Oxygenic photosynthesis occurs in cyanobacteria and green plants, and requires photosystem (PS) II to catalyze the photo-oxidation of water. Most previous molecular engineering work on PS II has been done with the cyanobacterium Synechocystis sp. PCC 6803, because of the well developed molecular genetics technology. However, our recent development of a streamlined system for manipulation of the psbA gene encoding the D1 peptide of PS II of Chlamydomonas reinhardtii makes this an attractive alternative. In contrast to cyanobacteria, this unicellular alga has a chloroplast with all the main characteristics of higher plants, including the same PS II, light harvesting apparatus, and oxygen evolving complex. Cells deficient in photosynthesis can be readily grown on acetate. Biophysical methods based on fluorescence and spectrophotometry are easier to use than with cyanobacteria because PS II is expressed at higher level, and has a chlorophyll (Chl) light harvesting apparatus, allowing techniques developed using chloroplasts to be applied directly to intact cells. We are using molecular engineering and biophysical assay to explore the relation between structure and function in PS II in C. reinhardtii.
Biochemical and biophysical evidence has shown that PS II spans the chloroplast membrane, with a water oxidizing site on the lumenal side, a plastoquinone reductase site on the stromal side, and a set of chlorophyll and pheophytin chromophores catalyzing the photochemical reaction which separates charge between these sites. The atomic structure of PS II is not known. However the weak sequence homology of PS II and bacterial reaction centers, along with much biochemical evidence, has been taken to indicated common features in the basic structure and mechanism. In our first models for the core of PS II (1-3) we used the solved structure of a bacterial reaction center (4) as a template, using residue substitution, and manual construction based on structural prediction to accommodate differences (such as extra loops) apparent from homologous alignment. Other models (5, 6) have incorporated inter-helical loops constructed by use of template matching from known structures, energy minimization and simulated annealing. More recent models of photosystem II have included a putative Mn-cluster and bicarbonate group. Others labs have concentrated on modeling the binding pockets for primary and secondary plastoquinone acceptors (QA and QB respectively) (7). There are substantial differences in structure between models in regions of mechanistic interest, reflecting the lack of experimental constraints.
Click here for a Chime tutorial on a model of the Photosytem II core from Jonathan Nugent and colleagues.
Biochemical evidence suggests an overall structure and complement of chromophores similar to that of the bacterial reaction centers (see 8-10 for recent reviews). On the acceptor side, the basic structure, with binding sites for QA and QB spaced by an Fe ligated by 4 His, seems well established, as does the probability that E130 in D1 provides a ligand to pheophytin. The histidine ligands to the ancillary BChl in the bacterial centers are not conserved in D1 or D2, leaving the Chl ligands uncertain. The lumenal side is even more problematical, since it is here that the major differences from the bacterial reaction center are found. The nature of P680 (the primary donor Chl) is still controversial. The dimeric special pair expected from bacterial centers has not been unambiguously demonstrated, although H198 of D1, and the equivalent in D2, are conserved as putative ligands, and the near IR band of P680+ is diagnostic. Several suggestions have been made that P680 might be monomeric (8-10), and more equivalent to the ancillary Chls of the bacterial reaction center. Important roles for Tyr-161 (D1) as donor to P680+, and for Tyr-160 (D2) as an ancillary donor, have been established (11-15). A Mn-cluster catalyzes O2-evolution (see below).
(PS II chromophores).
Kinetic measurements and studies of herbicide resistance lesions have suggested that the two-electron gate (see 16 for review of earlier work, 23-28) has the same general mechanism and basic structure in PS II and photosynthetic bacteria. In green plants, the kinetics of the individual partial reactions are well resolved by the fluorescence methods introduced by this lab (29-33). We have studied the mechanism in higher plants and in C. reinhardtii, and have characterized mutations in both which give rise to herbicide resistance and modified function (31-33). The atomic structure has been resolved in bacterial reaction centers (4, 91,92), and the mechanism of inhibition suggested from kinetic studies has been confirmed by the observation that herbicides can replace the quinone substrate at the QB-site. The atomic structures provide no dynamic information, but much work has attempted to relate structure to function through studies of mutant strains (see 28 for recent review). With higher plants and C. reinhardtii, similar kinetic work has been mainly with mutants selected for herbicide resistance (24-26, 29-34).
Advances in the past decade have come from the development of methods for biochemical dissection and biophysical characterization of the oxygen evolving complex (see 10 for comprehensive review, 35-54). Three extrinsic oxygen evolution enhancing proteins (OEEP) have been identified, and their role in lowering the concentrations of Ca2+ and Cl- required for an active complex demonstrated, but although they are required for efficient reaction under physiological conditions, oxygen evolution can occur rapidly without them (35-40). The Mn-cluster is therefore assumed to require only ligands from the intrinsic proteins (D1, D2, CP43, CP47, two subunits of cytochrome b-559, and possibly the product of psbI) (10), and the entire electron transfer mechanism is demonstrated to reside in this complex.
The structure of the donor-side can be probed by a variety of spectroscopic methods (reviewed in 8-10, 41). The Mn cluster and its ligand environment have been extensively characterized through the EPR multiline, g=4.1 (S2-state) and g=4.8 (S1) signals (reviewed in 9, 41, see also 46-51). An anomalous EPR-signal observed when the normal transition to the S3-state is perturbed by Ca2+ or Cl- depletion has been attributed to an organic radical tentatively assigned to an oxidized histidine side-chain (64-67). The structure of the Mn-cluster has been probed by EXAFS, and distances to neighboring Mn and [O, N] atoms derived (52-54). Shifts in the K-edge X-ray of Mn have been demonstrated for two of the major transitions (S0S1S2), and interpreted as showing oxidation of Mn (42-44), but oxidation on the S2S3 transition is still controversial (43, 45). The S0-state is probably Mn2+:Mn3+:Mn4+:Mn4+, progressing by 1-electron oxidations to Mn4+:Mn4+:Mn4+:Mn4+ or Mn4+:Mn4+:Mn4+:His+ in S3 (9, 10).
The spectroscopic methods reviewed above provide no kinetic information, and require a commitment of time, biochemical material and equipment which make the cost of detailed analysis of numerous mutant strains exorbitant. Fortunately, much detailed information can be obtained by much simpler and less expensive methods based on fluorescence and spectrophotometry. Kinetic parameters for the S-state transitions (55-62) have been characterized by measurement of absorbance changes in the near UV, and by kinetic methods based on fluorescence (63). Thermodynamic characteristics have been probed with respect to the acceptor-side by measurement of back-reaction kinetics (73), delayed fluorescence (70) and thermoluminescence glow-curves (74). EPR and UV-spectrophotometry, have been used to follow redox changes of the secondary donor (YZ, or Z), identified as Y-161 of D1 (10-15).
The protolytic reactions are also easily measured photometrically (68). In chloroplasts, the 1, 0, 1, 2 pattern for release of the protons from water (in transitions S0-S1-S2-S3-S0) (68-70) has been shown to reflect non-integer stoichiometries for the different transitions, which vary with pH (71, 72). The pK values suggest that dissociable groups of proteins are the immediate source of H+. This is in accord with the electron transfers indicated by redox transitions of the Mn cluster; extensive evidence suggests that the S2 and S3 states carry an excess positive-charge, so the sub-stoichiometric H+-release must involve coulombic effects (71).
Most work to date has been with the systems developed using the cyanobacterium Synechocystis sp. PCC 6803 for mutagenesis of D1, D2 and CP43 subunits (reviewed in 10, 24, 26).
Donor side. Potential ligands to the Mn-cluster have been explored by extensive mutagenesis of both D1 and D2 (10, 16, 19-21, 67, 75). Coordination of a 4-Mn cluster would be expected to involve 24 sites, of which 10 would be filled by di-µ-oxo bridges, leaving 14 for other groups. Surprisingly few residues have been identified as of critical importance in oxygen evolution, and much speculation has centered around the question of the missing ligands. In other proteins, in addition to polar side-chains, many metal ligands are provided by back-bone carbonyl >C=O groups, and by water bridges, and it would be difficult to identify these by specific residue change. In addition, Cl- may be a ligand in PS II. Exploration of polar side-chains has focused on inter-helical loops on the lumenal side of the proteins. In D1, residue changes giving altered O2-evolution or Mn-binding in loop ab (D59, D61, E65), loop cd (D170, E189, H190), and the C-terminal span (H332, E333, H337, D342, and the terminal -COO- of A344) identify potential ligands. Residue Y161 provides the secondary donor (Z), D170 is necessary for binding of Mn with high affinity in the assembly of the active OEC, but probably does not provide a ligand to the Mn-cluster (127, 144, 145), H198 is a possible ligand to the putative [Chl]2, and H190 is essential for O2-evolution, a possible candidate for the putative redox-active histidine and/or a potential Mn-ligand. H195 as non-essential. These residues are in spans where the structure is unfortunately ill-defined. In D2, only E70 in loop ab looks like a good candidate.
Molecular Engineering of the D1 protein in C. reinhardtii. The past few years have seen major advances in molecular genetics and biochemical protocols which make C. reinhardtii an attractive model system for a combined biophysical and molecular genetic analysis of green-plant photosynthesis (32, 76-90). Efficient chloroplast transformation in C. reinhardtii using a particle gun has opened the way for specific mutagenesis of the chloroplast genome, and molecular engineering of the chloroplast-encoded proteins involved in photosynthesis (81). Two chloroplast genes encode the core proteins of PS II; the psbA gene for D1 (equivalent to the L subunit of bacterial reaction centers) and psbD for D2 (equivalent to the M subunit). The psbA gene in C. reinhardtii has previously been cloned and sequenced (82), and was shown to be the allele on which several herbicide resistant mutations were mapped (83). One copy of the psbA gene is present in each of the inverted-repeats of the chloroplast genome, of which ~40 copies are estimated to be in a gamete cell (84). Unlike those in higher plants, the psbA gene in C. reinhardtii is interrupted by four large introns. Although the polyploid chloroplast genome and the presence of introns have made it difficult to manipulate the psbA gene, several groups have succeeded in introducing site-directed mutations in C. reinhardtii (76-79, 85-87, 127), by transforming wild-type strains using a plasmid-borne mutagenized fragment, with or without a separate plasmid carrying the 16S rRNA gene conferring spectinomycin resistance. These protocols have several drawbacks, which relate to the presence of introns, the need for co-transformation with a separate plasmid to introduce selectable markers, and the lack of tailored features to allow application of cassette mutagenesis protocols. Johanningmeier and Heiss (88) have reported construction of a plasmid containing cDNA of the psbA gene, which was shown to transform a psbA deletion strain, but with poor efficiency. We have constructed an intron-free psbA by PCR-splicing which shows a high efficiency of expression (90, and see below); the poor efficiency seen in (88) could be attributed to the truncated 5'-untranslated region in their construct. Mutations at two positions, H190 and H195, on the donor side in PS II of C. reinhardtii have been constructed by Sayre et al. (76-79), and this lab has collaborated with Dr. Sayre in their characterization (77, 78). As in Synechocystis 6803, H195 as non-essential, and H190 is essential for O2-evolution. Changes to H, T, P and N at D170, thought to be involved in assembly of the Mn complex, have been constructed by Erickson (87), and characterized in some detail (127). Because of the availability of fluorescence methods, kinetic studies of the two-electron gate are more advanced in green plants than bacterial systems, even though we do not have a defined atomic structure. Our previous work has been on kinetic characterization of mutants selected for herbicide resistance (29-33). We demonstrated using fluorescence techniques that we could characterize the changes in rate constants, pK values and binding constants which defined the affects of mutagenesis at the mechanistic level, and this allowed us to explore the structure-function interface. An extensive literature on C. reinhardtii mutants selected for herbicide resistance has been recently reviewed (24-28), and the report of a recent symposium on the topic provides a range of papers covering recent developments (34).
Construction of an intron-free psbA gene, with aadA gene in tandem, and unique restriction sites. In order to overcome the difficulties in manipulating the native gene discussed above, we have constructed an intron-free psbA gene incorporating useful restriction sites and the aadA gene as a spectinomycin resistance (specr) cassette in tandem, by using PCR splicing (89, 90). We used this approach because the nucleotide sequence of the psbA gene had already been determined (82) and the clone with the native sequence was available, and because the final construct could be made with long and intact 5'- and 3'-flanking regions. Details of the construction of plasmids pBA155, pBA157 and pBA158 are given in Minagawa and Crofts (90). The plasmids all transformed the ac-u- psbA deletion strain of C. reinhardtii, with expression of psbA and aadA showing ~100% linkage, to give specr and pho+ colonies in high yield.
Mutagenesis of psbA gene and molecular engineering of the D1 protein. Site-directed mutations have been constructed in pBA155, 157 or 158 by a two-step PCR method (89, 90). Fragments carrying mutations were used to replace the wild-type segments as cassettes. Where indicated (Table I), the recombinant plasmids were then shot into a psbA deletion mutant. Bombarded cells were screened for specr and pho+ phenotypes. All the specr transformants are expected to contain the mutagenized psbA gene. Since the recipient deletion strain does not contain the wild-type locus, we do not have to worry about contamination by a wild-type copy of the psbA gene.
Mutagenesis of the psbD gene. The psbD gene encodes D2, the second core protein of PS II. There have been no previous reports of site-directed mutagenesis of this gene in C. reinhardtii, although several groups have introduced mutations in Synechocystis 6803. To construct a plasmid suitable for molecular engineering, a HindIII-PstI fragment containing the psbD gene has been subcloned in a phagemid vector, pGEM3fZ(+). Using single-stranded templates of the resultant construct, pBD102, we have introduced site-directed mutations using the 5-methyl cytosine method, as summarized in Table I. In order to use this system for transformation, we need to introduce a selectable marker, and have concentrated on finding alternatives to aadA (conferring a specr phenotype), as described below. In parallel work, Hutchinson and Sayre have constructed a plasmid with the aadA gene inserted at the 5'-end of psbD, which they have made available to us.
Selectable genetic markers for expression in chloroplasts. In order to extend future work to more ambitious projects using parallel mutagenesis in more than one gene, we need an alternative selection protocol, and have looked for new selectable genetic markers which can be used and expressed in C. reinhardtii. We have used the pBA157 plasmid to introduce the following potential marker genes in place of the aadA gene: gfp (green fluorescence protein), luxAB (luciferase), tetQ (tetracycline resistance), and ermF (erythromycin resistance). Expression of the genes could be readily tested by selection of pho+ colonies after biolistic transformation of the ac-u- psbA deletion strain, followed by appropriate screening. The most promising results have been obtained with ermF, where transformation has given pho+ strains showing resistance to erythromycin, and possibly with expression of tetQ.
Preliminary characterization of strains with mutations in D1. We have demonstrated a number of effects not previously seen in cyanobacterial systems. Table I summarizes our progress in characterization of mutant strains (Donor-side residue changes). Points of interest are: (i) Y161F looks very like the similar cyanobacterial mutant, with a low fluorescence after a single flash, showing no rapid donor to P680+. We have shown that oxidation of Chl and carotenoid provides a source of electrons for slow reduction of P680+. (ii) P162, and conserved aromatics at F180, or F182, are not essential, but F186L has a strongly perturbed donor side, leading to severely reduced growth. We have generated several revertants to this mutation but haven't had a chance to characterize or sequence them. (iii) Mutation of E189 to Q or L showed minor effects but, surprisingly, E189D failed to grow, and showed a blocked donor side after 1 flash, indicating a defect in the S1 or S2 transition. This indicates that charge and polarity are not important, but length of side chain is, suggesting that E189 has no ligand function, but stabilizes local structure by extending to a polar phase which aspartate cannot reach. (iv) N191 mutants showed more subtle effects on the donor side, suggesting stabilization of oxidized products by polarity, but no liganding; (v) S264G in the QB-site showed atraziner and properties similar to the same mutation in higher plants. Electron transfer to QB showed a normal rate constant, but diminished stability of QB-; electron transfer to QB- was normal. (vi) H252 clearly has a critical function; it cannot be replaced by glutamine, which shows a similar size and polarity, suggesting that the protolytic function might be important. The mutation gave rise to a severely blocked electron transport (~100 x slower), and a pho- phenotype, although the QB-site was still functional, as judged by further inhibition on addition of DCMU.
(H252 and S264 in the Qb-pocket)
TABLE I Residues under investigation in D1 Residue=>to Oligo. Seq. Plasmid In Pho+ Fluor.rise Back 2-elec. gate Chlamy. reaction Y161 F + + pBA155 + - very small presumed rapid not detect. P162 L + + pBA155 + ++ NY NY NY F180 L + + pBA155 + +++ normal normal normal F182 L + + pBA155 + +++ normal normal normal F182 Y + + pBA155 + +++ normal normal normal F186 L + + pBA155 + + NY NY NY E189 Q + + pBA155 + +++ slow? rapid+norm.phases normal E189 L + + pBA158 + +++ normal rapid+norm.phases ambig E189 D + + pBA158 + - F1. normal normal ambig F2. truncated - - N191 A + + pBA158 + +++ normal slow normal N191 L + + pBA155 + +++ normal slow normal N191 D + + pBA155 + +++ normal slowed normal H252 Q + + pBA155 + - normal normal inhibited S264 G + + pBA155 + ++ normal normal perturbed R257 K,E,M + + pBA158 + +++ NY NY NY R257 Q + + pBA158 + - NY NY NY Random mutagenesis NY pBA158 + NA NY NY NY Selectable marker genes co-transformed with pBA155 Marker gene Plasmid In Chlamy. Pho+ Expression ErmF pBA157 + +++ Yes (ermr) TetQ pBA157 + +++ Weak tetr? Gfp pBA157 + +++ Not detectable LuxAB pBA157 NY NY NY Residues under investigation in D2 Residue =>to Oligo. Seq. Plasmid In Chlamy. Y160 F + + pBD102 NY Y160 L + + pBD102 NY P161 A + NY pBD102 NY F179 L + NY pBD102 NY F181 L + NY pBD102 NY F185 L + NY pBD102 NY F188 Y + NY pBD102 NY N190 D + NY pBD102 NY Notes: Olig, oligonucleotide synthesized; Seq, cassette sequenced; Plasmid, plasmid used as transformation vehicle; Chlamy, C. reinhardtii transformed, with growth on TAP medium; Pho+, photosynthetic growth (wild-type rates +++, lower rates indicated by fewer +, - is pho-); Fluor. rise, variable fluorescence rise in the sub-micro second time scale; Back Reaction, Charge recombination measured by decay of variable fluorescence in the presence of DCMU; 2-elec. gate, kinetics of QA- oxidation after 1 or 2 flashes, showing binary pattern (normal) or normal kinetics after 1 flash but no binary pattern (ambig); NY, not yet completed.