Phosphoinositide 3-kinase activation regulates cell division time by coordinated control of cell mass and cell cycle progression rate.

Cells must increase their mass in coordination with cell cycle progression to ensure that their size and macromolecular composition remain constant for any given proliferation rate. To this end, growth factors activate early signaling cascades that simultaneously promote cell mass increase and induce cell cycle entry. Nonetheless, the mechanism that controls the concerted regulation of cell growth and cell cycle entry in mammals remains unknown. The phosphatidylinositol 3-kinase (PI3K)/protein kinase B pathway regulates cell cycle entry by inactivating forkhead transcription factors and promoting cyclin D synthesis. PI3K/protein kinase B-derived signals also affect activation of p70 S6 kinase and the mammalian target of rapamycin, enzymes involved in cell growth control. We previously showed that enhancement of PI3K activation accelerates cell cycle entry, whereas reduction of PI3K activation retarded this process. Here we examined whether expression of different PI3K mutants affects cell growth during cell division. We show that diminishing or enhancing the magnitude of PI3K activation in a transient manner reduces or increases, respectively, the protein synthesis rate. Alteration of cell growth and cell cycle entry by PI3K forms appears to be concerted, because it results in lengthening or shortening of cell division time without altering cell size. In support of a central role for PI3K in growth control, expression of a deregulated, constitutive active PI3K mutant affects p70 S6 kinase and mammalian target of rapamycin activities and increases cell size. Together, the results show that transient PI3K activation regulates cell growth and cell cycle in a coordinated manner, which in turn controls cell division time.

Cell division is the process by which a cell duplicates its DNA content and cell mass to produce two daughter cells. In mammals, cell division is essential during development and in the adult for tissue regeneration. Most cell division is a symmetrical process that gives rise to two virtually identical daughter cells (1)(2)(3)(4)(5). To initiate symmetrical cell division, mitogens trigger a number of early signals that culminate in the activation of G 1 cyclin/CDKs (required for the G 1 -S transition) and induce an increase in cell mass. This increase is required to ensure that macromolecular composition and cell size are conserved in daughter cells (1)(2)(3)(4)(5). Signaling pathways that control the cell mass increase, referred to as cell growth, have recently been elucidated and include proteins such as phosphoinositide 3-kinase (PI3K), 1 p70 S6 kinase (p70 S6K), the mammalian target of rapamycin (mTOR), and the tuberous sclerosis complex (TSC1/2) (4,5). These proteins regulate synthesis of ribosomal components and activation of the translational machinery (4,5). It is nonetheless unknown how the two independent processes of cell cycle entry and cell growth are coordinated during cell division, particularly in mammals. Studies in Saccharomyces cerevisiae show that whereas blockage of cell division by inactivation of most cell division cycle (cdc) genes allows cell growth to continue, inhibition of cell growth impairs cell cycle progression (6). The cell cycle thus appears to be linked to cell growth. Similar results were obtained in Drosophila (7)(8)(9)(10).
PI3K is an enzyme that transfers phosphate to the 3-position of the inositol ring of membrane phosphoinositides. The PI3Ks are divided into three subclasses based on their primary structure and substrate specificity, but only class I enzymes generate phosphatidylinositol 3,4-bisphosphate and phosphatidylinositol 3,4,5-trisphosphate products in vivo. Basal levels of these lipids are very low in quiescent cells but increase rapidly and transiently following growth factor receptor stimulation, regulating a variety of cell responses including survival and division (11)(12)(13)(14). The 3-polyphosphoinositides recruit pleckstrin homology domain-containing proteins such as phosphoinositide dependent kinase-1 and protein kinase B (PKB), which mediate PI3K signal propagation (15)(16)(17)(18). Class IA PI3K is a heterodimer composed of a p85 regulatory and a p110 catalytic subunit (11-14, 19 -23). Activation of this enzyme following growth factor receptor stimulation controls cell cycle entry by regulating cyclin D synthesis and inactivation of FOX0 forkhead transcription factors, events required for G 1 -to-S transition (24 -27). Nonetheless, subsequent inactivation of PI3K is also important for completion of the cell cycle, because expression of constitutive active PI3K mutants inhibits forkhead activity in G 2 , which is required for mitotic progression (27). In addition to controlling these events, PI3K activation governs cell growth by regulating activation of p70 S6K and mTOR (28 -32).
In addition to the biochemical studies mentioned, genetic experiments in Drosophila support PI3K/PKB, p70 S6K, and mTOR involvement in cell growth control (7)(8)(9)(10). In mammals, deregulation of p70S6K, mTOR, or the PI3K/PKB pathway was shown to affect cell size (35,(51)(52)(53). Nonetheless, despite the well demonstrated function of PI3K/PKB, p70 S6K, and mTOR in cell growth control, little is known of how cell growth induction is linked to cell cycle progression. Based on the capacity of PI3K to regulate pathways that control cell growth and cell cycle entry, we hypothesized that PI3K activation may contribute to the concerted regulation of these processes during cell division in mammals.
We previously described the consequences on cell cycle progression of interfering with physiological PI3K regulation in NIH 3T3 cells by expressing different p85/p110 PI3K forms (27). These studies indicated that enhancement of PI3K activation in a transient manner accelerates cell cycle progression, whereas reduction of PI3K activation decreases this process (27). Here we show that expression of p65 PI3K , a mutant that enhances transient PI3K activation, augmented the protein synthesis rate of cycling cells. This increase was concerted with the cell cycle progression rate, because p65 PI3K expression shortened division time without altering cell size. Accordingly, expression of the recombinant p85␣ regulatory subunit, which reduces the magnitude of transient PI3K activation, increased cell division time without altering cell size. These observations illustrate the concerted regulation of cell growth and cell cycle progression rates by PI3K, thereby controlling cell division time. The key role of PI3K in growth control is supported by the observation that expression of a deregulated, constitutive active PI3K form altered p70 S6K and mTOR activation kinetics, giving rise to larger cells.
Extract Preparation and Western Blotting-Cells were lysed in 50 mM HEPES pH 8, 150 mM NaCl and 1% Triton X-100 containing phosphatase and protease inhibitors (27,58). For p70S6K immunoblotting, cells were lysed in 10 mM Hepes pH 7.8, 20 mM ␤ glycerol phosphate, 15 mM KCl, 1 mM EDTA, 1 mM EGTA, 10% glycerol, 0.2% Nonidet P-40 containing phosphatase and protease inhibitors (58). Protein concentration was estimated by the BCA assay (Pierce) and equal protein amounts were resolved in SDS-PAGE. Gels were transferred to nitrocellulose and probed with the indicated antibodies.
Cell Labeling-Cells were washed in methionine/cysteine-free RPMI (BioWhittaker) and incubated in this medium supplemented with 10% dialyzed fetal calf serum for 2 h prior addition of 35 S Met/Cys (20 Ci; Amersham Biosciences) for the times indicated. For 35 S Met/Cys labeling of cells in G 0 and G 2 , cells were incubated 16 h in serum-free medium or in medium containing 10% serum and 5 M etoposide, respectively, then labeled as above. For G 1 labeling, cells were incubated as for G 0 conditions, labeled, and then incubated with 10% dialyzed calf serum for 1 h. The cells were collected and lysed in Triton X-100 lysis buffer (50 mM HEPES pH 8, 150 mM NaCl and 1% Triton X-100 containing phosphatase and protease inhibitors, 58). Protein concentration was estimated and 20 g of total protein were resolved in SDS-PAGE and autoradiographed.
Cell Size Determinations-To examine cell size after transient transfection and sorting, cells were seeded in 60 mm dishes (2.5 ϫ 10 5 cells/plate), transfected the following day at 80% confluence using 0.5 g pEGFP C1 (Clontech) plus 2 g of plasmids encoding p110 caax or p70S6K (58), and incubated overnight. The cells were replated in 10-mm dishes, incubated alone or in the presence of rapamycin (20 nM, 72 h), harvested, and sorted for GFP expression. Forward scatter profiles were analyzed by live cell flow cytometry using a Becton Dickinson fluorescence-activated cell sorter.
To determine cell size in stable transfected cell lines, the cells were maintained in exponential growth, alone or in the presence of rapamycin (20 nM, 4 days). The cell diameters and volumes were determined using a particle size counter (CASY, Schä rfe System).

PI3K Deregulation Alters Cell
Growth-We previously examined the consequences on cell cycle progression of interfering with physiological PI3K activation kinetics by expressing different PI3K forms (27). These studies indicated that enhanced PI3K activation accelerates cell cycle entry, whereas decreased PI3K activation reduces this transition, supporting the role of PI3K in cell cycle entry. PI3K activation must nonetheless be transient to allow completion of the cell cycle, because expression of a constitutive active PI3K or PKB mutant deregulated forkhead transcription factor activity throughout the cell cycle, impairing mitotic progression (27).
Here we analyze whether PI3K regulates cell growth in exponentially dividing mammalian cells. We examined growth in NIH 3T3 cell lines expressing p110 caax , a constitutively active p110 catalytic subunit mutant (57), or p65 PI3K , a mutant of the PI3K p85␣ regulatory subunit that binds to p110 and enhances its activation by growth factors (56). We also studied cell lines expressing recombinant p85␣ at levels double those of the endogenous protein; this modification reduces the magnitude of endogenous p110 activation (56,57). The cells were maintained in exponential growth and labeled for short periods with [ 35 S]Met/Cys to compare protein synthesis rates. A 30-min labeling period was adequate to obtain sufficient labeling without saturation; Fig. 1 illustrates a representative experiment and quantification of several assays. Whereas p85␣ expression reduced [ 35 S]Met/Cys incorporation, the two activating PI3K mutations, p110 caax and p65 PI3K , increased the protein synthesis rate.
We also compared protein synthesis rates in these cell lines that had been arrested in different phases of the cell cycle. As reported, protein synthesis in control cells was maximal in G 1 (1, 3) and was moderate in G 0 -and G 2 -arrested cells (Fig. 2). Protein synthesis was also low in M phase arrested cells (not shown). G 0 -arrested p65 PI3K -expressing cells showed a higher rate of protein synthesis than control or p85␣expressing cells. This may be due to the modest basal activation of PI3K seen in p65 PI3K cells (56). Nonetheless, p110 caax -cells exhibited a remarkably higher level of protein synthesis than control cells (Fig. 2). In G 1 , both p65 PI3K and p110 caax cells showed a higher protein synthesis rate than controls, whereas [ 35 S]Met/Cys incorporation was lower in p85␣-expressing cells (Fig. 2). Finally, in all cell lines, protein synthesis was lower in G 2 than in G 1 and had a distinct protein labeling pattern but remained higher in p110 caaxexpressing cells (Fig. 2). Thus, compared with controls, p85␣or p65 PI3K -expressing cells showed increased and decreased protein synthesis rates, particularly in G 1 . p110 caax - We postulated that PI3K may control cell growth and cell cycle progression rates in a concerted manner, giving rise to cells that are normal in size but that divide more rapidly or more slowly, depending on the intensity of PI3K activation. We measured the size of the stable cell lines expressing the different PI3K forms by flow cytometry. Both p85␣-and p65 PI3Kexpressing cells showed a size similar to that of NIH 3T3 control cells (Fig. 3). Nonetheless, cells expressing the constitutive active p110 caax mutant were larger than controls (Fig. 3). We also analyzed cell diameter and volume using a particle size counter and found that only p110 caax cells showed a statistically significant volume and diameter difference compared with controls (Table I). We conclude that alteration of the magnitude of PI3K activation in a transient manner does not modify cell size; in contrast, constitutive activation of PI3K increases cell size.
PI3K Deregulation Alters Cell Division Time-We next examined the cell division time. Stable cell lines were seeded at similar density, and the division rate was calculated by cell counting at different time points after the initiation of culture. p65 PI3K protein expression reduced doubling time, whereas increased expression of the p85␣ form increased t1 ⁄2 (Fig. 4). Cells expressing mutants that alter the magnitude but not the transient kinetics of PI3K activation are thus able to alter cell cycle progression in concert with cell growth, inducing variations in t1 ⁄2 without significantly altering cell size. In contrast, cells expressing p110 caax , which exhibit sustained PI3K activity, show a t1 ⁄2 similar to that of wild type cells (Fig. 4), as well as a larger size (Fig. 3); this suggests that cell growth and cell cycle progression are not coordinated in these cells. Similar results were obtained (not shown) using CTLL2 cells expressing different PI3K forms (56,58).
p110 caax Inhibits Down-regulation of Pathways Controlling Cell Growth-To analyze the mechanisms by which PI3K deregulation affects cell growth, we examined the activation kinetics of the PI3K effector p70 S6K in p110 caax cells. p70 S6K activation is a multistep process that begins with phosphorylation of pseudosubstrate residues (Ser 411 , Ser 418 , Thr 421 , and Ser 424 ), followed by phosphorylation of Thr 389 (44,45). Once Thr 389 is phosphorylated, the enzyme is susceptible to Thr 229 phosphorylation by phosphoinositide-dependent kinase-1 kinase (48,49,55). Thr 389 phosphorylation is considered a limiting step and was analyzed as an indicator of p70 S6K activation in p110 caax and control cells at distinct cell cycle phases. Phosphorylation of Thr 389 was increased in G 1 after serum addition and was low in G 2 and M phase arrested control cells (Fig. 5A). In contrast, p110 caax -expressing cells retained a low level of p70 S6K Thr 389 phosphorylation in G 0 , and Thr 389 remained phosphorylated in G 2 and M. Similar results were obtained when we examined Thr 421 /Ser 424 phosphorylation (Fig. 5A). Thr 389 phosphorylation was transient in the p65 PI3K and p85␣ stable cell lines (not shown). p110 caax expression thus induces prolonged p70 S6K activation.
In addition to regulating p70 S6K activation, the PI3K/PKB pathway also regulates mTOR by controlling TSC2 phosphorylation (30 -32, 43, 59). mTOR in turn regulates 4EBP1 and p70 S6K phosphorylation (36,37,39). Thus, although p70 S6K was deregulated in p110 caax cells (Fig. 5A), this effect may reflect an mTOR defect or a direct effect of PI3K on p70 S6K activity, because PI3K also controls p70 S6K by mTOR-independent mechanisms (47). To analyze whether mTOR activity is altered in p110 caax cells, we thus examined 4EBP1 phosphorylation. We found a more slowly migrating, hyperphosphorylated 4EBP1 species in control cells following serum stimulation (G 1 ) (Fig. 5A). This species was nonetheless already found in serum-starved p110 caax cells and was more clearly detectable in p110 caax cells than in controls in all cell cycle phases (Fig.  5A). This suggests that p110 caax expression affects mTOR activation. These results show that deregulation of PI3K affects mTOR and p70 S6K activity.
To analyze whether p70 S6K deregulation was exclusively a consequence of defective mTOR inactivation, we overexpressed TSC1 and TSC2, which inhibit mTOR (30 -32). Transfection of the exogenous TSC1/2 complex in p110 caax cells reduced 4EBP1 mobility as well as p70 S6K activation levels in G 1 (Fig. 5B) but did not correct the prolonged activation kinetics of p70 S6K in G 2 /M (Fig. 5B). This supports mTOR deregulation as a contributory mechanism to cell mass increase in p110 caax cells. In addition, even when mTOR is inhibited by TSC1/2 expression, p110 caax induces prolonged p70 S6K activation, reflecting a direct PI3K effect on p70 S6K activation that may also contribute to increasing the size of p110 caax cells.  Enhanced Activity of p70 S6K and mTOR Mediates p110 caax Cell Size Increase-To examine whether increased p110 caax cell size was a consequence of enhanced p70 S6K and mTOR activation, we inhibited these enzymes using rapamycin (33). Control and p110 caax stable transfectants were cultured alone or with rapamycin, and their volumes were measured in a particle size counter (Table II). Rapamycin decreased the volume of control cells moderately (ϳ10%) and that of p110 caax cells more intensely (Ͼ20%) ( Table II). We also measured the volume of stable transfectants of the p70 S6K mutant D3E p70 S6K, an activating mutation with acidic substitutions in the pseudosubstrate region residues (Ser 411 , Ser 418 , Thr 421 , and Ser 424 ) but whose activity requires Thr 389 and Thr 229 phosphorylation, remaining sensitive to rapamycin (55). As for p110 caax cells, the D3E p70 S6K-expressing cells were larger, and their size decreased by ϳ20% following incubation with rapamycin (Table II).
In an alternative approach, the cells were transiently transfected with a vector encoding GFP and either a control vector or cDNA encoding p110 caax . The cells were then incubated alone or in the presence of rapamycin, and GFP-positive and -negative cells were isolated by cell sorting (transfection efficiency, ϳ60%). p110 caax -transfected cells were larger than control cells (Fig. 6). Nonetheless, control and p110 caax cells were similar in size when incubated with rapamycin (Fig. 6). The cells were also transfected with cDNA encoding p70 S6K, which gave rise to larger cells; this phenotype was also attenuated by rapamycin addition (Fig. 6). Similar results were obtained using the constitutive active p70 S6K mutant D3E p70 S6K (not shown). As the size of p110 caax cells decreases upon inhibition of p70 S6K and mTOR, these results indicate that PI3K increases cell growth by affecting mTOR and p70 S6K regulation. Our observations support the hypothesis that transient variations in the magnitude of PI3K activation modify growth and cell cycle progression rates in concert. In contrast, sustained PI3K activation deregulates cell growth machinery throughout the cell cycle, uncoupling the protein synthesis rate from cell cycle progression rates, giving rise to larger cells. DISCUSSION The observations presented show that alteration of endogenous PI3K activation by expression of PI3K-interfering forms (p65 PI3K , p85␣, or p110 caax ) affects the rate of cell growth in dividing mammalian cells. Moreover, expression of p65 PI3K or p85␣, which induces transient enhancement or reduction in the magnitude of PI3K activation, alters cell growth without significantly modifying cell size (Fig. 3). This shows that transient changes in the intensity of PI3K activation modify cell cycle progression in concert with cell growth rates. In fact, the regulated increase in PI3K activation induced by p65 PI3K accelerated cell division (decreased t1 ⁄2 ), whereas the p85␣-triggered reduction in the magnitude of PI3K activation delayed cell division (Fig. 4). These observations suggest that cells sense the magnitude of PI3K activation and establish a cell cycle progression rate that is proportional to the cell growth rate, ensuring that daughter cells maintain appropriate cell size. The critical role of PI3K in the control of cell growth during cell division is supported by the observation that expression of a deregulated, constitutive active PI3K mutant impairs coordination of these two processes, inducing a cell size increase. This is the first description that links transient PI3K activation to the concerted regulation of cell growth and cell cycle progression rates during cell division in mammals.
Expression of the constitutive active PI3K mutant p110 caax induces enlargement in cell size (Fig. 3). This mutant increases the protein synthesis rate (Fig. 1) and accelerates cell cycle entry but retards G 2 /M progression and cell cycle exit (27). A partial explanation for the lack of balance between cell growth and cell cycle progression rates in p110 caax cells may thus be the delayed transition through G 2 /M. We nonetheless show FIG. 5. Constitutive PI3K activation impairs p70 S6K downregulation. A, lysates (50 g) from control and p110 caax -expressing cells arrested at the indicated cell cycle phases were resolved by SDS-PAGE. The gels were analyzed in Western blot using anti-phospho-Thr 389 -p70 S6K, anti-phosphoThr 421 /Ser 424 -p70 S6K, anti-p70 S6K, and anti-4EBP1 antibodies. B, NIH 3T3 cells cultured in DMEM with 10% CS were transiently transfected with a vector encoding p110 caax alone or in combination with vectors encoding TSC1 and TSC2. At 24 h post-transfection, the cells were arrested in cell cycle phases as indicated, and the lysates (50 g) were examined in Western blot using anti-phosphoThr 389 -p70 S6K, anti-p70 S6K, anti-TSC1, anti-TSC2, and anti-4EBP1 antibodies. D3Ep70 S6K cDNA NIH 3T3 cell lines, stably transfected as indicated, were cultured in DMEM/10% CS alone or in the presence of rapamycin (20 nM) and harvested, and cell diameter and fluid volume were measured using a particle size counter. The mean of 10 determinations is shown. a ⌬ volume was calculated by subtracting the mean volume of each rapamycin-treated cell line from the mean volume of the same untreated cell line.

PI3K Controls Cell Division Time
that constitutive activation of PI3K also interferes with correct down-regulation of cell growth-promoting pathways. Accordingly, incubation with PI3K inhibitors reduces cell growth (35) and impairs cell cycle entry (24,27,35). We examined mTOR and p70 S6K and found that p110 caax expression extended p70 S6K activation kinetics to the G 2 /M phases and induced hyperphosphorylation of the mTOR effector 4EBP1 in G 0 . PI3K activation must thus be transient to allow correct control of cell cycle progression (27) and cell growth throughout the cell cycle.
The reduction in p110 caax cell size following rapamycin inhibition of mTOR and p70 S6K activity suggests that these PI3K effectors control growth in dividing cells. PI3K may regulate cell growth by additional mechanisms. This possibility is supported by the behavior of NIH 3T3 p65 PI3K -expressing cells, in which p70 S6K is transiently triggered and down-regulated, but whose activation levels are lower than in controls (not shown). This concurs with our previous observations showing that p85, but not p65 PI3K , forms a complex with p70 S6K and mTOR that is required for p70 S6K activation (58). Stable NIH 3T3 p65 PI3K cells express similar levels of p65 PI3K and of endogenous p85 (56), which accounts for the moderate p70 S6K activation observed in these cells. Because p65 PI3K cells have a higher protein synthesis rate and reduced p70 S6K activation; p70 S6K does not appear to be the main effector mediating enhanced cell growth in these cells. Activation of the p70 S6K 2 isoform (51, 60), TSC inactivation (30 -32), or an as yet undescribed mechanism may cooperate with p70 S6K to enhance cell growth in response to PI3K activation. In addition, it was recently reported that PI3K enhances 5Ј TOP mRNA translation independently of p70 S6K (50). We found that p65 PI3K -expressing cells have a higher proportion of rpL32 mRNA (5Ј TOP) (50) in heavy polysomes than control cells (not shown), suggesting that 5Ј TOP translation is enhanced in these cells.
The observations presented suggest that PI3K has an essential role in the concerted regulation of cell growth and cell cycle progression. Previous observations in yeast illustrated that inhibition of cell growth blocks cell cycle entry, whereas inhibition of cell cycle progression allows growth to continue (6). This shows that cell cycle entry is linked to the cell growth process. As to the signaling pathways that control cell growth in yeast, no class I PI3K homologues have been found in this organism; TOR function in control of cell growth is nonetheless conserved from yeast to mammals (61).
In the fruit fly Drosophila melanogaster, disruption of cell cycle regulatory genes (dE2F and cdc2) results in cell cycle arrest at a larger cell size (62,63). This shows that growth without division can also be observed in this organism, but division requires growth. With regard to the pathways that control cell growth and cell division, mutations in Inr, dp110, dIRS (Chico), dPTEN, and dRas affect cell growth and cell cycle simultaneously, whereas mutations in dTOR, d4EBP, and dS6K affect only cell size (reviewed in Refs. 4, 7-10, and 64). In addition, deletion of the negative regulator TSC1 (which participates in negative control of TOR) affects cell size (4,5,65). PI3K regulates the TSC complex and TOR (7, 30 -32), suggesting that one signaling branch downstream of PI3K regulates cell growth, and the other controls cell cycle progression. dAKT appears to regulate only cell size, suggesting that AKT lies in the growth branch of the PI3K pathway in flies (66). Another difference compared with mammals is that dS6K appears to lie in a pathway different from that of dPI3K (67), although the dPI3K pathway still controls cell growth and cell division. Most of the mutations mentioned above were described in the Drosophila wing imaginal disc, in which cell growth and cell cycle increase in parallel. The study of these processes in Drosophila has the additional difficulty that organ size is subject to internal regulatory mechanisms (reviewed in Refs. 4, 5, 7, and 64). Moreover, division is not coupled to growth in some organs; for example, in the pupal stage, postmitotic cells in the eye grow without undergoing division (4). This explains the observation that flies carrying a dp110 mutation exhibit a cell growth and cell division phenotype in the wing but only a growth phenotype in the eye (8).
In mammals, inhibition of cell growth also blocks cell cycle entry (3), although growth continues following inhibition of cell cycle entry (35). This also shows that growth in mammals can be separated from the cell cycle but that the cycle is linked to growth. Cell growth in mammals requires PI3K and TOR activities; in fact, expression of the p16 cell cycle inhibitor blocks the cycle in G 1 , but the resulting cells are larger (35). This cell size increase is partially blocked by TOR inhibition and even more clearly by PI3K inhibitors, illustrating the relevance of PI3K and TOR in cell growth control (35). Nonetheless, only PI3K, but not TOR, appears to mediate the concerted regulation of cell growth and cell cycle (Fig. 4). In contrast to the ability of PI3K mutants to regulate cell cycle progression and growth, activation of the mTOR pathway does not trigger cell division (5,9,35). PI3K is thus the first signaling pathway reported to link both processes. Two routes would be induced by PI3K, one branch involved in triggering cell cycle entry and the other in promoting cell growth. The branch regulating cell growth includes mTOR and its effectors, among others (5, 28 -32, 50). Regulation of cell cycle entry downstream of PI3K requires Rac, Cdc42, and PKB activation, which affects cyclin/CDK activities or stability (4, 24 -27, 62, 68). Nonetheless, PI3K involvement in coordinating cell growth and cell division was not observed in mice expressing constitutive active forms of PI3K/PKB in the heart (52, 53). Expression of constitutive active PI3K/PKB in post-mitotic FIG. 6. Incubation with rapamycin reduces p110 caax cell size. NIH 3T3 cells cultured in DMEM with 10% CS were transiently transfected with a PSG5 empty vector, a vector encoding p110 caax , or a vector encoding p70S6K, all in combination with a vector encoding GFP (4:1). At 24 h post-transfection, the cells were plated alone or with rapamycin (20 nM) and incubated for 72 h. The cells were harvested and sorted for GFP expression and then analyzed by live cell flow cytometry. Overlaid forward scatter profiles of GFP(ϩ) and GFP(Ϫ) cells are shown. Ctr, control. cells (cardiomyocytes) may mask the contribution of PI3K to triggering cell division (52,53). In contrast, the phenotype of mice expressing the transient p65 PI3K mutation as a transgene in T cells and retina revealed the contribution of the PI3K route in cell division in vivo (69,70).
The mechanism by which PI3K exerts concerted regulation on cell cycle progression and cell growth is incompletely understood. Induction of cell growth and cell cycle entry may simply occur in parallel. Because both cell growth and cell cycle entry are regulated by PI3K, the magnitude of PI3K activation may determine the extent of these processes. It is also possible that translation of a specific cell cycle entry component is sensitive to the availability of the translation machinery. In yeast, G 1 cyclin (Cln3) protein expression is highly dependent on the levels of the translation initiation complex, such that Cln3 levels define whether a cell has sufficient translation machinery to enter the cell cycle (71). It has also been shown that overexpression of cyclin D in Drosophila triggers cell growth (4), supporting the possibility that cyclin E rather than cyclin D acts as a growth sensor in this organism. In mammals, PI3K contributes specifically to inducing cyclin D and E synthesis and regulates E2F induction (24,72). Nonetheless, whether or not translation of mammalian G 1 cyclins mRNAs depends on PI3K-controlled translation machinery remains to be determined.
The fact that PI3K has a crucial role linking cell growth and cell cycle entry does not imply that this enzyme is in itself sufficient for either of these processes. For instance, 5Ј cap translation, which accounts for 85% of total translation, requires mTOR activation. Nonetheless, mTOR activity requires not only TSC inactivation by PI3K/PKB (30 -32) but also appropriate ATP and nutrient levels (40,41). Translation initiation is also regulated by mitogen-activated protein kinase-dependent pathways (73). For cell cycle entry, other signaling cascades in addition to PI3K also modulate cyclin D expression (74,75). The requirement for signals other than PI3K to induce cell growth or division explains why some receptors that activate PI3K can induce cell growth, whereas others trigger cell division (4,5,50,72). It is thus possible that the pathways that act in conjunction with PI3K to trigger cell growth and cell cycle entry also have a role in coordinating these two processes. Nonetheless, the concerted modification of cell cycle progression and cell growth rates observed after genetic alteration of PI3K points to this early signal as a central player for correct coordination.
In conclusion, alteration of t1 ⁄2 without modification of cell size or cell cycle profiles in p65 PI3K and p85␣-expressing cells illustrates the central role of PI3K in the concerted regulation of cell growth and cell cycle progression. The upstream position of PI3K in cell growth-and cell cycle-controlling signaling pathways makes this regulation possible. Coordination of both processes requires PI3K activation to be transient.