PHOSPHORYLATION REGULATES KSR1 STABILITY, ERK ACTIVATION, AND CELL PROLIFERATION

Kinase suppressor of Ras (KSR) is a molecular scaffold that interacts with the components of the Raf/MEK/ERK kinase cascade and positively regulates ERK signaling. Phosphorylation of KSR1, particularly at Ser(392), is a critical regulator of KSR1 subcellular localization and ERK activation. We examined the role of phosphorylation of both Ser(392) and Thr(274) in regulating ERK activation and cell proliferation. We hypothesized that KSR1 phosphorylation is involved in generating signaling specificity through the Raf/MEK/ERK kinase cascade in response to stimulation by different growth factors. In fibroblasts, platelet-derived growth factor stimulation induces sustained ERK activation and promotes S-phase entry. Treatment with epidermal growth factor induces transient ERK activation but fails to drive cells into S phase. Mutation of Ser(392) and Thr(274) (KSR1.TVSA) promotes sustained ERK activation and cell cycle progression with either platelet-derived growth factor or epidermal growth factor treatment. KSR1(-/-) mouse embryo fibroblasts expressing KSR1.TVSA proliferate two times faster and grow to a higher density than cells expressing the same level of wild-type KSR1. In addition, KSR1.TVSA is more stable than wild-type KSR1. These data demonstrate that phosphorylation and stability of the molecular scaffold KSR1 are critical regulators of growth factor-specific responses that promote cell proliferation.


Introduction
Specific signal transduction pathways in mammalian cells are potent regulators of mitogenesis. Aberrant regulation of such pathways can lead to oncogenesis. The protooncogene Ras is a small GTPase that couples signals from extracellullar stimuli to intracellular pathways controlling cell growth, differentiation, and survival (1)(2)(3). Activation of the Raf-MEK 1 -ERK (p42/p44) MAP kinase cascade by Ras modulates cell growth, differentiation, and survival (4)(5)(6).
The Ras-Raf/MEK/ERK pathway mediates signals downstream of a variety of extracellular stimuli. Distinct responses to different stimuli can be generated by modulating the duration of ERK activation. In PC12 cells, NGF induces sustained ERK activation to promote neuronal cell differentiation, whereas EGF induces transient ERK activation to promote cell proliferation (7,8). In contrast, EGF induces transient ERK activation in fibroblasts to promote survival, but not mitogenesis, while PDGF induces sustained ERK activation and promotes S phase entry and cell proliferation (7,9). It is unclear, however, how different extracellular ligands differentially regulate the duration of ERK activation. Different adaptors may be recruited to the PDGF and EGF receptors, which may change the duration or intensity of the signal directed at the Raf/MEK/ERK kinase cascade. As an alternate mechanism, molecular scaffolds regulate the intensity and duration of signaling output in a concentration-dependent manner (10)(11)(12)(13)(14)(15). Molecular scaffolds may also regulate the interaction, subcellular localization, or stability of the signaling components that they bind. 5 blocking KSR1 phosphorylation generated a PDGF-like response to EGF treatment, suggesting that KSR1 phosphorylation mediates ERK signaling specificity to EGF and PDGF stimulation.
In combination, these mutations also inhibited KSR1 turnover. These data suggest that proliferative responses through the Raf/MEK/ERK signaling cascade can be regulated by controlling the phosphorylation and stability of KSR1. AlexaFluor 680 goat anti-mouse IgG (Molecular Probes) or IRdye 800 (Rockland Immunochemicals). Blotted proteins were detected and quantified using the Odyssey infrared imaging system (LI-COR).

ERK Timecourses. Timecourses of ERK activation were performed using the In-Cell
Western blot analysis on the Odyssey system (LI-COR), which allows analysis and quantification of ERK phosphorylation in multiple samples by simultaneously probing fixed cells with antibodies against total ERK and ERK phosphorylated on Thr 202 and Tyr 204 . KSR1 -/-MEFs expressing KSR1 constructs were seeded in triplicate at 1x10 5 cells/well in a 96-well plate (BioSource International, Inc.). Twenty-four hours after plating, cells were serum-deprived for four hours and stimulated with EGF (100 ng/ml) or PDGF-BB (25 ng/ml) for the indicated timepoints. Cells were fixed in 3.7% formaldehyde for 20 min. at room temperature and stained for ERK/phosphoERK (Santa Cruz sc-94, 1:200, Cell Signaling Technology 9106, 1:200) according to manufacturers' instructions and as described previously (18). AlexaFluor 680 antirabbit IgG (Molecular Probes, 1:200) and IRDye 800 anti-mouse IgG (Rockland Immunochemicals, 1:200) were used as secondary antibodies. Plates were scanned using the Odyssey system. Results were analyzed using the In-Cell Western Blot Odyssey software (LI-COR) and Microsoft Excel.

Proliferation Assays and BrdU incorporation.
To assay growth rate, cells were seeded in a 12-well plate at 2.5 x 10 4 cells per well. Cells from triplicate wells were trypsinized, harvested, and counted each day for six days using a Beckman Coulter Counter.
Cells were assayed for cell cycle progression by monitoring incorporation of 5-bromo-2deoxyuridine (BrdU, BD Biosciences). Cells were plated at 1.6x10 4 cells per well on coverslips.
Four hours after plating, cells were serum-starved for 72 hours to induce quiescence. Cells were then stimulated with EGF (100 ng/ml), PDGF-BB (25 ng/ml), or serum (10% fetal bovine serum) in the presence of BrdU (10 µM) for 20 hours. The cells were fixed in methanol:acetone (1:1) at -20ºC for 8 min and were rehydrated with TBS for 10 min at room temperature. After treatment

Results
Phosphorylation of KSR1 modulates the duration of ERK activation. KSR1 expression promotes ERK activation (18). We hypothesized that phosphorylation of KSR1 contributes to its role in ERK activation by EGF and PDGF. To investigate this possibility, we expressed wild-type and mutated forms of KSR1 in KSR1 -/-MEFs (16) at levels comparable to those in wild-type MEFs (Fig. 1).
KSR1 -/-MEFs expressing ectopic KSR1 exhibited sustained ERK activation with PDGF stimulation ( Fig. 2A). ERK activation peaked at 5 minutes after treatment, but showed significant phosphorylation out to 90 minutes of stimulation. In contrast, upon treatment with EGF, the cells exhibited transient ERK activation (Fig. 2B), with ERK phosphorylation peaking at 5 minutes of stimulation but falling rapidly back to basal levels. KSR1.SA caused sustained ERK activation in cells treated with PDGF ( Fig. 2A). In cells treated with EGF, KSR1.SA promoted a burst of ERK activation that was sustained above the level obtained in cells expressing wild-type KSR1 (Fig. 2B). These data suggest that Ser 392 dephosphorylation enhances both the intensity and duration of ERK activation. ERK activation was decreased in both intensity and duration in KSR1 -/-MEFs expressing KSR1.TV upon treatment with either PDGF ( Fig. 2A) or EGF (Fig. 2B). These data demonstrate that blocking phosphorylation of this putative ERK phosphorylation site alone does not facilitate ERK phosphorylation.
To determine which KSR1 mutation was dominant (SA or TV), we expressed KSR1 mutated at both Thr 274 and Ser 392 (TVSA) in KSR1 -/-MEFs (Fig. 1). Cells expressing KSR1.TVSA demonstrated more sustained ERK activation upon PDGF stimulation ( Fig. 2A) compared to wild-type KSR1 expressed at the same level. Cells expressing KSR1.TVSA also exhibited markedly prolonged ERK activation upon stimulation with EGF (Fig. 2B). Therefore, the prolonged ERK activation generated by KSR1.SA was dominant to the decreased ERK activation induced by KSR1.TV. Furthermore, combined mutation of Ser 392 and Thr 274 synergistically promoted sustained ERK activation upon treatment with either EGF or PDGF.
These data suggest that the phosphorylation of KSR1 on Thr 274 and Ser 392 can modulate the time course of ERK activation in response to specific growth factors.
Phosphorylation of KSR1 modulates S-phase entry. In fibroblasts, sustained ERK activation allows quiescent cells to re-enter the cell cycle, whereas transient ERK activation does not (7,9). The phosphorylation of Thr 274 and Ser 392 on KSR1 is important for modulating the duration of ERK activation ( Fig. 2A and 2B). Therefore, the ability of KSR1 phosphorylation to regulate S-phase entry was measured by BrdU incorporation into DNA.
In comparison to untreated cells, fetal bovine serum and PDGF, but not EGF, were sufficient to promote S-phase entry in at least 90% of cells expressing ectopic KSR1 (Figs. 3 and S1). Both KSR1.SA and KSR1.TV were similarly competent to promote cell cycle progression in the presence of serum. KSR1.SA expression caused S phase entry in 85% of cells treated with PDGF and in 40% of cells treated with EGF. In comparison to control cells, EGF and PDGF had minimal ability to induce S phase entry in cells expressing KSR1.TV (Fig. 3).
To further investigate the role of KSR1 phosphorylation in mediating cell cycle progression, we assayed S-phase entry in cells expressing KSR1.TVSA. PDGF treatment of KSR1 -/-MEFs expressing KSR1.TVSA was sufficient to promote cell-cycle progression (Figs. 3 and S1). Upon expression of KSR1.TVSA, EGF treatment also induced S phase entry in 76% of cells. Thus, combined mutation of Ser 392 and Thr 274 in KSR1 conferred a PDGF-like response to cells treated with EGF. These data indicate that KSR1 phosphorylation can modulate growth factor-induced cell cycle progression by regulating the duration of ERK activation.

KSR1 phosphorylation regulates cell proliferation. To determine if KSR1
phosphorylation also regulates mitogenesis in cultured cells, we assessed proliferation rates of KSR1 -/-MEFs expressing KSR1 or KSR1 phosphorylation mutants. KSR1.SA induced a modest increase in proliferation rate compared to wild-type KSR1 (Fig. 4A). However, cells expressing KSR1.TVSA showed a two-fold increase in proliferation rate during log-phase growth compared to wild-type KSR1, and grew to an increased density (Fig. 4A). Expression of KSR1.TV alone slowed proliferation slightly in comparison to wild-type KSR1, consistent with its inhibitory effects on ERK activation. These data indicate that the phosphorylation of KSR1 on Ser 392 and Thr 274 are important regulators of a cell's proliferative potential.
KSR1 phosphorylation regulates oncogenic transformation. KSR1 -/-MEFs are resistant to oncogenic transformation by activated Ras V12 , but reintroduction of wild type KSR1 rescues the transformation defect in a dose-dependent manner (18). To determine if KSR1 phosphorylation also regulates cell transformation, we assayed anchorage-independent growth of MEFs on soft agar. When expressed in KSR1 -/-MEFs at levels comparable to those observed in wild-type MEFs, neither ectopic KSR1 nor KSR1.TVSA allowed growth in soft agar. These data suggest that the increased proliferative potential associated with KSR1.TVSA is not sufficient to induce cell transformation (Fig. 4B). We also generated MEFs expressing increasing amounts of KSR1 or KSR1.TVSA by sorting and selecting cells expressing increasing levels of GFP from the bicistronic vector. Overexpression of wild-type KSR1 at any level was not sufficient to promote cell transformation. However, moderate overexpression of KSR1.TVSA induced the anchorage-independent growth of KSR1 -/-MEFs, even in the absence of activated Ras V12 (Fig. 4B). Increasing expression of KSR1.TVSA inhibited cell transformation to basal levels, as observed previously in cells overexpressing wild type KSR1 (18). We next assayed the phosphorylation and activation of the ERK substrate p90 RSK1 by western blot. As a direct consequence of sustained ERK activation, the phosphorylation of the ERK substrate p90 RSK1 is also prolonged 120 minutes after EGF stimulation in the presence of KSR1.TVSA, compared to wild-type KSR1 (Fig. 5B). These data demonstrate that KSR1 phosphorylation, by mediating the duration of ERK activation, also regulates the activation of downstream targets of the Raf/MEK/ERK cascade. In addition to showing prolonged activation, the levels of total RSK1 are elevated in cells expressing KSR1.TVSA. Increased expression of p90 RSK1, a cyclin-responsive gene, is attributable to the elevated cyclin D1 activity associated with the increased proliferative rate of cells expressing KSR1.TV.S932A (37).
Phosphorylation of KSR1 modulates KSR1 stability. The expression level of the molecular scaffold KSR1 regulates its biological effects (18,24,30). Overexpression of KSR1 inhibits ERK activation and ERK-induced phenotypes (24,25,36). However, careful titration of  6A). Mutation of Thr 274 or Ser 392 individually was not sufficient to increase the stability of KSR1 (Fig. 6B). These data demonstrate that KSR1.TVSA is more stable than wild-type KSR1, and suggest that phosphorylation of KSR1 on Thr 274 and Ser 392 modulates KSR1 turnover and degradation. transformation. In addition, KSR1.TVSA was more stable than wild-type KSR1, suggesting that the phosphorylation of KSR1 regulates its turnover and, in turn, regulates the duration of ERK activation.

Discussion
The duration of ERK signaling is critical to generating specific biological responses to MAP kinase pathway signaling. Prolonged ERK activation promotes its sustained nuclear localization, stabilization of immediate early gene products, and cell-cycle progression (7,9,38).
To understand how KSR1 mutations that alter its phosphorylation may affect ERK activation, we studied the time course of ERK activation in cells expressing wild type or mutated KSR1. Cells expressing KSR1.SA showed robust, and somewhat prolonged, ERK activation upon treatment with EGF. Expression of KSR1.SA also induced a modest increase in proliferation rate compared to wild-type KSR1 (Fig. 4A), and modestly increased S-phase entry upon EGF stimulation (Fig. 3). This is consistent with data in Cos cells demonstrating that mutation of Ser 392 to Ala promotes MEK activation by Raf (32). Dephosphorylation of Ser 392 by PP2A is required for KSR1 release from 14-3-3 and translocation to the plasma membrane (35). Thus, KSR1.SA is already primed for membrane localization in the absence of a mitogenic signal, which explains how KSR1.SA promotes ERK activation and cell cycle progression.

Mutation of Thr 274 on KSR1 inhibited ERK activation in both intensity and duration in
response to treatment with either EGF or PDGF (Fig. 2). As a consequence, mutation of Thr 274 alone was not sufficient to promote cell proliferation (Fig. 4A) or cell cycle progression upon EGF or PDGF treatment (Fig. 3). However, the combined mutation of Ser 392 and Thr 274 induced sustained ERK activation upon treatment with either PDGF or EGF, promoted S-phase entry with either PDGF or EGF treatment (Fig. 3), and caused an increased rate of proliferation compared to cells expressing wild-type KSR1 (Fig. 4A). These data suggest that the dynamic phosphorylation and dephosphorylation of Ser 392 and Thr 274 work together to promote mitogenesis by regulating the duration of ERK activation to generate signaling specificity. The data indicate that, if Ser 392 is dephosphorylated, dephosphorylation of Thr 274 (or blocking its phosphorylation) can serve as an activating step in ERK signaling and is critical for determining the duration of ERK activation.
The kinases C-TAK and nm23-H1 both interact with and phosphorylate KSR1 on Ser 392 (32,33). The kinase that phosphorylates KSR1 on Thr 274 in intact cells is not known. However, Thr 274 resides in a consensus sequence for phosphorylation by ERKs and the MEK inhibitor PD98059 inhibits Thr 274 phosphorylation (30). The KSR1 kinase domain could autophosphorylate on Thr 274 . Though the kinase domain of mouse and human KSR1 lack evolutionarily conserved sequences that are typically required for catalytic activity (21), KSR1 has been reported to be an active kinase (39)(40)(41)(42). However, other evidence demonstrates that KSR1 function is independent of the kinase domain (26,29,32). Autophosphorylation on Thr 274 is unlikely, since KSR1 constructs that lack the entire KSR1 kinase domain are still phosphorylated on Thr 274 in immune complex kinase assays (31).
Cells expressing KSR1.TVSA are able to interpret signals induced by both PDGF and EGF binding as mitogenic. Consequently, cells expressing KSR1.TVSA may interpret more growth factor-induced signals as mitogenic than do cells expressing wild type KSR1. We believe this hyper-mitogenic signaling accounts for the increased proliferation of MEFs in culture expressing KSR1.TVSA.
A current model for KSR1 function in cells is that KSR1 brings MEK to Raf to be activated at the plasma membrane (32). Based on these observations, we hypothesized that KSR1 regulated the duration of ERK activation by regulating the duration of MEK activation. In cells expressing KSR1.TVSA, we observed that MEK phosphorylation upon EGF treatment is slightly elevated compared to cells expressing wild-type KSR1. However, phospho-MEK levels were not sustained following EGF treatment compared to wild-type KSR1 (Fig. 5A). While KSR1 may facilitate the phosphorylation of MEK by Raf, our data suggest that KSR1 phosphorylation may also regulate ERK activation independent of its effects on MEK.
The expression of KSR1 is required for the complete activation of the Raf/MEK/ERK kinase cascade. Deletion of KSR1 impairs growth factor-induced activation of ERK and prevents the transforming effects of activated Ras V12 in MEFs. Expression of ectopic KSR1 into KSR1 -/-MEFs at levels comparable to those found in wild-type MEFs restores ERK activation and Ras V12 -induced cell transformation without any effects on cell proliferation in the absence of an oncogene (18). In an experimental system, overexpression of KSR1 to levels that interact optimally with Raf, MEK and ERK is sufficient to enhance ERK activation and cell proliferation. Altered levels of KSR expression may play a role in regulating cell responsiveness. KSR1 is up-regulated during 1,25-dihydroxyvitamin D3-mediated differentiation of HL60 cells (43) and following NGF-induced differentiation of PC12 cells (26).
However, overexpression of a scaffold protein beyond a stoichiometric optimum results in the titration of its binding partners apart from each other and inhibits the biologic activity of its kinase cascade (10,44,45). These data suggest that the expression of a molecular scaffold such as KSR1 is a potential point of regulation in a cell to modulate the biochemical and biological output of a signaling cascade. Here we show that KSR1.TVSA has increased stability compared to wild-type KSR1 (Fig. 6), suggesting that phosphorylation of KSR1 at Thr 274 and Ser 392 negatively regulates KSR1 protein stability. In addition, KSR1.TVSA allowed sustained ERK activation upon EGF stimulation and consequently enhanced cell proliferation. The increased stability of KSR1.TVSA may promote sustained ERK activation by facilitating more productive interactions between phospho-MEK and inactive ERK, or by increasing the pool of KSR1 available for recruitment to a signaling complex. It is also possible that the increased pool of KSR1 binds and stabilizes phospho-ERK. KSR1 may protect ERK from dephosphorylation by MAP kinase phosphatases, either by steric regulation or by promoting ERK localization to an optimal site of action.
The mechanism for regulation of stability of KSR1 through phosphorylation has not been elucidated. While KSR1 transcript levels are high in all tissues except liver, KSR1 protein expression is only detectable in the brain, bladder, ovary, testis, and lung (46), suggesting the presence of a mechanism regulating KSR1 translation or stability. KSR1 binds the chaperone proteins Hsp90, Hsp70, and Hsp68 (26). These interactions may stabilize KSR1 protein expression. Treatment of cells overexpressing KSR1 with geldanamycin to inhibit Hsp activity increases the turnover of KSR1, either through a direct or indirect mechanism (26). It is possible that altering the phosphorylation state of KSR1 modifies its association with heat shock proteins, thereby regulating its stability. In addition, KSR1 has been shown to associate with the E3 ubiquitin ligase IMP (47), but IMP-mediated ubiquitination of KSR1 has not been detected 2 .
Phosphorylation of KSR1 on Ser 392 or Thr 274 may stimulate the turnover of KSR1 by promoting interaction with a ubiquitin ligase or by inhibiting interaction with a stabilizing protein. Based on its surrounding sequence, Thr 274 is a putative MAP kinase phosphorylation site (30). Thr 274 phosphorylation by active ERK might serve as a potential site of feedback regulation promoting the interaction of KSR1 with a destabilizing protein or dissociating KSR1 from a stabilizing protein.
Interestingly, increasing the expression level of KSR1.TVSA stimulated anchorageindependent growth in soft agar, a hallmark of cell transformation (Fig. 4B). It has been shown previously that KSR1 is required for transformation induced by expression of Ras V12 (18,48), but expression of wild-type KSR1 at any level is not sufficient to induce cell transformation on its own (Fig. 4B). Moderate overexpression of KSR1.TVSA promoted cell transformation, even in the absence of an oncogene. These data demonstrate that KSR1 phosphorylation is a potent regulator of a cell's proliferative and oncogenic potential. Moreover, the data suggest that anchorage-independent growth may require a sustained level of signal output from the Raf/MEK/ERK signaling cassette that can be attained by the combination of prolonged ERK activation provided by KSR1.TVSA with elevated levels of scaffold expression. The ability of KSR1 to function as an oncogene is dependent upon a defined level of expression. KSR1.TVSA fails to support anchorage-independent growth when expressed at high levels. Similarly, over expression of wild type KSR1 at comparable levels suppresses colony formation induced by oncogenic Ras V12 (18). This concentration-inhibitory effect of wild type KSR1 to inhibit Ras V12 -induced cell transformation occurs at levels of the scaffold that exceed those required for optimal interactions with Raf, MEK and ERK (18). Thus, the transforming potential of KSR1.TVSA remains subject to, and limited by, its function as a scaffold.
A GFP fusion protein of KSR1 cycles through the nucleus in a phosphorylation dependent manner (34). GFP-KSR1 is primarily cytoplasmic in quiescent cells. However, mutation of phosphorylation sites Thr 274 and Ser 392 causes the redistribution of GFP-KSR1 to the nucleus. The biological significance of KSR1's nuclear localization and the regulation of its subcellular trafficking remain to be characterized. However, data shown here suggest that nuclear localization of KSR1 may be a function of KSR1 stability and the duration of activity by the Raf/MEK/ERK kinase cascade.
The molecular scaffold KSR1 is a dynamic effector of signaling through the Raf/MEK/ERK kinase cascade. Its function is modified by its phosphorylation state and is subject to input from multiple pathways. These observations suggest that KSR1 may serve as an excellent control point for receiving input from intracellular mechanisms that positively or negatively regulate mitogenic potential and other aspects of cell fate affected by the Raf/MEK/ERK kinase cascade. constructs were treated with cycloheximide in serum-free medium for the indicated times. Cell lysates were analyzed by western blot using an anti-KSR1 antibody. KSR1 levels were