The serine/threonine protein kinase encoded by the Akt proto-oncogene is catalytically inactive in serum-starved primary and immortalized fibroblasts. Here we show that Akt and the Akt-related kinase AKT2 are activated by PDGF. The activation was rapid and specific, and it was abrogated by mutations in the Akt Pleckstrin homology (PH) domain. The Akt activation was also shown to depend on PDGFR beta tyrosines Y740 and Y751, which bind phosphatidylinositol 3-kinase (PI 3-kinase) upon phosphorylation. Moreover, Akt activation was blocked by the PI 3-kinase-specific inhibitor wortmannin and the dominant inhibitory N17Ras. Conversely, Akt activity was induced following the addition of phosphatidylinositol-3-phosphate to Akt immunoprecipitates from serum-starved cells in vitro. These results identify Akt as a novel target of PI 3-kinase and suggest that the Akt PH domain may be a mediator of PI 3-kinase signaling.

The protein kinase encoded by the Akt proto-oncogene is activated by phospholipid binding, membrane translocation and phosphorylation. To address the relative roles of these mechanisms of Akt activation, we have employed a combination of genetic and pharmacological approaches. Transient transfection of NIH3T3 cells with wild-type Akt, pleckstrin homology (PH) domain mutants, generated on the basis of a PH domain structural model, and phosphorylation site Akt mutants provided evidence for a model of Akt activation consisting of three sequential steps: (1) a PH domain-dependent, growth factor-independent step, marked by constitutive phosphorylation of threonine 450 (T450) and perhaps serine 124 (S124), that renders the protein responsive to subsequent activation events; (2) a growth factor-induced, PI3-K-dependent membrane-translocation step; and (3) a PI3-K-dependent step, characterized by phosphorylation at T308 and S473, that occurs in the cell membrane and is required for activation. When forced to translocate to the membrane, wild-type Akt and PH domain Akt mutants that are defective in the ®rst step become constitutively active, suggesting that the purpose of this step is to prepare the protein for membrane translocation. Both growth factor stimulation and forced membrane translocation, however, failed to activate a T308A mutant. This, combined with the ®nding that T308D/S473D double mutant is constitutively active, suggests that the purpose of the three-step process of Akt activation is the phosphorylation of the protein at T308 and S473. The proposed model provides a framework for a comprehensive understanding of the temporal and spatial requirements for Akt activation by growth factors.

The cytoplasmic serine-threonine protein kinase coded for by the c-akt proto-oncogene features a protein kinase C-like catalytic domain and a unique NH 2 -terminal domain (AH domain). The AH domain is a member of a domain superfamily whose prototype was observed in pleckstrin (pleckstrin homology, or PH, domain). In this communication, we present evidence that the AH/PH domain is a domain of protein-protein interaction which mediates the formation of Akt protein complexes. The interaction between c-akt AH/PH domains is highly specific, as determined by the failure of this domain to bind AKT2. The AH/PH domain-mediated interactions depend on the integrity of the entire domain. Akt molecules with deletions of the NH 2 -terminal portion (amino acids 11 to 60) and AH/PH constructs with deletions of the C-terminal portion of this domain (amino acids 107 to 147) fail to interact with c-akt. To determine the significance of these findings, we carried out in vitro kinase assays using Akt immunoprecipitates from serum-starved and serum-starved, plateletderived growth factor-stimulated NIH 3T3 cells. Addition of maltose-binding protein-AH/PH fusion recombinant protein, which is expected to bind Akt, to the immunoprecipitates from serum-starved cells induced the activation of the Akt kinase.

The Akt protooncogene encodes a serine-threonine protein kinase which is activated by growth factor-generated signals that are transduced via the phosphatidyl-

A cDNA clone that hybridizes to an mRNA (lA10) that accumulates to a substantial level in the axis of the germinating soybean seed was sequenced. The amino acid sequence of the clone indicates an almost perfect repeat of Pro-Pro-Val-Tyr-Lys resulting in a protein containing 40% proline and lacking serine and histidine. On the likelihood that such a protein might be a hydroxyproline-rich cell-wall-glycoprotein (HRGP), cell walls of a soybean cell culture were extracted by procedures used to obtain soluble basic cell-wall glycoproteins, and the proteins were fractionated and purified. A 33-kDa protein (and possibly a 28-kDa protein) was obtained that has an amino acid distribution similar to that of the cDNA clone. The protein lacks histidine and serine and contains 20% hydroxyproline and 20% proline. The HRGP is thus distinct both in its amino acid content and in its pentameric repeat of Pro-Pro-Val-Tyr-Lys, with half of the prolines being hydroxylated. (4) was maintained on a 14-day cycle with a 1:12 dilution of cells (7 ml and 77 ml of medium in a 250-ml flask) every 2 weeks.In the preparation described in Results,25 ,uCi (1 Ci = 37 GBq) of [14C]proline was added to 8-day cells, and the cells were allowed to continue growing for 4 hr, after which time they were filtered on a coarse scintered funnel and frozen in liquid N2. In more recent preparations, 8-to 9-day cultures have been collected directly without labeling and the purification steps have been monitored solely by following the absorbance at 280 nm. Frozen cells (4.6 g) were homogenized in a Polytron in 0.25 M sucrose/3 mM EDTA/40 mM Tris-HCI, pH 8/0.5 mM phenylmethylsulfonyl fluoride/5 mM dithiothreitol (2.2 ml/g), and centrifuged for 2 min at 750 x g. The pellet was recentrifuged for 10 min at 23,500 x g and then washed one time in 14 ml of 5 mM potassium phosphate (pH 6.8) and six times with 15 ml of deionized water, centrifuging each time for 2 min at 750 x g. The pellet was extracted twice with 12 ml of 0.2 M CaC12 (5, 6), each extraction lasting 45 min. After centrifuging for 5 min at 2000 x g, the supernatant was adjusted with 68% trichloroacetic acid to a final concentration of 10% trichloroacetic acid, and the suspension was kept at 0C overnight. After centrifuging for 15 min at 24,000 x g, the supernatant was dialyzed against five changes of water over a 5-hr period and then overnight against 25 mM Tris HCI (pH 8.0). The dialyzer tubing had a Mr 3000 cutoff (Thomas 3787-H45). A typical preparation yielded 1.9 mg of protein based on 4 A280 units for a solution of 1 mg/ml, a figure calculated from the presence of 13 mol % tyrosine (see Table 1) in comparison with bovine serum albumin, which has 3.3% tyrosine (7) and 1 A280 unit for a solution of 1 mg/ml. The sample was chromatographed on a column of CM-cellulose (1.2 x 10 cm) equilibrated with 25 mM Tris-HCl (pH 8.0) and eluted with 100 ml of a linear gradient of 0.4 M NaCl in the same buffer. The 33-kDa protein, eluting at ""0.14 M NaCI, was pooled (tubes 17-27 in Fig. 3), concentra...

We have resolved and analyzed two proline-rich proteins isolated from the walls of soybean cells in culture. The proteins are similar in amino acid content, containing 20% proline, 20% hydroxyproline, 20% lysine, 16% valine, 10% tyrosine, and 10% glutamate. The proteins undergo a rearrangement or a limited cleavage in dilute NaOH, but are otherwise remarkably stable to a high concentration of alkali. We have cloned and sequenced a cDNA from soybean axes germinated for 31 hours (1A10-2) coding for a protein that closely corresponds in its amino acid content to that of the proline-rich proteins. The cDNA sequence predicts a decameric repeat of Pro-Pro-Val-TyrLys-Pro-Pro-Val-Glu-Lys. Consequently, this class of proteins is referred to as repetitive proline-rich proteins, i.e., RPRP2 and RPRP3. We have also analyzed RNA gel blots with probes that discriminate between the new cDNA clone and a related cDNA previously reported [SbPRPl;Hong, Nagao, and Key (1987). J. Biol. Chem. 262, 8367-83761. Messenger RNAs from young seedlings and from soybean suspension cultures correspond primarily to the new RPRP clone (1A10-2), whereas the predominant mRNA accumulating later in the roots corresponds to SbPRPl.

We have resolved and analyzed two proline-rich proteins isolated from the walls of soybean cells in culture. The proteins are similar in amino acid content, containing 20% proline, 20% hydroxyproline, 20% lysine, 16% valine, 10% tyrosine, and 10% glutamate. The proteins undergo a rearrangement or a limited cleavage in dilute NaOH, but are otherwise remarkably stable to a high concentration of alkali. We have cloned and sequenced a cDNA from soybean axes germinated for 31 hours (1A10-2) coding for a protein that closely corresponds in its amino acid content to that of the proline-rich proteins. The cDNA sequence predicts a decameric repeat of Pro-Pro-Val-Tyr-Lys-Pro-Pro-Val-Glu-Lys. Consequently, this class of proteins is referred to as repetitive proline-rich proteins, i.e., RPRP2 and RPRP3. We have also analyzed RNA gel blots with probes that discriminate between the new cDNA clone and a related cDNA previously reported [SbPRP1; Hong, Nagao, and Key (1987). J. Biol. Chem. 262, 8367-8376]. Messenger RNAs from young seedlings and from soybean suspension cultures correspond primarily to the new RPRP clone (1A10-2), whereas the predominant mRNA accumulating later in the roots corresponds to SbPRP1.

Microarray analysis of human tissue is frequently hindered by the limited amount of RNA available. Although amplification protocols can be utilized, the relative representation of transcripts present in the starting material must remain unaltered. In this study, 200 ng of total RNA derived from cultured renal epithelial cells from tuberous sclerosis complex (TSC) carriers and control individuals was amplified by in vitro transcription with T7 RNA polymerase. The resulting Cy-labeled cDNAs (from total or amplified RNA (aRNA)) were analyzed as direct replicates and dye-flips on slides containing 10,000 human cDNAs. The Pearson correlation coefficients for the direct replicate experiments were 0.80 (20 microg total RNA), 0.85 (40 microg total RNA), and 0.93 (2 microg of aRNA). Comparisons between the array data revealed that the majority of genes expressed in total RNA (97% for 20 microg and 85% for 40 microg) were also detected in aRNA. The correlation coefficient of the expression ratios for genes detected in both total RNA (40 microg) and aRNA was 0.63. Further, Student's t-test indicated no significant difference (P = 0.83) between these ratios. These results indicate that the number of expressed genes detected with total RNA is proportional to the amount of RNA used and underscore the requirement of large amounts of total RNA for a comprehensive characterization of gene expression profiles. RNA amplification allows the detection of a large number of genes expressed in the starting RNA population without altering their relative intensities significantly. Thus, an RNA amplification step improves the quality of gene expression results obtained by microarray analysis. This study indicates that high quality microarray data can be generated from small amounts of RNA, including those extracted from limiting clinical samples and microdissected histological specimens.