Journal of Biochemical and Biophysical Methods.

17 (1988) 237-248 237

Elsevier

BBM 00710

A simple method for determination of rotational

correlation times and separation of rotational and
polarity effects from EPR spectra of spin-labeled
biomolecules in a wide correlation time range

Heinz-Jiirgen Steinhoff
lnstitut ]'fir Biophysik, Ruhr-Universitiit Bochum, 4630 Bochurn, F,R. G.

(Received 21 April 1988)

(Accepted 15 August 1988)

Summary

A method using nitroxide radical spin labels for determining both the isotropic rotational correlation
time ZR and the environmental polarity of the label is described. By means of a least square fitting
method, the values of an effective hyperfine tensor A' and of an effective g value tensor g ' of randomly
oriented spin labels are determined from X-band EPR spectra on the basis of an effective time-indepen-
dent Hamihordan. The traces of the tensors deliver the information about the environmental polarity of
the label and are not dependent on the rotational correlation time "rR. A new averaging parameter S
(~'R), calculated on the basis of the principal values of the tensor A', permits the evaluation of the
rotational correlation time z R in a very wide time range between 1 0 10 and 1 0 6 s. An application of
this method to spin-labeled methemoglobin over a large temperature range and in environments of
different polarity is discussed.

Key words: EPR; Rotational correlation time; Environmental polarity

Introduction

The sensitivity of spin-label spectra to the rate of nitroxide rotational motion has
long been known [1,2]. Therefore, if a spin label can rigidly bind to a macromolc-
cule, its electronic paramagnetic resonance (EPR) spectrum should directly reflect

Correspondence address: H.-J. Steinhoff, Institut fi.ir Biophysik, Ruhr-Universit~it Bochum, 4630 Bochum.
F.R.G.

0165-022X/88/$03.50 ,~2 1988 Elsevier Science Publishers B.V. (Biomedical Division)

238

the motion of the macromolecule itself or some residual motion (librational motion)
of the label with respect to the whole macromolecule [3-5]. Rotational correlation
times ri~ for isotropic motions of slow-tumbling spin labels or spin-labeled biopoly-
mers have been determined by use of appro:dmate methods as we11 as simulation
methods [2]. These approaches utilize the property' that the positions of the outer
peaks in the X-band spin-label spectra (their separation is approximately 2Azz,)
shift inward from their rigid limit position, 2Azz, if the rotational motion increases.
On the other hand, the principal values of the hyperfine tensor __Aand of the g value
tensor g of a nitroxide radical depend significar~tly on the polarity of its environ-
ment [6,7]. Since a protein possesses hydrophobic and hydrophilic regions, it must
be considered that the environment of a protein-bound spin label may change its
polarity during a temperature-induced conformational change of the protein.
Another explanation of an observed temperature dependence of the peak separation
in the spectrum ma t' be the N O ' - H X hydrogen bound formation by, the spin label
[8,9~.
A special problem has been to separate the spectral effects of nitroxide libra-
tional motion from those due to hydrogen bonding interactions or polarity' change.
This problem may' bc partially overcome if one considers not only the hypcrfine
separation 2Azz but also line-width values of the absorption spectra [4,5,13]. In the
present work it is dcmonstra:cd, that this problem of" separation may be especially
overcome through determination of thc complete tensors g and A of the spin-labeled
protein using a fast least square fitting method. In the case of polarity change of the
spin-label environment, the change of the hyperfine separation 2Azz is correlated to
the change of the hyperfine tensor (TrAy trace. On the other hand. if an observed
change of the hyperfine separation is due to librational motion of the labcl, the trace
of the tensor _A must net alter, which follows directly from the rotational invarianee
of the trace of a tensor [111.
Therefore the starting point of the separation method presented here must be the
determination of the tensors g and A also in the slow motional region. However, a
least square fitting method in the slow tumbling region (~R -- 10 7 S) on the basis of
Freeds prograln [2] is very time consuming. A ~ypicat fit with seven parameters
including ten steps will last up to 2300 s (depending on rR) on a Cyber 855,
delivering the tensors g and A and the rotational correlation time ~'R- Instead of
this, the method presented here describes the EPR spectra of tumbling spin labels
using an effective time-independent Hamiltonian. This is a drastic simplification
and its relevance to experiment and theory has to be checked very carefully. Some
points of practical interest arc to be mentioned. To ensure the uniqueness of the fits
the number of parameters is reduced to a minimum: the six principal values of the g
and _A tensor. The line width function is determined from the spectrum in an
independent step. The fitting pr(rcedure is very fast so that an implementation of the
method using a personal computer is possible.
As an example of an application of this method, the tensors g and A of the spin
1abel .. :-(1-oxvl-_. 6.6-tetramethvl-4-piperidinvl) iodoacetami-de (JAA6) bound
within methemoglobin in different environments are determined in a temperature
range between 80 and 330 K.
 239

Materials

Oxyhemoglobin was prepared from fresh horse blood samples by the method of
Benesch et al. [12]. Spin labeling (JAA6) followed the procedure of McConnell et al.
[13]. The oxidation of the labeled oxyhemoglobin to methemoglobin was achieved
by addition of a threefold amount of K 3 (Fe(CN)~,). The sample was desalted by
running it through a column of Sephadex G-25. Polycrystalline samples were
prepared by mixing the solutions of labeled methb with buffered anamonium sulfate
according to the method of Perutz [14].
EPR spectra were measured on home-made X-band and K-band (22 G H z )
spectrometers equipped with a modified Oxford ESR 9 variable temperature acces-
sory. The microwave power used was 0.1 mW, the modulation frequency was 50
kHz and the modulation amplitude 0.4 10 - 4 T. The magnetic field was measured
with an A E G Telefunken N M R instrumentation. 2,2-Diphenyl-l-picrylhydrazyl
(DPPH) powder served as a g value standard ( g = 2.0037). After analog-digital
conversion, the spectra were recorded in a personal computer (CBM 8296, Com-
modore) and then transmitted to a Cyber 855 (Control Data). There the tensors g
and _A were determined as described in the following section.

Methods

Determination of the tensors g and A from E P R spectra of amorphous or viscous

samples
The EPR spectra of a nitroxide radical with spin 1 / 2 having fixed orientations
can be accounted for with the basic spin Hamiltonian

It" =/~eSgH + SA1, (1)

where H is the external field vector, /~ the Bohr magneton, S the electron spin
operator, I the nuclear spin operator, g the electron g value tensor and /t the
electron nuclear hyperfine interaction tensor. To a good approximation [15], the
eigen values of the Hamiltonian Eq. (1) are given by

E = g ( O , ~ ) B e H M s + A ( O , c ~ ) M s M I, (2)
with

g(O,O) = g x x l z x + g v v l z y + g z z l z z ,

= ( A lz + A ylz + AhIz ),j2, (3)

where gxx, ,g'YY, gzz and Axx, A y y , Azz are the principal values of the tensors g
and A, respectively, a n d / z x , lzv, /zz are the squared direction cosines between the
nitroxide x, y, z axes and the laboratory magnetic field direction (z): /zx = ( - s i n 0
240

COS~) 2, /ZY = (sin0, sin~) 2, lzz = (cos0) 2- Ms and M l are the eigen values of Sz
and lz, respectively. Considering the selection rules Eq. (2) gives the resonance
frequencies of every" spin-label orientation in the external magnetic field. A F o r t r a n
program was written to calculate the corresponding E P R spectrum. The E P R line
positions are calculated in steps of 3 for 0 and q~. The intensity of each EPR line
of a r a n d o m distribution of spin-label orientations in the laboratory fixed coordi-
nate system is proportional so the n u m b e r of molecules d N, for which Eq. (3) holds:

1 d~2 - d N ~ No sin0 dO d4~. (4)

After summing the stick spectra within a 0.5 G range of H, the line shape is applied.
The line shape function may be evaluated from the first derivative of the experimen-
tal EPR absorption spectrum. In the case that the difference between the principal
values Azz and Axx, Ay~- is large c o m p a r e d to the line width A (a condition,
which is fulfilled in the case of nitroxide labels), the resonance lines of the first
derivative of the EPR spectrum at the positions H+ = (hv++_ A , c z ) / ( g z z B ~ ) are
identical with the line shapes of a single resonance tine [16]. Here we suppose that
the line width of the resonance lines does not depend strongly on M~ or g. This
assumption will be proved experimentally in the next section. The convolution of
the calculated stick spectrum with the first derivative of the line shape determined
from the experimental E P R spectrum yields the desired theoretical E P R spectrum f,
which is now a function of the six principal values of the tensors g and _A only. By
means of the subroutine S Q U F I T [17] the initially chosen values of g and A are
varied iteratively until a m i n i m u m of the chi-squared value Q is ob]ained (least
square fitting). Alternatively the determination of the line shape function m a y be
included into the fitting, increasing the n u m b e r of parameters from six to nine. A
complete fit including ten steps takes less than 10 s on a Cyber 855 (about 30 min
on a personal computer without floating point coprocessor).
The next step is to prove the applicability of the fitting method to experimental
spectra. The immobilization of the labeled methemoglobin molecules was achieved
by crystallization in a m m o n i u m sulfate solution, by freezing of a methemoglobin
solution or by lyophilization. As an example Fig. 1 shows the calculated and the

TABLE 1
VALUES OF THE TENSORS g' AND A', DETERMINED IN THE X- AND K-BAND AT T = 180 K

X-band K-band
(u = 9.00 GHz) (u = 22.00 GHz)
gxx 2.0089 + 2 x 10 -4 2.0089 + 5 x 10-4
gyy 2.0064+210 4 2.0060+510-4
gzz 2.0021 + 2 x 10-4 2.0021 + 5 x 10-4
Axx/MHz 17.4 +_1.2 18.2 +2.0
Ayy/MHz 21.3 *1.2 18.6 +2.0
Az~/MHz 104.5 _+0.1 104.0 + 2.0
1/3 TrA/MHz 47.7 _+0.4 46.9 +_2.0
 241

(3

I I I

0.315 0.320 0.325 0.330

B/Testa

I I I I
0.810 0.815 0.820 0.825
B/Testa
Fig. 1. EPR spectra of polycryst~fine methemoglobin-JAA6 at T=180 K. (a) t,o=9.0 GHz: (b)
% = 22.0 GHz. The dotted lines show the least square fittings to the experimental spectra.

experimental spectra of a polycrystalline sample at 180 K in the X- a n d K - b a n d .

Excellent agreement is o b t a i n e d between the calculated a n d the e x p e r i m e n t a l curves.
The tensors g a n d _A d e t e r m i n e d from both frequency b a n d s are shown in T a b l e l.
2a2

The corresponding values coincide. The greater standard deviation of the values
determined from the K-band compared to those determined from the X-band is due
to a lower signal to noise ratio. This result shows that the approximation in Eq. (2)
is valid in the K-band too.
No significant change in the values of g, _A and Q was observed when we
extracted the line shape function from the h~-gh field peak instead of the low field
peak. So the dependence of the line width on M~ and g is small and the assumption
of a single line shape function for the whole EPR spectrum is justified.

Properties of the effectice spin Hamiltonian

Because of the choice of a time-independent Hamiltonian Eq. (1) the introduced
method of determining the tensors can firstly be applied to strong immobilized
samples. From the work of Freed [2] it is known that the hyperfine separation and
the intrinsic line width of the first derivative of the absorption spectra is insensitive
to molecular motion, if the correlation time r is greater than 10 -7 s. Motional
fluctuations with corrdation times less than 10--7 s will produce partial averaging of
the tensor dements. We describe the spectrum in this time range introducing
effective g ' and A' tensors and this yields an effective time independent Hamiho-
nian H~ff. The questions now to be answered concern the properties and appLica-
tions of this effective time-independent Hamiltonian. A first consistency check is
that one must have TrA = TrA', which follows directly from the rotational invari-
ance of the trace of a tensor. We have calculated EPR spectra which are char-
acterized by motional fluctuations with correlation times between 6 l 0 - 7 and
10 10 s using the stochastic Liouville method of Freed published in [2] with
assumed values of the tensors g and A, which we shall call 'true' in the following.
These spectra were fitted with the method described in the preceeding section
leading to values of the effective tensors g ' and A', which depend on the values of
%. The relative depilation of TrA' from the ' true' value TrA, o.r~A, and the deviation
of Trg' from Trg, o.r~~, are shown in Fig. 2. In the time range of % above
0.3 ]0 -v s and below .10 -9 s the values of TrA' of the best fits deviate less than !%
from the given value of TrA, the deviation in Trg' is less than 5 10-5 in this
range. Even in the critical correlation time range between 10 -~ and 10 .-9 s the
systematic deviations in TrA' and Trg' are less than 3% and 10 -4, respectively. The
fluctuations of ~r~g are mainly due to the finite step width of the spin-label
orientations for 0 and q). While the traces of the fitted tensors are almost
independent of the correlation time %, their principal vahies change from the rigid
limit values to the complete averaged values Axx =A~. v = A z z = 1 / 3 TrA and
gxx = gw" = gzz = 1 / 3 Trg in the rapid motion limit: the value of AT~z changes
from !05 MHz to 47.6 MHz for example. The independency of TrA' on % is an
essential result and will be used to characterize motional effects on EPR spectra and
to separate other effects. We introduce an averaging parameter S characterizing the
degree of averaging on the basis of the tensor A:

Sz = A z z - 1/3 TrA'
Azz - 1 / 3 YrA ' (5)
 243

2
~'T. A.]0
/.
~.rg"10

-1 i i i i i i . i l i L , ,, i , ,L,~i[ i,.

16~ 16g 168 16; 16~

Fig. 2. The de,Aation of TrA' and Trg' from the 'true' traces of the tensors, TrA and Trg, respectively, as
a function of the correlation time ~'R for Brownian rotational diffusion. The deviations o are defined as
OTrA = (TrA' - T r A ) / T r A (X) and ol-rg= (Trg' - T r g ) / T r g (C). The curves were determined from
fittings of spectra calculated on the basis of a time-independent effective Hamiltonian to spectral
simulations using the exact stochastic Liouville method of Freed [2]. The parameters for the calculated
spectra in the notation of [2] are: L_<I6, K_<12, A x x = 7 . 4 G, Avv=6.1 G, AT,z=37.5 G,
gxx = 2.0089, gvv = 2.0063, gzz = 2.0023 and the peak-peak first derivative line width is 3.0 G.

or, i n c l u d i n g all p r i n c i p a l v a l u e s

Z IA~i- 1/3 TrA'[

i = X,V,Z
S = " (6)
]A,- 1/3 Tra]
i = x,y,z

T h e t e n s o r p a r a m e t e r s w i t h o u t p r i m e s a r e the rigid l i m i t values. S will c h a n g e f r o m

1 for the rigid l i m i t case, A~t = A , , to z e r o in the c a s e o f c o m p l e t e a v e r a g i n g . T h e
v a r i a t i o n o f S w i t h ~R is s h o w n in Fig. 3 for the B r o w n i a n d i f f u s i o n m o d e l . It
s h o u l d be n o t e d t h a t S is d e f i n e d in the w h o l e r a n g e o f ~'R f r o m 10 - 6 to 10 -10 s. So
t h e r a n g e o f ~R, w h i c h c a n b e e v a l u a t e d f r o m an e x p e r i m e n t a l s p e c t r u m u s i n g a
r e l a t i v e l y s i m p l e a n d fast m e t h o d , c a n b e e x t e n d e d to c o r r e l a t i o n t i m e s less t h a n
7 x 10 - 9 s, w h i c h h a d b e e n the l o w e r l i m i t for the s i m p l i f i e d m e t h o d g i v e n in [2].
T h e r e a s o n for t h a t l i m i t a t i o n is t h a t the o u t e r lines b e g i n to c o n v e r g e to the
m o t i o n a l l y n a r r o w e d s p e c t r u m a n d the p a r a m e t e r S d e f i n e d in [2] b e c o m e s u n d e -
f i n a b l e . C o m p a r e d to this, Fig. 3 c a n b e u s e d as an e m p i r i c a l c a l i b r a t i o n c u r v e for
244

0-9
08
07
06
0.5
0.4
0.3

0"2f
ol

16'0 10-9 ~a 70-7 10-%/s

Fig. 3. The parameters S z (o) and S (D) versus rotation',d correlation time. The curves were determined
from fittings of spectra on the basis of a zime-independent effective Hamihonian to spectral simulations
using the stochastic Liouville method. The parameters for the simulations are the same as in Fig. 2.

determining r R values from the effective hyperfine tensor ,4' in the whole time
range from 10 -6 to ]0 -1 s.
The values of the parameters S and TrA' shown in Figs. 2 and 3 were calculated
starting the fitting with varying initial sets of g ' and A'. The fitting converged in all
cases if the initial choice of g ' and A ' corres-ponds to S values, which deviate less
than 0.3 from the final (' true-~) value of S. The initial values of TrA' and T r g ' m a y
deviate from their true values more than 20% and 2 10 -3, respectively.

Application of the method." influence of temperature and enuironmental polarity on the

EPR spectrum of spin-labeled methemoglobin
The behaviour of the parameter A z z and of the trace of the hyperfine tensor,
1 / 3 T r A , with temperature for the w e t and lyophilized samples is shown in Fig. 4.
The value Azz of all samples decreases with increasing temperature. A b o v e 190 K
the rate of change of A z z of the wet samples is greater than that of the lyophilized
sample and increases with increasing temperature. However, the trace of the
hyperfine tensors of all samples is independent of temperature in the whole
temperature range investigated even in the critical temperature range between 220
and 300 K. However, the hyperfine separation as well as 1 / 3 T r A of the lyophilized
sample are less than the corresponding values of the wet samples. For a better
understanding of this behaviour the hyperfine tensor __Aof the spin label J A A 6 was
determined in model experiments in environments of different polarity at 120 K.
Fig. 5 shows the plot of 1 / 3 T r A against A z z for these samples. There is good
 245

Azz/MHz
o

106

10/,
/',
102

,LA
100

98

96

80 ~oo ,20 ,,o 40 ,8o ~oo 220 2,0 2~o 28o ~/K

lira
_ _ _ "b , . : i

MHz

5O

~2
I I

80 100 120 lt,0 160 180 200 220 240 260 280 T/K

Fig. 4. (a) The parameter Azz as a function of temperature for methemoglobin-JAA6, which has been
immobilized by different methods: (A) frozen in aqueous solution, (z~) crystallized in a m m o n i u m sulfate
solution of high ionic strength, (O) lyophilized. (b) The trace of the hyperfine tensor, 1 / 3 TrA, as a
function of temperature for the same samples shown in (a).

linear correlation between 1 / 3 TrA and Azz and the values of the spin label bound
to the methemoglobin molecule fit the curve well. So we conclude that the
correlated change of Azz and 1 / 3 TrA by removing the water from the methe-
moglobin-JAA6 is due to a change of the polarity of the spin-label environment. On
the other hand, the temperature dependent change of Azz of all samples is not
correlated with a change of the trace of the tensor, the trace is even independent of
temperature in the whole temperature range investigated. Therefore a temperature-
induced change of the polarity of the spin-label environment in the protein can be
246

~-Tr A
MHz

50

48

z,8

44

42

~0 I I l l I t

92 94 96 98 100 !02 104 106 108 Azz/MHz

Fig. 5. Correlation between the trace of the hyperfme tensor, 1/3 TrA, and Azz for the spin label JAA6
in systems of different polarity, determined at 120 K: ten) butanol, (11) ethanol, (o) lyophilizcd
methemoglobin-JAA6, (e) glycerol, (zx) methemoglobin-JAA6 in aqueous solution, (zx) crystallized
methemoglobin-JAA6.

e x c l u d e d a n d t h e o n l y e x p l a n a t i o n f o r t h e a b o v e r e s u l t s is t h e e x i s t e n c e o f a n i s o -
t r o p i c m o t i o n a l f l u c t u a t i o n s o f t h e l a b e l s e v e n in t h e f r o z e n s a m p l e s . A d e t a i l e d
a n a l y s i s o f t h e s e f l u c t u a t i o n s will b e g i v e n in a s u c c e s s i v e p a p e r .

Simplified description of the method and its applications

A method using ni',roxide radical spin labels for determining both the isotropic rotational correlation
time r R and the environmental polarity of the label is described. By means of a least square fitting
method an effective hyperfine tensor ,4' and an effective g value tensor g ' are determined from X-band
EPR spectra. The traces of these tensors deliver the information about the environmental polarity of the
label and are not dependent on the rotational correlation time rg in the time range investigated from
10 ~ to 10 - m s. An averaging parameter S(rR) calculated from the values of the effective tensors
permits the evaluation of the rotational correlation time from experimental EPR spectra in the whole
time range given above. Due to the algorithm used here these informations about spin labels and
spin-labeled biomolecules may be evaluated from their EPR spectra using a personal computer. This
method may be applied to investigations of spin-labeled biomotecules where a change of motional
freedom of the label due to a structural change may be accompanied by a change of environmental
polarity. In this case the only consideration of the hyperfine separation would lead to incorrect results.

Acknowledgements
S p e c i a l t h a n k s a r e d u e to P r o f e s s o r A. R e d h a r d t f o r t h e n u m e r o u s i n s p i r i n g
c o n v e r s a t i o n s . T h e a s s i s t a n c e o f C. B a s s a r i s a n d H. G r o t t h a u s m a n n i n t h e p r e p a r a -
t i o n o f t h e m a n u s c r i p t is g r a t e f u l l y a c k n o w l e d g e d .
 247

References

1 Stone, T.J., Buckman. T., Nordio, P.L. and McConnell, H.M. (1965) Proc. Natl. Acad. Sci. USA 54,
1010-1017
2 Freed, J.H (1976) in: Berliner, L.J. (Ed.), Spin Labeling TheoD' and Applications, Academic Press,
New York, pp. 53-132.
3 tlyde. J.S., Smigel, M.D., Dalton, L.R. and Dalton. L.A. (1975) J. Chem. Phys. 62, 1655-1667
4 Johnson, M.E. (1978) Biochemist~' 17, 1223-1228
5 Johnson, M.E. (1979) Biochemist~' 18, 378-384
6 McConnell, H.M. and McFarland, B.G. (1970) Q. Rev. Biophys. 3, 91-136
7 Lassmann, G., Ebert, B., Kuznetsov, A.N. and Damerau. W. (1973) Biochim. Biophys. Acta. 310,
298-304
8 Johnson. M.E. (1981) Biochemist~" 20, 3319-3328
9 Hwang, J.S.. Mason, R.P., Hwang. L.P. and Freed, J.H. (1975) J. Phys. Chem. 79, 489-511
10 Ruggiero. J.. Sanches, R., Tabak, M. and Nascimento, O.R. (1986) Can. J. Chem. 64. 366-372
11 Van, S.P., Birrell, G.B. and Griffith. O.H. (1974) J. Magn. Res. 15. 444-459
12 Bcnesch, R.E., Benesch, R., Renthal, R.D. and Maeda, N. (1972) Biochemist~, 11, 3576-3582
!3 McConnell, H.M., Deal, W. and Ogata, R.T. (1969) BiochemistD' 8, 2580-2585
24 PerutT, M.F. (1968) J. Cryst. Growth 2, 54-56
]5 Libertini. L.J. and Griffith, O.H. (1970) J. Chem. Phys. 53, ]359-1367
16 Weft, H.A. and Hecht. H.G. (1963) J. Chem. Phys. 38. 281-282
17 Braess. D. (1966) Computing 1,264-267