The deletion of petG in Chlamydomonas reinhardtii disrupts the cytochrome bf complex.

The 4-kDa protein encoded by chloroplast petG copurifies with the cytochrome bf complex of spinach and is found in a number of other photosynthetic organisms, including the eukaryotic alga Chlamydomonas reinhardtii. To determine whether petG is involved in the function or assembly of the cytochrome bf complex, the gene was cloned from C. reinhardtii, excised from the DNA fragment, and replaced with a spectinomycin resistance cassette. A petG deletion strain of C. reinhardtii was then obtained by biolistic transformation. The resulting homoplasmic petG deletion strains are unable to grow photosynthetically, and immunoblot analysis shows markedly decreased levels of cytochrome b6, cytochrome f, the Rieske iron-sulfur protein, and subunit IV. To verify that this phenotype was due to the removal of petG, we also constructed a strain with a deletion in the open reading frame (ORF56), which is found 25 base pairs downstream of petG. The ORF56 deletion strain grew photosynthetically and had wild-type levels of the four major cytochrome bf subunits. We conclude that the absence of the PetG protein affects either the assembly or stability of the cytochrome bf complex in C. reinhardtii.

The 4-kDa protein encoded by chloroplast petG copurifies with the cytochrome bf complex of spinach and is found in a number of other photosynthetic organisms, including the eukaryotic alga Chlamydomonas reinhardtii. To determine whether petG is involved in the function or assembly of the cytochrome bf complex, the gene was cloned from C. reinhardtii, excised from the DNA fragment, and replaced with a spectinomycin resistance cassette. A petG deletion strain of C. reinhardtii was then obtained by biolistic transformation. The resulting homoplasmic petG deletion strains are unable to grow photosynthetically, and immunoblot analysis shows markedly decreased levels of cytochrome b 6 , cytochrome f, the Rieske iron-sulfur protein, and subunit IV. To verify that this phenotype was due to the removal of petG, we also constructed a strain with a deletion in the open reading frame (ORF56), which is found 25 base pairs downstream of petG. The ORF56 deletion strain grew photosynthetically and had wild-type levels of the four major cytochrome bf subunits. We conclude that the absence of the PetG protein affects either the assembly or stability of the cytochrome bf complex in C. reinhardtii.
The cytochrome bf complex, found in green plants, eukaryotic algae, and cyanobacteria, serves to connect photosystem I to photosystem II in the chloroplast or cyanobacterial electron transfer chain. The complex functions as a plastoquinol:plastocyanin/cytochrome c 6 oxidoreductase, the rate-limiting step of the photosynthetic electron transport chain. These redox reactions are coupled to an efficient translocation of protons across the membrane; the resulting proton gradient provides a source of energy for the synthesis of ATP by the chloroplast or cyanobacterial ATP synthase (1)(2)(3).
The cytochrome bf complex contains four large subunits. Three of these (cytochrome b 6 , cytochrome f, and the Rieske iron-sulfur protein) bind redox-active prosthetic groups. The remaining large subunit (subunit IV), together with cytochrome b 6 , forms the binding sites for plastoquinone oxidation and reduction. In addition to the four large subunits, a number of smaller polypeptides have been found to be associated with the cytochrome bf complex. A 4-kDa protein copurifies with the cytochrome bf complex in spinach, and antibodies raised to a synthetic decapeptide derived from maize petG cross-react with preparations from spinach, tobacco, pea, wheat, and rice, although not to those from Chlamydomonas reinhardtii or Synechocystis 6803 (4). Several small proteins were found with the spinach and the C. reinhardtii cytochrome bf complex, among them the PetG protein (migrating at 4.8 kDa in spinach and at 4.1 kDa in C. reinhardtii) and another protein believed to be a nuclear gene product (migrating at 3.7 kDa in spinach and at 3.8 kDa in C. reinhardtii) (5,6). The function of these small polypeptides is unknown.
Previously, the sequence and location of petG in the C. reinhardtii chloroplast genome were reported (7). To investigate the function of petG, we deleted the gene from the chloroplast, inserting in its place a spectinomycin resistance cassette containing a Chlamydomonas atpA promoter region and the bacterial aadA gene (8). Our results show that the petG deletion strains are incapable of photosynthetic growth, but that they grow heterotrophically on acetate. Components of the cytochrome bf complex are markedly diminished in these strains, indicating that the petG gene product is required for either the stability or assembly of the complex.

EXPERIMENTAL PROCEDURES
Cell Strains and Culture Conditions-C. reinhardtii wild-type strains CC-124, CC-125, and CC-1928 were obtained from the Chlamydomonas Culture Collection (Duke University), and wild-type strain 137 was obtained from M. Goldschmidt-Clermont (University of Geneva, Switzerland). Cells were grown on plates or in liquid culture at 20 -25°C under dim light in TAP or HSA medium or, where noted, in HS minimal medium (9).
Cloning petG-Oligonucleotides were synthesized by the DNA Synthesis Facility (Barker Hall, University of California, Berkeley, CA). End labeling of the multiply degenerate oligonucleotide and Southern hybridization procedures were performed according to Sambrook et al. (10). Chlamydomonas chloroplast DNA was isolated on a cesium chloride gradient using bisbenzimide dye (11). Plasmid DNA was sequenced using Sequenase following the manufacturer's instructions (U. S. Biochemical Corp.) or a modification of this procedure (12). Chlamydomonas cell extracts were prepared, and polymerase chain reactions were performed as described previously (13). Sequencing of double-stranded polymerase chain reaction products involved removal of primers by washing the oil-free 50-l reaction mixture three times with 400 l of TE buffer (10 mM Tris, 1 mM EDTA, pH 7.5) followed by concentration in a 100-kDa cutoff filter unit (Ultrafree-MC, Millipore Corp., Bedford, MA). The final volume was adjusted to 42 l. 7 l of polymerase chain reaction product was combined with 10 pmol of sequencing primer, 1 l of dimethyl sulfoxide, and 2 l of Sequenase buffer and annealed by heating at 95°C for 3 min and immediately cooling on dry ice. Sequencing was then carried out following the modified Sequenase procedure (12).
The following degenerate oligonucleotide was designed from the Nterminal region of the 4.1-kDa protein isolated from C. reinhardtii (5): ATGGT(T/A)GAACC(T/A)CTTCTTG(G/C)(T/A)GGTAT(T/C)GT. A sizespecific library of PstI-HindIII-digested chloroplast DNA (from wildtype C. reinhardtii strain CC-125) was constructed in pBluescript and probed, and a 3.6-kb 1 insert containing petG was isolated. Sequencing of petG revealed a discrepancy in the N-terminal protein sequence at the position of the seventh amino acid; it is cysteine, in agreement with the report of Fong and Surzycki (7).
Construction of pG14G and p56A15-To construct the petG deletion, the 3.5-kb PstI-HindIII chloroplast DNA fragment containing petG was cloned into pBluescript SK ϩ to produce pG35 (Fig. 1A). pG35 was then cut with AflII, filled in, and recut with PstI. The 1.2-kb fragment released was cloned into the PstI-SmaI site of pUC19, destroying the SmaI site, but regenerating the AflII site to form pGA. pUC-atpX-aadA was cut with HindIII and filled in to generate a 1.4-kb cassette containing the atpA promoter region followed by the bacterial aadA gene, which confers spectinomycin resistance when inserted into the C. rein-hardtii chloroplast genome (8). pGA was partially digested with SspI to remove a 134-bp fragment containing the petG coding region, and blunt-ended atpX-aadA was ligated into the gap to create pG14. The orientation of the cassette in pG14 is such that aadA is transcribed in the same direction as for petG. pG35 was opened at the HincII site in the multiple cloning site, and an EcoRI linker was inserted there to form pG35E. pG35E was cut with AflII and EcoRI to release a 2.4-kb fragment, which was inserted into the AflII and EcoRI sites in pG14 to create pG14G (Fig. 1B).
For the ORF56 deletion, pUC-atpX-aadA was cut with ClaI and PstI, releasing a 1.5-kb fragment, which was cloned into pBS-KAS ϩ , a modification of pBluescript in which the SmaI site is replaced with AflII. This permitted the subsequent excision of the atpX-aadA cassette with ClaI and AflII. pG35 was partially digested with KpnI, and the 3Ј-ends of the isolated single-cut fragment were blunted and religated to destroy the KpnI site in the multiple cloning region, creating pG35mK. pG35mK was cut with KpnI and AflII to generate a 462-bp fragment and a very large fragment that contained the vector. Both fragments were isolated. The 462-bp fragment was further digested with TaqI to release a 71-bp piece containing part of the ORF56 coding region. The remaining 391-bp KpnI-TaqI fragment was ligated to the isolated large KpnI-AflII vector-containing fragment and the ClaI-AflII-cut cassette (see above) in a single reaction to create p56A15, taking advantage of the compatible cohesive ends generated by TaqI and ClaI (Fig. 1C).
Transformation of Chlamydomonas-Plasmid DNA was isolated for transformation using QIAGEN columns, and 1-3 g was precipitated onto either M10 tungsten particles (Analytical Scientific Instruments, Alameda, CA) or 1-m gold particles (Bio-Rad) by standard methods (14). Chlamydomonas cells were grown and plated for transformation as described by Boynton et al. (15). The transformation utilized a PDS-1000/He biolistic device (DuPont NEN) using conditions described by Whitelegge et al. (16).
Hybridization Probes-To identify psbL, petG, ORF56, and ORF712 sequences, the probes pPSBL, pGS, p56, and pO712, respectively, were used (Fig. 1A). In addition, the following fragments were isolated as probes: an 800-bp HindIII-AflII petA fragment, a 960-bp SacI PetC fragment, a 300-bp PstI-HindIII petD fragment containing the petD coding region and ϳ40 bases of the 5Ј-noncoding region, and a 580-bp AccI-PvuII petB fragment. An 800-bp NcoI-PstI fragment containing the aadA gene from pUC-atpX-aadA was cloned into the EcoRV and PstI sites of pBluescriptII (eliminating all Chlamydomonas sequence) to form pAADA-NP. The 800-bp fragment was then cut from pAADA-NP to use as a probe for the aadA sequence. All probes were labeled using [ 32 P]dCTP by the random primer method (10).
Northern Hybridization-Total RNA was isolated from 500 ml of mid-log phase Chlamydomonas cells by a method modified from Merchant and Bogorad (17). The cells were pelleted and resuspended in 10 ml of TEN buffer (50 mM Tris-HCl, 150 mM NaCl, 15 mM EDTA, pH 7.5) containing 2% SDS and 40 g/ml proteinase K by shaking at room temperature for 20 min. 10 ml of TEN buffer-saturated phenol was added, and the bottle was returned to the shaker for 20 min. The suspension was transferred to a sterile capped polypropylene tube and centrifuged at 20°C for 10 min at 11,000 ϫ g. Organic extractions of the aqueous phase and LiCl precipitation of RNA were as described (17). The final RNA pellet was resuspended in 0.1 M sodium acetate, 5 mM magnesium acetate, pH 5.2, and digested with RNase-free DNase for 30 min at 37°C. The RNA was then extracted once with phenol and once with chloroform and precipitated with 2.5 volumes of ethanol.
Southern Hybridization-Chlamydomonas DNA either was isolated using a DNA miniprep (18) or was recovered from the LiCl supernatant of the RNA preparation described above. The latter method utilized a standard ethanol precipitation and RNase treatment to purify the DNA (10). Conditions for electrophoresis, transfer, and hybridization were conventional (19).
Antibody Production-A peptide (SKVYDWEFEERLE-C-) was de- Map of the region of the chloroplast DNA containing petG. A, shown is a diagram of plasmid pG35 containing psbL, petG, ORF56, and part of ORF712. The striped bars indicate the DNA fragments cloned to serve as probes for these four reading frames. B, petG is replaced with the aadA cassette to form plasmid pG14G. C, ORF56 is disrupted with the aadA cassette to yield p56A15. The sizes of the fragments that would be generated from a PstI-HindIII digest are shown. The restriction sites depicted are not necessarily unique.
signed to correspond to the N-terminal region of C. reinhardtii cytochrome b 6 with an added cysteine. It was synthesized and coupled to keyhole limpet hemocyanin at the University of Kentucky Macromolecular Structure Analysis Facility (Lexington, KY). This peptide conjugate (0.25 mg/ml) was used with Ribi adjuvant to inoculate rabbits at several subcutaneous sites using a standard regimen (20).
Immunoblotting-Protein samples were run on denaturing 15% polyacrylamide gels by the method of Laemmli (21). Samples for electrophoresis were prepared by extracting concentrated Chlamydomonas cells (5-15 g of chlorophyll at 0.5-1.5 mg/ml) with 4 volumes of acetone and centrifuging for 5 min to pellet the protein. The pellet was rinsed with acetone and allowed to air-dry briefly. The pellet was then solubilized with 30 l of sample buffer (1% SDS, 12.5 mM Tris, pH 6.8, 100 mM dithiothreitol) and incubated at 65°C for 3 min. Loading dye (6 l; 60% glycerol, 10 mg/ml bromphenol blue), was added, and the sample was loaded immediately. Lanes were loaded on an equal chlorophyll basis (22). Following electrophoresis, protein was transferred to nitrocellulose and detected using enhanced chemiluminescence (ECL, Amersham Corp.). Incubation, washes, and development were carried out following the manufacturer's instructions. Rabbit sera containing polyclonal antibodies generated against spinach cytochrome f, spinach subunit IV, maize Rieske iron-sulfur protein (from Dr. A. Barkan, University of Oregon), or a synthetic peptide conjugate (corresponding to C. reinhardtii cytochrome b 6 ) were used at a 1:10,000 dilution. The horseradish peroxidase-coupled goat anti-rabbit antibody was used at a 1:5000 dilution.

RESULTS
Sequencing of petG cloned from C. reinhardtii strain CC-125 revealed a transposition of a guanosine and a thymidine, corresponding to an arginine, not a leucine, at position 30, in contrast to a previous report (7). Sequencing of two other wildtype strains (CC-124 and 137) by polymerase chain reaction confirmed this sequence (Fig. 2).
Construction of the petG and ORF56 Deletions and Southern Analysis-To ascertain the role of the PetG protein in the cytochrome bf complex, the gene was replaced with a cassette containing 650 bp of the Chlamydomonas chloroplast atpA promoter region followed by the bacterial aadA gene encoding spectinomycin resistance (Fig. 1B) (8). ORF56 contains a large number of rare codons and is unlikely to be translated, but because of its proximity to petG (25 bp), we also constructed an insertional inactivation of this reading frame using the same spectinomycin resistance cassette (Fig. 1C). This was done to ensure that any effect of the petG deletion was not instead a result of unintended disruption of some function of ORF56.
The large constructs pG14G and p56A15 were transformed into Chlamydomonas CC-1928 using a biolistic particle delivery system, and transformants were selected on TAP medium containing 150 g/ml spectinomycin and 50 g/ml ampicillin. Transformants were recloned three to four times on selective plates and then tested for homoplasmy by Southern analysis (Fig. 3). Overexposure of the blot in Fig. 3A shows no indication of any remaining wild-type copies of the chloroplast genome, as evidenced by the absence of the 3.6-kb band in lanes 2-5. The 2.3-and 2.6-kb bands arise from the insertion of the aadA cassette in the petG and ORF56 deletions strains, respectively. As expected, the aadA probe hybridizes to these bands (Fig. 3B, lanes 2-5), but to nothing in the wild-type strain (lane 1). These two independent transformants with the deleted petG gene (GF-3 and GF-6) and two with the partially deleted ORF56 reading frame (56-2A and 56-6A) were selected for further characterization (Fig. 3).
Photosynthetic Activity-The ability of the petG mutant strains to grow photosynthetically was determined by streaking the strains on HS minimal medium plates and growing under high light. Fig. 4 shows that the petG deletion strains were unable to grow under conditions that required photosynthetic competence. These strains grew as normal on acetatecontaining medium under low light. The strains containing the ORF56 deletion grew as well as the wild-type strain under both conditions.
Immunoblotting-Immunoblot analysis of the petG deletion strains GF-3 and GF-6 showed substantially reduced levels of cytochrome b 6 , subunit IV, cytochrome f, and the Rieske ironsulfur protein relative to the wild-type levels of these proteins (Fig. 5). The level of protein was estimated to be 5-20% of the wild-type level for cytochrome b 6 and subunit IV and 10 -20% of the wild-type level for cytochrome f and the Rieske iron-sulfur protein, with some variation between the two petG deletion strains. There is an obvious difference in the level of protein found in GF-6 versus GF-3, with GF-3 having consistently lower levels of cytochrome bf complex proteins (Fig. 5). Other independent GF-transformants were observed to fall between these extremes. The ORF56 strains appeared as wild type in the levels of the subunits of the cytochrome bf complex (Fig. 5).
Northern Blots-The loss of subunits of the cytochrome bf complex in the absence of petG, as determined by immunoblotting, either could be due to instability of the assembled complex, leading to degradation of the individual components, or could be caused by a problem in transcription. To rule out the latter possibility, total RNA was isolated from each of the strains and hybridized with DNA probes for each of the cytochrome bf genes (Fig. 6). As expected, the RNA transcript for petG was absent in the petG deletion strains, but present in the ORF56 deletion strains. Levels of RNA transcripts for petB, PetC, and petD were normal in the petG deletion strains GF-3 and GF-6. Surprisingly, the level of transcript for petA was diminished relative to wild-type levels in both petG deletion strains. Equal loading of the RNA was assessed by probing the blots with an EcoRI fragment derived from the 16 S rRNA gene (data not shown).
The level of RNA transcripts in the ORF56 deletions strains was similar to the wild-type level or only slightly diminished in 56-6A, whereas markedly reduced levels were found for all five pet genes in 56-2A (Fig. 6). There seemed to be no functional significance of the reduced transcript levels in this strain as its ability to grow photoautotrophically was comparable to the wild-type strain (Fig. 4).
In addition to transcript levels of pet genes, we analyzed transcripts from genes and reading frames in the neighborhood of petG and ORF56 to see if these were affected by the removal of petG or ORF56 and the insertion of the aadA cassette. The psbL gene (upstream from petG; see Fig. 1) generated transcript levels that varied among the independent isolates of each deletion: in GF-6 and 56-2A, the transcript level was diminished relative to the wild-type strain, but in GF-3 and 56-6A, the level was normal (Fig. 6). The situation with ORF56 and ORF712, however, was still more complex. In the wild-type strain, neither reading frame yielded a single defined transcript, but rather, very low levels of several large, possibly unprocessed, precursors (Fig. 7A). In the GF-and 56-strains, the size of several of these large transcripts increased, suggesting that there was initiation at the atpA promoter portion of the inserted aadA cassette. This was verified by hybridization with an aadA probe (Fig. 7B). Although it is unclear whether these large low-level transcripts have any function, their levels are undiminished in the GF-and 56-strains, and we have concluded that it is unlikely that they have any role in the observed nonphotosynthetic phenotype of the petG deletion strains.

DISCUSSION
C. reinhardtii petG was cloned from chloroplast DNA of strain CC-125 using an oligonucleotide probe to identify the gene. Sequencing of petG from three C. reinhardtii strains showed a transposition of guanosine and thymidine in codon 30, resulting in an arginine rather than a leucine at this position, in contrast to a previous report (7). This arginine is present in the derived PetG amino acid sequence from Chlamydomonas eugametos as well as Cyanophora paradoxa, Marchantia polymorpha, and all higher plants for which the sequence is known (Fig. 2). It is absent only in Euglena gracilis among all the PetG sequences determined to date.
To determine whether the PetG protein is essential for the functioning of the cytochrome bf complex, petG was deleted and replaced with a spectinomycin resistance cassette. This resulted in the loss of photosynthetic function and decreased levels of all of the cytochrome bf complex subunits. In contrast, C. reinhardtii mutants that contained a disruption of the open reading frame immediately (25 bp) downstream of petG, ORF56, were fully photosynthetic and showed normal levels of cytochrome bf subunits. Therefore, the loss of cytochrome bf activity and protein was not due to an indirect effect on an adjacent reading frame, but rather was the result of the loss of petG itself.
The diminished level of cytochrome bf subunits could result from an aberration in transcription or processing of RNA transcripts or may be a result of a problem with the assembly or stability of the assembled complex in the absence of the PetG protein. However, Northern blots of total RNA from the petG deletion mutants showed that the transcript level and size for the two chloroplast-encoded genes, petB and petD, and the nuclear-encoded PetC gene were all normal. The transcript level for petA was diminished, but this was unlikely to be the cause of the reduced level of PetA protein or the inability of this strain to grow photosynthetically because strain 56-2A had even greater loss of petA transcript and yet had a wild-type level of cytochrome f and normal photosynthetic capability. It appears then that the absence of petG directly affects either the assembly or stability of the cytochrome bf complex.
The altered level of petA transcript in the two petG deletion strains may be a direct consequence of the deletion of petG, or it may also be a random effect similar to the variability in the levels of psbL transcripts or the variability in transcript levels between the two ORF56 deletion strains. In the latter two cases, the variability is most likely accounted for by secondary mutations in the Chlamydomonas genome possibly induced by the growth of the cells in fluorodeoxyuridine as part of the transformation protocol. On the other hand, it was noted by Kuras and Wollman (32) that in a 5-min pulse labeling of ⌬petB and ⌬petD Chlamydomonas strains, there is an extensive decrease in the synthesis of cytochrome f relative to the wild-type strain. This is not observed with cytochrome b 6 in ⌬petA or ⌬petD strains or with subunit IV in ⌬petA or ⌬petB strains. However, they attributed this decrease in synthesis to either a cotranslational or early post-translational regulation (32), not as a result of a decrease in the level of petA transcripts, as observed here with the petG deletion strain.
Other photosynthetic complexes have been observed to be destabilized in the absence of a small polypeptide component. In Chlamydomonas photosystem II, disruption of psbI, which encodes a 4.8-kDa polypeptide, causes photosensitivity of the organism and an 80 -90% loss of photosystem II complex relative to wild-type levels, although photoautotrophic growth is still possible (33). In contrast, a psbK disruption strain of C. reinhardtii cannot grow photosynthetically; this 4.1-kDa protein is apparently required for the assembly and/or stability of photosystem II (34).
With the cytochrome bf complex, it has been observed that the deletion of any one of the subunits causes a large decrease in the level of the remaining subunits. In Chlamydomonas, the deletion of petA results in 5% cytochrome b 6 remaining, 5% subunit IV, trace amounts of the Rieske iron-sulfur protein, and no detectable PetG protein (32). With a petD deletion, 10% of cytochrome f remains, and again, 5% of cytochrome b 6 , trace amounts of the Rieske protein, and no detectable PetG protein.
The deletion of petB follows the same pattern, except that subunit IV is found in barely detectable amounts (32). When a full complement of subunits was not present, the levels of cytochrome b 6 and subunit IV were found to be regulated by degradation; cytochrome f, as noted above, is believed to be regulated by cotranslation or at early post-translation. Therefore, it is not surprising to find that in the absence of the PetG protein, the remaining subunits of the cytochrome bf complex are found at a markedly decreased level.
The cytochrome bf complex has both structural and functional similarities to its bacterial and mitochondrial respiratory counterpart, the cytochrome bc 1 complex. However, the total number of subunits varies extensively among the cytochrome bc 1 complexes (35). The bacterium Paracoccus denitrificans has the minimal three subunits necessary to contain the redox centers, whereas the yeast and bovine mitochondrial complexes have 10 subunits each. In addition to the three subunits containing the prosthetic groups, these cytochrome bc 1 complexes have two large core subunits essential for assembly and five small subunits with a molecular mass of Ͻ15 kDa.
The effect of the deletion of each of the five small subunits has been studied extensively in yeast. Deletion of subunit 7 or 8 causes a full loss of activity and the spectral loss of cytochrome b, suggesting that the entire complex is absent (35). In contrast, deletion of subunit 6, 9, or 10 affects only the activity of the complex; in the case of a subunit 9 deletion, the loss of activity is Ͼ95%, but the complex is fully assembled (35)(36)(37). This provides a strong contrast to our results with the cytochrome bf complex, where the deletion of petG results in a large loss of every other subunit.
Although the sequence of the PetG protein is not homologous to any of the yeast small subunits, it is tempting to speculate that the PetG protein is functionally homologous to one of the two yeast small subunits (subunit 9 or 10) that likewise contains a single transmembrane helix. This type of homology has been proposed for bovine subunit 11, which in spite of a very minimal sequence identity to yeast subunit 10, shares the structural motifs of a transmembrane helix, a charge distribution of basic amino acids at the N terminus and both acidic and basic amino acids at the C terminus (36). Although PetG shares with yeast subunits 9 and 10 the feature of having both acidic and basic amino acid residues at the C terminus, the sole charged residue at the N terminus is a glutamate, rather than 3 or 4 basic residues. This difference in charge would be expected to confer upon the PetG protein the opposite orientation in the membrane relative to the yeast subunits. This prediction is in agreement with experiments on spinach and maize thyla- FIG. 7. Northern analysis of the petG and ORF56 deletion strains. The method used and the samples contained in the lanes are as described in the legend of Fig. 6. A, hybridization with the ORF712 probe (see Fig. 1); B, hybridization with an aadA probe. The bands on both blots were very faint and required a lengthy exposure, in comparison to the hybridizations in Fig. 6. koid membranes, where the hydrophilic C terminus of PetG was found to be sensitive to stroma-accessible protease activity (4). Because of this inverse orientation, it seems unlikely that PetG and the yeast subunits have any functional homology.
We have shown that the presence of the PetG polypeptide is essential to the C. reinhardtii cytochrome bf complex for either assembly or stability of the complex. Work is underway to determine whether, in addition, it may have a function in catalysis.