Citation
Opiate receptor/agonist interactions

Material Information

Title:
Opiate receptor/agonist interactions evaluation of 15N NMR inept for resolving the clastic binding question
Creator:
Schilling, Kevin Harold
Publication Date:
Language:
English
Physical Description:
xi, 159 leaves : illustrations ; 29 cm

Thesis/Dissertation Information

Degree Divisions:
Department of Chemistry, CU Denver
Degree Disciplines:
Chemistry

Subjects

Subjects / Keywords:
Binding (Materials) ( lcsh )
Nuclear magnetic resonance ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 134-140).
General Note:
Submitted in partial fulfillment of the requirements for the degree of Master of Science, Department of Chemistry
Statement of Responsibility:
by Kevin Harold Schilling.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
18008657 ( OCLC )
ocm18008657

Full Text
OPIATE RECEPTOR/AGONIST INTERACTIONS:
EVALUATION OF 15N NMR INEPT FOR
RESOLVING THE CLASTIC BINDING QUESTION
by
Kevin Harold Schilling
B.S., Western Illinois University, 1978
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Master of Science
Department of Chemistry
1986
''ft
fi X B IT
*AU '


This thesis for the Master.of Science degree by
Kevin Harold Schilling
has been approved for the
Department of
Chemistry
by
Michael A. Mikita


iii
Schilling, Kevin Harold (M.S., Chemistry)
Opiate Receptor/Agonist Interactions: Evaluation of
NMR for Resolving the Clastic Binding Question
Thesis directed by Assistant Professor Michael A. Mikita
Clastic binding was the name given to the
hypothesis by Belleau and Morgan (1974) that
stereospecific opiates bind to their receptors as the
free base rather than in the N-protonated form.
The INEPT (Insensitive Nucleus Enhanced by
Polarization Transfer) pulse sequence can increase the
sensitivity of ^N to NMR by a theoretical factor of
9.87 when the nitrogen is directly coupled to a proton.
This enhancement may be used to determine if a nitrogen
is protonated.
A systematic enhancement vs. pH titration using
model compounds revealed that enhancements decrease
before mass-action considerations began to decrease the
protonated population as pH was raised. Variable
temperature studies and integral changes in the n/4J
INEPT delay established the dependence of polarization
transfer on proton-exchange dynamics, demonstrating the
inherent sensitivity of the method to dynamics-dependent
N-H contact times.


i v
An evaluation of the INEPT method applied to
resolving the clastic binding question of opiates has
been made. Attempts to mimic the receptor cavity using
crown ethers as well as systems that encourage ion pair
formation with the protonated amine have shown that the
physical confinement of a protonated nitrogen has the
ability of altering the proton dynamics to a situation
conducive to INEPT enhancement. These results suggest
that a study of INEPT enhancements from a solution of
solubilized receptor subtypes with a ^N-labeled opiate
may answer the question of clastic binding.


V
ACKNOWLEDGEMENTS
Robert Meglen and Larry Anderson for reading
the thesis. Robert Damrauer and Sandra Eaton for
comments. Steve Danahey for technical assistance.
Michael Mikita, who has fulfilled the highest standard
as an educator: having stimulated a lasting curiosity
in his student.


CONTENTS
CHAPTER
I. INTRODUCTION.................................... 1
Background.................................. 1
Ultimate Goal............................. 2
Thesis Problem.............................. 2
15N NMR..................................... 3
INEPT................................... 4
The rotating frame....... ............... 6
Vector description of INEPT.......... . 10
(
Optimization of delays.................. 15
Outline of Thesis Work..................... 20
II. RESULTS/DISCUS SION.......................... 22
INEPT Enhancement......................... 22
Enhancement vs. pH......................... 26
Concentration vs. activity.............. 37
The Dynamic Equilibrium................... 39
Lineshape analysis...................... 47
Temperature effects. ................... 59
Effects of altering INEPT delays..... 67
Solvent Effects............................ 73
Digression to general
solvent properties.................... 72


vii
Acetic acid............................. 75
Dimethyl sulfoxide..................... 78
Validity of pH measurements
in DMSO.............................. 81
Use of neat DMSO........................ 83
Crown ethers............................ 84
Ion-pairing............................. 88
Digression to a discussion
of ion pairing....................... 88
Observations............................ 91
Solvent acidity......................... 91
Solvent dimerization.................... 94
Receptor Binding........................... 98
Proton exchange and solvent exposure. 98
Specific interactions at receptor.... 102
Proton exchange in bound complex.... 104
INEPT and reaction (1)................. 110
Relaxation effects..................... Ill
III. CONCLUSION................................... 113
Summary................................ 113
Future directions...................... 115
IV. EXPERIMENTAL.................................. 117
Materials................................. 117
Instrumentation........................... 118
Methods................................... 120
INEPT..................................... 122


viii
Experimental Procedures..................... 123
Preliminary.............................. 123
Evaluation of delays................... 123
Enhancement............................ 124
Enhancement vs. pH...................... 125
Tropine................................ 126
Quinuclidine........................... 127
Ammonia................................ 127
Acetamide.............................. 128
Rate vs. pH.............................. 128
Lineshape analysis.................... 128
Temperature effects.................. .. 128
Altering INEPT delays................. 130
Solvent Effects......................... 130
Acetic acid........................... 130
DMSO................................... 130
Crown ethers........................... 132
Ion-pairing............................ 132
BIBLIOGRAPHY........................................ 134
APPENDIX
1. NMR SPECTRA.................................. 141


XX
TABLES
Table
1. Effect of pH on protonation
29


FIGURES
Figure
1. Precession of Nuclear Spins about B ....... 7
r o
2. Magnetization Vector in the Rotating Frame.. 7
3. a. 90 Pulse.................................... 8
3. b . 180 Pulse................................. 8
4. The Spin Echo Phenomenon......................... 11
5. a. 90 H........................................ 12
5. b. Delay 1................................... 12
6. a. 180 H....................................... 13
6 . b . 180 N....................................... 13
7 . a . Delay 2.................................. 13
7 b . 90 H........................................ 13
8. a. 90 N........................................ 15
8. b. Decoupling Delay............................. 15
9. a. Delay t 9 b . 90 H....................................... 16
10. Signal Intensity vs delay t...................... 16
11. The INEPT Pulse Sequence......................... 19
12. Tropine........................................ 24
13. Quinuclidine................................... 32


xi
14. Acetamide tautomeric equilibrium................ 34
15. Signal intensity vs. pH......................... 36
16. NH proton exchange rate in aq. HC1.............. 43
17. pH dependence of exchange rates of NH
protons with water........................... 45
18. N-H bond lifetime vs. pH........................ 73
19. Crown ether-cavity/cation complex............... 84


CHAPTER I
INTRODUCTION
Background
Prior to 1974 it was believed that
stereospecific opiates bind to their receptors in the
N-protonated form (Beckett, Casy, Harper 1956). At
that time Belleau and others established the importance
of the relative spatial orientation of the N lone pair
of electrons on analgesic activity (Belleau et al.
1974). From that it was suggested that the free base
may be the active form rather than the N-protonated
form (Belleau, Morgan 1974) with productive binding
including a stereospecific electron transfer to some
electrophilic site at the receptor. There has since
been much debate concerning this type of
structure-activity relationship of opiates at the
receptor level (Opheim, Cox 1976; Kolb 1984). At
present the problem is still unresolved.


2
Ultimate Goal
The ultimate goal of the opiate research in
this laboratory is to use ^N NMR signals from a
labeled opiate in an in-vitro environment of
solubilized opiate receptors to obtain direct evidence
as to whether or not the nitrogen is protonated at the
receptor site.
The noninvasive nature of NMR makes it
particularly well- suited for investigating interactions
of biological molecules without perturbing the
chemistry. Recent successes in the solubilization and
purification of opiate receptor subtypes (Chow, Zukin
1983) in active form from rat brain now make attempts
at this goal attractive.
Thesis Problem
The purpose of this thesis work was to
determine the feasibility of employing a particular
type of NMR experiment, the INEPT pulse sequence, to
the ultimate goal.


3
15n nmr
The basic theory of Nuclear Magnetic Resonance
is common to all experiments and all nuclei.
Familiarity with certain aspects of the theory aids
understanding how and why INEPT should work.
The fundamental properties involved are the
nuclear spin and the magnetogyric ratio, an expression
which describes the magnitude of the nuclear magnetic
moment. These properties are constants for a
particular type of nucleus. ^N and both have a
spin of 1/2 and so have two energy levels in an applied
magnetic field. The magnetogyric ratio of ^N is
smaller than that of by a factor of 9.87 and is
opposite in sign (Abraham, Loftus 1979, p.5).
The energy of interaction involved in an NMR
experiment is proportional to these two properties, at
a given applied field strength. The intrinsic
sensitivity of a nucleus depends on the Boltzmann
distribution resulting from the magnitude of the energy
difference between the nuclear Zeeman levels and, thus,
is largely determined by the very identity of the
nucleus. These factors, along with a natural abundance
of 0.365%, put the sensitivity of ^N relative to at
3.8x10 (Martin, Martin, Gouesnard 1981, p.3).


4
The abundance problem can be overcome by
introduction of ^~*N label into the opiate to be used.
The remaining, intrinsic insensitivity of ^~*N to NMR
actually forms the basis for choosing INEPT as a
possible tool for achieving the ultimate goal. This
will be discussed in detail later.
The quadrupolar nucleus experiences
randomly fluctuating electric fields with
14
characteristic frequencies at the N resonant
frequency which affords an efficient relaxation
mechanism. The short relaxation time resulting from
this electric quadrupole relaxation gives rise to
14
extremely broad lines making N unsuitable despite its
300-fold higher natural abundance.
INEPT
As mentioned above, nuclei with large
magnetogyric ratios (like ^H) are polarized to a
greater extent in a given magnetic field and, thus, are
much more sensitive than those with low magnetogyric
ratio, such as ^N. In 1979, a new method was reported
for enhancing the intensity of NMR signals from nuclei
of low magnetogyric ratio that are scalar coupled to
hydrogen (Morris, Freeman 1979). The enhancement


5
arises from the transfer of nuclear spin polarization
from the sensitive nucleus with large Boltzmann
population differences to the weak nucleus using scalar
coupling to force the weak nucleus energy levels to
follow the inversion of the sensitive nucleus. Morris
and Freeman adopted the acronym INEPT for "Insensitive
Nuclei Enhanced by Polarization Transfer" to this
method.
The method was later demonstrated specifically
for ^N coupled to hydrogen (Morris 1980). The
inherent insensitivity of ^N to NMR spectroscopy makes
its INEPT enhancement particularly striking. The
theoretical enhancement is equal to the ratio of the
magnetogyric ratios of the two coupled species which
for ^N coupled to hydrogen is 9.87.
This enhancement is a direct result of the
presence of a hydrogen coupled to the nitrogen and,
thus, may provide straightforward evidence of the
presence or absence of protonation.
A number of enhancement by polarization
transfer methods have been developed that utilize some
form of N-H coupling. These include Selective
Population Inversion (SPI) (Pachler, Wessels 1973),
Selective Population Transfer (SPT) (Jakobsen et al.
1974) and Nuclear Overhauser Enhancement (NOE)


6
(Abraham, Loftus 1979). INEPT is much 1 ess dependent
on precise knowledge of experimental conditions such as
the coupling constant and chemical shift values of
either nuclei than SPI and SPT (Garber 1983) and is not
subject to the possibility of signal nulling from
partial NOE. These considerations make INEPT
particularly suited for systems such as opiate-receptor
work in which coupling constants may be affected during
the course of various experiments.
The utility of INEPT and its theoretical basis
has been decsribed. A rigorous description of the
INEPT pulse sequence now follows. This description is
facilitated from the rotating frame of reference.
The rotating frame. When nuclei of spin 1/2
are placed in a magnetic field Bq they will precess
around Bq at a frequency characteristic of their
magnetogyric ratio, called the Larmor frequency, and
will be oriented either with or opposed to Bq as shown
in Figure 1. It is the Larmor frequency at which
resonance occurs. The applied field has also lifted
the spin degeneracy resulting in a slight Boltzmann
excess in the lower energy spin state. If the z-axis
is designated to be in the direction of Bq, then at
thermal equilibrium there will be a net magnetization


along the z-axis, and the net magnetization in the
xy-plane will be zero.
7
Figure 1. Precession of Nuclear Spins about Bq .
If the laboratory were rotated at the Larmor*
frequency around the z-axis, the nuclei would no longer
appear to precess and the Boltzmann excess could then
be represented by a single stationary magnetization
vector Mq aligned with Bq (Figure 2.). Since a nucleus
appears not to be precessing, it effectively
experiences a zero magnetic field.
Figure 2. Magnetization Vector in the Rotating Frame.
This condition is known as the rotating frame
of reference and makes description of the effects of


8
radiofrequency pulses and time delays of varying
duration straightforward.
A pulse of frequency equal to the Larmor
frequency applied along the stationary x-axis is
equivalent to applying a static field along the
rotating x-axis. In the rotating frame, this would
cause the net magnetization vector to rotate in a
clockwise direction around the x-axis in the zy-plane
at a frequency dependent on the magnetogyric ratio (and
the strength of the pulse). Particular pulse angles
are thus specific for each nucleus type and can be
described by the pulse duration needed to rotate Mq
through the desired angle. A 90 degree pulse is one of
correct duration to produce a rotation of Mq to the
y-axis (Figure 3a.). A 180 degree pulse is one that is
applied for twice as long and rotates Mq onto the
-z-axis (Figure 3b.).
Figure. 3 a. 90 Pulse.
b 180 Pulse.


9
The bulk magnetization vector Mq for a given
nuclear spin is a representation of all the individual
nuclei of that type, which actually precess at slightly
different frequencies due to inhomogeneities in the
applied field and interactions with other spins. The
individual components tend to separate in the xy-plane
after a 90 pulse as their frequencies differ from the
average Larmor frequency of M .
The spin-spin relaxation time, T^> describes
loss of bulk magnetization along the y-axis of the
rotating frame (after a 90 pulse) and is related to
the line-width of the resonance. The spin-lattice
relaxation time, T^, describes the re-establishment of
thermal equilibrium along the z-axis after a pulse that
displaces Mq from the z-axis. In many liquids, these
times are approximately equal and on the order of
seconds for and at least an order of magnitude
larger for (Martin et al. 1981).
Relaxation is considered complete after a
period equal to five times these values. The various
delays incorporated in the INEPT sequence are in the
millisecond range, so relaxation has a negligible
effect on magnetization during the actual sequence.


10
Vector description of INEPT. From a rotating
frame of reference, the INEPT pulse sequence can be
easily visualized using vectors to represent net
magnetization of nuclei under the influence of
heteronuclear scalar coupling. It has been shown that
the vector model can be considered a rigorous
description of the quantum-mechanical evolution
normally treated by a density matrix approach (Pegg,
Bendall, Doddrell 1981). The sequence is based on the
excitation of spin echos; a modulated proton spin
echo is used to invert one satellite of each proton
directly coupled to nitrogen, which leads to the full
Boltzmann spin polarization appearing across the
connected transitions (Morris 1980).
The spin echo phenomenon is illustrated in
Figure 4. After a 90 pulse, the spin system begins to
lose phase coherence due to spin-spin relaxation and Bq
inhomogeneities. Spin a_ is precessing faster and spin
b_ slower than the majority of the nuclei which are
represented by the bulk magnetization vector M The
frame of reference is rotating at the same frequency as
M After a time t, a 180 pulse is applied along the
x-axis, flipping the vectors over so that a is now
behind and b is ahead of M Because the individual
o
spins are still precessing at their individual


11
frequencies as before, after another identical time t,
a. will have caught up with M, which will have caught up
with _b, leading to a refocusing of the spins on the
-y-axis and formation of a spin echo at time 2t
(Abraham, Loftus 1979).
t ->
t -
Figure 4. The spin-echo phenomenon.
Spin echos are normally described in terms of
the vectors a_ and _b representing differing frequencies
due to inhomogeneities. Concerning INEPT, these two
vectors would represent the two frequencies of a
doublet .
The following description of the INEPT pulse
sequence is based on those provided, by Morris and
Freeman (1979); Garber (1983); and Brevard (1983, p.16)
and will be specific for (designated N) coupled to
a single proton (designated H). The frame of reference
is taken to be rotating at the ^N Larmor frequency in
the absence of proton coupling. Pulses are directed


12
along the x-axis and detection is in the y-axis.
Pulses can be phase shifted as necessary to be directed
along the y-axis.
After a 5T^ delay to allow buildup of full
polarization of proton spins, a net H magnetization
exists along the z-axis as in Figure 2 previously. A
O
90 H pulse is applied along the x-axis which rotates
the H magnetization onto the y-axis (Figure 5a.).
Because of coupling with N, this magnetization begins
to dephase into two components, H& and H^, whose
precessional frequencies differ by and appear to
move in opposite directions due to the intermediate
frequency of the frame of reference rotation. If we
allow a delay of t sec., where t=l/4J, the vectors
representing the two components will accumulate a
relative phase angle of 90 (Figure 5b.).
O
Figure 5 a. 90 H.
b. Delay 1.


13
A 180H pulse is then applied along the x-axis,
creating a mirror image about the xz-plane (Figure
6a.)- A simultaneous 180N pulse is applied which
interchanges the spin labels reversing their
direction, thereby maintaining coupling information
(Figure 6b.).
Figure 6 a. 180 H. b. 180N.
After another delay of t sec., spins coupled to
O
N .diverge and have a separation of 180 along the
x-axis (Figure 7a.). If a 90H pulse is applied along
the y-axis, these spins will now be aligned with the
z-axis (Figure 7b.).
Figure 7 a. Delay 2.
b. 90 H.


14
If relaxation during the two delays of time 2t
is neglected, the population across one H transition,
H^, is inverted, while the other is unchanged. Since N
and H share common energy levels, the N transitions now
have spin population differences possessed by H. If N
is observed with a 90 pulse along the x-axis at this
time (Figure 8a.), both components of the doublet will
be enhanced by the theoretical factor 9.87 with one
being inverted, no overall transfer of polarization
having occurred (Morris, Freeman 1979). In the absence
of J^, t*ie 180 N pulse in Figure 6b. would have had no
effect and the two H spins would have converged on the
-y-axis to form a spin-echo, losing the possibility of
polarization transfer.
If H decoupling were employed immediately after
the 90N pulse, the positive and negative components
would cancel and eliminate the signal. If a proper
delay is incorporated between the 90N pulse and
acquisition, the vectors representing the N spin states
could rephase to a single vector. Switching on the
proton decoupler at this time will produce a singlet
with a net magnetization equal to the sum of the
individual coupled vectors. The delay in this case of
nitrogen coupled to a single proton is also 1/4J.


15
After the 90N pulse, the vectors are opposed and along
the y-axis (Figure 8a.)- During the decoupling delay,
the two vectors converge on the x-axis (Figure 8b.).
Figure 8 a. 90N. b. Decoupling delay.
Optimization of delays. The optimum delay, t,
is a function of J and depends on the number of coupled
protons. For coupled to a single proton, the
enhancement factor varies sinusoidally with t, having
maximum absolute values at t=n/4J where n=l,3,5..., and
nulled signals if n=2,4,6..., with attenuated
enhancements for all other values (Rinaldi, Baldwin
1983). If, for example, the delay between the 180N
pulse in Figure 6b. and the 90H pulse in Figure 7b.
was too short,, the vectors would not be able to diverge
completely to the x-axis (Figure 9a.). Observation
with a 90N pulse would result in a signal with an
enhancement factor sin(r) of maximum where r is the


angle between the -y-axis and the spin vectors at the
time of the observe pulse (Figure 9b.).
Figure 9 a. Delay t The sine dependence of polarization transferred
on delay values allows inaccuracies to exist in the J
values assumed without, a great loss of INEPT
efficiency. This will be an asset to achieving the
ultimate goal as of the opiate may be affected upon
binding to the receptor. Graphical representation of
this dependence is shown in Figure 10.
Figure 10. Signal intensity vs. delay t.


17
A specific example of the effect of a small
error in setting the delays is demonstrated with the
nitrogen of pyrrole. Spectrum 1 was collected using
the optimum delay time corresponding to the literature
J of 96 Hz (Martin et al. 1981, p.220), while spectrum
2 represents the effect of using an inaccurate delay by
assuming a J of 81 Hz (Appendix 1.1 & 1.2). Although
the delay was in error by some 20%, the signal was only
reduced by approximately 5%. The spread of values of
one bond JMTI of aniline in various solvents was shown
to be less than 5% (Martin et al. 1981, p.66). Thus,
changes in from receptor interactions may be
disregarded.
The delay incorporated to allow refocusing
prior to decoupling and acquisition also takes the form
n/4J and exhibits exactly the same optimization
behavior.
INEPT spectra acquired without proton
decoupling will consist of two peaks for N coupled to a
single proton as would a conventional spectrum with the
exception that one of the peaks will be inverted. The
chemical shifts of the two peaks represent those of
multiplets so the coupling constant can be measured
on natural abundance compounds using INEPT (van Stein
et al. 1984). This is demonstrated using pyrrole.


18
Three undecoupled spectra of pyrrole were
collected using delays corresponding to values of
69, 96, and 119 Hz (appendix 1.3, 1.4, 1.5
respectively). All three spectra were doublets with a
peak separation of 96.4 Hz, in agreement with the
literature value cited previously of 96 Hz and with
only minor variations in signal intensity. This shows
that Jj^ can, indeed, be determined by INEPT using only
a reasonable guess for the delays. The value of J
determined can then be used to set the optimal delay
time for maximum polarization transfer in experiments
where J is not expected to change. If J does change,
however, the INEPT enhancement is not expected to be
l
suppressed to any great extent.
The INEPT pulse sequence can be written as:
(90xH)-t-(180xH)(180xN)-t-(90 H)(90xN)-t-(BB)-AQT,
where t is the appropriate delay based on J, BB
designates broadband decoupling at the proton
frequency, AQT is the commencement of acquisition, and
9C>xH represents, a 90 pulse of proton resonant
frequency applied along the x-axis, etc. For nitrogen
coupled to a swingle proton, all three delays are the
same. The vector representation of the entire sequence
is shown in Figure 11.


180 H.
180 N.
Delay 2 .
90 H.
90 N.
Decoupling delay.
Figure 11. The INEPT Pulse Sequence.


f
Outline of Thesis Work
20
The goal of the opiate research in this
laboratory is to resolve the clastic binding question
involving opiates and their receptors. The objective
of this work was to evaluate the ^N INEPT NMR method
for its utility to this end using model compounds.
The enhancement itself must be demonstrated as
sufficient to be used to differentiate between
protonated and nonprotonated nitrogens. An examination
of the effect of pH on enhancement is made and compared
with the predicted effect of pH on protonation. The
relationship between pH's effect on protonation and its
effect on proton exchange dynamics is then examined in
light of this comparison. The prevalent reaction type
is identified, one whose rates have been shown by
others to be comparable to the INEPT time scale.
Lineshape analysis was performed to extend their
results to our particular compouds.
Examination of INEPT enhancements in different
solvents is made in an attempt to predict what possible
effects changing the environment of the nitrogen part
of the molecule will have. These results are extended
to what is believed to take place at the receptor in a
discussion on Receptor Binding.


21
Clearly, dynamics may play a key role in the
utility of the INEPT method in answering the question
of whether or not an opiate is protonated at the
receptor. It was for this reason that a systematic and
careful analysis of dynamic effects on ^N INEPT
enhancements was undertaken.
Although the importance of dynamics on the
efficiency of INEPT had been suggested earlier (Marion
et al. 1982), this study was undertaken as the first
systematic examination of these effects.


CHAPTER II
RESULTS/DISCUSSION
INEPT Enhancement
An initial set of three experiments was
designed and carried out to quantify actual INEPT
enhancements obtained under various controlled
conditions.
As explained earlier, the maximum theoretical
INEPT enhancement for nitrogen coupled to a proton is a
factor of 9.87 over a spectrum created by a
conventional 90 N pulse experiment without NOE. Other
investigators report actual enhancements of 8 and 8.5
(Morris 1980; Live et al. 1984b). An example of the
INEPT enhancement was demonstrated in this laboratory
using the ammonium nitrogen of ^N enriched ammonium
nitrate. Two spectra were collected, one using INEPT,
and the other without INEPT. Decoupling was not
employed so the results show polarization transfer
enhancement only with no additional signal intensity
from multiplet collapse. All other conditions were


23
identical between.the two. Comparison of these two
spectra demonstrates the effect of INEPT on signal
intensity (Appendix 1.6 & 1.7). The gain in the signal
to noise ratio realized by using INEPT is approximately
8.3 times that of a conventional spectrum. This
enhancement is, therefore, judged to be of sufficient
magnitude, under these conditions, to demonstrate that
INEPT enhancement does, indeed, have diagnostic value
for the determination of protonation of a nitrogen.
Next, an experiment was conducted on the
nitrate nitrogen of the same compound to determine if
proton-free nitrogens are affected in any way. The
conditions are again identical for each of the two
spectra, one with and one without using INEPT. These
spectra, (Appendix 1.8 & 1.9), indicate that INEPT does
not enhance the signal from this nonprotonated
nitrogen.
The previous two experiments showed enhancement
from INEPT for a protonated nitrogen and lack of
enhancement for proton-free nitrogen when compared to
non-INEPT spectra. These comparisons required the
ability to obtain a signal from a non-enhanced ^N
nucleus in a reasonable time. The use of ^N enriched


24
samples enabled this to be done. The availability of
label limited these two experiments to dissimilar
nitrogen types, however. The third experiment of this
set integrated the enhancement concepts just
established by using the same nitrogen while changing
only its protonation. The tertiary bicyclic amine,
tropine (Figure 12.), was chosen due to its structural
similarity to opiates and its availability in quantity
as a non-controlled substance.
samples, at high and low pH, all else identical. In
this experiment, it was possible to observe the
nitrogen with INEPT while changing its protonation and
keeping the molecule type the same. The protonated
sample gave a generous signal, while no signal at all
was observed with its non-protonated counterpart
(Appendix 1.10 & 1.11).
Figure 12. Tropine.
An INEPT spectrum was collected from two
This set of experiments establishes the fact
that the signal from a protonated nitrogen nucleus is


25
enhanced to a significant degree when observed with
INEPT as compared to the corresponding non-INEPT
spectrum. It was also determined that the use of INEPT
has no enhancement effect on the signal from a
non-protonated nitrogen nucleus. Long range coupling
between N and H was thus demonstrated to be ineffective
for INEPT enhancement, at least when the interpulse
delays are set for polarization transfer through
one-bond coupling. From this, the statement can be
made that under these controlled conditions, any
significant difference in signal intensity between an
INEPT and non-INEPT spectrum is the result of direct
protonation, as expected.
The point is now made that non-protonated
nitrogens with or without INEPT and protonated
nitrogens without INEPT all show similar unenhanced
signal intensities. This seemingly redundant point may
be valuable later when the need to ensure against false
indications of protonation becomes a possibility as
experimental conditions become more complex. The
indication of protonation by use of INEPT will not be a
result of enhancement alone, but rather as a difference
between the INEPT and non-INEPT spectra.


26
Enhancement vs. pH
The in-vivo opiate receptor interaction takes
place at the normal physiological pH, approximately
7.4. Having established the qualitative effect of
protonation on INEPT enhancement, attention now turns
to evaluating INEPT at intermediate pH's. A comparison
is made between enhancement and predicted protonation
at several pH's.
One of the advantages of pulsed Fourier
Transform NMR is the ability to add consecutive scans
together to increase the sensitivity. The signal will
add coherently and its intensity will be directly
proportional to the number of scans taken. The noise,
being random, will be additive only as the square root
of the increase in the number of scans. The overall
improvement in the signal to noise ratio over a single
scan will,, thus, be according to the square root of the
number of scans if the spectrometer is stable in
tuning, field strength, etc. An improvement of
ten-fold in the signal to noise ratio requires an
increase in data accumulation by a factor of 100.
This, along with the relative insensitivity of
natural abundance ^ N, makes it possible to choose a
number of scans such that a signal would be


27
undetectable under conventional conditions but using
INEPT would give a sizable S/N for a protonated
nitrogen. That is, a non-protonated nitrogen would
make no apparent contribution to the signal because of
its insensitivity. Although the point made previously
that the indication of protonation would be enhancement
and not the mere presence of a signal makes this
condition unnecessary to the ultimate goal, it
facilitates the quantitative study of pH effects on
INEPT.
Even though the possible changes in chemical
shift with protonation may rule out contribution from
free base nitrogen to the protonated signal, using the
conditions outlined above gives assurance that signals
observed are representative of only the protonated
fraction.
The size of an NMR signal is directly
proportional to the number of nuclei producing it. If
spectra were accumulated using a predetermined number
of scans as described above, the size of the signal
will then be directly proportional to the number of
protonated nuclei alone.
In acidic solution (99+% protonated), the
signal will be at its maximum intensity. In basic
solution (<1% protonated), the signal will be at its


28
minimum intensity, undetectable under the chosen
conditions. In this realm, what is being observed is
the enhancement only since nonprotonated nitrogens
yield no observable signal nor would protonated
nitrogens if not for the enhancement. Since the
enhancement is a result of INEPT which in turn is
dependent on protonation, the signal and its intensity
is a direct result only of protonation.
Table 1 gives the fraction of the nitrogen
nuclei protonated as a function of pH for the compounds
quinuclidine, tropine, and ammonia. Their nitrogen
pK 's are 11.45^, 10.2^, and 9.25^, respectively.
3
These data were determined using a modified
Henderson-Hasselbalch equation: logR=pK -pH where R is
3
the ratio of protonated to non-protonated nitrogens,
HN/N. The fraction protonated, F, is then determined
by: F=HN/C,where C=HN+N, the total concentration of
the compound.
"''Alumni, S.jJencks, W. (1980) J. Am. Chem.
Soc. 102., 2052.
. ^Merck Index (1980) 10t'1 Ed. p.9446.
Merck & Co., Rahway, N.J.
^Handbook of Chemistry & Physics (1981-2)
62n Ed. p.D-144 CRC Press, Boca Raton, Fla.


29
Table 1. Effect of pH on protonation.
pH Fracti Quinuclidine on protonated Tropine Ammonia
1.0 >0.99 >0.99 >0.99
7.0 >0.99 >0.99 >0.99
8.0 >0.99 >0.99 0.95
9.0 >0.99 0.94 0.64
10.0 0.97 0.61 0.15
11.0 0.74 0.14 0.02
12.0 0.22 0.02 <0.01
13.0 0.03 <0.01 <0.01
14.0 <0.01 <0.01 <0.01
As a result of the conditions outlined and with
dynamic effects aside, a plot of signal intensities vs.
pH using INEPT should resemble that of an ordinary
acid-base titration curve with peaks of maximum
intensity at the low pH limit. The peak where the pH
is equal to the pK of the compound should have an
3
intensity exactly one-half that at the low pH limit.
Also, one would expect a baseline of maximum
absorptions of nearly the same size at several of the
lower pH's since the overall effect of changing. pH on
protonation is relatively small in the region far below
the pK (eg. 99.999 vs. 99.990% protonation). As pH is
3
increased, only as it begins to approach the pK should
the peaks start to become less intense (a compound with
a pK of 11 would still be 91% protonated at pH 10).
3


30
As mentioned before, the signal intensity is
actually a quantitative representation of enhancement
only and is directly proportional to the number of
protonated nitrogens with maximum intensity in the low
pH limit. The value F would then also represent the
expected magnitude of the signal intensity as compared
to maximum.
Several spectra were collected from these
compounds each following incremental increases in pH
starting at the low pH limit. Signal intensities at
several pH's for these compounds are shown in Appendix
1.12-1.14. A quantitative graph of signal intensity
vs. pH for these compounds will be seen later. Gross
observations are now discussed.
The nitrogen signals of tropine began to
decrease in intensity between the initial pH adjustment
from 0.2 to 1.5 and continued to decrease until the
peak was completely absent by pH 5. Further increases
in the pH of the solution to 6.6, 7.4, 9.3 and 12.4
still resulted in an unenhanced spectrum. According to
the calculations shown in Table 1, this loss of
enhancement took place at a pH range where nitrogen
protonation should be essentially 100%. Some possible
contributors to this observation must be considered.


31
Most opiates, including tropine, have an OH
group on one of the ring carbons. Dehydration
involving this OH would result in a structural change
possibly leading to an alteration of the nitrogen pK .
. 3
13
This structural change was ruled out with C NMR at
low and high pH.
Tropine should have two pK^'s, one for the
proton on the nitrogen and one for the phenolic
hydrogen. An investigation by Kaufman et al. (1975),
i
of several opiates showed the pK of the OH hydrogen to
3
be generally 1-1.5 units higher than that of the
nitrogen. The oxygen, thus, would tend to remain
protonated to a greater extent than the nitrogen as pH
increases.
In addition, the decrease in enhancement seen
at lower than expected pH's is. more consistent with
external proton demand. The amphoteric OH group may
also be protonated, which would tend to consume
protons. Since the basicity of ROH is less than HOH,
however, this phenomenon would be expected to make
little contribution in an aqueous solution. Thus, the
possibility that participation of the OH group in the
equilibrium makes any.contribution to the enhancement
vs. pH phenomenon observed seems unlikely in the pH
range in which the observations were made.


To further eliminate the question of the OH
group affecting the protonation of the nitrogen, the
32
INEPT enhancement vs. pH characteristics of the
compounds quinuclidine and ammonia were investigated.
tropine with the signal intensities decreasing at pH's
well below those required for significant
deprotonation. This commenced as pH was increased from
2.0 to 3.0, with the decrease continuing until
suppression was complete by pH 5.0. The initial
adjustment from pH 1.1 to 2.0 had no effect on the
enhancement.
Peak suppression began between the adjustment from pH
0.3 to 1.0, with total absence by pH 3.0. The initial
adjustment from 0.0 to 0.3 had no effect on the
enhancement.
For these three compounds, the signals began to
decrease in intensity and disappear completely at pH
Figure 13. Quinuclidine.
Quinuclidine exhibited the same behavior as did
The same observations were made for ammonia.


33
levels much lower than would be expected based on
equilibrium considerations alone, that is, at pH levels
where the nitrogens are still essentially fully
protonated.
Since the signal intensity is being affected as
a result of changing pH in a region where the degree of
protonation is staying essentially constant and the
equilibrium is not being altered by other functional
groups, it would seem likely that the signals are being
affected by pH itself. Under the assumed conditions,
the signal is actually an indication of enhancement
from the INEPT method, so pH-labile phenomena are
affecting the INEPT method itself even while full
protonation is effectively maintained.
Since the previous experiments were performed
on compounds with pK 's from 9-11, the next logical
3
step was to evaluate a compound with a nitrogen pK of
2 or less to determine more clearly the dependence of
the role of pH in the peak suppression. If the
phenomenon of enhancement suppression observed with the
other compounds is a characteristic of the pH range
only, then the enhancement vs. pH of a nitrogen with a
sufficiently low pK would likely exhibit peaks
- 3
representative of protonation. Acetamide was chosen
4
since it has a pK value of 0.63 .
4Ibid. D-139.


34
An enhancement vs. pH experiment, was conducted
on acetamide in the same manner as previous examples
(Appendix 1.15). A major difference in the results
surfaced in the low pH region, (pH 3.
initial pH of <0.0 to 0.4, the signal increased in
intensity, then it began to decrease and then
disappeared above pH 2.5. In previous experiments, a
constant signal intensity could be obtained at the low
pH limit which did not change with further pH
decreases. Acetamide gave maximum enhancement at a pH
approximately equal to pK Enhancement decreased at
higher and lower pH. The pH of maximum signal
intensity was 0.4, where the intensity would be
expected to be somewhat greater than one-half of
maximum.
Investigations of proton transfer reactions
involving acetamide indicate that oxygen is the primary
site of protonation in extremely acid solutions
(Berger, et al . 1959; Liler 1971) An equilibrium
exists between compounds A & B in Figure 14 with B
dominating as the pH becomes very low.

C n-h C N
sh CH^
Figure 14. A. B.
Acetamide tautomeric equilibrium.


35
The prevalence of species B at very low pH
suggests that protonation of oxygen is responsible for
the decrease in enhancement as pH is taken to low
levels by deprotonating nitrogen, or increasing the
rate involving nitrogen deprotonation.
Rate studies of proton exchange involving the
amino group in acetamides and amino acids reveal a
decrease in rates as pH is raised from very low levels,
then the rates increase as pH is raised further (Klotz,
Frank 1965; Buchner et al. 1978).
The investigation of the enhancement vs. pH
behavior of a nitrogen with a low pKg was carried out
in an attempt to characterize more fully the
relationship between pH, pK and enhancement. The
3.
results obtained are complicated by too many variables
to be of use in this thesis. For this reason, no
further investigation of this type was conducted nor
will additional explanation of the observations be
j
offered.
Figure 15 is a plot of INEPT signal intensity
as a function of pH.. The signals are placed
quantitatively along the abscissa at the measured pH at
which they were taken and the ordinate represents peak
height.


36
1.0
2.0
3.0
4.0
PH
Figure 15. Signal Intensity- vs. pH
This shows a true quantitative relationship between
enhancement and pH, whereas the spectra in the Appendix
are arranged to emphasize the existence of the baseline
maximum in the low pH limit.
As a result of these experiments it appears
that INEPT enhancement, although requiring protonation,
is not an accurate measurement of protonation as a
function of pH. Discussion now turns to why this is
so.


37
Concentration vs. activity. The effects of
electrolyte concentration on equilibria are well known
(Castellan, 1971). The presence of other ions in
solution changes the effective concentration of ions
involved in the equilibrium through electrostatic
interactions. As the pH was changed in conducting the
enhancement vs. pH experiments, certain amounts of
ionic salts were introduced that made the ionic
strength at the end of the experiment higher than it
was at the beginning. The concentration parameter
called the activity, which varies with ionic strength,
can be used instead of molar concentration to free the
dependence of K values to changes in ionic strength and
ensure accuracy in equilibrium calculations.
Typically, the.use of activities in ionic
solutions causes K to vary by a factor of 2 or 3, and
in extreme cases by as much as a factor of 10
(Blackburn 1969, p.6). Changing K by a factor of 10
would alter the pK by one unit. The enhancement vs. pH
curves in Figure 15 appear shifted to the left by some
six units, far more than could be accounted for by
assuming an incorrect value for K.
Grunwald (1963) applied the approximation of
Bunnett (1960) that logK = logK 0 + H + log[HCl] +
3 3.
wloga, to various amines, where H is the acidity
function, a is the water activity and w is a


38
molecular-structural parameter characteristic of a
tertiary amine. For the worst case of the highest
value of w found by Bunnett in HC1 and at a pH of <0.2,
the value of would decrease by approximately half.
This corresponds to the pH vs. enhancement curve being
shifted to the right by about one third of a pH unit.
This is in agreement with some typical values compiled
for the change in pK^ from at least one review
(Kolthoff 1979, Chap 13).
The magnitude of the effect of electrolyte
concentration on chemical equilibrium is dependent on
changes in the charges involved in the equilibrium,
becoming greater as the charges on one side of the
equilibrium increase. The reaction of primary interest
here is the transfer of a hydrogen ion from a nitrogen
to a water molecule, as will be described later. When
a proton is transferred to a neutral solvent molecule,
no separation of charge occurs, so this type of
equilibrium is inherently less sensitive to activity
effects. No changes in measured pH were observed upon
adding excess salt in the form of KC1 to a solution
typical to these experiments.
Any change in equilibrium values due to the
effect of using activity vs. molar concentration,
therefore, is concluded to be minimal.


39
The Dynamic Equilibrium
Observations made on resonances of amino
acids indicate that chemical exchange occurs at the
proper rate to affect the N-H scalar interaction under
certain conditions (Blomberg, Maurer, Ruterjans 1976).
The effects resulting from the dynamics of proton
exchange on INEPT spectra must now be addressed.
Previous work cited on similar amines will allow
evaluation of the possible equilibria taking place in
terms of their order of importance under certain
conditions and their rates as a function of pH. This
may allow a specific reaction(s) to then be studied in
some detail.
The equilibrium calculations made previously
assume the reactions involved to be reversible and the
system to be in the state of thermodynamic equilibrium.
The relative, amounts of constituents do not vary as a
function of time allowing the calculation of F
(fraction protonated) from pH and pK as was done. The
forward and reverse reactions continue, however, on a
molecular scale at equal rates. These rates and the
position of equilibrium are independent of one another.
There exists a continual transformation of reactants to
products and vice versa.


The phenomenon taking place here is the
transfer of a hydrogen ion from one molecule or ion to
another. Of particular concern are the protons
directly involved in the polarization transfer to
nitrogen. A proton can move to and from a number of
places. In a system of water, nitrogenous base, and
acid (HC1), the following protolytic reactions are
applicable:
R3N+H + B <- -> r3n + B+H (I)
R3N + B+H' <- - R3N+H' + B (ID
where B indicates a basic entity of any form.
In the pH range in which the enhancements were
affected, would be a short-lived intermediate due
to its low concentration which results from its
basicity. Using a pK of 11.45 for quinuclidine, the
concentration of free base at pH 3, where the effects
of exchange are beginning to become apparent, is
8
calculated to be <10 M. Type (I) reactions are,
therefore, more prevalent. They are of most interest
because of this, and, more importantly, since the
protonated nitrogen is involved.


41
The reactions of type (I) are summarized as follows:
R3N+H + 0H2 < * R3N + H+0H2 (1)
R3N+H + oh r3n + HOH (2)
r3n+h + NR3 4 r3n + H+NR3 (3)
R3N+H + oh2 + NR3 - '
r3n + H'OH + h+nr3 (4)
From kinetic studies of the protolysis of
methylammonium ion, Grunwald et al. (1957) deduced a
total rate of reaction (I) to be:
k=kl+(k2Kw+(k3+k4)Ka[R3N+H])/[H+]
=rate/[R3N+H],
where Kw=[H+][0H~], and L=[H+][R3N]/[R3N+H]. Their
work showed rate increases as the pH was varied from 3
to 5. The product k[H+] was approximately constant for
a given value of [R3N+H]. Many literature examples
exist that also show that the proton exchange rate
increases with pH (Emerson et al: 1960; Grunwald et al.
1956) possibly by OH and H20 base catalysis (Blomberg
et al. 1976). Sheinblatt and Gutowsky (1964) found the
R3N+H lifetime between exchanges to decrease as pH
increases.


42
Various studies on tertiary amines determined
the rate constants of the forward reactions to be on
the order of 10^ s for (3) and (4); 10^ s-'* for (2)
and value's from tenths to several hundred sec-'*' for
(1). One overall rate took the form: 10^k[H+] M s-*-
with the value increasing from 2 to 15 as the
concentration of the amine is increased (Grunwald et
al. 1956). Sheinblatt's work on amino acids yielded
values for the rate constant for reactions of the type
(1) from 110 to 470 sec. Proton exchange rates of
k[R^N+H][R^N] with k=10^ s ^ were determined for
imidazole ^ of histidine (Alei et al 1980).
At the hydroxide concentrations of these
experiments, reaction (2) would not be significant even
with the magnitude of its velocity constant (Emerson et
al. 1960). Reactions (3) & (4) have the same
dependence on pH and concentration and roughly the same
rate. Since the concentration of free base was
calculated to be very small, the contribution of these
reactions is diminished. Grunwald1s 1963 study gives
an overall rate =k^[R2N+H] + k^[R^N+H][R^N] where the
first-order term is predominant at low pH where the
increase in rate with pH must be interpreted as a
medium effect on that term. The constant k^ was
concluded to be a function of [H+], Figure 16 is his
plot of rate vs. pH for 0.2 M trimethylamine HC1.


43
It can be concluded that, due to the particular
experimental conditions, reaction (1-) is a major
contributor to the mechanism whereby protons are
abstracted from nitrogen.
Figure 16. NH proton exchange rate in aq. HC1.
Source: Grunwald, E. (1963) J. Chem. Phys.
67, 2211.
Emerson et al. (1960) found that the rate
constant for reaction type (1) decreases with
increasing electrolyte concentration. The 1963
Grunwald study concluded that the first-order rate
constant for that reaction using trimethylammonium ion
decreases markedly at pH<2 and the greater part of this
effect is due, however, to something other than a
neutral salt effect. The effect of electrolyte


44
concentration on rate was found to be much greater for
HC1 than that of other chlorides.
When an ion reacts with a neutral molecule, the
transition state has the same charge as the reactant
ion. The specific rate in this case will be
independent of ionic strength over a considerable
range, ignoring dispersion. The insignificance of the
effect of ionic strength on equilibrium position may
thus be extended to rates as well.
Due to the large difference between [H+] and
[H^O] at a pH of 3, the pH dependency of proton
abstraction by H 2 0 as in (1) is not readily apparent.
This reaction takes place by a 2-step mechanism which
will be described in detail later.
The enhancement vs. pH plots indicated a
correlation between the pH of complete signal
suppression and the pK of the particular compound.
3
The plots appear shifted along the pH axis by a number
of units similar to the difference in pK s. The trend
a
that appeared shows an approximate unit increase in the
pH of suppression with pK Calculation of rates from
3
values determined by Grunwald et al. (1960) show the
overall rate constant to increase with higher K^.
Nitrogen-15 NMR work on glycine and its ester by Cooper
et al. (1973) found the difference in pH units for


45
exchange dependent line-widths to closely approximate
the difference in pK^'s of the two species.
Blomberg et al. (1976) suggested that the sharp
increase in exchange rates is a measure of the basic
character of the nitrogen group very much like the
individual pK value. Figure 17 shows the pH dependence
of exchange rates for alpha amino protons with water
for glycine, lysine, and proline. The curve to the
right is that of proline, which has an alpha amino pK
approximately 1 unit higher than the other two
(Lehninger 1975, p.79).
Figure 17. pH dependence of exchange rates of NH
protons with water.
Source: Blomberg, F., et al. (1976) Proc.
Natl. Acad. Sci. USA. 73, 1409.
Close inspection reveals a strong correlation between
the pH at which rates sharply increase and the pK 's of
3


46
the nitrogens. Those pK^'s are similar to those used
in this investigation and the pH range in which
enhancements decreased matches where the rates showed
the sharp increase. The similarity with respect to the
pH axis between this plot and that in Figure 15 ties
increases in proton exchange rates of similar compounds
with loss of INEPT enhancement.
The preceeding discussion determined that
proton exchange can be a plausible explanation for the
pH effects on the INEPT enhancement from work done on
similar compounds. The mean lifetime of a proton on a
nitrogen can be expressed by 1/k^. Examples of typical
proton abstraction rates from nitrogen in the various
studies cited correspond to N-H bond lifetimes
comparable to the critical time dependence of INEPT
polarization transfer from proton to nitrogen. The
effect of pH on these rates was greatest in the range
where INEPT was affected. Lineshape analysis through
the use of proton NMR on the systems of this thesis
will determine if the same phenomena apply specifically
to these compounds.


47
Lineshape analysis. The observation of
spectroscopic line shapes to study rates of simple
proton transfer reactions depends on the uncertainty
principle. A state with a short lifetime has an
uncertainty in its energy depending on its lifetime.
If this state is involved in a spectroscopic
transition, the uncertainty of the energy of that
transition manifests itself in an uncertainty in the'
frequency of the absorption which leads to a broadening
of the spectroscopic linewidth (Bell, 1973, p.119). As
a result, exchange rates can be determined from line
broadening.
Proton magnetic resonance can provide a direct
means of measuring fast exchange rates due to its
sensitivity and its being a direct means of observing
the exchanging protons. It indicates which atoms are
exchanging and at what rate while the measured system
is in equilibrium (Hahn, Maxwell 1952).
As a result of reversible proton transfers,
such protons experience more than one discrete
resonance frequency, corresponding to their multiple
magnetic environment. The chemical shift and the spin
coupling of certain protons then becomes time
dependent.


48
If a proton can exist in two environments, the
observed proton spectrum depends on the rate of proton
exchange between them. If the exchange is slow on the
NMR time scale (or frequency difference between the two
environments) they will show up as two distinct peaks.
If the exchange rate is fast, there will be only a
single peak representing a weighted average of the
environments, the peak becoming, sharper with faster
rates as the uncertainty in the energy of the averaged
state becomes less. At intermediate rates, the
spectrum begins to be transformed from one type to the
other,the resolvability becoming blurred as the peaks
appear to coalesce. It is from this broadened spectrum
that information concerning rate constants of the
processes involved can be derived.
In acidic solutions, an amine group exists
predominantly in the RN+H form and proton exchange is
slow. In less acidic solution, the exchange rate
increases and resolution of the N-H interaction also
becomes less as the exchange rate approaches the NMR
time scale of the interaction.
Proton spectra were collected from quinuclidine
at pH's 0.8, 1.8, 2.8, 3.6, 5.3, and 7.5 (Appendix
1.16-1.21). At pH 0.8, a triplet is seen for the N-H
proton with a chemical shift 8.93. Each linewidth is


49
is approximately 18 Hz, and the splitting is well
resolved. The pH was increased to 1.8 and then to 2.8
with no noticeable effect on the splitting or
linewidth. At pH 3.6, the splitting was less
resolvable, with the peaks some 40 Hz wide. At pH 5.3,
a broad hump is seen with a linewidth approximately
equal to the entire multiplet seen before, about 188
Hz, and its center shifted from the center of the
triplet approximately 0.6 ppm toward the water
resonance. At pH 7.5, the hump is smaller in height
and shifted an additional amount.
At high pH the water line also includes protons
which are exchanging rapidly between nitrogens and
water, which would cause a shift in the water
resonance. This effect is not noticable due. to the
concentration difference between nitrogen and water.
The methylene absorptions appear to be little
affected through the entire pH range observed. This
was as expected, since they were assumed not to be
directly involved in the exchange with their only
perturbation being through long range coupling with the
N-H protons undergoing exchange. This interaction was
too weak to be observable.
The number- of lines of a mult-irpTet to a
first-order approximation is 2X1+1 for each set of


50
equivalent nuclei, where X is the number of equivalent
nuclei of spin I in each set which the nucleus in
question (H), is spin coupled (Witanowski, Webb 1973,
p.9). ^4N has a nuclear spin of 1 and so causes
spin-spin splitting into three components. Due to the
weakness of the interaction of N-H protons with any of
the other nuclei and the broadness of these peaks, this
triplet is the only splitting resolved for those
protons. The broadness of the absorption is attributed
14
to the efficient quadrupolar relaxation of the N
nucleus.
The coupling interaction of protons to
different nuclei is related to the magnetogyric ratios
of the two nuclei, giving the following relation:
J(14NH)=-0.7129 J(15NH). The value for J(15NH) using
15N INEPT was found to be 80 Hz (see EXPERIMENTAL,
CHAPTER IV). Using the above relation, JC^N) should
be 57 Hz, in perfect agreement with the 57 Hz between
the individual lines of the N-H triplet obtained at low
pH's (see Appendix 1.16).
The overall observation is that as the pH
increases, the N-H proton lines become broader, less
resolved, indicating an increase in the rate of
exchange involving those protons. This broadening
gives a measure of the mean lifetime of the N-H bond.


51
The value of the residence time, t, can be
inferred in the restricted interval where the exchange
rate has a noticeable influence on the resonances. If
h is the linewidth at half-maximum, in Hz, and is less
than the splitting, then the following relation holds:
7rh=l/T2 + 1/t (a)
where t is the average time that a proton is bonded to
a nitrogen of a given spin state (Emerson et al. 1960).
The natural linewidth of a resonance in the
absence of exchange is a function of the spin-spin
relaxation time, T^. The 1/t term accounts for the
additional broadening brought on by exchange and drops
out when exchange does not affect the linewidth. In
the low pH region, linewidth depends only on the
spin-spin relaxation time, T2> The consistent value of
18 Hz for the linewidth at pHs 0.8, 1.8, and 2.8
implies that linewidth is unaffected by exchange and
represents the natural linewidth. This corresponds to
a T2 of 0.018 sec., in agreement with 0.02 sec.
obtained by Grunwald et al. (1957). It also can be
stated that t<0.018 sec. at pH 2.8 and below.
Using equation (a) and a T2 of 0.018 sec., at
pH 3.6, where exchange modulated broadening first
becomes observable, a value of approximately 0.014 sec.


52
is calculated for the mean N-H bond lifetime from the
increase in linewidth to 40 Hz resulting from raising
the pH from pH 2.8. For quinuclidine, the INEPT delays
are 0.0031 sec. An INEPT experiment with decoupling
has three such delays, requiring N-H contact times of
0.0093 sec. for its completion.
For this discussion the mean lifetimes will
also represent the point where half the total number of
lifetimes are greater and half less. At pH 3.6, half
the N-H bond breaking reactions would involve rates
greater than 1/0.014 sec., and half less. Under
conditions that would yield a mean rate of 1/0.0093
sec., the required residence time for polarization
transfer, half the N-H bonds would exist for longer
than 0.0093 sec. Under those conditions, the
enhancement would be expected to be half of maximum.
Without going into detailed probability calculations,
it can be said that a mean lifetime of 0.0093 sec. will
occur at a pH slightly above pH 3.6, where the mean
lifetime is 0.014 sec.
A mean residence time of 0.0093 sec. is reached
as the ^H linewidth increase passes through the 50 Hz
region according to equation (a). This additional
information comes from the proton spectra that INEPT
enhancements af half maximum are expected at a pH just
over 3.6, where the linewidth is 40 Hz.


53
Examination of the enhancement vs. pH relation
of quinuclidine (Appendix 1.13), shows that the peak at
pH 3.0 was about 75% of maximum, and at pH 4.0, about
25% of maximum. From this alternate approach, the
enhancement of half maximum intensity, which according
to activation theory corresponds to a mean residence
time of 0.0093 sec., occurs at a pH between 3 & 4.
A relation also exists that applies exactly at
coalescence for a symmetric multiplet involving
protons, assuming all are exchanging with water
(Sandstrom 1982). At coalescence, the exchange rate
1/2
can be derived from the coupling by: k=irJ/2 Using
a multiplet separation of 57 Hz, the k obtained at
coalescence, which occurs between pH 3.6 and 5.3
(probably closer to 3.6) is 127 sec. ^, giving a mean
lifetime of 0.008 sec. The INEPT enhancement at pH 4.0
was approximately 35% of maximum. It is not
unreasonable to assume that a mean t of 0.008 sec.
would translate to 35% or so of the N-H bonds lasting
at least 0.0093 sec.
The lineshape of the ^N absorbance also can be
used to measure the mean lifetime of the N-H bond
before proton transfer to another nucleus. The line
width is inversely related to T£ and the residence
time, the average time that a proton is bonded to


54
nitrogen of a given spin state. The same broadening,
coalescence, etc. phenomena apply as in proton spectra.
Dynamic NMR work done by Blomberg et al.
1976 showed a marked effect on the lineshape as pH (and
exchange rate) was changed. Broadening of the
signal occurs as the rate of chemical exchange
approaches the coupling constants. Deviations from
that particular rate (from changing pH) resulted in a
2 2
sharp resonance when J t <<1 where t is the average
lifetime before exchange, (1/rate), where J is in
radians/sec.
Broadening of a coupled spectrum commences as
the rate of exchange, t (where t is the preexchange
time), approaches the spin-spin coupling constant.
Coalescence to a broad, single-line absorption occurs
2 2
when J t =1 (Cooper et al. 1963). For quinuclidine,
with a coupling constant of 80 Hz, coalescence will be
attained as t approaches approximately 0.002 sec. From
the proton spectra, and equation (a), this exchange
14
rate is estimated to occur at the pH that the N-H
proton absorbance is 177 Hz wide. This is estimated to
be near pH 5. Enhancements were drastically affected
below this pH and no peaks were seen at any higher pH,
meaning enhancements are reduced before this timescale
is reached with complete suppression upon attainment of


55
this condition. This should be expected, since it is
through this coupling that polarization transfer
occurs.
The above time scales involving coupling
constants describe coalescence of splitting patterns.
Since the spectra were collected under decoupling
conditions, it may be argued that these relations would
not apply. Regression to the vector description of
INEPT decoupling shows the purpose of the third,
decoupling delay of allowing the individual splitting
multiplets, & H^, to converge to a single vector
before observation (Figure 8.). The accuracy of
this convergence would be affected by the resolvability
of the multiplet, which would reveal itself in the
resulting peak. Since decoupling does not take place
until ^N acquisition at the end of the sequence, the
coupling interaction is present during the INEPT spin
evolution described.
Conventional ^N NMR spectra as a function of
pH usually also show a variation in chemical shift as a
result of the alteration in the environment of the
nitrogen that occurs due to the periodic presence and
absence of hydrogen. Changes in chemical shifts of up
to 73 ppm have been observed for imidazole N^ of
histidine (Alei et al. 1980). A downfield shift of


56
13.7 ppm on protonation of quinuclidine was reported by
Duthaler and Roberts (1978). For cases in which these
changes are observed, two.peaks are seen corresponding
to each nitrogen type at slow exchange.
Assuming a 13.7 ppm change in the chemical
shift upon protonation, lineshape changes in the
peak from this would require even higher rates than for
effects on splitting (J=80 Hz vs. 13.7 ppm=lll Hz at
8.1 MHz). This suggests that effects on lineshape by
changes in chemical shift would require even higher
rates than for loss of coupling, thus, would occur at a
higher pH.
The degree of line broadening may be described
by the following: l/To = l/T0 + P ^P,^w^(t + t, ) where
z z a b a b
I/T2 and 1/T2 are the respective full widths at half
height for the signals observed during and in the
absence of exchange, P^ and P^ are the relative
populations of the two species, and w is the frequency
separation (Blomberg et al. 1976). At pH 4, the value
for P^, the population of the free base form, is so
small as to eliminate the entire second term on the
right side of the expression leaving the lineshape
unaltered by exchange.
An additional factor of 9.87 must be included
with population factors when averaging involves


57
protonated nitrogens because of the peak magnification
by INEPT. The major difference in comparing this with
other systems involving chemical exchange is this
consequence that non-protonated nitrogens give no
signal due to the bias provided by selective INEPT
enhancement. Nonprotonated nitrogens, due to their low
population and insensitivity, are making no
contribution to any averaging in this pH range.
It is definitive that the lineshape of the ^N
peak would be broadened considerably if the enhancement
vs. pH disappearance was due to the geometric loss of
peak height through broadening. This, would be expected
to be more noticeable as the height decreased. Without
exception, all the ^N peaks of the enhancement vs. pH
spectra exhibited very small, completely random changes
in lineshape and chemical shift attributable to
experimental variation. These small changes showed no
variation with peak height. Suppression of enhancement
is occurring before the rate becomes rapid enough to
affect the ^N linewidth by conventional means through
the N-H coupling interaction or changing chemical shift
even if these conventional phenomena were not precluded
by the low base concentration.
Additionally, if. the signal suppression were
due to line broadening, the peak would be expected to


58
reappear at the fast exchange limit as found in other
studies (Irving, Lapidot 1975). There is no reason to
doubt that the exchange rate would continue to increase
with pH until this limit is reached. This would likely
occur before significant equilibrium reduction of
protonated quinuclidine. The fact that the signal does
not reappear can be attributed to the idea that loss of
polarization transfer was responsible for signal
disappearance in the first place.
Several factors have potential for affecting
INEPT peaks. These can be placed in order of
importance with increasing pH: N-H contact time
decreased to INEPT timescale, loss of J resolution,
broadening from change in chemical shift, reduction in
protonated fraction. Loss of polarization transfer
occurs some 6 pH units lower than significant
deprotonation.
It is postulated that as pH is increased in the
particular systems under study, the dynamics of
exchange affect the efficiency of the INEPT enhancement
before it would affect peaks in the conventional ways
discussed. The fact that enhancement is lost without
any line broadening in itself implies that exchange had
not yet reached the timescale of the coupling or
chemical shift when it affected INEPT. INEPT N-H


59
polarization transfer in typical opiate-like compounds
is more sensitive to proton exchange dynamics than are
the lineshapes.
Further support that the enhancement vs. pH
observations made are due to dynamics was obtained by
the results from altering the exchange rates by
changing the temperature of the solution.
Temperature effects. With very few exceptions
the rate of reaction increases, often very sharply,
with increases in temperature. A conservative rule of
thumb is that rates approximately double for each ten
degree rise in temperature (Moncrief, Jones 1977,
p.230). If the enhancement suppression of concern is
related to rates of proton abstraction, then
enhancement vs. pH spectra would be expected to be
markedly affected by temperature changes.
Enhancement vs. pH spectra were obtained for
quinuclidine as before, at two temperatures, 310K and
2 90 K, at pH's 1.0, 2.2, 3.5 and 4.1. The spectra of
the two experiments are shown in Appendix 1.22 & 1.23,
with both combined to a single enhancement vs. pH plot
for comparison in Appendix 1.24. All previous spectra
were collected at the ambient temperature of the probe,
which due to power dissipated from decoupling, is


60
maintained at 310K. This experiment in effect
attempts to reproduce the previous enhancement vs. pH
results while determining the effect of lowering the
temperature. Changes in linewidth or chemical shift
between temperatures were unobserved. No change in
the measured pH was seen.
At pH 1.0 and 2.2, peaks were of equal
intensity at both temperatures, within experimental
error. The enhancements were also of equal intensity
from one pH to the other. At pH 3.5, temperature began
to have an effect on the magnitude of enhancement. At
that pH, enhancements at both temperatures were
suppressed-, the peak at the higher temperature being
affected more. At pH 4.1, the lower temperature peak
was suppressed an additional amount from the previous
pH, while the peak at the higher temperature was
affected significantly more. A third spectrum
collected at pH 4.1 at.a temperature of 330K showed a
complete absence of enhancement. Enhancements at the
three temperatures: 290 310 and 330 K at pH 4.1
appear in Appendix 1.2.5, with a reference indicating
maximum enhancement for comparison.
When nuclei of spin 1/2 in thermal equilibrium
are placed in an external magnetic field, the spins
align in two possible orientations, with and against


61
the field. These orientations correspond to the
possible energy states, the one with the external field
being of lower energy. There will be an excess in the
lower energy state, the magnitude of which is
proportional to the energy difference and the
temperature according to the Boltzmann distribution
law: N /N =e(-hv/kT), where N & N, are the
a b a b
populations.of the lower and upper energy states, hv is
the. energy between the states, k is the Boltzmann
constant and T is absolute temperature. It is this
excess in the lower energy state that NMR observes. At
nitrogen resonance, the result.of changing the
temperature from 310K to 290K results in a change in
N /Nfa by a factor of 10 warranting no further
consideration.
The nitrate nitrogen of a labeled Na^NO^ was
observed by conventional NMR at temperatures 310 K
and 290K to observe any inherent instrumental factors
which change with temperature that might affect signal
intensities. These spectra, seen in Appendix 1.26,
show no obvious change in signal intensity at these two
different temperatures.
Equi 1 ibrium 'constants are known to be
temperature dependent. This dependence can be
quantified via the enthalpy change of the reaction


62
according to the Gibbs-Helmholz equation: d(lnK)/dT=
AH/RT The value of K for an endothermic reaction
increases with temperature and decreases with
increasing temperature for an exothermic reaction.
Assuming AC^ to be zero, (Maskill 1985, p.143), then
AH will be constant and the equation can be integrated
giving the change in InK with temperature as
AH/R(1/T -1/T). According to best estimates of AH for
a reaction of type (1) (Bell 1973, p.76), a 20K change
in temperature would change the value of K by 0.5 units
at the very most. The effect of temperature on
equilibrium populations is insignificant at the pH
range far removed from the pK&.
Molecules require a certain excess energy, the
energy of activation, in order to react. Activated and
unactivated molecules are in equilibrium. The rapid
increase in k with temperature is primarily due to the
greater fraction of collisions with energy sufficient
for reaction assuming E does not change appreciably
with temperature or pH.
The temperature dependence of the- INEPT
enhancements supports the conclusion reached earlier
that the reaction of type (1), which has a rate
consistent with activation control, is predominant in
this system over (2), (3) and (4), which are diffusion
controlled .


63
The relation between ,the rate constant k and
the temperature was first proposed by Arrhenius:
E /RT
k=Ae , where E is the activation energy and A is an
experimentally determined pre-exponential factor
(Castellan 1971, p.746). It is noted that the increase
in rate with temperature is exponential.. The larger
the activation energy of a reaction, the stronger its
temperature dependence.
Using the accepted rule of thumb and the
results obtained by the investigation of proton
abstraction involving trimethylammonium ion via a
reaction of type (1) by Grunwald (1963), a three-fold
change in rate can be conservatively assumed from the
310K to 290K temperature change. This factor, will be
used in evaluating the observed temperature effects on
enhancement...
From Appendix 1.24, it can be seen that at pH
2.2, temperature had little or no effect on
enhancement. The proton linewidth was measured as 18
Hz from pH-0.8-2.8 (Appendix 1.16-1.18). Since the
proton spectra show a consistent linewidth on either
side of this pH, the exchange rate is slow on the
proton time scale at pH 2.2. Rates, therefore, cannot
be determined by lineshape at this pH, because exchange
is too slow to affect lineshape. All that can be


64
inferred is that exchange is slow compared with the
observed resolution of 18 Hz, corresponding to a mean
lifetime >_ 0.018 sec. The lifetime is probably much
greater since enhancements are maximum here and the
mean value must exceed the INEPT requirement of 0.0093
sec. by an amount such that a negligible fraction of
the N-H bond lifetimes are <0.093 sec. The exchange
rate with cooling would make this time even longer, so
no change was expected or observed.
At pH 3.6, a mean lifetime of 0.014 sec. was
calculated from the proton spectrum at 310K. For the
moment this will be assumed to be the lifetime at pH
3.5 also. Lowering the temperature to 290K would make
the mean lifetime on the order of 0.042 sec. This is
longer than the required 0.0093 sec. of the INEPT
delays, but since this represents a mean value, a
finite fraction of the reactions will still involve
times <0.0093 sec.
The value for t at 310K and pH 3.5 of 0.014
sec. and the assumed tripling of this time at 290K to
0.042 sec. are both greater than 0.0093 sec., meaning
that at least partial enhancements are expected at both
temperatures. From Appendix 1.24, it is seen that at
pH 3.5 enhancement is affected at both temperatures,
although less so at the lower temperature. This


J
65
observation is consistent with the expected fraction of
reactions with ,t>0.0093 sec. at the two temperatures.
The reaction with a mean t value of 0.042 sec. would
have a smaller fraction at >0.0093 sec., hence, larger,
enhancement, than the 0.014 sec. reaction.
Since the enhancement at pH 3.5 and 290K is
somewhat less than the maximum, with the tripling
assumption putting the t at 0.042 sec., the value of t
for pH<2.8 is indeed substantially greater than 0.018
sec. and is actually >0.042 sec.
At pH 4.1 exchange rates are faster than at pH
3.6 and subsequently, enhancements are less. The
observation from Appendix 1.24 is that the effect of pH
on enhancement is less at the lower temperature. This
is expected from the exponential nature of
temperature's effect on rates. Since enhancement is
now being affected at both temperatures, the result of
changing the temperature will be more apparent.
A proton spectrum was not collected near enough
to pH 4.1 to obtain a reliable t value. For the sake
of arguement, assume a linewidth of 70 Hz can be
interpolated- from the 40 Hz at pH 3.6 and 188 Hz at pH
5.3. This gives a t value of 0.006 sec. at 310K and a
value of 0.018 sec at 290K from the rate tripling
assumption. A small fraction of the reactions at 310K


66
will take >0.0093., but the majority of the N-H bonds
will have lifetimes <0.0093 sec. with an overall mean
of 0.006 sec. The enhancement at pH 4.1 and 310K is
about 20% of maximum, which would agree with the
statement that 20% of the N-H bonds have lifetimes
>0.0093sec. when the mean t is 0.006 sec., not at all
unlikely.
At 290K the mean t was reasoned as 0.018 sec.,
less than the mean t of 0.042 sec. at pH 3.5 consistent
with its smaller peak. More than half the N-H bonds
woul4 have lifetimes >0.0093 sec. but a smaller
fraction than would have at pH 3.5. The enhancement at
pH 4.1 is about two-thirds maximum and less than at pH
3.5 and 290 K .
Examination of Appendix 1.24 shows the
enhancements at pH 3.5, 310K and pH 4.1, 290 K to be
approximately the same, implying equal exchange rates.
The t value for pH 4.1, 290K was estimated as 0.018
sec. The t value was measured as 0.014 sec. from the
proton spectrum at pH 3.6. This number would be
slightly greater at pH 3.5. The agreement is
remarkably close considering one came from a proton
spectrum and the other from a complex ^N spectrum.
The 0.014 sec. is a measured value. The value 0.1 pH
unit higher was estimated to be slightly higher,


67
possibly approaching 0.018 sec. The value it is
supposed to agree with was taken as 0.018 sec. This
value resulted from the tripling assumption of the
value of 0.006 sec., which itself was assumed from
interpolation of a 70 Hz proton linewidth.
The lifetime at a temperature of 330 K is
roughly calculated to be one third of the 0.006 sec. at
310K or, 0.002 sec. Appendix 1.25 shows no observable
enhancement at 330K and pH 4.1. The 0.002 sec. was
also estimated for 310 K and pH 5 from calculations
performed before with the proton spectra. No
enhancement has been observed for quinuclidine at pH 5,
meaning that a small fraction of N-H bonds last longer
than 0.0093 sec.
Results of changing the temperature of the
solution while observing the effect on INEPT
enhancements supports the notion that INEPT is
sensitive to exchange dynamics in the 2-5 pH range.
This was reasoned to be due to temperature's effect on
kinetics and not thermodynamics or instrumental
conditions.
Effects of altering INEPT delays. Another
method of altering the dynamics involves an indirect
approach by changing the sensitivity of polarization


68
transfer to N-H contact times. The delays incorporated
into the INEPT sequence for the case of nitrogen
coupled to a single proton has been explained
previously to be a function of n/4J, where n=l,3,...
The periodic nature of the delay times is easily
verified by inspection of the vector description
presented and is graphically represented in Figure 10.
There are three such delays of duration 0.0031 sec. in
a decoupled INEPT experiment of quinuclidine. If these
delays are increased by a factor of three, the required
N-H bond contact time for polarization transfer would
become 0.028 sec.
Spectra were collected using delays of 0.0093
sec. instead of 0.0031 sec. during the course of the
temperature study previously. The temperature used was
310K. The enhancements at pH's 1.0, 2.2, 3.5 and 4.1
with these longer delays is shown in Appendix 1.27,
with a reference indicating maximum enhancement for
comparison. The immediate observation is that
enhancements that depend on longer delays are more
sensitive to the increases in rate with pH.
The mean t value at pH 3.6 was said to be 0.014
sec. Compared with the required residence time of
0.028 sec., a small peak would be expected which is
what was observed. At pH 2.8 and below, with no


69
broadening, it was determined only that the t value is
0.042 sec. or greater. Consistent with previous
arguements based on activation theory, a maximum peak
would require that an overwhelming majority of the N-H
bonds last more than 0.028 sec. This would require a
mean t value much greater than 0.042 sec. Since the
peak at pH 1.0 is about equal to that at pH 2.2, the
rate probably did not change much between those pH's.
The assumption that times of 3/4J and 9/4J are equally
insignificant compared to relaxation still holds.
The unknown value of t at pH<^2.8 is >0.0093 and
also >0.028 sec. At the low pH limit, t is implied to
be sufficiently greater than 0.0093 for maximum
enhancements but not compared to 0.028 sec. Since
enhancements are the same between the two lower pH's
using the longer delay, rate probably lines out at this
pH region in agreement with Figure 17. showing proton
exchange rates as a function of pH.
The mean t value passes through 0.028 sec.
somewhere between pH 2.2 and 3.5 according to Appendix
1.27. At that point, the enhancement would be half
maximum, as half the lifetimes would be more than 0.028
sec. and half less.
The three spectra, 290K, 310K and the longer
delay at 310K, at each of four pH's are shown in


70
Appendix 1.28. It would appear that increasing pH from
1.0 to 2.2 does not change rates much. The fact that
lowering the temperature has no effect implies that
enhancement is not dynamics dependent at 310+20K here.
As exchange rates increase with pH,
enhancements decrease. This was much less radical at
the lower temperature. The effect of changing
temperature is much greater as pH increases. Lowering
the temperature decreases rates, which, as pH is
increased, delays the rate effect on enhancement,
effectively shifting the enhancement vs. pH plot to the
right along the pH axis. It is not difficult to
imagine supercooling a system to the point where the
enhancement vs. pH plot would represent protonation vs.
pH and not dynamics vs. pH. At that point, further
temperature decreases would have no effect on the plot.
The phenomenon of proton exchange dynamics was
explored from several different approaches. A number
of quantitative deductions were made to varying degrees
of accuracy, some based on measurement, others were
derived from these measured values and extended to
conditions not amenable to measurement using fairly
reliable assumptions.. Several consistencies between
these separate approaches have become apparent,
validating the methods and giving support to the


71
assumptions made in the process of estimating some of
the data.
The purpose of the preceeding rate experiments
was to characterize ^N INEPT enhancement of
opiate-type compounds as a function of pH in a manner
commensurate with the ultimate goal. For this reason,
additional temperature runs were not made. Enough
quantitative data was obtained in the process, however,
to permit a rough rate vs. pH curve for quinuclidine
(Figure 18.).
It is concluded that the loss of INEPT
enhancements at pH levels giving full protonation is
due to. proton exchange dynamics. The N-H contact time
for a given proton passes through' the INEPT time scale
at these pH levels. To achieve the ultimate goal, a
proton (if it exists) on the nitrogen of an opiate must
give an INEPT signal when the opiate is bound to the
receptor. At physiological pH, this signal would not
be observable in aqueous solution. However, the
environment of the N-(H) part of the opiate is much
different in the receptor cavity than, in aqueous
solution, surely altering the dynamics responsible for
loss of enhancement. INEPT enhancement will now be
evaluated in different solvents to establish if the act
of altering this environment by itself can allow an
INEPT signal to be observed.


72
Figure 18. N-H bond lifetime vs. pH.
A. The curve must pass into pH<2.8 region below
this point.
B. 0.014 sec., pH 3.6, calculated from (a).
C. pH 3.6, 290, tripling assumption from B.
D. Interpolation of proton spectra to a 50 Hz
linewidth.
E. Interpolation of proton spectra to.a 70 Hz
linewidth.
F. Tripling assumption from E.
G. Coalescence of proton spectra.
H. INEPT peaks at pH 3.5,.. 310=4.1, 290 from F.
I. From half enhancement of delay spectra.
J. Estimated from where proton linewidth would be
177 Hz calculated from (a) using 0.002 sec. at
NH coalescence.


Solvent Effects
73
To explore the effect of solvent identity on
proton exchange rates, two nonaqueous solvent were
utilized, acetic acid and dimethyl sulfoxide. The
choices were limited by solubility considerations;
participation in proton exchange with the solute forms
a large part of the solvation process. Some pertinent
concepts relating to the solvation phenomenon are
discussed.
Digression to general solvent properties. The
effect of the identity of the solvent on reaction rates
is governed by the solvation process, the interactions
of solute with the solvent molecules. This is in turn
determined by many of the solvent's properties,
including: structure, dielectric constant,
polarizability, dipole moment, viscosity, acidity or
basicity and hydrogen-bonding capability. The overall
solvent effect consists of the weighted contributions
of all the properties which affect solvation.
Interpretations based on these types of properties are
almost always complicated by the simultaneous
participation of several counteracting factors.
The influence of the solvent on reaction rate
depends on the degree of solvation of the reactants and
activated complex. In terms of solvation phenomena,


one must consider the difference between the solvation
free energies of the reactants and the transition
state, as the change in rate with medium is a
consequence of the differences in solvation free
energies (Popovych, Tompkins 1981). For a reaction
which solvates the reactants to a higher degree than
the intermediate complex or product, the reaction will
take place more slowly than in a solvent which solvates
the reactants less.
The dielectric constant of a solvent measures
its ability to separate opposite charges. A reaction
in which the products have more charges than the
reactants is favored by a solvent with a high
dielectric constant. The reaction between R^I^H and
l^O, as in (1), was previously reasoned to be the
primary mechanism of deprotonation under the conditions
of interest. The position of an equilibrium involving
acids that are positively charged to begin with is not
particularly sensitive to the dielectric constant of
the solvent, because when they transfer a proton to a
neutral solvent molecule, no separation of charge
occurs.
Changing the solvent will usually result in a
change in the dielectric constant of the solution,
among other things. Changing the dielectric constant


75
of the solvent by itself is not expected to have much
influence on the position of this equilibrium,
depending on subsequent changes in charge localization.
According to Bronsted's concepts, a proton is
only given up by an acid in the presence of a base that
can accept the proton. Hydrogen chloride dissolved in
benzene does not dissociate into ions but does so
readily in water. The extent of ionization and the
dynamics involved depends not only on the substance of
interest, but also on the nature of the solvent.
Water is an amphiprotic solvent, that is, it
can function as either an acid or a base, and undergoes
the following autoprotolytic reaction:
H20 + H20 $. H30+ + OH" (5)
Acetic acid. Water-free acetic acid is a far
less basic solvent than water, its tendency to accept a
proton is smaller. Acetic acid has been used to
differentiate the strengths of acids considered to be
equally strong in water. Nitric acid and perchloric
acid are both considered strong in water, but in an
acetic acid solvent, the ionization of nitric acid is
incomplete because of the acidity of the solvent. The
strongest base that exists in an acetic acid solution


76
is the acetate ion, just as the hydroxide ion is in an
aqueous solution.
Acetic acid has an autoprotolysis constant
similar to water, (pK =14.45). In contrast to water,
s
however, acetic acid does not have a matched acidity
and basicity (Popovych, Tomkins 1981, p.49). The very
low basicity of acetic acid means its Kg value must be
dominated by the high proton-donating tendency, (6).
The acidic properties of acetic acid solvent are
sufficient to cause bases of medium strength to react
more or less completely with the solvent:
B + HOAc S [B+H 0Ac~] % B+H + OAc (6)
HA + HOAc 5- [H2OAc+ A] S H20Ac+ + A- (7)
with the ion pair only partially dissociated because of
the low dielectric constant.
The completeness of the reaction is a function
of the dissociation constant of the solute species and
the autoprotolysis constant of the solvent. The
equilibrium position is the outcome of the competition
for protons. The difference in that position for
different solvents may imply different rates of proton
abstraction from the solute nitrogen. Acidic solvents,
such as acetic acid, magnify the basic properties of


77
solutes. Consequently, acids are made weaker in acetic
acid than they are in.water. Since water is a stronger
base than acetic acid, reaction (8) has a higher
equilibrium constant than reaction (9). As K is a
ratio of rates, the forward rate of (8) will likely be
greater than that of (9).
R3N+H + H2 $ R3N + H30+ (8)
r3n+h + HOAc 5 r3n + H20Ac+ (9)
The conjugate acid of HOAc,. (H2OAc+), is
stronger than H^0+, and will protonate R^N to a greater
extent than will.
A solution of quinuclidine hydrochloride in
glacial acetic acid was prepared at a concentration
similar to that used for previously described
experiments. Sufficient acetic anhydride was present
to ensure anhydrous conditions. The resulting INEPT
NMR spectrum of this example of (9) gave a large peak
(Appendix 1.29). No addition of acid or base or other
reagents was performed.
The INEPT spectrum resulting from reaction (9)
is also shown in Appendix 1.30 along with the spectrum
corresponding to reaction (8). These two spectra
represent quinuclidine hydrochloride in either pure
acetic acid or pure water.


78
Since no reagents were added to these solutions
to purposely adjust the pH, this experiment may be used
to demonstrate the direct effect of solvent identity on
INEPT enhancements, by that particular solvent's
ability to affect rates through its tendency to accept
protons. The result obtained shows that N-H contact
time, or, presence of INEPT enhancement, is potentially
sensitive to the nature of the entity with which the
N-H interacts. Additional experiments were performed
to evaluate this notion using a solvent that is more
different in the way it solvates the solute and less
active in its participation in the acid-base
equilibrium.
Dimethyl sulfoxide. The most significant
feature that distinguishes amphiprotic from aprotic
solvents is hydrogen bonding capability. Amphiprotic
liquids are characterized by their ability to
self-associate via hydrogen bonding and to solvate
anions and protophilic uncharged species by donating
hydrogen bonds to them to a greater extent than aprotic
solvents.
DMSO is an example of a solvent which does not
have the capability of donating strong hydrogen bonds
to' solute species but has a. high enough dielectric
constant, (46.7), (Popovych, Tomkins 1981, p.35) to
allow studies of ionic solutions.


79
INEPT spectra v/ere collected using quinu.clidine
hydrochloride in two different solvent systems: pure
H^O and 60% dimethyl sulphoxide (DMSO) in H^O. The pH
was adjusted to the same value (4.0) in each case (see
EXPERIMENTAL). The DMSO solution gave a much larger
INEPT enhancement than the water solution as seen in
Appendix 1.31. The same experiment was conducted using
tropine instead of quinuclidine with the same
qualitative result (Appendix 1.32). These results can
be explained in terms of solvation of reactants vs.
products.
The water molecule can act simultaneously as a
hydrogen ion donor and acceptor, so the free base form
of tertiary nitrogen compounds, as well as dissociated
HC1, are efficiently solvated in water. While both
DMSO & water H-bond cation acids, only water as a
H-bond donor can also H-bond the undissociated base.
This latter effect tends to increase acid strength of
B+H in water. DMSO is a poorer H-bond donor because it
lacks an active hydrogen and the positive end of its
dipole is buried in the molecule.
In DMSO, a protonated amine would have a lesser
tendency to lose a proton since the free base would be
less stable than the protonated form. The difference
in solvation energy between the protonated form and the


80
free base is greater in DMSO than in water.
Consequently, the pK of tertiary nitrogen compounds
would be expected to increase on going from water to
DMSO since the free base would be stabilized by
H-bonding.more in water. It has been previously
established that hydrogen bonding of solvent to the
conjugate base, would decrease acidity in DMSO
(Ritchie, Uschold 1967). The greater basicity of DMSO
would tend to counteract the observed results, however.
The pK^ for the cation acid triethylaraine is
10.7 in water vs. 9.0 in DMSO (Kolthoff et al. 1968).
This was reasoned due to the greater basicity of DMSO
over water. The effect of the presence of hydrogen
bonding of the free base in water vs. the lack thereof
in DMSO was implied to increase the pK on changing the
solvent from water to DMSO. Published values for the
pK of quinuclidine in DMSO were not found, although
some trends were.. The extent to which the pK values
are larger in water becomes less as the number and size
of alkyl groups increases with the pK becoming larger
in DMSO for ammonia (Kolthoff et al. 1968).
Quinuclidine has been described as a cyclic
tertiary amine with paraffin groups tied back leaving
the nitrogen atom more exposed and more reactive.
There is also steric strain which would be relieved by
protonation. This is supported by the higher aqueous


81
pK^ of quinuclidine (11.45) vs. that of triethylamine
(10.7). This would make the free base even more stable
in water compared with DMSO and make the pKQ lower in
water relative to DMSO. This fact would tend to
counteract the tendency for the decrease in pK in DMSO
due to solvent basicity.
For a reaction which solvates the reactants to
a higher degree than an intermediate complex or
product, the reaction will take place more slowly than
in a solvent which does not solvate the reactants.
Water solvates small ions with high, charge density
better than DMSO. This would favor dissociation of the
protonated amine in water relative to DMSO.
Validity of pH measurements in DMSO. The
difficulty in making interpretations based on pH
measurements in nonaqueous solvents cannot be
overstated. Nevertheless, experimental studies of the
glass electrode affirm the hydrogen ion response in
many media that contain at least a few percent water.
Practical measurements of acidity are then possible
with the assurance that the liquid-junction potential
between the nonaqueous solution and the aqueous salt
bridge does not change appreciably as the acidity of
the solution varies over a considerable range of
hydrogen ion activity (Bates 1969, p.81,2).


82
Transfer activity coefficients of the proton
are unique in that their values in a given solvent can
be used to correlate the scales in that solvent and
in water and to refer the emf series in that solvent to
the zero point of the standard hydrogen electrode in
water. The conventional aqueous p^ scale extends from
0 to p K g or, 0 to 14 for water, and from -logy to
(pKs~logy), where 7 is the transfer activity
coefficient from water to DMSO, for solvents on the
aqueous scale which is from 3.3 to 36.6 for DMSO. The
change in activity coefficient of the hydrogen ion in
DMSO means that a solution of a given pajj in DMSO would
have a higher pg^ (be more basic) on the aqueous pg^
scale by 3.3 units (Popovych 1981, p.193,4).
A quantitative correlation between pH
measurements in water and in DMSO is not meant to be
implied, only that the effect of this consideration
would tend to be in a direction of less enhancement in
DMSO instead of more. The correlation made in the last
paragraph is intended to be qualitative only in that
the DMSO solution is more basic than a solution of same
pH in water. The presence of an INEPT enhancement in
DMSO at a pH where none is found in water may have a
partial contribution from a greater basicity in DMSO at
a given measured pH. It was already mentioned that the


83
activity of the hydrogen ion is greater in DMSO than in
water. This would decrease the measured pH of a DMSO
solution of a given basicity. Measured pH would tend
to be lower in that solvent for a given [H+].
Use of neat DMSO. An experiment was conducted
in a manner similar to the previous solvent
investigation using NH^Cl in pure water vs. pure DMSO
with no pH adjustments. No peak was seen in the H2O
solution but there was an enhancement in the pure DMSO
solution (Appendix 1.33). This shows an influence of
the solvent type without introducing the complications
of pH measurements.
The preceeding set of experiments determined
that the nature of the solvent can influence the
dynamics of proton exchange and that this alteration
under certain circumstances is sufficient to affect the
presence or absence of INEPT enhancements. In the
preceeding systems exchange was still taking place, but
at a slower rate.
This experimental strategy continued with
attempts to reproduce more closely the environment of
the opiate nitrogen when bound to the receptor cavity.
The proton on the nitrogen (if it exists) will be
physically constrained when the opiate is bound and,
although it may move back and forth between the


84
nitrogen and the interacting site, it cannot diffuse
away. The first study of this type will utilize the
cation complexing cavity of crown ethers.
Crown ethers. Certain members of a particular
class of compounds called crown ethers possess the
ability to act as solubilizing agents in organic
solvents for many ionic salts which are otherwise
insoluble. A complex is formed as a result of
ion-dipole interactions between the cation and negative
charge associated with oxygen atoms in the ring (Figure
19). This unique cation complexing property of crown
ethers was sought as a model to mimic the receptor
cavity in an attempt to study in-cavity dynamics of
ligands.
-c/
0 C+ 0
\^-0 0-^
w
Figure 19. Crown ether-cavity/cation complex
The 18-crown-6 ether is known to complex the
ammonium ion and the dimensions, 2.84 angstroms for the
N+H^ ion and 4 angstroms for the hole of 18-6, lend
support that this should indeed occur (Pedersen 1970).


85
A series of experiments was performed using the
18-crown-6 ether in aqueous solutions. Various
concentrations of ammonium chloride in this solution
were examined with INEPT. The results are shown in
Appendix 1.34. An initial solution of ammonium
chloride in water with no crown gave no enhancement,
due to the rate of proton exchange with the solvent.
Introduction of 18-crown-6 ether to a 1:2 molar ratio
of N+H^:crown resulted in an enhancement. Increasing
the nitrogen concentration to a 2:2 molar ratio
resulted in a peak twice as intense. Further increases
in the ammonium fraction to 3:2 and 4:2 molar ratio did
not cause further increases in INEPT enhancements.
The addition of crown ether complexed the
ammonium ion, immobilizing to an extent the protons on
the nitrogen, resulting in N-H contact times that allow
INEPT enhancements. The mere addition of the
complexing crown enabled enhancements to occur. The
fact that the enhancements increased no further as the
ammonium concentration was increased to 3:2 and 4:2
molar ratios would tend to rule out viscosity as the
mediator of the change in dynamics.
The tendency for an ion to associate with the
solvent greatly affects complex stability. Ammonium
chloride is strongly solvated in water, so the ether
must compete with surrounding solvent for cation.


86
In the solutions described here, an additional
equilibrium is introduced due to the presence of crown,
the complexation reaction, (10). The protonated cation
can now enter into two possible equilibria, proton loss
to solvent and complexation with the crown. The value
of log K for the complexation equilibrium of N+H^ with
18-6 is 1.1 at 25 C where K=[Complex]/[N+H^][18-6]
(Frensdorff 1971).
N+H4 + 18-6 +i- Complex (10)
At equilibrium, K=12.59 for (10), and the
concentration of complexed ammonium ion will be
approximately 80% of the total ammonium in the
solution. For simple amines it was determined
previously that at the low pH limit, full INEPT
enhancement is attained in the sense that essentially
all nitrogen nuclei are protonated for a sufficient
amount of time to allow polarization transfer such that
lowering the pH further has no effect. An INEPT
spectrum of a solution of NH^Cl of identical
concentration to the 2:2 solution before was taken at
the low pH limit. The peak of the complexed ammonium
ion was approximately 82% as intense as the peak from
this solution as seen in Appendix 1.35, consistent with
this equilibrium hypothesis.


87
When the cation is complexed, it cannot undergo
proton exchange. INEPT enhancement can be seen as a
result of this proton immobilization. The cation crown
complexation may in some ways be viewed as analogous to
the opiate-receptor interaction. To the extent that
this analogy is a correct description of the
opiate-receptor complex, it would appear that physical
confinement of a protonated nitrogen from the solvent
can potentially immobilize the proton to allow INEPT
enhancements.
Due to the nature of the oxygen atoms of the
ether, this does not address the possibility of proton
transfer to a reactive entity within the receptor
cavity. It is possible that a reactive carboxylate or
sulfate oxygen may be involved in a possible
interaction with protonated nitrogen. The hypothesis
of Smythies (1977) concerning the structure of the
opiate receptor includes the detail that part of the
interaction involves a protonated nitrogen forming an
ion pair with a carboxylate oxygen. This type of
interaction would differ from the electrostatic
attraction to the lone pairs of the oxygen atoms of
crown ethers, introducing the chance for a more active
participation of the proton in the interaction and may
have an effect on INEPT not encountered in the crown


88
complex analog. Examination of INEPT enhancements of
quinuclidine in solvents that promote ion pair
formation involving the nitrogen were conducted to
evaluate this situation.
Ion-pairing. Reactions of organic acids and
bases in non-dissociative solvents produce ion pair
products. Reactivity is reduced because of the strong
attractive force exerted toward each other.
Experiments with quinuclidine free base in acetic acid
(HAc) and trifluoroacetic acid (TFA) were performed.
Both these solvents would tend to protonate the
nitrogen and inhibit dissociation of the ions formed.
Digression to a discussion of ion pairing.
Random thermal motions of ions in solutions of
electrolytes constantly bring cations and anions into
proximity, occasionally to contact distances. Such
pairs remain together until sufficiently large thermal
fluctuations send the ions apart. Dissolution of a
reasonably soluble salt as free ions requires
sufficient solvation energy to almost completely offset
the lattice energy. Dissolution as ion pairs requires
solvation energy only approximately equivalent to the
sublimation energy (Gordon 1975, p.371). Thus poor
solvation promotes ion association. Ion-pairing is


89
more prevalent as the dielectric constant of the
solvent decreases.
The ability of a solvent to separate charges is
measured by its dielectric constant. Through its
dielectric constant, the solvent will influence the
dipole-dipole interactions between reactants and their
ability to make contact. For a given ion, the
solvation energy decreases with an increase in
dielectric constant of the solvent. In a solvent of
low dielectric constant, a greater amount of energy is
required to separate a positively charged particle from
one with negative charge than in a solvent with a high
dielectric constant.
For quaternary ammonium salts, the logarithm of
the association constant was found to be linear
according to the inverse of the dielectric constant.
For a solvent with a dielectric constant of 10, log K
was approximately 4 (Hammett 1970, p.224). The degree
of ionization will vary with the stability of the ion
pair .
The term ion pair implies ions of opposite
charge which are in contact retaining essentially the
same internal electronic structure as the separated
ions. Two oppositely charged bodies having the
properties of real ions cannot be brought close