Eur. J. Biochem. 160, (1 986) 0 FEBS 1986

Transcription

1 Eur. J. Biochem. 160, (1 986) 0 FEBS 1986 Differential detergent-solubilizaon of integral thylakoid membrane complexes in spinach chloroplasts Localization of photosystem 11, cytochrome b6-fcomplex and photosystem I Peter J. MORRISSEY, Steven W. McCAULEY and Anastasios MELIS Division of Molecular Plant Biology, University of California, Berkeley (Received June 18,1986) - EJB Progressive solubilization of spinach chloroplast thylakoids by Triton X- was employed to investigate the domain organization of the electron transport complexes in the thylakoid membrane. Triton/chlorophyll ratios of 1 : 1 were sufficient to disrupt fully the continuity of the thylakoid membrane network, but not sufficient to solubilize either photosystem I (PSI), photosystem 11 (PSII) or the cytochrome b,-f(cyt b6-f) Complex. Progressive with the Triton concentration increase (Triton/Chl > 1 : l), a differential solubilization of the three electron transport complexes was observed. (a) Solubilization of the Cyt b6-f complex from the thylakoid membrane preceded that of PSI and apparently occurred early in the solubilization of stroma-exposed segments of the chloroplast lamellae. (b) The initial removal of chlorophyll (up to 40% of the total) occurred upon solubilization of PSI from the stroma-exposed lamella regions in which PSI is localized. (c) The tightly appressed membrane of the grana partition regions was markedly resistant to solubilization by Triton X-. Thus, solubilization of PSII from this membrane region was initiated only after all Cyt b6-fand PSI complexes were removed from the chloroplast lamellae. The results support the notion of extreme lateral heterogeneity in the organization of the electron transport complexes in higher plant chloroplasts and suggest a Cyt b6-f localization in the membrane of the narrow fret regions which serve as a continuum between the grana and stroma lamellae. The thylakoid membrane of higher plant chloroplasts shows distinct differentiation into grana and stroma-exposed lamellae [l]. In the grana, disc-shaped thylakoids are pressed against each other at the partition region. This thylakoid stacking is enabled upon screening of surface charges by divalent Mgz+ ions [2]. Stroma-exposed lamellae are interconnected with the grana via the narrow membrane in the fret region [3] thus helping to form an extensive and continuous thylakoid network. The functional significance of this membrane differentiation is clearly the localization of different photosystems in grana and stroma-exposed regions [4-71. It is now believed that most of PSII is segregated in the thylakoid membrane of the grana partition regions [6-91 whereas PSI is found exclusively in stroma-exposed thylakoids [7]. This lateral separation of PSII from PSI in the thylakoid membrane implies long-distance electron transport from the membrane of the grana partition regions to the stroma-exposed regions of chloroplasts. The average lateral distance of PSII from PSI is of the order of pm. As discussed by other authors [9-111, the nature of the electron shuttle between grana and stroma lamellae (plastoquinone or plastocyanin) largely depends on the localization of the intermediate cytochrome (Cyt) b6-f complex in the thylakoid membrane. Correspondence to A. Melis, Molecular Plant Biology, 31 3 Hilgard Hall, University of California, Berkeley, California, USA Abbreviations. Chl, chlorophyll; PS, photosystem; PTo0, reaction center of PSI; Qa, primary quinone of PSII; Cyt, cytochrome; SDS, sodium dodecyl sulfate. The distribution of the Cyt b6-f complex in grana and stroma thylakoids has been investigated in the past upon fractionation of thylakoid membranes with detergent, mechanical treatment and aqueous polymer two-phase separation. Chloroplast fractionation with Triton X- yielded resolved membranes from the grana partition region [12] that contained little or no Cyt b6-f complex [8] suggesting the exclusion of this complex from the membrane of the grana partition region. Mechanical disruption of chloroplasts by Yeda press, followed by aqueous polymer two-phase separation of inside-out and right-side-out vesicles resulted in a balanced distribution of the Cyt b6-fcomplex in the two types of vesicles [ The measurements with inside-out and right-side-out vesicles suggested even distribution of the Cyt b6-f complex between stacked and unstacked membranes. Immunocytochemical approaches to the localization of the Cyt b6-f complex place it almost evenly distributed in grana and stroma lamellae [ Based on a study with maize mesophyll (grana-containing) and bundle sheath (granalacking) chloroplasts, Ghirardi and Melis [l 11 concluded that both PSI and the Cyt b6-f complex are excluded from the membrane of the grana partition region. They suggested that the Cyt b6-f complex might be localized in a domain of the thylakoid membrane occurring in the vicinity of the PSIIcontaining grana partition regions but not as an integral component of the partition region itself [ll]. Recently, a chloroplast thylakoid membrane fragment was identified, distinct both from that of PSII and from that of PSI, which was enriched in the Cyt bs-fcomplex [19]. It was suggested that such a Cyt &$domain may be localized in the region between

2 390 the appressed and the non-appressed membranes, possibly in the membrane of the fret region [19]. In the present work we have addressed the question of the Cyt bs-f localization in the thylakoid membrane upon investigation of the selective fractionation of the thylakoid membrane by Triton X-. Incubation of chloroplast thylakoids with low Triton concentrations disrupted the continuity of the thylakoid network without solubilizing integral thylakoid membrane complexes. A distinct differential solubilization of the integral electron transport complexes is manifested upon Triton concentration increase. The Cyt b6-f complex is removed from the membrane prior to PSI, whereas the membrane of the grana partition region and PSII show a marked resistance to the detergent. The results are discussed in terms of different domains of the chloroplast thylakoid for the localization of PSII, PSI and Cyt b6-f complexes. MATERIALS AND METHODS Spinach (Spinacea oleracea L.) chloroplast thylakoid membranes were isolated from freshly harvested leaves of hydroponically grown plants. Deveined leaves were ground in a blender in buffer containing 50 mm Tricine/NaOH (ph 7.8), 0.4 M sucrose, 10 mm NaCl and 5 mm MgC12 for 10 s at 4 C. The slurry was filtered through four layers of miracloth. Chloroplasts were isolated by centrifugation of the filtrate at 5000 x g for 5 min. The chloroplast pellet was suspended in buffer containing 50 mm Tricine/NaOH (ph 7.8), 0.4 M sucrose, 10 mm NaCl and 5mM MgC12 and 5mM CaC12, using a Wheaton homogenizer, to a chlorophyll concentration of about 1.5 mg/ ml. Chlorophyll concentrations were determined in 80% acetone using the procedure of Arnon [20]. Triton X- fractionations of the thylakoid membrane system were performed by a modification of the procedure of Berthold et al. [12]. The thylakoid membranes were incubated with Triton X- at 0 C for 30 min. The Triton/Chl (w/w) ratios used ranged from 0.1:l to 34:l. After the incubation, the unsolubilized membranes were precipitated by centrifugation at xg for 30 min. The supernatant was carefully removed and the pellet was resuspended in the same buffer. The chlorophyll concentrations of both the supernatant and resuspended pellet were then determined. Quantification of the electron transport complexes (PSII, PSI and Cyt b6-j) was implemented spectrophotometrically by measuring the concentration of QA, and Cytf, respectively. It is generally accepted that each functional component (QA, P700, CytJ) occurs in a 1 : 1 stoichiometry with its associated complex (PSII, PSI, Cyt b6-a respectively). Measurement of the concentrations of QA and P700 were obtained from the amplitude of light-induced absorbance change at 320 nm (AA320) for QA and 700nm (da700) for P700 using a laboratory-constructed difference spectrophotometer [21]. Actinic light in the green region of the spectrum [22] was provided by a combination of Corning CS 4-96 and CS 3-69 filters. The optical pathlength of the cuvette for the measuring beam was 2.07 mm and for the actinic beam it was 1.46 mm. For the QA measurements, the reaction mixture contained 200 pm Chl, 20 pm 3-(3,4-dichlorophenyl)-l,l-dimethylurea (DCMU) and 2mM K3Fe(CN)6. The samples were preilluminated briefly in the presence of DCMU and ferricyanide to allow for the complete oxidation of Cyt f and P700. This preillumination eliminated absorbance change contributions of Cyt f and P700 to dajzo. Four subsequent illuminations I 1 I I I I Triton/Chl Fig. 1. The solubilization of chlorophyll from spinach thylakoids (A) and differentialflattening correction values at 320 nm (B) as a function of the TritonlChl (wlw) ratio. (A) Solubilization is defined as the percentage of total Chl found in the supernatant following incubation of chloroplasts with Triton for 30 min at 0 C and centrifugation at x g for 30 min. Experiments with summer (0) and winter (0) spinach are shown. (B) Flattening originates from the non-uniform pigment distribution in the cuvette and it depends on chloroplast size [23]. Note that a substantial reduction in the flattening correction values occurs prior to any significant chlorophyll solubilization per sample were administered at a rate of one every minute for the registration of the dajzo amplitude. The latter was corrected for the effect of particle flattening on absorbance difference measurements using the procedure of Pulles et al. [23]. In calculating the concentration of QA, a differential absorption coefficient of 13 mm- cm- was used [24]. For the P700 measurements, the reaction mixtures contained 200 pm Chl, 200 pm methyl viologen, 2 mm sodium ascorbate and 0.2% SDS. The amplitude of AA700 was obtained upon a single illumination per sample. Concentrations of P700 were calculated using a differential absorption coefficient of 64 mm-' cm-' [25]. The concentration of Cyt f was estimated from the amplitude of the reduced-minus-oxidized absorbance difference signal at obtained with an Aminco DW-2a instrument operated in the split-beam mode, using a differential absorption coefficient of 18 mm-' cm-' [26]. The optical path length of the cuvette was 1.0 cm and the halfband width of the measuring beam was set at 3.0nm. Isosbestic points used for the absorbance difference spectra were nm and 560 nm [26]. The reaction mixture contained 240 pm Chl, 1.9% Triton X- and 300 pm K3Fe(CN)6. Sufficient amounts of hydroquinone were added in the sample cuvette to yield a final concentration of 3 mm. Protein concentrations were calculated for both the pellet and supernatant fractions using the method of Lowry [27] on resuspended acetone-precipitated aliquots. RESULTS The extent of membrane solubilization by Triton was expressed as the percentage Chl remaining in the supernatant following centrifugation at x g for 30 min. Thus, solubilization is defined arbitrarily by the fraction of chlorophyll that cannot be precipitated by this centrifugation. Fig. 1 A shows the extent of membrane solubilization plotted

3 391 as a function of Triton/Chl ratio during the treatment. The scatter of points at Triton/Chl ratios greater than 10: 1 probably arises from minor variations in incubation time and temperature during the Triton treatment as well as from seasonal variations in the plant material. Chloroplast absorbance and absorbance difference spectra are distorted as a result of the non-uniform distribution of pigments in the particle suspension. Such distortion manifests itself as a lowering of the absorbance and is referred to as 'flattening' of spectra [23]. The extent of flattening depends on the wavelength of measurement, on the pigment density in the particle and on the particle size. In spinach chloroplasts, flattening correction values at any given wavelength depend on the structural integrity of the chloroplast thylakoid system. It is evident that a detergent-mediated solubilization of the thylakoid membrane will reduce or eliminate the correction for flattening. In Fig. 1 B, the differential flattening correction factors at 320 nm are plotted as a function of the Triton/Chl ratio of the treatment. It is observed that flattening correction values (C,,,, = 1.5 in untreated samples) are lowered progressively upon increasing the Triton/Chl ratio of the treatment. They approach a low value of 1.1 under conditions of little Chl solubilization (less than 5% of the total, see Fig. 1 A). Thus, the substantial reduction in the flattening correction values is completed prior to the onset of the Chl-protein complex solubilization from the thylakoid membrane. It is suggested that low Triton concentrations (Triton/Chl < 1 : 1) disrupt the continuity of the thylakoid membrane network and diminish the size of the light-absorbing particles but are not sufficient for the solubilization of complexes from the resulting membrane fragments. We monitored the concentration of the three electron transport complexes (PSII, Cyt b6$, and PSI) in the various pellet and supernatant fractions resulting from the Triton X- treatment. This was done by measuring the amounts of Cyt f, QA and P,,, which are integral components of the Cyt b6-a PSII and PSI complexes, respectively. Fig. 2 shows typical traces from such measurements. A representative absorbance difference spectrum for Cyt f is shown in Fig. 2 (upper). The absorbance difference band with a peak at 554 nm is characteristic of Cyt f. In unfractionated chloroplasts we determined average ChllCytf = 825: 1. Representative amplitude measurements of the absorbance change at 700 nm (for PTo0) and 320 nm (for QA) are given in Fig. 2 (middle and lower, respectively). In unfractionated thylakoids, we determined average Chl/P700 = 612:l and Chl/Q, = 370: 1, respectively. An important feature of our experimentation was accounting for all of the Cytf(Cyt b6-j), P700 (PSI), and QA (PSII) following the Triton fractionations. To eliminate the possibility that Triton exerts a destructive effect on any of the integral complexes, we routinely obtained a 'balance sheet' by measuring the concentration of each complex in the supernatant and pellet fractions after each treatment. The validity of our procedure was established in Fig. 3A where, upon increased membrane solubilization, the loss of Cyt f from the pellet fraction is paralleled by the concomitant recovery of it in the supernatant fraction. The amount of PSI, as measured by P700, is accounted for in a similar manner (Fig. 3B). Unlike the balance sheet of Cyt f and P7,,, we were unable to detect the photochemical activity of QA in the supernatant fractions following Triton treatment. This is attributed to inactivation of PSII upon removal from the thylakoid membrane Wavelength #= Time,s Fig. 2. Reduced minus oxidized absorbance difference spectrum of Cyt f (upper) and experimental traces of the light-induced absorbance changes at 700nm and 320nm. (Upper) Isosbestic points were at nm and 560 nm. The Cyt f concentration was determined from the amplitude of the spectrum at 554 nm using a differential absorption coefficient of 18 mm-' cm- '. The Chl/Cyt f = 825 is a typical value obtained with unfractionated thylakoids. (Middle) Experimental trace of the light-induced absorbance change at 700 nm (da,oo) of unfractionated thylakoids. A differential absorption coefficient of 64 mm-' cm-' was used to calculate Chl/P,oo = 612 from the amplitude of (Lower) Experimental trace of the lightinduced absorbance change at 320 nm of unfractionated thylakoids. A differential absorption coefficient of 13 mm ~ I cm- I was used to calculate Chl/Q = 370 from the amplitude of da32o Fig. 4A shows the fraction of each integral complex (Cyt b6-f, PSI and PSII) recovered in the pellet following treatment at various Triton/Chl ratios. The percentage of complex in the pellet is plotted as a function of the extent of chlorophyll solubilization. Clearly, there is a differential loss of complexes from the thylakoid membrane fragments that constitute the pellet. The removal of the Cyt b6-fcomplex precedes that of PSI while PSII shows a marked resistance to solubilization by Triton. In particular, the loss of 50% of the Cyt b6$complex from the pellet occurs along with the loss of only 15% of the chlorophyll (see also Fig. 3A). The loss of 50% of the PSI complex from the pellet is concomitant with the loss of 26% of the total chlorophyll (see also Fig. 3B). The loss of about 50% of PSII from the pellet is concomitant with the solubilization of 65% of the chlorophyll. It is important to observe in Fig. 4A that the loss of 90% of the Cyt b6-fcomplex from the pellet was accompanied by the loss of only 10% of PSII. As reported by Dunahay et al. [28], the Triton-mediated selective solubilization of Cyt &$and of PSI from the thylakoid membrane does not interfere with the integrity of the membrane of the grana partition regions. Therefore, the results of Fig. 4A support the conclusion that the Cyt 66-fcomplex is not interspersed with PSII in the thylakoid membrane of the grana partition region [ll].

4 392 - c u 80 $.- 60 X 40 s b a 40- ap 20-@ Oi % Chlorophyll Solubilization Fig. 3. Percentage of (A) total Cyt f and (B) total P700 found in pellet (0) andsupernatant (0) fractions as a function of Chl solubilization. The Chl solubilization was defined as the percentage of the total Chl that cannot be precipitated following incubation of chloroplasts with Triton and centrifugation at xg for 30 min. (A) The Cyt f lost from the pellet fraction with increasing Chl solubilization was recovered in the supernatant fraction. Note that 50% of Cyt f solubilization occurs concomitant with the release of only 15% of the total Chl. (B) The P700 lost from the pellet fraction with increasing Chl solubilization was recovered in the supernatant fraction. Note than 50% of PTo0 solubilization occurs concomitant with the release of 26% of the total chlorophyll Fig. 4A (0) suggests a biphasic pattern in the removal of PSII from the thylakoid membrane. The first phase accounts for the solubilization of about 20% of PSII and it occurs concomitant with the removal of the initial 40% of the chlorophyll. This phase very likely originates from the selective solubilization of PSIID which is localized in nonappressed thylakoid membranes [7]. Plotted in Fig. 4B are the Chl &hl b ratios for both the pellet and supernatant fractions versus percentage chlorophyll solubilization. For chlorophyll solubilization less than 40% of the total, the supernatant fraction has a Chl a/chl b ratio of approximately 5.4, which is typical of stroma lamella fractions [7]. Progressive with the chlorophyll solubilization above the 40% mark, the Chl a/chl b ratio of the supernatant is lowered until it reaches that of the unfractionated thylakoids (Chl a/chl b = 2.9) at % solubilization. The pellet fraction has a Chl a/chl b ratio of 1.9 at chlorophyll solubilizations greater than 40%. This value is very close to the Chl a/chl b ratio reported for the grana partition regions [8, 191. It may be concluded that the initial 40% of the total Chl solubilized is associated with PSI and PSII, in stroma-exposed thylakoids while the remaining 60% is associated with PSII, in the thylakoid membrane of the grana partition regions [15,29] a 4.0- > O loo %Chlorophyll Solubilizotion Fig. 4. Percentage of the integral thylakoid membrane electron-transport complexes remaining in the pellet fraction of thylakoid membranes (A) and Chl a/chl b ratios (B) after variable extent of Chi solubilization by Triton. (A) Note the differential removal of Cyt bs-f( A), PSI (0) and PSII (0) from the thylakoid membrane upon progressive extraction with Triton. Also note that the loss of 90% of Cyt b.5-f from the pellet fraction is accompanied by the loss of only 10% of PSII. (B) Chl a/chl b ratios of pellet (0) and supernatant (0) fractions, obtained after each Triton treatment, plotted as a function of the Chl solubilization. Note that at Chl solubilizations less than 40%, the supernatant fractions have a Chl a/chl b >5 which is characteristic of stroma lamellae while at solubilizations > 40%, the pellet fractions have a Chl a/chl b = 1.9 characteristic of grana partition regions Fig. 5. A three-dimensional representation of the structural and functional organization of the chloroplast thylakoid membrane showing schematically the three integral electron transport complexes. PSII (0) is shown residing in the membrane of the grana partition region, PSI (0) is localized in stroma lamellae and the Cyt b6-f(a) is localized mainly in the membrane of the fret regions which serve as a continuum between the granal and stromal membranes DISCUSSION The progressive fractionation of the chloroplast thylakoid membrane by Triton X- revealed a well defined sequence of events leading to the solubilization of all chloroplast electron transport complexes. Incubation at low Triton/Chl ratios (I 1 : 1) resulted in the disruption of the thylakoid membrane continuity, apparently yielding smaller membrane fragments. These probably consist of a mixed composition of paired membranes from the grana partition regions and of large stroma-exposed lamellae [28]. Such fragmentation of the thylakoid membrane system is evidenced in the gradual lowering of the flattening correction values (Fig. 1 B) because flattening correction values strongly depend on particle size and pigment density. It may be concluded that incubation of spinach thylakoids at a Triton/Chl = 1:l resulted in the

5 complete fragmentation of the thylakoid membrane system. This Triton concentration, however, was not sufficient to remove electron transport complexes from the thylakoid membrane and centrifugation at x g for 30 min sufficed to precipitate all membrane fragments and to account for the recovery of nearly % of the Cyt b6-& PSI and PSII complexes in the pellet. Following centrifugation of thylakoid membranes incubated at a hgher Triton/Chl ratio, the recovery of electron transport complexes in the pellet fraction is no longer %. This result comes about either because of the selective removal of complexes by Triton from the edges of the thylakoid membrane fragments, or because of the further fragmentation of the thylakoids into smaller vesicles that cannot be precipitated at xg for 30 min. In either case, a distinctly differential loss of complexes from the membrane is observed upon incubation at higher Triton/Chl ratios. Thus, the solubilization of the Cyt bs-f complex precedes that of PSI, whereas PSII and the membrane of the grana partition region shows a marked resistance to Triton solubilization (Fig. 4). The differential solubilization of the three complexes upon incubation with Triton supports the notion of the domain organization of the thylakoid membrane in chloroplasts [19]. Our results cannot preclude the localization of some Cyt b6-f complex in the membrane of the grana partition region, especially in the area adjacent to the fret region. However, they strongly suggest it is unlikely that any two integral thylakoid membrane complexes are randomly intermixed in the chloroplast lamella. This argument is particularly applicable to the Cyt bs-fand PSII complexes because we observed a greater than 90% solubilization of the Cyt b6-f complex prior to interference of Triton with PSII, (Fig. 4A). In support of this conclusion, we note that a nearly complete solubilization of Cyt b6-fcomplex by Triton did not affect other functional components in the membrane of the grana partition regions (Cyt b-559, plastoquinone [S]). A scheme of the domain organization of the chloroplast thylakoid membrane is presented in Fig. 5. This three-dimensional diagram of the chloroplast membrane system is based on the work of Wehrmeyer [3]. The three membrane domains are: the membrane of the grana partition region containing PSII complexes (0), the membrane of the fret region enriched in the Cyt b6-fcomplex (A) and the stroma lamellae serving as the locus of PSI (0). The localization of the Cyt b6-f complex in the near vicinity of the grana is in agreement with the isolation of a thylakoid membrane domain which is enriched in the Cyt bs-fcomplex but depleted in both PSII and PSI [19]. Recent immunocytochemical studies have indicated that the Cyt b,-fcomplex is localized randomly both in granaand stroma-exposed lamellae [ The interpretation of the immuno-gold results is at variance with the interpretation of our results so more work is required to delineate between the two alternatives. The work was supported by a United States Department of Agriculture Competitive Research Grant to A.M. The support of 393 a National Science Foundation predoctoral fellowship to P.M. is acknowledged. REFERENCES 1. Coombs, J. & Greenwood, A. D. (1976) in The intact chloroplast, vol. 1 (Barber, J., ed.) pp. 1-51, Elsevier/North-Hoiland, Amsterdam. 2. Barber, J. (1984) in Advances in photosynthesis research, vol. 111 (Sybesma, C., ed.) pp , Junk, The Hague. Wehrmeyer, W. (1964) Planta (Berl.) 63, Akerlund, H.-E., Andersson, B. & Albertsson, P.-A. (1976) Biochim. Biophys. Acta 449, Andersson, B. & Anderson, J. M. (1980) Biochim. Biophys. Acta 593, Andersson, B. & Haehnel, W. (1982) FEBS Lett. 146, Anderson, J. M. & Melis, A. (1983) Proc. Natl Acad. Sci. USA 80, Lam, E., Baltimore, B., Ortiz, W., Chollar, S., Melis, A. & Malkin, R. (1983) Biochim. Biophys. Acta 724, Haehnel, W. (1984) Annu. Rev. Plant Physiol. 35, Millner, P. A. & Barber, J. (1984) FEBS Lett. 169, Ghirardi, M. L. & Melis, A. (1983) Arch. Biochem. Biophys. 224, Berthold, D. A., Babcock, F. T. & Yocum, C. F. (1981) FEBS Lett. 235, Cox, R. P. & Andersson, B. (1982) Biochem. Biophys. Res. Commun. 103, Anderson, J. M. (1982) FEBS Lett. 138, Akerlund, H.-E. & Andersson, B. (1983) Biochim. Biophys. Acta 725, Allred, D. R. & Staehelin, L. A. (1985) Plant Physiol. 78, Goodchild, D. J., Anderson, J. M. & Andersson, B. (1985) Cell Biol. Int. Rep. 9, Shaw, P. J. & Henwood, J. A. (1985) PZanta (Berl.) 165, Melis, A., Svensson, P. & Albertsson, P.-A. (1986) Biochim. Biophys. Acta 850, Arnon, D. I. (1949) Plant Physiol. 24, Melis, A. & Hart, R. W. (1980) Carnegie Inst. Wash. Yearbook 79, Ghirardi, M. L. & Melis, A. (1984) Plant Physiol. 74, Pulles, M. P. J., Van Gorkom, H. J. & Verschoor, G. A. M. (1976) Biochim. Biophys. Acta 440, Van Gorkom, H. J. (1974) Biochim. Biophys. Acta 347, Hiyama, T. & Ke, B. (1972) Biochim. Biophys. Acta 267, Bendall, D. S., Davenport, H. E. & Hill, R. (1971) Methods Enzymol. 23A, Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193, Dunhay, T. G., Staehelin, L. A., Seibert, M., Ogilvie, P. D. & Berg, S. P. (1984) Biochim. Biophys. Acta 764, Melis, A. & Anderson, J. M. (1983) Biochim. Biophys. Acta 724,