Cytochrome c Methyltransferase, Ctm1p, of Yeast*

Cytochromes c from plants and fungi, but not higher animals, contain methylated lysine residues at specific positions, including for example, the trimethylated lysine at position 72 in iso-1-cytochrome c of the yeast Saccharomyces cerevisiae. Testing of 6,144 strains of S. cerevisiae, each overproducing a different open reading frame fused to glutathione S-transferase, previously revealed that YHR109w was associated with an activity that methylated horse cytochrome c. We show here that this open reading frame, denoted Ctm1p, is specifically responsible for trimethylating lysine 72 of iso-1-cytochrome c. Unmethylated forms of cytochromec but not other proteins or nucleic acids are methylatedin vitro by Ctm1p produced in S. cerevisiae orEscherichia coli. Iso-1-cytochrome c purified from a ctm1-Δ strain is not trimethylated in vivo, whereas the K72R mutant form, or the trimethylated Lys-72 form of iso-1-cytochrome c, are not significantly methylated by Ctm1p in vitro. Like apocytochromec, but in contrast to holocytochrome c, Ctm lp is located in the cytosol, consistent with the view that the natural substrate is apocytochrome c. The ctm1-Δ strain lacking the methyltransferase did not exhibit any growth defect on a variety of media and growth conditions, and the unmethylated iso-1-cytochrome c was produced at the normal level and exhibited the normal activity in vivo. Ctm1p and cytochrome c were coordinately regulated during anaerobic to aerobic transition, a finding consistent with the view that this methyltransferase evolved to act on cytochrome c.

CTM1 gene was disrupted by replacing the entire open reading frame of the gene with the kanMX4 gene. The polymerase chain reaction-generated fragment required for producing the ctm1-⌬::kanMX4 disruption was prepared by the method of Baudin et al. (21), using the pFA6-kanMX4 plasmid (22) as template and the following two synthetic oligonucleotides, where the underlined segments are the sequences homologous to the G418/kanamycin cassette, pFA6-kanMX4, and where A in the ATG initiation codon of CTM1 is assigned position 1: the upstream disruption primer (Ϫ64), 5Ј-AATAGTATAATTCGCCATCCT-CATAACCACTGAAAAATCGAAGTTAACAGCTGAAGCTTCGTA; and the downstream disruption primer (ϩ1845), 5Ј-CGCGTCAGATTGTT-CTTTGTGATGCTTTTAATGTAGAAGAAAGGACAACATCAGGCTCC-TGCAT. The following two synthetic oligonucleotides were used to identify the correct disruption by polymerase chain reaction: the upstream screening primer (Ϫ105), 5Ј-GTAGTTAGCAATGCTAGATCGT; and the downstream screening primer (ϩ1922), 5Ј-GTATTACACTAC-AAAGCTCCTC.
Construction of Epitope-tagged Ctm1p-Myc-The gene encoding the Ctm1p-Myc protein was constructed with the epitope tagged at the C terminus by using the following two synthetic primers, where the sequence in bold corresponds to the c-Myc epitope DMEQKLISEEDLN, and where the underlined segments correspond, respectively, to SalI and NheI restriction sites: the downstream primer, 5Ј-AGAAGAAAG-GTCGACTAAAATCAATTCAAGTCCTCTTCGGAAATGAGCTTCT-GCTCCATATCCTGAAAGAAA-3Ј; and the upstream primer, 5Ј-GAA-GGATAATACCGCTAGCTAGACAATCTTA-3Ј, with the position of the 5Ј-end at Ϫ200 nucleotides of the CTM1 gene. The product of the polymerase chain reaction, whose sequence was verified by DNA sequencing, contains the native promoter and the CTM1 gene fused to the segment encoding the c-Myc epitope at the C terminus. The ctm1-⌬ strain B-12331 was transformed with plasmid pAB2578 that was obtained by ligating the NheI-SalI fragment of CTM1-Myc to the SpeI-SalI sites of plasmid pAA625 (pRS316). The resulting CTM1-Myc strain, B-12588, was used for subcellular localization studies of Ctm1p.
Expression of CMT1 in Escherichia coli-Expression of CMT1 was carried out with the pET System, using the pET15b plasmid and the E. coli strain BL21, as described in the pET System manual (Novagen, Madison, WI).
Preparation of mRNA and Northern Blot Analysis-Total RNA was isolated from approximately 10 8 cells as described by Russo et al. (23). Enriched poly(A) RNA was isolated from a total of 1 mg of total RNA with the Oligotex mRNA kit (QIAGEN Inc., Valentia, CA) as recommended by the vendor. Northern blot analyses of different mRNAs were carried out as described previously (23). Messenger RNA levels were quantified by PhosphorImager analysis (Model 425E, Molecular Dynamics, Sunnyvale, CA) and finally normalized against ACT1 mRNA signals.
Subcellular Fractionation-The yeast cell extracts were prepared by disrupting the cells with glass beads in Buffer A containing 50 mM Tris, pH 7.5, 150 mM NaCl, 5 mM EDTA, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride. After vortexing 6 times for a total of 3 min, the lysates were clarified by low speed centrifugation. The membrane fraction was obtained by centrifuging the cell extracts at 13,000 ϫ g for 10 min at 4°C, and the resulting pellet was solubilized in Buffer A containing 1.0% Triton X-100. The supernatant contained cytosolic proteins. The mitochondrial protein fraction was prepared separately by the procedure described by Daum et al. (24).
Western Immunoblotting-The aliquots of the samples from cytosolic, membrane, and mitochondrial fractions, as well as total cell extract, were loaded on Laemmli 10% SDS polyacrylamide gels (25), and the separated proteins were transferred to ECL nitrocellulose filters (Amersham Pharmacia Biotech) with a standard running condition of 3 h at 75 v as described by Towbin et al. (26). Subsequently, the filters were probed with anti-Myc 9E10 monoclonal antibody (Babco, Richmond, CA). A total of 15 g of protein from total cell extracts and 5 g of protein from cellular fractions were loaded on the gel. After Western blotting, the filters were developed by using ECL reagents (Amersham Pharmacia Biotech).
Indirect Immunostaining of Yeast Cells-Yeast cells of the strain B-12588 were stained by standard indirect immunofluorescence technique with mouse anti-Myc antibody. Briefly, cells were grown to early log phase in a synthetic medium lacking uracil, collected, and fixed in formaldehyde. Subsequently the cells were treated with zymolyase for 5 min, at which time the reaction was blocked in phosphate-buffered saline/BSA solution. Spheroplasts were incubated for 1 h with anti-Myc antibody and for 30 min with anti-mouse IgG fluorescein isothiocyanate conjugate. After extensive washing, the cells were examined with an Olympus microscope containing a fluorescein filter.
Iso-1-cytochrome c Content-The amounts of iso-1-cytochromes c, as well as the other cytochromes, were determined by spectrophotometric recordings of intact cells at Ϫ196°C, as described previously (27).
Content of Trimethyllysine in Iso-1-cytochromes c-The degree of trimethylation of iso-1-cytochrome c was estimated by direct amino acid compositional analysis. Iso-1-cytochromes c were purified by two rounds of chromatography on weak cation-exchange BioRex70 column, first with 100 -200 mesh and subsequently with 200 -400 mesh (Bio-Rad) in potassium phosphate buffer, pH 7.0, with a 0 to 1.0 M KCl linear gradient. If necessary, the protein samples were concentrated with a Centricon-3 device after chromatography (Amicon-Millipore, Bedford, MA). The amino acid composition of 500-pmol samples of purified iso-1-cytochromes c were analyzed after hydrolysis at 110°C in vacuo for 24 h in 5.7 M HCl. Analyses were carried out with a Hewlett Packard system, using ODS Hypersil (C18 reverse phase type), 5 m, 200 ϫ 2.1-mm column. The peaks for each amino acid, including trimethyllysine, were identified by comparison of the HPLC profiles with the HPLC profiles produced with known amino acid mixtures.
Purification and Assay of Methyltransferase-Yeast strain MRM2122 was homogenized with glass beads and GST-Ctm1p was purified by glutathione-agarose affinity chromatography as described previously (17) The reaction mixtures were incubated at 30°C for 1 h, placed on ice, and 125 l of 10 mg/ml BSA and 950 l of 20% trichloroacetic acid were added. Tubes were quickly frozen with dry ice and thawed on wet ice and then centrifuged for 20 min at 4°C. The supernatants were discarded, and the pellets were washed with 1 ml of 5% trichloroacetic acid and dissolved in 150 l of 88% formic acid. A total of 3 ml of Eco-scint (National Diagnostics, Atlanta, GA) were added, and [ 3 H]methylated products were counted in a Beckman scintillation counter and corrected for controls lacking methyltransferase.

RESULTS AND DISCUSSION
Ctm1p Is a Methyltransferase That Specifically Acts on Cytochrome c-Martzen et al. (17) previously reported that the open reading frame, YHR109w, was associated with an activity that methylated horse cytochrome c, but not BSA, as measured by incorporation of [ 3 H]SAM into acid-precipitable counts. We have further investigated the role of the YHR109w, now denoted Cmt1p, as a methyltransferase by examining a wider range of substrates and by using more highly purified Ctm1p.
As summarized in Table II, GST-Ctm1p prepared from strain MRM2122 (Table I) acted on horse cytochrome c, but not significantly on eight other proteins, or on tRNA or bulk DNA. At low substrate concentration, the [ 3 H]methyl group incorporation in non-cytochrome c samples was less than 2% of the incorporation found with cytochrome c.
Furthermore, we established that Ctm1p encodes the observed methyltransferase activity, rather than simply co-purifying with the activity. First, overproduction of GST-Ctm1p in yeast yielded extracts with more activity, as expected if it were the rate-limiting component of activity. Second, expression of (His) 6 -Ctm1p in E. coli resulted in an activity acting with a high specificity on cytochrome c (Table II). In fact, approximately 75% pure preparations of GST-Ctm1p from yeast and of (His) 6 -Ctm1p from E. coli had approximately the same specific activities over a range of enzyme concentrations (Fig. 1). Thus, Ctm1p is sufficient by itself to act as a methyltransferase specifically for cytochrome c.
Comparisons of the deduced Ctm1p open reading frame, presented in Fig. 2A, to various data bases by using BLAST in different variations (28) did not reveal any significant overall similarity to any other protein. Also Ctm1p did not contain any of the known motifs (motif I and posts-I, -II, or -III) deduced from a large number of SAM-dependent methyltransferases (29,30). However, a Protein Data Bank Blast, which is based on 3-dimensional structures, revealed that a Ctm1p region spanning approximately 100 amino acids was similar to a region of the rRNA methyltransferase from Streptococcus pneumoniae that methylates an adenine of 23 S ribosomal RNA, preventing macrolide-lincosamide-streptogramin antibiotics from binding to the ribosome and thus causing antibiotic resistance (31). As shown in Fig. 2B, the 99 amino acid regions of Ctm1p and rRNA methyltransferase were 24% identical and 44% similar. However, the functional significance of this similarity is unknown.
Ctm1p Only Trimethylates Lysine at Position 72-The critical evidence that CTM1 encodes the methyltransferase that trimethylates Lys-72 in iso-1-cytochrome c came from the amino acid compositional analysis of iso-1-cytochrome c from the ctm1-⌬ strain. The ctm1-⌬ strain was prepared by replacing the CTM1 gene with kanMX4 gene (Table I). As discussed below, the ctm1-⌬ strain contained the normal level of iso-1-cytochrome c. However, as shown in Fig. 3, trimethyllysine could not be detected in iso-1-cytochrome c from the ctm1-⌬ strain. Furthermore, trimethyllysine was present at the normal level in iso-1-cytochrome c from strain B-12462 (Table I), in which the CTM1 gene was reintroduced in the ctm1-⌬ strain (data not presented).
We have also provided evidence that Ctm1p methylates exclusively Lys-72 in vitro. Iso-1-cytochrome c from strain B-6978, containing normal iso-1-cytochrome c with Tml72, is a much poorer substrate than that from B-12331, containing iso-1-cytochrome c that is not trimethylated (Table III). Presumably, the 4% low level of incorporation in iso-1-cytochrome c from the normal strain is due to the fact that cytochromes c from normal strains are not completely methylated in vivo and are approximately 5-10% unmethylated (15). More importantly, iso-1-cytochrome c with the K72R replacement (19) had no significant level of methylation (Table III).
The lack of methylation at any site other than Lys-72 is consistent with the results of Takakura et al. (32), who investigated the amino acid sequence requirement for trimethylation of Lys-72 by examining 21 altered iso-1-cytochromes c having single replacements in the region encompassing residues 67 through 77. Their study revealed that tyrosine 74 is critical for trimethylation of Lys-72, whereas replacements at other positions did not produce significant diminutions. However, other similarly spaced lysine and tyrosine residues at other sites in the protein did not result in trimethylation of the lysine residue. Thus, a properly situated aromatic residue, determined by the overall conformation of apocytochrome c in the vicinity of Lys-72, appears to be essential for trimethylation.
Ctm1p Is Located in the Cytosol-The cellular location of Ctm1p was investigated with strain B-12588 containing the epitope-tagged allele CTM1-Myc. Indirect immunofluorescence

. Specific activities (pmol of [ 3 H]methyl groups transferred per h per g) of either GST-Ctm1p prepared from S. cerevisiae (E) or (His) 6 -Ctm1p prepared from E. coli (q).
Both GST-Ctm1p and (His) 6 -Ctm1p were approximately 75% pure as determined by Coomassie staining of SDS gels. staining revealed that Ctm1p was uniformly distributed throughout the cell, consistent with the view that Ctm1p is located primarily in the cytosol (data not presented). More importantly, subcellular fractionation studies also revealed that Ctm1p is located in the cytosol, with only insignificant levels in mitochondria (Fig. 4). The slight difference in mobility of Ctm1p in the figure is due to a higher concentration of total protein in lane T, including membrane proteins and other cellular components, which reduces migration. Although the anti-Myc antibody produced a nonspecific background, similar Ctm1p bands were observed with cellular fractions using anti-GST antibody and strain MRM2122 (data not presented).
Finding Ctm1p in the cytosol is consistent with early studies suggesting that the methyltransferase acts on apocytochrome c in the cytosol before imported in mitochondria and before heme attachment. Holocytochrome c was found to be a poorer substrate than apocytochrome c, which was prepared either from holocytochrome c or by translation in vitro (15,33). Also, both apocytochrome c and holocytochrome c lost their substrate capability when bound to mitochondria in low ionic conditions, but this capability was restored when the cytochromes c were released from the mitochondria by KCl treatment. Interestingly, cycloheximide inhibited both protein synthesis and methylation (34). These results taken together suggest that nascent apocytochrome c is methylated in vivo, probably cotranslationally, before being associated with mitochondria (15,33).

The ctm1 ⌬ Strain Has No Observable Functional Defects-
The isogenic CTM1 and ctm1-⌬ strains did not differ in their growth as determined by testing diluted suspension on various media and conditions, including growth at 30 and 37°C on rich yeast extract-and peptone-containing media, or synthetic media with a variety of different carbon sources such as glucose, ethanol, lactate, or glycerol. In particular, no difference was observed on synthetic lactate medium, which is especially sensitive for revealing functional defects in cytochrome c (35). Also, the same comparative growth of the two strains was observed on media containing 1M NaCl or KCl and media adjusted to pH values of 8.0, 8.5, or 9.0.
As expected, the isogenic CTM1 and ctm1-⌬ strains contained the same level of iso-1-cytochrome c, as well as the same level of the other cytochromes (Fig. 5).
Thus, our studies, as well as previous studies, have not revealed the biological function and importance of cytochrome c methylation. The results of some experiments performed in vitro led to the suggestion that Tml72 is required for mitochondrial import of cytochrome c in yeast (33, 36 -38), a conclusion that was later refuted by Ceesay et al. (39). On the other hand, pulse-chase experiments suggested that Tml72 protected cytochrome c from proteolytic degradation in vivo (40). At the end of a 40-h chase period, the extent of degradation of the unmethylated form was three times higher than that of the methylated form. However, cycloleucine, used to inhibit cytochrome c methylation, could have also affected the protein degradation system.
Ceesay et al. (41) also investigated the turnover of iso-1cytochrome c having amino acid replacements at positions 71-74 and, consequently, different levels of trimethylation of Lys-72. Although the K72R iso-1-cytochrome c had the shortest half-life, 9 h compared with 23 h for the normal protein, there was no significant correlation between the half-life and the degree of trimethylation of Lys-72 among the 6 other iso-1cytochromes c having various replacements.
Pollock et al. (42) reported that the alkaline conformational transition of the Lys-72 iso-1-cytochrome c was approximately 0.6 pK a units lower than the Tml72 iso-1-cytochrome c and suggested that the role of Lys-72 methylation was to diminish an alkaline conformer at high pH values. However, the biological significance of this alkaline transition is unclear, and we did not observe difference of growth on lactate and other media at high pH values (see above).
The function of trimethylated Lys-72 has been investigated   in yeast by examining the consequences of K72R (19) and K72D (32) replacements. Spectroscopic measurements revealed that K72R and K72D iso-1-cytochromes c were at normal or near normal levels. Furthermore, growth in lactate medium indicated that the K72R iso-1-cytochrome c had normal or near normal activity in vivo (19). On one hand, the maintenance of Lys-72 in 96 different species, except Tetrahymena pyriformis (43), and the maintenance of the apparently specific methylase in numerous plant and fungal species indicate that Lys-72 or Tml72 is essential from an evolutionary point of view. On the other hand, the results with the K72R and K72D iso-1-cytochromes c indicate that a lysine or trimethylated lysine at position 72 is not absolutely required for biosynthesis, mitochondrial import, or activity, and the residue at this position is not critical. It should be noted that conserved residues at other sites in iso-1-cytochrome c also can be replaced without drastically effecting its biosynthesis or function (44). Because Lys-72 is at most only marginally critical for function, the role of methylation may be difficult to assess, especially from in vitro studies, and the biological function of trimethylation remains to be elucidated. The importance of Lys-72 for interaction of cytochrome c with various physiological partners, including cytochrome c 1 , cytochrome c oxidase, cytochrome c peroxidase, and cytochrome b 2 (45-48) suggests that Tml72 may serve a subtle role in one or more of these interactions.
Coordinate Regulation of Cytochrome c and Ctm1p-Liao and Sherman (49) previously reported that the activity of the methyltransferase was lower in extracts of yeast grown under conditions of catabolite repression or anaerobiosis, the same conditions under which cytochrome c is low. Also, during anaerobic to aerobic adaptation, the methyltransferase was induced with cytochrome c, indicating that the synthesis of cytochrome c and the methylase are at least partially coordinately regulated.
We have extended these studies by examining the levels of CTM1 and CYC1 mRNAs in cells grown under the same conditions used by Liao and Sherman (49). The results clearly revealed that the coordinate regulation of cytochrome c and Ctm1p occurs at the transcriptional level (Fig. 6). Even though the two studies were conducted with two different sets of isogenic strains, a comparison of our results with CTM1 and CYC1 mRNAs and the results of Liao and Sherman (49) with methyltransferase activity and iso-1-cytochrome c were remarkably similar (Fig. 7).
Although the mechanism of CTM1 transcriptional regulation remains to be determined, it is reasonable to suggest from upstream sequences that CTM1 contains Hap1p binding sites, functionally equivalent to the sites in the promoter region of CYC1 and other heme-induced genes (Fig. 8).
The high specificity of Ctm1p for methylating cytochrome c and the coordinate regulation of CTM1 and CYC1 further support the view that this methyltransferase evolved to act specifically and efficiently on cytochrome c. Although likely, it remains to be seen whether cytochrome c is the sole substrate of Ctm1p.