sulfoximine (BSO),
drastic decreases in tissue ascorbate levels and increases in DHA levels occurred (64).
Similar results were seen when GSH deficiency was induced in guinea pigs after BSO
administration (50), and when GSH monethyl ester, which is transported and converted
intracellularly into GSH, was administered to GSH deficient guinea pigs, the onset of scurvy
was delayed suggesting the sparing effect of GSH on ascorbate (65). In adult rrrice, which
are able to synthesize AA, GSH deficiency resulted in increased AA synthesis in the liver
(66). This result has also been seen in isolated murine hepatocytes (67). Meister and
coworkers have concluded that the drastic decrease in tissue ascorbic acid levels in GSH
deficiency demonstrated the importance of GSH dependent AA recycling (51 ). These studies
demonstrated a role of GSH in the regeneration of AA in vivo, but they did not resolve the

question of whether the regeneration of AA from DHA by GSH was enzymatic or chemical.

II. B. 3. Chemical vs. Enzymatic GSH-dependent Dehydroascorbic Acid Reduction.

GSH reduction of DHA to AA has been estimated to have a tm of 15 nrin at pH 7.5
and at 25°C (68). However, under the same conditions DHA degrades to non recyclable
products, like 2,3-diketogulonic acid with a t,,2 of 2 min (69). This strongly supports the
importance of catalytic regeneration of AA from DHA. Evidence for GS H-dependent DHA
reductases in. animal tissues have existed for many years (70- 72). Bigley et al. (73)
demonstrated that GSH-dependent DHA reductase activity was greatest in human

neutrophils, followed by monocytes, lymphocytes, and fibroblasts. Several GSH-dependent

l4

DHA reductases have been identified including the thioltransferases from vaccinia virus (74),
pig liver (75- 76), human placenta (63), and human neutrophil (77), the 32kDa-DHA
reductase isolated from rat liver (78) and human erythrocytes (79), bovine liver protein
disulfide isomerase (PDI) (63, 75, 76), the p52 enzyme from Trypanosoma cruzi (80), a
DHA reductase from spinach chloroplasts homologous to plant trypsin inhibitor (81), and
a 24 kDa DHA reductase from spinach leaves (82).

In a study by Park and Levine (7 7), the reducing activity of neutrophils after cell lysis
was non-dialysable and could not be accounted for by chemical reduction of DHA by GSH.
They found that the purified DHA reducing activity was thioltransferase (77). In rat liver
cytosol, up to 70% of the DHA reductase activity was immunotitrateable using an antibody
directed towards the rat liver 32 kDa DHA reductase (83). These experiments demonstrate
the importance of GSH-dependent DHA reductases in DHA reduction.

The emerging picture in the field of DHA reductases is one in which there are several
potential systems to regenerate AA from DHA. NADPH-dependent DHA reductases like
3K-hydroxysteroid dehydrogenase (62) or thioredoxin/thioredoxin reductase (63) and GSH-
dependent DHA reductases like thioltransferase (75) or the 32 kDa DHA reductase (78- 79)
may all play roles in vivo. It is likely that the importance of each enzyme in DHA reduction
will vary from species to species and cell to cell. Meister and coworkers have provided
strong evidence for GSH-dependent DHA reduction vs. NAD(P)H dependent DHA reduction
(50-51). However, the NADPH-dependent DHA reduction from the thioredoxin/thioredoxin
reductase system may also be important. Thioredoxin reductase (63) and the 32 kDa DHA
reductase (83) have been separately shown to be responsible for 33% and 70% of the DHA

reductase activity in rat liver, completely accounting for the DHA reductase activity in this

15

tissue. On the other hand, thioltransferase has been proposed to be the only DHA reductase
in human neutrophils (7 7).

The wide variety of DHA reductases, especially the GSH-dependent DHA reductases
raises several enzyme structure/function questions. Many of these enzymes have several
different activities including thioltransferase, thioredoxin/thioredoxin reductase, and PDI
which belong to the class of enzymes known as thiol :di sulfide oxidoreductases (84). Whether
each enzyme has an identical, similar or different catalytic mechanism for DHA reduction

is unknown.

III. Thiolzdisulfide Oxidoreductases.

Thiolzdisulfide oxidoreductases catalyze the oxidation of protein thiols and the
reduction and isomerization of protein disulfide bonds. These reactions are important for
both the structural integrity of proteins and the enzymatic activity of other enzymes (85).
Thiolzdisulfide oxidoreductases like PDI (86), DsbA and DsbB (87), thioredoxin (88), and
thioltransferase (glutaredoxin) (84), all contain the -CXXC- active site motif, which is
essential for their activity. GSH is the principal reductant of PDI and thioltransferase,
whereas NADPH, via thioredoxin reductase, is the principal reductant of thioredoxin (120).
The environment in which the -CXXC- active site exists in each of these enzymes dictates

its redox potential and the subsequent specialization of each enzymatic activity (89-90).

[11. A. DsbA, DsbB, and Protein Disulfide Isomerase.

DsbA, DsbB, and Dst in E. coli and PDI in eukaryotes are believed to be essential

l6

111

\"ai

the

[hit

(10

enzymes for the proper folding of newly synthesized proteins which contain disulfide bonds
(86-87). DsbA is a relatively oxidizing enzyme compared to other thiol:disulfide
oxidoreductases with a redox potential of -0.089 V (Figure 2; 91 ). Of the essential cysteines
for DsbA’s activity, the pKa of Cys-30 in DsbA is approximately 3.5 with Cys-33 being
relatively unreactive (92).

PDI is a highly abundant homodimer (57 kDa each, 114 kDa total) located in the
lumen of the endoplasmic reticulum (93-94). Like DsbA, PDI is also a relatively oxidizing
thiol:disulfide oxidoreductase with a redox potential of -0. l 1 V (Figure 2: 95). The sequence
of PDI reveals two domains that share approximately 30% amino acid identity to thioredoxin
(96). The N-terminal cysteines in each of the thioredoxin-like domains ofPDI possess a pK,
of 6.7 as determined by iodoacetic acid inactivation (97) and are essential for disulfide
isomerase activity ( 98). The essential function of PDI has been proposed to be its ability
to unscrarnble non-native disulfide bonds (99), but bovine liver PDI has also been
demonstrated to be a GSH-dependent DHA reductase which may also be an important

firnction of PDI (75).

III. B. T hioredoxin.

The thiol:disulfide oxidoreductase, thioredoxin, is a 12 kDa enzyme found in a wide
variety of organisms ranging from archaea to humans (100). Thioredoxins have the
conserved active site sequence of -Trp-Cys-Gly-Pro-Cys-, and a disulfide bond is formed in
the active site upon thioredoxin carrying out its function (100). The oxidized form of
thioredoxin is reduced to a dithiol by NADPH and the flavoprotein thioredoxin reductase

(100). The thioredoxin/thioredoxin reductase system has a wide variety of activities

17

pri

die

If

primarily as a catalyst for protein disulfide bond reduction or oxidation of a dithiol to a

disulfide (88), but the thioredoxin/thioredoxin reductase system also functions as an alloxan

Enzyme or chemical

Redox potential (V)

 

 

 

 

 

 

 

 

 

 

 

E. coli DsbA (91) -0.089
Bovine liver P01 (95) -O.11

Pig liver thioltransferase (149) -O.159

E. coli thioltransferase 3 (150) -0.198
cysteine (203) -O.22

E. colithioltransierase1 (150) -0.233
GSH (203) -0.24

E. colithioredoxin (88) -O.26
NADPH (88) -O.31
dihydrolipoate (203) -O.32

 

Figure 2. Redox potentials of chemicals and thiol:disulfide oxidoreductases.

reductase (101), a DHA reductase (63), and catalyzes the cleavage of S-nitrosoglutathione

to GSH and nitric oxide (102). Thioredoxin has been shown to increase proliferation of

human B-cell lines presumably due to its thiol:disulfide oxidoreductase activity (103). In

addition, thioredoxin/thioredoxin reductase have been demonstrated to function as an

endogenous regeneration system for heat shock factor during oxidative stress (104) and

pmr

15

promote AP-l activity and inhibit NF-KB activity in intact cells (105).

E. coli thioredoxin has a redox potential of -0.26V at pH 7.0, making it a relatively
strongly reducing thiol:disulfide oxidoreductase (Figure 2, 88). The essential amino acid for
E. coli thioredoxin’s activity is Cys-32 whose titration curve displays pK,s at 7.5 and 9.2
(106). The solvent inaccessible Cys-35 has a pKa of 1 1.1 in E. coli thioredoxin, but Cys-35
is believed to act as a nucleophile in the thioredoxin catalytic mechanism (107). Asp-26 is
a highly conserved residue in thioredoxins (100), and has been shown to act as a general
acid/base in the catalysis of E. coli thioredoxin (108). The proposed catalytic mechanism of
thioredoxin is shown in Figure 3 (108). To complete the mechanism, oxidized thioredoxin

is regenerated to reduced thioredoxin by thioredoxin reductase and NADPH (88).

III. C. T hioltransferase (Glutaredoxin).
The final well characterized thiol:disulfide oxidoreductase is thioltransferase,
otherwise known as glutaredoxin. The term thioltransferase was first used in 1974 by

Askelof et al. (109) to describe a partially purified rat liver enzyme which catalyzes the

 

following reaction:
9) RSSR + GSH > RSSG + RSH
10) RSSG + GSH > GSSG + RSH

 

Previously, the term transhydrogenase had been used to describe enzymes which catalyze the
combination of reactions (9) and (10)

l 1) RSSR + 2GSH > GSSG + 2RSH

 

(110-111), but Askelof, et al. (112), provided the first evidence of a two step mechanism.

Axelsson, et al., first purified thioltransferase from rat liver as an 11,000 dalton enzyme in

19

H (“202111

H—0
“’1‘
S
ASP26 03 HS
ASP26 OH C :Substrate \—< :‘f‘Substrate
O r S 1 O /S
- ( ) $35H S32
535H jaz +— K )
Trx
Trx
i (2)
’H (H202)n
H20+
ASP26 HS
OH :LJSubstrate AS1’26 Q'W H81.
HS Substrate
O J"

(3) ° S/H/\S/cs)

2’3” :— fi .2”

Trx

Figure 3: Proposed catalytic mechanism of thioredoxin from Chivers and Raines

(108). In this mechanism, Cys-32 of E. coli thioredoxin attacks a sulfur group

of the substrate froming a mixed disulfide between thioredoxin and the substrate

(reaction 1). After deprotonation by water (reaction 2), Asp-25 then deprotonates

Cys-35 of thioredoxin which attacks Cys-32 releasing reduced substrate and

forming oxidized thioredoxin (reaction 3). Oxidized thioredoxin is reduced by

thioredoxin reductase and NADPH.

20

 

1978 and characterized thioltransferase’s ability to catalyze the GSH-dependent reduction
of S-sulfoglutathione, S-sulfocysteine, L-cystine, L-homocystine, cystamine, oxidized
trypsin, ribonuclease, oxytocin, or insulin, and mixed disulfides of GSH and coenzyme A,
L-cysteine, cysteamine, and hen egg white lysozyme.

In 1976 Holmgren isolated an E. coli mutant strain lacking thioredoxin activity, but
retaining the ability to provide reducing equivalents to ribonucleotide reductase (113).
Holmgren named this heat stable factor glutaredoxin, and found it to catalyze the hydrogen
transfer from GSH to ribonucleotide reductase (113). E. coli glutaredoxin was determined
to be a protein of 1 1,600 daltons (1 14), and was shown to catalyze reaction (12) in an in vitro
reconstituted system with the B1 and B2 subunits of ribonucleotide reductase:

(12) 20511 + CDP

 

> GSSG + dCDP + H20 (115)
Glutaredoxin has also been suggested to play a role in ribonucleotide reductase function in
calf thymus (116) and vaccinia virus (11 7).

Until 1987, thioltransferase and glutaredoxin were considered distinct enzymes. In
198 7, Gan and Wells determined the primary structure of pi g liver thioltransferase, a 12 kDa
enzyme, and found it to be 82% identical to calf thymus glutaredoxin (118). Then in 1989,
when the amino acid sequence of calf thymus glutaredoxin was reexamined by mass
spectroscopy and compared to pig liver thioltransferase and rabbit bone marrow
glutaredoxin, each possible comparison yielded at least an 83% sequence identity (119). It
is now generally accepted that glutaredoxin and thioltransferase are two names for the same
enzyme (120, 84). For the purpose of the discussion, from this point on this class of

enzymes will be solely referred to as thioltransferase.

21

III. C. 1. Purification, Cloning, and Sequences of T hioltransferases.

Thioltransferase has been purified and its amino acid sequence determined from
human red blood cells (121), human placenta (122), pig liver (1 18), calf thymus (1 19), rabbit
bone marrow (123), rice (124), yeast (125), E. coli (126), vaccinia virus (127), and
bacteriophage T4 (128). Each enzyme possesses the common active site ~CPF/YC- (amino
acid alignment in Figure 4). Thioltransferase has been cloned from human brain (129),

human placenta (130), pig liver (131), rice (132), E. coli (133), and vaccinia virus (74).

IV. 3D Structures of Thiolzdisulfide Oxidoreductases.

Thioredoxins, PDI, DsbA, DsbB, and Dst are related enzymes based on the
similarity of their active site sequences (Figure 5). The conserved Asp at position 26 in E.
coli thioredoxin is a Glu in PDI and DsbA suggesting that a Glu at this position may function
as a general acid/base in PDI and DsbA catalysis. Thioredoxins, DsbA-C and PDI, also

show conserved 3D structural characteristics beyond the similarity of their active sites.

22

 

Figure 4. Amino acid sequence alignment ofthioltransferases. The active site -CPF/YC- is
in bold and underlined. Val-59, Thr-73, and Asp-74 in E. coli thioltransferase are also in
bold and underlined because these residues have been shown to be involved in binding GSH
(143).

23

 

 

 

 

 

 

 

 

 

 

 

 

 

Enzyme Active Site Sequence

E coli DsbA (137) EFFSFFCPHCYQFE

E coli DsbB (204) VMLLKCVLCIYER
E coli Dst (205) VFTDITCGYCHKLH
Rat PDI domain I (100) EFYAwaKALA
Rat PDI domain 11 (100) EFYAPmeQLA
E coli thioredoxin (100) DFWAESQGEKMIA
Mouse thioredoxin (100) DFSATWQQPQKMIK
Human thioredoxin (100) DFSATWCGPCKMIK

 

Figure 5. The active site sequences of the thiol:disulfide oxidoreductases DsbA, DsbB,
Dst, PDI and thioredoxin. The active sites of each enzyme are in bold and underlined.

24

IV. A. 3D Structure of T hioredoxin.

The crystal structure of the oxidized form of E. coli thioredoxin at 4.5 A was first
reported in 1974 by deerberg et al. (134), and further refined to 1.68A by Kati et al. (135)
in 1990 ( Figure 6). The secondary structure of E. coli thioredoxin is Blalflzmfla, 04 D, cg and
is composed of a core of a twisted B-sheet flanked on either side by helices (Figure 6, 135).
The active site forms a protrusion on the surface of the molecule and is at the amino terminal
end of helix (1,. The -8G including Cys-32 is solvent accessible, whereas Cys-35 is not. The
proximity of Cys-32 to the positive electrostatic field of the N-tenninal end of an a helix may
contribute to the low pK, of Cys-32 and its subsequent reactivity (135). In support of this
concept, Kortemme and Creighton (136) demonstrated with model a-helical peptides that
a decrease in the thiol pK, of cysteine occurs when cysteine is near the N-terminus of the a-

helix, whereas at the C-terminus the thiol pK, value increases slightly.

IV. B. 3D Structures of DsbA and Protein Disulfide Isomerase.

Oxidized DsbA has been crystallized and its structure determined (13 7). A portion
of the structure of DsbA closely resembles that of thioredoxin (Figure 6). In Figure 6, the
structure of DsbA is given with the thioredoxin-like structure facing the reader and on the
right hand side of the molecule. The a-helices behind the active site are not part of the
thioredoxin fold. In addition, the structure of the N-terminal domain, residues 1-120, of
human PDI in the oxidized form has been determined by NMR and it’s structure also closely
resembles that of thioredoxin (Figure 6, 138). In each of the described structures, the -
CXXC- active site is located at the N-terminus of an a helix most likely contributing to the

nucleophilicity of the N-terminal cysteine. The common structural motif seen in these

25

N-terminus of
E. coli thioredoxin Bovine Liver PDI

    

Figure6. Kabsch-Sanderrepresentationofthe structuresofE. colithioredoxin,

the N-terminus of protein disulfide isomerase, and E. coli DsbA. The active site
disulfideofeachenzymeis labeledandis formdatorneartheN-termimrs ofan

alpha helix (designated by the red cylinders).

26

enzymes of a central core composed of a [i-sheet surrounded by a-helices has become known
as the thioredoxin fold. The structure of reduced E. coli thioredoxin has been determined via
NMR, and the overall structure is very similar to the crystal structure of oxidized thioredoxin

(139).

IV. C. 3D Structures of T hioltransferases.

The 3D structures of oxidized pig liver thioltransferase (140), oxidized E. coli
thioltransferase (141), reduced E. coli thioltransferase (142), E. coli thioltransferase mutant
C148 in complex with GSH (143), and oxidized T4 bacteriophage glutaredoxin (144) have

been determined.

IV. C. 1. 3D Structure of Bacteriophage T4 T hioltransferase.

The structure of oxidized T4 thioltransferase is very similar to that of thioredoxin
with a central B-sheet surrounded by (at-helices except for the first 20 amino acids in
thioredoxin which form a B-strand and a-helix not found in T4 thioltransferase, and the
active site of E .coli thioredoxin protrudes more from the molecule whereas the active site
of T4 thioltransferase is located more in an occluded cleft (Figure 7, 144, 140). The active
site disulfide of Cys-14 and Cys-17 is found at the N-terminus of helix a1, and Cys-l7 is

solvent inaccessible (Figure 7, 144).

IV. C. 2. 3D Structures of E. coli T hioltransferase.
Both reduced and oxidized E. coli thioltransferase were found to have very similar

structures, except the solvent accessible surface area in the active site of the reduced enzyme

27

T4 thioltransferase Pig Liver thioltransferase

  
  

Active site

  

E. coli thioltransferase

Active site

 

Figure 7: Kabsch-Sander representations of the structures of T4 thioltransferase,
pig liver thioltransferase, and E. coli thioltransferase. The active site disulfide of
each enzyme is shown and is near the N-terminus of an alpha helix (represented by
red cylinders).

28

was larger than that of the oxidized enzyme (14]). Again, the active site cysteines were
located at the N-terminus of a-helix l and Cys l 1 is solvent accessible whereas Cys 14 is not
(Figure 7, 141). Xia et al. (141) found that the structures of oxidized E. coli thioltransferase
and oxidized T4 thioltransferase were very similar to each other with a BaBaflBa structure
and to E. coli thioredoxin, which is [1010010045001 possessing an extra 0 strand and a-helix at

the N-terminus (Figure 7).

IV. C. 3. 3D Structure of Pig Liver T hioltransferase.

Oxidized pig liver thioltransferase also possesses a very similar overall fold to other
thioltransferases and the thioredoxins, with a central B-sheet surrounded by a-helices (figure
7, 140). Major differences include E. coli thioredoxin having an additional B-strand prior
to or-helix l of pig liver thioltransferase, and pig liver thioltransferase having an a-helix prior
to B-strand 1 not found in T4 or E. coli thioltransferase. As a result, thioltransferases are also
referred to as having the “thioredoxin fold”, as seen in Figure 8 where the structures of E.
coli thioredoxin and pig liver thioltransferase are compared. Note how the active sites are
at the N-terminus of an a-helix and the overall structure is that of a central B-sheet
surrounded by a-helices.

The active site in pig liver thioltransferase of -Cy322-Pr023-Phe24-Cy825- is found
at the amino terminus of helix «2 and Cys 22 is solvent accessible and forms a hydrogen
bond with the main-chain nitrogen atom N 25 (3.3 A) (Figure 9, 140). Katti et al. (140)
compared the disulfide bridge of pig liver thioltransferase to those of E. coli thioredoxin and
T4 thioltransferase finding the 14-membered rings to be geometrically very similar. It is

believed that the l4-membered disulfide ring of -CXXC- is evolutionarily favored because

29

 

Pig Liver thioltransferase

E. coli thioredoxin

   

Figure 8. Kabsch-Sander representations of the structures of E. coli thioredoxin and
pig liver thioltransferase. The active site disulfide of each enzyme is shown and is
near the N-terminus of an alpha helix (represented by red cylinders).

30

 

Figure 9. Crystal structure of pig liver thioltransferase. The active site
disulfide 'sz‘x'x‘czr is shown as are LY319» A1325, 14327, 611157: Cys 78:
and Cys”.

31

IV. C. 4. Glutathione Binding to T hioltransferase.

The GSH binding site of thioltransferase has been explored in T4 thioltransferase and
E. coli thioltransferase. Modeling of GSH into the crystal structure of T4 thioltransferase
suggested that His- 12, Cys- 1 4, Val-15, Cys-17, Tyr-16, Thr-64, and Asp-80 may be involved
in GSH binding (146). Cys-14 was shown to be the key residue for thioltransferase activity
and residues His-12 and Tyr-l6 were shown to be important in GSH binding (146). The
mutation of Asp-80 to Ser resulted in enhanced GSH affinity, and furthermore, in other

thioltransferases the comparable residue is a threonine or serine (Figure 4, 146).

IV. C. 5. Solution Structure of the Mixed Disulfide Between E. coli Thioltransferase and
Glutathione.

The solution structure of E. coli thioltransferase active site mutant protein C148
forming a mixed disulfide with GSH has been solved (143). The -8G moiety was found to
be localized in a cleft (143), which is not seen in pig liver thioltransferase (140). The cleft
in E. coli thioltransferase is bound by residues Tyr-l3, Thr-5 8, Val-59, Tyr-72, Tyr-73, and
Asp-74. The overall structure of the C148-mixed disulfide closely resembles that of
oxidized E. coli thioltransferase, more so than reduced E. coli thioltransferase (143). The
strongest interactions of the enzyme with ~8G were hydrogen bonds seen between the
carbonyl oxygen atom of the Cys residue in -SG and the amide proton of Val-59 and the
carboxylate group of y-Glu in -SG and the amide proton of Thr-73 (143). In addition, the
carboxylate group of y-Glu of -SG was located at the N-terminus of a helix 3, and therefore
the partial positive charge located at the N-terminus of an a-helix may contribute to the

binding of GSH to the enzyme (143). Weaker interactions between the enzyme and -SG

32

included hydrogen bonds between the carboxylate group of -8G and the amide protons of
Tyr-72 and the OH of Thr-73 and salt bridges between the NH; group of Glu in -SG and the
side chain carboxylate of Asp-74 and between the carboxylate group of Gly in -8G and the
side chain NH3+ of Lys-45 (143). In other thioltransferases, Val and Asp are completely
conserved in positions comparable to Val-59 and Asp-74 in the E. coli enzyme (Figure 4 ).
Tyr-73 in E. coli thioltransferase corresponds to either a Tyr or a Ser in nearly all the other
thioltransferases (Figure 4). Lys—45 in E. coli thioltransferase does not appear to be
conserved, but in the crystal structure of pig liver thioltransferase Gln-57 seems to
correspond to the position of Lys-45 in E. coli thioltransferase (Figure 9).

Most of the interactions between GSH and the C 148 mutant E. coli thioltransferase
outside of the disulfide bond itself were between the y-Glu moiety of GSH and the enzyme
with only one contact between the Gly residue and the enzyme (143). Rabenstein and Mills
(14 7) studied the thioltransferase catalyzed GSH/GSSG interchange reaction by NMR using
pig liver thioltransferase finding the y-Glu-Cys moiety of GSSG and GSH-containing mixed
disulfides to be essential for recognition by thioltransferase. However, the rate constant for
catalysis of the GSH/GSSG interchange was 10 fold faster than the rate constant for the
catalysis of y-Glu-Cys/y-Glu-Cys disulfide interchange (I4 7). These results were supported
by Srinivasan et al. (148) who found the y-Glu—Cys moiety of GSH to be essential for
reactivity by human erythrocyte thioltransferase, but thioltransferase catalysis was enhanced
further by GSH. These results suggest that the structure of a mixed disulfide between GSH
and a -CXX8- mutant of a mammalian thioltransferase would show similar enzyme/GSH

interactions to those seen in E. coli thioltransferase and GSH (I43).

33

V. Function and Mechanism of Thioltransferase.

V. A. Redox Potentials and pKa of Active Site Cysteines in T hioltransferases.

Thioltransferases have redox potentials which are more reducing than DsbA or PDI,
but not as reducing as thioredoxin, ranging from -0. 159 V for pig liver thioltransferase (149)
to -0.233V for E. coli thioltransferase 1 (150) (Figure 2). A hallmark of thioltransferase as
compared to other thiol:disulfide oxidoreductases is the extremely low pK, of the N-terminal
active site cysteine. Cys-22 in pig liver thioltransferase has a pK, of 3.8 (151), Cys-22 in
human thioltransferase a pK, of 3.8 (152), and Cys-26 in yeast thioltransferase a pK, of 3.5
(153). The N-terminal cysteine of pig liver thioltransferase, Cys-22, has been demonstrated
to be essential for thioltransferase activity (151).

Several factors contribute to the extremely low pK, of Cys-22. In the crystal structure
of pig liver thioltransferase, Cys-22 is situated near the amino terminal end of helix «2,
which provides a partial positive charge to stabilize the negative charge on the sulfur atom
(140). Also, the sulfur atom of Cys-22 forms a hydrogen bond with the main-chain nitrogen
atom of Cys-25 (140). Yang and Wells carried out a site-directed mutagenesis study of pig
liver thioltransferase and determined that Arg-26 was a major contributor to the low pKl of
Cys-22 because the R26V and R26V:K27Q mutant proteins were weakly sensitive to
carboxyrnethylation with an undetermined pKa for Cys-22 that was pH independent from 2.5
to 8.5 (151). Katti et al. (140) concluded from the crystal structure that Arg-26 had the
potential to swing its side chain into the active site without much steric hindrance. The
combination of Arg-26, the position of Cys-22 at the amino-terminal end of helix «2, and the

sulfur atom hydrogen bond to main-chain nitrogen atom of Cys-25 most likely act

34

 

 

synergistically resulting in the low pK, of Cys-22 thiol.

V. B. Protein Deglutathionylation Activity of T hioltransferase.

Although thioltransferase is capable of catalyzing the GSH-dependent reduction of
small molecular weight molecules like S-sulfocysteine and the disulfides cystine and
hydroxyethyldisulfide (112, 152) this is probably not a significant or important role of
thioltransferase in vivo. When Axelsson et al. (112) first described thioltransferase, they
also demonstrated its ability to catalyze the reduction of a mixed disulfide of a protein such
as lysozyme and GSH. This process now named protein deglutathionylation is where

thioltransferase catalyzes the following reaction:

 

(l3) protein-SSG + GSH >protein-SH + GSSG
Protein glutathionylation is reversible, occurs as an early cellular response to oxidative stress
(154-159), and has been proposed to be an important mechanism to protect proteins from
irreversible modification by oxidative stress (160-161). The activities of many enzymes
including glycogen phosphorylase, glycogen synthase, phosphofi'uctokinase, pyruvate
kinase, and protein kinase are modified by thiol:disulfide exchange (162-163).
Thioltransferase has been demonstrated to catalyze the GSH-dependent reactivation
of oxidatively inactivated pyruvate kinase (164-1 65), omithine decarboxylase (I 66), human
and rat pi class GSH-S-transferases (167-168), phosphofiuctokinase (169), and fi'uctose 1,6-
bisphosphatase (1 70). Recently, Zheng et al. (1 71) demonstrated the autoregulation of the
E. coli transcription factor OxyR by E. coli thioltransferase. When E. coli cells were

subjected to peroxide stress, OxyR was activated by oxidation and induced the production

of thioltransferase (172). After induction of thioltransferase and restoration of the

35

GSH/GSSG ratio, OxyR was reduced and inactivated by thioltransferase (1 71). In none of
these cases was reactivation shown to be determined by deglutathionylation of these
enzymes.

Gravina and Mieyal (1 73) have argued that thioltransferase is a specific glutathionyl
mixed disulfide oxidoreductase. They proposed that thioltransferase catalyzes the following
reaction:

(14) GSSR + GSH >RSH + GSSG

 

where the first step in GSH-dependent RS SR reduction is nonenzymatic and forms the GS SR
substrate, which thioltransferase then catalytically reduces (I 73). Gravina and Mieyal
demonstrated that thioltransferase catalyzed the deglutathionylation of papain-S-SG, BSA-S-
SG, and oxyhemoglobin-S-SG but not papain-S-S-cysteine, BSA-S-S-cysteine, hemoglobin-
S-S-cysteine, or hemoglogin-S-S-cysteamine (1 73). Gravina and Mieyal’s work suggests
that the GSH-dependent reactivation of oxidatively inactivated enzymes catalyzed by
thioltransferase discussed previously is indeed deglutathionylation. In support of this
concept, thioltransferase reactivated inactivated glutathionylated HIV-I protease in a GSH-
dependent manner (1 74), and the DNA-binding activity of oxidatively inactivated nuclear
factor I (I 75). A shortcoming of Gravina and Mieyal’s work is they did not analyze
thioltransferase catalyzed GSH-dependent dethiolation of small molecular weight disulfides
like cystine or HED. It is possible that a mixed disulfide containing GSH, RSSG, is a
favored substrate due to the proposed GSH binding site in thioltransferase. With very large
substrates like protein mixed disulfides, steric hindrance may only allow thioltransferase to
catalyze the reduction of protein-S-SG bonds, but small disulfides like HED may not face

an active site accessibility problem. Most likely, thioltransferase is a specific glutathionyl

36

mixed protein disulfide oxidoreductase and also a catalyst of GSH-dependent reduction of
small molecular weight disulfides.

An additional role for thioltransferase and its catalysis of deglutathionylation has
been demonstrated in vitro in accelerating PDI catalysis of GSH-dependent folding of
ribonuclease A (I 76, 177). The role of thioltransferase in the refolding of ribonuclease
resides in thioltransferase’s ability to catalyze both the formation and reduction of mixed
disulfides containing GSH which are then converted rapidly to intrarnolecular disulfides in
the presence of PD] (1 7 7).

Under conditions of oxidative stress, thiol:disulfide oxidoreductases like
thioltransferase, thioredoxin, and protein disulfide isomerase may all play roles in protein
dethiolation. For example, human erythrocyte membranes damaged by diamide treatment
were reduced to free sulflrydryls in the presence ofthioltransferase (169). Thioltransferase
and thioredoxin appear to have different substrate specificities with respect to regenerating
oxidatively damaged proteins (178). Yoshitake et al. (178) found that thioltransferase
regenerated phosphofructokinase and glyceraldehyde 3-phosphate dehydrogenase inactivated
with mixed disulfide bonds, whereas thioredoxin regenerated phosphofi'uctokinase and
glyceraldehyde 3-phosphate dehydrogenase inactivated by monothioloxidation to sulfenic
or sulfinic acid. In support of these results, Jung and Thomas (149) also found
thioltransferase to more efficiently catalyze the deglutathionylation of protein mixed

disulfides than thioredoxin or protein disulfide isomerase.

V. C. The Catalytic Mechanism of T hioltransferase.

The catalytic mechanism of thioltransferase has been studied in detail. Yang and

37

Wells (151) demonstrated by site directed mutagenesis of the. gene encoding pig liver
thioltransferase that Cys at position 22 is the essential amino acid for activity. Cys-25 in pig
liver thioltransferase was not essential for activity because its mutation to a serine led to a
10% increase in relative activity, but the C25A mutant protein had a 91% loss in relative
activity (151). This result suggested that either a sulfur or an oxygen atom is required at
position 25 as part of the catalytic mechanism, but the placement of the hydrophobic amino
acid alanine at position 25 may have disrupted the stability of the enzyme complicating the
interpretation of these results (151). Arg at position 26 and Lys at position 27 were also
shown to be important in the mechanism because the R26V, K27Q, and R26V:K27Q mutant
proteins had a loss in relative activity of 68%, 33%, and 95%, respectively (151). As
discussed earlier, Arg-26 was determined to play a major role in the nucleophilicity of Cys-
22, but the role of Lys-27 in the mechanism was not determined.

Using proteins isolated from these site directed mutants, Yang and Wells (1 7 9) then
determined the catalytic mechanism of pig liver thioltransferase using S-sulfocysteine (Cys-
803‘) and the small molecular weight disulfides cystine and hydroxyethyldisulfide (HED).
Preincubation of wild type recombinant pig liver thioltransferase (RPLTT) and the C258,
R26V, and K27Q mutant proteins with iodoacetamide (IAM) yielded residual activities of
approximately 10%, 0%, 7 0%, and 15%, respectively (1 79). Thus, R26V was only partially
inactivated which was expected based on the importance ofArg-26 in the nucleophilicity of
Cys-22. Preincubation of RPLTT, C258, R26V, and K27Q with GSH prior to the addition
of IAM did not provide protection against IAM inactivation and actually potentiated the
inactivation of RPLTT, R26V, and K27Q (I 79). RPLTT, R26V, and K27Q were partially

or fully protected from IAM inactivation by preincubation with Cys-803', GSH + Cys-803',

38

and HED (1 79). In contrast, C258 was only protected from IAM inactivation by
preincubation with GSH + Cys-803' (1 79).

RPLTT, R26V, and K27Q have different isoelectric points in the oxidized and
reduced form because of the formation of an intramolecular disulfide bond upon oxidation
with the subsequent loss of the negative charge of Cys-22 (151). To analyze the oxidation
state of the enzymes and possibly demonstrate the presence of enzyme-substrate
intermediates from the preincubation mixtures, Yang and Wells (1 79) carried out isoelectric
focusing studies on RPLTT and the C258 mutant. RPLTT has a pl of 7 .0 and a pI of 8.0 in
the reduced and oxidized forms, respectively (151). Oxidized RPLTT was found by IEF
when reduced RPLTT was treated with HED, cystine, Cys-803', cystine + GSH, and GSH
+ Cys-803' (179). Reduced RPLTT was also detected when RPLTT was incubated with
Cys-803', cystine + GSH, and GSH + Cys-803' (1 79). Finally, a possible enzyme-substrate
intermediate was seen when RPLTT was incubated with Cys-803‘ (179). The protection
against IAM inactivation by preincubating RPLTT with HED, Cys-803' , and Cys-803' +
GSH was most likely due to the formation of oxidized enzyme, but may also be due to the
formation of an enzyme-substrate intermediate when RPLTT was preincubated with Cys-
SO,‘ (I 79). These results demonstrated the ability of the disulfide substrate (RSSR) to bind
to reduced enzyme forming oxidized enzyme and releasing 2 RSH (reactions 1 and 2 in
Figure 10).

The C258 mutant protein does not possess the ability to form an intramolecular
disulfide at the active site. IEF of C258 preincubated with HED, cystine, and Cys-803'
demonstrated the formation of enzyme-substrate intermediates, but HED and Cys-803'

preincubation with C258 did not protect against IAM inactivation (1 79). These results

39

suggested that the enzyme-substrate intermediates were in equilibrium with reduced enzyme
+ substrate, and reduced enzyme was susceptible to IAM inactivation (reactions 1 and 6 in
Figure 10). Enzyme-substrate intermediates of RPLTT and C258 were demonstrated by
radioactive labeling of each enzyme with L-['4C]-cystine and Cys-803' + [3H]-G8H but not
[3H]-GSH alone demonstrating the ability of the disulfide substrate to bind to thioltransferase
before GSH, a result supported by IAM inactivation studies (1 79).

Yang and Wells (179) propose a mechanism for thioltransferase that is shown in
Figure 10. In this mechanism, the disulfide substrate RSSR binds first to the enzyme
releasing RSH and forming the mixed disulfide ES-SR (reaction 1, Figure 10). Enzymes
possessing both cysteines at the active site could then undergo two possible pathways where
(1) Cys-25 attacks Cys-22 releasing RSH and forming an intramolecular disulfide (reaction
2, Figure 10) which is then attacked by GSH forming the mixed disulfide ES-SG (reaction
3, Figure 10), or (2) the first molecule of GSH attacks ESSR forming ES-SG and releasing
RSH (reaction 4, Figure 10). Lacking the second active site cysteine, the C258 mutant
protein can only undergo reaction 4. ES-SG is then reduced by the second molecule of GSH
to form ES' releasing GSSG to complete the catalytic cycle (reaction 5, Figure 10). Reaction
6 demonstrates the irreversible modification of reduced enzyme by IAM.

Gravina and Mieyal (173) proposed a different mechanism for thioltransferase as
shown in Figure l 1. Recall that Gravina and Mieyal provided evidence for thioltransferase
being a specific glutathionyl mixed disulfide oxidoreductase (1 73), but they only studied
protein mixed disulfides. They proposed that of reactions (15) and (16) thioltransferase only
catalyzes reaction (16)

(15) RSSR’ + GSH -- RSH + GSSR’

40

 

(l6) GSSR’ + GSH -= GSSG + R’SH
with reaction (15) occurring spontaneously (173). This may be true for protein mixed
disulfides due to steric constraints because of the size of the substrate, but Yang and Wells
provide strong evidence that thioltransferase catalyzes both reactions (15) and (16) with
small molecular weight molecules like Cys-803' or the disulfides HED and cystine (1 73).
The Gravina and Mieyal mechanism for thioltransferase (Figure 11) is a ping pong
mechanism consistent with the observed parallel lines in two-substrate kinetic studies (1 73).
As stated previously most likely a major in vivo function of thioltransferase is to catalyze
protein deglutathionylation, and therefore the mechanism proposed by Gravina and Mieyal

may be important.

V. D. The Dehydroascorbate Reductase Activity of T hioltransferase.

It was long believed that thioltransferase was only a GSH-dependent thiol:disulfide
oxidoreductase, but in 1990 Wells et al. (75) demonstrated that human placenta, bovine
thymus, and pig liver thioltransferases possessed GSH-dependent DHA reductase activity.
Subsequently, thioltransferase fi'om vaccinia virus (74) and human neutrophils (77) were
demonstrated to possess GSH-dependent DHA reductase activity. There are conflicting
reports concerning the DHA reductase activity of plant thioltransferase. Sha et al. (124)
reported that purified rice thioltransferase has DHA reductase activity, although the same
group reported that cloned rice thioltransferase does not have DHA reductase activity (132).
In addition, Morell et al. (180) reported that purified spinach thioltransferase lacks DHA
reductase activity.

The identification of thioltransferase as a DHA reductase raised the fundamental

41

y—Glu-Cys-Gly

52201420014142 3': -8
RPLTT 1-1/
$25H RSH m \l 7
(OH) 525
”' sam/ (3)
[A (6) mun/$32 \
RSH \25- H Y-Glu-Cys-Gly

I
| S . . .
RPLW<§<JSR ‘— (OR) RSH mfg film
(335*. f‘ V (OHIH
fiq/ /
\

    
 

(5)

GSSG

Figure 10. Proposed catalytic mechanism of thioltransferase by Yang and Wells
(1 79). The C258 mutant is represented by (OH) placed under Cys-25.

42

Y-Glu-Cy'Is-Gly (SR Y-Glu-Cys-Gly
.. l
S l y/ ‘-3
SR /322
RPLTT I
RSH (1) X \325
y-Glu-Cys-Gly /
g x’/
-Glu-Cys-G|y , ’
. ,4 (3)

l RPLTT
S- SR
1 \4 RSH \25 ' H
3&2

(2) l
m( (OR)

3st +—

y—Glu-Cys-Gly
I .
S
‘ J CuCys—Gly
I
/ ..
(4) RPLTT S
\
$25 ' H
GSSG

Figure 11. Mechanism of thioltransferase proposed by Gravina and Mieyal

(I 73). In this reaction mechanism, Gravina and Mieyal propose that GSH

chemically reacts first with the disulfide substrate RSSR forming RSSG and

releasing RSH (reaction 1). Cys-22 of thioltransferase then attacks the sulfur

group of GSH in the RSSG substrate releasing RSH forming a mixed disulfide

between enzyme and -SG (reaction 2). The mixed disulfide is then reduced by

a second molecule of GSH forming GSSG and reduced enzyme completing the
reaction cycle ( reaction 4).

43

enzyme structure/ function question, does thioltransferase have analogous/different catalytic
mechanisms for its DHA reductase and thiol:disulfide oxidoreductase activities? Yang and
Wells (151) provided initial evidence supporting analogous mechanisms when they
compared the relative DHA reductase activities of RPLTT to the variant proteins and found:
C228 (no activity), C258, (191% activity), R26V (30% activity), K27Q (73% activity),
R26V:K27Q and C25A (both with no activity). These results demonstrated that Cys-22 is
the essential amino acid for both the thiol:disulfide oxidoreductase and DHA reductase
activities of thioltransferase (151). In addition, Arg-26 and Lys-27 are important for both
activities and Ser can replace Cys at position 25 in both activities, but Ala cannot (151).
Surprisingly, the C258 mutant protein had nearly double the relative DHA reductase activity
of the wild type but only 10% more thioltransferase activity (151). Yang and Wells proposed
a DHA reductase mechanism analogous to that of thioltransferase as shown in Figure 12
(I 79). In this mechanism, they proposed the formation of a thiohemiketal intermediate
which is releasable by GSH, but not by attack of Cys-25 on Cys-22, forming AA and the
mixed disulfide ES-SG (1 79). Similarly, Mieyal et al. (84) proposed the formation of the
intramolecular disulfide and release of AA in the DHA reductase activity ofthioltransferase
as one possible mechanism. However, based on their studies which suggested that
thioltransferase is a specific glutathionyl mixed disulfide oxidoreductase, Mieyal et al. (84)

also proposed that the true substrate for thioltransferase is DHA-8G (Figure 13).

V. E. Other Potential Non-disulfide Substrates for Thioltransferase.
It is possible that other non-disulfide substrates for thioltransferase exist. For

example, the thioredoxin/thioredoxin reductase system has been shown to catalyze the

44

CHZOH

 

 

HO
O H
(1) y-Glu-Tys-Gly O
H
’
S_/\SZZ ’3
/ OH 0
RPLTT
32,11

RPLTT\
S25H / (2)
AA
(3) y-Glu-Cys-Gly

GSSG
m s

825H y-Glu-Cys-Gly

Figure 12. Mechanism of thioltransferase's DHA reductase activity proposed by
Yang and Wells (179). In this mechanism, reduced enzyme attacks C1 of DHA
forming a thiohemiketal intermediate (reaction 1). GSH then attacks Cys-22
releasing AA and forming a mixed disulfide between enzyme and -SG (reaction 2).

A second molecule of GSH then attacks the sulfur of the GSH in the mixed disulfide

releasing GSSG and restoring the reduced form of the enzyme (reaction 3).

45

CHZOH
on
H o

H
o ( .33
F (1)
y-Glu-Cys-Gly \

crrzorr
OH

O H I
AA
y-Glu-CyS-Gly Y -Glu -Cys-Gly
s_ l
/ 22 (2) (Dis
322
/ k,—

RP LTT S
l RP LTT I

O

SZSH
SZSH Y-Glu-Cys-Gly

(3)

GSSG

Figure 13. Alternative mechanism for the DHA reductase activity of
thioltransferase proposed by Mieyal et al.(84). In this mechanism, DHA reacts
with GSH chemically forming a thiohemiketal intermediate in reaction (1). Cys-22
of thioltransferase attacks the sulfur atom of GSH forming a mixed disulfide between
enzyme and -SG releasing AA in reaction (2). A second molecule of GSH attacks
the sulfur atom of the GSH in the mixed disulfide releasing GSSG and forming

reduced enzyme to complete the catalytic cycle.

46

NADPH dependent reduction of the diabetogenic agent alloxan to dialuric acid (101). Based
on the structural similarities between alloxan and DHA (Figure 14), thioltransferase may be

a GSH-dependent alloxan reductase.

V. E. I. Alloxan Toxicity to the 2-cells of the Pancreas.

Alloxan (2,4,5,6 [1H,3H] pyrimidinetetrone) is selectively toxic to the 2-cells of the
pancreas of certain animals, e. g., mice and rats (181-183). This toxicity has been attributed
in part to accumulation of alloxan in the islets (184. 185) and 2-cells (186). Furthermore,
the pancreas possesses relatively lower levels of enzymes such as catalase, Cu-Zn superoxide
dismutase (SOD), Mn-SOD, and glutathione peroxidase, which protect cellular components
against reactive oxygen species (I 85, I 87). The toxic mechanism of alloxan involves redox
cycling between alloxan and dialuric acid. Reduction of alloxan to dialuric acid by
physiological reagents has been demonstrated with NAD(P)H (I 88, 189) and GSH (190—193)
and enzymatically by the thioredoxin/NADPH-thioredoxin reductase system (EC. 1.6.4.5)
(101). Dialuric acid oxidation generates superoxide, hydrogen peroxide, and hydroxyl
radical, in the presence of an appropriate metal catalyst like iron (194-197). Specifically,
reaction of alloxan with GSH generated superoxide and hydrogen peroxide (190-193),
hydroxyl radical in the presence of iron (197), and released iron from ferritin and
subsequently formed hydroxyl radical (I 98). Evidence supporting this mechanism includes
protection against alloxan toxicity in isolated islets by SOD, catalase, hydroxyl radical
scavengers, metal chelators (199-201), and in mice, in vivo, by SOD (202). If
thioltransferase does catalyze the GSH-dependent reduction of alloxan to dialuric acid, then

thioltransferase may play a role in the toxicity of alloxan.

47

Oxidized Forms:

Alloxan

Reduced Forms:

yr
H—N>:2:O

HO OH

Dialuric Acid

CHQOH
H OH O
O
H
O O

Dehydroascorbic Acid

CHZOH

O
H
HO OH

L-Ascorbic Acid

Figure 14. Structural comparison of alloxan to dehydroascorbic acid
and dialuric acid to L-ascobic acid

48

VI. Objectives of this Thesis.

The objectives of my work in this thesis are threefold. First, I wish to develop a
reproducible GSH-dependent DHA reductase activity assay and to kinetically characterize
and compare the rapidly growing class of mammalian DHA reductases. Second I seek to
characterize the catalytic mechanism of the GSH-dependent DHA reductase activity
associated with thioltransferase. This will provide a foundation for studies of other GSH-
dependent DHA reductases and will answer the enzyme structure/function question, does
thioltransferase have analo gous/ different catalytic mechanisms for its GSH-dependent DHA
reductase and thiol:disulfide oxidoreductase activities. Third I will determine whether other
non-disulfide substrates for thioltransferase exist, specifically alloxan, and if thioltransferase
is a GSH-dependent alloxan reductase, might thioltransferase contribute to the toxicity of

alloxan.

49

REFERENCES

(l) Sauberlich, HE. (1997) in Vitamin C in Health and Disease (Packer, L. and Fuchs, J.
Eds.) pp. 1-24, Marcel Dekker, Inc., New York, NY.

(2) Cotran, R.8., Kumar, V., and Robbins, S.L. (1994) Pathologic Basis of Disease, W.B.
Saunders Company, Philadelphia, PA.

(3) Phillips, CL. and Yeowell, H.N. (1997) in Vitamin C in Health and Disease (Packer, L.
and Fuchs, J. Eds.) pp. 205-230, Marcel Dekker, Inc., New York, NY.

(4) Berg, RA. and Prockop, DJ. (1973) Biochem. Biophys. Res. Commun. 52, 1 15-120.
(5) Eyre, DR. (1980) Science 207, 1315-1322.
(6) Englard, 8., and Seifier, S. (1986) Ann. Rev. Nutr. 6, 365-406.

(7) Grove, S.M., Lindberg, K.A., Reynolds, G., Sivarajah, A., and Pinnell, SR. (1981)
Proc. Natl. Acad. Sci. USA 78, 2879-2882.

(8) Lyons, BL, and Schwarz, R.I. (1984) Nucleic Acids Res. 12, 2569-2579.
(9) Hsuanyu, Y. and Dunford, HE. (1990) Arch Biochem Biophys, 281, 282-286.
(10) Levine, M., Morita, K., and Pollard, H. (1985) J. Biol. Chem. 260, 12942-12947.

(1 1) Bolscher, B.G., Zoutberg, G.R., Cuperus, RA, and Wever, R. (1984) Biochim Biophys
Acta 784, 189-191.

(12) Marquez, L.A., Dunford, H.B., and van Wart, H. (1990).]. Biol. Chem. 265, 5666—
5670.

(13) Wells, W.W., Dou, C.Z., Dybas, L.N., Jung, C.H., Kalbach, H.L., and Xu, DP (1995)
Proc. Natl. Acad. Sci. U.S.A. 92, l 1869-11873.

(l4) Dou, C., Xu, DP, and Wells, W.W. (1997) Biochem. Biophys. Res. Commun. 231,
820-822.

(15.) Jung, CH. and Wells, W.W. (1997) Biochem. Biophys. Res. Commun. 239, 457-462.

(16) Halliwell, B. and Whiteman, M. (1997) in Vitamin C in Health and Disease (Packer,
L. and Fuchs, J. Eds.) pp. 59-74, Marcel Dekker, Inc., New York, NY.

50

(17) Yu, B.P. (1994) Physiol. Rev. 74, 139-162.
(18) Buettner, GR. (1993) Arch. Biochem. Biophys. 300, 535-543.

(19) Packer, L. (1997) in Vitamin C in Health and Disease (Packer, L. and Fuchs, J. Eds.)
pp. 95-108, Marcel Dekker, Inc., New York, NY.

(20) Kagan, .V.E., Yalowich, J.C., Day, B.W., Goldman, R., Gantchev, T.G., and
Stoyanovsky, D.A. (1994) Biochemistry 33 , 9651-9660.

(21) Nishikimi, M. (1975) Biochem. Biophys. Res. Commun. 63, 463-468.

(22) Chakrabarty, 8., Nandi, A., Mukhopadhyay, GK, and Chatterjee, I.B.( l 992) Mol. Cell.
Biol. 3, 41-47.

(23) Frei, B., Forte, T.M., Ames, B.N., and Cross, CE. (1991) Biochem. J. 277, 133-138.

(24) Frei, B.F., Stocker, R., and Ames, B.N. (1988) Proc. Natl. Acad. Sci. USA 85, 9748-
9752 .

(25) Frei, B., England, L., and Ames, B.N. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 6377-
6381.

(26) Sciamanna, M.A., and Lee, CF. (1993) Arch. Biochem. Biophys. 305, 215-224.

(27) Ozaki, M., Fuchinoue, 8., Teraoka, 8., and Ota, K. (1995) Arch. Biochem. Biophys.
318, 439-445.

(28) Miller, D.M. and Aust, SD. (1989) Arch. Biochem. Biophys. 271, 113-1 19.
(29) Packer, J.E., Slater, TE, and Wilson, R.L. (1979) Nature 278, 737-738.

(30) Doba, T., Burton, G.W., and Ingold, K.U. (1985) Biochim. Biophys. Acta. 835, 298-
303.

(31) Wefers, H. and Sies, H. (1988) Eur. J. Biochem. 174, 353-357.
(32) Halliwell, B., and Gutteridge, J.M.C. (1990) Arch. Biochem. Biophys. 280, 1-8.
(33) Halliwell, B., Wasil, M.., and Grootveld, M. (1987) FEBS Lett. 213, 15-18.

(34) Hu, M.L., Louie, 8., Cross, C.E., Motchink, P., and Halliwell, B. (993) J. Lab. Clin.
Med. 121, 257-262.

(35) Babior, BM. (1978) N. Engl. J. Med. 298, 659-668.

51

(36) Pietri, 8., Culcasi, M., Stella, L., and Cozzone, P. J. (1990) Eur. J. Biochem., 193, 845-
854.

(37) Pietri, 8., Séguin, J .R., Arbigny, P., and Culcasi, M. (1994) Free Rad. Biol. Med. 16,
523-528.

(38) Mehlhom, R.J. (1991).]. Biol. Chem. 206, 2724-2731.
(39) Zhang, Y. and Fung, W.M. (1994) Free Rad. Biol. Med. 16, 215-222.
(40) May, J .M., Qu, Z.C., and Whitesell, RR. (1995) Biochemistry 34, 12721-12728.

(41) Christine, L., Thomson, G., Iggo, B., Brownie, AC, and Stewart, GP. (1956) Clin.
Chim. Acta 1, 557-564.

(42) Hughes, RE. and Kilpatrick, GS. (1964) Clin. Chim. Acta 9, 241-244.

(43) Basu, 8., Som, 8., Deb, 8., Mukherjee, D., and Chatterjee, LB. (1979) Biochem.
Biophys. Res. Commun. 90, 1335-1340.

(44) Winterbourn, CC. and Vissers, M.C.M. (1983) Biochim. Biophys. Acta 763, 175-179.
(45) Hemiléi, H., Roberts, R, and Wikstréim, M. (1984) FEBS Lett. 178, 25-30.
(46) Washko, P.W., Wang, Y., and Levine, M. (1993) J. Biol. Chem. 268, 15531-15535.

(47) Welch, R.W., Wang, Y., Crossman, A., Park, J.B., Kirk, KL, and Levine, M. (1995)
J. Biol. Chem. 270, 12584-12592.

(48) Wang, Y., Russo, T.A., Kwon, 0., Chanock, 8., Rumsey, SC, and Levine, M. (1997)
Proc. Natl. Acad. Sci. USA 94, 13816-13819.

(49) Griffith, O.W. and Meister, A. (1979) J. Biol. Chem. 254, 7558-7600.
(50) Meister, A. (1994) J. Biol. Chem. 269, 9397-9400.
(51) Meister, A. (1992) Biochem. Pharm. 44, 1905-1915.

(52) Ménensson, J., Jain, A., Stole, E., Prayer, W., Auld, P.A.M., and Meister, A. (1991)
Proc. Natl. Acad. Sci. USA 88, 9360-9364.

(53) Jain, A., Mirrtensson, J ., Mehta, T., Krauss, A.N., Auld, P.A.M., and Meister, A. (1992)
Proc. Natl. Acad. Sci. USA 89, 5093-5097.

(54) Tsao, CS. (1997) in Vitamin C in Health and Disease (Packer, L. and Fuchs, J. Eds.)

52

pp. 25-58, Marcel Dekker, Inc., New York, NY.

(55) Coassin, M., Tomasi, A., Vannini, V., and Ursini, F. (1991) Arch. Biochem. Biophys.
290, 458-462.

(56) Villalba, J .M., Canalejo, A., Buron, M.I., Cordoba, F., and Navas, P. (1993) Biochem.
Biophys. Res. Commun. 192, 707-713.

(57) Bando, M., and Obazawa, H. (1994) Jpn. J. Ophthalmol. 38, 1-9.
(58) Patterson, J.W. (1950).]. Biol. Chem. 183, 81-88.

(59) Pillsbury, 8., Watkins, D., and Cooperstein, SJ. (1973) J. Pharmacol. Exp. T her. 185,
713-718.

(60) Bianchi, J., and Rose, RC. (1986) Toxicology 40, 75-82.

(61) May, J .M., Qu, Z.C., Whitesell, RR, and Cobb, CE. (1996) Free Rad. Biol. Med. 20,
543-551.

(62) Del Bello, B., Maellaro, E., Sugherini, L., Santucci, A., Comporti, M., and Casini, AF.
(1994) Biochem J. 304, 385-390.

(63) May, J .M., Mendiratta, 8., Hill, KB, and Burk, RF (1997).]. Biol. Chem. 272, 22607-
22610.

(64) Mrtensson, J., and Meister, A. (1991) Proc. Natl. Acad. Sci. USA 88, 4656-4660.

(65) Mirtensson, J ., Han, J ., Griffith, O.W., and Meister, A.(1993) Proc. Natl. Acad. Sci.
USA 90, 317-321.

(66) Miirtensson, J., and Meister, A. (1992) Proc. Natl. Acad. Sci. USA 89, 11566-11568.

(67) Braun, L., Csala, M., Poussu, A., Garzé, T., Maral, J., and Banhegyi, G. (1996) FEBS
Lett. 388, 173-176.

(68) Crook, EM. (1941) Ph. D. Dissertation, Cambridge.

(69) Ball, M. (1937) J. Biol. Chem. 118, 219-239.

(70) Hughes, RE. (1964) Nature 203, 1068-1069.

(71) Yamarnoto, Y., Sato, M., Ikeda, S. (1977) Bull. Jap. Soc. Sci. Fish. 43, 53-57.

(72) Yarnamoto, Y., Sato, M., Ikeda, S. (1977) Bull. Jap. Soc. Sci. Fish. 43, 59-62.

53

(73) Bigley, R., Riddle, M., Layman, D., and Stankova, L. (1981) Biochim. Biophys. Acta.
659, 15-22.

(74) Ahn, B.Y., and Moss, B. (1992) Proc. Natl. Acad. Sci. USA 89, 70600-7064.

(75) Wells, W.W., Xu, D.P., Yang, Y., and Rocque, RA. (1990).]. Biol. Chem. 265, 15361-
15364.

(76) Wells, W.W., Xu, DR, and Washbum, MP. (1995) Methods. Enzymol. 252, 30-38.
(77) Park, J.B., and Levine, M. (1996) Biochem. J. 315, 931-938.

(78) Maellaro, E., Del Bello, B., Sugherini, L., Santucci, A., Comporti, M., and Casini, AF.
(1994) Biochem J. 301, 471-476.

(79) Xu, D.P., Washbum, M.P., Sun, GP, and Wells, W.W. (1996) Biochem. Biophys. Res.
Commun. 221, 117-121.

(80) Moutiez, M., Que'me'neur, E., Sergheraert, C., Valerie, L., Tartar, A., and Davioud-
Charvet, E. (1997) Biochem. J., 322, 34-48.

(81) Triimper, 8., Follmann, H., and Hiiberlein, I. (1994) FEBS Lett. 352, 159-162.
(82) Hossain, M.A., and Asada, K. (1984) Plant Cell Physiol. 25, 85-92.

(83) Paolicchi, A., Pezzini, A., Saviozzi, M., Piaggi, 8., Andrevccetti, M., Chieli, E.,
Malvadi, G., and Casini, AF. (1996) Arch. Biochem. Biophys. 333, 489-495.

(84) Mieyal, J .J ., Gravina, S.A., Mieyal, P.A., Srinivasan, U., and Starke, D.W. (1995) in
Biothiols in Health and Disease (Packer, L., and Cadenas, E., Eds.) pp. 305-372,
Marcel Decker, Inc., New York, NY.

(85) Chivers, P.T., Prehoda, K.E., and Raines, RT. (1997) Biochemistry 36, 4061-4066.

(86) Freedman, R.B., Hirst, T.H., and Tuite, M.F. (1994) Trends Biochem. Sci. 19, 331-336.

(87) Loferer, H and Hennecke, H. (1994) Trends Biochem. Sci. 19, 169-171.

(88) Holmgren, A. (1985) Annu. Rev. Biochem. 54, 237-271.

(89) Siedler, F., Rudoplh-Bdhner, 8., Doi, M., Music], H.J., and Moroder, L. (1993)
Biochemistry 32, 7488-7495.

(90) Chivers, P.T., Prehoda, K.E., and Raines, RT. (1997) Biochemistry 36, 4061-4066.

54

(91)Wunder1ich, M, and Glockshuben, R(l993) Prot. Sci. 2, 717-726.
(92) Nelson, J .W., Creighton, TE. (1994) Biochemistry 33, 5974-5983.
(93) Freedman, RB. (1989) Cell 57, 1069-1072.

(94) Noiva, R., Kimura, H., Roos, J. and Lennar'z, W.J. (1991).]. Biol. Chem. 266, 19645-
19649.

(95) Hawkins, H.C., DeNardi, M., and Freedman, RB. (1991) Biochem. J. 275, 341-348.

(96) Edman, J.C., Ellis, L., Blacher, R.W., Roth, RA, and Rutter, W.J. (1985) Nature 31 7,
267-270.

(97) Hawkins, HQ and Freedman, RB. (1991) Biochem. J 275, 335-339.
(98) Walker, K.W., Lyles, M.M., and Gilbert, H.F. (1996)Biochemist1y 35, 1972-1980.

(99) Martha, C.A., Laboissiére, C.A., Sturley, S.L., and Raines, RT. (1995).]. Biol. Chem.
270, 28006-28009.

(100) Eklund, H., Gleason, F.K., and Homgren, A. (1991) Proteinssstructure,function, and
genetics 1], 13-28.

(101) Holmgren, A., and Lyckeborg, C. (1980) Proc. Natl. Acad. Sci. USA 77, 5149-5152.
(102) Nikitovic, D. and Holmgren, A. (1996).]. Biol. Chem. 271, 19180-19185.

(103) Biguet, C., Wakasugi, N., Mishal, A., Holmgren, A., Chovaib, 8., Tursz, T., and
Wakasugi, C. (1994).]. Biol. Chem. 269, 28865-28870).

(104) Jacquier-Sarlin, MR. and Polla, 8.8. (1996) Biochem. J. 318, 187-193.

(105) Schenk, H., Klein, M., Erdbiigger, W., Drdge, W., and Schulze-Osthoff, K. (1994)
Proc. Natl. Acad. Sci. USA 91, 1672-1676.

(106) Chivers, P.T., Prehoda, K.E., Volkrnan, B.F., Kim, B.-M., Markley, J .L., and Raines,
RT. (1997) Biochemistry 3 6, 14985-14991.

(107) LeMaster, D.M. (1996) Biochemistry 35, 14876-14881.
(108) Chivers, RT. and Raines, RT. (1997) Biochemistry 3 6, 15810-15816.

(109) Askeliif, P., Axelsson, K., Eriksson, 8., and Mannervik, B. (1974) FEBS Lett. 3 8, 263-
267.

55

(110) Racker, E. (1955).]. Biol. Chem. 217, 867-874.

(111) Nagai, S. and Black, 8. (1968).]. Biol. Chem. 243, 1942-1947.

(112) Axelsson, K., Eriksson, 8., and Mannervik, B. (1978) Biochemistry 1 7, 2978-2984.
(113) Holmgren, A. (1976) Proc. Natl. Acad. Sci. USA 73, 2275-2279.

(114) Holmgren, A. (1979).]. Biol. Chem. 254, 3664-3671.

(115) Holmgren A. (1979).]. Biol. Chem. 254, 3672-3678.

(116) Luthman, M., Eriksson, 8., Holmgren, A., and thelander, L. (1979) Proc. Natl. Acad.
Sci. USA 76, 2158-2162.

(117) Rajagopal, I., Ahn, B.Y., Moss, 8., and Mathews, CK. (1995) J. Biol. Chem. 270,
27415-27418.

(1 18) Gan, Z.R. and Wells, W.W. (1987).]. Biol. Chem. 262, 6699-6703.

(1 19) Papayannopoulos, I.A., Gan, Z.R., Wells, W.W., and Biemann, K. (1989) Biochem.
Biophys. Res. Commun. 159, 1448-1454.

(120) Wells, W.W., Yang, Y., Deits, TL, and Gan, Z. (1993) Adv. Enzymol. Relat. Areas
Mol. Biol. 66, 149-201.

(121) Papov, V.V., Gravina, S.A., Mieyal, J .J ., and Biemann, K. (1994) Prat. Sci. 3, 428-
434.

(122) Padilla, C.A., Martinez-Galisteo, E., Barcena, I.A., Spyrou, G., and Holmgren, A.
(1995) Eur. J. Biochem. 22 7, 27-34.

(123) Hopper, 8., Johnson, R.S., Vath, J.E., and Biemann, K. (1989).]. Biol. Chem. 264,
20438-20477.

(124) Sha, 8., Minatuchi, K., Kisaki, N., Sato, K., Ohtsuki, K., Kurata, A., Yoshikawa, H.,
Kotaru, M., Masumura, T., Ichihara, K., and Tanaka, K. (1997).]. Biochem. (Tokyo)
121, 842-848.

(125) Gan, Z.R., Polokoff, M.A., Jacobs, J.W., and Sardana, MK. (1990) Biochem.
Biophys. Res. Commun. 168, 944-951.

(126) Hodg, J.O., Jdmvall, H., Holmgren, A., Carlquist, M., and Persson, M. (1983) Eur.
J. Biochem. 136, 223-232.

56

(127) Johnson, G.P., Goebel, S.J., Perkus, ME, Davis, 8.W., Winslow, J .P., and Paoletti,
E. (1991) Virology 181, 378-381.

(128) deberg, BM. and Holmgren, A. (1972) J. Biol. Chem. 247, 8063-8068.

(129) Fernando, M.R., Sumimoto, H., Nanri, H., Kawabata, 8.1., Iwanaga, 8., Minakami,
8., Fukumaki, Y., Takehige, K. (1994) Biochim. Biophys. Acta 1218, 229-231.

(130) Meyer, EB and Wells (1995) FASEB J. 9, A/1195.
(131) Yang, Y., Gan, Z.R., and Wells, W.W. (1989) Gene 83, 339-346.

(132) Minakuchi, K., Yabushita, T., Masumura, T., Ichihara, K., and Tanaka, K. (1994)
FEBS Lett. 337, 152-160.

(133) H66g, J.O., van Bahr-Lindstrdm, H., Jdmvall, H., and Holmgren, A. (1986) Gene 43,
13-21.

(134) Séderberg, B.-O., Holmgem, A., Branden, C.-I. (1974).]. Mol. Biol. 90, 143-152.
(135) Katti, D.K., LeMaster, D.M., and Eklund, H. (1990).]. Mol. Biol. 212, 167-184.
(136) Kortemme, T. and Creighton, T.E. (1995) J. Mol. Biol. 253, 799-812.

(137) Martin, J.L., Bardwell, J.C.A., and Juriyan, J. (1993) Nature 365, 464-468.

(138) Kemmink, J., Darby, N.J., Dijkstra, K., Nilges, M., and Creighton, T.E. (1996)
Biochemistry 35, 7684-7691.

(139) Dyson, H.J., Gippert, G.P., Case, D.A,. Holmgren, A., and Wright, RE. (1990)
Biochemistry 29, 4129-4136.

(140) Katti, S.K., Robbins, A.H., Yang, Y., and Wells, W.W. (1995) Prot. Sci., 4, 1998-
2005.

(141) Xia, T.-H., Bushweller, J.H., Sodano, P., Billeter, M., Bjdmberg, 0., Holmgem, A.,
and Wiirthrich, K. (1992) Prot. Sci. 1, 310-321.

(142) Sodano, P., Xia, T.H., Bushweller, J .H., Bjdmberg, 0., Holmgem, A., Billeter, M, and
Wiirthrich, K. (1991).]. Mol. Biol. 221, 1311-1324.

(143) Bushweller, J.H., Billeter, M., Holmgren, A., and Wiirthrich, K. (1994).]. Mol. Biol.
235, 1585-1597.

(144) Eklund H., Ingelman, M., deerberg, B.O., Uhlin, T., Nordlund, P., Nikkola, M.,

57

Sonnersam, U., Joelson, T., and Petratos, K. (1992).]. Mol. Biol. 228, 596-618.

(145) Gleason, F.K., Lim, C.-J., Gerami-Nejad, M. , and Fuchs, J .A. (1990) Biochemistry,
29, 3701-3709.

(146) Nikkola, M., Gleason, F.K., Saarinen, M., Joelson, T., Bjdmberg, 0., and Eklund, H.
(1991).]. Bio.lChem.,266,16105-16112.

(147) Rabenstein, D.L. and Mills, K.K. (1995) Biochim. Biophys. Acta I 249, 29-36.
(148) Srinivasan, U., Mieyal, PA, and Mieyal, J.J. (1997) Biochemistry 36, 3199-3206.
(149) Jung, CH, and Thomas, J.A. (1996) Arch. Biochem. Biophys. 335, 61-72.

(150) Aslund, F ., Berndt, K.D., Holmgren, A. (1997).]. Biol. Chem. 272, 30780-30786.
(151) Yang, Y., and Wells, W.W. (1991).]. Biol. Chem. 266, 12759-12765.

(152) Mieyal, J. J., Starke, D. W., Gravina, 8. A., and Hocevar, B. A.( 1991)
Biochemistry 30, 8883-8891.

(153) Gan, Z.R., Sardana, M.K., Jacobs, J.W., and Polokoff, MA. (1990) Arch. Biochem.
Biophys. 282, 110-115.

(154) Collison, M.W., Beidler, D., Grimm, L.M., and Thomas, J.A. (1986) Biochim.
Biophys. Acta 885, 58-67.

(155) Collison, M.W. and Thomas, J .A. (1987) Biochim. Biophys. Acta 928, 121-129.
(156) Rokutan, K., Thomas, J.A., and Sies, H. (1989) Eur. J. Biochem. 179, 233-239.

(157) Miller, R.M., Sies, H., Park, E.M., and Thomas, J.A. (1990) Arch. Biochem. Biophys.
276, 355-363.

(158) Simplicio, P. and Rossi, R. (1994) Biochim. Biophys. Acta 1199, 245-252.
(159) Hiranruengchok, R. and Harris, C. (1995) T eratology 52, 196-204.
(160) Park, EM. and Thomas, J.A. (1988) Biochim. Biophys. Acta 964, 151-160.

(161) Chai, Y.C., Jung, C.H., Lii, C.K., Ashraf, 8.8., Hendrich, 8., Wolf, B., Sies, H., and
Thomas, J.A. (1991) Arch. Biochem. Biophys. 284, 270-278.

(162) Gilbert, HF. (1984) Meth. Enzymol.]07, 330-351.

58

(163) Ziegler, D.M. (1985) Ann. Rev. Biochem. 54, 305-29.
(164) Mannervik, B. and Axelsson, K. (1980) Biochem. .1. 190, 125-130.
(165) Axelsson, K., and Mannervik, B. (1983) FEBS Lett. 152, 114-118.

(166) Flamigni, F., Manniroli, S., Caldarera, C.M., and Guarnieri, C. (1989) Biochem. J.
259,111-115.

(167) Shen, H., Tamai, K., Satoh, K., Hatayama, 1., Tsuchida, 8., and Sato, K. (1991)Arch.
Biochem. Biophys. 286, 178-182.

(168) Terada, T., Maeda, H, Okamoto, K., Nishinaka, T., Mizoguchi, T., Nishihara, T.
(1993) Arch. Biochem. Biophys. 300, 495-500.

(169) Terada, T., Oshida, T., Nishimura, M., Maeda, H., Hara, T., Hosomi, 8., Mizoguchi,
T., Nishihara, T. (1992).]. Biochem. (Tokyo) 111, 688-692.

(170) Terada, T., Hara, T., Yazawa, H., and Mizoguchi, T. (1994) Biochem. Mol. Biol. Int.
32, 239-244.

(171) Zheng, M., Aslund, F., and Storz, G. (1998) Science 279, 1718-1721.
(172) Tao, K. (1997).]. Bacteriol. 179, 5967-5970.
(173) Gravina, 8A., and Mieyal, J .J . (1993) Biochemistry 32, 3368-3376.

(174) Davis, D.A., Newcomb, F .M., Starke, D.W., Ott, D.E., Mieyal, J .J ., and Yarchoan, R.
(1997).]. Biol. Chem. 272, 25935-25940.

(175) Bandyopadhyay, 8., Starke, D.W., Mieyal, J .J ., and Gronostajski, R.M. (1998).].Biol.
Chem. 273, 392-397.

(176) Lundstrdm-Ljung, J. and Holmgren, A. (1995) J. Biol. Chem. 270, 7822-7828.

(177) Ruoppolo, M., Lundstréim-Ljung, J., Talamo, F., pucci, P., and Marino, G. (1997)
Biochemistry 36, 12259-12267.

(178) Yoshitake, S., Nanri, H., Rohan, Fernando, M.R., and Minakami, S. (1994) J.
Biochem. (Tokyo) 116, 42-46).

(179) Yang, Y., and Wells, W.W. (1991).]. Biol. Chem. 266, 12766-12771.

(180) Morell, 8., Follman, H., and Hilberlein, I. (1995) FEBS Lett. 369, 149-152.

59

(181) Rerup, C. C. (1970) Pharmacol Rev 22, 485-518.
(182) Oberley, L. W. (1988) Free Radical Biol Med 5, 1 13-124.
(183) Scarpelli, D. G. (1989) T oxicol Appl Pharmacol 101, 543-554.

(184) Weaver, D. C., McDaniel, M. L., and Lacy, P. E. (1978) Endocrinology 102, 1847-
1855.

(185) Malaisse, W. J., Malaisse-Lagae, F., Sener, A., and Pipeleers, D. G. (1982) Proc.
Natl. Acad. Sci. USA 79, 927-930.

(186) Hammarstrbm, L., Hellman, B., and Ullberg, 8. (1966) Diabetalogia 2, 340-345.
(187) Grankvist, K., Marklund, S. L., and Téiljedal, I.-B. (1981) Biochem. J. 199, 393-398.
(188) Miwa, 1., and Okuda, J. (1982) Biochem. Pharmacol. 31, 921-925.

(189) Nukatsuka, M., Sakurai, H., and Kawada, J. (1989) Biochem. Biophys. Res.
Commun. 165, 278-283.

(190) Winterboum, C. C., and Munday, R. (1989) Biochem. Pharmacol. 38, 271-277.
(191) Lenzen, 8., and Munday, R. (1991) Biochem. Pharmacol. 42, 1385-1391.

(192) Sakurai, K., and Ogiso, T. (1991) Chem. Pharm. Bull. (Tokyo) 39, 737-742.
(193) Munday, R., Ludwig, K., and Lenzen, S. (1993).]. Endacrinol. 139, 153-163.

(194) Deamer, D. W., Heikkila, R. E., Panganamala, R. V., Cohen, G., and Comwell, D.
G.(1971) Physiol. Chem. Phys. Med. NMR. 3, 426-430.

(195) Cohen, 0., and Heikkila, R. E. (1974).]. Biol. Chem. 249, 2447-2452.
(196) Munday, R. (1988) Biochem. Pharmacol. 3 7, 409-413.

(197) Sakurai, K., Miura, T., and Ogiso, T. (1990) Chem. Pharm. Bull. (Tokyo) 38, 993-
997.

(198) Sakurai, K., and Miura, T. (1988) Chem. Pharm. Bull. (Tokyo) 36, 4534-4538.

(199) Grankvist, K., Marklund, 8., Sehlin, J ., and Taljedal, I.-B. (1979). Biochem. J. 182,
17-25.

(200) Fischer, L. J., and Hamburger, S. A. (1980) Diabetes 29, 213-216.

60

(201) Uchigata, Y., Yamamoto, H., Kawamura, A., and Okamoto, H. (1982).]. Biol. Chem.
257, 6084-6088.

(202) Grankvist, K., Marklund, 8., and Tiiljedal, I.-B. (1981) Nature 294, 158-160.

(203) Jocelyn, PC. (1972), Biochemistry of the SH Group, Academic Press, New York,
NY.

(204) Missiakas, D., Georgopoulus, C., and Raina, S. (1993) Proc. Natl. Acad. Sci. USA 90,
7084-7088.

(205) Frishman, D. (1996) Biochem. Biophys. Res. Commun. 219, 686-689.

61

Chapter II

Kinetic Characterization and Comparison of Mammalian Dehydroascorbate

Reductases

This chapter is the combination of the contributions made to the following papers:
(1) Wells, W.W., Xu, DR, and Washbum, MP. (1995) Methods. Enzymol. 252, 30-38 and
(2) Xu, D.P., Washbum, M.P., Sun, GP, and Wells, W.W. (1996) Biochem. Biophys. Res.
Commun. 221, 117-121.

62

ABSTRACT

Several glutathione (G8H)-dependent dehydroascorbate (DHA) reductases (EC
1.8.5.1) were purified to homogeneity and their DHA reductase activities kinetically
characterized. Recombinant pig liver thioltransferase (RPLTT) was the most robust DHA
reductase at pH 6.9 with a kW/Km for DHA of 2.43 :1: 0.85 x 104 M" sec" followed by human
erythrocyte 32 kDa DHA reductase with a kca/Km for DHA of 1.49 i 0.21 x 104 M" sec",
with bovine liver protein disulfide isomerase the weakest DHA reductase with a km/Km for
DHA of 93 :l: 14 M" sec". Human erythrocyte 32kDa DHA reductase and RPLTT were
equally robust DHA reductases at pH 7.2 with a kw/Km for DHA of 2.47 i 0.64 x 10" M"
sec’I and a kw/Km for DHA of 2.01 d: 0.13 x 104 M" sec". The DHA reductase activities of
these enzymes and several others from mammals are compared and contrasted. Based on
kinetic comparisons, the most efficient and most likely to play an important in viva role in

DHA reduction are thioltransferase and the 32kDa GSH-dependent DHA reductase.

63

Ascorbic acid (AA) is an important antioxidant (1-6) and electron donor for enzymes
like dopamine-B-hydroxylase, prolyl-4-hydroxylase, and 1ysy1hydroxylase(7). The function
of AA as an electron donor is highlighted by its importance in collagen biosynthesis (8) and
norepinephrine biosynthesis (9). In addition, AA has recently been demonstrated to be
essential for the release of insulin from the pancreatic islets of scorbutic guinea pigs (10) and
a cofactor for mitochondrial glycerol-3-phosphate dehydrogenase (11).

When AA carries out its function it undergoes two successive one electron oxidations
to semidehydroascorbic acid (semiDHA) and dehydroascorbic acid (DHA). A relatively
stable radical, semiDHA may rapidly disproportionate to AA and DHA (12) or be directly
reduced to AA by an NADH-semiDHA reductase (13-15). Currently, NADH-semiDHA
reductase has not been purified to homogeneity nor kinetically characterized.

DHA is toxic to cells and tissues as highlighted by its in vitro toxicity to pancreatic
islets (16-1 7) and erythrocytes (18). DHA is taken up preferentially by cells like human
erythrocytes (19), neutrophils (20), and fibroblasts (20) and rapidly reduced to AA. May et
al (19) determined that DHA reduction by human erythrocytes does not involve
semidehydroascorbate. Therefore, recycling of DHA is important especially in species like
humans whose only source of AA is from the diet.

DHA can be reduced to AA in an NADPH dependent manner by both 301-
hydroxysteroid dehydrogenase (21) and thioredoxin reductase (22). However, Glutathione
(GSH) deficiency in newborn rats, which compared with adults synthesize AA at a reduced
rate, caused drastic ascorbate decreases in tissues highlighting the importance of GSH in the
recycling of AA (23). GSH chemically reduces DHA to AA, especially at alkaline pH, and

three mammalian GSH-dependent DHA reductases have been identified: thioltransferase

64

(glutaredoxin, 24), protein disulfide isomerase (PDI, 24), and a 32kDa enzyme (32kDHAR)
identified in rat liver (25) and human erythrocytes (26). Whether DHA reduction in vivo is
primarily chemical, enzymatic, or both remains unclear with evidence existing for both
mechanisms (2 7-28). However, GSH reduction of DHA to AA has been estimated to have
a tm of 15 min at pH 7.5 and at 25°C (29), but under the same conditions the degradation of
DHA to non recyclable products like 2,3-diketogulonic acid has a tm of 2 minutes (30)
strongly supporting the importance of catalytic regeneration of AA from DHA. In addition,
recent work by Park & Levine (31) provided evidence for enzymatic recycling of DHA in
human neutrophils, catalyzed in large part by thioltransferase. In this chapter, the GSH-
dependent DHA reductase activities of several mammalian enzymes are kinetically
characterized and compared.
MATERIALS AND METHODS

Materials. EDTA, glutathione disulfide (GSSG), bovine pancreatic insulin, and
phenylmethylsulfonyl fluoride (PMSF) were purchased from Sigma, St. Louis, MO. GSH,
dithiothreitol (DTT), glutathione disulfide reductase, and NADPH were products of
Boehringer Mannheim, Indianapolis, IN. Bromine was purchased from Acros Organics,
Pittsburgh, PA. Hydroxyethyldisulfide (HED) was a product of Aldrich, Milwaukee, WI.
Isopropylthiogalactoside (IPTG) was acquired from Calbiochem, San Diego, CA. CM-
Sepharose, CM-Sephadex C50, DEAE-Sephacel, Sephacryl 8-200, and Sephadex G75 were
products of Amersham Pharmacia Biotech, Piscataway, NJ. Sodium phosphate was
purchased from EM Science, Gibbstown, NJ. Triton X-100 was a product of Research
Products International, Elk Grove Village, IL. S-sulfocysteine was prepared by the method

of Segel and Johnson (32). DHA was prepared as described previously (33). Human

65

erythrocyte 32 kDA GSH-dependent DHA reductase was purified by Dian Peng Xu, as
described (26).

Purification of Pig Liver T hialtransferase. Expression of the thioltransferase gene
and purification of RPLTT was carried out as described previously (34) with the following
modifications. After ammonium sulfate precipitation, the pellets were resuspended in 2 ml
of 20 mM sodium phosphate, pH 6.5 (buffer A), containing 2 mM DTT, and dialyzed
overnight (2 x 2L) against the same buffer. After dialysis, the sample was loaded onto a
Sephacryl 8-200 column ( 2.5 cm x 89 cm) equilibrated and eluted with buffer A containing
2 mM DTT. The pooled active fractions were incubated with 10 mM HED for 30 min at
room temperature and loaded onto a CM-Sepharose column (5 cm x 38 cm) equilibrated with
buffer A. The column was then washed with l L of buffer A followed by 1 L of 20 mM
sodium phosphate, pH 7.5 (buffer B). The bound disulfide enzyme was eluted with an 800
m1 0.0-1.0 M NaCl linear gradient in buffer B. The active fractions were pooled and
concentrated using Centriprep 3 concentrators and stored at '70°C until use. Protein
concentrations were determined by the bicinchoninic acid protein assay (Pierce Chenrical
Company) according to the manufacturer’s directions.

Assay of thioltransferase activity. During the purification, the enzyme activity was
assayed as described previously (34). Samples were assayed in 0.318 mM sodium
phosphate, 1.0 mM EDTA, 0.5 mM GSH, 0.34 mM NADPH, 5 ug GSSG reductase, and 2.5
mM S-sulfocysteine in a final volume of 500 )1]. Upon the addition of S-sulfocysteine , the
mixture was assayed at 340 nm for 3 min at 30°C. A control without sample was run
simultaneously for each assay and the difference ((AADS 340nm/min) x 0.5 x 1000/6.22)) gave

activity in nmoles/min.

66

Purification of Bovine Liver Protein Disulfide Isomerase. Bovine liver PDI was
purified through a modification of the procedure described by Hilson et a] (35). Frozen
bovine liver was thawed and washed in 0.9% w/v NaCl. It was homogenized in a Waring
blender in l L of 100 mM sodium phosphate, 5 mM EDTA, 1% Triton X-100, 0.5 mM
PMSF, pH 7.50. After straining of the homogenate through two layers of muslin, the sample
was centrifuged at 18,000 x g for 30 min. The supernatant was decanted through glass wool
and heated to 58°C for 15 min in a 70°C water bath. The supernatant was transferred to an
ice bath to cool and centrifuged at 18,000 x g for 40 min. Ammonium sulfate was added to
the supernatant from the heat denaturation to reach a final concentration of 55% saturation.
After stirring for 30 min the sample was centrifuged at 18,000 x g for 30 min and ammonium
sulfate was further added to the supernatant to reach 85% saturation. After stirring for 30
min the sample was centrifuged at 18,000 x g. The pellet was resuspended in a minimal
volume of 25 mM citrate, pH 5.3, and dialyzed overnight against the same buffer (2 x 5L).

Dialyzed extract was applied to a Sephadex G75 column equilibrated with 25 mM
citrate, pH 5.3, and eluted with the same buffer. Active fractions were pooled and applied
to a CM-Sephadex C50 column equilibrated with 25 mM citrate, pH 5.3 and eluted with the
same buffer. Active fractions were pooled and ammonium sulfate added to reach a final
concentration of 90%. After stirring for 30 min the sample was centrifuged at 18,000 x g for
30 min. The pellet was resuspended in a minimal volume of 25 mM sodium phosphate, 1
mM EDTA, 0.1 M NaCl, pH 8.0 and dialyzed 2 x 2L overnight against the same buffer.

Dialyzed sample was loaded onto a DEAE-Sephacel column equilibrated with 25 mM
sodium phosphate, 1 mM EDTA, 0.1 M NaCl, pH 8.0. After loading, the column was

washed with two volumes of 25 mM sodium phosphate, 1 mM EDTA, 0.1 M NaCl, pH 8.0.

67

PDI was eluted with 0.1-0.5 M NaCl gradient in 25 mM sodium phosphate, 1 mM EDTA,
pH 8.0. Active fractions were concentrated in Centriprep 10 concentrators, diluted with 25
mM sodium phosphate, 2.5 mM EDTA, pH 7.50, and concentrated again.

Assay of Protein Disulfide Activity. Bovine liver PDI activity was assayed by the
insulin reduction assay of Holmgren (36). Sample was assayed in 100 mM potassium
phosphate, 2 mM EDTA, pH 7.0, lmM DTT, and 0.06 mM insulin in a final volume of 500
ul. Upon the addition of insulin the sample was assayed at 650 nm at 30°C for 30 min. A
control without sample was run simultaneously for each assay. An increase in absorbance
at 650 nm was indicative of activity.

Dehydroascorbic Acid Reductase Assay and Enzyme Kinetics. DHA reductase
activity was assayed as described previously (33). Initial reaction velocities were compared
at various pH values using 200 mM sodium phosphate, 1 mM EDTA buffers between pH 5.5
and 7.5. Assays performed to determine the kinetic constants for RPLTT ( 0.72 ug), bovine
liver PDI (69ug), and human erythrocyte 32kDa-DHA reductase (1.85 ug for DHA profile
and 1.575 ug for GSH profile) were run in 200 mM sodium phosphate, 1 mM EDTA, pH
6.9, in a final volume of 500 ul at 30°C for 3 min and recorded at 265.5 nm. Assays
performed to determine the kinetic constants for RPLTT (1 u g) and human erythrocyte
32kDa DHA reductase (2.3 ug) were also run in 100 mM sodium phosphate, pH 7.2, in a
final volume of 500 ul at 30°C for 3 min and recorded at 265.5 run. A blank without enzyme
was run simultaneously with each assay and the difference gave the activity in nmoles/min.
Kinetic constants for DHA were determined by varying [DHA] and holding [GSH] = 3.0
mM, and kinetic constants for GSH were determined by varying [GSH] and holding [DHA]

= 1.5 mM. Km(app) and Vm,,(app) values were calculated by nonlinear least-squares fitting

68

to the velocity versus substrate concentration data using the PSI-Plot 3.5 software. The

values for k were calculated by dividing Vmax(app) by the molar concentration of the

C31

enzymes.

RESULTS

DHA reductase activity of bovine liver PDI, RPLTT, and human erythrocyte 32kDa
DHA reductase. The kinetic plots for the DHA reductase activity of bovine liver PDI at pH
6.9 (Figure 1A and IE), RPLTT at pH 6.9 (Figure 2A and 2B) and 7.2 (Figure 3A and 3B),
and human erythrocyte 32 kDa DHA reductase at pH 6.9 (Figure 4A and 4B) and at pH 7.2
(Figure 5A and SB) best fit the Michaelis-Menten equation (EZ-Fit Kinetic Program, 37).
The results of these kinetic analyses are found in Table 1. Although L-cysteine and
cysteamine chemically reduce DHA to AA, neither bovine liver PDI, RPLTT, or human
erythrocyte 32 kDa DHA reductase was able to catalyze this reaction demonstrating a
specificity of GSH.

DISCUSSION

The results in this chapter contribute to the growing body of literature concerning
DHA reductases. A large munber of DHA reductases in many species have been identified
and kinetically characterized. These include the GSH-dependent DHA reductases like
thioltransferases from vaccinia virus (39) pig liver (24, 33), human placenta (22), and human
neutrophil (31), the 32kDa-DHA reductase isolated from rat liver (25) and human
erythrocytes (26), bovine liver PDI (22, 24, 33), the p52 enzyme from T 1ypanasama cruzi
(38), a DHA reductase fi'om spinach chloroplasts homologous to plant trypsin inhibitor (40),
and a 24 kDa DHA reductase from spinach leaves (41 ). In addition, NADPH-dependent

DHA reductases have been identified which include rat liver thioredoxin/thioredoxin

69

rate (nmolelmln)

 

 

 

 

[DHA] (mM)
10 B
8 _ i
A L
.5 6:
E .
3 .
O .
E 4 -
C .
v
a u
g .
h 2 ..
o .-
................................
o 1 2 3 4 5 6
[GSH] (mM)

Figure 1. Kinetic analysis of the DHA reductase activity of bovine liver PDI at pH 6.9.
Bovine liver PDI was assayed for DHA reductase activity as described in the Methods
section. Kinetic constants for DHA were determined by varying [DHA] as shown and
holding [GSH] = 3.0 mM (A), and kinetic constants for GSH were determined by varying
[GSH] as shown and holding [DHA] = 1.5 mM (B). Each data point (-I-) is the average of
3 separate experiments. Error bars represent one standard deviation of the data and are not
seen if the values are smaller than the size of the symbols.

7O

W (Main)

 

 

 

 

 

Figure 2. Kinetic analysis of the DHA reductase activity of RPLTT at pH 6.9. RPLTT was
assayed for DHA reductase activity as described in the Methods section and Figure 1. Each
data point (-I-) is the average of 3 separate experiments. Error bars represent one standard

deviation of the data.

71

20r- A

rate (nmolelmln)

 

 

 

[DHA] (MM)

rate (nmolelmln)

 

 

 

[GSH] (mM)

Figure 3. Kinetic analysis of the DHA reductase activity of RPLTT at pH 7.2. RPLTT was
assayed for DHA reductase activity as described in the Methods section and Figure 1. Each

data (-I-) point is the average of 3 separate experiments. Error bars represent one standard
deviation of the data.

72

 

 

 

6 A
E
E
E
O
E
5
2
8
[DHA](mM)
12-
‘ B
11-
161
g -
E
2
O
E
5
3
E

 

 

 

 

[GSH] (mM)

Figure 4. Kinetic analysis of the DHA reductase activity of human erythrocyte 32kDa
DHA reductase at pH 6.9. Human erythrocyte 32kDa DHA reductase (1.85 ug for DHA
profile and 1.575 ug for GSH) profile was assayed for DHA reductase activity as described
in the Methods section and Figure 1. Each data point (-O-) is the average of 3 separate
experiments. Error bars represent one standard deviation of the data.

73

rate (nmolelmln)

 

 

[DHA] (mM)

20
18
16
14

12

 

10

rate (nmolelmln)

NJIOQ

 

 

[GSH] (mM)

Figure 5. Kinetic analysis of the DHA reductase activity of human erythrocyte 32kDa
DHA reductase at pH 7.2. Human erythrocyte 32kDa DHA reductase (2.3 ug) was assayed
for DHA reductase activity as described in the Methods section and Figure 1. Each data

point (-I-) is the average of 3 separate experiments. Error bars represent one standard
deviation of the data.

74

873mm 65 meme 2% 5:828:00 BEBE 323 3623 05 3 5 8833-282 805:8: .3 38.528 683 .8228 3.31% e
8ch mo cease—8:8 .838 65 .3 393 .8

as; wEEZe .3 poem—:28 203 829

683?; Wm

«o

.NK Ea E 25522 E Entomoe we Begum—e $53280 2352 e
6.0 In 8 25502 E Become—e me 8525on 3:828 anaco— e

 

 

2 .26 M2 42833 .2 3.84:: .24 :83? M24332; :8
2 «no .2 3334 SN .2 42.39% .2 48.32% .2 x Roam: SE
9-84.446 e2 ea
to 4 am No e 3 no a n... no 4 cm 28 a new 50
so 4 am So 4 mg 88 4 £8 86 4 Ed 2:. a mac $5
.325 emcee
I2 2 a 8m 83% 32m .3 RN 555.32
ecmfioEofl
32.235 bummeommSwEoEH eommeowmceb—omfi. eomflospom <$D summeospom <39
528m .834 mi 824 wE 0mm BEBE 0mm BEBE 58883
.53.—
oE>om

 

momfloscom BefoomaEPEQQ 5:25:82 new $66.53.— 03652
H 05mm.

75

reductase (22) and 3a-hydroxysteroid dehydrogenase (21 ). Many of these enzymes have
several different activities including thioltransferase, thioredoxin/thioredoxin reductase, and
PDI which belong to the class of enzymes known as thiol:disulfide oxidoreductases, which
contain -CXXC- at the active site (42).

The newly measured and reported kinetic constants of the known DHA reductases
are found in Tables I and 11. Of these enzymes, bovine liver PDI is the weakest DHA
reductase with a kW/Km for DHA of 93 :t 14 M’1 sec’1 (33) although reported values vary (22).
The most robust DHA reductases identified to date are RPLTT (33) and human erythrocyte
32kDa-DHA reductase (26) each with a km/Km for DHA of approximately 2 x 104 M" sec'l
at both pH 6.9 and 7.2. Even though there are a wide variety of enzymes represented,
analysis of the kinetic constants yield some common themes. The Klm values for DHA for
all the enzymes range from 0.2 to 4.6 mM with the majority of the enzymes having K“I values
for DHA in the range of 0.2 to 0.7 mM. The lower limit of Km values for DHA is apparently
0.2 mM, even for an enzyme like the 32kDa GSH dependent DHA reductase which is
presumably a specific DHA reductase having no known activities other than as a DHA
reductase (25, 26). In addition, no enzyme has a hat/K"1 for DHA in excess of 2.5 x 104 M’1
sec", a value that is substantially lower than the theoretical limit for a [Cw/Km of 1 x 108 M’'
sec". Based on kinetic comparisons, the 32kDa-DHA reductase from human erythrocytes
(26) would be the primary human DHA reductase because it had a kcat and a [cw/K“ for DHA
which were 6 fold higher than those of human placental thioltransferase (22) (both measured
at pH 6.9) and a kc,“ and a kc,,,/Km for DHA which were 5 fold higher than those of human

neutrophil thioltransferase (31 ) (measured at pH 7.2 and 7.4, respectively). However, the in

76

Kinetic Parameters for DHA Reductases

Table II

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Values for DHA Values for GSH
enzyme pH k... K... kca/Km K... Ira/K...
(min'l) (mM) (NI-1830-1) (mM) (NI-1300")

bovine liver PDI 6.9 16 2.8 93 2.9 91
bovine liver PDI (22) 6.9 94 1.8 870
RPLTT 6.9 374 0.26 24,300 3.5 1810
RPLTT 7.2 305 0.25 20,100 2.5 2050
Human placental Om 6.9 41 0.27 2531
(recombinant) (22)
Human neutrophil er

Native (31) 7.4 66 0.25 4400 2.0 550

Recombinant (31) 7.4 66 0.20 5500 2.0 550
Human erythrocyte 6.9 251 0.28 14,900 2.53 1650
32kDa-DHA reductase
Human erythrocyte 7.2 316 0.21 24,700 3.5 1510
32kDa-DHA reductase
Rat liver 32kDa-DHA 7.2 140 0.245 9520 2.8 833
Rase (25)
Tryp Cruzi p52 (38) 7.0 358 0.67 8910 2.6 2290

NADPH NADPH

Human Thioredoxin 7 .4 90 2.5 600
Reductase (TR) (22)
TR + rat liver 7.4 71 0.7 1690
thioredoxin (22)
Rat 3a-hydroxysteroid 6.9 58.1 4.6 211 0.0043 22,500
dehydrogenase (21 )

 

 

 

 

 

 

 

77

vivo concentrations of these enzymes are not known, which would better indicate the
importance of each enzyme in DHA reduction.

Perhaps the best characterized DHA reductase is thioltransferase. Thioltransferase
has been purified and characterized as a DHA reductase from pig liver (24, 33), human
placenta (22), human neutrophils (3]), rice (43), spinach leaves (44), E. coli. (24), and
vaccinia virus (39). Of these enzymes, even though pigs can synthesize AA and would
presumably not depend on this DHA reductase system, pig liver thioltransferase was the most
robust DHA reductase with a kw/Km for DHA of 2.4 x 104 M‘1 sec'l compared to a [rm/K“ for
DHA for the human enzymes of 2.5-5.5 x 103 M" sec’l (Table H) making the pig liver
enzyme 4-10 fold more efficient. In contrast, there are conflicting reports concerning the
activity of plant thioltransferase. Sha et al. (43) reported that purified rice thioltransferase
has DHA reductase activity although the same group reported that cloned rice
thioltransferase does not (45). In addition, Morell et al. (44) reported that purified spinach
thioltransferase lacks DHA reductase activity.

Strong evidence exists that both NADPH-dependent and GSH-dependent DHA
reduction is important in vivo. The selenoprotein thioredoxin reductase possesses NADPH-
dependent DHA reductase activity (22). Rats fed a selenium deficient diet had an 88%
reduction in thioredoxin reductase activity and a 33% reduction in liver ascorbate levels
providing evidence for selenium dependent enzymes, like thioredoxin reductase, playing a
role in DHA recycling (22). However, 67% of control levels of liver ascorbate remained in
selenium deficient rats, suggesting a role for GSH-dependent DHA reductases, whose
activities were not affected by selenium deficiency (22).

Unable to synthesize AA, when newborn rats were made GSH deficient by

78

administering the y-glutamylcysteine synthetase inhibitor, buthionine sulfoximine (BSO),
drastic decreases in tissue ascorbate levels occurred (23). Similar results are seen when GSH
deficiency is induced in guinea pigs afier BSO administration (46), and when GSH monethyl
ester, which is transported and converted intracellularly into GSH, was administered to GSH
deficient guinea pigs, the onset of scurvy was delayed suggesting a sparing effect of GSH on
ascorbate (48). In adult mice, which are able to synthesize AA, GSH deficiency resulted in
increased AA synthesis in the liver (4 7). These studies demonstrated a role of GSH in the
regeneration of AA, in vivo, but they did not resolve the question of whether the regeneration
of AA from DHA by GSH was enzymatic or chemical, suggesting a need for knockout
transgenic mice studies in the future.

GSH reduction of DHA to AA has been estimated to have a t,,2 of 15 min at a pH of
7.5 and at 25°C (29). However, under the same conditions DHA degrades to non recyclable
products, like 2,3-diketogulonic acid, with a tl,2 of 2 min (30), strongly supporting the
importance of catalytic regeneration of AA from DHA. Evidence for GSH-dependent DHA
reductases in animal tissues have existed for many years (49-51). Bigley et al. (52)
demonstrated that GSH-dependent DHA reductase activity was greatest in human
neutrophils, followed by monocytes, lymphocytes, and fibroblasts. In a study by Park and
Levine (31), the reducing activity of neutrophils after cell lysis was non-dialysable and could
not be accounted for by chemical reduction of DHA by GSH. In addition they found that the
purified DHA reducing activity was thioltransferase (31). In rat liver cytosol, up to 70% of
the DHA reductase activity was irnmunotitratable using an antibody directed towards the rat
liver 32 kDa DHA reductase (53). These experiments demonstrate the importance of GSH-

dependent DHA reductases in DHA reduction in vivo.

79

The emerging picture in the field of DHA reductases is one in which there are several
potential enzymatic systems to regenerate AA from DHA. NADPH-dependent DHA
reductases like 3a-hydroxysteroid dehydrogenase (21 ) and thioredoxin/thioredoxin reductase
(22), and the more catalytically efficient GSH-dependent DHA reductases thioltransferase
(24) and the 32 kDa DHA (25, 26) reductase may all play roles in vivo. It is likely that the
importance of each enzyme in DHA reduction will vary from species to species and cell to
cell. For example, thioredoxin reductase (22) and the 32 kDa DHA reductase (53) have been
separately shown to account for 33% and 7 0% of the DHA reductase activity in rat liver,
completely accounting for the DHA reductase activity in this particular tissue. On the other
hand, thioltransferase has been proposed to be the major DHA reductase in human
neutrophils (31).

The wide variety of DHA reductases, especially the GSH-dependent DHA reductases,
raises several enzyme structure/function questions. In addition, thioltransferase, protein
disulfide isomerase and thioredoxin all belong to the class of enzymes known as
thiol:disulfide oxidoreductases which contain the -CXXC- motif in their respective active
sites (42). Whether each of these enzymes has an identical, similar or different catalytic
mechanism for DHA reduction is unknown. The only catalytic mechanism information
available is derived from the DHA reductase activity of RPLTT. Yang and Wells (54)
compared the relative DHA reductase activities of equal amounts of C258, R26V, and K27Q
variants to that of the native RPLTT. With RPLTT activity defined as 100%, C258 had a
relative activity of 194%, R26V a relative activity of 30%, and K27Q a relative activity of
73% (54). Furthermore, the active site of RPLTT has been previously identified as Cys-22

which has a pKa of 3.8 (55) and is the essential amino acid mediating RPLTI‘ thioltransferase

80

and DHA reductase activities (54). In the following chapter the first detailed description of

a GSH-dependent DHA reductase mechanism is presented; i.e., that of thioltransferase.

81

REFERENCES

(1) Frei, 3., England, L., and Ames, B.N. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 6377-
6381.

(2) Halliwell, B., and Gutteridge, J .M.C. (1990) Arch. Biochem. Biophys. 280, 1-8.
(3) Buettner, GR. (1993) Arch. Biochem. Biophys. 300, 535-543.

(4) Sciamanna, M.A., and Lee, CF. (1993) Arch. Biochem. Biophys. 305, 215-224.
(5) May, J .M., Qu, Z., and Whitesell, RR. (1995) Biochemistry 34, 12721-12728.

(6) Ozaki, M., Fuchinoue, S., Teraoka, S., and Ota, K. (1995) Arch. Biochem. Biophys. 318,
439-445.

(7) Englard, S., and Seifter, S. (1986) Ann. Rev. Nutr. 6, 365-406.
(8) Lyons, BL, and Schwarz, RI. (1984) Nucleic Acids Res. 12, 2569-2579.
(9) Levine, M., Morita, K., and Pollard, H. (1985) J. Biol. Chem. 260, 12942-12947.

(10) Wells, W.W., Dou, C.Z., Dybas, L.N., Jung, C.H., Kalbach, H.L., and Xu, D.P. (1995)
Proc. Natl. Acad. Sci. U.S.A. 92, 11869-11873.

(11) Jung, CH. and Wells, W.W. (1997) Biochem. Biophys. Res. Commun. 239, 457-462.

(12) Bielski, B.H.J. (1982) In Ascorbic Acid: Chemistry, Metabolism, and Uses (Seib, P.A.
and Tolbert, B.M., Eds.) Advances in chemistry Series 200 (Comstock, M.J., Ed.),
pp. 81-100, American Chemical Society, Washington, DC.

(13) Coassin, M., Tomasi, A., Vannini, V., and Ursini, F. (1991) Arch. Biochem. Biophys.
290, 458-462.

(14) Villalba, J .M., Canalejo, A., Buron, M.I., C6rdoba, F., and Navas, P. (1993) Biochem.
Biophys. Res. Commun. 192, 707-713.

(15) Bando, M., and Obazawa, H. (1994) Jpn. J. Ophthalmol. 38, 1-9.
(16) Patterson, J.W. (1950).]. Biol. Chem. 183, 81-88.

(17) Pillsbury, S., Watkins, D., and Cooperstein, SJ. (1973) J. Pharmacol. Exp. T her. 185,
713-718.

82

(18) Bianchi, J., and Rose, RC. (1986) Toxicology 40, 75-82.

(19) May, J.M., Qu, Z.C., Whitesell, RR, and Cobb, CE. (1996) Free Rad. Biol. Med. 20,
543-551.

(20) Welch, R.W., Wang, Y., Crossman, A., Park, J .B., Kirk, KL, and Levine, M. (1995)
J. Biol. Chem. 270, 12584-12592.

(21) Del Bello, B., Maellaro, B., Sugherini, L., Santucci, A., Comporti, M., and Casini, A.F.
(1994) Biochem J. 304, 385-390.

(22) May, J .M., Mendiratta, S., Hill, K.E., and Burk, RF (1997) J. Biol. Chem. 2 72, 22607-
22610.

(23) Mrtensson, J ., and Meister, A. (1991) Proc. Natl. Acad. Sci. USA 88, 4656-4660.

(24) Wells, W.W., Xu, D.P., Yang, Y., and Rocque, P.A. (1990) J. Biol. Chem. 265, 15361-
15364.

(25) Maellaro, E., Del Bello, B., Sugherini, L., Santucci, A., Comporti, M., and Casini, A.F.
(1994) Biochem J. 301 , 471-476.

(26) Xu, D.P., Washbum, M.P., Sun, GR, and Wells, W.W. (1996) Biochem. Biophys. Res.
Commun. 221, 117-121.

(27) Wells, W.W., and Xu, D.P. (1994) J. Bioenerg. Biomem. 26, 369-377.

(28) Wells, W.W., and Jung, CH. (1997) in Vitamin C in Health and Disease (Packer, L.
and Fuchs, J. Eds.) pp. 109-121, Marcel Dekker, Inc., New York, NY.

(29) Crook, EM. (1941) Ph. D. Dissertation, Cambridge.

(30) Ball, M. (1937) J. Biol. Chem. 118, 219-

(31) Park, J.B., and Levine, M. (1996) Biochem. J. 315, 931-938.

(32) Segel, I.H., and Johnson, M.J. (1963) Anal. Biochem. 5, 330-337.

(33) Wells, W.W., Xu, D.P., and Washbum, MP. (1995) Methods. Enzymol. 252, 30-38.
(34) Yang, Y., and Wells, W.W. (1990).]. Biol. Chem. 265, 589-593.

(35) Hillson, D.A., Lambert, N., and Freedman, RB. (1984) Meth. Enzmol. 107, 281-294.

(36) Holmgren, A. (1979) J. Biol. Chem. 254, 9627-9632.

83

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

(3 8) Moutiez, M., Quéméneur, E., Sergheraert, C., Valerie, L., Tartar, A., and Davioud-
Charvet, E. (1997) Biochem. J., 322, 34-48.

(39) Ahn, B.Y., and Moss, B. (1992) Proc. Natl. Acad. Sci. USA 89, 70600-7064.

(40) Triimper, S., Follmann, H., and Haberlein, I. (1994) FEBS Lett. 352, 159-162.

(41) Hossain, M.A., and Asada, K. (1984) Plant Cell Physiol. 25, 85-92.

(42) Mieyal, J .J ., Gravina, S.A., Mieyal, P.A., Srinivasan, U., and Starke, D.W. (1995) in
Biothiols in Health and Disease (Packer, L., and Cadenas, B., Eds.) pp. 305-372,
Marcel Decker, Inc., New York, NY.

(43) Sha, S., Minatuchi, K., Kisaki, N., Sato, K., Ohtsuki, K., Kurata, A., Yoshikawa, H.,
Kotaru, M., Masumura, T., Ichihara, K., and Tanaka, K. (1997) J. Biochem. (Tokyo)
121, 842-848.

(44) Morell, S., Follman, H., and Haberlein, l. (1995) FEBS Lett. 369, 149-152.

(45) Minakuchi, K., Yabushita, T., Masumura, T., Ichihara, K., and Tanaka, K. (1994)
FEBS Lett. 337, 152-160.

(46) Meister, A. (1994).]. Biol. Chem. 269, 9397-9400.
(47) Miirtensson, J., and Meister, A. (1992) Proc. Natl. Acad. Sci. USA 89, 11566-11568.

(48) Méirtensson, J ., han, J ., Griffith, O.W., and Meister, A.(1993) Proc. Natl. Acad. Sci.
USA 90, 317-321

(49) Hughes, RE. (1964) Nature 203, 1068-1069.

(50) Yamamoto, Y., Sato, M., Ikeda, S. (1977) Bulletin of the Japanese Society 0 f Scientific
Fisheries 43, 53-57.

(51) Yamamoto, Y., Sato, M., Ikeda, S. (1977) Bulletin of the Japanese Society 0 f Scientific
Fisheries 43, 59-62.

(52) Bigley, R., Riddle, M., Layman, D., and Stankova, L. (1981) Biochim. Biophys. Acta.
659, 15-22.

(53) Paolicchi, A., Pezzini, A., Saviozzi, M., Piaggi, S., Andrevccetti, M., Chieli, B.,
Malvadi, G., and Casini, A.F. (1996) Arch. Biochem. Biophys. 333, 489-495.

84

(54) Yang, Y., and Wells, W.W. (1991) J. Biol. Chem. 266, 12759-12765.

(55) Gan, Z-R., and Wells, W.W. (1986) J. Biol. Chem. 261, 996-1001.

85

Chapter III

The Catalytic Mechanism of the Glutathione-Dependent Dehydroascorbate

Reductase Activity of Thioltransferase (Glutaredoxin)

This chapter has been submitted to Biochemistry and is currently under review.

86

ABSTRACT

The catalytic mechanism of the glutathione (GSH) dependent dehydroascorbic acid
(DHA) reductase activity of recombinant pig liver thioltransferase (RPLTT) was
investigated. Kinetic analysis of the GSH-dependent DHA reductase activities of native
RPLTT and the R26V, K27Q, and C258 variants indicated that Arg-26 and Lys-27 are
involved in catalysis, but Cys-25 appears to lower the catalytic capacity of the enzyme since
its substitution with a serine enhances the kc,,(app). Iodoacetamide (IAM) inactivated the
DHA reductase activities of RPLTT, and its C258, R26V, and K27Q variants confirming
the essential role of cysteine in the reaction mechanism. When preincubated with DHA,
RPLTT or the R26V and K27 Q but not C25S variants, proteins were protected against IAM
inactivation, suggesting that RPLTT and R26V and K27Q enzymes have the ability to
chemically reduce DHA forming AA and the intramolecular disulfide form of the enzyme.
Electrochemical detection of AA demonstrated the ability of reduced RPLTT and the R26V
K27Q, and C258 variants to chemically reduce DHA to AA in the absence of GSH. Also
in the absence of thiol substrate, RPLTT and the R26V and K27Q proteins had nearly
identical initial rates of DHA reduction which were 4 to 5 fold greater than that of the C25S
variant. Isoelectric focusing analysis revealed that the product of reaction of reduced RPLTT
and the R26V and K27Q enzymes but not the C25S variant with DHA was indeed oxidized
enzyme. Based on the experimental results, a catalytic mechanism for the DHA reductase
activity of RPLTT is proposed. This is the first description of a catalytic mechanism of a

glutathionezdehydroascorbate oxidoreductase (EC 1.8.5.1).

87

Ascorbic acid (AA) is an important antioxidant (1-6) and electron donor for enzymes
such as dopamine-[J-hydroxylase, prolyl-4-hydroxylase, and lysyl hydroxylase (7). The
fimction of AA as an electron donor is highlighted by its importance in collagen biosynthesis
(8) and norepinephrine biosynthesis (9). In addition, AA has recently been demonstrated to
be essential for the release of insulin from the pancreatic islets of scorbutic guinea pigs (10)
and as a cofactor for mitochondrial glycerol-3-phosphate dehydrogenase (I I).

When AA carries out its function, it undergoes two successive one electron
oxidations to semidehydroascorbic acid (semiDHA) and dehydroascorbic acid (DHA). A
relatively stable radical, semiDHA may disproportionate to AA and DHA (12) or be directly
reduced to AA by an NADH-semiDHA reductase (13-15). DHA can be reduced to AA in
an NADPH dependent manner by both 3a-hydroxysteroid dehydrogenase (16) or thioredoxin
reductase (1 7). Glutathione (GSH) chemically reduces DHA to AA, and three mammalian
GSH-dependent DHA reductases have been identified: thioltransferase (glutaredoxin) (18),
protein disulfide isomerase (PDI) (I8), and a 32kDa enzyme (32kDHAR) identified in rat
liver (19) and human erythrocytes (20). In vivo experiments demonstrated the role of GSH
in DHA reduction (21), but whether in vivo DHA reduction is primarily chemical, enzymatic,
or both remains unclear with evidence existing for both mechanisms (22,23). Recently, Park
& Levine (24) provided evidence for enzymatic recycling of DHA in human neutrophils,
catalyzed in large part by thioltransferase. Despite the potential importance of GSH-
dependent DHA reductases in the maintenance of cellular AA, the only catalytic mechanism
information available is derived fi'om the DHA reductase activity of recombinant pig liver
thioltransferase (RPLTT) (25).

Thioltransferase is an 11.7 kDa protein belonging to a class of enzymes, containing

88

the -CXXC- active site motif, known as thiol:disulfide oxidoreductases (26).
Thioltransferase catalyzes thiol/disulfide exchange reactions (for reviews see 26, 27), the
reduction of alloxan to dialuric acid (28), and the reduction of DHA to AA (18), all in a GSH
dependent manner. The active site of RPLTT has been previously identified as Cys-22 which
has a pK, of 3.8 (29) and is the essential amino acid mediating RPLTT thioltransferase and
DHA reductase activities (25 ).

In the present study, we investigated the catalytic mechanism of the DHA reductase
activity of thioltransferase using RPLTT and several site directed mutant proteins. The
results provided experimental evidence for a proposed mechanism for the DHA reductase
activity of RPLTT, which can be used as a model for studying other GSH-dependent DHA

reductases.

MATERIALS AND METHODS

Materials. Coomassie Brilliant Blue R-250, Iodoacetic acid (IAA), iodoacetamide
(IAM), EDTA, and glutathione disulfide (GSSG) were purchased from Sigma, St. Louis,
MO. GSH, dithiothreitol (DTT), glutathione disulfide reductase, and NADPH were products
of Boehringer Mannheim, Indianapolis, IN. Metaphosphoric acid and hydroxyethyldisulfide
(HED) were purchased from Aldrich, Milwaukee, WI. Thiourea was obtained from
Matheson Coleman and Bell, Cincinnati, OH, and bromine was purchased from Acros
Organics, Pittsburgh, PA. AA, ammonium sulfate, and NaCl were purchased from J .T.
Baker, Phillipsburg, NJ, and isopropylthiogalactoside (IPTG) was purchased from
Calbiochem, San Diego, CA. CM Sepharose, Sephacryl S-200, and Sephadex G25 were

products of Amersham Pharmacia Biotech, Piscataway, NJ. Centriprep-3 concentrators were

89

purchased from Amicon, Inc., Beverly, MA. Serva Servalyt Precotes for pH 3-10, and p1 test
mix were acquired fi'om Crescent Chemical Co., Inc., Hauppage, NY. DHA was prepared
as described previously (30). S-sulfocysteine was prepared by the method of Segel and
Johnson (31 ).

Expression of the T hioltransferase Gene and Purification of Pig Liver
T hioltransferases. Expression of the gene and purification of RPLTT and the C25S variant
were carried out as described previously (32), with the following modifications. After
ammonium sulfate precipitation, the pellets were resuspended in 2 ml of 20 mM sodium
phosphate, pH 6.5 (buffer A), containing 2 mM DTT, and dialyzed overnight (2 x 2L) against
the same buffer. After dialysis, the sample was loaded onto a Sephacryl S-200 column ( 2.5
cm x 89 cm) equilibrated and eluted with buffer A containing 2 mM DTT. The pooled
active fractions were incubated with 10 mM HED for 30 min at room temperature and loaded
onto a CM-Sepharose column (5 cm x 38 cm) equilibrated with buffer A. The column was
washed with l L of buffer A followed by l L of 20 mM sodium phosphate pH 7.5 (buffer B).
The bound disulfide enzyme was eluted with an 800 ml 00-10 M NaCl linear gradient in
buffer B. The active fractions were pooled and concentrated using Centriprep 3
concentrators and stored at -70°C until use. Protein concentrations were determined by the
bicinchoninic acid protein assay (Pierce Chemical Company) according to the manufacturer’s
directions. During the purification, the enzyme activity was assayed as described previously
(35).

The R26V and K27Q encoding genes were subcloned from the pKK233-2 vector (25)
into pET23d(+) (N ovagen, Inc.) using pET23d(+) HindIII and NcoI sites and standard

molecular biology techniques (33). In these constructs, R26V and K27Q proteins were

90

h— '—.-E

synthesized without the His-Tag®. The two proteins were produced in E. coli strain
BLD21(DE3) (Novagen, Inc.) transforrnants that were grown overnight at 37°C in Luria-
Bertani medium containing 50 ug/ml ampicillin. These cultures were diluted 1/200 in l L
of fresh LB medium with 50 ug/ml ampicillin and shaken at 37°C until an A550 of 0.4-0.7
was reached. Protein expression was induced by adding 0.45 mM IPTG to the culture, and
after 3.5 h the cells were harvested by centrifugation at 10,000 x g for 10 min. From this
point, R26V and K270 proteins were purified as described for RPLTT and C258 enzyme
except that for the R26V and K27Q variants, 20 mM sodium phosphate, pH 6.0, and 20 mM
sodium phosphate, pH 7.0, were used instead of buffers A and B, respectively.

Dehydroascorbic Acid Reductase Assay and Enzyme Kinetics. DHA reductase
activity was assayed as described previously (30). Assays performed to determine the kinetic
constants were run in 200 mM sodium phosphate, 1 mM EDTA, pH 6.9, in a final volume
of 500 a] at 30°C and recorded for 3 min at 265.5 nm. A blank without enzyme was run
simultaneously with each assay and the difference gave the activity in nmoles/min. Kinetic
constants for DHA were determined by varying [DHA] and holding [GSH] = 3.0 mM, and
kinetic constants for GSH were determined by varying [GSH] and holding [DHA] = 1.5 mM.
The following enzyme amounts were used in their respective kinetic studies: 0.72 ug
RPLTT, 2.12 ug R26V, 1.45 ug K27Q, and 0.44 ug C258. Km(app) and Vm,,(app) values
were calculated by nonlinear least-square fit to the velocity versus substrate concentration
data using the PSI-Plot 3.5 software. The values for kcat were calculated by dividing
Vm,,(app) by the molar concentration of the enzymes.

Preparation of Reduced T hioltransferases. RPLTT and the C258, R26V, and K27Q

variant samples, purified to homogeneity, were separately adjusted to 20 mM DTT and

91

reduced for 30 min at room temperature, subjected to Sephadex 625 (2.5 cm x 78 cm)
filtration, and eluted with triply distilled H20 to remove the excess DTT. Fractions with
thioltransferase activity were pooled and lyophilized. Lyophilized enzyme was resuspended
in triple distilled H20, and protein concentration were determined.

Iodoacetamide Inactivation of T hioltransferases DHA Reductase Activity. Reduced
RPLTT (712M) and its C25S ( 7uM), R26V (10 uM), and K27Q (10 12M) variants were
incubated with 0.1 mM IAM, separately, in 100 mM sodium phosphate, pH 7.5. Aliquots
were removed at various times up to 45 min and assayed for DHA reductase activity in 200
mM sodium phosphate, pH 6.85, containing 1 mM EDTA and 3.0 mM GSH. Upon the
addition of 1.5 mM DHA, reductase activity was recorded at 265.5 nm for 2 min at 30°C.
Percent activity was determined for each enzyme by running concurrent control with IAM
absent. The IAM inactivation experiments were also performed after preincubation of each
enzyme separately with 0.05 mM DHA plus 0.1 mM GSH, 0.05 mM AA, 0.05 mM GSSG,
0.5 mM HED, 0.05 mM GSH, or 0.5 mM DHA in 100 mM sodium phosphate, pH 7.5, for
10 min at room temperature. IAM was then added to 0.1 mM and the experiments conducted
as described above.

Reduction of DHA to AA by T hioltransferases in the Absence of GSH. Equal amounts
(6 12M) of reduced RPLTT or the C25S, R26V, and K27Q variants were separately incubated
in a 100 ,ul reaction volume with 600 uM DHA for 0.5, 1, 2, 5, and 10 min in 200 mM
sodium phosphate, 1 mM EDTA, pH 6.85. In addition, control reactions containing DHA
and no enzyme were run. At each time point, 100 ill of 10% metaphosphoric acid, 1 mM
thiourea, 1 mM EDTA was added to the reaction tube. The samples were centrifuged at

14,000 rpm for 2 min in an Eppendorf Microcentrifuge. AA concentrations were determined

92

on the supernatant as described previously (34). Separation and quantitative analysis were
accomplished using a 3.9 x 150 mm Waters Delta Pak-Su C18 reversed phase column with
a mobile phase of 50 mM sodium phosphate, pH 3.0, and a flow rate of 0.7 ml per min. AA
was detected using an ESA Model 5200A Coulochem II electrochemical detector , and an
ESA Model 5011 analytical cell. Detector one, -150 mV, detector two, + 125 mV, and a
guard cell, + 200 mV were optimal for quantification. The areas of the ascorbic acid peaks
were determined by using a Hewlett-Packard Model 3395 integrating recorder. Areas of
sample AA were compared with those from standard solutions of 50 nM AA to 3 11M AA
dissolved in 5% metaphosphoric acid, 0.5 mM thiourea, 1.0 mM EDTA, 100 mM sodium
phosphate (pH 2).

Isoelectric Focusing. Reduced RPLTT or its C25S, R26V, and K27Q variants, 12
ug each, were incubated alone and with 1 mM DTT, 1 mM HED, 1 mM DHA, 1 mM GSH,
0.5 mM DHA plus 1 mM GSH , 0.5 mM GSSG, or 0.5 mM AA separately in 40 mM sodium
phosphate, pH 7.5, for 30 min in a 30°C water bath. IAA was added to a final concentration
of 2 mM and the samples were incubated for 30 additional min at 30°C. Controls included
enzyme alone, enzyme + lmM DTT, enzyme + 1 mM HED, or enzyme + 1 mM DHA. Each
of the incubation mixtures (in a total volume of 10 ul) was then analyzed directly on a
Servalyt Precotes isoelectric focusing gel, pH 3-10 (150 pm thick, 125 mm x 125 mm),
following the manufacturer’s instructions for 2666 volt hours on an LKB 2217 Ultraphor
Electrofocusing Unit with a Lauda K-2/R cooling system. After focusing, the gel was fixed
with a solution of 30% ethanol and 10% glacial acetic acid, stained with a solution of 0.001%
Coomassie Brilliant Blue R-250, 10% glacial acetic acid, and 50% ethanol, and washed with

the fixing solution.

93

RESULTS AND DISCUSSION

Kinetic Characterization of Wild Type and Mutant RPLT T Dehydroascorbic Acid
Reductase Activity. Previously, Yang and Wells (25) compared the relative DHA reductase
activities of equal amounts of C25S, R26V, and K27Q mutant proteins to that of RPLTT.
With RPLTT activity defined as 100%, C25S had a relative activity of 194%, R26V a
relative activity of 30%, and K27Q a relative activity of 73% (25). In the current study, the
kinetic parameters for C25 S, R26V, and K27Q variants were determined and compared to
those of RPLTT (Table l). The kinetic plots for each enzyme were best fit by the Michaelis-
Menten equation (RPLTT (Figure 2 in Chapter 11), C258 (Figure 1), R26V (Figure 2), and
K27Q, (Figure 3), EZ-Fit Kinetic Program, 35). The catalytic constant, kc,,(app), for R26V
and K27Q proteins were 2.1 and 1.8 fold lower, respectively, than that of RPL'I'I‘. The
affinity constant, Km(app), for DHA of R26V enzyme was 1.6 fold higher than that of
RPLTT. As a result, the specificity constants (kw/Kn) of R26V protein were 3.3 fold lower
for DHA and 1.9 fold lower for GSH than those of RPLTT. Because of the decrease in
kw(app) for K27Q enzyme , the specificity constants were 1.9 fold lower for DHA and 1.5
fold lower for GSH than those of RPLTT. The C258 variant had a 2.7 fold higher kw (app),
a 3.2 fold higher Km(app) for DHA, and a 1.5 fold higher Km(app) for GSH than those of
RPLTT. As a result of the increased Km(app) for DHA, the specificity constant of C258
protein for DHA was not significantly different than that of RPLTT. However, the
specificity constant for GSH was 1.9 fold higher than that of RPLTT. Arg-26 and Lys-27
participated in catalysis because both R26V and K27 Q variants had lower values for kw(app)
than RPLTT. In contrast, a cysteine at position 25 lowered the catalytic capacity of the

enzyme since the placement of a serine at position 25 enhanced the kc,,(app).

94

rate (nmoldmln)

 

 

 

o 1 2 3 4 5
[DHA] ("I")

rate (nmolelmln)

 

 

A 1 1 A l A A I A l 1 A 1 A A A A
O 1 2 3 4 5 6 7 8 9 10 1 1
[GSH] (mM)

Figure 1. Kinetic analysis of the DHA reductase activity of the C25S enzyme. C258
variant ( 0.44 ug) was assayed for DHA reductase activity in 200 mM sodium phosphate, 1
mM EDTA, pH 6.9. A blank without enzyme was run simultaneously with each assay and
the difference gave the activity in nmoles/min (-I-). Kinetic constants for DHA were
determined by varying [DHA] as shown and holding [GSH] = 3.0 mM, (A) and kinetic
constants for GSH were determined by varying [GSH] as shown and holding [DHA] = 1.5
mM (B). Each data point is the average of 3 separate experiments. Error bars represent one
standard deviation of the data.

95

rate (nmolelmln)

 

 

 

0 1 2 3 4 5
[DI-IA] (mM)

rate (nmolelmln)

 

 

[GSH] (mM)

Figure 2. Kinetic analysis of the DHA reductase activity of the R26V variant. R26V
enzyme ( -I-) was assayed for DHA reductase activity and kinetic constants determined as
described in Figure l.

96

rate (nmolelmln)

 

 

 

0 1 2 3 4 5
[DHA] (mM)

25-
p

 

rate (nmolelmln)

 

 

 

[GSH] (mM)

Figure 3. Kinetic analysis of the DHA reductase activity of the K27Q variant. K27Q
enzyme (-I-) was assayed for DHA reductase activity and kinetic constants determined as
described in Figure 1.

97

A888: _8:333m Peon—.mfi ne‘er—8:9 modva a 53, 88> .5.qu m:_::ommo.:8 Bob 388.33: 33:85:39....
.82on .8838 .285 8.889 :38 e .23 oz? 8.8 38888 58. 8826 88885 ._.

.oSBtOm

Wm 873mm 05 mam: 38 88:80:00 08333 3.33 382? 05 2 5 0.88388— :8:=:o: 3 88328 0.83 838 3:32 e
.8888 05 .«o 88.5888 8.0:: 05 3 3:855; mEEE: as 88338 083 829 .s« o

.83 8:353: 33538: :03 08: .5.qu :8 3:388
282% .888 38:88 2: :3 8:88: 8 3:88:88 388m 8.88 8 85 .«o 8838 8888 a :88 2: 083 3:388 2852e

 

 

:2 x 3.0 a 2: :2 an mg a 85 :2 on $6 a 3 2 x No a 8 :8
3.2 $382 :2 x5 «2 .2x 51:: .2 88428 $5
9.83-23 53.3
$2.38 3a 3 to?” 3 831m :8
48 H Rd .58 a :3 .25 a ad 85 a 86 SE
.228 83$
*4 a gm *3 a. o: .. 2: a 82 om a 42 2758.3
.08: .58: .38 95.8 8588

 

388:88m 285% 08888525. 8>§ mE 83:2 :8 09$. 23?

H 2an

98

Proposed Mechanism of RPL TT’s DHA Reductase Activity. Based on our knowledge
of RPLTT’s DHA reductase activity (25) and on the reported mechanism of RPLTT’S thiol-
disulfide exchange activity (26, 27) a mechanism for RPLTT’s GSH-dependent DHA
reductase activity is proposed (Figure 4). Reaction 1 is the formation of a thiohemiketal
intermediate between RPLTT and DHA. With the wild type enzyme, the reaction can then
proceed in two possible ways: (i) in reaction 2, GSH displaces AA from RPLTT and a mixed
disulfide between RPLTT and GSH is formed; (ii) in reaction 3, Cys-25 attacks the mixed
disulfide at Cys-22 and displaces AA forming an intramolecular disulfide bond. Reaction
3 is followed by reaction 4 during which GSH attacks the intramolecular disulfide bond to
form a mixed disulfide between RPLTT and GSH at Cys-22. In reaction 5 a second
molecule of GSH attacks the mixed disulfide, ES-SG, forming GSSG and the original
reduced form of RPLTT. The alteration in the C258 mutant protein is indicated by (OH).
Since CZSS enzyme cannot proceed via reactions 3 and 4, only reactions 1, 2 and 5 are
feasible. In the thiol-disulfide exchange mechanism, a disulfide compound like HED or
GSSG, RSSR, would replace DHA, and RSH would be released upon formation of a mixed
disulfide between Cys-22 and RSSR; i.e., RPLTT-SSR (36). This species could have either
Cys-25 attacking Cys-22 and forming an intramolecular disulfide bond as in reaction 3, or
GSH attacking Cys-22 and displacing RSH as in reaction 2 (Figure 4) (36). Initial reaction
of RPLTT and RSSG would release RSH and form RPLTT-S-SG. This species would then
undergo reaction 5 to restore reduced enzyme. In vivo, thioltransferase would most likely

reduce protein glutathionated substrates, RSSG, to RSH and GSH (3 7, 38).

99

CHZOH

 

-HA
H OHO CHZOH
o H 0“
H
o
o o k (I) H HO
: o
m. W E/
2
S-CH -C 6 S_ \5"
E’ 2 \*—() IAM + E: GSH (0")
\SH 0 SH
(OH) (OH)
+ l—
GSSG
(5) S
GSH

GSH

 

Figure 4: Proposed mechanism of GSH-dependent DHA reduction catalyzed
by RPLTT

100

ilf'

 

Iodoacetamide Inactivation of the Dehydroascorbic Acid Reductase Activity of Wild
Type and Mutant RPLTT. To investigate substrate binding to RPLTT, IAM inactivation
studies of the DHA reductase activity of reduced RPLTT and its C258, R26V, and K27Q
variants were carried out. Only Cys-22 of pig liver thioltransferase reacted with iodo-[l-
l4C]acetic acid (39), and treatment of reduced RPLTT or its C258, R26V, and K27Q variants
inactivated each enzyme’s thiol:disulfide oxidoreductase activity (25). IAM inactivated the
DHA reductase activity of reduced RPLTT (Figure 5) as well as the reduced C258 (Figure
6), R26V (Figure 7), and K27Q (Figure 8) enzymes. For each enzyme, the inactivation
profile best fit the exponential decay function, y = y0 + Ale("‘"), with a k (the rate constant of
inactivation) = 0.163 t 0.02 min'l for RPLTT, versus 0.202 1 0.023 min‘l for C258, 0.196
:1: 0.092 min’l for R26V, and 0.146 i 0.028 min‘l for K27Q enzymes. Both R26V (Figure
7) and K27Q (Figure 8) variants were incompletely inhibited with 40% residual activity after
45 minutes of incubation with IAM. In contrast, RPLTT (Figure 5) and C258 enzyme
(Figure 6) had residual activities of 15% and 8%, respectively. Several attempts were made
to prepare reduced R26V and K27Q proteins to increase the inactivation of their respective
DHA reductase activities, but the results were the same as in Figures 7 and 8 with incomplete
inactivation of both mutants (data not shown). IAM inactivation of the DHA reductase
activity of reduced RPLTT and the C258, R26V, and K27Q variants confirmed the role of
cysteine as the key catalytic residue in the mechanism (25).
Preincubation of reduced RPLTT (Figure 5) or the C258 (Figure 6), R26V (Figure
7), or K27Q (Figure 8) variants for 10 minutes with 0.05 mM GSH enhanced the [AM
inactivation of RPLTT and K27Q enzymes and had no effect on the [AM inactivation of

C258 and R26V variants. The inactivation profiles best fit the exponential decay function

101

with a k = 0.133 :t 0.015 min‘I and an Al (the maximal amount of inhibition) = 93 i 4 % for
RPLTT, a k= 0.146 :1: 0.014 min" and an AI = 100 at l % for C25S enzyme, a k = 0.105 :1:
0.012 min'1 and an AI = 80 :t 4 % for R26V protein, and a k = 0.108 i 0.01 min’l and an A1
= 89 :1: 4 % for K27Q variant. The k for C258, R26V, and K27Q enzymes were significantly
lower than the k when enzyme was inactivated with IAM with no preincubation. However,
preincubation with GSH enhanced the % inactivation of K27Q protein by 30% (Figure 8),
and RPLTT by 8% (Figure 5). These results indicated that the resistance of K27Q enzyme
to IAM inactivation is due to this protein being in the oxidized form. GSH can reduce the
oxidized enzyme to expose Cys-22 to reaction with IAM. As shown in Figure 4, GSH could
react with the oxidized enzyme (reaction 4) forming ES-SG. A second molecule of GSH
could react with ES-SG (reaction 5) forming GSSG and the reduced enzyme, which is
susceptible to IAM inactivation (reaction 6). When studying the thiol:disulfide
oxidoreductase activity of RPLTT, Yang & Wells (36) found that GSH enhanced RPLTT and
K27Q variant inactivation by IAM, as seen in the current studies. In addition, Yang & Wells
(36) found the R26V protein to be partially resistant to IAM inactivation, with little effect
of GSH preincubation on IAM inactivation. Arg-26 has been determined to be the key amino
acid for the low pKa of Cys-22, however, its pKa is as yet undetermined (25). At pH 7.5,
Cys-22 in R26V protein will primarily be in the protonated form rendering Cys-22 resistant
to IAM inactivation because IAM will only react with the unprotonated or thiolate form of
cysteine.

Preincubation of reduced RPLTT (Figure 5) or its C25S (Figure 6), R26V (Figure 7),
or K27Q (Figure 8) variants for 10 minutes with 0.05 mM DHA + 0.1 mM GSH significantly

protected the enzyme’s DHA reductase activity from IAM inactivation. R26V and K27Q

102

variants were completely protected, whereas RPLTT and C25S enzyme retained 80 d: 11 %
and 50 i 7 % of their respective DHA reductase activities afier 45 min. When each enzyme
was separately preincubated with the reaction products, 0.05 mM AA (Figures 5-8) or 0.05
mM GSSG (Figures 5-8), only preincubation with GSSG protected the enzyme’s DHA
reductase activity from IAM inactivation. Again, the inactivation profiles when each enzyme
was preincubated with AA best fit the exponential decay equation with a k = 0.203 i 0.008
min‘1 and an A, = 80 i1% for RPLTT, a k = 0.146 at 0.011 min'1 and an A, = 92 d: 3 % for
C258 enzyme, a k = 0.205 3: 0.035 min”l and an A, = 63 i 5 % for R26V protein, and a k =
0.238 i 0.029 min‘1 and an A, = 59 :t 3 % for the K27Q variant. Preincubation of the
enzymes with another disulfide reagent, 0.5 mM HED, completely protected the enzyme’s
DHA reductase activity from IAM inactivation (data not shown). Reduced RPLTT or the
C25S, R26V, and K27Q variants were protected from IAM inactivation by forming a mixed
disulfide bond with GSSG or HED to generate ES-SG or ES-S—(CH2)2-OH, respectively.
When studying the thiol:disulfide oxidoreductase activity of thioltransferase, RPLTT or
R26V and K27Q proteins were protected from IAM inactivation by preincubation with the
substrate S-sulfocysteine or HED (36). In addition, RPLTT and C258 enzyme were shown
to be oxidized or form enzyme-substrate intermediates with the disulfide species cystine or
HED (36). Because Cys-25 is present, RPLTT or R26V and K27Q proteins could also be
protected from IAM inactivation by forming an intramolecular disulfide bond between Cys-
22 and Cys-25, i.e. oxidized enzyme, a form of RPLTT shown to exist by x-ray
crystallography (40).

In contrast to reduced RPLTT (Figure 5) or the R26V (Figure 6) or K27Q (Figure

7) variants, the DHA reductase activity of C25S protein (Figure 8) was not protected from

103

IAM inactivation by a 10 min preincubation with 0.5 mM DHA. The inactivation profile of
C25S enzyme best fit the exponential decay equation with a k = 0.195 t 0.04 min'l and an
A, = 90 2t 8 %. The fact that C25S enzyme was not protected suggests that the postulated
thiohemiketal intermediate (reaction 1, Figure 4) is not stable enough as an enzyme-substrate
intermediate to prevent IAM inactivation. If this intermediate exists, it probably dissociates

into the reduced enzyme and DHA, and IAM then reacted irreversibly with the reduced

enzyme (reaction 6, Figure 4).

104

%actlvlty

 

 

 

..\ u A
l 0 ‘-
0 1 fi T l ' T l ‘ l
0 1o 20 so 40 so
time (min)

Figure 5. IAM inactivation of reduced RPLTT with and without preincubation with
substrates. Reduced RPLTT (711M) was separately incubated either with 0.1 mM IAM (-O-)
in 100 mM sodium phosphate, pH 7.5 or preincubated with 0.05 mM GSH (-O-), 0.05 mM
GSSG (-I-), 0.05 mM DHA + 0.1 mM GSH (-|:1-), 0.5 mM DHA (-A-), or 0.05 mM AA (-

A- ) for 10 min in buffer prior to IAM addition. Aliquots were removed at the time points

shown and assayed for DHA reductase activity as described in the Methods section. Each
data point is the average of 3 separate experiments. Error bars represent one standard
deviation of the data and are not seen if the values are smaller than the size of the symbols.

105

o/o activity

 

 

 

 

 

 

time (min)

Figure 6. IAM inactivation of reduced C25S protein with and without preincubation with
substrates. Reduced C25S (7aM) was separately incubated either with 0.1 mM IAM (-O-)
in 100 mM sodium phosphate, pH 7.5 or preincubated with 0.05 mM GSH (-0-), 0.05 mM
GSSG (-I-), 0.05 mM DHA + 0.1 mM GSH (-E1-), 0.5 mM DHA (-A-), or 0.05 mM AA (-

A- ) for 10 min in buffer prior to IAM addition. Aliquots were removed at the time points

shown and assayed for DHA reductase activity as described in the Methods section. Each
data point is the average of 3 separate experiments. Error bars represent one standard
deviation of the data and are not seen if the values are smaller than the size of the symbols.

106

14o-

 

 

 

130:
1203
1103 '
100‘ g u’\ /
so: \\
ESQ-t: \u
.2 70‘ X
8 6°: -\
32 :3 3I\\ I
., .
30+1 $3; —_‘9
+ ‘33-—
20-
10:
o‘. . . . r a . . . . .
o 10 20 30 40 so
time(min)

Figure 7. IAM inactivation of reduced R26V protein with and without preincubation with
substrates. Reduced R26V (711M) was separately incubated either with 0.1 mM IAM (-O-)
in 100 mM sodium phosphate, pH 7.5 or preincubated with 0.05 mM GSH (-O-), 0.05 mM
GSSG (-I-), 0.05 mM DHA + 0.1 mM GSH (421-), 0.5 mM DHA (-A-), or 0.05 mM AA (-

A- ) for 10 min in buffer prior to IAM addition. Aliquots were removed at the time points

shown and assayed for DHA reductase activity as described in the Methods section. Each
data point is the average of 3 separate experiments. Error bars represent one standard
deviation of the data and are not seen if the values are smaller than the size of the symbols.

107

120-
110-

 

 

 

 

100 _ _
‘ K ‘ \zz ;:
E
> .‘.
E .
33 \‘\.\ ‘
30': -1 Q
20‘ .
- I
10: I

0-1
—L
O
M
O
to
O
A
O
0'!
0

time (min)

Figure 8. IAM inactivation of reduced K27Q protein with and without preincubation with
substrates. Reduced K27Q (7aM) was separately incubated either with 0.1 mM IAM (~O-)
in 100 mM sodium phosphate, pH 7.5 or preincubated with 0.05 mM GSH (-0-), 0.05 mM
GSSG (-I-), 0.05 mM DHA + 0.1 mM GSH (-[:1—), 0.5 mM DHA (-A-), or 0.05 mM AA (-

A- ) for 10 min in buffer prior to IAM addition. Aliquots were removed at the time points

shown and assayed for DHA reductase activity as described in the Methods section. Each
data point is the average of 3 separate experiments. Error bars represent one standard
deviation of the data and are not seen if the values are smaller than the size of the symbols.

108

Ability of Reduced Wild Type and Mutant RPL T Ts to Chemically Reduce
Dehydroascorbic Acid to Ascorbic Acid. Because they contain Cys-25, we hypothesized
that reduced RPLTT or its R26V and K27Q variants could chemically reduce DHA to AA
in the absence of GSH forming oxidized enzyme (reactions 1 and 3, Figure 4). Reduced
RPLTT and the C258, R26V, and K27Q proteins all could chemically reduce DHA to AA
in the absence of GSH (Figure 9). DHA alone was also spontaneously reduced to AA and
L-erythroascorbic acid as seen by Jung & Wells (41 ) with an initial rate of 122 i 34 nM
AA/min reaching a maximum AA concentration of 109 :1: 17 nM after 10 min. C258 enzyme
had an initial rate of 201 i 4 nM AA/ min and reached a maximum AA concentration of 277
i: 18 nM after 10 min. RPLTT and its R26V and K27Q variants were the most robust DHA
reductants with initial rates of 883 :t 76 nM AA/min, 896 i 144 nM AA/min, and 1053 :t 60
nM AA/min, respectively, and maximum AA concentrations of 3. 1 8 :1: 0.14 uM, 2.09 d: 0.79
uM, and 2.62 i 0.25 MM, respectively.

The results (Figure 9) indicated that reduced RPLTT and its R26V and K27Q variants
can undergo reactions 1 and 3 (Figure 4) where one mole of reduced enzyme reduces one
mole of DHA to one mole of AA and one mole of oxidized enzyme. While C258 protein
cannot undergo reaction 3 because it lacks Cys-25 to form an intramolecular disulfide bond,
it still reduced DHA to AA at a rate greater than that of DHA alone, but at a much lower rate
than that of RPLTT and its R26V and K27Q variants, suggesting a weak ability of C25S
protein to form an intermolecular disulfide bond upon the reduction of DHA.

The initial rates of reactions 1 and 3 (Figure 4) for RPLTT and its R26V and K27Q
variants were nearly identical. However, when RPLTT and its R26V and K27Q variants are

compared kinetically in the presence of GSH (Table I), R26V and K27Q mutant proteins

109

have decreased DHA reductase activity. These results suggest that Arg-26 and Lys-27 do not
play a role in DHA binding to the enzyme (reaction 1, Figure 4), or the conversion of the
postulated thiohemiketal intermediate to AA and oxidized enzyme (reaction 3, Figure 4).
Recently, Srinivasan et al. (42) provided evidence that the rate-limiting step in the
thiol:disulfide oxidoreductase activity of thioltransferase is the nucleophilic attack of GSH
on ES-SG forming GSSG and reduced enzyme (reaction 5, Figure 4). If this step is also the
rate limiting step of RPLTT’s DHA reductase activity, one would expect to see an effect of
the R26V and K27Q variants only at this point and not elsewhere in the reaction mechanism.
The evidence presented in this paper supports the hypothesis that the rate limiting step in
RPLTT’s DHA reductase activity is reaction 5 (Figure 4), because neither mutation of Arg-
26 to Val or Lys-27 to Gln affected the ability of the enzyme to undergo reactions 1 and 3

(Figure 4), while both mutant proteins had decreased DHA reductase activity (Table I).

110

 

[M1 ("M1

 

 

 

time(mln)

Figure 9. HPLC analysis of AA production by DHA incubation with reduced RPLTI‘ and
the C258, R26V, and K27Q variants. Equal amounts (6 uM) of reduced RPLTT (-O-) and
the C25S (—O-), R26V (-I-), and K27Q (-E1-, 10 aM) proteins were separately incubated in
a 100 ul reaction volume with DHA (600 14M) for various lengths of time in 200 mM sodium
phosphate 1 mM EDTA, pH 6.85. In addition, DHA alone control reactions were run (-O-).
At the time points shown, 100 #1 of 10% metaphosphoric acid, 1 mM thiourea, 1 mM EDTA
was added to the reaction tube and AA concentrations were determined by an HPLC
chromatography procedure as described in the Methods section. Each data point is the
average of at least three measurements of one to three separate experiments. Error bars
represent one standard deviation of the data and are not seen if the values are smaller than
the size of the symbols.

111

Isoelectric Focusing Analysis of RPLTT and the C25S, R26 V, and K2 7 Q variants.
The goal of the isoelectric focusing (IEF) studies was to determine the state of the enzyme
that afforded protection from iodoacetamide. In the IEF studies, we used iodoacetic acid
(IAA) to modify the enzymes instead of IAM. IAM would form carboxamidomethylated
enzyme which would be indistinguishable from oxidized enzyme in IEF analysis because the
net result would be the loss of the negative charge on Cys-22. For example, Yang et al (3 6)
determined the p1 of oxidized RPLTT and RPLTT treated with IAM to be identical, p1 = 8.0.
IAA and IAM behaved identically with respect to their inactivation of the DHA reductase
activities of RPLTT and the C25S, R26V, and K27Q mutant proteins (data not shown).
Using DTT to determine the p1 of the reduced state of each enzyme and HED to determine
the oxidized state of each enzyme gave similar results to those seen by Yang and Wells (25).
Reduced RPLTT (Figure 10, lane 3) and the C25S (Figure 11, lane 3), R26V (Figure 12, lane
3), K27Q (Figure 13, lane 3) variants, yielded pls of 7.0, 7.0, 5.7, and 5.9 respectively.
Oxidized RPLTT (Figure 10, lane 5) and the R26V (Figure 12, lane 5) and K27Q (Figure 13,
lane 5) mutant proteins, yielded pls of 8.0, 7.0, and 7.0 respectively. Oxidized RPLTT and
the R26V and K27Q variants yield more basic pls because of the formation of an
intramolecular disulfide bond between Cys-22 and Cys-25. HED treated C258 protein
focused to pls of 7 .0, 7.3 and 8.0 which corresponds to reduced enzyme (pl 7.0) and enzyme
with a mixed disulfide with HED (Figure 11, lane 5).

IAA treatment of each enzyme allowed us to trap the reduced form of the enzyme
because the process of isoelectric focusing was itself slightly oxidizing. Untreated RPLTT
(Figure 10, lane 1) and the R26V (Figure 12, lane 1) and K27Q (Figure 13, lane 1) proteins

focused to pls corresponding to both reduced and oxidized species, whereas C258 enzyme,

112

63

th

pr

in

or

31".

lacking the ability to form an intramolecular disulfide bond, focused only to a reduced p1
(Figure 11, lane 1). When treated with IAA afier a 30 minute preincubation in buffer,
RPLTT focused to bands of 5.9 and 4.9 (Figure 10, lane 2), the C25S variant focused to a pI
of 6.0 (Figure 1 1, lane 2), R26V protein focused to bands of 4.4, 4.9, and 5.7 (Figure 12, lane
2), and K27Q enzyme focused to bands of 4.6, 4.9, 5.0, 5.9, and 7.0 (Figure 13, lane 2). For
each enzyme pretreatment with DTT prior to 1AA treatment yielded similar results except
that reduced enzyme (p1 = 7.0) was also seen with RPLTT (Figure 10, lane 4) and C25S
protein (Figure 1 1, lane 4). In each case, the addition of IAA to the reduced enzyme resulted
in species which are unique from the reduced or oxidized form of the enzyme, and trapped
the reduced form of the enzyme preventing oxidation of enzyme during IEF itself.
Preincubation with HED, DHA, DHA + GSH, or GSSG protected against the IAM
inactivation of the DHA reductase activity of RPLTT and the R26V and K27Q mutant
proteins whereas preincubation with GSH or AA did not. In agreement with the IAM
inactivation studies, preincubation of R26V (Figure 12) or K27Q (Figure 13) proteins with .
HED (lane 6), DHA (lane 8), DHA + GSH (lane 10), or GSSG (lane 11) prior to IAA
addition resulted in the formation of oxidized enzyme (p1 = 7.0). The predominant form of
RPLTT (Figure 10) when preincubated with HED (lane 6), DHA + GSH (lane 10), or GSSG
(lane 11) was oxidized enzyme (pI 8.0), protecting against the formation of IAA modified
species, although a small amount of p1 = 4.9 species was seen when RPLTT was
preincubated with HED. The primary form of RPLTT (Figure 10), when preincubated with
DHA prior to IAA treatment (lane 7), was at pI 7.3, possibly indicating the trapping of an
enzyme-substrate intermediate. These results confirm that the formation of oxidized RPLTT

and its R26V and K27Q variants results in the protection of the enzyme from IAA

113

modification and IAM inactivation.

Preincubation of RPLTT (Figure 10) or the R26V enzyme (Figure 12) with GSH
(lane 9) or AA (lane 12) prior to IAA treatment resulted in the appearance of reduced enzyme
( p1 = 7.0 for RPLTT and pl = 5.7 for R26V protein) and IAA modified enzyme (pls = 4.9
and 5.9 for RPLTT and pls = 4.4 and 4.9 for R26V protein). In contrast to RPLTT and R26V
protein, K27Q enzyme treated with IAA (Figure 13, lane 2)and K27Q enzyme preincubated
with AA prior to IAA treatment (Figure 13, lane 12) yielded some oxidized enzyme,
indicating that the prepared reduced K27Q enzyme contained some oxidized enzyme.
Preincubation of the K27Q variant with GSH yielded species with pIs = 6.0 and 5.0 (Figure
13, lane 9) indicating the reaction of enzyme with IAA and supporting the finding that GSH
enhances the IAM inactivation of the DHA reductase activity of K27Q enzyme.

The DHA reductase activity of the C25S protein was protected from IAM inactivation
by HED, DHA + GSH, or GSSG but not by GSH, DHA, or AA. In agreement with these
results, when the C25S enzyme (Figure 11) was preincubated with GSH (lane 9), AA (lane
12), and DHA (lane 8) both reduced enzyme (pl = 7 .0) and the band indicative of
carboxymethylation (pl = 6.0) appeared in significant quantities. When the C25S mutant
protein was preincubated with HED (lane 6) or DHA + GSH (lane 10) prior to IAA
treatment, no pI 6.0 species was seen. Some 6.0 species was seen when the C25S enzyme
was preincubated with GSSG (lane 11) even though GSSG afforded complete protection
against IAM inactivation of the DHA reductase activity of C25S enzyme (Figure 6).

As shown in Figure 9, reduced RPLTT and the R26V and K27Q variants were
capable of chemically reducing DHA to AA in the absence of GSH. When reduced RPLTT

(Figure 10, lane 7) and the R26V (Figure 12, lane 7) or K27Q (Figure 13, lane 7) mutant

114

proteins were incubated with DHA alone, oxidized enzyme (p1 = 8.0 for RPLTT and p1 = 7.0
for the R26V and K27Q enzymes) was formed. This is consistent with the ability of reduced
RPLTT or the R26V and K27Q variants to chemically reduce DHA forming AA and
oxidized enzyme (Figure 4, reactions 1 and 3). Previously, DHA had been shown to be a

thiol oxidant of reduced bovine pancreatic ribonuclease (43-46).

115

:T‘

 

 

 

 

Figure 10. Isoelectric focusing analysis of incubation of reduced RPLTT with various
substrates. Reduced RPLTT (12ug) was separately incubated alone (lane 2), with 1 mM
DTT (lane 4), 1 mM HED (lane 6), 1 mM DHA (lane 8), 1 mM GSH (lane 9), 0.5 mM DHA
plus 1 mM GSH (lane 10), 0.5 mM GSSG (lane 11), or 0.5 mM AA (lane 12) in 40 mM
sodium phosphate, pH 7.5, for 30 min in a 30°C water bath. IAA was then added to a final
concentration of 2 mM and the samples incubated for 30 additional minutes at 30°C. No
IAA controls included untreated enzyme (lane 1), enzyme treated with 1 mM DTT (lane 3),
1 mM HED (lane 5), 1 mM DHA (lane 7). Isoelectric point standards are 5111 of Serva pI
test mix. The p] value for each of the reaction mixtures was measured on a Servalyt Precote
pH 3-10 isoelectric focusing gel according to the manufacturer’s instructions. The gel
shown is typical of two or more analyses.

116

LELNELA, -§_-.§,,.J__,_8 9 _

   

- “a. a; U- in» v1!” ‘1’

 

 

 

Figure 11. Isoelectric focusing analysis of incubation of reduced C25S with various
substrates. Reduced C25S (12ug) was separately incubated alone (lane 2), with 1 mM DTT
(lane 4), 1 mM HED (lane 6), 1 mM DHA (lane 8), 1 mM GSH (lane 9), 0.5 mM DHA plus
1 mM GSH (lane 10), 0.5 mM GSSG (lane 1 1), or 0.5 mM AA (lane 12) in 40 mM sodium
phosphate, pH 7.5, for 30 min in a 30°C water bath. 1AA was then added to a final
concentration of 2 mM and the samples incubated for 30 additional minutes at 30°C. No
IAA controls included untreated enzyme (lane 1), enzyme treated with 1 mM DTT (lane 3),
1 mM HED (lane 5), 1 mM DHA (lane 7). Isoelectric point standards are Sal of Serva pl
test mix. The pl value for each of the reaction mixtures was measured on a Servalyt Precote
pH 3-10 isoelectric focusing gel according to the manufacturer’s instructions. The gel
shown is typical of two or more analyses.

117

 

   
 
  

123 456789101112

- va w I“!

Figure 12. Isoelectric focusing analysis of incubation of reduced R26V with various
substrates. Reduced R26V (12ug) was separately incubated alone (lane 2), with 1 mM DTT
(lane 4), 1 mM HED (lane 6), 1 mM DHA (lane 8), 1 mM GSH (lane 9), 0.5 mM DHA plus
1 mM GSH (lane 10), 0.5 mM GSSG (lane 11), or 0.5 mM AA (lane 12) in 40 mM sodium
phosphate, pH 7.5, for 30 min in a 30°C water bath. IAA was then added to a final
concentration of 2 mM and the samples incubated for 30 additional minutes at 30°C. No -
IAA controls included untreated enzyme (lane 1), enzyme treated with 1 mM DTT (lane 3),
1 mM HED (lane 5), 1 mM DHA (lane 7). Isoelectric point standards are 5,121 of Serva pl
test mix. The pl value for each of the reaction mixtures was measured on a Servalyt Precote
pH 3—10 isoelectric focusing gel according to the manufacturer’s instructions. The gel
shown is typical of two or more analyses.

118

 

4:112 345678 9101112
y~ ivu- In";

na-fl- . “"

 

 

 

Figure 13. Isoelectric focusing analysis of incubation of reduced K27Q with various
substrates. Reduced K27Q (12,ug) was separately incubated alone (lane 2), with 1 mM DTT
(lane 4), 1 mM HED (lane 6), 1 mM DHA (lane 8), 1 mM GSH (lane 9), 0.5 mM DHA plus
1 mM GSH (lane 10), 0.5 mM GSSG (lane 11), or 0.5 mM AA (lane 12) in 40 mM sodium
phosphate, pH 7.5, for 30 min in a 30°C water bath. IAA was then added to a final
concentration of 2 mM and the samples incubated for 30 additional minutes at 30°C. No
IAA controls included untreated enzyme (lane 1), enzyme treated with 1 mM DTT (lane 3),
1 mM HED (lane 5), 1 mM DHA (lane 7). Isoelectric point standards are So] of Serva pl
test mix. The p1 value for each of the reaction mixtures was measured on a Servalyt Precote
pH 3-10 isoelectric focusing gel according to the manufacturer’s instructions. The gel
shown is typical of two or more analyses.

119

Proposed fimctions of R26 and K27 in thioltransferase. The DHA reductase activity
of both the R26V and K27Q variants were incompletely inhibited by IAM with 40% residual
activity after 45 minutes. GSH only enhanced the IAM inactivation of the DHA reductase
activity of the K27 Q protein but not that of the R26V protein. This suggested that the K27 Q
variant was protected by the presence of oxidized enzyme whereas the R26V mutant protein
was protected by Cys-22 protonation. The R26V variant (Figure 12) had no oxidized protein
present when reacted with IAA (lane 2) or when preincubated with AA (lane 12). Thus
R26V protein is afforded protection from IAM inactivation and IAA modification because
Cys-22 is protonated. This evidence further supports the conclusion that Arg-26 is the amino
acid primarily responsible for the exceptionally low pK, (3.8) of Cys-22 (25). Although Cys-
22 in the K27Q protein has a slight increase in pK1i of 4.3 vs. 3.8 for RPLTT, the role of Lys-
27 was previously unknown (25). The current studies suggest that the role of lysine at
position 27 is to stabilize the reduced form of RPLTT. In contrast to the other enzymes,
when the K27Q variant was treated with IAA (Figure 13, lane 2) or preincubated with AA
prior to IAA addition (Figure 13, lane 12), a significant amount of oxidized enzyme was
seen. Finally, K27Q protein had the greatest initial rate of chemical reduction of DHA to
AA, a reaction which includes the oxidation of the active site sulflrydryls (reaction 3, Figure
4). These data suggest that the K27 Q enzyme is much more readily oxidized than RPLTT
or the R26V mutant protein indicating a role of lysine at position 27 in the stabilization of
the reduced form of the enzyme.

CONCLUSIONS
The work described in this paper supports the catalytic mechanism proposed in Figure

4. When starting with the reduced form of the enzyme, IAM inactivation studies

120

. In
I

”Am—‘1”
2.1
IA
.

 

demonstrated the ability of DHA to bind to enzyme (reaction 1, Figure 4), prior to GSH
binding (reaction 3 or 4, Figure 4) because DHA protected against IAM inactivation whereas
GSH did not. The ability of RPLTT and the R26V and K27Q variants to chemically reduce
DHA to AA in the absence of GSH (reactions 1 and 3, Figure 4) was demonstrated by
electrochemical analysis and IEF. In addition, IEF analysis demonstrated the ability of
reduced RPLTT and the R26V and K27Q variants to react with GSSG and form oxidized
enzyme (reverse of reactions 5 and 4 in Figure 4). The relative inability of the C258 protein
to chemically reduce DHA to AA, but its robust DHA reductase activity strongly suggest that
reaction 2 (Figure 4) is a part of this mechanism. Further studies in progress involve the
detection of the thiohemiketal intermediate and the study of its catalytic capacity. In
addition, in the presence of GSH, whether reaction 2 or 3 predominates in the mechanism
of RPLTT and the R26V and K27Q variants has yet to be determined. Previously proposed
models of the DHA reductase activity of thioltransferase (22, 23, 26, 27, 36) and for the
DHA reductase activity of T rypanosoma cruzi p52 (47) have not included reactions 1 and
3 in Figure 4, but the work described in this paper clearly demonstrates the potential of

thioltransferase to undergo these reactions.

121

‘

. a“. . ...'

[:1

 

REFERENCES

(1) Frei, B., England, L., and Ames, B.N. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 6377-
6381.

(2) Halliwell, B., and Gutteridge, J.M.C. (1990) Arch. Biochem. Biophys. 280, 1-8.
(3) Buettner, GR. (1993) Arch. Biochem. Biophys. 300, 535-543.

(4) Sciamanna, M.A., and Lee, GP (1993) Arch. Biochem. Biophys. 305, 215-224.
(5) May, J.M., Qu, Z., and Whitesell, RR. (1995) Biochemistry 34, 12721-12728.

(6) Ozaki, M., Fuchinoue, S., Teraoka, S., and Ota, K. (1995) Arch. Biochem. Biophys. 318,
439-445.

(7) Englard, S., and Seifter, S. (1986) Ann. Rev. Nutr. 6, 365-406.
(8) Lyons, BL, and Schwarz, RI. (1984) Nucleic Acids Res. 12, 2569-2579.
(9) Levine, M., Morita, K., and Pollard, H. (1985) J. Biol. Chem. 260, 12942-12947.

(10) Wells, W.W., Dou, C.Z., Dybas, L.N., Jung, C.H., Kalbach, H.L., and Xu, D.P.
(1995) Proc. Natl. Acad. Sci. U.S.A. 92, 11869-11873.

(11) Jung, CH. and Wells, W.W. (1997) Biochem. Biophys. Res. Commun. 239, 457-462.

(12) Bielski, B.H.J. (1982) In Ascorbic Acid: Chemistry, Metabolism, and Uses (Seib, P.A.
and Tolbert, B.M., Eds.) Advances in chemistry Series 200 (Comstock, M.J., Ed.), pp.
81-100, American Chemical Society, Washington, DC.

(13) Coassin, M., Tomasi, A., Vannini, V., and Ursini, F. (1991) Arch. Biochem. Biophys.
290, 458-462.

(14)Vi11alba, J.M., Canalejo, A., Buron, M.I., C6rdoba, F., and Navas, P. (1993) Biodrem
Biophys. Res. Commun. 192, 707-713.

(15) Bando, M., and Obazawa, H. (1994) Jpn. J. Ophthalmol. 38, 1-9.

(16) Del Bello, B., Maellaro, E., Sugherini, L., Santucci, A., Comporti, M., and Casini, A.F.
(1994) Biochem J. 304, 385-390.

(17) May, J.M., Mendiratta, s., Hill, K.E., and Burk, RF. (1997) J. Biol. Chem. 272, 22607-
22610.

122

 

(18) Wells, W.W., Xu, D.P., Yang, Y., and Rocque, P.A. (1990).]. Biol. Chem. 265, 15361-
15364.

(19) Maellaro, E., Del Bello, B., Sugherini, L., Santucci, A., Comporti, M., and Casini, A.F.
(1994) Biochem J. 301 , 471-476.

(20) Xu, D.P., Washbum, M.P., Sun, GR, and Wells, W.W. (1996) Biochem. Biophys. Res.
Commun. 221, 117-121.

(21) Meister, A. (1994) J. Biol. Chem. 269, 9397-9400.
(22) Wells, W.W., and Xu, D.P. (1994).]. Bioenerg. Biomem. 26, 369—377.

(23) Wells, W.W., and Jung, CH. (1997) in Vitamin C in Health and Disease (Packer, L. P
and Fuchs, J. Eds.) pp. 109-121, Marcel Dekker, Inc., New York, NY. '

(24) Park, J.B., and Levine, M. (1996) Biochem. J. 315, 931-938. 11.
(25) Yang, Y., and Wells, W.W. (1991) J. Biol. Chem. 266, 12759-12765.

(26) Wells, W.W., Yang, Y., Deits, TL, and Gan, Z. (1993) Adv. Enzymol. Relat. Areas
Mol. Biol. 66, 149-201.

(27) Mieyal, J .J ., Gravina, S.A., Mieyal, P.A., Srinivasan, U., and Starke, D.W. (1995) in
Biothiols in Health and Disease (Packer, L., and Cadenas, E., Eds.) pp. 305-372, Marcel
Decker, Inc., New York, NY.

(28) Washbum, MP. and Wells, W.W. (1997) Free Rad. Biol. Med. 23, 563-570.

(29) Gan, Z-R., and Wells, W.W. (1986) J. Biol. Chem. 261, 996-1001.

(30) Wells, W.W., Xu, D.P., and Washbum, MP. (1995) Methods. Enzymol. 252, 30-38.

(31) Segel, I.H., and Johnson, M.J. (1963) Anal. Biochem. 5, 330-337.

 

(32) Yang, Y., and Wells, W.W. (1990).]. Biol. Chem. 265, 589-593.

(33) Sarnbrook, J ., Fritsch, ER, and Maniatis, T. (1989) Molecular Cloning: A Laboratory
Manual, 2“d ed., Cold Spring Harbor Laboratory Press, Plainview, New York.

(34) Xu, D.P. and Wells, W.W. (1996) J. Bioenerg. Biomem. 28, 77-85
(35) Perrella, F.W. (1988) Anal. Biochem. 174, 437-447.

(36) Yang, Y., and Wells, W.W. (1991).]. Biol. Chem. 266, 12766-12771.

123

(37) Gravina, S.A., and Mieyal, J .J . (1993) Biochemistry 32, 3368-3376.

(38) Jung, C.H., and Thomas, J .A. (1996) Arch. Biochem. Biophys. 335, 61-72.

(39) Gan, Z-R., and Wells, W.W. (1987) J. Biol. Chem. 262, 6704-6707.

(40) Katti, S.K., Robbins, A.H., Yang, Y., and Wells, W.W. (1995) Prot. Sci, 4, 1998-2005.
(41) Jung, CH. and Wells, W.W. (1998) personal communication.

(42) Srinivasan, U., Mieyal, P.A., and Mieyal, J .J . (1997) Biochemistry 36, 3199-3206.
(43) Venetianer, P. and Straub, F .B. (1963) Acta Physiol. Acad. Sci. Hung, 24, 41-53.

(44) Givol, D., Goldberger, RR, and Anfinsen, CB. (1964) J. Biol. Chem, 239, PC3114-
PC3116.

(45) Krause, E.G., Venetianer, P., and Straub, EB. (1965) Acta Physiologica, 27, 295-302.
(46) Venetianer, P. and Straub, EB. (1965) Acta Phsyiologica, 27, 303-315.

(47) Moutiez, M., Quéméneur, E., Sergheraert, C., Valerie, L., Tartar, A., and Davioud-
Charvet, E. (1997) Biochem. J., 322, 34-48.

124

 

Chapter IV

Glutathione Dependent Reduction of Alloxan to Dialuric Acid Catalyzed by

Thioltransferase (Glutaredoxin): a Possible Role for Thioltransferase in Alloxan

Toxicity

The majority of this chapter has been published: Washbum, MP. and Wells, W.W. (1997)

Free Rad. Biol. Med. 23, 563-570.

125

ABSTRACT

Recombinant pig liver thioltransferase (RPLTT) catalyzes the reduction of alloxan to dialuric

acid by glutathione (GSH). This is the second non-disulfide substrate, afier dehydroascorbic

acid, described for thioltransferase. The reaction kinetics, measured by a coupled assay

including glutathione disulfide reductase and NADPH, yielded a Km = 82 MM for alloxan,

a kc," = 37 sec“, and a kw/Km = 4.5 x 105 M“sec". The presence of RPLTT suppressed the

competitive formation of compound 305, an alloxan-GSH conjugate of unknown structure, _1

and at GSH concentrations between 0.05 mM and 1.5 mM, oxygen consumption was greater 3
1

than that recorded in the uncatalyzed reaction. Both superoxide dismutase and catalase ‘3

inhibited oxygen consumption in 1.0 mM GSH and 0.2 mM alloxan in the presence of

RPLTT. This study suggests that thioltransferase (glutaredoxin) plays a significant role in

the cytotoxicity of alloxan in vulnerable tissues.

 

 

126

 

Alloxan (2,4,5,6 [1 H,3 H] pyrimidinetetrone) is selectively toxic to the 2-cells of the
pancreas of certain animals, e.g., mice and rats (1-3). This toxicity has been attributed in part
to accumulation of alloxan in the islets (4, 5) and 2-cells (6). Furthermore, the pancreas
possesses relatively lower levels of enzymes such as catalase, Cu-Zn superoxide dismutase
(SOD), Mn-SOD, and glutathione (GSH) peroxidase, which protect cellular components
against reactive oxygen species (5, 7). The toxic mechanism of alloxan involves redox
cycling between alloxan and dialuric acid. Reduction of alloxan to dialuric acid by
physiological reagents has been demonstrated with NAD(P)H (8, 9), GSH (1 0-13), and
enzymatically by the thioredoxin/NADPH-thioredoxin reductase system (EC. 1.6.4.5) (I4).
Dialuric acid oxidation generated superoxide, hydrogen peroxide, and hydroxyl radical, in
the presence of an appropriate metal catalyst like iron (15-18) Specifically, reaction of
alloxan with GSH generated superoxide and hydrogen peroxide (10-13), hydroxyl radical in
the presence of iron (18), and released iron from ferritin and subsequently formed hydroxyl
radical (19). Evidence supporting this mechanism includes protection against alloxan
toxicity in isolated islets by SOD, catalase, hydroxyl radical scavengers, metal chelators (20-
22), and in mice by SOD (23).

The chemical reaction of alloxan and GSH yields two distinct Spectrophotometric
products, dialuric acid and compound 305 (IO-13). Dialuric acid, the reduced form of
alloxan, has a characteristic absorbance at 275 nm (13) and is rapidly formed upon the
addition of GSH to alloxan (10). Compound 305 is a more gradually appearing alloxan—GSH
conjugate of undetermined structure which absorbs maximally at 305 nm (10, 12). The
formation of compound 305 is in competition with the reduction of alloxan to dialuric acid

(10, 12), and the production of dialuric acid is favored when GSH is in excess of alloxan (10,

127

 

 

12).

Thioltransferase (EC 1 .8.4. 1) (glutaredoxin), a 12 kDa enzyme, catalyzes GSH-
dependent thiol-disulfrde exchange reactions (24) Furthermore, this enzyme isolated from
human placenta, bovine thymus, and recombinant pig liver thioltransferase (RPLTT) catalyze
the GSH-dependent reduction of dehydroascorbic acid (DHA) to ascorbic acid (25). The
active site of RPLTT has been determined to be -Cy322-Pro23-Phe24-Cy825-, with Cy322
the essential amino acid for both thioltransferase and DHA reductase activity (26).

Thioredoxin, shown to oxidize NADPH in the presence of thioredoxin reductase and
alloxan (14) shares a similar active site to that of thioltransferase, -CXXC- (24, 27).
Furthermore, alloxan and DHA and their reduction products, dialuric acid and ascorbic acid,
have similar structures (Figure 1). Based on thioredoxin/NADPH—thioredoxin reductase
system reactivity with alloxan and thioltransferase catalysis of DHA reduction, we examined
thioltransferase as a GSH-dependent alloxan reductase and tested the potential role of
thioltransferase in alloxan’s cytotoxicity.

MATERIALS AND METHODS

Materials. GSH, alloxan monohydrate, bovine erythrocyte SOD (4400 units/mg
protein), and bovine liver catalase (1600 units/mg protein) were purchased from Sigma
Chemical Company (St. Louis, MO USA). NADPH, dithiothreitol (DTT), bovine serum
albumin (BSA), and yeast glutathione disulfide (GSSG) reductase were from Boehringer
Mannheim Corporation (Indianapolis, IN USA). RPLTT was purified as described
previously (28). In all of the following studies RPLTT was stored in 20% glycerol, 20 mM
sodium phosphate, pH 7.5. Reduced RPLTT, C25S, and K27Q for IEF were prepared in the

absence of glycerol as described in Chapter III. GSSG reductase stock solution was 1.25

128

* .- .3; inns-mi,"

 

Oxidized Forms:

Alloxan

Reduced Forms:

0 H
>—l\(
H—N>—2=O
HO OH

Dialuric Acid

Figure 1. Structural comparison of alloxan to dehydroascorbic acid

CHZOH
H OH O
O
H
O O

Dehydroascorbic Acid

CHZOH

O
H
HO OH

L-Ascorbic Acid

and dialuric acid to L-ascobic acid

129

f3 # j‘" 9;. 7‘3- '

mg/ml in 1.3 M (NH4)2SO4. Stock solutions of alloxan were in N2 sparged 1 mM HCl and
kept on ice. This method of preparation and storage maintained alloxan for up to 1.5 h.

Kinetic characterization of RPLTT as an alloxan reductase. Kinetic analysis of
RPLTT catalysis of the GSH-dependent reduction of alloxan to dialuric acid was based on
the following assay system.

Thioltransferase

Alloxan + 2GSH -—‘ Dialuric Acid + GSSG

GSSG Reductase
GSSG + NADPH + H+ -——- 2GSH + NADP“

Initial reaction velocities were compared at various pH values using 200 mM sodium
phosphate, 1 mM EDTA buffers, between pH 5.5 and 7.5, and 200 mM Tris-HCl, 1 mM
EDTA buffers, between pH 8.0 and 9.5. The optimal assay conditions consisted of 1.0 mM
GSH, 0.5 mM NADPH, 8.75 ug GSSG reductase, 200 mM sodium phosphate, 1 mM EDTA,
pH 7.5 in a 500 ill reaction volume. This assay is similar to those used previously for
thioltransferase studies (29, 30). Kinetic studies were carried out in the presence of 0.375
ug RPLTT. The reaction was initiated upon the addition of varying amounts of alloxan and
followed as a decrease in absorbance at 340 nm at 25°C for 1.5 min. A blank without
enzyme was run simultaneously for each assay and the difference ((CAbsm nm/min) x 0.5 x
1000/ 6.22) gave activity in nmoles/min. Using this assay, the initial velocity of RPLTT was
linear over a range up to 2 mg protein/ml.

Spectrophotometric study of uncatalyzed and RPL TT catalyzed reactions. The effect

of RPLTT on the alloxan and GSH reaction from 250 nm to 320 nm was monitored (scan run

130

 

at 2 nm/sec). GSH (1.5 mM) and alloxan (0.2 mM) in 200 mM sodium phosphate, 1 mM
EDTA, pH 7.5, were mixed in a final volume of 500 pl in the presence or absence of 1.25
ug RPLTT at 23 °C. The difference in uncatalyzed and RPLTT catalyzed production of
dialuric acid increased linearly up to 2.5 ug/ml RPLTT. Spectra were taken at multiple time
points up to 50 min afier initiating the reaction by addition of alloxan.

The extinction coefficient for dialuric acid was determined to be 14,500 M"cm‘l at
275 nm by taking the absorbance multiple times of 50 MM alloxan reacted with 1.0 mM DTT
in 200 mM sodium phosphate, 1 mM EDTA, pH 7.5. The reaction was initiated by the
addition of alloxan and analyzed from 250 nm to 320 nm until reaching a peak absorbance
value at 275 nm.

Oxygen consumption during the GSH-dependent reduction of alloxan. Oxygen
consumption was determined at 25 °C using a YSI Model 533] oxygen probe and a YSI
Model 53 biological oxygen monitor. Baseline oxygen uptake was measured in a 3 ml
reaction volume of 200 mM sodium phosphate, 1 mM EDTA, pH 7.5, in the presence or
absence of 7.5 ag RPLTT and various concentrations of GSH. Alloxan was then added to
a final concentration of 0.2 mM and maximal rates of oxygen consumption were measured.
In experiments investigating the effects of SOD and catalase, the enzymes were dissolved
in 100 mM sodium phosphate, 0.5 mM EDTA, pH 7.5, and added prior to baseline oxygen
uptake measurement. Oxygen consumption in the presence of bovine serum albumin (BSA,
2.5 pig/ml) was measured as a control for non-specific protein interactions. Oxygen
consumption was measured on a Linseis Flatbed Recorder model L6522 B. Statistical
analysis was carried out with the GraphPAD InStat program using a one way analysis of

variance with selected comparisons subjected to the Bonferroni post hoc test.

131

I"—

Isoelectric Focusing. Reduced RPLTT and its C258 and K270 variants, 12 Mg each,
were separately incubated alone or with 1 mM alloxan or 0.5 mM alloxan + 1.0 mM GSH
in 40 mM sodium phosphate, pH 7.5, for 30 min in a 30°C water bath. Each of the
incubation mixtures (in a total volume of 10 al) was then analyzed directly on a Servalyt
Precotes isoelectric focusing gel, pH 3-10 (150 mm thick, 125 mm x 125 mm), following the
manufacturer’s instructions for 2666 volt hours on a LKB 2217 Ultraphor electrofocusing
unit with a Lauda K-2/R cooling system. After focusing, the gel was fixed with a solution
of 30% ethanol and 10% glacial acetic acid, stained with a solution of 0.001% Coomassie
Brilliant Blue R-250, 10% glacial acetic acid, and 50% ethanol, and washed with the fixing
solution.

RESULTS

Spectrophotometric study of the eflect of RPL T T on the GSH reduction of alloxan.
At 22 sec after reaction initiation, the uncatalyzed reaction yielded 0.1 1 :1: 0.01 mM dialuric
acid measured spectrophotometrically at 275 nm indicating a 55% reduction of alloxan to
dialuric acid (Figure 2A). The RPLTT catalyzed reaction yielded 0.16 :1: 0.01 mM dialuric
acid, equivalent to 82% reduction of alloxan to dialuric acid, 45% more dialuric acid than
in the uncatalyzed reaction (Figure 2A). By 4 min, the reduction of alloxan to dialuric acid
in both the RPLTT catalyzed and the uncatalyzed reaction was 90%-95% complete (Figure
2B). From 4-50 min, the levels of dialuric acid in the RPLTT catalyzed and the uncatalyzed
reaction were not significantly different and remained 90%-95% complete, never reaching
100% completion (Figure 2C—2D). The absorbance at 305 nm for the uncatalyzed and
RPLTT catalyzed reactions at various time points are plotted in Figure 3. The rate of

formation of compound 305 from 0-4 min was faster in the uncatalyzed reaction mixture than

132

._ 71

t? i—

I
t

 

in the RPLTT catalyzed reaction mixture. By 2 min, a significant difference in absorbance
at 305 nm occurred; the uncatalyzed reaction had an absorbance of 0.40 i 0.12 greater than
the RPLTT catalyzed reaction, and this trend continued for 50 min. However, from 4-50 min,
the rate of production of compound 305 was similar in both reaction mixtures.

Kinetics. The coupled assay system was chosen because under certain alloxanzGSH
ratios necessary for kinetic characterization, compound 305 generation interfered with
accurate measurement of dialuric acid formation at 275 nm, whereas no interference occurred
at 340 nm. Although NADPH does reduce alloxan to dialuric acid at a low rate (8, 9) this
was subtracted as part of the control without RPLTT, and RPLTT did not utilize NADPH
as a substrate. Up to 1 mM alloxan did not affect GS SG reductase activity (data not shown).
Optimal pH for the coupled assay was in the range of 7.5-8.0 (Figure 4). The kinetic profile
yielded a sigmoidal curve with inhibition at high substrate concentration (Figure 5). The best
fit was to the Hill equation (n=2.7) (EZ-Fit kinetic program (31)). The kinetic parameters
of the GSH dependent reduction of alloxan are reported in Table 1. Based on these kinetic
values, alloxan compares favorably with DHA, hydroxyethyldisulfide, and S-sulfocysteine

as a thioltransferase substrate.

133

1.
L1:

 

3.0 - 3.0 -

 

 

 

l l 214;; AL 1“... 00
250 250 270 28) 290 300 310 320 250

W (rm)

 

 

 

 

 

0.0b ‘4 i ‘ l ‘ L *4 AJé—j 00 LHJQI. JkALHJ
250 260 270 280 290 300 310 320 250 260 270 2% 29) 300 310 320

Wm“) mam)

 

Figure 2. Spectrophotometric (250 nm - 305 nm) time course study of the effect of RPLTT
on the alloxan and GSH reaction. GSH (1.5 mM) and 0.2 mM alloxan in 200 mM sodium
phosphate, 1 mM EDTA, pH 7.5, were mixed in a final volume of 500 ill in the presence
(solid lines) or absence (dashed lines) of 1.25 ag RPLTT] at 23°C and analyzed from 250
nm to 320 nm. Spectra were taken at multiple times over 50 min. The spectra were taken
at A) 22 see (first opportunity for measurement), B) 2 min, C) 4 min, and D) 30 min. Each
chemical and enzymatic spectra shown at each time point are the average of three or four

separate experiments. Error bars represent one standard deviation of the data at 275 nm and
305 nm.

134

3"
I

 

_ u‘ _ A-'fi"‘

 

2.5 -

2.01

1.5-é §/§/

Abeorbmce (3051111)
1-0-1
\
l-O—l
\

 

 

Figure 3. Time course study of the effect of RPLTT on the alloxan and GSH reaction at 305
nm. GSH (1.5 mM) and 0.2 mM alloxan in 200 mM sodium phosphate, 1 mM EDTA, pH
7 .5, were mixed in a final volume of 500 a] in the presence (-O-) or absence (-O-) of 1.25
ug RPLTT at 23°C and analyzed from 250 nm to 320 nm. Spectra were taken at multiple
times over 50 min. Data points are the averages of three or four separate experiments. Error
bars represent one standard deviation of the data at 305 nm and are not seen if the values are
smaller than the size of the symbols.

135

 

 

140-: /§\
1203
j? : I

103-j

80%

«)i / /$/ \9

20: 0 e \
:gée/ Q

0‘ I ' I j I i I ' I ' I j I ' I ' I
5.5 6.0 6.5 7.0 7.5 so as 9.0 9.5
d-I

Figure 4. pH profile of the GSH-dependent alloxan reductaes activity of RPLTT. RPLTT
(0.7491 ug) was assayed for alloxan reductase activity in 1.0 mM GSH, 0.34 mM NADPH,
5 ug GSSG reductase, and various buffer conditions in a 500 ul reaction volume. The
reaction was initiated by the addition of alloxan and was followed at 340 nm at 25°C for 15
min. The pH was varied as shown using 200 mM sodium phosphate containing lmM EDTA
between pH 5.5 and 7.5., and 200 mM Tris-HCl with lmM EDTA between pH 8.0 and 9.5.
Reaction rates in the presence (-O-) or absence (-O-) of RPLTT are shown. Data points are
the averages of at least three separate experiments. Error bars represent one standard
deviation of the data and are not seen if values are smaller than the size of the symbols.

136

 

 

 

100-
80 ¢
A . O>O‘—\——O‘—_O
.r: . O/ . o\'
E 604 _, O
E 40?
i 20,
d 1/
.IO
0% .............................
0o 01 02 03 04 05 06

Figure 5. Kinetic analysis of GSH dependent alloxan reductase activity of RPLTT. RPLTT
(0.375 Mg) was assayed for alloxan reductase activity in 1.0 mM GSH, 0.5 mM NADPH,
8.7 5 ug GSSG reductase, 200 mM sodium phosphate, 1 mM EDTA, pH 7.5, in a 500 ul
reaction volume. The reaction was initiated by the addition of alloxan and followed at 340
nm at 25°C for 1.5 min. Alloxan concentration was varied as shown and initial rates
calculated. Data points (-O-) are the averages of four separate experiments. Error bars
represent one standard deviation of the data and are not seen if the values are smaller than
the size of the symbols. Kinetic parameters were obtained via the Hill equation. The Hill
equation (-O-) (n = 2.7) yielded the best fit of a wide range of kinetic models tested in the
EZ-FIT kinetic program.

137

.—.‘- “in“, A- -

 

 

, .Uoom 8 E: own an

58 m 8m 326:8 can ova—8635028??— .8 o:_3m>oo.:=m-m Mo 92:22“ 05 com: 3855 803 28388 05 £822? Emma 805 5

2.23.2 :5 8d e5 .8883." 030 m: m .952 w: 3c .38 28 3 832: some. has one 56% é Basso case
05 8 .8266 8393 taxman 8338 a mum: 358.223 803 camoumzootsmm use ova—36350283: com 3:838 235M a.
.ANMV E: 0ch an Eon 05.80313 22:28 oEoESocgoboomm 826 a was» 820883 265V— 965 5890 8 com: 808% >33 .2

2.25:on 030:2 :m-Nm 2.: mafia BEE—Bow

0.83 25338 23:5 28 gum u a mo 320508 a 5:5 :23ch Ex 05 :23 22> E “mom .58 m4 com Down 3 E: Sm um 330:8

98 538—? CO 22:23 2: an @2225 33 corona.— 22. .0823 c0382 3 com a 5 .05 In Jon—hm 28 _ .anmmonm 8:68 2:.
com .898? 88 w: m; .522 28 no .50 28 on 5 egos 8988a 538% so Bane as a: $2 2152 .

 

 

.2 on ad 4 on o on em moo H G; aoaoseotam
.2 x as 4 me n H we 22 4 3m ”8236;53er
2 s. .2 on 3o 4 «.8 oo o No 85 H 36 .28 osooooeeaeoe
2 x Gd H as N H Hm mood H Sod .523
room-22 e33 foe .3 A25 EM 323%

 

$80883 2852 omfiommamb—omfi b>= mE “SEEEOoom

_ 2an

138

Oxygen consumption in the RPL T T catalyzed and uncatalyzed GSH reduction of
alloxan. Neither alloxan, RPLTT, GSH + RPLTT nor RPLTT + alloxan consumed oxygen.
By contrast, the addition of GSH to alloxan or to alloxan + RPLTT caused significant
oxygen consumption (Figure 6). The presence of RPLTF caused an increase of oxygen
consumption above the uncatalyzed reaction with 0.05 mM GSH to 1.5 mM GSH. At 2.0
mM GSH and above, oxygen consumption declined, and the uncatalyzed and RPL'IT
catalyzed oxygen consumptions converged. Maximal oxygen consumption occurred at 1.5
mM GSH where the presence of 2.50 rig/ml RPLTT was associated with an oxygen
consumption of 148 i 8 uM Oz/min. Maximal difference between the RPLTT catalyzed and
the uncatalyzed reactions (52.6 i 9.2 uM Oz/min) occurred at 0.5 mM GSH. Enhancement
of oxygen consumption was specific for RPLTT. Oxygen consumption experiments with
BSA (2.50 ,ug/ml BSA, data not shown) behaved identically to those without BSA.

Catalase and SOD inhibition of oxygen consumption in the RPLT T catalyzed and
uncatalyzed GSH reduction of alloxan. In the uncatalyzed and RPLTT catalyzed reactions,
catalase (5 ug/ml to 80 ug/ml) significantly inhibited oxygen consumption when compared
with the control (Figure 7A, p<0.0l in each case). In contrast, SOD did not inhibit oxygen
consumption in the uncatalyzed reaction, but did inhibit oxygen consumption in the RPLTT
catalyzed reaction (Figure 7B, 2.5 pig/ml to 80 ug/ml SOD, p<0.01 in each case). From 10
ug/ml to 80 ug/ml SOD there was no significant difference in oxygen consumption between
the uncatalyzed and RPLTT catalyzed reaction. Oxygen consumption in the RPLTT
catalyzed reaction with SOD (40 ng/ml) and catalase (4O pig/ml) was significantly lower than
with both SOD(40 ug/ml) alone and catalase (40 ug/ml) alone (Figures 7A and 7B,

p<0.00 l ). The addition of SOD (40 rig/ml) and catalase (40 ug/ml) to the RPLTT catalyzed

139

1.11::

A A. . [L., A'A."

4..

 

 

 

reaction was the only instance where oxygen consumption was decreased by 50%.
Furthermore, oxygen consumption in the RPLTT catalyzed reaction was not significantly
different from the uncatalyzed reaction with catalase (40 pig/ml), but it was significantly
different from the uncatalyzed reaction with SOD (40 ug/ml) + catalase (4O rig/ml) (p<0.05).
Finally, oxygen consumption in the uncatalyzed reaction with SOD (40 rig/ml) + catalase
(40 ng/ml) was not significantly different from catalase (40 ug/ml) alone, but it was
significantly lower than with SOD (40 ug/ml) alone (Figures 7A and 7B, p<0.01).

At 1.0 mM GSH and 0.2 mM alloxan, the lag time (time to maximal rates) was 25
i 2 see. In the presence of 2.50 lug/m1 RPLTT, the time of initial rates was 17 :h 2 sec. SOD
increased the lag time incrementally in the uncatalyzed reaction to 36 a; 4 sec at 80 ug/ml.
SOD had no effect on the lag time in the presence of RPLTT, and catalase had no efi‘ect on
the lag time in either case (data not shown). The maximal rate of oxygen consumption was

always completed 2 min after reaction initiation (data not shown).

140

180
-3

1404: §/§\

 

 

Max. 02 Won (plen)
§ 8‘
_ \
\.
.l-
l-O/IOJ//

Figure 6. Oxygen uptake in the uncatalyzed and RPLTT catalyzed alloxan + GSH reaction.
Oxygen consumption was determined at 25°C using a YSI Model 5331 oxygen probe and
a YSI Model 53 biological oxygen monitor. Baseline oxygen uptake was measured in a 3
ml reaction volume in 200 mM sodium phosphate, 1 mM EDTA, pH 7.5 in the presence (-
O-) or absence (-O-) of 7 .5 tag RPLTT and varied [GSH]. Alloxan was then added to a final
concentration of 0.2 mM and maximal rates of oxygen consumption calculated (values
plotted). Data points are the averages of three or four separate experiments. Error bars
represent one standard deviation of the data and are not seen if values are smaller than the
size of the symbols.

141

5l

.‘g-A‘._' :14) K
1;- .

 

Tfofz’o'a'o'Xo's'ofiéo'ioréo'

[Catalase] (LCM)

Max. 0, consumption Wain)
B 8

+03

/

*0:

"\f-

 

 

 

 

Max. 02 Consumption Wain)

Figure 7. Catalase and SOD inhibition of oxygen uptake in the uncatalyzed and RPLTT
catalyzed alloxan + GSH reaction. Oxygen consumption was determined as described in
Figure 6. Baseline oxygen uptake was measured in a 3 ml reaction volume in 200 mM
sodium phosphate, 1 mM EDTA, 1 mM GSH, pH 7.5 in the presence (-O-) or absence (-0-)
of 7.5 ug RPLTT and in the presence of varied [catalase] (A) or varied [SOD] (B). The
effect of both 40 pig/ml catalase and 40 ,ug/ml SOD on oxygen consumption is plotted on A
and B as -I- in the presence of RPLTT and -l:l- in the absence of RPLTT. Maximal rates
began 36 seconds or less after reaction initiation (80 ug/ml SOD in the chemical reaction).
The 0 pig/ml SOD and/or catalase data points are taken directly from the 1.0 mM GSH data
points in Figure 6. Data points are the averages of three or more separate experiments. Error
bars represent one standard deviation of the data and are not seen if the values are smaller
than the size of the symbols. (*Statistically different from oxygen consumption in the
absence of catalase or SOD; p<0.01 Bonferroni post hoc test.)

142

 

Isoelectric focusing of RPLT T and the C255 and K2 7Q variants. In the previous
chapter (111), reduced RPLTT and the K27Q protein but not the C258 protein were
demonstrated to have pronounced abilities to chemically reduce DHA to AA. The product
of this reaction was oxidized enzyme, which was identified via IEF. Because of the
structural similarities between alloxan and DHA, we investigated the ability of alloxan to
oxidize reduced protein. Reduced RPLTT and the C258 and K27Q variants were seperately
incubated alone or with 1 mM alloxan or 0.5 mM alloxan + 1.0 mM GSH (Figure 8). As
seen previously in Chapter III, IEF itself was slightly oxidizing with reduced RPLTT (lane
1) and reduced K27Q enzyme (lane 7) each focusing to reduced (7.0 for RPLTT and 5.9 for
K27Q enzyme) and oxidized pIs (8.0 for RPLTT and 7.0 for K27Q enzyme). Upon the
addition of alloxan to either RPLTT (lane 2) or the K27Q variant (lane 8) only oxidized
enzyme was seen (p1 = 8.0 for RPLTT and pl = 7.0 for K27Q enzyme). When RPLTT (lane
3) or the K27Q variant (lane 9) was incubated with alloxan + GSH, reduced enzyme was
primarily seen with a slight amount of oxidized enzyme seen with the K27Q variant. CZSS
enzyme incubated with alloxan resulted in focused pls of 7.0, corresponding to reduced
enzyme, and 6.0 (lane 5). When the C258 variant was preincubated with alloxan + GSH

(lane 6) it focused to pIs of 7.0 and 5.9.

143

23456789

  

8.3—>
8.0—>

735—) f
6.9-> ‘

6.0—> - i

4.5—)

 

 

Figure 8. Isoelectric focusing analysis of incubation of reduced RPLTT and the C258 and
K27Q variants. Equal amounts (12,ug) of reduced RPLTT and the C25S and K27Q variants
were separately incubated alone (lanes 1, 4, and 7, respectively), with 1 mM alloxan (lanes
2, 5, and 8, respectively), or with 0.5 mM alloxan plus 1.0 mM GSH (lanes 3, 6, and 9,
respectively). Isoelectric point standards are 5a] of Serva pI test mix. The pI value for each
of the reaction mixtures was measured on a Servalyt Precote pH 3-10 isoelectric focusing
gel according to the manufacturer’s instructions.

144

DISCUSSION

Direct spectral analysis demonstrated that RPLTT catalyzed the GSH-dependent
reduction of alloxan to dialuric acid. This is the second nondisulfide substrate, after DHA,
described for thioltransferase, and the second demonstration of enzyme catalyzed alloxan
reduction after the thioredoxin/NADPH-thioredoxin reductase system (14 ). Compared with
other substrates for RPLTT, including S-sulfocysteine, hydroxyethyldisulfide and DHA, the
kinetic parameters for alloxan represent the lowest K.,n and the largest kw/Km measured (Table
I). As in the case of alloxan, S-sulfocysteine, and hydroxyethyldisulfide, kinetic constants
were measured via a coupled assay with GSSG reductase, whereas DHA kinetic constants
were obtained in an assay monitoring the production of ascorbic acid directly (32). In the
direct assay, the best fit was to the Michaelis-Menten equation with substrate inhibition. The
kinetics for alloxan and S-sulfocysteine (29) in the coupled assay yielded a best fit to a Hill
plot with coefficients of n = 2.7 and n = 2.5, respectively. The sigmoidal nature of the curve
is most likely a side effect of the coupled assay and not a result of cooperativity.

The formation of compound 305, an alloxan-GSH conjugate of unidentified structure,
was in competition with the reduction of alloxan to dialuric acid (1 0-12). Since a greater rate
of reduction to dialuric acid was catalyzed by RPLTT between 0-4 min, there was more
alloxan available to form compound 305 in the uncatalyzed reaction. From 4—50 minutes,
however, there were equal and constant amounts of dialuric acid and alloxan in both
reactions, suggesting the reaching of reaction equilibrium with respect to GSH-dependent
dialuric acid formation. The remaining alloxan and alloxan formed from dialuric acid
oxidation in both reactions will either achieve equilibrium or form compound 305. Because

enzymes catalyze reactions but do not alter equilibrium constants, the rate of formation of

145

 

compound 305 was equal in both the RPLTT catalyzed and uncatalyzed reaction. The
evidence presented here supports the view that alloxan is a preferred substrate for RPLTT
and that compound 305 formation is the product of a competing side reaction. Although
RPLTT led to the suppression of the formation of compound 305, compound 305 increased
with time in both the uncatalyzed and RPLTT catalyzed reaction. Furthermore, 82% of the
alloxan was reduced independently to dialuric acid 22 sec after initiation of the reaction in
the presence of RPLTT with negligible absorbance at 305 nm. If compound 305 was a
substrate or intermediate in the reaction forming dialuric acid, an increased absorbance at
305 nm would be expected prior to development of the 275 nm peak.

To assess a potential role for thioltransferase in alloxan toxicity, oxygen uptake in
the uncatalyzed and RPLTT catalyzed reaction was analyzed. The presence of RPLTT
caused a significant increase in oxygen consumption compared to the chemical reaction from
0.05 mM GSH to 1.5 mM GSH. At higher concentrations of GSH (2.0 and 3.0 mM), there
was a decrease in oxygen consumption compared with the 1.0 and 1.5 mM GSH levels. This
was most likely due to the inhibitory effect of high GSH:alloxan ratios on dialuric acid
oxidation and oxygen consumption, as observed previously (12).

The production of reactive oxygen species is essential for the toxicity of alloxan in
the 2-cells of the pancreas (2, 3, 20-23). If thioltransferase is to play a role in the toxicity of
alloxan, not only should there be an increase in oxygen consumption, but superoxide and
hydrogen peroxide should be produced by the RPLTT catalyzed reaction mixture. In the
RPLTT catalyzed reaction, SOD and catalase significantly inhibited oxygen uptake. The
presence of both SOD (4O rig/ml) and catalase (40 ag/ml) further enhanced inhibition of

oxygen uptake, and was the only instance where maximal oxygen consumption was

146

‘1

G.— .mnTM

 

decreased by 50%. These results indicate that both superoxide and hydrogen peroxide were
generated by the RPLTT catalyzed reaction and subsequent oxidation of dialuric acid. In the
uncatalyzed reaction, only catalase inhibited oxygen consumption. These results appear to
conflict with previous work where both SOD and catalase inhibited oxygen consumption
during the chemical reaction (10, 11, 13) For example, Munday, et al. (13) determined that
in 50 mM potassium phosphate, pH 7.4, 1 mM GSH, 0.05 mM alloxan, and at 25°C, the
presence of 10 rig/ml SOD inhibited oxygen uptake by 91 %.

According to Winterbourn, et al. (33), dialuric acid oxidation can occur via two
mechanisms, a superoxide dependent chain mechanism and an autocatalytic mechanism. As
a result, under the appropriate conditions, dialuric acid oxidation can occur as rapidly in the
presence of SOD as in the absence of SOD because of the ability of dialuric acid to oxidize
via the autocatalytic mechanism (33). In the proposed reaction scheme for the alloxan-
dialuric acid cycle (Figure 9), SOD can suppress reactions 2 (02" + H” + Ale2 ~ 'Ale +
H202) and 4, but oxidation can persist if reaction 3 keeps pace with the alloxan radical
produced in reactions 1 (Ale2 + 02 ~ ’Ale + 02" + H) and 5 (33). In the presence of
SOD, maximal rates of oxidation occurred with increasing concentrations of alloxan and
dialuric acid and when [alloxan] = [dialuric acid] (33). In the experiments described in the
current paper a 4-fold higher level of alloxan was used than was used in references 10, I I,
and 13, resulting in favorable conditions for rapid dialuric acid oxidation by the autocatalytic
reaction in the presence of SOD.

In contrast to the uncatalyzed reaction , SOD inhibited oxygen consumption in the
presence of RPLTT because dialuric acid oxidation by the autocatalytic mechanism is

minimized in the presence of RPLTT. The autocatalytic mechanism depends on the

147

NADPH + H+ GSSG

GR TT

NADP+ 2 GSH Alx 3

Compound 305

 

02.-+ H+

SOD: 2oz"+ H+ ——> H202+02

catalase: 2 H202 ——-> 2 H20 + 02

Figure 9: Interactions between oxygen and dialuric acid. Reversible reactions are
designated by numbers placed near their reactants. Reaction 1 is Ale2 + 02—»

°Ale + 02'-+ H+ and reaction 2 is OZ°-+ H+ + Ale2 -—> 'Ale + H202.
GR, GSSG reductase: TT, thioltransferase;

148

 

production of alloxan radical by reactions 1 and 5, and the consumption of alloxan radical
by reaction 3 (Figure 9). However, in the presence of RPLTT the GSH-dependent reduction
of alloxan to dialuric acid was favored over the formation of alloxan radical by reaction 5.
Winterboum, et al. (33) report that the rate constant for reaction 5 (Alx + Ale2 ~ 2 'Ale,
Figure 8) is 45 M" sec‘l whereas in this paper we report the kw/Km for RPLTT catalysis of
Alx + ZGSH —~ Ale2 + GSSG as 4.5 x 105 M“ sec". Alloxan concentrations in the reaction
mixture were low since RPLTT rapidly catalyzed the GSH—dependent reduction of alloxan
to dialuric acid. Therefore, oxidation of dialuric acid depended predominantly on the
superoxide dependent chain mechanism because the comproportionation of alloxan and
dialuric acid forming alloxan radical was not favored. In the presence of both RPLTT and
SOD, oxygen consumption was inhibited because dialuric acid oxidation by the autocatalytic
mechanism was minimized.

In the previous chapter, reduced RPLTT and the R26V and K27Q variants but not
the C25S enzyme were demonstrated to be able to reduce DHA to AA in the absence of
GSH. The product of this reaction is oxidized enzyme, which was identified via isoelectric
focusing. Due to the structural similarities between DHA and alloxan, the ability of RPLTT
and the K27 Q and C258 variants to reduce alloxan in the absence of GSH was investigated.
When reduced RPLTT and K27Q enzyme were incubated with alloxan, oxidized enzyme
was formed, suggesting the ability of RPLTT and K27Q enzyme to reduce alloxan
chemically. This result suggests that thioltransferase has similar catalytic mechanisms for
both its DHA and alloxan reductase activities.

In contrast, the C25S enzyme was unable to form oxidized enzyme, and when

incubated with DHA no species besides reduced enzyme was seen. However, when the

149

 

C25S variant was incubated with alloxan, reduced enzyme and a species at p1 = 6.2 were
seen. The pl = 6.2 species may be the formation of a thiohemiketal intermediate. When the
C258 enzyme was preincubated with alloxan plus GSH, reduced enzyme and a species
focusing at a p1 = 5.9 was seen. The reaction of alloxan plus GSH has been demonstrated
to produce oxygen radicals (Figure 7 and refs 10—13). The species focusing to a p1 = 5.9 is
perhaps the formation of oxidized sulfur like sulfenic or sulfonic acid.

Reduction of alloxan to dialuric acid was previously demonstrated with GSH (1 0-13),
NAD(P)H (8, 9) and enzymatically by the thioredoxin/NADPH-thioredoxin reductase system
(14). In this paper, we demonstrated thioltransferase catalysis of the GSH dependent
reduction of alloxan to dialuric acid. We provided evidence suggesting that thioltransferase
may play a role in the toxicity of alloxan by enhancing redox cycling between alloxan and
dialuric acid. In studies of calf (34) and pig (35) thioltransferases, although the pancreas had
lower levels than other tissues like the stomach and lung, moderate activity was
demonstrated. Thioltransferase levels in the 2-cells of the pancreas is unknown. Eizirik, et
a1. (3 6) reported that human islets were not destroyed by alloxan, in vitro, nor when grafted
into nude mice, whereas they confirmed the toxicity of alloxan in rat and mouse islets. This
raises the question of species differences in susceptibility to alloxan and the reasons for these

species differences.

150

 

 

REFERENCES

(l) Rerup, C. C. (1970) Pharmacol Rev 22, 485-518.

(2) Oberley, L. W. (1988) Free Radical Biol Med 5, 1 13-124.

('3) Scarpelli, D. G. (1989) T oxicol Appl Pharmacol 101, 543-554.

(4) Weaver, D. C., McDaniel, M. L., and Lacy, P. E. (1978) Endocrinologz 102, 1847-1855.

(5) Malaisse, W. J ., Malaisse—Lagae, F., Sener, A., and Pipeleers, D. G. (1982) Proc. Natl.
Acad. Sci. USA 79, 927-930.

(6) Hammarstrdm, L., Hellman, B., and Ullberg, S. (1966) Diabetologia 2, 340-345.
(7) Grankvist, K., Marklund, S. L., and Téiljedal, I.-B. (1981) Biochem. J. 199, 393-398.
(8) Miwa, I., and Okuda, J. ( 1982) Biochem. Pharmacol. 31 , 921-925.

(9) Nukatsuka, M., Sakurai, H., and Kawada, J. (1989) Biochem. Biophys. Res. Commun.
165, 278-283.

(10) Winterboum, C. C., and Munday, R. (1989) Biochem. Pharmacol. 38, 271-277.

(11) Lenzen, S., and Munday, R. (1991) Biochem. Pharmacol. 42, 1385-1391.

(12) Sakurai, K., and Ogiso, T. ( 1991) Chem. Pharm. Bull. (Tokyo) 39, 737-742.

(13) Munday, R., Ludwig, K., and Lenzen, S. (1993).]. Endocrinol. 139, 153-163.

(14) Holmgren, A., and Lyckeborg, C. (1980) Proc. Natl. Acad. Sci. USA 7 7, 5149-5152.

(15) Deamer, D. W., Heikkila, R. E., Panganamala, R. V., Cohen, G., and Comwell, D.
G.(l97l) Physiol. Chem. Phys. Med. NMR. 3, 426-430.

(16) Cohen, G., and Heikkila, R. E. (1974) J. Biol. Chem. 249, 2447-2452.
(17) Munday, R. (1988) Biochem. Pharmacol. 3 7, 409-413.
(18) Sakurai, K., Miura, T., and Ogiso, T. (1990) Chem. Pharm. Bull. (Tokyo) 38, 993-997.

(19) Sakurai, K., and Miura, T. (1988) Chem. Pharm. Bull. (Tokyo) 36, 4534-4538.

151

(20) Grankvist, K., Marklund, S., Sehlin, J ., and Taljedal, I.-B. (1979). Biochem. J. 182, 17-
25.

(21) Fischer, L. J ., and Hamburger, S. A. (1980) Diabetes 29, 213-216.

(22) Uchigata, Y., Yamamoto, H., Kawamura, A., and Okamoto, H. (1982) J. Biol. Chem.
25 7, 6084-6088.

(23) Grankvist,I(., Marklund, s., and Téiljedal,I.-B.(1981)Nature294,158-160.

(24) Wells, W. W., Yang, Y., Deits, T. L., and Gan, Z. R. (1993) Adv. Enzymol. Relat.
Areas Mol. Biol. 66, 149-201.

(25) Wells, W. W., Xu, D.-P., Yang, Y., and Rocque, P. A. (1990) J. Biol. Chem. 265,
15361-15364.

(26) Yang, Y., and Wells, w. w. (1991).]. Biol. Chem. 266, 12759-12765.
(27) Holmgren, A. (1985) Annu. Rev. Biochem. 54, 237-271.

(28) Yang, Y., and Wells, w. w. (1990).]. Biol. Chem. 265, 589-593.
(29) Gan, Z.-R, and Wells, w. w. (1986).]. Biol. Chem. 261,996-1001.

(30) Mieyal, J. J., Starke, D. W., Gravina, S. A., and Hocevar, B. A.( 1991)
Biochemistry 30, 8883-8891.

(31) Perrella, F. W. (1988) Anal. Biochem. 174, 437-447.
(32) Wells, W. W., Xu, D.-P., and Washbum, M. P.(1995) Methods Enzymol. 252, 30-38.

(33) Winterboum, C. C., Cowden, W. B., and Sutton, H. C. (1989) Biochem. Pharmacol.
38, 611-618.

(34) Rozell, B., Barcena, J. A., Martinez-Galisteo, E., Padila, C. A., and Holmgren, A.
(1993) Eur. J. Cell. Biol. 62, 314-323.

(35) Gan, Z.-R, and Wells, W. W. (1988) J. Biol. Chem. 263, 9050-9054.

(36) Eizirik, D. L., Pipeleers, D. G., Ling, Z., Welsh, N., Hellerstrom, C., Anderson, A.
(1994) Proc. Natl. Acad. Sci. USA 91, 9253-9256.

152

 

Chapter V

Summary and Perspectives

153

' ..-..‘.-... fit";

,i a 1 .
.'

 

In the completion of this thesis I have developed a reproducible GSH-dependent
DHA reductase activity assay and kinetically characterized and compared mammalian DHA
reductases, a rapidly growing class of enzymes. Prior to this thesis, very little was known
about the catalytic mechanism of any DHA reductase. The only information known was that
Cys-22 was essential for the DHA reductase activity of thioltransferase (1). Furthermore,
Cys at position 25 was known not to be essential, and the placement of a serine at this
position enhanced the relative DHA reductase activity of thioltransferase (I ). I characterized
the catalytic mechanism of a GSH-dependent DHA reductase for the first time, that of
thioltransferase. Finally, I characterized the alloxan reductase activity of thioltransferase and
provided evidence that thioltransferase may contribute to the toxicity of alloxan. Alloxan
is the second non-disulfide substrate described for thioltransferase.

Chapter 11 describes the kinetic characterization of the DHA reductase activity of
RPLTT, bovine liver PDI , and the human erythrocyte 32kDA DHA reductase. The
emerging picture in the field of DHA reductases is one in which there are several potential
systems to regenerate AA from DHA. NADPH-dependent DHA reductases like 31(-
hydroxysteroid dehydrogenase (2) and thioredoxin/thioredoxin reductase (3) and GSH-
dependent DHA reductases like thioltransferase (4) and the 32 kDa DHA (5, 6) reductase
may all play roles in vivo. It is likely that the importance of each enzyme in DHA reduction
will vary from species to species and cell to cell. Based on their respective catalytic
activities, RPLTT and the human erythrocyte 32kDa DHA reductase were the most robust
DHA reductases when compared to all the known mammalian DHA reductases.

It is interesting that human and pig liver thioltransferases had very similar Km values

for DHA of 0.2 to 0.27 mM but their kcat values varied from 41 to 374 min" resulting in pig

154

liver thioltransferase having a kc,,/Km for DHA of 2.4 x 104 M" sec" compared to a kcm/K1m for
DHA for the human enzymes of 2.5 to 5.5 x 103 M" see" (Table 11 ,Chapter 11) making the
pig liver enzyme 4 to 10 fold more efficient. Both pig liver and human thioltransferases
contain 105 amino acids. With a sequence identity of 83% there are 18 amino acids different
between the two enzymes (Figure 4, Chapter I). A thorough site directed mutagenesis study
of the human enzyme should allow one to identify which amino acid(s) account for this
difference in activities.

When the 32kDa DHA reductase from human erythrocytes was purified by Xu et al.
(6), the following peptide fragments were sequenced (unpublished results): (1)
HEVININLK, (2) NKPEWFFKK, (3) MILELFSK, (4) PSLVGSF IR, (5) WQFLELYL, and
the C-terminal fragment (6) LKQIGALQ. A recent search in the data bank by Dr. Wells
identified fragments (l) to (5) to be identical to fragments in a partial clone of a human
homolog to a mouse protein that is differentially expressed in lymphoma cells that have
different susceptibilities to radiation induced apoptosis. The cloning was carried out under
the direction of Michael Story at the University of Texas MD Anderson Cancer Center. The
full length protein is 32 kDa (personal communication from Dr. Story), making it possible
that the 32kDa DHA reductase from human erythrocytes is differentially expressed in
lymphoma cells, and that its expression may provide defense’s against damage induced by
radiation.

Thioltransferase has been demonstrated to have a 4 fold enhanced expression in
MCF-7 adriamycin resistant cells compared to MCF-7 adriamycin sensitive cells (7). It was
postulated that the DHA reductase activity of thioltransferase contributed to the resistant

phenotype. In the Ph.D. thesis of Dr. Elizabeta Borer-Meyer, Dr. Borer-Meyer transfected

155

 

MCF-7 adriamycin sensitive cells with human thioltransferase and found that thioltransferase
transfection indeed increased the resistance of these cells to adriamycin, but not to the level
of resistance of the MCF-7 adriamycin resistant cells. Furthermore, the increased resistance
of the transfected cell lines was most likely not due to the DHA reductase activity of
thioltransferase. The latest results concerning the 32kDa DHA reductase suggest that cancer
cells may possess enhanced DHA reductase activities due to elevated levels of this enzyme,
and enhanced DHA reductase activity may provide cancer cells with enhanced resistance to
damage induced by radiation.

The wide variety of DHA reductases, especially the GSH-dependent DHA reductases,
raises several enzyme structure/flinction questions. Whether each enzyme has an identical,
similar or different catalytic mechanism for DHA reduction is unknown. Several catalytic
mechanisms for DHA reductases have been proposed prior to this thesis, but no one had
specifically studied the catalytic mechanism of a DHA reductase. In Chapter III, the first
description of a catalytic mechanism of a DHA reductase is presented.

When starting with the reduced form of the enzyme, IAM inactivation studies
demonstrated the ability of DHA to bind to enzyme, prior to GSH binding because DHA
protected against IAM inactivation whereas GSH did not. The ability of RPLTT and the
R26V and K27Q variants to chemically reduce DHA to AA in the absence of GSH was
demonstrated by electrochemical analysis and IEF. In addition, IEF analysis demonstrated
the ability of reduced RPLTT and the R26V and K27Q variants to react with GSSG and form
oxidized enzyme. The relative inability of the C258 enzyme to chemically reduce DHA to
AA, but its robust DHA reductase activity suggest that the postulated formation of a

thiohemiketal intermediate is reduced by GSH releasing AA and forming an enzyme mixed

156

disulfide with GSH.

Although Chapter 111 provides strong evidence for the proposed catalytic mechanism
of the DHA reductase activity of thioltransferase, many more studies will be needed to
completely determine the catalytic mechanism. The major aspect of the proposed
mechanism that needs to be experimentally demonstrated is the presence of a thiohemiketal
intermediate between enzyme and a non-disulfide substrate like DHA. It is clear that in
order to carry out this work, the C25S enzyme of the enzyme must be used since wild type
enzyme is capable of binding and chemically reducing DHA to AA forming oxidized
enzyme. It is also important to determine what the mechanism of thioltransferase is in the
presence of GSH. Does wild type enzyme form an intramolecular disulfide as an
intermediate, or does GSH displace DHA from the enzyme prior to Cys-25 attack on Cys-22
releasing AA. Finally, in E. coli thioredoxin, Asp-26 has been demonstrated to act as a
general acid/base and deprotonate Cys-35 which can then attack Cys-32 to release reduced
substrate forming oxidized enzyme (8). RPLTT was capable of chemically reducing DHA
to AA forming oxidized enzyme. It is likely that Cys-25 in RPLTT must be deprotonated
by a general base to attack Cys-22 to form an intramolecular disulfide in this reaction, but
a residue acting as a general base has not been found in thioltransferase.

During the study of the catalytic mechanism of the DHA reductase activity of
thioltransferase, a possible function of Lys-27 in thioltransferase was elucidated. The DHA
reductase activity of the K27 Q variant was incompletely inhibited by IAM with 40% residual
activity after 45 minutes. GSH enhanced the IAM inactivation of the DHA reductase
activity of the K27Q enzyme. Although Cys-22 in the K27Q enzyme has a slight increase

in pKa of 4.3 vs. 3.8 for RPLTT, the role of Lys-27 was previously unknown (1). The

157

 

current studies suggested that the role of lysine at position 27 is to stabilize the reduced form
of RPLTT. In contrast to the other enzymes, when the K27Q variant was treated with IAA
or preincubated with AA prior to IAA addition, a significant amount of oxidized enzyme
was seen. Finally, the K27Q enzyme had the greatest initial rate of chemical reduction of
DHA to AA, a reaction which includes the oxidation of the active site sulfhydryls. These
data suggest that the K27Q variant is much more readily oxidized than RPLTT indicating
a role of lysine at position 27 in the stabilization of the reduced form of the enzyme.

Many DHA reductases have several different activities including thioltransferase,
thioredoxin/thioredoxin reductase, and PDI which belong to the class of enzymes known as
thiol:disulfide oxidoreductases, which contain the -CXXC- active site (9). Nothing is known
about the catalytic mechanism of the DHA reductase activity of any other enzyme besides
thioltransferase. Because they are thiol:disulfide oxidoreductases, it will be interesting to
determine whether PDI or the thioredoxin/thioredoxin reductase system have identical,
similar or different catalytic mechanisms for DHA reduction, and how their respective
catalytic mechanisms compare to that of thioltransferase. Since nothing is known about the
active site of the recently identified 32kDa DHA reductase, this is perhaps the most exciting
enzyme to study, especially because of its potential role in cancer.

It was long believed that thioltransferase was only a GSH-dependent thiol:disulfide
oxidoreductase, but in 1990 Wells et al. (10) demonstrated that human placenta, bovine
thymus, and pig liver thioltransferases possessed GSH-dependent DHA reductase activity.
With the advent of this discovery, I asked the question, what other non-disulfide substrates
for thioltransferase might exist?

Reduction of alloxan to dialuric acid was previously shown to occur with GSH (11-

158

 

 

14), NAD(P)H (15, 16) and enzymatically by the thioredoxin/NADPH-thioredoxin reductase
system (17). Because thioltransferase is GSH dependent, thioltransferase and thioredoxin
share the active site -CXXC-, and alloxan and DHA share a similar triketone structure, I
investigated the GSH-dependent alloxan reductase activity of thioltransferase. In Chapter
IV, I demonstrated thioltransferase catalysis of the GSH dependent reduction of alloxan to
dialuric acid. I provided evidence suggesting that thioltransferase may play a role in the
toxicity of alloxan by enhancing redox cycling between alloxan and dialuric acid.
Presumably because of the similarities between alloxan and DHA, thioltransferase would
have similar catalytic mechanisms for their GSH-dependent DHA reduction. In support of
this, when alloxan was incubated with RPLTT or the K27Q variant , oxidized enzyme was
formed. When alloxan was incubated with the C25S enzyme and analyzed via IEF, both
reduced the C25S variant Q31 = 7.0) and a species focusing at a pI of 6.2 were seen (Figure
8, Chapter IV). This species at pl = 6.2 may be the postulated enzyme-substrate
intermediate thiohemiketal which would be very useful for future investigations.

Finally, the alloxan reductase activities of thioltransferase and the
thioredoxin/thioredoxin reductase system suggest that these enzymes may play a role in
alloxan induced diabetes. Alloxan must be reduced to dialuric acid in order to be toxic
because the oxidation of dialuric back to alloxan generates reactive oxygen species like
superoxide, hydrogen peroxide, and hydroxyl radical in the presence of an appropriate metal
catalyst (18-21). There may be other as yet unstudied redox cycling compounds similar in
structure to alloxan that may be substrates for thioltransferase and/or thioredoxin making
these enzymes important in their toxic action on cells and tissues.

The substrate profile of thioltransferase demonstrates that depending on the

159

 

 

conditions, the presence ofthioltransferase could be either beneficial or harmfiJl. As a DHA
reductase and a thiol:disulfide oxidoreductase, thioltransferase is believed to contribute to
the homeostasis of a cell. However, in the presence of a redox cycling substrate like alloxan,
thioltransferase could potentiate the compound’s toxicity. Therefore, depending on the
conditions, thioltransferase may either contribute to the protection or add to the damage of

a cell.

160

J
Lin:-

 

REFERENCES

(l) Yang, Y., and Wells, W.W. (1991).]. Biol. Chem. 266, 12759-12765.

(2) Del Bello, B., Maellaro, E., Sugherini, L., Santucci, A., Comporti, M., and Casini, A.F.
(1994) Biochem J. 304, 385-390.

(3) May, J .M., Mendiratta, S., Hill, K.E., and Burk, RF. (1997) J. Biol. Chem. 2 72, 22607-
22610.

(4) Wells, W.W., Xu, D.P., Yang, Y., and Rocque, P.A. (1990) J. Biol. Chem. 265, 15361-
15364.

(5) Maellaro, E., Del Bello, B., Sugherini, L., Santucci, A., Comporti, M., and Casini, A.F.
(1994) Biochem J. 301, 471-476.

FTnhAuo me. ‘1
.. iL

(6) Xu, D.P., Washbum, M.P., Sun, GP, and Wells, W.W. (1996) Biochem. Biophys. Res.
Commun. 221, 117-121.

(7) Wells, W.W., Rocque, P.A., Xu, D.P., Meyer, E.B., Charamella, L., and Dimitrov, NV.
(1995) Free Rad. Biol. Med. 18, 699-708.

(8) Chivers, RT. and Raines, RT (1997) Biochemistry 36, 15810-15816.
(9) Mieyal, J.J., Gravina, S.A., Mieyal, P.A., Srinivasan, U., and Starke, D.W. (1995) in

Biothiols in Health and Disease (Packer, L., and Cadenas, E., Eds.) pp. 305-372,
Marcel Decker, Inc., New York, NY.

 

(10) Wells, W.W., Xu, D.P., Yang, Y., and Rocque, P.A. (1990) J. Biol. Chem. 265, 15361- f
15364.

(11) Winterboum, C. C., and Munday, R. (1989) Biochem. Pharmacol. 38, 271-277.
(12) Lenzen, S., and Munday, R. (1991) Biochem. Pharmacol. 42, 1385-1391.

(13) Sakurai, K., and Ogiso, T. (1991) Chem. Pharm. Bull. (Tokyo) 39, 737-742.

 

(l4) Munday, R., Ludwig, K., and Lenzen, S. (1993) J. Endocrinol. 139, 153-163.
(15) Miwa, 1., and Okuda, J. (1982) Biochem. Pharmacol. 31, 921-925.

(16) Nukatsuka, M., Sakurai, H., and Kawada, J. (1989) Biochem. Biophys. Res. Commun.
165, 278-283.

161

(l7) Holmgren, A., and Lyckeborg, C. (1980) Proc. Natl. Acad. Sci. USA 7 7, 5149-5152.

(18) Deamer, D. W., Heikkila, R. E., Panganamala, R. V., Cohen, G., and Comwell, D.
G.(l97l) Physiol. Chem. Phys. Med. NMR. 3, 426-430.

(19) Cohen, G., and Heikkila, R. E. (1974) J. Biol. Chem. 249, 2447-2452.
(20) Munday, R. (1988) Biochem. Pharmacol. 37, 409-413.

(21) Sakurai, K., Miura, T., and Ogiso, T. (1990) Chem. Pharm. Bull. (Tokyo) 38, 993-997.

162

 

   

MICHIGAN smTE UNIV. LIBRQRIES
1|1111111111111111111111111111111111111111111
31293016825418