A novel 55-kDa regulatory subunit for phosphatidylinositol 3-kinase structurally similar to p55PIK Is generated by alternative splicing of the p85alpha gene.

Phosphatidylinositol 3-kinase, which is composed of a 110-kDa catalytic subunit and a regulatory subunit, plays important roles in various cellular signaling mechanisms. We screened a rat brain cDNA expression library with P-labeled human IRS-1 protein and cloned cDNAs that were very likely to be generated by alternative splicing of p85α gene products. These cDNAs were demonstrated to encode a 55-kDa protein (p55α) containing two SH2 domains and an inter-SH2 domain of p85α but neither a bcr domain nor a SH3 homology domain. Interestingly, p55α contains a unique 34-amino acid sequence at its NH terminus, which is not included in the p85α amino acid sequence. This 34-amino acid portion was revealed to be comparable with p55PIK (p55) in length, with a high homology between the two, suggesting that these NH-terminal domains of p55α and p55 may have a specific role that p85 does not. The expression of p55α mRNA is most abundant in the brain, but expression is ubiquitous in most rat tissues. Furthermore, it should be noted that the expression of p85α mRNA in muscle is almost undetectably low by Northern blotting with a cDNA probe coding for the p85α SH3 domain, while the expression of p55α can be readily detected. These results suggest that p55α may play an unique regulatory role for phosphatidylinositol 3-kinase in brain and muscle.

In this study, we isolated a novel alternatively spliced cDNA from the p85␣ gene by expression screening from a rat brain cDNA library using a 32 P-labeled human IRS-1 protein. This cDNA was demonstrated to encode a 55-kDa protein, which was designated p55␣, because it is partly identical to p85␣. In addition, we suggest changing the name of p55PIK to p55␥ to avoid confusion between p55PIK (p55␥) and p55␣. Herein, we compare the amino acid sequences of four isoforms of the regulatory subunit of rat PI 3-kinase and show their tissue distributions. These isoforms may be activated by different stimuli and/or at different intercellular locations.

MATERIALS AND METHODS
Preparation of Recombinant Human IRS-1-Human IRS-1 cDNA was obtained by screening the human genomic library using a 32 Plabeled DNA fragment. According to the sequence of human IRS-1 reported by Araki et al. (12), oligonucleotides were synthesized as follows: TCAATGCTGCAACAGCAGATGA as a forward primer and TCAGTGCCAGTCTCTTCCTCTCTG as a reverse primer. PCR amplification was performed using these primers, and a 321-bp fragment was obtained from human genomic DNA. A human genomic library (a generous gift from Dr. H. Hirai, Third Department of Internal Medicine, Faculty of Medicine, University of Tokyo) was screened using a 32 Plabeled 321-bp PCR fragment, and one positive clone was isolated. The coding region of human IRS-1 genomic DNA was subcloned into pBacPAK9 transfer vector, a baculovirus vector (Invitrogen), and the baculovirus was produced according to the manufacturer's instructions. The purification of IRS-1 from Sf-9 cells infected with the baculovirus containing IRS-1 DNA was performed as described previously (11). The insulin receptor was partially purified from human placenta on wheat germ agglutinin agarose as described previously (13). The 32 P-labeled IRS-1 probe was prepared by the incubation of IRS-1 with activated insulin receptor in the presence of Mn 2ϩ and 32 P-labeled ␥-ATP (11).
Expression Screening with Human [ 32 P]IRS-1 Protein-An oligo(dT)primed rat brain cDNA library was prepared in UNI-ZAP XR (Stratagene) according to the manufacturer's instructions. Sixty 15-cm diameter plates representing 3,000,000 independent plaques were plated and incubated for 7 h at 37°C. Then, the plates were overlaid with nitrocellulose filters that had been impregnated with 10 mM isopropyl-␤-D-thiogalactopyranoside and incubated for 8 h at 37°C. Hybridization of the filters with the [ 32 P]IRS-1 probe and washing were performed as described previously (11). The cDNA inserts in pBluescript were prepared by in vivo excision according to the manufacturer's instructions (Stratagene). The nucleotide sequences were determined using an ABI automatic sequencer. * This work was supported by a grant-in-aid for scientific research from the Japanese Ministry of Education. 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.
Northern Blotting-Northern blotting was performed using a commercially available filter made by Clontech (Palo Alto, CA). The 1-663nucleotide sequence of p85␣, 1011-2175 of p85␣, 1-2170 of p85␤, 96-1381 of p55␥, and 1-159 of p55␣ were labeled with [ 32 P]dCTP and used as probes. The filter was hybridized and washed according to the manufacturer's instructions (Clontech). Autoradiography was performed at Ϫ80°C for 12-48 h, and the radioactivities of the bands obtained were measured using a BAS2000 (Fuji).
To confirm the specificity of these antibodies, p85␣, p85␤, p55␥, and p55␣ were expressed in Sf-9 cells using the baculovirus system. These cDNAs coding the full amino acid sequence and the HA tag amino acid sequence (YPYDVPDYA) at each C terminus were subcloned into pBacPAK9 transfer vector, and the baculoviruses were prepared according to the manufacturer's instructions (Clontech). The Sf-9 cells infected with baculoviruses containing each of the four isoforms were cultured for 48 h and lysed in Laemmli buffer. The samples were subjected to SDS-PAGE, and immunoblotting was performed as described previously (14).
Immunoblotting of p85␣, p55␣, and p55␥ Expressed in Rat Brain-Rat brain was homogenized in ice-cold lysis buffer (1/10, w/v) containing 50 mM Hepes (pH 7.5), 137 mM NaCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 2 mM EDTA, 1% Nonidet P-40, 10% glycerol, 2 mg/ml aprotinin, and 34 mg/ml phenylmethylsulfonyl fluoride. Insoluble material was removed by centrifugation at 14,000 ϫ g for 60 min and incubated for 2 h at 4°C with ␣p85 PAN-UBI covalently coupled with protein A-Sepharose beads, which were also purchased from Upstate Biotechnology Inc. The beads were washed three times in lysis buffer, boiled in Laemmli buffer, and then removed by centrifugation. The supernatants were subjected to SDS-PAGE, and immunoblotting was performed as described previously (14).

RESULTS AND DISCUSSION
A human IRS-1 gene was successfully cloned from a human genomic library, and the complete nucleotide sequence of its coding region was determined. In comparison with the sequence reported by Araki et al. (12), two nucleotides were revealed to be different in our IRS-1 nucleotide sequence (C to G at 2166 bp and A to G at 3432 bp). The C to G change at 2166 bp caused a change in the amino acid sequence (C to W at 382). As these differences were thought to be due to polymorphism, we prepared recombinant IRS-1 protein using a baculovirus containing this IRS-1 DNA.
A rat brain cDNA expression library was screened with 32 Plabeled recombinant IRS-1, and 81 positive independent clones were isolated after three or four rounds of screening. They included cDNAs containing complete coding regions of p85␣, p85␤, and p55␥, of which nucleotide sequences were determined. In addition, we obtained three independent cDNAs containing the nucleotide sequence coding for the NH 2 -terminal SH2 domain of p85␣ and previously undocumented 166nucleotide sequence at its 5Ј-upstream side. These cDNAs contained an open reading frame of 1362 nucleotides, and the deduced amino acid sequence is shown in Fig. 1. The presence of this mRNA in rat brain was confirmed by reverse transcription PCR using the 5Ј-primer in the newly identified nucleotide sequence and the 3Ј-primer in the nSH2 domain or in the cSH2 domain found in p85␣ cDNA (data not shown). We designated this putative protein p55␣ on the basis of its molecular weight. p55␣ contains two SH2 domains and an inter-SH2 domain, which are identical to those of p85␣. Thus, p55␣ mRNA appears to be transcribed by alternative splicing from the p85␣ gene. The SH3 domain and bcr homology domain found in p85␣ are replaced in p55␣ by a unique 34-residue NH 2 terminus followed by a conserved proline-rich motif (PPALPPKPPKP). Interestingly, this 34-residue region of p55␣ is comparable in length to the corresponding NH 2 -terminal portion of p55␥ (11), and 16 of the 34 amino acids are identical in the two peptides. These conserved sequences suggest that their unique NH 2terminal portion may have a specific functional role, which p85 does not. Further study is needed to resolve this issue.
The levels of expression of p85␣, p55␣, p55␥, and p85␤ mRNAs in various rat tissues were investigated, and the results are shown in Fig. 2. Northern blotting with a 5Ј-unique 159-nucleotide sequence located in the 5Ј-untranslated region FIG. 1. Alignment of amino acid sequences of p85␣, p55␣, p55␥, and p85␤. The cDNAs coding for the complete peptides of p85␣, p55␣, p55␥, and p85␤ were isolated by screening a rat brain expression cDNA library with 32 P-labeled IRS-1 protein probe. The nucleotide sequences were determined with an ABI automatic sequencer. The amino acid residues for each protein, with the addition of gaps (-) to optimize the alignment, are numbered to the right of each sequence. Two SH2, the bcr homology, the prolinerich, and the SH3 domains are boxed.

55-kDa Regulatory Subunit for PI-3 Kinase from p85␣ Gene
and a coding region for the NH 2 -terminal 25-amino acid sequence in the NH 2 terminus of p55␣, neither of which is included in the p85␣ cDNA nucleotide sequence, revealed three mRNA species of 6.0, 4.2, and 2.8 kb in the brain (Fig. 2B). Among them, the 4.2-kb band was also detected in all of other tissues examined. Northern blotting with nucleotides coding for the p85␣ SH3 domain revealed two mRNA species of 7.7 and 4.2 kb (Fig. 2A). In addition, the cDNA probe coding for the p85␣/p55␣ nSH2 domains was also used for Northern blotting, and four mRNA species of 7.7, 6.0, 4.2, and 2.8 kb were observed (Fig. 2C). The 4.2-kb band was detected on all Northern blots, and the intensities of this band were compared in various tissues. The amount of the 4.2-kb mRNA detected with Northern blotting using a cDNA probe coding for the p85␣/p55␣ nSH2 domain is thought to be the sum of the amounts of the p55␣ and p85␣ mRNAs. The intensity of the 4.2-kb band among various tissues observed in blotting utilizing a cDNA probe coding for the p85␣/p55␣ nSH2 domain was revealed to be similar to that obtained with a cDNA probe coding for the p85␣ SH3 domain and differed significantly from that obtained with the p55␣ 5Ј-unique cDNA probe. This result may suggest that p85␣ mRNA is expressed more abundantly than p55␣ mRNA in most tissues, with the apparent exceptions of brain and skeletal muscle. However, it should be noted that in skeletal muscle the p85␣ mRNA expression level is almost undetectably low, while p55␣ mRNA can be readily detected. In muscle, the activation of PI 3-kinase is presumed to be involved in insulinstimulated glucose uptake through the translocation of GLUT4 to the plasma membrane (3). Therefore, it might be possible that p55␣ plays a more important role than p85␣ in the stimulation of glucose uptake by skeletal muscle. In brain, both p55␣ and p55␥ mRNAs are expressed abundantly, as are those of p85␣, suggesting that these regulatory subunits having neither bcr homology nor SH3 domains may have a function(s) different from that of p85. PI 3-kinase appears to be important, first, in that its activation is essential for neurite elongation of rat PC-12 cells, and in addition, VPS34, a yeast PI 3-kinase homologue, was shown to be involved in vacuolar protein sorting (15). Thus, PI 3-kinase may play an essential role in the secretion of neurotransmitters via regulation of vesicle sorting in the brain. Taken together, one or more of these four regulatory subunits might be essential for neuronal differentiation, while the others may be involved in the secretion of neurotransmitters.
In order to detect the endogeneous p85␣, p55␣, and p55␥ proteins in rat tissues, we prepared specific antibodies against each of the three. These antibodies did not recognize different isoforms of regulatory subunits, produced in the Sf-9 cell experiment using the baculovirus expression system (Fig. 3). As shown in Fig. 3A, by immunoblotting using the anti-HA antibody (12CA5), the electrophoretic mobility of p55␣ is essentially the same as that of p55␥. The rat brain lysates immunoabsorbed by the beads covalently coupled with ␣p85 PAN-UBI , which recognize all p85␣, p55␣, p55␥, and p85␤ expressed in Sf-9 cells because of the highly conserved amino acid sequence of these peptides (data not shown), were subjected to SDS-PAGE and immunoblotted with control antibody, ␣p85 PAN-UBI , ␣p85␣ SH3 , ␣p55␣, and ␣p55␥ (Fig. 3E). The immunoblot obtained with ␣p85 PAN-UBI revealed the two bands of 85 and 55 kDa, while that obtained with ␣p85␣ SH3 showed only the 85-kDa band. In contrast, ␣p55␣ and ␣p55␥ both showed the 55-kDa band. These results indicate the expression of these isoforms in brain.
Finally, to determine whether or not p55␣ is associated with PI 3-kinase activity, as in the case of p85␣, we immunoprecipitated the rat brain soluble fraction with each control antibody, ␣p85 PAN-UBI , ␣p85␣ SH3 , or ␣p55␣. PI 3-kinase activities in these immunoprecipitates were measured (Fig. 3F). The immunoprecipitates obtained with ␣p85 PAN-UBI , ␣p85␣ SH3 , or ␣p55␣ were demonstrated to contain significant PI 3-kinase activity, as compared with the control antibody immunoprecipitate. This result strongly suggests that p55␣ also exists as a heterodimer with a p110 catalytic subunit and that it functions as a regulatory subunit of PI 3-kinase.
In this study, we showed that there are at least four isoforms of the regulatory subunit for PI 3-kinase. All of the four isoforms contain two SH2 domains and the binding site for association with the p110 catalytic subunit, suggesting that these regulatory subunits of PI 3-kinase interact with phosphotyrosine residues on the receptors or receptor substrates through one or both of their SH2 domains, resulting in activation of the p110 catalytic subunit. However, SH3 and bcr homology domains found in p85␣ or -␤ are replaced in p55␣ or -␥ by unique 34-residue NH 2 termini. Although the functional roles of SH3 domain have not been understood yet, the association between SH3 domain and proline-rich segments in various signaling proteins (i.e. dynamin (16), paxillin (17), hSOS1 (18), p85 subunit of PI 3-kinase (19)) is reported. Thus, the differences in the NH 2 -terminal region observed among the regulatory subunit isoforms may contribute to differences in subcellular distributions and/or to varying degrees of PI 3-kinase activation in response to various growth factor receptors and oncogenic products.
In summary, we have identified a novel alternatively spliced regulatory subunit, which may have important functions in brain and muscle. Our future studies will focus on the variety of possible functions mediated by differences in the NH 2terminal portions of the regulatory subunits of PI 3-kinase.