- Email: [email protected]
New applications of electron spin resonance to problems in biochemistry and pharmacology
Pergamon Press
Life Sciences Vol . 13, pp. 1299-1314, 1973 . Printed in Great Britain
NEW APPLICATIONS OF ELECTRON SPIN RESONANCE TO PROBLEMS IN BIOCHEMISTRY AND PHARMACOLOGY Colin F . Chignell Section on Molecular Pharmacology, Pulmonary Branch, National Heart and Lung Institute, Bethesda, Maryland
20014
(Received in final form 8 October 1973) Summary---Some of the more recent applications of spin labeling to problems in biochemistry, pharmacology and molecular biology are discussed .
Spin-labeling is a spectroscopic technique that employs stable organic radicals as probes or reporter groups for biological macromolecules (1-5) . The first free radical to be used as a spin label was the cation radical of the phenothiazine drug chlorpromazine (I) .
Ohnishi and McConnell studied the
binding of the chlorpromazine cation radical to shear-oriented DNA and found that the aromatic plane of the drug was oriented nearly perpendicularly
C1 CH 2CH2CH2N (CH 3 ) 2 I to the helix axis of the nucleic acid (6) .
Since the chlorpromazine free
radical was stable over a rather limited pH range McConnell and coworkers sought other spin labels .
In 1961 Hoffmann and Henderson reported the syn-
1299
Electron Spin Resonance
130 0
Vol. 13, No. 10
thesia of di-t-butyl nitroxide (DTBN) (II), a free radical that was stable in
aqueous solution over a wide range of temperatures and pH values (7) .
Rozantsev
and his coworkers (8) subsequently synthesized a large number of free radicals in which the nitroxide group was part of a heterocyclic ring system (III) . R
0
These nitroxidea provided McConnell and his coworkers with the starting materials from which they prepared more specific spin labels .
Subsequently
McConnell, Rassat and many others have synthesized a wide variety of compounds containing the nitroxide free radical (1-5,8,9) . ZSR Spectroscopy of the NitroxIJI radical The spin of the unpaired electron on the nitroxide group generates a magnetic moment that can interact with a strong applied magnetic field.
When
an unpaired electron is placed in a magnetic field it may exist in one of two energy states in which its magnetic moment is aligned either parallel or antiparallel to the direction of the field .
Transitions between these two energy
states can be induced by the application of electromagnetic radiation of the appropriate energy .
The relationship between the magnetic field strength (H,
and the required frequency (v) is given by
1301
Electron Spin Resonance
Vol. 13, No . 10
where h is Planck's constant ; g is the so-called g-value for the electron ; and
P
is the Bohr magneton, a fundamental constant for the electron .
indicates that, at resonance, to the magnetic field (4 H), varied . changed .
Equation (1)
the applied frequency (v) is directly proportional
so that ESR can be observed when either H or v is
For experimental convenience it is usual to keep v constant while H is Most commercial ESR spectrometers operate in . the microwave frequency
range of 9 z 10 9 Ha (or 9 G8s)
so that H is approximately 3300 gauss (G) .
The ESR spectrum of the nitroxide radical, when present at low concentration in a non-viscoua solvent, consists of three equally spaced lines of about the same height (Fig . 1) .
FIG. 1.
This triplet results from the interaction of
The ESR spectrum of a nitroxide radical in aqueous solution .
the magnetic moment of the unpaired electron with the magnetic moment of the 14N
nucleus .
In qualitative terms, the magnetic moment of the nitrogen nucleus
can be aligned parallel, anti-parallel or perpendicular to the applied magnetic field .
Since the magnetic field experienced by the electron is the sum of the
external magnetic field and the local contribution provided by the N14 -nucleus it follows that the electron experiences three different magnetic field values
Electron Spin Resonance
1302
Vol. 13, No. 10
each of which gives rise to an absorption line in the spectrum (Fig . 1) .
For
technical reasons the commercially available ESR instruments display the spectrum as its first derivative instead of the simple absorption spectrum familiar to optical spectroscopists .
The spectrum is characterized by three parameters :
(1) the hyperfine splitting (Ao ), i .e . the distance (G) between adjacent lines, (2) the so-called g-factor (g o ), i .e . the position of the center line in the magnetic field and (3)
the peak-to-peak linewidth (G) .
When a nitroxide radical is incorporated into a host crystal, its ESR spectrum is found to depend on the orientation of the radical with respect to the magnetic field (Fig . 2) .
The largest hyperfine splitting (Azz ) is observed
H //X
H //Y
v
%,-- H //Z
FIG . 2 . The ESR spectra of 4 ;4'-dimethyloxazolidine-N-oxyl derivative of acetone oriented in the crystal 2,2,4,4tetramethyl-1,3-cyclobutanedione (adapted from reference 24) . when the magnetic field is parallel to the nitrogen ?r-orbital (z-axis) (Fig . 3) . The g-values are also dependent on the orientation of the nitroxide group (Fig . 2) .
At angles that lie between the three principle axes, the ESR spectra are
intermediate between those shown in Fig . 2 .
When the crystal of nitrcxide is
dissolved in a solvent of low viscosity the motion of the free radical is so fast that its ESR spectrum appears as a triplet (Fig . 1) in which the hyperfine
Electron Spin Resonance
Vol. 13, No. 10
FIG. 3.
1303
The molecular coordinate system of the nitroxide group.
splitting (Ao ) and g-value (go ) are each the average of the values seen in Fig. 2. The ESR spectrum of the nitroxide radical is susceptible to (1) the polarity of the environment of the radical,
(2) the mobility of the radical and
(3) the orientation of the radical with respect to the applied magnetic field . The ES R spectrum of the nitroxide group is modified by the presence of other paramagnetic species .
Furthermore,
the nitroxide group can be chemically re-
duced by a variety of agents with the concomitant loss of its ESR signal .
All
these properties make the nitroxide radical almost an ideal reporter group . should be borne in mind, however,
It
that the bulkiness of the nitroxide radical
may cause a large steric perturbation of its local environment . The Nitroxide Groupas an Indicator of Molecular Motion When the molecular motion of a nitroxide radical in dilute solution is decreased by increasing the solvent viscosity the ESR lines of the free radical broaden and the spectrum becomes asymmetric is known as the rigid glass,
(Fig . 4) .
The limiting line shape
powder or polycrystalline spectrum of the nitroxide
radical.
This spectrum can be thought of as a simple sum of all spectra of
Fig . 2.
As a result, the splitting between the outermost peaks of the rigid
glass spectrum is 2AZZ , corresponding to the bottom spectrum of Fig. 3 .
The rigid
Vol . 13, No . 10
Electron Spin Resonance
1304
FIG . 4.
The effect of viscosity on the ESR spectrum of 2,2,6,6-tetramethylpiperidine-l-oxyl (5 x 10-4M) dissolved in glycerol (adapted from reference 24) .
glass spectrum is encountered whenever the spin label is randomly oriented and molecular motion is either absent or very slow on the ESR time scale ; that is, when 1' -1 (at 9 .5
« I Azz
GHz)
- A ax
where
'r
I ti
7 x 10 7 sec-1 and T _l<
gzz
I jMh
l ti 3 x 10 7 sec-1
is the rotational correlation time .
The ESR spectrum of the nitroxide group can often provide useful information on the mobility of a spin label probe .
For example, when human erythro-
cyte ghost membranes are reacted with a maleimide spin label (IV), their ESR
0 IV
Vol. 13, No. 10
Electron Spin Resonance
1505
spectrum reveals the presence of at least two populations of spin labels that differ in their mobilities .
The spectrum of one group resembles the rigid glass
spectrum of the nitroxide group (ef.Fig . 4) with a splitting of 59G between low and high field extreme (Fig . 5, lines 1 and 5) .
Fig . 5.
mobilized .
The ESR spectrum of human erythrocyte ghost membranes spin labeled with IV .
In contrast a second group
lines 2 and 4)
These spin labels are highly im-
has a fairly sharp three line (Fig . 5,
spectrum that is characteristic of a mobile nitroxide group .
The
center line in Fig . 5 (line 3) contains contributions from all of the spin labels .
Holmes and Piette have suggested (10)
that the highly immobilized spin
labels are attached to sulfhydryl groups that are buried deep within the membrane where their motion is restricted .
These authors further postulate that
the more mobile spin labels are attached to surface sulfhydryl groups .
In
contrast to IV,the iodoacetamide spin label V, labels only the surface sulfhydryl
V groups
(10) .
Holmes and Piette have found that when erythrocyte ghosts labeled
with V are treated with chlorpromazine, a highly immobilized population of spin labels appears in the spectrum .
They have suggested that chlorpromazine induces
1308
Electron Spin Resonance
Vol . 13, No. 10
a conformation change in erythrocyte membranes so that spin labels which are on the outside of the membrane move into the interior
(10) .
Spin-labeled analogs of atearic acid (VI) have proved to be very useful probes for both natural and artificial membranes . CH3 (CH2)m
C
(CR2 )
n
For example,
COOR
N = 0
0 VI when VI
(m-12, n-3, R-H)
is incorporated into erythrocyte ghost membranes its
ESR spectrum resembles the rigid glass spectrum of the nitroxide group with a splitting of 57G between the low and high field extrema (Fig . 6) .
Fig. 6.
In contrast when
The ESR spectra of two stearic acid spin labels bound to human erythrocyte ghost membranes .
VI. (m-1, n-lá, R-í) is incorporated into the ghost membranes it$ ESR spectrum (Fig . 6)
indicates a high degree of motional freedom.
Since it seems most likely
that both labels are oriented with their ionized carboxyl groups at the membrane interface these observations suggest that near its surface, the membrane has a highly ordered rigid structure, whereas the interior of the membrane is fairly fluid in nature .
Similar results have been observed for other membrane systems
both natural and artificial (11-13) . Spin labels have been employed to study the effect of various perturbanta on membrane systems.
For example, Hubbell and coworkers have made use of
1307
Electron Spin Resonance
Vol. 13, No . 10
nitrox de analogs of methyl stearate (VI mw5, n lO, R-CH 3 ) and 17 ß-hydroxy-5aandrostane to study the interaction of the local anesthetics benzyl alcohol and lidocaine with erythrocyte ghost membranes (14) .
BSR measurements indicated
that, at low concentrations, benzyl alcohol produced a fluidizing effect on the membrane .
However, at high
(lytic) benzyl alcohol concentrations, the spin
labels became highly immobilized.
Hubbell and .coworkers suggested that at the
lytic concentrations, benzyl alcohol uncovered protein spin label binding sites that were covert in the unperturbed membrane (14) . The Nitrotide Group as
a
Probe for Molecular Orientation
The 8SR spectrum of the nitroöde radical is sensitive to the orientation of the radical with respect to the applied magnetic field (Fig . 2) . of the stearic acid spin label VI hydrocarbon chain .
(m"5, n-10, R-H)
The z-axis
is parallel to the long
Hubbell and McConnell have shown that when this spin label
is incorporated into shear-oriented . canine erythrocytes there is a larger splitting when the magnetic field is oriented perpendicular to the surface of the red cells. R-H)
This suggests that the preferred conformation of VI
(m"5, n-10
is one in which its long hydrocarbon chain is oriented perpendicular to
the membrane surface.
Similar observations have been made with artificial
membrane systems such as oriented phospholipid multilagers .
In their experiment
with phosphatidylcholine multilayers Griffith and coworkers have reported (15) that as the nitroxide group is moved farther and farther away from the carboxyl head group of stearic acid, the difference between spectra recorded with the field parallel and perpendicular to the plane of the multilayers decreases until with VI
(m-1, ns14, R-H)
there is little difference between the two orientations
These observations indicate that while an ordered multilemellar arrangement of hydrocarbon chains exists at the surface, the interior of the multilayer is quite fluid .
Other experiments by Hsia and coworkers (16-18) employing both a
cholestane spin label and VI (m-10, ¢5, R-H) have shown that cholesterol causes an increase in the rigidity of egg lecithin multilayers .
Similar results have
Vol . 13, No . 10
Electron Spin Resonance
1308
been reported by Kroes et al . for erythrocyte membranes isolated from control and cholesterol fed guinea pigs
(19) .
In contrast to cholesterol,
the general
anesthetics such as chloroform and butane decrease the organization of lecithin or brain lipid multilayers at very low concentrations (20) .
Local anesthetics
such as procaine or tetracaine increase order at low concentrations but decrease order at high concentrations
(20) .
The Nitroxide Group as a Probe for Binding Site Polarity Solvent effects on the ESR spectrum of the nitroxide radical are characterized by changes in both Ao and g parameters . o
DTBN in water yields Ao - 16 .7
G and go - 2 .0056 while DTNB in hexane is characterized by Ao - 14 .8 g and go 2.0061.
An extensive study by Dodd et al . has shown that Ao decrease while go
increases as the solvent polarity is decreased (21) .
In line shape studies of
moderately immobilized spin labels, the solvent dependence can be troublesome because it is difficult to estimate the effects on the principle values of A and g.
However, when small, rapidly tumbling spin labels are used, the solvent
effects can be useful . Hubbell and McConnell have reported that when VII is diffused into an
0 VII aqueous phospholipid dispersion or a rabbit vague nerve in Ringer solution, the high field line is replaced by two lines (22) .
Similar observations have been
made by Jost and Griffith with the DTA61-myelin system (23) .
The relative in-
tensities of the two high field lines provides a measure of the amount of spin label in hydrocarbon and aqueous environments .
McConnell and coworkers have
made use of this fact to develop a method for the estimation of the fraction of lipid in a biological membrane that is in a fluid state (24) .
130 9
Electron Spin Resonance
Vol. 13, No . 10
Some Other Applications Involvinjt Nitroxide Spin Labels Membranes It is obvious from the foregoing discussion that spin labels have played an important role in determining the structure of membranes (see also 25,26) . More recently, McConnell and coworkers have used spin-labeled phosphatidylcholine analogs to measure some of the dynamic properties of membranes .
For example,
Kornberg and McConnell have estimated the rate of inside-outside transitions occurring in egg lecithin vesicles with the aid of spin label VIII (27) .
In
these experiments vesicles labeled with VIII were treated with ascorbic acid at 0° .
The ascorbic acid quickly reduced the spin labels present in the external CR 2 .O .CO(CH 2 )
14 CR3
CH3 (CH2) 4CO .OCH CH 2.0 .p0 .0 .CHCH' 2 0'
+1 3 N 1 CH 3
VIII monolayer thereby abolishing their paramagnetism.
Since ascorbic acid is a
highly polar molecule it could not penetrate the vesicles and reduce the spin labels present in the internal monolayer.
When the internally oriented labels
re-oriented toward the outside they were reduced with a half time of 6 .5 hrs . From this experiment Kornberg and McConnell were able to measure the rate of phospholipid flip-flop across the artificial membrane . Scandella and coworkers have estimated the rate of phospholipid lateral diffusion in rabbit sarcoplasmic reticulum using spin labels VIII and I% (28) . When vesicles are prepared from either VIII or I% nitroxide-nitroxd.d e interactions CH 2 .O .CO(CH 2) nCH 3 CH 3
C
(CH 2) 14 CO .0 .
CH 2
t CH 2 .O .p0 .0 CH2CH 2N(CH 3) 3 0
Vol . 13, No. 10
Electron Spin Resonance
131 0
reduce the ESR spectrum to a single line (5) .
When such vesicles are added to
a sarcoplasmic reticulum preparation patches of the spin label become incorporated into the membrane .
As the spin labels diffuse into the membrane their ESA
spectrum changes until eventually the three line pattern appears .
From an
analysis of the line shapes and their rate of change with time Scandella and coworkers were able to estimate that the diffusion constant, D, of the spin labels was 6 x 10 -8 cm 2/sec at 37° (28) .
Grant and McConnell have used this
same approach to examine phospholipid diffusion in the membrane of Acholeplasma laidlawii (29) . Topographical studies of binding sites Chignell and coworkers have studied the topographies of the active sites of several mammalian erythrocyte carbonic anhydrases by means of a series of spin-labeled aromatic sulfonamide (%) inhibitors (30-33) .
The active site of
d --->
carbonic anhydrase is a deep crevice at the bottom of which is a single zinc atom.
When an aromatic sulfonamide inhibitor binds to the active site the
sulfonamide group is directly coordinated to the zinc atom .
Chignell and co-
workers prepared a series of spin labeled sulfonamides in which the distance, d, between the sulfonamide group and the pyrrolidine ring of the spin label was varied .
It was found that when d was small the spin label was highly immobilized
when the inhibitor bound to the enzyme active site .
As d was increased the
mobility of the spin label also increased until eventually the free radical demonstrated little or no interaction with the active .
Using this technique it
was possible to determine that the active site of human erythrocyte carbonic anhydrase C was funnel-shaped and about 14A deep .
Similar studies with the
human B isozyme and bovine carbonic anhydrase B suggested that the active sites
Vol. 13, No. 10
Electron Spin Resonance
131 1
of these enzymes were the same shape as human carbonic anhydrase C but somewhat deeper .
This "molecular dipstick" approach was originally devised by Hsia and
Piette who used the technique to study the topography of hapten binding sites on rabbit anti-2,4-dinitrophenyl immunoglobulins (34) . Free radical assay technique Leute and coworkers have combined spin labeling with the immunoassay technique to produce a procedure for the rapid determination of morphine and other analogs in urine, saliva and other biological fluids (35) .
(XI)
These
workers prepared an antigen (XII) by coupling morphine to bovine serum albumin (BSA) .
Antibodies were then raised against the antigen in rabbits .
XI XII
When the
R-H R - -CH2C0-BSA
XIII
spin-labeled morphine analog XIII bound to the antibodies the ESR spectrum of the nitroxide group became broad and assymmetric, indicating a high degree of immobilization at the immunoglobulin binding site .
When morphine was added to
the spin label-immunoglobulin complex the spin label was displaced and its ESR spectrum reverted to the sharp three-line pattern.
Leute and coworkers were
able to estimate the concentration of free spin label by measuring the amplitude of the low field peak .
When this amplitude was plotted as a function of added
morphine a calibration curve was obtained from which it was possible to estimate the concentration of morphine in any biological sample .
Morphine substitutes
such as methadone and propoxyphene and unrelated drugs such as barbiturates and amphatemines were not recognized by the antibody .
Thus the technique is well
131 2
Electron Spin Resonance
suited for use in heroin treatment programs .
Vol. 13, No. 10
Chignell and Starkweather have
recently shown (36) that this same approach can also be used when other drug binding proteins such as enzymes are available. Prognosis Spin-labeling is a versatile spectroscopic technique that can be used to probe the structure of biologically important macromolecules . see increasing application
The future should
of spin-labeling to biological problems particularly
those that involve the various receptor proteins .
Finally, when used in con-
junction with immunoassay procedures, spin labels provide a rapid method for the detection and quantitation of drugs and other small molecules present in biological fluids . References 1.
H. M. McCONNELL and B . G. McFARLAND, Quant .
2.
C . L. HAMILTON and H. M. McCONNELL, StAuctu&at Chemí.b,tAy and Motecuta&
B .iotogy,
Rev . Biophyb . 3 91-96 (1970) .
p 115-149 W .H .Freeman, San Francisco (1968) .
3.
0 . H . GRIFFITH and A. S. WAGGONER, Accounte Chem . Re6 .
4.
1 . C. P. SMITH,
2 17-24 (1969) .
Hiotogieae Apptí,eaüons o6 EZectun Spin Reaonanee SpeetAo-
beopy, p 483 Wiley/Interscience, New York (1972) . 5.
P . JOST and 0 . H. GRIFFITH, Method& in PhaAmacotogy, Vol. 2, p . 223-276, Appleton-Century-Crofts (1972) .
J . Amen . Chem . Soc . 87 2293 (1965) .
6.
S. OHNISHI and H. M. McCONNELL,
7.
A . K. HOFFMAN and A. T . HENDERSON, J . Amen . Chem . Soc .
8.
E. G. ROZANTSEV,
8 3 4671-4672 (1961) .
Fnee NitAoxyt Radi.catA, Plenum Press, New York/London,
(1970) . 9.
A. RASSAT, Bute . Soc . Chien .
FAance 3273 (1965) .
10 .
D. E. HOLMES and L . H. PIETTE, J . PhaAm . Exp . Then.
11 .
W. L. HUBBELL and H. M. McCONNELL, (1969) .
17 3 78-84 (1970) .
Pnvc . Nat . Acad . Sci .
(USA) 63 20-27
Vol. 13, No. 10
12 .
Electron Spin Resonance
131 3
D . CHAPMAN, M. D . BARRATT and V. B. KAMAT, Si.ochí.m . Siophyb .
Acts. 17 3
154-157 (1969) . 13 .
H. SIMPKINS, E . PANRA and S. TAY, J . MembAane Bí.ó.1 . 5 334-344 (1971) .
14 .
W. L. HUBBELL, J. C. METCALFE, S. M. METCALFE and H . M. McCONNELL,
Siochi.m .
Si.ophye . Acts 219 415-427 (1970) . 15 .
P . JOST, L. J. LIBERTINI, V. C. HEBERT and 0. H. GRIFFITH, J . Motec . Siot . 59 77-98 (1971) .
16 .
J . C. HSIA, H. SCHNEIDER and I . C. P. SMITH, Síochím . Siophye . Acta 202 399-402 (1970) .
17 .
J. C. HSIA and J . M. BOGGS, Siochim .
Siophyb . Aeta 266 18-25 (1972) .
18 .
J. C. HSIA, R. A. LONG, E. E . HRUSKA and H. D. GESSER, Si.oclhim . Siophyb .
Acts 290 22-31 (1972) . 19 .
J. C. KROES, R. OSTWALD and A. KEITH, Siochim. Siophy6 . Acts 274 71-74 (1972) .
Chinkca 25 350-360 (1971) .
20 .
I. C. P . SMITH,
21 .
G. H. DODD, M. D. BARRATT and L. RAYNER, FBBS LettM 8 286-288 (1970) .
22 .
W. L. HUBBELL and H. M. McCONNELL, PAoc . Nat . Acad . Sci .
(USA) 61 12-16
(1968) . 23 .
0. H. GRIFFITH, L . J. LIBERTINI and G. B. BIRREL, J . Phy6 . 3425
24 .
Chem . 75 3417-
(1971) .
H. M. McCONNELL, K . L. WRIGHT and B. G. McFARLAND, Siochi.m . Sí.ophy6 . Red .
Commun . 4 7 273-281 (1972) . 25 .
P . JOST, A. S . WAGGONER and 0. H. GRIFFITH, StAuctuke and
Function o6
Si.otog.i¢a.1 MembAaaneA, p . 83-144, Academic Press, New York (1971) . 26 .
R. J. MELHORN and A . D. KEITH, Membrane Motecutan Siotogy, p . 192-227, Sinauer Associates, Stamford (1972) .
27 .
R. D. KORNBERG and H. M. McCONNELL, Siochem .
28 .
C. J. SCANDELLA, P . DEVAUX and H. M. McCONNELL, PAoc . Nat . Acad . Scí. . 69 2056-2060 (1972) .
10 1111-1120 (1971) . (USA)
Electron Spin Resonance
1314
29 .
C. W. GRANT
and
Vol. 13, No . 10
H. M. McCONNELL, Pnoc . Nat . Acad . Sc í. .
(USA)
70 1238-1240
(1972) . 30 .
C. F. CHIGNELL, L(.Se Scí. (1)
10 699-706 (1971) .
31 .
R. H. ERLICH, D. K . STABKWBATHER
and C.
F. CHIGNELL, Molec . PhaJCmacol .
9
61-73 (1973) . 32 .
C. F. CHIGNELL, D. K. STARKWEATHER
and
R. H. ERLICH, B.í.och,ím .
Ikophy6 . Acta
271 6-15 (1972) .
and
D. K. STARKWEATHER, Comp . B,í.ochim . Pharrmacol .
33 .
C. F. CHIGNELL
34 .
J. C . HSIA
35 .
R. K. LEUTE, E . F. ULLMAN, A. GOLDSTEIN
and
L . H. PIETTE, AAch Biochem . B .í.ophy4 .
and
. (in prees)
12 9 296-307 (1969) .
L. A. HERZENBERG, Natulce (New
Bdnl .) 236 93-94 (1972) . 36 .
C. F. CHIGNELL
and
D. K. STARKWEATHER, PhaJUnawIogy 8 368-376 (1972) .
Life Sciences Vol . 13, pp. 1299-1314, 1973 . Printed in Great Britain
NEW APPLICATIONS OF ELECTRON SPIN RESONANCE TO PROBLEMS IN BIOCHEMISTRY AND PHARMACOLOGY Colin F . Chignell Section on Molecular Pharmacology, Pulmonary Branch, National Heart and Lung Institute, Bethesda, Maryland
20014
(Received in final form 8 October 1973) Summary---Some of the more recent applications of spin labeling to problems in biochemistry, pharmacology and molecular biology are discussed .
Spin-labeling is a spectroscopic technique that employs stable organic radicals as probes or reporter groups for biological macromolecules (1-5) . The first free radical to be used as a spin label was the cation radical of the phenothiazine drug chlorpromazine (I) .
Ohnishi and McConnell studied the
binding of the chlorpromazine cation radical to shear-oriented DNA and found that the aromatic plane of the drug was oriented nearly perpendicularly
C1 CH 2CH2CH2N (CH 3 ) 2 I to the helix axis of the nucleic acid (6) .
Since the chlorpromazine free
radical was stable over a rather limited pH range McConnell and coworkers sought other spin labels .
In 1961 Hoffmann and Henderson reported the syn-
1299
Electron Spin Resonance
130 0
Vol. 13, No. 10
thesia of di-t-butyl nitroxide (DTBN) (II), a free radical that was stable in
aqueous solution over a wide range of temperatures and pH values (7) .
Rozantsev
and his coworkers (8) subsequently synthesized a large number of free radicals in which the nitroxide group was part of a heterocyclic ring system (III) . R
0
These nitroxidea provided McConnell and his coworkers with the starting materials from which they prepared more specific spin labels .
Subsequently
McConnell, Rassat and many others have synthesized a wide variety of compounds containing the nitroxide free radical (1-5,8,9) . ZSR Spectroscopy of the NitroxIJI radical The spin of the unpaired electron on the nitroxide group generates a magnetic moment that can interact with a strong applied magnetic field.
When
an unpaired electron is placed in a magnetic field it may exist in one of two energy states in which its magnetic moment is aligned either parallel or antiparallel to the direction of the field .
Transitions between these two energy
states can be induced by the application of electromagnetic radiation of the appropriate energy .
The relationship between the magnetic field strength (H,
and the required frequency (v) is given by
1301
Electron Spin Resonance
Vol. 13, No . 10
where h is Planck's constant ; g is the so-called g-value for the electron ; and
P
is the Bohr magneton, a fundamental constant for the electron .
indicates that, at resonance, to the magnetic field (4 H), varied . changed .
Equation (1)
the applied frequency (v) is directly proportional
so that ESR can be observed when either H or v is
For experimental convenience it is usual to keep v constant while H is Most commercial ESR spectrometers operate in . the microwave frequency
range of 9 z 10 9 Ha (or 9 G8s)
so that H is approximately 3300 gauss (G) .
The ESR spectrum of the nitroxide radical, when present at low concentration in a non-viscoua solvent, consists of three equally spaced lines of about the same height (Fig . 1) .
FIG. 1.
This triplet results from the interaction of
The ESR spectrum of a nitroxide radical in aqueous solution .
the magnetic moment of the unpaired electron with the magnetic moment of the 14N
nucleus .
In qualitative terms, the magnetic moment of the nitrogen nucleus
can be aligned parallel, anti-parallel or perpendicular to the applied magnetic field .
Since the magnetic field experienced by the electron is the sum of the
external magnetic field and the local contribution provided by the N14 -nucleus it follows that the electron experiences three different magnetic field values
Electron Spin Resonance
1302
Vol. 13, No. 10
each of which gives rise to an absorption line in the spectrum (Fig . 1) .
For
technical reasons the commercially available ESR instruments display the spectrum as its first derivative instead of the simple absorption spectrum familiar to optical spectroscopists .
The spectrum is characterized by three parameters :
(1) the hyperfine splitting (Ao ), i .e . the distance (G) between adjacent lines, (2) the so-called g-factor (g o ), i .e . the position of the center line in the magnetic field and (3)
the peak-to-peak linewidth (G) .
When a nitroxide radical is incorporated into a host crystal, its ESR spectrum is found to depend on the orientation of the radical with respect to the magnetic field (Fig . 2) .
The largest hyperfine splitting (Azz ) is observed
H //X
H //Y
v
%,-- H //Z
FIG . 2 . The ESR spectra of 4 ;4'-dimethyloxazolidine-N-oxyl derivative of acetone oriented in the crystal 2,2,4,4tetramethyl-1,3-cyclobutanedione (adapted from reference 24) . when the magnetic field is parallel to the nitrogen ?r-orbital (z-axis) (Fig . 3) . The g-values are also dependent on the orientation of the nitroxide group (Fig . 2) .
At angles that lie between the three principle axes, the ESR spectra are
intermediate between those shown in Fig . 2 .
When the crystal of nitrcxide is
dissolved in a solvent of low viscosity the motion of the free radical is so fast that its ESR spectrum appears as a triplet (Fig . 1) in which the hyperfine
Electron Spin Resonance
Vol. 13, No. 10
FIG. 3.
1303
The molecular coordinate system of the nitroxide group.
splitting (Ao ) and g-value (go ) are each the average of the values seen in Fig. 2. The ESR spectrum of the nitroxide radical is susceptible to (1) the polarity of the environment of the radical,
(2) the mobility of the radical and
(3) the orientation of the radical with respect to the applied magnetic field . The ES R spectrum of the nitroxide group is modified by the presence of other paramagnetic species .
Furthermore,
the nitroxide group can be chemically re-
duced by a variety of agents with the concomitant loss of its ESR signal .
All
these properties make the nitroxide radical almost an ideal reporter group . should be borne in mind, however,
It
that the bulkiness of the nitroxide radical
may cause a large steric perturbation of its local environment . The Nitroxide Groupas an Indicator of Molecular Motion When the molecular motion of a nitroxide radical in dilute solution is decreased by increasing the solvent viscosity the ESR lines of the free radical broaden and the spectrum becomes asymmetric is known as the rigid glass,
(Fig . 4) .
The limiting line shape
powder or polycrystalline spectrum of the nitroxide
radical.
This spectrum can be thought of as a simple sum of all spectra of
Fig . 2.
As a result, the splitting between the outermost peaks of the rigid
glass spectrum is 2AZZ , corresponding to the bottom spectrum of Fig. 3 .
The rigid
Vol . 13, No . 10
Electron Spin Resonance
1304
FIG . 4.
The effect of viscosity on the ESR spectrum of 2,2,6,6-tetramethylpiperidine-l-oxyl (5 x 10-4M) dissolved in glycerol (adapted from reference 24) .
glass spectrum is encountered whenever the spin label is randomly oriented and molecular motion is either absent or very slow on the ESR time scale ; that is, when 1' -1 (at 9 .5
« I Azz
GHz)
- A ax
where
'r
I ti
7 x 10 7 sec-1 and T _l<
gzz
I jMh
l ti 3 x 10 7 sec-1
is the rotational correlation time .
The ESR spectrum of the nitroxide group can often provide useful information on the mobility of a spin label probe .
For example, when human erythro-
cyte ghost membranes are reacted with a maleimide spin label (IV), their ESR
0 IV
Vol. 13, No. 10
Electron Spin Resonance
1505
spectrum reveals the presence of at least two populations of spin labels that differ in their mobilities .
The spectrum of one group resembles the rigid glass
spectrum of the nitroxide group (ef.Fig . 4) with a splitting of 59G between low and high field extreme (Fig . 5, lines 1 and 5) .
Fig . 5.
mobilized .
The ESR spectrum of human erythrocyte ghost membranes spin labeled with IV .
In contrast a second group
lines 2 and 4)
These spin labels are highly im-
has a fairly sharp three line (Fig . 5,
spectrum that is characteristic of a mobile nitroxide group .
The
center line in Fig . 5 (line 3) contains contributions from all of the spin labels .
Holmes and Piette have suggested (10)
that the highly immobilized spin
labels are attached to sulfhydryl groups that are buried deep within the membrane where their motion is restricted .
These authors further postulate that
the more mobile spin labels are attached to surface sulfhydryl groups .
In
contrast to IV,the iodoacetamide spin label V, labels only the surface sulfhydryl
V groups
(10) .
Holmes and Piette have found that when erythrocyte ghosts labeled
with V are treated with chlorpromazine, a highly immobilized population of spin labels appears in the spectrum .
They have suggested that chlorpromazine induces
1308
Electron Spin Resonance
Vol . 13, No. 10
a conformation change in erythrocyte membranes so that spin labels which are on the outside of the membrane move into the interior
(10) .
Spin-labeled analogs of atearic acid (VI) have proved to be very useful probes for both natural and artificial membranes . CH3 (CH2)m
C
(CR2 )
n
For example,
COOR
N = 0
0 VI when VI
(m-12, n-3, R-H)
is incorporated into erythrocyte ghost membranes its
ESR spectrum resembles the rigid glass spectrum of the nitroxide group with a splitting of 57G between the low and high field extrema (Fig . 6) .
Fig. 6.
In contrast when
The ESR spectra of two stearic acid spin labels bound to human erythrocyte ghost membranes .
VI. (m-1, n-lá, R-í) is incorporated into the ghost membranes it$ ESR spectrum (Fig . 6)
indicates a high degree of motional freedom.
Since it seems most likely
that both labels are oriented with their ionized carboxyl groups at the membrane interface these observations suggest that near its surface, the membrane has a highly ordered rigid structure, whereas the interior of the membrane is fairly fluid in nature .
Similar results have been observed for other membrane systems
both natural and artificial (11-13) . Spin labels have been employed to study the effect of various perturbanta on membrane systems.
For example, Hubbell and coworkers have made use of
1307
Electron Spin Resonance
Vol. 13, No . 10
nitrox de analogs of methyl stearate (VI mw5, n lO, R-CH 3 ) and 17 ß-hydroxy-5aandrostane to study the interaction of the local anesthetics benzyl alcohol and lidocaine with erythrocyte ghost membranes (14) .
BSR measurements indicated
that, at low concentrations, benzyl alcohol produced a fluidizing effect on the membrane .
However, at high
(lytic) benzyl alcohol concentrations, the spin
labels became highly immobilized.
Hubbell and .coworkers suggested that at the
lytic concentrations, benzyl alcohol uncovered protein spin label binding sites that were covert in the unperturbed membrane (14) . The Nitrotide Group as
a
Probe for Molecular Orientation
The 8SR spectrum of the nitroöde radical is sensitive to the orientation of the radical with respect to the applied magnetic field (Fig . 2) . of the stearic acid spin label VI hydrocarbon chain .
(m"5, n-10, R-H)
The z-axis
is parallel to the long
Hubbell and McConnell have shown that when this spin label
is incorporated into shear-oriented . canine erythrocytes there is a larger splitting when the magnetic field is oriented perpendicular to the surface of the red cells. R-H)
This suggests that the preferred conformation of VI
(m"5, n-10
is one in which its long hydrocarbon chain is oriented perpendicular to
the membrane surface.
Similar observations have been made with artificial
membrane systems such as oriented phospholipid multilagers .
In their experiment
with phosphatidylcholine multilayers Griffith and coworkers have reported (15) that as the nitroxide group is moved farther and farther away from the carboxyl head group of stearic acid, the difference between spectra recorded with the field parallel and perpendicular to the plane of the multilayers decreases until with VI
(m-1, ns14, R-H)
there is little difference between the two orientations
These observations indicate that while an ordered multilemellar arrangement of hydrocarbon chains exists at the surface, the interior of the multilayer is quite fluid .
Other experiments by Hsia and coworkers (16-18) employing both a
cholestane spin label and VI (m-10, ¢5, R-H) have shown that cholesterol causes an increase in the rigidity of egg lecithin multilayers .
Similar results have
Vol . 13, No . 10
Electron Spin Resonance
1308
been reported by Kroes et al . for erythrocyte membranes isolated from control and cholesterol fed guinea pigs
(19) .
In contrast to cholesterol,
the general
anesthetics such as chloroform and butane decrease the organization of lecithin or brain lipid multilayers at very low concentrations (20) .
Local anesthetics
such as procaine or tetracaine increase order at low concentrations but decrease order at high concentrations
(20) .
The Nitroxide Group as a Probe for Binding Site Polarity Solvent effects on the ESR spectrum of the nitroxide radical are characterized by changes in both Ao and g parameters . o
DTBN in water yields Ao - 16 .7
G and go - 2 .0056 while DTNB in hexane is characterized by Ao - 14 .8 g and go 2.0061.
An extensive study by Dodd et al . has shown that Ao decrease while go
increases as the solvent polarity is decreased (21) .
In line shape studies of
moderately immobilized spin labels, the solvent dependence can be troublesome because it is difficult to estimate the effects on the principle values of A and g.
However, when small, rapidly tumbling spin labels are used, the solvent
effects can be useful . Hubbell and McConnell have reported that when VII is diffused into an
0 VII aqueous phospholipid dispersion or a rabbit vague nerve in Ringer solution, the high field line is replaced by two lines (22) .
Similar observations have been
made by Jost and Griffith with the DTA61-myelin system (23) .
The relative in-
tensities of the two high field lines provides a measure of the amount of spin label in hydrocarbon and aqueous environments .
McConnell and coworkers have
made use of this fact to develop a method for the estimation of the fraction of lipid in a biological membrane that is in a fluid state (24) .
130 9
Electron Spin Resonance
Vol. 13, No . 10
Some Other Applications Involvinjt Nitroxide Spin Labels Membranes It is obvious from the foregoing discussion that spin labels have played an important role in determining the structure of membranes (see also 25,26) . More recently, McConnell and coworkers have used spin-labeled phosphatidylcholine analogs to measure some of the dynamic properties of membranes .
For example,
Kornberg and McConnell have estimated the rate of inside-outside transitions occurring in egg lecithin vesicles with the aid of spin label VIII (27) .
In
these experiments vesicles labeled with VIII were treated with ascorbic acid at 0° .
The ascorbic acid quickly reduced the spin labels present in the external CR 2 .O .CO(CH 2 )
14 CR3
CH3 (CH2) 4CO .OCH CH 2.0 .p0 .0 .CHCH' 2 0'
+1 3 N 1 CH 3
VIII monolayer thereby abolishing their paramagnetism.
Since ascorbic acid is a
highly polar molecule it could not penetrate the vesicles and reduce the spin labels present in the internal monolayer.
When the internally oriented labels
re-oriented toward the outside they were reduced with a half time of 6 .5 hrs . From this experiment Kornberg and McConnell were able to measure the rate of phospholipid flip-flop across the artificial membrane . Scandella and coworkers have estimated the rate of phospholipid lateral diffusion in rabbit sarcoplasmic reticulum using spin labels VIII and I% (28) . When vesicles are prepared from either VIII or I% nitroxide-nitroxd.d e interactions CH 2 .O .CO(CH 2) nCH 3 CH 3
C
(CH 2) 14 CO .0 .
CH 2
t CH 2 .O .p0 .0 CH2CH 2N(CH 3) 3 0
Vol . 13, No. 10
Electron Spin Resonance
131 0
reduce the ESR spectrum to a single line (5) .
When such vesicles are added to
a sarcoplasmic reticulum preparation patches of the spin label become incorporated into the membrane .
As the spin labels diffuse into the membrane their ESA
spectrum changes until eventually the three line pattern appears .
From an
analysis of the line shapes and their rate of change with time Scandella and coworkers were able to estimate that the diffusion constant, D, of the spin labels was 6 x 10 -8 cm 2/sec at 37° (28) .
Grant and McConnell have used this
same approach to examine phospholipid diffusion in the membrane of Acholeplasma laidlawii (29) . Topographical studies of binding sites Chignell and coworkers have studied the topographies of the active sites of several mammalian erythrocyte carbonic anhydrases by means of a series of spin-labeled aromatic sulfonamide (%) inhibitors (30-33) .
The active site of
d --->
carbonic anhydrase is a deep crevice at the bottom of which is a single zinc atom.
When an aromatic sulfonamide inhibitor binds to the active site the
sulfonamide group is directly coordinated to the zinc atom .
Chignell and co-
workers prepared a series of spin labeled sulfonamides in which the distance, d, between the sulfonamide group and the pyrrolidine ring of the spin label was varied .
It was found that when d was small the spin label was highly immobilized
when the inhibitor bound to the enzyme active site .
As d was increased the
mobility of the spin label also increased until eventually the free radical demonstrated little or no interaction with the active .
Using this technique it
was possible to determine that the active site of human erythrocyte carbonic anhydrase C was funnel-shaped and about 14A deep .
Similar studies with the
human B isozyme and bovine carbonic anhydrase B suggested that the active sites
Vol. 13, No. 10
Electron Spin Resonance
131 1
of these enzymes were the same shape as human carbonic anhydrase C but somewhat deeper .
This "molecular dipstick" approach was originally devised by Hsia and
Piette who used the technique to study the topography of hapten binding sites on rabbit anti-2,4-dinitrophenyl immunoglobulins (34) . Free radical assay technique Leute and coworkers have combined spin labeling with the immunoassay technique to produce a procedure for the rapid determination of morphine and other analogs in urine, saliva and other biological fluids (35) .
(XI)
These
workers prepared an antigen (XII) by coupling morphine to bovine serum albumin (BSA) .
Antibodies were then raised against the antigen in rabbits .
XI XII
When the
R-H R - -CH2C0-BSA
XIII
spin-labeled morphine analog XIII bound to the antibodies the ESR spectrum of the nitroxide group became broad and assymmetric, indicating a high degree of immobilization at the immunoglobulin binding site .
When morphine was added to
the spin label-immunoglobulin complex the spin label was displaced and its ESR spectrum reverted to the sharp three-line pattern.
Leute and coworkers were
able to estimate the concentration of free spin label by measuring the amplitude of the low field peak .
When this amplitude was plotted as a function of added
morphine a calibration curve was obtained from which it was possible to estimate the concentration of morphine in any biological sample .
Morphine substitutes
such as methadone and propoxyphene and unrelated drugs such as barbiturates and amphatemines were not recognized by the antibody .
Thus the technique is well
131 2
Electron Spin Resonance
suited for use in heroin treatment programs .
Vol. 13, No. 10
Chignell and Starkweather have
recently shown (36) that this same approach can also be used when other drug binding proteins such as enzymes are available. Prognosis Spin-labeling is a versatile spectroscopic technique that can be used to probe the structure of biologically important macromolecules . see increasing application
The future should
of spin-labeling to biological problems particularly
those that involve the various receptor proteins .
Finally, when used in con-
junction with immunoassay procedures, spin labels provide a rapid method for the detection and quantitation of drugs and other small molecules present in biological fluids . References 1.
H. M. McCONNELL and B . G. McFARLAND, Quant .
2.
C . L. HAMILTON and H. M. McCONNELL, StAuctu&at Chemí.b,tAy and Motecuta&
B .iotogy,
Rev . Biophyb . 3 91-96 (1970) .
p 115-149 W .H .Freeman, San Francisco (1968) .
3.
0 . H . GRIFFITH and A. S. WAGGONER, Accounte Chem . Re6 .
4.
1 . C. P. SMITH,
2 17-24 (1969) .
Hiotogieae Apptí,eaüons o6 EZectun Spin Reaonanee SpeetAo-
beopy, p 483 Wiley/Interscience, New York (1972) . 5.
P . JOST and 0 . H. GRIFFITH, Method& in PhaAmacotogy, Vol. 2, p . 223-276, Appleton-Century-Crofts (1972) .
J . Amen . Chem . Soc . 87 2293 (1965) .
6.
S. OHNISHI and H. M. McCONNELL,
7.
A . K. HOFFMAN and A. T . HENDERSON, J . Amen . Chem . Soc .
8.
E. G. ROZANTSEV,
8 3 4671-4672 (1961) .
Fnee NitAoxyt Radi.catA, Plenum Press, New York/London,
(1970) . 9.
A. RASSAT, Bute . Soc . Chien .
FAance 3273 (1965) .
10 .
D. E. HOLMES and L . H. PIETTE, J . PhaAm . Exp . Then.
11 .
W. L. HUBBELL and H. M. McCONNELL, (1969) .
17 3 78-84 (1970) .
Pnvc . Nat . Acad . Sci .
(USA) 63 20-27
Vol. 13, No. 10
12 .
Electron Spin Resonance
131 3
D . CHAPMAN, M. D . BARRATT and V. B. KAMAT, Si.ochí.m . Siophyb .
Acts. 17 3
154-157 (1969) . 13 .
H. SIMPKINS, E . PANRA and S. TAY, J . MembAane Bí.ó.1 . 5 334-344 (1971) .
14 .
W. L. HUBBELL, J. C. METCALFE, S. M. METCALFE and H . M. McCONNELL,
Siochi.m .
Si.ophye . Acts 219 415-427 (1970) . 15 .
P . JOST, L. J. LIBERTINI, V. C. HEBERT and 0. H. GRIFFITH, J . Motec . Siot . 59 77-98 (1971) .
16 .
J . C. HSIA, H. SCHNEIDER and I . C. P. SMITH, Síochím . Siophye . Acta 202 399-402 (1970) .
17 .
J. C. HSIA and J . M. BOGGS, Siochim .
Siophyb . Aeta 266 18-25 (1972) .
18 .
J. C. HSIA, R. A. LONG, E. E . HRUSKA and H. D. GESSER, Si.oclhim . Siophyb .
Acts 290 22-31 (1972) . 19 .
J. C. KROES, R. OSTWALD and A. KEITH, Siochim. Siophy6 . Acts 274 71-74 (1972) .
Chinkca 25 350-360 (1971) .
20 .
I. C. P . SMITH,
21 .
G. H. DODD, M. D. BARRATT and L. RAYNER, FBBS LettM 8 286-288 (1970) .
22 .
W. L. HUBBELL and H. M. McCONNELL, PAoc . Nat . Acad . Sci .
(USA) 61 12-16
(1968) . 23 .
0. H. GRIFFITH, L . J. LIBERTINI and G. B. BIRREL, J . Phy6 . 3425
24 .
Chem . 75 3417-
(1971) .
H. M. McCONNELL, K . L. WRIGHT and B. G. McFARLAND, Siochi.m . Sí.ophy6 . Red .
Commun . 4 7 273-281 (1972) . 25 .
P . JOST, A. S . WAGGONER and 0. H. GRIFFITH, StAuctuke and
Function o6
Si.otog.i¢a.1 MembAaaneA, p . 83-144, Academic Press, New York (1971) . 26 .
R. J. MELHORN and A . D. KEITH, Membrane Motecutan Siotogy, p . 192-227, Sinauer Associates, Stamford (1972) .
27 .
R. D. KORNBERG and H. M. McCONNELL, Siochem .
28 .
C. J. SCANDELLA, P . DEVAUX and H. M. McCONNELL, PAoc . Nat . Acad . Scí. . 69 2056-2060 (1972) .
10 1111-1120 (1971) . (USA)
Electron Spin Resonance
1314
29 .
C. W. GRANT
and
Vol. 13, No . 10
H. M. McCONNELL, Pnoc . Nat . Acad . Sc í. .
(USA)
70 1238-1240
(1972) . 30 .
C. F. CHIGNELL, L(.Se Scí. (1)
10 699-706 (1971) .
31 .
R. H. ERLICH, D. K . STABKWBATHER
and C.
F. CHIGNELL, Molec . PhaJCmacol .
9
61-73 (1973) . 32 .
C. F. CHIGNELL, D. K. STARKWEATHER
and
R. H. ERLICH, B.í.och,ím .
Ikophy6 . Acta
271 6-15 (1972) .
and
D. K. STARKWEATHER, Comp . B,í.ochim . Pharrmacol .
33 .
C. F. CHIGNELL
34 .
J. C . HSIA
35 .
R. K. LEUTE, E . F. ULLMAN, A. GOLDSTEIN
and
L . H. PIETTE, AAch Biochem . B .í.ophy4 .
and
. (in prees)
12 9 296-307 (1969) .
L. A. HERZENBERG, Natulce (New
Bdnl .) 236 93-94 (1972) . 36 .
C. F. CHIGNELL
and
D. K. STARKWEATHER, PhaJUnawIogy 8 368-376 (1972) .