Partitioning of an Anchor Dipeptide in a Phospholipid Membrane - The

19 Nov 2009 - The results provide a descriptive picture of the depth and geometry of partitioning of a guest N-myristoylated methyl glycine anchor dip...
0 downloads 0 Views 762KB Size
Download PDF
16246

J. Phys. Chem. B 2009, 113, 16246–16250

Partitioning of an Anchor Dipeptide in a Phospholipid Membrane Victor V. Volkov* and Roberto Righini† European Laboratory for Nonlinear Spectroscopy (LENS), Via Nello Carrara 1, I-50019 Sesto Fiorentino, Italy ReceiVed: June 30, 2009

We explore the localization of a guest N-myristoylated methyl glycine anchor dipeptide in a phospholipid environment. The dipeptide is part of a conservative sequence, which ensures proper association of a wide group of proteins in living organisms with a cellular phospholipid membrane. Using linear and two-color anharmonic infrared spectroscopy, we measure relative degrees of hydration of the amide I modes of the dipeptide and of phospholipid carbonyls. The atomic density of water in dependence of the distance from the hydrophobic center of the bilayer (a result of an independent Neutron scattering experiment) allows us to determine the relative altitudes of the peptide carbonyls with respect to those of the phospholipid ones. Considering this, and the dimensions of the dipeptide molecular frame, we anticipate the average angle between the backbone of the dipeptide and the normal to the membrane surface. The results provide a descriptive picture of the depth and geometry of partitioning of a guest N-myristoylated methyl glycine anchor dipeptide into a phospholipid membrane. Introduction Structural and dynamic properties of small proteins are of practical importance since polypeptides demonstrate healing propensity in treatment of physiological malfunctions and various forms of cancer, in particular.1-4 Even though injected intravenously into blood plasma, the efficiency of such remedies is often settled upon partitioning into the phospholipid membrane. This is because the phospholipid bilayer either hosts the target sites or mediates the molecular transport to it.5,6 There are several forms of association with a bilayer. Some proteins, such as G-proteins, porins and the like, play both a constituent and a functional role, penetrating the bilayer as integrated helical cage systems or barrel-like structures and, at the same time, carrying important functions such as ion transport and/or signal transduction.7 Others may associate with it temporally through electrostatic and hydrophobic interactions with specific chemical moieties at the phospholipid membrane interface.7-9 According to the third format of association, polypeptide partitioning and its specificity are due to the presence of hydrophobic tails (anchors) attached to certain conservative structural segments of the polypeptide.7 We may appreciate well the significance of the latter mechanism of association since anchoring provides an opportunity for a stable, specific, and controllable form of interaction of polypeptides with the membrane surface (of certain composition and fluidity) on the time scale of cellular metabolism. Generally, our knowledge of the structural properties of the peptide-membrane association comes from the results of the experimental determination of macroscopic thermodynamic properties via calorimetric techniques and of statistical distributions of microscopic properties obtained by means of different techniques, X-ray, electron, and neutron diffraction, optical spectroscopy, and optical microscopy. Scattering experiments and conventional spectroscopy, in general, and optical spectroscopy, in particular, allow characterization of autocorrelation * To whom correspondence should be addressed. Phone 0039 (0)55 457 2494. E-mail: [email protected]. † E-mail: [email protected].

(two-point correlation) functions and, consequently, are unable of measuring intermolecular correlations. In result, an exact appreciation of intermolecular structural relations in liquid phospholipid membranes is, currently, in critical dependence on the predicting power of classical molecular dynamics (MD) simulations. However, this computational technique is far from producing a completely reliable picture of the physics and chemistry in supramolecular structures; in particular, it is known to fail to reproduce precisely the molecular and the atomic densities in phospholipid membranes. Recently, multidimensional time-resolved infrared10-12 and visible optical13 spectroscopy have been employed as powerful tools to probe experimentally the intermolecular structure and dynamics on the time scale of the atomic motion. Infrared (IR) multidimensional spectroscopy, being sensitive to local structural dynamics, is particularly helpful in the structural analysis of disordered mesoscopic molecular composites, such as liquid phospholipidmembranefragmentsunderphysiologicconditions.14-17 Recently, the technique was successfully applied to the structural and dynamical analysis of several polypeptides (including those of promising anticancer activity2) associated with a phospholipid bilayer.14,18 In this article, we demonstrate the resolving power of twocolor ultrafast nonlinear IR spectroscopy to determine the depth and geometry of partitioning of a guest N-myristoylated methyl glycine (MrG) dipeptide (Figure 1) associated with a hosting phospholipid bilayer made of 1-palmitoyl-2-linoleyl phosphatidylcholine (PLPC). The dipeptide is part of a very conservative sequence in many organisms.19,20 Being the result of posttranslational modification, the N-myristoylated terminal plays important roles in cellular methabolism.21,22 Of course, the structural and dynamic properties of the terminal, while in the membrane, are not known and are the subject of our investigation. To accomplish the task, we employ two IR pulses; the first one excites the water stretching modes, and the second one measures the changes in the spectrum of the amide I modes induced by the excitation of the water vibration. Two facts provide the basis for the structural analysis that we develop here. First, a well-defined structural correlation exists between the

10.1021/jp9082536 CCC: $40.75  2009 American Chemical Society Published on Web 11/19/2009

Anchor Dipeptide in a Phospholipid Membrane

J. Phys. Chem. B, Vol. 113, No. 50, 2009 16247

Figure 1. Structure of the N-myristoylated methyl glycine dipeptide and linear and nonlinear spectra of membrane fragments with this molecule as indicated in the panels.

carbonyl groups and the water molecules due to formation of effective hydrogen bonds upon partitioning of water into the membrane. Second, the water density distribution across the bilayer thickness is known from independent experiments. Accordingly, we can relate the two-color infrared differential signal to the localization of the CdO groups within the membrane. Methods Section The N-myristoylated methyl glycine dipeptide (see Figure 1) is synthesized by BIOSYNTAN GmbH, Berlin, Germany, using standard solid-state peptide synthesis. In order to avoid possible spectral degeneracy of the amide carbonyls, we employ 13Cisotope-labeled myristic acid in the synthesis of the dipeptide. Figure 1 shows the position of the carbon isotope in the resultant structure of the dipeptide. The purity of the dipeptide is better than 95%. We remove the trifluoroacetic acid (TFA, left after reverse-phase HPLC) by lyophilizing the acquired substance under low pH (pD) conditions three times. Soy L-R-phosphatidylcholine 95% extract is from AVANTI-POLAR-LIPIDS (product number 441601G). 1-Palmitoyl-2-linoleyl phosphatidylcholine (PLPC) is the main molecular component of the extract. In order to prepare membrane fragments, first we mix a solid powder of PLPC and MrG dipeptide in a ratio of 10:1 mechanically. Second, we dissolve the mixture in chloroform and dry it on a glass plate. Further, after addition of D2O, the suspension undergoes mechanical mixing (under elevated temperature: 60 °C) until the sample (50 µm film between two CaF2 windows) demonstrates proper optical quality. We recorded the steady-state IR spectra using a FTIR spectrometer (Shimadzu 8400S). FTIR spectral properties are particularly helpful in characterization of the molecular composition in the prepared samples. As the suspension is prepared upon addition of D2O, we do observe a trace of native water in the samples as a OH impurity. However, since the FTIR spectrum does not show any obvious optical density characteristic of the HOD bending mode, we disregard any possible presence of H2O molecules in the samples. In the course of preparation, we notice that one of the amide I modes (native) may experience a 2-3 wavenumbers blue shift within a few hours afterward. Therefore, we let the sample “age” for several hours to reach spectral stability. We detected the nonlinear IR spectra of MrG dipeptide in phospholipid membrane fragments using a spectrometer based on a Ti:Sapphire laser-amplifier system (Legend, Coherent Inc.) producing a 1 kHz train of 30 fs (fwhm), 3 mJ pulses at 800 nm. This output was split to generate a 2 µJ mid-IR pump pulse

(vertical polarization) in an optical parametric amplifier (OPA) TOPAS (Light Conversion Ltd., Vilnius, Lithuania) and a 1.5 µJ mid-IR probe pulse in a home-built OPA. The spectral width of both OPAs is about 200 cm-1 in the mid-IR spectral range. In the experiment, we excite the D2O stretching modes and probe the perturbed CdO vibration using 100 fs pulses centered at 2300 (for the pump) and at 1700 cm-1 (for the probe), respectively. The bandwidth of the pump pulse covers the low and central frequency components of the D2O stretching band (see Figure 1). This allows us to excite water associated in hydrogen bonding (including those with carbonyls), and at the same time, we avoid possible excitation of CH modes (centered at 2900 cm-1) of the phospholipid and dipeptide. A molecular filter (50 µm thick H2O layer) in the path of the pump pulse ensures that the red wing of excitation does not affect the probe process in the 6 µm spectral region. Before it reaches the sample, a fraction of the probe pulse is split off and used as a reference. After the sample, both probe and reference pulses are spectrally dispersed in a spectrometer (TRIAX 180, HORIBA Jobin Yvon, Milano, Italy) and imaged separately on a 32 channel double array mercury cadmium telluride detector (InfraRed Associates Inc., Florida, U.S.A.). The ratio of the probe to the reference spectra gives the differential absorption spectrum, which is recorded as a function of the pump-probe time delay. Results and Discussion Figure 1 (upper panels) shows the FTIR spectra of the sample in the excitation (right side) and detection (left side) spectral regions. The resonances at 1587 and 1633 cm-1 are due to the 13 C-labeled and native amide I modes of the dipeptide, respectively. The small peak at 1668 cm-1 is a combination band of the dipeptide. The dominant band in the linear spectrum at 1734 cm-1 is due to the collective carbonyl stretching modes of PLPC.23 We estimate the molecular composition of the sample from the integrated optical densities of the proper stretching modes in the FTIR spectrum. The integration ranges for 13C, for native amide I, for phospholipid CdO, and for water OD stretching modes are 1570-1610, 1630-1660, 1670-1760, and 1980-2780 cm-1, respectively. First of all, on the basis of the integrated intensities, we estimate that the lipid-to-water ratio in samples of the best optical quality is 1:6. A single phospholipid molecule may coordinate up to 14 water molecules in the first shell;24,25 the degree of hydration of the polar groups is then substantial. Second, the ratio of the integrated optical densities in the carbonyls stretching region of PLPC and MrG dipeptide is about 8-10:1. This value is consistent with the molecular ratio used in preparation. In result, we consider that

16248

J. Phys. Chem. B, Vol. 113, No. 50, 2009

the dipeptide molecule is present in the prepared sample as an isolated impurity. The left and the right side lower panels show the CdO and D2O responses, respectively, upon excitation of D2O modes. The differential spectrum in the right panel is the “diagonal” response obtained when both pump and probe pulses are resonant with the heavy water absorption; we described and rationalized these results in previous publications.15,16 The left lower panel represents the responses of carbonyls upon OD excitation; we observe sharp ripples of the differential optical density in the spectral regions characteristic for each moiety as marked in the FTIR spectrum. These two-color signals are more than 10 times weaker than the signal measured in the singlecolor experiment. The nonlinear response of the carbonyl moieties appears instantaneously upon excitation of the water stretching modes. We ascribe it to the off-diagonal anharmonic coupling between the stretching vibrations of the excited OD modes and of the CdO moieties involved in hydrogen bonding to water.26 In other words, excitation of the OD stretching mode perturbs (flattens) the potential surface of the interacting carbonyl groups, leading to the red shift of their frequencies, as observed in the differential signal. The instantaneous appearance of the signal allows us to rule out its possible attribution to thermal effects. In fact, appearance of a thermal contribution in the time-resolved response (due to the upset of the Boltzmann population of low-frequency states) implies energy relaxation and spatial redistribution; it is a multistep process that, even in its initial stage, is characterized by a time constant in the range 1.2-2 ps.16,17 We can then consider the two-color anharmonic signal to be informative of the carbonyl-water pairing at least within the first 300 fs, when the anisotropic heat contribution is below 10-15% of the overall response. On the other hand, at very short time delay, when pump and probe pulses overlap, multiphoton interactions and resonant cross-phase modulation complicate the spectral response. Therefore, considering the duration of pulses employed, we limit our analysis to the nonlinear spectra recorded at time delays longer than 200 fs. Since in this time interval the CdO response arises upon direct excitation of the stretching vibration of the OD bond directly involved in hydrogen bonding with the carbonyl, the two-color anharmonic signal has to be ascribed to the carbonyl-water pairs only. The relative amplitudes of the nonlinear spectral features at 1585, 1645, and 1725 cm-1 can then be taken as proportional to the degree of hydration of the peptide carbonyls and of the phospholipid CdO groups, respectively. Our first step is then that of integrating the absolute value of the nonlinear spectral features corresponding to the 13 C amide I, to the native amide I, and to the phospholipid carbonyl vibrations (left lower panel in Figure 1). Next, a normalization procedure is necessary in order to take into account the relative abundances of the three moieties and their absorption coefficients. This is achieved simply by dividing the amplitudes of the three integrated nonlinear signals by the integrated intensities of the corresponding FTIR bands (left upper panel in Figure 1). The normalized intensities of the amide nonlinear responses are finally normalized to that of phospholipid carbonyls. As a result, if the normalized integral nonlinear response of an amide I carbonyl is larger than that of the phospholipid CdO group (their ratio >1), then we consider the amide I carbonyl to be better hydrated than the phospholipid carbonyls. Figure 2 shows the histograms of normalized integral intensities of the signals of two amides relative to that of the phospholipid carbonyls. The data are collected in distinct

Volkov and Righini

Figure 2. Histograms of the normalized integral intensities of the signals of two amides relative to those of the phospholipid carbonyls (see Results and Discussion). The red and blue full lines are Gaussian fits to these distributions. xc and S.D. are the center of distribution and the standard deviation, respectively.

Figure 3. (A) Phospholipid carbonyl (black line) and water (blue line) atomic density distributions according to ref 27. The magenta line is the atomic density distributions of the hydrated phospholipid carbonyls. (B) Horizontal red dashed lines and vertical red bars: altitudes of MrG carbonyls and corresponding standard deviations with respect to the hydrophobic center of the bilayer.

experiments and for different delay times within 200 and 300 fs. The red and blue full lines are Gaussian fits of these distributions. Correspondingly, the standard deviations are 2×(2(ln 2))0.5 times smaller than the fwhm of the Gaussians. The results indicate that the 13CdO is, on average, about 2.6 times more hydrated than the phospholipid carbonyls, while the native CdO of the peptide has a degree of hydration ∼1.2 times higher than that of the phospholipid carbonyls. We may rationalize these results in terms of the relative altitudes (with respect to the hydrophobic center of the bilayer) of the three involved carbonyl moieties at the membrane interface. In order to obtain a quantitative estimate, we employ the results on the atomic density of the carbonyl and water in hydrated membrane fragments obtained from neutron scattering studies.27 Figure 3A represents the atomic densities of the water molecules (blue line) and of the phospholipid carbonyl groups (black line) in dependence of the distance from the center of the bilayer toward the polar interface, as reported by S. White in ref 27. The density of water decreases steeply moving toward the inner part of the bilayer as a consequence of the hydrophobic nature of the hydrocarbon tails. On the other side, the spatial distribution of CdO atomic density peaks at 15.7 Å and may serve as a good structural reference. The product of the two distributions (magenta line in Figure 3A) corresponds to the probability of finding a hydrated phospholipid CdO group when moving from the center to the surface of the membrane. The first momentum of this density, peaked at 17.2 ( 1.6 Å from the center (dash magenta line in Figure 3A), represents the

Anchor Dipeptide in a Phospholipid Membrane

Figure 4. Schematic reconstruction of the relative position of the MrG carbonyls with respect to the carbonyls of PLPC. The color codes for the atomic densities and for the levels of carbonyls are according to the definitions in Figure 3.

reference altitude, to which the relative degrees of hydration of the carbonyls of the dipeptide have to be compared. In particular, if the native amide carbonyl is 2.6 ( 0.44 times better hydrated than that of phospholipid (see Figure 2), then, taking into account the measured water density, we have to find the distance from the membrane center where the density of water is 2.6 ( 0.44 times higher than that at the reference altitude of 17.2 Å. As a result, we find that the first momenta of the hydrated peptide carbonyl distributions are located at 17.7 (s.d.: +0.7/ -0.9) and 20.1 (s.d.: +0.6/-0.7) Å from the membrane center. These values are indicated by the red dashed horizontal lines in Figure 3B; the vertical bars represent their standard deviations. Thus, we may summarize our findings; the 13C-labeled and native amide carbonyls of the dipeptide, in order to be 1.2 ( 0.35 and 2.6 ( 0.44 times better hydrated then CdO of the phospholipid, must be located about 2 and 4.4 Å, respectively, above the CdO groups of the phospholipid. This suggests that while the 13C-labeled group is not too far from the glycerol backbone of the phospholipid, the native moiety of the dipeptide may be found in the vicinity of the phospholipid phosphate. The schematic reconstruction of the relative position is shown in Figure 4. Are these conclusion on the localization of the amide groups consistent with the linear spectral properties? The FTIR spectrum of phospholipid membranes demonstrates that the inhomogeneous broadening dominates in the resonances characteristic of the phospholipid carbonyls. There are two contributions to this broadening, the variance of the local frequencies of the CdO oscillators according to the electric field generated by the charged moieties at the interface and the presence of coupling between neighboring CdO groups. The latter effect gives rise to excitonic dispersion. The two effects were disentangled and specified recently using 2DIR spectroscopy.23 The same experiment pointed out the quasi-static (on the picosecond time scale) character of this inhomogeneity, as the molecular reorientation in membranes takes place on a much longer time scale. In contrast, the shapes of the amide I FTIR bands reported here (see the linear spectrum in Figure 1) are rather close to Lorentzians. This suggests two important implications; first, the dipeptide backbone should experience

J. Phys. Chem. B, Vol. 113, No. 50, 2009 16249 only one conformation, thus excluding any possible spectral inhomogeneity due to structural variances, as reported, for instance, for a tripeptide in an aqueous environment.28 Second, the fact that homogeneous broadening dominates may be interpreted as the indication that the two amide groups of the dipeptide are more exposed to water than the phospholipid CdO groups, thus allowing an effective motional line narrowing. In this sense, the relative hydration plays an important role. In fact, since the density of water decrease rapidly toward the hydrophobic core of the bilayer, motional line narrowing does not contribute much to the line shape of the of phospholipid carbonyls. As we have shown in previous papers,15,16,29 when the degree of hydration is drastically reduced, the water molecules deeply trapped in the membrane polar compartment are practically frozen and demonstrate quasi-static spectral inhomogeneity, analogous to that of the carbonyl groups. Motional line narrowing due to the water solvent effect is then rapidly lost upon moving toward the inner part of the bilayer, where the aqueous network is disrupted and the water dynamics slows down dramatically. This is the case of the PLPC carbonyls, located at a lower altitude inside of the bilayer with respect to those of the dipeptide. Another issue is how hydration affects the intensity ratio of the amide I modes. In general, the extinction coefficient of the CdO stretching slightly increases when involved in hydrogen bonding due to charge displacement and the resulting delocalization. The FITR spectrum (Figure 1) shows instead that the intensity of the less-hydrated carbonyl is larger. This suggests that, in this case, the backbone conformation governs the intensity ratio.30 Finally, we wish to discuss the average orientation of the dipeptide backbone with respect to the normal to the surface of the bilayer. The average difference of the altitudes of the two amide carbonyls with respect to the membrane center is about 2.4 Å. On the other hand, anticipating a rather flat, unfolded conformation of the dipeptide backbone,30 we can estimate for the distance between the two amide carbonyls a value of about 4.1 Å. From these two values, we calculate the angle between the dipeptide backbone and the normal to the membrane surface to be θ = 54°. This corresponds to a quite tilted orientation of the dipeptide backbone. This is consistent with the general structural alignment of the phospholipid environment at the interface of the membrane. In particular, X-ray in phospholipid single crystal,31 solid-state 31P NMR measurements,32,33 and molecular mechanics studies34-37 demonstrate that the phosphate-choline vector is tilted by a rather large angle (it varies between 63 and 90°, depending on the method adopted) with respect to the normal to the bilayer surface. At the same time, 13C-1H solid state NMR suggests that the glycerol-phosphate linker is about 35-36° off of the same normal.38 A deeper review of the molecular physics governing the orientation of the dipeptide backbone would require further theoretical studies at the level of both classical and quantum theory. In conclusion, we show that two-color anharmonic spectroscopy does provide an experimental opportunity for structural analysis. In the present case, we also make use of the results of neutron scattering experiments. It is of some interest to consider the different characters of the two experimental methods. Neutron scattering probes atomic densities in dependence of the spatial coordinates. In disordered systems, of course, only statistical distributions, averaged over the interaction region, can be extracted. Instead, two-color IR nonlinear response is informative of how the excitation of one vibrational mode affects another vibration. Thus, if the involved modes are of local

16250

J. Phys. Chem. B, Vol. 113, No. 50, 2009

nature, we gain the opportunity of probing the intermolecular arrangement on the scale corresponding to the actual localization of the vibrations. This specificity is attractive in comparison to the outcome of neutron scattering. In this work, we get information on the relative exposure of carbonyls to ambient water; the specificity of the response is due to the local character of the hydrogen bond. At the same time, the same specificity excludes the contribution of those carbonyls which are not involved in hydrogen bonding to water. The red dashed lines in Figure 3b in fact correspond to the first momenta of the structural distributions of the hydrated CdO groups, and the vertical bars indicate their standard deviations. It is clear that the information obtained from neutron scattering and from twocolor IR nonlinear spectroscopy is rather complementary. Furthermore, here, we demonstrate how the combination of experimental data obtained from linear steady-state IR optical spectroscopy, third-order ultrafast IR optical spectroscopy, and neutron scattering allows extraction of structural information that none of the techniques alone can provide. The experimental approach employed in this work represents a valuable opportunity to learn about the molecular properties in a liquid membrane environment and to review and constrain the structural findings that can be obtained from MD simulations for systems like the one considered here. Acknowledgment. The work has been supported by EU Contract RII3-CT-2003-506350 and by the Italian Ministry of University (MIUR). The authors express thanks to Dr. S. White for permission to use the atomic densities as reported in ref 27. References and Notes (1) Anderson, H. J.; Coleman, J. E.; Andersen, R. J.; Roberge, M. Cancer Chemother. Pharmacol. 1997, 39, 223–226. (2) Bai, R.; Durso, N. A.; Sackett, D. L.; Hamel, E. Biochem. 1999, 38, 14302–14310. (3) Loganzo, F.; Discafani, C. M.; Annable, T.; Beyer, C.; Musto, S.; Hari, M.; Tan, X.; Hardy, C.; Hernandez, R.; Baxter, M.; Singanallore, T.; Khafizova, G.; Poruchynsky, M. S.; Fojo, T.; Nieman, J. A.; AyralKaloustian, S.; Zask, A.; Andersen, R. J.; Greenberger, L. M. Cancer Res. 2003, 63, 1838–1845. (4) Piekarz, R. L.; Robey, Rob.; Sandor, V.; Bakke, S.; Wilson, W. H.; Dahmoush, L.; Kingma, D. M.; Turner, M. L.; Altemus, R.; Bates, S. E. Blood. 2001, 98, 2865–2868. (5) Vail, D. M.; MacEwen, E. G.; Kurzman, I. D.; Dubielzig, R. R.; Helfand, S. C.; Kisseberth, W. C.; London, C. A.; Obradovich, J. E.; Madewell, B. R.; Rodriguez, C. O.; Fidel, J.; Susaneck, S.; Rosenberg, M. Clin. Cancer Res. 1995, 1, 1165–1170. (6) Anderson, P. Future Oncology 2006, 2, 333–343. (7) Luckey, M. Membrane Structural Biology with Biochemical and Biophysical Foundations; Cambridge University Press: New York, 2008.

Volkov and Righini (8) Efremov, R. G.; Nolde, D. E.; Volynsky, P. E.; Chernyavsky, A. A.; Dubovskii, P. V.; Arseniev, A. S. FEBS Lett. 1999, 462, 205–210. (9) Gordon, L, M.; Mobley, P. W.; Pilpa, R.; Sherman, M. A.; Waring, A. J Biochim. Biophys. Acta 2002, 1559, 96–120. (10) Mukamel, S. Principles of Nonlinear Optical Spectroscopy; Oxford University Press: New York, 1995. (11) Asplund, M. C.; Zanni, M. T.; Hochstrasser, R. M. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 8219–8224. (12) Garrett-Roe, S; Hamm, P. J. Chem. Phys. 2008, 128, 1045071104507-13. (13) Brixner, T.; Stenger, J.; Vaswani, H. M.; Cho, M.; Blankenship, R. E.; Fleming, G. R. Nature 2005, 434, 625–628. (14) Mukherjee, P.; Krummel, A. T.; Fulmer, E. C.; Kass, I.; Arkin, I. T.; Zanni, M. T. J. Chem. Phys. 2004, 120, 10215–10224. (15) Volkov, V.; Palmer, D. J.; Righini, R. Phys. ReV. Lett. 2007, 99, 0783021–0783024. (16) Volkov, V.; Palmer, D. J.; Righini, R J. Phys. Chem. B 2007, 111, 1377–1383. (17) Volkov, V. V.; Nuti, F.; Takaoka, Y.; Chelli, R.; Papini, A. M.; Righini, R. J. Am. Chem. Soc. 2006, 128, 9466–9471. (18) Volkov, V.; Hamm, P. Biophys. J. 2004, 87, 4213–4225. (19) James, G.; Olson, E. N. Biochemistry 1990, 29, 2623–2634. (20) Farazi, T. A.; Waksman, G.; Gordon, J. I. J. Biol. Chem. 2001, 276, 39501–39504. (21) Furuishi, K.; Misumi, S.; Shoji, S. Biochem. Biophys. Res. Commun. 1996, 222, 344–351. (22) Vilas, G. L.; Corvi, M. M.; Plummer, G. J.; Seime, A. M.; Lambkin, G. R.; Berthiaume, L. G. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 6542– 6547. (23) Volkov, V.; Chelli, R.; Zhuang, W.; Nuti, F.; Takaoka, Y.; Papini, A. M.; Mukamel, S.; Righini, R. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 15323–15327. (24) Pasenkiewicz-Gierula, M.; Takaoka, Y.; Miyagawa, H.; Kitamura, K.; Kusumi, A. J. Phys. Chem. A 1997, 101, 3677–3691. (25) Mrazkova, E.; Hobza, P.; Bohl, M.; Gauger, D. R.; Pohle, W. J. Phys. Chem. B 2005, 109, 15126–15134. (26) Volkov, V.; Takaoka, Y.; Righini, R. Phys. Chem. Chem. Phys. 2009, DOI: 10.1039B914511G. (27) Jacobs, R. E.; White, S. H. Biochemistry 1989, 28, 3421–3437. (28) Volkov, V.; Takaoka, Y.; Righini, R. J. Phys. Chem. B 2009, 113, 4119–4124. (29) Woutersen, S.; Pfister, R.; Hamm, P.; Mu, Y.; Kosov, D. S.; Stock, G. J. Chem. Phys. 2002, 117, 6833–6840. (30) Volkov, V.; Righini, R., to be submitted. (31) Vanderkooi, G. Biochemistry 1991, 30, 10760–10768. (32) Seelig, J.; Gally, H. U. Biochemistry 1976, 15, 5199–5204. (33) Griffin, R. G.; Powers, L.; Pershan, P. S. Biochemistry 1978, 17, 2718–2722. (34) Ha Duong, T.; Mehler, E. L.; Weinstein, H. J. Comput. Phys. 1999, 151, 358–387. (35) de Vries, A. H.; Chandrasekhar, I.; van Gunsteren, W. F.; Hunenberger, P. H. J. Phys. Chem. B 2005, 109, 11643–11652. (36) Hogberg, C.-J.; Lyubartsev, A. P. J. Phys. Chem. B 2006, 110, 14326–14336. (37) Siu, S. W. I.; Vacha, R.; Jungwirth, P.; Bockmann, R. A. J. Chem. Phys. 2008, 128, 125103/1–125103/12. (38) Hong, M.; Schmidt-Rohr, K.; Zimmermann, H. Biochemistry 1996, 35, 8335–8341.

JP9082536