Dec 10, 1993 - Wolfgang Haehnel, Thomas Jansen1,. Kathrin Gause, Ralf Bernd Klosgen1 ... Communicated by R.Herrmann. Mutant plastocyanins with Leu at ...
The EMBO Journal vol. 1 3 no.5 pp. 1028 - 1038, 1994
Electron transfer from plastocyanin to photosystem I
Wolfgang Haehnel, Thomas Jansen1, Kathrin Gause, Ralf Bernd Klosgen1, Bernd Stahl2, Doris Michl1, Barbel Huvermann, Michael Karas2 and Reinhold G.Herrmann1 Lehrstuhl fuir Biochemie der Pflanzen, Albert-Ludwigs-Universitdt, Schunzlestrasse 1, D-79104 Freiburg, 'Botanisches Institut, Universitiit Miinchen, Menzinger Strasse 67, D-80638 Miinchen and 2Institut fiir Medizinische Physik und Biophysik, Westfilische Wilhelms-Universitiit, Robert-Koch-Strasse 31, D-48149 Miinster,
Germany Communicated by R.Herrmann
Mutant plastocyanins with Leu at position 10, 90 or 83 (Gly, Ala and Tyr respectively in wildtype) were constructed by site-specific mutagenesis of the spinach gene, and expressed in transgenic potato plants under the control of the authentic plastocyanin promoter, as well as in Escherichia coli as truncated precursor intermediates carrying the C-terminal 22 amino acid residues of the transit peptide, i.e. the thylakoid-targeting domain that acts as a bacterial export signal. The identity of the purified plastocyanins was verified by matrixassisted laser desorption/ionization mass spectrometry. The formation of a complex between authentic or mutant spinach plastocyanin and isolated photosystem I and the electron transfer has been studied from the biphasic reduction kinetics of P700+ after excitation with laser flashes. The formation of the complex was abolished by the bulky hydrophobic group of Leu at the respective position of G10 or A90 which are part of the conserved flat hydrophobic surface around the copper ligand H87. The rate of electron transfer decreased by both mutations to < 20% of that found with wildtype plastocyanin. We conclude that the conserved flat surface of plastocyanin represents one of two crucial structural elements for both the docking at photosystem I and the efficient electron transfer via H87 to P700+. The Y83L mutant exhibited faster electron transfer to P700+ than did authentic plastocyanin. This proves that Y83 is not involved in electron transfer to P700 and suggests that electron transfer from cytochromef and to P700 follows different routes in the plastocyanin molecule. Plastocyanin (Y83L) expressed in either E.coli or potato exhibited different isoelectric points and binding constants to photosystem I indicative of differences in the folding of the protein. The structure of the binding site at photosystem I and the mechanism of electron transfer are discussed. Key words: electron transfer complex/photosystem I/ plastocyanin/site-directed mutagenesis
Introduction Plastocyanin is a single-copper protein which functions in photosynthetic electron transport as a mobile electron carrier 1028
in the thylakoid lumen. It is reduced by cytochromef in the cytochrome b6lf complex and oxidized by the reaction center chlorophyll P700 of photosystem I (PSI). Its crucial role, its well-defined structure of a Class I copper protein (Guss and Freeman, 1983; Guss et al., 1986) and the analogous functions of cytochrome c2 in photosynthetic bacteria (Tiede et al., 1993), of cytochrome c6 in cyanobacteria and to some extent of cytochrome c in the respiratory chain have generated substantial general interest in unravelling the reaction mechanism of this protein. In contrast to cytochrome c which bears a single electron transfer site at the surface of the molecule (Augustin et al., 1983), structural features (Guss and Freeman, 1983) and the interaction with inorganic complexes (Cookson et al., 1980; Sykes, 1985) have suggested at least two possible routes for electron transfer with the reaction partners of plastocyanin. One involves an outer sphere mechanism through H87 which is accessible from the flat hydrophobic surface at the 'northern' end of the molecule formed by residues that are highly conserved in plant plastocyanins (e.g. GI0, L12, G34, F35 and A90; Guss and Freeman, 1983). This hydrophobic area in contact with the solvent was proposed to act as a recognition site for redox partners or for orienting groups on the membrane. An alternative route includes Y83 which is surface exposed in a niche and surrounded by two 'acidic patches' of negatively charged carboxyl residues at positions 42-45 and 59-61. From Y83 a through-bond electron pathway could extend via C84 to the copper center (Lowery et al., 1993). A third possibility concerns the longitudinal axis in the plastocyanin interior which houses an invariant mesomeric system of aromatic amino acid residues (Y80, F41, F82 and F14; Guss and Freeman, 1983). The arrangement of these residues is almost perpendicular to each other so that a potential for strongly interacting ir-electron systems could be suspected, although the long distance and the contributions by inter-state coupling seem to be unfavorable for this electron transfer route. The possibility of different electron transfer sites suggests that plastocyanin could interact with cytochrome f and PSI at different sites. Chemical modification of the acidic patch (Beoku-Betts et al., 1983; Anderson et al., 1987) was found to decrease the electron transfer from cytochromef. It has been proposed that this patch is a docking site for the two electron carriers and that Y83 is involved in electron transfer from cytochrome f Measurements with mutant plastocyanin containing Phe or Leu at the position of Y83 (He et al., 1991) indicated that this residue forms part of the electron transfer route. However, the effects of other mutations lead to contradicting conclusions. In contrast to the chemical modification, the loss of a charge at the negative patch by site-directed mutation of D42 to Asn did not change the rate of the reaction (Modi et al., 1992). Moreover, L12 at the flat hydrophobic region which is very close to H87 was shown to exert a significant effect on the electron transfer from cytochromef The interaction of plastocyanin with PSI © Oxford University Press
Electron transfer from plastocyanin to PSI
is even less clear. Although the observation that the replacement of L12 by Glu inhibits electron transfer to P700 has suggested an important role of the 'northern' surface for this electron transfer (Nordling et al., 1991), the extra negative charge could also significantly change the orientation prior to the electron transfer between these reactants (see Haehnel, 1986). An interaction of negative charges of plastocyanin with positive residues of subunit III, the product of the gene psaF, has been shown to be crucial for both the binding of plastocyanin and fast electron transfer to PSI (Hippler et al., 1989). Covalent cross-linking of plastocyanin to PSI stabilizes the active complex as indicated by the fast electron transfer with a halftime of 12-14 As. The cross-linked product shows a 1: 1 stoichiometry of amino acid sequences from subunit IH and plastocyanin (Hippler etal., 1990). In principle, structure-function relationships can be approached in several ways: by chemical cross-linking and modifications, reconstitution, compartment-specific transformation (and in eukaryotes also by compartment-alien transformation) and by use of antisense RNA. We have used recombinant DNA technology combined with isolated PSI complexes and fast laser flash techniques as an approach to the outlined questions. The available information as well as the fact that plastocyanin originates in a single nuclear gene in spinach (Rother et al., 1986) formed the incentive for analyzing the effects of mutations on electron transfer in the reconstituted system. Since detailed studies about molecular mechanisms require substrates in non-limiting amounts, we have investigated various expression systems for their application in the large-scale synthesis of the protein. We have checked the mutations at the protein level by mass spectrometry and analyzed the binding and the rate constants of electron transfer to PSI from laser-induced reaction kinetics. Our study bears on two of the three abovementioned routes. Preliminary accounts of this work can be found in Jansen (1991) and Haehnel et al. (1992).
Results The characteristics of the plastocyanin gene constructs and their products were monitored in three ways: (i) by sequence
i (It N \ Nvr 15 .N
ti-
I
I
G N.I I1-2 153ll
_; I 5 al) , tl -
1
2 5
I- 3 O -
analysis at all relevant stages of analysis including transgenes in plant material, (ii) by in organello import with isolated spinach chloroplasts, and (iii) by Northern and Western analysis of total cellular RNA and protein respectively,from transgenic potato. 35S-labelled translation products made in vitro from transcripts of linearized (PstI) recombinant pBSC M13+ (T3 polymerase) in wheat germ or rabbit reticulocyte lysates could be imported by isolated spinach chloroplasts (Figure 1). All proteins were correctly translocated to the thylakoid lumen and processed, and were indistinguishable in size from the native form, as judged by their protease resistance in isolated thylakoids and co-electrophoresis of the imported protein with the authentic plastocyanin (Figure 1) in several gel systems. This finding established also that none of the mutants significantly impaired import and routing processes within the organelle (see Discussion). All translation products and imported proteins could be selectively precipitated with a monospecific polyclonal antiserum against spinach plastocyanin (data not shown). Expression in two different systems Several customary in vivo approaches initially examined in order to generate large amounts of mature (wildtype or mutant) spinach plastocyanin failed in our hands. These included three different E. coli expression vectors, namely ptacl lE (Amann et al., 1983), pDS12/RBSII, NcoI (Stueber et al., 1984; Bujard et al., 1987) and pDM1 (Ahlquist and Janda, 1984) and appropriate bacterial strains (DH5ci, DH5aDMl, M1000, KS303 and KS476) as well as two eukaryotic protein expression vectors, the baculovirus-based system pAc6lO (Doerfler, 1986) and the pBM272 vector for protein expression in yeast (Johnston and Davis, 1984). None of them yielded appropriate amounts of the intact protein (Jansen, 1991; and data not shown). Pulse - chase experiments showed that the proteins were efficiently synthesized after induction but rapidly degraded so that their stationary concentrations were low. Only the pDS vector produced a reasonable amount of plastocyanin, but this protein turned out to be both unstable and inactive after isolation. The reason for this is unknown. It is conceivable that mature plastocyanin was recognized as a heterologous component and degraded. Subsequent attempts to operate I
i
1
TvrIS3-1ceti 2 -5
15 3) e)
A\ I 1 9) -l CLI -
Irt
1
2 5
15 3())-
2 4. 0 20. 1 1 7.0-
*1
-
4-
p
1 4.4-
8 . 26. 2-
m
Fig. 1. In organello import of wildtype and mutant plastocyanin precursor proteins. Fluorographs of 10-19% SDS-polyacrylamide gels are shown. Thirty microliters of in vitro translation products (lanes tr) were incubated with intact chloroplasts isolated from spinach in the light in a total volume of 240 A1. After the times (in minutes) indicated above the lanes, aliquots of 40 Al were removed, and the proteins of the reisolated chloroplasts separated on protein gels. A further 30 min aliquot of each assay was treated for 20 min with 0.1 mg/ml trypsin prior to analysis (lanes +). In lanes tr, 0.25 yd translation mix was loaded. All other lanes contained proteins corresponding to 8 itg chlorophyll. The positions of the precursor (p) and mature (m) proteins, as well as of the characteristic degradation products (*) are indicated. The band migrating close to the position of mature plastocyanin found in some of the translation assays (lanes tr) is caused by translational initiation at the methionine codon close to the terminal processing site. This product is not import-competent (see Clausmeyer et al., 1993). The sizes of the molecular weight standards (lane M) are given in kDa.
1029
W.Haehnel et al.
with the precursor protein in E. coli (Hibino et al., 1991) also failed. Although the precursor precipitated inside the cells in form of inclusion bodies from which it could be readily purified, subsequent processing by chloroplast extracts containing the highly active, specific processing proteases (Robinson and Ellis, 1984; Kirwin et al., 1987) yielded only low amounts of apo-plastocyanin. Finally, two strategies turned out to be promising: (i) secreting the protein into the bacterial periplasmic space using constructs that carry the mature polypeptide and the thylakoid-targeting part of the plastocyanin transit peptide for translocation (this transit peptide domain can efficiently replace prokaryotic leader peptides and is correctly removed by the bacterial leader peptidase during this process; see below), and (ii) heterologous expression in planta, using the respective precursor proteins in transgenic plants (see Last and Gray, 1990; Jansen, 1991). Potato rather than tobacco was chosen for the heterologous expression of spinach plastocyanin because of a greater difference of isoelectric points (I.P.) of its plastocyanin and of the mutated spinach proteins (see Discussion). A difference of 0.097 and 0.093 as found for wildtype spinach plastocyanin and the mutant plastocyanin A9OL, respectively, appeared to be sufficient for a quantitative separation of the transgenic from the host plastocyanin by anion exchange chromatography as shown in Figure 2A. But an unexpected shift in the I.P. by about -0.02 after exchange of the neutral amino acids in the mutant plastocyanin GlOL and Y83L made this chromatographic separation difficult as shown in Figure 2B. Preparative isoelectric focusing on a narrow pH gradient (Figure 2C) was a prerequisite for quantitative isolation of these two mutant plastocyanins.
Characterization of mutated proteins Figure 3 presents normalized spectra of mutated plastocyanins. All spectra were almost identical to those of wildtype plastocyanin except for minor changes of the band at 460 nm and the spectrum of plastocyanin Y83L. The absorbance ratio between 278 and 597 nm was 1.2 for plastocyanins A9OL and G1OL, consistent with pure spinach plastocyanin and an extinction coefficient of the oxidized protein at 278 nm of 5.9 mM-Icm-I (Durell et al., 1990). In Figure 3 the spectra of these two plastocyanins are normalized at 278 nm. The spectrum for plastocyanin Y83L is normalized to an extinction coefficient of 4.4 mM- Icm- I at 278 nm taking into account a decrease of the extinction coefficient by 1.5 mM -cm-1, which is the averaged value for a tyrosine residue at 278 nm (Donovan, 1969; Durell et al., 1990). Within the accuracy of the experiments the absorbance at the band near 597 nm shows the same maximal value for all plastocyanins studied. A preliminary chemical determination of the copper content of plastocyanin Y83L was also consistent with the extinction coefficient of 4.9 mM-'cm-1 at this band (Katoh et al., 1962). This extinction coefficient provides the basis needed for a quantitative comparison of rate constants and dissociation constants in the functional measurements. It should be pointed out that the spectrum of oxidized plastocyanin Y83L shows a remarkable shift by 4 nm of the absorbance band to 593 nm and a low ratio of 0.95 of the absorbance at 278 nm/593 nm. The two mutations at the flat hydrophobic surface at the 'northern' end of the plastocyanin, plastocyanin A9OL and plastocyanin GIOL, as well as the mutation Y83L near the negative patch do not change the net charge at acidic pH. -
B
A
I
4 min
A9OL
Y83L
C
ApH __ 0--
I0.2
Fig. 2. Separation of mutated spinach plastocyanin from native potato plastocyanin. (A) Spinach plastocyanin A90L was separated from potato plastocyanin (first and second peak, respectively) on a MonoQ HR5/5 column with a gradient from 0.15 to 0.26 M NaCl within 27 min at 0.5 ml/min. The total amount of plastocyanin was 38 nmol. The bar represents an absorbance of 0.05; the trace was monitored at 278 nm. (B) Spinach plastocyanin GIOL could not be separated from potato plastocyanin under the conditions used in panel A (total amount of plastocyanin was 130 nmol). (C) Isoelectric focusing in a preparative Immobiline gel using a gradient from pH 3.5 to 4.5. The bands show the (unstained) blue color of oxidized plastocyanin. Left: spinach plastocyanin GIOL in the upper band, potato plastocyanin in the lower band. The total amount of plastocyanin was 447 nmol. Right: spinach plastocyanin Y83L expressed in Ecoli (100 nmol). The resolution was sufficient to separate plastocyanin from a band of contaminating E.coli protein (only seen after Coomassie stain).
1030
300
400 500 WAVELENGTH / nm
600
Fig. 3. Absorbance spectra of oxidized plastocyanin A9OL, thick line; plastocyanin GIOL, dashed line; plastocyanin Y83L, thin line; expressed in E.coli. The spectra are normalized at 278 nm with the assumption that the extinction coefficient of plastocyanin Y83L decreased from 5.9 to 4.4 mM -cm-I to account for the reduced number of Tyr residues. The maximum at 597 nm is shifted by 4 nm to shorter wavelength in the spectrum of plastocyanin Y83L.
Electron transfer from plastocyanin to PSI Table I. Differences between the isoelectric points of mutated and wildtype spinach plastocyanin
Expressed in
Mutation
However, minor changes of the I.P. by these mutations could be detected by isoelectric focusing in an Immobiline gradient of 1 pH unit. The results are summarized in Table I. Within the limits of resolution (of the relative position), which is better than 0.004 pH units, identical isoelectric points of 3.93 were found for native plastocyanin expressed in E. coli and wildtype spinach plastocyanin. The I.P. of the plastocyanin A9OL expressed in potato and E. coli, respectively, agrees excellently. However, the I.P. value of plastocyanin Y83L expressed in potato was higher and that expressed in E. coli lower than that of the wildtype protein. This difference was even greater for plastocyanin G1OL from the two sources. This finding could originate from posttranslational modifications, from different processing of the -
Wildtype A9OL Y83L GIOL
Potato
E. coli
-
-0.004 -0.005 +0.018 +0.028
-0.004 -0.025 -0.027
Values measured following separation by acrylamide with Immobiline, pH 3.5 -4.5. In this system the wildtype spinach plastocyanin has a pl of 3.932. and the wildtype potato plastocyanin has a pI of 3.835 (-0.097 relative to the former). -C-
--
AM = - 77
M+ = 10333
WT Potato
A
1\ -5l
4-
I AM=-47
-4
Tyr83-Leu
M+ = 10363
Tyr83-Leu
A
B
WT Spinach X'\ M+ = 10410
WT Spinach
'\ M+ = 10410
B
C
Ala9O-Leu
Ala9O-Leu
C
D I AM
"Glyl 0-Leu"
= -
AM = 60
54
GlylO-Leu
M+ = 10356
EI
M+
=
10470
D
E
10000
= 10363
11000
10000
11000
M/Z M/Z Left: expressed in transgenic potato plants. Fig. 4. Matrix-assisted laser desorption/ionization mass spectra of recombinant spinach plastocyanins. (A) Wild type potato plastocyanin; (B) spinach plastocyanin Y83L; (C) Wild type spinach plastocyanin (control); (D) spinach plastocyanin A9OL; (E) spinach plastocyanin GlOL. Right: expressed in E.coli. (A) Spinach plastocyaninY83L; (B) wildtype spinach plastocyanin; (C) spinach plastocyanin A9OL; (D) spinach plastocyanin GIOL. The shoulder or minor peak at the falling edge of the peaks results from an association of dihydroxybenzoic acid minus H20 with the protein (AM = 136 Da). 1031
W.Haehnel et al.
precursor proteins or from folding differences of the polypeptide chains in the two expression systems (see below). All mutant plastocyanins were further characterized by mass spectrometry. 'Time of flight' mass spectra after matrix-assisted laser desorption/ionization are shown in Figure 4 for the mutant plastocyanins expressed in potato (left) and in E. coli (right). The mass at the position of the peak gives the value for MH+. Besides the molecule ion a shoulder originating from MNa+ and at a mass which is formed by MH+ plus 136 by association of a matrix molecule of 2,5-dihydroxybenzoic acid (minus H20) is detected. The difference, AM, between the mass of the mutant plastocyanin and the wildtype protein is clearly resolved as visualized by the peak shift relative to the control marked by the dashed line. The deviation of the measured mass value from that expected for the amino acid sequence is given in Table II. The relative difference falls into 0.02% for all proteins except for plastocyanin GIOL expressed in potato which produced a mass 113 Da lower than that expected for this mutation. The reason for this difference is not known; it would be consistent with a 114 Da difference if the C-terminal Asn residue were missing. Thus, the difference in I.P. between the two GIOL plastocyanins results most likely from differences in primary structure. Note also that the Y83L plastocyanins from the two organisms possess identical masses, despite the difference in their isoelectric points (Table I) suggesting that the primary sequence of these mutant proteins appears to be correctly processed in both expression systems. Functional measurements Laser-induced absorbance changes of P700 in PSI particles (PSI-200) as a function of time in the presence of different spinach plastocyanins are shown in Figure 5. After oxidation by laser excitation the reduction of P700+ at the high concentration of 40 MtM wildtype plastocyanin displays two kinetic components (Figure 5A). The fast component with a halftime of 12-15 Mts is typical for the electron transfer within the complex between plastocyanin and PSI (Haehnel et al., 1980). The slow component follows pseudo first-order kinetics because the concentration of plastocyanin is about two orders of magnitude higher than that of P700 (0.29 MM). Its halftime is dependent on the plastocyanin concentration consistent with the second-order reaction between reduced plastocyanin and P700+. The reduction of P700+ by plastocyanin Y83L shows the same kinetic components as
Table II. Masses of mutated spinach plastocyanins Mutation
Expressed in Potato
Native
A90L Y83L GIOL
10 10 10 10
415 457 363 356
E. coli
(+2) (+2) ( 0O) (-113)
-
-
hv
kon Pcred + P700
=
PCred P700
-
PCred * P700+
ket -
PcOX * P700
koff LPcredJLP700J
koff
KD =
-
- =
kon
(1)
[Pcred P700] *
The constant ket of the electron transfer after excitation (hp) is in a first approximation given by the halftime of 12 Ms of the fast component (ket = ln2/t,/2) with a value of 5.8 x 104 s- 1. The contribution of this fast reaction to the overall second-order rate is small relative to the rate of complex formation. Therefore, assuming that kon is independent of the redox state of P700, the values of k2 resemble those of kon which can be estimated as described by Nordling et al. (1991).
-
10 456 (+1) 10 363 ( O) 10 470 (+1)
The mass of wildtype plastocyanin from spinach was 10 410 (-3), from potato 10 333 (+2) and from transgenic potato 10 331 ( I0). In brackets: difference from mass MH+ expected from amino acid sequence.
1032
the wildtype with a minute increase in the amplitude of the 12 Mts component (Figure SC). In contrast, both plastocyanin A9OL and plastocyanin GIOL exhibit only the slow component but with a halftime increased by a factor of approximately three and five in Figure 5B and D, respectively, relative to the control. Even at high concentrations ( < 60 MtM) of the two mutant plastocyanins no indication of a fast kinetic component was noted. This suggests strongly that the electron transfer occurs without prior forming a complex with a significant lifetime. Figure 6 shows the effect of increasing plastocyanin concentrations ([Pc]) on the slow kinetic component of P700+ reduction. The pseudo first-order rate constant, kl, is proportional to [Pc] up to 70 MtM protein. This is shown in Figure 6A and B for the mutant proteins expressed in E.coli and in potato plants, respectively. The slope of the fitted straight line (regression analysis) Ak1/A[Pc] provides directly the second-order rate constant, k2, for the reaction PCred + P700 + - PCox + P700 The second-order rate constants are summarized in Table III. It should be noted that the rate constants for the two Y83L plastocyanins are higher than that of the wildtype protein. The fraction of PSI forming a complex with plastocyanin is estimated from the amplitude of the fast kinetic component relative to that of the two plastocyanin dependent components which represents 90% of the total amplitude of P700 (see Figure 5). Figure 7 shows this relative amplitude as a function of the plastocyanin concentration. The PSI reaction centers that were accessible to plastocyanin were shown to form a complex by increasing the plastocyanin concentration up to 800 MAM (M.Hippler, F.Drepper and W.Haehnel, unpublished data). Extrapolation of the data in Figure 7 indicates a relative amplitude of 50% at 60-70 MLM wildtype plastocyanin, and at 25 M,M and 175 MM of mutant plastocyanin Y83L expressed in E. coli and potato, respectively. As a first approximation, these values represent the dissociation constant, KD, of the complex between reduced plastocyanin and PSI with reduced P700, Pcred * P700,
Discussion The outlined study served two principal purposes: (i) to evaluate structure-function relationships in the electron transfer from plastocyanin to PSI by analyzing the effect of defined amino acid replacements in a reconstituted system,
Electron transfer from plastocyanin to PSI
0
,,l l -W -
_*
^~~VV A9OL
CO)
0
-1
v-
Y83L~~~~~~~~~~~~~~~~~~~~~~~~~~~~
z
m -2
0
cn 0
B
0
m
0LL
(5
-1
z I
GlOL
0
-2
D 0
0.4
2
4
4,
0
0.4
2
6
4
8
TIME / ms Fig. 5. Absorbance changes at 703 nm in PSI particles as a function of time induced by a laser flash in the presence of spinach plastocyanin. (A) Wildtype plastocyanin; (B), plastocyanin A9OL; (C), plastocyanin Y83L; (D), plastocyanin GIOL. The concentration of plastocyanin was 40 jsM for A, B and D and -50 AM for C (see text). The time base was switched 0.410 ms after the flash from 0.5 to 2 (A and C), to 5 (B), or to 10 its/address (D). Fifty signals were averaged with a repetition rate of 0.8 Hz.
A
B 3000
3000 .C')
U,c
F-
z
z
H 2000
(I) z
Cl) z
0
2000 H
0 LuI
EHr
H 1000 _
v-_
1000 F
Ir
/t' ,4
A90Lp -
0-t. 0
20
40
60
[PLASTOCYANIN] / 106M
0
0
-
*I
20 40 [PLASTOCYANIN] / 106M
60
Fig. 6. (A) Pseudo first-order rate constant of the P700+ reduction as a function of the concentration of mutant spinach plastocyanin expressed in E.coli. 0, wildtype plastocyanin from spinach (control) PCs; 0, plastocyanin A9OL; V, plastocyanin Y83L; V, plastocyanin GIOL. (B) Pseudo first-order rate constant of the P700+ reduction as a function of the concentration of mutant spinach plastocyanin from transgenic potato plants. 0, wildtype plastocyanin from spinach (control) PCs; 0, plastocyanin A9OL; V, plastocyanin Y83L. Dotted lines give the linear regression curve. and
(ii)
to test
and
optimize
various
overexpression systems
for the large-scale synthesis of this protein. Expression of
mutant
plastocyanin
The initial serious drawbacks experienced with customary overexpression systems could be ultimately bypassed by two complementing strategies: 'export cloning', that is, secreting
the protein into the bacterial periplasm and by heterologous expression in planta. 'Export cloning' was based on constructs encoding the mature polypeptide that operate with the C-terminal (thylakoid-targeting) part of a bipartite chloroplast transit peptide in place of a leader peptide. This is in line with the observation of Seidler and Michel (1990) who successfully expressed the lumenal spinach 33 kDa
1033
W.Haehnel et al.
be noted that the mutant proteins could not be obtained with a yield equal to that of the host (see Figure 2A). Mutants that impair the interaction between the internal aromatic residues have not yet yielded sufficient protein for analysis in any of the expression systems. This could indicate an inherent instability of these plastocyanins or a counter selection of the mutant protein if its malfunction interferes with or even blocks electron transfer in the host cell chloroplast. The relative instabilities appear to be the same in all expression systems checked, indicative of a general
Table IH. Electron transfer from mutated plastocyaninIs to P700 + [second-order rate constant (107 M-' s-1)] Mutation
Expressed in
A9OL Y83L GIOL
Potato
E. c'oli
1.0 (20%) 5.6 (114%)
1.9 (39%) 7.1 (145%) 0.718 (16%)
-
Rates were taken from regression analysis of the data shown Figure 6. The rates relative to wildtype spinach plastoc-yanin 4.9 x 107 M-l s-I (= 100%)] are given in brackets.
0.7
I
I
Y83LE
w 0.6 :_ 0.5 -J
Pcsll
20.4 H 6 w 0.3 F < 0.2
w
in [k2 =
_-'vY83L
-S
H
*
_
_-
0.1 0.0
0
10
20 30 40 50 [PLASTOCYANIN] / 10-6 1
60
70
Fig. 7. Amplitude of the fast (12-14 As) kinetic comp reduction as a function of the concentration of wildtypce plastocyanin from spinach (0, 0), and plastocyanin Y83L expresseed in potato (V) and in Ecoli (V), respectively. Different batches of P'SI particles
were used for the measurements shown with open symbols than for those with closed symbols. The curves labelled PCSI and show the good reproducibility of the control measurements MVith spinach
PCSII
plastocyanin.
protein of the oxygen-evolving complex as a 'truncated' precursor protein in E. coli. Under our conditions the thylakoid-targeting part of the plastocyanin 4and of the 33 kDa polypeptide transit peptide proved to be ssuperior to the corresponding domains of the 23 and 16 kE)a proteins of the oxygen-evolving complex and also to the azurin leader sequence (data not shown; see Jansen, 1991)i. Wildtype as well as mutant plastocyanin GlOL, A9OL anId Y83L were exported and accumulated efficiently in the periiplasmic space of the E. coli cell. High resolution gel electrl ophoresis and mass spectrometry verified that the residua 1 part of the plastocyanin transit peptide is correctly and quantitatively removed by the leader peptidase. Remarkably. when copper was present it was efficiently incorporated. The heterologous expression in planta used the respective full-length precursor protein in transgenic potato. This host was favored over tobacco because its authenti(c plastocyanin can readily be separated by anion exchange chromatography or electrofocusing from the spinach protein. Its molecular mass (10 331 Da, Figure 4) is also smallerr than that of spinach plastocyanin (10 412 Da). This exclu(des the risk of contamination of the spinach protein with pot;ato plastocyanin. The expression system leads to the corr^ect import of the wildtype as well as the mutant spinach prcxteins into the thylakoid lumen of the potato chloroplast. How(ever, it should 1034
intrinsic recognition and turnover mechanism (Jansen, 1991). Translation and post-translational processes may modify protein structure, particularly in heterologous systems. Recent developments in mass spectrometry (Hillenkamp and Karas, 1990) have provided fast and precise tools for analyzing proteins. For all mutant spinach plastocyanins the correct primary sequence could be verified except for plastocyanin GlOL expressed in potato plants. Without this control showing an unidentified mutation of the protein the functional measurements (not shown) would have suggested misleading conclusions about the structure-function relationship. Thus, mass spectrometry can be an aid to increasing the validity of analysis and is of great advantage in particular with mutated proteins that yield incompletely understood results. Protein folding is generally determined by a global energy minimum and the physico-chemical properties of the plastocyanins made in the two expression systems should coincide. This is true for most of the mutants. However, differences in I.P. and in particular in the dissociation constant of the two mutant spinach plastocyanins Y83L (Table I and Figure 7, respectively) being either higher or lower than the value of the wildype are evidence for different
secondary structures. It is important to note that the masses and hence protein sequences were identical and that copper was incorporated in an active form in both instances. The observation that the folding process in E.coli results in a conformation different from that in potato plants suggests caution should be exercised in the selection of heterologous overexpression systems in order to avoid the risk of misinterpretations. It is not known whether the final structure of the protein depends on the interaction with chaperones. Interaction between plastocyanin and photosystem I Electron transfer from plastocyanin to PSI involves two essential steps, (i) binding of plastocyanin at the PSI complex and (ii) electron transfer from bound plastocyanin to P700+. The laser-induced reduction kinetics of P700+ provide detailed informations on these steps, the dissociation constant (see Figure 7), the fast electron transfer (12 its) within the complex (Figure 5), and the rate constant of the binding which is close to the second-order rate constant (Figure 6 and Table III). These reactions depend on the quality of the isolated PSI (Haehnel, 1986). In this study we used a preparation with a fraction of 80-90% of total PSI capable of forming a complex with plastocyanin which is close to that in intact chloroplasts (Haehnel et al., 1989). In all measurements we have found a dissociation constant of the complex between spinach plastocyanin and PSI of - 60 ,tM, but not values as high as 170 /%M (Nordling et al., 1991). A possible loss of subunit III of PSI (psaF gene product) would drastically change the electrostatic interaction and decrease the electron transfer rate between the reactants by three orders of magnitude (Ratajczak et al., 1988).
Electron transfer from plastocyanin to PSI
AiaQ,O.-Lelu A
-
Tyr83
.St ()1 '1 y1N!Iu u
i-)
Li
HIs8 11
Ala9O
W'., GlylO-L eu
Tyr83
Fig. 8. Left: Van der Waals surface of poplar plastocyanin viewed from above the 'northern' end (top) and from the side (bottom) after clockwise rotation by 300 around the z-axis from the position given by the coordinates in the Brookhaven Database (Guss and Freeman, 1983). The surface of the copper ligand H87 and of the residues GlO, Y83 and A90 which are conserved in all plant plastocyanins are highlighted and have been mutated to Leu in this work. The distance between the ca-C atom of GIO and the NH (imidazole) of H87 at the surface is 0.98 nm. Right: top, 'northern' half of mutant plastocyanin A9OL; bottom, 'northern' half of mutant plastocyanin GIOL. The program MOBY (Springer Verlag) was used to change A90 and GIO, respectively, to Leu and for energy minimization of the structure by the amber force field (Weiner et al., 1984). The non-bonding orbitals of the sulfur atom in Cys84 and Met92 were introduced as lone pairs. To compensate for the lack of force field parameters of Cu2+ the position of the copper atom and its four atomic ligands was kept invariable. Electrostatic interaction was included for atoms spaced up to 1.0 nm. The dotted line is drawn to illustrate the position of the flat surface around the copper ligand H87 in the wildtype structure.
One striking result of this study is that the electron transfer from plastocyanin Y83L is faster than that of wildtype plastocyanin. Participation of this residue in electron transfer to P700+ can therefore be excluded. This is in agreement with the previous suggestion (Nordling et al., 1991) derived from a 2-fold decrease of the electron transfer rate when Y83 is replaced by His with its aromatic heterocycle of imidazole. However, the spectrum of the mutant plastocyanin used in their study indicates a substantial contamination with other or inactive proteins. Our data exclude the possibility that the aromatic system of Y83 is involved in the electron transfer and that its OH-group contributes to the binding of plastocyanin to PSI, e.g. by an H-bridge, as proposed for the interaction with cytochrome f (He et al., 1991). This, in turn, implies that the route of electron transfer from cytochrome f to plastocyanin is different from that from plastocyanin to P700. Our finding that the interaction with PSI is hindered by Tyr relative to Leu in line with the conserved nature of Y83 in almost all plastocyanins suggests an important function in a reaction different from that with PSI. This is consistent with the conclusion that Y83 may be indispensable for electron transfer from cytochrome f. The shift in the dominant charge transfer band at 597 nm
by -4 nm to shorter wavelength (Figure 3) may be the first experimental evidence for the predicted long-range interaction between the copper center and Y83 via C84 and the peptide bond (Lowery et al., 1993) in support for this electron transfer route. The shift of the peak can be explained by a weaker S(Cys)p7r-Cu2+ transition at 606 tm in parallel with a stronger S(Cys)pa-Cu2+ transition at 560 nm of the metal-thiolate system relative to the wildtype (Penfield et al., 1981). The bulky hydrophobic residue of Leu was introduced in place of either G10 or A90 within the conserved flat hydrophobic surface in order to change the shape of this surface while minimizing changes of electrostatic interactions. Figure 8 illustrates the relative position of the mutated residues close to the imidazole of H87, the outer sphere ligand of the copper center (left), and how far the hydrocarbon chain of Leu protrudes from the flat surface at the two sites (right). The functional measurements with both mutants indicate that the formation of the complex is completely abolished at plastocyanin concentrations as high as 60-70 AM at which half of the PSI complexes would carry a bound wildtype plastocyanin. Thus, the data show unambiguously that the hydrophobic surface at the 'northern' 1035
W.Haehnel et al.
end with a diameter of 1.5-2.0 nm has to be flat to allow the formation of a complex between plastocyanin and PSI. This interaction suggests a complementary flat hydrophobic binding site at the PSI complex with a site of electron transfer at a central position similar to that of H87 (see Figure 8 left, top view). The second-order reduction rate of P700+ is smaller by a factor of three to five with the two mutant proteins than the control, providing additional evidence for the electron transfer via H87 to PSI. This electron transfer without preceding complex formation between reduced plastocyanin and PSI with reduced P700 may be facilitated by the positive charge of P700+ formed after laser excitation and the flexibility of the structures enabling a suitable contact. In terms of scheme (1) the effect of the Leu residue could be understood as an increase of the rate constant koff by the weakened narrow range hydrophobic interaction at an almost constant value of kon and unchanged electrostatic forces. The deduced route is entirely different from that proposed for the electron transfer from cytochrome f to plastocyanin (He et al., 1991). With respect to the structure of the docking pocket for plastocyanin at the PSI complex it is also the interaction of at least one negative carboxyl group of plastocyanin and a positive lysine residue of subunit IH which is essential for the formation of the complex as shown by cross-linking experiments (Hippler et al., 1989). This combination of electrostatic and hydrophobic interaction elucidates details of the reaction mechanism. The electrostatic interaction alone is not strong enough for the formation of a complex as seen with plastocyanin A9OL and GIOL in Figure 5. It needs in addition the tight fit of the hydrophobic surface to provide the delicately balanced strength for the complex between reduced plastocyanin and PSI. In such a balanced state the transition from reduced to oxidized plastocyanin would diminish the attracting negative charges at plastocyanin and favor the release of the oxidized molecule after electron transfer to P700+. This mechanism ensures a rapid turnover at the reaction center. An increase found in the redox potential of plastocyanin bound to PSI supports this model (F.Drepper, M.Hippler and W.Haehnel, unpublished experiments). Thus, a plastocyanin molecule approaching PSI is probably oriented in two steps (i) by long-range electrostatic interaction between its negative patches and the positive residues of subunit Ill (PsaF protein), and (ii) by a tight docking with its flat hydrophobic surface at a site probably constituted by the two large subunits Ia and lb (PsaA and PsaB protein, respectively) close to P700, the special pair of chlorophyll a molecules (Krauss et al., 1993). These steps are followed (iii) by a fast electron transfer with a tl, of 12 /ts close to the center of this hydrophobic area of contact, and (iv) a facilitated release of the oxidized plastocyanin.
Materials and methods Mutagenesis of spinach plastocyanin Site-directed mutagenesis of spinach plastocyanin was based on the gapduplex method (Kramer et al., 1984; Fritz et al., 1988) and a genomic clone that carries the entire coding region (Rother et al., 1986). The following alterations were introduced: (i) GIOL (codon GGA to CTG), (ii) Y83L (TAC to CTG) and (iii) A9OL (GCT to CTG). The correctness of these mutations was verified by nucleotide sequence analysis (Sanger et al., 1977) using Sequenase 2.0 (United States Biochemicals, Cleveland, OH).
1036
Expression of spinach plastocyanin in transgenic potato plants A 1646 bp SmaI-EcoRI fragment of the spinach plastocyanin gene (Rother et al., 1986), consisting of 834 bp of the plastocyanin promoter, 67 bp of 5' untranslated nucleotides, the region encoding the precursor protein (either wildtype or mutated) and 241 bp of 3' untranslated sequences including the polyadenylation signal, was inserted into the binary vector pBIN19 (Bevan, 1984) and transferred into Agrobacterium tumefaciens strain LBA4404 (Hoekema et al., 1983) using the helper plasmid pRK2013 (Figurski and Helinski, 1979). Potato leaf discs (cv. Desiree) were transformed according to Horsch et al. (1985). Transformed material was selected on MS medium (Murashige and Skoog, 1962) containing 2% sucrose, 2 mg/l zeatin riboside, 0.02 mg/l naphthyl acetic acid, 0.02 mg/l gibberellic acid, 50 mg/l kanamycin and 250 mg/l carbenicillin. Shoots of -2 cm were cut off and rooted on hormone-free MS medium containing 2% sucrose, 50 mg/l kanamycin and 250 mg/l carbenicillin. Rooted plantlets were transferred to soil and grown in a safety greenhouse according to the German regulations for handling transgenic plants. Expression of spinach plastocyanin in E.coli Plastocyanin is synthesized as a precursor molecule in the cytosol with a bipartite N-terminal transit peptide (Smeekens et al., 1986), a hydrophiic import domain for transfer across the chloroplast envelope and a hydrophobic region flanked by polar residues which resembles bacterial signal sequences and mediates translocation across the thylakoid membrane (von Heijne, 1985). This transit peptide is removed in two steps (Smeekens et al., 1986), the former by a stroma-located metalloprotease (Robinson and Ellis, 1984), the latter by a thylakoid-located protease during transport of the plastocyanin intermediate into the thylakoid lumen (Kirwin et al., 1987). For 'export expression' into the periplasmic space of E. coli, the DNA segment encoding the plastocyanin precursor was mutated using the polymerase chain reaction in such a way that an NdeI restriction site was introduced at nucleotide position + 137 from the translation start site. This restriction site, leading to an amino acid replacement of Leu by Met at residue 47 in the precursor molecule, was utilized to insert the DNA fragment for the thylakoid-targeting part of the transit peptide (residues 48-69) plus the entire mature polypeptide into the expression vector pAR3040 (Rosenberg et al., 1987). This plasmid carries the promoter, the 5' untranslated region and the start codon (with an NdeI site) of the T7 gene 10. The exchange of equivalent 417 bp NcoI-BstBI fragments between this clone bearing the authentic spinach plastocyanin and those carrying the segments for the mutant proteins outlined above yielded corresponding expression vectors for the mutant plastocyanins. The resulting plasmids were introduced into E.coli strain BL21(DE23) which carries the gene for T7 RNA polymerase under the control of the isopropyl-3-D-thiogalactopyranoside (IPTG) inducible lacUV5 promoter. Positive clones were selected as described (Studier and Moffatt, 1986). The respective recombinant bacteria were grown in LB medium supplemented with 4 g/l glucose and 40 mg/l ampicillin to an OD6W of 1.0. IPTG and CuS04 were then added to final concentrations of 1 mM and 10 AM, respectively, to induce the synthesis of wildtype or mutant plastocyanin. After incubation at 37°C for 4 h, the cells were harvested by sedimentation at 8000 g for 10 min. They were resuspended in 0.1 vol 20% sucrose, 20 mM Tris-HCl, 2 mM EDTA pH 8.0, and kept at 250C for 10 min to release periplasmic proteins into the supematant. The supematant was finally clarified by centrifugation for 10 min at 8000 g. Protein purification from E.coli The periplasmic fraction from a 2.5 1 culture was loaded on a DE23 anion exchange cellulose column (2 cm x 20 cm) equilibrated with 20 mM Tris-HCI, pH 7.6. This buffer was used in all isolation steps. After washing with buffer and buffered 50 mM NaCl, plastocyanin was eluted with a continuous salt gradient between 50 and 400 mM NaCl. Plastocyanincontaining fractions were detected by addition of K3[Fe(CN)6] and concentrated by ultrafiltration (YM05 membrane, Amicon). The concentrate was desalted (PD-10, Pharmacia) and loaded onto a MonoQ column (HR5/5, Pharmacia) equilibrated with 20 mM Tris-HCl, pH 7.6. After elution with a gradient from 0 to 300 mM buffered NaCl, the fractions containing plastocyanin were again concentrated by ultrafiltration (Centricon 3000, Amicon). The copper-containing (blue) protein was purified by preparative isoelectric focusing on a polyacrylamide gel of 5 mm thickness with a pH gradient from 3.5 to 4.5 formed by Immobiline mixtures (Pharmacia). The blue band was excised with a wide razor blade and the protein collected after electroelution in 10 mM Tris-borate buffer, pH 7.6, for 3 h at 200 V with a Bio-Trap (Schleicher & Schuell). After concentration by ultrafiltration the solution was desalted into distilled water and stored under liquid nitrogen until use. Plastocyanin concentrations were determined by using the extinction
Electron transfer from plastocyanin to PSI coefficient of 4.9 mM Icm-I at 597 nm for the oxidized protein (Katoh etal., 1962).
Protein purification from transgenic potato plants The procedure for isolating plastocyanin from potato leaves was optimized for yield. Grinding the leaves in the presence of cold acetone (0°C) as the initial step was superior to other homogenization techniques. Fifty grams of deveined leaves were ground in a precooled mortar at 4°C with 50 ml 0.1 M Tris-HCI buffer, pH 7.6, 50 ml cold acetone and a pinch of sand. After centrifugation at 10 000 g for 5 min, the supernatant was mixed with 1.16 vols of cold acetone and kept on ice for 5 min. The pellet formed after centrifugation under the same conditions was resuspended in 2 ml of 60 mM Tris-HCI, pH 7.6, and dialyzed overnight against the same buffer in dialysis tubing with a molecular mass cut-off of 5000. Insoluble material was removed by centrifugation at 10 000 g for 5 min. The supernatant was loaded on an anion exchange cellulose DE23 (Whatman) (2 x 20 cm) equilibrated with 20 mM Tris-HCl buffer, pH 7.6. After washing with buffer, proteins were eluted with a salt gradient from 0 to 0.3 M buffered NaCl. The plastocyanin-containing fractions, detected by addition of K3[Fe(CN)6], were concentrated by ultrafiltration (YM05 membrane, Amicon) and purified by gel filtration (Sephacryl S200HR, Pharmacia). Plastocyanin was concentrated, reduced by addition of sodium ascorbate, desalted and loaded onto a MonoQ column equilibrated with buffer. The potato and spinach plastocyanins were then eluted with a gradient from 150 to 260 mM buffered NaCl (see Figure 2A). This method failed to separate the transgenic mutant GIOL and Y83L spinach plastocyanins from the native potato plastocyanin (see Figure 2B). These mutant proteins were purified by preparative isoelectric focusing and electroelution as described above (see Figure 2C). PSI particles (PSI-200) were prepared from spinach leaves as described (Wynn and Malkin, 1988). Spinach plastocyanin was isolated as described in Christensen et al. (1991), wildtype potato plastocyanin essentially as in Plesnicar and Bendall (1970). Mass spectrometry Matrix-assisted laser desorption/ionization mass spectrometry followed published procedures (Hillenkamp and Karas, 1990; Hillenkamp et al., 1991; Chait and Kent, 1992). The protein (0.3 mM) was diluted with distilled water by a factor of 10. One microliter of this sample was mixed with 10-20 A1l 2,5-dihydroxybenzoic acid (10 g/l). Five different samples with up to 1 1l of this mixture each containing 0.75-1.5 pmol protein were applied to the same target. The samples were dried in a gentle stream of air. Desorption and ionization were induced by 3 ns pulses of 337 nm from a N2 laser (Laser Science Inc., Cambridge, MA) with an irradiance of 106- 107 W/cm2. The mass scale was calibrated with a mixture of bovine insulin and horse heart cytochrome c with masses for MH+ of 5734.5 and 12 360.2, respectively. Kinetic measurements Reaction mixtures contained PSI particles at a concentration of 100 jig chlorophyll/ml, 30 mM HEPES-NaOH buffer (pH 7.2), 5 mM MgCl2, 1 mM sodium ascorbate, 0.2 mM methylviologen, 0.1 mM diaminodurene, and plastocyanin as indicated. The absorbance changes of P700 were measured after excitation with a flash of saturating intensity from a Nd:YAG laser at 532 nm (5 ns full width at half maximum, FWHM). The measuring light of 703 nm (2.6 nm FWHM) passed through the cuvette containing a 0.3 ml sample with a light path of 1.2 mm and was detected by a silicon photodiode (1 cm2) protected by an interference filter of 702.9 nm. The output from the photodiode was amplified with an electrical bandwidth ranging from d.c. to 1 MHz. The d.c. offset was realized within 2 1is immediately before the trigger of the transient recorder by a combination of a 16 bit ADC and a 16 bit DAC. The amplified signal was digitized in a transient recorder (DL 912, Datalab) by using a dual time base switching from 0.5 Its/address to 2-20 As/address to capture the complete time course. The signals were directly transferred to a mini-computer (A-700, Hewlett Packard) and averaged. Triggering was under program control by the computer. Signal disturbance by fluorescence was monitored without measuring light and subtracted. Fifty signals were averaged with the repetition rate of 0.8 Hz. Analysis of the time course was by a program using a least squares fit of a two-exponential decay.
Acknowledgements Help with the spectroscopic measurements by Friedel Drepper, with the energy minimization of the modified structures by Udo Hoiweler and isolation of mutant plastocyanin from E. coli by Jan Reichert are gratefully acknowledged. This work was supported by Deutsche Forschungsgemeinschaft (SFB 143 and Ha 1083/5-3).
References Ahlquist,P. and Janda,M. (1984) Mol. Cell. Biol., 4, 2876-2882. Amann,E., Brosius,J. and Ptashne,M. (1983) Gene, 25, 167-178. Anderson,G.P., Sanderson,D.G., Lee,C.H., Durell,S., Anderson,L.B. and Gross,E.L. (1987) Biochim. Biophys. Acta, 894, 386-398. Augustin,M.A., Chapman,S.K., Davies,D.M., Sykes,A.G., Speck,S.H. and Margoliash,E. (1983) J. Biol. Chem., 258, 6405-6409. Beoku-Betts,D., Chapman,S.K., Knox,C.V. and Sykes,A.G. (1983) J. Chem. Soc. Chem. Commun., 20, 1150-1152. Bevan,M. (1984) Nucleic Acids Res., 12, 8711-8721. Bujard,H., Gentz,R., Lanzer,M., Stueber,D., Mueller,M., Ibrahimi,I., Haeuptle,M.-T. and Dobberstein,B. (1987) Methods Enzymol., 15, 416-433. Chait,B.T. and Kent,S.B.H. (1992) Science, 257, 1885-1894. Christensen,H.E.M., Conrad,L.S. and Ulstrup,J. (1991) Photosynth. Res., 28, 89-93. Clausmeyer,S., Klosgen,R.B. and Herrmann,R.G. (1993) J. Biol. Chem., 268, 13869-13876. Cookson,D.J., Hayes,M.T. and Wright,P.E. (1980) Nature, 283, 682-683. Doerfler,W. (1986) Curr. Top. Microbiol. Immunol., 131, 51-68. Donovan,J.W. (1969) In Leach,S.D. (ed.), Physical Principles and Techniques of Protein Chemistry, Part A. Academic Press, New York, pp. 101-170. Durell,S.R., Lee,C.-H., Ross,R.T. and Gross,E.L. (1990) Arch. Biochem. Biophys., 278, 148-160. Figurski,D.H. and Helinski,D.R. (1979) Proc. NatlAcad. Sci. USA, 76, 1648-1652. Fritz,H.-J., Hohlmaier,J., Kramer,W., Ohmayer,A. and Wippler,J. (1988) Nucleic Acids Res., 14, 6987-6999. Guss,J.M. and Freeman,H.C. (1983) J. Mol. Biol., 169, 521-563. Guss,J.M., Harrowell,P.R., Murata,M., Norris,V.A. and Freeman,H.C. (1986) J. Mol. Biol., 192, 361-387. Haehnel,W. (1986) In Staehelin,L.A. and Arntzen,C.J. (eds), Photosynthesis III (Encyclopedia of Plant Physiology, New Series, Vol. 19). Springer Verlag, Berlin, pp. 547-559. Haehnel,W., Pr6pper,A. and Krause,H. (1980) Biochim. Biophys. Acta, 593, 384-399. Haehnel,W., Ratajczak,R. and Robenek,H. (1989) J. Cell Biol., 108, 1397-1405. Haehnel,W., Gause,K., Jansen,T., Klosgen,R.B., Herrmann,R.G., Stahl,B., Karas,M., Witt,I. and Witt,H.T. (1992) In Murata,N. (ed.), Research in Photosynthesis. Kluwer Academic, Dordrecht, Vol. II, pp. 455 -462. He,S., Modi,S., Bendall,D.S. and Gray,J.C. (1991) EMBO J., 10, 4011-4016. Hibino,T., Douwe de Boer,A., Weisbeek,P.J. and Takabe,T. (1991) Biochim. Biophys. Acta, 1058, 107-112. Hillenkamp,F. and Karas,M. (1990) Methods Enzymol., 193, 280-295. Hillenkamp,F., Karas,M., Beavis,R.C. and Chait,B.T. (1991) Anal. Chem., 63, 1193-1203. Hippler,M., Ratajczak,R. and Haehnel,W. (1989) FEBS Lett., 250, 280-284. Hippler,M., Ratajczak,R. and Haehnel,W. (1990) In Baltscheffsky,M. (ed.), Current Research in Photosynthesis. Kluwer Academic, Dordrecht, Vol. II, pp. 675-678. Hoekema,A., Hirsch,P.R., Hooykaas,P.J.J. and Schilperoort,R.A. (1983) Nature, 303, 179-180. Horsch,R.B., Fry,J.E., Hoffmann,N.L., Wallroth,M., Eichholtz,D., Rogers,S.G. and Fraley,R.T. (1985) Science, 227, 1229-1231. Jansen,T. (1991) Untersuchungen zu Struktur-/Funktionsbeziehungen am Photosyntheseapparat h6herer Pflanzen: Molekulare Analyse und gezielte Modifikation kemkodierter Gene fur Plastocyanin und FerredoxinNADP+-Oxidoreduktase. Thesis, Ludwig-Maximilians-Universitiat, Munchen. Johnston,M. and Davis,R.W. (1984) Mol. Cell. Biol., 4, 1440-1448. Katoh,S., Shiratori,I. and Takamiya,A. (1962) J. Biochem., 51, 32-40. Kirwin,P.M., Elderfield,P.D. and Robinson,C. (1987) J. Biol. Chem., 262, 16386-16390. Kramer,W., Drutsa,V., Jansen,H.W., Kramer,B., Pflugfelder,M. and Fritz,H.-J. (1984) Nucleic Acids Res., 12, 9441-9456. Krauss,N., Hinrichs,W., Witt,I., Fromme,P., Pritzkow,W., Dauter,Z., Betzel,C., Wilson,K.S., Witt,H.T. and Saenger,W. (1993) Nature, 361, 326-331. Last,D.I. and Gray,J.C. (1990) Plant Mol. Biol., 14, 229-238. Lowery,M.D., Guckert,J.A., Gebhard,M.S. and Solomon,E.I. (1993) J. Am. Chem. Soc., 115, 3012-3013. Modi,S., Nordling,M., Lundberg,L.G., Hansson,O. and Bendall,D.S. (1992) Biochim. Biophys. Acta, 1102, 85-90.
1037
W.Haehnel et al. Murashige,T. and Skoog,F. (1962) Physiol. Plant., 15, 473 -497. Nordling,M., Sigfridsson,K., Young,S., Lundberg,L.G. and Hansson,O. (1991) FEBS Lett., 291, 327-330. Penfield,K.W., Gay,R.R., Himmelwright,R.S., Eickman,N.C., Norris,V.A., Freeman,H.C. and Solomon,E.I. (1981) J. Am. Chem. Soc., 103, 4382-4388. Plesni6ar,M. and Bendall,D.S. (1970) Biochim. Biophys. Acta, 216, 192-199. Ratajczak,R., Mitchell,R. and Haehnel,W. (1988) Biochim. Biophys. Acta, 933, 306-318. Robinson,C. and Ellis,R.J. (1984) Eur. J. Biochem., 142, 337-342. Rosenberg,A.H., Lade,B.N., Chui,D.-S., Lin,S.-W., Dunn,J.J. and Studier,F.W. (1987) Gene, 56, 125-135. Rother,C., Jansen,T., Tyagi,A., Tittgen,J. and Herrmann,R.G. (1986) Curr. Genet., 11, 171-176. Sanger,F., Nicklen,S. and Coulson,A.R. (1977) Proc. NatlAcad. Sci. USA, 74, 5463-5467. Seidler,A. and Michel,H. (1990) EMBO J., 9, 1743-1748. Smeekens,S., Bauerle,C., Hageman,J., Keegstra,K. and Weisbeek,P. (1986) Cell, 46, 365-375. Studier,F.M. and Moffatt,B.A. (1986) J. Mol. Biol., 189, 113-130. Stueber,D., Ibrahimi,j., Coulter,D., Dobberstein,B. and Bujard,H. (1984) EMBO J., 3, 3143-3148. Sykes,A.G. (1985) Chem. Soc. Rev., 14, 283-315. Tiede,D.M., Vashishta,A.C. and Gunner,M.R. (1993) Biochemistry, 32, 4515 -4531. von Heijne,G. (1985) J. Mol. Biol., 184, 99-105. Weiner,S.J., Kollman,P.A., Case,D.A., Singh,U.C., Ghio,C., Alagona,G., Profeta,S. and Weiner,P. (1984) J. Am. Chem. Soc., 106, 765-784. Wynn,R.M. and Malkin,R. (1988) Biochemistry, 27, 5863-5869.
Received on October 19, 1993; revised on December 10, 1993
1038
