Stark effect experiments in cytochrome c-type proteins: structural hierarchies

Please download to get full document.

View again

of 9
2 views
PDF
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Stark effect experiments in cytochrome c-type proteins: structural hierarchies
Document Share
Document Tags
Document Transcript
  Biophysical Journal Volume 71 July 1996 77-85 Stark Effect Experiments in Cytochrome c-Type Proteins: Structural Hierarchies M. Kohier,* J. Gafert,* J. Friedrich,* J. M.Vanderkooi, and M. Laberge *Physikalisches Institut und Bayreuther Institut fOr MakromolekCilforschung, Universitdt Bayreuth, 95440 Bayreuth, Germany, and  JohnsonResearch Foundation, Department of Biochemistry and Biophysics,School of Medicine, University of Pennsylvania, Philadelphia, PA 19104 USA ABSTRACT We performed hole-burningStark effect experiments on cytochrome c in which the iron of the heme was either removed or replaced by Zn. According to the experiments, thefree-base compound has an effective inversion center, even in the protein. The Zn compound, on theother hand, shows quitepeculiar features: in the low-frequency rangeof the inhomogeneous band, it definitely has a dipole moment, as indicated by a splitting of thehole in the external field. However, in the maximum of the inhomogeneous band, asevere charge redistribution occurs, as the experiments show. In additionto the Starkexperiments, we performed calculations of the electrostatic fieldsat the pyrrole rings and at themetal site of the heme group. We interpret our findings with a modelbased on structural hierarchies: the protein can exist in a few subconformations, which can be distinguished through the structureof the heme pocket. The different pocket structures support different structures of the chromophore, which, in turn, can be distinguished through their behavior in an external field. These distinct structures, in turn, correspond to a rather broad distribution of proteinstructures, which leave, however, the pocket structure largely unchanged. These structures show up in inhomogeneous broadening. INTRODUCTION What is so fascinating about proteins is the factthat they reflect organization andrandomness at the same time (Frauenfelder et al., 1988; Friedrich, 1995). It seems that both are needed for a sufficiently high level of complexity (Frauenfelder andWolynes, 1994). Organization is, for in- stance, reflected in x-ray diffraction, whereas randomness is most directly reflected in inhomogeneous line broadening. It is a challenging problem in theoptical spectroscopy of chromoproteins to unravel the nature of inhomogeneous line broadening. Inhomogeneous line broadening reflects all the conformational disorder in a protein. It has been suggested that,in proteins, disorder is organized in hierarchies (Ansari et al., 1985). It is an intriguing task to find out whether thevarious hierarchy levels correspond to different physicaland/or functional properties as has been shown, for instance, forthe so-called taxonomic states in myoglobin (Hong et al., 1990; Frauenfelder et al., 1991). For the spectroscopistthere are several problems. First, thedifferent hierarchy levels may not be separated clearly enough in the spectrum. Second, it is not always straight- forward to find a spectroscopically accessible parameter whose changes are sufficientlylarge to be clearly measured as onegoes from one level to another. Probably the most sensitive techniques of unraveling substructures in inhomogeneous line broadening and asso- ciated changes in physical properties are frequency-selec- Received for publication 18 December 1995 and in final form 19 March 1996. Address reprintrequests to Dr. Josef Friedrich, Physikalisches Institut, Universitat Bayreuth, D-95440 Bayreuth, Germany. Tel.: 49-921-55-3245; Fax: 49-921-55-3250; E-mail: josef.friedrich@uni-bayreuth.de. i 1996 by the Biophysical Society 0006-3495/96/07n77/09 2.00 tive optical techniques, such as hole burning or photon echo techniques. With these techniques, inhomogeneous effects can be measured on thescale of the homogeneous linewidth (Friedrich and Haarer, 1984; Friedrich, 1990, 1991; ThornLeeson and Wiersma, 1995; Thorn Leeson et al., 1994). Thispaper presents a hole-burning Stark effect study on cytochrome c-type proteins. The Stark effect is very sensi- tive to the localfield generated from the matrix at the probe site (Kohler et al., 1995; Gafert et al., 1995a,b). Given the protein structure, this field can be calculated by solving the Poisson-Boltzmann equation (Honig and Nicholls,1995). The high sensitivity of the hole-burning Starkeffect to the local matrix field gives riseto interesting aspects in protein spectroscopy; performing the experiment at several frequen- cies in the inhomogeneous band reveals whether the se- lected conformational substates are associated with changes inthelocal field at the chromophore site. SPECIFIC ASPECTSOFA HOLE-BURNING STARK EFFECTEXPERIMENT The Stark effect on spectral holes shows a quitespecific feature-it is, as a rule, always linear in the external field, even if the probe molecule has inversion symmetry and does not possess a permanent dipole moment. The reason is that the field generated at the site of the probe by the matrix is orders of magnitude larger than the external field needed to change the shape and thecenter frequency of the hole (Maier, 1986; Meixner et al., 1986; Schatz and Maier, 1987; Kador et al., 1990). An order of magnitudemeasure of the ratio of internal and external field is the ratio of the inho mogeneous width to the hole width in an external field. Three cases are of interest: 1) The chromophore has inversion symmetry, the matrix is random. The Stark effect is determined through the dif- 77  Volume 71July 1996 ference of the induced dipole moments in the groundand excited states. This difference vector is random in magni- tude as well asorientation. This randomness is not affected by the polarization of the hole-burning laser, and, conse- quently, a hole will broaden only in an external field, irrespective of the laserpolarization. 2) The chromophore does not have inversion symmetry. As a rule, there will be a dipole moment difference vector with well-defined orientation. Polarized laser excitation cre-ates a macroscopically anisotropic distribution of thisdif- ference vector, which can be controlled with respect tothe external field by choosing the laser polarizationin a proper way. If the main axis of this distribution is parallel to the Stark field, the hole will split. Hence, a splitting canbe considered asthe signature of a dipole moment difference vector with a well-defined orientation in the frame of the chromophore. 3) The chromophore has inversion symmetry, but the matrix is ordered.In this case, the matrix will induce a dipole moment with a well-defined orientation. This situa- tion is equivalent to the presence of a permanent dipole moment. Hence, the hole willalso split. For proteins,theinteresting conclusions come from a comparison with the behavior of the chromophore in a glass. From the glass spectra, it can be inferred whether the chromophore hasinversion symmetry. If it has inversion symmetry, it can bebroken by the protein in thesense that a macroscopic anisotropy of the dipole moment difference vector can be generated with polarized light excitation (Gafert et al., 1995a,b). If so, the field at the chromophore site must be dominated by the ordered structure of the protein and not by the randomness of the host glass. In this way, one gets information on the range of the relevant interaction of the chromophore with its environment. From a variation of the Stark pattern with burn frequency we learn that the structure of the protein must be different at the fiequencies considered. In this way, one gets information on the correlation between frequency and structure and on the organization of the energy landscape. MATERIALS AND METHODS Speroscopy and sample preparation Zn and free-base cytochrome c were prepared as previously described (Vanderkooi et al., 1976). In cytochrome c the vinyl groups of protopor- phyrin IX are condensed withcysteines to give a chromophore, as shown in Fig. 1 for the Zn and free-base derivatives. The Stark effect experiments were performed for Zn-cytochrome c and the respective free base. For both chromophores we also performed the comparativeexperiments in a hostglass. We chose dimethylformamide/ glycerol in a volume rado of 1:3. The protein was dissolved in a KH2P04 buffer at pH 7. To ensure good optical qualityof the sample, the buffer was mixed with glycerol in a volume ratio of 1:2. At the concentrations used, the ODs at the maximum of the long-wavelength band were 0.9 for Zn-cytochrome c (ZnCc) and 0.35 for free-base cytochrome c (H2Cc). For Zn-protoprophyrin IX (ZnPP) and free-base protoporphyrin IX (H2PP) dissolved in the glass matrix, the ODs were 0.9 and 0.12, respectively. The samples were sealed in glass cuvettes (1 X 6 X 10 mm3) and placed between two electrodes. The respectivevoltage was varied between 0 and H3C H3C H3C 3 H3s C _ CH3 HSC.N. . I H3C  CCH3 CH2CH2COOHCH2CH2COOH FIGURE 1 The chromophores. The letters designate the four porphyrin pyrrole rings. Note that the heme chromophores are covalently bound to the apoprotein through S linkages of cysteines. 7 kV, corresponding to a variation in the field between 0 and 11.7 kY/cm.The laser polarization was either parllel orperpendicular to the Stark field. Hole burning was performed at 1.6 K with asingle-frequency ring dye laser operatedwith rhdamine 60. Holes were detected in transmission. Power levels and burning times wereon the orderof 100 ,u&W and100 s, respectively. To get sufficiently good signals, the holes were burned to relative depthsof 30%. For the homogeneous linewidth experiments onZnCc, we burned at each selected wavenumber a series of holes whose relative depths varied from 2 to 20%.The width of the hole extrapolated to zeroarea is twice the homogeneous linewidth. The frequencies at which hole-burniing Stark effect experiments were performed are marked with arows (Fig. 2 and 5). For the zinc chro- mophore it suffices to show results at two different frequencies. For the free-base chromophore, no frequency dependence was found, either for the glass or forthe protein, and thus we show results for one frequency only. The usualdata evaluaton that we usein hole-burning Stark effect experiments, namely averaging over all the angnlar degrees of freedom as well as over some properly chosen distribuion of induced dipole moments (Schptz and MaTer, 1987), could not be applied to ZnCc because of the unusuallysharp photoproduct peaks (Fig. 4). lnese experienced a shift in the Stark field aswell, so that baseline problemsoccurrd.Because of this, we will not compare teoe numbers but rather focus on the characterisc qualitatve features. 78 Biophysical Joumal  Stark Effect Experiments 0 1680017000 1720017400 wavenumber 17600 nell et al., 1990), the x-ray waters were included. The minimization was performed by using a distance-dependent dielectric. The structure was first subjected to a steepest descent minimization to achieve a maximum derivative of less than21kJ mol-F, followed by a conjugate gradient using a Newton-Raphson algorithm for which a residual maximum gradient of 0.42 kJ mol was set as a conver- gence criterion. Calculation of electric potential and field The electrostatic potential and field of cytochrome c were calculated on the minimized structure at specific heme sites (Fe, NA, NB, NC, ND; see Fig. 1), using finite-difference Poisson-Boltzmann calculations as implemented in the Del- phi software package (Gilson et al., 1988; Nicholls and Honig, 1991). This method has recently become a choice approach to solving the Poisson-Boltzmann equation for solvated molecules with complex charge distribution pat- terns, and we have recently used it to simulate the electro- static field imposed by cytochrome c peroxidase on cyto- chrome c (Anni et al., 1995). The method has been fully described elsewhere (Gilson et al., 1988; Sharp andHonig, 1990). Inclassicalelectrostatics, thespatialvariation in potential 4 at position r is related to the charge distribution pand the position-dependent dielectric permittivity E as follows: VE(r)VOr) = -4irp/kT. (1) FIGURE 2 Inhomogeneously broadened long-wavelength absorption band of Zn-protoporphyrin (ZnPP) in a glass matrix and incorporated in cytochrome c. Arrows indicate frequencies where Stark effect experiments were performed. Fat arrows refer to the data shown in Figs.3 and 4. The black dots show zero-fluence extrapolatedhole width measurements. COMPUTATIONAL METHODS Energy minimization Replacing theiron in cytochrome c with a zinc atom does not significantly affect thecentral metal environment(Anni et al., 1995, 1996), and we accordingly performed our calculations using coordinates generated from the x-ray structure of horse heart cytochrome c: pdblhrc.ent(Bush- nell et al., 1990),obtained from the Brookhaven Protein Data Bank (Bernstein et al., 1977). The energy was mini- mized using the Discover module of Insight II (Biosym Technologies, SanDiego) on a Silicon GraphicsIRIS In- digo workstationwith the CVFF force field modified forthe heme group (Laberge et al., manuscriptsubmitted for pub- lication). Missing hydrogens were added by using Discover (Biosym Technologies), subject to van der Waals con- straints. In cytochrome c, only 7.5% of the total heme surface can directly interact with the solvent, and the heme group is deeply buried in the surrounding protein matrix, which eliminates the need to include the solvent explicitly in the minimization. However, because four of theproto- porphyrin carbon atoms are exposed to the solvent (Bush- Inthe presence of ions,the Poisson equation can be com- bined with the Boltzmann equation, yielding: VE(r)V(r) - EK2 sinh(O(P)) = -41rp;/kT. (2) The term K2 is equal to 1/A2 or 8-rrq2IlekT, where A is the Debye length,I is the ionicstrength of thesolution, q is the proton charge, and pf is the fixed charge density. 4, E, K, andp are all functions of the vector r, with the secondterm in the equation describing the salt effect. Two dielectric con- stants are used to take into account thesolvationeffect experienced by the charged groupsof the protein in an aqueous solvent.In the algorithm used to solve Eq. 2, the solute-solventvariables were mapped on a three-dimen- sional grid (653) and charge, dielectricconstant, and ionic strength (0.075 mM were assigned to each lattice point. The longest dimension of the protein as a percentage of grid was 66%. Because ionic strength was taken into account, boundary conditions were set using the Debye-Hiickel equa- tion (Gilson et al., 1988). In the Delphi solution, thePois- son-Boltzmann equation is then replaced by a series of finite-difference equations, which are solved by iteration. The electrostatic energies are calculated as AG = Iqii, (3) where qi is the charge and  pi is the electrostatic potential at each atom i. The solute atomic charges and the van der 79 ohler et al.  Volume 71 July 1996 Waals radii are plugged into the calculation, alongwith the solute and solvent dielectric constants, 2 and 80, respec- tively (Young et al., 1993). The probe radius was set to1.4 A, and calculations weredone on differentscales and lattice mappings to ensure the uniformity ofcomputational results. Graphicswereproduced using the GRASP software pro- gram (Nicholls et al., 1991). Two calculations were performed. In the first, formal charges were assigned as follows: -0.5 to the Asp and Glu carboxyl oxygens, -0.5 to the oxygens of the two heme propionates, +0.5 to the Arg N, and to bothHis N6 and Nf, and   1 to the Lys N,. Inthe second calculation, the pro- pionates were left uncharged. RESULTS Zn-cytochrome c Fig. 2 compares the inhomogeneously broadened long- wavelength absorption band of ZnPP in a glass with ZnCc. Whereas the spectrumof theglass sample is smooth, the spectrum of the protein is structured. To elucidatethenature of thisstructure, we measured the quasi-homogeneous hole width for several frequencies acrossthe two long-wave- length peaks of the inhomogenous band. It stays constant at 300 MHz. This value corresponds with a coherence time of >1 ns characteristic for a purely electronic srcin. Hence, the first two peaks arise mainly from electronic srcins, whereas the third and possibly higher peaks must corre- spond with vibrational transitions because we did not suc- ceed in burning narrow holes into this part of the spectrum. Fig. 3 shows holes burned at two different wavenumbers (bold arrows in Fig. 2) into the spectrumof the glass sample. Although the holes do not show a splitting, there is a clearlydiscernible tendency of a flattening of the shape of the hole for a polarization parallel to the Stark field as one goes from higher to lower frequencies. This flattening is indicative of a permanent dipole moment difference, al- though it is rathersmall.Fig. 4 correspondswith Fig. 3. It shows the respective data of the protein. There is a striking feature: in the red edge of the band(16994 cm-') the hole clearly splits if thepolarization of the laser is parallelto the Stark field. The high-frequency hole(17,155 cm-1), however, just broad- ens, irrespective of polarization. We want to point outa quite unusual feature: the spec- trum shows sharply distributed photoproduct peakswith antiholecharacter. At present it is not clear where these features come from. One possibility could be that they are due to rotational tunneling transitionsin the methyl sub- stituents,as they havebeen observed, for instance, in di- methyl-s-tetrazine in an n-alkane lattice (Orth et al., 1993). In any case, weknow for sure that apart from such a narrow-bandwidth phototransformation, there is also a pho- toreversible broadband transformation that mutually con- verts the red and blue srcins (Jira, 1995). H2-cytochrome c Fig. 5 corresponds with Fig. 2. It compares the inhomoge- neous absorption bands of H2PP in a glass with H2Cc. Again, theprotein spectrum is structured, whereas thespec- trumof theglass sample is smooth. Arrows indicate where hole burning was performed. Figs. 6and 7 show hole spectra in an electric field forthe two samples. In both cases there is broadening only, irre- spective of laser polarization, and there are no frequency- dependent features. DISCUSSION The inhomogeneous spectra Theinhomogeneous spectra show common features. For both chromophores the glass spectra are smooth, whereas FIGURE 3Stark effect on spectral holes of Zn- protoporphyrin IX in a glassmatrix for two differ- ent frequencies. Polarization of the laserfield is either parallel or perpendicular to the Stark field Es,. C~ 9.9kV/cm 4 7 kV/cm M0.0 kV/cm Zn-PPlglass V 17081 cm- EL   Est 1.7 K 4.7 kV/cm 0.0°- kV/cm Zn-PP/glass || 17081 cm- EL 1L E5,t 1.7 K -10 0 10 frequency / GHz 10.5 kV/cm 5.3 kV/cm 0.0 kV/cm Zn-PP/glass 17208 cm- F_ 11 F 1 7K I   9.9 kV/cm 4.7 kV/cm 0.0 kV/cm Zn-PP/glass 17208 cm' ELL st ES1 7 K -10 0 10 frequency / GHz 80 Biophysical Joumal I1. I rvII L . St  Stark Effect Experiments FIGURE 4 Stark effect on spectral holes of Zn- cytochrome c for two different frequencies. Polar- ization of the laser field is either parallel or per- pendicular to the Stark field Es,.   -10 0 10 frequency / GHz *.   11.7 kV/cm 5.3 kV/cm 0.0 kV/cm Zn-Cc 17155 cm-' EL 1 ESt 1.7 K t -10 0 10 frequency / GHz the protein spectra are structured. As forthe srcin of this structure in H2Cc, a straightforward argument would be that it corresponds with theinner ring tautomer states. However, the structure in the spectrum of ZnCc must be of a different   15800 160001620016400 wavenumber nature. The homogeneous linewidth experiments support the presence of two electronic srcins. They obviously cor- respond with two distinct structures of ZnCc. In this context it should be stressed that the  two srcin feature does not seem to be specific to ZnCc. We found two srcins in all proteins substituted with Zn- and Mg-porphyrin derivatives that we have investigated so far. We suggest that the two structures are related to the two possible locations of the metalwith respect to the molecular plane. The two positions can be distinguished through the asymmetry of the protein environment,which, in tum, induces an asymmetric charge distribution at the heme site. This can be demonstrated by calculating the electrostatic potential at theiron and at the   16600 FIGURE 5 Inhomogeneously broadenedlong-wavelength absorption of free base protoporphyrin in a glass matrix and of free base cytochrome c. Arrows indicate frequencies at which Stark effect experiments were per- formed. Bold arrow refers to the data shown in Figs. 6 and 7. -10 0 10 frequency / GHz FIGURE 6 Stark effect on a spectral hole of free baseprotoporphyrin IX in a glass matrix.Polarization of thelaser field is eitherparallel or perpendicular to the Stark field Est. 11.7 kV/cm 5.3 kV/cm 0.0 kV/cm Zn-Cc 17155 cm' EL11 E 1.7 K 11.7 kV/cm 5.3 kV/cm 0.0 kV/cm H2-PP/glass 15977 cm' IEL   st 1.7 K   St~~I r11.7 kV/cm 5.3 kV/cm 0.0 kV/cm H2-PP/glass 15977 cm' EL 1L Est 1.7 K I   m .3 kV/cm 0.0 kV/cm Zn-Cc 16994 cm-i EL   Fst 1.7 K kV/cm 5.3 kV/cm 0.0 kV/cm Zn-Cc16994 cm-i ELI 1.7 K Fst 81 Kohler et al.
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks