p32 Protein, a Splicing Factor 2-associated Protein, Is Localized in Mitochondrial Matrix and Is Functionally Important in Maintaining Oxidative Phosphorylation*

Human p32, originally cloned as a splicing factor 2-associated protein, has been reported to interact with a variety of molecules including human immunodeficiency virus Tat and complement 1q (C1q). p32 protein is supposed to be in the nucleus and on the plasma membrane for the association with human immunodeficiency virus Tat and C1q, respectively. None of the interactions, however, is proven to have a physiological role. To investigate the physiological function of p32, we determined the intracellular localization of p32. The fractionation of cells, fluorescent immunocytochemistry, and electron microscopic immunostaining show that p32 is exclusively localized in the mitochondrial matrix. We cloned a Saccharomyces cerevisiaehomologue of human p32 gene, referred to yeast p30 gene. The yeast p30 protein is also localized in the mitochondrial matrix. The disruption of the p30 gene caused the growth retardation of yeast cells in a glycerol medium but not in a glucose medium, i.e. the impairment of the mitochondrial ATP synthesis. The growth impairment was restored by the introduction of the human p32 cDNA, indicating that p30 is a functional yeast counterpart of human p32. Taken together, both p32 and p30 reside in mitochondrial matrix and play an important role in maintaining mitochondrial oxidative phosphorylation.

Human p32 has been cloned as a splicing factor 2-associated protein from human HeLa cells (1). This protein has been reported to interact with a variety of substances.
The p32 protein has been shown to interact with human immunodeficiency virus (HIV) 1 Tat protein (2)(3)(4) and Rev protein (5, 6) by using in vitro binding studies and two-hybrid analyses in yeast cells. In the cell transfected with the HIV Tat cDNA, the promoter-bound Tat can recruit a p32/VP16 fusion protein to the promoter (3). The transient expression of the murine homologue of p32, YL2, potentiates the function of Rev up to 4-fold (5). Hence, it is presumed that p32 protein might be a co-factor of HIV Tat and Rev proteins. The transcription factor IIB (7) and lamin B receptor (8,9) have also been reported to interact with the p32 protein in vitro.
In addition to these proteins, p32 can interact with globular heads of C1q (gC1q). Because the recombinant p32 protein inhibits C1q hemolytic activity (10,11), p32 is assumed to be a receptor for gC1q. Furthermore, the binding of p32 to H-kininogen (12), to hyaluronic acid (13), and to vitronectin (11) has been reported.
Irrespective of the binding of p32 to many substances, the functional aspects of these associations are not appropriately defined. For example, p32 is neither cross-linked to RNA nor effective to the splice site selection (1), casting a question about the contribution to spliceosome. p32 must be localized in the nucleus to functionally associate with SF2/ASF, HIV Tat, Rev, transcription factor IIB, and lamin B receptor. Although p32 is proposed to be in the nucleus (8), it is not unambiguous because the crude nuclear fraction, a 1,000 ϫ g pellet, is usually contaminated with other fractions of cells (15). To interact with gC1q, H-kininogen, hyaluronic acid, and vitronectin, p32 must exist at the extracellular side of plasma membrane. There is evidence for the presence of p32 in the membrane fraction (13). The reported membrane fraction, which is prepared by 30,000 ϫ g centrifugation of post-nuclear supernatant (13), would contain many organelles including mitochondria which are pelleted by 7,000 ϫ g centrifugation. More recently, most of p32 has been suggested to be localized in cytoplasm but neither in nucleus nor on plasma membrane (16).
In addition to uncertainty of p32 localization, the possibility is raised that p32, which is a highly acidic protein with pI 4.2, nonspecifically interacts with the basic region of nuclear proteins (4).
Thus, despite the many studies on p32, its physiological role remains to be clarified. To elucidate the function of p32, we cloned yeast p30 gene, a yeast counterpart of human p32, and precisely examined the intracellular localization of human p32 and yeast p30 proteins. Here we show that both p32 and p30 proteins are exclusively localized in the mitochondrial matrix and that they are functionally important in maintaining mitochondrial oxidative phosphorylation.

EXPERIMENTAL PROCEDURES
Fractionation of U937 Cells-U937 cells were cultured in RPMI 1640 media (Life Technologies, Inc.) containing 10% fetal bovine serum (Bio-Whittaker, Walkersville, MD), 100 units/ml penicillin, and 100 g/ml streptomycin at 37°C in a 5% CO 2 humidified incubator. After the growth reached 60 -70% confluence, the cells were harvested and homogenized in the homogenizing buffer (10 mM Tris-HCl, pH 7.4, 0.25 M sucrose, and 1 mM EDTA). The subcellular fractions were obtained as described (17). Briefly, the crude mitochondrial fraction (7,000 ϫ g pellet of 700 ϫ g supernatant) was layered on the discontinuous sucrose gradient consisting of successive 4.5 ml of 1.6 M, 1.0 M sucrose from the bottom and centrifuged at 100,000 ϫ g. The band between 1.6 and 1.0 M sucrose was collected for the mitochondrial fraction. The crude nuclear fraction (700 ϫ g pellet) was loaded on the top of discontinuous sucrose gradient made by successive layering of 4.5 ml of 2.0 and 1.6 M sucrose containing 1 mM MgCl 2 and centrifuged at 100,000 ϫ g for 60 min. The pellet at the bottom was collected and used for nuclear fraction.
Fractionation of Mitochondria-The mitochondrial fraction was further separated (17). Three mg of the mitochondrial fraction in 0.5 ml of buffer (10 mM HEPES, pH 7.5, 1 mM EDTA, 250 mM sucrose) was sonicated at output 5.0 for 10 s with an Ultrasonic processor (Tokyorika, Tokyo, Japan) 5 times and centrifuged at 100,000 ϫ g. The resulting pellets and supernatant were used as membrane and soluble fractions, respectively. To disrupt outer membrane, the mitochondrial fraction was suspended in a hypotonic buffer (10 mM HEPES, pH 7.4, 1 mM EDTA) and centrifuged at 14,000 ϫ g for 10 min. The pellets and supernatant were used as mitoplasts and intermembranous fractions, respectively.
Plasmid Construction for p32 cDNA-Human cDNA library (Human HeLa S3 MATCHMAKER cDNA Library, CLONTECH Laboratories, Inc., Palo Alto, CA) was used for amplification of human p32 cDNA.
The DNA fragment corresponding to the mature form of p32-(74 -282) was amplified by PCR using gene-specific primers H74 (5Ј-GAAT-TCTTTCGCGATGCTGCCTCT-3Ј) containing EcoRI site at its 5Ј-end and H282TGA (5Ј-GGATCCGCTCTACTGGCTCTTGAC-3Ј) containing BamHI site at its 5Ј-end and stop codon of p32. The PCR product was subcloned into the vector pCR™II (Invitrogen BV, De Schelp, NV Leek, The Netherlands) to generate pCRHm and then the sequence of the insert was examined to be correct.
The EcoRI-BamHI fragment from pCRHm was inserted into the vector pBluscriptII (Stratagene Ltd., Innovation Center, Cambridge). Then the vector was digested by XhoI and BamHI. The XhoI-BamHI fragment was inserted into XhoI-BamHI sites of the vector pET15b (Novagen, Madison, WI) to generate the expression vector pETHm for producing a mature form of human p32 with His-tag at its N terminus in Escherichia coli.
Vectors for Transfection-The vectors for expressing hemagglutinin (HA)-tagged p32 protein in PLC cells were constructed as follows. First, a vector containing cDNA for HA-tag (pBHA) was made. Oligonucleotide (5Ј-TTGGATCCGTACCCCTATGATGTGCCTGACTATGCTGGTA-GCTCTCATCATCATCATCATCATAGCAGCGGATGACTCGAGTT-3Ј) encoding the 12 amino acids of influenza HA epitope plus six histidine repeat was annealed with its antisense oligonucleotide. The BamHI-XhoI fragment of this DNA was inserted into BamHI-XhoI sites of pBluscriptII to generate pBHA.
Then, the full-length cDNA of human p32 was amplified by PCR from the cDNA library with gene-specific primers H1 (5Ј-GAATTCTTTC-CGCGATGCTGCCTCT-3Ј) containing EcoRI site at its 5Ј-end and H282 (5Ј-GGATCCTGGCTCTTGACAAAACT-3Ј) containing BamHI site at its 5Ј-end. To generate a DNA fragment coding a mature form of p32-(74 -282) without stop codon, a PCR reaction was performed using primers H282 and H73ATG (5Ј-GAATTCATGCTGCACACCGACGG-AGAC-3Ј) harboring the BamHI site at its 5Ј-end and ATG initiation codon. The two kinds of PCR products were cloned into pCR™II to generate pCRH1 and pCRH74, respectively. The sequences were examined to be correct. The EcoRI-BamHI fragments from pCRH1 and pCRH74 were inserted into pBHA (pBHA1 and pBHA74, respectively). Two expression vectors of p32 with HA-tag at the C terminus, pCIHA1 and pCIHA74, were constructed by inserting the XbaI-XhoI fragment from pBHA1 and pBHA74 into the vector pCIneo (Promega, Madison, WI).
Vectors for Yeast Transformation-A PCR reaction was performed with the primer H1 and H282TGA, and the PCR product was subcloned into pCR™II, the sequence was compatible with those previously reported (18). The EcoRI fragment containing p32 full-length cDNA was inserted into EcoRI site of pGAP (19), containing yeast replication 2 origin and glyceraldehyde triphosphates dehydrogenase promoter, to generate the expression vector pGAPH1 of 1-282 amino acid residues of human p32 in yeast cells.
Antibody Preparation-Recombinant human p32 and yeast p30 proteins were obtained by expressing under the control of T7 promoter in E. coli. Briefly, BL21 (DE3) cells harboring pETHm or pGEXY114 (the constructions are described in elsewhere). The cells were grown in 200 ml of LB containing 200 g/ml ampicillin until A 600 of culture medium reached 0.6. Isopropyl-␤-D-thiogalactoside was then added to a final concentration of 1 mM, and the cells were incubated at 37°C for 3 h. The cells were harvested, and the fusion protein was purified according to the manufacturer's protocol. To obtain polyclonal antibodies against human p32 protein and yeast p30 protein, 500 g each of the recombinant proteins expressed in E. coli was emulsified with complete Freund's adjuvant (Difco) and injected into a Japanese white rabbit. Two weeks later, the first booster injection (500 g) was given, followed by three booster injections at 2-week intervals. Sera were obtained 2 weeks after the last booster injection. The His-tag region of recombinant p32 protein was removed with thrombin. The recombinant p32 without His-tag was covalently linked to CNBr-activated Sepharose 4B (Pharmacia Biotech Inc.), which was used to purify the antibodies against p32. The antibodies against glutathione S-transferase (GST) portion in anti-GST-p30 fusion protein antiserum was absorbed in GST-Sepharose 4B.
Immunoblotting Analysis-Proteins were separated on 10 or 7.5% SDS-PAGE and transferred onto a nitrocellulose membrane. After the membrane was incubated in the blocking buffer (phosphate-buffered saline (PBS) containing 5% nonfat milk and 0.1% Tween 20) at room temperature for 2 h or at 4°C overnight, the membrane was then incubated with 0.5 g/ml affinity purified anti-p32 antibody, 5,000-fold diluted anti-p30 antiserum, or 0.2 g/ml anti-60-kDa heat shock protein monoclonal antibody (StressGen Biotechnologies Corp, Victoria, British Columbia, Canada) in the blocking buffer for 1 h and washed with PBS containing 0.1% Tween 20 (PBS-T) for 5 min three times. The membrane was then incubated in PBS-T containing 0.1 g/ml horseradish peroxidase-conjugated anti-rabbit or anti-mouse IgG (Amersham Int., Little Chalfont, Buckinghamshire, UK) for 30 min. The immunoreactive bands were detected using a chemiluminescent substrate as in manufacturer's protocol (Amersham Corp.). The signals was quantified using NIH image program (20,21). The absolute amount of p32 in the mitochondrial fraction and total lysate was determined by comparing signal intensities to those of serial concentrations of purified recombinant p32 protein (6,12,24,48, and 72 ng).
Immunocytochemistry-Indirect immunofluorescence staining of the cells was done as described by Larsson et al. (22) after the cells were fixed and permeabilized with 50% methanol and 50% acetone for 10 min at room temperature and rehydrated. Secondary antibody conjugated with fluorescein isothiocyanate was used against rabbit IgG (American Qualex, San Clemente, CA). Electron microscopic immunocytochemistry of mitochondria of Jurkat cells was done as described previously (17).
Transient Expression of p32 Protein with HA Tag-The pCIneo vectors containing p32 cDNA with HA tag were transfected into PLC cells as described by Kamura et al. (23) with minor modifications. Briefly, 2 g of each plasmid DNA and 4.5 l of cationic liposome (Promega, Madison, WI) were mixed in 200 l of PBS and added to the cells. After incubation for 60 min at 37°C, the cells were resuspended in 2 ml of complete medium containing 10% fetal calf serum and cultured on a 35-mm dish. For immunostaining, a part of the cells were plated onto a four-well tissue culture chamber (Division of Miles Laboratories, Inc,. Naperville, IL). After incubation for 48 h at 37°C, the cells on a four-well tissue culture chamber were fixed and stained. The cells cultured on a 35-mm dish were harvested and immunoblotted with polyclonal antibody raised against HA-tag (Babco, Richmond, CA) or purified anti-p32 antibody.
Data base sequence comparison was performed using FASTA and BLAST program as described (27) Vectors for Yeast p30 -The yeast p30 gene encoding 1-266 amino acids was amplified with a primer UY1 (5Ј-AAGCTTAAACAACAAAT-AATGTTCTT-3Ј) containing HindIII site at its 5Ј-end and a primer LY266 (5Ј-GGATCCAAGTGGAAAAACTTCTTCA-3Ј) containing BamHI site at its 5Ј-end. The PCR was performed on genomic DNA from yeast strain YPH499. The PCR product was subcloned into the vector pCR™II to generate pCRY1. Two independent clones in which the sequences completely matched those deposited in GenBank were used for yeast p30 cDNA. A vector of yeast p30 (pGAPY1) for the protein expression in yeast cells was constructed by inserting EcoRI fragment from pCRY1 into EcoRI site of pGAP.
To construct a cDNA for a GST-yeast p30-(114 -266) fusion protein, a DNA fragment encoding 114 -266 amino acids of p30 was amplified with a gene-specific primer MY (5Ј-GAATTCATGGACGTAGCTCA-GATTGCTAAT-3Ј) containing EcoRI site at its 5Ј-end and the primer LY266. The PCR product was subcloned into pCR™II to generate pCRY114. EcoRI-BamHI fragment from pCRY114 was inserted into pGEX-4TK vector (Pharmacia). The resulting expression vector pGEXY114 was used for producing a GST-yeast p30-(114 -266) fusion protein in E. coli.
Disruption of p30 Gene-Deletion of most of the p30 gene was carried out by a one-step replacement (28). A targeting vector for yeast p30 cDNA was constructed as follows. PCR was performed on genomic DNA from yeast strain YPH500 with primers TU1 (5Ј-GGTTGAAAAGGCT-TCCAGACAAC-3Ј) and TU2 (5Ј-CATATGTGATAGTTTATGTATCTT-3Ј). TU1 corresponds to 592 bp upstream of initiation codon. TU2 corresponds to 24 bp upstream of initiation ATG (A ϭ 0 bp) codon of yeast p30 open reading frame and contains NdeI site at its 5Ј-end. The PCR product was subcloned into pCR™II to generate pCR-U. Another PCR reaction was performed with primers TL1 (5Ј-TGGATCCTTATGT-TGATAGTGCTACTC-3Ј) and TL2 (5Ј-GAGCTCAGCTGCCTGTTGC-CAACTATT-3Ј). TL1 is complementary 261 bp upstream of the termination site and contains the adopter of BamHI site at its 5Ј-end. TL2 is complementary 258 bp downstream of termination TAA (A ϭ 0 bp) codon of yeast p30 open reading frame and contains the adopter of SacI site at its 5Ј-end. The PCR product was subcloned into pCR™II to generate pCR-L. The 1229-bp PvuII-BamHI fragment from the plasmid pGBT9 (CLONTECH Laboratories, Inc., Palo Alto, CA), which contains the TRP1 yeast auxotrophic marker, was inserted into PvuII and BamHI sites of pBR322 to generate pBR-T. The 1460-bp fragment containing TRP1 was excised from pBR-T with NdeI and BamHI. The fragment was inserted into NdeI and BamHI sites of pCR-U to generate pCR-UT. The 520-bp BamHI-SacI fragment from pCR-L was inserted into BamHI and SacI sites of pCR-UT to generate pCR-UTL. The DNA for targeting was liberated from the plasmid by EcoRI-SacI digestion and used to transform the haploid strain YPH500 to generate strain YPH500 (⌬p30::TRP1). To replace TRP1 by HIS3, the NdeI-SalI fragment containing TRP1 of the plasmid pCR-UTL was removed, and the resulting cohesive ends of the plasmid were filled and ligated to each other, and then a BamHI fragment from the vector pYAC3 (29) containing HIS3 was inserted into the BamHI site of the TRP1 genedeleted plasmid to generate pCR-UHL. The EcoRI-SacI fragment from the plasmid was used to transform the haploid strain YPH499 and generate strain YPH499 (⌬p30::His3). The gene disruptions of p30 were verified at the DNA level (30) and by Northern hybridization analysis (31).
Subfractionation of S. cerevisiae-Mitochondrial fraction was isolated from the strain YPH499 as described by McKee and Poyton (32) with minor modifications. Spheroplasts prepared by the treatment of zymolyase 20T (10 mg for 1 g of cells) (Seikagaku, Tokyo, Japan) were homogenized. The homogenate was centrifuged at 700 ϫ g. The resulting post-mitochondrial supernatant was collected and centrifuged at 100,000 ϫ g for 60 min. The supernatant was used as a cytosol fraction, and the pellet was used as a microsomal fraction. Purity of the subcellular fractions was analyzed by the activities of marker enzymes as described under "Fractionation of U937 Cells." The mitochondrial fraction was sonicated and centrifuged at 100,000 ϫ g for 60 min. The resulting supernatant was used as a mitochondrial soluble fraction, and the pellet was used as a mitochondrial membrane fraction. A mitoplast fraction was prepared by osmotic shock as described by Daum et al. (33).
Other Methods-The activity of lactate dehydrogenase was measured by the method of Bergmeyer et al. (34). One unit of the activity was defined as the absorbance change of 1.0/min. The activities of succinatecytochrome c reductase (35), insensitive NADPH-cytochrome c reductase (36), adenylate kinase (35), and fumarase (37) were measured as described. One unit of the activities for adenylate kinase and fumarase was defined as the absorbance change of 1.0/min. Protein concentrations were determined by the method of Lowry et al. (38) with bovine serum albumin as a standard.

RESULTS
Immunocytochemical Detection of p32 Protein-To assess the intracellular localization of p32 protein in cultured cells, we performed immunofluorescence staining using the affinity purified anti-p32 antibody (Fig. 1A). A coarse granular pattern was observed in the cytoplasm of U937 cells and Jurkat cells. In PLC cells, we observed a reticular, lacy cytoplasmic staining pattern. In the three cell lines, nuclei were hardly stained. These patterns of staining are typically observed when mitochondrial components are immunostained (15).
Intracellular Distribution of p32 Protein-To biochemically confirm the mitochondrial localization of p32, U937 cells were fractionated into cytosolic, mitochondrial, and microsomal fractions. The contamination of each fraction with the other fractions was less than 4%, as judged by the specific activities of marker enzymes, i.e. lactate dehydrogenase, succinate-cytochrome c reductase, and rotenone-insensitive NADPH-cytochrome c reductase for cytosol, mitochondria, and microsomes, respectively (Table I).
When the total homogenates were immunoblotted with anti-p32 antibody, a protein of 32 kDa was specifically detected (Fig.  1B, left panel). This band was not visible with preimmune serum (Fig. 1B, right panel). Recombinant p32 protein migrates slower than the endogenous p32 protein due to its additional His-tag (Fig. 1B). p32 was detected in the mitochondrial fraction but not in the cytosolic or microsomal fractions ( Fig. 2A). Compared with the known amount of recombinant p32 Protein in Mitochondria p32, the specific content of p32 in the mitochondrial frction was 1.43 ng/mg protein (Table I). In agreement with the mitochondrial localization of p32, the relative specific content of p32 in each fraction was essentially the same as the relative specific activity of succinate-cytochrome c reductase (Table I). On the basis of recovery of succinate-cytochrome c reductase activity in the mitochondrial fraction, the amount of p32 protein in mitochondria was calculated to account for virtually all (approximately 96%) of the total amount of p32 in cells. The reason why succinate-cytochrome c reductase activity was higher than p32 content in the microsomal fraction (Table I) may be the contamination of the fraction with disrupted mitochondrial inner membranes lacking matrix components.
Localization of p32 Protein in the Mitochondrial Matrix-The sublocalization of p32 in mitochondria was examined. When the mitochondrial fraction was separated into membrane and soluble fractions after sonication, p32 was recovered from the soluble fraction exactly in parallel with the activity of fumarase that resides in the mitochondrial matrix (Fig. 2B) (Table II). On the basis of recovery of the fumarase activity in the mitochondrial soluble fraction, the content of p32 in the soluble fraction accounted for about 96% of p32 in intact mitochondria.
To determine whether mitochondrial p32 is localized in the intermembranous space or in the matrix, we preferentially disrupted outer membranes of mitochondria by hypotonic treatment and obtained mitoplast and intermembranous fractions. p32 as well as the activities of succinate-cytochrome c reductase and fumarase was recovered from the mitoplast fraction with precisely the same degree of enrichment (Table II). p32 was negligible in the intermembranous space where the activity of adenylate kinase, a marker enzyme of the space, was enriched (Fig. 2C). On the basis of recovery of fumarase activity in the mitoplast fraction, the amount of p32 in mitoplasts accounted for about 102% p32 in intact mitochondria (Table II). These results indicate that p32 is exclusively localized in the mitochondrial matrix.
The localization of p32 in the matrix was further confirmed by electron microscopic immunostaining of mitochondria isolated from Jurkat cells by using anti-p32 antibody (Fig. 3). The signals were seen in the matrix (electron dense area). When non-immune serum was used, signals were not observed (data not shown).
p32 Protein in the Nucleus-We prepared the crude nuclear fraction and post-nuclear supernatant by centrifugation of the total cell lysate at 700 ϫ g. The resulting crude nuclear fraction was further purified by centrifuging the fraction on the discontinuous sucrose gradient bed. p32 was recovered from the post-   p32 Protein in Mitochondria nuclear supernatant (Table III). The recovered amount of p32 protein in the post-nuclear supernatant accounted for 107% p32 content in the total lysate according to the recovery of succinate-cytochrome c reductase activity in the post-nuclear supernatant. The relative specific content of p32 in the nuclear fraction was essentially the same as the relative specific activ-ity of succinate-cytochrome c reductase in the nuclear fraction. Transient Expression of Premature and Mature p32-It is known that the mature form of p32 lacks the N-terminal 73 amino acids compared with the full-length protein deduced from the cDNA. Considering the possibility that the N-terminal region is a signal sequence for the import to the mitochondrial matrix and processed there, we have expressed the premature and mature forms of p32 with HA-tag at their C termini in PLC cells.
In the cells transfected with the full-length cDNA, the mature form of 34-kDa protein was found to be much stronger than the premature form of 38-kDa protein both with anti-HA and anti-p32 antibodies (Fig. 4A, lane 2 in left and right panels), indicating that the HA-tagged premature p32 is efficiently processed to the mature form in the cells. Only a 34-kDa protein was detected in the cells transfected with the mature form of cDNA (Fig. 4A, lane 3 in left and right panels). Endogenous p32 protein (32 kDa) was detected in all the three cell lines with anti-p32 antibody (Fig. 4A, right panel).
In the cells transfected with the full-length cDNA, HAtagged p32 distributed granularly in the cytoplasm but not in the nucleus (Fig. 4B, left panel) as observed for endogenous p32 in Fig. 1A, indicating its mitochondrial localization. In contrast, the cytoplasm and, to a lesser extent, nucleus were homogeneously stained in the cells transfected with the mature form cDNA (Fig. 4B, right panel), irrespective of both cell lines predominantly containing the mature 34-kDa protein (Fig. 4A). These results suggest that the N-terminal region to be processed is necessary for p32 to be imported to mitochondria.
Intracellular Localization of Yeast p30 Protein-We have cloned a yeast gene homologous to human p32 gene, referred to yeast p30. The amino acid sequence of yeast p30 and human p32, which shows 60% similarity and 30% identity over all, is highly conserved particularly in their C termini (Fig. 5B). The human p32 and yeast p30 proteins shared well conserved structural features, i.e. (i) diffuse distribution of acidic residues with intercalated non-charged portions, (ii) small regions with positive charges at the N-and C termini, and (iii) a region with a periodic repeat of leucine residues (Fig. 5A).
When the total homogenate of wild type yeast cells was immunoblotted with anti-p30 antibody, the antibody specifically probed a protein with a molecular mass of 30.5 kDa (Fig.  6A, lane 1 in upper panel). Preimmune serum could not detect this 30.5-kDa band (data not shown). Total homogenates of wild type cells were separated into mitochondrial, cytosolic, and microsomal fractions. The contamination with mitochon- FIG. 4. Immunofluorescence staining and immunoblot analysis of transiently transfected cells. A, PLC cells were transfected with pCIHA1 encoding 1-282 amino acids of p32 (p32/1-282) or pCIHA74 encoding 74 -282 amino acids of p32 (p32/74 -282) as described under "Experimental Procedures." The cells were suspended in PBS and sonicated. Whole cell lysates (100 g/lane) were separated on 10% SDS-PAGE and immunoblotted using antibodies specific for HAtag (anti-HA) and p32 (anti-p32). B, the cells were fixed and stained with antibodies specific for HA-tag (original magnification, 400 ϫ). 5. Comparison of the amino acid sequences. A, amino acid sequences of human p32 and S. cerevisiae p30 are depicted schematically with boxed regions as indicated. B, comparison of the C-terminal amino acid of p32 from humans, YL2 protein (5) from mice, yeast p30 from S. cerevisiae, and F59A2.3 (similar to splicing factor-associated 32K chain) from Caenorhabditis elegans. Key: identity (͉), high similarity (:), similarity (⅐).

TABLE III p32 in nuclear and post nuclear fractions of U937 cells
The specific content of p32 and the specific activity of succinatecytochrome c reductase are expressed as percent of those in the mitochondrial fraction. Each value is a mean of two independent experiments. Nucleus, nuclear fraction; post-nuclear, post-nuclear supernatant fraction. p32 Protein in Mitochondria dria was less than 1% both in the cytosolic and microsomal fractions based on the specific activity of succinate-cytochrome c reductase (Table IV). The mitochondrial heat shock protein 60 (p66) was also detected only in the mitochondrial fraction (Fig.  6A, lane 1 in lower panel), further indicating that the fractions are well separated. When these fractions were immunoblotted with anti-p30 antibody, p30 was only detected in the mitochondrial fraction (Fig. 6A, upper panel). When the mitochondrial fraction was subfractionated (Table V), p30 was detected in the soluble fraction (Fig. 6B, left panel) and the mitoplast fraction (Fig. 6B, right panel) as observed in the case of human p32. The finding indicates that yeast p30 is also localized in the mitochondrial matrix.
Gene Targeting of p30 -To investigate the role of p30 in mitochondria, we disrupted the yeast p30 gene. When the total homogenate of the ⌬p30 strain (YPH499(⌬p30::HIS3)) was immunoblotted with anti-p30 antibody, the signal for p30 was not visible (Fig. 6C, upper panel), whereas the mitochondrial heat shock protein 60 (p66) was detected both in the wild type and in the ⌬p30 strain to a similar extent (Fig. 6C, lower panel). When the plasmid containing p30 cDNA was introduced into the ⌬p30 strain, the expression of p30 protein was restored (Fig. 6D, lane  2 in middle panel). The ⌬p30 strain grew normally in a medium containing glucose (YPDagar), slowly in the glycerol medium (YPGagar) at 30°C (data not shown), and very slowly at 18°C (Fig. 7A, ⌬p30), suggesting that the ⌬p30 strain has an abnor- ). An equal amount of each fraction was separated on 7.5% SDS-PAGE and immunoblotted with anti-HSP60 monoclonal antibody (anti-HSP60). The relative positions of yeast p30 (p30) and yeast HSP60 (p66) (42,43) are shown, respectively. B, mitochondrial fraction (mt) prepared from YPH499 was sonicated and separated into soluble fraction (mt soluble) and membrane fraction (mt membrane). Fifteen g of protein were separated on 10% SDS-PAGE and immunoblotted with anti-p30 antiserum (left panel). The mitochondria (mt) were incubated in the hypotonic solution and separated into intermembranous fraction (intermembrane) and mitoplast fraction (mitoplast), which were separated on 10% SDS-PAGE and immunoblotted with anti-p30 antiserum (right panel). C, total homogenates from wild type and the ⌬p30 strains were prepared and separated on 10% SDS-PAGE. Proteins were transferred to nitrocellulose membrane and immunoblotted by the immune serum for p30 (anti-p30). Equal amount of proteins was separated on 7.5% SDS-PAGE and immunoblotted with anti-HSP60 monoclonal antibody (anti-HSP60). The relative positions of yeast p30 (p30) and yeast HSP60 (p66) are shown. D, mitochondrial fraction prepared from U937 cells and total homogenates of the indicated strains were separated on 10% SDS-PAGE and immunoblotted with purified anti-p32 antibody (anti-p32) or anti-p30 antiserum (anti-p30). Equal amount of each protein was separated onto 7.5% SDS-PAGE and immunoblotted with anti-HSP60 monoclonal antibody (anti-HSP60). The relative positions of yeast p30 (p30), human p32 (p32), human HSP60 (p58) (44), and yeast HSP60 (p66) are shown, respectively. mality in maintaining mitochondrial ATP synthesis. The growth retardation of the ⌬p30 strain was partially restored by the introduction of p30 cDNA (Fig. 7A, ⌬p30 ϩ pGAPY1) but not by vector alone (Fig. 7A, ⌬p30 ϩ pGAP). The ⌬p30 haploid strain was crossed with another ⌬p30 haploid strain or with wild type. The growth rate of the ⌬p30/⌬p30 homozygote in glycerol medium was much slower than that of the wild type/ ⌬p30 heterozygote or wild type/wild type homozygote at 18°C (Fig. 7B). Thus, yeast p30 deletion mutant showed cold sensitivity in the nonfermentable carbon source medium.
We examined whether the introduction of human p32 cDNA could complement the growth retardation in the glycerol medium caused by the disruption of p30 gene. The ⌬p30 strain grew slowly in the glycerol medium (YPG) at 30°C, and the growth retardation of the ⌬p30 strain was partially restored by the introduction of p30 cDNA (Fig. 7C). It is not clear at present why the yeast p30 cDNA could not completely restore the growth rate of the deletion mutant, but one possibility is that the plasmid and the genome are different in the transcriptional control. In yeast cells, human p32 was processed to the same size as in human cells (Fig. 6D, lane 3 in upper panel). Human p32 restored the growth rate at the same degree as yeast p30 (Fig. 7C). DISCUSSION We have shown in the present study that p32 is localized in the mitochondrial matrix in a soluble form. p32 was not detected in membrane-associated fractions in our experiment. It seems unlikely that p32 works on plasma membrane as a receptor for C1q, H-kininogen, and hyaluronic acid. The small amount of p32 present in the nuclear fraction (Table III) may be contamination with mitochondria, because (i) the specific content of p32 in the nuclear fraction was almost the same as the specific activity of succinate-cytochrome c reductase in the nuclear fraction, (ii) nuclei were barely stained in immunocytochemistry, and (iii) p32 protein in the nuclear fraction was the mitochondrially processed mature form. p32 protein has been assumed to be biosynthesized as a proprotein of 282 amino acids and post-transcriptionally processed to the mature protein of 209 amino acids (fragment 74 -282) (18). The processing mechanisms have been unknown. Most mitochondrial proteins are synthesized with an N-terminal signal sequence, which targets these proteins to mitochon-   7. Growth properties of the yeast strains. A and B, growth properties of the indicated yeast cells on glucose (YPDagar) and glycerol (YPGagar) media are shown. Each plate was incubated at 18°C for 10 days. WT, YPH499; WT/WT, the diploid strain of YPH499/YPH500; WT/⌬p30, the diploid strain of YPH500/YPH499 (⌬p30::HIS3); ⌬p30/ ⌬p30, the diploid strain of YPH500 (⌬p30::TRP1)/YPH499 (⌬p30::HIS3). C, logarithmic plot of A 600 versus time in hours for the indicated strains grown in the glycerol medium (YPG) at 30°C. Cells were first grown in the synthetic medium with glucose at 30°C and then diluted into complete medium containing 3% glycerol. At the indicated times, samples were removed and their A 600 values were determined. ϫ, wild type strain (YPH499) carrying plasmid alone (pGAP); q, ⌬p30 strain (YPH499 (⌬p30::HIS3)) carrying plasmid alone; E, ⌬p30 strain carrying the yeast p30 plasmid (pGAPY1); Ⅺ, ⌬p30 strain carrying the human p32 plasmid (pGAPH1). dria and is excised in mitochondria. A program, PSORT, predicts that the N-terminal region of premature p32 has a structure typical to mitochondrial signal peptide, i.e. amphipathic helix with basic residues. Here we show that the N-terminal region is required for the import of p32 into mitochondria and is efficiently processed (Fig. 4).
We demonstrated that yeast p30 is localized in the mitochondrial matrix, and the disruption of the gene causes the slow growth in glycerol medium, but the slow growth is not lethal. The cold sensitivity on the glycerol medium as was seen in the p30 disrupted strain can be caused by abnormality of several genes. The two different mutations in the oligomycin sensitivity conferring protein decrease the activity of ATP synthase below a threshold required for the growth of yeast at 18°C (39). SSH1 (HSP70), a member of the heat shock protein family, is required for normal mitochondrial DNA replication (40). The impaired replication in ssh1 cells results in a defect of the ATP synthesis at a low temperature. Similarly, the yeast p30 gene may encode a protein that modulates the oxidative phosphorylation directly or indirectly.
The growth impairment in the glycerol medium of the p30 gene-disrupted strain was restored by the introduction of the human p32 gene to exactly the same extent as that of the yeast p30 gene. The p32 protein may play a physiologically important role in human mitochondria.
There is a report that the peripheral T lymphocytes from HIV1 carriers show dysfunction of mitochondria (41). The interaction of p32 protein with HIV Tat might affect mitochondrial functions. Although many proteins have been supposed to interact with the human p32 protein in nucleus or on plasma membrane, the interactions with p32 should be carefully reexamined from the view of its exclusive localization in mitochondria.