3 0 1

 

 

ABSTRACT

HYBRID COENZYME-SUBSTRATES
AS CONFORMATIONALLY RESTRICTED PROBES
OF DEHYDROGENASE ACTIVE SITES

BY

Jelmer And ries Miedema

Since the early discovery of the stereospecificity of enzymatic
reactions,1 many substrate-product studies have been performed with
a view towards rationalization of enzyme structure and reaction
mechanism. In the case of the alcohol dehydrogenases (E. C. 1. l. 1. l),
responsible for the nicotinamide adenine dinucleotide mediated inter-
conversion of hydroxylic and carbonyl substrates, much ingenious
experimental work by Prelog,2 Ringold,3 and others has gone into the
construction of composite aliphatic substrates in attempts to clarify
these points. Karabatsos4, Stamoudis5 and others have extended
these results to substrates in which hydrOphilic-hydrOphobic inter-
actions become significant.

Less work has been done regarding the orientation of the
resulting ”diamond lattice sections"2 with respect to the enzyme-
coenzyme binary complex. The solution of this problem is a necessary
complement to construction of these models if any mechanistic or

structural data derived from them are to be pr0perly interpreted.

Jelmer Andries Miedema

This interpretation will yield a predictive model of the binary complex-
substrate interaction and considerable stereochemical information
about the enzyme active site.

A model describing this interaction Cannot result from further
work with the conventional substrate-enzyme-coenzyme three-body
systems. Such systems permit far too many degrees of conforma-
tional freedom to allow the single transition state—-product correlation
implicit in the confirmation of any given model. Imposition of an
additional constraint would be required to make such a correlation
possible.

This is simply accomplished by physically linking the substrate
molecule to the coenzyme through an orientation restricting
methylene bridge. To this end the following hybrid coenzyme-sub-

strates were synthesized:

/ O O
c/ C” CH
\ C O \N I
N(CHZ)n H2 H (CI-12)n ('3 CH3
H H
H
N n=2,3,4,5 N n=3,4,5
+ I + | R: adenosine diphosphoribose
I II

These analogs of nicotinamide adenine dinucleotide showed full
activity in the horse liver alcohol dehydrogenase catalyzed oxidation

of ethanol, at a rate however 10-3-10"4 that observed in the natural

Jelmer Andries Miedema

system. None showed any reactivity with yeast alcohol dehydrogenase.

Hybrid coenzyme-substrates I(n =4, 5) and II(n =4, 5) also had
the ability to oxidize small but significant amounts of the attached
hydroxyl moiety.

Further studies of I(n = 4, 5) in semicarbazide hydrochloride
containing buffers indicated very significant effects on both Keq and
kinitial in the case of I(n =4) but only minor effects in that of I(n =5).
The meaning of this result is not yet clear.

An inhibition experiment gave strong but not conclusive
evidence for an intramolecular reaction in the case I(n =4) (i. e.
evidence that both the hydroxyl and coenzyme moieties participating
in the ternary complex were provided by the same hybrid coenzyme
analog). No conclusion regarding reaction mechanism could be drawn
from a similar experiment with I(n = 5), but models support the like-
lihood of an intramolecular mechanism in this case as well as the
previous. If these preliminary results are confirmed, I(n =4, 5) and

hOpefully also II(n =4, 5) will provide useful stereochemical probes of

the dehydrogenase active site.

Jelmer Andries Miedema

BIBLIOGRAPHY

E. Fischer and H. Thierfelder, Chem. Ber., 31, 2031 (1894).

 

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

 

H. J. Gingold, J. M. H. Graves, A. Clark, and T. Bellas,
Recent Progr. Horm. Res., _2__3_, 349 (1967).

G. J. Karabatsos, J. S. Fleming, N. Hsi, and R. H. Abeles,
Abeles, J. Amer. Chem. Soc., §§, 849 (1966).

 

Vassilios C. Stamoudis, Ph. D. thesis, Michigan State University,
(1973).

HYBRID COENZYME-SUBSTRATES
AS CONFORMATIONALLY RESTRICTED PROBES
OF DEHYDROGENASE ACTIVE SITES

BY

Jelmer And rie s Miedema

A DISSER TA TION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1977

TABLE OF CONTENTS

Page
LIST OF TABLES ............... . ......... iv
LIST OF FIGURES .................. v
INTRODUCTION . . . ....................... 1
EXPERIMENTAL .......................... 17
1. Synthesis .......................... 17
Preparation of potassium 3-pyridinecarboxy1ate . . . . 17
Preparation of 3-chloroformy1pyridine . . ....... 18
Preparation of potassium phthalimide .......... 18
Preparation of 5-chloro- Z-pentanone .......... 19
Preparation of 5-chloro-2-pentanol ...... . . . . . 20
Preparation of 2-(trimethylsiloxy)-5-chlor0pentane . . 21
Synthesis of 5-amino-2-pentanol— The
Gabriel Synthesis .................... 22
Preparation of N- (4-trimethylsiloxypenty1) -
phthalimide ..................... 2 2
Preparation of 4-trimethylsiloxypenty1-
ammonium phthalhydrazide ............. 23
Preparation of 5-amino-Z- pentanol . ....... 24

Preparation of 1-methy1cyclohexyl hydrOperoxide . . . 26
Preparation of 7-chloro-2-heptanone .......... 2.8
Preparation of 7-chloro-Z-heptanol ........... 29

Synthesis of 7-amino-2-heptanol —
The Gabriel Synthesis .................. 30

Preparation of N-(6 ~hydroxyhepty1)-phtha1imide. . 30

ii

Preparation of 6-hydroxyheptylammonium
phthalhydrazide. . . ..........

Preparation of 7-amino-2-heptanol . .

Synthesis of the N-(hydroxyalky1)-3-pyridine-
carboxamides......... ..... .....

Preparation of the trimethylsiloxyalkylamine.

Preparation of the N-(trimethylsiloxyalkyl)-
3-pyridinecarboxamide. . . . ...... . .

Preparation and isolation of the N-(hydroxy-
alkyl)-3-pyridinecarboxamide . . . . . .

II. Preparation and isolation of NAD+ analogs

III. High pressure liquid chromatographic monitoring
of analog synthesis . .

IV. Determination of analog activity in enzyme systems

V. Instrumentation. . . . . . . . . . . . . . . .....

RESULTS AND DISCUSSION.

1. Monitoring analog synthesis by high pressure
liquid chromatography .

II. Identification of the products of the analog synthesis . . .

III. The intramolecular reactivity of substrate-
coenzyme hybrids

BIBLIOGRAPHY

iii

Page

31
31

32
33

34

34

37

38
4O

43

46

46

50

53

66

Table

LIST OF TAB LES

Reactant aminoalcohols .

Melting points of solid N- (hydroxyalkyl)-
3- pyridinecarboxamides .

Buffer systems used in coenzyme analog
incubation with dehydrogenase . .

Qualitative effects of the addition of NAD+ and/or
semicarbazide HCl on substrate oxidation-

iv

Page
33

36

41

62

Figure

10.

11.

12.

13.

LIST OF FIGURES

Structure of nicotinamide adenine dinucleotide
Absolute configuration of NADH

Characteristic diamond lattice for HLAD,
after Prelog . . . . . ........

Modified diamond lattice for HLAD, A
after Ringold et al. , , , , . . . . .....

Prelog's model of carbonyl orientation
with respect to NADH. . . .

Kosower's model of the ternary complex

Graves' and Ringold's model of carbonyl
orientation with respect to NADH

Karabatsos' model of carbonyl orientation

withrespecttoNADH. . . . . . . . . . . . . .

Competition between hydrophilic-hydrophobic
and steric interactions for control of

substrate orientation. . ..... . . . . . . .

Tabulation of possible models of carbonyl

orientation with respect to NADH .......

Plot of percent analog synthesis vs. time .

Chromatographic isolation of pyridine

nucleotide II(n =5) . ..... . . . .......

Relative energy levels of the possible
reaction modes . . ........

00000

10

11

12

l3

13

49

51

63

INTRODUCTION

Since time immemorial the ability of organisms to interconvert
oxo and hydroxyl compounds has been the basis of and driving force
behind life. At first this process was not at all associated with
cellular metabolism, even in such a relatively obvious case as
fermenting yeast, but was ascribed to a mysterious and ill-defined
"geist". In fact, this important association was not made until
183 5-37,1 over one hundred and fifty years following van Leewenhoek's
initial microscopic observations of yeast and other cell life.

Once made, however, this association had a subtle and tena-
cious hold on scientific thought. It became a commonly accepted view
that intracellular chemical processes were mystically linked to the
”life force" of the cell and by definition could not occur outside the
cell. The strength of this belief was so great that its overthrow with
the 1897 demonstration of the enzymatic activity of a cell-free
(sterile) yeast cell extract was sufficient to win Edward Buchner the
1907 Nobel prize. The award was granted solely for the significance
of the experimental result; it was generally recognized that the work
involved only application of already well-known techniques.

This discovery removed much of the awe and mystery from

the processes of life. Interest in understanding these processes grew
apace with belief in their understandability. It was found that Buch-
ner's extract contained several components, one of which was termed
an alcohol dehydrogenase for its ability to oxidize alcohols and reduce
carbonyls through hydrogen and electron transfer (Batelli, Wieland,
and Stern, 1913).2 The significance of this type of enzyme became
clearly apparent when its presence was determined in tissues from
many sources, both plant and animal.

The structure of the coenzyme forming with the substrate the
reduction-oxidation couple catalyzed by dehydrogenase was postulated
in 1931 by Schenk and Euler and finally proved by synthesis in 1957

by Alexander Todd.3 This structure is shown in Figure 1.

O NHZ
C/
\
NH2 N
</ .N
N O O
CHZ— O-P-O—r- O-CHZ
9 o' o‘ 0

Figure 1. Structure of nicotinamide adenine dinucleotide.

The dehydrogenases and their associated coenzymes proved
to be important agents both in aerobic and anaerobic respiration and
in biological electron transport processes.4 Work on these enzymes
continued and quickly led to the, purification by crystallization of the
first alcohol dehydrogenase, that isolated from yeast (Negelein and
Wulff, 1937).5 Isolation of other dehydrogenases in pure form
followed and left work in the field at a point where efforts to clarify
the action of these enzymes could begin in earnest. Over the succeed-
ing years much information has been obtained, culminating in the
X-ray6 and partial amino acid sequence analyses7 of several
dehydrogenases. Some remarkable similarities have been noted
among these enzymes8 and the growing completeness of our under-
standing of their static structures will eventually lead to understanding
of their catalytic action.

As this research is concerned mainly with the dehydrogenase
isolated from horse liver, a summary of the presently known char-
acteristics of this enzyme is given below.

Horse liver alcohol dehydrogenase9 (E. C. 1. l. 1. 1.) is a
metallo-enzyme of wide substrate specificity utilizing Zn as a co-
factor and nicotinamide adenine dinucleotide as coenzyme. The
enzyme has a molecular weight of 80,000 and is composed of two
identical subunits. Although the separated subunits are inactive,

they provide two active sites when combined in the enzyme molecule.

The amino acid sequence of the subunits is known;7 as is the fact that
each contains two Zn atoms, one of which seems to have structural
and the other catalytic significance.6 The enzyme exists in as many
as twelve isoenzymatic forms which differ slightly in primary struc-
ture and in reactivity with acyclic versus steroidal substrates.
Crystal structure data to 2.919 resolution are available for one of the
isoenzymes.

Shortly following the initial crystallization of horse liver
alcohol dehydrogenase in more or less pure form in 194811 (the
existence of the isoenzymatic forms was not recognized for another
ten years), kinetic studies by H. Theorell and B. Chance established
the enzyme as ordered bi-bi in its action; that is the enzyme binds
coenzyme, then substrate, and releases product and then coenzyme

1
in obligatory order. 2 This reaction sequence is represented by

equation 1.

ROH+ENZ +NAD+:: ROH+ENZ::NAD+:: ROH::ENZ::NAD+:: (1)
C = O::ENZ::NADH+H+:: c = O+ENZ::NADH:: c = O+ENZ +NADH

Almost simultaneously with this clarification of the kinetics
of the catalysis some details of the mechanism became known.
Through the use of deuterium labeled substrate and coenzyme, West-
heimer, Vennesland and coworkers were able to show that the hydro-
gen transfer step of the redox equilibrium was both direct13 and

1
stereospecific 4 with respect to coenzyme and substrate. The

structure of the reduced coenzyme was as yet unknown, but these
studies demonstrated that the hydrogen (deuterium) atom leaving the
hydroxylic substrate was ultimately bonded to the reduced coenzyme
and that this same atom was transferred in the reverse step to yield
an alcohol stereoisomerically identical to the original. Any possi-
bility of participation by solvent protons or exchange with other
components of the system was thus eliminated.

Although the actual structure of the reduced coenzyme was
uncertain, it was known that reduction took place on the nictotinamide
moiety. The three positions ortho and para to the ring nitrogen which
could conceivably undergo reduction were similar in that all would
yield a methylene bearing diastereotopic hydrogens. That is, in all
three cases if the individual members of the pair of protons bonded
to the methylene were alternately replaced by any other atom, the
two new molecules generated would be neither identical nor mirror
images of each other. Clearly then, though bonded to the same
carbon, these protons were distinct and to be consistent with the
observations previously mentioned an enzyme could utilize only one
of the two protons. Thus the dehydrogenases were classified in two
sub-groups based on the proton utilized. Lacking any real structural
information these were arbitrarily labeled as type A and type B
enzymes. A fairly comprehensive listing of dehydrogenases by type

is given in reference 15. Horse liver alcohol dehydrogenase is of

type A.

In 1953 the methylene of the reduced coenzyme was shown to
be located in the position para to the ring nitrogen.16 It was another
nine years before the absolute configuration was determined and the
resulting assignment of the A and B protons made. After applying a
difficult enzymatic cleavage and chemical breakdown procedure,

Cornforth, Popjak, and coworkers arrived at the following structure.17

 

Figure 2. Absolute configuration of NADH.

With the elucidation of these basic details of the enzyme's
mechanism, attempts to probe the active site of this and other
dehydrogenases began to intensify. Initially this work focused on
attempts to determine the steric dimensions of the active site by
the construction of ”diamond lattice sections“. 18’ 19’ 20

In the case of the dehydrogenase isolated from Curvularia
falgata, V. Prelog approached this problem by determining the
stereoisomeric distribution of products arising from the reduction

1
of the isomeric decalin-l, 4-diones. 8 By aligning those products

forming at measurable rates in such a way that the CRZHOH tetrahedra

were coincident Prelog was able, by virtue of conformational
restrictions on the decalin skeleton, to sketch out the apparent
volume which could be occupied by substrate at the active site. By
also taking into consideration the differing rates of reaction of the
various substrates he was able to identify ”forbidden” positions
which if occupied would greatly reduce substrate reactivity. Similar
work reported in the same paper with horse liver alcohol dehydro-
genase as catalyst and methyl substituted cyclohexanols and decalin
derivatives as substrates resulted in the diamond lattice shown in

Figure 3.

V‘.

Figure 3. Characteristic diamond lattice
for HLAD, after Prelog.
This work was later greatly elaborated by Ringold and col-
laborators. These workers arrived at the modified diamond lattice
section for horse liver alcohol dehydrogenase shown in Figure 419

and performed detailed studies on the 3- hydroxysteroid dehydro-

genase isolated from E; testosteroni.

 

Figure 4. Modified diamond lattice for
HLAD, after Ringold e_ta_1..

Still, the precise significance of these lattice sections
remained somewhat obscure. If the enzyme active site were a rigid
structure these models could be considered simply as defining the
volume at the catalytic site not occupied by either solvent, bonded
ions, or protruding members of the enzyme protein. This explana-
tion clearly did not account for the possibility of the enzyme under-
going consecutive changes in conformation with the binding of
coenzyme and substrate molecules, changes which perhaps were not
independent of the substrate bound. Although this hypothesis has yet
to be completely confirmed or denied, it is implicit in the Theorell-
Chance mechanism and the enzyme is known to lose a symmetry
element on binding of nicotinamide adenine dinucleotide.6 The
likelihood of similar conformational changes upon substrate binding

cannot be ignored.

The existence of these diamond lattice models did, however,
suffice as a starting point for speculation regarding the orientation of
the carbonyl bond axis with respect to the coenzyme, or to be more
precise, with the coenzyme-enzyme binary complex. On the basis of
his work with the type B dehydrogenase associated with Curvularia
falgata, Prelog postulated the following steric relationship between

1
substrate and coenzyme. 8

 

Figure 5. Prelog's model of carbonyl orientation
with respect to NADH.

He supported this model by stating that this orientation, with
the larger substituent further from the bulky carboxamide and both
substituents neatly straddling the methylene bridge, minimized non-
bonded interactions in the transition state. Prelog carefully noted
that it was the structure of the enzyme protein which would play the
decisive role in the stereochemistry of this complex, thus acknowl-

edging both the probability of hydrOphilic-hydrophobic interactions

10

with the enzyme protein and the primacy of enzyme-substrate (not
coenzyme-substrate) steric interactions.

This model generally supported another, developed on inde-
pendent evidence, by E. M. Kosower.21 It had previously been noted
that the 340nm absorption peak of reduced nicotinamide adenine di-
nucleotide shifted to a lower wavelength, 325nm, upon binding of the
coenzyme to horse liver alcohol dehydrogenase. Kosower rationalized
this observation by postulating the presence of a quaternary nitrogen
about 31: from the nitrogen of the dihydrOpyridine ring. Model com-
pound studies indicated that such a structure would indeed shift the
energy of the excited state dipole relative to the ground state species
sufficiently to account for the experimental result. Incorporating this
ammonium ion, he hypothesized the following ternary complex with
the ion covalently linked to the enzyme protein and hydrogen bonded

to both substrate and coenzyme ribose.

H CH3 hydmphobic

\ / region
Ell/l C
’O H H H
\\c o
\‘
/ s
HZN I H
| 1. .
“Q
‘H O
O
\ J \
OH 0' \
9“ on ”/ §H
\ \ Zn
E \\\\ \ I],
E \\“ ’I/

Figure 6. Kosower's model of the ternary complex.

11

Clearly this model supported Prelog's conclusions regarding
the orientation of the carbonyl, but again no inference regarding the
relative positions of the carbonyl substituents could be drawn unless
further assumptions regarding non-bonded interactions and hydrophilic
and/or hydrophobic enzyme regions were made.

A third model, offered by Graves and Ringold,19 completely
contradicts the previous two. On the basis of their work with horse

liver alcohol dehydrogenase they arrived at the following structure.

 

Figure 7. Graves' and Ringold's model of
carbonyl orientation with respect
to NADH.

It is interesting to note that all of these models could predict
the same product structure in a given reduction. Thus it becomes a
doubly worthwhile problem to test the correctness of one or another
of these models. This question is complicated by the fact that it has
become clear that enzyme-substrate interactions do take precedence
over interactions with coenzyme and so the model may be different

for every enzyme. This is most succinctly shown by noting the

12

existence of two A type lactic acid dehydrogenases, one specific for

d-lactic acid and the other for l—lactic acid.22 It is also reflected in
23 . 24

the work of Karabatsos, Stamoudis, and Nunez who demonstrated

the fallacy of applying Prelog's model for g. falcata to alcohol de-

hydrogenase. They arrived at the model shown in Figure 8; the

Opposite of that of Prelog.

 

Figure 8. Karabatsos' model of carbonyl
orientation with respect to NADH.

In addition, through studies utilizing polar hydrogen bonding sub-
strates, these workers showed the significance of hydrophilic-
hydrophobic interactions and the ability of such forces to counter-
balance and even overwhelm those resulting from steric interactions.
If, as in Figure 9, one of the carbonyl substituents is sufficiently
hydrOphilic, the relative orientation predicted on steric grounds

may be completely reversed.

 

100% (ref. 22)

13

 

100% (ref. 25)

Figure 9. Competition between hydrophilic-
hydrophobic and steric interactions
for control of substrate orientation.

All these models seem reasonable on the basis of the evidence

offered in their defense.

constitute a proof of any given model.

when considering Figure 10.

 

Figure 10.

0:

2b

This is

3a

o_-__-c

3b

In no sense however, does this evidence

clea rly apparent

\ _
C_.O
/

4a

1.
\czo
S /

4b

Tabulation of possible models of carbonyl

orientation with respect to NADH.

14

L and S represent any two different substituents. 1, 2, 3, and 4a will
yield the same product and 1, 2, 3, and 4b its enantiomer. The sigma
spZ bonds are shown as projections on the dihydrOpyridine ring and
the C =O bond axis is shown aligned with the coordinate on which its
projection is longest. Coplanarity is not implied.

These eight models represent the possible orientations of the
carbonyl with reSpect to the enzyme-coenzyme binary complex. Each
represents an approximation to reality, for of course an infinite num-
ber of actual alignments are conceivable. Even ignoring this detail,
at most two possible products (enantiomeric alcohols) may be formed
from each carbonyl undergoing reduction. Thus a one to one cor-
respondence between product and model does not exist and it follows
that the choice of model cannot be made on the basis of product
analysis alone. This conclusion extends to the diamond lattice studies
of Prelog and Graves, based merely on the composites of individual
product studies, and to all other known work in this field. Quite
simply, not only is the previously available evidence insufficient to
come to any conclusion regarding the structure of a ternary complex
but the methods us ed to obtain the evidence are not suitable to this end.

Our incomplete knowledge of the enzyme active site can be
expanded only by the application of different techniques. In recent
years the X-ray studies of Branden,6 Ros sman,8 and other workers,

the work of Weiner and Mildvan with Spin labeled analogs of

15

nicotinamide adenine dinucleotide}6 and Takahashi'sZ7 analysis of
the fluorescence properties of Co-modified horse liver alcohol
dehydrogenase provide examples.

This thesis describes the development of another potentially
useful probe of the active site. Consideration of the weaknesses of
the product analysis based probes described previously quickly leads
to apprehension of the common flaw. In all cases the experiment is
carried out such that the ternary complex has an excessive number
of degrees of freedom. In consequence a single rationale of the
experimental data, and thus a single model, cannot be given. The
solution of the problem self-evidently lies in the removal of the
excess degrees of freedom. In principle this can be accomplished
simply by binding the reacting Species together and this has been the
approach taken in this instance.

Specifically, this research project has involved the synthesis
of and initial reactivity studies on analogs of nicotinamide adenine
dinucleotide in which the hydroxylic substrate is incorporated into
the coenzyme molecule. To the best of our knowledge this is the
first research undertaken involving such hybrid coenzyme-substrate

systems. The structures synthesized are shown on the folIowing

page.

O

:nz/‘é

(CHQnCH

n=2, 3,4,5

2

-OH

16

O
/
cg: 9H
H(CHZ)n C'I — CH3
H
N n=a4¢5
H
R R=adenosine

diphosPhoribose
II

EXPERIMENTAL

I. Synthesis

Preparation of potassium 3-pyridinecarboxylate

 

To a 2000ml three-necked round -bottomed flask equipped with
a mechanical stirring apparatus, powder funnel, and reflux condenser
was added 1200ml of 95% ethanol. With stirring and some cooling,
60g (1.07mol) potassium hydroxide was dissolved in the alcohol.
On complete dissolution of the base, stirring was continued and 123.1g
(1.00mol) nicotinic acid (Sigma Chemical Co.) was added. The flask
was then stoppered and brought to reflux. Reflux was continued for
one half hour, the product was cooled to room temperature, and the
clear solution was transferred to a 2000ml beaker. The solution was
cooled further to 0°, allowed to stand at this temperature overnight,
and the white crystalline product was filtered. After washing with
cold ethanol and oven drying at 120° a first crOp of 92g (0.571mol,
57% yield) of potassium 3-pyridinecarboxylate was obtained. Upon
reducing the total volume to 250ml a second crOp of 23g was obtained.

The overall yield was 115g (0.714mol, 71% of the theoretical amount).

17

18

Preparation of 3 - chloroformylpyridine28

To a 1000ml three—necked round-bottomed flask equipped with
a mechanical stirring apparatus, reflux condenser terminated with a
gas bubbler, and a 50ml addition funnel was added 400ml of carbon
tetrachloride. SuSpended with stirring in the carbon tetrachloride
was 64.8g (0.402mol) of potassium 3-pyridinecarboxylate which had
been dried under vacuum for 24 hours at 115°. The flask and addition
funnel were thoroughly flushed with dry nitrogen, 36.0ml (59.7g,
0.501mol) thionyl chloride (Fisher Scientific, purified) was pipetted
into the closed addition funnel, and the system was closed. The
thionyl chloride was added dr0pwise with stirring to the suspension
and the exothermic reaction was cooled as necessary. On completion
of the addition the mixture was heated to reflux. With the cessation
of SO2 evolution the bubbler was replaced with a drying tube and the
reflux was continued overnight. The solution was then cooled,
filtered, and the carbon tetrachloride was removed in a rotary evap-
orator. The crude material was refiltered. Distillation under
reduced pressure yielded 41.26g (0.29lmol, 72% theoretical) of
3-chloroformylpyridine, bp. 70° at 4-5mm. The reported bp. was

98- 100° at 21mm.

Preparation of potassium phthalimide2

 

The phthalimide used in this preparation was a very crude

technical grade product (Eastman Organic, practical). As a

19

preliminary step, 100g (0.68mol) of the crude phthalimide was re-
crystallized from 2000ml 95% ethanol.

The recrystallized material was redissolved in 2000ml fresh
95% ethanol boiling in a 3000ml Erlenmeyer flask on a magnetic
stirrer-hot plate. To the hot solution was added 46g (0.82mol)
potassium hydroxide in 300ml 95% ethanol. The combined solutions
were quickly mixed and immediately plunged into an ice bath. By
replenishing the ice as necessary, the temperature of the solution
was brought below 10° as quickly as possible. The product pre-
cipated immediately and was removed from solution by filtration,
then washed twice with 100ml portions of 95% ethanol and a 100ml
portion of acetone. After drying at 115°, 95g (0.51mol, 75% of

theoretical) of potassium phthalimide was obtained.

Preparation of 5- chloro- Z-Jaentanone

 

5-chloro-2-pentanone was prepared from 2-acetylbutyrol-

actone (Aldrich) by a procedure submitted to Organic Syntheses by

 

G. W. Cannon, R. C. Ellis, and J. R. Leal. 'The reaction was
carried out on a four molar scale. Vacuum distillation yielded 390g
(3.23mol, 81% of theoretical) of 5-chloro-2-pentanone, bp. 60-61°

at 10mm. The reported bp. was 70-72° at 20mm.

 

 

20

Preparation of 5- chloro - Z-pentanol

To a 500ml three-necked round-bottomed flask equipped with
a magnetic football stirrer, thermometer, and 100ml addition funnel
were added 175ml 95% ethanol and 28.50ml (30.14g, 0.250mol)
5-chloro-2-pentanone. The solution was cooled to 0° with stirring
in an ice slush.

During the cooling period a solution of sodium borohydride
(4.72g, 0.125mol, Metal Hydrides, Inc.) in 60ml of distilled water
was prepared. This was added dr0pwise to the stirring ketone solu-
tion at a rate to keep the reaction temperature below 20°. . On com-
pletion of the addition the solution was stirred vigorously for 12-14
hours at 0°.

The reaction mixture was titrated to neutrality with 20%
sulfuric acid, again with cooling, and the precipated salts were
removed by filtration. The solids were washed with 50ml of 95%
ethanol which was then combined with the filtrate. Distillation under
reduced pressure removed the bulk of the ethanol and some water
(28-30° at 85mm). The residue was washed into a 250ml separatory
funnel with 25ml ether to which was added sufficient extra water to
provide an aqueous phase of about 100ml volume. The organic layer
was separated and the remaining aqueous phase extracted twice with
50ml portions of ether. The combined ether fractions were washed

with 50ml saturated sodium chloride solution and dried over potassium

21

carbonate. After removal of the ether by rotary evaporation the

NMR Spectrum of the product showed no starting material and only
minor detectable impurities. Although distillation proved unnecessary
and is not recommended due to the ready cyclization of the product,

a bp. of 30° at 0.4-0.5mm was obtained. The reported bp. is 66-68°
at 3mm.31 The yield of crude 5-chloro-2-pentanol was 27.92g

(0.228mol, 91% of theoretical).

Preparation of 2- (trimethylsiloxy) - 5-chlor0pentane

 

An apparatus consisting of a dry 1000ml three-necked round-
bottomed flask equipped with a mechanical stirrer, thermometer, and
100ml addition funnel terminated with a drying tube was prepared and
flushed with dry nitrogen. To the flask were added 700ml anhydrous
ether, 27.92g (0.228mol) 5-chloro-2-pentanol, and 35.5ml (30.4g,
0.280mol) chlorotrimethylsilane (Aldrich). The solution was cooled
to 0° with stirring in an ice slush bath.

The addition funnel was closed, flushed with nitrogen, and
filled with 50.0ml (36.3gm, 0.359mol) of triethylamine (Mallinckrodt)
which had been dried over potassium hydroxide pellets. With the
flask remaining in the ice bath, the triethylamine was added dr0pwise
with stirring such that the reduction temperature did not exceed 20°.
Since the reaction was accompanied by formation of a copious pre-
cipitate, the stirrer speed was adjusted to maintain proper mixing.

On completion of the addition the reaction mixture was stirred for

22

an additional hour at room temperature.

The precipitate was removed by filtration and washed with
100ml anhydrous ether which was then combined with the filtrate.
The ether and most of the residual triethyl amine and chlorotri-
methylsilane were removed by rotary evaporation. This procedure
was accompanied by more precipitation and the solution was again
filtered. The solid was washed with 25ml of anhydrous ether which
was then added to the filtrate. Further rotary evaporation yielded
a crude product which showed no remaining hydroxyl and a strong
absorption in the tetramethylsilane region of the NMR Spectrum.
The 2-(trimethylsiloxy)-5-chlor0pentane was not purified but used

immediately to minimize hydrolysis losses.

Synthesis of 5-amino-2-pentanol -— The Gabriel Synthesis32

 

Preparation of N-(4-trimethylsiloxypentyl)-phthalimide--To a

 

500ml three-necked round-bottomed flask equipped with a stirbar,
thermometer, serum cap, and condenser terminated by a gas bubbler
was added 45g (0.243mol) of potassium phthalimide. This was fol-
lowed by addition of a solution of the crude 2-(trimethylsiloxy)-5-
chloropentane (0.228mol assumed) in 200ml of dimethyl formamide
that had been dried by storage over molecular Sieve type 4A (J. T.
Baker). The system was flushed with dry nitrogen, and the serum
cap and bubbler were replaced with a glass stopper and drying tube

respectively. The reaction mixture was stirred and heated gradually

23

to reflux at about 125°. After three hours of reflux the flask was
cooled to room temperature.

The contents of the flask were poured into a 2000ml separatory
funnel and diluted with 1000m1 distilled water and 300ml chloroform.
Shaking brought about the dissolution of all solids. The organic
layer was removed and the aqueous phase was re-extracted twice
with 100ml portions of chloroform. The chloroform layers were
combined and washed with 200ml of 0.1N sodium hydroxide in order
to remove excess phthalimide. Washes with 150ml distilled water
and 150ml saturated sodium chloride followed. The solution was
dried over anhydrous potassium carbonate and the chloroform re-
moved by rotary evaporation. The crude N- (4-trimethy1siloxypentyl)-
phthalimide was recovered as a yellowish oil. Some hydrolysis to
N-(4-hydroxypentyl)-phthalimide was inevitable and of no consequence.
Purification was not attempted.

Preparation of 4-trimethylsiloxypentylammonium phthal-

 

hydrazide -- To a 1000ml Single-necked round-bottomed flask equipped

 

with a stirbar and condenser were added the crude N-(4-trimethyl—
Siloxypentyl)-phthalimide (0.228mol assumed), 11.5m1 (11.85g,
0.237mol) hydrazine hydrate (Matheson, Coleman, and Bell), and
500ml methanol. The reaction mixture was refluxed for one and a
half hours. Upon cooling to room temperature a white asbestos-like

s olid was deposited.

24

The flask contents were cooled to 0° and the solid was filtered
and saved. Rotary evaporation of the methanol from the filtrate
resulted in the progressive precipitation of more product which was
periodically collected by filtration. No attempt was made to purify
the 4-trimethylsiloxypentylammonium phthalhydrazide from the
combined solids and the small amount of oily residue remaining after
the removal of the methanol.

Preparation of 5-amino-2-pentanol -- To a 1000ml Single-necked

 

round-bottomed flask equipped with a stirbar and condenser was added
the 4-trimethylsiloxypentylammonium phthalhydrazide (0.228mol
assumed) dissolved in 500ml distilled water. After stirring was
begun, 21ml of concentrated hydrochloric acid (8.31g HCl, 0.251mol)
was added to the clear yellow-green solution. A heavy precipitate of
free phthalhydrazide was immediately evident.

In order to complete hydrolysis of the silyl ether the mixture
was heated and then refluxed gently for two hours. After the mixture
was cooled to room temperature the phthalhydrazide was removed
by filtration, resuspended in 200ml distilled water, and refiltered.
The combined filtrates were subjected by rotary evaporation to reduce
the total volume to about 200ml. Any additional precipitate was again
removed by filtration.

The hydrochloric acid and salts in solution were neutralized

by the addition of 10g (0.250mol) of sodium hydroxide. Additional

25

acid or base was added as required to bring the pH of the reaction
mixture to 10. The mixture was poured into a 500ml separatory
funnel and twice extracted with 50ml portions of chloroform to remove
discoloring side-products. These chloroform layers were discarded
and the aqueous phase was made 30% in base by the addition, with
cooling, of 60g of sodium hydroxide.

This solution was extracted with 100ml portions of chloroform.
A few drops of each chloroform extract were evaporated on filter
paper and checked for amine concentration with a ninhydrin Spray
reagent (Nin-Sol®, Pierce Chemical Company). After eight to ten
extractions, the intensity of the developed color fell and the com-
bined extracts were dried over anhydrous potassium carbonate.
Following filtration the chloroform was removed by distillation
through a fifteen inch asbestos-wrapped Vigreaux column. Especially
in the latter stages care was taken not to overheat the product nor
entrain it with distilling solvent. Gradual application of a vacuum
was helpful. The product distilled at 56-59° at 0.2-0.4mm. The
reported bp. is 80-81° at 1mm. 31 The yield was 14.5g (0.14lmol,

62% of the theoretical based on 5-chloro-2-pentanol).

Anal. Calcd. for C H1

5 NO: C, 58.21; H,12.70; N,13.58.

3
Found: C, 58.06; H,12.81; N,13.40.

26

Preparation of 1-methylcjclohexyl hydroperoxide33

 

To a 2000ml three-necked round-bottomed flask equipped with
a long condenser, thermometer, and slit serum cap was added 1200m1
(923g, 9.40mol) of methylcyclohexane (Fisher Certified). A fritted
glass gas diffusion tube was inserted through the serum cap, which
should fit snugly, and connected to an air supply which was largely
dried and freed from carbon dioxide by passage through a drying
tower containing layers of potassium hydroxide and indicating
Drierite® (W. A. Hammond). The liquid was heated to 98-100° and
its level was marked on the flask. Air was then bubbled through the
material at a rate such that the aerosol dr0plets visible in the con-
denser air column seemed, as nearly as possible, neither to rise
nor fall. The rheostat controlling the heating mantle was adjusted
to maintain the required temperature, and then both rheostat and
flow rate were periodically readjusted until equilibrium was attained.
The reaction was allowed to continue for seven days, with additional
methylcyclohexane being added as required to maintain approxi-
mately the marked volume. No more than 100ml Should be lost in
the initial twenty-four hours and the rate of loss Should drop as the
reaction proceeds.

During the latter half of the reaction period the hydroperoxide
concentration was monitored by iodometric titration.34 (See method

of Wagner, Smith, and Peters.) The final hydroperoxide content

27

reached 1.06meq/ml. When no further increase in hydroperoxide
titer was noted, air flow was halted, a stirbar was added, the
condenser was replaced with a distillation head, and the excess
methylcyclohexane was immediately removed. Distillation at
atmospheric pressure was discontinued when the pot temperature
approached 105° and the collected distillate, mainly starting material
and water, was dried over molecular sieve and reused.

The product mixture was then transferred to an appropriately
sized single-necked round-bottomed flask equipped with a stirbar and
vacuum distillation head. During vacuum distillation it was essential
to cool the distillation flask to room temperature if for any reason
the vacuum had to be broken. Failure to take this precaution resulted
in product decomposition accompanied by violent deflagration.

Following a forerun of residual methylcyclohexane, a second
fraction was removed, bp. 27-28° at 2mm. The product distilled at
54-60°, 0.15mm. (literature bp. 38° at 0.03mm) Iodometric titration
showed the hydrOperoxide content to be 100% within experimental
error.

The yield of hydroperoxide was 104g (0.799mol, 8.5% of
theoretical). The product is a mixture of the desired 1-methylcyclo-
hexyl hydroperoxide (60%) and the isomeric 2, 3, and 4-methylcyclo-
hexyl hydroperoxides (40%).33 Thus the yield of the desired isomer

is 62.4g (0.479mol, 5% of theoretical). No attempt was made to

28

separate the l-methylcyclohexyl hydroperoxide from its isomers.

Preparation of 7 - chloro - 2-heptanone3 5

 

To a 2000ml three-necked round-bottomed flask equipped with
a thermometer, football stirrer, and 1000ml addition funnel was
added a solution of the mixed hydroperoxide product (0.720mol total,
0.432mol 1-methylcyclohexy1 hydroperoxide) in 21 Oml petroleum
ether (Matheson, Coleman, and Bell; bp. 30-60°). The solution was
placed in an ice bath and cooled to 0° with stirring.

In a 1000ml beaker equipped with a stirbar was prepared a
solution containing 207g (0.745mol) ferrous sulphate heptahydrate
(Matheson, Coleman, and Bell), 485ml distilled water, and 170ml
(2.04mol) concentrated hydrochloric acid. After cooling to room
temperature this solution was added dropwise to the cooled and
vigorously stirred hydrOperoxide mixture. The rate of addition was
controlled such that the temperature of the reaction medium rose to
and was maintained at l4-16°. The initially water-white organic
phase quickly turned a dark red-brown and then increasingly yellow-
green as the addition proceeded. Upon completion of the addition,
stirring was continued and the ice bath was removed. The reaction
mixture was brought to room temperature and stirred an additional
half hour.

After the aqueous and organic layers were separated by using

29

a 2000ml separatory funnel, the aqueous phase was reextracted with
75 and 50ml portions of petroleum ether. The combined ether layers
were washed with 25ml HZO’ 25ml saturated sodium chloride and
finally dried over anhydrous potassium carbonate. Following filtra-
tion the petroleum ether was removed by rotary evaporation. Vacuum
distillation of the residue yielded 51.19g (0.344mol, 80% of theoretical)

of 7-chloro-2-heptanone, bp. 78-80° at 3mm. The reported bp. is

101-102° at 16mm.

Preparation of 7- chloro- 2-heptanol

 

To a 500ml three-necked round-bottomed flask equipped with
a thermometer, stirbar, and 100ml addition funnel was added 51.19g
(0.344 mol) of 7-chloro-2-heptanone in 120 ml methanol. The mixture
was cooled to 0° by stirring in an ice bath.

A solution of 4.88g (0.129mol) of sodium borohydride (Metal
Hydrides, Inc.) in 70ml distilled water was prepared and added drop-
wise to the stirring methanolic ketone solution. Cooling was continued
to maintain the reaction temperature below 20° . On completion of the
addition the reaction mixture was stirred for 12-14 hours at 0° .

The homogenous product mixture was titrated to neutrality
with 30% sulfuric acid and the precipitated salts were removed by
filtration. The solide were washed with some methanol which was

then combined with the filtrate. The bulk of the methanol was

30

removed by using a rotary evaporator. Approximately 100ml of
distilled water was added to the residue and the mixture poured into
a 500ml separatory funnel. The evaporation flask was rinsed with
ether which was then poured into the funnel. The product was
extracted with three 75ml portions of ether. The combined ether
layers were washed with 50ml of saturated sodium chloride, dried
over potassium carbonate, and the ether removed by rotary evapora-
tion. The crude 7-chloro-2-heptanol weighed 46.55g (0.309mol, .

90% of theoretical) and boiled at 54-55°, 0.15mm.

Synthesis of 7-amino-2-heptanol - The Gabriel Synthesis32’36

 

Preparation of N-(6-hydro:gheptyl)-phthalimide -- To a 250ml

 

three-necked round-bottomed flask equipped with a stirbar, condenser,
and thermometer was added 33.7g (0.182mol) of potassium phthalimide.
A solution of 20.18g (0.134mol) of 7-chloro-2-heptanol in 100ml dry
dimethylformamide was prepared and added with stirring.

While stirring was continued, the mixture was heated to 120-
150° for three hours. The flask contents were then cooled to 100°
and 10ml of distilled water was added. The mixture was reheated
to reflux at 117° and left for 12-14 hours.

The flask contents were cooled to room temperature and
poured into a 1000ml separatory funnel. The reaction flask was

rinsed with 150ml of chloroform which was added to the separatory

31

funnel along with 500ml of distilled water. The mixture was shaken,
the chloroform layers were washed with 150ml 0.1N sodium hydroxide,
150ml distilled water, and 150 m1 saturated sodium chloride. The
solution was then dried over anhydrous potassium carbonate and the
chloroform and some dimethylformamide were removed by rotary
evaporation. The residual oil was used without purification.

Preparation of 6-hydroxyheptylammonium phthalhydrazide ~-
To a 500ml single-necked round-bottomed flask equipped with a stirbar
and condenser were added a solution of the crude N-(6-hydroxyheptyl)-
phthalimide (0.134mol assumed) in 300ml methanol, and 6. 6ml (6.8g,
0.136mol) hydrazine hydrate. After refluxing the mixture for one and
a half hours, most of the methanol and excess hydrazine hydrate were
removed by rotary evaporation. The residue was used without
purification.

Preparation of 7-amino-2-heptanol -- The solid 6-hydroxy-

 

heptylammonium phthalhydrazide (0.134mol assumed) was dissolved
in 250ml distilled water and the pH of the solution was adjusted to 10
by the addition of concentrated potassium hydroxide solution. The
mixture was poured into a 1000ml separatory funnel and extracted
twice with 50ml portions of chloroform to remove any colored re-
action side products. When the chloroform extracts were colorless
the solution was made 30% in potassium hydroxide and the product

was extracted with 100ml portions of chloroform. The aminoalcohol

32

content of the successive extracts was monitored as described under
the preparation of 5-amino-2-pentanol. The tendency for fine
needles of potassium phthalhydrazide to form during extraction was
counteracted by the addition as necessary of extra 30% potassium
hydroxide.

The combined extracts were dried over anhydrous potassium
carbonate and the chloroform was removed with efficient stirring
and careful distillation through a Vigreaux column to minimize
entrainment. The yield of 7-amino-2-heptanol was 11.79g (0.0898mol,
67% of theoretical based on 7-chloro-2-heptanol), bp. 68-69° at

0.10-0.15mm, mp. 37-39°.

Anal. Calcd. for C7H17NO: C,64.07; H, 13.06; N,10.68.

Found: C,64.01; H,l3.06; N, 10.53.

Synthesis of the N—(hydroxyalkyl)—3-pyridinecarboxamides

 

A procedure suitable for the preparatiOn of this series of
compounds from the corresponding aminoalcohols is described.

A list of the aminoalcohols and their sources follows in Table 1.

33

Table 1

Reactant aminoalcohols

 

 

Aminoalcohols Source

3 - aminop ropanol Aldrich

4- aminobutanol Chemicals Procurement Lab.
5-aminopentanol Aldrich; Pfaltz and Bauer
6-aminohexa'nol Pfaltz and Bauer
5-amino-2-pentanol this work

6-amino-2—hexanol Michael K. May, M. S. thesis,

M. S. U., 1973.

7-amino-2-heptanol this work

 

Preparation of the trimethylsiloxyalkylamine -- A 100ml bantam
three-necked round-bottomed flask fitted with a stirbar, serum cap,
and condenser terminated by a gas bubbler was flushed with dry
nitrogen. About 0.050mol of dry aminoalcohol was added and the
exact amount was determined by difference weighing. While keeping
the system under a blanket of dry nitrogen an equivalent (0.025mol,
5.32m1, 4.11g) of hexamethyldisilazane, (Aldrich) followed by a
catalytic amount of oven-dried ammonium chloride, was added.
Stirring was begun.

Nitrogen flow was discontinued and the heterogeneous reaction
mixture was slowly heated. Upon warming, the evolution of gaseous

ammonia began. Heating was continued until the oil bath temperature

34

reached 130°. At this point no reflux was Visible, ammonia evolution
had ceased, and the reaction mixture had become a homogeneous
water-white solution. The reaction flask was maintained at this
temperature for a brief period, the gas bubbler was replaced by a
drying tube, and the solution was allowed to cool. Purification of
the silyl ether amine was not necessary.

Preparation of N-(trimethylsiloxyalkyl) - 3-pyridinecarboxamide
-- Keeping the silyl ether amine under a blanket of dry nitrogen, 35ml
of pyridine (stored over molecular sieve type 4A) and 15ml of triethyl-
amine (stored over potassium hydroxide) were added to the previously .
described reaction flask. The flask was heated to 105° and an
equivalent (0.050mol, 7.20g, 5.57ml) of 3-chloroformylpyridine was
slowly added by using a disposable syringe equipped with a neoprene
plunger. On completion of the addition the serum cap was replaced
with an ungreased glass stopper and the mixture was refluxed for
one half hour.

After cooling to 0° the precipated triethylammonium hydro-
chloride was removed by filtration. The N-(trimethylsiloxyalkyl)3-
pyridinecarboxamide was not removed from the solvent.

Preparation and isolation of N-Lhydroxyalkyl)-3-pyridine-

 

carboxamide - A 100ml bantam round-bottomed flask was fitted with

 

a stirbar and distillation head. The N-(trimethylsiloxyalky1)-3-
pyridinecarboxamide solution was added and the remaining triethyla-

mine was removed. The distillation head was replaced by a reflux

35

condenser, 5ml of distilled water was added, and the solution was
refluxed for about an hour or until hydrolysis of the silyl ether was
complete. This reaction was easily monitored by taking successive
NMR spectra of neat samples of the mixture. Hydrolysis was
accompanied by a distinctive 7hz downfield shift of the protons in
the Silane region.

On complete hydrolysis the pyridine-water azeotrOpe and
remaining pyridine were completely removed by vacuum distillation
followed by several air flushes and depressurizations at about 120°.
The crude N-(hydroxyalkyl) -3-pyridinecarboxamide was consistently
obtained in 98-100% of the theoretical yield. Analysis by infra-red
Spectroscopy and thin-layer chromatography (silica gel, 8:2/CHC13:
CH3OH) showed the desired major product and minor amounts of
3-pyridinecarboxylic acid and other unidentified compounds. The
product was taken up in dry chloroform and purified by chromatog-
raphy on alumina.

A 2.5x80cm alumina column (Matheson, Coleman, and Bell,
Alcoa type F-20, 80-200 mesh) was prepared in 100:2/CHC13:
CH3OH solvent, an aliquot of chloroform solution containing about
5g crude product was applied, and the column was eluted with the
same solvent at a flow rate of 1.5-2.0ml/min. Due to the straight-

forward nature of the synthesis and lack of intermediate work-up,

isolated yields of 85-90% were typical. The products were viscous

36

yellowish oils. Some crystallized spontaneously and most could be
crystallized from various solvents. A table of melting points

follows.

Table 2

Melting points of solid N-(hydroxyalkyl)-
3-pyridinecarboxamides

 

N - (hyd roxyalky l) -

 

3-pyridineca rboxamide Solvent M. P.
N-(3-hydroxypr0pyl)>¢< Ethyl acetate Between 5-20°
N-(4-hydroxybutyl) Ethyl acetate 71-73°
N-(5-hydroxypenty1) Ethyl acetate 86-88°
N-(6-hydroxyhexyl)* Ethyl acetate 79-80°
N-(4-hydroxypentyl) Not yet obtained --
N-(S-hydroxyhexyl)** Acetone 45-51°
N-(6-hydroxyheptyl) Not yet obtained --

 

=5:
I would like to express my appreciation to Vassilios C.
Stamoudis for his preparation of these compounds.

**
Michael K. May, M. S. thesis, Michigan State University,
East Lansing, 1973.

37
. . . + 3
II. Preparation and isolat1on of NAD analogs

A five dram glass vial equipped with a polyethylene cap and
miniature stirbar served as an incubation vessel for this enzymatic
synthesis. The incubation was performed in a circulating water bath
thermostatically controlled at 37 at 025°.

To the vial were added 0.185g (0.251mmol based on a
hydrated Mw of 735) nicotinamide adenine dinucleotide (Sigma
Chemical Company, grade III) and 1.04mmol of the precursor nico-
tinamide derivative. This was followed with 0.53m1 1.0M potassium
phosphate buffer, pH 7.5, and 12ml of boiled distilled water. All
solids were dissolved with gentle stirring and the pH of the solution
was adjusted to 7.7 d: 0.1 by dr0pwise addition of concentrated potas-
sium hydroxide. After the vial was brought to temperature, 0.50g
(3.5 units; 1 unit= 1.0 mol per min at 37°, pH 7.3) of pig brain NADase
(Sigma, E. C. 3. 2. 2. 5., NAD glycohydrolase) suspended in solution
was added. The mixture was incubated with vigorous stirring for
three hours and the reaction was ended by making the solution 5% in
trichloroacetic acid. A procedure for monitoring the conversion of
NAD+ to analog will be described subsequently.

Denatured protein was removed by centrifugation at 10,000
rpm for ten minutes. The supernatant liquid was added to cold

acetone (1:5 volume ratio), let stand one minute and centrifuged at

38

5, 000 rpm for ten minutes. The acetone solution was discarded, the
centrifuge bottle was dried with a gentle stream of air, and the crude
product coating its bottom and sides in a barely perceptible film was
taken up in 10ml of boiled distilled water. This working solution was
stable for up to a year if maintained at 4°. The analog was isolated
by chromatography on polyethyleneimine cellulose. A lml aliquot of
the working solution containing about 20mg of solid was loaded on a
preconditioned 0.9x55cm PEI cellulose (Sigma, 1.17meq/g) column
and eluted with 0.0030M NH4HCO3 at a flow rate of 1.6ml/min. The
eluate was monitored at 254nm in a 0.1cm flow cell and collected in

5m1 aliquots. A sample separation is Shown in Figure 12. The

identification of the fractions will be discussed later.

III. High pressure liquid chromatographic
monitoring of analog synthesis

In order to establish reaction conditions for analog synthesis,
a method, preferably quantitative, was needed for following the
decrease in NAD+ and the corresponding increase in analog concen-
tration as the incubation of precursor with NADase progressed.
The usual procedures, analyses based on evaluation of the ultraviolet
spectrum of the reduced pyridine nucleotides or the coenzyme CN-

9

complexes, were not felt to be suitable as

1. it was not certain that the analogs would undergo reduction
at a rate sufficient to enable real- time analysis, and

39

2. it seemed likely that the Spectral properties of the
analogs and their complexes would differ little from
those of the parent nicotinamide adenine dinucleotide.

As a point of interest both of these objections proved to be valid.

Other attempts to directly follow the increase in analog con-
centration were futile. An indirect approach relying on high pressure
liquid chromatography to quantitate the uptake of N-(hydroxyalkyl)
nicotinamide was then develOped. As the nicotinamide derivatives
were present in four-fold excess, the errors involved were large,
but the method nevertheless proved useful.

The liquid chromatograph was equipped with a 37-50p, 2'x0.093”
Corasil II column (Waters Associates) eluted with 97:3/CHC13:MeOH
at 60% pump stroke. The chloroform had previously been purified
by passage through an alumina column.

Samples were prepared for analysis as follows: The enzy-
matic synthesis was carried out as described but with the abstraction
at intervals of aliquots somewhat over 0.5ml in volume. The samples
were immediately centrifuged to remove particulate matter and
0.500ml of the clear supernatant liquid was added to a tapered,
stoppered test tube. An internal standard solution containing 0.675g
3-acety1pyridine (Sigma) per 10ml solution had been prepared, and
after adding a 5-10p.l aliquot to the test tube the mixture was extracted
with 5.00ml of an 8:2/CHCL :MeOH solution. The organic phase was

3

removed, dried for five minutes over a reproducible volume of

40

anhydrous potassium carbonate, and filtered into a vial. Chromatog-
raphy of the mixture and comparison of the areas under the
nicotinamide and N-(hydroxyalkyl) nicotinamide peaks with the height
of the 3-acetylpyridine peak yielded information of progress of the

reaction. A discussion of results will follow.

IV. Determination of analog activity in enzyme systems

On enzyme-catalyzed reaction with a hydroxylic substrate,
nicotinamide adenine dinucleotide and the analogs herein described
undergo a l, 4 reduction of the pyridine moiety. This transformation
is easily observed through the appearance of an absorption at 340nm.
The presence and intensity of this absorption were taken to indicate
the degree of reaction.

The procedure was as follows: The coenzyme analogs were
isolated in aqueous ammonium bicarbonate solution following elution
from the cellulose ion exchange column. Depending on concentration,
0.50 to 1.00ml of this solution was used per sample and in each case
sufficient blank column eluate added to make the total volume 1.00ml.
To this was added 2.00ml of a stock solution of the buffer of interest,
whose molarity and pH were adjusted so as to yield the desired values
on dilution. Alcohol, if present, was added as 10pl 95% ethanol.
These mixtures we re made up and later incubated in one dram

screw-tOp glass vials with Teflon®- lined caps. An initial (t =0)

41

spectrum of each sample was taken within two minutes of the addition
by syringe of the enzyme. In all cases reaction was sufficiently slow
that the delay was insignificant. The spectrum was then rescanned
at the desired internals and the results were recorded either directly
as absorbance versus a constant reference, or indirectly as the
difference spectrum of two samples. Care was taken to establish
the same base-line for spectra taken some time apart.

The buffer stock solutions used were: (pH adjusted to the

indicated value with KOH).

Table 3

Buffer systems used in coenzyme analog
incubation with dehydrogenase

0.05M sodium pyrophosphate - pH 10

0.05M sodium pyrophosphate; 0.10% gelatine; 0.001M semicarbazide
hydrochloride - pH 10

0.05M sodium perphosphate; 0.01M semicarbazide hydrochloride
- pH 10

0.05M sodium pyrophosphate; 0.10% gelatine; 0.01M semicarbazide
hydrochloride - pH 10

0.075M glycine; 0.01M semicarbazide hydrochloride - pH 8 8t 10

0.075M glycine; 0.10% gelatine; 0.01M semicarbazide hydrochloride
- pH 10

0.075M Tris; 0.01M semicarbazide hydrochloride - pH 10

 

42

The horse liver alcohol dehydrogenase (Sigma, E. C.
1.1.1.1.) was obtained in vials containing lyophiliz ed protein equivalent
to 20 units activity (one unit converts 1p. mol ethanol per minute to
acetaldehyde at pH 8.8, 25°) and reconstituted with 0.01M potassium
phosphate buffer pH 7.50. Once reconstituted, the material was stored
in dry ice. Incubation was at ambient temperature (25°) with one unit
(0.05ml) of enzyme.

Yeast alcohol dehydrogenase (Sigma, E. C. 1.1.1.1.) was
Similarly reconstituted and the incubation with 40 units enzyme in
0.2m1 buffer per sample was otherwise carried out in the same way.

Boiled, house-supplied distilled water was used in all buffers.
An attempt to improve procedures by using water distilled from
alkaline permanganate resulted in no noticeable benefits and con-
siderable inconvenience.

Blanks containing co-eluted NAD+ in place of analog were run
with all samples in order to monitor contamination by oxidizable
impurities. Generally only small and quickly apparent background
absorptions were found. Interestingly, those analogs not reactive
in the absence of external ethanol also showed no blank, reflecting
the expected lesser activity of the analogs in the dehydrogenase

system.

43

V. Instrumentation

SLectrometers fiand Spectrophotom ete rs

 

NMR: Varian T-60
Mass: Hitachi Perkin Elmer RMU-6
Infra-red: Perkin Elmer 237B

Visible ultra-violet: Unicam SP. 800
Beckman DB-G

pH meter

Instrumentation Laboratory model 245 fitted with a Beckman

39183 probe.

Melting point appa ratus

 

Hoover capillary

Centrifuges

Waco Separator (WilkenS-Anderson Co.)

So rvall RC 2- B refrige rated

Temperature bath

 

Regulated temperature incubations were performed in a
stirred water bath set at 37 :1: 0.25°. A vermiculite insulated battery
jar was equipped with a mechanical stirrer (Talboys Instrument Co.,

model no. 104) and a Beckman differential thermometer. The bath

44

temperature was maintained within i0.01° of the set- point by a
mercury microset thermoregulator (Precision Scientific Co., model
no. 62541). The thermoregulator switched an electronic relay
(Precision Scientific Co., model no. 62690), energizing a 125 watt
blade heater (Cenco) through a Powerstat rheostat (The Superior
Electric Co.) set to equalize on and off times. A submersible mag-
netic stirrer (Henry Troemner, Inc.) was used to agitate the incu-

bation mixture.

Liquid chromatography

 

The low pressure liquid chromatography apparatus consisted
of a Buchler Polystaltic pump, Glenco column, and Buchler Fracto-
Mette 200 fraction collector, all maintained at 4°. Column eluate
was monitored with a Beckman DB-G spectrophotometer equipped
with a 0.1cm flow cell and Operating at 254 or 260nm. The output
was converted to the absorbance mode by a logarithmic amplifier
(see Appendix) and recorded on a Sargent model SR recorder. In
order to minimize dead-space and mixing effects, 0.027" i.d. Teflon®
tubing was used at all key points.

High pressure liquid chromatography utilized a Waters ALC
202* equipped with a 2'x0.093" Corasil II column (Waters). The

output of an ultraviolet absorption detector Operating at 254nm was

 

*We are indebted to Professor Harold Hart.

45

recorded on a Heath model EU-205-11 strip chart recorder driven

through a Heath EU-200-01 potentiometric amplifier.

RESUL TS AND DISCU SSION

I. Monitoring analog synthesis by high
pressure liquid chromatography

A major problem with analytical methods which monitor a
reaction through measurement of the uptake of a reactant present in
excess is that of minimizing error. For example, if a reactant
present in five-fold excess could be determined to :klO%, the error
in a percent completion calculation would be 50% and the analyst had
better seek elsewhere. The leverage present in the present situation
is similar; the reactant monitored is present in four-fold excess.
As a result it might seem more appealing to follow nicotinamide
generation rather than N-(hydroxyalkyl) nicotinamide uptake.
Unfortunately the equilibrium shown in equation (2) is very much
idealized and the NADase may cleave NAD+, freeing nicotinamide,

with no concommitant assembly of analog.

0 0
Co Co
\N

R

//O c://O
C
\Hz H NADase EH \Hz
0 + Q ——~.__. m
If N +1 N
+R R
(Z)

R = adenosine
diphosphoribose

46

47

In order to establish procedures for this analysis preliminary
work was done with aqueous dummy solutions containing N-(hydroxy-
alkyl) nicotinamide, nicotinamide, and 3-acety1pyridine. Two major
sources of error were found. The first originated in the sample
preparation procedure, specifically in volume reproducibility and
primarily in the approximately 2% limitation on the precision with
which the internal standard could be dispensed from a microliter
syringe. The second error lay in the peak integration process.

The 3-acety1pyridine, nicotinamide, and N(hydroxyalkyl)
nicotinamide eluted from the column in that order and as relatively
symmetrical peaks with little tailing. However, while the nicotina-
mide and N-(hydroxyalkyl)-nicotinamide gave broad and substantial
peaks having a high area/edge ratio and thus well-suited to cut-and-
weigh integration, the 3-acety1pyridine standard eluted as a sharp
spike. It seemed that application of the cut-and-weigh technique
here might lead to excessive error, and comparison of results using
cut-and-weigh with those obtained using the height of the 3-acetyl-
pyridine peak confirmed this possibility.

Data calculated for a series of seven injections of a standard
mixture indicated the following: For peak area calculations based
on reference peak weight the average deviation from the mean was
3.9%. For those based on reference peak height this value was 2.3%.

Expression of the standard deviation as a percentage of the mean

48

provided corresponding values of 4.6 and 3. 2%.

Clearly, the minimum error of about 3.2% was obtained if the
integrated peak areas were referenced to the height and not weight of
the reference peak. This experiment had been undertaken with no
specific precautions to assure Optimal equilibration of the chroma-
tograph. In a later study, careful conditioning of the column by
operation under experimental conditions for an hour or two prior to

sample injection resulted in a maximum deviation from the mean of

 

1.5% for calculations by this method. This marked increase in
precision was duly noted and every effort subsequently made to
precondition the column routinely. Reasonable equilibration through-
out a run was assumed if identical first and last injections agreed
within 2%. An error of 2% for this stage of the analysis was then
assumed.

By considering both error sources, the total error in a
percent completion calculation would be (4(2+2))% or 16%. This is
large, but certainly allows for sufficient precision to crudely Opti-
mize the incubation time. If necessary, greater accuracy could be
achieved through use of an electronic or ball and disc integrator.

A plot of results for two analog syntheses, Figure 11, inci-
cated that 180 min incubation resulted in a 40-60% conversion of NAD+
to analog. This degree of conversion conveniently left behind sufficient
NAD+ to serve as a contamination monitor during isolation procedures,

a use to be described later.

96 Re action

100-

90‘

80--I

20..

10"

 

---4

‘-

49

»--4
f

 

L----

d)-
p—
I-
L

S‘

1 I I 11
I T T 100

Minutes

Figure 11. Plot of percent analog synthesis
vs. time.

50

II. Identification of the products of the analog synthesis

0
/ 0 / OH
\ ‘N c
gym—12)!) CH2 - OH H( H2)n CI: — CI-I3
H
N n=2,3,4,5 N n=3,4,5
+| + l R: adenosine
R R diphosphoribose
I II

This compound was not studied.

The nicotinamide adenine dinucleotide analogs shown in
structures I and II were synthesized and then isolated by chromatog-
raphy on polyethyleneimine cellulose, as described in the experi-
mental section. A typical separation with experimental conditions
and fraction identifications is shown in Figure 12. All isolations
proceeded within the same range of yields (20-25% based on
reactant NAD+, assuming identical €260nm for analogs as for NAD+)
and, with the proviso that resolution deteriorated with decreasing

analog alkyl chain length, all analogs behaved similarly on

chromatography.

51

 

 

 

 

0.6 r-
0.5 '-
0.‘ "
m
Z
Z
D
Z .-
9 0.3
.—
Q.
8
8
n 0.2 b-
<
0.1 L-
3
O J 1 J 1 j 1 4 1

15 25 35 45 55 65 75 85 95 105115125135

ml ELUATE

Figure 12. Chromatographic isolation of
pyridine nucleotide II(n = 5).

Isolation of pyridine nucleotide II(n =5). About 20mg of the

crude acetone precipitate in lml H O was loaded on a preconditioned

2
0.9x55cm PEI cellulose column (1.17meq/g) and eluted with 0.0030M
NH4HCO3 at a flow rate of 1.6ml/min. The eluate was monitored at
254nm and collected in 5ml aliquots. The fractions were identified

as described -- (l) unresolved nicotinamide and N-(alkylhydroxy)

nicotinamide, (2) analog II(n=5), and (3) NAD+.

52

The fraction containing nicotinamide and N-(hydroxyalkyl)
nicotinamide was identified on the basis of UV spectra and by thin
layer chromatography against known samples (unactivated Silica gel
G developed with 5:4:1/chloroform:acetone:acetic acid or 73:27:1/

Ozconc. NH OH). The identity of the NAD+ fraction

1sobutyr1c ac1d:H2 4

was established by thin layer chromatography (PEI cellulose develOped
with 0.15M NH4C1) and further substantiated by its immediate reduc-
tion in buffers containing ethanol and either horse liver or yeast
alcohol dehydrogenas e.

The analogs were characterized as follows: Within experi-
mental error both the analogs and NAD+ were characterized by a
xmax at 260nm in an otherwise featureless UV spectrum. Both
formed fluorescent addition complexes with CN-, a reaction diagnostic
of the pyridinium Salts possessing a meta Sp2 carbon substituent.
Associated with this reaction in all instances was a decrease in the
extinction coefficient at 260nm and the appearance of a new absorp-
tion at about 325nm. The absence of any frequency shift in the newly
generated absorption in the case of the analogs confirms that these
are in fact nicotinamide nucleotides.

All the analogs underwent apparently complete reduction when
incubated in pyrophosphate buffer with horse liver alcohol dehydro-
genase and ethanol, but at a rate 10.3—10-4 that of the natural system.

As was the case with the natural coenzyme, this reaction was

53

indicated by the appearance of a second xmax at 340nm. Unlike
NAD+, however, none of the analogs has yet been observed to function

as a coenzyme with yeast alcohol dehydrogenase.

III. The intramolecular* reactivity of
substrate-coenzyme hybrids

The reactivity of these coenzymes in the ethanol-horse liver
alcohol dehydrogenase system, albeit kinetically unspectacular,
provided a convincing demonstration of the ability of the hybrids to
bind to the enzyme protein and be manipulated by it throughout the
conformational changes accompanying formation of the binary and
ternary complexes essential to enzymatic action. This sequence,
first postulated by Theorell and Chance12 and substantiated many
times since, is shown in equation 3.

ROH+ENZ+NAD+z ROH+ENZ::NAD+:: ROH::ENZ::NAD+::

(3)
c=o: :ENZ::NADH+H+:: C=O+ENZ::NADH2 C=O+ENZ+NADH

 

*In the context of this discussion the following definitions apply:

intramolecular reaction-reaction in which the ”ternary"
complex is a two-body system with a single hybrid coenzyme-
substrate molecule providing both reacting moieties.

intermolecular reaction-reaction in which the ternary
complex is a three-body system containing two hybrid coenzyme-
substrate molecules, one providing the reacting coenzyme moiety
and the other the substrate.

54

Discovery of this reactivity was important, but viewed in
light of the main goal of this project it meant only that a precondition
had been met without which this goal would have been unattainable.
Not only was it necessary for the coenzyme moiety of the analog to

bind the enzyme active site, but also essential was the intramolecular

 

binding and oxidation of the substrate moiety. This was a much
larger order.

Determination of intramolecular reduction- oxidation could
most logically be broken into two experimental steps, answering
two ordered questions. The first was: Knowing that the hybrid
coenzyme-substrates form binary complexes with horse liver alcohol
dehydrogenase capable of binding and oxidizing ethanol, is it also
possible to find evidence for the reactivity of one or more of these
binary complexes with the hydroxyl moiety of the hybrid coenzyme?
The second, naturally enough, was: Can any such reaction be shown
to be intramolecular? As iS often the case, the simplicity of the
answers was inversely related to that of the questions.

In principle, the first question could be answered quite
readily by the observation of reaction, or lack of it, in an enzyme-
hybrid coenzyme system free from added alcohol. Due to experi-
mental difficulties and our lack of SOphistication in dealing with
biochemical reactions it was some time before reliable results

were obtained.

55

Long incubations of 16 to over 24 hours proved to be
necessary, that resulted in occasional but unpredictable enzyme
instability. Another problem was that of contamination arising from
many potential sources. The presence of atmospheric ethanol or
acetone, incomplete removal of ethanol during processing of the
enzyme, and contamination of buffer were important considerations
and led us to doubt some early positive results because of an un-
comfortable realization of just how little such contamination would
be required to destroy an experiment Each trial was carried out
with 3 x10-7 mol of coenzyme, requiring only 1.5x 10-5g ethanol to
completely react. The real danger of course lay in the possibility
that even smaller amounts of some contaminant, or enzyme decom-
position, might raise false hopes by mimicking the small amount of
reaction we realistically might see. Clearly, we would never be
able to trust our results until we had develOped a method of detecting
such spurious "reductions”. Fortunately this was easily done.

Because the enzymatic synthesis of the hybrid analogs was
an equibilibrium reaction, some NAD+ always remained in the product
mixture even after optimization of analog yield. It was only necessary
to recover this NAD+ along with the coenzyme and use it in an incuba-
tion blank. The NAD+, isolated on the same column and almost
simultaneously with the hybrid, incubated in the same buffer, with

the same enzyme and under the same conditions, served as an ideal

56

monitor. These blank runs always indicated some contamination,

but with prOper technique and unless either the enzyme or coenzyme
had been stored improperly or too long, the amount was usually small.
For the purpose of these experiments the useful life of the enzyme was
a week or so with storage of the solution on dry ice. The isolated
NAD+, if stored under refrigeration in an air-tight container, often
gave low blanks even after two weeks.

With this means of verifying the validity of our results, we
were able to conclude that four of the hybrid coenzymes did Show
some reactivity over and above that caused by the oxidation of con-
taminants. These were I(n =4, 5) and II(n =4, 5), that is, those hybrid
coenzyme-substrated in which the hydroxyl was separated from the
coenzyme by chains of five and Six methylenes. Interestingly, those
hybrids which did not Show this reactivity showed no reaction at all,
i. e. not even the small contamination reaction of the NAD+ blank.
This reflects the expected lower activity of the dehydrogenases with
these analogs than with NAD+ and implies that the small reaction of
the blanks was not due to ethanol. The identity of the oxidizable
impurity has so far not been established.

The first question had been answered in the affirmative,
but the degree of coenzyme reduction found is worthy of further dis-
cussion. It corresponded to only 5-15% of the complete reduction

found on incubation with excess ethanol, and lesser amounts being

57

obtained with coenzymes of type II and the greater with those of type I.
We were naturally quite disappointed and at first tended to ascribe
these results to kinetic control of the reaction, i. e. deactivation of
the enzyme prior to reaction completion. Consistently, however,
we found the dehydrogenase to be still viable upon addition of ethanol,
even though further reduction did not occur in its absence. Of course
it is always possible that only one of the isozymes processes our
hybrid system and that for some reason this isozyme inactivates at
a greater rate than the others. Although some precedent for this
possibility does exist this explanation admittedly seems far-fetched.10
Much more likely is the presence of the true equilibrium point in the
5-15% reduction range. Attempts to test the fast-decaying isozyme
theory by periodic additions of fresh dehydrogenase unfortunately
failed. Contamination from inactivated enzyme made the results
impossible to interpret.

In an effort to compare our results with those to be obtained
in a more commonplace Situation, we located literature values for
the equilibrium constants on some alcohol-NAD+ systems. 39
Using equimolar ethanol-NAD+ and 2-propanol-NAD+ as models,
these values implied that at pH 10 and 25°, our incubation conditions,
24% and 89% respectively of the NAD+ would be reduced at equili-

brium. Clearly we were nowhere in sight of these values.

The discrepancy is difficult to explain, for even though the

58

model systems superficially seem very different from the synthe-
sized hybrids, this is not altogether true. Equilibrium constants
are controlled by Simple thermodynamic parameters. The component
parts of the redox pairs in both cases are essentially identical; both
systems oxidize aliphatic alcohols and reduce pyridinium carboxa-
mides. Linking the pieces together wouldn't be expected to have
much effect on the separately considered oxidation- red-uction
potentials, or through them on the equilibrium constant. Apparently
then, any such effect would have to be the result of some interaction
between the linked coenzyme and substrate moieties. Perhaps the
most reasonable explanation for the observed reduction in the
equilibrium constant would involve possible intramolecular hydrogen
bonding between the hydroxyl Of the internal substrate and the ribose
or pyrophosphate linkages of the analog coenzyme. Such bonding
would reduce the free energy of the oxidized coenzyme-alcohol
relative to the reduced coenzyme-carbonyl forms and lead to the
observed smaller amount of reaction to the analogs.

All speculation as to its cause aside, this small amount of
reaction was a potentially serious problem. Its significance was
redoubled because the final conclusions of this project would have to
be based on the relative reactivity of two diastereomers of structure
II; here we had a degree of reactivity mediocre even in an absolute

sense. Even though the experiments would be run with proper

59

controls, it would be difficult for us to present with complete con-
fidence conclusions based on such potentially small changes, and
even more so for outsiders to accept them. A means of displacing
the equilibrium and/or increasing the reaction rate was a necessity.
After trying and failing with a number of other approaches,
the expedient of including semicarbazide hydrochloride in the incuba-
tion buffer proved effective. Incubation of I(n =4) with dehydrogenase
in 0.05M glycine/KOH—0.01M semicarbazide hydrochloride, pH 10,
resulted in the establishment of an equilibrium in the range of 30-50%
reduction. An initial attempt with a buffer 0.001M in semicarbazide
hydrochloride had yielded no detectable results, thus the implicit
dependence of coenzyme reduction on semicarbazide concentration
could almost certainly be used to force the equilibrium even further.
As things were, this was a substantial and probably sufficient increase
in the extent of reaction. Even more remarkable was the effect of
semicarbazide on the rate of the reaction. After approximately four
hours incubation, an extent of reaction equivalent to a full twenty-four
hours incubation with a semicarbazide-free buffer had been attained.
Interestingly enough, these results were not repeated with
I(n =5). Any effects induced by semicarbazide on either the equili-
brium on the reaction rate were in this instance relatively small.
We will return to this observation shortly.

With the advent of success in obtaining substantial amounts

60

of reaction with our hybrid coenzyme-substrates, perhaps the most
important problem of all began to loom larger. This was really not
so much a problem as a question having only one acceptable answer.
Earlier in this discussion the second question posed was: Can any
such reaction be Shown to be intramolecular? A stereochemical
determination could not be made unless the observed reduction were
shown to occur intramolecularly. Our hope was that it would and the
problem was to find a means of determining if this were indeed the
case.

Our first thought was to use a labeling experiment. If co—
enzyme analogs of structure I were to be synthesized with the carbon
alpha to the hydroxyl nonstereospecifically monodeuterated, any
intermolecular reaction catalyzed by dehydrogenase would result in
deuterium scrambling and the consequent presence in the equilibrium
mixture of dideuterated and nondeuterated species. This would not
be the case in an intramolecular reaction, and subsequent cleavage
and mass Spectral analysis of the N-(hydroxyalkyl)nicotinamide
moiety would readily discriminate between the possibilities. Although
this experiment should eventually be performed and probably stands
alone in terms of realiability and freedom from possible misinter-
pretation, we were unable to do so due to difficulties in obtaining the
analogs in sufficiently large amounts. An alternate experiment

applying the principle of competitive inhibition was develOped in its

61

place. It should be noted that all results from this point on must be
considered preliminary.

Although the Theorell-Chance mechanism (equation (1)) for
horse liver alcohol dehydrogenase catalyzed hydrogen transfer
reactions has been accepted for many years, some of its kinetic
implications have recently been called into question.40 It now appears
that the kinetics of the reaction are not invariant, but may change with
the concentration of substrate relative to enzyme. Considering this

present uncertainty regarding the behavior of the dehydrogenase and
our lack of any quantitative kinetic data regarding these coenzyme-
substrates, it seems premature to enter into any prolonged discussion
of their reaction kinetics. Even from a simplistic vieWpoint, however,
it is clear that comparison of the reaction rates of analogs incubated
with and without addition of supplemental nicotinamide adenine
dinucleotide might constitute a useful probe into the mode of reaction.
If reaction occurred by an intermolecular mode the additional NAD+

might generate an increase in v.

f ' ' a a f
initialby unctlomng S sourceo

easily utilized coenzyme. On the other hand, an intramolecular
reaction might reveal its elf by inhibition brought about by competition
for enzyme active site between analog and NAD+.

The experiment was easily carried out. Three cuvettes,
one containing the analog in question, the second NAD+, and the

+
third amounts of both NAD and analog identical to those in the first

62

two, were prepared and incubated with dehydrogenase in glycine/KOH-
semicarbazide hydrochloride buffer, pH 10. Difference ultra-violet
SpectrOSCOpy was used to determine relative reaction rates. The
results were clear cut. Addition of NAD+ substantially in increased
the initial reaction rate of I(n = 5) and this rate increase fell with time.
With I(n =4), the NAD+ caused no change in the initial reaction rate;
after about four hours a very slight inhibition became apparent.

The effects of both semicarbazide hydrochloride and NAD+

on reactivity are tabulated below.

Table 4

+
Qualitative effects of the addition of NAD and/or
semicarbazide HCl on substrate oxidation

 

 

 

NAD+ +
coenzyme semicarbaz ide semica rbazide
Keq kinitial kinitial
I(n = 4) +++ +++ 0
I(n = 5) (a) 0 ++

 

(a) final equilibrium may not be attained before
dehydrogenase denaturation

The interpretation of these results is straight-forward. The
semicarbazide experiment indicates that the rate determining steps

of the two analog-dehydrogenase systems are quite different. This

63

difference in kinetics might also lead to speculation that the analogs

react by different mechanisms, one intermolecular and the other

intramolecular, but at this stage any such conclusion must be pure

conjecture.

+
The data from the NAD inhibition experiment may be ration-

alized on the following energy diagram.

 

 

 

I (n = 4) AG
A /\

NAD+

participating

paths

 

 

analog only A G
intermolecular
path
intramo] ecular
path
F i gu r e 1 3.

 

I (n25)

analog only

intra

 

and
inter-

 

/ t—

 

possible reaction modes.

molecular
paths

Relative energy levels of the

64

The energies shown are those of the rate limiting steps of the
systems' three possible modes of reaction: (1) intramolecular,

(2) intermolecular (analog only), and (3) intermolecular (NAD+
participation). All energy levels are relative to those of the inter-
molecular (NAD+ participation) modes, which are assigned the same
level as a matter of convenience only.

Addition of NAD+ brought about a definite increase in the
reaction rate of analog I(n = 5). Apparently, the energy level of the
newly available intermolecular (NAD+ participation) mode lay some-
what below that of ‘ppth analog-only modes, intra- and intermolecular.
On the basis of this evidence we can again reasonably place the
intermolecular (analog only) mode of coenzyme I(n =4) at an energy
above that of the corresponding intermolecular (NAD+ participation)
mode. Experimental results with this analog indicated that addition
of NAD+ had little or no effect on the initial rate of reaction. Clearly
then, some lower energy mechanism is followed both in the presence
and absence of NAD+. 7 This can only be the intramolecular mech-
anism. This conclusion is strengthened by a study of Space-filling
models of the hybrid substrate-coenzymes. The hydroxyl bearing
carbon of I(n = 4) is readily aligned in close proximity with the nico-
tinamide ring and in such a way that the carbon-oxygen bond orienta-

tion fulfills the requirements of the Prelog and Karabatsos models,

(see Figures 7 and 10).

65

Having come to this fortunate result for analog I(n =4) we are
left with the problem of analog I(n =5). At this stage we do not have
sufficient data to make a positive determination on whether the reac-
tion path followed here in the absence of NAD+ is intra or inter-

' molecular. Space filling models however Show no reason why an
intramolecular path would not be followed in this case also. The

six methylene chain allows what seems to be almost as good as
orientation with respect to the nicotinamide ring as the five methylene
chain. Analogs I(n = 2, 3) on the other hand, have alkyl chains too
Short to allow such an orientation and their non-reactivity with horse
liver alcohol dehydrogenase may be taken as additional evidence of
intramolecular reaction in those analogs which are reactive. Work
to confirm these initial results and extend them to analogs of struc-

ture II iS continuing.

10.

11.

BIBLIOGRAPHY
J. S. Fruton, ”Molecules and Life," pp. 44-46, Wiley-Inter-
science, New York, (1972).
J. S. Fruton, Ibid., pp. 318-319..

A. Todd, J. Roy. Inst. Chem., 82, 309 (1958).

 

A. L. Lehninger, "Biochemistry," pp. 333, 361, 391, Worth
Publishers, Inc., New York, (1970).

E. Negelein and H. Wulff, Biochem. Zh., 289, 436 (1937).

 

C. I. Branden, H. Eklund, B. Nordstrom, T. Boiwe, G. Soder-
lund, E. Zeppezauer, I. Ohlsson, and A. Akeson, Proc. Nat.
Acad. Sci. USA, 10, 2493 (1973).

 

 

D. J. Cannon and R. H. McKay, Biochem. BiOphys. Res. Comm.,
£2. 403 (1969).

H. Jornvall, Ibid., 35, 542 (1969).

 

H. Jornvall, Proc. Nat. Acad. Sci. USA, :9, 2295 (1973).

 

L. E. Webb, E. J. Hill, and L. J. Banaszak, Biochemistry, 12,
5101 (1973).

C. D. Bennett, Nature, (London), 248, 67 (1974).

M. Buchner, G. C. Ford, D. MoraS, K. W. Olsen, and M. G.
Rossman, Proc. Nat. Acad. Sci. USA, 10, 3052 (1973).

 

 

H. Sund and H. Theorell in ”The Enzymes,” vol. 7, p. 25,
ed. by P. D. Boyer, H. Lardy, and K. Myrback, Academic
Press, New York, (1963).

H. Theorell in ”Pyridine Nucleotide Dependent Dehydrogenases,"
pp. 121-26, ed. by H. Sund, Springer Verlag, Heidelberg, (1970).

R. Bonnichsen and A. Wassen,_Arch. Biochem. BiOphys., _l__8_,
361 (1948).

 

66

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

67

H. Theorell and B. Chance, Acta. Chem. Scand., _5_, 1105, 1127
(1951).

 

F. H. Westheimer, H. F. Fisher, E. E. Conn, and B. Vennes-
land, J. Amer. Chem. Soc., Z_3_, 2402 (1951).

 

H. F. Fisher, E. E. Conn, B. Vennesland, and F. H. West-
heimer, J. Biol. Chem., 202, 687 (1953).

 

R. Bentley, "Molecular Assymetry in Biology," vol. 2, pp. 6-9,
Academic Press, New York, (1970).

M. E. Pullman, A. San Pietro, and S. P. Colowick, J. Biol.
Chem., 206, 129 (1954).

 

J. W. Cornforth, G. Ryback, G. Popjak, C. Donninger, and
G. Schoepfer, Jr., Biochem. BiOphys. Res. Comm., 2, 371
(1962).

 

V. Prelog, Pure Appl. Chem., 2, 119 (1964).

 

J. M. H. Graves, A. Clark, and H. J. Ringold, Biochemistry,
4, 2655 (1965).

 

H. J. Ringold, J. M. H. Graves, A. Clark, and T. Bellas,
Recent Progr. Horm. Res., 23, 349 (1967).

 

E. M. Kosower, Biochim. BiOphys. Acta., 56, 474 (1962).

 

R. Bentley, "MolecularAssymetry in Biology,” vol. 2, p. 51,
Academic Press, New York, (1970).

G. J. Karabatsos, J. S. Fleming, N. Hsi, and R. H. Abeles,
J. Amer. Chem. Soc., 88, 849 (1966).

 

V. C. Stamoudis, Ph. D. thesis, Michigan State University, 1973.
H. Nunez, unpublished work, 1972.

C. Neuberg, Advan. Carbohfi. Chem., _4, 75 (1949).

 

A. S. Mildvan and H. Weiner, J. Biol. Chem., 244, 2465 (1969).

 

M. Takahashi and R. A. Harvey, Biochemistry, 12, 4743 (1973).

 

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

68

 

 

H. Zinner and H. Fiedler, Arch. Pharm., 291, 330 (1958).
N. Putokhin, Trans. Inst. Pure Chem. Reagents (Moscow),
6, 10 (1927).

G. W. Cannon, R. C. Ellis, and J. R. Leal in "Organic Syntheses,"
coll. vol. 4, p. 597, ed. by N. Rabjohn, John Wiley and Sons, Inc.,
New York, (1963).

R. C. Elderdield, W. J. Gensler, F. Brody, J. D, Head, S. C.
Dickerman, L. Wiederhold III, C. B. Kremer, H. A. Hageman,
F. J. Kreysa, J. M. Griffing, S. M. Kupchan, B. Newman, and
J. T. Maynard, J. Amer. Chem. Soc., _6_8_, 1579 (1946).

 

M. S. Gibson and R. W. Bradshaw, Ang. Chem. Int. Ed. Engl_.,
l. 919 (1968).

J. C. Sheehan and W. A. Bolhofer, J. Amer. Chem. Soc., 7_2_,
2786 (1950).

 

 

W. Pritzkow and K. H. Grobe, Chem. Ber., fl, 2156 (1960).

 

R. D. Mair and A. J. Graupner, Anal. Chem., 39, 194 (1964).

 

F. Minisci and G. Belvedere, Gazz. Chim. Ital., 22, 1299 (1960).

 

 

H. S. Mosher, J. Amer. Chem. Soc., 68, 1565 (1946).

N. O. Kaplan and M. M. Ciotti, J. Biol. Chem., 221, 823 (1956).

 

 

S. P. Colowick, N. O. Kaplan, and M. M. Ciotti, J. Biol. Chem.,
191. 447 (1951).

H. Sund and H. Theorell in "The Enzymes," vol. 7, p. 27, ed.
by P. D. Boyer, H. Lardy, and M. Myrback, Academic Press,
New York, (1963).

K. Bush, V. J. Shiner, Jr., and H. R. Mahler, Biochemistry,
_1_2_, 4802 (1973).

 

"lllllllllllllll