The effect of water on the rate of conformational change in protein allostery.

The influence of solvation on the rate of quaternary structural change is investigated in human hemoglobin, an allosteric protein in which reduced water activity destabilizes the R state relative to T. Nanosecond absorption spectroscopy of the heme Soret band was used to monitor protein relaxation after photodissociation of aqueous HbCO complex under osmotic stress induced by the nonbinding cosolute poly(ethylene glycol) (PEG). Photolysis data were analyzed globally for six exponential time constants and amplitudes as a function of osmotic stress and viscosity. Increases in time constants associated with geminate rebinding, tertiary relaxation, and quaternary relaxation were observed in the presence of PEG, along with a decrease in the fraction of hemes rebinding CO with the slow rate constant characteristic of the T state. An analysis of these results along with those obtained by others for small cosolutes showed that both osmotic stress and solvent viscosity are important determinants of the microscopic R --> T rate constant. The size and direction of the osmotic stress effect suggests that at least nine additional water molecules are required to solvate the allosteric transition state relative to the R-state hydration, implying that the transition state has a greater solvent-exposed area than either end state.

Steric barrier to bathorhodopsin decay in 5-demethyl and mesityl analogues of rhodopsin.

Absorbance difference spectra were recorded from 20 ns to 1 micros after 20 degrees C photoexcitation of artificial visual pigments derived either from 5-demethylretinal or from a mesityl analogue of retinal. Both pigments produced an early photointermediate similar to bovine bathorhodopsin (Batho). In both cases the Batho analogue decayed to a lumirhodopsin (Lumi) analogue via a blue-shifted intermediate, BSI, which formed an equilibrium with the Batho analogue. The stability of 5-demethyl Batho, even though the C8-hydrogen of the polyene chain cannot interact with a ring C5-methyl group to provide a barrier to Batho decay, raises the possibility that the 5-demethylretinal ring binds oppositely from normal to form a pigment with a 6-s-trans ring-chain conformation. If 6-s-trans binding occurred, the ring C1-methyls could replace the C5-methyl in its interaction with the chain C8-hydrogen to preserve the steric barrier to Batho decay, consistent with the kinetic results. The possibility of 6-s-trans binding for 5-demethylretinal also could account for the unexpected blue shift of 5-demethyl visual pigments and could explain why 5-demethyl artificial pigments regenerate so slowly. Although the mesityl analogue BSI's absorption spectrum was blue-shifted relative to its pigment spectrum, the blue shift was much smaller than for rhodopsin's or 5-demethylisorhodopsin's BSI. This suggests that increased C6-C7 torsion may be responsible for some of BSI's blue shift, which is not the case for mesityl analogue BSI either because of reduced spectral sensitivity to C6-C7 torsion or because the symmetry of the mesityl retinal analogue precludes having 6-s-cis and 6-s-trans conformers. The similarity of the mesityl analogue BSI and native BSI lambda(max) values supports the idea that BSI has a 6-s angle near 90 degrees, a condition which could disconnect the chain (and BSI's spectrum) from the double bond specifics of the ring.

Effect of NADPH on formation and decay of human metarhodopsin III at physiological temperatures.

Difference absorption spectra were recorded during the formation and decay of metarhodopsin III after sonicated membrane suspensions of rhodopsin were bleached at 37 degrees C. The data were analyzed using SVD, spectral decomposition and global exponential fitting. By comparison of the results in the presence or absence of 70 microM NADPH and those for bovine or human rhodopsin, a single comprehensive scheme was fit to all the data, including reduction of retinal to retinol by the intrinsic retinol dehydrogenase. On the time scale studied the mechanism involves two 382 nm absorbing species and two 468 nm, absorbing species, supporting the notion that human metarhodopsin III is not a homogeneous species. The results confirm that metarhodopsin III forms and persists sufficiently long in the human retina under physiological conditions that it could undergo secondary photoisomerization.

Structural constraints imposed by a non-native disulfide cause reversible changes in rhodopsin photointermediate kinetics.

Suspensions of bovine rhodopsin in 2% lauryl maltoside detergent were treated with Cu(phen)(3)(2+) to form a disulfide bridge between cysteines 140 and 222 which occur naturally in the bovine rhodopsin sequence. Absorption difference spectra were collected after excitation with a pulse of 477 nm light on the time scale from 1 micros to 690 ms, and the results were analyzed using global exponential fitting. Only two exponentials could be fit to data from the Cu(phen)(3)(2+)-treated rhodopsin, while three exponentials were needed to fit data either from untreated rhodopsin or from Cu(phen)(3)(2+)-oxidized rhodopsin after further dithiothreitol reduction. Dithiothreitol treatment of rhodopsin which had not been previously oxidized with Cu(phen)(3)(2+) had no effect on the observed kinetics. Since the 140-222 disulfide has previously been shown to block transducin activation, its effects on rhodopsin activation are of considerable interest. Cu(phen)(3)(2+) treatment favors formation of the meta I(380) intermediate relative to meta I(480) and slows formation of meta II from meta I(380). This suggests that the protein change involved in meta I(380) formation is similar to the structural constraint introduced by the 140-222 disulfide. These results show that formation of disulfides in rhodopsin has potential as a tool for discriminating between the three isochromic, 380 nm absorbing intermediates involved in rhodopsin activation and for gaining insight into how their structures differ.

Near-ultraviolet magnetic circular dichroism spectroscopy of protein conformational states: correlation of tryptophan band position and intensity with hemoglobin allostery.

The near-UV magnetic circular dichroism spectroscopy of the aromatic amino acid bands of hemoglobin was investigated as a potential probe of structural changes at the alpha(1)beta(2) interface during the allosteric transition. Allosteric effectors were used to direct carp and chemically modified human hemoglobins into the R (relaxed) or T (tense) state in order to determine the heme-ligation-independent spectral characteristics of the quaternary states. The tryptophan magnetic circular dichroism (MCD) peak observed at 293 nm in the R state of N-ethylsuccinimide- (NES-) des-Arg-modified human hemoglobin (Hb) was shifted to a slightly longer wavelength in the T state, consistent with the shift expected for tryptophan acting as a proton donor in a T-state hydrogen bond. Moreover, the increase observed in the T-state MCD intensity of this band relative to the R-state intensity was consistent with the effect expected for proton donation by tryptophan on the basis of the Michl perimeter model of aromatic MCD. The peak-to-trough magnitude of the R - T MCD difference spectrum is equal to 30% of the total R-state peak intensity contributed by all six tryptophans present in the human tetramer; the relative magnitude specific to the two beta37 tryptophans undergoing conformational change is estimated accordingly to be 3 times larger. The Trp-beta37 spectral shift, about 200 cm(-)(1), is in good agreement with the shifts observed in other H-bonded proton donors and provides corroborating spectral evidence for the formation in solution of a T-state Trp beta37-Asp alpha94 hydrogen bond observed in X-ray diffraction studies of deoxyHb crystals.

Multiple geminate ligand recombinations in human hemoglobin.

The geminate ligand recombination reactions of photolyzed carbonmonoxyhemoglobin were studied in a nanosecond double-excitation-pulse time-resolved absorption experiment. The second laser pulse, delayed by intervals as long as 400 ns after the first, provided a measure of the geminate kinetics by rephotolyzing ligands that have recombined during the delay time. The peak-to-trough magnitude of the Soret band photolysis difference spectrum measured as a function of the delay between excitation pulses showed that the room temperature kinetics of geminate recombination in adult human hemoglobin are best described by two exponential processes, with lifetimes of 36 and 162 ns. The relative amounts of bimolecular recombination to T- and R-state hemoglobins and the temperature dependence of the submicrosecond kinetics between 283 and 323 K are also consistent with biexponential kinetics for geminate recombination. These results are discussed in terms of two models: geminate recombination kinetics modulated by concurrent protein relaxation and heterogeneous kinetics arising from alpha and beta chain differences.

Characterization of equilibrium intermediates in denaturant-induced unfolding of ferrous and ferric cytochromes c using magnetic circular dichroism, circular dichroism, and optical absorption spectroscopies.

Protein unfolding during guanidine HCl denaturant titration of the reduced and oxidized forms of cytochrome c is monitored with magnetic circular dichroism (MCD), natural CD, and absorption of the heme bands and far-UV CD of the amide bands. Direct MCD spectral evidence is presented for bis-histidinyl heme ligation in the unfolded states of both the reduced and oxidized protein. For both redox states, the unfolding midpoints measured with MCD, which is an indicator of tertiary structure, are significantly lower than those measured with far-UV CD, an indicator of secondary structure. The disparate titration curves are interpreted in terms of a compound mechanism for denaturant-induced folding and unfolding involving a molten globulelike intermediate state (MG) with near-native secondary structure and nonnative tertiary structure and heme ligation. A comparison of the dependence of the free energy of formation of the MG intermediate on the redox state with the known contributions from heme ligation and solvation suggests that the heme is significantly more accessible to solvent in the MG intermediate than it is in the native state.

pH dependence of photolysis intermediates in the photoactivation of rhodopsin mutant E113Q.

Glutamic acid at position 113 in bovine rhodopsin ionizes to form the counterion to the protonated Schiff base (PSB), which links the 11-cis-retinylidene chromophore to opsin. Photoactivation of rhodopsin requires both Schiff base deprotonation and neutralization of Glu-113. To better understand the role of electrostatic interactions in receptor photoactivation, absorbance difference spectra were collected at time delays from 30 ns to 690 ms after photolysis of rhodopsin mutant E113Q solubilized in dodecyl maltoside at different pH values at 20 degrees C. The PSB form (pH 5. 5, lambda(max) = 496 nm) and the unprotonated Schiff base form (pH 8. 2, lambda(max) = 384 nm) of E113Q rhodopsin were excited using 477 nm or 355 nm light, respectively. Early photointermediates of both forms of E113Q were qualitatively similar to those of wild-type rhodopsin. In particular, early photoproducts with spectral shifts to longer wavelengths analogous to wild-type bathorhodopsin were seen. In the case of the basic form of E113Q, the absorption maximum of this intermediate was at 408 nm. These results suggest that steric interaction between the retinylidene chromophore and opsin, rather than charge separation, plays the dominant role in energy storage in bathorhodopsin. After lumirhodopsin, instead of deprotonating to form metarhodopsin I(380) on the submillisecond time scale as is the case for wild type, the acidic form of E113Q produced metarhodopsin I(480), which decayed very slowly (exponential lifetime = 12 ms). These results show that Glu-113 must be present for efficient deprotonation of the Schiff base and rapid visual transduction in vertebrate visual pigments.

Lutein and zeaxanthin are associated with photoreceptors in the human retina.

PURPOSE: Previous studies showed that lutein and zeaxanthin, the major human retinal carotenoids, are concentrated in the macula. In this study, the carotenoids in human macular and peripheral retina and the retinal pigment epithelium (RPE) were analyzed. They were also determined in the rod outer segments (ROS) before and after removal of extrinsic membrane proteins. METHODS: Carotenoids were extracted from the macular and peripheral sections of human retina and RPE with hexane in dim light and analyzed by high performance liquid chromatography (HPLC). ROS samples equivalent to the amount in a single retina were also analyzed. RESULTS: Retinal carotenoid amounts were similar to previous reports, but only low levels were detected in the RPE. Regional ratios of lutein:zeaxanthin were similar in the retina and RPE. Approximately 25% of the total retinal carotenoids were found in the ROS, indicating that a substantial portion of peripheral retinal carotenoids are present in the ROS. However, after removal of the extrinsic membrane proteins and subsequent analysis, carotenoids were not detected. CONCLUSIONS: Most of the carotenoids in the human peripheral retina are present in the ROS. These ROS carotenoids are associated with soluble or salt-dependently bound proteins.

Multiple pathways on a protein-folding energy landscape: kinetic evidence.

The funnel landscape model predicts that protein folding proceeds through multiple kinetic pathways. Experimental evidence is presented for more than one such pathway in the folding dynamics of a globular protein, cytochrome c. After photodissociation of CO from the partially denatured ferrous protein, fast time-resolved CD spectroscopy shows a submillisecond folding process that is complete in approximately 10(-6) s, concomitant with heme binding of a methionine residue. Kinetic modeling of time-resolved magnetic circular dichroism data further provides strong evidence that a 50-microseconds heme-histidine binding process proceeds in parallel with the faster pathway, implying that Met and His binding occur in different conformational ensembles of the protein, i.e., along respective ultrafast (microseconds) and fast (milliseconds) folding pathways. This kinetic heterogeneity appears to be intrinsic to the diffusional nature of early folding dynamics on the energy landscape, as opposed to the late-time heterogeneity associated with nonnative heme ligation and proline isomers in cytochrome c.

Spectroscopic evidence for nanosecond protein relaxation after photodissociation of myoglobin-CO.

Nanosecond time-resolved absorption and magnetic optical rotatory dispersion (MORD) measurements of photolyzed myoglobin-CO visible bands (500-650 nm) are presented. These measurements reveal a 400 ns process, spectrally distinct from ligand recombination, that accounts for 7% of the observed spectral evolution in the visible absorption bands and 4% in the MORD. The time-resolved MORD, more sensitive to heme coordination geometry than absorption, suggests that this process is most likely associated with protein relaxation on the distal side of the heme pocket, perhaps accompanying rehydration of the deoxymyoglobin photoproduct or accommodation of protein side chains to ligand escape.

The effects of octanol on the late photointermediates of rhodopsin.

Membrane suspensions of unperturbed rhodopsin and rhodopsin perturbed with 2.5 mM octanol were photolyzed with 477 nm laser pulses at 20 degrees C and 35 degrees C. Changes in absorbance were monitored at times ranging from 1 microsecond to 80 ms after excitation. The data were analyzed using singular value decomposition, global exponential fitting and kinetic modeling. A recently proposed model involving the photointermediate Meta-I380 (T. E. Thorgeirsson, J. W. Lewis, S. E. Wallace-Williams, and D. S. Kliger, Biochemistry 32, 13861-13872, 1993) fits data for samples with and without octanol. Comparison of the microscopic rates shows this alcohol accelerates the formation of Meta-II via Meta-I380. Activation and equilibrium thermodynamic parameters obtained from Arrhenius plots suggest that octanol reduces the entropy increase in forming both Meta-I380 and Meta-II. It also lowers the enthalpy of Meta-I380 relative to Lumi and of Meta-II relative to Meta-I480. To help determine whether octanol affects the protein directly or indirectly through the lipid bilayer, similar experiments were conducted using rhodopsin solubilized in 0.13% dodecyl maltoside with and without octanol. Spectral shifts in the presence of octanol suggest that a direct protein interaction exists in addition to previously reported effects dependent on membrane free volume.

Proton transfer reactions linked to rhodopsin activation.

Purified bovine rhodopsin solubilized in dodecyl maltoside was photolyzed at 20 degreesC with 477 nm light, and difference spectra were collected at time delays ranging from 10 micros to 10 ms after photolysis. Bromocresol purple was added to the samples to detect pH changes in the aqueous environment due to changes in the protonation state of rhodopsin. The data were analyzed using singular value decomposition and global exponential fitting, which revealed three exponential processes indicating the presence of at least four intermediates. Spectral changes of the indicator dye were separated from those of rhodopsin, and proton release and uptake rates were analyzed within the framework of rhodopsin photoreaction kinetics. Proton release occurred during Lumi decay to Meta-I380 followed by uptake upon Meta-I380 decay and by a more significant proton uptake with the time course of Meta-I480 decay. On the basis of the estimated number of protons released and taken up in each step of the rhodopsin photoreaction, we concluded that two forms of Meta-II are present. The two forms of Meta-II, Meta-IIa' and Meta-IIb, differ in protonation state from one another as do both from the earlier, 380 nm absorbing form, Meta-I380.

Effects of pH on rhodopsin photointermediates from lumirhodopsin to metarhodopsin II.

Time-resolved absorption difference spectra of membrane suspensions of bovine rhodopsin at pH 5, 6, 7, 8, 9, and 10 were collected in the time range from 1 micro s to 200 ms after laser photolysis with 7-ns pulses of 477-nm light. The data were analyzed using singular value decomposition (SVD) and global exponential fitting. At pH 7 the data agree well with previously obtained data (Thorgeirsson et al. (1993) Biochemistry 32, 13861-13872) with fits improved at all pH's by inclusion of a small component due to an absorbance change caused by rotational diffusion which is detectable even at magic angle polarization. A "square scheme" suggested to best explain the previous data, which involves two branches following decay of the lumi intermediate with pathways (1) lumi --> MI480 right harpoon over left harpoon MII and (2) lumi right harpoon over left harpoon MI380 --> MII, could be confirmed throughout the entire pH range. However, to account for the increased rate of the MII --> MI480 reaction in path 1 for rising pH values, we propose that the MII in the square scheme consists of deprotonated MII and protonated MIIH+ forms in rapid equilibrium with each other, resulting in an extended square scheme and increasing the number of 380-nm products from two to three. In addition to the kinetic processes described by the extended square scheme, above pH 8 fast ( approximately 10 micro s) and slow ( approximately 50 ms) components were found. The fast component was assigned to the decay of a blue-shifted lumi intermediate, and the slow component, resolvable only at pH 10, was assigned to formation of a 450 nm absorbing photoproduct.

Time-resolved circular dichroism studies of protein folding intermediates of cytochrome c.

The circular dichroism spectra of cytochrome c (cytc) in 4.6 M guanidine hydrochloride (pH 6.5) indicate that the secondary structure in reduced cytc is near-native, whereas in the CO-bound species (COCytc) it is substantially unfolded. Photolysis of COCytc should thus induce large changes in the secondary structure, which can be probed with time-resolved circular dichroism (TRCD) spectroscopy in the far-UV region. Time-resolved absorption (TROA) and TRCD methods were used to study the photolysis reaction of COCytc in efforts to identify structural intermediates in cytc folding on time scales from nanoseconds to seconds. TROA data from the Soret region, similar to previous studies, showed four intermediates with lifetimes of 2, 50, 225, and 880 micros. The 2-micros process is proposed to involve Fe(II)-Met80 coordination. Approximately 7% of the native CD signal was observed in the TRCD signal at 220 nm within 500 ns, with no significant additional secondary structure formation observed. Further folding after 2 micros may be inhibited by ligation of His26/His33 with Fe(II), which is suggested to be associated with the 50-micros phase. The two slowest components, tau = 225 and 880 micros, are attributed to CO rebinding on the basis of mixed-gas experiments. CO rebinding is expected to compete with protein folding and favor the unfolded state. However, when the two CO rebinding lifetimes are extended into milliseconds by reducing the CO concentration, there is still no significant increase in CD signal at 220 nm.

Early photolysis intermediates of gecko and bovine artificial visual pigments.

Nanosecond laser photolysis measurements were conducted on digitonin extracts of artificial pigments prepared from the cone-type visual pigment, P521, of the Tokay gecko (Gekko gekko) retina. Artificial pigments were prepared by regeneration of bleached gecko photoreceptor membranes with 9-cis-retinal, 9-cis-14-methylretinal, or 9-cis-alpha-retinal. Absorbance difference spectra were recorded at a sequence of time delays from 30 ns to 60 microseconds following excitation with a pulse of 477-nm actinic light. Global analysis showed the kinetic data for all three artificial gecko pigments to be best fit by two-exponential processes. These two-exponential decays correspond to similar decays observed after photolysis of P521 itself, with the first process being the decay of the equilibrated P521 Batho<-->P521 BSI mixture to P521 Lumi and the second process being the decay of P521 Lumi to P521 Meta I. In spite of its large blue shift relative to P521, iso-P521 displays a normal chloride depletion induced blue shift. Iso-P521's early intermediates up to Lumi were also blue-shifted, with the P521 Batho<-->P521 BSI equilibrated mixture being 15 nm blue-shifted and P521 Lumi being 8 nm blue-shifted relative to the intermediates formed after P521 photolysis. The blue shift associated with the iso-pigment is reduced or disappears entirely by P521 Meta I. Similar blue shifts were observed for the early intermediates observed after photolysis of bovine isorhodopsin, with the Lumi intermediate blue-shifted 5 nm compared to the Lumi intermediate formed after photolysis of bovine rhodopsin. These shifts indicate that a difference exists between the binding sites of 9- and 11-cis pigments which persists for microseconds at 20 degrees C.

Properties of early photolysis intermediates of rhodopsin are affected by glycine 121 and phenylalanine 261.

Glycine 121 in transmembrane (TM) helix 3 and phenylalanine 261 in TM helix 6 of bovine rhodopsin have been shown to be critical residues for creating an appropriate chromophore binding pocket for 11-cis-retinal [Han, M., Lin, S. W., Smith, S. O., and Sakmar, T. P. (1996) J. Biol. Chem. 271, 32330-32336; Han, M., Lin, S. W., Minkova, M., Smith, S. O., and Sakmar, T. P. (1996) J. Biol. Chem. 271, 32337-32342]. To further explore structure-function relationships in the vicinity of receptor helices 3 and 6, time-resolved absorption difference spectra of rhodopsin mutants G121A, G121V, and G121L/F261A were obtained at 20 degrees C. Data were collected from 30 ns to 690 ms after laser photolysis with 7 ns pulses (lambdamax = 477 nm) and analyzed using a global exponential fitting procedure after singular value decomposition (SVD). For each mutant, the decay of its bathorhodopsin photoproduct (batho) into an equilibrium with its blue-shifted intermediate (bsi) was too fast to resolve (<20 ns). The reaction scheme found for the mutants G121A and G121L/F261A was batho/bsi --> lumirhodopsin (lumi) --> metarhodopsin I (MI) --> metarhodopsin II (MII). For G121V, an additional early 380 nm absorber, with a back-reaction to lumi, had to be included in the above scheme. For the three Gly121 mutants, the main pathway to reach the active MII state is via lumi and MI. This is in contrast to rhodopsin where the main pathway in detergent samples is via lumi and an early 380 nm absorber, MI380. From the accelerated batho decay present in all three mutants, we conclude that Gly121 is likely to participate in the earliest chromophore-protein interactions. In addition, bsi decay is further accelerated in mutant G121L/F261A, suggesting that Phe261 is an essential determinant of the protein processes involved in bsi decay.

Chromophore structural changes in rhodopsin from nanoseconds to microseconds following pigment photolysis.

Rhodopsin is a prototypical G protein-coupled receptor that is activated by photoisomerization of its 11-cis-retinal chromophore. Mutant forms of rhodopsin were prepared in which the carboxylic acid counterion was moved relative to the positively charged chromophore Schiff base. Nanosecond time-resolved laser photolysis measurements of wild-type recombinant rhodopsin and two mutant pigments then were used to determine reaction schemes and spectra of their early photolysis intermediates. These results, together with linear dichroism data, yielded detailed structural information concerning chromophore movements during the first microsecond after photolysis. These chromophore structural changes provide a basis for understanding the relative movement of rhodopsin's transmembrane helices 3 and 6 required for activation of rhodopsin. Thus, early structural changes following isomerization of retinal are linked to the activation of this G protein-coupled receptor. Such rapid structural changes lie at the heart of the pharmacologically important signal transduction mechanisms in a large variety of receptors, which use extrinsic activators, but are impossible to study in receptors using diffusible agonist ligands.

Deriving reaction mechanisms from kinetic spectroscopy. Application to late rhodopsin intermediates.

A general algebraic approach to the kinetic analysis of time-dependent absorption data is presented that allows the calculation of possible kinetic schemes. The kinetic matrices of all possible reaction mechanisms are calculated from experimental eigenvalues and eigenvectors derived from the decay constants and amplitude spectra (b-spectra) of the global exponential fit to the time-dependence of the absorption data. The eigenvalues are directly related to the decay constants, and the eigenvectors are obtained by decomposing the b-spectra into spectral components representing the intermediates. The analysis method is applied to the late intermediates (lumi, meta I, meta I-380, and meta II) of the rhodopsin photoreaction. The b-spectra are decomposed into lumi, meta I, meta-380, and rhodopsin spectra. The meta-380 component is partitioned into isospectral meta I-380 and meta II components based on physical criteria. The calculated kinetic matrices yield a number of reaction mechanisms (linear scheme with back reactions, branched schemes with equilibrium steps, and a variety of square models) consistent with the photolysis data at 25 degrees C. The problems associated with isospectral intermediates (meta I-380 and meta II) are treated successfully with this method.