Purification and Complete Sequence of a Small Proteolipid Associated with the Plasma Membrane H+-ATPase of Saccharomyces cereuisiae*

The purified plasma membrane H’-ATPase of Schi-zosaccharomyces pombe and Saccharomyces cerevisiae display, in addition to the catalytic subunit of 100 kDa, a highly mobile component, soluble in chloroform/ methanol. Chloroform/methanol extraction of S. cerevisiae plasma membranes led to isolation of a low molecular weight proteolipid identical to that present in purified H+-ATPase. NHz-terminal amino acid se- quencing revealed a 38-residue polypeptide with a cal-culated molecular mass of 4250 Da. The polypeptide lacks the first two NH2-terminal amino acids as com- pared with the deduced sequence of the PMPl gene (for plasma membrane proteolipid) isolated by hybrid-ization with an oligonucleotide probe corresponding to an internal amino acid sequence of the proteolipid. The polypeptide is predicted to contain an NHz-terminal transmembrane segment followed by a very basic hy- drophilic domain. The plasma character-ized protons energizing the for and

The plasma membrane of fungi contains a well characterized ATPase that pumps protons out of the cell, energizing the membrane for nutrient uptake. The fungal H+-ATPase is structurally and functionally related to the Na+/K+-, Ca2+-, and H+/K+-ATPases of animal cell membranes as well as to the H+-ATPase of plant-cell plasma membranes. It contains a 100-kDa catalytic subunit whose organization within the membrane is similar to that of the other members of the cation-translocating ATPase family. Particularly striking is the presence of conserved stretches which include regions involved in ATP hydrolysis and formation of the &aspartylphosphate intermediate (see recent reviews by Serrano (1988), Nakamoto and Slayman (1989), and Goffeau and Green (1990)).
All the biochemical and genetic work carried out so far has been based on the implicit assumption that the fungal H+-ATPase contains no additional subunit. Upon classical Laem-*This work was supported by grants from the Services de la Politique Scientifique: Action Science de la Vie and the Fonds National de la Recherche Scientifique, Belgium. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted M77845.
to the GenBankTM/EMBL Data Bank with accession number(s) $Recipient of a fellowship from the Fonds National de la Recherche Scientifique.
Supported by a postdoctoral fellowship from the commission of the European Communities.
ll To whom correspondence should be addressed. Tel.:  mli gel electrophoresis, the purified enzyme displays a highly enriched Coomassie-stained band with a molecular mass of 100 kDa. It has been reported, however, that the Schizosaccharomyces pombe preparation contains a component moving slightly ahead of the tracking dye. This highly mobile component fails to stain with Coomassie Blue but develops a red color with periodic acid-Schiff stain, suggesting the presence of either glycoproteic or proteolipidic material (Dufour and Goffeau, 1978). The existence of a ninhydrin-positive component in the lipidic chloroform/methanol extract of the purified S. pombe H+-ATPase was later reported, but no indication was given as to the nature of this material and its relation to the fast moving compound observed during electrophoresis (Dufour and Goffeau, 1980).
In the present study, the possibility that the fungal H+-ATPase might possess a proteolipid component of high electrophoretic mobility was reexamined using novel electrophoresis conditions adapted to separating low molecular weight proteins (Schagger and von Jagow, 1987). Indeed, the purified H+-ATPase of S. pombe and Saccharomyces cerevisiue showed two low molecular weight, Coomassie-stained bands after electrophoresis on a Tricinel-SDS-polyacrylamide gel. Chloroform/methanol extraction of these polypeptides from S. cereuisiue plasma membranes established the lipid-like properties of these highly mobile compounds. Microsequencing of the purified proteins revealed the existence of a single, small 38-residue polypeptide in both bands. This polypeptide was extracted from purified H+-ATPase as well as from plasma membranes. The corresponding gene was isolated by hybridization with an oligonucleotide probe derived from a portion of the amino acid sequence. Gene sequencing confirmed the amino acid sequence with the exception of two additional amino-terminal residues.

6425
Purification and Sequence of a 5' . cerevisiae Proteolipid ml protein) were suspended in 25 volumes of a mixture of chloroform/ methanol (2:l) for 2 h a t 4 'C. Undissolved material was removed by filtration and the extract was washed with 0.2 volume of water. After centrifugation of the milky emulsion a t 2,000 X g for 30 min, the lower organic phase was concentrated to dryness under reduced pressure in a flash evaporator. The final residue was dissolved in a minimal volume of chloroform. Four volumes of diethyl ether were added to the chloroform extract, and the mixture was stored for 1 h at -20 "C. Precipitated proteins were collected by centrifugation.
Protein concentrations were determined by the method of Lowry et a/. (1951) after dilution of the samples in 0.4% sodium deoxycholate.
In order to detect very small polypeptides, Tricine-SDS-polyacrylamide gel electrophoresis was carried out as described by Schagger and von Jagow (1987). Proteins were fixed by a solution containing formaldehyde (Biirk et a/., 1983) and stained with 0.025% Serva blue G in 10% acetic acid.
Peptide Sequencing-Ether-precipitated proteolipid (60 pg) was dissolved in 25 pl of trifluoroacetic acid and diluted to 100 pl with water. After centrifugation a t 13,000 rpm for 5 min, the supernatant was directly loaded on a precycled Polybrene-coated glass filter. The amino acid sequence was determined using an Applied Riosystem sequenat.or (model 477A) equipped with an on-line I'TH-derivative analyzer (model 120A).
Reaction of Amino Croups with Phenyl Isocyanate and Cleavage with HNPS-skatole-Aliquots containing 108 pg of ether-precipitated proteolipid (30 nmol) were treated with phenyl isocyanate to block the NH, terminus (Fearnley et al., 1990). The modified protein was precipitated with ether, collected by centrifugation, and solubilized in 50% acetic acid. RNPS-skatole was added to give a final molar excess of 100 over tryptophan residues, and the mixture was incuhated for 24 h a t 25 "C. These conditions cleaved the polypeptide after tryptophan 28 (Fontana, 1972). The sample was then lyophilized and sequenced as described above.
Carboxyl End Determination-Ether-precipitated proteolipid (88 pg, 20 nmol) was suspended in 10 p1 of 1% SDS and incubated at 60 "C for 20 min. The concentration of SDS was then diluted to 0.05% by addition of 190 pl of 50 mM sodium citrate, pH 6.0. A 25-pI control aliquot was withdrawn, and the rest of the solution was incubated a t 25 "C with 1 pg of carboxypeptidase Y. A t various times from 30 s to 10 min, 2 5 p l aliquots were withdrawn and the reaction was stopped with 5 pl of acetic acid. The samples were lyophilized and soluhilized in 100 pl of 10 p M internal standard solution (norleucine). A 2O-pl aliquot of the sample was injected into a Waters amino acid analysis column of a Perkin Elmer 3H analyzer using o-phthalaldehyde as derivatization reagent (Dufour, 1986).  et al., 1988). Several clones hybridized with the probe under low stringency conditions (Ghislain et al., 1990). Double-stranded clones were sequenced by the dideoxy method (Sanger et al., 1977) with T 7 DNA polymerase (Tabor and Richardson, 1987), using the oligonucleotide prohe as a primer. A second oligonucleotide, complementary to the 5"flanking sequence of the proteolipid coding region, was used to determine the complete nucleotide sequence of the gene.

Electrophoretical Patterns of Purified H'-ATPase
and Chloroform/Methanol Plasma Membrane Extract-S. cerevisiae and S. pombe plasma membrane H+-ATPase, solubilized by lysolecithin and purified by centrifugation through a sucrose gradient Dufour et al., 1988), were subjected to a Tricine-SDS-gel electrophoresis (Fig. L4, lanes   3 and 4 ) . Both preparations display, in addition to the catalytic subunit of 100 kDa, two diffuse bands with apparent molecular mass lower than 10 kDa. Purified H+-ATPase from S. cerevisiae was treated with chloroform/methanol. This treatment extracted specifically the two highly mobile compounds which amounted to about 10% of total ATPase protein.
A more convenient extraction of the low molecular weight component was obtained from Saccharomyces cerevisae plasma membranes which yielded about 1.5% of total membrane protein by chloroform/methanol. The chloroform/ methanol-soluble fraction was subjected to a Tricine-SDS-gel electrophoresis (Fig. lR, lane 4 ) . This chloroform/methanol extract contains the two diffuse bands contaminating the purified H+-ATPase; their apparent molecular weights were of 7,500 and 4,000. Addition of diethyl ether to the chloroform extract solubilized most of the 4-kDa band (Fig. l R , lane 6), as well as free lipids including phospholipids, monoglvcerides, triglycerides, sterols, fatty acids, and sterol esters (Fig. 2 H ) . Consistently, the lower part of the 4-kDa band comigrates with pure lecithin (data not shown). In contrast, the 7.5-kDa band was precipitated with ether (Fig. lB, lane 5). No free lipids were detected in this ether-insoluble fraction (Fig. 2C).
Amino Acid Sequence of the Purified Polypeptide-NHzterminal amino acid sequencing was carried out on the etherprecipitated fraction which is completely devoid of free lipids. A 38-residue sequence was unambiguously determined except for the residue at position 16. This residue, which was suspected to be a serine since breakdown products of serinephenylthiohydantoin were detected, is produced in markedly low yield (Fig. 3C). The low yield of tryptophan 28 (Fig. 3C) is typical of this residue. The progressively decreasing yields from glutamine 35 to phenylalanine 38 (Fig. 3C) suggest that the COOH end of the polypeptide was reached. The protein was treated with phenyl isocyanate to block the NH, end and subsequently cleaved with BNPS-skatole, which cleaves after tryptophan residues (Fontana, 1972). Only one peptide was generated, the sequence of which corresponds exactly to that expected from cleavage after tryptophan 28 (Fig. 3A). As for the intact polypeptide, no further residue was obtained by Edman degradation after residue 38. Carboxypeptidase Y liberated only phenylalanine within the first 10 min, which is consistent with the assignment of phenylalanine 38 as the COOH-terminal residue. It therefore seems likely that the complete amino acid sequence of the polypeptide was determined.
Ether-soluble and -insoluble fractions of the chloroform extract from S. cereuisiae plasma membranes were found to be identical in NH2-terminal amino acid sequence (Fig. 3, A  and B ) . These results indicate that both fractions contain the same polypeptide. It therefore seems likely that the 7.5-kDa compound is a dimeric form of the 4-kDa polypeptide. The dimer would be precipitated with ether because of sponta-

10 15
neous association and mutual protection of the hydrophobic portions of the polypeptides while their hydrophilic portions are exposed to the outside. The size discrepancy between the two bands probably reflects the presence of lipids as well as anomalies in migration of low molecular weight peptides as previously reported (Huang and Mathews, 1990). Furthermore, diethyl ether was added to the chloroform extract of purified H'-ATPase from S. cereuisiae. The insoluble fraction was subjected to Edman degradation. The first 15 amino acid residues of the NH2 terminus correspond exactly with those determined for the proteolipid extracted from plasma membrane (data not shown).
The corresponding gene referred to as PMPl (for plasma membrane proteolipid) was isolated and sequenced. As compared with the deduced amino acid sequence, the purified polypeptide lacks the first two residues, found to be methionine and threonine (Fig. 4). A cysteine residue was found at position 18 of the nascent polypeptide. The presence of cysteine, undetectable by Edman degradation without pretreatment, easily explains the very low yield observed during amino acid microsequencing for position 16 of the mature polypeptide, which was supposed to be a serine. Searches for homologous proteins in current protein data banks, using the FASTA algorithm (Pearson and Lipman, 1988) failed to uncover any significant primary structural similarity to known proteins.
Hydropathy Analysis of the Mature Polypeptide-A hydropathy plot was derived for the 38-residue polypeptide according to Kyte and Doolittle (1982) (Fig. 5). Two structural domains  thy analysis was carried out using the algorithm of Kyte and Doolittle (1982) with a window size of 7 residues. A 20-residue region with a hydrophobicity index 21.6 is predicted to span the membrane.
Purification and Sequence of a S. cerevisiae Proteolipid are clearly defined. The NHz terminus (residues 1-24) is predicted to span the membrane. In contrast, the hydrophilic COOH end (residues 25-38) forms a very basic domain with 1 lysine and 4 arginine residues.

DISCUSSION
We have purified a small protein from the S. cereuisiae plasma membrane as well as from the purified H+-ATPase. This protein can be classified as a proteolipid since it is extracted by a mixture of chloroform/methanol and thus has lipid-like properties (Folch and Lees, 1951). When electrophoresed on Tricine-SDS-polyacrylamide gels, this proteolipid displays two bands. The molecular weights of the corresponding proteins are 7,500 and 4,000. The same NHz-terminal sequence (94% homogeneous) was determined by Edman degradation during microsequencing of these two bands, suggesting that the extracted proteolipid exists as a dimer and a monomer.
The proteolipid is firmly bound to the H+-ATPase of the S. cereuisiae plasma membrane and cannot be removed by any of the attempted purification procedures preserving ATPase enzyme activity. It takes a denaturing chloroform/methanol extraction to dissociate the proteolipid from the major 100-kDa subunit of the H+-ATPase. As pointed out in the introduction, the H+-ATPase purified from the fission yeast S. pombe seems to contain a low molecular weight compound with a mobility and lipid-like properties similar to those described here for S. cerevisiae (Dufour and Goffeau, 1978).
Moreover, previous quantitative studies on the subunit composition of the Neurospora crussa membrane H+-ATPase detected a very small oligopeptide that migrates in the tracking dye region (Scarborough and Addison, 1984). These authors could not exclude the possibility that this compound is a stoichiometric subunit of the H+-ATPase. Thus, the 4-kDa proteolipid might be a true component of the fungal H+-ATPase. It has also been reported that the mammalian Na+/ K+-ATPase has a 58-amino acid proteolipidic subunit, whose function remains unclear (Collins and Leszyk, 1987). The proteolipid described in this study might be a yeast analog to phospholamban, a small, 52-residue peptide regulating the sarcoplasmic reticulum Ca2+-ATPase (Tada and Katz, 1982). Particularly striking is the presence of single transmembrane span and highly charged hydrophilic domains predicted for phospholamban, Na+/K+ proteolipid and the proteolipid presented in this study (Mercer et al., 1990;Simmerman et al., 1986).
The yeast plasma membrane proteolipid undergoes some posttranslational modifications; as compared with the deduced sequence of the gene, the proteolipid lacks the first two NHz-terminal residues, suggesting that integration into the plasma membrane is achieved with minimal cleavage. Moreover, during NHz-terminal amino acid sequencing of the mature polypeptide, alanine 21 and threonine 22 were found to occur with a lower yield than most of the other residues. This raises the possibility that threonine might be modified, which could disturb the Edman degradation of the two neighboring amino acids. It is conceivable that a fatty acid binds threonine through an ester linkage, as reported for the myelin proteolipid (Stoffel et al., 1983). Alternatively, a palmitate might be covalently bound to the proteolipid via the cysteine in position 16 (in the mature polypeptide) as is generally the case for fatty acid-acylated proteins (see review by Olson (1988)). Such a covalent modification would probably account for the solubility of the small proteolipid in chloroform/methanol.