Active State of Sensory Rhodopsin II: Structural Determinants for Signal Transfer and Proton Pumping

doi:10.1016/j.jmb.2011.07.022

J. Mol. Biol. (2011) 412, 591–600 Contents lists available at www.sciencedirect.com

Journal of Molecular Biology j o u r n a l h o m e p a g e : h t t p : / / e e s . e l s e v i e r. c o m . j m b

Active State of Sensory Rhodopsin II: Structural Determinants for Signal Transfer and Proton Pumping Ivan Gushchin 1, 2 , Anastasia Reshetnyak 1, 2 , Valentin Borshchevskiy 1, 2, 3 , Andrii Ishchenko 3, 4 , Ekaterina Round 1, 2 , Sergei Grudinin 5, 6 , Martin Engelhard 7 , Georg Büldt 2, 3 and Valentin Gordeliy 1, 2, 3 ⁎ 1

Laboratoire des Protéines Membranaires, Institut de Biologie Structurale J.-P. Ebel, UMR5075 CEA-CNRS-UJF, 38027 Grenoble, France 2 Research–Educational Center “Bionanophysics,” Moscow Institute of Physics and Technology, 141700 Dolgoprudniy, Russia 3 Institute of Complex Systems, ICS-5: Molecular Biophysics, Research Center Juelich, 52425 Juelich, Germany 4 Institute of Crystallography, University of Aachen (RWTH), Jaegerstrasse 17-19, 52056 Aachen, Germany 5 NANO-D, INRIA Grenoble-Rhone-Alpes Research Center, 38334 Saint Ismier Cedex, Montbonnot, France 6 CNRS, Laboratoire Jean Kuntzmann, BP 53, Grenoble Cedex 9, France 7 Max Planck Institute of Molecular Physiology, 44227 Dortmund, Germany Received 9 March 2011; received in revised form 23 June 2011; accepted 13 July 2011 Available online 4 August 2011 Edited by R. Huber Keywords: bacterial rhodopsins; membrane protein; active state; signaling

The molecular mechanism of transmembrane signal transduction is still a pertinent question in cellular biology. Generally, a receptor can transfer an external signal via its cytoplasmic surface, as found for G-protein-coupled receptors such as rhodopsin, or via the membrane domain, such as that in sensory rhodopsin II (SRII) in complex with its transducer, HtrII. In the absence of HtrII, SRII functions as a proton pump. Here, we report on the crystal structure of the active state of uncomplexed SRII from Natronomonas pharaonis, NpSRII. The problem with a dramatic loss of diffraction quality upon loading of the active state was overcome by growing better crystals and by reducing the occupancy of the state. The conformational changes in the region comprising helices F and G are similar to those observed for the NpSRII–transducer complex but are much more pronounced. The meaning of these differences for the understanding of proton pumping and signal transduction by NpSRII is discussed. © 2011 Elsevier Ltd. All rights reserved.

Introduction Microbial rhodopsins, a family of at least 1000 members, 1 have been the focus of recent research

*Corresponding author. Laboratoire des Protéines Membranaires, Institut de Biologie Structurale J.-P. Ebel, UMR5075 CEA-CNRS-UJF, 38027 Grenoble, France. E-mail address: [email protected]. Abbreviations used: SRII, sensory rhodopsin II; BR, bacteriorhodopsin.

because of their functional diversity (encompassing ion pumps, channels, and photosensors) and their potential applications in energy conversion, computer memory and chip development, or optogenetics. 2,3 The common features of all pigments are a seventransmembrane helix scaffold, a retinal bound to the C-terminal helix G via a protonated Schiff base to a lysine residue, and a light-activated photocycle. Three archetypical microbial rhodopsins can be distinguished. 2 Bacteriorhodopsin (BR), a proton pump, has been investigated exhaustively in great detail and is therefore a reference point for all other microbial rhodopsins. Channelrhodopsin, only

0022-2836/$ - see front matter © 2011 Elsevier Ltd. All rights reserved.

592 recently discovered in algae, functions as a lightactivated ion channel. The third system, sensory rhodopsin II (SRII), together with sensory rhodopsin I, enables archaeabacteria to seek optimal light conditions for their energy needs while avoiding photooxidative damage. Both photoreceptors trigger a signal transduction chain that is closely related to that of the eubacterial two-component system. One essential but not yet completely understood problem concerns the molecular mechanism of signal transfer from receptors to cytoplasmic signaling components. To advance our understanding of signal transduction, we determined the crystal structure of the active state of uncomplexed SRII from Natronomonas pharaonis (NpSRII). NpSRII was chosen for these experiments because it has been characterized quite in detail (for reviews, see Klare et al. 2 and Sasaki and Spudich 4). In the cellular membrane, NpSRII binds to its cognate transducer, NpHtrII, in a 2:2 complex 5,6 to form the signaling module. The chain of events is triggered by light, thereby initiating a trans–cis isomerization of the retinylidene chromophore of NpSRII. As a consequence, the protein passes through several intermediates until it reaches back the original state in about 1 s. The functionally important conformational step occurs during M1 → M2 transition, which leads to the signaling state. 7,8 During this transition, the signal is transferred to NpHtrII through the surface of helices F and G, resulting in a rotary motion of transmembrane helix TM2. 6,9 This activation occurs at the level of the membrane. How this signal is transmitted to the signal domain, thereby modulating the activity of the histidine kinase CheA and triggering the cytoplasmic two-component signaling cascade, is not known yet. 2,5,9,10 Published X-ray and NMR structures of NpSRII 11–14 coincide with experimental errors and strongly resemble those obtained for BR. 15 The conformation of the backbone and that of conserved side chains in the retinal binding pocket are more or less identical. Moreover, the structure of the receptor 11–14 is almost the same as that of the NpSRII–NpHtrII complex. 5 These observations pose an interesting question. How does nature fine-tune common scaffolds to engender two completely different functions: sensor and ion pump? From previous work, it is evident that NpSRII and sensory rhodopsin I are capable of pumping protons, albeit with poor efficiency. 16–19 Similarly, BR can be converted into a sensor by just three mutations. 20 On the other hand, despite great effort, it has not been possible to change NpSRII into an efficient proton pump. 21 Apparently, the requirements for an effective ion pump are much more demanding than those for a functional sensor. A second observation relates to the inhibition of the proton pump upon binding of transducer NpHtrII to its cognate receptor NpSRII. 22,23 It has been discussed that, in the 2:2 complex, the

Active State of Sensory Rhodopsin II

cytoplasmic channel cannot open sufficiently, thereby altering proton uptake kinetics such that reprotonation of the Schiff base is faster from the extracellular side. This kinetic correlation would result in a futile proton cycle (reviewed by Sasaki and Spudich 24). Certainly, this proposal can only be verified by a comparison of NpSRII active states with and without a bound transducer, NpHtrII. Crystal structures are available for the NpSRII ground state 11–13 and for its K-intermediate, which is formed at room temperature in the nanosecond range, 25 as well as for the NpSRII–HtrII complex, for which data on the ground state and on its K-intermediate and M-intermediate (the active state) are available. 5,9 Despite several attempts, the structures of the late (active) states of NpSRII have not yet been obtained. The reason for this failure has been related to large conformational changes upon light-activated M-formation, which lead to a severe disturbance of crystal packing and, consequently, a substantial decrease in resolution. Similar difficulties arise in experiments with intermediate states of the visual pigment rhodopsin. 26,27 Here, we describe an approach that allows us to determine the structure of an intermediate in cases where conformational changes in a protein, upon its transition from the ground state to an intermediate state, are large and may disturb crystal packing dramatically, making crystals unsuitable for X-ray crystallography. We were able to solve the structure of the active state of NpSRII. The data provide information about the movement of the functionally important helices F and G. Furthermore, we provide the data on the water distribution in the cytoplasmic proton uptake channel, which relates directly to the proton pump properties of NpSRII alone and in complex with its transducer, NpHtrII.

Results A comparative analysis of structural changes in NpSRII alone and in the NpSRII–NpHtrII complex necessitated crystals grown under conditions close to those reported for the complex (as described by Gordeliy et al. 5 and Moukhametzianov et al. 9 and as discussed by Reshetnyak et al. 13). We managed to obtain the crystals under similar conditions. The addition of trehalose turned out to be not only important for obtaining high-quality crystals but also crucial for preserving diffraction quality during M-state trapping. Structure of the ground state Using these conditions, we obtained the structure of the NpSRII ground state. The diffraction data were integrated up to 1.9 Å (Table 1). The built

593

Active State of Sensory Rhodopsin II Table 1. Crystallographic data collection and refinement statistics (molecular replacement)

Data collection Space group Cell dimensions a, b, c (Å) α, β, γ (°) Resolution (Å) Rmerge (%) I/σI Completeness (%) Redundancy Refinement Resolution (Å) Number of reflections Rwork/Rfree (%) Number of atoms Protein Ligand/ion Water B-factors Protein Ligand/ion Water RMSD Bond lengths (Å) Bond angles (°)

NpSRII ground state

NpSRII active state

C2221

C2221

86.99, 128.09, 50.63 90, 90, 90 28–1.9 (2.0–1.9)a 7.0 (39.3)a 14.4 (3.8)a 96.9 (97.4)a 3.8 (3.8)a

86.44, 131.09, 51.23 90, 90, 90 43–2.6 (2.74–2.6)a 11.7 (83.7)a 6.5 (1.5)a 92.7 (86.1)a 3.4 (3.1)a

22–1.9 20,672 (1107c) 15.34/17.31

20–2.6, 2.9, 2.6b 7633 (401c) 25.12/27.58

1694d 375 73

3184 (1552e) 726 (352e) 88 (15e)

11.09 58.70 27.28

18.39 (18.08e) 54.30 (54.27e) 22.66 (27.15e)

0.013 1.248

0.012 1.393

Rmerge = ∑h∑i|I(h,i) − 〈I(h)〉|/∑h∑i|I(h,i)| × 100%, where I(h,i) is the intensity value of the ith measurement of h, and 〈I(h)〉 is the corresponding mean value of h for all I measurements of h. Summation is performed over all measurements. Rwork = ∑∣Fo − Fc∣/∣∑Fo∣ × 100%. Rfree was calculated for 5% of the observed reflections which were picked randomly within thin resolution shells and were omitted from the refinement and Rwork calculations. They are consistent between data sets. a Values in parentheses are for the highest-resolution shell. b For the M-state structure, anisotropic resolution limits were used. c Number of reflections not used for refinement (free reflections). d Including double conformers. e Number of atoms whose position was refined (atoms of the active-state model).

model contains residues 1–219, with poorly ordered A–B and C–D loop regions. The level of similarity between the new structure and the published X-ray structures is very high. The RMSDs of coordinates are 0.31 Å for the 1H68 structure, 12 0.43 Å for the 1JGJ structure, 11 and 0.27 Å for the structure reported by Reshetnyak et al. 13 (the position of the A–B loop was omitted from calculations; RMSD was calculated for backbone atoms). We observed differences in the position of the C–D loop (residues 93–97), which in our structure is displaced by about 1.5 Å, as compared to published structures. The D193 position is closer to R72 in our structure than in structures 1JGJ and 1H68 for 0.8 Å and 1.8 Å respectively. We did not observe water molecules 2017 and 2026 from the 1H68

structure but identified three additional water molecules: near the Y174 carboxyl group, near the Y73 hydroxyl group, and near D193. Also, a small amount (occupancy b35%) of water molecules at position 2013 of the 1H68 structure may be replaced by chloride ions, as in the 1H68 structure, since it would better fit the electron density. The cytoplasmic loop F–G, comprising residues 175–195, is ∼0.5 Å closer to the extracellular channel in our model. Finally, in the 1JGJ structure, a number of residues have a rotameric conformation different from that in our structure or in the 1H68 structure. These differences may be ascribed to different crystallization conditions and/or to a better resolution of our structure, and seem not to be important for protein function. As for the recently published NMR structure 2KSY, 14 the differences are much larger. Substantially, they include displacements of up to 1 Å of residues known to be important for signal transduction (Y199, Y174, and T204) and changes in their hydrogen bonds. These differences may probably be attributed to the lack of proper restraints in NMR structure determination, the absence of water molecules and retinal proton in the NMR model, or the fact that the protein was in detergent micelles and not in the lipid bilayer. We also observe a strong similarity between the ground-state structure of NpSRII alone and the ground-state structure of NpSRII in complex with NpHtrII. The RMSDs of coordinates are 0.53 Å for the 1H2S structure 5 and 0.45 Å for the structure in space group I212121 (A.I. et al., unpublished data). There are only slight differences in poorly ordered loop regions, which include also the position of the C–D loop. An important observation concerns D193, which occupies the same position in the structure of the complex and in the present structure, contrary to its position in structures 1H68 and 1JGJ. These data emphasize the importance of identical crystallization conditions. Additionally, it was possible to assign 31 hydrocarbon chains (Fig. 1). Two of them may be identified unambiguously as archaeal phytanyl chains by their electron densities. Both eubacterial and archaeal lipids may be present in the crystal, as NpSRII was expressed in Escherichia coli and subsequently reconstituted into Halobacterium salinarum polar lipids. Additionally, one molecule of octyl-β-D-glucopyranoside could be identified. Inclusion of the lipid molecules in the model turned out to be necessary for obtaining the active-state structure of NpSRII because it resulted in more consistent electron density maps and better R-factors. Heterogeneity of the active state The procedure employed in this work did not allow for the trapping of a pure M-state. The next

594

Active State of Sensory Rhodopsin II

Fig. 1. Lipid tails in the ground-state model of NpSRII. (a) The receptor binding face. (b) The face opposite to the receptor binding face. Lipid tails are shown in green stick representations. Surface potential was calculated using ABPS. 28 Blue corresponds to a potential value of +3 kT/e or higher, and red corresponds to a potential value of −3 kT/e or lower. Lipid bilayer was not included in the calculation. NpSRII helices are marked by corresponding letters. Tyrosine Y199 and the region of positive potential near helix F correspond to the place of NpHtrII transducer binding in the NpSRII–NpHtrII complex.

intermediate in the photocycle, the O-intermediate, could have been accumulated under these conditions as well. Indeed, spectroscopic data show that the amount of the O-state reaches up to two-thirds that of the M-state (see Supplementary Information and Supplementary Fig. 1). Independent refinement of the M and O structures is impossible due to a low data-to-parameter ratio, which is slightly over 1 when only one intermediate structure is refined. Nevertheless, independent data from the literature indicate that the structure of the M-state closely resembles that of the O-state. First, the photophobic signal is relayed to the transducer in both the Mstate and the O-state. 29,30 EPR experiments also show that reordering of helix F and TM2, which is believed to represent the signal, decays concomitantly with O-decay. 6 Fourier transform infrared spectroscopy studies 31–33 have shown that differences between the M-state and the O-state are much smaller than those between the M-state and the ground state. Transient grating analysis also shows little volume changes upon M-to-O transition compared to transition from the ground state to the M-intermediate. 34 The main difference concerns the backisomerization of the retinal to the alltrans conformation. 31–33 The global similarity of the late intermediate structures was also proposed for BR. 35 From these considerations, it can be assumed

that the structures of the M-intermediate and the Ointermediate are quite similar at the present resolution. It is therefore reasonable to treat the M-state and the O-state as one entity. This assumption is also justified by the fact that the occupancy of the active state obtained by crystallographic means is in good agreement with the results of spectroscopy, as the combined M-state and O-state amount to about 50%. Finally, it should be noted that, in the case of the NpSRII–NpHtrII complex, the trapped active state contained not only the M-intermediate but also some amount of the O-intermediate. Nonetheless, no large unaccounted difference in electron densities was seen when the data were fitted with one model representing the active state of the protein. The resolution of the structure was 2.0 Å; thus, the large structural deviation in the O-state, if at all present, would have been detected. 9 Trapping the active state in crystals Upon trapping of the active state in crystals, diffraction worsened anisotropically. The largest loss in resolution (1 Å) occurred along the b-axis, which is perpendicular to the membrane plane. At the same time, we observed large changes in unit cell dimensions. The most pronounced change was again observed in the b-axis direction, with

595

Active State of Sensory Rhodopsin II

the cell dimension increasing from 128 Å to 131 Å. These effects were stronger with the higher intensity of the light used in the trapping of the active state. Spectroscopy measurements showed that illumination of the crystals (excitation wavelength, 488 nm) with a laser power of 0.33 W for 2 s (the cryostream is blocked for that period) results in the accumulation of approximately 35% M-state and 25% O-state. A further increase in laser power slightly raises the amounts of the accumulated active states to the maxima of ∼ 45% M-state and 30% O-state. These data indicate that the more the active state is present in the crystal, the larger are the diffraction resolution losses. Therefore, optimum conditions, under which both active-state occupancy and resolution would be reasonable, had to be established. We found that at a laser power of 0.33 W, the resolution loss along the b-axis does not exceed 1 Å; meanwhile, the trapped amount of the active state allows for structure determination. A number of diffraction data sets with different NpSRII crystals were obtained by using such an approach. The resolution limits for the best crystal were 2.5 Å, 2.9 Å, and 2.6 Å for axes a, b, and c, respectively. For that crystal, the amount of the active state determined crystallographically was found to be approximately 45% (see Supplementary Information and Supplementary Fig. 2).

Conformational changes in the NpSRII active state As predicted from the loss of diffraction quality during trapping, we observe large conformational changes in the NpSRII active state (Fig. 2a). The structural changes reported below are reproducible for three different illuminated crystals displaying the highest resolution, for which refinement of the active state was possible. The changes are in accordance with difference Fourier electron density maps between the active state and the ground state. Thereafter, we discuss the changes as they are observed in the structure of the highest resolution. The alignment of structures for the comparison was performed by nonhydrogen atoms for the whole protein. Large conformational changes are observed at the transducer binding interface of helices F and G. Helix F moves towards the cytoplasm by about 0.3 Å, whereas helix G moves in opposite direction towards the extracellular side by 0.7 Å. Helix A moves together with helix G, although with twice-as-lower amplitude. The extracellular F–G loop moves out of the membrane plane. These movements are similar to those observed for the NpSRII–NpHtrII complex but display larger amplitudes (Fig. 2). Movement of the G helix with Y199, which forms crystal contact in case of NpSRII, seems to be the main

Fig. 2. Overall structural rearrangements in NpSRII upon transition to the active state. (a) Structural rearrangements in uncomplexed NpSRII. (b) Structural rearrangements in NpSRII in complex with NpHtrII 9 (Protein Data Bank ID: 2F95). The ground state is shown in blue, and the active state is shown in magenta. NpHtrII in (b) is shown in gray for both states. Lysine 205, retinal, and tyrosine 199 are shown as ball-and-stick models. NpSRII helices are marked by corresponding letters (larger letters correspond to front helices, and smaller letters correspond to back helices). Arrows show the movements of the corresponding helices. Alignment was performed by NpSRII backbone atoms.

596 cause of the increase in crystal b-dimension, as the increase (3 Å) corresponds well to 4 × 0.7 Å, a number which follows from geometrical considerations. On the cytoplasmic surface, the E–F loop changes its conformation and becomes more compressed. The movement of the C–D loop is questionable, since this loop is poorly ordered and its position differs, as exemplified above in the known groundstate structures. The alterations on the surface are paralleled by changes in the interior of light-activated NpSRII. The difference density maps clearly show that 13-cisretinal is present in the crystals, amounting to about 35% of the M-intermediate, as determined spectroscopically (see the text above). Interestingly, the position of T204 relative to Y174 and helix F does not change (Fig. 3). Unfortunately, the structure resolution does not allow for an analysis of changes in the position of water molecule, bonding to W171 and T204 backbone. W171, forming a steric conflict with the retinal C13 methyl group, is pushed away from retinal by ∼ 0.6 Å. W76 is displaced towards helix F in the membrane plane. Electron densities are observed only for few highly ordered water molecules in the extracellular channel. Nevertheless, the absence of a water molecule adjacent to the Schiff base (analogous to W402 of BR) is clearly seen, as there are no 2Fo − Fc densities in that region and, in addition, inclusion of this water molecule into the model results in pronounced negative difference electron density maps Fo − Fc. The disappearance of this water molecule is the result of Schiff base deprotonation

Fig. 3. Comparison of retinal environments in the NpSRII ground-state structure (blue) and active-state structure (magenta). The active state is a mix of the M-state and the O-state, and retinal is shown only as it would be in the M-state in 13-cis conformation. The positions of other residues represent the averages of those in the M-state and the O-state, and reflect the general conformational changes in this region. Hydrogen bonds are shown by broken lines. Water molecule 3 disappears in the active state.

Active State of Sensory Rhodopsin II

and was observed in the NpSRII–NpHtrII complex 9 and in BR 36 as well.

Discussion We discuss the results in three different contexts. First, we compare the changes in the active state of uncomplexed NpSRII to those in the NpSRII– NpHtrII complex. Second, we analyze the implications of the determined structure for understanding proton pumping by NpSRII. Finally, we discuss the methodological aspects of the observation of large conformational changes in proteins in crystals. Functionally important conformational changes in NpSRII The conformational changes observed for free NpSRII and for NpSRII in complex with its transducer are quite similar in nature (Fig. 2). Major movements are observed for helices F and G, which have been correlated to signal propagation from the receptor to the transducer. 6,9,37,38 Additionally, an inward movement of the extracellular part of helix C is observed. Although these conformational changes are detected in both cases, in the case of NpSRII alone, they have almost twiceas-large amplitudes (Fig. 2). Smaller changes in the complexed NpSRII cannot be attributed to crystal packing, as it is looser in the crystals of NpSRII– NpHtrII, especially in space group I212121 (A.I. et al., unpublished data). Thus, it means that the transducer opposes such structural rearrangements in the receptor. As this opposition leads to conformational changes in the transducer itself, it constitutes a signal transfer. In addition to helix movements, we observe conformational changes in the E–F loop, which are not present in the NpSRII–NpHtrII complex. They may result from crystal packing (E–F loops of adjacent proteins form a crystal contact) or may be a genuine conformational change in the active state of uncomplexed NpSRII. It has been shown by solidstate NMR spectroscopy, fluorescence spectroscopy, and EPR spectroscopy that the E–F loop interacts tightly with the transducer. 39–41 However, it is clear that the E–F loop is not crucial for signal transduction, as the NpSRII E–F loop deletion mutants, as well as the BR triple mutant (with a small disordered E–F loop), are all capable of signaling. 20,42 The remaining question is whether the E–F loop enhances the signal transduction from NpSRII to NpHtrII or just serves for NpSRII stabilization. So far, we have discussed the large conformational changes between the ground state and the active state of NpSRII. Unfortunately, due to the resolution of diffraction data, we are not able to observe either the minute changes in the structure

597

Active State of Sensory Rhodopsin II

or the movements of water molecules in the cytoplasmic channel. Nonetheless, we are able to conclude that there are no changes in the uncomplexed NpSRII active state in the relative positions of residues T204 and Y174 (Fig. 3), which are proposed to be important for signal generation in the NpSRII– NpHtrII complex. 43,44 Probably, this reflects the spectroscopically observed absence of alteration in T204-Y174 bonding in uncomplexed NpSRII. 45,46 Implications for NpSRII proton pumping Like BR, NpSRII is able to pump protons across the membrane, although less effectively. Upon formation of the NpSRII–NpHtrII complex, this function is inhibited. 22,23 This inhibition may originate from kinetics alteration, structural changes that prohibit pumping, or a combination of these factors. Here, we analyze the implications of structural changes. NpSRII active-state transition results in the disappearance of the water molecule, which makes a hydrogen bond to the retinal in the ground state (Fig. 3). The same is observed in the BR M-state, 47 as well as in the NpSRII–NpHtrII complex M-state. 9 There are also conformational changes of up to 1 Å in the cytoplasmic channel, which are probably associated with charge redistribution. A slight outward tilt of the G helix, combined with E–F loop changes, may facilitate the formation of a water chain, which is believed to be needed for retinal reprotonation and thus for completion of the pumping cycle. These changes are not observed in the M-state of complexed NpSRII and may account for the inhibition of NpHtrII proton pumping. There are two ways in which NpHtrII may affect NpSRII dynamics. The first possibility is that NpHtrII may directly influence NpSRII conformational changes. The other possibility is as follows. The hydrophobic/hydrophilic boundary of NpSRII is uneven in the region of helices F and G, as seen from the surface potential distribution (Fig. 1). Thus, the natural tendency of a lipid bilayer to be flat would result in an energetic penalty for the NpSRII ground state. In the active state, the mismatch is lowered; thus, the active state is preferred from the bilayer curvature energy point of view. NpHtrII shields helices F and G from the lipid bilayer and thus reduces its influence on the NpSRII structure. In that case, the energy of the active state is higher than in the absence of NpHtrII; thus, the conformational changes are lower. Observation of large conformational changes in the protein in crystals

measures should be undertaken to obtain tractable data. The parameters of intermediate-state accumulation should be thoroughly studied. The ideal loading procedure will result in not less than 40–50% of the intermediate. At the same time, multiple transitions from the ground state to the intermediate state, and back, should be avoided, as repetitive changes will cause more damage to the packing. Thus, a short loading time is preferred. Diffraction data should be measured from several crystals to sample for those that diffract better and to check the consistency of the observed changes. Data should be treated with all available techniques to obtain the best resolution possible. In our case, that meant utilization of anisotropic data and construction of the full model with as many incorporated lipid molecules as possible. Finally, intermediate-state occupancy should be evaluated for each crystal (using crystallography techniques) because, often, the crystals that seem to have better resolution in fact contain not many protein molecules trapped in the intermediate state.

Materials and Methods Protein preparation and crystallization Protein preparation and crystallization were performed as described previously. 48,49 The genes of NpSRII were cloned into a pET27bmod expression vector with a C-terminal 7× His tag. Protein was expressed in E. coli strain BL21(DE3) and purified as described previously. 48,49 After removal of imidazole by diethylaminoethyl cellulose chromatography, NpSRII-His was reconstituted into purple membrane lipids (protein-to-lipid ratio, 1:35). After filtration, the reconstituted protein was pelleted by centrifugation at 100,000g. For resolubilization, the samples were resuspended in a buffer containing 2% n-octyl-β-D-glucopyranoside and shaken for 16 h at 4 °C in the dark. The resolubilized protein was isolated by centrifugation at 100,000g. Solubilized protein (20 mg/ml) in crystallization buffer (150 mM NaCl and 25 mM Na/KPi, pH 5.1) was added to the lipid phase formed from monooleoyl (NU-CHEK PREP, Inc., USA). The precipitant used was 1 M salt Na/KPi (pH 5.1) with 0.3 M trehalose. Crystals were grown at 22.5 °C. The addition of trehalose improved the quality of grown NpSRII crystals. It cannot be excluded that the improvement is due to the “trapped” dynamic protein conformations. 50 Nevertheless, we should also mention two other possible sources of the influence of trehalose on the quality of membrane protein crystals grown in meso: improvement of cryoprotection and direct influence on the crystallization matrix. This ongoing study will help to better understand the observed effect. Trapping of intermediates

As was already mentioned, the observation of large conformational changes in the protein in crystals is complicated by the loss of long-range crystal order and the consequent diffraction loss. Therefore, special

For the loading of the intermediate state, the crystals were illuminated with an argon–krypton ion laser at a wavelength of 488 nm (Omnichrome). Cryostream was

598 blocked for 2 s. The illumination was turned off permanently after 1 s has passed after the recooling has started. X-ray data were collected in the dark. Spectroscopic characterization of the crystals The visual spectra of single crystals were measured with an IR–Vis spectrophotometer described elsewhere. 51 Spectra from single crystals were recorded in nitrogen cryostream under conditions identical with those used for X-ray measurements. Data were fitted by skewed Gaussians, with modified parameters from Chizhov et al. 52 Acquisition and treatment of diffraction data X-ray diffraction data (wavelength, 0.934 Å) for the ground state and the active state were collected at beamlines ID14-1 and ID23-1 of the European Synchrotron Radiation Facility (Grenoble, France), respectively. The image acquisition time was 10 s and the oscillation range was 0.9°, and a total of 105 images were collected in each data set. Diffraction patterns were processed with MOSFLM/SCALA (CCP4 53). Anisotropic correction was performed via an anisotropy server. 54 Crystallographic parameters and refinement statistics are summarized in Table 1. Crystallographic refinement For ground-state structure determination, the 1H68 model was iteratively refined using REFMAC5 55 and Coot. 56 For use in signalling-state structure determination, the resulting ground-state structure was minimized in CNS 57 in order to equilibrate the model with a CNS parameter set. q-Weighted difference density maps built with CNS were used for the preliminary analysis of structural changes.58 Most differences were located in the retinal binding pocket and along helices F and G. Isomerization and deviation of the retinal out of the F-helix, as well as displacement of W171, are clearly seen for all data sets. The changes in the different crystals were similar. Intermediate-state refinement was conducted as follows. First, the ground-state model was subjected to rigidbody refinement. The part of it corresponding to the ground state (occupancy, 55%) was fixed, and the other part corresponding to the intermediate state (occupancy, 45%) was subjected to simulated annealing. For the annealed part of the model, a less ordered loop and terminal residues (residues 1–2, 28–32, and 217–219), along with surface water molecules and small lipid fragments, were removed. Internal water molecules, expected to be present but not clearly seen at the present resolution, were harmonically restrained to positions in the ground-state structure. After simulated annealing, the conformations of some residues were manually changed towards ideal geometry. Accession codes Coordinates and structure factors have been deposited in the Protein Data Bank under accession codes 3QAP and 3QDC.

Active State of Sensory Rhodopsin II

Acknowledgements We would like to thank the Structural Biology Group of the European Synchrotron Radiation Facility. We are also thankful to C. Baeken for help with NpSRII preparation. This work was supported by the program “Chaires d'excellence” edition 2008 of ANR France, by a CEA(IBS)-HGF(FZJ) STC 5.1 specific agreement, and by a Marie Curie grant for the training and career development of researchers (FP7-PEOPLE-2007-1-1-ITN, project SBMPs). This work was performed in the framework of Russian State contract numbers 02.740.11.0299, 02.740.11.5010, 974 (in the framework of activity 1.2.2), and P211 (in the framework of activity 1.3.2) of the Federal Target Program “Scientific and academic research cadres of innovative Russia” for the years 2009– 2013. Financial support from the ONEXIM group and the Skolkovo foundation is gratefully acknowledged.

Supplementary Data Supplementary data to this article can be found online at doi:10.1016/j.jmb.2011.07.022

References 1. Sharma, A. K., Spudich, J. L. & Doolittle, W. F. (2006). Microbial rhodopsins: functional versatility and genetic mobility. Trends Microbiol. 14, 463–469. 2. Klare, J., Chizhov, I. & Engelhard, M. (2008). Microbial rhodopsins: scaffolds for ion pumps, channels, and sensors. In Bioenergetics: Energy Conservation and Conversion (Schäfer, G. & Penefsky, H. S., eds), pp. 73–122, Springer, Berlin. 3. Zhang, F., Wang, L. P., Brauner, M., Liewald, J. F., Kay, K., Watzke, N. et al. (2007). Multimodal fast optical interrogation of neural circuitry. Nature, 446, 633–639. 4. Sasaki, J. & Spudich, J. L. (2008). Signal transfer in haloarchaeal sensory rhodopsin–transducer complexes. Photochem. Photobiol. 84, 863–868. 5. Gordeliy, V. I., Labahn, J., Moukhametzianov, R., Efremov, R., Granzin, J., Schlesinger, R. et al. (2002). Molecular basis of transmembrane signalling by sensory rhodopsin II–transducer complex. Nature, 419, 484–487. 6. Wegener, A. A., Klare, J. P., Engelhard, M. & Steinhoff, H. J. (2001). Structural insights into the early steps of receptor–transducer signal transfer in archaeal phototaxis. EMBO J. 20, 5312–5319. 7. Wegener, A. A., Chizhov, I., Engelhard, M. & Steinhoff, H. J. (2000). Time-resolved detection of transient movement of helix F in spin-labelled pharaonis sensory rhodopsin II. J. Mol. Biol. 301, 881–891. 8. Bordignon, E., Klare, J. P., Holterhues, J., Martell, S., Krasnaberski, A., Engelhard, M. & Steinhoff, H. J. (2007). Analysis of light-induced conformational

599

Active State of Sensory Rhodopsin II

9.

10. 11.

12.

13.

14.

15. 16. 17.

18.

19.

20.

21.

22.

changes of Natronomonas pharaonis sensory rhodopsin II by time resolved electron paramagnetic resonance spectroscopy. Photochem. Photobiol. 83, 263–272. Moukhametzianov, R., Klare, J. P., Efremov, R., Baeken, C., Göppner, A., Labahn, J. et al. (2006). Development of the signal in sensory rhodopsin and its transfer to the cognate transducer. Nature, 440, 115–119. Hazelbauer, G. L., Falke, J. J. & Parkinson, J. S. (2008). Bacterial chemoreceptors: high-performance signaling in networked arrays. Trends Biochem. Sci. 33, 9–19. Luecke, H., Schobert, B., Lanyi, J. K., Spudich, E. N. & Spudich, J. L. (2001). Crystal structure of sensory rhodopsin II at 2.4 angstroms: insights into color tuning and transducer interaction. Science, 293, 1499–1503. Royant, A., Nollert, P., Edman, K., Neutze, R., Landau, E. M., Pebay-Peyroula, E. & Navarro, J. (2001). X-ray structure of sensory rhodopsin II at 2.1-Å resolution. Proc. Natl Acad. Sci. USA, 98, 10131–10136. Reshetnyak, A. B., Borshchevskiy, V. I., Klare, J., Moiseeva, E. S., Engelhard, M., Buldt, G. & Gordeliy, V. I. (2008). Comparative analysis of sensory rhodopsin II structures in complex with a transducer and without it. J. Surf. Invest. X-ray Synchrotron Neutron Tech. 2, 894–899. Gautier, A., Mott, H. R., Bostock, M. J., Kirkpatrick, J. P. & Nietlispach, D. (2010). Structure determination of the seven-helix transmembrane receptor sensory rhodopsin II by solution NMR spectroscopy. Nat. Struct. Mol. Biol. 17, 768–774. Luecke, H., Schobert, B., Richter, H. T., Cartailler, J. P. & Lanyi, J. K. (1999). Structure of bacteriorhodopsin at 1.55 Å resolution. J. Mol. Biol. 291, 899–911. Sasaki, J. & Spudich, J. L. (1999). Proton circulation during the photocycle of sensory rhodopsin II. Biophys. J. 77, 2145–2152. Bogomolni, R. A., Stoeckenius, W., Szundi, I., Perozo, E., Olson, K. D. & Spudich, J. L. (1994). Removal of transducer HtrI allows electrogenic proton translocation by sensory rhodopsin I. Proc. Natl Acad. Sci. USA, 91, 10188–10192. Iwamoto, M., Shimono, K., Sumi, M., Koyama, K. & Kamo, N. (1999). Light-induced proton uptake and release of pharaonis phoborhodopsin detected by a photoelectrochemical cell. J. Phys. Chem. B, 103, 10311–10315. Schmies, G., Lüttenberg, B., Chizhov, I., Engelhard, M., Becker, A. & Bamberg, E. (2000). Sensory rhodopsin II from the haloalkaliphilic Natronobacterium pharaonis: light-activated proton transfer reactions. Biophys. J. 78, 967–976. Sudo, Y. & Spudich, J. L. (2006). Three strategically placed hydrogen-bonding residues convert a proton pump into a sensory receptor. Proc. Natl Acad. Sci. 103, 16129–16134. Klare, J. P., Schmies, G., Chizhov, I., Shimono, K., Kamo, N. & Engelhard, M. (2002). Probing the proton channel and the retinal binding site of Natronobacterium pharaonis sensory rhodopsin II. Biophys. J. 82, 2156–2164. Schmies, G., Engelhard, M., Wood, P. G., Nagel, G. & Bamberg, E. (2001). Electrophysiological characterization of specific interactions between bacterial sensory

23.

24. 25.

26. 27. 28.

29.

30.

31.

32.

33.

34.

35. 36.

37.

rhodopsins and their transducers. Proc. Natl Acad. Sci. USA, 98, 1555–1559. Sudo, Y., Iwamoto, M., Shimono, K., Sumi, M. & Kamo, N. (2001). Photo-induced proton transport of pharaonis phoborhodopsin (sensory rhodopsin II) is ceased by association with the transducer. Biophys. J. 80, 916–922. Sasaki, J. & Spudich, J. L. (2000). Proton transport by sensory rhodopsins and its modulation by transducerbinding. Biochim. Biophys. Acta, 1460, 230–239. Edman, K., Royant, A., Nollert, P., Maxwell, C. A., Pebay-Peyroula, E., Navarro, J. et al. (2002). Early structural rearrangements in the photocycle of an integral membrane sensory receptor. Structure, 10, 473–482. Ruprecht, J. J., Mielke, T., Vogel, R., Villa, C. & Schertler, G. F. (2004). Electron crystallography reveals the structure of metarhodopsin I. EMBO J. 23, 3609–3620. Choe, H. W., Kim, Y. J., Park, J. H., Morizumi, T., Pai, E. F., Krauss, N. et al. (2011). Crystal structure of metarhodopsin II. Nature, 471, 651–655. Baker, N. A., Sept, D., Joseph, S., Holst, M. J. & McCammon, J. A. (2001). Electrostatics of nanosystems: application to microtubules and the ribosome. Proc. Natl Acad. Sci. USA, 98, 10037–10041. Taniguchi, Y., Ikehara, T., Kamo, N., Yamasaki, H. & Toyoshima, Y. (2007). Dynamics of light-induced conformational changes of the phoborhodopsin/ transducer complex formed in the n-dodecyl β-Dmaltoside micelle. Biochemistry, 46, 5349–5357. Yan, B., Takahashi, T., Johnson, R. & Spudich, J. L. (1991). Identification of signaling states of a sensory receptor by modulation of lifetimes of stimulusinduced conformations: the case of sensory rhodopsin II. Biochemistry, 30, 10686–10692. Hein, M., Wegener, A. A., Engelhard, M. & Siebert, F. (2003). Time-resolved FTIR studies of sensory rhodopsin II (NpSRII) from Natronobacterium pharaonis: implications for proton transport and receptor activation. Biophys. J. 84, 1208–1217. Bergo, V., Spudich, E. N., Spudich, J. L. & Rothschild, K. J. (2003). Conformational changes detected in a sensory rhodopsin II–transducer complex. J. Biol. Chem. 278, 36556–36562. Furutani, Y., Iwamoto, M., Shimono, K., Wada, A., Ito, M., Kamo, N. & Kandori, H. (2004). FTIR spectroscopy of the O photointermediate in pharaonis phoborhodopsin. Biochemistry, 43, 5204–5212. Inoue, K., Sasaki, J., Spudich, J. L. & Terazima, M. (2007). Laser-induced transient grating analysis of dynamics of interaction between sensory rhodopsin II D75N and the HtrII transducer. Biophys. J. 92, 2028–2040. Subramaniam, S. B. P., Tittor, J. O. D. & Lanyi, J. H. R. (1999). Protein conformational changes in the bacteriorhodopsin photocycle. J. Mol. Biol. 287, 145–161. Sass, H. J., Büldt, G., Gessenich, R., Hehn, D., Neff, D., Schlesinger, R. et al. (2000). Structural alterations for proton translocation in the M state of wild-type bacteriorhodopsin. Nature, 406, 649–653. Bergo, V. B., Spudich, E. N., Rothschild, K. J. & Spudich, J. L. (2005). Photoactivation perturbs the membrane-embedded contacts between sensory rhodopsin II and its transducer. J. Biol. Chem. 280, 28365–28369.

600 38. Yoshida, H., Sudo, Y., Shimono, K., Iwamoto, M. & Kamo, N. (2004). Transient movement of helix F revealed by photo-induced inactivation by reaction of a bulky SH-reagent to cysteine-introduced pharaonis phoborhodopsin (sensory rhodopsin II). Photochem. Photobiol. Sci. 3, 537–542. 39. Bordignon, E., Klare, J. P., Doebber, M., Wegener, A. A., Martell, S., Engelhard, M. & Steinhoff, H. J. (2005). Structural analysis of a HAMP domain. J. Biol. Chem. 280, 38767–38775. 40. Etzkorn, M., Seidel, K., Li, L., Martell, S., Geyer, M., Engelhard, M. & Baldus, M. (2010). Complex formation and light activation in membrane-embedded sensory rhodopsin II as seen by solid-state NMR spectroscopy. Structure, 18, 293–300. 41. Yang, C. S., Sineshchekov, O., Spudich, E. N. & Spudich, J. L. (2004). The cytoplasmic membraneproximal domain of the HtrII transducer interacts with the E–F loop of photoactivated Natronomonas pharaonis sensory rhodopsin II. J. Biol. Chem. 279, 42970–42976. 42. Sasaki, J., Nara, T., Spudich, E. N. & Spudich, J. L. (2007). Constitutive activity in chimeras and deletions localize sensory rhodopsin II/HtrII signal relay to the membrane-inserted domain. Mol. Microbiol. 66, 1321–1330. 43. Sudo, Y., Furutani, Y., Kandori, H. & Spudich, J. L. (2006). Functional importance of the interhelical hydrogen bond between Thr204 and Tyr174 of sensory rhodopsin II and its alteration during the signaling process. J. Biol. Chem. 281, 34239. 44. Ito, M., Sudo, Y., Furutani, Y., Okitsu, T., Wada, A., Homma, M. et al. (2008). Steric constraint in the primary photoproduct of sensory rhodopsin II is a prerequisite for light-signal transfer to HtrII. Biochemistry, 47, 6208–6215. 45. Furutani, Y., Sudo, Y., Kamo, N. & Kandori, H. (2003). FTIR spectroscopy of the complex between pharaonis phoborhodopsin and its transducer protein. Biochemistry, 42, 4837–4842. 46. Sudo, Y., Furutani, Y., Shimono, K., Kamo, N. & Kandori, H. (2003). Hydrogen bonding alteration of Thr-204 in the complex between pharaonis phoborhodopsin and its transducer protein. Biochemistry, 42, 14166–14172. 47. Lanyi, J. K. (2006). Proton transfers in the bacteriorhodopsin photocycle. Biochim. Biophys. Acta, 1757, 1012–1018.

Active State of Sensory Rhodopsin II 48. Hohenfeld, I. P., Wegener, A. A. & Engelhard, M. (1999). Purification of histidine tagged bacteriorhodopsin, pharaonis halorhodopsin and pharaonis sensory rhodopsin II functionally expressed in Escherichia coli. FEBS Lett. 442, 198–202. 49. Shimono, K., Iwamoto, M., Sumi, M. & Kamo, N. (1997). Functional expression of pharaonis phoborhodopsin in Escherichia coli. FEBS Lett. 420, 54–56. 50. Cordone, L., Ferrand, M., Vitrano, E. & Zaccai, G. (1999). Harmonic behavior of trehalose-coated carbon-monoxy-myoglobin at high temperature. Biophys. J. 76, 1043–1047. 51. Gordeliy, V. I., Schlesinger, R., Efremov, R., Büldt, G. & Heberle, J. (2003). Crystallization in lipidic cubic phases. In Membrane Protein Protocols: Expression, Purification, and Characterization (Selinsky, B, ed.), Humana Press, Totowa, NJ. 52. Chizhov, I., Schmies, G., Seidel, R., Sydor, J. R., Lüttenberg, B. & Engelhard, M. (1998). The photophobic receptor from Natronobacterium pharaonis: temperature and pH dependencies of the photocycle of sensory rhodopsin II. Biophys. J. 75, 999–1009. 53. Collaborative Computational Project Number 4 (1994). The CCP4 Suite: programs for protein crystallography. Acta Crystallogr., Sect. D: Biol. Crystallogr. 50, 760–763. 54. Strong, M., Sawaya, M. R., Wang, S., Phillips, M., Cascio, D. & Eisenberg, D. (2006). Toward the structural genomics of complexes: crystal structure of a PE/PPE protein complex from Mycobacterium tuberculosis. Proc. Natl Acad. Sci. 103, 8060. 55. Murshudov, G. N., Vagin, A. A. & Dodson, E. J. (1997). Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr., Sect. D: Biol. Crystallogr. 53, 240–255. 56. Emsley, P. & Cowtan, K. (2004). Coot: model-building tools for molecular graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 60, 2126–2132. 57. Brünger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998). Crystallography & NMR System: a new software suite for macromolecular structure determination. Acta Crystallogr., Sect. D: Biol. Crystallogr. 54, 905–921. 58. Ursby, T. & Bourgeois, D. (1997). Improved estimation of structure–factor difference amplitudes from poorly accurate data. Acta Crystallogr., Sect. A: Found. Crystallogr. 53, 564–575.