coupling photoisomerization of retinal to directional ...€¦ · the principal question of the...

doi:10.1006/jmbi.2000.3884 available online at http://www.idealibrary.com on J. Mol. Biol. (2000) 300, 1237±1255

Coupling Photoisomerization of Retinal to DirectionalTransport in Bacteriorhodopsin

Hartmut Luecke1,2,4*, Brigitte Schobert2,4, Jean-Philippe Cartailler1,4

Hans-Thomas Richter2,4, Anja Rosengarth1,4, Richard Needleman3

and Janos K. Lanyi2,4*

1Department of MolecularBiology and BiochemistryUniversity of California, IrvineCA, 92697, USA2Department of Physiology andBiophysics, University ofCalifornia, IrvineCA 92697, USA3Department of BiochemistryWayne State UniversityDetroit, MI 48201, USA4UCI program inMacromolecular StructureUniversity of California, IrvineCA 92697, USA

E-mail address of the [email protected]; [email protected].

{To distinguish it from the intermlight-adapted but unilluminated stabacteriorhodopsin is denoted as BR

Abbreviation used: FTIR, Fourier

0022-2836/00/051237±19 $35.00/0

In order to understand how isomerization of the retinal drives uni-directional transmembrane ion transport in bacteriorhodopsin, we deter-mined the atomic structures of the BR state and M photointermediate ofthe E204Q mutant, to 1.7 and 1.8 AÊ resolution, respectively. Comparisonof this M, in which proton release to the extracellular surface is blocked,with the previously determined M in the D96N mutant indicates that thechanges in the extracellular region are initiated by changes in the electro-static interactions of the retinal Schiff base with Asp85 and Asp212, butthose on the cytoplasmic side originate from steric con¯ict of the13-methyl retinal group with Trp182 and distortion of the p-bulge ofhelix G. The structural changes suggest that protonation of Asp85initiates a cascade of atomic displacements in the extracellular region thatcause release of a proton to the surface. The progressive relaxation of thestrained 13-cis retinal chain with deprotonated Schiff base, in turn,initiates atomic displacements in the cytoplasmic region that cause theintercalation of a hydrogen-bonded water molecule between Thr46 andAsp96. This accounts for the lowering of the pKa of Asp96, which thenreprotonates the Schiff base via a newly formed chain of water moleculesthat is extending toward the Schiff base.

# 2000 Academic Press

Keywords: bacteriorhodopsin, membrane proteins, X-ray crystallography,photocycle intermediates
*Corresponding authors
Introduction

The relative simplicity of ion translocation inbacteriorhodopsin (Haupts et al., 1997, 1999; Lanyi,1997, 1998a,b; Oesterhelt, 1998; Wikstrom, 1998),based on photoisomerization of the all-trans retinalchromophore to 13-cis,15-anti and the ensuing reac-tion cycle by which the retinal thermally reisome-rizes, has motivated three decades of intensivework with this protein. Indeed, from a vastamount of static and time-resolved spectroscopy ofvarious kinds, site-speci®c mutagenesis, and low-resolution three-dimensional and projection maps,a fairly complete description has emerged for the

ing authors:eduediate states, the

te of.

-transform infra-red.

distinguishing characteristics of intermediate statesand the nature of their interconversions in thelight-driven transport cycle.

The principal question of the mechanism of thispump is how the photoisomerization of the retinal,with only local vectoriality, can drive the direc-tional movement of the transported ion across thewidth of the membrane. The answer lies in thenature of the photocycle reactions, as well as thestructure of the protein and its changes during thetransport. In kinetic terms, the photochemical cycleof bacteriorhodopsin appears to be a linearsequence of the spectroscopically identi®ed statesBR,{ J, K, L, M, N and O (reviewed by Lanyi &VaÂroÂ , 1995). The K state is distinguished by a dis-torted 13-cis,15-anti con®guration of the retinal,evident from its large-amplitude hydrogen-out-of-plane vibrational bands (Braiman & Mathies 1982;Siebert & Mantele 1983), and changes of thehydrogen-bonding of water (Fischer et al., 1994).The retinal is partly relaxed in the L state, and

# 2000 Academic Press

1238 Directional Transport in Bacteriorhodopsin

hydrogen-bonds of protein groups and boundwater begin to change (Smith et al., 1986; Roepeet al., 1987; Maeda et al., 1992, 1994, 1999;Yamazaki et al., 1995, 1996; Zscherp & Heberle,1997). The M intermediate consists of at least threesubstates in the model L!M1!M2!M

02 (ZimaÂnyi

et al., 1992b; Brown et al., 1994a; Dickopf & Heyn1997), in which the retinal Schiff base comes to aprotonation equilibrium with the proton acceptorAsp85, located toward the extracellular side, andthis equilibrium is then shifted in two steps tonearly full deprotonation of the Schiff base. Thetransition of M1 to M2 is a reversible step of an asyet unknown nature, and the transition of M

02 is

associated with the release of a proton to the extra-cellular surface. This proton can be detected by pHindicator dyes, either covalently bound to the pro-tein or in the bulk (Heberle & Dencher, 1992;ZimaÂnyi et al., 1992b; Scherrer et al., 1994; Alexievet al., 1994; Brown et al., 1994a; Cao et al., 1995).Although there is disagreement over the origin ofthe released proton (Brown et al., 1995a; Essen et al.,1998; Rammelsberg et al., 1998; Luecke et al., 1998,1999b), it seems clear (Balashov et al., 1995, 1996;Richter et al., 1996a,c) that the decrease of the pKa

of the proton release site is coupled to the protona-tion of Asp85. Because M1 is in equilibrium withthe L state, but M2` is in equilibrium with the next,the N state, the reaction M1!M2 is often referredto as the ``protonation switch,`` that changes ``pro-ton access'' of the Schiff base from the extracellularto the cytoplasmic side. In a more recent model ofthe transport (Brown et al., 1998), however, theswitch cannot be localized to a single reaction ofthe photocycle, but consists of a combination ofevents in both extracellular and cytoplasmicregions, as well as the varying local access of theSchiff base.

The N state arises when the Schiff base is repro-tonated, while Asp96 to the cytoplasmic sidedeprotonates (Gerwert et al., 1989; BouscheÂ et al.,1991). In the D96N mutant, a late M state, MN, hasbeen identi®ed (Sasaki et al., 1992), which in somerespects resembles N. In the later M (whether it beM2, M

02 or MN) and N, projection maps and par-

tially three-dimensional maps (Dencher et al., 1989;Subramaniam et al., 1993,1999; Han et al., 1994;Sass et al., 1997; Kamikubo et al., 1996, 1997;Vonck, 1996; Oka et al., 1997, 1999; Hendricksonet al., 1998), as well as environmental changes ofspin-labels (Steinhoff et al., 1994; Pfeiffer et al.,1999), had indicated large-scale conformationalchanges at the cytoplasmic ends of helices F and Gand, to a lesser extent, at helices B and C. Thechange at helix F appeared to be an outward-directed rigid-body tilt, described also by a large(about 5 AÊ ) increase of the distances between twopairs of strategically placed spin-labels in N(Thorgeirsson et al., 1997). In view of the lack ofany obvious means by which a proton can betransferred, directly or indirectly, across the ratherapolar cytoplasmic region between Asp96 and theSchiff base (Grigorieff et al., 1996), and the sugges-

tive effects of increased osmotic pressure (Cao et al.,1991) and hydrostatic pressure (VaÂroÂ & Lanyi,1995) on the reprotonation of the Schiff base,it seemed that the functional rationale for theconformational changes in the cytoplasmic regionmay be to allow the in¯ux of water. However,a projection map from neutron diffraction ruledout a net increase of hydration in this region in theM state that accumulates in D96N (Weik et al.,1998).

In the N state, Asp85 is still protonated andAsp96 is deprotonated, and the retinal-bindingsite is now relaxed (Fodor et al., 1988a; Ames &Mathies 1990; PfefferleÂ et al., 1991; Sasaki et al.,1994), accommodating, preferentially (Dioumaevet al., 1998a), the 13-cis,15-anti con®guration.Because in N the Schiff base is reprotonated, thebarrier to double-bond rotation is lowered(Tavan et al., 1985), and thermal reisomerizationto all-trans is made possible. Decay of the Nstate to produce the O state includes both repro-tonation of Asp96 and reisomerization of theretinal. The driving force for these transitions isnot clear, but might be the last, unidirectionalstep of the photocycle, in which the initiallyvery low pKa of Asp85 is re-established, as thisresidue will deprotonate in a strongly uni-directional reaction to reprotonate the extracellu-lar proton release site (Richter et al., 1996b;Zscherp & Heberle 1997; Balashov et al., 1999;Zscherp et al., 1999; Li et al., 2000). When a pro-ton is not released from this site earlier in thecycle, as at pH below the pKa for release in thewild-type (ZimaÂnyi et al., 1992b), and in mutantsblocked at this step, e.g. in E204Q (Brown et al.,1995a), the proton is released from Asp85 to thesurface at the end of the cycle.

The recent determinations of the crystallographicstructure of the BR state to increasingly better res-olutions (Luecke et al., 1998; Essen et al., 1998;Belrhali et al., 1999; Mitsuoka et al., 1999), mostrecently to 1.55 AÊ (Luecke et al., 1999a), have sup-plied many of the missing speci®c moleculardetails for this picture. The structural modelsdescribe an extended three-dimensional hydrogen-bonded network of polar side-chains and orderedwater molecules of obvious functional signi®cancein the extracellular region of this seven-transmem-brane-helical protein, provide the rationale for thestability of the protonated Schiff base in theimmediate vicinity of the anionic Asp85 andAsp212, and con®rm the notable lack of such ahydrogen-bonded chain in the cytoplasmic region.Even more signi®cant is that the structures of twoof the intermediate states of the photocycle, K andM, have been now determined, to 2.1 and 2.0 AÊ

resolutions, respectively. The former was producedin a photostationary state mixture by illuminationof the wild-type protein at 110 K (Edman et al.,1999), the latter by conversion of the D96N mutantto the M state upon illumination at room tempera-ture and rapid cooling to 100 K (Luecke et al.,1999b). These structures yielded additional

Directional Transport in Bacteriorhodopsin 1239

insights, particularly into the coupling of eventsnear the retinal to the release of a proton to theextracellular surface.

Although a satisfactory understanding of themechanism of the transport will require that thestructures of the other intermediate states be deter-mined, and at such resolution and accuracy as toreveal all relevant molecular details, it is clearalready that the crystallographic approach haschanged the way one ought to regard bacteriorho-dopsin. Many of the outstanding questions in thephotocycle mechanism should be now framed inexplicitly structural terms, and the structuraldescriptions will require revisiting, and in somecases more explicitly interpreting the spectroscopicresults. Some of these questions are the following:what is the exact geometry of the distortion in theretinal in the K (Braiman & Mathies 1982; Siebert& Mantele 1983; Rothschild et al., 1984), L (Fodoret al., 1988b; Diller & Stockburger 1988; PfefferleÂet al., 1991), and later in the O state (Smith et al.,1983; Kandori et al., 1997), suggested by FTIR andRaman data? Is the rationale for these distortionsthe conservation of free energy for transport, or dothey mediate accessibility of the Schiff base in theextracellular and cytoplasmic directions (or both)?How does the relaxation of the con¯icts betweenthe photoisomerized retinal and the protein deter-mine the direction of the sequential ion transfers inthe transport cycle? Is a proton transferred directlyfrom the Schiff base to Asp85, as generallyassumed from the kinetic correlation between thedeprotonation of the Schiff base and the protona-tion of Asp85, or is this a more complex reactionwith possibly the participation of a hydroxyl ion(Luecke, 2000)? What is the nature and functionalrationale of the rigid-body movements of thehelices in the cytoplasmic region? How are thesemovements induced by molecular events in theearlier steps in the photocycle? Does water partici-pate in lowering the pKa of Asp96 and in transfer-ring a proton from Asp96, or the cytoplasmicsurface, to the Schiff base? Is there an increase ofhydration of the cytoplasmic region as the N stateis formed? Or alternatively, is there another pro-cess that couples these deprotonation and protona-tion reactions? In the last photocycle reaction, whatis the pathway of the proton from the dissociationof Asp85 from this buried residue to the protonrelease site near the surface?

The results we report here trace the origin of thecytoplasmic to extracellular direction of ion trans-location to the various con®gurational changes ofthe photoisomerized retinal during the photocycle,which affect the extracellular and cytoplasmicregions differently. Our approach to this questionwas to determine the structure of the M state pro-duced in a photostationary state of the E204Qmutant, because this mutation blocks the protonrelease step (Brown et al., 1995a), and we expectedthat the M state that accumulates will retain someof the characteristics of M2, or at least an earlier Mstate than the last one before N is formed. For this

reason, and to distinguish them from the inter-mediates de®ned by kinetic and spectroscopic cri-teria (ZimaÂnyi et al., 1992b; Brown et al., 1994a;Dickopf & Heyn 1997), we regard the M in thephotostationary state of E204Q as an ``early'' M,and that of D96N (Luecke et al., 1999b) as a ``late''M. Indeed, many of the structural changes weobserve in the putative early M are similar to, butof lesser magnitude than those in the previouslydetermined late M. In a satisfying way, the confor-mational changes in the extracellular and cyto-plasmic region correlate with the various changesof the retinal, and provide clues to the causes ofdirectionality in the transport cycle.

Results

Photoproduction and stabilization of an Mstate in the E204Q mutant

For wild-type bacteriorhodopsin, the photocyclereaction of crystals grown in cubic lipid phase hadbeen found to resemble that in purple membranesuspensions (Heberle et al., 1998). This is not sur-prising, because the p3 lattice of the membranesheets has the same packing arrangement and lat-tice constants as in the x-y crystallographic planeof the P63 space group of the three-dimensionalcrystals. In the E204Q mutant, the decay of the Ostate is considerably slowed (Brown et al., 1995a),and in membrane suspensions the photostationarystate produced by yellow light at ambient tempera-ture contains primarily the O intermediate. Thiswas not found to be the case for the three-dimen-sional crystals of the E204Q mutant, however. Asshown in Figure 1(a), in the crystals the BR state isconverted to an M state, as indicated by theappearance of a new absorption maximum atabout 410 nm, and the disappearance of absorptionabove 500 nm, comparable to the reference spec-trum for the M state in purple membrane suspen-sions (in Figure 1(b), data from Gergely et al.,1997). The amplitude of residual absorbance at570 nm indicates that less than 7 % of the BR statecould have remained in the photostationary stateof the illuminated crystals. In contrast, the O statewould have been detected as a high-amplitudered-shifted band with a maximum at about 630 nm(Gergely et al., 1997). The accumulation of M in theE204Q mutant is consistent with kinetic measure-ments of the photocycle in the crystals (notshown), which indicate the existence of a long-lived M state with a similar relaxation time-con-stant as in D96N, i.e. a slower M decay than in thewild-type. As in the case of the D96N mutant(Luecke et al., 1999b), the conversion of the BR tothe M state was completely reversed within a fewseconds at ambient temperature in the dark, andthe interconversions could be repeated manytimes.

Figure 1. Identi®cation of the photoproduct in crystalscontaining E204Q bacteriorhodopsin. (a) Spectra for acrystal, illuminated and unilluminated but light-adapted. Illumination with a yellow laser was for onesecond at ambient temperature, followed by rapid cool-ing to 100 K by a cryo-stream of nitrogen gas, anddetermination of the spectra. (b) Spectra of the BR andM states in wild-type purple membrane suspension, atroom temperature, from Gergely et al. (1997).


Determination of the structures of the BR andM states

We have solved the X-ray structure of the E204Qmutant in the BR state (our unpublished results) at1.7 AÊ resolution. In this mutant, the structuralrearrangements upon the conservative replacementof protonated glutamic acid with glutamine aremore extensive than in the case of the D96Nmutation (Luecke et al., 1999b). The changes arecon®ned to the extracellular side, and include largeside-chain displacements of Glu194 and Gln204,insertion of an additional water molecule betweenArg82 and Glu204, which results in generallyimproved order in this region, and less dramaticchanges of the Arg82 side-chain and Wat406. Nochange was observed in the region near the retinalor on the cytoplasmic side. The details will bedescribed elsewhere.

Now, we have determined the structure of theM state obtained by illumination of the E204Qmutant with yellow light at room temperature (as

in Figure 1(a)), followed by rapid cooling to 100 K,at 1.8 AÊ resolution. This M state structure is con-siderably better ordered than that previouslyreported for an M state produced in crystals of theD96N mutant (Luecke et al., 1999b). The increase inresolution from 2.0 AÊ for the late M structure to1.8 AÊ for the early M intermediate is accompaniedby dramatically improved electron density maps(Figures 2, 4 and 6), in which new water moleculescould be localized, and the crystallographic R-fac-tors are much better (Table 1). In addition, thecytoplasmic ends of helices F and G, which werehighly disordered in the M state of the D96Nmutant, are now well ordered.

Structural changes of the retinal and itsimmediate environment

Figure 2(a) and (b) show density maps of theregion of the retinal, Asp85, and other neighboringresidues in the BR and M states of the E204Qmutant. The rotation of the C131C14 double-bond,and lack of rotation of the C151N double-bondand C14ÐC15 single bond are both evident in themap for M (Figure 2(b)), as expected for the13-cis,15-anti con®guration. Also to be noted arethe changed hydrogen-bonding of water molecule501, the absence of water molecule 402 and theappearance of water molecule 503 in the M state.Figure 3 shows the details of these and otherchanges (the structure shown in red, for the M ofD96N, will be discussed below). The changed con-tour of the chain of the isomerized retinal causesupward buckling and moves C13 in the cyto-plasmic direction along the z-axis by 0.6 AÊ . TheSchiff base nitrogen atom is displaced by 0.7 AÊ ,and CE of the connected Lys216 chain by 1.3 AÊ .The retinal chain in the BR state is nearly comple-tely planar, but in the M state exhibits a small butdistinct out-of-plane curvature between C11 andC15 (not clearly seen from the viewing angle inFigure 3). The CB-CG bond of Asp85 has rotated(�w2 � � 48 �) and thus OD2 and OD1 havemoved relative to both the Schiff base NZ andOG1 of Thr89, to which Asp85 is hydrogen-bondedin the BR state. OD1, in particular, which is hydro-gen-bonded to Wat402 (and Wat401, not shown) inthe BR state, is displaced by 1.5 AÊ , and no longerparticipates in any hydrogen bonds. Asp212 (notshown) has changed much less, with OD1 andOD2 moving about 0.5 AÊ , although the hydrogen-bonding of these oxygen atoms is altered. OD1 ofAsp212 that lost its hydrogen bond to Wat402 isnow hydrogen-bonded to Wat401, in addition tothe OH of Tyr57. OD2 gained a hydrogen bondfrom NE1 of Trp86, in addition to its hydrogenbond with the OH of Tyr185. Movement of theside-chain of Lys216 that follows the con®gura-tional change of the retinal is accompanied bylarge displacement of its peptide atoms, asdescribed in more detail below.

The electron pair of the retinylidene nitrogenatom of this M state points no longer in the extra-

Figure 2. 2Fo ÿ Fc density mapsof BR and M states of E204Q, con-toured at 1s, of the region of theretinal, Lys216, Asp85, Trp182, andwater molecules 401, 402, 501, 502,and 503. (a) BR state; (b) M state.


cellular direction, as the N-H bond in the BR state(Belrhali et al., 1999; Luecke et al., 1999a), but it isnot turned as completely to the cytoplasmic direc-tion as in the M of D96N (Figure 3; Luecke et al.,1999b). It is oriented toward Thr89, and the Schiffbase NZ is now 3.0 AÊ from OG1 of Thr89. Whilethis putative hydrogen-bond donor is out of theplane of C151NZ-CE by about 60 �, a hydrogen-bond with the Schiff base nitrogen atom seemslikely to form for lack of another acceptor of theThr89 hydroxyl proton.

Structural changes in the extracellular region

Figure 4 shows a density map for the extra-cellular region in the M state of E204Q. As in theM state of D96N (Luecke et al., 1999b), a large dis-

placement moves the side-chain of Arg82 fromnear Asp85 and Asp212 toward Glu194, and theNH1 forms a hydrogen bond with the glutamateOE1. Figure 5(a) compares the models for the BRand M states of the E204Q mutant, and inFigure 5(b) the same comparison is made for theBR and M states of the D96N mutant (Luecke et al,1999b). As will be reported elsewhere in moredetail, the E204Q residue replacement causes con-siderable rearrangement of the water moleculesnear Gln204 in the BR state. The changes of aminoacid residues in the M state are similar in the twomutants, however. The largest movement in bothmutants is of the side-chain of Arg82, toward theextracellular surface, CZ being displaced by 2.3 AÊ

in E204Q and 1.7 AÊ in D96N. This disrupts thehydrogen bonds of NH1 and NH2 to Wat407, and

Table 1. X-ray data collection and re®nement statistics

BR state M stateA. Data reduction resolution range (AÊ ) 1.7-25.0 1.70-1.73 1.8-25 1.80-1.83

Total observations 257,539 214,773Unique structure factors 25,709 21,451Mosaicity (�) 0.42 0.33Average I/s(I) 29.6 2.9 28.2 3.3Completeness (%) 99.1 86.4 99.5 95.1Rmerge (%) 5.7 47.5 8.2 44.6

B. Refinement resolution range (AÊ ) 1.7-12.0 1.8-12.0

Number of structure factors 24,676 20,778Number of restraints 8217 8198Number of parameters 8300 8252Twin ratio 50:50 58:42Number of protein atoms 1747 1747Number of retinal atoms 20 20Number of water molecules 54 45Number of lipid atoms 310 310R-factor (%) for data F > 4s(F)/all data 12.0/13.2 13.1/14.2Rfree (%) for data F > 4s(F)/all data 17.7/19.1 19.0/20.3Average protein B (AÊ 2) 26.2 25.9Average retinal B (AÊ 2) 17.5 18.8Average water B (AÊ 2) 39.9 43.4Average lipid B (AÊ 2) 52.1 53.7Deviation from ideal bond lengths (AÊ ) 0.009 0.009Deviation from ideal bond angles (deg.) 0.026 0.024

Rmerge (I) � �hkl �i j Ihkl,i ÿ hIhkli j/�hkl �i jIhkl,ij, where hIhkli is the average intensity of the multiple Ihkl,i observations for symme-try-related re¯ections.

I/s(I), average of the diffraction intensities, divided by their standard deviations.R-factor � �hklj Fo ÿ Fc j/�hkljFoj, where Fo and Fc are observed and calculated structure factors, respectively.Rfree � �hkl e Tj Fo ÿ Fc j/�hkl e Tj Foj, where the test set (5 % of the data) is omitted from the re®nement in such a way that all

structure factors in each of several thin-resolution shells were selected to avoid bias due to the presence of merohedral twinning.


in E204Q (Figure 5(a)) NH1 forms a hydrogenbond with Glu194. This bond, not evident in the Mstate of D96N (Figure 5(b)), perhaps becauseGlu194 is displaced away from, rather thantoward, Arg82, might account for the much greatermovement of Arg82 in the M state of E204Q.Although the structure of this region is affected bythe E204Q mutation itself, the changed electrostaticenvironment of the Arg82 guanidinium group inthe M state is consistent with the notable asymme-try between these nitrogen atoms detected in theM of the wild-type protein by solid-state NMR(Petkova et al., 1999). It is noteworthy that theobserved NMR chemical shifts strongly argueagainst the possibility that Arg82 deprotonates atthis time in the photocycle.

Structural changes in the cytoplasmic region

Figure 6(a) and (b) show density maps of thecytoplasmic region for the BR and M states ofE204Q. While in the BR state the side-chains ofAsp96 and Thr46 are hydrogen-bonded to oneanother (Figure 6(a)), as in the wild-type (Lueckeet al., 1999a), in the M state a new water molecule,504, is intercalated between them (Figure 6(b)).Wat504 is hydrogen-bonded, in turn, to anothernew water molecule, 503, which is hydrogen-bonded also to Wat502, which is present in boththe M and the BR state.

The movements of the retinal induce thesechanges in the region of Asp96 by two differentmeans, as described in the following. Figure 7(a)shows a chain of covalent and hydrogen bondsthat extends in the BR state from the peptide O ofAla215 (hydrogen-bonded to Wat501), where thelocal interruption of the hydrogen-bonding patternof the a-helix was termed a p-bulge (Luecke et al.,1999a), continues with the hydrogen bond betweenthe peptide O of Lys216 and Wat502 to the peptideO of Thr46, and connects ®nally with Asp96through a hydrogen bond between OG1 of Thr46and the carboxyl OH of Asp96. The isomerizationof the retinal causes considerable local displace-ments at the p-bulge in helix G via the side-chainof Lys216 that connects them. In particular, CAand the O of Lys216 move by 0.9 and 2.1 AÊ ,respectively (resulting in a �c of �48 � for themain chain), and the O of Ala215 moves by 1.0 AÊ .As a result of this, and because of the 0.6 AÊ displa-cement of the indole ring of Trp182 by the upwardtravel of the 13-methyl group of the retinal (thatmoves Wat501 hydrogen-bonded to it from a dis-tance of 3.0 AÊ to the peptide O of Ala215 to 5.0 AÊ ),the hydrogen bond between the Ala215 O andWat501 is broken. Wat501, still connected to NE1of Trp182, is now hydrogen-bonded instead toOG1 of Thr178. On the other hand, Wat502remains hydrogen-bonded to the peptide O ofLys216, but it has moved by 1.3 AÊ , and acquired a

Figure 3. Models of BR state (carbon atoms colored beige, from E204Q), earlier M state (from E204Q, carbon atomscolored blue) and late M state (from D96N, carbon atoms colored red) of the region of the retinal, the side-chain ofLys216, Asp85, Thr89 and Wat402, shown in stereo view. The hydrogen bonds in the BR state are shown in gold, inthe M state (in this case only the earlier M from E204Q) in green.


new hydrogen bond to a new water molecule, 503,that is not evident in the BR state. The displace-ment of Wat502 in turn moves the peptide O ofThr46, and thereby its OG1. Breaking the connec-tion between Trp182 and helix G thus causes acascade of movements along the chain, and movesOG1 of Thr46 away from Asp96. The importanceof the contact between the 13-methyl group of theretinal and its counterpart in the binding site (nowidenti®ed as the indole ring of Trp182) is indicatedby the observation that in bacteriorhodopsin con-taining 13-desmethyl retinal the chromophore is in

Figure 4. Density map of the M state of E204Q, as in Figur

a complex mixture of various isomeric states, withall-trans only a small minority (GaÈrtner et al., 1983).

Figure 8(a) shows additional, coordinated large-scale side-chain movements that occur in the Mstate of E204Q, evidently as the consequence of thedisplacements of the indole ring of Trp182 andhelix G. The side-chain of Leu181 moves into theregion vacated by Ala215, with its CD1 and CD2moving as much as 1.7 and 2.4 AÊ , respectively,toward the retinal. At the same time, on the cyto-plasmic side of Trp182, OG1 of Thr178, and thephenyl ring of Phe219 move away from helices B

e 2, of the region of Arg82, Glu194 and Gln204.

Figure 5. Models of BR, earlier M (from E204Q) and late M (from D96N) of the region that includes Arg82, Glu204,Glu194, and several water molecules, shown in stereo view. Color code for the BR, and the earlier and late M states,as in Figure 3. (a) Comparison of the BR and M states of the E204Q mutant. (b) Comparison of the BR and M statesof the D96N mutant.


and C, by 0.4 and 0.9 AÊ , respectively. These move-ments are evident also, but with increased ampli-tudes, in the M state of D96N (Figure 8(b)). Thisconformational change constitutes the repackingof this region in response to its perturbation bythe upward buckling of the retinal at C13. It isimportant that the side-chain of Asp96 followsthese side-chain displacements (Figure 8(a)), andits carboxyl group is displaced toward helix F,OD1 and OD2 by 0.4 and 1.7 AÊ , respectively. Asdescribed below, another result of the side-chainmotions is the outward tilt of helix F from aboutresidue 179 up to the cytoplasmic surface.

The two kinds of structural changes describedabove move Thr46 and Asp96 apart, increasing thedistance of OG1 of Thr46 to the carboxyl OH ofAsp96 from 2.4 AÊ to 3.6 AÊ . A new water molecule,504, is intercalated between them, hydrogen-bonded to both as well as to Wat503 (Figures 6(a)and 7(a)). As discussed below, the appearance ofWat504 will have a profound effect on the protonaf®nity of Asp96.

Figure 7(a) shows also that in the M state theregion of Asp96 is connected to the main-chain of

helix G through hydrogen bonds formed by watermolecules 504, 503, and 502 to each other and tothe C1O of Lys216 and Ala215. There is now anetwork of ordered water molecules in the cyto-plasmic region that extends from Asp96 to within7.4 AÊ of the Schiff base. A third view of this region,from the cytoplasmic surface, is given in Figure 9.It illustrates the hydrophobic side-chains in thechannel that connects Asp96 and the retinal Schiffbase, and the partial network of water moleculesthat forms in the M state of E204Q. If a hydrogen-bonded chain between Asp96 and the Schiff base isto form, for facilitating proton transfer betweenthem during the rise of the N state, at least twomore water molecules will be required to completeit. This will be then the rate-limiting step for thereprotonation of the Schiff base.

Movements of helices

Figure 10 shows that the cytoplasmic ends ofhelices F and G are displaced in the M state,but helix E (as the other four helices, notshown) exhibit virtually no movement of this

Figure 6. Density maps of the (a) BR state and (b) M state of E204Q, as in Figure 2, but of the region of Asp96 andThr46.


kind. Although the atom-to-atom movementsare small (ca. 0.7 AÊ at the cytoplasmic end ofhelix F), it is clearly discernible that the changesare restricted to the cytoplasmic region, andbegin with residue 178 for helix F and residue223 for helix G. Because the better de®ned main-chain atom displacements of helix F are all inthe same direction, i.e. away from the center ofthe protein and toward helix E, and become lar-ger as the cytoplasmic surface is approached,they describe the rigid-body motion that corre-sponds to the outward tilt of this helix proposedearlier from projection maps. The location of thistilt suggests that it originates from the repackingof side-chains (Figure 8(a)) that results inincreased separation between helices G and F.This would explain the disorder at the cyto-plasmic end of helix F in the M state of D96N(Luecke et al., 1999b) as a further developmentof the outward tilt, the result of the much great-er movements of the side-chains in this state(Figure 8(b)).

Local displacements of main-chain atomswithin the helices are observed at the p-bulge ofhelix G (residues 215-216, shown in Figure 10),and near Arg82 and Asp96 of helix C (notshown).

Discussion

Kinetic measurements of the bacteriorhodopsinphotocycle in membrane suspensions had indi-cated that in the absence of proton release upon

protonation of Asp85, at pH < pKa for release(ZimaÂnyi et al., 1992b), as well as in the E204Q,E194Q, and R82A mutants (Balashov et al., 1993,1997; Brown et al., 1995a; Dioumaev et al., 1998b)the photocycle is not stalled at the L$M1$M2

equilibrium as expected for the simple linearmodel of L$M1$M2$M2

0. Instead, as in thewild-type protein at higher pH, the Schiff basebecomes nearly fully deprotonated, the result of astill poorly understood reaction of the photocycle.In some mutants the L$M1$M2 equilibrium isdramatically shifted in favor of L, and the transientaccumulation of M remains low even after theM2!M2

0 reaction at high pH (Brown et al.,1994a), but in others, like E204Q, the amount of Mdetected in the photocycle is as high as in the wild-type (Brown et al., 1995a).

Nevertheless, we assume that the M state ofE204Q retains some features of the M2 or an earlierM state of the wild-type photocycle. The M statefrom D96N, in turn, has been proposed to rep-resent the last M state, perhaps the MN state, pro-duced before the formation of N (Sasaki et al.,1992; Kamikubo et al., 1996, 1997). Based on theseassumptions, we reconstruct the crucial molecularevents in the cycle. In this analysis we attemptto distinguish the mechanistically meaningfuldifferences between the two M states from thosedifferences that re¯ect the different BR states thatgave rise to them. It is signi®cant that at the retinalthe differences between the BR states of the wild-type, D96N, and E204Q mutant are negligible, butnear Asp96 and particularly in the region ofGlu204, distinct structural differences that originate

Figure 7. Models of BR, earlier M (from E204Q) and late M (from D96N) of the region that includes the retinal,Trp182, Thr178, Lys216, Ala215, Thr46, Asp96, and water molecules 501, 502, 503 and 504, shown in stereo view.Color code for the BR, and the earlier and late M states, as in Figure 3. (a) Comparison of the BR and M states of theE204Q mutant. (b) Comparison of the BR and M states of the D96N mutant.


from the D96N (Luecke et al., 1999b) and E204Qmutations, respectively, were noted, and will beconsidered below.

As shown in Figures 2 and 3, there are twodistinct kinds of con¯icts that arise from thephotoisomerization of the retinal.

Electrostatic conflict at the Schiff base withAsp85 and Asp212

In the M state from E204Q, the Schiff basemoves 0.7 AÊ in the extracellular direction, and theplane of C151NZ-CE turns toward the cytoplasmicdirection, but not fully. In the M state from D96N,the Schiff base is displaced somewhat further, by0.9 AÊ , in the same direction, and the plane ofC151NZ-CE comes closer to being parallel withthe z-axis, the retinylidene nitrogen atom assumingthe position expected for 13-cis,15-anti from simplemodel building. This is obviously a more relaxed

con®guration than the earlier M. However, thenumerous hydrogen bonds in the BR state thatinvolve also water molecules 401 and 402 arebroken already in the former M state.

Steric conflict of the 13-methyl groupwith Trp182

The retinal chain undergoes progressive changesbetween the two M states we describe. Becausemovements of the b-ionone ring and the 13-methylgroup are hindered by protein residues, in the Mfrom E204Q the retinal remains extended in spiteof the intrinsically bent shape of the 13-cis,15-anticon®guration, and accommodates the lesser bendpartly by increasing the angle tended byC12-C131C14 (124 � versus 114 � in the BR state). Inthe small-molecule retinal crystal structure, andfrom quantum chemical calculations, this angle is117-118 � for all-trans retinal (Tavan et al., 1985).

Figure 8. Models of BR, earlier M (from E204Q) and late M (from D96N) of the region that includes the retinal,Trp182, Thr178, Lys216, Ala215, Leu181, Phe219 and Asp96, shown in stereo view. Color code as in Figure 3. (a)Comparison of the BR and M states of the E204Q mutant. (b) Comparison of the BR and M states of the D96Nmutant.


Another part of the strain is taken up by move-ments of the Lys216 chain, consistently with pre-dictions from the FTIR spectra of the M state insamples with isotopically labeled lysine (Takeiet al., 1994). In the M state from D96N, the in-planeupward bulge of the chain at C13, that begins inthe earlier M, becomes greater (Figure 3), and theC12-C131C14 angle decreases to 114 �, the value inthe BR state.

It seems clear that the ®rst increase in the pKa ofAsp85 that leads to the L$M1 equilibrium mustoriginate from the disruption of the hydrogenbonds that the anionic Asp85 received fromWat402 and Thr89, and the loss of several watermolecules (Luecke et al., 1999b). The orientation ofthe Schiff base away from Asp85 in the two Mstates points to one of the still unsolved questionsin the bacteriorhodopsin photocycle: how does theSchiff base become deprotonated and Asp85 proto-nated in this reaction? If the retinal is thought to

assume increasingly relaxed states in successivephotocycle reactions, one might expect that in ear-lier states, i.e. in M1 and L, it could have a moredrastically strained con®guration than in the M inFigure 3, with the Schiff base N-H bond pointingtoward Asp85. If this were the case, a direct hydro-gen bond between the Schiff base and Asp85, inthe absence of Wat402 could facilitate direct protonexchange between them. There is some experimen-tal support for such a mechanism. From theincreased frequency of the coupled C1N stretchand N-H bend mode, it had been suggested that inthe L state the Schiff base moves closer to its coun-ter-ion, and enters into a stronger interaction withit than in the BR state (Smith et al., 1984; Alshuth &Stockburger 1986). Also, it is known that theabsorption maximum of M

02 in the D96N mutant is

shifted from the 412 nm position of M1 to 407 nm(ZimaÂnyi et al., 1992a; Radionov et al., 1996), and asimilar shift is observed in wild-type monomeric

Figure 9. View of the structural model of the M state of E204Q from the cytoplasmic side downward, shown instereo view. The retinal, Trp182, Thr178, Lys216, Leu181, Phe219, Asp96, Thr46, and the four water molecules shownin Figure 7(a) are included to illustrate the formation of a partial aqueous network between Asp96 and the Schiffbase.


bacteriorhodopsin (VaÂroÂ & Lanyi 1991b; Milderet al., 1991). The direction of this shift is toward theabsorption maximum of retinal with unprotonatedSchiff base, when in solution unaffected by chargesor dipoles (about 390 nm). The direction and the

Figure 10. Main-chain atoms of helices E, F, and G, in theColor code, the BR state is colored according to atom type, tG; the arrows represent the alternating N! C directions of twith the views in Figures 7 and 8.

extent of the shift would be consistent thereforewith a weakening of the electrostatic interactionwith the dipole of Asp85 in the M1!M2!M2

0reaction. However, in this context it is puzzlingthat in the model for the K intermediate, that

BR and M states of the E204Q mutant, shown in stereo.he M state in blue. The helices are identi®ed as E, F andhe peptide chains. The retinal is included for comparison


arises much earlier than M in the photocycle,the N-H bond of the Schiff base is orientedalready fully to the cytoplasmic direction(Edman et al., 1999).

Alternatively, the highly polarized Wat402 mightmediate an indirect proton exchange between theSchiff base and Asp85. If in the L state the Schiffbase has the similar con®guration as we ®nd in M(Figures 2 and 3), and Wat402 is still near Asp85and Asp212, this water molecule might dissociate toprotonate Asp85, and the hydroxyl ion producedcould then move to the cytoplasmic side of theSchiff base and receive its proton (Luecke, 2000).This would have to be a concerted process, becausedeprotonation of the Schiff base and protonation ofAsp85 constitute a single kinetic step. Once Wat402is re-formed it would presumably diffuse awayfrom the hydrophobic region on the cytoplasmicside of the Schiff base. This scheme explains iontransfer in the same terms as chloride ion transfer inthe related protein, halorhodopsin (reviewed byOesterhelt, 1995, 1998), and the D85T mutant ofbacteriorhodopsin (Sasaki et al., 1994), where Clÿ

(instead of OHÿ) must move from the extracellularside of the protonated Schiff base to the cytoplasmicside. The equivalence of the protonation of Asp85and the transfer of a Clÿ from the extracellular tothe cytoplasmic side is indicated by the observationthat they both (in wild-type and D85T bacteriorho-dopsins, respectively) result in release of a proton tothe extracellular surface, that is abolished by theE204Q mutation (Brown et al., 1996). We note, how-ever, that this mechanism is also in con¯ict with thereported structure of the K state, because Wat402 isabsent near its location in the BR state (Edman et al.,1999).

The origin of the proton that is released to theextracellular surface upon protonation of Asp85 isnot clear, but the observed displacement of thepositively charged side-chain of Arg82 toward theextracellular side in the M state of D96N (Lueckeet al., 1999b) made it the likely candidate as theagent of the coupling between protonation ofAsp85 and deprotonation of the release site. Thus,when the positive charge of Arg82 moves awayfrom Asp85 and toward the proton release site, thepKa of the former is increased and that of the latteris decreased. This simple model explains why theproton release site deprotonates upon protonationof Asp85. It now appears that the changes of watermolecules 401, 402, and 406 near Asp85, as well asthe displacement of Arg82, are complete, even inthe absence of proton release, when the Schiff basehas deprotonated. As shown in Figure 5, the dis-placement of the side-chain of Arg82 in the extra-cellular direction in the M state with proton releaseblocked (in E204Q) is even greater than in the Mstate after proton release (in D96N). This is consist-ent with the effects of the E204Q and E194Qmutations on the large, cooperative kinetic deuter-ium isotope effects on the protonation of Asp85 inthe photocycle, which are even greater in theE194Q and E204Q mutants than in the wild-type

(Brown et al., 2000), and suggested that the isotopeeffects originate from the rearrangement of watermolecules and side-chains upon protonation ofAsp85, instead of the conduction of a proton to thesurface. Importantly, the ®nding we report here,that Arg82 exhibits the same kind of motion inE204Q (where proton release is blocked) as inD96N (where proton release occurs), is strong evi-dence that the direct cause of this side-chain dis-placement is not the proton release but theprotonation of Asp85.

The C13 region of the retinal shows progressivemovement from the BR state to the M before pro-ton release (E204Q mutant), and the M state afterproton release (D96N mutant) (Figure 7(a) and (b)).In particular, in the former M, the 13-methyl groupis displaced by 0.6 AÊ in the cytoplasmic direction,but in the latter M state by 1.3 AÊ . Thus, the kinkdue to the rotation of the C131C14 double bond ismore fully accommodated by upward ``buckling''of the region of C13. The C10-NZ distance is 6.9 AÊ

in the BR state, decreases to 6.6 AÊ in the former M,and to 6.3 AÊ in the latter M. The upward move-ment of the 13-methyl group creates steric con¯ictwith the indole ring of Trp182, which is movedprogressively upward (Figure 7). It is interestingthat the N-H stretch band of the indole nitrogenatom of Trp182 exhibits greatly increased ampli-tude in the L state, but not in K or M, or the laterintermediates (Yamazaki et al., 1995). The reasonfor this amplitude increase might be distortion ofthe hydrogen bond of NE1 of Trp182 with Wat501in the L state when the 13-methyl group begins itsupward travel, because, unlike in M, Wat501 couldbe still hydrogen-bonded also to the peptidecarbonyl group of Ala215.

It is the movements of the 13-methyl groupand the p-bulge of helix G that cause the struc-tural changes in the cytoplasmic region in thesecond half of the photocycle. As shown inFigures 7 and 8, in the earlier M the displace-ment of the side-chain of Lys216 and the localregion of the p-bulge of helix G connected to it,and the upward displacement of the indole ringof Trp182 increases the distance between NE1 ofTrp182 and the peptide O of Ala215 from 5.4 to6.7 AÊ . In the late M state, Trp182 moves furtherand tilts away from the retinal, Wat501 is nolonger ordered (or present) in the structure, andthe distance between NE1 of Trp182 and thepeptide O of Ala has increased to 7.7 AÊ . HereNE1 of Trp182 is hydrogen-bonded directly toOG1 of Thr178. The resulting displacements ofthe main-chain of helix G and Wat502 that con-nect this region to OG1 of Thr46, together withthe movement of the side-chain of Asp96 uponthe coordinated repacking of the side-chains ofThr178, Leu181 and Phe219, interrupt the hydro-gen bond between Thr46 and Asp96 in E204Q,and between Thr46 and Asn96 via Wat504 inD96N. We conclude that whether residue 96 isan aspartate or an asparagine residue, the pro-gressively increasing motion of the 13-methyl


group of the retinal in the earlier and late Mstates, is transmitted to it.

In the E204Q mutant, which is indistinguishablefrom the wild-type in the cytoplasmic region, themovements of the side-chains of Thr46 and Asp96in the M intermediate allow the intercalation ofWat504 between OG1 of Thr46 and OD2 of Asp96.What is the rationale for this? At the time the Mintermediate is present in the photocycle, the pKa

of the Schiff base is 8.2, determined from the effectof pH on Schiff base protonation during the photo-cycle of the D96N mutant (Brown & Lanyi, 1996).From the [M]/[N] ratio in the M

02$ N equilibrium

in the wild-type photocycle (VaÂroÂ & Lanyi 1991a),it can be estimated that the pKa of Asp96 shouldbe about 0.5-1 pH unit lower than the pKa of theSchiff base (Brown & Lanyi, 1996). Because the pKa

of Asp96 in the BR state is above 11 (SzaÂraz et al.,1994), this means that it will have to be loweredconsiderably. Indeed, this pKa becomes as low as7-8 at this time in the cycle (Cao et al., 1993;Balashov et al., 1999; Zscherp et al., 1999; Li et al.,2000). FTIR spectra suggest that the pKa of Asp96is lowered already in M. There is a downshift ofthe frequency of the C1O stretch of the proto-nated carboxyl group of Asp96 in the M state,from 1742 to 1736 cmÿ1, a shift observed mosteasily in the D115N mutant, where the overlappingband of Asp115 is absent (Sasaki et al., 1994). Asexpected from the change of the hydrogen-bondingstatus of Asn96 in the M state of D96N (Figure 8),the C1O stretch of this amide also shifts infrequency (Sasaki et al., 1992).

The involvement of the region between Trp182and Thr46 in the reprotonation of Asp96 andthe reisomerization of the retinal in the photo-cycle is strongly supported by mutational evi-dence. Of the well over 100 single-site mutantsexamined (unpublished results), extreme slowingof the decay of the N state (by three orders ofmagnitude) is caused by replacements of onlysix residues, Trp182 (Weidlich et al., 1996), Thr46(Marti et al., 1991; Brown et al., 1994b), Phe171(Brown et al., 1995b), Phe42, Thr90 (Marti et al.,1991), and Leu93 (Subramaniam et al., 1991), allin the region between the 13-methyl group ofthe retinal and Asp96 (Luecke et al., 1999a). Thisis as expected if the thermal isomerization of theretinal from 13-cis, 15-anti to all-trans is depen-dent on protonation of Asp96, as reported(Dioumaev et al., 1998a), and during the secondhalf of the photocycle the M, N, and O statesare interconverted in equilibrating reactions. Thephenotype of the T46V mutant, in particular,con®rms that OG1 of Thr46 modulates the pKa

of Asp96. The more rapid reprotonation of theSchiff base and the much slower reprotonationof Asp96 in the photocycle of this mutant (Martiet al., 1991; Brown et al., 1994b) indicate thatAsp96 is easier to deprotonate in the M to Nreaction, and imply that its proton occupancyduring the decay of N is much less, i.e. its pKa

is lowered when threonine is replaced with

valine (Brown et al., 1994b). In the BR state ofthe wild-type, because the OG1 of Thr46 is ahydrogen-bond donor to the peptide O of Phe42,it must be the hydrogen-bond acceptor to theOD1 of Asp96, stabilizing the latter in the proto-nated state. Presumably, in T46V the polar car-boxyl group of Asp96 binds a water moleculeequivalent to Wat504, as in the BR state of theD96N mutant (Figures 7(b) and 8(b)), and stabil-ize the anionic state of the carboxyl group.

Asp96 will not deprotonate, even with a loweredpKa, unless the proton can move to its acceptor,the Schiff base. Most suggested models forthe reprotonation of the Schiff base require ahydrogen-bonded chain of water molecules in thecytoplasmic region that serve this purpose(Papadopoulos et al., 1990; Cao et al., 1991;Grigorieff et al., 1996). Very few water moleculesthat would participate in such a chain could beobserved in the M state from D96N where thisregion differs from the wild-type (Luecke et al.,1999b), but in the M state from E204Q there seemsto be the beginnings of a network that extendsfrom Asp96 toward, but not quite to, the Schiffbase (Figures 7, 8, and 9). The 7.4 AÊ gap that hasto be bridged near the Schiff base to complete achain requires the participation of more water inthis region during the M to N transition. Thus, ifwater facilitates the reprotonation of the Schiffbase, as suggested by indirect evidence (Cao et al.,1991; VaÂroÂ & Lanyi, 1995), its rearrangementswould have to be the rate-limiting step in pro-ducing the N state.

Conclusions

Comparison of the M states of the E204Q andD96N mutants reveals the different kinds of con-¯icts at the photoisomerized retinal, and how theirrelaxations initiate sequential conformationalchanges, ®rst in the extracellular and then in thecytoplasmic region. These structural changes deter-mine the direction of the transmembrane move-ments of the transported protons. Disruption ofhydrogen-bonding and charge displacements atthe deprotonated Schiff base initiate changes in theextracellular region that lead to release of a protonto the surface. Movement of the 13-methyl groupduring relaxation of the strained retinal chain, andthe connected Lys216 side-chain, then initiate pro-gressive changes in the cytoplasmic region, leadingto lowering of the pKa of Asp96 and thus reproto-nation of the Schiff base from this side via a transi-ent hydrogen-bonded network.

Materials and Methods

Data collection

Crystals of the E204Q mutant of bacteriorhodopsingrown in cubic lipid phase (Landau & Rosenbusch, 1996)are isomorphous to wild-type crystals and form thin hex-


agonal plates, typically about 60 mm � 60 mm � 15 mm,with space group P63 and strong merohedral twinning(Luecke et al., 1998). After mechanical extraction of thecrystals, the adhering cubic lipid phase was removed bysoaking for several hours in 0.1 % octylglucoside solutioncontaining 3 M sodium phosphate at pH 5.6. Diffractiondata were collected from cryo-cooled (100�K) crystals atbeamline 9.2 at the Stanford Synchrotron Radiation Lab-oratory (SSRL) using a 2 � 2 array CCD detector (ADSC,San Diego, CA). For the BR (ground) state structure, asingle crystal was light-adapted prior to cryo-cooling asbefore (Luecke et al., 1998,1999a,b). For the M state struc-ture, a light-adapted crystal was ®rst cryo-cooled, thenthe cryo-stream was blocked for one second while thecrystal was illuminated with a randomly polarized yel-low laser (5 mW, 594 nm). The crystals became colorless,indicating virtually complete conversion to an M statewith deprotonated Schiff base. The spectrum inFigure 1(a) was determined with a home-constructedinstrument based on an Ocean Optics Inc. (Dunedin, FL)model 2000 Fiber Optic Spectrometer (75 mm beam-diameter). Diffraction images were integrated, scaledand merged with DENZO/SCALEPACK (Otwinowski,1993).

Refinement

The starting model (1C3W, Luecke et al., 1999a)was re®ned against BR-state and M-state structure fac-tors with SHELXL-97 (Sheldrick & Schneider, 1997),taking merohedral twinning into account. Fo ÿ Fc

maps were employed in locating additional watermolecules. Omit maps and 3Fo ÿ 2Fc maps wereemployed extensively in re®tting the retinal, side-chainconformations, and locating water molecules. Due tothe quality of the maps, the angle and planarityrestraints around the Schiff base (C13 to CE) wererelaxed to minimize bias due to stereochemicalrestraints. The ®nal E204Q M state model includesresidues 5 to 156, 162 to 231, the covalently linked 13-cis retinal, 45 water molecule, and 14 lipid molecules.

Protein Data Bank access codes

The coordinates of the E204Q ground state structureand the E204Q M state structure have been deposited inthe RCSB Protein Data Bank with accession codes 1F50and 1F4Z, respectively.

Acknowledgments

We are grateful to the beamline staff at the StanfordSynchrotron Radiation Laboratory (SSRL, BL9-2), theEuropean Synchrotron Radiation Facility (ESRF, ID13)and SPring-8 (BL41XU) for their help, and to G. VaÂroÂ forgiving us the ®les for constructing Figure 1(b). This workwas supported, in part, by grants from the NIH to H.L.(R01-GM59970) and J.K.L. (R01-GM29498), and from theDOE to J.K.L. (DEFG03-86ER13525).

References

Alexiev, U., Marti, T., Heyn, M. P., Khorana, H. G. &Scherrer, P. (1994). Covalently bound pH-indicatordyes at selected extracellular or cytoplasmic sites inbacteriorhodopsin. 2. Rotational orientation ofhelices D and E and kinetic correlation between Mformation and proton release in bacteriorhodopsinmicelles. Biochemistry, 33, 13693-13699.

Alshuth, T. & Stockburger, M. (1986). Time-resolved res-onance Raman studies on the photochemical cycleof bacteriorhodopsin. Photochem. Photobiol. 43, 55-66.

Ames, J. B. & Mathies, R. A. (1990). The role of back-reactions and proton uptake during the N! Otransition in bacteriorhodopsin's photocycle: akinetic resonance Raman study. Biochemistry, 29,7181-7190.

Balashov, S. P., Govindjee, R., Kono, M., Imasheva, E.,Lukashev, E., Ebrey, T. G., Crouch, R. K., Menick,D. R. & Feng, Y. (1993). Effect of the arginine-82 toalanine mutation in bacteriorhodopsin on darkadaptation, proton release, and the photochemicalcycle. Biochemistry, 32, 10331-10343.

Balashov, S. P., Govindjee, R., Imasheva, E. S., Misra, S.,Ebrey, T. G., Feng, Y., Crouch, R. K. & Menick,D. R. (1995). The two pKas of aspartate-85 andcontrol of thermal isomerization and proton releasein the arginine-82 to lysine mutant of bacterio-rhodopsin. Biochemistry, 34, 8820-8834.

Balashov, S. P., Imasheva, E. S., Govindjee, R. & Ebrey,T. G. (1996). Titration of aspartate-85 in bacterior-hodopsin: what it says about chromophore isomeri-zation and proton release. Biophys. J. 70, 473-481.

Balashov, S. P., Imasheva, E. S., Ebrey, T. G., Chen, N.,Menick, D. R. & Crouch, R. K. (1997). Glutamate-194 to cysteine mutation inhibits fast light-inducedproton release in bacteriorhodopsin. Biochemistry,36, 8671-8676.

Balashov, S. P., Lu, M., Imasheva, E. S., Govindjee, R.,Ebrey, T. G., Othersen, B. I., Chen, Y., Crouch, R. K.& Menick, D. R. (1999). The proton release group ofbacteriorhodopsin controls the rate of the ®nal stepof its photocycle at low pH. Biochemistry, 38, 2026-2039.

Belrhali, H., Nollert, P., Royant, A., Menzel, C.,Rosenbusch, J., Landau, E. M. & Pebay-Peyroula, E.(1999). Protein, lipid and water organization in bac-teriorhodopsin crystals: a molecular view of thepurple membrane at 1.9 AÊ resolution. Structure, 7,909-917.

BouscheÂ, O., Braiman, M. S., He, Y.-W., Marti, T.,Khorana, H. G. & Rothschild, K. J. (1991).Vibrational spectroscopy of bacteriorhodopsinmutants. Evidence that Asp-96 deprotonates duringthe M! N transition. J. Biol. Chem. 266, 11063-11067.

Braiman, M. S. & Mathies, R. A. (1982). ResonanceRaman spectra of bacteriorhodopsin's primaryphotoproduct: evidence for a distorted 13-cis retinalchromophore. Proc. Natl Acad. Sci. USA, 79, 403-407.

Brown, L. S. & Lanyi, J. K. (1996). Determination of thetransiently lowered pKa of the retinal Schiff baseduring the photocycle of bacteriorhodopsin. Proc.Natl Acad. Sci. USA, 93, 1731-1734.

Brown, L. S., Gat, Y., Sheves, M., Yamazaki, Y., Maeda,A., Needleman, R. & Lanyi, J. K. (1994a). The retinalSchiff base - counterion complex of bacteriorhodop-sin: changed geometry during the photocycle is


a cause of proton transfer to aspartate 85.Biochemistry, 33, 12001-12011.

Brown, L. S., Yamazaki, Y., Maeda, M., Sun, L.,Needleman, R. & Lanyi, J. K. (1994b). The protontransfers in the cytoplasmic domain of bacteriorho-dopsin are facilitated by a cluster of interacting resi-dues. J. Mol. Biol. 239, 401-414.

Brown, L. S., Sasaki, J., Kandori, H., Maeda, A.,Needleman, R. & Lanyi, J. K. (1995a). Glutamic acid204 is the terminal proton release group at theextracellular surface of bacteriorhodopsin. J. Biol.Chem. 270, 27122-27126.

Brown, L. S., VaÂroÂ , G., Needleman, R. & Lanyi, J. K.(1995b). Functional signi®cance of a protein confor-mation change at the cytoplasmic end of helix Fduring the bacteriorhodopsin photocycle. Biophys. J.69, 2103-2111.

Brown, L. S., Needleman, R. & Lanyi, J. K. (1996). Inter-action of proton and chloride transfer pathways inrecombinant bacteriorhodopsin with chloride trans-port activity: implications for the chloride transloca-tion mechanism. Biochemistry, 35, 16048-16054.

Brown, L. S., Dioumaev, A. K., Needleman, R. & Lanyi,J. K. (1998). Local-access model for proton transferin bacteriorhodopsin. Biochemistry, 37, 3982-3993.

Brown, L. S., Needleman, R. & Lanyi, J. K. (2000). Ori-gins of deuterium isotope effects on proton transfersof the bacteriorhodopsin photocycle. Biochemistry,39, 938-945.

Cao, Y., VaÂroÂ , G., Chang, M., Ni, B., Needleman, R. &Lanyi, J. K. (1991). Water is required for protontransfer from aspartate 96 to the bacteriorhodopsinSchiff base. Biochemistry, 30, 10972-10979.

Cao, Y., VaÂroÂ , G., Klinger, A. L., Czajkowsky, D. M.,Braiman, M. S., Needleman, R. & Lanyi, J. K.(1993). Proton transfer from Asp96 to the bacterior-hodopsin Schiff base is caused by decrease of thepKa of Asp96 which follows a protein backboneconformation change. Biochemistry, 32, 1981-1990.

Cao, Y., Brown, L. S., Sasaki, J., Maeda, A., Needleman,R. & Lanyi, J. K. (1995). Relationship of protonrelease at the extracellular surface to deprotonationof the Schiff base in the bacteriorhodopsin photo-cycle. Biophys. J. 68, 1518-1530.

Dencher, N. A., Dresselhaus, D., Zaccai, G. & BuÈ ldt, G.(1989). Structural changes in bacteriorhodopsinduring proton translocation revealed by neutrondiffraction. Proc. Natl Acad. Sci. USA, 86, 7876-7879.

Dickopf, S. & Heyn, M. P. (1997). Evidence for the ®rstphase of the reprotonation switch of bacteriorho-dopsin from time-resolved photovoltage and ¯ashphotolysis experiments on the photoreversal of theM-intermediate. Biophys. J. 73, 3171-3181.

Diller, R. & Stockburger, M. (1988). Kinetic resonanceRaman studies reveal different conformationalstates of bacteriorhodopsin. Biochemistry, 27, 7641-7651.

Dioumaev, A. K., Brown, L. S., Needleman, R. & Lanyi,J. K. (1998a). Partitioning of free energy gainbetween the retinal and the protein in bacteriorho-dopsin. Biochemistry, 37, 9889-9893.

Dioumaev, A. K., Richter, H. T., Brown, L. S., Tanio, M.,Tuzi, S., SaitoÃ , H., Needleman, R. & Lanyi, J. K.(1998b). Existence of a proton transfer chain inbacteriorhodopsin: participation of Glu-194 in therelease of protons to the extracellular surface.Biochemistry, 37, 2496-2506.

Edman, K., Nollert, P., Royant, A., Belrhali, H., Pebay-Peyroula, E., Hajdu, J., Neutze, R. & Landau, E. M.

(1999). High-resolution X-ray structure of an earlyintermediate in the bacteriorhodopsin photocycle.Nature, 401, 822-826.

Essen, L. O., Siegert, R., Lehmann, W. D. & Oesterhelt,D. (1998). Lipid patches in membrane protein oligo-mers: crystal structure of the bacteriorhodopsin-lipid complex. Proc. Natl Acad. Sci. USA, 95, 11673-11678.

Fischer, W. B., Sonar, S., Marti, T., Khorana, H. G. &Rothschild, K. J. (1994). Detection of a watermolecule in the active site of bacteriorhodopsin-hydrogen bonding changes during the primaryphotoreaction. Biochemistry, 33, 12757-12762.

Fodor, S. P., Ames, J. B., Gebhard, R., van der Berg,E. M., Stoeckenius, W., Lugtenburg, J. & Mathies,R. A. (1988a). Chromophore structure in bacterior-hodopsin's N intermediate: implications for theproton pumping mechanism. Biochemistry, 27,7097-7101.

Fodor, S. P., Pollard, W. T., Gebhard, R., van den Berg,E. M., Lugtenburg, J. & Mathies, R. A. (1988b).Bacteriorhodopsin's L550 intermediate contains aC14-C15 s-trans-retinal chromophore. Proc. Natl Acad.Sci. USA, 85, 2156-2160.

GaÈrtner, W., Towner, P., Hopf, H. & Oesterhelt, D.(1983). Removal of methyl groups from retinal con-trols the activity of bacteriorhodopsin. Biochemistry,22, 2637-2644.

Gergely, C., ZimaÂnyi, L. & VaÂroÂ , G. (1997). Bacteriorho-dopsin intermediate spectra determined over awide pH range. J. Phys. Chem. B, 101, 9390-9395.

Gerwert, K., Hess, B., Soppa, J. & Oesterhelt, D. (1989).Role of aspartate-96 in proton translocation bybacteriorhodopsin. Proc. Natl Acad. Sci. USA, 86,4943-4947.

Grigorieff, N., Ceska, T. A., Downing, K. H., Baldwin,J. M. & Henderson, R. (1996). Electron-crystallo-graphic re®nement of the structure of bacteriorho-dopsin. J. Mol. Biol. 259, 393-421.

Han, B.-G., Vonck, J. & Glaeser, R. M. (1994). Thebacteriorhodopsin photocycle: Direct structuralstudy of two substates of the M-intermediate.Biophys. J. 67, 1179-1186.

Haupts, U., Tittor, J., Bamberg, E. & Oesterhelt, D.(1997). General concept for ion translocation byhalobacterial retinal proteins: the isomerization/switch/transfer (IST) model. Biochemistry, 36, 2-7.

Haupts, U., Tittor, J. & Oesterhelt, D. (1999). Closing inon bacteriorhodopsin: progress in understandingthe molecule. Annu. Rev. Biophys. Biomol. Struct. 28,367-399.

Heberle, J. & Dencher, N. A. (1992). Surface-bound opti-cal probes monitor proton translocation and surfacepotential changes during the bacteriorhodopsinphotocycle. Proc. Natl. Acad. Sci. USA, 89, 5996-6000.

Heberle, J., BuÈ ldt, G., Koglin, E., Rosenbusch, J. P. &Landau, E. M. (1998). Assessing the functionality ofa membrane protein in a three-dimensional crystal.J. Mol. Biol. 281, 587-592.

Hendrickson, F. M., Burkard, F. & Glaeser, R. M. (1998).Structural characterization of the L-to-M transitionof the bacteriorhodopsin photocycle. Biophys. J. 75,1446-1454.

Kamikubo, H., Kataoka, M., VaÂroÂ , G., Oka, T.,Tokunaga, F. & Needleman, R., et al. (1996). Struc-ture of the N intermediate of bacteriorhodopsinrevealed by x-ray diffraction. Proc. Natl Acad. Sci.USA, 93, 1386-1390.


Kamikubo, H., Oka, T., Imamoto, Y., Tokunaga, F.,Lanyi, J. K. & Kataoka, M. (1997). The last phase ofthe reprotonation switch in bacteriorhodopsin: thetransition between the M-type and the N-type pro-tein conformation depends on hydration. Biochemis-try, 36, 12282-12287.

Kandori, H., Yamazaki, Y., Hatanaka, M., Needleman,R., Brown, L. S., Richter, H. T., Lanyi, J. K. &Maeda, A. (1997). Time-resolved Fourier transforminfrared study of structural changes in the last stepsof the photocycles of Glu-204 and Leu-93 mutantsof bacteriorhodopsin. Biochemistry, 36, 5134-5141.

Landau, E. M. & Rosenbusch, J. P. (1996). Lipidic cubicphases: a novel concept for the crystallization ofmembrane proteins. Proc. Natl Acad. Sci. USA, 93,14532-14535.

Lanyi, J. K. (1997). Mechanism of ion transport acrossmembranes. Bacteriorhodopsin as a prototype forproton pumps. J. Biol. Chem. 272, 31209-31212.

Lanyi, J. K. (1998a). Understanding structure and func-tion in the light-driven proton pump bacterio-rhodopsin. J. Struct. Biol. 124, 164-178.

Lanyi, J. K. (1998b). The local-access mechanism ofproton transport by bacteriorhodopsin. Biochim.Biophys. Acta Bio-Energet. 1365, 17-22.

Lanyi, J. K. & VaÂroÂ , G. (1995). The photocycles ofbacteriorhodopsin. Isr. J. Chem. 35, 365-386.

Li, Q., Bressler, S., Ovrutsky, D., Ottolenghi, M.,Friedman, N. & Sheves, M. (2000). On the proteinresidues that control the yield and kinetics of O630

in the photocycle of bacteriorhodopsin. Biophys. J.78, 354-362.

Luecke, H. (2000). Atomic resolution structures of bac-teriorhodopsin photocycle intermediates: the role ofdiscrete water molecules in the function of thislight-driven ion pump. Biochim. Biophys. Acta. In thepress.

Luecke, H., Richter, H. T. & Lanyi, J. K. (1998). Protontransfer pathways in bacteriorhodopsin at 2.3 AÊ res-olution. Science, 280, 1934-1937.

Luecke, H., Schobert, B., Richter, H. T., Cartailler, J.-P. &Lanyi, J. K. (1999a). Structure of bacteriorhodopsinat 1.55 AÊ resolution. J. Mol. Biol. 291, 899-911.

Luecke, H., Schobert, B., Richter, H. T., Cartailler, J.-P. &Lanyi, J. K. (1999b). Structural changes in bacterior-hodopsin during ion transport at 2 AÊ resolution.Science, 286, 255-260.

Maeda, A., Sasaki, J., Shichida, Y. & Yoshizawa, T.(1992). Water structural changes in the bacteriorho-dopsin photocycle: analysis by Fourier transforminfrared spectroscopy. Biochemistry, 31, 462-467.

Maeda, A., Sasaki, J., Yamazaki, Y., Needleman, R. &Lanyi, J. K. (1994). Interaction of aspartate 85 with awater molecule and the protonated Schiff base inthe L intermediate of bacteriorhodopsin: a FourierTransform infrared spectroscopy study. Biochemis-try, 33, 1713-1717.

Maeda, A., Tomson, F. L., Gennis, R. B., Ebrey, T. G. &Balashov, S. P. (1999). Chromophore-protein-waterinteractions in the L intermediate of bacteriorhodop-sin: FTIR study of the photoreaction of L at 80 K.Biochemistry, 38, 8800-8807.

Marti, T., Otto, H., Mogi, T., RoÈsselet, S. J., Heyn, M. P.& Khorana, H. G. (1991). Bacteriorhodopsinmutants containing single substitutions of serine orthreonine residues are all active in proton trans-location. J. Biol. Chem. 266, 6919-6927.

Milder, S. J., Thorgeirsson, T. E., Miercke, L. J. W.,Stroud, R. M. & Kliger, D. S. (1991). Effects of deter-

gent environments on the photocycle of puri®edmonomeric bacteriorhodopsin. Biochemistry, 30,1751-1761.

Mitsuoka, K., Hirai, T., Murata, K., Miyazawa, A.,Kidera, A., Kimura, Y. & Fujiyoshi, Y. (1999). Thestructure of bacteriorhodopsin at 3.0 AÊ resolutionbased on electron crystallography: implication ofthe charge distribution. J. Mol. Biol. 286, 861-882.

Oesterhelt, D. (1995). Structure and function in halo-rhodopsin. Isr. J. Chem. 35, 475-494.

Oesterhelt, D. (1998). The structure and mechanism ofthe family of retinal proteins from halophilicarchaea. Curr. Opin. Struct. Biol. 8, 489-500.

Oka, T., Kamikubo, H., Tokunaga, F., Lanyi, J. K.,Needleman, R. & Kataoka, M. (1997). X-ray diffrac-tion studies of bacteriorhodopsin. Determination ofthe positions of mercury label at several engineeredcysteine residues. Photochem. Photobiol. 66, 768-773.

Oka, T., Kamikubo, H., Tokunaga, F., Lanyi, J. K.,Needleman, R. & Kataoka, M. (1999). Confor-mational change of helix G in the bacteriorhodopsinphotocycle: investigation with heavy atom labelingand X-ray diffraction. Biophys. J. 76, 1018-1023.

Otwinowski, Z. (1993). SCALEPACK Data Collectionand Processing. In Data Collection and Processing(Sawyer, L., Isaacs, N. & Bailey, S., eds), pp. 56-62,SERC Daresbury Laboratory, Warrington, UK.

Papadopoulos, G., Dencher, N. A., Zaccai, G. & BuÈ ldt,G. (1990). Water molecules and exchangeablehydrogen ions at the active centre of bacteriorho-dopsin localized by neutron diffraction. Elements ofthe proton pathway. J. Mol. Biol. 214, 15-19.

Petkova, A. T., Hu, J. G. G., Bizounok, M., Simpson, M.,Grif®n, R. G. & Herzfeld, J. (1999). Arginine activityin the proton-motive photocycle of bacteriorhodop-sin: solid-state NMR studies of the wild-type andD85N proteins. Biochemistry, 38, 1562-1572.

PfefferleÂ, J.-M., Maeda, A., Sasaki, J. & Yoshizawa, T.(1991). Fourier transform infrared study of the Nintermediate of bacteriorhodopsin. Biochemistry, 30,6548-6556.

Pfeiffer, M., Rink, T., Gerwert, K., Oesterhelt, D. &Steinhoff, H. J. (1999). Site-directed spin-labelingreveals the orientation of the amino acid side-chainsin the E-F loop of bacteriorhodopsin. J. Mol. Biol.287, 163-171.

Radionov, A. N., Kalaidzidis, I. V. & Kaulen, A. D.(1996). Complicated character of the M decay pHdependence in the D96N mutant is due to the twopathways of the M conversion. FEBS Letters, 399,251-254.

Rammelsberg, R., Huhn, G., LuÈ bben, M. & Gerwert, K.(1998). Bacteriorhodopsin's intramolecular proton-release pathway consists of a hydrogen-bondednetwork. Biochemistry, 37, 5001-5009.

Richter, H. T., Brown, L. S., Needleman, R. & Lanyi, J. K.(1996a). A linkage of the pKa`s of Asp-85 andGlu-204 forms part of the reprotonation switch ofbacteriorhodopsin. Biochemistry, 35, 4054-4062.

Richter, H. T., Needleman, R., Kandori, H., Maeda, A. &Lanyi, J. K. (1996b). Relationship of retinal con®gur-ation and internal proton transfer at the end of thebacteriorhodopsin photocycle. Biochemistry, 35,15461-15466.

Richter, H. T., Needleman, R. & Lanyi, J. K. (1996c). Per-turbed interaction between residues 85 and 204 inTyr-185-Phe and Asp-85-Glu bacteriorhodopsins.Biophys. J. 71, 3392-3398.


Roepe, P., Ahl, P. L., Das Gupta, S. K., Herzfeld, J. &Rothschild, K. J. (1987). Tyrosine and carboxyl pro-tonation changes in the bacteriorhodopsin photo-cycle. 1. M412 and L550 intermediates. Biochemistry,26, 6696-6707.

Rothschild, K. J., Roepe, P., Lugtenburg, J., Pardoen, J.A. (1984). Fourier transform infrared evidence forSchiff base alteration in the ®rst step of the bacter-iorhodopsin photocycle. Biochemistry, 23, 6103-6109.

Sasaki, J., Shichida, Y., Lanyi, J. K. & Maeda, A. (1992).Protein changes associated with reprotonation ofthe Schiff base in the photocycle of asp96-asnbacteriorhodopsin. The MN intermediate withunprotonated Schiff base but N-like protein struc-ture. J. Biol. Chem. 267, 20782-20786.

Sasaki, J., Lanyi, J. K., Needleman, R., Yoshizawa, T. &Maeda, A. (1994). Complete identi®cation of C1Ostretching vibrational bands of protonated asparticacid residues in the difference infrared spectra of Mand N intermediates versus bacteriorhodopsin.Biochemistry, 33, 3178-3184.

Sass, H. J., Schachowa, I. W., Rapp, G., Koch, M. H. J.,Oesterhelt, D., Dencher, N. A. & BuÈ ldt, G. (1997).The tertiary structural changes in bacteriorhodopsinoccur between M states: X-ray diffraction and Four-ier transform infrared spectroscopy. EMBO J. 16,1484-1491.

Scherrer, P., Alexiev, U., Marti, T., Khorana, H. G. &Heyn, M. P. (1994). Covalently bound pH-indicatordyes at selected extracellular or cytoplasmic sites inbacteriorhodopsin. Proton migration along the sur-face of bacteriorhodopsin micelles and its delayedtransfer from surface to bulk. Biochemistry, 33,13684-13692.

Sheldrick, G. M. & Schneider, T. (1997). High resolutionre®nement. Methods Enzymol. 277, 319-343.

Siebert, F. & Mantele, W. (1983). Investigation of theprimary photochemistry of bacteriorhodopsin bylow-temperature Fourier-transform infrared spec-troscopy. Eur. J. Biochem. 130, 565-573.

Smith, S. O., Pardoen, J. A., Mulder, P. P. J., Curry, B.,Lugtenburg, J. & Mathies, R. A. (1983). Chromo-phore structure in bacteriorhodopsin's O640 photo-intermediate. Biochemistry, 22, 6141-6148.

Smith, S. O., Myers, A. B., Pardoen, J. A., Winkel, C.,Mulder, P. P. J., Lugtenburg, J. & Mathies, R. A.(1984). Determination of retinal Schiff base con®gur-ation in bacteriorhodopsin. Proc. Natl Acad. Sci.USA, 81, 2055-2059.

Smith, S. O., Hornung, I., van der Steen, R., Pardoen,J. A., Braiman, M. S., Lugtenburg, J. & Mathies,R. A. (1986). Are C14-C15 single bond isomerizationsof the retinal chromophore involved in the proton-pumping mechanism of bacteriorhodopsin? Proc.Natl Acad. Sci. USA, 83, 967-971.

Steinhoff, H.-J., Mollaaghababa, R., Altenbach, C.,Hideg, K., Krebs, M., Khorana, H. G. & Hubbell,W. L. (1994). Time-resolved detection of structuralchanges during the photocycle of spin-labeled bac-teriorhodopsin. Science, 266, 105-107.

Subramaniam, S., Greenhalgh, D. A., Rath, P.,Rothschild, K. J. & Khorana, H. G. (1991). Replace-ment of leucine-93 by alanine or threonine slowsdown the decay of the N and O intermediates inthe photocycle of bacteriorhodopsin: implicationsfor proton uptake and 13-cis-retinal! all-trans-retinal reisomerization. Proc. Natl Acad. Sci. USA,88, 6873-6877.

Subramaniam, S., Gerstein, M., Oesterhelt, D. &Henderson, R. (1993). Electron diffraction analysisof structural changes in the photocycle of bacterio-rhodopsin. EMBO J. 12, 1-8.

Subramaniam, S., Lindahl, I., Bullough, P., Faruqi, A. R.,Tittor, J., Oesterhelt, D., Brown, L. S., Lanyi, J. K. &Henderson, R. (1999). Protein conformationalchanges in the bacteriorhodopsin photocycle. J. Mol.Biol. 287, 145-161.

SzaÂraz, S., Oesterhelt, D. & Ormos, P. (1994). pH-induced structural changes in bacteriorhodopsinstudied by Fourier transform infrared spectroscopy.Biophys. J. 67, 1706-1712.

Takei, H., Gat, Y., Rothman, Z., Lewis, A. & Sheves, M.(1994). Active site lysine backbone undergoes con-formational changes in the bacteriorhodopsinphotocycle. J. Biol. Chem. 269, 7387-7389.

Tavan, P., Schulten, K. & Oesterhelt, D. (1985). Theeffect of protonation and electrical interactions onthe stereochemistry of retinal Schiff bases. Biophys.J. 47, 415-430.

Thorgeirsson, T. E., Xiao, W., Brown, L. S., Needleman,R., Lanyi, J. K. & Shin, Y.-K. (1997). Opening of thecytoplasmic proton channel in bacteriorhodopsin.J. Mol. Biol. 273, 951-957.

VaÂroÂ , G. & Lanyi, J. K. (1991a). Thermodynamics andenergy coupling in the bacteriorhodopsin photo-cycle. Biochemistry, 30, 5016-5022.

VaÂroÂ , G. & Lanyi, J. K. (1991b). Kinetic and spectro-scopic evidence for an irreversible step betweendeprotonation and reprotonation of the Schiff basein the bacteriorhodopsin photocycle. Biochemistry,30, 5008-5015.

VaÂroÂ , G. & Lanyi, J. K. (1995). Effects of hydrostaticpressure on the kinetics reveal a volume increaseduring the bacteriorhodopsin photocycle. Biochemis-try, 34, 12161-12169.

Vonck, J. (1996). A three-dimensional difference map ofthe N intermediate in the bacteriorhodopsin photo-cycle: part of the F helix tilts in the M to N tran-sition. Biochemistry, 35, 5870-5878.

Weidlich, O., Schalt, B., Friedman, N., Sheves, M.,Lanyi, J. K., Brown, L. S. & Siebert, F. (1996). Stericinteraction between the 9-methyl group of the reti-nal and tryptophan 182 controls 13-cis to all-transreisomerization and proton uptake in the bacterio-rhodopsin photocycle. Biochemistry, 35, 10807-10814.

Weik, M., Zaccai, G., Dencher, N. A., Oesterhelt, D. &Hauss, T. (1998). Structure and hydration of theM-state of the bacteriorhodopsin mutant D96N stu-died by neutron diffraction. J. Mol. Biol. 275, 625-634.

Wikstrom, M. (1998). Proton translocation by bacterio-rhodopsin and heme-copper oxidases. Curr. Opin.Struc. Biol. 8, 480-488.

Yamazaki, Y., Sasaki, J., Hatanaka, M., Maeda, A.,Kandori, H., Needleman, R., Shinada, T., Yoshihara,K., Brown, L. S. & Lanyi, J. K. (1995). Interaction oftryptophan182 with the 9-methyl group of theretinal in the L intermediate of bacteriorhodopsin.Biochemistry, 34, 577-582.

Yamazaki, Y., Tuzi, S., SaitoÃ , H., Kandori, H.,Needleman, R., Lanyi, J. K. & Maeda, A. (1996).Hydrogen bonds of water and C1O groups coordi-nate long-range structural in the L photointermedi-ate of bacteriorhodopsin. Biochemistry, 35, 4063-4068.

ZimaÂnyi, L., Cao, Y., Chang, M., Ni, B., Needleman, R.& Lanyi, J. K. (1992a). The two consecutive M sub-


states in the photocycle of bacteriorhodopsin areaffected speci®cally by the D85N and D96N residuereplacements. Photochem. Photobiol. 56, 1049-1055.

ZimaÂnyi, L., VaÂroÂ , G., Chang, M., Ni, B., Needleman, R.& Lanyi, J. K. (1992b). Pathways of proton releasein the bacteriorhodopsin photocycle. Biochemistry,31, 8535-8543.

Zscherp, C. & Heberle, J. (1997). Infrared differencespectra of the intermediates L, M, N, and O of the

bacteriorhodopsin photoreaction obtained by time-resolved attenuated total re¯ection spectroscopy.J. Phys. Chem. ser. B, 101, 10542-10547.

Zscherp, C., Schlesinger, R., Tittor, J., Oesterhelt, D. &Heberle, J. (1999). In situ determination of transientpKa changes of internal amino acids of bacterio-rhodopsin by using time-resolved attenuated totalre¯ection Fourier-transform infrared spectroscopy.Proc. Natl Acad. Sci. USA, 96, 5498-5503.

Edited by D. C. Rees

(Received 23 March 2000; received in revised form 15 May 2000; accepted 17 May 2000)

coupling photoisomerization of retinal to directional ...€¦ · the principal question of the...

Documents