Cloning, expression, and localization of 230-kDa phosphatidylinositol 4-kinase.

A phosphatidylinositol (PI) 4-kinase cDNA was cloned from a rat brain cDNA library. This cDNA encoded a protein of 2041 amino acids with a calculated molecular weight of 231,317. The deduced amino acid sequence shared the identity of 52.3 and 34.4% in the presumed catalytic domain with two yeast PI 4-kinases, STT4 and PIK1, respectively, and showed 31.7% identity to p110α subunit of rat PI 3-kinase in the same domain. In addition, a 3′ half coding region of the present cDNA was 89.6% identical to and its deduced amino acid sequence was 98.2% identical to the sequence for PI4Kα, a recently reported human PI 4-kinase of type II, suggesting that PI4Kα is an alternative form of the present PI 4-kinase molecule. The present cDNA contained sequences encoding the ankyrin repeat domain, lipid kinase unique domain, pleckstrin homology domain, presumed lipid kinase/protein kinase homology domain, proline-rich region, and SH3 domain. By examining PI kinase activity in transfected COS-7 cells using the epitope tag immunoprecipitation as well as the conventional way, the product phosphatidylinositol phosphate was identified as phosphatidylinositol 4-phosphate but not phosphatidylinositol 3-phosphate. This PI 4-kinase activity was markedly enhanced in the presence of Triton X-100 but relatively insensitive to inhibition by adenosine. By epitope tag immunohistochemistry, the immunoreactivity for this PI 4-kinase molecule was largely localized in close association with the membranes of the Golgi vesicles and vacuoles. By in situ hybridization analysis, the expression of mRNA for this PI 4-kinase was evident throughout the gray matter of entire brain with higher expression intensity in fetal brain. These data imply that this novel PI 4-kinase is involved in some processes essential to neuronal differentiation and maturation including the synaptogenesis and synaptic plasticity.

Several forms of mammalian PI 4-kinases have been purified and characterized from different tissues (6 -9) and classified into two types, type II and type III, on the basis of differences in sensitivity to adenosine and nonionic detergent. The type II enzyme has an apparent molecular mass of approximate 55 kDa, is inhibited by adenosine, and is stimulated by nonionic detergent, whereas the type III enzyme has a much larger apparent molecular mass of Ͼ 200 kDa, is resistant to inhibition by adenosine, and is stimulated by nonionic detergent.
Recent molecular cloning studies have identified two different cDNAs encoding yeast PI 4-kinases, PIK1 (10) and STT4 (11). On the assumption that there is a significant similarity in sequences between yeast PI 4-kinases and mammalian counterparts, we have undertaken the isolation and sequencing of cDNA encoding mammalian PI 4-kinase. During this attempt, Wang and Cantley (12) have cloned a cDNA from human placenta and fetal brain cDNA libraries by the polymerase chain reaction using primers encompassing partial human cDNA sequences from the unidentified gene Humxt010191, which encodes amino acids similar to PI 3-kinase p110 subunit. They have reported that the encoded protein termed PIK␣ has enzymatic properties of type II PI 4-kinase in their transfection experiment in insect cells (12). However, the predicted molecular mass, which was calculated to be 97 kDa in their study, is inconsistent with the size of major type II PI 4-kinase previously purified from human red cell (6), bovine brain (7), and uterus (8). In addition, there have so far been no reports on the molecular structure of mammalian type III PI 4-kinase. In this paper we report for the first time the molecular cloning of 230 kDa of protein from rat brain, which shows the enzymatic properties of type III PI 4-kinase, appears to be associated with the membrane, and localizes widely in the gray matter of the brain, especially dominantly in the fetal brain.

MATERIALS AND METHODS
Polymerase Chain Reaction-Total RNA was extracted from adult rat brain by guanidine thiocyanate/phenol/chloroform extraction (13). Poly(A) RNA was isolated by chromatography on an oligo(dT) cellulose column. First strand cDNA was prepared using First-Strand cDNA Synthesis kit (Pharmacia Biotech Inc.). The primers used for PCR were composed of two degenerate oligonucleotides that were made according to the amino acid sequences of conserved regions between putative lipid kinase domains in yeast STT4 and PIK1; the regions corresponded to the amino acid sequences (V/T)GDD(C/L)RQ (residues 1647-1654, 5Ј primer) and HIDFGF(I/M) (residues 1770 -1776, 3Ј primer) (the nucle-* This study was supported by Grants-in-aid for Scientific Research 07044216, 07457001, and 07278203 (to H. K.) and 07780670 (to K. G.) from the Ministry of Education, Science, and Culture of Japan and grants from Asaoka Eye Clinic Foundation, Hamamatsu, Japan and from Japan Chemical Research Pharmaceutical Co. Ashiya, Japan. 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 to the GenBank TM /EMBL Data Bank with accession number(s) D83538.
‡ To whom correspondence should be addressed: Dept. of Anatomy, Tohoku University School of Medicine, Seiryo-cho 2-1, Sendai 980, Japan. Tel.: 81-22-273-9029; Fax: 81-22-272-7273. 1 The abbreviations used are: PI, phosphatidylinositol; PI(4)P, phosphatidylinositol 4-phosphate; PI(3)P, phosphatidylinositol 3-phosphate; PIP, phosphatidylinositol phosphate; PCR, polymerase chain reaction; kb, kilobase pair(s); RACE, rapid amplification of cDNA ends. otide numbers represent those of STT4). The sequences for the primers were designed according to the mammalian codon usage: CGGAATTC-CGG(A/T/C)GA(T/C)GA(T/C)GA(T/C)T(G/T)(T/C)CG(G/C)CA(G/A)GA for 5Ј primer and CGGAATTAT(G/A)AA(G/A/T)CC(G/A)AA(G/A)TC(G/ A)AT(G/A)TG for 3Ј primer. The 5Ј ends of the 5Ј and 3Ј primers contained a EcoRI restriction sequence for the subsequent cleavage of cloned cDNA fragments. PCR amplification was performed by using the AmpliTaq DNA polymerase according to the following schedule: 94°C for 30 s, 45°C for 1 min, 72°C for 2 min for 35 cycles, followed by further incubation for 7 min at 72°C. PCR products were electrophoresed on a polyacrylamide gel, and an amplified DNA (about 400 base pairs) was excised from the gel and subcloned into pUC118 after digestion with EcoRI. Sequence analysis revealed that 4 of 20 clones (pMP1, 2, 12, and 18) were identical and showed homology with STT4 and PIK1.
cDNA Cloning-A rat brain cDNA library was constructed as described previously (14). Clones (3 ϫ 10 6 ) derived from the cDNA library were screened by hybridization with 382-base pair cDNA fragment of pMP1. Hybridization was carried out at 42°C in a buffer containing 50% formamide, 5 ϫ SSC (saline sodium citrate), 1 ϫ Denhardt's (0.02% Ficoll 400, 0.02% polyvinylpyrrolidone, 0.02% bovine serum albumin), 50 mM sodium phosphate (pH 7.2), and 250 g/ml heat-denatured salmon sperm DNA for 16 h. The membranes were washed twice at room temperature in 2 ϫ SSC, 0.1% SDS for 10 min, followed by two washes in 0.1 ϫ SSC, 0.1% SDS at 42°C for 30 min and finally at 55°C for 30 min. Among seventeen hybridization positive clones isolated, four clones containing large cDNA inserts were selected and subcloned into pUC118. The cDNA inserts of these clones showed an identical digestion pattern of the restriction enzymes except for some length difference in their extreme 5Ј portions, and one clone containing the largest cDNA inserts (pRPIK7, 4.2 kb) was chosen for further sequence analysis on both strands by the dideoxy chain termination method (15) with a 373A DNA sequencer (Applied Biosystem) according to the supplier's instructions. By another round of screening with a 5Ј portion of the pRPIK7 as a probe, a further elongated clone, pRPIK20 (6.4 kb), was isolated. The 5Ј missing end was obtained by a rapid amplification of cDNA ends (RACE) PCR (5Ј AmpliFinder RACE kit, Clontech). For the sequences obtained by RACE, four independent clones were completely sequenced.
RNA Extraction and Northern Blot Analysis-Total RNA was extracted from several tissues of adult rats as described previously (16). Each of the total RNA samples (30 g/lane) were denatured with formaldehyde and size-separated by agarose gel electrophoresis. The RNA was transferred and fixed to a nylon membrane (Nytran, Schleicher & Schuell) and hybridized with a probe corresponding to the 5Ј coding sequence (nucleotides 1277-3354) labeled with [ 32 P]dATP. This 5Ј coding sequence was not included in human PI4K␣ (12). Conditions for hybridization and washing were performed as described previously (14). Autoradiography was performed at Ϫ80°C for 3 days.
Transfection and PI Kinase Activity-The full-length cDNA for the newly identified molecule was subcloned into the expression vector, pSRE (pcDL-SRa 296 in Ref. 17) as modified by Sakane et al. (18). A region of cDNA for our molecule corresponding to that for human PI4K␣ was also subcloned into the vector for comparison of enzymatic characteristics. The vector alone or the constructs were transfected into COS-7 cells by DEAE-dextran (19). After incubation for 3 days in Dulbecco's modified Eagle medium supplemented with 10% fetal calf serum, transfected COS-7 cells were harvested and lysed by sonication in lysis buffer containing 20 mM Tris-HCl (pH 7.4), 0.25 M sucrose, 4 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 20 mg/ml leupeptin, 20 mg/ml aprotinin, 20 mg/ml pepstatin, 50 mg/ml soybean trypsin inhibitor, and 1 mM phenylmethylsulfonyl fluoride. After removal of undisrupted cells by centrifugation (550 ϫ g, 10 min), the supernatant, hereafter referred to as the lysate, was centrifuged at 100,000 ϫ g for 30 min to separate soluble and particulate fractions. Protein concentrations were determined by the method of Lowry et al. (20) with bovine serum albumin as a standard. PI kinase activity was measured by the method of Kato et al. (21) with some modifications. The reaction mixture (50 l) contained 0.3% Triton, 50 mM Tris-HCl (pH 7.4), 20 mM MgCl 2 , 1 mM EGTA, 5 mg/ml of phosphatidylinositol (Sigma), and 1 mM ␥ATP (5,000 cpm/nmol; ICN). The reaction was continued for 5 min at 30°C and was stopped with 80 l of 1 N HCl. The lipid was extracted with 160 l of 1:1 (v/v) chloroform: methanol. The organic layer was collected and analyzed by thin layer chromatography (TLC). Thin layer plates of silica gel (Merck) were pretreated in 40% methanol containing 1% potassium oxalate and 2 mM EDTA (pH 7.2) and then baked at 105°C for 60 min. The TLC was developed in a solution containing 70:100:15:25 (v/v/v/v) chloroform: methanol:28% ammonia solution:distilled water. The band of phosphatidylinositol phosphate (PIP) detected by autoradiography was scraped with a sharp spatula and collected for liquid scintillation counting. For further determination of the PIP product as PI(3)P or PI(4)P, thin layer plates precoated with trans-1,2-diaminocyclohexane-N,N,NЈ,NЈ-tetraacetic acid were used as described by Walsh et al. (22). Position of PI(3)P was determined by using as a sample A431 cell lysate, which had been reported to contain PI 3-kinase activity (23).
Epitope Tagging and Immunoprecipitation-An epitope tag composed of eight amino acids (FLAG marker peptide, Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys by Kodak Co.) was fused to the newly identified molecule by cloning the 24-base pair FLAG coding sequence next to the initiation codon ATG of the novel cDNA. The FLAG epitope-tagged molecule was expressed in COS-7 cells using the expression vector, pSRE, by DEAE/dextran method. Transfected COS-7 cells were harvested and lysed with 1 ml of 1% Nonidet P-40 in lysis buffer containing 20 mM Tris-HCl (pH 7.4), 0.25 M sucrose, 4 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 20 mg/ml leupeptin, 20 mg/ml aprotinin, 20 mg/ml pepstatin, 50 mg/ml soybean trypsin inhibitor, and 1 mM phenylmethylsulfonyl fluoride. Lysates were incubated for 40 min at 4°C with gentle rocking and centrifuged at 10,000 ϫ g for 20 min at 4°C. Equal quantities of soluble protein were immunoprecipitated for 1 h at 4°C with a monoclonal antibody (anti-FLAG-M2, Kodak) specific for the FLAG marker peptide. Protein A-Sepharose beads precoated with 2 mg/ml bovine serum albumin were then added to the cell lysates for 1 h of incubation. Beads were washed twice with 1% Nonidet P-40 in the lysis buffer and thrice with the lysis buffer. Thereafter, PI kinase assays were performed on beads by the ways described above.
In Situ Hybridization Histochemistry-Fresh frozen blocks of brains from adult (postnatal day 49) male and fetal (prenatal day 18) rats were sectioned at 30-m thickness on a cryostat. The sections were mounted on silane-coated glass slides, fixed with paraformaldehyde, and pretreated as described previously (14). The slides were prehybridized in a solution containing 50% formamide, 4 ϫ SSC, 1 ϫ Denhardt's, 1% sarcosyl, 0.1 M sodium phosphate (pH 7.2), 100 mM dithiothreitol, and 250 g/ml heat-denatured salmon sperm DNA for 2 h at room temperature. Hybridization was performed at 42°C for 16 h in a moist chamber in the same solution (50 l/slide), which contained an additional 10% dextran sulfate and 0.5-1.0 ϫ 10 6 cpm of the same cDNA probe as used for Northern analysis labeled with [ 35 S]dATP by nick translation or the same amount of a control probe (pBR322, the plasmid vector). After hybridization, the slides were sequentially rinsed in 2 ϫ SSC, 0.1% sarcosyl at room temperature for 30 min, three times in 0.1 ϫ SSC, 0.1% sarcosyl at 42°C for 40 min each, and dehydrated in 70 and 100% ethanol. The sections were exposed to Hyperfilm-␤ max (Amersham Corp.) for 2-3 weeks.
Immunohistochemistry and Immunoblotting-The cells were fixed with 2% paraformaldehyde/0.1 or 0.01% Triton X-100 and were incubated with the anti-FLAG antibody. Sites of antigen-antibody reaction were visualized using the avidin-biotinylated peroxidase complex (ABC) system (Vector Laboratories) with diaminobezidine as a substrate. Some of the specimens, after immunoreaction, were postfixed with OsO 4 and uranyl acetate and embedded in Epon. Ultrathin sections were examined under electron microscope. In the immunoblotting, the lysates of the overexpressed cells were boiled for 4 min in Laemmli's sample buffer and subjected to SDS/7.5% polyacrylamide gel electrophoresis (24). The proteins were then electrophoretically transferred to a nitrocellulose membrane (pore size, 0.45 m). After blocking the nonspecific binding sites in the 5% skim milk (w/v) in phosphatebuffered saline, the membrane was incubated for 2 h at room temperature with the antibody against FLAG and then treated with peroxidase-conjugated anti-rabbit IgG antibody for 1 h.

RESULTS
The cDNA clone was isolated through PCR-mediated DNA amplification by using primers for the amino acid sequences conserved between yeast STT4 and PIK1 and subsequent screening of a rat brain cDNA library. The primers used were degenerate oligonucleotides with mammalian codon usage against yeast amino acid sequences. The homologous DNA sequences were amplified from a rat brain cDNA library and subcloned into pUC118.
Sequence analysis of the PCR products revealed that four of 20 clones (pMP1, 2, 12, and 18) isolated were identical and showed amino acid sequence identities of 52.3 and 34.4% with yeast STT4 and PIK1, respectively. A rat brain cDNA library containing 3 ϫ 10 6 phage clones was screened by hybridization with the cDNA probe derived from the pMP1 cDNA under high stringency conditions. Four clones containing the large cDNA inserts were isolated from positive clones and subcloned into pUC118. Restriction map of these cDNA inserts revealed an identical pattern except for some size differences in the 5Ј end of the cDNA inserts. The nucleotide sequence of a representative clone containing the largest cDNA inserts (pRPIK7, 4.2 kb) was determined. Another round of screening and a RACE were performed to obtain the 5Ј end of the coding sequence.
The nucleotide and deduced amino acid sequences of the composite cDNA are presented in Fig. 1. The putative initiation codon was preceded by an in-frame stop codon at nucleotides Ϫ249, Ϫ305, and Ϫ323. The deduced amino acid sequence encoded a protein of 2041 amino acids with a calculated molecular weight of 231,317. By comparison with sequences in the protein and DNA data base, the nucleotide sequence of the composite cDNA from nucleotide 3562 throughout the rest of the 3Ј coding region starting with an ATG sequence was revealed to be 89.6% identical to and its deduced amino acid sequence was 98.2% identical to the sequence for the recently reported human PI 4-kinase of type II, termed PI4K␣ (12). As expected from this identity, the present cDNA contained sequences encoding the ankyrin repeat domain, lipid kinase unique domain, pleckstrin homology domain, and presumed lipid kinase/protein kinase homology domain in the same turn as PI4K␣ (12). The deduced amino acid sequence from the present composite cDNA also showed 52.3 and 34.4% identity in the presumed catalytic domain and 40.7 and 28.3% identity in the lipid kinase unique domain to STT4 and PIK1, respectively. In addition, this sequence showed 31.7% identity to p110␣ subunit of rat PI 3-kinase in the presumed catalytic domain and 26.7% identity in the lipid kinase unique domain (Fig. 2). As for the N-terminal half, which had no identity relation to the human PI4K␣, we found two proline-rich regions (amino acids 152-156 and 210 -218) and a SH3 domain (25). Five leucine-rich portions were also found in the sequence, each of which, however, did not form a complete ␣-helix by Chou-Fasman analysis program (26).
In Northern blot analysis of adult rat tissues using a probe composed of the nucleotide sequence from nucleotides 1277 to 3354 of the present composite cDNA, whose counterpart is not contained in the human PI4K␣, a single hybridization band of 7.8 kb in size consistent with nearly a full length of the present cDNA was detected intensely in the brain, kidney, and lung and less intensely in the small intestine, uterus, and adrenal gland, whereas it was detected weakly to faintly in the heart, skeletal muscle, thymus, spleen, and testis. No distinct hybridization band was detected in the liver (Fig. 3). For comparison, the same nylon membrane was hybridized with another probe composed of nucleotide sequence from 4206 to 6466, which corresponds to the sequence from nucleotides 654 to 2905 of the human PI4K␣. The hybridization pattern was the same as the former one, and no significant hybridization band was clearly detected at sizes about 3.5 kb, in which the smaller band had been described to occur for the human PI4K␣ in the previous study by Wang and Cantley (12).
The lipid kinase activity of this novel molecule was measured FIG. 1. a, map of cDNA encoding rat PI 4-kinase. Isolated clones and RACE product were combined to construct the composite cDNA. The open box and the lines represent the coding region and untranslated sequences, respectively. The PCR product (pMP1) used for screening a full-length cDNA, and the probes used for Northern blot analysis and in situ hybridization histochemistry are also indicated. b, nucleotide sequence of the composite cDNA and the deduced primary structure of rat PI 4-kinase. In-frame stop codons in the 5Ј-untranslated region are underlined. Proline-rich regions are doubly underlined. The leucine, isoleucine, and valine residues (shaded) making up leucine-rich portions are also shown. by transfection into COS-7 cells of the full-length cDNA including the noncoding regions. As controls, an expression vector, pSRE (17), was expressed in COS-7 cells. A lysate from COS-7 cells transfected with rat PI 4-kinase cDNA showed 4-fold higher activity with the substrate than the control lysate, whereas another lysate from COS-7 cells transfected with the epitope-tagged cDNA also showed a high kinase activity (Fig.  4a). In the immunoprecipitation using the antibody against FLAG-tag and lysates from COS-7 cells overexpressed with the epitope-tagged cDNA, the kinase activity in the lysates was at least 100-fold higher than that in the control. In the immunoblotting by the antibody against FLAG tag, a single immunoreaction band with an approximate molecular size of 230 kDa was clearly detected in the lysate from the transfected cells (Fig. 4b).
PIP products were further analyzed to separate PI(4)P and PI(3)P according to Walsh et al. (22). It was revealed that the products catalyzed by the present molecule were almost exclusively PI(4)P (Fig. 4c). The kinase activity was not changed in the presence or the absence of calcium ion (data not shown). After separation of the lysate into soluble and particulate fractions, the enzymatic activity for the present molecule was recovered predominantly from the particulate fraction (Fig.  4d). The predominant recovery in the particulate fraction was true in case of transfection with the epitope-tagged cDNA (data not shown). In the immunoblotting by the antibody against the FLAG tag, a single immunoreaction band was dominant in the particulate fraction (Fig. 4e). The activity for this molecule was markedly stimulated in the presence of Triton X-100 with increasing concentration from 0.1 to 0.2% but inhibited slightly as the concentration of Triton X-100 exceeded 0.2% (Fig. 5a). On the other hand, equal amounts of COS-7 cell lysates containing the FLAG-tagged cDNA for this molecule were immunoprecipitated with the anti-FLAG antibody and assayed under various concentrations of adenosine and a constant concentration of 0.3% Triton X-100. The activity for this molecule was relatively insensitive to increasing concentration of adenosine (Fig. 5b).
When COS cells were overexpressed with the epitope-tagged cDNA and immunostained for the FLAG tag, cells immunoreactive for FLAG accounted for approximately 2-5% of the total cell population, and they appeared randomly dispersed in each culture dish. The immunoreactive products were densely aggregated in forms of caps in juxtaposition to the nuclei (Fig. 6a). No significant immunoreactivity was discerned in any other cell regions such as the cell margins or nuclei. No immunoreactivity was detected in any cells when the transfection was made with the cDNA without the tag (data not shown). In immunoelectron microscopy, the immunoreaction products were localized in the cytoplasmic surface of the membranes of Golgi vesicle and vacuoles and their adjacent cytoplasm. No immunoreaction products were seen on the plasma membrane (Fig. 6c). As a reference, the FLAG epitope tag was fused to the 3Ј half of cDNA for the present molecule corresponding to the human PI4K␣-cDNA, and COS cells were transfected with the tagged cDNA and immunostained. The resultant light microscopic image was the same as that observed with the full-length cDNA, and the immunoreactive product was largely confined to perinuclear region (Fig. 6b).
By in situ hybridization histochemistry of brain on prenatal day 18, the expression for this novel molecule was detected intensely throughout the mantle zone of fore-, mid-, and hind brain. In the cerebrum, the expression was intense in the FIG. 2. a, linear representation of rat PI4K, PI4K␣, two yeast PI 4-kinases, STT4 and PIK1, and rat PI 3-kinase p110␣ subunit. The lipid kinase unique domain, pleckstrin homology domain, catalytic domain, proline-rich region, and SH3 domain are shown. b, identity of the catalytic domain among the PI kinases described above. Conserved residues are shaded, and sequences used for PCR amplification are doubly underlined. c, comparison of the SH3 domain of rat PI4K with the SH3 domains of phospholipase C␥, PI 3-kinase p85 subunit, c-src, and spectrin. Consensus residues of the SH3 domain, as defined by Musachio et al. (25), and identical ones are shaded. cortical plate, whereas it was weak in the vetricular zone, and no expression was seen in the intermediate zone (Fig. 7a). On postnatal day 49, the expression was evident more or less throughout the gray matter of the entire brain, among which the hippocampal pyramidal cells, the dentate granule cells, and the cerebellar granule cells expressed the mRNA intensely and the olfactory mitral and granule cells and the cerebral cortex expressed it moderately. Although the expression was weak in the diencephalon and brain stem, no significant expression was detected in the cerebellar medulla (Fig. 7b). When the expression signals in sections from fetal and adult brains were compared by simultaneous exposure of both sections to one and the same Hyperfilm-␤ max, the expression appeared much higher in the fetal brain than the adult brain, especially in the brain stem. When sections of brains at fetal and adult stages were hybridized with the control probe, a cDNA fragment of about 800 base pairs (PstI-HindIII) from the pBR322 plasmid vector without any insert cDNA, no significant hybridization signals were detected anywhere in the sections (data not shown). DISCUSSION We have cloned a mammalian PI 4-kinase species that localizes in hippocampal pyramidal and dentate granule cells and most neuronal cells throughout the gray matter of the brain. Judging from the enhanced activity of PI 4-kinase by Triton, the relative insensitivity to inhibition on the enzyme activity by adenosine, the calculated molecular weight in accord with that so far reported for the purified counterpart enzyme, and the membrane association of the enzyme activity, it is strongly suggested that this enzyme molecule is the first cloned mammalian PI 4-kinase of type III.
When compared with PI4K␣, a human type II PI 4-kinase reported by Wang and Cantley (12), this rat PI 4-kinase molecule has several distinct features: a much higher molecular weight (230 versus 97 kDa) and relative insensitivity to adenosine. On the other hand, several similarities are noted: an activation of kinase activity by detergents, especially an identity (89.6 and 98.2%, respectively), of the 3Ј half coding region of cDNA and its deduced amino acid sequence for the present rat PI 4-kinase to PI4K␣, and consequently a similar tissue distribution of mRNA between the two molecules except for testis. The similarity is noted in the Northern blotting with either one of the two cDNA probes for the present PI 4-kinase: its 5Ј coding region, whose counterpart is not contained in PI4K␣, and its 3Ј coding one, which corresponds to PI4K␣. Although only a single band of 7.8 kb size is evident in any tissues even with the 3Ј probe in the present study, this may represent a much lower expression of the smaller transcripts due to the species difference. The smaller transcript was weekly detected even in their Northern blotting of human materials.
The association with particulate fraction of this PI 4-kinase activity as observed in the COS expression system does not seem to mean that this enzyme is an intrinsic membrane protein, because no amino acid sequences sharing the significantly extended hydrophobicity is found in its primary structure by the hydropathy analysis (27).
Although there has been no careful analysis of type III versus type II PI 4-kinase activities in subcellular fractions, the activity of PI kinase has been found in intracellular membrane, such as Golgi apparatus and lysosomes in addition to the plasma membrane (28 -32). It has also been reported that brain coated vesicles have both type II and III PI 4-kinase activity, whereas red cell plasma membrane has only type II PI 4-kinase activity (6,7). In this regard, the present study provides the first clear evidence for the localization of PI 4-kinase having the enzymatic characteristics of type III in the Golgi apparatus membranes, although we have to admit a requirement for this conclusion; the exogenously expressed protein should be localized to its normal intracellular compartment in the transfected cells. If this localization is further confirmed in some normal in situ cells immunohistochemically using a specific antibody against this novel molecule itself, it implies that PI 4-kinase play direct roles in the vesicular traffic, the main Golgi function, as suggested by Liscovitch et al. (33), rather than in the ligand-stimulated receptor-mediated turnover of PI in the plasma membrane.
With regard to the tissue distribution by the present study, the intense expression of mRNA for this novel PI 4-kinase molecule is detected in the brain in the present Northern blotting. The in situ hybridization analysis reveals its expression to be evident in almost all neurons throughout the fetal brain and the expression to decrease in adult brain. This expression pattern in brain development implies that this novel PI 4-kinase may be involved in neuronal differentiation and maturation including the synaptogenesis and synaptic plasticity, although the detail mechanism in the involvement remains to be elucidated.
Because the 3Ј half coding region of cDNA for the rat PI 4-kinase molecule is identical to PI4K␣, a human PI 4-kinase recently cloned by Wang and Cantley (12), the pleckstrin homology domain is, as expected, contained in the present PI 4-kinase in the same sequence alignment with the catalytic and lipid kinase unique domains as PI4K␣. The presence of the pleckstrin homology domain in a given molecule has suggested a possible regulation of the molecule by G proteins (34,35). This seems to be the case for p110␥ isotype, unlike other isotypes, of human PI 3-kinase, which has recently been cloned as the first example containing the pleckstrin homology domain among PI kinases (36). Although the possibility of activation of this PI 4-kinase by G proteins through the pleckstrin homology domain remains to be elucidated, Wang and Cantley have already reported no significant changes of PI4K␣ activity when G protein ␤␥ is added to PI4K␣ immunoprecipitates.
Among consensus sequences revealed in the molecular structure of this PI 4-kinase molecule, the presence of two prolinerich regions in the N-terminal half, which has no identity relation to PI4K␣, may be noticeable in the molecular regulation mechanism of this PI 4-kinase, because the proline-rich sequences are possible sites to bind with the SH3 domain of  c, and d) or PI4K␣ (b). The inset shows Golgi region at a higher magnification. Note dense aggregates of the immunoreaction materials in forms of caps (arrows) in juxtaposition to the nuclei (N) and their close association to the membrane of the Golgi vesicles and vacuoles (G) in both of the cells. Also note that there is no immunoreaction on the plasma membrane (P). ϫ300 (a and b); ϫ5000 (c); ϫ10000 (d).

FIG. 7.
In situ hybridization of rat PI 4-kinase mRNA in the fetal (a) and adult (b) rat brains. Film autoradiographic image of parasagittal section. Note expression signals throughout the gray matter of entire brain, such as the olfactory bulb (OB), the cerebral cortex (Cx), the hippocampal pyramidal cell layer (Hip), the dentate granular cell layer (DG), and the cerebellar cortex (Cb). Note intense expression signals throughout the mantle zone of the entire brain on prenatal day 18. The ventricular zone (V) expresses the mRNA weakly, and no significant expression is seen in the intermediate zone (arrow). Also note the positive expression throughout the gray matter of the entire brain with various intensity in different loci on postnatal day 49. No expression signals are seen in the white matter, such as the corpus callosum (CC) and the cerebellar medulla (CM). CP, caudate putamen; Cx, cerebral cortex; P, pons; T, tectum; Th, thalamus. some yet unknown cytosolic molecules involved in the signal transduction. It may be possible that the proline-rich domain makes a self-association with the SH-3 domain found in this PI 4-kinase in an intramolecular manner.
As noted in the introduction, the calculated molecular weight of human PI4K␣, which was recently reported to show the activities of type II PI 4-kinase by Wang and Cantley (12), was much larger than that of most species of type II PI 4-kinase that had previously been purified from various tissues. In this regard, we should notice again the remarkable identity in the deduced amino acid sequence of human PI4K␣ to that for a 3Ј half of the coding region of the present rat PI 4-kinase molecule and the tissue distribution of the present hybridization band of 7.8 kb in the Northern blotting almost corresponding except for testis to that of a larger transcript (7.5 kb) for PI4K␣, suggesting that PI4K␣ is an alternatively spliced species of the present PI 4-kinase molecule. Several explanations are thus possible for these issues. One of them is that previously purified type II PI 4-kinase represents an alternatively spliced species of the present PI 4-kinase molecule. The other is that there exists a distinct gene encoding type II PI 4-kinase with the molecular weight of about 50 kDa. The previous finding by Endemann et al. (37) that a monoclonal antibody inhibiting the enzyme activity of type II PI 4-kinase from a variety of tissues and species failed to inhibit that of type III PI 4-kinase is in favor of the latter possibility. On this latter assumption, the search for other PI 4-kinase species is under way.