Alpha-Dialkyl amino acid transaminase


Material Information

Alpha-Dialkyl amino acid transaminase purification and mechanism
Physical Description:
viii, 122 leaves : ill. ; 29 cm.
Bailey, Gordon Burgess, 1934-
Publication Date:


Subjects / Keywords:
Transaminases   ( mesh )
Biochemistry and Molecular Biology Thesis Ph.D   ( mesh )
Dissertations, Academic -- Biochemistry and molecular biology -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1966.
Bibliography: leaves 114-121.
Statement of Responsibility:
by Gordon Burgess Bailey.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000887077
oclc - 24624175
notis - AEJ5294
System ID:

This item is only available as the following downloads:

Full Text





December, 1966



The author wishes to acknowledge the pl guidance

of Dr. W. B. Dempsey, not only during th *.. ':~tion of

this dissertation but throughout the author's pr<.octoral

tE training.

Sincere appreciation is also expresed- to the staff

and students of the Biochemistry Department and In partic-

ular, to Dr. A. L. Koch for a stimulating assocla '.:a

during the author's training.

o The perceptive criticisms offered by Dr. J. A. Deyrup

of the Department of Chemistry during the preparation of

z this manuscript are acknowledged.

>4 The financial support of this research was w-_vided

0 both by U. S. Public Health Service Fellowship,

1-FI-&G:--28252-01 a.-.d National Institutes of Health
Training Grant, 2-TI-GM-00706.

Finally, the author is most grateful to his wife,

Dorothy, whose patience and encouragement have made this

^ o accomplishment possible.








Statement of the Problem 1

ELckground 1

Mechanisms of Pyridoxal Catalysis 4

Faction Specificity of PLP Catalysis 13


Materials 18

Methods 23


Selection and Growth of Organisms 31

Enzyme Purification 39

Enzyme Purity 51

Co-factor Content of the Purified Enzyme 54

The Reaction Mechanism 59

Evidence for a Single Transaminase 70

Other Details of the Reaction Mechanism 80




The Reaction Mechanism 84

The Dependence upon Added PLP for
Transaminase Activity 88

The Natural Substrate 90

Control of Synthesis of the a-Dialkyl Amino
Acid Degrading System in Ps-1 92

On the Nature of the Reaction Specificity 93



VITA 122


Table Page

1. 14C- Carbon Dioxide Release from 1-14C-DL
-Isovaline by Cell Free Extracts of Ps-1. 36

2. Purification of a-Dialkyl Amino Acid
Transaminase. 43

3. Stoichiometry of the Enzyme-Catalyzed
Reaction of a-Dialkyl Amino Acids with
Pyruvate. 61

4. Stoichiometry of the Enzyme-Catalyzed
Reaction of a-Hydrogen Bearing Amino
Acids with Pyruvate. 68

5. Apparent Kinetic Constants for Substrates c-
the Purified Enzyme Preparation. 71

6. Relative Rates of Transamination of Iso\line
and L-Alanine with Several Different a-Keto
Acid Substrates. 73


Figure Page

1. Proposed structures of the aldimine and ketimine
complexes of pyridoxal with an amino acid. 6

2. Proposed courses of non-enzymatic pyridoxal
-catalyzed decarboxylation and decarboxylation
-transamination. 8

3. Proposed mechanism for one half reaction of
enzymatic transamination. 10

4. Reaction sequence followed in the synthesis of
2- C-DL-isovaline. 20

5. Gel filtration through Sephadex G-200 of the
dissolved 33 to 50 per cent ammonium sulfate
precipitate. 47

6. Gradient elution from DEAE-cellulose of the
most active fractions from the gel filtration
step outlined in Table 2. 49

7. Analytical ultracentrifugation and polyacrylamide
gel disc electrophoresis of the final enzyme
preparation. 52

8. Time course of acetone production in the reaction
of AIB with pyruvate catalyzed by the purified
enzyme. 56

9. Spectrum of the purified enzyme preparation. 57

10. Thin layer chromatograms of 2,4-dinitrophenyl-
hydrazones prepared from butanone and the pro-
duct of the enzyme-catalyzed reaction between
isovaline and pyruvate. 62

11. Thin layer chromatograms of isovaline, alanine,
and reaction mixtures of the enzyme-catalyzed
reaction of isovaline with pyruvate. 64

12. Heat stability of the a-carboxyl and a-hydrogen
labilizing activities in the purified preparation. 75

13. Gel filtration through Sephadex G-200 of
a-carboxyl and a-hydrogen labilizing activities. 76

14. Gradient elution of a-carboxyl and a-hydrogen
labilizing activities from DEAE-cellulose. 77

15. Exchange between 60 mM 1-14 C-L-alanine and
60 mM pyruvate, pH 7.3. 82

16. Proposed mechanism of a-dialkyl amino acid
transaminase. 85

17. Proposed conformations of pyridoxal-amino acid
aldimines oriented for release of the a-hydrogen,
the a-R group or the a-carboxyl. 94

18. Possible conformations of L-serine and D-alanine
at the active site of L-serine transhyd:roxymethyl-
ase as predicted from the model proposed by
Dunathan. 95

19. Possible mechanism of a PLP-dependrcnt rcese. 100

20. Possible conformations of AIB and L-alanine in
the active site of a-dialkyl amino acid trans-
aminase with the orientation of the a-carboxylate
anion of both substrates fixed relative to the
plane of the conjugated system or with the bond
to be cleaved fixed relative to this plane. 103

21. Some possible conformations of AIB and L-alanine
-pyridoxal aldimines, at the active site. 105

22. Possible conformations of D-alanine at the
active site. 108





The following is a list of abbreviations of chemical

compounds and enzymes which are used in the text.

AIB a-aminoisobutyric acid

PAL Pyridoxal

PAM Pyridoxamine

PLP Pyridoxal 5'-phosphate

PMP Pyridoxamine 5'-phosphate

GOT Glutamic-Oloacetic Transaminase



Statement of the Problem

The purpose of this research was to elu;ite the

mechanism of the enzymatic degradation of DL-a-2&thyl-a

-amino-n-butyric acid (isovaline) and related a-dC-lky1

amino acids. This was accomplished by 1) enrichlcn.t and

isolation of an organism containing an enzyme that acted

readily on these substrates; 2) purification of the enzyme

responsible for the first step of the degradation; 3)

partial analysis of both the reaction capabilities and

the substrate specificities of this protein; a-d (4)

interpretation of the reaction mechanism and enzyme

reaction specificity in terms of related systems.


The mechanisms of the most common reactions of amino

acid metabolism in mammals, e.g., transamination, oxidative

deamination, and certain eliminations, require cleavage of

the substrate a-hydrogen. Other reactions, such as the

reverse aldol cleavage of serine (1) or deamination of

histidine (2) do not involve the a-hydrogen but depend

upon an electron donating or withdrawing group attached

to the p-carbon. Amino acids with two alkyl groups attached

to the a-carbon have been found to be resistant to enzymatic

attack in animals (3-6), although Weissbach ct al. (7) noted

that a-methyl analogs of 3,4-dihydroxyphenylalanine and sev-

eral other aromatic amino acids were decarboxylated to the

appropriate amines by a pyridoxal 5'-phosphate (PLP)-depen-

dent aromatic amino acid decarboxylase from guinea pig kid-

ney at about one hundredth the rate of the parent compounds.

While the rate was too slow to consider this a physiologi-

cally significant reaction, the finding was mechanistically

significant because it did point out directly, the non

-essentiality of an a-hydrogen for PLP-catalyzed enzymatic

decarboxylat ion.

In 1908, Ehrlich (8) reported that a racemic mixture

of isovaline was resolved by yeast cultures. .;:c.y years

later the thesis work of den Dooren de Jong (9) as reported

by St-phenson (10) listed several common orga'niss which

were able to grow with isovaline or a-amino-isobutyric acid

(AIB) as their nitrogen source. They were B. pol-=xa, A.
aerogenes, B. herbicola, E. coli., and S. marcesens. These

reports have not been confirmed.

The most recent report of microbial metabolism of

a-dialkyl amino acids was that of Aaslestad and Larson

(11) who isolated a pseudomonad which utilized a-amino

isobutyrate (AIB) for carbon. Dialyzed cell extracts of

their organism catalyzed the release of carbon dioxide from

this amino acid when pyruvate and PLP were added to the re-

action solutions. Alanine and acetone were identified as

products. In a subsequent preliminary report (12) these

workers announced a partial purification and characteriza-

tion of the enzymatic system. The PLP and a-keto acid re-

quirement persisted through a ten-fold purification of the

decarboxylating activity. Decarboxylation-transamination

by a single enzyme was proposed.

From tL; evidence presented and from an understanding

of enzymatic transamination reactions the following sequence

can be written for the proposed reaction:

AIB + PLP-aEzyme carbon dioxide + acetone

+ PMP-Enzyme (1)

Pyruvate + PMP-Enzyme ---> alanine + PLP-Enzyme (2)

The details of the overall reaction which appears to have

no clear analogy in previously studied PLP-depsndent onzyze

systems were not reported.

A similar enzyme became the object of the research

presented here. In order to understand the mechanistic

implications of such a reaction sequence, it will be in-

structive to consider briefly, first, some of the available

information about the mechanisms of PLP-catalyzed reactions.

Mechanisms of Pyridoxal Catalysis


The discovery twenty years ago by Snell (13,14) that

a variety of amino acids reversibly transferred their amino

group to pyridoxal (PAL) during autoclaving led to his

suggestion that this ccrpound might serve as an amino group

carrier in enzymatic transamination. This fact was soon

established in his, and other, laboratories (15-17) and the

confirmation stimulated further exploration of the non-enzy-

matic reactions of pyridoxal with amino acids (18-20).

Using structural analogs, the components of the sub-

stituted pyridine ring which were essential for transamina-

tion were determined (21-23). These were founC to be the

fcr---l gc'p, the heterocyclic nitrC^en ortho or para to

the forzyl, as well as a free hydroxyl ortho, or possibly

para, to the formyl function (24). With 4-nitro-salicyl-

aldehyde it was found that the nitro group could replace

the protornted pyridine nitrogen, showing that the electron

withdra-f.-7 capacity of the substituted ring was an impor-

tant property which assisted transamination.

These several findings served as the experimental
foundation for the well-known proposals of the general

mechanisms of pyridoxal catalysis published in 1954 by

the workers in Snell's laboratory (21) and independently

by Braunstein and Shemyakin (25). These mechanisms have

been reviewed and extended considerably since their original

formulation (26-31) but the basic concepts have remained


The initial step in all such reactions is believed to

be the formation of an aldimine Schiff base between the

amino acid and the formyl of pyridoxal (I, Figure 1). The

other bonds to the amino acid a-carbon are weakcnd in this

comp--l: by the inductive effect of the electronegative

pyr.cL~~e nitrogen felt through the conjugated system which
ext'en' to the amino nitrogen. In a subsequent step one of

these bonds about the a-carbon is cleaved to form a ketimine

interrclate (II, Figure 1). This ketimine may be hydro-

lyz-C. l.7-ding to transamination, or may return to the

aldimine form prior to hydroly:sis.

t' '.two critical steps cccur in any PAL-catalyzed

reaction mechanism which determine the ultimate products.

They are first, labillzation of a particular bond about the

amino acid a-carbon in the aldimine to form the ketimine

and second, protonation of the ketimine either at the formyl

carbonQ leading to hydrolysis at the a-carbon and production

of carbonyl products and the amine form of the catalyst, or

protonation at the a-carbon followed by hydrolysis of the

resulting aldimine to yield amine products and the aldehyde

form of the catalyst.




/ \=.

F0 *
0 0

\7 /


o 00
4-) 'I0
0a 0

a o .. H

o) ,0c
0 :03
P.'t O

0. 0 H

0 NO O 1-

. 0 L4

S4 -4' 40
So 00 -ri
0 > 0 0

0i 10

P- 4 0

Qr-d 0 R
961,3 O


PAL-catalyzed, non-enzymatic decarboxylation of

a-hydrogen bearing amino acids is not easily demonstrated

because this reaction proceeds much more slowly than the

competing reactions which are initiated by removal of the

a-hydrogen atom (28). In order to avoid the complication

of these competing reactions Kalyarkar and Snell (20) stud-

ied non-enzymatic decarboxylction by PAL with the a-dialkyl

amino acid, AIB. Not only did the expected PAL-catalyzed

decarb-C".y'lon of this amino acid to isopropylamine and

carbon dioxide occur, but, a decarboxylation-transamination

was c''erccd in which pyridoxal was converted to pyridox-

amine (PAI), and the amino acid carbon skeleton to acetone

and car':-n dioxide. They suggested the reaction schemes of

Figure 2 to explain their findrins. The sequ-nce on the

right including intermediate III is the now acceptedd mecha-

nism of PAL-catalyzed straight a-decarboxylation initially

proposed by Snell's group (32,33) ard by Westheimer (34,35)

in which the a-carbor71 is cleaved directly to form the

ketimine. This mechanism, unlike that suggested by Braun-

stein -.-d '-?-yk'in (25) and Werle and Koch (36), would not

involve the c.--hydrogcn of a normally configured amino acid.

The sequence to the left, including intermediate IV, is the

propos::a mechanism of decarboxylation-transamination. The

formation of the ketimine is the same in both reactions;

the difference is the position of protonation of the

+ -


-o=z o

" \7/ \~
g \


o -



aI I

2 (- 1 2
-a I 2

z x
I 0o


\ \ /

* 0

I 0

i C
I --




a )


o 0


C -

0 0

n -
r 0 L


ketimine intermediate and consequently the site of hydroly-

sis which breaks the complex. Equation (1) of the reaction

proposed by Aaslestad and Larson (11) is the enzymatic

analogy for this non-enzymatic decarboxylation-transamina-


The non-involvement of the a-hydrogen in enzymatic

decarboxylation of natural amino acids was first indicated

when it was determined that deuterium was not lost from

a-deutero tyrosine during decarboxylation with tyrosine

decarboxylase (35). Further, Belleau and Burba (37) pre-

pared both stereo isomers of a-deutero tyrr-nro by decar-

boxzyltion of a-deutero tyrosine and unlabeled t-rozine

by tyrosine decarboxylase in rater and deuterium oxide,

respectively. They then compared the rates of oxidation

of those enzymatically prepared substrates by a stereo

-specific amine olidase with the rates of oxidation by

the cr c enzyme of chemically synthesized R and S a-C:utero

tyranine. The results proved that decarboxylation occurred

with r:-7-tion of configuration (i.e. the proton acquired

from the solvent was added at the configurational position

occupied by the a-carboxyl before cleavage to form the


Transaminat ion

Snell's proposed mechanism (28) of one half reaction

of enzyratic transamination is shown in Figure 3. The

formation of the initial aldimine (II) is represented as

0 0

I ly+





~ /1

a transaldimination, since it is currently believed that

the aldehyde function of PLP is bound as an aldimine to

a lysyl (-amino group of an enzyme when not completed with

substrate (38,39,28). Formation of the ketimine (III) in

this reaction occurs by cle:.v, ge of the a-hydrogen.

The essence of the transamination reaction is that

protonation of the ketimine may occur at either the formyl

ca.rbon or the a-carbon so that hydrolysis leads to tr-.nsfer

of the amino grcup to or from the co-factor. The second

half reaction of transamination, with the appropriate a-keto

acid substrate, is merely the reverse of the h lf reaction

as -'ritten.

It is now well established that the only transamination

mec'.'.-s consistent with abundance of data gathered (40-45)

are ones in which the PL? form of the enzyme is converted to

the PE"IP form by the amino acid substrate followed by regen-

eration of the PLP-enzyme by the a-keto acid sv.-strate.

Some of the most convincing data for this conclusion include

the separate isolation of PLP and PMP forms of glutamic

-oxaloacetic transaminase (GOT) af'er reaction with the

appropriate substrate (41), and the finding that the GOT

apoenzyme is activated by either form of the co-enzyme (42).

Elaborate kinetic studies have be:.r performed by Velick

and Va-.'ra (43), and Henson and Cleland (44) with GOT, and

by Hopper and Se-al (46) with glutamic-pyruvic transamin-

ase. Again the results were consistent only with a

shuttle mechanism in which the substrates interact one at

a time with the two different forms of the enzyme. In

addition, substrate inhibition studies by some of these

'.Trke-c (43,44) provided strong evidence that all substrates

co' the same bind-ng site sequentially.

Both half reactions of transamination are reversible

(40) and measure'snts of exch'iane have been made with homol-

c-_":, asino-a-keto acid pairs in which one member of each

p .- was initially radioactive (47,48).

AL. trransaminases studied thus far are specific for

on ne a-.ino acid ster-coce, nor.eally the T form (30),

altho':h a D-amino acid specific has been

purlfil-c frcm Bacillus subtil'Is (49). Rvecrsibility studies

(48) --iflied that the product amino acid of tr-a-sircation

has the same configuration about the a-'. be .s the sub-

str.te amino ac!i, This may be looked upon as retention

of confc'rcatlc' in the overall reaction, analogous with

enzymatic a-decarbe--lation.

.occnt spctr.l and optical rotatory dispersion m-ca-

suremcnts of the intcr-.ction of a-methylaspartate with GMC

(50) ha'v led "o the proposal of a detailed mechanism of
the revcr bible h.l-f reaction of enzymatic tra.nsEmnation 1i: a. refinc-ont of, but in c-zplete agreement with,

the essential steps of the '.rlier proposed ~echh.nlza.

Reaction Specificity of PLP Catalysis

IHon-cnzy mtic systems

As might be expected, in the non-enzymatic reactions

of PAL with amino acids, more than one reaction may occur

simultaneously leading to a variety of products (18,19).

The structure of the particular amino acid involved and

the reaction conditions, such as pH, ionic strength, or

inclusion of certain polyvalent metal ions determine which

bonds ".bout the a-carbon become cleaved.

r' c-~.--p?e, careful accounting of all species at the

en. c'f t"-'c reaction of a-mcthZlserine with PAL In the pr--

ence of Cu++ (20) revealed that .the amino acid underwent

el'-- car-b -.-lati on-tran'sanination which involved cleav-

age of the a-carboxyl, or cleavage of the a-methylol group.

In addition, the alanine produ-ce'. by7 the latter reaction

underwent subsequent '.-na L.-.tion which required labil-

ization c-' its a-hydrogen.

With the aid of the non-enzymatic model systems, as

much or more has been deduced about the detailed reaction

mechanisms of vitamin B6 catalysis than those of any co

-enzyme. However, the understanding of the steps of these

mesh.n!sms does not explain, and in fact, seems to make

more obvious the absence of an explanation for the reaction
specificity and greatly enhanced rates displayed in the
analogous enzymatic reactions.

The emerging understanding of protein structure and

concepts of macromolecular flexibility, and their influence

upon-binding and catalysis at the active site (5B,52,45),

is providing models to explain these phenomena. Changes

in the conformation of the enzyme protein, induced by groups

on the approaching substrate molecule may be involved in

many cne.ymo reactions (51) to aid binding and orien-ation

for catalysis. The presence of bulky substitucn'-" on the

subs'--te, which apparently are not directly in-olved in

the reaction mechnlIcs is often required in order for b",'-

ins or catalysis to be observed (51). Also, the fact that

enzymes are specific for one enantiomer of an optically

active amino acid pair and -may not be inhibited c'7 the

non-substrate member (53,54), illustrates the importance

of the configuration of groups about one atom to binding

as well as reaction.

Observations of the sensitive relationship between

total substrate structure and configuration, and enzyme

specificity are manifold. The recent work on the specific-

ity of glutamine synthetsase (55) and the studies with

citrate cleavage enzyme (56) are excellent cases in point.

The exploration by Jenkins et al. (40) of heat stabilization
and inhibition of GOT by substrate analogs demonstrated the

importance of both substrate carboxyl groups to binding by

this PLP enzyme.

Recently, Dunathan (57) proposed a model to explain

the reaction specificity of PLP enzymes. It drew in part

upon the conclusions made earlier by Perault et al. (27)

from extensive quantum mechanical calculations with the

molecular orbital method. These earlier workers argued

that all the bond labilizations observed to occur at the

a-carbon of an amino acid-PLP aldimine complex were ener-

getically favorable because of the gain in dclocali1ation

enoerg- acquired by extention of the conjugatid system as

a result of the bond cleavage.

Dunathan emphasized first that, if this gain in reso-

nance energy was to aid kinetically in the bond-bre-kinr

process, the sigma bond to be broken must have a geometry

in the transition state which allows overlap with the pi

-electron cloud of the conjugated PLP aldimine. Maximum

signm-pi overlap would occur when the bond to be cleaved

was oriented perpendicular to the plane of the pyridine

r-ing. The validity of this argument was well supported

by results of reaction rate studies with similar chemical

systems which were sterically restricted from such overlap

(58-60). This overlap is possible since the nature of the

hybrid bonding in the aldimine co.-.ex confines all the

atoms of the conjugated system and the a-carbon of the

amino acid to the same plane (see Figure 17).

Second, Dunathan made the reasonable assumption that

in the aldlminr the binding of the amino acid a-carborylate

anion in the enzyme active site controlled the orientation

of the bonds about the substrate a-carbon, relative to the

plane of the pi-electron cloud of the co-factor. By varia-

tion of the position, relative to the plane of the conju-

gated system, of this carboxylate binding site on different

enzymes, the proper bond about the a-carbon of each enzyme's

substrates could be specifically oriented in a position for

cleavage. Thus, the specificity of protein binding of the

a-carboxyl group dictated the reaction specificity of the

enzyme with regard to the first critical step, bond cleav-

age to form the ketimine. Presumably, additcnral factors

controlled the site of protonT.tion of the ketimine and

further steps.

This nodel is surely an oversimplified explanation

of the fact--ors which are involved in the control of the

reaction specificity of these enzymes, particularly in

that it neglects the obvious importance of the othar

valences of the a-carbon in binding (40). Nevertheless,

it is hard to deny the validity of the argument that cer-

tain geometries relative to the plane of the pyridine ring

will have a lower activation energy for bond cleavage and

that there is a reasonably restricted and specific site on

the protein for a-carboxylate binding which will influence

the orientation of the a-carbon bonds relative to the ring

plane. The predictions of this model appear to fit the

published data on the reaction specificities of individual

PL? enzymes.

The reaction mechanism of the a-dialkyl amino acid

transaminase which was the subject of this study exhibited

properties both of a decarboxylase and a transaminase. For

this resscn it was considered useful to investigate the

dt~..l3 of the mechanism of this enzyme. The reaction

specificity which was revealed appeared inco-sis -nt with

the predictions of the nodel proposed by Dunathan. The

effect of this apparent inconsistency on the validity and

restrictions of his speculations will be discussed after

the mechanism for the enzyme and some of its properties

have b'en described.




All commercial chemicals except isovaline were reagent

grade or better. Lower grade commercial isovaline, which

was determined to be greater than 90 per cent pure by ti-

tr".tion, showed only a single ninhydrin sensitive spot on

paror and thin layer chromatograms. This material was used

in growth =~dia but was crystallized from aqueous acetone

(61) for enzyme assays.

PLP'was purchased either from The California Corpora-

tion for 2iochemical Research or Sigma Chemical Co. PMP

was purch. '-ed from The California Corporation for Biochemi-

cal Research.

Identity and radio purity of the commercial isotopes

used were checked by chromatography and, when necessary,

for organic solvent extractible or volatile, acidic impuri-

ties. Radioactivity measurements were made in 10 ml Bray's

scintillation fluid (621 primarily with a Packard Tricarb

liquid scintillation spectrometer (Model 3003).

Chemical syntheses

One-14C- and 2- 14C-DL-isovaline were synthesized from

butanone, ammonia and potassium cyanide by a Strec'-ir

reaction (63) following the protocol of Levene and Steiger

(61) with slight modifications. K14CN was used for synthe-

sis of 1- 14C-DL-isovaline, whereas 2-14C-butanone prepared

from 1-14 C-acetate was used for synthesis of 2- C-DL-iso-


The separate reactions required for the synthesis of

2-14C-DL-isovaline are outlined in Figure 4. One-14C-acctyl

chloride was prepared in 75 per cent yield by addition of a

six-fold excess of freshly distilled benzoyl chloride to 60

.moles of dry sodium acetate (64) containing 2.0 me 1-14

-acets.e. The volatile product distilled as it was formed

and was collected in a receiving vessel which was submerged

in an acetone-dry ice bath and vented through a calcium

chloride containing drying tube. When the initial exo-

thermic reaction subsided, the mixture was held at 1000

until distillation ceased. The 14C-acetyl chloride was

stored, sealed at -20.

Freshly prepared ethylmagnesium bromide was converted

to the less reactive diethylcadmium, for ethyl addition to

the acetyl chloride, by reacting dry cadmium chloride with

the Grignard reagent. One hundred mmoles finely powdered

cadmium chloride was added.slowly with stirring to 180

mmoles (based on the. assumption of complete reaction of

the initial magnesium content) of iced ethylmagnesium

bromide. Th' resulting ethereal slurry was refluxed for




CH3CH2Br + K:





CH3CH 14coc3
1 3N



C3CH14 3

Figure 4. Reaction sequence followed in tho synthesis
of 2-14C-DL-isoval ine.

90 minutes with stirring. Ether was then removed under

reduced pressure while 90 ml xylene was being added to
replace the solvent.

The previously prepared 1-14 C-acetyl chloride was
brought to 60 mmoles total with unlabelled, freshly dis-
tilled acetyl chloride, diluted with 15 ml xylene and added

slowly with stirring to the organometallic cor-und. The

bath temperature was maintained at 0 throughout the addi-
tion, and then the reaction was refluxed for 60 minutes.
The addition product was hydrolyzed by adding, sequentially,

15 C ice and 40 ml 6 M sulfuric acid.
Th' xylene phase was saved and the aqueous phase was
saturated with sodium chloride and extracted twice with

75 ml fresh xylene. The combined xylene extracts were
washed tiice with 20 ml of 5 per cent sodium carbonate

and finally with 20 ml water. After drying over magnesium

sulfate this organic phase was submitted to fractional dis-

The fraction between 50 and 780 (butanone, bp = 79.60)

(65) contained 50 per cent of the radioactivity initially
present in the washed xylene. A 2,4-dinitrophenylhydrozone
prepared from a small amount of this material, melted
between 116-1180 (butanone 2,4-dinitrophcnylhydrazone,
mp = 1170) (65) and was radioactive. The 14C-but.none
was used in the Strecker reaction without further purifi-

cation. The yield was 30 psr cent based on 14C content.

A portion of the radioactive butanone (approximately

0.5 me) was diluted to 40 mmoles with freshly distilled,

unlabelleC butanone for the Strecker reaction which w,-

accomrlished as by Levene and Steiger (61) through isola-

tion c': the amino acid hydrochloride.

This product was dissolved in 50 ml of water, placed

on a 4 cm diameter column containing 450 g washed IRC-120

(H+) and crashed free of chloride with deionized water.

Ic_'07.aI.e was eluted with 2 N ammonium hydroz.IC and the

volatile solvent was removed aun.-r reduced pressure. The

residue was taken up in water, Cdcolorize.l by filtration

through charcoal, and the amino acid crystallized by addi-

tion of: acetone (61). The crystals were dried to constant

weight at 1100 and the specific activity determined by

liquid scintillation spectrometry. The yield was 60 per

cent based on butanone.

The infrared spectrum was taken in Nujol with a Pe-kin

-Elmer Infracord and compared with that of crystalized,

commercial isovaline. Chromatograms of the product on

paper and silica gel-G thin layers were identical to those

of the commercial amino acid. Confirmation of the identity

of these compounds was obtained during analysis of the

products of enzymatic degradation (below).

Eacteralj cu!t.-ros

All cu-':-.rce- cntaikn2 tho salts -:- orcken.-:-'. Yanof..

sky .r. " (06) as modified by Dc-.sey (67) but exclud-

ing the nitrogen source, ammonium sulfate. This salt solu-

tion, hereafter described as RYB ec.lts was supplemented

with appropriate carbon and nitrogen sources as dicta.ted

by the needs of each experiment. Stocks of the orgr.'. ---

isolated. for this study were. maintained on agar .s -:? of

FYB salts containing 0.2 per cent DL-iso:-.line or 0.2 per

cen-t "L-i-ovaline plus 0.2 per cent glucose. Li.cuid c:i-

tures of one liter or less were c.crated! by s.a':-.'. Ab-

sorbancy at 650 mu was used to me-&ur- cell growth. Unless

otherwise indicated, cultures 'were incub.ted at 300

For la-: scale cultures one liter of f s' 'l. g.:c :n

bacteria was used to ino--.lte each carbc- containing 16

liters of medium. Aeration was with sterile air intro2u.ed

threu.h a fritted glass sw-.r,a. Fno.-ns was v;::.r.s3ed

with CGeneral r'.ectric Antifoam 60. The cells were har-

vested in a 17'..:.rples centrifuge then suspended in 0.01 M

potassium '.h-s-'-.ate, pH 7.5, for --ph lli'ttion. Dried

cells ware ster-c at -200 until used for enzyme purifica-


D .'.y.e pri ancd fi a nP.... o:'.' c rl-

rn t' 1t'a p ',--.rilcc. r *'-- -, assay '-ol" -,.' res nt Cf.

the acetone produced by decarboxylation-trans~minat.on of

AIB with pyruvate. The neutral ketone p.u-:t was sep-

arated from unrcacted pyruv7te by extraction :?r- e.l-.-

linized reaction solutions into carbon tetrachloride.

Acetone 2,4-dinitrophcnylhyrr.cone was prepared in this

sol--rt by a notification of the procedure of Gr.enbcr,:

and Lester (68) and the abscrb_.ncy of the soluble deriva-

tive d-termined at 420 my. Acctcne production was linear

dur>'-- the first 10 per cent of the reaction and increas-ed

_1-_ :'-y with protein ccnce, -ration under the conditions


The r.c~sy contained 20 !,. sodit,= pyrT.7ate, 20 mM AIB,

50 mm potassium phosphate, pH 7.5, plus 0.1 :. _.P ,dded

fro- an acidified, aqueous stock sol-tion just prior to

asZ:-7. ,'3n the PLP concentration in the reaction solution

was critical a fresh solution was,r:T. action was

initiated with up to 0.05 ml c:.-_--, and c.rricc, out at 30

with either boiled or no enzyme in controls.

Up to 0.5 "1 of the reaction mixture was pipetted

directly into 0.2 ml 5 N sodium hydroxide which had been

previously placed in an 18 X 150 mm screw capped culture

tube. Acetone was extracted from this basic aqueous phase

with 10 ml carbon tetrachloride on a rotary shaker for 15

minutes. The aqueous phase was discarded and 3.0 ml 0.2

per cent 2,4-dinitrophenylhydrazine in 2 N hydrochloric

acid added. The tubes -were shaken another 20 minutes to

allow formation and extraction of the neutral acetone

2,4-dinitrophonylhydrazone. The aqueous phase was again

discarded and the organic phase washed tw-ce with 0.5 N

sodium hydroxide. The absorb.incy of the 2.4-dinitrophenyl-

hydrarone in carbon tctr chloride could be measured

directly in the reaction tube with a Bausch and Lo-b

Spectronic 20 colorimeter. An identically treated S-iple

from the control was used as a blank. .For more accurate

measurements, absorbancy was determined with a Beckman DU


A standard curve was constructed by following the same

isolation procedure with aliquots of the assy solution

cont .-:". known amounts of acetone. Absorbancy was con-

vertsed to tm-noes acetcno from the standard curve. One unit

of activity was defined as one Emole acetone produced per

minute, ..- specific activity was units p- r : protein.

Protein was determined by the mnthed of Lo r, et al. (69)

using a bovine serum albumin st.: C..

Butanone formation from,-lation-transamination

of isovaline could be measured similarly, however, a simpler

quantitation of this ketone was possible using 2-14C-DL

-isovaline as amino acid substrate. The ascay procedure

was the same as used to measure acetone production through

the termination of reaction with base. Ten ml toluene was

added at this point instead of carbon tetrachloride and the

ketone extracted into this solvent (using some cf the 2-14C

-butanone prepared for the synthesis of 2- 14C-DL-isovaline,

it was determined that 15 minutes' -hakirn with 10.0 ml

toluene quantitatively extracted the butanone from 0.5

ml or less of the alkalinized reaction solution). An

aliquot of the toluene layer was counted in Bray's sol>-

tion to quantitate the butanone. I:egligible quenching

was noted when up to 5 ml of the toluene 'a-se was counted.

For direct measurement of decarboxylation, the C

-carbon dio::ide released from 1-14C-amino acid substrates

was trapped in hyamine hydroxide and counted by liquid

scintillation spectrometry. A technique similar to that

devised by Cuppy and Crevasse (70) was employed, using

serum stoppered scintillation vil.-s as the reaction, and

cart':-e dioxide collection, vessels. The assay solution

in a volume of 0.8 ml was added and the vial sealed. Two

tenths ml of enzyme appropriately diluted with b--:fer was

injected from a syringe to st"-. the reaction. Reaction

was halted and carbon dioxide released by injecting 0.2

ml 2 I sulfuric acid. After c-.rbon dioxide capture, the

hyamine-conti'-.-:-g conterwell was wiped dry and placed

directly in Bray's scintillation fluid for 14C-carbon

dioxide determinTation. This solvent was slightly less

efficient than the toluene based scintillation fluid nor-

.ll:; used to measure radioactivity in lhaine hydroxide.

However, because it is more tolerant of water, less vari-

ability of counting rate due to aqueous condensation which

accumulated inside the centerwells during reaction and

carbon dioxide collection was observed with Bray's fluid.

All counts were corrected for quenching with 14C-tolumne

internal ct-.:-ndrd and were compared to the bac-l-:round of ncn-enzyme controls.

Exchange b t-tccn alanine and pyruvate was measure .

by ether extraction of the 1- C-pyruvate formed from

1-l- C-L-alanine. At the end of incubation u.' to 0.5 ml

of a reaction solution similar to that dosc-ib,' above

but c, .1..i:ng 1-4 C-alanine and vut.L.llcd pyruvate was

pipe'te'. into screw tr.o 0.2 ml 2 M

sul'uric acid. Fifteen ml c.--d.rous ether was aifed and

the aqueous phase extracted for 30 minutes on. a rotary

s~a.': Aliquots of the ether la7-r were counted in 10

ml of ,ray's solution. The efficiency of the pyruvate

extr..'tion procnC.ure was determined with reaction solutions

to which kno-mn amounts of 1- C-pyruvate had been added.

An approxination of the initial reaction rate could be

obtainc. if e-;trc'tion was acecr:c '.ct- before =cre th i

a few per c-nt of the radioactivity had been converted to


The less-ether-soluble 1-14C-a-keto butyrate, formed
by transamination of 1-14 C-L-amino-n-butyrate was sep-

arated from the labelled amino acid substrate on a Dowex

50 (H+) column. In these ac.Czys enzymatic reaction was

stopped with heat and 1.0 ml cf the reaction mixture

placed directly on top of a 1 X 12 cm column. The acids

were icihed through with 10 ml of water and 0.1 ml of the

eluate was counted'to measure the radCoactive a-keto acid


To measure transamination of non-radioactive amino

ac3.?d, 1-14C-pyruvate was used as the a-keto acid substrate,

and the 1-14C-alanine proCucci was separated from the

be'le.'7.d substrate on a similar column. After the unre-

ac c1 C-p:'r'uvte was washed from the- column the bound

amino P.cids were driven off with 2 N ammonium hydroxide

an- 0.1 ml of this eluate analyzed for radioactivity.

Corrections were made for the quenching caused by the

acid or base present in the column eluates.

S"-''?x G-200 was prcpr-7d for columns in cold 0.02

M po'a-h ua phosph-'-te, 0.05 4 potassium "'ior~-C, pH 7.5.

Depc- ."- on sample size either a 25 X 100 cm laboratory

column (7- rmscia fino Chemicals, Pisca'...-.7, New Jersey),

or a 4.5 X 85 cm plexiglas column construct ed locally, was

used for enzyme purification. After a column was poured

it was equilibrated with the same buffer containing 50 PM

PLP. Earlier columns ::ere elutc. by descenl: buffer

flowi however, more successful V'r7.tionatioens were achieved

later using prc-sized beads of ,';0-60 a diameter and ascer.-

in2 elution.

Removal cf salts, such as ammonium sulfate, fr-m

protein solution., and buffer chari-'s, were accomplisheC

on short columns of Sephadex G-75, equilibrated with the

aprrorriate buffer. The columns -:ere p.c:ed with a volume

of gel calculated to have an excluded volume equal to 120

per cent of the sample volume.

Ion exchange chromatography of proteins was accom-

plicK-. with Schleicher and Sch-':.i Typ. 40 DZ.E-cellulose.

The resin was washed as described by Peterson and Sober

(71) prior to equilibration with buffer.

Various size D.weE-cellulose col:_-n3 were prepared

depe-fi.>- on the amount of protein to be fractionated.

Generally, a ratio of 2.5 onJ protein per ml of packed

col u-a -as sought. Most successful enzyme elutions were

made .-ith a linear, potassium chloride -r2.dient front 0.05 M

to 0.5 C in 0.02 M pct-.ssium phosphate teair p1 7.5.

Si s. "r procedures with E .S-S->?.es, A-50 resulted

in smear'-: of the eluted enzyme- peal so th-.t this medium

was discarded in favor of cellulose.

Ascending paper chromatography of amino acids was

accon.mlihi on Whatman No. 1 filter paper using a luti-

dine:collidine:water:d.iethylamine solvent (72 ).

Thin layer chromatography of amino acids was on glass

plates coated with silica gel G (Research Specialties Co.,

Richmond, California) using phenol:water, containing soi'.i-m

cyanidc (73). Amino acids were revealed with ninhyzrin.

Butanone 2,4-dinitrophenylhydrazone was chromatocraphed

on the same support. The solvent systems used are pointed

out with the experimental results.

Radioactivity of non-quenching substances on thin

layer plates was accurately qvu:.ntitated by scraping 1 X 3

cm sections (approximately 50 -7) of dried uvp.~d gel

from the plates into a scintillation vial and suspending

this po0-?.r in 10.0 ml of a 4 per cent suspension of

thixotropic gel ponder (Cab-o-sil, Godfrey L. Cabot, Inc.).

There was virtually no quenching encountered with this


Washed Do-:ex 50 W- X 4 (H+) was used for amino acid

and a-kcto acid separations on 1 X 12 cm columns.


SelecctLon n?.rid ro Th of Crg .n.-. s

Fr.-'chment on ioo-nl.lne as sole carbon source

Soil from an area previ. :-7 found to yield successful

enrichments on isovaline had "-_-n kept in a bottle of tap

water to which fresh DL-isovaline was add-cd periodica.l"y.

The :".-st enrichment in this project was carried out with

this ss -en-zion. The soil inoculum was v.K..d to ?"B salts

ccn--.'.~"g 0.1 per cent ammonium. sulfate :nd 0.4 per cent

DL-isovoi1ne, Growth was visible aft-r 2 days an-' an ali-

quot was '-::. tr...:C -rrad to fresh liquid mclum ?.nd shaken

for 4 days. Samples of this well-grown culture di-

luted. in sterile s-:r.1o and streY.'---" cn plates containing

the same medium in 2 per cent agar. S--7.1 white colonies

became visible after about 4 days at 300 but rc.chcd full

size only after 10 days at the same tcmprrature. Growth

was the same at room temperature and even slower at 37.

A sinr-le colony was selected and restreaked on agar. An

apparently pure strain of a gram negative organism was

isolatc.. by ec-al such transfers and was then maintaine-'

on ca.-.r slants of the same nc.ium. Microscopic observa-

tions revealed long, (6-7_) actively motile rods. This

bacterium r 7c-" with a doubling tln: of 4.5 hours on bct.h

isovallne and A::. All at'--.c:.ts to incrcose th.

r .te on the a-dc.m1l-yl ci'no acid substrates by such rean~

as variation of substrate concentration, pH, temp.r.ture,

trace =ntal or vitamin content were unsuccessful. Only

when the cells were grown on a rich peptone-sugar medium

could the do.'ubli- time be chzr-t-od to 2 hours. Such a

medium was unsuitable for this project since presumably

it would lead to catabolite rp.rcc-ion of the a-dl..llrl

amino acid .degrading carns.

ir Tn-ints designed to detect the nature of the

cn-::--tic attack by this o.rptm- on the a-diallkl amino

acid -re un7ce'.c-ful. When -" C-DL-isovaline was us:d

in the cro:-:;t' C-i:.U, up to 90 -r c:,tt of the labelled

carbo-yl group -:as trapped in :-T-...'.n s-:'2'*.'d or li-

quid :-"'tures Gre:.- in sealed Tf?>. This re-cald that

it was in'ocd isov.7.1.n. which was bing metabwl'.-ed and

also th'-.t both enantiomers were utilized. Hc'--oer, no
14C-ca.rbn dioxide release could b- detected when tolu-

enized cell suspensions or extr.c'.3 prepared either with

a French Pressure Cell, by sonication, or by alumina

grinding were incubated with the radioactive substrate.

Addition of PLP and pyruvate (or several other a-lc':o

acids) which were required for the decarboxJ-lation-trans-

amination reaction (11) did not produce any detectable

alteration of the substrate. Various co-factors and

metals, a:.:of, c'iTh-r alone or in riborate l'rs, were

also i.:"foctive. 2~Thin layer chromc....r:.:. of the insu-

bated reaction mixtures ztci_.cted that no reaction had

occurred since the only radio..; ,pot present coinciCded

with that of the substrate.

When cells were sh-..:n with isova'line at a ccocntr-

tion which would to one-half full growth in liquid

culture and the re.aindCr of the carbon recuirer^-t was

provided in the form of rny one of a -....':_r of cc, po-tuid.s

suspected as in rr.dilates in the der?.d...tion of icovaline,

grow th reached only the c-tcet poss:'.ble with the amount of

iso.-line present.

T- fl.lure to d'.rr... t'o pat'-.-ay by which DL-iso-

cal..e as d-raded by this b.trium stimulated further

search for a "-ss resistant biological s --=. A --_ber

of available 7l-C-- -Cere screened for their

ability to do any of the follo-"n.:

1) utilize isovaline as car1--,n scurc2 for growth;

2) utilize this amino acid for nitrc-cn in a medium

co:t-i'-ning glucose but devoid of any other nitrogen


3) release C-carbon dioai--c from 1- 1C-DL-le-va.1ne

when allowed to grow in Difco Pen A-s.y medium

(Difoo Laboratories, Detroit, Michigan) plus
1-14C-DL- isoval ine.

B. hirbicola was not :.ailable for this s-r-:cy, however,

the f"'.-'.rs of den Dooren de JonZ (3) were not confirmed

with E. coli, A. aerogenes S. marcesens or B. .

S. cereveslae did not .c,( i-o although yeast

had b-on reported to do so (7).

A sin7.e organism was discovered which both utilized

isovaline for nitrogen and released the labelled carboxyl

as cr::b.-n dioxide but was unable to grow on isovaline for

carbon. No:ne of the other bacteria tested sho:-cd activity

of any sor;t.

The active org:e.nlsm, which had b-7.-n initially isolated

as a ctt' of stock cultures of E. col! B obtained

from Dr. A. L, Koch, was suggested to be in the -,us

Pseudoonas by the followi"- findings:

a) short, motile, g-=.- ne -.ti-. rods when grownm with + glucose and ElYB salts;

b) aerobic growth only a'-.d el,-aline reaction, on

mannrltol, rltose, lactose, Simmons citrate, and

Russell double s-...r acar and in malonate broth

(all prod.-.ts of Difco C.bzre.tories, Detroit,

c) urea not hydrolyzed.

This organism, desi-nnt: Ps-1, was ultimately used as the

source of enzyme for this project. When cells of Ps-1

.fully grc'- in glucose + isovallne were sedimented and

resuspendsd in either 0.4 per cent DL-isovali-e

or 0.,4 per cent AIB plus F2YB salts, no sim l" 1n'- c- ..:

c. .-: r,-'. e-":-- after a w- : of sh*'.'-:i '.-- '. bling .r

during logarithmic gro-th in the glucoec + isovaline me-

- -7 .s 2 =pro-''- lately 2 o" Ts, Ps-1, ras c ct

::?-, Lthe isooalinc utill,__ s co-.~.n isolat-" 1r-.

this labor.".ory and also from that L '.'._od o"- ;.-.".cc ..

and Larcon (11) which grc- on AIB as sole ...- .

EThe '' ". .- --' .

-Cor7 ation that Ps-1 ra" a'e! to acorc< the

amino gr"-o7 c- .oval^ne for nt: -.-- :ete-.'z .n and also

rolc.e.sec_ the a-car.'--1 as carbon d,'.onide ,- consistent

wit- a '-.r' '.tion-trn.n.n..--..tion 11'-- '.-.t proposed

by Ac.- -I? :.' and. Larmn (11. .-..- --.- P7-?

sti'r.' release of carbon dioxide in the o-tr syst.-,

these oc.--..-.:. ::or tested for '".:'. effect on the

ity of cell :C-- ?-'*-'-': cf Ps-1 to --';--lte iso17--

line (Tcble 1), The results sutp:c-'e. such a mechanism.

Prclmin:,zr, to att-.pts to purify the a-dieal qi.-.o

acid ....L ..:"..O from ..-l, were sought rhich

would :'.=:-...o E2.xmum ..tC.hesis of the cr"'z.

Variation of the specific oativity in crude o tracts

of cells Crown out in 0.2 per cent +icali.o + 0.2 per

cent glucose over the pH r2:.J'_ f::-- 6.0 to 7.5 was slic:t'

1..C.r.bon. ?_ioi.e Release from -- -Is--. .
b- C3ell .r: T-.-ts of Ps-la

-..ion Additions CPM X 10-2

Pyruvat e PLP

1 + + 122

2 + 77

3 6

a) L-c--.illCzed cells (0.4 g) were suspend-d in 16.0 ml
0.06 M potassium phcpsr1.te, -H 7.5 and sonicated at
10 fOr 15 Th; 43,000 X o E of
t-h'! sonicate was 7st.:'edo ,4
b) -:.tion 1 contained 60 umole 1-- C-DL-isoval!ne
(-0, 000 cpx), 60 ummole sodlixm :r.72.te, 0.. .ole
PLP, 1-O .0ole -'- --sium -*': sp ..'.e, I 75, a'nd 1.0
l2 cr.. xe:rc.ct in a fir--, volume of 3.0 :l.
cac-ctions 2 and 3 were id.entic. t-7 1 ith the ex-
clusion of the indicated con-'.cs Ine,.ubation was
for 5 ho-rs. B.o0activi- were cc!; to
t'oe of ac.':;'.eD enzyme cn .....nCn controls. The
'-'C-carbon dio.-L.le asc.:7 is described in 7Mcthods.

as was the Tariation when the initial concentration of

ZL-isovaline ranged from 0.1 to 0.4 per cent.

The activity producCd in cells grown with DL-a-amino

-n-butyrate was 50 per cent of that present in i-o-._'-Lne

grown cells. When ammonia was available, alone or v'fith

iso0allne, less than 2 per cent was prc--':. Thus, the

enzyme was bo-h in.duclle and subject to catabolite
r- "."7-'- ion.

'ro tenths per cent isovaline + 0.4 per cent glucose

were chosen as substrates for lar:e scale pro'--.tion of


L-..r -c.le c!'tu s "Tr er ~ -'c'ot. .-n

--? 100 liter cul-',res of Ps-1 were prep.-rc' ':.ring

the course of purif-cation studies. In the first, the

med.'.uo contained RYB salts plus 0.2 per cent DL-1iova.ine

and 0.4 per cent glucose. p5 was initially s--' at 6.0

because of a indication f-mo prelln-.'.?: studies

of greater enzy-e synthesis at this pH. Karvest was

accomplished at 36 hours when the gro'.h rate deore.ased

sh-.rply. The second 100 liter culture con';ains, in addi-

tion to RYB salts, 0.15 per cent DL-isovaline plus 0.15

per cent AIB for nitrog--n. The hC'her total nitrogen as

well o.s 0.5 per cent glu.scce was used to increase the

total cell yield in this Tvelume of medium. The initial

pH was also raised to 7.5 to preclude cessation of

due to acid prod&:tion in the higher fluco-o c-"ce:tZ .ion,

and also 'cI'-!:e the hoped for -.i--. c'--e FC.ld had not

been realized at the lower pH in *': first batch cult-re.

Cells again .-cre harvecstcd at 36 hours. The yield, after

.7ophilization, frem the fi-st culture was 90 g and, from

the second, 141 g dry cells.

The pros'mption of a higher, total enzyme yield in

the second culture was, unfortunately, short lived since

extracts c' these cells had only one-half the specific

activi'-y found in extracts prepared from the first culture.

It -oU.ld be unwise to form conjectures about the rea-

sons fcr the difference in specific activity in these t-o

cul"". :-- considering the cr-ness c:" control c':-r condi-

tic~.- in such lar-2 cultures. In view of -l pr:llinary

studies of the effect of pH and isovline cer-.-trateon

on indo in, one :-ould tend o li natc t ose factors

as causes in themselves. The ..-'-- ...ooS c..tration

in the second fe.ium may have been sufficient to make more

pronounced, existing. catabolite repression. Introduction

of the second nitrogen source, AID, in the sc-ond instance

was the most obvious ch.-ge a'r'l m. be related to the cause

of the reduced induction. C:rtt.-ly, in retrospect, a

more extensive study of the effects c; carbon source and

concentratfa-n, as well as of nitro.. source, on

synthesis, zhoul. have been conducted before largs scale

production was att-:=ted.

The-re is no published- purification of an a-.lalyl

p.ino acid trans .inc.:e. In the verbal presentation of

th'.r preliminary fr1ert (12), Aaslestad and Larson dis-

cussed. s'-ps which h?. allowed a ten-fold purification of

the r--.'.e studied in their laboratory. They mentioned

sensitivi-y of the activity in crude extracts to dCl.sis

and :'... option of activity by scC'.'. ions. The affect of

C!.!7ysis proc bly arose frcm loss of co-factor since activ-

ity could be partially recovered :: subseq-.o:.-.; addition of

PLP. TL-. report ced an a.p-arent -,L of 40 r".

Th.- inhibition by sodium ion could ts relieved by

over-helrmng concentrations of potassium ion. In fact,

they f --.*. that the c-.: :' was protected both di:ring di-

alys ^-.-'. het *'..-.- 'by reasc-z':.e high (0,5-1.0 M)'um chloride con-ntr.-tions.

Al-:.o':.h there appc.-rc'. to be some differences between

their sy.te s and that of Ps-1 there soonmd little reason to

doubt that the reaction catalyzcC was essentially the same.

r_ e.-.le, in Th. laboratcr:,- it was found that as soon

as the enzyme was removed front the cnviroi=ent of the --.11

extract, no activity was measurable witho-.t including PLP

in the P..:s:.. Tl:: app..rnt ::P found, using the crude

extract was c.pp:oxir.-tely 5 -. As a precaution against

dena'--.-:..tion, 50 a2 MLP was included w'-::.-7cr feasible

during the purification, altho~ ;h it was found later that

the purified enzyme roained reasonably stable when stored

for long periods of time in the absence of added co-factor.

A detailed study of the effect of -z-.rcu salts on

activity was not conducted; however, certain incidental

observations were made. The activity was not changcd when

concentrations of pota.zi'am chloride as high as 1.0 M --c7e

incl"'iod in protein solutions. Activity was not lower when

the sodium rather than potassium s-l2t of pyruvw.te was used

in the assay, but was rinishec 65 per cent when the assay

was c.:-'d-cted in mM sodium pyrophcphate b.'ffer. It was not

deter-inod whether the additional sodium ion or pyrophos-

pha.te ion was responsible.

.- r2irtain ionic stre-'"c during the -e.rious purifi-

cation steps and to decrease the possib!7.ity of inhibition

by sodium ion, an 0.05 .T nT c;a.sium chloride, 0.02 M potas-

sium phos:-t.te buffer at pH 7.5 was used during the enzyme

purification, except where pointed out. For brevity this

system will be designated KCl-phos or KCl-phos + PLP when 50

V1 PLP is also present. Unless note-d_ otherwise, all pro-

tein solutions were maintained bet-ecn 0 to 120 during

purification and stored at either -20 or -5.

Initial purification studies

A number of standard p:ocedurce which h.vje been useful

in batch en:yme preparations were tested on a small scale

with the hope of finding one which would allow at least

a two-fold purification from the supernatant of 43,000 X g

centrifuged sonicate without great loss of total enzyme

activity. This supernatant, adjusted to pH 7.5 and con-

taining approximately 25 mg protein per ml, was used for

the preliminary studies.

Procedures which were unsuccessful will be mentioned

only briefly, for reference during possible future purifi-


Neutral, aqueous 4 per cent solutions of either pro-

tamine sulfate or dihydrostreptomycin sulfate, added to

the crude supernatant in increasing amounts raised the

A280/A260 ratio measured after the precipitate was removed,

however, the treatment caused loss of up to one-half the

enzyme activity, which was not recoverable from the


Variable results were obtained with isoelectric pre-

cipitation. Activity began to come out of solution at

about pH 5.2 and over 60 per cent could be precipitated,

and later recovered, by lowering the pH to 4.9. However,

below this pH a large amount of the total protein commenced

to precipitate. It was noted for the first time, during

the pH studies, that no activity could be detected in the

dissolved precipitates unless PLP was added to the assay.

Generally, there was a 25 to 30 per cent loss of total

activity as a result of acid treatment. It appeared that

this method of purification, which sometimes gave about

two-fold purification over the crude supernatant, might

be refined sufficiently to become useful, but the results

of these small scale experiments were too variable to risk

loss of a large proportion of the enzyme in a scaled up


Potassium chloride (0.5 M), but no combination of sub-

strates, protected the enzyme in the crude extract, from

heat up to 530 However, little total protein was precipi-

tated after 30 minutes at this temperature.

Final enzyme purification

A summary of the most successful purification is shown

in Table 2. Specific activity and yield values refer to

that portion which was carried to further purification.

The final preparation from this purification was used for

purity determinations and studies of the reaction except

where noted. Cells from the second batch culture were used

in the purification.

Preparation of cell free extracts.---Lyophilized cells

were prepared for sonication by suspension in KCl-phos +

PLP at a concentration of 50 mg per ml. Twenty-five ml

volumes of the suspension were sonicated with the large

probe of an MSE 60 watt ultrasonic oscillator. The sus-

pension, in a Jacketed glass cell, was cooled by circu-

lating ice water. It was determined that 15 minutes'

4) q




. r-


0 .}- '0 C'\ \0 0 "o
o O 4 .-- 3 4-


or o (0 ~4 .) x

O Ct 0' ON N7
0 H

V^ ~ ~ \3O5 3
0 Cr- C0- C- U

00 0 cl 1 0 d 00 0 l

H\ H
9 H
Va- c^ o o N Nr

O O O r- try aI
r-i O r-r

s o
0 u-'1 S> CO LA H-
-- ^' C^ e-i r-S

O oo -r-'
^. h 0 0
C O 0)

C) %C N a. .)
S'. N H r-H 0
CS P- -- 0 C.'
0. C^ M 0 N\

v t'A
'-4 C / f-^ ^ ^
0 '
rVI VT ,-^ ^ r
H V X -'
0 Q (<' 0 :

, a N 0 0 c 0
rli N4 (' -^ \^ XC C'-


Ii !







J +?


t ti +,

so.ication as s' to release a r?.-r.n- c.-.nt of

I..--.. r-- ti. 3 ae"rial into t:e 43,0C0 X g :.?'--..;-.r.

All oen3;o :. 'ivity remained in "-- c this

- ntrif ation-. ::c:e than 90 per cent of the aotivi

re :. In the t..'' .:.'. .:wr coI-tri f..-c on of the

.hole scnica. o at 100,0C) X j for 90 minutes, and this

-" .'.ed a greater increase in specific actv"-:.- t.-n tLe

lo-er sc eeo centrif:. -ticn. T"*.- 1 0, OO X 7 ,-.- t';: '7" .- 7.,

".:. 'oc. Into wo'" size ol0?.s, -.c tor s. c -20- an.

so~' as :rr"; ateria. for all :?W.r.W': 1" ^ ci.

s'-',?. Th- aoti :- w-as stta..c for s c -' ...hs in this


A'eI-= r: sulfate ,-?.tione.tion_---Aa c '-: sculfacte

precipitation 'as used to concentrate protein sc-.utions

for .'c over *-cl filtration colu-m. s and also to Ceffect

a twc-r2 l- -hf-fold purification from the 100,000 X g
.----.atento A -."rrA'-.ted a.,monium sulfate soltion ras

.- pa.rea at room tcerature, ith ammonm :.;..--.o::ide d,

until fifty-fold di;..d -.'..e measured ? 7.5. C''- cc-

2 'tr."-.olution rss d o.-: c slowly, wth -.-:. into

the r':c.ei n solutionn which also had '-n adjusted to pH 7.5.

',-c> the -"-toin so.:.tion was 33 : cent satturated. adi-

tion ,:cs sto:--::, b'.t -'Irri .-: c:. ;inued fer 15 minmr" ,

The p::"cipi ta: was then ed by c -.t:.; e.nd

cat: .icn raised to 50 "ch caused a much

hea;-' r: poropitate to f c:. This 's i-olt.. as above

and both precipitates -"ere dissolved in a min~lzu vol"-_o

of KC1--:".0 + PLP.

Assa.y- of the dissolve- 33 per cent ar~onium sulfate

precipitate ..::. ..l re-:.z.led 1c- than 10 to 15 per cent

of t-. recovered enzyme activity, whereas 85 to 90 per

c-.t w:as fc'-.nd in the suspen.- 13 to 50 per -. .-.-_'inium

sulfate p7.ipitate. Invariably, however, up to 40 per

.....-, cf the t-:.-l, or?.inal .Lcit was lost- In aCd.:-

t-ic .., the total -^- ..-.'.:-. activity calcul"t-'. :":- ,.'_-ing

the 33 7 c",' sat-..-:.ted. s p tant prir to sa.'.':..tion

to 50 cp-: cent Ias less than the I-.-..ue csalc'u. tedc after

as---- of the dissolved. 33 to 50 per c.--t precipitate. The

cone nc.t.on of ammoni-um. sn ie actr"-.:r --.-- ."':-. clg

the r- of the 33 per cent cup rnat.nt '.as'.l.te to

be 14 r"B Possbl.7r this salt con-': .: r.I.tion pr.lucecl some

inhi'o In c:'ri:" the assay. Pesra.21b the r-oniu si-.-

fate c~ cr.tration in the as?..-- sol:tio:. of the dissolved

33 to 50 :.- c.--.t precipitate aCs n--.:.iole. Aasles-.

end. Larson (12) had re=arked that their enzyme was uninhib-

ited "-- r -"--T 1: ion; however, it is not lo-M-n if they "."-

the suTfate salt_, Thus it :h.t there is an at

least p!rt^>..e.1-r reversible, detrimental effect upon the

enzyme of aEronium sl tate.

Go fitration,--eTo determine if the eye '0zyn be

resolc... with Sephadex gels a portion 0' "-7e 100,000 X

supc :::..-.t was mixed. with slurries of G-25, 50, 75, 150

an d 2"0. which hiad bn oeqr-n.tcI "ith KCl-phos. Af'er

stir-n'- for 2 hours to allow equilib ::-C. of the enzyme

with a,1 avail.t..e -cl-:-:, the gels were allowed "o s-ttle

and an aliqr.ot- of the supernatant c .L- ed. C-: act:i -7

was lo,-r r:--y in the c.-'.e containing G-200, intie..-'-

accesos the eC:-:-.2 to the pc-: 2.'.-: of th'- C1 only.

Lhe dissolved 33 to 50 7--: cent amonir- -.7.' ate re-

cipi~t .'e wich ccnt:'.-. d 3 g. protein in 70 rm a cl-re:

of nso3luble material by cc .gtion and '.o'

incrcasc-. -th sucrose. This sample was or"-.-.. on th:

to? of the 4.5 cm c..-.-.rer covun (see Kethe3s) and oe.v.ttd

with Cl-phos + PLP at a flow rate of 20 ml per '.-. T

.e'ltion pat":ern of an C-.lT, smaller sc-..e run on te. 2.5

on. c..lic.'ter column is shown in Fisure 5, and is *":-ical of

el"it"nr p"tterns from .:th size colv.rns.

As can be seen in this figur,, the ---:. activity
jwasoll re ol2ed from the -."": of "- -.. neoleular we':';

-aterial cluting in the excy ld xn 7 t.r.'a-in e

had an clution -ol.01.e to excluded voli me ratio 7 1.7.

This r---o :-as used -- .-- (74) with a n'Mber of

st-n.rd -r".oins to develop a meeons of molecular wei,'-ht

station :'. gel filtration. Cc-_..p-r:..con c.. the ratio,

1.7, with data accumulated in that s':-.'-. allows a molecular

we -:-..t estimation of about 150,000 for the -dialk.yl asino

acid 1ecrading cn::--e,

S:-Y:..: ACTIVITY (units/Vbc



= /
2 0



ft"i *

0'1 0'r-O


; O H

?-; r-!
T^ 0: ft 0 C M
: c

l0 o -t- -
0 0 d

(N: r-' r O!

*~ 0 < A 0c
o. ^ ;a-
o ?- tO--! o~ o~

j:: 4.) y '-
-^ rz 0 0) H

: ~ LO C r. VI

K- ~~ T' H- ?0 0
-p (' ;,-pa
H -'2aZ)r'
0 < -
<,j- +S.C 2. 4J
4* *>0 F
0 t)(^e; )

(-. -/Ou6 ) Nil310od

Ch? .-:"t active fractions from t-" column were 01..:....

three-fold -.:.r.' -.-" the ammonium s':.-:te fr..-tion....

trea.":"-.t "-o-7'- to be --:. gentle since recovery tc'--led

essent-. .... 100 per cent of '-: activit- place on the

Ion exchange wl.h D,7".-0 1olligose.---T .^o -.-'" actit"

.- -tio;.s o-rz : the gel filtration step c .-.:.-i 967 -

-.'." in 180 ml. These were cc-:.'cd. a.nd a.oc'. rI-"C7

to a 2.5 X 100 nm cl--:2. -c!:I to 85 omn 1lt-h D E -cellulose

which had '%. equlli':it.:f. .Cl-:", :_ 7.5. ""

p.'- in -.c lon was allowed to soa: into the ccl v.1 0
b .... ..... ,,s ashcd w "-'-) 4.50 ml of the : iC br a "' ; or

to -,- -.e any poor7..- bound -":":"In hen ..:.:'.o-. "-.s no

lc-- c .-':ct-c. in the c: .1.' -rnt, the nzs-me s as 1.zite n with

a -11 r potass. c chloride grade-l' starti-- 1i,"0 ml

in Ce. n ...r.. -ir (.-.Ire 6). All the en-e o activity re-

00co- '- I -^ this column a-pparK :--" p peak

near the center of the major p-roe-i pc:f. S -we 95

cent of the ac.cec, activity was -rc.-.c: from the colan,

the exclusion of st:.. Z-. ::-- the ::-:fer system

duri-n el.-.ion a-T:---ed not to be 7-::-'-ently dCotrimental

to the :--me. 2 -. PL? -ich ma pres:ct in the

initial :-'ln : e was etcctD by its yellow color

and f scene uncor .ul tra ct .. L'.t ca. this c-pouCd

clutd- in a peal-k followL. -, but s .'.:-' 0 ". overlapplr.; "-e

er:yLe peal:. Althor '.-- cciscr te s-t'ein banc- were not

. .ZY -E ACTIVITY (:, /m':
0 E




C -



o |H;

4b -, C)
; f-; f
C- 0 0 <

0:: 0~ 0
F! P <>
C)- ^ C

.<: 0 ^)

<., r-s 0' -

... ^ .. G
-io -i a
.P C\ c3

0 C;r-

a -i y'4
;*'! c 0

;o 0) C

C,'./\; ^ a^


f- CO)
-f "d r

q0 0

(liu/7-..') N1310 dl



scra:?....ed on the column, the =cst active fractions were

p'rified four-and one-half-fold. over the last stop.

Cleaner separ.ation from cont;in..ting proteins was

ob"icI. b a co-nd. fractionation of the three '.-,t active

fractions of the first D':.;.-celluleo cole'mn on a smaller

but Id.::t'ic.l column. The three w-_ctin: were ccabined

cn.' returned to ICl-phoo by ";.-. : over Sophaes.r 5-75 with

th:t buffer. All the protein rccov"r-: from the G-75 col-

umn was applied with moc.crate .- pressure to an 0.5 X 18

inch -..-cellulose column and washed with 50 -1 of

ing buffer. Th? same r_ dic.; was used to elute 1-0t with

only 250 -1 in each reservoir to start. A si -1e ..-etoin

pec :-s eluted -'-.. cont" all the enzymatic activity,

but since all fr-actions did not have the same specific

activity there was no doubt of heto -;.n ity of the enzyne

p-,.':. C'-' two most active fractions, which b'.h had the

same specfo.c activity, were a-r-i -.tely tewo-old pu.rl-

fled tc'r the bet fr_-.tions of the first D_:,: column and

contained 22 'per cent of the enzme applied to this last


i--."'. ar monium sulfate f .ticnation.---The abo-: two

fracti.os, which contained slightly over 2 g M-::c-ein r

ml, were :-.--:ht to 50 per cent sa-.-. .tlton by sec- additionn

c-' so. amonium sulfate. The fine precipitate that

formed was sedimented at 27,0CO X (g for 30 minutes and

disso -. i 2.6 l of :C:l-. "pho. Sixty per ccr-t of the

enzyme r-.s -t-o'cd frcn .this s--.op -ith a 20 per cent

increoaso in spec.iic activity.

A one.. :--" seventy-fold p.rification f:rc- the

sz....t:-3 material h:.-. been achieved ct this point. Be-

cause of the relatively small a-mount e-:- protein ::-' '"

aei t'-e a-7-.r.3 d.c..trimental ?::-"ect of ammonium sulfa-,te

fL'.r'-W r purif-c-.ton steps or c'-" :. i.ton .-cre not

.-. ._}... ...'. -~
c.. :-! ul".t- -.- :-lf tion ifn a- n ED'i ncz0o :'o -.,'

E ' iltr.;rliCe,"- the one .-:.-?.-. sc-?-n.--fold
pi .......t..ion was passed "...' sra l Sephaoex

C-75 coln to rid it of roes.:-:.. a=moni'n s~I-,".- and to

chang svnt to 0.1 ':...zim c'.cric.e, 0.02 I

potassiU::a -o.-'-.. te, pH 7.5. C:- resulti-.-s protein con-

centration -as 5.0 "a per al. Sedimentation was at 52,

640 rpm and 21. hntc,;--.phs were taken at 8-rinute inter-

S r r-imu smec: was attained asnd those taken at

26 and C3 minu'v;e are sho-wn in Figure 7. :'-rly all th'.

protein sel:ented in a single, appa.-c.tly ..-. trical peak,

however, upon insp.-. ion a s..:-'ht con':r-...'.ting 'oak

can be seen =r.-:.- be:. ',the -- or ocnponent. The

1 Ultr- ...-..:'-r ft:ation r-. performed ,,.o Arth-_ ::o-.- s.



I I +
Figure 7. Analytical ultracentrifugation and polyacrylamide
gel disc electrophoresis of the final enzyme
preparation. Details of the procedures appear
in the text.

ob crt .i.c.r.tatIon c._ffici'.nt of the r..jor peak

found to be 8.3 S b7 the c..:.>i'c..1 method d:.:oir. ez. by (75). Icle'a:7..r welcht wtr to.'..tio by ultra-

cont.-i:'.-a,"L.t-Lon has nc- been at: .3-

;-- -:.?-ng the limita. .ons of .I-tic' ..-trac'- :--

fP:--.tion as a measure cf protein purity- (especic..-y -'..'

G- fLItr-.ticr., a techniatqe which also Iotionate pro-

:.:-s on the basis of s'l:.-r :- .7- iamic prop-';ies hs

beeoon used c'.:: purification), it was concluded f-.t a

large aero'nt of cer-".1.-ting ;:.en w"O not c'.:. 21 in

t'-: final p--: -.r--.ti2n b' this t ''-.'' .e

Gel ltz-hoeois

.-~~-~i-- i'.oe gel disc elec rehcrrz:- at a rrn.rnn

0:: of -bout 9 was accc-.:.f.rh:11 : the -c':.:. of Clark (7,).

Fif-:. -- "of the fI-.l prepa: ion in 0.2 ml 10 per cent

sucreo- on-li'c to the top replicate gel. and sub-

jectoe to a current of 5 ra r-- : -. The up-:r buff2.r

c:..-:- c.?--:.-.i 0.005 per rn':. Erom ...:.c Blue which the -.:fer and r--.o a :. ring" to

indicate .1- pror-:- of the '.-:ffer fnto '-- ele -:";

fie'd -,-as rc'ov". ',-hen the :-:.f&cr frc:it r:?.c.> the cnedic

end c the gel --I fast moving banIs I'e:e 1cficlt aftr
C-' (::.:'are 7). The 01o,;er no-i7jg : ':ein c'.-'. rIot

-:.r ;o repeosoe"nt a l.:: -... ... c the total, ol-

thc-: qantitatie meas.:.-.-- of the relati-e r2c1.-ities

o. the tro ban6cs were not .'.d.e

A"k-'--- "ts at disc el tr:.h.resis at pH 4.3 (77) :7ore

unsuccessful. In those exper-..-*:-.-.z the protein zigrat.:.

th.:cr-.:h l-he san-'le and spacer gels but not beycnd. the oi-

gin (anode) of the 7.5 Fpr co-; r-.': g, een p. "er

long ties in the electr'.c fcld.. Since the enzyme -ad.

prec'.it--.tced c= cr-.:o ext=--"r at -f 5.0 it was ,re-,ctc.

that it would bc.r a net positiL-v:o T-e at the lower elec-

tr-.'*.- etic It was not detcwrir.- n' .:-- the protein cU'..

r.. .:l--te into the running gel in the several oeperz1:nts

S -rch el electrophorcc~-s cas perefo--.'. '-- the method

of -'.- hies (78) at -.- 8.5. Tho 7:0e pa't-::.-. a1 o0sc:-c:-

in -".acrylamide gels was c'-.-- .',, -. this te-:- '.ue,

althc :- the second minor band ;acs mere diffuse in the

t ,,as assumed that the -.: r ultracentr.... al peak

a- elo :-t: .-oretic band were :-- the eo".-:.-z :-.'

st'-.-:'--. -- those two analytical -.:-teria the r_-.:-re apr--red

to be at least 90 per c::- .":-..

-'".cre was little doubt that PLP ras a co-factor in the

de ar' c.- lation-transamination c" A73. The extent of ace-

tone production fro- AIB ca;.-.'..7 :. z :- the purf i' 1 -:::.13

2".rch --..1 clectrc -'_' was conchucted by ",.r.s &ra
.s.-.:' :; *

is shown in Figure 8. can be seen no acetc7-.: was r-'-

.ch'. he- Tr -7ruvate or PLP was excluded frc t'h

reaction r'.':-;:, T':o apr'.c KtX for this reaction was

fo _n-d to be 5 -"` (:-.ble 5) D.-pite the indication from

these data s that the co-factor was not t:'.,.l bound to the

eonr7e, B6 T"as deo':-c-ed in the final pro.!ation from the
s- .trum and ~C microbiological asscay ? "had not been

ac..d to the sol--.," -:-.em since gel filtration c:. G-200,

S7. eq:.ent to that step the ra':..-...1: had :- ". fracti::a-.ted

o -n -cellulose twice, precipitated :'- sulfate

and d--.l'"ed with S- o x-.C.. G-75 t`7ee

S.......,S eriec tri

Ss-- .tr uma -as t'-. : in a 3ecka.n DU s 7ct ro'hotore-

'.r 1-,0--cn 250 4. ^.50 'with the purific .--:-n at pT

7.5 0 i H r'ssi chlori.'c, 002 :i potassium phccphate

at room temperature. In addition to the :--or protein poal:

at 280 r-. there 4s a shoulder at 295 and suggestions of

bread peaks centered around 330 401 Tr 405 ,m. A pe::a

which is not .-.~etic of free co-factor has been
o--.:7c.z between 400 r:r. 430 z- with -rcl7 all protein

-bound PLP "-:..-_ (28,78) and is prec-:~7:d to rc:,ult from

a >-Irogen bct-:'d aldimine between PL? 4 c'. the (-amino of

a -.s;- resid-c.. ~.. 405 m. peak observed here suggested

this linhg.o since free co-factor has its X.~'cs3t x

at 333 nr. .-.: PL, shows a wea peak at ap-.roximately


-~<-~ 0

- T, ~~2?? ~ ~

C, I

O> 4. oi

&r ::, 0,
.0 o4 (5

,u ;, e 0

E4 < a '4 o

*)i 0 -?,

l-i p- &< y4 i

'M 0 'f)P. -*0

o *- o ?; >
F- >,.'^ (41 0

Q-1 ..;; *<-' po T0
0 0 -3 73
0 ) **- t

0 E 0 4)


O -IS 0
0' -, t

c-1 T?*- .

fl*-;A. *4)

0 0 0 .0
c3o ca

I ; 4 7) 0.

o C ) d
r- C-' r. r o

0 :;^

o) <*
o 4 o

u .



0 ,0.4'4

0 -
/Z 0- 0

O 0




o ao

10 V" Oo
- o -,,

-3 od
F 1- c

Q 3

0 t> LC t. S

a I'

337 mp at neutral pH. A peak close to 330 mR has been
observed in the spectrum of purified glutamic decarboxylase

at pH 6.5 (79) but is not generally found in PLP enzymes at

neutral pH (28). A peak at 295 mu is typical of free PLP,

but only at pH values below 5-

A detailed spectral study of the enzyme under differ-

ent conditions of pH and during reaction with different

substrates has not been undertaken, but it is warranted

since it will surely serve as a useful tool to help reveal

the nature of co-factor binding to this enzyme and the

steps of the reaction mechanism with different substrates.

i~r biol 0 oical. assay

A sample of the final preparation was appropriately

diluted- and analyzed for the presence of B6 by the yeast

bloascay of Atkin et al. (80).3 The preparation was found

to contain 1 mole of co-factor per 400,000 gn protein.

This is pr-suiably PLP although the assay does not distin-

guish between several of the B6 vitamers (81).

If one assumes a minimum of 1 mole PLP bound per mole

of enzyme and a molecular weight of 150,000 (gel filtration)

the yeast assay data imply that the final preparation is a

maximum of 40 per cent pure. On the other hand, if co-fac-

tor were bound tightly by the enzyme one would expect no

requirement for added PLP during assay. In view of the PLP

3Dr. W. B. Dempsey performed this assay.

requirement during assay, and the ultracentrifugal and

electrophoretic evidence of purity well above 40 per cent,

it seems probable that the co-factor became lost from some

of the enzyme molecules during purification. Novogrodsky

and Meister (82) found that aspartate p-decarboxylase lost

up to 80 per cent of its PLP co-factor during purification.

The Reaction Mechanism

The reaction with a-dialkyl amino acids

It appeared obvious that the reaction catalyzed by

this enzyme between an a-dialkyl amino acid and pyruvate

was a decarboxylation-transamination. However, before the

mechanism could be described accurately it was essential

that the stoichiometry be determined. One-14C- and 2-14 C

-DL-isovaline had been synthesized (see Methods) and 1-1C

-pyruvate was available, therefore, these substrates were

chosen for a study of the details of the reaction. The

predicted reaction between isovaline and pyruvate is:

DL-Isovaline + pyruvate --> carbon dioxide +
butanone + alanine

To identify and quantitate the products three sets of

reaction solutions were prepared which were identical in

that each contained, in a volume of 0.5 ml, 10 pmole DL

-isovaline, 10 mole sodium pyruvate, 0.1 mole PLP, and

125 mole potassium phosphate, pH 7.5, but which differed

in that the first set contained 1-14C-DL-isovaline, the

second 2- 14C-DL-isovaline, and the third 1-14 C-pyruvate.

All were incubated in appropriate vessels at 30 with or

without 21 yg protein added to start the reaction. At the

end of 4 hours each set was treated by a procedure which

would selectively isolate and quantitate the anticipated

radioactive product. The results are shown as Experiment

1, Table 3.

The reaction containing 1- 14C-DL-isovaline was assayed

for 14C-carbon dioxide. To positively identify the radio-

active product from the reaction with 2-14C-DL-isovaline as

butanone the entire volume of replicates of this reaction

mixture were treated to form butanone 2,4-dinitrophenyl-

hydrazone (see Methods). In this case several additional

base washes were used to try to completely remove the

2,4-dinitrophenylhydrazone of pyruvate. The carbon tetra-

chloride was evaporated and the derivative dissolved in a

small volume of ethyl acetate. Aliquots of this sample

were chromatographed on silica gel G thin layers in 5 sol-

vent systems with known butanone 2,4-dinitrophenylhydrazone

(Figure 10). On 1 chromatogram a second yellow spot ap-

peared. This compound was not checked for radioactivity

but may be pyruvate 2,4-dinitrophenylhydrazone. Another

aliquot of the ethyl acetate solution was dried onto a

planchet, counted in a gas flow counter and found to be

highly radioactive. Aliquots from others of this reaction


Stoichiometry of the Enzyme-Catalyzed Reaction
of a-Dialkyl Amino Acids with Pyruvate

Products as % of Amino
Amino Acid Substrate Exp. Acid Substratea

CO Ketone Alanine

DL-isovaline 1 47 46 42

2 53 nmb 53

3 nm 80 nm

a-aminoisobutyrate 4 nm 3 nm

a) Details of assay and methods of identification and
c: .titation of products are described in the text.
b) A--cr-e values of replicate reactions are tabulated.
nm = not measured.

-.~ -.-.
cL ;~.



'V C


, .CM

M r-4

Q4 H
to 4

* 8 0
0 z



* .
o H
.z r-C
4-P Cn^
*0 00


0 0
O o
N -
C r- 0
0 >.,


0 0

N M **


N'-.--' -,

0 to ) 4-)
*.4 4.); .

tdo p..
o4 i to4

rd P

O Co

cto ge

Qo 4- 0 V

4.) 8 Ea

'd 'd r-t 'd
,CN N (
r-1 >> <
>>r-(^ .C

0 0

00 CS

kN Id

.to 0
0 k w H

4.) P4.1- 0 4.')
0Ht>'d~ 0

to M0~
90 9 I'l Z

*r<) cSH C)

H0 tv 0 to

& .r4 4.

E-O ,C 0 *r
0 d
0 z Q)

4 ,
CO 4- W P'l
Hd r 0
IO z 0d
a o >

E-4 0 S


.e f1
^ l b



? '' \

set were tken to quantitate the product either by the

toluene extraction method or by the absorbancy of its

2,4-diitrrophen71hydrazone. The radioactive product from

2- 14C-DL-sovaline displayed a number of properties which

reasonably established that it was the predicted ketone.

Replicates from the third set of reaction vessels

which initially contained 1-14 C-pyruvate were plunged into

boiling water to destroy the enzyme and the contents as-

say~d for 1- C-alanine on Dowex 50 w-X 4 (A+) columns.

In a separate, but similar, experiment (Experiment 2,

Table 3) alanine was positively identified as the amino

acid product by quantitative thin layer chromatography

(Figure 11). In this experiment the reaction was stopped

with 0.1 ml sulfuric acid. Five Rl of the reaction mixture

was sp't' ~d on silica gel G and chromatographed beside ala-

nine and. isovaline standards. Alanine and isovaline are

readily distinguished on thin layer plates, not only by

different RF values, but by their respective rates of reac-

tion with ninhydrin. Color from alanine appeared on sprayed

plates after about 4 minutes' heating at 70, whereas that

from isovaline did not appear until after 7 to 8 minutes at

this temperature. Both isovaline and alanine were found in

the incubation containing enzyme, while only isovaline was

present in the non-enzyme control. The radioactivity which

ri---ated was at the RF of alanine and was measured. On

chromatograms of the non-enzyme control all the radio

0 cs V
m Nd
Z @a 4>

P > 0
OH 00
Q C -P 4->4Q
0 P
$4 k
o s0
-CE-' 0
u -p oq-I'd
< 0 *P0

40 0

-' H 4 I4
w ^ is (D .>

H Z 4 cs
r- 4 co a
z Ps<- 0

-H co mP z
0 O

::-C-4 ^-
< C) r

1 To

0 > Q
t 04' 1)

0 o : 4-
03 0

o l4d n
o O.S 0

O4 0

o OH

f- >d 0
P -1 Z

0 0 1

0 ) H 0
4: a a-4Q

Sadi (

E410 cs z
t rto

.^c 0 -
E-44 a~ <3




. 4








activity rem--ined at the origin. The quantity of alanine

produced as determined by chromatography equalled the amount

of carbon dioxide released in identical reactions which con-

tained 1- 14C-DL-isovaline to start.

In subsequent experiments only one of the above methods

of identification and quantitation was employed to analyze

pro..cts of a-dialkyl amino acid transamination. In a num-

ber of these experiments it was established that the amount

of b.tnonD produced was equivalent to the amount of amino

gro:p transferred to pyruvate. No detectable carbon dioxide

was produced in the absence of the a-keto acid substrate,

and in the presence of the a-keto acid substrate a-carboxyl

release &.. not exceed amino group transfer or butanone pro-

duction. T-he-cfore, decarboxylation without subsequent

transamination did not occur to any extent.

Conversion of DL-isovaline in reactions such as those

described rarely exceeded 50 per cent, and then only by an

amount within the experimental error. Under similar condi-

tions AI3 was transaminated up to 98 per cent.(Figure 8 and

ES::eriment 4, Table 3). However, when 10 mole of 2-14 C-DL

-isov-.1l-ox was incubated for 6 hours with a four-fold excess

of pyruvate to drive the transamination and a twenty five

-fold excess (500 u;) of enzyme to increase the rate of

catalysis, over 80 per cent of the radioactivity was ex-

tractable into toluene (Experiment 3, Table 3). Thus-both

enanticoers of isovaline were attacked by the purified

preparation. It is likely that one isomer reacted at a

slow rate relative to its mirror image which would. explain

the close to 50 per cent conversion with smaller amounts of

enzyme and shorter incubation time.

One-amino-cyclopentane carboxylic acid, the cyclic

analog of diethylglycine was also a substrate. This find-

inr.- was in accord with the observation that both isomers of

isov?.ine were acted upon.

The reaction with a-hydrogen amino acids

As was pointed out in the introduction, the evidence

that the mechanism of transamination proceeds sequentially,

with conversion of PLP-enzyme to PMP-enzyme by the amino

acid st:2'--.te followed by regeneration of PLP-enzyme by

the a-koto acid substrate, is conclusive. In addition, the

two half reactions are independent and fully reversible,

such that the amino acid generated from the a-keto acid is

also a sub:':-?.te.

Although the enzyme studied here differs from other

amino acid transaminases in that decarboxylation of a-di-

al!'l amino acids is required to allow formation of the

intermediLate ketimine, the presumed regeneration step in-

volving pyruvate appears exactly analogous to that of other

transaminases. To elucidate the mechanism of this second

half reaction the ability of-a-hydrogen bearing analogs of

the a-dialkyl amino acid to react with pyruvate was

meas'urd using the purified preparation. The results of

these experiments are shown in Table 4.

After 4 hours' incubation under the same conditions

described for Table 3, but with unlabelled pyruvate and

1-14C-L-alanine as substrates, reaction was stopped with

acid and carbon dioxide collected (Experiment 1, Table 4).

Significantly, only a small amount of radioactivity above

the low control values was trapped in hyamine.

When hyamine was suspended over prepared solutions of

1-- C-pyruvate, the same percentage of the radioactivity

was trapped in the organic base. Therefore, it appeared

that little if any L-alanine was enzymatically decarboxy-

lateC and that the radioactivity trapped in hyamine was

probably due to non-enzymatic decarboxylation of the 1-14C

-pyruvate in acid solution.

However, 26 per cent of the total radioactivity was

extracted into ether from the acidified reaction mixture.

Essentially no pyruvate could be extracted from neutral or

basic reaction mixtures. One- 14C-alanine was not extrac-

table at any pH. The fact that the extracted radioactive

compound was acidic and contained the a-carboxyl of alanine

led to the necessary conclusion that it was pyruvate.

When radioactive pyruvate was a substrate, 1-14C

-alanine was produced as determined by thin layer chroma-

tography and ion exchange columns. Thus within the limits

of detection the only reaction catalyzed in these


Stoichlometry of the Enzyme-Catalyzed Reaction
--'of,---r: 3Be-' ''. .'m.nino Acids *:-th fyruvate

Products as % of Amino
Anino Ac. Sub-tr?.te ExD Acid Substratea

CO- a-Keto Acid Alanine

L-al!anne 1 <0.5 26 nmb

2 nm nm 39

D-alanIne 3 3 <0 <0.5 nm

4 nm nm < 2

DL--.-amino-n-butyrate 5 <0.4 29 35

L-a-arino-n-b':.tyrTte 6 nm nm 11

D-a-c o--butyrate 7 nm nm <0.3

a') ...." of assay and methods of identification and
quan'itation of products are described in the text.
rn- = not measured.

experiments was transamination between L-alanine and pyru-

vate potentiated by loss of the a-hydrogen of L-alanine.

Minute amounts above the non-enzyme control values of

acid volatile and acid extractable radioactivity were gen-

erated in the reactions with 1-14 C-D-alanine as amino acid

substrate (Experiment 3, Table 4), and minor amounts of

14C-alanine could be detected by the ion exchange technique

when unlabelled D-alanine and 1-14C-pyruvate were substrates

(Experimental 4, Table 4). The possibility that D-alanine

may be decarboxylated at a very slow rate requires further

testing, but on the basis of the results of these experi-

ments it can be concluded that D-alanine was neither de-

carbo=lated, nor transaminated to pyruvate, at a rate

comparable to that of other substrates.

L- but not D-a-amino-n-butyrate, an a-hydrogen bearing

analog of isovaline, was transaminated to a-keto butyrate

(Experiments 5-7, Table 4). The small amount of radioac-

tivity appearing as carbon dioxide when 1- 14-DL-a-amino

-n-butyrate was used could be explained as when 1- 14C-L

-alanine was the substrate.

The results of the experiments outlined in Tables 3

and 4 indicated that the same enzyme accomplished trans-

amination of a-dialkyl amino acids by initial cleavage

of the a-carboxyl group necessarily, and transamination

of the L-isomers of a-hydrogen bearing analogs by cleavage

of the a-hydrogen preferentially. Although the final

preparation with which these experiments were performed

appeared reasonably pure, it was essential to determine

by other means whether this indication was true.

Evidence for a Single Transaminase

Relative rates of a-carboxyl and a-hydrogen labilizing
trans .mination

Certain of the apparent kinetic properties have been

determined with the enzyme for several of the substrates

(Table 5). Apparent Km values and Vmax values were ob-

tained from double reciprocal plots (83) with essentially

saturating concentrations of the second substrate and the

co-factor. Turnover numbers were calculated from V..

values assuming the final preparation to be a pure enzyme

of 150,000 molecular weight. Under identical conditions

the Vmax for L-alanine transamination with pyruvate was

2.3 times that for decarboxylation-tr.. --amination of AIB

with pyruvate. Therefore, if the a-hydrogen labilizing

enzyme was a trace contaminant of the a-dialkyl transamin-

ase, this enzyme must have a much greater turnover number

than the latter.

Several more direct approaches were taken to test

the identity question.

Relative reactivity of a-keto acid substrate

First, since the amino acid substrate presumably is

not involved mechanistically in the half reaction of


Apparent Kinetic Constants
of the Purified Enzyme

for Substrates

Substrate Apparent Km Turnover Number


Pyridoxal 5'-phosphate 0.005a
Pyruvate 2b
a-aminoisobutyrate 80 15500

L-alanine 33c 36000

pH Optimum for AIB, Pyruvate = 8.0 8.5d

a) 20 mfl pyruvate, 20 mM AIB, or L-ala, 50 mM potassium
phosphate, pH 7.5, using purified preparation.
b) 20 mM AIB, 0.1 mM PLP, 50 mM potassium phosphate,
pH 7.5, using 100,000 X & extract. .
c) 20 mM pyruvate, 0.1 mM PLP, 50 mM potassium phosphate,
pH 7.5, using product-from ion exchange (Figure 14)
with specific activity of 5.6. Turnover calculations
assume pure enzyme has specific activity of 10.0 and
molec. wt. = 150,000.
d) potassium phosphate, pH 5.7 8.4; sodium pyrophos-
phate, pH 7.9 9.5, using purified preparation.

PMP-enzyme with the a-keto acid, one would expect, if a

single enzymewere responsible, the same relative efficiency

of different a-keto acid substrates for both transamina-

tions. Using the purified preparation and identical reac-

tion conditions, but with isovaline the substrate in one

set and L-alanine the substrate in the second, this cor-

relation was observed with all a-keto acids tested (Table

6). However, it was noted that, compared with pyruvate,

the ability of all other a-keto acids to transaminate L

-alanine appeared considerably poorer than their ability

to transaminate isovaline. This is understandable in view

of the apparent irreversibility of the a-dialkyl trans-

amination but ready reversibility of L-alanine pyruvate

exchange (below). Thus, even though reactions were allowed

to proceed less than a few per cent, the fact that pyruvate,

which was generated from L-alanine, was a much more active

a-keto acid substrate than the one being tested would re-

duce the measured extent of 1-14 C-L-alanine transamination

in all assays but the one in which pyruvate itself was

being tested.

Variable activity was observed with oxaloacetate even

when freshly made solutions were used. However, since no

measurable transamination of L-aspartate with 1-14 C-pyru-

vate could be detected in another experiment, it was sus-

pected that the activity measured with oxaloacetate was


Relative Rates of Transamination of Isovaline
and L-Alanine with Several Different
a-Keto Acid Substratesa

a-Keto Substrate Relative Rate

DL-Isovalineb L-Alaninee

Pyruvate 1.00 1.00

a-keto butyrate 0.95 0.65

a-keto valerate 0.49 0.14

Glyoxalate 0.12 0.09

a-keto isocaproate 0.06 0.03

a-keto phenylpyruvate 0.00 0.00

a-keto glutarate 0.00 0.00

a) Each reaction contained 20 mM a-keto acid, 45 mM
amino acid, 0.1 mM PLP, 50 mM potassium phosphate,
pH 7.5. The product from ion exchange(Figure 14),
was the enzyme source. Relative rates were calcu-
lated from the mean of duplicate reactions.
b) Assayed by toluene extraction of 2- 14C-butanone
produced from 2- C-DL-isovaline.
Q) Assayed by ethe extraction of 1-14C-pyruvate
formed from 1- C-L-alanine.

due to pyruvate which arose from decarboxylation of the

p-keto acid, and that oxaloacetate itself was not an amino

acceptor. It was also notable that the other most common

amino acceptor, a-keto glutarate, was inactive with both

amino acid substrates.

Heat stability

As a test of similarity of protein structure the heat

stability of the a-hydrogen and a-carboxyl labilizing activ-

ities was determined with the purified preparation. An

aliquot of the concentrated protein solution was diluted

with '.1C-phos and divided into 1 ml samples. These were

heated separately in water baths at different temperatures;

35 minutes at 53 and 5 minutes at 60, 65, and 70. After
heating, each solution was rapidly cooled to below 5 and

the activity with AIB and L-alanine measured. The dena-

turation profiles for both substrates are shown in Figure

12, and coincide.

Gel filtration and ion exchange

The elution patterns of protein and of a-hydrogen and

a-carboxyl labilizing activities from Sephadex G-200 and

DEAE-cellulose columns are shown in Figures 13 and 14.

The prop.ration used for the gel filtration experiment

consisted of pooled, low activity,fractions from a previous

Sephadex G-200 column. The active fractions from the G-200

column represented in Figure 13 served as the sample for
the ion exchange column of Figure 14.

I 0




4 & Ir+
-P *P c
4*4o e-44

4) 4) 4
0 a 4
Sso asU

bo q-4H t o
0 a P A
1 4)ff

,0C 4 1) ,

OW CS 4.1 4
0 H 0
r-IP r-
*H *- OErit-
w4 0 o
o0 0
w'w U)H 0 02

04)q 10>-40
0 UJ 9 43 (+ 0

(0 0o od d to
O ^ m fia
I- S;3-4)

o cr
< d 5 6
KC d w
w 43u 0
0 r 4.3 022

wi CH 43

4.3 a >b 0
a- "q 0
00) <44o 08 2d
o. a 8o


^ t P< 3 0$

-))0 ..P 0 2 $^
,s( I ^ ,o a

ro oH ^c
a) o EBo
>0)0 4
+-J 8f
w4^ 4.-
".4 t> i


o.~:~- K


( ~ -





ac I


x 0


3AZN3 3AV73i
3WdkZN3 3AILV'13H


4)O0 t
o 4 0 0

S go o
K I ,-

0 goo
0 034-4 03
I 4-
c 0 4 o

o o oa4-
I .0 0- 00
0 N. 0


)d EA

S co4.)
1 E
z 3 H 0

0 H0 0

H -4' csG



3 CS


CS 4.)4
b5) C+


\ r-I 0 43>
< -J-

0 o00
d *i 0
> 4 )
>\0 0 0

; 0
\ ra *r-
X^ C \0
^ 0\ W4
ca~~~ 4-2 \.r
^ s \ ** HO
'rX 4 r3r-


x I \ '-1O >
,' | *P~r4

The relative activity of the two transaminations was

the same in all fractions within the error limits of the

assay used. It had been considered likely that Ps-1 would

contain more than one enzyme capable of alanine pyruvate

transamination, even if the a-dialkyl transamination,acted

also on L-alanine. The presence of any of this other enzyme

in AIB transaminating fractions would have led to ambiguous
results. One- .C-DL-a-amino-n-butyrate and a-keto butyrate

were used as the a-hydrogen amino acid and a-keto acid sub-

strates, hoping that L-a-amino-n-butyrate would not be a

good substrate for such a contaminating transaminase.

However, the assay for a-amino-n-butyrate transamination

involved elution and quantitation of 14C-a-keto butyrate

from a Dowed 50 (H+) column, a procedure which, because of

dilution was inaccurate for rate studies when low activi-

ties were present. The more accurately assayed alanine

pyruvate transamination was employed to measure a-hydrogen

labilizing activity eluted from the DEAE-cellulose column,

and closer correlation of the two activities was observed.

More than one a-hydrogen labilizing transaminase did not

appear to be present in the starting sample. The conclusion

from these data was that the enzyme responsible for both

activities had similar size and charge characteristics.

Co-Induction of a-hydrogen and a-carboxyl labilizing

Both a-hydrogen and a-carboxyl labilizing activities

were present in extracts of cells induced with either DL

-i4ovaline or DL-a-amino-n-butyrate in the absence of other

nitrogen sources, while both activities were greatly re-

duced in cells grown with ammonia as nitrogen source. How-

ever, the ratio of a-hydrogen to a-carboxyl labilizing ac-

tivity was greater in extracts of the Dl-a-amino-n-butyrate

grown cells. The possibility of such a finding was antici-

pated since a second a-hydrogen labilizing transaminase

with somewhat different specificity might be induced by

the a-hydrogen bearing amino acid. It was for this reason

that the better substrate, but more common amino acid, L

-alanine; had not been used as the a-hydrogen bearing


T-e results of all the experiments described in this

section imply that one enzyme was responsible for both

types of transamination. There was no evidence contrary

to this interpretation and it was therefore concluded that

the purified preparation contained a PLP-dependent trans-

aminase which proceeded by different mechanistic steps with

a-dialkyl amino acids than with a-hydrogen bearing amino


Other Details of the Reaction Mechanism

Stereochemistry of the amino acid product

Before the reaction mechanism of this transaminase

could be accurately described it was essential to establish

the stereochemistry of the amino acid product, although the

L isomer was expected. This question was answered by a

direct and an indirect test.

First, the labelled alanine produced upon incubation

of AIB and 1- 4C-pyruvate with enzyme was isolated from a

Dowex 50 (H+) column, concentrated on a steam bath and used

as a substrate for the enzyme, which had been shown to be

specific for the L isomer. However, less than the calcu-

latec amount of labelled alanine exchanged with pyruvate

in this second incubation. The result allowed the con-

clusion that L-alanine was a product, but did not allow

the exclusion of D-alanine as a product, also. Rather

than pursue this type of proof, a second means was employed

to test whether only the L isomer was a product.

If both the substrate and product amino acids had the

same configuration, the rate of an exchange reaction between

alanine and pyruvate would remain constant throughout a

given Incubation; for, as each alanine molecule was con-

verted to pyruvate, a pyruvate would be converted to ala-

nine of the substrate form, so that the concentrations of

all substrates would remain constant. In addition, as was

pointed out by Jenkins and Sizer (48), if one of these

species was labelled at the outset of exchange, a first

order reversible equilibration of isotope would occur,

such that at equilibrium the specific activity of both

substrates would be the same, provided they were both

present in the same initial concentration. Jenkins and

Sizer showed further that a semi-log plot versus time of

the decrease of radioactivity to the equilibrium value

had a slope which was proportional to the true rate of

the exchange reaction. However, the slope of this plot

would be constant only if the product amino acid was an

equally active substrate.

The exchange reaction between equal amounts of 14C

-L-alanine and pyruvate catalyzed by a-dialkyl amino acid

transaminase is shown in Figure 15. Figure 15 A shows that

at equilibrium 50 per cent of the radioactivity exists as

pyruvate. A semilog plot of the loss of radioactivity from

alanine as equilibrium was approached versus time (Figure

15 B) was linear and therefore established that the product

amino acid was the predicted L-alanine.

The irreversibility of the a-carboxyl labilizing half

Reversibility of the half reaction involving cleavage

of the a-carboxyl of the a-dialkyl amino acids was tested

by attempting to measure exchange between isovaline, 14C

-carbon dioxide and butanone. Replicate reaction mixtures




co ,A

2 0I 2d -



^ r
<. -*--* --



4 MP
0 ~0P
S 4

0 o

to ,- 0Vo

0 0
C c
(D 4-4
E 0

(2 0 .
c0 0


31VAnfAd SV WdO %




were prepared which contained in 1.0 ml, 20 mole each of

DL-isovaline, sodium. 14C-bicarbonate (1.5 X 106 pm), and

butanone; 0.5 Pmole PLP, and 50 pmole potassium phosphate,

pH 7.5. At the end of 6 hours' incubation with 270 pg en-

zyme, carbon dioxide was liberated with sulfuric acid and

trapped in hyamine. An aliquot of the carbon dioxide free

reaction mixture was counted to detect any 14C which may

have become incorporated into isovaline during the incuba-


In a second experiment carbon dioxide exchange with

the co-factor bound intermediate (i.e. reversibility of

the pres-ued aldimine to ketimine, a-carboxyl cleavage

step) was tested analogously except that butanone was ex-

cluded from these reaction mixtures which however, con-

tained 0.2 mole pyruvate per ml. Pyruvate was included

to reconvert to PLP-enzyme, the PMP-enzyme which would be

forncd and be incapable of the exchange reaction in ques-


In both experiments the radioactivity remaining in

the reaction solutions after liberation of carbon dioxide

was equal to background indicating no detectable fixation

of 14C.


The Reaction Mechanism

The overall reaction mechanism for the a-dialkyl amino

acid transaminase, constructed from the evidence presented

in the preceding section is summarized in Figure 16. Ac-

cepted structural notations for major intermediates pre-

sumed to exist in PLP-catalyzed decarboxylation and trans-

aminatlon reactions have been used to illustrate the mech-

anisn. It is written as the reaction of an a-dialkyl amino

acid wi-'h an a-keto acid.

Clea-.:e of the a-carboxyl group of the aldimine (I)

leads to formation of the ketimine (II) which becomes pro-

tonated at the formyl carbon to produce the less conjugated

ketimine (III). This intermediate is hydrolyzed to produce

a ketone and PMP-enzyme. The mechanism of the first half

reaction appears to be the enzymatic analog of the non-enzy-

matic decarboxylation-transamination shown in Figure 2.

The second half reaction involving PMP-enzyme and the

a-keto acid is identical with other amino acid transamina-

tion mechanisms (Figure 3), in that it is fully reversible.

As such it presumably includes reversible tautomerization

of the ketimine (IV) to the aldimine (VI) via the fully
conjugated intermediate (V).



/ =
a *:: 2 0

03= Z

Sa: o




(CII 9 < T

6 -z
a I


o a:
0 \ (.-3
0- -z +4-


a: 2



I + U=o

0 0

1 o 4 2

' C0

0 0

T 4


S, 0

o 0

0 ,,0
o -



\ 2
4 zo-

The enzyme did not catalyze detectable exchange of the

a-carboxyl of isovaline with 14C-bicarbonate or exchange

between isovaline, butanone and 14C-bicarbonate. The re-

versibility of the hydrolysis of intermediate III was not

determined. Although the equilibrium constant for the over-

all reaction of an a-dialkyl amino acid transamination with

pyruvate was not measured, it is evident from the extent of

reaction with AIB (Figure 8) the equilibrium strongly favors


The unique feature, which distinguishes this enzyme

from other PLP-dependent enzymes that have been described

is its ability to transaminate amino acids with a tertiary

a-carbon by cleavage of the a-carboxyl group. However, the

most significant feature of the mechanism in terms of its

relationship to our understanding of PLP-dependent enzyme

mechanisms and reaction specificity would seem to be the

ability to potentiate transamination of a-hydrogen bearing

analog of these amino acids by cleavage of the a-hydrogen

in preference to the a-carboxyl. Because such a property

could be significant to the construction of concepts about

the factors controlling reaction specificity of this, and

possibly other B6 enzymes, it is important to consider the

justification of this conclusion from the experiments con-

ducted to test the question.

The experiments that were performed were designed to

reveal physical or chemical differences between the proteins

or catalytic sites involved in the a-hydrogen and a-carboxyl

labilizing transaminations. The results of these experi-

ments indicated that the enzyme which transaminated a-di-

alkyl amino acids had the following properties in common

with that responsible for alanine-pyruvate exchange:

1) simultaneous induction by the same compounds, with

either substrate configuration about the a-carbon;

2) the same size, charge, and denaturation properties

as measured by gel filtration, ion exchange, and

heat stability;

3) the same degree of co-factor requirement during

assay; and

4) the same relative activity with a variety of a

-keto acid substrates and no activity with a-keto

glutarate and probably oxaloacetate, the two most

common amino acceptors.

If two enzymes were involved one would be required to ex-

plain why a transaminase which was -able to tautomerize

the pyruvate-PLP ketimine intermediate to the aldimine (as

in the formation of alanine during regeneration of the PLP

form of the a-dialkyl transaminase) was unable to catalyze

the reverse tautomerization to form the same ketimine. The

evidence seemed overwhelming in favor of the conclusion

that a single enzyme was responsible for both reactions.

The Dependence upon Added PLP
for Transaminase Activity

The apparent KPLP for a-dialkyl amino acid transaminase

was approximately 5 P-, yet the final preparation contained

B6 even after the enzyme had undergone several procedures

during purification which should have removed free PLP or

a weakly bound co-enzyme. It is well understood that Km

can not be interrupted as an indication of binding, but if

there is independent evidence, as in this case, that the

co-factor remained bound to its active site, one can expect

a low Km or even no requirement for added co-factor during


There are at least two possible explanations for the

findings.with this enzyme. First, since 50 gM PLP was

added during the early purification steps, the co-factor

presu.bly bound to protein in the final preparation may

have become attached, during purification, to free amino

groups on the protein (84) and have, in fact, no connection

with the co-factor requirement for catalysis. Thus the

co-enzyme binding site could be essentially empty in the

finnl preparation. The weak peak in the spectrum of the

purified enzyme (Figure 9) about 405 ma is suggestive of

PLP present as an aldimine.

On the other hand, the requirement for added PLP dur-

ing catalysis could reflect the difference in binding prop-

erties of PLP and PMP since the latter is formed during the

transamination reaction. In several experiments with the

a-dialkyl transaminase in which enzyme was used in great

excess (270 Pg) of the normal concentration, it was noted,

incidentally, that after long incubation times the bright

yellow color of the added PLP (500 0M) became bleached

implying possible conversion of most of the PLP originally

present to PMP, free in solution. Such a loss of PMP from

the active site after transamination is an extreme example

of the type of inactivation observed of aspartate p-decar-

boxylase (82) when PLP or an a-keto acid were not present

during assay.

It is predicted then, that PMP will fulfill the

co-factor requirement for the a-dialkyl amino acid trans-

aminase but will have a much higher Km than PLP. One-half

the maximum activity possible with 50 uM PLP was observed

in an assay in which 200 1M PMP was used. However, it was

discovered by other workers in this laboratory that the

PMP preparation used contained 4 per cent PLP as a con-

taminant; this made the solution 8 uM in PLP, a concen-

tration sufficient to account for the observed activity.

Further experiments with pure PMP will be required to test

its ability to serve as co-factor. PAL at 200 uM was


The Natural Substrate

The findings in this study imply that, in nature, this

transaminase could have as its usual substrates c-dialkyl

amino acids and pyruvate, and functions to assimilate the

nitrogen of these or similar "unnatural" amino acids into

the mainstream of intermediatry metabolism. It has been

reported that AIB is found in the protein of horse muscle

(85), and as a component of an antibiotic (86). Confirma-

tion of either of these reports could justify the existence

of a bacterial enzyme designed specifically to degrade

these compounds.

Further, 1-aminocyclopentane carboxylic acid which,

in space, has a structure closely resembling both D and

L-isovaline, is a substrate for the enzyme. Reasoning

from this, it should be the case that 1-aminocyclopropane

carboxylic acid, the cyclic analog of AIB, and a natural

component of some fruits (87,88), should also be a sub-


Inducibility by a-hydrogen amino acids implies a use

in the cell of the enzyme for their transamination. Yet

L-alanine exchange with pyruvate was the only reasonably

rapid reaction observed with the a-hydrogen substrates

and the apparent Km for alanine was much greater than

for AIB (Table 5).

Conversely the apparent Km for pyruvate was lowest

of all substrates tested. The specificity indicated by

the several amino and c-keto acids tested, appears to be

limited to compounds with short non-polar a-R groups.

The rate of transamination decreased considerably when

the side chain length of the a-keto acid substrate was

increased from two to three carbons as in the shift from

a-keto butyrate to a-keto valerate (Table 6). The rate

decreased more sharply when the alkyl group was removed

altogether, as in the shift from pyruvate to glyoxalate.

The aromatic side chain of phenyl pyruvate and the terminal

carboxyl of a-keto glutarate prevented activity with these

compounds. Likewise, L-aspartate was not an amino acid

substrate. It should be a safe assumption, in view of the

evidence for reversibility of this and all other amino acid

transaminases which have a-hydrogen amino acid substrates,

that the amino acid corresponding to a particular a-keto

acid will show reactivity proportional to that shown by

the a-keto acid and vice versa. The only class of common

amino or a-keto acids not tested were those with positively

charged side chains. Unless, one of these is a good sub-

strate, the utility of this enzyme as a strictly a-hydrogen

amino acid transaminase is questionable.

Control of Synthesis of the c-Dialkyl
Amino Acid Degrading System in Ps-1

a-Dialkyl amino acid transaminase was induced in Ps-1

by DL-isovaline, AIB and 1-aminocyclopentane carboxylic

acid, as well as the a-hydrogen bearing analog of isovaline,

DL-a-amino-n-butyrate. Induction by alanine was not mea-

sured. There was no evidence of coordinate or sequential

induction of a pathway for the complete degradation of

a-dialkyl amino acids since the organism was unable to

utilize the ketone products for carbon. However, the en-

zyme exhibited some physical properties (e.g., sedimenta-

tion coefficient, heat stability) in common with the enzyme

studied by Aaslestad and Larson (11), which came from a

bacterium that used AIB for carbon. Both organisms were

tentatively assigned to the genus Pseudonomas. It could

be that the bacterium used in this laboratory was closely

related to their organism but had lost, through mutation,

the ability to degrade the ketones. Or possibly the nat-

ural "unnatural" amino acid substrate has not been dis-

covered but is one which could be degraded completely by

this organism.

The control of this enzyme may be of interest because

unlike most inducible, degradative enzymes, it has two

substrates, the amino and the a-keto acid. It is assumed

that induction requires only the amino acid substrate;

however, a simultaneous need for an a-keto acid for

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EQ81D5VIG_DAKJ02 INGEST_TIME 2012-02-29T19:18:48Z PACKAGE AA00009133_00001