The coenzymes cyclic adenosine 3',5'-monophosphate
and thiamine pyrophosphate : a quantumchemical
description
Scheffers - Sap, M.M.E.
DOI:
10.6100/IR25489
Published: 01/01/1979
Document Version
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the author’s version of the article upon submission and before peer-review. There can be important differences
between the submitted version and the official published version of record. People interested in the research are advised to contact the
author for the final version of the publication, or visit the DOI to the publisher’s website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
Citation for published version (APA):
Scheffers - Sap, M. M. E. (1979). The coenzymes cyclic adenosine 3',5'-monophosphate and thiamine
pyrophosphate : a quantumchemical description Eindhoven: Technische Hogeschool Eindhoven DOI:
10.6100/IR25489
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners
and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
• You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal ?
Take down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Download date: 16. Jun. 2017
THE COENZYMES
CYCLIC ADENOSINE 3', 5' -MONOPHOSPHATE
AND THIAMINE PYROPHOSPHATE
A quantumchemical description
M. M. E. SCHEPPERS-SAP
CO-ENZYMEN CYCLISCH ADENOSINE 3 1 ,5 1 -MONOFOSFAAT EN
C:.'fiAMINE PYROFOSFAAT.
EEN QUANTUMCHEMISCHE BESCHRIJVING.
Xaast water zijn proteinen (eiwitten) essentiële bestanddelen van alle levende organismen, van de meest eenvoudige
tot de meest complexe toe. In levende organismen hebben
proteinen verschillende functies; ze treden op als:
- e~zymen, dit zijn stoffen die scheikundige reacties kunnen
versnellen of vertragen. Het zijn dus katalysatoren.
- ~ntistoffen, die als wapen dienen in het arsenaal van
verdedigingsmechanismen van organismen.
- ~ou~stenen van lange eiwitketens.
- :ransportmiddeZ, d.w.z. ze zijn verantwoordelijk voor het
vervoer van belangrijke stoffen, o.a. zuurstof, in
levende organismen.
- scheikundige boodschappers. Een voorbeeld vormen de hormonen die chemische reacties laten plaatsvinden of ophouden zodanig, dat levensprocessen mogelijk worden.
:oals hierboven al is aangegeven, worden proteinen met
katalytische activiteit enzymen genoemd. Vele enzymen
hebben om werkzaam te kunnen zijn een, van proteinen verschillend, deeltje nodig. Daar deze deeltjes mede verantKoordelijk zijn voor het totale reactieverloop noemt men
ze co-enzymen. Het werk dat in dit proefschrift is beschreven heeft betrekking op twee co-enzymen, namelijk cyclisch
adenosine 3' ,5'-monofosfaat (afgekort: c-AMP) en thiamine
pyrofosfaat (afgekort: TPP).
CycZisch adenosine 3',5'-monofosfaat
In hogere organismen, zoals de mens, treedt c-AMP op als
tweede boodschapper van veel hormonen. Hormonen (eerste
boodschappers) die een celwand niet kunnen passeren, zijn
toch in staat processen in de cel te beïnvloeden door er
voor te zorgen, dat via de wand van de cel de boodschap
doorgegeven wordt, waardoor dan het c-AMP i<~ordt gevormd in
de cel. Het c-AMP activeert of inactiveert dan op zijn
beurt enzymen in de cel en is zo in staat indirect vele
processen te regelen. Na het bewerkstelligen van de gewenste effecten in de cel dient c-AMP uitgeschakeld te
worden. Dit gebeurt 6f doordat het door de cel onveranderd
\vordt uitgescheiden 6f doordat het door enzymen omgezet
wordt in adenosine 5'-monofosfaat (afgekort: 5'-ANP).
Deze laatste reactie wordt als volgt weergegeven:
R
0--~p/
0 _::,
~~·
05'
HO/
c-AMP
R=adenine
water
OH
OH
5'-AMP
Bij de vorming van 5'-AMP uit c-AMP komt veel warmte vrij.
De hoeveelheid die vrij komt is groter dan bij reacties
van soortgelijke verbindingen.
Met behulp van computerberekeningen aan eenvoudige
molecuulmodellen van c-M4P en 5'-AMP (R = H, waterstof)
is geprobeerd een verklaring te vinden voor die grotere
hoeveelheid warmte. De berekeningen zijn gedaan aan eenvoudigere moleculen omdat de rekentijd anders veel te
groot wordt. De gegevens verkregen uit de berekeningen
tonen aan dat het warmte-effect kan worden toegeschreven
aan twee factoren:
1) het verschil in ruimtelijke vorm van de vijfring in
c-AMP en 2) het verschil in aanhechting van watermoleculen (de stoffen zijn opgelost in water) aan verschillende kanten van de deeltjes 5 '-AMP en c-AMP.
In 5'-AMP kan zich een watermolecule bevinden tussen de
zuurstofatomen (0) op positie 1' en 5'. Dit is in c-AMP
niet mogelijk, omdat de afstand tussen deze twee atomen
te groot is.
:~iamine
pyrofosfaat
TPP is een algemeen voorkomend co-enzym in levende organismen. Het is een voor de mens noodzakelijke voedingsstof ter voorkoming van beriberi (verlammingsziekte). TPP
treedt op als co-enzym bij reacties van ketocarbonzuren
(bijvoorbeeld: pyruvaat deeltje CH 3COCOO-), waarbij kool
zuurgas vrijkomt. De manier waarop TPP functioneert is
ontdekt door R. Breslow, die vaststelde dat het reactieve
centrum (hier vinden de reacties plaats met andere stoffen) zich bevindt op het koolstofatoom (positie 2) tussen
stikstof (N) en zwavel (S). Uit proeven door R. Ereslow
uitgevoerd aan de vijfring van TPP en erop lijkende vijfringen is gebleken, dat de vorming van een reactief centrum op positie 2 bij een vijfring waarin S vervangen is
door een stikstofatoom (N, imidazolium systeem) moeilijker
verloopt dan bij de thiazolium ring (dit is de naam van de
vijfring in TPP). Deze bevindingen zijn echter tegen-
strijdig met spectrametrische gegevens van deze twee vijfringen, waaruit men namelijk kan afleiden dat de vorming
van het reactieve centrum in beide gevallen even snel zou
moeten gaan.
Aan eerder genoemde vijfringen en de te vormen vijfringen
met een reactief centrum zijn computerberekeningen gedaan
om te achterhalen welke factoren verantwoordelijk z n
voor de, relatief gezien, grote snelheid waarmee bij het
thiazolium systeem het reactieve centrum gevormd wordt.
De gegevens tonen aan, dat het thiazolium systeem de
electrenen (dit zijn de deeltjes die zorgen voor de bin-
dingen tussen de atomen) die betrokken zijn bij de vorming
van het reactieve centrum op een zodanige wijze heeft opgeborgen, dat ze gemakkelijker te gebruiken zijn.
M.M.E. Scheffers-Sap
Berlicum, 14 december 1979
THE COENZYMES
CYCLIC ADENOSINE 3', 5' -MONOPHOSPHATE
AND THIAMINE PYROPHOSPHATE
A quantumchemical description
PROEFSCHRIFT
ter verkrijging van de graad van doctor in de
technische wetenschappen aan de Technische
Hogeschool Eindhoven, op gezag van de
rector magnificus, prof. ir. J. Erkelens, voor
een commissie aangewezen door het college
van dekanen in het openbaar te verdedigen op
vrijdag 14 december 1979 te 16.00 uur
door
MARIA MARGARETHA ELISABETH SCHEFFERS-SAP
geboren te Tilburg
DRUK: WIBRD HELMONO
DIT PROEFSCHRIFT IS GOEDGEKEURD DOOR
DE PROMOTOREN
PROF. DR. H.M. BUCK
EN
PROF. DR. U.K. PANDIT
"Would you tell me, please, which way I
ought to go from here?"
"That depends a good deal on where you
want to get to," said the Cat.
"I don't much care where-" said Alice.
"Then it doesn't matter which way you
go," said the Cat.
"-so long as I get somewhere," Alice
added as an explanation.
"Oh, you're sure to do that," said the
Cat, "if you only walk long enough."
uit Alice's Adventures in Wonderland
Contents
Chapter I
Chapl:er 11
General introduetion
1.1 Coenzymes
I.Z Cyelio adenosine 3',5'monophosphate
I.3 Thiamine pyrophosphate
I.4 The va~idi
of quantumohemioal ealoulations
Heferenoes
Summary of the all-valenee methods
used, the Extended-Hückel, CND0/2
and ab-initio method
11.1 Introduetion
II.Z The Extended-Hüekel method
1!.2.1 Theory
II.2.2 Parameters
the
ordinary and iterative
EH calculations
II.3 The Complete NegZeet
Differential Overlap method
11.3.1 Theory
the CNDO
method
II.3.2 Parametrization for
the CND0/2 methad
II.3.3 The GEOMO program
11.4 Ab-initia calculations
11.5 Mulliken population ana
is
II.6 Calculation of the solvation
enthalpy
Heferences
11
12
16
17
20
21
23
25
26
29
30
31
32
33
36
Chapter 111
Chapter IV ,
The solvent effect on the
enthalpy of hydralysis of c-AMP
111.1 Introduetion
111.2 Geometriee of phoephate
dieetere
111.3 The effeat of the solvent
and the ribose ring puakering on the net enthalpies
of hydralysis
111.3.1 Calaulation of net
enthalpies of hydralysis and net solvation enthalpies
111.3.2 Ribose ring puakering
111.4 Diecuesion
Referenaei and notes
The influence of solvation and
ribose ring puckering on the
enthalpy of hydralysis of c-AMP
IV.1 Hydragen bonding
1V.2 The water dimer
1V.3 Hydration sahemes of models
of a-AMP • 5 1 -AMP and 3 '-AMP
1V.4 The aontribution of solvation and ribose ring
puakering to the net enthalpy of hydralysis of
a-AMP
1V.S Hydragen bonding ae a model
for the dynamias of enzymeaoenzyme aomplexes
Referenaes
38
41
46
51
51
55
58
60
62
68
69
72
Chapter V
The acidity of thiamine pyrophosphate and related systems
V.l IntPoduction
V.1.1 Ristoriaal back74
ground
V.1.2 Relation
structure
of TPP to the oatalytio aativity
V.2
77
CND0/2 calculations on 1,3azolium systems
V. 2. 1 d-Orbital aonjugation
80
V.2.2 Bonding and electron
V.3
V.4
densities
85
Solvation enthalpies
91
The use of an MO desaription
for the transition state and
an estimation of the aati
vation enthalpy
V.4.1 The aharaater of the
transition state
91
V.4.2 Estimation of the
aativation enthalpy
92
V.4.3 H-D exchange reaations
of arenes
Referenaes and notes
Chapter VI
Thiamine pyrophosphate-catalyzed
decarboxylation of pyruvate anion
VI.l Introduetion
VI.2 The reaation saheme for
95
97
100
the pyruvate deaarboxy-
101
lation reaation
VI.3
The net reaation enthalpies of pyruvate deoarboxylations
VI.4
~ith
1,3-
azolium systems
103
TPP as aocarboxylase
108
References
110
Appendix
A
112
Appendix
B
116
Summary
122
Samenvatting
124
Curriculum vitae
126
DankW"oord
127
The work described in this thesis has been financially
supported by Unilever Research, Vlaardingen, The Netherlands
CHAPTER I
General
I. 1
introduetion
Coenzymes
Protein molecules serve several functions in living
systems. Perhaps their most striking biochemical role is
their ability to affect in a specific and efficient manner,
the rates of the wide spectrum of reactions that constitute
the dynamic aspect of the process of life. Proteins possessing such catalytic activity are called enzymes 1 (for a
classification, see Table I). They are distinguished from
ordinary proteins by having active sites, which are responsible for the action of the enzymes. Some enzymes depend
for activity only on their structures as proteins, while
others also require one or more nonprotein compounds, called
co;actors. Cofactors fall into two groups, the metal cofactors and the organic cofactors. The latter group, which
Table 1.1
Classes of enzymes and types of reaction
catalyzed
enzyme
type of reaction catalyzed
oxireductases
transferases
hydrolases
lyases
oxidation-reduction
group transfer reactions
hydralysis
the addition of groups to double honds or
vioe versa
isomerizations
condensation of two molecules coupled with
cleavage of the pyrophosphate bond of ATP
is omeras es
ligases
11
are called ooenzymes 1 encompass a wide range of compounds
which are related to vitamins. The catalytically active
enzyme-cofactor complex is called the holoenzyme. When the
cofactor is removed, the remaining protein, which is
catalytically inactive by itself, is called apoenzyme. In
case of a very tightly bound enzyme-coenzyme complex, the
coenzyme is usually referred to as a prosthetio group.
lfuereas coenzymes regulate chemica! reactivity, enzymes
related to these coenzymes control the stereospecificity,
as is very impressively demonstrated for NADH 1 • All enzymes
exhibit various features that could conceivably be elements
in the reguiatien of their activity in living cells. The
rates of enzymatic reactions depend on: -the pH in the
cell, -the substrate concentrations, -the cofactors. Some
enzymes possess, in addition, properties that specifically
endow them with regulatory roles in metabolism. Such more
highly specialized forms are called regulatory enzymes.
One class cernprises the allosterio enzymes, whose catalytic
activity is modulated through the noneavalent binding of a
specific compound (cofactor and termed an allosterio
effector) at a site on the protein other than the catalytic
site. The mechanism of action of an allasterie effector can
be a direct or an indirect one.
In this thesis attention has been given to two important coenzymes, namely cyclic adenosine 3' ,5'-monophosphate and thiamine pyrophosphate.
1.2
Cyolio adenosine 3',5'-monophosphate
Cyclic adenosine 3',5'-monophosphate 2 (c-~1P, Figure
1.1) is a universally occurring nucleotide of immense biologica! importance. It has been first isolated in 1959 by
Butherland and coworkers as part of their investigation on
the mechanism of action of certain hormones, such as
adrenaline, in regulating carbohydrate metabolism. On basis
of their study they proposed that the immediate action of
adrenaline and many other horrnanes lies in the activatien
of the enzyme which is responsible for the production of
12
1.1
c-AMP
c-A~•P. In turn, c-AHP controls the activity of other enzymes,
frequently by an allasterie activation. Since c-AMP transmits and amplifies, within cells, the chemical signals
delivered via the blood by horrnanes (first messengers), it
is called a second messenger (Figure 1.2).
In the breakdown of glycogen to blood glucose in the
l i ver ce 11, c-AMP acts as an all os teric effector (Figure
1.3). The enzyme protein kinase is inactive until c-AMP is
present. The activated kinase perfarms the same function
for a related enzyme, phosphorylase kinase. This enzyme
activates in turn the phosphorylase. The result of this
final activatien is the breakdown of glycogen. Whenever
glycogen is degraded, it would be a waste of energy to
continue the synthesis of additional glycogen. A specific
enzyme, however, mediates the synthesis of glycogen, i.e.
glycogen synthetase. At the time when some c-AMP molecules
are initiating the reaction for the conversion of glycogen
into glucose, others are generating the inactive form of
glycogen synthetase.
The concentratien level of c-AMP in the cell is regulated by the action of two enzyrnes. c-AMP is formed from
ATP by the action of adenyl cyclase, a rnernbrane-bound
enzyrne, and it is converted into adenosine 5'-monophosphate
(5'-AMP) by a specific phosphodiesterase (Figure 1 .4). The
hydralysis of c-AMP into 5'-AMP is a highly exothermic
13
hormones
( first messenger l
adenyl cyclase
I receptor}
/
"""
ATP
c-AMP
I secend
messenger l
/!~
biochemica!
responses
(enzyme
gene
activation,
expression l
/l~
physiologicol
respon,ses
(glycogenolysis. membrone
permeability l
Figure 1.2
The seaond messenger aonaept
c
odenyl
~;1
5'-AMP
AMP
phosphodiesterase
'Y'~
92
+
IQSe
-PP
• 2P;
1
~
pyrophosphatase
adenylate
ATP
Figure 1.4
14
Synthesis and aonversion of a-AMP
kinase
receptor
cell
ATP
membrane
c-AMP +PPi
protein
L
kinase _ ____,_ protein
kinase
+
c-AMP-@
ATP + phosphorylase kinase - - - 2
( inactive)
Ca +
phosphorylase
kinase (active)
©
@:XID (inactive)
ATP
+
phosphohydrolase
(active)
b
+
phosphohydrolase
ADP
a
+
ADP
( active)
( inactive)
glycogen + Pi
glucose
~
glucose
eelt membrane
1 - ph os phate
6- phosphate -glucose
I
blood
Figure 1.3
Conversion of glycogen into glucose
+
Pi
glucose
reaction 3 • The large negative Gibbs free energy 3 (-37.2 kJ/
mole) and enthalpy 3 (-46.4 kJ/mole) provides a thermadynamie
harrier against the reversal through a phosphodiesterase.
I.3
Thiamine pyrophoaphate
When thiamine (vitamine B1 ) was isolated in 1911, the
chief concern was its role in nutrition. Since then its
structure has been elucidated, and its pyrophosphate ester
was identified as cofactor for the enzyme pyruvate decarboxylase4. Thiamine pyrophosphate (TPP, Figure 1.5),
CH3
~
Figure 1.5
\_...5
OH
I
OH
I
11
0
11
0
CH -CH 2 -0 -P-O-P- 0
2
_
TPP
cocarboxylase, serves as a coenzyme for two classes of
enzyrne-catalyzed reactions of the carbohydrate metabolism
in which aldehyde groups are removed and/or transferred:
(1) the decarboxylation of a-keto acids and (2) the formation or degradation of a ketales (Figure 1.6). In these
reactions the thiazole ring of TPP is a transient carrier
of a covalently bound "active" aldehyde group 5 •
The present view of the mechanisrn by which TPP
functions as coenzyrne has arisen frorn the discovery that
thiamine alone promotes nonenzyrnatic decarboxylation of
pyruvate to yield acetaldehyde and carbondioxide. Studies
of this model reaction disclosed that the hydragen at
position 2 of the thiazole ring ionizes readily to yield
a carbanion, which reacts with the carbonyl carbon atom
of pyruvate at elevated temperatures to yield carbondioxide and the hydroxyethyl derivative of the thiazole ring.
16
0
11
/
c
R
R,
o-
I
H-C-OH
0
I
c=o
~
co2
0
11
+
1.6
/
[ ~ +TPP
0
11
R-C
:/ !-''
R-C-H
I
R
+
R,-C-H
H+
~0
11
C-H
0
0
OH
11
11
I
R- C- OH
+
R-C-C-R
I
H
2
Basic pathway for TPP-dependent reactions
The hydroxyethyl group may then undergo hydrolysis, to
yield acetaldehyde, or react with an aldehyde to yield an
acyloin.
Thiamine must be supplied in the diet as precursor
for TPP. TPP is formed by a transfer, catalyzed by a
thiamine pyrophosphokinase, of the pyrophosphate group of
ATP. When the supply of thiamine is restricted, then one
or more enzymes requiring TPP will also be deficient.
Thiamine is widespread among foods, but there is little
synthesis by intestinal microorganisms, and symptons
readily appear after dietary deprivation.
I.4
The validity of quantumchemiaal calculations
Although molecular orbital (MO) calculations have been
performed on a number of problems relevant to biochemical
structure and the functions of biomolecules, it is well
worth to consider the objections which can be raised to
17
such studies.
Until quite recently, the principal problem in combining experimental and theoretica! approaches to various
subjects has been the gap between experimental data (solution of graatest interest) and theoretica! "free state"
re sul ts. A fel~ attempts have been publisbed in l i terature
which explicitly incorporated solvent effects into the
calculations. Yet it will be at least saveral more years
before such efforts can provide data of accuracy equal to
the experimental ones. Thus it is necessary to identify
those theoretica! results which are subject to solvent
effects.
In this study theoretica! results of two types are
presented: energies of reactions and electron densities.
In a reaction, there are two basic quantities of interest:
the relative energies of the compounds and the activatien
energy of the reaction. While both of these quantities
are subject to solvent effects without any doubt, the
relativa total energies of the reaetauts and products
will be difficult to predict in general.
While thermodynamic properties (equilibrium constauts
for, e.g. ionization, tautomerization and molecular association) may be very strongly solvent-dependent (as demonstrated by comparison of gas phase and solution basicities 6 )
it appears unlikely that the intrinsic sleetronie structures
of ions and molecules are dependent. It is well accepted
that many functional groups undergo subtie electronic
changes upon solvation, such as the spectroscopie changes
accompanying hydrogen bonding. Yet there is no evidence
from theory that attaching a hydrogen bonded water molecule,
the electronic structure of a large molecule seriously
changes. Furthermore, theoretica! "free state" calculations
are capable of reproducing solution speetral characteristics. For example, they accurately reflect changes as
gross as covalent attachment of a proton to a nucleic
base in water 7 , which is certainly a more significant
change than adding a neutral non-covalently bound solvent
molecule. It is then reasonable to assume that while the
18
solvent may cause subtle changes in the electrooie structure,
it is highly unlikely and unprecedented that the polarity
of a bond would be reversed on account of a change in solvent alone. One should also note in this context the many
successful correlations of experimental magnetic resonance
parameters with simple charge or spin densities calculated
for the "free state" of a system 8 •
In Chapter II of this thesis a description is given
of the quantumchemical methods which have been used.
Especially the semiempirical CND0/2 metbod has been applied.
Some results are supported by calculations using the ab~~tio method with the ST0-3G basis set. Furthermore, the
metbod by which the solvation enthalpy has been calculated
is discussed.
Chapter III and IV present a study on the large exothermic enthalpy of hydro
is of c-AMP. It was found that
a contribution to this large exothermic enthalpy is deli\·ered by a regio-specific hydratien in 5'-AHP and 3'-AMP
a~~ via loss of strain in the ribose ring.
In Chapter V the H-D exchange reactions of 1,3-azolium
cations have been stuclied in order to explain the rateenhancement for the 1,3-thiazolium cations. It is clearly
shown that the smaller amount of energy necessary for the
1,3-thiazolium cation to employ the appropriate a MO is
responsible for the relatively small difference in exchange
rate between the 1,3-oxazolium and 1,3-thiazolium cation 9 •
In Chapter VI the reaction path for the decarboxylation
of the pyruvate anion with 1,3-azolium cationsis described 9 •
19
Heferences
1. General information about enzymes and coenzymes: (a)
H.R. Mahler and E.H. Cordes, "Biologica! Chemistry",
Harper and Row, New York; 1971; (b) E. Buddecke,
"Grundriss der Biochemie", W. de Gruyter, Berlin; 1974.
2. J.P. Jost and H.V. Rickenberg, Ann. Rev. of Biochemistry,
40, 741 (1971).
3. J.A. Gerlt, F.H. Westheimer and J.M. Sturtevant, J.
Biol. Chem.,
5059 (1975).
4. Reference la, pp 401-406.
5. J.J. Mieyal, R.G. Votaw, L.O. Krampitz and H.Z. Sable,
Biochim. Biophys. Acta, lil• 205 (1967).
6. E.H. Ernett, Acc. Chem. Res., ~. 404 (1973).
7. W. Hug and I. Tinoco Jr., J. Am. Chem. Soc., 95, 2803
(1973); 96, 665 (1974).
8. J.B. Stothers, "Carbon-13 NMR Spectroscopy", Academie
Press, New York; 1972; Chapter 4.
9. M.M.E. Scheffers-Sap and H.M. Buck, J. Am. Chem. Soc.,
lQ!, 4807 (1979).
20
CHAPTER 11
Su nunary
methods
o:f the
used,
CNDO /2
all- valenee-elect rons
the Extended-Hückel,
and
ab-initio
metbod
Introduetion
II.1
The concept of molecular orbitals constructed from atomie
orbitals is suggested as early as 1929 by Lennard-Jones 1 and
subsequently referred to by Mulliken 2 as the ''linear combinattien
atomie orbitals" (L.C.A.O.-MO) approach.
The Hartree-Fock methad is a procedure for finding the
best many electron wave function o/ (2.1) as an anti-symmetrized
product of one electron orbitals ~-.
In the case of molecules,
l
the functions
(2.2) are molecular orbitals formed usually
from a L.C.A.O.-MO approximation.
\V
-
i
/-:')"
p
Z: ( -1 ) P [ 1); ( 1) a ( 1 )1); ( 2) 8 ( 2) .... ;f! Zn ( 2n) 8 ( 2n)]
1
2
p
f
\ .n.
l:c
]l
.Q
]ll
]l
( 2 . 1)
( 2. 2)
The set of initial atomie functions $ ]l is called the basis set.
Although the complete salution of the Hartree-Fock problom requires an infinite basis set, good approximations can be
achieved with a limited number of atomie orbitals. The coefficients c ;n. , which measure the contribution of each atomie orbital in the molecular orbitals, are parameters determined by
a variational procedure, i.e. chosen so as to minimize the
expression
E
( 2. 3)
21
where E represents the expectation value of the electronic
energy associated with the Hamiltonian H of the given molecule.
If only kinetic energy and Coulomb terms are taken into account and furthermore the Born-Oppenheimer approximation is
assumed to be valid, the Hamiltonian operator is given by
2n
L
H
~
Hcore(~) +
2n
L
~<v
1/r
~v
( 2. 4)
where Hcore(~) is a ene-electron operator, representing a sum
of kinetic and potential energy of all electrans in the core
and 1/r11\) is a two-electron operator, which represents the
mutual repulsion of the electrons in the atomie orbitals
~ and v. The variation theorem requires for each molecular
orbital i, that the coefficients c . satisfy the following
~1
sets of simultaneous equations:
Ec . (F
~
~1
~\)
-E.S
1
~\)
) = 0
v = 1 , ••••• , n
(2.5)
in which n is the number of basis set functions used and
(2.6)
with S~V the overlap integral, <~ ~ I~ V >.
A non-trivial solution of the secular equation exists if
IF
~V
-E.S
I =0
1 ~\)
(2.7)
with the values Ei being the eigenvalues.
Roothaan 3 has shown that for a closed shell system
is given by
F
~\)
H~v
+EL
pcr
Pp 0 [<J.lvJJpcr>- !<~pJJvcr>]
F~v
(2.8)
where
(2.9)
and
(2.10)
22
and P
is the total electrooie popuiatien in the overlap
90
region between atomie orbitals p and o:
ace
(2.11)
2 l: c pl.c (jl.
i
The salution of the secular equation (2.7) requires the
evaluation of the matrix terms F~v· The F~v's are functions
of the coefficients c . and are evaluated by solving the
~1
secular equation. The Hartree-Feek procedure thus requires to
make a preliminary guess of the values of the molecular population distribution terms Ppa; these values are then used to
calculate the matrix elements F)l\! and the next step is to
solve the secular determinant. This, in turn, provides a better
approximation to the wave function and an "improved" set of
values P . The process is repeated until no difference is
po
found between successive improved wave functions. Finally, it
may be shown that when such a calculation has been iterated
to selfconsistency, the total electronic energy E of a closed
shell molecule is given by
(2.12)
The main obstacles to the salution of this problem lie in the
farmidabie number of multicentered integrals <)lv/ /pa> which
arise even with the use of a minimal basis set, and the difficulty involved in their evaluation.
The CND0/2 approximation belongs to the SCF molecular
orbital methods, whereas the Extended-Hückel methad is referred
to as an approximate SCF-field theoryq.
In the Extended-Hückel as weJl as the CND0/2 calculations
the Slater 5 AO's of all valenee electrens are used as basis
set.
II.2
The Extended-Hüokel method
II.2.1 Theory
The Extended-Hückel (EH) theory, developed by Hoffmann 6
23
,
calculates a- and n-electron distributions simultaneously.
In this method, the basis set for the linear combination of
atomie orbitals is extended, with respect to the simple
Hückel method, including all valenee shell atomie orbitals.
The basis set used in the calculations consists of 1s orbital
of hydrogen, 2s and 2p orbitals of carbon, oxygen and
nitrogen, and 3s, 3p and 3d orbitals of sulphur and phosphorus. In the Hoffmann formulations H~~·s are chosen as
the negative values of the valenee shell ionization potential (VSIP) and the Wolfsberg-Helmholtz approximation is
used for estimating off-diagonal elements
(2.13)
The value of K, which is used as a sealing factor, is
chosen as 2.00 in accordance with earlier work 6 • 7 • The overlap matrix is internally computed and the Hamiltonian
matrix is constructed from it by equation (2.13). The
complete set of (2.5) is solved with two matrix diagonalizations. The resultant wave functions are subjected to a
Mulliken population analysis (see II.4.5), yielding overlap
populations and gross atomie populations. The total energies
are calculated, according to Hoffmann 6 , as
i
(2.14)
where €i and ni are the orbital energy and accupation number
of the ith molecular orbital, respectively.
Some objections against the use of semi-empirica!
methods like the EH metbod have been discussed 8 • The correct
symmetry and general shape of a molecule or ion might be
correctly calculated, but good precise bond angles, lengtbs
and force constants are not to be expected. In ordinary
EH calculations excessive charges accumulate on more electronegative centers. This shortcoming is corrected by a metbod
that assumes linear dependenee between the matrix elements
and the calculated net charges. The most common variant
of the EH metbod employs an iterative technique in which
24
the diagorral matrix elements are considered as a function of
the net atomie charges. The calculation is iterated to
charge consistency. Iterational treatment has led to improvement of the results for ionic species, but has given
no significant difference for neutral systems 9 • The iterative
EH method, as proposed by Rein et a~. 10 , calculates the
total energies according to equation (2.15).
Etot = ~.
lJ:: n.1 (s.+h.)
1 1 + Ecore-core
(2.15)
1
~
Khere h. = <~ lhl~ > and the second term is the care rel
\1
\1
pulsive energy, calculated by:
(2.16)
Ecore-core
- Aeff an d rAB b e1ng
·
·
e ff ect1ve
core c h arges o f atom A an d
distance between atoms A and B, respectively. In equation
(2.15), h represents the same one-electron operator of
kinetic energy and core attraction as the one in the HartreeFock method. The matrix elements H
have been calculated
\111
according to equation (2.17), as derived by Basch et al. 11 :
H
\1\1
Xq 2
+
Yq
+
Z
(2.17)
in which X, Y and Z are input parameters (Section II.2.2).
At each iteration the H
values are obtained from those of
11\1
the previous cycle by equation (2.18):
H
llll(n+1)
H
2
( )(1-;\) + À(Xq +Yq+Z)
J..lll n
(2.18)
with À, the damping parameter, taken as 0.1. Iterations are
continueduntil the atom charges remain constant to within
0.01 electrooie charge.
II.2.2
Parameters for the ordinary and iterative EH
catauZations
The orbital parameters (Table !1.1), except for sulphur,
25
entering the EH theory, namely, the valenee state ionization
potentials (VSIP) and orbital exponents, are the same as
those used by Boyd 12 in a molecular orbital study of ATP.
For sulphur the data have been taken from a study by
BartelZet al. 13 • The orbital exponents are just the Slater
values, exeept for the H ls and P 3d orbitals, for whieh
the values are taken from SCF optimization ealculations 14 •
The carbon and hydragen VSIP's are those in common usage 6 ,
and the phosphorus and oxygen values are taken from SCF
eigenvalues of P0 13 • For the iterative EH ealeulations the
constants for equation (2.17) are obtained from the atomie
speetral data as determined by Basehet al. 11 •
Table 11.1
element
H
c
c
N
N
0
0
p
p
p
s
s
s
11.3
11.3.1
Parameters in ordinary and iterative EH calculations
orb i tal
1s
2s
2p
2s
2p
2s
2p
3s
3p
3d
3s
3p
3d
VSIP
(eV)
13.60
21.40
11.40
26.00
13.40
37.59
14.62
18.57
13.98
8.48
23.06
10.36
7.00
orbital
exponent
1. 200
1. 625
1. 625
1. 950
1. 950
2.275
2.275
1. 600
1. 600
1. 400
1. 817
1. 817
1. 200
x
y
z
(eV)
(eV)
(eV)
13.62
3.47
3.47
3.49
3.44
3.47
3.46
1.77
1. 51
1. 77
1. 51
1. 7 5
1. 58
27. 18
17.56
14.65
20. 11
12.70
22.89
18.57
13. 2 3
15.25
1. 18
15.25
10.41
2.00
13.60
19.40
10.60
25.56
8.28
32.30
15.80
18.77
20.51
1. 15
20.51
12. 31
0.83
The fompZete EegZeet of QifferentiaZ QverZap methad
Theory of the CNDO methad
1f the full SCF equations are solved without any
approximations, then the calculated energies and electron
26
distributions are dependent on the choice of the coordinate
axis. The results must also be the same whether we choose
to take a linear combination of atomie orbitals, ar a
linear combination of hybridized orbitals. The results of
an SCF calculation are invariant to an orthogonal trans
formation of the atomie orbital basis. If one introduces
approximations to the SCF equations then the conditions of
rotational and hybridizational invariance must be conserved.
The approximations for the CNDO methad are 15 :
1. Only valenee electrans are treated explicitly, the inner
shells being treated as part of a rigid core.
2. cu 's are treated as if they farm an orthorrormal set;
thus
s ]JV
= éi
]JV
(Kronecker delta)
(2.19)
3. All two electron integrals which depend on the overlap
of charge densities of different orbitals are neglected.
This means that
éi
]JV
éi
y
pcr J.lP
(2.20)
4. The electron interaction integrals are assumed to depend
on
on the atoms to which the orbitals ~ ]J and ~ ') belang.
Thus yJ.lP is set equal to yAB' measuring an average repulsion between an electron in a valenee atomie orbital on
A and another in a valenee orbital on B.
~. The core matrix element H
contains the interaction
]J]J
energy of an electron in valenee orbital ~ on A with
]J
the care of A and with the cores of all other atoms B
H
(2.21)
]J]J
u)J]J
(2.22)
6. Core matrix elements H)JV , where ~ ]J and ~ V are different
but both belang to A, may in analogy to 4 be written:
H
)1\1
(2.23)
27
However, due to the mutual orthogonality of s, Px• Py
and p Z , U]JV is zero and the remaining terms are small,
so H].JV = 0 for lJ f v.
7. Core matrix elements H , where ~ is on atom A and ~P
lJP
lJ
is on atom B, will be considered proportional to the
overlap integral S]Jf) :
(2.24)
H
]Jf)
Under these approximations, the matrix elements of the
Fock Hamiltonian reduce to
FlJlJ
UlJlJ
+
(PAA-!PlJlJ)yAA
+
B(~A) (PBBYAB-VAB)
(2.25)
13~BS]JV - !PlJVYAB
(~lJ
lJfV
(2.26)
on A, $v on B)
The expression (2.26) applies even if 1J and v are on the
same atom. Then S]J\!
0 and yAB is replaced by yAA.
The total energy is given by the sum of monoatomie and
diatomic terros
E SA
A
Etot
with
A
EA
E p
jJ
AB
and
EAB
= EE
jJ\)
-1
]J]J u]J]J
+
E
A<B
+
e:
(2.27)
AB
AB
2)
!EE (P ]J]J p \)\) -lp
2 ]J\!
]J\!
(2P J.JV 13 lJV - 21 PlJV 2+y AB )
(2. 28)
+
(ZAZBrAB-PAAVAB-PBBVBA+PAAPBBYAB)
(2.29)
For large intermolecular separations, the potential integrals VAB' VBA and yAB all approximate rAB-l and with QA
ZA-PAA (QA: net atomie charge on A), the last group of
termsin (2.29) becomes QAQBrAB- 1 . This shows that the
theory takes proper account of the electrastatic interaction
between charged atoms in a molecule.
28
II.3.2
Parametrization for the CND0/2 method 16
From the previous Section one obtains a penetratien
integral term ZByAB-VAB in F~~· if one substitutes QB =
ZB-PBB in equation (2.25). This term gives rise to calculated bonding energies even when the bond orders connecting
two atoms are zero.
In the CND0/2 method this deficiency is corrected in
the simplest possible way by neglecting the penetratien
integrals. Thus
(2.30)
The core matrix elements can be estimated from atomie data
in tKo ways:
and
-I )J
(2.31)
-A
(2.32)
~
1\ith I the ionization potential, A the electron affinity
and ZA the effective nuclear charge. U
is in the CND0/2
~~
method the average of both estimations:
- 1
2
(1 ~ +A)(Z AZ
- 1 )y AA
~
(2.33)
The values used for the electronegativities -!(I +A) are
11
11
listed in Table 11.2. Initial estimates of the LCAO coefficients may be obtained by a HUckel-type theory using matrix
elements
(o)
-l(I +A)
F
( 2. 34)
~~
2.
11
~
(2.35)
and the final solution is obtained as described in Sectien
II.1. The bonding parameters S~B are approximated by
(2.36)
29
Table II.2
Values of parameters 15 •
17
!(Is+As)
HIP +Ap)
~(Id+Ad)
-BoA
co re
element
(eV)
(eV)
(eV)
(eV)
charge
H
7. 176
14.051
19.316
25.390
14.033
17.650
-
-
9
21
25
31
15
18
1
4
5
6
5
6
c
N
0
p
s
5. 572
7.275
9. 111
5.464
6.989
-
0.500
0.713
B~ and B~ are adjustable empirically determined parameters
and chosen to give the best agreement between CND0/2 and
ab-initio calculations.
The sp and spd type calculations differ only by the
omission of 3d functions from the basis set.
11.3.3
The GEOMO program 19
The GEOMO program perfarms LCAO calculations with any
usual semi-empirical formalism (CNDO, INDO, MINDO). The
algorithms in this program permit use of parametrization
and allow direct minimization of energy with respect to
any geometrie parameter. For our purposes the CND0/2 methad
is used.
A stable geometrie configuration of a molecule
corresponds, in the Born-Oppenheimer approximation, to the
minimum of the molecular energy, when the interatomie
distanees vary. In the CND0/2 method the total energy (2.27)
ean be decomposed into monoatomie terms, whieh do not depend
on geometry, and diatomic ones. The other terms determining
electron repulsing are negleeted. Therefore, the energy
variations arise from diatomie terros (2.29) only, wherein
the term rAB -1 is replaced by the nuclear repulsion term
f(rAB). When the variatien of an internal eoordinate qz
modifies the distance rAB' the variatien of EAB can be
calculated from:
30
5EAB
;;:
oB ..
AB
l:L rzp
•
ij
~
)lV
6qz
-1p2
2 )1\.J
6VAB
+
öf (r AB)
ZAZB
oyAB
PAAPBB-6qz
- PAA-6-qz
6qz
+
(2.37)
whereas the quantities P, defined from c )11. 's, have zero
derivatives.
The method for the minimization of the energy is the
classical conjugate gradient metbod with the variable
metric, developed by Murtagh and Sargent 20 • The SCF iteration procedure is performed until the energy converges within 10- 6 eV and the optimization is stopped when the minimum
relative quadratic difference allowed for two consecutive
values of atomie coordinates is smaller than 10 8
11.4
Ab-initia aaZoulations 21
The CNDO and EH method use a minimal basis set of
Slater-type atomie orbitals (STO's). Full Slater-type
ca:culations are, however, time consuming, largely because
of the evalustion of two-electron integrals. Replacing each
STO by a linear combination of a small number of Gaussiantype orbitals, is a possibility to reduce the computation
time, since integrals invalving Gaussian functions can be
evaluated analytically. The combination of K Gaussian-type
orbitals (K
2-6) are obtained for STO with ç = 1 and then
uniformly scaled. Thus
tjJ )1
r (
Ç'
where
tjJ 15
I (
1 >!)
1;3/2<1:)1' (1 ,1;.!:_)
K
E
k
<ilzs'(l,.!:_)
K
E
2 s ! ( 1 '.!:_)
K
E
<IJ
k
k
(2.38)
d1s,kg1s(a1k,
dzs,kg1s(a2k'
d2p,kg2p(a2k'
(2.39)
31
Here g 15 and g 2p are the Gaussian-type orbitals:
~
2
(2a/TI) 4 exp (-ar )
2
(128a 5/TI 3 )l r exp(-ar )cose
(2.40)
The constants d and a in (2.39) are chosen to minimize the
integrals
(2.41)
Values for a and d along with the corresponding E values
are given by Hehre et aZ. 21 • With the basis functions (2.39)
the total energy can be obtained as described in Section
II.1. In this study the Gaussian 70 program 22 is used with
an ST0-3G basis set.
MuZZiken popuZation anaZysis 23
II.S
The density matrix is defined such that, if
l:c .cp (~. is the ith MO),
f.l )..11
).I
~i
1
- the diagonal element of the density matrix is
P
).I )..I
M
2
l: n.c .
i= 1
1
(2.42)
)..11
where M is the number of occupied MO's
accupation number of the ith MO
~i
and ni is the
- and the off-diagonal element of the density matrix is
M
P
fl\!
l:
i= 1
n.c .c .
1
]..11
(2.43)
\!1
The customary Mulliken definitions qf population analysis
are:
- the net atomie QOpulation in atomie orbital ).I:
NAP(ll)
32
pllll (since sllll
1)
(2.44)
- the net atomie population on atom A:
NAP(A) = E
N
m
E P S
i::1 1111 1111
NAP(~)
)l
Khere Nm is the number of AO's on A
- a measure for the interaction between
(on atom B):
p
~
Jl
s
(on atom A) and
(2.46)
].IV 11V
- the !Otal
~verlap
TOP(AB) =
EOpulation between atoms A and B:
E
Jl,V
II.6
(2.45)
CaZcuLation
p
s
11V jlV
(2.47)
the solvation enthaLpy
The first approximation which has been used for the
solvation of monoatomie ions is the Born charging energy
term 24 • The model for the solvent effect proposed by Jano 25
considers the molecule to be enclosed in a sphere which is
embedded in a polarizable solvent. The medium is characterized as a continuurn by its dielectric constant Ë and the
solute molecule is represented by charge particles qA
situated at fixed points !.A inside the sphere, which has
radius a (Figure 2.1).
Based on the electrastatic energy 26 of a charge distribution (2.48), Jano 25 derived a similar equation (2.49)
for the solvation enthalpy as equation (2.50), which bas
been proposed by Boytink et aZ. 27 •
U
=l
p(A)V(A)dv
(2.48)
with - U the electrastatic energy of a charge distribution
p(A)
- V(A) potential at point A.
(2.49)
(2.50)
33
2.1
IZZustration of the cZassiaaZ eZeatrostatia
modeZ
with QA, QB: net charges on atoms A and B
distance between atoms A and B
effective radius, corresponding with the
spherical cavity of the Ath ion which depends
on the dielectric medium
E
dielectric constant of the solvent (EH 0 = 80)
2
Comparison of (2.49) and (2.50) gives an approximation for
the integral yAB ~ rAB-l' and for the monoatomie integral
yAA ~rA-l' BortreZ and GueriZZot 28 worked out the algorithms for the EH program 29 •
Equation (2.50) expresses solely the electrastatic
contribution from the total solvent-solute interaction.
Moreover, it does riot involve any solvent effect upon the
electronic structure of the solute molecule. This effect
can be taken into account by incorporation of the solvent
parameters into the Hamiltonian for solute molecules.
V
Comparison of both approaches by Miertu~ and KyseZ 30 shows
that the major part of solvation stabilization is due to
34
the electrastatic contribution. Thus as an approximation
of the solvation enthalpy, the electrastatic part, calculated
according to equation (2.50), has been used.
The calculations have been performed on the Burroughs
ï700 Computer at the Computing Centre, Eindhoven University
of Technology.
35
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
36
J.E. Lennard-Jones, Trans Faraday Soc.,~. 668 (1929).
R.S. Mulliken, J. Chem. Phys., l• 375 (1935).
C.C.J. Roothaan, Rev. Mad. Phys., 23, 69 (1951).
J.A. Pople, Trans Faraday Soc., 49, 1375 (1953).
J.C. Slater, Phys. Rev., ~. 509 (1930); 34, 1293 (1959).
R. Hoffmann, J. Chem. Phys., ~. 1397 (1963).
G. Govil, J. Chem. Soc. A, 2464 (1970).
W.C. Herndon, Progr. Phys. Org. Chem., ~. 154 (1972).
B.J. Duke, Theoret. Chim. Acta (Berl.), ~. 260 (1968).
R. Rein, N. Fukuda, H. Win, G.A. Clarke and F.E. Harris,
J. Chem. Phys., 45, 4743 (1966).
H. Basch, A. Visté and H.B. Gray, Theoret. Chim. Acta
(Berl.), l• 458 (1965).
D.B. Boyd, Ph.D. Thesis, Harvard University, Cambridge,
Mass., 1967.
L.S. Bartell, L.S. Su and H. Yow, Inorg. Chem., ~.
1903 (1970).
D.B. Boyd and W.N. Lipscomb, J. Chem. Phys., 46, 910
(1967).
J.A. Pople and D.L. Beveridge, "Approximate molecular
orbital theory", McGraw-Hill Book Company, New York,
N.Y., 1970.
See reference 15, pp 57-59.
J.N. Murrell and A.J. Harget, "Semi-empirica! selfconsistent field molecular orbital theory of molecules",
Wiley Intersciences, London, 1972, pp 34-101.
H.H. Jaffé, Acc. Chem. Res., ~. 136 (1969).
D. Rinaldi, Comput. and Chem., 1, 109 (1976); Program
290, Quanturn Chemistry Program Exchange, Indiana
University.
B.A. Murtagh and R.W.H. Sargent, Comput. J., ll• 185
(1970).
W.J. Hehre, R.F. Stewart and J.A. Pople, J. Chem. Phys.,
~. 2657 (1969).
22. Gaussian 70, program 236, Quanturn Chemistry Program
Exchange, Indiana Univers
23. R.S. Mulliken, J. Chem. Phys., Q, 1833, 1841, 2338,
2343 (1955).
24. W.M. Latimer, K.S. Pitzer and C.M. Klansky, J. Chem.
Phys., l• 108 (1939).
25. I. Jano, C. R. Acad. Sc. Paris,! 261, 103 (1965).
26. J.G. Kirkwood, J. Chem. Phys., l· 351 (1934).
27. G.J. Hoytink, E. de Boer, P.H. v.d. Mey and W.P. Weyland, Reel. Trav. Chim. Pays-Bas, ~. 487 (1956).
28. A. Bortrel and C.R. Guerillot, C. R. Acad. Sci. Ser. C,
27 ' 1663 (1973).
29. P. Dibout, EHT-SPD, program 256, Quanturn Chemistry
Program Exchange, Indiana University.
V
V
30. S. Miertus and 0. Kysel, Chem. Phys., 21, 27 (1977).
37
CHAPTER 111
The
solvent effect
of hydrolysis
III.l
on the
of
enthalpy
c-AMP
Int~oduation
The coenzyme cyclic adenosine 3',5'-monophosphate
(c-AMP), which acts as a "second messenger", has been recognized in the past years 1 as a key substance in the
regulation of many metabolic processes. lts level of concentration in the cell is controlled by the enzyme adenylate
cyclase, which catalyzes the conversion of ATP into c-AMP.
The conversion of c-AMP into adenosine 5'-monophosphate
(5'-AMP) takes place via a phosphodiesterase. Most phosphodiesterases degrade c-AMP solely into 5'-AMP. The isolation
of a phosphohydrolase from Enterobaater aerogenes 2 has made
hydralysis possible, which delivers a mixture of 5'-A}1P and
adenosine 3'-monophosphate (3'-AMP), as stuclied by WestheimeP et aZ. 3 • The hydralysis to either 3'-AMP or 5'-A,\1P
involves a large exothermic Gibbs free energy 3 (-37.2 kJ/
mole) and enthalphy 4 (-46.4 kJ/mole). Both values areabout
9 kJ/mole more negative than the values for the hydralysis
of "energy rich" ATP into ADP and inorganic phosphate under
the same conditions (Table III.1). The 3'-ester and 5'-ester
bond of c-AMP have been concluded to be "high energy" bonds 3 •
The discovery of phosphohydrolase from Enterobaater aerogenes
affords the possibilities to measure the enthalpies of
hydrolysis of monocyclic and acyclic phosphatediesters 4 •
Joint consideration of hydralysis enthalpies and geometries
of the alkyl phosphates and the nucleotides leads to some
useful and general conclusions. The hydralysis data 4
(Table III.l) obtained for acyclic and monocyclic phosphates
38
are in agreement with their structures. The more negative
Table III.1
Enthalpies of hydralysis and OPO bond angles
of phosphate diestersa
phosphate d.1ester b
llHobsd
a
(kJ/mole)
OPO bond
angle (degrees)
ATP
c-A~!Pd
-37.2
-46.4
-
cyclic guanosine 3 1 > 5 I monophosphate (_~)
cyclic uridine 3 t '5 t -
-43.9e
-
monophosphate CD
methyl S-D-ribofuranoside 3,5cyclic phosphate (_1)
cyclic adenosine 2 t '3'monophosphate
diethyl phosphate (_2)
ethylene phosphate (IJ
trimethylene phosphate (~)
tetramethylene phosphate (~)
dimethyl phosphate C.!Q)
-49.4
103
46.0
-38.1
-
7.5
-26.8
1 2. 5
- 9.2
- 7.3f
96
102
98
104
107g
10 5
~The values refer to the hydralysis of singly charged di-
esters to form singly charged monoesters. bFor structures
see Figure 3.1. 0 Hydrolysis enthalpy, measured at pH= 7.3,
25 °C by microcalorimetry 4 • dHydrolysis of c-AMP to 3'-AMP
and 5'-AMP with similar enthalpies. eReference 4. tReferenee
5. gReference 6.
of hydralysis of five-membered cyclic phosphates
entha
relative to acyclic phosphate esters have been attributed
to strain 7 • 8 , which is correlated with the OPO bond angles 4
(Table III.1).The exothermicities of the hydralysis of
cyclic 3',5'- and 2',3'-nucleotides suggest that these
phosphodiesters may be strained with the farmer being more
strained. In contrast to the enthalpies of hydrolysis,
several independent observations 9 - 13 show that cyclic 2' ,3'39
nucleotides are more strained than cyclic 3' ,5'-nucleotides
with respect to their products of hydrolysis.
OH
Y
R
3'
0
-~
/
:yP".
0
3.1
0
0
s'
'
5'
1 R = adenine
2 R = guanine
uridine
~ R
methyl
R
.i =
11 R = H
5 R
§_ R = ethyl
10 R =methyl
7
= adenine
n =2
8 n =3
9 n = 4
Struature of phosphate dieeters
So the large exothermic enthalpy of hydralysis of c-AMP is
rather unexpected and can not be explained from the geometries of cyclic 3' ,5'-nucleotides as well as strain
energy calculations 14 • In order todetermine the cause of
the pronounced exothermicity of the hydralysis of cyclic
3' ,5'-nucleotides, the methyl riboside cyclic phosphate (i)
has been examined by Westheiroer et aL. 4 , since it would
40
eliminate any possible effect of the heterocyclic bases
present in the nucleotides. The data for 2, ~ and ± clearly
show that the base is nat responsible for the large exothermic enthalpy of hydrolysis. As suggested by Westheimer
e~ aZ. ~. a contribution to the enthalpy of hydralysis by
so:vation may account for this phenomenon. In order to
give a qualitative basis to this idea, the effect of the
solvent on the enthalpies of hydra
is of various phosphate
diesters (~-~. lQ, ll) and of
of c-AMP, with
oxygen atoms replaced by methylene groups
3.2),
has been examined by the semi-empirical Extended-Hückel
methad (EH) and its iterative variant procedure (Chapter
I I) .
III.2
Geometries of phosphate diesters
The geometriesof ethylene phosphate 15 (2), trimethyl
ene phosphate 16 (~), the 5'-methylene analogue of c-AMP 17
(12), their products of hydrolysis, 3'-AMP 18 13 and the
3'-methylene analogue of 3'-AMP 19 (~) arebasedon X-ray
crystallographic data. Those for c-AMP 20 (l), 5'-AMP 21
diethy: (~) and dimethyl 22 phosphate (lQ) are based on
quanturn chemical calculations. The geometries of the 3'methylene analogue of c-AMP (12_) and the 5'-methylene
analogue of 5'-AMP C.U.) are estimated from c-AMP and 5'.-\~!P. The conformation of the ribose ring is taken to be
the same in the cyclic and acyclic compound. The geometries
of the 1' -methyl ene analogue of c-Al\1P, 5' -AMP and 3' -AMP
1~, 12• 18, respectively) are based on those of c-AMP and
its products of hydrolysis, wherein the ribose ring is replaced by a cyclopentane ring 23 •
is of X-ray crystallographic data, NMR studies and
theoretical calculations offer an understanding of the
and possible conformations of nucleotides. The
possible conformations of c-AMP, 5'-AMP and 3'-AMP will be
discussed here. The notations and conventions for the
internal rotations as proposed by SundaraZingam 24 are
adapted. From X-ray crystallographic and NMR data it is
41
OH
0
-""'
-,
3'
p/
o-'l' "
Y
R
.
os-
16
17
Figure 3.2
42
Methylene analogues of a-AMP, 5'-AMP and 3'-AMP
found that the torsional angle (for definition see Figure
3.3) about the glycosidic bond C(1')-~ (x), defining the
relative orientation of the base with respect to the sugar,
D
A
\
a
s-s-c
/
(al
re 3.3
(b)
Definition of rotatien angle a. Torsion angZe
about the bond B-C in the sequenoe of atoms
A-B-C-D is the angle through whioh the bond
C-D is rotated with respect to the near bond
A-B; a is oonsidered positive for a righthanded rotation. (a) viewed perpendicula:r> to
the bond;,
(b) Newman projection
is in the anti region (-90° <x< 90°), which corresponds
to x = 3.8°, 25.7° and 50° for 3'-AMP, 5'-AMP and c-AMP,
respectively. The ribose ring has a C(3')-endo conformation
in 5'-A~.IP
and 3'-AMP 24 , whereas that of c-AMP is the
C(~')-exo-C(3')-endo conformer 20 • 25 • 26 (Figure 3.4). The
conformation of the sugar ring in the 3'- and 5'-methylene
analogue of c-AMP and their products of hydralysis is
C(3')-endo C(Z')-exo and C(3')- ndo-C(4')-exo, respectively.
Literature data 20 • 25 • 26 reveal that the phosphate ring in
c-AMP and the 5-methylene analogue is fixed in a chair
conformation. The Newman projections 1-III, IV-VI and VIIIX, shown in Figure 3.5, illustrate the preferred conformations constrained along the C(4')-C(5'), C(5')-0(5')
and C(3')-0(3') bands, respectively, in nucleotides. The
conformational studies of nucleotides in recent years 27
show that 5'-AMP exists predominantly in the gauohe-gauche
conformation about the C(4')-C(S') bond (I) and the C(5')0(5') bond (IV), with dihedral angles of about 60°. An
24
43
r------,
I
I
I
I
I
/*Os·\
0,-
C3 •
Hs·b
Hs·a
1
I
I
I
0,· *Hs~ C
\
3•
Os
H••
I 99
Hs.
b
c,~c,
I
/*Hsb\
C3 •
Hs.a
Os'
H.·
H••
II gt
III tg
VI t'g'
kPo
c..
r-----,
o,.
y g't'
IV g'g·
HJ'
r------,
I
·
C2 -
3
o3 P~
C4 •
C2 •
P0 3
VII t#
F:gure 3.E
44
VIII g" ( -)
IX g" ( t)
Newman projectionsof the conformation of the
ribose-phosphate backbone chain in 5'-AMP and
3'-AMP (g: gauche; t: trans)
c•.
C{l.')
3
C(3')- endo- C(!,')- exo ( T4
?~g~re
3.4
exo
C{3') -endo( 4T
3
)
)
Puckering of the ribose ring
important stereochemical consequence of a S-S'-nucleotide
existing in the gg-g'g' conformation is that the atoms
H(4'), C(4'), C(S'), O(S') and Pare in the same plane and
that the four bond coupling path between H(4') and P is the
familiar "W" conformation (X). Hall et al-. 28 • 29 have shown
45
that the magnitude of the long-range coupling constant
4J(POCCH) exhibits a maximum of about 2.7 Hz fora planar
"\1!" conformation and that this magnitude decreases with
other conformations to zero coupling. For 5'-A.MP the observed values of 4J(PH(4')) are between 1.7-2.0 Hz 30 • The
mag_nitude of the 3J(HCOP) value 31 of 3'-A.MP indicates that
the conformation about the C(3')-0(3') bond corresponds to
that in which the phosphate group is gauche to H(3') (IX).
From X-ray crystallographic data 18 it has been found that
the orientation of the C(5')-0(5') bond in 3'-AMP with
respect to the ring bonds C(4')-0(1') and C(4')-C(3') is
gauche and trans, respectively, with dihedral angles
0(1')-C(4')-C(5')-0(5') of 57° and C(3')-C(4')-C(5')-0(5')
of -172° (II). The lowest energy conformation 32 of 5'-Ai\IP
and 3'-AMP is in good agreement with the structure determined by X-ray crystallography 18 • 33 and by 31 P- and 1HNMR studies of 3'- and 5'-nucleotides in solution 27 •
Because of the slight influence of the base on the hydrolysis4, ribofuranoside 3,5-cyclic monophosphate Cll) and
the corresponding products of hydralysis are chosen as a
simplified model for c-AMP, 5'-AMP and 3'-AMP, respectively.
Por the methylene analogues the same simplification is
adapted. The conventional numbering system 24 for c-AMP is
used.
III.3
The effect of the solventand the ribose ring
puckering on the net enthalpies of hydralysis
III.3.1
Calculation of net enthalpies of hydralysis and
net solvation enthalpies
As suggested inSection III.1, solvation may contribute4 to the large exothermic net enthalpy of hydralysis of
c-AMP. Molecular orbital calculations have been performed
on various phosphates (~-~. lQ-}l, 1±. ~) and their
products of hydrolysis, using the semi-empirical EH methad
and an iterative variant of this methad (Chapter II). The
charge distribution (for the net charges of some important
atoms in 11. 14 and 16 see Table III.2), determined by
46
both methods, tagether with the known atomie distances are
used to calculate the solvation enthalpy according to
equation (2.50).
Table III.2
compound
14
Net charges (in electron units)a
p
0.35
0.40
0.22
0. 31
0. 18
0.33
. b on
charge d ens1ty
0 ( 3')
0 ( 1 ')
0(5')
-0.83
-0.48
-0.85
-0.48
-0.84
-0.47
0.42
-0.37
-0.41
-0.41
-0.80
-0.39
-0.42
-0.39
-0.81
-0.40
-0.41
-0.42
0(6,7)
-0.67
-0.35
-0.66
-0.36
-0.67
-0.38
aObtained by a Mulliken population analysis. bThe values in
the first line are obtained by the EH method, those in the
second line by the iterative variant.
In Table III.3 the calculated net enthalpies of hydralysis
in the gas phase (~Hgca c ) and enthalpies of solvation
1
(6Hs 0 1 v) are given for the hydralysis reactions. From these
ca 1 c
values the net enthalpies of hydralysis in salution
(LHsoln) are calculated according to
oln = ~Hg
+
ca 1 c
~H~~Î~· Comparison of the results obtained with the EH
method and the iterative variant reveals that the results
of the latter metbod are in a better agreement with the
experimental data. The correlation lines between 6Hsoln
and the experimental data are shown in
3.6. Both
methods give good correlation between the experimental
and calculated values, the correlation coefficients being
0.995 and 0.998 for the normal and iterative EH method,
respectively. The slope of the correlation lines is 2.95
and 2.30, respectively. This is the multiplicative factor
by which the experimental and calculated values are interrelated. This factor can be ascribed to the EH methad as
shown by Herndon 3 5 •
47
-t::.
Hsoln
t
(kJ/motel
140
.
u-~
11-13"
120
100
80
60
4o
s-2o.
20
10
20
30
40
50
- t::. H ex p ( kJ I mo!e )
Figure 3.6
Correlation between 6H 8 oln and AHexp for the
hydralysis of phoaphate dieeters (· EH results,
x results for iterative variant). For numbering of aompounda see Figure 3.2 and J.?
The calculated net enthalpies of solvation in Table
III.3 indicate that solvation has an effect on the hydralysis enthalpy. Moreover, this effeot is aonsiderably
larger in the aase of the hydralysis of o-AMP with respect
to the other phosphate diesters. This larger exothermicity
is due, as is shown inSection III.4, to an extra stabiZization of the hydralysis product with reapeet to the
reactants by regio-specific hydragen bonding with water
molecules.
48
Table III.3
Experimental and calculated net enthalpies of hydralysis aml net solvation
enthalpiesa
with the Eli method
with the iterative variant
c llHsolv d .6.Hsoln e
.6.Hg
c llHsolv d llHsoln e Mig
.6.H exp b
calc
calc
calc
calc
hydralysis of
kJ/mole kJ/mole kJ/mole kJ/mole kJ/mole kJ/mole kJ/mole
model c-AMP
f
(J
ethylene phosphate
trimethylene phosphate
di ethyl phosphate
dimethyl phosphate
1'-methylene analogue c-AJv!Ph
i
3 -methyl ene analogue c-AMP
5 1 -methyl ene analogue c-AMP
1
-46.4
-46.4
-26.8
-1 2. 5
- 7.5
- 7. 3
-
-
-108.8
-112.1
- 55.6
- 33.9
- 3.8
- 3. 1
- 83.5
- 79.8
- 9 4. 1
- 6 8. 1
-21.8
-19.7
- 9.6
- 4.0
- 9. 2
- 9. 2
9.5
-10.8
-17.7
-11 . 6
-130.6
-131.8
- 65.2
- 37.9
- 13.0
- 1 2. 3
- 93.0
- 90.6
-111.8
- 79.7
-79.4
-81. 3
39.7
-19.8
- 1.5
- 1. 8
-48.3
-46. 1
-40.3
-52.5
-24.2
-25.0
-11.9
- 9. 8
-11.0
-11.2
-12.3
-13.9
-12.0
-18. 1
-103.6
-106.3
- 51.6
- 29.6
- 12. 5
13. 0
- 60.6
- 60.0
52.3
- 70.6
-
-
aAll values refer to the hydralysis of singly charged diesters to form singly charged monoesters (the lowest energy conformations of the compounds are given in Figure 3.7). bMeasured
net enthalpy of hydralysis in solution~ cNet enûhalpy of hydralysis in gas phase. dCalculated
net enthalpy of solvation. eNet enthalpy of hydralysis in solution 34 • fHydrolysis tomodelof
5 1 -AMP. gHydrolysis tomodelof 3 1 -AMP. hHydrolysis to 1 1 -methylene analogue of 5 1 -AMP.
iHydrolysis to 1'-methylene analogue of 3'-AMP.
,-10
Dol)/
H
'I
'/p""-
0
0
~~~c
'
1
~o-e,;
'c'
'\I
I
Q
l
'
R
c-c.-
'
11
R
= H, model c- AMP
\
:
H
H
R =H, model 5'- AMP
11.
H,
0
H-O
I
"- .,;o,'c/ G·'c/
P
;
A·~·l
'f\1
c- .01 " R
O '-'o
I
"....c,
I 0
I
H
R = H, model 3'- AMP
13
•
0-H
...__..
/
/C-C-,_
,0
,- •\
Oà.ö!.•
H-0
7
19
X
P-O
/
2- hydroxyethyl phosphote
•
\
'~-0/
H
'-è-c/
_,9,
-,J\
0-P-0
/
\'-,
~
. H-0/
8
Q,,
_"I
-
\
l
'p
0
20
/0-R
3- hydroxypropyl phosphate
Q,,,, /0-R
-
I
p
ót'l
Hl' a
0
R;/
mono-olkyl phosphate
R = ethyl
10 R = methyl
.§_
Figure 3.7
50
21
22
R = ethyl
R =methyl
Lowest energy structures of phosphate diesters
and their products of hydralysis
III.3.2
Ribose ring puckering
exoWestheimer et al.~ have shown that the
thermic enthalpy of hydralysis can not be ascribed to
strain in the phosphate diester ring. No attention has
been given hitherto to the difference in conformation of
the ribose ring of the nucleotides. The ribose ring has a
C(3')-endo conformation for 5'-AMP and 3'-AMP, whereas that
of c-AMP exists in a C(4')-exo-C(3')-endo conformation
(F
3.4). Calculation of the difference in total
enthalpy of a free tetrahydrofuran ring in the C(4')-exoC(3')-endo and C(3')-endo conformation shows that the
difference in total enthalpy is 22.4 kJ/male and 18.8
kJ/mole for the EH and the iterative variant, respectively.
The C(3')-endo conformer is the most stable one. This imp" aates that a part of the differenae in net enthalpies
o;
lysis can be explained by the
of ribose
pAakering in c-AMP, 5'-AMP and 3'-AMP. This conclusion is
underlined by the data of the methylene analogues of c-AMP.
The geometries of the ribose ring have been taken the same
in the cyclic compound and the product of hydrolysis. In
the next Chapter the contribution of both aspects, i.e.
solvation and ribose ring puckering, in the net enthalpy
of hydralysis will be considered.
III.4
Discussion
In Figure 3.7 the lowest energy conformations of the
various phosphate diesters and their products are drawn.
Hydragen bonding with water molecules in the phosphate diesters occurs between the oxygen ligands of phosphorus.
This is confirmed by ab-initia calculations on the dimethyl
phosphate anion (lQ), performed by Pullman et al. 36 •
Between the tetrahydrofuranyl oxygen and the phosphate
ester oxygens in c-AMP no hydragen bonding is possible
because of the large distance (4.0 R). A linear structure
of the hydralysis products has the lewest energy. The
hydragen bonding between the oxygen ligands remains. In
51
3-hydroxypropyl phosphate (lQ) and 2-hydroxyethyl phosphate
(~) no effective hydragen bond interaction is possible
between the hydroxyl oxygen and the phosphate ester oxygens.
Hydragen bonding via one water molecule between 0(1') and
0(5') is possible in 3'-AMP as wellas in 5'-AMP, with a
distance between these atoms of 3.0 ft and 2.9 ft, respectively (Figure 3. 8). This stabilization does not exist in c-A~IP
14 R
=H
Figure 3.8
11
R
=H
5'-AMP and 3'-AMP with one moZeauZe of water
between 0(1') and 0(5')
and other phosphate diesters or their products of hydralysis. The hydragen bonding via one water molecule in 3'AMP and 5'-AMP must be responsible for the larger net
exothermic solvation enthalpy for the hydrolysis of c-M1P.
The calculations of the enthalpy of solvation show that
the difference in net solvation enthalpies (10-12 kJ/mole
from the EH method and 12-15 kJ/mole from the iterative
variant) arises mainly from the hydragen bonding via one
molecule of water between 0{1') and 0(5'). This regiospecific solvation contributes namely 10 kJ/mole and 12
kJ/mole for the EH method and the iterative variant,
respectively. The net solvation enthalpies for the hydralysis of the methylene analogues of c-AMP underline this
52
conclusion. When either 0(1') or 0(5') is replaced by a
methylene group, the net solvation enthalpies diminish
by about 12 kJ/mole. Upon replacing the 0(3') atom by a
methylene group a difference of only 2 kJ/mole is observed.
This smal! difference may be due to the different hydratien
possibilities of 0(2') and 0(3'). Berthod and Pullman 37 have
shown by ab-initio calculations on a free ribose ring that
hydratien can occur on 0(2') and 0(3') separately, but
that a bridge position is also possible.
It seems that our proposed model is strongly supported
by 13 C-NMR spin lattioe rela:r:ation times (T 1 's), T1 data
give an opportunity to analyze intermolecular effects, e.g.
hydragen bonding 38 • Comparison of NT 1 values (:\: number of
directly attached protons to the carbon atom) indicate the
presence of anisotropy of segmental motion in a molecule.
As an example, Czarnieaki and Thornton 39 compared the NT 1
values of the exocyclic carbon atom of galactose (24) and
glucose (~) derivatives (Figure 3.9) of ganglioside head
groups with each other and with the ring carbon atoms.
They found that the exocyclic carbon atom of the glucose
compound is isotropie with respect to the ring carbon
atoms and explained this phenomenon by intermolecular
OH
HO
Figure 3.9
R,
R,
HO
a
R1
::
b
R,
= OCH 3 ;
H ; R2
= OCH3
R2 = H
Gluaose and galactose derivatives
hydragen bonding via one molecule of water between the exocyclic CH 20H group and the pyranose ring oxygen atom. For
53
5'-AMP the NT 1 values of the ribose carbon atoms, as determined by Norton and Allerhand 40 , are between 0.18-0.22 s,
whereas the value of the exocyclic C(5') atom is 0.20 s.
This implicates that the degree of rotational freedom for
C(5') is as large as that of the ring carbon atoms. This
iso~ropy underlines our proposed model for the occurrence
of regio-specific hydration between 0(1') and 0(5') in
5'-AMP via one molecule of water.
X-ray arystaZZographia data 41 of the aminoglycosyl
antibiotic puromycin dihydrochloride pentahydrate, which
is a modification of 3'-AMP, wherein the phosphate moiety
is replaced by p-methoxy-1-phenylalanyl amino group,
demonstrate that hydragen bonding does occur between 0(1 ')
and 0(5'). Although in this compound a phosphate group is
replaced by a large amino acid group, the conformation of
the 3'-amino-acyl nucleotide is in accordance with the
preferred conformations of acyclic nucleotides 24 • It seems
that puromycin provides a good conformational model for
this compound.
In eoncZusion, the large exothermia enthaZpy of
hydralysis of c-AMP is primarily due to bath the differenae
in aonformations of the ribose ring in c-AMP and its hydroZytia produats as well as the relatively large net negative
solvation enthalpy of the hydrolysis.
54
Beferences and Notes
1. J.P. Jost and H.V. Rickenberg, Annu. Rev. Biochem.,
4
741 (1971).
2. J.A. Gerlt and G.J.R. Whitman, J. Biol. Chem., 250,
5053 (1975).
3. 0. Hayiaishi, P. Greengard and S.P. Colowick, J. Biol.
Chem., 246, 5840 (1971).
4. J.A. Gerlt, F.H. Westheimer and J.M. Sturtevant, J.
Biol. Chem., 250,5059 (1975).
5. J.P. Guthrie, J. Am. Chem. Soc., 99, 3991 (1977).
6. C.L. Coulter, J. Am. Chem. Soc.,
, 4048 (1975).
;, P.C. Haake and F.H. Westheimer, J. Am. Chem. Soc.,
8.
9.
10.
11.
12.
13.
1+.
15.
16.
17.
18.
' 1102 (1961).
F.H. Westheimer, Acc. Chem. Res., l• 70 (1968).
S. Chang, D. McNally, S. Shary-Tehrany, M.J. Hickey and
R.H. Boyd, J. Am. Chem. Soc.,
, 3109 (1970).
C.C. Browne and F.D. Rossini, J. Phys. Chem.,
' 927
(1960).
J. Kumamoto, J.R. Cox, Jr. and F.H. Westheimer, J. Am.
Chem. Soc., 2..!!• 4858 ( 1956).
H.G. Khorana, G.M. Teuer, R.S. Wright and J.G. Moffat,
J. Am. Chem. Soc., .z.g,, 430 (1957).
P. Szabo and L. Szabo, J. Chem. Soc., 3758 (1960).
J.A. Gerlt, Ph. D. thesis, Harvard University, Cambridge, Mass. (1974).
T.A. Stertz and W.N. Lipscomb, J. Am. Chem. Soc.,~.
2488 (1965); in this paper the labelling of atoms in
Table I and Figure 1 is contradictory. In Table I the
set 0(2), 0(3), 0(4), 0(5), 0(6), C(7), C(8), should
be replaced by 0(3), 0(6), 0(7), 0(2), C(4), C(5),
0(8). The correct value for y 0(5) = -0.2137.
Mazhar-ul-Hague, C.N. Caughlan and W.L. Moat, J. Org.
Chem., ~. 1446 (1970).
S.M. Hecht and M. Sundaralingam, J. Am. Chem. Soc.,
~. 4314 (1972).
(a) M. Sundaralingam and L.H. Jensen, J. Mol. Biol.,
55
~f. Sundaralingam, Acta Cryst., t.J.,
495 (1966).
M. Sundaralingam and J. Abola, J. Am. Chem. Soc., 94,
5070 (1972).
J.N. Lespinasse and D. Vasilescu, Biopolymers, 1}, 63
(1974).
D. Vasilescu, J.N. Lespinasse, F. Camous and R. Cornillon, FEBS Letters, lZ• 335 (1972).
D.G. Gorenstein, J.B. Findlay, B.A. Luxon and D. Kar,
J. Am. Chem. Soc., 99, 3473 (1977).
Tables of interatomie distances and configuration in
molecules and ions, A.D. Mitchell and L.C. Cross Ed.,
Special Publication no. 11, The Chemical Society,
London, 1958; M 185.
M. Sundaralingam, Biopolymers,
821 (1969).
B.J. Blackburn, R.D. Lapper and I.C.P. Smith, J. Am.
Chem. Soc.,
, 2873 (1973).
R.D. Lapper, H.H. Mantsch and I.C.P. Smith, J. Am.
Chem. Soc.,
, 2878 (1973).
(a) C. Altona and M. Sundaralingam, J. Am. Chem. Soc.,
~. 2333 (1973); (b) F.E. Hruska, D.J. Wood, T.N.
McCaig, A. Smith and A. Holy, Ca~ J. Chem., ~. 497
(1974); (c) R.H. Sarmaand R.J. Mynott, J. Am. Chem.
Soc., ~. 1641 (1973); (d) c. Altona, J.H. van Boom,
J.R. de Jager, H.J. Koeners and G. van Binst, Nature,
247, 558 (1974).
L.D. Hall and R. B. Malcolm, Can. J. Chem., ~. 2092,
2102 (1972).
B. Donaldson and L.D. Hall, Can. J. Chem., ~. 2111
(1972).
R.H. Sarma, R.J. Mynott, D.J. Wood and F.E. Hruska,
J. Am. Chem. Soc., 95, 6547 (1973).
O.B. Davies, Progr. in NMR Spectroscopy, }l. 135 (1978).
J.M. Thornton and P.M. Bayley, Biochem. J., 149, 585
(1975).
J. Kraut and L.H. Jensen, Acta Cryst., ~. 79 (1963).
P. George, R.J. Witonsky, M. Trachtman, C. Wu, W. Dor-
.]1_, 914 (1965); (b)
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
56
z,
35.
36.
37.
38.
39.
~0.
41.
wart, L. Richman, W. Richman, F. Shurayh and B. Lentz,
Biochim. Biophys. Acta, 223, 1 (1970).
W.C. Herndon, Prog. Phys. Org. Chem., ~. 154 (1972).
B. Pullman, A. Pullman, H. Berthod and N. Gresh,
Theor. Chim. Acta (Berl.),
, 93 (1975).
H. Eerthad and A. Pullman, Theor, Chim. Acta (Berl.),
47' 59 (1978).
J.R. Lyerla, Jr. and G.C. Levy, Top. Carbon-13 NMR
Spectroscopy, l• 79 (1974).
~t.F. Czarniecki and E.R. Thornton, J. Am. Chem. Soc.,
~. 8279 (1977).
R.S. Norton and A. Allerhand, J. Am. Chem. Soc., 8
1007 (1976).
~1. Sundaralingam and S.K. Arora, J. Mol. Biol., .zl,
49 (1972).
57
CHAPTER IV
The
inttuenee
ribose
of
solvation
ring puckering
the
enthalpy
of hydrolysis
IV.1
Hydrogen bonding
of
and
on
c-AMP
The elucidation of the possible interactions in
solute-solvent systems is of fundamental importance for
the understanding of the structural and functional properties of biomolecules. In fact, several studies 1 are directed in an effort to introduce explicitly the solvent effect,
especially the influence of water into theoretica! computations on biomolecules. Many studies employ the "traditional" approach to the problem through the use of a
"continuum" model. This macroscopie representation, as is
used in the previous Chapter, tries to account for the
bulk effect of the surrounding medium. A second approach,
which has been much less explored, consists of a "discrete"
treatment in which one attempts to establish the individual
sites of interaction of the solvent molecules with the
system studied. The purpose of a macroscopie representation
of the solvent is to interpret or predict the molecular
behaviour in solution on a more direct basis. The essential
problem consists of determining the most probable sites of
binding of water molecules to biomolecules. Upon binding
of a water molecule with the system hydrogen bonds 2 • 3 are
formed. The hydrogen bond has been of special interest to
the chemists since Latimer and Rodebuah~ first pointed
out the significanee of hydrogen bond mechanism to describe the structure of water. The interest greatly increased when Watson and Crick 5 postulated hydrogen bonding
to be a key feature of the structure of DNA. The term
58
hydragen bond has, however, no universally accepted definition. According to the simple valenee bond theory, a hydragen atom should be capable of forming only one chemica!
bond. In many cases hydragen is attached to two atoms. In
such cases the additional bond is called a hydragen bond.
Pimental and McCleZZan 2 define it as "an interaction between
a group A-H and an atom or group of atoms B in the same or
a different molecule when there is evidence of bond
mation and that this new bond Zinking A-H and B, specificaZinvolves the hydragen atom already bonded to A". Of the
tKo functional groups taking part in the interaction, A-H
serves as a proton donor 6 and Bas a proton acceptor 6 •
~!os t commonly known groups acting as proton donors are:
the carboxyl, hydroxyl, amine and amide group. Moreover,
the hydragen atom attached to phosphorus, sulphur and
selenium, can take part in hydragen bonding. The dissociation enthalpy of hydragen honds varies from a few kJ/
mole to about 170 kJ/mole, depending on the different
functional groups which are involved.
Of the semi-empirica! methods used to study hydragen
bonding, the CND0/2 method (Chapter II) is far superior to
the EH and MIND0/3 method. The latter methods are not
suitable for the evaluation of hydragen bonding, as was
found by Murthy and Rao 7 for the EH method, and by Zie~inski et al. 8 and Klopman et al. 9 for the MIND0/3 method,
since no stabilization is observed upon water dimer
formation. Comparison of CND0/2 with ab-initia calculations
on dimers reveals that the CND0/2 results turn out to be
very satisfactory. Comparison 10 • 11 of interaction enthalpies
and intermolecular distances for dimers, calculated by
ab-initio and CND0/2 procedures, show that CND0/2 calculations
relatively good values. Hence the CND0/2 metbod
represents a reliable and relatively cheap procedure and is
as such very useful for the determination of interactions
between the solvent and large molecules, where more
accurate calculations are far outside the range accessible
by present computational facilities. For the CND0/2 equi-
59
librium geometries, however, the calculated interaction
enthalpies are too large and the intra- and intermolecular
distauces too short. CND0/2 calculations at fixed experimental geometries reproduce the available experimental
enthalpies of association very satisfactorily 12 • The
hydratien schemes of the models of c-AMP and the products
of hydralysis are stuclied with molecular orbital calculations, using the GEOMO (CND0/2) program (Chapter II).
IV.2
The water dimer
Interaction enthalpies which are more negative than
those corresponding to water-water interaction, point to
the existence of hydratien sites. In order to obtain insight into the possible sites, dimerization of water
molecules is first considered. Water dimers can exist in
three basic conformations, i.e. linear, cyclic and bifurcated (Figure 4.1). The linear hydragen bond has been
systematically studied, but less attention has been paid
to cyclic and bifurcated bonds 13 • This is surprising since
non-linear conformations occur frequently in the solid
state and very probably in biologica! systems 14 • 15 • There
linear
.Hi'\.
H;-
o··
I"
. .
s "
o.....- H
.4
6
cyclic
H3
bifurcated
Figure 4.1
60
Water dimers
has been a long-standing discussion 16 as to whether the
ground statesof simple dimers, e.g. (H 2o) 2 , are cyclic or
linear.
Calculations using the GEOMO-CND0/2 program, including
geometry optimizations, are performed on the water dimer
in the three conformations. The basic geometries, as found
to be the most stable ones by Morokuma and Peterson 17 in
ab-initia calculations, are used. In Table IV.1 the dimerization enthalpies are given with the equilibrium
oxygen-oxygen distance. The binding enthalpy of -26.0 kJ/mole
fcr the linear conformer agrees favourably with the experimental value of -21.0 kJ/mole for the water diroer in
Table IV.l
dimer
ge ometry
linear
bifurcated
cyclic
Water diroer results
R(0-0)
~
2.53
2.60
2.48
Mi dim
kJ/mole
-26.0
- 5. 5
-12.6
the gas phase 13 and of -23.9 kJ/mole per hydragen bond in
ice 13 • The enthalpy differences between the linear dimer
and the other two conformers are in excellent agreement
with the results of an ab-initia study by Morokuma and
Peterson 17 • The net charges of the atoms of the water
monoroer and dimers are listed in Table IV.2. Por the
ie
conformation the net charge transfer is zero, as is expected from the symmetry of the complex. The electron density
of the hydrogen atom, which is participating in the hydrogen
bond is reduced upon hydragen bond formation. This phenomenon
can be explained by electron pair repulsion. The electrans
of the H-0(4) bond are "pushed back" towards the 0(4) atom
by the lone pair of 0(1).
61
Table IV.2
atom
0( 1)
H(2)
H( 3)
0(4)
H(5)
H(6)
{;.b
Net charges (in e.u.)a
HzO
monomer
(H20lz
linear
(HzO)z
bifurcated
CHz0)2
cyclic
-0.266
0. 133
0. 133
-0.265
0. 151
0. 151
-0.319
0. 175
0.108
0.037
-0.268
0. 139
0. 138
-0.291
0. 141
0. 141
0.009
-0.281
0. 128
0.153
-0.281
0.153
0.128
-
-
-
o.ooo
aObtained by a Mulliken population analysis. bMagnitude
of electron transfer.
IV.3
Hydration sahemes of models of
J'-AMP
a-AMP~
5'-AMP and
Primarily the geometries of the nucleotides are optimized and subsequently kept constant throughout the
computations (for optimized geometries see Appendix A).
Interaction with one molecule of water is considered for
all possible hydratien sites. Whereas the probability that
a water molecule will detach itself from the bulk water
structure to bind directly to a hydragen atom of a methylene group is very small 18 , only hydratien sites involving the oxygen atoms and the hydroxyl hydragen atoms
are studied. The geometry parameters, which determine the
hydragen bonding of the water molecule with the nucleotides, are optimized for each interaction in order to
obtain the local minimum enthalpy. Interaction enthalpies
are calculated by substracting the sum of the enthalpies
of the isolated compounds from the enthalpy of the adduct.
In Table IV.3 the interaction enthalpies (~Hhydr), intermolecular hydragen bond lengtbs and net charges of the
water molecule are given for the models of c-AMP, 5'-AMP
62
and 3'-AMP, respectively. The optimized intermolecular
hydrogen bond distances are between 1.38 Rand 1.63 R.
These distauces are smaller than those obtained by abinitia calculations. The discrepancy can be ascribed to
the CND0/2 method, which can underestimate bond lengths up
to 10 • In general it is known that in hydrogen honds the
0 ... 0 distances range from 2.5 to 3.5 R. Less attention
has been paid to the range of the O... H hydragen bond
lengths. A~Zinger
has pointed out the difference between
van der Waals contact radii and van der Waals potential
minimum radii. The latter values rather than the former
for the criterium of a hydrogen bonding interaction are
used, the cut-off fora H... O bond is then (1.50 + 1.65)R.
All distauces of less than 3.15 Rare classified as bonding
interactions. This implies that, even when the 10% underestimation is taken into account, all hydrations considered
are bonding interactions.
The water molecule acts as a proton donor when it is
linked to an anionic, hydroxyl or phosphate ester oxygen
atom resulting in an overall transfer of electrans fróm
the nucleotide to the water molecule. If interaction with
the hydroxyl hydragen atom occurs, charge is transferred
in the reverse direction, which is characteristic for water
acting as proton acceptor 6 • A population analysis has been
made for both the complexes and the isolated monomers
(data, see Appendix B). Some interesting conclusions emerge
from these results:
- the hydrogen in the hydragen bond looses electron density
upon hydragen bonding;
the adjacent oxygen atoms gain electron density, the
oxygen atom of the proton donor molecule displays the
greatest change;
- the largest loss of electrous occurs at the hydrogen,
carbon or phosphorus atom, which is immediately attached
to the proton acceptor atom;
- all hydragen atoms, which are attached to the electronegative atom of the proton donor molecule and not in19
63
Table IV.3
numbera
of water
,molecule
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Calculated interaction enthalpies, intermolecular hydrogen bond distances and net charges on the water
molecule for hydrations of models of c-AMP, 5' -A~!P and 3' -AMP
hHhydr
b
t.Hhydr
b
t>Hhydrb
(kJ/mole)
c-AMP
(kJ/mole)
5 '-AMP
(kJ/mole)
3'-AMP
-121.8
-120.5
-126.0
55.0
- 67.2
- 68.0
63.4
- 43.3
- 39.9
-126.4
-118.9
-124.7
57. 1
- 66.4
- 66.8
- 60.9
- 41.2
- 44. 1
- 78.5
63.8
- 68.0
- 67.2
- 56.7
-126.8
118. 4
-124.7
- 34.4
66.4
- 66.8
62.6
- 54.6
42.4
78.5
62.6
- 70. 1
68.5
- 72.7
-
-----
--
-
-
I
1
t>Hhydr
a
(kJ/mole)
-115. 1
-11 5. 1
-113. 8
23.9
- 34.0
34.0
- 29.4
- 22.7
- 21.0
-
--
--
- 32.8d
- 34.4d
- 30.7/
-
31.9 6
g
o..... H distancesJ
net charges (e.u.) on water
molecule for
c-AMP
5'-MfP
3 '-AMP
-0.0818
-0.0765
-0.0836
-0.0688
-0.0480
0.0245
-0.0595
-0.0492
-0.0426
-0.0846
-0.0817
-0.0887
-0.0437
-0.0164
0.0300
-0.0174
-0.0324
-0.0362
0.0421
-0.0359
-0.0231
-0.0144
0.0341
-0.1073
-0.0845
-0.1102
-0.0641
-0.0163
0.0140
-0.0266
-0.0498
-0.0524
0.0379
-0.0401
-0.0215
-0.0325
0.0255
------
(.R)
c-AMP
5'-AMP
3'-AMP
1. 39
1.44
1. 38
1. 38 (0(7')) 1.61 (0(7')) 1.64
1.47
1. 38
1. 39
1. 45
1. 58
1. 43
1. 59 (0(2')) 1. 59 (0(2')) 1. 54
1. 53
1.53
1. 54
1.63
1.63
1.65
1. 47
1. 53
1. 49
1. 59
1.61
1. 53
1.62
1.64
1. 56
1. 58
1. 58 (0(1')) 1. 56
1. ss (0(5')) 1. 58
1.49
1. 59
-----
(0(7')
(0(2')
(0(1')
(0( 1 t)
--
aFigure 4.2 presents the numbering of the hydration sites; bhHhydr = Hadduet- (Hnucl ~ Hwater)' with Hnucl' Hwater'
Hadduet the total enthalpy of the nucleotide, the water molecule and adduct, respectively; 0 Results from other studies:
for the phosphate moiety ref. 20, for the ribose ring ref. 21; dThis work; 6 ~Hhydr = -30.7 kJ/mole as reference for the
H(0(3')) atom in 5'-AMP and hHhydr • -31.9 kJ/mole as reference for the H(O(S')) atom in 3'-AMP; fin parentheses the
oxygen atom which is considered.
model
c-AMP
R:H
model
?igure 4.2
5'- AMP
R:: H
model
3'- AMP
R =H
Hydration sehemes of modelsof e-AMP, 5'-AMP
and 3'-AMP (numbers of water moleauZes coP:respond to Table IV,3)
65
corporated in the hydrogen bond become more negative upon
hydrogen bond formation.
These results are in line with those for simple dimers 13 •
The calculated interaction enthalpies point to the
existence of a number of possible hydration sites with
interaction enthalpies more negative than the corresponding
water-water interaction (öHdim = -26.0 kJ/mole). However,
the.fact that the CND0/2 metbod neglects three- and fourcenter repulsions (Chapter II) favours conformations in
which atoms can approach closer, which leads to a larger
gain in binding enthalpy with respect to ab-initio results.
The results for the phosphate moiety are in line with those
obtained by ab-initio calculations for the dimethyl phosphate anion 20 (Table IV.3). The values for the ribose
unit are, however, more negative for the CND0/2 calculations
with respect to ab-initio computations for the free ribose
ring 21 (Table IV.3). The latter show, in contrast with our
results, that 0(1') and 0(5'), separately, do notact as
hydration sites, because the interaction enthalpy is less
negative than the dimerization enthalpy of water (öHdim =
-25.6 kJ/mole for ab-initia calculations 21 ) . The interaction enthalpies for the formation of a five- and sevenmembered ring between 0(1') and 0(5'), which are not
studied by BePthod and Pu~~man 21 , are as large as those
for the hydration of the H(0(2')) atom and the formation
of a five-membered ring between 0(2') and 0(3'). Therefore
it is assumed that the proposed hydrogen bonding between
0{1') and 0(5') exists. In addition ab-initio calculations
with the program Gaussian 70 22 , using a minimal ST0-3G
basis set, have been performed on a free ribose ring in
the C(3')-endo conformation, which is present in 5'-AMP
as well as in 3'-AMP. The geometry data of the ribose ring
in 3'-AMP 23 are employed. Formation of a five- and savenmerobered ring between 0(1') and 0(5') results in an interaction enthalpy of -34.4 kJ/mole and -33.2 kJ/mole,
respectively (Figure 4.3). The hydrogen bond distance
between 0(1') and H(H 20) was taken to be 1.74 Rand 1.83 R,
respectively. CND0/2 calculations on the roodels of c-AMP
66
4.J
Ribose ring in C(J')-endo conformation with
three main hydration sites
and the products of hydralysis have also been performed
for these distances and give results which are in better
agreement with the ab-initio results than the calculations
which perfarm optimization to obtain an equilibrium
distance. The interaction enthalpies are -43.3 and -41.6
kJ/mole for the five- and seven-membered ring between
0(1') and 0(5'). For the formation of a five-membered ring
between 0(2') and 0(3') the interaction enthalpy, obtained
with the Gaussian 70 program, equals -32.3 kJ/mole with
the 0(2')-H(H 20) distance 1.76 ~.
The ab-initio results for the formation of a five- or
seven-membered ring between 0(1') and 0(5') inthefree
ribose ring show that this interaction enthalpy is 8-9
kcal/mole more negative than that of the water dimer. This
value is in agreement with the value (10 kJ/mole) calculated
according to equation (2.50) for the solvation enthalpy.
The significanee in the interpretation of the physical
and chemical properties of nucleotides has recently become
increasingly evident. In an artiele publisbed by Bolton
and Kearna 24 , a model has been proposed for intermolecular
hydrogen bonding in c-AMP between the 2'-0H group of the
ribose ring and the free phosphate oxygen atoms (with the
0(2')-0(7') distances assumed to be 3.6 ~). They conclude
from 1H NMR spectra of cyclic nucleotides in aqueous and
mixed solvents that the 2'-0H proton is protected against
67
exchange with bulk water. On the other hand, these authors 24
could not find any crystallographic evidence for the proposed interaction. Our results and these obtained by
Berthod and Pullman 21 demonstrate, however, that hydrogen
bonding is possible via one molecule of water between the
0(2') and 0(3') atoms. Furthermore, crystallographic
evidence 21 is available to support this model. The distance
between the 2'-0H oxygen atom and the nearest phosphate
oxygen atom in c-AMP is found to be 5.0 R25 and not 3.6 R
(vide supra). As a corollary tothese data, an intermolecular hydrogen bonding, as proposed by Bolton and
Kearns, is unlikely, instead a bond between the 2'-0H group
and the 0(3') atom is indicated.
IV.4
The aontribution of solvation and ribose ring
puokering to the net enthalpy of hydralysis of a-AMP
The ab-initio results for the hydratien of 0(1') and
0(5') of the free ribose ring demonstrate that the continuurn approximation, as effered in the previous Chapter,
gives a good estimation of the enthalpy of solvation for
the position between 0(1') and 0(5'). Whereas ab-initio
computations give a better approximation of the experimental values, the difference in enthalpy between the two
conformers as present in c-AMP, on the one hand, and
5 '- AMP and 3' -M1P, on the ether, is determined by ab-ini ti o
Table IV.4
Calculated and experimental values
llHsolv (kJ/mole) llHribose puckering
calc
calc
iterative
(kJ/male) EH methad variant
(kJ/mole)
llHexp
a
b
-46.4
-46.4
-21.8°
-19.7
-24.2d
-25.0
-9.2
-9.2
trimethylene
phosphate
-12.5
-
- 9.8
-
c-AMP
4.0
aHydrolysis to 5'-AMP; bHydrolysis to 3'-AMP· 0 11Hsolv
'
calc
(0(1')-0(5'))=-10 kJ/mole·' dllHsolv
(0(1')-0(5'))=-12
kJ/mole.
calc
68
calculations with the program Gaussian 70. Tetrahydrofuran
in these two conformations is used as a model. The enthalpy
difference between the two conformers is 9.2 kJ/mole, with
the C(3')-endo conformer being more stable with respect to
the C(4')-exo-C(3')-endo conformer. In Table IV.4 the experimental and calculated data are listed for the hydralysis
of c-M-lP and trimethylene phosphate, respect i vely. Al though
one should be careful in camparing the calculated and experimental values, the data in Table IV.4 tagether with the
net solvation enthalpy for the position between 0(1') and
0(5') (L'lHsolv = -10 kJ/mole, obtained by the continuurn
calc
model, Chapter III) reveal that for c-AMP the net solvation
enthalpy and the difference in enthalpy obtained by puckering of the ribose ring are especially responsible for the
large exothermic net enthalpy of hydrolysis.
IV.5
Hydragen bonding as a model for the dynamica of
enzyme-coenzyme complexes
In the previous Chapter a seven-membered ring model
for the hydragen bonding between a water molecule and 5'-AMP
has been proposed. The similarity of the data for the
formation of a five- or seven-membered ring reveals that
the formation of the farmer is also possible (Figure 4.4).
The appearance of a five-membered ring makes a "through
water" interaction possible via the "free" proton of the
Figure 4.4
Transformation of intermoZecuZar hydragen
bonding
69
water molecule between 5'-A~W as well as 3'-AMP and other
enzymic sites and therefore it opens the possibility of
conformation change in the enzymes. The concentratien level
of c-AMP in cells is controlled by i ts conversion, via the
action of a phosphodiesterase, into 5'-AMP in a highly
exothermic reaction. c-AMP serves as a messenger that
regulates the enzymatic reactions in cells. It has also
been shown to stimulate the activity of the genes via the
synthesis of messenger RNA which in fact reproduces the
information contained in the DNA of the genes. On the other
hand, 5'-AMP acts as an inhibitor 26 for these processes. An
example is found in the enzymatic process, wherein the 3'
to 5' exonuaZease activity associated with both mammelian
and bacterial DNA polymerases is selectively inhibited 27
by nucleoside 5'-monophosphates, whereas the cyclic nucleotides and the nucleoside 3'-monophosphates show no inhibition.
It might be possible to explain these results with the
proposed regio-specific hydrogen bonding between 0(1') and
0(5'). Por 5'-AMP the information for the inhibition can be
given via the water molecule between 0(1') and 0(5'). This
kind of information is absent in c-AMP for geometrical
reasons. The fact that 3'-AMP differs considerably in its
molecular behaviour with respect to 5'-AMP is probably due
to the orientation of the binding sites for the enzyme involving the phosphate moiety, the base and the "free" proton
as characteristic for the enzyme interactions. It should be
stressed that a cooperative effect, for geometrical reasens
is probably only opera ti ve in 5' -AMP and absent in 3' -AMP.
In 1941, Lipmann 28 suggested that phosphate compounds
with large exothermic Gibbs free energies and enthalpies
of hydralysis could act as energy sourees for biologica!
systems. The data offered in the previous and this Chapter,
together with the known data of ATP, and results obtained
by Westheimer et al. 29 on cyclic 2',3'-nucleotides, indicate
that there are at least three different methods for storing
energy in phosphate compounds. These sourees are: (1) storage
in pyrophosphate bonds, as exemplified 30 by ATP and the many
70
kinase and phosphorylase enzyme systems with ATP as substrate
or product; (2) phosphate ring strain, found in the cyclic
2' ,3'-nucleoside monophosphates 29 involved in the ribonuclease-catalyzed hydralysis of RNA; (3) ribose ring enthalpy starage and solvation enthalpy, as for c-AMP. The
thermodynamically favoured hydralysis may simply ensure
that enzymatic hydralysis is an effective mechanism for
lowering the concentration of c-AMP and increasing that of
5'-A~IP, when an inhibitive activity is required.
71
References
1. P. Claverie, J.P. Daudey, J. Langlet, B. Pullman,
D. Piazzola and H.J. Huron, J. Phys. Chem., g, 405
(1978).
2." G.C. Pimenteland A.L. McClellan, "The hydragen bond",
W.H. Freeman editor, San Francisco, California; 1960.
3. "The hydragen bond", vol. I, II and III, P. Schuster,
G. Zundel and c. Sandorfy, eds., North-Holland
Publushing Co., Amsterdam; 1976.
4. W.M. Latimer and W.H. Rodebush, J. Am. Chem. Soc., 4 ,
1419 (1920).
5. J.D. Watson and F.H.C. Griek~ Nature, lil• 737 (1953).
6. M.D. Dolgushin and V.H. Pinchuck, Theoret. Chim. Acta
(Berl.), ~. 157 (1977).
7. A.S.N. Muthy and C.N.R. Rao, Chem. Phys. Lett., ~~
123 (1968).
8. T.J. Zielinski, D.L. Breen and R. Rein, J. Am. Chem.
Soc., 100, 6266 (1978).
9. G. Klopman, P. Andreozzi, A.J. Hopfinger, 0. Kichuchi
and M.J.S. Dewar, J. Am. Chem. Soc., 100, 6267 (1978).
10. P.A. Kollman and L.C. Allen, Chem. Rev., ~. 283 (1972).
11. P. Schuster, Z. Chem., ..:!.]_, 41 (1973).
12. P. Schuster, Int. J. Quanturn Chem., ~. 851 (1969).
13. Reference 3: Chapter 2.
14. Reference 3: Chapter 1.
15. R. Balasubramanian, R. Chidambaran and G. Ramachandian,
Biochim. Biophys. Acta, 21 182, 196 (1970).
16. Reference 3: Chapter 22.
17. K. Morokuma and L. Pederson, J. Chem. Phys., ~. 2275
(1968).
18. A. Pullman, Bull. Soc. Chim. Belg.,~. 963 (1976).
19. W.L. Allinger, Adv. Phys. Org. Chem., ..:!.]_, 17 (1976).
20. B. Pullman, A. Pullman, H. Berthod and N. Gresh,
Theoret. Chim. Acta (Berl.), 40, 93 (1975).
21. H. Berthod and A. Pullman, Theoret. Chim. Acta (Berl.),
i2_, 59 (1978).
72
22. Gaussian 70, program 236, Quanturn Chemistry Program
Exchange, Indiana University.
23. M. Sundaralingam, Acta Cryst., ~. 495 (1966).
24. P.H. Bolton and D.R. Kearns, J. Am. Chem. Soc., lQl,
479 (1979).
25. J.N. Lespinasse and D. Vasilescu, Biopolymers, ll•
63 (1974).
26. A.L. Lehninger, "Biochemistry" 2nd ed., Worth Publishers
Inc., :--lew York; 1975, pp 630, 712, 734.
27. J.J. Byrnes, K.M. Downey, B.G. Que, M.Y.W. Lee, V.L.
Black and A.G. So, Biochemistry, ~. 3740 (1977).
28. F. Lipmann, Adv. Enzymol., 1, 99 (1941).
~9. J.A. Gerlt, F.H. Westheimer and J.M. Sturtevant, J.
Biol. Chem., 250, 5059 (1975).
30. H.R. ~1ahler and E. H. Gordes, "Biological Chemistry",
Harper and Row, New York; 1971.
73
CHAPTER V
Thè
acidity
of thiaDJ.ine
and :related
V.1
V.1.1
py:rophosphate
systeDJ.s
Introduetion
Bistorical background
Thiamine pyrophosphate (TPP) serves as coenzyme for
various types of enzymatic reactions (Chapter I). The
nature of the molecule itself, with its combination of an
aromatic aminopyrimidine ring and a substituted aromatic
1,3-thiazolium ring, does not immediately suggest a unique
mode of reactions with substrates. Nevertheless, various
predictions have been made and tested until the currently
accepted mechanism for catalysis by TPP evolved. Langenbeek
and Hutschenreuter 1 have studied the ability of a variety
of amines to catalyze both decarboxylation and acyloin
condensation of a-keto acids, and they considered these
reactions as.models for enzymatic decarboxylations. Many
years later, Wiesner and VaZenta 2 proposed that the 4'amino group of TPP acted in an analogous manner in TPPcatalyzed reactions. They postulated Schiff base (imine)
formation, foliowed by formation of a carbanion on the
bridge carbon atom with subsequent rearrangement to form a
B-unsaturated acid, which could readily decarboxylate
(Figure 5.1 (1)). Both key points (imine formation and an
intermediate bridge methylene carbanion) in the suggested
mechanism have been shown to be highly unlikely 3 • Catalysis
of acyloin condensations by cyanide, such as the formation
of benzoin 4 , was long known, as was the condensation of
quaternary pyridinium compounds with aldehydes to yield
74
adducts at the position a to the quaternary nitrogen atom 5 •
The similarity between the thiazolium and the pyridinium
ring was recognized by Ugai et al. 5 • 6 • They found, however,
that insteadof forming adducts, the 1,3-thiazolium derivatives catalyzed acyloin condensations. These results made
7
~izuhara and Handler , and Breslow 8 suggest different
+/'-1.,
N
~§
(1)
RCHO
+
F\S
(2)
N:
x
H OH
R"'- ...rOH
c-
RCHO
5.1
+
--.R'
Proposed meohanisms of oata
l
(3)
is by TPF
75
mechanisms for the catalysis by TPP. Mizuhara and Handler
proposed that the carbonyl carbon atom of the substrate is
attacked by the tertiary nitrogen atom of the pseudo base
(Figure 5.1 (2)), forming a zwitterionic adduct that can
either cleave to give free aldehyde or condense with another aldehyde molecule to form an acyloin. However, the
acid chemistry of 1,3-thiazolium compounds has been described by Duolos and Haake 9 , and Metzler 10 , and they found
no appreciable amount of pseudo base present in solution,
at any pH. Hence this mechanism apparently did not explain
TPP catalysis. Breslow 8 stuclied various 1,3-thiazolium
compounds under the experimental conditions of Mizuhara
and Handler, measuring the formation of acetoin from
acetaldehyde and pyruvate. Ereslow concluded that the
bridge methylene group must be activated by the adjacent
quaternary nitrogen atom as well as by the adjacent aromatic
ring, and he proposed a mechanism whereby the electron
deficient carbonyl carbon atom of the substrate reacts with
a carbanion formed by dissociation of one of the protons
of the bridge methylene group (Figure 5.1 (3)). The adduct
would be cleaved or condensed in an analogous way as depicted by Mizuhara and Handler 7 • Breslow 11 established~
however, the laak of H-D exahange at the bridge methylene
CH3
CH 3
R,
+'i=<
020
R-NyS
25°
Figure 5.2
76
+H
y
R-N
S
D
H
pD 5
t , : 2 min
12
pD 7
t,
12
R1
CH3
R,
+M
R-N~ s
1
: infinitesimal
H-D exchange on paaition C(2) of 1,3-thiazolium ring
group, but
at the same time that the C(2) proton of
the 1,3-thiazolium ring did exchange readily with
5.2). He proposed the H-D exchange reaction
deu:::erium
to occur via a particularly stable conjugate base (l)· The
arguments in favour of this mechanism are supported by the
lability of the proton at position Z of the thiazolium
ring, which is demonstrated with the aid of infrared and
\MR data 12 • Breslow 12 • 13 adapted the mechanism for catalysis
of the benzoin condensation by the cyanide ion 4 to the
cata
is by TPP (Figure 5.3).
(1)
CH3
R,
.H
R-N
~
S
(2)
5.3
Acyloin condensation (1) and decarboxylation
of a-keto aaids (2)
V.1.2
Relation of struature of TPP to the aatalytic
activity
The generally accepted mechanisms for enzymatic
reactions with TPP as coenzyme suggest that the structure
of TPP has evolved so, as to stabilize carbanions on positien Z of the 1,3-thiazolium ring. In order to gain insight into the features which determine the formation of
the carbanion, Haakeet al. 1 "• 15 studied the rate constants
of deuterioxide-catalyzed generation of cations with
77
deuterium on position 2 (~) and the formation of carbanions
on position 5 (i) via the decarboxylation of N-methylated1 ,3-azolium-5-carboxylates (~). The results of this study
CH3
CH3
CH3 -~
·~
CH3-NYX
D
2
C02
.!!
b
~
x =NCH3
x =s
x ::: 0
l
-co 2
!!
x
.Q.
!:.
x =s
x :: 0
4
= NCH3
13
are shown in Table V.1. The similarity of both
c-H(2) and
13 c-IJ(5) coupling constants 14 (Table V.1) in homologous
1,3-imidazolium and 1,3-thiazolium cations indicates that
hydrogens at position 2 and 5, respectively. have similar
potential acidity in the cations. On the other hand. the
ohserved 11-D exchange and decarboxylation is much faster
for the 1.3-thiazolium cation.
Table V. 1
Rates of H-D exchange and decarboxylation. and
coupling constants
CH3
·
CH3-~
a
CH3
C02
CH 3 -N@X
x
0
1
103.5
105.5
x
kdecarb
rel
NCH 3
s
NCH 3
s
0
78
kH-D
rel
1
103.0
105.4
J(
13 C(2)-H)
220
216
246
J(
13 C(5)-H)
201
202
224
Until now, theoretica! contributions were only focussed
on the thermadynamie acidity. It appeared that polarization
of the a honds is principally responsible for stabilization
of the 1,3-thiazolium ylid. An additional stabilization
effect of sulphur on carbanions has usually been ascribed
to the possibility of (d-p)n 16 - 18 or (d-p)cr 18 backbonding
of the lone pair of the carbanion into the vacant d orbitals
of sulphur. However, ab initio calculations 19 performed on
simpler sulphur or oxygen containing anions (shown in
Figure 5.4), show that stabilization of a carbanion by an
y
H
H-J:)~y
H--f--H
H..__x
H
X = O,S
Y :: -, H,Li
?igure 5.4
adjacent sulphur atom is nat due to (d-p)n bonding, but to
the greater polarizability of sulphur.
In order to gain insight into the thermadynamie and
kinetic acidity of 1,3-azolium cations, d-orbital participation, solvation, bonding and electron densities are considered in more detail. In the H-D exchange reaction the
formation of the conjugate base 14 • 15 is the rate-determining
step, thus for the character of the transition state the
HOHO and penultimate occupied MO of the conj
bases are
studied. For the kinetic acidity the HO's of the cations
are taken into consideration. The systems investigated with
the GEOMO program, using the CND0/2 metbod (Chapter Il)
are shown in Figure 5.5. Although no H-D exchange rates are
available for 1,3-phosphazolium systems, the cation and
79
HWH
-H+
HWH
H
.§.
.f.
x = NH
x=0
x=s
d
x
È.
FiguPe 5.5
.·-
I
6
N~X
H
g_
... H-N~X
\
H
-H•
B-
... H-N~X
H
7
= PH
Systems investigated
related zwitterion are examined for the effect of d-orbital
participation. The geometry of the 1,3-thiazolium, 1,3imidazolium and 1,3-oxazolium system, determined by Sax e:
al. 20 , Rerat 21 and Albano et al. 22 , respectively, is used.
For the 1,3-phosphazolium cation and ylid the geometrical
parameters are taken from phosphole 23 and the 1,3-imidazolium system. All internal parameters in the structures
are optimized.
V.Z
V.2.1
CND0/2 aalaulations on 1.3-azolium systems
d-Orbital aonjugation
The enhanced acidity of protons adjacent to sulphur
in its various oxidation states has been known for nearly
a century 2 ' . Most werkers, with some exceptions 25 , have
preferred an explanation for this phenomenon which implies
a lowering of the energy of the transition state for proton
abstraction (in case of kinetic acidity) or of a carbanion
formation (in case of thermodynamic acidity) by (d-p)n
80
bonding, which is possible for atoms adjacent to second
ro~ atoms. In recent years, the validity of the d-orbital
model has been justified most frequently by reference to
some experimental observations by Doering and Hoffmann 16
and by Oae et al. 17 •
Computations with and without ioclusion of 3d orbitals
on the sulphur and phosphorus atom have been performed on
compounds
and ~. 6c and ~. and ]_s;_ and ~. in order to
test the role of d orbitals in carbanion stabilization by
sulphur as compared to nitrogen and oxygen. The results
in Table V.2 show that dorbitals have no effect on the
proton affinities of 6 and ]_, or conversely, on the C-H
acidity of the cations 5. The introduetion of d orbitals
just renders the basis set more flexible but lewers the
enthalpy of the cation and conjugate base by the same
amount.
Optimized bond lengtbs of the 1,3-thiazolium and
1,3-phosphazolium cations and conjugate bases are shown in
Table V.3. It is note worthy that the samebond lengths are
obtained with the two basis sets. So this geometrical parameter is probably not sensitive to the presence or absence
of d-type functions. The fact that the honds in the conjugate bases are langer than in the cations suggests the
absence of (d-p)n conjugative effects, because such effects
are expected to be manifested by a decrease in bond
length 2 6 •
The contribution of d-type functions to the MO's of
6c and ~. and 7c and d has been assessed by consideration
of the coefficient matrix and charge distribution in these
conjugate bases. In the coordinate system, shown in Figure
5.6, one (d-p)n and one (d-p)o interaction is possible on
symmetry grounds, viz. (C2p y -X3d yz )n and (C
X3d xz ) o.
For the presence of (d-p)n and (d-p)o conjugation it is
necessary that the coefficients of C2p , X3dyz and C2px,
1
X3dxz' respectively, are non-zero. Table V.4 lists the
coefficients of the C2p and X3d functions in the two
highest occupied MO's. The HOMO of the 1,3-thiazolium
81
Table V.2
compound
-Sa
-6a
-7a
-7b
-Sc
-6c
-7c
-Sd
-5d
-6d
6d
-7d
-7d
Calculated enthalpy differences between ylids
and cations
basis
set
total enthalpy
(a. u.)
sp
sp
sp
sp
sp
sp
spd
sp
spd
sp
spd
sp
spd
sp
spd
sp
spd
sp
-48.11892
-47.48897
-47.43444
-S4.0S414
-S3.48088
-S3.4211S
-46.52022
-46.24668
-4S.90095
-45.62671
-45.86165
-45.57796
-43.27787
-42.88092
-42.63474
-42.23641
-42.59755
-42.20079
a,b
r
(a. u.)
óH
.t;H a,a
r
(a. u.)
0.6299S
0.67448
O.S7326
0.63299
0.61927
0.61997
0.66857
0.66872
0.64313
0.64451
0.68032
0.68013
al'll-lr
Hylid-Hcation; bFormation of conjugate base on
position 2; ePermation of conjugate base on position 5.
conjugate base is essentially the carbon lone pair orbital
and would have been expected to exhibit the greatest (d-p)o
interaction. The (C2py-S3dyz)n interaction appears in the
penultimate occupied MO. In case of the 1,3-phosphazolium
conjugate base (C2py-P3dyz)n interaction occurs in the
HOHO and (C2px-P3dxz)o interaction in the penultimate
82
Table V.3
of 1,3-thiazolium and
Optimized bond 1
1,3-phosphazolium cations and conjugate bases,
computed bath with and without ioclusion of
d orbitals in the basis set
bond length (ll.)
compound
Sc
-6c
c2-x
cz x
CçX
CçX
spda
spa
spda
sp
1. 680
1. 685
1 . 711
1. 682
1.709
1 . 7 30
1. 713
1. 689
1. 698
1. 713
1.742
1 . 7 48
1 . 7 56
1. 691
1.709
1 • 6 80
1 . 70 4
1. 722
1 • 71 2
-/C
6d
7d
a
1.695
1. 7 21
1.744
1.749
1 . 761
Basis set.
(a)
coordinate
system
y
L:,
(b)
(d- p)Tt
conjugation
CJ x----c
0 Q
a \:J
Figure 5.6
( d-
plcr
conjugation
y
o L,
----x
The nature of (d-p)n and
Jo
ugation
in the aonjugate bases of 1,3-azoZium systems
with X = S, PH and C
C(2),
C(S)
83
Table V.4
Coefficients of C2p and X3d orbitals which are
appropriate for (d-p)rr and (d-p)cr interaction
in the HOMO and penultimate occupied MOa
coefficientsb
compound
MO
6c
--
HOMO
POMO
HOMO
POMO
HOMO
POMO
HOMO
POMO
-6d
-7c
-7d
X3d
0.065
0.041
-0.069
0.081
0.071
-0.091
0.081
-0.079
0
C2pd
(xz)
(yz)
(xz)
(yz)
(yz)
(xz)
(yz)
(xz)
0.532
-0.278
-0.627
-0.337
-0.375
-0.631
-0.420
-0.577
(x)
( )')
(x)
(y)
(y)
(x)
(y)
(x)
aPOMO: penultimate occupied MO; bSymbols in parentheses
refer to type of basis function; 0 X=S,P; dC2p refers to
C(Z) in 6 and to C(S) in 2·
occupied ~10. The data of Table V.4 indicate that the HOMO
has the greatest (d-p)cr interaction in case of 6c and 7c
and the greatest (d-p)rr interaction for 6d and 7d. However,
it is clear that the proper d-orbital coefficients are
substantially smaller than those of the C2p orbitals in
the MO, so that (d-p)rr and (d-p)cr conjugation can hardly
be considered to constitute an essential basis for the
explanation of the properties of the 1,3-thiazolium cation.
The net orbital populations of X3dxy and X3dxz' 0.057 and
0.070 in 6c, 0.048 and 0.059 in 6d, 0.062 and 0.058 in 7.!::.,
and 0.067 and 0.051 in 7d. These values are too small to
permit chemica! significanee to be attached to (d-p)rr and
(d-p)cr conjugation. The total electronic population of the
d orbitals is the same in the cations Sc and Sd as in the
conjugate bases 6c and 7c, and 6d and 7d (0.34 e and 0.41 e,
respectively). Thus the overall aonalusion is: 3d orbitals
on sulphur and phosphorus in 1.3-azolium systems aat as
84
?CZ~rization
funetions rather than as i
t
valenee
Some analyses have stressed the potential importsnee
of sulphur 3d orbitals in explaining the reactivity of the
1,3-thia:::olium cation and TPP 14 ' 18 ' 27 • 28 • For example, the
greater stability of the 1,3-thiazolium transition state
relative to the 1,3-imidazolium transition state, indicated
kinetic exchange rates, has been attributed, at least in
part, to ''(d-p)o overlap stabilization through interaction
of a d orbital at sulphur with the a orbital directed away
:rom the ring at the 2 carbon" 28 • The contention that
(d-p)ê bonding might play a significant role has been taken
as an explanation for the results of exchange rates studies
in 1,3-thiazolium ions and thiazoles 18 • MO calculations on
the potential importsnee of the sulphur 3d orbitals in
acidity have yielded conflicting results. Ab-initia calculations for an a-sulfinyl carbanion led to the conclusion
that there were no d-orbital contributions to the higher
occupied riO' s 2 9 • Streitwieser and Wi Zliams 3 0 concluded from
i iti
computations for the thiomethyl anion and its
ugate acid that sulphur 3d orbitals stabilized the acid
and the base to the same degree, and that sulphur stabilized
arbanions by polarization rather than by d-orbital conjugation. A similar conclusion was drawn from ARCANA calculations on the structures of thione esters 31 • On the other
hand Extended Hückel studies of thiamine and TPP by Jordan 32
indicated a large 3d-orbital participation. The ExtendedHückel methad is, however, known to overestimate net
(Chapter II). The conclusions effered in this
Section are therefore in accord with the conclusions of
b-initio calculations performed on simpler, sulphur-contain
anions 29 • 30 • In the following Sections only the
computations with d-type functions employed on sulphur and
phosphorus will be considered.
V.2.2
Bonding and electron densities
Table V.S and V.6 show the electron densities and the
85
Table V.S
At om
x,
c2
N3
c4
es
Hz
Hs
At om
x,
cz
N3
c4
es
Hz
Hs
a- and TI-electron densitiesa
a
3.527
2.862
3.527
2.883
2.883
-Sa
TI
1. 48
0.92
1. 48
1. 06
1. 06
--
0.897
0.906
a
4.039
3.048
3. S41
2.893
3.014
0.892
0.870
ain electron units.
TI
1 . 5 78
0.65
1.578
1. 096
1.096
a
3.460
3.492
3.460
2.889
2.889
1.756
0.841
1. 419
0.9S3
1. 0 32
a
4. 19S
3.543
3.479
2.8S4
2.977
a
3.498
2.84S
3.S31
2.802
3.S12
0.991
TI
1.467
1. 024
1 • 49 2
1 • 201
0.815
-6c
-1. 04 7
1f
1. 907
0.572
1 • s 18
o. 972
1. 033
(J
'IT
4.316
3.02S
3.S30
2.809
3.436
0.984
1. 737
0.900
1. 444
1.068
0.849
--
7b
-6b
--
0.997
'IT
-7a
a
4. 411
2.84S
3.S51
2.906
2.824
0.885
0.900
a
3.327
2.949
3.4S4
2.897
2.979
0.900
0.871
1f
1. 615
0.826
1 . 461
1. 042
1. 0 ss
a
4.433
3.4S4
3.477
2.919
2.829
1f
1. 67 4
O.S86
1. 578
1. 068
1.094
--
'IT
1. 482
1. 000
1.528
0.902
1. 08S
a
3.429
3.343
3.412
2.862
3.07S
-0.9S7
1f
1. 5 71
0.971
1. 430
1. 239
0.790
--
1. 002
-5d
a
4.464
2.803
3.569
2.0899
3.4S6
0.997
-6d
TI
1.6SS
0.764
1. ss 1
0.9S7
1. 073
a
3.S25
2.951
3.410
2.832
3.369
0.993
--
-7d
1f
1. 467
1. 033
1. S87
0.986
0.926
Table V.6
Bond
c 2-X
C2-N3
N3-c4
C4-C5
CçX
Bond
c -X
2
C2 N3
N3-c4
C4-C5
c x
Mulliken overlap populationa
-Sa
-7a
-Sb
(J
rr
û
7f
û
;r
0. 720
0. 720
0.673
0.798
0.673
0. 144
0. 144
0.091
0. 211
0.091
0.693
0.693
0.670
0.801
0.670
0. 1 39
0. 139
0.077
0.223
0.077
0.712
0.683
0.642
0.845
0.694
0. 136
0.143
0.083
0.232
0. 108
-5c
(J
0
6cb
TI
0
7f
0
û
TI
a
TI
û
7f
0. 110
0. 158
0.078
0.228
0.060
0.591
0.684
0. 6 79
0.819
0.570
0. 104
0. 153
0.072
0.234
0.056
0.613
0. 721
0.631
0.843
0.599
0. 104
0. 160
0.076
0.236
0.074
(J
1T
5db
6db
-
;r
0. 801 0. 196 0.744 0. 1 8 9 0.705 0.200
(0.175)(0.077)(0.182)(0.087)(0.178)(0.081)
0.715 0. 145 0.697 0. 146 0. 714 0. 14 4
0.673 0.090 0.677 0.084 0.652 0.080
0.793 0.203 0.830 0.215 0.861 0. 216
0.776 0. 136 0.708 0. 121 0.738 0. 1 3 2
(0.110) (0.076) (0.102) (0.077) (0.074) (0.074)
-7b
0.629
0.722
0.663
0. 810
0.560
7
-
6b
0
;r
0
1T
0.871 0.282 0.842 0.266 0.875 0.275
(0.193)(0.126)(0.213)(0.115)(0.131)(0.112)
0. 71 5 0. 10 7 0.710 0. 119 0.718 0. 104
0.676 0. 104 0.664 0.095 0.659 0.098
0.799 0. 19 5 0. 818 0.201 0.859 0. 214
0.815 0. 197 0.789 0. 1 7 5 0. 779 0. 18 7
(0.197)(0.080)(0.217){0.075)(0.213)(0.081)
acr,n: Total overlap population in o and n bond, respectively; bln parentheses total (a-d) and
d) overlap.
overlap populations in the 1,3-azolium cations and conjugate
bases. In the cations both the sulphur and phosphorus atom
have positive character, whereas the nitrogen and oxygen
atom are nearly neutral. The electropositive nature of
sulphur has been previously reported 31 • 33 • Upon deprotonation sulphur gains the largest amount of electron density.
None of the products formed has a classica! ylid structure 34
as was recently found by Aldrich et al. 35 in calculations
with the ARCANA semi-empirica! MO method. GaZlo and
Sable 36 • 37 have pointed out that the 13 c chemical shifts
for the 1,3-thiazolium carbon atoms of thiamine do not
correlate with the previously calculated rr-charge densities38•39. The net atomie charges calculated by the
CND0/2 methad provide a better correlation with the experimentally determined 13 c chemical shifts. The 13 c
resonance for C(2) is at lower field than that for C(4),
which is again at lower field than that for C(S). This is
consistent with the net positive charges of C(2) relative
to C(4) and C(S). To the extent that a correlation between
the 13 c chemica! shifts and the net atomie charges is to
be expected 40 , the CND0/2 results seem to be more successful at predicting the relative 13 c chemica! shifts of the
1,3-thiazolium ring atoms than the ARCANA calculations
carried out by Aldrich et al. 35 •
In Table V.7 the calculated net charges of the protonlike atoms at the carbon 2 (H(2)) and the nitrogen atom
Table V.7
Calculated net charges of H(2) and H(3) and
corresponding pK's
net charge
compound
Hz
Sa
-Sb
0. 1034
0.1151
0.1077
-
'
net charge
pKa
17
12
14
0.1951
0.2054
0.2024
aAcidic pK as measured by Haake and Bausher 41
15
as computed by Haake et al.
88
pKb
H3
7.52
1. 03
3.07
;
bBasic pK
(H(3)) are given with the corresponding acidic and basic
pK's. The calculated net charges correlate very well with
the pK values. The correlation coefficient of the H(2)
s vs the acidic pK and the H(3) charges vs the basic
pK is -0.96 and -0.99, respectively.
The TI bonding for the five-membered 1,3-azolium rings
is extensively delocalized. The largest amount of TI locali:ation is in the C(4) C(S) bond, and with the N-CH-X
fragment resembling a separate delocalized " netwerk. As is
expected on symmetry considerations, the data in Table V.6
indicate that rr-bond density remains nearly the same upon
generation of the conjugate base. The loss of electron
density indicates that although the TI-bond density for the
X CH-X and X-C-X fragments does not change much, the ~
bcnds are shifted from C(2) to X and N in ~ and from C(S)
toKards C(4) in l (Table V.6). Upon generation of the conjugate base the o framewerk displays the greatest change.
The o honds (C(2)-X) and (C(2)-N(3)) of the 1,3-azolium
systems show loss of electron charge density upon deprotonation of C(2}. The electron density is transferred
from the (N(3)-C(2)) o bond to C(2), whereas that of the
(C(2}-X) o bond is shifted to X. It is found that the
polarization of the (C(2)-X) o bond is stronger (for X =
PH or S) relative to the (C(2)-N(3)) o bond. The (C(Z)-X)
~ bond is more polarized for X = S or PH than for X
0
or NH, respectively. Upon generation of a negative charge
at C(S), the (C(4)-C(S)) o bond shows an increase of
electron charge density, whereas the (C(S)-X) o bond shows
a decrease when X is a second row atom and an increase when
X is a first row atom. Upon deprotonation, a-electron
density shifts from C(4) to C(S), C(S) to X (X= 0, S, PH)
and from X to C(S) (X
NH). The change in polarity in the
(C(S}-X) a bond is greater for X = PH or S than X = 0
the (C(2)-X) o bond upon deprotonation at C(Z)). The
above results can be ascribed to the much greater polarizability42 of both the sulphur (3.45 ~ 3 ) and phosphorus
(4.42 ~ 3 ) atom with respect to the carbon (1.75 ~ 3 ), oxygen
(0.73 ~ 3 ) and nitrogen (1.04 ~ 3 atom). The totaloverlap
89
population of the {C{2)-X) bond decreases upon generation
of the conjugate base, whereas the (C(S)-X) bond follows
the same line for X = S, PH, but increases for X = NH, 0.
The difference in overlap population between cations and
conjugate bases increases along the series of atoms X =S,
PH, .0, NH. This is in line with the greater polarizability' 3
of the (C-S) bond (1.88 R3 ) with respect to the (C-0) bond
(0.81 R3 ) and the (C-N) bond (0.57 R3 ). The larger decrease
in overlap population of the (C(2)-S) bond indicates that
this bond is more weakened with respect to the others. In
the present systems a decrease in overlap population is
accompanied by a lengthening of the bond (Table V.3 for
X= s, PH; Table V.8 for X= NH, 0). These findings agree
quite well with results of ab-initia calculations on a-oxa
and a-thia carbanions 19 , and methanethiol and ethane 30 •
Table V.8
Optimized bond lengths of 1,3-imidazolium and
1,3-oxazolium cations and conjugate bases
bond length
CR)
c2-X
CçX
-Sa
1. 350
1. 39 2
6a
7a
1. 36 2
1. 393
1.400
1. 3 71
1.372
1 • 38 3
compound
1. 352
1. 322
1. 340
1. 324
-
-6b
-7b
The data in Table V.2 show that the relative thermadynamie acidity~ (i.e. formation of a carbanion) has the
same trend as the kinetic acidity (i.e. rate of H-D exchange
of the cation), but the correlation is rather poor. Thus
the greater polarizability of sulphur as well as the (C-S)
bond which affords stabilization of the conjugate base of
the 1,3-thiazolium cation with respect to the others, does
,,,,,
;'1''tc1y account for the relative small difference in
Jl-D exchange between the 1,3-oxazolium and 1,3-thiazolium
cation.
4
90
V.3
SoZvation enthalpies
As is outlined in Chapter I, when discrepancies exist
betKeen results of calculations in the gas phase and experiments, it is worthwhile to consider the influence of
solvation. The solvation enthalpies of the reactions in
Figure 5.5, calculated according to equation (2.50), are
shmm in Table V. 9. The data clearly show that the net
solvation enthalpy does not account for the differences in
reaction rates.
Table V.9
Net solvation enthalpies of deprotonation of
1,3-azolium cations
deprotonation
of position
x
net solvation enthalpy (kJ/mole)
PH
0
NH
s
c ( 2)
c ( 5)
V.~
-58.0
49.1
-58.4
-48.6
-58.8
-48.8
-59.6
-49.7
The use of an MO desaription for the transition state
and an estimation of the aativation entha
V.J.1
The aharaater of the transition state
The H-D exchange reaction in 1,3-azolium systems is
assumed to occur via an ylid. It is generally known 45 that
in an endethermie reaction the transition state closely
resembles the reaction product. Thus, it is fairly reasonle to assume that in the deprotonation of the 1,3-azoZium
aations the transition state aZosely resembles the con14
15
• Therefore we describe the
character of the transition state by using the HOMO and
penultimate occupied MO of the conjugate base. The cal
culations show that only the conjugate bases on position
2 and 5 of the 1,3-thiazolium cation have a HOMO with predominant a contribution, whereas the HOMO's of the other
ugate bases mainly possess n contribution. The a HOMO
jugate base struature
•
91
C1
n HOMO
HOMO
y
~'
X
Figure 5.7
= NH,O,PH
The HOMO's of the conjugate bases
(Figure 5.7) of the 1,3-thiazolium conjugate base is principally the carbon lone pair orbital on the C(2) and C(5)
atom. The n HOMO of the other conjugate bases is in the
C(2) -zwitterion chiefly located on the C(2) atom and to
a minor extent on X, whereas in the C(S)-zwitterion a
similar situation is found for C(5) and C(4), respectively.
The penultimate occupied MO is in the latter cases a a MO,
i.e. the carbon lone pair orbital. The enthalpy difference
between the TI HOMO and the penultimate a MO is small,
namely 5.9 kJ/mole, 5.0 kJ/mole and 5.3 kJ/mole for 6a, 6b
and 6d, and 4.8 kJ/mole, 4.1 kJ/mole and 4.3 kJ/mole for
l!• 7b and 7d, respectively.
V.4.2
Estimation of the activation enthalpy
As mentioned before, it hae been propoeed that deprotonation of the aations and deuteration of the aonjugate
92
base take place in the plane of the 1,3-azolium systems 14
5.8). This invoZves that only a o MO is representative for the H-D exchange reaetion. All cations have, how-
ever, a HOMO with predominant TI character, whereas the
penultimate occupied MO has o character, which is appropriate
for the description of the H-D exchange on position 2. For
the deprotonation of the C(S) atom a lm.,rer lying o MO is
suitable, which is principally located on the atomie
_J(!:J\
y
H
Pigure 5.8
---
Aetivated complex for proton abstraction of
1,3-azolium cations
orbitals of C(S) and H(S). The enthalpy differences between
the rr HOMO and the appropriate o MO are shown in Table
V.10. Thus extra enthalpy will be necessary in order to use
the appropriate lower lying occupied cr MO. The amount of
activatien enthalpy in the H-D exchange reactions can be
approximated by: the enthalpy difference between the
conjugate base and the cation and an additional term for
the enthalpy difference between the TI HOMO and the
appropriate a MO of the cation. The calculated activatien
enthalpies (Table V.10) are in excellent agreement with
the observed H-D exchange rates and decarboxylation rates 44 •
The correlation coefficient for the relative k exp vs the
relative kcalc for the C(Z) and the C(S) position is 0.96
and 0.97, respectively. Comparison of the activatien enthalpies necessary for the generation of the C(Z) and C(S)
zwitterion shows that they do not fit the relative exchange
93
Table V.10
Calculated activatien enthalpies and relative
rates for the deprotonation of the C(2) and
C(5) atom in 1,3-azolium systems
rel kacalc
rel kdexp
(a. u.)
Hb
a
(a. u.)
1 ,3-imidazolium c 2
1, 3-oxazolium c 2
1,3-thiazolium c 2
1,3-phosphazolium c2
0. 1203
0. 11 39
0.0731
0.1010
0.7503
0.6872
0. 6923
0.7441
<<1
2.0 x 10 5
3
1. 0 x 10
<1
1
10 5. 5
103.S
1,3-imidazolium Cs
1,3-oxazolium CS
1,3-thiazolium c5
1,3-phosphazolium cs
0.2191
0.2316
0.2022
0.2172
0.8936
0.8646
0.8708
0.8975
<1
10 5
2.5 x 10 2
<1
1
10S.4
10 3. 0
t.Ha
aEnthalpy difference between 1r HOMO and
cation; bActivation enthalpy: Ha = t.H +
Table V.2); 0 Relative rates, calculated
enthalpies (T = 33 °C); dRelative rates
Haakeet aZ. 14 • 15 (pH 4-5, T = 33 °C).
-
-
appropriate cr NO in
t.Hr (t.Hr• see
from the activatien
as measured by
rates (1:10 4 ) as observed by Haakeet aZ. 1 q• 15 • The
difference in reaction enthalpy (Table V.2) is already too
large to underline the difference in exchange rate and rate
of decarboxylation. It should be kept in mind that rates of
decarboxylation are compared with the generation of carbanions at position 5. For the reverse of the reaction, i.e.
protonation, the transition state involves breaking the
H-0 bond of the neutral sol vent with generation of a
hydroxide or alkoxide ion. Data 46 for acetone and HCN give
an estimation for the rate constant for protonation by
water, namely kH 2o ~ 10 5 s -1
From 13 c-H coupling constants of 1,3-azolium cations
it was expected that the 1,3-thiazolium cation would have
a much lower exchange rate than the 1,3-oxazolium cation
94
and a similar rate as the 1,3-imidazolium cation. The observed discrepancies were mainly ascribed 1 ~• 15 to d-orbital
conjugation. In this Chapter it is, however, clearly demonstrated that not d-orbital participation but the smaller
amount of enthalpy necessary for the 1,3-thiazolium cation
to employ the appropriate a MO, is responsible for the
relatively small difference in exchange rate (factor 100)
between the 1,3-oxazolium and 1,3-thiazolium cation.
V.4.3
H-D exchange reaations of arenes
In the previous Section a hypothesis is offered for the
estimation of the activatien enthalpy for H-D exchange
reactions in aromatic 1,3-azolium cations. In ordertotest
this hypothesis GEOHO-CND0/2 calculations with complete
geometry optimization, have been performed for the generation of carbanions of various polycyclic aromatic hydroTable V.11
Calculated and experimental relative exchange
rates of ArDa
liH b
r
a
liHMO
d
Ar
benzene
2-biphenyl
3-biphenyl
4-biphenyl
1- na ph thaiene
2-naphthalene
9-phenanthrene
9-anthracene
2-anthracene
1-anthracene
0.958066
0.916066
0.864216
0.897166
0.930206
0.909513
0.913492
0.924754
0.913923
0.864646
0.0003
0.0419
0.0927
0.0601
0.0261
0.0476
0.0421
0.0603
0.0717
0.0912
0.958366
0.957966
0.956916
0.957266
0.956306
0.957116
0.955592
0.955054
0.955623
0.955846
figures and numbering, Figure 5.9;
kre'e
krel
calc
exp
(T=49.9°) (T=49.9°)
1.0
1. 4 7
4.08
2. 91
7.40
3.37
14.80
24.96
14.36
11.56
1.0
1.2
3.7
2.3
6. s
4. 1
1 7. 9
45. of
10.9
(Ar-)-H(ArH);
cllHMO=H~OMO(ArH)-H~ppropriate
MO(ArH); dH!=liHr+liHMO; eRef.
47; fThe most serious error is that for anthracene-9-d which
solubility problems produce a large error; gActivation
enthalpy.
95
:00: ·G-O·
5
:000:
5
Figure 5.9
10
5
4
4
6
6'
s'
3
Numbering of aromatie aompounds
carbons, e.g. benzene, biphenyl, naphthalene, anthracene,
and phenanthrene. The relative rates of deuterium exchange
with lithium cyclohexylamide, are determined by Streitwieser and Lawler~ 1 for various positions of the forementioned arenes. The experimental and calculated results
are listed in Table V.ll. Comparison of ~Hr and the experimental values delivers a poor correlation, the
correlation coefficient is -0.26. Taking into account the
enthalpy difference of the rr HOMO and the appropriate a ~10,
as is proposed in the previous Section, the experimental
data and Ha are in good agreement. The correlation coefficient is 0.969 and the slope of the correlation line is
0.53. The fact that the data, obtained by the equation
Ha = L'IHr + L'IHMO' are in good agreement with the experimental
results points to direct polarization of electrous in the
C-H bond to the carbon atom.
The conclusion can be drawn that, taking into account
the highest occupied MO's, it is possible to find fairly
good relative H-D exchange rates for aromatic compounds
with the CND0/2 method.
96
rences and notes
1. W. Langenbeek and Z. Hutschenreuter, z. Anorg. Allg.
Chem., 188, 1 (1930).
2. K. Wiesner and z. Valenta, Experienta,
, 190 (1956).
3. K.G. Stern and J.L. Melnick, J. Biol. Chem., 31 597
(1939).
4. A. Lapsworth, J. Chem. Soc.,~. 995 (1903); ~. 1209
(1904).
5. T. Ugai, S. Tanaka and S. Dokawa, J. Pharm. Soc. Jpn.,
63, 269 (1943).
6. T. Ugai, S. Dokawa and S. Tsubokawa, J. Pharm. Soc.
Jpn., &..±, 3 (1944}.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
1 7.
18.
19.
s.
~1izuhara
and P. Handler, J. Am. Chem.
Soc.,~,
571
{1954).
R. Breslow, Chem. Ind. (London). R28 (1956).
J.M. Duclos and P. Haake, Biochemistry,
, 5358 {1974).
D.E. Metzler, "The enzymes", Eds. P.D. Boyer, H. Lardy
and K. Myrbäck, 2nd ed., p 295, Academie Press, New
York; 1960.
R. Bres low, Chem. Ind. (London), 893 (1958).
R. Breslow, J. Am. Chem. Soc., !Q., 3719 (1958).
R. Ereslowand E. McNelis, J. Am. Chem. Soc.,~. 3080
(1959).
P. Haake, L.P. Eausher and W.B. Milier, J. Am. Chem.
Soc., 2_l, 111 3 (1969).
P. Haake, L.P. Eausher and J.P. McNeal, J. Am. Chem.
Soc., 2l_, 7045 (1971).
w. van E. Doering and A.K. Hoffman, J. Am. Chem. Soc.,
2..1_, 521 (1955).
s. Oae, w. Tagaki and A. Ohno, Tetrahedron, ~. 417
(1964).
R.A. Olofson and J.M. Landesberg, J. Am. Chem. Soc.,
8 ' 4263, 4265 (1966).
J.M. Lehn and G. Wipf, J. Am. Chem. Soc., 2§_, 7498
(1976).
97
20. M. Sax, P. Pulsinelli and J. Pletcher, J. Am. Chem.
Soc. , ~. 1 5 5 ( 19 7 4) •
21. C. Rerat, "Molecular structures and dimensions", Yol.
Al, N.V. Oosthoek, Utrecht (Netherlands), 1972, p 365.
22. V. Albana, P.L. Bellen, F. Pempa and V. Scatturin,
"Molecular structures and dimensiens", vol. Al, :\.V.
Oosthoek, Utrecht (Netherlands), 1972, p 230.
23. W. von Niessen, L.S. Cederbaum and G.H.F. Diercksen,
J. Am. Chem. Soc., 98, 2066 (1976).
24. (a) D.J. Cram, "Fundamentals of carbanien chemistry",
Academie Press, New York, 1965, pp 71-84; (b) H.A. Bend
in "Organic Chemistry of sulphur compounds", vol. 3,
N. Kharash and C.Y. Meyers, Eds., Pergamon Press, ~ew
York.
25. H.H. Jaffé, J. Phys. Chem., ~. 185 (1954).
26. L.l\1. Tel, S. Welfe and l.G. Czismadia, Int. J. Quanturn
Chem., J.., 475 (1973).
27. R. Coburn, J. Landerbur, 0. Kemp and R. Olefson,
Tetrahedren, ~. 685 (1970).
28. R. Breslow, Ann. N.Y. Acad. Sci., 98, 445 (1962).
29. S. Wolfe, A. Rauk and I.G. Czismadia, J. Am. Chem. Soc.,
89, 5710 (1967).
30. A. Streitwieser Jr. and J.E. Williams Jr., J. Am. Chem.
Soc., 2.7.., 191 (1975).
31. H.S. Aldrich, Int. J. Quanturn Chem., QB 2 (1975).
32. F. Jerdan, J. Am. Chem. Soc.,~. 3623 (1974).
33. M. Gelus, P.M. Vay and G. Berthier, Theor. Chim. Acta
(Berl.), 2_, 182 (1967).
34. A.W. Johnson, "Organic Chemistry Menegraphs Series",
vol. 7, "Ylide Chemistry", Academie Press, New York,
1966, pp 1, 251, 304.
35. H.S. Aldrich, W.L. Alwerth and N.R. Clement, J. Am.
Chem. Soc. , 100, 2362 (1978).
36. A. A. Gallo and H.Z. Sable, J. Biel. Chem., 249. 1382
(1974).
37. A.A. Gallo and H. Z. Sable, J. Biol. Chem. , 251, 2564
(1976).
98
38. B. Pullman and A. Pullman, "Quantum Chemistry", Interscience, New York, N.Y.; 1963, pp 636-656.
39. B. Pullman and C. Spanjaard, Biochem. Biophys. Acta,
46, 576 (1961).
40. A.J. Jones, D.M. Grant, M. Winkley and R.K. Robbins,
J. Am. Chem. Soc.,
, 4071 (1970).
41. P. Haake and L.P. Bausher, J. Phys. Chem., ll, 2213
(1968).
42. J. Thorhallson, C. Fisk and S. Fraga, Theoret. Chim.
Acta (Berl.),
, 388 (1968).
43. H.A. Stuart, "Molekülstruktur", Springer Verlag, Berlin,
1967; pp 423-426.
44. As is shown in reference 14 and 15 the activation
es and Gibbs free energies are almost the same
entha
while the activatien entropy is very small. Therefore,
it is reasonable to campare relative activation enthalpies and experimental k values.
45. G.S. Hammond, J. Am. Chem. Soc.,
334 (1955).
46. M. Eigen, Angew. Chemie, Intern. Ed. Engl., ~. 1
(1964).
47. A. Streitwieser Jr. and R.G. Lawler, J. Am. Chem. Soc.,
.!U_, 5388 (1965).
zz,
99
CHAPTER VI
ThiaJD.ine pyrophosphate- catalyzed
decarbox.y lation of pyruvate anion
VI.1
Introduation
The decarboxylation of pyruvate to acetaldehyde by the
enzyme pyruvate decarboxylase is a reaction which requires
the coenzyme thiamine pyrophosphate 1 (TPP). On the basis of
investigations of nonenzymatic model reactions, Breslow 2
has proposed that the enzymic reactions praeeed via 2-(1carboxy-1-hydroxyethyl)thiamine pyrophosphate (.l!_), which
is formed from TPP and pyruvic acid by reaction of the 1,3thiazolium ring, ionized at C(Z), with the carbonyl group
of pyruvic acid. Decarboxylation of this intermediate
yields 2-(1-hydroxyethyl)thiamine pyrophosphate (~), which
Q.
R=
CH 3
"=è
-<~ J
CH 2-
NH 2
Q
100
R=
CH 3
R1:::
H
;
~=
3-
C H OP 0
2 4 2 6
can loose acetaldehyde under formation of TPP. The proposed
intermediate Za has been isolated from reaction mixtures
Khich contained pyruvic acid and pyruvate decarboxylase
holoenzyme 3 • 4 • Moreover, it has been shown that pyruvate
decarboxylase apoenzyme catalyzes the formation of acetalfrom 2a 3 • 5 • Results similar tothese have also been
obtained for pyruvate dehydrogenase, the enzyme which
cat
the TPP-dependent oxidative decarboxylation of
pyruvic acid to acetyl coenzyme A6 •
Since most enzymic reactions in which TPP is a cofactor are mechanistically similar to the pyruvate decarboxylase reactions 7 , the decarboxylation of the pyruvate
anion by means of 1,3-azolium cationsis chosen as model
reaction.
VI.2
The reaction scheme for the pyruvate decarboxylation
re action
In
6.1 the mechanistic path;vay is depicted for
the entire reaction, which is consistent with the previous
modeland enzymic studies 8 • Kinetic studies 9 of the decarboxylation of 1b clearly show that the zwitterion is
the species which decarboxylates.
The most likely mechanism for the decarboxylation reaction is the one that yields as the initial product the
planar neutral enamine, which can be protonated in a subsequant rapid reaction. The mechanism of decarboxylation
closely resembles the one proposed for the decarboxylation
of 2-methyl 2-(2-pyridyl)butyric acid (1) 10 • The fact that
101
0
t==\
+f
\
-N
11
..
0
•I
11
-N
S
+ CH-C-C-OH
v..::______/'
3
y
0 0
11
11
CH - C-C- OH
3
\S
-o-C-COOH
I
CH
I
NI
3
0
\
11
S
+
)(
HO
Figure 6.1
CH -C-H
3
CH3
Meohanism for pyruvate deaarboxylation
this compound decarboxylates more rapidly in neutral
aqueous salution than in a strong acid or base suggests
that the zwitterion is the intermediate species. The intermediacy of a planar enamine is clearly shown by the finding
that decarboxylation of optically active 1 yielded racemie
X : 0,
102
TTPP
;
X : S,
TTTPP
2-s-butylpyridine 10 • The formation of enamines is also
supported by the synthesis of activated analogues for TPPdependent enzymic reactions, e.g. thiamine thiazolone pyrophosphate (TTPP) and thiamine thiothiazolone pyrophosphate
(TTTPP) 1 1 •
The tautomerization of the enamine to the dipolar
ion is presented as a proton transfer reaction from the
adjacent hydroxyl group to carbon. This may occur directly
or with the participation of another base. The final step
in the mechanism is elimination of the conjugate base from
the aleoholate anion.
In order to understand the mechanistic aspects of the
enzymic decarboxylation of the pyruvate anion in more
detail, GEOMO-CND0/2 calculations, with geometry optimizations have been performed for the reaction path as depicted in Figure 6.2.
The geometry parameters of the 1,3-azolium ring of 610 have been taken from the corresponding cations. Those
for the 2-(1-carboxy-1-hydroxyethyl) group of 6 are obtained
from lithium dihydrogen citrate 12 and for the 2-(1-hydroxyethyl) group of ~ from X-ray crystallographic data of 2(1-hydroxyethyl)3,4-dimethyl-1,3-thiazolium cation 13 • Por
the enamine, carbondioxide and acetaldehyde, known distances and bond angles are used 14 • The structure of the
enamine 2 has purposely been drawn with the hydroxyl group
and nitrogen cis to one another, sirree examination of
Dreiding models demonstrate that in the other geometrical
isomer there is considerable steric interaction between a
substituent of nitrogen and the vinylic methyl group. The
geometrical data for the pyruvate anion are obtained from
X-ray crystallographic data 15 • All internal parameters
have been optimized.
VI.3
The net reaction enthalpies of pyruvate decarboxylations with 1,3-azolium systems
The net reaction enthalpies are listed in Table VI.1.
Crosby et al. 9 have pointed out that the first step in the
103
decarboxylation has to be considered as a direct proton
transfer between the carboxylate anion and C(2) of the 1,3thiazolium cation (I, Figure 6.2). This reaction seems to
be the most exothermic for the 1,3-oxazolium system. In the
secend step (II) the ylid attacks the pyruvic acid rather
than the pyruvate anion, because the greater electronwithdrawing effect of the COOH group makes the keto carbon
atom of the acid more reactive. For the 1,3-azolium compounds with second row atoms as hetero-atom reaction II is
more exothermic.
Since it is assumed 11 that the enamine closely resembles the transition state for decarboxylation of ~. the
net reaction enthalpy of III is a measure for the relative
rates of decarboxylation. It is obvious that decarboxylation of 2-(1-carboxy-1-hydroxyethyl)-1,3-thiazolium (6d)
will be the most rapid one. This is in agreement with our
recently publisbed results 16 • In this study the lewest unoccupied MO (LUMO) of 1,3-azolium cations has been examined
in order to gain insight into the reactivity of the 1,3thiazolium cation as electron acceptor in biochemica! decarboxylation reactions. As is shown in Table VI.2, the
LUMO of the 1,3-thiazolium cation has the most negative
enthalpy value with respect to the other 1,3-azolium cations.
This means that the decarboxylation will occur most rapidly
Table VI.2
Enthalpy value of LUMO of 1,3-azolium cations
x
NH
0
E (a.u.)
-0.1343
-0.1789
PH
s
-0.1065
-0.2213
with TPP as catalyst. It should, however, be noticed that
all these data refer to gas phase conditions.
Starting from the enamine there are two possible ways
to obtain the aleoholate anion (~), from which the ylid
and the aldehyde can be formed, namely (1) tautomerization
of the enamine to the dipolar ion ~ and (2) formation of
2-(1-hydroxyethyl)-1,3-azolium compounds (~) and subsequent
ionization of the hydroxyl group. The data for reaction IV
104
IQ\
H-J8~
- - rl- NvX
'-../
li
5
--
III
fo
Ho-c-c""'
I
'oCH3
6
I
\
H-NI
X
----
0
H-J~~
-o-e
HOXCH3 IJ[
11
y
2.
+ CH -CH
3
H
I
CH 3
8
7
x -
I
\
x
H-NI
+
H+
YI
HO
CH3
l
H-N~
HO-C- H
I
CH 3
~
-H•
0
H-NfX
1TI
-o-c-H
I
CH 3
t=
x
0
.9.
NH
.Q
0
f
PH
CH 3
Q
s
10
Figure 6.2
H-NYX
HO-C-
I
Reaction pathway
105
....
0
en
Table VI.1
Net reaction enthalpiesa
6Hr
6Hr
t>Hr
6Hr
6Hr
6Hr
l:\H
r
t.Hr
a. u.
a.u.
a.u.
a. u.
a. u.
a. u.
a.u.
a.u.
II
III
V
VI
VII
VIII
-0.2322
-0.2293
-0.1980
-0.1652
-0.4833
-0.4478
-0.4520
-0.4372
-0.2509
-0.2294
-0.1800
-0.2540
1.0599
1.0205
1.1094
1 . 0 154
x
I
NH
-0.2166
-0.2673
-0.1974
-0.2213
0
PH
s
-0.2606
-0.2397
-0.4823
-0.4882
0.2171
0.1772
0.3738
0. 1560
IV
-0.2370
-0.2161
-0.1719
-0.2414
aNumbers of reactions are shown in Figure 6.2.
and VII indicate that the latter is more exothermic. For
the 1,3-thiazolium compounds this is in agreement with the
pK a values. The pK a for the ionization of the hydroxyl
group of 2-(1-hydroxyethyl)thiamine pyrophosphate in water
is probably about 12 17 •. The pK 8 for the dissociation of
the a hydrogen atom of this compound has not been determined. A crude estimation 9 gives a value of pKa = 17 for
this reaction. These pKa values show that in water the
proton transfer from the hydroxyl group to the a carbon is
favoured thermodynamically.
The mechanism prediets that the a hydrogen of 2-(1hydroxyethyl)thiamine pyrophosphate ought to exchange with
solvent deuterium through interconversion as is shown in
Figure 6.3. The net reaction enthalpies of VI and VIII
.F\X
~
-N~
HO
020
-------=-
....,---
C~H
I
-NI
x
HO)(CH
I
.. I
\
~
3
CH 3
6.3
H-D exchange of
-N
\
y
X
__,_
-=--
c-
HO /
I
-•N
' CH3
y
\X
HO-C-D
I
CH 3
2-(1-hydroxyethy~Jthiamine
pyrophosphate
show that most probably the 1,3-thiazolium compound will
exchange most readily. In Table VI.3 the net charges on
C(2) of~ are listed. The partial positive charge on C(2)
of~ stahilizes the adjacent a carbanion and this helps
to explain the ease of formation of the a carbanion bath
enzymatically 7 and in model studies 18 • These results stand
in contrast to the results of X-ray crystallographic
studies 13 of~. which place a partial negative charge on
C(2). The discrepancy probably reflects the more indirect
approach used in the X-ray crystallographic studies. The
107
Table VI.3
Net charges on C(2) of
9a
--
net charge
on C(2) (e.u.)
0.201
~
9b
0.231
9c
--
9d
--
0.189
0.249
charge density at C(2) is an important factor in determining kinetic acidity of (Ca)-H. The order of net charges
is the same as the order of net reaction enthalpies.
VI.4
TPP as aocarboxylaae
Haake et al. 17 have shown that the relative rates for
the formation of the ylids of 1,3-azolium cations are in
the order 4b:4d:4a = 10 5 · 5 :10 3 · 5 :1. If the rate of ylid
formation correlates with catalytic effectiveness, 1,3oxazolium compounds should be better catalysts than 1,3thiazolium compounds. There are, however, no data for in
vivo catalysis by 1,3~oxazolium compounds. Breelow 19 has
stuclied the benzoin condensation catalyzed by a number of
1,3-azolium species, including benzoxazolium compounds.
In all cases, the benzoxazolium cations have shown no
catalysis. This does, however, not implicate that 1,3oxazolium compounds can not act as catalysts. Zottewies
and Helmiek 20 have demonstrated that for benzothiazolium
ions, the primary reason for the diminished reactivity of
the 1-hydroxyethyl 0 chain is steric inhibition of resonance.
Interaction between the hydroxyl group of this chain with
the substituant bonded to nitrogen prevents maximum overlap, resulting in an effective delocalization of the
electrans from the reactive C-H bond into the positively
charged ring of the transition state. Codington and Wuerat 21
have synthesized an oxazolium analogue of thiamine. They
have found that this compound does not function as a cocarboxylase. Thus, the 1,3-oxazolium compounds seem to be
ineffective in model and in vivo reactions where thiamine
108
CH
OH
+
3
R1
.Hx
-N
v
CH 3
Hx
-NI
x
H
?igv.re 6.4
R1
OH
I
\x-
-NI
"eH
0
+ H+
~
Ring opening reaction of 1,3-azolium cations
is effective. Duclos and Haake 22 have considered the ringopening reaction of 1,3-azolium cations (Figure 6.4), since
it has been assumed by Breslow 18 that there has to be a pH
optimum, below which the concentratien of the ylid is too
low for observable catalysis. Duclos and Haake have established that at physiological pH (pH 7.4) the 1,3-oxazolium
cations are kinetically and thermodynamically unstable.
This explains why 1,3-oxazolium compounds are inactive in
{~vivo and in vitro: 1,3-oxazolium ions are unstable near
neutrality, they hydrolyse to the ring-opened form. The
1,3-thiazolium and 1 ,3-imidazolium compounds are thermodynamically stable at physiological pH, but as the pH increases, a significant amount of ring-opened form of 1,3thiazolium is present and above pH 9.5 this form predominates. The 1,3-imidazolium ions require very strongly basic
solutions for ring opening. They form, however, ylids more
than 10 3 times more slowly than the corresponding 1,3-thiazolium ions (vide supra). Therefore, of the 1,3-azolium
compounds, the 1,3-thiazolium ion is suited for cocarboxylase function. It is thermodynamically stable at pH 7.4,
and the rate of formation of the 1,3-thiazolium ylid
enables it to be an effective catalyst.
109
Referenoes
1. A.V. Morey and E. Juni, J. Biol. Chem., 243, 3009
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
110
(1968).
R. Breslow, Chem. Ind. (Londen), 893 (1957).
L.O. Krampitz, I. Suzuki and G. Gruell, Fed. Proc., 20,
971 (1961).
H. Holzer and K. Beaucamp, Biochim. Biophys. Acta, 46,
225 (1961).
H.W. Goedde, B. Ultich, C. Stahlmann and H. Holzer,
Biochem. Z., 343, 204 ( 1965).
P. Scriba and H. Holzer, Biochem. z., l• 473 (1961);
H.W. Goedde, H. Inouye and H. Holzer, Biochim. Biophys.
Acta, 50, 41 (1961); L.P. Hager and L.O. Krampitz,
Fed. Proc.,~. 536 (1963).
L.O. Krampitz, Ann. Rev. Biochim., ~. 213 (1969).
H.R. Mahler and E.H. Cordes, "Biologica! Chemistry",
Harper and Row, New York, 1971, sec. edn.; pp 401-406.
J. Crosby, R. Stone and G.E. Lienhard, J. Am. Chem.
Soc., 92, 2891 (1970).
W. van E. Doering and V.Z. Pasternak, J. Am. Chem. Soc.,
.zl, 143 (1950).
J.A. Gutowski and G.E. Lienhard, J. Biol. Chem., 251,
2863 (1976).
"f.1olecular structures and dimensions", vol. Al, N.V.
Oosthoek (Netherlands), 1972.
fv!. Sax, P. Pulsinelli and J. Pletcher, J. Am. Chem.
Soc., 96, 155 (1974).
"Tables of interatomie distances and configuration in
molecules and ions", A.D. Mitchell and L.C. Cross Eds.,
Special Publication no 11, The Chemica! Society London,
1958.
S.S. Tavale, L.M. Paul and A.B. Biswas, Acta Cryst.,
11· 1281 (1961).
M.M.E. Scheffers-Sap and H.M. Buck, J. Am. Chem. Soc.,
.!Q..!, 4807 (1979).
P. Haake, L. P. Bausher and J. P. ~icNeal, J. Am. Chem.
So c . , 9 3 , 7 0 4 5 ( 1 9 7 1 ) •
18. J.J.
~Iieyal,
R.G. Votaw, L.O.
Biochim. Biophys. Acta,
lil•
Krampitz and H.Z. Sable,
205 (1967).
19. R. Breslow, J. Am. Chem. Soc., 80, 3719 (1958).
20. J.A. Zoltewicz and L.S. Helmick, J. Org. Chem., 43,
1718 (1978).
21. J.F. Codington and H.H. Wuerst, Fed. Proc. Fed. Amer.
S o c . Exp • Bi o 1. , 1§_, 1 6 5 ( 19 5 7) .
22.
J.~1.
Duclos and P. Haake, Biochemistry,
.Jl.,
5358 (1974).
111
APPENDIX A
CND0/2 optimized geometry of c-AMP
R
at om
N
A
B
c
0 ( 1' )
1
2
3
4
5
6
7
8
0
1
2
3
4
5
6
4
7
0
0
0
0
1
0
2
3
1
5
5
6
7
7
2
3
3
2
3
4
5
5
3
1
2
2
c ( 1')
C(2')
C(3')
C(4')
c (5')
0(5')
0(3')
p
0(6)
0(7)
0(2')
ll(1~)
H( 1t)
H(2')
H(3')
H(4')
H(S')
a
H(5b)
H(0(2'))
9
10
11
12
13
14
15
16
17
18
19
20
9
9
3
2
2
3
4
5
6
6
12
1
1
5
6
6
1
4
4
1
2
3
1
1
2
R
o.ooo
1. 3 73
1. 502
1. 498
1. 522
1. 494
1.373
1.396
1. 610
1. 664
1. 6 71
1. 390
1.104
1.106
1 • 115
1 . 130
1. 125
1.099
1 • 10 1
0.999
ALP HA
degrees
o.o
0.0
105.9
106.6
95.6
105.5
110. 8
115.3
109.0
108.6
107.3
111 • 9
110. 2
117. 2
110. 2
110. 5
117. 4
114.6
114. 7
108. 1
BETA
degrees
0.0
0.0
0.0
11.3
18.3
152.9
171.5
-160.4
53.3
-144.6
85.7
130.4
-130.2
105.3
-110.7
- 86.9
81.8
- 68.7
51.8
105.7
N: number of atom; A: number of atom by which the bond
length R for atom N is defined; B: number of atom by which,
together with atom A, the bond angle ALPHA ( L NAB) is defined; C: number of atom by which, tagether with atom A and
B, the dihedral angle BETA (LL NABC) is defined.
112
c;.;D0/2 optimized geometry of 5'-AMP
R
ALP HA
BETA
c
~
degrees
degrees
a torn
~
A
0 ( 1' )
c ( 1 t)
1
0
0
0
0.000
0.0
0.0
2
1
0
0
1. 390
0.0
o.o
C(2 t)
3
2
1
0
1. 496
107.1
0.0
C(3')
4
3
2
1
1. 495
102.3
- 30. 1
C(.f')
5
4
3
2
1. 383
99.9
42.2
c ( 5')
6
5
4
3
1. 379
119. 5
-133.9
-
B
0(5')
7
6
5
4
1 . 4 71
108. 3
p
8
7
6
5
1 • 794
115. 2
1 77. 2
0(2')
9
3
2
1
1 . 432
111. 3
- 88.7
·o c3')
10
4
3
2
1. 390
108.3
170.0
0(6)
11
8
7
6
1. 6 7 3
106.3
17 7. 1
0 ( 7)
12
8
7
6
1. 6 71
108.9
50.3
0(8)
13
8
7
6
1.576
106.0
- 69.2
H( 1 ~)
14
2
3
4
1. 130
1 10 • 1
-121.9
H ( 1 b)
H(0(2'))
1s
2
3
4
1.137
110. 4
12 4. 1
16
9
3
2
1 . 0 55
106.8
- 58.4
H(0(3'))
17
10
4
3
1. 051
108.8
56.6
H(2')
18
3
2
1
1.124
11 0. 9
99.4
H(3')
19
4
3
2
1. 111
11 7. 1
- 37.5
79.9
78.1
H(4')
20
5
4
3
1.133
118. 3
H(5~)
21
6
5
4
1 . 128
11 3. 8
78.8
H(5b)
H(0(8))
22
6
5
4
1. 12 4
113. 8
- 78.8
23
13
8
7
0.929
119. 1
98.8
113
CND0/2 optimized geometry of 3'-AMP
at om
N
A
B
c
~
ALP HA
degrees
0(1')
1
0
0
C ( 1 I)
2
1
0
C(2')
C(3')
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2
1
0
0
0
3
4
5
6
4
8
9
9
9
3
2
2
13
4
5
6
6
7
12
3
2
1
3
4
5
5
4
8
8
8
2
2
0.000
1 . 391
1. 493
1. 497
1. 502
1.480
1. 390
1. 394
1.770
1. 663
1. 663
1 • 7 50
1. 390
0.991
1 . 100
0.878
0.930
0.948
0.952
1.012
1. 0 30
1. 032
1. 081
0.0
0.0
107.6
102.8
99.5
114.7
105.5
109.6
94.9
113. 5
106.2
101.9
110.8
111 • 6
106.9
105.5
110.3
112.6
116.6
107.4
103.3
109.8
111. 5
R
C(4')
C(5')
0(5')
0(3')
p
0(6)
0( 7)
0(8)
0 (2 t)
H( 1~)
H( 11,)
H(0(2'))
H(3')
H(4')
H(S~)
H(Si,)
H(0(5'))
H(0(8))
H(2')
114
3
1
1
5
4
4
4
1
1
3
3
4
5
5
5
5
2
2
3
4
4
6
5
9
2
8
1
1
BETA
degrees
o.o
o.o
o.o
- 21.6
36. 2
-156.5
56.8
-158.5
-123.0
51.3
180.7
- 65.4
97.4
117. 0
-127.6
-148.5
- 79. 1
77.7
50.4
- 68.8
148.9
- 78.0
-143.9
115
APPENDIX
B
Hydration scheme of c-MlP
numbera and
characterb of
water molecule
1
A
2
A
3
A
4
A
s
A
6
7
8
A
9
A
D
A
water molecule
net charge" on
He
0
H"
0.2420
0.2415
0.2424
0.2253
0. 2107
0.1674
0.2195
0. 1962
0.2004
-0.3855
-0.4036
1L0610.0756
0.0626
0.0"39
0.0886
-0.3896
-0.3680
-0.3473
-0.2906
-0.3639
0.14~6
0.08H
-0.3372
0.0918
-0.3395
0.0965
asee drawing; bA: proton acceptor, D: proton donor; eH involved in hydragen
bond; dH nat involved in hydragen bond, except for formation of seven-membered
ring (12); "in e.u.
116
c-A~!P
continued
atoms and net chargese on
atom and net chargese
on atom involved
in hydrogen bond
numbera
of water
molecule
the atorns a dj a een t to
proton acceptor or donor
1
0(6)
-0.4189
p
0.4308
2
0(6)/0(7)
-0.4602/-0.5116
p
0.4363
3
0(7)
-0.5299
p
0.4325
-1
0 (3')
-0.2883
C (3' )/P
0.2112/0.4012
-0.2499/-0.2386
C(2')/C(3')
5
0(2')/0(3')
6
H(0(2'))
0.2214
0(2')
0.1483/0.2018
-0.3516
7
0(2')
-0.2636
C(2' )/H(0(2'))
0.1475/0.1421
8
0(1
'l
-0.2455
C ( 1 ')/C ( 4')
0.1391/0.1017
9
0(5')
-0.2833
C(5' )/P
0.2156/0.3864
117
flydra ti on scheme of S '- M1P
..............
"
o,. o..
LJ
water molecule
net charge (e. u.) on
numbera and
characterb of
water molecule
1
2
3
4
5
A
A
A
A
A
Hd
0.2435
-0.3892
-0.3951
-0.3933
-0.3604
-0.3258
-0.2756
-0.3069
-0.3148
-0.3211
-0.2644
-0.3366
-0.3501
-0.3269
-0.2702
o. 0611
A
D
o. 1 s 19
D
7
A
8
A
A
10
11
12
13
14
0
0.1588
0. 2440
0. 2411
0.1594
0.1548
0.1972
0.1865
0.2215
0. 1102
0.2218
0. 1569
0. 2311
6
9
H"
D
A
A
Notes: see hydratien scheme of c-AMP
118
0.1546
0.0606
0.0756
0.1500
0. 1508
0.0923
0.0959
0.0634
0. 1121
0.0789
0. 1701
0.0814
0.1524
S '-A~lP, continued
numbera
of water
molecule
0(7)
-0.441
-0.4401/
-0.4612
-0.4599
0(3.)
-0.2412
0(2' )/
0(3.)
0(6)
0(6)/
0(7)
4
5
atoms and net chargese on
the atoms adjacent to
proton acceptor or donor
atom and net
on atom involved
in hydrogen bond
H(O(Z'))
p
0.4269
p
0.4269
p
0.4364
C(3')/!1(0(3')
0.0752/0.1302
-0.2727/
C(Z' )/
0.1508/
-0.2391
c (3.)
0.0645
0. 146 7
0(2')
-0.3144
0(2.)
-0.2813
C(Z' )/fl(O(Z'))
0.1469/0.1099
8
0(1 ')
-0.2509
C(l' )/C(4')
0.1278/0.2130
9
0(5')
-0.3501
C(S' )/P
0.1621/0.4064
10
H(0(8))
11
0(8)
-0.4936
11(0(8) )/P
0.1732/0.4113
0(1. )/
-0.2487/
C(1')/C(5')/
0.1280/0.1366/
0(
13
1~
')
0.2347
0(8)
-0.4652
-0 . .>601
C(4')/P
0.2131/0.3962
0(1' )/
-0.2412/
C(1')/C(5'l/
0.1271/0.1298
0 5')
-0.3556
C(4')/P
H(0(3'))
0. 1680
0(3')
0.2132/0.3944
-0.2818
119
llydration scheme of :i' -NIP
water molecule
net chargee on
numbera and
'characterb of
water molecule
1
A
z
A
3
A
4
A
5
A
6
D
7
A
8
A
9
A
10
D
11
12
13
14
A
A
A
D
H"
0
Hi
0.2412
0.1588
0.2424
0.2080
0.1682
0. 1397
0.1554
0.1991
0.1944
0. 1248
0.2054
0.1476
0.1831
0.1992
-0.4149
-0.3475
-0.4171
-0.3536
-0.2968
-0.2840
-0.2747
-0.3558
-0.3539
-0. 2911
-0.3339
-0.3281
-0.3119
-0.2805
0.0655
0.1542
0.0645
0.0905
0.1123
0.1303
0.1027
0.1069
0. 1071
0. 1284
0.0926
0.1590
0.0963
0.1068
Notes: see hydration scheme of c-MIP
120
3'-AMP, continued
numbera
i of
\\ater
molecule
1
2
atoms and net charges 8 on
the atoms adjacent to
proton acceptor ar donor
atom and net
on atom involved
in hydragen bond
-0.4576
p
0.4342
0(6)/
-0.4536/
p
0. 4151
0(7)
-0.4432
0(6)
3
() ( 7)
-0.4427
p
0.4349
4
0(3')
-0.3378
C(3')/P
0.1453/0.4424
0(2' )/
-0.3071/
-0.3357
c (2' )/
c (3')
0.1455/
0(3')
5
6
H[O(Z'))
0.2053
0(2')
o. 1426
-0.3400
7
0(2')
-0.3091
C(Z')/H(O(Z'))
0.1398/0.1818
8
0(1')
-0.2717
C(1')/C(4')
0.1429/0.1390
9
0(5')
-0.2596
C(S')/H(O(S'))
0(8)
0.1423/0.1142
-0.3762
10
H(0(8))
11
0(8)
-0.3599
P/1!(0(8))
0.4199/0.1093
12
0( 1 ')/
-0.2667/
C(1')/C(4')/
0.1430/0.1378/
0(5')
-0.2696
C(5')/P
0.1426/0.4081
0(1 ')/
-0.2731/
C(1 ')/C(4' ]/
0.1415/0.1301/
0 (5')
-0.2599
C(S')/P
13
14
H(O(S'))
0.1631
.
0. 1411
0(5')
0.1498/0.3987
-0.2987
121
Sun1D1ary
In this thesis a quantumchemical study of the molecular dynamics of the coenzymes cyclic adenosine 3' ,5'monophosphate (c-M·1P) and thiamine pyrophosphate (TPP) is
presented. Molecular orbital calculations have been performed using the Extended-Hückel and the CND0/2 method,
and the ST0-3G ab-initia procedure.
c-Ar.•P is a key substance in the regulation of enzymatic processes in the cell and it stimulates the
activity of the genes via the synthesis of messenger ~~A.
which in fact reproduces the information stored in the
DNA of the gen. The level of concentratien of c-AMP in
the cell is regulated via hydralysis with a phosphodiesterase, which yields adenosine 5'-monophosphate (5'-AMP).
The hydro
is to 5 '-M•P in vol ves a large exothermic
enthalpy (-46.6 kJ/mole). In particular, attention has
been focussed on the mechanistic aspects of this hydralysis. The solvent effect on the enthalpy of hydralysis
has been studied by the EH method for the hydrolysis of
c-AMP and related cyclic phosphate diesters. The results
show that the difference in enthalpy between c-AMP and
related phosphate diesters can be explained, in part, by
the difference in net solvation enthalpy. It has been
calculated that the large exothermic net solvation enthalpy
of the hydralysis of c-AMP can be ascribed to an extra
stabilization of the hydralysis product with respect to
the reactants and which is absent in the other cyclic
phosphate diesters. The extra stabilization is due to
122
hydragen bonding with water between 0(1') and 0(5') in
5'-AMP. From CND0/2 calculations, with a water molecule
situated between 0(1') and 0(5'), it is concluded that this
position indeed is an important hydratien site and that
formation of a five- as well as a seven-membered ring is
possible. The occurrence of the former opens the potentiality fora "through water" interaction of 5'-AMP with
enzymic sites. In addition to the solvent effect the conformational difference of the ribose ring in c-AMP, on the
one hand, and in 5'-A~1P, on theether hand, bas been taken
into account. It is clearly demonstrated that this difference also contributes to the overall exothermic enthalpy
of hydrolysis.
Furthermore, attention bas been paid to the coenzymatic
function of TPP. It serves as a coenzyme for enzymatic reactions: oxidative and nonoxidative decarboxylations of
a-keto acids and formation of a ketols. The mechanism of
the nonenzymatic reactions bas been observed with the
discovery of Ereslow that the aromatic hydrogen at position
2 of the 1 ,3-thiazolium
exchanges readily with solvent
deuterium. The H-D exchange reactions of 1,3-azolium cations
have been stuclied by the CND0/2 metbod with optimization
of all geometrical parameters, in order to explain the rate
enhancement for the 1,3-thiazolium cations. The following
results are obtained: (a) stabilization of a carbanion by
the adjacent sulphur atom is not due to (d-p) conjugation;
(b) the 1 ,3-thiazolium conjugate base is stabilized with
respect to the other conjugate bases by the greater polarizability of sulphur; (c) the smaller amount of energy
necessary for the 1,3-thiazolium cation, with respect to
the ether cations, to use the penultimate cr MO, gives an
explanation for the unique rate enhancement. Besides, the
decarboxylation of the pyruvate anion by means of 1,3azolium cations bas been examined with the CND0/2 method,
since most enzymic reactions in which TPP is a cofactor
are mechanistically similar to this reaction. The data
suggest that the decarboxylation with the 1,3-thiazolium
will be the most rapid one.
123
SaDtenva't'ting
In dit proefschrift wordt een quanturnchemisch onderzoek beschreven naar de moleculair-mechanistische aspecten
van de coenzymen cyclisch adenosine 3',5'-monofosfaat
(c-AMP) en thiamine pyrofosfaat (TPP). Molecular orbital
berekeningen zijn uitgevoerd met Extended-HOckei (EH),
CND0/2 en ST0-3G ab-initio procedures.
c-AMP speelt een belangrijke rol in de regulatie van
enzymatische reacties in de cel en het stimuleert de activiteit van de genen door de synthese van messenger RNA, dat
de informatie die opgeslagen ligt in het DNA van de genen
reproduceert. De concentratie aan c-AMP in de cel wordt
geregeld via een hydrolyse-evenwicht tussen c-AMP en adenosine 5'-monofosfaat (5'-AMP). De vorming van 5'-At-lP
blijkt sterk exotherm te zijn, de hydrolyse-enthalpie is
-46.4 kJ/mol. Speciale aandacht is uitgegaan naar de mechanistische aspecten van deze hydrolyse. De invloed van
het oplosmiddel op de hydrolyse-enthalpie is bestudeerd
met de EH methode voor de hydrolyse van c-AMP en verwante
cyclische fosfaat diesters. De resultaten tonen aan, dat
het enthalpieverschil tussen c-AMP en de fosfaat diesters
gedeeltelijk verklaard kan worden door het verschil in
netto solvatatie-enthalpie. ~e exotherme netto solvatatieenthalpie voor de hydrolyse van c-AMP kan worden toegeschreven aan een extra stabilisatie van het hydrolyseproduct ten opzichte van de uitgangsproducten. Deze
stabilisatie wordt bepaald door een regio-specifieke
hydratatie van de brugpositie tussen 0(1') en 0(5') in
124
5'-ANP. CND0/2 berekeningen, met een watermolecule geplaatst
tussen 0(1') en 0(5') in 5'-AMP ondersteunen de veronderstelling, dat deze positie een belangr ke hydratatie site
is. Het blijkt, dat zowel een vijf- als een zevenring gevormd kan worden. Het voorkomen van een
fring opent de
mogelijkheid tot een interactie van 5'-AMP met enzymen.
Verder is het verschil in conformatie van de ribose ring in
c-A~1P en 5' -AMP bestudeerd. Er is duidel
k aangetoond, dat
dit verschil ook bijdraagt tot de exotherme hydrolyseenthalpie van c-AMP.
Tevens is aandacht geschonken aan het werkingsmechanisme
van TPP, dat optreedt als coenzym voor reacties in het
carbohydraat metabolisme, waarin aldehyde-groepen worden
geactiveerd. Moleculair-mechanistisch gezien speelt hierbij
de C(2) positie van de 1,3-thiazolium ring een grote rol,
wordt ondersteund door H-D uitwisselingsexperimenten
Zow). De onverwacht grote H-D uitwisselingssnelheid van het 1,3-thiazolium kation ten opzichte van andere
1,3-azolium kationen is bestudeerd met behulp van CND0/2
berekeningen. De volgende resultaten zijn verkregen: (a)
stabilisatie van het carbanion door een naburig zwavelatoom kan niet worden toegeschreven aan (d-p) conjugatie;
(b) de geconjugeerde base van het 1,3-thiazolium kation
wordt ten opzichte van de andere geconjugeerde basen gestabiliseerd door de grotere polariseerbaarbeid van het
zwavelatoom en de zwavel-koolstof binding; (c) b
de H-D
uitwisselingsreactie kan slechts gebruik gemaakt worden
van een er ~10. Bij de 1, 3-azolium kationen is dit de op één
na hoogst bezette MO. De geringe hoeveelheid energie, die
het 1,3-thiazolium kation nodig heeft om gebruik te maken
van deze MO met betrekking tot de andere kationen, is verantwoordel k voor de unieke snelheid van de H-D uitwisseling in het 1,3-thiazolium kation. Bovendien is de decarboxylering van het pyruvaat anion met behulp van 1,3-azolium
kationen bestudeerd. Het blijkt dat de laagst onbezette MO
van het kation bepalend is voor de afsplitsing van kooldioxide. Het unieke karakter van het 1,3-thiazolium kation
wordt hiermee bevestigd.
125
Curriculum vitae
De schrijfster van dit proefschrift is geboren op
12 februari 1952 te Tilburg. Na het behalen van het diploma
gymnasium B aan het St. Pauluslyceum te Tilburg begon zij
in 1970 met de scheikunde-studie aan de Technische Hogeschool te Eindhoven. Het afstudeerwerk verrichtte zij bij
de vakgroep Organische Chemie van de afdeling Scheikundige
Technologie met als afstudeerdocent p:of. dr. H.M. Buck.
Op 29 oktober 1975 behaalde zij het ingenieursdiploma.
Vanaf 1 november 1975 tot 1 augustus 1979 is zij
verbonden geweest aan de Technische Hogeschool te Eindhoven
als wetenschappelijk ambtenaar. In deze periode werd onder
leiding van prof. dr. H.M. Buck het onderzoek, dat beschreven is in dit proefschrift, uitgevoerd.
Sinds 1 augustus 1979 is zij als lerares scheikunde
en natuurkunde werkzaam op het St. Pauluslyceum te Tilburg.
126
Dankwoord
Aan de totstandkoming van dit proefschrift hebben
velen een bijdrage geleverd. In het bijzonder op het gebied
van quanturnchemische berekeningen en de interpretatie van
zoKel literatuurgegevens als resultaten. Voor al deze hulp
en voor de prettige sfeer waarin ik heb kunnen werken wil
ik een ieder bedanken.
Verder ben ik diegenen zeer erkentel k, die een bijdrage hebben geleverd aan de uiteindelijke vormgeving van
dit proefschrift.
127
Stellingen
1. Bij de verklaring voor het niet uitwisselen van het 2'0H waterstofatoom van cyclisch adenosine 3' ,5'-monofosfaat, wordt door Bolton en Kearns ten onrechte geen
rekening gehouden met waterstofbrugvorming tussen 0(2')
en 0(3').
P.H. Bolton en D.R. Kearns, J. Am.
Chem. Soc., .lQJ_, 479 (1979).
2. Daar de nucleofiele aanval op een fosfaat-gesubstitueerd
fosfoniumion voornamelijk plaatsvindt op de fosfaatgroep, kan de alkyloverdracht in de door Ramirez et aZ.
beschreven modelstof beter verklaard worden door een
nucleofiele aanval op het vijfgecoördineerde intermediair.
F. Ramirez, Y.F. Chaw, J.F. Mareeek
en J. Ugi, J. Am. Chem. Soc., 96,
2429 (1974).
3. De verwachting dat er optische activiteit aantoonbaar
is na cyclodehydratatie van racemisch 4-hydroxycyclohexanon met behulp van een optisch actief carbodiimide
berust op een verkeerde interpretatie van de regels van
asymmetrische inductie.
R.R. Hiatt, M.J. Shaio en F. George,
J. Org. Chem., _ii, 3265 (1979).
4. Bij de in de literatuur veel voorkomende bewering dat
fluorescentie werd gemeten bij "druk nul"-omstandigheden wordt vrijwel nooit rekening gehouden met longrange energy transfer volgens het Förster-Dexter
mechanisme.
R.G. Milier en E.K.C. Lee, Chem. Phys.
Letters, 41, SZ (1976); K. Uchida,
I. YamazaKI en H. Baba, Chem. Phys.
Letters,~. 133 (1976).
S. Bij de bepaling van de bezettingsgraad van alkylchloorsilanen die aan een silica-oppervlak gebonden zijn,
worden koolstofatomen die geen deel uitmaken van de
lineaire alkylketen ten onrechte verwaarloosd.
H. Hemetsberger, ~1. Kellermann en
H. Rieken, Chromatographia, 10, 726
(1977).
--6. De beschrijving van electreneninteracties in spirageconjugeerde modelsystemen met behulp van correlaties
tussen berekende ladingsdichtheden en gemeten 13 c
chemical shifts kan gemakkelijk tot onjuiste conclusies
leiden.
S.Q.A. Rizvi, J. Foos, F. Steel en
G. Fraenkel, J. Am. Chem. Soc., 101,
4488 (1979); H. Dürr, K.H. Alberr-en
M. Karisch, Tetrahedron, 35, 1285
(1979).
-7. In artikelen over quanturnchemische onderwerpen dient de
algemene term energie gespecificeerd te worden naar de
gebruikte rekenmethode.
8. Op het identificatieplaatje van de nederlandse militair
wordt het land van herkomst ten onrechte met "Holland"
aangeduid.
9. Niet bij ieder slot betekent linksom open en rechtsom
dicht; niet bij iedere kraan betekent linksomdraaien
meer en rechtsomdraaien minder vloeistof. Dergel ke
situaties, die een potentiële bron van dagelijkse
ergernis zijn, dienen door normalisatie te worden opgeheven.
10. Het nederlandse handbal biedt te weinig mogelijkheden
voor de niet-prestatie gerichte recreant die buiten
competitieverband zijn sport wil beoefenen.
11. Doordat met name science-fiction schrijvers, futurologen en schrijvers van populair-wetenschappelijk werk
over een computer spreken als ware het een met rede
begaafd individu dat zelfstandig informatie kan verzamelen en beslissingen nemen, kan de mythevorming
rond de computer alleen maar aanzienlijk versterkt
worden. Het verdient dan ook aanbeveling te spreken
over een rekenautomaat en te vermijden dat de indruk
ontstaat dat het om een beslissingsmachine gaat.
12. In de promotiereglementen van Hogescholen en Universiteiten in Nederland dient duidelijker naar voren te
komen, dat de promovendus zowel van het mannel k als
het vrouwelijk geslacht kan zijn.
13. Om te trachten de afstand tot de wetenschap voor een
ieder kleiner te maken, verdient het aanbeveling bij
proefschriften een gepopulariseerde samenvatting te
voegen.
M.M.E. Scheffers-Sap
Berlicum, 14 december 1979