Deformation of Helix C in the Low Temperature L-intermediate of Bacteriorhodopsin
2004; Elsevier BV; Volume: 279; Issue: 3 Linguagem: Inglês
10.1074/jbc.m300709200
ISSN1083-351X
AutoresK. A. P. Edman, Antoine Royant, Gisela Larsson, Frida Jacobson, Tom Taylor, David van der Spoel, Ehud M. Landau, Eva Pebay‐Peyroula, Richard Neutze,
Tópico(s)Neuroscience and Neuropharmacology Research
ResumoX-ray and electron diffraction studies of specific reaction intermediates, or reaction intermediate analogues, have produced a consistent picture of the structural mechanism of light-driven proton pumping by bacteriorhodopsin. Of central importance within this picture is the structure of the L-intermediate, which follows the retinal all-trans to 13-cis photoisomerization step of the K-intermediate and sets the stage for the primary proton transfer event from the positively charged Schiff base to the negatively charged Asp-85. Here we report the structural changes in bacteriorhodopsin following red light illumination at 150 K. Single crystal microspectrophotometry showed that only the L-intermediate is populated in three-dimensional crystals under these conditions. The experimental difference Fourier electron density map and refined crystallographic structure were consistent with those previously presented (Royant, A., Edman, K., Ursby, T., Pebay-Peyroula, E., Landau, E. M., and Neutze, R. (2000) Nature 406, 645-648; Royant, A., Edman, K., Ursby, T., Pebay-Peyroula, E., Landau, E. M., and Neutze, R. (2001) Photochem. Photobiol. 74, 794-804). Based on the refined crystallographic structures, molecular dynamic simulations were used to examine the influence of the conformational change of the protein that is associated with the K-to-L transition on retinal dynamics. Implications regarding the structural mechanism for proton pumping by bacteriorhodopsin are discussed. X-ray and electron diffraction studies of specific reaction intermediates, or reaction intermediate analogues, have produced a consistent picture of the structural mechanism of light-driven proton pumping by bacteriorhodopsin. Of central importance within this picture is the structure of the L-intermediate, which follows the retinal all-trans to 13-cis photoisomerization step of the K-intermediate and sets the stage for the primary proton transfer event from the positively charged Schiff base to the negatively charged Asp-85. Here we report the structural changes in bacteriorhodopsin following red light illumination at 150 K. Single crystal microspectrophotometry showed that only the L-intermediate is populated in three-dimensional crystals under these conditions. The experimental difference Fourier electron density map and refined crystallographic structure were consistent with those previously presented (Royant, A., Edman, K., Ursby, T., Pebay-Peyroula, E., Landau, E. M., and Neutze, R. (2000) Nature 406, 645-648; Royant, A., Edman, K., Ursby, T., Pebay-Peyroula, E., Landau, E. M., and Neutze, R. (2001) Photochem. Photobiol. 74, 794-804). Based on the refined crystallographic structures, molecular dynamic simulations were used to examine the influence of the conformational change of the protein that is associated with the K-to-L transition on retinal dynamics. Implications regarding the structural mechanism for proton pumping by bacteriorhodopsin are discussed. Light-driven vectorial proton translocation is basic to the mechanism of energy transduction by photosynthetic systems. Bacteriorhodopsin (bR) 1The abbreviations used are: bR, bacteriorhodopsin; LT, low temperature; EC, extracellular; CCD, charge-coupled device; Wat, water; σ, root mean square electron density of the map. is the simplest known light-driven proton pump and has long served as a model system for understanding how protons may be transported "up hill" against a transmembrane proton motive potential. bR contains seven transmembrane α-helices that surround a proton translocation channel lined with strategically placed charged residues (3.Henderson R. Baldwin J.M. Ceska T.A. Zemlin F. Beckmann E. Downing K.H. J. Mol. Biol. 1990; 213: 899-929Crossref PubMed Scopus (2439) Google Scholar). Depending upon their protonation states, which change in a well orchestrated cascade as a proton is transported across the cell membrane, these charged residues can serve as either proton donors or proton acceptors. Light activation of the chromophore, an all-trans-retinal molecule covalently attached to Lys-216 in helix G via a protonated Schiff base (the primary proton donor) results in the 13-cis-retinal configuration with two-thirds quantum efficiency. Steric conflicts and mechanical stress resulting from photoisomerization initiate a sequence of conformational changes that can be characterized spectroscopically and that perturb the local environment of several key residues, strongly affecting their pKa values and creating transient pathways for proton transfer. The specific spectral intermediates of the bR photocycle have been well characterized, and a common reaction scheme is: bR570 → K590 ↔ L550 ↔ M412 ↔ N560 ↔ O640 → bR570 (sub-scripts denote the wavelengths of the respective absorption maxima). M412 is usually associated with two distinct intermediates (4.Varo G. Lanyi J.K. Biochemistry. 1991; 30: 5008-5015Crossref PubMed Google Scholar), commonly referred to as the early and late M-states, whose spectrally silent transition correlates with a conformational change on the cytoplasmic half of the protein. In general, at low temperatures the thermal motions that aid the crossing of energy barriers associated with specific transitions can be reduced or even frozen out, and light activation of three-dimensional crystals can populate specific photocycle intermediates (5.Schlichting I. Berendzen J. Phillips Jr., G.N. Sweet R.M. Nature. 1994; 371: 808-812Crossref PubMed Scopus (335) Google Scholar, 6.Genick U.K. Borgstahl G.E. Ng K. Ren Z. Pradervand C. Burke P.M. Srajer V. Teng T.Y. Schildkamp W. McRee D.E. Moffat K. Getzoff E.D. Science. 1997; 275: 1471-1475Crossref PubMed Scopus (365) Google Scholar, 7.Genick U.K. Soltis S.M. Kuhn P. Canestrelli I.L. Getzoff E.D. Nature. 1998; 392: 206-209Crossref PubMed Scopus (321) Google Scholar, 8.Chu K. Vojtchovsky J. McMahon B.H. Sweet R.M. Berendzen J. Schlichting I. Nature. 2000; 403: 921-923Crossref PubMed Scopus (226) Google Scholar, 9.Edman K. Royant A. Nollert P. Maxwell C.A. Pebay-Peyroula E. Navarro J. Neutze R. Landau E.M. Structure (Camb.). 2002; 10: 473-482Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). Using this approach, a number of x-ray and electron diffraction studies on light-driven structural changes in bR (1.Royant A. Edman K. Ursby T. Pebay-Peyroula E. Landau E.M. Neutze R. Nature. 2000; 406: 645-648Crossref PubMed Scopus (219) Google Scholar, 10.Edman K. Nollert P. Royant A. Belrhali H. Pebay-Peyroula E. Hajdu J. Neutze R. Landau E.M. Nature. 1999; 401: 822-826Crossref PubMed Scopus (293) Google Scholar, 11.Sass H.J. Buldt G. Gessenich R. Hehn D. Neff D. Schlesinger R. Berendzen J. Ormos P. Nature. 2000; 406: 649-653Crossref PubMed Scopus (301) Google Scholar, 12.Facciotti M.T. Rouhani S. Burkard F.T. Betancourt F.M. Downing K.H. Rose R.B. McDermott G. Glaeser R.M. Biophys. J. 2001; 81: 3442-3455Abstract Full Text Full Text PDF PubMed Google Scholar, 13.Schobert B. Cupp-Vickery J. Hornak V. Smith S. Lanyi J. J. Mol. Biol. 2002; 321: 715-726Crossref PubMed Scopus (179) Google Scholar, 14.Lanyi J. Schobert B. J. Mol. Biol. 2002; 321: 727-737Crossref PubMed Scopus (114) Google Scholar, 15.Lanyi J.K. Schobert B. J. Mol. Biol. 2003; 328: 439-450Crossref PubMed Scopus (137) Google Scholar, 16.Matsui Y. Sakai K. Murakami M. Shiro Y. Adachi S. Okumura H. Kouyama T. J. Mol. Biol. 2002; 324: 469-481Crossref PubMed Scopus (159) Google Scholar) and in bR mutants (17.Luecke H. Schobert B. Richter H.T. Cartailler J.P. Lanyi J.K. Science. 1999; 286: 255-261Crossref PubMed Scopus (472) Google Scholar, 18.Luecke H. Schobert B. Cartailler J.P. Richter H.T. Rosengarth A. Needleman R. Lanyi J.K. J. Mol. Biol. 2000; 300: 1237-1255Crossref PubMed Scopus (180) Google Scholar, 19.Vonck J. EMBO J. 2000; 19: 2152-2160Crossref PubMed Google Scholar), as well as bR mutant intermediate state analogues (20.Subramaniam S. Henderson R. Nature. 2000; 406: 653-657Crossref PubMed Scopus (367) Google Scholar, 21.Rouhani S. Cartailler J.P. Facciotti M.T. Walian P. Needleman R. Lanyi J.K. Glaeser R.M. Luecke H. J. Mol. Biol. 2001; 313: 615-628Crossref PubMed Scopus (87) Google Scholar), have been presented recently. When taken together with an acceptance of the limitations of the techniques of kinetic crystallography, most structures combine to yield a remarkably consistent picture of the structural mechanism of outwardly directed proton pumping by bR (22.Neutze R. Pebay-Peyroula E. Edman K. Royant A. Navarro J. Landau E.M. Biochim. Biophys. Acta. 2002; 1565: 144-167Crossref PubMed Scopus (153) Google Scholar). As with the majority of studies of the low temperature L-intermediate (LLT) of bR (15.Lanyi J.K. Schobert B. J. Mol. Biol. 2003; 328: 439-450Crossref PubMed Scopus (137) Google Scholar, 23.Becher B. Tokunaga F. Ebrey T.G. Biochemistry. 1978; 17: 2293-2300Crossref PubMed Google Scholar, 24.Braiman M.S. Mogi T. Marti T. Stern L.J. Khorana H.G. Rothschild K.J. Biochemistry. 1988; 27: 8516-8520Crossref PubMed Google Scholar, 25.Maeda A. Sasaki J. Yamazaki Y. Needleman R. Lanyi J.K. Biochemistry. 1994; 33: 1713-1717Crossref PubMed Google Scholar, 26.Maeda A. Kandori H. Yamazaki Y. Nishimura S. Hatanaka M. Chon Y.S. Sasaki J. Needleman R. Lanyi J.K. J. Biochem. (Tokyo). 1997; 121: 399-406Crossref PubMed Google Scholar, 27.Hu J.G. Sun B.Q. Petkova A.T. Griffin R.G. Herzfeld J. Biochemistry. 1997; 36: 9316-9322Crossref PubMed Scopus (60) Google Scholar, 28.Hendrickson F.M. Burkard F. Glaeser R.M. Biophys. J. 1998; 75: 1446-1454Abstract Full Text Full Text PDF PubMed Google Scholar, 29.Kandori H. Yamazaki Y. Shichida Y. Raap J. Lugtenburg J. Belenky M. Herzfeld J. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 1571-1576Crossref PubMed Scopus (72) Google Scholar), our previous work on the structural changes associated with a build up of L was performed at 170 K. In that study three-dimensional crystals of bR were illuminated for 30 s with green light, thereby populating primarily K170K, followed by a further 40-s relaxation in the dark (1.Royant A. Edman K. Ursby T. Pebay-Peyroula E. Landau E.M. Neutze R. Nature. 2000; 406: 645-648Crossref PubMed Scopus (219) Google Scholar) in which L170K evolved. Subsequently the crystals were quenched in liquid nitrogen. As stated in that study, this trapping protocol produced predominantly the L-intermediate, although the existence of other species was detected and acknowledged (1.Royant A. Edman K. Ursby T. Pebay-Peyroula E. Landau E.M. Neutze R. Nature. 2000; 406: 645-648Crossref PubMed Scopus (219) Google Scholar) and was later quantified as a mixture of K170K, L170K, and M170K in the ratio 1:3:1 (2.Royant A. Edman K. Ursby T. Pebay-Peyroula E. Landau E.M. Neutze R. Photochem. Photobiol. 2001; 74: 794-804Crossref PubMed Scopus (27) Google Scholar). Due to the mixture of spectral species the interpretation of our earlier structural result has been controversial (30.Balashov S.P. Ebrey T.G. Photochem. Photobiol. 2001; 73: 453-462Crossref PubMed Google Scholar, 31.Lanyi J.K. J. Phys. Chem. 2000; B104: 11441-11448Crossref Scopus (109) Google Scholar, 32.Lanyi J.K. Luecke H. Curr. Opin. Struct. Biol. 2001; 11: 415-419Crossref PubMed Scopus (133) Google Scholar). As such it is valuable to repeat the intermediate trapping experiment under conditions that produce pure L-intermediate in three-dimensional bR crystals. In this work we present spectral and structural results following red light illumination of three dimensional bR crystals cooled to 150 K. The spectral analysis, difference Fourier map, and refined crystallographic structure demonstrate that the light-induced structural changes previously reported (1.Royant A. Edman K. Ursby T. Pebay-Peyroula E. Landau E.M. Neutze R. Nature. 2000; 406: 645-648Crossref PubMed Scopus (219) Google Scholar) were correctly interpreted as characterizing the build up of LLT. It follows that the mechanistic model of vectorial proton transport by bR should incorporate the fact that significant rearrangements of water molecules on the EC half of the protein, a reorientation of the guanidinium group of Arg-82 toward the EC medium, and a local flex of helix C toward the proton translocation channel centered near Asp-85 all occur after retinal photoisomerization yet prior to the primary proton transfer event from the Schiff base to Asp-85. Building upon the refined structural models, molecular dynamic simulations are presented that illustrate how the geometry of the retinal is perturbed by the conformational change in the protein associated with the K-to-L transition, providing structural insight into how the L-intermediate sets the stage for the primary proton transfer from the Schiff base to Asp-85 in the L-to-M transition. Protein Crystallization—bR was crystallized using a lipidic cubic phase matrix as described previously (33.Landau E.M. Rosenbusch J.P. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 14532-14535Crossref PubMed Scopus (798) Google Scholar). These crystals grow as hexagonal plates typically 70 μm across and less than 10 μm thick. Crystals were harvested from the viscous lipidic environment by a lipase treatment (34.Nollert P. Landau E.M. Biochem. Soc. Trans. 1998; 26: 709-713Crossref PubMed Scopus (34) Google Scholar). For both microspectrophotometry and x-ray diffraction studies, crystals were light adapted (35.Balashov S.P. Imasheva E.S. Govindjee R. Ebrey T.G. Biophys. J. 1996; 70: 473-481Abstract Full Text PDF PubMed Google Scholar) at room temperature by several minutes of exposure to bright white light immediately prior to being mounted on cryoloops and flash frozen in liquid nitrogen. Single Crystal Microspectrophotometry—Single crystal microspectrophotometry (36.Hadfield A. Hajdu J. J. Appl. Crystallogr. 1993; 26: 839-842Crossref Scopus (82) Google Scholar, 37.Bourgeois D. Vernede X. Adam V. Fioravanti E. Ursby T. J. Appl. Crystallogr. 2002; 35: 319-326Crossref Scopus (68) Google Scholar) was used to characterize both the ground state (Fig. 1a) and spectral intermediate (Fig. 1b) trapped within three-dimensional crystals of bR. An Oxford Cryosystems nitrogen gas stream held the crystal at a constant temperature of 150 K. The reaction was initiated using a 20-milliwatt, 635 nm laser diode coupled to a 225-μm-diameter optical fiber, which guided light to within ∼300 μm of the crystal position incident at an angle ∼30° relative to that of the crystal plate. The optical fiber was decoupled relative to the red laser, and the light flux was measured using a Newport 818-UV photometer with the same red light intensity being used for both x-ray diffraction and microspectrophotometry experiments (estimated to be of the order of 5 watts/cm2 at the crystal position). The ground state spectrum (Fig. 1a) and the difference spectrum between that recorded during red light illumination and the ground state spectrum (Fig. 1b, discontinuous line) were recorded using a continuous wave halogen lamp with a probe spot diameter of ∼15 μm. The difference spectrum between the spectrum recorded following 30-s illumination and 40-s delay in the dark and the ground state spectrum (Fig. 1b, continuous line) was recorded using a 10-μs Xenon flash lamp. An Ocean Optics SD-2000 CCD spectrophotometer was used to record spectra enabling a broader wavelength domain to be sampled with significantly higher sensitivity than in previous work (2.Royant A. Edman K. Ursby T. Pebay-Peyroula E. Landau E.M. Neutze R. Photochem. Photobiol. 2001; 74: 794-804Crossref PubMed Scopus (27) Google Scholar). X-ray Data Collection—Light-adapted (35.Balashov S.P. Imasheva E.S. Govindjee R. Ebrey T.G. Biophys. J. 1996; 70: 473-481Abstract Full Text PDF PubMed Google Scholar) frozen bR crystals were illuminated at 150 K with red light (λ = 635 nm) for 30 s followed by a further 40-s delay in the dark and were then quenched in liquid nitrogen. Approximately 20 min following off-line trapping, crystals were mounted on the x-ray goniometer of beamline ID14-EH2 of the European Synchrotron Radiation Facility. A first (coarse) alignment of the crystal on the goniometer head was made using red light, and the final adjustment (a few seconds) was made using dim white light. An x-ray wavelength of 0.93 Å was used for data collection with an ADSC Quantum4 CCD detector. 120 frames of 1°/frame were collected with 30-s x-ray exposure/frame. Data were processed using the HKL package (38.Otwinowski Z. Sawyer L. Isaacs N.W. Bailey S. Data Collection and Processing. DL/SCI/R34. Daresbury Laboratory, Warrington, UK1993: 55-62Google Scholar) and the CCP4 suite (39.Bailey S. Acta Crystallogr. Sect. D Biol. Crystallogr. 1994; 50: 760-763Crossref PubMed Scopus (37) Google Scholar). Crystallographic data and refinement statistics are summarized in Table I.Table IDiffraction data and refinement statisticsResolution (Å)47—2.3Space groupP63No. of observations92,219No. of unique reflections9,848Completeness (%) (outer shell)aOuter shells were 2.42—2.30 Å96.9 (95.1)Rsym (%)bRsym = ΣjΣh|Ihj — 〈Ih〉|/ΣjΣhIhj (outer shell)aOuter shells were 2.42—2.30 Å7.2 (51.0)I/σ(I) (outer shell)aOuter shells were 2.42—2.30 Å8.1 (1.5)Twinning (%)27 (±2)Unit cell (Å)a = b = 60.9, c = 109.5RefinementAlternate conformationOccupancy of L150K (%)50Refined residues228Rcryst (%)cRcryst = Σh∥Fobs(h)| — |Fcalc(h)∥/Σh |Fobs(h)|25.2Rfree (%)dCalculated from a set of 5% randomly selected reflections that were excluded from refinement29.8Root mean square deviationsBond lengths (Å)0.0082Bond angles (o)1.26a Outer shells were 2.42—2.30 Åb Rsym = ΣjΣh|Ihj — 〈Ih〉|/ΣjΣhIhjc Rcryst = Σh∥Fobs(h)| — |Fcalc(h)∥/Σh |Fobs(h)|d Calculated from a set of 5% randomly selected reflections that were excluded from refinement Open table in a new tab Merohedral Twinning—The degree of merohedral twinning was determined using Britton plots and Yeates statistics (40.Yeates T.O. Methods Enzymol. 1997; 276: 344-358Crossref PubMed Scopus (352) Google Scholar) as described in Ref. 41.Royant A. Grizot S. Kahn R. Belrhali H. Fieschi F. Landau E.M. Pebay-Peyroula E. Acta Crystallogr. Sect. D Biol. Crystallogr. 2002; 58: 784-791Crossref PubMed Scopus (9) Google Scholar and yielded a value of 27% twinning. Crystallographic data were detwinned according to the formula (41.Royant A. Grizot S. Kahn R. Belrhali H. Fieschi F. Landau E.M. Pebay-Peyroula E. Acta Crystallogr. Sect. D Biol. Crystallogr. 2002; 58: 784-791Crossref PubMed Scopus (9) Google Scholar), Idetwin(h1)=[(1-x)Iobs(h1)-xIobs(h2)]/(1-2x)(Eq. 1) Idetwin(h2)=-[xIobs(h1)-(1-x)Iobs(h2)]/(1-2x)(Eq. 2) prior to further crystallographic analysis where Idetwin is the intensity corresponding to diffraction from an untwinned crystal, Iobs is the (twinned) measured intensity, the two reflections h1 and h2 are related by the twin operation but not by crystallographic symmetry, and x is the degree of merohedral twinning. When following this procedure the detwinned data is independent of any structural model, and the number of independent observations is not reduced, but the errors associated with detwinning increase with increasing twinning, approaching infinity as x approaches 0.5 (41.Royant A. Grizot S. Kahn R. Belrhali H. Fieschi F. Landau E.M. Pebay-Peyroula E. Acta Crystallogr. Sect. D Biol. Crystallogr. 2002; 58: 784-791Crossref PubMed Scopus (9) Google Scholar). Consequently when following this protocol the refined crystallographic R-factors are typically somewhat larger than might be expected. For perfectly twinned (or almost perfectly twinned) crystallographic data (from which several intermediate structures of bR have been reported (12.Facciotti M.T. Rouhani S. Burkard F.T. Betancourt F.M. Downing K.H. Rose R.B. McDermott G. Glaeser R.M. Biophys. J. 2001; 81: 3442-3455Abstract Full Text Full Text PDF PubMed Google Scholar, 13.Schobert B. Cupp-Vickery J. Hornak V. Smith S. Lanyi J. J. Mol. Biol. 2002; 321: 715-726Crossref PubMed Scopus (179) Google Scholar, 14.Lanyi J. Schobert B. J. Mol. Biol. 2002; 321: 727-737Crossref PubMed Scopus (114) Google Scholar, 15.Lanyi J.K. Schobert B. J. Mol. Biol. 2003; 328: 439-450Crossref PubMed Scopus (137) Google Scholar, 17.Luecke H. Schobert B. Richter H.T. Cartailler J.P. Lanyi J.K. Science. 1999; 286: 255-261Crossref PubMed Scopus (472) Google Scholar, 18.Luecke H. Schobert B. Cartailler J.P. Richter H.T. Rosengarth A. Needleman R. Lanyi J.K. J. Mol. Biol. 2000; 300: 1237-1255Crossref PubMed Scopus (180) Google Scholar)), x = 0.5, and it is not possible to retrieve the diffraction data corresponding to an untwinned crystal from the observations alone. Rather when calculating electron density maps the "detwinned" observations are recovered according to the formula (42.Redinbo M.R. Yeates T.O. Acta Crystallogr. Sect. D Biol. Crystallogr. 1993; 49: 375-380Crossref PubMed Google Scholar), Idetwin(h1)=[Iobs(h1)+Icalc(h1)-Icalc(h2)]/2(Eq. 3) Idetwin(h2)=[Iobs(h1)+Icalc(h2)-Icalc(h1)]/2(Eq. 4) or variations thereof, which depend explicitly upon intensities calculated from the structural model, Icalc(h) (the convention is that Idetwin and Icalc correspond to diffraction from only half of the crystal). In this case each observation Iobs(h1) and its associated twin Iobs(h2) cannot be measured independently (i.e. Iobs(h1) = Iobs(h2)), and the number of independent observations at any given resolution is reduced almost exactly by a half relative to data with low levels of twinning. Consequently refined crystallographic R-factors from perfectly twinned data are typically a factor of 1/√2 lower than for low (or un-)twinned data (40.Yeates T.O. Methods Enzymol. 1997; 276: 344-358Crossref PubMed Scopus (352) Google Scholar, 41.Royant A. Grizot S. Kahn R. Belrhali H. Fieschi F. Landau E.M. Pebay-Peyroula E. Acta Crystallogr. Sect. D Biol. Crystallogr. 2002; 58: 784-791Crossref PubMed Scopus (9) Google Scholar, 42.Redinbo M.R. Yeates T.O. Acta Crystallogr. Sect. D Biol. Crystallogr. 1993; 49: 375-380Crossref PubMed Google Scholar), and the explicit dependence of the detwinned data on the model itself (Equations 3 and 4) introduces additional model bias in electron density maps. Our work on the ground state (43.Belrhali H. Nollert P. Royant A. Menzel C. Rosenbusch J.P. Landau E.M. Pebay-Peyroula E. Struct. Fold. Des. 1999; 7: 909-917Abstract Full Text Full Text PDF Scopus (409) Google Scholar) and KLT-state (10.Edman K. Nollert P. Royant A. Belrhali H. Pebay-Peyroula E. Hajdu J. Neutze R. Landau E.M. Nature. 1999; 401: 822-826Crossref PubMed Scopus (293) Google Scholar) of bR was free from merohedral twinning, and our previous work on LLT (x = 0.26 (1.Royant A. Edman K. Ursby T. Pebay-Peyroula E. Landau E.M. Neutze R. Nature. 2000; 406: 645-648Crossref PubMed Scopus (219) Google Scholar)) as well as that presented here (x = 0.27) had sufficiently low levels of twinning such that data were detwinned according to Equations 1 and 2 prior to further structural analysis. Difference Fourier Analysis—Detwinned crystallographic data were scaled together with the ground state observations (Protein Data Bank entry 1QHJ) using the CCP4 suite (39.Bailey S. Acta Crystallogr. Sect. D Biol. Crystallogr. 1994; 50: 760-763Crossref PubMed Scopus (37) Google Scholar). The difference Fourier electron density map was calculated by Fourier transform of (Fexc-Fgnd)·exp(i·Φgnd) using the refined ground state model (Protein Data Bank entry 1QHJ) for phases (43.Belrhali H. Nollert P. Royant A. Menzel C. Rosenbusch J.P. Landau E.M. Pebay-Peyroula E. Struct. Fold. Des. 1999; 7: 909-917Abstract Full Text Full Text PDF Scopus (409) Google Scholar). The structural changes described here were interpreted directly from the difference electron density map and were confirmed by partial occupancy refinement. Structural Refinement—Partial occupancy refinement was performed, using CNS (44.Brunger A.T. Adams P.D. Clore G.M. DeLano W.L. Gros P. Grosse-Kunstleve R.W. Jiang J.S. Kuszewski J. Nilges M. Pannu N.S. Read R.J. Rice L.M. Simonson T. Warren G.L. Acta Crystallogr. Sect. D Biol. Crystallogr. 1998; 54: 905-921Crossref PubMed Scopus (16709) Google Scholar), by refining one excited state conformation (labeled LLT or L150K, which was allowed to change conformation), while the ground state model (with complementary occupancy) was held fixed. After the models were aligned as rigid bodies, a simulated annealing refinement of LLT was performed to escape bias toward the initial model. The crystallographic occupancy was determined by first selecting 10 evenly spaced (from 0 to 100%) values for the crystallographic occupancy, refining the structure of LLT with this occupancy held fixed, and then refining the occupancy with the model for LLT held fixed. The crystallographic occupancy converged from above and from below to a value of 50% (see Ref. 2.Royant A. Edman K. Ursby T. Pebay-Peyroula E. Landau E.M. Neutze R. Photochem. Photobiol. 2001; 74: 794-804Crossref PubMed Scopus (27) Google Scholar for details of the method). As with previous work (1.Royant A. Edman K. Ursby T. Pebay-Peyroula E. Landau E.M. Neutze R. Nature. 2000; 406: 645-648Crossref PubMed Scopus (219) Google Scholar, 9.Edman K. Royant A. Nollert P. Maxwell C.A. Pebay-Peyroula E. Navarro J. Neutze R. Landau E.M. Structure (Camb.). 2002; 10: 473-482Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar, 10.Edman K. Nollert P. Royant A. Belrhali H. Pebay-Peyroula E. Hajdu J. Neutze R. Landau E.M. Nature. 1999; 401: 822-826Crossref PubMed Scopus (293) Google Scholar), planar constraints were placed on the geometry of the retinal during refinement of LLT since no paired electron density peaks indicating the nature and extent of distortion of the retinal geometry were visible at a high confidence level in the difference Fourier map. Molecular Dynamic Simulations—Molecular dynamic simulations were used to investigate the influence of the K-to-L structural transition on retinal dynamics. Crystallographically resolved water molecules and additional water molecules predicted using Dowser (104.Zhang L. Hermans J. Proteins Struct. Funct. Genet. 1996; 24: 433-438Crossref PubMed Scopus (301) Google Scholar) were included in these simulations. Dowser assigns non-crystallographic water molecules by constructing a molecular surface from the input structure file and finding positions where a solvent probe (with a default radius of 0.2 Å) touches three atoms simultaneously. Energies were computed for buried waters, and those with a stabilization energy of at least -2 kcal/mol were included in the model. Lipid molecules were removed from the model, which was instead inserted into a slab of low temperature argon atoms (45.Aqvist J. Luzhkov V. Nature. 2000; 404: 881-884Crossref PubMed Scopus (369) Google Scholar) with enhanced Lennard Jones repulsions between argon and water (σ = 0.36 nm and ϵ = 0.85 kJ/mol) so as to mimic the hydrophobic membrane environment. Simulations of the intact membrane bilayer have been reported (46.Baudry J. Tajkhorshid E. Molnar R. Phillips J. Schulten K. J. Phys. Chem. B. 2001; 105: 905-918Crossref Google Scholar) but were not considered essential for this work. Time-dependent protein dynamic simulations of 500-ps duration were performed using the chemical simulation software package GROMACS (47.Lindahl E. Hess B. van der Spoel D. J. Mol. Model. 2001; 7: 306-317Crossref Google Scholar) with a time step of 1 fs. The Gromos-96 force field was used throughout (48.van Gunsteren W.F. Billeter S.R. Eising A.A. Hunenberger P.H. Kruger P. Mark A.E. Scott W.R.P. Tironi I.G. Biomolecular Simulation: the GROMOS96 Manual and User Guide. Hochschulverlag AG an der ETH Zurich, Zurich1996Google Scholar) and was modified to include parameters for the retinal and Schiff base where partial charges for simulations (49.Tajkhorshid E. Suhai S. J. Phys. Chem. B. 1999; 103: 5581-5590Crossref Google Scholar) were kindly provided by E. Tajkhorshid, 2E. Tajkhorshid, personal communication. and torsional displacements from planarity were subject to a harmonic energy penalty of ∼320 kJ·mol-1·radian-2. All protein Cα atoms were restrained to their crystal positions by a force constant of 100 kJ·mol·nm-1. Equilibration of this system for 20 ps resulted in models with well relaxed argon and water molecules, and the potential energy in any given simulation converged within 80 ps. The root mean square deviation of the non-hydrogen atoms of the average trajectory structures (from 100 to 300 ps) deviated from the crystallographic structures by ∼1.0 Å. Spectral Analysis—Fig. 1a shows the absorption spectrum of light-adapted bR (35.Balashov S.P. Imasheva E.S. Govindjee R. Ebrey T.G. Biophys. J. 1996; 70: 473-481Abstract Full Text PDF PubMed Google Scholar) in the ground state recorded from a three-dimensional bR crystal cooled to 150 K using single crystal microspectrophotometry (36.Hadfield A. Hajdu J. J. Appl. Crystallogr. 1993; 26: 839-842Crossref Scopus (82) Google Scholar, 37.Bourgeois D. Vernede X. Adam V. Fioravanti E. Ursby T. J. Appl. Crystallogr. 2002; 35: 319-326Crossref Scopus (68) Google Scholar). In Fig. 1b, two difference spectra (illuminated state spectrum minus the ground state spectrum) recorded at 150 K are presented: one recorded with simultaneous red light illumination contains a discontinuity at 635 nm due to scatter of the red laser light from the crystal and in agreement with difference spectra presented in Ref. 15.Lanyi J.K. Schobert B. J. Mol. Biol. 2003; 328: 439-450Crossref PubMed Scopus (137) Google Scholar and one recorded after 30-s red light illumination followed by a further 40-s delay in the dark and in agreement with difference spectra presented in Ref. 16.Matsui Y. Sakai K. Murakami M. Shiro Y. Adachi S. Okumura H. Kouyama T. J. Mol. Biol. 2002; 324: 469-481Crossref PubMed Scopus (159) Google Scholar (data are reproduced from Fig. 5c of Ref.
Referência(s)