K-vEPA
United States
Environmental Protection
Agency
Health Effects Research
Laboratory
Research Triangle Park NC 27711
EPA 600 1-79-035
September 1979
Research and Development
Comprehensive
Progress Report
for Fourier
Transform NMR
of Metals of
Environmental
Significance
EP 600/1
79-035 k
EDISOM, H.J. 08317 • '
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are.
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4 Environmental Monitoring
5. Soctoeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL HEALTH EFFECTS RE-
SEARCH series This series describes projects and studies relating to the toler-
ances of man for unhealthful substances or conditions. This work is generally
assessed from a medical viewpoint, including physiological or psychological
studies. In addition to toxicology and other medical specialities, study areas in-
clude biomedical instrumentation and health research techniques utilizing ani-
mals — but always with intended application to human health measures
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
EPA-600/1-79-035
September 1979
Comprehensive Progress Report
for
Fourier Transform NMR of Metals of Environmental Significance
by
Paul D. Ellis and Jerome D. Odom
Department of Chemistry
University of South Carolina
Columbia, South Carolina 29208
Grant R804359
Project Officer
Nancy K. Wilson
Environmental Toxicology Division
Health Effects Research Laboratory
Research Triangle Park, North Carolina 27711
LIBRARY ^
PROTBCWOIAQBRCI
-------
Disclaimer
This report has been reviewed by the Health Effects Research Laboratory,
U.S. Environmental Protection Agency, and approved for publication . Ap-
proval does not signify that the contents necessarily reflect the views and
policies of the U.S. Environmental Protection Agency, nor does mention of
trade names or commercial products constitute endorsement or recommendation
for use.
-------
Forward
The many benefits of our modern, developing, industrial society
are accompanied by certain hazards. Careful assessment of the relative
risk of existing and new man-made environmental hazards is necessary
for the establishment of sound regulatory policy. These regulations
serve to enhance the quality of our environment in order to promote
the public health and welfare and the productive capacity of our
Nation's population.
The Health Effects Research Laboratory, Research Triangle Park
conducts a coordinated environmental health research program in
toxicology, epidemiology, and clinical studies using human volunteer
subjects. These studies address problems in air pollution, non-
ionizing radiation, environmental carcinogenesis and the toxicology
of pesticides as well as other chemical pollutants. The Laboratory
develops and revises air quality criteria documents on pollutants
for which national ambient air quality standards exist or are proposed,
provides the data for registration of new pesticides or proposed
suspension of those already in use, conducts research on hazardous
and toxic materials, and is preparing the health basis for non-ionizing
radiation standards. Direct support to the regulatory function of
the Agency is provided in the form of expert testimony and preparation
of affidavits as well as expert advice to the Administrator to assure
the adequacy of health care and surveillance of persons having suffered
imminent and substantial endangerment of their health.
This report presents results of extensive studies of the inter-
actions of the toxic metals selenium and cadmium with biological sys-
tems, particularly metalloproteins. Improved analytical methods for
examination of these interactions, using nuclear magnetic resonance
spectroscopy, were developed and are described in this report.
F. Gordon Hueter
Director
Health Effects Research Laboratory
ill
-------
Abstract
Metals or metalloids of primary importance to this study are cadmium and
selenium. These nuclei are ideally suited for a multinuclear magnetic resonance
study involving direct observation of the metal nucleus by Fourier transform
techniques because of their respective receptiveness to an nmr experiment and
because they are spin \ nuclei. The principal objectives of this project included:
i) the study of the interactions of cadmium with various amino acids, ii) a
detailed examination of the nmr parameters of cadmium substituted metallo-
proteins, iii) the development of a systematic understanding of the relaxation
behavior of 77Se, and iv) the development of the necessary nmr technology
to bring these objectives to fruition on a reasonable time scale per experiment.
Our work with the cadmium/glutathione system indicate that the conditions
necessary to observe a 113Cd resonance are such that the results of such a
study would have little relevance to a real biological system. Hence, our
research with 113Cd nmr has oeen largely devoted to a critical examination of
three metalloproteins; concanavalin A, bovine superoxide dismutase, and
carboxypeptidase A. Our work to date indicates that this new spectroscopic
probe, i.e. 113Cd nmr, is providing new and significant information to the
biochemist with respect to the fundamental mode of action of these proteins
and the relationship between the function of the protein and the metal
associated with it.
Furthermore, we have been involved in an intensive effort to find other nmr
active spin \ nuclei that could be exploited in biological applications. Such a
nucleus is 77Se. Although extensive chemical shift information is available
for 77Se, little, if any relaxation time information was available before our
efforts in this area of research. The spin lattice relaxation time is a parameter
of paramount importance in determining the overall utility of a given nucleus
in a biological application, that is, it is intimately related to the signal-to-noise
per unit time. To this end, we have carried out an extensive study on the
nature of the specific mechanism(s) of spin lattice relaxation and the corresponding
values of T!. The systems that we have investigated to date are: organo_selenium
compounds, RSeR1, selenols RSeH, diselenides, RSeSeR', selenates, SeOi* 2,
and selenocysteamine. As a result of this research it is clear that 77Se may
be very useful to studying active site sulfhydryls in sulfhydryl proteins.
In order to pursue these research topics in the most efficient fashion, it
was necessary for us to develop some novel modifications of our nmr instrumenta-
tion. One of these is a unique nmr probe, capable of spinning 18 mm nmr
tubes, decoupling at any frequency, and observing any nmr active nucleus.
Further, the most important design parameter was signal-to-noise ratio per
unit volume. That is, we did not take the "milk-bottle" approach in our system.
But rather, we tried to optimize the signal-to-noise on a 5 ml coil volume. The
net result is that the probe leads to a timesaving of approximately a factor of
twelve over conventional 12 mm nmr systems.
Finally, we have solved a rather severe experimental problem in nmr spec-
troscopy. That is, the required power levels for efficient broad band hetero-
nuclear spin decoupling. This problem reaches critical proportions when examining
-------
biological systems, e.g. proteins in high salt concentrations, on nmr spectro-
meters employing superconducting magnets. The basic approach employs a
linear frequency modulation scheme, i.e. a "Chirp",followed by a 180° -phase
modulation of the "Chirp". The relative rates of "Chirp" to 180° -phase modulation
must be kept in a 4 to 1 ratio for the method to succeed. The net result is
that uniform decoupling can be achieved over a range in excess of 2kHz with
only 2 watts of power. For those systems not employing a Faraday shield,
efficient decoupling can be affected with power levels between 1 and 2 watts.
VI
-------
Table of Contents
Disclaimer ......................... ii
Forword ......................... iii
Abstract ......................... v
Table of Contents ......................... vii
List of Abbreviations and Symbols ......................... ix
Acknowledgement ......................... x
I . Introductory Comments ......................... 1
II. Interactions of Spin 1/2
Metals with Amino Acids ......................... 1
III. Cadmium- 113 NMR of Cadmium
Substituted Metalloproteins ......................... 1
A. Cadmium- 113 NMR Studies
of Cadmium Substituted
Concanavalin A ......................... 1
B. Cadmium-113 NMR Studies
of Cadmium Substituted
Superoxide Dismutase ......................... 7
C. Cadmium-113 NMR Studies
of Carboxypeptidase ......................... 18
IV. Selenium-77 NMR Investigations
of Model Biological Systems ......................... 21
V . Recent Advances in Multinuclear
NMR Techniques .........................
A. Development of a Multinuclear
18 mm Probe ......................... "
B. Utilization of Chirp Frequency
"Modulation with 180°-Phase
Modulation for Heteronuclear
Spin Decoupling ......................... 38
VI . Summary ......................... 52
VII
-------
VII. References 53
VIII. Figures and Figure Captions 63
VIII
-------
LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
Asp — aspartic acid
CD — circular dichroism
Con A — Concanavalin A
CW — continous wave
esr — electron spin resonance
FT — Fourier transform
His — histidine
MCD — Magnetic circular dichroism
nmr — nuclear magnetic resonance
NOE — Nuclear Overhauser Effect
ORD — optical rotary dispersion
ppm — parts per million
SOD — Superoxide Dismutase
SYMBOLS
HjH1 — Hameltonians for the spin system
I — spin I
1,1 ,1 — x,y, of z components of the I spin angular momentum.
J
resudual indirect nuclear spin coupling constant.
rotating frame transformations
spin S
spin lattice relaxation time
transverse relaxation time
Wj,u)g — resonance frequency for spin I or S in radians per second.
90°, 180°-- pulse flip angles
'R
R,R',V-
S
Ti
T2
IX
-------
Acknowledgement
The authors wish to acnowledge the work of Professors James Fee, Dave
Behnke as well as Drs. Bill Dawson, Vladameir Basus and graduate students
Dave Bailey, Allen Cardin, Allen Palmer, and Andy Byrd. The authors are
particularly grateful to Dr. Nancy Wilson of the Health Effects Research Laboratory
for her invaluable help at the early stages of this research.
-------
I. Introductory Comments
Within the scope of EPA Grant R804359 there are four principal objectives.
These objectives include: i) the study of interactions of spin 1/2 metals, e.g.
113Cd, with various amino acids, ii) a detailed examination of the nmr parameters
(chemical shifts, Ti and T2) of cadmium substituted metalloproteins, iii) the
development of a systematic understanding of the magnetic resonance parameters
of those spin 1/2 nuclei which are not well characterized and are of importance
to environmental research problems, e.g. 77Se, and iv) the development of the
necessary nmr technology to bring these objectives to fruition on a resonable
time scale per experiment. This progress report is divided into these four
categories.
II. Interactions of Spin 1/2 Metals With Amino Acids.
The main objective of this project was to examine the relaxation parameters
and chemical shifts of metals in the presence of various amino acids. If these
results were positive, then these same methods would be applied to a study of
metals interacting with hormones. Our initial efforts were directed toward a
study of the 113Cd/glutathione system. Glutathione was chosen because of the
existance of sulfhydryl residues in reduced form and disulfide bonds in the
oxidized form. The resultSof these experiments were very frustrating. That is,
the 113Cd nmr signal disappeared in the presence of the amino aicd. The only
way the resonance could be observed was to raise the sample temperature to
above 60°. Admittedly the system could then be studied, but the results would
have little, if any, relevance to a m vivo biological system. Further, most
of the hormones (which are very expensive) would denature or decompose at
these temperatures. Since it was important for these studies to be applicable
to biological environments we chose to defer this study until some better biological
models were developed.
Recently a short note on the 113Cd nmr of the cadmium/glutathione system
has appeared1. Their results are consistent with our observations.
III. Cadmium-113 NMR of Cadmium Substituted Metalloproteins.
We have examined the 113Cd nmr of three metalloproteins, Concanavalin A,
Carboxypeptidase A, and bovine superoxide dismutase. During the course
of these studies (to be described in detail below) three brief reports on the
113Cd nmr of metalloproteins have appeared 2~\ These studies dealt with
three proteins: Alkaline Phosphatase2' ** and Carbonic Anhydrase B2' 3 from
bovine and human sources. These studies demonstrated the utility of 113Cd
nmr in studying metalloproteins, even though the results of Armitage et al. 2
on carbonic anhydrase B are in direct disagreement with those of Sudmeier and
Bell3.
A. Cadmium-113 NMR Studies of Cadmium Substituted Concanavalin A
This research has been performed in collaboration with Professor David
Behnke of the University of Cincinnati Medical School.
5-7
-------
Concanavalin A (Con A), a lectin isolated from the jack bean, has been
extensively studied due to its unusual biological properties8"10. These pro-
perties include: preferential agglutination of suspended malignant cells as
compared to normal parent cells11, .stimulation of blastogenesis in lymphocytes12,
stimulation of cell mediated immunological responses13, and pronounced effects
on the phenomena of delayed hypersensitivityllf. As variant as the above
list appears, all of these properties seem to depend on the ability of Con A
to bind to cell surfaces, which in turn involves specific interactions between
saccharide binding sites on the cell surface with the Con A moiety15. Because
of this, the fundamental requirements for the saccharide binding activity of
Con A have been an area of intensive research over the last several years.
It has been known for some time that one of the requirements for the
saccharide binding activity of Con A involves certain metals bound to the
protein. The nature and location of these metals have been studied by X-ray
crystallography16"21, esr22, nmr23 3lf and circular dichroism33"35. These
studies reveal that Con A consists of a dimer (pH £6) or tetramer (pH >_7)
composed of identical subunits with a molecular weight of 25,50036'37. Each
subunit requires two metals to bind saccharide. One of the essential metals
binds to a site denoted SI. This site is occupied by Mn(II) in the native protein
but will also accept other transition metals as well3'39. The second site,
denoted S2, contains Ca(II) in the native protein but may also bind Cd(II),
and to a lesser extent Mn(II) 3£f/ 38. It has been proposed that an ion must have
a high affinity for nitrogen ligands to bind at SI and must have an ionic radius
of very nearly lAto bind at S239. Studies involving the removal and replacement
of various metals, both native and non-native reveal that the SI site must be
occupied before the second metal will bind in the S2 site, and that both sites
must be occupied to allow the protein to display saccharide binding activity38'1*0.
Although the location of the saccharide binding site has been the subject of
some controversy in the past37"1*0, it now appears that this site is about
11A distant from the SI metal21.
Recent developments in multinuclear Fourier Transform (FT) nmr techniques1*1""1*5
have given researchers the capability of using a variety of nmr nuclei as
probes to investigate chemical and biological systems. One application of interest
is the use of metal nuclides as probes of metal-protein interactions in metalloproteins
and enzymes. The native metals found in these proteins have, in general,
poor high resoultion nmr characteristics but may in many cases be replaced
by metals with more fo/orable properties. One substitute nuclide with ex-
cellent nmr properties is 113Cd. Several studies have been published that
have investigated the 113Cd nmr spectra of a variety of inorganic and organo-
metallic models systems 39~"1*3. More recently, Armitage and co-workers2'
have investigated the 113Cd nmr of Cd(II) substituted alkaline phosphatase,
human and bovine carbonic anhydrase B. These three metalloproteins are all
similar in that each has basically the same symmetry around the metal site
(four-coordinate or tetrahedral) and for each protein, three of the ligands
binding the metal are nitrogen. Despite these similarities, the reported 113Cd
chemical shifts of these proteins are in a range of over 40 ppm. Sudmeier and
Bell3 have also investigated the 113Cd nmr of Cd(II) human carbonic anhydrase
B with findings that differ considerably from the former work. The difference
in the chemical shifts reported is 80 ppm. In any case, the sensitivity of 113Cd
nmr as a probe metal environment has been clearly demonstrated.
-------
We wish to report here our findings on the 113Cd nmr of 113Cd(II) substituted
Con A. The general areas of concern in this report include: (i) the differences
between locked and unlocked Con A and (ii) the interaction of ConAwith saccharides
as viewed by 113Cd nmr.
Native Con A was isolated from jack bean meal (purchased from Sigma
Chemical Co. or Pfaltz and Bauer, Inc.) as previously described51. Enriched
113CdO (96%) was purchased from the Atomic Energy Commission (Oak Ridge
Laboratories) and converted to the chloride using a metal free HC1 solution.
Other metals salts used for metal competition studies (Pb(NO3)a, ZnCU, CaCU'
2H2O) were obtained commercially and used without further purification.
All buffers used in the dialysis of the apo-protein and for assays were rendered
metal free by dithizone extraction. Protein concentrations were measured
spectrophometrically at 280 nm using s0^^ = 1-2452 with the appropriate
dilution factor and a monomeric molecular weight of 2550050.
Apo Con A was prepared using a modified version of the procedure given
by Brown e^aL_31*. Approximately 1.0 g of native Con A was dissolved in
8 ml of deionized water and the pH lowered to 1.3 by dropwise addition of
1 M HC1 to the solution while stirring at 2°. After stirring for 45 minutes at
2° the solution was transferred to dialysis bags and dialyzed against deionized
water. After dialysis against water, the solution was perevaporated until
the total volume was approximately 10 ml. This step was found useful in
concentrating the Con A solutions. The concentrated protein was then dialyzed
against the appropriate metal-free buffer with the final dialysis against a DzQ
based buffer to provide an internal lock for nmr experiments. The solution
was centrifuged to remove precipitate and the Con A concentration of the supernatant
was determined. Typical protein concentrations run from 1-2.3 mM, in protomer.
The remetallated proteins were prepared by direct addition of the appropriate
salts to the apoprotein or by dialysis of the of the metal into the Con A pre-
parations. Saccharide binding activity was checked using the mannan assay
previously described56.
All 113Cd nmr studies were carried out on a highly modified Varian XL-100-15
spectrometer equipped with the Gyro Observe® option. Some experiments
were carried out using a frequency synthesizer mode of operation to be described
elsewhere1*5. All studies involved the use of a home-built multinuclear 18 mm
nmr probe1*1*. The sample volume necessary to observe 113Cd on this probe
is 5 ml. Due to the broad lines encountered in these systems, it was not found
necessary to spin the samples. All 113Cd chemical shifts are referenced to
an external standard of 0.1 M Cd(ClO1() 2 in 50/50 HaO/DzO1*8. A positive
chemical shift denotes resonances to lower shielding.
Brown et^al^31* have recently investigated the various conformational states
of Con A by the method of XH nuclear magnetic relaxation dispersion. They
found that Con A can be qualitatively described as existing in either of two
states, a metastable, unlocked form(s) or in a locked form(s). The unlocked
conformation(s) is characterized by rapid exchange of Ca(II) and Mn(II) with
the Con A moiety. The locked form(s) is characterized by slow exchange of
Mn(II) and Ca(II) with the protein. Further, it appears that the ground state
energies of both conformations are approximately equal with a relatively high
(22 kcal/mole) barrier separating the two states. This barrier allows the
conformers to be isolated and studied independently. Thus, at 5° the unlocked
conformer may be observed for a reasonably long time (several hours), whereas,
at 25° the metastable state will rapidly convert to the locked conformation.
Although the data described by these workers was only for the native metals
-------
(Mn(II) at SI and Ca(II) at S2), subsequent data Indicates that similar pro-
perties exist in preparations where Cd(II) is used as the S2 metal52. "" It
is of interest to investigate both conformations using 113Cd nmr techniques.
The Locked Conformation of Con A. Figure la depicts the 113Cd nmr spectrum
fora sample of Con A with two equivalents of 113Cd(II) per protomer. The
spectrum consists of three resonances at 68, 43, and -125 ppm. It is interesting
to note that there are three resonances in this system, and that the resonance
at -125 represents the most shielded 113Cd resonance reported to date.
Experiments using several different metals to compete for the 113Cd(II)
were used to assign these sites. Manganese, which occurs in the native state
at SI was used to compete for that site. Figure Ib shows the spectrum of Con
A to which 1.0 equivalent of Mn(II) and 2.2 equivalents of 113Cd(II) have
been added. Again peaks at 68, 46, and -125 ppm can be seen, although the
peaks show differences in intensity and lineshape from the peaks of the 2 Cd
spectrum. The resonance at 46 ppm is reduced in intensity, most likely due
to Mn(II) competition. This site was tentatively labeled SI. The peak at
-125 ppm shows increased linewidth. This could be due to paramagnetic
relaxation of the Cd(n) brought about by nearby Mn(II) or due to a slight
difference In chemical shift of the shielded site in those protein molecules
containing either Cd(II) or Mn(n) at SI. Finally the peak at 68 ppm shows
a greatly increased intensity and line broadening. The increase is greater
than can be explained by extra Cd{II) driven out of SI. Both the increase and
the broadening can be explained be efficient paramagnetic relaxation of 113Cd(II)
which originally had a relaxation time greater than the . 4 second recycle time
during data acquisition. Relaxation times of cadmium salts in solution are
much longer than the relaxation times of protein bound ions; this suggests
that the resonance at 68 ppm might be from non-bound ions.
To assign the S2 site, a calcium competition experiment was prepared.
Figure Ic represents Con A to which 2.2 equivalents of Cd and 1.0 equivalent
of Ca(II) have been added per protomer. The peak at -125 ppm has disappeared,
or is below the noise level. The peak at 46 ppm has gained intensity, and has
shifted to 43 ppm, leaving a small peak in the original position, addition of
an excess of Ca(II) causes the entire peak to shift to 43 ppm. The peak at
68 ppm has gained intensity, but shows no other change. These two experiments
indicate that the Mn(II) site (SI) is at 46 ppra, and the Ca(II) site (S2) is
at -125 ppm. They also imply that the site at 68 ppm is free in solution, or
at best, is nonspecifirally and loosely associated with the protein.
In order to confirm the non-bound nature of the resonance at 68 ppm,
an experiment using a varying chloride ion concentration was used. Chloride
ion in aqueous solution has a very high affinity for Cd (II), and has a large
deshielding effect on the 113Cd(II) chemical shift. A spectrum of 2 113Cd(II)
Con A in a chloride free buffer (.05M acetate, pH 5.2) had three peaks at
46, 8, and -125 ppm. The peaks at 46 and -125 ppm were independent of
the chloride concentration. The chemical shift of the resonance and the
chemical shift of 4mM 113Cd(NO3) 2 in the same buffer was plotted (Fig. 2)
as a function of chloride concentration. The moving resonance was shown to
have a chloride dependence almost exactly the same as the chloride dependence
of the resonance in the absence of protein.
Unlocked Conformation(s) of Con A. Attention will now be turned to the
the unlocked conformation of Con A. This conformer is easily prepared by
-------
adding the desired metals to cold apo Con A and keeping the resultant
remetallized protein at 5°C. As described earlier, unlocked Con A is characterized
by rapid exchange of the metals with the surrounding solution. So far these
experiments have been carried out only with mixed metal preparations of Con
A (Mn(II): Ca(II), or Mn(II): Cd(II)). It will be of interest to investigate
how the 2 Cd(II) protein will behave in the unlocked form.
Fig. 3a shows the spectrum of unlocked 2 Cd(II) Con A. The major difference
between this spectrum and that of the locked species is that only one resonance
(at 68 ppm) is observed. This unlocked species may be converted to the
locked by allowing the sample to stand at room temperature for a few hours.
The 113Cd nmr spectrum of the sample after standing is the same as that of
the locked species, shown in Fig. la. The origin of the resonance in Fig. 3a
may be due to one of several different possible environments for Cd(II).
These possibilities are: free Cd(II) with no interaction with the protein, Cd(II)
in the unlocked SI site only, Cd(II) in the unlocked S2 site only, or Cd(II)
in rapid exchange between all possible sites, including the free environment
outside the protein. The third possibility may be immediately ruled out on the
basis of the known binding order for Con A (SI before S2). To decide between
the remaining possibilities a metal competition experiment in which the 113Cd
nmr spectrum of an unlocked 1 Cd(II): excess Zn(II) Con A sample was obtained.
The spectrum, reproduced in Fig. 3b, shows a shift of the resonance to 72 ppm.
If the resonance was due to free Cd(II) only, no shift would have occurred.
Since the chemical shift in the presence of excess Zn(II) is deshielded with
respect to the resonance in the absence of Zn(II), implies that the chemical
shifts of 113Cd(II) in the SI sites of the locked and unlocked conformers are
similar. Thus, it appears that the resonance observed in the unlocked form
is due to Cd(II) in rapid exchange with all possible binding sites.
The magnitude of the contribution of the anion to the observed chemical
shift of Cd(II) in the unlocked species may be estimated by repeating the
2 Cd(II)-unlocked Con A experiment in a NO 3" containing buffer as opposed
to a Cl~ containing buffer solution. The results of this experiment are shown
in Fig. 3c. Again, only one resonance appears, however, now at -12 ppm.
This shift is similar in direction and magnitude to that seen in going from
CdClx~x species free in solution. In summary, it is evident that the unlocked
two Cd(II)-Con A species can be characterized as having cadmium in rapid
exchange with the unlocked protein. Further, the observed chemical shift
is highly dependent upon the nature of the salt employed in the buffer solution.
Interactions of Saccharides with Con A. The final portion of this report is
concerned with the 113Cd nmr of 2 Cd(II) Con A in the presence of binding
saccharides. Since the saccharide binding site has been determined to be
10-12 A from the SI site25 it should be expected that some change should
occur in the spectra especially in the SI and S2 resonances. First, as may
be anticipated, the 113Cd nmr spectrum of unlocked 2 Cd(II) Con A manifests
no change upon addition of saccharide. This result confirms that the protein
must be in the locked conformer to exhibit binding activity and has been in-
dependently measured using mannan rate assay and circular dichroism studies,
with excellent agreement among the methods.
Fig. 4 reproduces the 113Cd nmr spectrum of locked 2 Cd(II) Con A
to which a slight excess of the binding monosaccharide, methyl- a-D manno-
pyranoside has been added. The structure of this saccharide is given below.
-------
This sugar is known to bind tightly to Con A. The spectrum is similar to
the spectrum of Con A in the^ absence of sugars, but the -125 ppm resonance has moved
HO
to -133 ppm. The resonance at 46 ppm may again be assigned to the SI site
using methods analogous to those discussed above. As the saccharide binding
site is too far from the metal site for direct interaction (over 10A) 21, it appears from
this spectrum that a conformational change occurs upon binding the monosaccharide
which affects the S2 binding site. Edelman and coworkers21 have reported
dramatic conformational changes within Con A when the monosaccharide 2-deoxy-
2-iodo-methyl-crD-mannopyranoside binds. The most significant changes
occur in the metal-binding region when the entire region seems to move 4-6 A
with respect to the rest of the molecule. It is evident from the 113Cd nmr that
the monosaccharide methyl-orD-mannopyranoside is producing similar changes
within the 2Cd(II) Con A upon binding. There are suggestions in the literature56
that Con A binds differently to some monosaccharides than to some trisaccharides.
This is not supported by our own 113Cd nmr works; however, nmr experiments
using 13C and other nuclei are currently underway in our laboratory to further
Ust these observations.
Conclusions. Cadmium-113 nmr studies reported here demonstrate that two
113Cd(II)-Con A can exist in locked or unlocked forms as defined by Brown
et al.9*. The 113Cd nmr spectrum of the locked two 113Cd(II)-Con A preparation
yie!3ed three resonances, two of which can be assigned to cadmium bound to
the SI and S2 sites. The third resonance is attributed to 113Cd either not
bound or very loosely associated to the protein. The SI and S2 site are not
affected by changes in the buffer. Saccharide binding by those sugars that
Inhibit the activity of Con A may be monitored by a chemical shift at the S2
site, suggesting that a conformational change takes place when a sugar is
bound. In summary, it is clear that 113Cd nmr spectroscopy of cadmium
substituted metalloprotelns Is an excellent tool for delineating subtle changes
about the metal site in these proteins.
-------
B. Cadmium-113 NMR Studies of Cadmium Substituted Superoxide Dismutase56.
This research has been performed in collaboration with Professor James
Fee of the University of Michigan. A paper summarizing this work will be
submitted to the Proceedings of the National Academy of Sciences.
Superoxide dismutases, which are found in both aerobic and strictly anaerobic
organisms57 are a disparate class of proteins which possess Fe, Mn or Cu
as co-factors necessary for catalysis of the reaction
The protein isolated from bovine erthyrocytes has received the greatest
amount of attention and is the subject of this paper. This protein has a
molecular weight near 32,000 daltons and is made up of two identical subunits,
each of which binds a Zn(II) and a Cu(II) 58.
The Richardson's groupsaa)has determined the structure of bovine superoxide
dismutase by the methods of x-ray crystallography to a resolution on 3.2 A.
The metal binding site is shown in Scheme I.
Zn.
The two metals are approximately 6-7 A apart. The Cu(II) resides in a nearly
square planar arrangement of imidazole ligands and a water molecule appears
to act as an axial ligand directed toward the exterior of the protein 61. One
of the imidazoles has lost both its protons and forms a bridge to the Zn(II).
The Zn(II) resides in an irregular tetrahedron of ligand atoms supplied by
three histidines and a carboxyl group.
During catalysis of superoxide dismutation the Cu ion cycles between the
(II) and (I) valence states61 /62. A number of experiments support the contention 61*
that the bridge between the two metals is broken concomitant with the uptake
of a proton when Cu(II) is reduced and with the release of a proton when
Cu(I) is oxidized.
This communication is in part concerned with the question of whether
The Zn-N or the Cu-N coordinate bond is broken on reduction. (Scheme II)
Cu(D)_
-Or
-HD2
Cu(II) — ^±.—
! -01
-------
Already two lines of evidence suggest that it is the Cu-N bond that is broken
i.e. Ha. First, the optical spectrum of the 2Co(II) protein (Co(II) in the
Zn(II) site and the Cu site unoccupied) is very similar to that of the 2Co(II)-
2Cu(I) protein (Cn(I) in the Cu site)65'66. Thus one can infer from this
result that the Cu-N bond is broken in 2Co (II)-2Cu(I) protein resulting in
an isolation of the Cu(I) bonding site from the Co(II) site. The second line
of evidence comes from x-ray absorption studies67 which reveal that the
Zn(II) undergoes only minor structural perturbations upon reduction of Cu(II),
consistent with the Zn(II)-N bond remaining intact.
In this study, 113Cd(II) has been substituted for Zn(II) allowing direct
observation of the cadmium nucleus by NMR spectroscopy. Using this technique
we have examined the NMR properties of various cadmium derivatives of superoxide
dismutase, and one of our conclusions from these efforts is that the Cd(II)-N
bond is not broken upon reduction of the Cu(II).
Materials and Methods
Bovine superoxide dismutase (SOD) was isolated from bovine erythrocytes
according to the method of McCord and Fridovich68. Enriched 113CdO (96%
enrichment) was obtained from Oak Ridge National Laboratories and converted
to the chloride salt using metal free HC1. Copper sulfate was obtained commercially
and used without further purification. Doubly distilled water was used through
out and all other chemicals were of reagent grade.
The preparation of apoprotein was carried out using the method published 68 ,
and the Cd derivatives were prepared by adding the desired metals directly
to dilute apoprotein (1-2 mg/ml)in a 0.1 M acetate buffer the pH 5.5 and allowing
this solution to stand overnight at room temperature. The spectral properties
of the 2Cd(II)-2Cu(II) protein were identical to those described by Beem
et al.7,° and substitution levels of 90% were obtained. Protein concentrations
were determined using the method of Lowry 71 . Copper content of protein
sample was measured using e68o=150 M"1 , and cadmium content was deter-
mined by standard atomic absorption techniques.
The preparation of 2Cd(II)-2Cu(I) SOD from the oxidized protein was
carried out under strictly anaerobic conditions in a Kewaunnee inert atmosphere
glove box. First the 2Cd(II)-2Cu(II) SOD solution was rendered oxygen free
by purging the sample with N2. Solid sodium dithionite was then added directly
to the enzyme solutic- until the characteristic blue-green color of the oxidized
protein was completely blanched. The reduced enzyme was then extensively
dialysed against the appropriate buffer, also under rigidly oxygen-free con-
ditions, to remove the dithionite. Oxygen free D2O was added to the sample,
either by direct addition or by dialysis, to provide a deuterium lock for the
nmr experiments. After dialysis of the enzyme solution, the surface of the
dialysis bag occasionally exhibited a pink discoloration. The nature of this
material is not known, however, reoxidation of the sample caused the pink color
to disappear; the presence of this colored material had no effect on the outcome
of the experiments. The pH of the enzyme solution (uncorrected for deuterium
isotope effects and henceforth referred to as pH*) was measured inside the
inert atmosphere box using a Corning model 109 general purpose pH meter.
The pH* of the solutions was adjusted using either 1M H3POi» or 1M NaOH
with stirring. The reduced enzyme was finally transferred to an 18 mm nmr
tube which was sealed with a rubber stopper, silicone grease and Parafilm®.
8
-------
The inert atmosphere box was maintained at greater than atmospheric pressure
so that a positive pressure was present in the nmr tube during data accumulation.
The sample was reoxidized by the admission of air into the enzyme sample
after collection of the nmr data; the optical properties of the reoxidized protein
were identical to the original 2Cd(II)-2Cu(II) preparation.
All 113Cd nmr spectra were obtained on a highly modified Varian XL-100-15
spectrometer equipped with Gyro-Observe®. Some of the experiments were
carried out using a frequency synthesizer mode of operation to be described
elsewhere 72. All experiments involved the use of a home-built multinuclear
18 mm nmr probe 73. This probe requires 5 ml of sample to obtain a 113Cd
nmr spectrum. All 113Cd chemical shifts were referenced to an external sample
of 0.1 M Cd(CIO4)2 in 50/50 H2O/D2O. A positive shift denotes resonances
to lower shielding. Spin-lattice relaxation times (Ti) were measured using
the progressive saturation pulse sequence:
- (90° - T>n
where T is the recycle time (including data acquisition) and 90° refers to
the angle of the perturbing pulse 7" . No field gradient pulse was employed
in these experiments due to the broadness of the line, i.e., T*a«Ti. The
resulting data were fitted to the following equation using a nonlinear least
squares program
A(T) =A0exp ( -T/ T! )
where A(T ) is the intensity of the resonance obtained using the pulse interval
and A0 is the intensity obtained for r>_ 5 Ti. The use of the nonlinear program
obviates the necessity of determining AQ. The Nuclear Overhauser Effect,
NOE, was determined using the pulse sequence described above with T =7Ti.
The "enhanced11 spectrum was obtained with the proton decoupler on at all
times except during data acquisition. To minimize heating effects and line-
broadening, as well as decoupler interference, the NOE supressed spectrum
was obtained in the same manner. In the latter case, however, the decoupling
frequency was offset 20,000 Hz from the proton chemical shift range and the
noise modulation was turned off. The NOE was calculated using the equation:
MQE_A( "enhanced")
A("supressed)"
Results
The NMR spectra of three derivatives of 113Cd-substituted superoxide
dismutase were obtained: 2Cd(II), 2Cd(II)-2Cu(II), 2Cd(II)-2Cu(I) (Figures).
Some typical spectral parameters are given in Table I . For one of these
derivatives, 2Cd(II), a detailed study of the relaxation properties was made.
The results of progressive saturation and NOE experiments are shown in
Figures 6 and 7respectively.
Theory and Discussion
Relaxation Properties of 2Cd(ll)-Superoxide Dismutase. In any systematic
treatment of a molecular system by Fourier transform nmr methods, it is essential
to have a working knowledge of the spin lattice relaxation time, Ti, of the
nucleus of interest. It is well known that this parameter is a strong function
-------
Derivatives
Table I
Cadmium-113 Chemical Shifts for Various
Metal Derivatives of Super Oxide Dismutase
113Cd-Chemical Shift
2Cd(II) 311
2Cd(E)/23u(II) -d
33d(II)/2Cu(I) 320
Linewidth
27
-d
27
pH*/Buffer
5.5/.1M Acetate
4.7 or 8.0/ .1M
phosphate
5.5/.1M Acetate
Footnotes for Table I
(a)
(b)
(O
(d)
The designation m Metal refers to the number of metals per
dimer of protein
Chemical shifts in ppm with respect to 0.1M Cd(ClCU) 2-
A positive sign denotes resonances to lower shielding.
Linewidths are expressed in Hz.
No signal observed.
10
-------
of molecular structure and the motional dynamics of the system. At the very
least, a knowledge of TI is essential for optimum data collection rates. Cadmium-
metalloproteins are no exception, however it is worth noting that a single TI
and NOE experiment can represent a significant amount of instrument time,
i.e. approximately one week on our XL-100. In this section of the paper we
will discuss the measurement and interpretation of the 113Cd TI in 113Cd(II)-
substituted superoxide dismutase and its associated NOE.
There are several mechanisms available to an I=i nuclide for spin lattice
relaxation 75. However, when this spin is in a metalloprotein in a more or
less ionic environment there are basically two principal relaxation mechanisms
available: heteronuclear dipole-dipole (usually with protons) and chemical shift
anisotropy (CSA). Under ideal conditions these two mechanisms may be separated
by a knowledge of the NOE for the spin of interest. The experimental data
used to deduce the value of TI (1.2 seconds) and the NOE (0.66) are depicted
in Figures 6 and 7 respectively. It should be evident from the amount of noise
present in these spectra that the error associated with these numbers is
appreciable, i.e. approximately 20%. None the less, this error will not invalidate
the overall conclusions that one can draw from these experiments. o
From simple geometrical models 76, one can calculate the distance 2.8A,
between a hydrogen and the cadmium atom, in a cadmium-histidine complex.
If one makes the assumption that two hydrogens per histidine are equivalent
(with respect to their distance to the cadmium), then there will be six hydrogens
that could provide a dipolar relaxation pathway for the cadmium spin. It
is expected that the reorientational correlation time for the 113Cd bound to
a protein of 33,000 daltons will be in the range for 10 to 100 nsec. Using
these values of the correlation time and the geometrical model discussed above
one can predict the dipolar TI from standard equations 7?. These values
increase from approximately 2.3 seconds for a correlation time of 10 nsec to
approximately 16 seconds for a correlation time of 100 nsec. The experimental
value is 1.2 sec. Therefore one is forced to conclude that if the geometrical
arguments presented are valid, then the spin lattice relaxation is not dominated
by dipolar process, but rather that CSA processes are also contributing to the
TL
The amount of dipole vs. CSA can be estimated by a knowledge of the NOE.
The NOE in this type of spin system is not subject to a geometrical conjecture
concerning the structure of the active site. The NOE for a given pair of
heteronuclear dipolar coupled spins (such as 113Cd coupled to n 1H's) is only
a function of the correlation time describing the motion of the internuclear
vectors between the spins of interest. If the spin lattice relaxation mechanism
was dominated by CSA processes, then the NOE would be equal to one. The
largest algebraic value of the NOE if it was dominated by dipolar processes
is equal to 0.415 for a system out of the region of extreme narrowing. The
experimental value of 0.66 strongly suggests that the two relaxation pathways
(CSA and dipole-dipole) are about equally as efficient in providing spin lattice
relaxation for the 113Cd nucleus in superoxide dismutase. The value of 0.415
for the NOE represents the limiting value for a system with a correlation time
in excess of 50 nsec. If the correlation time for superoxide dismutase is shorter,
then the predicted NOE would be algebraicaliysmaller. This would imply a greater
importance of CSA process in the mechanism for spin lattice relaxation. Hence,
the 50/50 mixture represents a lower limit for the amount to CSA relaxation.
11
-------
Analysis of the 113Cd NMR of 2Cd{ll), 2Cd(ll)-2Cu(ll), and 2Cd(ll)-2Cu(l)
^uperoxide Dismutase. This section details the analysis of 113Cd nmr spectra
of the various metal derivatives of bovine superoxide dismutase. Before be-
ginning this analysis, however, it is prudent to review the known 113Cd nmr
data for relevant model complexes of cadmium and of metalloproteins. Attention
will be directed to the relationship between the chemical shift and the environment
of the 113Cd(II) nucleus, as well as the effects of external variables such
as anions and pH on these shifts.
From several reports in the literature the following generalizations are
made: octrahedral systems have the smallest chemical shifts (most shielded)
followed by progressively larger chemical shifts in tetrahedral and pentaco-
ordinate complexes. Within each of these groups, nitrogen ligands deshield
the resonance more than oxygen atoms. Sulfur ligands tend to cause a pro-
nounced deshielding of the cadmium nucleus.
Presently known 113Cd chemical shift data for metalloproteins are summarized
in Table "II. The most shielded 113Cd chemical shifts reported to date for a
metalloprotein are for the lectin Concanavalin A 78a. This protein exhibits
three 113Cd resonances of which those at -125 and 43 ppm are assigned to
the known S2 and SI sites, respectively 78^. The third resonance at 68 ppm
corresponds to "free" cadmium in solution. These sites have been shown by
x-ray crystallographic studies to have slightly distorted octahedral symmetry
with the majority of the first coordination sphere atoms being the oxygen of
amino acid residues or water 79. The Con A protein resonances do not exhibit
an anion dependency.
The majority of the proteins that have been investigated by 113Cd nmr
have four coordinate metal sites, although in one case five coordination is
possible . The 113Cd substituted metalloenzyme Carboxypeptidase A
exhibits no resonance unless an inhibitor such as 8-phenyl propionate is
added, in which case a resonance at 134 ppm occurs in the presence of chloride
anion. The crystal structure of this system shows the Cd(II) to be in a roughly
teterahedral environment formed by oxygen atoms for two glutamic acid residues
and a histidyl nitrogen with the fourth site open to solvent 80. In the above
cases, this fourth site is occupied either by water or chloride anion
Another metalloenzyme which has been investigated by 113Cd nmr is Cd(II)
substituted human Carbonic Anhydrase B, which Sudmeier and Bell 81, and
Armitage et al^82'83have studied. In the absence of anions, the latter investi-
gators report a chemical shift for the resonance of 146 ppm while the former
report 228 ppm for the same system. The nature of this discrepancy is not
clear, but may possibly be due to the amount of dissolved CO2 present in the
samples and /or different isozymes of the protein 81f . The x-ray structure
of this system shows the metal site to be a distorted tetrahedron composed
of three histidyl nitrogens with the fourth site open to solution 8 5 . In the
absence of anions, this site is occupied by HzO or OH, depending on the pH.
Further, the use of a variety of simple halide anions in the fourth site demon-
strated a chemical shift dependence of the Cd(II) resonance in this system,
with an observed variation of about 100 ppm 82 . Addition of
cyanide to this enzyme appears to form the pentacoordinate Cd(II) species
with a chemical shift of 410 ppm 81. The 113Cd chemical shifts of Cd(II)
substituted Alkaline Phosphatase and bovine Carbonic Anhydrase B have also
been determined 85. In the absence of competing effects, these metalloenzymes
have chemical shifts of 117 and 214 respectively. Although the crystal structures
12
-------
c
'55
S
a
_o
"to
aj
S
(»
_o
ro
n w
•Sc
Chemica
ppm)a
__ S— '
/•A "^
^j ,tz
S 5
1
"t/i
ojj
CD
•^3
QW
0)
CM
i — 1
1
O
LO
CM X)
I-H ro
,
O
C
o
c
V)
CD
oxygens
oxygens
nitro gen
^-J tD LO i— 1
o
o
o
1 — 1
1
o
o
en
i
i
i
i
i
oxygens
CD
T3
CO
CO
, — i
CO
0
>,
CO
0
^
nitrogens
03
t~*
oxygen ar
ubstrate
r— 1 CO
S
O
4-J K^
<«£"H
SH "70 OJ
O •$ tJ>
C en IX
0 3 X
T3 ^ O
m
CD*
i — !
CM
1
in
^T
r-H
CO
CD
>.
CO
0
nitrogens
CO
H
SH
O
O
ubstrate d
CO
H
O
<— 1
>1
i — 1
tO
3
co -p
3 C
sS 0
c O
0 >•
° X
(0 O
Cn
0"
o
i — i
i
LO
LO
co
0
>,
CO
0
^
nitrogens
ubstrate
CO CO
o
+->
ro
SH
O
C
O
T3
^
§S
CO >i
3 X
^- O
fO
u
1
ro
I
E
3
ro
U
II
8
o
0 0
+-> +-J
ro ro
G C
-rt -rH
T3 -O
SH SH
O O
O O
U O
X X
-rH -iH
CO CO
CN] i—I
CO CO
0
o
o
o
X
0
4-J
CO
c
•r—I
-a
o
o
o
o
M-l
CD
§
CD
4->
ro
C
-f-H
O
o
o
D
C
'3
4-1
O
a
_o
"m
ro
(0
(0
O
C
O
O
rd
PH
cu
C/3
tO
D.
0)
X
o
ra
O
CD
W
ro
S-i
U
•rH
SH
ro
O
CD
a,
w
O
0
I 1
rO
i—1
-------
m
US
m (/)
O
co
oo
i
o
i—<
CO
T3
cr
-=!<
to
(M
CO
oo
to
10
CD
w
o
u
to
\-
+J (A
DI Q)
o cn
o
co <— '
CD
w
(0
c
5
4»J
(0
c
£
s_
8
U
£
•1— 1
T5
^-*
O
O
Q
i-l
3
^
u
3
o
M-l
>^
1 — 1
,Q
•rH
CO
CO
o
D.
C
'3
+••
s
CO
CD
X
0
O f«
nS
Se
c
-rH
CD
C
.2
•M
^
13
+-"
CD
-------
Footnotes for Table II
a) Chemical shifts are in ppm with respect to external 0.1M Cd(ClOt).
A positive sign denotes resonances to lower shielding.
b) D.B. Bailey, P.D. Ellis, A.D. Cardin, and W.D. Behnke, J. Amer.
Ghem. Soc., 100, 5236 (1978). In this reference the assignment
of the SI and S2 sites was in error. The correct assignments are
given here (A.R. Palmer, P.D. Ellis, W.D. Behnke, D.B. Bailey
and A.D. Cardin, submitted to J. Amer. Chem. Soc.).
c) T. Drakenberg, B. Lindman, A. Cave1, and J. Parello, FEBSLett,
92, 346 (1978).
d) D.B. Bailey and P.D. Ellis, unpublished results. Both entries
denote 113Cd chemical shifts in the presence of g-phenylpropionate.
The entry for 133 ppm corresponds to the presence of Cl , whereas,
the resonance at 217 is in the presence of CIO if.
e) I.M. Armitage, R.T. Pajer, A.J.M.S. Miterkamp, J.F. Chelbowski,
and J.E. Coleman, J. Amer. Chem. Soc., 9_8 5710 (1976).
f) J.L. Sudmeier and S.J. Bell, J. Amer. Chem. Soc., 99, 4499
(1977). ~
g) I.M. Armitage, A.J.M.S. Miterkamp, J.F. Chelbowski, and J.E.
Coleman, J. Magn. Res., 29, 375 (1978).
h) present work
i) K.T. Suzuki and T. Maitani, Specialia, 34, 1449 (1978).
j) P.J. Sadler, A. Bakka, P.J. Beynon, FEES Lett., 94, 315 (1978).
k) J.D. Otvos and I.M. Armitage, personal communication.
15
-------
of these systems have not yet been determined, other work suggests that the
metal sites in both proteins are four coordinate, with three of the ligands in
each case being histidyl nitrogens
The 113Cd nmr spectrum of 2Cd(II) dismutase (Hg. 5a ) consists of one
resonance with a linewidth of 27 Hz. The appearance of only one resonance
indicates that if the Cd(II) sites in the two subunits differ, then this difference
only manifests a 113Cd chemical shift difference of less than 10 Hz (0.5 ppm).
Given the known sensitivity of the 113Cd chemical shifts to environmental
changes, this result suggests that the metal sites in the two subunits are
identical. The presence of the resonance at 311 ppm is in line with the metal
site being either four or possibly five coordinate. The crystallographic study
of this system reveals that the Zn(II) site (Cd(II) site) is four coordinate,
with three of the ligands due to histidyl nitrogens and the fourth site occupied
by an oxygen from an asparate residue 59. It should be noted that this
chemical shift is 90-150 ppm more deshielded than human Carbonic Anhydrase B,
which has the same ligand atoms (when no anions are present) and a similar
geometry at the Cd(II) site as the dismutase. This deshielding effect is in
all probability due to the greater covalent character of the aspartate-cadmium
bond in dismutase compared to the water-cadmium bond in Carbonic Anhydrase
or to a slightly different local symmetry about the cadmium.
No resonance was observed from the 2Cd(II)-2Cu(II) dismutase derivative.82
This result is to be expected as the paramagnetic Cu(II) atom is 6-7A from the
Cd(II) sites. Approximate calculations indicate that for our signal-to-noise
ratio the resonance will be broadened beyond recognition by the dipolar field
of the electron if the copper-cadmium interatomic distance is within 11A.
The reduced copper enzyme is of interest for a number of reasons. First,
the number of resonances that are observed will indicate the number of chemically
different Cd(II) sites in the enzyme. Second, the chemical shifts of this system
will indicate the amount of perturbation placed on the Cd(II) sites by the Cu(I).
Third, the dependence of this system on pH and buffer conditions will indicate
the accessibility of this site to solute molecules. Figure 5c depicts the 113Cd
nmr spectrum of 2Cd(II)-2Cu(I) dismutases. Only one resonance is observed
with a chemical shift of -320 ppm. Again, this result indicated that if there
is any difference between the two Cd(II) sites in the protein, that difference
only generates a 0.5 ppm chemical shift. Thus, as with the 2Cd (II)-protein,
the two sites are considered equivalent.
The observation of a small chemical shift between the Cu-free protein
and the 2Cd (II)-2Cu(T)-protein immediately suggests that only minor changes
occur at the Cd coordination site on the binding of Cu(I) to the Cu-site.
Mechanism (a) of Scheme II requires the breaking of the Cu(I)-imidazole bond
to provide the necessary proton while mechanism (b) has the Zn(II)-imidazole
bond breaking. If the correct mechanism were (a), and the Cd(II) system is
analogous to the Zn(II) system, the 113Cd nmr properties of the reduced pro-
tein would be similar to those of the copper free system. On the other hand,
if mechanism (b) were correct, the Cd(II) species would be required to break
the Cd-N bond which would either not be replaced (which is unlikely) or would
be replaced by water or some component from the solution. In any case, the
113Cd nmr properties of the reduced system would be dramatically different
from those of the copper free protein. On this basis, the 113Cd nmr results
indicate that mechanism (a) is the more likely. This point can be strengthened
by obtaining the 113Cd nmr data with a variety of pH and buffer conditions.
The following systems were therefore used: (i) 0.1 M acetate buffer, pH*
5.5 (ii) 0.1 M phosphate buffer, pH* 8.0, and (iii) 0.1 M phosphate buffer,
16
-------
pH* 4.7. Under all of these conditions there were no changes observed in
the 113Cd nmr spectrum of the 2Cd (II)-2Cu (I)-protein. An experiment was
also carried out in which the dithionite was not dialysed away from the protein.
Again, no change was noted in the chemical shift of the protein.
Several lines of evidence now support the contention that the imidazole
bridge between Zn and Cu is broken at the Cu site upon reduction. The
Cu(I) which remains very tightly bound to the protein86 would thus appear
to be coordinated to three atoms donated by the protein and to have at least
one position open to the solvent 87.
The certainity of this conclusion now allows speculative comment on the
role of the bridging imidazole in the catalysis process. Hodgson and Fridovich88
speculated that during catalysis the bridge breaks at Cu, a proton is then taken
up from the solvent as was shown by Fee and DiCorleto 6\ This proton
is subsequently donated to Cu(I) bound Oa" followed by electron transfer
from Cu(I) to the superoxide to form Cu(II)-O-O-H and to reform the bridge.
The present work in conjunction with other studies provides a rational structural
basis for such a mechanism but does not provide any evidence for its occurrence.
Two important facts which must be considered in asking how protons are donated
to O2 in the second step of catalysis 87 are: a-free aquo Cu-ions are ^ 4X
more efficient catalysts of superoxide dismutation then Cu-dismutase 8 9 , and
b_-removing the Zn from the protein lowers activity by at most a factor of two.
The question is raised as to whether a protein mediated proton transfer mechanism
is necessary in the catalytic process.
While this work was in progress, a report appeared by Armitage e_t al. 82
concerning similiar studies of SOD. These workers found that the copper
free system exhibited a chemical shift of 179 ppm. Further, upon reduction
of the 2Cd(II)-2Cu(II) preparation with dithionite, a chemical shift to 7 ppm
was observed, with a large increase in the linewidth of the resonance. These
observations are diametrically opposed to our own results. The possibility of
differences due to pH and buffer composition must be ruled out on the basis of
our own work. Further, the results of Armitage et_ al. suggest that a large
change occurs in the Cd(II) site between the copper free protein and the reduced
copper system. This change is incompatible with the mechanism proposed for
proton insertion by both ourselves and Armitage. Also, the results shown for
other systems indicate that the Zn(II) site is extremely stable and not likely
to exhibit such pronounced differences in the 113Cd nmr. On these bases, we
must conclude that our results are more indicative of the native system than
those of Armitage. The difference in results may be due to the method by which
the enzyme systems were prepared. The work referenced by Armitage describing
the preparation of the apoprotein does not include the perchlorate dialysis to
remove tightly bound EDTA 69. The presence of the EDTA has been previously
shown to produce erroneous results
17
-------
C. Cadmium-113 NMR Studies of Carboxypeptidase
Carboxypeotidase A is an enzyme that catalyzes the hydrolysis of peptides
and esters 90. It is an exopeptidase active toward the carboxyl terminus of
the peptide, where the residue is usually a phenyl derivative. The enzyme also
exhibits esterase activity toward esters containing phenyl groups. The protein
has a molecular weight of 34,000 and contains one zinc atom per molecule 91.
The zinc is in a roughly tetrahedral environment with one site open to the
environment. In a noncatalytic situation this site is filled by water or anion
such as Cl~. During catalysis this site is believed to coordinate the carbonyl
of the substrate. This conclusion follows from the X-ray work of Lipscomb and
his co-workers 92. In this research, the X-ray structure was determined on a
binary complex of the enzyme and the peptide glycyl-tyrosine. This pseudosubstrate
is not cleaved by the enzyme which, of course, raises some questions about
the relevence of this structure to the binding of typical reactive substrates.
The enzyme can also cleave esters of a-hydroxy acids, such as a-phenyllactic
acid; no X-ray work on Carboxypeptidase A with a found ester substrate or
pseudosubstrate has been done. A unified picture of the enzyme mechanisms
with Carboxypeptidase A has been recently proposed by Breslow and Wernick
They propose the following schemes:
93
(i) Peptide substrates bind so as to displace water from Zn(II).
molecule is then delivered by the glutamic car boxy late.
A water
E—CO 2
_xv />!
' H—O O
V
H
I
NH
I
I
Zn(II)-
I
(ii) Ester substrates may bind without displacing the water from Zn(II).
This puts the esters in position for a nucleophilic attack by the glutamate.
E-CO.
I
O
I
^—O H O-
I
H
I
Zn(II)-
I
A detailed picture of che peptide hydrolsis mechanism is envisioned by Breslow
to be the following:
M/
Zn(II)
R-C^IHR'
OH
u/
Zn(II)
R-C—NHR'
'*'H
O
COaH
E
\
1 H • • -O
a
Zn(II)
R—C-NH—R
W/
Zn(II)
O
+H2N-R'
18
-------
From the preceding mechanistic picture two points are relevant to the present
research: (1) ester substrates may not involve carbonyl coordination to the
Zn(II), and (2) as the products disassociate in the peptide hydrolysis mehcanism,
the resulting acid is coordinated to Zn(II) .
Byers and Wolfenden have examined several inhibitors of carboxypeptidase
A a1*. In the course of their research, they developed a so-called product
inhibitor of the enzyme. That is, the inhibitor, L-benzylsuccinate, resembles
the collected products of the peptide hydrolysis, and thus it is bound with an
affinity resembling their combined affinity.
If Breslow's detailed mechanistic picture93 is correct and if Wolfenden's
inhibitor indeed resembles disassociated products 9lt , then one would predict
that the Zn(II) would not be accessible to solvent water molecules or anions
such as Cl~ in the presence of the inhibitor. However, Byers and Wolfenden9lf
have demonstrated that L-benzylsuccinate can bind to apo carboxypeptidase
This binding is sufficiently tight to prevent Zn(II) from coordinating with the
enzyme. The results suggest that L-benzylsuccinate may not be bound to the
Zn(II) . Our research to date on carboxypeptidase has dealt with the question
of how L-benzylsuccinate binds to carboxypeptidase.
Cadmium-113 nmr spectroscopy should be able to shed some light on this
situation. Before we address ourselves to substrate binding effects, it would
be prudent to study carboxypeptidase in the absence of any inhibitors. The
results of these experiments have been exceptionally frustrating, in that to date
we have been unable to observe the 1:L3Cd resonance for the uninhibited enzyme.
This appears to be due to some intermediate exchange condition associated with
ligand exchange on and off the cadmium. Our initial efforts have been performed
on carboxypeptidase at approximately a 2 mM concentration. This concentration
was obtained by having the protein in a 3 M NaCl salt solution in which the
chloride ion would be expected to bind to the metal. We have rece_ntly found
that the enzyme can be solubilized by high concentrations of ClOit . This anion
would coordinate the cadmium to a lesser extent than Cl ion. We are presently
investigating the 113Cd nmr of carboxypeptidase A in the ClO^ solutions as a
function of the solution pH. An important point to make at this stage is that the
cadmium enzyme still maintains its esterase activity in the presence of ClOi* .
Our successful 113Cd research on carboxypeptidase A to date has been
achieved when the enzyme is in the presence of two inhibitors g-phenyl propionate
and D,L-benzylsuccinate. In the case of the g-phenyl propionate, a resonance
was observed at _134 ppm deshielded from 0.1 M Cd(ClOit) 2- The resonance was
obtained in ClO^ /buffer solution and the observed linewidth was approximately
32 Hz. Repeating the experiment using benzyl succinate yielded a resonance
at 140 ppm in the ClO^ /buffer solution with a linewidth of approximately 90
Hz. The 113Cd resonance, the D,L-benzylsuccinate inhibited enzyme in Cl /
buffer appeared at 217 ppm with an associated linewidth of 230 Hz. The
previous experiments were performed with a pH of 8 for the solution. The
corresponding experiment for D,L-benzylsuccinate in the Cl /buffer at a pH
of 9 yielded a resonance at 209 ppm with a linewidth of 138 Hz.
This large increase in linewidth between D,L-benzylsuccinate and 3-phenyl
propionate in the ClOij solutions is surprising. Since D,L-benzylsuccinate
has a KI approximately 136 times smaller than 3-phenyl propionate, and if both
inhibition are bound to the enzyme in_ the same way, we would have expected a
19
-------
linewidth narrower for the benzyl succinate than that observed for B-phenyl
propionate. These results clearly suggest that the cadmium in the benzyl
succinate inhibited enzyme is more accessible to water and ion exchange effects
than when g-phenyl propionate is inhibiting the protein. This assertion is_
further supported by the 113Cd nmr data of the inhibited enzyme in the Cl /
buffer solutions. That is, the linewidth is substantially greater in these solutions
aad it is pH dependent. Hence, one obvious conclusion that can be drawn from
the research done to date is that D,L-benzylsuccinate binds to carboxypeptidase
A in such a way that it does not coordinate the metal. Subsequent experiments
are needed to confirm this conclusion. Further, the present data cannot be
ased to support or refute the notion that the carbonyl oxygen of the inhibitor
g-phenyl propionate binds to the Cd(II).
20
-------
IV. Selenium-77 NMR Investigations of Model Biological Systems.
A preliminary account of this research has been published (W.H. Dawson
and J.D. Odom, J. Amer. Chem. Soc., 99, 8352 (1977)). The more detailed
summary, to be presented below, will be submitted to the Journal of the American
Chemical Society.
While Fourier transform nuclear magnetic resonance (FT NMR) spectroscopy
has been used to great advantage to study the chemistry of many elements of
the Periodic Table, those of Group VIA have received little attention. For oxygen,
and sulphur, the two lightest and most chemically prolific members of this group,
the only NMR active isotopes (17O and 33S) are quadrupolar nucleii which suffer
from very low natural abundance and relatively low sensitivity. Selenium and
tellurium, on the the other hand, both have spin-1/2 isotopes (77Se, 123Te,
125Te) with sufficient sensitivity to make their study readily accessible by
FT NMR. For the purposes of this investigation, our interest in selenium
stems from the active role it plays in many biological systems, not to mention
its increasing involvement in organic synthesis and an extensive inorganic
chemistry 95. Selenium-77 NMR has great potential as a means of exploring
the chemistry of this interesting element and as part of our continuing interest
in the applications of multinuclear NMR to biological systems96 we have initiated
a program aimed in this direction.
Early continuous wave (CW) nmr studies by Birchall, et_aL_,97 followed
by the INDOR studies of McFarlane and Wood98 demonstrated the large chemical
shift range of 77Se and the stereospecificity of its coupling constants. To date
there have been only a few reports concerned with the direct observation
of 77Se by FT NMR960'99-1.03 The nuclear spin-lattice relaxation time, T!, is a
critical parameter in determining the recycle time of FT NMR experiments.
More importantly, it can often be used as a powerful, diagnostic tool for the
determination of molecular structure, conformation and composition 1Q1* . It
can also be employed as a probe to investigate molecular motions and ineractions.
Dawson and Odom,96c Pan and Fackler,100 Gansow, et al.101 and Koch et aL_,102
have briefly examined spin-lattice relaxation times for 77Se in a variety of
chemical environments. The result of these studies indicated an importance of
spin-rotation (SR) and, to a lesser extent, chemical shift anisotropy (CSA)
mechanisms for spin-lattice relaxation.
Relaxation considerations assume an added significance when dealing with
large biological macromolecules that contain selenium. Those seleno-systems
which are involved in biochemical reactions will eventually be the focus of
greatest concern in our nmr studies. However, their high molecular weights
impart a motional sluggishness which, from the viewpoint of NMR spectroscopy,
requires operating near the limits of the so-called "region of extreme narrowing"
(UJT « 1) and complicates the task of extracting useful physico-chemical inform-
ation 105. It is advantageous to be forearmed with a knowledge of the mechanisms
governing TI and we have, therefore commenced our biological studies using
77Se NMR with an examination of the spin-relaxation of this nucleus as it appears
in a number of functional forms and in both aqueous and non-aqueous solvents.
Spin-lattice relaxation times were determined on a modified Varian XL-100-15
NMR spectrometer. Field-frequency stabilization was achieved by locking to
21
-------
the deuterium resonance of the solvent (either CDC13 or D2O). Selenium-77
resonances were initially detected at 19.1 MHz using the Gyrocode Observe®
accessory with the deuterium lock frequency at 15.400960 MHz. Under these
conditions, however, it was not possible to simultaneously proton-decouple
the sample at 100 MHz using either coherent or modulated (noise or squarewave)
irradiation without introducing a prohibitive amount of noise into the receiver.
This feature is apparently common to all XL-100 spectrometers100'106. A solution
to this difficulty was achieved by lowering the field slightly such that 77Se
resonates at 18.90 MHz with proton-decoupling at 99.41 MHz. The deuterium
lock frequency at this new field strength is exactly 15. 30 MHz. During the early
stages of this work this latter frequency was obtained by replacing the 15.400960
MHz master crystal by 15.3 MHz from a frequency synthesizer. Since the
Gyrocode Observe reference frequencies are ultimately derived from this master
frequency, the outputs of this unit using the new frequency were automatically
scaled to the required 99.41 MHz decoupling and 18.90 observe frequencies when
the "normal" 15.4 MHz Gyrocode settings were used. In the latter stages of
this study, the XL-100 was converted to a synthesizer-based instrument, wherein
the observe frequencies were derived from a General Radio 1061 frequency
synthesizer. The 15.3 MHz master frequency was generated by a home-built
frequency synthesizer which was phase locked to the GR 1061 synthesizer. The
standard components of the Varian 4412 probe were retunable from 15.4 to 15.3
MHz and 19.1 to 18.9 MHz without any perceptible performance losses.
The inversion-recovery pulse sequency, (180° - T- 90° - T)n, was used
to obtain the relaxation times (T i) when the magnitude of this parameter was
on the order of 5 sec of less . The 90° and 180° pulse widths were determined
in the usual manner with 90° pulse being on the order of 65 ysec. The waiting
time between pulses (T) was always set to be greater than 5 TI'S. When Ti's
longer than 5 sec were encountered, the homospoil saturation recovery sequence,
(PD- HS - 90° - HS - T- AT) was employed where PD, HS and AT are the
pulse delay (nominally 0.1 secj, the homospoil (Z-gradient) pulse, and the
acquisition time, respectively 108. In the case of either pulse sequence, at
least one spectrum was obtained with the value of T exceeding five times T i x °9.
The value of TI was extracted from the experimental data using a non-linear
least squares routine in the usual manner. The uncertainty in the Tx values
determined in this way is estimated to be within 10%, although repetition of
the measurements for selected samples used in this study indicated that the
values are reproducible to well within these limits.
Nuclear Overhauser effect enhancement factors (n) were determined as the
difference in integrated intensity between 1H-decoupled and -"-H-coupled spectra
and are reported as NOE = 1 + n 110. The measurements were taken from
two spectra, one in which the ^-decoupling frequency was set exactly on
resonance and the second with the decoupling frequency offset at least 10
KHz from this resonance frequency, with the modulation removed. The recycle
time between pulses in these experiments was at least seven times the value
~.t T> 111
Of Tl-
All spectra were obtained using 250 or 500 Hz spectral widths and trans-
formed using 8K data points (zero-filling as required). Chemical shifts are
reported relative to the resonance of dimethylselenide in CDCls (^1.0 M) in
the sense that a positive chemical shift denotes a resonance to lower shielding 112 .
22
-------
Temperature control was achieved using standard Varian accessories and
temperatures were measured with a copper-constantan thermocouple unit, the
"hot" end of which was inserted into an open sample tube containing the ap-
propriate solvent and held at a depth coincident with the receiver coils. The
tempertuares reported in this work are estimated to be accurate to ±1°.
All samples were degassed by several freeze-pump-thaw cycles and sealed
under dynamic vacuum in 12 mm NMR tubes. The solvents (CDC13 and DzO)
were treated with dithizone prior to being distilled in an all-glass apparatus.
Paramagnetic materials were removed from nmr tubes by allowing the tubes
to stand in a nitric acid bath for several days and rinsing them several times
with deionized water.
The following materials were obtained from commercial sources: selenium,
selenium dioxide, sodium selenate (Alfa-Ventron), d,l-selenomethione and
d,l-selenocystine (Sigma Chemical Co.,). Anhydrous hydrogen selenide was
a gift from A.J. Zozulin of this department and N,N-dimethylselenourea was
a gift from Dr. R.A. Zingaro (Texas A & M University) . Dimethyl selenide
(b.p. 54-57°; lit.113 57°) was prepared from the reaction of CH 3I and Na2Se
in aqueous base lllf. Di-n-butyl selenide (b.p. 196-197°; lit. 113 82-83°/13
torr) , di-n-octyl selenide115 (b.p. 157-158°/0.1 torr) and di-isopropyl selenide
(b.p. 136-137°; lit116 135°) were obtained from the reaction of the corresponding
alkyl bromides with Na2Se in liquid NH3116 . Methane selenol (B.p. 23-25°;
lit. 117 25.5°) was prepared from the reduction of dimethyl diselenide with
hypophosphorous acid118 . Ethane selenol (b.p. 50-53°;lit.116 52°) and
decane selenol (b.p. 125-130°/18 torr; lit.119 128-129°/13 torr) were synthesized
by the reduction of the corresponding dialkyl diselenides with elemental sodium
in liquid NHs116 • Dimethyl diselenide was obtained as an orange oil (b.p.
156-158°; lit.120 155-157°) from the reaction of CH3I and Na2Se2 in aqueous base1!1*
Debenzyl diselenide (m.p. 92-93°; lit.121 92-93°) was prepared from benzyl
chloride and Na2Se in ethanol122 . Diphenyl diselenide (m.p. 63.64°; lit.113
61-63°) was synthesized by the air oxidation of a solution of benzene selenol 123.
Didecyl diselenide (m.p. 12-16°; lit.121* 13-14°) was obtained as an orange
oil from the reaction of decyl bromide with Na2Se2 in liquid NH3116 . Trimethyl
selenonium iodide (m.p. 150-152°; lit.125 150-152°) and dimethylselentine
bromide (m.p. 89-92°; lit.111* 90°) were prepared by mixing cold diethyl
ether solutions of dimethyl selenide (20% molar excess) and the corresponding
alkyl halide and subsequently removing the volatile materials. Dimethylselenoxide
(m.p. 91-93°; lit.126 94°) was prepared from both the ozonolysis of dimethyl
selenide in CHC13113 and the reaction of silver oxide with dimethylselenium
dibromide in methanol114 . Selenocysteamine hydrochloride (m.p. 108-110°;
lit. 127 108-110°) was synthesized by the method of Klayman127 and methyl-
selenocysteamine hydrochloride (m.p. 149-151°; lit.128 149-151°) was prepared
according to the method of Tanaka et a_-_128 • In addition to the determination
of boiling points or melting points, the purity of all compounds was carefully
checked by XH and 77Se NMR.
23
-------
Diakyl Selenides
The data for this group of compounds is collected in Table IE. In each of
these molecules, scalar spin-spin coupling of the 77Se nucleus to nearby alkyl
protons is observed. Proton-decoupling reduces the resonances to single
sharp line, but in no case was a 1H-77Se NOEobserved. Thus, although the
presence of these nearby spin-1/2 nuclei must certainly give rise to a 1H-77Se
dipole-dipole relaxation, it is apparently very inefficient and its contribution
to the observed T i is minor relative to the other mechanisms. An earlier study1 °1
reported an NOE for (CH3) 2Se of 1.04. One can eliminate in principle a contri-
bution from scalar coupling relaxation since there is no means by which it could
be operative in these systems.
In order to distinguish which of the remaining two mechanisms is operative
(chemical shift anisotropy (CSA) or spin rotation (SR)) the 1 \ was determined
as a function of temperature and the data plotted as a semilog plot of Ri^Tx"-1
versus the reciprocal temperature (Fig. 8 ). The temperature dependencies
of the relaxation processes dictate that in such a plot the SR mechanism will
give a negative slop while the CSA mechanism will yield a positive slope, and
under the assumption (valid for these small molecules at the Larmor frequency
employed) that the extreme narrowing condition applies 129 . As may be seen
from Figure 8 , the negative slope and strict linearity of the plot for dimethyl-
selenide points to the fact that the spin-lattice relaxation of 77Se in this compound
is dominated by TI (SR) in the temperature range studies. This result is in
accordance with theoretical and practical expectations since the SR mechanism
requires a coupling of the nucleus to the valence electrons in a manner similar
to that which gives rise to the so-called "paramagnetic shielding" term of the
chemical shift . As a consequence, heavy, magnetically active nuclei which
possess a large chemical shift range will, in small molecules, usually, display
significant SR relaxation. This fact has been demonstrated experimentally for
a number of heavy nuclei, e.g., 113Cd96a, ^Hg9^ , and 205T 131 . The
dioctylselenide molecule should, by virture of its larger size, tumble less rapidly
in solution than dimethylselenide and, hence, SR should be less important in
this case. In fact, as is seen in Figure , only at the higher temperatures
does SR become important and at low temperatures the steep positive slope and
lack of an NOE imply a CSA mechanism. The increased importance of the SR
mechanism at higher temperatures is undoubtly due to the increased overall
motion of the molecule coupled with increased segmental motion of the selenium
atom within the chain. The quantitative separation of these motions is beyond
the scope of this report.
The slower tumbling experienced by selenium in dioctylselenide at -50°
contributes, in this case, to a diminution of the SR mechanism and an enhance-
ment of CSA relaxation. This aspect is of considerable interest to the study
of macromolecules containing an RSeR1 moiety because of the implication that
the CSA mechanism may be important in these systems. If one assumes that
the motion of the 77Se nucleus can be described by a single correlation time,
T, then the relaxation rate for CSA processes can be expressed1053 as
Se c
Here, YO is the magnetogyric ratio for selenium, HQ is the applied magnetic
-------
Table III. Selenium-77 Chemical Shifts and Spin-Lattice Relaxation Times
Chemical3
Shift TI Temp
Compound
(CH3)2Se
(n-C4Hg)2Se
,n-C8H17,2Se
(i-C3H7)2Se
d,£-CH,,SeCH9CH0CH
o ^ Z
wri rtO G(_/O ^wH «1N! ri «
(CH3Se)2
(C6H5CH2Se)2
(C6H5Se)2
(n-C1()H21Se)2
(ppm)
Ob
3
167
168
432
(NH2)COOH 75
44
50
281
412
481
316
(sec)
5.2
7.5
8.6
24.4
4.3
9.8
19.1
23.1
16.5
10.4
14.8
6.0
8.7
13.6
20.4
13.5
8.4
15.9
24
28
9
13
27
31
20
31
21
14
(°C) Conditions
32 CDC1-,
12 J
-60
40
030 acetone
40
0 CdCl3
-40
41
0 CDC1,
-42 J
41
0 CDC1,
-42 J
34 0.1 M, D2O, pD 4
43
0 0.5 M, CDC13
43
10 0.5 M, D0O, pD 4
4Q 0.5MCDC13
55 0.5 M, CDCU
18 ^
45 0.5 M, CDClo
0 J
43 0.5 M, CDC1-
0 J
25
-------
Compound
CH,SeH
J
O
C2H5SeH
n-C1()H21SeH
UC/-V/"*TI r"*u MHJ
no con ^on /vLMn «
d, £-HSeCH9CH(NH9)COOt
L* £
H2Se
(NH4)2Se
NaSeCH3
(CH3)3SeI
(CHJ JSeCBrJCH^pO'p'^c
Chemical
Shift
(ppm)
-130
39
.41
-7
-212
I -141
-288
-511
-330
258
140
3
"298
Ti
(sec)
1.3
3.3
3.7
1.7
4.8
9.5
1:1
7.7
1.9
3.4
7.1
-
0.7
1.0
-
16,3
13.7
1 8
II /
0.7
25
30
21
Temp
(°C)
40
-30
-45
40
-30
-60
40
-30
-83
42
0
32
-
34
- 10
-
43
32
32
- 1
-30
65
35
8
Conditions0
acetone- dc
f-»
\J
CDC13
acetone-dR
O
CDC13
0.5 M, D2O, pD 8.3
0.1 M, D2O, pD 5
l.OM, D90d
Lt
0.5M, D2Od
l.OM, D20d
1.0 M, D2O, pD 7
0.5 M, CDCU
O
0.5 M, D0O, pD 7
£
26
-------
Chemical
Shift Ti Temp
Compound (ppm) (sec) (°C) Conditions
H SeO3 1292
1307
1281
1307
Na2SeO4 1051
(CH3)2SeO 819
(CH3)2NC(Se)NH2 147
3.
2.
2.
8.
3.
2.
10
12
6
16
4.
8.
1
1
8
5
6
8
.0
.5
.5
.8
4
9
8.6
10.0
63
12
63
10
67
10
63
10
61
15
65
30
55
32
1.
0
M, D
•A
pH 9.
6
£j
1.
0.
0
5
M, D
9
£
0,
M,D00,
pH 1.
pH 9.8
5
r
Lt
0.
0.
0.
0.
5
5
5
5
M, D
M, D
M, D
M, D
9
Li
2
2
2
o,
o,
o,
o,
pH 1.
pH 6.
pH 7
pD 4
5
6
Relative to dimethylselenide in CDCl-. The chemical shifts are all temperature
dependent and those given are for temperatures between 32 and 42°C.
The exact Larmor frequency for (CHJJSe in CDCl- is 18.957787 MHz with
O L, O
2
the H lock frequency (internal CDClg) at 15.3000000 MHz.
Unless otherwise specified, the samples were 20% (v/v) in the specified
solvent.
See text.
27
-------
field strength, Aa is the chemical shift anisotropy and M_ is the resonance
frequency for selenium in radians per second. If one asl&unes a value of
10- sec/radian for a correlation time132 for the 77Se nucleus in dioctylselenide
at 50° then from Figure 8 and Eqn [ 1], one can estimate a value of Ao for
this system to be 971 ppm. This value is in accord with the value of Aa =
1260 ppm determined for a single, crystal of elemental selenium133. Hence,
for a protein with a correlation time in the range of 10 to 100 nsec, one would
expect Ti's in the range of 0.14 to 0.8 sec., respectively. Chemical shift
anisotropy contributions to the line width for this hypothetical system can
be calculated from Eqn. [2],
T2-l(CSA) = ^Y|eH* (Aa)2[8tc+ ^ H [2]
1 + a)SeT c
The expected linewidth would be between 5 Hz and 38 Hz for correlation times
between 10 and 100 nsec/radian, respectively. If the value of 971 ppm is
representative of the CSA for the 77Se nucleus in RSeR1 compounds, one can
expect that CSA relaxation mechanisms will dominate for molecules undergoing
slow molecular motion.
The molecules diisopropylselenide and dibutylselenide have sizes inter-
mediate between those of dimethylselenide and dioctylselenide and their Ti's
reflect this transition. The plot for diisopropylselenide (Fig.8) indicates
that. SR is the dominant mechanism, with the TI'S at equivalent temperatures
being marginally longer than for the smaller dimethylselenide. For dibutyl-
selenide the plot displays a pronounced curvature similar to that for di-
octylselenide but the Ti's are longer at all temperatures. Clearly, the SR
and CSA mechanisms, both of which are contributing here, are less efficient
in this compound, a situation that arises because the molecular tumbling is
too slow to provide an effective SR relaxation and too fast for an effective
CSA relaxation.
For completeness one must also consider the possibility that the change-
over in spin-lattice relaxation mechanisms between the dimethyl and dioctyl-
selenides may be due to a difference in either the anisotropy of the chemical
shift or the spin rotation coupling constant1053'130. However, these parameters
should not be radically different for homologous compounds like these dialkyl
selenides. The similai Ti's for dimethyl- and diiso- propylselenide whose
resonances are separated by a chemical shift difference of 340 ppm and for
which one might expect the differences in these parameters to be the greatest,
support the contention that motional arguments provide the most consistent
rationale of the data and probable cause of the difference.
The amino acid d, 1-selenomethionine, CH3SeCH2CH2CH(NH2)COOH, was
studied as a 0.1 M solution in DzO(pD*2). The TI value of 13±1 sec at 34°
falls between that for dimethyl- and dibutylselenide which is in accord with
the above data as regards its size relative to the simple dialkyl selenides.
The fact that the amino acid will be protonated at pD=2 may increase its
effective size as a result of solvation. To study this particular aspect as
it relates to the Ti's of dialkyl selenides in aqueous solvents we measured
the Tiof a similar compound methylselenocysteamine, CHs SeCKfeCHaNH^
in both CDCla and D2O (pD%2). The magnitude of the T value in CDC13
is similar to that of (CH s)2Se (Table III) and the temperature dependence of
this value implies that SR is the dominant mechanism. The compound appears
to be much more mobile in CDC13 than in D ^0 as evidenced by the much longer
28
-------
TX'S in acidic aqueous solution (Table H) . Thus, although the SR mechanism
again dominates TI, solvation in aqueous solution has increased its size and
slowed the tumbling of the molecule. The chemical shift of this compound
does not reflect this change to the same extent - resonance in D2O is very
similar to that of the amino acid selenomethionine and shifts by only 6 ppm on
changing the slovent to CDCls.
Dialkyl Diselenides
The spin-lattice relaxation of 77Se in the diselenides (Table III ) closely
resembles that of the dialkylselenides in that spin rotation is the dominant
mechanism for the smaller diselenides but CSA relaxation becomes important
when the size of the molecule becomes very large, as in didecyl diselenide.
These conclusions are based on the temperature dependencies of the TI'S and
the lack of a measureable 1H-77Se NOE for any of the compounds. The TI'S
for these compounds are of the same order of magnitude as for diphenyldisenlenide
has been previously reported 101. This is shorter than any value observed
in this study; however, without a knowledge of how the samples were prepared
for the NMR experiments in the previous study, a meaningful comparison can
not be made.
Selenols
Spin Rotation accounts for the relaxation in the alkane selenols (Table III
and Figure 9) . The Ti's are small relative to the dialkyl selenides and are
largely independent of the length of the alkyl chain, suggesting that in the
long chain compounds there is enough segmental motion at the end of the
chain to provide an efficient SR relaxation in spite of the molecules' long
overall correlation times. The absence of a measureable NOE from the directly
bound hydrogen in these compounds in in sharp contrast with what is commonly
experienced in 13C nmr, where the 1H-13C DD relaxation is a most important
mechanism 13It . This situation reflects not only the much greater efficiency
of SR relaxation for 77Se as compared with 13C, but also the inefficiency of
1H-77Se DD relaxation. If, under conditions of extreme narrowing, the 77Se
was relaxed exclusively by the DD process the maximum NOE enhancement of
2.61 (total intensity 3.61) would be observed 135 . The 1H-77Se DD relaxation
is diminished primarily because of the long Se-H bond distances (1.44 A in
ethanselenol1 3 6 compared to 1.09A for alkyl C-H bonds) and the inverse
sixth power dependence of the DD process on this distance (Eqn.[ 3])1"*0 .
2 ?
V Y
- [3]
r 6
rSeH
The lower gyromagnetic ratio for 77Se is also a factor and one may readily
calculate from Eqn. [3 ] that, for a single bound hydrogen, the ratio of
T! (DD) for 77Se to Ti (DD) for 13C will be 9.4. Low molecular weight
solutes have reorientational correlation times typically in the range of 1CT11
to 10" 13 sec /radian132 and for these conditions a single hydrogen 1.44A away
29
-------
will give rise to a 77Se TI (DD) between 325 and 4270 sec. Thus, because
the Ti(DD) in these compounds is on the order of hundreds of seconds it does
not compete favorably with the SR mechanism. The correlation times for proteins
are much longer137 (ca. 1CT8 to 10~7 sec/radian) in which case the 77Se Ti(DD)
will lie in the range of~0~. 3 to 2.0 sec. Under these conditions the Ti(DD)
should have a greater influence on the Ti(OBS).
The use of CDCla as solvent in these studies reduced the tendency of the
Se-H protons to undergo intermolecular exchange. Evidence in support of
this comes from the highly shielded chemical shift of the selenol protons
(5 =0.6 ppm in CDCls relative to TMS) and the sharpness of the lines as well
as the retention of the spin-coupling in both the 77Se spectrum and the XH
spectrum. lonization of this sort may be expected to take place in aqueous
solution, however, and considering that water is the most common biological
environment it was deemed of interest to examine some selenols in aqueous
solution to observe the effect on TI.
The simple alkane selenols are insoluble in water but dissolve in basic
solution. The TI for sodium methylselenolate (1.0 M in D2O, pD^lO) was
found to be 16.3 sec at 43°. The resonance is shifted considerably to higher
shielding compared to methaneselenol (-330 ppm relative to dimethyl selenide)
and is split into a well resolved quartet due to coupling with the methyl
hydrogens (JseHa6.6 Hz). The compound is apparently completely ionized
at pD = 10 with very little chemical exchange, as evidenced by the sharpness
of the lines and the shielded chemical shift.
A slightly more complicated situation was encountered with the amino
acid selenocysteine, HSeCHzCHfNHzJCOOH, which was prepared from the
reduction of selenocystine. A 0.5 M solution of this compound (pH^5, 50%
H2O/D2O) exhibited a broad resonance (v, /- = 150 Hz) centered at -141 ppm
relative to dimethylselenide . Proton-decoupling had no effect on the linewidth
which appears to be governed entirely by exchange effects. Unfortunately,
the low signal-to-noise ratio and sample limitations prevented an examination
of this compound in greater detail.
Instead, we investigated an analogous water soluble selenol, selenocys-
teamine, HSeCHaCHzNHa. The resonance for this compound is very pH dependent.
Its pKa has been determined to be 5.01128 and the graph of chemical shift
versus pH has the form of a typical weak acid titration plot (Fig. 10 ) . At low
pH the compound exists in the ammonium form 1^ while at high pH it is predominantly
HSeCH2CH2NH3
1 2
the Zwitterionic form 2 . The Ti for this compound (0.5 M in either H2O
or DzO) was determined to be 7.0 sec at pH 1.4 and 7.1 sec at pH 8.3 which
would indicate that the TiNs are the same for both forms 1 and 2. The presence
of chemical exchange is again evidenced by the breadth of the lines (v-, /2 = 13 Hz) .
From the graph it is apparent that the resonance for the RSeH species lies
approximately 150 ppm to lower shielding of the RSe form and dynamic exchange
between these sites could account for the observed decrease T2, which
30
-------
broadens the lines to the extent that any Hl-'^Se spin coupling is obscured.
The ionization process could also lead to an effect on T2 -through scalar
coupling relaxation but this mechanism is unlikely to affect T i since it may
be easily shown that, for the anticipated 1H-77Se one-bond coupling of 40-60
Hz, Ti(SC) will be negligible at the Larmor frequencies of the experiment.
Furthermore, the same TYs in both H zO and D 2O solutions serve to eliminate
the possibility that modulation of the scalar coupling ineraction affects T!.
Finally, it may be noted that the T! for this compound demonstrated a marked
sensitivity to dissolved oxygen. Prolonged exposure of an aqueous solution
to the solution to the atmosphere converts the compound to the diselenide,
selenocystamine, although the reaction requires several hours to reach com-
pletion128 . It was found, however, that for solutions in the pH range 4-8,
by simply opening the NMR tube to the atmosphere for 2 or 3 minutes, the
small amount of oxygen so introduced resulted in a change of T i from 7 sec.
to 0.1-0.3 sec.
In conjunction with this study we also prepared and investigated aqueous
solutions of hydrogen and ammonium selenide. The ammonium selenide solutions
were prepared on a vacuum line by co-dissolving hydrogen selenide and a
300% molar excess of anhydrous ammonia in D2O. The resonance for this sample,
at both 0.3 and 1.0 M, occurs at -511 ppm and is the most shielded selenium
resonance yet reported. The linewidth is broad (_V]/o = ^7 ^z^ no doubt due
to an exchange between the species HSe and Se2 , although as the shift would
imply the predominant form is probably solvated Se2 . Once again the breadth
of the resonance led to such low signal-to-noise that an accurate T i could
not be determined. A crude inversion recovery experiment indicated that
the TI is relatively long and certainly greater than 5 sec. Attempts to observe
a resonance for sodium selenide were fraught with difficulties as this compound
is almost totally insoluble in water.
Hydrogen selenide in water primarily H2Se and small amounts of HSe .
The sample exhibits a resonance at -288 ppm which may be compared with
that reported for neat H2Se, -226 ppm97' 9s. The TI'S for this sample were
found to be 0.9 and 0.7 sec at 10° and 34°, respectively. These values are
similar to those for the alkaneselenols and we tentatively attribute the slight
decrease in T i with an increase in temperature to a SR mechanism, although
the temperature dependence is not large and SR and CSA may well both be
contributing mechanisms.
Selenonium Compounds
Biologically important, naturally occurring forms of selenium in aqueous
solution are selenonium ions, RaSe+. Trimethylselenonium ion, for example,
is the normal excretory product arising from the ingestion and metabolism
of many forms of selenium: it constitutes about 20-50% of the selenium excreted
in the urine of rats fed a selenium diet 138. The resonance for a 1.0 M solu-
tion of trimethylselenonium iodide at 32° appears at 258 ppm and the TI
was found to be 13.7 sec. The material is insoluble in non-aqueous solvents
and so to aid the interpretation of the T! mechanisms, a more detailed analysis
of the selentine, (CH 3) 2Se(Br)CH 2CO 2C 2H 5 was undertaken. The Ti's for
a 0.5 M solution of this compound in CDCls are given in Table III . The values
of TI are relatively small and increase only slightly over the sixty degree
temperature range of the experiment. This behavior, along with the absence
of a measureable NOE, can be taken to indicate that CSA is an important me-
31
-------
chanism, although the temperature invariance suggests that CSA and SR may
both be contributing mechanisms for spin-lattice relaxation. The magnitudes
of the TI'S in aqueous solution are much greater than in the organic solvent
and the increase in T i up to 35° followed by a decrease suggests again that
both CSA and SR mechanisms are operative here. The greater rate of relaxation
in CDCla could reflect a lesser degree of mobility in the aqueous solution.
However, the resonance position shifts from 140 ppm in CDCls to 298 ppm
in DaO which indicates very different environments for the selenium in these
two solvents. The selenonium compounds are believed to be completely ionized
in aqueous solution1 ^ but some ion-pairing or even coordination of the bromide
is likely to occur in the organic solvent. In the latter case it is, therefore,
quite possible that the selenium is relaxing via a SC interaction with the bromine
nucleus. There are two magnetic nuclei of bromine, 79Br (I = 3/2,50.5% nat.
abundance) and 81Br(i= 3/2, 49.4% nat. abundance) and at the magnetic field
strength of these experiments, their Lannor frequencies are 24.89 and 26.83
MHz, respectively. Both of these nuclei have large electric quadrupole moments
and in an asymmetrical environment such as in the selentine, their relaxation
times, which are dominated by quadrupolar relaxation, will be on the order
of microseconds or less. For a selenium-bromine SC interaction, the 77Se Ti
is given by the expression x ° ^
S(S + 1) [4]
where J is the selenium-bromine scalar spin coupling constant, S is the bromine
nuclear magnetic quantum number, Tz is the spin-spin relaxation time of the
bromine nucleus, and Au> is the difference in Larmor frequencies for selenium
and bromine. Since the value of ( Au> Ta) 2 may conceivably approach unity in
the case of the selentine one can see that for reasonable estimates of J the
77Se Ti(SC) will be a few seconds or less. By analogy with the coupling of
77Se to nuclei such as 19F, 31P and 195Pt, wherein spin-spin coupling constants
between 500 and 1600 Hz have been observed ,9 7'1I|0one can anticipate a large
scalar coupling to 79Br and 81Br. In fact, by the judicious choice of J and
T2 one can reproduce on the basis of a SC relaxation alone the observed
77Se Ti and Ta of the selentine in CDCls. Further experimentation is required
to substantiate this argument but at this stage it should be pointed out that
SC relaxation is the most plausible explanation for the Ti's of the dialkylselenides.
Oxyacids
Inorganic salts of the oxyacids HaSeOs and H2SeOi, are common mineral
forms of this element. They are readily absorbed and metabolized by a number
of organisms, particularly the selenium accumulator plants, and likely to be
present in natural biological extracts ll*1. In addition, selenite ion is the
most prevalent means by which selenium is administered in animal tests.
Thus, an examination of some 77Se Ti's for this class of compounds seemed
32
-------
warranted.
It is apparent at the outset, given the polyprotic nature of these compounds,
that their solutions are unlikely to consist of only one discrete species. Differing
degrees of ionization are inevitable and for a given compound the composition
of the solution will depend on pH, concentration and temperature. Indeed, a
recent study103 of the pH dependence of the chemical shift of 77Se in aqueous
H2SeO3 demonstrated the selenium resonance was sensitive to various equilibria
and stepwise protonation or deprotonation. In our study, when selenous acid
was examined as a function of pH, concentration and temperature, the TI was
observed to vary sixfold over the range of conditions employed (Table III) .
The data thus serve to indicate that the TI sensitive to the changes in the average
environment of the selenous ion, a behavior that parallels that of 31P in phosphate
ion llt2. The complexity of the system renders an exact interpretation of the
data to be important relaxation mechanisms for the selenite ion. The possibility
of a SC mechanism from rapid exchange of the acidic protons is unlikely for
reasons presented earlier; it was also verified that the Ti's are virtually the
same in both H2O and DaO solutions. The temperature dependence of the TI
value for selenate ion (Table E ) resembles that for selenite ion but must be
regarded with the same discretion.
A very recent report by Koch, et^al^102 has appeared concerning 77Se
FT NMR studies of H2SeO3, Na2SeO7~and"NaHSeO 3 and NaaSeO^ in H2O. Their
NOE determinations are in complete agreement with this study in that, for H2SeO3
and NazSeCU, no NOE was observed. It is interesting to note, however, that
for a 4.0 molal solution of Na2SeO3, an enhancement of 0.4 was observed.
The T! measurements reported by Koch, e^al^.102 were not obtained under
comparable conditions and a direct comparison of the values is not possible.
For Na2SeOit (0.5 m, pH = 1-4) a Tx = 1-.2 sec. was obtained at ambient temperature
compared to our values of 6.5 sec (61°C) and 16.8 sec (15°C) using 0.5 M
Na2SeCU in D2Oat'pH = 6.6 For H2SeO3 (4.0 m, pH = 1-4) Koch, et al. re-
port102 a T! value of 1.1 sec (H2O) and 1.4 sec (D2O) at ambient temperature.
The concentration of H2SeOa in our study was held at either 1.0 M or 0.5 M
and for a 1.0 M H2SeO3 solution (pH = 1.5 in D2O) we obtained TI values of
2.8 sec (63°C) and 8.5 sec (10°C). Our Ti values are substantially longer
in all H2SeO3 samples studied. It is possible that the samples of Koch, et_ al.
contained dissolved oxygen which would shorten the TI values and it is also
known that at concentrations of ^4M or greater, pyroselenate ions, SezOs 2,
are formed. Thus, further studies of relaxation times of seleno oxyacids are
clearly needed.
The relaxation times of dimethylselenoxide and N ,N-dimethylselenourea
are included in Table in as further examples of multiply-bonded selenium. It
appears that SR is the dominant relaxation mechanism for 77Se in these cases.
The generality of this result awaits further experimentation, however, not
only because they are both low molecular weight examples of this class but
also hydrogen bonding, hydration and enolization are characteristic properties
of these functionalities and must be taken into consideration.
Summary and Conclusions
For the molecules studied in this work the spin-rotation and chemical
shift anisotropy mechanisms were found to be the most important means of
spin-lattice relaxation for the selenium-77 nucleus. The spin-rotation me-
chanism is very effective in reducing the Ti's of the small molecules but this
33
-------
mechanism requires very rapid molecular rotation and freedom of movement and
quickly becomes attenuated when the selenium atom is part of a large molecule.
Solvation and/or aggregation also tends to reduce the efficiency of SRand this
accounts for much of the difference in T i between organic and aqueous solutions
of these selenium compounds. With the exception of perhaps the selenols, this
mechanism will likely be unavailable for selenium-containing biopolymers.
It was, however, illustrated that in these systems the CSA mechanism, which
was evidenced in some of the larger molecules studied herein, will play an important
role in determining the TI.
The dipole-dipole mechanism, which is so beneficial to 13C NMR studies,
was found to be totally absent in the compounds studied. Thus, in spite of
a potentially large NOE, no enhancement was observed upon proton decoupling.
Th'is is a consequence of not only the dominance of the more efficient SR and
CSA mechanisms but also the marked inefficiency of the 1H~77Se DD mechanism.
The long Se-H internuclear distances, even for directly bound protons, account
for much of this ineffectiveness. This DD process has a strong dependence
on the molecular correlation time, however, and it was shown that in biological
macromolecules TI (DD) can be on the order of tenths of a second. These
considerations indicate that the study of such large molecules by 77Se FT
NMR will not be encumbered by exceptionally long recycle times. It may be noted
that under the conditions where DD relaxation becomes most effective (i.e.,
long correlation times) the maximum observable NOE is reduced to 1.1.
Prototropic ionization is a characteristic of several selenium compounds
of potential biological interest. It was illustrated that this property has a
definite effect on the o&served Tj, although secondary in the sense that it
does not contribute directly through a scalar coupling mechanism but controls
the nature of the species under investigation. It was beyond the scope of
the present study to provide a thorough interpretation of the data in terms
of all the relevant physicochemical aspects but the data serve adequately to
demonstrate the difficulties that may be encountered in analysing the 77Se
TI'S of ionizable compounds in aqueous solution. In a more optimistic vein,
this type of data could lead to an understanding of the pKa of the ionizable
group and information on the pH effects at a catalytic enzyme site containing
selenium. i* 3
-------
V. Recent Advances in Multinuclear NMR Techniques.
In the course of the research performed under the auspices of this grant
over the past two years, we have developed several innovative methods for
improving our instrumental capability to do multinuclear nmr experiments. One
facet of this research was the development of a multinuclear 18 mm nmr probe.
This account has been published (R.A. Byrd and P.D. Ellis, J. Magn. Resonance,
26, 169 (1977), and is briefly summarized below. Two other aspects of our
instrumental development research which are as yet unpublished deal with the
efficiency of heteronuclear decoupling and the development of a synthesizer
based modification of our XL-100 nmr spectrometer. The former has tremendous
significance in high field application of multinuclear nmr studies and is described
in detail following the 18 mm probe summary. The synthesizer based XL-100
facilitates the multinuclear magnetic resonance studies on any system, and it
will be briefly summarized here.
The synthesizer mode of operation was designed to minimize our dependence
upon Varian's rf observe related transmitters, local oscillators, and frequency
controllers. The age of our spectrometer (8 years) and the intensive use made
of it has resulted in an increasing amount of down time associated with these
components. Furthermore, the maze of back-plane wiring, the interconnecting
of these components, and their associated connections to the receiver and the
nuclei-select module is a continuing source of problems. This modification removes
our dependence upon these components.
Briefly, this project entailed the use of a General Radio synthesizer (GR-
1061) as the observe frequency source for the spectrometer. Of course one
must phase-lock the frequency synthesizer to the spectrometer. This was
accomplished by building a 15.4/15.3 MHz high spectral purity frequency synthe-
sizer that is phase-locked to the GR-1061. Thus, the output of this minisyn-
thesizer becomes the 2H-master clock for the spectrometer. The local oscillator
function is provided by a Hewlett-Packard 23OB tuned rf amplifier. A portion
of the observe rf is mixed with the 10.7 MHz if available within the spectrometer
and the desired mixing product, usually the first upper side band, is selected
by the tuned rf amplifier. The low level observe frequency is routed through
a network of rf switches, used for gating and phase alteration and then to a
broad-band power amplifier, the ENI-3100L. Hence, with this relatively "simple"
modification we have removed our dependency on over twenty rf modules and
several hundred feet of trouble prone cable. This modification should prove
to be very useful for any aging nmr spectrometer. A publication describing the
implementation of this approach will be submitted to the Journal of Magnetic
Resonance.
A. Development of a Multinuclear 18 mm Probe.
We would like to report the recent development of a probe for 18 mm spinning
sample tubes which is capable of operating over the frequency range of 10-100
MHz on a standard Varian XL-100-15 NMR spectrometer. The improvement in
signal-to-noise ratio achieved with this probe averages a factor of 3.4 over
the standard 12 mm tube throughout the operable frequency range. This
35
-------
corresponds to a reduction of approximately twelve in the time required for
a given experiment. Allerhand and co-workersll+1* have previously reported
a probe for 13 C FT nmr which utilized 20 mm sample tubes. Large sample
tube probes are now commercially available for the Bruker WH-180 and the
Varian XL-100 (using the new 4418 probe for 13C) . In order to facilitate the
investigation of various metal nuclides 96 and the more common nmr nuclides,
e.g. 2H, 13C, 31P, and 19F, in biological systems, we have developed a single
probe aimed at optimizing sensitivity subject to the constraint of minimizing
the amount of sample required. This probe requires no additional preamps
and is capable of performing any experiment already functional on a standard
XL-100 spectrometer, including heteronuclear decoupling of nuclei other than
protons, TI and T2 experiments, 2H, XH, and 19F lock for field stabilization.
The sensitivity improvements are measured relative to the identical experiment
performed in a 12 mm tube on our modified XL-100, which includes the improvement
attained through use of a crystal filter in a manner similar to that reported by
Allerhand, et^al^11"* . A brief description of the major design features of this
probe as well as some examples of its performance are presented here.
The 18 mm probe was constructed using the body of a Varian 4415 probe
and was incorporated into our XL-100-15 spectrometer. The probe design uses
the cross-coil configuration and interchangeable plug-in inserts for the receiver
coil. This feature provides tunability over the frequency range of 10-100
MHz in a manner analogous to the standard Varian XL-100 probes. The standard
transmitter matching networks are easily modified such that they tune both
our 4412 probe and the 18 mm probe for low power pulses, e.g. 100 W pulses.
The standard decoupler matching networks tune both probes without modification.
Due to the larger sample tube and the cross-coil configuration, a source of rf
pulse power on the order of 1-2 kW is desirable for optimal 90° pulse widths,
e.g. 20-30 ysec. We use an ENI Model 3100L side band amplifier, as a replacement
for the tuned 100 W amplifier boards in the XL-100, to drive a Heathkit Model
SB-220 amplifier. This combination produces approximately 1.6 kW of rf pulse
power which is continuously tunable over the range of 7-32 MHz. Also, the
Heathkit amplifier is easily modified to deliver 1 kW of rf power up to 40.5
MHz for observation of 31P. We have constructed our own high power matching
networks and attain representative 90° pulse widths of 54 ysec for 15N (10. 6 G),
22 ysec for 13 C(io.6 G) , and 30 ysec for 31P (4.8 G), all of which prove to
be quite adequate for observation of the respective nuclides at 23.5 kG. In
addition, we have co: structed the necessary components for a Varian V4420
kilowatt amplifier, in order that we may obtain approximately 900 W of rf pulse
power at 94 MHz. The proton rf field strength for decoupling is approximately
2.2 kHz (0.52 G) when the rf^power is 9 W. For studies of large biomolecules,
4-6 W is generally sufficient . The probe does not contain an upper dewar
assembly for wide range variable temperature control; however, the lower
dewar assembly and temperature controller of the standard 4415 probe are still
part of our 18 mm probe and allow us to operate over a range of 5-75°C.
Thermal stability of ±1°C is routinely achieved.
The spinner used on this probe was constructed in our machines shop and
is capable of spinning rates up to 40 Hz with no difficulties. We use precision,
thin walled 18 mm o.d. sample tubes supplied by Wilmad Glass Co., Inc., Buene,
New Jersey. Teflon vortex-preventing plugs are supplied by the same manufacturer.
Sample volume varies from 4-5 ml depending upon the insert and observe frequency
employed.
36
-------
We have encountered no problems with magnetic field inhomogeneity for
resonance frequencies of 40 MHz and below. Half-height linewidths equal to
those attainable on our 12 mm 4412 probe are routinely achieved with the 18
mm probe. However, for higher frequencies, e.g. 19F at 94 MHz, we have
observed approximately 0.5 Hz inhomogeneity broadening.
In Fig. 11 we show the natural-abundance 13C nmr spectrum of 10 mM aqueous
sucrose obtained on our 18 mm probe. This spectrum represents about 1 hr of
signal accumulation and has a rms S /N ratio of 13:1. We feel it is significant
that this spectrum using only 5 ml of sample represents slightly greater S/N
than that reported by Allerhand e_t al^144 using 12 ml of sample. Consequently,
the sensitivity per unit volume, whicFis a very important criteria when considering
nmr investigations of large biomolecules, is considerably higher in our probe.
Figures 12-14 demonstrate the sensitivity improvements attained for 19F, 31P and
113Cd with the 18 mm probe. The ability to observe metal nuclide resonances,
e.g. 113Cd, at low concentrations is important if one wishes to use nmr to direct-
ly probe the metal binding sites of metalloproteins 82 . The two spectra in
Fig. 16 are of 32 mM aqueous CdCl2; however, the lower spectrum was obtained
in 1 hr using an 18 mm sample tube whereas the upper spectrum was obtained,
using identical acquisition parameters, after 12 hrs of accumulation in a 12 mm
sample tube. These spectra indicate that detection of 113Cd resonances at
concentrations as low as 4 mM in isotopically enriched 113Cd (this equivalent to
32 mM in natural abundance) bound to a protein is quite practical using this
probe.
37
-------
B. Utilization of Chirp Frequency Modulation with 180°-Phase Modulation for
Heteronuclear Spin Decoupling.
Modulation of the decoupling rf with psuedo random noise as described by
Ernst ll*5 is now routinely used as a method for broadband heteronuclear spin
decoupling. However, there are some drawbacks to this technique. The main
problem arises because of the relatively high rf power levels that are necessary
to decouple all protons within a sample at 100 MHz, i.e. approximately 10 watts
of power. This situation can become critical when superconducting magnets
are employed, especially when high salt concentrations are employed with biological
samples. For such samples the dielectric heating due to the intense and continuous
rf decoupling fields is more efficient at the higher frequencies.
In an attempt to alleviate this experimental problem, Grutzner and Santini
have introduced a 180°-phase modulation scheme for heteronuclear spin decoupling.
They have shown that under the most common experimental situations that this
method is more efficient than noise decoupling. That is, with 10 watts of power
a reasonably uniform decoupling can be obtained over a bandwidth of approximately
1.3 kHz. By bandwidth we mean the range of decoupling frequencies that will
maintain the signal-to-noise ratio of a given nucleus, e.g. 13C, at a value of at
least 1/2 that of the on resonance coherent decoupling signal-to-noise. Although
their method does provide more uniform decoupling than noise decoupling, it
is clear that the achieved bandwidth is less than ideal for superconducting
magnets. Further, the required power levels still do not avoid the rf heating
problems.
Clearly what is needed is a decoupling method that has a bandwidth in
excess of 2KHz with lower power levels. Further, to have the most efficient
utilization of the available power the power spectrum should be uniform through-
out the decoupling bandwidth. A modulation scheme that meets this latter
requirement is a linear frequency modulation method, i.e. a Chirp . Originally
developed for radar applications. Chirp modulation has been shown to produce
an rf power distribution which is almost constant as a function of frequency
within the bandwidth of the Chirp, followed by a sharp drop off at frequencies
outside of the bandwidth lk7 . Furthermore, a Chirp pulse has been demonstrated
to be useful as a low rf power substitute for the usual high power short pulse
excitation of the nuclei in a Fourier transform (FT) nmr experiment . It
will be shown here, that Chirp frequency modulation in combination with a 180°-
phase modulation scheme is more efficient than phase modulation alone, and
further, that one can obtain decoupling bandwidths in excess of 2KHz with
significantly lower levels.
The coupled IS two-spin system has been analyzed by Anderson and Freeman11*9
for the case when two rf fields are being applied. One of these is the decoupling
rf field with its frequency near the Larmor frequency of the S nuclei, and the
other, the observing rf field, with its frequency near the Larmor frequency of
the I nuclei. In modern FT nmr experiments, the decoupling rf field is generally
strong and the observing rf pulse field averaged over its duty cycle is generally
weak. The Hamiltonian for this system can, therefore, be separated into two
parts H and H1 where the latter, the contribution from the observing rf pulse
field, can be treated as a small perturbation. The Hamiltonian is given by the
following expressions;
38
-------
H = -[oul + o)aS - 2nJ I • S + YTH9 (JY cosu9t - I sinu9t)
i Z S Z i\j r\, I 6 A £ y 9t)] [5]
o £ A £ y /•
and
H1 =-[YTH-(I cosu.t - I sincj.t) + Y H ( S cosu>,t - S sinu), t)] [6]
X 1 X i y i S * X J- y J.
where I is the spin of the I nuclei whose Larmor frequency is uw; S is the spin
of the S nuclei whose Larmor frequency is i.
In order to analyze the effect of amplitude and frequency modulation of the
decoupling rf field, we must assume that both H2 and 0)2 are time dependent.
Let us now transform the Hamiltonian to a rotating coordinate system defined
by the following transformation:
R = exp[-i u2(t) (Sz +Iz)t] [7]
The transformed Hamiltonian is given by:
HR = RHR~1 -i R R"1
= -(wj - u2(t) - to>2(t))Iz - (cog - u»2(t) - t u2(t))Sz
-2lrH'S-YIH2(t)Ix-Y8H2(t)Sx [8]
and
H1 R = RH'R'1 = -{yjHjI^cos (uj - u2(t))t - I sinjuj - a>2(t))t] +
+ YjjHj^cosUj - u)2(t))t - Sysin(o)1 - »2(t))t]} [B]
Since we are interested in heteronuclear decoupling, the weak coupling
approximation, |o>_ - , and assuming that |tco2(t) |«|
-------
where
- u2(t) - tu2(t) [12]
and
(t) = a) - u9(t) - toL(t) [13]
o o « Zi
At this point we no longer need to condider H'R, since it only involves
the sampling of the I magnitization. Hence, we will focus our attention of
HR. In order to bring the Hamiltonian in eq [ 10] to diagonal form, let us carry
out the transformation defined by:
V = exp[i e(m) Sy] [14]
where
6(m) - tan"1[ysH2(t)/(ns(t) - 27rjm) ] [15]
where m represents the eigenvalue of the operator Sz- Then the transformed
Hamiltonian has the following form:
Hv = -[(ns(t) - 2TTjm)2 + (YsH2)2]^Sz - ^(t)Iz
(t)/(fis(t) - 2TrJm)] + [y^t)^ (t)/(n,,(t) - 2irjm) 2] }
/($yt) - 2TrJm))2]"1 [16]
In order to make eq. [16] useful we must now consider explicit forms for fi(t) .
Chirp decoupling
Chirp modulation corresponds to a frequency modulation using a periodic
linear ramp. Thus, for one cycle of the Chirp, the frequency of the decoupling
rf field is given by:
tc te
w2(t) = u)Q + 2irrt for - ^^t < [17]
where 2irr is the rate of change of the angular frequency, given in rad/sec2,
and tc is the length of one cycle of the Chirp_in seconds. Using eq.[l7] and
eQ-[13]/ we get the following expression for & (t) :
s
n (t) = -2u9(t) - tu0(t) = -4irr [18]
S " Z
Substituting eq. [ 18] in eq. [16] , and making H2(t) independent of time,
the Hamiltonian with Chirp modulation is the following:
40
-------
- [(ng(t) - 2TrJm) + (YSH2) sz
-Sy 4TrrYsH2/[(ns(t) - 2TrJm)2 + (YSH2)2] [19]
If the coefficient of the S term in eq.[19] is small enough, it can be
neglected and the diagonalized Hamiltonian would have the form:
Hv = -nj(t)Iz - l(ng(t) - 2TrJm)2 + (YsH2)2]^Sz [20]
In order to find the conditions under which the coefficient of the S term
in eq.[19] can be neglected, we must carry out another diagonalizing transformation
defined by:
V1 = exp[i G^m) Sxl [21]
where
=tan {4irrYsH2/[ (ng(t) - 2TrJm) + (Y^) ] } [22]
Then, the z component of the transformed Hamiltonian will be:
2
(Hy )z = -flI(t)Iz - {(ns(t) - 2TrJm)+ (YSH2)
+[(4ur/YsH2)/(H- (ng(t) - 21rJm)2/(YsH2)2)]2}iSz
The eq. [20] will be valid under the following condition:
(YSH2)2 »{(4Trr/YsH2)/[l + (ng(t) - 2TrJm)2/(YsH2) 2] }2 [24]
however,
(4Trr/Y HJ/I1 + (fl (t) - 2TrJm) 2/(Y H?) 2] < 4Trr/YqH? [25]
o Z. S oti — oZt
since the minimum value for the expression in brackets in inequality [25] is
1.0. Then inequality [24] will be satisfied if the following inequality is satisfied:
(YQHJ2»(4in-/Y,H?)2 [26]
o Z« o £*
which reduces to:
(YsH2)2»47rr [27]
The Hamiltonian of eq[i9] cannot be completely diagonalized. However,
it can be shown that if inequality [27] is valid, then all further transformations
will lead to a further reduction in the value of coefficient of the Sy in eq.
[19]. Then, under the condition of inequality [ 27 ], eq[20] will be valid.
Substituting eq.[i6] in eq. [13] and eq.[i2] , and the result in eq.[20] ,
the following expression for the diagonalized Hamiltonian is obtained:
-------
- 4Trrt)Iz - A(m)Sz [28]
where
A(m) = [(o> - o + 4-nrt. For YsH2»2TrJ, with 1 = 2
and S = \, there are only two allowed transitions of frequencies:
«A = uj + i[A(i) - A(-i)] [30]
and
- A(-i)] [31]
[32]
Then, the residual coupling J_ is given by:
K
Replacing the values of A(2) and A(=i) from eq.[29] in eq.[32], the residual
coupling is given by:
uj) + (YsH2)]> [33]
The effect of Chirp modulation is similar to chemical exchange. That is,
the Chirp leads to a time dependent chemical shift term with eq. [25] . For
large enough Chirp frequencies, sharp resonances should be observed. Under
such conditions of large Chirp modulation frequencies, the observed residual
coupling will be given by the average of the residual coupling over one Chirp
cycle. This average coupling will be given by:
-------
(Jp) =~\2 Jp(t) dt
•** av T_ .1 K
r r t * ^^_j. T \ -l_ / O \ I
:r~ -[(oi - u) - 4trrt •+• irT) + (Y^HO) ] }dt [341
r\ t \ •*" o O —•>•.•. "Si SX L^-'-'-J
Evaluating the integral within eq. [34], the following expression is ob-
tained :
(Jp) = y—{A/ A2+ (YSHJ2 -B/"
K av 16Tr2rt S 2
c
-C/ C2+(YSH2)2 +D/ D2+(YSH2)2 +
, (A + / A2 + (YqH ) 2) (D + / D2 + (Y-HJ 2) ....
r HJ2ln[ M LJ ]}[35]
_ J--r— -T— —
(B + /BZ + (Y0Hj2) (C + /C2 + (Y-H,)2}
where A, B, C and D are given by the following expressions:
A = ^s ~ ^o " 27rrtc ~ ^
B = a) - a) + 2irrt - irj
SO C
C = oj - oj - 2-rrrt + irj
o O C
D = us - O)Q + 2irrtc + TrJ
Equation [35] will give the residual coupling provided that the Chirp modulation
frequency is high enough to give sharp lines and that inequality 1*7]is satisfied.
Square wave modulated decoupling
This type of modulation consists of a periodic phase reversal, that is,
the rf field is being phase modulated by a square wave so the rf decoupling
-------
field is described by:
H2(t) - -H2 -tc/2 <_t < 0 [36]
and
H2(t) = H2 O < t < tc/2 [37]
where t is the cycle time in seconds, and H2 is the magnitude of the rf field.
In eq. [16], fis(t) would be independent of time. However, H2(t) as defined by
eqs. [36]and [37] has a discontinuity for t = 0, making Ha(O) = ». Thus,
the analysis of Chirp decoupling cannot be used.
The method of analysis of Anderson and Freeman150 , based on a Fourier
series expansion of the modulating waveform requires low rf decoupling power
and a small modulation index. The function described by eqs.[36] and [ 37]is
certainly not one of small modulation index. Furthermore, we are presently
interested in high power rf decoupling fields. Therefore, the analysis provided
by Anderson and Freeman150 cannot be used for analysis of square wave
modulated decoupling. Hence, the theoretical conclusions reached by Grutzner
and Santini1"*9 should be considered as only zero-order approximations.
All spectra were obtained on a Varian XL-100-15 NMR spectrometer operating
at 25.2 MHz, in the FT mode. The decoupling frequency was generated using
a General Radio 1061 frequency synthesizer. The output frequency of the
synthesizer was partially controlled by an analog sawtooth function, in order
to create the Chirp. The average frequency was measured using a Hewlett
Packard 5328A universal counter, and the bandwidth of the modulation set
using an oscilloscope to match the extremes of the sawtooth modulating signal
to values predetermined using a DC voltage at the modulation port and measuring
the output frequency. Phase modulation was accomplished using a double
balanced mixer with a square wave at the modulation port. The DC offset
of the square wave was adjusted so that the absolute value of the first half
of the cycle was equal to that of the second half of the cycle, in order to
keep the rf amplitude constant throughout the cycle.
The decoupling coil matching network was tuned to less than 1% reflected
power as measured with a Bird Electronic Corp. model 4370 R.F. wattmeter.
The effective decoupling rf field was determined from residual splittings in
single frequency off resonance experiments, using the following equation:
TsH2/2^ = (Av2 - JR2/4)%2 - J2R)*/IR [38]
where Av is the difference between the proton resonance frequency and the
decoupling frequency in Hz. Equation[ 38] can be obtained by reformulating
the equation obtained for the residual coupling JR in the AX case from the
results given by Anderson and Freeman11*9 . For a decoupling power of 10
watts the value of the term ysH2/2Tr was determined to be 3300 Hz ± 100 Hz,
for a decoupling power of 4 watts it was 2300 Hz ± 100 Hz and for a decoupling
power of 2 watts it was 1640 Hz ± 50 Hz.
The sample used contained methyl formate and ^20% of deuterochloroform
for locking purposes. The frequency of the center of the formyl proton doublet
-------
was determined to ±1 Hz using single frequency decoupling with 0.5 W of rf
power, by changing the decoupling frequency until best decoupling was observed.
Chirp Decoupling
The residual coupling constant for the formyl carbon resonance of methyl
formate was measured for various Chirp rates and bandwidths. The results
are given in Table IV , along with the results predicted from eq.[35] .
Inequality [27], describes the conditions under which eq. [35] is valid.
It can be rewritten in the following form:
(Y0H9/2ir)2 »r/m [39]
S Lt
All experimental results of Table IV show good agreement with the results
of eq.[35] except for the case when the Chirp repetition frequency is 2.0
kHz and the bandwidth is of 10.0 kHz. For this case r/-rr= 6.4x 106 sec"2
and (YsH2/2ir)2 = 1.0 x 107 sec"2, and inequality[39]in not obeyed.
At Chirp repetition frequencies of 1.0 kHz or higher, the observed spectra
consisted of sharp lines, that is, the linewidth of the individual lines of the
observed doublet were the same as that for on resonance decoupling. At
lower Chirp rates, the resonances showed increased broadening especially when
the average decoupling frequency was very far from the center of the proton
doublet. At very low Chirp frequencies (less than 100 Hz), the 13C spectrum
disappeared completely!
None of these results can be considered as giving good decoupling. However,
these results are instructive in the understanding of the experiment using
Chirp and square wave modulation simultaneously, which will be discussed
subsequently. The best Chirp decoupling is observed when using the maximum
bandwidth allowed by inequality [ 39].
Chirp and Square Wave Modulation
It will be shown here that a combination of Chirp and square wave modulation,
applied simultaneously to the decoupling rf, decouples over a wider range than
with either type of modulation alone. As exploratory experiments, using
various ratios of Chirp to square wave frequencies, were carried out it became
quickly evident that the ratio of Chirp to_ square wave frequencies must bjs
four-tp-one in_ order to obtain good decoupling. This can be understood if
we look at the Fourier series: components of the modulated rf. A spectrum
analyzer was used to look at the frequency components of the resultant rf.
With Chirp modulation alone, the spectrum consisted of sidebands of equal
intensity going out from the center frequency VQ ± r c , where rtc is the
<£
bandwidth of the Chirp. These sidebands were equally separated, the separation
being equal to the modulating sawtooth frequency fc. Outside the range
vo ~ -n0-, the sidebands decreased in intensity very quickly.
With square wave modulation alone, the spectrum consisted of sidebands
with frequencies, v0 ± (2n + l)fs, where fg is the frequency of the square wave,
-------
and n is a positive integer. Since the Chirp is a frequency modulation, while
square wave phase modulation is equivalent to an amplitude modulation, square
wave modulation of the Chirp can be analyzed as a modulation of each of the
sidebands of the Chirp. Therefore, sidebands will be created at the following
frequencies:
v ± (nf + (2nr + l)f )
o c s
where n and n' are positive integers. Unless the ratio of frequencies is an
integer, a large number of unequal intensity sidebands will be generated within
the bandwidth of the Chirp.
At integral ratios, with f > f , the sideband separation will be that of the
Chirp alone. The same will be true if f It = 2 or fc/fs = 1. However, with
fc/fs = 4, the number of sidebands will DeSdoubled, their intensities being
equal to each other within the bandwidth of the Chirp since each line is the
sum of the first square wave modulation sideband of one Chirp sideband with
the second of another, the third of another and so on. With tclf - 3, the
number of sidebands will be 1.5 times the number of sidebands 01 the Chirp
alone for a fixed bandwidth. Careful examination will show that there will be
two different intensities of sidebands. One of these corresponds to the sum
of the odd square wave modulation sidebands of alternating Chirp sidebands,
and the other to the sum of the second sideband of one Chirp sideband with
the second of another and all higher even order sidebands of alternating Chirp
sidebands.
Any higher ratios of t /f will create a larger number of sidebands. However,
their intensities are not all equal since the first sideband of one Chirp cannot
coincide with the second of the next, therefore creating at least two different
intensities of sidebands. Analysis of the resulting rf modulated with both
Chirp and square wave, using a spectrum analyzer, showed that in fact only
with a ratio f /f =4, the sidebands within the bandwidth of the Chirp were
all of the same intensity.
In order to measure the effectiveness of the decoupling methods, the formyl
carbon resonance of methyl formate was used. This carbon should represent
a worse case due to its large value of JCH/ i-e- 226 Hz. The decoupling frequency
was changed until the observed carbon linewidth doubled. This would lead
to reduction in the signal-to-noise ratio of a factor of two. Hence, the resulting
decoupling bandwidth will correspond to the effective 3db points of the power
distribution. The linewidth with proton decoupling frequency exactly at the
center of the proton doublet was kept at 2.0 Hz. This means that the additional
broadening or residual coupling was of 2.0 Hz at the 3db point. The effective
decoupling bandwidth for the Chirp/square wave modulation scheme under
a variety of Chirp bandwidths and Chirp repetition frequencies are summarized
within Tables V (with 10 watts of power) and VI (with 4 watts and 2 watts
of power) . Further, Figure 15 depicts the results of a Chirp/square wave
decoupling experiment when the decoupling power is 10 watts. Figure 16summarizes
the same experiment with only 2 watts of power. Table VI summarizes the
effective decoupler bandwidth for decoupling experiments with square wave alone
for 10 and 4 watts of decoupling power. Figure 17 graphically demonstrates
the decoupling efficiency of the square wave method with 4 watts of decoupling
arid a square wave rate of 100 Hz.
-------
Table IV
Residual Coupling Constants for Chirp Decoupling
(a) Case! Cases IIOH Case E Casein JR(S.F.)(c)
AV {Hz) JR(calc.)(b)jR(dbs) jR(calc.)(b) jR(obs.) JR(dbs.)
100 5.9 5.5 2.2 4.0 2.6 6.7
200 11.8 12.0 4.3 9.0 5.5 12.5
300 17.7 17.3 6.4 12.5 7.5 19.5
500 29.4 31.0 10.8 20.0 11.8 33.3
1000 57.7 60.0 21.5 29.5 20.8 65.3
Case I: Chirp repetition frequency 1000 Hz, Chirp bandwidth 2.0 kHz.
Case H: Chirp repetition frequency 2000 Hz, Chirp bandwidth 10.0 kHz.
Case HI: Chirp repetition frequency 1000 Hz, Chirp bandwidth 10.0 kHz.
(a) A v = difference between the average decoupling frequency and the
center of the proton doublet.
(b) calculated using equation [35]
(c) results for single frequency decoupling
(d) the decoupling power for all these experiments was of 10 watts
(YsH2/2n = 3300 Hz).
Table V
Decoupling Range with Chirp and Square Wave
Modulation at 10 W of Decoupling Power
Chirp Freq. (Hz) Bandwidth (kHz) Decoupling range(a) (Hz)
790, 2.0 1600
400 2.5 1700
400 5.0 2100
1600 10.0 poor decoupling
800 10.0 2000
400 10.0 2200
(a) proton frequency range within which the residual coupling or broadening
of the observed carbon resonance is less than 2.0 Hz.
47
-------
Table VI
Decoupling Range with Chirp and Square Wave Modulation
at 4W(a)and 2VV(b)of Decoupling Power.
Chirp (Hz) freq. (kHz) Bandwidth (Hz) Decoupling range
400a 5.0 1800
400a 6.4 2000
400a 10.0 2200
800a 5.0 poor decoupling
348b 8.0 2200
404b 8.0 2100
500b 8.0 1100
(a) y H9/2TT = 3200 Hz
O L*
(b) TH/2^ = 1640 Hz
Table VII
Decoupling Range with Square Wave Modulation
Square wave rate (Hz) Decoupling power (watts) Decoupling range (Hz)
100 10 1400
202 10 1900
100 4 1300
200 4 600
40 2 500
50 2 700
60 2 600
-------
Examination of Tables V and VI shows that best decoupling was obtained
when the Chirp bandwidths was of 5.0 kHz or more, and when the Chirp
repetition frequency was of about 400 Hz. The reason for good decoupling
with large Chirp bandwidths can be understood from the theory and experiments
for Chirp modulation alone. Only at large bandwidths was the residual coupling
considerably reduced from that expected for single frequency decoupling.
Furthermore, if the condition of the inequality depicted by eq [35]is violated,
the spectra obtained showed very poor decoupling when both Chirp and square
wave modulation was used.
From Table VI the largest decoupling range at 10 W decoupling power was
of 1900 Hz for the square wave decoupling. From Table V , the largest decoupling
range at 10 W decoupling power for modulation with both Chirp and square wave,
is of 2.2 kHz, which is roughly a factor of 1.2 larger than the best decoupler
with square wave modulation alone.
From a comparison of Tables V and VH for Chirp rates of 400 Hz, the
results for 4 W of power are almost the same as the results at a decoupling power
of 10 W. However, with a Chirp bandwidth of 5.0 kHz, 4 W of decoupling power
and a Chirp repetition frequency of 800 Hz, conditions which do not violate
inequality [ 39], very poor decoupling was observed. At this Chirp repetition
frequency, the square wave rate, which is fixed by the four-to-one ratio, is
of 200 Hz. When square wave modulation alone was used, at a decoupling power
of 4W and rate of 200 Hz, the effective decoupling bandwidth was only 600 Hz.
This poor decoupling with square wave modulation alone may be the reason for
the poor decoupling observed when both Chirp and square wave modulation is used.
At 4 W of decoupling power, the largest decoupling range for modulation
with both Chirp and square wave is of 2200 Hz. The largest decoupling range
with square wave modulation alone, at this power level, was of 1300 Hz. Thus,
at 4 W of decoupling power, the decoupling range has been nearly doubled when
both Chirp and square wave modulation is used. Furthermore, at 2 W of
decoupling power, the largest decoupling range with Chirp and Square wave
was of 2000 Hz while the largest range with square wave alone was of only 700
Hz. At this power level the decoupling range with Chirp and square wave
modulation is a factor of three larger than with square wave modulation alone.
From these results, it would be anticipated that the same decoupling bandwidth
could be achieved on systems not employing a Faraday shield with only 1 to 2
watts of decoupling power.
Modification of commercial nmr spectrometers for decoupling with Chirp and
180°-phase modulation.
Generation of this decoupling scheme was done using a swept frequency
synthesizer as described in the experimental section. It would be more useful,
however, to use a simpler, less expensive method of generating the Chirp.
This can be done using a voltage controlled oscillator (VCO) with a phase
lock loop circuit which is locked to the 1.0 MHz reference used as a frequency
standard in many nmr spectrometers. With a low pass filter in the control
loop the modulation can be introduced as an additional voltage to the frequency
control input of the VCO. The loop circuit maintains the average frequency at a
fixed value.
-------
Such a circuit was developed here, as shown in Fig."18 using SN 74324 for
the VCO. The average frequency in this case is 5.0 MHz. For phase detection
this frequency is divided by a factor of 100, while the 1.0 MHz reference is
divided by a factor of 20. The 50 kHz signals go to the phase detector (MC 4044) .
The output of the phase detector goes through an integrator and a filter to
generate the control voltage for the VCO. A sawtooth wave is introduced
at this point through a summing network. Fig, 19 shows the circuit used to
generate the sawtooth wave and square wave, at exactly 1/4 of the frequency
of the sawtooth wave, to be used for the 180°-phase modulation.
While a frequency of 500 kHz could have been used for the phase detection,
it was found experimentally that 50 kHz provides a more stable output frequency.
Care must also be taken to isolate the VCO from the control circuits.
The 5.0 MHz output of this oscillator is mixed with a fixed frequency from
the original decoupler. The output is then I80°-phase modulated and filtered
to remove the component at a frequency 10.0 MHz from the desired decoupling
frequency. The output is then amplified in the normal way. Additional am-
plification may be needed at this stage.
The output of the 5.0 MHz oscillator was analyzed through a spectrum analyzer.
The error in the frequency of the modulated oscillator was of about ±50 Hz.
This uncertainty represents a reduction of about 100 Hz in the decoupling range,
which is a small reduction for a decoupling range of 2.0 - 2.2 kHz. This system
operated satisfactorily and is now routinely used in our laboratory.
A Non-trivial Application of the Modification
As an illustration of the utility of this method, the 19F decoupled 13C
spectrum of (CH3)2BC2F3 was obtained and is shown in Fig. 22 . The fluorine
spectrum of this compound consists of one multiplet at 6~79.6 ppm (relative to
CFC13) another at 6 -94.6 ppm and a third one at 5-184.5 ppm . At 94.1
MHz, the 19F spectrum is spread over a range of nearly 10.0 kHz. However,
the two multiplets to higher shielding are separated by 1.4 kHz, which is
certainly within the decoupling range of 2.2 kHz when Chirp and 180° phase
alternation are used at 4 watts of decoupling power under optimum Chirp rate
and bandwidth (2 watts would also be sufficient) . In order to also decouple
the deshielded resonance, a final 100% amplitude modulation with a sine wave
of 4.58 kHz was used. In this example 8 watts of rf power were used so that
there were 4 watts of decoupling power for each band (a reduction of the power
by a factor of 2 would have been sufficient to decouple all fluorines. The
Chirp rate used was 500 Hz and the Chirp bandwidth was 8 kHz.
As Fig. 20 illustrates, there are two broad resonances at 6.B ppm and 136.4
ppm, and one sharp resonance at 162.6 ppm relative to TMS. The reso*nance
atl36.'4 ppm can be assigned to the perfluorovinyl carbon bonded to boron. It
is directly bonded to boron, and is broadened by interaction with the nuclear
quadrupole of the boron nucleus. The resonance at 6.8 ppm can be assigned
to the methyl carbons. It is broad due to boron coupling, quadrupolar broadening
and coupling to the protons of the methyl group.
Perfluoromethylcyclohexane has a similar spread of 19F resonances 152 .
There are three groups of resonances. Two separated from each other by 1.4
kHz and the third 4.9 kHz away from the average of the first two as observed
at 84.64 MHz. Complete decoupling of all fluorine resonances was accomplished
by Ovenall and Chang 152 using noise modulation with a 100% amplitude modulation
50
-------
using a sine wave, in addition to the noise modulated output of a second synthesizer.
The total power used in these experiments as they have been reported was
of 30 watts! Block-Siegert shifts were also observed in those experiments and
were of the order of 4 kHz. Our present experiments for (CHa) 2B-C2F3/ using
only 8 watts of decoupling power, is estimated to have Block-Siegert shifts of
<1 kHz.
51
-------
VI. Summary
During the course of this research project we have investigated, in detail,
the 113Cd nmr of three metalloproteins; con-A, carboxypeptidase-A, and bovine
superoxide dismutase. Although we were not the first research group to publish
on the importance of this particular approach, it is evident from the progress
report that our studies are not merely reporting chemical shifts of new species.
But rather, our work on all three systems is providing new and significant
information to the biochemist with respect to the fundamental mode of action
of these proteins and the interelationship between the function of the proteins
and the metal associated with it.
Furthermore, we have been involved in an intensive effort to find other
nmr active spin 1/2 nuclei that could be exploited in biological applications.
Such a nucleus is 7'5e. Although extensive chemical shift information is
available for 77Se, little, if any, relaxation time information was available
before our efforts in this area of research. The spin lattice relaxation time
is a parameter of paramount importance in determining the overall utility of
a given nucleus in a biological application, that is, it is intimately related
to the signal-to-noise per unit time. To this end, we have carried out an
extensive study on the nature of the specific mechanism(s) of spin lattice
relaxation and the corresponding values of Ti . The systems that we have
investigated to date are; organoselenium compounds, RSeR1, selenols RSeH,
diselenides, RSeSeR1, selenates, SeO 4 ~2 , and selenocysteamine. As a result
of this research it is clear that 77Se nmr may be very useful to studying active
site sulfhydryls in sulfhydryl proteins.
In order to pursue these research topics in the most efficient fashion, it
was necessary for us to develop some novel modifications of our nmr instrumenta-
tion. One of these is a unique nmr probe, capable of spinning 18 mm nmr tubes,
decoupling at any frequency, and observing any nmr active nucleus. Further,
the most important design parameter was signal-to-noise on a 5 ml coil volume.
The net result is that the probe leads to a timesaving of approximately a factor
of twelve over conventional 12 mm nmr systems.
Finally, we have solved a rather severe experimental problem in nmr
spectroscopy. That is, the required power levels for efficient broad band
hetero-nuclear spin decoupling. This problem reaches critical proportions when
examining biological systems, e.g. proteins in high salt concentrations, on
nmr spectrometers employing superconducting magnets. The basic approach
employs a linear frequency modulation scheme, i.e. a "Chirp" followed by a
180° -phase modulation of the "Chirp". The relative rates of "Chirp" to 180°
-phase modulation must be kept in a 4 to 1 ratio for the method to succeed.
The net results is that uniform decoupling can be achieved over a range in
excess of 2kHz with only 2 watts of power. For those systems not employing
a Faraday shield, efficient decoupling can be affected with power levels
between 1 and 2 watts.
52
-------
VII.
References
1. B. Bergersson, R.E. Carter, and T. Drakenberg, J. Magn. Resonance,
28, 299 (1977).
2. I.M. Armitage, R.T. Pajer, A.J.M.S. Miterkamp, J.F. Chleborski, and
I.E. Coleman, J. Amer. Chem. Soc., 98, 5710 (1976).
3. J.L. Sudmeir and S.J. Bell , ibid, 99, 4499 (1977).
4. J.F. Chlebowski, I.M. Armitage and J.E. Coleman, J. Biol. Chem.
252, 7053 (1977).
5. D.B. Bailey, A.D. Cardin, P.D. Ellis, W.D. Behnke, J. Amer. Chem.
Soc., 100, 5263 (1978).
6. A.R. Palmer, P.D. Ellis, W.D. Behnke, D.B. Bailey, A.D. Cardin,
submitted to J. Amen.. Ghem. Soc.
7. A.R. Palmer, D.B. Bailey, P.D. Ellis, W.D. Behnke, and A.D. Cardin,
submitted to Proc. Nat. Acad. Sci.
8. G.L. Nicholson, Int. Rev. Cytol., 39, 90 (1974).
9. T.K. Choudhary, and S.K. Weiss, (eds), Adv. Exp. Med. Biol., 55
(1974).
10. Concanavalin A as a Tool, ed. H. Bittiger, (John Wiley and Sons, New
York. New York (1977).
11. M. Inbar, and Z. Sachs, Proc. Nat. Acad. Sci., U.S.A., 67, 1418(1969).
12. S.D. Douglas, R.M. Kamin, and H.H. Fridenberg, J. Immunol., 103,
1185-1195 (1969).
13. L. Stavy, A.J. Treves, and M. Feldman, Nature, 232, 56 (1971).
14. M.A. Leon, and H.J. Schwartz, Proc. Exp. Biol. Med., 131, 735 (1969).
15. I.J. Goldstein, C.E. Hollerman, and E.E. Smith, Biochemistry, 4, 876
(1965).
16. G.M. Edelman, B.A. Cunningham, G.N. Reekejr., J.W. Becker, M.J.
Waxdal, and J.S. Wang, Proc. Nat. Acad. Sci. U.S.A., 69. 2580 (1972).
17. K.D. Hardman, and C.L.F. Ainsworth,Biochemistry, LI, 4910-19 (1972).
18. J.W. Becker, G.N. Reekejr., J.L. Wang, B.A. Cunningham, and G.M.
Edleman, J. Biol. Chem., 250, 1513 (1975).
19. G.N. Reeke, Jr., J.W. Becker, and G.M. Edelman, J. Biol. Chem.,
1525 (1975).
53
-------
20. K. D. Hardman Adv. Exp. Med. Biol., 40, 103 (1973).
21. J.W. Becker, G.N. Reeke, Jr., B.Z. Cunningham, and G.M. Edelman,
Nature, 259, 406 (1976).
22. C. Nicolav, A.J. Kalb and J. Yariv, Biochem. Biophys. Acta, 194,71
(1969).
23. G.M. Alter, E.R, Pandolfino, D.J. Christie, and J.A. Magnuson, Biochemistry,
16, 4034 (1977).
24. J.J. Grimaldi and B.D. Sykes, J. Biol. Chem., 250, 1618 (1975).
25. C.F. Brewer, H. Sternlicht, D.M. Marcus, and A.P. Grollman, Proc.
Nat. Acad. Sci., U.S.A., 70, 1007 (1973).
26. C.F. Brewer, H. Sternlicht, D.M. Marcus and A.P. Grollman, Biochemis-
try, 12 4448 (1973).
27. J.J. Villafranca and R.E. Viola, Arch. Biochem. Biophys., 160, 465
(1974).
28. G.M. Alter and J.A. Magnuson, Biochemistry, _13, 4038 (1974).
29. B.H. Barber and J.P. Carver, J. Biol. Chem., 248, 3353 (1973).
30. J.J.Grimaldi and B.D. Sykes, J.Biol. Chem., 250, 1618 (1975).
31. E. Meirovitch and A.J. Kalb, Biochem. Biophys. Acta, 303, 258 (1972).
32. A.D. Sherry and G.L. Cottam, Arch. Biochem. Biophys., 156, 665 (1973).
33. B.H. Barber and J.P. Carver, Can. J. Biochem., 53,371 (1974).
34. R.D. Brown, III, C.F. Brewer and S.H. Koenig,Biochem. Biophys. Res.
Commun., 43, 660 (1977).
35. W.D. McCubbin, K. Oikawa and C.M. Kay, Biochem. Biophys. Res.
Commun., 43, 660 (1971) .
36. A.J. Kalb and A. Lustig, Biochem. Biophys. Acta, 168, 366 (1968).
37. W.D. McCubbin and C.M. Kay, Biochem. Biophys. Res. Commun., 44,
101 (1971). ~~
38. A.J. Kalb and A. Levitzki, Biochem. J., 109, 669 (1968).
39. M. Shoham, A.J. Kalb and I. Pecht, Biochemistry, 12, 1914 (1973).
40. A.D. Sherry, A.D. Newman and C.G. Gutz, Biochemistry, 14, 2191 (1973).
-------
41. C.S. Peters, R. Codrington, H.C. Walsh and P.D. Ellis, J. Mgn. Reson.,
11, 431 (1973).
42. D.D. Traficante, J.A. Simms, and M. Mulcay, J. Magn. Reson., 15
484 (1974).
43. A.G. Marshall, L.D. Hall, M. Hatton and J. Gallos, J. Magn. Reson.,
131, 392 (1974).
44. R.A. Byrd and P.D. Ellis, J. Magn. Reson., 26, 169 (1977).
45. R.A. Byrd, D. Sinclair, A.R. Garber and P.D. Ellis, manuscript in
preparation.
46. R.J. Kostelnik, and A.A. Bothner-By, J. Magn. Resonance, JL4, 141
(1974).
47. G.E. Maciel and M. Borzo, J. Chem. Soc. Chem. Commun., 394 (1973).
48. A.D. Cardin, P.D. Ellis, J.D. Odom, and J.W. Howard, J. Am. Chem.
Soc., 97, 1672 (1975).
49. R.A. Haberkon, L. Quz, W.O. Gillum, R.H. Holm, C.S. Zin and R.L.
Lord, Inorg. Chem., 15, 240 (1976).
50. B. Birgersson, R.F. Carter and T. Drakenberg, J. Magn. Reson. 28,
299 (1977).
51. B.B.L. Agrawal and I.J. Goldstein, Biochem. Biophys. Acta, 147, 262
(1967).
52. J. Yariz, A.J. Kalb and A. Levitzke, Biochem. Biophys. Acta, 165, 303
(1968).
53. J.L. Wang, B.A. Cunningham and G.M. Edelman, Proc. Nat. Acad. Sci.,
U.S.A. 70, 1625 (1971).
54. International Union of Pure and Applied Chemistry, Pure Appl. Chem.,
45, 219 (1976).
55. R.D. Brown, personal communication.
56. C.E. Richardson and W.D. Behnke, J. Mol. Biol., 102, 441 (1976).
57. D.B. Bailey, P.D. Ellis, and J.A. Fee, submitted to Proc. Nat. Acad.
Sci. (U.S.A.).
58. A.M.Michelson, J.M.McCord, and I. Fridovich (Eds.) "Superoxide and Superoxide
Dismutases" Academic Press, London, (1977).
55
-------
59. J.A. Fee, A.M. Michelson, J.M. McCord, and I. Fridovich (Eds.)
"Superoxide and Superoxide Dismutases" Academic Press, London, 173
(1977).
60. J.S.Richardson, K.A. Thomas, Bitt. Rubin, and D.C. Richardson,
Proc. Nat. Acad. Sci., 72, 1349 (1975).
61. J.S. Richardosn, K.A. Thomas, and D.C. Richardson, Biochem, Biophys.
Res. Commun, 63, 986 (1975).
62 B.P. Gaber, R.D. Brown, S.H. Koenig, and J.A. Fee Biochim. Biophys.
Acta, 271, 1 (1972).
63. D. Klug-Roth, I. Fridovich, and J. Rubin, J. Am. Chem. Soc., 95,
2786 (1973).
64. E.M. Fielder, P.B. Roberts, R.D. Bray, D.J. Lowe, G.N.Mautner,
G. Rotrlio. and L. Calabrese, Biochem. J. 139, 49 (1974).
65. J.A. Fee, and P.O. DiCorleto, Biochemisty, 12, 4893 (1973).
66. T.H. Moss and J.A. Fee ,Biochem. Biophys. Res. Commun. 615, 799
(1975).
67. M.E. McAdam, E.M.Fielder, F. Lavelle, L. Calabrese, D. Cocco, and
G. Rotrilio, Biochem. J. 167, 271 (1977).
68. W.E. Blumberg, J. Peisach, P. Eisenberger, and J.A. Fee, Biochemistry,
17_, 1842 (1978).
69. J.M.McCord and I. Fridovich, J.Biol. Chem. 244, 6049 (1969).
70. J.A. Fee, J. Biol. Chem., 248, 4229 (1973).
71. K.M. Beem, W.E. Rich, and K.V. Rajagopalan, J. Biol. Chem. 249,
7298 (1974).
72. O.H. Lowry, N.J. Rosebrough, A.L. Farr, and R.J. Randcoll, J. Biol.
Chem., 193, 265 (1951).
73. A.R. Garber, and P.O. Ellis, manuscript in preparation.
74. R.A. Byrd and P.D. Ellis, J. Magn. Reson., 2<6, 169 (1977).
75. R. Freeman and H.E.W. Hill, J. Chem. Phys., 54, 3367 (1971).
76. T.C. Farrar and E.D. Becker "Pulse and Fourier Transform NMR"
Acadmice Press, New York (1971).
77. "Interatomic Distances Supplement", Special Publication No. 18, Chemical
Society, London (1965).
56
-------
78. Reference 19 p 55.
79. a. D.B. Bailey, P.D. Ellis, A.D. Cardin, and W.D. Behnke, J. Araer.
Chem. Soc., 100, 5263:bA.R. Palmer, P.D. Ellis, W.D. Behnke, D.B.
Bailey, and A.D. Cardin, submitted to J. Amer. Chem. Soc.
80. G.M. Edelman, B.A. Cunningham, G.M. Reeke, Jr., J.W. Becker,
M.J. Waxdal, and J.L. Wang,Proc. Nat. Acad. Sci., 69, 2580 (1972).
81. G.N. Reeke, J.A. Hartsuck, M.L. Ludwig, F.A. Quiocho, R.A. Steitz,
and W.N. Kipscomb, Proc. Nat. Acad. Sci. (U.S.A.), 5Q, 2220 (1967).
82. J.L. Sudmeier and S.J. Bell, J. Am. Chem. Soc., 99, 4499(1977).
83. I.M. Armitage, A.J. M.Schoot-Uiterhamp, J.F. Chlebowski, and J.E.
Coleman, J. Magn. Res. 29, 375 (1978).
84. I.M. Armitage, R.T. Pajer, A.J.M.SchootJJiterkamp, J.F. Chlebowski,
and J.E. Coleman, J. Am. Chem. Soc., 98, 5710 (1976).
85. J.L. Sudmeier, personal communication.
86. K.K.Kannan B. Nostrano, K. Fridborh, S. Lougren, A. Ohilss and
M. Petef, Proc. Nat. Acad. U.S.A. 72, 51 (1975).
87. J.A. Fee and R.G. Briggs, Biochim. Biophys. Acta 400, 439 (1975).
88. J.A. Fee and R.L. Ward, Biochem. Biophys. Res. Commun. 7_1, 427
(1976).
89. E.K. Hodgson and I. Fridovich, Biochemistry, U, 5294 (1975).
90. J. Rabani, D. Klug-Roth, and J. Lilie, J. Phys. Chem., 77, 1169 (1973).
91. An excellent review of carboxyepetidase-A and its physical and chemical
properties has been published by M.L. Ludwig and W.N. Lipscomb, in
Inorganic Biochemistry, edited by G.I. Eichhorn, Elsevier, Chapter 15,
(1973).
92. B.L. Vallee and H. Neurath, J. Biol. Chem., 217, 253 (1955).
93. G.N. Reeke, J.A. Hartsuck, M.L. Ludwig, F.A. Quiocho, T.A. Stietz and
W.N. Lipscomb, Proc. Nat. Acad. Sci. U.S.A., 58, 2220,(1967);W.N. Lipscomb
J.A. Hartsuck, G.N.Reeke, F.A. Quiocho, Brokhaven Symp. Biol., 21,
24 (1968); W.N. Lipscomb, G.N. Reeke, J.A. Hartsuck, F.A. Quiocho,
and P.H. Bethge, Phil Trans Royal Soc., B257, 177 (1970).
94. R. Breslow and D.L. Wernick, Proc. Natl. Acad. Sci. U.S.A., 74,
1303 (1977).
57
-------
95. L.D. Byers and R. Wolfenden, Biochemistry, 12, 2070 (1973).
96.(a)D.L. Klayman and W.H.H. Gunther, Ed,, "Organic Selenium Compounds,
Their Chemistry and Biology", Wiley-Interscience, New York, 1973.
(b)R.A. Zingaro and W.C. Cooper, Ed., "Selenium", Van Nostrand-Reinhold,
New York, 1974.
97. (a)A.D. Cardin, P.D. Ellis, J.D. Odom and J.W. Howard, J. Am. Chem.
Soc., 97, 1672 (1975).
(b)M.A. Sens, N.K. Wilson, P.D. Ellis and J.D. Odom, J. Magn. Resonance,
19. 323 (1975).
(c) W.H. Dawson and J.D. Odom, J. Am. Chem. Soc., 99, 8352 (1977).
(d) D.B. Bailey, P.D. Ellis, A.D. Cardin and W.D. Behnke, J. Am. Chem.
Soc., 100, 5263 (1978).
98. T. Birchall, R.J. Gillespie and S.L. Vekris, Can. J. Chem., 43, 1672
(1965).
99. W.M. McFarlane and R.J. Wood, J. Chem. Soc., A, 1397 (1972).
IDO.(a)S. Gronowitz, I. Johnson and A.B. Hornfedlt, Chem. Scr., 3, 94 (1973).
(b)ibid., 8,8 (1975).
(c) A. Fredga, S. Gronowitz and A.B. Hornfeldt, ibid., 8, 15 (1975).
(d)S. Gronowitz, A. Konar and A.-B. Hornfeldt, Org. Magn. Resonance,
9, 213 (1977).
101. W-H. Pan and J.P. Fackler, J. Am. Chem. Soc., 100, 5783 (1978).
102. O.A. Gansow, W.D. Vernon and J.J.Dechter, J. Magn. Resonance,
32, 19 (1978).
103. W. Koch, O. Lutz and A. Noelle, Z. Naturforsch, 33a, 1025 (1978).
104. H. Kolshorn and H. Meier, J. Chem. Research (s), 338 (1977).
105.(a)J.R. Lyerla, Jr. and D.M. Grant in "International Review of Science.
Phys. Chem. Series", C.A. McDowell, Ed., Medical and Technical
Publishing Co., Bol. 4, Chap. 5 (1972).
(b)D. Shaw, "Fourier Transform NMR Spectroscopy", Elsevier, New York,
(1976).
106. (a)A. Abragam, "Principles of Nuclear Magnetism", Oxford University Press,
London, 1961.
58
-------
(b) E. Oldfield and A. Allerhand, J. Biol. Chem., 250, 6403 (1975).
(c)E. Oldfield, R.S. Norton and A. Allerhand, ibid., 250,6468 (1975).
(d)W.E. Hull and B.D. Sykes, J. Mol. Biol., 98, 121 (1975).
107. Personal com'munication,Varian Associates Applications Laboratory.
108.(a)R.L. Void, J.S.Waugh, M.P.Klein and D.E. PhelpsJ. Chem. Phys.,
48, 3831(1968).
(b)R. Freeman and H.D. W. Hill, ibid., 5_1, 3140 (1969).
109(a)J.L.Markley, W.J. Horsley and M.P.Klein, J. Chem. Phys., 55, 3604 (1971).
(b)G.C.McDonald and J.S. Leigh, J. Magn, Reson., 9, 358 (1973).
110. J.H. Noggle and R.E. Schirmer, "The Nuclear Overhauser Effect.
Chemical Applications" .Academic Press, New York, 1971.
lll.(a)M. Sass and D. Ziessow, J. Magn. Reson., 25, 263(1977).
(b)D. Canet, ibid., 23, 361(1976).
112. The choice of dimethylselenide as the 77Se chemical shift reference conforms
with that of McFarlane and Wood.9a The other common standard is seleno-
phene^which is reported to be 617 ppm deshielded relative to demethyl-
selenide. 101
113. G. Ayrey, D. Barnard and D.T. Woodbridge, J. Chem. Soc., 2089
(1962).
114. M.L. Bird and F. Challenger, J. Chem. Soc., 570 (1942).
115. G. Cohen, C.M Murphy, J.G. O'Rear, H. Rauner and W.A. Zisman,
ind. Eng. Chem., 45, 1766 (1953).
116. L. Brandsma and H.E. Wijers, Rec. Trav. Chim., 82,68 (1963).
117. G.E. Coates, J. Chem. Soc., 2839 (1953).
118. W.H.H. Gunther, J. Org. Chem., 31, 1202 (1966).
119. J. von Bruan, W. Teuffert and K. Weissbach, Liebigs Ann. Chem.,
472, 121 (1929).
120. H.J. Backer and W. van Dam, Rec. Trav. Chim., 54, 531 (1935).
121. T. Zincke and K. Fries, Liebigs Ann. Chem., 334, 342 (1904).
122. D.L. Klayman and T.S. Griffin, J. Am. Chem. Soc., 95, 197 (1973).
59
-------
123. P.G. Foster, Org. Syn., Coll. Vol. 3, 771 (1955).
124. E. Rebane, Arkiv. Kemi, 25, 363 (1966).
125. I.G.M. Campbell, R.C. Cookson and M.B. Hocking, J. Chem. Soc.,
2184 (1965).
126. R. Paetzold and G. Bochmann, Z. Anorg. Allg. Chem., 360, 293 (1968).
127. D.L. Klayman, J. Org. Chem., 30, 2454 (1965).
128. H. Tanaka, H. Sakurai and A. Yokoyama, Chem. Pharm. Bull. (Tokyo)
.18, 1015 (1970).
129. Reference 110 pages 34-37.
130. (a)C. Deverell, Mol. Phys., 18, 319 (1970).
(b)A.A. Maryott, T.C. Farrar and M.S. Molmberg, J. Chem. Phys.,
54, 64 (1971).
131(a)S.O. Chan and L.W. Reeves, J. Am. Chem. Soc., 96, 404 (1974).
(b)M. Bacon and L.W. Reeves, ibid., 95, 272 (1973).
(c)J.F. Hinton and R.W. Briggs, J. MAgn. Reson., 25, 379 (1977).
132. (a)Y.K.Levine, N.J.M.Birdsall, A.G. Lee, J.C. Metcalfe, P. Parrington
and G.C.K. Roberts, J. Chem. Phys., 6jO, 2890 (1974).
133.(a)A. Koma and S. Tanaka, Solid State Commun., W, 823 (1972).
(b)A. Koma, Stat. Sol., 57b, 299 (1973).
(c)M. Mehring, "NMR Basic Principles and Progress:, R. Diehl, E.Fluck,
and R. Kosfeld, eds., Springer Verlag, New York, 1976,Vol. 11, p. 191.
134. D. Doddrell, V. Glushko and A. Allerhand, J. Chem. Phys., 5_6, 3683
(1972).
135.(a)K.F. Kuhlmann and D.M. Grant, J. Am. Chem. Soc., 90, 7355 (1968).
(b)K.F. Kuhlmann, D.M. Grant and R.K. Harris, J. Chem. Phys., 52,
3439 (1970).
136. J.R. Durig and W.E. Bucy, J. Mol. Spectrosc., 64, 474 (1977).
137j[a)B.D. Sykes, Biochem. Biophys. Research Commun., 39, 508 (1970).
(b)B.D. Sykes, P.G. Schmidt and G.R. Stork, J. Biol. Chem., 245, 1180
(1970).
60
-------
138.(a)J.L. Byard, Arch. Biochem. Biophys., 130, 556 (1969).
(b) I.S. Palmer, D.D. Fisher, A.W. Halverson and O.E. Olson, Biochem.
Biophys. Acta, 177, 336 (1969).
(c)I.S. Palmer, R.P. Bonsalus, A.W. Halverson and O.E. Olson, ibid,
208, 260 (1970).
139. Reference l(a) pages 239-240.
140.(a)D.W. W. Anderson, E.A.V. Ebsworth, G.D. Meikle and D.W.H. Rankin,
Mol. Phys., 25, 381 (1973).
(b)W.M. McFarlane, Chero. Commun., 755 (1968).
141. Reference l(a) Chaps. XIII A-E.
142. T. Glonek, P.J. Wang and J.R. VanWazer, J. Am. Chem. Soc., 98,
7968 (1976).
143. Ref. 15, p. 58.
144. A. Allerhand, R.F. Childers, and E. Oldfield, J. Magn. Resonance,
Ij., 272 and references cited therein (1973).
145. R.R. Ernst, J. Chem. Phys., 45, 3845 (1966).
146. J.B. Grutzner and R.E. Santini, J. Magn. Res., IB, 173 (1975).
147. J.R. Kluander, A.C. Price, S. Darlington and W.J. Albersheim, Bell
System Technical Journal, XXXIX (4), 746 (1960).
148. J. Delayre and J.J. Dunand, Abstracts 17th Experimental NMR Conference.
149. R. Freeman and W.A. Anderson, J. Chem. Phys., 37, 85 (1962).
150. R. Freeman and W.A. Anderson, J. Chem Phys., 42, 1199 (1965).
151. E.J. Stampf and J.D. Odom, Journal of Organometallic Chemistry, 108,
1 (1976).
152. D.W. OvenaD and J.J. Chang, J. Magn. Res., 25, 361 (1977).
61
-------
VIII. Figures and Figure Captions
Figure Captions
Figure 1
(a) Cadmium-113 nmr spectrum of Concanavalin A to which 2.2 equivalents
of 113Cd C12 per monomer have been added to the previously metal-free
protein. The sample is 2.3 mM in Con A protomer. The buffer used in
these experiments was 0.2 M NaCl, 50 mM NaAc, pH = 5.2, with D 2O added
to provide an internal lock. The experimental conditions used to obtain
all spectra shown were: flip angle = 45°, recycle time = 0.4 sec,
spectral window = 10,000 Hz, number of data points collected = 1024.
These spectra required 70000 transients. A 5 kHz enlargement of the
spectra is displayed with Hz of line broadening for sensitivity en-
hancement. Resonances occur at 68, 46 and -125 ppm.
(b) Cadmium-113 spectrum of Con A containing 2.2 equivalents of 113Cd(II)
and one equivalent of Mn(II) . The resonances are at 68, 46, and -125 ppm.
(c) Cadmium-113 nmr spectrum of Con A containing 2.2 equivalents of 113Cd(II)
and one equivalent of Ca(II) . The large resonances occur at 68 and 41 ppm.
The small resonance is at 46 ppm.
Figure 2
(A) The chemical shift of the free 113Cd(II) resonance in a Cd:Cd Con A sample
vs chloride ion concentration. The sample contains 2.2 mM Con A protomer,
5.0 mM Cd(NO3)2, and .05 M sodium acetate buffer pH 5.2. Sodium Chloride
was added to give the desired Cl~ concentration.
(o) The chemical shift of 4 mM 113Cd(N03) 2 in the absence of protein but
in the same buffer as above.
Figure 3
(a) Cadmium-113 nmr spectrum of "unlocked" Con A containing two equivalents
of 113Cd(II). One resonance is observed at 68 ppm. The same experimental
conditions are used in Fig. 1 with a protomer concentration of 1.1 mM.
(b) Cadmium-113 nmr spectrum of 3a after the addition of Zn(II) . The resonance
occurs at 72 ppm.
(c) Cadmium-113 nmr spectrum of unlocked 2 Cd(II)-Con A in nitrate containing
buffer. The protomer concentration is 1.8 mM with the resonance occurring
at -12 ppm.
Figure U
(a) The cadmium-113 nmr spectrum of Con A containing 2 equivalents of 113Cd(II)
and an equivalent of methyl a-DOmanno-pyranoside. The sample is 1.6 mM
in Con A protomer. The resonances occur at 43 and -131 ppm.
(b) Cadmium-113 nmr spectrum of 2 Cd(II)-Con A for reference. The resonances
occur at 68, 43 and -125 ppm.
62
-------
Figure 5
(a) The 113Cd nmr spectrum on 2 Cd(II)-SOD. The spectrum was obtained
on a 5000 Hz spectral window using a 45° flip angle with a 0.8 second delay
between pulses. Five thousand transients were accumulated. The spectrum
shown is a 2000 Hz expansion of the transformed spectrum. One resonance
appears at 311 ppm.
(b) The 113Cd spectrum of 2 Cd(II)-2 Cu{II)-SOD ran under the same conditions
as (a).
(c) The 113Cd spectrum of 2 Cd(II)-2 Cu(I)-SOD obtained using the same
conditions as in (a). One resonance appears at 321 ppm.
Figure 6
The spectra obtained in a progressive saturation experiment on 2 Cu(II)-SOD.
The sequence (90°-T) was used where T included the data acquisition interval.
The spin lattice relaxation time was determined using a non-linear least
squares program.
Figure 7
The spectra obtained in determining the NOE enhancement for 2 Cd(II)-SOD.
The sequence (90°-PD)n was used where PD>; 7 TI. Fig. 2 (a) is proton coupled
spectrum and 2 (b) the proton decoupled spectrum.
Figure 8
Arrhenius plots of the temperature dependence of the observed relaxation
rates (R=l/Ti) of dialkylselenides.
Figure 9
Arrhenius plots of the temperature dependence of the observed relaxation
rates (Rnl/Ti) of alkaneselenols.
Figure 10
pH profile of the chemical shift of a 0.5 M solution of selenocysteamine in
DzO. The experimentally determined pK is indicated on the graph.
Figures 11-14
See figures.
Figure 15
The 1H decoupled 13C spectrum of methyl formate (formyl carbon) using
63
-------
Chirp and Square wave modulation at 10 watts of power (Chirp repetition
rate was 400 Hz and Chirp bandwidth was 10 kHz) . The left most signal
corresponds to an on resonance condition for the 1H decoupling. Each resulting
signal corresponds to a 100 Hz offset of the decoupler to higher frequency.
Figure 16
The 1H decoupled 13C spectrum of methyl formate (formyl carbon) using
Chirp and Square wave modulation with 2 watts of decoupling power (Chirp
repetition frequency equal to 300 Hz and a Chirp bandwidth of 6.0 kHz) .
The format of this figure is the same as that employed in Figure 2.
Figure 17
The -"-H decoupled 13C spectrum of methyl formate (formyl carbon) using
square wave modulation at 4 watts os decoupling power. The optimum modulation
rate of 100 Hz was employed. The format of this figure is the same as that
employed in Figure 2.
Figure 18
Circuit diagram for a frequency modulated 5.0 MHz oscillator.
Figure 19
Circuit diagram for a sawtooth function generator and synchronous square
wave at 1/4 of the sawtooth frequency.
Figure 20
The 19F decoupled 13C spectrum of perfluorovinyl dimethyl borane,
(CH 3) 2BC2F3. See text for chemical shifts.
64
-------
Figure 1
a
100
—T 1 1 1 r~
60 20 -20
-60 -100 -140
65
-------
Figure 2
o
o
66
-------
Figure 3
67
-------
Figure 4
(b)
68
-------
Figure 5
SUPEROXIDE DISIWASE
(c)
(B)
(A)
69
-------
Q
O
(f)
u
(D
70
-------
Figure 7
NOE EXPERIMENT (Cd2 + -SOD)
A) coupled
P) decoupled
(A)
71
-------
Figure 8
-1.50
-2.00
ex.
C -2.50
-3.00
-3.50
a Me2Se
o i - Pr 2 Se
•Bu2Se
Oct2Se
3.0
3.5
4.0 4.5
1/T X 103
5.0
72
-------
Figure 9
-0.5
-1.0
ae.
c
-1.5
-2.0
A MeSeH
• EtSeH
• DecylSeH
3.0
3.5 4.0 4.5
VT X 10'
5.0
73
-------
Figure 10
-250
pH
74
-------
Proton-decoupled natural-abundance 13C spectrum of aqueous sucrose at 32°C. This spectrum
represents 4096 accumulations using 90° rf pulses, 8192 points in the time domain, a spectral width of
4(KX) Hz, and a recycle time of 1.0 sec. Exponential multiplication equal to 1-Hz broadening was used
for this plot which is a 2000-Hz portion of the entire frequency sepectrum.
A
B
Rg. 12»'p spectra of trifluoroacetic acid (20 mA/ in 19F) at 94.1 MHz. Both spectra were recorded
using the same sample following a single 90 rf pulse, with a sweepwidthof 2000 Hzand0.3-Hz broad-
ening due to exponential multiplication. Spectrum (A) was recorded in a 12-mm lube and (B) was
recorded in an 18-mm tube.
B
****
A
Fig. 13MP spectra of 25-mAf aqueous potassium phosphate at 40.5 MHz. Both spectra were re-
corded using the same sample following a single 90 rf pulse, with a sweepwidth of 2500 Hz, 8192
points in the time domain, and 0.4-Hz broadening due to exponential multiplication. Spectrum (A)
was recorded in a 12-mm tube and (B) was recorded in an 18-mm tube.
75
-------
A
B
VWL/ iVv^/lV^V^1/)M^'f^\^^
Rq. 14 Neural-abundance "JCd spectra of 32-mA/ CdCI2 al 22.2 MHz. Both spectra were recorded
using 30 rf pulses, a 5000- Hz sweep width, XI 92 points in the time domain, and 0.8-H/ broadening due
to exponential muttipliealion. (A) Fifty thousand accumulations in a 12-mm tube. (B) Four thousand
accumulations in an IX-nim tube.
Figure 15
76
-------
0
3
01
i!
77
-------
Figure 18
1.0 n\i ivf. input
1N914
777 M
LED
78
-------
Figure 19
22k.n.
ioon
2.2 kn
i5on'
: f>
ioon
' 1 '
7 fTJ fj
L
L
L 3.4 >
7
-IS v
L
L
0 Q
7475
C Q
7400 V-
,-HSv
i.okn
>—*w o
-15V
2.2 k/X
«W
7400
i.okn
2.2k.n
—o-u
79
-------
Figure 20
80
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO. 2.
IPA-630/1 -79-035
TITLE AND SUBTITLE
30MPREHENSIVE PROGRESS REPORT FOR FOURIER TRANSFORM
MR OF METALS OF ENVIRONMENTAL SIGNIFICANCE
AUTHOR(S)
3aul D, Ellis and Jerome D. Odom
PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Chemistry
Jniversity of South Carolina
:olumbia, South Carolina 29208
. SPONSORING AGENCY NAME AND ADDRESS
tealth Effects Research Laboratory RTP, NC
Dffice of Research and Development
J.S. Environmental Protection Agency
Research Triangle Park, NC 27711
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
September 1979
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION! REPORT NO.
10. PROGRAM ELEMENT NO.
1EA615
1 1 . CONTRACT/GRANT NO.
Grant R804359
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA 600/11
. SUPPLEMENTARY NOTES
.ABSTRACT
Interactions of the metals cadmium and selenium with various biologically important
substrates were studied by nuclear magnetic resonance (NMR) spectroscopy. Cadmium-113
WR was used for a critical examination of three metalloproteins: concanavalin A,
Dovine superoxide dismutase and carboxypeptidase A. The NMR parameters of selenium-77
vere investigated, with a view to using this nucleus as a probe of active site
sulfydryl groups in proteins.
Several advances in NMR instrumentation were developed to further the aims of this
Jroject. One is a unique NMR probe, capable of spinning large (18 mm) NMR tubes,
iecoupling at any frequency, and observing any NMR-active nuclei. A decoupler modifi-
:ation, "Chirp" decoupling, was developed. This modification allows good experimental
-esults with approximately 1/10 the power required without modification.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
:ourier Analysis
:ourier Transformation
Juclear Magnetic Resonance
. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b. IDENTIFIERS/OPEN ENDED TERMS
19. SECURITY CLASS (This Report)
UNCLASSIFIED
20. SECURITY CLASS {This page/
UNCLASSIFIED
c. COSATI Field/Group
06BJ
21. NO. OF PAGES
91
22. PRICE
•A Form 2220-1 (Rey. 4-77)
PREVIOUS EDITION IS OBSOLETF
81
------- |