Crystal structure of sensory rhodopsin II at 2.4 angstroms: insights into color tuning and transducer interaction.

We report an atomic-resolution structure for a sensory member of the microbial rhodopsin family, the phototaxis receptor sensory rhodopsin II (NpSRII), which mediates blue-light avoidance by the haloarchaeon Natronobacterium pharaonis. The 2.4 angstrom structure reveals features responsible for the 70- to 80-nanometer blue shift of its absorption maximum relative to those of haloarchaeal transport rhodopsins, as well as structural differences due to its sensory, as opposed to transport, function. Multiple factors appear to account for the spectral tuning difference with respect to bacteriorhodopsin: (i) repositioning of the guanidinium group of arginine 72, a residue that interacts with the counterion to the retinylidene protonated Schiff base; (ii) rearrangement of the protein near the retinal ring; and (iii) changes in tilt and slant of the retinal polyene chain. Inspection of the surface topography reveals an exposed polar residue, tyrosine 199, not present in bacteriorhodopsin, in the middle of the membrane bilayer. We propose that this residue interacts with the adjacent helices of the cognate NpSRII transducer NpHtrII.

Coupling photoisomerization of retinal to directional transport in bacteriorhodopsin.

In order to understand how isomerization of the retinal drives unidirectional transmembrane ion transport in bacteriorhodopsin, we determined the atomic structures of the BR state and M photointermediate of the E204Q mutant, to 1.7 and 1.8 A resolution, respectively. Comparison of this M, in which proton release to the extracellular surface is blocked, with the previously determined M in the D96N mutant indicates that the changes in the extracellular region are initiated by changes in the electrostatic interactions of the retinal Schiff base with Asp85 and Asp212, but those on the cytoplasmic side originate from steric conflict of the 13-methyl retinal group with Trp182 and distortion of the pi-bulge of helix G. The structural changes suggest that protonation of Asp85 initiates a cascade of atomic displacements in the extracellular region that cause release of a proton to the surface. The progressive relaxation of the strained 13-cis retinal chain with deprotonated Schiff base, in turn, initiates atomic displacements in the cytoplasmic region that cause the intercalation of a hydrogen-bonded water molecule between Thr46 and Asp96. This accounts for the lowering of the pK(a) of Asp96, which then reprotonates the Schiff base via a newly formed chain of water molecules that is extending toward the Schiff base.

Structural changes in bacteriorhodopsin during ion transport at 2 angstrom resolution.

Crystal structures of the Asp96 to Asn mutant of the light-driven proton pump bacteriorhodopsin and its M photointermediate produced by illumination at ambient temperature have been determined to 1.8 and 2.0 angstroms resolution, respectively. The trapped photoproduct corresponds to the late M state in the transport cycle-that is, after proton transfer to Asp85 and release of a proton to the extracellular membrane surface, but before reprotonation of the deprotonated retinal Schiff base. Its density map describes displacements of side chains near the retinal induced by its photoisomerization to 13-cis,15-anti and an extensive rearrangement of the three-dimensional network of hydrogen-bonded residues and bound water that accounts for the changed pKa values (where Ka is the acid constant) of the Schiff base and Asp85. The structural changes detected suggest the means for conserving energy at the active site and for ensuring the directionality of proton translocation.

Structure of bacteriorhodopsin at 1.55 A resolution.

Th?e atomic structure of the light-driven ion pump bacteriorhodopsin and the surrounding lipid matrix was determined by X-ray diffraction of crystals grown in cubic lipid phase. In the extracellular region, an extensive three-dimensional hydrogen-bonded network of protein residues and seven water molecules leads from the buried retinal Schiff base and the proton acceptor Asp85 to the membrane surface. Near Lys216 where the retinal binds, transmembrane helix G contains a pi-bulge that causes a non-proline? kink. The bulge is stabilized by hydrogen-bonding of the main-chain carbonyl groups of Ala215 and Lys216 with two buried water molecules located between the Schiff base and the proton donor Asp96 in the cytoplasmic region. The results indicate extensive involvement of bound water molecules in both the structure and the function of this seven-helical membrane protein. A bilayer of 18 tightly bound lipid chains forms an annulus around the protein in the crystal. Contacts between the trimers in the membrane plane are mediated almost exclusively by lipids.

Do ATP4- and Mg2+ bind stepwise to the F1-ATPase of Halobacterium saccharovorum?

It is commonly believed that MgATP2- is the substrate of F1-ATPases and ATP4- acts as a competitive inhibitor. However, the velocity equation for such competitive inhibition is equivalent to that for a rapid equilibrium ordered binding mechanism in which ATP4- adds first and the binding of Mg2+ is dependent on the formation of the E x ATP4- complex. According to this ordered-binding model, solution formed MgATP2- is not recognized by the ATPase as a direct substrate, and the high-affinity binding of Mg2+ to the E x ATP4- complex is the key reaction towards the formation of the ternary complex. These models (and others) were tested with an F1- ATPase, isolated from Halobacterium saccharovorum, by evaluating the rate of ATP hydrolysis as a function of free [ATP4-] or free [Mg2+]. The rates were asymmetrical with respect to increasing [ATP4-] versus increasing [Mg2+]. For the ordered-binding alternative, a series of apparent dissociation constants were obtained for ATP4-(K(A)aPP), which decreased as [Mg2+] increased. From this family of K(A)aPP the true K(A) was retrieved by extrapolation to [Mg2+] = 0 and was found to be 0.2 mM. The dissociation constants for Mg2+, established from these experiments, were also apparent (K(B)aPP) and dependent on [ATP4-] as well as on the pH. The actual K(B) was established from a series of K(B)aPP by extrapolating to [ATP4-] = infinity and to the absence of competing protons, and was found to be 0.0041 mM. The pKa of the protonable group for Mg2+ binding is 8.2. For the competitive inhibition alternative, rearrangement of the constants and fitting to the velocity equation gave an actual binding constant for MgATP2- (K(EAB)) of 0.0016 mM and for ATP4- (K(EA)) of 0.2 mM. Decision between the two models has far-reaching mechanistic implications. In the competitive inhibition model MgATP2- binds with high affinity, but Mg2+ cannot bind once the E x ATP4- complex is formed, while in the ordered-binding model binding of Mg2+ requires that ATP4- adds first. The steric constraints evident in the diffraction structure of the ATP binding site in the bovine mitochondrial F-ATPase [Abrahams, J. P., Leslie, A. G. W., Lutter, R. & Walker, J. E. (1994) Nature 370, 621-628] tend to favor the ordered-binding model, but the final decision as to which kinetic model is valid has to be from further structural studies. If the ordered-binding model gains more experimental support, a revision of the current concepts of unisite catalysis and negative cooperativity of nucleotide binding will be necessary.

Enzymatic synthesis of ATP analogs and their purification by reverse-phase high-performance liquid chromatography.

The enzymatic syntheses of ATP analogs, such as tubercidin 5'-triphosphate, formycin A 5'-triphosphate, and etheno-ATP, from their respective mono- and diphosphate are described. The reaction products were purified by reverse-phase HPLC using a C-18 matrix and a volatile mobile phase at pH 7, with tributylamine as the ion-pairing agent. Each of the analogs required a buffer of somewhat different composition for the baseline separation of reaction product and reactants. The elutions were isocratic and allowed several successive runs without any intermediate equilibration of the column. After freeze-drying of the pooled fractions, the yield of the synthesized nucleoside triphosphate was approximately 70%. The described procedures are applicable either for analytical investigations or for semi-preparative purposes.

The effects of sulfite or nitrate on turnover-dependent inhibition in the ATPase from Halobacterium saccharovorum are related to the binding of the second metal ion.

The turnover-dependent inhibition of the Halobacterium saccharovorum ATPase is dependent on two parameters: pH and the concentration of the divalent cation present. At pH 6 and 1 mM Mn2+ the inhibition is small, but increases steeply with 6 mM Mn2+. In contrast, at pH 8.5 the inhibition is more than 90% at 1 mM Mn2+, and higher concentrations have little additional effect. A relationship between the occupation of a second metal ion binding site and turnover-dependent inhibition was postulated previously [Schobert, B. (1992) J. Biol. Chem. 267, 10252-10257]. The results lead to a model where this site (X-) can alternatively bind protons (XH), depending on the pH and the free metal ion concentration. The pKa of XH is estimated to be 9. The turnover-dependent inhibition is diminished by bisulfite, whereas sulfite is ineffective. The kinetics show that bisulfite and metal ion compete for the same site. In the proposed model, bisulfite binds via its negative charge to the site from which Pi was released and is arranged such as to interact with X- via its protonated group (X-HSO3-). In this way, formation of the inhibited enzyme species XMe is prevented. Inhibitory anions like nitrate, which carry a permanent dipole as a common feature, show uncompetitive inhibition vs metal ions. The data are compatible with a model in which these inhibitors bind to the vacant Pi site and position their positive charges near XH. As a consequence, the pKa of XH is decreased and X- is stabilized, which in turn favors the formation of XMe. The downshift in pKa was calculated to be 0.7 pH unit.

The catalytic site is located on subunit I of the ATPase from Halobacterium saccharovorum. A direct photoaffinity labeling study.

Nucleotide-binding sites of the ATPase from the halophilic archaebacterium Halobacterium saccharovorum were labeled by ultraviolet irradiation in the presence of [alpha-32P]ATP. A high-affinity site, located on subunit I (98 kDa), was identified as catalytic by the following criteria: ATP bound to subunit I was hydrolyzed and the cross-linked nucleotide was ADP; the specificity for ATP or ADP compared to that of other nucleotides was high; the tightly bound radionucleotide was exchangeable in the presence of excess unlabeled ATP and Mg2+; photolabeling of this site and enzyme inhibition due to tightly bound ADP were both dependent on the presence of Mg2+ and showed identical Kd values; treatment that restored the activity of the ADP-inhibited enzyme also led to the release of the tightly bound nucleotide from subunit I. In addition, a non-catalytic nucleotide-binding site was found, located on subunit II (71 kDa). This site did not hydrolyze ATP, its occupation was Mg2+ independent and the affinity for ATP and the nucleotide specificity were much lower than that of subunit I. We suspect that this site is nonspecific. These results indicate that H. saccharovorum ATPase is different from F1-ATPases which contain the catalytic site on the second largest subunit, but may be similar to other archaebacterial and vacuolar ATPases.

The binding of a second divalent metal ion is necessary for the activation of ATP hydrolysis and its inhibition by tightly bound ADP in the ATPase from Halobacterium saccharovorum.

A binding site for divalent metal ions on the ATPase from Halobacterium saccharovorum is proposed. This site is different from the catalytic site which binds ATP and a complexed divalent metal ion. Occupation of the second site greatly stimulates the rate of ATP hydrolysis and the affinity of the catalytic site for the metal ion-ATP complex. The time-dependent inhibition of the ATPase, which occurs during catalysis and which is known to be caused by the retention of ADP, is also dependent on the occupation of this metal ion binding site. The binding of the metal ion apparently induces extremely tight binding of ADP after the departure of Pi. Mg2+, Mn2+, Zn2+, Co2+, and Ca2+ were tested and showed both the activating and the inhibitory effects, although their binding constants for ATP and the second metal ion binding site were quite different. The characteristic shapes of the nonlinear ATP hydrolysis curves obtained with different metal ions, and different ratios of metal ion and ATP, could be explained with the established dissociation constants. On this basis, a model for the ATPase was developed with two catalytic cycles: one in which the second metal ion binding site is occupied, and another in which it is empty. These pathways are connected by metal ion-dependent equilibria.

F1-like properties of an ATPase from the archaebacterium Halobacterium saccharovorum.

For F1-ATPases from mitochondria and chloroplasts, tight binding of Mg2+ and ADP without Pi at a catalytic site had been reported as a cause of enzyme inhibition. The time dependence of this inhibition and the effect of various agents on this process have been described (Du, Z., and Boyer, P. D. (1990) Biochemistry 29, 402-407, and references therein). Similar results are now reported for the ATPase from Halobacterium saccharovorum. The nonlinear hydrolysis kinetics were modulated by nitrate, azide, sulfite, GTP, ADP in the absence of ATP, or Pi in characteristic ways, in good analogy with the effects of these agents on F1 enzymes. The similarity to the F1 systems suggests that it is tight ADP binding that is affected. Although these reactions of the H. saccharovorum ATPase occurred on different time and concentration scales than those of F1-ATPases, the two systems do not appear to be fundamentally different. The hydrolytic mechanism of the H. saccharovorum ATPase thus identifies this enzyme as a member of the F0F1-ATPase family.

Hysteretic behavior of an ATPase from the archaebacterium, Halobacterium saccharovorum.

The membrane-bound detergent-activated ATPase from Halobacterium saccharovorum was purified at a physiological salt concentration (4 M NaCl) in the presence of nonionic detergents. The preparation contains putative subunits of 110, 71, 31, 22, and 14 kDa. The enzyme activity required high salt concentration but was not dependent on any one specific monovalent cation or any anion. The hydrolysis of ATP was nonlinear with time; the data are consistent with a kinetic model where the enzyme is irreversibly converted from an initial into a final stable form during the first few minutes of the reaction. The model thus contains a rate constant (k) for the transition and hydrolytic rates, v1 and v2, for the two forms of the enzyme. We found that this hysteretic behavior was influenced differently by various conditions and inhibitors. The constant k was smaller with Mn2+ than with Mg2+ as the divalent cation, showed negative temperature dependence, and a distinct pH optimum between 7.5 and 8.5. Thiols decreased k, but nitrate, a specific inhibitor of archaebacterial ATPases, increased it. ADP showed competitive inhibition against both the initial and the final form of the enzyme. Nitrate reversibly inhibited only the latter and in a manner dependent on whether Mn2+ or Mg2+ was used. The kinetic data suggest that all agents tested, with effects on the hydrolytic activity, seem to act at or near the catalytic site of the enzyme.

Structure and orientation of halorhodopsin in the membrane: a proteolytic fragmentation study.

Halorhodopsin (HR), the light-driven chloride pump in halobacteria, was digested with various proteolytic enzymes. As expected, carboxypeptidase A removed 14 amino acids from the C-terminal tail of detergent-solubilized HR, producing a fragment of 25.2 kd in size. Membrane-associated HR could be digested as well, but not in right-side-out sealed cell envelope vesicles. We conclude, therefore, that the orientation of HR in the cytoplasmic membrane is such that the C-terminal tail faces the cytoplasmic side. Tryptic digestion of detergent-solubilized HR resulted in the removal of the same C-terminal segment, but also in the production of two more cleavage products (molecular masses of 20.9 and 16.8 kd respectively). These cleavage sites were determined by amino acid sequencing of the newly produced N termini, and they turned out to be within interhelical loops in an earlier proposed structural model for HR. Incubation with chymotrypsin and thermolysin yielded different sites of cleavage, but also in regions which were proposed to be accessible on the surface of the protein. Since the results show that three of six proposed interhelical loop segments contain proteolytic digestion sites, they support the proposed structural model for HR.

Effects of arginine modification on the photocycle of halorhodopsin.

Exhaustive reaction with phenylglyoxal removed 9 of the 12 arginine and 1 of the 2 lysine residues in detergent-solubilized halorhodopsin, without affecting the chromophore. The consequences of this extensive removal of positive charges on various chloride-binding equilibria and the photochemistry were evaluated. No significant effects were seen on the affinity of Site I to chloride and on the increase in the pKa of Schiff-base deprotonation, which is caused by the chloride binding at this site. No significant effects were seen on the affinity of Site II to chloride, either. However, the photocycle of the pigment was affected. Kinetic modeling of the observed changes in flash-induced absorption changes suggests that the modification increases the affinity of the main halorhodopsin photointermediate to chloride by about fourfold. If chloride translocation involves release of chloride from this intermediate during the transport cycle, the result might explain the observed partial inhibitory effects on chloride transport. Plausible models of chloride translocation include reversible binding of the anion by positively charged groups, strategically arranged in the protein. The results indicate that two of the three spectroscopically observable chloride-dependent equilibria do not depend on a large number of positively charged residues in the protein. To the extent that the unaffected equilibria represent association and dissociation which occur during chloride translocation, at least part of the chloride translocation might be accomplished with the participation of only a few positively charged residues.

Electrostatic interaction between anions bound to site I and the retinal Schiff base of halorhodopsin.

The influence of different anions on the deprotonation of the retinal Schiff base of halorhodopsin in the dark was investigated. We find that a large number of anions cause a significant increase of the pKa of the Schiff base, an effect attributed to binding to "site I" on the protein. The concentration dependencies of the spectroscopic shifts associated with the changes of the pKa yielded dissociation constants (and thus binding energies) for the anions, which were related to the Stokes radii. The data fit the predictions of electrostatic interaction between the anions and the positive charge associated with site I, if the latter is located within a few angstroms from the surface of the protein. The specificity of site I toward various anions is quantitatively explained by the differences in the change of Born energy upon transfer of the anions from water to the binding site. The changes in the deprotonation energy of the Schiff base upon the binding of anions, delta delta Gdeprot, could be calculated from the delta pKa at infinite anion concentration. Unexpectedly, the delta delta Gdeprot values were remarkably close to the energies of binding to site I. Thus, site I and the Schiff base are strongly electrostatically coupled, either because of close proximity or because of the possibility of allosteric energy transfer between them.

An NADH:quinone oxidoreductase of the halotolerant bacterium Ba1 is specifically dependent on sodium ions.

The rate of NADH oxidation by inverted membrane vesicles prepared from the halotolerant bacterium Ba1 of the Dead Sea is increased specifically by sodium ions, as observed earlier in whole cells. The site of this sodium effect is identified as the NADH: quinone oxidoreductase, similarly to the other such system known, Vibrio alginolyticus (H. Tokuda and T. Unemoto (1984) J. Biol. Chem. 259, 7785-7790). Sodium accelerates quinone reduction severalfold, but oxidation of the quinol, with oxygen as terminal electron acceptor, is unaffected. The sodium-dependent pathway of quinone reduction exhibits higher apparent affinity to extraneous quinone (Q-2) than the sodium-insensitive pathway, and is specifically inhibited by 2-heptyl-4-hydroxyquinoline N-oxide. ESR spectra of the membranes contain a feature at g = 1.98 which is tentatively identified as one originating from semiquinone. This feature is increased by NADH and decreased by addition of Na+, suggesting that, as proposed from different kinds of evidence for the V. alginolyticus system, sodium affects the semiquinone reduction step. As in the other system, the site of sodium stimulation in Ba1 probably corresponds to the site of sodium translocation, which was shown earlier (S. Ken-Dror, R. Shnaiderman, and Y. Avi-Dor (1984) Arch. Biochem. Biophys. 229, 640-649) to be linked directly to a redox reaction in the respiratory chain.

Effects of anion binding on the deprotonation reactions of halorhodopsin.

The retinal Schiff base of halorhodopsin deprotonates with a pKa of 7.4 in 0.5 M Na2SO4 in the dark. In the presence of various anions, such as chloride or nitrate, etc., the pKa is raised by up to 1.5 units. Analysis of the dependency of the pKa on anion concentration favors the model in which the anions do not bind to the positively charged Schiff base nitrogen, but to a site near it, and exert their effect on the pKa by direct (perhaps electrostatic) interaction. Adding nitrate, or one of several other anions, causes also a small blueshift in the visible absorption band of the chromophore. These effects on the pKa and the absorption band define an anion binding site in halorhodopsin, termed Site I. Chloride and bromide apparently bind in addition to another site, which is associated with a small red-shift of the absorption band and changes in the photocycle. This other anion binding site is termed Site II. Illumination of halorhodopsin samples results in the deprotonation of the Schiff base with a much lowered pKa, but at very low rates probably determined by the generation of a deprotonating photointermediate. Binding of Site I anions increases the pKa of deprotonation in the light also. The similarity of the responses of the apparent pKa in the dark and in the light to anion concentration suggests that anion binding to Site I influences deprotonation of the Schiff base similarly in the photointermediate and in the parent halorhodopsin molecule.

Evidence for a halide-binding site in halorhodopsin.

In attempting to describe a halide-binding site in halorhodopsin (P580), a light-driven chloride pump in halobacterial membranes, we have investigated the effects of chloride and bromide on flash-induced absorption changes in this pigment, and studied the effects of a diuretic drug, MK-473, on the photochemistry and the transport. We find that at high sulfate or phosphate concentrations, but in the absence of halide, the principal photointermediate is P660, whose half-life is about 1.5 ms. When chloride or bromide are added, the production of P660 is depressed and its half-life becomes longer (up to approximately 10 ms). With increasing halide concentration, the cycle proceeds more and more via the alternative photointermediate, P520, whose half-life varies with the halide concentration in a fashion similar to that of P660. Transport activity, measured during sustained illumination, increases in a manner parallel to the accumulation of P520 up to about 400 mM halide, but declines at concentrations above this value. The transport is inhibited by MK-473 with competitive kinetics, and the effects of this inhibitor on the photocycle are also consistent with displacement of halide ions from their binding site. The observations reported here suggest that chloride and bromide bind to P580, P660, and P520, and that this binding is to a distinct site on the protein.

Halorhodopsin is a light-driven chloride pump.

Light-dependent membrane potentials, ionic fluxes, and volume changes were measured in two kinds of Halobacterium halobium cell envelope vesicles: one containing bacteriorhodopsin and another halorhodopsin. Bacteriorhodopsin-containing vesicles extruded protons by a primary electrogenic mechanism and an energized volume decrease was observed. This was shown to be the consequences of sodium extrusion via proton/sodium antiport (which recirculated protons) and the accompanying passive chloride extrusion. Halorhodopsin-containing vesicles, in contrast, exhibited a volume increase during illumination, apparently caused by primary inward transport of chloride, and accompanied by passive cation (sodium or potassium, and proton) uptake. It was demonstrated that the chloride transport will occur against both electrical and concentration gradients across the vesicle membrane. Moreover, chloride was required on the vesicle exterior for the light-dependent generation of membrane potential, pH change, and swelling. These observations are inconsistent with an earlier proposal that halorhodopsin is an outward directed sodium pump, but suggest very strongly that it is an inward directed chloride pump. Quantitative arguments from the present work rule out a significant role of sodium in the functioning of halorhodopsin.

Spin label studies on osmotically-induced changes in the aqueous cytoplasm of Phaeodactylum tricornutum.

The effects of hyperosmotic stress and adaption on the aqueous cytoplasm of Phaeodactylum tricornutum have been studied with spin labels using 0.2M external Ni2+ to obtain spectra solely from labels within the cells. From partitioning of the TEMPO spin label between the internal aqueous phase and the membrane it is found that the internal volume of the cells decreased by approx. 50-60% in media of high osmotic strength (1.9 osmol/l). During the accumulation of proline in the cells (8.8 mg/ml packed cells) on incubation in the medium of high osmolarity for 3 days, the recovery of the volume was 80%. Further addition of proline to the medium resulted in an increase in the proline concentration in the cells (12.2 mg/ml packed cells) and a recovery in volume of 90%. Cells incubated in the absence of any nitrogen source showed very little recovery and were in a stressed state even in the absence of an osmotic gradient. From the rotational correlation times of the TEMPONE spin label it was found that the effective microviscosity in the cytoplasm of normal cells (approx. 3-8 cP) was considerably higher than that of the external medium (1 cP) and increased 1.5-2-fold under high osmotic stress (1.9 osmol/l). Adaption during the accumulation of proline only decreased the effective microviscosity by approx. 50% of the stressed-induced increase, a considerably smaller recovery than that of the cell volume.

Production of Hexanal and Ethane by Phaeodactylum triconutum and Its Correlation to Fatty Acid Oxidation and Bleaching of Photosynthetic Pigments.

In a light-dependent reaction (3.5 kilolux) at pH 5, the evolution of hexanal, ethane, and ethylene has been established with cell suspensions of the diatom, Phaeodactylum tricornutum. During this process, chlorophyll and carotenoids are partially bleached. Addition of 25 millimolar alpha-linolenic acid or 12 millimolar docosahexaenoic acid yield total pigment destruction and enhancement of ethylene and ethane formation (by about 150 and 7,600%, respectively), whereas hexanal production decreases by 70%. Eicosapentaenoic acid, the major polyunsaturated fatty acid in diatoms, stimulates both ethane and hexanal formation (by about 1,400 and 130%, respectively), but reduces ethylene production (by about 60%). This competition suggests that the production of the volatile compounds is closely connected, although hexanal and ethylene obviously possess different unsaturated fatty acids as precursors. Both the kind of the fatty acids and their relative amounts seem to determine the pattern of the evolved hydrocarbons. The presence of 10 millimolar propylgallate inhibits the evolution of the volatile compounds by about 80%, indicating that radical formation might play a key role in this light-dependent cascade of reactions.