Transduced p16^INK4a Peptides Inhibit Hypophosphorylation of the Retinoblastoma Protein and Cell Cycle Progression Prior to Activation of Cdk2 Complexes in Late G₁

Introduction

Progression of eukaryotic cells from the early and mid G₁ phase of the cell cycle through the G₁ restriction point, into late G₁ phase, through the G₁-S phase transition and then into S phase requires the concerted activities of several members of the Cdk 4 family, i.e., cyclin D:Cdk4/6, cyclin E:Cdk2, and cyclin A:Cdk2 complexes (1, 2, 3) . One important substrate of G₁ cyclin:Cdk complexes is the product of the retinoblastoma tumor suppressor gene, pRb, a negative regulator of early G₁ phase cell cycle progression, which contains 16 putative Cdk consensus phosphorylation sites (1) . pRb exists in two general phosphorylated forms in G₁. In early and mid G₁, pRb is present as an active, hypophosphorylated form that associates with cellular transcription factors and contains a ∼1:1 molar ratio of phosphate to pRb (4, 5, 6) . At the late G₁ restriction point, pRb becomes initially inactivated by hyperphosphorylation and contains a ∼10:1 molar ratio of phosphate to pRb, resulting in the release of transcription factors (1 , 4, 5, 6) . Thus, pRb appears to be both activated and inactivated in a nonredundant fashion by use of Cdk phosphorylation sites coupled to phosphate molar ratios.

In human oncogenesis, there is a strong selection for genetic alteration of one or more members of the p16^INK4a, cyclin D:Cdk4/6, and pRb pathway involved in regulating G₁ cell cycle progression (7) . However, the timing and duration of cyclin D:Cdk4/6 complexes involved in phosphorylating pRb and the requirement for G₁ cell cycle progression are presently poorly understood. Because of inherent limitations of cellular manipulation by methods such as transfection and viral vector infection, such as lag time for gene expression and low efficiencies, these experiments cannot address the critical issue of the temporal activity required by cyclin D:Cdk4/6 complexes for pRb phosphorylation and G₁ cell cycle progression. Therefore, to precisely investigate the kinetic requirement for early G₁ phase cell cycle progression by cyclin D:Cdk4/6 complexes, we transduced p16^INK4a peptidyl mimetics directly into ∼100% of cells and assayed for pRb phosphorylation and kinetics of cell cycle progression. Transduction of p16 mimetics results in rapid (>20 min) inactivation of Cdk4/6 complexes and thus allows for an analysis of Cdk4/6 kinetic requirements. We report here that cyclin D:Cdk4/6 complexes perform the activating hypophosphorylation of pRb in early and mid G₁ which is nonredundant with the inactivating hyperphosphorylation of pRb by cyclin E:Cdk2 complexes in late G₁. In addition, Cdk4/6 activity is required for early G₁ phase cell cycle progression up to, but not beyond, activation of cyclin E:Cdk2 complexes at the restriction point and into late G₁ phase.

Materials and Methods

Cell Culture and Flow Cytometry Analysis.

Human HaCaT keratinocytes were maintained as described (6) . For G₁ cell cycle arrest, HaCaT cells were contact inhibited by plating at high density (6 × 10⁶/10-cm dish) for 36–40 h in 10% FBS, trypsinized, and replated at low density (1.5 × 10⁵/well of a 6-well dish) and assayed for cell cycle position at various time points up to 30 h postreplating or treated with TAT-p16 peptides at 5, 10, and 15 h postreplating. DNA content FACS analysis was performed as described (6) .

Labeling and Immunoprecipitations.

G₁-arrested, contact-inhibited HaCaT cells were pretreated with TAT-p16 peptides for 1 h and then labeled in the presence of TAT-p16 peptides for 4 h with 3–5 mCi of [³²P]orthophosphate (ICN Biomedicals) per 10-cm dish or with 250 μCi [³⁵S]methionine (NEN) as described (6) . Cellular lysates were prepared, and pRb was immunoprecipitated by addition of G99-549 anti-pRb antibodies (PharMingen) that recognize only the fast migrating, un- and hypophosphorylated forms of pRb as described (6) . Immune complexes were collected on protein A-Sepharose (Pharmacia), washed three times, resolved by SDS-PAGE, transferred to nitrocellulose filters, and analyzed by phosphorimaging (Molecular Dynamics). After PhosphorImager analysis, nitrocellulose filters were immunoblotted with anti-pRb antibodies as described (6) .

Kinase Assay.

Rabbit anti-Cdk2 (Santa Cruz Biotechnology) immunoprecipitates were washed three times with ELB, followed by washes two times with kinase buffer [50 mm HEPES (pH 7.0), 10 mm MgCl₂, 1 mm DTT, and 1 μm unlabeled ATP] and suspended in 25 μl of kinase buffer plus 100 μCi of [γ-³²P]ATP (Amersham; 6000 Ci/mmol) plus 2 μg of histone H1 (Sigma) substrate. Reactions were incubated for 30 min at 30°C, stopped by addition of 2× SDS buffer, separated on SDS-PAGE, and analyzed by phosphorimaging (Molecular Dynamics). Equal amounts of rabbit antimouse antibodies were used as negative controls.

TAT-p16 Peptides.

Thirty-two-mer TAT-p16 peptides were synthesized so that each contained an NH₂-terminal 11-mer TAT protein transduction domain (single-letter code, YGRKKRRQRRR; Ref. 8 ) followed by a glycine residue and either a 20-mer wild-type p16 sequence (WT, DAAREGFLATLVVLHRAGAR; Ref. 9 ) or a charge-match control sequence (MUT, ARGRALTAHVDRLGEFVAAL). After synthesis and purification, peptides were resuspended in water. FITC-labeled TAT-p16 peptides were generated by fluorescein labeling (Pierce), followed by gel filtration on a S-12 column attached to an fast protein liquid chromatography (Pharmacia).

Results1

Transduction of p16 Peptidyl Mimetics into Cells.

To focus on the question of exactly when during the progression of G₁ phase of the cell cycle cyclin D:Cdk4/6 complexes are required, we chose to introduce peptidyl mimetics of p16^INK4a, a negative regulator of Cdk4/6 (2) , by the rapid method of protein transduction (8 , 10) . Treatment of cells with peptides and proteins containing the protein transduction domain from HIV TAT protein results in a rapid transduction into ∼100% of cells in a given population (primary or transformed cells) in a receptorless fashion (6 , 8 , 10 , 11) . In addition, because of its concentration dependency, TAT-mediated transduction results in a near equivocal intracellular concentration of the transduced protein from cell to cell in the population. Previously, Fahraeus et al. (9) demonstrated that a 20-amino acid peptidyl mimetic of p16 was capable of binding to and inactivating Cdk4/6 in vitro. We synthesized 32-mer peptides that consisted of an NH₂-terminal TAT protein transduction domain (11 amino acids; Ref. 8 ), followed by a glycine residue for free bond rotation and a COOH-terminal 20-mer of either functional wild-type p16^INK4a or charge-matched control sequences (Fig. 1A) ⇓ .

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Characterization of TAT-p16 peptides. A, structure of TAT-p16 wild-type (WT) and mutant (MUT) peptides represented in single-letter amino acid code. B, FACS analysis of FITC-labeled TAT-p16 peptide 20 min after addition to cells. Note narrow peak width of transduced cells. Control is FITC labeling reaction in the absence of peptide. C, asynchronous HaCaT keratinocytes were transduced with increasing concentrations of TAT-p16 wild-type (WT) or mutant (MUT) peptides for 30 h and analyzed for cell cycle position by propidium iodide staining for DNA content by FACS.

To monitor transduction into cells, TAT-p16 peptides were labeled with FITC and added to the media of cells. Flow cytometry analysis (FACS) of treated cells demonstrated that TAT-p16-FITC peptides transduced rapidly into ∼100% of cells, achieving maximum intracellular concentrations in <20 min (Fig. 1B) ⇓ . Confocal microscopy analysis confirmed transduction of TAT-p16-FITC peptides into ∼100% cells and revealed an intracellular location of the transduced peptide and not mere attachment to the cellular membrane (data not shown; Ref. 8 ). Thus, consistent with our TAT fusion proteins published previously (6 , 8 , 10 , 11) , TAT-p16 peptides transduce into ∼100% of cells in a rapid, concentration-dependent fashion.

To functionally test the ability of TAT-p16 peptides to elicit a G₁ phase cell cycle arrest, we used human HaCaT keratinocytes as a model cell culture system because of their sensitivity to p16^INK4a-mediated cell cycle arrest (6) . Asynchronous HaCaT keratinocytes were treated with 10, 50, or 100 μm TAT-p16 wild-type or control peptides for 30 h and then analyzed for cell cycle position by DNA content and FACS analysis (Fig. 1C) ⇓ . Treatment of HaCaT cells with TAT-p16 wild-type peptide resulted in a significant G₁ phase cell cycle arrest, whereas the TAT-control peptide had minimal effects on cell cycle position. Cell cycle arrest prior to the late G₁ restriction point has been shown to result in loss of hyperphosphorylated forms of pRb (4 , 6 , 12 , 13) . Consistent with an early G₁ cell cycle arrest, anti-pRb immunoblot analysis of asynchronous HaCaT cells treated with TAT-p16 peptides showed loss of hyperphosphorylated pRb (data not shown). Taken together, these observations demonstrate that TAT-p16 wild-type peptides rapidly transduce into ∼100% of cells and retain the previously associated properties of p16 to elicit an early G₁ phase cell cycle arrest.

Transduction of p16 Peptide into Cells Inhibits pRb Hypophosphorylation.

On denaturing SDS-PAGE immunoblots, unphosphorylated and hypophosphorylated pRb comigrate (6) and hence, are indistinguishable. Therefore, to analyze the influence of accumulated TAT-p16 peptides and hence, resultant inactivation of Cdk4/6 complexes on pRb hypophosphorylation, G₁ arrested contact-inhibited HaCaT cells present in 10% FBS, containing only hypophosphorylated pRb and no hyperphosphorylated pRb, were treated with 100 μm TAT-p16 wild-type or mutant peptides for 1 h, followed by the addition of [³²P]orthophosphate for 4 h. pRb was immunoprecipitated from cellular lysates with anti-pRb antibodies, resolved by SDS-PAGE, transferred to a filter, and analyzed for pRb ³²PO₄ content (Fig. 2A) ⇓ . The same filter was then normalized for pRb protein levels by anti-pRb immunoblot analysis (Fig. 2B) ⇓ . Treatment with TAT-p16 wild-type peptides resulted in a marked loss of pRb hypophosphorylation and the appearance of unphosphorylated pRb (Lane 2). In contrast, cells treated with TAT-p16 control peptides retained hypophosphorylated pRb, as did untreated control cells (Lanes 1 and 3). These observations suggest that in cells containing physiological concentrations of cyclin D:Cdk4/6 complexes, pRb is only hypophosphorylated in vivo and not hyperphosphorylated.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Transduction of TAT-p16 peptide results in loss of pRb hypophosphorylation. G₁ arrested, contact-inhibited HaCaT cells were treated with either 100 μm TAT-p16 wild-type (WT) or mutant (MUT) peptides for 1 h and then [³²P]orthophosphate labeled for 4 h in the presence of TAT-p16 peptides. pRb was immunoprecipitated, transferred to a filter, analyzed by phosphorimaging for ³²P content (A), and then the same filter was probed by anti-pRb immunoblot analysis to normalize for pRb protein levels (B). Control (ctrl) cells were untreated. Note the loss of pRb hypophosphorylation, appearance of unphosphorylated pRb, and their comigration on SDS-PAGE.

TAT-p16 Peptides Arrest Cells before Activation of Cdk2 Complexes in Late G₁.

To directly ascertain the kinetic requirement of cyclin D:Cdk4/6 activity in promoting G₁ phase cell cycle progression, we sought to transduce TAT-p16 peptides into synchronized HaCaT cells at various time points. HaCaT cells were contact arrested in early G₁ by plating at high density for 36 h in 10% serum. Cells were then released from arrest by replating at low density and analyzed for pRb phosphorylation status and Cdk2 kinase activity at various time points (Fig. 3A) ⇓ . Immunoprecipitation of pRb from [³²P]orthophosphate-labeled cellular lysates and immunoblot analysis showed that pRb was hypophosphorylated at replating (Fig. 2 ⇓ , t = 0). pRb remained hypophosphorylated at 5 and 10 h postreplating and initially became inactivated by hyperphosphorylation at 15 h postreplating (Fig. 3A ⇓ , top panel). Anti-Cdk2 immunoprecipitation-kinase analysis first detected Cdk2 activity, likely cyclin E:Cdk2 complexes, at 15 h postreplating (Fig. 3A ⇓ , bottom panel). On the basis of DNA FACS analysis, replated HaCaT cells progress from late G₁ phase into S phase at >22 h (data not shown). Thus, inactivation of pRb by hyperphosphorylation and activation of Cdk2 complexes occur concomitantly in late G₁.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Kinetics of TAT-p16-mediated G₁ arrest. A, contact-arrested HaCaT cells were released from the G₁ block by replating at low density, followed for pRb phosphorylation status by anti-pRb immunoblot analysis (top panel) and for Cdk2 activity by anti-Cdk2 immunoprecipitation-kinase analysis using histone H1 as a substrate (bottom panel). Hyperphosphorylated pRb first appears at 15 h postreplating, concomitant with activation of Cdk2 complexes. B, contact-arrested and released HaCaT cells from above (A) were transduced with increasing concentrations of TAT-p16 wild-type peptides at 5, 10, and 15, h postreplating and then analyzed for cell cycle progression by FACS analysis for DNA content. Cells treated at 5 and 10 h retain the ability to be arrested by TAT-p16 peptides; however, at 15 h, TAT-p16 peptides are unable to effect a G₁ arrest, consistent with activation of Cdk2 complexes. Control cells were released from G₁ arrest but untreated with peptides.

We next treated G₁ arrested, contact-inhibited HaCaT cells that were then released from G₁ arrest by replating at low density and treated with 10, 50, or 100 μm TAT-p16 wild-type peptide at 5, 10, and 15 h postreplating and analyzed for cell cycle position at 30 h postreplating by DNA content and FACS analysis (Fig. 3B) ⇓ . Transduction of TAT-p16 wild-type peptides into the synchronized cells were capable of eliciting a G₁ cell cycle arrest when transduced at 5 and 10 h postreplating. However, TAT-p16 wild-type peptides were unable to effect a G₁ arrest when transduced into cells at 15 h postreplating, consistent with the appearance of active Cdk2 complexes and hyperphosphorylated pRb. Taken together, these observations directly demonstrate, for the first time, that cyclin D:Cdk4/6 complexes are required for early and mid G₁ phase cell cycle progression up to, but not beyond, the point of Cdk2 activation and transition through the restriction point into late G₁.

Discussion

Advancement from the early G₁ phase of the cell cycle through the restriction point and into late G₁ phase requires the activity of several cyclin:Cdk complexes and hyperphosphorylation of pRb at the restriction point (1, 2, 3 , 6 , 11 , 12 , 14) . The requirement for both cyclin D:Cdk4/6 and cyclin E:Cdk2 activity for G₁ phase cell cycle progression has been shown previously (15, 16, 17) . However, the kinetic requirement for cyclin D:Cdk4/6 activity remained unclear. By use of the protein transduction methodology (8) , we were able to rapidly introduce the p16^INK4a-negative regulator of Cdk4/6 into ∼100% of cells in the population in a concentration-dependent fashion. This method presents a superior technique to perform both biological and in vivo biochemical assays on the entire cellular population in precise timing intervals. In addition, cellular manipulation by protein transduction avoids problems associated with transfection of a limited percentage of the cells and unregulated overexpression of cyclins that can lead to nonphysiological conditions within the cell.

We show here that introduction of a p16^INK4a peptidyl mimetic into synchronized keratinocytes by protein transduction results in a G₁ phase cell cycle arrest, provided that Cdk2 complexes have not become activated. In addition, inactivation of Cdk4/6-containing complexes results in the loss of pRb hypophosphorylation and the appearance of the unphosphorylated form of pRb. Thus, cyclin D:Cdk4/6 activity, presumably for hypophosphorylation of pRb (and likely other substrates), is required to progress to the restriction point but not through it. Transition across the late G₁ restriction point appears to require activation of Cdk2 complexes, likely cyclin E:Cdk2 complexes, that inactivate pRb by hyperphosphorylation, causing release of transcription factors from pRb (Fig. 4) ⇓ . These observations are entirely consistent with previous reports demonstrating the constitutive expression and activity of cyclin D:Cdk4/6 complexes in early G₁ of cycling cells (6 , 11 , 18, 19, 20) . In addition, work by Ikeda et al. (21) demonstrated that in G₀, E2F transcription factors are bound by a pRb-related pocket protein, p130. Thus, as the cell progresses from a G₁ arrested state into a G₀ state containing inactive Cdk4/6 complexes, pRb becomes dephosphorylated, allowing p130 access to bind and regulate this family of transcription factors.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Model of cyclin:Cdk complex kinetics in regulating pRb phosphorylation and G₁ phase cell cycle progression. In G₀, p130, a pRb-related pocket protein, binds E2Fs (30). Newly synthesized pRb is unphosphorylated and becomes activated (capable of binding transcription factors) when hypophosphorylated by cyclin D:Cdk4/6 complexes in early G₁. The initial inactivating hyperphosphorylation of pRb occurs by activation of cyclin E:Cdk2 complexes at the restriction point, resulting in release of transcription factors, such as E2Fs, from pRb that result in transcriptional activation. Thus, the functioning of cyclin D:Cdk4/6 and cyclin E:Cdk2 complexes are both nonredundant and diametrically opposed. After degradation of cyclin E from Cdk2 in S phase, cyclin A:Cdk2 complexes maintain pRb hyperphosphorylation into late G₂ phase, followed by cyclin B:CDC2 in M phase.

The data presented here demonstrate that the roles of cyclin D:Cdk4/6 and cyclin E:Cdk2 complexes in advancing early and late G₁ phase, respectively, are nonredundant and in fact diametrically opposed with respect to pRb regulation; cyclin D:Cdk4/6 complexes activate pRb, whereas cyclin E:Cdk2 complexes inactivate pRb. pRb oscillates between hypophosphorylated and hyperphosphorylated forms in every cell cycle. In addition, hypophosphorylated pRb forms associate with cellular transcription factors, such as E2Fs, whereas hyperphosphorylated forms do not (2 , 11 , 17) . Thus, given occupation of up to 16 Cdk sites and a ∼1:1 molar ratio of phosphate to pRb, hypophosphorylated pRb may be comprised of multiple isoforms that individually contain a combination of one or perhaps two phosphates. Indeed, two-dimensional isoelectric focusing of hypophosphorylated pRb shows the presence of >12 hypophosphorylated pRb isoforms in keratinocytes and human peripheral blood lymphocytes. 5 The generation of multiple isoforms of active hypophosphorylated pRb may serve to target subsets of hypophosphorylated pRb isoforms to specific transcription factors and, hence, result in promoter-specific regulation. Future experiments directed at characterizing specific pRb isoforms associated with specific transcription factors in cells that express subsets of cyclin D1, D2, D3:Cdk4, and Cdk6 complexes should help to understand the complex regulatory events that occur in early G₁ and at the restriction point.