Photophosphorylation as Function of A D P Concentration at Varying Transmembrane Proton Gradients Georg H einen and Heinrich Strotmann In stitu t für B iochem ie d e r P flanzen, H ein rich -H ein e-U n iv ersität D ü sseld o rf, U n iv e rsitä tsstraß e 1, D -4000 D ü sse ld o rf, B u n d esrep u b lik D eutschland Z . N a tu rfo rsch . 44c, 473—479 (1989); received Ja n u a ry 16, 1989 D edicated to P rofessor A c h im Trebst on the occasion o f his 60th birthday Iso la te d S pinach C h lo ro p lasts, P ho to p h o sp h o ry latio n , P roton-M otive F o rce, E nzym e K inetic C o n sta n ts R a te s o f p h o to p h o sp h o ry latio n w ere m easured at co n stan t sa tu ratin g p h o sp h a te c o n c e n tra tio n , varying A D P c o n c e n tra tio n , and varying light intensity. A s the tran sm e m b ra n e p ro to n g ra d ie n t is d e cre ased by p h o sp h o ry la tio n to d ifferent extents dep en d in g on the c o n ce n tra tio n o f A D P . rates o f A T P fo rm atio n o b tain e d at the d ifferent A D P c o n cen tratio n s w ere p lo tte d versus th e actual stead y sta te A p H (in the absence o f Ai|>) during the course o f the re ac tio n . A pH w as m o n ito re d by th e c alib ra ted 9-am inoacridine fluorescence technique. In secondary plots p h o sp h o ry la tio n as function o f A D P c o n ce n tra tio n at d ifferent constant A pH values w ere o b tain e d . T he results indicate M ichaelis-M enten kinetics. T he tru e K m for A D P is virtually u n affected by A p H w h ereas (a t A D P sa tu ra tio n ) strongly dep en d s on A pH . T he results are discussed in the fram ew o rk of a sim ple enzym e kinetic m odel which considers the in trath y lak o id al p ro to n (at c o n stan t ex tern al p H ) as a th ird su b stra te for A T P form ation. T he m odel is capable o f explaining th e re p o rte d resu lts as well as a variety o f im p o rta n t results from the lite ratu re. Introduction Photophosphorylation is catalyzed by a protontranslocating ATPase of the F0F] type which is a con­ stituent of the thylakoid membrane of the chloro­ plast. The mechanism of A|iH+-coupled phosphoryl transfer to A D P is a subject of intensive investiga­ tion. As the catalytic site is located on ß subunit or at the interface between a and ß [for review see 1—3], three catalytic entities must be assumed to be present per one A TPase molecule because the num ber of a and ß copies per enzyme is three [4—5]. Nevertheless Michaelis-M enten kinetics are usually observed when phosphorylation is measured as function of A D P or phosphate concentration. This fact may be explained within the framework of the three-site binding change mechanism propos­ ed by Boyer and his coworkers [6]. This hypothesis assumes that the three catalytic sites change their properties in a sequential mode, so that at a given mom ent only one of them is in a condition to take up the substrates. Substrate binding to one site faciliA b b revia tio n s: C hi, c h lorophyll; D C M U , 3-(3 ',4 '-d ich lo ro p h e n y l)-l,l-d im e th y lu re a ; PM S, phenazine m ethosulfate; T ric in e, N -[tris(hydroxym ethyl)m ethyl]glycine. R e p rin t re q u ests to P rof. D r. H . S trotm ann. V erlag d e r Z eitsc h rift fü r N a turforschung, D-7400 T übingen 0341 -038 2 /8 9 /0 5 0 0 - 0473 $ 0 1 .3 0 /0 tates product release from another site. A t very low A D P concentration (< 1 ^im), however, a deviation from monophasic Michaelis-Menten behaviour was observed [7], a result which was interpreted to indi­ cate the operation of a single site in this concentra­ tion range of A D P. Although im portant for the elu­ cidation of the catalytic mechanism, single site catalysis can be disregarded in a physiological sub­ strate concentration range. Manyfold determ ination of K m for A D P are re­ ported in the literature with rather diverging results. A systematic investigation showed that K m increases with increasing light intensity [8—10], decreases with increasing concentration of electron transport in­ hibitors like DCM U [8], increases or decreases in the presence of an uncoupler depending on the concen­ tration employed [9, 12]. Different conclusions were drawn from these experim ental findings. D avenport and McCarty [11] were right in criticizing that all these m easurem ents suffered from uncontrolled ex­ perim ental conditions with regard to A|iH+ during the phosphorylation reaction. Ap,H* is a function of the rate of electron transport, as well as the rates of pas­ sive and productive proton efflux. With increasing substrate concentration we can expect a progressive decrease of the electrochemical proton gradient even at constant light intensity, since the rate of phos­ phorylation is dependent on the concentrations of Unauthenticated Download Date | 10/2/16 3:48 PM 474 G . H ein en and H . S tro tm an n • P hoto p h o sp h o ry latio n as Function of A D P C o n c en tratio n the substrates. As electron transport is controlled by intrathylakoidal pH [12, 13] varying extents of accel­ eration of electron transport must be taken into ac­ count, and the rate of passive proton leakage which is a function of the internal proton concentration will be changed, too. Analysis of the effect of A D P [10] and phosphate concentrations [14] on the magnitude of steady state ApH indeed verified these predic­ tions. The goal of this study is the determ ination of the kinetic constants of the H*-ATPase at controlled ApH (in the absence of Aty) and to give an answer to the question w hether the true K m is constant or variable with ApH. This decision has an im portant implication for the understanding of the enzymatic mechanism. The experim ental approac!' includes m easure­ ments of photophosphorylation ai varying light in­ tensity or uncoupler concentration, respectively, and varying A D P concentration. By simultaneous regis­ tration of 9-aminoacridine fluorescence which was calibrated for the employed experimental conditions [15, 16], the rates of ATP form ation could be related to the actual ApH. K m and Vmax at constant actual ApH were ascertained by secondary plots. M ethods Thylakoids were isolated from spinach leaves as in ref. [17]. All m easurem ents were conducted in a cylindrical glass cuvette which was placed in a self-constructed fluorom eter [16]. The experim ental set-up perm itted easy addition of substrates and taking of aliquots for analysis. During the whole experim ent the fluores­ cence of 9-aminoacridine was recorded as a measure of ApH. The final cuvette volume was 2.5 ml, the tem pera­ ture 20 °C. The medium consisted of 25 mM Tricine buffer, pH 8.0, 50 mM KC1, 5 mM MgCl, 10 mM dithiothreitol, 50 piM PMS, 50 nM valinomycin, 10 mM glu­ cose and 30 units/ml hexokinase (salt-free, Sigma). A fter addition of thylakoids corresponding to 25 ^ig chlorophyll/ml, 5 ^im 9-aminoacridine was injected. Then full light was given for 2 min to bring the A T P ­ ase in its thiol-m odulated state. A fter dark relaxation of the fluorescence signal, the indicated concentrations of A D P were added. Since A D P caused some instantaneous artificial fluorescence quenching [15], the fluorescence emis­ sion obtained after A D P addition minus basal fluo­ rescence (absence of 9-aminoacridine) was used as standard (<J>0) for calculation of ApH. Subsequently, light of varying intensity was turned on to produce different initial ApH values. 2 min later, after reaching a steady fluorescence, photo­ phosphorylation was started by the addition of 32Plabeled phosphate. Initiation of ATP synthesis causes a decrease of ApH. The steady state fluores­ cence during phosphorylation (<I>) was used for cal­ culation of the actual ApH. For analysis of the formed 32P-labeled glucose-6-phosphate 0.15 ml samples were taken 15, 30, 45 and 60 s after addition of [3: P]Pj and deproteinized by H C104 (0.5 m final concentration). Organic 32P was separated from in­ organic 32P by precipitation [18] and m easured in a scintillation counter. The calculation of ApH was based on the calibra­ tion perform ed in ref. [16] under the same experi­ mental conditions. Because of the observed linear relationship between log («J^- ^ ) / ^ and ApH in the relevant ApH range between 2.5 and 3.5 [15, 16], the equation by Schuldiner et al. [13] could be employed. From the calibration curve [16] an apparent internal thylakoid volume of 30 ^il/mg chlorophyll was esti­ m ated. The empiric formula for calculation of ApH under the employed experimental conditions was ApH = log (<E>0—<t»)/<I> + 3.12. Results Fig. 1 shows rates of photophosphorylation as a function of the actual ApH at 6 different A D P con­ centrations ranging from 10 to 152 ^im. The data are taken from a large num ber of single independent ex­ periments; in every individual experiment up to 20 points were gained. As the activity of phosphoryla­ tion varied between the experiments, the rates were normalized by relating them to mean maximal rates. For every ADP concentration in every experim ent Vrnax values at saturating ApH were estim ated by extrapolation. For this purpose reciprocal rates (1/v) were plotted versus reciprocal internal proton con­ centration to the power n ([H +in]"). The exponent n was chosen, so that the coefficient of determ ination (r2) of a linear regression attained its maximum. Usu­ ally n was between 2.5 and 3. In order to ensure comparability of the ApH curves at different ADP concentrations, ApH was varied in one experim ent with not less than two ADP concentrations in an overlapping mode. Unauthenticated Download Date | 10/2/16 3:48 PM G . H ein en and H . S tro tm an n • P h o to p h o sp h o ry la tio n as Function o f A D P C oncentration 475 Fig. 1. R a te o f ph o to p h o sp h o ry latio n at 6 dif­ ferent A D P c o n cen tratio n s as function o f the actual steady state A pH m easu red during the course of the reaction. ApH was also changed by addition of varying con­ centrations of the uncoupler nigericin at constant (maximal) light intensity. The results (not shown) indicate that these data fit into the curves of Fig. 1. This finding which could be expected on the basis of the chemiosmotic theory, dem onstrates that solely the magnitude of ApH is critical, not the m ethod of its manipulation. Rates of ATP form ation as a function of ADP concentration at constant ApH values are shown in Fig. 2. The points were either taken directly as single or mean values from Fig. 1 or gained by linear inter­ polation between vicinal experimental points. Be­ cause of the low rates, a fair evaluation of data below pH = 2.6 is impossible. In a reasonable ApH range the plots of phosphorylation rate versus A D P con­ centration are essentially hyperbolic; hence Michaelis-M enten kinetics may be presupposed. In order to determ ine Vmax and K m the graphs were linearized according to Lineweaver-Burk (1/v versus 1/[ADP]), Eadie-Hoffstee (v versus v/[ADP]) and Hanes-W oolf ([ADP]/v versus [ADP]). The three modes of evaluation give different emphasis to the single experim ental points, depending on the ADP concentration and the obtained rates. Therefore ex­ perim ental scattering affects the determ ination of the kinetic param eters by the three methods in different ways. As the confidence of the estimates is restricted by the relatively low num ber of A D P concentrations employed, the evaluation by different plots may give o U) E (X 5 HM ADP Fig. 2. R ate of p h o to p h o sp h o ry latio n at different actual steady state A pH values as function of A D P c o n cen tratio n . T h e experim ental points w ere tak e n from Fig. 1. T he curves w ere com puted by the m odel p re sen te d in the D is­ cussion. additional certainty. To exclude any subjectivity in preparation of the plots, the graphs were computed from the experimental data by linear regression. The results are summarized in Table I. As expected, Vmax increases progressively with rising ApH. Between ApH 2.6 and 3.0 maximal velocity increases by a factor of about 7. On the other hand, K m is virtually unaffected by the magnitude of ApH in this range. Unauthenticated Download Date | 10/2/16 3:48 PM 476 G. H ein en and H . S tro tm an n • P h o to p h o sp h o ry la tio n as F unction o f A D P C o n c en tratio n T ab le I. D e term in a tio n o f Vmax and ATm(A D P ) for p h o to p h o s­ p h o ry latio n at different actual A pH values. T he d ata w ere taken from Fig. 1 and p lo tted according to L inew eaver-B urk (A ). E adieH o ffste e (B ), and H an es-W o o lf (C ). T he kinetic p aram eters w ere gained from lin ear regression analysis of the experim ental points. yy max [ixmol A T P /m g Chi p e r h] K m(A D P ) [[am] A ctu al A pH A B C m ean A B C m ean 2.60 2.65 2.70 2.75 2.80 2.85 2.90 2.95 3.00 55.6 66.5 60.3 63.5 61.8 70.8 55.5 56.1 50.9 44 37 46 42 62 63 71 66 67 76 88 76 94 96 102 97 131 123 169 141 209 184 217 203 196 205 238 213 270 247 303 274 265 277 351 298 58.3 60.9 48.1 49.3 54.4 72.9 47.9 55.4 42.1 43.6 62.5 56.9 51.3 46.9 59.6 50.9 45.7 43.1 65.0 76.1 75.2 56.8 84.1 79.8 67.6 67.3 67.4 The single estimates of K m obtained by the three different plots vary by a factor of about 2. However, they reveal no clear-cut increasing or decreasing tendency in dependence of ApH. Hence we conclude that ApH in the indicated range from 2.6 to 3 does not affect the Michaelis constant significantly. The reciprocal values of Table I were plotted versus the reciprocal internal H + concentrations to the pow er n, [H +jn]". The value of n was chosen so that the plot gave a linear approximation. Linear re­ lationships were found at values of n between 2.5 and 3, Fig. 3 shows a fit perform ed with n = 2.6. n is identical with the Hill coefficient for H +in which can be classified as the third substrate of photophos­ phorylation according to A D P + Pi + n H +in -> ATP + H 20 + n H +out. The reaction can be written as a unidirectional process, because A TP is recycled to ADP by the subsequent hexokinase reaction. The maximal veloc­ ity obtained at saturation of A D P and Pj as function of [H +in] can be expressed by the Michaelis-Menten equation as 1 / (M t-Tjn)2'6 Fig. 3. L in ew eav er-B u rk plot o f Vmax (sa tu ratin g A D P co n ­ c en tratio n ) versus in te rn al H + co n ce n tra tio n . T he p ro to n c o n ce n tra tio n is e xpressed in M 26 (see tex t). T he results w ere taken from T ab le I. T he points are m ean values of the bars indicate the lim its o f variation. V'max = [Et] •kc ■[H +in]”/([H +in]" + K (H +in)), where [Et] is the total enzyme concentration, kc is the rate-lim iting velocity constant and K (H +in) means an equilibrium constant for the internal proton with the dimension (mol/1)". Theoretically n is equal to the stoichiom etry of H + translocated through the A TP­ ase complex per ATP formed. The approximation towards 3 is in agreem ent with determ inations of H +/ ATP stoichiometry obtained by different experim en­ tal approaches [15, 19-22], The plot 1/Vmax versus l/[H +in]" permits the estimation of the maximal rate of ATP form ation at saturation of all involved sub­ strates. This param eter (V abs) was determ ined as 1000 to 2000 nmol/mg chlorophyll per h, depending on the chosen value for n , which agrees well with the reported turnover num ber of the enzyme [23], In Unauthenticated Download Date | 10/2/16 3:48 PM 477 G . H ein en and H . S tro tm an n • P h o to p h o sp h o ry la tio n as Function of A D P C o n cen tratio n Fig. 3, Kabs was extrapolated to 1250 [imol/mg Chlorophyll per h. The received value of K (H Tln) per­ mits the calculation of a half-maximal ApH of about 3.1. Discussion The here reported results show that the Michaelis constant for A D P is largely independent of ApH in a range from 2.6 to 3.0 when m easured at phosphate saturation and controlled thylakoid energization, i.e. at a constant actual pH gradient during variation of the A D P concentration. The evaluation proposes Michaelis-Menten type kinetics, an assumption which is based on the hyperbolic dependence of ve­ locity on A D P concentration in the employed range. We have to admit considerable experimental error due to scattering in ApH determ ination and the utili­ zation of results from diverse independent experi­ ments. Nevertheless, the large num ber of data and their treatm ent by different secondary plots confers some confidence on the determ ination of the kinetic param eters. The evaluated ApH range from 2.6 to 3.0 appears rather small; however, in this range the activity of phosphorylation increases by almost one order of m agnitude, which should be sufficient to ensure possible substantial changes in K m. Since no clear-cut tendency of an increase or decrease of could be detected in dependence of ApH, it appears highly probable that the true K m is ApH-independent. The same presumption was made by Quick and Mills [10] in explaining the results on apparent K m (A D P) at uncontrolled ApH on the basis of a reason­ able theoretical model. In variance to their assump­ tion (K m = 12 |j ,m ) , however, our data indicate a true K m which is 5 times higher (59 |j .m ) . In a formal enzyme kinetic model the internal H + can be considered as a kind of substrate in phos­ phorylation when the external H + concentration is kept constant as in the here conducted experiments. Invariability of ATm(A D P) at F max being increased by raising [H +in] may be interpreted by conceiving H +jn as an “inverse” non-com petitive inhibitor. In analogy to the formalism of non-competitive inhibition, H +in would act on a deenergized free enzyme to create an energized enzyme which is ready to bind substrate A D P, as well as on a deenergized form of the en­ zyme with bound A D P to create the energized enzyme-ADP complex. The active enzyme-ADP com­ plex interacts with phosphate in a reversible reaction to form the enzyme-product complex. In the model shown below, it is assumed that the active enzymeATP complex can also undergo transfer to a deener­ gized complex and that the reverse reaction depends on [H+in], too. The final reaction, the dissociation into the free enzyme and free A TP, can be written as an irreversible step since the product ATP is trapped by hexokinase. As a consequence of product release, the energized state of the enzyme is discharged. The model is a different view of a conformational cou­ pling hypothesis which is basically not far from the mechanism proposed by Boyer and his colleagues [ 6], E9 . nH+h/fi ADP k2 P, k3 ...- Ee ADP ^-----flHV, nH+*k, Ed-ADP E.-ATP /r_t E„+ATP /f_i Ed-ATP The respective energized and deenergized forms are designated by the indices e and d, the num ber of protons reacting from the inner phase of the thy­ lakoid is n. The value of n is assumed to be 3. Com­ putation of the model yields for the param eters Vmax and ATm(ADP): Kmax = E, • [H +in]" • k m k y k j ( [ H + in]'' • k\ + f c - O t t P i f o + * - 3 + * 4 ) and Km(A D P) = ( k . 2k - 3 + k^_k4 + [P;]*3*4)/ (£2[Pj]^3 + kik-T, + k2k4). The model which is similar but not identical to a scheme proposed by Quick and Mills [10, 14] is cer­ tainly incomplete. So, it does not include site cooperativity and explains ATP synthesis but not hydrolysis without making additional assumptions. Nevertheless, the model is capable of explaining a variety of im portant results reported in the literature as well as the findings of the present study. (1) The independence of the true /Cm(A D P) as well as the dependence of F max on the actual ApH or [H +in], respectively, is fulfilled by the model. As in the scheme of Quick and Mills [10], variability of the apparent K m obtained at different light intensities or uncoupler concentrations can be explained, too. (2) Energy of the electrochemical gradient is not invested directly in the chemical reaction at the Unauthenticated Download Date | 10/2/16 3:48 PM 478 G . H ein en and H . S trotm ann • P h o to phosphorylation as F un ctio n o f A D P C o n c en tratio n catalytic site [6]. Here the transformation of “deenergized” into “energized” enzyme forms is the only site of energy input. This assumption was also made by Schumann [24] in a comprehensive hypo­ thesis of the H +-coupled chloroplast ATPase. The deenergized species may be conceived as enzyme molecules characterized by “closed” i.e. unavailable catalytic sites, whereas the energized forms have “o pen” catalytic sites which are accessible to the sub­ strates of the medium. If a closed site is containing a nucleotide molecule, this ligand therefore appears “tightly bound” . In variance with the “energy-linked binding change mechanism” [6] the formation of tightly bound A D P or ATP would be a side reaction rather than an obligate interm ediate step of the catalytic sequence. The ratio of ATPase molecules with exchangeable nucleotides to those with tightly bound nucleotides depends on the energy state of the m em brane [8, 25] and the level of tightly bound nu­ cleotides changes inversely with the activity of photo­ phosphorylation as function of energy input [26]. W hen light is turned off, the level of tightly bound nucleotides increases to a maximum of one per A TP­ ase when A D P or ATP is present in the medium [27], Tight binding of A D P is related with inactivation of the enzyme [28]. On the other hand, the tightly bound nucleotides are released or exchanged upon reillum ination, a reaction which is related with en­ zyme reactivation [28]. U nder certain conditions the initial rate of nucleotide release matches the rate of A TP form ation [29, 30]. All these results can be easi­ ly explained by the model. H ence, the reactions re­ lated with enzyme activation are basically identical with the energy transfer reactions involved in the catalytic process and deactivation is a compulsory consequence of relaxation of the proton gradient leading to closure of the catalytic sites. From the standpoint of this model it is unnecessary to assume extra nucleotide binding sites for nucleotide-depend­ ent regulatory processes. This view is in line with the recent finding that tightly bound A D P and ATP in darkened thylakoids are on catalytic sites, a result which was obtained by the photoreactive nucleotide analogue 2-azido-ADP [32], The general structural identity of the sites containing tightly bound nucleo­ tides with those performing catalysis had been sug­ gested before on the basis of photolabeling experi­ ments with the same analogue [31]. Although up to 6 nucleotide-binding sites were identified in isolated CF! [33], the participation of all of them in catalysis and control of the m em brane-bound enzyme has never been proven. (3) It was shown by lsO exchange that decrease of substrate concentration increased the num ber of re­ versible cycles of ATP form ation at the catalytic site before the product ATP is released [34], This experi­ mental fact was concluded to indicate catalytic site cooperativity, i.e. release of the product is facilitated by binding of substrates to a second catalytic site. For the reason of simplicity, site cooperativity was om it­ ted, but could be integrated in the scheme without changing the general features significantly. The same effect on lfsO exchange was reported to occur when light intensity was reduced [35] and this result was taken as evidence for A|IH+ dependence of product release. In contrast, our model proposes that release of ATP from the energized enzyme is a spontaneous reaction. This conclusion which is derived from the formal enzyme kinetic interpretation of the reported results, is in line with the earlier finding that the affinity of the energized enzyme for ATP is by one to two orders of magnitude lower than for A D P [36]. On the energized enzyme, A TP acts as a competitive inhibitor of phosphorylation with a K\ as high as 4 m M [36], indicating that the dissociation equilib­ rium is far on the side of free A TP. In the framework of the model, the effect of low light intensity on 1kO exchange may be interpreted by assuming that the cycles of reversible ATP hydrolysis occur in the state of the enzyme where A TP is tightly bound (EdATP). Actually the exchange of [IS0 ] H 20 with tight­ ly bound ATP was dem onstrated experimentally [37], If k -\ > &4, the equilibrium may be on the side of E d-ATP rather than free A TP at low ApH, thus permitting repetitive lsO exchange with medium [180 ] H 20 . (4) The “kinetic com petence” of tightly bound ATP is an unsolved problem connected with the energy-linked binding change mechanism. In this re­ spect conflicting results were reported [38, 39]. It is quite certain that the exchange of at least part of the tightly bound ATP is too slow to be a step in the catalytic process [37]. In our model the exchange of tightly bound nucleotides is not necessarily in pace with the catalytic cycle. In this context it should be noted that for sake of simplicity the rate constants k\ and k_\ were assumed to be identical for the inter­ conversion of all deenergized to energized enzyme forms. They could well be different for the reversible transitions of E d, E d-ADP and E d-ATP, respectively. Unauthenticated Download Date | 10/2/16 3:48 PM G . H einen and H . S tro tm a n n ■P h o to p h o sp h o ry la tio n as Function o f A D P C oncen tratio n to the corresponding energized forms. Independence of K m on ApH would be fulfilled if the ratios of the forward and backward constants were the same. The curves drawn in Fig. 2 were com puted with the model by chosing the following constants: £ t -&4 = Vabs = 1250 ^imol/mg chlorophyll per h (Fig. 3), Km(A D P) = 59 [am (Table I), (*_3 + k4)/k3 = K m(P{) = 500 ^im [40], and k - X!k\ = 2000 |xm3. The latter value was calculated from the ApH at half-maximal phos­ phorylation rate which was found to be 3.1 (Fig. 3): k - x!kx = (10(31"8)-106)3 = 2000 ^m3. [1] H . S tro tm a n n and S. B ick e l-S a n d k ö tter, A n n u . Rev. P lant Physiol. 35, 9 7 - 1 2 0 (1984). [2] S. M erch an t and B. R. S elm an, P hoto sy n th . R es. 6, 3 - 3 1 (1985). [3] J. M. G alm ich e, G . G ira u lt, an d C. L em aire. Photochem . P h otobiol. 41, 7 0 7 -7 1 3 (1985). [4] K. H . Süss and O . S chm idt, FE B S L ett. 144, 213—218 (1982). [5] J. V. M o roney, L. L o p estri, B. F. M cE w en. R. E. M cC arty, and G . G . H a m m e s, F E B S L ett. 158, 58—62 (1983). [6] P. D. B oyer and W. E . K o h lb re n n er, in: E nergy C oupling in P hoto sy n th esis (B. R. Selm an and S. Selm an -R e im e r, e d s.), pp. 231—240, E lsevier/N orth H o l­ land In c ., New Y o rk 1981. [7] S. D . S tro o p and P. D . B o y er, B iochem istry 24, 2 3 0 4 -2 3 1 0 (1985). [8] S. B ick e l-S a n d k ö tter and H . S tro tm a n n , F E B S L ett. 125, 1 8 8 -1 9 2 (1981). [9] C. V in k ler, B iochem . B iophys. R es. C om m un. 99, 1 0 9 5 -1 1 0 0 (1981). [10] W . P. Q uick and J. D . M ills, B iochim . B iophys. A cta 893, 1 9 7 -2 0 7 (1987). [11] J. W. D a v en p o rt and R. E . M cC arty, Biochim . B io­ phys. A cta 851, 1 3 6 -1 4 5 (1986). [12] U . Siggel, in: P roceedings o f the T h ird In tern atio n al C ongress on P hoto sy n th esis (M . A v ro n , e d .). Vol. I, pp. 645—654, E lsevier, A m ste rd am 1975. [13] S. S chuldiner, H . R o tte n b e rg , and M . A v ro n , E u r. J. B iochem . 25, 6 4 - 7 0 (1972). [14] W. P. Q uick and J. D . M ills, B iochim . Biophys. A cta 932, 2 3 2 -2 3 9 (1988). [15] H . S tro tm a n n and D . L ohse, F E B S L ett. 229, 3 0 8 -3 1 2 (1988). [16] D . L o h se, R. T h e le n , and H . S tro tm a n n . Biochim . B iophys. A c ta, su b m itte d . [17] H . S tro tm an n an d S. B ick e l-S a n d k ö tter, Biochim . B iophys. A cta 460, 1 2 6 -1 3 5 (1977). [18] Y. Sugino and Y. M iyoshi, J. Biol. C hem . 239, 2 3 6 0 -2 3 6 4 (1964). [19] M . R a th e n o w and B. R u m b erg , B er. B unsenges. Phys. C hem . 84, 1 0 5 9 -1 0 6 2 (1980). [20] W . Ju n g e , B. R u m b erg , and H . S chröder. E ur. J. B iochem . 14, 5 7 5 -5 8 1 (1970). [21] A . R. P ortis and R. E. M cC arty, J. Biol. C hem . 249, 6 2 5 0 -6 2 5 4 (1974). 479 It should be mentioned that this value of course relies on the correct quantitative determ ination of ApH in our experiments. The problems related with calibration of the 9-aminoacridine fluorescence method by the phosphate potential in energetic equilibrium are discussed elsewhere [16]. Acknow ledgem ents This work was supported by grants from Deutsche Forschungsgemeinschaft (SFB 189) and Fonds der Chemischen Industrie. The authors thank Mrs. Rita Reidegeld for help in preparing the manuscript. [22] J. W. D a v en p o rt and R. E . M cC arty. J. Biol. C hem . 256, 8 9 4 7 -8 9 5 4 (1981). [23] P. G rä b e r, E . S chlodder, and H. T. W itt, Biochim . B iophys. A cta 461, 426—440 (1977). [24] J. S chum ann, Z . N a tu rfo rsc h ., subm itted. [25] S. B ickel-S andkötter, Biochim . B iophys. A cta 723, 7 1 - 7 7 (1983). [26] H . S tro tm a n n , S. B ick el-S an d k ö tter, U. F ra n e k , and V. G e rk e , in: E nergy C oupling in P hotosynthesis (B. R. Selm an and S. S elm an -R eim er. e d s.), pp. 187—196, E lsev ier/N o rth H ollan d In c ., N ew Y ork 1981. [27] H . S tro tm a n n , S. B ick e l-S a n d k ö tter. an d V. S hoshan. FE B S L ett. 101, 3 1 6 -3 2 0 (1979). [28] J. S chum ann and H . S tro tm a n n , in: P hotosynthesis II. E lectron T ran sp o rt and P h o to p h o sp h o ry la tio n (G . A k o yunoglou. e d .), pp. 881—892, B alaban Int. Sei. Service, P hiladelphia, Pa. 1981. [29] E. S chlodder and H . T. W itt, Biochim . B iophys. A cta 635, 5 7 1 -5 8 4 (1981). [30] H. S tro tm a n n , in: A dvances in P hotosynthesis R e ­ search (C. Sybesm a, e d .). Vol. II, pp. 477—484, M artinus N ijhoff/D r. W. Ju n k P ublishers, T he H ague 1984. [31] J. J. C zarnecki, M . S. A b b o tt, and B. R . Selm an, Proc. N atl. A cad. Sei. U .S .A . 79, 7 7 4 4 -7 7 4 8 (1982). [32] J.-M . Z h o u , Z . X u e, Z . D u , T . M elese, and P. D. B oyer, B iochem istry 27, 51 2 9 -5 1 3 5 (1988). [33] Z . X ue, J.-M . Z h o u , T . M elese. R. L. C ross, and P. D. B oyer. B iochem istry 26, 3749—3753 (1987). [34] D . D. H ack n ey , G . R o se n , and P. D . B o y er, Proc. N atl. A cad. Sei. U .S .A . 76, 3 6 4 6 -3 6 5 0 (1979). [35] S. D. S tro o p and P. D . B oyer, B iochem istry 26, 1479 -1 4 8 4 (1987). [36] U. F ran ek and H . S tro tm a n n , FE B S L ett. 126, 5 —8 (1981). [37] L. T. Sm ith, G . R o se n , and P. D . B oyer, J. Biol. C hem . 258, 10887-10894 (1982). [38] D. J. Sm ith and P. D . B o y er, Proc. N atl. A cad. Sei. U .S .A . 73, 4 3 1 4 -4 3 1 8 (1976). [39] C. A flalo and N. S havit, E u r. J. B iochem . 126, 61—68 (1982). [40] H. S tro tm a n n , S. N iggem eyer, and A .-R . M ansy, in: Progress in P hotosynthesis R esearch (J. Biggins, e d .). Vol. Ill, pp. 29—36, M artinus N ijhoff P ublishers, D o rd re ch t 1987. Unauthenticated Download Date | 10/2/16 3:48 PM