The Chloroplast-encoded α Subunit of Cytochromeb-559 Is Required for Assembly of the Photosystem Two Complex in both the Light and the Dark in Chlamydomonas reinhardtii *

The role of cytochrome b-559 in the photosystem two (PSII) complex has been investigated through the construction of a psbE null mutant by transformation of the chloroplast genome of the green alga Chlamydomonas reinhardtii. No PSII activity could be detected in this mutant either in oxygen evolution assays or by analysis of variable chlorophyll fluorescence. Immunoblotting experiments showed that the absence of PSII activity in the mutant was due to the loss of the PSII complex in both light-grown and dark-grown cultures. In contrast, the photosystem one reaction center polypeptide, PsaA, was present at wild-type levels in the mutant. RNA gel blot assays confirmed that the transcript levels for the psbA, psbD, andpsbF genes were unaffected by disruption of thepsbE gene, suggesting a post-transcriptional effect on their expression. Pulse-labeling experiments showed that either synthesis of PSII subunits was impaired in the psbE null mutant or there was extremely rapid degradation of newly synthesized subunits. Interestingly, the PsbE and PsbF subunits accumulated to wild-type levels in a psbA deletion mutant of C. reinhardtii, FuD7, which fails to synthesize D1 and assemble PSII. Our results provide evidence for a role for cytochromeb-559 in the early steps of assembly of the PSII complex, possibly as a redox-controlled nucleation factor that determines the level of PSII within the thylakoid membrane.

The role of cytochrome b-559 in the photosystem two (PSII) complex has been investigated through the construction of a psbE null mutant by transformation of the chloroplast genome of the green alga Chlamydomonas reinhardtii. No PSII activity could be detected in this mutant either in oxygen evolution assays or by analysis of variable chlorophyll fluorescence. Immunoblotting experiments showed that the absence of PSII activity in the mutant was due to the loss of the PSII complex in both light-grown and dark-grown cultures. In contrast, the photosystem one reaction center polypeptide, PsaA, was present at wild-type levels in the mutant. RNA gel blot assays confirmed that the transcript levels for the psbA, psbD, and psbF genes were unaffected by disruption of the psbE gene, suggesting a post-transcriptional effect on their expression. Pulse-labeling experiments showed that either synthesis of PSII subunits was impaired in the psbE null mutant or there was extremely rapid degradation of newly synthesized subunits. Interestingly, the PsbE and PsbF subunits accumulated to wild-type levels in a psbA deletion mutant of C. reinhardtii, FuD7, which fails to synthesize D1 and assemble PSII. Our results provide evidence for a role for cytochrome b-559 in the early steps of assembly of the PSII complex, possibly as a redox-controlled nucleation factor that determines the level of PSII within the thylakoid membrane.
Photosystem two (PSII) 1 is a multisubunit complex within the thylakoid membrane that catalyzes the light-induced oxidation of water to molecular oxygen as well as the reduction of plastoquinone to plastoquinol (1). At the heart of PSII is a the reaction center (RC) complex composed of the polypeptides D1, D2, PsbI, and a transmembrane b-type cytochrome termed cytochrome b-559 (2). As yet, the function of this cytochrome within PSII is unresolved, but it probably plays a role in protecting the PSII RC from photoinactivation, possibly by acting as a source of electrons to reduce long-lived chlorophyll cation species within the PSII RC (3), which unless reduced would oxidize nearby amino acids within the protein complex (4).
The structure of Cyt b-559 is quite unusual, as it is thought to be a heterodimer of two subunits arranged in parallel orientations: the ␣ subunit encoded by psbE and the ␤ subunit encoded by psbF (5). The haem molecule is likely to be ligated by the single histidine residues within each subunit (6). Although an ␣␤ heterodimeric structure for Cyt b-559 is generally favored, recent work has demonstrated that the genetic fusion of the N terminus of the ␤ subunit to the C terminus of the ␣ subunit in Synechocystis 6803 leads to the assembly of a functional cytochrome b-559 (7). Because the ␣␤ heterodimer model requires that the N termini of the monotopic ␣ and ␤ subunits lie on the same side of the membrane, the fusion of PsbE and PsbF without loss of PSII activity has been interpreted in terms of an ␣ 2 /␤ 2 homodimeric structure for Cyt b-559 with one haem bound to ␣ 2 and one to ␤ 2 (7). If this interpretation is correct, then Cyt b-559 would contain 2 haems and 2 copies each of ␣ and ␤. One way to investigate the roles of psbE and psbF in PSII function is to construct specific null mutants. In the transformable cyanobacterium Synechocystis sp. PCC 6803, as with all higher plants, psbE and psbF are cotranscribed with the downstream genes, psbL and psbJ (8), which also encode PSII subunits (9,10). The effect of mutating either psbE or psbF on PSII function is therefore complicated by possible polar effects on gene expression within this operon. This potential problem is highlighted by the work of Shukla and co-workers (11), which showed that the presence of a stop codon within the coding sequence for PsbE in a spontaneous mutant of Synechocystis 6803 led to loss of the entire psbEFLJ transcript.
Early studies showed that deletion of the entire psbEFLJ operon in Synechocystis led to loss of PSII activity (12,13) and in light-grown cultures to a failure to accumulate D2 and to a reduction in the levels of D1 and the PSII subunits CP47 and CP43 to about 50% of WT (11). Whether the loss of PSII activity in this mutant is due to increased rates of photoinactivation has not been resolved. Deletion of psbF also leads to loss of PSII activity (14), suggesting that the ␤ subunit alone is crucial for assembly of PSII. However, the stability of the psbEFLJ transcript was not investigated in the psbF mutant, so it remains possible that expression of the other genes within the operon had also been affected.
In addition to Synechocystis sp. PCC 6803, the green alga Chlamydomonas reinhardtii is also widely used to study PSII through genetic means. Because of its eukaryotic nature, Chlamydomonas is possibly a more appropriate model system with which to study photosynthesis in higher plants (15). Unlike cyanobacteria and higher plants, the psbEFJL operon is naturally disrupted in C. reinhardtii. The psbF gene is cotranscribed with the downstream psbL gene (16), whereas psbE is located approximately 7.3 kb away and transcribed in the opposite direction (16,17). This gene arrangement should thus allow the specific disruption of the psbE gene without affecting transcription of the psbF and psbL genes.
In this paper, we report on the construction and characterization of the first chloroplast psbE null mutant. Our results demonstrate the importance of the PsbE protein for assembly of PSII both in the light and the dark and suggest a model in which cytochrome b-559, besides having a role in electron transfer, also acts as a nucleation factor in the assembly of PSII. 2
Plasmid Constructions-Plasmid pF1 is a derivative of plasmid p78 (which contains the 15.1-kb PstI-4 fragment from the chloroplast genome of C. reinhardtii cloned into pBR322) (20) and was constructed by digesting p78 with EcoRI to remove a 5.8-kb fragment followed by religation. pF1 contains the psbE gene with upstream and downstream flanking sequences of approximately 8 and 2 kb, respectively ( Fig. 1).
To disrupt the psbE gene, a 1.9-kb spectinomycin-resistance cassette was isolated from plasmid pUC-atpX-AAD (21) after digestion with EcoRV and SmaI and ligated into the AlwNI site found within the psbE gene, after the ends were made blunt with the Klenow fragment of DNA polymerase I, to yield plasmid pF1aad. Standard methods were used to generate the recombinant DNA plasmids (22).
Transformation of Chlamydomonas-Chloroplast transformation experiments were performed according to Andronis et al. (23) using a particle gun (Shearline). Transformants were selected on TAP plates containing 100 g/ml spectinomycin. Spectinomycin-resistant colonies were restreaked three to four times on TAP plates containing spectinomycin before their genotypes were analyzed.
Isolation and Analysis of DNA and RNA-The isolation of total Chlamydomonas DNA and DNA gel blot analyses were performed according to Ref. 23. Total RNA was isolated from 100-ml cultures of mid-log phase Chlamydomonas cells as described in Ref. 24, in the presence of 100 M aurintricarboxylic acid (Sigma). RNA was electrophoresed in a 1.5% agarose gel containing formaldehyde, transferred to nylon membranes (Hybond-N, Amersham International), and hybridized to 32 P random-labeled specific DNA probes using the SSPE/formamide system as described (22).
To confirm the homoplasmicity of the psbE insertion mutants 3 and 4, a PCR analysis was performed using primers flanking the psbE gene. PCR reactions were performed in a Techne PHC-3 thermocycler following typical protocols (25) in the presence of Taq DNA polymerase (Promega). The oligonucleotides used as primers for PCR were SA1 (5Ј-TTGTTTCAATGGGGCATTAT-3Ј; located 407 base pairs upstream the psbE initiation codon), SA5 (5Ј-ATAGATGGTTTGAAAAGG-3Ј; located 301 base pairs downstream the psbE initiation codon), and B (5Ј-CACTGCCTCTAATAAAGTC-3Ј; located within the aadA cassette) (Fig. 1). For WT, primers SA1 and SA5 amplified a 0.75-kb fragment, whereas in strains 3 and 4, a 2.65-kb fragment was generated because of the presence of the 1.9-kb spectinomycin-resistance cassette (data not shown). No 0.75-kb band was amplified from strains 3 and 4 consistent with the absence of WT copies of the psbE gene. Using primers SA1 and B, a 0.5-kb fragment was amplified from strains 3 and 4 but not from the WT (data not shown).
Measurements of PSII Activities-Light-induced O 2 -evolution activity in whole cells was measured with a Clark-type electrode. The measurements were performed in TAP medium at 25°C, under saturating light conditions, in the presence of 1 mM potassium ferricyanide and 1 mM 2,6-dichloro-p-benzoquinone. Fluorescence induction kinetics of dark adapted cells, on TAP plates, were performed using a pulseamplitude modulation fluorometer (PAM 101, Walz) using a white actinic light of intensity (1000 E⅐m Ϫ2 s Ϫ1 ).
Protein Analysis-Thylakoid membranes were prepared as described (26) in the presence of a mixture of protease inhibitors (0.1 mM Pefab-lock (Boehringer Mannheim), 1 mM phenanthroline, 1 M pepstatin A, and 10 M leupeptin). The isolation was carried out essentially in the dark at 4°C, and throughout the procedure the chlorophyll concentration was measured according to the method of Arnon (27). The separation of protein by SDS-polyacrylamide gel electrophoresis and the immunodetection of protein was performed using methods described in Ref. 23. Visualization of bound antibody was performed by using a secondary antibody conjugated either to alkaline phosphatase (23) or to horseradish peroxidase, in the latter case followed by detection using enhanced chemiluminescence (ECL, Amersham).
Antisera specific for the ␣ and ␤ subunits of Cyt b-559 from C. reinhardtii were raised against 17-mer oligopeptides derived from the deduced N-terminal sequences of PsbE (AGKPVERPFSDILTSIR) and PsbF (acetyl-TTKKSAEVLVYPIFTVR) (17). Rabbits were immunized subcutaneously with the oligopeptides in the form of multiple antigenic peptides (purchased from Alta Bioscience, UK) following standard methodologies (28). Polyclonal antibodies specific for D1 (N-terminal fragment, spinach, gift of Roberto Barbato (29) (20) without sulfate but still containing trace elements that contain sulfur), resuspended in the same medium to 10 7 cells/ml, and incubated overnight under the original growth conditions. After another wash, the cells were resuspended in TAP-S (minus sulfur) (TAP containing no trace elements and made with Beijerinck's solution lacking sulfate) for 2 h, under 1000 E⅐m Ϫ2 s Ϫ1 , and then incubated with cycloheximide (Sigma) at 10 g/ml for 10 min. Pulse labeling of cells was performed with carrier-free Na 2 35 SO 4 (stock at 0.9 mCi/ml, ICN) at 0.1 mCi/ml for 5 min. Aliquots of cultures were immediately centrifuged, and cell pellets were frozen in liquid nitrogen and kept at Ϫ80°C. The cells were subsequently washed once with 50 mM MES/ NaOH, pH 6.5, containing 5 mM EDTA and a mixture of protease inhibitors (as described previously) and solubilized for 1 h in an equal volume of solubilization buffer. Proteins were separated by SDS-polyacrylamide gel electrophoresis containing 6 M urea, transferred onto nitrocellulose membranes, and exposed to x-ray film.

Construction of a psbE Insertion
Mutant-DNA sequencing was first performed to identify further genes in the vicinity of psbE. In agreement with Ref. 32, a putative rps9 gene, encoding the S9 subunit of the plastid ribosome, and the highly conserved ycf4 gene, which is involved in the assembly of PSI (32), lie downstream of psbE (Fig. 1). Northern blots indicate, however, that psbE is not cotranscribed with these genes so that disruption of psbE would not be expected to affect their expression (16,32,33). To generate a psbE null mutant of C. reinhardtii, a donor plasmid, designated pF1aad, was first constructed from the parental plasmid, pF1, by insertion of a spectinomycin-resistance cassette into an AlwNI site located 26 codons into the coding region of psbE (Fig. 1). This construct was introduced into the chloroplast of WT C. reinhardtii using the biolistic technique developed by Boynton and co-workers (34) (see "Experimental Procedures") and transformants selected on spectinomycin. Fig. 2 shows a Southern blot of total DNA isolated from three spectinomycin-resistant colonies (designated 1, 3, and 4) obtained from the transformation experiment. Two of the spectinomycin-resistant strains (3 and 4) gave a profile diagnostic for homologous recombination of the interrupted psbE gene into the chloroplast genome ( Fig. 2A). A psbE-specific probe hybrid-ized to a 1.9-kb HindIII fragment in the WT but to 1.4-kb and 0.9-kb HindIII fragments in these mutants. As expected, these mutants also contained the spectinomycin-resistance cassette (Fig. 2B). Spectinomycin-resistant strain 1 gave a WT-like profile and did not contain the spectinomycin cassette. Presumably, spectinomycin resistance arose in this strain through a spontaneous mutation. The data presented in Fig. 2 also indicate that mutants 3 and 4 were homoplasmic and did not contain WT copies of the psbE gene. Homoplasmicity of mutants 3 and 4 was also confirmed by PCR analysis using primers flanking the psbE gene (see "Experimental Procedures" and data not shown). Strains 3 and 4 therefore lack a functional psbE gene and were characterized further.
Biochemical Analysis of the psbE-inactivated Mutants-In contrast to WT, the psbE insertion mutants 3 and 4 were unable to grow photoautotrophically on minimal medium either on agar plates or in liquid (Table I). However, the psbE null mutants could be propagated in the presence of acetate either in the light or the dark, although their growth rates in the light were slightly slower than WT (Table I). Mutant cells grown on acetate in the light were tested for light-induced oxygen evolution using a Clark-type electrode. In contrast to WT, no oxygen evolution was detected in strains 3 and 4 either in the presence of bicarbonate (whole chain electron transport) or the artificial electron acceptors 2,6-dichlorobenzoquinone and ferricyanide (PSII-mediated electron transport) (Table I).
PSII activity was also investigated by measuring the chlorophyll fluorescence arising from cells of WT and mutant growing on TAP plates in the light. The ratio of variable fluorescence to maximum fluorescence (F v /F m ) is a measure of the efficiency of light capture by PSII. The psbE mutants showed no detectable variable fluorescence, consistent with the absence of functional PSII RCs, whereas the WT showed a F v /F m ratio of approximately 0.75.
Immunoblotting experiments were performed to assess if the PSII RC subunits, D1 and D2, accumulated in the thylakoid membrane in the absence of PsbE when mutant cells were grown in the presence of light. Fig. 3A shows that both D1 and D2 could not be detected under these conditions. To exclude the possibility that in the absence of PsbE, PSII can be assembled but is more prone to light-induced degradation, cultures grown heterotrophically in the dark were also examined. Fig. 3B shows that the levels of D1, D2, CP47, CP43, PsbF, and PsbO (also known as the 33-kDa extrinsic polypeptide) were drastically reduced in thylakoid membranes isolated from darkgrown cultures of the psbE mutants compared with WT. As judged from a dilution series, CP43 was present at approximately 1-10% of WT levels in strains 3 and 4 (Fig. 3B), and in overexposed blots D2 was found to accumulate to 1-5% of WT levels (data not shown). The amount of D1 protein in the psbE mutants was below the level of detection, which was estimated to be at 1% of WT levels. As expected PsbE was absent in the mutants (Fig. 3B). The CP47, PsbO (33-kDa polypeptide), and   -resistant colonies (1, 3, and 4), following digestion with HindIII. Controls were ϳ1 ng each of pF1, pF1aad, and a 2-kb aadA cassette (A, probed with a psbE-specific probe; B, probed with an aadA-specific probe).
PsbF subunits could not be detected in the mutant thylakoid membranes (Fig. 3B).
Transcript Levels in the psbE Insertion Mutant-RNA gel blots were performed to examine whether disruption of the psbE gene had effects on the accumulation of mRNA from psbF, located 7.3 kb away on the chromosome, or other PSII genes such as psbA or psbD. As shown in Fig. 4, the levels of the psbA, psbD, and psbF transcripts were similar in both psbE mutant 3 and WT when grown in either the light or the dark in the presence of acetate. As expected, the psbE null mutant failed to accumulate psbE mRNA (Fig. 4). Interestingly, in comparison with the psbA and psbD genes, the level of transcripts arising from the psbE and psbF genes in the WT appeared to be greater in the dark than the light.
Pulse Labeling Using [ 35 S]Sulfate-Pulse-labeling experiments were performed to investigate the rate of synthesis of plastid-encoded PSII subunits in the absence of the PsbE subunit. To allow visualization of chloroplast-encoded gene products, the pulse experiments were performed in the presence of cycloheximide, an inhibitor of cytoplasmic protein synthesis. Immunoblotting was performed to help assign the labeled bands. Fig. 5 shows the profile obtained with cells of WT, mutants 3 and 4, and FuD7, which is a psbA deletion mutant (19). As expected, the WT shows good labeling of the D1, D2, CP47, and CP43 proteins. For FuD7, no D1 synthesis was observed, but as previously documented (19,35), in the absence of D1 synthesis, some incorporation of the D2, CP43, and CP47 subunits into the thylakoid membrane was detected. Mutants 3 and 4 showed radiolabeling profiles more similar to FuD7 than WT. Of the major PSII subunits, D2 and CP43 were the most conspicuously labeled in the psbE mutants, although the de- gree of radiolabeling, as in FuD7, was much less than WT. These results suggested that the rate of synthesis of PSII subunits is reduced in the psbE null mutants or that there is rapid degradation of the subunits in the absence of PsbE. Similar results were also obtained when pulse labeling of the cells was performed in the dark (data not shown).
Although FuD7 fails to assemble a functional PSII complex and has depleted levels of D2, CP47, and CP47 compared with the WT (19,35), the PsbE and PsbF polypeptides could still be detected at WT levels in thylakoid membranes using specific antibodies (Fig. 6).

DISCUSSION
In this paper, we have described the construction of a psbE insertion mutant of C. reinhardtii. A particular advantage of using C. reinhardtii over Synechocystis 6803 to construct such a mutant is that pleiotropic effects on the transcription of the downstream psbFLJ genes in Synechocystis 6803 are avoided. Disruption of the psbE gene in C. reinhardtii leads to an inability to accumulate subunits of the PSII core complex within the thylakoid membrane in both dark-grown (Fig. 3B) as well as light-grown material (Fig. 3A). Loss of PSII is therefore not due to enhanced photoinactivation of the PSII complex, a possibility that was not excluded in the analysis of previous Synechocystis mutants affected in cytochrome b-559 biosynthesis (8,11,12).
Northern blots confirm that the psbF, psbD, and psbA transcript levels are unaffected in the psbE null mutants so clearly the inability to assemble PSII lies at a post-transcriptional step. Pulse-labeling experiments revealed that the synthesis of the major PSII core polypeptides was drastically impaired in the psbE null mutants (Fig. 5) consistent with either a block in synthesis or the rapid degradation of newly synthesized proteins because of a failure to assemble them into a stable core complex.
A number of PSII null mutants have now been constructed in Chlamydomonas through biolistic transformation. The psbH (36) and psbK (37) mutants, like the psbE mutant described here, also fail to accumulate PSII. However, in pulse-labeling assays, both the psbH and psbK null mutants show a WT-like profile, indicating that it is the degradation rather than the synthesis of PSII subunits that is affected in these mutants. The labeling pattern for the psbE null mutant on the other hand is very similar to that of FuD7, which contains a disrupted psbA gene. Our data therefore support a model in which the ␣ subunit is important for promoting synthesis of D1 or for retarding its degradation. As such the ␣ subunit appears to be an important component in the early steps of the assembly of the PSII complex.
A role for cytochrome b-559 in the early steps of assembly of PSII is also supported by recent in vitro pulse-chase experiments that have shown that radiolabeled D1 protein can be assembled into the PSII core complex through the stepwise addition of subunits (38). Whether D2 or cyt b-559 is the first partner for D1 following insertion into the thylakoids remains to be established. However, it is clear that the PsbI subunit found in the isolated PSII RC (2) is not required for assembly of PSII (39).
Early studies on the psbEFLJ deletion mutant of Synechocystis indicated that D1 and D2 were absent from the thylakoid membrane but that CP47 and CP43 accumulated to WT levels (12). Later work showed that when an additional protease inhibitor, leupeptin, was used to avoid inadvertant proteolysis of thylakoid membrane protein and care taken to avoid saturation of the signals in immunoblotting experiments, the levels of D1, CP43, and CP47 were approximately 50% of WT, and only D2 failed to accumulate (11). In the studies reported here, a mixture of protease inhibitors including leupeptin was used during the isolation of the thylakoid membranes, and in addition whole cell extracts were analyzed (Fig. 5). The absence of D2 in the Synechocystis psbEFLJ mutant has been interpreted in terms of a specific interaction between D2 and cyt b-559 during assembly of PSII (11). However, as noted in the Introduction, a psbEFLJ deletion mutant is not an appropriate strain to study the specific role of PsbE or PsbF in PSII assembly, and so it is unclear how far the results in the cyanobacterial psbEFLJ mutant can be compared with the algal psbE null mutant described here.
It is clear, however, that quite different phenotypes can be displayed by Synechocystis 6803 and Chlamydomonas for the same type of mutation. For example, null mutants of the Chlamydomonas psbH (36) and psbK (37) genes lead to a destabilization of the PSII core complex, reminiscent of the phenotype of the psbE null mutants presented here, whereas Synechocystis 6803 psbH (40) and psbK (41,42) null mutants assemble functional PSII and are still capable of photoautotrophy. The more drastic effect seen in the Chlamydomonas mutants is consistent with the presence of a more efficient proteolytic system in the chloroplast for the removal of mis-assembled PSII complexes.
We have prepared in the course of this work antibodies specific for the PsbE and PsbF proteins of Chlamydomonas, which has allowed for the first time unambiguous examination of the expression of these proteins in other Chlamydomonas PSII mutants. Interestingly, we found that both PsbE and PsbF accumulate to WT levels in FuD7, which cannot synthesize D1 and lacks detectable levels of D2 and CP47 (19,35). The cytochrome b-559 apopolypeptides therefore appear to be stable in the thylakoid membrane of Chlamydomonas in the absence of many of the PSII subunits but is required for their accumulation. Cyt b-559 is also present at approximately WT levels in a psbA null mutant of Synechocystis 6803 (43) and a psbE mutant of the same organism that is truncated at the C terminus by 31 residues yet contains only a low level of assembled PSII (44). The accumulation of Cyt b-559 in the absence of D1 and D2 is therefore likely to be a common feature among oxygenic photosynthetic organisms. Cytochrome b-559 may therefore act as a nucleation factor for assembly of PSII in both cyanobacteria (44) and chloroplasts (this work). Indeed, studies on the light-induced development of the chloroplast has indicated that cytochrome b-559, detected spectrophotometrically, accumulates in etioplasts in the absence of the majority of the PSII subunits (45). We suggest that the level of PSII within the thylakoid membrane may be controlled by the abundance of cyt b-559 and that its redox state, which is controlled by reactions within the thylakoid membrane, may be one of the mechanisms by which the assembly of PSII is regulated. Assembly of PSII may be closely coordinated with D1 synthesis, which is known to be under redox control (46).
In previous work (47), the stoichiometry of the ␣ and ␤ subunits to the D1 and D2 subunits within the isolated PSII RC from C. reinhardtii, as determined by 14 C labeling, was found to be approximately 1:1:1:1 with 1 haem present per PSII RC. This result led to the conclusion that either cyt b-559 was present as a ␣␤ heterodimer or else there was an equimolar mixture of PSII RCs containing either ␣ 2 or ␤ 2 haem-containing complexes (33,47). Because accumulation of ␤ within the thylakoid membrane is dependent on the presence of ␣ (Fig. 3), it is likely that each PSII RC contains both ␣ and ␤. Hence, the mutagenesis data combined with the isotopic labeling data provide support for the ␣␤ heterodimer model for cyt b-559. Such a conclusion is at odds with the recent finding that an active cytochrome b-559 was synthesized in a strain of Synechocystis 6803 in which the psbE and psbF genes were fused (7). One possible explanation is that the natural His ligand present in the psbF gene product at codon 23 has been functional replaced by an alternative ligand within the same subunit. Inspection of the psbF sequence suggests that Met-38 toward the C terminus is a possible candidate. Although replacement of the His ligand by the Met ligand may be expected to destabilize haem binding, the covalent fusion of the ␣ and ␤ subunits may enhance its stability.