', .UIIII‘IIIhiI:fITI1II

I ,. : III

I III

' I-
“III II -

 
    
   
 
  

“If!
“I"! ‘IIIIIII III?

III III IIIIIIII :
IIIIIIMy ,I. IN

I IIl . III I

 
 

I I

reg

  
    
  

     
      
  
  
   

I

 
  

  

‘- _.__:_:_:;:::
‘A.
q_.
“3’
W—-—«—_
-7—
_ .
_.
W __
_—'::_: 4-;_

fi_
~A~tw
fi—w.
‘-_————

      

 
  

1;,
IIIIM

 

 

    
        
   
  
       
  

  
     
   

     
  

if. 3?.
‘IIIW; ,6?“

4

II :3 gal-uh 33%?“

6 III? III." a?

I“ III, II

 
   

V I, “I“

g. 1}“ '31 54,431! Iin. I .2333]?
I} ‘ . 1 ml h“ l: a
II? :ImI I n I5 I
If: ii; In “Pry“: I11? Iii?“

.1,
I I v '9 u r. Ht.
If: I I'I' :31; .- ; :‘ {RITE
II I: IIIIIM" III IJIIIfI‘ ': 'I‘I‘IIIIILE: -,. I.
It III “MIL xiii”; "1".‘5 £.‘.’".I..I "’ ‘z ‘
3:”;ng ‘II‘ .1 -..a I“

’WLI‘JII‘L“ .‘I-{Sii {If _f.g;;,: f‘g
. N .é‘III’I'I'R‘I“ “LI'II‘II‘ .31.}; I7;- I
| .1 EilfivIw-HIHIIJ an ‘II “II :‘EII (:II: I? XII... III II

IIJJHIfi-“iq 1" III. ”II ”III”; i'I‘IiI'IH; “ll 11% ”“1
.III- 'LJI' 0'1. lIflf' I“L I ': i ,_ - ‘L'#
III“. “I IIIII ' I”:

III ' .1: w"
"I I II‘IIIII‘IIIII‘I” I

 

 

I I I ' It in“: I': 41". I 1‘3“};1".
t. { I I: III: ‘ ‘3‘: 1581!“? “(:3)” M" q'h'I.

   
   
   
 
  
 

 
 

     
 
  
  

 
   

  

   
   

"' {SM
:7 '

 

raw” -".~J=-

THESIS - --‘-«:3..__ .- ..
1.. In: W?
La Lift; gig” ‘Kd‘g 9"?

61"«143 (”91" A? “*5???-
r-"n fi/Qfifi‘

L UMV’Sfimgg-y J

m

 

This is to certify that the

thesis entitled

PURIFICATION AND CHARACTERIZATION OF AN
a-KETOISOCAPROATE OXIDASE FROM RAT LIVER

presented by

Patrick John Sabourin

has been accepted towards fulfillment
of the requirements for

 

 

Ph.D. degree in Biochemistry

 

 

 

Q/A‘,

Major professor

//

0-7 639

 

 

 

MSU

LIBRARIES
”—

 

 

RETURNING MATERIALS:
RTace in book drop to
remove this checkout from
your record. FINES will
be charged if book is
returned after the date
stamped be10w.

 

 

 

PURIFICATION AND CHARACTERIZATION OF AN
o—KETOISOCAPROATE OXIDASE FROM RAT LIVER

By

Patrick John Sabourin

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Biochemistry

1981

ABSTRACT

PURIFICATION AND CHARACTERIZATION OF AN
a-KETOISOCAPROATE OXIDASE FROM RAT LIVER

By

Patrick John Sabourin

Isopycnic sucrose gradient separation of rat liver organelles re-
vealed the presence of two distinct enzymes which decarboxylate a-keto-
isocaproate; the mitochondrial branched-chain a-keto acid dehydrogenase
and a cytosolic a-ketoisocaproate oxidase. The branched-chain a—keto
acid dehydrogenase uses a-ketoisocaproate, a-keto-B-methylvalerate, and
a-ketoisovalerate (the a-keto acids of leucine, isoleucine and valine) as
substrates and requires CoASH and NAD+ as cofactors. This enzyme has
been studied extensively and operates by a mechanism similar to that of
pyruvate dehydrogenase. The cytosolic a-ketoisocaproate oxidase uses
only a-ketoisocaproate and a-keto-Y-methiolbutyrate (the a-keto acids of
leucine and methionine) as substrates and does not require CoASH or
NAD+ for activity. The purpose of this study was to further character-
ize the cytosolic o-ketoisocaproate oxidase activity.

The cytosolic a-ketoisocaproate oxidase activity was partially char-
acterized in 70,000 x g supernatant fractions of rat liver homogenates
(cytosol preparation). Oxygen was required fbr enzymatic activity. Cy-
tosol preparations consumed oxygen when either a-ketoisocaproate or a-
ketOdY—methiolbutyrate were added. None of the other a-keto acids tested
stimulated oxygen consumption. a-Ketoisovalerate, a-keto-B-methylvaler-
ate, a—ketobutyrate and a-ketononanoate inhibit the a—ketoisocaproate
oxidase activity. Phenylpyruvate is a very potent inhibitor. a-Ketoiso-

caproate oxidase activity was only detected in liver and kidney.

In order to further characterize the cytosolic a—ketoisocaproate
oxidase (decarboxylase) activity, it was purified from rat liver. The
purified enzyme required Fe2+ and a sulfhydryl reducing reagent fbr
optimal activity. Other metal ions tested would not replace Fe2+.

During purification this enzyme was quite unstable. Inclusion of 5% mon-
othioglycerol in all buffers increased stability. Using a 180 fold puri-
fied preparation of the a-ketoisocaproate oxidase, the assay conditions
fbr this enzyme were optimized. Maleate, up to 0.2 M increased the
activity. This was not due to an ionic strength effect. Optimal activ-
ity was obtained at pH 6.0 in the presence of 1 mM FeSO4, 0.5 mM ascor-
bate and 1 mM dithiothreitol.

The Km of the a-ketoisocaproate oxidase for a-keto-meethiolbuty-
rate is about 6 times higher than that for a-ketoisocaproate (1.9 mM vs.
0.3 mM). This enzyme converts a-ketoisocaproate to a-hydroxyisovaleric
acid. Isovalerate is not a free intermediate of the reaction. 180
incorporation experiments indicate that this enzyme is a dioxygenase,
that inserts one 0 atom of 02 into the carboxyl group and the other 0
atom of 02 into the B-hydroxyl group of B-hydroxyisovaleric acid.

To Carol Lee

11

ACKNOWLEDGEMENTS

I would like to thank Dr. Loran Bieber for his guidance, encourage-
ment and critical evaluation of my experimental data. He is responsible
for increasing my ability to critically judge my work and that of others.
I would also like to acknowledge the help and friendship of my coworkers
who have made my stay here a pleasant and valuable experience. I am
deeply grateful to Betty Baltzer, who worked overtime on mass spectral
analysis and Sara Morrison-Rowe and Chris Vandenberg who helped me with
countless tedious assays.

I am grateful for the financial support of the National Institute of
Health and the College of Osteopathic Medicine at Michigan State Univer-
sity which made this research and my academic training possible.

I extend my deepest gratitude to my wife Carol. Her support, en-
couragement, understanding and patience have helped me through the dis-

couraging days and kept the fun in science.

iii

TABLE OF CONTENTS

LIST OF FIGURES . . . . . . . . . . . . . . . . . . .
LIST OF TABLES . . . . . . . . . . . . . . . . . . .
ABBREVIATIONS . . . . . . . . . . . . . . . . . . . .
INTRODUCTION . . . . . . . . . . . . . . . . . . . .
LITERATURE REVIEW . . . . . . . . . . . . . . . . . .
I. Metabolism of the Branched-Chain Amino Acids

A. Metabolic Pathway . . . . . . . . . . . .

3. Regulation and Interorgan Distribution of
Branched-Chain Amino Acid Metabolism . .

II. Effect of Carnitine on Branched-Chain Amino

kid Metab01 1 sm 0 C C O O . C C C O O O O O O 0

III. Specific Effects of Leucine on Metabolism . .
A. Effect of Leucine on Protein Turnover . .
B. Effect of Leucine on Insulin Secretion .

C. Effect of Leucine on Glucose and Pyruvate
MetabOI 1. 5'" I O O O O O O O O O O O 0 O 0

IV. Metabolic Diseases Associated with Branched-
Chain Amino Acid Metabolism . . . . . . . . .

V. Evidence for the Existence of an Extramito-
chondrial Branched-Chain a-Keto Acid Decar-
boxylaseActivity..............

MATERIALS 0 I O I O O O O O O O O O O O O O O O O O 0

METHODS O O 0 O O O O O O O O O O O O O O O O O O O 0

iv

Page

viii

xii

bah

10
10
12

12
13
16

18
19

Page

I. Sucrose Gradient Separation of Rat Liver
organe] 1 es 0 O O O 0 O O O O O O O O O O O O O O O 0 O O O 19

II. Preparation of Radioactively Labeled arKeto
kids 0 O O O O O O O O O O O O O O O O O O O O O O O O O O 20

III. Assay of arKetoisocaproate Oxidase Activity . . . . . . . . 21

A. Assay of a—[1-14C] Keto Acid Decarboxy-
lating Activities . . . . . . . . . . . . . . . . . . . 21

mthOd A O O O O O O I O O O O O O O O O O O O O O O O 21
MethOd B O O O O O O O O O . O O O O O O O O O O O O O O 22

B. Measurement of Oxygen Consumption in
Presence of apKeto Acids . . . . . . . . . . . . . . . 23

IV. Isolation of Rat Liver Cytosol Preparations . . . . . . . . 23

V. Purification of Rat Liver Cytosolic a-Keto-
isocaproate Oxidase . . . . . . . . . . . . . . . . . . . . 24

A. Purification A: Partial Purification of
Rat Liver Cytosolic a-Ketoisocaproate
OXidase O O O O O O O O O O O O O O O O O O 0 O O O O O 24

B. Purification B: Final Purification of
apKetoisocaproate Oxidase . . . . . . . . . . . . . . . 26

0. Purification C . . . . . . . . . . . . . . . . . . . . 29
VI. Analysis of Reaction Products . . . . . . . . . . . . . . . 31

VII. Polyacrylamide Gel Electrophoresis Using
Denaturing or Non-Denaturing Conditions . . . . . . . . . . 32

VIII. Determination of Molecular Height of the
arKetoisocaproate Oxidase by Sephacryl S-ZOO
Chmmtography I O O O O O O O O O O O O I O 0 O O O O O O 33

IX. Enzymatic Incorporation of 180 Into 8-
Hydroxyisovaleric Acid . . . . . . . . . . . . . . . . . . 34

A. Incubationswith1802................. 34
B. Incubations with H2130 . . . . . . . . . . . . . . . . 35
C. Isolation of B-Hydroxyisovaleric Acid . . . . . . . . . 35
RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

I.

II.

III.

IV.

V.

Subcellular Distribution and Partial Char-
acterization of an a-Ketoisocaproate Oxidase

Of Rat Li ver 0 O O O O O O O O O O O O O O O O O

A. Subcellular Distribution of Branched-
Chain arKeto Acid Decarboxylating
Activities in Rat Liver . . . . . . . . . .

B. Properties of Rat Liver Cytosolic a-
Ketoisocaproate Decarboxylase Activity . .

C. Effect of arKeto Acids on Cytosolic a-

Ketoisocaproate Oxidase Activity . . . . . .

0. Tissue Distribution of a-Ketoisocaproate
Oxidase in the Male Rat . . . . . . . . . .

Purification and Stabilization of the a-Keto-
isocaproate Oxidase Activity from Rat Liver . .

A. Stabilization of and Cofactor Requirements
of a—Ketoisocaproate Oxidase Activity . . .

8. Purification of asKetoisocaproate Oxidase
ktiVity O O O O O I O O I O O O O O O O 0

Physical Properties of Rat Liver arKetoiso-

caproate Oxidase . . . . . . . . . . . . . . . .

Kinetic and Catalytic Properties of the
Purified a-Ketoisocaproate Oxidase . . . . . .

A. Stability of the Purified arKetoisocap-
roate Oxidase . . . . . . . . . . . . . . .

8. Metal Requirement of the arKetoisocap—
mate OX‘Idase O O O O O O O I O O O O O O O

C. Activation of the a—Ketoisocaproate
OXidase O O O O O O O O O 0 O I O O O O O O

0. Optimal Assay Conditions for the or
Ketoisocaproate Oxidase . . . . . . . . .

E. Substrate Specificity of the a-Keto-
isocaproate Oxidase . . . . . . . . . . . .

Studies on the Mechanism of the arKetoiso-
caproate Oxidase . . . . . . . . . . . . . . .

A. Product Identification . . . . . . . . . .

vi

37

37
39
46
49
49
49
61
68
68
68-
74
78
85
96

99
99

Page

8. Evidence That Isovalerate Is Not An
Interm1ate O O O I O O O O O O O O O O O O O O O O O 105

C. Mechanism of Formation of B-Hydroxyiso-
valeric Acid by the a—Ketoisocaproate
0x1dase O O O O O O O O O O O O O O O O O O O O O O O O 1.08
DISCUSSION 0 O O O 0 O O O O O O O O O O O O O O O O O O 0 O O O O 113

LIST OF REFERENCES 0 O O O O O O O O O O O O O O O O O O O O O O O 125

vii

LIST OF FIGURES

Figure Page
1 Subcellular Distribution of Branched-Chain a-Keto

Acid Decarboxylases in Rat Liver . . . . . . . . . . . . . 38

2 The Effect of Ammonium Sulfate on Cytosolic a-
Ketoisocaproate Decarboxylase Activity . . . . . . . . . . 41

3 Effect of Time and Protein Concentration on Cyto-
solic arKetoisocaproate Decarboxylase Activity . . . . . . 44

4 Effect of Substrate Concentration on Cytosolic
a-Ketoisocaproate Decarboxylase Activity . . . . . . . . . 45

5 Stability of the a-Ketoisocaproate Oxidase in a
12-F01d Puri fied Preparation I I I I I I I I I I I I I I I 57

6 Stability of the a-Ketoisocaproate Oxidase in an
80-Fold Purified Preparation . . . . . . . . . . . . . . . 58

7 Effect of Monothioglycerol on a—Ketoisocaproate
0x1dase I I I I I I I I I I I I I I I I I I I I I I I I I 60

8 DEAE-Cellulose Chromatography of a-Ketoisocap-
roate OXidase I I I I I I I I I I I I I I I I I I I I I I 64

9 Phenyl-Sepharose Chromatography of a—Ketoiso-
caproate Oxidase . . . . . . . . . . . . . . . . . . . . . 66

10 Sephacryl 5-200 Chromatography of a-Ketoiso-
caproate Oxidase . . . . . . . . . . . . . . . . . . . . . 67

11 SOS-Polyacrylamide Gel ElectrOphoresis of «-
Ketoisocaproate Oxidase . . . . . . . . . . . . . . . . . 69

12 Native Gel Electrophoresis of a-Ketoisocap-
roate OXidase I I I I I I I I I I I I I I I I I I I I I I 70

13 Determination of the Subunit Molecular Height
of aPKetoisocaproate Oxidase . . . . . . . . . . . . . . . 71

14 Molecular Height of the asKetoisocaproate Oxidase
Determined by Sephacryl 3-200 Chromatography . . . . . . . 72

viii

Figure
15
16

17

18

19

20

21
22

23
24

25

26
27
28
29
30
31

32

Stability of the Purified a—KIC Oxidase . . . . .

Effect of Dithiothreitol and Ascorbate on Acti-
vation of the a-Ketoisocaproate Oxidase by FeSO4
or FEC‘B I I I I I I I I I I I I I I I I I I I I I

Effect of FeSO4 Concentration on a-Ketoisocapro-
ate OXIdase ktiVity I I I I I I I I I I I I I I I

Effect of Ascorbate on a-Ketoisocaproate Oxidase
Activity in the Presence of Optimal and Subopti-
mal Concentrations of FeSO4 . . . . . . . . . . .

Stability of the “Activated“ a-Ketoisocaproate
Oxidase in the Presence or Absence of FeSO4,
Ascorbate and Dithiothreitol . . . . . . . . . . .

Effect of Time and Protein Concentration on the
a-Ketoisocaproate Oxidase Activity . . . . . . . .

pH Optimum of the asKetoisocaproate Oxidase . . .

Effect of NaCl Concentration on the a-Ketoiso-
caproate Oxidase Activity . . . . . . . . . . . .

Effect of Maleate Concentration on the a-Keto-
isocaproate Oxidase Activity . . . . . . . . . . .

Effect of Preincubation Time on a-Ketoisocap-
roate Oxidase Activity . . . . . . . . . . . . . .

Effect of Substrate Concentration on a—Ketoiso-
caproate or arKetoAY-Methiolbutyrate Oxidase
MtiVity I I I I I I I I I I I I I I I I I I I I I

Gas Chromatograph Profile of the Major Product of
Q'KEEOIsocaproate OXIdaSE o o o o o o o o o o o 0

Mass Spectrum of the Major Product of a—Ketoiso-
caproate Oxidase . . . . . . . . . . . . . . . . .

Dowex-I Chloride Chromatography of the Reaction
Products of the Purified a-Ketoisocaproate Oxidase

Evidence That Isovalerate Is Not An Intermediate
of the a-Ketoisocaproate Oxidase Reaction . . . .

Enzymatic Incorporation of 180 from 1302 or H2130
into a-Hydroxyisovaleric Acid . . . . . . . . . .

Incorporation of 180 from 1302 or H2130 into B-
Hydroxyisovaleric Acid . . . . . . . . . . . . . .

Proposed Mechanism of the a-Ketoisocaproate Oxi-
dase Reaction . . . . . . . . . . . . . . . . . .

ix

79

80

84

87
91

93

94

97

100

101

102

104

107

110

118

120

 

10

11
12

13

LIST OF TABLES

Cofactor Requirements of Rat Liver Cytosolic ar
Ketoisocaproate Decarboxylase Activity . . . . . .

Effects of a—Keto Acids, Arsenite, and Oxygen on
a-Ketoisocaproate Oxidase Activity in Rat Liver
Cytosol Preparations . . . . . . . . . . . . . . .

Oxygen Consumption by Rat Liver Cytosol Prepar-
ations Using Various apKeto Acids as Substrates .

Tissue Distribution of arKetoisocaproate Decar-
boxylase (Oxidase) in the Male Rat . . . . . . . .

Partial Purification of a-Ketoisocaproate Oxidase
from Rat L1 ver I I I I I I I I I I I I I I I I I I

Effects of Possible Cofactors on a-Ketoisocapro-
ate Oxidase Activity in the Partially Purified
Preparation (Sephadex G-150 Pool) . . . . . . . .

Purification of a-Ketoisocaproate Oxidase from
Rat Liver: Purification C . . . . . . . . . . . .

Stability of a-Ketoisocaproate Oxidase in the
Presence of Monothioglycerol, Dithiothreitol, or
GlycerOI I I I I I I I I I I I I I I I I I I I I I

Purification of a-Ketoisocaproate Oxidase from
Rat Li ver I I I I I I I I I I I I I I I I I I I I

Effect of Various Metal Ions on a—Ketoisocaproate
OXIdase ktiVity I I I I I I I I I I I I I I I I I

Activation of asKetoisocaproate Oxidase by FeSO4 .

Stability of a-Ketoisocaproate Oxidase Assay
Mi Xture I I I I I I I I I I I I I I I I I I I I I

Effect of Various Buffers on a-Ketoisocaproate
Oxidase Activity . . . . . . . . . . . . . . . . .

X

47

48

50

51

53

55

59

62

75
82

89

92

Table Page

14 Effect of ADP or EDTA on a-Ketoisocaproate
OXIdase ktiVity I I I I I I I I I I I I I I I I I I I I I 95

15 Decarboxylation of arKetoisocaproate and arKeto-
Y-Methiolbutyrate by Purification Fractions . . . . . . . 98

16 Presence of 18O in Enzymatically Formed 3-
Hydroxyisovaleric Acid and Its Fragments . . . . . . . . . 111

xi

ADP
BCAA
BCKA
Buffer A
CoA
CoASH
DEAE
DTNB
DTT
EDTA
FAD
Hepes
B-HIVA
arKIC
a-KYMB
MES
MOPS
NAD+
NADH
NADP+
* NAOPH
SOS
TCA
Tris

ABBREVIATIONS

adenosine 5'-diphosphate

branched-chain amino acid(s)

branched-chain keto acid(s)

20 mM Tris HCl pH 7.8, 1% Isopropanol

coenzyme A

reduced coenzyme A

diethylaminoethyl

5,5'-dithiobis(2-nitrobenzoic acid)

dithiothreitol

(ethylenedinitrilo)-tetraacetic acid, disodium salt
flavin adenine dinucleotide
N-Z-hydroxyethylpiperazine-N'-2-ethanesulfOnic acid
B-hydroxyisovaleric acid

a-ketoisocaproate

a-keto-v-methiolbutyrate

2(N-morpholino)ethane sulfbnic acid
morpholinopropane sulfbnic acid

nicotinamide adenine dinucleotide (oxidized)
nicotinamide adenine dinucleotide (reduced)
nicotinamide adenine dinucleotide phosphate (oxidized)
nicotinamide adenine dinucleotide phosphate (reduced)
sodium dodecyl sulfate

tricarboxylic acid

tris(hydroxymethyl)aminomethane (Trizma base)

xii

INTRODUCTION

The branched-chain amino acids, leucine, isoleucine and valine are
essential in the diets of all mammals. The initial step in the catabo-
lism of these amino acids is a transamination to produce a-ketoisocapro-
ate, a-keto-a-methylvalerate and a-ketoisovalerate, respectively. The
a—keto acids are then decarboxylated by a branched-chain a-keto acid
dehydrogenase and converted to the acyl-CoA. A single enzyme, the
branched-chain a-keto acid dehydrogenase (EC 1.2.4.3 and 1.2.4.4), cata-
lyzes the decarboxylation and conversion of all three branched-chain
a-keto acids to their respective acyl-CoA's (1-4). However, there are
also reports in the literature of dehydrogenases and/or decarboxylases
which use only one or two of the three branched-chain a-keto acids as
substrates (5-10).

The occurrence of branched-chain acylcarnitines (11,12) and branched-
chain carnitine acyltransferase activities (11,13-15) in mammalian tis-
sues indicates that carnitine may be involved in metabolism of the carbon
skeletons of branched-chain amino acids. Solberg gt 31. (16) have shown
conversion of branched-chain GPKECO acids to branched-chain acylcarni-
tines by mitochondria of various rat tissues. Isobutyrylcarnitine is
metabolized by isolated beef and rat liver mitochondria (17).

A role for carnitine in branched-chain amino acid metabolism would be
particularly likely if some of the branched-chain acyl-CoA's were formed
outside of the matrix of mitochondria. The existence of glyoxylate:

1

leucine aminotransferase (18), carnitine acetyltransferase (19) and car-
nitine isobutyryl transferase (13) activities in peroxisomes indicates
that the metabolism of the branched-chain acyl residues may occur in more
than one cellular compartment. In order to investigate this possibility,
the subcellular distribution of the branched-chain a-keto acid decarboxy-
lating activities in rat liver was determined.

The data presented herein demonstrates the presence of at least two
branched-chain a-keto acid decarboxylating activities in rat liver: a
mitochondrial activity which decarboxylates all three branched-chain
a—keto acids, a-ketoisocaproate, a-keto-B-methylvalerate and a-ketoiso-
valerate and a soluble enzyme which decarboxylates only a-ketoisocaproate
and asketo-Yemethiolbutyrate.

The mitochondrial a-keto acid decarboxylating activity is due to the
mitochondrial branched-chain a-keto acid dehydrogenase. This enzyme
requires both CoA and NAD+ as cofactors. Recently, this enzyme has
been the subject of numerous investigations since it appears to be highly
regulated and the rate limiting step of branched-chain amino acid metabo-
lism in most tissues.

The cytosolic arketoisocaproate decarboxylating activity is different
from the mitochondrial branched-chain a-keto acid dehydrogenase. The
cytosolic enzyme does not require CoA or NAO+. Although there are pre-
vious indications of a soluble o-ketoisocaproate decarboxylase activity
(1,5,7-9), a detailed characterization of this activity is lacking.

The data presented herein shows that the cytosolic enzyme is an oxi-
dase which converts a-ketoisocaproate to B-hydroxyisovalerate. The pur-
ification of this enzyme and the determination of some of its physical

and kinetic properties are also presented. Speculation concerning the

3

possible role(s) of this enzyme in branched-chain amino acid metabolism

is covered in the discussion.

LITERATURE REVIEH

I. Metabolism of the Branched-Chain Amino Acids
A. Metabolic Pathway

The catabolism of leucine, isoleucine and valine occur via
very similar pathways. The amino acid is first transaminated
producing the a-keto acid. The apketo acid is then decarboxylat-
ed and converted to the corresponding acyl-CoA by a mitochondrial
branched-chain a—keto acid dehydrogenase. Further metabolism of
these acyl-CoA groups occurs inside the mitochondria. Leucine is
ketogenic since it is metabolized to acetyl-CoA and acetoacetate.
Isoleucine is only slightly ketogenic due to acetyl-CoA forma-
tion. Both isoleucine and valine are gluconeogenic since they
can be metabolized to intermediates of the tricarboxylic acid
(TCA) cycle. B-Hydroxy-B-methylglutaryl-CoA, formed during
leucine catabolism, may also be used in sterol synthesis
(20,21).

Transamination of leucine, isoleucine and valine is catalyzed
by a branched-chain amino acid transaminase. Ichihara gt 31.
(22-28) have isolated three isozymes of this enzyme. Isozyme I
uses all three branched-chain amino acids (BCAA) as substrates
and has been fbund in several tissues of rat and hog. Isozyme 11
uses only leucine and methionine as substrates and has only been
found in rodent liver. Isozyme III uses all three BCAA as

4

5
substrates and has only been found in brain, ovary and placenta
(29).

arKetoglutarate is the primary acceptor of amino groups from
the BCAA. Pyruvate and oxaloacetate are not utilized by the BCAA
transaminase purified from hog heart and brain or from rat liver
(24,30,31). Hhen rat diaphragm, atria or epididymal fat pad
slices are incubated with leucine, production of alanine and glu-
tamate plus glutamine increases (32-34). However, the transamin-
ase in rat skeletal muscle has a much greater affinity fer
a—ketoglutarate (Km 8 0.10 mM) than fer pyruvate (Km = 3.4
mM) (35). Therefbre, production of alanine is probably due to
transamination of glutamate with pyruvate. The branched-chain
a-keto acids, a-ketoisocaproate (a-KIC), a-ketoisovalerate and
arketo-s-methylvalerate can also serve as amino group acceptors
for the BCAA (31).

The branched-chain u-keto acids (BCKA) can be transported
into the mitochondria where they are decarboxylated and converted
to acyl-CoA's by the mitochondrial BCKA dehydrogenase. This
enzyme has been purified from rat liver (3) and frah bovine liver
and kidney (2,36). It uses all three BCKA as substrates, re-
quires both CoA and NAD+ as cofactors, and appears to operate
by a mechanism similar to that of pyruvate dehydrogenase.

Further metabolism of the acyl CoA's produced by the BCKA
dehydrogenase occurs in the mitochondria via a pathway very simi-
lar to B-oxidation. Little is known about the enzymes involved
in this pathway. Recently, a specific isovaleryl-CoA dehydrogen-

ase has been identified and isolated from pig and rat liver

B.

6
(37-39). This enzyme is different from the butyryl-CoA or gener-
al acyl-CoA dehydrogenases involved in B-oxidation. This enzyme,
therefore, may be specific fer metabolism of isovaleryl-CoA, a
metabolite of leucine. Isobutyryl-CoA, a metabolite of valine,

is not a substrate fer this isovaleryl-CoA dehydrogenase.

Regulation and Interorgan Distribution of Branched-Chain Amino
Acid Metabolism

The metabolism of the BCAA has been studied in a variety of
mammalian tissues both 1g 113g and lg 11339. Most amino acids
are taken up and metabolized by the liver (40), however, the BCAA
are a notable exception, showing very little uptake by or net
transport from the liver (41). Leucine, isoleucine and valine,
however, are rapidly taken up and oxidized by skeletal muscle
(42,43).

The initial transamination of BCAA occurs mainly in extrahe-
patic tissues (26,44,45). The tissue distribution of this enzyme
has been done in rat. Stomach and pancreas contain the highest
transaminase activity (46) followed by kidney, heart, brain, dia-
phragm, testes and skeletal muscle. Liver has very low trans-
aminase activity (26,44,45). On the other hand, the highest spe-
cific activity of the BCKA dehydrogenase in rat, monkey, guinea
pig and man occurs in liver and kidney (45,47,48). Skeletal mus-
cle has very little BCKA dehydrogenase activity. These data sug—
gest that the BCAA may be transaminated by peripheral tissues and
the resulting a-keto acids transported to the liver for fUrther

metabolism.

7

Although the BCKA dehydrogenase is apparently low in muscle
tissue, Goldberg gt al. (49) have suggested that this is the
major site fbr leucine catabolism, due to the large total mass of
muscle, and the low transaminase activity in liver. Indeed,
skeletal muscle is capable of oxidizing leucine completely to
002 (50,51). However, when rat atria or skeletal muscle are
incubated with leucine, approximately 20-40% of the a-KIC pro-
duced is released into the media (35,45,50,52,53). The BCKA
dehydrogenase, therefore, is probably the rate limiting step in
atria and skeletal muscle.

Livesey gt g1. determined arterio-venous differences of the
branched-chain a-keto acids in tissues of both rat (54) and man
(55) after protein ingestion or leucine infusion. In both cases
there was a net release of the BCKA by muscle and a net uptake by
liver. Noda and Ichihara (56) reported a significant depression
(14-60%) of total leucine oxidation in rats when the liver or
kidney vessels were ligated. These results are again consistent
with the involvement of both muscle and liver in BCAA metabo-
lism.

Recent investigations have revealed that the BCKA dehydrogen-
ase is highly regulated (57-67) most likely via a phosphoryla-
tion/dephosphorylation mechanism (59,61,62,66). Under the prOper
assay conditions, the BCKA dehydrogenase activity of heart or
muscle can be increased as much as 10 fold over basal levels
(60,67). Therefore, previous estimates of the BCKA dehydrogenase

activity in various tissues may have been carried out at

8
suboptimal conditions and must be reinvestigated under both phys-

iological and optimal conditions.

11. Effect of Carnitine on Branched-Chain Amino Acid Metabolism

Carnitine stimulates the oxidation of BCAA in rat skeletal muscle
(68) and heart (69) but not in liver (69,70). The lack of effect of
carnitine on the oxidation of BCAA by liver appears to be due to the
low level of BCAA transaminase in this tissue. Hhen a—KIC is used as
a substrate, carnitine does stimulate oxidation by liver homogenates
(70). Van Hinsbergh gt 31. (71) and May gt al. (72) have shown stim-
ulation of BCKA oxidation by carnitine in rat liver, heart, skeletal
muscle, and kidney. The effect of carnitine, therefbre, must be at
or distal to the BCKA dehydrogenase step.

Solberg and Bremer (73) have shown that when various rat or mouse
tissues are incubated with the BCKA and L-[methyl-3H1carnitine,
branched-chain acyl [3H-methyl] carnitines are formed. Branched-
chain acylcarnitines as well as branched-chain carnitine acyltrans-
ferase activities have also been identified in many tissues of beef
and rat (II-15,17). Since the branched-chain acyl-CoA's formed by
the BCKA dehydrogenase can cause product inhibition of this enzyme
(74), carnitine may stimulate BCAA metabolism by removing these acyl-
CoA groups.

Two hypotheses were proposed to account fbr the stimulatory
effect of carnitine. The first, proposed by Bieber and Choi (12) is
that carnitine may be involved in shuttling of acyl-CoA's produced by
the BCKA dehydrogenase across an intracellular permeability barrier

fer further metabolism. The second hypothesis, proposed by Van

9
Hinsbergh et al. (69), is that carnitine removes inhibitory branched-
chain acyl-CoA groups from the mitochondria and maintains free CoA
levels. Both of these effects would favor activation of the BCKA
dehydrogenase.

The first hypothesis would be supported if any of the BCKA dehy-
drogenase were located outside of the mitochondria or even located on
the outside of the inner mitochondrial membrane. Indeed, several
investigators have reported evidence of extramitochondrial BCKA
decarboxylase or oxidase activities (1,5-8). In addition the data of
Johnson and Connelly (5) suggests that the BCKA dehydrogenase is
located on the outside of the inner mitochondrial membrane.

Data presented herein, however, demonstrates that only one BCKA
dehydrogenase could be detected in rat liver and that this was ex-
clusively mitochondrial. Extramitochondrial a-keto acid decarboxy-
lating activities detected by other investigators (1,5-8) may be due
to an oxidase similar to the a-ketoisocaproate oxidase described in
this paper. Furthermore, Bremer and Davis (75) have shown that the
BCKA dehydrogenase must be located on the inside of the inner mito-
chondrial membrane since addition of CoA and NAD+ are not required.
fer optimal activity in intact mitochondria. Previous experiments by
Johnson and Connelly (5) utilized a ferricyanide assay which is quite
insensitive and subject to interfering reactions (76).

In light of these results, the second hypothesis seems more
attractive. Studies in rat heart (75), liver (72) and skeletal mus-
cle (71) indicate that carnitine stimulates the BCKA dehydrogenase by
removing inhibitory branched-chain acyl-CoA's. The addition of

isovalerylcarnitine decreases the oxidation of BCKA apparently due to

10

build up of intramitochondrial isovaleryl-CoA (71-72). Thus carni-

tine apparently stimulates BCAA metabolism by decreasing intramito-

chondrial acyl-CoA's and increasing the CoA/acyl-CoA ratio as well as

the availability of free CoASH. These affects would favor "activa-

tion“ of the BCKA dehydrogenase, which is believed to be the rate

limiting step in BCAA metabolism.

III. Specific Effects of Leucine on Metabolism

Leucine and/or its'metabolites affect several metabolic processes

which are not affected by isoleucine or valine. These include pro-

tein synthesis and degradation, insulin secretion, and glucose and

pyruvate oxidation. The a-KIC oxidase described herein uses the

a—keto analogue of leucine as a substrate, but does not use the

a—keto analogues of isoleucine and valine. Therefore, this enzyme

may be important in regulation of leucine levels separately from iso-

leucine or valine. In this section, therefore, I will describe the

recent research concerning the specific metabolic effects of leu-

cine.

A.

Effect of Leucine on Protein Turnover

Fulks, Li and Goldberg (77) fbund that all three BCAA togeth-
er, leucine alone, or a combination of leucine and isoleucine
stimulated protein synthesis and inhibited protein breakdown in
rat diaphragm. Buse gt al. (78), however, saw no effect of
valine or isoleucine alone on rat diaphragm. In heart, both leu-
cine and its metabolites stimulate protein synthesis and inhibit

protein degradation (79). In gastrocnemius muscle, however, only

11
leucine was effective; a-KIC, isovalerate and acetate had no
effect. Isoleucine and valine together had no effect on protein
turnover in either tissue (79). These results must be interpre-
ted with caution since uptake of leucine metabolites may differ
in heart vs. gastrocnemius muscle.

The exact mechanism of leucine's effect on protein turnover
is unknown. Goldberg and Tischler (80) fbund that cycloserine,
an inhibitor of leucine transamination, prevents the inhibition
of protein breakdown by leucine without affecting the stimulatory
effect of leucine on protein synthesis. Therefore, leucine
itself may regulate protein synthesis. Morgan gt al. (79) fbund
levels of protein synthesis to be approximately proportional to
intracellular leucine concentrations in heart and skeletal mus-
cle. Their studies indicate that leucine facilitates peptide
chain initiation as measured by a decrease in ribosomal sub-
units.

In order to exert its inhibitory effect on protein degrada-
tion, leucine must be metabolized. As mentioned previously,
inhibition of leucine transamination prevents leucine's inhibi-
tory effect on protein degradation. a-KIC mimics the effects of
leucine on protein degradation in rat diaphragm and heart
(79,80).

The catabolism of muscle protein is a serious problem in
patients with sepsis or trauma. Administration of a mixture of
BCAA has been found to be useful in preventing muscle wasting and

improving nitrogen balance in some of these patients (81).

B.

12

Effect of Leucine on Insulin Secretion

Leucine and a-KIC stimulate insulin secretion by rat pancre-
atic islets (82-84). Isoleucine, valine and their respective
a-keto acids, however, have no effect (82,85). It appears that
leucine must be converted to a-KIC in order to stimulate insulin
secretion (82) but the mechanism of action is unknown. It should
be kept in mind that in all of these studies, leucine or a-KIC
levels are above normal physiological values. Approximately 5 mM
a-KIC (82,84,86) or leucine (83) is required for significant
stimulation of insulin secretion. Normal plasma levels of leu-
cine vary between 0.1 to 0.5 mM and BCKA levels normally do not
exceed 0.08 mM (87).

Effect of Leucine on Glucose and Pyruvate Metabolism

Leucine inhibits formation of 14002 from D-[U-14C]
glucose and inhibits pyruvate oxidation in the fasted rat dia-
phragm and heart at concentrations between 0.2 and 1.2 mM
(50,88). At 0.5 mM, isoleucine and valine had no effect on pyru-
vate oxidation in these tissues. At higher concentrations, the
a-keto derivatives of both leucine and valine inhibit the oxida-
tion of pyruvate and/or a—ketoglutarate in a variety of tissues
(89-91).

Aminooxyacetate, an inhibitor of the BCAA transaminase, abol-
ishes leucine's ability to inhibit pyruvate oxidation. a-KIC,
however, inhibits pyruvate oxidation in the presence or absence
of aminooxyacetate (88). Thus, metabolism of leucine is required

fer its effect on pyruvate oxidation.

13
Using a perfused rat heart system, Haymack gt al. (60) have
shown that 0.5 mM a-KIC causes considerable inactivation of the
pyruvate dehydrogenase while activating the BCKA dehydrogenase.
The addition of 0.25 mM pyruvate to the perfusate caused appre-
ciable inhibition of the BCKA dehydrogenase and activated the
pyruvate dehydrogenase. Thus, these two dehydrogenases appear to

be regulated in an independent and reciprocal manner.

IV. Metabolic Diseases Associated with Branched-Chain Amino Acid Metabo-
lism

Several diseases associated with defects in branched-chain amino
acid metabolism have been recognized. These are characterized by the
accumulation of one or more of the BCAA and/or their metabolites in
blood or urine.

Maple syrup urine disease is a rare (but often misdiagnosed)
disease which occurs in early infancy. This disease is characterized
by a distinct maple syrup-like odor in the sweat and urine of the
child and the accumulation of the BCAA and BCKA in urine and serum
(92). This disease leads to mental retardation and usually death .
within a year. It is unknown whether the BCAA or BCKA are responsi-
ble fbr the clinical abnormalities. o-Ketoisocaproate and caketoiso-
valerate inhibit pyruvate and/or a-ketoglutarate oxidation by mito-
chondria isolated from rat brain, liver, kidney and heart (89-91).
Leucine is known to stimulate insulin secretion and may be responsi-
ble for the hypoglycemia seen in these patients. Snyderman (93) has
noticed that in all of the maple syrup urine disease patients they

studied, leucine was elevated to a much greater degree than the other

14
branched-chain amino acids. Of the three branched-chain amino acids,
leucine and its metabolites appear to be mainly responsible for the
observed neurological symptoms (94), hypoglycemia (95) and inhibition
of gluconeogenesis (96) seen in untreated cases.

Isovaleric acidemia has been reported in several children (97-
99). This disease is characterized by a “cheesy“ odor to the breath
and body fluids (92) and recurrent episodes of vomiting, acidosis and
coma occurring after excessive protein intake or infection. Fibro-
blasts from a patient showed a deficiency in the isovaleryl-CoA dehy-
drogenase (100,101) and the accumulation of isovaleric acid and its
glycine conjugate in the blood and urine (102). Interestingly, leu-
cine and a-ketoisocaproate do not accumulate (92). This may be due,
at least partly, to a loss of regulation and concomitant increase in
activity of the BCKA dehydrogenase (103). The isovaleryl-CoA dehy-
drogenase activity has been shown to be due to a specific enzyme
which can be separated from the butyryl-CoA and general acyl-CoA
dehydrogenase (37). This enzyme does not use isobutyryl-CoA, a meta-
bolite of valine, as a substrate. Therefore, the metabolic block
associated with this disease impairs leucine metabolism without
directly affecting isoleucine or valine metabolism.

Other metabolites also have been fbund to accumulate in patients
during isovaleric acidemic episodes. Tanaka 33 al. (104) detected
grossly elevated levels of lactate, acetoacetate, B-hydroxybutyrate
and B-hydroxyisovalerate in urine of these patients.

Other metabolites of the branched-chain amino acids have also
been found to accumulate in certain metabolic disorders. These dis-

orders are named by the metabolite that accumulates, fer example,

15
B-methylcrotonylglycinuria, propionic acidemia and methylmalonic
acidemia (105). In B-methylcrotonylglycinuria, the defect appears to
be at the a-methylcrotonyl-CoA carboxylase, which can be due to
either a defective enzyme (106) or a biotin deficiency (107). In one
patient, with no a-methylcrotonyl-CoA carboxylase activity, large
amounts of a—ketoglutarate and B-hydroxyisovalerate as well as
B-methylcrotonylglycine were excreted in the urine (106), but no
ketosis was present. In a biotin responsive patient, however,
a-hydroxybutyric acid, B-hydroxyisovaleric acid, lactic acid,
a-hydroxyisobutyric acid and s-hydroxypropionic acid were all elevat-
ed along with B-methylcrotonylglycine (105). This is due to the bio-
tin requirement of propionyl-CoA carboxylase and pyruvate carboxylase
as well as that of B-methylcrotonyl-CoA carboxylase.

Metabolic disorders of branched-chain amino acid metabolism may
also be secondary to other abnormalities. Landaas (108,109) has
shown increased levels of s-hydroxyisovaleric acid, a-hydroxyisobuty-
ric acid and 2-methyl-3-hydroxybutyric acid in patients with keto-
acidosis from a variety of causes. These B-hydroxy acids are fbrmed
from leucine, valine and isoleucine, respectively. A positive cor-
relation was found between urinary levels of B-hydroxybutyric acid
and a-hydroxyisovaleric acid (108). Increased levels of all three
branched-chain amino acids are also seen in ketoacidotic patients
(108,110). The mechanism by which branched-chain amino acids and

their B-hydroxy metabolites are elevated is unknown.

V.

16
Evidence for the Existence of an Extramitochondrial Branched-Chain
arKeto Acid Decarboxylase Activity.

The existence of a single mitochondrial BCKA dehydrogenase which
utilizes the a-keto acids of leucine, isoleucine and valine as sub-
strates is now widely accepted. However, dehydrogenases or decar-
boxylases, which use only one or two of the BCKA as substrates, have
been reported (5-10).

Hohlhueter and Harper (1) first reported decarboxylation of
a-[1—14CJ KIC by a soluble preparation of rat liver. They recog-
nized that this activity was not due to “leaked“ mitochondrial BCKA
dehydrogenase, since deliberate attempts to release the enzyme from
mitochondria were unsuccessful. Under their assay conditions, only
8% of the total a-[I-14C] ketoisocaproate decarboxylating activ-
ity was soluble, the rest was mitochondrial.

In bovine liver, Johnson and Connelly (5) identified a small but
significant amount of a-[1-14c] K10 and a—[I-14C] keto-3-
methylvalerate decarboxylating activity in the soluble fraction.
a-Ketoisovalerate was not decarboxylated by this fraction. This
activity did not require the addition of CoA. However, they did not
eliminate the possibility of endogenous CoA in their preparation.

Grant and Connelly (7) have shown that a—KIC, but not a-ketoiso-
valerate or a-keto-p-methylvalerate, is decarboxylated by cytosolic
preparations from liver and kidney of mouse, rat, rabbit, guinea pig
and beef and from the liver only of chicken. This decarboxylase
activity did not require added 00A or NAD+ and showed very little
activity with pyruvate, a-ketoglutarate, a-ketoisovalerate, or

a—keto-a-methylvalerate as substrates (8).

17

An arketoisocaproate decarboxylase activity has also been iso-
lated from beef liver (9). This activity does not require CoA and
NAO+ as substrates and does not use the other two BCKA as
substrates. This activity is also insensitive to arsenite, an inhib-
itor of the mitochondrial BCKA dehydrogenase. Properties of this
enzyme are very similar to those of the a-ketoisocaproate oxidase
described herein.

The cytosolic arketoisocaproate decarboxylating activity appears
to be widely distributed among mammals (7). Previous studies indi-
cate that this activity is quite low in comparison to the mitochon-
drial BCKA dehydrogenase (1,5). However, cytosolic decarboxylation
of a-ketoisocaproate has been shown to increase under conditions
which cause high circulating levels of the branched-chain amino
acids, fer example, high protein meal (111) or diabetes (112). This
enzyme activity therefore may be important in BCAA metabolism when

serum or tissue levels of the BCAA are elevated.

MATERIALS

The materials used in this study are listed below with the source and
catalog number. All other chemicals used were of analytical reagent
grade or the finest commercially available.

DEAE (diethylaminoethyl)-cellulose, Hhatman DE52, reeve angel, Clifton,

NJ

[1-14C] sodium isovalerate, ICN Pharmaceuticals Inc., Irvine, CA
(Cat. No. 12128)

L-[1-14cJ Leucine, ICN Pharmaceuticals Inc., Irvine, CA (Cat. No.
10088) or New England Nuclear, Boston, MA (Cat. No. NEG-169)

L-[U-14C])Leucine, ICN Pharmaceuticals Inc., Irvine, CA (Cat. No.
10089

L-[4,5-3H] Leucine, ICN Pharmaceuticals Inc., Irvine, CA (Cat. No.
20036) or Schwarz/Mann, Orangeburg, NY (Cat. No. 0332-313344)

L-[1-14C] Methionine, Amersham Corp., Arlington Heights, IL (Cat. No.
CFA 433)

Oxygen-1802, Merck and Company Inc., Rahway, NJ (Cat. No. MO-1454)
90 atom % 0-18, Lot No. 55-F

Phenyl-sepharose CL-4B, Sigma, St. Louis, MO (Cat. No. P-7892)

Sephacryl S-ZOO Superfine, Pharmacia, Piscataway, NJ (Cat. No. 17-0871-
01)

Sephadex G-150, Pharmacia, Piscataway, NJ (Cat. No. 17-0070-01)
L-[1-14c1 Valine, New England Nuclear, Boston, MA (Cat. No. NEC-17I)

Hater-180 (normalized) Merck and Company Inc., Rahway, NJ (Cat. No.
MO-1670) 97 atom % 180, Lot No. 2164-E

18

1.

METHODS

Sucrose Gradient Separation of Rat Liver Organelles

Male Sprague-Dawley rats, 300 gm, were fasted 2 days and then
sacrificed by decapitation. Subcellular organelles from liver were
separated by isopycnic density centrifugation as previously described
(113) except that a 1 to 10 (instead of 1 to 20) homogenate of rat
liver was used and the buffer for both the grinding medium and the
sucrose gradient was 1 mM sodium phosphate, pH 7.5. Organelles were
separated on a 600 mL zonal gradient (Figure 1B) or a 60 mL tube gra-
dient (Figure 1A). The zonal gradient was identical to the 60 mL
tube gradient except that the volume of all the sucrose solutions and
amount of rat liver homogenate applied to the gradient were increased
10 times. Fractions were collected and assayed.

Aliquots for assay of o-KIC decarboxylase activity were stored at
-80°C until assayed. a-Ketoisovalerate decarboxylase activity was"
measured on aliquots that were stored overnight at 4°C. Catalase
(114), fUmarase (115) or glutamate dehydrogenase (113), and NADPH-
cytochrome c reductase (116), the marker enzymes for peroxisomes,
mitochondria and microsomes, respectively, were assayed by the cited
methods with the following modifications. To rupture the organelles,
0.1% Triton X-100 was added to catalase and fumarase assays. The
assay of NADPH—cytochrome c reductase included 0.01 mM rotenone and

0.68 mM sodium azide to inhibit mitochondrial electron transport.
19

20
II. Preparation of Radioactively Labeled orKeto Acids

14C or 3H labeled o—keto acids were prepared from the cor-
respondingly labeled L-amino acids and purity determined according to
the method of Rudiger gt al. (117). Solutions of 1-14C-labeled ‘
arketo acids were made up to the desired concentration by addition of
unlabeled o—keto acids and stored at -80°C.

a-[4,5-3H] KIC was purified prior to use to remove contaminants
which eluted in the water wash and with isovaleric acid from Dowex-l
chloride columns. The contaminants amounted to approximately 20% of
the total cpm. To purify a-[4,5-3H] KIC, peak fractions from the
Dowex-H+ column (117) containing o-[4,5-3H] KIC were neutralized
and applied to a 0.5 x 4.5 cm Dowex-I chloride (100-200 mesh) col—
umn. The column was washed with water fbllowed by a 200 mL linear
gradient of 0-0.2 N HCl and 5 mL fractions were collected. Fractions
containing only a-[4,5-3H] KIC were pooled, pH adjusted to 5.0 and
lyophilized. In order to remove the large amount of salt present,
the residue was resuspended in 0.02 N HCl, saturated with NaCl, and
extracted with l-octanol. Ether was not used for extraction because
much of the QPKIC was decarboxylated, probably due to contaminating~
peroxides in the ether. The octanol layer was re-extracted with
water and adjusted to pH 6. The aqueous layer containing the GP
[4,5-3H] KIC, showed only one peak when rechromatographed on the
Dowex-l chloride column.

No contaminating substances were detected when QPEU-14C] KIC
was chromatographed on Dowex-I chloride columns. This compound was

therefore used without further purification.

21
III. Assay of drKetoisocaproate Oxidase Activity
A. Assay of a-[1-14C] Keto Acid Decarboxylating Activities

Assays for determination of arketo acid decarboxylating
activities were carried out in 1.5 cm diameter culture tubes,
which were tightly stoppered with a serum cap. A plastic cup
(Kontes) hanging from the serum cap contained 0.2 mL Hyamine
(methylbenzethonium hydroxide) to trap 14C02 released dur-
ing the reaction. The reaction conditions were varied throughout
this work in order to establish optimal conditions for the assay.
Two sets of reaction conditions were widely used and are de-
scribed below. Oecarboxylase assays using other conditions are

described in the figure and table legends.

Method A:

Reaction mixture A contained in a final volume of 0.4 mL; 0.2
M Tris base, 0.2 M maleic acid, pH adjusted to 6.5 with NaOH, 1.5
M ammonium sulfate, 1.0 mM NaZCO3, 1.0 mM ar[1-14C] KIC
(or other a-keto acid) (50-100 dpm/nmole), and 25 to 100 uL of
rat liver cytosol preparation or partially purified arKIC oxi-.
dase. This reaction mixture was used for initial studies of the
QPKIC oxidase in crude rat liver preparations and in initial
attempts at purification. All components of the reaction mixture
except the enzyme and c-[1-14C] keto acid were equilibrated
at 25°C. The enzyme was then added to all assays approximately
8-15 min before addition of the a-[1-14c1 keto acid.

22

All reactions were initiated by the addition of the o—[I-
14C] keto acid. In some cases the d—keto acid was injected
through the serum cap with a syringe. However, usually the
GPKECO acid was added directly to the reaction mix and then the
tube was quickly stoppered with a serum cap. A cup hanging from
this cap had previously been filled with 0.2 mL Hyamine.

Incubations were for 60 min (unless noted otherwise) in a
25°C shaking water bath. To terminate the reactions, 0.2 mL of
20% trichloroacetic acid was added and an additional hour with
shaking was allowed fbr collection of 14C02. The cup plus
Hyamine was put into a scintillation vial and counted in 10 mL of
scintillation fluid (5). Specific activity was determined by
releasing all of the 14coz from the a-[1-14c1 keto acid
using ceric sulfate (118). Nonenzymatic decarboxylation of the
a-keto acids was determined by replacing the rat liver prepara-
tion with buffer in the assay. The blank value determined in
this way was the same as when boiled enzyme preparations were

used and is subtracted from all values reported herein.

Method 8:

Reaction mixture B contained in a final volume of 0.4 mL; 0.2
M Tris base, 0.2 M maleic acid, pH adjusted to 6.5 with NaOH, 1.0
nM FeSO4, 0.5 nM ascorbic acid, 1.0 nM dithiothreitol, 1.0 nM
o—[1-14C] KIC (or other a—keto acid) (approximately 100
dpm/nmole), and 5 to 100 uL of the sample to be assayed. A stock
solution containing 16 mM FeSO4, 8 mM ascorbic acid, and 16 mM
dithiothreitol (cofactor mix) was prepared fresh daily. Addition

23
of 25 uL of this cofactor mix per assay gave the desired final
concentrations of FeSO4, ascorbic acid and dithiothreitol (1.0
mM, 0.5 mM, 1.0 mM). The reaction was preincubated for 1 hour at
25°C (unless noted otherwise) with all components excluding the
a-[I-14C] keto acid. The a-[1-14C] keto acid was then
added to initiate the reaction and the assay continued as in
Method A.

B. Measurement of Oxygen Consumption in Presence of QPKECO Acids
The reaction mix contained in a final volume of 3.0 mL; 0.2 M
Tris, 0.2 M maleate, pH 6.5, 1.5 M ammonium sulfate, 1.0 mM
Na2C03 and 0.5 mL of rat liver cytosol preparation (see
below). Oxygen consumption was measured with a Yellow Springs
Instruments oxygen monitor. After establishing a basal rate of
02 consumption the o—keto acid was added to give a final con-
centration of 1 MM and the change in the rate of 02 consumption
was recorded. All reactions were carried out at 25°C.
Oxygen solubility in the reaction mix was determined by the
method of Robinson and Cooper (119). A value of 96 1 10 nmol
OZ/mL was fbund. This value was used fer all polarographic

measurements.

IV. Isolation of Rat Liver Cytosol Preparations
Male Sprague-Dawley rats, 150-200 g, were fasted for 24-48 h and
then sacrificed. The livers were homogenized using a Potter-Elvehjem
glass homogenizer in 3 vol of 0.25 M sucrose buffer (0.25 M sucrose

in 2.5 nM Hepes [N-Z-hydroxyethyl pi perazine-N'-2-ethanesul fonic

V.

24
acid], pH 7.4, 0.25 m EDTA). The homogenate was centrifuged 12 min
at 500 x g and the pellet discarded. The 500 x g supernatant fluid
was centrifuged 12 min at 20,200 x g. The 20,200 x g supernatant
fluid was centrifuged 1.5 h at 70,000 x g. The supernatant fluid was
decanted and the volume reduced by ultrafiltration. The concentrated
fluid was dialyzed for 4 h against several changes of 33 mM potassium
phosphate buffer, pH 7.4. This preparation is referred to as rat
liver cytosol and was stored at -80°C. Protein was determined by the
Lowry method (120) as previously modified (121).

Purification of Rat Liver Cytosolic QPKECOISOCGPFOEtE Oxidase
Several purifications of the rat liver a—ketoisocaproate oxidase

were carried out. Those which are pertinent to this work are

described below. Protein was determined in purification fractions

using the method of Bradford (122). a—KIC oxidase activity was mon-

itored by decarboxylation of a-[1-14C] KIC as previously

described. All purification steps were carried out at 4°C unless

noted otherwise.

A. Purification A: Partial Purification of Rat Liver Cytosolic
drKetoisocaproate Oxidase
Sprague-Dawley rats were decapitated and the livers removed

and homogenized with a Potter-Elvehjem glass homogenizer in 5-10
vol of 0.25 M sucrose. The homogenate was centrifuged 12 min at
500 x g and 12 min at 20,000 x g and the supernatant saved. This
was stored at -80°C until used. The 20,000 x g supernatant frac-
tion was thawed and centrifuged at 70,000 x g for 1.5 h. Lipids

25
were aSpirated from the top of the sample and the supernatant
fluid was saved. This “70,000 x g supernatant" fraction was
stored at -80°C overnight. After thawing, one-tenth volume of
cold 2% protamine sulfate was slowly added while stirring at 4°C.
After stirring fbr 15 min, the preparation was centrifuged for 18
min at 13,000 x g. The “0.2% protamine sulfate supernatant" was
brought to 35% saturation with the slow addition of powdered
amnonium sulfate, stirred for 15 min at 4°C, and centrifuged 18
min at 13,000 x g. The “35% (NH4)ZSO4 supernatant" was
stored at ~80°C.

The d-KIC oxidase activity was stable at -80°C fbr several
weeks, but 80% of the activity was lost in 4 days at 4°C. One
percent isopropanol completely stabilized the activity in the
“35% (NH4)2SO4 supernatant fraction“ fer at least 1 week at
4°C. Up to 1% isopropanol in the assay reaction mix did not
affect the GPKIC oxidase activity.

The frozen “35% (NH4)ZSO4 supernatant“ fraction was
thawed and dialyzed against 10 liters of 20 nM Tris-HCl, pH 7.8,
1% isopropanol fer 48 h, with two changes of buffer. The dialy-
sate was centrifuged 30 min at 13,000 x g to remove precipitated
material. The supernatant, "pre-DEAE dialysate“ was stored for 4
days at -80°C and then applied to a 4 x 40-cm DEAE-cellulose
(Hhatman DESZ) column, equilibrated with 20 mH Tris-HCl, pH 7.8,
1% isopropanol (Buffer A). This was washed with 2 liters of the
equilibration buffer fellowed by a 2-liter linear gradient of 0
to 0.5 M NaCl in Buffer A. The OPKIC oxidase activity eluted at
approximately 0.8 M NaCl and peak fractions were pooled ("DEAE-

26
cellulose pool"). This was stored 4 days at -80°C and then con-
centrated to 9.8 mL with an Amicon PM 30 ultrafiltration filter
("concentrated DEAE pool"). The concentrated DEAE pool was cen-
trifuged at 13,000 x g for 10 min to remove precipitated material
and applied to a 4.8 x 82-cm Sephadex 6150 (40-120 um) column
which was equilibrated with Buffer A containing 0.2 M NaCl. The
column was eluted with the same buffer at a flow rate of approx-
imately 40 mL per hour and 9.4-mL fractions were collected. The
GPKIC oxidase activity was retained by the column and eluted
after catalase activity. The peak fractions were pooled and
stored at -80°C. After thawing, the sample was concentrated to
13.6 mL with an Amicon PM 30 filter and again stored at -80°C
("Sephadex G-150 pool").

Purification 8: Final Purification of o-Ketoisocaproate Oxidase
The 10,000 x g supernatant fractions (in 0.25 M sucrose, 1%
isoprOpanol) from a number of rat liver preparations were stored
at -80°C. These are stable for at least two years.
Powdered (NH4)ZSO4 was slowly added to 7.2 L of the
10,000 x g supernatant fraction while stirring until 45% of sat-
uration was achieved. After stirring for an additional 30 min,
the preparation was centrifuged 30 min at 10,000 x g. The “45%
(NH4)2504 supernatant“ was brought to 75% saturation with
the slow addition of powdered (NH4)ZSO4, stirred an addi-
tional 30 min, and centrifuged 30 min at 10,000 x g. The pellet
(45-75% (NH4)ZSO4 fraction) was resuspended in Buffer A.
This fraction was then dialyzed 60 hrs against 10 L of Buffer A,

27
with 5 changes of buffer and centrifuged 15 min at 8,000 x g to
remove precipitated material (Pre-DEAE dialysate).

The "Pre-DEAE dialysate" was applied to a 4.8 x 83 cm DEAE
cellulose (Hhatman 0E52) column, equilibrated with Buffer A.

This was eluted with 2.5 L of Buffer A fellowed by a 9 L linear
gradient of 0-0.2 M NaCl in Buffer A. The aPKIC oxidase activity
eluted at approximately 0.06 M NaCl and fractions containing
activity were pooled (“DEAE cellulose pool“), concentrated to 250
mL using an Amicon PM 10 ultrafiltration membrane (“concentrated
DEAE pool“), and stored at -80°C.

In previous purifications (i.e.: Purification A) large
losses of a-KIC oxidase activity occurred when using DEAE, phenyl
sepharose and sephacryl columns during later stages of the puri-
fication. In passage over Sephacryl S-ZOO only 10% of the
applied activity was recovered. Protein containing fractions
from the NaCl gradient elution of the initial DEAE column, which
eluted befbre and after the peak of o—KIC oxidase activity, were
pooled (“DEAE side fractions“). Hhen columns were pretreated
with the “DEAE side fractions“, 90% of the applied arKIC oxidase
activity was recovered. Therefore, all columns (except the first
DEAE column) were pre-treated with the “DEAE side fractions“ and
then washed extensively with the elution buffer until no more
protein was detected in the elluent.

Stability studies (see Results) also demonstrated that 5%
monothioglycerol stabilized the orKIC oxidase activity. Subse-
quently, 5% monothioglycerol was included in all buffers.

28

A portion of the "concentrated DEAE pool“ (116 mL) was
adjusted to 5% monothioglycerol and 2.5 M NaCl by slowly adding
these compounds while stirring (Pre-phenyl sepharose). This was
applied to a 4.0 x 40 cm phenyl sepharose CL-4B column (Pharma-
cia) which had been pretreated with 1.5 L Buffer A, 1.0 L of Buf-
fer A containing 5% monothioglycerol, 2.5 M NaCl, 1.8 L of ”DEAE
side fractions“ containing 2.5 M NaCl, 2 L Buffer A, and equili-
brated with 1.5 L Buffer A containing 5% monothioglycerol, 2.5 M
NaCl. The a—KIC oxidase activity was eluted with 1.5 L Buffer A
containing 5% monothioglycerol, 2.5 M NaCl, followed by a 2 L
linear gradient of 2.5-0 M NaCl in Buffer A containing 5% mono-
thioglycerol. Fractions containing a-KIC oxidase activity were
pooled and concentrated to 20 mL (Phenyl pool concentrate).

The “phenyl pool concentrate“ was then applied to a 4.8 x 82
cm Sephacryl S-200 column which had been pretreated with "DEAE
side fractions“ and equilibrated with Buffer A containing 5% mon-
othioglycerol, 0.1 M NaCl. The column was eluted with the equil-
ibration buffer at a flow rate of 13.6 ml/hr. Fractions contain-
ing orKIC oxidase activity were pooled (Sephacryl S-200 pool) and
stored at -80°C.

Since monothioglycerol at concentrations greater than 0.6%
inhibits the GPKIC oxidase activity (see Results), only small
aliquots (IO-20 mL) of the Sephacryl S-200 pool could be assayed
accurately. In some experiments requiring larger amounts of this
enzyme, the monothioglycerol was removed from the Sephacryl S-200
pool. This was accomplished by passing 0.5 mL of the Sephacryl
S-200 pool over a 0.75 x 14 cm Bio Gel P-6 (50-100 mesh) column

C.

29
which was pretreated with "DEAE side fractions“ and equilibrated
with Buffer A containing 0.1 M NaCl. Protein was monitored by
adding 50 uL of Coomassie Blue Reagent (122) to 10 uL of each
fraction and visually checking for blue color. To monitor mono-
thioglycerol, 1 mL of 0.3 mM DTNB (5,5'-dithiobis-Z-nitrobenzoic
acid) in .25 M glycylglycine pH 8.2, was added to the same assay
tubes and yellow color checked visually. Fractions containing
protein, but no monothioglycerol, were pooled and stored at -80°C
(P-6-Pool). The purified d-KIC oxidase appears to be stable for
at least 1 week in the absence of monothioglycerol if kept at

’800Co

Purification C

The GPKIC oxidase activity is fairly stable (20% loss in 10
days) in crude preparations, however this enzyme becomes very
unstable as it is purified. Several purifications were done as
part of a series of stability studies. One of these, which is
pertinent to the experiments in this thesis is described here.
This purification was attempted before Purification B and is
essentially the same except fer the fbllowing changes.

A portion (200 mL) of the “concentrated DEAE pool“ was thawed
and centrifuged to remove precipitated material. To this por-
tion, 50 mL of 95% ethanol at -20 to -30°C was slowly added while
stirring on an ice bath. The temperature of the solution was
maintained below -5°C. After complete addition of ethanol, the
solution was stirred an additional 20 min, centrifuged at 10,000
x g and the pellet discarded. To this "25% ethanol supernatant",

30
150 mL of 95% ethanol (-20 to -40°C) was added slowly. The solu-
tion was allowed to stir an additional 20 min after complete
addition of ethanol and centrifuged 20 min at 10,000 x g. The
pellet was resuspended in Buffer A containing 2.5 M NaCl by stir-
ring overnight at 4°C. A large portion of the pellet did not re-
dissolve. This fraction was then centrifuged 20 min at 10,000
x g and the supernatant (25-IOO% ethanol fraction) stored at
-80°C. The yields of o-KIC oxidase activity from the ethanol
precipitation step were variable; therefbre this step was omitted
in the final purification (Purification B).

The “ZS-100% ethanol fraction” was thawed and applied to a
phenyl sepharose column and the purification continued as in
Purification B. Buffers used to elute phenyl sepharose and
Sephacryl S-200 columns did not contain 5% monothioglycerol.
These columns were pretreated with the “DEAE side fractions".

The ”Sephacryl S-200 pool“ was dialyzed 35 hr against 400 mL
of Buffer A containing 1% monothioglycerol, with 3 changes of
buffer. The dialysate was aplied to a 0.8 x 12 cm DEAE (Hhatman
DE-SZ) column which had been pre-equilibrated with Buffer A
containing 1% thioglycerol. Note that this column was not
pretreated with “DEAE side fractions“. The column was washed
with 60 mL of Buffer A containing 1% monothioglycerol and eluted
with a 180 mL linear gradient 0-0.1 M NaCl in the same buffer.
Fractions containing o—KIC oxidase activity were stored

separately at -80°C (“2nd DEAE column").

31
VI. Analysis of Reaction Products

The reaction mix used for analysis of a—KIC oxidase reaction pro-
ducts contained 0.2 M Tris, 0.2 M maleate, pH adjusted to 6.5 with
NaOH, 1 nM dithiothreitol, 0.2 m‘i FeSO4, 0.4 nM ascorbic acid, 1 nM
Na2C03, 1 m a-kIc, 0.3 uCi a-[4,5-3H] KIC, 0.1 uCi a-[1-
14C] KIC, and 0.2 mL of the “Sephadex G—150 pool" (Table 5) in a
final volume of 0.8 mL. The reaction was followed by measuring the
loss of a-[1-14C] KIC and was terminated at 4.5 h with 0.4 mL of
20% trichloroacetic acid. The protein was removed by centrifugation
and the supernatant fluid was diluted with 20 mL of H20 and neu-
tralized with NaOH. This was applied to a 0.7 x 37-cm Dowex-I x 8
chloride (200-400 mesh) column and washed with 70 mL H20. The col-
umn was then eluted with a 300 mL linear gradient of 0 to 0.01 N HCl,
4 mL fractions were collected, and dpm of 3H and 14C were
determined by double-label counting techniques. The fractions cor-
responding to the largest 3H peak (peak I, Fig. 29) were pooled,
the pH adjusted to 7 and lyophilized. The residue was resuspended in
2 mL of 2 N HCl, saturated with NaCl, and extracted 10 times with 1
mL diethyl ether; 90% of the 3H was extracted into the ether layer.
The ether was then evaporated to approximately 50 uL under N2. The
sample was analyzed on a Hewlett-Packard 5830 A gas chromatograph
equipped with a stream splitter which divides the sample eluted from
the gas chromatograph column between the flame detector and an exter-
nal outlet tube. Samples were collected every minute from the outlet
tube by condensation in a cold Pasteur pipet. The sample was rinsed
from the pipet with 10 mL of scintillation fluid (5) into a scintil-
lation vial and counted. A 6-ft x Z-mn i.d. glass col um packed with

32

15% SP 1220, 1% H3PO4 on 100/120 Chromosorb HAH, was used with a
variable temperature program. The temperature was kept at 110°C for
5 min followed by a 3°C per minute increase up to 140°C. Injection
temperaure was 180°C and N2 was the carrier gas. The peak contain-
ing radioactivity was identified by using the same gas chromatograph
system (except the carrier gas was He) in line with a Hewlett-Packard
mass spectrometer. These analyses were performed by the mass spec-
trometry facilities at Michigan State University, under the supervis-

ion of C.C. Sweeley and J. Hatson.

VII. Polyacrylamide Gel Electrophoresis Using Denaturing or Non-Denatur-
ing Conditions

SOS (sodium dodecyl sulfate) polyacrylamide gel electrophoresis
was carried out on slab gels according to the procedure of Laemmli
(123). Gels contained 10% acrylamide. Staining with Coomassie blue
dye was carried out according to the procedure of Bonner and Laskey
(124).

Tube gels containing 7.5% acrylamide were prepared for native gel
electrophoresis as described (123) except SDS was omitted. Gels were
pre-run overnight at a constant current of 2 mA per gel. The cathode
chamber contained .375 M Tris, 0.08% L-cysteine pH 9.0 and the anode
chamber 2.5 mM Tris/glycine pH 8.3, 0.008% L-cysteine during this
pre-run. After this pre-run the buffers were replaced with 25 mM
Tris/glycine pH 8.3, 0.08% L-cysteine (cathode chamber) and 2.5 mM
Tris-glycine pH 8.3, 0.008% L-cysteine (anode chamber). The sample
was applied in a solution containing 10% glycerol, 0.003% bromophenol

blue. Electrophoresis was carried out at a constant current of 1 mA

33

per gel for 1 hr, 2 mA per gel for 1 hr, and 3 mA per gel for the
final 3 hrs. The gels were stained with Coomassie blue dye as
described (124). arKetoisocaproate oxidase activity was monitored in
some of the gels which were not stained. These gels were sliced into
2 mm sections, each section put into a separate test tube, and 50 uL
of 20 "M Tris HCl pH 7.8, 1% isopropanol, 0.1 M NaCl, 5% monothiogly-
cerol added. These were shaken for 2 days at 4°C and then assayed

for a-KIC oxidase activity (see Method B).

VIII. Determination of Molecular Height of the a-Ketoisocaproate Oxidase
by Sephacryl S-ZOO Chromatography
A 1.6 x 63 cm Sephacryl S-ZOO column was equilibrated with 20 mM
Tris HCl pH 7.8, 1% isopropanol, 0.1 M NaCl. The column was then
treated with 40 mL of the "DEAE side fractions" (see Section VB,
Methods) and washed extensively with equilibration buffer. To cali-
brate the column, bovine serum albumin, ovalbumin, and cytochrome c
were applied separately to the column, eluted with the equilibration
buffer and their elution volumes determined by measuring 00280
or 00410 (cytochrome c). The flow rate of the column was 1.7
ml/hr. In a separate experiment, 0.5 ml of the purified aPKIC oxi-
dase (Sephacryl S-200 pool, Purification B) was applied to the column
and eluted with the equilibration buffer containing 5% monothiogly-
cerol. d-KIC oxidase activity was measured in fractions by Method B.
The data was plotted and molecular weight determined by the method of
Andrews (127).

34
IX. Enzymatic Incorporation of 180 Into B-Hydroxyisovaleric Acid
A. Incubations with 1302

In 1802 incorporation experiments, the reaction mixture
contained in a final volume of 0.30 mL; 0.2 M Tris base, 0.2 M
maleic acid, pH adjusted to 6.5 with NaOH, 1 mM FeSO4, 0.5 mM
ascorbic acid, 1 mM dithiothreitol, 2 mM a-[u-14c1 KIC (.025
uCi) and 0.1 mL of purified a-KIC oxidase (P-6-Pool, see purifi-
cation B).

All components except the a—[U-14C] KIC were added to a
culture tube and frozen in a acetone, dry ice bath. Hhile froz-
en, the a-[u-14c1 KIC was added and this was also frozen.

The culture tube was then capped with a serum cap and the air
removed by evacuation. The tube was flushed with N2, thawed,
refrozen, and evacuated again. This process was repeated 4 times
to insure all 02 was removed. The culture tube was then filled
with 90 atom % 1302 (approximately 15 mL) via a syringe.

The contents of the culture tube were then thawed and incubated 3
h at 25°C. At the end of the reaction, a-hydroxyisovaleric acid
was isolated and analyzed as described below.

The ratio of N2 to 02 was determined in 1 mL of the gas
phase of the reaction vessel at the start and end of the reaction
to insure that air was not leaking into the reaction vessel.
Oxygen and nitrogen were analyzed on a Varian 3700 gas chromato-
graph using a 2 mm x 2 m stainless steel column packed with
Molecular Sieve 5A. The injection temperature was 40°C and col-
umh temperature was 40°C isothermal. The flow rate of helium,

the carrier gas, was 30 cm3/min. Detection was achieved with a

35
thermal conductivity detector at 100°C with a filament tempera-
ture of 190°C. The instrument output was calibrated with a
standard mixture of 1% 02:99% N2. 1802 dilution by

1502 due to leakage was calculated to be less than 3%.

Incubations with H2180

The reaction mixture fer H2180 experiments was the same
as for 1802 experiments except the reaction contained 93%
H2130. In order to reduce H2150, 0.187 mL of 0.32 M
Tris, 0.32 M maleic acid pH adjusted to 6.5 with NaOH, 18.7 uL of
a solution containing 16 nM FeSO4, 8 nM ascorbic acid, and 16
mM dithiothreitol, and 0.2 mL of purified a-KIC oxidase (P-6-
Pool) were combined, frozen and lyophilized. To the lyophilized
powder, 300 an. of 97% H2180 and 11 m. of 80 m4 a-[u-
14C] KIC were added and all components mixed well. The tube
was then capped with a serum cap, gassed with oxygen for two min
and incubated 3 h at 25°C. a-hydroxyisovaleric acid was then

isolated and analyzed as described below.

Isolation of p-Hydroxyisovaleric Acid

Isotopic oxygen incorporation experiments were terminated by
passing the reaction mixture over a 0.5 x 5.0 cm Dowex-I chloride
(200-400 mesh) column. The column was washed with 5 mL H20 and
a-hydroxyisovaleric acid eluted with six 1 mL portions of 0.02 N
HCl. The fraction containing B-[U-14C] hydroxyisovaleric
acid was saturated with NaCl, and extracted three times with an

equal volume of diethyl ether. The ether extractions were pooled

36
and concentrated to approximately 50-100 uL under N2. s-Hy-
droxyisovaleric acid was analyzed by gas chromatography-mass
spectrometry as described in Section VI of Methods except the gas
chromatograph column temperature was 150°C, isothermal. B-Hy-
droxyisovaleric acid had a retention time of 7.7 min under these

conditions.

RESULTS

I. Subcellular Distribution and Partial Characterization of an d-Keto-
isocaproate Oxidase of Rat Liver
A. Subcellular Distribution of Branched-Chain d-Keto Acid Decarboxy-
lating Activities in Rat Liver
Peroxisomes, mitochondria and microsomes from male rat liver
were separated on a sucrose gradient as shown in Figure 1. The
subcellular distribution of d-[1-14C] ketoisocaproate (a—KIC)
and a-[1-14C] ketoisovalerate decarboxylase activities in rat
liver using assay conditions optimized for the mitochondrial
branched-chain a-keto acid dehydrogenase are shown in Figure 1A.
All of the d-[1-14C] ketoisovalerate decarboxylating activ-
ities coincided with the mitochondrial marker, fumarase. None
was associated with peroxisomes (catalase marker) or microsomes
(NADPH-cytochrome c reductase marker). However, when a-
[1-14C] KIC was used as a substrate, two peaks of decarboxy-
lating activity were feund; one with mitochondria and the other
with the cytosol fractions.
Other studies showed that the mitochondrial activity requires
CoA and NAD+ (data not shown) and uses all three branched-chain
QPKEtO acids as substrates. This activity is due to the mito-
chondrial branched-chain o-keto acid dehydrogenase (E.C. 1.2.4.3

and 1.2.4.4).

37

38

 

 

 

Figure 1.

 
 
 
 

9..

 

 

comets/umn x no”
AME/m1! 2

 

 

 

 

 

 

Subcellular Distribution of Branched-Chain a-Keto Acid
Decarboxylases in Rat Liver. Marker enzymes for peroxisomes
(catalase), mitochondria (fumarase or glutamate dehydrogen-
ase), and microsomes (NADPH cytochrome c reductase) are
shown in the upper figures of A and B.

In Figure A, decarboxylase activity was measured using
conditions Optimized for the mitochondrial branched-chain
o-keto acid dehydrogenase (lower figure). Each assay con-
tained in a final volume of 0. 4 mL; 33 mM sodium0 ph05phate,
pH 7.2, 1 mM MgCl2, 0. 25 mM CaClz, 1 mM N3
mM CoASH, 1 mM NAD , either 0. 4 TM a-[1-10103ketoiso-
valerate (o-KIV) or 0.6 mM d-[1-4C] ketoisocaproate
(d-KIC) (100,000 cpm) and 0.1 mL of the fraction to be
assayed.

In Figure B, decarboxylation of d-[1-14C] ketoiso-
caproate was measured using assay conditions Optimized for
the cytosolic d-ketoisocaproate oxidase activity (see Meth-
ods, Method B). The assays also contained 5 mM sodiun
arsenite.

Abbreviations are Glu OH = glutamate dehydrogenase and
NADPH Cyt c Red = NADPH cytochrome c reductase.

B.

39

The cytosolic o-KIC decarboxylase activity is clearly differ-
ent from the mitochondrial activity, as will be shown. After
Optimizing assay conditions fOr this enzyme, the subcellular dis-
tribution was again determined (Figure 18). Sodium arsenite (5
mM), an inhibitor of the mitochondrial branched-chain d-keto acid
dehydrogenase was included in each assay. Again the d-KIC decar-

boxylase activity was detected only in the soluble fraction.

Properties Of Rat Liver Cytosolic o-Ketoisocaproate Decarboxylase
Activity

Cofactor requirements and other properties of the cytosolic
a-KIC decarboxylase activity were investigated after dialysis.
Table I shows that added MgZ+, Ca2+, Na2C03, phos-
phate, CoA, NAD+, NADP+, FAD, thiamine pyrophosphate, and
lipoic acid did not appreciably affect the activity. Although
phosphate was present in all Of the assays fOr Table 1, other
experiments (data not shown) using rat liver cytosol prepared in
Tris-maleate buffer, showed that phosphate is not required fOr
d-KIC decarboxylase activity.

The assay fOr the a-KIC decarboxylase activity was optimized
for concentration and pH of the buffer. Optimal activity was
Obtained in 0.2 M Tris, 0.2 M maleate buffer at pH 6.5. Activity
at pH 7.2 was approximately 80% of that at pH 6.5.

Connelly gt al. (6) reported that bovine liver o-ketoisoca-
proic:a-keto-B-methylvaleric acid dehydrogenase was activated by
ammonium sulfate. Hhen rat liver cytosol was assayed in 0.2 M

Tris, 0.2 M maleate, pH 6.5, 1.5 M ammonium sulfate stimulated

40

TABLE 1

COFACTOR REQUIREMENTS OF RAT LIVER
CYTOSOLIC o-KETOISOCAPROATE DECARBOXYLASE ACTIVITYa

 

a-KIC decarboxylase

 

Factor omitted or added activity (% of control)
None (control) (100)
- MgClz 102
- CaClz 95
- Na2C03 111
- NAD+ 88
- Coenzyme A 116
- Potassium phOSphate, + 33 M Tris 120
+ 0.5 nM Thiamine perphosphate 110
- NAD+ + 1 mM NADP+ 100
- NAD+ + 1 mM FAD 61
+ 0.5 mM DL-d-lipoic acid 86

0.5 mM Thiamine pyrophosphate,
+ 0.5 nM DL-a—lipoic acid 101

 

aThe assay conditions were the same as in Fig. 1A except fOr the
additions and deletions shown. Each assay contained 0.1 mL rat liver

cytosol

preparation. Control activity was 0.55 nmol/min/mg protein.

Figure 2.

41

 

8

 

 

nmole 14002/ min/ mg protein

 

 

0.47 1 L
(J 11) 213

Ammonium Sulfate (M)

The Effect of Ammonium Sulfate on Cytosolic o-Ketoisocapro-
ate Decarboxylase Activity. o-KIC decarboxylase activity
was assayed as in Methods (Method A) except that the ammon-
ium sulfate concentration was varied and the incubation time
was 30 min.

42

the a-KIC decarboxylase activity (Figure 2). This stimulation
was seen consistently, but the degree of stimulation was variable
in different rat liver cytosol preparations.

Figure 3 shows that with these assay conditions, the reaction
rate was linear fOr at least 90 min when 1.5 M ammonium sulfate
was present. In the absence Of amnonium sulfate, decarboxylase
activity was always lower and linear fOr about 60 min. o-KIC
decarboxylase activity was linear with protein concentration up
to 2.8 mg of rat liver cytosolic protein per assay (the largest
amount tested).

The apparent Km for the rat liver cytosolic d-KIC decar-
boxylase activity for o-KIC using these improved assay conditions
is 0.03 mM (Figure 4A). Hith the assay buffer used fOr the
sucrose gradient in Figure 1A, which was not at the Optimal pH
for the cytosolic a-KIC decarboxylase activity and did not
include amnonium sulfate, an apparent Km Of 0.2 nM was deter-
mined (Figure 48).

During the course Of these experiments, it was noted that
there is a slow increase in o-KIC decarboxylase activity with
time at 4°C. This is shown in Figure 4B where the two lines rep-
resent identical assays done on the same rat liver cytosol prepa-
ration but assayed 1 h apart. The apparent Km value was not
affected by this activation of the enzyme. In all experiments,
assays were grouped and started within 20-30 min Of each other.

The release Of 14002 from a-[1-14C] KIC was com-
pletely inhibited in the absence of oxygen (Table 2). This indi-

cated that the a-KIC decarboxylase activity uses oxygen as an

Figure 3.

43

Effect Of Time and Protein Concentration on Cytosolic a-Keto-
isocaproate Decarboxylase Activity. In the upper figure 100
uL of a cytosol preparation (2.8 mg protein) was assayed in
the presence (I) or absence (0) Of 1.5 M ammonium sulfate.

In the lower figure the amount of cytosol preparation was
varied.

44

 

 

 

 

 

 

 

 

 

 

1m

80
ca
2 60
.8
"' 4)
‘8
2
z 20

z) 40 60 80 100
MINUTES

SE

2

c?
‘0

a

2

:z

1 1 l l l
1.0 2.0

MG CYTOSOL PROTEIN

Figure 3

Figure 4.

45

 

Km =0.03mM

l

 

 

 

 

 

 

.20 -10 o 10 20 so
1/(«-KIC) (mu '1)

Effect Of Substrate Concentration on Cytosolic o-Ketoisocap-
roate Decarboxylase Activity. a-KIC decarboxylase activity
vs. the concentration of a-KIC was determined under two dif-
ferent assay conditions. Assays in A contained in a total
volume Of 0. 4 mL, 0. 2 M Tris-maleate pH 6. 5, 1. 5 M ammonium
sulfate, 1 mM NaZCO, 0.04-0.40 mM a-[1-14C] KIC

(630 cpm/nmol) and .7 mg/ml cytosolic protein. The incuba-
tion time was 15 or 30 min. Assays in B were done in a
total volume of 0. 4 mL containing 33 mM 1Rotassium phosphate,
pH 7.2, 1 mM Na CO3, 0.05-0.50 mM d-[1-1C] KIC

(1,100 cpm/nmoli,3 and 2.1 mg/ml cytosolic protein. Incuba-
tion time was 15 min. Two identical assays were done
approximately 1 h apart; (0) first set Of assays, (0) second
set of assays. Experimental points in both A and B were
plotted and the best- -fitting lines drawn by computer analy-
sis (HILKIN 2 PROGRAM, Michigan State University, East Lan-
sing, MI). Initial velocities (V0) are expressed as
nmol/min/mg cytosolic protein. Km and Vmax were

determined by the data weighting procedure suggested by
Hilkinson (144).

C.

46

electron acceptor. The second column of Table 2 shows that rat
liver cytosol preparations consume oxygen when d-KIC is added to
the assay. In several experiments, using either the rat liver
cytosol preparation or the partially purified o-KIC oxidase
(Sephadex G-150 pool, see Table 5), 0.7-0.9 molecules of 02

were consumed fOr each C02 molecule released from a-KIC.

Effect of a-Keto Acids on Cytosolic a-Ketoisocaproate Oxidase
Activity

Both assay methods gave similar results when the different
a-keto acids were tested as inhibitors of the a-KIC oxidase
(Table 2). 1 mM drKetOisovalerate, d-ketO-e-methylvalerate,
d-ketobutyrate, d-ketononanoate, and phenylpyruvate strongly
inhibited the decarboxylase and oxidase activity. Phenylpyru-
vate is an exceptionally strong inhibitor, showing complete inhi-
bition at 0.01 nM. a-Ketoisovalerate showed no inhibition at
0.01 mM. 1 mM d-ketoglutarate, pyruvate, o-ketO-Y-methiolbuty-
rate, and arsenite, an inhibitor Of dihydrolipoyl transacetylase,
had little effect at a concentration Of 1 mM.

The oxidation Of various a-keto acids by a rat liver cytosol
preparation is shown in Table 3. At a concentration Of 1 mM, the
two branched-chain d-keto acids, d-ketoisovalerate and arketO-B-
methylvalerate, as well as pyruvate, d-ketobutyrate, and a-keto-
nonanoate caused little 02 consumption. However, o-keto-Y-
methiolbutyrate, the keto analog of methionine, was oxidized by
rat liver cytosol preparations at a rate approximately 2 times

the rate Of a-KIC oxidation when both were assayed at a

.chaocq me\:Hs\uoe:m=ou me so comma—ac moueH .05: we ummmogaxm mmHHH>HHom Hocucoun

.;=_H we: OucocamoomHoumxiu we =o_uucucmu=oo aumm< .»u_>_uuo Hocucou as» we am :agu egos Lm>mc we:
moueuHquL camzama quHHAeHLm> .mxumme aHmOHwamc at» we come as» men mm:_u> .mcogumz Lave: umchummu mm
coHua52m=ou we go H< uogumzv omemHmc oueH An vmaemmu we: auH>HHuu mmevpxo mamocaeuomHoumxuom

 

47

 

. NoH 3.5.3.2 2: H
- H Hue m=e_zv
«H m oncogecocoumxno 25 H
aw om mueLAH=aouoxue :5 H
- om oueeae=a_o_gooe-»-eoog-o :5 H
o o oue>=caaHacmgaua zs Ho.o
o o mam>=caaHa=mgmua :5 H
mm moH muecoua—mouoxua :5 H
as m“ oum>=cxm as H
aH mm mumcmHo>HA=um51nuoumxua :5 H
mm mm muucmHm>omHouoxao :5 Ho.o
a mH wuacwHe>omHouoxua :5 H
aHme\e_e\_aEe em.ov eHms\e_e\HOEe o.HV
ooH ooH HHocucouv ocoz
HHocucou we av AHocacou mo my HcoHHonu gov copawcn<
noszmcou «o wanna—me oueH

 

aqu>Huoe «mauon oumocaauomHoumxuo

 

monh<¢<mmmm Hom0h>u ¢m>HH h<¢ zH
>hH>Hhu< mm<con mh<omm<oochhmxua zo zmu>xo az< .uthwm¢< .moHu< chmxna mo mhumuum

N u4m<h

48

TABLE 3

OXYGEN CONSUMPTION BY RAT LIVER CYTOSOL PREPARATIONS
USING VARIOUS a-KETO ACIDS AS SUBSTRATES

 

nmol 02 consumed/min/mg

 

Substratea cytosol proteinb
a-Ketoisocaproate 0.87
a-KetO-Y-methiolbutyrate 1.87
d-Ketoisocaproate

+ o-ketO-Y-methiolbutyrate 0.97
a-Ketoisovalerate 0.03
d-KetO-e-methylvalerate 0.04
Pyruvate 0.02
a-Ketoglutarate 0.02
B-Phenylpyruvate 0.02
o-Ketobutyrate 0.02
a-Ketononanoate 0.07

 

aAll additions gave a final assay concentration of 1 mM.

bAssays were carried out as described under Methods (Method A).
Values are the mean Of two replicate assays. Deviation between assays
was always less than 0.10 nmol 02 consumed/min/mg protein.

D.

49

concentration of 1 mM. d-[1-14C] keto-Y-methiolbutyrate is
decarboxylated by rat liver cytosol preparations at a rate about
2 times that of a-KIC (unpublished Observation). The addition Of
both 1 mM d-KIC and 1 mM a-ketO-Y-methiolbutyrate to cytosol
preparations produced a rate Of Oz consumption intermediate

between the addition Of either substrate alone.

Tissue Distribution Of o-Ketoisocaproate Oxidase in the Male Rat
Table 4 shows the distribution of the o-KIC oxidase in 27,000
x g supernatant fractions Of several rat tissues. The highest
specific activity occurs in liver. Activity was also fOund in
the kidney. Activity was not detected in brain, heart, muscle or
pancreas. The absence Of o-[1-14C] ketoisovalerate decar-
boxylating activity in all tissues indicates that the mitochon-
drial branched-chain a-keto acid dehydrogenase is not active with

the assay conditions used.

II. Purification and Stabilization of the o-Ketoisocaproate Oxidase

Activity from Rat Liver

A.

Stabilization Of and Cofactor Requirements of d-Ketoisocaproate
Oxidase Activity

In order to further characterize physical and kinetic proper-
ties Of the cytosolic a-KIC oxidase activity as well as identify
the products of the reaction, purification of this enzyme was
necessary.

a-KIC oxidase was partially purified from a 70,000 x 9 super-

natant fraction of a rat liver homogenate as shown in Table 5.

50

.333 mung—qua “Ho com... 95 $32, .Dmluomqmmc .2: H EB z: N .2: N mew:

uHOu OHnLoume use .Hoa_mcguoH;u_v .eommd we mcopumcucwucou Amman unauxw .Hm vegumzv mcogum: :H mm toxemme

mew: muH=He aceueccoazm m x coo.NN use .xu_>Huum wmmuon mumocaeuomHoquua wsu mw~HHHneum HoceqocaomH

.HoceaoLQOmH «H OH woumznue we: acmaeccoaam esp .vmuguumHu aoHHma as» ecu cHE me so» a x coc.NN an venue

-_cu:mu we; muucmmosos ugh .uoe as mew: mucoEHoocu HH< .coucmHn mcpcoz a :H cam mH cow uoNHcmmoec; we:

mHOmzz .Lm~_:mmoso: mmm—m sangm>Hu Leave; a :_ mommeq N ha cage—How we: m_:h .Lm~_:wmosoz :ocucha a gap:
.Omocuzm z mN. mo mosaHo> m o» N =_ muccuwm on Low mOEHH m um~wcmmoeos we: mHumzs uqmuxm mammHH :ucu

 

 

mo.ov No.ov mcmcucaa
mo.ov No.ov w—omsz
eo.oV No.cv acme:
vo.oV No.ov :Hocm
mc.ov NH.o HucuHx
No.o mm.o cw>HH
ouncOHu>cmHoumx nuvHuHuuo mucocamOOmHouwx mueHuHuuo mama.»

 

chmuoca me\:_s\umo:vcca NoucH «Hosev
AHH>HHO< mcpueHzxoacuomo uHu< ouoxuo

 

h<m mH<z mzh zH Amm<aonv mm<4>xom¢<oua mh<o¢¢<uomH0hmea mo onhamthmHo mammHh

e m4m<h

.czosm we m=OHHOHmu Lo :oHHHuua Lew ugmuxm H< vozuuzv muogumz gave: me umaommo we: Omeupxo uquoo

.25 m.c we: uqua we copumcucmocou Hc=HH mzu Hamuxm H<
vogumzv mcogam: gave: me :OHHu—xxoncauov uHx mueHiHuuo m:_camems x: umxemmo ma: Omouon uquee

 

51

a~.mH eHN aem.e aceu on e.mH oeeacoome :5 e.o + .eomoa
:5 ~.o + .Hoe_oceeo_eo_e :5 H +
.eommHezzv Ssoeoez easema<

 

ee.m ac ace.H new om e.mH eOWNHezzv S=o=o_z eoxeaw<
m.~ m ~e.o oe om e.mH Hoea cmH-u xoeeeaom .H
e.m 4N mm.H on cow m.m Hooa m<ma eooaeeeooeou .e
o.m Hm Ho.~ ace cow mmN Hooa omoH=HHoo-u<ma .m
e.H on oe.o ase oomH oak ooemHHeHa m<wo-oca .e
~.H om a~.o omoH comm omm oeeeaecoasa «omNHezzH awn .m
m.~ mHH ~e.o oeNN comm on Seeoeecoasa oSaCH=m oeHEaeoee u~.o .~
o.H ooH m~.o ome ooom wee Seaoeecoesa a x ooo.o~ .H
mesa xgm>ouem aAmE\=_s\HoE:v «Ac—stoe:v Hmav HHEV copuuucm

vHou HHH>HHO~ HHH>HHO~ =Hmuoca OzaHo>

oHCHoeam .eoo» Heed»

 

mu>HH h<m team mm<ono uh<om¢<uomHohmxia no onh<uHuHm=a H<Hb¢<m
m uHm<h

52

The total activity of d-KIC oxidase is doubled by protamine sul-
fate precipitation. Hhen 35% ammonium sulfate was added to the
0.2% protamine sulfate supernatant, no additional protein was
precipitated and there was a large loss Of activity.

Column chromatography using DEAE cellulose or Sephadex caused
a large loss of activity. This could be due to the removal of
low molecular weight materials necessary fOr enzyme activity.
Therefore, various compounds were added to the assay mixture in
an effort to restore activity; see Table 6. The activity Of the
d-KIC oxidase was increased severalfold by the addition Of fer-
rous iron, an electron donor (ascorbate, NADH, or NADPH) and a
sulfhydryl reducing compound (CoASH or dithiothreitol). All
three factors are required fOr optimum activity. At higher con-
centrations Of Fe2+, ascorbate is not required (see Figure
18). a-Ketoglutarate, a cofactor for some dioxygenases (125),
was not required. Catalase, which increases the activity Of some
oxygenases, also had no effect on the GPKIC decarboxylase activ-
ity. However, some catalase is present in the enzyme prepara-
tion.

Optimal concentrations Of FeSO4, ascorbate and dithiothrei-
tol were found to be 1.0 mM, 0.5 mM and 1.0 mM, respectively when
using either the "DEAE-cellulose pool" or "Sephacryl S-200 pool"
as a source Of a-KIC oxidase. Despite the inclusion Of these
compounds in all assays, large losses of activity occurred during
purification when DEAE-cellulose, phenyl sepharose or sephacryl
columns were used. Hhen 0.4 mL of the concentrated DEAE pool

(see Table 5) was applied to a 1.3 x 56 cm Sephacryl S-200

53

TABLE 6

EFFECTS OF POSSIBLE COFACTORS ON arKETOISOCAPROATE OXIDASE ACTIVITY
IN THE PARTIALLY PURIFIED PREPARATION (SEPHADEX G-ISO POOL)

 

o-KIC oxidase activity

 

 

nmol 14C02 Percentage
Additionsa min/mg/protein of control
Expt 1
None 1.24 100
CoASH, FeZI, ascorbate, o-KG 3.45 278
CoASH, Fez , a-ke 1.35 108
CoASH, ascorbate, a-KG 1.80 145
, ascgrbate, d-KG 1.11 89
CoASH, Fe , ascorbate, a-KG,

DTT 3.56 286
Fe2+, ascorbate, o-KG, DTT 4.00 322
GPKG 1.13 91

Expt 2
None 1.09 100
011, Fe2+, ascorbate, a—KG 4.74 433
011, Fe2+, ascorbate 4.36 399
DTT, Fe2+, ascorbate, d-KG,

catalase 4.33 397
DTT 0.91 83
011, Fe2+, NADH, a-KG 2.77 254
011, Fe2+, NADPH, a-ke 2.86 262

 

aConcentration Of additions were as fellows; CoASH, 1 mM; FeSO4
(Fe2+ ), 0. 2 mM; ascorbate, 0. 4 mM; a-ketoglutarate (o-KG), 0. 5 mM;
dithiothreitol (DTT),1m; catalase, 100 U/assay; NADH, O. 5 mM; NADPH,
O. 5 mM.

b45537 0f c-[1-14C] KIC decarboxylase was carried out as under
Materials and Methods (Method A) except the amonium sulfate was omitted
and the concentration of o-ketoisocaproate was 0.5 mM. Each assay
included 25 or 50 uL of the “Sephadex G-150 Pool“ (Purification A, Table
5). Values are mean Of duplicate assays.

54

column, less than 10% Of the applied activity was recovered.
Elution with high concentrations Of salt removed more protein
from the column but did not increase the yield Of a-KIC oxidase.
Pretreatment of the Sephacryl S-200 column with protein contain-
ing fractions recovered from the DEAE-cellulose column (step 5,
Table 5), which did not contain o-KIC oxidase activity ("DEAE
side fractions“) increased recovery to 82%. Kaufman and Fisher
(126) used a similar procedure in the purification of phenylala-
nine hydroxylase from rat liver.

Results Of a second attempt to purify rat liver d-KIC oxidase
are shown in Table 7. FeSO4, ascorbate, and dithiothreitol
were included in all assays (Method B). Phenyl sepharose and
Sephacryl S-200 columns were pretreated with “DEAE side frac-
tions“.

Ammonium sulfate fractionation resulted in a 2 fold increase
in specific activity with a 85% recovery. DEAE cellulose columns
did not require pretreatment with protein, due to the large
amount of protein applied to the column, and always gave a 5 to 6
fold increase in specific activity. Ethanol fractionation gave a
6 fold increase in specific activity, however, yields were quite
variable from one purification to another. Therefore, this step
was omitted from later purifications.

Despite pretreatment Of phenyl sepharose and sephacryl col-
umns with "DEAE side fractions”, large losses of a-KIC oxidase
activity were still apparent. Note that 60% of the d-KIC oxidase
activity was lost while concentrating the phenyl sepharose pool

to 7.2 mL (Pre-Sephacryl S-ZOO); very little activity was lost on

55

.HOON . coev Hoom.oe . HSH>HS6< .aooev 4 H2283. a

coma we; sHooa m<uo vmuecucmucous On» No HHm wH ng>oumc umaomaxm zozm on umuomccoo mH xgm>ouwmn

.Hm uosumzv mcosuwz :_ me vocsmams we: AHH>HHOa Omeuon muuoeaeuomHoumxuoa

".O.H .coHHauHeHcsa one zmzoczu uchceo

 

mmN.H

Humpoca Ho: mew: mcoHuocewv

 

NmH-ee no eH-~m me - ass—co memo oew .oH
mHH aoH o.mm oeH.m mm He .ooa co~-m choaeaom .a
He aNH H.Hm om~.m ONH ~.H eoeeeeeooeeo Hooa .Heeea .w
~.Hm ecu m.ee oco.m omH oom.H .ooa omoceeaoa Hseoee .H
m.He awe e.Hm oom.eH wee mm eoHSoece Haeaeuu nooH-m~ .e
.cowumu: 7:3 93 $0 Log—35mg mg» coach.» umtgou ma: spoon m<uo umumgucmucou: as» we .2. 8m 3.5
H.oH on mH.m oom.om ooo.~ ace .ooe u<mg uooeeeeooeoa .m
e.oH em Ne.m oom.mm oe~.e oeH.H Hooa oao_=__oo-u<ua .e
m.~ Hm HH.H oom.~m coo.me cm~.H oeemxmewe m<ma-oca .m
m.H mm ea.o omm.Hm oom.~m own.H eoHSoecc eom H :zv umH-me .N
Ho.HH HooHv Hm.o oom.oe ocm.~HH ooc.m oeeoeeeoaaa a x ooo.ooH .H
«can xcm>ouox «Hms\:Hs\o_oe=v aH=HE\uHos:v Hmsv HHEV
eHoL a »e_>_oo< NSH>Hoo< eHoSoca oe=_e>
oHCHooam _aooe Haeoe

 

o onh<uHmH¢=m "mm>HH h<¢ zozm mm<o~xo mp<cxm<uomHohmxua mo onh<uHuH¢=¢

N m4m<~

56

the sephacryl column itself. The a-KIC oxidase activity is fair-
ly stable at 4°C in less pure preparations such as the "concen-
trated DEAE pool" (9% loss of activity in 4 days, Figure 5).
Purified preparations Of a-KIC oxidase, however, rapidly lose
activity (Figure 6; 50% loss Of activity in 4 days). This loss
of activity does not appear to be related to the protein concen-
tration, since dilution of the "concentrated DEAE pool" to 1.68
mg protein/mL did not substantially alter stability (Figure 5).
The o-KIC oxidase activity was quite stable at -80°C fOr up to 20
days in all purification fractions (Figure 5 and 6), however, at
-20°C or room temperature activity was rapidly lost. d-KIC oxi-
dase activity was more stable at 4°C than at -20°C or room tem-
perature.

The addition of 5% monothioglycerol (0.6 M) stabilized d-KIC
oxidase activity at 4°C for at least 6 days (Table 8). One per-
cent monothioglycerol (0.12 M) was not effective in preventing
loss Of activity. Dithiothreitol at 1 nM or 5 mM and in the
presence or absence of 5% glycerol also did not prevent loss Of
activity. Although monothioglycerol prevents the loss of a-KIC
oxidase activity, this compound caused a decrease in the measured
initial activity of the d-KIC oxidase (Table 8). Figure 7A shows
that d-KIC oxidase activity is inhibited by assay concentrations
of monothioglycerol greater than 0.6%. Increasing the amount Of
FeSO4 in the assay mixture did not prevent this inhibition
(Figure 7B). The inhibitory effect of monothioglycerol is most

likely due to a depletion Of 02 in the assay mixture.

Figure 5.

57

 

168 mg proiein/mL A
..--.g. ------- °°80°C

  
   

 

 

 

 

 

 

>.
E m 0.4 4 4°C
2 E 0.3“ “‘1 20°C
1': O 2 . 7‘"
8 s
:8 o 0.1? Room
X 3 Temp.
CD 53 . E3
.93 3; 0.5 _ l68mg proiem/mL
E o ””.D- ----- “-O-BOOC
g- 0.4 "‘7
gin 03 o
53 15 ' 41C:
'§_ § 0.2
1‘? 0.1 “~.-20°c
U . .
0 IO 20

Stability of the a-Ketoisocaproate Oxidase in a 12-Fold Pur-
ified Preparation. Aliquots Of the “concentrated DEAE Pool"
(Purification B) (B) or a 1:10 dilution Of the “concentrated
DEAE Pool" (A) were stored at the temperatures indicated.
a-KIC oxidase activity was monitored as in Methods (Method
B). Frozen samples were only thawed once, then discarded.
Values are the mean Of duplicate assays.

58

 

(‘5
o

01
O .

 

Activity Remaining

 

 

 

%

120°C

 

 

a-Keioisocoprooie Oxidase

IO 20 3O 40
Days

Figure 6. Stability Of the a-Ketoisocaproate Oxidase in an 80-Fold

Purified Preparation. Aliquots Of pooled fractions from the

‘“2nd DEAE column" Of Purification C (Table 7) were stored at
the temperatures indicated. o-KIC oxidase activity was
measured as in Methods (Method B). Frozen samples were
thawed only once, then discarded. 100% is 42 nmole/min/mg
protein for -80°C experiments and 31.7 nmole/min/mg protein
for 4°C and -20°C experiments. Values are mean of duplicate
assays.

59

.eHesem we HeHeaeLe a: mNV H: mN ee=Heeeee

Hemme seem .maemme eueeHHeeL N we gees ego mu_:mem .HaaH>Hue< He:_uv uoe we emeceum whee w Le a ceywe

ece HHHH>HHO< HeHHHeHV xHeueHeeEEH Hm eezuezv meeguez :_ we emcemees we: HHH>HHee emeeer eeeeceeeemHeuexie

.HHE NH.oV eseHe> HeeHe eEem es» e» ecumenee wee: meeeeHHe HH< .ezesm mceHHHeee mg» :HHz uoe He eeeeeee:_
ace: HH oHaaH oomv e eoHHeoHLHL=a co .ee=_eo “(as new. oeH sate aeoHHoece eoHooa Lo HHe mH.ov mooseHH<

 

 

an e me.e H.H~ HooHoceHoHeH_e 2E m + HocooHHe am
mm m e.HH H.H~ HoHHoceHoHeHHQ :5 H + HoeooaHe um
HoH e H.H~ a.e~ HoeooHHHoHeHoeez am
aN e mH.H m.e~ HocooHHmoHeHoeoz HH
NM e ma.H e.m~ HoHHoceeoHeHHo :5 m
Hm e ee.a m.m~ HoHHoeeeoHeHHQ :5 H
ee e e.HH e.e~ oeoz

eeHeHeeoe eae He omecoem HeeHd He HHeH eoHH_ea<

»u_>Hue< a ee mama :Heuece ma\=_s\eeeeeece cu H e—e2=

auH>Hue< emeeon eueeceeee Heuexue

 

Homuu>Hw «o .HOhHmmzththn .Hommu>4ocH:»ozoz
mo muzmmume uzh zH mm<ono mh<oae<oomHohuxia mo >hHHHm<hm

w m4m<h

Figure 7.

6O

 

Y
53
m

1

p
.p
j

 

 

 

4 .8 L2— 16 2:0 24
[Monothioglycerol] "/6

 

1I2 L I I
’8.-HH-9_-_§|9..aidl.telon
0.8 e3

O-Ketoisocaproate Oxidase Activit
umole 14002 produced/hrlmg

+Monothioglycerol
A

 

 

 

 

IO 20
[F135K14] IHWNA

Effect Of Monothioglycerol on drKetOisocaproate Oxidase.
e-KIC oxidase was assayed as in Methods (Method B) with the
following modifications. In A, monothioglycerol was added
as indicated. In B, the final assay concentration Of

FeSO? was varied from 1 to 20 mM and assays were carried
out n the presence or absence Of 1.25% monothioglycerol as
indicated. Each assay contained 50 uL (31 ug protein) of
pooled fractions from the "2nd DEAE column" of Purification

C (Table 7).

B.

61

Purification of a-Ketoisocaproate Oxidase Activity

Table 9 shows the results of an d-KIC oxidase purification
using conditions which stabilize this enzyme. The initial steps
in the purification (steps 1-5) did not include monothioglycerol.
This may be the basis fOr the low recovery (50%) in the “concen-
trated DEAE pool”. Phenyl sepharose and sephacryl S-200 columns
were pre-treated with "DEAE side fractions“ and 5% monothiogly-
cerol was included in all buffers used.

The specific activity of GPKIC oxidase was increased 2.2 fold
by 45-75% (NH4)zSO4 fractionation with only a 25% loss of
total activity. This fraction was then dialysed and applied to a
DEAE column (Figure 8). A 15% loss Of activity occurred during
dialysis. Loss of activity during the DEAE column chromatography
was due to pooling of peak fractions Off this column.

Hhen applied to phenyl sepharose in a buffer containing 2.5 M
NaCl, o-KIC oxidase activity was retarded and elutes just behind
the column void volume (Figure 9). A large amount Of protein
remained bound to the column. The purification at this step may
be improved by increasing the concentration of NaCl in the sample
and elution buffer to promote a stronger hydrophobic interaction
between the aPKIC oxidase and phenyl sepharose.

Sephacryl S-200 chromatography (Figure 10) increased the spe-
cific activity of the o-KIC oxidase almost 5 fold. Three peaks
of protein were fOund. The drKIC oxidase activity migrated wfith
the first Of these protein peaks. The overall yield Of o-KIC
oxidase activity was 27% with a final specific activity of 104

nmole/min/mg protein.

62

.HeHH . omNHHoee.em . HHH>HHo< Heeoev . o>ooom a H.o._ .eoHHaoHHHesa oeH emaoeee eoHccao

a;
seen em: =Heee m<uo eeuecueeeeeee es» ee HHe NH age>eeec eeueeexe seam ea eeueeccee mp xce>eeeme

.m eeguez Ne emcemeee qu>HHee eee Hm
eeHHeeHcheev meegue: :H eenpcemee we ce>HH we; see; eeHeHcee we: HHH>HHee emeeer eueeceeeemHeuexie

 

eeH eHN eeH ome.H HH ee Hooe eo~-m HHgoaeaom .a
m.oe am” e.- eoH.eH eHe ON eooaceeooeoo Hooa HHeoea .e
H.4m eem H.mH emH.a «Hm who Hooa owoeaeaoa HHeoea .H
m.~H eme He.e oee.~H eem.H ouH omoceeaem HHeoea-oee .e

ceHHeOH$HL=e me» we LeeuweEec ezu :meecgu eeHLcee me: =Heee N<uo eeaecueeeeeee use me HE eHH aHeo

 

«.mH em mm.e eo~.m~ oe~.e omw Hooa H<ea eoHeEHeooeoe .m
e.mH Nm ~e.~ oom.em eoo.e oHH.H .663 omoHsHHoo-e<Hg .e
o.m mm He.H oee.em ooo.o~ see.H oeemamaHe e<mo-oea .m
~.~ «H e~.H eoH.me eoH.em eo~.H eeHHoece com Hezzv amH-me .N
Ho.HH HeoHH em.e ooe.em oe~.eoH oo~.H Heeueecoasm m x ooo.eH .H
ecee Nge>eeem Hms\=_s\eHee:v H=_s\eHee=v Hmsv HHEV eeHHOeLm

eHom a HHH>HHo< aHHsHHo< eHoHoca oesHo>

oHeHooam HeHoH HeHoH

 

mm>HH h<¢ tome mm<ono mh<ome<uomHOFNxte mo onh<uHuH¢=e

a uHm<h

3

6

.Hm aoeHezv meoeeez =_ we eoaemma age He eoHH
neewwwcee .meezuez eemv Heeweecm Hue: z N.ouo e new: eeueHe we: auw>wuee emeewxe uere .cseHee
ewe—e—Hee u<ma e e» eewHeee ecu eeNAHewe ece: :ewueeewaeecw euewHem seweesse amNume He eeeweune

emeewxe onue we meewueceeece eeeeu .emeewxo eueeceeeemweuexue we HeeecmeueEeczo emeHeHHeuum<ma

.w ecemwe

64

x—K (VV) ICXIV

 

 

 

04 _.

.9 4 C?
0—0 (000092) a 00010qu 940098
00 to - N
9 9 . 9?

 

 

.H 1w/U1wmaanpwd Z0:1,. elowu

,(nmpv asepgxo OIDOJdDOOSIOIGM-D

 

200

IOO
FRACTION MMBER

Figure 8

.meegue: ew me
vexemme eLe: eweuece use Hm vegeezv emeewxe uHxie .eeeewueee :ewueHe eee eeuceem we: HeceeaHm
-65er um .Hoeaoao: H: .3 1.. He: at: z... 8 E 8...: z e.e-m.~ we .2383 28:: as: N
a E 545.. 83an 2 .882 2 we Hosea—8:88... am .Hoeeaoaofl H: .3 .a a: 6...: He 8
saw: eeeeHe me: ase—cu ezw .Hm eewueewwwcee .meeguez eemv cseHee emecegeem ereze e e» eewpeee
ece .Heceemeewgaeeee um .Huez z m.N ea e: eees .eeuecueeueee ece: Hm eceawev es=Hee emeHeH

-Heenu<ue ezu wwe meewueeww xeee .emeewxo eeeeceeeemweuexue we ageecmeeesecgu emeceseemuHaeeze .m eeemwe

66

o—o [tn/Bu: ugade

80 I00 izo

66'

FRACTION NUMBER

40

 

 

 

20

 

1 #L J—
O O 0
N3

N _
o—o wugm/paanpmd 200,, spam
MgAgpv asopgxo emonooosgmex-p

Figure 9

Figure 10.

a-Ketoisoccprooie Oxidase Activity
nmole "cog produced/min/mL H
g

67

 

8
9
4:

  
  

 

 

 

m :20

so" so
FRACHON NUMBER

Sephacryl S-200 Chromatography of a-Ketoisocaproate Oxidase.
Peak fractions from the phenyl sepharose column (Figure 9)
were concentrated, applied to a sephacryi S-200 column and
eluted as described in Methods. a-Ketoisocaproate oxidase
was measured as in Methods (Method B) and protein monitored
by absorbance at 280 nm.

68

III. Physical Properties of Rat Liver a—Ketoisocaproate Oxidase

IV.

505 gel electrophoresis of the "Sephacryl S-200 pool"
(Purification B, Table 9) revealed one major protein band with
several minor protein bands (Figure 11). When electrophoresis was
carried out under non-denaturing conditions, QPKIC oxidase activity
migrated with the major protein band (Figure 12) indicating that this
protein is the a—KIC oxidase.

The subunit molecular weight of a—KIC oxidase determined by $05
gel electrophoresis (Figure 11) was 46,000 (Figure 13). Molecular
weight of the GPKIC oxidase was also determined under non-denaturing
conditions, that is, Sephacryl S-200 chromatography, according to the
method of Andrews (127). A molecular weight of 51,000 was obtained
by this method (Figure 14). This enzyme appears to be a monomer.

The 10% difference in molecular weights determined under denaturing
and non-denaturing conditions may be due to assymetry of this

protein.

Kinetic and Catalytic Properties of the Purified arKetoisocaproate

Oxidase

A. Stability of the Purified o-Ketoisocaproate Oxidase

The purified a—KIC oxidase (Sephacryl S-200 pool, Table 9)

was quite stable at 4°C in the presence of 5% monothioglycerol.
Only 15% of the initial activity was lost over 3 weeks (Figure
15). By 70 days, however, all of the arKIC oxidase activity was
lost. This may be due to the gradual autooxidation of the mono-

thioglycerol which stabilizes the enzyme.

69

,
”New - --

 

Figure 11. SDS-Polyacrylamide Gel Electrophoresis of a-Ketoisocaproate
Oxidase. Gel electrophoresis of 60 pg (A), 24 pg (8) or 12
pg (C) of the purified a—KIC oxidase (“Sephacryl S-200 Pool,
Table 9) was carried out as in Methods. Molecular weight
standards (column 0) are glycogen phosphorylase a (94,000),
bovine serum albumin (66,000), ovalbumin (45,000),
trypsinogen (24,000) and B-lactoglobulin (18,400).

Figure 12.

70

 

@—

 

 

 

Reloiive Absorbmce 01 600 nm

 

wk 3

 

 

(I)
I

 

 

 

mole/W251
63

0.2 of do 0T8 LO
Reloiive Mobility

Native Gel Electrophoresis of a-Ketoisocaproate Oxidase. A
30 uL (36 pg protein) aliquot of the purified a—KIC oxidase
(Sephacryl S-200 pool, Table 9) was applied to two separate
gels and electrophoresis carried out as in Methods. Protein
(upper figure) and a-KIC oxidase activity (lower figure)
were monitored as in Methods.

Figure 13.

71

 

(D
9

a

Mr, dolions x I0'3

95

 

 

 

 

 

'05—'62 oi4 o"—‘*.6 ofe |.O
. RM

Determination of the Subunit Molecular Weight of a-Ketoiso-
caproate Oxidase. The data were obtained from the SDS-poly-
acrylamide gel shown in Figure 11. Relative mobility (Rm)
was determined by measurement of the migration of the pro-
tein relative to migration of bromophenol blue. Abbrevia-
tions and molecular weights are PHOS, glycogen phosphorylase
a (94,000); BSA, bovine serum albumin (66,000); OVA, oval-
bumin (45,000); TRYP, trypsinogen (24,000); LACT, e-lacto-
globulin (18,400); LYS, lysozyme (14,300).

72

 

 

 

 

 

8
BSA
6' a KIC 0 'd (5| 000)
" XI 086 ,
/
OVA
4-
MW
23..
Cyi c
70 8'0 910 IOO

Eluiion Volume (ml)

Figure 14. Molecular Weight of the a-Ketoisocaproate Oxidase Determined
by Sephacryl S-200 Chromatography. The molecular weight of
the purified a-KIC oxidase (Sephacryl S-200 pool, Table 9)
was determined by Sephacryl S-200 chromatography as
described in Methods.

Y

Figure 15.

E?

(I)
C)
l

n/mg protein

m
c?

4s
‘9

CI-Keioisocoprooie Oxidase Aciivii
"‘002 produced/m
8

73

 

 

 

 

 

20 4'0 60
DAYS

Stability of the Purified a—KIC Oxidase. Aliquots of the
purified a-KIC oxidase (Sephacryl S-200 pool, Table 9) were
stored at 4°C or -80°C. a-KIC oxidase activity was measured
at various time intervals as described in Methods (Method
B). Values are mean of duplicate assays.

B.

74

At -80°C, there is initially a 15% loss of a—KIC oxidase
activity, possibly due to freezing and thawing, but then the
activity is stable fer at least 70 days. Routinely, small ali-
quots of the purified QPKIC oxidase (Sephacryl S-200 pool) were

stored separately at ~80°C and thawed immediately prior to use.

Metal Requirement of the u-Ketoisocaproate Oxidase

The effect of a wide variety of metal ions on the a—KIC oxi-
dase activity was tested (Table 10). In the presence of ascor-
bate and dithiothreitol, ferrous iron (Fe504 and FeClz) and
ferric iron (FeCl3) activated the a—KIC oxidase activity. In
the absence of added iron, but with ascorbate and dithiothreitol
included in the assay, activity was 25% of that in the presence
of iron. If ascorbate and dithiothreitol as well as iron are
omitted, very little activity is observed. GPKIC oxidase activ-
ity, observed in the presence of ascorbate and dithiothreitol
when iron is omitted from the assay, may be due to contaminating
iron in the buffers used. o-Phenanthroline, an iron chelator,
abolishes most of the GPKIC oxidase activity obtained in the
presence of ascorbate and dithiothreitol. A high non-enzymatic
decarboxylation of arKIC (Blank Activity) is observed in the
presence of 1 nM o-phenanthroline. This high rate of non-enzy-
matic decarboxylation occurred only when both o-phenanthroline
and ascorbate were included in the assay mixture, but the mechan-
ism is unknown.

CsClz, CaClz, MgClz, MnClz, ZnClz, CdClz and
NiClz had no effect on o—KIC oxidase activity at a

75

TABLE 10

EFFECT OF VARIOUS METAL IONS 0N a-KETOISOCAPROATE OXIDASE ACTIVITY

 

 

a-KIC Oxidase Blank % of
Activity Activity Control
Addition (or deletion) nmole/min/mg nmole/min/mg Activity
None 23.5 1.7 25
None (-ascorbate, -DTT) 1.3 2.2 1
1 nM o-phenanthroline 5.1 41.1 5
1 mM Fe504 93.1 0.6 (100)
1 mM FeClz 99.3 0.6 106
1 mM FeCl3 92.7 0.6 99
1 mM CoClz 19.8 1.8 21
1 nM HgClz 15.1 0.8 16
1 mM CaClz 24.8 1.5 27
1 mM CuClz 9.4 1.5 10
1 nM CuCl 9.7 1.5 10
1 mM CuSO4 9.7 1.4 10
1 mM MgCl2 24.8 2.0 27
1 mM MnClz 25.1 1.7 27
1 mM ZnClz 27.0 1.7 29
1 mM CdClz 26.2 1.1 28
1 mM NiClz 20.0 2.2 21

 

a—Ketoisocaproate oxidase activity was assayed as in Methods (Method
B) except FeSO4 was replaced by the addition shown. All assays con-
tained 1 nM dithiothreitol (DTT) and 0.5 nM ascorbate except where indi-
cated otherwise. Each assay contained 10 uL (12 pg protein) of Sephacryl
S-200 pool (Purification B). Values are mean of 2 replicate assays.
Blank activity was measured in duplicate assays in which the Sephacryl
S-200 pool was replaced by 10 uL 20 mM Tris HCl pH 7.8, 1% isopropanol,
0.1 M NaCl, 5% monothioglycerol. Activity is expressed in nmole/min/mg
protein.

76

concentration of 1 mM. HgClz and cuprous or cupric chloride
were inhibitory. The aPKIC oxidase, therefore, appears quite
specific in its requirement fOr iron.

The ability of ferric iron (Fe3+) to replace ferrous iron
(Fe2+) in the activation of the a-KIC oxidase could be due to
reduction of Fe3+ to Fe2+ by ascorbate. To test this
possibility, the effect of ascorbate and dithiothreitol on QPKIC
oxidase activity in the presence of Fe504, FeCl3 or no added
iron was determined (Figure 16).

In the presence of ascorbate and dithiothreitol, Fe504 and
FeCl3 produce approximately the same activation of the a-KIC
oxidase. In the absence of added iron the activity is 72% lower,
as seen previously (Table 10). The omission of dithiothreitol
from the assay has no significant affect on these results. If
ascorbate is omitted from the assay, activation of the enzyme by
Fe504 decreases slightly (17%), whereas activation by FeCl3
decreases 50%. In the absence of iron very little activity is
found. The omission of both ascorbate and dithiothreitol from
the assay produce the same results as the omission of ascorbate'
alone.

Previous studies (see Table 6) indicated that the a—KIC oxi-
dase required a sulfhydryl reducing reagent, such as dithiothrei-
tol, for activation of the QPKIC oxidase activity. No effect of
dithiothreitol on the enzyme activity is apparent in Figure 16.
The enzyme sample being assayed in Figure 16 contained 0.6 M mon—
othioglycerol. When 10 uL of this sample was assayed, a final

concentration of 15 mM monothioglycerol was introduced into the

Figure 16.

a-Ketoisocaproate Oxidase Activity
umole1 002 produced/hrlmg

77

a 1 mM Peso,

lll'linNAIFeCHae
D No Fe added

(3)

is

‘.
ti
'(2
p.
‘
l
.'
'3.
4'»
I

l\)

 

Ascorbate DTT Ascorbate N0
+ DTT Addition

Effect of Dithiothreitol and Ascorbate on Activation of the
d-Ketoisocaproate Oxidase by Fe504 or FeCl3. d-KIC oxi- .
dase activity was measured as in Methods (Method B) except
Fe504, ascorbate and dithiothreitol were only added as
indicated. Final assay concentrations were: FeS04, 1 mM;
FeClz, 1 mM; ascorbic acid, 0.5 mM; dithiothreitol (DTT),

1 mM. Each assay contained 10 uL (12 pg protein) of the
d-KIC oxidase (Sephacryl S-200 pool; Table 9). The mono-
thioglycerol in the Sephacryl S-200 pool gave a final assay
concentration of 0.125% (15 mM) monothioglycerol. Values
are mean of duplicate assays.

C.

78

assay mixture. Therefbre, monothioglycerol most likely provided
the reduced sulfhydryl groups.

These results suggest that the reduced f0rm of iron,

Fe2+, is utilized by the enzyme. In the presence of ascor-

bate, ferric iron (Fe3*) may be reduced to supply the enzyme
with FeZT. In the absence of ascorbate and dithiothreitol

ferric iron still activates the d-KIC oxidase to some extent.

The reason fer this is presently unclear. Monothioglycerol,
which is present in all assays at a concentration of 15 mM due to
carry over into the assay with the enzyme, may reduce Fe3+ to
Fe2+.

Fe504 gave optimal activation of the QPKIC oxidase between
1.0 MM and 5.0 mM (Figure 17). Concentrations above 5 mM became
progressively inhibitory.

The effect of ascorbate on GPKIC oxidase activity was tested
at suboptimal (0.05 mM) and optimal (2 nM) concentrations of
FeSO4 (Figure 18). In the presence of 0.05 nM FeSO4, 1.0 nM
ascorbate increased QPKIC oxidase activity 67%. However, in the
presence of 2 mM Fe504, ascorbate had very little effect on
d-KIC oxidase activity. The stimulatory effect of ascorbate,
therefore, may be solely due to its capacity to keep iron in the
reduced, ferrous state. As noted earlier (Table 6), NADH and

NADPH can replace ascorbate as a source of reducing equivalents.

Activation of the drKetoisocaproate Oxidase
In routine assays of a—KIC oxidase activity, the enzyme was

preincubated with Fe504, ascorbate and dithiothreitol in the

Figure 17.

a-Ketoisocoprooie Oxidase Activity
umole1 4CO2 produced/hrlmg

79

 

.h

 

 

 

 

l 1 1 I I I
| 2 ‘5! 5 IC> I5 2C>
Eesog mM

Effect of Peso? Concentration on d-Ketoisocaproate Oxidase
Activity. d-K C oxidase activity was measured as in Methods
(Method B) except for the variation of the Fe504 concen-
tration. Each assay contained 10 uL (12 ug Protein) of the
d-KIC oxidase (Sephacryl S-200 pool; Table 9). Values are
mean of duplicate assays.

80

 

 

 

 

 

 

  

 

 

 

' o
o
.z ' N
12.5 g 6 2.0 mM F9804
<I 1:
£ -
3% '0
£3 <0
. o
a a, 4-
59 Ii “‘-<0
0‘“ 0.05 mM F9804
if)
.22 2 2
5; <3
9; s
t!
1 {0 L T 1 L 20
[Ascorbate] :1le
Figure 18. Effect of Ascorbate on d-Ketoisocaproate Oxidase Activity in

the Presence of Optimal and Suboptimal Concentrations of
Fe504. GPKIC oxidase activity was measured as in Methods
(Method B) except each assay contained either 2 mM (solid
circles) or 0.05 mM FeSO4 (open circles) and the concen-
tration of ascorbate was varied. Each assay contained 10 uL
(12 pg protein) of the d-KIC oxidase (Sephacryl S-200 pool;
Table 9). Values are mean of duplicate assays.

81

assay mixture fer one hour. d-[1-14C] KIC was then added to

the assay and release of 14002 monitored. From this type

of assay it was not possible to determine whether Fe504 was

only required for activation of the enzyme (i.e. during the pre-
incubation) or if FeS04 was required during the enzymatic reac-
tion. The experiments shown in Table 11 were designed to answer
this question.

In these experiments the purified d-KIC oxidase preparation
(Sephacryl S-200 pool, Table 9) was incubated at 25°C for 1 hour
with or without 0.5 mM FeSO4, 1.0 mM dithiothreitol (activation
mixture). Ascorbate was omitted from the reactions to reduce the
stimulation of activity caused by carry over of the activation
mixture into the assay mixture. After the 1 hr activation peri-
od, 10 uL of the activation mixture was assayed fer d-KIC oxidase
activity in the presence or absence of Fe504 and dithiothrei-
tol. Hhen FeSO4 and dithiothreitol were omitted from the assay
mixture, the assay concentrations of Fe504 and dithiothreitol
were 0.0125 and 0.025 mM, respectively, due to carry over from
the "activation mixture”.

If the d-KIC oxidase is incubated without FeS04 and dithio-
threitol during the “activation period“, very little activity is
obtained when these compounds are omitted from the assay mixture
(Experiments 6 and 8). The inclusion of FeS04 and dithiothrei-
tol in the assay mixture increased GPKIC oxidase activity, as
expected (Experiments 5 and 7).

when the GPKIC oxidase is incubated with FeSO4 and dithio—
threitol during the "activation period“, d-KIC oxidase activity

82

.mxammu
muau_pa=u we come mew mm:_a> .aummu mgu cusp ocaux_s copam>wuuo use soc; cm>o mecca on use au.>_uoe eczocm
$.08 3:2: 3 32¢ E .2 m5 3 v3.69. come“. 2:. .mucmswcmaxm $05 59¢ 33.28 mm: oaofiouvd .w
can m mucms_coaxm mo madman o» wound no: _ou_ocguopza_u new owed co “cacao damn mwga .mmmodczd .ocacoo
com .umxammm m=_ma msx~cm 0:» Edge cm>o xucmu on can me: e new ~ mucma_cquo co mnemmu =_ uemmmcg ~oupmcgu
lopsu_u use eomom .czogm mm umpcm> ma: cargo; copuanao:_mca as» can .m>ona umumuwvcw mcopuacuzmucou mgu
:. ucmmmca mew: pouwmcguopgupu use comm; .uouu_eo ma: ounncoumu unmuxo Am vogumzv avenue: :_ mm umxommo
an; d_asam we» co 4: OH ..eoccaa =o_ca>_uaa. m_=c cauc< .Ahcav .ou_ac;uo_gu_e :5 o.~ ecu eomae :5
m.c c=o=c_z co =»_; u.m~ an 2; a Lac cacan=o=_ was Am d_n~c .m eacdad_c.c=av ._ooa co~-m .xgaagaam. deb

 

 

m~.o om muc.o mm~o.o 8: m
~a.m co o.~ m.o oz K
mo.o e mmc.o mu~o.o oz m
ww.~ e O.“ m.o oz m
~a.~ om mmo.o mNHo.o mm» v
o~.m on o.H m.o mm» m
um.“ e mmo.o mm~o.o mm» N
~a.~ e o.~ m.o mm» H
cwmuoca me\c;\m_osz Acpsv Azav Azav copcod u=m5_cqum
aup>puu< gawk pea comm; copuu>puu< maven: acmmmcd
mmocpxo u_xid copuMmmmuwacd bps :5 a .eommu :5 m.o

 

copuacucmucoo Numm< .mcpm

 

v0mm... >m mm<o~xo mh<om¢<oom~ohmxia mo zcuh<>~hu<
Hg u4m<h

83

is significantly elevated compared to the activity when these
compounds are omitted during the "activation period" (compare
Experiments 1-4 to 5-8). If the GPKIC oxidase activity is mea-
sured in the absence of Fe504 and dithiothreitol very little
activity is observed (Experiments 6 and 8). However, if the
enzyme is first “activated“ with Fe504 and dithiothreitol a
large increase in activity is seen (Experiments 2 and 4). QPKIC
oxidase activity is always greater when FeSO4 and dithiothrei-
tol are included in the assay mixture itself (Experiment 1 vs. 2
or 3 vs. 4). Activity was also greater when the enzyme was pre-
incubated in the assay mixture for 60 min as opposed to 4 min.

These experiments indicate that the d-KIC oxidase can be
activated by FeSO4 and dithiothreitol prior to the enzymatic
decarboxylation of GPKIC. This "activated enzyme" retains its
activity even when assayed in the absence of Fe504 and dithio-
threitol (Experiment 2 and 4).

The "activated“ GPKIC oxidase has higher activity when 0.5 nM
FeSO4 and 1.0 mM dithiothreitol are included in the assay mix-
ture, than in the absence of these compounds. The lower activity
in the absence of FeSO4 and dithiothreitol was fOund to be due
to a time dependent loss of deIC oxidase activity after removal
of these compounds (Figure 19). After activation of the d-KIC
oxidase (Sephacryl S-200 Pool, Table 9) with 1 mM FeSO4, 0.5 mM
ascorbate, 1 mM dithiothreitol, the activity was monitored vs.
time in an assay mixture containing 1 mM FeSO4, 0.5 mM ascor-
bate, 1 mM dithiothreitol (solid squares), or in an assay mixture

without these compounds (solid circles). a-KIC oxidase activity

Figure 19.

84

 

 

 
 

 

 

 

a
z; I
8 g FeS Ascorbate DTT /’
4'- o //
'6 g 3; ,1
I
2 N ,/
I
3. 8 2- /’ J
g:
_ , No Addition pH 6.5
*3 E ' 1—-*-~£No Addition pH 28
5 1' 8 1 flJ
20 40 60

Assay Incubation Time (Min)

Stability of the "Activated" drKetoisocaproate Oxidase in
the Presence or Absence of FeSO4, Ascorbate and Dithio-
threitol. The d-KIC oxidase was “activated" by incubation
for 1 hr at 25°C in a mixture containing in a final volume
of 0.25 mL: 0.2 mL (0.24 mg protein) purified d-KIC oxidase
(Sephacryl S-200 pool, Table 9; solid symbols)'or 0.2 mL
buffer (20 M Tris HCl pH 7.8, 1% isopropanol, 0.1 M NaCl,-
5% thioglycerol, open symbols); 1.0 mM FeSO4, 0.5 nM
ascorbic acid and 1.0 nM dithiothreitol. Then, 10 uL of
this mixture was assayed vs. time as in Methods (Method 8)
except the preincubation was omitted and the incubation time
was varied (solid and open squares). In two sets of experi-
ments, FeSO4, ascorbate and dithiothreitol were omitted

from the assay (open and closed circles and triangles) and
in one of these experiments, the pH of the Tris-maleate buf-
fer was adjusted to 7.8 with NaOH (open and closed trian-
gles). Open symbols designate the “blank“ activity. The
blank activity has not been subtracted from total activity
(solid symbols). Values are mean of duplicate assays.

85

was linear vs. time for 60 min when FeSO4, ascorbate, and
dithiothreitol were included in the assay mixture. The initial
activity of the "activated“ GPKIC oxidase in the absence of
FeSO4, ascorbate, and dithiothreitol was slightly greater than
in the presence of these compounds. However, the activity was
rapidly lost when these compounds were omitted from the assay
mixture. The QPKIC oxidase activity was even lower when the
“activated" enzyme was assayed in the absence of FeSO4, ascor-
bate and dithiothreitol at pH 7.8 (solid triangles).

In summary, the d-KIC oxidase is "activated“ in the presence
of FeSO4, ascorbate and dithiothreitol. After removal of these
compounds (i.e. by dilution) the GPKIC oxidase remains active for
a short period of time bef0re returning to an inactive state. In
order to carry out reactions which are linear with time for up to
1 hr, FeSO4, ascorbate and dithiothreitol must be included in

the assay mixture.

Optimal Assay Conditions f0r the d-Ketoisocaproate Oxidase

The GPKIC oxidase activity was linearly related to time (up
to 60 minutes) and protein concentration (up to 6 ug); see Figure
20. The arEl-14CJ KIC decarboxylating activity in the
absence of enzyme (Figure 20; Blank) was negligible compared to
the enzymatic activity.

Hhen a large number of assays were undertaken, the apparent
d-KIC oxidase activity, as measured by release of 14C02 from
ae[1-14C] x1e, decreased gradually. Since the purified a—KIC

oxidase is stable for several weeks at 4°C, this loss of activity

86

Figure 20. Effect of Time and Protein Concentration on the d-Ketoisocap-
roate Oxidase Activity. d-KIC oxidase activity was measured
in the Sephacryl S-200 pool (Table 9) as in Methods (Method
B) except for the variation of time (upper figure) or protein
concentration (lower figure).

87

 

 

  
  

_O_

60

4O

20

 

 

nu
co

m w m
ceased N8... c8:

33:3 0830 mcoocaooowcmxid‘

 

Time (minutes)

 

 

 

 

w m ...
2838258 «8.. case
223 $025 eceedsowaaxic

 

Figure 20

88

is most likely not due to instability of the enzyme, but rather
instability of some component of the assay mixture. Indeed, when
the cofactor mixture (containing FeSO4, ascorbate and dithio-
threitol; see Methods) was added to the assay mixture 100 min
befbre addition of the enzyme, a 15% loss of activity was
observed (Experiment C, Table 12) compared to when the enzyme was
added immediately (Experiment A) or 49 min after addition of the
cofactor mixture (Experiment B). The longer the cofactor mixture
remained in the Tris, maleate buffer before addition of enzyme,
the lower the observed QPKIC oxidase activity (unpublished obser-
vation). The cofactor mixture itself was not unstable. The use
of cofactor mixture which was set at 4°C for 101 min (Experiment
0) gave the same results as fresh cofactor mixture (Experiment A
and E). The enzyme (d-KIC oxidase) was also stable during this
120 min period (compare Experiments A and E). Hhen added to the
assay buffer, the cofactor mixture may gradually remove a vital
component of the d-KIC oxidase reaction or slowly produce an
inhibitor. It is quite likely that the cofactor mixture depletes
the assay mixture of 02 or a complexed ferm of 02 which is
required by the GPKIC oxidase for activity. Hhen assays are all
carried out within 50 min, no apparent loss of GPKIC oxidase
activity was seen (Experiment B, Table 12). Therefore, the
cofactor mixture could be added to all assays at the same time.
However, when a large number of assays were carried out, the
cofactor mixture was added to each assay immediately (within 30

sec) bef0re the enzyme.

89

.wm: weewwe ewe HoH eew wow so wwm we; xwz eeueeweu
we» .a Hewswewexw cw mewewgz .ews m :quw; ewe: eee gmwew eweeewee me: xwz eeueeweu .m Hewswewexw :He

.Am eeguwzv meeguwz cw eweweemwe we wee ewweeee wewz maemme we» .wsxwew we eewwweee ewuw< .weeust Hemme
ow eo_e_eea acewwa e.e we eeac wee: xwz eeceewee can He w_aee mSea ee~-m .eeeaeeeme eEHNea seem .mwee
eu w>wueHwe wee mwswu ewsue HH< .wswu eewN wen we uwm me: u Hewswewexm e» xwz eeueeweu we eewpweee wzw

.maemme wueew—eee we eewe wee wwe—3 $3332.32. .E. H eee a: m5 .75 H
we :ewueeueweeee Hemme Heeww e 9.33 Heewwefiewfiwe r... eH use .38 eweeeeme z: m 43$ 7... m: eweweweee
xwz eeweeweu we» .weeust xemme egg on xwz eeaeeweo we 4: mm we eewuweee wee ewHwe mwewu meewee> we eweee

we: wsxuew egg was» Hewexw Hm eeguwzv meegwwz cw me eweemews we: muw>wuee wmeewxe weeeeeeeemweewxie

 

 

on Hw.o m.o m.omH eowH m
moH mm.e m.o m.HoH HoH a
ma w~.m ooH ooH o u
mm wH.m me cm H m
HeeHv e~.e m.e m.~ N <
Heeueeu ewwueee ms\eg\wHeE: Hewsv Hewev _Aewav Hewswewexm
we a haw>wuee wmeewxe uHxid wEANew eee xwz wsa~em xwz eeueeweu
eeueeweu we eewuweee we :ewuwee< we eewuwee<
ewwzuwe wewu we Heeee< we wsww we weww

 

masthz ><mm< mm<ono wh<ome<uomHohmxie we >h—HHm<hm
NH u4m<h

90

The pH optimum of the QPKIC oxidase was determined using sev-
eral different buffers (Figure 21). The ionic strength was kept
constant in these experiments. The pH optimum in a Tris-maleate
buffer was at 6.0. a—KIC oxidase activity, however, was much
lower when MES, MOPS, or Tris buffers were used. Above pH 7.0
the assay mixtures turned a reddish brown color presumably due to
oxidation of FeSO4.

The effect of the concentration of the Tris-maleate buffer
and other buffers on GPKIC oxidase activity was determined (Table
13). Increasing concentrations of Tris-maleate caused an
increase of d-KIC oxidase activity. When 50 nM MES (ZEN-Morpho-
lino]ethylene sulfOnic acid) or 50 mM Bis Tris Propane (1,3-bis-
[tris(hydroxymethyl)-methylamino]propane) are used as buffer, the
d-KIC oxidase activity is lower than in 20 mM Tris, 20 mM maleate
buffer. In the presence of phosphate buffer very little GPKIC
oxidase activity is observed.

Ionic strength is not responsible fOr the variable a-KIC oxi-
dase activity in different buffers. Increasing concentrations of
NaCl had little effect on d-KIC oxidase activity (Figure 22). ~
However, maleate concentration had a dramatic effect on activity
(Figure 23). d-KIC oxidase activity increased 57% when the con-
centration of maleate was increased from 50 rm to 0.1 M. Optimal
activity was obtained at 0.1 to 0.2 M maleate. Concentrations of
maleate above 0.4 M were inhibitory. The concentration of Tris
base had no effect on QPKIC oxidase activity. QPKIC oxidase in

0.2 M Tris, 0.2 M maleate, pH 6.5 was the same as in 50 nM Tris,

Figure 21.

oproote Oxidase Activity

oisoc

6...
00: M Tris,QlM Mdediei

a, 5_ 002M MES
E 002 M MOPS
g AOZM Tris
a 4'
9
Q. 3...
8"
5! 23L.

(1- Ket

91

 

  
     

 

 

 

 

pH Optimum of the deKetoisocaproate Oxidase. d-KIC oxidase
activity was assayed as in Methods (Method B) except the
following buffers were used: 0.1 M Tris, 0.1 M maleic acid
(solid circles); 0.2 M MES (open circles): 0.2 M MOPS (Open
squares); or 0.2 M Tris (open triangles). The pH of the
buffers was adjusted with NaOH or HCl. The final ionic
strength of all buffers was adjusted to 0.47 with NaCl.
Each assay contained 10 uL (12 ug protein) of the Sephacryl
S-ZOO pool (Table 9). Values are mean of duplicate assays.

92

TABLE 13

EFFECT OF VARIOUS BUFFERS 0N arKETOISOCAPROATE OXIDASE ACTIVITY

 

deKetoisocaproate Oxidase Activity

 

Assay Buffer umole/hr/mg protein
0.2 M Tris, 0.2 M maleic acid pH 6.5 5.93
0.1 M Tris, 0.1 M maleic acid pH 6.5 4.40
50 mM Tris, 50 mM maleic acid pH 6.5 2.96
20 M Tris, 20 nM maleic acid pH 6.5 2.31
50 mM sodium phosphate pH 6.5 0.46
50 nM MES pH 6.5 1.90
50 mM BisTris propane pH 6.5 1.54

 

deKetoisocaproate Oxidase activity was assayed as in Methods (Method
B) except the buffer was varied as shown above. Each assay contained 10
uL (12 ug protein) of the Sephacryl S-200 Pool (Purification B). Abbre-
viations are: MES, ZEN-MorpholinOJethane sulfonic acid; BisTris propane,
(1,3-bis[tris(hydroxymethyl)-methylamino]-propane. The pH of Tris:maleic
acid buffers was adjusted with NaOH. Values are mean of duplicate
assays.

Figure 22.

a-Ketolsocaproote Oxidase Activity

93

 

 

 

 

 

E.
E +
1,
§
if
cu
O
:to l-
.'2 4
[NaCl] M

Effect of NaCl Concentration on the d-Ketoisocaproate Oxi-
dase Activity. d-KIC oxidase activity was assayed as in
Methods (Method 8) except the buffer used was 50 mM Tris, 50
mM maleic acid pH 6.5. The concentration of NaCl in the ,
assays was varied as shown. Each assay contained 10 uL (12
pg protein) of the Sephacryl S-200 pool (Table 9). Values
are mean of duplicate asays.

Y

Figure 23.

a-Ketoisocoproote Oxidase Activlt
umole "C02 produced/Hr/mg
4>

94

 

00

g»
o

[\J
I

 

 

 

05 LC
flVloleote] M

Effect of Maleate Concentration on the d-Ketoisocaproate
Oxidase Activity. d-KIC oxidase activity was assayed as in
Methods (Method B) except the buffer used was maleic acid
adjusted to pH 6.0 with Tris base. The concentration of
maleate in the assay was varied as shown. Each assay
contained 10 uL (12 ug Protein) of the Sephacryl S-200 pool
(Table 9). Values are mean of duplicate assays.

95

TABLE 14

EFFECT OF ADP 0R EDTA ON a-KETOISOCAPROATE OXIDASE ACTIVITY

 

QPKECOTSOCGPFOGCE Oxidase Activity

 

Addition umole/hr/mg protein
None 1.93
1 "M ADP 0.78
2 mM ADP 0.58
5 nM ADP 0.59
10 mM ADP 0.79
1.1 M EDTA 0.10
2 mM EDTA 0.11
5 MM EDTA 0.09

 

d-Ketoisocaproate oxidase activity was measured as in Methods (Method
B) except 50 mM MES pH 6.0 was substituted fer 0.2 M Tris, 0.2 M maleate
pH 6.5. ADP and EDTA were added as indicated. Each assay contained 10
uL (12 ug protein) of the Sephacryl S-200 pool (Purification B, Table 9).
Values are mean of duplicate assays.

E.

96

0.2 M maleate, pH 6.5 (5.78 vs. 5.75 umole/hr/mg protein of Seph-
acryl S-200 pool).

Maleate may activate the d-KIC oxidase by forming a chelate
of Fe2+ which is favorable fer the catalytic reaction. EDTA
and ADP, other iron chelators, were tested in the presence of 50
mM MES, a buffer which has very little tendency to bind metal
ions (128). These compounds caused considerable inactivation of
GPKIC oxidase activity (Table 14). Therefore, the d-KIC oxidase
either prefers the Fe-maleate complex or maleate activates this
enzyme by some other mechanism.

The effect of time of preincubation of the e-KIC oxidase with
FeSO4, ascorbate and dithiothreitol in the assay mixture befOre
initiating the enzymatic reaction is shown in Figure 24. Optimal
activation of the d-KIC oxidase was achieved after 60 min. Fur-
ther preincubation, for up to 200 min, caused very little change

in e-KIC oxidase activity.

Substrate Specificity of the deKetoisocaproate Oxidase

Crude preparations of rat liver cytosol (Table 3) oxidatively
decarboxylated both GPKIC and d-keto-Y-methiolbutyrate (GPKYMB).
In order to determine whether or not both of these d-keto acids
are decarboxylated by the same enzyme, decarboxylation of a-[l-
14C] KIC and d-[1-14C] KYMB were monitored in the various
purification fractions (Table 15). The ratio of decarboxylation
of d-KIC to that of d-KYMB was approximately 1.0 and did not vary
significantly throughout the purification. This indicates that

97

 

  
 
 

 

2‘

2E

gist

a§4

58‘“

e?-

{3 .

 

 

 

I00 200
Preincubation Time

Figure 24. Effect of Preincubation Time on a-Ketoisocaproate Oxidase
Activity. d-KIC oxidase activity was assayed as in Methods
(Method B) except for variation of the preincubation time.
Each assay contained 10 uL (12 ug protein) of the Sephacryl
S-200 pool (Table 9). Values are mean of duplicate assays.

98

TABLE 15

DECARBOXYLATION 0F orKETOISOCAPRDATE AND arKETO-Y-METHIOLBUTYRATE
BY PURIFICATION FRACTIONS

 

 

nmole 14C02/min/mg protein Ratio

Fraction a-[1-14c1 :IgStrfifl-lm me 'aTaxviiIiE
10,000 x g supernatant 0.61 0.53 1.15
DEAE-cellulose pool 7.60 7.24 1.05
Phenyl pool concentrated 23.4 23.4 1.00
Sephacryl S-200 pool 94.5 95.2 0.99

 

Decarboxylation of arEl-14C] ketoisocaproate (GPKIC) and o—[l-
14C] keto-Y-methiolbutyrate (o-KYMB) were measured in purification
fractions (see Table 9) by Method B. Values are mean of 2 determina-
tions.

99
one enzyme catalyzes the decarboxylation of both of these sub-
strates.

The apparent Km's of the GPKIC oxidase fer GPKIC and QPKYMB
were determined using optimized assay conditions (Figure 25).
The apparent Km for oeKIC was 0.32 i 0.02 mM. The apparent
Km for QPKYMB, however, was almost 6 times higher, 1.90 t 0.12
mM. The apparent Vmax determined in the presence of oeKYMB
is higher than when determined in the presence of oeKIC as the

substrate (247 t 6 vs. 130 t 3 nmole/min/mg protein).

V. Studies on the Mechanism of the deKetoisocaproate Oxidase
A. Product Identification

An 18 fold purified preparation of d-KIC oxidase (Sephadex
G-150 pool; Purification A) was used to determine the reaction
products. d-[4,5-3H] KIC was incubated with the partially pur-
ified enzyme and the products were separated by anion exchange
chromatography. A profile similar to that shown in Figure 29A
was obtained. Some radioactivity, possibly due to 3H20,
eluted in the water wash (not shown in Figure 29A). Hhen e-[4,5-
3H] KIC was incubated under identical conditions, but without
the enzyme preparation, only one peak cochromatographing with
a-KIC was seen (data not shown).

The major product, peak I of Figure 29A, was analyzed by gas-
liquid chromatography (Figure 26). One major peak with a reten-
tion time of 15.1 min was observed. Greater than 80% of the
injected radioactivity was associated with this peak. Using this

system, authentic o-hydroxyisovaleric acid, 3,3-dimethylacrylic

100

.HeeHv cemewaw: we ewumwmmzm weeeweeee mewugaww: ewee wee mewme AH: .mewmeew umem

.xwwmew>we= weepm eemwgewz .z<xuoze zHgHsz mwmaHeee ewueesee an eweHEewuwe wew: xe5> eee

ex we mweHe> .meewueeHEewewe e we wmeew>e wee mweHe> .He wHeewv Heee oouim Haeeegewm wee

we Hewweeee a: my H: m eweweeeee xemme :eem .meewaweamewe zuweeHw> Hewuwew wesmew ea ewwee>
we: wewe eewueeeeew wee eee e:egm we ewwee> wew: mz»x mueHiHuie new uHx nueHngie we
meeweeeweweeee we» .wmee wwew saw: o.m :e on ewamenee ewee ewe—es z ~.e we: ewwwee wee new we:
N oo eomeocoeH me: we.» eowoeesoe_oee wee .meowoao_w_eoe me_3o__ow ego gee: He eoeeozv aooeeoz
ew me ewxemme we: Hmv mz>x mueHiHuie ee H<V qu mucHiHuie we eewueHaxeeeeewo .auw>wwe<

 

       

 

 

 

Qmmflwxc OHQLEH:£—O..£Hfl=—l>l°00¥ld LO muflOc—Qfluom POHOXIO CO COFHQLHCNUCOU @Hflsumnam $0 Humwhm
:5 @5er .22 82.3
e. w w. _
e .35. 6:5)
acrezeeoawsagxwy\. no mw m. e .mr

 

 

 

geeexoecawteuuxo;
2e «8 newts:

 

 

%wmwmwmd 300,.

m’
8

 

 

 

 

 

 

.mN oe=e_e

Figure 26.

101

 

 

 

54 2&7
'9' 242

 

 

CPM x i0'3

 

 

 

IO 20 30
TIME (MINUTES)

Gas Chromatographic Profile of the Major Product of o-Keto-
isocaproate Oxidase. The upper figure shows the gas-chroma-
tographic profile of the major product of o-KIC oxidase iso-
lated as described under Methods. Below this is shown the
radioactive profile of the gas chromatograph effluent.

102

 

 

 

  
     

 

 

m. b} cg-clo’ n-nyaro-yuonuuc octa
1‘ methyl. «to:

30, 59

60
N ‘ noggin-cue .3

«‘1 9 ”$3“ «no
, a «an. .,
._ 301 1!:
- J)
m «HAHH
j: 40 so so :00 ten :40
:- 53
2 1”} Unknown 1, (to. acid
“J 83‘ 59
> 0 85
— 6“ no al.-8.03
P Jr
.¢ 4m 1%
d 1}
I.“
a an

0.

4. C. I. 1. 110 140
M/E

Figure 27. Mass Spectrun of the Major Product of o-Ketoisocaproate Oxi-
dase. The mass spectrum of standard a-hydroxyisovaleric
acid, methyl ester (Michigan State University mass spectrum
library No. 9052), is shown along with the mass spectrum of
the free acid of the unknown compound with a retention time
of 15.1 min on the gas chromatograph (Figure 26).

103

 

.ceee one» e. see e H

Heweu wee we wmeueweewe ex» ea mewwwe xewe zeew w>eee ewee=e wgw .ueeewHe 43 H e cw eweHEewuwe
euw>wweeeweee eee ewpewHHee wew: we m.H we meewueeew .Hu: 2 ~.o we omH eee Hewweeem Ho: 2
o~.o-~o.o eewewH e we .5 omH .uewweeem He: 2 No.eio eewewH e we we omH .ow: 42 w new: ween—w
me: caeHee wsw .ee:Hee Hgmwe eemiooH. iHuiHixw:ee so Nu x m.o e e» ewwHeee we: eewueHem mwgu
we we ewes» .Hs H.m ea ewumenee wseHe> eee o.e ea ewumenee :e we» .eewuemewweaewe he ew>eewe
we: waeuwewewee ewwueee wgw .ewmewpwe NoueH eeeu e» ewe me ewe HwH eee ewee ewuweeeeeng

iweu new we 4: cm we :ewuweee we ewueeHEeww me: eewueewe wgw .uomm we me: n ewueeeeew we: «web
.weHEex: we ~.o neweweueee eee ewumeHe eewmees e eHw; sews: eee Eeewm e new: eweeee gene we:
wee» wzw .eweee me: uHx HueHizuie we» .st eewueewe wee we mueweeeeee ewgae HHe we new»
ieeeeewwee ewe me e ewpw< .Hmeegww: wwmv ew>eewe Heewex—mewsueees sew: wmeewxe uHxie we eewuee
-aeoee .e o.eee. ..oee oe~-m .ecoaeeom. ago a. ..eoa-e-a. one ..oe.. eoeaee. ..oo.-e-e. eo._oa
co .oe.. e..oa. .Hooa-e-.= wo .ea eHV d: mm eeo ...e= ee.ev oeaoceeoeo_eooe HewH-=m-o

2.. m6 .Heuwwefiewfiwe we. o.H .Eee eweeeeme 7... m6 .eomww r... c.H £02 5:: me e... e 3.5.3
:e ewee ewe—es z ~.o .mHew z ~.o "we oH.o we wasHe> Heeww e e. meweweweee st eeweeewe e :w
wmeewxe weeeeeeeemweuwxie ewwwweee wee sew: ewueeeeew me: weeeeeeeemweuwx muvHiauie .wmee

iwxo wueeeeeeemweuwXLe ewwwweew we» we mueaeeee eewueewm wsw we ageeemeueEeegu weweequ Hixw:ea

.eN ocsmwe

104

em oe=m_.

8952 8.62...
.unxm

 

 

 

 

 

 

 

nqu. nxu~w nxu.
H I 11- III H \I\\ttI\/>I\|/\..cwacttskl i:
r\
V.I\\/olt \(x l N.o
0"]!!! \\l\
II/ «‘1 O\ON— .
> .23.... HR - oe
Izcxe/C so! a. .w
flamh .. H .
02.6 H 4:“ wwe
e m
\
<>_.+mH .e
<on
-m:
H
“MN

 

 

e”0| x uouoouxwdp PM

105

acid and isovaleric acid had retention times of 18.4, 6.9 and 3.9
min, respectively.

Mass spectral analysis of the peak at 15.1 min of Figure 28
gave a spectrum which identifies the compound as a-hydroxyisoval-
eric acid (Figure 27). Comparison of the mass spectrum to that
of the methyl ester of authentic B-hydroxyisovaleric acid shows
identical peaks at mass 43, 59 and 85. The difference between
the peaks at 117 and 103 is 14 mass units, which corresponds to a
CH2 group. This is expected since the unknown sample was in
the free acid form whereas the standard was methylated.

The reaction products fermed from o-[U-14C] KIC by the
purified a-KIC oxidase (Sephacryl S-200 pool, Table 9) were also
analyzed. Dowex-l chloride chromatography showed the formation
of 2 products (Figure 28). No radioactivity was detected in the
H20 wash or at the position of peak II in Figure 29A. Peak I
was identified as B-hydroxyisovaleric acid by gas chromatography.
Peak III has not been identified, is present in minor amounts and
is not produced in the absence of the o-KIC oxidase. This
compound migrates similar to isovaleric acid during Dowex-I

chloride chromatography.

Evidence That Isovalerate Is Not An Intermediate

Figure 29 shows the Dowex-l chloride column profile of pro-
ducts obtained from o-KIC oxidase under various conditions. In
Figure 29A the partially purified aPKIC oxidase (Sephadex G-150
pool, Purification A) was incubated with oe[4,5-3H] KIC.

Fe2+, ascorbate, and dithiothreitol were omitted from the

106

reaction mixture. The profile obtained is the same when these
compounds are included in the reaction mixture. This experiment
has been repeated several times and five peaks of radioactivity
are always obtained. One peak is in the H20 wash (not shown).
Peak I has been identified as B-hydroxyisovaleric acid. Peaks II
and 111 have not been identified. Peak II is not found when pur-
ified o—KIC oxidase is used (see Figure 28). Peak III is found
in variable amounts in different experiments. Up to 30% of the
3H-labeled products migrate with peak 111 in some experiments.
Peak 1V_is the unreacted o-[4,5-3H] KIC. This experiment has
been repeated several times with recovery of 85-100% of the
applied radioactivity.

Peak III migrates the same as isovaleric acid on the Dowex-I
chloride column. To determine if isovalerate is a free intermed-
iate in the fermation of s-hydroxyisovalerate, the partially pur-
ified arKIC oxidase was incubated simultaneously with o—[4,5-
3H] KIC and [1-14C] isovalerate (Figure 298). The 14C
label (dotted line) remained associated with the isovaleric acid
peak, whereas 55% of the 3H label (solid line) was incorporated
into B-hydroxyisovaleric acid. This experiment shows that iso-
valerate is not a free intermediate in e-hydroxyisovalerate for-
mation from o-KIC.

Another possibility is that peaks II and III arise from the
metabolism of o-[4-3H] hydroxyisovalerate. Hhen B-[4-3H]
hydroxyisovalerate was incubated with the partially purified
o-KIC oxidase only a-hydroxyisovalerate was found, indicating

that it is a terminal product of this system (Figure 29C).

Figure 29.

107

 

iii“ "

flu in H10 vim-01.

28x
IIIKK:

oPM/ML x 10'3

 

 

1 O 200
FRACTION NUMBER

Evidence That Isovalerate Is Not An Intermediate of the
deKetoisocaproate Oxidase Reaction. Reactions contained
0.2 M Tris-maleate, pH 6.5, 1 mM Na2C03, 25 uL of o—KIC
oxidase (Sephadex G-150 pool, see able 5), and 0.1 mM

o-[4 5-3H] KIC (A); 0.1 mM a-[4,5- H] KI and 0.1 an

[1-1 C] isovalerate (B); or 1.0 nM a-[4- H] hydroxy-
isovalerate (C) in a final volume of 0.1 mL. Experiments 8
and C also contained 0.2 mM FeSO4, 0.4 mM ascorbate, and 1
mM dithiothreitol. The reactions were incubated for 3 h (A)
or 1 h (B and C) at 25°C, then terminated by adding 50 uL
20% trichloroacetic acid. Protein was removed by centrifu-
gation and the supernatant neutralized. This was applied to
a 0.5 x 22 cm Dowex-I x 8 chloride (100-200 mesh) column and
washed with 5 mL H 0. The column was then eluted with a
150-mL linear grad ent of 0-0.02 N HCl, followed by another
150-mL linear gradient of 0.02-0.20 N HCl, followed by 150
mL of 0.20 N HCl. Fractions of 1.5 mL were collected and
the dpm of 3H (solid line) or 14C (dotted line) were
determined in every other fraction. The number abgve each
peak refers to the percentage of the total 3H or 1 C

dpm in that peak. Abbreviations are o-ketoisocaproic acid
(KIC), isovaleri acid (IVA), and s-hydroxyisovaleric acid
(HIVA). 8-[4,5- H] Hydroxyisovalerate was 1501 ted from a
previous reaction of GPKIC oxidase with GP[4,5- H] KIC.

Co

108

Mechanism of Formation of B-Hydroxyisovaleric Acid by the aeKeto-
isocaproate Oxidase

Previous experiments indicated that 02 was required fer
GPKIC oxidase activity (Table 2). In order to determine the
source of the oxygen incorporated into s-hydroxyisovaleric acid,
enzymatic incorporation of 180 from 1802 or H2130
was determined.

Figure 30A shows the mass spectrum of enzymatically fermed
e-hydroxyisovaleric acid when no isotopically labeled oxygen
(other than natural abundance) is present. The ion at m/e 103
contains both oxygen atoms which are incorporated into the newly
fermed B-hydroxyisovaleric acid. The removal of one of the car-
boxyl oxygen atoms, as H20, leads to the fermation of the ion
at m/e 85. One of the carboxyl oxygen atoms originates from the
substrate, o—KIC, while the other carboxyl oxygen atom is incor-
porated from 02 or H20 during the enzymatic reaction. Since
the two carboxyl oxygen atoms are equivalent, one half of any
180 incorporated into the carboxyl group of A (Table 16);
will be lost when fragmented to B (Table 16). 180 incorpor-
ated into the B-hydroxyl group will not be lost during this frag-
mentation. Fragment C (ions, m/e 59,61) contains only the
hydroxyl oxygen of B-hydroxyisovaleric acid. Therefore, only
incorporation of 180 into the B-hydroxyl group of e-hydroxy-
isovaleric acid can lead to an increase in m/e 61. Fragment 0
(ions, m/e 43, 45) also lacks the carboxyl group, should give

identical results as those seen with fragment C, and does.

109

Figure 30. Enzymatic Incorporation of 180 from 1302 or H2130 into
B- -Hydroxyisovaleric Acid. 3- -Hydroxyisovaleric acid was pro-
duced enzymatically in the Opresence of no igotopically
labeled oxygen (A), 90%1802 (B) or 93% H210 (C), iso-
lated, and the mass spectrum determined as described in
Methods. In experiment 0, 3- -hydroxyisovalerate was incubat-
ed in the complete mixture (minus a—KIC) with 93% H2130.

 

110

A

 

 

H
m
9...
H
Mac
COMO
m h
a u
w H.
CI?
0
H
3 3
H e H
CuMic
H
3
m 4
a
.3
H
c
w w

 

140

120

100

40

43

105

 

 

 

 

q 1 4 1 1 l

n

87

107
105

 

51

45

1

4

43

w H. a ..

a. Emzmhz. m>Ffimm

14 1 4 1 4

1'“?

139

100

30

so

103

 

59

 

43

 

190.

1 {l4 1‘14

1mg»... a

a

13. 140

10.

M/E

Figure 30

111

 

 

HH «H om om NN fl me
+0 UIMIU .9
mm Km mm m“ OR o me
«:9
HH ea mm mm mm a He :cuuumzu .0
mm c» on HN HR 0 am
0 H mm m m N am :Ouuumzu
N V am me «e H Km :9 .m
u
m¢ cm Hm Ne me o mm o
~ m mm mH mH ~ No“ =o-u-m=u
N m H¢ as Hm H meg ~19 .<
=o-u
Na Na m m m o mag o
ouoLm—a>om—
-xxcguagum gap: omawz N N wuqu d wuqxu msoum m\e newsman;
a z 1o o o z o o
comwonaucu o“ a” ”H mfi owfi

 

mucaucana a

 

mhzmzw<mu mhu nz< o~u< o~mm4<>om~>xoao>zuu nutmeg >44<u~h<z>~zm z~ owa mo uuzummza

cfi u4m<~

112

when B-hydroxyisovaleric acid was produced in the presence of
1302 (Figure 308), 13% of the molecules incorporated
180 into the B-hydroxy position. Greater than 94% of the
B-hydroxyisovaleric acid molecules contained at least one 180
atom. Since the 1302 used in this experiment contained 10%
1502, the value of 5-8% at m/e 103 indicates that all of
the B-hydroxyisovaleric acid was labeled with at least one
180 atom.

Hhen produced in the presence of H2130 (Figure 306),
60% of the B-hydroxyisovaleric acid molecules incorporated
180 into the B-hydroxyl group. Virtually all of the mole-
cules contained at least one 180 atom, therefbre the carboxyl
group was almost completely labeled with one 180 atom.

When the previously formed product, B-hydroxyisovaleric acid
was incubated with H2130 under conditions identical to
those used for the enzymatic formation of a-hydroxyisovaleric
acid (except QPKIC was omitted), no incorporation of 180 into

the molecule was detected (Figure 300).

DISCUSSION

The results presented herein clearly demonstrate the existence of two
separate aPKIC decarboxylating activities in rat liver; one mitochon-
drial, the other cytosolic. No activity was detected in peroxisomes or
microsomes.

The mitochondrial QPKIC decarboxylase activity is due to the
branched-chain asketo acid (BCKA) dehydrogenase (EC 1.2.4.3 and 1.2.4.4)
which has recently become the fbcus of numerous investigations (1-4,
56-67). This enzyme uses all three branched-chain d-keto acids, GPKIC,
a-ketoisovalerate and arketo-s-methylvalerate, as substrates and requires
both CoA and NAD+ as cofactors. The mechanism of this enzyme is simi-
lar to that of the pyruvate and dpketoglutarate dehydrogenase complexes
(2).

The cytosolic GPKIC decarboxylase activity does not utilize arketo-
isovalerate and arketo-B-methylvalerate as substrates and does not
require CoA or NAD+. This enzyme activity is due to an oxidase which
decarboxylates and hydroxylates aPKIC to form B-hydroxyisovalerate (B-
HIVA). This enzyme also utilizes arketo-Yemethiolbutyrate, the a—keto
analogue of methionine, as a substrate.

Previous investigations indicated that rat liver cytosolic GPKIC
decarboxylating activity was insignificant in comparison to the activity
of the mitochondrial BCKA dehydrogenase (1). However, even when using

assay conditions which had been optimized for the mitochondrial BCKA

113

114
dehydrogenase, almost equal amounts of the mitochondrial and cytosolic
GPKIC decarboxylase activities were fbund (see Figure 1). The cytosolic
arKIC oxidase activity in rat liver is markedly affected by the nutri-
tional and hormonal status of the animal. Dixon and Harper have fbund
that when rats are fed a 50% casein diet for 6 days, cytosolic a-KIC
decarboxylase activity increases approximately 3.5 times, whereas the
mitochondrial BCKA dehydrogenase activity remains constant (111). Dia-
betes also elevates rat liver cytosolic QPKIC decarboxylase activity
without affecting the mitochondrial BCKA dehydrogenase (112). Therefore
it appears that, at least in some situations, the cytosolic a-KIC oxidase
may have an important role in leucine metabolism. The purification and
characterization of this enzyme as well as the identification of the pro-
ducts of the reaction should serve as a starting point for understanding

the function(s) of this enzyme.

The apparent Km of the purified cytosolic d-KIC oxidase fbr a—KIC
was fbund to be 0.3 mM when Optimal assay conditions were used (Figure
25). This is 20 fold higher than the apparent Km of the mitochondrial
BCKA dehydrogenase fbr a-KIC, which is 15 pH (3). The Km of the cyto-~
solic OPKIC oxidase for QPKIC, however, is quite dependent on the assay
conditions employed. Using a crude rat liver cytosolic preparation in
the presence of 1.5 M ammonium sulfate, the apparent Km of the a-KIC
oxidase for GPKIC was 0.03 mM (see Figure 4). The actual affinity of
this enzyme fbr GPKIC in 1112, therefore, may be quite different from
that determined in 11559.

The relative proportion of d-KIC utilized by the cytosolic QPKIC

oxidase as compared to the mitochondrial BCKA dehydrogenase may also be

115
affected by factors other than Km. May gt al. (112) and Dixon and Har-
per (111) have reported an increase in mitochondrial BCKA dehydrogenase
activity after freeze-thawing of mitochondria from rat liver. This may
be due to an increase in the activity of BCKA dehydrogenase due to
freeze-thawing or to an increase in accessibility of the BCKA dehydrogen-
ase to its substrate. If the latter possibility is true then transport
of the a-keto acids into the mitochondria may be rate limiting in some
situations. A transporter(s) of the BCKA has been identified in rat liv-
er (129,130) and may be very important in regulation of BCAA metabolism.
The Km for transport of aPKIC into isolated rat liver mitochondria is
0.2-0.5 mM (131), which is several fbld greater than the Km of the
mitochondrial BCKA dehydrogenase for d-KIC (3).

The subcellular pool in which arKIC is produced nay also determine
which enzyme (GPKIC oxidase or BCKA dehydrogenase) metabolizes this com-
pound. The branched-chain amino acid transaminase is located both in the
mitochondrial and cytosolic compartments of rat liver (26). Therefore,

BCKA's can be formed in at least two different intracellular pools.

The activation of the d-KIC oxidase by ferrous iron (Fe2+),
ascorbate (NADH or NADPH), and a sulfhydryl compound is typical of many
non-heme iron oxygenases (132). Other metal ions tested were unable to
substitute for Fe2+. Ferric iron (Fe3+), however, was equally
effective as Fe2+ in activation of the d-KIC oxidase when ascorbate
was present. This may be due to the reduction of Fe3+ to Fe2+ by
ascorbate. The stimulatory affect of ascorbate on the a-KIC oxidase

appears to be solely due to reduction of iron to the ferrous state

116
(Fe2+). In the presence of high concentrations of Fe2+, ascor-
bate has no effect. Other reducing agents such as NADH and NADPH can
replace ascorbate.

FeSO4 activation of the arKIC oxidase persists fbr up to 10 minutes
after removal of FeSO4 by dilution (see Figure 19). However, d-KIC
oxidase activity quickly diminishes thereafter if FeSO4 is not present
in the assay mixture. This may be explained by a weak association of
Fez+ with the enzyme, either at the active site or at another site
which may be necessary to provide the proper conformation of the enzyme
for catalytic activity.

The GPKIC oxidase activity requires the presence of a reduced sulf-
hydryl compound, such as dithiothreitol or CoASH. Monothioglycerol at a
concentration of 0.6 M (5%) stabilizes the purified enzyme when stored at
4°C. These results indicate that the enzyme contains sulfhydryl groups
which must be reduced in order fbr the enzyme to express catalytic activ-
ity. The stabilization of the a—KIC oxidase by monothioglycerol may not
be entirely due to its ability to keep enzyme sulfhydryl groups reduced.
Monothioglycerol at a concentration of 0.12 M (1%) should be a good sulfe
hydryl reducing agent, yet does not prevent d-KIC oxidase inactivation.
The higher concentration of monothioglycerol (0.61 M) may also protect
the enzyme from inactivation by removing oxygen from the solution. Sev-
eral dioxygenases are known to be inactivated by oxygen (132,133).

Optimal activity of the purified d-KIC oxidase occurs at pH 6.0 in
the presence of 0.2 M maleate, 1 mM FeSO4, 0.5 mM ascorbate and 1 mM
dithiothreitol. The low pH optimum of this enzyme may be due to the fact
that Fe2+ is rapidly oxidized at higher pH to ferric hydroxide. In

117

reactions carried out at pH 7.0 or greater, assay mixtures developed a
reddish-brown color. At pH 6.0 or 6.5 this was not detectable.

Increasing maleate concentration from 50 uM to 0.2 M gave almost a 2
fold increase in the QPKIC oxidase activity. This increased activity was
not due to ionic strength of the buffer. Maleate may complex with
Fe2+ in a fbrm which is favorable for catalytic activity. Other met-
al chelators tested, EDTA and ADP, could not replace maleate and actually
inhibited uPKIC oxidase activity.

Our results indicate that the cytosolic a-KIC oxidase consumes 0.7-
0.9 mol of 02 per mole of C02 released from a—KIC (Table 2). This
ratio is consistent with the formation of s-hydroxyisovalerate from
d-KIC. The expected value of 1.0 was probably not achieved due to the
fbrmation of small amounts of products other than B-hydroxyisovalerate
(i.e., peaks II and III, Figure 29A). Isovalerate is not a free (undis-
sociated) intermediate in the conversion of a-KIC to a-hydroxyisovaler-
ate. Partially purified d-KIC oxidase preparations converted d-KIC but
not isovalerate to B-hydroxyisovalerate. 5,3-Dimethylacrylic acid also
failed to yield s-hydroxyisovalerate when incubated with the purified
d-KIC oxidase (unpublished observation).

In order to determine the enzymatic mechanism of B-hydroxyisovalerate
formation, incorporation of 180 from 1802 or H2180 was
determined (see Figure 31). When GPKIC was incubated in the presence of
1302 gas, greater than 92% of the e-hydroxyisovalerate molecules
incorporated one 180 atom into the carboxyl group. In 15% of the

molecules, one atom of 130 was also incorporated into the a-hydroxyl

CHOOH
CH2
'18

02 CH3-C—OH
CH3

18

~15%

c”oon

CH2
CHa-cflou

CH3

H2180

~60%

Figure 31.

118

CHOOH

CH2
CH3-C—OH

CH3

”85%

CHOOH

CH2
CHa-C-OH

CH3

~30%

Incorporation of 130 from 1302 or

s-Hydroxyisovaleric Acid.

COOH

CH2
CH3-C-OH

CH3

<896

COOH

CH2
CH3-C-0H

CH3

<696

H2180 into

119
group. These results indicate that both oxygen atoms incorporated into
a-hydroxyisovalerate originate from 02. The incorporation of one
180 atom from H2130 into the carboxyl group is not at variance
with this proposal, since GPKEtO acids exchange the ketonic oxygen with
water (134). The low incorporation of 180 from 1302 into the
hydroxyl position of B-HIVA could be explained by exchange with water.
This is indeed the case. When GPKIC is incubated with the arKIC oxidase
in the presence of H2130, 60% of the a-HIVA molecules contain one
180 atom in the hydroxyl group. 180 from gaseous 1302 can
not exchange into this hydroxyl position (134), therefore, incorporation
of molecular 1302 must be enzymatic.

when B-hydroxyisovalerate is incubated with H2130 under identi-
cal reaction conditions used fbr other 130 experiments (minus GPKIC)
no incorporation of 180 occurs. Therefbre, exchange of the hydroxyl
oxygen with H20 must occur with an intermediate of the reaction.

A mechanism for the OPKIC oxidase reaction, consistent with the
results is shown in Figure 32. This mechanism was proposed by Hamilton
for dioxygenases which require an d-keto acid as a cosubstrate (135,136).
In the case of a—KIC an intramolecular reaction mechanism may be involved
with the d-keto group of d-KIC itself serving as the cosubstrate. This
mechanism involves the fbrmation of a peracid from the reaction of enzyme
bound FezT, 02 and d-KIC, as shown. The peracid could then act to
insert an oxenoid oxygen at an unactivated C-H bond in the way that
nitrenes and carbenes insert. The products of the reaction would then be
s-hydroxyisovalerate, C02 and the regenerated enzyme-Fe2+ complex.

Lindblad gt al. have proposed a similar mechanism for the enzymatic

formation of homogentisate from p-hydroxyphenylpyruvate by

nucleophilic attack
at the carbonyl

I I I IV
perisovalerate

 

H20
H202
minor reaction attack of oxenoid
oxygen at ficarbon
v .
'0 '1') 0H
M M
0 O

V VI

isovalerate 29.-hydroxyisovalerate

Figure 32. Proposed Mechanism of the a-Ketoisocaproate Oxidase
Reaction.

121
p-hydroxyphenylpyruvate hydroxylase (134). This enzyme also requires
Fe2+ and ascorbate fbr maximal activity. The results of their
180 incorporation experiments using 1302 and H2130 are
very similar to the results reported herein for the uPKIC oxidase.

The regeneration of Fe2+ during the enzymatic reaction would
explain why the a-KIC oxidase retains activity even after the unbound
Fe2+ is removed from the assay by dilution (Figure 19). The loss of
activity by 20 min may be accounted for by a slow dissociation of iron
from the enzyme due to weak binding at the active site.

The enzymatic formation of isovalerate (peak III, Figure 28 and 29)
could be explained by the mechanism proposed. The breakdown of periso-
valerate before formation of B-HIVA would lead to the formation of iso-
valerate. This uncoupling of the decarboxylation and hydroxylation reac-

tions has been observed with prolyl hydroxylase (137).

The d-KIC oxidase activity is inhibited by other branched-chain and
long-chain a-keto acids. Phenylpyruvate is an especially potent inhibi-
tor of this activity. In patients with phenylketonuria, phenylpyruvate
accumulates in the body fluids. An a-KIC oxidase similar to that de-
scribed herein has also been fbund in human liver (unpublished observa-
tion). Inhibition of this enzyme by phenylpyruvate may adversely affect
leucine catabolism in patients with phenylketonuria.

a-Hydroxyisovaleric acid has been identified in the urine of patients
with several different types of clinical disorders (104,105,108,109,138).
It is currently believed that this compound is produced by hydration of
B-methylcrotonyl-CoA (139) followed by hydrolysis of the resulting

122
a-hydroxyisovaleryl-CoA fbrmed. The cytosolic QPKIC oxidase, however,
can catalyze the fbrmation of a-hydroxyisovalerate directly from a-KIC.
This enzyme has only been fbund in the liver and kidney of rats. No
activity was detected in brain, heart, skeletal muscle or pancreas. The
OPKIC oxidase may function as a ”safety valve“ to prevent buildup of
GPKIC, which is quite toxic. The tissue distribution of this enzyme is
consistent with such a function.

Tanaka gt 31. identified high amounts of a-hydroxyisovaleric acid in
the urine of patients with isovaleric acidemia (104). This disease is
due to a block at the isovaleryl-CoA dehydrogenase (IOO), therefore, no
B-methylcrotonyl-CoA is fbrmed. Since B-hydroxyisovalerate was presumed
to be fbrmed from a-methylcrotonyl-CoA, this observation was puzzling.
The presence in humans of the a-KIC oxidase which forms a-hydroxyisoval-
erate directly from a-KIC could explain this inconsistency.

Landaas (108,109) has identified B-hydroxyisobutyric acid, B-hydroxy-
isovaleric acid, and d-methyl~3-hydroxybutyric acid in the urine of keto-
acidotic patients and has shown that these acids arise from the metabo-
lism of valine, leucine, and isoleucine, respectively. He found a posi-
tive correlation between the concentration of these acids and the degree
of ketoacidosis in a variety of clinical disorders. The branched-chain
amino acids are metabolized similarly via transamination, decarboxyla-
tion, and dehydrogenation to fbrm methylacryl-CoA (from valine), a-meth-
ylcrotonyl-CoA (from leucine), and tiglyl-CoA (from isoleucine). Hydra-
tion of these compounds would yield the B-hydroxy acids, which are seen.
However, there are several reasons to question this pathway for formation
of the B-hydroxy acids. If the fbrmation of these B-hydroxy acids was

due to a block in branched-chain amino acid metabolism, as indicated by

123
the high serum levels of the branched-chain amino acids, then one would
expect other intermediates, such as B-methylcrotonylglycine or isovaleric
acid (100,105,138) to accumulate. Landaas notes that although leucine,
isoleucine, and valine accumulate in the serum of ketoacidotic patients,
other metabolites of branched-chain amino acid metabolism are not seen
(109). A cytosolic QPKIC oxidase activity with properties similar to
those described fbr the rat liver enzyme has also been identified in
human liver (unpublished observation). This enzyme may be responsible,
at least partially, fbr production of the B-hydroxyisovalerate which
accumulates in these patients.

The rat liver cytosolic a-KIC oxidase also uses arketo-Y-methiolbuty-
rate (the keto analog of methionine) as a substrate. The affinity of
this enzyme for a-keto-Y-methiolbutyrate (apparent Km = 1.9 all) is much
less than that for a—KIC (apparent Km 2 0.3 mM). The product fbrmed by
the aeKIC oxidase from arketo-Y-methiolbutyrate has not yet been identi-
fied, but migrates identically to s-hydroxyisovalerate using Dowex-l
chloride chromatography (unpublished observation; see Figure 29). If the
reaction is similar to the GPKIC oxidase reaction, the expected product
would be 3-hydroxy-3-methylthiopropionate. The product of this reaction
is not 3-methylthiopropionate. 3-Methylthiopropionic acid can be extrac-
ted into diethyl ether (140). The product of the GPKIC oxidase reaction
using aeketo-Y-methiolbutyrate as a substrate cannot be extracted into
diethylether (unpublished observation).

Steele and Benevenga (140) showed that rat liver homogenates can
transaminate methionine to form arketo-Y-methiolbutyrate and decarboxy-
late acketo-Y-methiolbutyrate to fbrm 3-methylthiopropionate. In rat
liver, 75% of the decarboxylase activity is associated with the

124

mitochondrial fraction and 15% is cytosolic (141). They reported a high
Km (>1 mM) for the cytosolic arketo-Y-methiolbutyrate decarboxylase
activity. This is consistent with the results presented here and indi-
cates that this activity may only be of importance when methionine or
c-keto-Y-methiolbutyrate levels are extremely elevated, as in hypermeth-
ioninaemia (142).

The mitochondrial arketo4Y-methiolbutyrate decarboxylase activity has
a Km of 0.1-0.6 mM fOr its substrate (141). This activity appears to
be due to the BCKA dehydrogenase (143). The utilization of d-keto-Y-
methiolbutyrate by the same cytosolic and mitochondrial enzymes that met-
abolize GPKIC is interesting, especially since rat liver also contains an
enzyme that transaminates both leucine and methionine to fbrm these

a—keto acids (25).

LIST OF REFERENCES

10.

11.

12.

13.

14.

15.

125

LIST OF REFERENCES

Hohlhueter, R.M. and Harper, A.E. (1970) J. Biol. Chem. 245, 2391-
2401.

Petit, F.H., Yeaman, S.J. and Reed, L.J. (1978) Proc. Natl. Acad.
Sci. U.S.A. 15, 4881-4885.

Parker, P.J. and Randle, P.J. (1978) FEBS Lett. 29, 183-186.

Johnson, H.A. and Connelly, J.L. (1972) Biochemistry 11,
1967-1973.

Connelly, J.L., Danner, D.J. and Bowden, J.A. (1968) J. Biol.
Chem. gig, 1193-1203.

Grant, H.D. and Connelly, J.L. (1974) Fed. Proc. 33, 1570.
Grant, H.D. and Connelly, J.L. (1975) Fed. Proc. 34, 640.

Goedde, H.H. and Keller, H. (1967) in Amino Acid Metabolism and
Genetic Variation (Nyhan, H.L., ed.) pp. 191-215, McGraw-Hill Book
Company, New York.

Kean, E.A. and Morrison, E.Y.St.A. (1979) Biochim. Biophys. Acta
561, 12-17.

Choi, Y.R., Fogle, P.J., Clarke, P.R.H. and Bieber, L.L. (1977) J.
Biol. Chem. 252, 7930-7931.

Bieber, L.L. and Choi, Y.R. (1977) Proc. Natl. Acad. Sci. U.S.A.
14, 2795-2798.

Bieber, L.L., Sabourin, P.J., Fogle, P.J., Valkner, K. and
Lutnick, R. (1980) in Carnitine Biosynthesis, Metabolism, and
Functions (Frenkel, R. and McGroarty, J.D., eds.). pp. 159-176,
Academic Press, New York.

Choi, Y.R., Fogle, P.J. and Bieber, L.L. (1979) J. Nutr. 109, 155-
161.

Fogle, P.J. and Bieber, L.L. (1979) Biochem. Med. 22, 119-126.

16.

17.

18.
19.

20.

21.

22.

23.

24.

25.

26.
27.
28.

29.

30.

31.

32.

33.

34.

126

Solberg, H.E. and Bremer, J. (1970) Biochim. Biophys. Acta 222,
372 -380 o

Choi, Y.R., Clarke, P.R.H. and Bieber, L.L. (1979) J. Biol. Chem.
251. 5580-5583.

Hsieh, B. and Talbert, N.E. (1976) J. Biol. Chem. gg1, 4408-4415.

Markwell, M.A.K., McGroarty, E.J., Bieber, L.L. and Tolbert, N.E.
(1973) J. Biol. Chem. 248, 3426-3432.

Rosenthal, J., Angel, A. and Farkas, J. (1974) Am. J. Physiol.
226, 411-418.

Stillway, L.H., Heigand, D.A. and Buse, M.G. (1979) Lipids 14,
127-131.

Ichihara, A., Noda, C. and Ogawa, K. (1972) Adv. Enzyme Regul. 11,
155-166.

Aki, K., Yoshimura, T. and Ichihara, A. (1973) J. Biochem. 14,
779-7840

Aki, K., Yokojima, A. and Ichihara, A. (1969) J. Biochem. 65, 539-
544.

Ikeda, T., Konishi, Y., and Ichihara, A. (1976) Biochim. Biophys.
Acta 425. 622-531.

Ichihara, A. and Koyama, E. (1966) J. Biochem. 59, 160-169.
Shirai, A. and Ichihara, A. (1971) J. Biochem. 19, 741-748.

Ogawa, K., Yokojima, A. and Ichihara, A. (1970) J. Biochem. 68,
901-911.

Ichihara, A., Noda, C. and Tanaka, K. (1981) Dev. Biochem. 18,
227-231.

Aki, K., Ogawa, K. and Ichihara, A. (1968) Biochim. Biophys. Acta
152, 276-284.

Taylor, R.T. and Jenkins, R.T. (1966) J. Biol. Chem. 241, 4396-
4405.

Tischler, M.E. and Goldberg, A.L. (1980) Am. J. Physiol. 238,
E487-E493.

Odessey, R., Khairallah, E.A. and Goldberg, A.L. (1974) J. Biol.
Chem. gig, 7623-7629.

Tischler, M.E. and Goldberg, A.L. (1980) J. Biol. Chem. 255, 8074-
8081.

35.
36.

37.

38.
39.

40.

41.
42.

43.

44.
45.

46.

47.

48.

49.

50.

51.

52.

53.
54.

127
Odessey, R. and Goldberg, A.L. (1979) Biochem. J. 118, 475-489.

Danner, D.J., Lemmon, S.K., Besharse, J.C. and Elsas, L.J. (1979)
do 8101. Chem. 334-, 5522'5526.

Noda, C., Rhead, H.J. and Tanaka, K. (1980) Proc. Natl. Acad. Sci.
U.S.A. 11, 2646-2650.

Hall, C.L. (I981) Dev. Biochem. 11, 35-40.

Ikeda, Y., Noda, C. and Tanaka, K. (1981) Dev. Biochem. 1g,
41-46.

Miller, L.L. (1962) in Amino Acid Pools (Holden, J.T., ed.). pp.
708-721, Elsevier, Amsterdam.

Felig, P. (1975) Annu. Rev. Biochem. 35, 933-955.

Abumrad, N.N., Patrick, L., Rannels, S.L. and Lacy, N.N. (1981)
Dev. Biochem. 1g, 317-322.

Pardridge, R.M., Casanello-Ertl, D. and Duducgian-Vartavarian, L.

Awapara, J. and Seale, B. (1952) J. Biol. Chem. 191, 497-502.

Shinnick, F.L. and Harper, A.E. (1976) Biochim. Biophys. Acta 111,
477-486.

Ichihara, A., Noda, C. and Goto, M. (1975) Biochem. Biophys. Res.
Commun. 61, 1313-1318.

Khatra, 8.5., Chawla, R.K., Sewell, C.H. and Rudman, D. (1977) J.
Clin. Invest. 52, 558-564.

Dancis, J., Hutzler, J. and Levitz, M. (1961) Biochim. Biophys.
A4313 2,! 60-640

Goldberg, A.L. and Odessey, R. (1972) Am. J. Physiol. 131, 1384-
1391.

Tischler, M.E. and Goldberg, A.L. (1980) Am. J. Physiol. 118,
E480-E486.

Odessey, R. and Goldberg, A.L. (1972) Am. J. Physiol. 121,
1376-1383.

Chang, T.H. and Goldberg, A.L. (1978) J. Biol. Chem. 153,
3677-3684.

Odessey, R. and Goldberg, A.L. (1979) Biochem. J. 118, 475-489.
Livesey, G. and Lund, P. (1980) Biochem. J. 188, 705-713.

55.
56.
57.

58.
59.
60.

61.

62.
63.

64.

65.
66.

67.
68.

69.

70.
71.

72.

73.

74.
75.

128
Elia, M. and Livesey, G. (1981) Oev. 8iochem. 18, 257-262.
Noda, C. and Ichihara, A. (1976) J. Biochem. 88, 1159-1164.

Buffington, C.K., DeBuysere, M.S. and Olson, M.S. (1979) J. Biol.
Chem. ggg, 10453-10458.

Odessey, R. (1980) FEBS Lett. 181, 306-308.

Naymack, P.P., DeBuysere, M.S. and Olson, M.S. (1980) J. Biol.
Chem. 888, 9773-9781.

Randle, P.J., Lau, K.S. and Parker, P.J. (1981) Dev. Biochem. 18,
13-220 '

Odessey, R. (1981) Dev. Biochem. 18, 23-28.

Danner, D.J., Sewell, E.T. and Elsas, L.J. (1980) Dev. Biochem.
1_8_, 29-34.

Aftring, R.P., May, M.E. and Buse, M.G. (1981) Dev. Biochem. 18,
67-72.

Patel, T.8. and Olson, M.S. (1981) Dev. Biochem. 18, 79-84.

Buffington, C.K., Haymack, P.P., Debuysere, M.S. and Olson, M.S.
(1981) Dev. Biochem. 18, 85-90.

Van Hinsbergh, V.H., Veerkamp, J.H., Engelen, P.J.M. and Ghijsen,

Van Hinsbergh, V.N.M., Veerkamp, J.H. and Zuurveld, J.G.E.M.
(1978) FEBS Lett. 88, 100-104.

Paul, H.S. and Adibi, S.A. (1978) Am. J. Physiol. g1, 15494-5499.

Van Hinsbergh, V.H.M., Veerkamp, J.H. and Cordewener, J.H.G.

May, M.E., Aftring, R.P. and Buse, M.G. (1980) J. Biol. Chem. 888,
8394-8397.

Solberg, H.E. and Bremer, J. (1970) Biochim. Biophys. Acta 888,
372-380.

Parker, P.J. and Randle, P.J. (1978) Biochem. J. 111, 751-757.

Bremer, J. and Davis, E.J. (1978) Biochim. Biophys. Acta 888, 269-
275.

76.

77.

78.
79.

80.

81.

82.

83.

84.

85.
86.

87.
88.

89.

90.

91.

92.

93.

129

Craine, J.E. and Connelly, J.L. (1970) Anal. Biochem. 88,
539-546.

Fulks, R.M., Li, J.B. and Goldberg, A.L. (1975) J. Biol. Chem.
3.5.9. 290-298.

Buse, M.G. and Reid, M. (1975) J. Clin. Invest. 88, 1250-1261.

Morgan, H.E., Chua, 8.H., Boyd, T.A. and Jefferson, L.S. (1981)
Dev. Biochem. 18, 217-226.

Goldberg, A.L. and Tischler, M.E. (1981) Dev. Biochem. 18, 205-
216.

Blackburn, G.L., Desai, S.P., Keenan, R.A., Bentley, 8.T.,
Moldawer, L.L. and Bistrian, 8.R. (1981) Dev. Biochem. 18,
521-526.

Hutton, J.C., Sener, A. and Malaisse, N.J. (1980) J. Biol. Chem.
888, 7340-7346.

Malaisse, H.J., Hutton, J.C., Carpinelli, A.R., Herchuelz, A. and
Sener, A. (1980) Diabetes 88, 431-437.

Hutton, J.C., Sener, A. and Malaisse, H.J. (1979) Biochem. J. 184,
303-31 1 o

Lenzen, S. (1978) Biochem. Pharmacol. 81, 1321-1324.

Holze, S. and Panten, U. (1979) Biochim. Biophys. Acta 88, 211-
218.

Shinnick, F.L. and Harper, A.E. (1977) J. Nutr. 181,, 887-895.

Chang, T.H. and Goldberg, A.L. (1978) J. Biol. Chem. 888, 3696-
3701.

Lysiak, R., Stepinski, J. and Angielski, S. (1970) Acta Biochim.
Pol. 11, 131-141.

Bowden, J.A., McArthur III, C.L. and Fried, M. (1971) Biochem.
Med. 8, 101-108.

Bowden, J.A., Brestel, E.P., Cope, R.T., McArthur III, C.L.,
Hestfall, D.N. and Fried, M. (1970) Biochem. Med. 1, 69-76.

Dancis, J. and Levitz, M. (1972) in The Metabolic Basis of
Inherited Disease (Stanbury, J.8., Hyngaarden, J.8. and
Fredrickson, D.S., eds.) p. 426-439, McGraw-Hill Book Company, New
York.

Snyderman, S.E. (1967) in Amino Acid Metabolism and Genetic
Variation (Nyhan, H.L., ed.) pp. 171-183, McGraw-Hill Book
Company, New York.

130

94. Snyderman, S.E., Norton, P.M., Roitman, E. and Holt, L.E. (1964)
Pediatrics 81, 454-472.

95. Donnell, G.N., Liverman, E., Shaw, K.N.F. and Koch, R. (1967) Am.
J. Dis. Child. E, 60-63.

96. Greenberg, R. and Reaven, G. (1966) Pediatrics 81, 934-941.

97. Tanaka, K., Budd, M.A., Efron, M.L., Isselbacher, K.J. (1966)
PT‘OC. Nat]. kado 5C1. U.S.A. ii, 236-2420

98. Shih, V.E., Mandell, R. and Tanaka, K. (1973) Clin. Chim. Acta 88,
437-439.

99. Tanaka, K., Mandell, R. and Shih, V.E. (1976) J. Clin. Invest. 88,
164-172.

100. Rhead, H.J. and Tanaka, K. (1980) Proc. Natl. Acad. Sci. U.S.A.
71, 580-5830

101. Rhead, R., Dubiel, B. and Tanaka, K. (1981) Dev. Biochem. 18,
395-400 0

102. Tanaka, K. and Isselbacher, K.J. (1967) J. Biol. Chem. 818, 2966-
2972.

103. Hishnick, M.M. and Gray, C. (1981) Dev. Biochem. 18, 389-394.

104. Tanaka, K., Orr, J.C. and Isselbacher, K.J. (1968) Biochim.
Biophys. Acta 188, 638-641.

105. Chalmers, R.A., Lawson, A.M. and Hatts, R.M.E. (1974) Clin. Chim.
Acta 88, 43-51.

106. Finnie, M.D.A., Cottrall, K., Seakins, J.H.T. and Snedden, H.
(1976) Clin. Chim. Acta 18, 513-519.

107. Gompertz, 0., Bartlett, K., Blair, 0. and Stern, C.M.M. (1973)
Arch. Dis. Child. 88, 975-977.

108. Landaas, S. (1974) Clin. Chim. Acta 88, 39-46.
109. Landaas, S. (1975) Clin. Chim. Acta 88, 143-154.

110. Felig, F.H., Marliss, E., Ohman, J.L. and Cahill, G.F. (1970)
Diabetes 18, 727-729.

111. Dixon, J.L. and Harper, A.E. (1981) Fed. Proc. 88, 900.

112. May, M.E., Mancusi, V.J. and Buse, M.G. (1979) Fed. Proc. 88,
1028.

113. Murphy, P.A., Krahling, J.8., Gee, R., Kirk, J.R. and Tolbert,
N.E. (1979) Arch. Biochem. Biophys. 188, 179-185.

114.
115.
116.

117.

118.
119.
120.

121.

122.

123.

124.

125.

126.

127.
128.
129.

130.
131.

132.

133.

134.

131
Tolbert, N.E. (1974) Methods Enzymol. 81, 734-746.
Racker, E. (1950) Biochim. Biophys. Acta 1, 211-214.

Pederson, T.C., Buege, J.H. and Aust, 5.0. (1973) J. Biol. Chem.
858, 7134-7141.

Rudiger, H.H., Langenbeck, U. and Goedde, H.H. (1972) Biochem. J.
188, 445-446.

Meister, A. (1952) J. Biol. Chem. 181, 309-317.
Robinson, J. and Cooper, J.M. (1970) Anal. Biochem. 88, 390-399.

Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951)
J. Biol. Chem. 188, 265-275.

Bieber, L.L., Abraham, T. and Helmrath, T. (1972) Anal. Biochem.
88, 509-518.

Bradfbrd, M.M. (1976) Anal. Biochem. 18, 248-253.

Laemmli, U.K. (1970) Nature 888, 680-685.

Bonner, R.M. and Laskey, R.A. (1974) Eur. J. Biochem. 88, 83-88.
Abbot, M.T. and Undenfriend, S. (1974) in Molecular Mechanisms of
Oxygen Activation (Hayaishi, 0., ed.), pp. 167-214, Academic
Press, New York.

Kaufman, S. and Fisher, 0.8. (1970) J. Biol. Chem. 818,
4745-4750.

Andrews, P. (1965) Biochem. J. 88, 595-606.
Good, N.E. and Izawa, S. (1972) Methods Enzymol. 81, 53-68.

Patel, T.8., Haymack, P.P. and Olson, M.S. (1980) Arch. Biochem.
Biophys. 881, 629-635.

Mackay, N. and Robinson, 8. (1981) Dev. Biochem. 88, 227-231.

Williamson, J.R., Halajtys-Rode, E. and Coll, K.E. (1979) J. Biol.
Chem. 881, 11511-11520.

Nozaki, M. (1974) in Molecular Mechanisms of Oxygen Activation
(Hayaishi, 0., ed.), pp. 135-165, Academic Press, New York.

Hebber, S., Harzer, G. and Whiteley, J.M. (1980) Anal. Biochem.
188, 63-72.

Lindblad, 8., Lindstedt, G. and Lindstedt, S. (1970) J. Am. Chem.
Soc. 88, 7446-7449.

135.

136.

137.

138.

139.

140.

141.
142.
143.
144.

132

Hamilton, G.A. (1974) in Molecular Mechanisms of Oxygen Activation
(Hayaishi, 0., ed.), pp. 405-451, Academic Press, New York.

Walsh, C., ed. (1979) Enzymatic Reaction Mechanisms. pp. 501-521,
W.H. Freeman and Company, San Francisco.

Hayaishi, 0., Nozaki, M. and Abbott, M.T. (1975) in The Enzymes
(Boyer, P.D., ed.) 3rd Ed., Vol. 12, pp. 119-189, Academic Press,
New York.

Charles, B.M., Hosking, G., Green, A., Pollit, R., Bartlett, K.
and Taitz, L.S. (1979) Lancet 8, 118-120.

Coon, M.J., Robinson, W.G. and Bachhawat, B.K. (1955) in A
Symposium on Amino Acid Metabolism (McElroy, W.D. and Glass, 8.,
eds.), pp. 431-441, Johns Hapkins Press, Baltimore.

Steele, R.D. and Benevenga, N.J. (1978) J. Biol. Chem. 888, 7844-
7850.

Dixon, J.L. and Benevenga, N.J. (1979) J. Nutr. 188, xxvi.
Perry, T.L. (1967) Can. Med. Assoc. J. 81, 1067-1072.
Livesey, G. (1981) Dev. Biochem. 18, 143-148.

Wilkinson, G.N. (1961) Biochem. J. 88, 324-332.