FACTORS CONTROLLING PRODUCT
STEREOSPECIFICITY IN THE REDUCTION
OF CARBONYL COMPOUNDS WITH
ALCOHOL DEHYDROGENASE AND
REDUCED NICOTINAMIDE ADENINE
DINUCLEOTIDE

Thesis for the Degree of Ph. D.
MICHIGAN STATE UNIVERSITY
VASSILIOS CHRISTOS STAMOUDIS
1973

 
      
  

  
 
  

IIIIIIIIIIIIII

300

IIIIII I

46¢.9 312

"’ Lia .4RY

Michig (1 State
Univ ersity

III

S . I ' This is to certify that the
‘ ' thesis entitled
‘ CONTROLLING PRODUCT STEREOSPECIFICITY IN THE REDUCTION

OF CARBONYL COMPOUNDS WITH ALCOHOL DEHYDROGENASE AND
REDUCED NICOTINAMIDE ADENINE DINUCLEOTIDE

 

presented by

Vassilios Christos Stamoudis

has been accepted towards fulfillment
of the requirements for

M degree in MST/17

i {fkmegflé II

 

Major professor :
i _ I
m + B); - ~.
A I

0-7639

  

RIDING IY

NO“ I: SDII S'
BOOK IIIIIIEIIY IIIB.

Hm IINDEIS

 
 

  
 

”I III." Ilflllfll

 

137*

 

“A

triage-«RY f
blithigan State I
University

      
   

ABSTRACT

FACTORS
CONTROLLING PRODUCT STEREOSPECIFICITY IN THE REDUCTION
OF CARBONYL COMPOUNDS NITH ALCOHOL DEHYDROGENASE AND
REDUCED NICOTINAMIDE ADENINE DINUCLEOTIDE.

By
Vassilios Christos Stamoudis

(Baoixns Xpionou Etayoédns)

Although phytochemical and yeast reductions1 of a large variety of
carbonyl compounds have been known for many years, complete product
stereospecificity studies, especially with purified enzymes and co-
enzymes, are few. Karabatsos2 suggested a simple Model K (L and 5,
large and small groups) indicative of the spatial relationship between
coenzyme and substrate at the transition state in the reduction of car-
bonyl compounds with A-type dehydrogenases and NADH. Model K fits
the data for all aldehydes and those ketones which do not contain polar

groups. This model is the opposite of the one Prelog3

suggested for
A-type dehydrogenases, indicating that his assumption of steric co-
enzyme-substrate interactions is incorrect. It is known that hydroxy-

4

acetone and pyruvic acid5 have the opposite substrate-coenzyme spatial

relationships (g and g, respectively) of that of D and L lactaldehyde6
(1). To further evaluate the steric substrate-enzyme interactions 3;,
the hydrophilic-hydrophobic substrate-enzyme interactions at the active
site, we carried out reductions of methyl ethyl ketone and chloro-
acetone, substrates very similar in size to both E and g, with L-ADH
and NADH. In doing this, we established simple reaction conditions

and methods for isolating the product alcohols. The stereospecificity

of horse liver alcohol dehydrogenase, (1.1.1.1), (L-ADH) towards methyl

Vassilios C. Stamoudis

 

H81,” HA
2 0 CONHZ
CH R"
3
\
I
R
Model K I (100%) g, R'=-CH20H (100%) g, "=-CH2CH3(72%)

for A-type enz.

Q: R'=-COOH (100%) Q, R"=-CH2CT (53.4%)

ethyl ketone (é) yielding as the major product (+)-(S)-2-butanol (72.0:,8%)
proved lower (44.0 :_l.6%) than the stereospecificity reported5
for yeast alcohol dehydrogenase (1.1.1.1), (Y-ADH). This finding is

7 that the "steric

in accord with Model K and supports the suggestion
hindrance" provided by the two enzymes at the active site is different,
being greater for Y-ADH. The stereospecificity of L-ADH toward chloro-
acetone (g), yielding as the major product (+)-(S)-1-chloro-2-propanol
(53.4%), was found to be only 6.8 :_.4%, surprisingly low for an enzymatic
reaction involving a substrate with two different substituents.

From the fact that all aldehydes and those ketones in which the
substituents are only alkyl and phenyl groups follow Model K, we con-
cluded that for these substrates the steric factors seem to be predominant
and Model K holds. Further, based on configurations l, g, g, 1’ and i
we postulate that for bifunctional substrates the hydrophilic substituent
of the carbonyl compound tends to occupy the location of the small group
in Model K, with the hydrophobic group taking the place of the large group.
Thus, hydrophilic-hydrophobic substrate-enzyme interactions may, as in
g and Q, invert the substrate-coenzyme spatial relationship (predicted

by Model K). However, in the case of l the steric factor is so big

Vassilios C. Stamoudis

(H gs. -8:-CH3) that the hydrophilic-hydrophobic interactions are not
sufficient to invert the predicted relationship. In the case of chloro-
acetone (3) the hydrophilicity of the chloromethyl group, compared to
that of hydroxymethyl or carboxyl, appears to be insufficient to invert
the spatial relationship, but its influence leads to an almost racemic

product.

REFERENCES

 

—l
O

C. Neuberg, Adv. Carbohyd. Chem. 4, 75 (l949).

 

2. G. J. Karabatsos, J. S. Fleming, N. Hsi, and R. H. Abeles, J. Amer.
Chem. Soc., 66, 849 (1966).
3. V. Prelog, Pure Appl.Chem., 9(1), llg (1964).

 

4. G. J. Karabatsos and H. Nunez, unpublished results (1971).
5. J. Van Eys and N. 0. Kaplan, J. Amer. Chem. Soc., 12, 2782 (1957).

 

6. B. Zagalak, P. A. Frey, G. J. Karabatsos, and R. H. Abeles, J. Biol.
Chem,, £41, 3028 (1966).

7. For a discussion see: R. Bentley, "Molecular Asymmetry in Biology",
Vol. 11, Academic Press, New York, I970, pp. 20-22.

FACTORS
CONTROLLING PRODUCT STEREOSPECIFICITY IN THE REDUCTION
OF CARBONYL COMPOUNDS HITH ALCOHOL DEHYDROGENASE AND
REDUCED NICOTINAMIDE ADENINE DINUCLEOTIDE

By

Vassilios Christos Stamoudis
(BooiAns Xpi'omu Ztauodons)

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Chemistry
1973

/ .

." f) k? f

'9'».

p
c"‘

.‘Wh

To

My Mother Korepivn (Katherine)
and

My Wife A{a (Lea)

ii

AKNOHLEDGMENTS

 

I wish to express my sincere appreciation to my academic advisor,
Professor Gerasimos J. Karabatsos, for his guidance and encouragement
throughout this investigation and Professor H.Reusch for helpful dis-

cussions while serving as second reader of this thesis.

I wish to thank Professor J. Speck for the availability of the
facilities in his laboratory and my wife Lea for her patience, under-

standing and helpful discussions throughout this research.

The helpful discussions with Dr. Hernan Nunez and his friendship

are greatly appreciated.

I also wish to express my appreciation to the Department of Chemistry
of Michigan State University for providing the opportunity to gain
teaching experience and the National Science Foundation for partial

financial assistance.

The technical assistance provided by the undergraduates Randy Lievertz,
Sam Speck and especially Doug Beyer and David Abbott is also greatly

appreciated.

Finally, I wish to thank all my undergraduate students in either labs
or recitations, especially the classes of CEM 132, Spring 71; CEM 241,
Summer 71; CEM 242, Summer 72; and all six classes of CEM 351, 352 and 353
for the academic years 7l-72 and 72-73 because their real interest toward

learning, overwhelmingly demonstrated by their massive and active parti-
cipation in open help sessions, made the period of this investigation much

more meaningful, interesting, and enjoyable.

iii

TABLE OF CONTENTS

ACKNOWLEDGMENTS ........................
LIST OF TABLES ........................

LIST OF FIGURES ........................

INTRODUCTION .........................

EXPERIMENTAL .........................

II.

Equipment ........................

Materials ........................
Treatment of Data ....................
Reaction of methyl ethyl ketone with L-ADH and NADH. . . .
a. Procedure ......................

b. Separation of product and excess substrate from
enzyme and coenzyme ..................

c. Isolation and detection of product 2—butanol .....
d. Optical activity measurements .............

e. Determination of the concentration (X%) of product
2-butanol in the ethanolic solution ..........

f. Resolution of 2-butanol ................
The reaction of chloroacetone with L-ADH and NADH .....
a. Procedure .......................

b. Isolation and detection of the product l-chloro-
Z-propanol ......................

c. Optical activity measurements .............

d. Determination of the concentration of the (X%) solution
of product l-chloro-Z-propanol in solution ......

iv

Page
iii
vi

vii

16
16
17
18
18

21
21
22

25
27
32
32

32
35

37

TABLE OF CONTENTS (Continued)

e. Determination of the enantiomeric purity of product
l-chloro-Z-propanol by NMR with a chiral Europium

shift (optishift) reagent ...............

RESULTS AND DISCUSSION ....................

A. Enantiomeric purity of product 2-butanol .........
B. Enantiomeric purity of product l-chloro-Z-propanol

1) From specific rotation ................

2) From NMR data with Eu(hfpc)3 optishift reagent . . . .

C. General comments about reactions 1 and 2 .........

D. Stereospecificity .....................

1) Methyl ethyl ketone ..................

2) Chloroacetone .....................

E. Steric enzyme-substrate interactions ...........

F. Hydrophilic-hydrophobic substrate-enzyme interactions. . .

G. Steric gs, hydrophilic-hydrophobic interactions ......

CONCLUSION ..........................

BIBLIOGRAPHY .........................

Page

39

46
46
46
46
47
48
50
50
52
53
57
59

60
61

LIST OF TABLES
TABLE PAGE

1 Changes in pH and A340 with Time for the Reaction of
Methyl Ethyl Ketone with L-ADH and NADH ......... l9

2 GC analysis of the fractions collected with respect to
the content of 2-butanol ................ 22

3 Optical rotations of product 2-butanol, (X%) solution,
in absolute ethanol ................... 23

4 Optical rotations of product 2-butanol solution in
absolute ethanol .................... 23

5 Relation between GC peak height and 2-butanol con-
centration ....................... 25

6 Data summarizing the recrystallizations for the purifi-
cation of (S)-2-butyl brucine phthalate ......... 29

7 Optical rotations for a 6.38% solution of resolved
(+)-(S)-2-butanol in absolute ethanol .......... 3O

8 Changes in pH and A340 with time for the reaction of
Chloroacetone with L-ADH and NADH ............ 33

9 Optical rotations of product l-chloro-Z-propanol (X%)
solution in chloroform ................. 35

10 Summary of data to determine the relationship between
l-chloro-Z-propanol peak heights in GC and % concentration 37

11 Average weights (104g) of the cut-outs of the four peaks
of the methyl doublets for two product l-chloro-Z-propanol
spectra (A and B) and one commercial compound spectrum (C).
The sweep width of the spectra was 50 Hz ........ 4O

12 Reduction of carbonyl compounds by purified enzyme and
coenzyme (NADH) ..................... 54

vi

LIST OF TABLES (Continued)

TABLE PAGE

13 Carbonyl compound reductions by actively fermenting
yeast .......................... 55

14' Reductions of carbonyl compounds possessing at least
one polar group by purified enzyme and coenzyme (NADH). . 57

vii

FIGURE

10

11

12

LIST OF FIGURES

PAGE
The structure of pyridine nucleotides ......... 2
Absorbance in 340 nm (A340) vs, time for the reaction
of methyl ethyl ketone with L-ADH and NADH ...... 20
Plot of optical rotation gs, wavelength for product
2-butanol in absolute ethanol ............. 24
Plot of peak height vs, % concentration of 2-butanol
in absolute ethanol .................. 26
Plot of optical rotation vs, wavelength for a 6.38%
solution of resolved (+)-(S)-2-butanol in absolute
ethanol ........................ 31
Absorbance at 340 nm (A340) vs, time for the reaction
of Chloroacetone with L-ADH and NADH ......... 34
Plot of optical rotation vs, wavelength for product
l-chloro-Z-propanol in chloroform ........... 36
Plot of peak height vs, % concentration of l-chloro—
2-propanol in chloroform ............... 38
The NMR spectra of product l-chloro-Z-propanol in
chloroform before (P0) and after (P1) the addition
of 50 mg Eu(hfpc)3 .................. 41
The NMR spectra of commercial 1-chloro-2-propanol in
chloroform before (C0) and‘after (C) the addition of
50 mg Eu(thC)3 .................... 42

The NMR spectrum of product l-chloro-Z-propanol in
chloroform after the addition of 100 mg Eu(hfpc)3. . . 43

Magnified methyl doublets (sweep width 50 Hz) of two
different runs of the NMR spectrum of product l-chloro-
2-propanol after the addition of 100 mg Eu(hfpc)3. . . 44

viii

LIST OF FIGURES (Continued)
FIGURE PAGE

13 Magnified methyl doublets (sweep width 50 Hz) of the
NMR spectrum of commercial l-chloro-Z-propanol after
the addition of 50 mg Eu(hfpc)3 ............. 45

G9

14 Absorption spectra of NAD and NADH ........... 49

ix

INTRODUCTION

Stereospecificity in enzymatic hydrogen transfer reactions has
received considerable attention for many years. Availability of
pure enzymes, coenzymes, coenzyme analogs and isotopically labelled
substrates and coenzymes has made the study of such enzymatic reactions
easier. Dehydrogenases, which utilize pyridine nucleotides as coenzymes
and zinc metal as cofactor, are an important class of enzymes catalyzing
oxidation-reduction reactions involving the transfer of hydrogen. The
remarkable work1 of Harburg and Cristian (1931-36), and of Komberg (1950)
led to the complete structure elucidation of the pyridine nucleotides
which are summarized in Figure 1. Their structures were confirmed by
synthesis in Todd's laboratories in 1957.

In the past, most of the enzymatic oxidation reduction reactions
involving substrate alcohols and carbonyl compounds were done by using
fermenting yeast preparations. The only carbonyl reduction studied
stereochemically with purified NADH and yeast alcohol dehydrogenase
was that of l-deuteroacetaldehyde done by Hestheimer's group2 in 1951.
Vennesland and Nestheimer~ showed that in a dehydrogenase
catalyzed reaction the hydrogen atom from the metabolite is transferred
directly to the 4-position of the nicotinamide ring, without mixing
with the hydrogen ions in the solution. They also demonstrated that
the NADH reduction of acetaldehyde with yeast alcohol dehydrogenase
(Y-ADH) is stereospecific in both substrate and coenzyme.12eaction(1)

 

 

H
CONH2
GI
I
R
:leG NADH
(Nicotinamide Adenine Dinucleotide) (Reduced NAD)
or DPIIB or DPNH
(Diphosphopyridine Nucleotide) (Reduced DPN)
I \ ONHZ
63/
I
RI
NADPe NADPH
(Nicotinamide Adenine Dinucleotide Phosphate) (Reduced NADP)
or TPIP or TPNH
(Triphosphopyridine Nucleotide) (Reduced TPN)

NH,
N
c) o / N
II II (\ICS)
o Ha—o—r-0—-r—o—cn. o N N

00 09
H
OH on on ox

o N
R-. if x = H; R-. if x = -P-OH
0H

Figure 1. The Structure of Pyridine Nucleotides.

may be stated as follows by using Hanson's3

4

nomenclature and having

and the established absolute con-

figuration of ethanol-l-d by Lemieuxs. "The pro-R (or A) hydrogen

in mind the work of Levy et. al.

atom at C-4 of NADH is transferred to the re-face of acetaldehyde-l-d".

 

 

H
,CONH D
Y-ADH “\~ 2 \\ CH (1)
/CH + \ C/ 3
D\C 3 A-type enzyme (3/ H“ \OH
I
R

(-)-(S)-Ethan01-1-d

From the absolute configuration6

at C-4 of the pyridine nucleus of
NADD, it is now known that both yeast and liver alcohol dehydrogenases
cause transfer of the pro-R (I) or A(II) proton, whereas the a(xial)
and e(quatorial) (cyclic) ketone reductases from Curvularia falcata
and pig liver, respectively, cause transfer of the pro-S or B proton;
the former are called A-type enzymes, the latter B-type enzymes. A

tabulation of A-type and B-type dehydrogenases is given in Reference 7.

 

(I) (II)

So far, we have discussed the coenzyme stereospecificity, which
is a priori expected in view of the fact that the methylene protons
in NADH (I or II) are diastereotopic* (the two faces in NADIa (III) are

 

NADe (III)

diastereotopic as well) and, hence, they are distinguished easily
by the enzyme. Let us now discuss the problem of substrate stereo-
specificity.

Although it was known for a long time that enzymes were chiral
reagents, their ability to distinguish between the two a ligands in

8

a system Caabg was clearly understood only after Ogston's postulate.

2 classic example already discussed here demonstrates

Hestheimer's
this clearly in the case of ethanol-acetaldehyde in which the two

methylene protons of the alcohol (pro-R, pro-S) or the two faces

*
By symmetry, two atoms, groups or faces (sides) are diastereo-

topic if they are not interchanged by any symmetry operation. By the
substitution criterium two atoms or groups are diastereotopic if iso-
topic substitution yields diastereomers and two faces are diastereotopic
if the addition of an achiral reagent across the two faces yields

diastereomers.

(re and si)3 of acetaldehyde are enantiotopic* and, hence, distinguish-
able by the enzyme (Y-ADH). This is a classic example of asymmetric
induction and may be depicted as follows (with the involvement of the

coenzyme as well). The two transition states I and II are diastereomeric

 

 

 

r ‘
D
”M
H,”’c/D Coenz 5
re face I \or H"'Z'\\OH
MC: 0 NADD CH3 1 , CH3 (R)
Y-ADH , 0“
fl
C 3 si face H’mC/O [W
I \ ’ .c\
CH Dvaoenz H124 0(5)
3 II a 3

sole product
and, hence, they have different energy. This thesis is a contribution
to continuous efforts to a) understand the spatial relationship of
coenzyme-substrate in the transition state in the oxidation-reduction
reactions involving pyridine nucleotides, L-ADH or other dehydrogenases
and a variety of aldehydes and ketones and b) devise a simple model

which would be consonant with the facts.

*Two atoms, groups or faces (sides) are enantiotopic (prochiral) if the
environment of the one is the mirror image of the environment of
the other. By symmetry, two atoms, groups or faces are enantiotoPic
if they are interchanged by an Sn axis of symmetry. By the sub-
stitution criterium two atoms or groups are enantiotopic if isotopic
substitution yields enantiomers and two faces are enantiotopic if the

addition of an achiral reagent across the two faces yields enantiomers.

9’10 studied the enzymatic reductions with

Prelog and his group
dehydrogenases (a-oxidoreductases) isolated from Curvularia falcata
(B-type enzymes) of a large variety of cyclohexanones and decalones
by determining the absolute configuration of the alcohols produced,

which was as shown.

5 L L
‘\C

1’
H‘ ‘0” 5 small group

large group

I

‘\

Prelog postulated the following Model P indicative of the relative
spatial relationship between coenzyme and substrate at the transition

state, when B-type dehydrogenases are used.

 

 

~—€P a
enzyme \\‘
HB OH
Model P (S)-configuration

Prelog justified his model as follows:

1) Steric coenzyme-substrate interactions are smaller if
the carbonyl group points toward the pyridine ring nitrogen11
of NAD (or NADP) and the large group L is over the hydrogen
whereas the small group 5 is over the carbamido group. Prelog
predicted that A-type enzymes should give products of the

opposite configuration, as Model II would apply.

   

Model P for B-type enzymes Model II for A-type enzymes

2) The position of groups L and s are further controlled by
their interactions with hydrophilic and hydrophobic (lipo-
philic) regions in the enzyme. During the reaction, then,
the most lipophilic group will be in the hydrophilic region.

12 and Kosowern as

This had been previously suggested by Hestheimer
well. (For a thorough discussion of Prelog's model and the "diamond
lattice section" postulate for L-ADH see references 7, 9, 10, 13, 14).
The prediction by Prelog concerning the absolute configuration of
products from reactions involving A-type enzymes should be incor-
rect. The reduction of acetaldehyde-l-d with yeast alcohol dehydro-
genase (Y-ADH) and L-ADH (A-type enzymes) produces (-)-(S)-Ethanol-l-d

15

which fits Model K suggested by Karabatsos for A-type enzymes.

   

C
ON“? Y-ADH D\ /CH3
C" c + NAD‘F
3 or L-A H ‘0‘ ‘
H OH
III (-)-(S)—Ethanol-l-d

Karabatsos' Model K is exactly the opposite of that of Prelog's

Model (II) for A-type enzymes. The following findings support Model

16 69

K further. Van Eys and Kaplan reported that Y-ADH and NAD react

with (S)-2-octanol but not with (R)-2-octanol. This fits Model K as

shown in IV.

Y-ADH 3 CH3\ /C6Hl3 Q
(____ CC + NAD

Hy F011

 

IV (S)-2-octanol

21

Gunther et. al. have recently demonstrated, in a very clever enzy-

matic preparation of the (R) and (S)-enantiomers of l-propanol-l-d,
that Y-ADH in the presence of HANS, NADH and diaphorase, all in 020,

forces the prng-protium of the C-1 of the propanol to exchange with
deuterium. Their results fit V and VI.

 

v
””1,
O CH H
ONH2 CAD” \c/ 26 3
D H2013 -——-> \n
I I

' OH
R O (S)

20 had shown that when geraniol-l-t was oxidized

Donninger and Ryback
with L-ADH and NAUs, it was the (R) isomer that lost the radioactivity.
Hence, L-ADH must remove the 9:9:R-hydrogen at C-1 of geraniol, a

fact that fits Model K (VII).

CONH H R'
2 L-ADH \C /
‘\
——> m
r OH

 

(R)—geraniol—l-t
VII RI 4W

From the evidence presented thus far, we see that Karabatsos15
correctly questioned the importance of repulsive interactions between
the CONH2 of the pyridine nucleotide and groups attached to the carbonyl.
Prelog's first assumption is, therefore,incorrect.

Further evidence supporting Model K comes from fermenting yeast
reductions. Mosher et. al.17’19 used fermenting yeast (A-type enzymes)
to reduce various l-d-aldehydes, namely trimethylacetaldehyde-l-d,
benzaldehyde-a-d, and butyraldehyde-l—d. The absolute configuration of
the resulting alcohols was (S) as predicted by Model K.

They also reduced18

with fermenting yeast a large variety of
ketones, including all combinations of the substituents methyl, ethyl,
n-propyl, n-butyl and phenyl. The enantiomorph of the carbinol pro-
duced in excess had the (S)-configuration, with one controversial
exception, that from the reduction of ethyl n-butyl ketone. Stereo-
selectivity in the reductions of the ketones varied from values as high

as 90% in the aromatic series to as low as 12% for ethyl n-propyl ketone.

The lack of 100% stereoselectivity during fermentive reduction of these

10

ketones was rationalized by Mosher et. al. on the assumption that "it
is the difference in steric requirements of the L (large) and 5 (small)
groups which is important and not any absolute preference of one group
or the other for a particular enzymic site". They further assumed
that ADH was the only enzyme involved in these reductions.

Let us now examine bifunctional substrates, such as lactic acid
or lactaldehyde. These substrates present two primary binding sites
to the system in a way not possible for a simple ketone with one func-
tional group and inert alkyl substituents. From the stereoselectivity
of those substrates we may examine the importance of hydrophilic-hydro-
phobic regions in the enzyme and then test Prelog's second assumption.

Karabatsos et. al.22’23

determined the absolute configuration at C-1

of 1,2-propanediol-l-d obtained from the reduction of D- and L-lactal-
dehyde with A-NADD and L-ADH in order to evaluate the importance of
hydrophilic and hydrophobic enzyme regions. Since the methyl group 15
hydrophobic and the hydroxyl hydrophilic, they anticipated the possibility
that the polar hydroxyl group might force NADD and lactaldehyde into

IX - opposite to VIII of acetaldehyde - rather than X.

 

They found that the absolute configuration at C-1 of 1,2-propanediol-l-d

was (R). Thus, as far as product stereospecificity is concerned,

11

lactaldehyde, either D or L, and acetaldehyde have the same substrate-

coenzyme relationship (X and VIII) and hydrophilic interactions, if

H HCH
L-ADH % \ c Q
D‘oo \

3
H
OH I

 

 

(R)

present, were insufficient to overcome the steric interactions in—

volving substrate and enzyme. Van Eys and Kaplan16

had reported that
pyruvic acid (XIV) with Y-ADH and NADH yielded lactic acid which was
reoxidized with D(-)-lactic acid-specific lactic dehydrogenase (D-LDH)
and acetyl pyridine-HAD, but not with L-LDH and the coenzyme. The
product was, thus, D-(-)-lactic acid, which has the (R) configuration.
Consequently, the substrate-coenzyme spatial relationship for pyruvic

acid is XI, in which the hydrophilic carboxyl group occupies a position

 

XI (R)-(-)-lactic acid

opposite to that occupied by the hydrophilic hydroxyl group in X. One
may say that hydrophilic-hydrophobic interactions are more important
than steric interactions in this case. Similarly, Neuberg and Nord24’25

reported that the phytochemical reduction of biacetyl yields (-)2,3-

12

butanediol, in which both carbons have the (R)-configuration. The

spatial relationships are therefore, X11 and XIII.

 

,O
’ CH3
CH"
yeast I 3 \c /
\
{S“b
PI H
(R)
(¥E%,OH
yeast:J \C /CH3
\ Y\
‘0‘ ‘3 \,
H H '
(RT

26

Levene and Halti reported that reduction of 3-keto-l-butanol yielded
(-)-l,3-butanediol, which has the (R)-configuration. Again the spatial

relationship (XIV) suggests that hydrophilic—hydrophobic interactions

H

HQ”

H
2 [.l 0" 2 yeast 3 HOCH2CH2\c/CH3

CH
\‘l 3 \o“
lo ” H
R

(R)‘

HO-CH -CH

2

XIV

29 studied

are more important than steric ones. Lemieux and Gigeurre
several s-keto monocarboxylic acids and their data, shown in XV, are

Iagain in accord with what has been said up to now; so is the preparation

13

 

of (-)-l,2-propanediol3O (R-configuration) with fermenting yeast from
the reduction of acetol (XVI).

 

 

CHZOCH//CH3
yeastCCL
7 \\ a
{0 ‘D’
H OH
XVI (R)-(-)-l,2-propanediol
Grzycki27 however, reported (+)-l,3-butanediol (S-configuration)

from the reduction of 3-ketobutyraldehyde (XVII), a fact which (if

correct) suggests either XVII or XVIII.

H CH0
Q33,/5 2

C
\
10“ 10:5

(5)"

     

C CH 2CH 2OH
CONHZ yeast N§C//2
CHZCHZOH E \

\\\
H“ H

XVIII (5)

14

Since XVIII is the opposite of XIV, it follows that either different

enzymes are involved or XVII (not XVIII) is the reduction path. Fujise28
obtained (-)-L-malate (S-configuration) by yeast reduction of oxalacetate
(added as the ester) (XIX), but (+)—L-s-hydroxybutyrate (S-configuration)

from acetoacetate (XX).

  

H,” H ZOOH
ONH2 yeast Q /COOH
HOZC-CHZ OOH .____,, CC
. H8 \H
R
x1x (S)-(-)-malate
H H
C02 H
CONHZ yeast CH\3 C/c H2
CH3 CHZCOOH _____,. :\
N 11‘“ C‘
I o .
R (3)]
xx

For all the phytochemical (yeast) reductions mentioned above, the
specific enzyme involved in each case has not been determined. Mosher18
assumed, as already mentioned, that the only enzyme acting in the re-
duction of the various aldehydes and ketones was Y-ADH (A-type enzyme)
with a pyridine nucleotide as coenzyme. For the bifunctional or poly-
functional substrates, in which at least one of the R], R2 groups is
polar, we may assume that dehydrogenases other than Y-ADH might be
involved in the reduction of the keto group. Furthermore, in these

yeast reductions we cannot rule out the possibility that coenzymes other

than NADH may also be involved.

15

To understand better the forces governing product stereospecificity
in these reactions, studies with purified enzymes and coenzymes, in
which the extent of product stereospecificity is accurately determined,

3] reduced hydroxyacetone by using

are needed. Karabatsos and Nunez
purified glycerol dehydrogenase (Gly-DH), an A-type enzyme isolated

from Aerobacter aerogenes32, and NADH. The isolated product (R)-(-)-l,
2-propanediol was 100% optically pure. The spatial substrate-coenzyme

relationship was as in XXI, which shows that hydrophobic-hydrophilic

QE$?H’CH3
SH“) i“ NAD‘a
H OH

 

(R)-(-)-l,2-propanediol
interactions, rather than steric interactions, govern product stereo-
specificity in this case. To further evaluate the importance of hydro-
philic and hydrophobic XE: steric interactions we thought that it would
be appropriate to use ketones, RICORZ, where the groups R1 and R2 are
comparable in size to those of hydroxyacetone but not as different

in their hydrophilicities. He chose methyl ethyl ketone (XXII) and
Chloroacetone (XXIII).

XXII XXIII

CH3 CHZCH3 CH

This thesis describes our findings with these two ketones.

EXPERIMENTAL

 

Equipment

All NMR spectra were recorded on a Varian T-60 NMR Spectrometer.
Usually deuterochloroform was used as solvent and TMS was the standard.
All the pH measurements were done with an instrumentation Lab. Inc.
pH/mv Electrometer, Model 245.

Gas chromatographic analyses were performed on a Heulett-Packard
F and M scientific 700 laboratory chromatograph. The chromatograph
was equipped with a thermal conductivity detector and a Sargent, Model SR,
recorder. The carrier gas was helium, and the column used was a 12'

x 1/4" stainless steel coil, packed with 20% carbowax in chromosorb w,
60/80. For liquid chromatographic separations we developed a system
consisting of a chromatographic column packed with the appropriate
material, a Beckman DB-G grating spectrophotometer equipped with a

0.1 cm flow cell and a Sargent, Model SR, recorder (with range plug from
1.25 up to 125 mV) and a Buchler Fracto-mete 200 fraction collector
connected with the recorder through a selenoid to click the tubes.

A UNICAM SP 800 uv spectrophotometer was used to follow the reactions
spectrophotometrically, as well as for other spectrophotometric
measurements. The optical rotations for the successively recrystallized
2-butyl brucine phthalates were measured in a Perkin-Elmer, Model 141,
Polarimeter with an automatic read-out output. The other optical rota-
tions were measured by using a Zeiss precision Polarimeter with a 0.3 ml

cuvette, 1 cm path. Usually four readings for each of the five wave-

16

17

lengths 578, 546, 436, 405 and 365 nm were taken at 25°. Then, the
optical rotation for 589 nm (0 sodium line) was found either by extra-

polation of the plot “:5 lg. A, or by using the following equation35:

“57s

 

“546 ' “578

 

(l) “589 “546

“57s

a _ a + 1.3727
546 578

 

Mass spectrometry was performed on either a LKB 9000 mass spectrometer
(70eV), with spectra being recorded as bar graphs by means of an on-

35

line data acquisition and processing program , or a Hitachi, Ltd.

RMU-60 mass spectrometer.

Materials

L-ADH, V-ADH, Gly-DH (from A. aerogenes,1yophilized powder), NADH

 

and HAD$ (both grade III, 98%). glycylglycine (glygly) and DEAE-

 

cellulose were purchased from Sigma Chemical Company. Methyl ethyl

 

ketone was a "Baker analyzed" reagent. 2-Butanol was purchased from

 

Mallinckrodt Chemical works. Chloroacetone, from Pfaltz and Bauer,

 

Inc., was purified by fractional distillation (b.p., ll9.5°). l-Chloro-
2-propanol, from Aldrich Chemical Company (97%), was purified by frac-

 

tional distillation (b.p., 124-6°). Buffer 0.01 M glygly pH 7.2 was

 

prepared by using preboiled distilled water and crystalline glygly
(free base). The pH was adjusted by using dilute sodium hydroxide
solution.

The DEAE-cellulose columns were prepared according to Bio-Rad

 

Laboratories booklet instructions by using the above described buffer.

18

Treatment of Data

 

The reported average values are the mean average of N independent
determinations. The uncertainty indicated is the standard error 0.

N
(I/IN-iizIxi-iizil/z

i=1

Q
ll

X
11

observed value

x
II

mean value

For ratios, like a/c, the uncertainty indicated is the standard

error obtained from the following relationship.

a = a/c[<°a/a)2+(°c/c)21"2

I. Reaction of Methyl Ethyl Ketone With L-ADH and NADH

 

a. Procedure
A solution of 250 ml of 0.01 M glygly buffer, pH 7.20, and 2.5 g
(3.53 mmol) of NADH were placed in a 500 ml Erlenmeyer flask. To it
was added 2.1 g (28.1 mmol) of methyl ethyl ketone.
After that 40 units of L-ADH (26.8 mg of protein) was added. In
Table l is summarized the progress of the reaction by measuring the pH
and the maximum absorption at 340 nm (A340). Figure 2 gives the PIOt

of A 0 gs, time.For the A340 measurements a 50 ul of reaction mixture

34
was added in a 3 ml uv cuvette containing 2.95 ml buffer. The pH range

19

Table 1. Changes in pH and A340 with Time for the Reaction of Methyl
Ethyl Ketone with L-ADH and NADH.

 

 

 

 

 

 

 

 

Time in Hours pH A34o nm
0 (no enz. added) 7.20 1.80
5 (min) 7.27 1.74
0.75 7.50 1.66
1.0 7.57 1.61
1.5 7.78 1.52
1.5 pH adjusted to 5.85
2.0 5.95 1.46
4.0 6.26 1.24
7.0 6.71 0.93
10.5 8.22 0.62
10.5 pH adjusted to 6.56

11.5 6.97 0.58-
14.0 7.46 0.42
15.0 7.54 0.40
16.0 7.60 0.36
17.0 7.64 0.35
18.0 7.66 0.34

 

 

 

 

 

during the reaction was 5.85-8.22. The pH was adjusted by adding drops

of 1N HCl. The temperature range was 26-29°. Gentle and continuous

nm

at 340

Absorbance

20

 

 

 

I 1 I U I I

0 2. 4. 6 a 10 12 14 16 18

Time (hours)

Figure 2. Absorbance in 340 nm (A340) gs, time for the

reaction of methyl ethyl ketone with L-ADH
and NADH.

 

21

stirring was maintained by using a magnetic stirrer. After 18 hours the
reaction reached 81% completion (based on A340). The reaction mixture

was then kept for half an hour at 4°.

b. Separation of product and excess substrate from enzyme and coenzyme
The above reaction mixture was added all at once on the top of an
anion exchange column, kept at 4°, and containing 20 g DEAE-cellulose.
The column was 34 cm long and 2.5 cm wide. The flow rate of the column
was 1.5 ml per 2 min. After the reaction mixture had passed through
the column, elution was continued with 0.01 M glygly buffer. The first

3 fractions consisted of 40 ml each and the rest of 55 ml each.

c. Isolation and detection of the product 2-butanol

 

To each fraction collected above was added sodium chloride(15-20%).
After extracting each fraction twice with 70 cc ether, the ether extracts
were combined and dried over magnesium sulfate. The ether then was re-
moved by fractional distillation through a 2.5 cm long, 1.7 cm wide Vigreux
column and perpendicular condenser so that most of the condensed liquid
returned into the column. The heating was done in an oil bath kept at
45-50°. The residues (usually 0.8-1.1 ml) were then analyzed by gas
chromatography. The analysis was done at 77° with a helium flow of
25 cc/min (14psi). The retention time for 2-butanol was 31.3 min., for
ether 5.5 min. and for methyl ethyl ketone 18.1 min. The results are
summarized in Table 2.

Residues 4 through 8, which contained the 2-butanol, were combined
and the 2-butanol was collected by gas chromatography. A total of 27
injections, 95-98 ul each, were used at 83° and 2-butanol (Rf of 23 min)

was collected in a trap dipped into 2-propanol-dry ice. The collected

22

Table 2. GC analysis of the fractions collected with
respect to the content of 2-butanol.

 

 

Amount Fraction Attenuation 2-Bu0H(R =31.3)
Injected # peak hgight
10 ul 1-3 1 0

5 ul 4 1 9.2

5 u1 5+6 2 6.4

5 ul 7+8 2 8.5

10 u1 9+10 1 0

10 H1 G.I 1 16.6

10 ul G2 1 0

 

 

 

 

 

 

G1 + residues 4-8 combined.

G2 + residue from reextraction of all water layers combined.

2-butanol was dissolved in 0.7 m1 ether and after 8 injections it was
recollected. The final product was then dissolved in 0.3 ml absolute

ethanol.

d. Optical activity measurements

 

The optical rotations of the above solution, at two wavelengths, are
given in Table 3. Because of a leak in the cuvette, readings at other
wavelengths were not taken; such readings were taken with a more dilute

solution from another run (Table 4). From the two values, “578 = +0.0400°

and “526 = +0.0445°, reported in Table 3, and by using Equation 1 (p.17)

25 25

we calculate o589 = +0.0386° :_0.0007°. By this procedure 6589 is about

23

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 3. Optical rotations of product 2-butanol, (X%) solution,
in absolute ethanol.
25 o o o o 0 Average
“x x1 x2 x3 x4 x5 25
0
A
eggs +.040 +.039 +.040 +.041 +.040 +.0400°:.0007
62526 +.045 +.044 +.045 +044 - +.0445°:.0006
Table 4. Optical rotations of product 2-butanol solution
in absolute ethanol.
(125 X ° X ° X ° X ° Average
A 1 2 3 4
103 250
x “A
61:38 +0.010 +0.010 +0.010 +0.010 +10.00:O.6
0125 +0.01O +0.011 +0.011 LI-O.011 +10.75j_0.6
546
0125 +0.015 +0.016 +0.016 +0.015 +15.75:O.6
436
0125 +0.019 +0.020 +0.020 H0.019 H19.501-_O.6
405
0125 +0.025 +0.026 +0.025 *0.026 +25.50j-_0.6
365

 

 

 

 

 

 

 

 

obtained by plotting in Figure 3 the data of Table 4 and extrapolating

25
to “589 o

4% less than 6:5

78'

This 4% difference in the value of 0:39 is also

 

24

 

 

 

 

025- ’

O

C

O

._ .020- .

O

‘6

8

_ .015 .

O

U

2

o O
.010- \.\‘~
.005 I l T I I .

350 400 450 500 550 600

VVaire~|eIiggt h nln

Figure 3. Plot of optical rotation gs, wavelength for product
2-butanol in absolute ethanol: 0 , measured; 5,
extrapolated.

25

e. Determination of the concentration (X%) of product 2-butanol in

 

the ethanolic solution

 

Three standard solutions 5.64%, 6.42% and 7.12% of commercial
2-butanol in absolute ethanol were prepared by diluting 5.64 g, 6.42 g
and 7.12 g of 2-butanol in absolute ethanol to a total volume of 100 ml
each. 0f each of these standard solutions, 10 ml was repeatedly in-
jected in the gas chromatograph at a temperature of 82° and attenuation

16. The data are summarized in Table 5.

Table 5. Relation between GC peak height and
2-butanol concentration.

 

 

 

 

 

% Concentration g of 2-butanol/1OO ml solution
5.64 6.42 7.12 xx
9.15 10.10 10.95 10 10
I: 9.15 10.30 11.25 10.15
01
£3 9.30 10.40 11.25 10.30
E 9.40 10.60 11.35 10.35
0.

9.50 10.65 11.50 10.55
s3 9.30 10.41 11.26 10.29

 

 

 

 

 

 

From a plot of the data of Table 5 (Figure 4) the concentration

of 2-butanol in the sample is found to be 6.36 :_0.l7%. Since this

solution had a rotation (3:39 = 0.0386° 1 0.0007°, its specific rota-

[“325 _ +0.0386° 1 0.0007° x 100 g
0.1 x (6.36 10.17)

tion is calculated to be:
+6.07° :_0.20° (or13.29%).

 

cm

height

Peak

26

 

11,5~

11.0-

10.5-

10.0-

9.5-

 

‘/

 

8.5

T T I l I

5.5 6.0 6. 5 7, O 7. 5

0/0 2-BuOH in ethanol

Figure 4. Plot of peak height gs. % concentration of 2-butanol
in absolute ethanol: 0, standard solutions; 5, un-
known solution.

 

27

37-41

f. Resolution of 2-butanol

 

Preparation of 2-buty] hydrogen phthalate. A 1000 ml one neck

 

round-bottomed flask was equipped with a magnetic stirrer and a reflux
condenser with a calcium chloride tube at the top. The entire apparatus
was dried. To the flask were added 74.1 g (1.00 mol) 2-butanol and
148.1 g (1.00 ml) phthalic anhydride (Matheson Coleman and Bell, 99.5%)
and the mixture was refluxed under continuous stirring for 12 hrs at
100-110°. After cooling, the reaction mixture was dissolved in 2 liters
of water by addition of excess sodium carbonate. The solution was ex-
tracted three times with 500 ml portions of ether. The aqueous layer
was then purged of ether by bubbling air through it for 4 hrs.

The solution was then acidified with concentrated hydrochloric
acid. The resulting white emulsion was immediately extracted with 1.5
chloroform in three portions. The combined extracts were washed several
times with water. After the solution was dried with calcium chloride
and filtered, the chloroform was distilled. The last traces of it were
removed under vacuum. The residue was recrystallized from 1200 ml 30-60
petroleum ether and 400 m1 90-120 petroleum ether. The crystals were
identified as 2-butanol hydrogen phthalate by NMR and mass spectrometry;
m.p. 58-6l°; yield 144.0 g (65%).

Preparation of the brucine salt of 2-butyl hydrogen phthalate. A
1000 ml three-necked round-bottomed flask was equipped with a magnetic
stirrer and a reflux condenser with a calcium chloride tube at the top.
The entire apparatus was dried. To the flask were added 44.5 g (0.20 mol)
2-buty1 hydrogen phthalate and 300 m1 acetone. After the solution was
heated at 40-50°, 78.9 g (0.20 mol) of anhydrous brucine (brucine

28

tetrahydrate, from Aldrich Chemical Company, was placed in a 50° oven

and a 10 mm vacuum was applied until the weight of the brucine remained
stable)was added in small portions allowing each portion to dissolve

before the next was added. When half of the brucine was added the solution
turned milky. The volume of acetone was then increased to 800 m1 and the
addition of brucine was continued. After all the brucine had been added
the reaction mixture was heated to reflux. The reflux was maintained

until the reaction mixture turned clear. Then, the hot reaction mixture
was filtered and allowed to crystallize. The crystals were filtered

and dried. m.p. 138-148°; yield 95.3 g (78.2%).

Purification of the brucine salt of (S)—2-buty1 hydrogengphthalate.
Table 6 summarizes all data on the successive recrystallizations of the
brucine salt of 2-butyl hydrogen phthalate to give the pure stereoisomer.
The original 95.3 g (155 mmol) of the brucine salt of 2-butyl hydrogen
phthalate is labelled product A. Product A was recrystallized several
times from methanol yielding successively purer products 8, C, 0, etc.
The recrystallizations continued up to the point where the specific rota-
tion of the product remained constant. The measurements were done in 4%
solutions in methanol for products A, B, C, and D and in 4% solutions in
ethanol for products 0 through L. The yield based on product L was
5.47 g (6%).

Isolation of (S)-2-butanol from the brucine salt of (S)-2-butyl

hydrogen phthalate. To a small, compact distillation apparatus equipped

 

with a 100 m1 round-bottomed flask, magnetic stirrer and a Vigreux column
(8 cm long, 1.5 cm diameter) was added 5.47 g (8.9 mmol) of the brucine
salt of (S)-2-butyl hydrogen phthalate and 40 ml water containing 1.1 g

29

 

 

 

Table 6. Data summarizing the recrystallizations for the purification
of (S)-2-buty1 brucine phthalate.
Product Weight m.p. Specific 4% in Volume solvent used
9 rotation solvent to dissolve it
- - - - - 800 ml acetone
A 95.30 138-148 -6.25 MeOH 270 m1 MeOH
B 37.26 154-155 -2.35 MeOH 100 ml MeOH
C 29.72 156-158 -l.07 MeOH 70 m1 MeOH
D 26.06 156-158 -0.09 MeOH 50 ml MeOH
D - - -4.43 EtOH -
E 19.95 - -4.40 EtOH 50 ml MeOH
F 17.06 - -4.05 EtOH 50 m1 MeOH
G 15.69 - -3.85 EtOH 50 m1 MeOH
H 14.00 156-157 —3.65 EtOH 50 ml MeOH
I 12.12 - -3.58 EtOH 50 ml MeOH
J 9.75 - -3.50 EtOH 50 m1 MeOH
K 7.38 156-157 -3.48 EtOH 50 m1 MeOH
L 5.47 - -3.475 EtOH 50 ml MeOH

 

 

 

 

 

 

 

(27.5 mmol) sodium hydroxide. The reaction mixture was then heated gradu-
ally up to boiling, with constant stirring, to start the distillation. To
the first 15 m1 distillate (88-100°) was added sodium chloride (20%).

After extracting the resulting solution twice with 40 ml ether, the ether
extracts were combined and dried over magnesium sulfate. The ether was

distilled by using a Vigreaux column, and the residue (about 0.8 ml) was

analyzed by GC and found to be about 50:50 ether: 2-butanol.

30

After a

few injections (95 ml each), about .35 9 (yield 65%) of highly pure

2-butanol was collected.

Two solutions of 6.55% and 6.38% in absolute

ethanol of the 2-butanol were prepared and their specific rotations

were determined as described previously.

6.38 :_0.05% solution are given in Table 7.

The optical rotations for the

From Figure 5 (plot of

 

 

 

 

 

 

Table 7. Optical rotations for a 6.38% solution of resolved (+)-(S)-
2-butanol in absolute ethanol.
Average
25 o o o o 3 25a
“A X1 X2 X3 X4 10 X (1A
a25 +0.095 +0.095 +0.095 +0.094 94.75 :.0.5
578
“S26 +0.112 +0.113 +0.113 +0.112 112.50 :_0.6
“236 +0.1s3 +0.182 +0.183 +0.132 182.50 :_O.6
“335 +0.217 +0.217 +0.218 +0.217 217.25 :_0.5
“€25 +0 275 +0.275 +0.275 +0.276 275.25 :_0.5

 

 

 

 

 

 

 

 

data of Table 7) the 6:39 was determined to be

from this the specific rotation was determined

: 0.19°.

+0.088° :_0.001°, and
to be: [ajggg = + 13.79°

Similarly from the 6.55% solution the specific rotation deter-

mined was [“]539 = + l3.89° :_0.18°. Since the previously reported

value for the specific rotation3

, in the same solvent, was [“]§5 =

+ 11.58° (C = 4.8% in absolute ethanol) we believe that the (+)-(S)-2-

butanol prepared in our work has high optical purity.

The highest values

Rotation

Opti cal

31

 

 

 

 

 

.30-
\.
.25- I“;
”“11
O
.20- \ J ..
O
.154
O
.10- \
O\
. 5 . 1 u u u
350 400 450 500 550 600 \
Wo v o I o n g t h n m

Figure 5. Plot of optical rotation gs. wavelength for a 6.38%
solution of resolved (+)-(S)-2-butanol in absolute
ethanol: 0 , measured; 0, extrapolated.

32

ever reported are those of Pickard and Kenyon37, [“130 + 13.87° (neat)
and [6137 + 13.52° (neat).

II. The Reaction of Chloroacetone Hith L-ADH and NADH

 

a. Procedure

To 170 m1 of 0.01 M glygly buffer, pH 7.59, containing 2.00 g
(2.82 mmol) NADH in a 300 m1 Erlenmeyer flask was added 0.518 g (5.6
mmol) of Chloroacetone. Then, 20 units of L-ADH (13.4 mg of protein)
was added. In Table 8 the progress of the reaction is summarized by
measuring the pH and the absorption at A340. Figure 6 gives the plot
of A340 gs, time. The measurements were done as in the case of methyl
ethyl ketone. The pH range for the reaction, which was adjusted by
adding drops of 1 N hydrochloric acid, was 5.74-8.16. The temperature
range was 28-33°. The stirring was very gentle and continuous by use
of a magnetic stirrer. After 70 minutes the reaction reached 85% comple-

tion (based on A340).

b. Isolation and detection of the product l-chloro-Z-propanol

The reaction mixture was divided in five portions. After addition
of sodium chloride, (ls-20%) each portion was extracted twice with about
double amounts of ether. All ether extracts were combined and dried over
magnesium sulfate. The ether then was removed by fractional distillation
by using a 25 mm long, 1.7 cm wide Vigereaux column to give a residue of
2.5 ml. Gas chromatographic analysis of the residue at oven temperature
141 :_1° gave a chromatogram with three peaks:

1) R

f 2.8 min., (ether) peak height 20.5 cm, Attenuation 512.

2) Rf 9.6 min., (Chloroacetone) peak height 18.9 cm, Attenuation 8.

33

Table 8. Changes in pH and A340 with time for the reaction
of Chloroacetone with L-ADH and NADH.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Time Past min pH A340 nm

0 (no enz.) 7.59 1.63

2 7.65 1.56

3 7.77 1.53

6 7.87 1.45

9 7.96 1.39 E; I
12 pH adjusted to 5.74 “
15 6.12 1.23

17 6.45 1.19

22 Addition of 20 Units

(13.4 mg) of enzyme L-ADH

25 7.33 1.05

30 7.65 .87

35 7.86 .75
40 8.02 .61 \
44 pH adjusted to 7.02
45 7.05 .50

50 7.30 .44

55 7.45 .37

60 7.56 .34

65 7.60 .30

70 7.62 .28

 

 

 

 

at 340 nm

Absorbance

34

 

 

 

 

I ‘5 I Y

O 10 20 30 40 50 60 70

Time (min)

Figure 6. Absorbance at 340 nm (A340) gs, time for the reaction
of Chloroacetone with L-ADH and NADH: O , more enzyme
added.

35

3) Rf = 13.4 min, (l-chloro-Z-propanol) peak height 15.0 cm,
Attenuation 8.

The product 1-chloro-2-propanol was then purified by making 16 injections,
95-98 ul each, at 143-4 :_1° into the gas chromatograph and collecting
the l-chloro-Z-propanol (Rf = 12.2-12.8 min) in a trap dipped into ice-
water. The collected product weighed 0.0685 g. The product was further
identified by nuclear magnetic resonance and mass spectrometry, and it
was purified by injecting and recollecting as described above. The re-
collected l-chloro-2-propanol (0.0290 9, 0.0271 ml) was dissolved in
0.5696 g (0.387 ml) of chloroform to give a solution of about 7.0%.

c. Optical activity measurements

 

The optical rotations of the solution at five wavelengths, are

given in Table 9. From Figure 7 (plot of data of Table 9), by

Table 9. Optical rotations of product l-chloro-2-propanol (X%) solution
in chloroform.

 

Average
25 0 ° ° ° ° 3 25 o

 

25 +0.011 +0.010 +0.010 +0.011 +0.010 +10.40 :_0.6

 

 

 

 

G25 +0.011 +0.011 +0.0011 +0.012 - +11.25 :_0.5
546

Q25 +0.014 +0.015 +0.014 +0.015 - +14.50 :_0.6
_ 436

G25 +0.017 +0.018 +0.018 +0.018 - +17.50 :_0.6
405

25 +0.025 +0.027 +0.027 +0.028 - +27 00 :_0.8

“365

 

 

 

 

 

 

 

 

 

Rotation

Optical

36

 

 

 

 

 

030~
\.
.025-
.020 '1
o
.015 - \
o
o\
.010 -‘ .~‘\
.005 g I I - | r
350 400 450 500 550 600
\Navolongth (nm)
Figure 7. Plot of optical rotation g_s_. wavelength for product

l-chloro-Z-propanol in chloroform: O , measured;
_5 , extrapolated.

37
extrapolation, we determined the 0:39 = 0.0l00° :_0.0006° for the (X%)
solution of product l-chloro-Z-propanol in chloroform.

d. Determination of the concentration of the (X%) solution of product

l-chloro-Z—propanol in solution.

 

Three standard solutions, 6.01%, 7.00% and 8.01%, of commercial
l-chloro-Z-propanol in chloroform were prepared by diluting 3.006 g,
3.50l g and 4.004 g of l-chloro-Z-propanol with chloroform to a total
volume of 50 ml in each case. Several 10 pl portions of each of these
standard solutions were injected in the gas chromatograph at 143° :_l°
and attenuation l6. The data are summarized in Table 10.

Table l0. Summary of data to determine the relation-

ship between l-chloro-Z-propanol peak
heights in GC and % concentration.

 

 

 

 

 

 

 

 

5.01% 7.00% 8.0l% xx
.. 10.9 12.9 15.1 12.7
‘55
-; 11.1 12.8 15.2 12.7
I
-§ 11.2 13.2 15.0 12.9
33
10.3 12.9 15.3 -
a 11.00 12.95 15.15 12.77
>
<

 

 

From Figure 8 (plot of data of Table 10) the concentration of
product l-chloro-Z-propanol in the solution was found to be 6.9 1.0.l%.
Hence the specific rotation for this particular solution of the product

is [51:39 = 00i0l0?; 3 2‘80$%:x 100 = +1.45° :_0.09° (or :_5.2%).

 

 

 

 

 

16-
0/
l5-
5
U
I4T
.3 ‘3‘ A”
O
" 12—
.8
° ll - o
O
a. /
IO -1
9 I I I I
5 6 7 8

°/a concentration

Figure 8. Plot of peak height y_s_. % concentration of l-chloro-
2-propanol in chloroform: o , standard solutions;
a , unknown solution.

39

9. Determination of the enantiomeric purity of product l-chloro-Z-

propanol by NMR with a chiral Europium shift (optishift) reagent.

 

The sample of l-chloro-Z-propanol used for this determination was
the product of a reaction carried out to 25% completion (9 hrs) with
glycerol dehydrogenase (Gly-DH) and then to 85% completion (1 hr) with
added L-ADH. Otherwise the reaction conditions were exactly the same
as the ones already described in Section 11a. The product was isolated
in the same manner as that from the reaction of methyl ethyl ketone
(section I, b and c). It was purified by injecting it twice in GC and
collecting it back and then it was dissolved in chloroform. The solution
was dextrorotatory, as expected.

NMR spectra were taken for the product (P) and commercial l-chloro-
2-propanol (C), both of which had similar concentrations in chloroform.
First, the NMR spectra of the two solutions were taken without adding the
optishift reagent. Then tris[3-(heptafluoropropylhydroxymethylene)-d-
camphorato]-europium(III), (Eu(hfde), obtained from Willow Brooks Labora-
tories, was added in 25 mg portions until the separation of the two methyl
doublets was satisfactory. NMR spectra were taken for various sweep
widths, both in the forward and reverse direction. The peaks of the two
methyl doublets, obtained at 50 Hz sweep widths were cut and weighed.

Each reported peak weight is the average of 10 independent measurements.
Two independent tracings A and B (from different spectra) were obtained
for the product sample (Figure l2) and one (C) for the commercial compound
(Figure 13). Table ll summarizes all the data. The four peaks were
numbered l, 2, 3 and 4 from a downfield to upfield direction. The down-
field doublet (peaks 1 and 3) represents one enantiomer and the upfield

(peaks 2 and 4) represents the other.

40

Table 11. Average weights (104g) of the cut-outs of the four peaks
of the methyl doublets for two product l-chloro-Z-pro-
panol spectra (A and B) and one commercial compound
spectrum (C). The sweep width of the spectra was 50 Hz.

 

Downfield Doublet

 

 

Spectrum Peak l Peak 3 Peaks l + 3
A 393.0 :_4.8 202.5 :_4.8 648.3 :_9.6
B 374.4 :_4.3 246.0 :_4.9 620.4 :_9.2
C 323.1 :_3.7 325.3 :_4.3 648.4 :_8.0

 

 

 

 

Upfield Doublet

 

 

Spectrum Peak 2 Peak 4 Peaks 2 + 4
A 235.5 :_3.6 334.6 :_4.2 570.1 :_7.8
B 236.8 :_4.2 314.4 :_3.8 551.2 :_8.0
C 322.0 :_3.9 331.0 :_3.9 653.0 1 7.8

 

 

 

 

 

 

41

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 9. The NMR spectra of product l-chloro-Z-
propanol in chloroform before (P0) and
after (P1) the addition of 50 mg Eu(hfpc)3.

42

 

W W I T v T1 ‘ —v 9"?‘r‘l‘1—‘V"'\"t"n T'l T’FP‘T-fi
g-Wr‘fil—r-l-wI-v-s-Er—r . a - ---_-‘v-.-- - .- q—p. r...
m . ' . 11L: Glue-1
l ; 1 .
. .\ -

I

I

0

I

I 1 M.

. ' i I! D

1 z ‘ . t i 1

«I I a n a . : W

t , ' I 1 1

If. 2 ; 3 MI 1 1 - 2 ,

a- . s : . . I

: ‘ :' . . ' v 3

5 . ‘- i ; ; i Z

I. . . . . ; . 3. 1 < . .

i 1 ‘ . . ,

g ; ; ' 1 ' 1 :

1 t. ..

I . -‘ 1 .

. - i

i 1 1.
0

j I

, ,

   
 

.. —' a?‘-'-—' . _-‘*—_—

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:1-
h
:5
3

' - 1. - -.

Figure 10. The NMR spectra of commercial l-chloro-Z-
propanol in chloroform before (C0) and after
(C) the addition of 50 mg Eu(hfpc)3.

43

:owuvcum mcu gmaya Egomogopgu c. Focmaogn1N1ogopgo-F ouzuosa mo Ezeuomam «:2 as»

.mfiuaegvsm as cop to

._P mesmea

 

 

 

. 1

...
,. a

. . , a
- Q...— a-‘O- .w-o-

.......

 

 

 

 

 

.060

.II aineptioiollli ill-bll'llllo I‘ll.-

 

 

. g . a
-.. -gor -—¢ oar-0’ . '
a-
n

m.

n,

.--

p

ld‘ol'r'0
. P n - L b

«I- - o‘-

1.1

    

 

 

.

n
.... a
u . -
.7.
m .
~..

 

 

III '0'.

-- a ...
——o--o con—..—

4 ...—q --.—-
i
3
l
l
1?

 

 

o chl if. 4
o .0195 ...mWI! 31. .4 .A

 

 

 

.....

-. .
u
“ .I .m ... ‘1 ...“.
. m
(0.1. 1
u
.u
“ ....
.
I. I'. . nc‘u

 

 

 

 

44

no - ........- .
‘ C
.' ... ’ - 0--Do-¢.;- .... . .
u..oco¢ -...

 

-..-uo.|
.o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 12. Magnified methyl doublets (sweep width 50 Hz)
of two different runs of the NMR spectrum of
product l-chloro-Z-propanol after the addi-

tion of 100 mg Eu(hfpc)3.

45

 

 

 

 

 

 

 

 

I
.....

Figure'k3. Magnified methyl doublets (sweep
width 50 Hz) of the NMR spectrum
of commercial l-chloro-Z-propanol
after the addition of 50 mg Eu(hfpc)3.

RESULTS AND DISCUSSION

A. Enantiomeric purity of product 2—butanol

 

As pointed out, the specific rotation of the product 2-butanol was

25
[“1589

= + 6.07° :_0.20° and the specific rotation of the resolved (+)—(S)—
Z-butanol [91239 = + 13.79° :_0.19°. Both values were determined from
6.38% solutions of the corresponding Z-butanols in absolute ethanol.

From these values the anantiomeric purity (E.P.) of the product 2-butanol

from the reduction of methyl ethyl ketone with L-ADH and NADH is as follows:

 

x 100 = 72.0 1 0.8%(+)-(S)-2-butanol

and, consequently, 28.0 :_0.8% (-)-(R)-2-butanol.

B. Enantiomeriggpurity of product l-chloro-25propanol

 

1) From specific rotation.

As mentioned previously, the specific rotation of the product 1-
chloro-Z-propanol was [01229 = + 1.45° :_0.09° (c = 6.90% in chloroform).
The only literature value for the specific rotation of (-)-(R)-l-chloro-
2-propanol (I) is [a]§gg = -19.l9° (c = 5.17% in chloroform)42. The

authors note that the sample used for their measurement was contaminated
43

with 5% (+)-(S)-2-ch10ro-1-propan01 (11) whose specific rotation was
42 .
[a]§39 = + l7.39° (neat) and [“1539 = + 15.74° (c = 5.07% 1n chloro-

form). They further note that the 15.74° value was obtained from a

sample of II contaminated with 5% of I. From the above one may calculate

46

47

25
589

chloro-Z-propanol. From this value the enantiomeric purity of the product

that [a] = + 21.12° (c = 5.17% in chloroform) for pure (+)-(S)-l-

1-chloro-2-propanol from the reduction of Chloroacetone with L-ADH and

NADH is:

5 +
E.p. = 2i-'2 1.0-'5 _x 100 = 53.4 1.0.22 (+)-(S)-1-chloro-2—pro-

 

panel and, consequently, 46.6 :_0.2% (-)-(R)-l-chloro-2-propanol.

2) From NMR data with Eu(hfpc), optishift reagent

 

From the data summarized in Table 11 and the fact that the product
l-chloro-Z-propanol was dextrorotatory, we conclude that the downfield
doublet (peaks 1 and 3) corresponds to the (S) isomer and the upfield
doublet (peaks 2 and 4), to the (R) isomer. The enantiomeric purity may
thus be calculated as follows:

From Spectrum C (control):

 

 

Downfield doublet a 648.4 + 8.0 = 0 = 49.8 + 0.4
Upfield doublet 553.0 :_7.8 0°993 i-°°°‘7' 0r 0' 50T2‘1‘074’

From Spectrum A:

 

 

 

 

(s = 548.3 + 9.5 _ s _ 53.2 + 0.5
T8) 570.1 :_7.8 ‘ 1°‘37 i-0'023’ °r R ' 45.8 1.0.5
From Spectrum 8:
s _ 520.4 + 9.2 _ s _ 53.0 + 0.5
(Ri" 551.2 :_8.0 ‘ 1°‘25 i-0'023’ °r (fii" 47.0 :,0.5

The average values of 53.1 :_0.5% (S)-isomer and 46.9 :_0.5% (R)-
isomer agree well with those obtained from the optical rotation

measurements .

48

C. General comments about reactions 1 and 2

 

 

L-ADH
- _ 2W---” - _ -
C”3.9-012 CH3 + NADH + ”a pH = 5.9 - 8.2 CH3 2:” CH2 CH3 + 111409
0 81% in 18 hrs OH
(1)
72% (S)
28% (R)
L-ADH
G 28-33°C
85% in 70 min 0H
(2)
53.4% (S)
45.5% (R)

Reactions 1 and 2 were followed by measuring the decrease in ab-
sorption at 340 nm (A340) which corresponds to the decrease of NADH. The
uv spectrum of NADH shows two maxima in absorption, one at 260 nm ( e=
14,500) attributed to the adenine portion of the molecule and another at
340 nm (5 = 6,220) attributed to the dihydropyridine portion. The uv
spectrum of NAD‘9 shows only one maximum absorption at 259 nm ( c= 17,000),
attributed to both adenine and nicotinamide portions of the molecule
(Figure 15). From Figures 2 and 6 we conclude that both reactions fol-
lowed typical Michelis-Menten enzyme kinetics. As seen, reactions 1 and
2 progress by consuming Hei Thus, by using a low capacity buffer, as we
did, the reactions may be followed by measuring the increase in pH. A
point of interest is that L-ADH reacts faster with Chloroacetone than
methyl ethyl ketone under the same conditions. The reaction of chloro-
acetone was more than 80% complete within one hour, whereas the reaction

of methyl ethyl ketone required at least 18 hours. It should be noted

 

49

ant-o
a.“

‘D

at

(«It

 

 

 

 

MILLIMOLAR EXTINCTION COEFFICIENT
;.

WAVE LENGTH (mu)

Figure 14. Absorption spectra of NAD9 and NADH (Ref. 51)

too that in the case of methyl ethyl ketone the molar ratio of substrate
to coenzyme (NADQ) was 8:1 , whereas, in the case of chloroacetone the
ratio was only 2:1. He found that fresh enzyme, and to a lesser extent,
fresh coenzyme were important for getting a faster reaction with larger
yields. In some cases we got only 65% completion of the latter reaction
(1) within 42-48 hours. Control reaction mixtures showed very little
change in the A340 absorption. The reason for keeping the reaction mix-
ture cold after the completion of the reaction and later during the
process of separation of the product was to keep "alive" the unreacted
NADH and NADe which were kept tightly bound on the DEAF-cellulose column.

The coenzymes were recovered by the method of Pastore and Friedkin48;

then NAD9 was enzymatically reduced to NADH with Y-ADH and alcohol”.

The overall yield was 50-60%.

50

All methods for the separation of the products were first tested
extensively by using known mixtures. As we demonstrated in the isola-
tion of 1-chloro-2-propanol in the reaction of chloroacetone with L-ADH
and NADH,the separation of excess substrate and product from the coen-
zymes (and enzyme) by using the DEAF-cellulose column is not necessary
if one is not concerned about recovering the (expensive) coenzymes.
Based on the decrease in the absorption at 340 nm, Gly-DH gave a slow
reaction with chloroacetone under the same conditions. Large amounts
of enzyme (Gly-DH) and large excess substrate gave a considerably faster

decrease in A340.

0. Stereospecificity

 

1) Methyl ethyl ketone

L-ADH is an A-type enzyme believed44 to be involved in direct transfer
of a hydride ion, in the rate-controlling step, from the dihydropyridine
ring (HA) to the carbonyl of methyl ethyl ketone. The hydride ion
attacks preferentially the re-face of the substrate to give (+)-(S)-2-
butanol as the major product (72%). This corresponds to a stereospeci-
ficity of 44%.

This is the first time in which methyl ethyl ketone was reduced
enzymatically by using purified enzyme L-ADH and purified coenzyme.
Neuberg and Nord45 reported that methyl ethyl ketone was converted by
actively fermenting yeast to (+)-2-butanol, which possessed approximately
25% of the rotation of the pure dextro isomer. This value corresponds
to an enantiomeric composition of 62.5% (+)-2-butanol and 37.5% (-)-2-

18

butanol in the product. Mosher et a1. , using more refined methods of

isolation and purification, reported that methyl ethyl ketone was reduced

51

by actively fermenting yeast to (+)-(S)-2-butanol which was 64-67%
optically pure. This value corresponds to an enantiomeric composition
of 82-84% (+)-(S)-2-butanol and 16-18% (-)-(R)-2-butanol. Finally,

it is of interest to note here that in the reverse reaction, namely

the oxidation of various carbinols with purified Y-ADH and NAD+, Van Eys

‘6 have observed that approximately one-half the stoichiometric

and Kaplan
amount of NAD+ was consumed in the oxidation of dl-Z-butanol (.11 out
of .25 mmol) and dl-2-octanol (.12 out of .25 mmol). In the case of
2-octanol the (+) isomer consumed the quantitative amount of NAD+ while
the (-) isomer was not oxidized, thus demonstrating a 98-100% stereo-

46’47 made the assumption

selectivity in the oxidation. Some investigators
that the same might happen with the (+) isomer and (—) isomer of 2-butanol.
But we argue that the extrapolation from 2-octanol to 2-butanol is in-
correct, especially in view of the fact that in the d1-2-butanol oxidation
by purified Y-ADH + NAD+ only 0.11 mmol out of the 0.25 mmol NAD+ was
consumed (which accounts only for 88% of the half amount). It is also
interesting to note that Van Eys and Kaplan in their paper do not discuss
anything more than the above stated fact about 2-butanol. Then 100% stereo—
selectivity in the above oxidation of 2-butanol cannot be justified by
the evidence presented.

The lack of 100% stereospecificity during the fermenting yeast re-
duction of methyl ethyl ketone (and a large variety of other ketones) was

18 as due to the fact that the substrates

rationalized by Mosher et al.
are unnatural to the only enzyme, Y-ADH, assumed to act in these reductions.
If the above assumption holds, then in view of our results, we have to
assume that L-ADH is less stereospecific than Y-ADH with respect to sub-

strate methyl ethyl ketone. However, if we assume that L-ADH and Y-ADH

52

behave the same way, then we have to assume that another factor, not

yet known, is present in fermenting yeast responsible for the larger
stereospecificity towards methyl ethyl ketone. In view of the evidence50
that Y-ADH possessed a greater degree of "steric hindrance" at the active
site than L-ADH, the first alternative seems more attractive. Indeed,
L-ADH will provide more freedom to the molecule (methyl ethyl ketone)

to orient or reorient itself with respect to the coenzyme, whereas,

Y-ADH will give a more fixed transition state because of space limitation

and so the stereospecificity will be greater.

2) Chloroacetone

 

The enzymatic reduction of chloroacetone has not been reported in
the literature. He again assume that during the reaction a hydride is
directly transferred in the rate controlling step, from the dihydro-
pyridine ring (HA) to the carbonyl of chloroacetone. The hydride ion
attacks preferentially (very poorly) the re-face of the substrate to give
(+)—(S)-1-chloro-2—propanol as major product (53.4%). This corresponds
to a stereospecificity of 6.8%. And so here is an enzymatic reaction,
relatively very fast, involving pure enzyme and coenzyme in which the
stereospecificity is very poor. Of course the substrate is unnatural,
but there should be some very special reason to have complete stereo-
specificity31 (hydroxyacetone) or a fair amount of it (methyl ethyl
ketone) with substrates very similar in size with chloroacetone, but
only very little stereospecificity with it.

Because of the fact that both the reaction of chloroacetone with L-ADH
and the reaction of the same substrate with L-ADH and Gly-DH give product
with the same (considering error) enantiomeric purity 53.4% and 53.1%

(S)-isomer, respectively, we can safely assume that both enzymes L-ADH

53

and Gly-DH possess the same stereospecificity with respect to chloro-
acetone.

There is no possibility‘of racemization in the process of separation.
The two samples were isolated with different techniques. The reaction
conditions were neutral so any racemization mechanism is unlikely. The
possibility of racemization in the GC is also excluded, because after
reinjection and collection, the product did not lose optical activity.

The product was identified by means of GC, NMR and mass spectrometry.

E. Steric enzyme-substrate interactions

 

we already presented in our introduction (p. 9) evidence that Prelog's
steric coenzyme-substrate interactions assumption is incorrect because
it contradicts the available facts. These facts agree rather with

Karabatsos' model K for A-type enzymes than Prelog's model for A-type

 

Prelog's model for Karabatsos' model (K) Prelog's model (P)
A-type enzymes for A-type enzymes for B-type enzymes

enzymes. Briefly, the facts are:

1) Enzymatic reductions, by using purified enzyme-coenzyme systems,

of acetaldehyde-l-d (Y-ADH, and L-ADH), propanaldehyde-l-d

2)

Table 12.

54

(Y-ADH), geraniol-l-t(L-ADH) and methyl n-hexyl ketone (Y-ADH)

(Table 12).

Fermenting yeast (assuming that only Y-ADH acts) reductions of

trimethylacetaldehyde-l-d, benzaldehyde-a-d, butyraldehyde-l-d

and ten ketones with all the possible combinations of methyl,

ethyl, n-propyl, n-butyl and phenyl substituents (Table 13).

enzyme (NADH). RICOR2 + RICH(0H)R2 (S)-isomer.

Reduction of carbonyl compounds by purified enzyme and c0-

 

 

 

 

 

 

Carbonyl Compound Enzyme Stereospecificity % 885* 30° Ref.
R1= s R2= L ([a]obS/[a]max) x100 cal/mol

0 CH3 L-ADH, Y-ADH 100 12,52
0 -CH2—CH Y-ADH ~100 21
T 1633\3e32‘ Y-ADH ~100 20
CH3 n-hexyl Y-ADH 98-100 16
CH3 -CH2CH3 L-ADH 44.0 :_1.5 583 1,25
CH3 -CH2-C1 L-ADH, Gly-DH 5.8 1‘. .4 85 i 9

 

 

 

 

 

 

All the above compounds upon enzymatic reduction yield the (S)-isomer as

sole or predominant product.

Our results also show that the enzymatic

reductions, by using purified enzyme-coenzyme systems, of methyl ethyl

 

55

Table 13. Carbonyl compound reductions by actively fermenting
yeast. RICOR2 + RICH(OH)R2(S)-isomer.

 

 

 

 

 

 

 

[Cerbonyl Compound Stereospecificity % AAGf 30° Ref,
R = s SR = L r
l 2 ([alobS/[almaxl x 100 cal/mo1 a
D n-Pr 100 19
0 t-Bu 100 17 j
0 Ph 100 19 F
Me Et 64-67 920-990 18
Me n-Pr 61-64 860-920 18
Me n-Bu 82 1740 18
Et n-Pr 12-23 140-280 18
Et n-Bu 13-27 160-330 18
n-Pr n-Bu O 0 18
Me Ph 59 1030 18 \
Et Ph 63-72 890-1120 18
n-Pr Ph 84-90 1500-1780 18
n-Bu Ph 86-89 1580-1720 18

 

 

 

 

 

 

 

ketone (L-ADH) and chloroacetone (L-ADH and Gly-DH) yield the (S)-isomer
as the predominant product (Table 12). So we have to admit that steric

interactions are very important for determining the stereospecificity

56

in the enzymatic reduction of carbonyl compounds, as long as both sub-
stituents to the carbonyl group are non-polar (H, D, T, alkyl, phenyl).
In Tables 12 and 13 we see that all aldehyde-l-d-(R-CDO) reductions
performed up to now with either purified ADH and NADH or fermenting
yeast give 100% or nearly so stereospecificity by yielding the (S)-
isomer. As we will see later this is true even if the R substituent
contains polar groups22 (lactaldehyde). For ketones (RlCORZ) the FakfiJ
highest (98-100%)stereospecificity reported involves 2-octanone (Y-ADH),
with phenyl alkyl ketones next (69-90% stereospecificity by yeast) and

 

dialkyl ketones last (12-82% stereospecificity by yeast). In the latter E3
ketones the bigger the difference between the two substituents the greater

the stereospecificity (2-octanone vs, n-propyl n-butyl ketone), provided

that there is no branched substituent present because then the reduction

does not proceed. In view of these facts we conclude that the very low
stereospecificity obtained in the case of chloroacetone cannot be rational-

ized on steric grounds, because then we would expect a stereospecificity

close to that of methyl ethyl ketone. So, by replacing the ethyl group

with a (polar) chloromethyl group the stereospecificity is reduced con-

siderably, to the extent that the enzymatic reduction of chloroacetone, \
a ketone with two considerably different groups, yields almost racemic

product, a fact that is very rare for an enzymatic reaction. The very

low stereospecificity of substrates like ethyl n-propyl ketone, ethyl

n-butyl ketone and the complete lack of stereospecificity of n-propyl

n-butyl ketone is not surprising because of the great similarity of their

substituents.

57

F. Hydrophilic-hydrophobic substrate-enzyme interactions

 

We discussed in our introduction (p. 10-11) how Karabatsos et al.,
using lactaldehyde as substrate, demonstrated that as far as product
stereospecificity is concerned, lactaldehyde (either D or L) and acetal-
dehyde have the same substrate-coenzyme spatial relationship (X and VIII,

p. 10, respectively). In Table 14 we summarize the data for the known

Table 14. Reductions of carbonyl compounds possessing at least one
polar group by purified enzyme and coenzyme (NADH).

 

 

 

 

RICOR2 +. RICH(0H)R2
Carbonyl Compound Enzyme Stereospecificity % AAGf 30° Ref.
cal/mol
R1 R2 ([aJObs/[GJmax) XIOO
-CH2-0H -CH3 Gly-DH (R) 100 31
-CH3 -CH2C1 L-ADH, Gly-DH (S) 6.8 i .4 85 i 9
-D -CH-CH3 (D) L-ADH (S) 100 15.23
CH \
-0 -CH-CH3 (L) L-ADH (5) 100 15.23 \‘
0H
-CO0H -CH3 Y-ADH (R) 100 16

 

 

 

 

 

 

 

 

reductions of bifunctional substrates with purified enzyme and coenzyme
systems. From the spatial substrate-coenzyme relationships indicated
for hydroxyacetone (2) and pyruvic acid (4) we conclude that the hydro-
philic hydroxymethyl and carboxyl groups overcome the steric effect and
completely invert the spatial relationship predicted by Model K.

58

 

 

As far as lactaldehyde is concerned, we can rationalize the result on the
basis that the difference is size between H and -gH-CH3 is so great that
the steric factors predominate, keeping the same sgatial relationship

as the corresponding groups in acetaldehyde. In fact,none of the alde-

hydes known to be reducted enzymatically53

with pyridine nucleotide depen-
dent DH's violate Model K. Since the very low stereospecificity of chloro-
acetone (1) cannot be explained on the basis of steric enzyme-substrate
interactions, we have to consider that the chloro-group is responsible

for the remarkably low stereospecificity exhibited by this enzymatic

reduction.

59

G. Steric vs. hydrophilic-hydrophobic interactions

Based on configurations 1; 2) 3; 4 and X (p. 10) we postulate that
for bifunctional substrates the hydrophilic substituent of the carbonyl
compound tends to occupy the location of the small group in Model K,
with the hydrophobic group taking the place of the large group. Thus,
hydrophilic-hydrophobic substrate-enzyme interactions may, as in 2 and
4, invert the substrate-coenzyme spatial relationship (predicted by
Model K). However, in the case of X, the steric factor is so big
(H vs, -CH-CH3) that the hydrophilic-hydrophobic interactions are not
sufficient to invert the predicted relationship. In the case of chloro-
acetone (1) the hydrophilicity of the chloromethyl group, compared to
that of hydroxymethyl or carboxyl, appears to be insufficient to invert
the spatial relationship, but its influence leads to an almost racemic
product. We explain then the very low stereospecificity of chloroacetone
as being a competition in which the two opposing interactions, namely,
steric and hydrophilic-hydrophobic enzyme-substrate ones, have almost
equal influence, with the steric interactions predominating slightly.

Throughout this thesis we have assumed that the carbonyl is pointing
"down" (the oxygen of the carbonyl is toward the nitrogen of the nico-

tinamide) according to Kosower“.

Of course, this is only one possibility
and we cannot exclude other orientations such as the carbonyl pointing

“up" or to any other direction. It is important to note here that our
postulate, concerning the relation between steric and hydrophilic-hydro-
phobic interactions is irrelevant to the direction in which the carbonyl

is pointing. The solution to this problem will naturally help to determine
more accurately the location of the hydrophilic and hydrophobic regions

of the enzyme with respect to the coenzyme54

60

CONCLUSION

 

In conclusion, we think that our work contributes to a better under-
standing of the factors controlling product stereospecificity in the
enzymatic reductions of carbonyl compounds with pyridine nucleotides
dependent dehydrogenases. From the known facts, and the new facts we
have contributed, we think that in order to establish the correct spatial
coenzyme-substrate relationship one must bear in mind both steric factors
and hydrophilic-hydrophobic regions of the enzyme. For all aldehydes
and for ketones whose substituents are non-polar, like alkyl or phenyl
groups, steric factors play the important role for determining stereo-
specificity and the major product can be predicted by a simple Model (K,
for A-type enzymes, P, for B-type enzymes). For ketones whose substituents
contain polar groups, both factors play a role with the hydrophilic-hydro-
phobic factor being extremely important when dealing with substrates pos-
sessing very polar groups. For these substrates the hydrophilic group
occupies the space of the small group and the hydrophobic group occupies
the space of the large group.

Finally, we would like to add that the problem of the polyfunctional
substrates is not yet clear because most of the facts available today come
from phytochemical or yeast reductions and the enzymes and coenzymes in-
volved cannot be safely assumed.

With hopes that this work will inspire some investigators to use
pure enzymes as conventional reagents and that more light will be given
soon to the active site of the ADH enzymes, either because of this work
and more to be done in this laboratory, or from X-ray crystallography,

we close this thesis.

10.
11.
12.

13.
14.

15.

BIBLIOGRAPHY

 

J. S. Fruton, "Molecules and Life", Wiley-Interscience, pp. 353-5
(1972).

F. H. Hestheimer, H. F. Fisher, E. E. Conn, and B. Vennesland,

 

J. Amer. Chem. Soc., 12’ 2403 (1951).
K. R. Hanson, J. Amer. Chem. Soc., 88, 2731 (1966).

 

H. R. Levy, F. A. Leowus and B. Vennesland, J. Amer. Chem. Soc., 12,

 

2949 (1957).

 

R. U. Lemieux and J. Howard, Can. J. Chem., 41, 308 (1963).
J. N. Cornforth, Bioch. Bioph. Res. Comm., 2’ 371 (1962).

 

R. Bentley, "Molecular Assymmetry in Biology", Vol. II, pp. 6-9,
Academic Press, (1970).
A. G. Ogston, Nature, London, 162, 963 (1948).

W

V. Prelog, Ind. Chim. Belge, 21, 1309 (1962).

 

V. Prelog, Pure Appl. Chem., 9(1), 119 (1964).

 

E. M. Kosower, Bioch. Bioph. Acta, 56, 474 (1962).

 

F. A. Loewus, F. H. Hestheimer and B. Vennesland, J. Amer. Chem. Soc.,

 

75, 5018 (1953).
M
V. Prelog, Chem. Nat. Prod., 3, 119 (1964).

 

H. Sund, "Biological Oxidations", Singer (Ed.) Intersciences Publ.
(1968), pp. 675-8.
G. J. Karabatsos, J. S. Fleming, N. Hsi, and R. H. Abeles, J. Amer.

Chem. Soc., 88, 849 (1966).

61

16.
17.

18.

19.

20.
21.

22.

23.

24.
25.
26.
27.
28.
29.
30.

31.
32.
33.

62

J. Van Eys and N. 0. Kaplan, J. Amer. Chem. Soc., 22, 2782 (1957).

 

V. E. Althouse, K. Ueda, and H. S. Mosher, J. Amer. Chem. Soc.,

 

82, 5938 (1960).

’VI;

R. MacLeod, H. Prosser, L. Fikentscher, J. Lanyi, and H. S. Mosher,
Biochem. 2, 838 (1964).

V. E. Althouse, D. M. Feigl, N. A. Sanderson, and H. S. Mosher,

J. Amer. Chem. Soc., 22, 3595 (1966).

 

C. Donninger, and G. Ryback, Biochem. J., 22, 11P (1964).

 

H. Gunther, F. Biller, M. Kellmer, and H. Simon, Angew. Chem.

 

Int. Ed. Vol. 12, 146 (1973).
P. A. Frey, G. J. Karabatsos, and R. H. Abeles, Bioch. Bioph.

 

Res. Commun., 18, 551 (1965).
V1.1

 

B. Zagalak, P. A. Frey, G. J. Karabatsos, and R. H. Abeles, J. Biol.
Chem., 241, 3028 (1966).

—'—’\/VD

C. Neuberg, Advan. Carbohyd. Chem., 1, 75 (1949).

 

C. Neuberg and F. F. Nord, 855., 222, 2248 (1919).
P. A. Levene, and A. Halti, J.Biol. Chem., 22, 361 (1933).

 

 

. Grzycki, Biochem. Z., 222, 195 (1933).
Fujise, Biochem. 2., 222, 231 (1931).

 

”MU,

U. Lemieux, and J. Giguerre, Can. J. Chem., 22, 678 (1951).

 

P. A. Levene and A. Nalti, in "Organic Synthesis" (A. H. Blatt, Ed.)
C011. V01. 2, p. 545, N11ey, N. Y., (1943).

G. J. Karabatsos and H. Nunez, unpublished results (1971).

E. C. C. Lin and B. Magasanic, J. Biol. Chem., 222, 1820 (1960).

 

N. 0. Kaplan and M. M. Ciotti, J. Biol. Chem., 222, 823 (1956) and

 

references cited therein.

34.
35.

36.

37.

38.
39.

40.
41.

42.

43.

44.

45.

46.

47.

48.
49.

50.

63

H. Zimmer and H. Fiedler, Arch. Pharm., 291/63, 330 (1968).
'VWVVD

 

C. C. Sweeley, B. 0. Ray, W. J. Wood, J. F. Holland, and M.
Kritchersky, Anal. Chem., 32, 1505 (1970).

 

F. L. J. Sixma and H. Wynberg, "A Manual of Physical Methods in
Organic Chemistry", John Wiley and Sons, Inc., (1964), pp. 172-3.
R. H. Pickard and J. Kenyon, J. Chem. Soc., 103, 1937 (1913);

 

99, 45 (1911).
V);
M. M. Sprung and E. S. Wallis, J. Amer. Chem. Soc., 22, 1717 (1934).

 

E. W. Hoffman, R. J. Ess, and R. P. Usinger, Jr., J. Amer. Chem. Soc.,

 

18. 5820 (1955).
S. W. Kantor and C. R. Hauser, J. Amer. Chem. Soc., 22, 1744 (1953).

 

A. W. Ingersoll, "Organic Reactions" Vol. II, John Wiley and Sons,
Inc., N.Y. (1944), pp. 400-404.
J. Nakajima, S. Suga, T. Sugita and K. Ichikawa, Bull. Chem. Soc.

 

032., $2, (12), 2980 (1967).
W. Fickett, H. K. Garner and J. J. Lucas, J. Amer. Chem. Soc., 22,

 

5063 (1951).
R. H. Abeles, R. F. Hutton and F. H. Westheimer, J. Amer. Chem. Soc.,

 

79, 712 (1957).

M

C. Neuberg and F. F. Nord, Ber. 22, 2237 (1919).

See Ref. 18, p. 843.

T. C. Bruice and S. J. Benkovic, "Bioorganic Mechanisms“, Vol. II,
p. 309, W. A. Benjamin, Inc., New York, (1966).

E. J. Pastore and M. Friedkim, J. Biol. Chem., 22g, 2314 (1961).

 

G. Rafter and S. P. Colowick, "Methods in Enzymology" Vol. 2, p. 887
(1957).
For a discussion see ref. 7 pp. l9-22 and references cited therein.

51.

52.
53.

S4.

64

J. M. Siegel, G. A. Montgomery, and R. M. Bock, Arch. Biochem.

Biophys. 22, 288 (1959).
H. R. Levy and B. Vennesland, J. Biol. Chem., 222, 85 (1957).

 

 

G. Popjak, in "The Enzymes" (P. D. Boyer ed.), Vol. II, 3rd ed.,
Academic Press, New York, N.Y., 1970, pp. 140-3.

We have already started investigating this problem. Further
research is currently being conducted by J. Miedema and M. May

in Karabatsos' laboratory.

“III'IIIIIIIIIII