Proc. Natl. Acad. Sci. USAVol. 92, pp. 2116-2120, March 1995Biophysics

Stark-effect experiments on photochemical holes inchromoproteins: Protoporphyrin IX-substituted myoglobin

(chromophore-protein interaction/order-disorder phenomena/spectral-structural correlations)

J. GAFERT*, J. FRIEDRICH*, AND F. PARAKt*Physikalisches Institut und Bayreuther Institut fCir Makromolekulforschung, D-95440 Bayreuth, Germany; and tPhysik-Department, Technische UniversitatMunchen, D-85748 Garching, Germany

Communicated by Hans Frauenfelder, Los Alamos National Laboratory, Los Alamos, NM, November 15, 1994

ABSTRACT We performed comparative Stark-effect ex-periments on spectral holes in a protein and a glass sample.The protein was protoporphyrin IX-substituted myoglobin ina glycerol/water solvent. The glass sample was a protopor-phyrin IX-doped mixture of dimethylformamide/glycerol. Asexpected, in both cases the spectral holes varied linearly withthe electric field. Yet, whereas in the protein the holes showeda clear splitting, they showed no splitting in the glass sample,irrespective of the chosen polarization of the laser. In bothsamples the hole broadened in the applied field. The magni-tude of the broadening was about the same in both cases. Thefollowing conclusions were drawn. The absence of a splittingin the glass signals an effective global inversion symmetry ofthe chromophore, despite its low symmetry group. The dipolemoment changes are random. In the protein the inversionsymmetry is broken through the spatial correlation of theprotein building blocks, leading to a molecular frame-fixeddipole moment difference and, hence, to the observed splitting.Despite these symmetry-breaking properties, the local struc-tural randomness is of the same magnitude in the glass andin the protein, as is obvious from the broadening. The distinctdifference in the Stark pattern shows that the range of therelevant chromophore interactions is confined to typicaldimensions of the protein.

The characterization of the solid state of proteins is notstraightforward. Proteins do not fit into the usual categories ofsolid-state phases. They have many "in-between" properties.At sufficiently high temperatures, say above 200 K, they showa behavior in between those of liquids and crystals (1-10). Atsufficiently low temperatures their properties are in betweenthose of crystals and glasses. For instance, their specific heatat low temperature is glass-like (11, 12), their dielectric as wellas ultra-sound absorption is glass-like (11), and many of theiroptical as well as their Mossbauer properties are glass-like (2,4, 5, 13-18). On the other hand, their x-ray diffraction patternis well resolved and crystal-like (19). Yet, even the x-raydiffraction reveals glass-like properties when the Debye-Waller factor is analyzed (20, 21): the mean-square displace-ment (x2) averaged over the atoms of an amino acid residuevaries along the backbone and, as the absolute temperaturegoes to zero, shows large deviations from the usual zero-pointvibrational amplitudes. These findings show that there is anadditional structural uncertainty in the positions of the proteinbuilding blocks which arises from a broad distribution ofconformational substructures.

In this paper we will show by using hole-burning Starkspectroscopy (22-27) that, on the one hand, the randomness inthe structure of a protein is comparable to the randomness ofa glass. Yet, on the other hand, there is a high spatialcorrelation in the protein which leads to a very specific Stark

pattern. From this specific pattern we will draw an importantconclusion as to the range of the interaction of the chro-mophore with its environment: this range does not exceed thetypical dimensions of the protein. The influence of the hostglass can be neglected. Hence, the chromophore is indeed aprobe of the protein. This is an important statement: it implies,for instance, that inhomogeneous line broadening in chro-moproteins is due to conformational substates; it confirms ina direct way that the properties measured in optical experi-ments-for instance, the compressibility (28-30)-are reallyproperties of the protein. Stark-effect experiments on proteinshave been performed so far on photosynthetic reaction centers(for a review, see ref. 31), which are characterized by strongcharge transfer states. Experiments in solution with biophys-ical relevance to signal transduction in visual pigments werecarried out by Mathies and Stryer (32) on retinal, but to ourknowledge hole-burning Stark experiments on proteins havenot been published.

Basic Aspects of Hole-Burning Stark Spectroscopy inProteins

To sketch the characteristic features of the Stark effect, westart with the assumption that the molecular probe can bedescribed approximately as a point dipole with dipole momentilg in the ground state and IUe in the excited state. An externalfield Eo induces a frequency shift Av of the transition underconsideration:

hAv = -fAi0E. - -f2 E.A .- E0, [1]

where Ap,,, is the difference of the dipole moments betweenground and excited states,

/lo = Ile - lAg, [2]

and A& is the difference of the polarizability tensors in therespective states.faccounts for the fact that in a host matrix the local field Ej.c

may be quite different from the external field. In most casesone uses the so-called Lorentz field correction, which is basedon the assumption that the molecular probe sits in a sphericalhole of a dielectric continuum which is polarized through theexternal field. In this case

3+ 2f= 3 '

[3]

with e being the dielectric constant. The problems with thepoint dipole approximation and the Lorentz correction werediscussed by Hanson et al. (33). As to proteins, nothing isknown about the quality of these approximations, but twoproblems are immediately obvious: (i) the medium, as seenfrom the chromophore, is no longer homogeneous and (ii) thecontinuum approximation for a protein seems to be question-

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

2116

Dow

nloa

ded

by g

uest

on

July

14,

202

1

Proc. Natl Acad Sci. USA 92 (1995) 2117

able. But for the qualitative features of our experiments this isof no concern.

In all our experiments the observed spectral features arelinear in Eo, hence, the quadratic term in Eq. 1 can safely beneglected.

Next, we have to include the influence of the matrix in moredetail. This influence is twofold. As mentioned above, thematrix modifies the external field via the f factor (Eq. 3) butcan generate a field of its own as well. The matrix field, in turn,can induce a dipole moment pi (i stands for induced) in thechromophore. This induced dipole moment may contributesignificantly to the observed Stark pattern because the matrixfields can be extremely high (34, 35). Note that ,iL contains thepolarizability tensor a, although the effect is still linear inthe external field Eo. If the matrix were a perfectly orderedcrystal, it would change the molecular parameter A,I.( by anadditional contribution Ap,i with a constant magnitude andspatial direction. Both contributions would add up, but itwould not be possible to separate them experimentally. In casethe matrix is disordered, A,ui attains a random character, whichgives rise to a broadening linear in E..

Let us describe the general situation of an asymmetricchromophore in a random matrix in more detail. AIpO is avector fixed in the frame of the chromophore. As a conse-quence there is a fixed angle y between A1.o and the transitiondipole moment ILeg. Hence, the direction of AIpO with respectto Eo can be selected by choosing the polarization of the laserin a proper way. As long as Ai0o is selected so that it has itsmajor component parallel to Eo, a spectral hole would split ina symmetric fashion since, in a random matrix, there is withequal probability a component which is antiparallel to thefield. From the splitting the quantity fASzo in Eq. 1 can bedetermined. On the other hand, if the polarization of the laseris chosen so that Aix, is mainly oriented orthogonal to theexternal field, only a broadening results.

In addition to the characteristic features resulting from Ag.o,there is AliL, which is random and, hence, can have anyorientation in the frame of the chromophore. Moreover, alsoits magnitude is subject to a variation. For the distribution ofthe magnitude Ali we chose a Maxwell-Boltzmann-like dis-tribution (25, 36), with the restriction that this distribution isconfined to two dimensions, taking into account the disk-likepolarizability tensor of a porphyrin molecule (26):

2 j2_

P(Agi) = (r)PA2ti expi - ( I) [4]

where or is the respective width. It is a parameter characteristicfor the randomness of the guest-host interaction. If we takeinto account the influence from a random matrix on theeffective dipole moment, Eq. 1 changes to

(Fig. 1) dissolved in 0.1 M NaOH. In the reconstitutediron-free Mb, the absorption ratio at 405 and 280 nm was 3.26.(For details of the sample preparation, see ref. 37.) The finalsample was obtained by mixing 2.4 ml of water-free glyceroland 1 ml of water containing 290 mg of Mb. Since protopor-phyrin IX is practically insoluble in glycerol/water, this finalsolution did not contain any significant amount of free chro-mophores. For comparison, protoporphyrin IX was also dis-solved in a glass matrix. We chose a mixture of dimethylform-amide and glycerol with a volume ratio of 1:3.

Spectroscopic Measurements. Holes were burnt in proto-porphyrin IX-substituted Mb at three different frequencies inthe inhomogeneous band located at 15,818, 16,021, and 16,195cm-1 as indicated by arrows in Fig. 4. In the glass sample weperformed hole burning at one frequency only, 15,977 cm-1(Fig. 2). The field Eo under burning conditions was -5.835kV-cm-'. The holes were burnt at 1.7 K to relative depths ofabout 30%. They were saturated to some extent, yet they couldstill be well fitted to Lorentzians. The hole-burning lasersystem was a ring dye laser pumped by a 6-W argon ion laser.Burning times were on the order of 15 s at radiation powerlevels of 50 ,W. The bandwidth of the laser was on the orderof a few megahertz. The holes were detected by scanning thelaser over the burnt-in hole and measuring the transmission.The sample was kept in cuvettes with a thickness of 1 mm.

The optical density varied between 0.09 and 0.7, depending onsample and spectral range. The cuvette was put between twoplane electrodes. The voltage across the electrodes was variedbetween -3.5 and 3.5 kV. The distance between the electrodeswas 6.5 mm.

Polarization Selection and Fitting Procedure. An importantaspect in a hole-burning Stark experiment concerns the laserpolarization. In a weakly bleached sample, the hole is stronglyanisotropic. The distribution of selected molecules is -cos2 E,where E is the angle between the polarization direction of thelaser field and the transition dipole moment Pieg. As a conse-quence, the Almo distribution is also strongly anisotropic,because the angle y between ILeg and Ai.0 is fixed. By choosingthe polarization of the burning laser properly, the Aim. distri-bution can be optically selected so that its main componentfalls either parallel or perpendicular to the external field Eo.We chose the laser field parallel as well as perpendicular to Eo,because it takes at least two different polarization directions toget unambiguous results from the fit. In addition, both patternshave to be fitted with the same parameters simultaneously. Forthe whole fit procedure we used the formulas as given bySchatz and Maier (36). For the protein sample, the resultingelectric-field features of the hole were distinctly different: theparallel orientation led to a splitting and a broadening of thehole, whereas the orthogonal orientation led to a broadeningonly.

hAv = -fE(A\u cos 0 + Aiii cos Oj), [5]

with 0 and Oi being the angles between the external field E.,and the respective dipole moment differences A,IO and A,ui.The Stark spectra are obtained from a convolution of the

burnt-in hole with the homogeneous lineshape function shiftedin the electric field according to Eq. 5. The convolution is thenaveraged over all possible orientations of the chromophorewith respect to the laser field and over the distribution as givenin Eq. 4 (36).

Materials and Methods

Sample Preparation, The natural heme group does notallow for narrow-bandwidth hole burning because of its ultra-short lifetime. Hence, horse myoglobin (Mb) (Sigma) wasdeprived of the heme group by a butanone extraction atpH 2.3.The apoprotein was reconstituted with protoporphyrin IX

/CH2HC CH3

-CH

CH2

CH3

HO-C-CH2CH2II0

CH 2CH2-C-OHII0

FIG. 1. Protoporphyrin IX.

Biophysics: Gafert et aL

Dow

nloa

ded

by g

uest

on

July

14,

202

1

Proc. Natl Acad ScL USA 92 (1995)

vb = 16,021 cm-1

15,600 16,000V, cm-1

-5

Frequency, GHz

o V

0

Frequency, GHz

FIG. 2. Deformation of the shape of a spectral hole in an externalelectric field E.. 0 indicates the hole without external field, 11 indicatesthe hole with the polarization of the laser parallel to E., and I withthe polarization perpendicular to E.. The sample was protoporphyrinIX in a dimethylformamide/glycerol glass. Temperature was 1.7 K. E.was 11.67 kV-cm-'. (Inset) Inhomogeneous absorption band with theburn frequency indicated. a.u., Arbitrary units.

Results

Our goal was to find out whether the ordered structure of theprotein leads to characteristic features in the Stark pattern ofthe chromophore as compared with a random environment.

Fig. 2 shows a hole burnt into the glass system as it deformedunder an external field E.. The field was varied from -5.835to +5.835 kV*cm-l. The polarization of the burning laser wasparallel (11) or perpendicular (I) to Eo. There was no essentialdifference: the hole broadened in a symmetric fashion in bothbases. Fig. 2 Inset shows the inhomogeneous spectrum and thespectral position where the hole-burning experiment wascarried out.

Fig. 3 shows the broadening of the hole in the glass as afunction of Eo. The respective dependence is perfectly linear.The linearity indicates a first-order Stark effect.

Fig. 4 shows the deformation of a hole under an electric fieldin the protein system. This figure should be compared with Fig.2. There is a striking difference: when the laser polarizationwas chosen parallel to Eo, the hole showed a clear splitting inaddition to a broadening. On the other hand, when the laserpolarization was perpendicular to E., there was a broadeningonly. This is quite in contrast to the glass sample, where nosplitting was found.

Fig. 4 Inset shows the inhomogeneous spectrum with thepositions indicated where Stark experiments were performed.The experiment shown is for the band center.

Fig. 5 shows the splitting, fAl0oEo/h, and the broadening,

FIG. 4. Deformation of a spectral hole in an external field E.. 0,without field; 11, laser polarization parallel to E.; I, laser polarizationperpendicular to Eo. The sample was protoporphyrin IX in myoglobin.Temperature was 1.7 K. E. was 11.67 kV-cm-1. (Inset) Inhomogeneousband with the positions indicated where Stark experiments wereperformed.

faE./h, as a function of Eo. In addition, the angle y betweenthe transition dipole moment IUeg and A,IO is shown as afunction of the field. -y did not show any field dependence, asexpected. Again, the data shown refer to the band center.Our observations can be summarized as follows: (i) all

spectral changes are linear in the external field EO; (ii) in theglass, we have broadening only; (iii) in the protein, we have a

splitting and a broadening as well; (iv) broadening in the glassand that in the protein are of comparable magnitudes; (v)although the magnitude of the broadening is different for thethree frequencies measured in the protein, the qualitativeaspects of the Stark spectra do not depend on frequency.

Discussion

The characteristic features of a Stark spectrum-namely, thesplitting and the broadening-depend on two parameters ofthe model: Al.i, the molecular frame-fixed change of thedipole moment, and a-, the mean value of the matrix-induceddipole moments.The Glass Sample. There are two interesting and surprising

observations. In the glass there is no splitting but a broadeningonly, irrespective of the polarization, and the broadening is

0

E., kV cm-1

FIG. 3. Broadening of a spectral hole in the glass sample as afunction of the external field Ee. The plotted quantity is foaE0/h (seeEqs. 1 and 4).

35 O0

30 L20

125 ?r0 4 8 12

E., kVcm-1

FIG. 5. Splitting (fAA0E./h) and broadening (foEo/h) of a spec-tral hole as a function of the external field. The sample was proto-porphyrin IX in myoglobin. Temperature was 1.7 K. Burn frequencywas 16,021 cm-'. Also shown is the calculated angle -y between Al,oand the transition dipole moment Ileg.

= 0.650

* 0.6010

0o

0.55

5

- ~~~~~~.* *

1 1K S

2118 Biophysics: Gafert et al.

Dow

nloa

ded

by g

uest

on

July

14,

202

1

Proc. NatL Acad Sc. USA 92 (1995) 2119

comparable to the respective one in the protein. From theabsence of any measurable splitting we have to conclude thatAgo, the molecular frame-fixed change of the dipole moment,is very small compared with ur. Since we are dealing withallowed transitions, the symmetry of the excited state isdifferent from that of the ground state. Hence, the straight-forward conclusion to satisfy the observation Ag. = 0 wouldbe to assume that the dipole moment in the ground state aswell as in the excited state is close to zero. In this case, themolecular IT electron system must have approximate inversionsymmetry. That is, the asymmetric substituents-namely, thepropionic acid groups and the methyl as well as the vinylgroups-are of a negligible influence on the 7relectron system.(Although this argument seems to be straightforward, we haveto keep in mind that in the protein, things seem to bedifferent.) Then, if the symmetry argument above is right, wehave to conclude that the total contribution to the Starkbroadening comes from the matrix. The matrix destroys thesymmetry of the Xf system by inducing a dipole moment. Sincethe solvent cage around the dye probe in the glass varies in acompletely statistical fashion, the induced dipole moment israndom, giving rise only to a broadening.The Protein Sample. Splitting and spatial correlation. Con-

trary to the glass, we observe a splitting of the hole in theprotein; hence, we have to assume that Al0o 0. How can thisassumption be reconciled with what we concluded from theexperiment in the glass-namely, a nearly inversion symmetryfor the IT electron system of the chromophore?Our argument is based on the strong spatial correlation of

chromophore and protein environment. In other words, it isbased on the long-range structural order of the protein. Unlikein the glass, the immediate environment around the dye is thesame in all protein molecules. In this environment there arepolar and even charged groups. They create an electric field atthe chromophore. Because of the strict spatial correlation, thisfield has a definite direction in a chromophore fixed frame. Itinduces a well-defined dipole moment. Similar observationshave been made for a chromophore in a lipid bilayer (35).Hence, there is a A,uo vector which has a fixed direction in themolecular frame. (Although, here AI,o is a quantity which isinduced by the pocket field of the protein, it acts like amolecular quantity. Hence, we use the respective notation.) Itis this molecular frame-fixed quantity A1io which leads to theobserved splitting: AILO has a definite orientation (angle y)with respect to the transition dipole moment. Through thepolarized laser field an anisotropically oriented ensemble ofprotein molecules is selected. This selected ensemble gives riseto an anisotropic A/,o distribution (in the laboratory frame).If this anisotropic distribution has its major component par-allel to the external field, a splitting results. In the oppositecase there is only a broadening. Along these lines of reasoningit is clear that the splitting signals structural order and spatialcorrelation in the protein as opposed to the glass.There is another important aspect to be stressed. In the glass

as well as in the protein sample the interaction which domi-nates the Stark-effect features is the same: namely, the dipole-induced dipole interaction, which falls off asR6. We can drawa definite conclusion on the range of this interaction. Theexperiments clearly demonstrate that it dies out on a lengthscale which cannot significantly exceed the typical dimensionof the protein. For Mb this scale is on the order of 15 A. If theinteraction range exceeded significantly the dimension of theprotein, we would not observe the dramatic difference in theStark pattern of the holes. The chromophore would feel theinfluence of the host glass and, consequently, the spectral holewould not show such a pronounced splitting but would broadenonly. We consider this finding important with respect toinhomogeneous line broadening and related phenomena suchas pressure effects, because it demonstrates that the relevantinteractions are those between chromophore and protein and

not those between chromophore and host glass. In otherwords, the chromophore is indeed a probe for the structureand for the dynamics of the protein.Broadening and structural disorder. Inhomogeneous broad-

ening always reflects structural disorder. Stark broadening isinhomogeneous. Hence, we can take the width oa as a measureof structural randomness. Unfortunately, our experiments donot directly yield this quantity. Instead, we get the scaledquantity f.o, where f corrects for the local field. f may bedifferent for the protein and the glass, but for a rough estimatewe can safely assume thatf is of the same order of magnitudein both samples. Then we can directly compare the broadeningin both samples, and we have the astonishing result that it is ofthe same order of magnitude. In other words, the fluctuationof the induced dipole moment in the spatially highly correlatedprotein is comparable to that in the totally random glass phase.It is only the average value of the dipole moments which isclose to zero in the glass. We stress that these results from theStark experiment fit beautifully to the observations frompressure broadening experiments (30), where it was alsoconcluded that the scales of randomness in a protein and aglass are comparable, despite the high spatial correlation of theprotein building blocks.Summary and Conclusion. We have presented hole-burning

Stark experiments on protoporphyrin IX-substituted Mb andon a protoporphyrin IX-doped glass. We found a clear splittingof the hole in the protein, depending on the laser polarization,but not in the glass. The Stark broadening was of the sameorder of magnitude in the glass and protein sample, respec-tively.We conclude that (i) in the glass the relevant symmetry of

protoporphyrin IX must have an approximate inversion cen-ter; (ii) the splitting in the protein sample is induced throughthe high spatial correlation of chromophore and proteinbuilding blocks; and (iii) the specific Stark pattern in theprotein sample shows that the range of the relevant chro-mophore-protein interaction does not exceed typical dimen-sions of the Mb protein.

We gratefully acknowledge enlightening discussions with LotharKador (University of Bayreuth) and G. Ulrich Nienhaus (Universityof Illinois, Urbana). We acknowledge support from the DeutscheForschungsgemeinschaft (SFB 213, Graduiertenkolleg NichtlineareDynamik und Spektroskopie) and from the Fonds der ChemischenIndustrie.

1. Frauenfelder, H., Parak, F. & Young, R. D. (1988) Annu. Rev.Biophys. Biophys. Chem. 17, 451-479.

2. Parak, F. & Knapp, E. W. (1984) Proc. Natl. Acad. Sci. USA 81,7088-7092.

3. Frauenfelder, H. (1984) Helv. Phys. Acta 57, 165-187.4. Parak, F., Hartmann, H. & Nienhaus, G. (1987) in Protein

Structure: Molecular and Electronic Reactivity, eds. Austin, R.,Buhks, E., Chane, B., DeVault, D., Dutton, P. L., Frauenfelder,H. & Goldanskii, V. I. (Springer, New York), pp. 65-84.

5. Parak, F., Heidemeier, J. & Knapp, E. W. (1988) in Biological andArtificial Intelligence Systems, eds. Clementi, E. & Chin, S.(ESCOM, Leiden), pp. 23-47.

6. Iben, E. T., Braunstein, D., Doster, W., Frauenfelder, H., Hong,M. K., Johnson, J. B., Luck, S., Ormos, P., Schulte, A., Steinbach,P., Xie, A. H. & Young, R. D. (1989) Phys. Rev. Lett. 62,1916-1919.

7. Frauenfelder, H., Sligar, S. G. & Wolynes, P. (1991) Science 254,1598-1603.

8. Nadler, W. & Stein, D. L. (1991) Proc. Natl. Acad. Sci. USA 88,6750-6754.

9. Doster, W., Cusack, S. & Petry, W. (1989) Nature (London) 337,754-756.

10. Rasmussen, B. F., Stock, A. M., Ringe, D. & Petsko, G. A. (1992)Nature (London) 357, 423-424.

11. Singh, G. D., Schink, H. J., von Lohneysen, H., Parak, F. &Hunklinger, S. (1984) Z. Phys. B Condens. Matter 55, 23-26.

12. Yang, I.-S. & Anderson, A. C. (1986) Phys. Rev. B 34,2942-2944.

Biophysics: Gafert et al.

Dow

nloa

ded

by g

uest

on

July

14,

202

1

Proc. NatL AcadJ Sci USA 92 (1995)

13. Frauenfelder, H. (1986) in Structure and Dynamics of NucleicAcids, Proteins and Membranes, eds. Clementi, E. & Chin, S.(Plenum, New York), pp. 169-177.

14. Kohler, W. & Friedrich, J. (1989) J. Chem. Phys. 90, 1270-1273.15. Friedrich, J. (1991) in Light in Biology andMedicine, eds. Douglas,

R. H., Moan, J. & Ronto, G. (Plenum, New York), Vol. 2, pp.345-356.

16. Friedrich, J. (1990) Mol. Cryst. Liq. Cryst. 183, 91-103.17. Zollfrank, J., Friedrich, J., Fidy, J. & Vanderkooi, J. M. (1991)

Biophys. J. 59, 305-312.18. Zollfrank, J., Friedrich, J., Fidy, J. & Vanderkooi, J. M. (1991) J.

Chem. Phys. 95, 3134-3136.19. Huber, R. (1988) Ang. Chemie 100, 79-89.20. Frauenfelder, H., Petsko, G. A. & Tsernoglou, D. (1979) Nature

(London) 280, 558-563.21. Parak, F., Hartmann, H., Aumann, K. D., Reuscher, H., Ren-

nekamp, G., Bartunik, H. & Steigemann, W. (1987) Eur. Biophys.J. 15, 237-249.

22. Marchetti, A. P., Scozzafava, M. & Young, R. H. (1977) Chem.Phys. Lett. 51, 424-426.

23. Saimolenko, V. D., Razumova, N. V. & Personov, R. L. (1982)Opt. Spectrosc. (Engl. Transl.) 52, 346-348.

24. Bogner, U., Schatz, P. & Maier, M. (1983) Chem. Phys. Lett. 102,267-271.

25. Maier, M. (1986) Appl. Phys. B 41, 73-90.

26. Meixner, A. J., Renn, A., Bucher, S. E. & Wild, U. P. (1986) J.Phys. Chem. 90, 6777-6785.

27. Kador, L., Haarer, D. & Personov, R. I. (1987)J. Chem. Phys. 86,5300-5307.

28. Gafert, J., Friedrich, J. & Parak, F. (1993) J. Chem. Phys. 99,2478-2486.

29. Friedrich, J., Gafert, J., Zollfrank, J., Vanderkooi, J. M. & Fidy,J. (1994) Proc. Natl. Acad. Sci. USA 91, 1029-1033.

30. Gafert, J., Friedrich, J., Vanderkooi, J. M. & Fidy, J. (1994) J.Phys. Chem. 98, 2210-2214.

31. Boxer, S. G., Goldstein, R. A., Lockhart, D. J., Middendorf,R. T. & Takiff, L. (1989) J. Phys. Chem. 93, 8280-8294.

32. Mathies, R. & Stryer, L. (1976) Proc. Natl. Acad. Sci. USA 73,2169-2173.

33. Hanson, D. M., Patel, J. S., Winkler, I. C. & Morrobel-Sosa, A.(1983) in Spectroscopy and Excitation Dynamics of CondensedMolecular Systems, eds. Agranovich, V. M. & Hochstrasser, R. M.(North-Holland, Amsterdam), pp. 621-679.

34. Kohler, B. E., Personov, R. I. & Woehl, J. C. (1995) in LaserTechniques in Chemistry, eds. Rizzo, T. & Myers, A. (Wiley, NewYork), in press.

35. Schmidt, S. & Reich, R. (1972) Ber. Bunsen Ges. Phys. Chem. 76,1202-1208.

36. Schatz, P. & Maier, M. (1987) J. Chem. Phys. 87, 809-820.37. Overkamp, M., Twilfer, H. & Gersonde, K. (1976) Z. Naturfor-

sch. C Biosci. 31, 524-533.

2120 Biophysics: Gafert et aL

Dow

nloa

ded

by g

uest

on

July

14,

202

1