- Email: [email protected]
Variable fluorescence of photosystem I particles and its application to the study of the structure and function of photosystem I
ARCHIVES
OF BIOCHEMISTRY
Vol. 235, No. 2, December,
AND
BIOPHYSICS
pp. 449-460, 1934
Variable Fluorescence of Photosystem I Particles and Its Application the Study of the Structure and Function of Photosystem I B. C. TRIPATHY, Department
J. E. DRAHEIM, of Biochemistry,
Received February
G. P. ANDERSON,
Ohio State University,
AND
to
E. L. GROSS’
Columbus, Ohio 4.9210
21, 1934, and in revised form August ‘7, 1934
Chlorophyll a fluorescence in Photosystem I (PSI) particles isolated according to the method of Bengis and Nelson [J Bid Chem. 252, 4564-4569 (1977)] was found to be dependent on the redox state of both P700 and X (an acceptor on the reducing side of PSI). Addition of dithionite plus neutral red to PSI caused an increase in fluorescence intensity and a shift of the main fluorescence peak from 689 to 674 nm. Addition of electron acceptors such as ferredoxin and methyl viologen decreased the fluorescence yield when added to PSI incubated under anaerobic conditions in the presence of excess dichlorophenol indophenol (DCIPHa). The K,,, for ferredoxin agreed with that determined from direct measurements of ferredoxin reduction, showing that X is a quencher of fluorescence. P700 was also found to be a quencher of fluorescence, since electron donors such as DCIPHB, TMPD, and plastocyanin decreased fluorescence with Km’s nearly identical to those observed for P700+ reduction. Chemical modification of PSI (with ethylene diamine + a water-soluble carbodiimide) to make it positively charged increased the fluorescence yield and shifted the 689-nm peak to 674 nm. The Km’s for DCIPHz and ferredoxin were decreased. In contrast, modification of PSI with succinic anhydride, which increased the net negative charge, increased the Km for ferredoxin. Salts affected the interaction of methyl viologen with PSI. Both anion and cation selectivity were observed. Limited proteolysis increased the Km for both methyl viologen and ferredoxin, indicating that their binding site on PSI was altered. These results suggest that the binding site for ferredoxin is on either the 70- or the 20-kDa subunit of PSI. o 1984 Academic PRESS, IW. Chlorophyll a fluorescence in chloroplasts is emitted mainly from Photosystern II. Its intensity is dependent on the redox state of both P680’ and Q (l-3). Therefore, it has been used to study elec’ To whom correspondence should be addressed. ‘Abbreviations used: PSI, Photosystem I; X, acceptor for Photosystem I; P700, chlorophyll which acts as primary electron donor for PSI; P’700+, oxidized P700; P630, chlorophyll which acts as primary electron donor for PSII: Q, primary electron acceptor for Photosystem II; PC, plastocyanin; MV, methyl viologen; EDA, ethylene diamine; PMSF, phenylmethylsulfonyl fluoride; Hepes, I-(2-hydroxymethyl)1-piperazineethanesulfonic acid; Mes, 4-morpholineethanesulfonic acid; Taps, 3-{[2-hydroxy-l,l-bis(hy-
tron transport in chloroplasts. There have been reports of variable fluorescence from Photosystem I (PSI) in PSI-enriched particles (4-8). The reports quoted above, however, do not clearly elucidate the redox dependence of variable fluorescence in PSI. Our first objective will be to document the existence and redox dependence of variable fluorescence in PSI particles isolated according to the method of Bengis
droxymethyl)ethyl]amino}-1-propanesulfonic acid; DCIPH, dichlorophenol indophenol; TMPD, N,N, N’,N’-tetramethyl-p-phenylene diamine; DCIP, dichlorophenol indophenol. 449
0003-9861/84 $3.00 Copyright All rights
0 19% by Academic Press, Inc. of reproduction in any form resewed.
450
TRIPATHY
and Nelson (9, 10). We will then use variable fluorescence to study the effects of cations, chemical modification, and proteolysis on the properties of the reducing side of PSI. Cations have been shown to have multiple effects on chloroplasts (11, 12). In particular, they have been shown to promote the interaction of plastocyanin (and other negatively charged electron donors) with P700 (13-1’7). A similar enhancement of the reaction with plastocyanin was obtained when PSI particles were made positively charged by modification with ethylenediamine in the presence of a water-soluble carbodiimide (18). Cations have also been shown to act on the reducing side of PSI (19), but this has not been well studied due to the difficulty of directly measuring effects on the reducing side of PSI. We will show that variable fluorescence is a useful tool for these and other studies. MATERIALS
AND
METHODS
Isolation of Photosystem I particles. The PSI particles were isolated according to the method of Bengis and Nelson (9). The chlorophyll/P700 and the chlorophyll a/b ratios were 80 and greater than 18, respectively. Chlorophyll was determined according to the method of Arnon (23). The PSI particles were dialyzed for 12 h against 50 mM Tris-HCl buffer (pH 8.0). Chemical modZfication and trypsin treatment. Chemical modification of PSI particles with EDA and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide was carried out according to the method of Burkey and Gross (21) as derived from Means and Feeney (24). Chemical modification with succinic anhydride was carried out as follows. Isolated PSI particles were dialyzed against 1.0 M sodium carbonate buffer (pH 8.0), containing 0.05% Triton X-100, for 2 h. The PSI particles were then modified by adding 35.7 mmol succinic anhydride/pmol P700 over a l-h period. The reaction was then allowed to go to completion for 0.5 h. The reaction was then quenched using 10 mM Tris-HCl (pH 8.2), containing 0.05% Triton X-100, added in sufficient amount to reduce the chlorophyll concentration to 30 pg/ml. The modified PSI particles were then dialyzed against the same solution. Trypsin treatment of PSI particles was carried out as described previously (25) by adding trypsin directly to the reaction cuvette containing 3 ml 10 mM Hepes/NaOH buffer (pH 7.8) at a chlorophyll: trypsin ratio of 21 by weight. The trypsin inhibitor
ET AL. phenylmethylsulfonyl fluoride (PMSF) was added after 10 min of incubation with trypsin. Chlorophyll a jluorescence measurements. The chlorophyll a fluorescence excitation and emission spectra of PSI particles were measured using a Fluorolog 1902 spectrofluorimeter (Spex Industries). A Tracer NS-570A digital signal analyzer was used to store and analyze the fluorescence spectra. The samples were excited at 436 nm (slit width = 5 nm). The excitation intensity was 500 pW/cm*. The emission spectra were corrected for photomultiplier sensitivity. The slit width for the emission was 5 nm. The fluorescence excitation spectra were not corrected for lamp and photomultiplier sensitivity. However, they were done under identical conditions and, therefore, can be compared. The PSI particles were suspended in 10 mM Hepes/NaOH buffer (pH 7.8) at a chlorophyll concentration of 2 pg/ml for the fluorescence measurements. All experiments were carried out at 20°C. Anaerobic conditions were maintained by adding 20 mM glucose, 40 units/ml glucose oxidase, and 600 units/ml catalase to the reaction mixture. Then, argon was bubbled through the reaction mixture for 5 min, after which the cuvettes were sealed. As glucose oxidase has a weak fluorescence tail in the region 670-690 nm, the amount of fluorescence quenching was calculated by subtracting the background fluorescence of glucose oxidase. Ferredoxin isolation and reductirm. Crude ferredoxin was isolated according to Davis and San Pietro (26), and was dialyzed over night against 50 mM Tris-HCl buffer (pH 8.0). Ferredoxin reduction was carried out under anaerobic conditions in sealed cuvettes as described above. Ferredoxin reduction was measured as a decrease in absorbance at 420 nm, upon continuous illumination, using the Aminco DW-2a spectrophotometer in the double-beam mode. Assay mixtures contained PSI particles (at 10 pg/ ml chlorophyll), 10 mM Hepes/NaOH buffer (pH 7.8), 5 mM sodium ascorbate, 0.33 mM DCIP, 0.05% Triton X-100, and various concentrations of ferredoxin. A 300-W projector lamp was used to illuminate the sample. Red actinic light (4.7 X 10’ ergs cm-r s-i) greater than 650 nm was isolated by placing 3 cm of water and a red long-pass cutoff filter (Corning CS-2-64) between the lamp and the sample compartment. A 620-nm short-pass cutoff filter (Bausch and Lombe NO 90-l-620) was used to prevent the actinic light from reaching the photomultiplier. The rates were calculated using an extinction coefficient of 9.7 mrt-’ cm-i at 420 nm for ferredoxin (27). RESULTS
The eflect of electron donors and acceptors on chlorophyll a jluorescence. First,
we determined whether the chlorophyll a fluorescence spectrum was dependent on
VARIABLE
FLUORESCENCE
the redox state of the reaction center. To do this, we determined the fluorescence emission spectrum under three illumination conditions. These three conditions corresponded to the following states of the reaction center: P+X, PX, and PX-. In each case, P refers to P’700 and X refers to any one of the acceptors on the reducing side of Photosystem I (PSI). In condition (a), the PSI particles were illuminated in the presence of oxygen and 50 PM ascorbate. The rate of P+ reduction by ascorbate was too slow to reduce the P+ produced in the light. In the aerobic state, oxygen acted as the terminal electron acceptor (1’7). Thus, the particles were in the P+X state under these conditions. In state (b), P700+ was reduced using high concentrations of the electron donor dichlorophenol indophenol (DCIPHz), converting the reaction center to the P’7OOX state. In case (c), the PSI particles were illuminated in the presence of dithionite and neutral red, which reduced both the P700 and X (8) and also maintained the sample in an anaerobic state. The reaction center was in the P7OOX- state under these conditions. Figure 1 shows the fluorescence emission spectrum for PSI particles illuminated under these three conditions. In case (a), the PSI particles showed a peak at 689 nm with a shoulder at 678 nm and another small peak at 730 nm. Addition of an electron donor (case b) caused a decrease in the emission at 678 and 730 nm with no shift in the emission spectrum. The fluorescence changes were complete within 20 s (not shown). In case (c), in which both P and X were reduced, the fluorescence intensity was increased and the 689nm peak was shifted to 674 nm. These results suggest that both P and X are quenchers of chlorophyll a fluorescence in PSI particles. Since we have shown that both P700 and X are quenchers of fluorescence, we can use variable fluorescence to study the effects of cations, pH, and chemical modification on both the oxidizing and reducing sides of PSI. This technique is particularly useful for the reducing side of PSI due to the paucity of methods for measuring it directly.
IN PHOTOSYSTEM
900
451
I
-
800
WAVELENGTH
(nm)
FIG. 1. Fluorescence emission spectra of PSI particles under various conditions. The excitation wavelength was 436 nm. Other conditions were as described under Materials and Methods. (- - -) Control PSI (no added electron donors or acceptors; ( * * * ) PSI treated with 20 pM DCIPH2; (p) PSI particles with 20 mM dithionite + 2 pM neutral red added.
In each case, we want to determine the effect of various treatments on the Km and I’,,,,, for electron transport using different donors and/or acceptors. Therefore, we need to be certain that there is a direct relationship between these quantities determined from variable fluorescence measurements and those determined by more conventional means; namely, measurements of P700+ or ferredoxin reduction. In particular, we must make sure that the Km’s are independent of light intensity. We have measured variable fluorescence in terms of percentage quenching. Percentage quenching is defined according to %Q = lOO*(F_n - Fn)/F-,,,
PI
where FD and FpD are the fluorescence intensities obtained in the presence and absence of an electron donor respectively. We have chosen this formulation because we can measure %Q more easily
452
TRIPATHY
and accurately under all conditions than we can measure the constant (FO) and variable (Fv) fluorescence levels. Moreover, %Q was proportional to the concentration of reduced P700 (Fig. 2). Under conditions of high light intensity, l/‘%Q is related to l/[D] according to &={L$}.{1+~},
[2]
where D is the donor concentration, Pt is the total amount of P700, 9, is the quantum yield for constant fluorescence (F), and \k, is a pseudo-quantum yield for the variable part of fluorescence (Fv) with M-l units. Kd is the concentration of donor required for half-maximal quenching and should be equal to the K, determined by other means. Q,,,,,, which is the percentage quenching observed at infinite donor concentration, is related to the maximum rate of electron transport. This equation is derived in the Appendix. An analogous equation can be derived for fluorescence quenching by an electron acceptor.
FIG. 2. The effect of DCIPHa on fluorescence quenching and the steady-state level of P700 oxidation. Light-induced P700 oxidation was carried out according to the method of Gross (17). The reaction mixture contained 10 mM Hepes buffer, pH 7.8, 2 mM ascorbate, and PSI particles containing 2 pgg/ml chlorophyll. The PSI particles were titrated with increasing concentrations of DCIP. The steady-state level of the light-induced P700 oxidation was determined after each addition. Fluorescence quenching was determined as described for Fig. 4. The light intensity was adjusted to 400 pw m-a for both parts of the experiment.
ET AL.
We determined the Kd as a function of light intensity for the electron donors DCIPHB and plastocyanin. In the case of plastocyanin, the Kd was 3.3 and 2.9 pM at 250 and 760 PW cm-’ intensity, respectively. For DCIPHz, Kd was 33 PM at both light intensities. Thus, we can conclude that the Kd)s are independent of light intensity. If Eq. [2] is correct, then other electron donors to PSI should also quench fluorescence. In each case, the Kd’s should be the same as the K,‘s determined by other methods. This is true for both plastocyanin and TMPD (Table I). Furthermore, plastocyanin quenched fluorescence only in the presence of Mgaf ions, a condition which was also required for P’700 reduction. In addition, the electron acceptors methyl viologen and ferredoxin also quenched fluorescence. As in the case of DCIPHz, both the 680- and 730-nm peaks decreased in intensity (not shown). Also, the fluorescence changes were complete within 20 s (not shown). In the case of ferredoxin, the extent of quenching was essentially the same using both blue (430 nm) and red (660 nm) exciting light. We also measured ferredoxin reduction directly by monitoring absorption changes at 420 nm under anaerobic conditions. We obtained a K, of 12.5 PM and a V,,, of 16 pmol ferredoxin reduced mg chlorophyll-’ h-l. This agrees very well with the value of 12 PM obtained from fluorescence measurements (Table I.). This agreement shows that fluorescence quenching is a good measure of activity on the reducing side of PSI. The question arises as to whether the electron donors and acceptors could be quenching fluorescence by reacting directly with the pigment bed. This is unlikely in the case of the large watersoluble molecules such as plastocyanin and ferredoxin, which would not be able to penetrate the membrane. In the case of DCIPHz, we found that 100 PM was unable to quench fluorescence of chlorophyll a in 80% acetone [See also (28)]. The eflect of chemical rnoo!i&xtion on Photosystem Ijtwrescence. Since we have
VARIABLE
FLUORESCENCE
IN PHOTOSYSTEM
TABLE
I
THE EFFECT OF ELECTRON DONORS AND ACCEPTORS ON THE QUENCHING FLUORESCENCE
Electron donor or acceptor DCIP * TMPD
Plastocyanin -Mg2 +3.3 mM MgCla Ferredoxin’ Methyl viologen
453
I
IN PHOTOSYSTEM
OF CHLOROPHYLL
a
I PARTICLES
Q Kia (PM)
(?6 Quk:hing)
Knad (PM)
Reference
(17)
32 32
30 29
29 29
No effect 2.6 12.5 460
30 36 24
3.2 12
(25)
a Fluorescence quenching was determined as a function of donor or acceptor concentration, after which from double reciprocal plots of fluorescence quenching vs donor or the Kd and Q,,,- were determined acceptor concentration. * Experiments using electron donors were done under aerobic conditions as described for Fig. 2. c Experiments with electron acceptors were done under anaerobic conditions as described under Materials and Methods. Each concentration represents a separate sample. Other conditions were as described for Fig. 3. dThe Kd’s were compared with the K,,,‘s determined from either the reduction of P’700+ or ferredoxin. Where indicated, the Km values were taken from the literature.
shown that variable fluorescence is a measure of the ability of electron acceptors such as MV and ferredoxin to interact with X, we can use this technique to study the effects of cations and net charge on PSI. We can change the charge on PSI by three methods: chemical modification (21), addition of salts to screen the existing charges (18, 29, 30), and changing the pH. We have examined the effect of two types of chemical modification on the fluorescence properties of PSI. In the first case, we modified carboxyl groups with ethylenediamine in the presence of a water-soluble carbodiimide (21). This reaction changes the isoelectric point of PSI from pH 4.7 to 9.5, converting it into a positively charged protein. The second modification involved the reaction of amino groups with succinic anhydride. This reaction adds a negative charge to the pK to 4.5 amino groups, lowering (Anderson and Gross, unpublished experiments). EDA modification shifted the 689-nm emission peak to 674 nm and also increased the intensity of both the 674- and ?‘30-nm peaks as compared to the unmodified control (Fig. 3). As in the case of
control PSI, addition of dithionite and neutral red caused an increase in the fluorescence intensity (Fig. 3). However, 9000
I ’
I
A’
8000 -
640
660
680
700
WAVELENGTH
720
740
760
(“ml
FIG. 3. The effect of DCIPHI and dithionite on the fluorescence emission spectrum of EDA-PSI particles. Conditions were as described for Fig. 2. (- - -) Con+ 20 mM dithitrol; ( + . . ) + 20 pM DCIPHr; (p) onite and 2 PM neutral red.
454
TRIPATHY
ET AL.
in contrast to dithionite-treated control P700+ reduction (1’7). Modification also PSI, there was no additional blue shift of caused an increase in the Q,,,,, from 33 to the 674-nm emission peak. 42%. The increase in Q,,, reflects an inChemical modification with EDA caused crease in the I’,,, for electron transport. a decrease in the Kd for DCIPHz from 32 We also examined the effect of EDA to 12 PM (Fig. 4). This was expected since modification on the interaction of PSI a negatively charged electron donor such with the electron acceptors ferredoxin and methyl viologen. The Km for ferredoxin as DCIPHz should react more favorably with positively charged PSI than with decreased from 12 to 4 PM as expected. No negatively charged PSI. These results are quenching was observed with methyl violconsistent with those reported earlier for ogen due to steric repulsion between positively charged MV and positively charged EDA-PSI. I 1 I 1 I I The shape of the fluorescence emission spectrum was not altered upon succinic anhydride modification (not shown). However, the Km for negatively charged ferredoxin was increased from 12 to 29 I.LM as expected. No effect was observed for methyl viologen, since the reaction with negatively charged control PSI was already favorable. The e#ect of salts on fluorescence quenching. The Guoy-Chapman theory (29, 30) predicts that salts such as NaCl and MgS04 should inhibit the favorable reaction between MV and control PSI. Saltinduced inhibition of fluorescence quenching was observed (Fig. 5A). However, there was a cation selectivity which is not predicted by Guoy-Chapman theory. For example, magnesium salts were more effective on a cation concentration basis than were calcium salts. Also, NaCl was less effective than predicted by theory. Conversely, anions should screen the positive charges on EDA-PSI, facilitating its interaction with MV. Some salts did promote fluorescence quenching (Fig. 5B). 0 IO 20 30 40 50 60 However, anion selectivity was observed [ DCIPH,] (PM) in that acetate anions were more effective than chloride ions. Moreover, magnesium FIG. 4. DCIPHr concentration dependence of salts were totally ineffective at promoting quenching of variable fluorescence in control (-) and EDA-PSI (- - -). Inset; Lineweaver-Burk plot fluorescence quenching. One possible exfor variable fluorescence as a function of DCIPHc planation is that magnesium ions compete concentration for control (-) and EDA-PSI (- - -). with MV for its binding site. The PSI particles were titrated with increasing In contrast to the results obtained with concentrations of DCIP in the presence of 59 PM Na MV, those obtained with DCIPH2 did obey ascorbate. The fluorescence intensity was determined the Guoy-Chapman theory. Both NaCl after each addition. The excitation wavelength was and MgS04 increased the interaction be436 nm. The emission wavelengths were 689 and 674 tween DCIPHz and negatively charged nm for control and EDA-PSI, respectively. Other PSI at pH 8.0. A plot of log percentage conditions were as described under Materials and quenching (%Q) vs the ionic strength to Methods.
cI
VARIABLE
FLUORESCENCE
SALT CONCENTRATION I
I
I
I
(mM) /
I
2 B No,SO, CH,COONo NoCl 4 so.
c:
WA IO
20
30
40
50
SALT CONCENTRATION
60
70
(mM)
FIG. 5. The effect of salts on methyl viologeninduced quenching of fluorescence for both control and EDA-modified PSI particles. The ability of methyl viologen to quench fluorescence was determined as a function of salt concentration for both control and EDA-modified PSI particles. PSI particles were maintained under anaerobic conditions with glucose, glucose oxidase, and catalase. Ascorbate and DCIP were added as described for ferredoxin reduction in order to maintain P700 in the reduced state and to promote electron transport resulting in the reduction of X. Other conditions were as described under Materials and Methods. (A) Control PSI; (B) EDA-PSI.
IN PHOTOSYSTEM
455
I
The Kd for the DCIPH2-induced quenching increased with increasing pH above pH 7 for control PSI and decreased over the same pH range for EDA-modified PSI. The results were corrected for the pH dependence of the interaction of ascorbate with PSI (13). These results can be explained on the basis of the effect of pH on the charge of DCIPH2. It is neutral below a pH of 7 (31) and negatively charged above it. Thus, above pH 7, the negative charge would facilitate its interaction with EDA-PSI while inhibiting that with control PSI. Qmax increased with increasing pH for both control and modified PSI. The pH dependence for Q,,, corresponds to a change in V,,, which reflects changes in the electron transfer reactions themelves (17). &ma,was greater for EDA-PSI than control PSI at all pH values tested. This increase in QmaXis probably related to the increase in the I’,,,,, observed for P700+ reduction under the same conditions (21). The pH dependence for ferredoxin-induced quenching is shown in Fig. 7B. There was a gradual increase in the Km for the interaction of ferredoxin with control PSI as the pH was increased, reflecting the increasing negative charge on both the PSI particles and ferredoxin. In contrast, Kd for MV was pH independent
the power l/2 (IS-“.5) produced a straight line according to (Fig. 6). log %Q = log %Q, - 0.078q~(IS)-~.~,
[3]
where %Q and %Q,, are the values in presence and absence of salt, respectively; x is the charge on the electron donor, q is the net charge on the particle, and IS is the ionic strength. A slope of -0.073 (Fig. 6) corresponds to a net charge (q) on the oxidizing side of PSI particles of -0.94 PC! cme2. Identical results were obtained using MgS04. Our results agree with those of Itoh (16) and Tamura et al. (18), who obtained -0.84 PC cmp2. The effect of pH on jluwescence quenching. Figure 7A shows the effect of pH on DCIPHz-induced fluorescence quenching.
I 0
/ 2
I 4
I 6
1 0
I IO
[Is]-“2 FIG. 6. Guoy-Chapman plot of fluorescence quenching as a function of ionic strength. Fluorescence quenching was determined as a function of NaCl concentration for control PSI particles using DCIPHa as the electron donor. Other conditions were as described for Fig. 3 and under Materials and Methods. The data were plotted according to Eq.
u51.
TRIPATHY
456
ET AL.
viologen on ferredoxin reduction in order to determine whether they interact at the same site on the reducing side of PSI. The observation (Fig. 8) that MV is a competitive inhibitor of ferredoxin reduction indicates that the two electron acceptors do act at the same site.
The eflect of proteolysis on j-luorescence quenching. Limited proteolysis is a useful
‘\ IO
IO
i
o-o 6
7
9
8 PH
40
0.8
C 0.6
30
r;, ” a d
Gl 20
0.4
Qm*
IO 0H 6
r 2 x’
technique for determining which polypeptides in a multisubunit complex such as PSI are required in order for a particular electron transport reaction to occur. We have previously shown that tryptic digestion of the Bengis and Nelson PSI preparation affected both the 70- and 20-kDa polypeptides (25). A decrease in the rate of P700+ reduction was also observed. Trypsin treatment had only a small effect on the fluorescence emission spectrum, consisting of a 10% decrease in emission intensity at 680 nm and a 10% increase in the emission at ‘740 nm (25). Tryptic digestion increased the Kd for both ferredoxin and methyl viologen as measured by the variable fluorescence technique (Table II). There was almost no effect on the V,,, in either case. Thus,
0.2
7
PH
s
9
0
FIG. 7. The effect of varying the pH on fluorescence quenching using control (p) and EDA-modified (- - -) PSI particles. (A) DCIPHa was the electron donor under aerobic conditions; (B) ferredoxin was the electron acceptor under anaerobic conditions; (C) methyl viologen was the electron acceptor under anaerobic conditions. The PSI particles were dialyzed into the appropriate buffers. Mes buffer was used between pH 5.5 and 6.7, Hepes buffer was used between pH 6.8 and 7.9, and Taps buffer was used between pH 8.0 and 9.0; all buffers were 10 mM. Other conditions were as described for Table I and under Materials and Methods.
(Fig. 7C), as expected if the interaction of MV was already favorable at low pH. Qmax’v however, did increase between pH 8 and 9.
Reduction of ferredoxin of MV: We studied
in the presence
the effect of methyl
'L f 5 i
0.8 "
lg 0.6 x
2 0.4 2 iG 0 0.2
E
2. [Ferredoxinl-
I (PM)-’
FIG. 8. The effect of methyl viologen on the rate of ferredoxin reduction. The rate of ferredoxin reduction was determined as a function of ferredoxin concentration in the presence (A) and absence (0) of 0.5 mM methyl viologen, after which the data were plotted in double-reciprocal form. Other conditions were as described under Materials and Methods section.
VARIABLE
FLUORESCENCE TABLE
IN PHOTOSYSTEM
457
I
II
THE EFFECX OF TRYPSIN TREATMENT AND SUCCINIC ANHYDRIDE MODIFICATION ON THE REDUCING SIDE OF PHOTOSYSTEM I
Electron Methyl
Control
acceptor
viologen
Kd (mM)”
Ferredoxin
Trpsin-treated PSI
PSI
Succinic anhydride modified
Q
0.46 21
2.66 26
0.47 31
~:PM) Qmar
12 31
24
29
36
50
a Fluorescence quenching was determined as a function of methyl viologen or ferredoxin concentration for either control or trypsin-treated PSI particles, after which the Kd and Q,.. were calculated. Other conditions were as described under Materials and Methods.
trypsin treatment interferes with the binding of both electron acceptors to PSI. We can conclude that one or both of the affected polypeptides (i.e., the 70- and 20kDa polypeptides) are required for the interaction of these electron acceptors with PSI, and one of them may contain the binding site for ferredoxin. Excitation spectra of control and EDAPSI particles in the presence and absence of dithionite. The excitation spectrum for the 730-nm emission peak of control PSI shows excitation maxima at 436 and 674 nm, with carotenoid bands at 470 and 495 nm (Fig. 9). The excitation spectrum for the 689-nm peak was the same as for the
\ 400
440
._--_480
---% -520
730-nm peak (not shown). Both excitation spectra match the absorption spectra (not shown). The absorption spectrum also agrees with that obtained earlier by Bengis and Nelson (9). In EDA-PSI particles, the excitation peaks were blue-shifted to 432 and 666 nm, respectively. Again, they corresponded to the absorption peaks (not shown). The blue shift in excitation and absorption suggests a structural reorganization of the PSI particles. A similar blue shift was observed upon addition of MgC& to control PSI (32). The blue shift was not due to release of free chlorophyll since modification increased the ability of the particles to utilize 650-nm light to
5M)
WAVELENGTH
6&a
640
660
720
(nm)
FIG. 9. Excitation spectra of control (p) and EDA-modified (- - -) PSI particles. The emission wavelength was 730 nm. Other conditions were as described under Materials and Methods. The fluorescence intensities were normalized at the blue excitation peak.
458
TRIPATHY
oxidize P700 (21). Moreover, the Chl/P700 ratio did not change upon modification (Burkey and Gross, unpublished observations). In contrast, treatment of control PSI with dithionite caused no blue shift in the absorption spectrum (not shown), although a blue shift was observed in the emission spectrum. Thus, there is a fundamental difference between dithionite treatment and EDA modification. In the case of dithionite, there was a shift in emission without a change in excitation or absorption (i.e., a change in Stoke shift). On the other hand, EDA modification caused both an 8-nm change in the excitation wavelength and a 15-nm change in emission wavelength. DISCUSSION
AND
CONCLUSIONS
The results presented above show that the fluorescence intensity of PSI particles isolated according to the procedure of Bengis and Nelson (9, 10) is dependent on the oxidation state of both X and P700. The evidence that X is a quencher of fluorescence consists of the following. A high fluorescence intensity was observed when PSI was illuminated in the presence of DCIPHz under anaerobic conditions. This produces the PX- state. Addition of the electron acceptors methyl viologen and ferredoxin decreased the fluorescence which would be expected if X were a quencher (i.e., they produce the PX state). Moreover, the observation that the Kd determined for ferredoxin from fluorescence quenching measurements agrees with the Km obtained from direct measurements of ferredoxin reduction supports this interpretation. Illumination of PSI in the presence of dithionite and neutral red produced an even higher intensity than illumination in the presence of DCIPH2 because it is more effective at producing the PX- state. Our results agree with those of Telfer et al. (8), who also observed that X (rather than X-) was a quencher of fluorescence. This is analagous to fluorescence quenching by Q in Photosystem II. The evidence of fluorescence quenching by P consists of the following. When PSI particles were illuminated under aerobic
ET AL.
conditions in the absence of electron donors, the P+X state was created since Xcan react with oxygen under these conditions (17). Addition of electron donors such as reduced DCIP, TMPD, and plastocyanin quenched the fluorescence by producing the PX state. The Kd’s for these donors agreed with the Km’s determined from measurements of P700+ reduction. In addition, the absolute requirement for M$+ ions in the case of plastocyanin was the same as for P700+ reduction. Thus, our results lead to the conclusion that it is P which is the quencher of fluorescence. This is reasonable in that fluorescence should be less when the trap is open and capable of doing photochemistry. However, in contrast to our results, both Telfer et al. (8) and Ikegami (7) found P700+ to be a quencher. One way to reconcile these results is that both P and P+ are quenchers of fluorescence, but that P is a greater quencher under the conditions of our experiments. Three peaks were observed in the emission spectrum of control PSI, at 674, 689, and 730 nm. Reduction with dithionite increased the 674- and 730-nm peaks. Addition of electron donors and acceptors affected these same peaks. We can conclude that both the 674- and 730-nm peaks belong to the core antenna and can respond to the state of the reaction center. Our 674-nm peak may be analogous to the 695-nm peak observed by Mullet et al. (33). However, they observed no changes in the 730-nm peak. Chemical modification with EDA caused a blue shift in both the absorption and emission spectra. EDA modification may cause a conformational change in the PSI particles, which shifts more of the chlorophyll into the “core” antenna. In this paper, we have shown not only that P and X act as quenchers of fluorescence in PSI particles, but also that variable fluorescence is a powerful technique for examining electron transport in PSI. In particular, it is useful for studying the reducing side of PSI which is difficult to study by other methods. In our preliminary investigations, reported herein, we have shown that salts affect the reducing side of PSI in a manner not totally con-
VARIABLE
FLUORESCENCE
sistent with the Guoy-Chapman theory. Moreover limited proteolysis experiments in conjunction with variable fluorescence measurements have helped us to narrow down the location of the binding site for ferredoxin. In the future, we will use this technique to study other types of modification on the structure and function of PSI. This technique should also prove useful for examining the effect of different growth conditions on the state of PSI. APPENDIX
The fluorescence intensity of PSI depends upon two components. One component is the constant fluorescence, F,. The other component is the variable fluorescence, F,. The measured fluorescence intensity of PSI before and after addition of an electron donor are designated by F-,, and F,, respectively. Thus
FD = F, + F, = Qo.I+
XJXVsI.P+,
[l]
where a0 is the quantum yield for F,, 9, is a pseudo-quantum yield for F, (in M-’ units), 1 is the intensity of light, and P+ and P are the molar concentrations of oxidized and reduced P700, respectively. Although P+ might be a quencher of PSI fluorescence, it is our contention that P must be a more effective quencher. If this is true, then the percentage quenching, %Q, should be proportional to the concentration of reduced P700. Due to the ease of accurately measuring FD and F-D, we shall define %Q as
%Q = { F-;-DFD}dOO.
[Z]
IN PHOTOSYSTEM
459
I
1 -- 1 P, + (%3/Q”) Pt - P+ %Q - 100
P+I%P+ and
Fl
k-2 P+ + D c (P+D) 2 P + D+.
P, = P + P+ + (P+D),
[71
In this description kl is a second-order rate constant for the formation of P+; kz is the rate constant for the formation of (P+D); and k3 and k4 are the rate constants for the decay of (P+D) to reactants and products, respectively. In the steady state we have dP/dt = d(P+D)/dt dP/dt = -kJP
PI
= 0,
PI
+ k,(P+D),
and d(P+D)/dt = &P+D - (k3 + k,)(P+D) = 0. After combining
and rearranging
[lo]
we have
klIKd + kJD + k,D P’ klIKd + kJD
WI where
Kd = (k4 + k,)/k, = K, If we let
B = klIKd + kJD + k4D klIKd + kJD ’
WI
condenses to
In the absence of an electron donor, P+ equals the total amount of P700 (PJ. The fluorescence prior to the addition of donor is, therefore,
where
[51
If we consider light as a substrate, then the kinetics of formation and decay of P’ and (P+D) can be simply described by
then Eq. [ll]
F-D = @-I+ \k,*I*Pt
*
P31
Substitution of Eq. [13] into Eq. [5] gives us 131 1 B(D + Kd) ?CQ= (B l)(Kd + B)D [41
and where (P+D) is the complex formed between P+ and the electron donor D. Combining Eqs. [l] through [4] gives us
x
1
i
+
(@O/Q”)
x
pt
1
1
100 .
P41
TRIPATHY
460
In general, a plot of (l/%Q) vs (l/D) will be both nonlinear and a function of light intensity. However, the intercept on the 2 axis (where (l/%Q) = 0) is B (B - l)(&
+ @Do
where (l/Do) is the intercept on the negative x axis. When (Do) is equal to Kd, the left hand side of Eq. [15] is identically zero. Thus, the intercept should be independent of light intensity and equal to the Km determined by other methods. In the special case where kl*I $s k4, B - 1.0 and Eq. [14] simplifies to
1+v}.t
[16]
Using Eq. [16], a plot of (I/%&) vs (l/D) will now be linear and light intensity independent. It will also have the form of the familiar Lineweaver-Burk plot. Since the measured curves are linear (see text), the klI 9 k4 condition probably holds under our experimental conditions. ACKNOWLEDGMENTS This work was supported in part by an Ohio State University postdoctoral fellowship to B. C. Tripathy. We wish to thank Dr. Robert T. Ross and Dr. Michael Marchiarullo for their helpful suggestions in the preparation of this manuscript. REFERENCES 1. DUYSENS, L. N. M., AND SWEERS, H. E. (1963) in Studies in Microalgae and Photosynthetic Bacteria (Japanese Society of Plant Physiol., eds.), pp. 353-372, Univ. of Tokyo Press, Tokyo. 2. OKAYAMA, S., AND BUTLER, W. L. (1972) B&him. Biophys. Acta 267, 523-527. 3. GOEDHEER, J. C. (1972) Annu. Rev. Plant PhysioL 23,87-112. 4. KARAPETYAN, N. V., KLIMOV, V. V., AND KRAYNOVSKII, A. A. (1973) Photosynthetica 7, 330337. 5. KARAPETYAN, N. V., AND RAKHIMBERDIEVA, N. A. (1981) in Proceedings of the 5th International Photosynthesis Congress (AkoyunogIOU, G., ed.), Vol. 1, pp. 337-346, Balban, Philadelphia. 6. SHUVALOV, V. A., KLIMOV, V. V., AND KRASNOVSKII, A. A. (1976) Mol. BioL 10, 326-339. 7. IKEGAMI, I. (1976) B&him. Biophya Acta 449, 245-258.
ET AL. 8. TELFER, A., BARBER, J., HEATHCOTE, P., AND EVANS, M. C. W. (1978) Biochim Biophys. Acta 504, 153-164. 9. BENGIS, C., AND NELSON, N. (1975) J. BioL Chem. 250, 2783-2788. 10. BENGIS, C., AND NELSON, N. (1977) J. BioL Chem. 252,4564-4569. 11. BARBER, J. (1980) Biochim. Biophys. Acta 594, 253-308. 12. BARBER, J. (1982) Annu. Rev. Plant Physiol 33, 261-195. 13. LIEN, S., AND SAN PIETRO, A. (1979) Arch B&hem. Biophys. 195,128-137. 14. DAVIS, D. J., KROGMANN, D. W., AND SAN PIETRO, A. (1980) Plant PhysioL 65, 697-702. 15. ITOH, S. (1978) B&him Biophys. Acta 504. 324340. 16. ITOH, S. (1979) Biochim. Biophys. Acta 548, 579595. 17. GROSS, E. L. (1979) Arch. B&hem. Biophys. 195, 198-204. 18. TAMURA, N., YAMAMOTO, Y., AND NISHIMURA, M. (1980) Biochim Biophys. Acta 592, 536-545. 19. HAEHNEL, W., PROPPER, A., AND KRAUSE, H. (1980) Biochim. Biophys. Acta 593, 384-399. 20. PROCHASKA, L. J., AND GROSS, E. L. (1977) Arch. B&hem Biophys. 181, 147-154. 21. BURKEY, K. O., AND GROSS, E. L. (1981) Biochemistry 20,2961-2967. 22. OLSEN, L. F., AND Cox, R. P. (1980) Photo&&em Photobiophys. 1, 147-153. 23. ARNON, D. I. (1949) Plant PhysioL 49, 1-15. 24. MEANS, G. E., AND FEENEY, R. E. (1971) in Chemical Modification of Proteins, pp. 144147, Holden-Day, San Francisco. 25. BHARDWAJ, R., BURKEY, K. O., TRIPATHY, B. C., AND GROSS, E. L. (1982) Plant PhysioL 70,424429. 26. DAVIS, D. J., AND SAN PIETRO, A. (1979) Anal B&hem. 95, 254-259. 27. BUCHANAN, B. B., AND ARNON, D. I. (1971) in Methods in Enzymology (San Pietro, A., ed.), Vol. 23, pp. 413-440, Academic Press, New York. 28. AMESZ, J., AND FORK, D. C. (1967) B&him. Bie phys. Acta 143,97-107. 29. OVERBECK, J. J. A. (1950) in Colloid Science (Kruyt, H. R., ed.) Vol. 1, pp. 115-193, Elsevier, Amsterdam. 30. RUBIN, B. T., AND BARBER, J. (1980) B&him. Biophys. Acta 592, 87-102. 31. GIBBS, H. D., COHEN, B., AND CANNAN, R. K. (1925) Public Health Records 40,649-659. 32. GROSS, E. L., AND GRENIER, J. (1978) Arch. B&hem. Biophys. 187,387-398. 33. MULLETT, J. E., BURKE, J. J., AND ARNTZEN, C. J. (1980) Plant PhysioL 65, 814-822.
OF BIOCHEMISTRY
Vol. 235, No. 2, December,
AND
BIOPHYSICS
pp. 449-460, 1934
Variable Fluorescence of Photosystem I Particles and Its Application the Study of the Structure and Function of Photosystem I B. C. TRIPATHY, Department
J. E. DRAHEIM, of Biochemistry,
Received February
G. P. ANDERSON,
Ohio State University,
AND
to
E. L. GROSS’
Columbus, Ohio 4.9210
21, 1934, and in revised form August ‘7, 1934
Chlorophyll a fluorescence in Photosystem I (PSI) particles isolated according to the method of Bengis and Nelson [J Bid Chem. 252, 4564-4569 (1977)] was found to be dependent on the redox state of both P700 and X (an acceptor on the reducing side of PSI). Addition of dithionite plus neutral red to PSI caused an increase in fluorescence intensity and a shift of the main fluorescence peak from 689 to 674 nm. Addition of electron acceptors such as ferredoxin and methyl viologen decreased the fluorescence yield when added to PSI incubated under anaerobic conditions in the presence of excess dichlorophenol indophenol (DCIPHa). The K,,, for ferredoxin agreed with that determined from direct measurements of ferredoxin reduction, showing that X is a quencher of fluorescence. P700 was also found to be a quencher of fluorescence, since electron donors such as DCIPHB, TMPD, and plastocyanin decreased fluorescence with Km’s nearly identical to those observed for P700+ reduction. Chemical modification of PSI (with ethylene diamine + a water-soluble carbodiimide) to make it positively charged increased the fluorescence yield and shifted the 689-nm peak to 674 nm. The Km’s for DCIPHz and ferredoxin were decreased. In contrast, modification of PSI with succinic anhydride, which increased the net negative charge, increased the Km for ferredoxin. Salts affected the interaction of methyl viologen with PSI. Both anion and cation selectivity were observed. Limited proteolysis increased the Km for both methyl viologen and ferredoxin, indicating that their binding site on PSI was altered. These results suggest that the binding site for ferredoxin is on either the 70- or the 20-kDa subunit of PSI. o 1984 Academic PRESS, IW. Chlorophyll a fluorescence in chloroplasts is emitted mainly from Photosystern II. Its intensity is dependent on the redox state of both P680’ and Q (l-3). Therefore, it has been used to study elec’ To whom correspondence should be addressed. ‘Abbreviations used: PSI, Photosystem I; X, acceptor for Photosystem I; P700, chlorophyll which acts as primary electron donor for PSI; P’700+, oxidized P700; P630, chlorophyll which acts as primary electron donor for PSII: Q, primary electron acceptor for Photosystem II; PC, plastocyanin; MV, methyl viologen; EDA, ethylene diamine; PMSF, phenylmethylsulfonyl fluoride; Hepes, I-(2-hydroxymethyl)1-piperazineethanesulfonic acid; Mes, 4-morpholineethanesulfonic acid; Taps, 3-{[2-hydroxy-l,l-bis(hy-
tron transport in chloroplasts. There have been reports of variable fluorescence from Photosystem I (PSI) in PSI-enriched particles (4-8). The reports quoted above, however, do not clearly elucidate the redox dependence of variable fluorescence in PSI. Our first objective will be to document the existence and redox dependence of variable fluorescence in PSI particles isolated according to the method of Bengis
droxymethyl)ethyl]amino}-1-propanesulfonic acid; DCIPH, dichlorophenol indophenol; TMPD, N,N, N’,N’-tetramethyl-p-phenylene diamine; DCIP, dichlorophenol indophenol. 449
0003-9861/84 $3.00 Copyright All rights
0 19% by Academic Press, Inc. of reproduction in any form resewed.
450
TRIPATHY
and Nelson (9, 10). We will then use variable fluorescence to study the effects of cations, chemical modification, and proteolysis on the properties of the reducing side of PSI. Cations have been shown to have multiple effects on chloroplasts (11, 12). In particular, they have been shown to promote the interaction of plastocyanin (and other negatively charged electron donors) with P700 (13-1’7). A similar enhancement of the reaction with plastocyanin was obtained when PSI particles were made positively charged by modification with ethylenediamine in the presence of a water-soluble carbodiimide (18). Cations have also been shown to act on the reducing side of PSI (19), but this has not been well studied due to the difficulty of directly measuring effects on the reducing side of PSI. We will show that variable fluorescence is a useful tool for these and other studies. MATERIALS
AND
METHODS
Isolation of Photosystem I particles. The PSI particles were isolated according to the method of Bengis and Nelson (9). The chlorophyll/P700 and the chlorophyll a/b ratios were 80 and greater than 18, respectively. Chlorophyll was determined according to the method of Arnon (23). The PSI particles were dialyzed for 12 h against 50 mM Tris-HCl buffer (pH 8.0). Chemical modZfication and trypsin treatment. Chemical modification of PSI particles with EDA and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide was carried out according to the method of Burkey and Gross (21) as derived from Means and Feeney (24). Chemical modification with succinic anhydride was carried out as follows. Isolated PSI particles were dialyzed against 1.0 M sodium carbonate buffer (pH 8.0), containing 0.05% Triton X-100, for 2 h. The PSI particles were then modified by adding 35.7 mmol succinic anhydride/pmol P700 over a l-h period. The reaction was then allowed to go to completion for 0.5 h. The reaction was then quenched using 10 mM Tris-HCl (pH 8.2), containing 0.05% Triton X-100, added in sufficient amount to reduce the chlorophyll concentration to 30 pg/ml. The modified PSI particles were then dialyzed against the same solution. Trypsin treatment of PSI particles was carried out as described previously (25) by adding trypsin directly to the reaction cuvette containing 3 ml 10 mM Hepes/NaOH buffer (pH 7.8) at a chlorophyll: trypsin ratio of 21 by weight. The trypsin inhibitor
ET AL. phenylmethylsulfonyl fluoride (PMSF) was added after 10 min of incubation with trypsin. Chlorophyll a jluorescence measurements. The chlorophyll a fluorescence excitation and emission spectra of PSI particles were measured using a Fluorolog 1902 spectrofluorimeter (Spex Industries). A Tracer NS-570A digital signal analyzer was used to store and analyze the fluorescence spectra. The samples were excited at 436 nm (slit width = 5 nm). The excitation intensity was 500 pW/cm*. The emission spectra were corrected for photomultiplier sensitivity. The slit width for the emission was 5 nm. The fluorescence excitation spectra were not corrected for lamp and photomultiplier sensitivity. However, they were done under identical conditions and, therefore, can be compared. The PSI particles were suspended in 10 mM Hepes/NaOH buffer (pH 7.8) at a chlorophyll concentration of 2 pg/ml for the fluorescence measurements. All experiments were carried out at 20°C. Anaerobic conditions were maintained by adding 20 mM glucose, 40 units/ml glucose oxidase, and 600 units/ml catalase to the reaction mixture. Then, argon was bubbled through the reaction mixture for 5 min, after which the cuvettes were sealed. As glucose oxidase has a weak fluorescence tail in the region 670-690 nm, the amount of fluorescence quenching was calculated by subtracting the background fluorescence of glucose oxidase. Ferredoxin isolation and reductirm. Crude ferredoxin was isolated according to Davis and San Pietro (26), and was dialyzed over night against 50 mM Tris-HCl buffer (pH 8.0). Ferredoxin reduction was carried out under anaerobic conditions in sealed cuvettes as described above. Ferredoxin reduction was measured as a decrease in absorbance at 420 nm, upon continuous illumination, using the Aminco DW-2a spectrophotometer in the double-beam mode. Assay mixtures contained PSI particles (at 10 pg/ ml chlorophyll), 10 mM Hepes/NaOH buffer (pH 7.8), 5 mM sodium ascorbate, 0.33 mM DCIP, 0.05% Triton X-100, and various concentrations of ferredoxin. A 300-W projector lamp was used to illuminate the sample. Red actinic light (4.7 X 10’ ergs cm-r s-i) greater than 650 nm was isolated by placing 3 cm of water and a red long-pass cutoff filter (Corning CS-2-64) between the lamp and the sample compartment. A 620-nm short-pass cutoff filter (Bausch and Lombe NO 90-l-620) was used to prevent the actinic light from reaching the photomultiplier. The rates were calculated using an extinction coefficient of 9.7 mrt-’ cm-i at 420 nm for ferredoxin (27). RESULTS
The eflect of electron donors and acceptors on chlorophyll a jluorescence. First,
we determined whether the chlorophyll a fluorescence spectrum was dependent on
VARIABLE
FLUORESCENCE
the redox state of the reaction center. To do this, we determined the fluorescence emission spectrum under three illumination conditions. These three conditions corresponded to the following states of the reaction center: P+X, PX, and PX-. In each case, P refers to P’700 and X refers to any one of the acceptors on the reducing side of Photosystem I (PSI). In condition (a), the PSI particles were illuminated in the presence of oxygen and 50 PM ascorbate. The rate of P+ reduction by ascorbate was too slow to reduce the P+ produced in the light. In the aerobic state, oxygen acted as the terminal electron acceptor (1’7). Thus, the particles were in the P+X state under these conditions. In state (b), P700+ was reduced using high concentrations of the electron donor dichlorophenol indophenol (DCIPHz), converting the reaction center to the P’7OOX state. In case (c), the PSI particles were illuminated in the presence of dithionite and neutral red, which reduced both the P700 and X (8) and also maintained the sample in an anaerobic state. The reaction center was in the P7OOX- state under these conditions. Figure 1 shows the fluorescence emission spectrum for PSI particles illuminated under these three conditions. In case (a), the PSI particles showed a peak at 689 nm with a shoulder at 678 nm and another small peak at 730 nm. Addition of an electron donor (case b) caused a decrease in the emission at 678 and 730 nm with no shift in the emission spectrum. The fluorescence changes were complete within 20 s (not shown). In case (c), in which both P and X were reduced, the fluorescence intensity was increased and the 689nm peak was shifted to 674 nm. These results suggest that both P and X are quenchers of chlorophyll a fluorescence in PSI particles. Since we have shown that both P700 and X are quenchers of fluorescence, we can use variable fluorescence to study the effects of cations, pH, and chemical modification on both the oxidizing and reducing sides of PSI. This technique is particularly useful for the reducing side of PSI due to the paucity of methods for measuring it directly.
IN PHOTOSYSTEM
900
451
I
-
800
WAVELENGTH
(nm)
FIG. 1. Fluorescence emission spectra of PSI particles under various conditions. The excitation wavelength was 436 nm. Other conditions were as described under Materials and Methods. (- - -) Control PSI (no added electron donors or acceptors; ( * * * ) PSI treated with 20 pM DCIPH2; (p) PSI particles with 20 mM dithionite + 2 pM neutral red added.
In each case, we want to determine the effect of various treatments on the Km and I’,,,,, for electron transport using different donors and/or acceptors. Therefore, we need to be certain that there is a direct relationship between these quantities determined from variable fluorescence measurements and those determined by more conventional means; namely, measurements of P700+ or ferredoxin reduction. In particular, we must make sure that the Km’s are independent of light intensity. We have measured variable fluorescence in terms of percentage quenching. Percentage quenching is defined according to %Q = lOO*(F_n - Fn)/F-,,,
PI
where FD and FpD are the fluorescence intensities obtained in the presence and absence of an electron donor respectively. We have chosen this formulation because we can measure %Q more easily
452
TRIPATHY
and accurately under all conditions than we can measure the constant (FO) and variable (Fv) fluorescence levels. Moreover, %Q was proportional to the concentration of reduced P700 (Fig. 2). Under conditions of high light intensity, l/‘%Q is related to l/[D] according to &={L$}.{1+~},
[2]
where D is the donor concentration, Pt is the total amount of P700, 9, is the quantum yield for constant fluorescence (F), and \k, is a pseudo-quantum yield for the variable part of fluorescence (Fv) with M-l units. Kd is the concentration of donor required for half-maximal quenching and should be equal to the K, determined by other means. Q,,,,,, which is the percentage quenching observed at infinite donor concentration, is related to the maximum rate of electron transport. This equation is derived in the Appendix. An analogous equation can be derived for fluorescence quenching by an electron acceptor.
FIG. 2. The effect of DCIPHa on fluorescence quenching and the steady-state level of P700 oxidation. Light-induced P700 oxidation was carried out according to the method of Gross (17). The reaction mixture contained 10 mM Hepes buffer, pH 7.8, 2 mM ascorbate, and PSI particles containing 2 pgg/ml chlorophyll. The PSI particles were titrated with increasing concentrations of DCIP. The steady-state level of the light-induced P700 oxidation was determined after each addition. Fluorescence quenching was determined as described for Fig. 4. The light intensity was adjusted to 400 pw m-a for both parts of the experiment.
ET AL.
We determined the Kd as a function of light intensity for the electron donors DCIPHB and plastocyanin. In the case of plastocyanin, the Kd was 3.3 and 2.9 pM at 250 and 760 PW cm-’ intensity, respectively. For DCIPHz, Kd was 33 PM at both light intensities. Thus, we can conclude that the Kd)s are independent of light intensity. If Eq. [2] is correct, then other electron donors to PSI should also quench fluorescence. In each case, the Kd’s should be the same as the K,‘s determined by other methods. This is true for both plastocyanin and TMPD (Table I). Furthermore, plastocyanin quenched fluorescence only in the presence of Mgaf ions, a condition which was also required for P’700 reduction. In addition, the electron acceptors methyl viologen and ferredoxin also quenched fluorescence. As in the case of DCIPHz, both the 680- and 730-nm peaks decreased in intensity (not shown). Also, the fluorescence changes were complete within 20 s (not shown). In the case of ferredoxin, the extent of quenching was essentially the same using both blue (430 nm) and red (660 nm) exciting light. We also measured ferredoxin reduction directly by monitoring absorption changes at 420 nm under anaerobic conditions. We obtained a K, of 12.5 PM and a V,,, of 16 pmol ferredoxin reduced mg chlorophyll-’ h-l. This agrees very well with the value of 12 PM obtained from fluorescence measurements (Table I.). This agreement shows that fluorescence quenching is a good measure of activity on the reducing side of PSI. The question arises as to whether the electron donors and acceptors could be quenching fluorescence by reacting directly with the pigment bed. This is unlikely in the case of the large watersoluble molecules such as plastocyanin and ferredoxin, which would not be able to penetrate the membrane. In the case of DCIPHz, we found that 100 PM was unable to quench fluorescence of chlorophyll a in 80% acetone [See also (28)]. The eflect of chemical rnoo!i&xtion on Photosystem Ijtwrescence. Since we have
VARIABLE
FLUORESCENCE
IN PHOTOSYSTEM
TABLE
I
THE EFFECT OF ELECTRON DONORS AND ACCEPTORS ON THE QUENCHING FLUORESCENCE
Electron donor or acceptor DCIP * TMPD
Plastocyanin -Mg2 +3.3 mM MgCla Ferredoxin’ Methyl viologen
453
I
IN PHOTOSYSTEM
OF CHLOROPHYLL
a
I PARTICLES
Q Kia (PM)
(?6 Quk:hing)
Knad (PM)
Reference
(17)
32 32
30 29
29 29
No effect 2.6 12.5 460
30 36 24
3.2 12
(25)
a Fluorescence quenching was determined as a function of donor or acceptor concentration, after which from double reciprocal plots of fluorescence quenching vs donor or the Kd and Q,,,- were determined acceptor concentration. * Experiments using electron donors were done under aerobic conditions as described for Fig. 2. c Experiments with electron acceptors were done under anaerobic conditions as described under Materials and Methods. Each concentration represents a separate sample. Other conditions were as described for Fig. 3. dThe Kd’s were compared with the K,,,‘s determined from either the reduction of P’700+ or ferredoxin. Where indicated, the Km values were taken from the literature.
shown that variable fluorescence is a measure of the ability of electron acceptors such as MV and ferredoxin to interact with X, we can use this technique to study the effects of cations and net charge on PSI. We can change the charge on PSI by three methods: chemical modification (21), addition of salts to screen the existing charges (18, 29, 30), and changing the pH. We have examined the effect of two types of chemical modification on the fluorescence properties of PSI. In the first case, we modified carboxyl groups with ethylenediamine in the presence of a water-soluble carbodiimide (21). This reaction changes the isoelectric point of PSI from pH 4.7 to 9.5, converting it into a positively charged protein. The second modification involved the reaction of amino groups with succinic anhydride. This reaction adds a negative charge to the pK to 4.5 amino groups, lowering (Anderson and Gross, unpublished experiments). EDA modification shifted the 689-nm emission peak to 674 nm and also increased the intensity of both the 674- and ?‘30-nm peaks as compared to the unmodified control (Fig. 3). As in the case of
control PSI, addition of dithionite and neutral red caused an increase in the fluorescence intensity (Fig. 3). However, 9000
I ’
I
A’
8000 -
640
660
680
700
WAVELENGTH
720
740
760
(“ml
FIG. 3. The effect of DCIPHI and dithionite on the fluorescence emission spectrum of EDA-PSI particles. Conditions were as described for Fig. 2. (- - -) Con+ 20 mM dithitrol; ( + . . ) + 20 pM DCIPHr; (p) onite and 2 PM neutral red.
454
TRIPATHY
ET AL.
in contrast to dithionite-treated control P700+ reduction (1’7). Modification also PSI, there was no additional blue shift of caused an increase in the Q,,,,, from 33 to the 674-nm emission peak. 42%. The increase in Q,,, reflects an inChemical modification with EDA caused crease in the I’,,, for electron transport. a decrease in the Kd for DCIPHz from 32 We also examined the effect of EDA to 12 PM (Fig. 4). This was expected since modification on the interaction of PSI a negatively charged electron donor such with the electron acceptors ferredoxin and methyl viologen. The Km for ferredoxin as DCIPHz should react more favorably with positively charged PSI than with decreased from 12 to 4 PM as expected. No negatively charged PSI. These results are quenching was observed with methyl violconsistent with those reported earlier for ogen due to steric repulsion between positively charged MV and positively charged EDA-PSI. I 1 I 1 I I The shape of the fluorescence emission spectrum was not altered upon succinic anhydride modification (not shown). However, the Km for negatively charged ferredoxin was increased from 12 to 29 I.LM as expected. No effect was observed for methyl viologen, since the reaction with negatively charged control PSI was already favorable. The e#ect of salts on fluorescence quenching. The Guoy-Chapman theory (29, 30) predicts that salts such as NaCl and MgS04 should inhibit the favorable reaction between MV and control PSI. Saltinduced inhibition of fluorescence quenching was observed (Fig. 5A). However, there was a cation selectivity which is not predicted by Guoy-Chapman theory. For example, magnesium salts were more effective on a cation concentration basis than were calcium salts. Also, NaCl was less effective than predicted by theory. Conversely, anions should screen the positive charges on EDA-PSI, facilitating its interaction with MV. Some salts did promote fluorescence quenching (Fig. 5B). 0 IO 20 30 40 50 60 However, anion selectivity was observed [ DCIPH,] (PM) in that acetate anions were more effective than chloride ions. Moreover, magnesium FIG. 4. DCIPHr concentration dependence of salts were totally ineffective at promoting quenching of variable fluorescence in control (-) and EDA-PSI (- - -). Inset; Lineweaver-Burk plot fluorescence quenching. One possible exfor variable fluorescence as a function of DCIPHc planation is that magnesium ions compete concentration for control (-) and EDA-PSI (- - -). with MV for its binding site. The PSI particles were titrated with increasing In contrast to the results obtained with concentrations of DCIP in the presence of 59 PM Na MV, those obtained with DCIPH2 did obey ascorbate. The fluorescence intensity was determined the Guoy-Chapman theory. Both NaCl after each addition. The excitation wavelength was and MgS04 increased the interaction be436 nm. The emission wavelengths were 689 and 674 tween DCIPHz and negatively charged nm for control and EDA-PSI, respectively. Other PSI at pH 8.0. A plot of log percentage conditions were as described under Materials and quenching (%Q) vs the ionic strength to Methods.
cI
VARIABLE
FLUORESCENCE
SALT CONCENTRATION I
I
I
I
(mM) /
I
2 B No,SO, CH,COONo NoCl 4 so.
c:
WA IO
20
30
40
50
SALT CONCENTRATION
60
70
(mM)
FIG. 5. The effect of salts on methyl viologeninduced quenching of fluorescence for both control and EDA-modified PSI particles. The ability of methyl viologen to quench fluorescence was determined as a function of salt concentration for both control and EDA-modified PSI particles. PSI particles were maintained under anaerobic conditions with glucose, glucose oxidase, and catalase. Ascorbate and DCIP were added as described for ferredoxin reduction in order to maintain P700 in the reduced state and to promote electron transport resulting in the reduction of X. Other conditions were as described under Materials and Methods. (A) Control PSI; (B) EDA-PSI.
IN PHOTOSYSTEM
455
I
The Kd for the DCIPH2-induced quenching increased with increasing pH above pH 7 for control PSI and decreased over the same pH range for EDA-modified PSI. The results were corrected for the pH dependence of the interaction of ascorbate with PSI (13). These results can be explained on the basis of the effect of pH on the charge of DCIPH2. It is neutral below a pH of 7 (31) and negatively charged above it. Thus, above pH 7, the negative charge would facilitate its interaction with EDA-PSI while inhibiting that with control PSI. Qmax increased with increasing pH for both control and modified PSI. The pH dependence for Q,,, corresponds to a change in V,,, which reflects changes in the electron transfer reactions themelves (17). &ma,was greater for EDA-PSI than control PSI at all pH values tested. This increase in QmaXis probably related to the increase in the I’,,,,, observed for P700+ reduction under the same conditions (21). The pH dependence for ferredoxin-induced quenching is shown in Fig. 7B. There was a gradual increase in the Km for the interaction of ferredoxin with control PSI as the pH was increased, reflecting the increasing negative charge on both the PSI particles and ferredoxin. In contrast, Kd for MV was pH independent
the power l/2 (IS-“.5) produced a straight line according to (Fig. 6). log %Q = log %Q, - 0.078q~(IS)-~.~,
[3]
where %Q and %Q,, are the values in presence and absence of salt, respectively; x is the charge on the electron donor, q is the net charge on the particle, and IS is the ionic strength. A slope of -0.073 (Fig. 6) corresponds to a net charge (q) on the oxidizing side of PSI particles of -0.94 PC! cme2. Identical results were obtained using MgS04. Our results agree with those of Itoh (16) and Tamura et al. (18), who obtained -0.84 PC cmp2. The effect of pH on jluwescence quenching. Figure 7A shows the effect of pH on DCIPHz-induced fluorescence quenching.
I 0
/ 2
I 4
I 6
1 0
I IO
[Is]-“2 FIG. 6. Guoy-Chapman plot of fluorescence quenching as a function of ionic strength. Fluorescence quenching was determined as a function of NaCl concentration for control PSI particles using DCIPHa as the electron donor. Other conditions were as described for Fig. 3 and under Materials and Methods. The data were plotted according to Eq.
u51.
TRIPATHY
456
ET AL.
viologen on ferredoxin reduction in order to determine whether they interact at the same site on the reducing side of PSI. The observation (Fig. 8) that MV is a competitive inhibitor of ferredoxin reduction indicates that the two electron acceptors do act at the same site.
The eflect of proteolysis on j-luorescence quenching. Limited proteolysis is a useful
‘\ IO
IO
i
o-o 6
7
9
8 PH
40
0.8
C 0.6
30
r;, ” a d
Gl 20
0.4
Qm*
IO 0H 6
r 2 x’
technique for determining which polypeptides in a multisubunit complex such as PSI are required in order for a particular electron transport reaction to occur. We have previously shown that tryptic digestion of the Bengis and Nelson PSI preparation affected both the 70- and 20-kDa polypeptides (25). A decrease in the rate of P700+ reduction was also observed. Trypsin treatment had only a small effect on the fluorescence emission spectrum, consisting of a 10% decrease in emission intensity at 680 nm and a 10% increase in the emission at ‘740 nm (25). Tryptic digestion increased the Kd for both ferredoxin and methyl viologen as measured by the variable fluorescence technique (Table II). There was almost no effect on the V,,, in either case. Thus,
0.2
7
PH
s
9
0
FIG. 7. The effect of varying the pH on fluorescence quenching using control (p) and EDA-modified (- - -) PSI particles. (A) DCIPHa was the electron donor under aerobic conditions; (B) ferredoxin was the electron acceptor under anaerobic conditions; (C) methyl viologen was the electron acceptor under anaerobic conditions. The PSI particles were dialyzed into the appropriate buffers. Mes buffer was used between pH 5.5 and 6.7, Hepes buffer was used between pH 6.8 and 7.9, and Taps buffer was used between pH 8.0 and 9.0; all buffers were 10 mM. Other conditions were as described for Table I and under Materials and Methods.
(Fig. 7C), as expected if the interaction of MV was already favorable at low pH. Qmax’v however, did increase between pH 8 and 9.
Reduction of ferredoxin of MV: We studied
in the presence
the effect of methyl
'L f 5 i
0.8 "
lg 0.6 x
2 0.4 2 iG 0 0.2
E
2. [Ferredoxinl-
I (PM)-’
FIG. 8. The effect of methyl viologen on the rate of ferredoxin reduction. The rate of ferredoxin reduction was determined as a function of ferredoxin concentration in the presence (A) and absence (0) of 0.5 mM methyl viologen, after which the data were plotted in double-reciprocal form. Other conditions were as described under Materials and Methods section.
VARIABLE
FLUORESCENCE TABLE
IN PHOTOSYSTEM
457
I
II
THE EFFECX OF TRYPSIN TREATMENT AND SUCCINIC ANHYDRIDE MODIFICATION ON THE REDUCING SIDE OF PHOTOSYSTEM I
Electron Methyl
Control
acceptor
viologen
Kd (mM)”
Ferredoxin
Trpsin-treated PSI
PSI
Succinic anhydride modified
Q
0.46 21
2.66 26
0.47 31
~:PM) Qmar
12 31
24
29
36
50
a Fluorescence quenching was determined as a function of methyl viologen or ferredoxin concentration for either control or trypsin-treated PSI particles, after which the Kd and Q,.. were calculated. Other conditions were as described under Materials and Methods.
trypsin treatment interferes with the binding of both electron acceptors to PSI. We can conclude that one or both of the affected polypeptides (i.e., the 70- and 20kDa polypeptides) are required for the interaction of these electron acceptors with PSI, and one of them may contain the binding site for ferredoxin. Excitation spectra of control and EDAPSI particles in the presence and absence of dithionite. The excitation spectrum for the 730-nm emission peak of control PSI shows excitation maxima at 436 and 674 nm, with carotenoid bands at 470 and 495 nm (Fig. 9). The excitation spectrum for the 689-nm peak was the same as for the
\ 400
440
._--_480
---% -520
730-nm peak (not shown). Both excitation spectra match the absorption spectra (not shown). The absorption spectrum also agrees with that obtained earlier by Bengis and Nelson (9). In EDA-PSI particles, the excitation peaks were blue-shifted to 432 and 666 nm, respectively. Again, they corresponded to the absorption peaks (not shown). The blue shift in excitation and absorption suggests a structural reorganization of the PSI particles. A similar blue shift was observed upon addition of MgC& to control PSI (32). The blue shift was not due to release of free chlorophyll since modification increased the ability of the particles to utilize 650-nm light to
5M)
WAVELENGTH
6&a
640
660
720
(nm)
FIG. 9. Excitation spectra of control (p) and EDA-modified (- - -) PSI particles. The emission wavelength was 730 nm. Other conditions were as described under Materials and Methods. The fluorescence intensities were normalized at the blue excitation peak.
458
TRIPATHY
oxidize P700 (21). Moreover, the Chl/P700 ratio did not change upon modification (Burkey and Gross, unpublished observations). In contrast, treatment of control PSI with dithionite caused no blue shift in the absorption spectrum (not shown), although a blue shift was observed in the emission spectrum. Thus, there is a fundamental difference between dithionite treatment and EDA modification. In the case of dithionite, there was a shift in emission without a change in excitation or absorption (i.e., a change in Stoke shift). On the other hand, EDA modification caused both an 8-nm change in the excitation wavelength and a 15-nm change in emission wavelength. DISCUSSION
AND
CONCLUSIONS
The results presented above show that the fluorescence intensity of PSI particles isolated according to the procedure of Bengis and Nelson (9, 10) is dependent on the oxidation state of both X and P700. The evidence that X is a quencher of fluorescence consists of the following. A high fluorescence intensity was observed when PSI was illuminated in the presence of DCIPHz under anaerobic conditions. This produces the PX- state. Addition of the electron acceptors methyl viologen and ferredoxin decreased the fluorescence which would be expected if X were a quencher (i.e., they produce the PX state). Moreover, the observation that the Kd determined for ferredoxin from fluorescence quenching measurements agrees with the Km obtained from direct measurements of ferredoxin reduction supports this interpretation. Illumination of PSI in the presence of dithionite and neutral red produced an even higher intensity than illumination in the presence of DCIPH2 because it is more effective at producing the PX- state. Our results agree with those of Telfer et al. (8), who also observed that X (rather than X-) was a quencher of fluorescence. This is analagous to fluorescence quenching by Q in Photosystem II. The evidence of fluorescence quenching by P consists of the following. When PSI particles were illuminated under aerobic
ET AL.
conditions in the absence of electron donors, the P+X state was created since Xcan react with oxygen under these conditions (17). Addition of electron donors such as reduced DCIP, TMPD, and plastocyanin quenched the fluorescence by producing the PX state. The Kd’s for these donors agreed with the Km’s determined from measurements of P700+ reduction. In addition, the absolute requirement for M$+ ions in the case of plastocyanin was the same as for P700+ reduction. Thus, our results lead to the conclusion that it is P which is the quencher of fluorescence. This is reasonable in that fluorescence should be less when the trap is open and capable of doing photochemistry. However, in contrast to our results, both Telfer et al. (8) and Ikegami (7) found P700+ to be a quencher. One way to reconcile these results is that both P and P+ are quenchers of fluorescence, but that P is a greater quencher under the conditions of our experiments. Three peaks were observed in the emission spectrum of control PSI, at 674, 689, and 730 nm. Reduction with dithionite increased the 674- and 730-nm peaks. Addition of electron donors and acceptors affected these same peaks. We can conclude that both the 674- and 730-nm peaks belong to the core antenna and can respond to the state of the reaction center. Our 674-nm peak may be analogous to the 695-nm peak observed by Mullet et al. (33). However, they observed no changes in the 730-nm peak. Chemical modification with EDA caused a blue shift in both the absorption and emission spectra. EDA modification may cause a conformational change in the PSI particles, which shifts more of the chlorophyll into the “core” antenna. In this paper, we have shown not only that P and X act as quenchers of fluorescence in PSI particles, but also that variable fluorescence is a powerful technique for examining electron transport in PSI. In particular, it is useful for studying the reducing side of PSI which is difficult to study by other methods. In our preliminary investigations, reported herein, we have shown that salts affect the reducing side of PSI in a manner not totally con-
VARIABLE
FLUORESCENCE
sistent with the Guoy-Chapman theory. Moreover limited proteolysis experiments in conjunction with variable fluorescence measurements have helped us to narrow down the location of the binding site for ferredoxin. In the future, we will use this technique to study other types of modification on the structure and function of PSI. This technique should also prove useful for examining the effect of different growth conditions on the state of PSI. APPENDIX
The fluorescence intensity of PSI depends upon two components. One component is the constant fluorescence, F,. The other component is the variable fluorescence, F,. The measured fluorescence intensity of PSI before and after addition of an electron donor are designated by F-,, and F,, respectively. Thus
FD = F, + F, = Qo.I+
XJXVsI.P+,
[l]
where a0 is the quantum yield for F,, 9, is a pseudo-quantum yield for F, (in M-’ units), 1 is the intensity of light, and P+ and P are the molar concentrations of oxidized and reduced P700, respectively. Although P+ might be a quencher of PSI fluorescence, it is our contention that P must be a more effective quencher. If this is true, then the percentage quenching, %Q, should be proportional to the concentration of reduced P700. Due to the ease of accurately measuring FD and F-D, we shall define %Q as
%Q = { F-;-DFD}dOO.
[Z]
IN PHOTOSYSTEM
459
I
1 -- 1 P, + (%3/Q”) Pt - P+ %Q - 100
P+I%P+ and
Fl
k-2 P+ + D c (P+D) 2 P + D+.
P, = P + P+ + (P+D),
[71
In this description kl is a second-order rate constant for the formation of P+; kz is the rate constant for the formation of (P+D); and k3 and k4 are the rate constants for the decay of (P+D) to reactants and products, respectively. In the steady state we have dP/dt = d(P+D)/dt dP/dt = -kJP
PI
= 0,
PI
+ k,(P+D),
and d(P+D)/dt = &P+D - (k3 + k,)(P+D) = 0. After combining
and rearranging
[lo]
we have
klIKd + kJD + k,D P’ klIKd + kJD
WI where
Kd = (k4 + k,)/k, = K, If we let
B = klIKd + kJD + k4D klIKd + kJD ’
WI
condenses to
In the absence of an electron donor, P+ equals the total amount of P700 (PJ. The fluorescence prior to the addition of donor is, therefore,
where
[51
If we consider light as a substrate, then the kinetics of formation and decay of P’ and (P+D) can be simply described by
then Eq. [ll]
F-D = @-I+ \k,*I*Pt
*
P31
Substitution of Eq. [13] into Eq. [5] gives us 131 1 B(D + Kd) ?CQ= (B l)(Kd + B)D [41
and where (P+D) is the complex formed between P+ and the electron donor D. Combining Eqs. [l] through [4] gives us
x
1
i
+
(@O/Q”)
x
pt
1
1
100 .
P41
TRIPATHY
460
In general, a plot of (l/%Q) vs (l/D) will be both nonlinear and a function of light intensity. However, the intercept on the 2 axis (where (l/%Q) = 0) is B (B - l)(&
+ @Do
where (l/Do) is the intercept on the negative x axis. When (Do) is equal to Kd, the left hand side of Eq. [15] is identically zero. Thus, the intercept should be independent of light intensity and equal to the Km determined by other methods. In the special case where kl*I $s k4, B - 1.0 and Eq. [14] simplifies to
1+v}.t
[16]
Using Eq. [16], a plot of (I/%&) vs (l/D) will now be linear and light intensity independent. It will also have the form of the familiar Lineweaver-Burk plot. Since the measured curves are linear (see text), the klI 9 k4 condition probably holds under our experimental conditions. ACKNOWLEDGMENTS This work was supported in part by an Ohio State University postdoctoral fellowship to B. C. Tripathy. We wish to thank Dr. Robert T. Ross and Dr. Michael Marchiarullo for their helpful suggestions in the preparation of this manuscript. REFERENCES 1. DUYSENS, L. N. M., AND SWEERS, H. E. (1963) in Studies in Microalgae and Photosynthetic Bacteria (Japanese Society of Plant Physiol., eds.), pp. 353-372, Univ. of Tokyo Press, Tokyo. 2. OKAYAMA, S., AND BUTLER, W. L. (1972) B&him. Biophys. Acta 267, 523-527. 3. GOEDHEER, J. C. (1972) Annu. Rev. Plant PhysioL 23,87-112. 4. KARAPETYAN, N. V., KLIMOV, V. V., AND KRAYNOVSKII, A. A. (1973) Photosynthetica 7, 330337. 5. KARAPETYAN, N. V., AND RAKHIMBERDIEVA, N. A. (1981) in Proceedings of the 5th International Photosynthesis Congress (AkoyunogIOU, G., ed.), Vol. 1, pp. 337-346, Balban, Philadelphia. 6. SHUVALOV, V. A., KLIMOV, V. V., AND KRASNOVSKII, A. A. (1976) Mol. BioL 10, 326-339. 7. IKEGAMI, I. (1976) B&him. Biophya Acta 449, 245-258.
ET AL. 8. TELFER, A., BARBER, J., HEATHCOTE, P., AND EVANS, M. C. W. (1978) Biochim Biophys. Acta 504, 153-164. 9. BENGIS, C., AND NELSON, N. (1975) J. BioL Chem. 250, 2783-2788. 10. BENGIS, C., AND NELSON, N. (1977) J. BioL Chem. 252,4564-4569. 11. BARBER, J. (1980) Biochim. Biophys. Acta 594, 253-308. 12. BARBER, J. (1982) Annu. Rev. Plant Physiol 33, 261-195. 13. LIEN, S., AND SAN PIETRO, A. (1979) Arch B&hem. Biophys. 195,128-137. 14. DAVIS, D. J., KROGMANN, D. W., AND SAN PIETRO, A. (1980) Plant PhysioL 65, 697-702. 15. ITOH, S. (1978) B&him Biophys. Acta 504. 324340. 16. ITOH, S. (1979) Biochim. Biophys. Acta 548, 579595. 17. GROSS, E. L. (1979) Arch. B&hem. Biophys. 195, 198-204. 18. TAMURA, N., YAMAMOTO, Y., AND NISHIMURA, M. (1980) Biochim Biophys. Acta 592, 536-545. 19. HAEHNEL, W., PROPPER, A., AND KRAUSE, H. (1980) Biochim. Biophys. Acta 593, 384-399. 20. PROCHASKA, L. J., AND GROSS, E. L. (1977) Arch. B&hem Biophys. 181, 147-154. 21. BURKEY, K. O., AND GROSS, E. L. (1981) Biochemistry 20,2961-2967. 22. OLSEN, L. F., AND Cox, R. P. (1980) Photo&&em Photobiophys. 1, 147-153. 23. ARNON, D. I. (1949) Plant PhysioL 49, 1-15. 24. MEANS, G. E., AND FEENEY, R. E. (1971) in Chemical Modification of Proteins, pp. 144147, Holden-Day, San Francisco. 25. BHARDWAJ, R., BURKEY, K. O., TRIPATHY, B. C., AND GROSS, E. L. (1982) Plant PhysioL 70,424429. 26. DAVIS, D. J., AND SAN PIETRO, A. (1979) Anal B&hem. 95, 254-259. 27. BUCHANAN, B. B., AND ARNON, D. I. (1971) in Methods in Enzymology (San Pietro, A., ed.), Vol. 23, pp. 413-440, Academic Press, New York. 28. AMESZ, J., AND FORK, D. C. (1967) B&him. Bie phys. Acta 143,97-107. 29. OVERBECK, J. J. A. (1950) in Colloid Science (Kruyt, H. R., ed.) Vol. 1, pp. 115-193, Elsevier, Amsterdam. 30. RUBIN, B. T., AND BARBER, J. (1980) B&him. Biophys. Acta 592, 87-102. 31. GIBBS, H. D., COHEN, B., AND CANNAN, R. K. (1925) Public Health Records 40,649-659. 32. GROSS, E. L., AND GRENIER, J. (1978) Arch. B&hem. Biophys. 187,387-398. 33. MULLETT, J. E., BURKE, J. J., AND ARNTZEN, C. J. (1980) Plant PhysioL 65, 814-822.