Anion binding to the Schiff base of the bacteriorhodopsin mutants Asp-85----Asn/Asp-212----Asn and Arg-82----Gln/Asp-85----Asn/Asp-212----Asn.

Studies of bacteriorhodopsin have indicated that the charge environment of the protonated Schiff base consists of residues Asp-85, Asp-212, and Arg-82. As shown recently (Marti, T., R?sselet, S. J., Otto, H., Heyn, M. P., and Khorana, H. G. (1991) J. Biol. Chem. 266, 18674-18683), in the double mutant Asp-85----Asn/Asp-212----Asn chromophore formation is restored in the presence of salts, suggesting that exogenous anions function as counterions to the protonated Schiff base. To investigate the role of Arg-82 and of the Schiff base in anion binding, we have prepared the triple mutant Arg-82----Gln/Asp-85----Asn/Asp-212----Asn and compared its properties with those of the Asp-85----Asn/Asp-212----Asn double mutant. Regeneration of the chromophore with absorption maximum near 560 nm occurs in the triple mutant in the presence of millimolar salt, whereas in the double mutant molar salt concentrations are required. Spectrometric titrations reveal that the pKa of Schiff base deprotonation is markedly reduced from 11.3 for the wild type to 4.9 for the triple mutant in 1 mM NaCl and to 5.5 for the double mutant in 10 mM NaCl. In both mutants, increasing the chloride concentration promotes protonation of the chromophore and results in a continuous rise of the Schiff base pKa, yielding a value of 8.4 and 7.6, respectively, in 4 M NaCl. The absorption maximum of the two mutants shows a progressive red shift, as the ionic radius of the halide increases in the sequence fluoride, chloride, bromide, and iodide. An identical spectral correlation in the presence of halides is observed for the acid-purple form of bacteriorhodopsin. We conclude, therefore, that upon neutralization of the two counterions Asp-85 and Asp-212 by mutation or by protonation at low pH, exogenous anions substitute as counterions by directly binding to the protonated Schiff base. This interaction may provide the basis for the proposed anion translocation by the acid-purple form of bacteriorhodopsin as well as by the related halorhodopsin.     

Aspartic acid 85 in bacteriorhodopsin functions both as proton acceptor and negative counterion to the Schiff base.

In bacteriorhodopsin Asp85 has been proposed to function both as a negative counterion to the Schiff base and as proton acceptor in the early stages of the photocycle. To test this proposal further, we have replaced Asp85 by His. The rationale for this replacement is that although His can function as a proton acceptor, it cannot provide a negative charge at residue 85 to serve as a counterion to the protonated Schiff base. We show here that the absorption spectrum of the D85H mutant is highly sensitive to the pH of the external medium. From spectroscopic titrations, we have determined the apparent pK for deprotonation of the Schiff base to be 8.8 +/- 0.1 and the apparent pK for protonation of the His85 side chain to be approximately 3.5. Between pH 3.5 and 8.8, where the Schiff base is protonated, and the His side chain is deprotonated, the D85H mutant is completely inactive in proton transport. Time-resolved studies show that there is no detectable formation of an M-like intermediate in the photocycle of the D85H mutant. These experiments show that the presence of a neutral proton-accepting moiety at residue 85 is not sufficient for carrying out light-driven proton transport. The requirements at residue 85 are therefore for a group that serves both as a negatively charged counterion and as a proton acceptor.     

Inversion of proton translocation in bacteriorhodopsin mutants D85N, D85T, and D85,96N.

Proton translocation activity of bacteriorhodopsin mutants lacking the proton acceptor Asp-85 was investigated using the black lipid membrane technique. Mutants D85N, D85T, and D85,96N were constructed and homologously expressed in Halobacterium salinarium to yield a membrane fraction with a buoyant density of 1.18 g/cm3, i.e., identical to that of wild-type purple membrane. In all mutants, the absorbance maximum was red-shifted between 27 and 49 nm compared with wild type, and the pKa values of the respective Schiff bases were reduced to between 8.3 and 8.9 compared with the value of > 13 in wild type. Therefore, a mixture of chromophores absorbing at 410 nm (deprotonated form) and around 600 nm (protonated form) exists at physiological pH. In continuous blue light, the deprotonated form generates stationary photocurrents. The currents are enhanced by a factor of up to 50 upon addition of azide in D85N and D85,96N mutants, whereas D85T shows no azide effect. The direction of these currents is the same as in wild type in yellow light. Yellow light alone is not sufficient to generate stationary currents in the mutants, but increasing yellow light intensity in the presence of blue light leads to an inversion of the current. Because all currents are carried by protons, this two-photon process demonstrates an inverted proton translocation by BR mutants.     

Effects of Asp-96----Asn, Asp-85----Asn, and Arg-82----Gln single-site substitutions on the photocycle of bacteriorhodopsin

Bacteriorhodopsin (BR) with the single-site substitutions Arg-82----Gln (R82Q), Asp-85----Asn (D85N), and Asp-96----Asn (D96N) is studied with time-resolved absorption spectroscopy in the time regime from nanoseconds to seconds. Time-resolved spectra are analyzed globally by using multiexponential fitting of the data at multiple wavelengths and times. The photocycle kinetics for BR purified from each mutant are determined for micellar solutions in two detergents, nonyl glucoside and CHAPSO, and are compared to results from studies on delipidated BR (d-BR) in the same detergents. D85N has a red-shifted ground-state absorption spectrum, and the formation of an M intermediate is not observed. R82Q undergoes a pH-dependent transition between a purple and a blue form with different pKa values in the two detergents. The blue form has a photocycle resembling that for D85N, while the purple form of R82Q forms an M intermediate that decays more rapidly than in d-BR. The purple form of R82Q does not light-adapt to the same extent as d-BR, and the spectral changes in the photocycle suggest that the light-adapted purple form of R82Q contains all-trans- and 13-cis-retinal in approximately equal proportions. These results are consistent with the suggestions of others for the roles of Arg-82 and Asp-85 in the photocycle of BR, but results for D96N suggest a more complex role for Asp-96 than previously suggested. In nonyl glucoside, the apparent decay of the M-intermediate is slower in D96N than in d-BR, and the M decay shows biphasic kinetics. However, the role of Asp-96 is not limited to the later steps of the photocycle. In D96N, the decay of the KL intermediate is accelerated, and the rise of the M intermediate has an additional slow phase not observed in the kinetics of d-BR. The results suggest that Asp-96 may play a role in regulating the structure of BR and how it changes during the photocycle.     

Mechanism of primary proton transfer in bacteriorhodopsin

Recent structures of putative intermediates in the bacteriorhodopsin photocycle have provided valuable snapshots of the mechanism by which protons are pumped across the membrane. However, key steps remain highly controversial, particularly the proton transfer occurring immediately after retinal trans-->cis photoisomerization. The gradual release of stored energy is inherently nonequilibrium: which photocycle intermediates are populated depends not only on their energy but also on their interconversion rates. To understand why the photocycle follows a productive (i.e., pumping), rather than some unproductive, relaxation pathway, it is necessary to know the relative energy barriers of individual steps. To discriminate between the many proposed scenarios of this process, we computed all its possible minimum-energy paths. This reveals that not one, but three very different pathways have energy barriers consistent with experiment. This result reconciles the conflicting views held on the mechanism and suggests a strategy by which the protein renders this essential step resilient.     

The titrations of Asp-85 and of the cation binding residues in bacteriorhodopsin are not coupled

An outstanding problem relating to the structure and function of bacteriorhodopsin (bR), which is the only protein in the purple membrane of the photosynthetic microorganism Halobacterium salinarium, is the relation between the titration of Asp-85 and the binding/unbinding of metal cations. An extensively accepted working hypothesis has been that the two titrations are coupled, namely, protonation of Asp-85 (located in the vicinity of the retinal chromophore) and cation unbinding occur concurrently. We have carried out a series of experiments in which the purple blue equilibrium and the binding of Mn2+ ions (monitored by electron spin resonance) were followed as a function of pH for several (1-4) R = [Mn2+]/[bR] molar ratios. Data were obtained for native bR, bR mutants, artificial bR and chemically modified bR. We find that in the native pigment the two titrations are separated by more than a pKa unit [delta pKa = pKa(P/B)-pKa(Mn2+) = (4.2-2.8) = 1.4]. In the non-native systems, delta pKa values as high as 5 units, as well as negative delta pKas, are observed. We conclude that the pH titration of cation binding residues in bR is not directly related to the titration of Asp-85. This conclusion is relevant to the nature of the high affinity cation sites in bR and to their role in the photosynthetic function of the pigment.     

Structural investigation of the active site in bacteriorhodopsin: geometric constraints on the roles of Asp-85 and Asp-212 in the proton-pumping mechanism from solid state NMR

Constraints on the proximity of the carboxyl carbons of the Asp-85 and Asp-212 side chains to the 14-carbon of the retinal chromophore have been established for the bR(555), bR(568), and M(412) states of bacteriorhodopsin (bR) using solid-state NMR spectroscopy. These distances were examined via (13)C-(13)C magnetization exchange, which was observed in two-dimensional RF-driven recoupling (RFDR) and spin diffusion experiments. A comparison of relative RFDR cross-peak intensities with simulations of the NMR experiments yields distance measurements of 4.4 +/- 0.6 and 4.8 +/- 1.0 A for the [4-(13)C]Asp-212 to [14-(13)C]retinal distances in bR(568) and M(412), respectively. The spin diffusion data are consistent with these results and indicate that the Asp-212 to 14-C-retinal distance increases by 16 +/- 10% upon conversion to the M-state. The absence of cross-peaks from [14-(13)C]retinal to [4-(13)C]Asp-85 in all states and between any [4-(13)C]Asp residue and [14-(13)C]retinal in bR(555) indicates that these distances exceed 6.0 A. For bR(568), the NMR distance constraints are in agreement with the results from recent diffraction studies on intact membranes, while for the M state the NMR results agree with theoretical simulations employing two bound waters in the region of the Asp-85 and Asp-212 residues. The structural information provided by NMR should prove useful for refining the current understanding of the role of aspartic acid residues in the proton-pumping mechanism of bR.     

Photocycle of dried acid purple form of bacteriorhodopsin.

The photocycle of dried bacteriorhodopsin, pretreated in a 0.3 M HCl solution, was studied. Some properties of this dried sample resemble that of the acid purple suspension: the retinal conformation is mostly all-trans, 15-anti form, the spectrum of the sample is blue-shifted by 5 nm to 560 nm, and it has a truncated photocycle. After photoexcitation, a K-like red-shifted intermediate appears, which decays to the ground state through several intermediates with spectra between the K and the ground state. There are no other bacteriorhodopsin-like intermediates (L, M, N, O) present in the photocycle. The K to K' transition proceeds with an enthalpy decrease, whereas during all the following steps, the entropic energy of the system decreases. The electric response signal of the oriented sample has only negative components, which relaxes to zero. These suggest that the steps after intermediate K represent a relaxation process, during which the absorbed energy is dissipated and the protein returns to its original ground state. The initial charge separation on the retinal is followed by limited charge rearrangements in the protein, and later, all these relax. The decay times of the intermediates are strongly influenced by the humidity of the sample. Double-flash experiments proved that all the intermediates are directly driven back to the ground state. The study of the dried acid purple samples could help in understanding the fast primary processes of the protein function. It may also have importance in technical applications.     

Estimated acid dissociation constants of the Schiff base, Asp-85, and Arg-82 during the bacteriorhodopsin photocycle.

The pK(a) values of D85 in the wild-type and R82Q, as well as R82A recombinant bacteriorhodopsins, and the Schiff base in the D85N, D85T, and D85N/R82Q proteins, have been determined by spectroscopic titrations in the dark. They are used to estimate the coulombic interaction energies and the pK(a) values of the Schiff base, D85, and R82 during proton transfer from the Schiff base to D85, and the subsequent proton release to the bulk in the initial part of the photocycle. The pK(a) of the Schiff base before photoexcitation is calculated to be in effect only 5.3-5.7 pH units higher than that of D85; overcoming this to allow proton transfer to D85 requires about two thirds of the estimated excess free energy retained after absorption of a photon. The proton release on the extracellular surface is from an unidentified residue whose pK(a) is lowered to about 6 after deprotonation of the Schiff base (Zimanyi, L., G. Varo, M. Chang, B. Ni, R. Needleman, and J.K. Lanyi, 1992. Biochemistry. 31:8535-8543). We calculate that the pK(a) of the R82 is 13.8 before photoexcitation, and it is lowered after proton exchange between the Schiff base and D85 only by 1.5-2.3 pH units. Therefore, coulombic interactions alone do not appear to change the pK(a) of R82 as much and D85 only by 1.5-2.3 pH units. Therefore, coulombic interactions alone do not appear to change the pK(a) of R82 as much as required if it were the proton release group.     

Coupling of the reisomerization of the retinal, proton uptake, and reprotonation of Asp-96 in the N photointermediate of bacteriorhodopsin

In the N to O reaction of the bacteriorhodopsin photocycle, Asp-96 is protonated from the cytoplasmic surface, and coupled to this, the retinal isomerizes from 13-cis,15-anti back to the initial all-trans configuration. To dissect the two steps, and to better understand how and why they occur, we describe the properties of two groups of site-specific mutants in which the N intermediate has greatly increased lifetime. In the first group, with the mutations near the retinal, an unusual N state is produced in which the retinal is 13-cis,15-anti but Asp-96 has a protonated carboxyl group. The apparent pK(a) for the protonation is 7.5, as in the wild-type. It is likely that here the interference with N decay is the result of steric conflict of side-chains with the retinal or with the side-chain of Lys-216 connected to the retinal, which delays the reisomerization after protonation of Asp-96. In the second group, with the mutations located near Asp-96 or between Asp-96 and the cytoplasmic surface, reprotonation of Asp-96 is strongly perturbed. The reisomerization of the retinal occurs only after recovery from a long-living protein conformation in which reprotonation of Asp-96 is either entirely blocked or blocked at low pH.     

Properties of Asp212----Asn bacteriorhodopsin suggest that Asp212 and Asp85 both participate in a counterion and proton acceptor complex near the Schiff base.

The gene coding for bacteriorhodopsin was modified in vitro to replace Asp212 with asparagine and expressed in Halobacterium halobium. X-ray diffraction measurements showed that the major lattice dimension of purple membrane containing the mutated bacteriorhodopsin was the same as wild type. At pH greater than 7, the Asp212----Asn chromophore was blue (absorption maximum at 585 nm) and exhibited a photocycle containing only the intermediates K and L, i.e. a reaction sequence very similar to that of wild-type bacteriorhodopsin at pH less than 3 and the blue form of the Asp85----Glu protein at pH less than 9. Since in the latter cases these effects are attributed to protonation of residue 85, it now appears that removal of the carboxylate of Asp212 has similar consequences as removing the carboxylate of Asp85. However, an important difference is that only Asp85 affects the pKa of the Schiff base. At pH less than 7, the Asp212----Asn protein was purple (absorption maximum at 569 nm) but photoexcitation produced only 15% of the normal amount of M and the transport activity was partial. The reactions of the blue and purple forms after photoexcitation are both quantitatively accounted for by a proposed scheme, K in equilibrium with L1 in equilibrium with L2----BR, but with the addition of an L1 in equilibrium with M reaction with unfavorable pKa for Schiff base deprotonation in the purple form. The latter hinders the transient accumulation of M, and the consequent branching at L1 allows only partial proton transport activity. The results are consistent with the existence of a complex counterion for the Schiff base proposed earlier (De Groot, H. J. M., Harbison, G. S., Herzfeld, J., and Griffin, R. G. (1989) Biochemistry 28, 3346-3353) and suggest that Asp85, Asp212, and at least one other protonable residue participate in it.     

Vibrational spectroscopy of bacteriorhodopsin mutants: light-driven proton transport involves protonation changes of aspartic acid residues 85, 96, and 212

Fourier transform infrared (FTIR) difference spectra have been obtained for the bR----K, bR----L, and bR----M photoreactions in bacteriorhodopsin mutants in which Asp residues 85, 96, 115, and 212 have been replaced by Asn and by Glu. Difference peaks that had previously been attributed to Asp COOH groups on the basis of isotopic labeling were absent or shifted in these mutants. In general, each COOH peak was affected strongly by mutation at only one of the four residues. Thus, it was possible to assign each peak tentatively to a particular Asp. From these assignments, a model for the proton-pumping mechanism of bR is derived, which features proton transfers among Asp-85, -96, and -212, the chromophore Schiff base, and other ionizable groups within the protein. The model can explain the observed COOH peaks in the FTIR difference spectra of bR photointermediates and could also account for other recent results on site-directed mutants of bR.     

Structural and energetic determinants of primary proton transfer in bacteriorhodopsin.

In the light-driven bacteriorhodopsin proton pump, the first proton transfer step is from the retinal Schiff base to a nearby carboxylate group. The mechanism of this transfer step is highly controversial, in particular whether a direct proton jump is allowed. Here, we review the structural and energetic determinants of the direct proton transfer path computed by using a combined quantum mechanical/molecular mechanical approach. Both protein flexibility and electrostatic interactions play an important role in shaping the proton transfer energy profile. Detailed analysis of the energetics of putative transitions in the first half of the photocycle focuses on two elements that determine the likelihood that a given configuration of the active site is populated during the proton-pumping cycle. First, the rate-limiting barrier for proton transfer must be consistent with the kinetics of the photocycle. Second, the active-site configuration must be compatible with a productive overall pumping cycle.     

Bacteriorhodopsin mutants D85N, D85T and D85,96N as proton pumps

Proton translocation in the BR mutants D85N, D85T and D85,96N was studied by attachment of purple membranes to planar lipid bilayers. Pump currents in these mutants were measured via capacitive coupling and by use of the appropriate ionophores. All mutants have a reduced pK of their Schiff bases around 8-8.5 in common. At physiological pH, a mixture of chromophores absorbing at 410 nm (deprotonated form) and around 600 nm (protonated form) coexists. Excitation with continuous blue light induces in all three mutants an outwardly directed stationary pump current. These currents are enhanced upon addition of azide in D85N and D85,96N by a factor of 50, but no azide enhancement is observed in D85T. Yellow light alone induces transient inwardly directed currents in the mutants but additional blue light leads to a stationary current with the same direction. All the observed currents are carried by protons, so that the consecutive absorption of a yellow and a blue photon leads to inverted stationary photocurrents by the mutants, as observed with halorhodopsin (HR). A mechanistic model describing the inversion of proton pumping is discussed by the cis-trans, trans-cis isomerization of the retinal and the different proton accessibility of the Schiff base from the extracellular or the cytoplasmic side of the membrane.     

Titration of aspartate-85 in bacteriorhodopsin: what it says about chromophore isomerization and proton release.

Titration of Asp-85, the proton acceptor and part of the counterion in bacteriorhodopsin, over a wide pH range (2-11) leads us to the following conclusions: 1) Asp-85 has a complex titration curve with two values of pKa; in addition to a main transition with pKa = 2.6 it shows a second inflection point at high pH (pKa = 9.7 in 150-mM KCl). This complex titration behavior of Asp-85 is explained by interaction of Asp-85 with an ionizable residue X'. As follows from the fit of the titration curve of Asp-85, deprotonation of X' increases the proton affinity of Asp-85 by shifting its pKa from 2.6 to 7.5. Conversely, protonation of Asp-85 decreases the pKa of X' by 4.9 units, from 9.7 to 4.8. The interaction between Asp-85 and X' has important implications for the mechanism of proton transfer. In the photocycle after the formation of M intermediate (and protonation of Asp-85) the group X' should release a proton. This deprotonated state of X' would stabilize the protonated state of Asp-85.2) Thermal isomerization of the chromophore (dark adaptation) occurs on transient protonation of Asp-85 and formation of the blue membrane. The latter conclusion is based on the observation that the rate constant of dark adaptation is directly proportional to the fraction of blue membrane (in which Asp-85 is protonated) between pH 2 and 11. The rate constant of isomerization is at least 10(4) times faster in the blue membrane than in the purple membrane. The protonated state of Asp-85 probably is important for the catalysis not only of all-trans <=> 13-cis thermal isomerization during dark adaptation but also of the reisomerization of the chromophore from 13-cis to all-trans configuration during N-->O-->bR transition in the photocycle. This would explain why Asp-85 stays protonated in the N and O intermediates.     

Decoupling of photo- and proton cycle in the Asp85-->Glu mutant of bacteriorhodopsin.

Surface bound pH indicators were applied to study the proton transfer reactions in the mutant Asp85-->Glu of bacteriorhodopsin in the native membrane. The amino acid replacement induces a drastic acceleration of the overall rise of the M intermediate. Instead of following this acceleration, proton ejection to the extracellular membrane surface is not only two orders of magnitude slower than M formation, it is also delayed as compared with the wild-type. This demonstrates that Asp85 not only accepts the proton released by the Schiff's base but also regulates very efficiently proton transfer within the proton release chain. Furthermore, Asp85 might be the primary but is not the only proton acceptor/donor group in the release pathway. The Asp85-->Glu substitution also affects the proton reuptake reaction at the cytoplasmic side, although Asp85 is located in the proton release pathway. Proton uptake is slower in the mutant than in the wild-type and occurs during the lifetime of the O intermediate. This demonstrates a feed-back mechanism between Asp85 and the proton uptake pathway in bacteriorhodopsin.     

Influence of the Charge at D85 on the Initial Steps in the Photocycle of Bacteriorhodopsin

Studies have shown that trans-cis isomerization of retinal is the primary photoreaction in the photocycle of the light-driven proton pump bacteriorhodopsin (BR) from Halobacterium salinarum, as well as in the photocycle of the chloride pump halorhodopsin (HR). The transmembrane proteins HR and BR show extensive structural similarities, but differ in the electrostatic surroundings of the retinal chromophore near the protonated Schiff base. Point mutation of BR of the negatively charged aspartate D85 to a threonine T (D85T) in combination with variation of the pH value and anion concentration is used to study the ultrafast photoisomerization of BR and HR for well-defined electrostatic surroundings of the retinal chromophore. Variations of the pH value and salt concentration allow a switch in the isomerization dynamics of the BR mutant D85T between BR-like and HR-like behaviors. At low salt concentrations or a high pH value (pH 8), the mutant D85T shows a biexponential initial reaction similar to that of HR. The combination of high salt concentration and a low pH value (pH 6) leads to a subpopulation of 25% of the mutant D85T whose stationary and dynamic absorption properties are similar to those of native BR. In this sample, the combination of low pH and high salt concentration reestablishes the electrostatic surroundings originally present in native BR, but only a minor fraction of the D85T molecules have the charge located exactly at the position required for the BR-like fast isomerization reaction. The results suggest that the electrostatics in the native BR protein is optimized by evolution. The accurate location of the fixed charge at the aspartate D85 near the Schiff base in BR is essential for the high efficiency of the primary reaction.     

External electric control of the proton pumping in bacteriorhodopsin

Comparative analysis of the photoelectric response of dried films of purple membranes (PM) depending on their degree of orientation is presented. Time dependence of the photo-induced protein electric response signal (PERS) of oriented and non-oriented films to a single laser pulse in the presence of the external electric field (EEF) was experimentally determined. The signal does not appear in the non-oriented films when the EEF is absent, whereas the PERS of the oriented PM films demonstrates the variable polarity on the microsecond time scale. In the presence of the EEF the PERS of the non-oriented film rises exponentially preserving the same polarization. The polarization of the PERS changes by changing the polarity of the EEF with no influence on the time constant of the PERS kinetics. The EEF effect on the PERS of the oriented films is more complicated. By subtracting the PERS when EEF ≠ 0 from the PERS when EEF = 0 the resulting signal is comparable to that of the non-oriented films. Generalizing the experimental data we conclude that the EEF influence is of the same origin for the films of any orientation. To explain the experimental results the two-state model is suggested. It assumes that the EEF directionally changes the pK_a values of the Schiff base (SB) and of the proton acceptor aspartic acid D85 in bacteriorhodopsin. Because of that the SB→D85 proton transfer might be blocked and consequently the L→M intermediate transition should vanish. Thus, on the characteristic time scale τ _{L → M} ≈ 30 μs; both intermediates, the M intermediate, appearing under normal conditions, and the L intermediate as persisting under the blocked conditions when D85 is protonated, should coexist in the film. The total PERS is a result of the potentials corresponding to the electrogenic products of intermediates L and M that are of the opposite polarity. It is concluded that the ratio of bacteriorhodopsin concentrations corresponding to the L and M intermediates is driven by the EEF and, consequently, it should define the PERS of the non-oriented films. According to this model the orientation degree of the film could be evaluated by describing the PERS.     

Pathways of proton release in the bacteriorhodopsin photocycle.

The pH dependencies of the rate constants in the photocycles of recombinant D96N and D115N/D96N bacteriorhodopsins were determined from time-resolved difference spectra between 70 ns and 420 ms after photoexcitation. The results were consistent with the model suggested earlier for proteins containing D96N substitution: BR hv----K----L----M1----M2----BR. Only the M2----M1 back-reaction was pH-dependent: its rate increased with increasing [H+] between pH 5 and 8. We conclude from quantitative analysis of this pH dependency that its reverse, the M1----M2 reaction, is linked to the release of a proton from a group with a pKa = 5.8. This suggests a model for wild-type bacteriorhodopsin in which at pH greater than 5.8 the transported proton is released on the extracellular side from this as yet unknown group and on the 100-microseconds time scale, but at pH less than 5.8, the proton release occurs from another residue and later in the photocycle most likely directly from D85 during the O----BR reaction. We postulate, on the other hand, that proton uptake on the cytoplasmic side will be by D96 and during the N----O reaction regardless of pH. The proton kinetics as measured with indicator dyes confirmed the unique prediction of this model: at pH greater than 6, proton release preceded proton uptake, but at pH less than 6, the release was delayed until after the uptake. The results indicated further that the overall M1----M2 reaction includes a second kinetic step in addition to proton release; this is probably the earlier postulated extracellular-to-cytoplasmic reorientation switch in the proton pump.